60^th EASD Annual Meeting of the European Association for the Study of Diabetes

Abstracts

Index of Oral Presentations

OP 01 New mechanisms and outcomes of treatments for diabetic kidney disease

OP 02 Which day of the week is your insulin day?

OP 03 Liver: the master regulator

OP 04 Beyond glucose in diabetes and prediabetes

OP 05 Can you change what you are by what you eat?

OP 06 Maternal influences and prediction in type 1 diabetes

OP 07 Cardiovascular complications in type 1 diabetes: burden and mechanisms

OP 08 Genetics and epigenetics in type 2 diabetes

OP 09 Drugs you thought you knew all about, but what else do they do?

OP 10 Hypothalamic regulation of systems metabolism

OP 11 Risk factors and treatment opportunities in neuropathy and hypertension

OP 12 Fatal attraction: immune cells and beta cells

OP 13 GLP-1-based polypharmacology: we are not done yet!

OP 14 Incretins: How many targets do you need?

OP 15 Cardiovascular risk in patients with type 1 diabetes: What do we know today, how can we modify it tomorrow?

OP 16 Emerging risk factors and treatment options for diabetic eye disease

OP 17 Type 2 diabetes prediction

OP 18 How to blow the low

OP 19 Prediction and prognosis of type 1 diabetes

OP 20 Are you afrAID?

OP 21 Cardiovascular risk in diabetes: in search of the holy grail

OP 22 Actions of incretins on the brain and pancreas you might not know about

OP 23 Why exercise?

OP 24 Islets forever: survival of the beta cells

OP 25 Incretin receptor agonists: better and better

OP 26 Love goes through the stomach: new roles for GLP-1

OP 27 Predictors of diabetic kidney disease progression from omics to ethnicity

OP 28 New kids on the block: stem cell islets

OP 29 More on hormones

OP 30 Causes and consequences of gestational diabetes

OP 31 Macrovascular outcomes of diabetes: a holistic view

OP 32 Fine tuning CGMing

OP 33 Emerging science in diabetic kidney disease from fasting to autophagy

OP 34 [B]reaking [F]at

OP 35 Insulin: ready, steady, go!

OP 36 Inside information on beta cell regulation

OP 37 Protection in islet transplantation

OP 38 New insights into CNS regulation of energy and glucose metabolism

OP 39 Novel insights in monogenic diabetes

OP 40 Around liver fat: the fight is on

OP 41 Turning back the biological clock - or merely slowing it?

OP 42 New factors in the control of glucose tolerance

OP 43 Treating type 1 diabetes: What else does the future hold?

OP 44 The unforgettable alpha cells

OP 45 How smart an insulin pen can be?

OP 46 Pregnancy: taking care of the future

OP 47 Inflammatory mechanisms regulating systems metabolism

OP 48 Managing diabetic foot disease: clinical studies and omics predictors

Index of Short Oral Discussions

SO 001 Genetic insights into type 2 diabetes: from risk variants to personalized treatment strategies

SO 002 Epigenetics and gestational diabetes: from maternal health to newborn outcomes

SO 003 The many facets of weight and type 2 diabetes risk

SO 004 Diet, lifestyle, and metabolic health: insights into diabetes risk and prevention

SO 005 Unraveling monogenic diabetes: from genetic variants to innovative diagnostic tools

SO 006 Advancements in predicting type 1 diabetes

SO 007 Global perspectives on type 2 diabetes: prevalence, risk factors, and predictive tools

SO 008 Advances type 1 diabetes: genetic, biomarker, and machine learning approaches

SO 009 Treatment of type 1 diabetes

SO 010 Complications of type 1 diabetes

SO 011 Disease progression and mortality in diabetes

SO 012 Health interactions and risks in diabetes: from cancer survival to chronic conditions

SO 013 Prevention of type 2 diabetes

SO 014 Early factors and family determinants of diabetes

SO 015 Genetics and biomarkers for prediction of type 2 diabetes

SO 016 Molecular predictors of type 1 diabetes

SO 017 Long-term complications and outcomes in type 2 diabetes: vascular health, kidney function, and mortality

SO 018 Diabetes management and outcomes: societal, behavioral, and clinical influences

SO 019 Catching signals to follow the right pathway in islets

SO 020 Holding hands with alpha cells in the islets

SO 021 Insulin service providers for the body

SO 022 Cells in attacking mode and beta cell victims

SO 023 Keeping islet cell identity under different circumstances

SO 024 Improving islet transplantation

SO 025 Breaking bad: destruction of beta cells

SO 026 Pregnancy and type 1 diabetes

SO 027 Gestational diabetes: clinical perspective

SO 028 Biomarkers and mechanisms in gestational diabetes

SO 029 On the young ones...

SO 030 Investigating impact of gestational diabetes

SO 031 All the reasons why exercise is good!

SO 032 Insulin action in peripheral tissues

SO 033 Understanding insulin sensitivity: What can we learn from human studies?

SO 034 Novel mechanisms of insulin resistance

SO 035 Non-insulin hormones

SO 036 Keeping your beta cells alive

SO 037 New approaches in carbohydrate metabolism

SO 038 Novel pharmacological approaches for the treatment of obesity and diabetes

SO 039 Improving metabolic outcome by bariatric surgery

SO 040 Now the brown fat talks

SO 041 Towards a better understanding of the adipose tissue - liver axis

SO 042 How lifestyle affects our health

SO 043 Inflammation and beta cell homeostasis

SO 044 Lipid signatures and ectopic fat

SO 045 Preclinical advances to better understand energy, glucose and lipid metabolism

SO 046 What’s new on white fat function?

SO 047 Novel mechanisms regulating energy and glucose metabolism

SO 048 Novel mechanisms controlling glucometabolic health

SO 049 Novel clinical insights into adipose tissue function

SO 050 Understanding and combating adipose tissue dysfunction

SO 051 Metabolic alterations in obesity and diabetes-linked co-morbidities

SO 052 Preclinical advances to understand glucose metabolism

SO 053 Oral agents to control glucose metabolism

SO 054 Improving metabolism via the gut

SO 055 Improving metabolism through nutrition

SO 056 Thinking outside the box to treat obesity and diabetes

SO 057 Novel therapeutic avenues for metabolic disease

SO 058 SGLT2 inhibition: focus on the heart and the kidneys

SO 059 Pleiotropic effects of SGLT2 inhibition

SO 060 Novel incretins: it is all about clinical outcomes

SO 061 Novel incretins: focus on kidney, liver and cardiovascular system

SO 062 Hot of the press novel incretins

SO 063 More combo’s less insulin!

SO 064 Incretins have more effects than you can imagine

SO 065 Advanced real life in type 1 diabetes

SO 066 Pumping without bumping

SO 067 Breaking away with tradition or how to give insulin differently

SO 068 Playing with food intake and energy expenditure

SO 069 HbA1c vs CGM: a tie break?

SO 070 De-combining CGM and insulin in type 2 diabetes

SO 071 How universal can CGM be?

SO 072 The future of diabetes science is here

SO 073 Diabetes technology - but at a higher level

SO 074 It is very hard to beat the system - or is it?

SO 075 A lesson in diabetes geography

SO 076 Up or down and back

SO 077 Continuous monitoring of CGM

SO 078 More stress on distress

SO 079 From microbiology to MRI in diabetic foot disease

SO 080 Diabetic kidney disease: clinical insights from childhood to frailty

SO 081 Advances in diabetic kidney disease from cells to histology

SO 082 Predicting and treating renal complications. What’s new in 2024?

SO 083 The four M’s of microvascular complications: MRI, Mitochondria, Models, and iMaging

SO 084 Mechanisms and treatment for microvascular complications; nerves, eyes, and blood vessels

SO 085 Clinical data and emerging markers in diabetic foot disease and neuropathy

SO 086 Novel markers and predictors for neuropathy in diabetes

SO 087 Clinical studies on the crosstalk between neuropathy and other systems

SO 088 Fatty liver in diabetes: pathways and relationships

SO 089 Brain matters in diabetes

SO 090 Skeletal muscles and subcutaneous adipose tissue: important players in diabetes management

SO 091 A new look at well-known drugs in type 2 diabetes

SO 092 Divination from a glass ball or something more? Predictive models, biomarkers, genome analysis

SO 093 A non-classical look at type 1 diabetes

SO 094 Cohorts, trials and cardiovascular disease

SO 095 Other complications of type 2 diabetes

SO 096 A non-traditional look at type 2 diabetes

SO 097 Pathogenesis of diabetes complications

SO 098 Cancer and diabetes

SO 099 A non-traditional look at type 1 diabetes

SO 100 Correlates of fatty liver disease in diabetes

SO 101 MASLD: now an established complication of diabetes

SO 102 Addressing risk and consequences of liver pathology in diabetes

Late-breaking Abstracts

Index of LBA Oral Presentations

LBA OP 01 Advancements in diabetes management and cardiovascular health

LBA OP 02 Exercise, metabolism and hormonal responses in diabetes

LBA OP 03 Incretin-based therapies and metabolic insights in diabetes

LBA OP 04 Diabetes management: protecting kidney function and preventing retinopathy

LBA OP 05 Diabetes, mental health and neuroprotection: emerging evidence and mechanisms

LBA OP 06 Transformative islet therapies and genetic discoveries

LBA OP 07 Emerging technologies in diabetes: from closed loop systems to non-invasive screening

Index of LBA Short Oral Discussions

LBA SO 01 Ailing heart, broken vessels

LBA SO 02 How NOT to get from fat to sweet

LBA SO 03 One size down, three (or more) to go

LBA SO 04 Never leave the liver (especially the fatty one)!

LBA SO 05 Incretins up and roaring

OP 01 New mechanisms and outcomes of treatments for diabetic kidney disease

1

Comparative renal outcomes of matched cohorts of patients with type 2 diabetes receiving SGLT-2 inhibitors or GLP-1 receptor agonists under routine care: the DARWIN-Renal study

G. Fadini ¹, E. Longato¹, M. Morieri², S. Del Prato³, A. Avogaro⁴, A. Solini⁵;

¹University of Padova, Padova, Italy, ²University Hospital of Padua, Padova, Italy, ³Sant’Anna School of Advanced Studies, Pisa, Italy, ⁴Department of Medicine, University of Padova, Padova, Italy, ⁵University of Pisa, Pisa, Italy.

Background and aims: Chronic kidney disease (CKD) in diabetes represents a major unmet need, a primary cause of end-stage kidney disease, and a powerful cardiovascular risk factor. Here, we compared the effects of sodium-glucose cotransporter 2 inhibitors (SGLT2i) and glucagon-like peptide-1 receptor agonists (GLP-1RA) on renal outcomes in patients with type 2 diabetes mellitus (T2D), focusing on changes in estimated glomerular filtration rate (eGFR), kidney function loss, and albuminuria.

Materials and methods: This was a multicentre retrospective observational study conducted at 50 diabetes outpatient clinics using electronic health records from an initial cohort of 48,593 patients. Patient characteristics were assessed before and after propensity score matching. The primary endpoint, change in eGFR, was analysed using mixed-effects models. Secondary endpoints included loss of kidney function, changes in albuminuria, HbA1c, body weight and blood pressure. Subgroup and sensitivity analyses were performed to assess the robustness of the findings, especially in patients without CKD at baseline.

Results: After matching, 5,701 patients were included in each group. Patients were predominantly male, aged 61 years, with a 10-year duration of diabetes, a baseline HbA1c of 8.0%, and a mean BMI of 33 kg/m². CKD was present in 23% of patients, defined as a reduced eGFR or a raised albuminuria. The newly initiated SGLT2i were dapagliflozin (53%), empagliflozin (39%), canagliflozin (8%), ertugliflozin (<0.1%). The newly initiated GLP-1RA were dulaglutide (52%), liraglutide (31%), exenatide (11%), semaglutide (4%), lixisenatide (2%). eGFR decline was significantly lower in the SGLT2i group compared to the GLP-1RA group throughout the observation period (median 2.2 years) with a difference of 1.19 ml/min/1.73 m². (95% C.I. 0.47-1.90; p=0.001). The rate of creatinine doubling was lower in the SGLT2i than in the GLP-1RA group (HR 0.64; 95% C.I. 0.41-0.99; p=0.047). The change in albuminuria was not different between groups, as was the rate of new-onset macroalbuminuria. Despite a 0.2% lesser HbA1c decline, SGLT2i demonstrated favourable changes in body weight (-1.3 kg) and systolic blood pressure (-1.1 mm Hg) compared to GLP-1RA. Results were robust in sensitivity analyses. Superiority of SGLT2i in preserving eGFR was confirmed in patients without baseline CKD (n=4416/group).

Conclusion: In patients with T2D, treatment with SGLT2i was associated with better renal outcomes compared to GLP-1RA, as evidenced by slower decline in eGFR and reduced risk of kidney function loss, even in primary prevention. These findings suggest the potential of SGLT2i as preferred agents for renal protection in T2D over GLP-1RA.

Supported by: the Italian Diabetes Society and AstraZeneca

Disclosure: G. Fadini: Employment/Consultancy; Lilly, Novo Nordisk. Lecture/other fees; Abbott, AstraZeneca, Boehringer, Lilly, Novo Nordisk, MSD, Mundipharma, Sanofi, Servier, Takeda.

2

Extracellular vesicles as messengers: how canagliflozin improves diabetic podocytes ferroptosis

G. Qin, X. Fan, Y. Song, F. Guo, S. Ma, M. Shao, W. Zhang, Y. Luo, F. Wei, W. Guo, Y. Zhao;

Endocrinology and metabolism, Henan Province Zhengzhou, Zhengzhou, China.

Background and aims: Ferroptosis, a lipid peroxidation-driven form of cell death dependent on iron, has been implicated in the pathogenesis of podocyte injury in diabetic kidney disease (DKD). Canagliflozin (CANA) has been shown to preserve renal podocyte function in a glucose-independent manner, although the mechanism remains unclear. Extracellular vesicles (EVs), nanovesicles derived from cells, are crucial mediators of intercellular communication. The current study explored the mechanisms by which extracellular vesicles derived from proximal tubular epithelial cells (PTECs) mediated by CANA enhance the resistance of diabetic podocytes to ferroptosis.

Materials and methods: The study involved co-culturing human proximal renal tubular epithelial cell line (HK-2) with human renal podocyte cell line (HPC) and administering CANA as an intervention. EVs were isolated from the culture medium of HK-2 cells using ultracentrifugation. Forty 8-week-old male db/db mice were randomly assigned to one of four groups: DKD, low-dose CANA, high-dose CANA, and GW4869; ten 8-week-old male db/m mice were included as the normal control group. The intervention was maintained for 12 weeks. Integrate GEO data sets with bioinformatics to analyze the candidate molecule influencing the secretion of EVs by PTECs following treatment with CANA. Utilize dual-luciferase reporter and ChIP assays to validate the interaction between FOXO1 and VPS53. Conduct 4D-label free proteomics and high-resolution non-targeted absolute quantitative lipidomic analyses on EVs isolated from HK-2 cells. Confirm the interaction between PCBP1 and ACSL4 through dual-luciferase reporter and RIP assays.

Results: Canagliflozin (10 μM) increased the number, but not the size, of EVs secreted by HK-2 cells in a high-glucose environment (CH-EVs), and the vesicles could be taken up by podocytes. CANA reduced blood glucose, creatinine and urine albumin creatinine ratio (UACR) levels in db/db mice (P < 0.05), improved renal pathological changes, and down-regulated the expression levels of GPX4 and SLC7A11. However, UACR failed to be significantly improved after CANA combined with GW4869. Enhancement of FOXO1 transcriptional activity by CANA resulted in increased secretion of EVs by HK-2 cells. CANA intervention or FOXO1 overexpression up-regulated VPS53 expression, indicating a direct binding relationship between FOXO1 and VPS53. Proteomic analysis revealed that CH-EVs were enriched in pathways related to ferroptosis. Lipidomic analysis suggested that CANA may modulate podocytes lipid homeostasis by altering the lipid composition of EVs. The downregulation of PCBP1 expression was observed in HPC cells cultured in high glucose conditions, but was subsequently upregulated following treatment with CH-EVs. Furthermore, the decreased expression of PCBP1 in HPC cells was found to elevate ACSL4 levels and enhance ferroptosis in these cells. PCBP1 was shown to directly interact with ACSL4, influencing the stability of the latter’s mRNA.

Conclusion: Canagliflozin enhances the generation and secretion of PTECs-derived EVs through the up-regulation of the FOXO1-VPS53 axis. Additionally, it suppresses podocyte ACSL4 mRNA stability and improves the condition of diabetic podocytes by modulating the EVs content of PCBP1 to mitigate ferroptosis.

Supported by: the National Natural Science Foundation of China [Grant numbers 81974110 and 82170839]

Disclosure: G. Qin: Grants; 81974110, 82170839.

3

Effects of semaglutide, empagliflozin and their combination on renal diffusion-weighted magnetic resonance imaging and total kidney volume in patients with type 2 diabetes

L. Vernstrøm ¹, S. Gullaksen¹, S.S. Sørensen¹, S. Ringgaard², C. Laustsen², H. Birn¹, K. Funck¹, E. Laugesen¹, P.L. Poulsen¹;

¹Aarhus University Hospital, Aarhus, Denmark, ²Aarhus University, Aarhus, Denmark.

Background and aims: Glucagon-like peptide-1 receptor agonists (GLP-1RAs) and sodium-glucose cotransporter 2 inhibitors (SGLT2Is) both provide protection against progression of kidney disease in diabetes. However, the mechanisms are incompletely understood. The apparent diffusion coefficient (ADC) derived from diffusion-weighted magnetic resonance imaging (DWI-MRI) is a measure of water diffusion and has been proposed as a measure of kidney microstructure including kidney fibrosis. In advanced kidney disease, the kidneys often become atrophic, however, in the initial phase of type 2 diabetes, there is an increase in renal size. In this study, we examined the effect of semaglutide, empagliflozin and their combination on renal ADC and total kidney volume (TKV).

Materials and methods: This was a substudy of a randomized clinical trial on the effects of semaglutide and empagliflozin alone or in combination. Eighty patients with type 2 diabetes and high risk of cardiovascular disease were randomised into four groups (n=20 in each) receiving either placebo, semaglutide, empagliflozin, or their combination for 32 weeks. We analysed the effects of treatment on changes in ADC (cortical, medullary, and ΔADC (medullary ADC subtracted from cortical ADC)), as well as TKV measured by MRI.

Results: Both semaglutide and empagliflozin decreased cortical ADC significantly compared with placebo (semaglutide -0.20 x 10^-3 mm²/s (95%CI -0.30;-0.10), p<0.001, empagliflozin -0.15 x 10^-3 mm²/s (95%CI -0.26;-0.04), p=0.01). No significant change was observed in the combination group (-0.05 x 10^-3 mm²/s (95%CI -0.15;0.05), p=0.29 compared to placebo). The changes in cortical ADC were not associated with changes in GFR, albuminuria, TKV, or markers of inflammation. Further, there were no changes in medullary ADC in any of the groups compared to placebo. Only treatment with semaglutide changed ΔADC significantly from placebo (-0.13 x 10^-3 mm²/s (95%CI -0.22;-0.04), p=0.01). Compared to placebo, TKV decreased by -3 % (95%CI -5;-0.3), p=0.04, -3 % (95%CI -5;-0.4), p=0.02 and -5 % (95%CI -8;-2), p<0.001 in the semaglutide, empagliflozin and the combination group respectively. The reductions in TKV were associated with reductions in GFR, albuminuria and HbA1c.

Conclusion: In a population with type 2 diabetes, semaglutide and empagliflozin significantly reduced cortical ADC compared to placebo, indicating microstructural changes in the kidneys. These changes were not associated with changes in GFR, albuminuria, or inflammation. Further, we found a decrease in total kidney volume in all active treatment groups possibly mediated by a reduction in hyperfiltration. Our findings suggest that DWI-MRI may serve as a promising tool for investigating the underlying mechanisms of medical interventions in patients with diabetes but may reflect effects not related to fibrosis.

Clinical Trial Registration Number: EUDRACT 2019-000781-38

Supported by: The Novo Nordisk Foundation (grant no. NNF170C0029064), the Central Denmark Region Research Fund (grant no. A3192).

Disclosure: L. Vernstrøm: Grants; the Novo Nordisk Foundation (grant no. NNF170C0029064), the Central Denmark Region Research Fund (grant no. A3192), the Danish Medical Associations Research Foundation (grant no. 2019-3780/41).

4

Gastric bypass and sleeve gastrectomy improve albuminuria more than iSGLT-2/GLP-1RA based medical therapy: a randomised clinical trial

J.P. Valderas ¹, C.A. Boza², C. le Roux³;

¹Department of Medical Sciences, Universidad de Antofagasta, Antofagasta, Chile, ²Clinica Meds, Santiago, Chile, ³University College of Dublin, Dublin, Ireland.

Background and aims: The ADA/EASD consensus Management of Hyperglycemia in Type 2 Diabetes (T2D), recommended the use of SGLT-2 inhibitors (iSGLT-2) and GLP-1 receptor agonist (GLP-1RA) in patients with Chronic Kidney Disease (CKD). Bariatric Surgery, especially Gastric Bypass (GB) have shown to reduce albuminuria. Until now, GB and Sleeve Gastrectomy (SG) have not been compared with optimal medical therapy.Aim:To compare the effect of GB, SG and the combination of iSGLT-2 and GLP-1RA treatment (MED) in patients with CKD, T2D, and grade I obesity.

Materials and methods: PRODIGIES is a prospective randomized controlled trial of medical versus surgical treatment of CKD (defined by urine albumin-creatinine ratio [uACR]> 30 mg/g) and T2D in patients with BMI between 30 to 35kg/m2.

Results: 24 patients were randomized to GB, 24 to SG and 23 to MED. At baseline there were not differences between groups in age (years , 51.9 ± 8.9 in GB; 51.5 ± 7.9 in SG; 54.9 ± 6.5 in MED) BMI ( kg/m², 32.2 ± 2.6 in GB; 31.6 ± 2.6 in SG; 31.3 ±2.8 in MED) and HbA1c (%, 9.0 ± 1.4 in GB; 9.3 ± 1.2; 9.1 ± 2.1 in MED. Baseline uACR (median and IQR, mg/g) and eGFR (mean ± SD, mL/min/1.73m² ) was 96.0 (47.1-149.8) and 95.7 ± 15.0 in GB; 59.1 (45.0-142.4) and 95.4 ± 21.0 in SG, and 98.0 (61.3-204.1) and 93.4 ± 15.8 in MED, without differences between groups. After 12 months, 6/24 patients in MED were on GLP-1RA alone, 1/24 on iSGLT-2 alone, and 16/24 on the combination of GLP-1 and SGLT2i. 4/24 patients in SG used iSGLT-2 and no surgical patients used GLP-1RA. Remission of albuminuria (uACR < 30 mg/g) occurred in 75% of patients after GB, 66,7% after SG, and 34,8% after MED (p=0.03). uACR reduction was 97.1 ± 128 in GB ( p < 0.0001) 80.3 ± 132 in SG (p < 0.0001) and 20.2 ± 137,mg/g in MED (p=0.2). Weight loss (kg) was greater in both surgical groups compared to MED, with no difference between GB and SG (19.9 ± 6.0 in GB, p < 0.0001; 16.7 ± 5.4 in SG, p < 0.0001; 4.3 ± 3.9 in MED, p= 0.3). HbA1c reduction (%) was similar between the three groups (2.5 ±1.7 in GB; 2.0 ±1.9 in SG; 2.1 ±1.8 in MED; p < 0.0001 for all groups). Blood pressure changes were similar. There were 3 surgical complications: 1 leak and sepsis in GB, 1 bowel perforation and 1 mesenteric thrombosis in SG. All patients underwent a full recovery. No serious adverse effects occurred in MED.

Conclusion: After 12 months GB and SG was superior at inducing remission of albuminuria, in patient with CKD, T2D and class I obesity. Magnitude of weight loss and not changes in HbA1c or blood pressures were associated with improvements in albuminuria.

Clinical Trial Registration Number: NCT01974544

Supported by: FONDECYT Project N° 1120877 and Johnson & Johnson

Disclosure: J.P. Valderas: Grants; Johnson & Johnson, FONDECYT Project N° 1120877.

5

SGLT2-inhibitor treatment for nephroprotection can be targeted using eGFR / albuminuria categories and risk scores: precision medicine in type 2 diabetes

T. Jansz ¹, K. Young², R. Hopkins³, B. Shields⁴, A.T. Hattersley⁴, R. Oram⁵, J. Dennis¹;

¹University of Exeter, Exeter, UK, ²College of Medicine and Health, University of Exeter, Exeter, UK, ³Institute of Biomedical & Clinical Science Research, University of Exeter, Exeter, UK, ⁴University of Exeter Medical School, Exeter, UK, ⁵Institute of Biomedical and Clinical Science, University of Exeter Medical School, Exeter, UK.

Background and aims: Despite current guidelines recommending SGLT2-inhibitors for nephroprotection to a large proportion of people with type 2 diabetes, the absolute benefit of these therapies has not been clearly established for many with early-stage chronic kidney disease (CKD). We therefore combined data from trial meta-analysis and internationally validated kidney disease progression risk scores to identify subgroups of people with type 2 diabetes who would have significant absolute benefit for nephroprotection if treated with SGLT2-inhibitors.

Materials and methods: We studied 134,420 adults with type 2 diabetes and eGFR ≥20mL/min, without pre-existing cardiovascular disease or heart failure, who started SGLT2-inhibitors (34%), DPP4-inhibitors (41%), or sulfonylureas (25%) in UK primary care (Clinical Practice Research Datalink [CPRD], 2013-2020). We validated the previously published relative risk reduction for kidney disease progression (sustained ≥40% decline in eGFR, end-stage kidney disease, or death from renal causes) from SGLT2-inhibitor trial meta-analysis. We then estimated individual-level absolute risk of kidney disease progression using CKD Prognosis Consortium (CKD-PC) risk scores. Combining the relative risk reduction and absolute risk predictions, we estimated the 3-year number needed to treat (NNT) with SGLT2-inhibitors in subgroups defined by eGFR and urinary albumin/creatinine ratio (uACR). We assessed the implications of current guideline recommendations in a contemporary population-representative UK cohort of people with type 2 diabetes (CPRD, n=388,115).

Results: SGLT2-inhibitors had an adjusted HR for kidney disease progression of 0.62 (95%CI 0.51-0.76) compared with DPP4-inhibitors/sulfonylureas, which was similar to previously published trial meta-analyses (0.62, 95%CI 0.56-0.68) and consistent across eGFR/uACR subgroups (interaction p 0.86 and 0.48, respectively). CKD-PC risk scores predicted kidney disease progression well after recalibration (C-statistic 0.766). The NNT was low for people with a uACR ≥30mg/mmol and any eGFR (NNT=28); an eGFR <60mL/min and uACR 3-30mg/mmol (NNT=28); and an eGFR <60mL/min and uACR <3mg/mmol (NNT=60) (8% of the study cohort). The NNT was 85 in people with an eGFR ≥60mL/min, uACR 3-30mg/mmol, and CKD-PC risk score ≥90th percentile (8% of the study cohort). The NNT was high for other people with an eGFR ≥60mL/min and uACR 3-30mg/mmol (NNT=271); and for people with an eGFR ≥60mL/min and uACR <3mg/mmol (NNT=478) (84% of the study cohort). This former group represents 46% of people with type 2 diabetes that current KDIGO and ADA/EASD guidelines recommend treating with SGLT2-inhibitors for nephroprotection .

Conclusion: SGLT2-inhibitor nephroprotective benefit is minimal (NNT>250) in almost half of people with type 2 diabetes that current guidelines recommend treating. Guidelines should take subgroup-specific NNTs into account for treatment recommendations.

Supported by: MRC grant (MASTERMIND consortium)

Disclosure: T. Jansz: None.

6

Effect of empagliflozin on urinary albumin excretion and hypoxic injury biomarkers of early phase diabetic kidney disease (EUROBEAT study)

H. Makino ¹, M. Kasahara², S. Kasama², N. Ozu², R. Takashima², M. Noguchi¹, H. Park³, Q. Chen⁴, K. Tsuruya², K. Hosoda¹;

¹National Cerebral and Cardiovascular Center, Suita, Japan, ²Nara Medical University, Kashihara, Japan, ³Kyowakai Hospital, Suita, Japan, ⁴Takatsuki Hospital, Takatsuki, Japan.

Background and aims: Growing evidence showed sodium glucose cotransporter 2 (SGLT2) inhibitor could prevent the progression of diabetic kidney disease (DKD). However, the reno-protective effects of SGLT2 inhibitor in DKD with microalbuminuria has not yet been fully proven. Furthermore, the precise mechanisms of SGLT2 inhibitor on reno-protective effect has been still unclear. In this study, we hypothesized that SGLT2 inhibitor prevents DKD with microalbuminuria via reduction of renal hypoxia-induced angiogenic factor overproduction.

Materials and methods: In this multicenter, prospective, randomized, double blinded clinical trial, people with type2 diabetes with microalbuminuria (urinary albumin creatinine ratio (ACR): 30-299 mg/gCr) with estimated glomerular filtration rate (eGFR) > 30 ml/min/1.73m² were randomized equally to empagliflozin (10 mg/day) (n = 40), or placebo (n = 39). The primary endpoint was change in ACR and urinary liver type fatty acid binding protein (LFABP) excretion from baseline to 24 weeks. Major secondary outcome was change in serum vascular endothelial growth factor (VEGF), angiopoietin like protein 2 (ANGPTL2), angiopoietin like protein 4 (ANGPTL4), and adrenomedullin levels from baseline to 24 weeks.

Results: Although log ACR significantly decreased in the empagliflozin group at 4 and 12 weeks, the difference of change at 24 weeks between two groups was not statistically significant (difference between the groups: -0.364 mg/gCr, 95% CI: -0.757 to 0.029, p=0.0686) (Figure 1A). In participants with eGFR < 60 ml/min/1.73m², the reduction in log ACR was significantly greater in empagliflozin group than placebo group (difference between the groups; -0.747 mg/gCr, 95%CI: -1.497 to 0.003), while in participants with GFR > 60 ml/min/1.73m², the reduction in log ACR tended to be greater in empagliflozin group than placebo group, but not statistically significant (Figure 1B). There was no difference in urinary LFABP excretion between empagliflozin and placebo groups. Among hypoxia induced factors, serum VEGF and ANGPTL2 decreased significantly more in the empagliflozin group (difference between the groups; -39.668 pg/mL, 95%CI: -76.088 to -3.249, p = 0.0332, difference between the groups; -0.532 ng/mL, 95%CI: -1.023 to -0.040, p = 0.0344, respectively), whereas there were no significant differences in adrenomedullin and ANGPTL4.

Conclusion: This study demonstrated that empagliflozin treatment decreased ACR in DKD with microalbuminuria. This study also demonstrated that empagliflozin treatment suppressed the VEGF and ANGPTL2, which might contribute to the mechanisms of reno-protective effects of this agent.

Clinical Trial Registration Number: jRCTs051200147

Supported by: This study was supported by Boehringer Ingelheim and Eli Lilly and Company.

Disclosure: H. Makino: Grants; Boehringer Ingelheim, Eli Lilly.

OP 02 Which day of the week is your insulin day?

7

Safety, tolerability, and pharmacodynamics properties of a novel once-weekly insulin GZR4 in healthy subjects

W. Chen ¹, L. Wan², M. Zhu², C. Hao¹, J. Zhao¹, A. He¹, T. Xie¹, Y. Li¹, Z.-R. Gan¹, C. Tang²;

¹Gan & Lee Pharmaceuticals, Beijing, China, ²Bishan Hospital of Chongqing Medical University, Chongqing, China.

Background and aims: GZR4, an innovative ultra-long-acting insulin analog designed for single once-weekly subcutaneous administration, has demonstrated significant antidiabetic bioactivity in preclinical investigations. The objective of the present study was to assess the safety, tolerability, pharmacokinetics (PK), and pharmacodynamics (PD) of a single ascending dose of GZR4 in healthy Chinese participants.

Materials and methods: In this randomized, double-blind, placebo-controlled, sequential, dose-escalation phase Ia study, healthy male adults aged 18-45 (BMI 19-24 kg/m²) were randomized into five cohorts. In cohorts 1-4, participants were randomized in a 4:1 ratio to once-weekly subcutaneous injection of GZR4 (doses: 1 ,3 ,6 ,12 nmol/kg) or placebo, and cohort 5 administered with 0.4 U/kg of insulin degludec (IDeg) as a comparator. A 24-hour glucose clamp procedure was executed on day 2 and day 7 respectively post-GZR4 administration in cohorts 2-4, and on day 1 following IDeg administration. The primary endpoints were safety and tolerability of GZR4. Blood samples were collected before and after administration of GZR4 for PK and PD analyses.

Results: Forty-three participants were enrolled, with no instance of mortality, serious adverse events (SAEs), or discontinuations related to the investigational products. The most common treatment-emergent adverse events (TEAEs) were hypoglycemia, which occurred only in subjects receiving the 12 nmol/kg dose of GZR4, and all TEAEs were mild to moderate in severity. The maximum plasma concentrations (C_max) of GZR4 increased dose-dependently from 43.15 ± 7.42 ng/mL to 535.75 ± 107.72 ng/mL, while the area under the concentration-time curve (AUC) from time zero to the last measurable concentration (AUC_0-last) ranged from 5535.8 ± 222.9 to 70134.2 ± 10856.2 h*ng/mL across cohorts 1 through 4. The glucose infusion rate (GIR) indicated that the glucose-lowering effect of GZR4 persists for approximately one week, with the area under the GIR-time curve of day 7 (AUC_GIR,144-168h) accounts for about 80% of day 2 (AUC_GIR,24-48h) in the cohort receiving a 12 nmol/kg dose of GZR4. Model-predicted daily distribution of the glucose-lowering effect of GZR4 evenly within 1-week dosing interval. Meanwhile, the AUC_GIR,144-168h for the 6 nmol/kg dose of GZR4 was closely comparable to the AUC_GIR,0-24h for 0.4 U/kg of IDeg (37.84 ± 9.86 versus 40.589 ± 14.39 h*mg/kg/min, respectively), implying that the average daily potency of GZR4 is nearly equivalent to that of a single dose of IDeg (1.14-fold), which suggests that the potency of GZR4 is approximately 2.5-fold greater than that of IDeg when evaluating based on weekly dosage.

Conclusion: GZR4 exhibits good tolerability in healthy subjects and a superior glucose-lowering effect compared to IDeg. The safety, PK, and PD profiles of GZR4 support further evaluation of once-weekly dosing regimens for glycemic management in patients with diabetes.

Clinical Trial Registration Number: CTR20222273

Disclosure: W. Chen: None.

8

Effects of insulin Efsitora alfa on frequency and severity of hypoglycaemia under conditions of increased hypo risk compared to glargine in type 2 diabetes

T. Heise ¹, G. Andersen¹, E.J. Pratt², J. Leohr², T. Fukuda², Q. Wang², C. Kazda², J. Bue-Valleskey², R. Bergenstal³;

¹Profil, Neuss, Germany, ²Eli Lilly and Company, Indianapolis, IN, USA, ³International Diabetes Center Park Nicollet, Minneapolis, MN, USA.

Background and aims: Efsitora is an insulin receptor agonist designed to have a flat pharmacokinetic profile and long half-life enabling weekly dosing. While these features may provide stable glucose levels, their impact on hypo risk is less clear. A phase 1 study was conducted to assess hypo risk using controlled, experimental conditions that mimic hypo risk situations that may be encountered in daily life.

Materials and methods: This single-site, open-label, 2-period, fixed-sequence (glargine-efsitora) study was conducted in 54 participants with T2D previously on basal insulin (BMI 21.8-39.7 kg/m², HbA1c 6.5 -9.4% [48-79 mmol/mol]). After titration to stable fasting glucose (FG) with glargine or efsitora, the incidence of hypo was assessed during 3 test conditions: prolonged fasting 24-hrs (PF), PF with exercise (EX), and after receiving a double dose (DD) of study insulin.

Results: Mean FG at start of tests was 6 mg/dL (0.3 mmol/L) lower with PF and EX and 10 mg/dL (0.6 mmol/L) lower with DD in the efsitora group compared to glargine. Incidence of Level 1 hypo (≥54 to <70 mg/dL [≥3.0 to <3.9 mmol/L]) was not significantly different under any test condition: incidence efsitora vs glargine, difference in proportion (95%CI) for PF: 44.7 vs 42.6%, 2.1 (-17.2, 21.4); EX: 65.9 vs 50.0%, 15.9 (-3.0, 34.8); DD: 68.1 vs 61.7%, 6.4 (-12.8, 25.6). Level 1 hypo resolved spontaneously or after 15 g oral glucose. Level 2 (<54 mg/dL [3.0 mmol/L]) was infrequent in both treatments and all test conditions. No severe hypo occurred in this study. Mean nadir glucose for hypo was similar between treatments and test conditions ranging from 62.8 to 66.3 mg/dL [3.5 to 3.7 mmol/L]. Duration of hypo events was also similar between treatments ranging from 76.6 to 115.2 mins depending on the test condition.

Conclusion: Once weekly efsitora did not increase the incidence, duration, or severity of hypo compared to once daily glargine during periods of provocation in patients with T2D.

Disclosure: Initially submitted at ADA 2024.

Disclosure: T. Heise: Grants; Adocia, AstraZeneca, Altimmune, Biocon, Bioton, Civica Foundation, Crinetics Pharmaceuticals, Eli Lilly, Enyo Pharma, Gan & Lee Pharmaceuticals, Genova, Nanexa AB, Novo Nordisk, SamChunDang Pharm. Co., Sanofi, Zealand Pharma. Honorarium; Gan & Lee Pharmaceuticals. Lecture/other fees; Eli Lilly.

9

No evidence of increased physical activity-related hypoglycaemia with once-weekly insulin icodec vs once-daily basal insulin in type 1 diabetes: ONWARDS 6

H. Sourij ¹, R. Bracken², M. Asong³, L. Carstensen³, S. Kehlet Watt³, A. Philis-Tsimikas⁴;

¹1Interdisciplinary Metabolic Medicine Trials Unit, Division of Endocrinology and Diabetology, Medical University Graz, Graz, Austria, ²Applied Sport, Technology, Exercise and Medicine Research Centre, Swansea University, Swansea, UK, ³Novo Nordisk A/S, Søborg, Denmark, ⁴Scripps Whittier Diabetes Institute, San Diego, CA, USA.

Background and aims: ONWARDS 6 compared the efficacy and safety of once-weekly insulin icodec (icodec) and once-daily insulin degludec (degludec) in adults with type 1 diabetes (T1D). This post hoc analysis examined physical activity-related hypoglycaemia in ONWARDS 6.

Materials and methods: During the trial, participants experiencing hypoglycaemic episodes were asked to indicate any relationship to physical activity in a digital diary. Overall hypoglycaemia rate was also assessed based on baseline physical activity level recorded using the International Physical Activity Questionnaire.

Results: The proportion of physical activity-related hypoglycaemic episodes relative to overall hypoglycaemic episodes was similar or lower for icodec versus degludec (Figure A); odds of having physical activity-related clinically significant or severe hypoglycaemia were similar between groups (OR: 1.04; 95% CI: 0.75-1.46; p = 0.8060). Most physical activity-related clinically significant or severe hypoglycaemic episodes did not lead to further hypoglycaemic episodes within 24 hours in either group (Figure B). Overall hypoglycaemia rates (including those unrelated to physical activity) were comparable within each treatment group regardless of baseline physical activity level (data not shown).

Conclusion: There was no evidence of increased risk of physical activity-related hypoglycaemic episodes with icodec versus degludec in adults with T1D.

Clinical Trial Registration Number: NCT04848480

Supported by: Study funding: Supported by Novo Nordisk A/S. Acknowledgements: Medical writing support provided by Oxford PharmaGenesis, funded by Novo Nordisk A/S

Disclosure: H. Sourij: Grants; Boehringer Ingelheim, Eli Lilly, Novo Nordisk, Sanofi. Honorarium; Amgen, Amarin, Bayer, Boehringer Ingelheim, Eli Lilly, Daiichi Sankyo, Novartis, Novo Nordisk, Cancom.

10

Associations between mean fasting glucose levels and adherence to app-based dose guidance for once-weekly insulin icodec in insulin-naive type 2 diabetes: post hoc analysis of ONWARDS 5

H. Bajaj ¹, A.M. Donatsky², S. Engberg², J.H. Martiny², A. Vianna³, I. Lingvay⁴;

¹LMC Diabetes & Endocrinology, Brampton, ON, ON, Canada, ²³Centro de Diabetes Curitiba, Curitiba, Brazil, ⁴UT Southwestern Medical Center, Dallas, TX, USA.

Background and aims: Inadequate titration of basal insulin contributes to suboptimal glycaemic control in clinical practice. In ONWARDS 5, a 52-week, phase 3a trial, participants in the insulin icodec (icodec) group used a dosing guide app to assist titration. This post hoc analysis assessed adherence to app-based dose guidance and its association with pre-breakfast self-measured blood glucose (SMBG).

Materials and methods: The app provided weekly algorithm-driven icodec dose guidance based on pre-breakfast SMBG. Regardless, investigators could suggest manual dose changes. Estimated mean pre-breakfast SMBG over time was compared across app-based guidance adherence subgroups. The subgroups were defined by the fraction of weeks adherent to app-based guidance out of the weeks with app usage (weeks adherent per 4 weeks: low ≤1; medium >1-≤3; high >3).

Results: On average 5.0% of administered icodec doses differed from app guidance (dose change by participant). Of 541 participants assessed, 56.7% had ≥1 manual dose change during the trial; most changes (82%) were increases to the dose recommended by the app. Estimated mean pre-breakfast SMBG improved from baseline irrespective of adherence to app guidance. Medium/high adherence subgroups had statistically significantly lower mean pre-breakfast SMBG (p <0.0001) with earlier achievement of target level (4.4-7.2 mmol/L) than the low adherence subgroup (Figure).

Conclusion: Most administered icodec doses were adherent with app guidance; medium/high vs low adherence to app guidance was associated with improved estimated SMBG. The findings support the use of dose guidance apps to assist real-world insulin titration.

Clinical Trial Registration Number: NCT04760626

Supported by: Novo Nordisk A/S

Disclosure: H. Bajaj: Grants; Novo Nordisk. Non-financial support; Novo Nordisk.

11

Efficacy and safety of once-weekly insulin icodec vs once-daily basal insulin in type 2 diabetes according to baseline sodium-glucose cotransporter-2 inhibitor use: ONWARDS 1-5

R. Goldenberg ¹, K.O. Bangsgaard², A. Fu², M. Kellerer³, S.B. Søgaard², T. Vilsbøll^4,5;

¹LMC Diabetes & Endocrinology, Concord, ON, Canada, ²Novo Nordisk A/S, Søborg, Denmark, ³Marienhospital Stuttgart, Stuttgart, Germany, ⁴Clinical Research, Steno Diabetes Center Copenhagen, Herlev, Denmark, ⁵Department of Clinical Medicine, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark.

Background and aims: To assess the treatment effects of once-weekly insulin icodec (icodec) vs once-daily basal insulin comparators ± concomitant sodium-glucose cotransporter-2 inhibitor (SGLT2i) use in ONWARDS 1-5.

Materials and methods: A post hoc analysis by trial of insulin-naïve (ONWARDS 1, 3, 5) and insulin-experienced (ONWARDS 2, 4) adults with type 2 diabetes (T2D).

Results: Participants on icodec vs once-daily comparators had greater or comparable HbA_1C reductions from baseline to end of treatment irrespective of SGLT2i use; there was no statistically significant treatment by subgroup interaction in HbA_1C changes (Table). Among subgroups in ONWARDS 1-3 and 5, overall rates of clinically significant or severe hypoglycaemia were low (<1 event/patient-year of exposure) across arms, with a numerically higher rate with icodec; in all trials, the overall hypoglycaemia rates were lower or similar among SGLT2i users vs non-users across arms (Table). There was no statistically significant treatment by subgroup interaction in terms of attainment of HbA_1C <7% without clinically significant or severe hypoglycaemia in any trial (p ≥ 0.31 for all trials).

Conclusion: The efficacy and safety of icodec vs once-daily comparators was generally consistent among adults with T2D, irrespective of baseline SGLT2i use.

Clinical Trial Registration Number: ONWARDS 1 (NCT04460885); ONWARDS 2 (NCT04770532); ONWARDS 3 (NCT04795531); ONWARDS 4 (NCT04880850); ONWARDS 5 (NCT04760626)

Supported by: Study funding: Supported by Novo Nordisk A/S. Acknowledgements: Medical writing support provided by Oxford PharmaGenesis, funded by Novo Nordisk A/S

Disclosure: R. Goldenberg: Employment/Consultancy; Boehringer Ingelheim, Eli Lilly, Janssen, Novo Nordisk, Sanofi. Grants; AstraZeneca, Boehringer Ingelheim, Eli Lilly, Janssen, Novo Nordisk, Sanofi. Honorarium; Boehringer Ingelheim, Eli Lilly, Janssen, Novo Nordisk, Sanofi. Lecture/other fees; Boehringer Ingelheim, Eli Lilly, Janssen, Novo Nordisk, Sanofi.

12

Efficacy and safety of once-weekly insulin icodec vs once-daily basal insulin in type 2 diabetes according to baseline glucagon-like peptide-1 receptor agonist use: ONWARDS 1-5

T. Vilsbøll ^1,2, K.O. Bangsgaard³, A. Fu³, M. Kellerer⁴, S.B. Søgaard³, R. Goldenberg⁵;

¹Clinical Research, Steno Diabetes Center Copenhagen, Herlev, Denmark, ²Department of Clinical Medicine, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark, ³Novo Nordisk A/S, Søborg, Denmark, ⁴Marienhospital Stuttgart, Stuttgart, Germany, ⁵LMC Diabetes & Endocrinology, Concord, ON, Canada.

Background and aims: To assess the treatment effects of once-weekly insulin icodec (icodec) vs once-daily (OD) basal insulin comparators ± concomitant glucagon-like peptide-1 receptor agonist (GLP-1 RA) use in ONWARDS 1-5.

Materials and methods: A post hoc analysis by trial of insulin-naïve (ONWARDS 1, 3, 5) and insulin-experienced (ONWARDS 2, 4) adults with type 2 diabetes (T2D).

Results: Participants on icodec vs OD comparators had larger or similar HbA_1C reductions from baseline to end of treatment irrespective of GLP-1 RA use; there was no statistically significant treatment by subgroup interaction in HbA_1C changes (Table). Overall rates of clinically significant or severe hypoglycaemia were low (<1 event/patient-year of exposure) across arms among ONWARDS 1-3 and 5 subgroups, with numerically lower rates among GLP-1 RA users vs non-users; in ONWARDS 4 (basal-bolus trial), the hypoglycaemia rate was similar among GLP-1 RA users vs non-users in the icodec arm (Table). There was no statistically significant treatment by subgroup interaction in any trial for the attainment of HbA_1C <7% without clinically significant or severe hypoglycaemia (p ≥ 0.21 for all trials), nor body weight change from baseline to end of treatment (p ≥ 0.17 for all trials).

Conclusion: The efficacy and safety of icodec vs OD comparators was generally consistent among adults with T2D, irrespective of baseline GLP-1 RA use.

Clinical Trial Registration Number: ONWARDS 1 (NCT04460885); ONWARDS 2 (NCT04770532); ONWARDS 3 (NCT04795531); ONWARDS 4 (NCT04880850); ONWARDS 5 (NCT04760626)

Supported by: Study funding: Supported by Novo Nordisk A/S. Acknowledgements: Medical writing support provided by Oxford PharmaGenesis, funded by Novo Nordisk A/S

Disclosure: T. Vilsbøll: Employment/Consultancy; Antaros Medical, Amgen, AstraZeneca, Boehringer Ingelheim, Eli Lilly, GSK, Novo Nordisk, Zealand Pharma, Sanofi, Sun Pharmaceuticals. Grants; Antaros Medical, Amgen, AstraZeneca, Boehringer Ingelheim, Eli Lilly, GSK, Novo Nordisk, Zealand Pharma, Sanofi, Sun Pharmaceuticals. Honorarium; Antaros Medical, Amgen, AstraZeneca, Boehringer Ingelheim, Eli Lilly, GSK, Novo Nordisk, Zealand Pharma, Sanofi, Sun Pharmaceuticals.

OP 03 Liver: the master regulator

13

Pharmacological inhibition of fructose metabolism reduces liver fat and improves insulin sensitivity in participants with MASLD

E.J.C. Koene ¹, J. Basset Sagarminaga¹, K. Brouwers², J. Mevenkamp², Y.M.H. Op den Kamp-Bruls², E. Phielix¹, V. Schrauwen-Hinderling², P. Schrauwen¹, M.C.G. Brouwers³;

¹Nutrition and Movement Sciences, Maastricht University, Maastricht, Netherlands, ²Radiology and Nuclear Medicine, Maastricht University Medical Center+, Maastricht, Netherlands, ³Internal Medicine/Endocrinology, Maastricht University Medical Center+, Maastricht, Netherlands.

Background and aims: Previous studies have shown that fructose intake is associated with metabolic dysfunction-associated steatotic liver disease (MASLD), which is a risk factor for type 2 diabetes (T2D). Of interest, experimental studies demonstrated that pharmacological inhibition of ketohexokinase (KHK), the first committed step in fructose metabolism which converts fructose to fructose-1-phosphate, effectively lowers de novo lipogenesis and intrahepatic lipid (IHL) content, which have recently been confirmed in phase II clinical studies. Here, we investigated the effect of a KHK inhibitor, PF-06835919, on insulin sensitivity.

Materials and methods: We conducted a randomized double-blind, placebo-controlled cross-over trial in which fifteen overweight/obese participants with MASLD (IHL ≥ 5.56%) were treated with PF-06835919 (300 mg, once daily) and placebo. Each 6-week treatment arm was evaluated by assessment of: 1) in vivo intrahepatic fructose phosphorylation (measured with ³¹P- Magnetic Resonance Spectroscopy (MRS) after an oral fructose load) 2) IHL content (quantified with ¹H-MRS), and 3) insulin sensitivity (M-value assessed with two-step hyperinsulinemic-euglycemic clamp).

Results: Hepatic phosphomonoester concentration increased upon an oral fructose load in the placebo arm, which was completely abolished after PF-06835919 treatment (p<0.001), showing that the drug inhibits normal fructose metabolism. IHL content was lower after PF-06835919 versus placebo (absolute difference: -2.5 %; 95% CI: -3.3, -1.8; p<0.001). Body weight remained stable between both periods (difference: 0.37 kg; 95% CI: -0.12, 0.83 kg; p=0.12). Whole body insulin sensitivity (M-value) was significantly higher (p=0.007) after 6 weeks of PF-06835919 (mean: 5.2 mg/kg/min) as compared to placebo (mean: 4.4 mg/kg/min).

Conclusion: Six weeks of KHK inhibition suppresses intrahepatic fructose metabolism, reduces IHL content, and improves insulin sensitivity in MASLD participants.

Clinical Trial Registration Number: NCT05463575

Supported by: Pfizer

Disclosure: E.J.C. Koene: None.

14

Osteoprotegerin regulates hepatic glucose metabolism and insulin sensitivity through mTORC1 signalling pathway

S. Qiu, G. Yang, M. Yang;

Department of Endocrinology, The Second Affiliated Hospital, Chongqing, China.

Background and aims: Osteoprotegerin (OPG), a soluble glycoprotein, belongs to the tumor necrosis factor (TNFα) receptor superfamily and is also an inflammatory cytokine receptor associated with bone remodeling, responsible for inhibiting osteoclast formation. In recent years, clinical studies have indicated that OPG may be a risk factor for atherosclerotic cardiovascular disease (CVD) and is associated with other metabolic diseases, including obesity, polycystic ovary syndrome (PCOS), nonalcoholic fatty liver disease (NAFLD), and type 2 diabetes mellitus (T2DM). However, the results of previous studies are inconsistent. For instance, it has been reported that serum OPG concentrations in patients with obesity, T2DM, PCOS, or NAFLD were either increased or decreased compared to those in normal adults. Although prior studies have demonstrated that inhibition of the RANK-RANKL-OPG system can improve hepatic insulin resistance (IR) and blood glucose levels, whether OPG in hepatocytes can regulate glucose metabolism, and its mechanism remains unknown. Therefore, understanding OPG’s role in glucose metabolism regulation is crucial for developing effective diabetes treatments.

Materials and methods: We utilized OPG global knockout (OPG-/-) mice and adeno-associated virus-mediated hepatic OPG overexpression mice to investigate the effect of OPG on glucose metabolism. To further explore the impacts of OPG on age-associated metabolic phenotypes, male wild-type (WT) and OPG-/- mice fed with either a normal chow diet (NCD) or a high-fat diet (HFD) were observed for 18 months. We examined phenotypic changes in body weight, glucose metabolism, insulin sensitivity, and energy expenditure. To accurately evaluate glucose metabolism and insulin sensitivity in vivo, we performed hyperinsulinemic-euglycemic clamps (HECs), considered the gold standard for evaluating IR, using tracer dilution methodology in conscious 5-month-old OPG-/- and WT mice. The interaction between OPG and raptor was analyzed by co-immunoprecipitation (CO-IP), and the regulation of raptor protein phosphorylation by OPG was examined using Phos-tag SDS-PAGE.

Results: We found that in obese mice, DNA hypermethylation inhibits hepatic OPG expression. Overexpression of OPG in the liver reduced the basal metabolic rate and aggravated glucose metabolism disorders in mice under HFD, whereas deletion of OPG increased the basal metabolic rate and improved glucose metabolism. Additionally, OPG deficiency consistently improved glucose metabolism and diet-induced IR during aging. OPG regulated glucose metabolism through the mTORC1 (Raptor)-S6K-IRS1/Akt signaling pathway. Mechanistically, the interaction between OPG and Raptor led to the phosphorylation of Raptor at the Ser⁸⁶³ or Ser⁸⁷⁷ sites, resulting in a band shift.

Conclusion: These data provide new insights into the understanding of the connection between OPG and glucose metabolism and insulin sensitivity, as well as into developing intervention strategies for targeting IR and T2DM.

Supported by: his work was supported by grants from the National Natural Science Foundation (82170816, 82300922), Chongqing Natural Science Foundation Project - Pos

Disclosure: S. Qiu: None.

15

Investigating whole-body fructose catabolism in vivo using stable isotopes to identify new anti-obesity targets

M. Rios-Morales, F. Westerbeke, D. Van Harskamp, B. Groen, M. Nieuwdorp;

Amsterdam University Medical Centers (Location AMC), Amsterdam, Netherlands.

Background and aims: Dietary fructose consumption has increased tremendously over the last 30 years and epidemiological studies have linked this to increased incidence of obesity, fatty liver disease, diabetes and cardiovascular disease. Fructose is classically described to be catabolized in the liver, where it enters glycolysis bypassing the insulin-dependent regulation that controls glucose catabolism. The fructose-derived metabolites are then processed into fatty acids in an unregulated fashion, which may lead to ectopic fat accumulation in diverse peripheral organs. Although the hepatic biochemical pathway of fructose has been elucidated, how and where fructose is majorly metabolized in the whole-body context has been poorly described. Recently it has been shown in mice, that host and microbial intestinal fructose metabolism plays a key role in the whole-body effect of fructose. In humans, however, it is still not clear whether these processes also take place upon fructose intake. Thus, here we present an update on an ongoing clinical trial that aims to study intestinal and whole-body fructose catabolism in patients with metabolic syndrome and the effect of fecal microbial transplantation (FMT) on its kinetics.

Materials and methods: Fructose catabolism was measured through a fructose challenge test in patients with metabolic syndrome before and after FMT. Subjects ingested a high dose of fructose (1gr/kg bodyweight) followed by a small dose of ¹³C-Fructose (120 mg) used as a tracer. For up to 6 hours plasma and breath samples were collected to measure concentrations and ¹³C incorporation into different metabolites. Fractional dose contributions were calculated using a computational model to estimate the changes in fructose catabolism upon FMT. Of note, due to current blinding we don’t know whether subjects received allogeneic or autologous FMT.

Results: The ingestion of high fructose dose leads to high production of H₂, an intestinal bacterial fermentation gas, and fast appearance of ¹²C/¹³C-fructose in the peripheral circulation. Uric acid was significantly decreased after FMT with a mean of 435 μM, compared to 537 μM before FMT (p=0.03). Fructose ingestion increased peripheral glucose levels up to 7.4 mM upon ingestion. However, after FMT plasma glucose levels changed from reaching its peak at 90 min to 30 min after fructose consumption. The majority of the ¹³C-label appears in CO₂ in the breath samples and no label incorporation was found in amino acids, organic acids, and fatty acids in blood.

Conclusion: The data suggest that 1gr/kg bodyweight of fructose saturates the human small intestine for its absorption leading to its microbial fermentation and its appearance in peripheral circulation. Fructose can also be converted to glucose and/or be oxidized after fasting. This glucose production seems to change from hepatic to intestinal upon fecal microbial transplant. This would state that humans can also activate intestinal gluconeogenesis upon fructose consumption and that this process depends at least in part on the gut microbiota. This is in line with the decrease in hepatic uric acid production after FMT. These results would pinpoint intestinal host/microbial fructose catabolism as an important novel target for treating obesity and thus, type 2 diabetes and other related diseases.

Clinical Trial Registration Number: NL-OMON24239

Supported by: ZonMw Vici grant 2020 to M. Nieuwdorp (09150182010020)

Disclosure: M. Rios-Morales: None.

16

Increased postprandial succinate levels are associated with hepatic gluconeogenesis regulation via SUCNR1

A. Marsal-Beltran ^1,2, L. Salmerón-Pelado¹, A. Ribas-Latre¹, M.-M. Rodríguez-Peña^1,2, C. Núñez-Roa^1,2, M. Repollés-de-Dalmau^1,2, E. Novoa^3,4, R. Nogueiras^3,4, R. Bosch⁵, J. Vendrell^1,2, V. Ceperuelo-Mallafré^1,2, S. Fernández-Veledo^1,2;

¹Research Unit, University Hospital of Tarragona Joan XXIII, Pere Virgili Health Research Institute (IISPV), Universitat Rovira i Virgili (URV), Tarragona, Spain, ²CIBER de Diabetes y Enfermedades Metabólicas Asociadas (CIBERDEM) - Instituto de Salud Carlos III, Madrid, Spain, ³Center for Research in Molecular Medicine and Chronic Diseases (CIMUS), University of Santiago de Compostela- Instituto de Investigación Sanitaria, Santiago de Compostela, Spain, ⁴CIBER de Fisiopatología de la Obesidad y Nutrición (CIBEROBN) - Instituto de Salud Carlos III, Madrid, Spain, ⁵Department of Pathology, Oncological Pathology and Bioinformatics Research Group, Hospital de Tortosa Verge de la Cinta, IISPV, Tortosa, Spain.

Background and aims: Diabetes mellitus poses a substantial challenge in healthcare, affecting multiple processes, notably hepatic gluconeogenesis. However, the exact molecular mechanisms causing this dysregulation have yet to be uncovered. In this context, succinate emerges as a signaling metabolite that rises in circulation after nutrient intake and in patients with diabetes, for which exploring its connection with gluconeogenesis offers potential for discovering new therapeutic targets.

Materials and methods: We examined Sucnr1 expression in the fasted-to-fed transition in murine livers and in hepatic cell lines. Furthermore, we generated mice with hepatocyte-specific deficiency of Sucnr1 (Albumin-Cre model) to investigate their metabolic phenotype, with littermate mice carrying Sucnr1 flanked by loxP sequences serving as controls. Additionally, we assessed the gluconeogenic response in murine primary hepatocytes isolated from WT and Sucnr1 knockout (KO) mice, as well as in hepatic cell lines treated with SUCNR1 agonists and antagonists.

Results: Sucnr1 mRNA expression was significantly enhanced in the livers of WT mice when fed ad libitum or after 24 hours of refeeding, as compared to mice that were fasted for 24 hours or subjected to a 4-day caloric restriction. We confirmed in vitro that glucose was responsible for an increase in succinate secretion from hepatocytes and Sucnr1 expression. Furthermore, mice lacking Sucnr1 in hepatocytes exhibited an accentuated gluconeogenic capacity in response to fasting, pyruvate tolerance test and ITT. Accordingly, primary hepatocytes isolated from global Sucnr1 KO mice showed higher glucose production and increased expression of gluconeogenic enzymes. We further confirmed these results using a SUCNR1 antagonist in THLE-2 cells and a SUCNR1 agonist in AML12 cells. At a mechanistic level, we demonstrated that SUCNR1 agonism decreased p-AMPKα and the expression of SIRT1 and FGF21. By contrast, we observed that the gluconeogenic phenotype of mice lacking Sucnr1 in hepatocytes was reverted over time, an effect that was concomitant with an altered expression pattern of hepatokines, decreased weight and adipocyte area in white adipose tissue depots, and increased number and size of pancreatic islets, suggesting a mechanism of compensation induced by the liver.

Conclusion: In essence, our research underscores the crucial involvement of SUCNR1 in regulating the liver’s response to fasting and feeding. Deleting Sucnr1 in hepatocytes leads to heightened gluconeogenesis, necessitating compensatory mechanisms outside the liver to mitigate disruptions in glucose homeostasis.

Supported by: LCF/PR/HR20/52400013; PID2021-122480OB-I00

Disclosure: A. Marsal-Beltran: Grants; FPU20/05633.

17

Whole-body insulin sensitivity is lower after 4-week high fructose compared to 4-week high saturated fat consumption in overweight individuals

K.H.M. Roumans ¹, P.M.G. van Lier¹, E. Phielix¹, B. Havekes², P. Schrauwen¹, V.B. Schrauwen-Hinderling³;

¹Nutrition and Movement Sciences, Maastricht University, Maastricht, Netherlands, ²Internal Medicine, Division of Endocrinology, Maastricht University Medical Center, Maastricht, Netherlands, ³Radiology and Nuclear Medicine, Maastricht University Medical Center, Maastricht, Netherlands.

Background and aims: High intake of fructose and/or saturated fat (SFA) has been suggested to impair insulin sensitivity. It has been hypothesized that diets high in fructose and/or SFA induce hepatic lipid accumulation, thereby impairing hepatic insulin sensitivity. Recent evidence suggests that next to intrahepatic lipid (IHL) content also its composition, i.e. the fraction of hepatic saturated and unsaturated fatty acids, might play a role in determining insulin sensitivity. However, it is so far unknown whether indeed hepatic SFA per se negatively affects hepatic insulin resistance. In that context, both high fructose and high SFA intake can lead to increased hepatic SFA, albeit via different pathways. Here, we investigated the effect of a 4-week high SFA diet compared to a 4-week high fructose diet on whole body and hepatic insulin sensitivity, while determining hepatic SFA fraction and IHL content.

Materials and methods: Thirteen overweight individuals (BMI 27-38 kg/m², age 45-75y) consumed a 4-week high fat/high SFA diet (SFA 23 En%) and a 4-week isocaloric high carbohydrate/high fructose diet (fructose 21 En%) in randomized order, with a 6-week washout in-between. Hepatic and whole-body insulin sensitivity were measured by a two-step hyperinsulinemic-euglycemic clamp (10 and 40 mU/m²/min) at the end of both arms. IHL content and composition were determined at the start and end of both arms using proton magnetic resonance spectroscopy.

Results: Whole-body insulin sensitivity (Δrate of glucose disappearance) was significantly lower upon 4 weeks of high fructose (14.7±2.2 μmol/kg/min), compared to 4 weeks of high SFA (18.0±2.4 μmol/kg/min, p=0.004). Hepatic insulin sensitivity (suppression of endogenous glucose production) was not significantly different between diets (fructose; 68.0±4.9%, SFA; 70.7±4.2%, p=0.51). Hepatic SFA fraction tended to increase by 3.0 percentage points (pp) on high fructose (p=0.06) and increased by 2.5 pp on high SFA (p=0.01), without differences between diets (p=0.45). IHL content and body weight were not different between diets (p=0.17 and p=0.48, respectively).

Conclusion: Our data shows that a high-fructose diet is more detrimental for whole-body insulin sensitivity compared to an isocaloric diet high in SFA. This difference cannot be explained by differences in hepatic insulin sensitivity, hepatic SFA fraction or IHL content. Other mechanisms, affecting peripheral insulin sensitivity, are likely to explain the negative effect of fructose over SFA on whole-body insulin sensitivity.

Clinical Trial Registration Number: NCT05017675

Supported by: EFSD/Lilly European Diabetes Research Programme

Disclosure: K.H.M. Roumans: None.

18

Cytoskeleton protein Leupaxin promote hepatic gluconeogenesis through activating HNF4α transitional activity

Z. Zhijian, L. Xiaomin, L. Fang, P. Yongde;

Department of Endocrinology and Metabolism, Shanghai General Hospital, Shanghai Jiaotong University, Shanghai / Huangpu District, China.

Background and aims: Hepatic gluconeogenesis plays an important role in regulation of liver glucose output , and maladaptation of this physiological effect contributes to the development of diabetes and associated comorbidities. Leupaxin (LPXN) is a cytoskeleton protein that is homologous to the focal adhesion protein, paxillin. Even though this gene has been reported to be associated with tumor development and immunity function, its role in the regulation of hepatic gluconeogenesis is unknown yet. We aim to further assess the expression pattern of LPXN gene in mouse liver tissue, and investigate its regulatory impact on hepatic gluconeogenesis.

Materials and methods: In our study, we observed the expression pattern of LPXN gene in mouse and human liver. This was followed by mouse experiments using adenovirus-mediated LPXN over-expression and adeno-associated virus-ShRNA mediated LPXN knock down in mouse liver. We investigated the effect of LPXN on hepatic gluconeogenesis and overall glucose homeostasis in vivo and ex vivo, as well as the possible underlying molecular mechanism.

Results: We observed that LPXN gene was routinely expressed in human and mouse liver and increase with liver fat deposit. Furthermore, over-expression of LPXN in liver of mice resulted in rising of blood glucose level and increasement of gluconeogenesis associated gene expression, down-expression of LPXN in liver of db/db mice and ob/ob mice resulted in improvemen of glucose level and reduction of gluconeogenesis related gene expression. Ex vivo experiments demonstrated that LPXN overexpression incrase gluconeogenesis in liver cell, LPXN downexpression resulted in reduction of gluconeogenesis related gene expression. Mechanistically, LPXN could translocate into nucleus interact with hepatic nuclear factor-4 (HNF4α) in liver and increase PEPCK expression.

Conclusion: Taken together, we provide the evidence for a novel role of LPXN in modulating hepatic gluconeogenesis, thereby reinforcing the fact that targeting LPXN may be a potential approach for the treatment of diabetes and associated complications.

Supported by: NSFC 81970751

Disclosure: Z. Zhijian: None.

OP 04 Beyond glucose in diabetes and prediabetes

19

Effects of a randomised controlled dietary intervention high in unsaturated fat, plant protein and fiber on intrahepatic lipid levels at 12 and 36 months

E. Álvarez Hernández ^1,2, M. Lazaratos^3,1, C. Wernicke¹, A. Pohrt⁴, L. Pletsch-Borba¹, N. Meyer¹, J. Machann⁵, A.F.H. Pfeiffer¹, J. Spranger^1,6, K. Mai^1,2;

¹Department of Endocrinology and Metabolism (Diabetes and Nutritional Medicine), Charité Universitätsmedizin, Berlin, Germany, ²Human Nutrition, German Institute of Human Nutrition, Nuthetal, Germany, ³Department of Experimental Diabetology, German Institute of Human Nutrition, Nuthetal, Germany, ⁴Institute of Biometry and Clinical Epidemiology, Charité Universitätsmedizin, Berlin, Germany, ⁵German Center for Diabetes Research, Institute for Diabetes Research and Metabolic Diseases (IDM) of the Helmholtz Center Munich, Tübingen, Germany, ⁶German Center for Diabetes Research, München-Neuherberg, Germany.

Background and aims: Previous short- to medium-term studies suggest that dietary protein and unsaturated fatty acids (UFA) promote improvements in intrahepatic lipid levels (IHL) independent of weight loss. Recent research suggests that dietary patterns have a greater impact on health outcomes than individual macronutrients. The effect of a dietary pattern high in protein and UFA on IHL at 12 and 36 months was analyzed. The aim was to elucidate the long-term effects of a dietary intervention on IHL. To our knowledge, this is the first study to evaluate dietary effects on IHL for such a long duration.

Materials and methods: In a 36-month randomized controlled trial, 502 participants aged 50-80 years with at least one risk factor for unhealthy aging were assigned to an intervention group that consumed a diet high in plant protein (15%-25% total energy intake (TEI)), fiber ( ≥30 g/ day), mono-UFA (15- 20% TEI) and poly-UFA (10-15% TEI) or to a control group that followed the dietary recommendations of the German Nutrition Society ((GNS) fat 30% TEI, carbohydrates 55% TEI, protein 15% TEI). IHL were assessed in 397 participants using magnetic resonance spectroscopy over a 36-month period. Results were analyzed using a linear mixed effects model.

Results: Preliminary results show a decrease in IHL at both 12 and 36 months (12 months -0.37; 95% CI: -0.54 - -0.20, p<0.001; 36 months -0.27; 95% CI: -0.46 - -0.08, p= 0.005). Fittingly, age also led to a decrease in IHL. Of note, there was a decrease in IHL when considering the interaction between intervention and time at 36 months (-0.31; 95% CI: -0.57- -0.04, p=0.024) and a trend towards reduced IHL when considering the interaction between intervention and time at 12 months (-0.23; 95% CI: -0.48 - 0.02, p=0.07).

Conclusion: The preliminary results of this study indicate that a dietary pattern emphasizing high intake of plant protein, fiber, and UFA may result in greater reductions in IHL at 12 and 36 months compared to a diet based on the GNS recommendations. These findings suggest that this type of dietary pattern may have long-term benefits that increase over time. This highlights the need for further research to understand the underlying mechanisms.

Clinical Trial Registration Number: DRKS00010049

Supported by: German Federal Ministry of Education and Research (BMBF), German Center for Diabetes Research and German Research Foundation

Disclosure: E. Álvarez Hernández: None.

20

Slower liver growth in-utero predicts the risk of prediabetes at 18 years of age, a DOHaD jump from birth size to organogenesis: novel human data in the Pune Maternal Nutrition Study (PMNS)

C.S. Yajnik ¹, A. Kinare¹, M. Chinchwadkar¹, D. Shetty¹, K. Coyaji¹, P. Yajnik¹, R. Ladkat¹, R. Wagh¹, C. Osmond², C.H. Fall²;

¹Diabetes Unit, KEM Hospital Research Centre, Pune, India, ²MRC-LEU, University of Southampton, Southampton, UK.

Background and aims: DOHaD studies usually report on associations between birth size (weight) and later risk of disease. We investigated if the rate of liver growth in-utero predicted the risk of diabetes in later life.

Materials and methods: In the PMNS (1993-96), we measured fetal liver growth by ultrasound between 28 and 34 weeks of gestation. We measured glucose tolerance (75 gm OGTT, ADA), glucose-insulin indices (HOMA), lipids and liver size at 18 years of age in the offspring (2012-15). We studied associations of in-utero liver growth with cardiometabolic risk factors at 18y.

Results: At 18y, BMI was 19.2 kg/m2 (±3.2), liver size was 12.75 cms (±1.42) and 30% had prediabetes. Fatty liver was uncommon (3.5%). Larger liver was associated with higher BMI, fasting plasma glucose and triglycerides, and lower disposition index and HDL. Birthweight of these participants was 2.7 kg (±0.3). In-utero, liver grew at 0.075 cms/day (±0.037, n=188) in the 3rd trimester. Slower liver growth in-utero was associated with lower insulin sensitivity at 18y. Those with slower liver growth in-utero had higher risk of prediabetes (OR 4.3, 1.3 to 11), especially if they had larger liver size at 18y (Figure 1).

Conclusion: The PMNS provides the first human evidence that slower liver growth in-utero is associated with an increased risk of prediabetes in later life, acting through insulin resistance. The findings advocate a primordial pre-conceptional intervention to improve fetal organ growth to prevent diabetes.

Supported by: Wellcome Trust & MRC, UK; ICMR and DBT, India; and intramural funding

Disclosure: C.S. Yajnik: None.

21

Excessive free fatty acids sensing in pituitary lactotrophs elicits liver steatosis by decreasing prolactin levels

Y. Bi, X. Ji, H. Yin;

Department of Endocrinology, Affiliated Drum Tower Hospital, Medical School of Nanjing University, Nanjing, China.

Background and aims: Abnormal nutrient sensing, ectopic deposition of lipid and the imbalance of coordinated regulation among organs are the key factors in nonalcoholic fatty liver disease (NAFLD) development. According to previous findings, low pituitary prolactin (PRL) level is a risk factor for the occurrence and development of NAFLD, but the reason and mechanism for the reduction of PRL level are still unclear. As we all known that the ectopic deposition of lipids leads to organ damage. However, the pituitary as the endocrine and metabolic organ of human body, we know little about whether it can sense the nutritional signal of high fatty acids, then lead to the impairment of pituitary function. It has been known that overnutrition-associated elevation of serum free fatty acids (FFA) exerts toxic effects on hepatic lipid metabolism. Therefore, this study aimed to explore whether the pituitary lactotrophs can sense high fatty acid signals, and then cause hepatic steatosis by reducing the level of PRL.

Materials and methods: Clinical cohort study and dynamic diets intervention were performed to clarify the sequential relationships among the increase of serum FFA level, the decrease of serum PRL level and the occurrence of liver lipid accumulation. Stereotactic pituitary FFA injection was conducted to assess the effect of FFA on PRL level. Lipid uptake assays was employed to verify the lipid uptake behavior of pituitary lactotrophs. Virus transfection analysis was utilized to study the mechanism of pituitary lactotrophs sensing high fatty acid signal. Stereotactic pituitary AAV injection and pituitary targeted drug delivery were conducted to observe the effect of inhibiting fatty liver formation by targeted inhibition of FFA sensing in pituitary.

Results: A cohort including 328 biopsy-proven subjects were analyzed. We demonstrated that elevated levels of FFA increased the risk of NAFLD (OR 1.645, P < 0.05), which was mediated by decreased PRL levels with a mediation effect of 29.1%. Moreover, elevation of FFA contributed to decreased serum PRL levels thus promoted liver steatosis both in mice with dynamic diets intervention and stereotactic pituitary FFA injection. Mechanic studies showed that increased CD36 expression in pituitary lactotrophs inhibited the synthesis of PRL by enhancing FFA sensing through lipid uptake assays and virus transfection analysis. Importantly, silencing of pituitary CD36 by stereotactic virus injection, as well as by a pituitary-targeted delivery system of CD36 inhibitor, efficiently elevated PRL levels and alleviated liver steatosis.

Conclusion: We reported previously unknown mechanisms of excessive free fatty acids sensing in pituitary lactotrophs elicits liver steatosis by decreasing prolactin levels. This study provided a new insight into the understanding of inter-organ mechanisms of hepatic lipid homeostasis, and targeted inhibition of FFA sensing in pituitary may be a potential therapeutic target for liver steatosis

Disclosure: Y. Bi: None.

22

The effect of 12-month treatment with glucagon-like peptide-1 receptor agonists, sodium-glucose cotransporter-2 inhibitors and their combination on cardiac markers and mitochondrial function

L. Pliouta ¹, E. Maratou², G. Pavlidis¹, A. Kountouri¹, E. Korakas¹, J. Thymis³, K. Katogiannis³, I. Ikonomidis³, V. Lambadiari¹;

¹2nd Dept. of Internal Medicine – Propaedeutic, Research Institute and Diabetes Center Attikon Univ, Attikon hospital, Athens, Greece, ²Department of Clinical Biochemistry, Attikon hospital, Athens, Greece, ³Second Department of Cardiology, Attikon hospital, Haidari, Greece.

Background and aims: We investigated the effects of insulin, glucagon like peptide-1 receptor agonists (GLP-1RA), sodium-glucose cotransporter-2 inhibitors (SGLT-2i) and their combination on N-terminal pro B-type natriuretic peptide (NT-proBNP), growth differentiation factor (GDF)-15, mitochondrial-derived peptide-c (MOTS-c) and oxidative stress in type-2 diabetes mellitus (T2DM) patients

Materials and methods: A hundred-sixty T2DM patients were randomized to insulin (n=40), liraglutide (n=40), empagliflozin (n=40) or their combination (GLP-1RA+SGLT-2i) (n=40) as add-on to metformin. We measured at baseline, at 4 and 12 months of treatment: a)NT-proBNP, GDF-15, and MOTS-c, b)thiobarbituric acid reactive substances (TBARS) and 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS), and c)left ventricular global longitudinal strain (GLS) and global work index (GWI) using speckle-tracking imaging.

Results: Twelve months after treatment, GLP-1RA, SGLT-2i and their combination showed a greater reduction of NT-proBNP (-43.1% vs. -54.2% vs. -56.9% vs. -14.7%) and GDF-15 (-14.8% vs. -15.2% vs. -16.1% vs. -0.9%) than insulin (p<0.05), despite a similar glycosylated hemoglobin reduction. Only treatment with SGLT-2i and GLP-1RA+SGLT-2i improved MOTS-c (p<0.05). GLP-1RA or GLP-1RA+SGLT-2i provided a remarkable reduction of TBARS and increase of GLS and GWI compared to insulin or SGLT-2i. In all patients, the reduction of NT-proBNP was associated with the improvement of TBARS, GLS and GWI, while the decrease of GDF-15 was correlated with the increase of ABTS and MOTS-c at 12 months (p<0.05).

Conclusion: Twelve-month treatment with GLP-1RA, SGLT-2i and their combination resulted in reduction of cardiac biomarkers and oxidative stress than insulin. SGLT-2iand the combined treatment appear more effective in the improvement of markers of neurohumoral activation and of MOTS-c.

Clinical Trial Registration Number: NCT03878706

Disclosure: L. Pliouta: None.

23

Prolactin deficiency drives diabetes-associated cognitive dysfunction by inducing microglia-mediated synapse loss

Y. Bi, Z. Zhang, J. Jiang, J. Wang;

Department of Endocrinology, The Affiliated Drum Tower Hospital, Nanjing, China.

Background and aims: Cognitive dysfunction has been increasingly recognized as an important comorbidity of diabetes. Synapse loss is fundamental to the pathophysiology of diabetes-associated cognitive dysfunction. However, little is known about the pathogenesis of synapse loss in diabetes. As resident macrophages of the central nervous system (CNS), microglia drive synapse loss via ingestion of synapses. Hormone serves a critical function in regulating the phagocytic function of microglia. The pituitary hormone prolactin (PRL), an important metabolic hormone in diabetes, has been widely reported about its role in peripheral macrophage, whereas literature regarding the effects of PRL on microglia is sparse. Therefore, this study aimed to investigate whether PRL acts on microglia-mediated synapse loss and further contributes to cognitive impairment in diabetes.

Materials and methods: A total of 1421 participants who completed serum hormone levels assessments, comprehensive cognitive evaluations, and 3.0T magnetic resonance imaging scans were enrolled. The PRL gene of C57BL6J mice was knocked out by CRISPR/Cas9 technology. Dynamic diet interventions were applied in both chow diet mice and high-fat diet (HFD)-fed diabetic mice. Microglial PRLR-cKO mice were generated after we treated Cx3cr1^creERT2: PRLR^fl/fl mice with tamoxifen. Intracerebroventricular infusion of PRL recombinant protein or its vehicle was performed in diabetic mice after 12 weeks of HFD feeding.

Results: Serum PRL level was a protective factor for mild cognitive impairment (MCI) (OR 0.921 95% CI: 0.867-0.978), and their levels were positively associated with hippocampal volumes (r=0.219, P <0.05) after adjusted for confounders. Further we generated PRL KO mice and immunofluorescence assays showed that the deficiency of PRL led to hippocampal synapse loss on account of increased microglial engulfment of synaptic structures. In HFD-fed diabetic model, based on the time course experiment, we found that decreased PRL level predated cognitive impairment and hippocampal synapse loss. Furthermore, microglial PRLR-cKO diabetic mice had worse cognitive performance and severe synapse loss than diabetic mice, indicating that the deleterious effects of PRL deficiency in diabetes-associated cognitive dysfunction were mediated by microglia. Moreover, infusion of PRL reduced the microglia-mediated synapse loss, further alleviating cognitive impairment in diabetic mice.

Conclusion: We reported a previously unknown mechanism in which prolactin deficiency drives diabetes-associated cognitive dysfunction by inducing microglia-mediated synapse loss. This study enriched the neurocognitive effects of the pituitary hormone PRL and provided an attractive therapeutic target for promoting cognitive health in diabetes.

Disclosure: Y. Bi: None.

24

When does metabolic memory starts? Insights from the AMD Annals initiative

G. Russo ¹, A. Nicolucci², G. Lucisano², M. Rossi², A. Ceriello³, F. Prattichizzo³, V. Manicardi⁴, A. Rocca⁵, P. Di Bartolo⁶, S. De Cosmo⁷, G. Di Cianni⁸, R. Candido⁹, AMD Annals Study Group;

¹University of Messina, Messina, Italy, ²CORESEARCH, Pescara, Italy, ³IRCCS MultiMedica, Milan, Italy, ⁴Associazione Medici Diabetologi, Roma, Italy, ⁵G. Segalini H. Bassini Cinisello Balsamo ASST Nord, Milan, Italy, ⁶Ravenna Diabetes Center - Romagna Local Health Authority, Ravenna, Italy, ⁷Dept Medical Sciences, Scientific Institute Casa Sollievo della Sofferenza, San Giovanni Rotondo (FG), Italy, ⁸Diabetes and Metabolic Deseases, Livorno Hospital, Livorno, Italy, ⁹Department of Medical Surgical and Health Sciences University of Trieste, Trieste, Italy.

Background and aims: Early, intensive glycemic control in T2D patients is associated with long-term benefits on cardiovascular disease (CVD) development. Evidence on the potential benefits obtained by achieving more stringent glycated targets, as close as possible to normal HbA1c value, i.e. HbA1c < 5.7%, is very limited to date, especially in real-world studies in patients with T2D.

Materials and methods: Newly diagnosed patients free of CVD at baseline were stratified according to the average HbA1c attained during the first 12, 24 and 36 months from diagnosis and the incidence of CVD in the following years was assessed. Overall, 251,339 subjects were identified in an Italian clinical registry, including information on all patients with T2D receiving care at over 300 diabetes clinics in Italy from January 2004 to December 2022. We adopted three definitions of early exposure periods (0-1, 0-2 and 0-3 years). Mean HbA1c values during the exposure periods were categorized into HbA1c < 5.7%, 5.7-6.4%, 6.5-7.0%, 7.1-8.0%, and >8.0%. The outcome was the incidence of major cardiovascular events (composite of myocardial infarction, stroke, coronary or peripheral revascularization, and coronary or peripheral bypass).

Results: At multivariate Cox regression analysis, compared with mean HbA1c <5.7% during the first year after diagnosis, the increase in the risk of incident CVD was 24% (HR=1.24; 95%CI 1.13-1.35), 42% (HR=1.42; 95%CI 1.30-1.56), 49% (HR=1.49; 95%CI 1.36-1.64), and 56% (HR=1.56; 95%CI 1.42-1.72) for patients with HbA1c of 5.7%-6.4%, 6.5%-7.0%, 7.1%-8.0%, and >8.0%, respectively. The same trend was documented in all exposure periods.

Conclusion: Legacy effect is a well- recognized phenomenon documented in several studies, suggesting that poor glycemic control after T2D diagnosis promotes an enduring damage on the vasculature. Our data support the need to safely achieve ambitious glycemic targets, immediately after diagnosis, to reduce cardiovascular risk in patients with T2D.

Supported by: The study was promoted by AMD. The analysis was partly supported by Eli Lilly SpA

Disclosure: G. Russo: Employment/Consultancy; Novo Nordisk, AstraZeneca, Sanofi, Boehringer, Lilly, Mundipharma, Sanchio.

OP 05 Can you change what you are by what you eat?

25

Unlocking the potential of urolithins: novel approaches to inhibit IAPP aggregation in diabetes

S.G. Ferreira ^1,2, A.F. Raimundo^3,4, N. Saraiva¹, C.N. Santos³, R. Menezes^1,3;

¹CBIOS - Center for Biosciences & Health Technologies, Universidade Lusófona - Centro Universitário de Lisboa, Lisboa, Portugal, ²Universidad de Alcalá, Escuela de Doctorado, Madrid, Spain, ³NOVA Medical School, Faculty of Medical Sciences, Lisboa, Portugal, ⁴iBET - Instituto de Biologia Experimental e Tecnológica, Oeiras, Portugal.

Background and aims: Pathological aggregation of Islet Amyloid Polypeptide (IAPP) is a significant contributor for β-cell dysfunction in diabetes onset and progression. However, its exploitation as a therapeutic target remains poorly explored. The pleiotropic action of (poly)phenols (PP) toward diabetes is well-documented. Notably, distinct classes of dietary PP have been shown to interfere with different steps of protein aggregation. Most of these compounds, however, have no to little bioavailability in the target tissues.

Materials and methods: Docking studies using AutoDock Vina and Pymol software were performed to screen an in-house collection of bioavailable PP metabolites towards their ability to interact with IAPP (NMR structure 2L86 from Protein Data Base). The best hits were tested in cell-free systems using synthetic IAPP by means of Thioflavin-T assays, Filter-tap and Transmission Electronic Microscopy (TEM). PP metabolites-mediated protection in INS-1 832/3 pancreatic β-cells challenged with toxic IAPP aggregates was measured by Lactate Dehydrogenase Test, PI/Annexin V staining and Glucose Stimulated Insulin Secretion (GSIS) assays. Mechanistic effects underlying metabolite protection were determined through RNAseq analysis, measurement of intracellular calcium levels using the Fluo-4 Direct dye and evaluation of oxygen consumption by Seahorse Extracellular Flux Analyzer.

Results: In computational studies, Urolithin A (UroA) and Urolithin B (UroB) emerged as highly effective molecules capable of physically interacting with IAPP. Both compounds displayed remarkable potential in altering IAPP fibril kinetics, size, and morphology. In INS-1 832/3 cells challenged with pre-formed IAPP aggregates, UroB demonstrated significant protection against IAPP-induced cytotoxicity. This effect was attributed to its impact on various cellular mechanisms, including redox homeostasis, KCl-induced calcium mobilization, increased spare respiratory response, and improved GSIS. Also, transcriptomic analysis revealed that IAPP aggregates formed in the presence of UroA modulated important cellular pathways related to metabolism, oxidative stress and energy regulation.

Conclusion: Our findings highlight the ability of UroA and UroB to directly interact with IAPP, thereby influencing the formation of amyloid structures. Furthermore, both metabolites were found to bolster cellular proteostasis mechanisms, empowering β-cells to more effectively cope with the challenges posed by IAPP. This research suggests a promising avenue for using dietary urolithins as potential inhibitors of IAPP aggregation, with implications for diabetes management.

Supported by: FCT under the [UIDB/04567/2020] and [UIDP/04567/2020]. COFAC/ILIND (grant COFAC/ILIND/FAZER+ EXLORER).

Disclosure: S.G. Ferreira: None.

26

Effects of a hypocaloric time-restricted feeding diet programme in individuals with metabolic dysfunction-associated steatotic liver disease: preliminary data of a randomised controlled trial

S. Tsitsou ¹, T. Bali², M. Adamantou², A. Saridaki², K.-A. Poulia¹, E. Petsiou³, E. Papakonstantinou¹, E. Cholongitas²;

¹Laboratory of Dietetics and Quality of Life, Department of Food Science and Human Nutrition, Agricultural University of Athens, Athens, Greece, ²First Department of Internal Medicine, Laiko General Hospital of Athens, Medical School, National and Kapodistrian University of Athens, Athens, Greece, ³Henry Dunant Hospital Center, Athens, Greece.

Background and aims: Time-restricted feeding (TRF) seems to have favorable effects on weight loss and cardiometabolic risk factors improvement in individuals with metabolic dysfunction-associated steatotic liver disease (MASLD). This study aims to investigate the impact of the time of the day in which feeding is restricted on several biochemical and anthropometric parameters.

Materials and methods: This 12-week randomized controlled trial in MASLD individuals with overweight/obesity consisted of three groups, all following a hypocaloric diet: control (C) without any time restriction in feeding, early 14:10 (14 hours of fasting) TRF (eTRF, eating between 08:00-18:00), and late 14:10 TRF (lTRF, eating between 12:00-22:00). All participants underwent liver elastography with controlled attenuation parameter (CAP) measurement to estimate hepatic steatosis. Recording of anthropometric and laboratory characteristics was performed at baseline and the end of the intervention, while the appropriate statistical analysis tests were run.

Results: Forty-nine MASLD individuals (22 males, 44.9%; mean age 53±1.7 years; mean BMI 32±0.6 kg/m²) have already finished the trial (C, n=16; eTRF, n=16; lTRF, n=17). Type 2 diabetes, hypertension, and dyslipidemia were prevalent at 36.7%, 46.9%, and 79.6%, of the individuals, respectively, without differences between groups at their baseline characteristics. In all groups, the following significant results were observed at 12 weeks, compared to baseline: a) weight loss (C: 8.2%, eTRF: 8.1%, lTRF: 7.9%, p_all<0.001), b) body fat loss (C: -5.4kg, eTRF: -5.2kg, lTRF: -5.5kg; p_all<0.001), c) waist circumference reduction (C: -8.7cm, eTRF: -7.6cm, lTRF: -8.6cm; p_all<0.001), d) neck circumference reduction (C: -1.7cm, eTRF: -1.7cm, lTRF: -1.5cm, p_all<0.010), and e) systolic/diastolic blood pressure reduction (C: -9.3/-5.5mmHg, eTRF: -15.5/-9.4mmHg, lTRF: -12/-5mmHg; p_all<0.010). Hepatic steatosis was reduced in the C (CAP: -28.3Db/m, p=0.001) and the eTRF (CAP: -30.6Db/m, p=0.002) groups. The three groups differed in total cholesterol (TC) (p=0.049), LDL (p=0.039), and triglycerides (TG) (p=0.053) levels at 12 weeks, compared to the baseline with significant improvements observed in C (TC: -15.4mg/dL, p=0.016); LDL: -13.6mg/dL, p=0.004) and lTRF (TC: -19.5mg/dL, p=0.002; LDL: -12.3mg/dL, p=0.031; TG: -19.8mg/dL, p=0.028) groups. Glucose metabolism and insulin resistance were improved in the eTRF group (HbA_1c: -0.3%, p=0.009; fasting insulin: -5.5mU/L, p=0.003; HOMA-IR: -1.8, p=0.049; QUICKI: +0.02, p=0.003; fasting glucose to insulin (FGI) ratio: +3.8, p=0.009).

Conclusion: A hypocaloric TRF diet could result in clinically significant weight loss (>5%), independently of the time of the day food is restricted, while limiting food consumption earlier in the day may lead to ameliorated glycemic control. Consequently, TRF could be an alternative and acceptable nutritional strategy in MASLD individuals.

Clinical Trial Registration Number: NCT05866744

Disclosure: S. Tsitsou: None.

27

A fasting-mimicking diet programme reduces liver fat and inflammation/fibrosis measured by MRI-derived biomarkers in patients with type 2 diabetes

E.L. van den Burg ¹, M.P. Schoonakker¹, P.G. van Peet¹, M.E. Numans¹, H. Pijl², H.J. Lamb³;

¹Department of Public Health and Primary Care, Leiden University Medical Centre, Leiden, Netherlands, ²Department of Internal Medicine, LUMC, Leiden, Netherlands, ³Department of Radiology, Leiden University Medical Centre, Leiden, Netherlands.

Background and aims: The aim of the present study was to assess effects of a fasting-mimicking diet (FMD) program for 5-consecutive days per month for twelve months in patients with type 2 diabetes on liver fat, measured by proton density fat-fraction (PDFF), and liver inflammation/fibrosis, measured by iron-corrected T1 (cT1). Furthermore, we aimed to assess whether changes in PDFF or cT1 are associated with changes in markers of metabolic control.

Materials and methods: Here we report secondary outcome measures of the Fasting In diabetes Treatment (FIT) trial, which was a randomised, controlled, assessor-blinded trial in which people with type 2 diabetes using metformin only and/or diet alone for glycaemic control were randomised to receive 5-consecutive day cycles of an FMD monthly as adjunct to usual care or usual care only for twelve months. Laboratory measurements and MRI were performed at baseline, 6 months and 12 months.

Results: Data were available of 89 participants who completed the baseline visits of the FIT trial, including MRI. Intention-to-treat analyses, using linear mixed models, revealed significant adjusted estimated treatment effects of the FMD on PDFF (-2.8%, 95% CI -4.7 to -0.8, p<0.01) and cT1 (-29.9 ms, 95% CI -51.8 to -8.0, p<0.01) at 12 months. Every percent decrease in PDFF was associated with a decrease in HbA1c of 0.75 mmol/mol (95% CI 0.51 to 1.00), fasting glucose of 0.14 mmol/L (95% CI 0.08 to 0.21), triglycerides of 0.04 mmol/L (95% CI 0.02 to 0.07) and total cholesterol of 0.04 mmol/L (95% CI 0.01 to 0.06). Every millisecond decrease in cT1 was associated with a decrease in HbA1c of 0.05 mmol/mol (95% CI 0.02 to 0.08), fasting glucose of 0.01 mmol/L (95% CI 0.00 to 0.02) and triglycerides of 0.00 mmol/L (95% CI 0.00 to 0.01).

Conclusion: Following an FMD program for 5-consecutive days per month for twelve months reduces both liver PDFF and cT1 MRI-derived biomarkers in patients with type 2 diabetes, indicating a reduction in liver fat and liver inflammation/fibrosis. The changes in PDFF and cT1 are associated with changes in HbA1c, fasting glucose and triglycerides. The change in PDFF was also associated with change in total cholesterol. Monthly cycles of an FMD appear to be a valuable adjunct to regular treatment of type 2 diabetes.

Clinical Trial Registration Number: NCT03811587

Supported by: Health~Holland, Top Sector Life Sciences & Health, the Dutch Diabetes Foundation and L-Nutra

Disclosure: E.L. van den Burg: Grants; The project was co-funded by Health~Holland, Top Sector Life Sciences & Health, the Dutch Diabetes Foundation and L-Nutra.

28

Effects of 12-week supplementation with coffee diterpene cafestol in healthy subjects with increased waist circumference: a randomised, placebo-controlled trial

F.D. Mellbye ¹, M.D. Nguyen¹, K. Hermansen², P.B. Jeppesen², Z.K. Al-Mashhadi¹, S. Ringgaard³, S. Gregersen¹;

¹Steno Diabetes Center Aarhus, Aarhus University Hospital, Aarhus, Denmark, ²Department of Endocrinology and Internal Medicine, Aarhus University Hospital, Aarhus, Denmark, ³The MR Research Centre, Aarhus University, Aarhus, Denmark.

Background and aims: Despite the established inverse correlation between coffee consumption and type 2 diabetes, the precise compounds responsible and their mechanisms of action remain unclear. Cafestol, a bioactive constituent of coffee, has exhibited glucose-lowering, insulin stimulating and insulin sensitivity improvement features in cell- and animal models. Most recently, cafestol appeared acutely to reduce glucose levels during glucose tolerance tests. However, the long-term impacts of cafestol on glucose metabolism in humans remain unexplored. This study aimed to evaluate the effects of purified cafestol on insulin sensitivity and adipose tissue distribution in healthy individuals at risk of developing type 2 diabetes.

Materials and methods: Forty healthy participants with abdominal obesity and heightened risk of type 2 diabetes were randomly assigned to receive either cafestol (6 mg) or placebo capsules twice daily for 12 weeks in a double-blinded, parallel design. Prior to and following the intervention, we conducted octreotide infusion insulin suppression tests to evaluate glucose disposal and insulin sensitivity via steady state plasma glucose (SSPG) levels. Additionally, MRI scans were performed to quantify visceral and subcutaneous adipose tissue volumes, while MRI-spectroscopy was utilized to assess liver fat content. Between-group differences post-intervention were analyzed using one-way ANCOVA, with adjustments made for baseline measures and sex.

Results: All participants completed follow-up visits. No significant between-group differences were observed in SSPG during hypoinsulinemic stage 1 (β_cafestol 0.18 mmol/l, 95%CI -0.21 - 0.56 mmol/l, p =0.4) or hyperinsulinemic stage 2 (β_cafestol 0.44 mmol/l, 95%CI -1.14 - 2.02 mmol/l, p=0.6). Visceral fat volume was significantly reduced in the cafestol group by 0,4L compared to placebo (β_cafestol -0.39 L, 95%CI -0.7 - -0.1 L, p=0.014). Concomitantly, body weight was significantly reduced in the cafestol group compared to placebo (β_cafestol -1.79 kg, 95%CI -3.45 - -0.13 kg, p=0.035). No significant between-group differences were detected in subcutaneous fat volume or liver fat fraction.

Conclusion: This study represents the first longer-term human intervention in healthy subjects with increased waist circumference comparing purified cafestol to placebo. While cafestol in small dosages (6 mg twice daily) for 12 weeks does not improve insulin sensitivity, we observed a significant visceral fat- and body weight reduction to cafestol compared to placebo. These observations are novel and hold important beneficial implications. Our findings warrant further investigation and cafestol may serve as a potential substance for weight and visceral fat reduction. To gain further insight future studies should apply higher cafestol dosages and/or longer treatment duration.

Clinical Trial Registration Number: NCT05672433

Supported by: Aarhus University, Steno Diabetes Center Aarhus, Helsefonden, Lægefonden / A.P. Møller Fonden and Civilingeniør Frode Nyegaard og Hustru’s Fond.

Disclosure: F.D. Mellbye: None.

29

Meal timing, its relation to metabolism, and heritability in German adult twins

O. Pivovarova-Ramich ^1,2, J. Vahlhaus^1,3, B. Peters^1,4, S. Hornemann¹, A.-C. Ost^1,5, M. Kruse^1,2, A. Busjahn⁶, A. Pfeiffer^2,1;

¹German Institute of Human Nutrition, Nuthetal, Germany, ²Charité – Universitätsmedizin Berlin, Berlin, Germany, ³University of Lübeck, Lübeck, Germany, ⁴German Center for Diabetes Research (DZD), München-Neuherberg, Germany, ⁵University of Leipzig, Leipzig, Germany, ⁶HealthTwiSt GmbH, Berlin, Germany.

Background and aims: Meal timing is a crucial factor influencing metabolic health which can be explained by the tight interaction between circadian clocks and metabolic homeostasis. Recent evidence suggests that meal timing pattern might contribute to obesity and diabetes risk, but this topic is still insufficiently studied. This work aimed (i) to investigate the link between individual meal timing pattern and diabetes-related metabolic traits as well as (ii) to explore the contribution of genetic and environmental factors to the meal timing architecture.

Materials and methods: The research was conducted in a German NUtriGenomic Analysis in Twins (NUGAT) cohort (n = 89) comprising 32 monozygotic and 12 dizygotic twin pairs with a BMI of 22.8 ± 2.73 kg/m² and median age of 25.0 years (IQR = 22.0). Glucose metabolism was assessed using the oral glucose tolerance test. Data on the meal timing architecture (eating window duration, caloric midpoint, start of the first and last meal, meal frequency etc.) were extracted from six-day food records. Chronotype, wake and sleep timing data were collected by the Munich chronotype questionnaire. Heritability of meal timing components was estimated by the ACE model, and bivariate correlation analyses were performed to assess their relation to metabolic traits.

Results: Correlation analysis revealed a number of associations between meal timing components and metabolic traits. Most association were found for the calorie midpoint calculated as a time point at which 50% of daily calories were consumed. Calorie midpoint positively correlated with BMI (r = 0.264, p = 0.013), waist (r =0.247, p = 0.007) and hip (r = 0.211, p = 0.049) circumference, fasting insulin (r = 0.245, p = 0.021) and negatively with Matsuda index of insulin sensitivity (β = -0.243, p = 0.023). The association of calorie midpoint with BMI and waist circumference remained significant after the model adjustment for gender and age (β = 0.276, p = 0.004 and β = 0.208, p = 0.010, respectively) and additional adjustment for energy intake and sleep duration (β = 0.256, p = 0.007 and β = 0.194, p = 0.016, respectively), whereas association with Matsuda index was lost in the last model. Further, the duration of meals was positively associated with visceral fat content and waist circumference, while the gap between meals showed negative associations with BMI and waist circumference. Heritability analysis estimated heritability of 12.5% for the caloric midpoint and of 35.9 % for the timing of the first meal to 41.8 % for the last meal. Meal timing components demonstrated a strong relation to the individual sleep-wake timing and chronotype, the heritability of which in turn ranged from 30.5% for chronotype to 40.3 % for wake and 51.6 % for sleep time.

Conclusion: Meal timing pattern is associated with diabetes-related metabolic traits and share a common genetic architecture with sleep behavior. Our findings help to differentiate (none-)modifiable factors driving individual meal timing which is important for development of effective nutritional strategies to combat obesity and diabetes.

Clinical Trial Registration Number: NCT01631123

Supported by: DFG RA 3340/4-1 to OP-R, project number 530918029; EASD Morgagni Prize 2020 to OP-R

Disclosure: O. Pivovarova-Ramich: None.

30

Liver fat as a key target in dietary management of type 2 diabetes and prediabetes using novel diet therapies: a randomised clinical trial (The NAFLDiet study)

U. Riserus ¹, M. Fridén¹, F. Rosqvist¹, M. Martinell¹, P.-O. Carlsson², L. Johansson³, H.-E. Johannson¹, F. Rorsman⁴, J. Vessby⁴, R. Landberg⁵, L. Lind⁴, M. Orho-Melander⁶, J. Kullberg⁷, H. Ahlström⁸;

¹Public Health & Caring Science, Uppsala University, Uppsala, Sweden, ²Medical Cell Biology, Uppsala University, Uppsala, Sweden, ³Antaros Medical, Gothenburg, Sweden, ⁴Medical sciences, Uppsala University, Uppsala, Sweden, ⁵Life Sciences, Chalmers University of technology, Gothenburg, Sweden, ⁶Dept of Clinical Sciences, Lund University, Malmo, Sweden, ⁷Department of Surgical Sciences, Uppsala University, Uppsala, Sweden, ⁸Surgical sciences, Uppsala University, Uppsala, Sweden.

Background and aims: Targeting liver fat by diet might improve glucose control and metabolic disorders of type 2 diabetes (T2D). We have previously shown that plant-derived polyunsaturated fatty acids (PUFA) in place of saturated fat reduces liver fat, whereas the effect of a novel “anti-lipogenic diet” replacing carbohydrates with PUFA has not yet been tested. In addition, a healthy Nordic diet (HND) low in saturated fat and rich in whole-grains has shown promising metabolic effects in overweight individuals, but its impact on liver fat and glycemic control in T2D is unknown. We aimed to investigate the effects on liver fat content (primary outcome) and related cardiometabolic risk factors after 12 months in individuals with T2D or prediabetes, comparing three different diets: a low carbohydrate diet high in polyunsaturated fat (LCPUFA) versus a healthy Nordic diet (HND), a LCPUFA diet versus usual care (UC) (recommended to follow the national dietary recommendations) and a HND versus UC.

Materials and methods: A three-arm parallel ad libitum randomized 1-year trial was conducted. Men and women (30-75 years) with T2D (53%) or prediabetes (n=148) were randomized to one of the three diets for 12 months. Apart from dietary advice, participants in all groups received key food items on a monthly/bimonthly basis. The fat sources in all diets, including the LCPUFA diet, were mainly plant-based. Liver fat was assessed by MRI. Dietary adherence was assessed by food records and dietary biomarkers (e.g. plasma fatty acids). Intention-to-treat effect was estimated by general linear models.

Results: Drop-out rate was very low (6%) and dietary adherence was high in all diet groups, as assewsed by food records and changes in plasma fatty acids (e.g. PUFAs). Liver fat (primary outcome) was reduced to a similar extent in the LCPUFA and the HND group compared to the UC group (-1.46%; 95% CI: -2.42, -0.51) and (-1.76 %; 95% CI: -2.96, -0.57), respectively. HbA1c decreased more in the HND compared to the other diets. Similar reductions in LDL-C were observed for the HND and the LCPUFA group compared to UC, but only the HND reduced triglycerides and CRP levels compared with UC. Despite all diets were ad libitum, body weight decreased more in the HND compared to both LCPUFA and UC (LCPUFA vs HND: 2.46 kg; 95% CI: 0.71, 4.21 and HND vs UC: -2.77 kg; 95% CI: -4.64, -0.90), respectively. No difference was observed between LCPUFA and UC: -0.31 kg; 95% CI: -2.18, 1.57). No significant differences between diets were shown for blood pressure, HDL-C or HOMA-IR.

Conclusion: A LCPUFA diet and a HND both effectively reduces liver fat as compared with UC, possibly through different mechanisms. A Healthy Nordic Diet however appears particularly clinically useful for the management of both T2D and NAFLD, considering the improvements in glycemic control, blood lipids, inflammation and sustained weight loss.

Clinical Trial Registration Number: NCT04527965

Supported by: Swedish Diabetes Foundation, Swedish Research Council Formas, Swedish Heart and Lung Foundation, Excellence of Diabetes Research in Sweden (EXODIAB),

Disclosure: U. Riserus: None.

OP 06 Maternal influences and prediction in type 1 diabetes

31

Is the relative protective effect of maternal vs paternal type 1 diabetes on risk of type 1 diabetes explained by differences in predisposing genes potentially induced by selective foetal loss?

G. Tapia ¹, A.-K. Rantala², C. Page¹, T. Skrivarhaug³, N. Lund-Blix⁴, K. Størdal⁵, L. Stene¹, HEDIMED investigators;

¹Norwegian Institute of Public Health, Oslo, Norway, ²University of Oulu, Oulu, Finland, ³Dept. of Pediatric, Diabetes Research Center Ullevål University Hospital, Oslo, Norway, ⁴Oslo University Hospital, Oslo, Norway, ⁵University of Oslo, Oslo, Norway.

Background and aims: The phenomenon that children of mothers with type 1 diabetes have lower risk of type 1 diabetes compared to children of fathers with type 1 diabetes is well known but the explanation remains elusive. Risk genotype-dependent fetal loss inducing differences in frequencies of predisposing genes in children is one proposed mechanism, which has not been investigated for non-HLA genes. We investigated whether the frequency of established type 1 diabetes susceptibility genes differed in offspring of mothers versus fathers with type 1 diabetes.

Materials and methods: The Norwegian Mother, Father and Child Cohort study has genotyped over 76 000 children born 2000-2009 and followed to 2021. Type 1 diabetes diagnosis was ascertained by linkage to nation-wide registries (n=461 children, 939 fathers, and 709 mothers).

Results: The odds ratio of type 1 diabetes conferred by maternal type 1 diabetes was confirmed to be lower than that for paternal diabetes (OR=6.5,95%CI:4.10-10.25 vs. OR=8.3,95%CI:6.98-11.41). The risk allele frequencies of the HLA DQ8 tag SNP rs7454108 and of the DQ2 tag SNP rs2187668 was similar in offspring of mothers with type 1 diabetes versus fathers with type 1 diabetes (26.2% vs 26.0%, and 20.0% vs 19.6%, respectively). Furthermore, the frequencies of the risk alleles of the INS (rs689: 77.9% vs 76.5%), PTPN22 (rs2476601, 84.7% vs 85.6%), and CTLA4 (rs3087243: 62.3% vs 62.1%) were similar in offspring of mothers compared to fathers with type 1 diabetes.

Conclusion: We found no support for a difference in frequencies of the most important type 1 diabetes susceptibility genes in offspring of mothers with type 1 diabetes compared to offspring of fathers with type 1 diabetes. Selective fetal loss depending on fetal genes predisposing to type 1 diabetes is therefore not a likely explanation for the differential association of maternal versus paternal type 1 diabetes.

Supported by: European Union Horizon 2020 research and innovation programme, grant agreement No 874864

Disclosure: G. Tapia: None.

32

Maternal type 1 diabetes confers long-term relative protection against type 1 diabetes in the offspring: results from five cohort studies

L.A. Allen ¹, P.N. Taylor¹, A. Carlsson², W.A. Hagopian³, E. Hedlund², A.V. Hill⁴, A.G. Jones⁵, J. Ludvigsson⁶, G.L. Mortimer⁷, S. Onengut-Gumuscu⁸, M.J. Redondo⁹, S.S. Rich⁸, K.M. Gillespie⁷, C.M. Dayan¹, R.A. Oram⁵;

¹Diabetes Research Group, Cardiff University, Cardiff, UK, ²Department of Clinical Sciences, Lund University, Lund, Sweden, ³Department of Pediatrics, Indiana University School of Medicine, Department of Medicine, University of Washington School of Medicine, WA, USA, ⁴Endocrinology and Diabetes, Royal Devon University Healthcare NHS Foundation Trust, Devon, UK, ⁵Department of Clinical and Biomedical Sciences, University of Exeter Medical School, Exeter, UK, ⁶Crown Princess Victoria Children’s Hospital and Division of Pediatrics, Department of Biomedical and Clinical Sciences, Linköping University, Linköping, Sweden, ⁷Translational Health Sciences, Bristol Medical School, University of Bristol, Bristol, UK, ⁸Center for Public Health Genomics, University of Virginia, Charlottesville, VA, USA, ⁹Pediatric Diabetes and Endocrinology, Texas Children’s Hospital. Department of Pediatrics, Baylor College of Medicine, Houston, TX, USA.

Background and aims: Type 1 diabetes is reportedly twice as common amongst the offspring of men compared with women with type 1 diabetes. This is believed to be the result of a relative protection afforded by maternal type 1 diabetes as compared with paternal type 1 diabetes.

Materials and methods: We performed a meta-analysis across 5 cohorts of individuals with type 1 diabetes (BOX, Better Diabetes Diagnosis, TrialNet Pathway to Prevention Study, Type 1 Diabetes Genetic Consortium, StartRight) (total n=11,475, diagnosed age 0-88 years), comparing the proportion with affected fathers versus mothers. We explored whether this comparison was altered by age at diagnosis, and timing of parental diagnosis relative to offspring birth. We compared the type 1 diabetes genetic risk score (GRS2) between individuals with mothers and fathers with type 1 diabetes.

Results: Almost twice as many individuals had an affected father compared with mother (overall OR 1.79 (95%CI 1.59,2.03), p<0.0001). The proportion of individuals with an affected father compared with an affected mother was higher both amongst individuals diagnosed with type 1 diabetes >18 years (OR 1.64 (95%CI 1.14,2.37), p<0.0001) and ≤18 years (OR 1.80 (95%CI 1.58, 2.05), p<0.0001). An excess of individuals with affected fathers was only observed if parental diagnosis was before offspring birth (OR 1.92 (95%CI 1.30,2.83) p=0.001 vs OR 1.28 (95% CI 0.94,1.75) p=0.12 after birth). Age at diagnosis and type 1 diabetes-free survival curves were comparable amongst offspring of affected fathers and mothers (p=0.31, 0.94 respectively). GRS2 was not significantly different between individuals with affected fathers versus mothers (p=0.31).

Conclusion: We report, in the largest study and first meta-analysis, evidence of relative maternal protection against type 1 diabetes in offspring as compared with paternal type 1 diabetes, evident even into adulthood. We provide novel insights into possible underlying mechanisms, indicating that in utero exposure to maternal type 1 diabetes appears critical.

Supported by: Wellcome Trust [Ref:225446/Z/22/Z]

Disclosure: L.A. Allen: None.

33

Maternal psychosocial stress and risk of childhood-onset type 1 diabetes

L.C. Stene ¹, A.-K. Rantala^1,2, M.C. Magnus¹, G. Tapia³, T. Skrivarhaug^4,5, N.A. Lund-Blix^5,1, K. Størdal^1,4, the HEDIMED Investigators;

¹Norwegian Institute of Public Health, Oslo, Norway, ²University of Oulu, Oulu, Finland, ³Epidemiology, Norwegian Institute of Public Health, Oslo, Norway, ⁴University of Oslo, Oslo, Norway, ⁵Oslo University Hospital, Oslo, Norway.

Background and aims: Psychological stress can influence the immune system and has been controversially linked to onset of type 1 diabetes but only recently investigated prospectively in birth cohorts. We aimed to investigate whether maternal psychosocial stress during pregnancy or the child’s first three years of life is associated with increased risk of type 1 diabetes in a cohort larger than all the previous studies in the field combined.

Materials and methods: We studied up to 91,000 mother-child pairs in the Norwegian Mother, Father and Child Cohort study born 2000-2009 and followed to 2021. Information on anxiety/depression (Likert scale range 1-4), negative life events (range 0-8) was obtained from repeated validated questionnaires. Type 1 diabetes diagnosis before 18 years of age was ascertained in the Norwegian Childhood Diabetes Registry (n=551). We estimated relative risks (aRR) with 95% CIs per unit increase, using binary log-linear regression adjusting for maternal type 1 diabetes, age, parity, education, smoking and pre-pregnancy BMI.

Results: Maternal symptoms of depression/anxiety (aRR=0.91,95%CI:0.71-1.17) and negative life events reported during pregnancy (aRR=1.04,95%CI:0.96-1.12), were not associated with type 1 diabetes. There were also no significant associations for these measures of stress assessed at child age 6 months, 18 months or 36 months (e.g. negative life events at 3 years: aRR=1.05, 95%CI: 0.94-1.16), or for other aspects of stress including job stress, social support and self-esteem or satisfaction with life.

Conclusion: In this so far largest cohort study to investigate maternal psychological stress in relation to childhood-onset type 1 diabetes, we found no significant association for any aspect of maternal stress at any time from pregnancy to child’s age three years.

Supported by: European Union Horizon 2020 research and innovation programme, grant agreement No 874864

Disclosure: L.C. Stene: None.

34

Trajectory of type 1 diabetes risk shifts after age 10 years between at-risk males and females

E.L. Templeman ¹, L. Ferrat², E. Rideout³, N. Thomas⁴, L. Allen⁵, M. Redondo⁶, C. Evans-Molina⁷, J. Sosenko⁸, R. Oram⁹, E. Sims¹⁰;

¹Department of Clinical and Biomedical Sciences, Exeter, UK, ²University of Geneva, Geneva, Switzerland, ³University of British Columbia, Vancouver, BC, Canada, ⁴Level 3, Exeter University, Exeter, UK, ⁵Diabetes Research Group, Cardiff, UK, ⁶Texas Children’s Hospital, Houston, TX, USA, ⁷Indiana Diabetes Research Center, Indiana University, Indianapolis, IN, USA, ⁸Miller School of Medicine, Miami, FL, USA, ⁹Institute of Biomedical and Clinical Science, University of Exeter Medical School, Exeter, UK, ¹⁰Indiana University School of Medicine, Indianapolis, USA.

Background and aims: In contrast to most autoimmune diseases, male sex is a risk factor for type 1 diabetes (T1D). This raises the hypothesis that either immune, metabolic, or other differences between sexes may impact risk or progression through stages of T1D. We aimed to assess the risk and rate of progression for individuals in the TrialNet Natural History study.

Materials and methods: We studied 235,765 relatives of people with T1D screened for autoantibody (AB) positivity in the TrialNet Pathway to Prevention (57% female). We estimated Kaplan-Meier survival curves per strata, and tested differences between strata with log-rank tests. Risk is stated as estimated 5 year risk [95% CI] for females and males respectively. Cox-proportional hazard models determined the effect of sex on the outcome, T1D, after adjusting for confounders. Partial dependence plots were used to display interactions between sex and age at screening with risk of disease.

Results: The proportion of individuals who screened positive for ABs was higher in males (females: 0.05 [0.049 - 0.051], males: 0.054 [0.053 - 0.056], p < 0.001). Of these individuals, males were more likely to screen positive for multiple ABs (females: 0.018 [0.017 - 0.018], males: 0.026 [0.025 - 0.027], p < 0.001). Absolute 5 year risk of progression to T1D in single AB positive individuals was significantly higher in males (females: 14% [12 - 16], males: 21% [18 - 23], p < 0.001), however similar risk was displayed across sexes when presenting with stage 1 (at least 2 ABs) (females: 38% [33 - 42], males: 38% [33 - 42]) or stage 2 (AB and dysglycaemia) (females: 57% [54 - 60], males: 59% [62 - 56]). Risk remains significantly higher in single AB positive individuals when adjusting for the presence of a first-degree relative with T1D and age (hazard ratio: 1.27 [1.07 - 1.50], p=0.007). A large decrease in 5 year T1D risk is displayed in females when screened, and autoantibody positive, before age 10 years as compared to after 10 years old (fig. 1). In contrast, a steady decline in 5-year T1D risk is displayed in males as age at screening increases.

Conclusion: Risk of T1D is significantly higher in males than females when presenting with a single autoantibody. Risk is similar between males and females in childhood, with the risk diverging at age 10. Risk in females then dramatically decreases whereas risk is sustained in males. This suggests sex appears to be linked with AB development indicating the importance of incorporating sex in the assessment of risk.

Supported by: Work was funded by NIH R01 DK121843-0, R01 DK124395, JDRF 3-SRA-2019-827-S-B. Erin Templeman is a PhD student funded by Randox.

Disclosure: E.L. Templeman: None.

35

Screening and monitoring for presymptomatic type 1 diabetes: the DiaUnion project

J. Hviid Klæbel ¹, A. Lind², S. Hamdan², M.N. Scherman², F. Kristensen³, J.C. Antvorskov¹, D.N. Agardh², F. Pociot¹;

¹Translational Type 1 Diabetes Research, Steno Diabetes Center Copenhagen, Herlev, Denmark, ²Department of Clinical Sciences, Lund University CRC, Malmö, Sweden, ³Medicon Valley Alliance, Copenhagen S, Denmark.

Background and aims: Type 1 diabetes (T1D), celiac disease (CD) and autoimmune thyroid disease (AITD), abbreviated as TRIAD diseases, are three common autoimmune diseases with onset in childhood and adolescence. CD and AITD share high-risk genes with T1D, thus a T1D patient will have an increased risk of additionally being diagnosed with these diseases. TRIAD diseases are predictable in the prodromal phases with the potential to stratify risk of disease by early autoimmune biomarkers. Autoantibodies (AAB) are key biomarkers for identification of at risk-individuals prior to appearance of clinical symptoms. The future goal is to establish a public screening program identifying people who will develop a TRIAD disease later in life. Thus, this study aimed to test a recruitment model, home capillary blood sampling and laboratory analyses.

Materials and methods: A total of 1,420 first degree relatives (FDR) (0-18 years, established biobank cohort of siblings) to newly diagnosed childhood/adolescent T1D patients in Denmark, and 2,271 children (two age cohorts; 7-8 and 13-14 years) from the general population in Skåne were included in the study. Participants from the general population in Skåne, received a home capillary kit for blood sampling with a video guide and afterwards returned the sample by public mail. The Antibody Detection by Agglutination-PCR (ADAP) multiplex technology as first-stage screening was compared with the gold standard Radio Binding Assay (RBA), for the analysis of 6 AAB (GADA, IAA, IA-2A, ZnT8, TPOA and tTGA). All samples were screened using the ADAP technology and AAB positivity confirmed by RBA.

Results: From the Danish samples, 192 (13.5%) were positive for a TRIAD AAB, 107 (7.5%) were positive for a T1D AAB, 55 (3.9%) were positive for a CD AAB, 55 (3.9%) were positive for a AITD AAB and 65 (4.6%) were positive for ≥2 T1D AAB. In the Swedish study, 19,593 children were invited, and of those 2,271 individuals were included. From these, 211 (9.3%) were positive for a TRIAD AAB, 60 (2.6%) were positive for a T1D AAB, 61 (2.7%) were positive for a CD AAB, 99 (4.4%) were positive for a AITD AAB and 14 (0.6%) were positive for ≥2 T1D AAB. For 2,271 samples ADAP and RBA were compared, and agreement ranged from 97.1% to 99.3%.

Conclusion: The study confirms the high prevalence of associated autoimmunity in FDRs, which supports the relevance of screening for multiple autoimmune diseases in this population. From our results, we can conclude that we have established an effective recruitment model with the participants being able to collect the screening sample at home. Thus, our model has the potential to be implemented as a large-scale screening program identifying high-risk individuals. Moreover, it was shown that the ADAP method was comparable to the conventional RBA for the detection of AAB associated with TRIAD and can be applied as a first-stage screening method.

Clinical Trial Registration Number: H-22053116

Supported by: EU Interreg ÖKS, RegH Denmark, Region Skåne, Lund University, SDCC, Barndiabetesfonden, NN Foundation

Disclosure: J. Hviid Klæbel: Grants; EU Interreg ÖKS, Capital Region of Denmark, Region Skåne, Lund University, Steno Diabetes Center Copenhagen, Barndiabetesfonden, Novo Nordisk Foundation.

36

Population screening for type 1 diabetes and celiac disease: autoimmunity screening for kids (ASK)

M. Rewers ¹, C. Geno Rasmussen¹, B. Frohnert¹, L. Yu¹, K. Simmons¹, A. Steck¹, H. O’Donnell¹, M. Stahl², E. Liu²;

¹Barbara Davis Center for Diabetes, University of Colorado, Aurora, CO, USA, ²Pediatrics, University of Colorado, Aurora, CO, USA.

Background and aims: Screening for autoantibodies (abs) can prevent significant morbidity associated with delayed diagnosis of type 1 diabetes (T1D) or celiac disease (CD). We are reporting on 7-year experience with screening for early T1D and CD in the U.S. general pediatric population children aged 1-17 y.

Materials and methods: Between 2017-2023, ASK has screened for pre-symptomatic T1D 34,110 children 1-17 y of age. Their age, sex, and race/ethnicity reflected Colorado population. Only 6% of participants had a first-degree relative with T1D. Most participants were screened in pediatric care clinics and outpatient labs. Islet abs to insulin, GAD, IA-2, and ZnT8 as well as transglutaminase abs to detect CD or celiac autoimmunity were measured using both standard radiobinding assay (RBA) and a more specific electrochemiluminescence assay (ECL) distinguishing predictive abs (high-affinity) from non-predictive ones (low-affinity). Screening positive children were invited to a confirmation visit that included abs by RBA and ECL, A1c, random blood glucose, and detailed review of symptoms. Children confirmed persistent for multiple islet abs or single abs by both methods received follow-up with education to prevent DKA, metabolic monitoring (CGM, OGTT, A1c), psychological support, and referrals to prevention trials or clinical services.

Results: At the initial screening or confirmation, 0.54% of the children had multiple islet abs, predicting a 70% 10-y risk of T1D and 0.42% had a single high-affinity ab (RBA+ and ECL+), predicting a 30% 10-y risk of T1D. Nearly 90% of the screening-detected high-risk children did not have a relative with T1D. Figure summarizes the results of screening, confirmation, and follow-up. During a 3-year follow-up, on average, 77 children were diagnosed with clinical diabetes: 7 at screening, 54 (29.2%) of those with multiple abs+, 9 (6.3%) of those with single high-affinity abs+, and 4 (0.6%) of those with single ab by one method, usually RBA. Of the 33,076 screening-negative children, 3 were diagnosed, on average 5 y later. In addition, ASK has identified 105 children at stage 2 T1D who have not yet progressed to stage 3. ASK re-screened 1,627 autoantibody negative children more than a year later; 2 have developed multiple and 4 a single high-affinity islet abs. Among children who completed confirmation, 4.5% had DKA at diagnosis, while 59% would be expected in our population. At the initial screening, ASK detected transglutaminase abs in 792 (2.3%) children who were not previously diagnosed with CD. Persistent celiac autoimmunity or CD was confirmed in 93% and 218 children have been newly diagnosed with CD; of those, only 11% had a family history of CD.

Conclusion: This population-based screening program for the two most common childhood autoimmune diseases has demonstrated high prevalence of pre-symptomatic T1D and CD and effectiveness in early diagnosis and prevention of DKA.

Supported by: JDRF 2-SRA-2022-1270-S-B, Helmsley Charitable Trust G-2109-04817

Disclosure: M. Rewers: Grants; JDRF, Helmsley Charitable Trust, Sanofi U.S., Janssen R&D, ProventionBio.

OP 07 Cardiovascular complications in type 1 diabetes: burden and mechanisms

37

Rapid vs slow plasma glucose decline has no impact on systolic function in individuals with type 1 diabetes

P.G. Hagelqvist ¹, K. Maytham¹, C. Rode¹, M. Sengeløv², U. Pedersen-Bjergaard³, T. Biering-Sørensen², F.K. Knop⁴, T. Vilsbøll¹, A. Andersen¹;

¹Clinical Research, Steno Diabetes Center Copenhagen, Herlev, Denmark, ²Department of Cardiology, Gentofte Hospital, Hellerup, Denmark, ³Department of Endocrinology and Nephrology, Nordsjællands Hospital Hillerød, Hillerød, Denmark, ⁴Center for Clinical Metabolic Research, Gentofte Hospital, Hellerup, Denmark.

Background and aims: Individuals with type 1 diabetes have an increased risk of cardiovascular disease. Rapid fluctuations of plasma glucose associated with insulin treatment may be a contributing factor. We investigated the impact of rapid vs slow plasma glucose decline on echocardiographic-derived measures of systolic function in individuals with type 1 diabetes.

Materials and methods: Twenty adult men with type 1 diabetes of which ten were well-regulated (mean (±SD) age 29.8±7.2 years, HbA1c 48.4±4.2 mmol/mol, BMI 23.7±2.1 kg/m²) and ten were dysregulated (age 31.0±8.4 years, HbA1c 72.3±7.5 mmol/mol, BMI 24.7±4.7 kg/m²) underwent two separate hyperinsulinaemic hyperglycaemic-euglycaemic clamp days in randomised order: Clamp Rapid involving a rapid plasma glucose decline from hyperglycaemia to euglycaemia (>0.15 mmol/l/min) and Clamp Slow involving a slow plasma glucose decline (<0.08 mmol/l/min). Cardiac function was evaluated by echocardiography during baseline hyperglycaemia (15 mmol/l) and again when euglycaemia (5 mmol/l) was reached.

Results: For both the well-regulated and dysregulated group, neither rapid nor slow plasma glucose decline induced changes in left ventricular ejection fraction nor global longitudinal strain from baseline hyperglycaemia, and, thus, no changes were observed between the two test days in any of the groups (Table 1).

Conclusion: Our findings suggest that a rapid plasma glucose decline from hyperglycaemia to euglycaemia has no impact on measures of systolic function in individuals with type 1 diabetes.

Clinical Trial Registration Number: NCT04800536

Supported by: DFF

Disclosure: P.G. Hagelqvist: None.

38

Effects of rapid vs slow plasma glucose decline on haemostasis in individuals with type 1 diabetes

P.G. Hagelqvist ¹, C.R. Andreasen¹, K. Maytham¹, F.K. Knop², P.I. Johansson³, U. Pedersen-Bjergaard⁴, A. Andersen¹, T. Vilsbøll¹;

¹Steno Diabetes Center Copenhagen, Herlev, Denmark, ²Center for Clinical and Metabolic Research, Gentofte Hospital, Hellerup, Denmark, ³Department of Clinical Immunology, Rigshospitalet, Copenhagen, Denmark, ⁴Nordsjællands Hospital Hillerød, Hillerød, Denmark.

Background and aims: People with type 1 diabetes are at increased risk of thrombosis compared to the general population, which to some extent may be explained by rapid and large plasma glucose fluctuations. In the present study, we investigated the acute effects of rapid vs slow plasma glucose declines on thrombelastography (TEG)-assessed haemostasis, hypothesising that rapid declines would induce coagulation activation, potentially aggravating the prothrombotic state of diabetes. Additionally, we explored the influence of glycaemic control on these effects.

Materials and methods: We included 20 adult men with type 1 diabetes in a 1:1 distribution of dysregulated (HbA1c ≥63 mmol/mol) and well-regulated diabetes (HbA1c ≤53 mmol/mol) in a randomised crossover study. The participants underwent two separate hyperinsulinaemic hyperglycaemic-euglycaemic clamp days with a rapid (>0.15 mmol/l/min) and slow (<0.08 mmol/l/min) plasma glucose decline, respectively. TEG, providing data of the entire process of coagulation activation (R-time), clot formation (α-Angle), clot strength (MA) and fibrinolysis (LY-30), was performed during hyperglycaemia (15 mmol/l) (baseline) and euglycaemia (5 mmol/l) on both test days. The effects of the plasma glucose decline rates on haemostasis were assessed in an overall analysis including all participants (n=20) and in a sub-group analysis with well-regulated (n=10) and dysregulated individuals (n=10) being analysed separately.

Results: In the overall analysis, a rapid plasma glucose decline induced an increased rate of clot formation (α-Angle) with a mean increase from baseline of 0.64 degrees [95% CI 0.05;1.23]. No changes in TEG parameters were observed during the day with a slow plasma glucose decline. Nevertheless, no significant difference in Δα-Angle value was observed between the two test days. In the sub-group analysis, a rapid plasma glucose decline resulted in an increased coagulation activation (decreased R-time from baseline of -1.03 minutes [-2.03;-0.03]) and an increased rate of clot formation (increase in α-Angle from baseline of 1.10 degrees [0.20;2.00]) in the sub-group with dysregulated diabetes. No differences in ΔR-time or Δα-Angle were observed between the two test days (Figure). For the sub-group with well-regulated diabetes, no changes in TEG parameters were observed within or between test days.

Conclusion: The present findings suggests that acute lowering of plasma glucose from hyperglycaemia to euglycaemia can induce acute procoagulant changes in haemostasis, particularly in individuals with dysregulated diabetes. Whether these findings can be explained by the plasma glucose decline rate per se or due to differences in plasma glucose levels remains unclear.

Clinical Trial Registration Number: NCT04800536

Supported by: DFF

Disclosure: P.G. Hagelqvist: None.

39

Coronary arteries assessment in asymptomatic high risk patients with type 1 diabetes using optical coherent tomography

M. Dubsky ¹, R. Roland², N. Marhefková¹, P. Wohlfahrt³, V. Karmazin², P. Novodvorsky¹, J. Kautzner², M. Haluzík¹, M. Pazdernik²;

¹Diabetes Centre, Institute for Clinical and Experimental Medicine, Prague, Czechia, ²Department of Cardiology, Institute for Clinical and Experimental Medicine, Prague, Czechia, ³Department of Preventive Cardiology, Institute for Clinical and Experimental Medicine, Prague, Czechia.

Background and aims: Data about the cardiovascular (CVD) risk of people with type 1 diabetes mellitus (T1DM) is lacking. Most of the CVD risk calculators were created for people with T2DM and simple non-invasive examinations like calcium score or carotid ultrasound are not part of guidelines. The aim of our study was to assess coronary arteries in high-risk T1DM patients by the use of optical coherent tomography (OCT) to estimate their real CVD risk.

Materials and methods: Sixty-two individuals with T1DM, duration of DM at least 10 years, age 30-67 years, without prior history of atherosclerotic CVD or target organ damage were examined by carotid ultrasound and calcium score over 7 months. Twelve subjects were recognized as high-risk defined as a calcium score above 400 and/or the presence of two or more plaques in carotids. Afterwards, those people were examined by coronary angiography and OCT. All subjects were characterized by STENO1 Risk Engine Calculator with mean STENO1 10-year risk score 25.7 %. Seven out of twelve (58.3%) people were treated with low-, 2/12 (16.7%) with moderate- and only 1/12 (8.3%) by high-intensity statin therapy. We assessed quantitative (maximal diameter stenosis, minimal lumen area [MLA]) and qualitative (thin-cap fibroadenoma [TFCA], very high-risk plaque [defined as TCFA with MLA < 3.5 mm², lipid arch > 180° and macrophages], macrophage accumulation [Figure-C], presence of thrombus or cholesterol cleft) OCT parameters.

Results: Mean age was 64.5±1.8 years, T1DM duration 36.1±11.5 years, HbA1c 58.5±8.8 mmol/mol and baseline LDL 2.3±0.6 mmol/l. Mean calcium score was 950±976 and mean number of carotid plaques was 2.8 ±1.1. Coronary angiography showed obstructive coronary artery disease in 5/12 (41.7%) people, one of them underwent percutaneous coronary intervention. TCFA was identified in 7/12 (58.3%) people [Figure-A] and very high-risk plaque was present in 4/12 (33.3%) people [B]. We saw intraluminal thrombus [D] and cholesterol crystal [E] in 3/12 (25%) people. Mean MLA at left anterior descending artery (LAD) was 3.18 mm^2. Pearson correlation coefficient revealed a strong correlation between OCT maximal diameter stenosis and calcium score at LAD (r=0.75, p=0.005). We observed no significant correlation between the number of plaques in carotids and TCFA at LAD.

Conclusion: Our study showed that people with T1DM without previous history or signs of CVD with high calcium score and carotid plaques had very severe OCT findings. We observed high-risk features like TCFA which is associated with a very high risk of plaque rupture and also a fatal CVD event. Therefore these people should be treated as very high-risk individuals.

Supported by: the project CarDia (Programme EXCELES, Project No. LX22NPO5104) - Funded by the European Union - Next Generation EU.

Disclosure: M. Dubsky: None.

40

Automated insulin delivery systems and vascular reactivity in type 1 diabetes: a prospective study

R. Lupoli, C.P. Petrosino, C. Rainone, G. Vallefuoco, I.L. Calcaterra, M.N.D. Di Minno, L. Bozzetto;

Federico II University of Naples, Naples, Italy.

Background and aims: Mechanisms underlying the increased cardiovascular risk in type 1 diabetes (T1D) are not completely known. Hyperglycemia, hypoglycemia, and glucose variability could all favor the development of atherosclerosis in T1D. Hybrid artificial pancreas (HAP) considerably improves glucose control and variability. Patients shifting from usual care to HAP represent a unique setting to test the impact of improved glucose control on vascular health. The aim of the present study was to evaluate the relation between glucose metrics and endothelial function and to assess the impact of 3-months use of HAP in patients with T1D.

Materials and methods: In a prospective cohort study we enrolled patients with ≥3-year T1D duration without macrovascular complications, regularly using continuous glucose monitoring (CGM), candidates to HAP. All enrolled subjects underwent ultrasound vascular assessment including flow mediated dilation (FMD), arterial stiffness, and intima-media thickness (IMT) before and after 3-month use of HAP. Time in range (TIR) 70-180 mg/dl, time above range (TAR) 181-250 mg/dl and >250 mg/dl, time below range (TBR) 69-54 mg/dl and <54 mg/dl, coefficient of variation (CV) as evaluated by CGM were recorded.

Results: A total of 24 T1D individuals (40% males, mean age 40 ± 18 years) were enrolled in the study. Nine patients (37%) were on multiple daily injections (basal-bolus) and 15 (63%) on insulin-pump (no automated systems). At baseline, FMD of brachial artery was 5.4 ± 3.5%, arterial stiffness was 5.5 ± 1.6 m/s, and IMT 0,6 ± 0.2 mm. A significant inverse correlation was found between FMD values and % TAR 181-250 mg/dl (rho = -0.538, p = 0,007). The assessment after 3-month HAP use (n = 13) showed a significant increase in TIR (from 52 ± 12 to 72 ± 17%, p < 0.001) and a significant reduction in TAR 181-250 mg/dl (from 30 ± 14 to 19 ± 11%, p = 0.005) and in glucose variability (from 36 ± 6% to 32 ± 4, p = 0.019). A significant improvement in FMD values (from 5.4 ± 3.5 to 9.3 ± 3.5%, p < 0.001) was observed after 3-month HAP while arterial stiffness and IMT remained substantially unchanged. The 3-month changes in FMD inversely correlated with changes in % TAR 181-250 mg/dl (rho = -0.566, p = 0,045).

Conclusion: In individuals with T1D vascular function was inversely associated with TAR. The use of HAP for 3 months induced a significant improvement in endothelial function associated with improvement in TAR. The benefits on long-term cardiovascular outcomes need to be investigated.

Disclosure: R. Lupoli: None.

41

Incidence and risk of myocardial infarction subtypes in type 1 diabetes

P. Smidtslund ^1,2, V. Harjutsalo^1,2, V. Thorn³, F. Jansson Sigfrids^1,2, P.-H. Groop^1,4, L.M. Thorn^1,5;

¹Folhälsan Research Center, Helsinki, Finland, ²Research Program in Clinical and Molecular Metabolism, University of Helsinki, Helsinki, Finland, ³Department of Cardiology, University of Helsinki and Helsinki University Hospital, Helsinki, Finland, ⁴Department of Nephrology, University of Helsinki and Helsinki University Hospital, Helsinki, Finland, ⁵Department of General Practice and Primary Health Care, University of Helsinki and Helsinki University Hospital, Helsinki, Finland.

Background and aims: Type 1 diabetes is associated with higher risk of myocardial infarction (MI). In the general population, the incidence of ST-elevation MI (STEMI) has decreased rapidly in comparison with non-ST-elevation MI (NSTEMI), but no studies have evaluated the incidence of MI subtypes in type 1 diabetes. Thus, our aim was to evaluate the incidence of MI subtypes in individuals with type 1 diabetes, and to assess the risk related to the presence of diabetic microvascular complications.

Materials and methods: Observational follow-up study of 4,217 individuals (51.8% men, median age 36.7 [28.1-45.6] years), with no prior history of MI or coronary revascularization, from the Finnish Diabetic Nephropathy Study. The baseline study visit included a comprehensive characterisation of the participants. The MIs during follow-up until death or end of 2017 were identified from the National Care Register for Health Care and death certificates, and all MIs were confirmed and classified based on ECG criteria from medical records and death certificates. We conducted Fine-Gray analyses to calculate subdivision hazard ratios (SHRs) for MIs, adjusted for traditional CVD risk factors. Each type of MI and death served as competing risk.

Results: We identified 449 (10.6%) incident MIs during a median follow-up time of 16.7 (IQR 13.7-18.7) years. The 15- and 20-year cumulative incidence rates for all MIs were 8.7% (95% CI 7.9-9.6) and 15.4% (12.0-19.1). The 20-year cumulative incidence of STEMI (n=84) was 2.4% (1.9-2.9), of NSTEMI (n=297) 10.9% (7.8-14.6), and 1.8% (1.4-2.4) for unclear MIs (n=68). Sex did neither affect the cumulative incidence of MIs nor its subtypes. In Fine-Gray analyses, severely increased albuminuria (SHR 2.01 [95% CI 1.47-2.74]), kidney replacement therapy (2.99 [2.06-4.36]), eGFR <60 ml/min per 1.73 m² (1.72 [1.28-2.30]), and severe retinopathy (1.58 [1.19-2.09]) were associated with higher risk of MI. Regarding STEMI, only eGFR <60 ml/min per 1.73 m² (2.15 [1.13-4.08]) was associated with higher risk. For NSTEMI, severely increased albuminuria (1.83 [1.25-2.69]), kidney replacement therapy (2.75 [1.75-4.34]), and severe retinopathy (1.54 [1.09-2.18]) were associated with higher risk, but eGFR <60 ml/min per 1.73 m² was not associated with higher risk. Moderately increased albuminuria was not associated with an elevated risk of any MI.

Conclusion: We observed a high cumulative incidence of confirmed MIs in individuals with type 1 diabetes, with NSTEMIs accounting for approximately two thirds. Diabetic kidney disease and diabetic retinopathy were associated with increased risk of all MIs and particularly NSTEMI. eGFR <60 ml/min per 1.73 m² increased the risk of STEMI. The risk of any MI was not increased in those with moderately increased albuminuria.

Supported by: Folkhälsan Research Foundation, Helsinki University Hospital state funding, Liv and Hälsa Society, Stockmann Foundation, Sigrid Jusélius Foundation

Disclosure: P. Smidtslund: Grants; Finnish Medical Society, Aarno Koskelon säätiö, The Finnish Foundation for Cardiovascular Research.

42

Whole genome sequencing reveals genetic variation associated with coronary artery disease in individuals with type 1 diabetes

A.A. Antikainen ^1,2, J. Haukka^1,2, A. Syreeni^1,2, S. Mutter^1,2, V. Harjutsalo^1,2, P.-H. Groop^3,2, N. Sandholm^1,2;

¹Folkhälsan Institute of Genetics, Folkhälsan Research Center, Helsinki, Finland, ²Research Program for Clinical and Molecular Metabolism, Faculty of Medicine, University of Helsinki, Helsinki, Finland, ³Abdominal Center, Nephrology, University of Helsinki and Helsinki University Hospital, Helsinki, Finland.

Background and aims: Diabetes increases risk of coronary artery disease (CAD) significantly, and the lifetime risk is particularly high among individuals with type 1 diabetes (T1D). Known CAD risk loci found within the general population increase CAD risk also among individuals with diabetes, but diabetes-specific genetic loci may also exist. CAD genetics have been studied in individuals with diabetes focusing on common genetic variants, but here we perform the first rare genetic variant (minor allele frequency, MAF≤1%) study for CAD in T1D.

Materials and methods: The study comprises 2,302 individuals with T1D from the Finnish Diabetic Nephropathy (FinnDiane) study. Participants were diagnosed with T1D before the age of 40 years and with insulin treatment initiated within one year from diagnosis. Whole-exome sequencing (WES) and whole-genome sequencing (WGS) data were available for 485 and 1,845 non-overlapping study participants, respectively. WES was performed with Illumina HiSeq2000 platform, and WGS either with Illumina HiSeq X or Illumina NovaSeq platform. We aligned WES with GATK4 pipeline and WGS with GraphTyper to GRCh38 human reference genome. A total of 250 and 728 individuals in the WES and WGS cohorts, respectively, have been diagnosed with CAD after their T1D diagnosis according to Finnish National Registry Data (ICD10: I20-I25). We required controls to be ≥35 years old and to have had diabetes for ≥15 years. We performed single variant and gene burden analysis with Firth’s correction using SAIGE 1.3.3; and meta-analysed WES and WGS results with METAL. For gene burden analysis, we aggregated putative loss-of-function (pLoF) variants within (1) low-frequency (MAF≤5%) and (2) rare (MAF≤1%) variant categories.

Results: No variant surpassed the genome-wide significance threshold (p-value<5×10^-8). However, we found one intronic variant on SLC45A4 to be suggestively associated with CAD in individuals with T1D (rs113455452, reference/alternative=G/A, MAF=1%, OR=0.088 [0.035-0.22], p-value=3.24×10^-7). SLC45A5 codes for the solute carrier family 45 member 4 membrane protein, which has been hypothesized to be involved in sucrose transport. No genes surpassed the multiple testing corrected significance threshold in gene burden analysis (p-value<2×10^-5). However, the burden of low-frequency pLoF variants on one gene, DPYD, were suggestively associated with CAD (MAF≤5%, OR=0.95 [0.92-0.97], p-value=1.54×10^-4). DPYD codes for a pyrimidine catabolism enzyme, whose inhibition has been shown to be protective against liver fibrosis.

Conclusion: We identified two suggestive CAD risk-loci: SLC45A5 intron and LoF variants on DPYD, in individuals with T1D. For replication and to increase the statistical power, we will next perform corresponding analyses in UK BioBank individuals with probable T1D (~2,000 individuals).

Supported by: Novo Nordisk Foundation, Sigrid Jusélius Foundation, The Finnish Foundation for Cardiovascular Research

Disclosure: A.A. Antikainen: None.

OP 08 Genetics and epigenetics in type 2 diabetes

43

Blood-based epigenetic markers associate with future macrovascular events in individuals with type 2 diabetes

S. Garcia-Calzon ^1,2, A. Maguolo^2,3, A. Perfilyev², M. Maziarz⁴, E. Ahlqvist⁵, I. Gonçalves^6,7, C. Ling²;

¹Dept. of Food Science and Physiology /Center for Nutrition Research, University of Navarra, Pamplona, Spain, ²Epigenetics and Diabetes Unit, Lund University, Malmö, Sweden, ³Department of Surgery, Dentistry, Pediatrics, and Gynecology, University of Verona, Verona, Italy, ⁴Bioinformatics Unit, Lund University, Malmö, Sweden, ⁵Genetics and Diabetes Unit, Lund University, Malmö, Sweden, ⁶Cardiovascular Research Translational Studies, Lund University, Malmö, Sweden, ⁷Department of Cardiology, Skåne University Hospital, Malmö, Sweden.

Background and aims: Type 2 diabetes (T2D) increases future risk for macrovascular events, as heart attacks and strokes. However, available tools for prediction of macrovascular events in individuals with T2D are suboptimal. We therefore aimed to discover blood-based epigenetic biomarkers associated with future macrovascular events in individuals with T2D.

Materials and methods: Individuals from the ANDIS and ANDIU Swedish cohorts were included. DNA methylation was analyzed in blood from 752 newly-diagnosed individuals with T2D free of macrovascular events, among whom 102 developed events (incident macrovascular events) during ~7 years. Multivariable-adjusted weighted-Cox regression models were performed, and ROC curves were generated.

Results: DNA methylation of 461 sites was associated with incident macrovascular events in newly diagnosed individuals with T2D (Hazard ratios range from 0.44 to 2.28 per 1 SD increase in methylation, FDR<0.05). These sites were annotated to 422 genes, including KCNQ1, PSEN1 and/or GAD1, which previously have been associated with diabetes or cardiovascular function. We found that DNA methylation sites associated with future macrovascular events generated an AUC of 0.82 (95% CI: 0.77 to 0.86). Interestingly, a ROC curve with only clinical risk factors measured at inclusion in ANDIS and ANDiU, including age, sex, HbA1c, BMI, smoking, diabetes medication, lipid-lowering medication and antihypertensives, generated a worse AUC of 0.69 (95%CI: 0.63 to 0.75) for prediction of future macrovascular events. We also calculated established well-known cardiovascular risk scores (UKPDS, Framingham’s, ASCVD and MESA) and these had a poor predictive capacity of future macrovascular events in our cohort with AUCs ranging from 0.57 to 0.60.

Conclusion: Blood-based epigenetic markers were robustly associated with future macrovascular events in newly-diagnosed individuals with T2D, indicating their potential for further development in precision medicine for diabetes.

Supported by: EFSD/Novo Nordisk Foundation Precision Diabetes Medicine Award Programme

Swedish Research Council, Region Skåne, Swedish Foundation for Strategic, H2020-Marie-Curie, Juan de la Cierva-Incorporación, Swedish Diabetes Foundat

Disclosure: S. Garcia-Calzon: None.

44

An updated meta-analysis of epigenome-wide association study of incident type 2 diabetes reveals 234 novel CpG sites

R.D. Triatin ^1,2, R. Hillary³, E. Fraszczyk⁴, A. Spijkerman⁵, M. Luijten⁵, R.E. Marioni³, I.M. Nolte¹, C.H.L. Thio^1,6, H. Snieder¹;

¹Department of Epidemiology, University Medical Center Groningen, Groningen, Netherlands, ²Department of Biomedical Sciences, Faculty of Medicine, Universitas Padjadjaran, Sumedang, Indonesia, ³Centre for Genomic and Experimental Medicine, Institute of Genetics and Cancer, University of Edinburgh, Edinburgh, UK, ⁴Institute of Bioorganic Chemistry, Polish Academy of Sciences, Poznan, Poland, ⁵Centre for Prevention and Health Services Research, National Institute for Public Health and the Environment, Bilthoven, Netherlands, ⁶Department of Population Health Sciences, Institute for Risk Assessment Sciences (IRAS), Utrecht University, Utrecht, Netherlands.

Background and aims: DNA methylation (DNAm) is known to be associated with type 2 diabetes (T2D) However, previous studies were limited by small sample size or relied on cross-sectional T2D data. For improved precision, we integrated longitudinal results from Generation Scotland cohort into our previous epigenome-wide association study (EWAS) of incident T2D (iT2D), and aimed to elucidate biological mechanisms.

Materials and methods: We performed individual EWAS of iT2D using DNAm measured before T2D onset in Generation Scotland (N=12,982) and in contributing cohorts from our previous meta-EWAS (N=3,200; KORA, KORA2, Doetinchem, EPIC-Norfolk, and ESTHER). We conducted a fixed-effect inverse-variance weighted meta-analysis of EWASs (meta-EWAS). Quality checks on individual cohort-level results were performed prior to meta-EWAS, and inflation was controlled using BACON. Significant differentially methylated CpG sites (p-value <1.1x10^-7) were taken forward to bi-directional two-sample Mendelian randomization (MR) using summary statistics from the GoDMC database, cross-sectional T2D GWAS, and mQTL analysis within the Generation Scotland. Sensitivity analyses using MR Egger, weighted-median, steiger filtering, and genetic colocalization were performed. Finally, we explored pathway enrichment of CpGs significant in MR by look-ups of biological pathways in Gene Ontology (GO), KEGG, and Reactome databases, using methylGSA package. MR estimates and pathway enrichment were considered significant at false discovery rate (FDR) <0.05.

Results: We uncovered 234 novel CpGs (p-value < 1.1x10^-7) and validated 75 previously identified CpGs with more precise estimates compared to our previous study. MR provided evidence of DNAm at 15 CpGs as causally upstream of T2D, 79 as downstream of T2D, and 2 bidirectional CpGs. The 15 upstream CpGs were enriched in the KEGG insulin signaling pathway (FDR: 0.006). The 79 downstream CpGs were enriched in 44 GO pathways, including 12 lipid-related pathways such as response to estradiol (FDR: 9x10^-72), lipid transporter activity (FDR: 0.003), and phospholipid transporter activity (FDR: 0.003). No significant enrichment was found in the Reactome database.

Conclusion: This study enhanced the understanding of the role of DNAm in iT2D by identifying 234 novel differentially methylated CpGs. We found suggestive evidence for causality of DNAm at 96 CpGs, primarily as a consequence rather than causative of T2D. These CpGs were enriched in key pathways related to insulin and lipids, aligned with the known biological mechanisms of T2D providing valuable insights for future research and therapeutic interventions.

Supported by: Research was supported by a PhD scholarship awarded to RDT by LPDP (no. 202108220807760).

Disclosure: R.D. Triatin: None.

45

Allelic expression imbalance in pancreatic islets indicates parent-of-origin effects on type 2 diabetes risk

J. Andersson ¹, G. Hatem¹, M. Maziarz¹, M. Lehtovirta², L. Hakaste³, T. Tuomi⁴, R. Prasad B⁵;

¹Lund University Diabetes Centre (LUDC), Malmö, Sweden, ²University of Helsinki, Helsinki, Finland, ³Samfundet Folkhälsan i svenska Finland rf, Helsinki, Finland, ⁴Helsinki University Central Hospital, Helsinki, Finland, ⁵Lund University Diabetes Center, Malmö, Sweden.

Background and aims: Type 2 diabetes (T2D) is characterized by defective insulin secretion by pancreatic β-cells due to increased insulin resistance. While genome-wide association studies (GWAS) for T2D have identified over 1100 SNPs, the precise mechanisms by which these variants contribute to disease risk remain unclear, with eQTLs shedding some light on the link. To investigate the cis-regulatory potential in islets with focus on putative cause or consequence of T2D, we leveraged gene expression data from islets exposed to chronic or acute hyperglycaemia, comparing them to normoglycemia. Our objective was to identify allelic expression imbalance (AEI), indicating the unequal expression of maternal and paternal alleles. Additionally, we aimed to investigate subsequent parental-specific influences of genetic variation on islet function and T2D risk.

Materials and methods: We utilized two islet RNAseq datasets. The first comprised data from T2D (n=33) and non-diabetic islet donors (n=155), while the second dataset included data from hyperglycemic (n=14) and normoglycemic donor islets (n=31) treated with physiological (5.5mM) and high (18.9mM) glucose. AEI and differential AEI analyses were conducted using the ASEP R package. Beta cell expression was looked up in published single-cell RNAseq data. Selected genes were further examined for parent-of-origin effects (POE) on insulin secretion and T2D risk using GWAS from the Botnia family cohort (n~9200). Significance thresholds were set at FDR<0.05.

Results: 3188 out of 11,161 genes (28%) showed significant AEI in 188 donor islets, 1624 out of 7807 (20.8%) genes in T2D, whereas 3020 of 10808 (27.9%) in non-T2D. Differential AEI analysis between islets exposed to normal or high glucose identified 137 genes out of 8063 (1.7%), 22 of 4518 (0.5%) genes in hyperglycemic and 63 of 7024 (0.9%) genes in normoglycemic samples. To identify potential genes causing T2D, we focused on those exhibiting AEI changes between T2D and non-diabetic donors while remaining unchanged during acute hyperglycemia. Among the initial 2098 candidates, 705 genes exhibited a negative correlation with insulin expression and were also expressed in β-cells. These included PDLIM5, GLB1L2 and RRAGD. GABARAPL1 and DHFR2 showed change in AEI response to high glucose in both chronic and acute hyperglycemia and were considered potential consequential genes. Among the known imprinted genes, 99 showed islet expression of which 25 showed AEI in chronic but not in acute hyperglycemia, suggesting they could be causal genes. After additional scrutiny for negative correlation with insulin secretion and beta cell expression, we identified 9 potential candidates which included GLIS3 and GRB10. No consequential genes were found among the imprinted ones. GLIS3 variants showed significant parent-of-origin effects (POE) with the maternal allele associating with a decrease in insulin secretion (CIR).

Conclusion: Many genes show AEI, some potentially due to POE. These novel genes provide clues into T2D pathophysiology and demonstrate the significance of AEI in disease relevant tissues.

Supported by: Swedish Research Council, Director Albert Påhlsson Foundation, Crafoord Foundation

Disclosure: J. Andersson: None.

46

Beta cell dysfunction polygenic risk associates with risk of gestational diabetes and postpartum abnormal glucose intolerance in Chinese

G. Yu ^1,2, C.H.T. Tam^1,2, C.K. Lim^1,2, M. Shi^1,2, C.C. Wang^3,4, J.C.N. Chan^1,4, W.H. Tam^3,5, M. Weedon⁶, R. Oram⁶, R.C.W. Ma^1,2;

¹Department of Medicine and Therapeutics, The Chinese University of Hong Kong, Hong Kong, Hong Kong, ²Li Ka Shing Institute of Health Sciences, Hong Kong, Hong Kong, ³Department of Obstetrics and Gynaecology, The Chinese University of Hong Kong, Hong Kong, Hong Kong, ⁴The Chinese University of Hong Kong, Hong Kong, Hong Kong, ⁵CUHK Medical Centre, Hong Kong, Hong Kong, ⁶University of Exeter Medical School, Exeter, UK.

Background and aims: Gestational diabetes mellitus (GDM) affects around 15% of pregnancies and is associated with 7-fold risk of type 2 diabetes (T2D). Whether genetic factors driving pathophysiological pathways for T2D contribute towards GDM is unclear.

Materials and methods: 963 Chinese women with singleton pregnancy (median age = 31.3 years) recruited as part of the Hyperglycemia and Adverse Pregnancy Outcome (HAPO) study are included. All women underwent 75-g oral glucose tolerance test (OGTT) at gestational 24-32 weeks and GDM was diagnosed according to WHO 2013 criteria. Postpartum abnormal glucose tolerance (AGT) at seven-year follow-up, was defined as impaired fasting glucose, impaired glucose tolerance or diabetes mellitus. Partitioned T2D polygenic risk scores (pPRSs) were calculated for each subject using two sets of newly developed hard- and soft-clustering pPRSs to assess individual genetic risk of specific pathophysiological pathways underlying T2D. Total T2D PRSs in hard- and soft-clustering were respectively summed by all single pPRSs. The associations between pPRSs and GDM at baseline and AGT at seven-year postpartum follow-up were examined using logistic regression adjusted for age at assessment and top four genetic principal components. Bonferroni correction was used to correct for multiple comparisons.

Results: 149 (15.5%) women were diagnosed with GDM at baseline. The pPRS for Beta cell+PI (OR (95% CI) = 1.76 (1.46, 2.12), P = 3.1×10^-9), related to beta cell dysfunction and higher proinsulin, and Total T2D PRS (OR (95% CI) = 1.83 (1.52, 2.23), P = 4.6×10^-10) from hard-clustering were strongly associated with higher risk of GDM (Figure 1A). The pPRS for Beta cell 1 (OR (95% CI) = 1.33 (1.12, 1.60), P = 0.002) from soft-clustering, related to beta cell dysfunction and glucose homeostasis, was also associated with higher risk of GDM (Figure 1B). At the seven-year follow-up, 129 (13.9%) women developed AGT. Total T2D PRS (OR (95% CI) = 1.91 (1.57, 2.35), P = 2.0×10^-10) and multiple pPRSs from hard-clustering were significantly associated with higher risk of AGT. Among the clusters, Beta cell+PI again showed the strongest association (OR (95% CI) = 1.53 (1.26, 1.86), P = 1.6×10^-5). We did not find significant results for soft-clustering pPRSs.

Conclusion: T2D genetic burden increased the risk of GDM and postpartum AGT. Genetic risk of beta cell dysfunction appears to be the predominant pathophysiological pathway underlying GDM and postpartum AGT in Chinese.

Supported by: The RGC of HKSAR (CUHK 473408, 471713, 14118316, 14118718, 14102719), CUHK Grants Matching Scheme and Internationalisation Faculty Mobility Scheme

Disclosure: G. Yu: None.

47

Loss-of-function mutations of the GIP receptor impair GIP-mediated insulin secretion

H. Kizilkaya ¹, M. H. Michaelsen², H. Reeh², C. Pless³, F. Koefoed-Hansen⁴, C. Nielsen⁵, S. Madsbad⁶, B. Hartmann⁷, M. Christensen⁸, F.K. Knop⁹, T. Hansen¹⁰, J.J. Holst¹¹, N. Grarup¹², M.M. Rosenkilde⁷, L. Gasbjerg¹³;

¹University of Copenhagen, Copenhagen, Denmark, ²Copenhagen University Hospital, Hellerup, Denmark, ³Copenhagen University Hospital, Copenhagen, Denmark, ⁴Center for Clinical Metabolic Research, Hellerup, Denmark, ⁵Center for Clinical Metabolic Research, Gentofte Hospital, Hellerup, Denmark, ⁶Hvidovre Hospital, Hvidovre, Denmark, ⁷⁸Clinical Pharmacology, Bispebjerg Hospital, Copenhagen, Denmark, ⁹Clinical Metabolic Physiology, Gentofte Hospital, Hellerup, Denmark, ¹⁰Novo Nordisk Foundation Center for Basic Metabolic Research, University of Copenhagen, Copenhagen, Denmark, ¹¹Biomedical Sciences, University of Copenhagen, Copenhagen, Denmark, ¹²The Novo Nordisk Foundation Center for Basic Metabolic Research, University of Copenhagen, Copenhagen E, Denmark, ¹³University of Copenhagen, Copenhagen N, Denmark.

Background and aims: GIP is a gut-derived hormone released in response to meal intake and involved in glucose, lipid and bone metabolism. Genetic variants of the GIP receptor (GIPR) are associated with cardiometabolic traits incl. low adiposity, altered blood pressure, and low bone mineral density. Here, we investigated the incretin effect and beta cell secretory responses to GIP and GLP-1 infusions in carriers with in vitro-confirmed loss-of-function (LoF) GIPR variants from the Danish population-based Inter99 cohort, and matched control individuals.

Materials and methods: In a randomized, single-blind, placebo-controlled, crossover study, we subjected 10 heterozygous carriers of LoF GIPR variants (6 men, 4 women, mean (±SD); age 63±5.8y) and 10 age- and gender-matched controls (mean age 66±5.4y) to five experimental days: a 75-gram oral glucose tolerance test (OGTT), an isoglycaemic intravenous (i.v.) glucose infusion (IIGI), and three 10 mmol/l hyperglycaemic clamps with i.v. infusion of GIP (4 pmol/kg/min), GLP-1 (1 pmol/kg/min) and saline, respectively.

Results: There was no difference in the incretin effect as assessed by C-peptide/glucose ratios during OGTT and IIGI between carriers of LoF variants and controls (59±21% vs 59%±19%, p>0.05). The lowest incretin effects (7% and 45%) were observed in LoF variant carriers with the E288G GIPR variant. During hyperglycaemic clamp with GIP infusion, carriers of LoF GIPR variants had a reduced beta cell responsiveness to glucose, assessed by (C-peptide/glucose ratios,) compared to controls (mean C-peptide/glucose ratio (30-150 minutes) for LoF 275 + 34 pmol/mmol vs controls 346 + 41 pmol/mmol, Fig. 1a). There was no difference in GLP-1-induced insulin secretion (Fig. 1b).

Conclusion: Compared with matched healthy controls, LoF GIPR variant carriers exhibit impaired beta cell response to GIP infusion during hyperglycaemic clamp.

Clinical Trial Registration Number: NCT06194955

Disclosure: H. Kizilkaya: None.

48

Change in telomere length and its implications in type 2 diabetes: the Fremantle Diabetes Study Phase II

M.L.H. Huang ¹, C.-H. Yang¹, W. Davis², K. Peters³, A.J. Jenkins¹, T. Davis²;

¹Diabetes and Vascular Medicine Laboratory, Baker Heart and Diabetes Institute, Melbourne, Australia, ²Medical School, University of Western Australia, Fremantle, Australia, ³Proteomics International, Perth, Australia.

Background and aims: Several studies support associations between relative telomere length (rTL), a novel biomarker of biological aging, and type 2 diabetes (T2D). Our aims were to characterise the association between baseline subject demographics and the 4-year change in rTL (∆rTL), and to investigate whether ∆rTL predicts all-cause mortality in adults with T2D in the Australian community-based Fremantle Diabetes Study Phase II (FDS2).

Materials and methods: FDS2 T2D participants (n=819) had peripheral blood collected at two timepoints, baseline (2008-11) and Year 4 (2012-15), a mean±SD 4.2±0.4 years apart, for rTL assay by quantitative PCR (intra- and inter-assay CVs 0.56% and 2.69%, respectively). ∆rTL was expressed as percentage change per annum as previously reported and categorised as: Shortening (<-2.69%), Unchanged (-2.69% to +2.69%), and Lengthening (>+2.69%). Multiple logistic regression identified significant baseline determinants of membership of the Shortening vs. Lengthening groups, excluding those Unchanged. ∆rTL as a i) continuous, and ii) categorical variable was added to the most parsimonious Cox regression model of baseline predictors of all-cause mortality to end-2021.

Results: ∆rTL was normally distributed (mean±SD 8.4±21.0%; range -93.8% to 115.7%) with 25.5% (n=209) Shortening, 10.5% (n=86) Unchanged, and 64.0% (n=524) Lengthening. Telomere Shortening vs. Lengthening was independently associated with age, sex, central obesity, use of lipid-modifying medication, ln(serum γ-glutamyl transferase), and ln(serum bilirubin) (see Table). During 5,945 person-years (7.3±2.1 years) of follow-up after the Year 4 assessment, 213 (26%) participants died (30% in both the Shortening and Unchanged groups, 24% in the Lengthening group, log-rank test P=0.180). After adjusting for conventional baseline risk factors, ∆rTL, as either a continuous or categorical variable, did not improve prediction of all-cause mortality (HR (95% CI): 1.00 (0.99, 1.01) for a 1% change; 0.90 (0.56, 1.44) for Shortening and 0.89 (0.58, 1.36) for Lengthening).

Conclusion: In representative adults with T2D, only one quarter of subjects had their telomeres shorten over four years. ∆rTL was significantly associated with key demographic and cardiometabolic variables but was not an independent prognostic indicator. rTL change was associated with sex, CHD, lipid-lowering drug use, smoking status, liver function tests, and central adiposity. Some similarities exist with a similar study in the FinnDiane Type 1 diabetes cohort. The addition of ∆rTL to traditional risk factor models does not enhance mortality prediction in T2D likely due to common modulating factors.

Supported by: Australian NHMRC Project Grants

Disclosure: M.L.H. Huang: None.

OP 09 Drugs you thought you knew all about, but what else do they do?

49

The effect of sodium-glucose cotransporter-2 inhibitors on graft survival and cardiorenal outcomes in patients with diabetes and heart transplantation

C.-M. Hwu ¹, F.-S. Yen², J.-Y. Huang³, C.-C. Hsu⁴, W.-Y. Cheng¹, J.C. Wei³;

¹Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan, Province of China, ²Dr. Yen’s Clinic, Taoyuan, Taiwan, Province of China, ³Chung Shan Medical University, Taichung, Taiwan, Province of China, ⁴National Health Research Institutes, Miaoli, Taiwan, Province of China.

Background and aims: To compare the risk of graft survival, cardiovascular events, dialysis, and all-cause mortality between the use and non-use of sodium-glucose cotransporter-2 inhibitors (SGLT2i) for the treatment of diabetes mellitus in patients with heart transplantation.

Materials and methods: In this study, we analyzed the TriNetX collaborative network data. We identified 6494 patients who underwent heart transplantation between January 1, 2015, and December 31, 2022. From these patients, 1063 pairs of SGLT2i users and non-users were selected based on their respective propensity scores. We used the Kaplan-Meier method and Cox proportional hazards models to assess differences in the risk of outcomes between the two groups.

Results: In the matched groups, SGLT2i users had a reduced risk of heart transplant failure and rejection (hazard ratio [HR]: 0.873, 95% confidence interval [CI]: 0.774-0.985), dialysis (HR: 0.566, 95% CI: 0.385-0. 833), all-cause hospitalizations (HR: 0.822, 95% CI: 0.739-0.916), and all-cause mortality (HR: 0.767, 95% CI: 0.627-0.938) than non-users. However, there were no significant differences between the two groups in the risk of post-transplant sepsis or infection (HR: 0.891, 95% CI: 0.739-1.075), heart failure exacerbation (HR: 0.915, 95% CI: 0.733-1.144), and ischemic heart disease (HR: 1.044, 95% CI: 0.939-1.161).

Conclusion: This multicenter cohort study showed that heart transplant recipients with diabetes mellitus treated with SGLT2i had a significant reduction in the risk of graft failure, need for dialysis, all-cause hospitalization, and mortality compared to those not treated with SGLT2i.

Supported by: funded by Chung Shan Medical University Hospital (CSH-2022 -A-023)

Disclosure: C. Hwu: None.

50

Efficacy and safety of canagliflozin in children and adolescents with type 2 diabetes: a multicentre, randomised, double-blind, parallel-group, phase 3 trial

S.R. Ali ¹, U. Nadgir², J. Gogate¹, W. Shaw¹, S. Fonseca¹, J. Antunes¹;

¹Janssen Research & Development, LLC, Raritan, NJ, USA, ²Center of Excellence in Diabetes and Endocrinology, Sacramento, CA, USA.

Background and aims: Treatment options for children with type 2 diabetes mellitus (T2DM) are limited due to frequent treatment failure and poor adherence to injectables. Canagliflozin (CANA), a sodium glucose co-transporter 2 inhibitor, has been shown to improve glycemic control in adults with T2DM. This global study assessed the efficacy and safety of CANA versus placebo (PBO) on glycemic control in children with T2DM.

Materials and methods: This randomized, double-blind, PBO-controlled, 2-arm, parallel-group, multicenter (104 sites in 10 countries) study enrolled children aged ≥10 to <18 years with inadequate glycemic control (HbA1c ≥6.5% to ≤11%) who were on diet and exercise alone or in combination with metformin monotherapy, insulin monotherapy, or metformin ± insulin. Study participants received CANA or PBO for 52 weeks (26-week core treatment period followed by 26 weeks of extended treatment; at Week 13, participants on CANA 100 mg with HbA1c ≥7% and eGFR ≥60 mL/min/1.73 m² were rerandomized 1:1 to CANA 100 mg/PBO or CANA 300 mg/PBO). The primary objectives were to assess the effects of CANA versus PBO on HbA1c at Week 26 in the overall population, as well as in the subset on background metformin ± insulin, and the safety and tolerability of CANA. Key secondary efficacy objectives included change from baseline in HbA1c at Week 52 in all participants; fasting plasma glucose (FPG); proportion of participants with HbA1c <7.5%, <7%, and <6.5% at Weeks 26 and 52; proportion of participants requiring rescue therapy; and time to rescue therapy.

Results: Baseline characteristics were similar between the CANA (n=84) and PBO (n=87) arms (overall: mean age, 14.3 years; 68.4% female; mean HbA1c, 8.0%). A statistically significant reduction from baseline in HbA1c was seen with CANA versus PBO at Week 26 (least square mean difference [95% CI]: -0.76% [-1.25, -0.27]; P=0.002). The HbA1c reduction was observed as early as Week 6 and maintained up to the end of trial. HbA1c trends were similar in the subset on background metformin ± insulin. FPG was significantly improved with CANA versus PBO at Weeks 26 and 52. At Week 52, a larger proportion of CANA versus PBO participants achieved HbA1c <7% (39/84 [54.9%] vs 17/87 [22.7%]) and <6.5% (26/84 [36.6%] vs 9/87 [12.0%]). Fewer participants on CANA (12%) versus PBO (46%) received rescue medications, and those on CANA had a longer time to rescue therapy. Changes in Tanner stage characteristics appeared similar between groups. Treatment-emergent adverse events (TEAEs) were reported in 65/84 (77.4%) and 65/87 (74.7%) participants receiving CANA and PBO, respectively; serious TEAEs were reported in 8/84 (9.5%) and 5/87 (5.7%) participants, respectively. One participant in each group had an AE leading to discontinuation. None of the serious TEAEs or AEs leading to discontinuation were considered related to CANA. The incidence of hypoglycemia TEAEs was higher with PBO versus CANA.

Conclusion: In children aged ≥10 to <18 years with T2DM, CANA led to clinically meaningful HbA1c reduction versus PBO in all study participants and those on metformin ± insulin. CANA was generally well tolerated, with a safety profile similar to that observed in the adults. CANA may offer a new treatment option for this age group with T2DM.

Clinical Trial Registration Number: NCT03170518

Supported by: This study was supported by Janssen Research & Development, LLC.

Disclosure: S.R. Ali: Employment/Consultancy; Janssen Research & Development, LLC.

51

Sodium-glucose cotransporter 2 inhibitor use and risk of dementia and parkinson’s disease among patients with type 2 diabetes

H. Kim ^1,2, G. Biessels³, M. Yu^1,4, N. Hong^1,2, Y.-H. Lee^1,5, B.-W. Lee^1,2, E. Kang^1,2, B.-S. Cha^1,2, M. Lee^1,2;

¹Internal Medicine, Yonsei University College of Medicine, Seoul, Republic of Korea, ²Institute of Endocrine Research, Seoul, Republic of Korea, ³UMC Brain Center, University Medical Centre Utrecht, Utrecht, Netherlands, ⁴Department of Internal Medicine, Yonsei University College of Medicine, Seoul, Republic of Korea, ⁵Yonsei University College of Medicine, Institute of Endocrine Research, Seoul, Republic of Korea.

Background and aims: Despite the mechanistic potential of sodium-glucose cotransporter 2 inhibitor (SGLT2i) to attenuate neurodegeneration, the efficacy of SGLT2i in neurodegenerative disorders among patients with type 2 diabetes is not well-established. This population-based cohort study aimed to investigate the association of SGLT2i use with risks of neurodegenerative disorders, in particular Alzheimer’s disease (AD), vascular dementia (VaD), and Parkinson’s disease (PD) in patient with type 2 diabetes.

Materials and methods: A total of 1,348,362 participants aged ≥ 40 years, diagnosed with type 2 diabetes, and who started antidiabetic drugs from 1 September 2014 and 31 December 2019 were evaluated, using the National Health Insurance Service Database. Propensity score matching (1:1; SGLT2i to other oral antidiabetic drugs [OAD]) produced a cohort of 358,862 Primary outcomes were the incidence of AD, VaD and PD. Secondary outcomes were all-cause dementia (AD, VaD, and other dementia) and composite of all-cause dementia and PD.

Results: From the 358,862 participants analyzed (mean [SD] age, 57.8 [9.6] years; 57.9% male), 6,837 incident dementia or PD events occurred. Regarding the primary endpoints, SGLT2i use was associated with reduced risks for AD (aHR, 0.81 [95% CI 0.76─0.87]), VaD (aHR, 0.69 [95% CI 0.60─0.78]), and PD (aHR, 0.80 [95% CI 0.69─0.91]). Use of SGLT2i was associated with a 21% lower risk for the all-cause dementia (adjusted HR [aHR], 0.79 [95% CI 0.69─0.90]) and 22% lower risk all-cause dementia and PD than use of other OADs (adjusted HR [aHR], 0.78 [95% CI 0.73─0.83]) with a 6-month drug use lag period. The effect of SGLT2is on these neurodegenerative disorders was not affected by age, sex, body mass index, blood pressure, glucose, lipid profiles, kidney function, health behaviors, Charlson comorbidity index, diabetic complications, comorbidities, and medications, yielding consistent results in sensitivity analyses.

Conclusion: In this nationwide population-based study, SGLT2i use significantly reduced the risks for neurodegenerative disorders in patients with type 2 diabetes independent of various factors including comorbidities and bioclinical parameters.

Disclosure: H. Kim: None.

52

Cardiovascular effectiveness of empagliflozin and glucagon-like peptide-1 receptor agonists combination therapy in adults with type 2 diabetes

P. Htoo ¹, D.J. Wexler², H. Tesfaye¹, S. Schneeweiss¹, R. Glynn¹, C. Shay³, N. Schmedt³, L. Koeneman⁴, J.M. Paik¹, E. Patorno⁵;

¹Brigham and Women’s Hospital, Boston, MA, USA, ²Diabetes Unit, Department of Medicine, Massachusetts General Hospital (MGH) Diabetes Research Center, Boston, MA, USA, ³Boehringer Ingelheim, Ingelheim am Rhein, Germany, ⁴Eli Lilly, Bad Homburg vor der Höhe, Germany, ⁵Brigham and Women’s Hospital, Boston, USA.

Background and aims: Due to the demonstrated cardiovascular benefits of empagliflozin, a sodium-glucose cotransporter 2 inhibitor (SGLT2i), and glucagon-like peptide-1 receptor agonists (GLP-1RA), clinical guidelines recommend their use in combination. However, current evidence is scarce on whether such combination therapy has additive cardiovascular benefits compared to either therapy alone. We aimed to compare the cardiovascular effectiveness of patients initiating combination therapy with empagliflozin-GLP-1RA vs. empagliflozin-dipeptidyl peptidase-4 inhibitors (DPP-4i).

Materials and methods: Using Medicare, Optum, and Marketscan databases (2014-2022), we identified patients with type 2 diabetes aged ≥18 years (≥65 years in Medicare), who initiated empagliflozin and had not used SGLT2i, GLP-1RA (including GLP-1RA for weight loss and tirzepatide), or DPP-4i in the previous 6 months. We further identified patients who went on to initiate GLP-1RA or DPP-4i while on treatment with empagliflozin. Follow-up started on the day after initiation of GLP-1RA or DPP-4i in an on-treatment scheme until discontinuation of either drug class or empagliflozin, a gap in insurance coverage, death, or end of the study period. After 1:1 propensity score matching, we estimated hazard ratios (HR) and rate differences (RD) per 1,000 person-years, evaluating a composite of myocardial infarction (MI) or stroke, hospitalization for heart failure (HHF), and mortality.

Results: We identified 19,551 matched pairs of patients initiating combination therapy with empagliflozin-GLP-1RA vs empagliflozin-DPP-4i. After matching, rates of MI-stroke were similar between treatment groups and had similar risk of MI-stroke [HR 1.00 (0.77, 1.29); RD: -0.04 (-3.08, 2.99) ]. Patients on empagliflozin-GLP-1RA had lower risks of HHF vs. those on empagliflozin-DPP-4i, [HR: 0.82 (0.71, 0.96) ); RD: -6.47 (-11.55, -1.42) ]. When we defined HHF using diagnosis codes only in the primary hospital discharge position, HR was 0.49 (0.34, 0.71) and RD was -4.34 (-6.61, -2.16) . The risk of all-cause mortality was also lower in patients on empagliflozin-GLP-1RA vs. empagliflozin-DPP-4i: [HR 0.61 (0.46, 0.81); RD: -4.79 (-7.56, -2.09) ].

Conclusion: In patients with type 2 diabetes in US clinical practice, patients initiating empagliflozin-GLP-1RA had lower risks of HHF and all cause-mortality compared to those initiating empagliflozin-DPP-4i, while the risks of MI-stroke were similar. Patients using combination therapy with both empagliflozin and GLP-1RA may have additive benefits compared to using empagliflozin alone.

Supported by: BI

Disclosure: P. Htoo: Grants; American Diabetes Association.

53

Metformin reduces fasting blood glucose independently of endogenous glucose production by promoting glycolysis and glucose clearance in well-controlled type 2 diabetes

T. Sarabhai ¹, T. LaMoia², S. Friesl³, M. Jonuscheit⁴, V. Schrauwen-Hinderling³, K. Petersen⁵, G. Shulman⁵, M. Roden³;

¹AG Energy, German Diabetes Center, Düsseldorf, Germany, ²Department of Internal Medicine, Yale School of Medicine, New Haven, CT, USA, ³German Diabetes Center, Düsseldorf, Germany, ⁴AG Metabolic Imaging, German Diabetes Center Düsseldorf, Düsseldorf, Germany, ⁵Section of Endocrinology, Yale University School of Medicine, New Haven, CT, USA.

Background and aims: Metformin is known to improve glycemic control in individuals with poorly-controlled type 2 diabetes (T2D) mainly by reducing rates of endogenous glucose production (EGP). Here, we investigated the yet unknown mechanisms by which metformin reduces blood glucose levels in well-controlled individuals with T2D.

Materials and methods: We examined hepatic glucose and energy metabolism in individuals with T2D who were near-normoglycemic (N-T2D; n=10, age 59±7 years, BMI 29±2 kg/m², HbA1c 7±1%) or hyperglycemic (H-T2D; n=8, age 62±2 years, BMI 28±1 kg/m², HbA1c 9±1%). All participants were studied with (M+; 1000 mg bid) or without (M-) metformin treatment for two weeks. Rates of substrate turnover were determined using [²H₇]glucose, [3-¹³C]lactate and [¹³C₃]glycerol, while hepatic energy metabolism was monitored by ¹H/³¹P/¹³C-magnetic resonance spectroscopy.

Results: In N-T2D and H-T2D, metformin treatment reduced fasting plasma glucose concentrations by ~30% (p<0.005 vs. M-), which could mainly be attributed to an ~25% increase in rates of glucose clearance (p<0.05 vs. M-). Increased glucose clearance could in turn be attributed to increased glycolysis as reflected by increased lactate production (+14%, p<0.01 vs. M- in N-T2D) resulting in increased plasma lactate concentrations (~40%; p<0.01 vs. M-). Only in H-T2D, but not in N-T2D, metformin also reduced rates of EGP (-13%; p<0.05 vs. M-). In both groups, metformin treatment resulted in an ~30% increase in hepatic glycogen content (p<0.05 vs. M-) and a ~25% decrease in hepatic ATP content (p<0.05 vs. M-). In contrast, metformin treatment had no effect on hepatic lipid content.

Conclusion: Independent of the degree of hyperglycemia, metformin lowers fasting plasma glucose concentrations by stimulating glucose clearance and promoting glycolysis. In addition, metformin reduces fasting plasma glucose concentration by reducing rates of EGP, but only in individuals with poorly-controlled T2D.

Disclosure: T. Sarabhai: None.

54

Impact of timing of metformin administration on the plasma lactate response to intraduodenal glucose infusion in type 2 diabetes

C. Xie, P. Iroga, Y. Sun, W. Huang, M.J. Bound, J. Grivell, K. Jones, M. Horowitz, C.K. Rayner, T. Wu;

The University of Adelaide, Adelaide, Australia.

Background and aims: Lactate, long regarded as a metabolic waste product, has emerged as a regulator of a number of metabolic processes, including enhancement of glycogen synthesis, inhibition of lipolysis, and suppression of appetite. Accordingly, increases in circulating lactate and its metabolites have been suggested to contribute to the metabolic benefits of metformin (e.g. weight loss). We have recently reported that metformin, when administered 30-60 min before an intraduodenal glucose infusion, is more effective in lowering the glycaemic response than when given at the onset of infusion, in metformin-treated type 2 diabetes (T2D). We have now evaluated whether the timing of metformin administration affects the plasma lactate response to intraduodenal glucose.

Materials and methods: Sixteen subjects with relatively well controlled T2D by metformin monotherapy (14 males, age 69.9 ± 1.9 years, BMI 28.7 ± 1.0 kg/m², HbA1c 6.6 ± 0.1%, duration of known diabetes 10.4 ± 2.6 years) were evaluated on 4 separate days in a double-blind, randomised, crossover design. Participants on immediate-release metformin (n = 11) were instructed to withhold their morning dose on each study day. Those taking extended-release metformin (n = 5) withheld the evening dose before, and the morning dose on, each study day. On each day, metformin (1000 mg) dissolved in 50 mL 0.9% saline, or 0.9% saline only (control), was administered at t = -60, -30, or 0 min via a nasoduodenal catheter, followed by an intraduodenal glucose infusion (3 kcal/min) between t = 0 - 60 min. ‘Arterialised’ venous blood was sampled every 30 min for measurements of plasma lactate. The plasma lactate incremental area under the curve between t = 0 to 120 min (iAUC_0-120min) was calculated using the trapezoidal rule and compared using repeated measures ANOVA. Data are means ± SEM.

Results: All subjects tolerated the protocol well. Prior to intraduodenal glucose infusion (t = -60 to 0 min), plasma lactate levels remained stable and did not differ between study days. In response to intraduodenal glucose infusion, plasma lactate increased on all study days. Compared with control, administration of metformin at t = 0, -30 and -60 min augmented the iAUC_0-120min for plasma lactate by 71%, 99% and 152% respectively (P < 0.05 each). The rise in plasma lactate was greater when metformin was administrated at t = -60 min than at t = -30 and 0 min (P < 0.001 and = 0.01 respectively).

Conclusion: In well-controlled metformin-treated T2D, metformin increases the plasma lactate response to an enteral glucose load, and this effect is much greater when metformin is given at an interval before glucose administration. Our observations suggest that the timing of metformin intake has a profound impact on postprandial lactate production. Further studies are warranted to clarify the potential clinical implications.

Clinical Trial Registration Number: ACTRN12621000878875

Supported by: The Hospital Research Foundation

Disclosure: C. Xie: None.

OP 10 Hypothalamic regulation of systems metabolism

55

Prolactin act on POMC neurons to regulate lipid metabolism of liver

Y. Zhu, Y. Bi;

Department of Endocrinology, Affiliated Drum Tower Hospital, Medical School of Nanjing University, Nanjing, China.

Background and aims: There is a negative association between prolactin (PRL) and the presence of metabolic-associated fatty liver disease (MAFLD), effects of PRL on lipid metabolism are primarily mediated by prolactin receptor (PRLR), but the specific organs and regulatory mechanisms involved are unclear. The latest research indicates that the central nervous system (CNS) plays an important regulatory role in liver lipid metabolism, and PRLRs are widely distributed in the CNS. However, the mechanism by which the CNS regulates hepatic lipid metabolism remains unclear. Therefore, this study aims to investigate whether PRL improves liver lipid metabolism through neuromodulation and to elucidate the specific neural circuits and mechanisms involved.

Materials and methods: The PRL systemic knockout mouse model was used to elucidate the role of PRL in hepatic lipid metabolism. I.c.v. prolactin was used to determine whether PRL improves hepatic lipid metabolism through central pathways. PRLR whole-brain immunofluorescence screening and electrophysiological experiments were used to elucidate the specific neurons targeted by PRL. The POMC neuron-specific PRLR knockout mouse model was used to reveal the role of PRL in neuromodulation of hepatic lipid metabolism. PRV and cre-dependent tracing virus were used to reveal the neural circuits through which PRL centrally improves hepatic lipid metabolism.

Results: PRL-KO mice are more prone to liver lipid deposition when induced by high-fat-diet (HFD), while intracerebroventricular (i.c.v.) injection of PRL can reduce the degree of liver lipid deposition in PRL-KO mice. Whole brain immunofluorescence (IF) experiments indicate high expression of PRLR in the arcuate nucleus (ARC) of the hypothalamus. Electrophysiological experiments show that PRL mainly activates POMC neurons in ARC. Neural circuit tracing experiments indicate that PRLR-positive POMC neurons project to the liver. Conditional knockout of PRLR in POMC neurons makes mice more susceptible to liver lipid deposition when induced by HFD.

Conclusion: Our research has uncovered a new role of PRL as a neural regulator, controlling POMC neuron activity and thus preventing MAFLD through neuromodulation. Revealing the neural circuits and molecular mechanisms through which the CNS regulates hepatic lipid metabolism, providing evidence for exploring central targets for the treatment of MAFLD.

Supported by: National Natural Science Foundation of China

Disclosure: Y. Zhu: Grants; This work was supported by grants from the National Natural Science Foundation of China (82030026).

56

Influence of insulin sensitivity on food cue evoked brain response in children

L. Semeia ¹, R. Veit¹, S. Zhao¹, S. Luo^2,3, B. Angelo^2,3, A.L. Birkenfeld^1,4, H. Preissl^1,4, A.H. Xiang⁵, S. Kullmann^1,4, K.A. Page^2,3;

¹Institute for Diabetes Research and Metabolic Diseases of the Helmholtz Center Munich at the University of Tübingen, German Center for Diabetes Research (DZD), Tübingen, Germany, ²Division of Endocrinology, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA, ³Diabetes and Obesity Research Institute, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA, ⁴Department of Internal Medicine, Division of Endocrinology, Diabetology and Nephrology, Eberhard Karls University Tübingen, Tübingen, Germany, ⁵Department of Research and Evaluation, Kaiser Permanente Southern California, Pasadena, CA, USA.

Background and aims: Insulin resistance in childhood is a significant risk for developing type 2 diabetes and related co-morbidities later in life. Research in adults indicates that insulin resistance impacts brain functions related to eating behavior and metabolic control. Yet, the relationship between brain activity related to food cues and peripheral insulin sensitivity is not well understood in children.

Materials and methods: We included 53 children (36 girls), aged 7-11, and assessed their peripheral insulin sensitivity (ISI) through an oral glucose tolerance test. Using functional magnetic resonance imaging, we measured brain responses before and after glucose ingestion. We compared brain activity and functional connectivity (FC) during a food cue task, involving high-caloric food and non-food images, between children with low and high ISI. Food cue evoked brain response (FCR) was calculated by contrasting the response between high-caloric and non-food pictures. Using a general linear model, FCR was evaluated by a full-factorial design with ISI, sex, and glucose intake (before vs. after) as factors, and age and BMIz as covariates. Statistical significance was set at p<0.05 family-wise error corrected for multiple comparisons at a whole-brain cluster level.

Results: Independent of glucose intake, children with lower ISI had increased FC between the insula and nucleus caudate, and decreased connectivity between the insula and mid temporal cortex compared to those with higher ISI when evaluating high caloric food cues. Sex differences were found in insular connectivity based on both prandial state and ISI. No significant differences were found between children with high compared to low ISI in overall brain responses to food cues.

Conclusion: Children with lower ISI showed distinct patterns of brain connectivity, particularly in insular functional connections, which is vital for gustatory and reward processes. When exposed to high caloric food cues, children with low ISI showed stronger insular communication to reward regions and lower communication to cognitive related regions. These differences may influence eating behavior and future risk of developing obesity and diabetes.

Supported by: ADA (#1-14-ACE-36); NIDDK (R03DK103083; R01DK116858); NIH (UL1TR001855); BMBF/DZD (01GI0925)

Disclosure: L. Semeia: None.

57

Regulation and mechanism of hypothalamic MST1 signalling on energy homeostasis and glucose/lipid metabolism

M. Yang ¹, L. Li²;

¹Department of Endocrinology, The Second Affiliated Hospital, Chongqing Medical University, Chongqing, Chongqing, China, ²The Key Laboratory of Laboratory Medical Diagnostics in the Ministry of Education and Department of Clinical Biochemistry, College of Laboratory Medicine, Chongqing Medical University, Chongqing, China.

Background and aims: Mammalian Sterile-20-like kinase 1 (MST1) is a serine-threonine kinase composed of 487 amino acids and is an important component of the "Hippo" signaling pathway. Previous studies have shown that MST1 is extensively involved in metabolic diseases related to obesity, aging, and insulin resistance. However, the role of MST1 in regulating metabolism in the central nervous system remains unclear. Therefore, this study aimed to explore whether MST1 is involved in the regulation of energy metabolism in the hypothalamus and its molecular mechanism.

Materials and methods: We injected AAV9-GFP/MST1 into the arcuate nucleus (ARC) of the hypothalamus of C57BL/6J mice to observe the metabolic phenotype and the expression of central neuron peptides related to energy metabolism in the mouse hypothalamus. To further explore the relationship between MST1 in POMC neurons and energy metabolism, we constructed mouse models with specific knockout and overexpression of MST1 in POMC neurons and examined their metabolic phenotype. POMC-FoxO1 KO mice generated by crossing POMC-Cre mice with FoxO1^fl/fl mice were used to determine the involvement of FoxO1 in hypothalamic MST1 signaling. Additionally, we used chemogenetic viruses to activate neurons in specific brain regions to further understand the mechanism of action of MST1 in the brain.

Results: We found that in WT mice fed a high-fat diet (HFD), overexpression of MST1 in the ARC reduces POMC expression. In addition, the injection of AAV9-MST1 into the arcuate nucleus of the hypothalamus significantly increased food intake, body weight, and blood glucose level; worsened insulin resistance; and promoted hepatic gluconeogenesis and lipid synthesis. Immunofluorescence staining showed that overexpression of MST1 in the ARC led to an increase in the level of forkhead box protein O1 (FoxO1), while other metabolic-related signaling pathways remained unchanged. Furthermore, specific knockout of MST1 in POMC neurons promoted energy expenditure, improved systemic metabolism, and reduced the level of FOXO1 in the ARC. In contrast, selective overexpression of MST1 in POMC neurons resulted in reduced energy expenditure and exacerbated metabolic dysfunction and was accompanied by downregulation of FoxO11 expression. Specific knockout of FoxO1 in POMC neurons counteracted the inhibitory effect of MST1 on POMC expression, thereby suppressing the metabolic phenotype changes caused by MST1 overexpression. Chemgenetical activation of POMC neurons in the ARC significantly reversed the promoting effect of MST1 overexpression on feeding and the inhibitory effect of MST1 overexpression on anal temperature in POMC neurons.

Conclusion: MST1 in the hypothalamus inhibits the expression of POMC neurons through the FoxO1-POMC pathway, thereby affecting whole-body energy metabolism.

Supported by: National Natural Science Foundation (82370852 and 82270853), Scientific and Technological Research Program of Chongqing Municipal Education Commission

Disclosure: M. Yang: None.

58

A hypothalamus-liver-skeletal muscle axis controlled by JNK1 and FGF21 mediates olanzapine-induced insulin resistance in an intraperitoneal treatment in male mice

V. Ferreira ¹, C. Folgueira², A. Montes¹, A.B. Hitos¹, R.J. Davis³, G. Sabio⁴, P. Rada¹, Á.M. Valverde¹;

¹Instituto de Investigaciones Biomédicas Sols-Morreale (IIBM), Madrid, Spain, ²CNIC, Madrid, Spain, ³Chan Medical School, University of Massachusetts, MA, USA, ⁴CNIO, Madrid, Spain.

Background and aims: Olanzapine (OLA), a widely used second-generation antipsychotic (SGA), causes metabolic dysfunctions in patients. Recently, we demonstrated that the oral treatment of male mice with OLA induces weight gain and hepatic steatosis. However, mice receiving OLA via intraperitoneal (i.p.) lost weight without steatosis, an effect related to higher OLA levels reaching the hypothalamus. Since clinical studies have shown insulin resistance in patients under OLA treatment, we investigated the impact of the i.p. treatment on peripheral insulin sensitivity as well as potential effects of the inhibition of Protein Tyrosine Phosphatase 1B (PTP1B), a negative regulator of insulin signaling.

Materials and methods: Wild-type (WT) and PTP1B-deficient (PTP1B-KO) mice, a preclinical model protected against insulin resistance, were treated with OLA (10 mg/kg) via i.p. for 8 weeks or received a single intrahypothalamic injection (15 nmol). Parameters related to glucose homeostasis were evaluated. Mechanistic studies were conducted in vagotomized mice and in mice lacking JNK in hypothalamus or liver and mice overexpressing hepatocyte FGF21.

Results: OLA i.p. treatment induces systemic insulin resistance (p<0.001, n=4) and pyruvate intolerance (p<0.05, n=16-24), and attenuates insulin signaling in both liver (p<0.05, n=4) and skeletal muscle (p<0.05, n=4), an effect that concurs with enhanced JNK phosphorylation (p<0.01, n=4-5) and IRS1 serine 612 phosphorylation (p<0.001, n=5-7) in the liver. Similar diminished peripheral insulin response was found 48 h after a single OLA intrahypothalamic injection that was preceded by the activation of both hepatic JNK (p<0.01, n=7-9) and IRS1 serine phosphorylation (p<0.01, n=7-10). Deletion of either hepatic or hypothalamic JNK1 or surgical vagotomy prevented OLA-induced impairments in insulin responses in liver and skeletal muscle. Importantly, downregulation of hepatic Fgf21 was found in mice treated with OLA via i.p. (p<0.01, n=5-8) or receiving an intrahypothalamic injection (p<0.05, n=7) and this decrease was prevented by deletion of hypothalamic JNK1. A step further, in mice with FGF21overexpression in the liver, OLA-induced insulin resistance was abolished only in skeletal muscle, but not in the liver, suggesting that OLA-induced deficits in insulin response in the skeletal muscle are secondary to the hepatic phenotype. Neither insulin resistance nor pyruvate intolerance were found in OLA-treated PTP1B-KO mice highlighting the inherent protection against insulin resistance in this model.

Conclusion: Despite of the beneficial effect of OLA administered via i.p. in avoiding weight gain, it induces peripheral insulin resistance through a hypothalamus-liver axis controlled by hypothalamic JNK1 activation which, in turn, activates hepatic JNK via the vagus nerve. This axis results in hepatic insulin resistance and a reduction of both hepatic and circulating FGF21 that ultimately impairs insulin response in the skeletal muscle. Importantly, the protection conferred by PTP1B deficiency against OLA-induced insulin resistance strongly suggests that targeting PTP1B might prevent metabolic comorbidities in patients under OLA treatment in a personalized manner.

Supported by: PID2021-122766OB-100 by MCIN/AEI/10.13039/501100011033 and H2020 Marie Sklodowska-Curie ITN-TREATMENT (Grant Agreement 721236, European Commission)

Disclosure: V. Ferreira: None.

59

In vivo phenotyping of glucose-dependent insulinotropic polypeptide receptor neurocircuitries in distinct hypothalamic nuclei

N.S. Figueredo Burgos ¹, A. Roberts¹, A. Adriaenssens¹, F. Reimann²;

¹UCL, London, UK, ²Dept. of Clinical Biochemistry, University of Cambridge, Cambridge, UK.

Background and aims: The glucose-dependent insulinotropic polypeptide receptor (GIPR) is extensively expressed throughout the brain and is necessary for the therapeutic effects of GIP/GLP-1 agonists in lowering body weight. Nonetheless, uncertainty remains about the signalling pathways mediating this effect. Studies comparing neurocircuitries engaged by Gipr neurons in different brain regions highlight the heterogeneity of the central GIPR signalling axis and the need to better define and dissect them. This study aimed to explore whether Gipr neurons in distinct nuclei of the hypothalamus leverage different mechanisms to affect feeding using chemogenetic-assisted phenotyping.

Materials and methods: A viral vector packaging the designer receptor exclusively activated by designer drugs (DREADD) DIO-hM3D(Gq)-mCherry, was stereotaxically injected into the caudal mediobasal hypothalamus (cMBH) and paraventricular nucleus of the hypothalamus (PVH) of Gipr-Cre mice (n=8 per group). The effects of chemogenetically activating Gipr neurons in the PVH (Gipr^PVH-Dq) or cMBH (Gipr^cMBH-Dq) on net food intake (FI), meal patterning, and energy expenditure (EE) were measured using Feeding Experimental Device (FED3) and the Promethion continuous monitoring system. Histology experiments were conducted for target validation and viral-assisted circuit tracing.

Results: Activation of Gipr neurons in the PVH (Gipr^PVH-Dq) and the cMBH (Gipr^cMBH-Dq) both increased EE and ambulatory activity. Analysis of ingestion, cumulative and hourly food intake, revealed that activation of Gipr neurons in both the cMBH and PVH led to a significant decrease in food intake 2 hours after injection of Clozapine-N-Oxide (CNO, specific ligand for DREADDs) (PVH Veh: 0.69g ± 0.04 vs CNO: 0.35g ± 0.091 P = 0.0043; cMBH Veh: 0.66g ± 0.075 vs CNO: 0.51g ± 0.065, P = 0.03)). Meal patterning analysis revealed distinct differences in microstructural feeding parameters affected by, Gipr^PVH-Dq and Gipr^cMBH-Dq neurocircuits. Acutely activating Gipr^PVH-Dq neurocircuits significantly decreased meal number (Veh: 6.6 ± 1.0 vs CNO:4.0 ± 0.9, P=0.005).—a measure of satiety. In contrast, activating Gipr^cMBH-Dq neurocircuits decreased meal duration (Veh: 52.2 s ± 3.5 vs CNO: 38.0 s ± 7.1, P = 0.04 ).--a measure of satiation. In addition to differing effects on meal microstructure Gipr^PVH and Gipr^cMBH neurocircuits exhibited separate and distinct connectivity and projection patterns. Gipr^PVH neurons projected to the median eminence and nucleus tractus solitarius, whereas Gipr^cMBH neurons projected to the lateral septum and bed nucleus of the stria terminalis. Viral-assisted monosynaptic input labelling analysis characterized the molecular identity of presynaptic neuron populations lying upstream of both Gipr^PVH and Gipr^cMBH cells.

Conclusion: Our data demonstrate that Gipr neruons in discrete hypothalamic nuclei exert distinct effects on feeding behaviour: activation of Gipr^PVH neurons elicits increased satiety, whereas Gipr^cMBH neurons promote satiation. Gipr^PVH and Gipr^cMBH neurons also exhibit divergent neurocircuitry, suggesting that Gipr neurons in the PVH engage distinct neuronal populations from Gipr neurons in the cMBH. Collectively, our data suggest that hypothalamic Gipr neurons are heterogenous, and engage separate circuits and signalling mechanisms to alter feeding.

Supported by: EFSD/Novo Nordisk Foundation Future Leaders Award Programme

Disclosure: N.S. Figueredo Burgos: None.

60

Brain-to-periphery crosstalk: exploring sympathetic innervation and dopamine signalling during diet-induced obesity or caloric restriction in male Wistar rats

F. Hukema ¹, C. Kagios¹, S. Hetty¹, S. Zelleroth², M.J. Pereira¹, E. Roman^2,3, J.W. Eriksson¹;

¹Clinical Diabetes and Metabolism, Uppsala University, Uppsala, Sweden, ²Pharmaceutical Biosciences, Uppsala University, Uppsala, Sweden, ³Animal Biosciences, Swedish University of Agricultural Sciences, Uppsala, Sweden.

Background and aims: Obesity and type 2 diabetes (T2D) display impaired insulin sensitivity in peripheral tissues, such as adipose tissue (AT), muscle, and liver. In addition, the brain’s glucose sensing and regulating functions may be altered. Brain-to-periphery communication partly occurs via tyrosine hydroxylase (TH)-expressing sympathetic nerves releasing norepinephrine (NE), and to a lesser extent dopamine (DA), regulating muscle, liver, and AT glucose and lipid metabolism. DA can also originate from adrenals or immune cells. DA acts via five receptors (DRD1-5) throughout the brain, influencing reward, behavior, and energy balance. DRDs are also found in insulin-sensitive tissues. This study explored the impact of diet-induced obesity (DIO) and caloric restriction on sympathetic innervation of AT, as well as DA signaling in AT and the brain. Ongoing work addresses the liver and muscle.

Materials and methods: Brains, epididymal and inguinal white AT (eWAT, iWAT), muscle, and liver were collected from male Wistar rats after 12 weeks of either a calorie-restricted, ad libitum control, or high-calorie cafeteria diet (DIO group; n=8-12/group). DRD1, DRD2, and TH protein levels were analyzed in peripheral tissues by Western Blot. In brain areas important for reward and homeostasis, e.g., caudate putamen (Cpu), and hypothalamus (Hth), DRD1 and DRD2 expression levels were examined by immunofluorescence (n=2-3/group). In these brain areas, also DA levels were obtained using matrix-assisted laser desorption/ionization mass spectrometry imaging (n=6/group).

Results: Throughout the experiment, control rats gained on average 42% body weight (BW). DIO rats gained 66% of BW, while caloric-restricted rats gained only 24% of BW (both p<0.0001 vs control). DIO, but not caloric restriction, reduced TH protein in iWAT by 61% (p<0.05), while TH protein levels in eWAT remained unchanged by both dietary interventions. iWAT TH levels correlated negatively with BW, BW gain, and relative eWAT amount (p<0.001). DIO decreased DRD1 protein levels in eWAT by 48%, and iWAT by 41% (both p<0.05), compared to control. Surprisingly, also caloric restriction reduced DRD1 protein levels in both depots (by 41%; p<0.05). Both DIO and caloric restriction increased DRD2 protein levels, but only in iWAT (both p<0.05). In the Cpu and the Hth, neither DIO nor caloric restriction altered DA levels. However, caloric restriction showed a trend towards increased DRD1 expression in the Cpu (p=0.09). DRD2 expression in the Cpu was increased by both DIO (by 10%) and caloric restriction (by 16%), compared to control (both p<0.05).

Conclusion: Both DIO and caloric restriction downregulated DRD1 in AT, while only caloric restriction increased DRD1 in the Cpu. This may imply a compensatory mechanism toward changes induced by caloric restriction. Moreover, dietary intervention seems to affect DA signaling in the Cpu via receptor modulation rather than altering DA levels. DRD2 was upregulated only in iWAT by DIO and caloric restriction, indicating a fat-depot-specific effect. Reduced TH expression may reflect a blunted NE and DA release in iWAT contributing to adiposity, and upregulation of DRD2 in this depot could be an adaptive response.

Supported by: Diabetesfonden, Novo Nordisk Fonden, UDC, ALF, European Commission Horizon PAS GRAS

Disclosure: F. Hukema: None.

OP 11 Risk factors and treatment opportunities in neuropathy and hypertension

61

Understanding the impact of diabetic peripheral neuropathy and neuropathic pain on quality of life and mental health in people with diabetes

M. Borbjerg ^1,2, A.-M.L. Wegeberg³, A. Nikontovic¹, C.D. Moerch², N. Ejskjaer^4,1, C. Brock^3,1, P. Vestergaard^1,4, J. Røikjer^1,2;

¹Steno Diabetes Center North Denmark, Aalborg University Hospital, Aalborg, Denmark, ²Integrative Neuroscience, Aalborg University, Aalborg, Denmark, ³Mech-Sense, Aalborg University Hospital, Aalborg, Denmark, ⁴Department of Endocrinology, Aalborg University Hospital, Aalborg, Denmark.

Background and aims: Diabetic peripheral neuropathy (DPN) affect up to 50% of people living with diabetes; it leads to decreased Quality of Life (QoL) and increased risk of depression and anxiety. However, little is known about the impact of concomitant neuropathic pain on these factors. Therefore, this study aimed to investigate QoL, mental health and socioeconomic factors in people with diabetes with and without DPN and neuropathic pain.

Materials and methods: 7,743 participants with diabetes participated in this observational, cross-sectional survey study. Participants with incomplete socioeconomic and mental health factors (n = 781) were excluded. Participants were grouped into people with (n = 1,601) and without (n=5,361) DPN based on the Michigan Neuropathy Screening Instrument questionnaire (MNSIq). Participants with DPN were subsequently divided into people with (n=1,085) and without (n=516) concomitant neuropathic pain based on the modified Douleur Neuropathique en 4 Questions (DN4)-interview. All participants answered The Short Form Health Survey SF(36), and Hospital Anxiety and Depression Scale (HADS). Socioeconomic factors were self-reported and included education level, employment status and annual household income. Differences between groups were assessed using the Mann-Whitney U-test based on the data type and normality. The normality of data was assessed using QQ plots and density plots. The level of statistical significance was 0.01. Results are presented as median and interquartile range (IQR).

Results: QoL-, depression- and anxiety scores all differed significantly between participants with and without DPN (all p<0.001). The SF36 medians was 55.1 [IQR 36.7 , 73.6] and 82.2 [IQR 63.6 , 90.9] for the two groups, respectively, while the HADS depression scores were 4.00 [IQR 1 , 8] and 1.00 [IQR 0 , 3], and the HADS anxiety scores were 5.00 [IQR 2 , 9] and 2.00 [IQR 1 , 5], respectively. Education level did not differ between the two groups, but a significant difference was seen for both work and salary (p-value < 0.001). For people with DPN and concomitant neuropathic pain, both QoL-, depression-, and anxiety scores were further attenuated with SF36 scores of 50.7 [IQR 34.8 , 69.8] for people with painful DPN and 61.2 [IQR 45.2 , 79.0] for people with painless DPN (p-value <0.001). HADS depression scores were 4.00 [IQR 1 , 8] and 1.00 [IQR 0 , 3], while HADS anxiety scores were 6.00 [3 , 10] and 4.00 [IQR 1 , 8] for the two groups, respectively (p-value <0.001). Socioeconomic factors did not differ between the two groups, except for education level (p-value <0.001).

Conclusion: People with DPN have reduced QoL and increased levels of anxiety and depression compared to people with diabetes without DPN. In addition, people with concomitant neuropathic pain further attenuated QoL, and increased levels of anxiety and depression compared to people with DPN without pain.

Disclosure: M. Borbjerg: None.

62

Impact of diabetes on heart rate variability among acutely admitted patients

R. Hadad ¹, A.V.R. Hviid², M.H. Domínguez¹, F.E. Nielsen³, B.S. Larsen¹, O. Wendelboe¹, A. Sajadieh¹, S.B. Haugaard²;

¹Dept. of Cardiology, Copenhagen University Hospital Bispebjerg and Frederiksberg, Copenhagen, Denmark, ²Dept. of Endocrinology, Copenhagen University Hospital Bispebjerg and Frederiksberg, Copenhagen, Denmark, ³Dept. of Acute Medicine, Copenhagen University Hospital Bispebjerg and Frederiksberg, Copenhagen, Denmark.

Background and aims: Diabetes is known to activate the sympathetic system resulting in higher heart rate and in low beat-to-beat variation of the heart (SDNN) a marker of cardiac autonomic neuropathy. It is less investigated if people with diabetes during admission to the hospital for acute medical therapy exhibit increased sympathetic tonus compared to people without diabetes.

Materials and methods: For 3 years we obtained informed consent from 961 unselected adult people acutely admitted to a larger Danish Hospital (covering 460,000 citizens) for medical treatment. 240 patients were excluded from analyses primarily due to non-sinus rhythm (80%). Heart rate variability (HRV) recordings were performed for 10 minutes within 24 hours from time of admission by Finapres Nova (Finapres, Amsterdam, Netherlands). The system provided automatic analysis of the HRV and both time and frequency domain components of HRV were recorded. The standard deviation of normal-normal beats (SDNN) was the primary marker, and secondary markers of autonomic function were the root mean square of successive normal-normal-interval differences (RMSSD), low frequency (LF), and high frequency (HF).

Results: The number of patients included with analyses of HRV was 721 among these 110 with diabetes (8 T1D, 102 T2D). People with diabetes were older, more often males, had higher Charlson Comorbidity Index, and higher pulse rate but similar smoking habits compared to people without diabetes (Table 1). SDNN was reduced by 25% (P<0.0001) in people with diabetes, and after adjustment for sex, age, smoking, BMI, resting heart rate and Charlson Comorbidity Index this reduction remained significant at 14% (P=0.03). Also, the secondary markers of sympathetic activation were impaired in people with diabetes (Table 1) and the difference remained robust after adjusting for age and sex but not for the full model.

Conclusion: The study demonstrated significantly increased sympathetic activation of people with diabetes, who were acutely admitted for medical treatment. The clinical impact and possible therapy of sympathetic activation in people with diabetes in the setting of daily life and during acute admission to hospital warrants investigated.

Clinical Trial Registration Number: NCT03934775

Supported by: A Bispebjerg Hospital Research Grant; Læge Sofus Carl Emil Friis og Hustru Olga Doris Friis’ Legat

Disclosure: R. Hadad: None.

63

A prognostic tool to predict the risk of 5-year all-cause mortality in diabetic patients with hypertension: evidence from two perspective cohorts

Y. Liu ¹, H. You¹, J. Yuan¹, S. Wu², Y. Wu², W. Yang¹;

¹State Key Laboratory of Cardiovascular Disease, Fuwai Hospital and National Centre for Cardiovascular Diseases, Beijing, China, ²Kailuan Hospital, North China University of Science and Technology, Tangshan, China.

Background and aims: Among diabetic patients, those with coexisting hypertension are considered to have a high risk of all-cause mortality. Effective risk stratification to optimize accurate management is an urgent need. The metabolic status, renal function, and cardiovascular conditions can profoundly affect the prognosis of diabetic patients. This study was performed to explore the association of obesity, deteriorating renal function, and higher triglyceride-glucose (TyG) index with the 5-year all-cause mortality risk in patients with diabetes-hypertension comorbidity based on the National Health and Nutrition Examination Survey (NHANES) database from the United States and the Kailuan cohort from China.

Materials and methods: The study involved 1397 patients from the NHANES cycles spanning 2003-2014 and 1239 patients from the Kailuan cycles spanning 2010-2015. The cut-off values of BMI and Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) eGFR were based on guidelines and clinical practice. Multivariable Cox regression was used to determine the association between predictors and 5-year all-cause mortality. Cumulative event-free survival estimates and survival curves were derived by the Kaplan-Meier method to assess the prognostic value stratified by the number of positive predictors. Groups were compared using the log-rank test.

Results: The 5-year all-cause mortality rate was 8.3% in the NHANES database and 9.0% in the Kailuan cohort. The median TyG index was 9.5 (IQR: 9.2-9.7) in the NHANES database after sample weighting, clustering, and stratification. Multivariable Cox regression demonstrated that a BMI of >30 kg/m² (HR: 1.86, 95% CI: 1.29-1.73), TyG index of >9.5 (HR: 1.64, 95% CI: 1.39-2.72), and CKD-EPI eGFR of <60 mL/min per 1.73m² (HR: 2.03, 95% CI: 1.58-4.01) were independent predictors of 5-year all-cause mortality in diabetic patients with hypertension after adjusting for age, sex, history of coronary artery disease, education, marital status, smoking history, and alcohol use. These results were reconfirmed in the Kailuan cohort for BMI (HR: 1.99, 95% CI: 1.63-2.45), TyG index (HR: 2.13, 95% CI: 1.86-3.02), and CKD-EPI eGFR (HR: 2.38, 95% CI: 1.87-4.35). The Kaplan-Meier survival curves showed that the all-cause mortality rates gradually increased with the number of positive predictors (NHANES database: 93.8% vs. 88.3% vs. 84.1% vs. 70.3%; Kailuan cohort: 95.1% vs. 87.8% vs. 81.5% vs. 72.7%; both log-rank P < 0.0001).

Conclusion: In diabetic patients with hypertension, an increased BMI, elevated TyG index, and lower CKD-EPI eGFR were independent predictors of 5-year all-cause death based on two prospective cohort studies from the United States and China. This novel prognostic tool has strong potential for effective risk stratification of wider populations in clinical practice.

Disclosure: Y. Liu: None.

64

Neuropathic pain symptoms in patients with painful diabetic neuropathy followed for 12 months after treatment with high concentration (179 mg) capsaicin patch: a retrospective cohort study

M. Eerdekens ¹, M.A. Überall², S. Engelen¹, T. Fajri³, S. Allen⁴, R. Freitas⁵, L. Garcia Guerra⁶, T. Quandel¹;

¹Grünenthal GmbH, AACHEN, Germany, ²O.Meany Medical Data & Project Management GmbH, Nuernberg, Germany, ³Laboratoires Grünenthal S.A.S., Paris, France, ⁴Averitas Pharma Inc., Morristown, NJ, USA, ⁵Grünenthal S.A., Lisbon, Portugal, ⁶Grünenthal Pharma S.A., Madrid, Spain.

Background and aims: Painful diabetic peripheral neuropathy (pDPN) represents a significant burden. However 50% of diabetes patients positively screening for pDPN, never receive a diagnosis, and 22% avoid discussing pain symptoms with health care professionals. Research of the predominant symptoms of pDPN, their treatment in clinical practice, and the evolution of the symptoms following treatment with high concentration (179 mg) capsaicin patch (HCCP), is thus of interest.

Materials and methods: This is an analysis of a non-interventional, retrospective 12 months cohort study using anonymized routine medical care data from the German Pain eRegistry of patients diagnosed with pDPN, exposed to treatment with HCCP at least once and followed up for 12 months. This analysis is part of a larger study assessing the correlation between patient baseline characteristics and responder rates. The current analysis focuses on the assessment of pDPN symptoms by means of the Pain Detect Questionnaire 7 items (PDQ-7) score. Patients are asked to rate their neuropathic pain symptoms on a 0-5 point scale from “never=0” up to “very strongly=5”. Assessments were done at baseline and after each successive HCCP treatment over the 12 month observation period. The use of concomitant pain medication was assessed before and after HCCP initiation.

Results: In total, 826 pDPN patients were included. All received 1 (100%) HCCP treatment; 653 (79%), 464 (56%), and 279 (34%) had a 2nd, 3rd and 4th treatment respectively. The mean (SD) age was 66.8 (13.1) years and pain duration was 5 (3.6) years. The baseline mean (SD) 24 hour average pain score (0-100) was 57.5 (18.2) and decreased to 16.0 (13.2) in patients who received 4 treatments (p<0.001). At baseline all patients were using concomitant analgesic medication; use decreased over the course of the 12 months. Of those who received 4 HCCP treatments, 28.7% no longer received concomitant analgesic medication at Month 12. The PDQ-7 total score (0-35) at baseline was mean (SD) 22.1 (3.5) and decreased to 18.2 (2.8), 17.8 (2.8), 17.8 (2.8) 17.8 (2.6) after 1, 2, 3, 4 treatments respectively. The most prominent symptoms at baseline were sensations of burning, prickling and tingling and numbness. Individual symptom scores at baseline and after 4 treatments are presented in Figure 1. HCCP was well tolerated.

Conclusion: This cohort study included pDPN patients with high baseline pain scores although all received pain medication. With continued HCCP treatment, pain intensity and pain medication use decreased significantly. Neuropathic pain symptoms significantly improved also the most prominent ones: sensations of burning, prickling/tingling and numbness.

Disclosure: M. Eerdekens: Employment/Consultancy; Employee of Grünenthal GmbH.

65

Risk factor phenotypes for cardiovascular autonomic neuropathy based on indices of heart rate variability in type 1 diabetes

B. Braffett ¹, L. El ghormli¹, I. Bebu¹, C. Martin², R. Pop-Busui², DCCT/EDIC Research Group;

¹George Washington University, Bethesda, MD, USA, ²University of Michigan, Ann Arbor, MI, USA.

Background and aims: Indices of heart rate variability (HRV) from standard electrocardiograms (ECG’s) were previously validated as sensitive surrogate measures of cardiovascular autonomic neuropathy (CAN). We determined risk factors associated with ECG-derived CAN, as well as the impact of CAN on subsequent cardiovascular disease (CVD), in adults with type 1 diabetes enrolled in the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC) study.

Materials and methods: Standard 10-second, 12-lead resting ECGs were obtained annually in 997 EDIC participants (mean±SD age 56.2±6.8 years; duration 34.7±7.9 years; 48% female) between 2014 and 2023 (EDIC years 22-29). Two HRV indices were derived: standard deviation of normally conducted R-R intervals (SDNN), and root mean square of successive differences between normal-to-normal R-R intervals (rMSSD). ECG-derived CAN was defined as SDNN<17.13 and rMSSD<24.94 ms, based on a previous validation study in this cohort using gold standard cardiovascular reflex tests (CARTs). Generalized estimating equation models assessed the association of risk factors with CAN over time. Two CVD outcomes were considered: major adverse cardiovascular events (MACE, defined as CVD death or nonfatal acute MI or stroke) and any-CVD (MACE plus silent MI, angina, percutaneous transluminal coronary angioplasty/coronary artery bypass graft, or congestive heart failure). Cox proportional hazards regression models assessed the association between CAN status in 2014 and CVD risk during the follow-up period.

Results: The prevalence of CAN was stable between 2014 and 2023 (average prevalence 64.4%). Higher pulse rate was the most significant risk factor for CAN (OR [95% CI] 1.27 [1.23,1.32] per 1 bpm), followed by higher mean HbA1c (1.63 [1.47,1.80] per 1%), older age, longer duration of type 1 diabetes, male sex, angiotensin receptor blockers use, and a history of hypertension. Among the event-free participants at baseline (i.e., year 2014), there were 73 incident cases of any-CVD (rate=1.4 events/100PY) and 30 incident cases of MACE (rate=0.5 events/100PY) during follow-up. The cumulative incidence of the first occurrence of any-CVD and of MACE (Figure) was significantly higher in participants with vs. without CAN at baseline (HR [95% CI] any-CVD 2.6 [1.5,4.5]; MACE 4.2 [1.5,12.2]). The associations remained significant after adjustment for mean HbA1c and age.

Conclusion: The prevalence of and risk factors for CAN, defined by ECG-derived indices of HRV, were similar to previously reported estimates based on CARTs. CVD risk was significantly higher in those with vs. without CAN at baseline. These data demonstrate and further support the utility of these indices as important point of care measures of CAN.

Clinical Trial Registration Number: NCT00360893

Supported by: NIDDK U01DK094176

Disclosure: B. Braffett: None.

66

Blood pressure polygenic risk scores associate with treatment-resistant hypertension in individuals with type 1 diabetes

R. Lithovius ^1,2, H. Sánez Tähtisalo³, A. Antikainen^1,2, S. Mutter^1,2, P.-H. Groop^1,2, K. Kontula³, T. Hiltunen³, N. Sandholm^1,2;

¹Folkhälsan Research Center, Helsinki, Finland, ²Department of Nephrology, University of Helsinki and Helsinki University Hospital, Helsinki, Finland, ³Department of Medicine, University of Helsinki and Helsinki University Hospital, Helsinki, Finland.

Background and aims: Hypertension is a major contributing and modifiable risk factor for CVD and kidney complications, but the treatment targets are often not met. Individuals with elevated BP despite intense treatment are diagnosed with treatment-resistant hypertension (TRH), a condition which further increases the risk. Early detection of these high-risk individuals may lead to more efficient BP control. It remains unknown to which extent TRH is explained by genetic factors predisposing to essential hypertension. Therefore, we explored if polygenic risk scores (PRSs) for systolic BP (SBP) and diastolic BP (DBP) predict TRH in two cohorts with and without type 1 diabetes: The FinnDiane and the LIFE-Fin.

Materials and methods: FinnDiane is an ongoing, observational nationwide study aiming at identifying genetic and clinical risk factors for complications in adults with type 1 diabetes. TRH was defined as SBP ≥130 or DBP ≥85 mmHg and the use of ≥3 antihypertensive drugs, or SBP <130 and DBP <85 mmHg while using ≥4 antihypertensive drugs (476 cases). Controlled hypertension was defined as <3 antihypertensive drugs and SBP <130 and DBP <85 mmHg (302 controls). LIFE-Fin is a randomized, double-blind study evaluating long-term effects of losartan compared with atenolol in adults with hypertension and signs of left ventricular hypertrophy. TRH was defined as SBP ≥140 or DBP ≥90 mmHg and the use of ≥3 antihypertensive drugs (173 cases) and controlled hypertension as SBP <140 and DBP <90 mmHg (205 controls) at the four-year follow-up visit of the study. Normal kidney function was defined as eGFR ≥60 ml/min/1.73 m². PRSs were calculated for each participant using imputed genomic data and the weights of 1,098,015 genetic variants; the variant-specific data were derived from UK Biobank based PRS-CS-method computed PRSs.

Results: In the FinnDiane cohort PRS for both SBP and DBP were associated with higher odds of TRH (ORs 1.55 [95% CI 1.30, 1.86], P <0.0001, and 1.27 [1.06, 1.52], P=0.01, per one-SD increase of the PRS), adjusted for age, sex, BMI, eGFR and albuminuria. However, when we split individuals into two groups by their kidney function (i.e., eGFR ≥60 vs. <60 ml/min/1.73 m²), only in those with eGFR ≥60 was the SBP PRS (1.76 [1.42, 2.19], P<0.0001) and DBP PRS (1.46 [1.18, 1.81], P=0.0006) associated with TRH. Neither SBP nor DBP PRSs were significantly associated with TRH in LIFE-Fin (P≥0.4). Of note, LIFE-Fin cases and controls had higher mean PRS for SBP and DBP than cases and controls in the FinnDiane study (P<0.0001 for each), suggesting that LIFE-Fin participants had high genetic risks for hypertension.

Conclusion: As PRS for BP associate with higher risk of TRH in individuals with type 1 diabetes and normal kidney function, early risk detection would be feasible. Conversely, individuals in the LIFE-Fin have already higher BP PRSs than in the FinnDiane, which may explain the non-significant association with TRH.

Supported by: Finnish Foundation for Cardiovascular Research

Disclosure: R. Lithovius: None.

OP 12 Fatal attraction: immune cells and beta cells

67

Beta cell extracellular vesicle surface PD-L1 is induced by IFN-α treatment and correlates with residual C-peptide in plasma of children with recent-onset type 1 diabetes

E. Sims ¹, C. Rao¹, D. Cater¹, J. Xu¹, A. De Olivera¹, C. Evans-Molina¹, D. Eizirik², R. Mirmira³;

¹Indiana University School of Medicine, Indianapolis, IN, USA, ²Dept. of Experimental Medicine, Free University Brussels (ULB), Brussels, Belgium, ³Dept. of Medicine, Section of Endocrinology, Diabetes and Metabolism, University of Chicago, Chicago, IL, USA.

Background and aims: Surviving beta cells in type 1 diabetes respond to inflammation by upregulating programmed death-ligand 1 (PD-L1) to engage immune cell programmed death-1 (PD-1) and limit destruction by self-reactive immune cells. Extracellular vesicles (EVs) and their cargo can serve as biomarkers of beta cell health and contribute to islet intercellular communication. We hypothesized that the inflammatory milieu of type 1 diabetes increases PD-L1 in beta cell EV cargo and that EV PD-L1 may protect beta cells against immune-mediated cell death.

Materials and methods: INS-1 and EndoC-βH1 beta cell lines and human islets were treated with IFN-α for 24 h to model the early proinflammatory type 1 diabetes microenvironment. EVs were isolated using ultracentrifugation or size-exclusion chromatography and analysed for PD-L1 cargo via immunoblot, flow cytometry, and ELISA. EV PD-L1: PD-1 binding was assessed using a PD-1/PD-L1 competitive binding assay. Plasma EV and soluble PD-L1 were assayed in plasma of children with recent-onset type 1 diabetes and compared to age, sex, and BMI-matched non-diabetic controls.

Results: PD-L1 protein colocalized with EV associated proteins intracellularly and was detected in beta cell EVs. IFN-α treatment yielded a 2-fold increase in EV PD-L1 cargo, without a corresponding increase in number of EVs. IFN exposure predominantly increased PD-L1 expression on the surface of beta cell EVs and beta cell EV PD-L1 showed a dose-dependent capacity to bind PD-1. Plasma EV PD-L1 levels from children with recent-onset type 1 diabetes were similar to non-diabetic controls. However, in children with type 1 diabetes, but not controls, plasma EV PD-L1 positively correlated with circulating C-peptide levels, suggesting that higher EV-PD-L1 could be protective for residual beta cell function.

Conclusion: IFN-α exposure increases PD-L1 on the beta cell EV surface. Beta cell EV PD-L1 has the capacity to bind PD-1 and circulating levels correlate with residual C-peptide in recent-onset type 1 diabetes. These findings suggest that EV PD-L1 binds PD-1 and could be exploited as a means to inhibit immune-mediated beta cell death.

Supported by: NIH

Disclosure: E. Sims: Grants; Grant recipient, National Institutes of Health.

68

T cells drive beta cell senescence in type 1 diabetes

J. Pipella, P. Thompson;

Physiology and Pathophysiology, University of Manitoba, Winnipeg, MB, Canada.

Background and aims: In 2022, an anti-CD3 immunotherapy Teplizumab received FDA approval to delay symptomatic Type 1 Diabetes (T1D). Studies show that anti-CD3 immunotherapy improves beta (β) cell function, suggesting an impact on β cell stress, yet the mechanism underlying the improvement of β cell function remains unclear. Recent studies indicate that a subpopulation of β cells can enter a cellular senescent fate and accumulate in the context of T1D. The aim of this study was to determine if anti-CD3 immunotherapy exerts its effects by mitigating the amount of β cell destruction by cytotoxic T cells and/or reducing the accumulation of stressed senescent β cells during the development of T1D.

Materials and methods: Female non-obese diabetic (NOD) mice received weekly doses of anti-CD3 monoclonal antibody or IgG control treatment at the prediabetic stage. Intraperitoneal glucose tolerance tests (IPGTTs) were conducted to assess β cell function prior to isolation of whole pancreas or pancreatic islets (n=18). Senescent β cell frequency and protein expression were assessed using immunohistochemistry (IHC) (n=6) and gene expression was measured in islets (n=18). The anti-CD3 treatment was validated and further immunophenotyping was conducted using flow cytometry (n=18). Published studies show that the administration of ABT-199, a senolytic compound that causes apoptosis in senescent cells, can halt the progression of T1D in NOD mice. Therefore, we proposed a combination therapy, incorporating anti-CD3 immunotherapy with ABT-199. Thus, prediabetic NOD mice were treated with anti-CD3/ABT-199, IgG/ABT-199, or vehicle treatment and similar β cell function (n=6), β cell senescence characterization (n=6), and immunophenotyping experiments were completed (n=6).

Results: Significant depletion of total CD3+ T cells (p<0.0005), decreases in antigen-specific insulin tetramer T cells (p<0.05), decreases in total CD4+ T cells (p<0.0005) and no changes in regulatory T cells were observed in the anti-CD3 treated mice compared to the IgG controls. IHC revealed β cell preservation and slowed disease progression, indicated by increased insulin+ stained area (p<0.05), and reduced insulitis (p<0.05), as well as a decrease in senescent β cells (p<0.0005) in the anti-CD3 mice. Remarkably, no difference in markers of β cell identity, proliferation, or unfolded protein response pathways were observed. We also observed a decrease in senescent gene expression (p<0.05) in anti-CD3 treated mice versus control. Notably, the combination treatment showed a synergistic effect, as we observed further improvement in glucose tolerance in anti-CD3/ABT-199 mice (p<0.05), compared to IgG/ABT-199 and vehicle groups, a decrease in senescent β cells in anti-CD3/ABT-199 mice (p<0.005) and IgG/ABT-199 mice (p<0.05) compared to vehicle, and a reduction in insulitis in CD3/ABT-199 treated mice (p<0.05) compared to other groups.

Conclusion: Taken together, these findings in NOD mice suggest a novel mechanism for the action of anti-CD3 at the late pre-diabetic stage. Specifically, the modulation of antigen specific T cells with anti-CD3 appears to impede disease progression by limiting the accumulation of senescent β cells. To further investigate whether T cells are driving healthy β cells towards a stressed senescent state, future studies will explore the interplay between T cells and β cell senescence using co-culture assays.

Supported by: Research Manitoba, CIHR

Disclosure: J. Pipella: None.

69

Insulitis and exocrinitis in antibody-positive non-diabetic individuals: role of HLA genotypes individuals: role of HLA genotypes

E. Larger, M. Diedisheim, D. Dubois-Laforgue, R. Mallone;

Diabetes Department Cochin Hospital, APHP, Paris Cité University /Université Paris Cité, Paris, France.

Background and aims: Type 1 diabetes (T1D) is characterized by the presence of autoantibodies on a genetic background largely determined by HLA class II haplotypes. Stage 1 of the disease is characterized by the presence of autoantibodies and normoglycemia. We quantified pancreatic immune infiltrates at stage 1 T1D, according to HLA genotypes.

Materials and methods: We analyzed CD3+ lymphocyte infiltration of both endocrine and exocrine pancreas in 117 sections from 30 non-diabetic individuals positive for at least one autoantibody from the Network for Pancreatic Organ Donors with Diabetes (nPOD). HLA haplotypes were classified as at risk (DQ2/DQB1*02:01 and/or DQ8/DQB1*03:02), protective (DQ6/DQB1*06:02) or neutral (other HLA-DQ alleles).

Results: Among these 30 individuals (median age 25 years, IQR 21-39; median BMI 24 kg/m², IQR 21-30), 23 were single autoantibody-positive and 7 were positive for 2 autoantibodies. β-cell mass was normal in all. HLA-DQ allele distribution was similar to that of autoantibody-negative non-diabetic nPOD donors, and differed from that of nPOD donors at stage 3. Insulitis was identified only in one case. CD3+ lymphocyte densities in islets did no correlate with HLA status or autoantibody number, and correlated with densities in the exocrine pancreas.

Conclusion: These autoantibody-positive donors including 7 with 2 autoantibodies had normal β-cell mass and no significant insulitis, suggesting an underlying autoimmune process progressing very slowly even in the presence of HLA genetic risk

Supported by: FFRD

Disclosure: E. Larger: None.

70

The Hippo pathway terminal effector YAP potentiates enteroviral infection in human islets

F. Atawneh ¹, S. Geravandi¹, H. Liu¹, O. Zabad¹, H. Pahwa¹, M. Madduri¹, A. Pugliese², A. Ardestani^1,3, K. Maedler¹;

¹Center for Biomolecular Interactions Bremen (CBIB), Bremen, Germany, ²Department of Diabetes Immunology & The Wanek Family Project for Type 1 Diabetes, Arthur Riggs Diabetes & Metabolism Research Institute, City of Hope, Duarte, CA, USA, ³Biomedical Institute for Multimorbidity (BIM), Centre for Biomedicine, Hull York Medical School, University of Hull, Hull, UK.

Background and aims: The risk of developing autoimmunity and type 1 diabetes (T1D) has been associated with recurrent enteroviral infections, especially with those from coxsackieviruses B (CVBs). Enteroviral RNA infection is markedly increased in the pancreas of organ donors with T1D as well as in those with T1D associated autoimmunity. The Hippo pathway terminal effector, Yes-associated Protein (YAP), plays a crucial role in balancing the host’s antiviral response pathway, either by boosting or inhibiting viral amplification and innate immunity responses. However, its mechanism in human islets is completely unknown.

Materials and methods: Isolated islets and exocrine cells from organ donors were infected with coxsackievirus CVB3 or CVB4 with and without adenoviral infection of YAP or the YAP-TEAD inhibitor Verteporfin (VP), which inhibits YAP-TEAD interaction and downstream YAP transcriptional activity. Viral infection was analyzed by the virus capsid protein (VP1) and β-cell survival by double staining for TUNEL and insulin. Results were confirmed by Western blotting.

Results: Both CVB3 and CVB4 enteroviral infections of human islets induced β-cell apoptosis, which was further potentiated by YAP overexpression (2-/3.5-fold increase compared to CVB3/CVB4 control-infected islets; p<0.05). The pro-viral effect of YAP was confirmed by the elevated levels of β-cell infection, as evidenced by immunolabeling of the enterovirus-specific viral capsid protein VP1 upon YAP overexpression (9-/3.5-fold increase compared to CVB3/CVB4 control-infected islets; p<0.05), as well as the significantly increased CVB3 and CVB4 intracellular genomic RNA levels (3-/5-fold increase compared to CVB3/CVB4 control-infected islets; p<0.05). Given that exocrine cells, particularly ductal cells, naturally express YAP, we investigated whether endogenous YAP has a similar pro-viral effect in a culture of human pancreatic cells consisting of both exocrine and endocrine islets from the same donor. Verteporfin, a chemical inhibitor of the YAP-TEAD complex that blocks YAP downstream signals, reduced both exocrine and endocrine CVB3 and CVB4 replication. Verteporfin significantly decreased the number of VP1-CK19 (ductal) double-positive cells by 55% and 60% and VP1-insulin double-positive cells by 40% and 35%, compared to CVB3/CVB4-infected controls (p<0.05). YAP-expressing cells exhibited virus positivity or were localized near virus-infected cells in T1D. Our findings demonstrate that YAP overexpression enhances CVB replication and promotes β-cell apoptosis, whereas its inhibition halts viral replication in primary pancreatic cells.

Conclusion: Here, we show that YAP overexpression enhances CVB replication and promotes β-cell apoptosis, while inhibiting its downstream signaling halts viral replication in both exocrine and β-cells, indicating a YAP-dependent viral replications in these cells. Our work reveals YAP as a modulator of enteroviral infection in human islets. Given YAP expression in the exocrine pancreas, we have identified a pathway that may act as an early driver of autoimmunity and T1D.

Supported by: DFG (German Research Foundation)

Disclosure: F. Atawneh: None.

71

An mRNA vaccine induces immune tolerance and protects against disease in NOD mice

L.P. Pallo ^1,2, V. Ng³, S. Chen³, J. Zaifman³, A. Nguyen³, P. Singh³, S. Zhang⁴, Y.C. Tam³, H.C. Denroche^3,1, C.B. Verchere^1,4;

¹BC Children’s Hospital Research Institute, Vancouver, BC, Canada, ²Pathology & Laboratory Medicine, The University of British Columbia, Vancouver, BC, Canada, ³Integrated Nanotherapeutics, Burnaby, BC, Canada, ⁴Surgery, The University of British Columbia, Vancouver, BC, Canada.

Background and aims: Antigen-specific immune tolerance therapy (ASIT) is a promising approach to restore tolerance to beta-cell autoantigens and halt autoimmunity in type 1 diabetes (T1D). Despite extensive efforts, past T1D ASIT strategies have yet to succeed clinically. To develop an improved ASIT, we leveraged lipid nanoparticles (LNPs) to co-encapsulate tolerogenic immunomodulators with mRNA encoding individual or multiple T1D autoantigens (T1D-tolLNPs).

Materials and methods: To test if T1D-tolLNPs could prevent spontaneous diabetes, we intraperitoneally (i.p.) immunized female NOD mice (n=15/group) biweekly from 5 to 15 weeks of age with T1D-tolLNPs or control LNPs containing T1D antigen without immunomodulators (T1D-antigen only), irrelevant antigen (ovalbumin; ova) with/without immunomodulators, or buffer control. Body weight and blood glucose (BG) were monitored weekly. Mice were euthanized upon diabetes onset (two consecutive BG > 20mM) or at 30 weeks of age. To evaluate the metabolic and immunological effects of T1D-tolLNPs, we i.p. injected a second cohort of female NOD mice (n=5/group) biweekly from 5 to 13 weeks of age with T1D-tolLNPs and control formulations. We collected blood at baseline, 4-, and 24-hours post-initial injection to assess acute cytokine response, conducted i.p. glucose tolerance tests at 12 and 14 weeks, and examined T-cell phenotypes and frequencies at 16 weeks.

Results: Female NOD mice treated with T1D-tolLNPs had significantly reduced diabetes incidence relative to buffer-injected controls (13% vs 67% diabetes-free at 30 weeks, p=0.002). Notably, disease protection was antigen-dependent as ova+immunomodulator LNP-treated mice exhibited higher incidence (53% vs 13%, p=0.018) than T1D-tolLNP mice. Immunomodulators were also needed for efficacy, as T1D-tolLNP-treated mice had lower incidence than T1D antigen-only counterparts. Interestingly, mice treated with multiple T1D antigen tolLNPs had lower incidence than single T1D antigen tolLNPs (13% vs 46%). At 12 weeks, T1D-tolLNP-treated mice trended toward sustained glucose tolerance compared to ovalbumin- and buffer-treated mice. This trend became more exaggerated at 14 weeks as buffer-treated mice were most glucose intolerant while T1D-tolLNP-treated mice were most glucose tolerant. The immunomodulatory effect of T1D-tolLNPs was supported by decreased circulating levels of inflammatory cytokines IL-6 (p=0.0079) and MCP-1 (p=0.0018) relative to T1D-antigen-only LNPs four hours post-injection. In addition, T1D-tolLNP-treated mice had a trend towards increased frequencies of CD25+FoxP3+ T cells (Tregs) and PD1+ Tregs frequencies compared to buffer-injected control mice at 16 weeks of age.

Conclusion: We demonstrate the efficacy of mRNA-encoding multiple T1D antigens as a vaccination approach to prevent autoimmunity in a preclinical T1D model. This requires multiple T1D antigens encoded by mRNA and co-encapsulated immunomodulators in an LNP and likely involves increased Treg function. These findings support further preclinical development and optimization of T1D-tolLNP formulation for clinical translation.

Supported by: JDRF

Disclosure: L.P. Pallo: None.

72

Verapamil in combination with low-dose anti-thymocyte globulin reverses hyperglycaemia in newly diagnosed diabetic NOD mice

P.-J. Martens, M. Viaene, L. Degroote, C. Mathieu, C. Gysemans;

Clinical and Experimental Endocrinology (CEE), KU Leuven, Leuven, Belgium.

Background and aims: Recent compelling evidence demonstrates that the initiation of verapamil, a calcium channel blocker with beta cell protective and T cell targeting properties, exhibits significant promise for the preservation of beta cell function at symptom onset in type 1 diabetes (T1D). However, at symptom onset, a highly immunogenic process ensues, nonetheless leading to the progressive destruction of beta cells. Considering this challenge, potent systemic immunomodulatory agents, such as low-dose anti-thymocyte globulin (ATG), have shown promising effects as single agents at T1D onset. Therefore, our aim was to combine the beta cell protective properties of verapamil with the potent immunomodulatory effects of low-dose ATG to establish enduring immune tolerance and facilitate sustained remission of T1D.

Materials and methods: We combined continuous administration of verapamil (1 mg/mL in drinking water) with low-dose murine (m)ATG (250 μg per day on day 0 and 3; i.v.) to study their efficacy in new-onset diabetic non-obese diabetic (NOD) mice.

Results: Verapamil stably reversed disease in 20% of mice (n = 3/15). Low-dose mATG reversed T1D in 39% of mice (n = 7/18) 7 days after therapy start, but the effect waned to 22% of mice (n = 4/18) by 8-weeks follow-up. Verapamil combined with low-dose mATG induced durable disease remission in 45% of mice (n = 9/20). Especially in mice with mild hyperglycaemia (<350 mg/dL) at disease onset, combination reversed 75% (n = 6/8) of mice compared to 33% in either verapamil (n = 2/6 mice) or low-dose mATG (n = 3/9 mice). Only combination therapy was able to preserve C-peptide levels and decrease insulitis severity compared to untreated mice. Mechanistically, either low-dose mATG-treated group induced lymphocyte and monocyte depletion 3 days after therapy start, recovering by day 14. Flow cytometry analysis revealed a decreased percentage of CD8⁺ T cells in the blood of either low-dose mATG-treated group at day 3, with a recuperation towards a CD8⁺ effector memory (CD44^highCD62L^-) phenotype by day 14. Moreover, all treated (verapamil, low-dose mATG or the combination) groups were associated with an increased frequency of FoxP3⁺(CD25⁺) regulatory T cells (Tregs) by day 3, that persisted until day 14 in the peripheral blood and pancreatic draining lymph nodes (PLN). Interestingly, while all low-dose mATG-treated groups were associated with an increased ratio of Tregs over activated (CD44^high) CD8⁺ T cells in the peripheral blood by day 3, this effect did not persist until day 14. However, in the PLN, where an unbalanced immune state serves as one of the main drivers of T1D, only combination therapy resulted in an increased ratio of Tregs over activated (CD44^high) CD8⁺ T cells that persisted until day 14.

Conclusion: Here we present the first evidence that continuous administration of verapamil in combination with a short course of low-dose mATG protected the beta cells and induced a transient imbalance in the frequency of immune cells favouring Tregs, which may be sufficient to establish long-term tolerance and confer permanent T1D remission. This innovative approach not only holds a significant therapeutic promise but also represents a paradigm shift towards combination therapies in the future management of T1D.

Clinical Trial Registration Number: 151/2020

Supported by: KU Leuven and IMI 2 Joint Undertaking INNODIA and INNODIA HARVEST

Disclosure: P. Martens: None.

OP 13 GLP-1-based polypharmacology: we are not done yet!

73

CagriSema driven weight loss in diet-induced obese rats prevents counter-regulation of weight loss associated reduction in energy expenditure

R.E. Kuhre ¹, J.M. Jacobsen¹, B. Hald¹, J.J. Fels¹, I. Blom², J. Halling², S. Larsen², S. Lundh¹, M.K. Gerstenberg¹, A. Secher¹, K. Raun¹;

¹Global Drug Discovery, Novo Nordisk A/S, Maaloev, Denmark, ²University of Copenhagen, Copenhagen, Denmark.

Background and aims: CagriSema (CS), a combination of long-acting analogues of amylin (cagrilintide) and glucagon-like peptide-1 (semaglutide), is in clinical development for treatment of obesity and type 2 diabetes. CS induces weight loss by suppressing food intake, but it is unknown whether counteracting mechanisms on metabolic adaptation (MA; weight-independent suppression of energy expenditure [EE] during weight loss) are involved.

Materials and methods: We studied the food intake and energy expenditure effects of CS in diet-induced obese male rats (Sprague Dawley, [~750 g]) receiving daily treatment for three weeks (co-administration; 2 nmol/kg s.c. of each). To separate the effects of weight loss from CS treatment effects, a group of weight-matched (WM) rats was included. Weight loss in WM group was driven by calorie restriction and matched the trajectory in the CS group. In a separate study, using a similar protocol, cagrilintide and semaglutide were administered individually at higher doses (10 and 30 nmol/kg) to produce weight loss equivalent to CS. Non-adjusted total energy expenditure (TEE) data are reported (kcal/h).

Results: CS and WM rats lost ~11% body weight (~16% vehicle adjusted, n=10-11). By end of treatment, cumulative food intake was ~50% lower than in the vehicle group (p<0.0001). Cumulative TEE was not different between CS and vehicle groups (P=0.13), but WM group was ~14% lower than CS group (P<0.05). Inducing comparable body weight loss in WM and CS rats required ~15% fewer calories in WM. The mean TEE during treatment was ~21% lower in WM than in vehicle rats (WM: 2.54±0.10 kcal/h; vehicle: 3.21±0.10 kcal/h, p<0.001), whereas TEE in CS rats was not different from vehicle (CS: 2.93±0.13 kcal/h, p=0.20). The lower TEE in WM rats resulted from reduced EE during both the inactive phase and the active phase. Body composition was similar between WM and CS rats by end of treatment. This suggests that the reduction in TEE in WM rats was driven by active downregulation of TEE (MA), and that CS treatment counter-acted this effect. As increases in TEE may increase core body temperature (Tb), we repeated the study and measured Tb in addition to TEE. CS again counter-acted the reduction in TEE observed in the WM group, but without increasing Tb (p=0.67 between vehicle and CS). Mono-treatment with cagrilintide and semaglutide also counter-acted the reduction in TEE in WM rats (~10%, p<0.01 compared to vehicle), and TEE was with both treatments not different from vehicle (P≥0.73). Additional analysis by the end of the study showed that 1) plasma concentration of triiodothyronine (T3) was elevated in CS compared with WM rats but was not different from vehicle, 2) plasma thyroxine (T4) was decreased in WM but not in CS rats, and 3) ex vivo liver mitochondrial proton leak (a major component of resting EE) tended to be lower in WM rats compared to CS and vehicle rats.

Conclusion: CS prevents MA associated with weight loss. This effect appears to be mediated by both cagrilintide and semaglutide. The underlying mechanisms could involve elevation of circulating triiodothyronine and prevention of decrease in plasma thyroxine. Human trials are needed to determine if the counteracting actions of CS on MA observed in rats also occur in humans.

Disclosure: R.E. Kuhre: Employment/Consultancy; employee of Novo Nordisk. Other; this study was funded by Novo Nordisk A/S.

74

Safety, tolerability and weight reduction findings of oral amycretin: a novel amylin and glucagon-like peptide-1 receptor co-agonist, in a first-in-human study

A. Gasiorek ¹, A. Heydorn¹, K. Kirkeby¹, C. Key², S. Toubro¹, L.H. Schefe¹, K. Dahl¹, J.B. Hjerpsted¹, A. Vegge¹;

¹Novo Nordisk A/S, Måløv, Denmark, ²San Antonio Clinical Research Unit, ICON plc, San Antonio, TX, USA.

Background and aims: Obesity is a major global health burden with limited treatment options. Amidst investigations into the potential synergy beyond glucagon-like peptide-1 receptor agonists (GLP-1RAs), the combined efficacy of the amylin analogue cagrilintide and the GLP-1RA semaglutide is being explored as a subcutaneous treatment solution. This study introduces amycretin, a novel protein-based unimolecular amylin and GLP-1 receptor co-agonist for once-daily oral administration. The aim of this first-in-human study was to investigate the safety, tolerability, pharmacokinetics and weight loss effects of single and multiple oral doses of amycretin in participants with overweight or obesity.

Materials and methods: In this first-in-human, single-centre, placebo-controlled, double-blinded study, participants with a BMI of 25.0-39.9 kg/m² without diabetes were randomised to receive once-daily oral amycretin (n=95) or placebo (n=29). The study consisted of three parts: single-ascending dose (Part A, 1-25 mg), 10-day multiple-ascending dose (Part B, 3-12 mg) and 12-week multiple-ascending dose (Part C, 3-2×50 mg) parts. The primary endpoint was the number of treatment-emergent adverse events (AEs), while secondary endpoints included AUC and C_max for plasma amycretin. Change in body weight from baseline was evaluated as a key exploratory endpoint.

Results: At week 12, the mean change in body weight (amycretin 50 mg, -10.4% [SD, 4.6]; amycretin 2×50 mg, -13.1% [SD, 4.8]) was significantly greater with oral amycretin versus placebo (-1.1% [SD, 2.6]; p<0.0001 for both doses) (Fig. 1). The majority of AEs reported were mild to moderate in severity and related to gastrointestinal discomfort (i.e. nausea and vomiting), and decreased appetite. These AEs occurred in a dose-proportional manner, with low tolerability for dose levels ≥18 mg when amycretin was administered as a single dose and ≥12 mg when administered for 10 consecutive days. However, by introducing stepwise dose escalation, all tested dose levels up to and including 2×50 mg had acceptable safety and tolerability profiles.

Conclusion: Daily oral amycretin treatment in adult participants with overweight or obesity and without diabetes demonstrated acceptable safety and tolerability profiles and led to remarkable reductions in body weight over only 12 weeks. Furthermore, the lack of weight loss plateauing indicates the possibility of achieving further weight reductions with extended treatment. The results underscore the promising potential of amycretin as an anti-obesity medication and may pave the way for a novel patient-centred weight-management option. Investigations of amycretin in larger and longer studies are being planned to fully assess its efficacy and safety profile.

Clinical Trial Registration Number: NCT05369390

Disclosure: A. Gasiorek: Employment/Consultancy; Novo Nordisk. Stock/Shareholding; Novo Nordisk. Other; This study was funded by Novo Nordisk A/S.

75

CagriSema improves insulin sensitivity in diet-induced obese rats

A. Secher ¹, C.L. Brand², K. Raun¹;

¹Global Drug Discovery, Novo Nordisk A/S, Søborg, Denmark, ²Global Drug Discovery, Novo Nordisk A/S, Måløv, Denmark.

Background and aims: CagriSema (CS), a combination of the long-acting analogues of amylin (cagrilintide) and glucagon-like peptide-1 (semaglutide), is in clinical development for treatment of obesity and type 2 diabetes. Rodent studies have shown that CS treatment results in weight loss. It is unknown whether CS also leads to improvement of insulin sensitivity. The aim of this study was to investigate changes in insulin sensitivity with CS in diet-induced obese (DIO) rats.

Materials and methods: DIO male rats were allocated to three groups (n=6 per group): vehicle control (VC), CS and a vehicle group that was weight-matched (WM) to CS by calorie restriction. The groups were stratified by body weight (BW), fat mass and fat-free mass. The rats received daily s.c. doses of either 2 nmol/kg CS or vehicle, and food availability in the WM group was adjusted daily to achieve similar weight loss as in the CS group. After 21 days of treatment, rats were instrumented under general anaesthesia (isoflurane) with carotid and jugular catheters for sampling and infusions, respectively. After 9-11 days of recovery, rats were subjected to a 180-minute hyperinsulinaemic (15 pmol/kg/min) and euglycaemic (~5.7 mM) clamp. ³H-glucose and ¹⁴C-2-deoxy-glucose (2DG) were used to measure hepatic glucose production (HGP), glucose uptake (GU) and tissue-specific 2DG uptake.

Results: CS-treated rats and WM treated rats lost ~10% BW (baseline-adjusted). Inducing a comparable BW loss between WM rats and CS rats required a supply of ~10% less food (accumulated) to the WM rats, suggesting an additional effect by CS in metabolic adaptation. FI in the CS and WM groups returned to baseline levels at the end of treatment without an increase in BW. No change in fasting plasma glucose was observed between groups, but baseline plasma insulin decreased by 40% with CS compared with VC (p<0.03) suggesting an increase in insulin sensitivity. Consistently, compared with the VC group, clamp glucose infusion rates (GIR) increased by 4.2- and 2.6-fold in the CS (p<0.001) and WM (p=0.097) groups, respectively. There were no significant differences in HGP between groups; however, compared with the VC group, GU increased by 3.3- and 1.6-fold in CS (p=0.002) and WM (NS) groups, respectively. To further explore the GU in specific tissues, 2DG uptake in muscle and fat was evaluated and found to be higher in muscle (but not in fat) in the CS group compared to the VC group.

Conclusion: Treatment with CS increased insulin sensitivity and improved glucose metabolism compared with VC in DIO rats. This improvement in insulin sensitivity could only partly be explained by weight loss as the GIR was numerically, but not significantly (65%, p=0.079), larger in the CS group compared to the WM group. A larger, ongoing study will investigate the effects of CS and the two components of CS (cagrilintide and semaglutide) individually on insulin sensitivity in DIO rats.

Disclosure: A. Secher: Employment/Consultancy; Novo Nordisk A/S. Stock/Shareholding; Novo Nordisk A/S. Other; Funded by Novo Nordisk A/S. Editorial and formatting support was provided by Sophie Bruce, of Apollo, OPEN Health Communications, funded by Novo Nordisk A/S.

76

Survodutide and semaglutide both induce weight loss but show different effects on food preference and dyslipidaemia in the free choice diet-induced obese hamster model

F. Briand ¹, R. Augustin², K. Bleymehl², B. Bajrami², T. Zimmermann², T. Klöckener², T. Klein²;

¹Physiogenex, Escalquens, France, ²Boehringer Ingelheim Pharma GmbH & Co. KG, Biberach, Germany.

Background and aims: Semaglutide, a GLP-1 receptor agonist, and survodutide, a GCGR/GLP-1 receptor dual agonist, achieves significant weight loss in patients living with obesity. To test whether this weight loss is associated with changes in food preference and improvements in dyslipidemia, we evaluated semaglutide and survodutide in the free choice diet-induced obese and dyslipidemic hamster, a model with a human-like lipoprotein metabolism.

Materials and methods: Obesity and dyslipidemia were induced with a 20-week free choice diet, which presents hamsters with a choice between control chow diet or a high fat/cholesterol diet, and plain water or 10% fructose water. Obese hamsters were then kept on this free choice diet and treated subcutaneously once daily with vehicle, semaglutide 15nmol/kg or survodutide 15nmol/kg for 5 weeks. The semaglutide dose was selected from previous studies showing significant body weight loss in this hamster model for obesity. The dose of survodutide was chosen based on glucagon receptor target engagement studies in this same model, demonstrating significant reductions in plasma levels of glucagon and amino acids.

Results: Compared with vehicle, semaglutide and survodutide led to a 17% and 11% body weight reductions, respectively (both p<0.001 vs vehicle). Compared with vehicle, semaglutide significantly lowered both chow diet and high fat diet intakes during the first week of treatment, but then returned to baseline. However, semaglutide significantly increased plain water intake and reduced fructose-enriched water during the whole treatment period (both p<0.001). Survodutide did not alter chow diet and plain water intakes, but significantly reduced high fat diet and fructose-enriched water intakes during the 5-week treatment period. Semaglutide and survodutide significantly reduced plasma insulin levels and the HOMA-IR index of insulin resistance (both p<0.01 vs. vehicle). While only semaglutide reduced plasma triglyceride levels significantly (-50%), both semaglutide and survodutide significantly reduced plasma total cholesterol levels by 24% and 41%, respectively (both p<0.001 vs. vehicle). The cholesterol-lowering effect of semaglutide was limited to HDL-cholesterol levels (-25%, p<0.001) without effects on LDL-cholesterol. In contrast, survodutide treatment was associated with cholesterol lowering in all lipoprotein fractions, including LDL-cholesterol (-39%, p<0.001).

Conclusion: Semaglutide and survodutide both induce weight loss and lower HOMA-IR but have different effects on food preference and dyslipidemia in the present obese hamster model, effects which may attributed to the glucagon component of survodutide. These preclinical data highlight the potential benefits of the GCGR/GLP-1R dual agonist survodutide for the treatment of obesity and related comorbidities.

Disclosure: F. Briand: Employment/Consultancy; Physiogenex.

77

Impact of GLP-1/GIP receptor signalling on human pancreatic adipose tissue organoids

E. Lorza-Gil ^1,2, O. Strauss¹, E. Ziegler¹, L. Sandforth^1,2, A. Sandforth^1,2, F. Berlth³, S. Singer⁴, A. Mihaljevic³, A. Birkenfeld^1,2, F. Gerst^1,2;

¹Institute for Diabetes Research and Metabolic Diseases of the Helmholtz Center Munich at University of Tübingen, Tübingen, Germany, ²Department of Internal Medicine IV, University Hospital Tübingen, Tübingen, Germany, ³Department of General, Visceral and Transplant Surgery, University Hospital Tübingen, Tübingen, Germany, ⁴Department of Pathology and Neuropathology, University Hospital Tübingen, Tübingen, Germany.

Background and aims: Pancreatic fat has been positively correlated with impaired insulin secretion in individuals at high risk for T2D, whereas recovery of beta cell function during T2D remission was associated with a reduction in pancreatic fat. GLP1R agonists and the dual GLP1R/GIPR agonist tirzepatide emerged as powerful drugs to rescue insulin secretion and induce weight loss in patients with T2D. However, data is missing on the role of incretins (GLP-1 and GIP) and of pharmacological modulation of their receptors (GIPR and GLP1R) in pancreatic fat accumulation. In this study, we aimed to investigate whether incretin hormones regulate adipogenesis and function of pancreatic adipocytes, using a novel human organoid model that closely mimics pancreatic adipose tissue phenotype.

Materials and methods: Mature adipocytes and stromal vascular fraction (SVF) cells were obtained following collagenase digestion of peripancreatic fat biopsies from non-diabetic patients (n=8). SVF spheroids were differentiated invitro for 19 days. Expression of adipogenic markers was analysed by RT-qPCR. Lipolytic activity was assessed by fluorometric detection of free fatty acid (FFA) and glycerol release, and by western blotting of P-thr197-PKA and P-ser660-HSL. Adipokines and chemokines secretion was measured with multiplex assay.

Results: During differentiation of adipocyte organoids, the mRNA levels of PPARG and ADIPOQ, two adipocyte differentiation markers, were upregulated 27-fold and 22000-fold respectively, reaching levels similar to those of native pancreatic adipocytes. The adipogenic capacity of pancreatic organoids, indicated by ADIPOQ mRNA level, negatively correlated with the donor’s body weight (p=0.01). GIPR mRNA level was upregulated upon organoid differentiation, whereas GLP1R mRNA was undetectable. Short-term treatment with 100 nM GIP and 10 nM tirzepatide stimulated lipolysis to a similar extent, as suggested by the increased release of FFA (15-fold) and of glycerol (1.7-fold), and the increased P-PKA and P-HSL, whereas exendin-4 was effectless. Chronic treatment of pancreatic organoids with GIP and tirzepatide did not alter expression of adipogenic markers nor desensitized lipolysis for stimulation with GIP/tirzepatide. Secretome analysis of GIP/tirzepatide-treated organoids indicated no change in adiponectin and leptin release. However, a 2-fold reduction of IL1B release was induced by GIP, suggesting a GIP-mediated impact on the spheroid’s resident CD68-positive immune cells.

Conclusion: This study reveals a broader metabolic impact of incretin signalling within the pancreas, i.e. pancreatic fat tissue, transcending the traditional focus on islet function. Activation of GLP1R/GIPR signalling by tirzepatide regulates lipolysis and cytokine release in pancreatic adipocytes, highlighting an additional therapeutic angel of this signalling axis in the treatment of obesity and T2D.

Supported by: BMBF 01GI0925

Disclosure: E. Lorza-Gil: None.

78

Petrelintide induced potent additional weight loss in combination with either semaglutide or tirzepatide in DIO rats

B. Vestergaard ¹, C. Wenander¹, J. Skarbaliene², J. Griffin¹;

¹Zealand Pharma A/S, Søborg, Denmark, ²Pharvaris GmbH, 6300 Zug, Switzerland.

Background and aims: Currently approved pharmacotherapy for weight management includes semaglutide and tirzepatide, both utilizing a backbone of GLP-1. Combination therapy consisting of hormone pharmacotherapies with complementary modes of action may add even further weight loss to that of semaglutide or tirzepatide without compromising safety and tolerability. Petrelintide, a novel once-weekly amylin analog currently in phase 1 clinical testing, induces meaningful reductions in body weight in lean and overweight subjects. In these studies, we investigated the combined weight loss effects of petrelintide plus semaglutide and petrelintide plus tirzepatide in diet-induced obese (DIO) rats.

Materials and methods: The Pharmacological effects were assessed in both studies by measurement of body weight and food intake. Study I: DIO rats were treated for 2 weeks with either vehicle or semaglutide 8 nmol/kg by once daily s.c. injection. From day 14 to day 34, vehicle-treated rats continued either on vehicle or were switched to petrelintide 10 nmol/kg, every 2nd day by s.c. injection. Likewise, semaglutide-treated rats continued either on semaglutide or were switched to semaglutide plus petrelintide. Study II: DIO rats were treated for 2 weeks with either vehicle or tirzepatide 10 nmol/kg by once daily s.c. injection. From day 14 to day 34, vehicle-treated rats continued either on vehicle or were switched to petrelintide 10 nmol/kg, every 2nd day by s.c. injection. Likewise, tirzepatide-treated rats continued either on tirzepatide or were switched to tirzepatide plus petrelintide.

Results: Study I: Combined treatment with petrelintide and semaglutide achieved a significant, sustained, and greater reduction in cumulative food intake and a corresponding significant, sustained and greater body weight reduction compared to vehicle- and each mono-treated group (-4.7% ±0.8 petrelintide, -8.0% ± 0.8 semaglutide, -16.1% ± 0.7 combination, +7.0% ± 1.1 vehicle; relative to initial body weights ± SEM). Study II: Combined treatment with petrelintide and tirzepatide achieved a significant, sustained, and greater reduction in cumulative food intake and a corresponding significant, sustained and greater body weight reduction compared to vehicle- and each mono-treated group (-5.2% ±0.7 petrelintide, -8.2% ± 1.1 tirzepatide, -15.3% ± 0.9 combination, 6.5% ± 0.4 vehicle; relative to initial body weights ± SEM).

Conclusion: In conclusion, combining petrelintide with either semaglutide or tirzepatide in DIO rats showed significantly greater body weight loss compared to the mono treatments. Combining petrelintide with either semaglutide or tirzepatide may potentially be a future therapy approach for weight management of people living with overweight and obesity. Petrelintide is currently being investigated in a multiple ascending dose (MAD) trial to assess the clinical potential for weight management.

Disclosure: B. Vestergaard: None.

OP 14 Incretins: How many targets do you need?

79

Short-acting glucagon-like peptide-1 receptor agonist therapy delays gastric emptying during insulin-induced hypoglycaemia as well as euglycaemia in type 2 diabetes

R.J. Jalleh ^1,2, D. Quast², M. Kamruzzaman², J. Grivell², S. Hatzinikolas², T.A. Murthy², K.L. Jones^1,2, M. Horowitz^1,2, C.S. Marathe^1,2;

¹Endocrine and Metabolic Unit, Central Adelaide Local Health Network, Adelaide, Australia, ²Adelaide Medical School, The University of Adelaide, Adelaide, Australia.

Background and aims: Insulin-induced hypoglycaemia accelerates gastric emptying (GE) substantially in people with and without diabetes. In contrast, glucagon-like peptide-1 receptor agonists (GLP-1RAs) may delay GE markedly, particularly when short-acting. It is not known whether GLP-1RAs attenuate the counter-regulatory acceleration of GE that occurs in response to insulin-induced hypoglycaemia; if so, this may predispose to postprandial hypoglycaemia when a GLP-1RA is combined with insulin. The aim of this double-blind, placebo-controlled, randomised, cross-over study is to evaluate if the GLP-1RA exenatide twice daily (BD) attenuates the acceleration of GE during insulin-induced hypoglycaemia in adults with type 2 diabetes (T2D).

Materials and methods: Seven adults with T2D managed either with diet and/or metformin [age 62.7±1.6 years, 4M, 3F, HbA1c 6.8±0.1% (50.8±1.3 mmol/mol), BMI 31.7±0.9 kg/m², median duration of T2D 5 years (IQR 4.3-10 years)] each participated in 4 study visits. They were initially randomised to either placebo (0.9% saline s. c.) or exenatide BD (uptitrated to 10 microgram s. c.) for 4 weeks and then randomised to either a euglycaemic (6 mmol/L) or hypoglycaemic (2.6 mmol/L) clamp. After 60 min at the target blood glucose concentration, they consumed a beef patty (270 kcal) labelled with 20MBq 99mTc sulphur colloid and 150ml unlabelled water. GE was monitored by scintigraphy for over 3 hours. After 3 days, the alternate clamp study was performed. After completing both clamp studies, there was a 1-week washout period and then all participants crossed over treatment arms (placebo/exenatide) and had the clamp studies repeated in the same manner. Both participants and investigators were blinded to the study drug used and participants were blinded to the clamp protocol conducted. The primary outcome was the difference in the percentage of food retained at 100 min between hypoglycaemic and euglycaemic clamps. Data, analysed using a two-tailed, paired t test, are presented as mean±SEM. As a secondary outcome, the areas under curve (AUCs) of GE were assessed with a two-way ANOVA.

Results: Exenatide BD delayed GE during both euglycaemia (placebo 48.8±12.9% food retained vs exenatide 90.7±6.0%, P=0.01) and hypoglycaemia (placebo 34.8±14.6% vs exenatide 78.7±12.1%, P=0.02). During hypoglycaemia, the intragastric retention of food at 100 min was reduced by 17.8±7.1% with placebo and 14.1±8.1% with exenatide (P=0.77). There was a difference in the AUCs for the four conditions (P<0.0001) and a treatment x time effect for total GE (P<0.0001).

Conclusion: In adults with T2D, exenatide BD delays GE markedly during both euglycaemia and hypoglycaemia, but the magnitude of the acceleration of GE by insulin-induced hypoglycaemia is comparable.

Clinical Trial Registration Number: ACTRN12622001506785

Supported by: RAH Clinical Project Grant (RAH2169)

Disclosure: R.J. Jalleh: None.

80

Neither exogenous nor endogenous glucose-dependent insulinotropic polypeptide affects appetite or energy intake acutely in healthy older people

E. Hosseini Marnani ¹, S. Hatzinikolas¹, A.C. Slavotinek¹, M. Kamruzzaman¹, R.J. Jalleh¹, K. Lange¹, S. Veedfald², L.S. Gasbjerg³, M.M. Rosenkilde², J.J. Holst², T. Wu¹, C.K. Rayner¹, C.S. Marathe¹, M. Horowitz¹, K.L. Jones¹;

¹The University of Adelaide, Adelaide, Australia, ²Copenhagen University, Copenhagen, Denmark, ³University of Copenhagen, Copenhagen, Denmark.

Background and aims: The incretin hormones, glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP), released following a meal, are pivotal to the regulation of postprandial blood glucose. GLP-1 and related agonists also reduce appetite and energy intake to facilitate weight-loss in individuals with obesity, with and without type 2 diabetes. Tirzepatide, a recently available co-agonist of GLP-1 and GIP receptors, is effective in the management of obesity. Unlike GLP-1, the effect of GIP on energy intake is uncertain, in part because of the lack of a specific GIP antagonist. The first GIP receptor antagonist suitable for use in humans, GIP(3-30)NH₂, has been developed by our group. We have evaluated the acute effects of both exogenous and endogenous GIP on appetite and energy intake following a nutrient load in healthy older people.

Materials and methods: After an overnight fast (12 hours), 18 healthy older participants (13F, 5M; mean age 67.7±0.7 years; BMI 28.1±0.9 kg/m²) attended the Clinical Research Facility (CRF) at the University of Adelaide on 3 separate occasions. Participants were randomised to receive an intravenous infusion of either (i) GIP (1.5 pmol/kg/min), (ii) the specific GIP receptor antagonist, GIP(3-30)NH₂, (800 pmol/kg/min) or (iii) 0.9% saline (placebo) in a double-blind, cross-over design, for 240min (t=-30-210min). At t=-2 min participants ingested a 75g glucose drink (300ml) and at t=180min, they were presented with a standardised buffet style meal and allowed to eat ad-libitum over 30min (until t=210min). Measurements of appetite (hunger and fullness) were obtained at regular time intervals, using validated visual analogue scales, from immediately before the infusion until t=210min. Energy intake (kcal) at the buffet meal was quantified using Foodworks software. Data are presented as mean values ± SEM.

Results: All study days were well-tolerated and there were no adverse events. Baseline hunger and fullness were comparable on each day. Neither exogenous nor endogenous GIP affected sensations of hunger or fullness (either fasting, after the glucose drink, or following the meal; Figure), or energy intake at the buffet meal (Placebo: 679 ± 68 kcal vs GIP: 699 ± 87 kcal vs GIP(3-30)NH₂: 634 ± 53 kcal; P=0.51).

Conclusion: In healthy older people, neither exogenous nor endogenous GIP has an acute effect on appetite or energy intake. These outcomes suggest that GIP does not modulate appetite in humans.

Supported by: Royal Adelaide Hospital Project Grant; NHMRC Project Grant ID: 1163910

Disclosure: E. Hosseini Marnani: None.

81

Novel unimolecular tetra-agonist peptides targeting GLP-1, GIP, amylin, and calcitonin receptors with enhanced metabolic benefits in animal models of obesity

C.M. Rondinone ¹, M.P. Valdecantos², P. Rada², A. Valverde², S. Ghosh³;

¹Pep2tango Therapeutics, Potomac, MD, USA, ²Instituto de Investigaciones Biomedicas Sols-Morreale, Madrid, Spain, ³Doon Associates LLC, San Diego, CA, USA.

Background and aims: The prevalence of obesity and associated co-morbidities necessitates innovative approaches for safe and efficacious therapies. Herein, we describe the characterization of PTT-A, a novel long-acting peptide agonist designed to target the GLP-1, GIP, Amylin and Calcitonin receptors, in rodent models and assessed its efficacy against the dual GIPR/GLP-1R agonist Tirzepatide.

Materials and methods: In vitro assays using recombinant cell lines expressing individual receptors were used to assess GLP-1, GIP, Amylin and Calcitonin receptor agonism. Multiple metabolic endpoints were examined in rodents, including acute food intake and calcium regulation effects of PTT-A in lean rats, acute glucose-lowering effects in lean mice, and its chronic effects in diet-induced obesity (DIO) rats compared to Tirzepatide.

Results: In in vitro assays, PTT-A demonstrated potent agonistic activity towards GIP, GIP, Amylin, and Calcitonin receptors for cAMP production. In acute in vivo studies, PPT-A elicited significant reductions in blood glucose levels (44.2±1.2% at 120 min; p<0.001 vs. baseline, n=5) in lean mice, and a decrease in calcium levels (6.4±1.5% at 8 h; p<0.01 vs. vehicle, n = 5) in lean rats. Furthermore, PTT-A exhibited a dose-dependent decrease in cumulative food intake in lean rats after a single dose. Chronic administration in DIO rats over 14 days led to substantial reductions in cumulative food intake (40.3±16.1%; p<0.001 vs. vehicle, n=5) and body weight (15.0±0.7%; p<0.001 vs. vehicle, n=5), primarily driven by reductions in fat mass with no loss of muscle mass. Inguinal and epididymal fat mass were reduced by 43.4±7.1% (p<0.001 vs. vehicle, n=5) and by 23.3±7.9% (p<0.01 vs. vehicle, n=5) respectively in PTT-A treated animals. In contrast, Tirzepatide, at equivalent doses, induced a reduction in body weight by 9.9 ±0.5% (p<0.001 vs. vehicle, n=5) with a decrease in both fat and muscle. Epididymal fat was reduced (22.5±4.2%; p<0.01 vs. vehicle, n=5) with no significative changes in inguinal fat. Gastrocnemius and soleus muscle weight were reduced by 28.6±4.6% (p<0.001 vs. vehicle, n=5) with Tirzepatide, whereas there was no reduction in muscle weight after treatment with PTT-A. Moreover, PTT-A demonstrated robust efficacy in glucose and plasma lipid lowering, insulin sensitization, and liver health, evidenced by reductions in plasma insulin (78.7±4.8%; p<0.001 vs. vehicle, n=5), alanine aminotransferase (48.8±3.7%; p<0.001 vs. vehicle, n=5), plasma triglycerides (54.7 ± 3.5%; p<0.001 vs. vehicle, n=5), and hepatic triglycerides (64.49 ± 7.0%; p<0.001 vs. vehicle, n=5), surpassing Tirzepatide’s effects at equivalent doses.

Conclusion: PTT-A emerges as a promising novel tetra-agonist, offering superior outcomes in weight loss, glycemic control, insulin sensitization, and liver health compared to clinically relevant GLP-1/GIP dual agonists. These findings underscore the potential of tetra-agonists as a multifaceted therapeutic option for addressing the complex challenges associated with obesity-related metabolic disorders

Disclosure: C.M. Rondinone: Employment/Consultancy; Pep2tango Therapeutics Inc. Stock/Shareholding; Pep2tango Therapeutics Inc.

82

Characterising the pharmacology of glucagon receptor agonism in a dual GCGR/GLP-1R agonist

T. Kloeckener, T. Zimmermann, P. Haebel, B. Bajrami, R. Augustin;

Boehringer Ingelheim Pharma GmbH & Co. KG, Biberach an der Riß, Germany.

Background and aims: Survodutide is a dual GCGR/GLP-1R agonist. The contribution of GCGR agonism to the pharmacological activity of survodutide has been based on target engagement markers.

Materials and methods: Here, we aimed to investigate the pharmacology of a long-acting GCGR agonist in vitro and in mice. BI2130086 is a synthetic peptide, which selectively activates the GCGR.

Results: BI2130086 agonizes the human and mouse GCGR in 100% plasma with a potency (EC₅₀) of 4.1 nM and 1.23 nM, respectively. Application of BI2130086 to lean mice did not significantly affect food intake, but significantly decreased plasma amino acids. As expected for GCGR agonism, BI2130086 significantly increased glucose excursion after an intraperitoneal glucose challenge. The body weight-lowering efficacy of BI2130086 (10, 30 and 100 nmol/kg) was assessed in diet-induced obese mice. After 4 weeks of daily dosing, 30 and 100 nmol/kg showed a reduction of body weight of 6% and 34% compared to vehicle (p≤0.001 vs vehicle). Mechanistically, the body weight-lowering effect of BI2130086 was attributed to an increase in energy expenditure (measured at 100 nmol/kg). The body weight-lowering effect of 100 nmol/kg BI2130086 was associated with a 62% reduction of fat mass, a 94% reduction in liver triglycerides and 76% reduction of plasma cholesterol (p≤0.001 vs vehicle). However, sub-chronic agonism of the GCGR was associated with significant hyperglycemia due to increased gluconeogenesis.

Conclusion: We conclude that GCGR agonism alone provides body weight lowering efficacy by increasing energy expenditure, but is associated with a hyperglycemia risk, which can be mitigated by the glucose-dependent insulinotropic activity of GLP-1R agonism. Survodutide has been shown to be a balanced GCGR/GLP-1R dual agonist that combines both GCGR and GLP-1R agonism in one molecule. It is currently being investigated in the SYNCHRONIZE trials in people living with overweight and comorbidities, or with obesity and without type 2 diabetes.

Disclosure: T. Kloeckener: None.

83

Effects of Liraglutide on GLP-1, GIP and glucagon concentrations in persons with type 2 diabetes with multiple daily insulin injections during a 24-week period

M. Koci ¹, P.-A. Jansson², P. Fatulla², K. Bergqvist³, S. Seyed Ahmadi², H. Imberg², I. Hirsch⁴, O. Mellander⁵, M. Lind²;

¹Department of Endocrinology and Diabetes, Sahlgrenska University Hospital, Gothenburg, Sweden, ²Department for Molecular and Clinical Medicine, Gothenburg University, Göteborg, Sweden, ³Department of Infectious Diseases, Gothenburg University, Göteborg, Sweden, ⁴Division of Metabolism, Endocrinology and Nutrition, UW Medicine Diabetes Institute, Seattle, WA, USA, ⁵Department of Clinical Sciences Malmö, Lund University, Malmö, Sweden.

Background and aims: Although extensive knowledge has been gained of metabolic and cardiovascular effects of GLP-1 analogues, placebo-controlled trials are sparse examining how endogenous incretin hormones are affected during therapy. The aim of this study was to evaluate the concentrations of GLP-1, GIP and glucagon in liraglutide treated persons with type 2 diabetes (T2D) and associations with changes in glycemic control.

Materials and methods: The current analysis included 106 patients with available biobank samples of the 124 persons with T2D treated with multiple daily insulin injections (MDI) who took part in the double-blind placebo controlled MDI liraglutide trial. Differences in GLP-1, GIP and glucagon between treatment groups were evaluated using analysis of covariance, adjusting for baseline values. Linear regression was used to evaluate baseline concentrations of GLP-1, GIP, and glucagon as possible predictors for a greater or smaller treatment effect of liraglutide on HbA1c.

Results: At 12 weeks GLP-1 was 25% lower (95% CI 9-37%, p=0.003) and GIP 39% higher (95% CI 8-79%, p=0.012) in the liraglutide group compared to the placebo group. At 24 weeks differences were sustained. Glucagon showed a tendency towards lower concentrations in the liraglutide group at 12 weeks (p=0.052) but reached similar levels as the placebo group after 24 weeks (p=0.98). In the liraglutide group, the total insulin dose decreased from a mean of 104 units (SD 45) at baseline to 85 units (SD 41) at 24 weeks. Similarly, mean Hba1c decreased from 9.0% to 7.4%. Each 30% decrease in glucagon concentration at baseline was associated with 0.3% greater reduction in HbA1c at 24 weeks of treatment with liraglutide compared to placebo (95% CI 0.1-0.5%, p=0.001). Baseline GLP-1 and GIP levels had no association with liraglutide effect.

Conclusion: An essential mechanism of metabolic effects of GLP-1 analogues is potentially mediated via increased GIP levels in persons with T2D. According to our findings serum glucagon levels may be an important biomarker for GLP-1 analogue responders.

Disclosure: M. Koci: None.

84

Effects of endogenous and exogenous glucose dependent insulinotropic polypeptide on blood pressure, heart rate, gastric emptying and blood glucose responses to oral glucose in healthy older people

R.J. Jalleh ^1,2, S. Hatzinikolas², A.C. Slavotinek², M. Kamruzzaman², K. Lange², C.H. Malbert³, S. Veedfald⁴, L.S. Gasbjerg⁴, M.M. Rosenkilde⁴, J.J. Holst⁴, C.S. Marathe^1,2, T. Wu², C.K. Rayner², M. Horowitz^2,1, K.L. Jones^2,1;

¹Endocrine and Metabolic Unit, Royal Adelaide Hospital, Adelaide, Australia, ²The University of Adelaide, Adelaide, Australia, ³Aniscan, INRAE, Saint-Gilles, France, ⁴University of Copenhagen, Copenhagen, Denmark.

Background and aims: Postprandial hypotension (PPH), a fall in blood pressure (BP) >20mmHg following a meal, occurs in ~ 15-20% of healthy people > 65 years and ~35% of people with type 2 diabetes. It is associated with a markedly increased risk of falls and is an independent predictor of mortality. The fall in BP reflects the increase in splanchnic blood flow and is greater when gastric emptying is more rapid. Glucagon-like peptide-1 (GLP-1) and GLP-1 receptor agonists attenuate the postprandial fall in BP, in part by slowing gastric emptying. In contrast to GLP-1, the effects of glucose dependent insulinotropic polypeptide (GIP) on BP and gastric emptying are poorly defined. We have used the specific GIP receptor antagonist, GIP(3-30)NH₂, to assess the effect of GIP on the BP, heart rate (HR), gastric emptying and blood glucose responses to a glucose drink in healthy older people.

Materials and methods: 18 healthy older participants (13F, 5M; mean age 67.7±0.7 years; BMI 28.1±0.9 kg/m²) attended the laboratory on 3 occasions after an overnight fast. Intravenous cannulae were inserted into both arms for blood sampling and intravenous infusion of (i) GIP (1.5 pmol/kg/min), (ii) GIP(3-30)NH₂ (800 pmol/kg/min) or (iii) 0.9% saline (placebo) for 240min (t=-30-210min) in a randomised, double-blind, cross-over design. Participants were seated and BP, mean arterial pressure (MAP) and HR monitored using an automated device. At t=-2 min participants ingested a 75g glucose drink (300ml) radiolabelled with 20MBq ^99mTc-calcium phytate to assess gastric emptying by scintigraphy, and at t=180min they were offered a buffet style lunch. Data are means±SEM.

Results: All studies were well-tolerated. There was a modest slowing of gastric emptying by exogenous GIP (P<0.05) with no effect of GIP(3-30)NH₂. Exogenous GIP and GIP(3-30)NH₂ had no effect on MAP, HR or blood glucose during the fasting period. Following the glucose drink, MAP fell (P<0.01) on all study days. The magnitude of the fall in MAP (P=0.03), and rise in HR (P=0.01) was less after GIP(3-30)NH₂ compared to placebo (Figure). Peak blood glucose was greater (P<0.001) after GIP(3-30)NH₂ (12.2±0.4 mmol/L), with no effect of exogenous GIP (10.6±0.4 mmol/L) when compared to placebo (10.6±0.5 mmol/L).

Conclusion: In healthy older people, the fall in BP and rise in HR after oral glucose are attenuated when GIP release is blocked. Accordingly, endogenous GIP, beyond its established role in glucose metabolism, modulates postprandial cardiovascular responses.

Supported by: Royal Adelaide Hospital Project Grant and NHMRC Project Grant ID: 1163910

Disclosure: R.J. Jalleh: None.

OP 15 Cardiovascular risk in patients with type 1 diabetes: What do we know today, how can we modify it tomorrow?

85

All-cause and cause-specific mortality, cardiovascular diseases, and prognostic factors in adult-onset type 1 diabetes

Y. Wei ¹, T. Andersson^1,2, T. Tuomi^3,4, T. Nyström⁵, S. Carlsson¹;

¹Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden, ²Stockholm County Council, Stockholm, Sweden, ³Department of Endocrinology, Abdominal Center, Helsinki University Hospital, Helsinki, Finland, ⁴Folkhälsan Research Center, Helsinki, Finland, ⁵Department of Clinical Science and Education, Karolinska Institutet, Stockholm, Sweden.

Background and aims: Type 1 diabetes (T1D) with adult-onset is prevalent yet scarcely investigated. Our aim was to provide new knowledge on the prognosis of adult-onset T1D. Specifically, we investigated risks of all-cause mortality, cause-specific mortality, and major adverse cardiovascular events (MACE) and assessed the role of different prognostic factors.

Materials and methods: This nationwide, population-based cohort study included people diagnosed with T1D (n=10,184) and type 2 diabetes (T2D, n=375,523) at age ≥18 years in 2001-2020 from the Swedish National Diabetes Register, which continuously collected data on lifestyle and clinical characteristics since diabetes diagnosis. We also included 509,172 population controls matched to people with T1D. Participants were followed for death and MACE (cardiovascular death or nonfatal myocardial infarction/stroke) through linkages to National Patient and Causes-of-Death Registers until 2022.

Results: During 10.2 (median) years of follow-up, we identified 816 deaths and 408 MACE events in people with T1D. Compared to population controls, people with T1D had higher risks of all-cause mortality (HR: 1.72; 95% CI 1.65, 1.85), cardiovascular deaths, non-cardiovascular deaths (deaths due to cancer and infection), and MACE (HR: 1.31 [ 1.18, 1.45]). They also had higher risks of all-cause mortality and non-cardiovascular deaths (deaths due to cancer, infection, diabetes) but lower MACE incidence than individuals diagnosed with T2D at similar ages. Smoking, BMI, HbA1c, blood pressure, and triglycerides were associated with mortality and MACE while physical activity was associated with mortality in T1D. There was a dose-response relationship between the number of such risk factors and MACE/mortality in people with T1D (Figure 1).

Conclusion: Adult-onset T1D was associated with elevated risks of all-cause mortality, cause-specific mortality, and MACE. Unhealthy lifestyle factors and uncontrolled metabolic profiles all conferred worse prognosis, indicating the need to intervene on such modifiable factors in people with T1D.

Supported by: The Swedish Research Council, Swedish Diabetes Foundation, FORTE, CSC

Disclosure: Y. Wei: None.

86

Preclinical carotid atherosclerosis is independtly associated with incident cardiovascular events in patients with type 1 diabetes

C. Milad ¹, C. Sola¹, M. Claro¹, K. Castillo¹, T. Serés-Noriega¹, L. Boswell², C. Viñals¹, E. Esmatjes¹, V. Perea³, J. Blanco¹, A. Mesa^1,4, I. Conget¹, M. Giménez¹, A. Amor Fernandez¹;

¹Hospital Clinic de Barcelona, Barcelona, Spain, ²Althaia Universitary Health Network, Manresa, Spain, ³Hospital Universitari Mútua de Terrassa, Terrassa, Spain, ⁴Hospital Santa Creu i Sant Pau, Barcelona, Spain.

Background and aims: Vascular ultrasound is recommended for cardiovascular (CV) risk reclassification in the general population, although its use in type 1 diabetes (T1D) patients has not been studied. We aimed to assess whether carotid atherosclerosis predicts incident CV events (CVE) in this population.

Materials and methods: Cohort study in T1D patients in primary prevention undergoing a CV risk assessment program (including vascular carotid ultrasound). Inclusion criteria for the program were: ≥40 years, diabetic kidney disease (DKD), and/or ≥10 years of T1D duration plus another CV risk factor. The primary composite endpoint includes the first cardiovascular event (CVE): coronary heart disease (CHD), stroke, peripheral artery disease (PAD) or heart failure (HF). The combination of CVE and total mortality was considered the secondary endpoint of our study.

Results: We followed n=484 individuals (46% females, mean age 48.9±10.2 years, T1D duration 27.6±10.1 years) for a mean of 4.7 years. We observed 18 CVEs (44% PAD, 39% CHD, 11% stroke, 6% HF) and 5 deaths (unrelated to CVEs). Patients with a CVE had more frequently microvascular complications (DKD, 33.3% vs. 10.1%; diabetic retinopathy, 61.1% vs. 37.8%) ) and higher levels of triglycerides (101 [81-148] vs. 75 [59-101] mg/dL), remnant cholesterol (20 [16-28] vs. 15 [12-20] mg/dL), and total leukocytes (8390±1860 vs. 6930±1960 per mm3; p<0.05 for all comparisons) at baseline. While plaque presence was not significantly associated with the primary or secondary endpoint (log-rank test=0.137 and 0.084, respectively), having ≥3 plaques was associated with future CVE (p=0.01) and the combined endpoint of CVE plus death (p=0.002). Adjustment for age and sex (Cox-regression models; HR 3.53 [1.15-10.78]; p=0.027) and for other CV risk factors (hypertension, statin treatment, past or current smoking and diabetes duration) and microvascular complications did not change the association with CVE (HR 3.21 [1.01-10.23]; p=0.048).

Conclusion: High atherosclerotic burden (≥3 plaques) is independently associated with incident CVEs in T1D. This is the first study that establishes the role vascular ultrasound as a reclassification tool in this high CV risk population.

Disclosure: C. Milad: None.

87

People with type 1 diabetes living with overweight or obesity are at increased cardiovascular risk: insights from the SFDT1 cohort

L. Salle ¹, J. Julla², G. Aguayo³, S. Tatulashvili⁴, H. Hanaire⁵, L. Sablone⁶, N. Servy⁷, E. Renard⁸, J.-F. Gautier⁹, G. Prevost¹⁰, N. Jeandidier¹¹, B. Guerci¹², E. Cosson¹³, G. Fagherazzi¹⁴, J.-P. Riveline⁹;

¹CHU Limoges, Limoges, France, ²Institut Necker Enfants Malades, INSERM U1151, CNRS UMR 8253, IMMEDIAB Laboratory, Paris, France, Department of Endocrinology and Diabetes, Centre Universitaire du diabète et de ses complications, A, Paris, France, ³Luxembourg Institute of Health, Strassen, Luxembourg, ⁴Ile de France, Université Sorbonne Paris Nord, Bobigny, France, ⁵Diseases & Nutrition, Toulouse Hospital, Toulouse, France, ⁶Société francophone du Diabète, Paris, France, ⁷Sanoia, Aubagne, France, ⁸Dept. of Endocrinology, Diabetes, CHU Montpellier, Montpellier, France, ⁹University Paris 7, Lariboisiere Hospital, Paris, France, ¹⁰CHU ROUEN, Rouen, France, ¹¹Endocrinology, Diabetes and Nutrition,, CHRU Strasbourg, Strasbourg, France, ¹²Diabetologie, Maladies Metaboliques & Nutrition, Hôpital Brabois, CHU, Vandoeuvre Lès Nancy, France, ¹³Diabetology-Endocrinology-Nutrition, Université Sorbonne Paris Nord, Bobigny, France, ¹⁴Deep Digital Phenotyping Research Unit, Luxembourg Institute of Health, Strassen, Luxembourg.

Background and aims: Individuals with type 1 diabetes (T1D) are faced with a growing prevalence of overweight or obesity, a recognized determinant of cardiovascular (CV) risk. This study aims to evaluate the metabolic profile, CV risk, and diabetes-related complications across body mass index (BMI) classes in individuals with T1D in primary CV prevention.

Materials and methods: Using data from SFDT1, a multicentric French cohort study, persons with T1D were stratified into 3 BMI classes: normal weight (18.5-24.9 kg/m²), overweight (OW) (25-29.9 kg/m²), and obesity (OB) (BMI ≥30 kg/m²). Statistical analyses were performed with mean ± SD or frequency (%), with group comparisons through Kruskall-Wallis, variance analysis, and Chi-squared tests. A validated insulin sensitivity (IS) score was applied to estimate IS across BMI classes. CV risk and end-stage kidney disease risk were evaluated with the 5 and 10-year Steno T1 Risk Engine score (ST1RE). Correlations between CV risk and BMI, waist circumference and IS were performed with Spearman correlation.

Results: Among 2,554 participants, 7.3% had a CV history. The study was conducted on 2,367 patients in primary prevention including 50.7% with normal weight, 30.7% with OW and 18.5% with OB. Age (33, 39 and 40 years), duration of diabetes (18, 21 and 23 years), high blood pressure (19.8, 32.4 and 41.2%) and dyslipidemia (38.9, 51.8 and 65.6%) increased with BMI class (p<0.0001). Tobacco use decreased significantly (24, 22.3 and 12.8%, p<0.0001). The mean HbA1c remained comparable (7.5% overall, p=0.597) despite the use of more complex insulin administration methods (open and hybrid closed loop) in patients with OW and OB (p<0.0001). With higher BMI class, IS score decreased progressively (9.57, 7.49 and 5.53, p<0.0001). CV risk at 5 and 10 years rose from 3.78 to 5.26% and from 7.24 to 9.93% between normal weight and OB (p<0.0001). Five and 10-year ST1RE correlated positively with BMI (r=0.161, p<0.001) and waist circumference (r=0.328, p<0.0001) and negatively with IS (r= -0.329, p<0.001). End-stage kidney disease risk estimated with ST1RE increased significantly with BMI class (1.15, 1.31 and 1.37%, p<0.0001).

Conclusion: T1D persons with OW and OB had higher CV risk scores at 5 and 10 years. Visceral adiposity and IS were associated with increased CV risk scores.

Disclosure: L. Salle: None.

88

Increased body mass index and risk of developing cardiovascular complications in a cohort of people with type 1 diabetes

E. Soto-Pedre ¹, L.D. Petty², R.J. McCrimmon³, E.R. Pearson¹;

¹Division of Population Health & Genomics, University of Dundee, Dundee, UK, ²University of Dundee, Dundee, UK, ³Division of Systems Medicine, University of Dundee, Dundee, UK.

Background and aims: Obesity is reported to have increased adverse cardiometabolic consequences in people with Type 1 diabetes (T1D). However, the extent of the association between increased Body Mass Index (BMI) and cardiovascular complications in these patients remains unclear, as it comes mostly from cross-sectional analyses. To address this knowledge gap, we conducted a study to estimate this risk in a large cohort of T1D individuals based on their BMI.

Materials and methods: Retrospective follow-up study included people with T1D aged more than 18 years at diabetes diagnosis from Tayside/Fife (Scotland, UK). The period of follow-up for the analysis was 1995-2019. Electronic medical records-linkage technology was used. Last available BMI measurement within 3 years post- diabetes diagnosis was collected as the baseline data. All cardiovascular outcomes were incident and defined from hospital admissions. Hypertension was defined as systolic blood pressure ≥140mmHg and/or diastolic blood pressure ≥90 mmHg on at least 3 occasions and/or the use of antihypertensive medications. Other covariates considered were age, sex, social deprivation, smoking history, renal disease, and baseline blood pressure, HbA1c and serum cholesterol. Survival analyses were used to accommodate varying follow-up durations among patients. Adjusted Cox models were also stratified by the year of diabetes diagnosis. BMI was coded both as binary and as dummy variables to account for three exposure levels (normal: BMI<25, overweight: BMI=25-29, and obese: BMI≥30 Kg/m²). Competing risk regression models were used to test robustness, where death was the competing event. Data analysis was performed using STATA/MP® version 15.1 software, with statistical significance set at P <0.05.

Results: The study identified 1,973 patients with T1D (42% female, mean age=34.3 years), and 18.9% were obese at baseline. During a mean follow-up of 8.5 years, the Kaplan-Meier estimates showed a lower survival probability of remaining free of hypertension for those overweight and obese at baseline. After 10 years of follow-up, this probability was 57% for those with overweight and 41% for those obese, compared to 74% for those with normal BMI. Adjusted Cox models showed that overweight patients had an increased hazard ratio (HR) for developing hypertension (HR= 1.66, 95%CI 1.12-2.45) which nearly doubled in patients with obesity (HR= 3.18, 95%CI 1.93-5.24). Patients with T1D and obesity had also increased risk of Major Adverse Cardiovascular Events- MACE (HR=3.36, 95%CI 1.10- 10.29), acute myocardial infarction (HR=3.33, 95%CI 1.31- 8.53) and diabetic retinopathy (HR=1.35, 95%CI 1.02- 1.79). Competing risk regression models showed similar results.

Conclusion: Although we acknowledge certain limitations in our study as BMI was evaluated only at baseline, this large cohort study underscores a significant higher risk of developing MACE (acute myocardial infarction), retinopathy and arterial hypertension in T1D individuals with higher BMI. These results emphasize the importance of incorporating weight management into the overall management and guidance of individuals with T1D. To implement this effectively in clinical practice, further prospective studies assessing the safety and efficacy of weight loss in these patients are warranted.

Supported by: IMI 2 grant agreement No 875534

Disclosure: E. Soto-Pedre: None.

89

Evaluation of the Steno-risk in predicting cardiovascular events in young type 1 diabetes and its association with the progression of microvascular complications progression

I.C. Paliares, P.M. Dualib, L.S.N. Torres, P.M.T. Aroucha, J.R. de Sá, S.A. Dib;

Department of Medicine, Federal University of Sao Paulo, Sao Paulo, Brazil.

Background and aims: Type 1 diabetes (T1D) is associated with a two to four times greater risk of a first cardiovascular (CV) event compared to the age-adjusted population without diabetes, occurring more than a decade earlier than expected. Identifying individuals at higher CV risk through stratifying tools is crucial for implementing more aggressive preventive strategies. The Steno Type 1 Risk Engine (ST1RE) was developed to aid clinical decisions in primary prevention for patients with T1DM, as existing CV risk models for the general population and type 2 diabetes tend to underestimate CV risk in T1DM. However, the applicability of ST1RE in different populations remains uncertain, as prediction models developed for one population may not accurately estimate risk in another. This study aims to evaluate the performance of the ST1RE in predicting CV among ethnically mixed young T1D individuals and its association with the progression of microangiopathy complications.

Materials and methods: A retrospective survey of 435 adults with T1D, free of CV events at baseline, was assessed by ST1RE for chronic diabetes complications at 5 and 10 years of follow-up. Individuals were stratified into three risk groups (<10%: low; 10-20%: moderate; ≥ 20%: high) for ten-year cardiovascular risk estimation. The estimated CV risk rates were compared with the observed rates at 5 and 10 years using statistical analyses, including t-tests, c-statistics, Hosmer-Lemeshow tests, and logistic regression.

Results: Among 435 patients (mean age: 25 years; interquartile range (IQR): 21-32) with T1D duration of 13 years (IQR: 9-18 years), 84% were in the low-risk group (LRG), 11% in the moderate-risk group (MRG), and 5% in the high-risk group (HRG) for CV risk as determined by ST1RE. During a follow-up period of 8.5 ± 2.6 years, 5.5% experienced a CV event (1.6%, 14.9%, and 50% respectively from the LRG, MRG, and HRG). The odds ratio (OR) for experiencing a CV event was higher in the HRG compared to the LRG (OR: 60.0, 95% CI: 18.8-191.7, p<0.001) and compared to the MRG (OR: 5.7, 95% CI: 1.8-16.2, p<0.001), as well as in the MRG compared to the LRG (OR: 10.5, 95% CI: 3.4-32.8, p<0.001). The ST1RE-estimated CV event rates were similar to the expected rates both at 5 years (3.45% vs. 3.4%, p=0.84) and at 10 years (6.1% vs. 5.8%, p=0.41) of follow-up. Progression of microangiopathies was higher in the HRG than in the LRG according to ST1RE for retinopathy (p=0.002) and nephropathy (p<0.001), and higher in the MRG than in the LRG for autonomic neuropathy (p=0.018) and nephropathy (p=0.019), Figure 1.

Conclusion: ST1RE performs well in predicting CV events at 5 and 10 years of follow-up, and a higher ST1RE score was associated with the progression of microangiopathy complications in this genetically heterogeneous T1D population

Disclosure: I.C. Paliares: None.

90

Cumulative glycaemic exposure is associated with coronary artery disease in type 1 diabetes: a call for action

R. Bergdal ^1,2, V. Harjutsalo^1,3, P.-H. Groop^1,4, S. Mutter^1,2, the FinnDiane Study Group;

¹Folkhälsan Research Center, Helsinki, Finland, ²Department of Nephrology, University of Helsinki and Helsinki University Hospital, Helsinki, Finland, ³Research Program for Clinical and Molecular Metabolism, Faculty of Medicine, University of Helsinki, Helsinki, Finland, ⁴Department of Diabetes, Central Clinical School, Monash University, Melbourne, Australia.

Background and aims: Hyperglycaemia and dyslipidaemia are well-known risk factors for coronary artery disease (CAD) in type 1 diabetes. The impact of long-term cumulative exposure to these risk factors is less explored. We investigated the relationship between cumulative glycaemic and lipid exposure and CAD in adults with type 1 diabetes. Further, the relationship between cumulative glycaemic exposure and metabolic memory was explored.

Materials and methods: This longitudinal study included 3,495 adults with type 1 diabetes from the FinnDiane cohort. Adults with end-stage kidney disease and history of CAD or stroke before the study baseline were excluded. Total cumulative glycaemic exposure (CGE_tot), cumulative hyperglycaemic exposure (CGE_hg) (accounting only for time spent above an HbA_1c of 53 mmol/mol (7%)), cumulative exposure to LDL cholesterol, triglycerides, and non-HDL cholesterol were calculated from diabetes diagnosis until the first-ever CAD event, death, or censoring at the end of year 2020, whichever occurred first. Logistic regression models were used to assess the association between cumulative measures and CAD. In a subcohort (N = 1,121) with data on HbA_1c prior to baseline, Cox proportional hazards models were used to investigate metabolic memory by examining how cumulative glycaemic exposure from diabetes diagnosis to the baseline of the study affected the incidence of CAD.

Results: During a median follow-up of 19.38 years, 534 participants had their first-ever CAD event. At baseline, individuals with incident CAD were older (44.74 ±10.83 vs 35.47 ±10.91, p <0.001), had a longer diabetes duration (28.65 ±10.96 vs 19.02 ±11.14, p <0.001), and higher HbA_1c (72.69 ±15.43 vs 67.76 ±16.20, p <0.001). CGE_hg (OR 1.03 [95% CI 1.02-1.05], p <0.001) and cumulative exposure to LDL cholesterol (OR 1.09 [95% CI 1.06-1.11], p <0.001), triglycerides (OR 1.71 [95% CI 1.31-2.22], p <0.001), and non-HDL cholesterol (OR 1.11 [95% CI 1.08-1.13], p <0.001) significantly increased the odds for incident CAD. Compared to the lowest tertile of CGE_hg, the highest one was associated with a 2-fold increase in the odds (OR 2.00 [95% CI 1.48-2.70], p <0.001) for incident CAD. CGE_tot was not significantly associated with CAD after adjusting for cumulative lipid exposure. In the subcohort, before baseline CGE_hg was significantly associated with incident CAD during 10 years (HR 1.03 [95% CI 1.01-1.06], p = 0.017) and both before baseline CGE_tot and CGE_hg during the full follow-up (HR 1.05 [95% CI 1.03-1.08], p <0.001 and HR 1.03 [95% CI 1.02-1.05], p <0.001).

Conclusion: Cumulative glycaemic exposure above the treatment goal and cumulative lipid exposure independently increase the odds of CAD in type 1 diabetes. Further, prior higher cumulative glycaemic exposure predicts future CAD. These findings emphasize the importance of reaching an HbA_1c below 53 mmol/mol (7%) and minimizing lipid exposure to reduce CAD risk calling on health care professionals to not settle for suboptimal care and to provide encouragement for individuals living with type 1 diabetes to reach their care goals.

Supported by: Folkhälsan Fdn., Stockmann Fdn., Liv och Hälsa Soc., Juselius Fdn., Novo Nordisk Fdn., Gov. funding

Disclosure: R. Bergdal: None.

OP 16 Emerging risk factors and treatment options for diabetic eye disease

91

Semaglutide does not affect central retinal thickness beyond the effect explained by glycaemic change: continuous glucose monitoring in persons with type 2 diabetes

S. Gullaksen ¹, L. Vernstrøm Hald¹, S. Skovgaard Sørensen¹, K. Funck², L. Petersen³, T. Bek³, E. Laugesen⁴, P. Poulsen¹;

¹Aarhus University Hospital, Steno Diabetes Center, Aarhus, Denmark, ²Aarhus University Hospital, Aarhus N, Denmark, ³Dept. of Ophthalmology, Aarhus University Hospital, Aarhus, Denmark, ⁴Department of Endocrinology and Internal Medicine, Aarhus University Hospital, Aarhus C, Denmark.

Background and aims: Glycaemic control is important for prevention of diabetic retinopathy (DR). However, large, and rapid improvements in glycemia is linked to worsening of DR. Semaglutide was in one study associated with worsening of DR. It remains a matter of discussion whether this was driven by rapid glycemic improvement (early worsening) or could be an inherent effect of semaglutide. In a randomized trial, we previously observed a small but significant increase in Central Retinal Thickness (CTR), an early marker of DR, in semaglutide treated diabetes participants (3.8 μm, 95 % CI [0.9;6.7] μm, p = 0.009). We hypothesize the association might be driven by glycaemic change. In this sub-study we assessed if Time in Range (TIR (3.9-10.0 mmol/L)) assessed by continuous glucose monitoring (CGM) is linked to CTR.

Materials and methods: This study was a post-hoc analysis of a 32 weeks randomised clinical trial investigating the separate and combined effects of semaglutide (glucagon-like peptide 1 receptor agonist) and empagliflozin (sodium-glucose co-transporter 2 inhibitor) on target organ damage. Participants were randomised into four groups (semaglutide, empagliflozin, the combination, or placebo). The total study population included 120 persons with type 2 diabetes, age ≥ 50 years. Time in range was assessed using CGM during 7-8 days in a predefined random subgroup of 40 participants (10 in each arm).

Results: 36 participants (90 %) were men. The mean age was 71(±7) years and the median HbA1c was 58 mmol/mol(51-65). 62 eyes had no retinopathy (grade 0), 10 eyes had mild non-proliferative retinopathy (NPDR) (grade 1), 2 eyes had moderate NPDR (grade 2) and 2 eyes had proliferative retinopathy (grade 4). 4 eyes were missing for technical reasons. We found no association between TIR (%) and CRT (μm) at baseline (p=0.247). During 32 weeks of treatment, increased TIR was associated with increased CRT (0.07 μm, 95%CI[0.03; 0.12] μm, p=0.002). In the 10 semaglutide treated participants CRT increased to the same extent as in the main study (3.8μm, 95%CI[-0.32; 7.85] μm, p=0.07) compared to placebo. However the effect was attenuated with adjustment for TIR (2.59 μm, 95%CI[-1.50; 6.60] μm, p=0.21).

Conclusion: During 32 weeks of treatment, increased time in range as measured by CGM was associated with increased CRT. Semaglutide treatment did not affect central retinal thickness beyond the effect explained by glycemic change. This could support the link between tight glycaemic control and worsening of DR.

Clinical Trial Registration Number: EUDRACT 2019-000781-38

Supported by: NNF, CDR, DMAR

Disclosure: S. Gullaksen: Grants; Danish Medical Association.

92

Is early worsening of diabetic retinopathy a concern when starting hybrid closed loop therapy in people with type 1 diabetes?

S.A.D. Mathara Diddhenipothage ¹, R. Blackhurst², E.M. Dutton², H. Smith², K. Hards¹, H. Lipinski³, M. Matheou¹, S. Fatum⁴, P.H. Scanlon⁵, A.N. Lumb¹, G.D. Tan¹;

¹Oxford Centre for Diabetes, Endocrinology and Metabolism, Oxford, UK, ²School of Medicine and Biomedical Sciences, Oxford, UK, ³Oxford Diabetic Eye Screening Programme, Oxford, UK, ⁴John Radcliffe Hospital, Oxford, UK, ⁵Gloucestershire Retinal Research Group, Gloucestershire Hospitals NHS Foundation Trust, Cheltenham, UK.

Background and aims: Rapid tightening of glycaemic control is associated with an increased risk of early worsening of diabetic retinopathy (EWDR). Initiation of HCL can lead to a rapid decrease in HbA1c and increase in time in range (TIR), but whether this is associated with EWDR is unclear.

Materials and methods: Retrospective audit of everyone with T1D started on HCL therapy in a UK tertiary care diabetes centre between Dec 2019 and June 2023.

Results: 74 people were included [64% female, mean age was 43±16y (mean ±SD), duration of T1D was 27±14.7y]. Before starting HCL, most (83.7%) people were treated with insulin pump therapy and were on either continuous or flash glucose monitoring (91.8%). 57 people (77%) were started on Tandem t:slim X2™ with Control-IQ Technology; 16 (21.6%), MiniMed™ 780G Advanced HCL System; 1 (1.3%), CamAPS FX system. Prior to HCL initiation, HbA1c was 62±18mmol/mol and TIR 56±20%. 15% of people had hypertension and 46% were treated with a statin. Before starting HCL, diabetic retinopathy (DR) assessment showed background retinopathy (R1) in 23 (31%) people; pre-proliferative retinopathy (R2) in 6 (8%); stable proliferative retinopathy (R3S) in 7 (9.4%); active proliferative retinopathy (R3A) in 2 (2.7%); and maculopathy (M1) in 13 (17.6%). Data was missing in 14 (20%) of people. HCL initiation was associated with a decrease in HbA1c of 10mmol/mol and an increase in TIR of 18.7%. 23 (31.1%) of people starting a HCL experienced an EWDR; 8 (10.8%) showed an improvement; and 22 (30%) were missing DR data. Of note, 42.2% of low-grade DR (R0/R1) and 26.7% of high-grade DR (R2/R3) at baseline, had EWDR. Baseline characteristics of people whose DR progressed were not different to those whose DR did not progress, in terms of, age, sex, baseline glycaemic control (HbA1c or TIR), DR grade, duration of diabetes or use of a statin. The magnitude of the change in HbA1c or TIR was not different in the group in whom DR progressed compared to the group in whom DR did nor progress. Similarly, there was no difference in the rate of EWDR between the different HCL systems. 8 (10.8%) people received specific treatment for DR after starting HCL (all had laser therapy, and 4 received additional anti-VEGF): 2 newly commenced on treatment, whilst 5 continued to require active treatment that was commenced pre-HCL.

Conclusion: A third of people with T1D starting HCL experienced EWDR, and up to 10.8% required specific therapy for DR. In this small cohort, there were no features identified to be specifically associated with EWDR.

Disclosure: S.A.D. Mathara Diddhenipothage: None.

93

Diabetes-related microvascular complications and periodontitis: Health in Central Denmark

F.V. Bitencourt ^1,2, A. Andersen², L. Bjerg², A. Sandbæk², H. Li³, G.G. Nascimento³, R. Spin-Neto¹, M.A. Peres³, F.R.M. Leite⁴;

¹Dentistry and Oral Health, Aarhus University, Aarhus, Denmark, ²Steno Diabetes Center Aarhus, Aarhus, Denmark, ³National Dental Research Institute Singapore, National Dental Centre Singapore, Singapore, Singapore, ⁴National Dental Centre Singapore, Singapore, Singapore.

Background and aims: Inconsistent findings indicate a potential relationship between diabetes-related microvascular complications and periodontitis. Small sample sizes and lack of adjustment to confounding variables have limited reliable estimates from previous studies. Furthermore, the joint impact of microvascular complications and dyslipidemia on periodontitis has not been examined. Thus, this study evaluated the link between individual and combined diabetic microvascular complications (i.e., neuropathy and retinopathy) and the severity of periodontitis in a Danish population-based study. We also explored the influence of dyslipidemia on these relationships.

Materials and methods: This study included 15,922 individuals with type 2 diabetes from the Health in Central Denmark study. Multinomial logistic regression was employed to determine the odds ratios (ORs) and 95% confidence intervals (CIs) for both individual and combined complications related to microvascular diabetes. Adjustments were made in the models for potential confounding variables, which included sociodemographic status, lifestyle habits, and health conditions. To equalize the measured confounding factors among participants with and without periodontitis, inverse probability of treatment weighting (IPTW) was applied. The robustness of the results was verified through sensitivity analyses, which involved calculating E-values for unmeasured confounders and modifying the definitions of microvascular complications.

Results: Post-IPTW adjustments, the models demonstrated an association between diabetic neuropathy (OR 1.36, 95% CI 1.14 - 1.63) and retinopathy (OR 1.21, 95% CI 1.03 - 1.43) and ’moderate/severe’ periodontitis. The presence of concurrent microvascular complications was linked to a 1.5 times increase in the likelihood of ’moderate/severe’ periodontitis (OR 1.51, 95% CI 1.23 - 1.85). We observed an additive scale effect modification by dyslipidemia on periodontitis, with relative excess risks of interaction between 0.11 and 0.44. The sensitivity analysis confirmed that neither unmeasured confounders nor definitions of microvascular complications influenced the results.

Conclusion: Diabetic neuropathy and retinopathy, independently and concomitantly, were associated with the severity of periodontitis. Furthermore, dyslipidemia exhibited an additional positive effect modification of microvascular complications on ’moderate/severe’ periodontitis. These results could help to identify which subgroups are more susceptible to diabetes-associated microvascular complications and periodontitis, enhancing strategies to reduce the impact of these conditions.

Disclosure: F.V. Bitencourt: None.

94

Impact of sleep duration at night on microvascular complications in patients recently diagnosed with type 2 diabetes

M.S. Johansen ¹, J.V. Stidsen¹, P. Asyaei¹, F.N. Pedersen^2,3, J. Grauslund^1,2, M.H. Olsen^4,5, K. Højlund^1,3, T.B. Olesen¹;

¹Steno Diabetes Center Odense, Odense University Hospital, Odense, Denmark, ²Department of Ophthalmology, Odense University Hospital, Odense, Denmark, ³Department of Clinical Research, University of Southern Denmark, Odense, Denmark, ⁴Department of Medicine and Steno Diabetes Centre Zealand, Holbaek Hospital, Holbaek, Denmark, ⁵Department of Clinical Medicine, University of Copenhagen, Copenhagen, Denmark.

Background and aims: Microvascular complications, such as retinopathy and nephropathy, are major contributors to morbidities associated with Type 2 Diabetes Mellitus (T2DM). Increasing evidence suggests that variations in sleep duration may influence the risk of developing diabetes-related complications. This study aims to explore the relationship between sleep duration and the presence of microvascular disease in individuals newly diagnosed with T2DM.

Materials and methods: This cross-sectional analysis uses data from The Specialist Supervised Individualized Multifactorial Treatment of New Clinically Diagnosed Type 2 Diabetes in General Practice (IDA) cohort. Sleep duration at night was measured using Axivity AX3 accelerometers, worn by participants for a period of 10 days. Sleep duration at night was classified into three categories: short (<7 h), optimal (7 to <9 h), and long (≥9 h) according to the most recent consensus statement. Microvascular disease was defined as either urine albumin/creatinine ratio (UACR) ≥ 30 mg/g or presence of diabetic retinopathy (DR) assessed by either mydriatic retinal imaging or ophthalmoscopy. Logistic regression analysis using optimal sleep duration as the reference group adjusted for age, sex, BMI, systolic blood pressure, smoking habits, HbA1c, duration of diabetes, and antihypertensive treatment was used to estimate odds ratios (ORs).

Results: In total, 396 participants had valid sleep duration measurements, UACR measurement and eye examination. The median age was 62 years (IQR: 53-68) with a mean diabetes duration of 3.5 ± 2.7 years, and 175 were females (44%), The cohort predominantly consisted of overweight individuals, with a median BMI of 31 (IQR: 28-35), and 68% (n=285) were on antihypertensive medication. Distribution of sleep duration was 12% (n=49) with short sleep duration, 60% (n=238) with optimal sleep duration, and 28% (n=109) with long sleep duration. Prevalence of microvascular damage was 38%, 18%, and 31% in the short, optimal, and long sleep duration groups, respectively. Short sleep duration was significantly associated with microvascular disease; OR 2.63 (95% CI: 1.32-5.26). Similarly, long sleep duration was independently associated with an increased risk of microvascular disease; OR 2.29 (95% CI: 1.32-3.97). Furthermore, the association between short sleep duration and microvascular disease was accentuated by age. Interestingly, for participants <62 years OR was 1.23 (95% CI: 0.43-3.55), whereas for those >62 years OR was 5.67 (95% CI: 2.10-15.27, P for interaction = 0.044).

Conclusion: In newly diagnosed T2DM patients, both short and long sleep durations are associated with a higher prevalence of microvascular disease compared to optimal sleep duration at night. Age amplifies the association between short sleep duration and microvascular disease, suggesting increased vulnerability among older individuals.

Supported by: Danish Agency for Science (09-067009, 09-075724), the Region of Southern Denmark, the Novo Nordisk Foundation and the University of Southern Denmark

Disclosure: M.S. Johansen: None.

95

Pericyte MDM2-knockout prevents retinal fibrosis in the development of diabetic retinopathy via regulating ITGB8-TGF-β1 axis

J. Lin, Y. Chen, Z. Xu, B. Krämer;

5th Medical Department, Medical Faculty Mannheim, Mannheim, Germany.

Background and aims: Retinal fibrosis is a critical pathological process in the development of diabetic retinopathy (DR). Pericytes, as myofibroblast precursors, play important role in vascular stabilization and tissue fibrosis. . Mouse double minute 2 homolog (MDM2) is an E3 ubiquitin ligase and involved in many cellular processes including tissue fibrosis. Integrin β8 (ITGB8) activates TGF-β signaling, which is the master regulator of fibrosis, and subsequently triggers the epithelial to mesenchymal transition (EMT) processes. However, the molecular mechanism of diabetic retinal fibrosis and signaling pathways remain unknown. This study aims to investigate the role of pericyte MDM2 in retinal fibrosis of diabetic retinopathy.

Materials and methods: To investigate the influences of MDM2 in the development of DR, we generated pericyte specific MDM2 knockout (MDM2-KO) mice and induced diabetes by streptozotocin. Animals were divided into four groups: wildtype (WT-NC), MDM2-KO control (KO-NC), WT-diabetic control (WT-DC), and MDM2-KO diabetes (KO-DC). After six months diabetes, vascular damage was evaluated by retinal morphometry. Neuroretinal function was assessed by ERG. In vitro experiments, human retinal pericytes were cultured in high glucose condition and transfected with control or MDM2 short interfering RNA (siRNA). Immunocytochemistry (ICC) was used to detect protein expression. The regulation of TGFβ-signaling (SMAD3) and fibrosis related proteins (collagens and cadherins) were detected with immunohistochemistry. Gene expression was evaluated by quantitative PCR.

Results: Pericyte MDM2-KO avoided the vascular damage in diabetes by reducing acellular capillary formation and pericyte loss. Pericyte MDM2-KO prevented apoptotic damage from hyperglycemia, but not protected the neuroretinal functions. Moreover, MDM2-KO in pericytes attenuated hyperglycemia-induced retinal fibrotic injury with reduced deposition of extracellular matrix protein and EMT. In diabetic retinas, we observed that the expression of ITGB8 was reversely regulated by MDM2. In vitro data confirmed that high glucose (HG)-induced accumulation of MDM2 led to the reduction of ITGB8 in retinal pericytes and promoted the activation of SMAD3. On the contrary, MDM2-knockdown resulted in ITGB8 recovery and reduction of TGF-β downstream signaling.

Conclusion: PPericyte conditional MDM2-KO prevents the development of diabetic retinopathy (DR). Targeting MDM2-ITGB8-TGFβ1 axis may be a novel therapeutic strategy for the treatment of DR, especial early phase of retinal fibrosis.

Supported by: German Research Foundation

Disclosure: J. Lin: None.

96

Topical administration of sitagliptin prevents neurodegeneration in an experimental model of steroid-induced glaucoma: clinical implications for treating diabetic macular oedema

R. Simo ¹, P. Bogdanov¹, A. Duarri², M.J. Canz¹, D. Sabater¹, H. Isla-Magrané², H. Ramos¹, A. Deàs-Just¹, C. Hernández¹;

¹Diabetes and Metabolism Research Unit. CIBERDEM, Vall d’Hebron Research Institute, Barcelona, Spain, ²Ophthalmology Research Group, Vall d’Hebron Research Institute, Barcelona, Spain.

Background and aims: Steroid-induced glaucoma is an important clinical problem related to the use of intravitreal glucocorticoid therapies for treating diabetic macular edema. Currently, this adverse effect is treated with conventional drugs addressed to reduce the intraocular pressure (IOP). Glaucoma is a neurodegenerative disease that leads to progressive dysfunction and loss of retinal ganglion cells (RGCs), but a substantial proportion of patients (25-50%) have a normal IOP level. On this basis we wanted to examine whether sitagliptin (a DPP-IV inhibitor) administered in eye drops, which was useful in diabetes-induced retinal neurodegeneration, could also be useful in preventing steroid-induced glaucoma.

Materials and methods: We used a mouse model of primary open-angle glaucoma induced by periocular administration of dexamethasone phosphate (DEX) once a week for 5 weeks. Eye drops of sitagliptin (10 mg/mL) vs. vehicle were administered from day 14 to day 35. Untreated mice were used as control group. After euthanasia, retinal sections and wholemount, transverse sections of optical nerve head (ONH) were performed. Measurements in vivo: IOP and fundoscopy. Measurements post-mortem: confocal microscopy immunofluorescence [IF] (GFAP, Iba-1, RBPMS, TUJ1, neurofilament (NFH), γ-synuclein, and galectin-3 and Olig2).

Results: We observed a significant loss of both RBPMS and TUJ1 in the neuroretina of DEX-group compared to controls (p<0.05), thus indicating the decrease of cell bodies and axon bundles of RGC in the DEX-vehicle group. This effect was prevented by topical administration of sitagliptin. In addition, a significant decrease in RGC axons marked by neurofilament was found in mice treated with DEX in comparison with control group (p<0.001), thus indicating reduced axonal integrity. Treatment with sitagliptin eyedrops prevented this deleterious effect. In addition, topical administration of sitagliptin significantly prevented the overexpression of GFAP and Iba-1 in the ONH observed in DEX-vehicle group, thus ameliorating the macro and microglial activation (p<0.05). Sitagliptin also inhibited the overexpression of galectin-3 (MAC-2) and gamma-synuclein (γ-Synuclein) in the ONH (p<0.001), two important mediators of inflammation and apoptosis, respectively. Finally, eyedrops of sitagliptin prevented the oligodendrocyte loss induced by DEX.

Conclusion: Topical administration of sitagliptin exerts a powerful neuroprotective action in glaucoma induced by dexamethasone. Our results pave the way for a new treatment of glaucoma based in neuroprotection, which could be particularly useful in those patients with diabetic macular edema who develop glaucoma after intravitreal injections of corticosteroids.

Supported by: Grant from the Spanish Ministerio de Ciencia e Innovación (PDC2021-121638-I00)

Disclosure: R. Simo: Other; RS and CH are inventors of the patent PCT/EP2017/060234.

OP 17 Type 2 diabetes prediction

97

Screening for type 2 diabetes and population mortality over 20-year follow-up (ADDITION-Cambridge): a cluster-randomised controlled trial

T.F.A. Marshall-Andon ^1,2, C.E. Boothby¹, J.B. Echouffo Tcheugui^1,3, L.A. Sargeant¹, K.M. Williams², T. Prevost^2,4, A.-L. Kinmonth², N.J. Wareham¹, S.J. Griffin^1,2;

¹MRC Epidemiology Unit, University of Cambridge, Cambridge, UK, ²Primary Care Unit, Department of Public Health and Primary Care, University of Cambridge, Cambridge, UK, ³Division of Endocrinology, Diabetes and Metabolism, Johns Hopkins School of Medicine, Baltimore, MD, USA, ⁴King’s College London, London, UK.

Background and aims: Type 2 Diabetes (T2D) is associated with premature mortality, but many patients remain undiagnosed. Detection of diabetes via screening may permit earlier management of hyperglycaemia and related cardiovascular risk factors and improve outcomes. Furthermore, screening may provide an opportunity for cardiovascular risk assessment and management which could impact population cardiovascular mortality. However, there remains uncertainty concerning the overall benefits and harms of screening. ADDITION-Cambridge, a cluster randomised controlled trial, explored the effect of a single round of systematic screening for T2D on all-cause, grouped-cause, and diabetes-related mortality, in high-risk individuals, to provide insights into the overall benefits and harms of a diabetes screening programme, at the population level. This work assesses mortality over 20 years of follow-up.

Materials and methods: General practices were randomised to control (n=5) and screening (n=28) arms by minimisation. However, one screening practice withdrew from the study. Diabetes risk scores were calculated for individuals aged 40-69 in both study arms using routine primary care data. Individuals in screening practices with a risk score in the top 25% of the population distribution were invited for stepwise diabetes screening between 2002 and 2006. An intention-to-treat analysis including high-risk individuals in both arms was conducted using a Cox proportional hazards model to assess all-cause mortality and competing risk regression to assess grouped-cause and diabetes-related mortality. Sensitivity analyses looking at pre-existing disease, missing covariate data, and screening attendance were conducted.

Results: 4,137 and 16,558 high-risk individuals were identified in control and screening practices. Models were adjusted for age, sex, BMI, and prescription of anti-hypertensives and steroids. Over 20 years of follow-up, the offer of screening was not associated with significantly lower all-cause mortality (HR 1.01, 95% CI 0.92-1.11) or diabetes-related mortality (HR 0.89, 95% CI 0.69-1.15). The offer of screening was associated with higher accident/suicide/violent mortality (HR 1.55, 95% CI 1.05-2.27, p=0.03), but was only associated with significantly lower cardiovascular mortality (HR 0.87, 95% CI 0.79-0.97, p=0.01) when a parallel cohort of 22 additional screening practices was included. Screening attendance was associated with significantly lower all-cause mortality (HR 0.69, 95% CI 0.64-0.74, p<0.001) and cardiovascular mortality (HR 0.69, 95% CI 0.62-0.77, p<0.001).

Conclusion: Systematic screening likely benefits those diagnosed with T2D but, in common with most screening programmes, does not influence mortality at the population level. However, screening attendance is significantly associated with personal characteristics, and may be lower in those at greatest risk of premature mortality. Attention to access and uptake in underserved populations must be ensured to mitigate increases in inequalities associated with screening.

Clinical Trial Registration Number: ISRCTN86769081

Supported by: Wellcome Trust, Medical Research Council, NHS R&D support, the National Institute for Health Research, the University of Aarhus (Denmark), and Bio-Rad

Disclosure: T.F.A. Marshall-Andon: Employment/Consultancy; ALK was a National Institute for Health Research (NIHR) Senior Investigator. Grants; JBET was funded by a scholarship from the Gates Cambridge Trust, SJG received support from the Department of Health NIHR Programme Grant funding scheme (RP-PG-0606-1259), The Primary Care Unit is supported by NIHR funds, TP was supported by the NIHR Biomedical Research Centre based at Guy’s and St Thomas’ NHS Foundation Trust and King’s College London, ADDITION-Cambridge was funded by the NIHR under the Programme Grants for Applied Research programme (RP-PG-0606-1259). Honorarium; SJG has received honoraria from AstraZeneca and Lilly for contributions to postgraduate education meetings. Non-financial support; Bio-Rad provided equipment to undertake capillary glucose screening by HbA1c in general practice.

98

A multi-country study of trends in the incidence of end-stage kidney disease among people with diabetes

J.W. Sacre ¹, J.I. Morton^1,2, L. Chen¹, M.E. Pavkov³, E.W. Gregg⁴, J.E. Shaw¹, D.J. Magliano^1,2, GLOBODIAB Writing Group;

¹Baker Heart and Diabetes Institute, Melbourne, Australia, ²Monash University, Melbourne, Australia, ³Centers for Disease Control and Prevention, Atlanta, GA, USA, ⁴Imperial College, London, UK.

Background and aims: Escalating worldwide prevalence of end-stage kidney disease (ESKD) is well-known, but the incidence of ESKD among people with diabetes is a more relevant metric for monitoring diabetes-related ESKD risk. This multi-country study examined trends over time in the incidence of ESKD treatment.

Materials and methods: Nine high-income jurisdictions contributed registry / administrative diabetes population data linked to the outcome of treated ESKD (initiation of kidney replacement therapy [KRT]) for different periods spanning the early 2000s to 2021. Incidence rates were derived from age-period-cohort Poisson models. Average annual percent changes in age-standardised rates were determined overall, and by sex.

Results: Incident cases of KRT totalled 288,668 during ~98 million person-years of follow-up. Age-standardised incidence rates (Figure part A) showed geographic variation throughout the study period, but decreased over time in six of the nine jurisdictions (Figure part B). The exceptions (Australia, France and Italy) were among the jurisdictions with the lowest KRT incidence at their respective study start dates and showed relatively unchanged rates over time. Trends were similar for men and women.

Conclusion: Incidence of KRT among people with diabetes has been declining in most participating jurisdictions. Continued surveillance may be important with the expected emerging impact of newer drugs (e.g. sodium-glucose cotransporter 2 inhibitors) at the population level.

Supported by: US Centers for Disease Control and Prevention

Disclosure: J.W. Sacre: None.

99

How frequently is early-onset type 2 diabetes misclassified in South Asian and White individuals?

S.A.W. Jones, H. Cheng, S. Pandya, C. Goodall, S. Misra;

Department of Metabolism, Digestion & Reproduction, Imperial College London, London, UK.

Background and aims: Accurate diabetes subtype classification is crucial for optimal treatment choice but can be challenging in young adults. Differences in prevalence of early-onset type 2 diabetes across ancestry groups further challenge classification. Use of biomarkers and genetic testing may help but are not uniformly utilised. We compared the accuracy of clinician diagnosis of diabetes subtype in a cross-sectional UK-based population of young-onset (<30 years at diagnosis) diabetes in South Asian (who are more likely to have early-onset type 2 diabetes) versus White ethnicities when assessed against unbiased, biomarker-defined diagnostic criteria.

Materials and methods: Individuals diagnosed with any type of diabetes before the age of 30-years were recruited. Participants were categorised as White or South Asian if the ethnic origin of all four grandparents matched. Clinical and anthropometric data were collected, and the diabetes type according to the clinical team was recorded. All individuals with a detectable C-peptide >27pmol/L AND 3 negative diabetes auto-antibodies underwent targeted next generation sequencing to identify MODY. Using biomarkers, Type 2 diabetes was defined by absence of 3 pancreatic antibodies AND a fasting C-peptide >200pmol/L or random C-peptide >600 pmol/L; Type 1 diabetes was defined by the presence of 1 or more auto-antibodies OR C-peptide <200 pmol/L; monogenic diabetes was defined by the presence of confirmed pathogenic mutation following targeted next generation sequencing. An indeterminate group was also included in which C-peptide was 200-599 pmol/L, 3 antibodies were negative AND no pathogenic mutations were detected.

Results: 653 White European (mean age 38.5 years, mean diabetes duration 22 years) and 618 South Asian individuals with early-onset diabetes (mean age 30.6 years, mean diabetes duration 14 years) were studied. Amongst individuals initially clinically diagnosed with Type 2 Diabetes (n=208 South Asian, n=43 White), 18% (n=37) of South Asian and 21% (n=9) of White (p=0.62) individuals were reclassified as Type 1 diabetes using biomarker criteria. White individuals were more likely to be reclassified with monogenic diabetes (12%, vs. 1.4% South Asians, p<0.001).12% of South Asian and 7% of white individuals had indeterminate diabetes subtype with C-peptide 200-600 pmol/L but no antibodies positive (p=0.38). Overall of those initially labelled with type 2 diabetes under 30 years 19% (n=40) of South Asian and 33% of White individuals were reclassified as having type 1 diabetes or monogenic diabetes (p=0.17). Applying the biomarker criteria to everyone, of those classified with Type 2 Diabetes (n=47 White, n=169 South Asian), White individuals were more likely to have been clinically diagnosed with Type 1 diabetes (27.7% vs. 13.6% in South Asians, p=0.046), whilst South Asian individuals were more likely to have been clinically diagnosed with Type 2 Diabetes (72% vs. 51% White Europeans, p=0.006).

Conclusion: When using an unbiased, standardised biomarker approach to classification ~1 in 3 White individuals and 1 in 5 South Asian individuals labelled with ‘type 2 diabetes’ under 30 years have an alternative subtype risking incorrect treatment. Early-onset diabetes should be systematically classified in all adults before a label of type 2 diabetes is given, as phenotypes can overlap considerably with type 1 diabetes and monogenic diabetes in both South Asian and White individuals.

Supported by: EFSD Future Leaders Mentorship Programme for Clinical Diabetologists supported by AstraZeneca

EFSD/Novo Nordisk Foundation Precision Diabetes Medicine Award Programme

Wellcome Trust Clinical Career Development Award to SM (223024/Z/21/Z), DRWF Sutherland Earl Fellowship

Disclosure: S.A.W. Jones: None.

100

Diabetes medication is not associated with cancer occurrence in type 2 diabetes patients: a study of the entire Danish population 1996-2022

B. Carstensen ¹, S. Friis², F. Persson¹, V. Kosjerina¹, M. Gamborg², L.S. Mørch²;

¹Steno Diabetes Center Copenhagen, Herlev, Denmark, ²Danish Cancer Institute, Copenhagen, Denmark.

Background and aims: During the last 15 years, the diabetes drug classes DPP-4i, GLP-1RA and SGLT-2i have been marketed, but the real-world evidence on long-term cancer safety of these drugs is still limited. With population data available to 2022, we had potential exposure for 15, 13 and 9 years for these three drug classes, so we aimed to quantify to which extent the long-term cancer occurrence was associated with use of diabetes medications.

Materials and methods: The entire Danish population was followed from 1996 through 2022 incl., using population registers of diabetes and cancer occurrence and of prescription fillings. Cancer events and person-years were divided in 1-year intervals, and in each interval diabetes status (noDM, T1DM, T2DM) and the exposure status to each of the drugs (metformin, SU, DPP-4i, GLP-1RA, SGLT-2i, insulin) was classified as ever exposed and time since first exposure. Models for cancer incidence rates were fitted using Poisson likelihood separately for each sex, comparing T2DM with noDM. For all cancer and each of 25 types of cancer, incidence rates were modeled by age, calendar time, date of birth (APC-model), time since start of each drug class and diabetes duration, all using natural splines. Rate-ratios relative to persons without drugs but with diabetes were reported, as well as relative to persons without diabetes.

Results: The number of cancers / person-years were (75,700 / 3.7 mil.) in T2DM, and (811,644 / 137 mil.) in noDM. For T2DM the RR between persons on drugs and persons not on drugs were between 0.9 and 1.1 for both women and men. Compared to persons without diabetes the RRs were between 1.0 and 1.2, all with negligible confidence intervals. Including duration of T2DM and time since start of each drug class showed RRs elevated at the start of diabetes and to a lesser extent at start of drug exposure. At start of diabetes, persons on metformin, SU or SGLT-2i had a RR of 2, on DPP-4i and GLP-1RA of 1.5 and on insulin of 3.0. Persons starting SU at duration of diabetes of 1, 3 and 5 years showed a RR of 1.5 at start, for insulin a RR of 2; and for the remaining drug classes only small RRs were seen. For all drugs, the RRs with increasing duration of diabetes and time since start of drug exposure waned during the first 2 years. There were 5,980 breast cancers in T2DM, and 75,722 in noDM; the overall RR associated with insulin was 0.88 (95% CI.: 0.83-0.94), none of the other drug classes showed any association, neither overall nor by time since start of drug exposure. There were 8,000 prostate cancers in T2DM, and 99,915 in noDM; SGLT-2i had an overall RR of 0.89 (0.82-0.98) with a decrease by increasing time since initiation. Insulin had an RR of 0.83 (0.78-0.88), stable after 4 years since initiation. None of the other drug classes showed any association with prostate cancer, neither overall nor by duration, except DPP-4i that showed an increase with increasing time since initiation to RR of 1.1 at 10 years. Relative to the non-diabetes population we saw the well-known overall RR of 0.9.

Conclusion: In this study based on the entire Danish population in 27 years, there is a clear indication that occurrence of cancer is not elevated long-term after initiation of any of the major diabetes drugs, relative to the non-exposed population. For all cancers the rates deviate less than 5% between persons with and without diabetes, regardless of type of drug class initiated.

Disclosure: B. Carstensen: Stock/Shareholding; Novo Nordisk.

101

Remnant cholesterol and incident type 2 diabetes: a community-based 14-year prospective cohort study

Y.-C. Hwang, J. Jun, I.-K. Jeong, K. Ahn, H. Chung;

Kyung Hee University Hospital at Gangdong, Seoul, Republic of Korea.

Background and aims: Recent epidemiologic studies have shown the association between remnant cholesterol level and future cardiovascular events. Apart from as a risk factor for cardiovascular disease, lipid measurements are predictive of incident diabetes.

Materials and methods: A total of 7,702 people aged 40 to 69 years free of diabetes at baseline were recruited from the Korean Genome and Epidemiologic Study. Remnant cholesterol was calculated and LDL cholesterol level was estimated by using Sampson-National Institutes of Health (NIH)2 equation. Oral glucose tolerance test was performed at every visit.

Results: During 14 years of follow-up, 22.0% (1,694/7,702) of people developed incident diabetes. In Cox proportional hazard model, remnant cholesterol was associated with an increased risk of diabetes independently of conventional risk factors for diabetes (hazard ratio per 1-SD increment (95% confidence interval): 1.15 (1.20 - 1.30); P<0.001). In subgroup analysis, remnant cholesterol was predictive of incident diabetes regardless of age groups, sex, body mass index, glycemic status, and presence of metabolic syndrome at baseline; however, it was not in people with high baseline triglyceride level. In comparison with other lipid measurements, the performance of remnant cholesterol in predicting incident diabetes outweighs those of total cholesterol, LDL cholesterol, and non-HDL cholesterol; however, it was comparable to that of triglyceride.

Conclusion: Remnant cholesterol level was an independent predictor of incident type 2 diabetes. In addition to stratifying cardiovascular risk, remnant cholesterol might be used to stratify future diabetes risk.

Disclosure: Y. Hwang: None.

102

Screening for type 2 diabetes is cost-effective: evidence from a pilot screening study in Denmark

N. C. Mateu ¹, P. Rossing², K. Neergaard², T. Thybo¹;

¹Danish Diabetes Association, Glostrup, Denmark, ²Steno Diabetes Center Copenhagen, Herlev, Denmark.

Background and aims: Studies on the cost-effectiveness of screening for type 2 diabetes mellitus (T2DM) are scarce, inconclusive, or obsolete. HbA1c as a diagnostic criterion, novel methods to collect and store blood samples, as well as the quality of available national digital registers allowing for targeted screening may substantially have decreased the screening cost per screen detected. The aim of the study is to evaluate whether targeted screening for T2DM using at-home self-sampling HbA1c tests is cost-effective in Denmark.

Materials and methods: From October 2022 to June 2023, the Danish Diabetes Association mailed free capillary at-home self-sampling HbA1c tests to 8,000 randomly selected individuals aged 50 to 75 years, who had not had their HbA1c measured within the past two years. The Clinical Laboratory Information Register, including laboratory measurements covering hospitals and GPs, was used to identify the population.

Results: 3,040 individuals (38%) returned a blood sample. This is in line with other Danish screening studies. Participants with HbA1c of 42-47 mmol/mol constitutes 11%, while participants with HbA1c ≥48 mmol/mol constitutes 1.7%. The age-adj. shares are consistent with the estimated age specific shares of individuals with unknown T2DM and prediabetes in Danmark. The screening costs was 1200 € in list prices per screen detected with T2DM. Of the screen detected, 49% had a HbA1c (mmol/mol) of [48;53[, 24% of [53;63[, and 27% of ≥63. This is similar to the HbA1c distribution of individuals with T2DM diagnosed through the Danish Health Care System (DHCS). However, individuals between 50-59 years of age constituted 51% of the participants with HbA1c ≥48 mmol/mol compared to 40%, when diagnosed through the DHCS. It indicates, that the screen detected were diagnosed earlier than through the DHCS. A recent Danish study finds that a three-year delay in diagnosis and treatment of T2DM patients entails direct and indirect costs of 259 €, 591 €, and 1061€ per patient with a HbA1c of [48;53[ mmol/mol, [53;63[ mmol/mol, and ≥63, respectively, while prior evidence suggest that screening move diagnosis forward by 3 to 6. A targeted screening program, moving diagnosis forward by 3 years across all HbA1c-levels, results in average cost savings of 1911€ given the distribution of screen detected across HbA1c as detected in the screening study. Hence, the total social return ratio of the screening program is 1911 €/ 1200 €= 1.59.

Conclusion: The study suggests that targeted screening for T2DM using at-home self-sampling HbA1c tests in Denmark is cost-effective. The study does not include non-pecuniary effects of early diagnosis and treatment, such as an increased quality of life for the patient and their relatives. The study further excludes the potential cost-savings from preventing T2DM among individuals who are identified with prediabetes. Existing evidence suggests that cost-savings are substantial. If the benefits of detecting individuals with prediabetes are included, the screening costs constitute 173 € per screen detected with prediabetes or T2DM. Hence, the social return ratio of screening for T2DM may be higher than 1.59. Finally, it is noted that due to risk for harvesting effects, a screening program should be carried out annually at five-year intervals, where citizens are offered screening in the years, they turn 50, 55, 60, 65 and 70 years old.

Disclosure: N. C. Mateu: None.

OP 18 How to blow the low

103

Real-world burden of severe hypoglycaemic events in type 1 diabetes in Germany

R. Ziegler ¹, Q. Zhang², P. Callahan², C. Vetter³, E. Garal-Pantaler⁴, A. De Remigis², L. Chen², W. Rathmann⁵;

¹Diabetes Clinic for Children and Adolescents, Muenster, Germany, ²Vertex Pharmaceuticals Incorporated, Boston, MA, USA, ³IQVIA Commercial GmbH & Co. OHG, Frankfurt, Germany, ⁴Team Gesundheit GmbH, Essen, Germany, ⁵Institute for Biometrics and Epidemiology, German Diabetes Center, Duesseldorf, Germany.

Background and aims: Severe hypoglycemic events (SHE) in adults with type 1 diabetes (T1D) are events requiring the assistance of another person for recovery and are associated with increased risk of death. Limited evidence exists on the healthcare resource utilization (HCRU) and costs of adults with T1D and SHEs living in Germany.

Materials and methods: Adults (≥18 years of age) with T1D were selected from a German BKK insurance claims database, which includes 5.7 million people with statutory health insurance, between January 1, 2013 and December 31, 2020. T1D was defined as ≥1 specific diagnostic code (ICD-10) in ≥2 quarters of a year (M2Q), a ratio of >0.5 for T1D over T2D M2Q diagnoses, ≥1 insulin prescription, and no non-insulin glucose-lowering drug prescriptions except for off-label use. For people with T1D and an inpatient SHE (HYPO+IP cohort), date of the first observed hospital visit with a primary hypoglycemia diagnosis or secondary hypoglycemia diagnosis with a concurrent Diagnosis Related Group code K60 was defined as the index date. People with T1D but no inpatient SHE (matched cohort) were matched with the HYPO+IP cohort on age, sex, index year, and geographic region in up to a 5:1 ratio and assigned the same index date as their matched pairs. Patients selected for the study were required to have 12 months’ continuous enrollment before (baseline) and after (follow-up period; except for death) the index date. Those with secondary/gestational diabetes were excluded. Per patient per year (PPPY) HCRU and cost outcomes were examined and compared for both cohorts.

Results: After matching, baseline demographics (mean age [42 years], men [53%], and geographic regions) were balanced for the HYPO+IP (n=736) and matched (n=3,680) cohorts. Similar number of patients were indexed each year (2013-2020) in the HYPO+IP cohort. The mean Charlson Comorbidity Index scores (HYPO+IP: 2.3 vs. matched: 2.2) were also similar at baseline. During the one-year follow-up period, the HYPO+IP and matched cohort had a mean (standard deviation [SD]) of 1.7 (3.4) vs. 0.3 (1.0) inpatient hospital admissions PPPY, respectively (p<0.001), with a higher mean [SD] length of stay (10.6 [7.4] vs 9.0 [10.7] days, p<0.001) for the HYPO+IP cohort. Furthermore, HYPO+IP cohort also had more outpatient visits, prescriptions, and sick leaves PPPY (p<0.05). The average total annual healthcare costs were substantially higher in HYPO+IP cohort (€13,482 vs €7,500, p<0.001; Figure).

Conclusion: Adults with T1D and an inpatient SHE living in Germany had higher HCRU and incurred substantially higher healthcare costs than matched adults with T1D without an inpatient SHE. This study highlights the presence of high unmet need in these patients.

Disclosure: R. Ziegler: None.

104

Association of type 1 diabetes duration or duration of advanced technology utilisation with severe hypoglycaemic event frequency

P. Callahan ¹, A. Boateng-Kuffour¹, K. Chandarana¹, D. Barry¹, L. Chen¹, C.S. Kelly², H. Nguyen², K. Chapman², E. Cornelius², W.A. Wolf², W. Polonsky³;

¹Vertex Pharmaceuticals Incorporated, Boston, MA, USA, ²T1D Exchange, Boston, MA, USA, ³University of California, San Diego, Encinitas, USA.

Background and aims: Adults with type 1 diabetes (T1D) may experience severe hypoglycemic events (SHE) and impaired awareness of hypoglycemia (IAH) despite use of advanced technologies. It is unclear how a longer duration of T1D or continuous glucose monitor (CGM) use is associated with SHE

Materials and methods: Adult CGM users in the T1D Exchange Registry completed an online survey including SHE frequency, IAH, duration of CGM use, and years since T1D diagnosis. SHE cohorts were created based on SHE frequency in the past 12 months and presence/absence of IAH (defined in the Table). Descriptive results (mean [M], standard deviation) were reported. Bivariate correlations were used to test associations between duration of T1D or duration of CGM use with SHE frequency using Pearson’s correlation.

Results: Among 1,847 CGM users, the mean age was 45.9 years. Mean T1D duration was 29.0 years. Most participants (55.0%) used CGM for ≥5 years (Table). There were no significant associations between SHE frequency and T1D duration (r=0.03; p=0.25) or duration of CGM use (r=-0.03; p=0.15). The problematic SHE cohort reported slightly longer T1D duration (M=31.9 years) than other cohorts (M=26.8-28.6 years).

Conclusion: Severe hypoglycemia remains an important clinical risk for adults with T1D regardless of time spent living with T1D or duration of advanced monitoring technology use.

Disclosure: P. Callahan: Employment/Consultancy; Vertex Pharmaceuticals Incorporated.

105

Residual C-peptide is associated with new and persistent impaired awareness of hypoglycaemia in type 1 diabetes

R. Varkevisser ¹, T. Sas^2,3, H.-J. Aanstoot², B. Wolffenbuttel¹, M. van der Klauw¹;

¹University Medical Center Groningen, Groningen, Netherlands, ²Diabeter Netherlands, Rotterdam, Netherlands, ³Department of Pediatrics, Erasmus Medical Center - Sophia Children’s Hospital, Rotterdam, Netherlands.

Background and aims: Impaired awareness of hypoglycaemia (IAH) is an unwanted complication of insulin therapy that increases the risk of hypoglycaemic events. The aim of this study was to describe the change in IAH over time in the Dutch Type 1 Diabetes biomarkers cohort, and to investigate factors associated with change in awareness status.

Materials and methods: A prospective cohort of individuals with type 1 diabetes, with a C-peptide < 300 pmol/L, who had completed the Clarke questionnaire to determine IAH status at baseline and at 2-years follow up were included in this study. C-peptide was measured by an immunoradiometric assay (IM3639, Beckman Coulter, Brea, California, USA). Changes in awareness status were defined as unchanged normal awareness (NAH), unchanged IAH, new IAH, and reversal of IAH. To investigate factors associated with awareness changes, two logistic regression analyses were conducted. Firstly, those with unchanged NAH and new IAH were compared. Secondly, individuals with unchanged IAH and reversal of IAH were compared. In both comparisons, univariate and age-adjusted odds ratios were calculated. After these analyses, three multivariate models were fitted using forward and backward stepwise selection using a 0.10 P-value cut-off, and stepwise backward selection using AIC criteria. Sensitivity analyses were conducted by excluding individuals whose blood samples had not been taken in a fasting state.

Results: Of the 611 individuals participating in the cohort, 431 had a C-peptide < 300 pmol/L and had completed the Clarke questionnaire at baseline and 2 years follow-up. The 178 excluded individuals were comparable to those included for analysis. Baseline prevalence of IAH was 17%, and 20% after 2-years follow-up. The incidence proportion of new IAH was 9.5%, and 31% for the reversal of IAH. When multivariate models were fitted, log-transformed C-peptide (OR 0.60, CI 0.34, 0.89) and diastolic blood pressure (OR 1.07, CI: 1.02, 1.12) were independently associated with new IAH in all three models. However, change in body mass index over time (OR 1.66, CI: 1.09, 2,91) was associated with reversal of IAH, in all three multivariate models. Sensitivity analysis excluding non-fasting blood draws yielded comparable results.

Conclusion: In this study, we not only reaffirm the previously reported independent association between low C-peptide and IAH, but also demonstrate that low C-peptide is independently associated with new incidence of IAH. This may suggest a potential temporal relationship wherein lower levels of C-peptide may predispose individual to the onset IAH. Diastolic blood pressure likely behaves as a proxy for vascular stiffness, and may indicate a relationship between microvascular damage, changes in perfusion and IAH. An increase in BMI over time was associated with reversal of IAH. This may be a result of compensatory behaviours, such as preventive snacking, leading to weight gain.

Clinical Trial Registration Number: NCT04977635

Supported by: JDRF 3-SRA-2014-291-M-R, DFN 2015.16.1856

Disclosure: R. Varkevisser: None.

106

A single experimental hypoglycaemic event induces acute and prolonged transcriptional changes in circulating monocytes of people with type 1 diabetes and healthy controls

M.S. Hendriksz ¹, R. Siebeler², J.I.P. van Heck¹, C.E.M. Verhulst¹, M.A. Hoeksema², M.P.J. de Winther², B.E. de Galan^1,3, C.J. Tack¹, R. Stienstra^1,4;

¹Department of Internal Medicine, Radboud University Medical Centre, Nijmegen, Netherlands, ²Department of Medical Biochemistry, Experimental Vascular Biology, Amsterdam UMC, University of Amsterdam, Amsterdam, Netherlands, ³Department of Internal Medicine, Maastricht University Medical Centre, MUMC+, Maastricht, Netherlands, ⁴Division of Human Nutrition and Health, Wageningen University, Wageningen, Netherlands.

Background and aims: A single experimental hypoglycaemic event in people with type 1 diabetes and in healthy controls increases absolute numbers of circulating monocytes, an effect that is sustained up to 7 days. However, the acute and prolonged impact of hypoglycaemia on monocyte function remains largely unknown. Therefore, we used RNA sequencing to study gene expression changes in circulating monocytes during hypoglycaemia and after 7 days in people with type 1 diabetes and healthy controls.

Materials and methods: Adults with type 1 diabetes (n=25) and matched controls (n=8) underwent a hyperinsulinaemic-euglycaemic hypoglycaemic (2.8 ± 0.1 mmoL/L [49.9 ± 2.3 mg/dL]) glucose clamp for 60 minutes. During euglycaemia, hypoglycaemia and 7 days later, blood was drawn, RNA was isolated from monocytes and used for RNA sequencing.

Results: In the healthy control group, data analyses revealed 219 and 117 differentially expressed genes (FDR < 0.05, Log2FoldChange > 0.5) during hypoglycaemia and 7 days later, respectively, when compared to euglycaemia. In the type 1 diabetes group, 91 differentially expressed genes were found during hypoglycaemia and 226 after 7 days. Pathways that were most significantly enriched (FDR < 0.05) in the healthy control group during hypoglycaemia included regulation of cell adhesion and several inflammatory pathways. After 7 days, the highest enriched pathways were mainly associated with mitochondrial metabolism, transcription and translation. These significantly enriched pathways were not observed in the type 1 diabetes group neither during hypoglycaemia nor 7 days later. Instead, most significantly enriched pathways (FDR < 0.05) during hypoglycaemia were related to protein tyrosine kinase signaling, and transcription and methylation after 7 days.

Conclusion: A single experimental hypoglycaemic event induces acute and prolonged transcriptional changes in circulating monocytes of both people with type 1 diabetes and healthy controls. However, groups differed with respect to significantly enriched pathways, e.g. associated with mitochondrial metabolism, that were seen in controls, but not in people with diabetes, possibly due to prior exposure to hypoglycaemia in the latter. Further research is needed to identify the signals that drive monocyte transcriptional changes and to determine the downstream consequences for monocyte function.

Clinical Trial Registration Number: NCT03976271

Supported by: This study has received funding from the Innovative Medicines Initiative 2 Joint Undertaking (JU) under grant agreement no. 777460.

Disclosure: M.S. Hendriksz: None.

107

Pathophysiological mechanisms of postprandial hypoglycaemia after gastric bypass in type 2 diabetes

D. Trico ¹, L. Sacchetta¹, E. Rebelos¹, N. Cimbalo¹, M. Chiriaco¹, D. Moriconi¹, L. Nesti¹, G. Nesti¹, S. Frascerra¹, M. Scozzaro¹, G. Daniele¹, S. Baldi¹, A. Mari², M. Nannipieri¹, A. Natali¹;

¹University of Pisa, Pisa, Italy, ²National Research Council, Padova, Italy.

Background and aims: Postprandial hypoglycemia (PPHG) is an underrecognized late complication of gastric bypass in people with type 2 diabetes (T2D), whose mechanisms remain unclear. Using an integrated physiology approach, we examined the prevalence of post-bypass PPHG and dissected its pathophysiological determinants in T2D.

Materials and methods: In this case-control study, subjects with history of T2D treated with Roux-en-Y gastric bypass (RYGB) 1 to 4 years before enrollment were characterized by a dual-labeled, 300-min, 75g oral glucose tolerance test (OGTT). Plasma glucose, glucose tracers, insulin, C peptide, GLP-1, GIP, glucagon, epinephrine, norepinephrine, and cortisol were measured. PPHG was defined as a plasma glucose nadir <3.0 mmol/L without symptoms or <3.3 mmol/L with symptoms of hypoglycemia. Mathematical models were implemented to estimate glucose metabolic fluxes and β-cell function parameters. At fasting and glucose nadir, 12-lead, 3-min continuous ECG and brachial blood pressure were recorded to detect alterations induced by counterregulatory sympathetic activation.

Results: A total of 24 subjects completed the assessments (14 women, age 54±9 years, BMI 29.3±6.3 kg/m², HbA1c 36±6 mmol/mol). PPHG occurred in 12 (50%) subjects, mostly (n=11) without warning symptoms (Panel A). Compared with those without hypoglycemia, PPHG subjects were more frequently women and had better glycemic control, with a greater rate of diabetes remission (100% vs 58%, p=0.04), despite similar age and body mass. PPHG was associated with higher tracer-derived glucose clearance (Panel B) due to early glucose-stimulated hyperinsulinemia and higher insulin sensitivity. Hyperinsulinemia was driven by enhanced β-cell function and reduced insulin clearance, while incretin hormones were similar between groups. PPHG subjects also had inadequate counterregulatory hormone responses to hypoglycemia, which prevented the physiological compensatory increase in endogenous glucose production. The independent effect of glucose clearance on PPHG was confirmed in multivariable analysis adjusted for potential confounders and after the exclusion of subjects with diabetes relapse in the control group. Morphology and duration of ECG waves, intervals, rhythm, heart rate variability, and blood pressure were within the normal range and similar between groups at fasting and glucose nadir.

Conclusion: PPHG frequently occurs in patients with T2D remission after gastric bypass, often without warning symptoms, driven by excessive glucose clearance due to the combination of early hyperinsulinemia and higher insulin sensitivity, along with defective counterregulation hindering the compensatory increase in glucose production. Post-bypass PPHG should be actively searched and managed in people with T2D for its potential negative health consequences.

Supported by: EFSD Rising Star Fellowship Programme supported by Novo Nordisk

European Society for Clinical Nutrition and Metabolism (ESPEN) Research Fellowship

Disclosure: D. Trico: None.

108

Metoclopramide improves hypoglycaemic counterregulation in diabetic rats

S.J. Fisher ¹, A.N. Iles¹, L.A. Schoeder¹, M.B. Music¹, B.V. Patel¹, A.R. Marksbury¹, M.M. Wooten¹, A.M. Woodcox¹, Z. Beckner¹, E.L. Macon², M.M. Devore¹;

¹University of Kentucky, Lexington, KY, USA, ²Animal Sciences, Texas A&M University, College Station, TX, USA.

Background and aims: To investigate the potential of metoclopramide, a dopamine antagonist, in mitigating impaired hypoglycemic counterregulation in a diabetic animal model.

Materials and methods: Diabetes was induced in 10-week-old Sprague-Dawley rats with streptozotocin (STZ, 65 mg/kg IP). Rats were randomized into three groups: 1) diabetic controls treated with recurrent saline (STZ+RS, n=6), 2) diabetic rats preconditioned with recurrent hypoglycemia (STZ+RH, n=7), and 3) diabetic rats preconditioned with recurrent hypoglycemia and metoclopramide (STZ+RH+MET: 3mg/kg IP, n=7). Following 3 days of preconditioning, all animals underwent a hyperinsulinemic (50mU/kg/min) hypoglycemic (~45 mg/dl) clamp.

Results: Glucose infusion rates to maintain hypoglycemia for STZ+RS (19±0.8 mg/kg/min) were higher in STZ+RH (27±0.9 mg/kg/min, p<0.0001 vs STZ+RS) and then reduced in STZ+RH+MET (24±0.1 mg/kg/min, p<0.05 vs STZ+RH). The glucagon response to hypoglycemia in control STZ+RS (p<0.05 vs basal) was abrogated in STZ+RH (p=NS vs basal) and restored in STZ+RH+MET (p<0.05 vs basal).

Conclusion: In this model of insulin deficient diabetes, the detrimental effect of recurrent hypoglycemia on glucoregulatory and counterregulatory hormonal response to hypoglycemia was reversed with metoclopramide treatment. It is concluded that the impaired response to hypoglycemia may be regulated via dopaminergic signaling.

Supported by: DK118082

Disclosure: S.J. Fisher: None.

OP 19 Prediction and prognosis of type 1 diabetes

109

Pioglitazone amplifies the decrease in HbA _1c and prevents the increase in plasma ketone caused by dapagliflozin in type 1 diabetes patients

R. DeFronzo ¹, M. Abdul-Ghani², G. Baskoy¹, A. Nakhleh³, S. Abdelgani¹, F. Al-Mulla⁴, M. Abu-Farha⁵, F. Alajmi⁵, T. Alessa⁴, N. Shehadeh⁶;

¹Diabetes Division, UTHSA, San Antonio, TX, USA, ²Medicine\Diabetes, University of Texas Health Science Center at San Antonio, San Antonio, TX, USA, ³The Azrieli Faculty of Medicine, Bar-Ilan University, Safed, Israel, Institute of Endocrinology, Diabetes and Metabolism, Rambam Health Care Campus, Haifa, Israel, ⁴⁵Dasman Diabetes Institute, Kuwait, Kuwait, ⁶Haifa, Israel.

Background and aims: SGLT2 inhibitors (SGLT2i) lower the plasma glucose concentration in T1DM patients. However, they cause an increase in plasma ketone concentration and risk of ketoacidosis. The aim of the present study is to examine whether pioglitazone amplifies the decrease in HbA1c and prevents the increase in plasma ketone concentration caused by dapagliflozin in T1DM patients.

Materials and methods: After a 4-week run in period, 60 T1DM patients received dapagliflozin 10 mg for 12 weeks. At week 16 patients were randomized to receive in double blind fashion pioglitazone (45 mg) or matching placebo for an additional 16 weeks. If at any stage of the study, the plasma ketone concentration exceeded 1 mM, open label pioglitazone (45 mg/day) was initiated for 16 weeks. The primary end point was the change from baseline to end of study (EOS) in HbA1c. The change from baseline to EOS in plasma ketone concentration was a secondary outcome. A linear model for repeated measures for HbA1c in terms of arm, visit, and arm by visit interaction with adjustment to baseline characteristics was used to contrast the difference in HbA1c and ketones between the two groups.

Results: T1DM patients were 42±3 years of age, 30% female, BMI=26.8±0.7, HbA1c=8.5±0.2%, insulin dose= 63±4 units, and eGFR=114±6 ml/min. The difference in HbA1c between the two treatment groups was 0.50%±0.23 (95% CI=-0.96, -0.05, p=0.03), suggesting significant decrease in HbA1c in subjects receiving dapagliflozin plus pioglitazone versus dapagliflozin plus placebo. The difference in plasma ketone concentration between the two treatment arms was -0.21±0.07 (95% CI= -0.34, -0.08, p<0.002), suggesting that pioglitazone significantly reduced the increase in plasma ketone concentration caused by dapagliflozin in T1DM patients. A total of 8 subjects had their plasma ketone concentration exceeded 1 mM while receiving dapagliflozin and prior to randomization to pioglitazone/placebo. Therefore, they were placed on open label pioglitazone (45 mg). The mean duration between initiating dapagliflozin and the increase in plasma ketones to >1 mM was 5.4±1.2 weeks. Initiating pioglitazone produced a precipitous decline in the mean plasma ketone concentration to a level below baseline (from 1.23±0.18 mM to 0.24±0.05 mM, p<0.002).

Conclusion: Addition of pioglitazone to SGLT2i in T1DM patients amplifies the decrease in HbA1c and prevents the increase in plasma ketone caused by SGLT2i. This combination provide a therapeutic option adjunctive to insulin in T1DM, and allows long-term cardiovascular and renal outcome studies to be carried out safely in T1DM patients.

Supported by: JDRF

Disclosure: R. DeFronzo: Grants; Boehringer Ingelheim, AstraZeneca, 89Bio, Amgen, Medality, CORCEPT. Other; Advisory Board: AstraZeneca, Novo Nordisk, Boehringer Ingelheim, Intarcia, Aardvark, Renalytix, CORCEPT, Alnylam, Speakers Bureau: AstraZeneca, CORCEPT, Renalytix.

110

Pancreatic imaging biomarkers of increased fibro-inflammation, low volume and low fat characterise type 1 diabetes: a UK Biobank study

H. Thomaides-Brears ¹, E. Jackson¹, A. Triay-Bagur¹, F. McCann², T. Mahon², H. Al-Mossawi², M. Pansini^1,3;

¹Perspectum, Oxford, UK, ²Immunocore, Oxford, UK, ³Clinica Di Radiologia EOC, Lugano, Switzerland.

Background and aims: Type 1 diabetes (T1D) is an autoimmune disease characterized by chronic inflammation of the pancreatic islets, beta cell loss leading to pancreatic atrophy due to increased immune cell infiltration of the exocrine pancreas, parenchymal fibrosis can also be evident. Biopsy of the pancreas is challenging because of its location, size and consequent risk of injury to the organ. We aimed to characterise T1D using non-invasive multi-parametric magnetic resonance imaging (mpMRI) markers to evaluate fibro-inflammation, fat and size in the pancreas compared to healthy individuals and those with type 2 diabetes (T2D) from the general population.

Materials and methods: Pancreas mpMRI metrics of fibro-inflammation (srT1), fat (proton density fat fraction, PDFF) and volume were extracted from abdominal imaging scans in the UK Biobank (application 9914) using a semi-automated pipeline. Three sub-groups were selected for comparison: T1D [n=106, 59% male, mean age: 64 (SD: 8) years, BMI 29 (6) kg/m², HbA1c 56 (14) mmol/mol, T1D duration 13 (13) years], T2D [n= 1190, 67% male, age: 66 (7) years, BMI 31 (5) kg/m², HbA1c 46 (11) mmol/mol, T2D duration 7 (7) years] and healthy controls without metabolic syndrome (non-metS) [n=17433, 44% male, age: 63 (8) years, BMI 25 (3) kg/m², HbA1c 34 (3) mmol/mol].

Results: Pancreas srT1 was significantly higher (p<0.001) in individuals with T1D (813 (91) ms) than in healthy controls (759 (73) ms) or in those with T2D (763 (77) ms). Pancreatic volume (T1D: 59 (17) ml, T2D: 68 (18) ml, non-metS: 66 (15) ml, p<0.001) and fat were significantly lower (T1D: 2.99 (2.04) %, T2D: 4.75 (3.34) %, non-metS: 3.65 (2.65) %, p<0.001). There was a negative correlation between srT1 and volume (r = -0.48 (-0.49, -0.47 p < 0.001). In individuals with T1D, srT1 tended to decrease with increasing pancreatic fat (r = -0.15 (-0.37, 0.08) p = 0.2), and this relationship was reversed in the healthy controls and T2D populations (r = 0.18 (17, 20) p < 0.001 and 0.08 (0.01, 0.16) p < 0.05, respectively). After accounting for age, sex, BMI and waist circumference in a linear model, T1D was associated with a 10 (1.4) ml reduction in pancreatic volume and a 57 (7) ms increase in srT1 (p< 0.001). After also accounting for reduced volume in individuals with T1D there was a 35 (6) ms increase in srT1 (p< 0.001). Figure left: A colourimetric map of srT1 from an individual with type 1 diabetes with elevated srT1 characteristic of fibro-inflammation. Right: srT1 vs pancreatic fat values.

Conclusion: Using mpMRI we found that the pancreata of adults with type 1 diabetes are characterised by increased fibro-inflammation, less fat and smaller size. Non-invasive imaging could be used to evaluate the pancreas, and measure effects of T1D treatments in clinical trials.

Disclosure: H. Thomaides-Brears: Employment/Consultancy; Perspectum.

111

Mortality and factors associated with it in Finnish patients with type 1 diabetes

E. Putula ^1,2, T. Kauppala³, S. Vanhamäki³, J. Haapakoski³, T. Laatikainen^3,4, S. Metso²;

¹Faculty of Medicine and Health Technology, Tampere University, Tampere, Finland, ²Department of Internal Medicine, Unit of Endocrinology, Tampere University Hospital, Tampere, Finland, ³Finnish Institute for Health and Welfare, THL, Helsinki, Finland, ⁴University of Eastern Finland, Faculty of Health Sciences, Finland.

Background and aims: According to earlier studies, life expectancy for individuals with type 1 diabetes is almost ten years lower than for the general population. The aim of the study was to assess the combined effect of mental and behavioural disorders, cardiovascular diseases (CVD), end-stage renal disease (ESRD), diabetic foot complications, HbA1c, LDL-cholesterol and chronic kidney disease according to estimated glomerular filtration rate (eGFR) and urine albumin-creatinine ratio (U-ACR) on all-cause mortality.

Materials and methods: The study included all 45,801 patients with diabetes type 1 from the Finnish Diabetes Registry during 2018-2022, and their diagnoses and laboratory values from the Patient Data Repository of Kanta services. The information on the time of death was obtained from the Digital and population data services agency. Mortality with patients with type 1 diabetes was compared with mortality in the non-diabetic population in Finland by estimating standard mortality rates (SMRs). SMRs were calculated as the ratio of the number of observed deaths in the study population and the number of expected deaths in general Finnish population. Poisson regression model was used to estimate the effect of risk factors on the SMR.

Results: Patients’ median age was 39 years. Median HbA1c was 64 mmol/mol (8%) and median LDL-cholesterol was 2.4 mmol/l. Seventeen percent of the patients had prevalent CVD, 21% had mental or behavioural disorder, 5% had foot complications, and 8% belonged to high or very high risk according to the U-ACR or eGFR. A total of 2,469 patients died during the follow-up. SMR for the total cohort was 1.82 (95% CI 1.75-1.89). SMRs according to age group formed a u-shaped curve that peaked in the age group of 40-49 years with SMR of 3.73 (95% CI 3.20-4.32). The absence of comorbidities was protective factor from death (RR 0.60 [95% CI 0.49-0.75]) while the risk ratio for HbA1c ≥64 mmol/mol was 1.27 (95% CI 1.14-1.42), LDL-cholesterol ≥2,6 mmol/l 1.33 (95% CI 1.17-1.51), high or very high risk according to eGFR or U-ACR 3.57 (95% CI 3.15-5.05), prevalent ESRD 2.56 (95% CI 2.22-2.96), CVD 1.46 (95% CI 1.23-1.74), foot complications 1.48 (95% CI 1.30-1.68), and mental or behavioural disorder 1.42 (95% CI 1.24-1.62), of which alcohol and substance abuse had the most significant contribution to mortality, RR 2.17 (95% CI 1.80-2.61).

Conclusion: There is substantial excess mortality due to type 1 diabetes peaking at the age of 40-49 years in Finland. Chronic kidney disease is the most important risk factor for premature death. On top of the traditional complications of diabetes, mental and behavioural disorders increase the risk of mortality with similar risk ratio than cardiovascular diseases and foot complications. Therefore, interventions should focus on improving both the traditional renal and cardiovascular risk factors but also mental health and especially alcohol and substance abuse.

Supported by: This study was financially supported by the State funding for university-level health research, Tampere University Hospital, Project No. MK335

Disclosure: E. Putula: Employment/Consultancy; This study was financially supported by the State funding for university-level health research, Tampere University Hospital, Wellbeing services county of Pirkanmaa / Project No. MK335.

112

Influence of the type 1 diabetes genetic risk score on response to immunotherapies for type 1 diabetes prevention

M.J. Redondo ¹, B. Bundy², H. Parikh², L. You², T. Triolo³, L. Ferrat⁴, E. Templeman⁴, M. Tosur¹, A.K. Steck³, P. Gottlieb³, S. Onengut-Gumuscu⁵, S.S. Rich⁵, R. Oram⁴, K.C. Herold⁶, J.P. Krischer²;

¹Baylor College of Medicine, Houston, TX, USA, ²University of South Florida, Tampa, FL, USA, ³University of Colorado, Aurora, CO, USA, ⁴Exeter University, Exeter, UK, ⁵University of Virginia, Charlottesville, VA, USA, ⁶Yale University, New Haven, CT, USA.

Background and aims: Teplizumab has shown efficacy at delaying the onset of type 1 diabetes (T1D) but markers of response to treatment are needed to increase risk-benefit ratio. The T1D genetic risk score-2 (GRS2) combines 67 HLA and non-HLA single nucleotide polymorphisms (SNPs) that increase T1D risk. We aimed to test the hypothesis that the likelihood of response to immunotherapies for T1D prevention increases with higher T1D GRS2.

Materials and methods: We studied autoantibody-positive participants in the TrialNet Anti-CD3 Prevention study (TN10) (n=44 teplizumab, n=32 placebo), TrialNet Abatacept Prevention study (TN18) (n=101 abatacept, n=111 placebo) and TrialNet Oral Insulin Prevention study (TN07) (n=283 oral insulin, n=277 placebo). Martingale residual scatterplots (MRS) based on the Cox Model were employed to assess T1D risk.

Results: Except for TrialNet Anti-CD3 Prevention (TN10), MRS smoothing curves demonstrated that GRS2 is directly correlated with T1D risk. Analyzing TN10 by arm, the smoothing curve showed that T1D risk increases for the placebo arm while, for the treated arm, the risk appears to decrease, creating a crossover point at GRS=12.8 near zero (relative risk=1). We then fitted a Cox Model including GRS2 dichotomized at 13, arm and an interaction term. The interaction term was significant (p=0.03) and the hazard ratio of T1D risk (teplizumab to placebo) for GRS2 ≥13 was 0.263 (95%CI=0.123-0.562) but 0.898 (95^th% CI=0.295, 2.74) for GRS2<13. Analyses were adjusted for age.

Conclusion: In clinical trials with agents that failed to prevent progression to stage 3 (clinical) T1D (e.g., abatacept and oral insulin), higher T1D GRS2 (reflecting higher genetic burden of T1D risk) was associated with higher risk of progression, as expected. However, teplizumab modified the relationship between GRS2 and T1D risk suggesting a greater beneficial therapeutic effect in participants with GRS2 ≥13. In contrast, participants with lower GRS2 had decreased response to teplizumab. These results support that genetics can help select individuals who will respond to treatments to prevent T1D.

Supported by: National Institutes of Health (NIH), NIDDK R01DK121843

Disclosure: M.J. Redondo: None.

113

Islet Antigen-2A (IA-2A) autoantibody positivity increases risk of progression within and across established stages of type 1 diabetes

E. Sims ¹, D. Cuthbertson², L. Ferrat³, E. Bosi⁴, C. Evans-Molina⁵, L.A. DiMeglio¹, B. Nathan⁶, H. Ismail¹, L. Jacobsen⁷, M. Redondo⁸, R. Oram⁹, J. Sosenko¹⁰;

¹Indiana University School of Medicine, Indianapolis, IN, USA, ²University of South Florida, Tampa, FL, USA, ³University of Geneva, Geneva, Switzerland, ⁴Internal Medicine, San Raffaele Scientific Institute, Milan, Italy, ⁵Indiana Diabetes Research Center, Indiana University, Indianapolis, IN, USA, ⁶University of Minnesota, Minneapolis, MN, USA, ⁷University of Florida, Gainesville, FL, USA, ⁸Texas Children’s Hospital, Houston, TX, USA, ⁹Institute of Biomedical and Clinical Science, University of Exeter Medical School, Exeter, UK, ¹⁰University of Miami, Miami, FL, USA.

Background and aims: Accurate understanding of type 1 diabetes risk is critical to optimize counseling, monitoring, and interventions, yet even within established staging classifications, individual time to clinical disease varies. Based on work suggesting islet antigen-2A autoantibody positivity (IA-2A+) is associated with increased type 1 diabetes progression, we asked if IA-2A would consistently identify individuals at higher risk across the natural history of autoantibody positivity.

Materials and methods: Genetic, autoantibody, and metabolic data from adult and pediatric autoantibody-negative (n=192) and autoantibody-positive (n=4577) relatives of individuals with type 1 diabetes followed longitudinally in the Type 1 Diabetes TrialNet Pathway to Prevention Study were analyzed. Cox regressions were used to compare cumulative incidences of clinical diabetes by autoantibody profiles and disease stages.

Results: Compared to IA-2A-, IA-2A+ individuals had higher genetic risk scores and clinical progression risk within single autoantibody-positive, Stage 1, and Stage 2 type 1 diabetes categories. People with a single IA-2A+ showed increased metabolic dysfunction and diabetes progression compared to people who were autoantibody-negative, other single autoantibody-positive, and IA-2A- Stage 1 individuals. When considering all autoantibody positive individuals, progression risk was similar for euglycemic IA-2A+ persons and dysglycemic IA-2A- persons.

Conclusion: IA-2A positivity is consistently associated with increased progression risk throughout the natural history of type 1 diabetes development. Single IA-2A+ individuals have a greater risk of disease progression than those who meet Stage 1 criteria but are IA-2A- and should be considered as having high risk early-stage disease for interventions, and undergo more frequent monitoring.

Supported by: NIH, JDRF

Disclosure: E. Sims: Employment/Consultancy; Consultant, Sanofi, DRI HealthCare. Lecture/other fees; Lecture, American Diabetes Association, Medscape, HealthMatters CME, MedLearning Group.

114

Risk of developing type 1 diabetes in individuals with other autoimmune conditions

S. Edelman ¹, D. Agardh², N. Cui³, L. Hao³, X. Li³, M. Wieloch³, L. Meneghini³;

¹University of California San Diego, San Diego, CA, USA, ²Celiac Disease and Diabetes Unit, Lund University, Lund, Sweden, ³Sanofi, Paris, France.

Background and aims: Autoimmune conditions, such as thyroid diseases and celiac disease, are more prevalent in individuals with autoimmune type 1 diabetes (aT1D) than in the general population. However, few studies have examined the risk of developing aT1D in those with preexisting autoimmune diseases. The objective of this study was to compare the risk of developing aT1D in individuals with specific autoimmune conditions to that of individuals without autoimmune conditions.

Materials and methods: This was a retrospective observational matched-cohort study based on real-world data from the Optum Clinformatics claims database. Inclusion criteria for the autoimmune condition cohorts included at least 1 ICD diagnosis of celiac disease, hyperthyroidism (including Graves’ disease), or hypothyroidism (including Hashimoto’s thyroiditis) during the identification period (January 1, 2017, to September 30, 2023); at least 1 month of follow-up after the date of earliest diagnosis of the autoimmune condition of interest; no diagnosis of T1D or T2D or treatment for diabetes during the preceding 12-month baseline period; and continuous medical and pharmacy enrollment during the baseline and follow-up periods. The control cohort included individuals without any autoimmune conditions during the study period. Individuals from the autoimmune and control cohorts were matched 1:1 on propensity scores based on baseline demographics and clinical characteristics. A Cox proportional hazards model was used to compare the risk of T1D between cohorts. A random sample analysis was conducted for 5% of the total number of hypothyroid patients identified.

Results: Mean (SD) ages at baseline for the individuals with celiac disease, hyperthyroidism, and hypothyroidism were 48.4 years (21.8), 61.0 years (18.2), and 62.1 years (17.4), with median (IQR) follow-ups of 725 days (320-1420), 743 days (325-1442), and 814 days (349-1593), respectively. Baseline characteristics and follow-up time were similar between the autoimmune condition cohorts and their respective controls. T1D developed in 0.14% (68/47,099) of individuals with celiac disease compared to 0.06% (27/47,099) of controls. The median (IQR) time to T1D development was 436 days (187-790) for those with celiac disease and 542 days (213-827) for controls. Of those with hyperthyroidism, T1D developed in 0.17% (281/164,830) compared to 0.06% (99/164,830) of controls. The median (IQR) time to T1D development was 399 days (136-937) and 538 days (272-889) for those with hyperthyroidism and controls, respectively. Of those with hypothyroidism, T1D developed in 0.16% (118/74,095) compared to 0.05% (38/74,095) of controls. The median (IQR) time to T1D development was 470 days (138-837) for those with hypothyroidism and 678 days (354-1444) for controls. The risk of developing T1D was significantly greater for each of the autoimmune condition cohorts compared to their respective controls (celiac disease: HR=2.54 [95% CI: 1.63, 3.97; p<0.0001]; hyperthyroidism: adjusted HR=2.98 [95% CI: 2.37, 3.75; p<0.0001]; hypothyroidism: HR=3.19 [95% CI: 2.22, 4.61; p<0.0001]).

Conclusion: The risk of developing autoimmune T1D was 2.5- to 3-fold higher for individuals with celiac and thyroid autoimmunity compared to those without autoimmune conditions. These findings may support screening of individuals with autoimmune conditions for early T1D.

Supported by: This study and medical writing support was funded by Sanofi.

Disclosure: S. Edelman: Employment/Consultancy; AstraZeneca, Dexcom, Eli Lilly, MannKind, Merck, Novo Nordisk, Sanofi, Senseonics. Lecture/other fees; Eli Lilly, Sanofi, Xeris.

OP 20 Are you afrAID?

115

Quality-of-life benefits for adults living with type 1 diabetes using the Omnipod 5 System compared with insulin pump therapy: results from a randomised controlled trial

J.-P. Riveline ¹, E. Renard², R.S. Weinstock³, T.T. Ly⁴, The OP5-003 Research Group;

¹Hôpital Lariboisière, APHP, Université Paris-Cité, Paris, France, ²Montpellier University Hospital, Montpellier, France, ³SUNY Upstate Medical University, Syracuse, NY, USA, ⁴Insulet Corporation, Acton, MA, USA.

Background and aims: Beyond their glycemic benefits, automated insulin delivery (AID) systems have been shown to improve quality of life in individuals living with type 1 diabetes (T1D). In a recent randomized controlled trial (RCT), the Omnipod® 5 AID System demonstrated safety and superior efficacy compared with insulin pump therapy plus continuous glucose monitoring (CGM) in adults with T1D. The purpose of this sub-analysis was to evaluate quality-of-life and psychosocial measures assessed during the study.

Materials and methods: In this multicenter RCT, adults with T1D across the United States and France were randomized (2:1) to use either the Omnipod 5 System with Dexcom G6 CGM (intervention arm) or the participant’s current insulin pump with Dexcom G6 CGM (control arm, no automation) for 13 weeks. Participants completed the following quality-of-life and psychosocial questionnaires at baseline and end of study: the European Quality of Life 5 Dimensions 3 Level Version (EQ-5D-3L), DAWN2 Impact of Diabetes Profile (DIDP), Pittsburgh Sleep Quality Index (PSQI), System Usability Scale (SUS), and Insulin Dosing Systems: Perceptions, Ideas, Reflections, and Expectations (INSPIRE).

Results: A total of 196 participants were enrolled, with 132 randomized to intervention and 62 randomized to control. Questionnaire results are provided in the Table. The EQ-5D-3L Visual Analogue Scale and Index Scores demonstrated significant improvements in overall quality of life (mean adjusted difference [95% CI] of 3.8 [0.8, 6.7]; p=0.0132 and 0.04 [0.02, 0.07]; p=0.0024, respectively) with Omnipod 5 use compared with standard pump therapy. Improvement in the perceived impact of diabetes on quality of life (DIDP) with Omnipod 5 was also observed (-2.8 [-4.7, -0.9]; p=0.0043). Participants using the Omnipod 5 System saw additional benefits with respect to sleep quality (-0.8 [-1.5, -0.1]; p=0.0308) and system usability (15.7 [11.2, 20.2]; p<0.0001) compared with those on standard pump therapy. Notably, INSPIRE questionnaire scores were high at baseline and study end in the intervention arm (mean 83.7 compared with 81.9, p=0.3011; optimal score: 100), suggesting positive perceptions of the Omnipod 5 System.

Conclusion: In addition to the safety and glycemic efficacy of the Omnipod 5 System, these results reveal significant improvements across quality-of-life and psychosocial measures for adults with T1D with Omnipod 5 use. These findings provide additional evidence to support that the Omnipod 5 System can alleviate some of the burden of diabetes self-management compared with standard pump therapy.

Clinical Trial Registration Number: NCT05409131

Supported by: This study was funded by Insulet Corporation

Disclosure: J. Riveline: Grants; Abbott, Air Liquide Santé International, Sanofi-Aventis, Novo Nordisk. Other; Sanofi-Aventis MSD, Eli Lilly, Novo Nordisk, AstraZeneca, Abbott, Dexcom, Alphadiab, Medtronic.

116

Early real-world performance of the Omnipod® 5 automated insulin delivery (AID) system in 9,900 people with type 1 diabetes in Europe

E.G. Wilmot ¹, T. Biester², F.M. Campbell³, D. Deiss⁴, J. Elliott⁵, F. Gibb⁶, J. Kröger⁷, L.M. Huyett⁸, J.J. Méndez⁸, I. Hadjiyianni⁸, L.R. Conroy⁸, T.T. Ly⁸;

¹University of Nottingham, School of Medicine, Royal Derby Hospital, Derby, UK, ²Auf der Bult Hospital for Children and Adolescents, Hannover, Germany, ³Leeds Children’s Hospital, Leeds, UK, ⁴MEDICOVER Berlin-Mitte, Berlin, Germany, ⁵University of Sheffield, Sheffield, UK, ⁶University of Edinburgh, Edinburgh Centre for Endocrinology & Diabetes, Edinburgh, UK, ⁷Center for Digital Diabetology, Hamburg, Germany, ⁸Insulet Corporation, Acton, MA, USA.

Background and aims: The tubeless Omnipod 5 AID System, which allows for personalized therapy through customizable glucose targets from 110-150mg/dL (6.1-8.3mmol/L) in 10mg/dL (0.55mmol/L) increments, recently became commercially available for people with type 1 diabetes (T1D) aged 2 years and older in the United Kingdom (UK) and Germany. This study aimed to evaluate the early real-world performance of the Omnipod 5 System in the first cohort of European users.

Materials and methods: A retrospective analysis of continuous glucose monitoring (CGM) and insulin data from Omnipod 5 users with T1D aged ≥2y and using ≥5 units of insulin per day in the UK and Germany that provided consent and had ≥90 days of data available in the cloud-based data management system was conducted.

Results: Data from 9,900 users (86.2% from the UK, 13.8% from Germany) with sufficient CGM data (≥75% of days with ≥220 readings), aged 2 to <18y (n=5,146) and ≥18y (n=4,754) were available at the time of analysis, with >1.5 million user-days of data in total. Outcomes are shown in the Table for average glucose targets. In adult users, median time in target range (70-180mg/dL; 3.9-10.0mmol/L) was 70.5%, 64.5%, and 54.1% for the 110mg/dL [6.1mmol/L], 120mg/dL [6.7mmol/L], and 130-150mg/dL [7.2-8.3mmol/L] targets, respectively, and time spent in hypoglycemia (<70mg/dL; <3.9mmol/L) was low (median 1.16%, 0.82%, and 0.57%, respectively) across glucose targets. Similar trends were seen in pediatric users who achieved a median time in target range of 66.4%, 64.9%, and 60.8% for the 110mg/dL [6.1mmol/L], 120mg/dL [6.7mmol/L], and 130-150mg/dL [7.2-8.3mmol/L] targets, respectively, with a low percentage of time spent in hypoglycemia across glucose targets (median 1.63%, 1.59%, and 1.42%, respectively). Use of the lowest target (used by 53.4% of the total population) was associated with the highest TIR, minimal time in hypoglycemia, and the highest proportion of users achieving clinical targets in both age groups.

Conclusion: Collectively, these early real-world results of Omnipod 5 use in almost 10,000 people with T1D in the UK and Germany demonstrate that the highly favorable glycemic outcomes first reported in the United States are achievable across populations. Additionally, these findings support the idea that users who are seeking to improve their TIR should consider decreasing their glucose target toward the lowest setting whenever possible.

Supported by: This study was funded by Insulet Corporation.

Disclosure: E.G. Wilmot: Grants; Abbott Diabetes Care, Embecta, Insulet Corporation, Novo Nordisk, Sanofi. Lecture/other fees; Abbott Diabetes Care, AstraZeneca, Dexcom, Eli Lilly, Embecta, Insulet Corporation, Medtronic, Novo Nordisk, Roche, Sanofi, Ypsomed.

117

Real-world glycaemic and person-reported outcomes one year after Tandem Control IQ ^TM initiation in children and adults with type 1 diabetes

J. De Meulemeester ¹, L. Valgaerts¹, C. De Block², L. Van Huffel³, D. Ballaux⁴, K. Spincemaille⁵, Y. Taes⁶, B. Lapauw⁷, G. Massa⁸, M. den Brinker², A. Messaaoui⁹, I. Gies¹⁰, B. Keymeulen¹⁰, C. Mathieu¹, P. Gillard¹;

¹University Hospitals Leuven – KU Leuven, Leuven, Belgium, ²University Hospital Antwerp, Edegem, Belgium, ³OLV Hospital Aalst, Aalst, Belgium, ⁴Vitaz, Sint-Niklaas, Belgium, ⁵AZ Delta, Roeselare, Belgium, ⁶AZ Sint-Jan, Brugge, Belgium, ⁷Ghent University Hospital, Ghent, Belgium, ⁸Jessa Hospital, Hasselt, Belgium, ⁹Universitair Kinderziekenhuis Koningin Fabiola, Brussels, Belgium, ¹⁰University Hospital Brussels, Brussels, Belgium.

Background and aims: Real-world (RW) evidence complements outcomes from RCTs on the use of hybrid closed-loop systems in the management of people with type 1 diabetes (T1D). This 12-month prospective observational study evaluated the RW impact of Tandem Control IQ^TM on glycemic control and person-reported outcomes (PROs) in children ≥6 years and adults with T1D.

Materials and methods: Between Oct 2021 and Dec 2022, all children (n = 114) and adults (n = 473) with T1D who started Tandem Control IQ^TM were recruited at 18 Belgian centers. Data were prospectively collected at start, and 4, 8 and 12 months of standard follow-up. PROs were evaluated through validated questionnaires. Data are reported as mean ± SD or least-squares mean (95% CI).

Results: Most children were girls (61%), white (79%), with a mean age of 12 ± 3 years. They had T1D for 6 ± 4 years and the majority (81%) used an insulin pump before. TIR increased from start to 4 months (51.6% [47.6-55.5] to 67.0% [63.8-70.2], p<0.001) and was sustained up to 12 months (64.4% [61.2-67.5], p<0.001). After 12 months, HbA1c decreased from 7.8% (7.6-8.1) to 7.1% (6.9-7.3), time <70 mg/dL from 3.9% (3.1-4.8) to 2.7% (1.9-3.5), time >180 mg/dL from 44.1% (39.8-48.5) to 32.9% (29.2-36.5), and time >250 mg/dL from 21.7% (17.9-25.5) to 13.0% (10.7-15.3) (all p<0.001). Parents reported less work absenteeism (408.0 vs 95.5 days/100 patient years, p<0.001) and children less school absenteeism (620.2 vs 328.1 days/100 patient years, p=0.001) one year after start of Tandem Control IQ^TM compared to one year before. Most adults were female (57%), white (95%), with a mean age of 39 ± 13 years. They had T1D for 20 ± 13 years and 53% used an insulin pump before. TIR increased from start to 4 months (58.7% [56.6-60.8] to 72.0% [70.2-73.9], p<0.001) and remained stable up to 12 months (70.9% [69.1-72.8], p<0.001). After 12 months, HbA1c decreased from 7.4% (7.3-7.5) to 6.7% (6.6-6.8), time <70 mg/dL from 4.2% (3.9-4.6) to 1.9% (1.8-2.1), time <54 mg/dL from 1.0% (0.9-1.2) to 0.4% (0.3-0.4), time >180 mg/dL from 37.0% (34.9-39.2) to 27.1% (25.3-28.9), and time >250 mg/dL from 13.2% (11.8-14.6) to 7.3% (6.3-8.4) (all p<0.001). Figure 1 shows the impact of improved glycemic control on PROs in both adults, children and their parents.

Conclusion: One year RW use of Tandem Control-IQ^TM in children and adults with T1D is associated with better glycemic control, more diabetes-related quality of life, and less worries about hypoglycemia for parents.

Clinical Trial Registration Number: NCT04414280

Supported by: A research grant was received from Tandem Diabetes Care (San Diego, CA, USA) and Dexcom (San Diego, CA, USA)

Disclosure: J. De Meulemeester: None.

118

Sustained 1-year improvement of glucose control under free-life hybrid closed-loop insulin therapy in children with type 1 diabetes: national observatory of closed-loop in France (OB2F)

E. Renard ¹, J.-P. Riveline², J.-B. Julla², F. Dalla-Vale¹, A. Spiteri³, L. Mathivon⁴, J.-F. Gautier², E. Bonnemaison⁵;

¹Montpellier University Hospital, Montpellier, France, ²Lariboisiere Hospital, Paris, France, ³Grenoble-Alpes University Hospital, Grenoble, France, ⁴Meaux General Hospital, Meaux, France, ⁵Saint Jean Clinic, Saint Jean de Védas Montpellier Métropole, France.

Background and aims: The efficacy of hybrid closed-loop insulin therapy (HCL) on glucose control in children with type 1 diabetes (T1D) has been reported in clinical trials. The OB2F nation-wide observatory aims at assessing HCL outcomes on glucose control following its initiation in common practice in this population during the year 2022.

Materials and methods: 480 participants were included in 23 pediatric departments. Clinical characteristics, metrics of continuous glucose monitoring (CGM) and HbA1c levels were collected at the initiation of HCL and after 6-month and 1-year follow-up. Severe Hypoglycemic and ketoacidosis events were collected during the previous year before inclusion and the 6-month and 1-year follow-up. Data is presented as median [IQR]. Mann-Whitney and Chi2 tests were used for statistical analyses.

Results: Participant characteristics at initiation of Tandem Control-IQ (64%) or MiniMed 780G (36%) HCL systems were: age 12 yrs [7-17], 53% boys, diabetes duration: 6 yrs [4-8]. Collected data at initiation, after 6 and 12 months was: HbA1c (% / mmol/mol) : 7.5 [7.0-8.0] / 58 [53-64], 7.0 [6.6-7.4] / 53[49-57] and 7.0 [6.7-7.4] / 53 [50-57]; % time in range 70-180 mg/dl (TIR): 54.0 [45.0-62.8], 66.6 [60.4-73.0] and 66.6 [60.7-73.0] ; % time>180 mg/dl: 42.7 [35.0-50.3], 31 [24.0-37.7] and 31 [23-37]; % time <70 mg/dl (TBR): 2.0 [1.1-3.5], 2.0 [1.0-3.3] and 2.0 [1.0-3.6] ; % coefficient of variation: 40.8 [37.0-44.0], 39.0 [36.0-43.0] and 39.7 [36.0-43.0], respectively. All differences but for TBR were significant between initiation and 6 months (p<0.0001), and not significant between 6 and 12 months. The % of participants who experienced at least 1 severe hypoglycemia moved from 2.7 to 0.9 and 0 (not significant), while the % of participants who experienced at least 1 ketoacidosis moved from 3.1 to 0.6 and 1.3 (p<0.0071), respectively. The % of participants reaching ISPAD goals of >70% TIR and <4% TBR moved from 8.3 at initiation to 26.5 after 1 year, while the % of participants reaching HbA1c <7% (53 mmol/mol) moved from 24 to 47.8. Only 1% of children stopped HCL before 6 months and 2% between 6 and 12 months.

Conclusion: This data show sustained 1-year improvement of glucose control while using HCL in a large nation-wide population of children with T1D, with no safety issue and few drop-outs in common practice.

Disclosure: E. Renard: None.

119

Glucose control in free-life under hybrid closed loop 6 and 12 months after initiation from a nation-wide survey

J.-P. Riveline ¹, J. Julla², C. Campinos³, S. Favre⁴, I. Tauveron⁵, S. Borot⁶, L. Kessler⁷, A. Carlier⁸, M. Giraud⁹, M. Joubert¹⁰, N. Hamamouche¹¹, S. Lablanche⁶, M. Huet¹², J.-F. Gautier¹³, E. Renard¹⁴;

¹Department of Endocrinology and Diabetes, Centre Universitaire du diabète et de ses complications, A, Paris, France, ²Institut Necker Enfants Malades, INSERM U1151, CNRS UMR 8253, IMMEDIAB Laboratory, Paris, France, Department of Endocrinology and Diabetes, Centre Universitaire du diabète et de ses complications, A, Paris, France, ³Hôpital René Dubos, Pontoise, France, ⁴Switzerland, ⁵Diabetes, CHU G. Montpied, Clermont-Ferrand, France, ⁶⁷CHRU de Strasbourg, Strasbourg, France, ⁸APHP - Bichat, Paris, France, ⁹CH Vannes, Vannes, France, ¹⁰Endocrinology Unit, University Hospital of Caen, Caen, France, ¹¹SANOIA, Paris, France, ¹²SFD, Paris, France, ¹³University Paris 7, Lariboisiere Hospital, Paris, France, ¹⁴Dept. of Endocrinology, Diabetes, CHU Montpellier, Montpellier, France.

Background and aims: Automated insulin delivery (AID) systems show metabolic efficacy in clinical trials, but these often involve strict selection and monitoring, which may not mirror real-world conditions. This extensive nation-wide survey, named OB2F, was conducted to assess the initiation process, follow-up modalities, and the metabolic effectiveness of AID during real-life use, following six and twelve months of continuous usage.

Materials and methods: All adults who initiated AID between January 1st and December 31st, 2022, were included in 79 centers, regardless of the model used. The description of the follow-up modality at AID initiation by the OB2F participating centers was retrospectively collected through a questionnaire. Clinical data, continuous glucose monitoring parameters, and HbA1c were collected at AID initiation, after 6 and 12 months of follow-up. The median results [IQR] of patients with available data at initiation (n=2154), 6 months (n=1993), and 12 months (n=1143) are presented. Comparisons were conducted using Mann-Whitney tests and Chi-square tests depending on the type of variables.

Results: The characteristics of individuals at baseline are as follows: age 43 years [32-54], 42% male, BMI 25.4kg/m2 [22.8-29.3], diabetes duration 23 years [15-32]. 51.9% of the patients were initiated on the Medtronic 780G model, while the other half were started on the Tandem IQ. Loop closure occurred in an outpatient setting for 69% of patients and during inpatient stays for the remainder. Telemonitoring was implemented for 70% of patients. The organization and metabolic evolution at 6 and 12 months following the initiation of AID are presented in Table 1. Clinical characteristics and organizational modalities showed no significant impact on metabolic and clinical outcomes at 6 months and 1 year (p>0.05).

Conclusion: After 12 months of routine and regular follow-up, AID enables the majority of patients to achieve metabolic goals, particularly in time spent within target range, and significantly reduces severe hypoglycemic events with only a minority of patients discontinuing AID.

Supported by: SFD

Disclosure: J. Riveline: Employment/Consultancy; Sanofi, MSD, Eli Lilly, Novo Nordisk, AstraZeneca, Abbott, Dexcom, Alphadiab, Medtronic, Air Liquide, Adelia.

120

Impact on GRI and other metrics in type 1 adult patients switching to advanced hybrid closed-loop systems: a one-year real-life experience

E. Resmini, E. Zarra, S. Dotti, G. Rotondi, A.V. Cornaghi, S. Madaschi, E. Cimino, G. Massari, L.C. Pezzaioli, C. Buoso, M. Sandri, A. Girelli;

ASST Spedali Civili di Brescia, Brescia, Italy.

Background and aims: The introduction of Advanced Hybrid Closed-Loop system (AHCL) has profoundly changed type 1 diabetes therapy. This study primarily aimed to assess the impact on Glycemia Risk Index (GRI) and other continuous glucose monitoring (CGM) metrics when switching from one of four insulin strategies to AHCL in type 1 adult patients.

Materials and methods: A single-center, retrospective pre/post observational study; 198 patients (age 44.4±12.7years, 115females/83males, diabetes duration 24.7±11.6years, HbA1c 7.5±0.9%), treated with different insulin therapies (MDI, CSII, SAP with PLGS, HCL) were assessed before and after switching to an AHCL (MiniMed 780G, Diabeloop Roche, Tandem Control-IQ) at 1, 3, 6, and 12 months. Mixed-effects multivariable regression models were used to estimate the mean pre/post variations at different time points, adjusted for potential confounders.

Results: A month after the switch, there was a significant improvement in CGM metrics and HbA1c for all patients: GRI -10.7, GMI -0.27%, CV -2.1%, TAR>250 -3.7%, TAR180-250 -5.6%, TIR +9.7%, HbA1c -0.54% (all p<0.001). This improvement was maintained throughout the observational period (at 3, 6, and 12 months, with all p-values <0.001). When improvements across the 780, Diabeloop, and Tandem CIQ devices were compared: the Diabeloop demonstrated significantly better performance in terms of GRI, GMI, CV, TAR>250 at T1 (for all p<0.01); the 780 recorded the highest average decrease in TAR180-250 (p=0.020), while Tandem achieved the most significant reduction in TBR54-69 (p=0.004).

Conclusion: Adopting an AHCL leads to a rapid and sustained improvement in GRI and other parameters of metabolic control for up to a year, regardless of prior insulin therapies, baseline conditions or brands.

Disclosure: E. Resmini: None.

OP 21 Cardiovascular risk in diabetes: in search of the holy grail

121

Atherosclerotic cardiovascular disease in familial hypercholesterolemia and diabetes: a nationwide registry study

D. Eriksson Hogling, K. Littmann, G. Kindborg, J. Brinck;

Department of Medicine Huddinge, Karolinska Institutet, Stockholm, Sweden.

Background and aims: Familial hypercholesterolemia (FH) is a hereditary dyslipidemia associated with severely increased risk of premature atherosclerotic cardiovascular disease (ASCVD). Novel cholesterol lowering therapies and diabetes mellitus (DM) treatments are expected to decrease incident ASCVD in these populations. The aim of this study was to describe DM prevalence and ASCVD risk in patients with DM in a nationwide Swedish FH cohort.

Materials and methods: A novel national ICD code for FH (E78.0A) was introduced in Sweden on 01 Jan 2019, which enabled the identification of the Swedish FH cohort from the Swedish National Patient Registry (NPR). The NPR was used to extract data on ASCVD and DM (1987 until 2021). DM was defined as having received any DM diagnosis until the date of the E78.0A diagnosis. Date and cause of death was extracted from the Swedish cause of death register. Register data allowed calculation of hazard ratio (HR) with 95 % confidence intervals (CI) for coronary artery disease (CAD) and an ASCVD composite endpoint (CAD, ischemic stroke, peripheral artery disease and cardiovascular death) between the age of 30-70 years in patients born 1952-1987. HRs were adjusted for sex and hypertension. Part of the cohort was matched on age, sex and region with a population control group. Two sample T-test, Pearson´s Chi square test and Cox regression were used for statistical analyses.

Results: 2947 patients with FH were identified. Of these, 197 (6.7 %) had DM (any type). Mean age was 44.8 ± 21.3 years in patients without DM and 63.0 ± 14.5 years in patients with DM (P<0.001). 51.5 % (no DM) versus 52.8 % (DM) were women (P = 0.72). ASCVD was present in 23.5 % (no DM) versus 56.3 % (DM), P<0.001. Part of the cohort (n=2538) was individually matched on age, sex and region with a control group (n=25380). DM prevalence was 7.7 % in this matched FH cohort versus 4.3 % in controls (P <0.001). HRs were calculated in patients born 1952-1987 (n=111 with DM + 1508 without DM). HR for CAD was 1.58 (95% CI 1.21-2.06) and HR (adjusted) 1.47 (95% CI 1.10-1.98). HR for ASCVD was 1.57 (95 % CI 1.22-2.03) and HR (adjusted) 1.49 (95% CI 1.12-1.97), see Figure 1.

Conclusion: In a Swedish nationwide FH cohort DM prevalence was 6.7 %. DM prevalence was higher in the FH cohort than in the general population. ASCVD risk at 35-70 years of age was increased in patients with DM and FH compared to FH patients without DM, independent of age, sex and hypertension.

Supported by: The study was supported by an investigator initiated study grant from Amgen

Disclosure: D. Eriksson Hogling: Grants; The study was supported by an investigator initiated study grant from Amgen.

122

Association of plasma concentrations of MGO-apoB _3184-3194 , a signature peptide of glycated apoB100, with cardiovascular events in people living with type 2 diabetes

S. Hadjadj ^1,2, C. Chevalier^1,3, A. Rodrigues Oliveira^1,2, M. Wargny^1,4, L. Thenaisie⁴, V. Pakulska⁵, Y. Coute⁵, P.-J. Saulnier⁶, A. Pinheiro³, C. Le May¹, B. Cariou¹, M. Croyal^1,2;

¹UMR 1087, L’institut du Thorax, Nantes, France, ²M-shark, Nantes, France, ³SEBIA, Evry, France, ⁴Clinique des données, CHU de Nantes, INSERM CIC 1413, Pôle Hospitalo-Universitaire 11 : Santé Publique, Nantes, France, ⁵Université Grenoble Alpes, INSERM, UA13 BGE, CNRS, CEA, FR2048, Grenoble, France, ⁶CHU Poitiers, Poitiers, France.

Background and aims: Chronic hyperglycemia induces the non-enzymatic glycation of LDL particles, leading to an increase in their pro-atherogenic properties. Our proteomic analysis enabled the identification of a signature peptide of methylglyoxal (MGO)-mediated glycation of the LDL structural protein apolipoprotein B100 (apoB100): MGO-apoB_3184-3194. We assessed the association of plasma concentrations of MGO-apoB_3184-3194 and cardiovascular (CV) disease in people living with type 2 diabetes (pwT2D).

Materials and methods: Major CV events (MACE-3P: myocardial infarction (MI), stroke and/or CV death), coronary artery disease (CAD: MI and/or coronary artery revascularization) and CV death were collected in a prospective single-center cohort of 1425 pwT2D (58% men, age 66 ± 12 years, without end-stage renal disease). MGO-apoB_3184-3194 peptide was quantified in the plasma of patients at baseline by mass spectrometry after enzymatic proteolysis. The association between MGO-apoB_3184-3194 plasma concentrations and CV events was evaluated by Cox models and expressed as Hazard Ratio (HR) for an increment of one standard deviation in the concentration.

Results: After a median follow-up of 7 years, 380 cases of MACE-3P (37.8/1000 patient-year), 223 cases of CAD (22.8/1000 patient-year), and 285 CV deaths (27.1/1000 patient-year) have been reported. After multple adjustment on age, sex, HbA_1c, personal history of CV disease, estimated glomerular filtration rate, urine albumin/creatinine ratio, non-HDL cholesterol, MGO-apoB_3184-3194 concentrations were significantly associated with MACE-3P, CAD and CV death (adjusted HR MGO-apoB_3184-3194 = 1.27 [95% confidence interval (CI), 1.14-1.42]; p < 0.001); 1.25 [95% CI, 1.09-1.44]; p = 0.002) and 1.23 [95% CI, 1.09-1.39]; p = 0.001), respectively. These associations remained significant even after further adjustment on apoB100 concentration.

Conclusion: MGO-apoB_3184-3194, a signature peptide of MGO-mediated glycation of apoB100, is independently associated with the occurence of CV events in pwT2D, opening new perspectives for its use as a biomarker in clinical practice and epidemiological studies.

Supported by: ANR PRCE 2021

Disclosure: S. Hadjadj: Employment/Consultancy; Bayer, Mundipharma, Novo Nordisk, Valbiotis. Grants; AstraZeneca, Asten Santé, Air Liquide Health Science, Bayer, Boehringer Ingelheim, Eli Lilly, Isis Diabète LVL, Nestle Home Care, Novo Nordisk, Pierre Fabre, Sanofi, Valbiotis, Vitalaire. Honorarium; AstraZeneca, Bayer, Boehringer Ingelheim, Eli Lilly, Mundipharma, Novo Nordisk, Novartis, Sanofi, Servier, Valbiotis. Non-financial support; invitation to congress : AstraZeneca, Bayer Boehringer Ingelheim, Eli Lilly, Novo Nordisk, Sanofi, Valbiotis.

123

The importance of muscle health in type 2 diabetes: adverse muscle composition and all-cause mortality

J. Linge ¹, M. Petersson², O. Dahlqvist Leinhard¹;

¹AMRA Medical AB, Linköping, Sweden; Linköping University, Linköping, Sweden, ²AMRA Medical AB, Linköping, Sweden.

Background and aims: As pharmacological treatments for type 2 diabetes (T2D) and obesity are becoming more and more effective in achieving larger weight loss, the need to better understand aspects of muscle health in these groups is growing. Adverse muscle composition assessed by magnetic resonance imaging (MRI) has previously been linked to poor functional performance, comorbidity, and all-cause mortality in metabolic dysfunction-associated steatotic liver disease (MASLD) and general population. The aim of this study was to investigate the associations of all-cause mortality with adverse muscle composition within T2D.

Materials and methods: 56,109 UK Biobank participants were imaged using a 6-minute neck-to-knee protocol. Images were analyzed by AMRA® Researcher for quantification of thigh fat-free muscle volume and muscle fat infiltration (MFI), and calculation of a personalized sex-, weight-, and BMI-invariant muscle volume z-score (z-MV). Participants were partitioned according to muscle composition phenotype (adverse muscle composition, only high MFI, only low z-MV, normal muscle composition) using previously published cut-offs for high MFI (8.82/7.69% for females/males) and low z-MV (-0.68 standard deviations for both females and males). The associations of muscle composition with all-cause mortality were investigated using Kaplan-Meier survival curves and Cox proportional-hazard modelling with normal muscle composition as referent. Models were adjusted for low hand grip strength, sex, age, BMI, other diseases (cancer, coronary heart disease), and lifestyle factors (smoking, alcohol consumption, physical activity).

Results: 2,864 participants had T2D and complete data (33.7% women, mean (±SD) age 67.8 (±7.3) years and BMI 29.6 (±5.3) kg/m²). During a follow-up of 3.8 (±2.2) years and 10,897 person years at risk 92 deaths were recorded. Modelling showed that adverse muscle composition (27.9% prevalence) was significantly associated with all-cause mortality (hazard ratio [95% CI] 3.31 [1.92;5.69], p<0.01), but not only high MFI and only low z-MV (1.73 [0.94:3.17], p=0.079 and 1.15 [0.50;2.63], p=0.740) (Figure 1). Adverse muscle composition remained significant when including low hand grip strength (3.36 [1.92;5.87], p<0.01), sex, age, and BMI (2.44 [1.35;4.39], p<0.01), other diseases (2.50 [1.37;4.57], p<0.01), and lifestyle factors (2.37 [1.29;4.35], p<0.01). Other significant factors in the fully adjusted model were sex, age, cancer, smoking, and physical activity.

Conclusion: Adverse muscle composition was common in T2D and significantly associated with all-cause mortality. A significant association was found only when participants had both low muscle volume z-score and high muscle fat infiltration indicating the importance of assessing the amount of muscle fat when evaluating muscle health.

Disclosure: J. Linge: Employment/Consultancy; AMRA Medical AB (employment), Eli Lilly (consultation). Lecture/other fees; BioMarin Pharmaceuticals. Stock/Shareholding; AMRA Medical AB.

124

Reduction in ischaemic events with icosapent ethyl in patients with diabetes and prior CABG: REDUCE-IT diabetes-prior CABG

S. Verma ¹, D.L. Bhatt², P.G. Steg³, M. Miller⁴, E.A. Brinton⁵, T.A. Jacobson⁶, S.B. Ketchum⁷, A. Lira Pineda⁷, D. Mueller-Wieland⁸, J.-C. Tardif⁹, C.M. Ballantyne¹⁰, REDUCE-IT Invesitgators;

¹Cardiac Surgery, St. Michael’s Hospital, Toronto, ON, Canada, ²Mount Sinai Fuster Heart Hospital, New York, NY, USA, ³Université Paris-Cité, AP-HP, and INSERM, Paris, France, ⁴Univ. of Pennsylvania School of Medicine, Philadelphia, PA, USA, ⁵Utah Lipid Center, Salt Lake City, UT, USA, ⁶Emory Univ. School of Medicine, Atlanta, GA, USA, ⁷Amarin Pharma, Inc., Bridgewater, NJ, USA, ⁸RWTH Univ. Hospital Aachen, Aachen, Germany, ⁹Montreal Heart Institute, Montreal, QC, Canada, ¹⁰Baylor College of Medicine, Houston, TX, USA.

Background and aims: Patients with diabetes mellitus (DM) are at high cardiovascular (CV) risk. Atherosclerotic cardiovascular disease (ASCVD) is the principal cause of mortality among these patients. Significant residual ischemic risk remains after coronary artery bypass grafting (CABG) surgery. The effect of icosapent ethyl (IPE) on CV risk in patients with DM and history of CABG at baseline is unknown.

Materials and methods: In REDUCE-IT (Reduction of Cardiovascular Events With Icosapent Ethyl-Intervention Trial), a multicenter, placebo (PBO)-controlled, double-blind trial, statin-treated patients with controlled LDL-cholesterol and mild to moderate hypertriglyceridemia were randomized to 4 g daily IPE or PBO. They experienced a 25% reduction in risk of a primary efficacy endpoint (composite of CV death, myocardial infarction, stroke, coronary revascularization, or hospitalization for unstable angina) and a 26% reduction in risk of a key secondary efficacy endpoint (composite of CV death, myocardial infarction, or stroke) compared with PBO. This post hoc analysis reports on the subgroup of patients with DM and CABG history at baseline.

Results: Of 8179 patients randomized in REDUCE-IT, 5785 (70.7%) had established ASCVD, of whom 2401 (41.5%) had baseline DM. Of these, 799 (33.3%) had history of CABG, with 387 patients randomized to IPE and 412 to PBO. In the PBO arm, patients with history of CABG had numerically higher rates of ischemic events vs those without history of CABG. Randomization to IPE was associated with a significant reduction in the primary endpoint (HR 0.72; 95% CI 0.56-0.93; P=0.01) and in the key secondary endpoint (HR 0.59; 95% CI 0.43-0.80; P=0.0007) compared with PBO. Absolute risk reduction was 7.9% and 9.4% in first events with a number needed to treat of 13 and 11 for the primary and key secondary endpoints, respectively, during median follow-up of 4.6 years (Figure). Patients with established ASCVD and DM but without history of CABG experienced comparable risk reductions. In patients with established ASCVD, DM and history of CABG, hospitalizations for atrial fibrillation or flutter occurred in 5.4% (IPE) and 3.4% (PBO) of patients (P=0.24). Serious bleeding events occurred in 4.9% (IPE) and in 4.4% (PBO) of patients (P=0.74).

Conclusion: In REDUCE-IT patients with history of DM and CABG at baseline, IPE treatment was associated with significant reductions in ischemic events.

Clinical Trial Registration Number: NCT01492361

Supported by: The study was funded by Amarin Pharma, Inc.

Disclosure: S. Verma: Employment/Consultancy; Amgen (Ad Board), AstraZeneca (Ad Board, RCT), Bayer (Ad Board), Boehringer-Ingelheim (Ad Board, RCT), Eli Lilly Research (Ad Board, RCT), HLS Therapeutics (Ad Board), Janssen (Ad Board), Novartis (Ad Board), Novo Nordisk (Ad Board, RCT). Grants; Amarin, Amgen, AstraZeneca, Bayer, Boehringer-Ingelheim, Eli Lilly Research, HLS Therapeutics, Novo Nordisk, Pfizer, PhaseBio. Honorarium; Amarin, AstraZeneca, Bayer, Boehringer-Ingelheim, Canadian Medical and Surgical Knowledge Translation Research Group, Eli Lilly Research, HLS Therapeutics, Janssen, Novartis, Novo Nordisk, Pfizer, PhaseBio, S & L Solutions Event Management Inc., Sanofi. Other; Dr. Verma holds a Tier 1 Canada Research Chair in Cardiovascular Surgery, He is the President of the Canadian Medical and Surgical Knowledge Translation Research Group, a federally incorporated not-for-profit physician organization.

125

Risk prediction of type 2 diabetes complications using metabolomic data from the whole UK Biobank cohort

K. Schut, S. Kerminen, L. Jostins-Dean, P. Würtz, J.C. Barrett;

Nightingale Health Plc., Helsinki, Finland.

Background and aims: Individuals diagnosed with type 2 diabetes are at an increased risk of complications, such as eye problems, cardiovascular events and death. Accurate personalised risk predictions for such complications could help to prevent them. We used nuclear magnetic resonance (NMR)-based metabolomic data from half a million UK Biobank participants to train and test risk models predicting eye problems, myocardial infarction, stroke, heart failure, kidney disease and all-cause mortality among those diagnosed with type 2 diabetes.

Materials and methods: In our main analysis, we included individuals with a prevalent type 2 diabetes diagnosis without any of the complications assessed in this study present at the time of the blood sampling (N = 6015). We trained Cox proportional hazards models with LASSO regularisation, 10-fold cross-validation and 10 years of follow-up on age, sex, and 38 NMR-based metabolomic biomarkers in half of the dataset, as well as a model with age, sex and haemoglobin A1C (HbA1c).

Results: More than a third of the individuals in the test set were diagnosed with one or more complications within 10 years of follow-up (N = 1099). The metabolomic risk scores significantly predicted all types of complications independent from age and sex with HRs per 1-SD increment between 1.45 (95% CI=1.29-1.62; p=5.67E-10) for eye problems and 2.14 (95% CI=1.98-2.30; p=2.28E-86) for kidney disease (Figure 1A). The use of metabolomics offered improvements over clinically measured HbA1c for the prediction of heart failure, all-cause mortality and kidney disease. HRs for myocardial infarction, stroke, and eye problems were similar between the two models. Cumulative incidence rates within top risk deciles were consistently higher for the metabolomic models compared to the HbA1c models, with the most notable differences for kidney disease and all-cause mortality (Figure 1B).

Conclusion: Metabolomic data can be used to predict the incidence of different types of type 2 diabetes complications. Especially the metabolomic prediction of kidney disease and all-cause mortality offers possibilities beyond the use of HbA1c. The ability of metabolomic risk scores to predict a wide range of complications, could aid in implementation of prediction and prevention tools in diabetes care.

Disclosure: K. Schut: Employment/Consultancy; All authors are employees of Nightingale Health Plc. Stock/Shareholding; All authors hold stock (options) in Nightingale Health Plc.

126

Diabetes and trends in survival after acute myocardial infarction, 2011 to 2020

P.-S. Song ¹, M. Kim¹, S. Lee²;

¹Chungnam National University Hospital, Daejeon, Republic of Korea, ²Chonnam National University Hospital, Gwangju, Republic of Korea.

Background and aims: Patients with diabetes mellitus (DM) are at high risk for mortality after acute myocardial infarction (AMI). Despite an overall trend of reduced mortality after AMI, the mortality gap between AMI patients with and without DM did not decrease over time in previous analyses. We assessed the most recent trends in outcomes for patients with AMI according to DM status.

Materials and methods: We analyzed data from the 2 Korean Acute Myocardial Infarction Registries (KAMIRs), KAMIR-NIH (enrolled between November 2011 and December 2015, early period) and KAMIR-V (enrolled from January 2016 to June 2020, late period), contemporary registries of AMI patients treated in tertiary hospitals with facilities for primary percutaneous coronary intervention (PCI).

Results: The study comprised 28,990 AMI patients, including 10,128 (34.9%) with DM. The prevalence of DM in the study cohort tends to be increased from 34.5% in early period to 35.3% in late period (adjusted odd ratios [OR], 1.045; p=0.079 for trend). The provision of secondary prevention agents rose significantly in both groups between early and late period; Reperfusion with PCI rates also increased over the study period in both groups. Thanks to advances in evidence-based therapies and their delivery between 2011 and 2020, impressive declines in in-hospital mortality occurred in those with (from 5.2% in early period to 4.0% in late period; adjusted OR, 0.748; p=0.003 for trend) and without DM (from 3.5% in early period to 2.9% in late period; adjusted OR, 0.794; p=0.006 for trend). These improvements were sustained at 12 months in both groups (from 12.3% in early period to 7.8% in late period; adjusted OR, 0.628; p<0.001 for trend in patients with diabetes; from 7.7% in early period to 5.1% in late period; adjusted OR, 0.659; p<0.001 for trend in patients without diabetes).

Conclusion: The in-hospital and 12-months mortality gaps between AMI patients with and without DM narrowed significantly from 2011 to 2020. These data suggest that despite the continually increasing prevalence of DM and the persistent incremental cardiovascular risk associated with DM, sustained efforts at implementing evidence-based therapies and improving the care afforded to these patients can help close the gap in adverse cardiovascular outcomes in patients with DM.

Disclosure: P. Song: None.

OP 22 Actions of incretins on the brain and pancreas you might not know about

127

Single-dose GLP-1-based pancreatic gene therapy durably maintains body composition and glycaemia after semaglutide withdrawal in a murine model of obesity

A.L. Fitzpatrick ¹, S. Wang¹, R. Reese¹, N. Picard¹, E. Cozzi¹, T. Kieffer^1,2, J. Caplan¹, H. Rajagopalan¹;

¹Fractyl Health, Inc., Burlington, MA, USA, ²University of British Columbia, Vancouver, BC, Canada.

Background and aims: Metabolic diseases, including obesity and type 2 diabetes, have risen dramatically and are attributed to the chronic overnutrition and sedentary lifestyle of modern society. Glucagon-like peptide 1 (GLP-1) therapies have proven to be effective for metabolic disease; however, treatment cessation often leads to rapid weight regain and glucose rebound. We have developed a novel, adeno-associated virus, pancreatic gene therapy (PGTx) platform enabling durable islet production of GLP-1-based peptides. Here, we assessed body composition and glycaemic parameters of single-dose PGTx versus chronic semaglutide (Sema) in a murine diet-induced obesity model.

Materials and methods: C57BL/6 mice were fed a 60% high-fat diet and were randomized to: single-dose intraperitoneal PGTx (1e13 vector genomes [VG], n=10), daily subcutaneous Sema (10 nmol/kg/d x 4 weeks, n=10), PGTx vehicle control (VC) (n=8), or Sema VC (n=8). Sema was withdrawn at week 4, and mice were randomized to PGTx (5e12 VG, n=5) or VC (n=5). Mean body composition and fasting glucose were assessed. A 0-5-point histopathology score was used to evaluate islet inflammation and integrity. Statistical comparisons were made to cohort baselines.

Results: At week 4, PGTx reduced fat mass by 21% and lean mass by 5% versus 16% fat mass and 5% lean mass with Sema (all, p<0.0001). Sema withdrawal led to a fat and lean mass regain to 1% and 2% below baseline, while Sema-withdrawn mice treated with PGTx maintained reductions of 17% (p<0.01) and 5% (p<0.001) at week 8. In PGTx and Sema groups, fasting glucose reduced by 18% at week 4 (both, p<0.0001). Sema-withdrawal resulted in fasting glucose rebound to baseline, while PGTx and PGTx-treated Sema-withdrawn mice maintained 21% (p<0.0001) and 22% (p<0.001) fasting glucose reductions at 8 weeks. No significant evidence of pancreatic inflammation was observed with mean islet histopathology scores of <0.5 in all groups evaluated.

Conclusion: Single-dose PGTx can durably improve fat mass and glycaemia and maintain these metabolic parameters upon Sema withdrawal without causing pancreatic inflammation. The data indicate that PGTx has the potential to advance GLP-1-based therapies toward durable efficacy in obesity and type 2 diabetes.

Disclosure: A.L. Fitzpatrick: Employment/Consultancy; Fractyl Health, Inc. Stock/Shareholding; Fractyl Health, Inc.

128

Brain circuitry activated by cagrilintide, semaglutide and CagriSema in mice

T.A. Lutz ¹, G. Lommi¹, C. Le Foll¹, A. Secher², K. Raun², M. Hankir¹;

¹University of Zurich, Zurich, Switzerland, ²Novo Nordisk, Måløv, Denmark.

Background and aims: Cagrilintide is a long acting, dual amylin and calcitonin receptor agonist (DACRA) that induces weight loss in people living with obesity. The combination of cagrilintide and the long-acting glucagon-like peptide 1 receptor (GLP-1R) agonist, semaglutide (CagriSema) is currently in clinical development for treatment of obesity and type-2-diabetes. The brain pathways involved in energy intake have been extensively studied for semaglutide but the brain feeding circuitry targeted by cagrilintide, and their combination CagriSema is poorly understood.

Materials and methods: We therefore assessed the impact of these drugs on food intake and neuronal activity in 2 separate cohorts of mice by performing acute feeding experiments in an automated feeding system and brain cFos immunohistochemistry analysis, respectively.

Results: In lean chow-fed wild-type C57BL/6 mice, intraperitoneal (I.P.) injection of cagrilintide (10 nmol/kg, n=6) reduced food intake compared to vehicle (n=6) during the dark phase (1.8 ± 0.2g vs. 2.5 ± 0.5g; p=0.004); semaglutide (10 nmol/kg, n=6) had a more potent effect (0.8 ± 0.3g) compared to cagrilintide and that was similar to CagriSema (2 nmol/kg + 2 nmol/kg, n=6; 0.7 ± 0.2g). Hence, I.P. injection of CagriSema at a five-fold lower dose than each drug alone caused an equipotent suppression of food intake to semaglutide. cFos immunohistochemistry analysis of brains from lean chow-fed calcitonin receptor-Cre tandem dimer-Tomato (CalcR-tdTomato) reporter mice that received acute subcutaneous injections of vehicle (n=8), cagrilintide (10nmol/kg, n=7), semaglutide (10nmol/kg, n=8) or CagriSema (2nmol/kg + 2nmol/kg, n=8) revealed strong activation of CalcR-expressing neurons in the hindbrain area postrema (AP) and nucleus tractus solitarius (NTS) for all drugs, whereas only cagrilintide and semaglutide activated downstream projection sites in the lateral parabrachial nucleus (lPBN), a brain region involved in the development of aversion. Notably, semaglutide and CagriSema also activated neurons in the arcuate nucleus of the hypothalamus (ARC).

Conclusion: Our findings shed light on the brain feeding circuitry activated by cagrilintide, semaglutide and CagriSema, and suggest that CagriSema achieves potent appetite suppression at much lower doses than treatments with single compounds. Future work will identify the exact neuronal pathways and molecular identity of the neurons activated by these drugs.

Supported by: Swiss National Science Foundation; Novo Nordisk research grant

Disclosure: T.A. Lutz: Grants; Research collaboration with Novo Nordisk.

129

A trans-ancestry GLP1R candidate locus study for mental and physical health

M.M.E. Hayman, R.J. Strawbridge, P. Welsh;

University of Glasgow, Glasgow, UK.

Background and aims: Metabolic diseases and cardiovascular complications are a leading cause of premature mortality in individuals with mental ill-health (MIH). Recent studies show the bi-directional associations between mental and physical illness, however the identification of the shared pathways have yet to be established. Understanding shared pathways would improve the management of both mental and physical symptoms and reduce cardiometabolic risk. Observationally, glucagon-like peptide 1 receptor agonists (GLP1RA), used for both type 2 diabetes and obesity management, have shown some evidence of off-target behavioural effects such as smoking cessation. Similarly, early evidence from murine studies demonstrated that with the use GLP1RA was associated with decreased depressive and anxious symptoms and might promote reversal of cognitive defects. A genetic association study focused on the GLP1R locus might ameliorate the understanding of potential long-term impacts of GLP1RA in humans. Therefore, we carried out a candidate locus study to better define the genetic architecture of the GLP1R locus with respect to mental ill-health and cardiometabolic disease using UK Biobank.

Materials and methods: Genetic variation in the GLP1R locus (encoding gene ±500kb) was identified and analysed for impact on a range of MIH [e.g., addiction, anhedonia, Bipolar Disorder, Generalised Anxiety Disorder, Major Depressive Disorder, mood instability, neuroticism score, and risk-taking behaviour] and cardiometabolic traits, using PLINK 1.9. Analyses were conducted separately in unrelated individuals of self-reported white British ancestry (n=408,774), European ancestry (n=50,314), South Asian ancestry (n=7,667), African-Caribbean ancestry (n=7,641), and multiple ancestry (n=10,437). Bonferroni correction based on the number of independent variants was used to define the significance threshold for each ancestry. All ancestries were subsequently assessed in an inverse variance-weighted fixed effects meta-analysis. Cross-trait conditional analyses was used to determine the interdependence of signals for MIH and cardiometabolic disease and expression data was used to investigate specificity of GLP1R effects and causality. Follow up in-silico analysis was done using the Genotype-Tissue Expression (GTEx) Portal.

Results: Twenty-one signals were associated with cardiometabolic (N_signal = 15) or MIH (N_signal = 6) traits. Of these signals, eight were in white British individuals, the largest sample. Cardiometabolic associations were observed in all the ancestries, while all but one ancestry had MIH associations. The trans-ancestry meta-analysis showed mainly concordant effect directions for blood pressure, type 2 diabetes, and body mass index. The signal associated with risk-taking behaviour showed discordant directionality but requires further validation in larger non- European cohorts. All signals associated with MIH and cardiometabolic phenotypes were conditionally independent. Follow up in-silico analysis showed that of the observed signals available in GTEx (N_signal =10) none of the MIH endophenotypes were within the coding region of GLP1R, while three cardiometabolic signals were.

Conclusion: MIH associated signals were located outside the coding region of GLP1R and we found no evidence of these signals influencing GLP1R gene expression, suggesting that these signals act via non-GLP1R mechanisms. It is therefore likely that observed behavioural changes with GLP1RA are not acting through GLP1R.

Disclosure: M.M.E. Hayman: None.

130

Assessing GLP1 receptor occupancy in the CNS and pancreas using PET

A. Khalil ¹, I. Velikyan², M. Bossart³, M. Wagner⁴, O. Eriksson⁵;

¹Department of Medicinal Chemistry, Uppsala University, Uppsala, Sweden, ²Department of Surgical Sciences, Uppsala University, Uppsala, Sweden, ³⁴Germany, ⁵Department of Medicinal Chemistry, Translational PET Imaging, Uppsala University, Uppsala, Sweden.

Background and aims: GLP-1R agonism, either alone or as part of multireceptor agents, has demonstrated significant efficacy in managing type 2 diabetes and obesity. GLP-1R agonists have been shown to have both peripheral and centrally mediated effects in preclinical models. However, there is a lack of tissue-specific biomarkers of GLP-1R target engagement for clinical application. This study used [⁶⁸Ga]Ga-DO3A-exendin-4 PET imaging to non-invasively quantify the GLP-1R occupancy in brain regions and pancreas of SAR441255 (triple GLP1/GIP/GCG receptor agonist) and tirzepatide (dual GLP-1R/GIPR agonist).

Materials and methods: Healthy piglets (n=9) were injected with 0.5-1 MBq/kg of [⁶⁸Ga]Ga-DO3A-exendin-4 and scanned by PET/CT over the entire body including pancreas and brain. Each pig was then administered a GLP-1R agonist at variable doses: 1-100 μg/kg of SAR441255 (n=5), 50-450 μg/kg of tirzepatide (n=3), or 6 μg/kg of unlabeled Exendin-4 (n=1, used as positive control). Another [⁶⁸Ga]Ga-DO3A-exendin-4 PET examination was then performed after 2.5 hours to measure the GLP-1R blockade induced by each drug at this time point. The tracer uptake was assessed as a standardized uptake value (SUV), and occupancy was defined as the decrease in tracer GLP-1R binding compared to baseline. An in vitro autoradiography assay was performed on frozen sections of cell pellets to confirm [⁶⁸Ga]Ga-DO3A-exendin-4 specificity to GLP-1R, with and without co-incubation using SAR441255 or tirzepatide as blocking agents.

Results: [⁶⁸Ga]Ga-DO3A-exendin-4 demonstrated high binding in the pancreas as well as defined brain regions and circumventricular organs including the medial hypothalamic area (which incorporates the arcuate nucleus), mamillary body, and pituitary. SAR441255 demonstrated a strong dose-dependent GLP-1R occupancy (>70% at the highest dose) in e.g. pancreas, pituitary, and mamillary body and ~60% occupancy in the medial hypothalamic area. Tirzepatide demonstrated lower GLP-1R occupancy at comparable doses in most tissues, presumably as it had not reached C_max. The in vitro autoradiography assay confirmed that both SAR441255 and tirzepatide compete with [⁶⁸Ga]Ga-DO3A-exendin-4 for the same binding site of GLP-1R.

Conclusion: We demonstrate in vivo target engagement of SAR441255 and tirzepatide at the GLP-1R using PET in a pig model. Dose-dependent GLP-1R drug occupancy was seen in both pancreas and known appetite centers in the brain. This PET technology can readily be translated for clinical studies of GLP-1R target engagement and occupancy.

Disclosure: A. Khalil: None.

131

Retatrutide, an agonist of GIP, GLP-1, and glucagon receptors, improves markers of pancreatic beta cell function and insulin sensitivity

M.K. Thomas ¹, J. Rosenstock², T. Coskun¹, M.L. Hartman¹, J. Lou¹, Q. Wu¹, Y. Du¹, S. Gurbuz¹, K. Mather¹, Z. Milicevic¹;

¹Eli Lilly and Company, Indianapolis, IN, USA, ²Velocity Clinical Research at Medical City, Dallas, TX, USA.

Background and aims: Retatrutide (RETA), an agonist of GIP, GLP-1 and glucagon receptors, significantly reduced HbA1c up to 2.2% (0.5 mmol/mol) and body weight up to 17% at 36 weeks in a Phase 2 trial in participants with Type 2 diabetes (T2D). In a second Phase 2 trial in participants with obesity or overweight but without T2D (OB), RETA significantly reduced HbA1c by up to 0.5% (-18.1 mmol/mol) and body weight by up to 24% at 48 weeks. To explore mechanisms by which RETA improved glycemic control, we assessed markers of pancreatic beta-cell function and insulin sensitivity.

Materials and methods: Mixed models for repeated measures evaluated fasting biomarkers from these two Phase 2 double-blind randomized placebo-controlled trials: T2D (281 subjects, 36 weeks, RETA 0.5, 4, 8, and 12 mg vs placebo and 1.5 mg dulaglutide) and in OB (338 subjects, 48 weeks, RETA 1, 4, 8, 12 mg vs placebo).

Results: Fasting glucose and insulin levels decreased over time from baseline with higher doses of RETA in T2D and OB. Homeostatic model assessment (HOMA2)-IR index (computed with insulin decreased over time from baseline with RETA 12 mg reaching reductions of 39% in T2D at 36 weeks and 52% in OB at 48 weeks, indicative of reductions in insulin resistance in both populations. Adiponectin, a marker of insulin sensitivity, increased with RETA from baseline up to 52% in T2D and up to 70% in OB (p<0.001). HOMA2-B index (computed with C-peptide), a marker of beta-cell function, rapidly increased with RETA up to 88% from baseline in T2D but did not significantly increase in OB. Proinsulin levels and proinsulin/C-peptide ratios, measures of beta-cell stress and dysfunction, decreased from baseline with RETA, by up to 71% and 62%, respectively, in T2D (p<0.001). Proinsulin levels also decreased from baseline with RETA in OB.

Conclusion: Enhanced glycemic control with RETA was associated with improvements in markers of both beta-cell function and insulin sensitivity in T2D and OB.

Disclosure: Previously accepted at ADA 2024.

Clinical Trial Registration Number: NCT04867785

Disclosure: M.K. Thomas: Employment/Consultancy; Eli Lilly and Company. Stock/Shareholding; Eli Lilly and Company.

132

Direct contact between endosomal GLP-1R, ER VAP-B and the AKAP SPHKAP triggers PKA-dependent MIC19 phosphorylation and beta cell mitochondrial remodelling

A. Tomas ¹, G. Austin¹, L. ElEid¹, A. Oqua¹, Y. Manchanda¹, Y. Poliakova², A. Montoya³, B. Jones¹, S. Millership¹, I. Prokopenko⁴;

¹Department of Metabolism, Digestion and Reproduction, Imperial College London, London, UK, ²School of Public Health, Imperial College London, London, UK, ³MRC Laboratory of Medical Sciences, London, UK, ⁴University of Surrey, Guildford, UK.

Background and aims: GLP-1R agonists (GLP-1RAs) ameliorate mitochondrial health by increasing its turnover and improving its quality control. While the GLP-1R is well known to stimulate cAMP production leading to activation of Protein Kinase A (PKA) and Exchange Protein Activated by cyclic AMP 2 (Epac2) signalling, there is a lack of understanding of the molecular mechanisms linking GLP-1RA-induced signalling with mitochondrial remodelling and improved mitochondrial function.

Materials and methods: Co-immunoprecipitation, LC-MS/MS, siRNA, confocal microscopy, insulin secretion assay, nanoBRET assay, subcellular PKA signalling with localised biosensors, mitochondrial function and remodelling assays, TUNEL assay, proximity ligation assay, GWAS meta-analysis.

Results: Here we show that active GLP-1R triggers localised mitochondrial cAMP/PKA signalling in pancreatic β-cells at membrane contact sites (MCSs) between endosomes, ER and outer mitochondrial membranes, due to direct GLP-1R interaction with ER-resident VAP-B and mitochondria-localised A-kinase anchoring protein (AKAP) Sphkap, a protein for which we demonstrate that genetic variants are associated with increased prevalence of type 2 diabetes and higher BMI as well as increased random glucose levels. Interaction between active GLP-1R and VAP-B requires GLP-1R internalisation and is differentially regulated by biased GLP-1R agonists associated with distinct trafficking profiles. Recruitment of Sphkap to GLP-1Rs at the ER-mitochondrial interface requires prior association of Sphkap with VAP-B via its VAP-B-binding phospho-FFAT domain. Localised mitochondrial PKA signalling in response to GLP-1R activation requires Sphkap, with disruption of this pathway associated with loss of GLP-1R-induced PKA-dependent phosphorylation of the mitochondrial contact site and cristae organizing system (MICOS) complex component MIC19. This phenotype is accompanied by defective mitochondrial remodelling, loss of GLP-1R-induced membrane potential and ATP increases and reduced GLP-1R-induced potentiation of insulin secretion (4.9 +/- 0.5 versus 3.1 +/- 0.3-fold in Control versus Sphkap RNAi-treated INS-1 832/3 cells) and β-cell survival under ER-stress (79 +/- 7% versus 56 +/- 3% survival).

Conclusion: This body of work demonstrates that, following GLP-1RA stimulation in pancreatic β-cells, the GLP-1R interacts with the ER MCS organising factor VAP-B from an endocytic location to engage SPHKAP, an AKAP linked to T2D and obesity in genome-wide association studies (GWAS), triggering a pool of mitochondrially localised PKA signalling that phosphorylates MIC19, and enables GLP-1RA-induced mitochondrial remodelling and optimal β-cell function.

Disclosure: A. Tomas: None.

OP 23 Why exercise?

133

Effects of exercise intensity on the insulin sensitivity and lipid composition in skeletal muscle under the same calorie consumption exercise

S. Kakehi, Y. Tamura, T. Funayama, R. Kawamori, H. Watada;

Juntendo University, Graduate School of Medicine, Tokyo, Japan.

Background and aims: Intramyocellular lipids (IMCL), altered by physical exercises, are associated with insulin sensitivity in skeletal muscle. However, the relationship between changes in IMCL composition due to exercise intensity and insulin sensitivity, along with the underlying molecular mechanisms, is not fully understood. The aim of this study is to investigate the impact of exercise intensity on skeletal muscle lipid composition differences in improving insulin sensitivity through exercise.

Materials and methods: We recruited 20 Japanese men with moderate insulin resistance (HOMA-IR≥1.6) and assigned them to a low-intensity exercise (LIE) (40% VO2 peak, n=11) or high-intensity exercise (HIE) (70% VO2 peak, n=9) group. Participants engaged in their assigned exercise protocols for five days, targeting 300 kcal/day of energy expenditure. Insulin resistance was assessed using hyperinsulinemic euglycemic clamps, and IMCL levels in the tibialis anterior (TA-IMCL) were measured before and after the exercise program. Muscle biopsy samples were analyzed for lipidomics using liquid chromatography-mass spectrometry and for skeletal muscle gene expression using RNAseq.

Results: Exercise induced differential changes in TA-IMCL levels between LIE and HIE groups (LIE -21%, HIE +55.0%), with a significant reduction observed in the LIE group. Both exercise intensities similarly increased insulin sensitivity. Notably, exercise affected skeletal muscle lipid profiles, with phospholipid species undergoing significant changes with exercise intensity. Specifically, lysophosphatidic acid (22:6), phosphatidylcholine (40:6) and phosphatidylserine (40:6) including docosahexaenoic acid (22:6), and some phosphatidylinositol (PI) and cardiolipin (CL) were significantly increased in the HIE group. We also found that many lysophospholipids were increased in the LIE group. The results of the orthogonal partial least squares discriminant analysis showed that PIs and CLs, changed by HIE, were extracted as lipids contributing to the difference in exercise intensity. Moreover, accompanying changes in lipid composition, the expression of Phospholipase A2 Group IV (PLA2G4), involved in lysophospholipid production in LIE, and ELOVL Fatty Acid Elongase 2 (ELOVL2) and Fatty Acid Desaturase 1 (FADS1), involved in docosahexaenoic acid biosynthesis in HIE, were altered.

Conclusion: These suggested that exercise intensity contributes to the improvement of insulin sensitivity in skeletal muscle through the regulation of different lipid compositions.

Disclosure: S. Kakehi: None.

134

Myofiber Piezo1 regulates muscle hypertrophy and metabolism in response to exercise

T. Inoue ¹, Y. Hirata¹, K. Nomura^1,2, N. Kuramoto¹, T. Nishigaki¹, K. Hozumi¹, K. Sugawara¹, W. Ogawa¹;

¹Division of Diabetes and Endocrinology, Department of Internal Medicine, Kobe University Graduate School of Medicine, Kobe, Japan, ²Department of Nutrition and Metabolism, Institute of Biomedical Sciences, Tokushima University Graduate School, Tokushima, Japan.

Background and aims: Skeletal muscle, the largest insulin target organ in the human body, plays a crucial role in whole-body metabolism. While exercise exerts a crucial influence on both the quantity and quality of muscle, its underlying mechanism remains not fully understood. We have previously revealed that the downregulation of the mechanosensory Ca²⁺ channel Piezo1 is an essential step of muscle atrophy during immobilization. The current study aims to explore the role of Piezo1 in exercise-induced remodeling of muscle.

Materials and methods: Piezo1 disruption was achieved by administering tamoxifen to inducible myofiber-specific Piezo1 knockout mice, generated by crossing Piezo1-flox and HSA-Cre-ERT2 mice. As a model of muscle hypertrophy induced by exercise, synergistic muscle ablation, involving the resection of the gastrocnemius and soleus muscles, was performed. 10 days after the resection of these muscles, the plantaris muscle, which underwent compensatory hypertrophy, was isolated and subjected to various assays. Voluntary exercise was facilitated by housing mice with running wheels for 3 months, with tamoxifen administered every 3 weeks to maintain continuous Piezo1 disruption.

Results: Within 10 days of synergistic muscle ablation, the mass and Piezo1 expression in plantaris muscles increased approximately 1.6- and 2-fold, respectively. DNA microarray analysis revealed Piezo1 as one of the most prominently upregulated genes among those encoding membrane Ca²⁺ channels during this process. Myofiber-specific Piezo1 disruption inhibited muscle hypertrophy induced by synergistic muscle ablation by approximately 40%, without affecting the alteration of genes involved in hypertrophy. However, it suppressed the phosphorylation of S6K/S6 proteins, markers of protein synthesis. In C2C12 myotubes, the Piezo1 agonist Yoda1 increased S6K/S6 phosphorylation, which was inhibited by a CaMKK inhibitor. Mice housed with running wheels showed upregulation of mitochondria-related genes, including Ppargc1a and Nr4a3, in muscle; this upregulation was hindered by myofiber-specific Piezo1 disruption. In C2C12 myotubes, Yoda1 synergistically upregulated both Ppargc1a and Nr4a3 with the β2-adrenergic agonist clenbuterol.

Conclusion: Our results indicate that Piezo1 in myofibers contributes to exercise-induced muscle hypertrophy likely via stimulating protein synthesis. Additionally, Piezo1 is involved in the regulation of genes related to muscle metabolism during aerobic exercise. Thus, Piezo1 plays a multifaceted role in skeletal muscle during exercise, suggesting its potential as a therapeutic target for muscle-related disorders.

Supported by: Japan Agency for Medical Research and Development under grant number 22ek0210159h0002

Disclosure: T. Inoue: None.

135

Exercise-induced exosomes contribute to the sustained metabolic changes after cessation of exercise training in insulin resistance states

L. Mastrototaro ¹, M. Apostolopoulou¹, S. Hartwig¹, K. Strassburger¹, P. Lipaeva¹, N. Trinks¹, Y. Karusheva¹, S. Gancheva¹, S. Trenkamp¹, S. Lehr¹, H. Al-Hasani¹, J. Szendroedi², M. Roden¹;

¹German Diabetes Center, Düsseldorf, Germany, ²Department of Medicine I and Clinical Chemistry, University Hospital of Heidelberg, Heidelberg, Germany.

Background and aims: High-intensity interval training (HIIT) not only consistently enhances whole-body oxidative capacity (VO₂max) in humans regardless of their insulin sensitivity, but also increases insulin sensitivity and induces release of small extracellular vesicles (sEV; exosomes) in insulin resistant individuals with and without type 2 diabetes. However, it remains unclear whether the HIIT-induced metabolic changes persist similarly among individuals with varying degrees of glucose tolerance after the cessation of the HIIT intervention and what are the potential underlying mechanisms.

Materials and methods: Individuals with type 2 diabetes (T2D: n=20) as well as age- and BMI-matched glucose tolerant humans (insulin resistant, IR: n=10; insulin sensitive, IS: n=12) performed a 12-wk supervised cycle ergometer training protocol and a subsequent 4-wk detraining period. To assess tissue-specific insulin sensitivity, VO₂max and ectopic lipid content, hyperinsulinemic-euglycemic clamps, spiroergometry, and magnetic resonance spectroscopy were performed in all participants before intervention (baseline), after 12-wk HIIT, and after 4-wk detraining. Finally, sEV were isolated from serum by size exclusion chromatography and characterized for number and size as well as proteomic profile using nanoparticle tracking analysis and mass spectrometry, respectively.

Results: After detraining, VO₂max declined in T2D (p<0.001) and IR (p<0.01) as compared to HIIT, but remained higher in all groups compared to baseline (p<0.001 and p<0.05 in T2D and glucose tolerant individuals, respectively). Of note, HIIT-induced improvements in hepatic insulin sensitivity and lipid content were maintained in T2D and IR (p<0.01 and p<0.05 vs baseline, respectively), whereas M-value returned to baseline values only in T2D (p<0.05 vs HIIT). Following detraining, T2D and IR demonstrated a sustained rise (p<0.01 and p<0.05 vs baseline, respectively) in the number of circulating sEV, which was positively associated with visceral fat volume in T2D (β=0.13, p=0.06). Proteomic analysis identified 1114 proteins in the isolated sEV, of which 186 proteins exhibited differential abundance across the three groups following detraining, with an upregulation of proteins associated with inflammation (myeloperoxidase, MPO; annexin 1/3) and oxidative stress (glutathione peroxidase 3; peroxiredoxin 1/2) in sEV isolated from T2D and IR as compared to IS. Finally, correlation analysis revealed that specific sEV proteins associate with metabolic parameters in T2D (e. g. vanin2 with M-value, r=-0.945, p=0.0150) and IR (e. g. MPO with BMI, r=0.993, p=0.0006).

Conclusion: This study shows that sEV proteins released upon HIIT at least partly account for the changes in whole-body oxidative capacity and insulin sensitivity observed after detraining in insulin resistant individuals irrespective of diabetes, highlighting the role of sEV for exercise-induced inter-organ crosstalk.

Clinical Trial Registration Number: NCT02039934

Disclosure: L. Mastrototaro: None.

136

Effects of light-intensity physical activity snacking on cardiometabolic parameters in young adults with overweight/obesity: a randomised controlled crossover trial

S. Sanfilippo ¹, O. Moser^1,2, P. Zimmermann¹, J. Schierbauer¹, A. Grothoff¹, N. Wachsmuth¹, T. Voit¹, A. Rössler², H.K. Lackner², S. Hoffmann¹;

¹University of Bayreuth, Bayreuth, Germany, ²Medical University of Graz, Graz, Austria.

Background and aims: Sedentary behaviour (SB) is an essential risk factor for overweight and obesity, cardiovascular disease, and type 2 diabetes (T2D). Though certain levels of physical activity (PA) may attenuate detrimental effects of SB, cardiometabolic responses involved are still not fully understood.

Materials and methods: We conducted a four-armed randomized controlled crossover trial investigating the effects of prolonged sitting and interrupted prolonged sitting with PA on cardiometabolic parameters in overweight/obese young adults. Participants were recruited in an university setting through notices (digital or paper form; also, in courses and lectures) and randomly assigned to the following conditions which were tested during an 8-hrs simulated workday: prolonged sitting (SIT), sitting interrupted with standing (SIT-STAND, accumulated standing time: 2.5 hrs per day), continuous standing (STAND), and continuous walking (1.0 mph/1.61 km/h; WALK). Interventions were conducted at the research lab and for all trial days, participants received a standardized breakfast and lunch. The primary outcome was to investigate how changes in body position (alternate sitting and standing, continuous standing or walking) affect the 8-hrs mean glucose curve compared to uninterrupted sitting. Secondary outcomes include the effects on postprandial glucose concentration as well as 8-hrs/24-hrs heart rate (HR) and heart rate variability (HRV) parameters regarding the respective study arm. Glucose concentrations were measured hourly during the visits. HR and HRV were assessed for 24-hrs including the 8-hrs intervention phase, and a home-phase by means of Holter electrocardiogram (ECG). Data were analysed by means of ANOVA with adjusted post-hoc testing (p<0.05).

Results: Seventeen individuals (eight females, mean age 23.4 ± 3.3 years, BMI 29.7 ± 3.8 kg/m²) completed all four trial arms. Compared with SIT (89.4 ± 6.8 mg/dL), 8-hrs mean glucose was lower in all other conditions (p<0.05), and reached highest statistical significance compared to WALK (86.3 ± 5.2 mg/dL, p=0.034). 2-hrs postprandial mean glucose after breakfast was ~7% lower for WALK (88.5 ± 8.6 mg/dL; 4.9 ± 0.5 mmol/L) compared with SIT (95.3 ± 9.2 mg/dL; 5.3 ± 0.5 mmol/L; p=0.002). Furthermore, significant time x condition effects for HR and HRV parameters favouring light-intensity walking were observed (p<0.001).

Conclusion: In overweight and obese young adults light-intensity PA snacking (walking) showed a significant blood glucose-lowering effect and improved HRV during an 8-hrs work environment compared to prolonged sitting, alternate sitting and standing, as well as compared to continuous standing.

Clinical Trial Registration Number: DRKS00031425

Disclosure: S. Sanfilippo: None.

137

The response in insulin sensitivity to different modalities of acute exercise

N.S. Nielsen ¹, J.O. Hvidemose¹, R. Aarvig¹, N.G. Skougaard¹, J.P. Hartmann¹, K.S. Husted², R.J.F. Loos², R.M.G. Berg¹, M. Ried-Larsen¹;

¹Centre for Physical Activity Research (CFAS), Copenhagen, Denmark, ²University of Copenhagen, Copenhagen, Denmark.

Background and aims: It is well established that insulin sensitivity is upregulated in the hours and days following exercise. However, most observations rely on clamp procedures. No studies have investigated the time-course of the glucose disposal following a single acute exercise bout, across various exercise types, utilizing a test where glucose homeostatic mechanisms remain intact, such as the OGTT. Thus, we aim to test the time-course of increased insulin sensitivity in the days after an acute bout of different exercise modalities. Additionally, we aim to assess inter- and intraindividual variability in insulin sensitivity.

Materials and methods: In a randomized cross-over study, physically inactive participants with BMI≥25 kg/m² and >40 years were recruited. The participants underwent three sets of experiments in a random order, each including a 50 min exercise bout of either 1) continuous aerobic exercise (64-76% of the maximal heart rate (HR_max)), 2) high-intensity interval training (HIIT; 5x4 min at >85% HR_max) or 3) strength training (3 sets with 10 repetitions of 70% of 1 repetition max in 4 exercises). Each experiment comprised an exercise bout followed by a 2-hour OGTT 24 and 48 hours post-exercise and was repeated the subsequent week. There was a one week between each exercise bout, resulting in 19 study days across 7 weeks. The primary outcome was change in Matsuda index (increment indicate greater insulin sensitivity). These preliminary data are descriptive with no statistical analysis conducted.

Results: We randomized 24 participants (58% female) with a mean (SD) age of 54 (8) years and BMI of 32 (4) kg/m². The baseline median (Q1-Q3) insulin sensitivity was 3.3 (2.8-4.4) arbitrary units (A.U.). At 24h and 48h following the acute exercise bout, the median (Q1-Q3) Matsuda index was 3.6 (2.5-4.6) and 4.0 (2.9-4.7), after continuous aerobic exercise, 3.8 (2.7-4.9) and 3.9 (2.7-5.0) after HIIT, and 3.6 (2.6-5.2) and 4.1 (2.7-5.7) A.U. after strength training. The interindividual variation in exercise-induced changes in insulin sensitivity during an OGTT 24 hours post-exercise is visually represented in Figure 1. This plot illustrates a range of change from baseline spanning from -52% to 110%, indicating heterogeneity in response to the interventions, with predominantly positive values. The average intraindividual variation in insulin sensitivity between all three interventions had a CV (SD) of 22.6 (11.3) %, while the average intraindividual variation within the same intervention was 17.5 (13.0) %.

Conclusion: These findings indicate subtle increases in insulin sensitivity following an acute exercise bout. However, we observed vast inter- and intraindividual variation, with no apparent group differences between interventions. Identifying the factors contributing to these individual variations is essential for determining effective, personalized exercise for prevention.

Clinical Trial Registration Number: NCT05718089

Supported by: Trygfonden

Disclosure: N.S. Nielsen: None.

138

Integrated analysis of the skeletal muscle methylome and transcriptome following exercise in females with type 2 diabetes and normal glucose tolerance

K. MacGregor ¹, J. Zierath², A. Krook³;

¹Karolinska Institutet, Solna, Sweden, ²Dept. of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden, ³Integrative Physiology Group, Karolinska Institutet, Stockholm, Sweden.

Background and aims: Epigenetic modifications through DNA methylation contribute to the regulation of skeletal muscle metabolism. While regular physical activity is a successful intervention for the treatment of type 2 diabetes, the underlying molecular mechanisms remain incompletely understood. This study aimed to identify whether alterations to skeletal muscle DNA methylation and transcriptional expression in response to exercise differ in type 2 diabetes, and to examine whether identified differences in DNA methylation are associated with transcriptional expression.

Materials and methods: Females with type 2 diabetes (N=20) or normal glucose tolerance (N=14) completed 35 min cycler exercise at 85 % maximal heart rate. Vastus lateralis muscle biopsies were collected before (pre), immediately (post) and 3 hr after exercise (recovery). Total RNA and DNA were isolated from biopsy tissue to perform transcriptome and methylation arrays.

Results: Exercise induced distinct, temporal alterations to global DNA methylation profiles in skeletal muscle of females with either type 2 diabetes or normal glucose tolerance. Females with type 2 diabetes had a consistently larger number of differentially methylated CpG sites and differentially methylated regions (DMR) compared to females with normal glucose tolerance, with the greatest DNA methylation responses at rest compared to pre (CpG sites = 11,187, DMR = 3,143). Pathway analysis of differentially methylated CpG sites and DMRs in response to exercise identified similar enrichment of immune response, leukocyte function, and T cell activation in skeletal muscle of females with type 2 diabetes or normal glucose tolerance. In females with type 2 diabetes, a further enrichment was observed in pathways relating to cellular development, including cell signalling, cell activation and cytoskeleton organisation. A quantitatively greater transcriptional response was observed in females with type 2 diabetes at recovery vs pre, with 3,839 genes unique to this contrast. Differentially expressed genes (DEG) in response to exercise were enriched in gene ontology pathways related to antigen processing and presentation in females with type 2 diabetes at recovery vs post and recovery vs pre. A significant intersection between DMRs and the corresponding DEG in response to exercise was noted only in females with type 2 diabetes between rest and pre (N=626, 20 %), with an enrichment in gene ontology pathways relating to immune response and cellular remodeling. Further analysis from this subset of intersecting genes identified 8 DMRs that were correlated with the corresponding DEG; ARHGAP15, BIN2, DOCK2, DOCK8, GNAI2, HLA-DMB, PLEK and WIPF1.

Conclusion: Skeletal muscle DNA methylation and transcriptome profiles differ between females with type 2 diabetes and normal glucose tolerance, with a heightened immunogenic signature emerging in response to exercise in females with type 2 diabetes. Alterations in DNA methylation in response to exercise function to regulate immune responses and cellular remodeling in females with type 2 diabetes.

Supported by: NNKI Postdoctoral Fellowship Programme

Disclosure: K. MacGregor: None.

OP 24 Islets forever: survival of the beta cells

139

Covid-19 and type 1 diabetes in Denmark: a longitudinal study on beta cell function

M. Bjerregaard-Andersen ¹, P. Emilie Petersen², K. Kolnes³, M. Bangshaab⁴, A. Razvan Andries⁵, J. Da Silva⁶, E. Carvalho⁶, T. Krarup Hansen⁴, P. Vestergaard⁷, F. Pociot², K. Højlund³, C.B. Juhl¹;

¹Department of Endocrinology, University Hospital of Southern Denmark, Esbjerg, Denmark, ²Steno Diabetes Center Copenhagen, Copenhagen, Denmark, ³Steno Diabetes Center Odense, Odense, Denmark, ⁴Steno Diabetes Center Aarhus, Aarhus, Denmark, ⁵University Hospital of Southern Denmark, Kolding, Denmark, ⁶CNC-UC - Center for Neuroscience and Cell Biology, Coimbra, Portugal, ⁷Steno Diabetes Center North Denmark, Aalborg, Denmark.

Background and aims: A possible association between the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and type 1 diabetes (T1D) development has been reported. It remains unknown whether SARS-CoV-2 infection affects the loss of beta cell function in T1D over time. We investigated this among individuals with new-onset T1D in Denmark.

Materials and methods: Inclusions of adult T1D patients occurred at six Danish hospitals. We report data from November 2020 to February 2024. The following assessments were done within approximately one month of T1D diagnosis: medical history, previous COVID-19, vaccinations, glycated hemoglobin (HbA1c), and SARS-CoV-2 antibodies (Cov2IgG) measurement. A mixed-meal tolerance test (MMTT) was done after an overnight fast at baseline, after 12 and 24 months, consisting of a standardized weight adjusted liquid meal. Blood for plasma glucose and C-peptide was collected at 30-60 minutes intervals over the next 120 or 150 minutes. The primary outcome was C-peptide area under the curve (AUC) during the MMTT (adjusted for test duration), as a measure of residual beta cell function. The median AUCs were compared between Cov2IgG positive and negative individuals, using Mann-Whitneys U-test due to non-normal distribution.

Results: A total of 97 individuals with T1D were included, 74 with available MMTT and Cov2IgG results. 34% (33/97) were female and the mean age was 31.6 (SD 10.1) years. At enrolment, the mean HbA1c was 108.4 (SD 24.8) mmol/mol. Of those with available Cov2IgG result, 19% (15/81) were Cov2IgG positive at enrolment, indicating previous COVID-19. The baseline median AUC_C-peptide was 655 (25-75%: 490-949) pmol/l for Cov2IgG negative vs. 684 (564-1099) for Cov2IgG positive individuals (P=0.49). After 12 months, (N=63), the AUC_C-peptide was 687 (396-911) vs. 564 (492-930) in the two groups, respectively (P=0.93). After 24 months, (N=25), the AUC_C-peptide was 436 (225-646) vs. 418 (273-677) (P>0.99). Cov2IgG status did not significantly alter AUCglucose or HbA1c at baseline or follow-up.

Conclusion: A relatively high proportion of newly diagnosed adult T1D patients were Cov2IgG positive. Assessed by MMTT, COVID-19 did not affect the loss of insulin secretion over time.

Clinical Trial Registration Number: NCT04623697

Supported by: Region of Southern Denmark Post-doc grant

Disclosure: M. Bjerregaard-Andersen: None.

140

Regular at-home monitoring of stimulated dried blood spot C-peptide levels can detect changes in beta cell function early in individuals with recently diagnosed type 1 diabetes

A.E.J. Hendriks ¹, M.L. Marcovecchio¹, P. Barker², M. Evans³, C. Mathieu⁴, on behalf of the INNODIA consortium;

¹Department of Paediatrics, University of Cambridge, Cambridge, UK, ²NIHR Cambridge Biomedical Research Centre, Core Biochemistry Assay Laboratory, Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK, ³Institute of Metabolic Science and Department of Medicine, University of Cambridge, Cambridge, UK, ⁴Clinical and Experimental Endocrinology, Department of Chronic Diseases and Metabolism, KU Leuven, Leuven, Belgium.

Background and aims: Early detection of changes in beta cell function in individuals with newly diagnosed type 1 diabetes (T1D) is essential for informing the design of clinical trials. However, the gold standard for this, the mixed-meal tolerance test (MMTT), is an invasive inpatient procedure that cannot be conducted frequently enough for this purpose. Here we assessed whether regular at-home measured dried blood spot (DBS) C-peptide levels, both fasting and after a standard liquid meal, can predict changes in beta cell function early.

Materials and methods: The INNODIA Newly Diagnosed cohort consists of people with T1D (aged 1-45 yrs) recruited within 6 weeks from diagnosis. Participants measured finger prick DBS C-peptide monthly fasting and 60 min post standard liquid meal (Ensure Plus®). Participants had plasma C-peptide levels measured at baseline and an MMTT was performed at 12 months to calculate plasma C-peptide area under curve (AUC) and measure peak plasma C-peptide. Linear mixed model was used to predict plasma C-peptide AUC and peak plasma C-peptide at 12 months using DBS C-peptide slope measured over the first 6 months after diagnosis with correction for age and baseline C-peptide levels.

Results: 292 people with newly diagnosed T1D (58% male) were studied, this included 125 children (<10 yrs), 124 adolescents (10-18 yrs) and 43 adults (>18 yrs). Median [IQR] number of DBS card pairs was 6.5 [3-10] over 12 months. Children had lower DBS C-peptide levels (fasting: median [IQR] 157 [87-242] pmol/l, p<0.01) than adolescents and adults (fasting: 303 [195-471] and 347 [265-523] pmol/l). DBS C-peptide significantly declined over time (p<0.001), but slope was not different between age groups. Plasma C-peptide AUC at 12 months was significantly lower in children (193 [154-420] pmol/l/min, p<0.001) than adolescents and adults (591 [310-989] and 857 [484-1092] pmol/l/min). Fasting DBS C-peptide measured over the first 6 months could not predict plasma C-peptide AUC or peak plasma C-peptide at 12 months. However, post liquid meal DBS C-peptide in the first 6 months after diagnosis significantly predicted plasma C-peptide AUC and peak plasma C-peptide at 12 months (p<0.01) (Figure 1A/1B).

Conclusion: Frequently at-home measured DBS C-peptide after a standard liquid meal can reliably detect changes in beta cell function in the first 6 months after diagnosis and is an outcome measure worth considering in clinical trial design.

Supported by: The Union’s Horizon 2020 research and innovation programme, ‘EFPIA’, ‘JDRF’ and ‘The Leona M. and Harry B. Helmsley Charitable Trust

Disclosure: A.E.J. Hendriks: Grants; The Union’s Horizon 2020 research and innovation programme, EFPIA, JDRF, The Leona M. and Harry B. Helmsley Charitable Trust.

141

Characterisation of residual proinsulin and C-peptide after 50 years of insulin dependence: the Joslin Medalist study

M. Yu, T. Chokshi, S. Jangolla, N. Ziemniak, E. Viebranz, I.-H. Wu, H. Shah, S. Bonner-Weir, G. King;

Research Division, Joslin Diabetes Center, Boston, MA, USA.

Background and aims: Comparative analysis of serum proinsulin and C-peptide for assessing residual beta cell function has not been well-studied in long-duration Type 1 diabetes (T1D). We evaluated whether a subset of the Joslin ”Medalists” (T1D≥50 years) retained the ability to secrete proinsulin, and whether this was associated with beta-cell function, islet autoantibodies (AAb) and morphology, and genetic risks for T1D and monogenic diabetes, in comparison to C-peptide.

Materials and methods: Medalists (n=245) underwent clinical evaluation including analysis of HLA T1D risk alleles, islet AAb, and fasting serum C-peptide and proinsulin. Beta-cell responses to metabolic stimuli were evaluated by mixed-meal tolerance tests (MMTT) and hyperglycemia/arginine clamps. Known monogenic diabetes variants were screened and classified as “likely pathogenic” if they had a REVEL (rare exome variant ensemble learner) score>0.75. Postmortem analysis of pancreases from deceased Medalists was also performed.

Results: In this Medalist subset, 18.5% and 30.9% exhibited detectable serum proinsulin and C-peptide, respectively. While those with detectable C-peptide (>0.05 ng/mL) had higher median proinsulin (0.054 vs. 0.049 pmol/L, p<0.01) compared to those with undetectable C-peptide, 10.1% of those with undetectable C-peptide, interestingly, also retained detectable proinsulin. In those with MMTT (n=167), 15.6% responded with doubling of baseline C-peptide. MMTT responders had higher prevalence of detectable fasting proinsulin (61.5% vs. 12.1%, p<0.01) and higher median fasting proinsulin (0.061 vs. 0.049 pmol/L, p<0.01) compared to non-responders. Medalists with baseline C-peptide ≥0.1 ng/mL (n=17) also underwent a hyperglycemic/arginine clamp, to which 9 responded with doubling of C-peptide. Clamp responders had consistently higher median stimulated proinsulin compared to non-responders, both after hyperglycemia stimulation alone (0.072 vs. 0.056 pmol/L, p<0.01) as well as after arginine stimulation (0.077 vs. 0.057 pmol/L, p<0.01). In terms of T1D genetic and autoimmune features, median proinsulin was lowest in HLA+/AAb+ Medalists, compared to HLA+/AAb- and HLA-/AAb- Medalists (0.049 vs. 0.051 vs. 0.052 pmol/L, respectively; p=0.02). In the HLA-/AAb- group, those who were positive for known monogenic diabetes variants (REVEL score>0.75) had higher prevalence of detectable proinsulin (80% vs. 6.7%, p<0.01) and higher median proinsulin (0.070 vs. 0.049 pmol/L, p<0.01) compared to those who tested negative. Postmortem examination of 73 Medalist pancreases showed three categories: A) only scattered insulin-positive cells in singlets/doublets (n=12); B) scattered plus few insulin-positive cells in some islets (n=46); and C) scattered plus some lobes with islets having many insulin-positive cells (n=15). The prevalence of both detectable serum proinsulin and C-peptide increased going from Category A to C (8.3% vs. 8.7% vs. 33.3%, p=0.04; and 8.3% vs. 28.4% vs. 60%, p=0.01; respectively).

Conclusion: People with long-duration T1D can retain the ability to secrete both proinsulin and C-peptide, with proinsulin being slightly less detectable. Higher proinsulin was associated with greater insulin positivity in islets, greater beta-cell response to metabolic stimuli, and lower T1D genetic and autoimmune risks, and may provide markers of beta-cell function in long-duration T1D.

Supported by: Mary K. Iacocca Foundation

Disclosure: M. Yu: None.

142

Assessment of beta cell function and glucose metabolism after near total pancreatectomy: implications of 80% pancreas reduction

L. Soldovieri ¹, G. Di Giuseppe¹, M. Zidda¹, M. Brunetti¹, G. Ciccarelli¹, S. Moffa¹, F. Cinti¹, A. Mari², G. Quero¹, S. Alfieri¹, A. Giaccari¹, T. Mezza¹;

¹Fondazione Policlinico Universitario Agostino Gemelli, Rome, Italy, ²Institute of Neuroscience, National Research Council (CNR) of Padova, Padova, Italy.

Background and aims: Hyperglycemia typically develops on a background of insulin resistance when the endocrine pancreas fails to cope with increased insulin demand with adequate insulin secretion. Many studies have shown that diabetes appears when the loss of pancreatic β cell mass is approximately between 50% and 80% in animals and between 41% and 65% in humans. Near total pancreatectomy (NTP) is a surgical procedure that removes 80% of pancreatic mass. In this study we aim to investigate the metabolic effects of NTP in a cohort of non-diabetic subjects on residual β cell function and glucose metabolism.

Materials and methods: We recruited 7 non-diabetic subjects undergoing near total pancreatectomy. Participants underwent an oral glucose tolerance test (OGTT) and a hyperglycemic clamp (HC) followed by arginine stimulation to estimate β cell responses and glucose tolerance before and 3 months after surgery. We calculated the area under the curve of glucose (AUC-Glu) and insulin (AUC-Ins) responses during OGTT. Therefore, we estimated the β cell glucose sensitivity (βCGS), an index of β cell function, and the rate sensitivity (RS), an index of first phase insulin secretion. Further, we calculated the total area under the curve of insulin responses (AUC-Itot) and arginine-stimulated insulin secretion (AUC-Arg) during HC.

Results: Compared to presurgical OGTT, we observed a 18% increase in glucose responses during OGTT after surgery (AUC-Glu pre 19761 vs post 23421 mg·dl^-1·min) with a 57% reduction of insulin responses (AUC-Ins pre 6499 vs post 2788 μU/ml·min p<0.05). Further, we observed a 36% reduction in the βCGS (pre 61.2 vs post 38.7 pmol min^-1m^-2mmol^-1L) and a 82% reduction in the RS (pre 763 vs post 139 pmol m-2mmol-1L p<0.05) after surgery. Both total insulin secretion and arginine-stimulated insulin secretion during HC were reduced (AUC-Itot pre 15776 vs 6489 μU/ml·min and AUC-Arg pre 7144 vs post 2520 μU/ml·min), respectively by 58.8% and 64.7%. According to the ADA diagnostic criteria, 3 of the 7 subjects developed diabetes.

Conclusion: In conclusion, despite previous reports, 80% β cell mass reduction is not always sufficient to develop hyperglycemia. Additional studies will be necessary to understand possible mechanisms playing a compensating role in the preservation of glucose tolerance despite near-total β cell mass loss.

Disclosure: L. Soldovieri: None.

143

EGFR inhibitor erlotinib downregulates high fat mediated beta cell inflammation, lipid accumulation and augments insulin secretory function

O. Mukherjee, S. Nag, S. Mandal, N. Das, R. Kundu;

Department of Zoology, Visva- Bharati University, Santiniketan, India.

Background and aims: The epidermal growth factor receptor (EGFR), a receptor tyrosine kinase, plays a pivotal role in regulating cellular growth, proliferation and survival. Existing literature indicates that circulating free fatty acids (FFAs) activate EGFR and may lead to various metabolic disorders including type 2 diabetes. Fetuin-A, a hepato-adipokine that is also secreted from pancreatic beta cells, is known to augment FFA signaling by binding with TLR4 and causing beta cell dysfunction. However, the precise underlying mechanisms remain elusive. The present study thus aims to elucidate some hitherto unknown effects of EGFR activation in pancreatic beta cells under hyperlipidemic conditions and the effect of its inhibition by Erlotinib (a tyrosine kinase inhibitor) on insulin secretory function.

Materials and methods: Immunoblot and qPCR analysis of mouse insulinoma (MIN6) cell line were performed to study the expression of EGFR, pEGFR, Pdx1 and several lipogenic and inflammatory markers in palmitate(0.75 mM)+fetuin-A(100 ug/ml) treated cells pre-incubated with Erlotinib(10 uM). EGFR-siRNA was used to study the changes in gene expression under above treatment conditions. Oil O Red staining was performed to visualize lipid accumulation in beta cells. Immunofluorescence studies were done to observe the expression and cellular localization of various proteins. For in vivo experiments, 8-week-old C57BL/6J male mice were randomly divided into 3 groups (n=6) that received standard diet, high fat diet (HFD) and HFD+Erlotinib(100 mg/kg/day) for 16 weeks. Erlotinib was administered during the last 4 weeks of treatment period. Pancreatic islets from treated mice were isolated by Percoll gradient and subjected to immunoblot, qPCR analysis and Oil O Red staining. Immunofluorescence of pancreatic sections was performed. Insulin, IL-1b and fetuin-A levels were checked by ELISA analysis. GTT, ITT, HOMA-IR and GSIS assays were also performed.

Results: MIN6 cells treated with palmitate+fetuin-A showed increased EGFR and pEGFR expression (~2 folds) compared to palmitate alone (p<0.05). Expression of lipogenic genes SREBP1, FASN and PPARg were elevated in palmitate+fetuin-A treatment (~1.7 folds), which were significantly reduced by Erlotinib (~1.6 folds, p<0.05). EGFR inhibition by Erlotinib and EGFR-siRNA showed increased Pdx1 expression in palmitate+fetuin-A treated cells (~2 folds). Co-localization studies further depicted reduced lipid accumulation in Erlotinib treated cells, along with an increase in insulin content. Erlotinib also alleviated inflammatory markers such as TLR4, NFkB, pNFkBp65, fetuin-A and IL-1b (~1.5 folds). Pancreatic islets of HFD+Erlotinib treated C57BL/6J mice exhibited decreased expression of SREBP1, FASN, PPARg, TLR4, NFkB and pNFkBp65 (~1.7 folds) compared to HFD mice (p<0.05). Additionally, serum levels of fetuin-A and IL-1b were downregulated after Erlotinib treatment. In isolated islets, Erlotinib led to attenuation of lipid accumulation and subsequent improvement of insulin sensitivity in mice, as evident from GTT, ITT and HOMA-IR studies.

Conclusion: The results indicate that EGFR inhibition significantly improves high fat induced pancreatic beta cell dysfunction by reducing inflammation, lipid accumulation and insulin secretory defects and could be a possible therapeutic target in diabetes treatment.

Supported by: SERB grant no. CRG/2022/007004

Disclosure: O. Mukherjee: Grants; SERB grant no. CRG/2022/007004.

144

Role of autophagy and intermittent fasting in beta cell mass and identity in K _ATP -induced neonatal diabetes

E. Castelblanco, Z.A. Shyr, I. Ramirez-Sotero, M.S. Remedi;

Department of Medicine, Division of Endocrinology, Metabolism and Lipid Research, Washington University School of Medicine in St Louis, Saint Louis, MO, USA.

Background and aims: Loss of pancreatic β-cell mass is a hallmark of both type 1 and type 2 diabetes. Autophagy is a cellular process that serves as a cornerstone in maintaining cellular homeostasis by clearing misfolded proteins and damaged organelles. Impaired β-cells autophagy has been demonstrated in both human subjects and animal models of type 1 and type 2 diabetes. However, the role of autophagy in loss of functional β-cell mass and identity in diabetes remains elusive. In this study, we delve into the intricate role of autophagy in modulating β-cell mass and identity to shed light on the underlying mechanisms driving diabetic pathology.

Materials and methods: In this study we utilized adult male and female tamoxifen-inducible K_ATP-gain-of-function (GOF) mouse model of human monogenic neonatal diabetes, with overactive K_ATP channels and lack of glucose-stimulated insulin secretion (GSIS). Both K_ATP-GOF and littermate control mice were injected with tamoxifen for five days, and then subjected to ad-libitum feeding (AL) or alternate-day intermittent fasting (IF). Our methodology encompassed an array of techniques, from physiological measurements of glucose and insulin levels, and glucose and insulin tolerance tests to electron microscopy for ultrastructural analysis, and western blotting and immunofluorescence for protein expression profiling.

Results: When fed AL, K_ATP-GOF mice recapitulated the human diabetic phenotype promoted by lack of GSIS. As diabetes progresses, AL-K_ATP-GOF mice demonstrate a marked decrease in insulin content, oxidative and endoplasmic reticulum (ER) stress, and progressive loss of β-cell mass and identity. Our studies revealed elevated LC3II message and protein levels in islets from K_ATP-GOF mice, suggesting autophagosome formation. However, despite this, we found a significant decrease in the LC3II chloroquine/LC3II saline ratio, indicating impaired autophagy. Consistent with this, electron microscopy images demonstrate increased autophagosome abundance/accumulation, distended ER and swelled-rounded mitochondria with deteriorated cristae in β-cells K_ATP-GOF mice. Importantly, mice subjected to alternate-day intermittent fasting (IF) exhibited improved glycemic control and glucose tolerance, enhanced insulin content, reduced ER stress, and restoration of mitochondrial morphology. Notably, IF mitigated autophagosome accumulation in islets of K_ATP-GOF mice and normalized autophagy to control levels. Furthermore, the heightened co-localization of LC3II:LAMP1 observed in β-cells from AL-fed K_ATP-GOF mice was restored to control levels in islets from IF-K_ATP-GOF mice. Strikingly, IF not only restored autophagy in islets from K_ATP-GOF mice but also prevented loss of β-cell mass and identity.

Conclusion: Our study underscores the critical role of autophagy dysregulation in non-obese and insulin secretory-deficient diabetes monogenic diabetes. Additionally, it demonstrates, for the first time, the therapeutic potential of IF in preserving β-cell mass and identity through autophagy improvement. These findings offer critical insights into the mechanisms underlying loss, and restoration, of β-cell mass and identity, highlighting a key role of autophagy modulation in this process.

Supported by: NIH R01DK123163

Disclosure: E. Castelblanco: None.

OP 25 Incretin receptor agonists: better and better

145

Comparative efficacy and tolerability of incretin-based therapies for adults with overweight or obesity: a network meta-analysis

T. Karagiannis ^1,2, K. Malandris^1,2, P. Kakotrichi¹, D. Vasilakou^1,2, I. Avgerinos^1,2, K. Tsapa¹, D. Tsapa¹, A. Liakos^1,2, E. Bekiari^1,2;

¹Clinical Research and Evidence-Based Medicine Unit, Aristotle University of Thessaloniki, Thessaloniki, Greece, ²Diabetes Centre, Second Medical Department, Aristotle University of Thessaloniki, Thessaloniki, Greece.

Background and aims: There is paucity of randomised controlled trials (RCTs) directly comparing tirzepatide, subcutaneous (s.c.) semaglutide and liraglutide in adults with overweight or obesity. We did a network meta-analysis to compare these three medications in terms of their efficacy and tolerability in this population.

Materials and methods: We searched PubMed and the Cochrane Central Register of Controlled Trials for RCTs that assessed tirzepatide 10 or 15 mg once-weekly, s.c. semaglutide 2.4 mg once-weekly, or liraglutide 3.0 mg once-daily versus placebo, in adults with overweight (body mass index [BMI] ≥27 kg/m² to <30 kg/m²) and at least one weight-related comorbid condition, or obesity (BMI ≥30 kg/m²). We did frequentist random-effects network meta-analyses and calculated mean differences (MDs) for % change from baseline body weight, and risk ratios (RRs) for achievement of ≥5% weight reduction and for incidence of gastrointestinal adverse events (nausea, vomiting, and diarrhoea).

Results: Thirty-six RCTs (35,378 participants) contributed data to the network meta-analysis. Tirzepatide 15 mg was the most efficacious in terms of % body weight reduction versus placebo, followed by tirzepatide 10 mg, semaglutide 2.4 mg, and liraglutide 3.0 mg (Figure). In between-drug comparisons, tirzepatide 15 mg was more efficacious compared to the other three comparators, while tirzepatide 10 mg and semaglutide 2.4 mg were more efficacious than liraglutide 3.0 mg. Versus placebo, the RR for achieving ≥5% weight reduction was 2.65 (95% CI, 2.06 to 3.42) for tirzepatide 15 mg, 2.61 (1.92 to 3.55) for tirzepatide 10 mg, 2.48 (2.02 to 3.03) for semaglutide 2.4 mg, and 2.07 (1.77 to 2.42) for liraglutide 3.0 mg, with no differences in between-drug comparisons. As compared to placebo, all medications increased risk for gastrointestinal adverse events, with tirzepatide 10 mg yielding the highest RR for nausea (3.32; 95% CI, 2.74 to 4.04), and tirzepatide 15 mg yielding the highest RR for vomiting (5.98; 4.14 to 8.65) and diarrhoea (2.95; 2.39 to 3.65). Notable between-drug differences were observed for tirzepatide 15 mg versus liraglutide in vomiting, and for both tirzepatide doses versus semaglutide and liraglutide in diarrhoea.

Conclusion: Tirzepatide outperformed s.c. semaglutide 2.4 mg and liraglutide 3.0 mg in terms of weight reduction in adults with overweight or obesity but was associated with a higher risk for gastrointestinal adverse events.

Disclosure: T. Karagiannis: None.

146

Is semaglutide as effective at reducing major cardiovascular events in the presence of impaired kidney function in people with overweight or obesity? A prespecified analysis from the SELECT trial

H.M. Colhoun ¹, S.E. Kahn², P.M. Brown³, E.T.B. Olesen³, R. Bravo³, J. Deanfield⁴, A. Goudev⁵, G. Latkovskis⁶, M. Lehrke⁷, A.M. Lincoff⁸, N. Rathor³, C.-C. Wu⁹, I. Lingvay¹⁰, On behalf of the SELECT investigator group;

¹University of Edinburgh, Edinburgh, UK, ²VA Puget Sound Health Care System and University of Washington, Seattle, WA, USA, ³Novo Nordisk A/S, Søborg, Denmark, ⁴University College London, London, UK, ⁵Medical University of Sofia, Sofia, Bulgaria, ⁶University of Latvia, Riga, Latvia, ⁷University of Aachen, Aachen, Germany, ⁸Cleveland Clinic Lerner College of Medicine of Case Western Reserve University, Cleveland, OH, USA, ⁹National Taiwan University Hospital, Taipei, Taiwan, Province of China, ¹⁰UT Southwestern Medical Center, Dallas, TX, USA.

Background and aims: In the SELECT trial, semaglutide (sema) 2.4 mg reduced major adverse cardiovascular events (MACE) by 20% vs placebo (pbo) in participants (pts) with overweight or obesity (BMI ≥27 kg/m²) and prior CVD but without diabetes. This prespecified analysis examined if the treatment effect on MACE and a combined endpoint of MACE or death from any cause was maintained in pts with reduced eGFR or micro/macroalbuminuria.

Materials and methods: Pts were randomised to once-weekly s.c. sema 2.4 mg (n=8,803) or pbo (n=8,801). eGFR (mL/min/1.73 m²) was estimated using the serum creatinine-based CKD-EPI 2009 equation. Urinary albumin-to-creatinine ratio (UACR; mg/g) was calculated using single urine samples. MACE was a composite of death from cardiovascular causes, nonfatal MI or nonfatal stroke in a time-to-first-event analysis. Cox proportional hazard models estimated treatment effect on MACE according to eGFR and albuminuria status at randomisation. Treatment effect for the combined endpoint of MACE or death from any cause was also estimated.

Results: At randomisation, 1,908 (11%) pts had eGFR <60 and 2,281 (13%) had UACR ≥30. Median follow-up: 3.5 years. In pts with eGFR ≥60, MACE occurred in 6.0% of sema vs 7.3% of the pbo arm (HR=0.82 [95% CI 0.72-0.92]). In pts with eGFR <60, MACE occurred in 9.7% of sema vs 13.5% of the pbo arm (HR=0.69 [0.52-0.90]; p interaction=0.22) (Figure 1). For MACE or death from any cause, HRs for sema vs pbo were 0.82 (0.74-0.92) for eGFR ≥60, vs 0.67 (0.53-0.84) for eGFR <60 (p interaction=0.12). In pts with UACR <30, MACE occurred in 5.9% of sema vs 7.3% of the pbo arm (HR=0.80 [0.70-0.90]). In pts with UACR ≥30, MACE occurred in 9.9% of sema vs 12.3% of the pbo arm (HR=0.80 [0.62-1.02]; p interaction=0.97) (Figure 1). For MACE or death from any cause, HRs for sema vs pbo were 0.79 (0.71-0.89) for UACR <30 and 0.81 (0.65-1.01) for UACR ≥30 (p interaction=0.88).

Conclusion: The effect of sema in reducing MACE, and reducing MACE or death from any cause, was consistent in pts with eGFR <60 vs ≥60 mL/min/1.73 m² or UACR <30 vs ≥30 mg/g. There was a higher rate of MACE associated with reduced eGFR and albuminuria in pts with overweight or obesity but without diabetes. Thus, the absolute treatment benefit on MACE with sema is higher in the presence of impaired kidney function.

Clinical Trial Registration Number: NCT03574597

Disclosure: H.M. Colhoun: Employment/Consultancy; consultancy for Novo Nordisk and Bayer. Grants; research funding from Sanofi, Roche, and IQVIA; grants from Chief Scientist Office, Diabetes UK, European Commission, Juvenile Diabetes Research Foundation, and Medical Research Council. Stock/Shareholding; stock in Roche Pharmaceuticals and Bayer. Other; funded by Novo Nordisk; SELECT ClinicalTrials.gov number, NCT03574597; speaker’s bureau for Novo Nordisk.

147

ZT002, a novel ultra long-acting GLP-1 receptor agonist in adults with overweight or obesity: a randomised, placebo-controlled, multiple ascending dose phase 1c study

L. Ji ¹, H. Zhou², X. Li², X. Xu³, A. Wong³, Y. Zu³, X. Song³, Y. Zhang³, S. Lin³, X. Zhang³, Y. Zhang³;

¹Department of Endocrinology and Metabolism, Peking University People’s Hospital, Beijing, China, ²Clinical Trial Research Center, The First Affiliated Hospital of Bengbu Medical University, Benbu, China, ³Beijing QL Biopharmaceutical Co., Ltd., Beijing, China.

Background and aims: ZT002 is a novel ultra long-acting glucagon-like peptide-1 receptor agonist (GLP-1RA) (t1/2: ~13 days) that is being developed for weight management with monthly subcutaneous administration. We aimed to examine the safety, tolerability, pharmacokinetics, and pharmacodynamics of ZT002 in adults with overweight or obesity.

Materials and methods: A phase 1c, randomized, placebo-controlled, double-blind, multiple ascending dose study was conducted in adults with overweight (24 kg/m²≤ body-mass index [BMI] <28 kg/m²) accompanied by at least one co-morbidity or obesity (28 kg/m²≤BMI≤40 kg/m²). This is a 12-14 week multiple ascending dose study comprised of 2 cohorts, and in each cohort, eligible participants were randomly assigned (10:4) to biweekly subcutaneous ZT002 (40 or 80 mg) or matched placebo. In each cohort, the dose of ZT002 was escalated in 2-week intervals to the desired dose over 4-6 weeks, participants were treated at the target dose for 10 weeks, and then followed up for 4 weeks.

Results: Total of 28 participants (N=21 [men]/7 [women]; median age=26 years; median BMI=33.2 kg/m2) were randomly assigned to ZT002 (40 mg group n=10; 80 mg group n=10) or placebo (n=8) and included in the safety and full analysis datasets. Over the 12 or 14-week treatment period, ZT002 was generally well tolerated without any treatment-related discontinuations or drug-related severe adverse events. The most frequent side effects reported were gastrointestinal disorders (nausea, vomiting, diarrhea) and decreased appetite, which were mostly mild to moderate in severity. The exposure at steady state was dose proportional to ZT002 doses. The mean percent change from baseline to week 12 in body weight was -9.6 % (SE 1.9) for participants receiving ZT002 in the 40 mg cohort and -0.8% (0.7) for participants receiving placebo. The mean percent change from baseline to week 14 in body weight was -13.1% (SE 1.7) for participants receiving ZT002 in the 80 mg cohort and -1.8% (1.3) for participants receiving placebo. Additionally, compared with placebo, ZT002 achieved more profound reductions in waist circumference. Cardiometabolic risk factors improved in all treatment groups.

Conclusion: This MAD trial demonstrated biweekly treatment with ZT002 induced clinically meaningful weight loss and cardiometabolic improvements with favorable tolerability and safety profiles. Further clinical evaluation of ZT002, including monthly subcutaneous administration for overweight/obesity, is ongoing.

Disclosure: L. Ji: Employment/Consultancy; Beijing QL Biopharmaceutical Co., Ltd.

148

A phase 3 evaluation of cAMP signalling biased GLP-1 analogue ecnoglutide in adults with type 2 diabetes

S. Xu ¹, D. Zhu², S. Bing³, Q. Zheng³, L. Guan³, F. Jiang³, X. Liu³, Y. Bu³, J. Ning³, Z. Zhu³, L. Yang³, M. Yang³, M. Fenaux¹, M.K. Junaidi¹, Ecnoglutide for T2DM Study Team;

¹Sciwind Biosciences, San Ramon, CA, USA, ²Nanjing University Medical School, Nanjing, China, ³Sciwind Biosciences, Hangzhou, China.

Background and aims: Ecnoglutide (XW003) is a cAMP signaling biased GLP-1 analog being developed for the treatment of type 2 diabetes mellitus (T2DM) and obesity. The primary objective of this study was to evaluate the efficacy of ecnoglutide administered for 24 weeks in adults with T2DM.

Materials and methods: We conducted a Phase 3 randomized, double-blind, placebo-controlled study of ecnoglutide, enrolling 211 adults with T2DM at 33 sites in China. Participants were randomized to receive 0.6 or 1.2 mg ecnoglutide or placebo as once weekly injections for 24 weeks, including dose escalation. All participants then received ecnoglutide (0.6 or 1.2 mg) for a total duration of 52 weeks. Change in mean HbA1c, body weight, and BMI, as well as safety and tolerability were evaluated.

Results: At baseline, participants had mean HbA1c of 8.54, 8.51, and 8.51% and BMI of 27.2, 26.4, and 27.2 kg/m2, for ecnoglutide 0.6, 1.2 mg, and placebo groups, respectively. After 24 weeks, participants receiving ecnoglutide achieved significant HbA1c reductions of 1.96 to 2.43% from baseline (P≤0.0003 for both cohorts vs placebo). At 24 weeks, 76.1% of participants receiving 1.2 mg ecnoglutide achieved HbA1c ≤6.5%, 35.2% had HbA1c <5.7%, and 43.7% had body weight reductions ≥5% from baseline. Ecnoglutide was safe and well tolerated. The proportion of participants reporting any adverse event (AE) ranged from 77.5 to 78.3% for ecnoglutide groups and 60.6% for placebo. Four (2.9%) treatment-related ≥Grade 3 AEs and one (0.7%) treatment-related serious AE occurred in the ecnoglutide groups. One participant from each cohort discontinued due to an AE. The most frequently reported AEs were decreased appetite, diarrhea, and nausea, which were mostly mild to moderate and transient.

Conclusion: Ecnoglutide resulted in robust HbA1c declines of up to 2.43% from baseline after 24 weeks of treatment in adults with T2DM, with up to 35.2% of participants reaching normoglycemia (HbA1c <5.7%) and 43.7% with weight reductions ≥5%.

Clinical Trial Registration Number: NCT05680155

Disclosure: S. Xu: Employment/Consultancy; Sciwind Biosciences.

149

Efficacy and safety of mazdutide in Chinese participants with overweight or obesity (GLORY-1)

L. Ji ¹, H. Jiang², H. Li³, J. Tian⁴, D. Liu⁵, Y. Zhao⁶, W. Qiu⁷, Y. Bi⁸, J. Gu⁹, Z. Liu⁹, H. Deng⁹, Y. Wang⁹, L. Li⁹, L. Qian⁹, The GLORY-1 Investigators;

¹Peking University People’s Hospital, Beijing, China, ²The First Affiliated Hospital and Clinical Medicine College, Henan University of Science and Technology, Luoyang, China, ³Zibo Municipal Hospital, Zibo, China, ⁴Luoyang Third People’s Hospital, Luoyang, China, ⁵Nanyang Medical College First Hospital, Nanyang, China, ⁶Shenyang Fifth People’s Hospital, Shenyang, China, ⁷Huzhou Central Hospital, Huzhou, China, ⁸Nanjing Drum Tower Hospital, Nanjing, China, ⁹Innovent Biologics, Inc, Suzhou, China.

Background and aims: The once-weekly glucagon like peptide-1 and glucagon receptor dual agonist mazdutide showed robust weight reduction and improvement on multiple cardiometabolic risk factors in phase 2 studies. GLORY-1 is a randomized, double-blind, placebo-controlled, 48-week phase 3 trial that evaluated the efficacy and safety of mazdutide in Chinese adults with overweight or obesity.

Materials and methods: We assigned 610 Chinese adults with a BMI ≥28 kg/m², or ≥24 kg/m² and at least one weight-related comorbidity, in a 1:1:1 ratio to receive once-weekly, subcutaneous mazdutide 4 mg, mazdutide 6 mg or placebo for 48 weeks. Co-primary endpoints were the percentage change in weight from baseline and a weight reduction ≥5% at week 32. Key secondary endpoints included the percentage change in weight from baseline at week 48, weight reduction ≥5%, ≥10% and ≥15% at week 48, change in waist circumference from baseline at week 48, as well as change in systolic blood pressure, lipids, serum uric acid and alanine aminotransferase from baseline at week 48. A mixed-effect model for repeated measures and a logistic regression model were conducted for continuous and binary endpoints, respectively.

Results: At baseline, mean body weight was 87.2 kg and mean BMI was 31.1 kg/m². Percentage change from baseline in weight was significantly greater with mazdutide 4 mg and 6 mg than with placebo at week 32 and 48. Compared with placebo, significantly more participants in mazdutide groups achieved weight reduction ≥5% at week 32, ≥5%, ≥10% and ≥15% at week 48 than those in the placebo group. Change in waist circumference was significantly greater in mazdutide groups than in the placebo group (Table 1). Furthermore, mazdutide (4 mg and 6 mg combined) demonstrated superiority versus placebo on change from baseline to week 48 in systolic blood pressure (estimated treatment difference −6.75 [95%CI: −8.57, −4.94]), triglycerides (−0.52 [−0.66, −0.39]), total cholesterol (−0.45 [−0.57, −0.34]), LDL cholesterol (−0.31 [−0.40, −0.23]), serum uric acid (−50.75 [−61.77, −39.74]) and alanine aminotransferase (−10.00 [−13.19, −6.80]) (P<0.0001 for all comparison). The most frequently reported treatment-emergent adverse events were gastrointestinal, mostly mild to moderate in severity.

Conclusion: In Chinese adults with overweight or obesity, mazdutide provides significant reductions in body weight and cardiometabolic risk factors.

Clinical Trial Registration Number: NCT05607680

Supported by: Innovent Biologics, Inc.

Disclosure: L. Ji: None.

150

A triple GIP, GLP-1 and glucagon receptor agonist, retatrutide, decreases inflammatory CV risk biomarkers in people with overweight or obesity, with or without type 2 diabetes

G. Ruotolo, C. Harris, Y. Lin, M.K. Thomas, Y. Qu, J. Wilson, K. Duffin, M.L. Hartman, T. Coskun, Z. Milicevic, A. Haupt;

Eli Lilly and Company, Indianapolis, IN, USA.

Background and aims: Retatrutide (RETA), an agonist of glucose-dependent insulinotropic polypeptide (GIP) receptor, glucagon-like peptide-1 (GLP-1) receptor, and glucagon receptor, demonstrated weight and HbA1c reduction in people with obesity and T2D. Biomarkers of inflammation, including high-sensitivity C-reactive protein (hsCRP) and interleukin-6 (IL-6), have been shown to be associated with increased cardiovascular risk.

Materials and methods: Data from adult participants in two phase 2 studies of RETA were included: Study 1(OB) -participants with overweight or obesity, without T2D treated with RETA (1, 4, 8 and 12 mg) vs placebo (PBO); Study 2 (T2D)-participants with T2D and BMI 25-50 kg/m² with RETA (0.5, 4, 8 and 12 mg) vs PBO and dulaglutide 1.5 mg. Adiponectin, leptin, hsCRP, IL-6, and TNF-α were measured by immunoassay in fasting plasma samples collected at baseline and at 24, and 36 (T2D) or 48 (OB) weeks. A mixed model repeated measures (MMRM) was used to analyze the percent change from baseline (i.e., changes after log-transformation) and compare the difference between treatment and PBO.

Results: At baseline, geometric least square mean hsCRP levels were higher in overall participants with T2D than in those with overweight or obesity (4.3 mg/L vs 3.4 mg/L, p=0.005). RETA reduced hsCRP levels from 31% (RETA 1 mg) up to 55% (RETA 8 mg) from baseline to week 48 in OB and from 24% (RETA 0.5 mg) to 41% (RETA 12 mg) from baseline to week 36 in T2D (Table). In the subgroup of participants with baseline hsCRP >2mg/L, hsCRP levels were reduced from 35% (RETA 1 mg) up to 63% (RETA 8 mg, p<0.001 vs PBO) in OB and from 28% (RETA 0.5 mg) up to 46% (RETA 4 mg, p=0.09 vs PBO) in T2D. RETA also reduced IL-6 levels up to 30% (RETA 8 mg, p<0.01 vs PBO) in OB and 30% in T2D (RETA 8 mg, p=0.3 vs PBO). RETA reduced leptin levels from 22% (RETA 1 mg) up to 72% (RETA 12 mg) and from 18% (RETA 0.5 mg) up to 53% (RETA 8 mg), while increasing adiponectin levels from 19% (RETA 1 mg) up to 71% (RETA 8 mg) and from 11% (RETA 0.5 mg) up to 57% (RETA 8 mg), in OB at 48 weeks and in T2D at 36 weeks, respectively.

Conclusion: RETA robustly reduced hsCRP and IL-6 levels in participants with overweight or obesity, without T2D in a dose-dependent manner. Similarly, RETA reduced hsCRP and IL-6 levels in patients with T2D, even though not significantly vs placebo. These data suggest that RETA may reduce cardiovascular risk in people with overweight or obesity, with or without type 2 diabetes, through inflammation reduction as evidenced by reduced circulating inflammatory biomarkers.

Disclosure: Initially submitted at European Society of Cardiology -ESC Congress 2024

Clinical Trial Registration Number: NCT04867785; NCT04881760

Disclosure: G. Ruotolo: Employment/Consultancy; Eli Lilly and Company. Stock/Shareholding; Eli Lilly and Company.

OP 26 Love goes through the stomach: new roles for GLP-1

151

Alpha-lipoic acid’s role in improving GLP-1 agonist treatment tolerability in young overweight patients with prediabetes

L. Jashi ¹, T. Peshkova¹, T. Shervashidze², R. Kvanchakhadze³, K. Dundua¹, L. Beridze¹;

¹Dept of Endocrinology, Avicenna - Batumi Medical University, Batumi, Georgia, ²David Metreveli Medical Center, Tbilisi, Georgia, ³

Background and aims: Obesity is a significant factor in the development of diabetes, which is often accompanied by prediabetes (hyperinsulinemia, insulin resistance). One solution to delay the onset of diabetes and improve impaired tolerance is to lose weight. While new drugs are effective for weight loss, patients often experience gastrointestinal side effects that make it challenging to achieve the target dose. This study investigated the effect of alpha-lipoic acid on the tolerability of GLP-1 agonists in overweight patients with prediabetes.

Materials and methods: The study included 18 patients with prediabetes between the ages of 37 and 50, averaging 45 ± 3.4 years. Eight men and ten women were selected for the study, with HbA1c levels between 5.9 and 6.4%, fasting glucose levels between 110 and 125 mg/dl, and BMI between 28 and 35kg/m². The patients were divided into two groups, and all were prescribed liraglutide to reduce body weight. The starting dose was 0.6, followed by titration of one dose. The drug’s tolerability was assessed using a special questionnaire that listed the main side effects, including headache, nausea, vomiting, diarrhea, and constipation. The first group received additional therapy of 1.2 U of alpha-lipoic acid 600 mg in the morning on an empty stomach, while the second group did not receive any further treatment.

Results: The results showed that the maximum dose of liraglutide 3.0 U in group I reached 44.45% (n = 8), and only 5.56% (n = 1) remained at 2.4 U. In group II, 16.7% (n = 3) reached 3.0 U, and 22.22% (n = 4) patients remained at 2.4 U. In 11.11% (n = 2), the dose stopped at 1.8 U. In terms of side effects, headache was lower in group I by 5.56% (n = 1), in group II by 33.33% (n = 6), nausea and vomiting in group I by 16.7% (n = 3), and 11.11% (n = 2), II in group 22.22% (n = 4) and 44.45% (n = 8), respectively, diarrhea in group I by 5.56% (n = 1) vs group II 16.7% (n = 3), constipation was observed only in group II 11.11% (n = 2). Body weight loss in group I was better at 5.1-7.2 kg compared to 3.5-4.9 kg in group II. Spearman correlation revealed a positive association between liraglutide dose and alpha-lipoic acid r = 0.58, p = 0.012. Among the side effects, a significant positive correlation with alpha-lipoic acid was observed only with vomiting r = 0.67, p = 0.002. The correlation with other side effects is positive but unreliable.

Conclusion: In conclusion, adding alpha-lipoic acid improves the tolerability of an adequate target dose of GLP-1 agonist, reduces the incidence of side effects, and increases the effect of reducing body weight.

Disclosure: L. Jashi: None.

152

Higher GLP-1R expression in the pituitary of males with type 2 diabetes compared to females determined by [ ⁶⁸ Ga]Ga-NODAGA-exendin-4 PET/CT imaging

S. Tokgöz ¹, L. Deden^2,3, R.I. Meijer⁴, C. Tack⁴, B.E. de Galan^4,5, M. Boss¹, M. Gotthardt¹;

¹Medical Imaging, Nuclear Medicine, Radboud University Medical Center, Nijmegen, Netherlands, ²Metabolic Surgery, Rijnstate Hospital, Arnhem, Netherlands, ³Vitalys, Elst, Netherlands, ⁴Internal Medicine, Radboud University Medical Center, Nijmegen, Netherlands, ⁵Internal Medicine, Maastricht University Medical Center, Maastricht, Netherlands.

Background and aims: Glucagon-like peptide-1 receptors (GLP-1Rs) expressed in the brain contribute to weight loss by stimulating satiety, upon binding of a GLP-1R agonist. GLP-1R agonist treatment is more effective in terms of weight loss in females compared to males. The hypothalamus-pituitary-gonadal axis is suggested to interact with the metabolic effects of GLP-1 analogues, but whether this interaction explains the different GLP-1 effects in men and women is unknown. To further explore sex dynamics in the brain, we used PET/CT imaging with the radiolabelled stable GLP-1 analogue, [⁶⁸Ga]Ga-NODAGA-exendin-4, to determine GLP-1R expression in the pituitary of men and women with type 2 diabetes.

Materials and methods: This analysis included 22 females and 13 males with type 2 diabetes. Age (55±9.8 vs 60±9.0 years), BMI (36±6.8 vs 32±6.0 kg/m²), fasting c-peptide (0.74 (0.41-1.5) vs 1.2 (0.48-1.3) nmol/L) and HbA_1c levels (61±12 vs 61±14 mmol/mol) were similar in men and women. All participants underwent a PET/CT scan 1 h after infusion with 100±5 MBq radiolabelled exendin. Tracer uptake of radiolabelled exendin, as a measure for GLP-1R expression, was determined in the pituitary. The maximum standardised uptake value (SUV_max [unitless]) of the PET scan was used to express tracer uptake. The SUV_max in the pituitary was correlated to descriptive data.

Results: We observed significantly higher uptake of radiolabelled exendin in the pituitary of males compared to females (SUV_max 2.9 (2.1-4.7) vs 5.3 (4.1-8.0), p=0.0054). Tracer uptake in the pituitary was positively correlated with BMI both in females (r=0.60, p=0.0031) and males (r=0.61, p=0.027). Although males had a relatively longer diabetes duration compared to females (16 (12-21,5) vs 8.5 (3.3-18.3) years, p=0.027), we did not observe a correlation between diabetes duration and tracer uptake in the pituitary (females r=-0.37, p=0.09; males r=-0.40, p=0.18).

Conclusion: Our data indicate sex differences in GLP-1R expression in the pituitary, with more pronounced pituitary uptake of radiolabelled exendin in men compared to women. To what extent our findings explain the differences in weight loss between men and women due to GLP-1R agonist treatment requires further study.

Disclosure: S. Tokgöz: None.

153

Serum GLP-1 and thrombospondin-2 predict improved metabolism and attenuated liver fibrogenesis in severely obese patients undergoing an intensive weight-loss therapy

N. Rohmann ¹, R. Surabattula², J. Koppenhagen¹, A. Beckmann¹, P. Wietzke-Braun¹, T. Hollstein^1,1, S.R. Myneni², K. Hartmann¹, L. Kruse¹, K. Schlicht¹, C. Geisler¹, K. Türk¹, D.M. Schulte^1,1, D. Schuppan^2,3, M. Laudes^1,1;

¹University Medical Center Schleswig-Holstein, Kiel, Germany, ²University Medical Center of the Johannes Gutenberg University, Mainz, Germany, ³Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA, USA.

Background and aims: Glucagon-like peptide-1 (GLP-1) is one of the most promising biological targets in weight loss therapy due to its involvement in the central regulation of food intake and satiety. Thrombospondin-2 (TSP-2) is a novel key serum fibrosis marker that reflects fibrosis and especially the dynamics of liver fibrogenesis in patients with fatty liver disease. We investigated serum GLP-1 and TSP-2 levels in association to intensive weight loss with special regard to liver inflammation and fibrogenesis upon lifestyle intervention.

Materials and methods: The Kiel Intervention Cohort is constituted of 180 patients with severe obesity who underwent a multimodal intensive interdisciplinary weight loss program for 6 months at a university medical center in Germany. At baseline, after 10 weeks of a very low calorie diet, and after another 16-week of a normo-caloric diet, serum concentrations of GLP-1 were measured using a valid commercial ELISA, and those of TSP-2 using a validated in-house ELISA for TSP-2. GLP-1 and TSP-2 were correlated with anthropometric and clinical laboratory parameters.

Results: A significant decrease of BMI (-16.8 ± 6.2 %, p < 10^-5) was achieved by the weight loss therapy, which significantly correlated with a decrease in TSP-2 (-14.32 ± 18.72 %, r_S = 0.218, p = 0.03). There was also a significant correlation between GLP-1 und TSP-2 at baseline (r_S = 0.225, p = 0.004). However, when assessing individual GLP-1 and TSP-2 courses, a lot of variation was noted, indicating the need for differentiated further analyses. Indeed, sex-specific differences were apparent with men having higher GLP-1 and TSP-2 values than women (both p < 0.01). Furthermore, when stratifying participants into quartiles of baseline GLP-1 (Overall range: 3.66 - 600.94 pg/ml), an approximation of very low and very high baseline GLP-1 but not consistent decrease or increase can be observed during the course of weight loss among these groups. In terms of functional and structural liver improvements, liver enzymes, stiffness assessed by liver elastography, and TSP-2 decreased similarly during weight loss. Still, when comparing these measures, deviations in risk estimation became apparent as reflected by the absence of a correlation between elastography and BMI.

Conclusion: Our results confirm the modulation of both the metabolic parameter GLP-1 and of obesity-related liver fibrogenesis via the matricellular fibrosis marker TSP-2 through multimodal weight loss therapy. Both parameters correlated only mildly with baseline BMI and efficient weight loss, indicating that improved metabolism will not always result in attenuated (liver) fibrogenesis. Further analyses will address the more complex association of fibrogenesis with body composition, metabolic status, and inflammatory activity.

Supported by: NR received DDG general project grant; DS received project-related grant by EU Horizon 2020 (project 777377 -LITMUS); NR/RS + DS/ML equal contribution

Disclosure: N. Rohmann: None.

154

Serotonin is a potent stimulator of GLP-1 secretion from the rat perfused intestine

T.A. Cookson ¹, I.M. Modvig¹, J.J. Holst^1,2;

¹Biomedical Sciences, University of Copenhagen, Copenhagen, Denmark, ²Faculty of Health and Medical Sciences, Novo Nordisk Foundation Center for Basic Metabolic Research, Copenhagen, Denmark.

Background and aims: Serotonin (5-HT) is involved in intestinal motility, fluid and electrolyte secretion, and visceral sensitivity. As 5-HT is secreted in response to mechanical pressure and nutrients, it serves as a robust physical and chemical sensor. Our aim was to investigate the effects of 5-HT on intestinal GLP-1 secretion in situ.

Materials and methods: Rats were anesthetized and either the entire small intestine or the colon were surgically isolated. For small intestine experiments, the blood supply to the colon was ligated and was excised. The small intestine was perfused through a catheter in the superior mesenteric artery. For colon experiments, the distal jejunum, ileum, cecum, colon, stomach, and spleen were excised, blood supply to the kidneys was ligated, and the colon was perfused through an aortic catheter. In both preparations, output was collected from the portal vein. Perfusions were performed using a single-pass system with a modified Krebs-Ringer bicarbonate buffer including the non-selective phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine to prime cells for hormone secretion. The buffer was pH-adjusted to 7.4, heated to 37°C, and continuously carbogenated with 95% O₂-5% CO₂. Arterial flow was clamped at 3mL/min or 7.5mL/min for colon and small intestine perfusions, respectively. Animals were euthanized and the organs were allowed to stabilize for 25 minutes. 5-HT was infused intravascularly at 0.1nM, 1nM, or 10nM for 5-minute periods. KCl (50 mM) was infused as a positive control at the end of all experiments. In separate experiments, 1nM and 100nM of GR113808 (selective 5-HT₄ antagonist), ondansetron hydrochloride (selective 5-HT₃ antagonist), and methiothepin mesylate (non-selective 5-HTR antagonist) were individually co-infused with 5-HT. Arterial pressure was monitored via transducer. Peristaltic motility index was calculated for the small intestine. GLP-1 and somatostatin were measured by radioimmunoassay.

Results: In the small intestine, 0.1nM, 1nM, and 10nM 5-HT increased secretion of GLP-1 (+58.4%, +91.2%, 48.2%; n=4, Figure), but not somatostatin. The motility index was significantly increased with 0.1nM, 1nM, and 10nM 5-HT (+54.1%, + 265.1%, +259.8%). In the colon, 0.1nM, 1nM, and 10nM 5-HT stimulated GLP-1 (+41.5%, +37.3%, +46.5%; n=4, Figure) and somatostatin (+70.7%, +53.1%, +54.5%; n=4) secretion. 5-HT antagonists did not blunt colonic GLP-1 secretion when applied individually.

Conclusion: 5-HT potently stimulates GLP-1 secretion from the small intestine and colon, as well as colonic somatostatin. The mechanisms of this remain unclear and may occur through multiple 5-HT receptor subtypes to provide redundancy to the actions of 5-HT.

Supported by: Dansk Frie Forskningsråd

Disclosure: T.A. Cookson: None.

155

Insulin resistance elicits GLP-1 secretion from pancreatic alpha cells in mice

G. Di Giuseppe ^1,2, S. Lind Jepsen², J. Vergara Ucin², D. Bjørklund Andersen², B. Hartmann², A. Giaccari¹, J.J. Holst³, T. Mezza¹;

¹Endocrinology and Diabetology, Fondazione Policlinico Universitario Agostino Gemelli IRCCS, Rome, Italy, ²Department of Biomedical Sciences, University of Copenhagen, Copenhagen, Denmark, ³NNF Centre for Basic Metabolic Research, University of Copenhagen, Copenhagen, Denmark.

Background and aims: It has been recently hypothesized that alpha cells are a potential source of intra-islet GLP-1, eventually exerting paracrine effects on islet function. Further, glucagon receptor knockout mice exhibit significant pancreatic 7-36 GLP-1 amide secretion and expression. However, it is still unknown whether intraislet GLP-1 production and secretion is modulated by insulin resistance and glucose dysmetabolism. In this study, we aim to explore the impact of insulin resistance on intra-islet proglucagon-derived peptides in a population of Diet-Induced Obese (DIO) mice.

Materials and methods: We performed a 75 min pancreatic perfusion on n=8 DIO vs. n=8 C57BL/6 control mice at low (3.5 mM) and medium (7 mM) glucose concentrations + 3.5 mM Vamine and 10 mM arginine stimulation; at the end, we infused insulin receptor antagonist S961 (n=4) or GLP-1 receptor antagonist Exendin 9 (n=4) to investigate possible involvement of insulin and GLP-1 receptor signaling. Total GLP-1 levels were measured using RIA; Insulin and Glucagon levels were measured through commercial ELISA kit (Mercodia).

Results: Compared to controls, DIO mice showed a significant secretion of total GLP-1 overtime during perfusions (p<0.0001), with no differences between 3.5 and 7 mM glucose and with a preserved response to arginine (Figure A). Further, S961 did not affect total GLP-1 secretion, whereas Exendin 9 led to increased intra-islet total GLP-1 secretion in DIO mice, as in a positive feedback-control mechanism (Figure B); no effects were detected in control mice. As expected, DIO mice showed increased insulin secretion at 7 mM glucose and during arginine stimulation, while no differences were observed in the glucagon secretion. S961 and Exendin 9 did not affect insulin nor glucagon secretion in the 2 groups.

Conclusion: Our data demonstrate that intraislet GLP-1 secretion is significantly increased in a rodent model of obesity and insulin resistance. In particular, this secretion pattern seems to be regulated by the activity of GLP-1 receptor itself. Further studies will be necessary to clarify the pathophysiologic action of intrapancreatic GLP-1 on insular environment, as well as its possible pharmacological implications.

Supported by: EFSD Future Leaders Mentorship Programme for Clinical Diabetologists supported by AstraZeneca

EFSD Albert Renold Travel Fellowship Programme

EASD Mentorship Programme 2023

Disclosure: G. Di Giuseppe: None.

156

Lipidated interleukin-22 reduces body weight through reduced food intake and increased fecal energy loss in a complementary manner to GLP-1 receptor agonists

M. van de Bunt ¹, D. Haß², M. Rohm², A. Kjølbye¹, R. Jorgensen¹;

¹Cytoki Pharma, Søborg, Denmark, ²Helmholtz Center Munich, Neuherberg, Germany.

Background and aims: The atypical cytokine interleukin-22 (IL-22), which acts on epithelial cells, has shown promise as a novel treatment option that would improve long-term outcomes for patients with obesity and diabetes. Here we demonstrate the mode of action underlying the body weight reduction of a long-acting lipidated IL-22 agonist and its combination potential with existing weight loss therapies.

Materials and methods: The effect of lipidated IL-22 on obesity at 2 dose levels was investigated in two diet-induced obese (DIO) models - one at thermoneutrality to characterise the mode of action underlying the weight loss effect, and one at standard conditions to determine the combination potential with a GLP-1 receptor agonist.

Results: Lipidated IL-22 and 30 nmol/kg semaglutide dosed for 5 weeks in DIO mice led to significantly greater reduction in body weight (-18.9% and -28.5%) compared to vehicle (-3.1%). Mice treated with a combination of both compounds from study start reduced body weight by -40.4%, demonstrating clear additivity for the two modes of action. This was further supported by 2 groups that were treated with semaglutide for the first 2 weeks before being switched to either a combination of semaglutide and lipidated IL-22 or lipidated IL-22 alone. The switch combination group showed weight loss (-38.2%) in line with the group that received the combination from study start. The group switched to lipidated IL-22 alone did continue to lose -2.9% weight over the next 3 weeks, albeit at a slower pace than the group receiving semaglutide alone for the entire study duration. To establish how lipidated IL-22 induces weight loss, animals were treated with 2 dose levels of lipidated IL-22 for 4 weeks at thermoneutrality. Lipidated IL-22 led to dose-dependent reductions in food intake, which previous studies with weight-matched control have shown to be insufficient to fully explain the body weight loss. Energy expenditure measurement through indirect calorimetry at weeks 1 and 4 found no increase in energy expenditure for lipidated IL-22. Bomb calorimetry showed a significant dose dependent increase in 24h fecal energy content on day 15 in the lipidated IL-22 groups (17.1 and 18.1 kJ/g) compared to vehicle animals (14.6 kJ/g). Fecal energy loss relative to energy intake through food therefore increased from 6.5% in vehicle animals to 10.0% and 12.8% in the lipidated IL-22 groups. The reduction in body weight in both studies were accompanied by significant dose dependent improvements in fasting plasma insulin, lipids, and liver enzymes.

Conclusion: Long-acting lipidated IL-22 significantly reduces body weight through a novel intestinal mode of action that is complementary to incretin-based therapies. The body weight reduction is accompanied by direct benefits on key drivers of cardiovascular outcomes.

Disclosure: M. van de Bunt: Employment/Consultancy; Cytoki Pharma.

OP 27 Predictors of diabetic kidney disease progression from omics to ethnicity

157

Redefining the eGFR-albuminuria relationship using genetics: setting stage for precision monitoring

C.H. Naaman ¹, T.W. Hansen^1,2, F. Persson¹, P. Rossing^1,2, T.S. Ahluwalia^1,2;

¹Steno Diabetes Center Copenhagen, Herlev, Denmark, ²University of Copenhagen, Copenhagen, Denmark.

Background and aims: Chronic kidney disease (CKD) poses a global challenge and a financial burden with susceptibility to other chronic vascular complications especially among persons with type 2 diabetes (T2D). Early indicators of CKD based on Kidney Disease Improving Global Outcomes (KDIGO) guidelines include dependency on kidney function markers a. estimated glomerular filtration rate (eGFR) and b. albuminuria (ALB). Albeit multiple genetic loci for these two traits have been known, the interrelationship between common loci is incompletely understood. Mapping eGFR-ALB interrelationship using genes will unravel target mechanisms and help develop effective CKD monitoring and treatment options. The current study aims to map the genetic basis of eGFR-ALB relationship.

Materials and methods: 432,292 individuals of mainly European descent (94%) with information on genetic and phenotypic data from the UK Biobank (UKB) population were included. Phenotypic data included sex, age, urinary albumin, serum creatinine, and systolic blood pressure. T2D was defined in 19,203 individuals. eGFR was calculated using the CKD-EPI creatinine equation from 2021. Data preprocessing involved outlier correction, removal of missing values and normalization. ALB was log transformed. Multiple linear regression was employed to a. validate known eGFR loci using the race free eGFR 2021 equation in the entire UKB population and T2D subgroup, and b. to investigate the relationship between eGFR loci and ALB levels (all and T2D subgroup), considering significance at p ≤ 0.05 before and after multiple testing correction.

Results: Among 424 known eGFR SNPs tested, 367 associated with eGFR (p ≤ 0.05). 185 of 367 also associated with albuminuria before (p ≤ 0.05), and 54 after bonferroni correction. The top associations included SHROOM3 locus rs28817415 T-allele, with eGFR (Beta = -0.38, p = 2.4 × 10^-108) and ALB (Beta = -0.02, p = 1.2 × 10^-23). This denotes a strong association with lower eGFR and ALB. PDILT gene rs77924615 A-allele associated with high eGFR (Beta = 0.57, p = 6.3 × 10^-240) and low ALB (Beta = -0.011, p = 1.6 × 10^-7). While some loci demonstrated a consistent negative or positive impact on both eGFR and ALB, others exhibited a negative-positive or positive-negative relationship. In the T2D population, 173 loci associated with eGFR (p ≤ 0.05), before and 42 after bonferroni correction. 15 loci associated with ALB before and one after bonferroni correction, specifically the PLEKHA1 gene locus rs7088058 T-allele (eGFR_creatinine, Beta = 0.17, p = 3.4 × 10^-22; ALB, Beta = -0.037, p = 1.7 × 10^-4).

Conclusion: Our study elucidates the complex interplay between eGFR and ALB loci, revealing diverse relationships characterized by varying effects. These findings underscore the multifaceted and pleiotropic nature of genetic loci in shaping the eGFR-ALB relationship for the general and T2D populations. These differences in their relationship can have important implications for the patient risk monitoring and require further investigation, also in relation to diabetes and metabolic status. The current study validates 185 eGFR loci (15 in T2D population) and identifies its association and interrelationship with ALB. These insights provide a foundation for future research utilizing Polygenic Risk Score-based precision studies, which potentially can lead to targeted interventions and personalized treatments for CKD.

Supported by: Novo Nordisk Foundation (Grant Number NNF18OC0052457)

Disclosure: C.H. Naaman: Grants; NNF18OC0052457.

158

Impact of ethnicity on faster progression of kidney function loss and onset of advanced chronic kidney disease in an ethnically diverse cohort of people with type 2 diabetes

T. Alobaid ¹, J. Karalliedde², S. Ayis³, M. O’Connell¹;

¹King’s College London, London, UK, ²King’s College London, London, UK, ³School of Population Health & Environmental Sciences, King’s College London, London, UK.

Background and aims: People with type 2 diabetes who develop chronic kidney disease (CKD) can often have differing and heterogeneous patterns/rates of progression to advanced stages of CKD. Our aim was to investigate eGFR trajectories in a large cohort of ethnically diverse urban people with type 2 diabetes (n= 14572, non-Caucasian 54.7 %) to study the impact of ethnicity in people with faster progression of CKD and to identify related risk factors associated with faster loss of kidney function.

Materials and methods: A total of 14,572 people with type 2 diabetes attending routine secondary care outpatient clinics in two large teaching hospitals with baseline eGFR >45 ml/min/1.73m2 were studied from 2004 to 2018 (median follow up of 4 years), The cohort was ethnically diverse with 5 different ethnicity categories: 45% White, 37.8% Black, 9.1% Asians, 2.7% mixed, and 5.2% others. Group Based Trajectory models (GBTM) were used to explore various patterns of progression of eGFR over duration of follow up in each of the main five ethnic groups. Identification of distinct trajectories was based on 4 criteria, 1) Highest Entropy, 2) Lowest Bayesian Information Criteria (BIC), 3) Low Akaike Information Criteria (AIK), and 4) meaningful composition of participants. Multinomial logistic regression was used to compare clinical/biochemical features associated with faster progression to stage CKD 5 (end stage kidney disease) between ethnic group with the majority (White) group treated as the reference.

Results: Of the 14,572 people with type 2 diabetes mellitus, 17.4% (n=2,531, 62.5% Male) demonstrated fastest progression of eGFR loss (defined as the trajectory group with greatest profile of eGFR decline over time). This fastest progression group, comprised, white ethnicity (19.4% n=1279) has mean (standard deviation) age 66.4(10) years and 33.9% female. Similarly, 15.1% (n=820) were of Black ethnicity with mean age 63.1 (11.4) years and 43.8% female, 18.0% Asian (n=239) with mean age 64.3 (10) and 35.2% female. There was 19.9% (n=78) of mixed ethnicity with mean age 62.2 (11.5) years and 39.7% female, and 15.0% (n=115) other ethnicity with mean age 64.2 (10.8) years and 35.75 % female. Baseline eGFR was 56.67, 55.85, 56.33, 57.44, 58.52 ml/min/1.73m² for the 5 ethnic groups respectively. Of this group of 2531, we observed that n=368 (14.5%) people progressed to stage 5 CKD during follow up. When we compared the clinical features, of this highest risk group, by differing ethnic groups we observed, after adjusting for age, blood pressure, albuminuria, HbA1c, gender, and baseline eGFR or CKD stage, that those reaching stage 5 CKD differed significantly by ethnicity. The relative risk ratio (95% confidence interval) of reaching CKD stage 5, after adjusting for the above traditional risk factors was 1.64 (1.22-2.21, p<0.05) for the Black; 2.39 (1.57-3.64, p<0.05) for the Asian; 1.97 (2.11-6.57, p<0.05) the mixed; and 3.72 (2.11-6.57, p<0.05) for the others, compared to the White group.

Conclusion: These results emphasize the need for tailored interventions and healthcare strategies that address the specific risk profiles of different ethnic groups and emphasize the importance of early detection and prompt management of modifiable risk factors to mitigate the rapid progression of CKD and its associated morbidity and mortality.

Disclosure: T. Alobaid: None.

159

Proteomic changes prior to onset and progression of albuminuria and end-stage kidney disease in type 1 diabetes

S. Mutter ^1,2, E. Valo^1,3, E. Dahlström^1,2, V. Harjutsalo^1,3, P.-H. Groop^1,4, N. Sandholm^1,3, FinnDiane Study Group;

¹Folkhälsan Research Center, Helsinki, Finland, ²Department of Nephrology, University of Helsinki and Helsinki University Hospital, Helsinki, Finland, ³Research Program for Clinical and Molecular Metabolism, Faculty of Medicine, Helsinki, Finland, ⁴Department of Diabetes, Central Clinical School, Melbourne, Australia.

Background and aims: Biomarker studies have identified circulatory proteins associated with diabetic kidney disease (DKD) and its progression. New high-throughput proteomics allows for extensive characterization of the serum proteome at an unprecedented scale, facilitating molecular discoveries and improving the identification of individuals at DKD risk.

Materials and methods: This FinnDiane Study followed 823 adults (61% men) with type 1 diabetes (T1D) for 9±5 years. At baseline, 5,415 serum proteins were measured (Explore HT proximity extension assay) and normalized by intensity normalization. Samples that did not pass sample or assay quality control were excluded. We analyzed 3 subsets: 448 adults with normal AER (<30mg/24h) of whom 164 developed albuminuria (≥30mg/24h), 150 adults with moderate albuminuria (30-300mg/24h) of whom 61 progressed to severe albuminuria (>300mg/24h), and 225 adults with severe albuminuria of whom 99 progressed to end-stage kidney disease (ESKD). In each subset, (unadjusted) P-values were based on t-tests or Mann-Whitney U tests. For the 5,073 proteins with normal distribution, ANOVA-based P-values were adjusted for sex, age, diabetes duration, estimated glomerular filtration rate, sample storage time, and number of sample thaw cycles. All P-values were adjusted using the false discovery rate (FDR).

Results: At baseline, the mean age was 41 years, mean HbA_1c 8.5%, and in ANOVA 20% of proteins differed between those with and without normal AER (top signals were MMP7, associated with kidney fibrosis in mice, and renin [both P_ANOVA=1.6×10^-31]). The baseline levels of 70 proteins (67 in ANOVA) were associated with incident albuminuria (Figure 1). Top signals were monocyte differentiation antigen CD14 (P_ANOVA=8.1×10^-6) and kidney injury molecule-1 (KIM-1) (P_ANOVA=1.8×10^-5). KIM-1 (P_ANOVA=0.001) was also among the top 3 signals for the progression to severe albuminuria together with PCSK9 (P_ANOVA=0.001), and ASGR1 (P_ANOVA=3.5×10^-5). While PCSK9 is a well-known lipid-lowering target, targeting ASGR1 in mice has also been shown to lower cholesterol. Finally, 21% (4% in ANOVA) of the proteins were associated with progression to ESKD, with the endoplasmic reticulum stress-induced CRELD2 as the top signal (P_ANOVA=3.1×10^-5).

Conclusion: In T1D we observed major changes in the proteome prior to the onset and progression of albuminuria and ESKD. We aim to leverage them to identify individuals at risk and to get molecular DKD insights. Figure 1. Proteome differences for 448 adults with T1D and normal AER of whom 164 developed albuminuria. P-Values from t-tests or Mann-Whitney U tests and FDR adjusted.

Supported by: Novo Nordisk (NNF23OC0082732), Folkhälsan Fnd, Stockmann Fnd, Academy of Finland, Juselius Fnd, Liv och Hälsa Soc, Med. Soc of Finland, Gov Res. Fund

Disclosure: S. Mutter: None.

160

Association between serum REG Iα and progression of diabetic kidney disease: a prospective cohort study

N. Huang ^1,2, S. Chen^1,2, Z. Shu^1,2, L. Li^1,2;

¹Department of Endocrinology, Southeast University Zhongda Hospital, Nanjing, China, ²Pancreatic Research Institute, Southeast University School of Medicine, Nanjing, China.

Background and aims: Regenerating protein Iα (REG Iα) is a 16 kDa protein primarily secreted by the pancreas. The levels of serum REG Iα are elevated in Type 2 diabetes mellitus patients with renal impairment, and serve as a novel biomarker for the screening of diabetic kidney disease (DKD). To date, limited studies have revealed the relationship between serum REG Iα and disease progression in DKD. The aim of this study is to validate the correlation between baseline serum REG Iα levels and the progression of renal dysfunction in patients with DKD.

Materials and methods: A prospective cohort study was carried out in an academic medical center, and followed up from October 2021 to January 2024. The levels of serum REG Iα were quantified by enzyme-linked immunosorbent assays. Demographics information and clinical biochemical parameters such as fast blood glucose (FBG), serum creatinine (Scr), blood urea nitrogen (BUN), uric acid (UA), estimated glomerular filtration rate (eGFR) and urine albuminuria creatine ratio (UACR) were collected at baseline and follow-up time. The kidney progression events in patients with DKD were defined as a decline of 30% or more in eGFR, or an increase of twofold or more in UACR, or progression to end-stage kidney disease, or death due to renal causes. Correlations between serum REG Iα and renal function were presented by spearman’s correlation analyses. Cox regression analyses were used to explore the influencing factors of renal function decline. Interaction analyses were used to adjust the potential confounding factors. Kaplan-Meier survival analysis was carried out to compare the incidence of DKD progression.

Results: A total of 456 patients with DKD were enrolled in this study, divided into quartile groups according to levels of baseline serum REG Iα, Q1 (<29.39 ng/mL), Q2 (29.39ng/mL-76.05 ng/mL), Q3 (76.06ng/mL-177.73 ng/mL), Q4 (>177.73 ng/mL). The median follow-up duration was 16.10±6.30 months. There were 99 (21.71%) patients reached the kidney progression events in total, 12 (10.53%) in group Q1, 11 (9.65%) in group Q2, 29 (25.44%) in group Q3, and 47 (41.23%) in group Q4 (p<0.001) . No differences were found in age, gender, hypertension, and cardiovascular disease (CVD) between these four groups (all p>0.050). The levels of serum REG Iα gradually increased as renal function declined progressively (p<0.001). Serum REG Iα levels were negatively associated with eGFR at baseline (R²=0.321, p<0.001) and at the endpoint of follow-up (R²=0.258, p<0.001). In multivariate cox regression models, levels of baseline serum REG Iα were significantly associated with an increased risk of DKD progression, after adjusting for potentially confunding factors, including age, gender, hypertension, CVD, FBG, Scr, BUN, and UACR. Comparing the levels of serum REG Iα in group Q1 as a reference, the hazard ratio (HR) for group Q2 was 1.071 (95%CI: 0.355-3.227, p=0.903), HR for group Q3 was 2.522 (95% CI: 1.098-6.370, p=0.040), and HR for group Q4 was 4.750 (95% CI: 1.883-11.985, p=0.001). Kaplan-Meier survival curves demonstrated that patients with higher levels of baseline serum REG Iα (>76.05ng/mL) were more likely to experience kidney progression events (Log-rank χ²=22.690, p<0.001).

Conclusion: The serum REG Iα is significantly associated with the progression of diabetic kidney disease, and may serve as a novel predictor for renal function decline in patients with DKD.

Clinical Trial Registration Number: ChiCTR2300072247

Supported by: NSFC (No. 81970717, No. 82000740, No. 82170845 and No. 82200929);

Disclosure: N. Huang: None.

161

Albuminuria trajectories over six years and their determinants in type 2 diabetes: the Fremantle Diabetes Study Phase II

W.A. Davis, T.M.E. Davis;

Medical School, University of Western Australia, Fremantle, Australia.

Background and aims: Urinary albumin:creatinine ratio (uACR) is a variable that can exhibit significant temporal changes, but few studies have characterised transition patterns between categories (normo-, micro- and macroalbuminuria) and their determinants in type 2 diabetes. The aim of this study was to use group-based trajectory modelling (GBTM) to identify clusters of people exhibiting distinct trajectories of uACR category in a community-based cohort.

Materials and methods: We analysed longitudinal data from the observational Fremantle Diabetes Study Phase II (FDS2) which recruited representative participants between 2008 and 2011. GBTM identified discrete groups of participants with type 2 diabetes and ≥2 biennial uACR measurements over 6 years. Independent baseline associates of group membership were assessed using multinomial regression.

Results: Of 1482 FDS2 participants with type 2 diabetes, 1145 (77.3%; baseline mean age 65.4 years, 53.3% males, median diabetes duration 8.0 years, geometric mean uACR 3.2 mg/mmol) had ≥2 biennial uACR measurements and were included in the present analyses. The optimum GBTM model comprised 5 categories: “normal”, “regressing”, “progressing”, “microalbuminuria”, and “macroalbuminuria” (see Figure for plots of individual serial changes by GBTM group). The stable normoalbuminuria group was the largest (>1/3 of participants). Two of equal size (≈10%) had regressing and progressing patterns (normal from/to microalbuminuria). The second largest group (also >1/3 of participants), centred on microalbuminuria, had substantial intra-individual variability in uACR over time. The stable macroalbuminuria group was the smallest. Compared with membership of the normal group, being Australian Aboriginal significantly increased the odds of macroalbuminuria group membership (OR (95% CI): 8.20 (2.42, 27.8)). For each decade of age, the odds of not being in the normal group increased by 30-40% across all other groups. Females were twice as likely as males to be in the regressing vs normal group. For each 5-year increment in diabetes duration, the odds of micro- or macroalbumnuria group membership increased by 24% and 37%, respectively. HbA1c, serum triglycerides, hypertension, and cardiovascular disease were increasingly associated with membership of the regressing to macroalbumniuria groups. Stage 3B chronic kidney disease or worse was only significant for the microalbuminuria and macroalbuminuria groups, substantially increasing the odds of membership vs the normal group (3.06 (1.28, 7.32) and 8.71 (2.89, 26.3), respectively).

Conclusion: People with type 2 diabetes exhibiting uACR fluctuations centred on microalbuminuria appear phenotypically on the pathway to persistent macroalbuminuria and its adverse sequelae. Early introduction of renoprotective therapies such as SGLT2 inhibitors in the regressing and progressing groups may change the natural history of nephropathy in these individuals.

Supported by: Australian National Health and Medical Research Council Project Grants

Disclosure: W.A. Davis: None.

162

HbA _1c variability is independently associated with progression of diabetic kidney disease in people with type 1 diabetes

A. Muthukumar ¹, L. Badawy¹, A. Mangelis¹, P. Vas², S. Thomas², A. Gouber¹, S. Ayis¹, J. Karalliedde^1,2;

¹King’s College London, London, UK, ²Guy’s and St. Thomas’ NHS Foundation Trust, London, UK.

Background and aims: The role of HbA1c-variability on the progression of diabetic kidney disease is unclear with most studies in Caucasian populations and limited data on its role in predicting advanced kidney outcomes. Our aim was to evaluate if long term intra-individual HbA1c-variability is a risk-factor for endpoint of kidney disease-progression (defined as estimated glomerular filtration rate (eGFR) decline ≥50% from baseline with final eGFR <30 ml/min/1.73m²) in an ethnically heterogeneous cohort of people with type 1 diabetes, with preserved eGFR ≥45 ml/min/1.73m² at baseline.

Materials and methods: Electronic health records (EHR) data from people attending out-patient clinics between 2004-2018 in two large university hospitals in London was collected. HbA1c-variability assessed by three distinct methods: 1. HbA1c standard deviation (SD), 2. Visit-adjusted SD: SD-HbA1c/√n/(n-1), ‘n’ is the number of HbA1c-measurements per patient, 3. coefficient of variation (CV)- SD-HbA1c/mean-HbA1c (HbA1c-CV). All people had ≥6 follow-up HbA1c measurements. eGFR was measured using CKD-Epi equation and clinical/biochemical results from routine care extracted from EHR.

Results: In total 3466 (50% Female, 78% Caucasian, 13% African-Caribbean, 6% Other ethnicity) people were followed for median (interquartile range) of 8.2 (4.2 - 11.6) years. Of this cohort, 249 (7%) developed kidney disease progression. Higher HbA1c-variability (as measured by all three methods) was independently associated with higher risk of kidney disease progression with hazard ratios (HR) and (95% confidence interval) of 7.76 (4.54-13.26), 2.62 (1.75-3.94), 5.46 (3.40-8.79), (lowest versus highest HbA1c-variability quartile). Increasing age, baseline HbA1c, systolic blood pressure, and urinary albumin-creatinine ratio were also associated with kidney disease progression (p<0.05 for all). African-Caribbean ethnicity was associated with increased risk of kidney disease progression HRs (95% CIs) of 1.47 (1.09-1.98), 1.76 (1.32-2.36), 1.57 (1.17-2.12), across all three methods for HbA1c variability respectively and this effect was independent of glycaemic variability and other traditional risk factors.

Conclusion: We observed an independent association between HbA1c-variability, evaluated by three distinct methods, and significant kidney disease progression, in a multi-ethnic type-1 diabetes cohort. Further studies are needed to elucidate the mechanisms that may explain our results and evaluate if HbA1c variability is a modifiable risk factor to prevent kidney disease progression.

Supported by: Guys and St Thomas Charity

Disclosure: A. Muthukumar: None.

OP 28 New kids on the block: stem cell islets

163

Metabolic maturation of stem cell derived beta cells upon murine engrafment

E. Vähäkangas ¹, T. Barsby¹, H. Montaser¹, S. Eurola¹, P. Katajisto¹, J. Saarimäki-Vire¹, T. Otonkoski²;

¹University of Helsinki, Helsinki, Finland, ²Biomedicum Helsinki, University of Helsinki, Helsinki, Finland.

Background and aims: Stem cell-derived pancreatic islets (SC-islet) may offer a solution for the limited availability of donor islets for cell therapy of type 1 diabetes. However, as we and others have previously reported, SC-islets are not a perfect facsimile of human islets, especially in their metabolism, despite their dynamic insulin secretion capability. We thus aimed to study whether this immature metabolic phenotype persists upon protracted engraftment in mice.

Materials and methods: This study was conducted by differentiating the H1 human embryonic stem cell line into pancreatic islets (SC-islets) using a well-established seven-stage differentiation protocol. SC-islets were implanted under the kidney capsule of immunocompromised NSG mice for 1 to 4 months. Metabolic and morphological assays of SC-islets pre- and post- implantation were performed (in parallel to cadaveric donor islets) using LC-MS-metabolomics, electron microscopy, immunohistochemistry analyses and dynamic insulin secretion assays.

Results: SC-islets have the capability of dynamically controlling mouse blood glucose levels by 3 months post-implantation. Murine engraftment leads to maturation of various metabolic and morphological aspects of SC-islets, leading them to resemble human donor islets more closely. These shifts occurred early in engraftment (1 month) and included increases in insulin granule crystallisation, TCA cycle metabolic activity, mitochondrial number, and lowered reactivity to pyruvate. Increased insulin granule crystallization can be seen by the percentage of non-crystallised (immature) insulin granules lowering from 54% in vitro to 30% at 1 month post-engraftment (p = <0.001). Glucose-sensitive TCA cycle activity increased in high glucose concentrations, with the enrichment of labelled carbons into citrate increasing from 12.6% in vitro to 21.8% 1 month post-engraftment (p = 0.04). Mitochondrial number increased from a mean of 0.38 mitochondria/μm² in vitro to 0.59 following 1 month of engraftment (p = 0.001). Conversely, changes in mitochondrial morphology were cell-specific and correlated more with the insulin granule crystallisation status of a given cell than the age of the SC-islet sample. SC-beta cell reactivity to pyruvate is higher than to glucose in vitro, this aberrant reactivity lowers significantly upon implantation. Additionally, some differences to human islets persisted even after 4 months of SC-islet implantation. These discrepancies included the lack of glucose-sensitivity of anaplerotic TCA cycle activity, and the bias of metabolites derived from glutamate.

Conclusion: An in vivo environment is beneficial for the metabolic maturation of SC-islets, leading to them more closely resembling human donor islets than prior to engraftment. Thus, we believe that the aberrant metabolite trafficking pathways seen in SC-islets are rapidly and robustly lessened following engraftment. These data will aid in ongoing safety and efficacy studies of SC-islet maturation in diabetic patients that are receiving SC-islet cell replacement therapies.

Disclosure: E. Vähäkangas: None.

164

Pro-inflammatory stimuli induce profound transcriptome and translatome changes in iPSC-derived beta cells that mimic type 1 diabetes beta cell gene signatures

C. Vinci ¹, F. Szymczak², S. Thomaidou³, H. Shakeri³, Y. Tong⁴, I.M. Akçay¹, X. Yi⁵, A. Zaldumbide⁶, D. Eizirik⁷, M. Cnop⁸;

¹Université libre de Bruxelles, Brussels, Belgium, ²ULB Center for Diabetes Research, Université Libre de Bruxelles, Anderlecht, Belgium, ³⁴ULB Center for Diabetes Research, Brussels, Belgium, ⁵Faculty of Medicine, ULB (Université Libre De Bruxelles) Center for Diabetes Research, Brussels, Belgium, ⁶Molecular Cell Biology - Virus and Stem Cell Biology, Leiden University Medical Center, Leiden, Netherlands, ⁷Dept. of Experimental Medicine, Free University Brussels (ULB), Brussels, Belgium, ⁸ULB Center for Diabetes Research, Universite Libre de Bruxelles, Brussels, Belgium.

Background and aims: Human induced pluripotent stem cells (iPSCs) differentiated into β-like cells represent an exciting model to investigate type 1 diabetes (T1D) pathogenesis. The unlimited source of such “neonatal” iPSC-β cells enables comprehensive omics studies. Here, we characterized the iPSC-β cell response to inflammatory stimuli by analyzing transcriptome and translatome.

Materials and methods: iPSCs were differentiated into β cells and purified by MACS using the surface marker CD49a. Cells were exposed for 24h to cytokines involved at different stages of T1D (IFNα, IFNγ, IL1β, TNFα, IFNγ+IL1β). RNA-seq was used to chart transcriptomes, and Ribosome profiling was used to uncover the translatome after IFNγ+IL1β treatment.

Results: MACS purification generated 80±7% pure iPSC-β cells. IFNγ+IL1β had the greatest impact on the β cell transcriptome (5657 differentially expressed genes); among the individual cytokines, IFNs had a greater impact (IFNy 2993 and IFNα 500 differentially expressed genes). In all 3 conditions, pathways related to β cell function were downregulated (regulation of insulin secretion, regulation of slit/robo signaling and respiratory electron transport). At the translatome level, RIBO-seq showed decreased translation of genes related to Ca²⁺ dynamics and insulin secretion, including CALB2, MRLN, KCNK16. iPSC-β cell glucose-stimulated insulin secretion was inhibited by cytokines, from 3-fold in control to 1.5-fold in IFNα and IFNγ+IL1β condition, without changes in insulin content. An increase of translation initiation in iPSC-β after exposure to IFNγ+IL1β was identified for 290 genes, mostly belonging to immune reaction pathways. When comparing transcriptomes of cytokine-exposed iPSC-β cells with FACS-purified β cells from T1D patients, high overlap was seen in particular for IFNα and IFNγ that shared respectively 2163 and 1216 genes with T1D, while all the other conditions shared less than 1000 genes.

Conclusion: The high transcriptome overlap between T1D β cells and cytokine-exposed iPSC-β cells validates the model. Transcriptome, translatome and functional analyses suggest a preferential translation of genes which increase visibility of β cells to the immune system at the expense of β cell function.

Supported by: INNODIA, WIN2WAL

Disclosure: C. Vinci: None.

165

Recruitment of thymic regulatory T cells to stem cell-derived beta cells expressing CCL22

E. Filatov ^1,2, S. Sasaki^2,3, M. Haque², S. Kar^2,4, V. Fung², M. Mojibian², N.A.J. Krentz^1,2, K.N. MacDonald², G. Soukhatcheva², D. Dai², P.C. Orban², M.K. Levings^2,5, C.B. Verchere^2,4, F.C. Lynn^2,5;

¹Cell and Developmental Biology, University of British Columbia, Vancouver, BC, Canada, ²Diabetes Research Group, BC Children’s Hospital Research Institute, Vancouver, BC, Canada, ³Department of Surgery, University of British Columbia, Vancouver, BC, Canada, ⁴Department of Pathology and Laboratory Medicine, University of British Columbia, Vancouver, BC, Canada, ⁵School of Biomedical Engineering, University of British Columbia, Vancouver, BC, Canada.

Background and aims: Islet transplantation can provide insulin independence for persons with type 1 diabetes (T1D); however, the procedure requires lifelong immunosuppression, which increases risk of infections and is cytotoxic to the transplanted cells. Regulatory T cells (Tregs) have the potential to attenuate rejection of transplanted β-cells by suppressing allo- and auto-reactive T cells. Tregs are recruited to sites of inflammation by secreted chemokines and cytokines such a CCL22. We hypothesized that transplanted stem cell-derived β-cells (sc-βcells) engineered to express the chemokine CCL22 would cause Treg migration to the graft and attenuate graft rejection.

Materials and methods: Human embryonic stem cells (hESCs) were genetically modified to express CCL22 (CCL22-GFP) and differentiated into sc-βcells using a 7-stage protocol. CCL22-GFP sc-βcells were characterized in vitro and compared to control INS-GFP sc-βcells that do not express CCL22. CCL22-GFP sc-βcells were characterized in vivo through transplantation under the kidney capsule of immunodeficient (NSG) mice made hyperglycemic by streptozotocin (STZ). The bioactivity of CCL22 secreted from cells was tested in vitro by Treg Transwell migration assay and in vivo by infusing luciferase-expressing human thymic Tregs into NSG mice that received CCL22-GFP sc-βcell transplants in the anterior chamber of the eye.

Results: CCL22 expression did not adversely impact the differentiation of hESCs into sc-βcells, as assessed by gene expression analysis and immunostaining for markers of sc-βcell maturity. STZ-diabetic mice that received CCL22-GFP sc-βcell transplants achieved normoglycemia, demonstrating function of CCL22-GFP sc-βcells. Tregs migrated towards CCL22-containing media collected from CCL22-GFP stem cells, and this chemotaxis was repressed when the CCL22 receptor, CCR4, was blocked. Finally, mice that received CCL22-expressing sc-βcells had nearly double the luciferase activity at their graft sites (P < 0.01), signifying Treg recruitment to CCL22-GFP sc-βcell grafts. Treg recruitment to CCL22 expressing sc-βcell grafts was confirmed by FoxP3 immunostaining.

Conclusion: Transplantation of CCL22-expressing sc-βcells allows diabetic mice to regain normoglycemia and induces Treg migration to the graft site. Future work will assess whether Treg recruitment to CCL22-expressing sc-βcells confers protection of grafts from destruction by T cells in models of T1D.

Supported by: Stem Cell Network: AWD-014029 STECELNE 2019; FY18/DT15; FY17/DT4, BCCHR, Health Research BC, JDRF, Manpei Suzuki Diabetes Foundation, CIHR, CIRTN

Disclosure: E. Filatov: None.

166

Knock-out of the activating ligand repertoire of human stem cell-derived beta cells avoids chronic allogeneic rejection by NK cells in a diabetic humanised IL-15 NOG mouse model

G. Siracusano ¹, R. Chimienti¹, M. Certo¹, C. Volpi¹, G. Lunetta¹, L. Perani², A. Palmisano³, T. Canu², R. Melzi⁴, F. Deambrogio⁴, V. Sordi⁵, M. Malnati⁶, L. Piemonti¹;

¹Diabetes Research Institute, Vita-Salute San Raffaele University, Milan, Italy, ²Experimental Imaging Center, San Raffaele Scientific Institute, Milan, Italy, ³Experimental Imaging Center, Vita-Salute San Raffaele University, Milan, Italy, ⁴Diabetes Research Institute, San Raffaele Scientific Institute, Milan, Italy, ⁵Diabetes Research Institute (DRI), San Raffaele Scientific Institute, Milan, Italy, ⁶Division of Immunology, Transplantation and Infectious Disease, San Raffaele Scientific Institute, Milan, Italy.

Background and aims: Natural Killer (NK) cells represent an heterogenous population of lymphocytes that regulate several aspects of the alloimmune response in solid organ transplantation. Reactivity of NK cells depends on the repertoire of ligands expressed by target cells. Recently, MHC-I^-/- stem cell-derived β cells have been proposed as an alternative source for Type 1 Diabetes (T1D) treatment. Although MHC-I abrogation prevents T cell attack, it triggers NK cells via missing self-recognition, leading to graft loss. As we demonstrated that the NK activating ligands CD276 and CD155 mediate MHC-I-null killing, we propose their knock-out to escape NKs and ensure long-term survival and function of transplanted β cells in diabetic humanized mice.

Materials and methods: We generated luciferase-expressing wild type (WT), MHC-I^-/- and B2M/CD276/CD155 triple knock-out (T-KO) human induced pluripotent stem cells (iPSC), then differentiated them into β cells (iBeta). We transplanted iBeta in the hindlimb muscles of hIL-15 NOG mice humanized with donor-derived NKs. Survival of iBeta was evaluated up to 8 months by in vivo imaging system. Graft morphology and perfusion of the area were investigated by magnetic resonance imaging (MRI) and contrast-enhanced ultrasound, respectively, and by immunohistochemistry. Human c-peptide was measured after alloxan-induced diabetes.

Results: T-KO iBeta successfully escaped the NK-mediated response as proved by only a slight reduction of bioluminescence early upon transplantation, comparable to the WT iBeta. Conversely, MHC-I^-/- iBeta were quickly recognized and rejected by NKs. Moreover, T-KO iBeta long-term persisted (240 days for T-KO vs 10±6 days for MHC-I^-/-, p < 0.0001) and properly maintained their endocrine function in the 60% of transplanted mice, as confirmed by the c-peptide levels and the ability to restore normal blood glucose after diabetes induction. Both MRI and histopathological analysis excluded tumour formations, whereas the perfusion analysis showed a robust revascularization of the graft area 8 months after transplantation.

Conclusion: We proposed the genetic manipulation of the NK activating ligands as a novel strategy to make MHC-I-null grafts invisible to human NKs in vivo, offering a new cell therapy approach in T1D treatment.

Supported by: EFSD/JDRF/Lilly European Programme in Type 1 Diabetes Research

JDRF Strategic Research Agreement (SRA) 3-SRA-2022-1254-S-B 2022

Disclosure: G. Siracusano: None.

167

Multi-omic analysis reveals a dual role of PTPN2 in protecting beta cells from inflammatory-induced cell death and promoting stem cell-derived beta like cell maturation

V. Vandenbempt ¹, J. Negueruela¹, F. Ribeiro-Costa¹, M. Nunes¹, S. Singh², L. Martelotto³, E. Gurzov¹;

¹Signal Transduction and Metabolism Laboratory, Université Libre de Bruxelles, Anderlecht, Belgium, ²IRIBHM, Université Libre de Bruxelles, Anderlecht, Belgium, ³Single Cell and Spatial-Omics Laboratory, University of Adelaide, Adelaide, Australia.

Background and aims: Protein tyrosine phosphatases (PTPs) play a key role in diabetes development and β-cell function. PTPN2 is a candidate gene for type 1 diabetes (T1D). We previously found that PTPN2 regulates interferon signaling and ER stress response in pancreatic β-cells in autoimmune diabetes. However, the role of PTPN2 during β-cell development is unclear.

Materials and methods: Single-cell RNA-sequencing (scRNA-Seq) human datasets were integrated for stem cell-differentiated β-like cells, and islets from T1D, T2D, and control organ donors to elucidate the PTP expression in diabetes . PTPN2^+/+ and PTPN2^-/- H1 embryonic stem cells were treated with IFNγ (1000U/mL) for 24h. The samples were divided for ATAC-seq and bulk RNAseq and analyzed by HOMER and DESeq2, respectively. Bulk RNA-seq performed in H1 stem cells was compared with our RNA-seq of EndoC-βH1 cells treated with IFNγ (1000U/mL) using RRHO. Differentiated β-like cells were used for scRNA-seq and implanted under the kidney capsule for pancreatic maturation in vivo. After 16 weeks, fasted human C-peptide levels were measured before streptozotocin (STZ, 150mg/kg) injection.

Results: Meta-analysis of scRNA-Seq in islets from T1D and healthy (males, age 20±12, BMI 23±13, n=4) or obese and T2D (females, age 49±15, BMI 35±6, n=3-4) organ donors showed decreased PTPN2 by long-term T1D, while the expression increased in T2D compared to the respective controls (p<0.001). In addition, PTPN2 expression was decreased during β-like cell differentiation in stages 5 to 7 and mouse transplantation (n=2-5). ATAC-seq and RNA-Seq on H1 stem cells (n=3) showed that PTPN2 knockout affects the inflammatory response upon cytokine treatment by modulating interferon-regulating factors. We compared these results with data obtained from human EndoC-βH1 β-cells and found that PTPN2 plays a similar anti-inflammatory role in stem cells and β-cells. PTPN2^-/- and PTPN2^+/+ H1 stem cells were differentiated into β-like cells and analyzed by scRNA-seq. Late and early β-like cells, polyhormonal cells, endocrine progenitor cells, and pancreatic endoderm cells were found among the clusters. A reduction in late β-like cells was found in PTPN2-knockout cells (67% PTPN2^-/- vs 77% PTPN2^+/+). Early β-like cells showed decreased expression of the pancreatic markers CHGA and INS (p<0.001) and decreased pancreatic β-cell identity in PTPN2^-/- cells (p<0.001). PTPN2-deficient and control islet organoids were transplanted under the kidney capsule of NOD/SCID mice to elucidate the role of PTPN2 during maturation. Fasted human C-peptide levels showed no difference after 16 weeks of maturation. However, when the graft functioned as the sole insulin source post-STZ, the PTPN2-knockout β-like cells did not maintain normoglycemia compared with the controls (n=3-4, p<0.05).

Conclusion: Our study demonstrates that PTPN2 is a master anti-inflammatory protein in cells. PTPN2 depletion decreases differentiation efficiency and hinders maintenance of normoglycemia after transplantation in NOD/SCID mice. Thus, PTPN2 plays a dual role, protecting β-cells from inflammatory-induced cell death and contributing to β-cell maturation after transplantation.

Supported by: F.R.S-FNRS PhD Aspirant scholarship

Disclosure: V. Vandenbempt: None.

168

Monoclonal antibodies against the immune checkpoints B7-H3 and CD155 enhance engraftment of iPSC-derived pancreatic progenitors into mice humanised with allogeneic PBMCs

R. Chimienti ¹, G. Siracusano¹, C. Volpi¹, M. Certo¹, G. Lunetta¹, F. Deambrogio², V. Sordi², L. Piemonti¹;

¹Diabetes Research Institute (DRI), Vita-Salute San Raffaele University, Milan, Italy, ²Diabetes Research Institute (DRI), IRCCS San Raffaele Scientific Institute, Milan, Italy.

Background and aims: Human induced pluripotent stem cells (iPSCs) represent a renewable source of β cells for cell therapy of Type 1 Diabetes (T1D). However, their clinical application is hindered by immune-mediated response. Phase I/II clinical trials proposed stem cell-derived pancreatic progenitors (PPs) for transplantation as they are characterized by null expression of MHC-I. Nevertheless, although the absence of MHC-I makes these cells less immunogenic towards a T cell-related response, it can trigger NK cells via missing-self recognition mechanisms. Our previous investigations highlighted the role of the immune checkpoints B7-H3 and CD155 in recognition of MHC-I-null cells by NK cells. Since down-modulation of MHC-I surface expression occurs, the iPSC-derived PPs may result in higher susceptibility to NK cell recognition and killing. Therefore, we aimed at assessing engraftment and survival of iPSC-derived PPs after transplantation into humanized mice treated with monoclonal antibodies (mAbs) antagonizing the human B7-H3 and/or CD155 ligands.

Materials and methods: We generated luciferase-expressing iPSCs and differentiated them into PDX1⁺/NKX6-1⁺ PPs in vitro. We transplanted the iPSC-derived PPs in hindlimb muscles of hIL-15 NOG mice humanized through intraperitoneal injection (i.p.) of 5·10⁶ donor-derived PBMCs. Engraftment of immune cells was assayed by flow cytometry and survival of graft was evaluated up to 21 days by in vivo imaging system (IVIS). Mice were treated with five injections of 1.25 mg/Kg of α-B7-H3, 2,5 mg/Kg of α-CD155, or both mAbs every 3 days. Untreated mice were used as control. In vivo differentiation was evaluated by immunohistochemistry.

Results: The hIL-15 NOG mouse model ensured persistence of both CD4⁺ and CD8⁺ T lymphocytes, as well as of human CD56^brightCd16^- (precursor) and CD56^dimCD16⁺ (effector) NK cells, over 28 days upon injection of human PBMCs. We confirmed that iPSC-derived PPs are rejected in untreated humanized mice within one week after transplantation, compared to the mAb-treated groups (p < 0.001). Moreover, treatment with α-B7-H3 alone did not confer total protection against immune recognition, as we observed a decrease of bioluminescence signal, and the graft was rejected between 14 and 21 days after transplantation. Conversely, in humanized mice treated with α-CD155 alone or both α-B7-H3 and α-CD155, the transplanted PPs survived up to 21 days (p < 0.001) and properly differentiated in vivo into endocrine islets.

Conclusion: We demonstrated that blocking mAbs against B7-H3 and CD155 axes can be exploited to increase the immune compatibility of iPSC-derived pancreatic derivatives without lasting genetic manipulation, thus increasing chances for their future application in cell therapy of T1D.

Supported by: EFSD/JDRF/Lilly European Programme in Type 1 Diabetes Research

Disclosure: R. Chimienti: None.

OP 29 More on hormones

169

Postprandial secretion of GLP-1 and GIP is reduced in recent-onset type 2 diabetes but increased in steatotic liver disease

M. Huttasch ^1,2, T. Mori^2,3, S. Kahl^1,4, G. Heilmann^1,2, N. Trinks^1,2, M. Schön^1,2, S. Trenkamp^1,2, Y. Kupriyanova^1,2, V. Schrauwen-Hinderling^1,2, N.J. Wewer Albrechtsen⁵, V. Burkart^1,2, R. Wagner^1,4, M. Roden^4,1, for the GDS group;

¹Institute for Clinical Diabetology, German Diabetes Center, Leibniz Institute for Diabetes Research at Heinrich Heine University, Düsseldorf, Germany, ²German Center for Diabetes Research (DZD e.V.), München-Neuherberg, Germany, ³Institute for Biometrics and Epidemiology, German Diabetes Center, Leibniz Institute for Diabetes Research at Heinrich Heine University, Düsseldorf, Germany, ⁴Department of Endocrinology and Diabetology, Medical Faculty and University Hospital Düsseldorf, Heinrich Heine University, Düsseldorf, Germany, ⁵Department of Clinical Biochemistry, Copenhagen University Hospital - Bispebjerg and Frederiksberg Hospital, Bispebjerg, Denmark.

Background and aims: The impaired capacity of the incretin hormones glucagon-like peptide 1 (GLP-1) and glucose-dependent insulinotropic peptide (GIP) to stimulate postprandial insulin secretion is a hallmark of type 2 diabetes (T2D). This phenomenon is commonly attributed to diminished signaling of the incretins at the pancreatic β cell. However, although reduced secretion of these hormones might also contribute to diminished incretin effect in T2D, study results currently remain heterogeneous. Furthermore, the association of incretin secretion with the presence of metabolic dysfunction-associated steatotic liver disease (MASLD) is still unclear. Thus, we compared postprandial GLP-1 and GIP secretion in people with and without recent-onset T2D and MASLD.

Materials and methods: Participants of the German Diabetes Study (GDS) with T2D (n=50; known diabetes duration <1 year) or with normal glucose tolerance (NGT; n=50), matched for age, sex and BMI, underwent mixed meal tolerance (MMTT; 365 kcal) and hyperinsulinemic-euglycemic clamp tests. Postprandial glucose and hormone secretion were quantified by incremental areas under the time curve (iAUC_0-180min) and by hepatic lipid content using ¹H-magnetic resonance spectroscopy. Associations of T2D and MASLD with incretin secretion were assessed by multiple linear regression.

Results: People with T2D and NGT had similar age (52±8 vs. 54±8 years, p=0.115), sex (72% vs. 60% male, p=0.205) and BMI (29.5±5.4 vs. 28.8±4.9 kg/m², p=0.531). T2D featured lower insulin sensitivity (M-value: 7.2±2.7 vs. 9.8±3.1 mg/kg/min, p<0.001), but higher HbA_1c (6.3±0.8 vs. 5.3±0.3 %; 43.9±8.3 vs. 34.2±2.9 mmol/mol), p<0.001) and hepatic lipid content(6.6±6.8 vs. 4.0±5.0 %, p=0.008) as well as - by trend - more MASLD (41% vs. 24%, p=0.056). During MMTT, people with T2D showed greater glucose excursions (iAUC_0-180min: 274±173 vs. 73±58 mmol/L*min, p<0.001), while iAUC_0-180min of GLP-1 (837±539 vs. 1084±561 pmol/L*min) and GIP (8736±4729 vs. 10,729±4098 pmol/L*min) were lower than in NGT (23% and 19% reduction, both p=0.027). The presence of MASLD associated with ~39% and 26% higher release of GLP-1 (β=371, 95%CI [134, 608], p=0.002) and GIP (β=2496, 95%CI [518, 4475], p=0.014), irrespective of age, sex, BMI and T2D diagnosis.

Conclusion: Postprandial secretion of GLP-1 and GIP is reduced in people with recent-onset T2D. Surprisingly, the presence of MASLD is associated with elevated incretin secretion irrespective of the presence of T2D, indicating that liver lipid accumulation might preserve incretin release in recent-onset T2D.

Clinical Trial Registration Number: NCT01055093

Disclosure: M. Huttasch: None.

170

Disparities in GLP-1 and GIP responses to small intestinal glucose infusion between people with well- and poorly-controlled type 2 diabetes

Y. Sun, C. Xie, M. Bound, K.L. Jones, M. Horowitz, R. Young, C.K. Rayner, T. Wu;

The University of Adelaide, Adelaide, Australia.

Background and aims: The incretin hormones, glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP), released by enteral nutrients, modulate energy and glucose homeostasis in both health and type 2 diabetes (T2D). Studies employing either an oral glucose load or a mixed meal have been inconsistent as to how incretin secretion differs in health and T2D, but have been confounded by the wide inter-individual variation in the rate of gastric emptying and (hence) delivery of nutrients to the small intestine. By infusing glucose directly into the duodenum, we previously showed that both GLP-1 and GIP responses to glucose are greater in T2D than in health. We have now compared the GLP-1 and GIP responses to intraduodenal glucose infusion between people with well- (HbA1c < 7.0%), and poorly- (HbA1c > 8.5%), controlled T2D.

Materials and methods: 24 diet-controlled T2D participants, including 12 with good glycaemic control (4M/8F; age 65.6 ± 1.7 years; BMI 27.4 ± 1.2 kg/m²; HbA1c 6.3 ± 0.2%, duration of known diabetes 5.0 ± 0.9 years) and 12 with poor glycaemic control (11M/1F; age 57.7 ± 2.6 years; BMI 31.0 ± 1.6 kg/m²; HbA1c 10.9 ± 0.6%, duration of known diabetes: 5.8 ± 1.5 years), were studied after an overnight fast. Blood glucose was maintained at 5 mmol/L by intravenous infusion of insulin for 90 min (t = -30 to 60 min). Between t = 0-30 min, a 150 mL solution containing 30g glucose and 3g 3-O-methylglucose (3-OMG, a marker of glucose absorption) was infused directly into the duodenum (i.e. 4 kcal/min). Venous blood was sampled every 10 min to measure plasma total GIP and GLP-1 and serum 3-OMG concentrations. Data are means ± SEM.

Results: Fasting plasma total GIP and GLP-1 levels were comparable between the two groups. In response to intraduodenal glucose infusion, the plasma GIP response was greater, and the GLP-1 response less, in participants with poorly-controlled T2D (P = 0.02 and P = 0.007 for time*group interaction, respectively). Serum 3-OMG concentrations did not differ.

Conclusion: Deterioration of glycaemic control in T2D is associated with reduced GLP-1, but augmented GIP responses to a small intestinal glucose load, independent of any change in intestinal glucose absorption.

Disclosure: Y. Sun: None.

171

Human GLP-1R expression in several organs outside the pancreas assessed by a biodistribution study of radiolabeled exendin PET/CT imaging

L. Deden ¹, M. Gotthardt², S. Tokgöz², E. Aarts¹, H. de Boer³, C. Tack⁴, M. Buitinga^2,5, I. Janssen¹, E. Hazebroek¹, B.E. Galan^6,7, M. Boss²;

¹Vitalys, Rijnstate, Arnhem, Netherlands, ²Medical Imaging, Radboud University Medical Center, Nijmegen, Netherlands, ³Internal Medicine, Rijnstate, Arnhem, Netherlands, ⁴Radboud University Medical Center, Nijmegen, Netherlands, ⁵Nutrition and Movement Sciences, Maastricht University, Maastricht, Netherlands, ⁶Internal Medicine, Radboud University Medical Center, Nijmegen, Netherlands, ⁷Internal Medicine, Maastricht University, Maastricht, Netherlands.

Background and aims: Knowledge about the actions of glucagon-like peptide-1 (GLP-1) in humans is increasing, however understanding of GLP-1 receptor (GLP-1R) expression and its relation with GLP-1 signaling remains incomplete. Therefore, in this study GLP-1R expression in human organs was assessed through measurement of accumulation of ⁶⁸Ga-NODAGA-exendin-4 (⁶⁸Ga-exendin) by positron emission tomography/computed tomography (PET/CT).

Materials and methods: Sixty-two participants were included from ongoing studies involving insulinoma detection, type 1 diabetes, and obesity, type 2 diabetes, and post-bariatric surgery outcomes. Radiotracer distribution was assessed visually and quantitatively in abdominal (all participants), chest (N=57), head (N=41) and pelvic (N=8) imaging.

Results: Clear radiotracer uptake was observed in the pancreas and duodenum in all participants. In addition, the majority of participants showed uptake in salivary glands (95%) and the pituitary (78%). Uptake in myocardium was observed in four participants (7%). Of the female participants, half showed uptake in glandular breast tissue (47%) and uterus and ovary (4 out of 8). No accumulation higher than blood pool was observed in liver, adipose tissue and skeletal muscles.

Conclusion: Our data confirm accumulation of ⁶⁸Ga-exendin in organs with well-known GLP-1R expression, such as pancreas and duodenum. In addition, accumulation was observed in organs in which the physiological role of GLP-1R expression remains unclear, such as the pituitary gland, salivary glands, glandular breast tissue and myocardium. These findings suggest that ⁶⁸Ga-exendin PET/CT is a promising tool to increase our knowledge about physiological GLP-1R signaling and to elucidate pathophysiological mechanisms underlying (variable) treatment responses of GLP-1R agonists in humans.

Disclosure: L. Deden: None.

172

Blocking the effects of endogenous GIP increases fasting plasma glucose levels in individuals with obesity

F. Koefoed-Hansen ¹, H. Kizilkaya², M. Helsted³, N. Sørum⁴, F. Dela⁵, M.M. Rosenkilde⁶, J. Holst⁶, B. Hartmann⁶, F.K. Knop⁷, L. Gasbjerg⁸;

¹Center for Clinical Metabolic Research, Gentofte Hospital, Hellerup, Denmark, ²University of Copenhagen, Copenhagen, Denmark, ³Gentofte Hospital, Hellerup, Denmark, ⁴Center for Clinical Metabolic Research, Gentofte Hospital, University of Copenhagen, Hellerup, Denmark, ⁵Medical Physiology, University of Copenhagen, Copenhagen, Denmark, ⁶⁷Clinical Metabolic Physiology, Gentofte Hospital, Hellerup, Denmark, ⁸University of Copenhagen, Copenhagen N, Denmark.

Background and aims: The incretin hormone, glucose-dependent insulinotropic polypeptide (GIP), is secreted from small intestinal enteroendocrine K cells in response to nutrient stimulation and constitutes a primary driver of postprandial potentiation of glucose-induced insulin secretion (the incretin effect). Exogenous GIP has effects on insulin secretion, glucagon secretion, and lipid homeostasis. It has also anabolic actions in bone and increases heart rate. Little is known about the role of circulating GIP between meals. Using the selective GIP receptor antagonist GIP(3-30)NH₂, we investigated the effects of endogenous GIP in obese individuals during fasting.

Materials and methods: In this double-blind, placebo-controlled, crossover study, 12 participants with obesity (7 men, 5 women; mean (±SEM) age 35±4 years; BMI 38 ±1.6 kg/m²; body fat percentage 37 ± 2.2%) underwent two experimental days, preceded by a 10-hour overnight fast, in randomised order and interposed by a minimum of one week,with 220-minute infusion of GIP(3-30)NH₂ (1,000 pmol/kg/min) and placebo (isovolumic saline), respectively. Blood samples were drawn repeatedly and at time 180 min, an ad libitum meal was provided. All data were compared using two-way rmANOVA. Baseline values and baseline-subtracted AUCs (bsAUCs) were compared using Student’s t test.

Results: GIP(3-30)NH₂ reached steady-state after 20 minutes with mean plasma levels of 85 ± 6.0 nmol/l. Baseline plasma glucose were similar (placebo 5.1 ± 0.12 mmol/l vs GIP(3-30)NH₂ 5.0 ± 0.11 mmol/l, p=0.12) and fell during both infusions (p<0.0001). Plasma glucose was higher during GIP(3-30)NH₂ infusion compared to placebo (mean bsAUC_0-180 (±SEM) -53.6 ± 12.2 vs -31.7 ± 6.6 min×mmol/l, p=0.031) (Figure 1). Plasma insulin, C-peptide, glucagon, ketone bodies, free fatty acids, glycerol, systolic blood pressure, diastolic blood pressure, heart rate and ad libitum food intake were unaffected by GIP(3-30)NH₂ compared to placebo (p≥0.05). Additionally, levels of endogenous incretins, GIP (mean_0-180 11 ± 0.4 pmol/l (placebo) vs 11 ± 0.3 pmol/l (GIP(3-30)NH₂), p=0.53) and GLP-1 (12 ± 0.9 pmol/l (placebo) vs 13 ± 2.2 pmol/l (GIP(3-30)NH₂), p=0.09) were stable. The insulin sensitivity indices HOMA-IR and QUICKI were unaffected by the GIP(3-30)NH₂ infusions (p>0.05).

Conclusion: In obese individuals, endogenous GIP could be involved in glucose homeostasis in the fasted state by reducing plasma glucose levels, but the underlying mechanisms remain elusive.

Disclosure: F. Koefoed-Hansen: None.

173

Nutrient sensing mechanisms in human cholecystokinin-secreting cells

M. Santos-Hernandez, N. Guccio, C. Alcaino, C.A. Smith, F. Reimann, F.M. Gribble;

Institute of Metabolic Science, Metabolic Research Laboratories, Cambridge, UK.

Background and aims: Nutrient sensing in the gut is crucial for the regulation of digestion, glucose homeostasis and food intake. Fatty acids are sensed by receptors (GPCRs) on enteroendocrine cells (EECs), triggering the secretion of gut peptide hormones such as the incretins glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP), and cholecystokinin (CCK). CCK plays a fundamental role in nutrient digestion, absorption and post-prandial satiety, and has recently been implicated in pathways underlying the development of preference for sugar and/or fat-containing foods. However, the molecular mechanisms underlying CCK secretion have not been completely elucidated in humans. This study aims to evaluate the signalling pathways (GPCRs) involved in the regulation of CCK secretion in human organoids in response to nutritional stimuli.

Materials and methods: Human duodenal organoids were engineered for live-cell analysis of CCK-producing I-cells. CRISPR-Cas9-mediated homology-donor repair was used to insert the yellow fluorescent protein Venus sequence, or the cAMP FRET-based reporter EpacS-H187 at the 3’ end of the CCK coding sequence. Bulk RNA-sequencing was performed in Venus-positive and negative cells after fluorescence-activated cell sorting (FACS). The CCK-Venus reporter line was used for live single-cell calcium imaging of I-cells, using the Ca²⁺-sensitive fluorescent indicator fura2-AM. The CCK-EpacS-H187 line was used for live single-cell cAMP imaging of I-cells.

Results: Transcriptomic analysis revealed that the fatty acid receptors FFAR1 and FFAR2, the receptors GRP119 and GPBAR, and the amino acid receptors CASR and GPR142 are differentially expressed in I-cells compared with Venus-negative control cells. I-cell intracellular calcium levels were increased upon stimulation by different chain-length fatty acids (acetate, butyrate, decanoate, and laurate (each at 1 mM), eicosapentaenoic acid (EPA, 200 microM) and AM1638 (FFAR1 agonist ,10 microM), whereas the FFAR4 agonist (TUG-891, 1 microM) did not increase intracellular calcium levels. I-cell intracellular cAMP levels were increased after stimulation with agonists of GPBAR1 and GRP119. Additionally, some fatty acids (acetate, butyrate and EPA) also induced an increase in cAMP likely due to the expression of the olfactory receptors OR51E1 and OR51E2, which were identified in Venus positive and negative populations.

Conclusion: In this study, the expression of GPCRs involved in nutrient sensing was identified in human CCK-secreting I-cells, contributing to a better understanding of I-cell physiology. Consistent with the expression data, nutrients and metabolites such as amino acids, bile acids and different length-chain fatty acids were found to directly activate I-cells by elevating intracellular calcium and/or cAMP. Future CCK secretion studies will be performed to validate the molecular mechanisms underlying CCK secretion, potentially identifying new potential therapeutic targets for controlling food intake and energy homeostasis.

Supported by: Wellcome Trust

Disclosure: M. Santos-Hernandez: None.

174

The effect of hepatic fat content on the hepatic response to glucagon during postprandial conditions

H.E. Christie, S. Mohan, A.M. Egan, A. Vella;

Mayo Clinic, Rochester, MN, USA.

Background and aims: Elevated glucagon concentrations that are inappropriate for the prevailing glucose concentrations are the hallmark of certain forms of prediabetes and of type 2 diabetes. Whether this represents innate α-cell dysfunction, a response to selective hepatic resistance to glucagon’s actions, or a combination of both is unknown. We therefore sought to determine the effect of rising glucagon concentrations on hepatic glucose metabolism in people with varying degrees of hepatic steatosis.

Materials and methods: Participants in this study had varying degrees of hepatic fat quantified using the proton density fat fraction (PDFF) measured with an MRI liver elastogram and were studied on one occasion after an overnight fast. Femoral vein (FV), femoral artery (FA), and hepatic vein (HV) catheters were placed under fluoroscopic guidance and Insulin was infused at 0.8 mU/kg/min (0 - 240 min) alongside glucagon (1.5 ng/kg/min 0 - 120 min; 3.0 ng/kg/min 120 - 240 min). A hyperglycemic clamp maintained a peripheral glucose of ~9.0 mmol/L from 0 min to 240min during which a mixture of amino acids (Clinisol (15%, 0.003ml/kg/min; 51% essential AA, 18% branched-chain AA, 9% aromatic AA; Baxter, Healthcare, Deerfield, IL)) was also infused.

Results: Glucose turnover was measured using the tracer dilution technique ([3-³H] glucose). Splanchnic blood flow (SBF) was measured using indocyanine green infusion. Net splanchnic glucose balance (NSGB) was calculated as the difference in glucose concentrations in the FA vs HV divided by splanchnic blood flow. Splanchnic glucose uptake (SGU) was calculated from the splanchnic extraction of glucose multiplied by SBF. Splanchnic glucose production (not shown) was calculated as NSGB - SGU. Data are presented as Mean ± Standard Error of the Mean (SEM). Subjects were classified as having hepatic steatosis if PDFF > 5.0%. Current results are from nine patients (BMI: 29 ± 2 kg/m²)(Table 1).

Conclusion: The presence of hepatic steatosis does not alter hepatic glucose metabolism in response to glucagon and insulin in concentrations intended to mimic postprandial conditions.

Clinical Trial Registration Number: NCT05500586

Supported by: DK116723

Disclosure: H.E. Christie: None.

OP 30 Causes and consequences of gestational diabetes

175

The effect of a reduced energy diet upon maternal weight and pregnancy outcomes in women with gestational diabetes: the DiGest trial

C.L. Meek ¹, D.L. Jones², N. Atta², E.H. Turner², L. Oude Griep³, K. Rennie³, E. De Lucia Rolfe³, S. Sharp³, H. Murphy⁴, R. Taylor⁵, L.C. Kusinski¹;

¹Leicester Diabetes Centre, University of Leicester, Leicester, UK, ²MRC-WT Institute of Metabolic Science, University of Cambridge, Cambridge, UK, ³MRC Epidemiology Unit, University of Cambridge, Cambridge, UK, ⁴University of East Anglia, Norwich, UK, ⁵Magnetic Resonance Center, Newcastle upon Tyne, UK.

Background and aims: Women with gestational diabetes are at increased risk of early onset type 2 diabetes postnatally, mediated in part through maternal BMI and excessive gestational weight gain. Although gestational diabetes is managed with medical nutrition therapy, little evidence exists to guide dietary advice during or after gestational diabetes. Reduced-energy diets improve clinical outcomes in type 2 diabetes but are untested in pregnancy. We hypothesised that a reduced energy diet in pregnant women with gestational diabetes would be associated with improved glycaemia antenatally and reduced risk for type 2 diabetes postnatally.

Materials and methods: In a randomized, controlled, double-blind, whole-diet replacement trial, women with gestational diabetes and BMI >25 kg/m² were assigned (1:1) to receive a reduced-energy diet (1200 kcal/day) (intervention) or notional standard-energy diet (2000 kcal/day; control) from enrolment around 29 weeks until delivery. The diet was provided as weekly dietboxes containing 40% carbohydrate, 35% fat and 25% protein. Diagnosis used the National Institute of Health and Care Excellence (NICE) and interim UK Covid-19 criteria. Primary outcomes were maternal weight change (enrolment to 36 weeks) and offspring standardised birthweight. Secondary outcomes included maternal antenatal and postnatal glycaemia, treatment requirements and neonatal outcomes.

Results: 423 women were randomized at 29wks gestation; (n=211 control; n=212 intervention). Outcome data were available for 386 (91.3%) participants at 36wks and 358 (84.6%) at delivery. Intervention and control groups had similar maternal weight change to 36wks (mean 0.41(SD 4.25) vs 0.50(4.15) kg; baseline-adjusted difference intervention vs control -0.148 (95%CI -0.95 to 0.68); p=0.741). The intervention reduced requirements for short-acting (odds ratio 0.37 (0.14 to 0.97); p=0.044) and long-acting (0.38 (0.19 to 0.75); p=0.006) insulin therapy at 36wks- but did not influence offspring outcomes. Women who lost weight (155/392, 39.5% of cohort; mean 3kg / 3%) had improved antenatal glycaemia (CGM TIR 3.5-6.7 mmol/l at 36 weeks OR 6.8 % increase (2.3 to 11.3); p=0.003), postnatal glycaemia (3-month HbA1c Coeff -3.6 mmol/mol (-5.7 to -1.4); p=0.002) and had reduced rates of large-for-gestational age (OR 0.52 (0.28 to 0.94); p=0.031).

Conclusion: Supply of a reduced energy diet did not guarantee weight loss in women with gestational diabetes. However, women who lost 3kg in late pregnancy had improved antenatal glycaemia and pregnancy outcomes with reduced postnatal HbA1c. Weight loss in women with gestational diabetes offers new opportunities to prevent or delay the development of type 2 diabetes.

Clinical Trial Registration Number: ISRCTN 37866

Supported by: EFSD/Novo Nordisk Foundation Future Leaders Award Programme

Diabetes UK 17/0005712

Disclosure: C.L. Meek: Grants; Dexcom Inc.

176

Reduction of type 2 diabetes can be achieved by lifestyle intervention in young women with previous gestational diabetes

J. Tuomilehto ¹, H. Liu², W. Li², L. Wang², S. Zhang², J. Leng², G. Liu², P. Shao², F. Zhang², H. Tian², Y. Shen³, S. Yang³, E. Gunderson⁴, X. Yang⁵, G. Hu³;

¹Finnish Institute for Health and Welfare, Helsinki, Finland, ²Tianjin Women’s and Children’s Health Center, Tianjin, China, ³Pennington Biomedical Research Center, Baton Rouge, LA, USA, ⁴Kaiser Permanente Northern California, Oakland, CA, USA, ⁵School of Public Health, Tianjin Medical University, Tianjin, China.

Background and aims: Few randomized controlled trials have shown that postpartum lifestyle intervention can prevent or delay the onset of type 2 diabetes in women with prior gestational diabetes. This trial aimed to assess the extent to which type 2 diabetes can be prevented or postponed by lifestyle intervention in women with prior gestational diabetes.

Materials and methods: We randomly assigned 1180 young women aged 20-49 years with prior gestational diabetes at 2.26 years postpartum on average to either the lifestyle intervention or control group. Each participant in the intervention group received individualized counseling aimed at reducing weight among overweight women, increasing physical activity, and keeping appropriate intakes of fat and carbohydrates including dietary fibre. Major elements of the intervention included six face-to-face meetings with the study dietitians in the first year, and two additional sessions in each subsequent year. The mean follow-up was 4.5 years for women who finished the oral glucose tolerance tests (OGTT) to diagnose diabetes, and 7.6 years for those who either took the OGTT or had the physician-diagnosed diabetes. The follow-up lasted up to 11 years. The OGTT was performed annually. The self-reported physician-diagnosed diabetes information was collected from participants missing an OGTT.

Results: Participants in the intervention group achieved significantly more often the predefined lifestyle targets for weight reduction, physical activity, and fibre intake than women in the control group. The overweight women in the intervention group compared with those in the control group lost more weight (2.0 vs. 0.8 kg, respectively), and weight reduction >5% was 30.2% vs. 17.5%, respectively; p= 0.003. During the follow-up the incidence of diabetes based on an OGTT alone was 9.0 and 15.3/1000 person-years and based on either an OGTT or physician diagnosis 9.1 and 14.3/1000 person-years in the lifestyle intervention and control groups, respectively. The lifestyle intervention reduced the incidence by 46% (95% CI 11-67%, p= 0.017) using an OGTT alone to diagnose diabetes, and by 40% (95% CI 11-59%, p= 0.011) using either an OGTT-based or the physician-based diagnosis of diabetes (Figure).

Conclusion: A healthy lifestyle management significantly reduced the incidence of diabetes among women with prior gestational diabetes. To prevent one case of diabetes during a 4.5-year intervention period, 30 women with prior gestational diabetes need to participate in the lifestyle intervention programme.

Clinical Trial Registration Number: NCT01554358

Disclosure: J. Tuomilehto: Grants; European Foundation for the Study of Diabetes.

177

Achievement of glycaemic control after gestational diabetes diagnosis is associated with a risk of obesity in the offspring similar to that in children of individuals without gestational diabetes

A. Ferrara, R.F. Chehab, M. Greenberg, C. Lee, A.L. Ngo, Y. Zhu;

Center for Upstream Prevention of Adiposity and Diabetes Mellitus, Division of Research, Kaiser Permanente Northern California, Oakland, CA, USA.

Background and aims: Gestational diabetes (GDM) increases the risk of childhood obesity in the offspring. The role of glycemic control on the obesity risk in the offspring is unclear and its trajectories have not been studied. We examined the association of glycemic control trajectories with childhood obesity risk in the offspring. We hypothesized that the obesity risk in children of GDM individuals with better glycemic control trajectories will be similar to that in children of individuals without GDM.

Materials and methods: Prospective cohort study of 258,064 mother-child dyads in 2011-2023. Among 17,316 GDM individuals, we identified four glycemic control trajectories from GDM diagnosis to delivery using self-monitoring of blood glucose measurements: stably optimal (39.2%), rapidly improving to optimal (32.3%), slowly improving to near optimal (16.7%), and slowly improving to suboptimal (11.8%). We defined childhood obesity as sex-specific body mass index (BMI)-for-age ≥95^th percentile using growth charts from the U.S. Center for Disease Control and Prevention. We calculated adjusted relative risk (aRR) and 95% CI of childhood obesity by glycemic control trajectories in GDM individuals compared to individuals without GDM using modified Poisson regression models adjusted for sociodemographic factors, parity, pre-pregnancy BMI, smoking and alcohol use.

Results: Obesity prevalence at 2-4 years was 15.1% in children of individuals without GDM, and 15.9%, 18.7%, 20.9% and 24.6% in children of GDM individuals in stably optimal, rapidly improving to optimal, slowly improving to near optimal, and slowly improving to suboptimal glycemic control trajectories, respectively. The childhood obesity risk increased with increasing levels of glycemia (p for trend ≤0.016). At 2-4 years, compared to children of individuals without GDM, children of GDM individuals in stably optimal and rapidly improving to optimal trajectories had a similar risk of obesity (aRR [95% CI]: 1.01 [0.95, 1.07], and 1.04 [0.98, 1.10], respectively), whereas children of GDM individuals in slowly improving to near optimal and slowly improving to suboptimal trajectories had a higher risk of obesity (1.13 [1.04, 1.22] and 1.23 [1.13, 1.34], respectively). At 5-7 years, only children of GDM individuals in stably optimal trajectory had an obesity risk similar to that observed in children of individuals without GDM, whereas children of individuals with less optimal glycemic control had higher risk of obesity (Table).

Conclusion: Early achievement of glycemic control after GDM diagnosis may mitigate the risk of childhood obesity associated with prenatal exposure to GDM to levels similar to those observed in children who were not prenatally exposed to GDM. Studies are needed to understand the barriers to achieving glycemic control early after GDM diagnosis.

Supported by: NIDDK P30-DK092924 RFC; NIHLBI R01HL157666 YZ

Disclosure: A. Ferrara: None.

178

Effectiveness of a family-based health promotion intervention on type 2 diabetes risk markers in women with prior gestational diabetes: 1-year follow-up from the Face-it RCT

K. Kragelund Nielsen ¹, I.K. Dahl-Petersen¹, P. Damm^2,3, D. Jensen^4,5, P. Ovesen^6,7, E. Mathiesen^2,3, U. Kampmann^7,8, C. Vinter^4,9, S. Knorr^6,8, L.L. Andersen⁴, H. Støvring⁸, H.T. Maindal^1,6, Face-it Study Group;

¹Steno Diabetes Center Copenhagen, Herlev, Denmark, ²Rigshospitalet, Copenhagen, Denmark, ³University of Copenhagen, Copenhagen, Denmark, ⁴Odense University Hospital, Odense, Denmark, ⁵Steno Diabetes Center Odense, Odense, Denmark, ⁶Aarhus University, Aarhus, Denmark, ⁷Aarhus University Hospital, Aarhus, Denmark, ⁸Steno Diabetes Center Aarhus, Aarhus, Denmark, ⁹University of Southern Denmark, Odense, Denmark.

Background and aims: Since the relative risk of type 2 diabetes (T2D) among women with prior gestational diabetes mellitus (GDM) is highest within 3-6 years after the GDM affected pregnancy, there is an unmet need for identification of effective prevention means within the first years after delivery. This study aimed to evaluate the effectiveness of a co-produced, family-based health promotion intervention on T2D risk markers in women with recent GDM.

Materials and methods: Multicenter randomized controlled trial across Denmark with 277 women with recent GDM randomly allocated in a 2:1 ratio to intervention group (n=184) or usual care group (n=93) at 10-14 weeks after delivery. The intervention group received (i) 3 home visits with family-based tailored counseling by a trained primary care health visitor, (ii) continuous digital health coaching, (iii) discharge letter from obstetric ward with information on the GDM diagnosis. Evaluation was one year after the delivery with assessment of risk markers for T2D, including OGTT. Primary outcome was body mass index (BMI).

Results: Median BMI at baseline was 27.7 kg/m² in the intervention group vs. 27.9 kg/m² in the usual care group. The intervention group did not significantly improve BMI compared to the usual care group (-0.44 kg/m²; 95%CI -0.98 to 0.11). However, a post hoc analysis demonstrated that the intervention was effective among the subgroup of 158 women with BMI ≥ 25 kg/m² (-0.86 kg/m²; (-1.58 to -0.14)). Analyses of secondary and tertiary outcomes in the whole cohort showed significant improvements in 2-h insulin (-94.25 pmmol/L; 95%CI -167.91 to -20.58) and fasting triglycerides (-0.18 mmol/L; 95%CI -0.30 to -0.05), and reduced odds of impaired fasting glucose (OR 0.33; 95%CI 0.12 to 0.91) in the intervention group at follow-up.

Conclusion: This family-based health promotion intervention with home visits and continuous digital health coaching did not result in significantly reduced BMI in the whole cohort but did result in a small significant reduction in BMI among women with BMI above 25 kg/m². A reduction in selected secondary outcomes, including cardiometabolic risk markers, was observed.

Clinical Trial Registration Number: NCT03997773

Supported by: NNF17OC0027826

Disclosure: K. Kragelund Nielsen: Other; Employed at Steno Diabetes Center Copenhagen, a regional public hospital and research institution, which is partly funded by grant from Novo Nordisk Foundation.

179

Partitioned polygenic scores and extremes of polygenic risk provide mechanistic insights into type 2 diabetes and gestational diabetes in British Pakistani and Bangladeshi populations

S. Hodgson ¹, A. Williamson¹, D. Stow¹, B.M. Jacobs¹, M. Samuel¹, J. Gafton¹, J. Zollner², M. Spreckler¹, C. Langenberg¹, D.A. van Heel¹, R. Mathur¹, M.K. Siddiqui¹, S. Finer¹, Genes & Health Research Team;

¹Queen Mary University of London, London, UK, ²University College London, London, UK.

Background and aims: South Asians experience a higher risk of early-onset Type 2 diabetes (T2D) with normal BMI. However, since genetic research is largely focussed on white Europeans, the reasons for this are poorly understood. We used 12 recently derived multi-ancestry partitioned polygenic risk scores (pPS) to identify the aetiological pathways underlying T2D, gestational diabetes mellitus (GDM), earlier onset, progression to complications and insulin dependence, and treatment response in a south Asian cohort of British Pakistani and Bangladeshi individuals.

Materials and methods: Using electronic health record and genetic data from 51,108 British Pakistani and Bangladeshi individuals with T2D (n = 11,673) and GDM (n = 1,965) in the Genes & Health study, we explored associations between pPS, T2D, GDM, diabetes complications, and treatment response using sex- and ancestry-adjusted multivariable regression and Cox proportional-hazards models. We compared genetic burden of pPS between British South Asians and white Europeans using UKBiobank.

Results: A pPS representing insulin deficiency was most strongly associated with T2D per standard deviation (OR: odds ratio):1.46, 95%CI:1.42-1.50), GDM (OR:1.27, 95%CI: 1.20-1.34) and age at T2D diagnosis (beta = -1.7 years, 95%CI: -1.5 to -1.9), followed by a pPS representing an unfavourable fat distribution (lipodystrophy). Individuals at high genetic risk of both insulin deficiency and lipodystrophy were diagnosed with T2D 8.2 years earlier with BMI 3 kg/m2 lower compared to those at low genetic risk. The insulin deficiency pPS was associated with poorer response to metformin, thiazolidinediones, and SGLT2i (post-treatment HbA1c increased from baseline by 0.51%, 1.83%, and 1.13% respectively), while the liver-lipid pPS was associated with good response to metformin. Higher Insulin deficiency and lipodystrophy pPS were also associated with faster progression to insulin dependence and microvascular and macrovascular complications. Using UK Biobank, we found that south Asians had a greater genetic burden of both these pPS compared to white Europeans.

Conclusion: In British Pakistani and Bangladeshi individuals, genetic predisposition to insulin deficiency, obesity, and lipodistrophy are associated with T2D and GDM, and earlier onset of type 2 diabetes. However only insulin deficiency and lipodystrophy are associated with faster progression to complications and insulin dependence, and only insulin deficiency with worse treatment response. Our findings identify a genetic sub-group of south Asians who do not respond to first-line diabetes treatment and accelerate faster toward insulin dependence and micro and macrovascular complications.

Supported by: Multiple including Wellcome, MRC, HEFC, NIHR, HDRUK. Full funding statement available at genesandhealth.org

Disclosure: S. Hodgson: None.

180

Breast cancer in women with previous gestational diabetes: a nationwide register-based cohort study from Denmark

M.H. Christensen ^1,2, C.A. Vinter^2,3, T.B. Olesen^1,3, M.H. Petersen¹, E.A. Nohr^2,3, K.H. Rubin^4,5, M.S. Andersen^6,3, D.M. Jensen^1,3;

¹Steno Diabetes Center Odense, Odense University Hospital, Odense, Denmark, ²Department of Gynecology and Obstetrics, Odense University Hospital, Odense, Denmark, ³Department of Clinical Research, University of Southern Denmark, Odense, Denmark, ⁴Research unit OPEN, Department of Clinical Research, University of Southern Denmark, Odense, Denmark, ⁵OPEN - Open Patient data Explorative Network, Odense University Hospital, Odense, Denmark, ⁶Department of Endocrinology, Odense University Hospital, Odense, Denmark.

Background and aims: Gestational diabetes mellitus (GDM) is a pregnancy complication characterised by insulin resistance and it affects millions of women worldwide every year. The longterm maternal health has been shown to be compromised after GDM with regard to development of a range of morbidities related to insulin resistance. In addition, a link has been suggested between insulin resistance and breast cancer which is the most common cancer and one of the leading causes of death. As such, women with previous GDM may also be at increased risk of incident breast cancer, however, the existing evidence is conflicting. We therefore aimed to explore this association further.

Materials and methods: The study was a nationwide register-based cohort study using register data (ICD-10 diagnosis codes) on the complete cohort of women giving birth in Denmark from 1997 to 2018. We excluded women with preexistence of diabetes, breast cancer and carcinoma in situ in the breast from the study population. Data were handled as time-to-event data and we performed Cox regression analyses in order to investigate the association between GDM as exposure and incident breast cancer as outcome. We adjusted for demographic and socioeconomic confounders in the analyses and performed a series of sensitivity analyses including adjustment for selected pregnancy complications.

Results: The study population comprised 708,121 women of whom 24,140 (3.4%) were diagnosed with GDM in ≥1 pregnancy. Median age at index pregnancy was 28 years both in women with and without GDM. The women were followed for 11.9 years (median; range 0-21.9 years) with a total risk time of 8,135,323 years. The crude hazard ratio (HR) for breast cancer was 0.99 (95% CI 0.85-1.15) in women with previous GDM compared to women without previous GDM. Adjustment for age, parity, ethnicity, income, occupation, education and preexisting comorbidity did not alter the findings substantially (adjusted HR 0.96 [95% CI 0.83-1.12]), nor did the sensitivity analyses produce results that supported a significant association.

Conclusion: No association between GDM and incident breast cancer was found in this large population-based cohort from Denmark with a follow-up of up to 22 years. Hence, data from this study indicate that the impaired future health profile after GDM does not include a higher breast cancer risk. Examination for and early detection of breast cancer should be prioritised regardless of GDM history.

Disclosure: M.H. Christensen: None.

OP 31 Macrovascular outcomes of diabetes: a holistic view

181

The relationship between serum HDL-cholesterol, cardiovascular disease and mortality in community-based people with type 2 diabetes: the Fremantle Diabetes Study phase II

T.M.E. Davis, W.A. Davis;

Medical School, University of Western Australia, Fremantle, Australia.

Background and aims: Older general population epidemiological studies found a simple inverse association between serum HDL-cholesterol and both cardiovascular disease (CVD) events and mortality, but a range of data over the last decade has suggested a U-shaped relationship. Whether this applies to type 2 diabetes is uncertain. The aim of this study was to assess the prognostic significance of serum HDL-cholesterol concentrations in representative, community-based participants from the Fremantle Diabetes Study Phase II (FDS2).

Materials and methods: The FDS2 is a longitudinal observational study conducted in an urban community of 157,000 people. We followed 1,479 participants with confirmed type 2 diabetes (713 females, mean age 65.6 years; 763 males, mean age 65.9 years) from study entry (2008-2011) to death or end-2021. The major adverse cardiovascular events non-fatal myocardial infarction (MI), non-fatal stroke and cardiovascular death (3-point MACE), and all-cause mortality, were ascertained from detailed prospectively collected data at regular assessments and validated linked administrative databases. Given sex differences in fasting serum HDL-cholesterol concentrations, independent associates of 3-point MACE, excluding participants with prior MI/stroke, were assessed using Cox and competing risk models with sex-specific quintiles of HDL-cholesterol added to the most parsimonious models. Predictors of all-cause mortality were identified using Cox proportional hazards modelling.

Results: In females, with baseline serum HDL-cholesterol quintile 2 (1.04-1.22 mmol/L) as reference, both quintiles 1 (<1.04 mmol/L) and 5 (>1.59 mmol/L) were significant independent predictors of 3-point MACE (P<0.027) and all-cause death (P<0.019) after adjustment for a full range of demographic, clinical and laboratory variables (see Figure). In males, serum HDL-cholesterol quintile did not add to the most parsimonious model for 3-point MACE, but quintile 1 (<0.90 mmol/L) was a significant predictor of death (P=0.026) with quintile 4 (1.15-1.31 mmol/L) as reference after adjustment (see Figure). Competing risk analyses for 3-point MACE showed similar results to the Cox models for both sexes.

Conclusion: There is a significant U-shaped relationship between serum HDL-cholesterol and subsequent 3-point MACE and all-cause death in community-based females with type 2 diabetes after adjustment for confounders. There was no such relationship for 3-point MACE in males but a low HDL-cholesterol was associated with all-cause mortality. These data have sex-specific implications for assessment of serum lipid profiles and their relevance to clinical management in type 2 diabetes.

Supported by: NHMRC (grants 513781, 1042231)

Disclosure: T.M.E. Davis: None.

182

Association between metabolic dysfunction-associated steatotic liver disease and ischemic stroke in patients with type 2 diabetes

M. Jang ¹, G. Kim¹, K.-N. Lee², K. Han³, R. Oh¹, S. Cho¹, J. Kim¹, Y.-B. Lee¹, S.-M. Jin¹, K. Hur¹, J. Kim¹;

¹Department of Medicine, Samsung Medical Center, Seoul, Republic of Korea, ²Department of Biomedicine & Health Sciences, The Catholic University of Korea, Seoul, Republic of Korea, ³Department of Statistics and Actuarial Science, Soongsil University, Seoul, Republic of Korea.

Background and aims: Although both type 2 diabetes mellitus (T2DM) and metabolic dysfunction-associated steatotic liver disease (MASLD) are associated with an increased risk of ischemic stroke, evidence is lacking as to whether the presence of MASLD confers an additional risk of ischemic stroke in patients with T2DM. Our aim is to investigate the association between MASLD and the risk of ischemic stroke in patients with T2DM.

Materials and methods: Among adults aged 20 or older who underwent nationwide health screening between 2009 and 2012, we identified 2,746,079 patients with T2DM. Participants were categorized into five groups: (1) No steatosis; (2) MASLD; (3) MASLD with other combined disease (Concomitant liver disease or Alcohol-related liver disease); (4) MASLD with increased alcohol intake (MetALD); (5) Alcohol-related liver disease (ALD) with metabolic disease. SLD was determined by fatty liver index (FLI) ≥30. We used Cox proportional hazards models and 95% confidence intervals to analyze the association between MASLD categories and the risk of ischemic stroke. We further estimated the risk of ischemic stroke according to degree of hepatic steatosis and alcohol intake.

Results: During a median of 7.1 years of follow-up, a total of 80,141 (3.7%) incident ischemic stroke cases were developed among 2,171,072 participants. Compared with participants without steatosis at the baseline, those with MASLD, MASLD with combined disease, MetALD, and ALD had an increased risk of ischemic stroke with adjusted hazard ratios (aHRs) of 1.15 (95% confidence interval [CI], 1.13-1.16), 1.14 (95% CI, 1.10-1.27), 1.19 (95% CI, 1.15-1.22), and 1.41 (95% CI, 1.34-1.42), respectively, after adjustment for confounding variables. Compared with the group with FLI<30, the group with or FLI 30-60, or ≥60 showed increased aHRs of development of ischemic stroke, and excessive alcohol consumption showed further increased risk of ischemic stroke compared with non-alcoholic group.

Conclusion: All categories of MASLD, MetALD, or ALD were significantly associated with an increased risk of ischemic stroke in patients with T2DM. And there was a gradual association between higher FLI values and greater incidence of ischemic stroke outcomes. Although the risk of stroke according to alcohol intake did not show a positive correlation, but the risk of ischemic stroke was significantly increased in the excessive alcohol drinking group.

Disclosure: M. Jang: None.

183

Risk for cardiovascular disease and all-cause mortality in type 1 compared with type 2 diabetes patients: a nationwide register-based study

V. Patsoukaki ¹, L. Lind¹, R. Kristófi¹, S. Radhi¹, K. Eeg-Olofsson², B. Eliasson², J.W. Eriksson¹;

¹Medical Sciences, Uppsala University, Uppsala, Sweden, ²University of Gothenburg, Gothenburg, Sweden.

Background and aims: Diabetes is a major risk factor for cardiovascular disease (CVD) and premature mortality. Both type 1 and type 2 diabetes (T1D, T2D) carry elevated risks compared to individuals with no diabetes, but there is no research directly comparing these two major diabetes forms. This study aimed to directly compare T1D with T2D regarding risk for cardiovascular disease and mortality.

Materials and methods: Data from the National Diabetes Register of Sweden (NDR) identified diabetes patients at the age between 18 and 84 years at the index date 01 January 2016. Follow-up was performed during five years, until 31 December 2020. Endpoints, including Myocardial Infarction (MI), Heart Failure (HF), Stroke, Cardiovascular (CV) related and All-Cause Mortality, were collected from National Patient Registry and Cause of Death Registry. Cox Proportional Hazards Survival Analysis adjusted for age and sex (and other factors as indicated) were used to assess the Hazard Ratios (HR) and corresponding 95% CI to compare the risk of the studied CVD and mortality outcomes in T1D and T2D patients. Stratification into different age groups was performed. Analysis included the entire population but also patients with or without previous CVD, respectively, in order to elucidate risk for first CVD manifestation as well as recurrence.

Results: A total of 38,351 T1D and 365,675 T2D patients were included. In age- and sex- adjusted Cox regression analysis of the entire population, greater risk for CVD and mortality events was observed in T1D compared to T2D. However, T2D under age 50 displayed higher risk for all studied CVD (HR 1.26; 95% CI 1.13-1.41), higher risk for MI (HR 1.22; 95% CI 1,05-1.43) and higher risk for HF (HR 1.62; 95% CI 1.34-1.95) as compared to T1D. In contrast, at ages above 60, T1D exhibited a significantly higher risk compared to T2D for all the studied outcomes, e.g. in age group 60-70 for all CVD (HR 0.89; 95% CI 0.83-0.95), for MI (HR 0.75; 95% CI 0.69-0.82), for all-cause mortality (HR 0.90; 95% CI 0.82-0.99). In patients with any previous CVD, T1D patients of all ages had a greater risk of a new event of CVD or death (Figure 1). In the whole population, a Cox regression model adjusted for established CV risk factors, revealed a greater risk for all studied outcomes in T2D compared to T1D patients. Regarding the duration of diabetes, for each year passing the risk of events increased by 1-1.5%. Interestingly, when diabetes duration was removed from the Cox model, T1D again displayed a significantly higher risk that T2D for all outcomes except Stroke.

Conclusion: This study shows that T1D and T2D differ in their impact on CVD and mortality, T2d being worse among younger and T1D among older individuals. Diabetes duration per se is a key risk factor of particular importance in T1D. Awareness of various modifiable risk factors is crucial for optimizing prevention of CVD, in particular among older T1D patients.

Disclosure: V. Patsoukaki: None.

184

Prospective associations between markers of low-grade inflammation and major cardiovascular events and mortality in recently diagnosed type 2 diabetes: a Danish nationwide cohort study

S.L. Domazet ^1,2, A. Sharfo^3,2, T.B. Olesen¹, J.V. Stidsen¹, J.S. Nielsen¹, R.W. Thomsen², N. Jessen⁴, P. Vestergaard⁵, M.K. Andersen⁶, T. Hansen⁶, C. Brøns⁷, V.H. Jensen⁷, A.A. Vaag^7,8, M.H. Olsen³, K. Højlund¹;

¹Steno Diabetes Center Odense, Odense, Denmark, ²Aarhus University Hospital, Aarhus, Denmark, ³Steno Diabetes Center Zealand and Department of Internal Medicine, Holbæk Hospital, Holbæk, Denmark, ⁴Steno Diabetes Center Aarhus, Aarhus, Denmark, ⁵Steno Diabetes Center North Denmark, Aalborg, Denmark, ⁶University of Copenhagen, Copenhagen, Denmark, ⁷Steno Diabetes Center Copenhagen, Herlev, Denmark, ⁸Lund University, Malmö, Sweden.

Background and aims: Persons with type 2 diabetes (T2D) experience an increased risk of CVD, even when classical risk factors are optimally managed. This residual CVD risk may be due to systemic low-grade inflammation not targeted by current pharmacological prevention of CVD. The aim of this study was to investigate the prospective associations of circulating IL-6, TNF-alpha, and hsCRP levels with major adverse cardiac event (MACE) and all-cause mortality in persons recently diagnosed with T2D.

Materials and methods: Using Meso Scale Discovery (MSD) assays, we measured plasma levels of the inflammatory biomarkers IL-6, TNF-α and used immunofluorometric assays to measure serum levels of hsCRP in >9000 study participants from the Danish Center for Strategic Research in Type 2 Diabetes (DD2) cohort. We constructed a five-point MACE including non-fatal myocardial infarction, non-fatal stroke, coronary revascularization, heart failure hospitalization, and cardiovascular death, while all-cause mortality was a secondary outcome. We applied Cox proportional hazards model with adjustment for classical CVD risk factors.

Results: Out of 9,660 study participants without previous CVD and at least one measured inflammatory biomarker, 1,006 experienced a first-time MACE and 920 died during a median follow-up time of 7.1 years and 6.9 years, respectively. The median age of study participants was 60.4 years (IQR: 51.7-67.7) and 44% were females. In models adjusted for age, sex, diabetes duration, waist circumference, LDL cholesterol, systolic blood pressure, physical activity, and smoking, a one SD increase in log-transformed levels of IL-6, TNF-alpha, and hsCRP was associated with an increased risk of MACE, with HRs of 1.78 (95% CI: 1.32-2.41), 1.94 (1.06-3.54), and 1.46 (1.19-1.78), respectively. Likewise, a one SD increase in IL-6, TNF-alpha, and hsCRP was associated with an increased risk of all-cause mortality, with HRs of 2.80 (2.00-3.92), 3.73 (1.90-7.31), and 1.45 (1.13-1.87). Associations were similar when further adjusted for pharmacological prevention of CVD (i.e. glucose-lowering, lipid-lowering, antihypertensive, and anticoagulant medication) as well as when we excluded study participants with serum hsCRP≥10 mg/L indicating an acute infection.

Conclusion: In persons with recently diagnosed T2D, markers of systemic low-grade inflammation were persistently associated with both MACE and all-cause mortality, independent of classical CVD risk factors.

Supported by: The Danish Agency for Science (grant no. 09-067009 and 09-075724) and The Novo Nordisk Foundation (grant no. NNF20OC0063292)

Disclosure: S.L. Domazet: None.

185

The interaction between type 2 diabetes and chronic kidney disease on heart failure outcomes: observational study from the Swedish Heart Failure and the Swedish National Diabetes Registries

A. Merolla ^1,2, V. Valente¹, L. Benson¹, U. Dahlström³, L.H. Lund^1,4, S. Gudbjörnsdottir^5,6, F. Cosentino^1,4, G. Savarese^1,4, G. Ferrannini^1,7;

¹Division of Cardiology, Department of Medicine, Karolinska Institutet, Stockholm, Sweden, ²Vita-Salute San Raffaele University, Milan, Italy, ³Department of Cardiology Linköping University Hospital and Department of Health, Medicine and Caring Sciences, Linköping University, Linköping, Sweden, ⁴Heart and Vascular Theme, Karolinska University Hospital, Stockholm, Sweden, ⁵National Diabetes Registry, Centre of Registries, Gothenburg, Sweden, ⁶Department of Molecular and Clinical Medicine, University of Gothenburg, Gothenburg, Sweden, ⁷Internal Medicine Unit, Södertälje hospital, Södertälje, Sweden.

Background and aims: Heart failure (HF), chronic kidney disease (CKD) and type 2 diabetes (T2D) are closely intertwined, contributing to each other’s onset and progression. Real-world evidence on large contemporary cohorts addressing this interplay and related outcomes is lacking.

Materials and methods: Patients with HF enrolled in the Swedish Heart Failure Registry from January 1, 2017, to December 31, 2021, were analyzed. Through the linkage with the National Diabetes Registry, we obtained T2D status and baseline characteristics, then stratified by absence/presence of T2D and eGFR ranges (<30, 30-44, 45-59, ≥60 ml/min/1.73m²). The primary outcome was the composite of time to a first HF hospitalization (HHF) or cardiovascular (CV) death. Secondary outcomes were major adverse CV events (MACE, i.e. CV death, non-fatal myocardial infarction and stroke), CV death and all-cause death. To assess the associations between eGFR strata and outcomes according to T2D, multivariable Cox regression models were performed, adjusting for relevant confounders and assessing T2D interaction with eGFR. A subgroup analysis was conducted by ejection fraction (EF).

Results: Of 36597 patients included, 8053 (22%) had T2D, 53% had HF with reduced EF, 25% HF with mildly reduced EF and 22% HF with preserved EF (HFpEF). Patients with vs without T2D had higher prevalence of CKD across eGFR and EF spectra. They were younger, more often male, had higher CV comorbidities and higher use of cardio-renal drugs as renin-angiotensin-system and sodium-glucose transporter 2 inhibitors. Lower eGFR categories, compared to reference ≥60 ml/min/1.73m², were progressively associated with higher risk of the primary outcome, independently of T2D status (Figure 1). This was consistent across EF, except for HFpEF with eGFR<30 ml/min/1.73m², where the association in T2D was not significant (p-interaction<0.01). Risks of MACE, CV death and all-cause mortality were higher for lower eGFR categories, with higher hazards in non-T2D group (p-interaction<0.01).

Conclusion: In a contemporary HF cohort, decreased kidney function was associated with a progressively higher risk of HHF/CV death and T2D was not a risk modifier. Renal protection should therefore be implemented in HF regardless of T2D.

Supported by: The Horizon Europe programme (project number 101095479 - More-EUROPA) and the Swedish Heart and Lung Foundation (project number 20220680)

Disclosure: A. Merolla: None.

186

Prevalence and prognostic significance of diabetes-related complications in patients with newly diagnosed type 2 diabetes in Sweden

S. Karayiannides, N. Widén, N. Rajamand-Ekberg, S.-B. Catrina;

Academic Specialist Center, Center for Diabetes, Karolinska Institutet (KI), Stockholm, Sweden.

Background and aims: Diabetes mellitus type 2 poses a substantial burden of complications at the time of diagnosis, with varying prevalence across studies. This study aims to present the characteristics and prevalence of comorbidities and diabetes-related complications within a nationwide cohort of newly diagnosed type 2 diabetes individuals in Sweden. Additionally, this study evaluates the prognostic significance of these complications concerning cardiovascular events and all-cause mortality.

Materials and methods: Information was collected on all patients with newly diagnosed diabetes mellitus type 2 that were registered in the Swedish National Diabetes Register (NDR) from 2006 until 2015. Information regarding comorbidities up to inclusion was retrieved from the National Patient Register (NPR), while pharmacotherapy details from 5 months before inclusion until 1-month post-inclusion were obtained from the Prescribed Drug register (PDR). The patients were followed for the incidence of combined event [CE (first of myocardial infarction, ischaemic stroke, or all-cause death) until 31 December 2016 (median follow-up time 4.6 years). Cox regression analysis was used to calculate Hazard Ratios (HR) and their corresponding 95% confidence intervals (CI) after adjustments for age and sex.

Results: The study cohort comprised 121 135 individuals with newly diagnosed type 2 diabetes, including 49 329 females (40.7%). At diagnosis, the median age was 62 years (IQR: 53-70), median HbA1c was 52 mmol/mol (IQR: 46-65), and median BMI was 30.4 (IQR: 27.2-34.3). Notably, comorbidities and diabetes-related complications such as ischaemic heart disease (12.6%), stroke (5.6%), heart failure (5.5%), atrial fibrillation (7.2%), retinopathy (15.0%), albuminuria (18.2%) and neuropathy (12.7%) were already present at the time of diagnosis. Presence of any one of retinopathy/neuropathy or albuminuria at diagnosis was associated with an increased risk of CE (HR 1.25;95% CI 1.11-1.40), while the presence of two out of three (HR 1.81;95% CI 1.53-2.15) or presence of all three microvascular complications (HR 1.97; 95% CI 1.34-2.88) was associated with an even higher increase in risk compared to individuals with no complications. Similarly, the presence of any macrovascular complication (HR 1.57; 95% CI 1.37-1.80) and both any microvascular and macrovascular complications (HR 2.54; 95% CI 2.31-2.79) were associated with increased risk for CE compared to individuals without diabetes-related complications, after adjusting for age and sex (Figure).

Conclusion: These findings underscore the substantial burden of comorbidities and diabetes-related complications present at the time of type 2 diabetes diagnosis. The concurrent presence of complications at diagnosis, significantly heightens the risk of adverse cardiovascular events and all-cause mortality, particularly when both microvascular and macrovascular complications coexist.

Disclosure: S. Karayiannides: None.

OP 32 Fine tuning CGMing

187

How variable is the glycaemic variability? Analysis of CGM data over one year

E. Schupp, R. Stienstra, C.J. Tack, R.I. Meijer;

Radboudumc, Nijmegen, Netherlands.

Background and aims: Glycaemic variability (GV) describes the fluctuations in blood glucose levels and may be relevant for the development of complications. However, GV itself can vary as well. The current guideline is to calculate GV over a period of two weeks, but it remains unclear whether the duration of this period impacts the GV and how much GV varies. The goal of this study was to determine the periodic effects of weekdays and week vs weekend (time of week) and months and seasons (time of year) on GV metrics. We further determined how much GV metrics calculated per two weeks change over a year.

Materials and methods: Glucose measurements from continuous glucose monitors of 262 participants with type 1 diabetes over the complete year 2023 were collected. The coefficient of variation (CV), time in range (TIR), standard deviation (SD) and mean glucose were calculated per day, and grouped per weekday and week vs weekend, per month and per season. One-way ANOVA was used to test whether these subdivisions had an effect on different GV metrics. Besides, GV metrics were also calculated per two weeks (following the current guideline) and the average, SD and range (difference between maximum and minimum value) of these GV metrics were calculated per participant and then taken together.

Results: The different weekdays, weekend vs week, months, and seasons, all significantly affected CV, TIR, SD and mean (p<0.001 for all groups and metrics, except for SD grouped by season p=0.03). CV was the highest on Saturdays (mean 34%, SE 0.093%), in weekends (mean 34%, SE 0.065%), in June (mean 34%, SE 0.12%) and in summer (mean 33%, SE 0.069%). TIR was the lowest on Sundays (mean 55%, SE 0.22%), in weekends (mean 56%, SE 0.16%), in December (mean 55%, SE 0.29%) and in autumn (mean 57%, SE 0.17%). SD was the highest on Sundays (mean 3.2 mmol/l, SE 0.011 mmol/l), in weekends (mean 3.2 mmol/l, SE 0.0081 mmol/l), in December (mean 3.3 mmol/l, SE 0.015 mmol/l) and in winter (mean 3.2 mmol/l, SE 0.0088 mmol/l). Mean glucose was the highest on Sundays (mean 9.8 mmol/l, SE 0.023 mmol/l), in weekends (mean 9.7 mmol/l, SE 0.016 mmol/l), in December (mean 9.9 mmol/l, SE 0.030 mmol/l) and in winter (mean 9.7 mmol/l, SE 0.018 mmol/l). CV for two week results of all participants was on average 38%, the SD 3.2% and the range 12%. Average TIR was 56%, the SD 6.7% and the range was 25%. Average SD was 3.7 mmol/l, the SD was 0.38 mmol/l and the range was 1.4 mmol/l. Average mean was 9.7 mmol/l, the SD 0.72 mmol/l and the range 2.7 mmol/l.

Conclusion: Both time of week (different weekdays, week vs weekend) and time of year (different months, seasons) have a significant effect on the GV metrics. While the differences of the time of week/year are significant, they are also small, and thus the clinical significance should be the topic of further research. Besides, when calculating the GV metrics per two weeks, all of these metrics vary throughout a year. In conclusion, for scientific studies aimed at the importance of GV, it appears important in what exact timeframe this parameter is calculated.

Supported by: Dutch Diabetes Research Foundation

Disclosure: E. Schupp: None.

188

The relationship between time in tight range and the presence of chronic complications in adults with type 1 diabetes: a retrospective cross-sectional real-world study

J. De Meulemeester ¹, S. Charleer¹, M. Visser¹, C. De Block², C. Mathieu¹, P. Gillard¹;

¹University Hospitals Leuven – KU Leuven, Leuven, Belgium, ²University Hospital Antwerp, Edegem, Belgium.

Background and aims: Recent guidelines advocate to report time in tight range (TITR; 70-140 mg/dL) in clinical trials. Focusing on TITR may emerge as a crucial metric in clinical care aimed at reducing the risk of chronic diabetes complications. This study evaluates the association between TITR and the presence of chronic complications in a real-world population of adults with type 1 diabetes (T1D).

Materials and methods: The occurrence of microvascular (diabetic retinopathy [DR], diabetic nephropathy [DN], diabetic peripheral neuropathy [DPN]) and macrovascular complications according to sensor-measured TITR was analyzed cross-sectionally in 808 adults with type 1 diabetes. Binary logistic regression was used to evaluate the association between TITR and the presence of complications without adjustment, with adjustment for HbA1c, and with adjustment for HbA1c and other confounding factors including gender, age, diabetes duration, BMI, blood pressure, lipid profile, smoking, and use of statins and renin-angiotensin-aldosterone system inhibitors.

Results: Of the 808 people included, 51.7% were male, had a mean age of 44.8 ± 15.2 years and T1D duration of 23.1 ± 13.6 years. Mean TITR was 33.9 ± 12.8%. Overall, 46.0% had any microvascular complication (34.5% DR, 23.8% DN, 16% DPN) and 16.3% suffered from any macrovascular complication. The prevalence of any microvascular complication (p for trend <0.001), DR (p for trend <0.001), DN (p for trend =0.009), and a cerebrovascular accident (CVA) (p for trend =0.005) decreased with increasing TITR quartiles (figure 1). For each additional 10% increase in TITR, the odds of having any microvascular complication decreased by 23.8% (OR 0.762, 95% CI [0.679-0.855], p<0.001), DR by 24.3% (OR 0.757, [0.670-0.856], p<0.001), background DR by 24.0% (OR 0.760, [0.655-0.882], p<0.001), severe DR by 14.6% (OR 0.854, [0.731-0.998], p=0.048), DN by 20.1% (OR 0.799, [0.699-0.915], p=0.001), DPN by 16.3% (OR 0.837, [0.717-0.977], p=0.026) and CVA by 34.9% (OR 0.651, CI [0.470-0.902], p=0.010). The independent association of TITR with any microvascular complication (OR 0.867, [0.762-0.988], p=0.032), DR (OR 0.837, [0.731-0.959], p=0.010), background DR (OR 0.831, [0.705-0.979], p=0.027), and CVA (OR 0.619, [0.426-0.899], p=0.012) persisted after adjustment for HbA1c. Similar results were obtained when controlling for HbA1c and other confounding factors as when controlling for HbA1c alone.

Conclusion: TITR is inversely associated with the presence of microvascular complications and CVA in adults with T1D. Although this study was not designed to establish a causal relationship, this analysis adds validity to the use of TITR as measure of glycemic control.

Clinical Trial Registration Number: NCT02601729 and NCT02898714

Disclosure: J. De Meulemeester: None.

189

Inpatient accuracy of continuous glucose monitors in people with type 1 diabetes: a multi-centre retrospective study

R. Wang ^1,2, A. Connell^2,3, C. Chiang^1,4, J. Tjahyadi², M. Kyi^1,5, S. Fourlanos^1,5;

¹Dept Diabetes & Endocrinology, Royal Melbourne Hospital, Parkville, Australia, ²Dept Endocrinology & Diabetes, Eastern Health, Box Hill, Australia, ³Dept Chemical Pathology, Eastern Health, Box Hill, Australia, ⁴Dept Chemical Pathology, Royal Melbourne Hospital, Parkville, Australia, ⁵Dept Medicine, Royal Melbourne Hospital, Parkville, Australia.

Background and aims: Continuous glucose monitor (CGM) use is widespread, however approval for inpatient CGM use remains limited given accuracy concerns in acutely unwell hospitalised patients. Following implementation of a national CGM subsidy for people with type 1 diabetes (T1D) in Australia in 2017, increasing numbers of people with T1D use CGM during hospitalisation. We aimed to determine CGM accuracy metrics amongst hospitalised individuals with T1D in a real-world setting.

Materials and methods: In this multi-centre retrospective observational study, we compared interstitial fluid glucose via CGM with reference blood glucose (capillary, blood gas, plasma) in patients with T1D requiring multi-day admissions during 2020-2023 across four hospitals in Melbourne, Australia. Capillary blood glucose was measured with point-of-care (POC) devices, Nova StatStrip or Accu-Chek Guide; blood gas (GAS) was measured with ABL90/ABL800 FLEX or GEM5000; and plasma glucose was measured with Architect c16000 or Cobas c502/Cobas c702. FreeStyle Libre 2, Guardian 3 and Dexcom G6 CGM were used. CGM and reference glucose results were time-matched (within 7 minutes).

Results: 5,371 paired CGM and pooled reference glucose pairs over 196 admissions were available for analysis. A majority of patients were men (56%), median age 48 years (IQR 31 - 62), median Charlson comorbidity index 2 (IQR 1-5), mean HbA1c 73 mmol/mol (8.8%), and median length of stay was 3.3 days (IQR 1.9 - 6.2). Aggregate population mean absolute relative difference (MARD) was 11.8%. CGM accuracy data compared with POC and GAS glucose are presented in Table 1. Comparing CGM measures with pooled reference glucose, a high proportion of glucose pairs fell within Zones A and B of the Clarke Error Grid (Libre 2 98.5%; Guardian 3 98.7%; Dexcom G6 97%), but was just below the ISO 15197:2013 standard (≥99%). Given specific accuracy concerns in haemodialysis (HD), we additionally analysed CGM data for eleven inpatients receiving HD. Analysis using pooled reference glucose demonstrated no significant MARD difference between HD and non-HD patients by Wilcoxon rank sum test (p=0.13). MARD with HD was 8.35% (285 pairs) and 10.45% (310 pairs) for Libre 2-POC and Dexcom G6-POC glucose pairs, respectively.

Conclusion: In a real-world cohort of inpatients with T1DM a high proportion of CGM measures accord with blood glucose (especially when POC blood glucose >3.9 mmol/L) including the scenario of HD. However, CGM falls short of ISO 15197:2013 accuracy standards. In most instances treatment decisions can likely be based on CGM measures, however more reporting of real-world hospital CGM data is required to validate accuracy and safety for inpatient CGM use.

Supported by: Australian Government RTP scholarship, Fred Knight scholarship, Gordon P Castles scholarship, Rowden White scholarship

Disclosure: R. Wang: Grants; Australian Government RTP scholarship, Fred Knight scholarship, Gordon P Castles scholarship, Rowden White scholarship.

190

Impact of continuous glucose monitoring on hospitalisations in people with type 2 diabetes: real-world analysis

S. Garg ¹, I.B. Hirsch², E. Repetto³, J. Snell-Bergeon⁴, B. Ulmer³, C. Perkins³, R. Bergenstal⁵;

¹University of Colorado School of Medicine and Barbara Davis Center for Diabetes, Aurora, CO, USA, ²School of Medicine, University of Washington, Seattle, WA, USA, ³Roche Diagnostics, Indianapolis, IN, USA, ⁴Barbara Davis Center, Department of Medicine (J.S.-B.), University of Colorado, School of Medicine, Aurora, CO, USA, ⁵International Diabetes Center, HealthPartners, Minneapolis, MN, USA.

Background and aims: Recent data shows that the increasing prevalence of diabetes in the US continues to drive a steady rise in healthcare resource utilization. Studies have also shown that acquisition of continuous glucose monitoring (CGM) lowers glycated hemoglobin (HbA1c) and hypoglycemia and reduces healthcare resource utilization in individuals with insulin-treated diabetes. However, the real-world impact of CGM on hospitalizations within broader type 2 diabetes population is not completely understood.

Materials and methods: In this retrospective analysis, we used Optum’s de-identified Market Clarity database of >79 million people to evaluate CGM use in 74,264 people with type 2 diabetes, who were treated with non-insulin medications (NIT; n=25,788), basal insulin (BIT; n=25,292), and prandial insulin therapy (PIT; n=23,184). The primary outcomes were changes in all-cause hospitalizations (ACH), acute diabetes-related hospitalizations (ADH), and acute diabetes-related emergency room visits (ADER) during the 6- and 12-monts post-index periods. The index period was defined as the date of the first CGM acquisition between June 30, 2019 to January 5, 2022.

Results: Reductions were observed in ACH in the 6 and 12 months (post-index period) in all three treatment groups: NIT -14% and -10%, BIT -25% and -23%, PIT -25% and -19%, respectively (all p<0.0001). ADH, and ADER were also significantly reduced in all three groups during post-index period (all p<0.0001). (Figure)

Conclusion: The current study demonstrated that use of CGM in real-world settings in people with type 2 diabetes treated with noninsulin and insulin regimens was associated with significant reductions in all-cause hospitalizations, acute diabetes-related hospitalizations and emergency room visits. Our findings suggest that expanding access to CGM for people with type 2 diabetes who are generally not considered to be eligible may help reduce hospitalizations and overall healthcare costs.

Supported by: Roche Diagnostics

Disclosure: S. Garg: Employment/Consultancy; Medtronic, Roche Diabetes Care, Merck, Lexicon, Novo Nordisk, Sanofi, MannKind, Senseonics, Zealand, Eli Lilly and Company. Grants; Eli Lilly and Company, Novo Nordisk, Merck, Lexicon, Medtronic, Dario, National Cancer Institute, T1D Exchange, National Institute of Diabetes and Digestive and Kidney Diseases, JDRF, Animas, Dexcom, Sanofi.

191

Impact of continuous glucose monitoring use on glycated haemoglobin (HbA _1c ) in people with type 2 diabetes: real-world analysis

R.M. Bergenstal ¹, I.B. Hirsch², E. Repetto³, J. Snell-Bergeon⁴, B. Ulmer³, C. Perkins³, S.K. Garg⁵;

¹International Diabetes Center, HealthPartners Institute, Minneapolis, MN, USA, ²School of Medicine, University of Washington, Seattle, WA, USA, ³Roche Diagnostics, Indianapolis, IN, USA, ⁴Barbara Davis Center, Department of Medicine (J.S.-B.), University of Colorado, School of Medicine, Aurora, CO, USA, ⁵Barbara Davis Center, University of Colorado Denver, Aurora, CO, USA.

Background and aims: Long-term suboptimal glycemic control leads to microvascular and macrovascular complications. Studies have shown that the use of continuous glucose monitoring (CGM) lowers glycated hemoglobin (HbA1c) and hypoglycemic episodes, particularly in individuals using insulin therapy. However, the real-world impact of CGM on HbA1c within the broader type 2 diabetes population, with and without the use of insulin therapy, has not been well studied.

Materials and methods: In this retrospective analysis, we used Optum’s de-identified Market Clarity database of >79 million people. Changes in HbA1c medications were assessed in a subgroup of 6,004 patients who had continuous enrollment in their health plan and HbA1c values during the 6-month pre-index and 6- and 12-month post-index periods and were treated with non-insulin medications (NIT; n=1,584), basal insulin (BIT; n=2,535), and prandial insulin therapy (PIT; n=1,885). The index period was defined as the date of the first CGM acquisition between June 30, 2019 to January 5, 2022

Results: The HbA1c values decreased by 0.9% at ~3 months in NIT, BIT and PIT groups (all p<0.0001). (Figure) The HbA1c values remained lower throughout the 12-months (-1.1%, -1.0%, and -1.0%, respectively (all p<0.0001]. These HbA1c improvements, were noted for each treatment group, in those individual who made no medication changes and in those individuals who added or subtracted diabetes medications over the 12 months after CGM acquisition.

Conclusion: The current study demonstrated that the use of CGM in real-world settings in people with type 2 diabetes treated with noninsulin and insulin regimens was associated with significant and sustained reductions in HbA1c values. Our findings suggest that CGM use improves glucose control for all medically-treated people with type 2 diabetes regardless of their therapy.

Supported by: Roche Diagnostics

Disclosure: R.M. Bergenstal: Employment/Consultancy; Abbott Diabetes Care, Ascensia, CeQur Corporation, DexCom, Hygieia, Insulet, Johnson & Johnson, Lilly, Medtronic, Novo Nordisk, Onduo, Roche, Sanofi, United Healthcare, Zealand.

192

Time above range and not time below range is associated with mortality in older patients with type 2 diabetes: results from the HYPOAGE study

S. Smati ¹, A.-S. Boureau², B. Guyomarch³, I. Allix⁴, C. Annweiler⁵, N. Cervantes⁶, I. Delabrière⁷, P. Gourdy⁸, S. Guyonnet⁹, R. Litke⁷, M. Paccalin¹⁰, A. Penfornis¹¹, P.-J. Saulnier¹², S. Hadjadj¹, B. Cariou¹;

¹Dept. of Endocrinology, L’institut du Thorax - Nantes Université, CHU Nantes, CNRS, INSERM, Nantes, France, ²Dept. of Gerontology, L’institut du Thorax - Nantes Université, CHU Nantes, CNRS, INSERM, Nantes, France, ³L’institut du Thorax - Nantes Université, CHU Nantes, CNRS, INSERM, Nantes, France, ⁴Dept. of Endocrinology, CHU Angers, Angers, France, ⁵Dept. of Gerontology, CHU Angers, Angers, France, ⁶Dept. of Gerontology, CHSF, Corbeil-Essonnes, France, ⁷Dept. of Gerontology, CHRU Lille, Lille, France, ⁸Dept. of Diabetology, CHU Toulouse, Toulouse, France, ⁹Dept. of Gerontology, CHU Toulouse, Toulouse, France, ¹⁰Dept. of Gerontology, CHU Poitiers, Lille, France, ¹¹Dept. of Diabetology, CHSF, Lille, France, ¹²Dept. of Diabetology, CHU Poitiers, Lille, France.

Background and aims: There is a lack of real-life data on blood glucose control in older patients with type 2 diabetes (T2D). We aimed to determine the frequency and predictors of hyperglycemia in older patients with T2D on insulin therapy by using continuous glucose monitoring (CGM). We also sought to compare the impact of time spent in hyperglycaemia or hypoglycaemia on 1-year mortality.

Materials and methods: HYPOAGE is prospective multicenter study including 146 insulin-treated older adults aged 75 and older, with T2D for at least 1 year and treated with insulin for at least 6 months. All patients benefited from ambulatory blinded CGM for 28 consecutive days with FreeStyle Libre Pro® sensor. All participants underwent a geriatric and diabetic assessment. Multivariable logistic regressions were used to identify factors associated with time above range (TAR2) [glucose concentration >250 mg/dL] > 10%. The cumulative incidence of 1-year mortality was estimated using the Kaplan-Meier method.

Results: The mean age of the 141 analyzed patients was 81.5±5.3 years and 56.7% were male, 72.3% were considered as complex and 27.7% as healthy after geriatric assessment. The mean baseline HbA1c was 7.9±1.0% and 43 patients had TAR2 > 10%. Compared to patients with TAR2 ≤ 10%, they were older (83.5[5.7] vs 80.6[4.9] years), with lower BMI (28.0[5.5] vs 31.3[6.1] kg/m2), higher HbA1c (8.6[0.9] vs 7.6[0.9] %), and a less frequent healthy geriatric status (14.0 vs 33.7%). In multivariate analyses, TAR2 > 10% remains only significantly associated with HbA1c (OR: 3.08 for 1% increase [1.72-5.50], p=0.0002) and BMI (OR: 0.92 [0.84-1.00], p=0.0496). TAR > 10% was inversely associated with 1-year survival (HR: 0.30 [0.09-0.94], p=0.03), but not time below range (TBR) [glucose concentration < 70 mg/dL] > 1% (HR: 0.85 [0.23-3.15], p=0.81).

Conclusion: In this real-life study, 30% of older patients are above the TAR target recommended by the ATTD 2019 consensus (< 10% for level 2). Unlike TBR, increased TAR2 is significantly associated with mortality. Reducing the time spent in hyperglycaemia should therefore become a priority goal for T2D management in older patients.

Clinical Trial Registration Number: ClinicalTrials.gov ID NCT03020264

Supported by: Investigator Inititative Study sponsored from SANOFI

Disclosure: S. Smati: None.

OP 33 Emerging science in diabetic kidney disease from fasting to autophagy

193

Study of the A _2b adenosine receptor in profibrotic activation of glomerular parietal epitelial cells in experimental diabetes

C. Jara ¹, D. Moscol¹, P.D. Sastre¹, A. Torres-Arevalo², I. Arias¹, C. Cappelli¹, R. San Martin¹;

¹Instituto de Bioquímica y Microbiología, Universidad Austral de Chile, Valdivia, Chile, ²Facultad de Medicina Veterinaria y Recursos Naturales, Universidad Santo Tomás, Talca, Chile.

Background and aims: Diabetic nephropathy (DN) is a complication of diabetes that remains incurable characterized by proteinuria and progressive loss of kidney function. Podocytes are terminally differentiated non dividing cells that form part of the glomerular filtration barrier (GFB). At the glomerular level, loss of podocytes disrupts the GFB leading to proteinuria and glomerulosclerosis. Recently, the parietal epithelial cells (PEC) of Bowman’s capsule have been identified as podocyte progenitor cells, however, some studies have shown that in DN, PEC undergo a profibrotic transition activating them and limiting their differentiation into the podocyte lineage. Notably, the molecular mechanisms associated with this process are still unclear. In our laboratory, the progression of DN has been correlated with high levels of adenosine and antagonism of adenosine receptor A_2B (A_2BAR) attenuates fibrosis and proteinuria in experimental DN rats. Thus, this work addresses the effect of A_2BAR antagonism on the profibrotic activation of PECs, to understand how podocytopenia is perpetuated in DN and identify possible targets for intervention.

Materials and methods: In vitro primary culture of PECs from control Sprague Dawley rats were treated with TGFβ 10 ng/ml and high D-glucose 25 μM for 48h. After treatment total RNA was extracted and the expression of PEC markers such as CD24 and CD56 and activated PEC markers such as CD9, CD44 and αSMA were evaluated using specific primers through RT-qPCR. For in vivo studies, experimental diabetes was induced in Sprague Dawley strain rats using a single i.v. dosis of streptozotocin (STZ) 55 mg/Kg. 4 weeks after diabetes induction, the A_2BAR antagonist MRS1754 (1mg/kg/48h i.p.) was administered to diabetic rats for another 4 weeks. All animal procedures were approved by the Ethic Committee resolution n° 465/2022. The PEC CD24⁺CD56⁺ subpopulation and PEC activated CD9⁺CD44⁺αSMA⁺ cells undergoing profibrotic activation were analyzed through immunohistochemistry and western blot in samples obtained from kidneys of control, diabetic and treated rat groups. The statistical analysis utilized was Student t test for n=3.

Results: In vitro studies in PECs evidenced a decrease in CD24 and CD56 transcript levels in cells exposed to TGF-β and high D-glucose which was blocked by exposition to MRS1754. On the other hand, profibrotic markers such CD9, and αSMA were up regulated in PECs exposed to TGF-β and high D-glucose which was also blocked by the MRS1754 exposition. These results correlate with our results in vivo where we determined that CD24 and CD56 levels were decreased while an increase of the profibrotic markers CD9, CD44 and αSMA occurred in the bowman´s capsule of diabetic rats. The use of A_2BAR receptor antagonist MRS1754 in diabetic rats also attenuated the increase of CD9, CD44 and αSMA and increased the levels of CD24 and CD56.

Conclusion: The antagonism of A_2BAR blocks the profibrotic activation of PEC, which may preserve their potential to differentiate toward podocyte.

Supported by: Fondecyt Postdoctorado N°3220711 and Fondecyt Regular N°1211613.

Disclosure: C. Jara: None.

194

WWP2-induced pericytes to myofibroblast transition contributes to the progression of fibrosis in CKD in diabetes

F. Conserva ¹, A. Cicirelli¹, G. Sclavo¹, M. Venneri¹, R. Franzin¹, A. Stasi¹, E. Squiccimarro¹, A. Gallone¹, M. Fiorentino¹, F. Pesce², L. Gesualdo¹, P. Pontrelli¹;

¹University of Bari Aldo Moro, Bari, Italy, ²Fatebenefratelli Hospital Isola Tiberina-Gemelli Isola, Rome, Italy.

Background and aims: CKD in diabetes is characterized by different clinical and histological phenotypes driven by diverse molecular pathways. CKD in diabetes includes different phenotypes, such as diabetic Nephropathy (DN), non-diabetic renal disease (NDRD) and mixed forms (DN+NDRD). DN represents the primary cause of end stage renal disease (ESRD) and our group demonstrated that renal damage in DN is characterized by the specific activation of lysine63 (K63)-ubiquitination (Ub) pathway, promoting the progression of renal fibrosis. WWP2 is an E3 ubiquitin-protein ligase that modulates pro-fibrotic myofibroblast activation in cardiac fibrosis and the metabolic reprograming of myofibroblasts in CKD. The aim of this study was to evaluate the involvement of WWP2 in the progression of kidney fibrosis in diabetes and its involvement in K63-Ub pathway

Materials and methods: WWP2, K-63Ub, α-sma and PDGFβrec expression were evaluated in: i) 22 kidney biopsies from patients with a biopsy proven diagnosis of DN (n=11), NDRD (n=6) and CKD without diabetes (n=5), by immunohistochemistry and immunofluorescence; ii) streptozotocin (STZ)-treated DBA/2J mice, a model of human DN, in the presence or absence of a specific inhibitor of K63Ub (NSC697923) by immunohistochemistry (n=3 for each group); iii) in HK2 tubular cells, EAHY926 endothelial cells and in human-derived pericytes under hyperglycemic conditions by western blotting and qPCR.

Results: Immunohistochemistry showed that WWP2 was expressed at higher levels in DN patients’ biopsies when compared to patients with NDRD (p<0.05 and p<0.01 respectively), both at tubular, glomerular and in the vascular compartment. Immunofluorescence confirmed that WWP2 increased expression in DN patients, was associated to accumulation of K63Ub proteins in the tubular compartment as well as in the vascular compartment in particular in pericytes, cells that envelop the surface of vessels and have been described as the major source of scar-forming myofibroblasts in CKD. The involvement of WWP2 in the K63-Ub pathway was confirmed in vivo in DBA2J diabetic mice in which the specific inhibition by NSC697923 of the K63Ub pathway, significantly reduced WWP2 expression in kidney tissues from diabetic mice (p<0.05). These results were confirmed in vitro where specific inhibition of K63Ub pathway by NSC697923 significantly reduced hyperglycaemia-induced WWP2 expression in HK2 cells by western blotting and qPCR (p<0.01). Hyperglycaemic conditions did not influence WWP2 expression in EAHY926 endothelial cells (p=n.s.). Conversely, using qPCR, in pericytes we observed an hyperglycemia-induced expression of both WWP2 levels (RR>2) and αsma levels (RR>3) at different time points. Moreover, flow cytometry showed that hyperglycemia also induced a decrease of NG2 marker of perycites, and an increase in vimentin thus suggesting their transition to myofibroblasts.

Conclusion: These findings demonstrate the influence of WWP2 on the K63-Ub pathway thus driving fibrogenesis in DN patients, in particular through pericyte to myofibroblast transition. WWP2 and K63-Ub pathway could represent novel potential targets for therapeutic intervention in the treatment of chronic kidney disease in diabetes.

Disclosure: F. Conserva: None.

195

Mechanisms of PERK/ATF4 axis-targeted ATG4B autophagy signalling pathway in promoting diabetic kidney disease

Q. Huang, X. Fei, J. Gong, Y. Chen, X. Wu;

Zhejiang Provincial People’s Hospital, Hangzhou, China.

Background and aims: The PERK/ATF4 axis plays an important role in autophagic regulation, but the regulatory mechanisms of the ATG4B-related pathway in DKD is unclear. This study aims to perform proteomic analysis to observe the relationship between autophagy proteins and DKD, and to establish a model of high glucose-induced HK2 cell to explore and clarify the regulatory mechanisms of PERK/ATF4 axis on ATG4B autophagic pathway.

Materials and methods: A high glucose-injured HK2 cell model was constructed, the ROS activation or inhibition experiments were performed, and CCK-8, DCFH-DA fluorescent probe, transmission electron microscope, flow cytometry , RT-PCR,Western-Blot , and IF were used to analyze the HK-2 cell proliferation viability, ROS production, autophagosomes formation, apoptosis levels, mRNA and protein expressions of the PERK/ATF4 axis and ATG4B pathway, and the cellular localization and expression of ATG4B, LC3 and p62. Then, HK2 cells model of ATF4 knockdown and ATG4B overexpression were constructed, and PERK inhibitors was administrated. The cell proliferation viability, ROS production, cell apoptosis, the expression levels of proteins, and the intracellular localization were detected by abovementioned methods. Lastly, constructing DM and DKD mouse models, infecting mice with AAV-ATG4B and AAV shATF4 lentivirus, collecting mouse kidney tissue samples, preparing FFPE sections,observed the pathological changes of renal tubule.

Results: After high glucose treatment, the proliferation viability of HK2 cells decreased, the apoptosis rate and ROS production increased (P<0.05 or 0.01),the intracellular autophagosome formation was reduced. An increaseing of PERK phosphorylation, up-regulation of the ATF4, CHOP and p62 expressions and gene transcription with the down regulation of ATG4B and LC3 protein were observed (P<0.05 or 0.01). ROS activators have the similar effects as the high-glucose cultures, while ROS inhibitors have a reversible effect. PERK inhibitors increased proliferation viability, and decreased ROS production and apoptosis treated with high glucose level (P <0.05 or 0.01), and they promoted intracellular localized expressions of ATG4B and LC3, and inhibited p62 expression. PERK inhibitors could down regulate ATF4 and up-regulate the ATG4B protein expressions, and also decrease the PERK and Eif2α phosphorylation, CHOP expression (P <0.05 or 0.01). Furthermore,ATF4 knockdown increased the proliferation viability of HK2 cells, decreased ROS production and apoptosis level (P <0.05 or 0.01). SiATF4 also promoted located expressions of cellular ATG4B and LC3 and suppressed the p62 expression. As expected, ATG4B overexpression had no effects on the protein expressions of PERK/ATF4 axis. Microscopy showed that compared with normal control group, mesangial cell proliferation, renal tubule degeneration, protein exudation and interstitial lymphocyte infiltration increased in DKD model group. the proliferation of glomerular cells in shATF4 and ATG4B groups was weakened, the protein degeneration of renal tubules and the number of interstitial lymphocytes was reduced.

Conclusion: The ATG4B autophagy pathway in renal tubular cells is involved in the pathogenic mechanisms of DKD. High glucose culture triggers the increasing of oxidative stress and apoptosis of HK2 cells, reduces cellular autophagy, and this is mainly inhibited by PERK/ATF4 axis-targeted ATG4B autophagy signaling, which may play an important role in the pathogenesis of DKD.

Disclosure: Q. Huang: None.

196

Renal gluconeogenesis maintains body fluids during fasting

K. Kaneko, M. Yamato, H. Katagiri;

Department of Diabetes, Metabolism and Endocrinology, Tohoku University Graduate School of Medicine, Sendai, Japan.

Background and aims: During fasting, endogenous glucose production (EGP) is essential for distributing glucose throughout the body as a fuel. Cytosolic phosphoenolpyruvate carboxykinase (PCK1) is a major rate-limiting enzyme in the gluconeogenic pathway. The liver is regarded as playing a central role in EGP, but during prolonged fasting, renal gluconeogenesis, taking place exclusively in proximal tubule (PT) cells, is reportedly activated to the extent that its contribution to EGP can even surpass that of the liver. Considering that the liver is anatomically much larger and presumably has far more glucose-producing capacity, we hypothesized that renal gluconeogenesis has intrinsic and unique purposes. We thus aim to uncover the physiological significance of renal gluconeogenesis during fasting.

Materials and methods: We produced mice with inducible PT-specific deficiency of PCK1 (K-Pck1KO mice) and analyzed their phenotypes in comparison with control mice during 24-hour (24hr) fasting.

Results: PT PCK1 deletion decreased renal glucose contents during fasting (n=6-8, P<0.05), but circulating glucose levels were not affected. Lactate administration enhanced glucose elevations in K-Pck1KO as compared with control mice (n=10, P<0.05), indicating that compensatory responses of enhanced gluconeogenesis in other organs, such as the liver, maintain systemic glucose levels. During ad libitum feeding, the kidney of K-Pck1KO mice showed no apparent histological abnormalities. In contrast, 24hr fasting induced marked dilatation of not only PTs but also distal nephron tubules in K-Pck1KO mice. Urinary volume and sodium excretion, both of which were physiologically decreased in control mice (n=6, P<0.05), conversely increased in K-Pck1KO mice during a 24hr fasting (n=6, P<0.05), leading to hypovolemic states. In addition to reduced expression of renal AQP1 (mainly functioning in PTs) (n=5-7, P<0.05), urinary responses to the administration of pharmacological agents targeting distal nephrons, such as thiazide diuretic and desmopressin, were absent in K-Pck1KO mice. Moreover, urinary excretions of albumin and glucose were increased in K-Pck1KO mice (n=6, P<0.05). Thus, the absence of PCK1-mediated renal gluconeogenesis injures not only PTs but also distal nephron segments, as fasting is prolonged, resulting in the inability to reabsorb glomerular filtrates: water, minerals and nutrients. These findings suggest that renal gluconeogenesis while fasting is a prerequisite for maintaining renal multi-tubular functions.

Conclusion: Considering that it would not be unusual for wild animals to experience food deprivation for periods exceeding 24 hours, this study has unveiled intrinsic roles of renal gluconeogenesis which is physiologically essential for survival: maintaining body water and solutes. In addition, given that renal gluconeogenesis is altered by diabetes and renal diseases, we anticipate that the functions elucidated herein will enhance understanding of the pathophysiology of various renal disorders.

Supported by: Grants-in-Aid for Scientific Research from JSPS

Disclosure: K. Kaneko: None.

197

Is NOX5 potentially a better therapeutic target than NOX4 in renal complications of diabetes?

J. Jha ¹, S. L-Trevino¹, A. Dai¹, J. Vincent², J. Meister³, M.E. Cooper¹, K. Jandeleit-Dahm¹;

¹Department of Diabetes, School of Translational Medicine, Monash University, Melbourne, Australia, ²Universite de Geneve, Geneva, Switzerland, ³Institute for Clinical Diabetology, German Diabetes Center, Düsseldorf, Germany.

Background and aims: Diabetes induced excessive production of reactive oxygen species (ROS) in the kidney contribute to development and progression of renal complications of diabetes. Chronic hyperglycaemia disrupts the renal ecosystem, leading to enhanced inflammation, progressive albuminuria and fibrosis in diabetic kidney disease (DKD). NADPH oxidase-NOX5, an isoform present in humans but not in rodents, has been identified as a significant contributor to renal ROS in humans. This study aimed to examine the specific contribution of NOX5 in renal pathology using NOX5 transgenic mouse as well as human kidney organoid model of DKD. This study also compares NOX5 with the previously characterized NADPH oxidase, NOX4, in terms of their relative roles in DKD.

Materials and methods: We examined the expression of human NOX5 and NOX4 as well as ROS production in human kidney biopsies and in human renal organoids and cells in diabetic conditions. We assessed the effect of NOX5 or NOX4 inhibition either alone or combined using genetic silencing and pharmacological inhibition approaches in human renal cells and organoids exposed to diabetic milieu environment. In animal experiments, we assessed the effect of human NOX5 overexpression independent of NOX4 in Nox5 transgenic mice in the presence or absence of diabetes. In addition, we also examined the level of urinary NOX5 in various stage of diabetic subjects with nephropathy.

Results: In comparison to NOX4, genetic silencing or inhibition of NOX5 showed robust attenuation of high glucose induced gene expression of markers of fibrosis (CTGF and collagen IV) and inflammation (MCP-1 and TLR4) as well as downregulation of ROS-sensitive factors EGR-1 and PKC-α and cell cycle regulator, P21 in cellular and organoid models of DKD. We also observed that NOX5 is upstream of NOX4 and that NOX5 inhibition also downregulates NOX4, but not vice versa. In vivo, overexpression of NOX5 independent of NOX4 pathway demonstrated an enhanced renal pathology as evidenced from increase in albuminuria, renal fibrosis and inflammation in diabetic mice. Additionally, we observed a positive correlation between urinary NOX5 excretion and levels of albuminuria in diabetic individuals with nephropathy.

Conclusion: These findings underscore the potential therapeutic significance of NOX5 inhibition over NOX4 in human DKD. The study proposes that inhibiting NOX5 alone could be a sufficient for mitigating the progression of DKD and strengthens the case for the development of NOX5-specific inhibitors as a potential therapeutic intervention. NOX5 could also represent a biomarker for diabetic kidney disease.

Disclosure: J. Jha: None.

198

Does the expression of glucagon-like peptide-1 receptor on bone marrow derived cells influence the onset and progression of diabetic kidney disease?

K.C. Sourris ¹, C. Bertuzzo Veiga², P. Kantharidis¹, C.J. Rosado¹, M. Bose¹, A. Dimitropoulos¹, Y. Xu², Y. Zhang², M.T. Coughlan¹, M. Aziz¹, P. Morgan², D.J. Drucker³, A.J. Murphy², K.A.M. Jandeleit-Dahm¹, M.E. Cooper¹;

¹Central Clinical School, Monash University, Melbourne, Australia, ²Baker Department of Cardiometabolic Health and Division of Immunometabolism, Baker Heart and Diabetes Institute, Melbourne, Australia, ³Lunenfeld-Tanenbaum Research Institute, Toronto, ON, Canada.

Background and aims: There are a number of therapies which target the GLP-1 pathway to improve glucose control in type 2 diabetic subjects with increasing evidence from clinical trials that such approaches may protect against end-organ complications including diabetic kidney disease. We have recently demonstrated anti-inflammatory actions of GLP-1 agonists ¹ which could explain in part their renoprotective actions. With widespread distribution of GLP-1 receptors (GLP-1R) including in immune cells we determined the influence of bone-marrow (BM) derived GLP-1R expression on progression of diabetic kidney disease

Materials and methods: We sub-lethally irradiated 8 week old C57BL/6 (WT) mice and subsequently reconstituted the mice with bone marrow from either syngeneic or allogeneic mouse genotypes (n=10/group). After 6 weeks of recovery, these mice were randomised to controls or rendered diabetic using a low dose streptozotocin protocol (5x 55mg/kg) and followed for 20 weeks. In total we had four groups with the following donor/recipient combinations: WT/WT, GLP-1R/WT controls and diabetic as well as concurrently run non-transplanted WT control and diabetic mice followed for 20 weeks. At the completion of the study, kidneys, bone marrow, plasma and urine were collected for subsequent analysis.

Results: Metabolic parameters including glycated haemoglobin and blood glucose were elevated in all diabetic groups (p<0.05) but not significantly different among the various genotypes. All diabetic WT/WT mice exhibited increased albuminuria (53.0 ± 9.8 vs control 6.4 ± 1.1 μg/24hrs). Interestingly, WT mice that received GLP-1R(-/-) bone marrow (GLP1R/WT), exhibited increased albuminuria (83.7 ± 13.3 μg/24hrs p<0.05) reflecting worse renal disease. Furthermore, bone marrow analysis of these mice demonstrated that the GLP-1R/WT chimeric mice had elevated levels of common myeloid progenitor cells (0.19 ± 0.01% vs. 0.14 ± 0.02 % of bone marrow, p<0.05) when compared to the WT/WT group.

Conclusion: Our data indicate that selective deletion of the GLP-1R from bone marrow derived cells, significantly aggravates renal injury in experimental diabetes. This may reflect an anti-inflammatory role for the GLP-1/GLP-1R axis involving BM derived cells which influence renal structure and function. These findings provide a further mechanistic explanation as to how GLP-1 agonists may confer renoprotection in diabetes.

Disclosure: K.C. Sourris: None.

OP 34 [B]reaking [F]at

199

Metabolic remodelling in adipose tissue: insights from Roux-en-Y gastric bypass surgery on fructose production

A. Reis-Costa ¹, P. Katsogiannos², M. Pereira², J. Jones¹, J. Eriksson²;

¹University of Coimbra, Institute for Interdisciplinary Research, Coimbra, Portugal, ²Clinical Diabetes and Metabolism, Uppsala University, Uppsala, Sweden.

Background and aims: Consumption of a hypercaloric diet rich in glucose and fructose is a risk factor for developing Type 2 Diabetes (T2D), and these sugars have become widely used as sweeteners in processed foods. Fructose is a well-known adipogenic and lipogenic nutrient and its metabolism is not insulin-regulated. In humans, hyperglycaemia is estimated to activate endogenous fructose production via the polyol pathway (POP) in many tissues, contributing to the development of many Diabetes comorbidities. However, its impact on the adipose tissue is still unclear. Our goal was to assess the impact of weight loss intervention with bariatric surgery on key enzymes involved in fructose metabolism in adipose tissue of subjects with T2D. We used a cohort undergoing Roux-en-Y gastric bypass (RYGB) surgery, showing metabolic improvements and T2D remission alongside weight loss.

Materials and methods: Subjects were recruited at the outpatient clinic of the Uppsala University Hospital. Thirteen patients were assessed at baseline, 4, 24 and 104 weeks after RYGB (preceded by 4 weeks of low-calorie diet). Subcutaneous adipose tissue needle biopsies were collected at each timepoint to be analysed for gene expression of key factors involved in fructose metabolism by transcriptomics. Statistical analysis consisted of mixed effects analysis and data is shown as mean with standard error.

Results: Aldose reductase, the rate-limiting enzyme of POP, gene expression significantly decreased at 104 weeks post-surgery compared to baseline (9.5±0.2 FPKM vs 11.8±0.5 FPKM, p=0.02). The gene expression of glucose 6-phosphate dehydrogenase and 6-phosphogluconolactonase, which are enzymes of the oxidative branch of the pentose phosphate pathway, significantly decreased compared to baseline (11.9±0.4 FPKM vs 14.4±0.5 FPKM, p=0.02; 16.3±0.6 FPKM vs 18.4±0.9 FPKM, p=0.04). Additionally, the gene expression of deoxyribose-phosphate aldolase, an enzyme that can feed glyceraldehyde 3-phosphate to the reducing end of the pentose phosphate pathway, is significantly increased at 104 post-RYGB compared to baseline. Glycerate kinase, an enzyme of fructolysis, is significantly increased from week 24 post-RYGB compared to baseline (4.6±0.6 FPKM vs 2.8±0.2, p=0.02; 3.8±0.4 FPKM vs 2.8±0.2, p=0.04).

Conclusion: The observed results indicate that RYGB surgery caused a significant downregulation of the polyol pathway, suggesting decreased endogenous fructose production that was associated with a shift in the pentose phosphate pathway away from its oxidative mode. Moreover, they hint at increased demand for glyceraldehyde 3-phosphate and 3-phosphoglycerate, which are both intermediates of glycolysis and glyceroneogenesis. Altogether, these findings indicate that RYGB leads to T2D remission partly by promoting glucose utilization in adipose tissue via glycolysis and lipogenesis, instead of POP.

Clinical Trial Registration Number: NCT02729246

Supported by: Swedish Diabetes Foundation, EXODIAB, ALF, Ernfors foundation, Uppsala County Council, Uppsala University Hospital, FCT2022.10983.BD

Disclosure: A. Reis-Costa: None.

200

Exercise training regulates long noncoding RNAs expression in subcutaneous WAT of obese participants

P. Nigro, T. Caputo, L.K. Simpson, M.M. Columbus, J. Ding, M.F. Hirshman, R.J.W. Middelbeek, L.J. Goodyear;

Joslin Diabetes Center, Boston, MA, USA.

Background and aims: Obesity can result in extensive metabolic dysfunction to white adipose tissue (WAT), changes that contribute to metabolic diseases. In contrast, studies performed primarily in rodents show that exercise training causes robust beneficial adaptations to WAT and that these adaptations play a critical role in the benefits of exercise on metabolic health. The molecular mechanisms by which exercise and obesity affect WAT in humans are poorly understood, but such knowledge could reveal novel strategies for treatment of obesity and metabolic disease. Here, our aim was to determine the effects of exercise training and obesity on long non-coding RNAs (lncRNAs), an emerging, yet understudied area of metabolism research. lncRNAs have been widely investigated in other chronic diseases where they have been shown to mediate critical cellular functions including gene expression through epigenetic, transcriptional, and translational modifications.

Materials and methods: Participants (14F/9M; 25-55 years; normoglycemic) were divided into two groups based on BMI. Lower BMI group were <27, 23.4±4.9kg/m² mean ± SEM; Higher BMI were >28, 30.2±3 kg/m². Subcutaneous WAT (scWAT) biopsies and clinical measurements were obtained before and after 10 wks of moderate-intensity endurance training, training that resulted in similar increases in VO_2peak (Lower BMI=5.6% p<0.03; Higher BMI=6.9% p<0.01). scWAT samples were used for RNAseq.

Results: We identified 4161 annotated lncRNAs which accounted for 88% of all non-coding RNA species sequenced. At baseline (pre training), higher BMI individuals exhibited differential expression of 145 lncRNAs (44 down, 101 up; p<0.05) compared to lower BMI. KEGG analysis showed prominent involvement of these obesity regulated lncRNAs in multiple pathways including insulin signaling, inflammation, hypoxia, and cellular senescence. Exercise training in higher BMI altered 160 lncRNAs (41 down, 119 up; p<0.05), and KEGG analysis predicted involvement of these lncRNAs in AMPK, mTOR, and longevity pathways. Exercise training in lower BMI altered 341 lncRNAs (45 down, 296 up; p<0.05), similarly in higher BMI participants KEGG analysis predicted involvement in increased insulin, mTOR, and neurotrophic signaling, and specific to lower BMI participants, decreased ECM signaling. Next, we searched for lncRNAs that were upregulated in higher BMI and downregulated with exercise training, as well as highly correlated with anthropometric and biochemical parameters. Of these, Tumor Protein P63 Regulated 1 (TPRG1-AS1), an antisense RNA with no known function, was elevated by 23% in higher BMI participants and decreased by 40% with training (p<0.05). TPRG1-AS1 correlated with both % body fat (r = 0.27, p< 0.01) and BMI (r = 0.17, p=0.05). To predict which cell types in scWAT express TPRG1-AS1, we performed deconvolution analysis and found that TPRG1-AS1 is exclusively expressed in mature hypertrophic adipocytes (p<0.0001). Interestingly, pathway analysis of genes co-expressed with TPRG1-AS1 suggested its potential involvement in ECM organization and lipid metabolism. Thus, TPRG1-AS1 is a novel adipocyte-specific lncRNA, upregulated in higher BMI subjects, modulated by exercise training, and likely involved in multiple adipose tissue functions.

Conclusion: Exercise training has dramatic effects on the lncRNAs profile in human scWAT, suggesting that this class of RNAs is critical for regulation of scWAT phenotype.

Supported by: R01DK099511 and R01DK101043 (to LJG) K23DK114550 (to RJWM)

Disclosure: P. Nigro: None.

201

Multi-omics analysis of human brown fat reveals novel candidates for the thermogenic activation of adipose tissue

M. Balaz ¹, L. Petriskova¹, D. Olesova¹, A. Ghosh², A. Kvasnicka³, N. Palesova¹, P. Stefanicka^4,5, L. Varga⁵, A. Jancovicova¹, P. Makovicky¹, D. Dobesova³, D. Friedecky³, C. Wolfrum², L. Balazova¹;

¹Biomedical Research Center, Slovak Academy of Sciences, Bratislava, Slovakia, ²Institute of Food, Nutrition and Health, ETH Zurich, Schwerzenbach, Switzerland, ³University Hospital Olomouc and Faculty of Medicine and Dentistry, Palacky University Olomouc, Olomouc, Czechia, ⁴Bory Hospital, Bratislava, Slovakia, ⁵Faculty of Medicine and University Hospital, Comenius University, Bratislava, Slovakia.

Background and aims: Since energy expenditure increases as a consequence of thermogenesis, pharmacological induction of this process represents an intriguing therapeutic approach. Despite the identification of several bioactive molecules with the potential to activate adipose tissue thermogenesis to date, none of them has proven effective or produced significant health benefits in clinical trials. Therefore, our aim is to identify novel candidates with the potential to increase the thermogenic activity of adipose tissue through multi-omics analysis of paired human brown and white adipose tissue biopsies.

Materials and methods: Deep neck brown adipose tissue samples, exhibiting enrichment of UCP1 mRNA and multilocular lipid droplet morphology, along with adjacent subcutaneous white adipose tissue biopsies from 15 patients undergoing thyroid surgery, underwent transcriptomic (next generation RNA sequencing), metabolomic, and lipidomic (Mass Spectrometry) analyses. These analyses were followed by extensive bioinformatics, including pathway analyses utilizing the GO (Gene Ontology) enrichment, KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway, and BioPAN (Bioinformatics Methodology for Pathway Analysis).

Results: Using this approach, we identified 4174 transcripts (FDR˂0.01), 59 metabolites (p˂0.05) and 240 lipids species (p˂0.05) differentially regulated between human brown and white fat. Robust evidence shows enrichment of membrane and mitochondrial lipids in brown fat, while neutral glycerolipids and fatty acids predominate in white adipose tissue. Pathway analysis on the lipidomic data (BioPAN) identified activation of phosphatidylserine synthesis and triglyceride reesterification in brown fat, and proposed several lipid-processing genes, including DGAT2, PISD, PTDSS1, CHPT1 and PEMT as interesting candidates. In addition, all five identified candidates are among the differentially regulated genes identified by the transcriptomic analysis. Furthermore, multi-omics factor analysis identified two factors which explain almost 60% of the variability between brown and white adipose tissue, and purine metabolites as the most prominent features contributing to these factors.

Conclusion: Our study providing comprehensive multi-omics analysis of human brown and white adipose tissue biopsies identified several novel candidate transcripts, metabolites, and lipids with the potential to increase the thermogenic activity of adipose tissue. Further studies are needed to uncover their role in the control of non-shivering thermogenesis.

Supported by: SASPRO2 No. 1148/01/02 and APVV-22-0291

Disclosure: M. Balaz: None.

202

A multi-omics approach to understand the underlying mechanisms of obesity

J. Niu, C. Gieger, A. Peters, S. Sharma, H. Grallert;

Helmholtz Munich, Neuherberg, Germany.

Background and aims: Body mass index (BMI) associated changes in gene expression and proteins in whole blood reflect molecular responses to adiposity variation, shedding light on its potential mechanisms and target spots. To characterize these changes, mRNA and protein analyses were done.

Materials and methods: We analyzed 13,320 genes in RNA sequencing (RNA-Seq) data and 809 proteins obtained in whole blood using a Liquid chromatography-mass spectrometry (LC-MS) method in 1,414 participants from the Cooperative Health Research Area Augsburg (KORA) FF4 cohort (2013/2014). Using two groups categorized as obese and non-obese, and considering BMI as the outcome, we performed differential gene expression analysis using "DESeq2" and "limma". To investigate whether changes in gene expression lead to protein alterations, we calculated correlations between mRNAs and proteins and constructed a linear regression model for proteomics. Additionally, we performed gene enrichment analysis and protein-protein interaction (PPI) analysis to explore potential mechanistic pathways involved. Mendelian randomization (MR) was used with Expression Quantitative Trait Loci (eQTL) and Genome-Wide Association Study (GWAS) data to assess the relationship between mRNAs and obesity.

Results: In the obese and non-obese groups, we identified 459 genes in the basic model corrected for age and sex only, while in the full model including additional covariates, sex, age, smoking status, physical activity, high-density lipoprotein cholesterol (HDL-C), systolic blood pressure, triglycerides, fasting glucose, and leukocytes, 312 genes showed significant changes. For the BMI-related genes, the basic model was replicated, revealing 3,983 genes. Functional enrichment analysis based on 439 overlapping genes in two basic models revealed several obesity-related processes, including erythrocyte metabolism, inflammation, and oxidative stress. ZNF827 was identified as the gene most significantly correlated with protein levels, and S100A9 was validated in proteomics via logistic regression. MR analysis revealed 14 genes influenced by obesity, with DNAJB2 showing a bidirectional causal relationship.

Conclusion: Our study provides a comprehensive framework of genes, proteins, and pathways associated with obesity, with findings replicated or reported in other previous studies. These results contribute to a deeper understanding of the molecular mechanisms underlying obesity and potential gene-targeted interventions.

Disclosure: J. Niu: None.

203

HOOK1: A potential novel target gene for obesity?

S.I.A. Svensson ¹, S. Saeed¹, T. Valderhaug², M. Dahl¹, B. Bai³, A. Cayir³, T. Rønningen¹, Y. Böttcher¹;

¹Department of Clinical Molecular Biology, University of Oslo, Lørenskog, Norway, ²Endocrinology, Akershus University Hospital, Lørenskog, Norway, ³Medical Division, Akershus University Hospital, Lørenskog, Norway.

Background and aims: Omental visceral adipose tissue (OVAT) associates with an increased risk of metabolic diseases compared to subcutaneous adipose tissue (SAT). To identify novel targets for obesity, we here applied a combination of ATAC seq (Assay for Transposase-Accessible Chromatin using sequencing), RNA seq, DNA methylation and gene expression analysis in paired samples of human OVAT and SAT.

Materials and methods: Intra-individually paired biopsies of OVAT and SAT were collected from obese patients (BMI >35; n= 78) during bariatric surgery, from which DNA and RNA were isolated. Previously generated in-house datasets of ATAC seq and RNA seq from OVAT and SAT were then integrated with publically available data on genome-wide DNA methylation levels from whole blood. Gene expression levels were measured using TaqMan Assays for RT-qPCR while a Q48 PyroMark was used for pyrosequencing to generate DNA methylation levels including a total of 12 CpG sites. Spearman´s correlation analysis together with regression models and paired sample t-tests was utilised for the statistical analysis.

Results: The integrated analysis of datasets including RNA-seq and genome-wide DNA methylation provided a list of 39 genes that are upregulated in OVAT. Further integration of the ATAC-seq and DNA methylation data generated a list of 19 targets upregulated in OVAT and presenting open chromatin in the promoter region. A novel target gene for obesity (HOOK1) was selected based on open chromatin regions along with differential gene expression levels in the discovery cohort used for the previously generated in-house datasets. A significantly higher expression level in OVAT compared to SAT was confirmed to be consistent throughout the larger cohort (p <0.001). Spearman´s correlation analysis of gene expression with clinical traits related to obesity returned a negative correlation with BMI in OVAT (p = 0.044). We further identified that DNA methylation levels at several CpG sites correlate with BMI (p = 0.022), hip circumference (p = 0.031), and markers of liver function such as ALAT (p = 0.002), ASAT (p = 0.002), ALP (p = 0.005) and GGT (p = 0.022) as well as thyroid function (TSH (p = 0.039) and FT4 (p = 0.033). Finally, when taking into account potential covariates such as age and gender, we find a positive association for markers related to liver function with DNA methylation at several CpG sites (p = 0.001, p = 0.004).

Conclusion: We present an identification strategy to identify a novel target gene for obesity and related clinical traits and its correlation with DNA methylation.

Clinical Trial Registration Number: 489516

Supported by: Helse Sør-Øst #2020002, Norway

Disclosure: S.I.A. Svensson: None.

204

Differences in adipose tissue-, liver- and jejunal gene expression between obese patients with and without glomerular hyperfiltration

S.T. Jagroep, M. Koning, V.E.A. Gerdes, M. Nieuwdorp, A.S. Meijnikman, D.H. van Raalte;

Amsterdam UMC, Amsterdam, Netherlands.

Background and aims: Glomerular hyperfiltration (GH) plays a pivotal role in the onset of obesity-related glomerulopathy (ORG), a frequent complication of obesity and driver of chronic kidney disease (CKD). Associations between adipose tissue compartments and adipose tissue-derived hormones and kidney disease have been published. In contrast, less is known about factors driving glomerular hyperfiltration. In addition to intrakidney factors, other organs such as intestine, liver and adipose tissue may affect glomerular hyperfiltration. Therefore, we aimed to investigate the differences in subcutaneous (SAT) and visceral adipose tissue (VAT), liver- and jejunal gene expression between obese patients with- and without glomerular hyperfiltration.

Materials and methods: This was an observational cohort study set in the The Netherlands. Included participants were men or women scheduled to undergo bariatric surgery, aged <65 years with an BMI >35 kg/m² . Glomerular filtration rate (GFR) was estimated using the Cockgroft gault creatine clearance and was corrected for body surface area (BSA). Patients were divided into 3 tertiles according to eGFR and differences in gene expression were calculated for the highest GFR tertile in relation to the lowest GFR tertile. VAT, SAT, liver biopsies and jejunal biopsies were all collected during scheduled bariatric surgery. RNA sequencing was performed at Novogene on an HiSeq with 150 bp paired-end reads and 10G data/sample. Differential gene expression was used to identify differences in gene expression. Main outcomes measured gene expression differences in liver, jejunal, VAT, and SAT between hyperfilterers and those without hyperfiltration.

Results: 415 obese patients were included of whom 75 were men. Median age [IQR] was 48.00 [39.50, 54.00] years. Median eGFR [IQR] in the lowest tertile and highest tertile were 145.13 [125.11, 156.01] and 225.15 [205.09, 248.36] respectively. In the liver, jejunum, SAT and VAT fat gene expression analysis showed 505, 294 242 and 57 differentially expressed genes (DEGs), respectively with an adjusted P-value <0.05. Liver tissue showed upregulation of 300 genes and downregulation of 205 genes. Upregulated- and downregulated genes included genes coding for transcriptions factors and the complement system respectively. Jejunal tissue showed upregulation of 78 genes and downregulation of 218 genes some of which associated with the complement system. SAT showed upregulation of 100 genes and downregulation of 29 genes. Upregulated genes included the gene coding for albumin. VAT showed upregulation of 28 genes and downregulation of 29 genes. Downregulated genes included the Lactase coding gene. Functional analyses through KEGG pathway enrichment showed that patients with hyperfiltration exhibited upregulation of the MAPK signaling pathway, oxidative phosphorylation, ERBB signaling, and chemokine signaling pathway compared to patients without hyperfiltration.

Conclusion: Multiple genes in the liver, jejunum and adipose tissue were significantly upregulated or downregulated in patients with obesity and hyperfiltration compared to patients with obesity without hyperfiltration. Moreover, these genes are involved in important immune and metabolic functions. This provides us with evidence that altered gene expression in obesity in liver, fat and intestine could be associated with glomerular hyperfiltration.

Disclosure: S.T. Jagroep: None.

OP 35 Insulin: ready, steady, go!

205

P21-activated kinase: a novel regulator of the FGF1/PDE4D anti-lipolytic pathway and insulin resistance

J. Seigner ¹, J. Krier¹, D. Spähn², J.L. Nono¹, R. Lukowski², A.L. Birkenfeld¹, G. Sancar¹;

¹Department of Internal Medicine IV, Diabetology, Endocrinology and Nephrology, IDM, Universitätsklinikum Tübingen, Tübingen, Germany, ²Department of Pharmacology Toxicology and Clinical Pharmacy, Institute of Pharmacy, University of Tübingen, Tübingen, Germany.

Background and aims: Type 2 diabetes (T2D) is one of the most frequent chronic diseases worldwide with an estimated prevalence of roughly 10 %. While underlying causes of T2D are multifactorial, adipose tissue dysfunction, lipotoxicity and insulin resistance play a major role. Hence, identifying novel factors regulating adipose tissue lipolysis and insulin signaling would enable novel therapeutic options to treat T2D. We have previously identified the Fibroblast growth factor-1 (FGF1) / Phosphodiesterase 4D (PDE4D) pathway that can lower blood glucose via suppression of lipolysis in adipose tissue and by regulating hepatic glucose production. While we showed that phosphorylation of PDE4D is critical for its antilipolytic and antidiabetic activities, the signaling cascade regulating PDE4D phosphorylation is unknown. In this study, we aimed to uncover the molecular mechanisms regulating FGF1/PDE4D-dependent lipolysis and ultimately new targets for the treatment of T2D.

Materials and methods: We have performed a targeted screen for kinase inhibitors that interfere with the antilipolytic role of FGF1 in adipocytes. Candidates are investigated for the effect on PDE4D phosphorylation (western blotting), lipolysis (free fatty acid (FFA) measurements), adipogenesis (gene expression with qPCR and lipid staining with Nile red), and insulin signaling (western blotting). Moreover, FRET-based cAMP biosensors are used to determine the effect of the candidate kinases on cAMP levels in live adipocytes.

Results: Our targeted screen in adipocytes revealed p21-activated kinase (PAK) as the regulator of the antilipolytic function of FGF1, PDE4D phosphorylation, and PDE4D activity. Acute inhibition of PAK reversed the suppression of lipolysis by FGF1 (Cntrl 1.97 μMol FFA/mg protein, FGF1 0.48 μMol FFA/mg protein vs PAK inhibitor FGF1 2.11 μMol FFA/mg protein, p< 0.001). Moreover, FGF1-induced PDE4D phosphorylation is reversed in a dose-dependent manner. Over-expression of PDE4D suppresses lipolysis in adipocytes (Cntrl 1.99 μMol FFA/mg protein vs PDE4D over-exp 0.65 μMol FFA/mg protein, p< 0.001) which can be reversed by PAK inhibition (1.59 μMol FFA/mg protein, p< 0.01). Inhibition of PAK increased the response to isoproterenol-induced cAMP production in adipocytes (p < 0.001) supporting the involvement of PAK in regulation of lipolysis. Chronic (2-day) inhibition of PAK activity decreased gene expression of adipogenic markers such as Pparg, Glut-4, adiponectin, Adrb3, and Hsl (>50%, p <0.01) and decreased adipogenesis by 35 % (p < 0.001). Finally, chronic inhibition of PAK activity resulted in a 75 % decrease in insulin signaling (p < 0.01) as assessed by AKT phosphorylation.

Conclusion: In this study, we identified PAK as novel component of the FGF1/PDE4D signaling cascade. The dependence of the antilipolytic function of FGF1 on PAK activity indicates a potential antidiabetic role for PAK in vivo. Moreover, our study suggest that PAK plays an important role in adipogenesis and insulin resistance.

Supported by: DZD Grant 2024

Disclosure: J. Seigner: None.

206

Investigating insulin-lowering mechanisms of SGLT2-inhibition in individuals with postbariatric hypoglycaemia

E. Faggionato ¹, D. Herzig², C. Dalla Man¹, L. Bally², M. Schiavon¹;

¹Department of Information Engineering, University of Padova, Padova, Italy, ²Bern University Hospital, University of Bern, Bern, Switzerland.

Background and aims: Roux-en-Y gastric bypass (RYGB) surgery increases postprandial glucose variability with higher peak and lower nadir values, predisposing to postbariatric hypoglycemia (PBH). The SGLT2-inhibitor empagliflozin was shown to reduce glucose excursions and lower insulin exposure in patients with PBH. In this work, we aimed to explore the insulin-lowering mechanisms of empagliflozin 25mg, by evaluating the effects on insulin secretion and extraction using the Oral C-peptide and Insulin Minimal Model (OCIMM).

Materials and methods: In a randomized two-period double-blind crossover trial, 22 post-RYGB adults with biochemically confirmed PBH were randomized to empagliflozin 25mg or placebo once daily over 20 days. After each 20-day period, participants underwent a mixed-meal tolerance test with frequent measurement of plasma glucose, C-peptide, and insulin over 2 hours. The OCIMM assumes that insulin and C-peptide secretion is made up of a static component, controlled by glucose level, and a dynamic component, driven by glucose rate of change, and that the first pass hepatic insulin extraction is a piece-wise linear function with coefficients to be estimated from the data. The model was identified assuming C-peptide and post-hepatic insulin kinetics to be identical between visits; while, for each visit, the model provided indices of basal (Φ_b) and total (Φ_tot) β-cell responsivity to glucose, and basal (HE_b) and total (HE_tot) hepatic insulin extraction. Finally, results for the two periods were compared.

Results: Results are reported in Table 1. The reduced insulin exposure observed with empagliflozin compared with placebo was attributable to a significant increase in total hepatic insulin extraction (HE_tot, p=0.006). Conversely, Φ_b, Φ_tot, and HE_b were not altered with empagliflozin vs placebo.

Conclusion: In addition to the well-established non-insulin-dependent glucose-depleting effects, empagliflozin was shown to reduce insulin exposure in post-RYGB by increasing total hepatic insulin extraction, without affecting insulin secretion.

Clinical Trial Registration Number: NCT05057819

Supported by: Swiss National Science Foundation PCEGP3_186978

Disclosure: E. Faggionato: None.

207

Insulin-degrading enzyme couples glucagon and insulin-mediated regulation of mitochondrial dynamics in hepatocytes

P. Cámara-Torres, S. Hernández de la Red, B. Merino, J. Santo-Domingo, M. de la Fuente, A. Alonso, I. Cózar-Castellano, G. Perdomo;

Institute of Biology and Molecular Genetics, Valladolid, Spain.

Background and aims: Mitochondrial dynamics play a crucial role in the pathophysiology of type 2 diabetes (T2D) by regulating glucose homeostasis and insulin sensitivity. In individuals with T2D, significant alterations in mitochondrial dynamics are observed, in addition to decreased hepatic insulin-degrading enzyme (IDE) levels. IDE is a ubiquitous metalloprotease present in two isoforms, cytosolic (IDEcyt) and mitochondrial (IDEmt), but the contribution of each isoform to the regulation of mitochondrial homeostasis in hepatocytes remains unknown. Here, we aimed to dissect the roles of IDEcyt and IDEmt on mitochondrial dynamics in response to glucagon and insulin in hepatocytes.

Materials and methods: A stable IDE knockout clone from human HepG2 cell line (HepG2-IDE-KO) was generated by CRIPR/Cas9. HepG2-IDE-KO cells were reconstituted with IDE by transfection. Confocal Z-stacks of cells expressing the matrix-targeted red fluorescent protein Cox8(MTS)-RFP were acquired and mitochondrial morphology and network was analyzed with ImageJ/Fiji Mitochondrial Analyzer plugin.

Results: Mitochondrial dynamics in HepG2 were shown to be mediated through IDE. Upon glucagon stimulation, augmented, elongated, and more interconnected mitochondria were observed. However, in the absence of IDE (HepG2-IDE-KO cells), glucagon had no impact on the mitochondrial network. Interestingly, when IDEmt was reconstituted in HepG2-IDE-KO cells, the effects of glucagon on mitochondrial dynamics were restored. However, IDEcyt had no effect on glucagon-mediated effects on mitochondrial dynamics. In support of these results, lower Drp1 levels were observed in cells reconstituted with IDEmt. Taken together, our findings indicate a role of mitochondrial IDE isoform in governing mitochondrial dynamics regulation. On the other hand, in response to insulin stimulation, HepG2 cells do not undergo changes in the mitochondrial network. However, in the absence of IDE, elongation and interconnection of mitochondria occurs. Reconstitution of IDEmt isoform leads to mitochondrial fragmentation, whereas the IDEcyt isoform causes the opposite effect, promoting mitochondrial fusion.

Conclusion: Our findings suggest distinct roles for IDE isoforms in regulating mitochondrial dynamics, indicating IDE’s role in maintaining mitochondrial integrity under insulin and glucagon stimulation. Understanding the mechanisms underlying these mitochondrial changes may open new therapeutic strategies for the treatment of T2D, focusing on restoration of mitochondrial function and regulation of mitochondrial dynamics as potential targets for these regulatory mechanisms.

Supported by: MICIN/AEI

Disclosure: P. Cámara-Torres: None.

208

Activation of the hypoxia-inducible factor pathway by roxadustat improves glucose metabolism in human primary myotubes

S. Mäkinen ¹, S. Sree¹, T. Ala-Nisula², H. Kultalahti¹, P. Koivunen², H.A. Koistinen¹;

¹Minerva Foundation Institute for Medical Research, Helsinki University Hospital, Helsinki, Finland, ²Faculty of Biochemistry and Molecular Medicine, University of Oulu, Oulu, Finland.

Background and aims: Hypoxia-inducible factor prolyl 4-hydroxylase (HIF-P4H) enzymes regulate adaptive cellular responses to low oxygen concentrations. Inhibition of HIF-P4Hs leads to stabilization of hypoxia-inducible factors (HIFs) and activation of the HIF pathway affecting multiple biological processes to rescue cells from hypoxia. As evidence from animal models suggests that HIF-P4H inhibitors could be utilized to treat metabolic disorders associated with insulin resistance, we examined whether roxadustat, a HIF-P4H inhibitor approved for the treatment of renal anemia, would have an effect on glucose metabolism in primary human myotubes.

Materials and methods: Primary skeletal muscle cell cultures, established from M. vastus lateralis biopsies from men with normal glucose tolerance (n=5) and type 2 diabetes (n=8), were treated with 10 μmol/l roxadustat for 24 h. Induction of HIF target gene expression was detected with quantitative real-time PCR. Glucose uptake and glycogen synthesis were investigated with radioactive tracers. Glycolysis and mitochondrial respiration rates were measured with a Seahorse analyzer.

Results: Exposure to roxadustat stabilized nuclear HIF1α protein expression in human myotubes. Treatment with roxadustat led to induction of HIF target gene mRNAs for GLUT1, HK2, MCT4, and HIF-P4H-2 in myotubes from donors with normal glucose tolerance, with a blunted response in myotubes from donors with type 2 diabetes. mRNAs for LDHA, PDK1 and GBE1 were induced similarly in myotubes from donors with normal glucose tolerance or type 2 diabetes. Exposure of myotubes to roxadustat led to an increase in glycolytic rate by 1.4-fold in myotubes from men with normal glucose tolerance (p=0.0370) and 1.7-fold in myotubes from donors with type 2 diabetes (p=0.0044), with no difference between the groups (p=0.1391). Exposure to roxadustat led to a reduction in basal mitochondrial respiration in both groups (p<0.01). Basal glucose uptake rates were similar in myotubes from donors with normal glucose tolerance (20.2±2.7 pmol mg^-1 min^-1) and type 2 diabetes (25.3±4.4 pmol mg^-1 min^-1, p=0.4205). Treatment with roxadustat enhanced insulin-stimulated glucose uptake to myotubes from donors with normal glucose tolerance (1.4-fold when compared to insulin-only condition, p=0.0023). Basal rate of glucose incorporation into glycogen was lower in myotubes from donors with normal glucose tolerance (233±12.4 nmol g^-1 h^-1) than in myotubes from donors type 2 diabetes (360±40.3 nmol g^-1 h^-1, p=0.0344). Insulin increased glycogen synthesis 1.9-fold (p=0.0025) in myotubes from donors with normal glucose tolerance, whereas roxadustat did not affect their basal or insulin-stimulated glycogen synthesis. Insulin increased glycogen synthesis by 1.7-fold (p=0.0031) in myotubes from donors with type 2 diabetes. While basal glycogen synthesis was unaffected by roxadustat, roxadustat pretreatment enhanced insulin-stimulated glycogen synthesis in myotubes from donors with type 2 diabetes (p=0.0345).

Conclusion: Roxadustat increases glycolysis and inhibits mitochondrial respiration in primary human myotubes regardless of diabetes status, and it may improve insulin action on glycogen synthesis in myotubes from donors with type 2 diabetes.

Supported by: Helsinki University Hospital, Finska Läkaresällskapet, Liv och Hälsa

Disclosure: S. Mäkinen: None.

209

Characterisation of new tissue-specific insulin resistance clusters in persons with and without diabetes and their relationship to comorbidities

K.B. Bódis ^1,2, K. Prystupa², O.P. Zaharia^1,2, M. Schön^1,2, V. Schrauwen-Hinderling², G.J. Bönhof^1,2, H.J. Hauner³, S. Meyhöfer⁴, A.L. Birkenfeld⁵, A.F.H. Pfeiffer⁶, M. Blüher⁷, S.R. Bornstein⁸, J. Szendrödi⁹, M. Roden^1,2, R. Wagner^1,2;

¹Department of Endocrinology and Diabetology, Medical Faculty and University Hospital Düsseldorf, Heinrich Heine University Düsseldorf, Düsseldorf, Germany, ²Institute for Clinical Diabetology, German Diabetes Center, Leibniz Center for Diabetes Research at Heinrich Heine University Düsseldorf, Düsseldorf, Germany, ³Institute of Nutritional Medicine, School of Medicine and Health, Technical University of Munich, Munich, Germany, ⁴Department of Medicine I, University Hospital of Schleswig-Holstein, Luebeck, Germany, ⁵Institute for Diabetes Research and Metabolic Diseases, University of Tübingen, Tübingen, Germany, ⁶Dept. Endocrinology, Diabetes and Nutrition, Charité Universitätsmedizin Berlin, Berlin, Germany, ⁷Department of Medicine, Endocrinology and Nephrology, University of Leipzig, Leipzig, Germany, ⁸Department of Internal Medicine III, Dresden University of Technology, Dresden, Germany, ⁹Department of Medicine I and Clinical Chemistry, University Hospital of Heidelberg, Heidelberg, Germany.

Background and aims: Insulin resistance constitutes a key pathological aspect in the development of type 2 diabetes (T2D), and its presence varies across type 1 diabetes (T1D) and prediabetes. This study seeks to unveil distinctive profiles of insulin resistance in specific tissues, aiming to unravel its role in metabolic diversity.

Materials and methods: In this study, k-means clustering was performed in 759 individuals (223 T1D, 346 T2D, 190 glucose-tolerant controls, CON) of the German Diabetes Study (GDS) with recent diabetes diagnosis (<1 year). Clusters were defined using glycemia (HbA1c), whole-body (mainly muscle) insulin sensitivity (M-value, Botnia Clamps with D-[6,6-²H₂]glucose), fasting endogenous glucose production and non-esterified fatty acids. Diabetes-related outcomes such as nephropathy, peripheral neuropathy, and hepatic lipid content (HLC) were assessed during 5-year follow-ups as part of the in-depth metabolic phenotyping of the GDS. Specific events, such as newly occurring myocardial infarctions and the initiation of insulin therapy, were collected annually by telephone interviews. Adjustments for age, gender, BMI, and diabetes type were applied to all findings.

Results: Three distinct clusters emerged: a whole-body insulin sensitive (WISE, consisting of 73 T1D, 41 T2D, 151 CON), and two whole-body insulin resistant clusters, one of these displayed elevated insulin resistance in adipose tissue (AIRE, composed of 50 T1D, 210 T2D, 39 CON), while the other exhibited increased insulin resistance in the liver (HIRE, comprising 100 T1D, 95 T2D, 0 CON) when compared to the other two clusters. AIRE was associated with elevated visceral adiposity and HLC compared to WISE and HIRE, but was also linked to lower peripheral nerve and renal function at the 5-year follow-up. HIRE exhibited higher HLC than WISE, but lower HLC than AIRE and was more likely to receive insulin treatment during the 5-year follow-up period.

Conclusion: Irrespective of diabetes, measurement of tissue-specific insulin sensitivity allows to define clusters with different risks for diabetes-related comorbidities. This underscores the importance of considering tissue-specific insulin resistance in characterizing metabolic heterogeneity and predicting outcomes.

Clinical Trial Registration Number: NCT01055093

Disclosure: K.B. Bódis: Grants; German Federal Ministry of Health (BMG), Ministry of Culture and Science of the State North Rhine-Westphalia (MKW NRW), German Federal Ministry of Education and Research (BMBF).

210

The gut epithelium-specific mTORC1 activation improves glucose metabolism selectively in male mice

R. Kawahara ¹, Y. Fujita^1,2, T. Yanagimachi², S. Kusui¹, S. Ida¹, N. Ohashi¹, K. Murata¹, K. Morino^1,3, S. Kume¹;

¹Medicine, Shiga University of Medical Science, Otsu, Shiga, Japan, ²Endocrinology and Metabolism, Graduate School of Medicine, Hirosaki University, Hirosaki, Aomori, Japan, ³Medicine, Kagoshima University, Kagoshima, Japan.

Background and aims: Recently, the intestine has been appreciated for playing crucial roles in glucose homeostasis through various actions, including the absorption of nutrients and the secretion of gut hormones. mTORC1 functions as a sensor of extracellular nutrition and intracellular energy balance. mTORC1 is activated in the liver and the kidney in diabetes mellitus, contributing to the development of insulin resistance and diabetic kidney disease. Several reports showed that mTORC1 was involved in tumorigenesis and inflammation in the intestine. However, limited papers indicated that mTORC1 in the intestine might contribute to glucose homeostasis, and results were still controversial. Moreover, several studies have shown sex-based differences in the mTOR signaling pathway across different tissues. In this study, we generated mice (Tsc1-VKO) with the gut epithelium-specific loss of the Tsc1 gene, which exclusively suppresses mTORC1, developing the model of mTORC1 overactivation in the intestinal epithelium. To clarify the roles of intestinal mTORC1 in metabolism, we investigated body weight changes and glucose metabolism and compared them to the male and female Tsc1-flox mice.

Materials and methods: The gut epithelium-specific Tsc1 knockout mice (Tsc1-VKO) were generated by crossbreeding Tsc1-flox mice with Villin-Cre mice using the Cre/loxP system. First, to confirm mTORC1 overactivation in the gut epithelium, we collected jejunum and ileum from Tsc1-VKO and the control mice. Then, we evaluated the Phosphorylation of S6 expression, a downstream effector of mTORC1, by immunohistochemical analysis. Next, we monitored body weight changes and performed oral and intraperitoneal glucose tolerance tests (OGTT and ipGTT), measuring blood glucose and insulin levels in both male and female mice.

Results: Immunohistochemical analysis revealed that Phosphorylation of S6 was enhanced from the villi to the crypts in the jejunum and the ileum of Tsc1-VKO mice compared to the control mice. This indicated that the Tsc1 gene ablation in the intestinal epithelium drove mTORC1 overactivation. In male Tsc1-VKO mice, body weight gain was comparable to that in the control mice. However, Tsc1-VKO mice showed significantly lower glucose excursion (AUC; Tsc1-flox 22775±1559 vs. Tsc1-VKO 16430±1156 mg/dl*min, p<0.01) without significant differences in plasma insulin levels (AUC; Tsc1-flox 47.0±5.2 vs. Tsc1-VKO 57.3±9.8 ng/ml*min, p=0.55) during the OGTT. Tsc1-VKO mice showed significantly lower glucose excursion during the ipGTT. Still, unlike OGTT, the mice presented a significant elevation of plasma insulin levels with a steep peak at 15 minutes (Tsc1-flox 0.53±0.055 vs. Tsc1-VKO 1.31±0.27 ng/ml, p<0.01). Female Tsc1-VKO mice demonstrated that neither blood glucose nor plasma insulin levels were significantly different compared to the control mice during the OGTT and ipGTT. In contrast, female Tsc1-VKO mice gained body weight considerably more than that of the control mice.

Conclusion: Our findings suggest that mTORC1 overactivation in the intestinal epithelium may improve glucose metabolism in male mice but not affect female mice, and these effects are independent of weight gain.

Supported by: Grant-in-Aid for Scientific Research(C)

Disclosure: R. Kawahara: None.

OP 36 Inside information on beta cell regulation

211

Structure, interaction, and nervous connectivity of beta cell primary cilia

A. Mueller ¹, N. Klena², S. Pang³, L. Galicia Garcia¹, O. Topcheva¹, S. Aurrecoechea Duran¹, H. Mziaut¹, J. Weitz⁴, M. Distler⁴, T. Kurth⁵, D. Schmidt⁶, H.F. Hess⁷, C. Xu³, G. Pigino², M. Solimena¹;

¹Paul Langerhans Institute Dresden, Dresden, Germany, ²Human Technopole, Milan, Italy, ³Yale School of Medicine, New Haven, CT, USA, ⁴University Hospital Carl Gustav Carus, Dresden, Germany, ⁵Center for Molecular and Cellular Bioengineering (CMCB), Dresden, Germany, ⁶Helmholtz Imaging, Max Delbrück Center for Molecular Medicine (MDC), Berlin, Germany, ⁷Janelia Research Campus, Ashburn, VA, USA.

Background and aims: Primary cilia are sensory organelles present in many cell types. Based on an array of microtubules termed axoneme they form a specialized membrane compartment partaking in various signaling processes. Primary cilia of pancreatic islet beta cells play a role in autocrine and paracrine signaling and their dysfunction is linked to diabetes. Yet, the structural basis for their functions is unclear.

Materials and methods: We used volume electron microscopy to image mouse and human beta cell cilia within the islet context and performed three-dimensional reconstructions of their axonemes and neighboring cells and tissues. We furthermore performed expansion microscopy of mouse pancreas to obtain quantitative data on cilia length and connectivity.

Results: We present three-dimensional reconstructions of complete mouse and human beta cell cilia, revealing a disorganized 9+0 axoneme structure (Figure 1a and b). Within the islet, cilia are spatially confined within deep ciliary pockets or squeezed into narrow extracellular spaces between adjacent cells. Beta and alpha cell cilia physically interact with neighboring islet cells pushing and strongly bending their plasma membranes. Furthermore, beta cells can contain multiple cilia that can meet with other islet cell cilia in the extracellular space. Additionally, beta cell cilia establish connections with islet-projecting nerves (Figure 1c). Close to some of these connections we could observe synaptic vesicles (Figure 1c’) and invaginations of the axon’s plasma membrane indicative of synaptic vesicle release (Figure 1c”).

Conclusion: These findings highlight the pivotal role of beta cell primary cilia in islet cell connectivity, pointing at their potential functional role in integrating islet intrinsic and extrinsic signals. We show for the first time that beta cell cilia connect to the islet innervation pointing to a role that goes beyond paracrine signaling. These novel insights contribute to understanding their significance in health and diabetes.

Supported by: grant agreements no. 115881 (RHAPSODY) and no. 115797 (INNODIA)

Disclosure: A. Mueller: None.

212

Somatostatin secretion at synapse like structures triggers primary cilia signalling to modulate beta cell function

C. Incedal Nilsson ¹, O. Dumral², B. Xie³, A. Müller⁴, G. Sanchez¹, M. Solimena⁴, H. Ren³, O. Idevall Hagren¹;

¹Medical Cell Biology, Uppsala University, Uppsala, Sweden, ²Medical Biochemistry, Karolinska Institutet, Stockholm, Sweden, ³Center for Quantitative Biology, Peking University, Beijing, China, ⁴Molecular Diabetology, TU Dresden Faculty of Medicine, Dresden, Germany.

Background and aims: Primary cilia, despite being continuous with the plasma membrane and cytosol, act as cellular antennae for sensing and integrating signals. Enriched with GPCRs, they provide alternative pathways for signal processing alongside canonical receptor signalling. The hedgehog (Hh) pathway, vital for pancreas development and islet cell function, relies on functional cilia. Hh stimulation leads to ciliary, Gi-dependent cAMP lowering and GLI transcription factor activation. Somatostatin receptor type-3 (SSTR3) are also Gi-coupled and localize to cilia, but whether these receptors crosstalk with the Hh pathway is not known. The aim of this study was to elucidate the role of primary cilia in somatostatin (SST) action in islet β-cells.

Materials and methods: The spatial arrangement of β-cell cilia and SST-secreting δ-cells in mouse islets was determined using FIB-SEM and confocal microscopy. A cilia-targeted SST sensor and confocal microscopy was used to measure SST release onto cilia in intact mouse islets. TIRF microscopy of mouse islets and MIN6 pseudo-islets expressing cilia-targeted cAMP and Ca²⁺ sensors were used to record SST-induced signal transduction. Functional evaluation of ciliary somatostatin signalling was performed in MIN6 pseudo-islets following shRNA-mediated knockdown of SSTR3.

Results: FIB-SEM images from different axial positions of a mouse islet revealed islet δ-cells in direct contact with β-cell primary cilia. Similar observations were made in transgenic mouse islets with δ-cell-specific expression of td-tomato or wild-type islets immunostained against SST, with each δ-cell making contact with on average 3.2 cilia, compared to 1.8 cilia for non-δ-cells (n=35; N=7; P=1.0E-5). To test if SST is released onto primary cilia, we expressed a SST sensor (SST1.0) based on SSTR5. The sensor localized to both plasma membrane and cilia, and stimulation with 10 mM glucose or 100 nM ghrelin triggered robust increases in SST1.0 fluorescence that were most prominent in the primary cilia. This indicate that ciliary SSTRs can sense endogenously released SST. Using cAMP sensors, we next showed that 100 nM SST induced cAMP lowering in both the cytosol and cilia, but only the lowering in the cilium was prevented in MIN6 pseudo-islets with reduced SSTR3 expression. Prolonged exposure to SST (18h) cause a slight shortening of primary cilia (9±0.08%, n=214, N=4, P<0.001) and induced the translocation of GLI2 from the primary cilium to the nucleus. The magnitude of the translocation mirrored that induced by Hh agonist SAG and was prevented by SSTR3 knockdown (WT-control: 25%, WT-SST: 56%, WT-SAG: 52%, KD-control: 43%, KD-SST: 42%; N=3, n=250-300 cells). These results indicate functional crosstalk between SST and the Hh-pathway. Consistent with long-term effects of SST on β-cell function, we find that 18h SST or SAG exposure both lowers resting cAMP levels and strongly attenuate GLP-1-induced cAMP elevations in β-cells (resting cAMP: P_SAG< 0.0001, P_SST= 0.005; 10 nM GLP-1: P_SAG <0.001, P_SST< 0.0001; n=140; N=3).

Conclusion: Our findings highlight the primary cilium as an alternative target of SST action in islet cells and suggest roles of SST beyond acute inhibition of hormone release.

Supported by: Novo Nordisk Foundation NNF19OC0055275 and Swedish Research council 2019-01456

Disclosure: C. Incedal Nilsson: None.

213

Primary cilia regulate GLP-1 signalling in pancreatic beta cells

I. Melena, L. Zhu, J. Cho, S. Adamson, J. Hughes;

Endocrinology, Metabolism & Lipid Research, Washington University School of Medicine, Saint Louis, MO, USA.

Background and aims: Glucagon-like peptide 1 receptor agonists (GLP-1RAs) have garnered attention for their pleiotropic benefits on metabolism and have become the leading prescribed drugs for treatment of both diabetes and obesity. A major part of their therapeutic benefit derives from enhanced postprandial insulin secretion in pancreatic beta cells, but the signaling mechanism of GLP-1 in beta cells and its potential compartmentalization remain incompletely understood. In this study, we aimed to determine whether cilia loss in beta cells is sufficient to disrupt GLP-1-augmented insulin secretion via altered cAMP and Ca²⁺ signaling.

Materials and methods: Mice with cilia deletion in pancreatic beta cells (βCKO; Ins1^Cre/Cre x Ift88^fl/fl) and littermate control mice (WT; Ins1^Cre/Cre and Ift88^fl/fl) were used for islet functional imaging and secretion assays. Glucose- and GLP-1-stimulated cAMP dynamics were assessed by live-cell imaging using virally transduced EpacSH187 sensor in islets from βCKO and control mice. Quantitative analysis of glucose- and GLP-1-dependent Ca²⁺ flux was performed using a double transgenic mouse model carrying cilia deletion and a genetically encoded beta cell-specific calcium reporter (βCKO-GCaMP; Ins1^Cre/Cre x Ift88^fl/fl x GCaMP6f) and littermate control mice (WT GCaMP; Ins1^Cre/Cre x GCaMP6f). Functional consequences of cilia loss on GLP-1 signaling were assessed by static and dynamic glucose- and GLP-1-stimulated insulin secretion in intact islets collected from cilia KO and WT mice.

Results: Islets lacking beta cell primary cilia exhibited decreased whole-cell cAMP response to stimulation with glucose and the GLP-1RA liraglutide (AUC (mean±SEM): 14.8±0.4 in βCKO vs 19.7±0.4 in WT, p=0.0008). Whole-cell Ca²⁺ dynamics including both first-phase activation and second-phase oscillations were dampened in cilia KO GCaMP reporter islets in response to glucose and GLP-1RA (First peak duration (mean±SEM): 1769±110 vs 2679±41, p<0.0001, #oscillation peaks/10 min: 6±0.6 vs 12±0.9, p<0.0001). Insulin secretion in response to co-stimulation with glucose and GLP-1 was compromised in cilia KO islets by 50% in dynamic insulin secretion (AUC (mean±SEM): 226±7 vs 118±13, p=0.002).

Conclusion: Our findings reveal that the primary cilium regulates GLP-1R signaling in beta cells by acting as a permissive signal for GLP-1-dependent cytosolic cAMP and Ca²⁺ activation toward insulin secretion. This represents a previously unappreciated subcellular mechanism of incretin signaling which may be targeted by modulators of primary cilia function.

Supported by: NIH NIDDK R01DK138974

Disclosure: I. Melena: None.

214

GPR180 receptor controls insulin production and secretion

M. Antal ¹, P. Makovicky¹, T. Dahlby², C. Wolfrum², M. Balaz¹, L. Balazova¹;

¹Biomedical Research Center of the Slovak Academy of Sciences, Bratislava, Slovakia, ²Lab. of Translational Nutrition Biology, ETH Zurich, INFH, Schwerzenbach, Switzerland.

Background and aims: Our knowledge of the function of the ubiquitous GPR180 receptor is limited to the regulation of adipocyte thermogenesis or lipid accumulation in the liver. Since we previously reported disrupted glucose homeostasis in Gpr180 global knockout mice that could not be fully explained by adipose tissue dysfunction, we aimed to investigate the involvement of the GPR180 receptor in beta cell function given its critical role in the regulation of glucose metabolism.

Materials and methods: To investigate the function of GPR180 in pancreas physiology, we studied beta cell specific Gpr180 knockout mice utilizing Cre/LoxP system with constitutive Cre recombinase expressed under insulin promoter. In addition, we employed MIN6 cell line with beta cell characteristics to study the effect Gpr180 silencing on insulin secretion in vitro.

Results: Beta cell specific ablation of Gpr180 gene in Gpr180^fl/fl x Rip-Cre mice is associated with both impaired glucose tolerance (n = 5 for both genotypes; AUC p ˂ 0.05) and glucose stimulated insulin secretion (n = 8 for both genotypes; Gpr180^fl/fl x Rip-Cre vs Gpr180^fl/fl 5 min after exogenous glucose load p ˂ 0.01). We observed less Langerhans islets determined by staining of insulin positive cells in pancreas sections (n = 12 for both genotypes; p ˂ 0.05) as well as decreased Ins2 gene transcription in isolated islets (n = 8 for both genotypes; n = p ˂ 0.05) of mice lacking Gpr180 in beta cells. Gpr180 silencing in MIN6 cells (n = 6 for both groups; p ˂ 0.001) resulted into impaired glucose stimulated insulin secretion (n = 6 for both groups; 25 mM glucose non-targeting siRNA vs siGpr180 p ˂ 0.05) without significant alteration in insulin gene expression (n = 6 for both groups; p > 0.05). Our preliminary results indicate altered glycolytic rate based on decreased ATP production in response to high glucose in MIN6 cells with Gpr180 knockdown (n = 5 for both groups; p ˂ 0.01).

Conclusion: Dysfunctional GPR180 receptor in beta cells leads to development of prediabetes like phenotype in chow fed mice. This results from combination of altered Insulin gene transcription and lower pancreatic islets quantity. The latter might originate from altered prenatal islets formation given the character of experimental model (constitutive knockout driven by promoter active during embryogenesis). However, utilizing in vitro functional studies revealed additional defects in the mechanism of insulin secretion independently of its production when manipulating the Gpr180 gene expression in MIN6 cell line. Thus, GPR180 receptor is an important element in the regulation of pancreas physiology at multiple levels.

Supported by: SASPRO2 No.1260/02/02 and VEGA no. 2/0128/23

Disclosure: M. Antal: None.

215

The type 2 diabetes-associated K ⁺ channel TALK-2 reduces beta cell endoplasmic reticulum Ca ²⁺ stores and cytosolic Ca ²⁺ entry, which limit insulin secretion

J.R. Dobson, P.K. Dadi, A.Y. Nakhe, D.A. Jacobson;

Molecular Physiology and Biophysics, Vanderbilt University, Nashville, TN, USA.

Background and aims: KCNK17, which encodes TALK-2, is one of the most abundant β-cell K⁺ channel transcripts. Polymorphisms in or near KCNK17 are associated with an increased predisposition for developing type 2 diabetes. Progress on the physiological and pathophysiological functions of TALK-2 have been limited because the mouse genome does not contain the KCNK17 gene. Thus, the objective of this study was to illuminate the function(s) of human β-cell TALK-2.

Materials and methods: Immunofluorescent staining as well as TALK-2-GFP constructs in combination with organelle fluorescent reporters were utilized to determine the cellular localization of TALK-2. TALK-2 function was assessed with cellular Ca²⁺ imaging of preparations including: 1. Inducible TALK-2 cells, 2. Adenoviral KCNK17 shRNA or Scramble shRNA expression in dispersed human islets, and 3. β-cell selective TALK-2 knockdown (KD) or control human pseudoislets. Insulin secretion was measured from human pseudoislets with perifusion and ELISA.

Results: TALK-2 protein displayed both plasma membrane and ER localization in β-cells within human pancreatic sections. Additionally, TALK-2-GFP colocalized with an ER-RFP marker and a plasmalemma dye in HEK293 cells. TALK-2 expression in HEK293 cells accelerated ER Ca²⁺ release (by 62±13%; p<0.05; n=3), reduced ER Ca²⁺ storage (by 67±12%; p<0.05; n=3) and increased basal cytosolic Ca²⁺ (by 11±0.5%; p<0.05; n=3). In human β-cells TALK-2 KD increased ER Ca²⁺ storage monitored as ER Ca²⁺ release following SERCA inhibition (by 16±5%; p<0.05; n=7). Moreover, TALK-2 KD increased β-cell Ca²⁺ influx at 11mM glucose in dispersed human islets (by 20±6%; p<0.05; n=7) and in pseudoislet preps (by 208±64%; p<0.01; n=4). Pseudoislets with β-cell selective TALK-2 KD displayed reduced basal insulin secretion at 1mM (by 51±6.7%; p<0.01; n=4) and 5.6mM glucose (by 60±4.7%; p<0.01; n=4). β-cell selective TALK-2 KD pseudoislets also showed enhanced glucose-stimulated insulin secretion (with 11 mM glucose, by 200±15%; p<0.05; n=4). Finally, KCl-mediated depolarization stimulated increased insulin secretion in the β-cell TALK-2 KD pseudoislets compared to controls (by 245±50%; p<0.05; n=4).

Conclusion: These data support that ER and plasmalemma localized TALK-2 channels function to reduce β-cell ER Ca²⁺ stores and limit glucose-stimulated Ca²⁺ influx respectively. Furthermore, TALK-2-mediated augmentation of ER Ca²⁺ leak likely enhanced basal insulin secretion by increasing cytosolic Ca²⁺. Additionally, TALK-2 hyperpolarization of the plasma membrane may reduce glucose-stimulated insulin secretion by limiting voltage-dependent Ca²⁺ channel activation. Therefore, polymorphisms in KCNK17 that increase TALK-2 activity or expression would be predicted to diminish β-cell ER Ca²⁺ stores, impair glucose-stimulated Ca²⁺ influx, limit glucose-stimulated insulin secretion, elevate basal insulin secretion, and thus enhance insulin resistance.

Supported by: DK129340-S1, DK129340, DK136768

Disclosure: J.R. Dobson: None.

216

Mitochondrial fission process 1 regulates glucose stimulated insulin secretion in pancreatic beta cells

S. Sarwat ¹, C. Zhang¹, R.M.A. Alonso¹, A. Mihalovits¹, M. Yang¹, G. Rutter², T.A. Rodriguez¹, A. Martinez-Sanchez¹;

¹Imperial College London, London, UK, ²Cell Biology & Functional Genomics, Imperial College London, London, UK.

Background and aims: Mitochondrial function is essential for integrating glucose metabolism and insulin secretion in pancreatic β-cells. Mitochondrial dysfunction disrupts β-cell secretory function and/or survival and may contribute to type 2 diabetes (T2D) progression. Mitochondrial fission process 1 (MTFP1) is an inner mitochondrial membrane protein that regulates mitochondrial fission and/or function in a cell-specific manner. In β-cells, MTFP1 is directly repressed by miR-125b, which plays a negative role in insulin secretion through multiple aspects of β-cell biology, including oxygen consumption, mitochondrial morphology and insulin production and secretion. Here, we seek to study the role of MTFP1 in β-cell function including glucose stimulated insulin secretion (GSIS) and mitochondrial morphology.

Materials and methods: We generated and used lentivirus expressing Cas9 under the control of a rat insulin promoter and gRNAs targeting MTFP1 and lentivirus expressing shRNAs targeting MTFP1 for knockout/knockdown of MTFP1 in EndoC-βH3 cells and human islets, respectively. MTFP1 gain-of-function was achieved with an adenoviral vector expressing human MTFP1. GSIS, mitochondrial morphology and DNA content and DRP1 (dynamin-related protein 1) phosphorylation were assessed in these cells using HTRF, mitotracker green staining, qPCR and western blot.

Results: Overexpression of MTFP1 in both EndoC-βH3 cells and human islets resulted in significant increases in GSIS (1.2-fold, p = 0.0055 and 1.5-fold, p = 0.0205, respectively). Conversely, CRISPR/Cas9-mediated elimination of MTFP1 in EndoC-βH3 cells and shRNA-mediated MTFP1 knockdown in human islets strongly decreased GSIS (2.5-fold, p = 0.0031; and n = 2; 1.4-fold, p = 0.0782, respectively). These secretory defects occurred in the presence of significant changes in mitochondrial morphology, as overexpression of MTFP1 in EndoC-βH3 cells decreased mitochondrial elongation (1.3-fold, p = 0.0006) while knockout of MTFP1 had the opposite effect (1.5-fold, p = 0.0064). Preliminary data suggest that phosphorylation of the important fission regulator DRP1 and mitochondrial DNA content were also altered.

Conclusion: Our data reveal a novel, critical role for MTFP1 in β-cells insulin secretion. Ongoing work will provide important additional insights into the mechanism of action of this protein in β-cells and its role in glucose homeostasis.

Supported by: DUK: 321/0006358, MRC: MR/X009912/1

Disclosure: S. Sarwat: None.

OP 37 Protection in islet transplantation

217

Islet transplantation vs insulin alone in patients with type 1 diabetes and a kidney transplant: a French nationwide study on behalf of the Trepid group

M. Maanaoui ¹, R. Lenain¹, Y. Foucher², F. Buron³, S. Lablanche⁴, G. Blancho⁵, S. Caillard⁶, L. Kessler⁷, A. Brodin-Sartorius⁸, M. Chetboun⁹, J. Kerr-Conte¹⁰, T. Berney¹¹, M.-C. Vantyghem¹², M. Hazzan¹, F. Pattou⁹;

¹Department of Nephrology, CHU Lille, Lille, France, ²Centre d’Investigation Clinique, CHU Poitiers, Poitiers, France, ³Department of Nephrology, CHU Lyon, Lyon, France, ⁴Department of Endocrinology, CHU Grenoble, Grenoble, France, ⁵Department of Nephrology, CHU Nantes, Nantes, France, ⁶Department of Nephrology, CHU Strasbourg, Strasbourg, France, ⁷CHRU de Strasbourg, Strasbourg, France, ⁸Department of Nephrology, Kremlin-Bicêtre Hospital, APHP, Paris, France, ⁹Department of Endocrine Surgery, CHU Lille, Lille, France, ¹⁰Department of Biotherapy, CHU Lille, Lille, France, ¹¹Islet of Langerhans Transplantation Program, Div of Transplantation, Nephrology and Clin. Immunology, University of Lyon Medical Center, Lyon, France, ¹²Endocrinology and Metabolism, Lille University Hospital, Lille, France.

Background and aims: Islet transplantation is associated with a benefit on glycaemic control compared to optimized insulin therapy in recent clinical trials. However, there is a lack of evidence concerning the long-term impact of islet transplantation on type 1 diabetic kidney transplant recipients’ prognosis.

Materials and methods: Every type 1 diabetic recipient transplanted with a kidney in France between 2000 and 2017 was included. Patients transplanted with pancreatic islets were compared to controls treated with insulin alone according to a matching method based on time-dependent propensity scores (using the following variables : year of transplantation, donor age, and recipient age, serum creatinine, HBA1c, BMI, cardiovascular background) which allow to ensure patients comparability at the time of islet transplantation. The primary outcome was graft failure, defined by death or return to dialysis.

Results: Among 2393 type 1 diabetic patients transplanted with a kidney during the study period, 327 were eligible to islet transplantation, including 47 that were actually transplanted with islets. Median time for islet transplantation was 34.8 months [21.8-48.4]. Probabilities of insulino-independence and islet graft survival at 1, 5 and 10 years were respectively 63.8% [51.5-79.2], 46.3% [33.9-63.2], 38.7% [25.9-57.8] and 89.4% [81.0-98.6], 87.2% [78.2-97.3], 78.2% [66.2-92.4]. Forty IAK were successfully matched to 80 KA. IAK was associated with a better metabolic control over time compared to KA, with a significant reduction of HBA1c up to five years (IAK: 6.50% (0.91) vs KA: 7.94% (1.73), p < 0.01). We found a significant benefit of islet transplantation compared to kidney alone on patient-graft survival, with a HR of 0.44 (95%CI: 0.23-0.88), mainly explained by a protective effect on the risk of death (HR: 0.41 (95%CI: 0.13-0.91)). This resulted in a significant increase of life-expectancy with a functioning kidney graft of +16.9 months (95% CI: +1.2 ; +33.5) and of the overall life-expectancy of +16.7 months (95% CI: +0.9 ; +35.4) after 14 years of follow-up.

Conclusion: We observe a significant benefit of islet transplantation on the risk of graft failure and death in type 1 diabetic kidney transplant recipients. These results provide incentives to promote islet transplantation in this population.

Disclosure: M. Maanaoui: None.

218

Effect of islet transplantation on the incidence of diabetic complications and mortality in patients with unstable type 1 diabetes

Q. Perrier ¹, C. Jambon-Barbara¹, L. Kessler², O. Villard³, F. Burron⁴, F. Pattou⁵, B. Guerci⁶, M. Roustit¹, E. Berishvili⁷, L. Rakotoarisoa², M.-C. Vantyghem⁵, E. Renard³, T. Berney⁴, P.-Y. Benhamou¹, S. Lablanche¹;

¹University Hospital, Grenoble, France, ²University Hospital, Strasbourg, France, ³University Hospital, Montpellier, France, ⁴University Hospital, Lyon, France, ⁵University Hospital, Lille, France, ⁶University Hospital, Nancy, France, ⁷Islet Isolation Centre, Geneva, Switzerland.

Background and aims: Islet transplantation alone (ITA) has become the standard therapeutic approach for unstable type 1 diabetes (T1D). Lot of data are available on metabolic control after ITA, but few on its impact on long-term diabetic complications and long-term survival. We aim to investigate the impact of ITA on delaying clinically significant events.

Materials and methods: This retrospective study utilized data from ITA recipients transplanted in France between 1999 and 2013. The control group consisted of a synthetic cohort identified from the French Health Insurance claims database (SNDS). The SNDS provides administrative and reimbursement data but lacks metabolic information. Consequently, this group comprised T1D patients treated solely with insulin, aged over 18 years, without prior kidney transplantation, and with more than two hospitalizations for diabetes-related complications in the year preceding inclusion. Each patient in the experimental group was matched using greedy propensity score (standardized mean difference (SMD) was used to assess matching efficacy) matching at a ratio of up to 1:10 with regards to covariates including sex, age, history of dialysis, amputation, stroke, transient ischemic attack, and myocardial infarction. The primary outcome was defined as the first occurrence of a composite criterion comprising death, dialysis, amputation, non-fatal stroke, non-fatal myocardial infarction, or transient ischemic attack. The secondary outcome was the occurrence of cancer. Survival analysis (log-rank test) was employed to compare event occurrence between groups.

Results: 61 patients with ITA were matched with 610 T1D but non-ITA patients (median follow-up at 10.4 years), achieving SMD < 10% (except for stroke -11.4) indicating negligible differences in potential confounders and balanced matched cohorts. The primary composite outcome occurred less frequently in patient with ITA compared to those without ITA (HR=0.39 [0.21-0.71], p=0.001), with significant differences observed only for the mortality sub-criterion (HR=0.22 [0.09-0.54], p=0.0003, table 1). Cancer occurrence did not differ between groups (HR=1.10 [0.59-2.06], p=0.8, table 1).

Conclusion: This study suggests long-term benefits of islet transplantation on mortality and diabetes complications. Furthermore, despite the use of immunosuppressive drugs following islet transplantation, we observed no significant increase in the risk of cancer. Altogether, these findings suggest a favourable risk-benefit ratio of islet transplantation in managing patients with unstable T1D.

Supported by: AOR Greffe-Agence de la Biomedecine, AO interne CHU Grenoble Alpes

Disclosure: Q. Perrier: None.

219

Ado12, a non-fibrotic immunoprotective hydrogel containing human islets shows efficient and sustained in vivo functionality for clinical use

A.-L. Gaffuri ¹, X. Gaume², J. Brun³, R. Besnard², C. Gautier², O. Jouannot¹, N. Laurent⁴, R. Eloy², K. Bouzakri⁵, O. Soula²;

¹Adocia, Lyon, France, ²³France, ⁴USA, ⁵UMR DIATHEC - EA 7294, University of Strasbourg, Strasbourg, France.

Background and aims: Ado12 is an easily implantable and fully retrievable scaffold for islet transplantation aiming to cure diabetes by cell therapy without requiring immunosuppression. It is based on a permselective hydrogel film allowing insulin diffusion while preventing immune cell invasion and antibody contact. Allogenic transplant of rat islets in immunocompetent diabetic rats has already demonstrated encapsulated islets survival, insulin secretion, and glycemic regulation. Compatibility with human islets, good local tolerance, and scalability to deliver a therapeutic dose were studied to meet requirements for a first in human.

Materials and methods: Device - Ado12 is a soft biomaterial with 95% water, synthetized from non-degradable polymers cross-linked by bio-orthogonal chemistry. It is reinforced with a surgical mesh embedded in the hydrogel and sandwiched in a suturable silicon ring. In vitro encapsulation - Human islets were encapsulated in Ado12. Their functionality was evaluated using perifusion up to 7 months in vitro. In vivo - In two independent studies, human islets were encapsulated in Ado12 in 1.4cm-diameter disks. NXG mice were implanted with 2 implants in the peritoneum for at least two months for a total dose of 2.2-3.1 kIEQ/mouse. In vivo functionality was evaluated by quantifying Human C-peptide secretion. They were explanted after 2 months. Explant viability and functionality were analyzed in vitro and by histology. Domestic pig was implanted by laparoscopy with Ado12 for a proof of concept of surgery.

Results: In vitro, Ado12 Human Islets has shown gluco-responsive insulin secretion up to 4 months. Ado12 tolerance was evaluated in the intraperitoneal site of immunocompetent rodent species, with or without human islets: Ado12 was well tolerated and did not trigger inflammatory reaction or fibrosis. In vivo functionality of Ado12 Human Islets was evaluated in immunodeficient mice. Human C-peptide secretion increased during the first two weeks to reach an average of 700 ± 164 pM/day and was maintained in 100% of treated mice until explantation (n=6 mice). In vitro secretion indexes of the explants (n=6 explants analyzed at D85 post implantation) were also maintained, with 65% ± 25 insulin secretion levels compared to preimplantation values. These results were confirmed in one additional study (n=4 mice). Therefore, Ado12 favors human islets in vivo functionality. A human sized Ado12 with the potential to deliver a therapeutic islet dose was successfully manufactured and implanted by laparoscopy in domestic pig.

Conclusion: Ado12 Human Islets have shown sustained survival and functionality of human islets in vitro for over 4 months. This long-term survival has been confirmed in vivo in immunodeficient mice, with a stable Insulin or C-peptide secretion over 3 months, absence of fibrotic encapsulation and survival of functional islets.

Disclosure: A. Gaffuri: Employment/Consultancy; Adocia. Stock/Shareholding; Adocia.

220

Functional islet transplantation into immunocompetent diabetic animals using hydrogel encapsulation without the need for immunosuppression

H. Stöver;

Allarta Life Science Inc., Dundas, ON, Canada.

Background and aims: Type 1 diabetes (T1D) is an autoimmune disorder where the body’s immune system mistakenly attacks insulin-producing beta cells in the pancreas, leading to a chronic deficiency in insulin and a lack of blood glucose control. Islet cell transplantation promises to replace lost cells and cure T1D; however, donor islet cell transplantation requires systemic immune suppression, whose side effects remain a concern. Biomaterials can provide physical immune protection and have the potential to enable allogeneic islet implantation as a potential cure. We have developed non-immunogenic, chemically crosslinked, synthetic hydrogels that are robust and shape-agnostic. Encapsulated rat and human donor islets remained viable and functional in vitro and in vivo.

Materials and methods: Micropipette aspiration, chemical stress tests (e.g., citrate), and in vivo persistence demonstrated the hydrogels’ mechanical robustness. Fluorescently labeled dextran (10, 70, 250, and 500 kDa), IgG (150 kDa), and insulin (5.8 kDa) showed diffusion limits and protein release profiles. Different polymer modifications were tested for toxicity to human induced pluripotent stem cells (hiPSCs) and for fibrosis in healthy mice and pigs. Glucose-stimulated insulin secretion (GSIS) showed rat and human donor islet functionality after encapsulation at different densities. Islet loads were tested for fibrosis and functionality (e.g., HbA1c, C-peptide, and blood glucose) following implantation into streptozotocin (STZ)-induced diabetic immunocompetent mice, rats, Yorkshire pigs, and Göttingen minipigs.

Results: The hydrogels exclude IgGs and larger molecules while allowing rapid insulin out-diffusion as low as ~ 1 min t_1/2. High cell viability and GSIS indices (avg. of 12x increase after stimulation) from rat and human islets demonstrate that the hydrogels maintained islet function in vitro. Capsules were retrieved intact after ~300 days in vivo. Encapsulated rat islets enabled rapid and persistent blood glucose control (<250 mg/dL) for 140+ days in immunocompetent STZ-diabetic C57BL/6 mice and 180+ days in immunocompetent rats. We noticed a dose-dependent effect on in vivo function despite similar in vitro GSIS indices. Viable human donor islets showed controlled blood glucose and increased C-peptide readings (up to 45 pmol/L) for 45 days in mice (~8x longer than reported for free human islets) and in pigs. When explanted, the encapsulated islets showed good functionality and viability, and the hydrogels had no-to-minimal tissue response while retaining integrity.

Conclusion: Our novel, proprietary, and retrievable hydrogels demonstrate immune protection of allogeneic and xenogenic islet transplants into large and small immuno-competent animals without immunosuppression. Currently, we are scaling to therapeutic doses in large animals in pursuit of a curative human allogeneic therapy for T1D without immunosuppression. Figure: The technology is shape-agnostic with centred placement of viable islets in high surface devices.

Supported by: JDRF: 2-IND-2023-1449-I-X

Disclosure: H. Stöver: Employment/Consultancy; CEO/CSO of Allarta Life Science Inc. Grants; JDRF. Stock/Shareholding; Founder and shareholder of Allarta Life Science Inc.

OP 38 New insights into CNS regulation of energy and glucose metabolism

221

Effects of activation of peripheral sympathetic nerve on glycaemic control in standard chow- and high-fat-fed rats

D. Sato ¹, L. Miyamoto², R. Banno^3,4, M. Kusunoki³;

¹Department of Biochemical Engineering, Yamagata University, Yonezawa, Japan, ²Department of Nutrition and Life Science, Kanagawa Institute of Technology, Atsugi, Japan, ³Research Center of Health, Physical Fitness and Sports, Nagoya University, Nagoya, Japan, ⁴Department of Endocrinology and Diabetes, Nagoya University Hospital, Nagoya, Japan.

Background and aims: Sympathetic nerve activity controls not only blood pressure but peripheral glucose uptake. We previously reported that electrical microstimulation (MS) of peripheral sympathetic nerve reduced blood glucose (BG) level independently of insulin action within 30 s in rats. However, this reduction was only temporal, and the mechanisms of glycemic control during the sympathetic stimulation have not been clarified yet. In the present study, we evaluated the effects of peripheral sympathetic activation on glycemic control.

Materials and methods: Male Wistar rats were divided into two groups: standard laboratory chow- (SLC) and high-fat-fed (HFF) groups at 6 weeks of age. Six weeks after, we detected peripheral sympathetic nervous signal in the unilateral sciatic nerve in both groups with a microelectrode under anesthetic condition. The MS was conducted via the electrode, and glucose uptake was assessed with euglycemic clamp. To evaluate effects of the MS on glucose metabolism and lipid utilization in peripheral tissue, we conducted MS with the same manner in both groups for 60 min, and measured BG and plasma levels of insulin and irisin. After the MS, we compared mRNA expression in the skeletal muscle and liver, glycogen content in the liver, and triglyceride (TG) content in the white adipose tissue (WAT) with non-stimulated, control groups.

Results: The euglycemic clamp revealed that the MS significantly elevated glucose infusion rate (GIR) in both groups (P<0.01), and this effect remained even after the MS (P<0.01). However, the GIR was significantly lower in the HFF group than in SLC group throughout the MS (P<0.01). As a result of the MS for 60 min, we observed no significant change in BG in the SLC group but slight increase in HFF group (P<0.05). While the MS did not affect change in plasma insulin level in both groups, irisin levels significantly increased only in the SLC group (P<0.05). Regarding mRNA expression in the skeletal muscle, the MS resulted in significant elevation of Pgc-1α expression in the soleus and extensor digitorum longus in the SLC group (P<0.05) while such difference in this expression was not detected in HFF group. In the liver in SLC group, the MS significantly enhanced G6Pase expression (P<0.05), whereas the MS did not affect such gene expression in the HFF group. The MS also significantly increased hepatic glycogen content in SLC group (P<0.05) while no significant difference was found in the HFF group. In addition, the MS significantly lowered TG content in the WAT in SLC group (P<0.05) although no significant difference was observed in the HFF group.

Conclusion: Activation of peripheral sympathetic nerve increased glucose uptake even under insulin resistance. In the standard chow-fed rats, BG was maintained by hepatic glucose production although elevation of glucose uptake with the MS. Also, the MS induced Pgc-1α expression in the skeletal muscle and led to irisin secretion. Thus, the increase in hepatic glycogen content may be attributed to the irisin. Since the MS reduced TG in the WAT, glycerol derived from TG might be supplied to the liver as a substrate of gluconeogenesis. In contrast, high-fat feeding suppressed these effects on glycemic control.

Supported by: Japan Society for the Promotion of Science

Disclosure: D. Sato: None.

222

Genetic dissection of serum pro-neurotensin implicates its causal effects on brain structure

J. Breitfeld ¹, K. Horn², V. Witte³, F. Beyer³, A. Velluva⁴, J. Pott², D. LeDuc⁴, R. Baber⁵, M. Stumvoll¹, O. Melander⁶, P. Kovacs¹, M. Scholz², A. Tönjes¹;

¹Department of Endocrinology, Nephrology and Rheumatology, University of Leipzig, Leipzig, Germany, ²University of Leipzig, Leipzig, Germany, ³Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany, ⁴Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany, ⁵Institute of Laboratory Medicine, University of Leipzig, Leipzig, Germany, ⁶SUS Malmo, Lund University, Malmö, Sweden.

Background and aims: Neurotensin (NT) is a 13-amino acid peptide primarily expressed in the central nervous system (CNS) and the small intestine. Tissue-specific processing may lead to different forms of NT. In the CNS, NT acts as neurotransmitter regulating various processes like body temperature, nociception, pituitary hormone secretion, and dopaminergic transmission. In response to fat intake, NT facilitates fatty acid absorption based on food lipid content through the gut. Peripherally, NT promotes fat absorption and weight gain, whereas central NT signalling supresses feeding and favours weight loss. Studies in NT-deficient mice showed protection from obesity, hepatic steatosis and insulin resistance associated with high fat consumption. Human studies showed that obese and insulin-resistant humans have elevated pro-NT levels, which could serve as an indicator of obesity risk later in life.

Materials and methods: Meta-analysis of genome-wide association studies (GWAS) for serum pro-NT from four independent cohorts (LIFE-Adult and Sorbs from Germany, Malmö Diet Cancer (MDC) and Malmö Offspring Study (MOS) from Sweden; N=10,096) was conducted. Potential functional variants of the NT locus were included in Mendelian Randomization (MR) analyses to assess possible causative chains between NT and correlated metabolically relevant traits. Further, MRI-based outcomes of the reward network using connectivity strengths (CS) and cluster coefficients (CC) of fractional anisotropy (FA) and number of streamlines (NOS) of white matter connections between major hubs of the reward network (N=1090; LIFE-Adult) were analyzed.

Results: Meta-GWAS followed by secondary analyses revealed three loci with significant associations: i) one on chromosome (chr) 4 - rs6822751; ii) on chr 11 - rs41392245 and iii) on chr 12 - rs2723889. MR analyses using the three variants suggested a causal effect of pro-NT on pallidum brain structure (P=7.4x10^-5 in the penalized and penalized-robust MR-Egger). Moreover, there was a significant causal effect on brainstem, in the robust and penalized-robust IVW methods (P=0.009). In addition, in a subsample with MRI measurements (N=1090), BMI was not significantly different between genotype groups after correcting for age and sex, for all three variants (full - null model comparison, all F<2.5, all P>0.08). Except for NOS CC (P=0.016), simple linear regressions adjusted for age and sex confirmed that a higher BMI was related to less reward network coherence, i.e. to lower FA CS, NOS CS, and FA CC (all t<-4.2, all P<3.4x10^-5).

Conclusion: The present genetic study not only revealed genetic variation associated with serum pro-NT, but also highlights causal effect of the pro-NT on human brain structure.

Supported by: DFG, German Research Foundation – Projektnummer 209933838 – SFB 1052

Disclosure: J. Breitfeld: None.

223

Characterising central cell types accessed by peripherally-administered incretin receptor analogues

A. Roberts ¹, C. Blouet², J. Broichhagen³, D. Hodson⁴, A. Adriaenssens⁵;

¹Department of Neuroscience, Physiology and Pharmacology, University College London, London, UK, ²Institute of Metabolic Science & MRC Metabolic Diseases Unit, University of Cambridge, Cambridge, UK, ³LMU Munich, Munich, Germany, ⁴Oxford Centre for Diabetes, Endocrinology and Metabolism (OCDEM), University of Oxford, Oxford, UK, ⁵Department of Neuroscience, Physiology, and Pharmacology, University College London, London, UK.

Background and aims: Incretin-based therapeutics have revolutionized the treatment of diabetes and obesity. The advent of dual incretin agonists, targeting both the glucose-dependent insulinotropic polypeptide receptor (GIPR) and glucagon-like peptide-1 receptor (GLP1R), represents a key advance for incretin-based pharmacology. Expression of Gipr and Glp1r in the central nervous system is critical for incretin-induced weight loss. However, the mechanisms by which incretin-based pharmacology works centrally to reduce body weight are not fully understood. Specifically, which key subpopulations of Gipr- and Glp1r-expressing cells within the brain are accessed by peripherally-administered incretin analogues is poorly characterized, thereby representing an important hurdle for elucidating the central actions of this drug class. Circumventricular organs (CVOs) are regions within the brain that are highly vascularized with a modified blood brain barrier (BBB), and have privileged access to the circulation. This accessibility confers a pivotal function for CVOs in facilitating neuroendocrine communication between the peripheral circulation and the brain. Previous work has shown that peripherally-administered GIP or GLP-1 peptide analogues localize to the median eminence (ME) and area postrema (AP)—two CVOs located in the midbrain and hindbrain, respectively. The aim of this study was to define the molecular identity of cell types located within these CVOs that bind peripherally-dosed incretin analogues.

Materials and methods: Mice were dosed peripherally with fluorescently-labelled incretin analogues and transcardially perfused prior to tissue collection. 30 μm thick slices of brain tissue were stained for cell markers including vimentin, sox10, myelin basic protein and vasopressin. Z-stacks of stained tissue were captured using a Leica SP8 confocal microscope at 63x magnification. 3D reconstructions of peptide-labelled or antibody-stained cells were rendered using the Surfaces function in IMARIS 10.1.1.

Results: We identified multiple cell types in the ME that have access to the GIPR agonist, sGIP648, and the GLP-1R antagonist, Luxendin647. sGIP648 signal in the ME strongly colocalised with immunoreactivity for vimentin, a cell marker for tanycytes and vascular cells, and myelin basic protein. sGIP648 signal was also found in close apposition to sox10, a marker for oligodendrocyte cell bodies. Luxendin647 signal in the ME strongly colocalized with myelin basic protein and vasopressin-expressing fibers, specifically in the dorsal ME.

Conclusion: We show that neuronal and non-neuronal cells alike bind incretin-based peptides in CVOs. Within the ME, GIP-based peptides principally localize to non-neuronal cells, including tanycytes and oligodendrocytes—both of which have been implicated in the control of energy balance. GLP-1-based peptides exhibit greater localization with AVP neuron fibres. Future work is needed to define a role for these cell types in mediating the effects of incretin-based pharmacology on the regulation of body weight and food intake.

Supported by: EFSD/Novo Nordisk Foundation Future Leaders Award Programme

Disclosure: A. Roberts: None.

224

fNIRS demonstrates over activation of the dorsolateral prefrontal cortex following OGTT in youths with obesity and reduced insulin sensitivity

M. Manco ¹, G. Scozia¹, A. Aureli¹, D. Menghini¹, D. Fintini¹, S. Lasaponara²;

¹Bambino Gesu Hospital, Rome, Italy, ²University La Sapienza, Rome, Italy.

Background and aims: The prefrontal cortex (PFC), particularly the dorsolateral prefrontal cortex (DLPFC), is important for maintaining focus and attention while controlling food impulses and inhibiting impulsive eating behaviors. The present study aimed to investigate PFC activity at fasting and following an oral glucose tolerance test (OGTT) in response to cognitive tasks and in relation to peripheral insulin sensitivity (ISI) in youths with obesity.

Materials and methods: Cross-sectional study of 28 patients (age 7-15 y). ISI quantified by Matsuda Index; and PFC activity as oxygenated, deoxygenated, and total hemoglobin (HbO, HbR, HbT) concentration using near-infrared spectroscopy (fNIRS, Artinis Brite 24). Brain activity was estimated in fasting state (F) and during the first 60 min of the OGTT; in resting (R) condition and during the series of cognitive tasks (Ts, Verbal Fluency, Digit Span, Non-Verbal Stroop) administered before (F-Ts), and 15 minutes after the glucose load (OGTT-Ts). The 15-minute resting state of the post-load condition (R-OGTT) was divided into five-time windows (three minutes each), then the three-minute time window was cut and averaged to obtain 5-time segments of one minute each (P1-P2-P3-P4-P5).

Results: Significant differences in the PFC activity were found between fasting and absorptive states, both at rest and during tasks, in both hemispheres, particularly in the DLPFCs. During resting, over-activation was observed bilaterally in DLPFC and right frontal pole, with the highest activation after OGTT compared to fasting. Analyzing the 5-time segments of one minute each, significant differences (p<0.005 FDR corrected) were found at P2 [P2_S3D1 (HbO) R-F: 1,47028E-06 vs. R-OGTT: 4,1364E-06 mmol/l], P3 [S8D5 (tHb) R-F: 1,15305E-06 vs. R-OGTT:3,36876E-06 mmol/] and P5 [S3D1 (HbO) R-F: 1,47028E-06 vs. R-OGTT: 4,34665E-06 mmol/l, and S8D5 (tHb) R-F: 1,15305E-06 vs. R-OGTT: 3,07201E-06 mmol/l]. There was a significant negative correlation between activity registered at the channel of the left DLPFC (S3D1) and ISI (r -0.61, p < 0.001 ). In addition, the hemodynamic response function was predicted by ISI (F(1,26) = 15.351, p = 0.001, r² = 0.38). Differences in activity were also found during tasks between fasting and OGTT, but they were not correlated with participants’ ISI. At the behavioral level, enhanced cognitive performance in the absorptive vs. fasting state was observed exclusively in the Stroop Congruent condition (F: 29 ± 7.3 vs. OGTT: 34 ± 10, p < 0.05).

Conclusion: fNIRS demonstrated overactivation of DLPFC following a standard OGTT in youths with obesity that was inversely associated with peripheral ISI: the lower the ISI, the higher the difference between fasting and glucose load activation. Whether this association signifies a protective mechanism to reduce the burden of peripheral insulin resistance in the growing age child needs clarification.

Supported by: Finanziato dall’UE– Next Generation EU – PNRR M6C2 - Investimento 2.1 Valorizzazione e potenziamento della ricerca biomedica del SSN PNRR-MAD-2022-12

Disclosure: M. Manco: None.

OP 39 Novel insights in monogenic diabetes

225

MODY is polygenic as well as monogenic! Polygenic background strongly contributes to genetically confirmed MODY and clinical cases without a monogenic cause

J. Murray Leech ¹, A.M. Arni¹, K. Chundru¹, R.N. Beaumont¹, K. Colclough², A.R. Wood¹, A.T. Hattersley¹, M.N. Weedon¹, K.A. Patel¹;

¹Department of Clinical and Biomedical Sciences, University of Exeter, Exeter, UK, ²Exeter Genomics Laboratory, Royal Devon University Healthcare NHS Foundation, Exeter, UK.

Background and aims: The influence of common genetic variants on maturity-onset diabetes of the young (MODY) has yet to be determined. It is also not known if individuals thought to have MODY, where no monogenic cause is found, have an increased polygenic risk. We have conducted the largest study to date with the aim of quantifying and identifying the pathways through which common genetic variants contribute to MODY.

Materials and methods: We performed a genome-wide association study (GWAS) in 982 genetically confirmed beta cell MODY probands (age diagnosis: 20.9 y, BMI: 24.9 kg/m², HNF1A-600, HNF4A-196, HNF1B-128, KCNJ11-11, ABCC8-32, RFX6-15) against 7,800 ancestry-matched controls. We replicated our findings in 396 independent probands, 813 controls. We used 4,955 type 2 diabetes (T2D) cases, 298 type 1 diabetes (T1D) cases as negative controls. We constructed polygenic risk scores (PRS) for MODY, T2D, T1D, fasting glucose, and ten other diabetes-related traits. We also examined 319 patients (age diagnosis: 21.6y, BMI: 25.1 kg/m²) referred with possible MODY where no monogenic cause was found on full genetic testing.

Results: The significant role of common genetic variants was shown by a genome-wide SNP-based heritability of 30.37% for MODY (95% CI: 24.67-36.07%, p<0.0001). The MODY PRS derived from the GWAS was 2.93 standard deviations higher in 386 independent MODY probands than independent controls (p<0.0001), replicating the contribution of common variants. The T2D PRS in MODY probands was higher than controls but lower than T2D cases (mean: 0.17 vs -0.32 (controls) vs 0.43 (T2D), p<0.0001). A similar pattern was seen for insulin secretion, fasting glucose and beta cell function PRS’s (all P<0.0001), whereas MODY cases had similar BMI, fasting insulin, lipodystrophy and T1D PRS’s to controls (p>0.05). One standard deviation increase in the T2D PRS was associated with 0.75 years earlier onset of diabetes in genetically confirmed MODY. The T2D PRS was greatly increased in clinical MODY cases without a known monogenic (mean: 0.99, p<0.0001). They also had higher fasting glucose, waist-hip ratio, BMI, fasting insulin, lipodystrophy and MODY PRS’s compared to controls (mean difference 0.21-1.33, all P<0.001).

Conclusion: A substantial proportion of the phenotypic variance in genetically confirmed beta cell MODY is explained by common genetic variation through beta cell pathways. MODY without a known mutation is likely to have an extreme polygenic background of multiple T2D related pathways.

Supported by: Wellcome Trust

Disclosure: J. Murray Leech: None.

226

A specific de novo ACTB variant is a novel cause of syndromic neonatal diabetes

S. Ahmed ¹, V. Lewis¹, J. Russ-Silsby¹, M.N. Wakeling¹, A.T. Hattersley², K. Patel¹, S.E. Flanagan², E. De Franco³;

¹Department of Clinical and Biomedical Sciences, University of Exeter, Exeter, UK, ²University of Exeter Medical School, Exeter, UK, ³Molecular Genetics, University of Exeter Medical School, Exeter, UK.

Background and aims: Neonatal diabetes is a genetic disease diagnosed <6 months of age. Pathogenic variants in >30 causative genes have been reported, but >10% of patients remain without a known aetiology. Identifying novel genes causing neonatal diabetes can improve our understanding of the mechanisms regulating beta-cell development and survival. Our aim was to identify novel causes of neonatal diabetes.

Materials and methods: We performed whole-genome-sequencing in 39 probands with neonatal diabetes and their unaffected, unrelated parents. Genes with de novo variants in ≥2 patients were prioritised for follow-up,