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

Abstracts

Index of Oral Presentations

OP 01 Tirzepatide: all things must SURPASS

OP 02 Present, future and future future of insulin therapy

OP 03 Pregnancy and diabetes: lessons from cohort studies

OP 04 Diet and exposome

OP 05 Walking towards better treatments for diabetic feet and neuropathy

OP 06 Adipocytes and adipose tissue plasticity

OP 07 Regulation of islet function through gene editing and modification

OP 08 Towards curing obesity

OP 09 Treatment beyond metformin

OP 10 Activating two or three G´s; one of us is lonely

OP 11 Be(a)ware of hypoglycaemia unawareness

OP 12 Diabetes and the brain

OP 13 Concerning insulin resistance

OP 14 Regulation of islet hormone secretion in type 1 diabetes and type 2 diabetes

OP 15 Incretins and glucagon in action

OP 16 What new in type 1 diabetes treatment: focus on the islet cells

OP 17 New approaches: a magical mystery tour

OP 18 Highlights in type 1 diabetes

OP 19 Why does diabetes make one sad?

OP 20 NAFLD and omics

OP 21 New strategies for diabetic retinopathy from mice to the clinic

OP 22 Improving the diagnosis and prediction of kidney disease progression

OP 23 More news from SGLT2s

OP 24 All around pregnancy

OP 25 Muscles at work

OP 26 Many digits in digital diabetes

OP 27 Adipocyte impact on obesity and metabolic health

OP 28 Novel signalling pathways regulating islet function

OP 29 Cardiac complications in type 1 and type 2 diabetes

OP 30 Years and years

OP 31 Future for type 1 diabetes treatment

OP 32 Obesity, liver and beyond

OP 33 Are we what we eat?

OP 34 Diabetic neuropathy from the brain to the bowels

OP 35 Omics and more: unbiased approaches to tackle obesity

OP 36 Cardiovascular risk, complications and fitness

OP 37 Inflammation: foe or friend?

OP 38 How dreaming about artificial pancreas enters the real world (data)

OP 39 Pregnancy, diabetes and metabolic health (of mother and child)

OP 40 What is new in 2023 for the treatment of diabetic kidney disease?

OP 41 SNPs, risk scores and gene therapy

OP 42 Risk factors and biomarkers of CVD

OP 43 Old dogs, new tricks

OP 44 Checking in on beta cell function

OP 45 Thinking outside the box to protect the kidney and the foot

OP 46 Endothelium and diabetes complications

OP 47 Senescence and age

OP 48 Different approaches to map signal pathways in islet cells

OP 49 Caring for care delivery

OP 50 Fatty liver, faulty liver

OP 51 "With a little help from my friends": therapies and combinations

OP 52 Looking to the future: risks and benefits of glycaemic control in diabetic retinopathy

OP 53 Elevating and (making) exciting diabetes education

OP 54 Risk factors for complications and mortality

OP 55 How the beta cells die and can be rescued

OP 56 The brain in control

Index of Short Oral Discussions

SO 01 Weight across the lifespan

SO 02 Lifestyle matters

SO 03 From single nutrients to dietary patterns

SO 04 Viruses and infections

SO 05 Autoimmunity in diabetes

SO 06 Diagnosis and prediction

SO 07 Mechanistic insights

SO 08 Coming from the heart

SO 09 Novel risk factors and treatments

SO 10 Body and brain

SO 11 Obesity and prediabetes

SO 12 Epigenetics

SO 13 Monogenic diabetes

SO 14 Heterogeneity of diabetes

SO 15 Transcriptional regulation in islet cells

SO 16 Different approaches to study hormone secretion regulation

SO 17 Modulators of beta cell signal transduction: How do they work?

SO 18 Role of immune cells in type 1 diabetes and type 2 diabetes

SO 19 Transplantation of islet and stem cells today and tomorrow

SO 20 Roads leading to apoptosis in beta cells

SO 21 Protection of beta cells

SO 22 Inflammation in type 1 diabetes

SO 23 Drug interventions in diabetes and associated disease

SO 24 Grades of insulin sensitivity

SO 25 Exercise as medicine

SO 26 Hormones from the gut and the liver

SO 27 Glucagon secretion and peripheral hormone action

SO 28 Pregnant with type 1 diabetes and adverse pregnancy outcomes

SO 29 Gestational diabetes

SO 30 Insulin secretion: for better or for worse

SO 31 It's in the brain

SO 32 Keeping focus on GLP1

SO 33 Weight loss rules

SO 34 Metabolic flames

SO 35 Sex and time matters

SO 36 Obesity and diabetes across generations

SO 37 Hormonal control of obesity

SO 38 Fantastic fats and where to find them

SO 39 Fat across organs

SO 40 Metabolism in humans

SO 41 Novel players in adipocyte biology

SO 42 Tirzepatide: SURPASSing the vibe check

SO 43 Small molecules to combat a large disease

SO 44 I heard the news today: novel incretins

SO 45 Dual agonism: two of us

SO 46 GLP-1 receptor agonists: action beyond glycaemia

SO 47 Incretin trials: trying new directions

SO 48 Incretins: a day in the life

SO 49 Therapies: here, there and everywhere

SO 50 Type 1 diabetes

SO 51 SGLT2 inhibitors: trying new directions

SO 52 Targeting muscle, kidney and liver

SO 53 Miscellaneous: The White Album

SO 54 Carbohydrates: time matters

SO 55 Nutrition: strawberry fields or glass onions?

SO 56 Devising diabetes devices

SO 57 How to make a good script for insulin therapy for the demanding audience

SO 58 To hell with the hypo

SO 59 How does AID aid the patients?

SO 60 How bright is the hybrid?

SO 61 Refining and redefining insulin therapy in type 1 diabetes

SO 62 If it is Tuesday, it is time for insulin injection

SO 63 Patient-centered perfecting of insulin therapy

SO 64 Quantifying hypoglycaemia: from mild to mad

SO 65 Yin and yang of CGMing

SO 66 Insulin therapy can be better and cheaper

SO 67 Un/successful fitting diabetes in a health care system

SO 68 Where do the happiest people with diabetes live?

SO 69 "When I'm sixty-four" (low and not-so-young)

SO 70 Will rise of the machines terminate diabetes?

SO 71 Making it all work

SO 72 When diabetes goes digital

SO 73 More or less likely but still important determinants of diabetes care

SO 74 Clinical observations to improve the prediction and progression of diabetic kidney disease

SO 75 Clinical outcomes and biomarkers in diabetic kidney disease

SO 76 Diabetic kidney disease pathophysiology from cells to mice, and humans

SO 77 New insights to the drivers of diabetic retinopathy progression

SO 78 Biomarkers and risk factors for diabetic foot disease and peripheral neuropathy

SO 79 Diabetic complications: observational studies from nerves to blood vessels

SO 80 Beyond the clinical examination: emerging diagnostics in diabetic neuropathy

SO 81 Diabetes and the heart

SO 82 Trends in cardiovascular complications

SO 83 Predicting cardiovascular events in diabetes

SO 84 Macrovascular disease: basic science

SO 85 Diabetes, lipids and the vessels

SO 86 Diabetic complications from head to toe

SO 87 Diabetic complications: the heart, the kidney and the bone

SO 88 Pathogenic mechanism of diabetic complications

SO 89 Diabetes: a systemic disease

SO 90 Epidemiology of cardiovascular disease in diabetes: What is new?

SO 91 Experimental models of diabetes: related complications

SO 92 Cognitive impairment in diabetes

SO 93 The Liver: from cell to clinic

SO 94 Diabetes and NAFLD

SO 95 NAFLD: relevant biomarkers

SO 96 NAFLD: future treatment

OP 01 Tirzepatide: all things must SURPASS

1

Tirzepatide achieves significant weight loss without adverse effects on muscle composition (SURPASS-3 MRI)

J. Linge ¹, I.J. Neeland², O. Dahlqvist Leinhard¹, N. Sattar³, L. Fernández Landó⁴, R. Bray⁴, A. Rodriguez⁴;

¹AMRA Medical AB, Linköping, Sweden, ²Harrington Heart and Vascular Institute, University Hospitals Cleveland Medical Center and Case Western Reserve University School of Medicine, Cleveland, OH, USA, ³Institute of Cardiovascular and Medical Sciences, BHF Glasgow Cardiovascular Research Centre, University of Glasgow, Glasgow, UK, ⁴Eli Lilly and Company, Indianapolis, IN, USA.

Background and aims: Aging and weight loss commonly result in loss of muscle mass, which is exacerbated in type 2 diabetes (T2D). Tirzepatide (TZP), a glucose-dependent insulinotropic polypeptide (GIP)/glucagon-like peptide-1 (GLP-1) receptor agonist, reduced body weight with a beneficial effect on fat distribution pattern in the SURPASS-3 MRI substudy. The aim of this exploratory analysis was to examine the effect of TZP on muscle composition and compare observed changes with TZP (5/10/15 mg) or insulin degludec (IDeg) to changes predicted from the longitudinal UK Biobank (UKB) imaging study.

Materials and methods: Thigh muscle fat infiltration (MFI), fat-free muscle volume (FFMV), and sex-, height-, and weight-invariant FFMV z-scores were quantified from MRI using AMRA Researcher. Predicted changes for MFI, FFMV, and FFMV z-score in the SURPASS-3 MRI substudy (N=296 participants with T2D inadequately controlled on metformin with/without sodium-glucose co-transporter-2 inhibitors and fatty liver index ≥60) were calculated using data from UKB. All prediction models were fitted to N=2942 UKB participants (scanned ~2.2y apart and included regardless of observed weight change) and adjusted for sex, baseline age, change in age, interaction between T2D status and weight change, as well as baseline value of either MFI, FFMV or FFMV z-score. Paired t-tests were used to compare observed with predicted change.

Results: MFI was significantly reduced across TZP arms (mean [SD] difference -0.23 [0.77]/-0.42 [0.61]/-0.44 [0.81] percentage points [pp] for TZP 5/10/15 mg, respectively), but not for IDeg (+0.03 [0.40] pp); Changes were greater than predicted by UKB (Fig1, left). FFMV was significantly reduced for TZP 5/10/15 mg (-0.44 [0.57]/-0.71 [0.60]/-0.76 [0.74] litres), but not for IDeg (+0.16 [0.54] litres); Changes were overall similar to those predicted by UKB (Fig1, right). FFMV z-score was significantly reduced for TZP 5/10/15 mg (-0.12 [0.33]/-0.23 [0.48]/-0.30 [0.47] SD, but not for IDeg (+0.06 [0.43] SD); Changes were greater than predicted by UKB, but the difference was only significant for TZP 15mg (p=0.004).

Conclusion: In SURPASS-3 MRI, TZP resulted in significantly improved body weight and fat levels without apparent adverse effects on muscle composition, with greater than predicted reduction in MFI and as-predicted reduction in FFMV when compared to changes described by data from the longitudinal UKB imaging study.

Clinical Trial Registration Number: NCT03882970

Supported by: Eli Lilly and Company

Disclosure: J. Linge: Employment/Consultancy; Employee - AMRA Medical AB.

2

Tirzepatide reduces muscle fat infiltration relative to insulin deglucec in people with type 2 diabetes (SURPASS-3 MRI)

N. Sattar ¹, J. Linge², I.J. Neeland³, O. Dahlqvist Leinhard², L. Fernández Landó⁴, R. Bray⁴, Á. Rodríguez⁴;

¹Institute of Cardiovascular and Medical Sciences, BHF Glasgow Cardiovascular Research Centre, University of Glasgow, Glasgow, UK, ²AMRA Medical AB, Linköping, Sweden, ³Harrington Heart and Vascular Institute, University Hospitals Cleveland Medical Center and Case Western Reserve University School of Medicine, Cleveland, OH, USA, ⁴Eli Lilly and Company, Indianapolis, IN, USA.

Background and aims: Skeletal muscle fat infiltration (MFI) is frequent in obesity and type 2 diabetes (T2D) and is associated with higher risk for cardiovascular disease and mortality. Tirzepatide (TZP), a glucose-dependent insulinotropic polypeptide (GIP)/glucagon-like peptide-1 (GLP-1) receptor agonist, reduces weight and has a beneficial effect on fat distribution pattern. The aim of this exploratory analysis was to determine the effect of pooled TZP doses on MFI compared with insulin degludec (IDeg) in participants in the SURPASS-3 MRI substudy.

Materials and methods: Participants with T2D (on metformin with/without sodium-glucose co-transporter-2 inhibitors [SGLT-2i], HbA_1c 53-91 mmol/mol [7-10.5%], BMI ≥25 kg/m², fatty liver index ≥60) were treated with once-weekly TZP 5 mg (N=71), 10 mg (N=79), 15 mg (N=72), or once-daily titrated IDeg (N=74). Body composition, including MFI, was assessed using MRI at baseline and week 52. Changes from baseline and treatment differences were tested with paired/standard t-tests, respectively. Associations of changes in MFI with changes in body composition and biomarkers were assessed using Spearman correlation coefficients.

Results: At baseline, participants had mean age of 56.2 years, T2D diagnosis duration of 8.3 years, HbA_1c of 67 mmol/mol (8.2%), weight of 94.4 kg, BMI of 33.5 kg/m², and 30% were on SGLT-2i. Baseline MFI (whole cohort: 7.96% [SD 2.42]) was similar for pooled TZP and IDeg arms. TZP showed a significant decrease in MFI (absolute: -0.36 [SD 0.74] percentage points [pp], relative: -4.48% [8.78], both p<0.001) while IDeg showed no significant change. The difference in mean absolute change between pooled TZP and IDeg was -0.39 pp (95% CI -0.56 to -0.23, p<0.001). For TZP, absolute change in MFI was significantly correlated (p<0.001) with change in weight (r=0.310), visceral fat (r=0.480), and abdominal subcutaneous fat (r=0.373), but not with liver fat (r=0.096, p=0.288) or muscle volume (r=0.085, p=0.293). In addition, absolute change in MFI was significantly correlated with change in triglycerides (r=0.197, p=0.014), HDL-cholesterol (r=-0.231, p=0.004), adiponectin (r=-0.196, p=0.046), and free fatty acids (r=0.304, p=0.002).

Conclusion: In people with T2D, treatment with TZP significantly reduced MFI relative to IDeg, with reductions correlating with favourable changes in weight, abdominal fat, and lipids.

Clinical Trial Registration Number: NCT03882970

Supported by: Eli Lilly and Company

Disclosure: N. Sattar: Employment/Consultancy; Consulting fees from Afimmune, Amgen, AstraZeneca, Boehringer Ingelheim, Eli Lilly and Company, Hanmi Pharmaceutical, Merck Sharp & Dohme, Novartis, Novo Nordisk, Pfizer, and Sanofi. Grants; AstraZeneca, Boehringer Ingelheim, Novartis, and Roche Diagnostics.

3

Tirzepatide reduces albuminuria in patients with type 2 diabetes: post-hoc pooled analysis of SURPASS 1-5

R.J. Wiese ¹, H.J.L. Heerspink², K.R. Tuttle³, I. Pavo¹, A. Haupt¹, Z. Yang¹, A. Hemmingway¹, D.Z.I. Cherney⁴, N. Sattar⁵;

¹Eli Lilly and Company, Indianapolis, IN, USA, ²Department of Clinical Pharmacy and Pharmacology, University Medical Centre Groningen,, University of Groningen, Groningen, Netherlands, ³Providence Health Care, University of Washington, Spokane, WA, USA, ⁴Division of Nephrology, Department of Medicine, Toronto General Hospital, University Health Network, University of Toronto, Toronto, ON, Canada, ⁵University of Glasgow, Glasgow, UK.

Background and aims: In SURPASS 4, the GIP/GLP-1 receptor agonist tirzepatide (TZP) showed a potential kidney protective effect in people with type 2 diabetes (T2D) and high CV risk by slowing the rate of eGFR decline and reducing urine albumin-creatinine ratio (UACR) vs insulin glargine over 2 years. In this post-hoc analysis, we explored effects of TZP on UACR changes in SURPASS 1-5 trials.

Materials and methods: UACR (% difference) for TZP (5, 10, 15 mg) vs comparators (COMPs) was analysed. Analyses were conducted in the pooled entire SURPASS 1-5 population and populations pooled by COMP: placebo (SURPASS 1 & 5); active (SURPASS 2 [semaglutide 1 mg] & SURPASS 3-4 [insulins]); and insulins. In each pooled population, data were examined in all patients and in subgroups defined by baseline UACR ≥3.39 mg/mmol or eGFR <60 mL/min/1.73m². Mixed model for repeated measures was used to analyse on-treatment data from baseline up to the end of treatment visit.

Results: UACR data was available in 6263 patients of whom 1846 had UACR ≥3.39 mg/mmol and 537 had eGFR <60 mL/min/1.73m². UACR decreased more with TZP 5, 10, and 15 mg vs COMPs in pooled SURPASS 1-5 and consistently across pooled placebo, active, and insulin COMP studies (Table). UACR lowering appeared more pronounced in subgroups with baseline UACR ≥3.39 mg/mmol or eGFR <60 mL/min/1.73m².

Conclusion: In people with T2D, including those with reduced kidney function, TZP decreased UACR vs COMPs to a clinically relevant degree, supporting a potential kidney protective effect.

Clinical Trial Registration Number: NCT03954834, NCT03987919, NCT03882970, NCT03730662, NCT04039503

Supported by: Eli Lilly and Company

Disclosure: R.J. Wiese: Employment/Consultancy; Eli Lilly and Company. Stock/Shareholding; Eli Lilly and Company.

4

Time spent with sustained HbA _1c ≤6.5% (47.5 mmol/mol) and weight loss ≥5% after initiating tirzepatide vs semaglutide: an exploratory analysis of SURPASS-2

B.K. Bergman ¹, J. Rosenstock², T. Garvey³, R.L. Batterham⁴, Y. Chen¹, M. Liu⁵, V.T. Thieu¹;

¹Eli Lilly and Company, Indianapolis, IN, USA, ²Velocity Clinical Research at Medical City, Dallas, TX, USA, ³University of Alabama at Birmingham, Birmingham, AL, USA, ⁴University College London, London, UK, ⁵Tigermed-BDM Inc., Somerset, NJ, USA.

Background and aims: Tirzepatide is a first-in-class once weekly (QW) glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide-1 (GLP-1) receptor agonist approved for treatment of people with type 2 diabetes (T2D) and under investigation for chronic weight management. In SURPASS-2, tirzepatide (5 mg, 10 mg, 15 mg) demonstrated significant and clinically meaningful reductions in HbA1c and improvements in body weight in adults with T2D compared to selective GLP-1 receptor agonist, semaglutide 1 mg QW. Here, we compare the continuous time spent in glycemic control with weight loss between tirzepatide and semaglutide in SURPASS-2.

Materials and methods: Participants were randomised to tirzepatide (5, 10, or 15 mg) or semaglutide (1 mg) with the primary endpoint at 40 weeks. The composite endpoint of continuous time spent with an HbA1c ≤6.5% (47.5 mmol/mol) and ≥5% weight loss was analysed in all patients on treatment without rescue medication, with missing data imputed. Logistic regression was used to compare the proportion of patients reaching glycemic/weight target at any time. Wilcoxon rank sum test was used to compare the median duration of continuous time spent in glycemic/weight control.

Results: Participants on all doses of tirzepatide spent longer time with an HbA1c ≤6.5% alone and a longer time with weight loss ≥5% alone than those on semaglutide (p<0.001 vs semaglutide) (Figure). Furthermore, an HbA1c ≤6.5% and weight loss ≥5% was achieved in 66-85% of participants treated with tirzepatide and 54% of participants treated with semaglutide. Participants on all doses of tirzepatide spent longer time with an HbA1c ≤6.5% and weight loss ≥5% than those on semaglutide (p<0.001 vs semaglutide). The median time spent with HbA1c ≤6.5% and weight loss ≥5% ranged from 35-60% of the trial duration (14-24 weeks) with tirzepatide treatment at all doses (5 mg, 10 mg, and 15 mg) vs 7% of the trial duration (2.7 weeks) with semaglutide. The time periods meeting these HbA1c and weight loss targets were greater even at the lowest dose of tirzepatide (5 mg) compared with semaglutide 1 mg and increased further at the 10 and 15 mg doses.

Conclusion: Treatment with tirzepatide led to a longer continuous time spent with an HbA1c ≤6.5% (47.5 mmol/mol) and weight loss ≥5% over the duration of the SURPASS-2 trial compared with semaglutide 1 mg in adults with T2D. Additionally, treatment with tirzepatide led to a longer continuous time spent with an HbA1c ≤6.5% alone and with weight loss ≥5% alone compared to semaglutide.

Clinical Trial Registration Number: NCT03987919

Supported by: Funded by Eli Lilly and Company

Disclosure: B.K. Bergman: Employment/Consultancy; Eli Lilly and Company. Stock/Shareholding; Eli Lilly and Company.

5

Tirzepatide compared to subcutaneous semaglutide for type 2 diabetes: a network meta-analysis

T. Karagiannis ¹, I. Avgerinos¹, K. Malandris¹, A. Stamati², A. Liakos¹, N. Kakaletsis¹, A. Tsapas^1,3, E. Bekiari¹;

¹Clinical Research & Evidence-Based Medicine Unit, Aristotle University of Thessaloniki, Thessaloniki, Greece, ²School of Medicine, Aristotle University of Thessaloniki, Thessaloniki, Greece, ³Harris Manchester College, University of Oxford, Oxford, UK.

Background and aims: There is paucity of randomised controlled trials (RCTs) directly comparing tirzepatide with subcutaneous (SC) semaglutide. We did a network meta-analysis to compare tirzepatide with SC semaglutide in terms of their efficacy and safety in people with type 2 diabetes.

Materials and methods: We searched Medline and the Cochrane Library for RCTs that assessed a maintenance dose of tirzepatide 5, 10, or 15 mg once-weekly or SC semaglutide 0.5, 1.0, or 2.0 mg once-weekly for at least 12 weeks. Comparators were any of the eligible doses of tirzepatide and semaglutide, placebo, and also other glucose-lowering drugs provided they were directly compared to tirzepatide and semaglutide in at least one trial. We did frequentist random-effects network meta-analyses and calculated mean differences (MDs) for change from baseline in HbA1c and body weight, and risk ratios (RRs) for incidence of gastrointestinal adverse events (nausea, vomiting, diarrhea) and serious adverse events.

Results: Twenty-two RCTs (18,472 participants) contributed data to the network meta-analysis. Tirzepatide 15 mg was the most efficacious in reducing HbA1c versus placebo (MD -2.00%; 95% CI -2.16 to -1.84), followed by tirzepatide 10 mg (-1.86%; -2.02 to -1.84) and semaglutide 2.0 mg (-1.62%; -1.96 to -1.28). Each of the three tirzepatide doses reduced HbA1c more than the respective low-, medium-, and high-dose of semaglutide. Versus placebo, tirzepatide was more efficacious in reducing body weight (Figure). Regarding between-drug comparisons, both tirzepatide 10 and 15 mg were more efficacious in lowering body weight versus either semaglutide 1.0 or 2.0 mg, while tirzepatide 5 mg was more efficacious versus semaglutide 0.5 and 1.0 mg. As compared to placebo, all doses of tirzepatide and semaglutide increased risk for gastrointestinal adverse events, with tirzepatide 15 mg yielding the highest RR for nausea (RR 3.57; 2.56 to 4.76), vomiting (4.35; 3.03 to 6.25) and diarrhea (2.04; 1.56 to 2.63). Between-drug comparisons yielded non-significant results, except for tirzepatide 15 mg versus semaglutide 1.0 mg (1.39; 1.03 to 1.85) and 0.5 mg (1.85; 1.35 to 2.56) in vomiting, and versus semaglutide 0.5 mg (1.45; 1.09 to 1.92) in nausea. There was no difference between tirzepatide, semaglutide and placebo regarding risk for serious adverse events.

Conclusion: In people with type 2 diabetes, tirzepatide 5, 10, and 15 mg were more efficacious in reducing HbA1c compared to semaglutide 0.5, 1.0, and 2.0 mg, respectively. Superiority of tirzepatide over semaglutide was more pronounced in terms of body weight loss. High-dose tirzepatide (15 mg) was associated with increased risk for vomiting versus low- and medium-dose of semaglutide.

Disclosure: T. Karagiannis: None.

6

“Surpass(ing)” an era of basal-bolus insulin therapy: tirzepatide vs insulin lispro tid added-on to poorly controlled basal insulin-treated type 2 diabetes

J.P. Frías ¹, J. Rosenstock², H.W. Rodbard³, S. Tofé⁴, E. Sears⁵, R. Huh⁵, L. Fernández Landó⁵, H. Patel⁵;

¹Velocity Clinical Research, Los Angeles, CA, USA, ²Velocity Clinical Research at Medical City, Dallas, TX, USA, ³Endocrine and Metabolic Consultants, Rockville, MD, USA, ⁴University Hospital Son Espases, Palma de Mallorca, Spain, ⁵Eli Lilly and Company, Indianapolis, IN, USA.

Background and aims: Tirzepatide (TZP) is a once weekly GIP and GLP-1 receptor agonist for the treatment of type 2 diabetes (T2D). Efficacy and safety of TZP vs insulin lispro (iLispro) was assessed in people with T2D with inadequately controlled glycemic control on insulin glargine (iGlar) ± metformin (MET).

Materials and methods: In this 52-wk open-label, multicenter, Phase 3b study, 1428 patients (mean baseline age 59 y; T2D duration 14 y; HbA1c 8.8% [72.7 mmol/mol]; BMI 33 kg/m²; iGlar dose 46 IU/day [pooled TZP: 47.1 IU/day (0.53 IU/kg/day), iLispro: 47.5 IU/day (0.54 IU/kg/day)]) were randomised (1:1:1:3) to TZP (5, 10, 15 mg) or iLispro TID, add-on to iGlar ± MET. Insulin doses were titrated to a target fasting and pre-prandial glucose of 100-125 mg/dL (5.6-6.9 mmol/L). Primary efficacy measure was change in HbA1c from baseline for TZP (pooled) vs iLispro at 52 wk. Secondary measures were change in body weight (BW) from baseline, and proportion of patients achieving HbA1c and BW loss goals (TZP pooled and each dose vs iLispro). Measures tested for superiority were HbA1c, BW, and HbA1c target <7.0% (controlled for Type 1 error) and HbA1c targets ≤6.5%, <5.7% and BW loss ≥5% (not controlled for Type 1 error). Analyses included on treatment data prior to initiating rescue therapy from all patients who received at least 1 dose of study drug, excluding patients discontinuing study drug due to inadvertent enrollment.

Results: At 52 wk, TZP (pooled) was superior vs iLispro in change in HbA1c from baseline, achieving a mean HbA1c of 6.50% vs 7.56% (47.6 vs 59.1 mmol/mol) (Table). LSM treatment differences (95% confidence interval) were -0.89% (-1.08, -0.70) (-9.7 mmol/mol [-11.8, -7.7]) for TZP 5 mg, -1.11% (-1.30, -0.92) (-12.1 mmol/mol [-14.2, -10.1]) for TZP 10 mg, and -1.30% (-1.49, -1.11) (-14.2 mmol/mol [-16.3, -12.1]) for TZP 15 mg (p<0.001, all doses vs iLispro). Glycemic improvement was achieved with substantially less insulin use (iGlar: 13 IU/day vs 42 IU/day [iLispro 62 IU/day]) and BW loss of 10 kg vs gain of 4 kg, respectively. The most common adverse events were gastrointestinal and mostly mild to moderate in severity (nausea: 14-26% vs 1%, diarrhoea: 11-15% vs 2%, vomiting: 5-13% vs 1%). Hypoglycaemia events (<54 mg/dL or severe) were reported in 12% (0.6 event rate/patient-year), 9% (0.4 event rate/patient-year) and 11% (0.2 event rate/patient-year) of patients treated with TZP vs 48% (4.4 event rate/patient-year) with iLispro. Three patients (1%; TZP 10 mg) reported 17 episodes of severe hypoglycaemia (0.05 event rate/patient-year).

Conclusion: TZP demonstrated clinically meaningful and superior glycaemic and BW control vs iLispro associated with substantially less hypoglycaemia and insulin use.

Clinical Trial Registration Number: NCT04537923

Supported by: Funded by Eli Lilly and Company

Disclosure: J.P. Frías: Employment/Consultancy; Akero, Altimmune, Axcella Health, Becton Dickenson, Boehringer Ingelheim, Carmot Therapeutics, Echosens, 89bio, Eli Lilly and Company, Gilead, Intercept, Metacrine, Merck, Novo Nordisk, Pfizer, Sanofi. Grants; Akero, Boehringer Ingelheim, BMS, 89bio, Eli Lilly and Company, Intercept, IONIS, Janssen, Madrigal, Merck, NorthSea Therapeutics, Novartis, Novo Nordisk, Oramed, Pfizer, Poxel, Sanofi. Other; Akero, Altimmune, Axcella Health, Becton Dickenson, Boehringer Ingelheim, Carmot Therapeutics, Echosens, 89bio, Eli Lilly and Company, Gilead, Intercept, Metacrine, Merck, Novo Nordisk, Pfizer, Sanofi.

OP 02 Present, future and future future of insulin therapy

7

The effectiveness of closed-loop systems is maintained after one year of use

A. Romero Gregori, L. Albert Fàbregas, D. Subías Andújar, I. Capel Flores, A. Cano Palomares, I. Mazarico, M. Luchtenberg, M. Rigla;

Endocrinology and Nutrition, Hospital Parc Taulí, Sabadell, Spain.

Background and aims: Few real-world studies of individuals using closed-loop systems (CLS) exist, mainly employing a single model of CLS for six months or less. The purposes of this study were: (1) To evaluate the long-term real-world efficacy of three CLS (Minimed 780G, DBLG1 -Diabeloop- and Control-IQ Tandem); (2) to assess the factors related to the achievement of near-normal glycaemic control after 12 months of CLS use; (3) to evaluate the relationship between time-in-range (TIR), HbA1c level, and the glucose management indicator (GMI).

Materials and methods: A retrospective longitudinal study was performed. It included type 1 diabetes adults who initiated a CLS between April 2021 and March 2022 in our hospital and provided data for ≥ 12 months. We evaluated the initial, 3-month, 6-month, and 12-month data regarding glycaemic outcomes and their baseline characteristics. The statistical analysis of the association of characteristics with 12-month HbA1c consisted of univariate analysis followed by multivariable analysis using linear regression with continuous 12-month HbA1c as the dependent variable.

Results: One-hundred and thirty-two CLS users were included (90 female, mean age 45.2 ± 11.9 years, mean weight 75.5 ± 15.6 kg, mean BMI 27.2 ± 4.8 kg/m2, mean baseline HbA1c 7.5 ± 0.9%). 38% were previously under multiple daily injections, 92% used continuous glucose monitoring (CGM) and 54% used an insulin bolus advisor. Different commercial CLS were used: 59 DBLG1, 40 Minimed 780G, and 33 Control-IQ. Three-month HbA1c significantly decreased to 6.8% ± 0.7% (-0.68% ± 0.70%, p 0.000), and 12-month HbA1c remained stable (HbA1c 6.6% ± 0.6%; 77% achieved HbA1c <7%, 17% HbA1c 7-7.5%, and 6% HbA1c ≥7.5%). There was a good correlation between 3-month HbA1c and 12-month HbA1c (r= 0.64, p 0.000). At 12 months mean GMI was 6.9%±0.4% (54% GMI<7%, 44% GMI 7-7.5%, 2% GMI>7.5%), mean TIR was 74% ± 9% (69% TIR>70%) and mean time below range <70 mg/dl (TBR) was 1.9% ± 1.4%.After 12 months we observed better glycemic control with the Control-IQ system (HbA1c 6.4% ± 0.7%, GMI 6.8% ± 0.3%, and TIR 76% ± 9%) and the Minimed 780G system (HbA1c 6.7% ± 0.5%, GMI 6.6% ± 0.3%, and TIR 79% ± 7%) compared with the DBLG1 system (HbA1c 6.8% ± 0.5%, GMI 7.1% ± 0,3%, and TIR 69% ± 7%).In univariate analyses initial-HbA1c, 3-month HbA1c, 6-month HbA1c, and 12-month GMI were positively associated with 12-month HbA1c; whereas 12-month TIR was inversely associated with 12-month HbA1c (all p 0.000). Regarding patient characteristics, Control-IQ use, a higher education level, a higher sensor use, and the number of daily carbohydrates were associated inversely with 12-month HbA1c (all p<0.04). We did not find any association with gender, age, initial weight, exercise, DM duration, prior DM treatment, previous use of insulin bolus advisor, or insulin doses. In the multivariable analysis 3-month HbA1c, the CLS model, and higher education levels remained associated with 12-month HbA1c levels. The correlation between TIR and GMI (r = -0.91, p 0.000) was stronger than between TIR and HbA1c (r= -0.56, p 0.000).

Conclusion: In almost all cases the long-term use of a CLS improved glycemic control substantially. Factors related to 12-month HbA1c levels are the CLS model used, a higher education level, and the 3-month HbA1c levels.

Disclosure: A. Romero Gregori: None.

8

Better glycaemic control with once-weekly insulin icodec versus once-daily insulin degludec in adults with insulin-naïve type 2 diabetes (ONWARDS 3)

I. Lingvay ¹, M. Asong², C. Desouza³, P. Gourdy⁴, S. Kar⁵, A. Vianna⁶, T. Vilsbøll^7,8, S. Vinther², Y. Mu⁹;

¹Division of Endocrinology, Department of Internal Medicine and Peter O’Donnell Jr. School of Public Health, University of Texas Southwestern Medical Center, Dallas, TX, USA, ²Novo Nordisk A/S, Søborg, Denmark, ³University of Nebraska Medical Center, Omaha, NE, USA, ⁴CHU de Toulouse & UMR1297/I2MC, Inserm, Université Toulouse 3, Toulouse, France, ⁵Novo Nordisk Service Centre India Private Ltd., Bangalore, India, ⁶Curitiba Diabetes Centre, Department of Endocrine Diseases, Hospital Nossa Senhora das Graças, Curitiba, Brazil, ⁷Clinical Research, Steno Diabetes Center Copenhagen, University of Copenhagen, Herlev, Denmark, ⁸Department of Clinical Medicine, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark, ⁹Department of Endocrinology, The First Medical Center of Chinese People's Liberation Army General Hospital, Beijing, China.

Background and aims: Efficacy and safety of once-weekly insulin icodec (icodec) versus once-daily insulin degludec (degludec) in adults with insulin-naïve type 2 diabetes were examined in ONWARDS 3, a 26-week, randomised, double-blind, double-dummy, treat-to-target, phase 3a trial.

Materials and methods: Participants (n=588) were randomised 1:1 to icodec or degludec. Key outcomes are shown in the table. The primary endpoint was HbA_1c change from baseline to week 26.

Results: Estimated mean HbA_1c change from baseline was greater with icodec (−1.57%; baseline: 8.55%) versus degludec (−1.36%; baseline: 8.48%); estimated treatment difference: −0.21%-points (95% CI −0.34, −0.08), confirming non-inferiority (p<0.0001) and superiority (p=0.0016). Level 2 or 3 hypoglycaemia rates were numerically but not statistically significantly higher with icodec versus degludec in the on-treatment period (table). Odds of achieving HbA_1c <7% without level 2 or 3 hypoglycaemia were statistically significantly higher with icodec versus degludec. There were no significant differences in weekly insulin dose from week 24 to 26 or body weight change from baseline to week 26 between arms. No unexpected safety findings were observed.

Conclusion: Overall, once-weekly icodec demonstrated superior HbA_1c reduction versus once-daily degludec with low level 2 or 3 hypoglycaemia rates in both arms. Icodec may facilitate insulin treatment acceptance and adherence by reducing basal insulin injections from ≥365 to 52 per year.

Clinical Trial Registration Number: NCT04795531

Supported by: Medical writing support was provided by Oxford PharmaGenesis, UK, funded by Novo Nordisk A/S

Disclosure: I. Lingvay: Employment/Consultancy; Novo Nordisk, Eli Lilly, Sanofi, AstraZeneca, Boehringer Ingelheim, Johnson and Johnson, Intercept, TARGETPharma, Merck, Pfizer, Valeritas, Zealand Pharma, Shionogi, Carmot Therapeutics, Structure Therapeutics, Bayer, Translational Medical Academy, WebMD. Other; Novo Nordisk, Sanofi, Merck, Pfizer, Mylan, Boehringer Ingelheim.

9

Efficacy and safety of once-weekly insulin icodec versus once-daily insulin degludec in type 1 diabetes: ONWARDS 6

D. Russell-Jones ¹, T. Babazono², R. Cailleteau³, S. Engberg³, C. Irace⁴, M.I.S. Kjaersgaard⁵, C. Mathieu⁶, J. Rosenstock⁷, V. Woo⁸, D.C. Klonoff⁹;

¹Department of Diabetes and Endocrinology, University of Surrey, Guildford, UK, ²Department of Internal Medicine, Tokyo Women's Medical University School of Medicine, Tokyo, Japan, ³Novo Nordisk A/S, Søborg, Denmark, ⁴Department of Health Science, University Magna Græcia, Viale Europa, Località Germaneto, Catanzaro, Italy, ⁵Novo Nordisk A/S, Aalborg, Denmark, ⁶Clinical and Experimental Endocrinology, University of Leuven, Leuven, Belgium, ⁷Velocity Clinical Research at Medical City, Dallas, TX, USA, ⁸University of Manitoba, Winnipeg, MB, Canada, ⁹Diabetes Research Institute, Mills-Peninsula Medical Center, San Mateo, CA, USA.

Background and aims: Efficacy and safety of once-weekly insulin icodec (icodec) vs once-daily insulin degludec (degludec) were assessed in adults with type 1 diabetes (T1D) in a randomised, open-label, treat-to-target phase 3a trial (26-week main phase + 26-week safety extension + 5-week follow-up).

Materials and methods: Participants (n = 582) were randomised 1:1 to icodec (n = 290) or degludec (n = 292), both in combination with insulin aspart. The primary endpoint was change in HbA_1c from baseline (BL) to week 26, tested for non-inferiority (0.3% margin).

Results: At week 26, from BL values of 7.59% (icodec) and 7.63% (degludec), estimated mean HbA_1c changes were -0.47% and -0.51%, respectively, confirming non-inferiority of icodec vs degludec (estimated treatment difference [ETD]: 0.05%-points [95% CI: -0.13 to 0.23]). From week 22 to 26, percentage of time in range 3.9-10.0 mmol/L (70-180 mg/dL) and time spent >10.0 mmol/L (>180 mg/dL) were not statistically significantly different between treatments, while percentage of time spent <3.0 mmol/L (<54 mg/dL) was statistically significantly higher with icodec vs degludec (p=0.0014). Overall rate of combined level 2 or 3 hypoglycaemia (BL to week 26) was statistically significantly higher with icodec vs degludec (19.9 [icodec] vs 10.4 [degludec] events per patient-year of exposure; estimated rate ratio: 1.9 [95% CI: 1.5; 2.3]; p<0.0001). The rate was also statistically significantly higher with icodec vs degludec when evaluated over 57 weeks. Estimated mean weekly basal dose was statistically significantly higher for icodec than degludec from week 24 to 26 (170 U/week [~24 U/day] vs 151 U/week [~22 U/day]; estimated treatment ratio [ETR]: 1.12 [95% CI: 1.07 to 1.18]; p<0.0001), while estimated mean weekly bolus dose was statistically significantly lower (132 U/week [~19 U/day] vs 161 U/week [~23 U/day]; ETR: 0.82 [95% CI: 0.74 to 0.90]; p<0.0001). Similar trends in insulin doses were reported for week 50 to 52.

Conclusion: In adults with T1D, once-weekly icodec showed non-inferiority in HbA_1c reduction vs degludec at week 26 with significantly higher rate of combined level 2 or 3 hypoglycaemia. For icodec, time spent <3.0 mmol/L (<54 mg/dL) and <3.9 mmol/L (<70 mg/dL) was either below or close to internationally recommended targets at both time points.

Clinical Trial Registration Number: NCT04848480

Supported by: This study was funded by Novo Nordisk A/S

Disclosure: D. Russell-Jones: Employment/Consultancy; Professor of Diabetes and Endocrinology and Consultant Physician, Royal Surrey Hospital and University of Surrey, Diabetes independent Consultant advisor to UK Civil Aviation Authority, Secretary of State’s Advisory Panel for Driving with diabetes. Grants; AstraZeneca, Dexcom, Eli Lilly, Novo Nordisk, Sanofi. Honorarium; AstraZeneca, Dexcom, Eli Lilly, Novo Nordisk, Sanofi.

10

Once-weekly insulin efsitora alfa and daily degludec achieve comparable glycaemic control in 3 patient populations during phase 2: clinical insights from CGM assessments

C. Kazda ¹, J. Bue-Valleskey¹, Q. Zhang¹, A. Haupt¹, J. Frias², D. Dahl³;

¹Eli Lilly and Company, Indianapolis, IN, USA, ²Velocity Clinical Research, Los Angeles, CA, USA, ³Gemeinschaftspraxis fur Innere Medizin und Diabetologie, Hamburg, Germany.

Background and aims: Once-weekly insulin efsitora alfa (efsitora, LY3209590, basal insulin Fc [BIF]) is an insulin receptor agonist that combines a novel single-chain analogue of insulin with a human IgG2 Fc domain. It is specifically designed for once weekly subcutaneous administration. CGM-derived parameters from efsitora phase 2 studies were reviewed and summarized to evaluate similarities and differences in glucose during treatment with daily or weekly basal insulin administration.

Materials and methods: Three randomized, open-label, treat-to-target phase 2 studies of 26 or 32 week duration assessed 912 people with T2D previously treated with basal insulin (T2D-Switch), new to insulin therapy (T2D-Naïve), or T1D on multiple daily injections (T1D). Studies assessed once-weekly efsitora vs once daily degludec (DEG). Key results were previously disclosed and met the primary endpoint of non-inferiority in HbA1c change from baseline (non-inferiority margin=0.4%). Two studies had CGM during entire study and T2D-Naïve had blinded, intermittent CGM using 14-day sessions.

Results: Patients had baseline HbA1c of 65.02 mmol/mol (8.1%), 63.93 mmol/mol (8.0%), and 58.46 mmol/mol (7.5%) for T2D-Switch, T2D-Naïve, and T1D respectively. Figure shows average hourly mean glucose at endpoint and time in range for T2D-Switch and T1D (inset). Time in range for T2D-Naïve: efsitora 76.0%; -1.4 (-5.4,2.7), p=.578; DEG 77.4%. Glucose profiles appeared comparable between treatments with 24-hour average glucose for efsitora vs DEG in T2D-Naïve: 7.6 mmol/L vs 7.05 mmol/L (137 vs 127 mg/dL) p=.042 and T1D: 175 vs 169 p=.075 (T2D-Switch not available). Average duration of hypoglycaemia (3.9 mmol/L [<70mg/dL]) per patient in minutes for efsitora vs DEG: T2D-Switch 31 vs 35 p=NS, T2D-Naïve 51 vs 91 p<.001, and T1D 35 vs 35 p=NS.

Conclusion: Once-weekly efsitora achieved similar glycaemic control vs DEG as seen in several CGM derived measures providing useful and reassuring insights into the use of once-weekly efsitora in these patient populations.

Clinical Trial Registration Number: NCT04450394, NCT04450407, NCT03736785

Supported by: Eli Lilly and Company

Disclosure: C. Kazda: Employment/Consultancy; Eli Lilly and Company. Stock/Shareholding; Eli Lilly and Company.

11

Efficacy and safety of a functionally selective insulin: a 26-week randomised, double blind trial versus insulin glargine in people with type 2 diabetes

T. Heise ¹, H. Haraldsson², T. Herbrand¹, L. Johansson², J. Jordan³, E. Nishimura⁴, L. Plum-Mörschel¹, J. Tank³, Ó. Thórisdóttir⁵, K. Lyby⁵;

¹Profil, Neuss, Germany, ²Antaros Medical, Mölndal, Sweden, ³Institute of Aerospace Medicine, German Aerospace Center, Cologne, Germany, ⁴Research & Early Development, Novo Nordisk, Måløv, Denmark, ⁵Medical & Science, Novo Nordisk, Søborg, Denmark.

Background and aims: NNC-965 is an acylated insulin analogue with an innovative mechanism of action selectively stimulating the insulin receptor with greater Akt and less MAPK activation. In diabetic rats, at doses with equivalent blood glucose lowering effects, NNC-965 improved vascular function and insulin sensitivity, reduced hepatic lipid accumulation in the liver, and induced less weight gain compared to basal insulin with a standard signalling profile. These positive effects could potentially reduce long-term cardiovascular risk. Therefore, we tested the hypothesis that NNC-965 improves Flow-Mediated Dilation (FMD), a measure of endothelial function, in people with T2D compared to current basal insulin therapy.

Materials and methods: People with T2D on basal insulin +/- metformin/DPP4i/SGLT2i were randomised 1:1 to NNC-965 or insulin glargine once daily for 26 weeks in a double-blind design. Insulin doses were titrated weekly aided by a pre-specified algorithm with a target fasting plasma glucose of 4.4-7.2 mmol/L. Before and at the end of the treatment period, participants were submitted to detailed cardiovascular assessments including FMD, the primary endpoint, and Pulse Wave Velocity, Leg Blood Flow, Retinal Arteriolar Dilation and Left Ventricular Ejection Fraction.

Results: Eighty-six (86) participants were randomized and exposed (mean±SD age 63±7.9 years, BMI 30.7±4.3 kg/m², HbA1c 7.6±0.9 %). Both treatments similarly improved glycaemic control (table). NNC-965 also improved FMD and insulin sensitivity (Rate of Glucose Disposal in an insulin tolerance test) compared to insulin glargine. Liver Fat decreased similarly on both treatments. However, NNC-965 increased body weight and body fat mass compared to insulin glargine. No statistically significant differences between NNC-965 and insulin glargine were identified for any of the other supportive secondary endpoints (Pulse Wave Velocity, Leg Blood Flow, Retinal Arteriolar Dilation, Left Ventricular Ejection Fraction, hypoglycaemic episodes).

Conclusion: With similar glycaemic control the NNC-965 novel functionally selective insulin analogue improved endothelial function measured by FMD and improved insulin sensitivity compared to insulin glargine, however it increased body weight and body fat mass. The finding further supports the idea that a functionally selective insulin could have a beneficial effect on long-term cardiovascular risk.

Clinical Trial Registration Number: NCT04575181

Supported by: Novo Nordisk A/S

Disclosure: T. Heise: Grants; Adocia, AstraZeneca, Biocon, Crinetics, Eli Lilly, Genova, Novo Nordisk, Sanofi, Zealand. Honorarium; Novo Nordisk. Lecture/other fees; Eli Lilly, Gan & Lee Pharmaceuticals, Novo Nordisk.

12

Ultra-rapid lispro is more protective against exercise-associated hypoglycaemia over lispro in adults with type 1 diabetes on continuous subcutaneous insulin infusion

J. Leohr ¹, A. Abibol², E. Smith LaBell¹, A. Ghosh¹, L.V. Turner³, D. Shakeri³, T. Fukuda¹, M.C. Riddell^2,3;

¹Eli Lilly and Company, Indianapolis, IN, USA, ²LMC Healthcare/Centricity Research, Toronto, ON, Canada, ³Muscle Health Research Centre, York University, Toronto, ON, Canada.

Background and aims: Individuals with type 1 diabetes (T1D) on continuous subcutaneous insulin infusion (CSII) typically have a large reduction in blood glucose levels during aerobic exercise, even when an insulin basal rate reduction (BRR) occurs at or before the onset of exercise. Ultra rapid lispro (URLi) is a new insulin lispro formulation with a faster onset of action and shorter duration of action compared to currently available rapid-acting insulin analogues. URLi’s shorter duration of insulin action could protect against exercise-induced hypoglycaemia when individuals on CSII perform a BRR prior to exercise. The aim of this study was to assess if using URLi resulted in less exercise-induced glucose lowering compared to Lispro with two different BRR approaches during CSII therapy.

Materials and methods: This was a double-blind, 4-period crossover, randomized (1:1:1:1) controlled trial in adults with T1D on standard CSII to evaluate URLi vs Lispro during either a 50% BRR started 60 minutes or a 100% BRR started 15 minutes prior to the onset of aerobic exercise. Primary endpoint was the change in glucose during exercise, with insulin lispro PK as the secondary endpoint. Physically active men and women with T1D (VO2max =41.6 ±7.21 ml/kg/min) underwent basal/bolus optimization with Lispro prior to randomization. Each period, exercise (treadmill walking, for 60min ~55% VO2max) was performed 4 hours after a standardized meal at lunch with their usual prandial insulin dose.

Results: Twenty-five participants (52% male) had a mean ±SD age 36.7±10.3 years, and HbA1c 50.4± 8.51 mmol/mol (6.8 ±0.78%). Mean ±SE plasma glucose levels prior to the exercise onset were similar between URLi and Lispro for the 50% BRR (8.6 ± 0.43mmol/L vs 9.5 ± 0.38mmol/L; p=0.15) and for the 100% BRR (9.2 ± 0.47mmol/L vs 9.0 ± 0.42 mmol/L; p=0.74). Exercise-induced glucose lowering was less pronounced with URLi vs Lispro, independent of BRR (Figure). During the 50% BRR, URLi attenuated the reduction in mean glucose level compared to Lispro (-1.5 ± 0.41 mmol/L vs -2.2 ± 0.43 mmol/L). Similarly, URLi also attenuated the reduction in mean glucose during the exercise compared to Lispro during the 100% BRR (-2.6 ± 0.35 mmol/L vs -3.4 ± 0.44 mmol/L). In both the cases, the differences between study treatments for each BRR were found to be statistically significant (p<0.05). Likewise, less hypoglycaemic events were observed with URLi (12%) compared to Lispro (32%) during the exercise period for both BRR approaches.

Conclusion: URLi demonstrated an attenuation of exercise-induced glucose lowering compared to Lispro during aerobic exercise using either a 50% or 100% BRR approach. These data suggest that URLi may protect against exercise-associated hypoglycaemia in adults with TID on standard CSII therapy.

Clinical Trial Registration Number: NCT05262387

Disclosure: J. Leohr: None.

OP 03 Pregnancy and diabetes: lessons from cohort studies

13

Pregnancy and neonatal outcomes following bariatric surgery in Queensland, Australia: a data-linkage matched cohort study

J. Eccles-Smith;

Mater Research Institution- The University of Queensland, Brisbane, Australia.

Background and aims: In Australia, bariatric procedures doubled between 2005 and 2015 with 80% performed on women of child bearing age. Pregnancy following bariatric surgery is associated with mixed maternal and fetal outcomes. Lower rates of hypertensive disorders and gestational diabetes but higher rates of small for gestational age infants and pre-term delivery are reported predominantly from women with previous roux-en y gastric bypass. Limited data are available regarding pregnancy outcomes after laparoscopic sleeve gastrectomy. This data-linkage project analysed the pregnancy and neonatal outcomes of women who delivered between 2013-2018 following bariatric surgery against matched controls within Queensland, Australia.

Materials and methods: Singleton pregnancies with prior maternal bariatric surgery (n=1,282) registered in the Queensland Hospital Admitted Patient Data Collection were control-matched (n = 12,820) with deliveries during 2013-2018 in the Perinatal Data Collection. Women were matched 1:10 by maternal BMI, smoking, age and parity with the first pregnancy following bariatric surgery for each woman used for analysis. Pregnancy and neonatal outcomes were analysed using paired t-tests, Chi-square or Fishers exact test as appropriate.

Results: Of 1282 women with a singleton delivery after bariatric surgery, 89% had undergone laparoscopic sleeve gastrectomy). The proportion of assisted reproductive technology use was higher in women with previous bariatric surgery 10.7% (n=137) vs 8.0% (n=1,022) matched women (p<0.001). Rates of preterm (<37 weeks) birth were higher in women with previous bariatric surgery at 10.5% (135) vs 7.8% (1004) in matched women. Neonates born to women with previous bariatric surgery had lower absolute birthweights (3223g ± 605g vs 3418g ± 595g; p<0.001), lower rates of large for gestational age (LGA) (8.6% vs 14.1%; p<0.001) and higher rates of small for gestational age (SGA) (10.7% vs 7.3%; p<0.001). Rates of gestational diabetes mellitus (GDM) and pregnancy induced hypertension were lower in women with previous bariatric surgery (15% vs 20%; p<0.001) and (4% vs 5%; p=0.01). Pre -existing diabetes, pre-existing hypertension, pre-eclampsia and admission to special care or neonatal intensive care nursery rates were not different between groups.

Conclusion: Our results suggest that pregnancy outcomes following maternal bariatric surgery differ from matched controls. There was lower proportions of GDM and LGA, but higher proportions of pre-term delivery, SGA and lower absolute birth weights in women with previous surgery to matched women. Despite BMI matching women with previous bariatric surgery had lower rates of GDM. Current screening for GDM following surgery is not evidence based, this may lead to fewer women being appropriately identified. Possible physiologic incretin changes following surgery may also contribute to altered glucose homeostasis.

Supported by: DARP

Disclosure: J. Eccles-Smith: Other; Research Training Program Tuition Fee Offset.

14

Altered postprandial glucose metabolism and enteropancreatic hormone response during pregnancy following Roux-en-Y gastric bypass: a prospective cohort study

L.L. Stentebjerg ¹, L.R. Madsen^2,3, R.K. Støving⁴, B. Hartmann⁵, J.J. Holst⁵, C.A. Vinter⁶, C.B. Juhl⁷, K. Højlund¹, D.M. Jensen^1,6;

¹Steno Diabetes Center Odense, Odense University Hospital, Odense, Denmark, ²Steno Diabetes Center Aarhus, Aarhus University Hospital, Aarhus, Denmark, ³Danish Diabetes and Endocrine Academy, Odense, Denmark, ⁴Department of Endocrinology, Odense University Hospital, Odense, Denmark, ⁵Novo Nordisk Center for Basic Metabolic Research, University of Copenhagen, Copenhagen, Denmark, ⁶Department of Gynecology and Obstetrics, Odense University Hospital, Odense, Denmark, ⁷Department of Endocrinology, Hospital of South West Jutland, Esbjerg, Denmark.

Background and aims: Roux-en-Y gastric bypass surgery (RYGB) increases the risk of postprandial hypoglycemia. In contrast, pregnancy increases insulin resistance, which could counteract hypoglycemia. We examined if RYGB performed prior to pregnancy altered the postprandial glucose metabolism and enteropancreatic hormone responses to a mixed meal test in pregnancy.

Materials and methods: Twenty-three pregnant women with RYGB and 23 pregnant women matched on BMI and parity underwent a 4-hour liquid mixed meal test (200 ml Nutridrink®, 300 kcal, 49 E% carbohydrate, 16E% protein, and 35 E% fat) in the first and third trimester with measurement of circulating levels of glucose, insulin, C-peptide, glucose-dependent insulin peptide (GIP), glucagon-like peptide 1 (GLP-1), and glucagon. Hypoglycemia was defined as plasma glucose < 3.5 mmol/L, and the updated HOMA2 insulin resistance index was calculated based on C-peptide.

Results: Women with RYGB had earlier and higher peak glucose (First trimester: 8.2 vs. 6.3 mmol/L, p<0.01. Third trimester: 8.3 vs. 6.6 mmol/L, p<0.01), lower nadir glucose (First trimester: 3.7 vs. 4.2 mmol/L, p<0.01. Third trimester: 3.9 vs. 4.2 mmol/L, p<0.01) and a higher frequency of hypoglycemia (First trimester: 39 vs. 4%, p=0.01. Third trimester: 24 vs. 0%, p=0.02) than women without RYGB. The lower nadir glucose levels were preceded by markedly elevated peak GLP-1 (First trimester: 104 vs. 30 pmol/L, p<0.01. Third trimester: 124 vs. 30 pmol/L, p<0.01) and peak insulin (First trimester: 1,337 vs. 603 pmol/L, p<0.01. Third trimester: 1,326 vs. 781 pmol/L, p<0.01) in women with RYGB, whereas GIP levels were unaltered. Fasting levels of serum insulin (First trimester: 40 vs. 97 pmol/L, p<0.01. Third trimester: 57 vs. 144 pmol/L, p<0.01) and C-peptide (First trimester: 590 vs. 788 pmol/L, p=0.01. Third trimester: 837 vs 1,268 pmol/L, p<0.01) increased in both groups from first to third trimester (all p<0.001). In accordance, HOMA2 insulin resistance increased from first to third trimester in both groups (First trimester: 1.3 vs. 1.8, p<0.01. Third trimester: 1.8 vs. 2.7, p<0.01), but was lower throughout pregnancy among the women with RYGB compared to controls (all p<0.01). Finally, compared to controls, AUC of glucagon were lower in women with RYGB (First trimester: 1,542 vs. 2,555 pmol/L·min, p<0.01. Third trimester: 1,694 vs. 2,123 pmol/L·min, p=0.05).

Conclusion: These results provide novel insights into the combined effects of RYGB and pregnancy on the post-meal glucose metabolism and enteropancreatic hormone responses during pregnancy, and how these changes associate with an increased risk of postprandial hypoglycemia. Despite an increase in insulin resistance during pregnancy, hyperinsulinemic hypoglycemia occurred in both first and third trimester of pregnancy.

Clinical Trial Registration Number: NCT03713060

Supported by: L.L. Stentebjerg: Grants; Region of Southern Denmark.

Disclosure: L.L. Stentebjerg: None.

15

Screening of gestational diabetes in early pregnancy by oral glucose tolerance test and biomarkers: a multi-centre cohort study

E.A. Huhn ^1,2, T. Fischer³, M. Todesco Bernasconi⁴, M. Kreft⁵, M. Kunze⁶, E. Dölzlmüller³, H. Jaksch-Bogensperger³, S. Heldstab⁴, N. Ochsenbein-Koelble⁵, A. Richter⁵, E. Bäz⁶, B. Winzeler⁷, A. Tura⁸, I. Hösli¹, C.S. Göbl⁹;

¹Department of Obstetrics and Gynaecology, University and University Hospital Basel, Basel, Switzerland, ²Clinic of Obstetrics and Prenatal Medicine, University Hospital Hamburg-Eppendorf, Hamburg, Germany, ³Department of Obstetrics and Gynaecology, Paracelsus Medical University, Salzburg, Austria, ⁴Department of Obstetrics and Gynaecology, Cantonal Hospital Aarau, Aarau, Switzerland, ⁵Department of Obstetrics and Gynaecology, University Hospital Zurich, Zürich, Switzerland, ⁶Department of Obstetrics and Gynaecology, University Hospital Freiburg, Freiburg, Germany, ⁷Department of Endocrinology and Diabetology, University and University Hospital Basel, Basel, Switzerland, ⁸CNR, Institute of Neuroscience, Padova, Italy, ⁹Department of Obstetrics and Gynaecology, Medical University of Vienna, Vienna, Austria.

Background and aims: According to most clinical recommendations Gestational Diabetes Mellitus (GDM) is diagnosed after 24 gestational weeks (GW). The identification of mothers with particularly high risk before 24 GW may be beneficial to improve pregnancy outcomes, however, there is no clear consensus about an accurate risk stratification at early pregnancy. Therefore, this study aims to evaluate the predictive performance of a 75 oral glucose tolerance test (OGTT) and several biomarkers at first trimester of pregnancy. Their association with insulin action and β-cell function were additionally assessed.

Materials and methods: In the setting of a prospective cohort study, we included 703 pregnant women in six middle European centres. Patient history as well as a blinded 75 g OGTT, and biochemical markers were assessed at a median gestational age of 13.4 GW (interquartile range: 12.7-14.1). A second 75g OGTT was performed between 24-28 GW to identify women with GDM according to the World Health Organisation criteria. A detailed investigation of glucose homeostasis was performed at both visits in a subgroup of 204 women (study participants included in Vienna).

Results: The later development of GDM occurred in a total of 87 women (12.4%) and was fairly predicted by glucose concentrations during the early OGTT (OGTT-Glucose 0 minutes: 70.3%, 95%CI: 63.8-76.8; OGTT-Glucose 60 minutes: 75.0%, 95%CI: 69.6-80.5; OGTT-Glucose 120 minutes: 72.9%, 95%CI: 66.5-79.2). In addition, some biomarkers showed significant but modest predictive accuracy as assessed by the area under the receiver operating characteristic curve (fructosamine: 56.9%; adiponectin: 61.2%; glycosylated fibronectin: 58.6%, triglycerides: 60.1%). A detailed assessment of the association with parameters of glucose homeostasis showed close association between early gestational OGTT glucose concentrations and dynamically assessed insulin resistance (as assessed by the composite Index), β-cell dysfunction (as assessed by the “oral” disposition index) as well as maternal HbA1c levels.

Conclusion: Although, recommendations for an accurate GDM diagnosis before 24 GW are missing dynamically assessed glucose concentrations during an early OGTT showed strong association with parameters of glucose homeostasis and a fair predictive accuracy for the later development for GDM. These can be used to distinguish between low and high-risk mothers.

Clinical Trial Registration Number: NCT02035059

Supported by: Swiss National Fund, Diabetes Society Basel, Gottfried and Julia Bangerter-Rhyner-Foundation

Disclosure: E.A. Huhn: None.

16

Implementation of a screening protocol for the detection of Maturity Onset Diabetes of the Young (MODY) in pregnancy: a retrospective preliminary analysis

R. Farrelly ¹, I. Blinkhorn¹, M. Rathod², E. Scott¹;

¹University of Leeds, Leeds, UK, ²Obstetrics Department, St James University Hospital, Leeds, UK.

Background and aims: Maturity onset diabetes of the young (MODY) is a monogenic, autosomal dominant condition, often misdiagnosed during pregnancy as gestational diabetes mellitus (GDM). Previous groups have suggested that a fasting glucose >5.5 mmol/l in women with a body mass index (BMI) <25 kg/m2 successfully identifies those who cases of GDM that should have MODY genetic testing. The aims of this service evaluation were to: 1) explore the impact of applying these screening criteria to an ethnically diverse population of women diagnosed with GDM to determine the numbers needing to be tested; 2) generate preliminary data on whether demographic or clinical characteristics differed between those women with GDM who fulfilled the criteria for MODY screening, compared to those who did not.

Materials and methods: GDM pregnancies diagnosed by NICE criteria on 75g oral glucose tolerance test (OGTT) at a single tertiary referral hospital between September 2021 and September 2022 were identified from electronic patient records held in a centralised database (PPM+ and K2). Only those patients with complete datasets available were included in the final analysis. Those suitable for MODY testing were identified within the GDM cohort if they had a BMI <25kg/m² at booking and a fasting glucose value of ≥5.5mmol/L at the time of 75g OGTT. Statistical analysis was performed using Microsoft Excel and GraphPad Prism. Differences and associations were considered statistically significant when p<0.05.

Results: A total of 678 GDM pregnancies with complete datasets were identified. Of these 17 (2.5%) were identified using the screening protocol to have potential MODY and thus, be eligible for genetic screening. There was no significant difference in maternal age between the two groups (p=0.3915) nor was there significant difference between the ethnicity of the two populations (p=0.8085). This supports that age and ethnicity should not be added to screening criteria for MODY in pregnancy.

Conclusion: Applying simple criteria, using BMI and fasting glucose, to our population with GDM, suggests that 2.5% of women should receive MODY genetic testing. This is comparable to that seen in previous screening implementation studies, despite our cohort being more ethnically diverse. Further observational studies implementing the protocol in real time with confirmatory genetic testing is now required.

Disclosure: R. Farrelly: None.

17

Maternal overweight and obesity is associated with metabolic changes and foetal overgrowth in the absence of gestational diabetes

G. Kotzaeridi ¹, C. Monod^1,2, T. Linder¹, G. Yerlikaya-Schatten¹, S. Wegener³, B. Mosimann², W. Henrich³, A. Tura⁴, C.S. Göbl¹;

¹Department of Obstetrics and Gynecology, Medical University of Vienna, Vienna, Austria, ²Department of Obstetrics and Gynecology, University Hospital Basel, Basel, Switzerland, ³Clinic of Obstetrics, Charité-Universitätsmedizin Berlin, Berlin, Germany, ⁴CNR Institute of Neuroscience, Padova, Italy.

Background and aims: Previous studies indicated an association between fetal overgrowth and maternal obesity independently of gestational diabetes mellitus (GDM). However, the underlying mechanisms beyond this possible association are not completely understood. Therefore, this study investigates fetal development and its association with subclinical metabolic alterations in overweight and obese mothers who remained normal glucose tolerant during gestation.

Materials and methods: In this prospective cohort study 893 women who did not develop GDM were categorized according to their pregestational BMI: 570 normal weight, 220 overweight, 103 obese. Study participants received a broad metabolic evaluation before 16 weeks and were followed up until delivery to assess glucose levels during the oral glucose tolerance test (OGTT) at mid gestation as well as fetal biometry in ultrasound and pregnancy outcome data.

Results: Increased maternal BMI was associated with an adverse metabolic profile at beginning of pregnancy including a higher degree of insulin resistance as well as elevated triglycerides. Despite not fulfilling diagnosis criteria for GDM, overweight and obese mothers showed higher glucose levels at fasting as well as during the OGTT. Finally, we observed increased measures of fetal subcutaneous tissue thickness in ultrasound as well as higher proportions of large for gestational age (LGA) infants in overweight (18.9%, p=0.021) and obese mothers (21.0%, p=0.027) vs. normal weight controls (11.8%). The risk for LGA development was further determined by OGTT glucose (60 min: p=0.013, 120 min: p=0.015) and maternal triglyceride levels (p=0.036).

Conclusion: Mothers affected by overweight or obesity but not GDM had higher risk for fetal overgrowth. An impaired metabolic milieu related to increased maternal BMI as well as higher glucose levels at mid gestation may impact fetal development in women still in the range of normal glucose tolerance.

Disclosure: G. Kotzaeridi: None.

18

Effect of an interdisciplinary lifestyle and psychosocial intervention on metabolic and mental health outcomes in women with gestational diabetes: the MySweetheart trial

J.J. Puder, D.Y. Quansah, L. Gilbert, A. Arhab, E. Gonzalez-Rodriguez, D. Hans, J. Gross, S. Lanzi, B. Stuijfzand, A. Lacroix, A. Horsch;

CHUV, Lausanne, Switzerland.

Background and aims: Gestational diabetes (GDM) increases the risk of depression, diabetes and cardiovascular disease. We tested the effect of a pre- and postpartum complex interdisciplinary lifestyle and psychosocial intervention on metabolic and mental health outcomes in women with GDM.

Materials and methods: This single blinded randomized controlled trial was conducted at the Lausanne Hospital between September/2016 and October/2021. 211 pregnant women with GDM were randomly assigned (1:1) to the intervention (n=105) or a guidelines-based usual-care (n=106). In addition to usual-care, the intervention consisted of four interdisciplinary lifestyle visits during pregnancy and four in the postpartum, two peer-support workshops, and a bimonthly telemedicine coach support. It focused on tailored behavioral and psychosocial strategies to improve diet, physical activity, mental health, social support, adherence to gestational weight gain (GWG) and weight retention recommendations. Primary outcomes were between-group differences in the decrease in weight and attenuation of depression symptom scores between inclusion and 1-year postpartum. Assessors were blinded to outcomes.

Results: 84 women in the intervention and 95 in the usual-care completed the study. There were no differences in the decrease in weight (β=-0·40 [95% CI:-2·1, 1·2]) or depression scores (β=-0·95 [95% CI:-2·46, 0·54]). Outcomes of body composition, glycemic parameters, aerobic and muscular fitness did not differ. The intervention led to an increase in fat-free mass at 1-year postpartum (β=0·02 [95% CI:-0·05, -0·02]) after covariate adjustment. It decreased insulin use, GWG (β=-1·20 [95% CI: -2·14, -0·26]), and the rate of weekly weight gain (β=-0·14 [95% CI: -0·25, -0·03]) and led to a higher proportion of women without weight retention at 1-year postpartum (34·1% vs 20·8%, p=0·034).

Conclusion: Compared to an active guidelines-based usual-care, the intervention did not lead to additional decreases in weight or attenuation of depression symptoms at 1-year postpartum. However, it decreased insulin use, GWG during pregnancy and increased the proportion of women without weight retention.

Clinical Trial Registration Number: NCT02890693

Supported by: SNSF. Gottfried und Julia Bangerter-Rhyner-Foundation

Disclosure: J.J. Puder: Grants; Swiss National Science Foundation, Gottfried und Julia Bangerter- Rhyner-Foundation.

OP 04 Diet and exposome

19

Exploring the use of metabolomics to evaluate compliance to a mediterranean diet intervention for cardiometabolic health

E. Smith ¹, U. Ericson¹, S. Hellstrand¹, M. Orho-Melander¹, P.M. Nilsson^1,2, C. Fernandez¹, P. Antonini³, S. Di Somma^4,3, F. Ottosson^5,1, O. Melander^1,2;

¹Department of Clinical Sciences, Lund University, Malmö, Sweden, ²Department of Emergency and Internal Medicine, Skåne University Hospital, Malmö, Sweden, ³GREAT Health Sciences, Rome, Italy, ⁴Department of Medical-Surgery Sciences and Translational Medicine, University of Rome Sapienza, Rome, Italy, ⁵Department of Congenital Disorders, Statens Serum Institut, Copenhagen, Denmark.

Background and aims: Healthy dietary intake is one of the main modifiable risk factors of development of type 2 diabetes and cardiovascular disease. Metabolomics has shown promise in estimating the quality of dietary intake by the measurement of plasma metabolites. We have previously identified a healthy dietary metabolic signature (HDMS), which was inversely associated with future type 2 diabetes in two separate prospective cohort studies. Our aim was to test if dietary metabolites and the HDMS were altered by a 7-day dietary intervention.

Materials and methods: We included 59 Swedish individuals in a 7-day Mediterranean diet (MD) intervention study conducted in the Cilento region in Italy. Participants were served food prepared with local ingredients and recipes for all meals. Profiling of pre- and post-interventional plasma metabolites was performed using an UHPLC-QTOF. Changes in metabolome was investigated with paired T-test adjusted for multiple testing using false discovery rate (FDR).

Results: The intervention resulted in an improvement in the diabetes-predictive HDMS in all 59 participants, with a large average effect size (mean increase 1.3 standard deviations (SD), 95% confidence interval 1.1-1.4, p = 6E-25) (figure 1). Out of 109 individual metabolites measured, 66 were significantly altered (FDR adjusted p < 0.05). The most significant increased levels were seen in pipecolate (1.6 SD, 95% CI 1.5-1.8, p = 7E-28), hippurate (1.3 SD, CI 1.1-1.5, p = 2E-18), caffeine (1.2 SD, CI 1.0-1.4, p = 7E-14), homostachydrine (1.1 SD, CI 0.9-1.3, p = 2E-13), acylcarnitine C11:0 (1.1 SD, CI 0.9-1.4, p = 9E-12), beta carotene (0.9 SD, CI 0.6-1.1, p = 5E-10) and 7-methylguanine (0.7 SD, CI 0.5-0.9, p = 7E-10). The most significant decreased levels were seen in piperine (-1.4 SD, 95% CI -1.4- -1.2, p = 2E-18), 3-methylhistidine (-1.4 SD, CI -1.7 - -1.1, p = 1E-12) and creatine (-0.7 SD, CI -0.9 - -0.5, p = 3E-09).

Conclusion: The HDMS, which we have previously found to associate with decreased incidence of type 2 diabetes was profoundly ameliorated by a 7-day MD. Interestingly, several of the individual metabolites known as markers of vegetables, fruits and berries, whole grain and legumes increased, whereas markers of meat intake decreased after the one-week intervention. Our study suggests that monitoring of the HDMS and several of its individual metabolites, which are markers of intake of specific food groups, can be a valuable tool in evaluating the effectiveness of dietary interventions and promoting healthy dietary adherence.

Supported by: VR, HLF

Disclosure: E. Smith: None.

20

Diet in the management of type 2 diabetes: an umbrella review of the evidence of randomised controlled trials

E. Szczerba ^1,2, J. Barbaresko¹, T. Schiemann¹, A. Stahl-Pehe^1,2, L. Schwingshackl³, S. Schlesinger^1,2;

¹Institute for Biometrics and Epidemiology, German Diabetes Center, Düsseldorf, Germany, ²German Center for Diabetes Research (DZD), Partner Düsseldorf, Munich-Neuherberg, Germany, ³Institute for Evidence in Medicine, Faculty of Medicine and Medical Center- University of Freiburg, Freiburg, Germany.

Background and aims: Diet is a cornerstone in the management of type 2 diabetes (T2D). However, it is still unclear which dietary factors are truly effective in managing cardiometabolic risk factors and reducing the risk of diabetes complications. Therefore, the aim of our study was to conduct an umbrella review of systematic reviews with meta-analyses of randomised controlled trials (RCTs) on any dietary factors and health outcomes in persons with T2D.

Materials and methods: We performed a systematic literature search in four online databases up to June 2022 by two researchers independently. We included systematic reviews with meta-analyses of RCTs (intervention time ≥12 weeks) reporting summary effect estimates on the impact of dietary factors (dietary patterns, food groups, nutrients, dietary supplements) regarding any health outcome in persons with T2D. Summary data were extracted by two investigators independently. We recalculated the summary effect estimates with 95% confidence intervals (CIs) using a random-effects model if the original publication provided insufficient information. Methodological quality was evaluated with the AMSTAR-2 tool and certainty of evidence (CoE) with the GRADEpro approach.

Results: Overall, we included 88 publications with 312 meta-analyses. Methodological quality was high to moderate in 22% and low to very low in 78% of the included publications. CoE was high for 2% of the meta-analyses, moderate for 13%, and low or very low for 85%. We found high CoE for the effects of liquid meal replacement on BMI (mean difference (MD): -0.87 kg/m²; 95%CI: -1.32, -0.43; n=8 primary studies) and body weight reduction (MD: -2.37 kg; 95%CI: -3.30, -1.44; n=9) as well as for low-carbohydrate diets (<26% of total energy) on HbA1c (MD: -0.47%; 95%CI: -0.60, -0.34; n=17) and triglycerides levels (MD: -0.30 mmol/L; 95%CI: -0.43, -0.17; n=19). Moderate CoE changes was found for plant-based diets and reduced anthropometric measures (BMI: -1.13 kg/m^2; 95%CI -1.88; -0.38; n=3, waist circumference: -2.41cm; 95%CI -3.50, 1.32; n=2), for liquid meal replacement diets, fibre, ginger, anthocyanins or probiotics with improved glycaemic measures (fasting blood glucose reduction of -63.0 to -1.78 mmol/l, HOMA-IR reduction of -0.82 to 1.08 units), and for high-protein diets, soya or fibre on improved blood lipids (triglycerides reduction of -0.13 to -41.0 mmol/l, total cholesterol reduction of -0.10 to -0.24 mmol/l, low density cholesterol reduction of -0.19 to -0.26 mmol/l).

Conclusion: Evidence shows that diet plays a multifaceted role in the management of T2D. An energy-restricted diet can be recommended for persons with T2D to reduce body weight and improve cardiometabolic markers. Beyond energy restriction, dietary approaches such as a plant-based, Mediterranean, low carbohydrate and high-protein diet can also be beneficial for cardiometabolic health for persons with T2D. To provide evidence-based recommendations for the effectiveness of other dietary factors on T2D management, more future primary studies as well as systematic reviews with meta-analyses are needed.

Disclosure: E. Szczerba: None.

21

The associations between metabolomic responses to a liquid mixed meal and incidence of type 2 diabetes: Are there novel biomarkers beyond glucose and triglyceride response?

R. Li-Gao ¹, C. Lu², D.A. Hughes³, J.B. van Klinken⁴, R. de Mutsert¹, F.R. Rosendaal¹, N.J. Timpson³, D.O. Mook-Kanamori¹, J.J. Goeman², H. Putter², K. Willems van Dijk⁴;

¹Department of Clinical Epidemiology, Leiden University Medical Center, Leiden, Netherlands, ²Department of Medical Statistics and Bioinformatics, Leiden University Medical Center, Leiden, Netherlands, ³MRC Integrative Epidemiology Unit, School of Social and Community Medicine, University of Bristol, Bristol, UK, ⁴Department of Human Genetics, Leiden University Medical Center, Leiden, Netherlands.

Background and aims: Disturbed postprandial plasma metabolite levels and responses (such as glucose and triglyceride) have been suggested as risk factors for type 2 diabetes (T2D). However, this remains to be fully investigated. We aim to leverage the Netherlands Epidemiology of Obesity (NEO) study to systematically examine the associations between metabolomic responses and incidence of T2D.

Materials and methods: We analyzed 5876 participants with a mean (SD) age of 56 (6) years, body mass index (BMI) of 29.8 (4.7) kg/m², 53% women, without T2D at baseline. In fasting and postprandial (t=150 min after a liquid mixed meal; 400 ml with 2.5MJ) plasma samples, 148 metabolites were measured using the 1H NMR platform. The response of a metabolite was defined as the difference between postprandial and fasting levels, which was standardized for a cross-metabolite comparison. A Cox proportional-hazard model was used to estimate hazard ratios (HR) of T2D, adjusted for confounders including age, sex, BMI and glucose levels. False discovery rate was applied in each meal state respectively.

Results: Over a median follow-up of 6.7 years, 302 new cases of T2D were detected. Glucose response showed the strongest association to T2D risk, with one SD change of glucose response leading to 80% (95% CI: 61%-101%) increase of disease risk. After adjusting for glucose levels, the responses of alanine and histidine showed the associations to T2D, with one SD change of alanine and histidine response leading to 26% (95% CI: 15%-35%) and 22% (95% CI: 19%-31%) decrease of T2D risk respectively. In addition, one SD change of responses of MVLDL, LVLDL, XLVLDL and XXLVLDL showed a higher risk of T2D (17%-24%).

Conclusion: In this middle-aged population, certain metabolite responses were associated with incidence of T2D, independent of glucose levels, which suggests the clinical utility of postprandial measurements in predicting T2D progression.

Clinical Trial Registration Number: NCT03410316

Supported by: This project is supported by "Innovation for Data Analytics" grant from Leiden University Medical Center, the Netherlands

Disclosure: R. Li-Gao: Employment/Consultancy; Ruifang Li-Gao is a part-time contractor for Metabolon Inc.

22

A randomised controlled trial of intermittently-scanned continuous glucose monitoring as an adjunct to lifestyle modification in impaired glucose tolerance

E. Chow, N. Chu, J. He, H. Leung, C. So, E. Poon, R. Chan, A. Kong, J. Chan;

The Chinese University of Hong Kong, Hong Kong, Hong Kong.

Background and aims: Continuous glucose monitoring (CGM) has been shown to promote behavioural and lifestyle changes in people with diabetes. Benefits of CGM in people with intermediate hyperglycemia remain unknown. In this study, we evaluated the effect of intermittently scanned continuous glucose monitoring (isCGM) as an adjunct to a lifestyle modification program (LMP) in individuals with impaired glucose tolerance (IGT).

Materials and methods: This was a 12-month prospective, open-label, non-masked, randomized controlled trial where participants with IGT were allocated 1:1 to isCGM as an adjunct to LMP (intervention group) versus LMP alone (control). Participants were aged 18-65 years, body mass index (BMI) between 18-40 kg/m² with 2-hour plasma glucose (2hPG) between 7.8-11.1 mmol/L on 75g oral glucose tolerance test (OGTT). Both groups participated in a 12-month structured LMP modelled on the Diabetes Prevention Program with weekly dietary consultations in the first 4-month intensive phase and monthly in the 8-month maintenance phase. The intervention group additionally wore an isCGM two-weeks on, two-weeks off in the first 4 months only. Fasting plasma glucose, 1h and 2hPG at OGTT, HbA1c, body weight, 3-day dietary records, physical activity were assessed at 4-monthly intervals. The primary outcome was improvement in glucose tolerance by intention-to-treat analysis.

Results: We included 177 participants (mean±SD age 57±7 years, BMI 26.7±4.0 kg/m², 59.3% female, 1hPG 10.9±1.6 mmol/L, 2hPG 8.4±1.4 mmol/L) who were randomized to CGM+LMP (n=89) or LMP (n=88) arms. Average sensor active time was 86±12% with mean 9.5±5.0 scans per day in the CGM group. At month 4, 1hPG was 9.7±2.1 versus 10.1±1.19 mmol/L in the intervention versus control group (-1.2±1.8 vs -0.7±1.8 mmol/L, treatment difference -0.46 mmol/L, 95% confidence interval -1.04, 0.13). Reductions in daily calorie intake were similar, but the CGM+LMP group had a greater reduction in total sugar intake. In per-protocol analysis (including those with >70% sensor use and/or attended >70% of consultations), the CGM+LMP group showed a greater reduction in 1hPG (-1.3±1.8 vs -0.7±1.7 mmol/L, p=0.05) and % body weight (-5.1±3.5 vs -4.2±3.0 %) at month 4. Higher sensor active time correlated with greater percentage weight loss in the CGM+LMP group (r= -0.364, p=0.002).

Conclusion: isCGM as an adjunct to LMP may facilitate early improvements in glucose tolerance in people with IGT. The ongoing trial will address whether these improvements are sustained at month 12 and beyond.

Clinical Trial Registration Number: NCT04588896

Supported by: HMRF Investigator Initiated Research Grant

Disclosure: E. Chow: None.

23

Enteroviral infections are not associated with type 2 diabetes

K. Maedler, H. Liu, S. Geravandi, A. Pugliese;

University of Bremen, Bremen, Germany.

Background and aims: For more than a century, enteroviral infections have been associated with autoimmunity and Type 1 Diabetes (T1D). The increased presence of enteroviral RNA in the pancreas from organ donors with T1D has just been confirmed in large meta-analyses. Virus responses evoke chemokines and cytokines, the “cytokine storm” circulating through the body and attack cells especially vulnerable to inflammatory destruction. Uncontrolled viral response pathways repeatedly presented during childhood highly correlate with autoimmunity and T1D. Whole intra-islet inflammation is a major trigger of beta-cell failure in both T1D and T2D. The genetic contribution of islet inflammation pathways is apparent in T1D, with several mutations in the interferon system. In contrast, in T2D, gene mutations are related to glucose homeostasis in beta-cells and insulin target tissue and rarely within viral response pathways. Therefore, the current study evaluated whether enteroviral RNA can be found in the pancreas from organ donors with T2D and its association with disease progression.

Materials and methods: Pancreases from well-characterized 29 organ donors with T2D and 16 age- and BMI matched controls were obtained from the network for Pancreatic Organ Donors with diabetes (nPOD) and were analyzed in duplicates. Single molecule FISH analyses were performed using three probe sets to detect positive strand enteroviral RNA; pancreas sections were co-stained by classical immunostaining for insulin and glucagon.

Results: There was increased islet lymphocytic infiltration and decreased beta-cell area in T2D pancreases, but no difference in the presence or localization of enteroviral RNA in control nondiabetic and T2D pancreases; viral infiltration showed large heterogeneity in both groups ranging from 0-94 virus⁺ cells scattered throughout the pancreas, most of them in the exocrine pancreas. Very rarely, a single virus⁺ cell was found within islets or co-stained with CD45⁺ immune cells. Only one single T2D donor presented an exceptionally high number of viruses, similarly as seen previously in T1D, which correlated with a highly reduced number of beta-cells.

Conclusion: Based on the analysis of enteroviral RNA by smFISH of currently available organ donors from the nPOD cohort, no association of enteroviral infection and T2D diabetes could be found. Despite great similarities in inflammatory markers in islets in T1D and T2D, long-term enteroviral infiltration is a distinct pathological feature of T1D-associated autoimmunity and in T1D pancreases.

Supported by: JDRF, DFG

Disclosure: K. Maedler: None.

24

Low dose radiation exposure and the risk of diabetes: a 10-year retrospective cohort study

H. Hu, T. Okubo;

National Institute of Occupational Safety and Health, Kawasaki, Japan.

Background and aims: Limited evidence exists on the impact of radiation exposure on diabetes risk. We examined the association between low-dose radiation exposure and risk of diabetes in adults who performed emergency work at the Fukushima Dai-ichi nuclear plant.

Materials and methods: This is a retrospective analysis of baseline data from the Epidemiological Study of Health Effects in Fukushima Nuclear Emergency Workers (NEWS). A total of 5,326 participants from the NEWS study were included. Diabetes was defined as a fast plasma glucose level of at least 126 mg/dL, an HbA1c level of at least 6.5%, or self-reported diagnosis of diabetes. Incident diabetes that occurred between 2012 and 2021 is identified if the age at onset is older than that in 2011. Radiation dose received at emergency work (2011/03-2011/12) was measured using the individual pocket alarm dosemeter for external exposure and the whole-body counter for internal exposure. The Cox proportional hazards regression analysis was used to assess the association between incident diabetes and cumulative radiation exposure, with adjustment for a wide range of covariates.

Results: Between 2012 and 2021, 392 participants developed diabetes. After adjusting for age, body mass index, smoking, alcohol consumption, leisure-time physical activity, employment history, dyslipidemia, and hypertension, the hazard ratios (95% confidence intervals) for diabetes were 1.06 (0.78, 1.45), 1.47 (1.12, 1.94), 1.33 (1.00, 1.77), and 0.98 (0.62, 1.56) for the 5-9 mSv, 10-19 mSv, 20-49 mSv, and ≥50 mSv groups, respectively, compared with the 0-4 mSv group (P for tend = 0.06).

Conclusion: We observed an increased risk of diabetes among nuclear emergency workers. A continuation of the NEWS study will result in an even clearer picture of the diabetes risk at low doses.

Supported by: Industrial Disease Clinical Research Grants

Disclosure: H. Hu: None.

OP 05 Walking towards better treatments for diabetic feet and neuropathy

25

Efficacy of autologous cell therapy in the treatment of chronic limb-threatening ischaemia in people with diabetes: a randomised controlled trial

M. Dubsky, J. Husakova, D. Sojakova, V. Fejfarova, R. Jarosikova, A. Nemcova, V. Lovasova, K. Sutoris, V. Woskova;

Institute for Clinical and Experimental Medicine, Prague, Czech Republic.

Background and aims: There are still many patients with the most severe stages of chronic limb-threatening ischemia (CLTI) who are not suitable for any kind of standard revascularization. Autologous cell therapy (ACT) is a novel treatment approach for patients with diabetes and no-option CLTI (NO-CLTI). The aim of our study was to analyze the impact of ACT on NO-CLTI in comparison with standard conservative treatment (ST) in a randomized controlled trial.

Materials and methods: Diabetic patients with NO-CLTI treated in our centre over 6 years were randomized to receive either ACT (n=28) or ST (n=27). After 12 weeks, those in the ST group, who did not improve were treated with ACT. The effect of ACT on ischemia and wound healing was assessed by changes in transcutaneous oxygen pressure (TcPO₂) and the number of healed patients at 12 weeks. Pain was evaluated by Visual Analogue Scale (VAS). Amputation rates and amputation-free survival (AFS) were assessed in both groups.

Results: During the first 12 weeks, TcPO₂ increased in the ACT group from 19.6±9.1 to 40.2±17.6 mm Hg (p=0.004) whereas there was no change in the ST group (from 20.4±10.9 to 23.1±13.7 mm Hg). Difference in TcPO₂ in the ACT group compared to ST group was 17.9 mm Hg (p=0.038) after 12 weeks. In the period from week 12 to week 24, when patients from ST group received ACT, the TcPO₂ in this group increased from 20.4±10.9 to 39.7±13.6 (p=0.006) while it did not change significantly in the ACT in this period. At 24 weeks, there was no significant difference in mean TcPO₂ between the two groups. Wound healing was greater at 12 weeks in the ACT group in non-amputated patients compared to the ST group (8/23 vs. 1/19, p=0.008). Pain measured using VAS was significantly reduced in the ACT group after 12 weeks compared to ST group (p<0.001). There was no difference in rates of major amputation and AFS between both groups at 12 weeks.

Conclusion: This study showed that ACT in patients with no-option CLTI and diabetic foot significantly improved limb ischemia and wound healing after 12 weeks compared to conservative standard therapy. Larger randomized controlled trials are needed to confirm the benefits of ACT in patients with NO-CLTI and diabetic foot disease.

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

Disclosure: M. Dubsky: None.

26

Effect of macrophage-regulating new drug in treatment of diabetic foot ulcers: from phase 3 MRCT to real-world

K. Lin ¹, C. Chen¹, Y. Tarng¹, M. Kuo², J. Chen², Y. Wu², C. Li³, S. Chang⁴;

¹Department of Orthopaedic surgery, Kaohsiung Veterans General Hospital, Kaohsiung, Taiwan, ²Oneness Biotech Co., Ltd, Taipei, Taiwan, ³Chiron Clinic, Pingtung, Taiwan, ⁴Division of Plastic Surgery, Department of Surgery, Integrated Burn and Wound Care Center, Taipei Medical University-Shuang Ho Hospital, Taipei, Taiwan.

Background and aims: Macrophage regulation plays an essential role in Diabetic foot ulcer (DFU) healing. By recognizing the therapeutic potential in M1/M2-macrophage-regulation for DFUs, we aim to investigate the efficacy and safety of ON101, a topical new drug with a M1/M2-macrophage-regulating mechanism in phase 3 MRCT across the US, China, and Taiwan.

Materials and methods: There are 236 eligible subjects with Wagner Grade 1 or 2 DFU(s), randomized 1:1 to receive ON101 or comparator for up to 16 weeks, followed by a 12-week follow-up.

Results: The incidence of complete healing was 60.7% by ON101 and 35.1% by the comparator (P=0.0001). The time to complete ulcer healing was faster in the ON101 group (P=0.001). Since the high levels of HbA1C, ulcer persistence, ulcer location, and amputation history as the risk factors leading to poor DFU outcomes, ON101 also showed demonstrated robust efficacy and safety in hard-to-heal ulcers.

Conclusion: ON101, approved by Taiwan FDA in 2021 and by Singapore in 2023, was shown with excellent therapeutic efficacy and safety in the treatment of DFUs. In a real-world scenario, ON101 further supports its excellent ability to promote wound healing in DFU patients with various severities, including Wagner grade 3 and 4, in the combination with other therapies. Furthermore, based on this novel mechanism, it reveals that ON101 promotes traumatic wounds healing and scar maturation/remodelling. This enlightens the future treatment in Wagner grade 1 to 4 DFUs by macrophage immunomodulation new drug as a promising solution.

Clinical Trial Registration Number: NCT01898923

Supported by: Oneness Biotech Co., Ltd.

Disclosure: K. Lin: Lecture/other fees; submission fees, Instructor fees.

27

Demographics and baseline characteristics of patients with Charcot foot: a multi-centre cohort analysis

N. Tentolouris ¹, C. Siafarikas¹, M. Dubský², J. LaFontaine³, K. Marie-France⁴, T. Tankova⁵, J. Pappachan⁶, A. Rastogi⁷, E. Jude⁸;

¹First Department of Propaedeutic Internal Medicine, Medical School, National and Kapodistrian University of Athens, Laiko General Hospital, Athens, Greece, ²Institute for Clinical and Experimental Medicine, Prague, Czech Republic, ³Department of Podiatry, UTRGV School of Podiatric Medicine, Texas, TX, USA, ⁴Diabetes and Endocrinology, Leicester General Hospital, Leicester, UK, ⁵Department of Endocrinology, Medical University of Sofia, Sofia, Bulgaria, ⁶Diabetes and Endocrinology, Lancashire Teaching Hospitals NHS Foundation Trust, Preston Road, Chorley, UK, ⁷Diabetes and Endocrinology, Post Graduate Institute of Medical Education and Research, Chandigarh, India, ⁸Diabetes and Endocrinology, Tameside and Glossop Integrated Care NHS Foundation Trust, Ashton under Lyne, UK.

Background and aims: Charcot foot (CF) is a rare but severe complication of diabetes mellitus (DM), that progressively leads to bone deformities of the lower extremities. The aim of this study was to record the baseline characteristics of patients with CF.

Materials and methods: A multicenter retrospective cohort study, including all patients diagnosed with CF from eight Diabetic Foot Clinics in 6 countries between 01/01/1996 and 31/12/2022, was performed. Data regarding duration and complications of DM, anthropometric, clinical and laboratory parameters were obtained from the medical records.

Results: A total of 774 patients were included. The mean age of patients was 56.5 ±10.7 years and the mean age at CF diagnosis was 54.5 ± 11.7 years, while 70.5% were men. Type 2 DM was more prevalent (83.2%), the mean duration of DM at diagnosis was 17.3 years and the median glycated hemoglobin was 8% (Interquartile range-IQ 6.8-9.3%) at diagnosis of CF. Regarding anthropometric features, mean height, weight and body mass index (BMI) were 169.40 (±16.5) cm, 84.4 (±21.0) kg, 28.97 (±7.8) kg/m², respectively. Most of the patients had hypertension (74.5%), a quarter of them were active or ex-smokers (25.2%) or alcohol users (23.3) and the minority of them had previous medical history of peripheral artery disease (19.7%), coronary artery disease (17.5%) or stroke (5.4%). Neuropathy was the most prevalent (89%) amongst the microvascular complications followed by retinopathy (60%) and nephropathy (45%). A foot ulcer was present in about half of the patients (48.5%) while CF was equally distributed between feet (right foot 49%-left foot 44%-bilateral 6.5%) and the majority of them were in midfoot (65%), forefoot 18%, hindfoot 13.9% and multiple joint involvement (2.6%).Patients with type 1 in comparison with those with type 2 DM were diagnosed with CF at younger age (46.9 ± 11.0 vs 57.9 ± 10.2 years, p<0.001), had lower BMI (26.6 ± 6.0 vs 29.5 ± 8.0 kg/m², p=0.001) and longer diabetes duration [median (IQ): 29.0 (21.0-38.0) vs 14.0 (8.0-20.0) years, p<0.001]. No differences were found between type 1 and type 2 DM in HbA1c. No significant gender differences were found in terms of age at CF diagnosis, BMI or HbA1c; however, the duration of DM at CF diagnosis was lower in men vs women 15.0 (1.0-22.0) vs 17.0 (10.0-25.0) years, p=0.037].

Conclusion: This is the largest cohort of patients with CF globally. CF is predominant in males, occurs in people with long-standing DM, is often accompanied by microvascular complications, half of the patients have foot ulcers at CF diagnosis, affects equally both feet, and the more frequent location is in the midfoot. Further studies are needed to look at outcomes and mortality.

Disclosure: N. Tentolouris: None.

28

Anti-RANKL antibody for active charcot neuroarthropathy of foot in chronic kidney disease: randomised, placebo controlled study

A. Rastogi, R. Dangwal, S.K. Bhadada;

Endocrinology, Post Graduate Institute of medical Education and Research, Chandigarh, India.

Background and aims: Charcot neuroarthropathy (CN) is characterised with increased osteoclastic activity that can be curbed with anti-resorptive agents. Chronic kidney disease (CKD) precludes the use of bisphosphonates for acute CN, but anti-RANKL antibody, Danosumab can be contemplated in CKD. We investigated Danosumab for CN remission in active diabetic CN of foot in people with CKD.

Materials and methods: Patients of diabetes and CKD presenting with unilateral, active CN of foot were randomized (1:2) to receive 60 mg Danosumab (sub-cutaneous) once at diagnosis (group A) or equal volume of normal saline (sub-cutaneous) (group B) in addition to standard if care as total contact cast (TCC). Participants and outcome assessor were blinded to treatment arms. Active CN was considered in the presence of swelling, erythema and temperature difference of >2⁰C [infrared dermal thermometry (FLIR Systems Inc, Orlando, USA) with a pixel resolution of 4800 (80*60), thermal sensitivity of <0.15°C and range of detection -20°C to 250°C] between two foot corroborated by MRI, if required. CKD was considered if eGFR<60 ml/min/1.73m² (CKD-EPI) and/or urine protein>300mg/24 hour. Details pertaining to demographic characteristics, foot presentation and diabetic complications were recorded. Anatomical region of interest of CN was classified according to Sanders and Frykberg (SF) classification. Patients were followed 2 weekly for 8 weeks, subsequently 4 weekly until remission and 2 monthly for further 6 months for recurrence of CN. Remission was defined as absence of temperature difference (<2⁰C) between two feet obtained 2 weeks apart with clinical resolution of signs of inflammation. Primary outcome measure was proportion of patients achieving remission and time to remission during follow up.

Results: During the study period, 246 persons with CN were identified, 102 of whom met the criteria for the first screening, and 78 were finally enrolled (n=26, group A, n=52 group B) and followed for a median of 48 weeks. Median age was 56.5 (48.8-65) and 57 (48.5-61.2) years, p=0.57; duration of diabetes [16(10-25.3) and 14.9 (10-19 years), p=0.151; HbA1c 8.1% (6.8-9.3%) and 8.5% (7.4-9.9%), p=0.294; eGFR 44.8 ml/min/1.73m² (21.1-65.6) and 45.7 ml/min/1.73m² (32.9-55.7) (p=0.771); 24 hour urine protein 557.9 mg (264.2-1112.1) and 343.6 (105.1-902.3) in group A and B respectively. Median temperature difference at presentation between affected and opposite foot was 3.4⁰C (2.7-6.9) and 3.2⁰C (2.2-4.0), respectively. A total of 56% had SF 2 and 3, 32% had SF 4, 8% had SF 1 and 4% had SF 5 localization of CN. All 26 (100%) patients achieved remission in group A compared to 42 (80.8%) patients in group B (p=0.026), with median time to remission of 15 (11-25) and 17.5 (14-31.5) weeks (p=0.229), respectively. A higher baseline haemoglobin (10.2 gm/dl versus 8.4 gm/dl) and 25 (OH)D3 (16 ng/mL versus 7.8 ng/ml) were associated with remission of active CN. Overall 23.1% and 32.7% patients died during follow up (p=0.380), in group A and B respectively considering background CKD.

Conclusion: Anti-RANKL antibody Danosumab added to standard of care (total contact cast) induces remission in greater proportion of patients of acute CN of foot and chronic kidney disease.

Disclosure: A. Rastogi: None.

29

LX9211 in individuals with painful diabetic peripheral neuropathy: results from a randomised, double-blind, placebo-controlled, parallel-group, multicentre study

R. Pop-Busui ¹, A. Patel², C. Sang³, C. Granowitz⁴, P. Banks⁴, P. Pierce⁴, F. Sun⁴, S. Gopinathan⁴;

¹Endocrinology & Diabetes, University of Michigan, Ann Arbor, MI, USA, ²Conquest Research, Winter Park, FL, USA, ³Harvard Medical School, University of Michigan, Boston, MA, USA, ⁴Lexicon Pharmaceuticals, Inc., The Woodlands, TX, USA.

Background and aims: Painful diabetic peripheral neuropathy (DPN) is a debilitating condition affecting millions of people. The currently available therapies often provide inadequate pain relief in most individuals with painful DPN.LX9211 is a potent, novel, orally administered inhibitor of adaptor-associated protein kinase 1 (AAK1), a non-opioid target for neuropathic pain. The aim of this study was to demonstrate the efficacy and safety of LX9211 in people with painful DPN.

Materials and methods: RELIEF-DPN 1 was a randomized, double-blind, placebo-controlled, parallel-group, multicenter clinical trial in which participants with painful DPN were randomized to receive placebo, LX9211 10 mg (with a 100 mg loading dose on Day 1), or LX9211 20 mg (with a 200 mg loading dose on Day 1), once daily for 6 weeks. The double-blind period was followed by a 5-week blinded safety follow-up period during which subjects received placebo once daily. The primary endpoint was the difference in change in average daily pain score (ADPS) between LX9211 and placebo at 6 weeks, as evaluated by mixed model repeated measures analysis.

Results: Among 319 adults randomized, baseline mean (SD) age was 62 (10) years, 41% were women, 76% White, 18% Black, mean HbA1C 7.7 (1.28) %, 61(14) mmol/mol, BMI 32.07 (4.45) kg/m2, and average daily pain score (ADPS) after placebo run-in period 6.55 (1.089). At Week 6, LX9211 was associated with achieving the primary endpoint (change from placebo in the ADPS) in the low dose arm (Low Dose: -1.39, p=0.007), and with significant improvement in: Neuropathic Pain Symptom Inventory (NPSI) total score (Low Dose: -5.18, p=0.064; High Dose: -7.22, p=0.008), burning pain (Low Dose: -1.40; p<0.001; High Dose: -0.89, p=0.017), pain interference on sleep (Low Dose: -0.96, p=0.005; High Dose: -1.04, p=0.002) and Patient Global Impression of Change score (Low Dose: -0.35, p=0.031; High Dose: -0.15, p=0.351) compared to placebo. LX9211 was safe (no treatment-related SAEs or deaths) and did not affect key cardiometabolic parameters (body weight, blood pressure, glucose control, cholesterol levels).

Conclusion: RELIEF-DPN 1 demonstrated significant improvements in validated neuropathic pain instruments and in other patient reported outcomes suggesting that LX9211 is a promising, opioid independent new avenue to treat people with painful DPN. Confirmation of these positive results are planned in future Phase 3 trials

Clinical Trial Registration Number: NCT04455633

Supported by: Lexicon Pharmaceuticals, Inc

Disclosure: R. Pop-Busui: Grants; Lexicon Pharmaceuticals.

30

The impact of obstructive sleep apnoea treatment on peripheral neuropathy in patients with type 2 diabetes: results from a 2-year feasibility rct

A.A. Tahrani ^1,2, E. Makhdom¹, A. Maher³, R. Ottridge³, M. Nicholls¹, R. Mahto⁴, M. Patel⁵, J. Young⁶, A.U. Nayak⁷, M.Z. Chen⁸, J. Kyaw-Tun⁹, S. Gonzalez¹⁰, R. Gouni¹¹, A. Subramanian¹², S. Patel³;

¹Institute of Metabolism and System Research, University of Birmingham, Birmingham, UK, ²Centre for endocrinology, diabetes and metabolism, Birmingham health partners, Birmingham, UK, ³Birmingham Clinical Trials Unit, University of Birmingham, Birmingham, UK, ⁴South Warwickshire NHS Foundation Trust, Warwickshire, UK, ⁵University Hospital Southampton NHS FT, Southampton, UK, ⁶Royal Wolverhampton hospitals NHS Trust, Wolverhampton, UK, ⁷University Hospitals of North Midlands NHS Trust, Stoke on Trent, UK, ⁸St George's University Hospitals NHS FT, London, UK, ⁹Calderdale and Huddersfield NHS FT, Huddersfield, UK, ¹⁰Bradford Teaching Hospitals NHS FT, Bradford, UK, ¹¹Nottingham University Hospitals NHS Trust,, Nottingham, UK, ¹²Institute of Applied Health Research, University of Birmingham, University of Birmingham, Birmingham, UK.

Background and aims: Obstructive Sleep Apnoea (OSA) is associated with an increased risk of Peripheral Neuropathy in patients with T2D (DPN). Hence, continuous positive airway pressure (CPAP) might be beneficial in reducing the burden of DPN in T2D.

Materials and methods: We conducted an open-label multicentre (13 centres) feasibility randomised control trial (RCT) of patients with T2D and OSA, where they were randomised to CPAP vs no CPAP over 2 years. The primary outcomes of this trial were related to feasibility. In this abstract, we report a secondary outcome of the trial, the impact of CPAP on DPN. Participants with resting Oxygen saturation <90%, central apnoea index >15/hrs or Epworth Sleepiness Score (ESS) >= 11 were excluded. DPN was assessed using the Michigan Neuropathy Screening Instrument (MNSI) (MNSIe2 or MNSIq), Neuropad, Vibration perception threshold (VPT>25V), and 10g monofilament. The follow-up procedures were amended due to the COVID-19 pandemic.

Results: Eighty-three patients were randomised to CPAP vs no CPAP (43 vs 40) with a median [IQR] follow-up of 645 [545,861] days. The study population mean (SD) age was 62.5 (10.9) years, BMI was 35.4 (7.2), and diabetes duration was 12.2(7.9) years. 89.1 % (n=74) were white European ethnicity, 28.9% (n=24) were women, and 48.2% (n=40) were prescribed insulin. The prevalence of DPN at baseline was 52.6% (n=41) based on MNSI. Only 26/43 patients used CPAP, with a median (IQR) usage of 3:40 (hours: minutes) [0:06, 4:45] per night. The intention-to-treat analysis is summarised in Table 1. At the study end, being randomised to CPAP was associated with a possible favourable impact on DPN compared to no CPAP based on MNSI, the Neuropad, VPT and 10g monofilament.

Conclusion: CPAP might have a favourable impact on DPN in patients with T2D. However, this requires a to be tested in the fully dedicated RCT.

Clinical Trial Registration Number: the ISRCTN registry ( https://www.isrctn.com/ISRCTN12361838, Registered 04 April 2018, Protocol version: v5.0 02.12.19)

Supported by: Clinician Scientist Fellowship awarded to the Chief Investigator by the National Institute for Health Research (CS-2013-13-029)

Disclosure: A.A. Tahrani: Grants; Novo Nordisk and Sanofi.

OP 06 Adipocytes and adipose tissue plasticity

31

Functional genomics approach for discovery of novel regulators of human adipose tissue plasticity

M. Verma ¹, N. Tolstrup¹, S. Hatch², G. Ettridge¹, J. Ast¹, L. Veizaj¹, C. Sun^1,3, M.Z. Mendoza¹, D. Ebner², A. Epanchintsev¹, A. Agrawal¹, V. Jurtz⁴, E.M. Toledo¹, B. Gilboa¹, C. Ämmälä¹;

¹Novo Nordisk Research Centre Oxford, Oxford, UK, ²Target Discovery Institute, University of Oxford, Oxford, UK, ³Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford, UK, ⁴Novo Nordisk A/S, Malov, Denmark.

Background and aims: Human adipose tissue response (i.e., hypertrophy vs. hyperplasia) to chronic over-nutrition is a critical determinant of an individual’s systemic insulin resistance. Adipose progenitor (AP) cell fate programmes are key to this metabolic plasticity and influence human health. We aim to describe genes associated with human AP identity and investigate their role in regulating AP fate and function.

Materials and methods: Human adipose scRNAseq analyses identified genes associated with AP identity. A functional genomics screen was established to investigate the role of these genes in regulating AP fate and function. Screening utilised CRISPR mediated loss-of-function (LoF) in human immortalised dedifferentiated (DFAT) APs, high content imaging, proximity ligation based adipokine assays, transcriptomics and machine learning (ML). We performed a proof-of-concept (PoC) screen involving LoF of genes known for regulating AP fate and function and non-targeting controls (NTCs).

Results: scRNAseq identified 608 genes associated with human APs. Disease annotation revealed type 2 diabetes (T2D) as top hit (150 genes; FDR = 4.5e^-11) in these genes. Knowledge graph-based novelty profiling suggested 2 clusters of well-known (144) and under-explored (464) genes that associates with metabolic diseases. CRISPR validation achieved 97% reduction in HPRT mRNAs and 67% reduction in 6-thioguanine toxicity in DFAT APs (p<0.001); 81% reduction in PPARG mRNAs and 91% reduction in lipid accumulation in in vitro differentiated DFAT APs (p<0.001). PoC screen at scale was enabled by an automation workflow. Cellular phenotyping utilised a) a novel deep learning model that generated neutral lipid-stained “like” images from brightfield images for longitudinal observation of lipid accumulation and >300 feature estimation; and b) adiponectin, leptin and MCP-1 secretion levels. Unsupervised ML analyses on cellular phenotyping data, revealed >3 clusters vs. NTCs. Supervised ML models used cluster annotations for pairwise (cluster vs. NTCs) feature reduction and classification. Several genes e.g., PTEN, PPARG, IRS-1, CEBPA, PAPOLA, POLA1, RB1 (p = 4e-⁸ - 1e^-53) known to regulate multiple aspects of APs biology were distinguished from NTCs in PoC screen.

Conclusion: Genes associated with human AP identity may modulate T2D risk. This functional genomic workflow may identify novel regulators of human AP fate and function, with direct impact for patient health.

Disclosure: M. Verma: Employment/Consultancy; MV, NT, GE, JA, LV, MLZM, AE, AA, EMT, BG, CA are employed by Novo Nordisk, Ltd., UK. VJ is employed by Novo Nordisk A/S Denmark. DE has received research funding from Novo Nordisk.

32

Estrogen receptor in adipocyte precursor cells: implications for diabetes after breast cancer therapy

E.A. Wellberg ¹, R.L. Scalzo², N.S. Thomas¹;

¹University of Oklahoma Health Sciences Center, Oklahoma City, OK, USA, ²University of Colorado Anschutz Medical Campus, Aurora, CO, USA.

Background and aims: Women who were treated for breast cancer have a 23% greater risk for type 2 diabetes than those who were not, and up to 48% of diabetes cases in breast cancer survivors are attributable to anti-estrogen (endocrine) therapies. Estrogen receptor (ER) signaling maintains metabolic homeostasis through central and peripheral mechanisms; therefore, any treatment that depletes estrogens or blocks ER will impact tissues all over the body. We recently identified a role for estrogen signaling in maintaining adipocyte progenitor cells in obese female mice, and hypothesize that endocrine therapies disrupt progenitor cell renewal during a positive energy balance, increasing diabetes risk. A balance of mature adipocyte hypertrophy and stromal progenitor cell hyperplasia is required for healthy adipose tissue expansion, to prevent ectopic lipid deposition that can cause type 2 diabetes. The aim of this study was to define the role of ER in adipocyte precursor cells and healthy adipose tissue expansion.

Materials and methods: We developed a murine model that uses a high fat diet, thermoneutrality, and low dose sustained tamoxifen or estrogen deprivation to accurately replicate relevant biological variables. The impact of endocrine therapies on body mass, composition, and glucose tolerance were evaluated. Anti-diabetic therapies were administered as interventions to attenuate the adverse metabolic effects of breast cancer treatments. ER agonists or antagonists were evaluated in cultured mouse adipocyte precursor cells using flow cytometry, gene expression profiling, and differentiation assays.

Results: In contrast to previous reports, tamoxifen treatment did not promote adipose tissue browning or rapid weight loss in mice. In high fat-fed females, endocrine therapies associated with weight gain, fewer adipocyte progenitors, larger mature adipocytes, hepatic steatosis, and glucose intolerance, which is consistent with human studies. Single cell RNA sequencing data from subcutaneous adipose stroma implicated stromal Wnt1-inducible signaling pathway protein 2 (Wisp2/Ccn5) in mediating the beneficial effects of estrogen on adipose tissue expansion. In isolated subcutaneous adipocyte precursor cells, we found that ER is more highly expressed in progenitors compared to the more committed preadipocytes. Estradiol treatment induced Wisp2 expression in an ER-dependent manner. Treatment of cultured cells with estradiol or Wisp2 increased progenitor cell proportions. Conversely, the selective ER modulator tamoxifen depleted progenitors and increased the fraction of preadipocytes.

Conclusion: This study highlights a potential mechanism through which ER signaling benefits whole body metabolism in a clinically relevant model of breast cancer endocrine therapy and the associated side effects. The regulation of hyperplastic adipose tissue expansion by estrogen signaling may underlie the elevated risk for type 2 diabetes in breast cancer survivors.

Supported by: HESI THRIVE; NIH NCI

Disclosure: E.A. Wellberg: None.

33

Ehd2 regulates plasma membrane integrity and downstream insulin signalling events

K.G. Stenkula ¹, C. Fryklund¹, A. Borreguero-Muñoz¹, H. Taylor², F. Kopietz¹, H. Ardalani³, O. Rogova³, P. Spegel³, N. Bryant⁴, G. Gould², M. Neuhaus¹;

¹Experimental Medical Science, Lund University, Lund, Sweden, ²University of Strathclyde, Glasgow, UK, ³Department of Chemistry, Lund University, Lund, Sweden, ⁴Department of Biology and York Biomedical Research Institute, University of York, York, UK.

Background and aims: Adipocyte dysfunction is a crucial driver of insulin resistance and type 2 diabetes. We have identified EH domain-containing protein 2 (Ehd2) as one of the most highly upregulated genes at the early stage of adipose tissue expansion. Ehd2 is a dynamin-related ATPase influencing several cellular processes, including membrane recycling, caveolae dynamics and lipid metabolism. The aim with the current study was to investigate the role of Ehd2 in adipocyte insulin signaling and glucose transport.

Materials and methods: Male C57BL6/N Ehd2 knockout and wildtype mice were fed 2 weeks of high-fat diet to reflect changes at the early stage of adipose tissue expansion. Primary adipocytes isolated at termination were analyzed to assess i) gene and protein expression of caveolin-1, insulin receptor and downstream targets by qPCR and western blot, respectively; ii) insulin-stimulated glucose uptake using glucose-tracer assay; and iii) lipid composition and cholesterol content by lipidomic analysis. In addition, we employed siRNA-induced Ehd2 knockdown in cultured 3T3-L1 adipocytes to examine insulin-stimulated complex formation of Soluble N-ethylmaleimide-sensitive factor-attachment protein receptors (SNARE's), which are essential for GLUT4 vesicle fusion, using proximal ligation assay.

Results: We demonstrate that under short-term high-fat diet-challenged conditions, Ehd2 deficiency caused a more severe deterioration of insulin signal transduction and impaired insulin-stimulated GLUT4 translocation. Furthermore, lack of Ehd2 was associated with altered plasma membrane lipid and protein composition, reduced insulin receptor expression, and diminished insulin-dependent SNARE protein complex formation.

Conclusion: These data highlight the importance of Ehd2 for maintenance of plasma membrane integrity, insulin receptor stability, and downstream insulin receptor signaling events, which all are essential for mediating insulin-stimulated glucose transport.

Supported by: NN, SRC, SDF

Disclosure: K.G. Stenkula: None.

34

Gpr146 deficiency preserves a lean, metabolically healthy phenotype in mice during diet-induced obesity

S. Brachs ^1,2, A. Harutyunyan^1,2, J. Sbierski-Kind¹, R. Sun^1,2, A. Liu^1,2, M. Brachs¹, L. Spranger¹, S. Kunz³, S. Ki⁴, L. Ehrlich⁴, K. Hu He⁵, M. Bielohuby⁵, P. Wohlfart⁵, K. Mai^1,2, J. Spranger^1,2;

¹Department of Endocrinology and Metabolism, Charité-Universitätsmedizin Berlin, Corporate Member of Freie Universität Berlin, Humboldt-Universität zu Berlin, Berlin, Germany, ²DZHK (German Centre for Cardiovascular Research), partner site Berlin, Berlin, Germany, ³CF Electron Microscopy, Max-Delbrück-Center for Molecular Medicine, Berlin, Germany, ⁴Institute for Cellular and Molecular Biology, Department of Molecular Biosciences, The University of Texas at Austin, Austin, TX, USA, ⁵Research and Development, Sanofi-Aventis Deutschland GmbH, Frankfurt am Main, Germany.

Background and aims: The G protein-coupled orphan receptor 146 (GPR146) is highly expressed in murine and human adipose tissue. Previously, we showed that Gpr146-deficient 3T3-L1 cells were not able to differentiate into adipocytes. Furthermore, we found GPR146 expression upregulated over a 3 months weight reduction phase in human adipose tissue. Thus, we examined GPR146 expression in our MAINTAIN study and further investigated metabolic function of Gpr146 in a Gpr146-deficient mouse model (GPR146^-/-).

Materials and methods: Correlations between adipose GPR146 expression and the recorded clinical parameters were evaluated. Furthermore, we investigated male GPR146^-/- mice challenged by diet-induced obesity (DIO), feeding a high-fat or western diet compared to a standard control diet for 3 months. Here, we examined glucose and insulin homeostasis, changes in body composition as well as lipid and energy metabolism. Additionally, we evaluated lipogenic and lipolytic capacity of adipose tissue in GPR146^-/- mice and analyzed the contribution of Gpr146 in adipocyte differentiation.

Results: In our MAINTAIN study, adipose GPR146 expression negatively correlated with BMI, but positively with ISI during hyperinsulinemic-euglycemic clamps, and was associated with enhanced insulin-mediated suppression of free fatty acid release after weight reduction (r = -0.43; p < 0.001). Moreover, lower adipose GPR146 expression predicted a reduced body weight regain after 18 month follow up (Coefficients: 0.117, 95% CI: 0.011 - 0.222, R²: 0.078, p = 0.031). GPR146^-/- mice gained significantly lower body weight and were protected against fat mass increase during DIO. Plasma concentrations of cholesterol, triglycerides and leptin were reduced. Under DIO, GPR146^-/- mice showed lower activity but increased energy expenditure with sustained glucose and insulin handling. RNA sequencing and immune cell analysis indicated less adipose tissue inflammation. Moreover, alterations of lipid uptake and release were observed in adipose tissue. Finally, a significant induction of Gpr146 expression was detected during ex vivo adipocyte differentiation.

Conclusion: We observed sustained body weight and lower fat mass gain with reduced activity and higher energy expenditure in GPR146^-/- mice during obesity challenge. GPR146^-/- mice showed a preserved metabolically healthy, lean phenotype. Our results indicate an important metabolic function of Gpr146/GPR146 in murine and human adipose tissue.

Clinical Trial Registration Number: NCT00850629

Supported by: Charité-Sanofi Joint lab, DFG KFO 218/1, DDG Menarini Preis SB, BMBF (DZHK; BER5.1) JS

Disclosure: S. Brachs: Grants; BMBF (DZHK; BER5.1), Deutsche Diabetes Gesellschaft (Menarini Preis 2022), DFG (KFO 218/1), Joint lab between Charité and Sanofi.

35

Arctic diet prevents postprandial hyperglycaemia in Tbc1d4-deficient mice

A.K. Scheel ¹, S. Pirseyedi¹, L. Espelage¹, H. Al-Hasani^1,2, A. Chadt^1,2;

¹Institute for Clinical Biochemistry and Pathobiochemistry, German Diabetes Center, Leibniz Center for Diabetes Research, Duesseldorf, Germany, ²German Center for Diabetes Research (DZD e.V.), Neuherberg, Germany.

Background and aims: The Rab-GTPase activating protein (RabGAP) TBC1D4 (AS160) is a key regulator of energy flux in oxidative skeletal muscle and adipose tissue. In previous studies, Tbc1d4-deficient (D4KO) mice showed decreased insulin-stimulated glucose uptake into skeletal muscle and white adipocytes and impaired contraction-mediated glucose uptake into skeletal muscle. In humans, a number of genetic variants in the TBC1D4 gene are associated with insulin resistance. In arctic populations, a common skeletal muscle-specific TBC1D4 p.Arg684Ter loss-of-function variant defines a specific subtype of type 2 diabetes (T2D). This variant is associated with postprandial hyperglycemia and a substantially elevated T2D risk and is hypothesized to represent an evolutionary adaptation to a traditional lifestyle of indigenous populations that is characterized by a diet low in carbohydrates and a high degree of daily physical activity.

Materials and methods: We generated whole-body D4KO mice as a model for genetically induced insulin resistance. The mice were subjected to high-fat diets differing in carbohydrate and protein content to mimic a traditional Arctic (low-carbohydrate high-protein) or a Western (high-carbohydrate low protein) diet, respectively. Longitudinal analyses of skeletal muscle, adipocyte function and glycemic control in response to dietary interventions were performed.

Results: Our findings indicate that Western diet feeding induces postprandial hyperglycemia in whole-body D4KO mice (WT vs. D4KO: 203 ± 8 mg/dl vs. 324 ± 28 mg/dl, p<0.01) and impaired insulin-stimulated glucose uptake into oxidative Soleus muscle (WT vs. D4KO: 3.11 ± 0.44 vs. 1.9 ± 0.23 nmol/mg/20min, p<0.017) and primary white adipocytes (WT vs. D4KO: 174.19 ± 26.56 vs. 52.5 ± 9.78 cpm/ mg lipid, p<0.0001). Interestingly, mice subjected to the Artic diet presented postprandial normoglycemia (WT vs. D4KO: 148 ± 5 mg/dl vs. 130 ± 5 mg/dl, p<0.315) and improved insulin-stimulated glucose uptake into skeletal muscle (Soleus WT vs. D4KO: 2.68 ± 0.33 vs. 3.03 ± 0.21 nmol/mg/20min, p<0.999).

Conclusion: Western diet feeding induces postprandial hyperglycemia in whole-body D4KO mice as observed in homozygous TBC1D4 p.Arg684Ter variant carriers. In contrast, a diet rich in protein and lipids but low in carbohydrates protects from insulin resistance as induced by the Tbc1d4-deficiency. Therefore, we hypothesize that an Arctic diet may prevent the onset of T2D in homozygous TBC1D4 p.Arg684Ter gene variant carriers and thus serves as a potential treatment in precision medicine.

Supported by: RTG 2576 "vivid - In vivo investigations towards the early development of type 2 diabetes"

Disclosure: A.K. Scheel: None.

36

WITHDRAWN

OP 07 Regulation of islet function through gene editing and modification

37

Establishment of pancreatic beta cell-specific gene knockout system based on CRISPR-Cas9 technology with AAV8-mediated gRNA delivery

K. Ueki ¹, Y. Nishida¹, H. Uzawa¹, A. Kanai¹, T. Miyatsuka², H. Watada¹;

¹Department of Metabolism and Endocrinology, Juntendo University Graduate School of Medicine, 2-1-1, Hongo, Bunkyo-ku, Tokyo, Japan, ²Department of Endocrinology, Diabetes and Metabolism, Kitasato University School of Medicine, 1-15-1 Kitasato, Minami-ku, Sagamihara, Kanagawa, Japan.

Background and aims: The cre-loxP system has provided useful resources to analyze the importance of tissue-specific gene knockout, including pancreatic beta cells associated with the pathogenesis of diabetes mellitus. However, it is expensive and time-consuming to generate transgenic mice harboring floxed genes of interest and cross them with pancreatic beta cell-specific Cre-expression mice. Therefore, we focused on the CRSIPR-Cas9 technology and aimed to generate pancreatic β-cell-specific gene knockout mice more conveniently by introducing adeno-associated virus (AAV8) into pancreatic β-cells.

Materials and methods: We crossbred CAG-LSL(loxP-Stop-loxP)-Cas9-P2A-EGFP mice and Ins1-Cre mice, in which the stop cassette flanked by loxP sequences is removed by Cre recombinase to generate pancreatic β-cell-specific Cas9-expressing (βCas9) mice. We attempted to knock out the genes specifically in pancreatic β-cells by intraperitoneal administration of AAV8 expressing sgRNAs to the targeted genes. The islets of the obtained gene knockout mice were analyzed with flow cytometry and fluorescence microscopy. As a proof of concept, we knocked out Pdx-1, a culprit gene for maturity-onset diabetes of the young 4 (MODY4), and examined the glucose tolerance and histopathology in the islets of the mice.

Results: The co-localization of insulin and EGFP expression was observed in βCas9 mice, indicating that Cas9 was expressed specifically in β-cells. Intraperitoneal glucose tolerance test (ipGTT) showed that glucose tolerance was comparable between βCas9 and control mice (1.00 vs. 0.91±0.05, fold change, n=4, p-value=0.08, AUC in ipGTT). Next, when AAV8 expressing sgRNA against EGFP was administered intraperitoneally to βCas9 mice, the expression of EGFP was decreased by approximately 80 percent compared to the controls (98.3% vs. 21.8±2.06%, fold change, n=3, p-value<0.0001, positive for EGFP expression estimated by flow cytometer) 4 weeks after administration, demonstrating that sgRNA expression with AAV8 effectively knocked out the targeted gene in βCas9 mice. Finally, we knocked out Pdx-1 by injecting AAV8 expressing sgPdx-1 and showed impaired glucose tolerance compared to the controls at 4 weeks after the administration (1.00 vs. 1.52±0.07, fold change, n=4, p-value<0.001, AUC in ipGTT). Histopathologically examining islets from the Pdx1 KO mice showed that β cells lacking Pdx1 expression showed lower insulin with glucagon expression, suggesting the transdifferentiation of the β cells by successful deletion of Pdx-1.

Conclusion: We successfully established the system in which pancreatic β cell-specific gene knockout is performed by AAV8-mediated sgRNA delivery. The method we show here can generate β-cell-specific gene KO mice more rapidly and conveniently than the conventional Cre-LoxP system or trans-pancreatic duct injection of AAV. It should be also noted that our Pdx-1 KO mice might show impaired glucose tolerance due to transdifferentiation of the beta cells with glucagon expression, requiring further validation of our gene KO system by other sgRNAs with no upregulation of glucagon.

Disclosure: K. Ueki: None.

38

Depletion of 14-3-3ζ in pancreatic beta cells by antisense oligonucleotides promotes beta cell proliferation and increases beta cell mass

G.E. Lim ¹, F. Paré¹, S. Djerroud¹, T.P. Prakash², E. Pirie², N. Wijesekara²;

¹Cardiometabolic axis, CRCHUM, Montreal, QC, Canada, ²Ionis Pharmaceuticals, Carlsbad, CA, USA.

Background and aims: We previously demonstrated that constitutive deletion of 14-3-3ζ in murine β-cells enhanced insulin secretion due to increased mitochondrial function. Moreover, 14-3-3ζ deletion also promoted β-cell proliferation. With the knowledge that diabetes is associated with a profound loss of functional β-cell mass, defined by insulin secretory capacity and the number of β-cells, these findings suggest that targeted depletion of 14-3-3ζ in mature β-cells could have beneficial effects on metabolic control by increasing functional β-cell mass. It is known that targeting β-cells in vivo is difficult due to the localization of islets and the heterogenous endocrine cell populations that comprise the islets of Langerhans. We previously reported the ability of targeting β-cells in vitro and in vivo with antisense oligonucleotides (ASOs) that are conjugated to GLP-1. As this novel, β-cell-centric approach was effective in ensuring the depletion of target proteins, we sought to examine if it could be used to reduce 14-3-3ζ expression in primary murine β-cells, thereby altering β-cell mass, proliferation, or function.

Materials and methods: ASOs against Ywhaz (ASO_Ywhaz) were conjugated to human GLP-1 (GLP-1-ASO_Ywhaz), and C57BL/6J mice were injected once weekly with a control ASO (GLP-1-ASO_Con) or GLP-1-ASO_Ywhaz for 4 weeks (1mg/kg s.c.), followed by i.p. glucose tolerance tests or measurements of insulin secretion (2mg/kg glucose). At endpoint, islets and tissues were harvested to assess specificity of islet 14-3-3ζ depletion by quantitative PCR, and β-cell mass and proliferation were also measured by immunohistochemistry. Isolated C57BL/6J islets were also exposed to GLP-1-ASO_Con or GLP-1-ASO_Ywhaz (100 nM) for 48- and 72-hours to assess knockdown efficiency, gene expression, and glucose stimulated insulin secretion (4 mM vs. 16 mM glucose; GSIS) in vitro.

Results: Treatment of isolated islets with GLP-1-ASO_Ywhaz significantly reduced mRNA levels of Ywhaz by 73% (p<0.05), in addition to parallel increases in Ins2 and Sst mRNA levels (164% and 111%, respectively; p<0.05). In vitro treatment of mouse islets with GLP-1-ASO_Ywhaz also potentiated GSIS by up to 1.7-fold (p<0.05). In vivo administration of GLP-1-ASO_Ywhaz reduced islet-specific levels of Ywhaz mRNA by 75% (p<0.05) with no discernable effects on mRNA levels of Ins1, Ins2, Sst, and Gcg. Analysis of Ywhaz mRNA levels in extra-pancreatic tissues, including heart, adipose, skeletal muscle, and liver, revealed no off-target effects of GLP-1-ASO_Ywhaz. Although no differences in in vivo insulin secretion were observed, a transient and modest impairment in glucose tolerance was observed in GLP-1-ASO_Ywhaz-treated mice, which persisted for up to 4 weeks after the last injection. Depletion of Ywhaz with GLP-1-ASO_Ywhaz significantly increased β-cell mass by 1.4-fold (p<0.05) when compared to GLP-1-ASO_Con-injected mice. This increase in β-cell mass corresponded with a 2.3-fold increase (p<0.05) in β-cell proliferation, as measured by EdU-positive β-cells.

Conclusion: In vivo depletion of 14-3-3ζ in β-cells with GLP-1-ASO_Ywhaz was effective in increasing pancreatic β-cell mass, despite a transient impairment in glucose tolerance. Collectively, these findings demonstrate the feasibility of in vivo targeting of 14-3-3ζ in β-cells with ASOs and potentially represent a novel approach to increase β-cell mass in vivo.

Supported by: CIHR, Canada Research Chairs Program

Disclosure: G.E. Lim: None.

39

Direct beneficial effects of truncated C1QL1 on mouse and human islet function in vitro

M. Billert, P. Atanes, I. Ruz-Maldonado, M. Zhao, S. Persaud;

Department of Diabetes, King's College London, London, UK.

Background and aims: We have identified that C1QL1, which encodes a secreted complement protein, is one of the most abundant GPCR ligand mRNAs expressed by human islets and we have shown that a truncated 101 amino acid fragment of C1QL1, termed C1QL1Δ4, stimulates insulin secretion from MIN6 beta cells and protects them against apoptosis. We have now quantified the effects of C1QL1Δ4 in mouse and human islets and investigated the mechanisms through which it exerts its effects.

Materials and methods: All experiments were performed using islets isolated from male CD1 mice or human organ donors, with ethical permission. Insulin secretion and apoptosis were quantified by radioimmunoassay and Caspase-Glo luminescence, respectively. Intracellular calcium levels were determined by single cell microfluorimetry following loading of dissociated islet cells with 10 μM Fura-2, and IP₁ and cAMP generation was determined using HTRF assays.

Results: 20 nM C1QL1Δ4 had no effect on basal insulin release but significantly potentiated glucose-stimulated insulin secretion from mouse islets (2 mM glucose: 0.24±0.1 ng/islet/h; +20 nM C1QL1Δ4: 0.27±0.03; 20 mM glucose: 1.6±0.2; +20 nM C1QL1Δ4: 2.7±0.2, n=17-20, P<0.01) and similar observations were seen in experiments with human islets (2 mM glucose: 0.8±0.1 ng/islet/h; +20 nM C1QL1Δ4: 1.2±0.1; 20 mM glucose: 1.6±0.3; +20 nM C1QL1Δ4: 3.0±0.4, n=8-11, P<0.01). In perifusion experiments with mouse and human islets 20 nM C1QL1Δ4 also reversibly increased glucose-induced insulin secretion (P<0.001, n=4). 48h exposure to 20 nM C1QL1Δ4 significantly protected islets from apoptosis induced by inflammatory cytokines (cytokine-induced apoptosis +20 nM C1QL1Δ4, % maximal: mouse islets, 75.0±3.5; human islets, 60.7±9.2, n=4, P<0.01) and 500 μM of the saturated long chain fatty acid palmitate (palmitate-induced apoptosis +20 nM C1QL1Δ4, % maximal: mouse islets, 53.9±3.7; human islets, 42.5±5.5, n=6, P<0.001). The protection observed with C1QL1Δ4 against cytokine- and palmitate-induced apoptosis was not significantly different from that seen with 20 nM of the GLP-1 mimetic exendin-4 (n=6, P>0.2). C1QL1Δ4 reversibly increased intracellular calcium in Fura-2-loaded islet cells, which was likely to be via calcium influx since it did not stimulate inositol phosphate generation in islets (IP₁ nM: basal, 0.04±0.03; +100 nM C1QL1Δ4, 0.04±0.01; +100 μM carbachol, 40.3±4.6, n=3-6, P<0.0001 carbachol vs basal). Quantification of islet cAMP levels suggests that C1QL1Δ4 activates a Gs-coupled GPCR since it elevated cAMP (cAMP nM: basal, 0.2±0.06; +20 nM C1QL1Δ4, 14.5±3.5; +20 nM exendin-4, 22.6±4.1, n=3, P<0.01 and P<0.001 C1QL1Δ4 and exendin-4 vs basal, respectively), and it did not inhibit forskolin-induced cAMP production.

Conclusion: We have demonstrated that a fragment of C1QL1, consisting of the first 101 amino acids (C1QL1Δ4), potentiates glucose-induced insulin release from mouse and human islets and protects them from cytokine and palmitate-induced apoptosis, with similar efficacy to exendin-4. The mechanism through which C1QL1Δ4 exerts these beneficial effects has not been fully identified, but it may be via a Gs-coupled GPCR since it elevates cAMP in islets.

Supported by: Diabetes UK

Disclosure: M. Billert: None.

40

Suppression of microRNA-33 improves pancreatic beta cell function and mass in murine and human islets

R. Ortega ¹, I. Ruz-Maldonado², P. Atanes¹, N. Price², M. Zhao¹, C. Fernandez-Hernando², S. Persaud¹, B. Liu¹;

¹Department of Diabetes, Faculty of Life Sciences & Medicine, King's College London, London, UK, ²Comparative Medicine & Pathology, Yale University School of Medicine, New Haven, CT, USA.

Background and aims: TargetScan has revealed that microRNA-33 (miR-33) targets a network of genes that are known to play key roles in the regulation of beta-cell mass. We have previously reported that expression of miR-33 was significantly down-regulated in islets isolated from mice fed with a high-fat diet where beta-cell proliferation was increased by 5.3±0.4-fold prior to the onset of a diabetic phenotype. We have also shown that inhibition of miR-33 in isolated mouse islets promoted glucose-stimulated insulin secretion and protected islets from cytokine-induced apoptosis. The aim of this study was to investigate whether miR-33 regulates beta-cell function and mass in islets isolated from human donors and miR-33^-/- mice.

Materials and methods: Expression of miR-33 in human islets isolated from lean (BMI<25 kg/m2) and obese (BMI≥30.0 kg/m2) donors was measured using Affymetrix miRNA profiling. To suppress the expression of miR-33, isolated human islets were transduced with lentiviral miR-33 inhibitors at multiplicity of infection of 5 or control vectors under the same experimental conditions. MiR-33^-/- mice were generated following CRISPR/Cas9-mediated excision of miR-33 using targeted guide sequences toward intron 16 of the sterol-regulatory element-binding protein-2 gene. The effects of miR-33 inhibition on islet caspase-3/7 activities were quantified using a luminescent assay. Insulin content and insulin secretion following miR-33 inhibition or deletion were determined by radioimmunoassay or insulin ELISA. Immunohistochemistry was used to quantify beta-cell proliferation in fixed pancreases from miR-33^-/- mice.

Results: Expression of miR-33 was down-regulated in islets isolated from obese donors (log2 normalised probe intensity; lean vs obese donor islets; 1.12±0.45 vs 0.56±0.08; n=3 biological replicates). In addition, lentivirus-mediated inhibition of miR-33 in human islets resulted in enhanced insulin secretion in response to 20mM glucose, without having any effect on basal insulin release (ng/islet/hour; SC: human islets transduced with scrambled control lentiviral inhibitor: 2mM glucose: 0.39±0.02; 20mM glucose: 0.64±0.03; LentV-miR-33: human islets transduced with miR-33 lentiviral inhibitor: 2mM glucose: 0.47±0.03; 20mM glucose: 1.01±0.05; n=7; P<0.01). Downregulation of miR-33 in human islets also increased insulin content (ng/islet; SC vs LentV-miR-33; 21.54±0.59 vs 26.43±2.09; n=6; P<0.05), and reduced islet apoptosis (luminescence units; 95129±3847 vs 84354±2427; n=8; P<0.05). Deletion of miR-33 in mice was associated with significant increases in basal insulin secretion (wildtype vs miR-33^-/-; 0.03±0.00 vs 0.05±0.01; n=8; P<0.05), insulin secretion in response to 20mM glucose (0.17±0.01 vs 0.23±0.01; n=8; P<0.01), insulin content (79.07±3.27 vs 120.59±2.79; n=8; P<0.001) and number of Ki67-positive beta-cells per islet (0.14±0.02 vs 0.54±0.11; n=35-50 islets; P<0.0001).

Conclusion: Data presented in this study are consistent with miR-33 playing a pivotal role in regulating beta-cell secretory function, beta-cell apoptosis and proliferation. Inhibition of miR-33 expression therefore represents a novel strategy for improving beta-cell function and maintaining beta-cell mass for the treatment of diabetes.

Supported by: Diabetes UK

Disclosure: R. Ortega: None.

41

Human pancreatic islet miRNA-mRNA regulatory networks associated with insulin secretion and type 2 diabetes development

A. Karagiannopoulos ¹, E. Cowan¹, M. Maziarz², J.L.S. Esguerra^1,3, L. Eliasson¹;

¹Islet Cell Exocytosis, Lund University Diabetes Centre, Department of Clinical Sciences - Malmö, Lund University, Lund, Sweden, ²Unit of Bioinformatics, Lund University Diabetes Centre, Lund University, Lund, Sweden, ³Novo Nordisk A/S, Copenhagen, Denmark.

Background and aims: A contributing cause of type 2 diabetes (T2D) is dysregulated insulin secretion from the β-cells within the pancreatic islets of Langerhans. MicroRNAs (miRNAs) are small non-coding RNAs that post-transcriptionally regulate gene expression essential for insulin secretion. The important role of individual miRNAs in type 2 diabetes (T2D) is well-studied, but one miRNA can exert its effects on multiple gene targets and single genes can be regulated by many miRNAs. Hence, a comprehensive view on how the interactions of miRNAs with their target genes can lead to T2D pathogenesis is missing. Here we aim to characterize the miRNA profile of T2D human pancreatic islets, identify confident miRNA targets and investigate the miRNA-mRNA networks implicated in the modulation of insulin secretion.

Materials and methods: Large-scale small RNA sequencing was performed in human pancreatic islets from non-diabetic (ND) and T2D donors and data were analyzed with an in-house pipeline. Differential expression analysis was performed with DESeq2 to identify miRNAs with altered expression in T2D. RNADisease (v4.0) and HMDD (v3.0) databases were utilized for miRNA-disease enrichment analysis. By integrating RNA-seq data from the same donors we identified negatively correlated miRNA-mRNA pairs which were supported by publicly available data on predicted (TargetScan, miRDB) or validated (TarBase, miRTatBase) pairs. MiRNA-target co-expression clusters were detected with Weighted Correlation Network Analysis (WGCNA) and were functionally annotated with Gene Ontology (GO) enrichment analysis.

Results: We found 70 differentially-expressed (DE) miRNAs (34 up-/36 downregulated) in human islets from T2D donors (N=52 ND, 9 T2D) (adj. p<0.05). Only miR-187-3p was found to be commonly DE in previous studies of similar design. MiRNA-disease enrichment analysis showed T2D to be strongly enriched only with the upregulated T2D miRNAs. Target analysis showed miR-101-3p to be associated with the highest number of genes (766) of which 25% were previously defined as T2D-related genes. Conversely, CHD9 is targeted by the highest number of miRNAs (20). Remarkably, genes with very high expression in the human islet seem to be subjected to limited miRNA regulation as they are targeted by none or very low number of miRNAs. WGCNA followed by GO term enrichment analysis revealed a cluster of co-expressed genes that are involved in insulin secretion, consisting exclusively of targets of upregulated T2D miRNAs. Lastly, when correlations between miRNA expression and 1^st and 2^nd phase insulin secretion were performed, 8 DE miRNAs were found to be significantly correlated with both traits (among them miR-101-3p, miR-9-5p and miR-141-3p).

Conclusion: Human islets from T2D donors display altered expression of miRNAs which are involved in miRNA-mRNA networks that affect insulin secretion. Significant correlation of the expression of DE miRNAs in T2D with both 1^st and 2^nd phase insulin secretion measurements suggests that these miRNAs are involved in the impairment of insulin secretion associated with the development of T2D.

Supported by: Royal Physiographic Society of Lund, Swedish Research Council, SRA-Exodiab, Swedish Foundation for Strategic Research (LUDC-IRC)

Disclosure: A. Karagiannopoulos: None.

42

Can long non-coding RNA derived micropeptides impact the pathogenesis of type 1 diabetes?

J. Mentxaka Salgado ^1,2, K. García Etxeberria³, L. Bergara¹, L. Mendoza Gómez², A. Castellanos Rubio^4,5, I. Santin^1,2;

¹Biochemistry and molecular biology, UPV/EHU, Leioa, Spain, ²Biocruces Bizkaia Health Research Institute, Barakaldo, Spain, ³Biodonostia Health Research Institute, Donostia, Spain, ⁴Genetics and Human Anthropology, UPV/EHU, Leioa, Spain, ⁵IKERBASQUE Basque Foundation for Science, Bilbao, Spain.

Background and aims: Type 1 diabetes (T1D) is a multifactorial disease whose development is heavily affected by both genetic susceptibility and environmental factors, such as enteroviral infections. Most genetic T1D-associated variants identified by GWAS fall within no coding regions of the genome, many of them being long non-coding RNA(lncRNA) genes. Although traditionally regarded as incapable of producing protein products due to their lack of coding potential, more recent evidence has proved that they can bind to the ribosome and generate stable and functional peptides from short open reading frames. Our hypothesis is that in pancreatic ß cells, upon viral infection induced stress, infrequent binding of non-coding transcripts to the ribosome might be promoted. As a result, translation of micropeptides could be favored under inflammatory conditions that could act as novel β cell autoantigens. Thus, the aim of this study was to identify lncRNA transcripts from which small peptides could potentially be translated in ß cells in response to a viral insult.

Materials and methods: Polyinosinic-polycytidylic acid (PIC), a synthetic double-stranded synthetic RNA, was transfected into a human pancreatic ß cell line (EndoC-ßH1) to mimic a viral infection. Ribosome complex-bound RNA from PIC-treated and untreated cells was isolated by an affinity purification method relying on active ribosomes incorporating a methionine homolog molecule onto nascent peptides. Ribosome complex-bound RNA was sequenced by high-throughput-sequencing and the coding potential of identified long non-coding transcripts of interest was assessed by means of bioinformatical tools. The coding ability of some promising candidates was further tested in vitro by Western blot and immunofluorescence.

Results: RNA sequencing of ribosome complex-bound RNA showed that around 27% of detected transcripts were lncRNAs, slightly more in the PIC treated condition. When intersecting the lncRNA genes in our dataset with publicly available ribosome profiling data in EndoC-ßH1, we obtained 140 hits of lncRNA genes with reasonable probability to be translated into a peptide. Additionally, we identified another lncRNA gene in our dataset which bears a T1D-associated single nucleotide polymorphism. The bioinformatical analysis revealed a relatively conserved 330 base pair open reading frame with high coding potential within the transcript and strong Kozak consensus sequence. Gene expression analysis of this RNA transcript showed that it is mainly cytosolic in EndoC-ßH1 cells, in line with described micropeptides in literature. Finally, the presence of the predicted 109 amino acid peptide was confirmed in vitro by immunofluorescence and Western blot.

Conclusion: Our results validate our approach for the quick identification of lncRNA gene candidates with potential peptide generating ability. Furthermore, we have been able to confirm one such candidate in in vitro experiments. It will be future work to elucidate the function of this novel peptide in pathogenesis of type 1 diabetes, as well as to discover new key players in this new field.

Supported by: Basque Government Department of Health

Disclosure: J. Mentxaka Salgado: None.

OP 08 Towards curing obesity

43

Effects of gastric bypass on regional brain glucose uptake during normo- and hypoglycaemia

G. Fanni ¹, S. Radhi¹, S. Kvernby², M. Sundbom², S. Haller², E. Roman³, J. Wikström², M. Lubberink², J.W. Eriksson¹;

¹Department of Medical Sciences, Uppsala University, Uppsala, Sweden, ²Department of Surgical Sciences, Uppsala University, Uppsala, Sweden, ³Department of Anatomy, Physiology and Biochemistry, Swedish University of Agricultural Sciences, Uppsala, Sweden.

Background and aims: Obese and insulin resistant individuals display higher brain glucose uptake during experimental hyperinsulinemic normoglycemia, and this is associated with higher plasma glucose and endogenous glucose production, abnormal beta-cell function, and lower M-value. This is reversed by Roux-en-Y gastric bypass (RYGB). Regional changes in brain glucose uptake after RYGB have been showed when assessed as response to OGTT and standardised meal, but not yet in response to hypoglycemia, which is common in post-GBP patients. We aimed at characterizing the RYGB-induced changes in regional brain glucose uptake during a normo- and hypoglycemic clamp and explore their association with anthropometric and metabolic outcomes as well as with changes in neuroendocrine pathways, functional connectivity, and cerebral blood flow.

Materials and methods: Post-hoc analyses of regional brain glucose uptake were performed on 10 individuals with obesity and no diabetes (age 25-49, BMI 35.2-45.4 kg/m²) undergoing RYGB. They were investigated with combined 18F-FDG-PET and fMRI (functional magnetic resonance imaging) of the brain during hyperinsulinemic normo- and hypoglycemic clamp, approximately one month before and four months after RYGB. Patients were thoroughly characterized with anthropometrics, hormonal assessments, glucose homeostasis indices, cognitive tests, hypoglycemic symptoms scale, cerebral blood flow, and resting-state functional connectivity analyses. Glucose uptake in the brain, assessed as 18-FDG clearance, was assessed during hyperinsulinemic normo- and hypoglycemia in selected brain regions, including both cortical (frontal, parietal, temporal, occipital) and subcortical areas (dorsal striatum, globus pallidus, amygdala, hippocampus). Spearman correlation analyses were performed to identify associations with post-surgical changes in other assessed variables.

Results: Glucose uptake under hypoglycemia was increased in globus pallidus after RYGB (p=0.02). No other significant changes of glucose uptake could be detected in the other examined brain regions during either normo- or hypoglycemia. However, the magnitudes of BMI reduction and increase in insulin sensitivity (M-value) following RYGB were significantly associated with lowering of glucose uptake during normoglycemia in subcortical regions, including amygdala, caudate, hippocampus, and putamen (r<-0.67, p<0.05 for all), but not in any cortical area (all p>0.05). There were no significant associations between the regional glucose uptake changes after RYGB and changes in cognitive performance, hypoglycemic symptoms, or counterregulatory hormonal responses.

Conclusion: Following RYGB, lower glucose uptake in subcortical nuclei, but not in cortical areas, was associated with BMI reduction and improved insulin sensitivity. These results may support a role of glucose metabolism in subcortical nuclei in mediating some favourable metabolic effects of bariatric surgery.

Clinical Trial Registration Number: Dnr2015/514

Supported by: SDF, Exodiab, Ernfors, Zetterling, Novo Nordisk, SSMR, TREATMENT, UAS ALF

Disclosure: G. Fanni: Grants; Swedish Diabetes Foundation, Exodiab, Ernfors Foundation, Zetterling Foundation, Swedish Society for Medical Research, Novo Nordisk Foundation, Marie Curie Network TREATMENT, Uppsala University Hospital ALF grants.

44

The metabolomic signature of weight loss interventions: insight from clinical trials

M.L. Smith ¹, L.J. Goudswaard¹, L.J. Corbin¹, The By-Band-Sleeve Trial Management Group, D.A. Hughes¹, A. McConnachine², P. Welsh³, R. Taylor⁴, M.E.J. Lean⁵, N. Sattar³, N.J. Timpson¹;

¹Bristol Medical School, University of Bristol, Bristol, UK, ²Institute of Health and Wellbeing, University of Glasgow, Glasgow, UK, ³School of Cardiovascular and Metabolic Health, University of Glasgow, Glasgow, UK, ⁴Translational and Clinical Research Institute, Newcastle University, Newcastle upon Tyne, UK, ⁵School of Medicine, Dentistry and Nursing, University of Glasgow, Glasgow, UK.

Background and aims: Weight loss is a key intervention for patients with type 2 diabetes (T2D) with proven efficacy on metabolic outcomes. Use of metabolomics technologies within clinical trials of weight loss interventions allows us to characterise the molecular signatures of weight loss and concurrent changes in T2D status. We have characterised the metabolomic consequences of a non-surgical weight management programme in patients with T2D and compared this profile to that seen in bariatric surgery.

Materials and methods: A total of 574 serum samples from the Diabetes Remission Clinical Trial (DiRECT) were collected at baseline and 12 months and analysed using both untargeted mass spectrometry (MS) and targeted ¹H-NMR spectroscopy (NMR). A linear mixed model was fitted to evaluate the effect of the intervention on metabolite levels. The same analysis was performed on 250 samples collected at baseline and 36 months post-randomization in the By-Band-Sleeve trial (BBS), where patients underwent bariatric surgery. Metabolites found to be associated with the interventions form the basis of exploratory analyses to assess the relevance of weight loss and T2D remission status.

Results: All DiRECT participants had T2D at baseline, with 46% of the intervention arm achieving T2D remission at 12 months. In DiRECT, 89 MS metabolites and 68 NMR metabolites were altered by the intervention (Holm-corrected P<0.05). Branched chain amino acids (BCAAs) decreased in the intervention group, whilst increases were seen in the amino acid glycine and some lipids including sphingolipids and plasmalogens. These changes were correlated with improvements in clinical indicators of metabolic health, including biomarkers of liver health. In BBS, 47 MS metabolites were altered by bariatric surgery (Holm-corrected P<0.05), of which 15 overlapped with those altered in DiRECT, including some phospholipid metabolites and glycine (Figure). NMR-measured BCAAs were also decreased following bariatric surgery.

Conclusion: We have characterised the metabolomic effects of an integrated weight management programme previously shown to deliver weight loss and diabetes remission and found consistencies with the metabolomic footprint of bariatric surgery. Metabolites commonly altered by both interventions are likely to be effects of weight loss, whilst those only altered in one study are intervention specific effects that may give detailed insight in the intervention’s mechanisms of action. Future work will explore the role that these intermediate metabolites play in weight loss and the remission of T2D.

Clinical Trial Registration Number: ISRCTN03267836; ISRCTN00786323

Supported by: Wellcome Trust, NIHR Biomedical Research Centre, Diabetes UK, NIHR HTA Programme

Disclosure: M.L. Smith: None.

45

Remission of type 2 diabetes 5 years after gastric bypass versus sleeve gastrectomy: the Oseberg RCT

J.W. Hauge ¹, H. Borgeraas¹, J.K. Hertel¹, M.C. Småstuen¹, R. Sandbu¹, L.K. Johnson¹, M. Svanevik¹, K.I. Birkeland², R.L. Kolotkin³, T.G. Valderhaug⁴, H.L. Gulseth⁵, J. Hjelmesæth^1,2, D. Hofsø¹;

¹Vestfold Hospital Trust, Tønsberg, Norway, ²University of Oslo, Oslo, Norway, ³Duke University School of Medicine, Durham, NC, USA, ⁴Akershus University Hospital, Lørenskog, Norway, ⁵Norwegian Institute of Public Health, Oslo, Norway.

Background and aims: Bariatric surgery causes weight loss, ameliorates hyperglycemia, and improves quality of life. Long-term data comparing standard gastric bypass (GBP) with sleeve gastrectomy (SG) in type 2 diabetes mellitus (T2DM) are limited. We aimed to assess if GBP provides superior 5-yr diabetes remission compared with SG.

Materials and methods: Single-centre RCT of adult patients with obesity and T2DM assigned (1:1) to GBP or SG with 5-yr follow-up. Main clinical outcome was diabetes remission, assessed as HbA1c ≤ 6% (42mmol/mol) without antidiabetic drugs. Predefined secondary outcomes included body weight loss, quality of life, and adverse events. The treatment policy estimand assessed effects including data from patients converted from SG to GBP (missing data handled by multiple imputation). The trial product estimand assessed treatment effects excluding data in patients converted to GBP from time of conversion. Outcomes were analysed using generalized linear mixed models for repeated measures with identity link (continuous outcomes) and log-link (binary outcomes). All reported results are for the trial product estimand, unless otherwise specified.

Results: From Jan 2013 until Feb 2018, 109 patients were randomised to GBP (n=54) or SG (n=55). Ninety-three (85%) patients completed 5-yr follow up (last visit Dec 2022). Three SG patients were converted to GBP due to insufficient weight loss at 2-yr (n=1) and 3-yr (n=2). None of these were in remission at conversion, whilst all three were in remission at 5-yr. Rates of patients achieving diabetes remission were higher after GBP compared with SG: Treatment policy estimand; 50% after GBP vs 26% after SG, RR= 1.79 [95% CI 1.05 to 3.06; p=0.032]. Trial product estimand; 50% after GBP vs 20% after SG, RR= 2.44 [95% CI 1.28 to 4.68; p=0.007]. Remission rates were also higher in GBP than SG when applying the new consensus definition for remission (HbA1c <6.5% without antidiabetic drugs), 63% after GBP vs 30% after SG, RR= 2.13 [95% CI 1.29 to 3.54; p=0.003]. HbA1c decreased similarly in both groups. Body weight loss was 22% [95% CI 20 to 24] after GBP vs 17% [95% CI 15 to 19] after SG, treatment difference 5.0% [95% CI 2.4 to 7.7; p<0.001]. Weight-related quality of life (IWQOL-Lite) score improved 34 [95% CI 29 to 38] after GBP vs 28 [95% CI 23 to 32] after SG, difference 6.2 [95% CI 0.02 to 12.5; p=0.049]. Adverse events (≥6 weeks) occurred in 40 patients (86 events) after GBP and 37 patients after SG (93 events). One death occurred in the SG group (malignancy).

Conclusion: This trial showed GBP to be superior at inducing remission of T2DM, providing weight loss, and increasing quality of life compared with SG during 5-yr follow-up.

Clinical Trial Registration Number: NCT01778738

Supported by: Educational grant (PhD) from South-Eastern Norway Regional Health Authority

Disclosure: J.W. Hauge: None.

46

Synergistic effects of tirzepatide and pioglitazone on insulin sensitivity and beta cell function in diet-induced obese rats

M. Abu-Farha ¹, I. Al-Khairi¹, M. Qaddoumi², S. Al-Sabah³, A. AlSabbagh¹, V. Lopez¹, T. Alessa¹, J. Kumar¹, N. Abukhalaf¹, P. Cherian¹, S. Al-Sabah⁴, F. Al-Mulla¹, J. Abubaker¹, M. Abdul-Ghani⁵;

¹Dasman Diabetes Institute, Kuwait City, Kuwait, ²Pharmacology and Therapeutics Department, Kuwait University, Kuwait City, Kuwait, ³Kuwait Ministry of Health, Kuwait City, Kuwait, ⁴Department of Pharmacology and Toxicology, Kuwait University, Kuwait City, Kuwait, ⁵University of Texas, San Antonio, TX, USA.

Background and aims: Tirzepatide, a dual GLP-1 and GIP receptor agonist, augments glucose-stimulated insulin secretion, significantly lowers HbA1c, and promotes weight loss. The improvement in insulin sensitivity is thought to be mediated by both weight loss-dependent and weight loss-independent mechanisms. Given that Tirzepatide and pioglitazone improve insulin sensitivity via distinct mechanisms (GIP agonist vs. PPARγ activation), we hypothesized that their combination would produce additive improvements in insulin action.

Materials and methods: A total of 32 male Sprague Dawley rats aged 6-8 weeks were fed a high-fat cafeteria diet for 12 weeks and then divided into four groups treated with either placebo, 10 mg/kg pioglitazone, 100 nmole/kg Tirzepatide, or a combination of pioglitazone and Tirzepatide for 8 weeks. Weight, FBG, and calorie consumption were monitored weekly. OGTT and ITT tests were performed at the end of the study, and liver, adipose tissue, muscle, and pancreatic tissues were collected.

Results: Body weight significantly increased during the 8-week treatment period in animals receiving vehicle (+62±3 grams) and pioglitazone (+83±4 grams) (both p<0.01 versus baseline), also significant for pioglitazone vs vehicle (p<0.05). Conversely, body weight significantly decreased in animals receiving tirzepatide (-67±3 grams) and in animals receiving tirzepatide plus pioglitazone (-75±3 grams). Despite the opposing effect on body weight, the incremental increase area under the plasma glucose concentration during the OGTT (∆G0-120) was comparable in animals receiving tirzepatide or pioglitazone (7.2±0.5, 8.1±0.6 mM.h) (p=ns), and it was significantly smaller than that in animals receiving vehicle (11.0±0.9 mM.h) (both p<0.01). Additionally, animals receiving tirzepatide plus pioglitazone experienced robustly greater insulin sensitivity index than animals receiving pioglitazone or tirzepatide monotherapy (both p<0.01). finally, animals receiving tirzepatide plus pioglitazone had the smaller fat cell size than animals receiving tirzepatide monotherapy., suggesting an additive effect of pioglitazone and tirzepatide on fat cell function.

Conclusion: Tirzepatide mitigated the weight gain associated with pioglitazone, and the combination exhibited a synergistic improvement in insulin action. This study supports the potential of combinational therapy with Tirzepatide and pioglitazone for enhancing insulin sensitivity in the context of obesity and diabetes.

Disclosure: M. Abu-Farha: None.

47

High-throughput proteomics of human adipose tissue reveals exercise and type 2 diabetes induced changes

R. Kruse ^1,2, J.K. Larsen³, N. Sahebekhtiari¹, R.M. Justicia³, G. Gomez³, M.H. Petersen¹, M.E. de Almeida⁴, N. Ørtenblad⁴, A.S. Deshmukh³, K. Højlund¹;

¹Steno Diabetes Center Odense, Odense University Hospital, Odense C, Denmark, ²Department of Clinical Research and Department of Molecular Medicine, University of Southern Denmark, Odense C, Denmark, ³University of Copenhagen, Novo Nordisk Foundation Center for Basic Metabolic Research, Copenhagen, Denmark, ⁴Department of Sports Science and Clinical Biomechanics, University of Southern Denmark, Odense M, Denmark.

Background and aims: White adipose tissue (WAT) plays a critical role in whole-body energy homeostasis, including insulin sensitivity; however, the proteome of human WAT has only scarcely been investigated. Moreover, it remains largely unknown whether adaptations in the human WAT proteome contributes to the exercise-induced improvements in metabolic health, including insulin sensitivity, in humans. Here, we examined the WAT proteome in obese men with type 2 diabetes (T2D) and activity-matched lean and obese men before and after eight weeks of high-intensity interval training (HIIT).

Materials and methods: Using a high-throughput and reproducible mass spectrometry-based proteomics pipeline, we compared the proteomic signatures of abdominal subcutaneous WAT biopsies collected from glucose-tolerant lean (n=18) and obese (n=15) individuals and obese patients with T2D (n=15) at baseline and in response to eight weeks of HIIT combining rowing and cycling.

Results: We identified a total of 3.773 proteins of which 2.106 were quantified in each sample for differential analysis of protein abundance. At the single protein level, we found 48 proteins to be significantly regulated (25 up-regulated, 23 down-regulated) in T2D versus lean men at baseline after correction for multiple testing. These changes, together with gene set enrichment analysis, demonstrated; (1) a down-regulation of proteins involved in several mitochondrial processes including oxidative phosphorylation, tricarboxylic acid (TCA) cycle, and fatty acid metabolism and (2) up-regulation of proteins involved in immune and stress responses as well as extracellular matrix (ECM) components in WAT of men with T2D. Although the HIIT intervention markedly improved insulin sensitivity as well as VO₂max and body composition in all groups, it only had minor effects on the WAT proteome. Interestingly, HIIT induced a substantial increase in WAT ferritin subunit levels, a protein involved in iron metabolism, in all groups, and these increases were strongly correlated with individual insulin sensitivity (both p<0.001). On the contrary, HIIT caused a reduction in serum ferritin levels in all groups, and this was negatively correlated with insulin sensitivity. Furthermore, gene set enrichment analysis revealed that gene ontology (GO)-terms related to mitochondria were down-regulated in the lean and obese men following HIIT.

Conclusion: In the present study, we report baseline differences in the WAT proteome between lean, obese, and T2D men, which extend previous transcriptomic signatures of adipose tissue dysfunction in T2D. Importantly, our results emphasize a previously unknown role of human WAT iron metabolism in exercise training adaptations and association to insulin sensitivity.

Clinical Trial Registration Number: NCT03500016

Supported by: EFSD/NNF Future Leaders Awards Programme, NNF, Odense University Hospital, University of Southern Denmark

Disclosure: R. Kruse: None.

48

GUBamy: a novel long-acting amylin analogue with significant therapeutic potential for mono- and combination therapy in obesity

M. Tozzi, M. Cavalera, N. Sonne, L. Holm, M. Ougaard, S. Rasmussen, H.B. Hansen, M. Kræmer, T. Porsgaard, N. Vrang, S. Jørgensen;

Gubra, Copenhagen, Denmark.

Background and aims: The pancreatic hormone amylin plays a central role in regulating energy homeostasis and glycemic control. Amylin potently reduces appetite making amylin mimetics attractive for the treatment of metabolic diseases. Here, we evaluated preclinical anti-obesity effects of a long-acting injectable amylin analogue (GUBamy) alone or in combination with a GLP-1 receptor agonist.

Materials and methods: For pharmacokinetic analysis, chow-fed male Sprague-Dawley (SD) rats were given a single intravenous (IV) dose of GUBamy (15 nmol/kg, n=3) or semaglutide (50 nmol/kg, n=3). Effects on real-time food intake were characterized in chow-fed male SD rats (n=5-6 per group) for 96h after administration of a single subcutaneous (SC) dose of vehicle or GUBamy (1, 3, 10 or 30 nmol/kg). Effects on body weight, food intake and body composition were investigated in diet-induced obese (DIO, 45% kcal fat) male SD rats (n=8 per group) after SC (QD) administration of vehicle, GUBamy (10 nmol/kg,) or semaglutide (10 nmol/kg) as monotherapy and combination treatment, respectively, for 2 weeks.

Results: GUBamy showed circulating half-life (T½_iv) and mean residence time (MRT) of 29h and 42h, respectively. For semaglutide, T½_iv and MRT was 10h and 13h. GUBamy markedly and dose-dependently suppressed cumulative food intake after single dosing, with all doses showing sustained anorectic efficacy over the entire 96h monitoring period (p<0.01). GUBamy and semaglutide monotreatment induced a robust and sustained body weight loss in DIO rats compared to baseline (GUBamy, 10.0 ± 0.6%; semaglutide 7.4 ± 0.7%, both p<0.001). Notably, GUBamy and semaglutide combination therapy promoted additive effects on weight loss (16.0 ± 1.0%, p=0.001 vs. monotreatment with semaglutide or with GUBamy) and food intake inhibition in DIO rats. For mono- and combination therapy, body weight loss following GUBamy and semaglutide subchronic administration was driven by reduced fat mass.

Conclusion: Our preclinical data demonstrates that GUBamy is a potent long-acting amylin peptide analogue with marked anti-obesity effects in DIO rats holding potential for combination therapy with GLP-1 receptor agonists. GUBamy is currently in development as a once-weekly, injectable amylin analogue for the treatment of obesity.

Disclosure: M. Tozzi: None.

OP 09 Treatment beyond metformin

49

The genetics of the response to second-line drug therapies for type 2 diabetes: genome-wide association studies in the GRADE clinical trial

L. Szczerbinski ¹, J.H. Li², H. Liu³, M. Tripputi³, A. Kretowski¹, S.E. Kahn⁴, J.M. Lachin³, N. Younes³, J.M. Mercader⁵, J.C. Florez², the GRADE Research Group;

¹Medical University of Bialystok, Bialystok, Poland, ²Diabetes Clinical Center, Massachusetts General Hospital, Boston, MA, USA, ³Department of Biostatistics and Bioinformatics, The Biostatistics Center, George Washington University, Rockville, MD, USA, ⁴Department of Medicine, VA Puget Sound Health Care System and the University of Washington, Seattle, WA, USA, ⁵Broad Institute of MIT and Harvard, Cambridge, MA, USA.

Background and aims: Adjuvant therapy optimization for metformin-based pharmacotherapy of type 2 diabetes (T2D) can be a challenge for practitioners. Currently, drug selection is largely algorithmic and does not target therapy to individual patients based on their genetic profiles. To address this gap, we performed the largest genome-wide association study (GWAS) to date in a randomized clinical trial for long-term response to insulin glargine U-100, glimepiride, liraglutide, or sitagliptin, when added to metformin.

Materials and methods: We analyzed data from the Glycemia Reduction Approaches in Type 2 Diabetes: A Comparative Effectiveness (GRADE) Study, a randomized head-to-head trial in patients with T2D from diverse ancestries, comparing the metabolic effects of the four medications adjuvant to metformin with a mean follow-up of 5 years. 4,689 participants underwent genome-wide genotyping with the Illumina Infinium GSA 3.0 array, and high-quality imputation was performed with the TOPMed reference panel, resulting in 15.5 million variants. We performed a GWAS of treatment failure, defined by meeting GRADE outcome criteria (primary: HbA1c ≥7%; secondary: HbA1c ≥7.5%) We implemented Cox proportional hazards regression models, adjusting for baseline HbA1c, BMI, age, sex, and 10 ancestry principal components, to test associations with the incidence of primary and secondary outcomes in each treatment arm. Variants were filtered to minor allele count ≥5 and imputation quality R² ≥0.80.

Results: We identified multiple loci associated with treatment response, defined as meeting the primary (glargine: 12 loci; glimepiride: 16 loci; liraglutide: 32 loci; sitagliptin: 7 loci) or secondary (glargine: 13 loci; glimepiride: 7 loci; liraglutide: 15 loci; sitagliptin: 13 loci) outcomes, at standard genome-wide significance (p<5.0×10^-8). Among them, we found that the T allele of rs115566325, located near the Zinc finger protein 217 gene (ZNF217), was associated with an increased risk of treatment failure to liraglutide (primary outcome: HR 2.15, p=4.6×10^-8; secondary outcome: HR 2.66, p=1.7×10^-9). Other interesting associations were found for response to glimepiride, where the C allele of rs78324530 located near IQCA1 (primary outcome: HR 3.54, p=2.9×10^-8; secondary outcome: HR 4.83, p=8.3×10^-10) and the A allele of rs11925227 near SLC2A2 encoding the glucose transporter 2 (primary outcome: HR 1.45, p=4.9×10^-9; secondary outcome: HR 1.31, p=5.2×10^-4), were associated with an elevated risk of treatment failure.

Conclusion: We identified new genetic loci associated with response to second-line adjuvant medications for T2D. The findings highlight the potential of genetic variation to guide therapy selection. Independent validation and translational research are needed to bring these findings to the bedside.

Clinical Trial Registration Number: NCT01794143

Supported by: NIDDK: U01DK098246; U34DK088043; R01DK123019; ADA; NHLBI, CDC, Polish Ministry of Health

Disclosure: L. Szczerbinski: Grants; Ministry of Health of the Republic of Poland (project entitled “Center for Artificial Intelligence at the Medical University of Bialystok”).

50

Effect of glucagon-like-peptide 1 analogues and sodium-glucose transport 2 inhibitors on MACE in an elderly population: a target trial emulation in Danish register data

V. Kosjerina ¹, M. Parsa¹, B. Carstensen¹, M.Z. Ankarfeldt², K.K. Sorensen³, B. Brock¹, H.R. Christensen⁴, J. Rungby¹, D. Vistisen¹, K. Clemmensen⁵;

¹Steno Diabetes Center Copenhagen, Herlev, Denmark, ²Department of Clinical Pharmacology and Center for Clinical Research and Prevention, Bispebjerg Hospital, Copenhagen, Denmark, ³Department of Cardiology, Nordsjaellands Hospital, Hillerod, Denmark, ⁴Department of Clinical Pharmacology, Bispebjerg Hospital, Copenhagen, Denmark, ⁵Novo Nordisk, Soborg, Denmark.

Background and aims: The elderly population is often underrepresented in randomized controlled trials (RCT). It has been suggested that explicitly emulating a hypothetical RCT (target trial) reduces bias in real-world data analysis. The aim of this study is to estimate the effect of glucagon-like peptide-1 analogues (GLP-1) and sodium-glucose transport 2 inhibitors (SGLT2i) on 3-point major adverse cardiovascular events (MACE), in a real-world elderly population with type 2 diabetes (T2D) by emulating a hypothetical target trial.

Materials and methods: By linking Danish nationwide registers, we identified a population of elderly individuals with a T2D diagnosis that were new users of either GLP-1, SGLT2i or dipeptidyl peptidase-4 inhibitors (DPP4i) at an age of ≥70 years. DPP4i was used as an active comparator, as it has shown a neutral effect on 3-point MACE. Key exclusion criteria were a MACE event (defined as myocardial infarction, stroke, or all-cause mortality) in the 6 months prior to index date or use of any of the study drugs 90 days prior index date. Propensity scores were calculated on baseline covariates using multinomial logistic regression. In an intention-to-treat analysis, we analyzed the incidence of MACE by fitting a Poisson regression model with the index drug as exposure and natural splines logit of the propensity score for GLP1 and SGLT2i and age at follow-up. In a second Poisson model, we allowed for an interaction between age and index drug.

Results: A total of 37,493 individuals met the eligibility criteria (GLP-1 [N = 5,683], SGLT2i [N = 5,000], and DPP4i [N = 26,810]). The age at T2D diagnosis was higher in the DPP4i group (mean [SD]) 69 (8) years compared to GLP-1 63 (8) years and SGLT2i (SD) 65 (8) years. We found a significant reduction in 3-point MACE incidence for GLP-1 with a rate ratio (RR) of 0.90 (95% CI 0.85-0.95) and for SGLT2i 0.83 (95% CI 0.78-0.90), both compared to DPP4i. Figure 1 illustrates the age-specific rate ratios from the interaction model for 3-point MACE, for GLP1 and SGLT2i vs DPP4i.

Conclusion: Both GLP-1 and SGLT2i reduced the incidence of MACE compared to DPP4i in this real-world elderly population with T2D. However, we found a reduction of this effect with increasing age at follow-up. Further analysis will include adjustments for relevant covariates during follow-up.

Disclosure: V. Kosjerina: None.

51

Lower risk of severe hypoglycaemia with SGLT2 inhibitor compared to DPP4 inhibitor use in subjects with type 2 diabetes: a propensity score-matched cohort study

S. Moon ¹, E. Kim², H.-S. Kwon³, K. Han⁴, Y.-B. Ahn⁵, K.-H. Song³;

¹Division of Endocrinology and Metabolism, Department of Internal Medicine, The Catholic University of Korea, Incheon, Republic of Korea, ²Internal Medicine, The Catholic University of Korea, Incheon, Republic of Korea, ³Division of Endocrinology and Metabolism, Department of Internal Medicine, The Catholic University of Korea, Seoul, Republic of Korea, ⁴Department of Statistics and Actuarial Science, Soongsil University, Seoul, Republic of Korea, ⁵Division of Endocrinology and Metabolism, Department of Internal Medicine, The Catholic University of Korea, Suwon, Republic of Korea.

Background and aims: Severe hypoglycemia has been linked to substantial health consequences such as cardiovascular events and mortality. This study investigated the risk of severe hypoglycemia associated with new-use sodium-glucose transport protein 2 (SGLT2) inhibitors compared with dipeptidyl peptidase-4 (DPP4) inhibitors.

Materials and methods: Data from the National Health Insurance Service of Korea was used to identify initiators of SGLT2 or DPP4 inhibitors from 2014 to 2017. By employing a 1:1 propensity score matching technique to control for confounding variables and guarantee a fair comparison, our study included new initiators of SGLT2 inhibitors (n=57,021) and DPP4 inhibitors (n=57,021). The Cox proportional hazards model estimated hazard ratios (HRs) with 95% confidence intervals (CIs) for developing severe hypoglycemia in the matched sample. Exploratory subgroup analyses assessed the consistency of the treatment effects on the primary outcome.

Results: During a 1-year follow-up, the incidence rate of severe hypoglycemia was 1.88 per 1,000 person-years in patients treated with SGLT2 inhibitors and 3.28 per 1,000 person-years in patients treated with DPP4 inhibitors. SGLT2 inhibitors were associated with a significantly lower risk of severe hypoglycemia (HR, 0.57; 95% CI, 0.45-0.73). Subgroup analyses showed that compared with their counterparts, SGLT2 inhibitors significantly decreased the risk of hypoglycemia in high-risk groups for hypoglycemia such as women, patients with peripheral artery disease, and patients on sulfonylurea.

Conclusion: The new use of SGLT2 inhibitors was associated with a 43% lower risk of severe hypoglycemia than DPP4 inhibitors. SGLT2 inhibitors may be safer in glycemic control than DPP4 inhibitors, especially in subjects at an increased risk of severe hypoglycemia.

Disclosure: S. Moon: None.

52

Benefit of dual therapy with GLP-1 RA and SGLT2i on renal outcomes in type 2 diabetes

K.K.B. Clemmensen ¹, B. Zareini², T. Gerds³, K.K. Sørensen², K. Kvist¹, J.-P. David¹, C. Torp-Pedersen²;

¹Novo Nordisk A/S, Søborg, Denmark, ²Department of Cardiology, Copenhagen University Hospital - North Zealand Hospital, Hillerød, Denmark, ³Section of Biostatistics, University of Copenhagen, Copenhagen, Denmark.

Background and aims: There is increased use of dual glucagon like peptide-1 receptor agonist (GLP-1 RA) and sodium-glucose co-transporter-2 inhibitor (SGLT2i) therapy, but real-world evidence on the renal benefit is lacking. The aim is to compare the dual use of GLP-1 RA and SGLT2i to other dual second-line type 2 diabetes therapies with respect to renal outcomes.

Materials and methods: From 2010-2021 patients from the Danish nationwide registries were followed from start of dual second-line type 2 diabetes treatment. Outcomes were chronic renal disease, end-stage renal disease and >50% eGFR decrease from baseline. The estimated risk following 1 of 4 dual therapy combinations was determined using a longitudinal causal inference framework assuming all patients followed one dual therapy for 5 years.

Results: In total 87,201 persons were included (GLP-1 RA and SGLT2i: 14,831, GLP-1 RA and DPP4/SU/TZD: 20,417, SGLT2i and DPP4/SU/TZD: 22,803, dual DPP4/SU/TZD: 29,150). The 5-year risk ratio (95%CI) of dual GLP-1 RA and SGLT2i therapy compared to reference (dual DPP4/SU/TZD) for chronic renal disease: 1.03 (0.97;1.08), end-stage renal disease: 0.12 (0.09;0.15) and >50% decrease in eGFR: 0.52 (0.45;0.59).

Conclusion: Dual therapy with GLP-1 RA and SGLT2i compared to other dual therapies results in greater risk reduction of end-stage renal disease and eGFR decline, while dual SGLT2i and DPP4/SU/TZD showed greater risk reduction of chronic renal disease.

Disclosure: K.K.B. Clemmensen: Stock/Shareholding; Share holder in Novo Nordisk.

53

Benefit of dual therapy with GLP-1 RA and SGLT2i on cardiovascular outcomes in type 2 diabetes

B. Zareini ¹, T. Gerds², K.K. Sørensen¹, K.K.B. Clemmensen³, K. Kvist³, J.-P. David³, C. Torp-Pedersen¹;

¹Department of Cardiology, Copenhagen University Hospital - North Zealand Hospital, Hillerød, Denmark, ²Section of Biostatistics, University of Copenhagen, Copenhagen, Denmark, ³Novo Nordisk A/S, Søborg, Denmark.

Background and aims: Evidence on the cardiovascular benefit of dual glucagon like peptide-1 receptor agonist (GLP-1 RA) and sodium-glucose co-transporter-2 inhibitor (SGLT2i) therapy is lacking. The aim is to compare the benefit of dual GLP-1 RA and SGLT2i therapy to other dual type 2 diabetes therapies with respect to heart failure and major cardiovascular adverse events (MACE).

Materials and methods: From 2010-2021 patients from the Danish nationwide registries were followed from start of dual second-line type 2 diabetes treatment. Primary outcome was heart failure, and secondary outcomes were MACE and all-cause death. The estimated risk following 1 of 4 dual therapy combinations was determined using a longitudinal causal inference framework assuming all patients followed one dual therapy for 5 years.

Results: A total of 87,201 persons were included (GLP-1 RA and SGLT2i: 14,831, GLP-1 RA and DPP4/SU/TZD: 20,417, SGLT2i and DPP4/SU/TZD: 22,803, dual DPP4/SU/TZD: 29,150). The 5-year risk ratio (95% CI) of dual GLP-1 RA and SGLT2i therapy compared to reference (dual DPP4/SU/TZD) for heart failure was: 0.93 (0.87;1.00), for MACE: 0.91 (0.87;0.95) and for all cause death: 0.78 (0.74;0.82), see Figure.

Conclusion: Dual therapy with GLP-1 RA and SGLT2i showed a benefit on heart failure compared to other dual therapies. Dual therapy with GLP-1 RA and SGLT2i showed a benefit on MACE and all-cause death compared to the reference treatment.

Disclosure: B. Zareini: None.

54

Precision medicine in type 2 diabetes: development and validation of a treatment selection algorithm for SGLT2-inhibitors and GLP1-receptor agonist therapies

P. Cardoso ¹, A.T. Nair², K.G. Young¹, R. Hopkins¹, A.G. Jones¹, N. Sattar³, R.R. Holman⁴, E.R. Pearson², A.T. Hattersley¹, B.M. Shields¹, T.J. McKinley¹, J.M. Dennis¹, on behalf of the MASTERMIND consortium;

¹University of Exeter, Exeter, UK, ²University of Dundee, Dundee, UK, ³University of Glasgow, Glasgow, UK, ⁴University of Oxford, Oxford, UK.

Background and aims: Choosing between SGLT2-inhibitors (SGLT2i) and GLP1-receptor agonists (GLP1-RA) requires understanding each drug class's relative benefits and risks for individual patients. We develop and validate a treatment selection model to predict glucose-lowering efficacy of each therapy class for individuals with type 2 diabetes based on their routine clinical characteristics. We further evaluate the impact of glycaemia-based treatment targeting on short/long-term outcomes.

Materials and methods: State-of-the-art Bayesian Causal Forests were used to develop a treatment selection algorithm to predict individual-level differences in 12-month glycaemic response (change from baseline in HbA1c) for the two therapies based on routine clinical characteristics. Model development was conducted in UK primary care data (CPRD Aurum), including 27,319 patients with type 2 diabetes initiating SGLT2i or GLP1-RA (baseline characteristics: HbA1c 77.0±15.5 [SD] mmol/mol, age 58.4±10.4 years, 60% male). The model was validated by comparing predicted versus observed differential treatment effects in independent hold-back CPRD data (n=20,850) and Scottish primary care data (Tayside & Fife, n=2,252). In CPRD, we explored the impact of targeting treatment based on HbA1c on 12-month weight change, 6-month discontinuation (a proxy of tolerability).

Results: Although glycaemic response was on average similar on both therapies (mean 12-month response SGLT2i 12.1 mmol/mol; GLP1-RA 12.3 mmol/mol, marked individual-level heterogeneity in response was evident, with 2,282 individuals predicted a >5mmol/mol benefit on SGLT2i (SGLT2i-optimal subgroup), and 3,376 individuals predicted >5mmol/mol benefit on GLP1-RA (GLP1RA-optimal subgroup). Compared to GLP1RA-optimal patients, SGLT2i-optimal patients were more male (86% versus 20%), younger (mean age 48.8 versus 66.1), with higher baseline HbA1c (90 versus 73 mmol/mol) and eGFR (109 versus 81 ml/min/1.73m2). The model performed well in CPRD hold-back validation, with average treatment benefits (versus alternative treatment) of 5.4 (CI: 3.4-7.4) mmol/mol with SGLT2i in the SGLT2i-optimal subgroup and 3.6 (CI: 1.7-5.5) mmol/mol with GLP1-RA in the GLP1RA-optimal subgroup, and in Scotland (SGLT2i-optimal group benefit 7.4 (CI: 0.1-14.1) mmol/mol; GLP1RA-optimal subgroup benefit 5.6 (CI: 0.8-12) mmol/mol). Weight change did not differ by predicted glycaemic benefit. Short-term discontinuation was highest for individuals predicted to have a glycaemic benefit from GLP1-RA but who received SGLT2i (GLPRA-optimal subgroup 6-month discontinuation: SGLT2i 26.5%, GLP1-RA 18.9%).

Conclusion: A validated treatment selection model for glycaemic response supports individualised treatment decisions when choosing between SGLT2i and GLP1-RA therapy. Optimising treatment choice based on glycaemia may also improve short-term tolerability and lower longer-term risk of new-onset microvascular disease. Precision type 2 diabetes medicine based on routine clinical features presents a low-cost and equitable approach to optimising clinical care applicable worldwide.

Supported by: EFSD Rising Star Fellowship Programme, UK MRC (MR/N00633X/1), NIHR Exeter BRC

Disclosure: P. Cardoso: None.

OP 10 Activating two or three G´s; one of us is lonely

55

CT-388, a novel once-weekly dual GLP-1 and GIP receptor modulator, is safe, well-tolerated, and produces more than 8% weight loss in 4 weeks in overweight and obese adults

M.V. Chakravarthy ¹, F.A. Arguelles-Tello², A.A. Sun³, M. Elliott¹, L. Acosta¹, J.E. Rankin⁴, S.K. Hansen¹;

¹Carmot Therapeutics, Berkeley, CA, USA, ²Avante-Sante Research Center, San Pedro Garza García, Mexico, ³Linear Clinical Research, Perth, Australia, ⁴Syneos Health, Adelaide, Australia.

Background and aims: GLP-1 and GIP can deliver complementary pharmacology when combined. CT-388 is a biased dual GLP-1 and GIP receptor modulator that exhibits no to minimal β-arrestin coupling at either receptor, designed for once-weekly subcutaneous administration. Primary objective was to investigate safety and tolerability of CT-388.

Materials and methods: A Phase 1, placebo-controlled, randomized, double-blind study in three parts was conducted in overweight and obese adults. Part 1 was a single ascending dose (SAD) study with doses ranging from 0.5 mg to 7.5 mg. Part 2 was a 4-week multiple ascending dose (MAD) study comprised of 3 independent cohorts with doses ranging from 5-12 mg administered via weekly up-titration. Part 3 is a 12-24 week multiple dose study with 3 additional cohorts (with one of the cohorts comprised of obese adults with T2D) with doses up 22 mg administered via titration. The current abstract outlines the results from Parts 1 and 2. Additional obese participants in Part 3 have completed enrollment in the 12-week cohort of CT-388/placebo dosing; treatment is currently ongoing, and results from these cohort(s) are anticipated to be available for presentation at the time of the EASD meeting.

Results: In parts 1 and 2, a total of 64 participants [N=28 (men)/36 (women); median age=34 years; median BMI=33 kg/m²) received at least 1 dose of CT-388 or placebo. Pharmacokinetic profile was investigated over a wide dose range (0.5-12 mg) and supports once-weekly administration. Percent change in body weight from baseline at Day 29 was dose responsive and significantly greater in CT-388 treated participants versus placebo (p <0.0001) across the 3 MAD cohorts with placebo-adjusted Least Square Mean difference [95% CI] of -4.8% [-6.3, -3.3], -6.4% [-7.8, -5.0], and -8.5% [-10.4, -6.7]. Mean percent decrease from baseline at Day 23 in fasting glucose (↓8-10%), insulin (↓20-26%), HOMA-IR (↓26-33%), AUC_0-120min glucose (↓26-28%) and AUC_0-120min insulin (↓33-58%) during oral glucose tolerance tests were seen in cohorts dosed up to 12 mg of CT-388, suggestive of improved insulin sensitivity. Over the 4-week treatment period, CT-388 was generally well tolerated without any treatment-related discontinuations in the MAD cohorts. The most frequent side effects reported were gastrointestinal (decreased appetite, nausea, vomiting, diarrhea), which were mostly mild in severity.

Conclusion: In summary, CT-388 delivers clinically meaningful weight loss and metabolic control with a favorable tolerability profile. These data warrant further clinical evaluation of CT-388, possibly with minimal to no titration, for the treatment of obesity, T2DM, and other weight-related comorbidities.

Clinical Trial Registration Number: NCT04838405

Disclosure: M.V. Chakravarthy: Employment/Consultancy; employment.

56

Retatrutide, a triple GIP/GLP-1/glucagon receptor agonist, provides robust HbA _1c and body weight reductions to people with type 2 diabetes: a 36-week, phase 2 study

T. Coskun ¹, J. Rosenstock², J. Frias³, A. Jastreboff⁴, Y. Du¹, J. Lou¹, S. Gurbuz¹, M.L. Hartman¹, A. Haupt¹, Z. Milicevic¹;

¹Eli Lilly and Company, Indianapolis, IN, USA, ²Velocity Clinical Research, Dallas, TX, USA, ³Velocity Clinical Research, Los Angeles, CA, USA, ⁴Yale School of Medicine, New Haven, CT, USA.

Background and aims: Along with improving glycaemic control, body weight reduction is increasingly recognised as central to the treatment of type 2 diabetes (T2D). Retatrutide (RETA) is a single peptide with agonist activity at the glucose-dependent insulinotropic polypeptide (GIP), glucagon-like peptide-1 (GLP-1), and glucagon receptors. We examined the efficacy and safety of RETA vs placebo (PBO) in people with T2D.

Materials and methods: In this 36-week, phase 2, randomised, double-blinded, double-dummy, parallel, placebo- and active comparator-controlled study, participants with T2D treated with diet and exercise alone or with a stable dose of metformin, an HbA_1c (glycated haemoglobin) ≥53 to ≤91 mmol/mol (≥7% to ≤10.5%), and body mass index (BMI) of 25 to 50 kg/m² were eligible to participate. Participants were randomised to PBO, dulaglutide 1.5 mg (DULA), or maintenance doses of RETA 0.5 mg, 4 mg, 8 mg, or 12 mg. Gradual dose escalation was employed to improve tolerability; two starting doses (2 mg vs 4 mg) were explored for doses greater than 0.5 mg. Treatments were administered once weekly via subcutaneous injection. The primary endpoint was change in HbA_1c from baseline to 24 weeks, and secondary endpoints included change in HbA_1c and body weight at 36 weeks, and several additional measures relevant for cardiovascular risk.

Results: At 24 weeks, RETA decreased HbA_1c in a dose-dependent manner, ranging from -4.7 to -22.1 mmol/mol (-0.4 to -2.0%) vs -0.1 mmol/mol (<0.1%) with PBO and -15.4 mmol/mol (-1.4%) with DULA. Findings were consistent for HbA_1c at 36 weeks [Table]. The proportion of participants achieving HbA_1c <7.0%, ≤6.5%, or <5.7% at 36 weeks was higher in the RETA-treated participants vs PBO. Body weight decreased dose-dependently with RETA at 36 weeks, with 16.9% reduction with RETA 12 mg. Fasting glucose, waist circumference, non-HDL cholesterol, triglycerides, and systolic blood pressure also decreased to a greater extent with RETA vs PBO. Mild-to-moderate gastrointestinal-related adverse events, including nausea, diarrhoea, vomiting, and constipation, were reported in 13-50% of RETA-treated participants, 13% with PBO, and 35% with DULA; these events occurred primarily during dose-escalation, and were mitigated by a lower starting dose (2 mg vs 4 mg). There were no reports of clinical meaningful or severe hypoglycaemia. Dose-dependent increases in pulse rate of up to 3.9 bpm were observed with RETA at 36 weeks, vs -3.2 bpm for PBO (p<0.05) and 1.8 bpm for DULA.

Conclusion: In people with T2D, RETA demonstrated clinically meaningful improvements in glycaemic control and robust clinically meaningful body weight loss vs PBO and dulaglutide, with a safety profile consistent with GLP-1 receptor agonist and GIP/GLP-1 receptor agonist classes.

Clinical Trial Registration Number: NCT04867785

Supported by: Eli Lilly and Company

Disclosure: T. Coskun: Employment/Consultancy; Employed by Eli Lilly and Company. Stock/Shareholding; Minor shareholder of Eli Lilly and Company.

57

Once-weekly triple receptor (GIP/GLP-1/glucagon) agonist retatrutide (LY3437943): efficacy and safety in a 48-week obesity phase 2 trial

A.M. Jastreboff ¹, L.M. Kaplan², J.P. Frias³, Q. Wu⁴, Y. Du⁴, S. Gurbuz⁴, T. Coskun⁴, A. Haupt⁴, Z. Milicevic⁴, M.L. Hartman⁴;

¹Yale School of Medicine, New Haven, CT, USA, ²Harvard Medical School, Boston, MA, USA, ³Velocity Clinical Research, Los Angeles, CA, USA, ⁴Eli Lilly and Company, Indianapolis, IN, USA.

Background and aims: Nutrient-stimulated hormone-based pharmacotherapeutics are advancing the treatment of obesity. Retatrutide (RETA) is a novel triple-agonist of the GIP, GLP-1 and glucagon receptors. A phase 1b study of RETA demonstrated up to 9.0 kg placebo-adjusted weight reduction over 12 weeks in participants with type 2 diabetes. We aimed to examine the efficacy and safety of RETA versus placebo for weight reduction in adults with obesity.

Materials and methods: This phase 2 trial randomly assigned adults aged 18-75 years with BMI ≥30 kg/m², or ≥27 kg/m² and ≥1 weight-related condition (type 2 diabetes excluded), to 48 weeks of once-weekly subcutaneous RETA (1, 4, 8 or 12 mg maintenance doses) or placebo. Gradual dose escalation was employed to improve tolerability; two starting doses (2 mg and 4 mg) were explored for doses greater than 1 mg. Primary outcome was % weight change vs. placebo from baseline to 24 weeks. Secondary outcomes assessed at 48 weeks included % change in body weight vs. placebo, % of participants reaching weight reduction targets and several additional measures relevant for cardiovascular risk. Differences in body weight change were explored in subgroups based on sex and BMI.

Results: Participants included 338 adults (48.2% female) with mean age 48.2 years and BMI 37.3 kg/m². The mean % change in weight at 24 weeks was −7.2 (RETA 1 mg), −12.9 (4 mg), −17.3 (8 mg), and -17.5 (12 mg) and −1.6 (placebo) (all p<0.001 versus placebo). The mean % change in weight at 48 weeks was −8.7 (RETA 1 mg), −17.1 (4 mg), −22.8 (8 mg), and −24.2 (12 mg) and −2.1 (placebo) (all p<0.001 versus placebo). At 48 weeks, weight reduction ≥10% and ≥15% was achieved in 91 and 75% (RETA 8 mg), and 93 and 83% (12 mg) and in 9 and 2% (placebo) of participants (all p<0.001 versus placebo). Trends for greater % weight reduction were observed for women compared to men and for participants with BMI ≥35 kg/m² compared to those with BMI <35 kg/m². Improvements in waist circumference, lipids, HbA_1c, and blood pressure were observed with RETA. Most common adverse events were gastrointestinal (nausea, diarrhoea, vomiting), were mild-to-moderate in severity, occurred primarily during dose-escalation, and were mitigated by a lower starting dose (2 mg vs 4 mg). Dose-dependent increases in pulse rate (bpm) were observed at 48 weeks: 1.7 (RETA 1 mg), 3.1 (4 mg), 4.7 (8 mg), and 6.0 (12 mg) and 0.4 (placebo) (all except 1 mg group p<0.05 vs. placebo).

Conclusion: In participants with obesity, RETA was well-tolerated and provided substantial and clinically meaningful reductions in body weight and improvements in cardiovascular risk factors.

Clinical Trial Registration Number: NCT04881760

Supported by: Eli Lilly and Company

Disclosure: A.M. Jastreboff: Employment/Consultancy; Amgen; Scholar Rock. Grants; NIH/NIDDK. Other; scientific advisor for Boehringer Ingelheim, Eli Lilly, Intellihealth, Novo Nordisk, Pfizer, Rhythm Pharmaceuticals, Structure Therapeutics, Terns Pharmaceuticals, and WW (formerly WeightWatchers).

58

Efficacy of cotadutide dual GLP1-glucagon receptor agonist on albuminuria and glycaemic control in patients with diabetic kidney disease

D. Robertson ¹, V. Selverajah¹, L. Hansen², L. Jermutus¹, K. Smith¹, C. Frickleton², A. Coggi¹, J. Sanchez³, Y.-T. Chang², V.E.R. Parker¹;

¹AstraZeneca, Cambridge, UK, ²AstraZeneca, Gaithersburg, MD, USA, ³AstraZeneca, Gothenburg, Sweden.

Background and aims: Cotadutide is a dual GLP1-glucagon receptor agonist under development for NASH and DKD. Achieving optimal glycaemic control can be especially challenging in patients with DKD. We evaluated the effect of cotadutide on both renal and glycaemic endpoints in patients with DKD

Materials and methods: In this randomised, double-blind, phase 2b study, patients with DKD on insulin and/or oral therapy with HbA1c ≥ 6.5 and ≤ 10.5, eGFR ≥ 20 and < 90 ml/min/1.73m², urinary albumin creatinine ratio (UACR) ≥ 50 mg/g and BMI ≥ 25 (23 in Japan) kg/m² were treated for 26 weeks. Participants were randomised (n = 45 per arm) to receive once-daily SC cotadutide titrated up to 100, 300 or 600 μg, or placebo. Investigators were advised to adjust insulin doses as required and CGM was performed for 26 weeks. Study endpoints included UACR, measures of glucose control, and body weight

Results: The primary endpoint was met and dose-dependent reductions in UACR from baseline were observed after 26 weeks treatment with 300 and 600 μg of cotadutide, -39.9% (95% CI -52.5, -23.9) and -46.0% (95% CI -57.1, -32.1) vs placebo (P <0.001). Clinically and statistically significant reductions in HbA1c were seen at all dose levels (-0.89 -0.92%, p<0.05) alongside reduced 10-day average glucose, glycaemic variability, and increased time spent in target range. There was no significant increase in time spent in hypoglycaemia. Insulin doses were reduced by 13-25% and body weight loss adjusted for insulin dose reduction was -5.4% and -7.35% at 300 μg and 600 μg respectively at 26 weeks (p<0.003). SAEs were balanced across all arms and there were fewer AE related discontinuations at 100 and 300 ug versus placebo, but more discontinuations were observed at 600 μg in comparison to 100 μg, 300 μg, and placebo.

Conclusion: In patients with DKD, cotadutide promoted clinically important effects on UACR, glycaemic control, and body weight. The results suggest cotadutide has a potential to provide benefit on renal outcomes with holistic benefits on body weight and glycaemic control in patients with DKD, but larger studies are warranted.

Clinical Trial Registration Number: NCT04515849

Disclosure: D. Robertson: Employment/Consultancy; Employee of AZ. Stock/Shareholding; Shareholder at AZ.

59

A Phase II, randomised, double-blind, placebo-controlled, dose-finding study of BI 456906 in people living with overweight/obesity

C.W. Le Roux ¹, O. Steen², K.J. Lucas³, L. Borowska⁴, A. Unseld⁵, E. Startseva⁴;

¹School of Medicine, University College Dublin, Dublin, Ireland, ²LMC Diabetes & Endocrinology, Toronto, ON, Canada, ³Diabetes & Endocrinology Consultants PC, Morehead City, NC, USA, ⁴Boehringer Ingelheim International GmbH, Ingelheim, Germany, ⁵Boehringer Ingelheim Pharma GmbH & Co. KG, Biberach, Germany.

Background and aims: Approved treatments for obesity include agents targeting the glucagon-like peptide-1 receptor (GLP-1R), which result in reduced food intake and delayed gastric emptying. BI 456906 is a dual agonist that acts on glucagon receptors, to increase energy expenditure, and the GLP-1R, and may therefore improve therapeutic efficacy. This dose-finding study evaluated the efficacy and safety of BI 456906 in participants with overweight/obesity.

Materials and methods: This Phase II, double-blind, placebo-controlled study was conducted in adults with BMI ≥27 kg/m² and without diabetes. Eligible participants were randomised 1:1:1:1:1 to weekly subcutaneous BI 456906 (four dose groups: 0.6, 2.4, 3.6, 4.8 mg) or placebo. The 46-week treatment period comprised a 20-week dose escalation phase, with increases once every 2 weeks (fixed, Weeks 1-10, no dose adjustment permitted; flexible, Weeks 11-20 [dose adjustment permitted based on tolerability of gastrointestinal adverse events]), followed by a 26-week maintenance phase (fixed dose). The primary endpoint was bodyweight change (%) from baseline at Week 46 to characterise the dose-response relationship for BI 456906. Secondary endpoints included absolute changes in bodyweight (kg), waist circumference (cm), systolic BP (mmHg) and diastolic BP (mmHg) from baseline at Week 46. A mixed model for repeated measurements (MMRM) was used to generate covariate adjusted fixed-effect estimates of absolute change over time for each continuous secondary endpoint.

Results: In total, 387 participants were randomised (treated set [TS], N=386; full analysis set [FAS], N=384; n≈77 per arm). Baseline demographics and clinical characteristics at trial entry were similar between study arms (FAS): 68.2% of participants were female, 78.4% were White, mean (SD) age was 49.1 (12.9) years, BMI 37.1 (6.1) kg/m², bodyweight 105.7 (20.4) kg, waist circumference 113.4 (14.5) cm, systolic BP 125.6 (13.4) mmHg and diastolic BP 81.3 (7.8) mmHg. At Week 46, all tested BI 456906 doses yielded a substantial placebo-corrected reduction from baseline in absolute bodyweight and waist circumference; the greatest mean reductions were seen in the 4.8 mg dose group (−15.8 kg and −12.1 cm, respectively; Table). Substantial placebo-corrected reductions in BP were also achieved at Week 46; the greatest mean reductions were −6.2 mmHg for systolic BP (3.6 and 4.8 mg dose groups) and −2.9 mmHg for diastolic BP (4.8 mg dose group; Table).

Conclusion: Over 46 weeks, BI 456906 doses ≥2.4 mg substantially improved cardiometabolic outcomes in participants with overweight/obesity.

Clinical Trial Registration Number: NCT04667377

Supported by: The study and medical writing support (Anna Wydra, OPEN Health Communications) were funded by Boehringer Ingelheim.

Disclosure: C.W. Le Roux: Lecture/other fees; has received personal fees from Boehringer Ingelheim, GI Dynamics, Herbalife, Johnson & Johnson, Keyron, Eli Lilly and Novo Nordisk outside the submitted work.

60

Phase 2, randomised, placebo-controlled trial of pemvidutide, a GLP-1/glucagon dual receptor agonist, in subjects with overweight or obesity: a 24-week interim analysis

L. Aronne ¹, M.S. Harris², M.S. Roberts², J. Suschak², S. Tomah², L. He², J. Yang², J.P. Frias³, S.K. Browne²;

¹Weill Cornell Medicine, New York, NY, USA, ²Altimmune, Inc, Gaithersburg, MD, USA, ³Velocity Clinical Research, Los Angeles, CA, USA.

Background and aims: Pemvidutide is a long-acting GLP-1/glucagon (GCG) dual receptor agonist in development for treatment of obesity and non-alcoholic steatohepatitis (NASH). It combines the anorexigenic effects of GLP-1R agonism with the increased energy expenditure and lipometabolic effects of GCG receptor agonism in a 1:1 potency ratio. Robust weight loss has been observed in Phase 1 studies, with a 76% reduction in liver fat content in subjects with non-alcoholic fatty liver disease (NAFLD). We report a prespecified interim analysis of the MOMENTUM trial, an ongoing Phase 2, randomized, placebo-controlled trial of subjects with overweight and obesity, that was conducted when 160 of 320 planned subjects completed 24 of 48 weeks of treatment.

Materials and methods: Subjects with obesity (BMI > 30.0 kg/m²) or overweight (BMI 27.0-29.9 kg m²) with at least one obesity-related comorbidity were randomized 1:1:1:1 to pemvidutide (1.2, 1.8, 2.4 mg) or placebo administered subcutaneously weekly with adjunct lifestyle modification. The 1.2 and 1.8 mg doses were administered without dose titration. A short 4-week titration was employed at the 2.4 mg dose, but dose reduction for adverse events (AEs), as employed in trials of other GLP-1 based agents, was not allowed.

Results: Median age, body weight and BMI of subjects were 48 yrs, 100 kg, and 36 kg/m², respectively, with approximately 75% of subjects being female. All dose groups achieved significant weight loss at Week 24 (7.3%, 9.4% and 10.7% at 1.2, 1.8, and 2.4 mg, respectively vs. 1.0% in placebo, (Table 1). Approximately 50% achieved >10% and 20% achieved >15% weight loss at the 1.8 and 2.4 mg doses. Significant improvements or trends in cardiometabolic risk factors, including waist circumference and serum lipids, were observed at all doses, with a > 5 mm Hg decrease in systolic blood pressure in the absence of an increase in heart rate at the 2.4 mg dose. Mild to moderate nausea occurred at rates consistent with other agents in the GLP-1 class and accounted for most adverse events (AEs), with dose-related AE discontinuations mostly early in treatment. No significant changes in glucose homeostasis (fasting glucose, HbA1c) were observed at any dose.

Conclusion: Pemvidutide induced significant reductions in body weight and other cardiometabolic risk factors after 24 weeks of a 48-week treatment period. The combination of significant weight loss and liver fat reduction is particularly attractive for patients with obesity and at-risk NASH. As in trials of other GLP-1 based agents, AE discontinuations could be reduced in future trials by allowance for dose reduction.

Clinical Trial Registration Number: NCT05295875

Disclosure: L. Aronne: Employment/Consultancy; Allurion, Altimmune, Atria, Gelesis, Jamieson Wellness, Janssen Pharmaceuticals, Jazz Pharmaceuticals, Novo Nordisk, Pfizer, Optum, Eli Lilly, Senda Biosciences, Versanis. Grants; Allurion, AstraZeneca, Gelesis, Janssen Pharmaceuticals, Novo Nordisk, Eli Lilly. Stock/Shareholding; Allurion, ERX Pharmaceuticals, Gelesis, Intellihealth, Jamieson Wellness, Myos Corp. Other; ERX Pharmaceuticals, Intellihealth, Jamieson Wellness.

OP 11 Be(a)ware of hypoglycaemia unawareness

61

ZT-01, a somatostatin receptor 2 antagonist, increases glucagon response to hypoglycaemia in a phase 1b glycaemic clamp study in type 1 diabetes

A. Abitbol ¹, R. Liggins², S. Peers³, E. Simonson², M.C. Riddell⁴, M. Midmer³;

¹LMC Healthcare/Centricity Research, Toronto, ON, Canada, ²Zucara Therapeutics, Vancouver, BC, Canada, ³Zucara Therapeutics, Toronto, ON, Canada, ⁴York University, Toronto, ON, Canada.

Background and aims: A novel somatostatin receptor 2 antagonist, ZT-01, has the potential to prevent hypoglycemia by improving the defective counterregulatory response in T1D. This study evaluated the effect of ZT-01 on glucagon counterregulation during hypoglycemia in adults with Type 1 diabetes (T1D).

Materials and methods: ZT-01 pharmacodynamics were evaluated in a Phase 1b randomized, three-arm, double-blind crossover trial in 23 hypoglycemia aware adults with T1D. Participants underwent up to three hypoglycemic clamp procedures with 3 or 20 mg ZT-01 or placebo dosed 1-2 h prior to hypoglycemia in a stepwise glucose clamp using variable rate IV insulin infusion: euglycemia (5.0 ± 0.5 mmol/L, pre- and post-dose clamp periods A and B, respectively), level 1 (3.5 ± 0.3 mmol/L, clamp period C), and level 2 hypoglycemia (2.6 ± 0.2 mmol/L, clamp period D). Blood glucose and plasma glucagon were measured at 5 and 15 minute intervals, respectively, and 3 glucagon measures were determined in the final 15 minutes ending each clamp step, called the “stable phase”.

Results: Glycemic targets were achieved in the stable phase of each clamp step (figure A). The average clamp duration was 4 h 15 m with placebo, and 4 h 35 m for either ZT-01 dose, but not statistically different. Following ZT-01, stable phase glucagon increased by 8.2 ± 7.7 and 17.0 ± 12.3 pg/mL, in level 1 and level 2 hypoglycemia, respectively, compared with no increase in level 1 and a 5.1 ± 6.0 pg/mL increase in level 2 hypoglycemia with placebo (p<0.0001, figure B). The insulin dose required to induce hypoglycemia increased by 50% with ZT-01 (20 mg) compared to placebo (p=0.025). A counterregulatory response of increased glucose level coinciding with an increased glucagon level was observed more frequently with ZT-01 than placebo. No clinically significant effects on other counterregulatory hormones or hypoglycemia symptom scores were observed with ZT-01. A transient rise in glucagon and blood glucose was also observed post-dose following ZT-01 in euglycemia (period B). No SAEs were reported and no TEAEs were observed with ZT-01.

Conclusion: This study demonstrates proof-of-concept that ZT-01 significantly enhances the defective glucagon counterregulatory response in people with T1D during level 1 and level 2 hypoglycemia.

Clinical Trial Registration Number: NCT05007977

Disclosure: A. Abitbol: Grants; Zucara Therapeutics.

62

HIT4HYPOS CGM analysis: high intensity interval training reduces nocturnal hypoglycaemia in people with type 1 diabetes and impaired awareness of hypoglycaemia

C.M. Farrell ¹, G. Cappon², D. West³, A. Facchinetti², R.J. McCrimmon¹;

¹Systems Medicine, University of Dundee, Dundee, UK, ²Department of Information Engineering, University of Padova, Padova, Italy, ³Faculty of Medical Sciences, Newcastle University, Newcastle, UK.

Background and aims: The benefits of exercise are unquestionable and regular exercise is recommended for people with T1D. However, fear of hypoglycaemia is a well-recognised barrier. Impaired awareness of hypoglycaemia (IAH) in Type 1 diabetes (T1D) is a major risk factor for severe hypoglycaemia. Exercise is generally considered to increase the risk of hypoglycaemia, but this may not apply to all forms of exercise. High intensity interval training (HIIT) is an effective type of exercise where the effect on hypoglycaemia risk in people with T1D and IAH has not been reported to date. To address this question, we carried out a post-hoc analysis of continuous glucose monitoring (CGM) data collected during a 4-week pilot, single centre, randomised parallel-group study using HIIT in people with T1D and IAH: HIT4HYPOS.

Materials and methods: Individuals (n=18) underwent a 4-week run-in period of insulin optimisation, before randomisation to HIIT (3 session per week [4 x 30 s cycle sprints separated by 2 min of recovery] to reach an intensity corresponding to ≥ 90% peak heart rate (HR) reached during VO₂Peak assessment) for 4 weeks plus use of real-time continuous glucose monitoring (rtCGM), (HIIT+rtCGM) or Control (rtCGM alone). Participants in the HIIT+rtCGM group were given advice about insulin dose adjustment and carbohydrate intake for exercise based upon expert consensus guidance. HIIT sessions were carried out in a ‘real-world’ setting, unsupervised, at home or a gym, at a time that suited the participant. Raw CGM data were processed and analysed using the open-source software AGATA (Automated Glucose dATa Analysis), this allowed cleaning and data preparation, extract CGM-derived metrics in compliance with literature definitions and compare HIIT+rtCGM vs. rtCGM via ad-hoc statistical tests.

Results: Full CGM data sets were analysed for all participants over the 4-week intervention period. All participants in the HIIT+rtCGM group completed 12 HIIT sessions and were able to achieve ≥ 90% peak HR. There were significantly more level 1, glucose 3.9-3.0 mmol/L, (1.58 vs. 0.21 events/week, p < 0.05) and level 2, glucose < 3.0 mmol/L, nocturnal hypoglycaemic events (0.45 vs. 0.00 events/week, p < 0.05) in rtCGM compared with HIIT+rtCGM and duration was longer (level 1: 33.33 vs. 25.86 min, p = 0.34; level 2: 24.38 vs. 22.50 min, p = 0.50). We found no correlation between HIT and hypoglycaemia up to 4 hours post exercise. There was an increase in nocturnal time in hyperglycaemia (level 1: 23.16 vs. 26.55 %, p = 0.22; level 2: 7.11 vs. 14.86 %, p = 0.06) associated with HIIT. However, overall, during the 4-week intervention period, there was no significant difference in average glucose (8.92 vs. 8.92 mmol/l, p = 0.27), glucose mean indicator (7.15 vs. 7.16, p = 0.27), or glucose variability, in terms of coefficient of variation (36.24 vs. 35.66, p = 0.64), between the two groups.

Conclusion: These findings suggest that HIIT does not increase the risk of post exercise or nocturnal hypoglycaemia. There was no increase in hypoglycaemia following HIIT or significant deterioration in overall glycaemic control. Together these findings suggest that HIIT may be a safe form of exercise in people with T1D and IAH and that it may protect from nocturnal hypoglycaemia.

Clinical Trial Registration Number: ISRCTN15373978

Supported by: Diabetes UK and JDRF

Disclosure: C.M. Farrell: Grants; Diabetes UK 17/0005591, JDRF 3-SRA-2017-485-S-B.

63

Targeted metabolomics of impaired awareness of hypoglycaemia in individuals with type 1 diabetes

B.H.R. Wolffenbuttel ¹, R.D.M. Varkevisser¹, A. Cecil², C. Prehn², D. Mul³, H.-J. Aanstoot³, M.M. Van der Klauw¹;

¹Dept. of Endocrinology, University of Groningen, University Med Ctr Groningen, Groningen, Netherlands, ²Metabolomics and Proteomics Core, Helmholtz Zentrum München, Neuherberg, Germany,

³Diabeter, Center for Focussed Diabetes Care and Research, Rotterdam, Netherlands.

Background and aims: Impaired awareness of hypoglycaemia (IAH) is a complication of insulin therapy that affects 20-40% of individuals with type 1 diabetes. IAH is partly explained by the absence of the sympathoadrenal response and adaptations of the brain; however, the exact pathophysiology behind this attenuation is not yet fully understood. Metabolomics is a rapidly evolving field that uses high-throughput data to help identify patterns in disease states in a non-hypothesis-driven approach. In this study, we aimed to identify metabolic signatures in IAH to elucidate potential pathophysiological pathways.

Materials and methods: Individuals in the Dutch type 1 diabetes biomarker study who completed the Clarke questionnaire were included in this study. Individuals were excluded if they had cardiovascular disease or were non-western European. Plasma samples from individuals with IAH and age-sex-matched individuals without IAH (NAH) were analysed using the targeted Biocrates AbsoluteIDQ® p180 assay. Metabolite data were analysed using MetaboAnalyst 5.0. After quality control, metabolite data were normalised, and general linear models adjusted for sex and diabetes duration were fitted using the limma package in R. Genome-wide association studies (GWAS) were then conducted with metabolites with p <0.05 and SNPs reaching genome-wide significance were re-introduced in the linear models as confounders.

Results: In total, 578 individuals completed the Clarke questionnaire. After exclusion, 67 individuals with IAH and 108 with NAH were selected for analysis. The median age was 44 years, with a diabetes duration of 23 years. Insulin dosage was 50 units per day, and the mean HbA1c was 60mmol/mol (7.6%). When adjusted for sex and diabetes duration, 10 metabolites were found to have p-values < 0.05. GWAS of these 10 metabolites led to genome-wide significant findings in two metabolites, sphingomyelin C24:1-OH and phosphatidylcholine diacyl C36:6, rs2071499 (G>A) on chromosome 1 and rs2876585 (C>T) on chromosome 6, respectively. When adjusted for the significant SNPs, 11 metabolites reached p<0.05. No metabolites reached significance when p-values were adjusted for false discovery rate. These 11 metabolites were sphingomyelins and phosphatidylcholines.

Conclusion: Phosphatidylcholines are metabolites important in the biosynthesis of sphingomyelin, which is an important component of the myelin sheath in nerve cells. However, potentially more importantly, these two metabolites are involved in the sphingomyelin cycle; this is in metabolic flux with the ceramide/diacylglycerol pool, which has previously been implicated in diabetes-related complications.

Supported by: JDRF, DFN, FoDF

Disclosure: B.H.R. Wolffenbuttel: Grants; Juvenile Diabetes Research Foundation, DiabetesFonds NL, Friends of Diabeter Foundation.

64

ZT-01, a somatostatin receptor 2 antagonist (SSTR2a), increases pulsatile glucagon release during hypoglycaemia in type 1 diabetes

R. Liggins ¹, A. Abitbol², M.C. Riddell³, S. Peers⁴, E. Simonson¹, M. Midmer⁴;

¹Zucara Therapeutics, Vancouver, BC, Canada, ²LMC Healthcare/Centricity Research, Toronto, ON, Canada, ³York University, Toronto, ON, Canada, ⁴Zucara Therapeutics, Toronto, ON, Canada.

Background and aims: Glucagon is released in a pulsatile manner to regulate glycemia, preventing hypoglycemia. Defective glucagon responsiveness to hypoglycemia is observed in many people living with Type 1 diabetes (T1D) within 5 years of diagnosis. A novel SSTR2, ZT-01, has the potential to increase glucagon secretion in hypoglycemia. The purpose of this study was to examine if endogenous glucagon pulse frequency and pulse amplitude could be increased in hypoglycemia in T1D by the novel SSTR2a, ZT-01.

Materials and methods: ZT-01 was evaluated in a Phase 1b study, in which impaired glucagon counterregulation was increased during hypoglycemia with treatment. In this randomized, three-arm, double-blind crossover trial in 23 adults with T1D of at least 5 years’ duration, participants underwent hypoglycemic clamp procedures with 3 or 20 mg ZT-01, or placebo, dosed 1-2 h prior to stepwise hypoglycemia induction with variable rate IV insulin infusion without exogenous glucose: euglycemia (5.0 ± 0.5 mmol/L), Level 1 (3.5 ± 0.3 mmol/L), and Level 2 hypoglycemia (2.6 ± 0.2 mmol/L). Blood glucose and plasma glucagon were measured at 5- and ~15-minute intervals, respectively, along with 3 glucagon measurements within 15 minutes once target glycemia was achieved at each step. The frequency and amplitude of glucagon increases (pulses) were determined.

Results: Pulsatile increases in glucagon were observed in most participants during hypoglycemic clamp steps. Pulse frequency and amplitude were treatment-dependent. Glucagon pulse amplitude causing a rise in blood glucose had a median (IQR) value of 15.8 (12.1 - 29.3) pg/mL. Glucagon pulse amplitude and frequency of pulses of at least 10 pg/mL were both increased, and tended to occur earlier in the clamps with treatment. The blood glucose threshold for glucagon release (level at the start of each pulse) was not affected by treatment. However, a rise in glucose level (≥0.8 mmol/L over 20 minutes), coinciding with a glucagon pulse, was observed in 50% of 20 mg ZT-01 treatment clamps compared to 20% of clamps with placebo. Increases in blood glucose values coinciding with the peak glucagon value during the Level 2 hypoglycemia clamp step also increased for ZT-01 versus placebo (2.8 vs 3.3 mmol/L, p<0.05). Glucagon exposure during the Level 2 hypoglycemia clamp step (average plasma level over time) increased over baseline to 14.4 pg/mL with ZT-01 (20 mg dose) compared to 1.9 pg/mL with placebo (p<0.0001). These average values do not fully reflect the glucagon response, as peak (Cmax) values during Level 2 hypoglycemia clamp step increased on average to 27.1 pg/mL with ZT-01 compared to 8.4 pg/mL (placebo) (p<0.0001).

Conclusion: ZT-01 enhanced the magnitude, amplitude and onset of glucagon pulsatility during hypoglycemia in adults with T1D, a novel finding associated with observations of glucose counterregulation.

Clinical Trial Registration Number: NCT05007977

Disclosure: R. Liggins: Employment/Consultancy; Zucara Therapeutics Inc.; Stock/Shareholding; Zucara Therapeutics Inc.

65

Meta-genome-wide association studies identify potential SNPs and enriched pathways for impaired awareness of hypoglycaemia

M.M. van der Klauw ¹, R.D.M. Varkevisser¹, N. Ali², A. Alaswad³, IAH GWAS research group, Dutch Diabetes Pearl Genomics consortium, C.-J. Xu³, Y. Li³, C.J. Tack², B.H.R. Wolffenbuttel¹, B.E. De Galan^2,4;

¹Department of Endocrinology, University Medical Center Groningen, Groningen, Netherlands, ²Department of Internal Medicine, Radboud University Medical Center, Nijmegen, Netherlands, ³Department of Computational Biology of Individualized Infection Medicine, Helmholtz Centre for Infection Research / Centre for Individualized Infection Medicine, Hannover, Germany, ⁴Department of Internal Medicine, Maastricht University Medical Center, Maastricht, Netherlands.

Background and aims: There is evidence that genetic factors play a role in impaired awareness of hypoglycemia (IAH), but findings are limited and partly controversial. The aim of the present study was therefore to identify genetic determinants of IAH in individuals with diabetes mellitus using genome-wide association.

Materials and methods: Genome-wide association meta-analyses were conducted for the presence of IAH. Three independent cohorts were involved including the Radboud University Medical Center type 1 diabetes cohort, the Biomarkers of heterogeneity in type 1 diabetes cohort, and the type 2 diabetes Dutch Diabetes Pearl cohort. IAH was assessed using the clamp-validated modified Dutch Clarke questionnaire score, and the blood glucose levels at which patients start to feel hypoglycaemic symptoms. DNA samples were genotyped with the Illumina GSA v1 or v3 chip, imputed to the Haplotype Reference Consortium, and meta-analysed using METAL.

Results: A total of 1,087 patients with type 1 and 1,734 patients with type 2 diabetes were included. In individuals with type 1 diabetes a SNP on chromosome 5 coding for CTC-546K23.1 reached genome-wide significance for the question “How low is your blood glucose before you feel hypoglycaemia symptoms?” (p=3.32e-8). Although a function has yet to be described, GTEx (v8) expression data shows higher expression of CTC-546K23.1 in brain-related tissue compared to other tissues. In the meta-analyses for the presence of IAH and the Clarke score no genome-wide significant SNPs were found. However, SNPs of potential genome-wide interest (P<5e-6) were identified, and lead SNPs mapped to protein-coding genes. Enrichment analysis of genes mapped with SNPs of genome-wide interest revealed some potentially enriched KEGG and Reactome pathways, as well as gene ontology (GO) terms. Amongst the enriched pathways were those involving neural regulation and development, and genes involved in the renin-angiotensin system. Similarly, enriched GO terms included processes of neural development and regulation, and regulation of systemic arterial blood pressure.

Conclusion: Genome-wide association meta-analyses of IAH show that there are genetic determinants of IAH, whose function remains largely unknown.

Supported by: JDRF, DFN, FoDF, AZ, NN, P, SA, NWO

Disclosure: M.M. van der Klauw: Grants; Juvenile Diabetes Research Foundation, Dutch Diabetes Research Foundation, AstraZeneca, Novo Nordisk, Pfizer, Sanofi Aventis, Netherlands Organisation for Scientific Research.

66

Reduced defensive responses to hypoglycaemia are associated with a loss of neuronal activation in the dorsomedial hypothalamus in mice

L. Sheppard ¹, E. Staricoff¹, E. Naden¹, M. Josipovic¹, C. Riches¹, B. de Galan², B. Thorens³, U. Pedersen-Bjergaard^4,5, A. McNeilly⁶, R. McCrimmon⁶, M. Evans¹;

¹University of Cambridge, Cambridge, UK, ²Department of Internal Medicine, Radboud university medical centre, Nijmegen, Netherlands, ³University of Lausanne, Lausanne, Switzerland, ⁴Department of Endocrinology and Nephrology, Nordsjællands Hospital, Hillerød, Denmark, ⁵Institute of Clinical Medicine, University of Copenhagen, Copenhagen, Denmark, ⁶University of Dundee, Dundee, UK.

Background and aims: Recurrent insulin-induced hypoglycaemia can result in impaired hormonal responses to subsequent events of low blood glucose, often termed HAAF (Hypoglycaemia Associated Autonomic Failure). To determine possible underlying central mediators, we validated and examined a mouse model of HAAF for neuronal activation patterns.

Materials and methods: Male C57BL/6J mice 6-8 weeks of age underwent surgery for the implantation of carotid artery and jugular vein catheters. Following recovery, mice underwent 4 sequential daily SC injections of either Aspart insulin (1.5 U/kg) to create recurrent hypoglycaemia (RH, n=6) or saline control (SAL, n=6). On day 5, free moving conscious mice underwent a 2h hyperinsulinaemic (Aspart, 10 mU/kg/min) hypoglycaemic clamp (plasma glucose 3.1 + 0.1 mM) with donor erythrocyte infusion to allow sample collection for later hormonal measures. Following perfusion-fixation at the end of the clamp studies, brains were collected, cleared and immunolabelled, after which they were rehydrated and embedded in paraffin. Hypothalamic sections were examined for cFos, a marker serving as proxy for neuronal activation.

Results: Consistent with the development of HAAF, adrenaline (0.7 + 0.2 vs 1.2 + 0.1 ng/ml, p=0.01) and glucagon (137 + 65 vs 393 + 45 pg/ml, p=0.01) responses to hypoglycaemia were lower in RH compared with SAL mice. We observed hypoglycaemia-induced discrete cFos activation in the dorsal area of dorsomedial hypothalamus (dDMH) in SAL mice but with marked blunting of responses in RH. Furthermore, the levels of cFos activation positively correlated with adrenaline responses to hypoglycaemia (Spearman p=0.02).

Conclusion: Our data identify the dDMH as a novel hypoglycaemia-responsive brain area with neuronal activation correlated to adrenaline responses to low glucose. We speculate that HAAF may be explained by RH-induced adaptations in this area, potentially offering a future therapeutic target.

Disclosure: L. Sheppard: Grants; This work was supported by the Innovative Medicines Initiative 2 Joint Undertaking (JU) under grant agreement No 777460. The JU receives support from the European Union’s Horizon 2020 research program.

OP 12 Diabetes and the brain

67

Associations of lower time in range with cognitive dysfunction and hippocampal damage in adults with type 2 diabetes

Z. Zhang, C. Yu, Y. Bi;

Endocrinology, Endocrine and Metabolic Disease Medical Center, Affiliated Drum Tower Hospital, Medical School, Nanjing University, Nanjing, China.

Background and aims: Accumulating evidence has suggested time in range (TIR) is associated with the risk of diabetic microvascular and macrovascular complications. This study aimed to investigate the cross-sectional associations between TIR and cognitive dysfunction, as well as the underlying hippocampal manifestations in patients with type 2 diabetes.

Materials and methods: A total of 510 participants with type 2 diabetes who completed continuous glucose monitoring and comprehensive cognitive evaluations were enrolled. Among them, 100 patients with normal cognition underwent 3.0T magnetic resonance imaging to assess brain structural and functional changes.

Results: The prevalence of mild cognitive impairment (MCI) was found to be significantly associated with lower TIR quartiles (P for trend <0.001). The multivariable-adjusted odds ratios (ORs) of MCI associated with TIR quartiles (Q1: ≥88% [reference group], Q2: 71-87%, Q3: 51-70%, Q4: <51%) were 1.00, 1.60 (95% CI: 0.66-3.83), 3.02 (95% CI: 1.14-8.04), and 7.67 (95% CI: 2.56-23.02) respectively (P for trend <0.001). Furthermore, the multivariable-adjusted OR for every 10% decrease in TIR was 1.36 (95% CI 1.20-1.52) for MCI. Decreased TIR was associated with cognitive decline in immediate memory, language, attention, delayed memory and processing speed. Importantly, cognitively normal patients with lower TIR showed reduced volume of left hippocampal subregions and decreased activation of the right parahippocampal gyrus, and these hippocampal alterations mediated the association between lower TIR and cognitive dysfunction

Conclusion: Lower TIR is associated with the elevated prevalence of cognitive impairment in individuals with type 2 diabetes, and such association is mediated by hippocampal damage in type 2 diabetes. These findings underscore the clinical relevance of better TIR attainment could be beneficial for preserving cognitive function and brain integrity in type 2 diabetes.

Clinical Trial Registration Number: NCT02738671

Supported by: National Natural Science Foundation of China Grant Awards (82000775)

Disclosure: Z. Zhang: None.

68

Decreased cerebral glucose and increased cerebral fatty acid utilisation during fasting in type 2 diabetes

M. Schain ¹, E. Johansson¹, I. Laitinen¹, A. Frödén Löwenmark¹, M. Lubberink², A. Gummesson³, P. Nuutila⁴, R. Esterline⁵, J. Oscarsson⁶, L. Johansson¹, K. Heurling¹;

¹Antaros Medical AB, Mölndal, Sweden, ²Department of Surgical Sciences, Radiology, Uppsala University, Uppsala, Sweden, ³Department of Molecular and Clinical Medicine, University of Gothenburg, Gothenburg, Sweden, ⁴Turku PET Centre, University of Turku and Turku University Hospital, Turku, Finland, ⁵AstraZeneca, Gaithersburg, MD, USA, ⁶AstraZeneca, Gothenburg, Sweden.

Background and aims: Epidemiological studies show that type 2 diabetes (T2D) increases the risk of developing Alzheimer’s Disease (AD). Alzheimer’s disease is associated with brain hypometabolism. It has been hypothesized that T2D and peripheral insulin resistance also manifest brain hypometabolism, and that central metabolic alterations constitutes a link between T2D and AD. The aim of the current study was to assess whether metabolic alterations are present in the brain of obese individuals with or without T2D using PET imaging with [¹⁸F]FDG and [¹⁸F]FTHA (radiolabelled glucose, and fatty acid analogues).

Materials and methods: Thirty-eight subjects were included (10 with T2D, 13 with obesity but without clinical diagnosis of T2D, and 15 lean controls). Subjects diagnosed with AD or cognitive dysfunction (MMSE<27) were excluded. Participants underwent [¹⁸F]FDG and [¹⁸F]FTHA brain PET imaging following 6 hours of fasting, to measure glucose and fatty acid utilization, using a 3T PET/MR system. Cerebral perfusion was assessed via arterial spin-labelling MRI, and single-voxel 1H-MRS was employed to measure metabolites (N-acetyl aspartate, choline, myo-inositol) in selected regions. Several markers of neurodegeneration were derived from cerebrospinal fluid (CSF) samples, including P-Tau, T-Tau, Abeta 1-40 and Abeta 1-42. Subjects were genotyped for the APoE4 allele.

Results: Glucose uptake in grey matter (GM) (Fig 1A) was lower among T2D than in controls (p=.01) and obese participants (p=.004), and correlated negatively with insulin resistance measured with HOMA-IR (p<.05). Fatty acid uptake rate in white matter (WM) was higher in the T2D group than in the controls (p=.03), but no statistical difference was found in GM (Fig 1A). Cerebral perfusion was lower in both GM and WM in T2D (p=.003 and .01), and obese (p=.006 and .02) participants compared to controls (Fig 1B). No differences were observed on any of the metabolites studied MRS or on the CSF markers.

Conclusion: Cerebral glucose hypometabolism during fasting is well-documented in AD. We here show that brain glucose utilization was also lower in participants with T2D without cognitive impairment and correlated with peripheral insulin resistance. Reduced blood flow may be a consequence of the lower oxygen demand following decreased glucose consumption. Increased uptake of fatty acids could be a compensatory mechanism for the impaired glucose utilisation. Metabolic imaging could be useful to study whether antidiabetic drugs attenuate brain metabolic alterations and thereby prevent or revert cognitive dysfunction.

Clinical Trial Registration Number: 2019-03454 (EC Sweden)

Supported by: This study was partially funded by AZ

Disclosure: M. Schain: None.

69

The role of microglial SCAP in diabetes-associated cognitive dysfunction

W. Zhu ¹, H. Zhang², T. Niu¹, K. Liu¹, S. Wang¹;

¹Department of Endocrinology, Southeast University, Nanjing, China, ²Department of Endocrinology, The First Affiliated Hospital of USTC, Division of Life Sciences and Medicine, University of Science and Technology of China, Hefei, China.

Background and aims: Diabetic patients with disrupted cholesterol metabolism are at significantly increased risk of cognitive impairment, and Sterol Regulatory Element-Binding Protein (SREBP) cleavage-activating protein (SCAP) is a cholesterol sensor that may serve as a signal hub integrating cholesterol metabolism and inflammatory immune regulation with broad functional effects in metabolic disorders. Here, we examined the effect of microglia-specific SCAP in diabetes-associated cognitive dysfunction (DACD)

Materials and methods: We generated microglia-specific knockout of SCAP (mSCAP) using the Cre-loxP recombinase system, as well as SCAP^flox/flox controls from the same litter. Both groups were fed either a normal diet or a 60% high-fat diet for 24 weeks to induce DACD. We evaluated learning and memory function by Morris water maze, Open Field Test, Y maze and Novel object recognition test; We used confocal microscopy and image J to scan and 3D reconstruct microglial cells to assess microglial cell morphology and polarization phenotype in the hippocampal tissue of each group of mice; We isolated primary microglial cells from adult mice using Miltenyi Biote mouse CD11b separation beads and measured expression of inflammatory responses; We examined neuronal and synaptic damage, cholesterol synthesis and inflammatory responses by detecting the expressions of related proteins and RNA. In addition, we transfected BV2 cells with a slow virus containing SCAP fragments for in vitro validation of in vivo experimental results. Finally, we used immunofluorescence and chromatin immunoprecipitation to confirm the direct interaction mechanism between the SCAP/SREBP2 and NF-kB inflammatory signaling pathways.

Results: In DACD mice, hippocampal and microglial SCAP levels were abnormally elevated, while microglial-specific knockout of SCAP showed a protective effect on cognitive function and neuronal damage induced by diabetes. Interestingly, mSCAP mice were shown to effectively increase microglial branching and reduce metabolic inflammation in DACD mice, as well as reduce M1 polarization of microglial cells. Further molecular analysis revealed that SCAP specifically recruited the transcription factor SREBP2 to the Golgi apparatus, further mature SREBP2 entered the nucleus and directly activated the NF-kB inflammatory signaling pathway, thereby promoting the release of inflammatory factors.

Conclusion: SCAP/SREBP2 is involved in the occurrence and development of DACD by directly activating the NF-kB signaling pathway and mediating microglial inflammation. Inhibiting microglial SCAP may become a new strategy for treating DACD.

Supported by: NSFC No. 81870568

Disclosure: W. Zhu: None.

70

Intracerebrospinal fluid AAV-FGF21 gene therapy prevents high fat diet-associated cognitive decline in mice

I. Grass ^1,2, I. Elias^1,2, V. Sacristan^1,2, A. Ribera^1,2, C. Jambrina^1,2, E. Casaña^1,2, V. Jimenez^1,2, F. Bosch^1,2;

¹Center of Animal Biotechnology and Gene Therapy (CBATEG), Universitat Autonoma de Barcelona, Bellaterra, Spain, ²Centro de Investigación Biomédica en Red de Diabetes y Enfermedades Metabólicas Asociadas, (ciberdem), Spain.

Background and aims: Type 2 diabetic and obese patients have a higher risk of dementia and loss of cognitive ability. Currently, there is no effective treatment for these associated cognitive deficits. Fibroblast growth factor 21 (FGF21) has been described to reduce cognitive decline in pre-clinical animal models. Here, we assessed the therapeutic potential of AAV-mediated gene transfer of FGF21 to the CNS, by intra-cerebrospinal fluid (CSF) administration, to treat obesity, diabetes, and the associated cognitive deficits.

Materials and methods: Two-month-old C57Bl6 mice were fed either a chow or a HFD for 14 months. Animals were administered intra-CSF with adeno-associated viral (AAV) vectors carrying a murine FGF21 coding sequence (AAV-FGF21). Nine months after the treatment different behavioural tests were performed.

Results: In addition to reversion of obesity and insulin resistance, intra-CSF administration of AAV-FGF21 in HFD-fed mice also counteracted the obesity-associated cognitive decline. Treatment with AAV-FGF21 prevented short-term and long-term memory loss, improved learning capacity and decreased anxiety-like behaviour. These therapeutic benefits were mediated, at least in part, by reduction of astrogliosis, increase of BDNF levels, and enhancement of PI3K-PDK1-AKT signalling pathway in the hippocampus.

Conclusion: All these results underscore the potential of brain-derived AAV-FGF21 gene therapy to treat obesity, insulin resistance, T2D, and associated cognitive decline.

Supported by: MINECO and FEDER, EU (SAF2017-86266R and PRE2018-084309) PID2020-113864RB-I00;2017 SGR1508; ICREA

Disclosure: I. Grass: None.

71

Remnant cholesterol is an independent risk factor for dementia in patients with type 2 diabetes: a nationwide population-based cohort study

H. Jung ¹, J. Huh¹, E. Rho¹, S. Lee¹, S.-H. Ihm¹, K.-D. Han², J. Kang¹;

¹Hallym University Sacred Heart Hospital, Anyang-si, Gyeonggi-do, Republic of Korea, ²Soongsil University, Seoul, Republic of Korea.

Background and aims: No study to date has longitudinally evaluated the effect of remnant cholesterol (remnant-C) on incident dementia in patients with type 2 diabetes (T2D). This retrospective study investigated the association between remnant-C levels and the development of dementia in Korean patients with T2D.

Materials and methods: Participants were 2,621,596 patients with T2D aged ≥40 years who underwent the national health screening in 2009-2011 and they were followed until 2020. Cox regression analyses were used to calculate hazard ratios (HRs) for all-cause, Alzheimer's, and vascular dementia in each quartile of remnant-C, adjusting for conventional risk factors for dementia.

Results: During a median follow-up of 10.3 years, 146,991 (5.6%), 117,739 (4.5%), and 14,536 (0.6%) developed all-cause, Alzheimer's, and vascular dementia, respectively. The risk of dementia increased progressively with higher remnant-C levels. Compared to the lowest quartile of remnant-C, HRs in the highest quartile were 1.11 (95% confidence interval [CI], 1.09-1.13), 1.11 (1.08-1.13), and 1.15 (1.09-1.21) for all-cause, Alzheimer's, and vascular dementia, respectively. The increased risk of dementia with higher remnant-C was more pronounced in younger than in older patients (HRs for all-cause dementia, 1.52 [95% CI, 1.36-1.69] in age 40-49, 1.17 [1.11-1.23] in age 50-59, 1.11 [1.08-1.14] in age 60-69, and 1.08 [1.05-1.10] in ≥70 years, respectively).

Conclusion: Higher remnant-C levels were associated with an increased risk of dementia in patients with T2D. This association was independent of traditional risk factors for dementia and was more pronounced in the younger patients. This implicates remnant-C as a novel risk factor for dementia.

Disclosure: H. Jung: None.

72

An assessment of factors contributing to the decline in risk of stroke complicating type 2 diabetes over recent decades: the Fremantle Diabetes Study

T.M.E. Davis, W.A. Davis;

University of Western Australia, Fremantle, Australia.

Background and aims: We have reported previously that the 5-year incidence of fatal/non-fatal stroke almost halved in the 15 years between the Fremantle Diabetes Study Phase 1 (FDS1: n=1291 recruited 1993-96; mean age 64 years, 49% males) and Phase 2 (FDS2: n=1509 recruited 2008-11; mean age 65 years 52% males). The aim of this study was to examine whether this reflected temporal changes in cerebrovascular disease risk factors and their management.

Materials and methods: The FDS is a community-based longitudinal observational study of diabetes involving representative cohorts from a postcode-defined typical urban Australian population. FDS1 and FDS2 participants with type 2 diabetes were followed for first ever hospitalisation for stroke or death over 10 years from study entry. Cox regression generated hazard ratios (HRs) for incident stroke in FDS1, FDS2 and the combined cohort. The proportional hazards assumption was assessed and, when violated, adjusted for by adding significant time-varying covariates.

Results: The 10-year incidence rates (95% CI) for stroke in FDS1 versus FDS2 were 12.4 (10.4, 14.7) versus 6.6 (5.3, 8.2) /1,000 person-years (incident rate ratio for FDS2 versus FDS1 0.54 (0.40, 0.71), P<0.001). The independent baseline associates for first stroke are shown in the table (*denotes time-varying variable). For FDS1 and FDS2 considered separately, common risk factors were increasing age and ln(urinary albumin:creatinine (uACR)). Other independent associates were atrial fibrillation/flutter (AF), warfarin use and Charlson Comorbidity Index (CCI) ≥3 in FDS1 and heart rate and prior cerebrovascular disease (CVD) in FDS2. In the pooled FDS1+FDS2 cohort, older age, ln(uACR), AF and prior CVD were independent associates, as were use of lipid modifying medication (protective) and peripheral arterial disease. When FDS2 was added to the pooled model, it was statistically significant and displaced lipid-modifying medication use.

Conclusion: These long-term data confirm the substantial reduction in stroke complicating type 2 diabetes between FDS phases, but indicate that microalbuminuria and AF remain potentially modifiable risk factors. The independent association between FDS2 participation and incident first stroke in the combined analysis suggests that there is a true FDS phase-specific difference after adjusting for confounding variables. This is partly explained by the greater use of lipid-modifying medications in FDS2 versus FDS1.

Supported by: Raine Foundation University of Western Australia, NHMRC project grants 513781 and 1042231

Disclosure: T.M.E. Davis: None.

OP 13 Concerning insulin resistance

73

Insulin resistance is associated with unbalanced GIP and GLP-1 secretion

G. Di Giuseppe ^1,2, G. Ciccarelli¹, L. Soldovieri¹, M. Brunetti¹, F. Cinti¹, S. Moffa¹, U. Capece¹, A. Mari³, J.J. Holst⁴, A. Giaccari¹, T. Mezza⁵;

¹Endocrinology and Diabetology Unit, University Hospital Agostino Gemelli, Rome, Italy, ²Department of Biomedical Sciences, University of Copenhagen, Copenhagen, Denmark, ³National Research Council of Italy, Padua, Italy, ⁴Novo Nordisk Foundation Centre for Basic Metabolic Research, University of Copenhagen, Copenhagen, Denmark, ⁵Pancreas Unit - Digestive Disease Center, University Hospital Agostino Gemelli, Rome, Italy.

Background and aims: The incretin hormones GIP and GLP-1 exert a fundamental role in glucose homeostasis after a meal, since their insulinotropic effect on beta-cells. However, it is still debated if modifications in GIP and GLP-1 secretion can affect the metabolic state, as well as their relative contribution to the dynamics of incretin-induced insulin secretion. In this study we aimed to explore the relative contribution of GIP and GLP-1 secretion and GIP/GLP-1 ratio on glucose metabolism and beta-cell function in a cohort of non-diabetic and newly diagnosed type 2 diabetic (T2D) subjects.

Materials and methods: 41 subjects with no previous diagnosis of T2DM and not on antidiabetic treatments were divided into 3 groups according to glucose tolerance calculated during an oral glucose tolerance test (OGTT): non glucose tolerant (NGT, n=12), impaired glucose tolerant (IGT, n=12), and type 2 diabetic (T2D, n=17). All subjects underwent a 4h mixed meal test (MMT) and glucose, insulin, c-peptide, GIP and GLP-1 levels were assessed every 30 min. To compare incretin secretion among the 3 groups, we performed a third-level interaction analysis by including a product term of time × glucose tolerance in the model. Further, we calculated the GIP/GLP-1 Secretion Ratio (GIP/GLP-1 SR) for each timepoint normalized for basal levels and the AUC of the ratio of GIP and GLP-1 secretion. Further, all subjects underwent a euglycemic hyperinsulinemic clamp to measure whole-body peripheral glucose utilization (M-value, mg·kg^-1·min^-1), an index of insulin sensitivity.

Results: As expected, during MMT we observed higher glucose levels in T2D subjects compared to NGT and IGT (p<0.01 for interaction), while insulin and c-peptide secretion showed a non-significant decreasing trend in T2D subjects compared to the other 2 groups (p=0.10 and p=0.21 respectively for interaction). Further, we observed a trend of decreasing GIP secretion in IGT and T2D subjects compared to NGT, but not statistically significant (p=0.08 for interaction), whereas no differences in GLP-1 secretion were detected among the 3 groups (p=0.65 for interaction). Moreover, IGT and T2D groups showed a reduced GIP/GLP-1 SR compared to NGT (p=0.02 for interaction). Interestingly, we found that the AUC of GIP/GLP-1 SR was positively correlated with the M-value (R=0.45, p=0.01).

Conclusion: Our data show that, rather than an altered secretion of the incretin hormones per se, insulin resistance is associated with an imbalance in the dynamics of incretin secretion overall after a meal. In particular, the GIP/GLP-1 SR appears to be an early marker of impaired glucose homeostasis in the time course towards type 2 diabetes. Further studies are needed to investigate whether the altered GIP/GLP-1 secretion is the cause or the effect of altered beta-cell incretin sensitivity and the potential metabolic and extrapancreatic effects of unbalanced incretin secretion.

Supported by: EFSD/AZ Mentorship Programme

Disclosure: G. Di Giuseppe: None.

74

The relative contribution of insulin secretion and insulin sensitivity in the progression to type 2 diabetes in Asians

M.H. Lee ¹, E. Febriana¹, M. Lim^1,2, S. Jin³, A.R. Cook³, S. Baig¹, I.Y.H. Ang³, A. Nastar⁴, J.B. Halter^1,5, F. Magkos⁶, S.-A. Toh^2,7;

¹Medicine, National University of Singapore, Singapore, Singapore, ²Medicine, National University Hospital, Singapore, Singapore, ³National University of Singapore, Singapore, Singapore, ⁴Alexandra Hospital, Singapore, Singapore, ⁵University of Michigan, Ann Arbor, MI, USA, ⁶Nutrition, Exercise and Sports, University of Copenhagen, Copenhagen, Denmark, ⁷NOVI Health, Singapore, Singapore.

Background and aims: Type 2 diabetes (T2D) is highly prevalent in Asia. A clearer understanding of its pathophysiology will have far-reaching implications for reducing disease burden. However, currently there are conflicting reports on the roles of insulin secretion and action in T2D progression. In this study, we assessed the relative contribution of inadequate insulin secretion and insulin resistance in the progression to T2D over 3 years in an Asian population using dynamic gold-standard measures.

Materials and methods: We recruited 1679 non-diabetic Asians from a multi-ethnic population in Singapore. Participants underwent an oral glucose tolerance test, an intravenous glucose challenge and a hyperinsulinemic-euglycemic clamp procedure to determine glucose tolerance, insulin secretion (acute insulin response) and insulin sensitivity (M-value over insulin concentration) at baseline and follow-up. 1248 out of the 1679 participants completed the study and were divided into four groups, based on the absence (n=477) or the presence of impaired insulin secretion (IIS, n=200), insulin resistance (IR, n=477), or combined IIS-IR (n=94). Cox’s proportional hazards model was used to estimate hazard ratios (HR) for T2D. The population attributable fraction (PAF) due to IIS, IR, or combined IIS-IR were also calculated. Repeated measures ANOVA was used to assess the longitudinal changes of insulin secretion and sensitivity between the baseline and follow-up visits.

Results: 77 out of the 1248 completers progressed to T2D. The mean follow-up period was 3.0 years. At baseline, isolated IIS was associated with higher percentage of females and a lower BMI (~23 kg/m²), whereas IR, whether isolated or combined with IIS, was associated with a higher BMI (>26 kg/m²). Participants with combined IIS-IR had the highest blood glucose concentrations. The HRs (95% CI) for T2D adjusted for age, sex, BMI and ethnic group, were 19.4 (4.4, 85.1) for isolated IIS, 9.9 (2.3, 42.3) for isolated IR, and 67.9 (16.2, 285.3) for combined IIS-IR. The corresponding PAFs (95% CI) were 18.5% (15.1, 19.3), 30.4% (19.2, 33.0), and 43.5% (41.4, 44.0), respectively. The ability to upregulate insulin secretion in response to a decrease in insulin sensitivity in the normal and the isolated IIS groups, or to upregulate insulin sensitivity in the isolated IR and combined IIS-IR groups, was associated with non-conversion to T2D.

Conclusion: While isolated IIS increases risk for T2D conversion by about 2-fold compared to isolated IR, both defects are critically important in the progression of the disease, with their combination conferring the greatest risk. Depending on the underlying pathophysiological defect, different intervention strategies may be needed to prevent T2D progression in Asians.

Clinical Trial Registration Number: NCT02838693

Supported by: NMRC and Ministry of Health IAF (Singapore), and Janssen Pharmaceuticals Inc. (USA)

Disclosure: M.H. Lee: None.

75

Dep-1 is a novel brain insulin receptor phosphatase and transcriptome regulator

S. Chopra ¹, J. Ulke², L. Schmidt³, O.L.J. Kadiri¹, M. Krüger³, K. Kappert², A. Kleinridders¹;

¹Molecular and Experimental Nutritional Medicine, University of Potsdam, Nuthetal, Germany, ²Institute of Laboratory Medicine, Clinical Chemistry and Pathobiochemistry, Charité - Universiätsmedizin Berlin, Berlin, Germany, ³CECAD Research Center/Institute for Genetics, University of Cologne, Cologne, Germany.

Background and aims: Obesity and diabetes are conditions characterized by insulin resistance, which not only affects peripheral tissues but is known to occur in the brain, where the dysregulation of insulin signaling is a feature of metabolic syndrome. Protein tyrosine phosphatases (PTPs) are crucial negative modulators of insulin action, and a subset of PTPs including the ubiquitously expressed transmembrane receptor-like PTP, density-enhanced phosphatase-1 (DEP-1), dephosphorylates phospho-tyrosine residues of the insulin receptor in peripheral tissues. Thus, to understand the impact of DEP-1 overall in neurons, we investigated DEP-1 dependent metabolic effects in vitro.

Materials and methods: To study the functional consequences of DEP-1 deficiency on insulin action, we established a DEP-1 KO in mouse undifferentiated dopaminergic-like Neuro-2a (N2A) cells with CRISPR/Cas9 technology and determined insulin signaling using immunoblotting. Furthermore, to gain insights into the function of DEP-1 in neurons, we used a phospho-proteomic approach to determine protein phosphorylation in control and DEP-1 deficient cells under basal and insulin resistant conditions. RNA sequencing served for in-depth transcriptome analysis of control and DEP-1 KO cells.

Results: DEP-1 KO cells revealed not only a significant increase in insulin receptor (IR) phosphorylation (n=4, Mann-Whitney U test=p<0.05) but also increased activation of insulin signaling nodes including increased ERK phosphorylation upon acute insulin challenge of 100nM (n=8, unpaired student t-test=p<0.05) shown by immunoblotting analysis. This demonstrates that DEP-1 is a phosphatase of neuronal IR and that DEP-1 deficiency in vitro leads to enhanced insulin signaling. Moreover, the phospho-proteomic analysis revealed that while unstimulated DEP-1 KO cells exhibited only subtle differences in phosphorylation patterns compared to control cells, palmitate-induced insulin-resistant conditions (400 μM palmitate for 16 hours) caused marked differences in phosphorylation patterns of over 333 phospho-sites from 5131 recognized phospho-sites between control and DEP-1 KO cells (n=6, log Welch’s t-test analysis= p>1.3, Welch’s t-test difference= p<-1.5 or >1.5). Interestingly, the STRING database revealed that a few of these phospho-sites were seen to be involved in the regulation of the insulin receptor signaling pathway as well (Irs2, Eif4ebp1, Smarcc1, Phip, and Ptpn1), supporting that DEP-1 impacts insulin signaling. Strikingly, a substantial proportion of these phospho-sites were also seen to alter biological functions related to RNA splicing (Polr2a, Srsf1, Srrm1), and DNA methylation (Dnmt3a, Brca1, Mbd3). Moreover, RNA sequencing revealed 34 significant alternative splicing events which occurred in DEP-1 KO cells compared to control cells (n=4, unpaired student t-test=p<0.05), underlying the observations seen in the phospho-proteomics.

Conclusion: Neuronal DEP-1 represents an IR phosphatase, with DEP-1 deficiency not only modulating insulin signaling but also exhibiting profound effects on cellular metabolism likely through epigenetic modifications and regulation of splicing.

Supported by: DFG

Disclosure: S. Chopra: None.

76

Renal cortical glucose uptake rate is decreased in insulin resistance and correlates inversely with serum free-fatty acids, ketone bodies and citrate levels

E. Rebelos ^1,2, A. Mari³, M.-J. Honka¹, L. Pekkarinen¹, S. Laurila¹, A. Latva-Rasku¹, J. Rajander⁴, E. Saukko⁵, H. Iida¹, P. Salminen⁶, E. Ferrannini⁷, P. Nuutila¹;

¹Turku PET Centre, University of Turku, Turku, Finland, ²Department of Clinical and Experimental Medicine, University of Pisa, Pisa, Italy, ³Institute of Neuroscience, CNR, Padova, Italy, ⁴Abo Akademi, Turku, Finland, ⁵Department of Radiology, Turku University Hospital, Turku, Finland, ⁶Division of Digestive Surgery and Urology, Turku University Hospital, Turku, Finland, ⁷Institute of Clinical Physiology, CNR, Pisa, Italy.

Background and aims: We have recently demonstrated that positron emission tomography using ¹⁸F-labeled-2-fluoro-2-deoxy-D-glucose ([¹⁸F]FDG-PET) can be used to analyse renal glucose uptake rates (GU), and that the renal cortex is an insulin sensitive tissue. In the present study, we aimed to gain insight into the determinants of renal glucose metabolism by testing whether circulating metabolites and renal sinus fat mass are related to renal GU.

Materials and methods: 18 subjects with obesity and 18 sex- and age-matched non-obese controls were studied with [¹⁸F]FDG-PET during insulin clamp conditions (40mU/m²/min). Renal scans were obtained immediately after [¹⁸F]FDG injection and again ~60 min later. Renal GU was measured using fractional uptake rate and after correcting for residual intratubular [¹⁸F]FDG. Circulating metabolites were measured using high-throughput proton NMR metabolomics. In 18 subjects renal sinus fat (RSF) was analysed using magnetic resonance imaging.

Results: We found that the corrected cortical GU is higher in healthy non-obese controls compared to subjects with obesity, and that it associates positively with the degree of insulin sensitivity (M value) (r=0.42, p=0.01). Moreover, corrected cortical GU was inversely associated with circulating β-OH-butyrate (r=-0.58, p=0.009), acetoacetate (r=-0.48, p=0.008), citrate (r=-0.44, p=0.01) and free fatty acids (FFA) (r=-0.68, p<0.0001), even when accounting for the M value. Corrected cortical CGU was also inversely associated with RSF (r=-0.58, p=0.01), irrespective of the degree of insulin sensitivity. On the contrary, medullary GU was associated only with urinary measures [¹⁸F]FDG.

Conclusion: This data suggest that when circulating citrate, FFA and ketones are relatively increased - or less suppressed by insulin - renal cortical tissue uses less glucose via glycolysis and produces more de novo glucose via gluconeogenesis.

Clinical Trial Registration Number: NCT04343469

Supported by: EFSD/MSD European Research Programme, grant number:96402

Disclosure: E. Rebelos: None.

77

The mechanisms of duodenal mucosal resurfacing procedure in reversing type 2 diabetes

N. LiJuan ¹, C. Zhe², S. YaoHuan³, Y. XinYi², Z. XiQiao⁴;

¹School of Medicine and Holistic Integrative Medicine, Nanjing University of Chinese Medicine, Nanjing, China, ²The first clinical medical college of Nanjing University of Chinese Medicine, Nanjing, China, ³Department of Medical Care, School of Nursing, Nanjing University of Chinese Medicine, Nanjing, China, ⁴b Department of Endocrinology, Affiliated Hospital of Nanjing University of Chinese Medicine, Jiangsu Province Hospital of Chinese Medicine, Nanjing, China.

Background and aims: Clinical studies have found that endoscopic duodenal mucosal resurfacing (DMR) procedure can ablate and destroy the abnormally thickened duodenal mucosa in patients with type 2 diabetes (T2D), which can significantly improve the metabolic disorder and achieve the reversal of T2D after the mucosa is rejuvenated. However, the role of duodenal mucosa in the pathogenesis of T2D and the mechanism of metabolic improvement by DMR procedure remain unclear. We are aim to study the mechanisms of DMR procedure to improve metabolism at the animal level, thereby providing a theoretical basis for the promotion of clinical application of DMR procedure.

Materials and methods: Six week old SD rats were fed with high-fat diet and injected intraperitoneally with streptozotocin (STZ, 30mg/kg) to establish a T2D model, and were randomly divided into DMR group and sham group. DMR procedure involves inserting a self-made DMR catheter into the duodenum through a gastric incision to perform mucosal abrasion, with a length of 6 cm. Oral glucose tolerance tests (OGTT) were performed at 3 and 6 weeks after surgery. The duodenum was evaluated by histology, immunohistochemistry, immunofluorescence and western blotting. the levels of plasma glucose, insulin, Glucose-dependent insulinotropic polypeptide (GIP), glucagon like peptide-1 (GLP-1), TG, and TC were measured.

Results: Compared with sham group, DMR procedure reduced fasting blood glucose (DMR group: 9.12±0.72, Sham group: 20.43±2.45, p<0.001), AUC_OGTT (DMR group: 39.63±5.32 mmol/L, Sham group: 59.56±5.85 mmol/L, p<0.001), HbA1c (DMR group: 158.12±15.13 ng/ml, Sham group: 219.83±58.02 ng/ml, p=0.05), Lipopolysaccharide (LPS, DMR group: 21.52±4.64 pg/ml, Sham group: 31.93±9.88 pg/ml, p=0.04), and increased serum insulin (DMR group: 17.01±3.32 mIU/ml, Sham group: 13.72±3.51 mIU/ml, p=0.125), GIP (DMR group: 72.73±24.43 pg/ml, Sham group: 38.27±17.79 pg/ml, p=0.002), and GLP-1(DMR group: 2.30±0.55 ng/ml, Sham group: 1.67±0.37 ng/ml, p=0.043) levels in rats with T2D at 6 weeks after surgery. Both duodenal villi length and crypt depth were significantly increased in T2D rats, and DMR procedure reversed these changes. In addition, DMR procedure enhanced the expression of tight junction proteins Occludin and ZO-1 in the duodenal mucosa of T2D rats.

Conclusion: Our results suggest that DMR can significantly improve metabolic disorders in T2D rats in a variety of ways, possibly related to reducing villi length and crypt depth to reduce nutrient absorption, enhancing duodenal barrier function to reduce LPS leakage, and reducing insulin resistance by increasing serum GIP and GLP-1 levels. Further animal studies are needed to validate these findings.

Disclosure: N. LiJuan: None.

78

¹³ C glucose breath test analysis as an accurate index of insulin resistance in people with type 1 diabetes

J. Mertens ^1,2, R. Braspenning¹, L. Roosens³, S. Francque^4,2, C. De Block^1,2;

¹Endocrinology, Diabetology and Metabolism, University Hospital Antwerp, Edegem, Belgium, ²Laboratory of Experimental Medicine and Paediatrics, University of Antwerp, Wilrijk, Belgium, ³Clinical Biology, University Hospital Antwerp, Edegem, Belgium, ⁴Gastroenterology and Hepatology, University Hospital Antwerp, Edegem, Belgium.

Background and aims: Insulin resistance (IR) is increasingly present in people with type 1 diabetes (T1D), and is associated with an increased risk of cardiovascular disease. The hyperinsulinemic-euglycemic clamp (HEC) is the gold standard to assess IR but is unsuited for clinical practice. Thus, easily-applicable, non-invasive alternatives are needed. The estimated glucose disposal rate (eGDR) is a proposed alternative but is an indirect measurement. We propose a ¹³C breath test as a direct index.

Materials and methods: We performed a 240 minute HEC (40 mU/min/m² with a priming dose of 80 mU/min/m²) with blood glucose clamped at 90 mg/dl (5 mmol/l). Final 30 minutes of the glucose infusion rate were used to determine the M-value (mg/kg/min) according to the protocol of DeFronzo et al. A ¹³C glucose breath test was performed, using 25 mg of universally labeled glucose (Cambridge Isotope Laboratories Inc., USA), obtaining quarter-hourly breath samples up to 150 minutes after ingestion. Samples were analyzed using an isotope ratio mass spectrometer (Abca 2, Sercon). The expired ¹³CO₂ after test drink ingestion was compared with the baseline value. Results are expressed as absolute increase (Δ %) in ¹³C after 90 minutes. We performed pairwise comparisons of the diagnostic accuracy of the ¹³C breath test and the eGDR (mg/kg/min) using the HEC as reference method. eGDR = 21.158 − (0.09 * waist circumference [WC]) − (3.407 * hypertension) − (0.551 * HbA1c [% DCCT].

Results: We included 48 subjects (62 % male, mean age 45 ± 15 years, HbA1c 7.3 ± 0.8 %, BMI 27.5 ± 5.0 kg/m²). M-value was 5.50 ± 3.05 mg/kg/min, Δ¹³C was 6.25 ± 2.21 %, eGDR was 5.83 [5.34] mg/kg/min. A significant correlation between Δ¹³C and M-value (r = 0.73, p < 0.001, figure) was found. The correlation between M-value and eGDR was weaker (Spearman r = 0.66, p < 0.001). Both M-value (r = -0.72, p < 0.001) and Δ¹³C (r = -0.67, p < 0.001) correlated stronger with waist circumference than with BMI (M-value r = -0.64, p < 0.001; Δ¹³C r = -0.49, p < 0.001). No significant correlation with HbA1c was present, both for M-value and Δ¹³C. Linear regression showed a significant association between Δ¹³C and M-value, adjusted for age, HbA1c, and gender (B = 1.11, adjusted R² = 0.52, p < 0.001). We found an AUROC of 0.78 for Δ¹³C to identify subjects with IR (M-value < 4.9 mg/kg/min). The optimal cut-off point was < 6.61 Δ % (sensitivity 0.77, specificity 0.62). DeLong pairwise comparison did not find a significant difference between AUROC^13C and AUROC^eGDR (p = 0.369).

Conclusion: The ¹³C breath test shows important potential as an indicator for IR in T1D and correlates more strongly with the HEC compared to the eGDR.

Clinical Trial Registration Number: NCT04623320

Supported by: Bijzonder Onderzoeksfonds (BOF), research grants from AstraZeneca and Boehringer Ingelheim

Disclosure: J. Mertens: Grants; Research grant AstraZeneca, Research grant Boehringer Ingelheim, Bijzonder Onderzoeksfonds (BOF).

OP 14 Regulation of islet hormone secretion in type 1 diabetes and type 2 diabetes

79

Loss of electrical beta to delta cell coupling underlies impaired hypoglycaemia-induced glucagon secretion in type 1 diabetes

R. Gao ¹, T. Hill¹, A. Benrick², L. J. B. Briant¹, L. Kothegali², H. Dou², Q. Zhang¹, P. Rorsman^1,2;

¹OCDEM, University of Oxford, Oxford, UK, ²Institute of Neuroscience and Physiology, University of Gothenburg, Gothenburg, Sweden.

Background and aims: Glucagon secreted from pancreatic α-cells is a principal glucose counterregulatory hormone. In type 1 diabetes (T1D), hypoglycaemia-induced glucagon secretion is severely impaired, leading to prolonged and fatal hypoglycaemia. How this defect develops is unknown but it may be due to aberrant intra-islet paracrine tone. Somatostatin released by δ-cells is a potent paracrine inhibitor of α-cell secretion thus somatostatin hypersecretion could result in glucagon failure in T1D. Using a non-obese diabetic (NOD) mouse model, we studied the islet-centric mechanisms that contribute to glucagon secretory defects in T1D.

Materials and methods: Islet hormone secretion was assessed using static incubation. Intracellular Ca²⁺ was monitored using live imaging in islets expressing GCaMP6f specifically in δ-cells. Islet cell membrane potential was recorded using perforated patch whole-cell technique. Plasma glucose and glucagon were measured in diabetic (NOD) and non-diabetic (ND) mice during insulin tolerance test.

Results: Following the onset of diabetes, NOD islet insulin content was <10% of ND controls. This was accompanied by impaired low glucose-stimulated glucagon secretion (45% of ND) and elevated basal somatostatin release (>400% of ND). Treatment with a somatostatin receptor antagonist (SSTRa) rescued glucagon secretory failure in NOD islets by exerting a more profound stimulation on glucagon secretion (316% in ND; 772% in NOD). Consistently, insulin-induced hypoglycaemia in vivo failed to elevate plasma glucagon in NOD mice but was rescued by SSTRa. These data indicate that excessive somatostatin release at low glucose may contribute to defective glucagon secretion in T1D. In human T1D islets, lowering glucose from 6 to 1 mM did not stimulate glucagon secretion as that in human ND islets (89% in ND; 3.5% in T1D), whereas basal somatostatin secretion was significantly elevated. To investigate the underlying mechanisms of dysregulated islet hormone release in T1D, we mimicked the immune attack of islets by pretreatment with a cytokine cocktail (CC). CC pretreatment for 24h replicated the secretory defects in T1D and reduced Gjd2 expression (encodes connexin-36 (CX-36) that forms gap junctions in β- and δ-cells) by 60% (p=0.048). Antagonising CX-36 with mefloquine enhanced somatostatin secretion (p=0.009) and δ-cell Ca²⁺ activity (p=0.017) at 1 mM glucose, suggesting that CX-36-mediated electrical coupling contributes to normal basal δ-cell activity. Using an optogenetic mouse model expressing light-gated Cl^- pumps specifically in β-cells, we found δ-cells were responsive to light radiation with an outward current (1.5±0.2pA), membrane repolarisation and reduced electrical activity. These data suggest hyperpolarisation of β-cells at low glucose may spread into neighbouring δ-cells via CX-36 and function as an ‘electric brake’; reduction in gap junction conductivity or β-cell destruction in T1D may remove the β-cell hyperpolarising influence, contributing to basal somatostatin hypersecretion.

Conclusion: CX-36-mediated electrical coupling between β- and δ-cells plays a role in appropriate basal somatostatin secretion and normal islet paracrine tone. Disruption of this communication contributes to somatostatin hypersecretion and aberrant paracrine inhibition of α-cell glucagon secretion seen in T1D.

Supported by: MRC, Helmsley Trust

Disclosure: R. Gao: None.

80

Glucose inhibits glucagon secretion of mice with alpha cell specific deletion of KATP channels

M. Parambath, B. Lai, H. Chae, E. Gatineau, P. Gilon;

Pôle d'Endocrinologie, Diabète et Nutrition, Université Catholique de Louvain, Brussels, Belgium.

Background and aims: Glucose stimulates insulin and somatostatin (SST) secretion by closing β- and δ-cell K_ATP channels. By contrast, it inhibits glucagon release by largely unknown mechanisms. The role of α-cell K_ATP channels in the control of glucagon release and their implication in the glucagonostatic effect of glucose are hotly debated. They are particularly difficult to evaluate because α-, β- and δ-cells mutually influence their secretions. To study the role of α-cell K_ATP channels in the control of glucagon release, we generated α-K_ATPKO mice that lack K_ATP channels specifically in α-cells and compared their glucagon secretion in response to a change of the glucose concentration between 1 and 7 mM to that of control islets. To evaluate the role of SST that is released from δ-cells upon K_ATP channel closure, we used two mouse models that lack SST, i.e. SSTKO and SSTKO/ α-K_ATPKO mice.

Materials and methods: The following mouse models were used: Kir6.2flox (control), α-K_ATPKO (GluCre x Kir6.2flox), SSTKO and SSTKO/α-KATPKO. To measure [Ca²⁺]_c (by confocal microscopy) in α-cells within islets, we used the following mice expressing the fluorescent Ca²⁺ sensor, GCamP6, in α-cells: α-GCamP6 (GluCre x LSL-GCamP6) and α-GCamP6/α-KATPKO (GluCre x Kir6.2flox x LSL-GCamP6). All experiments were performed in the presence of a 6 mM amino acid mixture (2 mM alanine, 2 mM arginine and 2 mM glutamine).

Results: Increasing the glucose concentration from 1 to 7 mM inhibited glucagon release of α-K_ATPKO and control islets expressing or not SST, suggesting that glucose can control glucagon secretion independently of α-cell K_ATP channels and SST. These changes in glucagon secretion were associated with parallel but modest changes in α-cell [Ca²⁺]_c. The addition of the K_ATP channel opener, diazoxide (250 μM), at 1 mM glucose suppressed glucagon release and decreased α-cell [Ca²⁺]_c of control islets. Interestingly, it stimulated glucagon secretion and increased α-cell [Ca²⁺]_c of α-K_ATPKO islets, whereas it was without effect in SSTKO/α-K_ATPKO islets. This suggests that pharmacological opening of α-cell K_ATP channels suppresses glucagon release whereas opening of δ-cell K_ATP channels stimulates glucagon secretion by alleviating the glucagonostatic effect of SST released by δ-cells.

Conclusion: Glucose can inhibit glucagon secretion independently of α-cell K_ATP channels and independently of SST. The strong inhibitory effect of glucose associated with a very modest decrease in α-cell [Ca²⁺]_c suggests that glucose controls glucagon release by a mechanism that is partly independent from α-cell [Ca²⁺]_c changes. The observation that diazoxide stimulates glucagon secretion of α-K_ATPKO islets but not that of SSTKO/α-K_ATPKO islets illustrates a strong tonic inhibition of glucagon release by SST, even at low glucose.

Supported by: FNRS, EFSD/JDRF/Lilly Programme, ARC, SFD

Disclosure: M. Parambath: None.

81

Recombinant PSP/reg treatment improves beta cell function in islets isolated from glucose intolerant and type 2 diabetic donors

I. Louvet ¹, M. Moreno-Lopez¹, A. Acosta-Montalvo¹, V. Gmyr¹, J. Kerr-conte¹, T. Reding², C. Saponaro¹, F. Pattou¹, R. Graf², C. Bonner¹;

¹Université de Lille, CHU Lille, Institut Pasteur de Lille, Inserm UMR1190 - EGID, Lille, France, ²Department of Surgery, University Hospital Zurich, Zurich, Switzerland.

Background and aims: Pancreatic islet dysfunction and demise are key characteristics of type 1 diabetes (T1D), type 2 diabetes (T2D), and HNF1A-MODY, respectively. Pancreatic Stone Protein / Regenerating Protein (PSP/reg) is regulated at the transcription level by the Hepatocyte Nuclear Factor family of proteins and physiologically secreted from pancreatic acinar cells. Upon focal or systemic extra-pancreatic inflammation, such as sepsis, PSP/reg is strongly increased, but also highly elevated in the serum of T1D, TD2, and HNF1A-MODY patients. Although, PSP/reg is expressed and induced in islets of diabetic mice, the mechanisms involved remain largely unknown. Therefore, the aim of this study was to investigate the expression and regulation of PSP/reg in the human pancreas and islets isolated from donors covering a wide spectrum of metabolic disease compared to normoglycemic controls.

Materials and methods: PSP/reg protein expression was analysed from human paraffin embedded pancreatic sections and isolated islets embedded in histogel by immunofluorescence. PSP/reg gene and protein expression was analysed by qPCR and Western Blotting, respectively. Perifusion techniques were used to determine glucose-stimulated-insulin-secretion (GSIS) in islets isolated from donors with glucose intolerance or T2D in response to a chronic treatment with recombinant PSP/reg (rPSP/reg).

Results: PSP/reg protein was highly enriched in the acinar cells of all donor pancreata studied. Islets embedded in histogel revealed that PSP/reg colocalized with insulin in beta cells, somatostatin in delta cells and, to a lesser extent, with glucagon in alpha cells. Of note, PSP/reg protein expression was significantly induced in islets isolated from T2D donors with obesity compared to those of obese (normoglycemic), glucose intolerant and normoglycemic donors. Moreover, the chronic treatment of rPSP/reg (24 h) enhanced both, first and second phase GSIS from islets isolated from glucose intolerant donors, while only the second phase was enhanced from those of T2D donors.

Conclusion: Collectively, these data indicate that PSP/reg may play an important role in preserving beta cell function during the early stages of the disease. They also suggest that rPSP/reg could be a promising therapy for the treatment of diabetes, but further studies in diabetic mice to determine its efficacy to reduce hyperglycemia and improve islet function are warranted.

Supported by: EFSD/NNF Diabetes Precision Medicine Award Programme

Disclosure: I. Louvet: None.

82

Functional genomics highlights type 2 diabetes-associated purinergic receptor P2RY1 as a new modulator of insulin secretion in humans

A. Dance ¹, M. Derhourhi¹, M. Boissel¹, M. Marre^2,3, G. Charpentier⁴, A. Khamis^1,5, P. Froguel^1,5, A. Bonnefond^1,5;

¹59, UMR1283 EGID, Lille, France, ²75, Inserm U1138, Centre de Recherche des Cordeliers, Paris, France, ³75, CMC Ambroise Paré, Paris, France, ⁴91, CERITD (Centre d'Étude et de Recherche pour l'Intensification du Traitement du Diabète), Evry, France, ⁵Department of Metabolism, Imperial College, London, UK.

Background and aims: Genome-wide association studies (GWAS) have shown that the P2RY1 locus is associated with type 2 diabetes (T2D) risk and with glucose levels. P2RY1 is a G-protein coupled receptor (GPCR) activated by ATP and ADP, which is highly expressed in pancreatic islets, particularly in β cells where, ATP /ADP stimulate insulin secretion by activating the ATP-dependent potassium channels. Through functional genetics, we aimed to analyze the putative contributions of genetic variants of P2RY1 to T2D risk and to other metabolic traits. We then characterized the downstream signaling pathways of P2RY1 in human pancreatic beta cells and deciphered its putative role in insulin secretion.

Materials and methods: P2YR1 was sequenced in 6,348 French adults including cases with T2D and normal glucose controls. To assess the functional effect of each identified variant, we performed 1/ luciferase assays (NFAT-RE system) on HEK293 cells overexpressing each variant, followed by P2YR1 activation by increasing doses of MRS2365 (methanocarba-2MeSADP) that is a specific agonist of P2RY1, and 2/ immunofluorescence to assess cellular localization of each mutant. We also performed expression quantitative trait loci (eQTL) analysis in 116 donors based on RNA-seq of pancreatic islets and DNA microrrays genotyping. In human pancreatic β cells (EndoCβH5), we performed glucose-stimulated insulin secretion (GSIS) assays coupled to P2YR1 activation by MRS2365. GSIS on EndoCβH5 cells transfected with siRNA P2RY1 (or control) are also being performed.

Results: We identified 17 rare missense variants in P2RY1 (including 6 novel variants). Our in vitro analyses highlighted 5 loss-of-function mutations. 80% of the mutation carriers presented with T2D. In the UK Biobank we found 10 rare null variants (nonsense or frameshift) that were mostly carried by patients with T2D (75%). Expression QTL analyses showed that a block of single nucleotide polymorphisms (SNPs; n=63) located in an enhancer of human islets was significantly associated with increased islet expression of P2RY1 and decreased T2D risk while another block of SNPs (n=51) was significantly associated with decreased expression of P2RY1 in human islets and increased T2D risk. In EndoCβH5, the GSIS assay showed that the P2RY1 specific agonist MRS2365 led to a 30% increase of insulin secretion when stimulated with 20mM of glucose. A kinome analysis (through PAMGENE technology) and transcriptomic analysis (through RNA sequencing) of the P2RY1 pathway in the EndoCβH5 cells in response to MRS2365 agonist or P2RY1 siRNA is in progress at high glucose levels.

Conclusion: Our genetic, functional genomic and pharamacological studies suggest that P2RY1 dysfunction is causatively associated with T2D riskand that P2RY1 contributes to the activation of the insulin secretion pathway. P2RY1 potent and selective agonists that cannot cross the blood-brain barrier are available, and could be tested as a potential new class of insulin secretagogues.

Disclosure: A. Dance: None.

83

A peptide-amino acid transporter axis regulates pancreatic beta cell maintenance and function

M. Nakazato, W. Zhang, H. Sakoda;

Faculty of Medicine, University of Miyazaki, Miyazaki, Japan.

Background and aims: Postprandial elevations in plasma amino acids enhance insulin release from pancreatic β cells. Several amino acid transporters (AAT) are expressed in β cells and stimulate postprandial insulin secretion by mediating the influx of amino acids. No endogenous substance that modulates the activity of AATs to regulate β-cell function and maintenance has been identified.

Materials and methods: We generated apoaequorin transgenic mice to detect intracellular Ca²⁺ mobilization under the control of the CAG promoter. Using the pancreata of these transgenic mice, we identified a novel 19-amino acid peptide named neuroendocrine regulatory peptide (NERP) derived from a granin protein VGF. VGF has dual functions in β-cells: regulating secretory granule formation and functioning as a multiple peptide precursor.

Results: Immunogold electron microscopy revealed colocalization of NERP with insulin in the storage granules of β cells. Immunohistochemistry and chromatography combined with radioimmunoassay for NERP verified the presence of NERP in human islets. NERP stimulated Ca²⁺ influx in β-cell-derived MIN6 cells. NERP enhanced glucose-stimulated insulin secretion (GSIS) from human and C57BL/6J mouse pancreatic islets and MIN6 cells. To identify NERP target candidates in β cells, we used ligand-receptor glycocapture technology, LRC-TriCEPS, a highly specific and sensitive pull-down assay involving binding to glycosylated membrane proteins. The volcano plot analysis showed a neutral amino acid transporter satisfied the criteria for specific binding to TriCEPS. Overexpression of the AAT in HEK293 cells augmented NERP binding. NERP binding to the AAT stimulates glutamine, alanine, and proline uptake into β-cells, thereby enhancing mitochondrial ATP production and reducing oxidative and endoplasmic reticulum stress. Kinetics study using Michaelis-Menten analysis indicates that NERP acts as a positive allosteric modulator on the AAT by enhancing the affinity of the AAT to import amino acids. Interestingly, NERP does not stimulate amino acids uptake under the low glucose condition. Silencing of the AAT in MIN6 cells abolished NERP−induced GSIS and protective effects on mitochondrial respiration and cell apoptosis. NERP expression in β cells of obese, diabetic db/db mice was reduced, and NERP administration recovered impaired β-cell maintenance and insulin secretion in db/db mice. In MIN6 cells, NERP recovered β cell dysfunctions induced by siVgf silencing and cytotoxin administration. NERP-deleted Vgf (Vgf^{NERP-4 KO}) knockin to Vgf-knockdown MIN6 cells exhibited insufficient vesicle biogenesis in β cell compared with full-length Vgf (Vgf^full) knockin to Vgf-knockdown. NERP-deleted mice generated by the CRISPR-Cas9 system exhibited impaired proinsulin processing and the second-phase insulin secretion.

Conclusion: NERP is secreted together with insulin following glucose stimulation, and positively modulates the AAT activity to import amino acids into β cells. These findings provide a novel autocrine mechanism via the peptide-amino acid transporter axis in β-cell maintenance and function, and extend the understanding of the underpinnings regulating glucose metabolism.

Disclosure: M. Nakazato: None.

84

Genome-wide screening for regulators of degradation of insulin secretory granules with a fluorescent reporter

A. Kanai, Y. Nishida, K. Ueki, H. Uzawa, H. Watada;

Department of Metabolism and Endocrinology, Juntendo University Graduate School of Medicine, 2-1-1, Hongo, Bunkyo-ku, Tokyo, Japan.

Background and aims: To control the amount of insulin secretory granules (ISGs) in pancreatic beta cells, ISGs are degraded in lysosomes in beta cells, coined as crinophagy. Recently, several mechanisms associated with the degradation of ISGs have been proposed; however, the details of the molecular mechanisms and the pathophysiological significance are remained to be revealed. To clarify these issues, we performed genome-wide screening to identify canonical molecules regulating ISG degradation with a fluorescent reporter monitoring for the degradation.

Materials and methods: We constructed a fluorescent probe for ISG degradation, where two fluorescent proteins EGFP and mCherry, with different pH-sensitivities, tandemly conjugated to a cytoplasmic lesion of ZnT8, an ISG membrane-localized protein. After validation of the reporter, we performed CRISPR/Cas9-based genome-wide screening to identify regulating genes for ISG degradation.

Results: When an ISG is contained in an acidic condition of lysosomes, the fluorescence of EGFP quenches due to low pH, whereas mCherry remains intact; thus EGFP/ mCherry ratio decreases along with ISG degradation. We calculated the EGFP/mCherry ratio at each time point after KRB (Krebs Ringer Buffer) stimulation for ISG degradation, and the EGFP/mCherry ratio after KRB stimulation gradually decreased compared to controls, suggesting that EGFP fluorescence was quenched following degradation in lysosomes (EGFP/mCherry ratio at 8 hrs, 1.0 vs. 0.3±0.02, p<0.001). In addition, bafilomycin A₁ inhibited this decay of EGFP after KRB stimulation (EGFP/mCherry ratio, 0.3±0.02 vs. 1.0, p<0.001), demonstrating that the reporter reflects the ISG degradation process mediated by lysosomes. Next, with MIN6 cells expressing both Cas9 and the fluorescent reporter, we investigated the involvement of conventional Atg5/7-dependent autophagy. We knocked out canonical genes for regulating Atg5/7-dependent autophagy, such as Vps34, Beclin-1, Atg5, FIP200 and Atg9. Those cells showed the accumulation of p62 and inhibition of the conversion of LC3-I to LC3-II, showing the impairment of conventional macroautophagy. Interestingly, ISG degradation evaluated by the ZnT8-mCherry-EGFP probe was not disturbed in the cells lacking the canonical autophagy genes, suggesting that ATGs regulating conventional autophagy is not involved in ISG degradation. Finally, we performed genome-wide gRNA screening to identify genes regulating ISG degradation. We transduced lentivirus expressing genome-wide gRNA library in MIN6 cells stably expressing both Cas9 and ZnT8-mCherry-EGFP and enriched the cells lacking the ISG degradation with a flow cytometer. Next generation sequence (NGS) revealed enriched gRNAs to identify genes essential for ISG degradation, and we found no genes associated with conventional macroautophagy among 44 genes identified.

Conclusion: By performing the genome-wide screening in two replicates, we identified candidate genes for regulating ISG degradation. Interestingly, in gene ontology analysis, the pathways related to conventional autophagy are not enriched. The results suggest that alternative regulation of autophagy, instead of conventional macroautophagy, is involved in the degradation process of ISGs.

Supported by: Japan Society for the Promotion of Science

Disclosure: A. Kanai: None.

OP 15 Incretins and glucagon in action

85

Signs of glucagon resistance after a two-week hypercaloric diet intervention

M.P. Suppli ¹, J.I. Bagger¹, G. van Hall², M.B. Christensen¹, N.J. Wewer Albrechtsen³, J.J. Holst⁴, F.K. Knop¹;

¹Center for Clinical Metabolic Research, Gentofte Hospital, Hellerup, Denmark, ²Clinical Metabolomics Core Facility, Department of Clinical Biochemistry, Rigshospitalet, Copenhagen, Denmark, ³Department of Clinical Biochemistry, Bispebjerg and Frederiksberg Hospital, Copenhagen, Denmark, ⁴Novo Nordisk Foundation Center for Basic Metabolic Research, Faculty of Health and Medical Sciences, Copenhagen, Denmark.

Background and aims: Hyperglucagonaemia is observed in individuals with obesity and contributes to the hyperglycaemia of persons with diabetes. Hyperglucagonaemia may develop due to impaired hepatic amino acid turnover, which results in elevations of circulating glucagonotropic amino acids. We recently observed glucagon resistance and hyperglucagonaemia in individuals with obesity and hypothesised it to be a consequence of hepatic steatosis. Here, we evaluated whether glucagon resistance could be induced by a short-term hypercaloric diet intervention designed to increase hepatic fat content in healthy individuals.

Materials and methods: We recruited 20 healthy, lean individuals (BMI: 23 ± 0.3 (mean ± SD) kg/m², age: 28 ± 1 years) to follow a hypercaloric diet (~5,000 kcal/day) and a sedentary lifestyle for two weeks. Amino acid concentrations in response to infusion of high physiological glucagon (4.0 ng/kg/min) were assessed during a pancreatic clamp with somatostatin (450 μg/h) and basal insulin (0.1 mU/kg/min). The participants were examined before and after the intervention and again after eight weeks to assess the reversibility of any metabolic changes. Hepatic steatosis was assessed by magnetic resonance spectroscopy. Data were compared using paired t tests and were adjusted for multiple testing using the method of Benjamini and Hochberg, which controls the false discovery rate.

Results: The intervention led to increases in body weight (3.5 [2.8;4.2] (mean [95% CI]) kg, P < 0.0001) and hepatic fat content (382 [206;705]%, P < 0.001). Insulin resistance was evident with unchanged fasting blood glucose levels and a 56% increase in fasting insulin (P < 0.001). Glucagon infusion led to a decrease in the concentration of total amino acids on all experimental days, but the percentage change in total amino acids was reduced (−2.5 ± 0.5 (mean ± SEM) vs. −0.2 ± 0.7%, P = 0.015) and the average slope of the decline in the total amino acid concentration was less steep (−2.0 ± 0.3 vs. −1.2 ± 0.3 μM/min, P = 0.016) after the intervention compared with baseline (Fig. 1). Body weight remained elevated at follow-up, while all other metabolic changes were normalised.

Conclusion: Our results indicate that short-term unhealthy behaviour, which increases hepatic fat content, causes a reversible resistance to the effect of glucagon on amino acid concentrations in healthy individuals, which may explain the hyperglucagonaemia associated with obesity and diabetes.

Clinical Trial Registration Number: NCT04859322

Supported by: Danish Diabetes Academy, which is funded by the Novo Nordisk Foundation, grant number NNF17SA0031406

Disclosure: M.P. Suppli: None.

86

Estimation of insulin resistance and beta cell function by i.v. glucagon-insulin tolerance test in novel subgroups of adult-onset diabetes in Vaasa diabetes registry

A. Laitinen ^1,2, A. Käräjämäki¹, A. Käräjämäki¹, H. Liisa^2,3, S. Kurki^2,4, L. Kaj¹, E. Ahlqvist⁴, T. Tuomi^3,5;

¹Vaasa Central Hospital, County of Ostrobothnia, Vaasa, Finland, ²University of Helsinki, Helsinki, Finland, ³Folkhälsan Research Center, Helsinki, Finland, ⁴Lund University, Malmo, Sweden, ⁵Helsinki University Hospital, Helsinki, Finland.

Background and aims: Five cluster-based subgroups of adult-onset diabetes are based on age at diagnosis, BMI, HbA1c, GAD-autoantibodies, and HOMA indices calculated from plasma glucose (PG) and C-peptide. We aimed to estimate their long-term differences in insulin secretion, insulin resistance, and liver fat content.

Materials and methods: From a regional Finnish diabetes registry Direva (N>7000) we invite 50-100 participants per subgroup with cluster assigned at registration <2 years from diagnosis and duration 4-10 years. Methods: an i.v. glucagon (0.5mg at time 0 min) - insulin (0.05 IU/kg at 40 min) tolerance test with C-peptide and PG measured at 0 and 6 minutes post-glucagon (insulin secretion), and PG at 0, 5, 11, 20 min post-insulin for K_ITT (glucose disappearance) to estimate insulin resistance; liver fat content with FibroScan®.

Results: 88 patients have participated (median duration 6.2 years): 21 SAID, 9 SIDD and 22 SIRD (severe autoimmune/insulin-deficient/insulin-resistant diabetes), 16 MOD and 20 MARD (mild obesity-/age-related diabetes). HbA1c decreased from baseline in SIDD [88.0 (95% CI 38-124) vs. 54 (34-80) mmol/mol; p=0.046, Wilcoxon signed-rank test], but increased in all other groups by 2.5-7.0 mmol/mol. C-peptide response to glucagon was greatest in SIRD with a significant difference against SAID and MARD [1.15 (0.84-1.42) vs. 0.41 (0.06-0.81 nmol/l), p<0.001; and 0.59 (0.44-0.95), p=0.029, respectively] using one-way ANOVA. SIDD group had a median of 0.66 nmol/l (0.34-1.21, not significant). Similarly, HOMA2β was highest in the SIRD [128.9 (99.7-177.5)] and lowest in the SAID [58.3 (CI 27.9-100.2), p<0.001] and SIDD [58.4 (49.8-99.6), p<0.001] groups (Kruskal-Wallis test). The C-peptide response correlated moderately with the HOMA2β (r=0.52, Spearman, p<0.001). K_ITT was highest in the SAID and SIDD groups but the variation was high and no significant differences were observed between the groups. HOMA2-IR was highest in the SIRD [2.96 (2.23-3.77)] and lowest in the SAID group [1.75 (0.61-2.22), p=0.001 (Kruskal-Wallis test)]. K_ITT had a weak negative correlation with HOMA2-IR [r=-0.24 (Spearman´s correlation, p=0.04]. The liver fat content was highest in the SIRD [321 Db/m, (95% CI 289-346 Db/m)] and lowest in the SAID group [281 Db/m (95% CI 277-302 Db/m), p=0.074, one-way ANOVA], while in the MOD group mean was 301 Db/m (95% CI 262-348 Db/m). Liver fat content correlated better with the HOMA2-IR index than with K_ITT [r=0.447, p<0.001 and r=-0.28, p=0.016, respectively, Spearman´s correlation].

Conclusion: The preliminary data of the ongoing study indicate that the subgroups differ notably in insulin secretion and resistance even years after diagnosis. Especially the SIRD group stood out having higher insulin secretion and resistance by every method, as well as high liver fat content. Moreover, our results strengthen the hypothesis that HOMA2-IR reflects the liver insulin resistance and liver fat content better than K_ITT derived from insulin tolerance test. The correlation between the C-peptide response to glucagon and HOMA2-β index was moderate.

Supported by: Vaasa Medical Foundation, The Finnish Medical Foundation, Turunmaa Duodecim Foundation, Tuovisen säätiö Foundation

Disclosure: A. Laitinen: Grants; Vaasa Medical Foundation, The Finnish Medical Foundation, Turunmaan Duodecim Foundation, Tuovisen Säätiö Foundation.

87

Role of K _ATP channels in regulating glucagon secretion under low and high glucose conditions

E. Lee ¹, R. Hatano¹, T. Noda¹, S. Acreman², Q. Zhang², T. Miki¹;

¹Chiba University, Chiba, Japan, ²Churchill Hospital, Oxford, UK.

Background and aims: K_ATP channel is essential for sensing both hypoglycemia and hyperglycemia to regulate insulin secretion from pancreatic β-cells. However, its role in regulating glucagon secretion from α-cells in response to glycemic alterations remains elusive. We previously reported that brain K_ATP channel is critical for regulating glucose homeostasis during severe glucose deprivation through regulating glucagon secretion. However, several papers also suggest that K_ATP channel in α-cells plays an important role in sensing hypoglycemia and triggering glucagon secretion. To further explore the involvement of brain K_ATP channels in regulating glucagon secretion under different glucose conditions, we utilized β-specific, neuron-specific, and null-K_ATP channel deficient mice and analyzed brain activation and secretion of glucagon and insulin.

Materials and methods: We generated β-specific, neuron-specific, and null-K_ATP channel deficient mice (βKO, nKO, and null KO, respectively) by tissue-specifically disrupting Kir6.2 gene that encodes the pore-forming subunits of K_ATP channels. To evaluate its role on glucagon secretion during neuroglycopenia, the mice were administered with 2-deoxyglucose (2DG, 400 mg/kg) intraperitoneally (i.p.) and plasma glucagon levels were measured in the portal vein. Neural activation in response to glycemic changes, c-fos immunostaining was performed in the hypothalamus. We also investigated the role of brain K_ATP channel in suppressing glucagon secretion in response to hyperglycemia by measuring plasma glucagon levels after i.p. glucose administration (2g/kg).

Results: In fed state, there was no significant difference observed in body weight and blood glucose levels in βKO, nKO, and null-KO compared to those in WT. Glucagon secretion in response to 2DG was found to be increased compared with normal saline (NS) injection similarly between WT (NS vs 2DG, p>0.001) and βKO (NS vs 2DG, p<0.001). However, it was significantly impaired in nKO and null-KO. Consistent with these results, 2DG-induced c-fos induction in the nucleus tractus solitarius (NTS) was found to be attenuated in nKO and null-KO compared with that in WT (NS vs 2DG, p<0.001). Intriguingly, plasma glucagon levels were significantly suppressed by i.p. glucose in WT (NS vs glucose, p<0.001), βKO (NS vs glucose, p<0.01) and nKO (NS vs glucose, p<0.05), but there was no suppression in null-KO. In addition, insulin levels following i.p. glucose increased significantly in WT (NS vs glucose, p<0.001) and nKO (NS vs glucose, p<0.001), but not in βKO and null-KO.

Conclusion: These findings suggest that neural K_ATP channel plays critical role in regulating glucagon secretion during neuroglycopenia by activating the neurons in the NTS. In addition, K_ATP channels in the cells other than neurons or β-cells is essential for glucose-induced glucagon suppression. In the meantime, K_ATP channels in β-cells rather than those in neurons play the primary role in regulating glucose-stimulated insulin secretion.

Supported by: Scientific Research Grants from the Ministry of Education, Culture, Sports, Science and Technology

Disclosure: E. Lee: None.

88

Glucose-dependent insulinotropic polypeptide and glucagon-like peptide-1 do not account for the entire incretin effect in healthy humans

S. Veedfald ^1,2, Y. Sun¹, C. Xiang¹, C. Xie¹, T. Wu¹, M. Bound¹, J. Grivell¹, B. Hartmann², K. Jones¹, M. Horowitz¹, J.J. Holst², C.K. Rayner¹;

¹University of Adelaide, Adelaide, Australia, ²Department of Biomedical Sciences, University of Copenhagen, Copenhagen, Denmark.

Background and aims: Glucose, given orally or enterally, elicits much greater insulin secretion than an intravenous (IV) glucose infusion yielding the same level of glycemia. This ‘incretin effect’ is driven by the release of insulinotropic peptides (‘incretins’) from the gut. We aimed to determine whether the known incretins, glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide-1 (GLP-1), account for the entire incretin effect in healthy humans, using specific antagonists for their receptors.

Materials and methods: Twelve healthy participants (10 males/2 females; (mean (SD); 29 (5) years; BMI 24 (4) kg/m²;) were studied on five occasions in a randomized, double-blinded fashion, following an overnight fast. On each day, a naso-duodenal catheter was positioned and blood glucose was maintained 5 mmol/L above fasting concentrations by a variable IV glucose infusion (t = 0-165min). On four study days, receptor antagonists for GIP (GIP_3-30, 800pmol/kg/min), GLP-1 (Exendin 9-39 (Ex_9-39), 600 pmol/kg/min), both, or neither (saline) were infused IV (t = 45-165min). On these four days, glucose (45g) was also infused intraduodenally (ID) (3kcal/min, between t = 75-135 min). On a fifth day, saline was infused IV and ID, while maintaining elevated blood glucose as on the other days, to determine baseline insulin (C-peptide) secretion. ‘Arterialized’ venous blood was sampled at t=0, 30, 45, 60, 75, 80, 90, 105, 120, 135, 150 and 165min for measurements of plasma C-peptide. The incretin effect was evaluated by the incremental area under the C-peptide curve between t = 75-165 min (iAUC_75-165min) after subtracting baseline secretion, and contributions of GIP and GLP-1 assessed by comparing iAUC_75-165min with the relevant antagonists, using ANOVA.

Results: Glycemia was maintained at mean 9.3 (0.9) mmol/L (mean (SD) between t = 0-165 min and was comparable between the 5 study days. C-peptide secretion was reduced by the incretin receptor antagonists (Ex_9-39+GIP_3-30 vs saline; 71(20) vs 227 (43) nmol/L x min (mean(SEM), p=0.001). The reduction in the incretin effect did not differ between the GLP-1 and GIP antagonist days (155(32) vs 142 (33) nmol/L x min (mean(SEM), p=0.95.

Conclusion: In healthy humans, GIP and GLP-1 contribute approximately equally to the incretin effect. When antagonizing both GIP and GLP-1 receptors, a substantial incretin effect remains, suggesting the existence of other as of yet unknown incretin(s).

Clinical Trial Registration Number: ACTRN12619000382178

Supported by: NHMRC, IRF

Disclosure: S. Veedfald: None.

89

Restoration of impaired portal glucose sensing by targeted manipulation of GLP-1r density in a translational model of insulin resistance

C. Malbert ¹, R. Allouche², M. Horowitz³, K.L. Jones³;

¹Human Nutrition, Aniscan, INRAE, Saint-Gilles, France, ²Paltech, Paris, France, ³Adelaide medical school, Adelaide, Australia.

Background and aims: The portal glucose sensor informs the brain of changes in glucose inflow via vagal afferents that are dependent on the glucagon-like peptide-1 (GLP-1) receptor (GLP-1r). We have shown that GLP-1r expression within the portal vein is markedly reduced in a translational model of insulin resistance (IR), associated with altered glucose signaling to the brain. We now investigated the potential for restoring reduced portal GLP-1r expression in IR animals using a targeted infusion of bioactive molecules, which have been demonstrated to increase GLP-1r expression in vitro.

Materials and methods: Five groups, each of five miniature Yucatan minipigs, aged three years, were used. One group was maintained lean and insulin sensitive, while the remaining four were made IR by a high fat-high sucrose diet for 4 months. Insulin sensitivity was determined in all animals with a euglycaemic-hyperinsulinaemic clamp. The precise portal location of the low-density GLP-1r area (compared to lean animals) was initially defined using PET/CT imaging after the administration of 68Ga-DO3A- exendin-4 radioligand and a catheter, exiting in the portal connective tissue, was then fixed at this location during laparoscopy. This catheter was used to infuse continuously either saline, dihydrotestosterone (DHT, 10 μg/kg/24H), metformin (MET, 2 mg/kg/24H), or exenatide (EX, 0.01 μg/kg/24H), i.e., molecules known to increase GLP-1r density in vitro. After 2 months continuous infusion, PET/CT imaging was repeated in all animals using the same GLP-1r radioligand. Vt coded images, the quantitative metric of receptor density, were obtained from PET/CT concurrently with monitoring of the arterial input function extracted from an arteriovenous shunt and radioHPLC of the authentic ligand in the plasma. Duodenal and pancreatic Vt were also computed as references of large GLP-1r expression organs.

Results: In IR animals, there was a marked reduction in GLP-1r density at the portal vein (p<0.05), but not in the pancreas or duodenum (see table). Treatment with DHT increased GLP-1r density at the portal vein substantially to be comparable to that in lean animals. The other treatments had no effect on portal GLP-1r density. Furthermore, no treatment affected the GLP-1r density in the pancreas or duodenum.

Conclusion: Localized administration of DHT, but not MET or EX, normalises portal GLP-1r density in IR animals, without affecting GLP-1r density in other organs. Accordingly, it is possible to restore impaired glucose sensing in IR animals. The implications for the pathogenesis and optimal management of insulin resistance /type 2 diabetes in humans now require evaluation, especially as DHT can be delivered for extended periods using a controlled drug release pellet.

Supported by: This work was made possible by a grant from Paltech

Disclosure: C. Malbert: None.

90

Body weight independent effects of CT-868, a signalling biased dual GLP-1/GIP receptor modulator, on glucose homeostasis in overweight and obese adults with type 2 diabetes

M. Hompesch ¹, M. Elliott², M. Hernandez¹, S. Hansen², A. Macias¹, M. Chakravarthy²;

¹Prosciento Inc., San Diego, CA, USA, ²Carmot Therapeutics, Berkeley, CA, USA.

Background and aims: GLP-1 and GIP can deliver complementary pharmacology when combined. CT-868 is a biased dual GLP-1 and GIP receptor modulator that exhibits no beta-arrestin coupling or receptor internalization at either receptor. The primary aim of this study was to examine the body weight-independent effects of CT-868 on insulin (INS) secretion via graded glucose infusion (GGI), glucose (GLUC) homeostasis in response to a mixed meal tolerance test (MMTT), gastric emptying (GE) by acetaminophen absorption, and ad libitum food intake (FI), in people with type 2 diabetes (T2D).

Materials and methods: A total of 20 overweight and obese adults with T2D were split into 2 groups: CT-868 vs placebo (PL), n=7 (group 1) and CT-868 vs PL vs liraglutide (Lira), n = 13 (group 2). CT-868 vs PL in both groups 1 and 2 was studied in a randomized, double-blind 2-period crossover fashion, and Lira in group 2 was administered in a fixed 3^rd period. Mean age and BMI were 52.2 years and 32.7 kg/m² with 55% males. Treatments were administered daily for only 4 days to minimize weight loss with a washout of 14 days in between each period.

Results: GGI data showed robust and significantly increased INS secretion for CT-868 vs PL [mean change (SD) in ISR/GLUC at GIRmax: CT-868 1.1(0.1) vs PL 0.1(0.2), p <0.01; Lira 1.0(0.3) pmol/kg/min]. MMTT based GLUC AUC_0-240min was significantly reduced for CT-868 vs PL [72(46) vs 392(48) mmol/L*min; p <0.05] and numerically lower vs. Lira [187(55) mmol/L*min]. GLUC lowering was accompanied by significantly decreased INS AUC_{0- 240min} in the CT-868 treated group [1178(846) mU/L*min], versus both PL [5833(873) mU/L*min] and Lira [4613(1253) mU/L*min; p <0.05]. As both CT-868 and Lira had similar delayed GE vs PL, the concomitantly reduced GLUC and INS excursions seen with CT-868 suggests improved INS sensitivity or enhanced INS independent GLUC disposal. CT-868 demonstrated minimal suppression of glucagon during MMTT vs Lira that showed glucagon lowering. Food intake (both absolute amount and total calories) was lower for CT-868 vs PL accompanied by reduction in hunger and appetite scores as assessed by visual analog scales. There were no significant body weight changes in any treatment period, as anticipated by design. The most frequent side effects reported for CT-868 were gastrointestinal (nausea, diarrhea), which were mostly mild and transient. There were no discontinuations in the study due to treatment-related adverse events.

Conclusion: Taken together, these data support a robust weight-independent effect of the fully biased dual GLP-1/GIP modulator CT-868 on glucose disposal with minimal suppression of glucagon, as compared to liraglutide, an unbiased GLP-1 receptor agonist. Further delineation of CT-868’s longer term effects in overweight and obese patients with both T2D and T1D is underway.

Clinical Trial Registration Number: NCT04973111

Disclosure: M. Hompesch: Employment/Consultancy; employment.

OP 16 What new in type 1 diabetes treatment: focus on the islet cells

91

Treatment of 355 early type 1 diabetes with a combination of multiple immune modulatory drugs

Z. Talaeerad, K. Khodabakhshi Pirkalani;

Mehr Medical Group, Tehran, Islamic Republic of Iran.

Background and aims: Type I DM is regarded as not reversible. We and others have shown elsewhere, that residual beta cell activity remains for a long time. Three interconnected but not identical aims can be explained in early disease. Calming the auto immune process as exemplified by reduction of autoantibodies titer; increasing the Insulin output as exemplified by C-Peptide and reestablishment of a normal metabolic state as exemplified by reduced Insulin need and better HbA1C level.

Materials and methods: Three hundred and fifty five consecutive patients were enrolled. Seven landmarks including serum fasting insulin and C-Peptide, 2hpp Insulin and C-Peptide HbA1C and anti-Gad and anti-Islet cell antibody were measured on day zero and every 3 months for at least 12 months. Patients received a combination of Pioglutazone, Metformin, Empaglifolizine as pancreas calming protocol and Atorvastatin, Vitamin A and Azithromycin as immune. The immune modulatory effect of all three mentioned agents are known and they were selected because of low side effect profile on long term, synergistic effect in preclinical trials and action on class I-MHC and eventually Class II.

Results: The astonishing results included an increase in C-Peptide and Insulin both fasting and 2hpp in almost 83% of the patients, a reduction of Anti-Gad and Anti-Islet Cell AB in up to 67% of the patients and a reduction of HbA1c in almost 93%. About 100 % could reduce their Insulin consumption, about 77% could discontinue their short acting Insulin and about 31% of the patients could discontinue their insulin injections completely.

Conclusion: To our knowledge, this is the first available and most successful treatment in early type I DM that uses ordinary medications with the greatest efficacy and denotes that multiple low risk specific immune modulatory drugs might be more effective than highly immune suppressing ones. With a follow up of up to three years we have not encountered major relapses though minor fluctuations in one or some of the markers were seen and were responded by us with aggressive reactions. The reduced need for insulin injection with cessation of short acting insulin injections in a high percentage of patients is another major advantage of this approach.

Disclosure: Z. Talaeerad: None.

92

Using a somatostatin receptor-2 antagonist to restore hypoglycaemia-induced glucagon secretion in type 1 diabetes

T.G. Hill ¹, R. Gao¹, J.G. Knudsen², N.J.G. Rorsman¹, Q. Zhang¹, A. Clark¹, P. Rorsman¹;

¹Oxford Centre for Diabetes, Endocrinology, and Metabolism, University of Oxford, Oxford, UK, ²Department of Biology, University of Copenhagen, Copenhagen, Denmark.

Background and aims: In type 1 diabetes (T1D), counterregulatory alpha-cell glucagon secretion becomes insufficient, contributing to potentially fatal hypoglycaemia. In streptozotocin-treated diabetic rats, somatostatin receptor-2 (SSTR2) antagonists have been proposed to restore counterregulatory glucagon secretion. Here, we investigated the ability of the SSTR2 antagonist, CYN154806, to correct glucagon release during hypoglycaemia using the spontaneous, autoimmune non-obese diabetic (NOD) mouse model of T1D. We also examined the presence of SSTR2 expression in islets from T1D and normoglycaemic donors.

Materials and methods: Islet hormone secretion was assessed in adult female diabetic (blood glucose ≥ 30.0 mmol/l) and normoglycaemic pre-diabetic control (blood glucose 5.4-8.3 mmol/l) NOD/ShiLtJ mice by in situ pancreas perfusion. Changes in alpha- and delta-cell SSTR2 expression were examined with quantitative immunocytochemistry in NOD diabetic compared with pre-diabetic NOD control, as well as in human T1D compared with normoglycaemic (ND) subjects.

Results: Both diabetic and pre-diabetic NOD islets contained immune cell infiltration, with NOD and human diabetic islets exhibiting a marked reduction in insulin-positive area/islet. Whereas hypoglycaemia normally increased glucagon secretion from the perfused control NOD pancreas, this response was abolished in the diabetic pancreas (1 vs 10 mmol/l glucose; control, 20.9±1.65 vs 8.97±0.45; diabetic, 6.18±0.63 vs 6.42±0.58 pg/min, n=8). Somatostatin secretion at 1 mmol/l glucose was abnormally elevated in the perfused pancreas of diabetic NOD mice (control vs diabetic, 2.13±0.09 vs 3.88±0.35 fmol/min, p<0.001, n=8). Application of the SSTR2 antagonist, CYN154806 (200 nmol/l), restored hypoglycaemia-induced glucagon secretion from the diabetic NOD pancreas (1 mmol/l glucose + CYN154806 vs 1 mmol/l glucose, 33.3±3.77 vs 6.18±0.63 pg/min, p<0.001, n=8). Alpha-cells in T1D displayed significantly higher co-localisation of glucagon with SSTR2 when compared with islets from ND (Pearson correlation coefficient, 0.596 vs 0.353, p < 0.0001, n = 2 T1D and ND), suggesting an increase in alpha-cell SSTR2 expression.

Conclusion: Alpha-cell hypersensitivity to somatostatin contributes to the inhibition of glucagon secretion in T1D hypoglycaemia which can be alleviated by an SSTR2 antagonist. These data support the use of SSTR2 antagonists as a promising adjunct therapy to insulin for reducing hypoglycaemic risk in T1D patients.

Supported by: Novo Nordisk-Oxford Fellowship; Leona M. and Harry B. Helmsley Charitable Trust; Medical Research Council

Disclosure: T.G. Hill: None.

93

Modulating the Hippo regulator YAP promotes beta cell regeneration in a model of type 1 diabetes

M. Madduri, S. Baskaran, M. Elawour, S. Rafizadeh, A. Ardestani, K. Maedler;

University of Bremen, Bremen, Germany.

Background and aims: Beta-cell regenerative therapy should replace what is missing in both type 1 and type 2 diabetes (T1D/T2D); highly functional beta-cells. The transcriptional regulator of the Hippo pathway Yes Associated Protein (YAP) controls proliferation but is epigenetically silenced and “disallowed” in mature pancreatic beta-cells. YAP re-expression in human beta-cells leads to a 3-fold induction of proliferation, which is accompanied by full preservation of insulin secretory function and beta-cell identity genes. Consistently, beta-cell-specific YAP-overexpressing mice showed dramatic induction in beta-cell proliferation, together with a remarkable beta-cell mass expansion. Here we identified beta-cell-specific YAP-induced reversal of diabetes in a mouse model of T1D. A small molecule activator of YAP downstream signals showed similar effects on beta-cell proliferation induction in human and mouse islets.

Materials and methods: Doxycycline inducible active YAP overexpressing mice (Teto-YAP^S127A) mice were cross-bred with mice carrying the tTA tetracycline transactivator under the control of the insulin promoter (RIP-rtTA) to achieve inducible beta-cell specific Rip-Ins2-TetO-hYAP1-S127A (beta-YAP) mice. In-vivo beta-cell destruction was induced by a single high-dose streptozotocin (STZ) injection and glucose homeostasis, beta-cell proliferation, apoptosis, and beta-cell specific gene expression were analyzed during the 4-week experiment.

Results: Two different experimental intervention strategies, namely YAP induction for 2 weeks one week after STZ injection and for 2 weeks with an interim pause, both showed a robust reduction of glucose levels throughout the experiment and improvement of glucose tolerance by YAP induction. In both experimental approaches, beta-cell area, -mass, -survival, and -proliferation were robustly improved. Pharmacological induction of YAP downstream signals significantly induced beta-cell proliferation in human islets. A screen for important functional genes for glucose-stimulated insulin secretion and key beta-cell transcription factors in human islets revealed consistent upregulation of most genes, including MafA and PDX1, even under the situation of highly increased proliferation in control and importantly in islets from a T2D donor. Also, the transcription factor forkhead box M1 (FOXM1), which is critical for postnatal beta-cell proliferation and mass expansion and identified as the YAP target gene was highly upregulated.

Conclusion: Our results show that transient re-expression of YAP promotes beta-cell regeneration and β-cell mass restoration. YAP has a strong pro-proliferative activity in human islets in vitro and in mice in vivo and might be a novel target for beta-cell regenerative therapy to prevent loss (rapid in T1D, progressive in T2D) of functional pancreatic beta-cell mass in diabetes.

Supported by: JDRF, DFG

Disclosure: M. Madduri: None.

94

Higher levels of autoantibodies directed against IFN-alpha in latent autoimmune diabetes in adults compared to type 1 diabetes

R. Amendolara, L. D'Onofrio, C. Castrucci, E. Maddaloni, S. Di Giulio, C. Mignogna, A. Siena, G. Giannini, R. Buzzetti;

Sapienza University of Rome, Rome, Italy.

Background and aims: Differently from type 1 diabetes (T1D), individuals with latent autoimmune diabetes in adults (LADA) show a slow decline in beta-cell reservoir over time. Immunological factors associated with the different natural history of LADA and T1D may unveil novel protective pathways for target therapies. IFN-alpha is recognized as a key trigger of deadly autoimmune attack against beta-cells during the early stages of T1D pathogenesis. We hypothesized that autoantibodies against IFN-alpha (Abs-IFN-alpha) might neutralize the detrimental effect of IFN-alpha towards beta-cells and therefore might be associated with the less aggressive natural history of autoimmune diabetes observed in LADA. In this study, we aimed to investigate the presence of Abs-IFN-alpha in sera of individuals with LADA compared to T1D.

Materials and methods: In this cross-sectional study, we enrolled 41 individuals with LADA and 41 individuals with T1D matched 1:1 for sex and BMI. The IFN-alpha neutralizing titer was tested with the aid of HEK-Blue^TM IFN-alpha/beta reporter cells that expresses alkaline phosphatase (AP) after stimulation with IFN-alpha2b. Clinical and biochemical data of people with LADA and T1D were retrieved from medical records.

Results: The comparison of the main clinical, anthropometric and biochemical features between individuals with LADA and T1D is summarized in Table 1. The frequency of Abs-IFN-alpha was 59% among participants with LADA, while 32% of individuals with T1D tested positive for these autoantibodies (p=0.015). Participants with LADA showed significantly higher levels of neutralizing Abs-IFN-alpha than individuals with T1D (1.5 [0-3.2] vs 0 [0-2.6] ng/mL, respectively; p=0.028).

Conclusion: Abs-IFN-alpha may be found in sera of subjects with autoimmune diabetes. Individuals with slowly progressive autoimmune diabetes are more prone to develop autoimmunity against IFN-alpha than individuals with T1D. Future studies should unveil whether the presence of Abs-IFN-alpha might be related to a slower beta-cell destruction in people with LADA.

Supported by: Italian Ministry of University and Research (PRIN 2017: 20175L9H7H)

Disclosure: R. Amendolara: None.

95

Beta cell autophagy in type 1 diabetes pathogenesis: real-time in vivo discoveries

O. Melnyk ¹, C. Muralidharan², J.J. Crowder¹, B.E. Duffett¹, M.M. Martinez³, A.K. Linnemann^1,2;

¹Departments of Pediatrics, Indiana University School of Medicine, Indianapolis, IN, USA, ²Departments of Biochemistry & Molecular Biology, Indiana University School of Medicine, Indianapolis, IN, USA, ³Departments of Medicine, Indiana University School of Medicine, Indianapolis, IN, USA.

Background and aims: Autophagy is a catabolic process used by all cells, including pancreatic β cells, to degrade old and damaged cellular components. Functional autophagy is critical for cell homeostasis and stress response. Disruptions in autophagy have been demonstrated in the context of multiple autoimmune diseases and were recently established by our group in the context of human type 1 diabetes (T1D). We observed evidence of dysfunctional autophagic components in the islets of autoantibody positive human donors as well as in pre-diabetic nonobese diabetic (NOD) mice, suggesting defective autophagy prior to the hyperglycemia onset. However, our prior studies were limited by evaluation of a single time point of autophagic degradation. Therefore, we developed a method to study autophagy in β cells in real-time in vivo and evaluated autophagic flux in the pre-diabetic stage.

Materials and methods: For in vivo imaging of mice, wild type C57Bl/6J mice (RRID:IMSR_JAX:000664), male and female NOD mice (RRID:IMSR_JAX:001976), and NSG mice (RRID:IMSR_JAX:005557) were intraperitoneally injected with a β cell selective dual fluorescent biosensor (AAV8.RIP1.mCherry.EGFP.LC3B). The biosensor marks β cell autophagosomes and its function is based on the different pH sensitivities of eGFP and mCherry proteins. Fusion of eGFP/mCherry labeled autophagosomes with acidic lysosomes during autophagic flux causes quenching of the eGFP signal, enabling real-time observation of autophagic flux in vivo. Three weeks post sensor injection, animals were anesthetized, and their pancreata externalized for intravital imaging using a LEICA SP8 DIVE multi-photon microscope. We injected mice with either saline or chloroquine (autophagy inhibitor) and observed the basal level of autophagic flux. Next, we injected animals with IFN-α, a cytokine associated with T1D pathogenesis. The collected images were analyzed, and the Pearson correlation coefficient of eGFP:mCherry punctae was calculated. In addition, the rate of the mCherry/colocalized puncta was evaluated over time to understand the autophagic flux change in response to IFN-α. For in vivo imaging of human islets, islets were infected with the autophagy biosensor in vitro and then transplanted under the kidney capsule of NSG mice. Two weeks later, intravital imaging was performed on the kidney containing islets using a similar approach as above.

Results: We observed rapid stimulation of β cell autophagic flux by IFN-α in the absence of chloroquine in the C57Bl/6J and NSG mice. These effects were completely blocked in the presence of chloroquine. Alternatively, IFN-α caused little to no response in the female and male prediabetic NOD mice, and chloroquine had no effect. Human islets had much more variable levels of baseline autophagy that was donor-dependent, and had a heterogeneous response to IFN-α stimulation.

Conclusion: We report rapid stimulation of pancreatic β cell autophagic flux in response to IFN-α exposure in vivo in mouse and human β cells. Further, we demonstrate defects in β cell autophagic flux before diabetes onset in a NOD mouse model. Future studies will determine both the mechanism and timing of the autophagic decline and its specific role in autoimmune diabetes pathogenesis.

Supported by: NIH P30DK097512, NIH R01DK124380

Disclosure: O. Melnyk: None.

96

Heterogeneity of islet cell physiology in the pathogenesis of type 1 diabetes

C.M. Cohrs, A.K. Mojica Avila, S. Speier;

Institute of Physiology, Dresden, Germany.

Background and aims: Type 1 diabetes is characterized by progressive immune cell mediated beta cell destruction, eventually leading to chronic hyperglycemia. Recent studies show that islet and beta cell mass is unequally distributed throughout the entire pancreas. Furthermore, islet immune cell infiltrations in T1D exhibit a rather patchy distribution throughout the pancreas during the pre-diabetic phase. However, it is currently still unclear if differences in islets and immune cell lesions lead to a preferential regional loss of beta cell mass and function in type 1 diabetes pathogenesis. We, therefore, aimed to investigate the regional heterogeneity of the functional beta cell mass in the head, body, and tail of the pancreas under physiological conditions as well as in correlation to immune cell infiltration during type 1 diabetes development.

Materials and methods: Live pancreas tissue slices were obtained from the different regions of the pancreas of control (NOD.Scid) and diabetes-prone NOD female mice aged 3, 6, 9, and 12 weeks (n=3-10 mice/group). Tissue slices were subjected to dynamic insulin secretion measurements upon various stimuli and were subsequently fixed for further analyses. Fixed slices were immunostained against insulin, glucagon, and CD3 for a detailed 3D morphometric analysis. Insulin staining was additionally used to normalize beta cell volume to insulin secretion assessed before. Finally, calcium imaging was employed to investigate differences in cellular responsiveness between beta cells from the two mouse strains.

Results: The assessment of insulin secretion exhibited age-related but no regional differences in glucose responsiveness for both NOD. Scid and NOD mice. Subsequent immunohistochemistry analysis showed no major changes in endocrine volume and islet composition, thus resulting in comparable normalized secretion patterns in NOD. Scid mice. While these observations also account for most of the prediabetic NOD mice until 12 weeks of age, we could observe an increasing number of single pancreatic regions in mice (up to 21% at 12 weeks) depicting unresponsiveness to glucose that may indicate beta cell dysfunction prior the development of overt diabetes. Interestingly, also in normoglycemic NOD mice, no overall differences in 3D morphology or degree of infiltration were observed between the three regions. Additionally, [Ca²⁺]_i imaging on tissue slices could show that the islets of NOD mice display - independent of the degree of infiltration - a lower and desynchronized proportion of glucose-responsive cells within the islets compared to NOD. Scid, further suggesting a beginning islet dysfunction. Recent onset hyperglycemic NOD females, however, exhibit an almost blunted insulin secretion and a near total loss of beta cell volume within the pancreas tissue slices. Surprisingly, the remaining beta cell volume showed a strikingly high secretory capacity which might hint at a compensatory mechanism to counteract the elevated glucose levels in vivo around the onset time.

Conclusion: In summary, this study provides a paralleled view of islet morphology, function, and immune cell infiltration in the different regions of the NOD mouse pancreas during type 1 diabetes progression. We show that a functional decline precedes a significant beta cell loss. Although no apparent differences suggest region-specific effects on functional beta cell mass, our data instead suggest a strong heterogeneity on the lobular level.

Disclosure: C.M. Cohrs: None.

OP 17 New approaches: a magical mystery tour

97

The effect of 14-day exogenous ketone supplementation on fasting metabolic outcomes in adults with type 2 diabetes: a randomised controlled trial

K. Falkenhain ¹, H. Neudorf¹, K. Madden², J. Walsh³, J. Little¹;

¹University of British Columbia Okanagan, Kelowna, BC, Canada, ²University of British Columbia Vancouver, Vancouver, BC, Canada, ³McMaster University, Hamilton, ON, Canada.

Background and aims: Emerging research has shown an acute glucose-lowering effect of exogenous ketones, suggesting therapeutic potential of these supplements in individuals with impaired glucose metabolism. However, it is unknown how repeated ingestion of exogenous ketones affect metabolic health in individuals with type 2 diabetes (T2D).

Materials and methods: We conducted a randomized, counter-balanced, double-blind, placebo-controlled crossover trial to test the effects of thrice daily pre-meal ingestion of an exogenous ketone monoester supplement drink in adults with T2D on fasted glycemic and metabolic health outcomes. Adults with T2D (N = 15, 40% female, mean [SD] age of 57 [9] years) completed two 14-day intervention periods consuming ketone monoester [(R)-3-hydroxybutyl (R)-3-hydroxybutyrate containing 15 g β-OHB] or calorie-free placebo 15 minutes prior to each meal for 14 days, separated by at least 14 days of washout. All meals were provided to participants with caloric intake matched between conditions. Data were analyzed using a linear mixed effects model with condition (ketone monoester supplement vs placebo), time point (pre- vs post-intervention), period (which condition the participant engaged in first), and baseline value included as fixed factors as well as a random intercept for each participant. The interaction effect estimate (time by condition) alongside corresponding 95% confidence interval is reported as the main analysis of interest.

Results: A single dose of the ketone monoester supplement (containing 15 g of β-OHB) provided during the screening visit elevated blood β-OHB to 1.9 mM ± 0.7 mM within 30 minutes. Adherence to the supplement regimen throughout the 14-day intervention periods was high with >93% of supplements consumed in both conditions. Self-reported gastrointestinal distress questionnaires assessing nausea, urge to vomit, bloating, belching, and cramping across the first 3 days of each intervention period showed greater incidence of symptoms following ingestion of the ketone supplement compared to placebo (P < 0.05). There were no differences between groups in change in glycemic outcomes of fasting glucose (0.6 mM; 95% CI: -0.6 to 1.9 mM; P = 0.31) or HbA1c (0.0 %; 95% CI: -0.2 to 0.2%; P = 0.97). Similarly, no change was observed in fasting β-OHB, non-esterified fatty acids, insulin, or blood lipids (all P > 0.1).

Conclusion: In the context of this free-living, controlled-feeding randomized controlled trial in adults with T2D, thrice daily pre-meal ingestion of a ketone monoester supplement did not improve glycemic control or affect other fasting metabolic outcomes. Future studies aiming to explore the mechanistic basis for the (lack of) glucose-lowering effect of exogenous ketone monoester ingestion in different populations are warranted.

Clinical Trial Registration Number: NCT05155410

Supported by: Canadian Institutes of Health Research (CIHR) project grant (PJT-169116), Michael Smith Foundation for Health Research (MSFHR) Scholar Award (16890)

Disclosure: K. Falkenhain: None.

98

ENT-03 rapidly decreases glucose prior to weight loss and activates brain regions involved in satiety and appetite regulation: comparison with semaglutide

D. Barbut ¹, J. Hecksher-Sorensen², J. Baur³, P. Titchenell³, J. Davis³, A. Fleming⁴, M.B. Zemel⁴, A. Moses⁵, M.A. Zasloff¹;

¹Enterin, Inc, Philadelphia, PA, USA, ²Gubra, Inc, Copenhagen, Denmark, ³Medicine, University of Pennsylvania, Philadelphia, PA, USA, ⁴Kinexum, Inc., Harper's Ferry, WV, USA, ⁵Memor consulting, Inc., Boston, MA, USA.

Background and aims: We discovered ENT-03 in newborn mouse brain, with potent glucose and weight lowering potential. This spermine-bile acid appears postnatally at a time when the maternal supply of milk is inadequate to support maximal growth, requiring mobilization of endogenous stores of energy. ENT-03 has potent PTP1B inhibitory activity and acts centrally to influence glucose metabolism and weight. In mice, ENT-03 phenocopies the PTP1B neuronal KO, reducing food intake, adiposity, insulin resistance, and hyperglycemia. The current study explores the impact and potential mechanisms of ENT-03 on glycemia and weight in Diet Induced Obesity (DIO) mice.

Materials and methods: ENT-03 (25mg/kg; n=13), semaglutide (0.04mg/kg or 0.12mg/kg; n=5 and n=8 respectively) or vehicle (n=13) were administered s.c. x2 wkly for 10 wks to DIO mice fed a high fat diet. Weight, glucose and metabolic parameters were monitored. In a separate experiment, C57bl/6jR mice were administered a single dose of vehicle, ENT-03S or semaglutide and were euthanized 2 hrs later. Brains were stained for cFos and imaged at single cell resolution using light sheet microscopy. cFos data from individual brains were mapped onto an average mouse brain atlas and the number of cFos labelled cells was quantified in 286 regions.

Results: In ENT-03-treated DIO mice, non-fasting glucose rapidly fell into the range seen in nonobese animals within 2 wks and prior to significant weight loss. Weight fell substantially thereafter. In contrast, in semaglutide-treated mice, glucose fell in proportion to weight loss. ENT-03 caused C-Fos activation (p<0.001) compared to vehicle in 40 regions within the hypothalamus and brainstem involved in appetite suppression, food-entrained circadian rhythm, autonomic function and growth. ENT-03 also activated brainstem nuclei which enable autonomic responses. Semaglutide stimulated 12 regions (p<0.001) over vehicle of which 10 were also significantly activated by ENT-03. ENT-03, but not semaglutide, stimulated the hypothalamic centers involved in energy homeostasis, including the POMC region of the arcuate nucleus, the lateral preoptic area, the paraventricular nucleus, the tuberal nucleus, and the dorsomedial nucleus (Figure 1).

Conclusion: ENT-03 is a novel, endogenous, centrally acting steroid with PTP1B inhibitory activity which rapidly normalizes glucose independent of body weight by acting on hypothalamic and brainstem circuits that regulate energy and metabolism in mice. The neural circuits affected by ENT-03 appear to differ from those impacted by semaglutide suggesting a different mechanism of action. Phase 1 studies in subjects with obesity and diabetes are underway.

Disclosure: D. Barbut: None.

99

New pharmacological approaches to improve insulin secretion in type 2 diabetes

B. Merino ¹, A. Sanz-González¹, I. Cózar-Castellano^1,2, C. Broca³, J. Sabatier³, G. Acosta^4,5, M. Royo^4,5, C. Hernándo-Muñoz⁶, T. Torroba⁶, G. Perdomo¹;

¹Institute of biomedicine and molecular genetics (IBGM), University of Valladolid - CSIC, Valladolid, Spain, ²Biomedical Research Networking Center in Diabetes, Madrid, Spain, ³IRMB Hopital St Eloi, Montpellier, France, ⁴Biomedical Research Network Center in Bioengineering, Biomaterials and Nanomedicine, Barcelona, Spain, ⁵Department of Surfactants and Nanobiotechnology, Institute for Advanced Chemistry of Catalonia, Barcelona, Spain, ⁶Department of Chemistry, University of Burgos, Burgos, Spain.

Background and aims: Type 2 diabetes (T2D) is a complex polygenic disease that involved chronic low-grade inflammation and immunological imbalance. Preimplantation factor (PIF) is a peptide expressed in the embryo and its placenta, and its synthetic form sPIF is a safe FDA-approved drug for use in humans, able to mimic the functions of natural PIF and has immunomodulatory and anti-inflammatory properties. PIF targets insulin-degrading enzyme (IDE), a protease with a high affinity for insulin, albeit degrades other small peptides. IDE is linked to increased susceptibility to develop T2D and its pancreatic and hepatic expression levels are altered in people with T2D. We demonstrate that IDE is necessary for the regulation of insulin secretion in β-cells. The aim of this work is to demonstrate that sPIF-mediated pharmacological activation of IDE improves glucose-stimulated insulin secretion (GSIS) and glucose tolerance in preclinical models of T2D.

Materials and methods: We used a preclinical murine model of T2D induced by high-fat diet feeding (HFD) for 12 weeks. This model was treated with saline or sPIF (1mg/kg/day) for 25 days using subcutaneous minipumps. Glucose tolerance, circulating insulin and C-peptide levels, and glucose-stimulated insulin secretion were assessed. In addition, the Min-6 cell line was used in order to test the effect of sPIF in vitro on GSIS and IDE activity. The effect on GSIS was also tested in human islets.

Results: Obese mice treated with sPIF showed improved glucose tolerance and a significant increase in circulating insulin and C-peptide levels, with no changes in hepatic insulin clearance. Pancreatic islets from sPIF-treated obese animals recovered the physiological ability to secrete insulin in response to high glucose. In addition, MIN-6 cells showed increased GSIS in parallel to a significant increase in IDE activity, suggesting that the proteolytic activity is necessary for the effects on insulin secretion. Increased GSIS was also observed in sPIF-treated human islets.

Conclusion: sPIF is able to recover the glucose-stimulated insulin secretion capacity of β-cells in a preclinical model of T2D and human islets. This effect is in turn related to an increase in IDE activity. Therefore, sPIF is a potential secretagogue for its use in people with T2D.

Supported by: EFSD-Novo Nordisk Rising Star, SED, Ministerio Ciencia y Universidades, La Caixa, Generalitat de Catalunya, CIBER-BBN, IBGM, Junta CyL

Disclosure: B. Merino: None.

100

Pancreatic gene therapy durably improves glycaemia and delays disease progression in a murine model of type 2 diabetes

H. Rajagopalan, C. Lubaczeuski, E. Cozzi, N. Picard, J. Wainer, R. Reese, J. Caplan, J. West, A.P. Liou;

Fractyl Health, Inc., Lexington, MA, USA.

Background and aims: Beta cell (β-cell) failure is a key pathophysiological driver of type 2 diabetes (T2D) progression. Preservation of β-cell function is essential to maintaining glycaemic control and preventing T2D complications. Glucagon-like peptide-1 receptor agonists (GLP1RAs) have demonstrated clinical efficacy and are thought to improve β-cell function. However, the requirement for chronic administration and systemic side effects limit their long-term effectiveness. Localized pancreatic GLP1RA production may enable durable islet-cell benefit and glycaemic control while diminishing side effects. We developed a novel adeno-associated virus (AAV) gene therapy delivery platform to enable local and durable production of therapeutic proteins by the pancreas. In this study, we investigated the effects of an optimized β-cell-restricted GLP1RA gene therapy on glucose metabolism and disease progression in the db/db mouse compared to semaglutide (Sema), a well-established GLP1RA.

Materials and methods: To establish a baseline response to chronically administered systemic GLP1RA in the db/db model, Sema (10 nmol/kg) or vehicle were given daily to 5-week-old mice for 10 weeks (n=8/group). We utilized β-cell and islet-based, in vitro, screening assays to identify GLP1RA transgene candidates optimized for β-cell-driven GLP1RA production, protein secretion, and glucose-stimulated insulin secretion. To evaluate the impact of a single-dose GLP1RA gene therapy, 1) optimized β-cell-restricted AAV-GLP1RA, 2) AAV control, or 3) vehicle were given to 5-week-old db/db mice (n=8/group). Fasting blood glucose (FBG), fasting insulin (FI), and glucose tolerance via an intraperitoneal glucose tolerance test (IPGTT) were evaluated to determine treatment effect. To confirm gene therapy islet restriction, GLP1RA pancreatic and serum-protein expression were determined by immunohistochemistry and liquid chromatography with tandem mass spectrometry.

Results: Sema-treated mice had a 64% decrease in FBG (p<0.0001) and a 3.5-fold increase in FI (p=0.01) across the time course. Glucose tolerance was significantly improved during IPGTT at weeks 3, 6, and 9 (p<0.001). Remarkably, AAV-GLP1RA treatment decreased FBG levels by 70% (p<0.0001) and elevated FI by 1.9 fold (p<0.01) for 10 weeks compared to controls. Glucose tolerance was improved during IPGTT at weeks 3, 5, 7, and 9 (p<0.01). AAV-GLP1RA protein was undetectable in serum and restricted to the islet, suggesting improved glycaemic control was due to intra-islet GLP1RA production and signaling.

Conclusion: In this study, we establish that a single-dose, β-cell-restricted, GLP1RA gene therapy can induce similar improvements in glycaemia and delay disease progression compared to chronic systemic Sema administration in a murine model of T2D. Taken together, these data provide proof of concept that an islet-restricted pancreatic gene therapy has the potential to durably treat metabolic diseases, including T2D.

Disclosure: H. Rajagopalan: Employment/Consultancy; Fractyl Health, Inc. Stock/Shareholding; Fractyl Health, Inc.

101

Electroporation therapy (ReCET) of the duodenum combined with GLP-1RA to replace insulin in type 2 diabetes patients: 12 months results from the first-in-human study

C.B.E. Busch ¹, S. Meiring², A.C.G. van Baar², F. Holleman³, M. Nieuwdorp³, J.J.G. Bergman²;

¹Amsterdam UMC, Amsterdam, Netherlands, ²Gastroenterology and Hepatology, Amsterdam UMC, Amsterdam, Netherlands, ³Internal Medicine, Amsterdam UMC, Amsterdam, Netherlands.

Background and aims: Exogenous insulin is the final treatment option for controlling hyperglycemia in patients with type 2 diabetes (T2D) but contributes to weight gain and deterioration of metabolic health. Studies have shown that duodenal mucosal ablation (DMR) results in better glycemic control by improving insulin resistance, the root cause of T2D and the metabolic syndrome. Re-Cellularization via Electroporation Therapy (ReCET), is a novel endoscopic procedure that uses electroporation to elicit cell apoptosis with subsequent re-cellularization in the duodenal mucosa. In this first-in-human study we aimed to eliminate insulin treatment in patients with T2D using a single ReCET procedure combined with a GLP-1 receptor agonist (GLP-1RA). Safety, feasibility, and efficacy of ReCET combined with GLP-1RA were assessed.

Materials and methods: First-in-human study in 14 patients with T2D, 28-75 years old, body mass index (BMI) 24-40 kg/m², with a glycosylated hemoglobin (HbA1c) ≤ 64 mmol/mol, basal insulin dose < 1 U/kg/day, c-peptide ≥ 0.2 nmol/l. All patients underwent ReCET under deep sedation, followed by a 2 week isocaloric liquid diet. Thereafter semaglutide (GLP-1RA) was introduced and titrated up to 1 mg/week. Primary feasibility endpoints (procedure time [time from catheter-in to catheter-out], technical success rate, % of patients tolerating GLP-1RA) and safety endpoints ((S)AEs including hypoglycemic events) were assessed. Primary efficacy endpoint was the number of subjects off insulin at 6 months with HbA1c ≤ 58 mmol/mol. At baseline, 6 and 12 months follow-up glycemic and metabolic data and treatment satisfaction scores were collected and assessed with the Wilcoxon paired signed-rank test. Data are presented as median with interquartile range (IQR).

Results: Median age was 62 (IQR 54 - 67) years, 57% was male and median units of daily insulin were 27 (IQR 22 - 33). ReCET showed a technical success rate of 100% with a median axial treatment length of 12 cm (IQR 11 - 14) centimeter. Procedure time was 58 (IQR 49-73) minutes. Maximum semaglutide dosage was tolerated by 13 (93%) patients. No device related SAEs were observed. One patient experienced a hypoglycemic event without need for third party assistance. Up until 12 months, 12 (86%) patients were off insulin, whilst showing significant improvements in glycemic control (HbA1c, fasting plasma glucose, time in range of glucose values), metabolic parameters and improved treatment satisfaction (Table 1).

Conclusion: These results suggest that ReCET with the Endogenex system is safe and feasible. ReCET in combination with semaglutide is a promising new therapeutic option that may effectively eliminate insulin therapy in selected T2D patients, while improving glycemic control and overall metabolic health.

Clinical Trial Registration Number: NL9482

Supported by: Unrestricted research grant from Endogenex.

Disclosure: C.B.E. Busch: None.

102

Study of MultI-eLectrode EndovaScular denervaTiOn in patieNts with type 2 diabetes mEllitus (MILESTONE): 6-month analysis from the first-in-human proof-of-concept study

T. Pan, Z. Wang, G.-J. Teng;

Southeast University, Nanjing, China.

Background and aims: To assess safety and glycemic indices at 6 months in a first-in-human proof-of-concept study of a minimally invasive, catheter-based endovascular denervation (EDN) procedure, in patients with type 2 diabetes mellitus (T2DM).

Materials and methods: Using a novel six-electrode catheter system, EDN was conducted on the celiac artery (CA), and aorta between CA and the superior mesenteric artery (SMA) in T2DM patients at a single medical center. The major inclusion criteria are T2DM patients diagnosed over 5 years, with glycosylated hemoglobin (HbA1c) over 7.5%. The primary outcomes were evaluated by the safety, HbA1c, homeostasis-model assessment of insulin resistance (HOMA-IR), fasting plasma glucose (FPG) and 2-hour postprandial plasma glucose (2hPG), oral glucose tolerance test (OGTT) based 3-hour C-peptide test at interval of 3, 6, 12, 18 and 24 months, respectively. The antidiabetic medication, lifestyle, physical condition, blood pressure (BP), plasma norepinephrine, angiotensin II, liver biochemistry, and plasma lipids were also recorded.

Results: A total of 11 T2DM patients were included for analysis. The technical success was 100% and no severe treatment-related adverse events or major complications were observed. Both HbA1c and HOMA-IR significantly reduced at 6 months, from 9.9±1.6 to 8.0±2.4% (P = 0.005), and from 13.3 (IQR 5.9-46.1) to 6.0 (IQR 3.1-11.9) (P = 0.016), respectively. At 1, 3, and 6 months of follow-ups, FPG levels were 10.2±0.8, 10.1±1.2, and 9.6±1.7 mmol/L, respectively, compared to 15.4±1.6 mmol/L at baseline (P = 0.001, < 0.001, and < 0.001, respectively), and 2hPG levels were 11.8±3.0, 11.4±4.0, and 11.0±5.9 mmol/L, respectively, compared to 17.9±6.0 mmol/L at baseline (P = 0.001, 0.001, and 0.002, respectively). OGTT based 3-hour C-peptide release test showed improved beta-cell function (AUC 0.23 (IQR 0.18-0.32) vs. 0.28 (IQR 0.21-0.38) pmol/mL, P = 0.046). A reduction of daily insulin injection was also observed (24 (IQR 15.5-47) vs. 19 (IQR 9-27.5) IU, P = 0.018) without changes in lifestyle. Improvements of liver function were observed although physical conditions, BP, plasma norepinephrine, angiotensin II, and blood lipids were not changed during follow-ups.

Conclusion: The 6-month analysis from this trial shows that EDN using the novel six-electrode catheter system at the new sites of CA and aorta between CA and SMA elicits a clinically significant improvement in hyperglycemia in patients with T2DM, with good tolerability. More patients registered with a long-term follow-up, and the insight mechanism of EDN are warranted.

Clinical Trial Registration Number: NCT 04086043

Disclosure: T. Pan: None.

OP 18 Highlights in type 1 diabetes

103

Incidence trends of type 1 and type 2 diabetes in children and adolescents in Germany, 2002-2020

A. Stahl-Pehe ^1,2, C. Bächle^1,2, S. Lanzinger^3,2, C. Kamrath⁴, R.W. Holl^5,2, J. Rosenbauer^1,2;

¹Institute for Biometrics und Epidemiology, German Diabetes Center (DDZ), Düsseldorf, Germany, ²German Center for Diabetes Research (DZD), Munich-Neuherberg, Germany, ³Institute of Epidemiology and Medical Biometry, ZIBMT, University of Ulm, Ulm, Germany, ⁴Center of Child and Adolescent Medicine, Justus Liebig University, Giessen, Germany, ⁵Institute of Epidemiology and Medical Biometry, ZIBMT, University of Ulm, Düsseldorf, Germany.

Background and aims: Estimates of diabetes incidence in children and adolescents are an important basis for allocating health resources and defining prevention measures. The aim of the study was to estimate incidence trends of type 1 diabetes (T1D) and type 2 diabetes (T2D) in children and adolescents under the age of 20 years for the period 2002 to 2020 in Germany.

Materials and methods: The diabetes incidence register of North Rhine-Westphalia (NRW), the most populous German federal state (22% of the total population of Germany) records newly diagnosed children and adolescents under the age of 20 years via three data sources: the prospective hospital-based active surveillance system ESPED, annual practice surveys and the nationwide Diabetes Prospective Follow-up Registry (Diabetes-Patienten-Verlaufsdokumentation, DPV). Completeness of ascertainment was estimated using the capture-recapture-method. Point and interval estimates (95% CIs) of incidence (per 100,000 person-years) were based on the person-years method and age- and/or sex-standardised by the direct method using equal weights. Poisson regression analyses accounting for overdispersion were used to estimate temporal trends.

Results: Between 2002 and 2020, 14,980 (8,165 boys) and 670 (284 boys) children and adolescents were newly diagnosed with T1D and T2D, respectively. Ascertainment was 99.0% (T1D) and 92.5% (T2D) complete. Corrected for ascertainment, the incidence of T1D and T2D was 22.9 (22.6; 23.3) and 2.0 (1.9; 2.2), respectively. In boys, compared with girls, T1D incidence was higher (24.3 [23.8; 24.8] vs. 21.5 [21.0; 22.1]), but T2D incidence was lower (1.7 [1.5; 1.8] vs. 2.4 [2.2; 2.6]). The T1D incidences for the age groups 0-4, 5-9, 10-14 and 15-19 years were 17.8 (17.1; 18.4), 30.6 (29.8; 31.5), 31.4 (30.6; 32.3) and 11.9 (11.4; 12.4), respectively. The T2D incidences for the age groups 10-14 and 15-19 years were 1.8 (1.6; 2.0) and 2.3 (2.1; 2.5), respectively. The incidence of T1D and T2D increased annually by 2.0% (1.5%; 2.4%) and 4.9% (3.2%; 6.6%), respectively, with only small differences between boys and girls (T1D: 2.1% [1.6%; 2.7%] vs. 1.8%, [1.1%; 2.4%] and T2D: 5.1% [2.5%; 7.9%] vs. 4.7% [2.4%; 7.0%]). T1D incidence has plateaued in the youngest age group (0.5% [-0.5%; 1.4%] annual increase), but increased in the age groups 5-9, 10-14 and 15-19 years by 2.2% (1.5%; 2.9%), 2.7% (2.0%; 3.4%) and 1.6% (0.5%; 2.6%) annually, respectively. T2D incidence increased annually by 6.7% (4.0%; 9.4%) among 10- to 14-year-olds and by 3.5% (1.3%; 5.8%) among 15- to 19-year-olds.

Conclusion: The incidence of T1D in children and adolescents in Germany is significantly higher than the incidence of T2D. However, the incidence of T2D has increased more than the incidence of T1D over the last two decades.

Disclosure: A. Stahl-Pehe: None.

104

Secular trends in the occurence of islet autoimmunity and type 1 diabetes by the age of 5 years in genetically susceptible children: a birth cohort analysis

A. Kyrönniemi ¹, T. Pokka¹, T. Valtanen¹, J. Lempainen^2,3, H. Hyöty^4,5, J. Ilonen³, J. Toppari^2,3, J. Kero^2,3, M. Knip^5,6, R. Veijola¹;

¹Department of Pediatrics, Research Unit of Clinical Medicine, Medical Research Center, University of Oulu and Oulu University Hospital, Oulu, Finland, ²Department of Pediatrics, Turku University Hospital, Turku, Finland, ³Institute of Biomedicine, University of Turku, Turku, Finland, ⁴Department of Virology, Tampere University, Tampere, Finland, ⁵Tampere Centre for Child Health Research, Tampere University Hospital, Tampere, Finland, ⁶Pediatric Research Center, Children’s Hospital, University of Helsinki, Helsinki, Finland.

Background and aims: The incidence of type 1 diabetes increased in Finland until mid-2000´s. The secular trend in the occurence of islet autoimmunity or type 1 diabetes in genetically susceptible children has not been studied before. Our aim was to determine whether there were differences between various birth cohorts in the occurrence of islet autoimmunity or type 1 diabetes by the age of 5 years in Finnish children with HLA-conferred increased risk.

Materials and methods: The Type 1 Diabetes Prediction and Prevention (DIPP) study is a prospective follow-up study of Finnish children with HLA class II risk genotype for type 1 diabetes. HLA inclusion criteria were tightened in 2010. Islet autoantibodies to insulin (IAA), glutamic acid decarboxylase (GADA) and insulinoma associated antigen-2 were analysed regularly. Zinc transporter 8 autoantibodies were analysed from multipositive children until Apr 2019, and from every follow-up sample thereafter. Positivity for the same islet autoantibody in two consecutive samples was considered confirmed. The presence of linear trend in the incidences of T1D and autoantibodies by the age of 5 years with respect to increase in birth year in four birth cohorts (Nov 1994-Jul 1997, Jan 2003-Aug 2004, Mar 2010-Dec 2015, and Jan 2016-Nov 2017) was tested by Linear-by-Linear test (IBM SPSS Statistics for Windows, version 29. Armonk, NY: IBM Corp).

Results: In the four birth cohorts, 5631 children were followed until the age of 5 years or until diagnosis of type 1 diabetes before 5 years of age. The frequency of confirmed positivity for single or multiple autoantibodies, the appearance of GADA as the first autoantibody, and the occurrence of type 1 diabetes increased towards the two most recent birth cohorts, and were highest in the most recent birth cohort. The trend for IAA as the first appearing autoantibody was not statistically significant. (Table)

Conclusion: In the four birth cohorts from the DIPP study an increasing secular trend in the occurrence of both islet autoimmunity and type 1 diabetes by the age of 5 years was observed. The differences between the first two birth cohorts and the third one could be due to tightened HLA inclusion criteria in 2010. However, for the two most recent birth cohorts, similar HLA inclusion criteria were applied, and the observed trend likely reflects a true increase in islet autoimmunity and type 1 diabetes.

Clinical Trial Registration Number: NCT03269084

Supported by: JDRF, Academy of Finland, Finnish Foundations and University Hospitals.

Disclosure: A. Kyrönniemi: None.

105

Residual insulin after anti-viral treatment with pleconaril and ribavirin in new onset type 1 diabetes: a randomised, placebo-controlled, double blind, clinical trial

I. Mynarek ¹, L. Krogvold¹, E. Ponzi², F.C.B. Mörch^3,4, T.W. Hessel^3,4, T. Roald¹, N. Lindblom⁵, J. Westman⁵, P. Barker⁶, J. Ludvigsson⁷, K.F. Hanssen¹, H. Hyöty^8,9, J. Johannesen^3,4, K. Dahl-Jørgensen^1,10;

¹Oslo University Hospital, Oslo, Norway, ²Clinical Trial Unit, Oslo University Hospital, Oslo, Norway, ³Department of Clinical Medicine, Steno Diabetes Center, Copenhagen, Denmark, ⁴Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark, ⁵Apodemus AB, Stockholm, Sweden, ⁶NIHR Cambridge Biomedical Research Centre, Cambridge, UK, ⁷Linköping university, Linköping, Sweden, ⁸Faculty of Medicine and Health Technology, Tampere University, Tampere, Finland, ⁹Fimlab laboratories, Tampere, Finland, ¹⁰University of Oslo, Oslo, Norway.

Background and aims: Previous studies have shown a clinically significant association between enterovirus infection and onset of clinical diabetes and a low-grade enterovirus infection has been demonstrated in the pancreatic islets of patient with newly diagnosed type 1 diabetes. The aim of this study was to determine the effect of antiviral treatment with the combination of pleconaril and ribavirin on beta cell function in children and adolescents from onset of type 1 diabetes.

Materials and methods: In this phase-II, placebo-controlled, double-blind, parallel-group trial, 96 children (6-15 years) were randomly assigned to receive oral antiviral treatment (pleconaril and ribavirin) (n=47, 19 females) or placebo (n=49, 21 females) for 6 months, started less than 3 weeks after diagnosis of type 1 diabetes (baseline). Primary endpoint was endogenous insulin production at 12 months, measured by area under the concentration-time curve (AUC) for C-peptide level in response to a 2-hour mixed-meal tolerance test (MMTT) in the full analysis set population. A mixed model for repeated measures was used, and the treatment effect was estimated as the average marginal effect at 12 months (AME). Secondary endpoints included preserved peak C-peptide level > 0.2 pmol/mL during MMTT, insulin dosage, HbA1c, glycated albumin, and severe hypoglycemic events.

Results: At 12 months, mean 2-hour C-peptide AUC was significantly higher in the antiviral group than in the placebo group (AME 0.057, p=0.04, adjusted for baseline level). This equates to a 37% higher C-peptide level at 12 months in the antiviral group, see figure. 36/42 vs. 30/45 of the subjects, respectively, had maximal C-peptide > 0.2 pmol/mL (p=0.04) at 12 months. There were no significant differences regarding HbA1c, glycated albumin, insulin dosage, severe hypoglycemic events or adverse events at 12 months.

Conclusion: Among children with newly diagnosed type 1 diabetes, a 26-weeks course with two antiviral drugs partially preserved stimulated C-peptide secretion 12 months after diagnosis and a higher proportion of participants with clinically relevant preserved C-peptide secretion than placebo. Further studies are needed to optimize the anti-viral treatment.

Clinical Trial Registration Number: EudraCT number2015-003350-41; ClinicalTrial.gov number NCT04838145

Supported by: Health Region South East, Norway; JDRF, US; EU, INNODIA and INNODIA Harvest

Disclosure: I. Mynarek: None.

106

Time trends in the incidence of diabetic ketoacidosis among adults with type 1 diabetes: a nationwide Danish register study

E. Buur Stougaard ¹, H. Amadid², E. Søndergaard³, B. Carstensen², M. Eika Jørgensen⁴, K. Nørgaard⁵, P. Rossing¹, F. Persson¹, D. Vistisen²;

¹Complication Research, Steno Diabetes Center Copenhagen, Herlev, Denmark, ²Clinical Epidemiology, Steno Diabetes Center Copenhagen, Herlev, Denmark, ³Steno Diabetes Center Aahus, Aarhus, Denmark, ⁴Steno Diabetes Center Greenland, Nuuk, Greenland, ⁵Diabetes Technology Research, Steno Diabetes Center Copenhagen, Herlev, Denmark.

Background and aims: Diabetic ketoacidosis (DKA) is a life-threatening but preventable complication of diabetes that occurs most frequently in persons with type 1 diabetes. We aimed to quantify the incidence of DKA according to age and describe the time trend of DKA among adults with type 1 diabetes in Denmark.

Materials and methods: Individuals aged ≥18 years with type 1 diabetes were identified from a nationwide Danish Diabetes register. Hospital admissions due to DKA were ascertained from the National Patient Register. Follow-up period was from 1996 to 2020.

Results: The cohort consisted of 24.718 adults with type 1 diabetes of which 14.457 (58%) were men. The incidence rate of DKA per 100 person years (PY) decreased with increasing age for both men and women. From 20 to 80 years of age the DKA incidence rate decreased from 3.27 to 0.38 per 100 PY. From 1996-2008 the incidence rate of DKA increased for all age groups with a subsequent minor decrease in incidence rate until 2020. From 1996-2008 the incidence rates increased from 1.91 to 3.77 per 100 PY for a 20-year old man and from 0.22 to 0.44 per 100 PY for an 80-year old man. From 2008-2020 the incidence rates decreased from 3.77 to 3.27 per 100 PY for a 20-year old man and from 0.44 to 0.38 per 100 PY for an 80-year old man.

Conclusion: The incidence rates of DKA are declining according to age and with an overall decline from 2008 for both men and women. This likely reflects improved diabetes management for persons with type 1 diabetes in Denmark.

Disclosure: E. Buur Stougaard: None.

107

Enhanced immune response to cow’s milk is associated with the development of islet autoimmunity and type 1 diabetes: the TRIGR birth cohort study

S. Niinistö ¹, D. Cuthbertson², M.E. Miettinen¹, L. Hakola³, A. Nucci⁴, T. Korhonen¹, H. Hyöty³, J.P. Krischer², O. Vaarala⁵, M. Knip⁵, E. Savilahti⁶, S.M. Virtanen¹, TRIGR Investigators;

¹Finnish Institute for Health and Welfare, Helsinki, Finland, ²University of South Florida, Tampa, FL, USA, ³Tampere University, Tampere, Finland, ⁴Georgia State University, Atlanta, GA, USA, ⁵University of Helsinki, Helsinki, Finland, ⁶Helsinki University Hospital, Helsinki, Finland.

Background and aims: Prospective cohort study findings and their meta-analysis consistently suggest that the high amount of cow’s milk consumed during childhood is a risk factor of islet autoimmunity and type 1 diabetes. Further, increased humoral immune responses to cow’s milk proteins have been associated with increased risk of developing type 1 diabetes in one small cohort study and several case-control studies. Our aim was to study whether humoral immune responses to cow’s milk are associated with type 1 diabetes -related islet autoimmunity or type 1 diabetes in a large, international randomized infant feeding trial.

Materials and methods: The study was carried out in the Trial to Reduce IDDM in the Genetically at Risk (TRIGR) cohort comprising children born between 2002-2007 in 15 countries. The TRIGR study is double-blind randomized clinical trial in 2159 infants with increased genetic risk for type 1 diabetes and a first-degree relative with type 1 diabetes. The participants were randomized to an extensively hydrolyzed casein formula or a regular cow’s milk protein formula (control). Children were followed for the development of type 1 diabetes -related autoantibodies up to 10 years of age, and serum samples were collected at 3 to 12 months’ intervals. Antibodies to cow’s milk proteins were analyzed in 2069 children from samples taken from birth to 3 years of age. Milk consumption after 18 months up to 10 years of age was assessed biannually with a food frequency questionnaire. Associations between antibody levels against milk antigens and milk consumption with the risk of islet autoimmunity and type 1 diabetes were analyzed by Cox time-dependent survival model.

Results: Cow’s milk IgG antibody levels both in cord blood and longitudinally from birth to 3 years of age were associated with increased risk of developing islet autoimmunity [cord blood HR 1.30 (95% CI 1.05-1.61); longitudinal 1.39 (1.07-1.81)] and type 1 diabetes [cord blood 1.32 (1.02-1.71); longitudinal 1.43 (1.04-1.96)] during the 10-year follow-up period. Further, intake frequency of liquid milk after infancy was associated with increased risk of islet autoimmunity [1.18 (1.05-1.33)].

Conclusion: Increased levels of cow’s milk antibodies and high intake of liquid milk after infancy are associated with the development of islet autoimmunity and/or type 1 diabetes in the TRIGR study cohort supporting the view that aberrant mucosal immune response to bovine proteins is a feature linked to type 1 diabetes and related autoimmunity.

Clinical Trial Registration Number: NCT00179777

Supported by: NIH, NICHD, NIDDK, CIHR, JDRF, CEC, EFSD/JDRF/Lilly Programme, AKA, DDRF, Finnish Diabetes Research Foundation

Disclosure: S. Niinistö: None.

108

Spatio-temporal association between COVID-19 incidence and type 1 diabetes incidence in children and adolescents in Germany

J. Rosenbauer ^1,2, A. Stahl-Pehe^1,2, C. Baechle^1,2, S. Lanzinger^3,2, C. Kamrath⁴, O. Kuß^1,2, R.W. Holl^3,2;

¹Institute for Biometrics and Epidemiology, German Diabetes Center, Leibniz Center for Diabetes Research at Heinrich Heine University Duesseldorf, Duesseldorf, Germany, ²German Center for Diabetes Research (DZD), Munich-Neuherberg, Germany, ³Institute of Epidemiology and Medical Biometry, ZIBMT, University of Ulm, Ulm, Germany, ⁴Center of Child and Adolescent Medicine, Justus Liebig University Giessen, Giessen, Germany.

Background and aims: During the COVID-19 pandemic an increase in the incidence of type 1 diabetes (T1D) in children and adolescents has been observed in many regions worldwide. So far, underlying causes of this increase have not been elucidated. The aim was to assess the spatio-temporal association of the COVID-19 incidence with the T1D incidence in children and adolescents in Germany during the pandemic.

Materials and methods: For the period from March 2020 to June 2022, nationwide data of incident T1D in children and adolescents aged <20 years and COVID-19 in the total population were selected from the Diabetes Prospective Follow-up Registry (DPV) and the Robert Koch Institute (RKI), respectively. Data were aggregated across 400 districts and seven time periods defined according to pandemic waves. Area deprivation and urban-rural indicators at the district level were obtained from the RKI and official federal statistics. District level incidence rates were indirectly standardized (by age and sex for T1D) using national data as the reference to calculate standardized incidence ratios (SIRs). Spearman correlation analysis and a Bayesian spatio-temporal Poisson model accounting for spatial and temporal auto-correlation were applied to assess the association between COVID-19-SIR and T1D-SIR. The Poisson model adjusted for deprivation quintiles and an urban-rural indicator.

Results: During the study period, 9,596 children and adolescents newly diagnosed with T1D were registered, and the median number of T1D cases per district was 17.5 (IQR 10.0-29.0). The overall T1D incidence was 26.8 (95% CI 26.3-27.4) per 100,000 person-years (PY), and the median T1D incidence and T1D-SIR across all districts and time periods were 24.1 (IQR 10.8-36.8) per 100,000 PY and 0.90 (IQR 0.41-1.37), respectively. A total of 28.4 million persons with COVID-19 were documented, and the median number of persons with COVID-19 per district was 53,725 (IQR 36,566-81,042). The overall COVID-19 incidence was 146.19 (95% CI 146.13-146.24) per 1,000 PY, and the median COVID-19 incidence and COVID-19-SIR across all districts and time periods were 49.9 (IQR 6.0-198.8) per 1,000 PY and 0.34 (IQR 0.04-1.36), respectively. Correlation analysis across all spatio-temporal data points showed no significant correlation between levels of T1D-SIRs and COVID-19-SIRs (ρ=0.009 (95%CI -0.028-0.046). The Bayesian spatio-temporal Poisson model confirmed this finding, a doubling of the COVID-19-SIR was associated with a non-significant 1.007-fold (95%CI 0.994-1.020) increase in T1D-SIR.

Conclusion: The results showed no convincing evidence of an association between the COVID-19 incidence and the T1D incidence in children and adolescents in Germany. However, the fact that we used a coarse temporal resolution and did not consider time lags between the COVID-19 incidence and the T1D incidence may have masked possible associations.

Supported by: German Federal Ministry for Education and Research within the German Center for Diabetes Research (grant no. 82DZD02E3G)

Disclosure: J. Rosenbauer: None.

OP 19 Why does diabetes make one sad?

109

Circulating metabolites as mediators in the association between depression and type 2 diabetes

N. Wang, Y. Yu, Y. Sun, Y. Wang, B. Wang, Y. Lu;

No.639 Zhizaoju Road, Shanghai, China.

Background and aims: The association between depression and diabetes has been recognized for many years, but the nature of this relationship remains uncertain. Mechanisms underlying the role of circulating metabolites in depression leading to increased risk of type 2 diabetes mellitus (T2DM) are unclear. We aimed to discover specific circulating metabolites that are mediators of the association between depression and the risk of incident T2DM.

Materials and methods: A total of 95,990 participants free of T2DM at baseline with complete information available for depression and circulating metabolites from the UK Biobank cohort study were included in the present study. We used a mediation model to investigate the association of depression with T2DM mediated by circulating metabolites. Furthermore, we used two-step Mendelian randomization (MR) to identify and estimate the median effects of depression on T2DM through circulating metabolites using summarized genetic data.

Results: Differential correlation network analysis revealed that perturbations in circulating metabolite coregulation existed in participants with depression. In the observational analyses, 58 out of 168 circulating metabolites was associated with both depression and T2DM. In general, the associations of depression with circulating metabolites and of these circulating metabolites with risk of incident T2DM showed consistent patterns. In detail, total lipids in very large VLDL, phospholipids in very large VLDL, triglycerides in small HDL, cholesteryl esters, free cholesterol in chylomicrons and extremely large VLDL explained the highest median effects in subclasses of lipid lipoproteins, respectively. In addition, two-step MR showed evidence of causal association of depression with T2DM through 3 (triglycerides in small VLDL, triglycerides in small HDL, and average diameter for VLDL particles) out of 58 circulating metabolites, with a mediated proportion of 8.2-11.7% of the total effect.

Conclusion: Depression causes disturbances in circulating metabolites, which may lead to a higher risk of incident T2DM.

Disclosure: N. Wang: None.

110

The correlation between diabetes severity and depression

Y. Cho ¹, K. Jo¹, E. Kim¹, J. Han¹, K. Han², S.-D. Moon¹;

¹Incheon St. Mary's Hospital, Incheon, Republic of Korea, ²Department of Statistics and Actuarial Science, Soongsil University, Seoul, Republic of Korea.

Background and aims: Given the high prevalence of both diabetes mellitus (DM) and depression, it is important to determine which patients with DM are at high risk of developing depression. We investigated whether the risk of depression increased as diabetes severity increased.

Materials and methods: 2,067,017 adults with type 2 DM excluding those diagnosed with depression before and within 1 year after index date were identified from a nationwide population-based cohort in Korea. DM severity scores consisted of insulin use, DM duration ≥ 5 years, using 3 or more oral hypoglycemic agents, presence of chronic kidney disease, cardiovascular diseases (CVD), or diabetic retinopathy. Each of these characteristics was scored as one point of diabetes severity and their sum was defined as a diabetes severity score from 0-6.

Results: We identified 407,047 cases of major depression during a median follow-up of 6.2 years. The incidence rate of depression showed a significant linear association with an increase in DM severity score (Figure 1). After adjusting for age, sex, hypertension, dyslipidemia, smoking, alcohol consumption, regular exercise, and BMI, the HRs of depression were 1.16 (95% CI, 1.15 to 1.17) in subjects with one component of diabetes severity score (1 score), 1.29 (95% CI, 1.27 to 1.30) in those with 2 points , 1.46 (95% CI, 1.45 to 1.48) in those with 3 points, 1.72 (95% CI, 1.69 to 1.76) in those with 4 points, 1.95 (95% CI, 1.88 to 2.02) in those with 5 points, and 2.11 (95% CI, 1.88 to 2.38) in those with 6 points, respectively.

Conclusion: Diabetes severity was associated with an increased risk for developing major depression. Our findings provide evidence for the feasibility of focusing on depression screening in people with high diabetes severity scores.

Disclosure: Y. Cho: None.

111

Risk of mental illness in new-onset type 1 diabetes in adulthood: nationwide cohort study

S. Kim ¹, S. Cho², R. Oh², Y.-B. Lee², S.-M. Jin², G. Kim², J. Kim^2,1;

¹Department of Clinical Research Design and Evaluation, Samsung Advanced Institute for Health Sciences and Technology, Sungkyunkwan University, Seoul, Republic of Korea, ²Division of Endocrinology and Metabolism, Department of Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea.

Background and aims: Type 1 diabetes mellitus (T1DM) can develop in people of all ages. Compared to general population, individuals with type 1 diabetes had a high prevalence of mental illness that was estimated to be 1.61 odds ratio (95% CI; 1.07, 2.48). Previous studies have shown that mental illnesses in people with T1DM negatively affect glycemic control. In this population-based study, we aimed to determine the incidence of mental illness in adults with new-onset T1DM compared to general population.

Materials and methods: We used a nationwide Korean National Health Insurance (NHI) database which covers about 97% of the Korean population (50 million). We extracted T1DM patients aged more than or equal to 18 diagnosed between January 2009 and December 2015 from the NHIS database using diagnostic codes and insulin prescriptions (N=10,391) and matched age, sex, and year of index date with up to five reference individuals without diabetes from the general population (N=51,995). Mental illness including depression, mood disorders, anxiety and stress-related disorders, eating disorders, personality and behavior disorders, and alcohol and drug misuse disorders were defined by ICD-10 codes at least twice a year from index date to December 2020, outpatient or inpatient. Hazard ratios (HRs) with 95% confidence interval (CI) for incident mental illness were estimated using multivariable Cox proportional hazard regression models.

Results: The mean age of study participants was 56.9 ± 12.9 years and the median follow-up duration was 8.1 years. The incidence rate of mental illness was more than twice as high in patients with new-onset T1DM (55.3 per 1000 person-years) compared to individuals without diabetes (24.7 per 1000 person-years). New-onset T1DM was independently and positively associated with the development of mental illness (adjusted HR(aHR)=2.1 [95% confidence interval (CI)=2.0-2.2]), depression (aHR=3.1 [95% CI=3.0-3.3]), mood disorders (aHR=2.7 [95% CI=2.5-3.0]), anxiety and stress-related disorders (aHR=1.9 [95% CI=1.8-2.0]), eating disorders (aHR=3.0 [95% CI=1.9-4.8]), personality and behavior disorders (aHR=3.4 [95% CI=2.0-5.7]), alcohol and drug misuse disorders (aHR=4.0 [95% CI=3.3-5.0]), compared to non-diabetes subjects. Stratified by age of onset into quintiles, the risk of developing mental illness excluding anxiety- and stress-related disorders, was highest when T1DM onset was between the ages of 46 and 53 years.

Conclusion: New-onset T1DM in adulthood was significantly associated with a higher risk of mental illness including depression, eating disorders, personality and behavior disorders, and alcohol and drug misuse disorders compared to general population.

Disclosure: S. Kim: None.

112

Association between chronic complications of diabetes and diabetes distress: Can we prove it?

J. Konečná ^1,2, D. Lacko^3,4, E. Horová¹, N. Křesáková¹, K.D. Riegel^5,2;

¹3rd Department of Medicine - Department of Endocrinology and Metabolism, General University Hospital in Prague, Prague, Czech Republic, ²Dept of Psychiatry, 1st Faculty of Medicine, Charles University and General University Hospital, General University Hospital in Prague, Prague, Czech Republic, ³Institute of Psychology, Czech Academy of Sciences, Brno, Czech Republic, ⁴Department of Psychology, Masaryk University, Brno, Czech Republic, ⁵Dept of Addictology, 1st Faculty of Medicine, Charles University and General University Hospital, General University Hospital in Prague, Prague, Czech Republic.

Background and aims: The level of negative stress related to diabetes is captured mostly by the Diabetes Distress Scale (DDS). The result reflects the situation of patients during the last month, nevertheless, a high level of diabetes distress (DD) is considered a risk factor in developing chronic complications of diabetes. To prove the causality of DD and diabetes complications is a challenge, and the way to explore it is to focus on aspects more stable in time - personality traits of patients. The study aims to extend the previous knowledge of personality traits and its association with DD with the focus on a connection between chronic complications of diabetes and the prevalence of high levels of DD and complications among Czech people with diabetes.

Materials and methods: The sample comprised 347 participants with diabetes mellitus (DM) (63.1 % T1D, 36,9 % T2D). The level of DD was assessed by DDS, personality traits by Personality Inventory for DSM-5 (PID-5), and all subjects completed an anamnestic questionnaire focused on treatment and complications of diabetes.

Results: Elevated or high level of DD was confirmed in 53 % of people with T1D and 26,6 % of T2D. Higher levels of pathological personality traits (Negative affectivity; p<.001) were found in the group of patients with chronic complications of diabetes. The combination of complications and high levels of DD did not show any statistically significant difference in comparison to the group with complications and without high levels of DD. In the sample of Czech patients, people with T1D experienced higher level of DD (p<.001). Between frequency of co-morbid chronic complications and the level of DD was found weak correlation (r = 0.12).

Conclusion: Personality traits were associated with chronic complications of diabetes, however the association between the level of DD and chronic complications was not confirmed. Experiencing the burden of diabetes is affected more by the stable traits of the patient rather than the situation experienced in past weeks. In the Czech sample, people with T1D experienced higher levels of DD than patients with T2D which could be explained by different demands in diabetes treatment - only 9 from the T2D group were treated with intensive insulin therapy.

Supported by: Ministry of Health, Czech Republic - conceptual development research of organization [64165]

Disclosure: J. Konečná: Grants; Ministry of Health, Czech Republic [GJIH-1599-04-1-180] – conceptual development of research organization [64165] General University Hospital in Prague, Czech Republic.

113

Mental health coping strategies and glycaemic management among people with diabetes in Europe

E. Cox, E. Lin, T. Bell, T. Bristow, R. Wood;

dQ&A, San Fransisco, CA, USA.

Background and aims: People with diabetes (PWD) experience higher rates of anxiety and depression than the general population, though research on mental health coping strategies and diabetes management is limited. Thus, this research aimed to identify ways that PWD cope with mental health disorders and examine the relationship between mental health support and diabetes self-management among PWD with comorbid anxiety or depression.

Materials and methods: From October-November 2022, n=3,444 adults with diabetes in France, Germany, Italy, Netherlands, Sweden, and the United Kingdom took an online survey in which they self-reported their depression and anxiety status and their most recent HbA1c, if they knew it. Glucose sensor users also reported their Time in Range (TIR) (70-180 mg/dl). All respondents self-reporting anxiety (n=647) and/or depression (n=584) completed a set of questions regarding their preferred mental health coping mechanisms. The subsequent responses (61% type 1, 63% female) were analyzed using two-proportion Z-tests and reported at the 95% confidence level.

Results: In the overall sample, 17% of PWD self-report depression and 20% report anxiety; 8% report comorbid anxiety and depression. PWD with anxiety and/or depression have worse clinical outcomes than their counterparts. Compared to those who do not self-report anxiety and/or depression diagnoses, a greater proportion of PWD with either mental health disorder have an HbA1c above 7 (42% vs. 54%) or a TIR less than or equal to 70% (52% vs. 64%). Among those self-reporting anxiety and/or depression in Europe overall, the top mental health coping strategies are support from loved ones (44%), exercise (41%), and psychiatric medication (40%), though this differs by country. Psychotherapy is most common among those in the Netherlands (43%) and Germany (42%), and is least common in the UK, France and Italy (18%, 18%, and 18%). Respondents in Italy report the lowest level of psychiatric medication use (21%), while respondents in the Netherlands or the UK report the highest (50%). Relative to other countries, use of alternative medicine is most common in France (16%), while reliance on support groups is most common in the Netherlands (15%). Preferred avenues for mental health support also differ by diabetes type and mental health disorder diagnosis. Compared to Type 2s, a greater proportion of Type 1s rely on psychotherapy (27% vs. 20%) for support. Type 2s, however, have a higher rate of psychiatric medication use compared to T1s (46% vs. 36%). Respondents who have anxiety alone rely most on exercise (48%) or support from loved ones (43%), while those with depression alone or comorbid anxiety and depression report highest levels of psychiatric medication usage (47% and 51%, respectively). Lastly, this research reveals a relationship between glycemic management and mental health support. In the EU overall, respondents with an HbA1c above 7 or a TIR less than or equal to 70% report higher psychiatric medication usage (43% vs. 33%), (44% vs. 33%) and lower reliance on exercise for mental health management (34% vs. 50%), (35% vs 46%) than their counterparts.

Conclusion: This research concludes that the use of mental health coping tools differs by country, diabetes type, mental health status, and glycemic management. Further research is needed to determine the effectiveness of each strategy for patients with diabetes and comorbid anxiety or depression and to identify avenues for integrated treatment.

Disclosure: E. Cox: None.

114

Acceptability and feasibility of CGM for people with severe mental illness: a mixed-methods study

J.V.E. Brown ¹, P.A. Coventry^1,2, N. Siddiqi^1,3, R.A. Ajjan^4,5;

¹Department of Health Sciences, University of York, York, UK, ²York Environmental Sustainability Institute, University of York, York, UK, ³Bradford District Care NHS Foundation Trust, Bradford, UK, ⁴School of Medicine, University of Leeds, Leeds, UK, ⁵Leeds Teaching Hospitals NHS Trust, Leeds, UK.

Background and aims: People with serious mental illness (SMI; schizophrenia, schizoaffective disorder, bipolar disorder, psychosis) are at higher risk of type 2 diabetes (T2D) and experience worse outcomes than the general population with diabetes. People with SMI are often excluded from research into novel treatments and care pathways, including continuous glucose monitoring (CGM), perpetuating profound inequality. Therefore, we undertook a quantitative and qualitative exploration of the acceptability and feasibility of using CGM in this population.

Materials and methods: Participants in the DIAMONDS programme (community-dwelling adults [18 years and over] with confirmed diagnoses of SMI and T2D) were offered blinded CGM for a 14-day wear period. Information about device uptake and data completeness was collected, as well as summary blood glucose data. Qualitative interviews were conducted with participants about their experience wearing the sensor. Interviews were recorded, transcribed verbatim, and data analysed.

Results: Twenty-one individuals were invited to try blinded CGM for one 14-day wear period (43% female; mean±SD age 50.7±8.1 years). BMI was 36.9±8.7; HbA_1c 66.0±13.8 mmol/mol and 24% were on insulin therapies. 62% were White with the rest predominantly Asian Pakistani, Asian Indian, or Black. Of the 21 enrolled, 16 accepted and five declined to wear the sensor. In total, 21 sensors were used with these 16 participants. Of the 21 sensors, 14 were returned with data; two were lost/misplaced by participants during/after the wear-period; three were lost by staff; and two were discarded without any data recorded due to problems during fitting. Whenever feasible, a second sensor was used with participants who did not successfully return their first sensor with sufficient data. Out of the 14 sensors returned, all but one provided data for five or more days. Mean glucose management indicator was 61.9±14.6 mmol/mol, time in range was 56.1±28.6% with glucose variability, assessed as coefficient of variation, at 26.7±7.5%. Hypoglycaemic exposure in those on insulin therapies was 5±5%. Ten participants took part in an in-depth qualitative interview lasting an average of 23 minutes (range 5 to 45). Preliminary analyses indicate that participants had a positive experience using CGM with minimal complications, describing CGM as “easy” and “convenient”. Importantly, individuals were interested in learning about glucose levels and engaging with their diabetes self-management according to glucose data collected by the sensor.

Conclusion: In a small sample of people with SMI and T2D with high BMI and elevated HbA_1c, there was a signal that CGM is likely to be acceptable in this particularly vulnerable population with high illness burden. These findings suggest that the utility of CGM should be evaluated further in people with SMI as a tool to improve diabetes self-management and outcomes in this high-risk population.

Clinical Trial Registration Number: ISRCTN15328700

Supported by: NIHR PGfAR RP-PG-1016-20003

Disclosure: J.V.E. Brown: None.

OP 20 NAFLD and omics

115

Efficacy and safety of efinopegdutide in patients with nonalcoholic fatty liver disease and type 2 diabetes: results from an active-comparator-controlled study

S.S. Engel, R.R. Shankar, R.L.H. Lam, K.D. Kaufman;

Merck & Co., Inc., Rahway, NJ, USA.

Background and aims: Currently, there are no approved therapies for nonalcoholic steatohepatitis (NASH). While GLP-1/glucagon receptor coagonists can provide substantial reduction in liver fat, the contribution of glucagon-mediated effects could be impacted by the relative hyperglucagonemia present in patients with T2DM. We assessed the effects of efinopegdutide (EFI), a coagonist with ~2:1 GLP-1/glucagon receptor potency, relative to the selective GLP-1 receptor agonist semaglutide (SEMA), on liver fat content (LFC) in the type 2 diabetes mellitus (T2DM) subgroup of a study of patients with nonalcoholic fatty liver disease (NAFLD).

Materials and methods: This was a Phase 2a, randomized, active-comparator-controlled, parallel-group, open-label study in subjects with NAFLD (18-70 years, BMI 25-50 kg/m², stable body weight, without T2DM, or with T2DM with an A1C ≤ 8.5% controlled by diet ± stable dose of metformin). During a 4-week screening period, an MRI-PDFF was performed to determine LFC. Participants with an LFC of ≥10% were randomized in a 1:1 ratio to open-label SC EFI 10 mg Q1W or SC SEMA 1.0 mg Q1W for 24 weeks, stratified according to concurrent diagnosis of T2DM. Both drugs were titrated to the target dose over an 8-week period. The primary efficacy endpoint was the least squares (LS) mean relative reduction from baseline in LFC (%) after 24 weeks of treatment.

Results: Among 145 randomized subjects, 48 had T2DM (24 in each treatment group). In the T2DM subgroup at baseline, the mean A1C was 6.4%, mean FPG was 122.1 mg/dL and mean LFC was 18.6% and was similar in the two treatment groups. The LS mean relative reduction from baseline in LFC at Week 24 was significantly (p=0.010) greater with EFI (66.8% [90% CI 57.2, 76.3]) than with SEMA (45.6% [90% CI 36.0, 55.1]) (figure). Median relative reductions from baseline in LFC at Week 24 were 77.7% with EFI and 41.8% with SEMA. Greater proportions of participants had relative reductions from baseline at Week 24 in LFC of ≥30%, ≥50% and ≥70% with EFI (79.2%, 70.8% and 62.5%, respectively) compared with SEMA (70.8%, 45.8% and 20.8%, respectively). A greater proportion of participants achieved normal LFC (< 5%) at Week 24 with EFI (62.5%) compared with SEMA (25.0%). Both treatment groups had an LS mean relative reduction from baseline in body weight at Week 24 (EFI 7.8% vs SEMA 7.1%, p=0.621). Mean changes in A1C and FPG were 0.1% and 2.5 mg/dL, respectively, with EFI, and -0.8% and -22.8 mg/dL, respectively, with SEMA. There was no meaningful difference in the safety profile in the T2DM subgroup relative to all participants in each treatment group.

Conclusion: In T2DM patients with NAFLD, treatment with EFI 10 mg weekly led to a significantly greater reduction in LFC than SEMA 1 mg weekly. While glycemic improvement was attenuated with GLP-1/glucagon coagonism vs selective GLP-1 agonism, substantial benefits of coagonism on hepatic fat metabolism were retained.

Clinical Trial Registration Number: EudraCT: 2020-005136-30; NCT: 04944992

Supported by: MSD

Disclosure: S.S. Engel: Employment/Consultancy; Merck & Co., Inc. Stock/Shareholding; Merck & Co., Inc.

116

Social jetlag induces fatty liver via perturbed serum prolactin rhythm mediated circadian reprogramming of hepatic lipid metabolism

P. Zhang, X. Ma, Y. Bi;

Nanjing Drum Tower hospital, Nanjing, China.

Background and aims: The pandemic of circadian misalignment especially social jetlag (SJL) is associated with metabolic disorders. However, the relevant effect of SJL on liver has been underestimated.

Materials and methods: Two hundred and thirty participants who underwent liver biopsy were recruited, whose serum levels of prolactin (PRL) at 8:00, 16:00, and 24:00 were detected and subjected to circadian analysis. Liver tissue were used for RNA-seq. Using a novel machine learning algorithm called ZeitZeiger, we reconstructed the circadian transcriptome atlas of human liver tissues. C57BL/6 mice were exposed to an experimental jetlag light/dark cycle condition for 16 weeks to mimic SJL in humans. Lipid content in mice liver was evaluated on MRI scanning and histological analysis. Mice liver tissue were collected at 4-hour interval during 24 hours for RNA-seq. After 16 weeks of jetlag, mice were administrated with PRL via different delivery strategies for 2 weeks.

Results: SJL was a risk factor for biopsy-confirmed NAFLD (OR: 1.017, p<0.05, which was mediated through dampened serum PRL rhythm in humans (mediation effect: 0.039, p<0.05) . Employing a jetlag model in wild-type and Prl knockout mice, we confirmed that jetlag induced hepatic steatosis by affecting PRL rhythm in vivo. the P value of rhythmicity test for serum PRL in control and jetlagged mice is <0.05 and >0.05, respectively. Using luciferase report assay, EMSA and ChIP experiments, we revealed that Prl expressions were transcriptionally manipulated by core clock gene retinoid-related orphan receptor-alpha (RORα) , which could be perturbed by jetlag. Hepatic RNA-seq data of subjects showed that 721 differential expression genes (DEGs) between the control and SJL groups (331 were up-regulated and 390 were down-regulated), and 2955 DEGs between NAFLD with controls (1397 were up-regulated and 1558 were down-regulated). These DEGs were annotated to Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database, and 25 common pathways between DEGs from group1 and group 2 were identified, including pathways of PRL receptor signaling, circadian rhythm and entrainment and non-alcoholic fatty liver disease. Importantly, we for the first time illustrated circadian features of hepatic genes were remolded in SJL subjects based on a supervised machine learning informatics (mean absolute error: 2.65 hour). In particular, hepatic circadian transcriptome of PRL receptor (PRLR) signaling pathway was significantly reprogrammed. Accordingly, the oscillation of circadian rhythm and lipid metabolism in the liver were rewired and the resulting imbalance between lipogenesis and fatty acid oxidation contribute to SJL induced hepatic steatosis. Finally, realigning normal rhythmicities of PRL was more efficient in improving hepatic steatosis in jet-lagged than that treated with PRL continuously delivered over a 24-h period via a osmotic pump.

Conclusion: Here, we established circadian transcriptome atlas in human liver for the first time and examined the effect of SJL on human liver. Strikingly, reprogrammed hepatic PRLR signaling pathway with concomitant dysregulated lipid metabolism homeostasis was a causative mechanism of fatty liver under SJL, which was mediated through derailed serum PRL rhythm. Restoring PRL rhythm was more effective in alleviating SJL induced fatty liver compared with non-circadian administration, bringing a new field of chronotherapy in NAFLD.

Clinical Trial Registration Number: NCT03296605

Supported by: NSFC(82030026, 81770819 and 81900787)

Disclosure: P. Zhang: None.

117

Chronic treatment of mice with hepatocyte-released lipotoxic small extracellular vesicles causes liver and pancreas inflammation and alters insulin action

I. Garcia-Martinez ^1,2, R. Alén^1,2, N. Cobo-Vuilleumier³, B.R. Gauthier^3,2, A.M. Valverde^1,2;

¹IIBm Alberto Sols, CSIC-UAM, Madrid, Spain, ²Centro de Investigación Biomédica En Red de Diabetes y Enfermedades Metabólicas (CIBERDEM), Madrid, Spain, ³Cell Therapy and Regeneration, CABIMER, Sevilla, Spain.

Background and aims: Non-alcoholic fatty liver disease (NAFLD) is a common feature of obesity and type 2 diabetes mellitus (T2D). Under lipotoxic stress, hepatocytes release small extracellular vesicles (sEVs) which act locally and contribute to NAFLD progression, but also can reach the circulation and target metabolic tissues including the pancreas. We investigated whether hepatocyte-secreted sEVs under lipotoxic conditions impact on liver and pancreas inflammation and subsequent effects in both tissues.

Materials and methods: sEVs released by primary hepatocytes from C57BL6j male mice exposed to palmitic acid (sEV^PA) or BSA (sEV^C) were purified by differential ultracentrifugation. C57BL/6J male mice were injected intravenously with sEVs twice a week for four-weeks to study chronic effects of lipotoxic sEVs. Inflammation in liver and pancreas was analyzed by flow cytometry, immunofluorescence and RT-qPCR. Glucose tolerance test (GTT) and in vivo glucose-stimulated insulin secretion (GSIS) were performed to evaluate whether sEV^PA alter glucose homeostasis. Islets and hepatocytes from sEV-injected mice were isolated to study insulin secretion and insulin signalling, respectively.

Results: Chronic injection of sEV^PA induced liver inflammation characterized by enhanced monocyte, neutrophil and Kupffer cells recruitment (p<0.05). This effect was absent in mice injected sEV^C. CD3⁺ T lymphocytes were also elevated in sEV^PA-injected mice, being CD4⁺/CD8⁺ ratio lower (p<0.05) than that of mice injected sEV^C. These results were confirmed by immunofluorescence and mRNA analysis of inflammatory markers in the liver. Sirius red staining and Acta2 mRNA confirmed features of fibrosis in sEV^PA-injected mice. Moreover, sEV^PA treatment elevated plasma triglycerides, total cholesterol and ALT (p<0.05). sEV^PA-treated mice presented higher plasma insulin in both fed (p<0.05) and fasting, as well as HOMA-IR index, pointing to peripheral insulin resistance. In fact, hepatocytes from mice injected sEV^PA showed decreased insulin-induced AKT phosphorylation compared to those from sEV^C-injected mice. Whereas no differences were found in the GTT, in vivo GSIS increased in sEV^PA-treated mice compared to mice injected sEV^C (AUC p<0.05). In the pancreas, analysis of immune populations revealed an increase in total immune cells (CD45⁺) in sEV^PA-treated mice. Moreover, while the total pool of macrophages did not differ between groups, M2-polarized macrophages were reduced in mice receiving sEV^PA. Those mice showed a slight increase in CD3⁺ T cells and also presented reduced CD4⁺/CD8⁺ ratio due to an elevation of CD8⁺ cells (p<0.01). Immunofluorescence confirmed the presence of CD3⁺ infiltrates within islets and across the exocrine pancreas in mice injected sEV^PA. In line with in vivo GSIS, islets from sEV^PA-treated mice showed a two-fold increase in insulin secretion (p<0.05) and upregulation of Ins1 and Ins2 mRNA (p<0.05) compared to those receiving sEV^C, pointing to a compensatory response.

Conclusion: We have identified a liver-pancreas axis mediated by hepatocyte-released lipotoxic sEVs likely contributing to low-grade tissue inflammation, a key feature in beta-cell dysfunction in T2D.

Supported by: EFSD/Boehringer Ingelheim European Research Programme, Research funds of Community of Madrid

Disclosure: I. Garcia-Martinez: None.

118

Targeting NAFLD with umbilical cord-derived mesenchymal stromal cells: a preclinical study

M.J. Meneses ^1,2, I. Sousa-Lima¹, I. Ferreira¹, H. Cruz³, J.M. Santos³, M.P. Macedo^1,4;

¹iNOVA4Health, NOVA Medical School, Faculdade de Ciências Médicas, NMS, Universidade NOVA de Lisboa, Lisbon, Portugal, ²DECSIS II Iberia, Évora, Portugal, ³ECBio, Investigação e Desenvolvimento em Biotecnologia, S.A, Amadora, Portugal, ⁴APDP-ERC - Diabetes Portugal, Education and Research Center, Lisbon, Portugal.

Background and aims: Mesenchymal Stromal Cells (MSCs) have emerged as a promising therapeutic tool for their immunomodulatory, anti-inflammatory, and regenerative properties. Among the various sources of MSCs, a specific subset of MSCs isolated from umbilical cord tissue (UC-MSCs) has been found to be safe and effective in preclinical tests for several immune- and cardiovascular-related conditions. However, the mechanism of action of UC-MSCs in the context of diet-induced obesity and non-alcoholic fatty liver disease (NAFLD) remains unclear. In this study, we hypothesize that UC-MSCs treatment positively affects both glucose and lipid homeostasis, thereby alleviating NAFLD.

Materials and methods: To test this hypothesis, six-weeks old C57Bl/6J male mice were fed either a normal diet or a hypercaloric diet (HFat), with a third group receiving the hypercaloric diet and a UC-MSCs treatment (HFat-MSCs). The UC-MSCs were weekly administered at a dose of 10E6 cells (i.p.) from 11 to 18 weeks of age. Insulin sensitivity and glucose tolerance were assessed (by ITT and GTT, respectively) after 5 weeks of diet and at the end of the study. We also evaluated the levels of insulin signaling pathway-related proteins and lipid metabolism-related genes.

Results: The HFat diet led to body weight gain, hyperglycemia, and glucose intolerance compared to the control group, indicating a prediabetic state. On the contrary, the HFat-MSCs group showed protection against body weight gain compared to the untreated HFat group, without affecting food nor caloric intake. Moreover, the HFat-MSCs group had lower random glucose levels and improved insulin sensitivity and glucose tolerance compared to the HFat group (p<0.05). The HFat-MSCs group also had a decreased liver/body weight ratio compared to the control group. Additionally, the HFat-MSCs group had decreased hepatic cholesterol and triglyceride content (p<0.05) compared to the HFat group, concurrent with less lipid droplets and NAFLD score as assessed by a pathologist. These changes might be a consequence of decreased liver pyruvate kinase (p<0.001), acetyl-CoA carboxylase (p<0.05), and Srebp2 (p<0.001) mRNA expression levels in the HFat-MSCs group compared to the HFat group. Interestingly, MSCs treatment decrease resulted in a restoration of these mRNA levels to normal diet levels (n.s.).

Conclusion: Our results indicate that UC-MSCs have a protective role in diet-induced obesity, prediabetic profile, and NAFLD. UC-MSCs protect against hepatic lipid accumulation by restoring the levels of lipid metabolism-related genes such as Srebp2. These findings provide a steppingstone for further investigation of these patented UC-MSCs as a potential therapy for metabolic disorders.

Supported by: FCT UIDB/Multi/04462/2020; Portugal2020, ALT20-03-0247-FEDER-113469 and LISBOA-01-0247-FEDER-113469

Disclosure: M.J. Meneses: None.

119

Change in Fibrosis-4 Index is associated with risk of liver events, cardiovascular events, and all-cause mortality in patients with obesity and/or type 2 diabetes

K. Khunti ¹, T.L. Berentzen², L.M. Nitze², M. Jara², A. Jensen², M.S. Kjær², K.K. Mangla², J.M. Tarp², Q.M. Anstee^3,4;

¹Diabetes Research Centre, University of Leicester, Leicester General Hospital, Leicester, UK, ²Novo Nordisk A/S, Søborg, Denmark, ³Translational & Clinical Research Institute, Faculty of Medical Sciences, Newcastle University, Newcastle upon Tyne, UK, ⁴Newcastle NIHR Biomedical Research Centre, Newcastle upon Tyne Hospitals NHS Trust, Newcastle upon Tyne, UK.

Background and aims: There is a lack of tools to monitor and assess risk of morbidity and mortality in people with non-alcoholic steatohepatitis (NASH). We evaluated the association of 12-month changes in FIB4 (ΔFIB4) and risk of developing severe NASH-related clinical events.

Materials and methods: A longitudinal cohort study was conducted using data from the UK Clinical Practice Research Datalink linked with Hospital Episodes Statistics and Office for National Statistics data. Patients were aged ≥18 years with obesity and/or T2D, ≥2 FIB4 measurements, and no alcohol-related disorders, chronic liver diseases other than non-alcoholic fatty liver disease or prescriptions of drugs inducing liver disease. ΔFIB4 was calculated using score at baseline and after 12 (±3) months. Patients were followed from second measurement until time of first liver event; time of first cardiovascular (CV) event (hospitalization/death); all-cause mortality; database migration; 10 years’ follow-up; or 1 Jan 2020, whichever was first. Aalen-Johansen and Cox proportional hazards models were used to estimate cumulative incidence and hazard ratios (HRs).

Results: Among 20,443 included patients, there were 466 liver events. In patients with high baseline FIB4 (>2.67), risk of a liver event after 10 years was 12.8%; after 12 months, the risk was 18.5% and 10.1% in those with FIB4 increase and decrease, respectively (Figure). Increased risk with increasing FIB4, and vice versa, was also seen in patients with an indeterminate (1.30-2.67) or low baseline FIB4 (<1.30) (Figure). Risk of a liver event was positively associated with ΔFIB4 in Cox models adjusted for sex, age, and baseline FIB4, with the association depending on baseline FIB4. Compared with patients with low baseline FIB4 and no change in FIB4 (reference), the HR (95% confidence interval) was 24.27 (16.98, 34.68) for those with high baseline FIB4 and a 1-unit FIB4 increase, and 10.90 (7.90, 15.05) for those with high baseline FIB4 and a 1-unit decrease. Compared with the reference, those with indeterminate and low baseline FIB4 and 1-unit FIB4 increase/decrease also had significantly higher/lower risk. Similar results were seen with CV events and mortality, though the association with CV events was attenuated after adjustment for age.

Conclusion: In patients with obesity and/or T2D, changes in FIB4 at 12 months were positively associated with risk of NASH-related clinical events across all FIB4 groups, highlighting the potential of using FIB4 to identify patients at risk of severe events.

Supported by: Funded by Novo Nordisk

Disclosure: K. Khunti: Employment/Consultancy; Abbott, Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Lilly, Merck Sharp and Dohme, Novo Nordisk, Roche, Sanofi-Aventis, Servier. Grants; AstraZeneca, Boehringer Ingelheim, Lilly, Merck Sharp & Dohme, Novo Nordisk, Sanofi-Aventis.

120

Oral insulin alleviates liver fibrosis and reduces fat in patients with NASH and type 2 diabetes: results of phase II clinical trial

M. Kidron ¹, Y. Ishay², J. Neutel³, Y. Kolben², R. Gelman², O. Sneh Arbib², O. Lopez⁴, H. Katchman⁵, R. Sotolongo⁶, Y. Ilan⁷;

¹Oramed Pharmaceuticals, Jerusalem, Israel, ²Hadassah Medical Center, Jerusalem, Israel, ³Orange County Research Center, Tustin, CA, USA, ⁴Integrium LLC, Tustin, CA, USA, ⁵Sourasky Medical Center, Tel Aviv, Israel, ⁶Research Institute of South Florida, Miami, FL, USA, ⁷Hebrew University of Jerusalem, Jerusalem, Israel.

Background and aims: NAFLD is the most common chronic liver disorder globally. NASH is an advanced form of NAFLD and a leading cause of end-stage liver disease and indication for liver transplantation. Insulin resistance and inflammation underlie the pathogenesis of the disease. This double-blind, randomized, placebo-controlled, multicenter study aimed to determine the effect of oral insulin in patients with NASH and T2DM.

Materials and methods: Patients were treated twice daily for 12 weeks with an 8 mg insulin (n=21) or placebo (n=11) capsule. Safety was monitored throughout the study. At screening and after 12 weeks of treatment, liver fat content was measured by MRI-PDFF and liver fibrosis and steatosis levels by FibroScan®.

Results: No severe drug-related adverse events were reported during the study. After 12 weeks of treatment, mean percent change from baseline whole-liver (-11.2±22.8% vs. -6.5±15.3%, respectively) and liver segment 3 (-11.7±21.1% vs. +0.1±17.3%, respectively) fat content were higher in the insulin as compared to the placebo arm. Patients receiving insulin showed a median -1.1 kPa and -21.0 dB/m reduction from baseline fibrosis and steatosis levels, respectively. In comparison, placebo-treated patients showed median increases of 0.3 kPa and 13.0 dB/m, respectively. At Week 12, HbA1c levels decreased by a mean -0.27±1.3% in the oral insulin cohort and increased by a mean 0.23±0.73% in the placebo cohort. Mean percent changes from baseline ALT and AST levels were -0.1±30.4% and -0.8±28.6%, respectively, in the oral insulin arm and 3.0±26.7% and 13.4±46.6%, respectively, in the placebo arm. Over the treatment period, mean total cholesterol, HDL, LDL, and triglycerides levels were consistently lower than baseline in the active arm and higher than baseline in the placebo arm.

Conclusion: The results of this feasibility study support the safety and potential therapeutic effect of orally delivered insulin on liver fibrosis, fat accumulation, and inflammatory processes.

Clinical Trial Registration Number: NCT04618744

Supported by: Oramed Pharmaceuticals

Disclosure: M. Kidron: Employment/Consultancy; Oramed Pharmaceuticals. Stock/Shareholding; Oramed Pharmaceuticals.

OP 21 New strategies for diabetic retinopathy from mice to the clinic

121

Fasting prevents cholesterol crystal formation and the development of diabetic retinopathy

J. Busik ¹, T.F. Dorweiler¹, Y. Adu Agyeiwaah², N. El-Darzi³, N. Mast³, G. Abela⁴, I. Pikuleva³, M.B. Grant⁵;

¹Michigan State University, East Lansing, MI, USA, ²The School of Optometry, University of Alabama at Birmingham, Birmingham, AL, USA, ³Department of Ophthalmology & Visual Sciences, Case Western Reserve University, Cleveland, OH, USA, ⁴Department of Medicine, Michigan State University, East Lansing, MI, USA, ⁵Ophthalmology, University of Alabama at Birmingham, Birmingham, AL, USA.

Background and aims: With the advancement of retinal imaging techniques, several types of hyperreflective deposits were identified in the diabetic retina. These include hyperreflective crystalline deposits that were recently identified as cholesterol crystals (CCs). We have demonstrated that fasting normalizes retinal cholesterol metabolism in db/db type 2 diabetic mouse model through Sirtuin 1 (SIRT1)-LXR dependent activation of reverse cholesterol transport (RCT). This study tested the hypothesis that fasting-mediated activation of SIRT1-LXR-RCT will effectively reduce CC formation and the development of diabetic retinopathy.

Materials and methods: Lepr^db/db(db/db) type 2 diabetes mouse model Lepr^db/m(db/m) controls were used in the experiments. HbA1C was measured every 3 months. Long term (6 months) intermittent fasting regiment (12h/12h) was used to normalize cholesterol metabolism. Retinal cholesterol input was measured using D₂O, and D₇-cholesterol. CC were analyzed by modified Scanning Electron Microscopy (SEM) and immunohistochemistry. Inflammatory markers were analyzed using real time qPCR. Retinal vascular permeability was measured using FITC-albumin. Data were analyzed by ANOVA with post-hoc analysis by Tukey’s range test.

Results: The total retinal cholesterol levels were significantly increased in diabetic mice (4±0.5 mg/g retina, n=5) as compared to controls (2.9±02 mg/g retina, n=5; P<0.05). The increase was due to retinal uptake of systemic cholesterol as determined by D₇-cholesterol measurements showing 35±7%, n=5 of cholesterol uptake in control and 81±9%, n=5 in 6 months diabetic mice. Increased cholesterol levels were associated with CC formation in diabetic retinas (19.7±7 crystals/0.01mm², n=5) No CC were observed in age-matched control animals. As fasting was effective at SIRT1-LXR and RCT activation, we next determined the effect of fasting on CC formation in the retina of db/db mice. Strikingly, fasting removed CC from diabetic mouse retinas (0.5±0.7 , crystals/0.01mm²,n=5, P<0.01), reduced inflammatory phenotype as indicated by expression of inflammatory markers TNF-α, IL-1β and ICAM-1 in diabetic retinas, and normalized retinal vascular permeability.

Conclusion: This study demonstrated cholesterol accumulation and CC formation in diabetic retinas. CC can activate key pathogenic events in diabetic retina. Fasting regiments shown to be effective at correcting retinal cholesterol imbalances removed CC from diabetic retinas and normalized long term diabetes-induced retinal inflammatory markers and vascular permeability.

Supported by: NIH NEI

Disclosure: J. Busik: Employment/Consultancy; Ceramedix, Inc.

122

NOD1 deficiency ameliorates the progression of diabetic retinopathy via modulating bone marrow-retina crosstalk

Y. Duan, J. Wu, W. Chen, D. Liu;

The Second Affiliated Hospital of Chongqing Medical University, Chongqing, China.

Background and aims: The nucleotide-binding oligomerization domain-containing protein 1 (NOD1) plays a critical role in provoking metabolic inflammation in diabetes. Additionally, its ligand has been reported to disturb the balance of bone marrow-derived hematopoietic stem/progenitor cells, which are crucial in the pathogenesis of diabetic retinopathy (DR). We hypothesized that NOD1 depletion would prevent the progression of DR by repairing bone marrow dysfunction.

Materials and methods: To investigate the role of NOD1 in DR, we generated NOD1^-/--Akita (genotype: NOD1^-/--Ins2^{WT/ C96Y}) double mutant mice and age-matched control groups. Retinal phenotypes were accessed by OCT, ERG and quantification of acellular capillaries. We also performed bone marrow transplantation to generate chimeric mice with bone marrow-specific depletion of NOD1 and analyzed bone marrow-derived hematopoietic cell types using flow cytometry.

Results: We found elevated circulating NOD1 activators in Akita mice after 6 months of diabetes. Retinal thickness was reduced in Akita mice compared to WT controls. However, the retinal thickness of the NOD1^-/--Akita mice was similar to that of WT and NOD1^-/- mice, suggesting NOD1 depletion partially restored diabetes-induced structural changes in cross-section retina. Additionally, NOD1^-/--Akita mice showed an increase in b-wave amplitudes under both scotopic and photopic conditions measured by ERG when compared to Akita alone, suggesting NOD1 depletion improved retinal cell responses to light stimulation in diabetes. A significant increase in acellular capillary numbers was observed in the retina from Akita mice, but NOD1 knockout largely ameliorated the progression of diabetic retinal vascular degeneration. We also found that the percentages of ST-HSCs and LT-HSCs were partially restored in the bone marrow from NOD1^-/--Akita mice compared to Akita alone. NOD1 depletion corrected the shift of hematopoiesis toward a more proinflammatory state in diabetes, as measured by both flow cytometry and CFU assay, suggesting NOD1 depletion-prevented DR progression is associated with the alteration of bone marrow phenotype. Further, hematopoietic-specific NOD1 deletion mitigated diabetes-induced alterations in retinal cross-section structure and improved the electrical responses of diabetic retina toward light stimuli. Enumeration of acellular capillaries revealed that diabetes-mediated increase in such pathologic change was alleviated by hematopoietic NOD1 ablation. Finally, loss of hematopoietic NOD1 showed a protective effect in DR, as fewer bone marrow derived monocytes infiltrated into the retina in the NOD1^-/--Akita→Akita chimeric mice.

Conclusion: These findings suggest that NOD1 may serve as a critical trigger for hematopoietic imbalance, leading to the progression of DR. Targeting NOD1 in bone marrow may represent a promising strategy for the prevention and treatment of DR.

Supported by: NSFC (No. 81900758) and CSTC (cstc2021jcyj-msxmX0250) to Y.D.

Disclosure: Y. Duan: None.

123

Anti-ceramide immunotherapy: a systemic treatment for diabetic retinopathy

T.F. Dorweiler ¹, A. Singh², G.J. Blanchard³, R.N. Kolesnick², J.V. Busik¹;

¹Department of Physiology, Michigan State University, East Lansing, MI, USA, ²Memorial Sloan Kettering Cancer Center, New York, NY, USA, ³Department of Chemistry, Michigan State University, East Lansing, MI, USA.

Background and aims: The diabetic metabolic insult leading to retinal vascular degeneration involves initial endothelial cell damage due to low-grade chronic inflammation that is then inadequately repaired due to compromised availability and functionality of bone marrow derived circulating angiogenic cells (CACs). The molecular metabolic link connecting both initial inflammation in the retina and dysfunctional CACs involves stress-induced activation and secretion acid sphingomyelinase (ASM), converting sphingomyelin into pro-inflammatory and pro-apoptotic ceramide in the outer leaflet of endothelial cells. Ceramide spontaneously coalesces into ceramide-rich platforms (CRP) (0.5-5.0 mm diameter). The outcome of this process in retinal endothelial cells (REC), is pro-inflammatory and apoptotic signaling. To intervene therapeutically in CRP-mediated microvascular inflammation and apoptosis, we developed a set of anti-ceramide antibodies (Abs). In this study, anti-ceramide Abs were used to restore endothelial homeostasis in animal and cell culture models of diabetic retinopathy.

Materials and methods: Bovine retinal endothelial cells (BRECs) were used as an in vitro model for RECs. Apoptosis was assessed using bisbenzimide. In vitro CRPs were quantified in BRECs and CACs using confocal imaging. Membrane fluidity was measured using fluorescence recovery after photobleaching and confocal microscopy ex vivo. Streptozotocin was utilized to induce hyperglycemia as a type-1 diabetes model. Anti-ceramide Abs were administered in vitro or via a single dose into the vitreous. Pro-inflammatory markers were analyzed via qRT-PCR. Vascular degeneration was captured using FITC-albumin technique.

Results: BRECs generate CRPs upon acute stimulation with 20 ng/mL TNFα/IL-1β after 45 sec (40/25%) and was fully ameliorated when treated with anti-ceramide Abs. 10ng/mL TNFα/IL-1β dose-dependently induced REC apoptosis (40/30%) in BRECs over 24 hrs and was fully mitigated when treated with anti-ceramide Abs. Intravitreal injections of anti-ceramide Abs was next tested in STZ-induced rat diabetes model. A single injection at the onset of hyperglycemia prevented diabetes-induced increase in retinal vascular permeability (n=10-18 per group; p<0.001, ANOVA). Ex vivo analysis of CACs showed a significant increase (42%) in CRP formation starting on day 3, with time-dependent increases on day 7, 14, 42, and 180 (70%) post-hyperglycemia in STZ mice compared to controls (n=9-12 animals per group per time point, p<0.05, Students T-test). CACs treated with 10 ng/mL TNFα showed reduced fluidity compared to vehicle (p<0.05, ANOVA). CACs were protected when treated with 100 μg/mL Abs (n=12-15, p<0.05, ANOVA).

Conclusion: Diabetes and diabetogenic stimuli lead to CRP formation, inflammatory signaling and apoptosis in REC as well as CAC in in vivo, ex vivo and cell culture models of diabetic retinopathy. Inhibition of CRPs by anti-ceramide Abs treatment returns REC and CAC to homeostasis as indicated by inhibition of inflammatory signaling, normalized retinal vascular permeability, and membrane fluidity and function of CACs. Collectively, our results present anti-ceramide immunotherapy as a potential treatment to alleviate chronic inflammation and restore microvascular repair via CACs to halt or reverse DR progression.

Supported by: R01EY030766 awarded to RNK and JVB.

Disclosure: T.F. Dorweiler: None.

124

Glucose tolerance and insulin resistance/sensitivity associate with retinal layer characteristics: the LIFE-Adult-Study

T. Ebert ¹, F.G. Rauscher², T. Elze³, M. Francke⁴, Y. Li³, K. Wirkner⁵, A. Tönjes¹, C. Engel², J. Thiery⁶, M. Blüher⁷, M. Stumvoll¹, T. Kirsten⁸, M. Loeffler², M. Wang³;

¹Medical Department III – Endocrinology, Nephrology, Rheumatology, University of Leipzig Medical Center, Leipzig, Germany, ²Leipzig University, Institute for Medical Informatics, Statistics, and Epidemiology (IMISE) and Leipzig Research Centre for Civilization Diseases (LIFE), Leipzig, Germany, ³Schepens Eye Research Institute of Mass Eye and Ear, Department of Ophthalmology, Harvard Medical School, Boston, MA, USA, ⁴Leipzig University, Institute for Medical Informatics, Statistics, and Epidemiology (IMISE), Leipzig, Germany, ⁵Leipzig University, Leipzig Research Centre for Civilization Diseases (LIFE) and Institute for Medical Informatics, Statistics, and Epidemiology (IMISE), Leipzig, Germany, ⁶Leipzig University, Institute of Laboratory Medicine, Clinical Chemistry and Molecular Diagnostics, Leipzig, Germany, ⁷Helmholtz Institute for Metabolic, Obesity and Vascular Research (HI-MAG) of the Helmholtz Zentrum, Helmholtz Zentrum München at the University of Leipzig, Leipzig, Germany, ⁸Medical Data Science, University of Leipzig Medical Center, Leipzig, Germany.

Background and aims: As the prevalence of insulin resistance and glucose intolerance is increasing throughout the world, diabetes-induced eye diseases are a global health burden. We aim to identify distinct retinal layers, which are susceptible to changes in insulin and glucose metabolism, using non-invasive, high-resolution spectral-domain optical coherence tomography (SD-OCT) in a population-based dataset.

Materials and methods: The LIFE-Adult-Study randomly selected 10,000 participants from the population of Leipzig, Germany. Standardized phenotyping included the assessment of metabolic risk markers and SD-OCT-derived thicknesses of 10 retinal layers. Retinal layer thicknesses were investigated in 7,384 healthy eyes stratified by normal glucose tolerance, prediabetes, and diabetes. The association of retinal layer characteristics with different indices of glucose tolerance, insulin resistance, or insulin sensitivity was determined using multivariable linear regression analyses for the individual markers adjusted for age, sex, and refraction. Various sensitivity analyses validated the observed findings.

Results: In the study cohort, most investigated retinal layers showed significant sex- and glucose tolerance-dependent differences in layer thicknesses. Multivariable linear regression analyses revealed a significant, independent, and inverse association between markers of glucose tolerance and insulin resistance with the thickness of the outer nuclear layer (ONL) and myoid zone (MZ). Conversely, insulin sensitivity positively and independently associated with both layers. Early-Treatment-Diabetic-Retinopathy-Study-based sectors confirmed our global associations with retinal layer thicknesses. Sensitivity analyses further validated our findings after stratifying the cohort by glucose homeostasis.

Conclusion: An impaired glucose homeostasis associates with a thinning of glucose-responsive layers ONL and MZ. Changes in ONL and MZ thicknesses might predict early metabolic retinal alterations. Furthermore, implementing retinal SD-OCT analyses into clinics and research requires adjustment for markers of glucose homeostasis.

Supported by: EU, ESF, Saxony, Lions Foundation, NIH, DFG, EFSD/AZ Mentorship Programme, DDG, Grimshaw-Gudewicz Foundation; Research to Prevent Blindness; BrightFocus Foundation

Disclosure: T. Ebert: None.

125

Risk factors for retinopathy in persons with type 1 diabetes followed from diagnosis and onward

S. Seyed Ahmadi ^1,2, J. Ludvigsson³, T. Nyström⁴, M. Lind^1,5;

¹Institute of Medicine, Gothenburg, Sweden, ²Department of Internal Medicine, Sahlgrenska University Hospital, Gothenburg, Sweden, ³Department of Biomedical and Clinical Sciences, Crown Princess Victoria Children’s Hospital, and Division of Paediatrics, Linköping University, Linköping, Sweden, ⁴Department of Clinical Science and Education, Internal Medicine, Södersjukhuset, Karolinska Institutet, Stockholm, Sweden, ⁵Department of Medicine, NU Hospital Group, Uddevalla, Sweden.

Background and aims: It is well known that persons with type 1 diabetes and poor glycemic control have an increased risk of retinopathy. However, the impact of other biomarkers and their relative importance for the development of retinopathy is less well understood.

Materials and methods: Children and adults with type 1 diabetes were included from the Swedish National Diabetes Registry and followed from January 1, 2001, to December 31, 2017. The association between risk factors and retinopathy was evaluated with the gradient of risk per 1 SD (standard deviation) to estimate the impact of each risk factor. The risk factors were evaluated as mean of longitudinally collected values from diagnosis and onwards. Retinopathy was defined as simplex retinopathy or worse. In all analyses of non-glucose related risk factors adjustments were made for age, sex and mean HbA1c-level.

Results: A total of 10398 persons (43.4 % females) with type 1 diabetes were included in the analysis. They had a mean age of 14.7 years at first visit and mean HbA1c varied between 7.9-8.3% (62.4-66.7 mmol/mol) for various follow-up times. For the risk of any retinopathy, HbA1c was the strongest risk factor with an odds ratio of 1.73 (95% CI 1.64-1.82, p<0.0001). Other risk factors of retinopathy were systolic blood pressure OR 1.20 (95% CI 1.15-1.26, p<0.0001), diastolic blood pressure 1.18 (95% CI 1.12-1.23, p<0.0001), BMI 1.21 (95% CI 1.15-1.27, p<0.0001) and smoking 1.06 (95% CI 1.01-1.11, p=0.0088). Increased HDL was associated with reduced risk of retinopathy, OR 0.94 (95% CI 0.89-0.99, p=0.013). LDL, triglycerides, and cholesterol showed no association. For more severe forms of retinopathy (nonproliferative and proliferative), other risk factors than HbA1c did not reach statistical significance due to relatively few events.

Conclusion: The current study supports that healthy lifestyle habits are important for the development of retinopathy in children and young adults. Individuals with type 1 diabetes, followed from diagnosis over at least 10 years, with high BMI, hypertension, smokers and those with low HDL developed retinopathy to a greater extent.

Disclosure: S. Seyed Ahmadi: None.

126

Incidence and risk factors for diabetic retinopathy at diagnosis of type 2 diabetes: an observational study of 77681 patients from the Swedish National Diabetes Registry

S. Sofizadeh ¹, C. Zhou², K. Eeg-Olofsson², M. Lind³;

¹Närhälsan Dalaberg primary health care, Uddevalla, Sweden, ²The Swedish National Diabetes Register, Gothenburg, Sweden, ³Department of Molecular and Clinical Medicine, Institute of Medicine, Gothenburg, Sweden.

Background and aims: The aim was to assess the prevalence of diabetic retinopathy (DR) in persons with newly diagnosed type 2 diabetes (T2D) to understand a potential need for intensified screening for early detection of T2D. We also evaluated if certain populations were at specifically high risk.

Materials and methods: Individuals from the Swedish national diabetes register (NDR) with a retinal photo <2 years after diagnosis of T2D were included. The proportion of patients with retinopathy (simplex or worse) was assessed. Patient characteristics and risk factors at diagnosis were analyzed in relation to DR with logistic regression.

Results: In total 77,681 individuals with newly diagnosed T2D, mean age 62.6 years and 41.1% females were included between January 2015 to December 2019. Of these, 17.2%, n=13329 had DR. In a subgroup of 72681 (94% of the total cohort) with information of patient characteristics DR was more common in older persons OR 1.03 (95% CI 1.01-1.05) and men compared with women, OR 1.10 (95% CI 1.06-1.15). Other variables associated with DR were lower education OR 1.08 (95% CI 1.02-1.14), previous stroke 1.18 (95% CI 1.07-1.30), chronic kidney disease 1.30 (95% CI 1.07 - 1.56), treatment with acetylsalicylic acid 1.14 (95% CI 1.07-1.21) and ACE-inhibitor 1.12 (95% CI 1.05-1.18). With respect to geographic areas DR was more common in individuals born in Asia 1.16 (95% CI 1.08-1.25) compared with those born in Sweden. DR was monotonically associated with higher HbA1c OR 1.75 (95% CI 1.66-1.85) at HbA1c >70 mmol/mol compared to the group with HbA1c below 48 mmol/mol. In a subgroup of 73% of patients who in addition had information of blood pressure and BMI at <3 months after diagnosis of T2D increased systolic BP > 140 mmHg OR 1.33 (95% CI 1.21-1.47) was associated with increased risk of DR compared to systolic BP110 mmHg whereas BMI showed an inverse association OR 0.72 (95% CI 0.66-0.78) for BMI >35 kg/m2.

Conclusion: Intensified screening of T2D and prediabetes is needed in Sweden since almost 1/5 of persons diagnosed with T2D have retinopathy already at diagnosis of T2D indicating long-standing hyperglycaemia before diagnosis. The prevalence of DR was somewhat higher in certain patient groups including men, those with previous stroke, kidney disease, hypertension and born in certain regions outside Sweden.

Disclosure: S. Sofizadeh: None.

OP 22 Improving the diagnosis and prediction of kidney disease progression

127

Receptor-interacting protein kinase (RIPK)3 regulates podocyte structure and function in diabetic kidney disease

E. Lee ^1,2, J. Kang¹, N.-J. Cho¹, S. Lee², J.-H. Lee³, J. Lee⁴, M. Choi², S. Son², M. Park², J. Park⁴;

¹Department of Internal Medicine, Soonchunhyang University Cheonan Hospital, Cheonan, Republic of Korea, ²Department of Medicine and BK21 FOUR Project, Soonchunhyang University, Cheonan, Republic of Korea, ³Department of Pathology, Soonchunhyang University Cheonan Hospital, Cheonan, Republic of Korea, ⁴Department of School of Life Sciences, Gwangju Institute of Science and Technology, Gwangju, Republic of Korea.

Background and aims: Receptor-interacting protein kinase (RIPK)3 is an essential molecule for necroptosis and its role in kidney fibrosis has been investigated using various kidney injury models. However, the relevance and the underlying mechanisms of RIPK3 of podocyte injury in diabetic kidney disease (DKD) are poorly understood.

Materials and methods: Diabetic animal model was induced by a high fat diet in Ripk3 knockout (KO) mice. For in vitro research, we exposed mouse or human podocytes to high glucose (30 mM), with or without GSK872, a RIPK3 inhibitor. The role and the underlying mechanism of RIPK3 in diabetic podocyte injury were analyzed in vivo, in vitro, and human DKD specimens.

Results: We used single-cell RNA sequencing on kidney cortex to characterize cell-type-specific gene expression. Cell trajectory analysis identified that Ripk3 had a function associated with the mitochondrial pathway. RIPK3 deficiency in DKD mice improved albuminuria, podocyte numbers, and renal histopathological features including foot process effacement and glomerular basement membrane (GBM) thickening. Increased mitochondrial fragmentation, upregulated mitochondrial fission-related proteins such as phosphoglycerate mutase family member 5 (PGAM5) and dynamin-related protein 1 (Drp1), and mitochondria dysfunction were decreased in RIPK3-depleted diabetic podocytes both in vitro and in vivo. By contrast, RIPK3 overexpression was sufficient to increase PGAM5 expression and mitochondrial fragmentation due to mitochondrial translocation of Drp1. Our data showed that expression of RIPK3 in diabetic podocytes was upregulated and that such upregulation mediated the development of diabetic podocytopathy, likely by regulating mitochondrial fission via PGAM5-Drp1 signaling thorough MLKL and by remodeling actin cytoskeleton. Furthermore, we found that RIPK3 level was upregulated in podocytes and plasma from human DKD cohort. Such upregulation was correlated with podocyte loss, albuminuria, and poor renal outcome in humans.

Conclusion: In summary, our results demonstrate that RIPK3/MLKL pathway is activated in podocytes from diabetic kidney. In addition, this pathway can regulate mitochondrial fission and contribute to podocytes injury through PGAM5-Drp1 signaling. Inhibition of RIPK3 could protect podocytes against mitochondrial dysfunction through suppression of PGAM5 and Drp1, suggesting that RIPK3 might be a novel therapeutic target for preventing DKD.

Supported by: NRF-2020R1A2C2003438

Disclosure: E. Lee: Grants; National Research Foundation (NRF) of Korea grant funded by the Korea Government (Ministry of Science and ICT) (2020R1A2C2003438).

128

Renal function based on serum cystatin C versus creatinine in determining mortality risk in type 2 diabetes: the Fremantle Diabetes Study Phase II

W.A. Davis ¹, P. Chubb², D.G. Bruce¹, T.M.E. Davis¹;

¹Medical School, University of Western Australia, Fremantle, Australia, ²PathWest Laboratory Medicine WA, Murdoch, Australia.

Background and aims: To determine whether estimated glomerular filtration rate (eGFR) based on serum cystatin C (cysC; eGFR_cysC) is superior to conventional serum creatinine-based eGFR (eGFR_creat) in determining 10-year mortality risk in community-based people with type 2 diabetes.

Materials and methods: Ninety-nine percent (1,466) of the 1,482 participants in the Fremantle Diabetes Study Phase II type 2 diabetes cohort (recruited 2008-11) were ≥17 years at entry and had both serum cysC and creatinine measurements. eGFR was calculated using the respective Chronic Kidney Disease Epidemiology Collaboration equations. Ten-year mortality was ascertained to end-2021 through validated data linkage. Receiver operating characteristic curves were compared for eGFR_cysC and eGFR_creat versus death. Cox regression was used to identify the most parsimonious model of independent associates of death excluding eGFR. eGFR_cysC and eGFR_creat were then entered separately as continuous and categorical variables. The likelihood ratio test (LRT) was used to discriminate between models.

Results: At baseline, the participants were (mean±SD) 66±11 years old, 52% were males, with median [interquartile range] diabetes duration 9 [3-16] years; eGFR_cysC and eGFR_creat were 64.9±23.4 and 79.7±21.4 mL/min/1.73m², respectively. Forty-three percent of participants were in the same eGFR category irrespective of equation, but 55% were in a higher eGFR category when calculated using creatinine (see Table). A quarter had eGFR_cysC <60 but eGFR_creat ≥60 mL/min/1.73m² while 4% had eGFR_cysC <30 but eGFR_creat ≥30 mL/min/1.73m². During 10 years of follow-up, 384 (26%) died. The area under the curve for eGFR_cysC was greater than that for eGFR_creat (difference 0.053, P=0.006). Age, sex, Asian ethnicity, being married, current smoking, diabetes duration, obesity, heart rate, systolic blood pressure, warfarin use, urine albumin:creatinine ratio, peripheral arterial disease, heart failure and Charlson Comorbidity Index ≥3 were independently associated with mortality; all models that included eGFR were better than this most parsimonious model (LRT, P≤0.0006). Continuous eGFR_cysC performed best but was not significantly better than categorical eGFR_cysC or categorical eGFR_creat (P≥0.145). The continuous eGFR_creat performed worst including versus categorical eGFR_creat (P=0.002). The relationship between categorical eGFR_cysC and risk of death was linear with a hazard ratio (HR) of 2.75 (95% CI (1.55, 4.86)) for <30 versus ≥90 mL/min/1.73m². There was a J-shaped relationship between categorical eGFR_creat and death with a nadir at 60-89 mL/min/1.73m².

Conclusion: These data suggest that use of eGFR_creat rather than eGFR_cysC may underestimate kidney function loss in type 2 diabetes, while eGFR_cysC provides a simpler and more statistically robust prediction of mortality than eGFR_creat after adjusting for other relevant variables.

Supported by: Australian NHMRC Project Grants

Disclosure: W.A. Davis: Grants; Australian NHMRC Project Grants.

129

Preliminary results from iBEAt ancillary-study: a novel classification of renal damage in diabetes

P. Pontrelli ¹, M. Rossini¹, F. Conserva¹, F. Pesce¹, T. Ardillo¹, M. Moschetta¹, F. Giorgino¹, L. Laviola¹, S. Sourbron², K.M. Gooding³, M.F. Gomez⁴, M. Kretzler⁵, L. Gesualdo¹;

¹University of Bari Aldo Moro, Bari, Italy, ²University of Sheffield, Sheffield, UK, ³University of Exeter, Exeter, UK, ⁴Lund University, Malmö, Sweden, ⁵University of Michigan, Ann Arbor, MI, USA.

Background and aims: Renal damage in diabetes includes diverse histological patterns, from true diabetic glomerulosclerosis (DN) to non-diabetic renal disease (NDRD); hence, kidney biopsy remains the gold standard for appropriate diagnosis. The iBEAt study within the BEAt-DKD project (https://www.beat-dkd.eu/; GA 115974) aims to identify and implement renal imaging biomarkers by magnetic resonance-MRI and ultrasound-US to dissect DKD heterogeneity. In this iBEAt sub-study, we aimed to explore the correlations between imaging, clinical, molecular and histopathological data of DKD patients to dissect the different phenotypes.

Materials and methods: Up to March 2023, we enrolled 75 patients to perform kidney biopsy, US and MRI, along with biofluids and clinical data collection. 64 out of 75 patients underwent kidney biopsy and three renal cores were collected (one core was formalin-fixed and paraffin embedded, the second was used for immunofluorescence and electron microscopy and the third was used for molecular analysis)

Results: According to the KDIGO guidelines our cohort included: 13 patients in stage A1 (1 G1, 1 G2, 2 G3a, 7 G3b, 2 G4); 26 patients in stage A2 (7 G1, 6 G2, 4 G3a, 7 G3b, 2 G4) and 25 patients in A3 (3 G1, 7 G3a, 7 G3b, 8 G4). According to the histological classification by Mazzucco et al (2002), 31% of DKD patients belong to class 1 (diabetic glomerulosclerosis); 33% to class 2 (vascular and ischemic glomerular changes); 4% to class 3a (glomerular diseases superimposed on DN); 31% to class 3b (other glomerulonephritis in the absence of DN). Pathologist assessment evidenced the presence of heterogeneous lesions among patients included within the same Mazzucco class. Therefore, we proposed a novel DKD classification integrating renal pathology and pathogenetic drivers. We first identified 7 different histological classes of DKD: i) I: pure DN; ii) II: DN and nephroangiosclerosis; iii) III: DN and acute tubular necrosis; iv) IV: DN and Focal Segmental Glomerulosclerosis (FSGS); v) V: Nephroangiosclerosis; vi) VI: FSGS and other GN; vii) VII: FSGS and nephroangiosclerosis. According to the pathogenic drivers of DKD, we then grouped these classes in pure metabolic damage (class I), metabolic damage and other drivers (including classes II, III, IV), pure vascular damage (class V) and immunological damage (classes VI, VII). By applying the newly proposed classification, we observed a significantly different distribution among classes of key clinical parameters, uACR (p=0.02), eGFR (p=0.012), serum creatinine (p=0.002) and proteinuria (p=0.004). PAS glomerular staining positivity confirmed the more severe glomerular damage in the metabolic classes (p=0.02); the vascular wall-to-lumen ratio was able to significantly discriminate patients with vascular vs. immunological damage (p=0.002). Finally, the renal resistive index measured through US significantly discriminated pure DN from mixed forms (p=0.02).

Conclusion: The integration of renal pathology, imaging and molecular data within the same patient has the potential to unlock new diagnostic criteria to specifically address the heterogeneity of renal phenotypes, providing a new classification of renal damage in diabetes.

Supported by: BeatDKD (grant agreement No 115974)

Disclosure: P. Pontrelli: None.

130

External validation of kidney failure risk equation in a multi-ethnic cohort of more than 7000 people with type 2 diabetes and chronic kidney disease

A. Goubar ¹, A. Mangelis¹, S. Thomas², P. Vas², J. Collins², S. Ayis¹, J. Karalliedde³;

¹King's College London, London, UK, ²King's Health Partners, London, UK, ³King's Health Partners, King's College London, London, UK.

Background and aims: The four variable Kidney failure risk equation (KFRE) which uses (age, sex, eGFR by CKD-EPI equation and urine albumin to creatinine ratio), is recommended to estimate the risk of end stage kidney disease (ESKD) and guides prompt referral of high risk people to kidney clinical services. There is limited data on the performance of KFRE in ethnically diverse cohorts of people with type 2 diabetes (T2DM). Our aim was to evaluate the performance of the KFRE in multi-ethnic cohort of people with T2DM and chronic kidney disease (CKD).

Materials and methods: We included 7,296 people with T2DM and CKD (eGFR 15 to 60 ml/min) attending routine clinical care at two large hospitals in South London between 2004 and 2018. Clinical data were collected from electronic patient records. The median follow-up was 10.2 years. The primary endpoint for KFRE's validation was time to ESKD (eGFR <15 ml/min). Ability of KFRE to discriminate individuals who experienced ESKD from those who did not was assessed using concordance index (C-index). KFRE's calibration was assessed by plotting the average predicted risk against the observed risk for each of the ten risk groups defined by tenths of predicted risk. Calibration metrics such as the median and the 90^th percentile of the absolute difference between smoothed observed and predicted risks (AD) were used.

Results: We evaluated 7,296 (30% female) with mean (standard deviation) age 65 (12) years of whom 2,955 (41%) were African-Caribbean, 3,298 (45%) Caucasian and 1,043 (14%) other ethnicity. Median (interquartile range) eGFR was 47.6 (38.2, 54.2) ml/min. Of the cohort, 1,130 (15.5%) people reached primary endpoint, with 270 (3.7%) over 2 years and 547 (7.5%) over 5 years. The discrimination of KFRE was satisfactory but not ideal for both 2 and 5 years ESKD risk prediction (C-index 0.843, 95% CI 0.839 to 0.847 for 2 and 0.801 95% CI 0.798 to 0.804 for 5 years). The KFRE 'under-predicted' the risk of ESKD in all the ten risk groups. The median (95% CI) and 90^th percentile (95% CI) of the AD were 0.022 (0.018 to 0.027) and 0.073 (0.061 to 0.084) for 2 years and 0.049 (0.40 to 0.057) and 0.138 (0.120 to 0.156) for 5 years ESKD risk prediction. Similar 'under performance' of KFRE was observed in all ethnic groups.

Conclusion: KFRE demonstrated satisfactory but not ideal prediction of ESKD in an ethnically diverse cohort of people with T2DM and CKD. We could not identify unique feature of the cohort that explained lower performance of the KFRE. Further studies are needed to identify variables/biomarkers that can improve the performance of KFRE.

Supported by: Guy’s and St Thomas’ Charity

Disclosure: A. Goubar: None.

131

Utility of molecular markers in urine to differentiate diabetic kidney disease from non-diabetic kidney disease and possible role in pathogenesis

S. Ghosh, M. Basu, A. Raychaudhury, N. Bhattacharyya, P. Mukhopadhyay;

IPGME & R CALCUTTA, Kolkata, India.

Background and aims: Renal involvement in Type 2 diabetes can be due to diabetes per se (Diabetic Kidney Disease, DKD) or other than diabetes (Non-diabetic Kidney Disease, NDKD). Renal biopsy remains the gold standard for diagnosis. None of the currently available clinical parameters are good to preselect individuals for renal biopsy. In animal models with diabetes and kidney disease, several gene alterations have been reported. Long non-coding RNAs are emerging as regulators of diverse molecular functions and contributors of human disease. Limited data available in animal models suggests possible alteration of lncRNA, which may contribute to the pathogenesis of DKD by modulation of protein-coding genes. This has not been explored in humans with biopsy-proven kidney disease in T2 DM. With this background, we aimed to determine the expression of long noncoding RNA, proteins, and protein-coding genes from urine to evaluate if this could help differentiate DKD from NDKD. To determine the possible relation of that selected lncRNA in the regulation of the identified proteins.

Materials and methods: We recruited patients with renal involvement (eGFR >30ml/min/m² to <60ml/min/m² and/or ACR>300mg/mcg who underwent renal biopsy and histopathologically classified (ISN/RPS). We also included subjects with T2DM without renal involvement and healthy controls. 5ml of second-morning urine sample in fasting state was collected. The expression of selected proteins was determined by the Enzyme-linked immunosorbent assay (ELISA) method (with creatinine normalization). Expression of several lncRNAs was determined in urinary cells and urinary exosomes. Differences in expression between groups were determined by quantitative real-time PCR with respect to internal control. Mann-Whitney U and Receiver operating characteristic (ROC) curves) were performed. The correlation between lncRNA and mRNA/protein was determined and further validated in the cell culture model.

Results: Amongst 110 subjects who underwent biopsy, 73 (66.4%) had DKD; 20 (18.2 %) had NDKD; and 17 (15.4 %) had mixed kidney disease. Expression of MALAT1, PVT1, NEAT, and HOTAIR levels was higher in the urinary cells of subjects with DKD, this expression was further validated in exosomes. In exosomes, PVT1 and MALAT1 had the best discriminatory ability to differentiate DKD from NDKD (p<0.01), (p<0.05) with 90% sensitivity and specificity of 82 %. Expression of different tubular glomerular markers was identified. Urinary NGAL (expression of protein level as well expression at gene level) was able to differentiate DKD from NDKD (P<0.001, P<0.01) which in turn was positively correlated with expression of lncRNAs (MALAT1 & PVT1) (P<0.05). Expression of Epithelial-mesenchymal transition (EMT) markers and markers of fibrosis of which expression of TGF β was different amongst the groups and it showed a positive correlation with MALAT1(r=0.767, p<0.001) and PVT1(r=0.670, p<0.01). We further validated our findings in the cell culture model.

Conclusion: Long-noncoding RNA expression along with its associated genes and protein expression from urine may help differentiate DKD from NDKD

Supported by: RSSDI, ICMR, West Bengal Science and technology, RSSDI West Bengal

Disclosure: S. Ghosh: None.

132

Nox5: a promising therapeutic target for diabetic kidney disease

J.C. Jha ¹, A. Dai¹, H.R. Kankanamalage¹, J. Meister², J. Vincent³, M.E. Cooper¹, K. Jandeleit-Dahm^1,2;

¹Department of Diabetes, Central Clinical School, Monash University, Melbourne, Australia, ²Institute for Clinical Diabetology, German Diabetes Centre, Leibniz Centre for Diabetes Research at Heinrich Heine University, Dusseldorf, Germany, ³Universite de Geneve, Geneva, Switzerland.

Background and aims: Diabetes is the leading cause of end stage kidney failure. Oxidative stress due to excessive production of reactive oxygen species plays a critical role in diabetic kidney disease (DKD). Pro-oxidant enzyme NADPH oxidase-NOX5 is considered as a major contributor of ROS and thereby aggravating renal injury in DKD. We aim to identify the pathogenic role of NOX5 and associated ROS-sensitive pathways in DKD using various experimental models including human kidney biopsies, human renal organoids and cells, humanised NOX5 transgenic mice and rabbit models as well as novel NOX5-specific inhibitors.

Materials and methods: We examined the ROS production as well as expression of NOX5 in association with ROS-sensitive factors including a protein kinase, PKC-α, a transcription factor, EGR1 (early growth response 1) and a key metabolic gene involved in redox balance, TXNIP (thioredoxin-interacting protein) in human kidney biopsies, human renal organoids and cells. We assessed the effect of NOX5 inhibition using genetic manipulation and pharmacological inhibition approaches in human renal cells and organoids exposed to diabetic milieu environment. In vivo, we also assessed the effect of endothelial cell specific -Nox5 overexpression independent of NOX4 in humanised Nox5 transgenic mice in the presence or absence of streptozotocin induced diabetes.

Results: We identified increased expression of renal NOX5 in diabetic patients and in high fat fed rabbits. We also found a positive association of renal NOX5 with upregulation of EGR-1, PKC-α and TXNIP. In vivo, overexpression of human NOX5 independent of NOX4 pathways demonstrated an increase in albuminuria, glomerulosclerosis, renal fibrosis and inflammation in association with upregulation of EGR-1, ERK1/2, PKC-α, PKC-ε and TXNIP via enhanced ROS production in comparison to diabetic mice not expressing Nox5. In vitro, pharmacological inhibition of Nox5 attenuated high glucose induced gene expression of markers of fibrosis and inflammation as well as downregulation of EGR-1, PKC-α and TXNIP.

Conclusion: The findings of this study suggest that NOX5 plays a key role in human DKD, thereby providing the impetus for the fast track validation of NOX5 specific inhibitors to improve kidney health in diabetes.

Supported by: NHMRC

Disclosure: J.C. Jha: None.

OP 23 More news from SGLT2s

133

SGLT2 inhibitor dapagliflozin enhances skeletal muscle fatty acid uptake in subjects with type 2 diabetes: a positron emission tomography study

A. Latva-Rasku ¹, J. Tuisku¹, A. Bhowmik¹, H. Keskinen¹, E. Rebelos¹, S. Laurila¹, J. Koffert¹, L. Nummenmaa¹, K. Heurling², J. Oscarsson³, P. Nuutila¹;

¹Turku PET Centre, University of Turku, Turku, Finland, Turku, Finland, ²Antaros Medical, Mölndal, Sweden, ³AstraZeneca AB, Mölndal, Sweden.

Background and aims: Sodium-glucose cotransporter 2 inhibitors (SGLT2i) have been shown to alter tissue metabolism in humans and animal models, with fatty acid and ketone bodies thought to become preferred fuels over glucose. The aim of this study was to investigate the effects of dapagliflozin on tissue fatty acid metabolism in subjects with type 2 diabetes (T2D) by using positron emission tomography (PET).

Materials and methods: In this randomized, double-blinded placebo-controlled trial, 53 subjects with inadequately controlled T2D received either dapagliflozin 10 mg or placebo daily for 6 weeks. Tissue fatty acid uptake was assessed at baseline and at the end of treatment in skeletal muscle, the brain and small intestine with PET by using [¹⁸F]-6-thia-heptadecanoic acid ([¹⁸F]-FTHA) tracer. The scans were performed in the afternoon after a minimum of 6 hours of fasting and 6 hours after taking the study drug. Changes in treatment effects were analysed using ANCOVA, with treatment group as the independent variable and baseline values as covariates, and the results are reported as least square means (95 % confidence interval for the difference between groups). For the brain, a similar statistic was used with statistical parametric mapping.

Results: A total of 40 subjects (dapagliflozin N=21, placebo N=19) completed the [¹⁸F]-FTHA studies. 55 % were males, and they had a mean (SD) age of 64.8 (6.9) years, BMI of 29.9 (3.6) kg/m² and HbA1c of 6.6 (0.6) % with no significant differences between groups. At 6 weeks, HbA1c and BMI were decreased in the dapagliflozin group (-0.2 [-0.4, -0.04] %, p=0.016, and -0.5 [-0.7, -0.2] kg/m², p <0.001).

Both skeletal muscle and brain fatty acid uptake increased (+1.0 [0.07, 2.0] micromol/100 g/min, p=0.036 in skeletal muscle, see Figure 1 for the brain, global mean increase 32 %). In the small intestine the minor increase in uptake was not statistically significant (+0.5 [-0.08, 1.2], p=0.08). Level of serum free fatty acids during the scans was not significantly affected by the treatment (p=0.27), whereas serum hydroxybutyrate level increased (+0.04 [0.01, 0.08] micromol/L, p=0.023] and fasting plasma glucose decreased (-0.62 [-1.1, -0.1] mmol/L, p=0.016) in the dapagliflozin group compared to placebo.

Conclusion: In this study, we showed by using direct assessment of tissue metabolism by PET that 6 weeks of treatment with dapagliflozin increases muscle and brain fatty acid uptake. This is in line with previous hypotheses and indirect measurements showing enhanced fatty acid metabolism in response to SGLT2i treatment, and demonstrate how widespread the shift from glucose to fatty acid utilization is.

Clinical Trial Registration Number: NCT03387683

Supported by: The study was funded by AstraZeneca

Disclosure: A. Latva-Rasku: None.

134

Anti-inflammatory and antioxidant effects of SGLT-2 inhibitors in epicardial adipose tissue of subjects with severe heart failure

M. Mraz ¹, B.J. Kasperova², S. Stemberkova-Hubackova², P. Svoboda², D. Hlavacek³, J. Mahrik⁴, O. Kuda⁵, T. Cajka⁵, I. Pleyerova², K. Reznickova², P. Novodvorsky¹, P. Ivak³, V. Melenovsky⁶, I. Netuka³, M. Haluzik¹;

¹Department of Diabetes, Institute for Clinical and Experimental Medicine, Prague, Czech Republic, ²Experimental Medicine Centre, Institute for Clinical and Experimental Medicine, Prague, Czech Republic, ³Department of Cardiovascular Surgery, Institute for Clinical and Experimental Medicine, Prague, Czech Republic, ⁴Department of Anesthesia and Resuscitation, Institute for Clinical and Experimental Medicine, Prague, Czech Republic, ⁵Institute of Physiology, Czech Academy of Sciences, Prague, Czech Republic, ⁶Department of Cardiology, Institute for Clinical and Experimental Medicine, Prague, Czech Republic.

Background and aims: The exact mechanisms behind cardioprotective effects of SGLT-2 inhibitors (SGLT-2i) are still not fully understood. Using complex transcriptomic, metabolomic and immunochemical methods we analysed epicardial (EAT) and subcutaneous (SAT) adipose tissue of patients with severe heart failure treated with SGLT-2i in order to identify possible cardioprotective pathways.

Materials and methods: Twenty six subjects with severe heart failure with reduced ejection fraction (NYHA III-IV) treated with SGLT-2i (18 with type 2 diabetes mellitus - T2DM) and 26 age-, BMI- and left ventricular ejection fraction-matched control subjects (8 with T2DM) scheduled for heart transplantation or mechanical support implantation were included into the study. The complex analysis of EAT and SAT included liquid chromatography in combination with mass spectrometry for metabolomics, RT-qPCR for transcriptomics and flow cytometry and immunochemistry for detection of immune cells and mitochondria.

Results: Compared with controls, EAT of SGLT-2i subjects showed marked reduction in macrophage infiltration along with a decrease in Th1 and Th2 cells and mRNA levels of macrophage marker ADGRE1. This was not the case for SAT, where only reduced mRNA levels of inflammatory markers including IL6 and MCP1 were detected. Moreover, mitochondrial staining revealed improvement of mitochondria morphology in EAT of SGLT-2i group accompanied by decreased oxidative stress. Finally, enhanced enrichment of ether lipids with oleic acid noted on metabolomic analysis suggests reduced disposition to ferroptosis (iron-dependent lipid peroxidation-regulated cell death) further contributing to decreased oxidative stress in EAT of SGLT-2i subjects.

Conclusion: SGLT-2i treatment is associated with reduced inflammation and oxidative stress in epicardial adipose tissue, which could at least partially explain their cardioprotective effects.

Supported by: CarDia (Programme EXCELES, Project No. LX22NPO5104) - Funded by the European Union - Next Generation EU; NV19-02-00118; IKEM, IN 00023001

Disclosure: M. Mraz: None.

135

Mapping the metabolic reprogramming induced by sodium-glucose cotransporter 2 inhibition

G. Leibowitz ¹, Y. Riahi¹, O. Mosenzon¹, I. Abramovich², B. Agranovich², L. Hinden³, L. Kadosh¹, R. Ben-Haroush Schyr⁴, D. Kleiman⁴, E. Cerasi¹, D. Ben-Zvi⁴, E. Gottlieb², J. Tam³, A. Kogot-Levin¹;

¹Diabetes Unit and Endocrine Service, Hadassah-Hebrew University Medical Center, Jerusalem, Israel, ²The Laboratory for Metabolism in Health and Disease, Technion-Israel Institute of Technology, Haifa, Israel, ³The Institute for Drug Research, School of Pharmacy, Faculty of Medicine, The Hebrew University of Jerusalem, Jerusalem, Israel, ⁴Developmental Biology and Cancer Research, Hadassah-Hebrew University Medical School, Jerusalem, Israel.

Background and aims: SGLT2 inhibitors (SGLT2i) reduce the risk for kidney disease, congestive heart failure and mortality in patients with and without diabetes. We have previously shown that the beneficial effects of SGLT2i are mediated via inhibition of the nutrient sensor mTORC1 in renal proximal tubular cells (RPTCs). We hypothesize that RPTCs function as a metabolic hub, regulating nutrient(s) metabolism in various organs. Here, we aimed to decipher the metabolic alterations that occur in the kidney, liver, and heart in diabetes and in response to SGLT2 inhibition, which prevent glucose reabsorption in RPTCs.

Materials and methods: We performed in vivo metabolic labeling with ¹³C-glucose in normoglycemic wild-type and diabetic Akita mice treated with or without the SGLT2 inhibitor dapagliflozin (10 mg/kg/day) or degludec insulin (1-4 units/day) for one week, followed by simultaneous metabolomics and metabolic flux analyses in the kidney cortex, liver, heart, and the plasma. mTORC1 and AMPK activities were analyzed by Western blotting.

Results: We found that in diabetes, glycolysis and glucose oxidation are impaired in the kidney, liver, and heart, evident by decreased levels of ¹³C-labeled pyruvate, lactate and TCA cycle metabolites. Treatment with dapagliflozin failed to rescue glycolysis and further inhibited pyruvate kinase activity in the liver. On the contrary, SGLT2 inhibition increased glucose oxidation in all organs. In addition, dapagliflozin increased plasma and liver ketone body β-hydroxybutyrate levels. Diabetes was associated with altered methionine cycle metabolism, evident by decreased betaine and methionine levels, whereas treatment with SGLT2i increased hepatic betaine along with decreased homocysteine levels. The activity of mTORC1 was inhibited by SGLT2i along with stimulation of AMPK in both normoglycemic and diabetic animals. In contrast to SGLT2i, treatment with insulin enhanced glycolysis, glucose-derived amino acid synthesis and inhibited ketogenesis.

Conclusion: SGLT2i induce metabolic reprogramming orchestrated by the key nutrient-sensing pathways AMPK and mTORC1, leading to enhanced energy production through increased glucose and fatty-acid oxidation, whereas insulin promote energy-consuming anabolic pathways. In addition, SGLT2i prevent dysregulated methionine cycling in diabetes. These effects may explain the beneficial effects of SGLT2i with important implications for diabetes and aging.

Supported by: EFSD/Novo Nordisk Programme for Diabetes Research in Europe, ISF

Disclosure: G. Leibowitz: None.

136

Tolerability and effectiveness of gliclazide MR-SGLT2i combination in type 2 diabetes people fasting during Ramadan: a subgroup analysis of the real-world DIA-RAMADAN study

M. Hassanein ¹, A. Durocher², V. Cortese²;

¹Dubai Hospital, Dubai, United Arab Emirates, ²Global Medical and Patient Affairs, Servier, Suresnes, France.

Background and aims: DIA-RAMADAN study was a prospective, observational, international study that explored in real-world settings safety and effectiveness of gliclazide MR in people with Type 2 Diabetes (T2D) fasting during Ramadan. The present sub-analysis was conducted in the subgroup of patients treated with gliclazide MR combined with a SGLT2i, with or without other oral antidiabetic drugs (OADs).

Materials and methods: The DIA-RAMADAN study enrolled patients across 9 countries in culturally diverse regions: India/Pakistan/Bangladesh, Egypt/Kuwait/UAE/Saudi Arabia, Indonesia/Malaysia. Data were collected during Ramadan 2019. Main eligibility criteria were adult T2D patients treated at stable doses with gliclazide MR for >90 days prior to enrolment. Patients requiring insulin and with HbA1c ≥9% were excluded. Patients were examined at inclusion (V0) 6-8 weeks before Ramadan. Patients remained on Gliclazide MR once daily for 14-18 weeks. A second visit (V1) was conducted 4-6 weeks after Ramadan. Main study outcomes were the proportion of patients reporting ≥1 confirmed hypoglycemia event (HE), changes in HbA1c and body weight. Changes pre- to post-Ramadan were analyzed using paired t-tests or Wilcoxon tests. Adherence (%) was calculated as (sum of number of intakes taken during Ramadan/sum of number of intakes to be taken during Ramadan) x 100.

Results: Of 1214 patients included, 102 received a gliclazide MR-SLGT2i based therapy. These patients were in average younger (mean ± SD: 51.4 ± 10.4 years vs 54.3 ± 10.4 years, respectively), with longer disease duration (5.8 ± 4.5 years vs 5.4 ± 5.7 years), higher HbA1c (7.7 ± 1.0% vs 7.5 ± 0.9%) and had more comorbidities (hypertension 58.8% vs 33.7%, dyslipidemia 51.0% vs 26.8%) compared to those not receiving the combination. Adherence to the combination therapy was ≥80% in 95.1% of patients. Most patients were treated with ≥3 OADs (92.2%). 4/102(3.9%) patients reported ≥1 confirmed HE and no severe HEs were reported. No patient broke his fast due to hypoglycemia. A significant reduction in HbA1c (-0.6 ± 1.0 %; p<0.001) and body weight (-0.6 ± 5.9 kg; p<0.001) was observed between V0 and V1.

Conclusion: This sub-analysis of the DIA-RAMADAN study suggests that the combination of gliclazide MR with a SGLT2i is a well tolerated and effective treatment option for patients with Type 2 Diabetes who are fasting during Ramadan. The study found a low incidence of hypoglycemic events and significant reductions in HbA1c and body weight. These findings in real-life settings provide important insights into the management of T2D during Ramadan and could inform clinical practice in this population.

Clinical Trial Registration Number: NCT04132934

Disclosure: M. Hassanein: Employment/Consultancy; Servier consultant.

137

Risk of diabetic microvascular complications, heart failure, hospitalisation and all-cause mortality with SGLT2i and GLP1-ra in type 2 diabetes: a real-world data study

A. Eleftheriadou ¹, D. Riley¹, P. Austin², G. Hernandez², U. Alam¹;

¹Department of Cardiovascular & Metabolic Medicine, University of Liverpool, Liverpool, UK, ²TriNetX LLC, Cambridge, MA, USA.

Background and aims: To assess the relationship of sodium-glucose cotransporter-2 inhibitors (SGLT2i) and glucagon-like peptide-1 receptor analogues (GLP1-ra) on 5-year risk of diabetic neuropathy, autonomic neuropathy, diabetic retinopathy and macular oedema, hospitalisation and all-cause mortality in type 2 diabetes.

Materials and methods: This is a retrospective cohort analysis of six million people with type 2 diabetes across 85 healthcare organisations using a global federated health research network (TriNetX, Boston, USA). Two intervention cohorts (SGLT2i, n=126,171 ; GLP1-ra, n=164,024 ) were compared against a control cohort (no SGLT2i/ GLP1-ra, n=1,665,412 ). Propensity score matching for age, sex, ischaemic heart disease, microvascular complications, HbA1c <7% and ³7%, and use of pioglitazone was used to balance cohorts 1:1. A sub-analysis comparing the two intervention cohorts was also performed.

Results: Intervention cohorts (both SGLT2i and GLP1-ra) demonstrated a reduced relative risk (RR, 95% confidence interval [95%CI]) compared to the control cohort at 5 years for macular oedema (SGLT2i: 0.55, 0.51-0.59. GLP1-ra: 0.80, 0.76-0.84), heart failure (SGLT2i: 0.55, 0.54-0.57. GLP1-ra: 0.63, 0.61-0.65), hospitalisation (SGLT2i: 0.64, 0.64-0.65. GLP1-ra: 0.65, 0.65-0.66) and all-cause mortality (SGLT1i: 0.38, 0.37-0.40. GLP1-ra: 0.39, 0.38- 0.40). Contrary to the GLP1-ra cohort, the SGLT2i cohort demonstrated a reduced risk compared to the control cohort at 5 years for diabetic neuropathy (SGLT2i: 0.86, 0.84-0.89. GLP1-ra: 1.19, 1.16-1.22), autonomic neuropathy (SGLT2i: 0.83, 0.76-0.92. GLP1-ra: 1.08, 1.00-1.17) and diabetic retinopathy (SGLT2i: 0.71, 0.67-0.74. GLP1-ra: 1.06, 1.02-1.10). When directly comparing to SGLT2i, the GLP1-ra cohort demonstrated an increased RR at 5 years in neuropathy (1.42, 1.39-1.47), autonomic neuropathy (1.28, 1.17-1.40), diabetic retinopathy (1.48, 1.41-1.55), macular oedema (1.49, 1.38-1.60), heart failure (1.16, 1.12-1.20), hospitalisation (1.04, 1.03-1.06) and all-cause mortality (1.09, 1.06-1.13).

Conclusion: SGLT2i and GLP1-ra both reduce the risk for macular oedema, heart failure, hospitalisation and all-cause mortality in people with type 2 diabetes over 5 years. Additionally, SGLT2i therapy reduces the risk for further microvascular complications including neuropathy, autonomic neuropathy and retinopathy. SGLT2i therapy was associated with the greatest risk reduction in diabetic microvascular complications as well as heart failure, hospitalisation and all-cause mortality. Future randomised controlled trials of SGLT2i and GLP1-ra should incorporate sensitive surrogate biomarkers of diabetic microvascular disease to validate these findings and if validated, in particular SGLT2i should be considered higher in the treatment algorithm for the general patient with type 2 diabetes.

Supported by: TriNetX LLC

Disclosure: A. Eleftheriadou: None.

138

Socioeconomic dispartities in the use of SGLT-2 inhibitors and GLP-1 receptor antagonists in adults with type 2 diabetes in Germany

M. Auzanneau ^1,2, J. Seufert³, S. Zimny⁴, T. Haak⁵, A. Zeyfang⁶, I. Hugenberg⁷, M. Pavel⁸, J. Rosenbauer⁹, R.W. Holl^1,2;

¹Ulm University, Ulm, Germany, ²German Center for Diabetes Research (DZD), Munich-Neuherberg, Germany, ³Division of Endocrinology and Diabetology, Department of Medicine II, Medical Center – University of Freiburg, Faculty of Medicine, University of Freiburg, Freiburg, Germany, ⁴Department of General Internal Medicine, Endocrinology and Diabetes, Helios Klinik Schwerin, Schwerin, Germany, ⁵Diabetes Center Mergentheim, Bad Mergentheim, Germany, ⁶Department of Internal Medicine, Geriatric Medicine, Palliative Medicine and Diabetology, medius Klinik Ostfildern-Ruit and Nürtingen, Ostfildern, Germany, ⁷RoMed Klinik Prien am Chiemsee, Prien am Chiemsee, Germany, ⁸Department of Medicine 1, Friedrich-Alexander University Erlangen-Nürnberg, Erlangen, Germany, ⁹Institute for Biometrics and Epidemiology, German Diabetes Center (DDZ), Leibniz Center for Diabetes Research at Heinrich Heine University Düsseldorf, Düsseldorf, Germany.

Background and aims: SGLT-2 inhibitors (SGLT2i) and GLP-1 receptor agonists (GLP1-RA) are broadly recommended for type 2 diabetes (T2D) treatment, but real-world usage is variable. We aimed to examine if demographic and regional socioeconomic factors are associated with their usage in a large population of adults with T2D in Germany.

Materials and methods: All visits of adults with T2D documented in 2021 in the nationwide Diabetes Prospective Follow-up (DPV) registry were included. Regional deprivation was assessed using the German Index of Socioeconomic Deprivation categorized into quintiles, from Q1 (least deprived districts) to Q5 (most deprived districts). We applied logistic regression models adjusted for gender, age, and diabetes duration to investigate the association between regional deprivation and SGLT2i and GLP-1RA usage in diabetes treatment.

Results: The study population comprised 30,016 adults with T2D (56% males; median age: 69.5 years [59.5-78.9]; median diabetes duration: 10.7 years [4.6-18.9]). Men used SGLT2i more frequently than women (24.5% vs. 16.2%, p<0.001), whereas GLP1-RA use did not differ between men and women (14.7% vs. 14.5%, p= 0.05). The use of both SGLT2i and GLP1-RA was highest in the 40-<60 years age group and lowest in the ≥ 80 years age group (SGLT2i: 27.7% vs. 9.6%, p<0.001; GLP1-RA: 22.1% vs. 4.3%, p<0.001). With increasing deprivation from Q1 to Q5, the use of SGLT2i significantly decreased from 26.0% [24.7-27.2] to 16.3% [15.4-17.2] (p<0.001) and the use of GLP1-RA from 17.1% [16.0-18.2] to 12.6% [11.8-13.4] (p<0.001). In contrast, the proportion of persons with T2D treated with insulin monotherapy increased between Q1 and Q5 from 8.3% [7.5-9.1] to 14.7% [13.8-15.6] (p<0.001).

Conclusion: Data from real-world care are essential to examine evidence-based access to new treatment options. Our data indicate that the use of SGLT2i and GLP1-RA in adults with T2D is associated with socioeconomic deprivation in Germany. Impaired access to these new cardiorenal protective glucose-lowering medications may increase the risk for renal and cardiovascular morbidity and mortality in socially deprived persons with T2D. Further research is necessary to investigate more precisely which socioeconomic barriers may limit this access to novel treatment options in Germany, a country with believed universal and comprehensive health care coverage.

Supported by: DZD

Disclosure: M. Auzanneau: None.

OP 24 All around pregnancy

139

Prediabetes and diabetes 11 years after pregnancy in ethnic minority groups with and without gestational diabetes

C. Waage ¹, I. Mdala¹, A. Brænd¹, A.K. Jenum¹, K.I. Birkeland^2,3;

¹Department of General Practice, University of Oslo, Oslo, Norway, ²Department of Transplantation, University of Oslo, Oslo, Norway, ³Department of Endocrinology, Obesity and Preventive Medicine, Oslo University Hospital, Oslo, Norway.

Background and aims: Women of South Asian origin living in Western countries have higher risk of developing diabetes compared to Europeans. We conducted a community-based study with unselected OGTT-screening for gestational diabetes (GDM) in three multi-ethnic districts of Oslo during 2008-2011 and included 823 women. We invited women still living in Oslo and nearby counties for re-examination in 2019-2022 to determine the prevalence of prediabetes (HbA1c 39-47 mmol/mol) and diabetes.

Materials and methods: Data were collected at four visits at (mean) gestational week 15 and 28, 14 weeks after birth (V1-V3) and at 11 years after birth (V4). At V4, we collected data from interviews, clinical examinations including anthropometrics, blood pressure measurements, objectively measured physical activity and blood samples. We assessed ethnic differences in dysglycemia using the Pearson Chi-square, and one-way analysis of variance (ANOVA) with Bonferroni correction for mean differences in HbA1c, weight gain and caliper-measured subcutaneous fat. We performed supervised machine learning using the least absolute shrinkage and selection operator (LASSO) logistic regression model to select predictors and create a parsimonious model.

Results: At V4 we examined 206 Europeans (mean±SD age 41.3±4.8 years, BMI 26.2±5.2), 86 South Asians (age 40.1±4.3 years, BMI 27.5±4.3) and 68 women of other ethnic minority origin (East Asia, Middle East and Africa) (age 41.7±4.7 years, BMI 28.1±5.2). The overall prevalence of dysclycemia was significantly higher in South Asians (43%) and other ethnic groups (34%) compared with Europeans (17%), P<0.001. For women with previous GDM, the prevalence of dysglycemia was significantly higher in South Asians (43.2%) compared to Europeans (22.2%), P=0.04. Mean weight gain from prepregnancy to V4 was significantly higher in South Asians (8.7±6.5 kg) compared to Europeans (5.3±7.4 kg), P=0.002. Women from South Asia and other minorities had a higher increase in in sub-scapular fat from V1-V4 (7.9 ±8.9 mm) compared to Europeans (3.1±7.4 mm), P<0.001. Preliminary results from the LASSO logistic regression model, indicate that the strongest predictors of prediabetes/diabetes after 11 years (ordered by predictive power) were ethnicity, HBA1c at V1, prepregnancy BMI and age. Our preliminary prediction model had an area under the ROC-curve of 0.7741 (95% CI: 0.7196, 0.8286).

Conclusion: We found that 34-43% of 41-year old women with ethnic minority background had developed prediabetes or diabetes 11 years after the index pregnancy. The most important predictor variables were South Asian ethnicity and HbA1c measured early in pregnancy.

Clinical Trial Registration Number: 2018/2517

Supported by: DAM

Disclosure: C. Waage: Grants; DAM foundation Norway.

140

Metabolites of diacylglyceride and unsaturated fatty acid synthesis improves prediction of gestational diabetes

E.C. Francis, R.K. Johnson, C.C. Cohen, D. Dabelea, W. Perng;

University of Colorado, Aurora, CO, USA.

Background and aims: Gestational Diabetes (GDM) is an increasingly prevalent condition that impacts maternal-child health across their lifecourse. Early prediction of GDM is an important research area with potentially substantial impact on intervention timing. Since GDM is a metabolically heterogeneous condition, untargeted metabolomics holds potential to improve this prediction. Prior studies have focused mainly on individual metabolites. However, the ratios of two metabolites may point to disruptions in pathways that are more characteristic of a phenotype than individual metabolites alone. Our objective was to compare fasting untargeted metabolomic ratios and individual levels to fasting glucose in the prediction of GDM.

Materials and methods: We used data from 989 pregnant women in the Healthy Start cohort. GDM was diagnosed at 24-26 weeks using the Carpenter Coustan criteria and abstracted from medical records. Untargeted metabolomics were measured with liquid chromatography mass spectrometry from fasting blood collected at approximately 17 weeks. We used principal component analysis of 2,574 metabolomic features and the elbow method to determine the important components (n=11). We selected the top 20 highest loading features from each component for further analysis (n=220 metabolomic features). We preformed enrichment analysis of the 220 metabolomic features using the original 2,574 as a reference to identify classes of metabolites that were over-represented. From the 220 metabolomic features, we calculated ratios of all possible metabolite pairs and chose the top 20 ranked ratios (based on P-values) to be used with the non-ratio features in prediction of GDM. We used a random forest algorithm to classify and rank features and evaluated feature importance using Monte-Carlo cross validation with a false-discovery rate (FDR) P-value of 0.05. We compared the area under the curve (AUC) from a model predicting GDM based on fasting glucose alone to a model that included fasting glucose and metabolomic features using leave-one-out cross-validation.

Results: The 220 selected metabolomic features were enriched in pathways of diacylglyceride metabolism and unsaturated fatty acid synthesis. A prediction model of GDM that included only fasting glucose resulted in an AUC of 0.74 (95% CI: 0.66, 0.81). Eight ratios and two individual metabolomic features were selected as the variables with high importance for GDM prediction (FDR P-value<0.05). A model that included fasting glucose and the selected metabolomic variables resulted in an AUC of 0.83 (0.77, 0.89), which was a significant improvement over a model of fasting glucose alone (P-value=0.002). Most metabolite ratios included metabolites of phosphatidylcholines (14:0/14:1-22:0/22:1), phosphatidylethanolamine (16:0/16:1 - 22:0/22:1), very-long chain acylcarnitines, and a-tocopherol phosphoric acid. A metabolite without annotation with a retention time and mass of 15.73 545.3400m/z was selected in three of the ratios, and as an individually important feature.

Conclusion: A select set of metabolomic features that were mostly from diacylglyceride and unsaturated fatty acid synthesis measured in mid-pregnancy improved the prediction of GDM diagnosed clinically in late pregnancy, compared with fasting glucose alone. Further study of untargeted metabolomics, especially among women with GDM, may identify different processes by which the maternal metabolome in pregnancy impacts maternal-child health.

Supported by: NIH: K99HD108272, R25HL146166, R01DK076648, UH3OD023248

Disclosure: E.C. Francis: None.

141

Continuous glucose monitoring-derived glycaemic metrics and adverse pregnancy outcomes among women with gestational diabetes

J. Zheng;

Westlake University, Hangzhou, China.

Background and aims: To investigate associations of continuous glucose monitoring (CGM)-derived metrics during pregnancy with pregnancy outcomes among participants with gestational diabetes mellitus (GDM).

Materials and methods: We included 1302 pregnant women (mean: 26.0 gestational weeks) with GDM. These participants received a 14-day CGM measurement at recruitment and were followed until delivery. The primary outcome was any adverse pregnancy outcome, defined as having ≥1 of the outcomes: preterm birth, small- or large-for-gestational-age (SGA or LGA) birth, fetal distress, premature rupture of membranes and primary cesarean delivery. The secondary outcomes were above individual outcomes. We performed multivariable logistic regression to evaluate associations of CGM-derived metrics with these outcomes.

Results: Per 1-SD difference in time above range (TAR), glucose area under the curve (AUC), nighttime and daytime mean blood glucose (MBG) was associated with higher risk of any adverse pregnancy outcome, with odds ratio: 1.24 (95%CI 1.09, 1.41), 1.21 (95%CI 1.08, 1.36), 1.17 (95%CI 1.04, 1.31) and 1.19 (95%CI 1.06, 1.34), respectively. The time in range TAR, AUC, nighttime and daytime MBG and mean amplitude of glucose excursions were positively associated, while time blow range were inversely associated with risk of LGA.

Conclusion: The CGM-derived metrics during pregnancy are associated with adverse pregnancy outcomes. These CGM biomarkers could serve as potential new intervention targets to maintain a healthy pregnancy status among patients with GDM.

Clinical Trial Registration Number: NCT04060056

Supported by: This study was funded by the National Key R&D Program of China (2022YFA1303900)

Disclosure: J. Zheng: None.

142

Methylation risk score for gestational diabetes on DNA methylation patterns associated with type 2 diabetes

T. Linares-Pineda ¹, N. Fragoso Bargas², M. Picón¹, M. Molina-Vega¹, A.K. Jenum³, L. Sletner⁴, S. Lee-Ødegård⁵, J.O. Opsahl⁵, G.-H. Moen⁶, E. Qvigstad⁷, R.B. Prasad⁸, K.I. Birkeland², S. Morcillo¹, C. Sommer⁹;

¹IBIMA, Málaga, Spain, ²Department of Endocrinology, Morbid Obesity & Preventive Medicine, Oslo University Hospital, Oslo, Norway, ³Department of General Medicine, Institute of Health and Society, University of Oslo, Oslo, Norway, ⁴Department of Pediatric and Adolescents Medicine, Akershus University Hospital, Lørenskog, Norway, ⁵Institute of Clinical Medicine, University of Oslo, Oslo, Norway, ⁶Institute of Clinical Medicine, University of Oslo, Oslo, Norway, ⁷Department of Endocrinology, Morbid Obesity and Preventive Medicine, Oslo University Hospital, Oslo, Norway, ⁸Lund University Diabetes Centre, Malmö, Sweden, ⁹Department of Endocrinology, Morbid Obesity and Preventive Medicine, Oslo, Norway.

Background and aims: Gestational diabetes mellitus (GDM) and type 2 diabetes mellitus (T2DM) share many pathophysiological factors, but the exact underlying mechanisms are largely unknown. Genetic predisposition for T2DM only explains a small percentage of T2DM occurrence or heritability, suggesting a role for environmental factors where epigenetics may play an important role. Whether T2DM and GDM share DNA methylation marks has not yet been tested. The aim was to test whether a Methylation Risk Socre (MRS) for T2DM is associated with increased risk for GDM across ancestry and GDM criteria

Materials and methods: MRS was constructed from 42 CpGs associated with T2DM in a meta-analysis of 5 European cohorts. The B-value from those CpGs were extracted from two cohorts of pregnant women with GDM and non-GDM from different ethnic origins and diagnostic criteria.

Results: EPIPREG and EPIDG cohorts are two pregnancy cohorts that have been studied to investigate DNA methylation and gestational diabetes mellitus (GDM). EPIPREG includes 480 healthy women of different ethnic origins (European and South Asians) who attended Child Health Clinics in Oslo, Norway during 2008-2010. EPIDG, includes 230 Mediterranean pregnant women attending the unit of diabetes and pregnancy at the University Hospital Virgen de la Victoria, Málaga, Spain. DNA methylation in both cohorts were quantified with Infinium MethylationEPIC BeadChip. R studio v.4.0.4 were used to perform all the statistical analysis. The study analysed the DNA methylation of the cohorts at gestational week 28+-2. With a logistic mixed model regression we elucidate that MRS were associated with GDM in our three groups, ancestry was treated as random intercept. The unadjusted and adjusted model by age, pre-pregnant BMI, family history of diabetes and smoking status were associated with GDM risk: O.R:1.3; 95%C.I: 1.14-1.79, P.val= 0.002 and OR: 1.4, 95%C.I: 1.10-1.74, P.val= 0.00014 respectively. Also, linear mixed model was performed in continuous clinical variables. The unadjusted linear mix model MRS was associated with C-peptide, fasting glucose and pre-pregnant BMI. In the adjusted linear mix model, only glucose was associated with MRS (P.val=0.0086).

Conclusion: For the first time, we have tested a novel MRS for T2DM that is significantly associated with GDM, suggesting shared epigenetic mechanisms between GDM and T2DM. This may help explain the increased risk of T2DM after GDM. Furthermore, understanding which factors that modify methylation of these genes, and the role of the genetic variants in disease development may help to improve prevention and management of GDM and T2DM in the future

Clinical Trial Registration Number: 2015/1035

Supported by: SOR-OST, NDA, ISCIII

Disclosure: T. Linares-Pineda: None.

143

Kidney disease in women with previous gestational diabetes: a nationwide, register-based cohort study

M.H. Christensen ^1,2, C. Bistrup^3,4, K.H. Rubin^5,6, E.A. Nohr^2,4, C.A. Vinter^2,4, M.S. Andersen^4,7, S. Möller^5,6, D.M. Jensen^1,4;

¹Steno Diabetes Center Odense, Odense University Hospital, Odense, Denmark, ²Department of Gynecology and Obstetrics, Odense University Hospital, Odense, Denmark, ³Department of Nephrology, Odense University Hospital, Odense, Denmark, ⁴Department of Clinical Research, University of Southern Denmark, Odense, Denmark, ⁵Research unit OPEN, Department of Clinical Research, University of Southern Denmark, Odense, Denmark, ⁶OPEN - Open Patient data Explorative Network, Odense University Hospital, Odense, Denmark, ⁷Department of Endocrinology, Odense University Hospital, Odense, Denmark.

Background and aims: Women with gestational diabetes mellitus (GDM) have an increased risk of developing diabetes and hypertension later in life and may be at higher risk of incident kidney disease compared to women without GDM. We explored the association between GDM and long-term incident chronic and acute kidney disease. Further, we quantified the potential mediating effect of diabetes and hypertension on the association and investigated the potential impact of level of insulin resistance on the future risk of kidney disease.

Materials and methods: This nationwide register-based cohort study included all women delivering in Denmark during 1997-2018. The exposure was GDM and outcomes were incident chronic and acute kidney diseases based on diagnosis codes. We used presence of GDM and need of insulin treatment during pregnancy as proxies for the level of insulin resistance and stratified according to subsequent development of diabetes or hypertension. Cox regression and mediation analyses were performed.

Results: In the study population of 697,622 women, 3.4% were diagnosed with GDM. Median follow-up was 11.9 years (range 0-21.9). Women with GDM had a 92% increased risk of developing chronic kidney disease (adjusted hazard ratio (aHR) 1.92 [95% CI 1.67-2.21]) compared to women without GDM. No significant association was found between GDM and incidence of acute kidney disease (aHR 1.08 [95% CI 0.90-1.29]). The association between GDM and chronic kidney disease was mediated by subsequent diabetes and hypertension by 75.7% (95% CI 61.8-89.6) and 30.3% (95% CI 25.2-35.4), respectively. Increased levels of insulin resistance were paralleled by increased risk of chronic kidney disease. This pattern was evident both in women with and without subsequent development of diabetes or hypertension.

Conclusion: Women with GDM had higher risk of developing chronic but not acute kidney disease. Subsequent diabetes and hypertension partly mediated the increased risk. Level of insulin resistance appeared to be associated with the future risk of chronic kidney disease.

Supported by: Research grant from the Danish Diabetes Academy which is funded by the Novo Nordisk Foundation

Disclosure: M.H. Christensen: Grants; Danish Diabetes Academy, Region of Southern Denmark, University of Southern Denmark.

144

Replication of interaction between PTPN22 susceptibility variants and caesarean section in type 1 diabetes risk: the Norwegian mother, father and child cohort study

L. Stene ¹, K. Størdal^1,2, T. Skrivarhaug³, G. Tapia¹;

¹Department of Chronic Diseases, Norwegian Institute of Public Health, Oslo, Norway, ²Institute of Clinical Medicine, University of Oslo, Oslo, Norway, ³Division of Paediatric and Adolescent Medicine, Oslo University Hospital, Oslo, Norway.

Background and aims: Type 1 diabetes (T1D) results from a combination of genetic susceptibility variants and largely unknown environmental exposures. Most suggested environmental risk factors have shown weak association with T1D, including birth by caesarean section. We previously discovered an interaction between the established PTPN22 coding variant rs2476601 (R620W/1858C>T) and caesarean section in a case-control study. We set out to replicate this in an independent cohort.

Materials and methods: The Norwegian Mother, Father and Child Cohort Study (MoBa) include over 100,000 mother and child pairs and T1D was ascertained in the Norwegian Childhood Diabetes Registry. Mode of delivery was obtained from the Medical Birth Registry of Norway and DNA extracted from umbilical cord vein blood was genotyped using an Illumina custom chip with valid data from 120 children who developed T1D and 373 randomly selected controls in MoBa.

Results: Caesarean section occurred among 21.7% of cases and 9.9% of controls. Like our previous finding, we found that carrying at least one PTPN22 risk allele was strongly associated with T1D in those born vaginally (OR=2.24, 95%CI 1.37-3.65), p=0.001, but not in those born by caesarean section (OR=0.63, 95%CI: 0.20-1.96, p=0.42), p(interaction)=0.045. Results were consistent after excluding those with maternal T1D. We further explored and confirmed similar patterns for both a non-HLA genetic risk score based on >40 established T1D susceptibility SNPs, and for celiac disease as the outcome in the same cohort.

Conclusion: There seem to be a consistent pattern of negative interaction where birth by caesarean section reduce or nullify the effect of genetic susceptibility variants in PTPN22 and other loci. Further work must clarify potential mechanisms, but we speculate that caesarean section somehow influences pathways involved in genetic susceptibility, perhaps by early life microbiota.

Supported by: Norwegian Research Council

Disclosure: L. Stene: None.

OP 25 Muscles at work

145

Deletion of the type 2 diabetes candidate gene SLC16A11 worsens whole body insulin sensitivity via impaired skeletal muscle metabolism in mice

N.N. El-Agroudy ^1,2, T. Schumann³, A. Kurzbach⁴, C. Hermann³, L. Sandforth¹, B. Hamilton⁵, G.I. Shulman⁶, A.L. Birkenfeld¹;

¹Institute for Diabetes Research and Metabolic Diseases (IDM), Helmholtz Zentrum München at the Univ, Tübingen, Germany, ²Pharmacology & Toxicology, Faculty of Pharmacy, Ain Shams University, Cairo, Egypt, ³Technische Universität Dresden, Dresden, Germany, ⁴University of Konstanz, Konstanz, Germany, ⁵Department of Metabolic Diseases, Boehringer Ingelheim Pharma GmbH & Co KG, Biberach an der Riss, Germany, ⁶Departments of Internal Medicine and Cellular & Molecular Physiology, Yale School of Medicine, New Haven, CT, USA.

Background and aims: Genome-wide association studies identified the monocarboxylate transporter SLC16A11 as a susceptibility gene for Type 2 Diabetes (T2D). Previous studies suggested that SLC16A11 plays a role in hepatic lipid metabolism. However, these studies were controversial, raising a debate of whether SLC16A11 variants reflect loss- or gain-of-function mutations. Here, we aimed at elucidating the role of SLC16A11 in the development of insulin resistance and metabolic dysfunction.

Materials and methods: Murine and human SLC16A11 were expressed in HEK293 cell lines to study their transport function and kinetics. SLC16A11 knockout mouse model was generated using CRISPR/Cas9 and metabolically characterized upon NCD and HFD feeding. Moreover, SLC16A11 mRNA expression was measured in different insulin responsive tissues of ob/ob and db/db mice.

Results: Our experiments in HEK293 cells reveal pyruvate and lactate as substrates for murine and human SLC16A11. Moreover, SLC16A11 mRNA expression was reduced in skeletal muscles of ob/ob and db/db mice (-69.8% and -22.8%; p<0.05) compared to littermate controls. Body weight and composition, energy expenditure, food intake, basal insulin and glucose levels were similar in SLC16A11 KO and WT littermate mice fed NCD or HFD. Interestingly, HFD-fed SLC16A11 KO mice required lower glucose infusion rate during hyperinsulinemic euglycemic clamp (SLC16A11 WT= 36.4±3 mg/kg/min; SLC16A11 KO= 24.7±2 mg/kg/min; P<0.01), indicating reduced whole-body insulin sensitivity. No changes were detected in hepatic glucose production. Besides, liver fat content as well as mitochondrial citrate synthase flux (V_CS) and pyruvate carboxylase flux (V_PC) were unaltered in livers of HFD-fed SLC16A11 KO mice compared to WT littermates. Instead, clamped 2-deoxy-glucose uptake was significantly lower in skeletal muscle (-41%; p<0.05) of SLC16A11 KO, together with increased membrane/cytosolic DAGs levels and membrane PKCθ activation, confirming skeletal muscle insulin resistance. The later was accompanied by significant reduction in CI- and CII-linked mitochondrial OXPHOS capacity, mitochondrial maximal respiration capacity (ETS) as well as complex-I levels; indicating impaired mitochondrial respiration in skeletal muscle of SLC16A11 KO mice.

Conclusion: Our data confirm for the first time that deletion of SLC16A11 deteriorates skeletal muscle insulin sensitivity without affecting hepatic metabolism. Accordingly, we propose that increasing SLC16A11 expression or activity could hold beneficial therapeutic effects for T2D.

Disclosure: N.N. El-Agroudy: None.

146

A long non-coding RNA regulating skeletal muscle protein metabolism

I. Sen, E. Caria, M. Savikj, J. Smith, J. Zierath, A. Krook;

Karolinska Institute, Stockholm, Sweden.

Background and aims: Long non-coding RNAs (lncRNAs) are important regulators of skeletal muscle physiology, and their expression is modulated in type 2 diabetes (T2D). In this study, we identify differentially expressed skeletal muscle-derived lncRNAs in individuals with T2D in comparison to the ones with normal glucose tolerance (NGT) and investigate the role of this/these lncRNA(s) in the regulation of skeletal muscle metabolism.

Materials and methods: 19 individuals with T2D and 17 matched controls with NGT were recruited to this study. Vastus lateralis skeletal muscle biopsies were collected from these individuals and total RNA samples were isolated. RNA-SEQ was performed for the discovery of differentially expressed skeletal muscle-derived lncRNAs. Several metabolic and molecular biology related assays were used for the functional characterization.

Results: RNA-SEQ data revealed one differentially expressed lncRNA in T2D subjects compared to the NGT control group. This was a human specific, novel, antisense lncRNA, which was significantly downregulated in T2D group (p=0,013). When we silenced the lncRNA in human skeletal muscle cells we observed a significant decrease in protein content (p<0,05) and protein synthesis (p<0,05) as tested by SunSET assays. Furthermore, we show that this lncRNA regulates protein synthesis through the modulation of ribosomal biogenesis confirmed with qPCR assays for ribosomal RNAs and protein coding genes. With RNA pulldown experiments followed by Mass Spectroscopy, we identified putative protein binding partners of this lncRNA. Using CLIP-SEQ data from ENCORI database, a list of unique interactors was created, which were enriched for IGF2BP1 and its interactors. These proteins were shown to bind and regulate the mRNA stability and/or translation of the master regulator of ribosomal biogenesis, MYC. We tested whether our candidate lncRNA was involved in any of these processes and observed that upon its knockdown MYC translation was decreased significantly (p<0,05).

Conclusion: We identify and functionally characterize a novel lncRNA, significantly downregulated in T2D, which regulates ribosomal biogenesis by modulating MYC translation possibly via physical interaction with RNA-binding IGF2BP proteins. This process may be important for the maintenance of skeletal muscle mass and by extension muscle function in people with T2D.

Clinical Trial Registration Number: 2013/647-31/3

Supported by: EFSD Rising Star Fellowship, EFSD/Lilly Young Investigator Research Award

Programme

Disclosure: I. Sen: None.

147

The insulin-sensitising effect of high-intensity interval training (HIIT) is associated with increased hydroxyphenyllactic acid levels in plasma and skeletal muscle

P.M. Møller ¹, M.H. Petersen¹, M.E. de Almeida², N. Ørtenblad², J. Havelund³, N.J. Færgeman³, K. Højlund⁴;

¹University of Southern Denmark, Odense University Hospital, Odense C, Denmark, ²Sports Science and Clinical Biomechanics, University of Southern Denmark, Odense, Denmark, ³Biochemistry and Molecular Biology, University of Southern Denmark, Odense, Denmark, ⁴Steno Diabetes Center Odense, Odense University Hospital, Odense C, Denmark.

Background and aims: Insulin resistance is associated with altered levels of blood and tissue metabolites. High-intensity interval training (HIIT) improves metabolic health, including insulin sensitivity, in obesity and type 2 diabetes (T2D). Here, we examined if the beneficial effects of HIIT are linked to changes in metabolite levels and metabolic pathways in blood, skeletal muscle, and adipose tissue and if these changes are preserved in obesity and T2D.

Materials and methods: Glucose-tolerant lean (n=16) and obese (n=15) men and obese men with T2D (n=13) completed an 8-week HIIT protocol combining cycling and rowing. HIIT improved insulin sensitivity, increased VO₂max, and improved body composition, as observed by a decrease in fat mass and an increase in lean body mass in all groups. Resting plasma, skeletal muscle, and subcutaneous adipose tissue samples collected after an overnight fast before and 48-h after the HIIT protocol were used for large-scale untargeted metabolomics analyses.

Results: HIIT altered >300 muscle metabolites across all groups, including metabolites within the lysine, tryptophan, and tricarboxylic acid cycle (TCA) metabolic pathways, but did not significantly impact the metabolome of the adipose tissue. Muscle levels of citric acid and cis-aconitic acid, the first two substrates of the TCA, increased with HIIT in all groups (adj. p<0.05), while baseline levels tended to be decreased in obese and T2D compared to lean men. Interestingly, we did not observe any changes in the other detected substrates of the TCA. In plasma, we found a few metabolites regulated by HIIT. Among these was hydroxyphenyllactic acid, a metabolite related to the tyrosine- and phenylalanine metabolism. HIIT increased hydroxyphenyllactic acid in all groups (obese & T2D adj. p<0.05, lean unadj. p=0.003), and this change was conserved in muscle as well (lean & obese adj. p<0.05, T2D adj. p=0.065). Both plasma and muscle metabolite levels correlated with insulin sensitivity (r>0.59, p<0.001), VO₂max (r>0.44, p<0.003), and BMI (r<-0.41, p<0.002). To test whether hydroxphenyllactic acid plays an active role in the insulin-sensitizing effect of HIIT, we treated cultured myotubes with hydroxyphenyllactic acid and measured levels of proteins involved in exercise and insulin signaling. Preliminary data suggest that hydroxyphenyllactic acid increases the abundance of AMPK, AKT, and TBC1D4, proteins involved in exercise and insulin signaling to glucose transport.

Conclusion: HIIT mainly changed the muscle metabolome with intact responses in obese and T2D men. These included, among others, altered levels of metabolites involved in oxidative energy production. In addition, hydroxyphenyllactic acid, a metabolite regulated by HIIT in both plasma and muscle, was associated with improved insulin sensitivity and may contribute to this process by increasing the abundance of exercise- and insulin-signaling enzymes.

Clinical Trial Registration Number: NCT03500016

Supported by: Novo Nordisk Foundation, SDCO PhD grant, Region Syddanmark PhD grant

Disclosure: P.M. Møller: Stock/Shareholding; Novo Nordisk A/S.

148

The effect of 12 weeks of high intensity interval training (HIIT) on 24h rhythmicity in substrate metabolism of individuals at risk for developing type 2 diabetes

M. Kotte, J.-F. Harmsen, I. Habets, S.S. de Kam, J. Jorgensen, F. Bosschee, K. Frenken, J.F. Cissen, T. van de Weijer, M. Buitinga, J. Hoeks, P. Schrauwen;

Universiteit Maastricht, Maastricht, Netherlands.

Background and aims: We previously reported that young healthy lean males display 24h rhythmicity in substrate metabolism and skeletal muscle mitochondrial function. This rhythmicity was blunted in older males at risk for developing type 2 diabetes. Here, we investigated whether 12 weeks of HIIT could restore 24h rhythmicity in substrate metabolism, plasma substrate levels and skeletal muscle mitochondrial function in males at risk for developing type 2 diabetes.

Materials and methods: We here report the preliminary analysis of 8 males at risk for developing type 2 diabetes (40-75 years, BMI >25 kg/m²). Indirect calorimetry was performed at 8AM, 1PM, 6PM, 11PM and 4AM, blood draws were taken two-hourly and muscle biopsies were obtained at 8:30AM, 1:30PM and 11:30PM prior to and following 12 weeks of HIIT. Body composition was measured using air displacement plethysmography and exercise capacity was measured by an incremental cycling test.

Results: 12 weeks of HIIT significantly decreased fat mass (p<0,05), improved exercise capacity(p<0,05), and displayed a trend for enhanced mitochondrial respiration (p=0,07). However, no changes in day-night rhythmicity of whole-body substrate metabolism, plasma substrate levels or skeletal muscle mitochondrial function were observed. Additional data, including data on clock gene expression in muscle, will be presented at the conference.

Conclusion: We show that 12 weeks of HIIT improved body composition and exercise capacity but did not influence day-night rhythmicity in substrate metabolism, plasma substrate levels and skeletal muscle mitochondrial function. Future studies could investigate whether different exercise models or time-restricted eating interventions can restore 24h metabolic rhythmicity in individuals at risk for developing type 2 diabetes.

Clinical Trial Registration Number: NCT04565418

Supported by: EFSD

Disclosure: M. Kotte: None.

149

High-fat diet impairs temporal dynamics of acute exercise in mice

L.A. Pendergrast ¹, L. Dollet^2,3, S.P. Ashcroft³, A.M. Ehrlich³, J.T. Treebak³, A. Krook², J.R. Zierath^1,3;

¹Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden, ²Department of Physiology and Pharmacology, Karolinska Institutet, Stockholm, Sweden, ³University of Copenhagen, Copenhagen, Denmark.

Background and aims: Exercise is empirically supported as a risk-reducing therapeutic for type 2 diabetes. Time of day regulates post-exercise metabolism in multiple organs, including liver, skeletal muscle, and adipose tissue. However, metabolic disease disrupts homeostatic function of the circadian clock. We aimed to characterize the timing-specific effects of acute exercise in a rodent model of obesity.

Materials and methods: C57BL6/NTac male mice (6wks of age) were fed a standard chow or high-fat diet for 5 weeks. At week 5, mice were subjected to a 60-minute (16m/min, 5% incline) running bout (or sham) during the early rest (day) or early active (night) phase. Tissue and serum samples were collected immediately post-exercise (n=6 /group). Basal and isoproterenol-stimulated adipose tissue lipolysis was assessed ex vivo for 1 hour following exercise (or sham). Immediately post exercise, in vivo glucose oxidation was measured after oral administration of ¹³C-glucose via ¹³CO₂ exhalation analysis in metabolic cages. Circulating exercise-responsive molecules were assessed in serum. Gene expression was analyzed in liver, inguinal white adipose tissue (iWAT), and gastrocnemius. Data was analyzed by 3-way ANOVA (main effects of diet, exercise, and phase) or a 2-way ANOVA (main effects of exercise and phase) using R.

Results: Expression of core clock genes Nr1d1 and Per2 in the iWAT; Clock and Arntl in the gastrocnemius; and Clock in the liver were altered in high-fat fed mice (3-way ANOVA; phase x diet interaction effect; p<0.05)—indicating a less distinct core clock between activity phases. Glucose oxidation was reduced at the early active phase (3-way ANOVA; phase effect; p<0.0001) and after exercise (3-way ANOVA; exercise effect; p=0.002). Chow mice did not display a post exercise decrease in glucose oxidation at the early active phase (2-way ANOVA; phase x exercise interaction; p=0.06), suggesting that timing influences postexercise substrate metabolism, but only in healthy mice. To assess if somatic substrate provision varies according to time of day, ex vivo lipolysis was determined in subcutaneous adipose tissue. In chow mice only, adipose tissue lipolysis was greater at the early active phase (2-way ANOVA; phase effect; p=0.051) and after exercise (2-way ANOVA; exercise effect; p=0.037). Isoproterenol-induced lipolysis was greater at the early active phase (2-way ANOVA; phase effect; p=0.015) in chow mice only. These data suggest a phase-specific reliance on fatty acids following exercise unique to healthy mice, which may rely upon a timing-specific sensitivity to adrenergic signaling in adipose tissue. Expression of fatty acid transporter 1 (Slc27a1) in gastrocnemius was greater during the early active phase only in chow mice (2-way; phase effect, p=0.017)—suggesting obesity impairs time-of-day-specific fatty acid transport.

Conclusion: Obese mice show a disturbance of core clock gene expression within tissues that are central to substrate turnover. In the context of timing, high-fat diet mice lack a distinct responsiveness to exercise at the whole-body level (glucose oxidation), as well as the tissue-specific level (via adipose tissue lipolytic function). In conclusion, obesity dampens diurnal metabolism, which may also influence the response to exercise.

Supported by: KID Karolinska Institutet Doctoral grant; Novo Nordisk Foundation; Swedish Diabetes Foundation

Disclosure: L.A. Pendergrast: None.

150

Acute exercise simultaneously increases GDF15 and unfolded protein response/integrated stress response signalling in skeletal muscle in obesity and type 2 diabetes

R. Sabaratnam ¹, J.M. Kristensen¹, A.J.T. Pedersen¹, R. Kruse¹, A. Handberg², J.F.P. Wojtaszewski³, K. Højlund¹;

¹University of Southern Denmark, Odense C, Denmark, ²Department of Clinical Biochemistry & Department of Clinical Medicine, Aalborg University, Aalborg, Denmark, ³The August Krogh Section for Molecular Physiology, Department of Nutrition, Exercise and Sports, University of Copenhagen, Copenhagen, Denmark.

Background and aims: Regular exercise is important in the prevention and management of cardiometabolic diseases including obesity and type 2 diabetes (T2D). Exercise-induced factors (exerkines) may be involved in the beneficial metabolic effects of exercise. Conversely, an impaired response to exercise and recovery due to attenuated expression/secretion of exerkines may contribute to cardiometabolic diseases. We aimed to investigate the effect of acute exercise on growth/differentiation factor 15 (GDF15) expression and circulating levels and its putative upstream regulators, the unfolded protein response (UPR)/integrated stress response (ISR) in muscle in patients with T2D and weight-matched individuals.

Materials and methods: Skeletal muscle biopsies and serum samples were obtained from 14 obese glucose-tolerant men and 13 obese men with T2D immediately before, after exercise and 3-h into recovery. Circulating and muscle transcript levels of GDF15 and key markers of UPR/ISR were determined. Furthermore, protein/phosphorylation levels of major regulators in UPR/ISR were investigated.

Results: In weight-matched controls and patients with T2D, acute exercise increased muscle GDF15 mRNA expression (P<0.001) and serum GDF15 levels (P<0.001), which remained elevated 3-h into recovery (P<0.001) compared with basal levels. Immediately after exercise, transcript levels of the markers of UPR/ISR, ATF4, CHOP, EIF2K3 (encoding PERK) and PPP1R15A (encoding GADD34) were increased in both groups (all P<0.05) of which only CHOP expression remained elevated 3-h into recovery (P=0.003). The UPR-specific markers including XBP1-U, XBP1-S and EDEM1 were increased after exercise in both groups compared with basal levels (all P<0.05). In weight-matched controls, the phosphorylation levels of eIF2α-Ser51, a core molecule in the UPR and ISR, increased after exercise in controls (P<0.001) but decreased 3-h into recovery in both groups (P<0.05). Exercise did not affect the expression of HSPA5/GRP78, the phosphorylation levels of PERK or the gene expression or protein abundance of ATF6.

Conclusion: Our findings reveal that acute exercise elicits pronounced and overall similar changes in the muscle expression and circulating levels of GDF15 in patients with T2D and weight-matched controls. In both groups, an exercise-induced activation in muscle of specific components from the UPR/ISR signalling cascade was observed. Collectively, our data suggest that exercise-mediated regulation of GDF15 and its putative upstream regulators is not impaired in patients with T2D.

Supported by: Danish Diabetes Academy supported by the Novo Nordisk Foundation

Disclosure: R. Sabaratnam: None.

OP 26 Many digits in digital diabetes

151

Correlation of glycaemic variability in the ICU, with mortality in critically ill patients, based on data visual analytics of real world evidence

G.E. Dafoulas ¹, I. Kalamaras², K. Votis², A. Bargiota¹;

¹Faculty of Medicine, University of Thessaly, Larisa, Greece, ²Centre for Research & Technology - Hellas, Thessaloniki, Greece.

Background and aims: Evidence suggests a role of glycemic variability (GV) in intensive care unit (ICU) mortality. Various metrics of GV for implementing glycemic control in critical patients, have been suggested and evidence is required to determine the most suitable metric. Coefficient of Variation (CoV) of glucose has been suggested as a measure for glucose variability in ICU. CoV is derived as ratio of blood glucose Standard Deviation (SD) over mean blood glucose (MBG): CoV=SD /MBG. Another indicator of glycemic variability, the J-Index, is defined as J= 0.001(MBG+SD)² .This study investigated the CV in mixed type of ICUs, as a prognostic marker for mortality in critically ill patients.

Materials and methods: We performed a retrospective cohort study using the Medical Information Mart for Intensive Care IV credentialed, anonymized database (MIMIC-IV), based on the data of 18071 ICU admissions between 2008 and 2019 at Beth Israel Deaconess Medical Center, USA. The selected admissions were those with at least 10 glucose measurements. The study protocol was approved by the respective Institutional Review Boards. From the total of 18071 admissions, 15354 (85%) were survivors and 2717 (15%) were non-survivors. Male subjects were 10961 (60.7%) in total, while the overall subject age was 65.2 +/- 15.3 years (mean +/- SD). From each admission, we extract two indicators of glycemic variability: CoV and the J-Index. We use the two indicators as the x and y coordinates, respectively, to visualize ICU admissions. We use logarithmic scales to demonstrate the shape of the distributions more clearly. Values at the bottom-left correspond to small variability, as measured by each indicator, while values at the top-right correspond to large variability. Rather than visualizing each admission as a point, we visualize the density of the points, using kernel-density estimation. We highlight the differences in density between survivors and non-survivors, by visualizing the difference in density between ICU admissions where the persons did not survive (yellow areas), versus those where the persons survived (blue areas).

Results: The results are shown in Panel 1 and 2 of the Figure below, for female and male subjects, respectively. The results are stratified per tertile (33% quantile) of OASIS (Oxford Acute Severity of Illness Score) and SOFA (Sequential Organ Failure Assessment) scores.

Conclusion: For small and medium SOFA scores, the non-survivor distribution exhibits a shift to the right and, for higher OASIS, to the top, indicating that non-survivors exhibit higher glycemic variability, measured with J-Index and CoV. The situation is similar in both genders, although more pronounced in male subjects. As SOFA and OASIS increase (bottom-right charts), this tendency disappears, with no clear shift towards higher variability, indicating that glycemic variability is not a critical factor for survival any more.

Supported by: Co-financed by Greece & the EU - ESF through the Operational Programme Human Resources Development, Education and Lifelong Learning 2014-2020

Disclosure: G.E. Dafoulas: None.

152

A machine learning model for basal insulin titration of people with type 2 diabetes: preliminary results

C.H.N. Thomsen ^1,2, T. Kronborg^1,2, S. Hangaard^1,2, P. Vestergaard^2,3, O. Hejlesen¹, M.H. Jensen^1,2;

¹Department of Health Science and Technology, Aalborg University, Aalborg, Denmark, ²Steno Diabetes Center North Denmark, Aalborg, Denmark, ³Department of Endocrinology, Aalborg University, Aalborg, Denmark.

Background and aims: Insufficient glycaemic control among insulin-treated people with type 2 diabetes (T2D) is a known issue caused by clinical inertia-induced behaviour among patients and clinicians. The titration process is further complicated because the individual's optimal basal insulin dose differs, primarily due to variations in pancreatic insulin secretion and insulin sensitivity. Therefore, basal insulin titration is a pursuit of an individual's optimal dose. These circumstances may lead to prolonged and conservative titration, leading to treatment fatigue and non-adherence. A tool able to foresee the effect of a change in basal insulin dose on fasting blood glucose (FBG) during titration may ease some of these challenges by enabling more optimal and individualised titration. Thus, this study aims to develop a machine learning model to predict the effect of a change in basal insulin dose on FBG in people with T2D.

Materials and methods: Data from a clinical trial investigating the efficacy and safety of insulin treatment in 786 adults with T2D was used. The data included age, sex, diabetes duration, self-reported FBG, self-reported basal insulin dose, hypoglycaemia, trial site, HbA_1c, and various blood and urine samples at baseline. Using a ranking approach, 80% and 20% of data were stratified into a training and test dataset. The output variable was total change in FBG during six weeks of titration. Fivefold cross validation was used to select features using forward selection and to train a multiple linear regression model optimised on relative absolute error (RAE). The model was tested on the independent test dataset to determine the model's performance. Pearson correlation coefficient (R) and RAE were reported.

Results: Subjects had a mean age of 57±9.3 years, mean duration of diabetes of 12±6.7 years, mean HbA_1c of 8.1±0.7%, and 50% were men. The total change in basal insulin dose over six weeks was 10.6±10.2 U [range: -31-40 U]. The selected features were FBG at baseline, total change in basal insulin dose over six weeks, trial site, haemoglobin level in blood at baseline, and alkaline phosphatase level in serum at baseline. During validation, the model obtained a mean RAE=0.64 [95% CI: 0.62-0.66] and R=0.75 (p<0.001) [95% CI: 0.73-0.77] and a mean RAE= 0.67 and R=0.73 (p<0.001) during test. Figure 1 shows the observed vs predicted change in FBG.

Conclusion: It is possible to predict the change in FBG due to a change in basal insulin dose with respect to individual characteristics in people with T2D using machine learning. However, a better model fit is expected if insulin doses collected by digital insulin pens are used instead of self-reported doses due to the known issue of non-adherence. Implementation of a similar model in clinical practice may reduce clinical inertia since glycaemic targets may be reached timelier.

Clinical Trial Registration Number: NCT01819129

Supported by: This work is supported by Steno Diabetes Center North Denmark, which is founded by the Novo Nordisk Foundation.

Disclosure: C.H.N. Thomsen: None.

153

Improvement in time in range after smart pen initiation in routine clinical practice

P. Adolfsson ¹, N.V. Hartvig², A. Kaas², N.N. Knudsen², J.K. Mader³;

¹Department of Pediatrics, Institute of Clinical Sciences, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden, ²Novo Nordisk A/S, Bagsværd, Denmark, ³Division of Endocrinology and Diabetology, Department of Internal Medicine, Medical University of Graz, Graz, Austria.

Background and aims: In 2020, a pilot study conducted in Sweden reported improved glycaemic control following the introduction of a smart insulin pen. The aim of the current multi-country study was to investigate whether people living with type 1 diabetes or type 2 diabetes (PLWD) on a multiple daily insulin regimen and using a continuous glucose monitoring (CGM) device had an improvement in their glycaemic control after initiating a smart insulin pen for injecting their bolus insulin in routine clinical practice.

Materials and methods: Data were collected from adults (≥18 years) with insulin-treated diabetes who were already using a CGM and started administering bolus insulin using a smart insulin pen (NovoPen 6) together with a CGM app to upload their injection data. PLWD were enrolled from 18 countries. To evaluate only time where the connectivity function of the pen was in active use, days with CGM data were included in the analysis, only if an individual had at least two injection data uploads to an app within the last 14 days. Key glycaemic outcomes included time in range (TIR; 3.9-10.0 mmol/L) and time below range (TBR; <3.9 mmol/L) in the overall dataset and for subgroups by level of baseline TIR for those who had 3 months of baseline CGM data.

Results: Data from 8931 adults were used for the analyses, comprising 580,101 days with CGM data and 221,596 days with recorded injections. The mean age (SD) was 42.2 (15.5). Mean TIR increased 1% from baseline to month 3 (95% CI 0.5, 1.4; p < 0.001) and 0.9% at month 6 (95% CI -0.0, 1.7; p = 0.054). Mean TBR decreased -0.3% from baseline to month 3 (95% CI -0.4, -0.1; p < 0.001) and -0.1% at month 6 (95% CI -0.3, 0.1; p = 0.352). When evaluating outcomes in subgroups based on level of TIR in the 3-month baseline period (N=3720), PLWD with TIR of <40% in this baseline period (N=703) had a significant increase in TIR of 5.1% after 3 months (p < 0.001; Figure). The observed changes in TBR were not significant for any of the subgroups.

Conclusion: Among PLWD initiating a smart insulin pen for their bolus insulin, there was a significant increase in TIR after 3 months, with the greatest improvements seen in those who had the lowest TIR in the baseline period. In addition, there was a significant decrease in TBR after 3 months. Similar changes in TIR and TBR were seen at 6 months but were not statistically significant owing to fewer PLWD having data at this timepoint. Further analysis will be conducted to investigate outcomes over a longer observation period. These real-world data showed that glycaemic outcomes improved after smart insulin pen initiation, especially in those with the highest unmet need. Combining a smart insulin pen with additional support such as app-based training or with education may further improve glycaemic control.

Supported by: This study was funded by Novo Nordisk A/S

Disclosure: P. Adolfsson: None.

154

Effect of smartphone app-based intervention on physical activity and glycemic control in patients with type 2 diabetes: a randomised controlled trial

R. Oh ¹, S. Kim², S. Cho¹, Y.-B. Lee¹, S.-M. Jin¹, G. Kim¹, J. Kim¹;

¹Division of Endocrinology and Metabolism, Department of Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea, ²Department of Clinical Research Design and Evaluation, Samsung Advanced Institute for Health Sciences and Technology, Sungkyunkwan University, Seoul, Republic of Korea.

Background and aims: We investigated the effects of a physical activity encouragement intervention based on a smartphone personal health record (PHR) application on step count increase, glycemic control and body weight in patients with type 2 diabetes.

Materials and methods: In this 12-week, single center, randomized controlled, and 12-week extension study, patients with type 2 diabetes who were overweight or obese were randomized to 1:2 to group using smartphone personal health record app (control group) or group using the app and received individualized motivational text messages (intervention group) for 12 weeks. Text message intervention were not applied to either group in the extension study conducted during the following 12 weeks. The primary outcome was change in daily step count after 12 weeks. Secondary outcomes included HbA1c, fasting glucose, and body weight.

Results: Of 200 participants, 62 (93.9%) in the control group and 118 (88.1%) in the intervention group completed the 12-week main study. Change in daily step count at week 12 was -766 ± 3570 for the control and -200 ± 4160 for the intervention group, showing no significance between two groups (P = 0.365). Among subjects with baseline step counts less than 7500 steps per a day, the change in mean daily step count at week 12 in intervention group (1319 ± 3020) was significantly larger than that in control group (-139 ± 2309) (P=0.009). At week 12, HbA1c in intervention group (6.7 ± 0.5%) was significantly lower than that in control group (6.9 ± 0.6%, P=0.041) and at week 24, changes in HbA1c from baseline were significant in both groups but, comparable between groups. Decrease in HbA1c from baseline to week 12 of intervention group was greater in subjects with baseline HbA1c ≥7.5% (-0.81 ± 0.84% to week 12) compared with those with baseline HbA1c <7.5% (-0.22 ± 0.39% to week 12) (P for interaction = 0.014). Significant body weight reduction from baseline to week 24 was seen in both groups without significant between-group differences.

Conclusion: App-based individualized motivational intervention on physical activity showed the improvement in step count increment and glycemic control in individuals who were physically inactive or who had baseline HbA1c>7.5% in patients with T2DM.

Disclosure: R. Oh: None.

155

Effectiveness of a mobile platform on pregnancy outcomes in women with gestational diabetes

J. Wang ¹, Y. Liu², Y. Gong³, T. Wei⁴, X. Li¹, Q. Wang¹, S. Luo¹, X. Zheng¹, J. Weng¹;

¹Department of Endocrinology and Metabolism, Division of Life Sciences and Medicine, The First Affiliated Hospital of USTC, University of Science and Technology of China, Hefei, China, ²Department of Obstetrics and Gynecology, The First Affiliated Hospital of USTC, University of Science and Technology of China, Hefei, China, ³Department of Endocrinology and Metabolism, The First Affiliated Hospital of USTC, Hefei, China, ⁴Southeast University, Nanjing, China.

Background and aims: Mobile health (mHealth) has increasingly shown promise for the management of gestational diabetes mellitus (GDM). TangMama is an mHealth platform designed for GDM women. The aim of the study was to evaluate the impact of mHealth intervention with TangMama platform on glycemic control and pregnancy outcomes in women with GDM.

Materials and methods: This is a retrospective, single-center, case-control study. A total of 420 patients who were diagnosed with GDM during 24-28 weeks of gestation were recruited, with 206 women in the mHealth group and 214 in the control group. The two groups were matched on age, parity, pre-pregnancy body mass index, 75-g oral glucose tolerance test values, and fasting blood glucose level in the first trimester. The mHealth group received multidiscipline management with TangMama application including glucose monitoring, education, dietary counseling, weight management, and medical counseling. The control group received standard care. Our outcomes included the rate of large-for-gestational-age (LGA) newborns, glycosylated hemoglobin (HbA_1c) level in the third trimester, gestational weight gain (GWG), cesarean delivery, and suboptimal pregnancy outcomes.

Results: Compared with the control group, the mHealth group demonstrated lower HbA_1c level in the third trimester (5.3±0.2% vs 5.5±0.5%, p=0.020), lower GWG (10.4±5.5 kg vs 12.1±8.0 kg, p=0.017), and lower rate of cesarean delivery (45.1% vs 55.1%, p=0.041). The rates of LGA (11.2% vs 19.2%, p=0.023), 1 min Apgar score <7 (0 vs 4.2%, p=0.003) were lower in the mHealth group. The length of maternal hospital stay was shorter (4.2±2.2 days vs 4.7±3.0 days, p=0.037) in the mHealth group. The rates of preterm birth, 5 min Apgar score <7, neonatal hypoglycemia, neonatal hyperbilirubinemia, neonatal respiratory distress, macrosomia, shoulder dystocia showed no significant difference between the groups.

Conclusion: The application of an mHealth platform improved maternal glycemic control, lowered the rate of LGA newborns, improved maternal and infant health in women with GDM.

Clinical Trial Registration Number: ChiCTR2200055766

Supported by: Program for Innovative Research Team of The First Affiliated Hospital of USTC, CXGG02

Disclosure: J. Wang: None.

156

Efficacy of a digital diabetes logbook: findings from a randomised controlled trial

B. Kulzer ¹, D. Ehrmann¹, S. Silbermann², J. Kober³, K. Finke-Groene¹, T. Roos¹, V. Schäfer⁴, I. Vesper², N. Hermanns¹;

¹FIDAM - Research Institute Diabetes Academy Mergentheim, Bad Mergentheim, Germany, ²Roche Diabetes Care GmbH, Mannheim, Germany, ³mySugr GmbH, Vienna, Austria, ⁴Roche Diabetes Care Deutschland GmbH, Mannheim, Germany.

Background and aims: The mySugr PRO app is a digital diabetes diary designed to help people with diabetes manage their glucose management and to reduce complexity and burden of diabetes management. The efficacy of the app was tested in a randomised controlled trial (RCT).

Materials and methods: An open-label, parallel group, RCT with a 3-month follow-up was conducted in 41 study sites across Germany. Eligible participants had to have type 1, type 2 or gestational diabetes and regularly perform self-monitoring of blood glucose. Randomisation was done in a 2:1 ratio, either to the intervention group, using the digital diabetes logbook (mySugr PRO) for 3 months, or to the treatment-as-usual control group without using an app. Primary outcome was reduction in diabetes distress at the 3-month follow-up as assessed with the Problem Areas in Diabetes (PAID) scale. Based on the power analysis, 396 evaluable participants were needed. Analysis with robust regression analysis was based on the intention-to-treat population.

Results: A total of 424 people with diabetes were randomized, 282 to the intervention and 142 to the control group (12.5% type 1 diabetes, 68.2% type 2 diabetes, 18.9% gestational diabetes, age: 51.7 ± 15.2 years, 50% female, diabetes duration: 9.5 ± 10.8 years, HbA1c: 7.1 ± 1.5%, PAID: 21.6±17.4). The drop-out rate was very low with 6.4% and 397 participants completed the 3-month follow-up. Intention-to-treat analysis showed that diabetes distress improved from a median score of 18·59 to 15·00 in the intervention group and slightly increased from 16·25 to 16·56 in the control group. Baseline-adjusted diabetes distress at follow-up was significantly lower in the intervention group compared to the control group (model-based treatment effect: -2·20, 95% CI -4·02 to -0·38, p = 0·0182) indicating that the intervention app significantly reduced diabetes distress (figure).

Conclusion: The popularity of diabetes apps seems to be inversely related to their evidence base of methodologically sound studies. This multi-centre RCT showed a significant reduction of ca. 10% in diabetes distress in users of the mySugr PRO app; this reduction was more pronounced in people with initially elevated diabetes distress. Thus, efficacy of the digital health intervention was demonstrated regarding improvement in psychosocial well-being.

Clinical Trial Registration Number: DRKS00022923

Supported by: The study was funded by Roche Diabetes Care

Disclosure: B. Kulzer: Employment/Consultancy; Beckton Dickinson, Novo Nordisk, Sanofi, Roche Diabetes Care, Dexcom, Berlin-Chemie. Lecture/other fees; Beckton Dickinson, Novo Nordisk, Sanofi, Roche Diabetes Care, Dexcom, Berlin-Chemie.

OP 27 Adipocyte impact on obesity and metabolic health

157

Metabolic crosstalk between beta cells and adipocytes in obese youths with primary insulin hypersecretion

D. Trico ¹, M. Chiriaco¹, J. Nouws², A. Vash-Margita², R. Kursawe³, E. Tarabra², A. Galderisi², A. Natali¹, C. Giannini⁴, M. Hellerstein⁵, E. Ferrannini⁶, S. Caprio²;

¹University of Pisa, Pisa, Italy, ²Yale University, New Haven, CT, USA, ³The Jackson Laboratory, Farmington, CT, USA, ⁴University of Chieti, Chieti, Italy, ⁵University of California, Berkeley, Berkeley, CA, USA, ⁶National Council of Research, Pisa, Italy.

Background and aims: Excess insulin secretion more than compensating for the degree of insulin resistance, defined as "primary" hypersecretion, is associated with obesity and an unfavorable metabolic phenotype. To unravel the complex metabolic crosstalk between β-cells and adipocytes, we examined the mutual influence and longitudinal trajectories of model-derived β-cell function and adipose tissue distribution, cell morphology, and endocrine function in youths with obesity.

Materials and methods: In a multiethnic cohort of 100 non-diabetic adolescents with overweight/obesity (age 15.8 ± 2.7 years, 39% females, BMI z-score 2.1 ± 0.7), primary insulin hypersecretion was defined as the upper tertile of the residuals’ distribution of the OGTT-derived insulin secretion rate (ISR) by insulin sensitivity (WBISI) best-fit curve. Hepatic fat fraction (HFF) and visceral (VAT) and subcutaneous adipose tissues (SAT) were measured by abdominal MRI. Adipocyte size distribution was assessed in periumbilical SAT biopsies. Triglyceride turnover and de novo lipogenesis (DNL) were measured in vivo by a long-term ²H₂O labeling method. Metabolic assessments were repeated after a 2-year follow-up in 68 subjects.

Results: Compared with normosecretors, hypersecretors had increased ISR (fasting +36 ± 11 pmol · m^-2 · min^-1, p = 0.002; OGTT +47 ± 6 nmol/m², p < 0.0001) and β-cell function (glucose sensitivity +43 ± 17 pmol · min^-1 · m^-2 · mmol/L^-1, p = 0.004; rate sensitivity +1085 ± 291 pmol · m^-2 · mmol/L^-1, p = 0.0003), but worse glucose tolerance (2-h glucose +0.6 ± 0.3 mmol/L, p = 0.006), despite similar demographics, BMI, fat mass, and WBISI. Hypersecretors had also higher HFF (+5.3 ± 1.8%, p = 0.024) and relative VAT proportion (+2.0 ± 1.0%, p = 0.048), enlarged adipocytes (peak diameter +7.3 ± 4.2 μm, p = 0.036; nadir diameter +6.4 ± 3.3 μm, p = 0.025), increased lipid turnover (+5.3 ± 2.5 %, p = 0.047) with similar DNL, and higher fat mass-adjusted FFA (+4.3 ± 2.2 μm/kg_FM, p = 0.049) and leptin levels (0.3 ± 0.1 ng · mL^-1 · kg_FM^-1, p = 0.002). At follow up, hypersecretors had increased fat mass and a 3-fold higher risk for altered glucose tolerance (OR 2.92, 95%CI 1.04 - 8.24; p = 0.047), while individuals with larger adipocytes or higher leptin levels showed increased β-cell glucose sensitivity.

Conclusion: In youths with obesity, primary insulin hypersecretion is associated with ectopic fat accumulation, hypertrophic adipocytes, higher leptin and FFA levels per fat mass, and greater lipid turnover, independent of whole-body adiposity and insulin resistance. The adipo-insular axis may influence the trajectories of glucose tolerance, adiposity, and β-cell function over time.

Clinical Trial Registration Number: NCT03195400

Supported by: NIH/NIDDK, EFSD Rising Star Fellowship/EFSD Mentorship Programme

Disclosure: D. Trico: None.

158

IGF2 promotes healthy white adipose tissue expansion by increasing preadipocyte proliferation early in life

I. Altun ¹, X. Yan¹, V. Miok¹, R. Karlina¹, S. Ussar^1,2;

¹Helmholtz Zentrum Muenchen, Neuherberg, Germany, ²German Center for Diabetes Research (DZD), Neuherberg, Germany.

Background and aims: The development of obesity is closely linked to an imbalance between energy intake and expenditure. White adipose tissue (WAT) plays a crucial role in the onset of metabolic syndrome associated with obesity. When the lipid storage capacity of WAT is impaired, lipids accumulate in skeletal muscle and liver, leading to insulin resistance and lipotoxicity. To improve the metabolic health of obese individuals, researchers have traditionally studied transgenic mouse models, comparing diseased mice (fed with high fat diet) to predisposed control mice fed with chow diet. Conversely, our approach is to investigate the different mechanisms underlying the apparently healthy expansion of WAT in early life compared to adult tissue expansion. The underlying hypothesis being that rapid expansion of adipose tissue is crucial for healthy infant development, whereas sustained growth in adulthood leads to impaired systemic metabolism, including insulin resistance and metabolic syndrome.

Materials and methods: We conducted single-cell RNA sequencing (scRNAseq) analysis of the stromal vascular fraction from subcutaneous, perigonadal, and brown adipose tissues in pre-weaned and adult wild type C57Bl6/j mice. Our goal was to understand the compositional differences between adipose tissue depots and age groups. Cell biological studies were conducted using primary in vitro differentiated adipocytes supplemented with IGF2 or IGF2 neutralizing antibodies. We established a novel IGF2-EGFP knock-in mouse model for our in vivo studies and combined this with various immunofluorescence antibody and in situ hybridizations.

Results: Our scRNAseq analysis showed that insulin-like growth factor 2 (Igf2) was the most differentially expressed gene in subcutaneous preadipocytes of pre-weaned mice, which we confirmed by qPCRs. A variety of gain or loss of function experiments using primary cell cultures of adult and pre-weaned subcutaneous preadipocytes in vitro showed that unlike insulin and IGF1, IGF2 had very little effects on adipogenesis per se. Conversely, we show that IGF2 enhances proliferation of adult preadipocytes. This suggests a potential role of IGF2 in regulating the preadipocyte pool size rather adipocyte differentiation. A more detailed analysis in vivo revealed that IGF2 expressing stromal cells are located in the connective tissue surrounding subcutaneous adipose tissue, an area that has previously been reported to be rich in stem-like DPP4⁺ adipocyte progenitor cells.

Conclusion: Our data support a model where IGF2 in pre-weaning preadipcoytes regulates the expansion of the adipose precursor pool. An increased preadipocyte pool during development could promote “healthy” adipose tissue expansion later in life by promoting hyperplastic versus hypertrophic adipose tissue growth.

Disclosure: I. Altun: None.

159

Loss of glucose control is precipitated in formerly obese mice re-exposed to obesogenic diet

C. Rebière, C. Rouault, F. Merabtene, K. Clément, G. Marcelin;

INSERM, Sorbonne University, Nutriomics, UMR_1269, Paris, France.

Background and aims: Obesity is a disease known to be the main risk factor for a number of comorbidities, particularly type 2 diabetes (T2D). It is defined as excessive fat accumulation and is accompanied by histopathological remodeling, including inflammation and fibrosis. White adipose tissue (WAT) plays a major role in regulating the systemic metabolism. However, obesity leads to the loss of its integrity and functions, favoring insulin resistance and the onset of T2D. Consequently, the management of patients with obesity and T2D relies on glucose-lowering therapy but also on weight intervention. Due to the critical role of WAT in metabolic health, we first aimed to characterize WAT remodeling after weight loss and assess whether weight loss improves adipocyte functions. Then, we investigated glucose handling and adipose tissue phenotype in formerly obese mice re-exposed to a new obesogenic episode.

Materials and methods: We set up a mouse model of weight loss in which high-fat diet (HFD) fed obese mice lost weight by changing to a low-caloric diet (chow). Thus, we compared the features of WAT (inguinal and visceral perigonadal) among three groups of mice: lean mice (only chow diet), obese mice (only HFD) and formerly obese (Fob) mice. Then, Fob mice were re-exposed to HFD to examine the consequences of obesity recurrence.

Results: In Fob animals, with body composition normalized to the level of lean mice, only the visceral perigonadal WAT exhibited higher fibrosis deposition on histological sections, a loss of perilipin-1 expression and a decrease in adiponectin mRNA levels in obese and Fob mice compared to lean mice. We also investigated lipolysis and found an increase in basal lipolysis and a loss of sensitivity to the β3-receptors agonist CL316-243 in obese and Fob mice compared to lean mice. We then examined the mechanisms underlying the increase in fibrosis deposition and analyzed the mRNA expression of different fibrosis markers. Surprisingly, while fibrosis increased in Fob WAT sections, the expression of mRNA encoding for fibrosis markers decreased to the level measured in lean mice, indicating an uncoupling between fibrosis production and deposition. We then re-exposed Fob mice to HFD for 4 weeks and observed that, despite similar fat mass regain, glucose handling was more severely impaired in Fob mice re-exposed to HFD compared to age-matched controls exposed to HFD for the first time. We believe that the unresolved WAT dysfunctions may be important in this process as insulin sensitivity was markedly lowered in the visceral WAT in Fob mice re-exposed to HFD.

Conclusion: In summary, although weight loss is metabolically favorable, functional and histological alterations established during obesity persist in the visceral WAT. Because the defects are not resolved by weight loss, they may limit the metabolic improvement triggered by weight loss and accelerate the reappearance and the severity of insulin resistance upon weight regain. We believe that unresolved fibrosis could be instrumental in the maintenance of adipocyte dysfunction following weight loss. The uncoupling between fibrosis marker mRNA expression and fibrosis deposition remains to be elucidated, but our first results suggest an alteration in collagen degradation processes that involve the macrophages and the progenitors in the visceral WAT.

Supported by: AFERO, FRM, ESFD, SFN, ANR, CAPES COFECUB

Disclosure: C. Rebière: None.

160

Relatively low plasma leptin concentrations are associated with a more favourable metabolic phenotype

M. Chiriacò ¹, L. Nesti¹, A. Flyvbjerg², A. Golay³, J.A. Nazare⁴, C.H. Anderwald⁵, A. Mitrakou⁶, R. Bizzotto⁷, A. Mari⁷, D. Tricò¹, A. Natali¹, EGIR-RISC Study Group;

¹University of Pisa, Pisa, Italy, ²Steno Diabetes Center, Copenhagen, Denmark, ³University Hospitals of Geneva, Geneva, Switzerland, ⁴Université Claude Bernard Lyon 1, Lyon, France, ⁵Medical University of Vienna, Vienna, Austria, ⁶National and Kapodistrian University of Athens, Athens, Greece, ⁷National Research Council, Padova, Italy.

Background and aims: Leptin is an adipokine that regulates energy balance and metabolism. Its plasma levels are proportional to fat mass but with a wide inter-individual variability. We tested whether individuals with different leptin levels, independently of adiposity, have different metabolic characteristics.

Materials and methods: We analyzed 1372 healthy adults from the RISC study cohort (30-60 years, M/F 595/777, BMI 26±4 kg/m²), followed-up after 3.5-years and characterized for body composition, metabolic variables with a 75-g OGTT, euglycemic-hyperinsulinemic clamp, intravenous glucose bolus, β-cell function, insulin clearance, hepatic and adipose insulin sensitivity and lipidome composition by quantification of non-esterified fatty acids (NEFA).

Results: Individuals were divided into Hypoleptinemic (HypoL), Normoleptinemic (NormoL) and Hyperleptinemic (HyperL), based on the residuals’ distribution of the sex-specific leptin-fat mass correlation, thus identifying individuals with comparable fat mass but with different plasma leptin concentrations. Compared to HyperL, HypoL had markedly lower leptin levels (6.5 vs 21.9 ng/ml, p<0.001) and, despite similar glucose tolerance, showed substantially lower fasting and OGTT insulin levels (-36% and -40% respectively, p<0.0001), along with higher insulin sensitivity (+34%, p<0.001). Both fasting and OGTT insulin clearance were higher in HypoL compared to HyperL (+14% and +27% respectively, p<0.0001), independently of insulin secretion and insulin sensitivity. Compared to HyperL, HypoL also showed lower β-cell function (glucose sensitivity: -11%; acute insulin response to iv glucose: -24%, p<0.05) and higher hepatic insulin sensitivity both in the fasting state (+41%, p<0.001) and during the hyperinsulinemic clamp (+24%, p<0.05). Also, HypoL showed lower NEFA levels during the clamp steady state compared to HyperL (0.03 vs 0.08 μmol/l, p<0.0001), indicating an increased adipose insulin sensitivity. There was no significant difference in lipidome composition except for a higher palmitoleate proportion (2.46 vs 2.21 %, p<0.0001) in HypoL vs HyperL. Overall, NormoL and HyperL showed similar characteristics and the observed differences were stronger in men and in lean individuals. After 3.5 years, changes in weight and metabolic characteristics were similar across the 3 groups.

Conclusion: For a given degree of adiposity, individuals with lower plasma leptin concentrations, show a more favorable metabolic phenotype characterized by increased insulin sensitivity and insulin clearance, along with reduced plasma insulin concentrations.

Supported by: The EGIR-RISC study was supported by EU grant (QLG1-CT-2001-01252)

Disclosure: M. Chiriacò: None.

161

Homozygous nonsense mutation in LAMA4 in two siblings with partial lipodystrophy and severe metabolic syndrome

Y. Karusheva ^1,2, M. Muso^1,2, J. Tadross^1,2, B. Lam^1,2, S. Lockhart^1,3, S. Gancheva^4,5, M. Bombrich^4,5, D. Savage^1,2, M. Roden^4,5, S. O'Rahilly^1,2;

¹MRC Metabolic Diseases Unit, Wellcome-MRC Institute of Metabolic Science, University of Cambridge, Cambridge, UK, ²NIHR Cambridge Biomedical Research Centre, Cambridge, UK, ³NIHR Cambridge Biomedical Research Centre, Cambridge, Germany, ⁴Institute for Clinical Diabetology, German Diabetes Center, Leibniz Center for Diabetes Research at Heinrich Heine University, Duesseldorf, Germany, ⁵German Center for Diabetes Research (DZD), München-Neuherberg, Germany.

Background and aims: Mutations in over 15 genes have been reported to cause recessive or dominant forms of human lipodystrophy. Whether there are further monogenic undetected forms of lipodystrophy is unclear.

Materials and methods: Two female siblings from a consanguineous family of Turkish origin were diagnosed in their teenage years with partial lipodystrophy characterized by loss of adipose tissue on arms, legs and hips, and excessive adipose tissue on the neck and trunk. The siblings, now aged 33, BMI 21.3 kg/m² and 26, BMI 25.3 kg/m² both have insulin resistant Type 2 diabetes, hepatic steatosis, hypertriglyceridemia and (in one sibling) recurrent acute pancreatitis. The siblings underwent whole exome sequencing to look for mutations potentially causative of partial lipodystrophy.

Results: No pathogenic or possibly pathogenic variants could be detected among the known lipodystrophy genes AGPAT2, AKT2, BANF1, BSCL2, CAV1, CAVIN1, CIDEC, LIPE, LMNA, LMNB2, PIK3R1, PLIN1, POLD1, PPARG, PSMB8, SPRTN, ZMPTSE24. Whole exome sequencing data from both siblings was interrogated for the presence of rare (gnomAD minor allele frequency < 0.1%) homozygous or compound heterozygous mutations. The siblings were both homozygous for an early nonsense mutation (Q45*) in the coding sequence of LAMA4, a gene encoding one of the components of the trimeric structural basement membrane protein Laminin. Laminin alpha-4 is highly expressed in adipose tissue, is involved in the attachment, migration and organization of cells and interacts with components of the extra cellular matrix. Two independent strains of mice lacking LAMA4 have been reported and both have an adipose tissue phenotype comparable to when other human lipodystrophy genes are deleted.

Conclusion: LAMA4 is a reasonable candidate to be the causative genetic defect in these siblings and mutations in this gene should be sought in cases with no molecular diagnosis. The findings highlight the potential importance of the adipose tissue extra cellular matrix to its healthy development and function.

Disclosure: Y. Karusheva: None.

162

Metabolic effects of sodium-glucose-co-transporter-4 (SGLT4) inhibition: from mouse studies to clinical translation

M. Moreno-López, C. Saponaro, A. Acosta-Montalvo, I. Louvet, G. Pasquetti, J. Thevenet, V. Gmyr, V. Raverdy, J. Kerr-Conte, F. Pattou, C. Bonner;

INSERM U1190 Translational Research for Diabetes, Lille, France.

Background and aims: Would inhibiting sodium-glucose-co-transporter 4 (SGLT4) reverse obesity and prevent the progression to type 2 diabetes (T2D)? Unlike some of its better-known SGLT family members, SGLT4 also exhibits a Na⁺-dependent alpha-methyl-D- glucopyranoside transport system with a Km of 2.6 mM, suggesting that it is a low affinity- type transporter, similar to that of SGLT2. Studies using radiolabeled mannose suggested that an 'SGLT4-like transport system' might be physiologically relevant for intestinal absorption. But unlike SGLT1 and SGLT2, which transport only glucose, SGLT4 transports naturally occurring sugars with a rank order of mannose, glucose, fructose, 1.5AG (artificial sweeteners), and galactose, all enriched in the Western diet (WD), and dangerously elevated in the blood of obese individuals with and without T2D. Therefore, we hypothesized that SGLT4 inhibition would lower the serum concentrations of these sugars, which would be a "game-changer" for the treatment of such pathological conditions.

Materials and methods: SLC5A9 gene (encoding SGLT4 protein) regulation was analyzed in the intestine of patients before and after weight-loss surgery. RNAscope analysis was used to determine the precise location of SLC5A9 in the human intestine and pancreas. Sglt4 knock-out (KO) mice were created using CRISPR/Cas techniques, allowing us to study changes in their metabolic phenotype for months while they were fed the WD. Sibling wild-type mice served as controls. Islets were isolated from both groups of mice, and a glucose-stimulated insulin secretion (GSIS) assay was performed by perifusion techniques.

Results: From obese patients with and without T2D (n =50), we observed a significant reduction in BMI (p<0.0001) post-surgery. This was accompanied by a marked reduction in glycemia after a glucose challenge (OGTT). We also found that SLC5A1 (SGLT1) mRNA was significantly reduced in both cohorts. However, we found a more profound decrease of SLC5A9 (SGLT4) mRNA expression (70%) levels post-surgery, compared to SLC5A1 (40%), while SLC5A2 (SGLT2), SLC2A2 (GLUT2), and SLC2A5 (GLUT5) mRNA expression levels remained unchanged. Intriguingly, we also discovered that Sglt4 KO mice were protected against obesity and T2D after only three months of feeding the WD when compared to WT mice fed the same diet. Using RNAScope, we revealed that both SLC5A1 and SLC5A9 mRNA levels were highly expressed in the apical membrane of the intestine of a person with obesity. Moreover, we also observed that SLC5A9 is mainly expressed in the exocrine pancreas and induced by obesity and T2D. We observed that islets isolated from WT mice fed the WD showed impaired GSIS, compared to those of Sglt4 KO mice.

Conclusion: Collectively, these data demonstrate that SLC5A9 mRNA levels are induced in the apical membrane of the intestine and exocrine pancreas in persons with obesity and T2D. Furthermore, Sglt4 deficiency slows the onset of obesity and hyperglycemia in mice fed the WD, improving insulin sensitivity, lending credence to the idea that SGLT4 inhibition improves beta cell function.

Supported by: PreciDIAB

Disclosure: M. Moreno-López: None.

OP 28 Novel signalling pathways regulating islet function

163

Driving the pancreatic beta cell function through nanoscale features: the influence of nuclear mechanotransduction

A. Galli ¹, N. Dule¹, P. Marciani¹, M. Castagna¹, P. Milani², C. Lenardi³, G. Tedeschi⁴, C. Perego¹;

¹Department of Excellence of Pharmacological and Biomolecular Sciences, University of Milan, Milan, Italy, ²Department of Physics, University of Milan, Milan, Italy, ³Cimaina, University of Milan, Milan, Italy, ⁴Department of Veterinary Medicine, University of Milan, Milan, Italy.

Background and aims: Nowadays, to accelerate the development of cell-replacement strategies in diabetes, there is a desperate need to understand mechanisms and stimuli regulating the β-cell development and function. We have recently demonstrated that the extracellular nanotopography promotes β-cell survival, differentiation, and function through the activation of a mechanotransductive pathway, however, the underlying molecular mechanisms need to be further investigated. In the last years, we focused on how mechanical forces might impact the nuclear organization, and thus the transcriptional and translational activities in pancreatic β-cells.

Materials and methods: Human isolated islets were cultured for 20 days on nanostructured zirconia substrates with a controllable ECM-like roughness (ns-ZrO_x), while flat zirconia substrates were used as controls (flat-ZrO₂). The activation of nuclear mechanotransduction was assessed by super-resolution fluorescence microscopy, western blot technique and confirmed by shot-gun proteomic.

Results: We found that the nanotopography, by reducing the cytoskeletal tension, modulates the nuclear shape and architecture (nuclear area and circularity; p<0.001 ns-ZrO_x vs flat-ZrO₂). These modifications are paralleled by the spatial reorganization of the nuclear lamina, and a shift of lamin isoforms in the islets grown on the nanosubstrates (p<0.05 ns-ZrO_x vs flat-ZrO₂). Changes in nuclear lamina might affect the access of transcription factors to the nucleus leading to changes in the transcriptional and translational activities. In accordance, the proteomic analysis revealed the upregulation of chromatin-associated proteins (GO:0031497), RNA binding proteins (GO:0003723), spliceosome (GO:0005681) and heterogeneous nuclear ribonucleoprotein (GO:0030529) complexes in the islets grown on the nanostructure (p<0.05 ns-ZrO_x vs flat-ZrO₂). Furthermore, we observed that the nanostructure modulates the expression of YAP/TAZ, a soluble transcription factor strictly controlled by mechanical forces and critically involved in the regulation of the β-cell fate (p<0.05 ns-ZrO_x vs flat-ZrO₂).

Conclusion: Our data indicate for the first time that the extracellular nanotopography controls the transcriptional and translational activities of pancreatic β-cells, further supporting the hypothesis of a mechanical-induced β-cell reprogramming. These findings provide a better understanding of the mechanotransductive-regulatory mechanisms in β-cells, revealing novel pathways and molecular effectors that may be helpful for designing engineered substrates that can improve the efficacy of replacing therapies in diabetes.

Disclosure: A. Galli: None.

164

Leader beta cells within the pancreatic islets are a functionally stable subpopulation in vitro and in vivo

L. Delgadillo Silva ¹, P. Chabosseau¹, F. Yong^2,3, L. Lopez-Noriega¹, S. Larouche¹, A. Provencher-Girard¹, A. Prat¹, G. Rutter^1,4;

¹CR-CHUM, Montreal, QC, Canada, ²Nanyang Technological College, Singapore, Singapore, ³Imperial College London, London, UK, ⁴Faculty of Medicine, Imperial College London, London, UK.

Background and aims: β-cell coordination within the islet is critical for normal glucose-stimulated insulin secretion. Recently, we and others have identified subpopulations of β-cells termed “leaders”, as well as highly-connected “hubs”. Whereas the former are the first cells to show an increase in intracellular Ca²⁺ during glucose-induced Ca²⁺ waves, the latter are essential for the coordination of Ca²⁺ oscillations. Dysfunction of either population may thus contribute to defective insulin secretion in some forms of diabetes. Whether these groups simply reflect transient functional states or are stable subpopulations with a defined lifetime, has been unclear up to now.

Materials and methods: To address these questions, we used islets from 10-15-week old C57BL6 Ins1Cre.GCaMP6f^LoxP mice which express the Ca²⁺ sensor GCaMP6f selectively in β-cells. Fast (3-6 Hz) confocal imaging and cell-tracking were used to record calcium dynamics with during stimulation in vitro at 11 mM glucose on a Zeiss LSM900 Airyscan microscope. Adenoviral delivery of the cryptic PA-mCherry allowed the specific labelling of leader cells using targeted UV irradiation (“photopainting”).

Engraftment of islets into the anterior chamber of the mouse eye (ACE) was used to assess stability by longitudinal imaging in vivo. During the three-week engraftment period, allowing islet revascularization and innervation, spontaneous photoactivation of pA-mCherry led to the labelling of a small number of cells which provided “way points” to orient islets during Ca²⁺ imaging sessions. Image acquisition was performed in anaesthetized mice (glycemia 12-17 mM) using a Zeiss LMS780. Leader cells were identified by inspection of movies and hub cells after data binarization and co-activity assessment running scripts under MATLAB.

Results: Examined in vitro, the identity of leader cells remained the same at day 0 and when imaged 24 h later (day 1) in 80% (8/10) islets (n=4 mice). Connectivity analysis identified a separate population of highly-coactivated hub cells, representing 11.2% of cells at day 0 (n=3 islets imaged in 3 separate mice). 30.6% of these remained as hub cells at day 1, with an overall increase in the proportion of cells identified as hubs in vitro to 34.4%. The average connectivity of each hub cell remained unchanged from day 0 to day 1, however (p=.25).

After engraftment into the ACE, Ca²⁺ waves in vivo initiated from the same leader cells (usually two per islet) at days 0, 1, 7 and 14 (n=6 islets imaged in 3 separate mice). In contrast, whereas 9.5% of cells were defined as hubs when examined in vivo at day 0, this fraction increased to 18.3% at day 1, as did the average connectivity of each hub cell (p<.001 day 1 versus day 0). However, only about 1/3 (32.7%) of the original hubs at day 0 remained in this role on day 1. The proportion of hubs / islet then decreased progressively (at day 7 and day 14: 8.8% and 1.1% of the total, respectively).

Conclusion: These findings demonstrate that leader β-cells are a highly stable subpopulation in vitro and in vivo, consistent with the distinct transcriptomes and localization of these cells within the islet. The behavior of hubs is more complex, with the proportion and identity of these cells varying substantially in vivo over a two-week time frame.

Supported by: CR-CHUM start-up funding, Welcome Trust Investigator Award

Disclosure: L. Delgadillo Silva: None.

165

Acute and long term somatostatin signalling via the beta cell primary cilium

T. Incedal Nilsson, Ö. Dumral, G. Sanchez, O. Idevall-Hagren;

BMC box 571, Uppsala University, Uppsala, Sweden.

Background and aims: Primary cilia are microtubule-based organelles that protrude from the surface of β-cell. They are enriched for G protein-coupled receptors (GPCRs) that enable the transduction of extracellular signals in parallel with canonical receptor signalling from plasma membrane receptors. The cilium thereby offers alternative pathways for signal processing, but the functional relevance of these pathways is largely unknown. The hedgehog (Hh) pathway requires functional cilia and is important for both pancreas development and maintenance of islet cell function in adulthood. Hh-binding to its receptor Smoothened (Smo) initiates G_i-dependent lowering of cilia cAMP and activation of GLI transcription factors. Other cilia-localized G_i-coupled receptors, such as somatostatin receptor 3 (SSTR3), possibly contribute to this regulation. The aim of this study was to identify putative signalling pathways downstream of ciliary SSTRs.

Materials and methods: Immunostaining and confocal microscopy were used to determine the distribution of islet cell cilia and cilia-localized proteins in mouse and human islets of Langerhans. Cultured mouse islets and MIN6 pseudo-islets expressing a novel cilia-targeted cAMP sensor and TIRF microscopy were used to record ciliary cAMP dynamics. Functional evaluation of ciliary somatostatin signalling was performed in MIN6 pseudo-islets following siRNA-mediated knockdown of SSTR3.

Results: Real-time recordings of cytosolic and ciliary cAMP in mouse islet β-cells revealed similar resting cAMP levels and the addition of forskolin resulted in cAMP increases in both compartments that exhibited similar amplitudes (forskolin: cilia 0.16±0.01, cyto 0.18±0.01, n=19) and kinetics (difference in rise time, 4±4 s; n=19, NS). Exposure to 10 nM GLP-1 stimulated cAMP production in the cytosol, which propagated into the cilium with a lag-time of 10±1s (n=17, P=0.013) and stabilized at a level similar to that of the cytosol (0.21±0.02 vs 0.18±0.01, n=17), indicating free cAMP diffusion. Immunostaining of mouse and human islets revealed strong enrichment of SSTR3 and SSTR5 in primary cilia. Addition of 100 nM sst suppressed forskolin-induced cAMP elevations in both cilia and cytosol; an effect was selectively diminished in cilia following siRNA-mediated knockdown of SSTR3 (control: 99±8% inhibition, siSSTR3: 54±7% inhibition, n=20, P=0.004), demonstrating that activation of ciliary SSTRs induces local changes in cAMP. Stimulation with sst also evoked cilia-specific Ca²⁺ increases that were sensitive to pertussis-toxin treatment and suppressed following SSTR3 knockdown, showing that ciliary sst signaling also involves Ca²⁺. Prolonged exposure to sst (100 nM, 18h) or the Smo agonist SAG caused nuclear accumulation of GLI2 (control: 25%, sst: 56%, SAG: 52%, sst+SAG: 53%, 200-250 cells for each group), indicating crosstalk between the sst and Hh pathways. Importantly, prolonged exposure did not result in loss of ciliary SSTR3 nor sst-induced ciliary Ca²⁺ increases, demonstrating a lack of receptor desensitization in the cilium.

Conclusion: Our results show that the primary cilium is a target of both acute and long-term sst action in islets and indicates roles of sst beyond acute inhibition of hormone secretion. Understanding these mechanisms will be important for understanding sst-regulation of islet function in health and diabetes.

Supported by: Novo Nordisk

Disclosure: T. Incedal Nilsson: None.

166

Identification of transcriptional regulators controlling beta cell differentiation

D. Balboa ^1,2, S. Moratinos¹, J. García-Hurtado^1,2, M.A. Maestro^1,2, J. Ferrer^3,4;

¹Center for Genomic Regulation, Barcelona, Spain, ²CIBER de Diabetes y Enfermedades Metabólicas Asociadas, Instituto de Salud Carlos III, Barcelona, Spain, ³Center for Genomic Regulation, Regulatory genomics and diabetes. CIBER de Diabetes y Enfermedades Metabólicas Asociadas, Instituto de Salud Carlos III, Barcelona, Spain, ⁴Genetics and Genomics Section, Department of Metabolism, Digestion and Reproduction, National Institute for Health Research (NIHR) Imperial Biomedical Research Centre, London, UK.

Background and aims: Our knowledge on the regulatory programs that govern human islet endocrine cell fate determination, development and maturation remains incomplete. To better interpret the genetic variants associated with diabetes and improve pancreatic islet generation from stem cells, it is necessary to identify the genes and regulatory elements involved in these processes. Indeed, an unresolved challenge in the differentiation stem cells to islets is the inability to precisely control the proportion of cells allocated to each major endocrine cell lineage. In this study, we aim to identify and characterize the function of transcriptional regulators involved in islet beta cell differentiation.

Materials and methods: To pinpoint putative candidate genes participating in beta cell differentiation, we interrogated single-cell transcriptome and accessibility profiles of human fetal and adult pancreas and islets, together with stem cell-derived pancreas progenitors and islets. We investigated their function by genome engineering knockout human pluripotent stem cell lines with CRISPR-Cas9 and differentiating them towards functional islet cells. We characterized these differentiated cells using flow cytometry, qRT-PCR, bulk and single cell RNAseq, single cell ATACseq and CHIP-seq.

Results: We identified the Achaete-Scute Family BHLH Transcription Factor 2 gene, ASCL2, as a putative transcriptional regulator of beta cell differentiation. ASCL2 expression peaks in fetal pancreas trunk epithelial cells and in stem cell-derived endocrine progenitor cells. Stem cell-islets (SC-islets) differentiated from ASCL2-KO lines displayed a 48,67% reduction in INS expression compared to controls (p=4.43 E-04), while GCG, PDX1 and MNX1 expression levels were not altered. We observed a 62% reduction in the number of C-PEP+NKX6-1+ cells in SC-islets (10.60±1.38%SEM in KO vs 28.09±3.32%SEM in controls, p=5E-04). ASCL2 transcription is regulated by transcription factors that harbor monogenic diabetes mutations, acting through a distal enhancer that directly interacts with the ASCL2 promoter. Disruption of this enhancer resulted in abrogation of ASCL2 expression and a 45% reduction in the number of C-PEP+NKX6-1+ cells in SC-islets (17,68±2.17% SEM enhancer deleted vs 32.20±3.55% SEM in controls; p= 9.20E-03). Finally, the ASCL2 locus harbors independent genetic signals associated with type 1 and type 2 diabetes.

Conclusion: Our results indicate a previously unrecognized crucial role of transcription factor ASCL2 in beta cell differentiation.

Supported by: EFSD Rising Star Fellowship; Ramon y Cajal Fellowship RYC2021-033131-I

Disclosure: D. Balboa: None.

167

Nutrient sensor OGT & mTORC1 crosstalk to coordinate beta cell mass and function

S. Jo, E. Alejandro;

University of Minnesota, Minneapolis, MN, USA.

Background and aims: A crucial aspect of type 2 diabetes (T2D) is the failure of β-cells to secrete sufficient insulin to tightly regulate blood glucose. Nutrient sensor, mechanistic target of rapamycin complex 1 (mTORC1), balances cell growth and degradative processes (autophagy) to maintain β-cell health and function. Changes to mTORC1 activity and autophagy are linked to T2D and β-cell failure; however, the underlying mechanisms driving these relationships remain unclear. Our recent work highlights nutrient sensor O-GlcNAc Transferase (OGT) as a critical regulator of β-cell health; β-cell OGT deletion (βOGTKO) causes diabetes due to severe β-cell mass loss and insulin secretion deficits. In response to changes in nutrient levels, OGT glycosylates key target proteins to impact various cellular processes. We hypothesize that under nutrient stress, OGT directly modulates the mTORC1 pathway, and thereby, autophagy-dependent β-cell function.

Materials and methods: First, we tested whether mTORC1-driven autophagy affects OGT-dependent β-cell function. To test this, we deleted ULK1 (autophagy initiator downstream of mTORC1) in βOGTKO mice (βULK1/OGTKO) (n=7-9 for in vivo, n=5-6 ex vivo analysis). Subsequently, to fully delineate the signaling crosstalk between OGT and mTORC1 pathways, we generated a mouse model with increased mTORC1 activity by deleting TSC2 (mTORC1 inhibitor) in βOGTKO mice (βTSC2/OGTKO) (n=9 for in vivo, n=4-5 ex vivo analysis).

Results: βOGTKO islets exhibited decreased mTORC1 activity and increased autophagy. Supporting our hypothesis, by attenuating autophagy via deletion of autophagy activating ULK1, βULK1/OGTKO mice showed improved glucose tolerance (AUC; 1-way ANOVA p<0.01) with attenuated hyperglycemia (2-way ANOVA p<0.01) compared to the βOGTKO mice by increasing β-cell function (2-way ANOVA p<0.05) but not mass. By fully restoring mTORC1 activity, βTSC2/OGTKO mice exhibited delayed onset of the diabetic phenotype of the βOGTKO mice. Normalization of blood glucose was associated with a full rescue of the β-cell mass deficit (1-way ANOVA p<0.05) but not improvements in secretory function, which suggests non-overlapping biological processes between OGT and mTORC1. To further study signaling transduction regulated by OGT and mTORC1, we performed phospho-protein antibody array (Full Moon BioSystems) on islets from control, βOGTKO, βTSC2/OGTKO, and βTSC2KO mice, revealing diverging, differential regulation of MAPK and calmodulin signaling by OGT and mTORC1 to modulate β-cell health and function.

Conclusion: Altogether, our data show a previously unknown and important role of OGT as a master nutrient-sensor affecting downstream mTORC1 and autophagy to drive β-cell mass and function.

Supported by: F31 (F31DK131860) to SJ. R01 (R01DK115720) to EUA.

Disclosure: S. Jo: None.

168

Ex vivo mapping of beta cell glutamate pathways by in situ enzymatic activities and metabolic profiling using ToF-SIMS in mouse islet cryosections

Y. Zhou ¹, N. Baez¹, T. Fu², T. Brun¹, P. Maechler¹;

¹Dpt of Cell Physiology and Metabolism, University of Geneva, Geneva, Switzerland, ²Laboratory of Advanced Technology and Dpt of Quantum Matter Physics, University of Geneva, Geneva, Switzerland.

Background and aims: In the beta-cell, glucose-stimulated insulin secretion is governed by various signals, among which elevation of intracellular calcium is mandatory. Glutamate pathways have been shown to modulate metabolism-secretion coupling, although it is debated whether this amino acid is consumed or produced by the glutamate dehydrogenase (GDH). In vitro experiments have provided contradictory results and in vivo assessment of beta-cell glutamate levels remains unrealistic. Here, we developed in situ enzyme-targeted nitroblue tetrazolium (NBT) assays coupled with metabolic profiling using ToF-SIMS to allow ex vivo assessment of GDH activity and metabolite levels, respectively, in cryopreserved tissues.

Materials and methods: We used wild type mice and beta-cell specific GDH KO mice (Beta-Glud1^-/-) of 12 weeks of age. Mice were sacrificed under different feeding conditions: 6h fasted, starved overnight, and fed. Non-fixed pancreata were collected and stored at -80°C before cryosectioning and analysis. Dithizone insulin staining was performed on pancreatic sections to detect beta-cells and assess islet versus acinar zones. The NBT assay was applied on cryopreserved sections to study the following dehydrogenase (DH) activities: glyceraldehyde-3-P-DH (GAPDH) as a readout of glycolysis, lactate-DH (LDH) for potential beta-cell dedifferentiation, succinate-DH (SDH) for mitochondrial activity, and GDH for glutamate pathway. The NBT signal was measured through calibrated optical density in islets with the neighboring acinar cells as reference. On the adjacent cryosections, we performed time-of-flight secondary ion mass spectrometry (ToF-SIMS) to examine metabolites related to glutamate pathways. Plasma glucose, insulin and glucagon levels were determined for each mouse.

Results: Blood parameters of mice upon the various feeding conditions showed the expected elevation of plasma insulin levels once fed (2.5-fold, p<0.01). Applied ex vivo to non-fixed pancreatic cryosections, DTZ insulin staining revealed islets for the in situ assessment of enzyme activities. As anticipated, LDH activity was very low in beta-cells compared with non-islet acinar zones. Islet GAPDH exhibited robust glycolytic activity in the fed mice. Surprisingly, mitochondrial enzymes SDH and GDH were not significantly modified by the feeding state. In situ metabolic profiling by ToF-SIMS showed that, compared to the 6h fasted mice, the specific glutamate fragment intensities were significantly increased in islets of control mice upon feeding (C₄H₆NO₂^-, m/z=100.040, p=0.009; C₅H₇O₄^-, m/z=131.035, p=0.003). This in situ elevation of glutamate levels was absent in Beta-Glud1^-/- mice. Interestingly, pyruvate levels (m/z=87.011) were not influenced by the feeding state, indicating that the flux rather than the absolute levels of this key metabolic intermediate plays a role in metabolism-secretion coupling.

Conclusion: Upon feeding, when there is a rise in glycemia and plasma insulin levels, glutamate increases in beta-cells. This increase is not observed in the absence of GDH in the beta-cells. These ex vivo results argue for an anabolic activity in the glutamate pathway upon glucose-stimulated insulin secretion with GDH-dependent elevation of glutamate.

Supported by: SNF

Disclosure: Y. Zhou: None.

OP 29 Cardiac complications in type 1 and type 2 diabetes

169

Trends in survival after first myocardial infarction in people with diabetes

L. Wedén ¹, T. Andersson^2,3, M. Lind^4,5, B. Eliasson⁶, R. Hofmann¹, T. Nyström¹;

¹Department of Clinical Science and Education, Karolinska Institutet Södersjukhuset, Stockholm, Sweden, ²Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden, ³Center for Occupational and Environmental Medicine, Stockholm, Sweden, ⁴Department of Molecular and Clinical Medicine, Institute of Medicine University of Gothenburg, Gothenburg, Sweden, ⁵Department of Medicine, NU Hospital Group, Uddevalla, Sweden, ⁶Institute of Medicine University of Gothenburg, Gothenburg, Sweden.

Background and aims: Earlier studies have demonstrated that people with diabetes have a worse outcome after myocardial infarction (MI), compared to people without diabetes. In contrast, recent studies have shown a substantially decline in mortality and cardiovascular (CV) outcome among people with diabetes. Trends in survival after a first MI in people with diabetes and without diabetes is still not extensively studied.

Materials and methods: Between 2006 to 2020 we identified 2,527 individuals with type 1 diabetes (T1D), 48,321 individuals with type 2 diabetes (T2D) and 243,170 individuals without diabetes with a first MI in national healthcare registries. Outcomes were trends in survival, i.e. death and CV death, and major cardiovascular events (MACE), i.e. non-fatal stroke, non-fatal MI, CV death and hospitalised heart failure. Multivariable-adjusted proportional hazards ratio (HR) models with 95% confidence intervals (CIs) were separately applied within people without diabetes (control group) and people with T1D, and T2D, with a maximum follow-up period of 1095 days. HRs were estimated in three-year blocks, as well as continuously by the date of entering the cohort.

Results: Individuals with T1D were younger (62 years, SD ±12.2) and more often women (43.6%) compared to individuals with T2D (75 years, SD ±10.8) women (38.1%), and to control group without diabetes (73years, SD ±13.2) women (38.1%). Proportion of ST-elevated MI (STEMI vs. Non-STEMI) between groups were T1D (29% vs 71%), T2D (30% vs 70%) and control group (39 vs 61%), respectively. Three-year trends in death, CV death and MACE between groups are shown in the Table. During follow-up and after multiple adjustments (sex, age, comorbidities, socioeconomic factors and medication) there was a significant decreased annual incidence trend (HR-1 * 100 %) for all cause death in control group -1.9% (-2.1 to -1.7%) and individuals with T2D -1.3% (-1.7 to -0.9%), with no such trend in individuals with T1D 0.0 (-1.9 to 1.9%). Corresponding numbers for CV death were for control group -2.0% (-2.1 to -1.8%), T2D -1.6% (-2.0 to -1.1%), and T1D -0.5% (-2.6 to 1.2%), and for MACE control group -2.3% (-2.5 to -2.2%), T2D -1.9% (-2.3 to -1.6%) and T1D -0.6% (-2.3 to 1.1%), respectively.

Conclusion: During the last 15 years, the trend in death and MACE in people without diabetes and with T2D having a first time MI has decreased significantly. In contrast, such a decreased trend was absent in people with T1D. This finding highlights the urgent need for understanding the cardiovascular risk in people with T1D.

Supported by: ALF agreement between Stockholm County Council and Karolinska Institutet (FoUI961018).

Disclosure: L. Wedén: Grants; AstraZeneca (ML), Dexcom (ML), Novo Nordisk (ML). Honorarium; AstraZeneca (TN), Merck Sharp & Dohme (TN), Novo Nordisk (TN, ML), Eli Lilly and Company (TN), Boehringer Ingelheim (TN), Abbot and Amgen (TN), Dexcom (ML), Medtronic (ML), Rubin Medical (ML). Lecture/other fees; Amgen (BE), AstraZeneca (BE), Boehringer Ingelheim (BE), Eli Lilly (BE), Merck Sharp and Dohme (BE), Mundipharma (BE), Novo Nordisk (BE), RLS Global (BE), Sanofi (BE).

170

Acute myocardial infarction and heart failure in young patients with type 1 and type 2 diabetes

P. Dikaiou ¹, M. Lind¹, J. Ludvigsson², G. Lappas¹, L. Björck¹, A. Rosengren¹;

¹Molecular and Clinical Medicine, Institution of Medicine, Sahlgrenska Academy, Gothenburg University, Gothenburg, Sweden, ²Biomedical and Clinical Sciences, Division of Pediatrics, Linköping University, Linköping, Sweden.

Background and aims: Type 1 and 2 diabetes are different in several ways but both types are strongly associated with high risk of cardiovascular complications. With rising incidence of type 2 diabetes in adolescents and young adults an increasing proportion of young people with diabetes will be type 2 diabetes. Young-onset type 2 compared to type 1 diabetes has been considered the more lethal phenotype of the two but few formal comparisons between the two types exist. The aim of the present study was to compare risk factors between young patients with type 1 diabetes and type 2 diabetes and outcomes with respect to acute myocardial infarction (AMI) and heart failure (HF).

Materials and methods: We used data from the Swedish National Diabetes Register and National Patient Register to identify all patients with type 1 and 2 diabetes aged 18 to 34 years 1996-2019. Data on primary outcomes (acute myocardial infarction (AMI) and heart failure (HF)) was collected between 1998 and 2019.

Results: In total, we included 43,897 patients with diabetes, 32,208 patients (73.4% with type 1 diabetes and 11,689 (26.6%) with type 2 diabetes. Mean age in patients with type 1 and 2 diabetes was 23.1 (standard deviation (SD) 5.5) and 29.1 (4.4) years, respectively, with mean duration of diabetes in type 1 diabetes 8.6 (SD 7.2) and in type 2 1.7 (3.0 years). Patients with type 2, compared to type 1 diabetes had a substantially higher mean body mass index (BMI) of 34.6 (SD 9.1) and 24.3 (SD 6.2) kg/m², respectively, and a more adverse risk factor profile with respect to lipids, blood pressure, and microalbuminuria, but lower mean HbA1c; 59.3 (SD 21.6) mmol/mol compared to 65.6 (SD 18.9) mmol/mol. During a mean follow-up of 12.2 (7.0) and 7.9 (5.6) years, for type 1 and type 2, respectively, 332 (1.0%) and 243 (2.1%) patients with type 1 and type 2, respectively, developed AMI and 142 (0.44%) and 87 (0.74%) heart failure. Cumulative incidence adjusted for age and BMI is shown in the Figure.

Conclusion: Of the two types of diabetes, type 2, as expected, had the worst cardiovascular risk factor profile and more micro/macroalbuminuria despite shorter duration. In absolute terms type 2 also had higher risk of AMI and HF even after adjustment for age and BMI.

Disclosure: P. Dikaiou: None.

171

Effects of dapagliflozin in type 2 diabetes and heart failure with mildly reduced or preserved ejection fraction across the background of glucose-lowering therapy in DELIVER

M. Vaduganathan ¹, S.E. Inzucchi², B.L. Claggett¹, I. Kulac¹, R.A. de Boer³, A.S. Desai¹, K.F. Docherty⁴, A.F. Hernandez⁵, M.N. Kosiborod⁶, P.S. Jhund⁴, C.S.P. Lam⁷, F.A. Martinez⁸, S.J. Shah⁹, J.J.V. McMurray⁴, S.D. Solomon¹;

¹Brigham & Women's Hospital, Boston, MA, USA, ²Yale University, New Haven, CT, USA, ³University Medical Center, Groningen, Netherlands, ⁴University of Glasgow, Glasgow, UK, ⁵Duke University Medical Center, Durham, NC, USA, ⁶Brigham & Women's Hospital, Kansas City, MO, USA, ⁷National Heart Centre, Singapore, Singapore, ⁸University of Cordoba, Cordoba, Argentina, ⁹Northwestern University Feinberg School of Medicine, Chicago, IL, USA.

Background and aims: Patients (pts) with T2D often require >1 medication to control their HbA1c. In DELIVER, the SGLT2i dapagliflozin (dapa) improved heart failure (HF) outcomes in 6263 pts with HF with mildly reduced or preserved ejection fraction (HFmrEF/HFpEF). Whether the effects of dapa on clinical outcomes in DELIVER were influenced by the background type or number of anti-hyperglycemic therapies (AHT) is unknown.

Materials and methods: 6263 pts with HFmrEF/HFpEF were randomized to dapa or placebo. Of these, 3150 (50.2%) had T2D. We evaluated the effects of dapa on the 1° outcome (worsening HF/CV death), HF hospitalisation, CV death, and all-cause death, using Kaplan-Meier estimates and Cox proportional hazards models. Treatment effects were assessed across the spectrum of baseline AHT number as a continuous variable.

Results: Within the T2D subgroup, 42% were female, mean age (±SD) 71±9, BMI 30.9±6.2, HbA1c 7.4±0.6%, eGFR 59.5±19.5, and left ventricular EF 54.0±8.7%. The mean AHT number at baseline was 1.3±1.0, with 22.9% on diet alone, 58.5% metformin, 30.0% insulin, 21.5% sulphonylurea (SU), 16.8% DPP4i and 2.2% GLP1RA (combination therapy being used in 40.6%.) The relative effects of dapa vs. placebo on most outcomes were consistent irrespective of background use of most AHT. 1° & key 2° outcomes in those taking vs. not taking the two most common AHT, metformin and insulin, are shown in the Figure. We observed nominal heterogeneity for the 1° outcome by SU use, with those on an SU experiencing larger benefits. However, this was not consistent across all outcomes. Furthermore, after accounting for other pre-specified subgroups and other AHT, there appeared to be no significant overall heterogeneity in the response to dapa. There was also no interaction with the background use of DPP4i or GLP-1 RA. Finally, in our analysis of pts taking different numbers of AHT (0-4), there were consistent effects of dapa across all HF outcomes.

Conclusion: In pts with T2D with HFmrEF/HFpEF, the treatment benefit of dapa was consistent across most types and numbers of AHT. These data support newer T2D treatment guidelines that no longer mandate foundational therapy with metformin. Furthermore, the benefits of dapa are not mitigated by concurrent insulin therapy, which may have sodium-retentive properties. The observed nominal statistical interaction with SU may reflect the play of chance but warrants further investigation.

Clinical Trial Registration Number: NCT03619213

Supported by: AstraZeneca

Disclosure: M. Vaduganathan: Employment/Consultancy; American Regent, AstraZeneca, Bayer, Baxter Healthcare, Boehringer Ingelheim, Cytokinetics, Lexicon, Novartis, Pharmacosmos, Relypsa, Roche Diagnostics, Sanofi, Galmed, Bayer, Occlutech, Impulse Dynamics. Lecture/other fees; AstraZeneca, Novartis, Roche.

172

Association of clinical characteristics with sudden cardiac arrest in people with type 2 diabetes with and without CVD: a longitudinal case-control study on primary care data

P.P. Harms ^1,2, L.H. van Dongen^3,4, F.C. Bennis⁵, K.M.A. Swart⁶, M. Hoogendoorn⁵, J.W.J. Beulens^7,2, H.L. Tan^3,8, P. Elders^1,2, M.T. Blom^1,2, RESCUED;

¹General Practice Medicine, Amsterdam UMC, Amsterdam, Netherlands, ²Amsterdam Public Health research institute, Amsterdam, Netherlands, ³Clinical and Experimental Cardiology, Amsterdam UMC, Amsterdam, Netherlands, ⁴Amsterdam Cardiovascular Sciences, Amsterdam, Netherlands, ⁵Computer Science, Vrije Universiteit Amsterdam, Amsterdam, Netherlands, ⁶PHARMO institute for Drug Outcomes Research, Utrecht, Netherlands, ⁷Epidemiology and Data Science, Amsterdam UMC, Amsterdam, Netherlands, ⁸Netherlands Heart Institute, Utrecht, Netherlands.

Background and aims: Sudden cardiac arrest (SCA) is a substantial public health problem and the causative event of up to 50% of cardiac deaths, accounting for 20% of mortality in high-income countries. People with type 2 diabetes have a twofold increased risk of SCA. However, predictors of SCA risk are not fully understood, specifically for those without a CVD history. We aimed to assess the (differences in) longitudinal associations of clinical characteristics recorded in primary care with SCA in people with type 2 diabetes with and without a CVD history.

Materials and methods: A case-control study with SCA cases with type 2 diabetes from the AmsteRdam REsuscitation STtudies (ARREST) registry of out-of-hospital resuscitation attempts by emergency medical services in the Dutch region of Noord-Holland (2005-2019). We included cases with presumed cardiac cause registered at General Practice (GP) practices from the PHARMO Data Network and the academic network of general practice at Amsterdam UMC (ANHA). Cases were matched (age, sex, T2D, GP practice) with up to five non-SCA controls. From the GP files, we collected relevant clinical measurements, medication use, and medical history from before the SCA date of the case-control pairs. These measurements were grouped by year until 5 years prior the case’s SCA date, and then categorized including a 'unknown' category for measurements that were not recorded that year. We analysed the associations of clinical characteristics with SCA in the total sample and in subgroups with and without CVD using univariable and multivariable time-dependent Cox-regression (Hazard Ratios, 95% confidence intervals).

Results: We included 815 cases and 3,696 controls. In multivariable models, only insulin use (2.01 (1.56-2.60)) was associated with SCA in both people with and without CVD. In people with CVD (409 cases/1,392 controls), moderate (1.84 (1.19-2.84)), severe (1.96 (1.05-3.66)) and unknown (1.80 (1.21-2.68)) albuminuria, and heart failure (2.22 (1.78-2.76)) were associated with SCA. In people without CVD (406 cases/2,304 controls), current (1.68 (1.28-2.20)) and unknown (1.61 (1.30-2.00)) smoking behaviour, high LDL cholesterol (>2.6 mmol/l: 1.54 (1.00- 2.37)), and QTc prolonging anti-psychotic (2.52 (1.68- 3.77)), prokinetic (1.94 (1.23- 3.05), and anti-biotic medication use (1.41 (1.02- 1.97) were associated with SCA.

Conclusion: In people with type 2 diabetes with CVD, albuminuria and heart failure are signs of increased SCA risk, while in people without a CVD history current and unknown smoking behaviour, high LDL cholesterol, and QTc prolonging medication use are indicators of SCA risk.

Supported by: Dutch Heart Foundation grant CVON2017-15 RESCUED, EU H2020 grant no 733381 ESCAPE-NET, COST Action PARQ grant No CA19137

Disclosure: P.P. Harms: None.

173

Duration of type 2 diabetes increases risk of third-degree atrioventricular block

S. Haxha ^1,2, K.K. Sørensen², S.B. Haugaard^3,4, M.H. Ruwald⁵, A. Halili^1,2, B.T. Philbert⁶, L. Køber⁶, G.H. Gislason⁵, C. Torp-Pedersen², C.N. Bang¹;

¹Department of Cardiology, Bispebjerg og Frederiksberg Hospital, Copenhagen, Denmark, ²Department of Cardiology, North Zealand Hospital, Hillerød, Denmark, ³Department of Endocrinology, Bispebjerg og Frederiksberg Hospital, Copenhagen, Denmark, ⁴Institute of Clinical Medicine, Faculty of Health and Medical Sciences,, Copenhagen, Denmark, ⁵Department of Cardiology, Herlev-Gentofte University Hospital, Copenhagen, Denmark, ⁶Department of Cardiology, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark.

Background and aims: Type 2 diabetes mellitus (T2DM) is a major contributor to morbidity and mortality, with an increased risk of sudden cardiac death. Duration of T2DM is associated with a higher rate of pacemaker implantation, which suggests that the duration of T2DM may be related to third-degree atrioventricular block (3AVB) as an indication of pacemaker. We aimed to investigate, whether duration of T2DM is associated to greater incidence of 3AVB.

Materials and methods: This is a nested case-control study, based on prospectively collected data in nationwide registries in Denmark. All 5 342 797 inhabitants of Denmark older than 18 years between 1995-2018, from which 24 177 nested cases with 3AVB were identified and matched randomly 1:5 at the risk set of each case on sex and birth year with 120 915 controls without 3AVB. Conditional logistic regression with time-dependent exposure and time-dependent covariates was used to calculate hazard ratios for 3AVB with duration of T2DM association split into four groups: 0-5 years, 6-10 years, 11-15 years, and >15 years. Multivariate analyses were adjusted for relevant comorbidities and medicine.

Results: T2DM duration 0-5 years, 6-10 years, 11-15 years and >15 years were associated with a stepwise increase in higher hazard ratio of third-degree atrioventricular block of 1.24 (95% confidence interval 1.16 to 1.33), 1.41 (95% confidence interval 1.31 to 1.52), 1.79 (95% confidence interval 1.64 to 1.95), and 2.20 (95% confidence interval 2.01 to 2.40), respectively, compared to patients free of diabetes, after adjusting for atrioventricular nodal blocking agents and comorbidities known to be associated with 3AVB. (Figure 1.)

Conclusion: This study shows a stepwise increase in the hazard ratio of 3AVB with T2DM duration, independent of atrioventricular nodal blocking agents and comorbidities known to be associated with 3AVB. These findings add to the increasing body of knowledge on T2DM being independently associated with 3AVB, suggesting that physicians need to be more vigilant of the cardiac conductions system in people with T2DM, especially when presenting with relevant symptoms in combination with an extensive history of T2DM.

Supported by: This work was funded by the independent research foundation Skibsreder Per Henriksen, R. og Hustrus Fond.

Disclosure: S. Haxha: Grants; Educational grant from Medtronic covering conference fee only to attend European Heart Rhythm association (EHRA) Congress 2022 in Copenhagen.

174

Glycaemia and cardiac arrhythmias in people with type 1 diabetes: a prospective observational study

P.G. Hagelqvist ¹, A. Andersen¹, K. Maytham¹, C.R. Andreasen¹, S. Engberg², T. Lindhardt³, J.L. Forman⁴, U. Pedersen-Bjergaard⁵, F.K. Knop⁶, T. Vilsbøll¹;

¹Clinical Research, Steno Diabetes Center Copenhagen, Herlev, Denmark, ²Steno Diabetes Center Copenhagen, Herlev, Denmark, ³Department of cardiology, Nordsjællands Hospital Hillerød, Hillerød, Denmark, ⁴Department of biomedical sciences, University of Copenhagen, Copenhagen, Denmark, ⁵Department of Endocrinology and Nephrology, Nordsjællands Hospital Hillerød, Hillerød, Denmark, ⁶Center for Clinical Metabolic Research, Gentofte Hospital, Hellerup, Denmark.

Background and aims: The impact of hypoglycaemia, hyperglycaemia and glycaemic variability on cardiac arrhythmia susceptibility in people with type 1 diabetes is uncertain. We performed a one-year prospective observational study employing continuous glucose monitoring and implantable loop recorders to investigate potential associations between glycaemia and cardiac arrhythmias.

Materials and methods: Thirty adults with type 1 diabetes ([mean ± SD] age 63 ± 8 years, BMI 26 ± 5 kg/m², HbA1c 57 ± 12 mmol/mol [7.3 ± 1.1%]) and with no history of cardiac arrhythmias were included. Daytime and night-time incidence rate ratios (IRR) for cardiac arrhythmias were determined both for hypoglycaemia (interstitial glucose (IG) <3.9 mmol/l) and hyperglycaemia (IG >10.0 mmol/l) in comparison with euglycaemia (IG 3.9 - 10.0 mmol/l). Furthermore, we explored the risk of arrhythmias during an increase in IG above the subject-specific mean IG and during increased glycaemic variability as estimated by coefficient of variation (CV).

Results: Hypoglycaemia was not associated with cardiac arrhythmias when compared to a combination of euglycaemia and hyperglycaemia (Fig.1A+B). However, during daytime the subject-specific hourly incidence of arrhythmias tended to increase with increasing time spent in hypoglycaemia compared to euglycaemia (Fig. 1A). Daytime hyperglycaemia was associated with increased risk of arrhythmias compared to euglycaemia (Fig. 1A+B). No association was found between nighttime hypoglycaemia or nighttime hyperglycaemia and risk of arrhythmias. A daytime increase in IG above the subject-specific mean IG showed an increased risk of arrhythmias (IRR 1.10 [ 95% CI 1.03 - 1.17] per 1 mmol/l increase). During daytime, no association was found between increased CV and the risk of arrhythmias (IRR 0.91 (95% CI [0.79 - 1.05] per 5% increase), whereas during nighttime an inverse relationship was observed (IRR 0.81 [95% CI 0.69 - 0.96] per 5% increase).

Conclusion: The present study suggests that acute hypoglycaemia and hyperglycaemia during daytime may contribute to an increased risk of arrhythmias in individuals with type 1 diabetes. Since no similar link was found during nighttime, this may indicate diurnal differences in arrhythmogenic susceptibility.

Clinical Trial Registration Number: NCT04011683

Supported by: Novo Nordisk Foundation

Disclosure: P.G. Hagelqvist: None.

OP 30 Years and years

175

Effect of low dose aspirin on incident diabetes among older adults: post hoc analysis of the ASPREE randomised placebo-controlled trial

S. Zoungas ¹, Z. Zhou¹, A.J. Owen¹, A.J. Curtis¹, S.E. Espinoza², M.E. Ernst³, R.L. Woods¹, S.G. Orchard¹, J.J. McNeil¹, A.M. Murray⁴, M.R. Nelson⁵, C.M. Reid⁶, J. Ryan¹, R. Wolfe⁷;

¹School of Public Health and Preventive Medicine, Monash University, Melbourne, Australia, ²Sam and Ann Barshop Institute, UT Health San Antonio, San Antonio, TX, USA, ³College of Pharmacy, The University of Iowa, Iowa, IA, USA, ⁴Berman Centre for Outcomes and Clinical Research, Hennepin Healthcare Research Institute, Minneapolis, MN, USA, ⁵Menzies Institute for Medical Research, University of Tasmania, Hobart, Australia, ⁶School of Public Health, Curtin University, Perth, Australia, ⁷School of Public Health and Preventive Medicine, Monash University, Melbourne, VIC, Australia.

Background and aims: The effect of aspirin on incident type 2 diabetes among older adults remains uncertain. This study investigates the randomised treatment effect of low dose aspirin on incident diabetes and fasting plasma glucose (FPG) levels among older adults.

Materials and methods: ASPREE was a double-blind, placebo-controlled trial of aspirin. Community-dwelling individuals aged ≥65 years and free of cardiovascular disease, independence-limiting physical disability and dementia were enrolled. Participants were randomised 1:1 to 100 mg daily enteric-coated aspirin or placebo. Incident diabetes was defined as self-report of diabetes, commencement of glucose lowering medication and/or FPG ≥7.0 mmol/L at annual follow-up visits. Presence of diabetes at baseline was an additional exclusion criterion for this analysis. Cox proportional-hazards and linear mixed-effects regression models assessed the effect of aspirin on incident diabetes and FPG levels respectively.

Results: 16,209 participants were included in the analysis (8,086 randomised to aspirin and 8,123 to placebo). Over a median follow-up of 4.7 years, 995 incident diabetes cases were recorded (aspirin: 459, placebo: 536). Compared with placebo, the aspirin group had a 15% reduction in incident diabetes (hazard ratio 0.85, 95% CI 0.75-0.97; p=0.01) and a slower rate of increase in FPG (difference [95% CI] in annual FPG change: -0.006 mmol/L [-0.009 to -0.002], p=0.004).

Conclusion: Aspirin treatment reduced incident diabetes and slowed the increase in FPG over time among initially healthy older adults. Given the increasing prevalence of type 2 diabetes among older adults, the potential for anti-inflammatory agents like aspirin to prevent type 2 diabetes or improve glucose levels needs further study.

Clinical Trial Registration Number: NCT01038583

Supported by: NIA & NIH (U.S.), NHMRC (Australia)

Disclosure: S. Zoungas: None.

176

Incident type 2 diabetes among users of denosumab and alendronate: a Danish cohort study

R. Viggers ^1,2, J. Starup-Linde^3,4, I. Leegaard¹, P. Vestergaard^1,2;

¹Steno Diabetes Center North Denmark, Aalborg University Hospital, Aalborg, Denmark, ²Department of Clinical Medicine, Aalborg University, Aalborg, Denmark, ³Steno Diabetes Center Aarhus, Aarhus University Hospital, Aarhus, Denmark, ⁴Department of Endocrinology and Internal Medicine, Aarhus University Hospital, Aarhus, Germany.

Background and aims: Alendronate and denosumab are two commonly used anti-resorptive therapies in the management of osteoporosis. Alendronate is a bisphosphonate and is suggested as a protecting agent against the development of type 2 diabetes (T2D). Denosumab is an antibody against the receptor activator of nuclear factor-κB ligand (RANKL) and higher RANKL levels are reported associated with an increased risk of T2D. We aimed to compare initiators of denosumab and alendronate and hypothesized no difference in the risk of incident T2D.

Materials and methods: We conducted a population-based Danish cohort study by accessing all discharge diagnoses (ICD-10 system) from the National Danish Patient Registry and all redeemed drug prescriptions (ATC classification system) from the Health Service Prescription Registry. The main cohort consisted of anti-osteoporotic treatment naïve users of alendronate or denosumab between 2011-2018 without any history of diabetes and age ≥ 50 years at index. We matched 5 alendronate users to 1 denosumab user by propensity score (ps) matching with replacement. The outcome was incident T2D. Incidence rates and cox proportional hazard risk ratios (HRs) for T2D were estimated with alendronate as the comparator and with 95% CIs. We assessed exposure time, total defined daily doses (DDDs), and drug adherence based on the medication possession ratio (MPR). Evaluation of response relationships was assessed by interaction terms and predicted HRs.

Results: After ps matching, the cohort consisted of 3,361 denosumab and 12,815 alendronate users. In total, 544 subjects developed T2D during the study period: 101 (3.01%) denosumab users and 443 (3.46%) alendronate users. The incidence rates per 1000 person-years were 11.72 (10.67; 12.86) and 9.45 (7.78; 11.49) for alendronate and denosumab, respectively. The drug adherence (MPR ≥ 0.8) was high both among denosumab (63.46%) and alendronate (83.53) users. The HR for developing T2D among denosumab users was 0.82 (0.66; 1.03). When restricting the analysis to those with MPR ≥ 0.8, the HR for developing T2D was 0.55 (0.41; 0.74) compared to alendronate users. Figure 1 illustrates the predicted HRs for T2D stratified by exposure time and DDDs.

Conclusion: These results suggest a lower incidence and risk of T2D among denosumab compared to alendronate users, though the estimates did not reach statistical significance. It is possible that denosumab and alendronate have short- and long-term protective effects on incident T2D, respectively. Lastly, high adherence to denosumab may decrease the development of T2D. With these findings, we hope to encourage further clinical research in the crosstalk between bone and glucose metabolism.

Supported by: This work was supported by Steno Collaborative grant, Novo Nordisk Foundation, Denmark (Grant no. NNF18OC0052064).

Disclosure: R. Viggers: None.

177

Prescribing trends of anti-diabetes medications near end of life among older adults with type 2 diabetes: a nationwide cohort study

A. Kutz, D.H. Kim, E. Patorno;

Department of Medicine, Brigham and Women`s Hospital, Boston, MA, USA.

Background and aims: Although deprescribing anti-diabetes medications towards end of life has been suggested to be safe data on the extent, patterns and secular changes in discontinuation of these agents near end of life is still scarce. We therefore assessed prescribing trends of anti-diabetes medications in the last year of life among older adults with type 2 diabetes and explored whether frailty is associated with differential prescribing patterns.

Materials and methods: Using nationwide claims data from Medicare fee-for-service beneficiaries, we conducted a cohort study including older adults (≥67 years) with type 2 diabetes who died between 2015 and 2019. We plotted 6-monthly trends in prevalent and incident users of anti-diabetes medications overall and stratified by level of frailty. We also assessed differences in prescribing patterns within the last year of death (9 to 12 months vs 0 to 4 months before death).

Results: Among 975,407 Medicare beneficiaries with type 2 diabetes (mean [SD] age, 80.9 [8.2] years; 54.3% female; 34.9% frail), filled prescriptions for any anti-diabetes medication within 1 year of death increased from 71.4% in 2015 to 72.9% in 2019 (p<0.01). The most pronounced increase in use was observed for metformin (40.7% in 2015 to 46.5% in 2019, p<0.01), whereas the use of sulfonylureas decreased over time (37.0% to 31.8%, p<0.01). The proportion of people on any anti-diabetes medication decreased throughout the last year of life from 66.1% in the 9-12 months before death to 60.8% (p<0.01) in the 0-4 months before death. This decline was mainly driven by a decrease in use of non-insulin medications, whereas short-acting insulins were more likely to be prescribed towards the end of life (from 18.5% to 20.0%, p<0.01). Prescribing trends among incident users of anti-diabetes medications were similar, with short-acting (4.1% to 7.4%, p<0.01) and long-acting insulin (6.0% to 8.2%, p<0.01) more frequently initiated closer to end of life. Trends by level of frailty were similar, with a larger proportion of more frail individuals using short- and long-acting insulin near end of life. Newer medications with proven cardiovascular benefit were the least frequently prescribed agents at end of life, though their use became more frequent in recent years.

Conclusion: In this cohort study, the proportion of older adults with type 2 diabetes who used an anti-diabetes medication within 1 year of death increased over time. However, there was evidence of an increase in medication discontinuation towards the end of life, in accordance with practice guidelines.

Disclosure: A. Kutz: Grants; Postdoc. Mobility grant (P400PM_194479 / 1) from the Swiss National Science Foundation, Educational grant from Novo Nordisk.

178

Age discrepencies in the use of glucose-lowering medications: a nationwide cohort study

K. Johansson ¹, T. Petersen¹, M. Christensen^1,2;

¹Department of clinical pharmacology, Bispebjerg and Frederiksberg Hospital, Copenhagen, Denmark, ²Copenhagen Center for Translational Research, Bispebjerg and Frederiksberg Hospital, Copenhagen, Denmark.

Background and aims: Various glucose-lowering medications are used to lower glycated haemoglobin (HbA1c) in patients with type 2 diabetes. Of the available medications, only glucagon-like peptide-1 receptor agonists (GLP-1RAs) and sodium-glucose co-transporter-2 inhibitors (SGLT-2is) have documented preventive effects on cardiovascular disease events within a few years of treatment. For older patients, guidelines recommend a less strict HbA1c-target and discourage treatment with sulphonylureas due to their inherent risk of hypoglycaemia. We hypothesized that these guidelines are yet to be fully implemented and that older patients receive fewer and other glucose-lowering medications than younger patients.

Materials and methods: We described the usage of glucose-lowering medication for all patients with type 2 diabetes in Denmark from 2018 to 2020 and related this to age and Hba1c by using regression models adjusting for: sex, renal function, comorbidities, level of education, gross household income and country of birth. A patient was considered a medication user if they redeemed a prescription of that medication class (fourth level ATC code) within three months following their Hba1c measurement.

Results: We identified 290,890 patients with type 2 diabetes in the study period. The number of medications used peaked at the age of 70, whereafter it decreased (Figure 1A). For example, when comparing 80- and 90-year-old patients with 50-years-old patients at an Hba1c value of 48 mmol/mol (6.5%), 80-year-old patients used 5% (95% CI 3-6%) fewer glucose-lowering medications and 90-year-old patients 31% (95% CI 30-33%) fewer glucose-lowering medications. Older patients were generally less likely to be treated with GLP-1RAs and SGLT-2is (Figure 1B). In contrast, they were more likely to be treated with insulin, dipeptidyl peptidase-4 inhibitors, and sulphonylureas. These discrepancies were attenuated with an increasing Hba1c. For example, compared with a 50-year-old patient, an 80-year-old patient with an Hba1c-value of 48 mmol/mol (6.5%) was 59% less likely to receive either a GLP-1RA or a SGLT-2i (RR 0.41, 95% CI 0.38-0.44), but 137% more likely to receive sulphonylureas (RR 2.37, 95% CI 2.19-2.58).

Conclusion: In Denmark, older patients were treated with fewer glucose-lowering medications than younger patients at the same Hba1c-level. Older patients were also more likely to receive medications with a higher risk of hypoglycaemia, such as sulphonylureas and insulin, and less likely to use GLP-1RA and SGLT-2i. Our findings suggest a potential for treatment improvement and stress the importance of frequently re-evaluating glucose-lowering treatment in all patients, independent of age.

Disclosure: K. Johansson: None.

179

UKPDS blood pressure study 34 year follow-up

R.L. Coleman ¹, A.I. Adler¹, P. Clarke², J. Leal², W. Whiteley^3,4, R.R. Holman¹;

¹Radcliffe Department of Medicine, University of Oxford, Oxford, UK, ²Nuffield Department of Population Health, University of Oxford, Oxford, UK, ³University of Edinburgh, Edinburgh, UK, ⁴University of Oxford, Oxford, UK.

Background and aims: The UK Prospective Diabetes Study (UKPDS) demonstrated that individuals with hypertension and diabetes randomly allocated to tight vs. less tight blood pressure control experienced a clinically important reduction in diabetes-related outcomes including microvascular disease, stroke and diabetes-related death during a median 8.4 years follow-up. These differences attenuated during the 10-year post-trial monitoring phase, with risk reductions no longer achieving clinical or statistical significance, although the risk reduction for peripheral vascular disease (PVD) became statistically significant (Hazard Ratio 0.50, 95%CI 0.28-0.92, P=0.02). We used NHS administrative data to extend follow-up for the surviving cohort to a maximum of 34 years.

Materials and methods: Participants who completed post-trial monitoring in September 2007 were linked to routinely-collected NHS England administrative data, extending potential follow-up to March 2021, and increasing person-years of follow-up by 16% to 18,084 years. Data for Scotland and Northern Ireland are awaited. We re-evaluated the 7 pre-specified UKPDS aggregate outcomes in intention-to-treat analyses using Kaplan Meier time-to-event and log rank analyses.

Results: After a median 15.8 years follow-up from randomization, 783 (86%) of the 911 participants whose vital status was known in 2021 had died. The number of additional events collected during the 14-year administrative follow-up period were evenly distributed between randomized groups. Overall, there were no clinical or statistically significant differences between the groups allocated previously to tight vs. less-tight blood pressure control for any of the aggregate outcomes: Any diabetes-related endpoint (Risk Ratio [RR] 0.93, 95%CI 0.80-1.08 P=0.32); Diabetes-related death (RR 0.91, 95%CI 0.74-1.11, P=0.34); All-cause mortality (RR 0.94, 95%CI 0.81-1.09, P=0.39); Myocardial infarction (RR 0.96, 95%CI 0.78-1.18, P=0.68); Stroke (RR 0.79, 95%CI 0.58-1.06, P=0.11); PVD (RR 0.59, 95%CI 0.34-1.03, P=0.062); and Microvascular disease (RR 0.87, 95%CI 0.67-1.12, P=0.28).

Conclusion: Longer-term analysis of the UKPDS cohort shows no persisting or legacy effects of the within-trial benefits of tight blood pressure control, or evidence of harm. This study highlights the potential use of administrative data to provide extended follow-up of clinical outcome studies.

Disclosure: R.L. Coleman: None.

180

MODY in older onset diabetes is common and identification can improve treatment: analysis of >72,000 people

L. Sharp ¹, L. Mirshahi², K. Colclough¹, J. Haley², A. Hattersley¹, D. Carey², M. Weedon¹, K. Patel¹;

¹University of Exeter, Exeter, UK, ²Geisinger, Danville, PA, USA.

Background and aims: MODY genetic testing currently excludes people with diabetes onset after 40y, despite studies showing MODY mutations can be found in older adults. We therefore performed large scale genetic studies of 72,557 people to identify the prevalence, clinical features, and clinical utility of monogenic diabetes genetic testing in people with older onset diabetes (diagnosed after >40y).

Materials and methods: We used whole exome sequence data from people with diabetes diagnosed >40y. This included 38,967 from the UK Biobank, a population cohort from the UK and 33,590 people from a health care system-based cohort from the USA with higher rate of obesity (50%). We analysed the 12 most common MODY genes including m.3243A>G (only in UK Biobank) and identified pathogenic/likely pathogenic variants in these genes. We then compared the clinical features of people with MODY versus those with non-MODY diabetes.

Results: Out of the 38,967 people diagnosed with diabetes after the age of 40 in the UK cohort, 188 people were found to have MODY, giving a prevalence of 0.48% (1 in 207, CI 0.42-0.56%). The prevalence was similar in people with insulin treated and non-insulin treated diabetes (0.50% vs 0.47%). The most common cause of MODY was GCK (44.7%) followed by HNF1A/4A (17%) and m.3243A>G (10.6%). Of 188 people with MODY, 48 (25%) received inappropriate treatment for their genotype. MODY cases had a lower age of diagnosis (55 vs 58, p<10^-07), lower BMI (28.1 vs 31.6, p<10^-17), higher frequency of gestational diabetes (females only, 11% vs 2%, p<10^-5) and lower T2D-GRS (0.04 vs 0.33, p<10^-4). The prevalence of MODY doubled to 1.1% (1 in 91) for people with BMI <27 (<median of the cohort). Analysis of 33,590 from a health care-based cohort with higher background rates of obesity (50%) showed lower prevalence of MODY (n=81, 0.25%,1 in 394, CI 0.2-0.3%). Of 81 individuals with MODY, 40 (50%) received inappropriate treatment for their genotype.

Conclusion: At least one in 207 people with older onset diabetes (>40y) have MODY. 25%-50% of MODY individuals have received inappropriate treatment for their genotype, highlighting the clinical utility of genetic testing. Further studies are needed to develop an optimal strategy to select older adults for genetic testing.

Supported by: Wellcome Trust

Disclosure: L. Sharp: None.

OP 31 Future for type 1 diabetes treatment

181

Exploring the role of transcription factors HNF1A, HNF1B, and HNF4A in human induced pluripotent stem cell-derived pancreatic islet cell differentiation

L. Unger, A. Friestad, T.A. Legøy, L.M. Ghila, S. Chera;

University of Bergen, Bergen, Norway.

Background and aims: Induced pluripotent stem cells (iPSCs) have become an important tool for diabetes research due to their ability to differentiate into various cell types, including pancreatic beta cells. These beta cells are responsible for producing insulin in the pancreas, and their dysfunction or loss is a hallmark of diabetes. iPSCs are a valuable tool for both diabetes research and potential treatment options, providing new avenues for understanding and addressing this disease. Monogenic diabetes is a form of diabetes caused by mutations in a single gene. One of the most common genes associated with monogenic diabetes is HNF1A, which codes for a transcription factor that plays a crucial role in regulating glucose metabolism in the pancreas. Mutations in HNF1A result in impaired insulin secretion and reduced beta cell function, leading to the development of diabetes. Previous studies suggest that HNF1A acts as critical genetic switch controlling pancreatic insulin-producing β-cell differentiation. Genetic cell tracing, timed conditional gene expression and diverse dynamic omics assays revealed a critical role of HNF1A in islet cell identity selection, maintenance, and crosstalk. Recent studies from our lab indicated the crucial role of the interplay between HNF1A and HNF4A and their ratio between expression level in restricting human islet cell identity. The aim of this study is to investigate the impact and time of occurrence of HNF1A, HNF1B and HNF4A along islet cell fate acquisition by using a seven-stage in vitro differentiation protocol, which successfully generated insulin-producing cells from human induced pluripotent stem cells (iPSC).

Materials and methods: In vitro differentiations were performed following previously established protocols. We introduced by CRISPR-Cas9 the hotspot mutation P291fsinsC in a commercially available iPSC line (noncarrier). Cellular analyses were performed via immunofluorescence staining, while the proteome was investigated via Tandem Mass Tag 11-plex labelling.

Results: The expression pattern for HNF1A, HNF4A and HNF1B during differentiation was established and an interplay between all the different HNFs revealed. To Investigate the effect of a MODY mutation on the expression pattern of the HNFs, an iPSC cell line carrying the MODY-3 hotspot mutation HNF1A P291fsinsC was differentiated towards pancreatic cell fate. The HNF1A P291fsinsC mutation shows an altered expression pattern and changes in morphology during pancreatic differentiation. After maturation of stem cell islets, the MODY-3 mutant showed impaired insulin production and an altered molecular landscape.

Conclusion: Using proteomics, immunofluorescent assays, glucose stimulated insulin secretion tests and bulk RNA sequencing, we were able to reveal the differences in the molecular landscapes between mature stem cell islets differentiated from MODY-3 and non-carrier iPSC.

Supported by: Novo Nordisk

Disclosure: L. Unger: Grants; Novo Nordisk.

182

Both MAFA and MAFB are required for acquisition of glucose-stimulated insulin secretion (GSIS) in human pluripotent stem cell-derived islets

A. Allouch, A. Diane, R. Abdulmuumin, H.H. Al-Siddiqi;

Qatar Biomedical Research Institute, Hamad Bin Khalifa University, Doha, Qatar.

Background and aims: Diabetes mellitus is a chronic disease characterized by impaired glucose metabolism resulting from an absolute or relative insulin deficiency. Current pharmacological treatments for type 1 diabetes involves exogenous insulin injection that cannot fully replicate rigorous glycemic control provided by endogenous β-cells. Cadaveric islet transplantation using Edmonton protocol has served as an effective intervention for restoring normoglycaemia. However, this approach is circumscribed by a serious scarcity of donor tissues. Emerging β-cell replacement with human pluripotent stem cell (hPSC)-derived β-cells may provide alternative solution. Most in vitro differentiation protocols have generated hPSC-derived β-cells with either impaired or weakened GSIS or immature phenotype lacking appropriate expression of key islet-enriched transcription factors such as MAFA and MAFB. However, their role in β-cell function is not fully characterized. This project aimed to study the expression pattern of MAFA/MAFB during in vitro differentiation and to investigate their role in the functionality of hPSC-derived β-cells.

Materials and methods: Differentiation of human embryonic (HUES8) and induced (iPSC824)-PSC into β-cells was carried out in a stepwise 2D differentiation protocol adapted from Hogrebe et al. (2021). The efficiency of relevant stage-specific marker expression as well as MAFA/MAFB expression pattern were tested using flow cytometry, immunostaining and RT-PCR. Finally, GSIS was performed following MAFA and MAFB silencing to determine their role on the functionality.

Results: Flow cytometry, immunostaining and RT-PCR data for relevant stages confirmed a directed differentiation into insulin-expressing β-cells for both cell lines. From stage 0 to stage 6, MAFA and MAFB showed differential expression pattern. GSIS data showed an increased c-peptide release in response to high (20 mM) glucose as compared with low (2 mM) glucose challenge, indicating the generation of functional β-cells. Interestingly, siRNA- mediated MAFA or MAFB suppression impaired GSIS.

Conclusion: our data demonstrated that both MAFA and MAFB are crucial for the functionality of hPSC-derived islets.

Supported by: QNRF

Disclosure: A. Allouch: None.

183

Role of arginine methylation of Neurogenin 3 in hESCs during development into the pancreatic endocrine cell in vitro

Y.-M. Han;

Graduate School of Medical Science & Engineering, KAIST, Daejeon, Republic of Korea.

Background and aims: Neurogenin3 (NGN3) is a key transcription factor in the cell fate determination of endocrine progenitors (EPs) in the developing pancreas. It is known that the stability and activity of NGN3 are regulated by serine phosphorylation. However, the role of NGN3 methylation remains elusive.

Materials and methods: We generated inducible PRMT1-knockout (P-iKO) human embryonic stem cells (hESCs) and differentiated them toward pancreatic endocrine cells (ECs) with doxycycline treatment. Using in vitro methylation assay, we found PRMT1 specifically methylates NGN3 at arginine 65. We also investigated the importance of arginine methylation on NGN3 stability and activity.

Results: Here, we report that protein arginine methyltransferase-1 (PRMT1)-mediated arginine 65 methylation of NGN3 is required for the pancreatic endocrine development of human embryonic stem cells (hESCs) in vitro. We found that inducible PRMT1-knockout (P-iKO) hESCs did not differentiate from pancreatic EPs to ECs in the presence of doxycycline. A paucity of PRMT1 led to NGN3 accumulation in the cytoplasm of EPs and decreased the transcriptional activity of NGN3. We also found that PRMT1 specifically methylates arginine 65 of NGN3, which is a prerequisite for ubiquitin-mediated degradation.

Conclusion: Our results indicate that arginine 65 methylation of NGN3 is a key molecular switch in the transition of EPs to ECs during human pancreatic development.

Supported by: Korean Fund for Regenerative Medicine (KFRM)

Disclosure: Y. Han: None.

184

Islet allotransplantation into pre-vascularised Sernova Cell Pouch™: results from the first patient cohort

M. Ogledzinski, W. Lin, S. Gondek, K. Milejczyk, B. Juengel, L. Basto, L. Perea, L.-J. Wang, M. Tibudan, R.N. Barth, J.J. Fung, P. Witkowski;

University of Chicago, Chicago, IL, USA.

Background and aims: The goal of the study was to assess safety, tolerability and effectiveness of the prevascularized Sernova Cell Pouch (SCP) for pancreatic islet transplantation.

Materials and methods: Each of 6 patients with T1DM and problematic hypoglycemia received 2 cadaveric islet transplants into SCPs pre-implanted onto the muscles of the abdominal wall. Each transplant involved open surgical access and infusion of islets into two 8-channel SCPs and one mini sentinel SCP. Patient received thymoglobulin/basiliximab for induction, tacrolimus and mycophenolate for maintenance immunosuppression.

Results: SCP implants and the combination of SCP with human cadaveric islets were both well tolerated with current durations exceeding 3 years. Wound infection resulting in device removal occurred after only one of 31 (3.2%) surgical procedures. Detectable islet function (≥0.3ng/ml peak serum c-peptide in mixed meal tolerance test) was achieved in 4 patients following islet transplant to SCP. It was correlated with transplanted islet mass of ≥3,000 IEQ/kg per procedure. Transplantation of islet mass <3,000 IEQ/kg did not lead to detectable c-peptide secretion. C-peptide was detectable for periods ranging from days up to 12 months after the procedure. Three patients with suboptimal levels of immunosuppression due to non-compliance experienced antibody mediated rejection based on de novo donor specific antibodies (DSAs). One patient with DSAs had detectable stimulated c-peptide at 90 days post-transplant. Histological assessment of sentinel SCPs retrieved ≥90 days post-transplant revealed surviving functional islets within vascularized SCP channels via positive immunofluorescence staining of insulin, c-peptide, glucagon and somatostatin in 5 of 6 patients. Five patients received supplemental intraportal islet transplants (IPITx) and all of them remain insulin independent (>3Y, >1Y, 5M, 4M, 3M). The sixth patient is awaiting an IPITx in the coming months as per protocol. Safety and dose-response observations from the first cohort of 6 patients led to the implementation of 10-channel SCPs with 50% greater transplant capacity than the 8-channel configuration in a second study cohort. Three patients were recently enrolled in the second cohort and implanted with 10-channel SCPs. Two of them are currently recovering from their first islet transplantation. Belatacept was introduced to maintenance immunosuppression in order to lower dose and toxicity of tacrolimus and myfortic and for improved protection from immunologic rejection.

Conclusion: Interim results confirm long-term safety and tolerability of Cell Pouch for pancreatic islet transplantation with histologically confirmed islet graft survival in five out of six patients.

Clinical Trial Registration Number: NCT03513939

Supported by: Sernova/ JDRF

Disclosure: M. Ogledzinski: None.

185

ADO12, a new allogenic islet encapsulation device for islet transplantation without immunosuppression

A.-L. Gaffuri ¹, A. Geissler¹, O. Jouannot¹, X. Gaume¹, C. Gautier¹, Y. Courbebaisse¹, R. Eloy¹, K. Bouzakri², O. Soula¹;

¹Adocia, Lyon, France, ²CEED, Strasbourg, France.

Background and aims: Islet transplantation is an effective cure for T1D but the requirement for immunosuppression makes its application limited. ADO12 is a new implantable and fully retrievable scaffold for islet transplantation without immunosuppression. Studies in vitro and in diabetic rat were performed to demonstrate the efficacy of the implant.

Materials and methods: Device - ADO12 is a permselective hydrogel film that allows the diffusion of insulin, prevents antibodies and immune cell invasion. It is a soft biomaterial with 95% water, synthetized from non-degradable polymers cross-linked by bio-orthogonal chemistry. It is reinforced with a suturable scaffold. In vitro encapsulation - Human islets were encapsulated in ADO12. Their functionality was evaluated using perifusion assay up to 4 months in vitro. In vivo - Diabetes was induced in outbred Wistar rats by injection of 75mg/kg of streptozotocin. Wistar rat islets were isolated and encapsulated in ADO12 at a density of 10-21kIEQ/mL in 1.6cm-diameter. 3 to 6 implants per rat were implanted in intraperitoneal and intramuscular sites for at least 41 days. Negative control diabetic rats received empty ADO12. C-peptide and glycemia were monitored. In one study, they were explanted after 2 months, and animals were monitored for at least 30 extra days. Explant viability and functionality were analyzed in vitro and by histology.

Results: Encapsulated human islets displayed a gluco-responsive insulin secretion comparable to naked islets in perifusion experiments after 7 days and 1 month from 20-50kIEQ/ml. Functionality was maintained up to 4 months after encapsulation. After implantation of ADO12 in diabetic rats, circulating insulin levels in vivo were on average 130pM/day over 41 days (n=25 rats), a significant increase versus negative control (43pM/day, n=13 rats) (N=7 studies, M.W. t-test, p<0.0001). Moreover, diabetic rats having received ADO12 (21kIEQ/mL) displayed sustained 627pM baseline-corrected C-peptide secretion for 70 days. The transplanted rats showed normal levels of fasting glycemia. After implant removal, C-peptide secretion dropped to basal levels, and rats displayed fasting hyperglycemia (FIG). Implants were well tolerated with no inflammation, nor fibrosis, nor immune cell infiltration, and were still able to secrete insulin in vitro.

Conclusion: ADO12 encapsulation system offers long-term viability and functionality of encapsulated human islets in vitro, as well as efficacy and long-term survival of islets in vivo. It is well tolerated, easily retrievable and isolating from the immune system. Upscaling to clinical application is ongoing.

Disclosure: A. Gaffuri: Employment/Consultancy; Adocia. Stock/Shareholding; Adocia.

186

Multi-component pancreatic islet cells-encapsulated microfiber grafts for treatment of diabetes

Z. Huan, L. Li;

Department of Endocrinology, Zhongda Hospital, School of Medicine, Southeast University, Nanjing, China.

Background and aims: Exogenous insulin supplementation is currently the most common treatment for type 1 diabetes, end-stage type 2 diabetes and other insulin deficiencies. However, frequent insulin injections not only cause physical discomfort to patients, but also make it difficult to avoid adverse effects such as hypoglycaemia, posing new challenges for glycaemic control in diabetes. With advances in cell transplantation and tissue engineering research, cell replacement therapy for diabetes is expected to offer new strategies for restoring endogenous insulin secretion. In order to maintain longer survival of cells transplanted into the body, more and more researches are encapsulating islet cells in hydrogels to achieve immune isolation and longer lasting graft survival.

Materials and methods: In this research, we obtained cell-laden hydrogel microfibers by applying a microfluidic approach, encapsulating a mixture of pancreatic islet alpha cells and beta cells into a sodium alginate-methacrylate-hyaluronic acid solution, collecting into a calcium chloride solution and using blue light irradiation for further solidification. Scanning electron microscopy (SEM) was used to observe the microstructure of the microfibers. Calcein AM-PI staining, CCK8 and other assays are used to detect the cytotoxicity and biocompatibility of microfibers. The permeability of the microfibers was verified using FITC-BSA. The glucose responsiveness of the α and β cells within the microfibers was demonstrated by high- and low-glucose stimulation, respectively. Cell-laden microfibers were further transplanted into diabetic mice to observe the glycaemic status and intraperitoneal glucose tolerance test (IPGTT) and intraperitoneal insulin tolerance test (IPITT) assays were performed to assess the therapeutic effect. Finally, the transplanted site tissues were sectioned and stained for HE, insulin and glucagon immunohistochemistry to see how the grafts functioned in vivo.

Results: The results showed that we could prepare continuous and controlled microfibers and that the cells could maintain good viability and proliferation within the microfibers, indicating that our hydrogels are biocompatible. The intensity of FITC-BSA fluorescence within the microfibers diminished over time, indicating that the smaller molecular weight of insulin and glucagon secreted by the cells could also be released and the pancreatic islet cells within the microfibers maintain a normal glucose responsiveness. Multi-cell laden microfibers were transplanted into diabetic model mice and achieved glycemic control for 6 weeks. Immunohistochemical staining of the grafts showed a large number of insulin and glucagon positive areas, indicating that our grafts acted as a glycemic control agent in vivo.

Conclusion: In conclusion, we have prepared biocompatible microfibers containing pancreatic α and β cells using microfluidic methods and verified its glycemic control in diabetic mice, providing new ideas and strategies for the study of islet transplantation and glycemic control.

Supported by: 82170845

Disclosure: Z. Huan: None.

OP 32 Obesity, liver and beyond

187

The effect of five days eucaloric changes in dietary carbohydrate and fat availability in the regulation of hepatic lipid content in obese men

A. London ^1,2, M.M. Richter¹, K.A. Sjøberg², N.J.W. Albrechtsen^3,4, M. Povazan⁵, L. Madsen^6,7, J. Øyen⁷, S. Madsbad¹, J.J. Holst⁸, G. van Hall⁹, H.R. Siebner^5,10, E.A. Richter², B. Kiens², A. Lundsgaard², K.N. Bojsen-Møller¹;

¹Dept. of Endocrinology, Copenhagen University Hospital, Hvidovre, Denmark, ²Sect. for Molecular Physiology, Dept. of Nutrition, Exercise and Sports, Uni. Of Copenhagen, Copenhagen, Denmark, ³Dept. of Clinical Biochemistry, Bispebjerg Hospital, Copenhagen, Denmark, ⁴NNF Center for Protein Research, Uni. of CPH, Copenhagen, Denmark, ⁵Danish Research Centre for Magnetic Resonance (DRCMR), Centre for Functional and Diagnostic Imaging and Research, Hvidovre Hospital, Hvidovre, Denmark, ⁶Laboratory of Genomics and Molecular Biomedicine, Dept. of Biology, Uni. of CPH, Copenhagen, Denmark, ⁷Institute of Marine Research, Bergen, Norway, ⁸Dept. of Biomedical Sciences, Uni. Of Copenhagen, Copenhagen, Denmark, ⁹Clinial Metabolomics Core Facility, Clinical Biochemistry, Rigshospitalet, Copenhagen, Denmark, ¹⁰Dept. of Clinical Medicine, University of Copenhagen, Copenhagen, Denmark.

Background and aims: Increased liver triacylglycerol (TG) content is associated with hepatic insulin resistance and dyslipidemia. Carbohydrate-restricted diets lower liver TG, but are often associated with weight loss which will also lower liver TG. The primary aim of this study was to investigate the effect of five days eucaloric changes in dietary carbohydrate intake on liver TG and secondarily, effects on hepatic insulin sensitivity and whole-body lipid metabolism.

Materials and methods: 11 obese normoglycemic men (BMI 31.6±3.7 (SD) kg/m², age 39±8 years, HbA1c 35±3 mmol/mol) were randomized to five days low-carbohydrate/high-fat diet (LC:14E% carbohydrate/70E% fat/16E% protein) or high-carbohydrate/low-fat diet (HC: 70E% carbohydrate/14E% fat/16E% protein) in a cross-over-design with 2-weeks washout. Liver TG was measured by ¹H-Magnetic Resonance Spectroscopy, and hepatic glucose production (HGP) was assessed by glucose tracer infusion during hyperinsulinemic euglycemic clamp (insulin infusion of 12 mU/m² body surface area/min). Indirect calorimetry was performed at fasting and during clamp.

Results: Liver TG decreased by 35±6% (mean±SE) after LC and was unchanged during HC. Body weight decreased by 1.0±0.2% during LC and changes in weight and liver TG did not correlate. The hepatic insulin sensitivity index (HISI; inverse of HGP_fasting x [C-peptide]_fasting) was increased 24±8% after LC, but was unchanged after HC. HGP during the clamp was unchanged after LC and HC. LC lowered fasting plasma TG by 35±9% and fasting respiratory exchange ratio (0.76±0.01 to 0.72±0.03), and lowered the plasma fatty acid marker of de novo lipogenesis (C16:1n-7) by 61±5%. LC increased fasting plasma ß-hydroxybutyrate by 73%±26. Fasting plasma growth differentiation factor 15 (GDF15) was 10±4% higher after LC compared with after HC. Peripheral glucose rate of disposal during clamp (Rd_clamp/[insulin]_clamp) did not differ during LC and HC.

Conclusion: In obese men, five days eucaloric substitution of dietary carbohydrates with fat lowers liver TG, improves hepatic insulin sensitivity and decreases circulating plasma TG through mechanisms involving increased fat oxidation and ketogenesis and decreased de novo lipogenesis. These results underscore that liver lipid and glucose metabolism respond rapidly to restrictions in carbohydrate intake by mechanisms which are likely independent of weight loss.

Clinical Trial Registration Number: NCT04581421

Supported by: DDA (ID PhD010-19), Independent research Fund Denmark (0134-00176B), Novo Nordisk Foundation excellence grant NNF19OC0032330

Disclosure: A. London: None.

188

Epigenetic and gene regulatory effects on liver metabolism in systemic dyslipidaemia and insulin resistance in a preclinical model

P. Fahlbusch ^1,2, A. Nikolic^1,2, N.-K. Riffelmann^1,2, S. Jacob^1,2, S. Hartwig^1,2, U. Kettel^1,2, H. Al-Hasani^1,2, J. Kotzka^1,2, B. Knebel^1,2;

¹German Diabetes Center (DDZ), Leibniz Center for Diabetes Research, Düsseldorf, Germany, ²German Center for Diabetes Research (DZD), München-Neuherberg, Germany.

Background and aims: The subtype classification of patients with type 2 diabetes also allows for prediction of the prevalence for comorbidities. In this study, we aim to identify hepatic epigenetic and gene expression signatures of the progressive severe insulin-resistant diabetes (SIRD)-cluster under minimized confounder variables using a preclinical model.

Materials and methods: The animal model used in this study is the lipodystrophic aP2-SREBP-1c mouse (aP2), which exhibits systemic insulin resistance and ectopic lipid accumulation without additional intervention. Female animals at the age of 24 vs 54 weeks (equivalent to 20-30 vs 40-50 years of human age) are investigated compared to C57Bl6 mice (C57). Hepatic energy metabolism was investigated using extracellular flux measurements in fractions of enriched liver mitochondria (n=7-8 per group). Liver biopsies were analyzed for methylated DNA content, proteome and transcriptome differential expression (n=3-5 per group). Statistical analyses of functional data were performed with Mann-Whitney test for pairwise comparisons. Omics analyses were conducted with Spectronaut® (Biognosis), Dragen™ Bio-IT platform (DESeq2-based, Illumina) and gene ontology enrichment analyses (geneontology.org), respectively.

Results: In comparison to C57, aP2 animals showed increased body weight with aggravation of the fatty liver phenotype when they age (difference C57 vs aP2 of mean liver weight/ body weight: young 1.57% vs aged 5.19%, p < 0.0001). In functional studies, a complex II-specific dysfunction of mitochondrial respiration was observed with increased oxidative phosphorylation (aP2 aged: 876 (CI 95%: 641-1110) vs C57 aged: 596 (505-687) oxygen consumption rate (OCR, pmol/min/U), p < 0.05) and decreased uncoupled respiration (aP2 aged: 399 (358-441) vs C57 aged: 772 (744-800) OCR, p < 0.001). The comparison of longitudinal development in the proteome data showed increased abundance of methionine cycle components (p<0.01) in young aP2 animals, while in aged animals components of methionine (p<0.01) and folate cycle (p<0.05) were downregulated. The analysis of the activity of regulatory enzymes of epigenetic DNA modifications showed a reduction in young aP2 livers compared to C57 (sirtuins: -17%; methyltransferase: -20%), with simultaneous hypomethylation of hepatic DNA (-30%). Consequently, hepatic gene expression changes between aP2 and C57 during aging from metabolism-related genes in the young to genes of cell division and remodeling processes in aged animals: In young animals differential expression (n=463 up/453 down) associated with regulation of fatty acid, triglyceride, cholesterol, glucose or fructose metabolic processes (all p <0.01), while in the aged animals the differentially expressed genes (n=1096 up/422 down) indicate an accumulation of chromosome condensation, kinetochore protein assembly and mitotic spindle assembly processes (all p<0.001).

Conclusion: Our data show that the metabolic status in the longitudinal perspective of the preclinical SIRD model can be attributed to an epigenetic hypomethylation of the DNA in the young with a resulting change in the expression pattern of genes of metabolism towards genes for cellular division and remodeling processes.

Supported by: DDG

Disclosure: P. Fahlbusch: None.

189

Preadipocyte-secreted Cilp1 mediates the communication of adipose with liver and promotes hepatic steatosis

Y. Bi ^1,2, S. Huang^1,2;

¹Department of Endocrinology, Endocrine and Metabolic Disease Medical Center, Affiliated Drum Tower Hospital, Medical School, Nanjing University, Nanjing, China, ²Branch of National Clinical Research Centre for Metabolic Diseases, Nanjing, China.

Background and aims: Adipose tissue, an essential endocrine organ, secretes many factors that contribute to the regulation of hepatic lipid metabolism. Preadipocytes are one of the dominant cell fractions and were previously thought to only undergo adipogenic differentiation. Recent studies have shown that preadipocytes can also secrete bioactive factors to coordinate immune and metabolic functions, and thus exert a local paracrine effect in adipose tissue. However, little is known about whether these factors can gain access to the circulation to cause an endocrine effect. Therefore, this study aimed to investigate whether preadipocyte-secreted factors could mediate inter-organ communication between adipose tissue and liver in the development of non-alcoholic fatty liver disease (NAFLD).

Materials and methods: Adoptive transfer assay was utilized to study the secretory function of adipose progenitor cell (APC) in vivo. Co-culture assay was conducted to assess the effect of APC on hepatocellular triglyceride accumulation. Single-cell RNA sequencing of human APC from visceral adipose tissue was performed to identify APC clusters and their roles in NAFLD. Serum, omental adipose tissue and liver biopsy were collected from 156 Chinese participants with or without NAFLD. Adeno-associated virus (AAV)-treated high fat-diet (HFD) mice were employed for functional study.

Results: Cell adoptive transfer and in vitro co-culture experiments demonstrated that APC from high-fat diet feeding mice or NAFLD patients could induce hepatic lipid accumulation, indicating the deleterious effects of APC on hepatic lipid metabolism in NAFLD. Further single-cell RNA sequencing data identified four APC subtypes in human visceral adipose tissues, in which CD9⁺CD55^lowAPC was found as a novel subset and increased dramatically in NAFLD. Based on fluorescence-activated cell sorting (FACS) analysis, gene expression profiling, and mass spectrometry analysis, we found cartilage intermediate layer protein 1 (Cilp1) was enriched in CD9⁺CD55^lowAPC and dramatically upregulated in the adipose tissue and plasma of NAFLD patients. After adjustment for confounders, serum Cilp1 level was the independent risk factor for NAFLD, and their levels was positively associated with the degree of liver steatosis. Furthermore, AAV-mediated Cilp1 downregulation in adipose tissue could prevent the development of HFD-induced liver steatosis and dyslipidemia, whereas Cilp1 overexpression promoted hepatic lipid accumulation in mice. At the mechanistic level, we initially found that recombinant Cilp1 protein significantly upregulated lipogenic genes in mouse primary hepatocytes.

Conclusion: We reported previously unknown mechanisms of adipose tissue-liver communication in which preadipocyte-secreted Cilp1 could promote hepatic steatosis in obesity. This study provided a new insight into the understanding of inter-organ mechanisms of hepatic lipid homeostasis, and targeting Cilp1 might provide a promising strategy for treating NAFLD.

Supported by: NSFC, China

Disclosure: Y. Bi: None.

190

Normal-weight visceral obesity promotes a higher 10-year atherosclerotic cardiovascular disease risk in patients with type 2 diabetes

J. Zheng ¹, Y. Hu¹, H. Xu¹, Y. Lei¹, J. Zhang², Q. Zheng³, L. Li⁴, W. Tu⁵, R. Chen⁶, Q. Guo⁷, X. Zang⁸, Q. You⁹, Z. Xu¹⁰, Q. Zhou¹¹, X. Wu¹;

¹Zhejiang Provincial People's Hospital, Hangzhou, China, ²Fenghua District Traditional Chinese Medicine Hospital of Ningbo, Ningbo, China, ³Yuhuan Second People's Hospital, Taizhou, China, ⁴Ningbo First Hospital, Ningbo, China, ⁵Shaoxing Shangyu People's Hospital, Shaoxing, China, ⁶Lishui People's Hospital, Lishui, China, ⁷The People’s Hospital of Putuo Zhoushan, Zhoushan, China, ⁸Yueqing People's Hospital, Wenzhou, China, ⁹Shaoxing People's Hospital, Shaoxing, China, ¹⁰Xianju people’s hospital, Taizhou, China, ¹¹The First Hospital of Jiaxing, Jiaxing, China.

Background and aims: Visceral obesity is associated with high cardiovascular events risk in type 2 diabetes. Whether normal-weight visceral obesity will pose a higher atherosclerotic cardiovascular disease (ASCVD) risk than BMI-defined overweight or obese counterparts with or without visceral obesity remains unclear. We aimed to investigate the relationship between general obesity and visceral obesity and 10-year ASCVD risk in type 2 diabetes patients.

Materials and methods: A total of 6997 type 2 diabetes patients who satisfied the requirements for inclusion were enrolled in this study. The participants were divided into six groups based on BMI and visceral fat area (VFA). The odd ratios for high 10-year ASCVD risk for different combinations of BMI and VFA were analyzed by stepwise logistic regression. The receiver operating characteristic (ROC) curves for diagnosing the high 10-year ASCVD risk were built, and the areas under the ROC curves were calculated. Potential nonlinear relationships between the levels of VFA and high 10-year ASCVD risk were examined with restricted cubic splines. Multilinear regression was used to identify factors affecting VFA in type 2 diabetes patients.

Results: In type 2 diabetes patients, subjects with visceral obesity had the higher 10-year ASCVD risk compared with those with normal visceral fat area, regardless of BMI category (all P<0.05). Of note, normal-weight visceral obesity patients had the highest 10-year ASCVD risk in six groups, which had more than a 2-fold or 3-fold higher OR than those who were overweight or obese according to BMI but did not have visceral obesity (OR=2.429, 95% CI: 1.413-4.175, P=0.001; OR=3.792, 95% CI: 1.897-7.579, P<0.001). The threshold of VFA for high 10-year ASCVD risk was 90 cm². Multilinear regression revealed that there were statistical difference in the effect of age (B=0.202, t=3.837, P<0.001), hypertension (B=17.070, t=14.918, P<0.001), drinking (B=9.812, t=7.491, P<0.001), fasting serum insulin (B=0.057, t=2.456, P=0.014), fasting plasma glucose (B=5.838, t=14.278, P<0.001), 2-hour postprandial C-peptide (B=0.284, t=2.424, P=0.015), triglyceride (B=3.716, t=9.144, P<0.001), total cholesterol (B=1.531, t=2.147, P=0.032), high-density lipoprotein cholesterol (B=-9.469, t=-6.141, P<0.001), and low-density lipoprotein cholesterol (B=2.513, t=2.955, P=0.003) on VFA in type 2 diabetes patients.

Conclusion: Type 2 diabetes patients with normal-weight visceral obesity had a higher 10-year ASCVD risk than BMI-defined overweight or obese counterparts with or without visceral obesity, which should initiate standardized management for ASCVD primary prevention.

Clinical Trial Registration Number: NCT04866667

Supported by: NSFC (81970714), the Joint Funds of the Zhejiang Provincial Natural Science Foundation of China under Grant No. LHDMZ23H070001

Disclosure: J. Zheng: None.

191

Glycaemic control determines the relationship of whole-body insulin resistance with liver mitochondrial capacity in obesity

S. Kahl ^1,2, K. Straßburger^3,2, G. Pacini⁴, N. Saatmann^1,2, K. Pafili^1,5, T. Sarabhai^1,5, S. Trenkamp^1,2, I. Esposito⁶, M. Schlensak⁷, F.A. Granderath⁷, M. Roden^1,5;

¹Institute for Clinical Diabetology, German Diabetes Center, Duesseldorf, Germany, ²German Center for Diabetes Research (DZD e.V.), Neuherberg, Germany, ³Institute for Biometrics and Epidemiology, German Diabetes Center, Duesseldorf, Germany, ⁴Independent Researcher, Padova, Italy, ⁵Department of Endocrinology and Diabetology, Heinrich-Heine-University Duesseldorf, Duesseldorf, Germany, ⁶Institute of Pathology, Heinrich-Heine-University Duesseldorf, Duesseldorf, Germany, ⁷Adipositas- und Refluxzentrum, Krankenhaus Neuwerk, Moenchengladbach, Germany.

Background and aims: While lower mitochondrial respiration tightly relates to insulin resistance in skeletal muscle, hepatic mitochondrial capacity can be higher in people with non-alcoholic fatty liver diseases (NAFLD), but reduced in type 2 diabetes. We hypothesized that worsening of glycemia may associate with lower hepatic mitochondrial capacity already in obese insulin resistant people without overt diabetes.

Materials and methods: We analyzed 42 humans with obesity grade 3 (BMI 50±7 kg/m², HbA1c 5.4±0.4% (35±4mmol/mol) undergoing bariatric surgery and intraoperative liver biopsy. Glucose tolerance and whole-body insulin sensitivity, given as PREDIcted M index (PREDIM), were assessed by oral glucose tolerance tests. Hepatic maximal ADP-stimulated mitochondrial respiration ([ETF/CI/CII]_P) was measured by high-resolution respirometry. All regression analyses were adjusted for age, sex and BMI.

Results: Applying simple regression analysis did not detect an association of PREDIM with [ETF/CI/CII]_P (p=0.5339). Further, interaction analyses revealed that the regression coefficient of [ETF/CI/CII]_P was depending on fasting blood glucose (FBG, p for interaction 0.0049; figure 1). Respective slopes were negative for FBG ≤88 mg/dl, but positive for FBG >88 mg/dl. Also, liver fat content interacted with the association of PREDIM and [ETF/CI/CII]_P (p=0.0440). However, when combining both interactions in one model, only the interaction with FBG remained significant. Performing regression analyses in eu- and dysglycemic subgroups revealed a negative association of PREDIM with [ETF/CI/CII]_P in euglycemia (p=0.0396), but not in dysglycemia (p=0.8564).

Conclusion: Onset of dysglycemia disrupts the adaptation of hepatic mitochondrial capacity to prevailing insulin resistance in people with grade 3 obesity. These findings support a role for antihyperglycemic treatment to preserve hepatic mitochondrial integrity and thereby liver health in this collective.

Clinical Trial Registration Number: NCT01477957

Supported by: The research of the authors is supported in part by grants from the German Federal Ministry of Health (BMG), the Ministry of Culture and Science of th

Disclosure: S. Kahl: Grants; German Diabetes Association (DDG; Menarini Projektförderung; Hans-Christian-Hagedorn-Projektförderung).

192

Hepatology referral of outpatients with type 2 diabetes and suspected NAFLD: comparison between EASL and AGA algorithms

T. Vidal Trecan ¹, L. Castera², T. El Khoury², J.-B. Julla¹, J.-P. Riveline¹, V. Paradis³, D. Valla², J.-F. Gautier¹;

¹Centre Universitaire du Diabete et de ses Complications, Lariboisiere Hospital, Paris, France, ²Hepatology, Beaujon Hospital, Clichy, France, ³Anatomopathology, Beaujon Hospital, Clichy, France.

Background and aims: There is a high prevalence of NAFLD (non alcoholic fatty liver disease) among patients with type 2 diabetes (T2D). Advanced fibrosis is the main prognostic driver associated with overall mortality. Recently, two algorithms have been proposed by EASL and American gastroenterological association (AGA) to screen patients with T2D at risk for NAFLD and advanced fibrosis and need a referral to a hepatologist. Screening of patients with T2D, who are the most at risk of advanced fibrosis is an unmet need. We aimed to compare the performance of these two algorithms in term of hepatology referral in a large cohort of T2D outpatients seen in a diabetes clinic.

Materials and methods: 1572 T2D patients seen in our diabetes clinic between October 2019 and February 2022, were systematically screened for NAFLD (steatosis and/or elevated ALT) and underwent FIB-4 and liver stiffness measurement (LSM) by vibration-controlled transient elastography (VCTE). Excessive alcohol consumption and viral hepatitis B and C were excluded. According to EASL algorithm, patients with FIB-4 ≥ 1.3 should undergo VCTE and those with LSM ≥ 8 kPa should be referred to a hepatologist. According to AGA algorithm, patients with FIB-4 > 2.67 or with FIB-4 between 1.3 and 2.67 and LSM ≥ 8 kPa should be referred to a hepatologist.

Results: The characteristics of the 1572 patients were as follows : male 60%; median age 61 years; BMI 29 kg/m2; waist circumference 103 cm; HbA1c 7.6%; AST 28 IU/L; ALT 28 IU/L. A liver biopsy (LB) was performed in 163 patients (advanced fibrosis 30%). Comparison of the two algorithms is shown in the Figure. According to EASL, 186 (12%) should be referred to a hepatologist (LSM ≥ 8 kPa). The 1386 remaining patients were considered at low risk. According to AGA, 227 patients (15%) should be referred (132 at high risk (FIB-4 >2.67 or LSM > 12 kPa) and 95 at intermediate risk). The 1345 remaining patients were considered at low risk. Among the 163 patients with LB, the rate of false positive was higher than that of false negative and similar between algorithms (38%, and 15%, respectively).

Conclusion: EASL and AGA algorithms performed similarly allowing to identify 12 to 15% of intermediate - high risk T2D patients, who should be referred a hepatologist. These numbers are higher than reported in the general population, highlighting the urgent need for screening these patients.

Disclosure: T. Vidal Trecan: None.

OP 33 Are we what we eat?

193

The effect of a dietary intervention on insulin requirements in type 1 diabetes: a 12-week randomised clinical trial

H. Kahleova ¹, T. Znayenko-Miller¹, K. Smith¹, R. Holubkov², N. Barnard¹;

¹Physicians Committee for Responsible Medicine, Washington, DC, USA, ²University of Utah, Salt Lake City, UT, USA.

Background and aims: A plant-based has been shown to improve insulin resistance and glycemic control and reduce medication requirements and in people with type 2 diabetes. The aim of this study was to compare the effect of this diet with a portion-controlled diet in people with type 1 diabetes.

Materials and methods: Fifty-eight people with type 1 diabetes were randomly assigned to follow a low-fat vegan diet (vegan group, n=29), or a portion-controlled diet (portion-controlled group, n=29) for 12 weeks. The content of all meals was recorded using the Cronometer mobile application. Insulin sensitivity was assessed, using the carbohydrate to insulin ratio, which was calculated as the number of grams of total dietary carbohydrate to total units of insulin administered. Total insulin dose was calculated as a sum of basal and bolus insulin units injected per day.

Results: The consumption of carbohydrates increased in the vegan group by an average of 118 g/day (p=0.002), compared with no significant change in the portion-controlled group (-18 g/day; p=0.11; treatment effect +137 [95% CI, +65 to +208]; p<0.001). Insulin sensitivity increased in the vegan group by 6.2 on average (p=0.002), compared with no significant change in the portion-controlled group (-0.8; p=0.32; treatment effect +7.0 [95% CI, +3.1 to +10.9]; p=0.001). Total daily dose of insulin decreased by 28%, specifically by 12.1 units/day in the vegan group (p=0.007), compared with no significant change in the portion-controlled group (-1.4 units/day; p=0.66; treatment effect -10.7 [95% CI, -21.3 to -0.2]; p=0.046). HbA1c decreased by 0.8% in the vegan group (p<0.001) and by 0.6% in the portion-controlled group (p=0.002; treatment effect -0.2 [-0.7 to +0.2]; p=0.34). Total cholesterol decreased by 32.3 mg/dL in the vegan group (p<0.001) and by 10.9 mg/dL in the portion-controlled group (p=0.03; treatment effect -21.4 mg/dL (-35.6 to -7.2; p=0.004). LDL cholesterol decreased by 18.6 mg/dL in the vegan group (p<0.001) and did not change significantly in the portion-controlled group (treatment effect -9.1 mg/dL (-22.6 to +4.5; p=0.18). Triglycerides did not change significantly in either group. Blood urea nitrogen decreased by 6.0 in the vegan group (p<0.001) and did not change significantly in the portion-controlled group (treatment effect -5.2 mg/dL (-7.9 to -2.5; p<0.001). Blood urea nitrogen to creatinine ratio decreased by 5.1 mg/dL in the vegan group (p<0.001) and did not change significantly in the portion-controlled group (treatment effect -6.7 (-10.7 to -2.6; p=0.002).

Conclusion: This study suggests that a low-fat vegan diet may have beneficial effects on insulin sensitivity, insulin requirements, glycemic control, and markers of cardiovascular and renal health in people with type 1 diabetes compared with a portion-controlled diet and may be used in the management of type 1 diabetes. Larger trials are needed to confirm these findings.

Clinical Trial Registration Number: NCT04944316

Supported by: PPCRM and Institute for Technology in Healthcare

Disclosure: H. Kahleova: None.

194

The effect of carbohydrate intake on glycaemic control in type 1 diabetes: a randomised cross-over study

S. Sterner Isaksson ^1,2, A.F. Ólafsdóttir¹, S. Ivarsson¹, H. Imberg^3,4, E. Toft^5,6, S. Hallström^1,7, U. Rosenqvist⁸, M. Ekström², M. Lind¹;

¹Institution of Medicine, Department of Molecular and Clinical Medicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden, ²Department of Medicine, NU Hospital Group, Uddevalla, Sweden, ³Statistiska konsultgruppen, Gothenburg, Sweden, ⁴Department of Mathematical Sciences, Chalmers University of Technology and University of Gothenburg, Gothenburg, Sweden, ⁵Department of Medicine, Ersta Hospital, Stockholm, Sweden, ⁶Institution of Medicine, H7, Karolinska Institute, Stockholm, Sweden, ⁷Sahlgrenska University Hospital, Department of Medicine, Gothenburg, Sweden, ⁸Department of Internal Medicine, Motala Hospital, Motala, Sweden.

Background and aims: Few studies have examined the effects of lower carbohydrate diets in persons with type 1 diabetes (T1D) and results are conflicting. The aim of this study was to evaluate the effects of carbohydrate intake on glucose control in adults with T1D.

Materials and methods: 54 adults with T1D and inadequate glycemic control (HbA1c ≥58mmol/mol) were randomized to start with either a moderate carbohydrate diet (30% of total energy from carbohydrates) or a traditional diabetes diet with higher carbohydrate content (50% of total energy from carbohydrates) for 4 weeks with a wash-out period of 4 weeks in a cross-over design. Masked continuous glucose monitoring was used throughout the study to evaluate effects on glucose control. The primary endpoint (difference in mean blood glucose levels between the diet periods) and secondary endpoints (including time in range, TIR, and time above range, TAR) were defined before enrolment.

Results: 50 patients were included in the full analysis set with a mean (SD) baseline HbA1c 69 (11) mmol/mol, BMI 29 (5), age 48 (14)years, and 50 % were females. The difference in mean blood glucose levels between moderate carbohydrate diet and traditional diabetes diet was -0.6 mmol/L (95% CI -0.9 to -0.3, p=0.001). There was also a significant difference in TIR; 4.7% (95% CI 1.3 to 8.0, p=0.008), TAR (>10 mmol/L) -5.9% (95% CI -9.6 to -2.2, p=0.003), and no significant difference in time in hypoglycemia<3.0 mmol/L 0.6% (95% CI -0.4 to 1.6, p=0.26).

Conclusion: This study shows that a diet with moderate amount of carbohydrates is more effective than the traditional diabetes diet with higher amount of carbohydrates to decrease mean blood glucose levels, time above range and increase time in range without increased risk of hypoglycemia.

Clinical Trial Registration Number: NCT03400618

Supported by: The Healthcare Board, VGR, The Dr. P Håkansson Foundation, The Swedish state (ALF agreement)

Disclosure: S. Sterner Isaksson: None.

195

Effect of intestinal exposure to endogenous bile acids on the incretin axis in type 2 diabetes

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

Level 5, Adelaide Health and Medical Sciences Building, The University of Adelaide, Adelaide, Australia.

Background and aims: Bile acids, previously regarded simply as ‘intestinal detergents’ for fat digestion, are now recognised as regulators of glucose homeostasis. Exogenous bile acids stimulate secretion of the incretin hormone, glucagon-like peptide-1 (GLP-1), but the impact of endogenous bile acids on the incretin axis remains to be determined. We evaluated the effects of both proximal and distal intestinal exposure to endogenous bile on blood glucose and plasma C-peptide, insulin, glucagon, GLP-1, glucose-dependent insulinotropic polypeptide (GIP) and fibroblast growth factor 19 (FGF-19, a hormone coupled with ileal bile acid absorption) in type 2 diabetes (T2D).

Materials and methods: Ten subjects with well-controlled T2D managed by diet and/or metformin (9 males, age 68.4 ± 2.5 years; BMI 28.5 ± 1.2 kg/m², HbA1c 6.6 ± 0.2%, duration of known diabetes 10.5 ± 2.3 years) were studied on 3 separate days in a single-blind, randomised, crossover design. On each day, a multi-lumen naso-intestinal catheter was positioned with an occlusive balloon in the distal duodenum, followed by continuous jejunal infusion of a glucose solution at 3 kcal/min over 90 min (i.e. t = 0 - 90 min). Between t = 0 - 180 min, endogenous bile was aspirated proximal to the balloon, and was either (i) withheld to exclude the intestine from exposure, or re-infused distal to the balloon into (ii) the jejunum (~50 cm from pylorus) or (iii) proximal ileum (~190 cm from pylorus). Blood was sampled frequently for measurements of blood glucose and plasma total GLP-1 and GIP, insulin, glucagon, and FGF-19. Repeated measures ANOVA was used to compare the differences between exclusion of endogenous bile and jejunal exposure to endogenous bile, and between jejunal and ileal exposure to endogenous bile. Data are means ± SEM.

Results: Excluding the small intestine from exposure to endogenous bile reduced plasma FGF-19, GLP-1, insulin and C-peptide concentrations substantially when compared to jejunal exposure (each P < 0.05), but did not affect blood glucose, plasma GIP or glucagon concentrations (all P > 0.1). There were no significant differences in blood glucose, plasma FGF-19, GLP-1, GIP, insulin, C-peptide or glucagon between jejunal and ileal exposure to endogenous bile (each P > 0.1).

Conclusion: Exposure of the small intestine to endogenous bile in T2D stimulates FGF-19 and GLP-1 substantially, to boost insulin secretion in response to luminal glucose; these effects appear independent of the region of the small intestine exposed.

Clinical Trial Registration Number: ACTRN12619001086156

Supported by: Diabetes Australia/ Royal Adelaide Hospital

Disclosure: C. Xie: Employment/Consultancy; Royal Adelaide Hospital. Grants; Diabetes Australia.

196

The effect of dietary branched-chain amino acid restriction during forearm immobilisation on whole-body and muscle insulin sensitivity

M.L. Dirks ^1,2, T.S.O. Jameson¹, L. Domingos¹, R.C. Andrews³, F.B. Stephens¹;

¹Public Health and Sports Sciences, University of Exeter, Exeter, UK, ²Human and Animal Physiology, Wageningen University, Wageningen, Netherlands, ³Institute of Biomedical and Clinical Science, University of Exeter Medical School, Exeter, UK.

Background and aims: Periods of muscle disuse, e.g. during illness or following injury, lead to muscle atrophy and the development of muscle insulin resistance (IR). This disuse-induced insulin resistance occurs rapidly and is accompanied by accumulation of branched-chain amino acids (BCAAs) in skeletal muscle. Importantly, increased circulating BCAA concentrations are predictors of IR and the later development of type 2 diabetes. The aim of the present study was to investigate the effect of dietary BCAA restriction on whole-body and forearm IR during 2 days of forearm immobilisation.

Materials and methods: Twenty-four healthy males and females were randomised into parallel control (CON; 4/8 f/m, 21±1 yrs, BMI 23.2±0.6 kg/m²) or BCAA restricted (BCAA-; 5/7 f/m, 21±1 yrs, BMI 22.7±0.7 kg/m²) groups. All participants underwent 48 h of unilateral forearm cast immobilisation, an experimental model of IR, with metabolic test days conducted prior to and immediately following immobilisation. For two days prior to the baseline test day all participants received an isocaloric low protein diet supplemented with 4 amino acid (AA) drinks per day (24±1 g per drink), such that the total diet provided 56±1 en% carbohydrate, 28±0 en% fat and 16±0 en% protein (1.5 g/kg body mass (BM)/day protein; 10.3±0.2 g leucine, 4.3±0.1 g isoleucine and 5.5±0.1 g valine per day). The same diet was consumed by CON during immobilisation, whereas the BCAA- group received drinks without BCAA on both days (total diet providing 1.5 g/kg BM/day protein; 0.4±0.0 g leucine, 0.2±0.0 g isoleucine, 0.4±0.0 g valine per day). On each test day whole-body glucose disposal rate (GDR) was assessed during a 2.5 h hyperinsulinaemic (50 mU/m²/min)-hyperaminoacidaemic (0.11 g/kg/h)-euglycaemic clamp. Whole-body substrate oxidation via indirect calorimetry and forearm glucose uptake (FGU) via the arterialized venous-deep venous forearm balance method were assessed at basal and under clamp conditions.

Results: Whole-body GDR during the last 30 mins of clamp was unchanged by immobilisation in CON (from 45.2±4.4 to 47.2±4.1 mg/kg; P>0.05) and decreased in BCAA- (from 42.3±4.1 to 36.0±3.2 mg/kg; P=0.022). Immobilisation led to a respective 103±28% and 59±12% decrease in basal and insulin stimulated FGU in CON (P<0.001) and was unaffected by diet. Immobilisation increased and clamps suppressed whole-body fat oxidation (both P<0.05) with no effect of diet. Whole-body energy expenditure and fasting blood glucose were unaffected by immobilisation or diet.

Conclusion: A BCAA-restricted diet consumed during 2 days of forearm immobilisation induces whole-body IR during a hyperinsulinaemic-hyperaminoacidaemic-euglycaemic clamp. This was not associated with an exacerbated decline in FGU of the immobilised forearm, suggesting a BCAA-restricted diet in healthy individuals may impair insulin sensitivity of non-immobilised skeletal muscle in response to amino acid infusion.

Clinical Trial Registration Number: NCT04626401

Supported by: Wellcome Trust 209198/Z/17/Z to Dr Marlou Dirks

Disclosure: M.L. Dirks: None.

197

The effect of adding different volumes of exercise training to diet-induced weight loss on body composition in persons with type 2 diabetes

C. Durrer ¹, M.P.P. Lyngbæk¹, G.E. Legaard¹, N.S. Nielsen¹, T.P. Almdal², M.A.V. Lund³, B. Liebetrau¹, C. Ewertsen⁴, C. Lauridsen⁴, T.T. Soloman⁵, K. Karstoft¹, B.K. Pedersen¹, M. Ried-Larsen¹;

¹7641, Rigshospitalet, Copenhagen, Denmark, ²Department of Endocrinology, Rigshospitalet, Copenhagen, Denmark, ³Department of Biomedical Sciences, University of Copenhagen, Copenhagen, Denmark, ⁴Department of Radiology, Rigshospitalet, Copenhagen, Denmark, ⁵Blazon Scientific, London, UK.

Background and aims: Weight loss has been added to clinical practice guidelines as a treatment target for hyperglycemia and is also implicated in type 2 diabetes (T2D) reversal; however, it is likely the loss of fat-mass (FM) rather than weight loss per se, that is beneficial. Diet-induced weight loss leads to reduced fat-free mass (FFM), which could be detrimental for glycemic control in T2D. As such, combining exercise with diet-induced weight loss to retain FFM could lead to more favorable changes in body composition than diet alone. Nevertheless, data regarding changes in FM and FFM in the context of a hypocaloric diet with exercise is sparse and it is unclear whether exercise can offset a hypocaloric diet concerning FFM loss. Accordingly, the aims of this secondary analysis were to assess the effects of different exercise volumes in combination with diet-induced weight loss on FM percentage (FM%) and FFM in persons with diagnosed T2D.

Materials and methods: In this secondary analysis of a 16-week four-armed randomized trial, 82 persons (35% females) living with T2D were randomly allocated to standard care (n=21, CON), caloric restriction (25% energy deficit; n=20, DCON), DCON + exercise 3 times/week (n=20, MED), or DCON + exercise 6 times/week (n=21, HED). The primary outcome was the change in FM% points (FM%p). Secondary outcomes included changes in FFM (absolute) and visceral adipose tissue (VAT) volume. Data were analyzed via a constrained baseline mixed effects linear model. If data were log-transformed, mean differences are presented as percentages.

Results: Baseline mean age was 58.2 ± 9.8 years, T2D duration was 4.0 years (interquartile range 1.9 to 5.5), and body weight was 101.4 ± 14.6 kg. Following the intervention, FM%p was changed by -3.5%p [95% CI -5.6 to -1.4]; p=0.002 in DCON, -6.3%p [95%CI -8.4 to -4.1]; p<0.001 in MED, and -8%p [95% CI -10.2 to -5.8]; p<0.001 in HED. It was further reduced in MED (-2.8%p [95% CI -4.9 to -0.7]; p<0.001) and HED (-4.5%p [95% CI -6.6 to -2.4]; p<0.001. Absolute FFM was reduced following the intervention in DCON (-2.7% [95% CI -4.4 to -0.9]; p=0.003) and MED (-1.9% [95% CI -3.7 to -0.1]; p=0.04) compared to CON; however, there was no reduction in absolute FFM in HED (-0.2% [95% CI -2.0 to 1.7]; p=0.9). VAT volume was also reduced in all intervention groups compared to CON (DCON: -666.0cm³ [95% CI -1109.4 to -313.1], MED: -1264.0cm³ [95% CI -2129.3 to -725.2], HED: -1786.4cm³ [95% CI -2380.5 to -1131.5]). There were also greater VAT reductions in MED (-598.1cm³ [95% CI -1444.1 to -9.0]) and HED (-1120.4cm³ [95% CI -1664.9 to -332.2]) compared to DCON.

Conclusion: All interventions were superior in reducing FM%p and VAT compared to CON. Adding exercise to caloric restriction was superior in reducing FM and VAT compared to caloric restriction alone. FFM was also reduced following caloric restriction alone and with 3 sessions of exercise a week but was unchanged when participating in 6 exercise sessions per week. These findings suggest that the negative effects of caloric restriction on FFM can be ameliorated by relatively large volumes of concomitant exercise.

Clinical Trial Registration Number: NCT03769883

Supported by: TrygFonden, Svend Andersen fonden, CIHR

Disclosure: C. Durrer: Grants; TrygFonden, Svend Andersen Fonden.

198

The effect of circuit training, empagliflozin or "vegeterranean diet" on metabolic and physical functions in older adults with type 2 diabetes (CEV-65 trial)

A. Buch ^1,2, O. Kis¹, Y. Greenman¹, N. Stern³, R. Eldor¹;

¹Institute of Endocrinology, Metabolism, Diabetes and Hypertension Tel Aviv Medical Center, Tel-Aviv, Israel, ²Nutritional Sciences, School of Health Sciences, Ariel University, Ariel, Israel, ³The Sagol Center for Epigenetics of Metabolism and Aging, Tel Aviv Sourasky Medical Center, Tel-Aviv, Israel.

Background and aims: Older adults with type 2 diabetes (T2DM) are at high risk for functional decline and sarcopenia. Aim: to assess the effects of three interventions associated with a negative energy balance: 1) circuit resistance training (CRT); 2) empagliflozin, and 3) plant-based Mediterranean diet ("vegeterranean") on metabolic and physical functions in older subjects with T2DM.

Materials and methods: 100 older adults (≥65 years) with T2DM and low levels of physical activity were randomized to three arms for 10 weeks: minimally-supervised (two sessions) CRT consisting of 3 home sessions/week; "vegeterranean" diet (VegMedD) - carbohydrates limited, with avoidance of red meat and poultry and daily recommendation of 1-2 servings of legumes or 10 mg/day empagliflozin. Compliance was monitored using a self-reported log (>67% of intentional workouts), return of remaining tablets (≥ 80% considered as good compliance), and a validated Mediterranean-diet adherence questionnaire. Efficacy was determined by metabolic measures (glycemic control and body composition) and functional parameters.

Results: 83 participants (58% women, n=29 for empagliflozin and VegMedD and n=25 for CRT) aged 69.5± 4.6 years with a mean BMI of 32±6 kg/m²; 42±7 % fat and mean HbA1c levels of 7.4±1.2 % completed 10 weeks. Compliance was as follows: 1) 81% in CRT; 2) 94% in empagliflozin 3) 100% in VegMedD (mean score 9.25 at baseline increased to 11.92). HbA1c change was similar between empagliflozin and VegMedD (-0.5%; -0.9% in men receiving empagliflozin), with no change in CRT (p<0.05). Insulin decreased mostly by empagliflozin (-2.19 MCU/ML vs. -0.49 in VegMedD and +0.48 CRT, p<0.05). VegMedD resulted in the highest decrease of fat compared to empagliflozin and CRT (-2.18 kg vs. -0.75 kg vs. +0.21 kg, respectively, p<0.05). However, skeletal muscle mass slightly decreased in VegMedD (-0.05 kg) compared to gains in empagliflozin and CRT (+0.17 kg, +0.25 kg). All metabolic changes were more prominent among men. Functional ability improved in CRT participants -mainly among women- with an increase in 30-sec chair raise [from 10.88 repetitions to 12.38 (+1.8 among women)]. In contrast, a decline was noticed in VegMedD and empagliflozin (-0.15 and -0.56 repetitions, respectively, p<0.05). Knee extension only in CRT (+1.33 kg) as compared to a decrease in knee extension strength in VegMedD and empagliflozin groups (-1.66 and -2.07 kg, respectively, p<0.05). Subjects who reduced their HbA1c more than the mean decline in the total cohort lost more fat, less muscle and handgrip strength, and improved more their functional ability.

Conclusion: Older adults with T2DM receiving short-term (10 weeks) empagliflozin or VegMedD improved similarly regarding metabolic control. These therapies, though, had a functional decline tradeoff as moderate declines in muscle strength and functional ability were observed. 10 weeks of CRT was inefficient for glycemic control but resulted in functional ability gains. Older adults with T2DM receiving glycemic potent therapies such as diet or empagliflozin may benefit from the addition of resistance training in a circuit manner in terms of physical function preservation.

Clinical Trial Registration Number: NCT03560375

Disclosure: A. Buch: None.

OP 34 Diabetic neuropathy from the brain to the bowels

199

Preservation of thalamic mitochondrial and neuronal function in Painful-Diabetic Peripheral Neuropathy: a multi-modal cerebral imaging study

G. Sloan ¹, D. Selvarajah¹, A. Anton², K. Teh², I.D. Wilkinson², S. Tesfaye¹;

¹Department of Oncology and Metabolism, University of Sheffield, Sheffield, UK, ²Academic Unit of Radiology, University of Sheffield, Sheffield, UK.

Background and aims: Previous neuroimaging studies have highlighted thalamic structural and vascular alterations in Painful-Diabetic Peripheral Neuropathy (DPN). In this study, we present a multi-modal MRI study to determine the structure, and neuronal and mitochondrial function of the thalamus in Painful-DPN.

Materials and methods: A total of 49 participants were enrolled, 38 with type 2 diabetes (9 without DPN [No-DPN]; 11 with Painless-DPN; and 18 with Painful-DPN) and 11 healthy volunteers (HV). Participants underwent detailed clinical and neurophysiological assessments, and MR imaging of the thalamus including: Proton-Magnetic Spectroscopy (1H-MRS); 31-Phosphorus MRS (31P-MRS); and brain morphometric analysis. We calculated the metabolite ratios N-acetyl aspartate to choline (NAA:Cho), a measure of neuronal function, and inorganic phosphate (Pi) to ATP (Pi:ATP), a measure of mitochondrial function.

Results: Here were no significant differences in thalamic volume between groups (ANOVA, p=0.927). There were significant group effects in both metabolite ratios: NAA:Cho (ANOVA, p=0.013); Pi:ATP (p=0.021). The NAA:Cho was numerically the lowest in Painless-DPN, reaching significance compared with HV (p=0.001) and Painful-DPN (p=0.019). Additionally, the Pi:ATP was significantly higher in Painless-DPN compared with HV (p=0.013) and Painful-DPN (p=0.008). There was also a significant correlation between the NAA:Cho and Pi:PCr (Pearson’s r -0.336, p=0.034).

Conclusion: This is the first study to perform 1H- and 31P-MRS in human DPN. We have demonstrated altered 1H- and 31P-MRS ratios in Painless- but not Painful-DPN, suggesting that mitochondrial dysfunction may underly thalamic neuronal dysfunction in Painless-DPN. However, in Painful-DPN, preservation of thalamic mitochondrial and neuronal function may be involved in the maintenance of neuropathic pain signaling.

Disclosure: G. Sloan: None.

200

Association of microvascular complications with reduced neuronal activity and cognitive impairment in type 2 diabetes: a resting-state fMRI study

W. Wang ¹, X. Yang¹, B. Lu¹, P. Gu¹, X. Hui², J. Shao¹;

¹Department of Endocrinology, Jinling Hospital, School of Medicine, Nanjing University, Nanjing, China, ²School of Biomedical Sciences, The Chinese University of Hong Kong, Hong Kong, China.

Background and aims: Adults with type 2 diabetes are at an increased risk of developing microvascular complications including diabetic nephropathy, retinopathy and neuropathy. Recently, more and more imaging studies of type 2 diabetes mellitus have reported structural and functional abnormalities in a variety of spatially diverse brain regions. The complex network in the human brain is the physiological basis for information processing and cognitive expression. Increasing amounts of data suggest that diabetes-related microvascular dysfunction is associated with a higher risk of cognitive impairment. However, the underlying mechanism is still unclear. This study aimed to investigate whether microvascular complications were associated with cognitive performance and brain dynamics in patients with type 2 diabetes mellitus (T2DM) based on resting-state fMRI analysis.

Materials and methods: Ninety eight patients with T2DM (55 diabetic patients without microvascular complications(DM) and 43 diabetic patients with microvascular complications(DM-C)) were recruited for Montreal Cognitive Assessment test (MoCA), Mini-Mental State Examination (MMSE), Self-Rating Anxiety Scale (SAS), Self-Rating Depression Scale (SDS) and neural activity of brain which was measured by rest-state function MRI (rs-fMRI).

Results: Diabetic patients with microvascular complications displayed a significantly decreased score on Montreal Cognitive Assessment test (MoCA), compared to the DM group (P=0.048). No notable disparity in Mini-Mental State Examination (MMSE), Self-Rating Anxiety Scale (SAS) and Self-Rating Depression Scale (SDS)was observed between the two groups when analyzed(P>0.05). Among rs-fMRI parameters, no significant difference was observed in gray matter volume. And meanwhile, mean amplitude of low-frequency fluctuations (mALFF) significantly decreased in the right cuneus, right calcarine, right superior occipital gyrus, right inferior occipital gyrus,and bilateral precentra in diabetic patients with microvascular complications, compared to DM patients. Futhermore, compared with DM patients, Diabetic patients with microvascular complications patients showed a significantly decreased mean regional homogeneity (mReho) in the left hippocampu, bilateral calcarine, left inferior occipital gyrus, and increased mReho in the left middle frontal gyrus. The mReho were positively correlated with MoCA (P = 0.009, r = 0.264) in diabetic patients.

Conclusion: Our findings suggest that diabetic patients with microvascular complications displayed more impaired cognition. Decreased neuronal activity in visual network and memory center was well correlated with diabetes-related microvascular dysfunction.

Disclosure: W. Wang: None.

201

Lipid parameters and risk of diabetic polyneuropathy in type 2 diabetes: a Danish cohort study

F.P.B. Kristensen ¹, D.H. Christensen¹, B.C. Callaghan², H.T. Sørensen¹, R.W. Thomsen³;

¹Department of Clinical Epidemiology, Aarhus University and Aarhus University Hospital, Aarhus, Denmark, ²Department of Neurology, University of Michigan, Ann Arbor, MI, USA, ³Department of Clinical Epidemiology, Aarhus University, Aarhus, Denmark.

Background and aims: Diabetic polyneuropathy (DPN) is a serious complication of type 2 diabetes (T2D) and a major cause of foot ulcers and amputations. Updated clinical guidelines from the American Diabetes Association emphasize that lowering blood lipids may aid in prevention of DPN. However, the precise association between specific blood lipid parameters and DPN remains unknown. We investigated the association of triglycerides, high-density lipoprotein (HDL) cholesterol, low-density lipoprotein (LDL) cholesterol, and total cholesterol with hospital-diagnosed DPN.

Materials and methods: Using population-based medical registries, we conducted a cohort study of 59,137 newly diagnosed T2D patients residing in Northern Denmark during 2005-2017. Lipid values were assessed one year after T2D diagnosis (the study index date for start of follow-up). DPN was identified using validated hospital diagnosis codes. We followed patients for incident DPN for up to 15 years, until death, emigration, or study end (April 30, 2021). We used Cox regression and restricted cubic splines to compute hazard ratios (HRs) for DPN, adjusted for age, sex, calendar year, hypertension, obesity, smoking, lipid- and glucose-lowering drug therapy, HbA1c, eGFR, comorbidities, and markers of health care utilization.

Results: During median follow-up of 7.3 years, 1,544 patients had hospital-diagnosed DPN (incidence rate = 3.6 per 1000 person-years). LDL and total cholesterol had no association with DPN, neither in continuous splines analyses nor using guideline cut-off levels. Compared with LDL cholesterol <1.8 mmol/L, the adjusted HR was 1.04 (95% CI 0.91-1.19) for LDL cholesterol between 1.8-2.59 mmol/L and 1.00 (95% CI 0.87-1.15) for levels ≥2.6 mmol/L. In contrast, we observed a linearly increasing association with DPN for triglyceride levels above 1.4 mmol/L and for HDL cholesterol levels above 1.5 mmol/L. When using guideline-specific cut-off levels, the adjusted HRs for DPN were 1.04 (95% CI 0.91-1.19) for triglyceride levels of 1.7-2.3 mmol/L, and 1.22 (95% CI 1.07-1.38) for triglyceride levels ≥2.3 mmol/L, compared with triglyceride levels <1.7 mmol/L. When we additionally adjusted for HDL and LDL cholesterol values, the triglyceride association increased to HRs of 1.12 (95% CI 0.97-1.28) for levels of 1.7-2.3 mmol/L and 1.34 (95% CI 1.17-1.53) for levels ≥2.3 mmol/L. For HDL cholesterol, levels <1.0 mmol/L were not associated with increased DPN (adjusted HR 0.98 [95% CI 0.87-1.12]), compared with guideline cut-off ≥1.0 mmol/L. Additionally adjusting for triglycerides and LDL cholesterol did not materially change the association. The results were similar among statin users and nonusers.

Conclusion: This study provided evidence that elevated triglyceride levels and possibly high HDL cholesterol, but neither LDL nor total cholesterol, are associated with increased risk of DPN. Our findings support a link between elevated triglycerides and DPN and may enable clinicians to improve DPN risk assessment in patients with early T2D.

Supported by: Aarhus University

Disclosure: F.P.B. Kristensen: None.

202

Preclinical efficacy of the novel GPR40 agonist CPL207-280 in neuropathy

T. Jaworski ¹, M. Janiszewski¹, P. Buda¹, V. Vialichka¹, K. Bazydło-Guzenda¹, M. Mach¹, K. Gałązka¹, J. Dominowski¹, M. Górka¹, K. Sałat², P. Popik³, K. Dubiel¹, J. Pieczykolan¹, M. Wieczorek¹;

¹Celon Pharma, Kazun Nowy, Poland, ²Jagiellonian University, Kraków, Poland, ³Maj Institute of Pharmacology, Kraków, Poland.

Background and aims: Neuropathic pain is a chronic condition resulting from the dysfunction of the somatosensory nervous system producing allodynia and hyperalgesia. It is the most common complication of diabetes, among other conditions. We have synthesized the CPL207280-51 compound, a GPR40 receptor agonist, with potent antidiabetic activity in T2D models. Since GRP40 agonists reduce allodynia in animal models of neuropathic pain, we analyzed the efficacy of CPL207280-51 in neuropathy.

Materials and methods: The in vivo distribution of orally administered CPL207280-51 (metformin salt) (3, 30, and 100 mg/kg) to the brain and spinal cord, and its pharmacokinetics were analyzed in BALB/c mice and Wistar rats using LC-MS/MS. The effect of CPL207280-51 on tactile allodynia in neuropathic (STZ-treated) mice was tested using the von Frey test. The effect of CPL207280-51 on the heat nociceptive threshold in STZ-treated mice was assessed using the hot plate test. Baseline (predrug) values of pain sensitivity were obtained before the mice were orally pretreated with CPL207280-51. Postdrug values were obtained at 30 min, 2 h, 7 h, 12 h, and 24 h and compared to pregabalin. Potential CPL207280-51 abuse was studied in Sprague-Dawley rats by Conditioned Place Preference and Drug Discrimination induced by morphine or THC. CPL207280-51 either alone or in combination with morphine or THC was administered before conditioning sessions.

Results: CPL207280-51 was quickly absorbed reaching higher plasma and CNS concentrations in mice than in rats within 15 min and declining thereafter. In STZ-neuropathic mice, oral administration of CPL207280-51 (30mg/kg) increased mechanical nociceptive threshold after 2h (p < 0.01), 7h (p < 0.05) and 12 h (p < 0.0001 vs. predrug paw withdrawal threshold). This antiallodynic effect of CPL207280-51 was separated from its metformin content. Mechanistically, the antiallodynic effect of CPL207280-51 was lowered by pretreatment with Rimonabant (CB1 receptor antagonist) (p < 0.01) and Naltrexone (μ-opioid receptor antagonist) (p < 0.001). In contrast, the heat nociceptive threshold was not affected by CPL207280-51. Next, a screen of 11 models of pain (acute & tonic, neuropathic, inflammatory, post-operative, and visceral pain) in rats indicated an analgesic action of CPL207280-51 in the oxaliplatin model of neuropathic pain. We confirmed this in a separate study in oxaliplatin-treated rats receiving CPL207280-51. CPL207280-51 (100 mg/kg) increased mechanical nociceptive threshold after 2h (p < 0.001), 4h (p < 0.0001), and 7 h (p < 0.0001) compared to the vehicle. In the drug abuse study, CPL207280-51 alone produced neither rewarding nor aversive effects. When given instead of morphine or THC, CPL207280-51 did not produce morphine- or THC-like responses. Interestingly, when given prior to morphine or THC, CPL207280-51 partially inhibited morphine- or THC-like responses. In none of the treatments, CPL207280-51 caused adverse effects and was well tolerated throughout the studies.

Conclusion: CPL207280-51 displays potent anti-allodynic activity in models of neuropathy involving the action of CB1 and opioid receptors without producing addiction.

Supported by: POIR.01.01.01-00-1323/20

Disclosure: T. Jaworski: None.

203

Prevalence of diabetic neuropathies in individuals with type 1 and type 2 diabetes in a tertiary outpatient clinic: the DANES cohort study

H. Mizrak ¹, H. Kufaishi¹, S.K. Hecquet¹, T.W. Hansen¹, R. Pop-Busui², P. Rossing¹, B. Brock¹, C.S. Hansen¹;

¹Steno Diabetes Center Copenhagen, Herlev, Denmark, ²Department of Internal Medicine, Division of Metabolism, Endocrinology and Diabetes, Ann Arbor, MI, USA.

Background and aims: Distal symmetrical polyneuropathy (DPN) and diabetic autonomic neuropathy (DAN) are common and severe complications to diabetes. Global prevalence estimates vary in part due to variations in screening modalities.

Materials and methods: We investigated the prevalence of DPN, cardiovascular autonomic neuropathy (CAN) and gastroparesis in individuals with type 1 diabetes (T1D) and type 2 diabetes (T2D) using a broad array of diagnostic modalities. DPN was evaluated using vibration perception threshold (VPT), sural nerve function, touch sensation, pain sensation, thermal sensation and sudomotor function. Definite DPN was defined based on the Toronto Consensus Criteria based on a combination of abnormal sural nerve conduction velocity/amplitude or abnormal sudomotor function and signs (abnormal VPT/touch sensation/pain sensation) or symptoms of DPN. Painful DPN was evaluated using Douleur Neuropathique 4 Questions (DN4) (DN4≥ 4). CAN was defined as two or more abnormal cardiovascular reflex tests, and gastroparesis was assessed using gastroparesis cardinal symptom index (≥ 1.9).

Results: In total, 1.721 individuals with T1D (n= 822; mean age (±SD) 54 ±16 years, median diabetes duration [IQR] 26 [15-40] years) and T2D (n= 899; mean (±SD) age 67 ±11 years, median diabetes duration [IQR] 16 [11-22] years) were included. Prevalence of DSPN varied from 2 -74% for T1D and 5-78% for T2D (Figure 1). Prevalence of definite DPN was 54% and 68% for T1D and T2D, respectively. Painful DPN (DN4≥ 4) was present in 5% and 15% (T1D and T2D, respectively). CAN was present in 14% and 19% with T1D and T2D, and gastroparesis in 6% and 9% with T1D and T2D (Figure 1). For isolated presence of neuropathies in T1D, 45% had only DPN without any other neuropathy (assessed with VPT > 25V or touch sensation), 4% had only painful DPN, 18% had only CAN, and 7% had only gastroparesis. For type T2D, the numbers were 50% DPN, 3% painful DPN, 11% CAN and 4% gastroparesis. For most diagnostics modalities, individuals with T2D had a higher prevalence of DPN, painful DPN and gastroparesis than those with T1D (Figure 1).

Conclusion: In this large contemporary cohort of people with both T1D and T2D treated according to current standard of care, we found a high prevalence of DPN and DAN, but with wide variation depending on diagnostic modality. DPN and DAN were only co-existing in approximately 50%, thus screening for both types of neuropathies is important.

Clinical Trial Registration Number: NCT05043636

Disclosure: H. Mizrak: None.

204

Overall gastrointestinal symptom burden is associated with diabetic autonomic neuropathy: a DANES cohort study

H. Kufaishi, H.I. Mizrak, B. Brock, P. Rossing, C.S. Hansen;

Steno Diabetes Center Copenhagen, Herlev, Denmark.

Background and aims: Up to 40% of people with diabetes report disabling gastrointestinal (GI) symptoms. Gastroparetic symptoms appear twice as frequent among persons with diabetes than in the general population. Diabetic autonomic neuropathy (DAN) is a cause of GI dysmotility that can lead to GI-symptoms, but symptoms and signs rarely correlate. We investigated the association between overall GI symptom burden and DAN assessed by different modalities in persons with type 1 diabetes (T1D) and type 2 diabetes (T2D).

Materials and methods: Cross-sectional study including 943 individuals with T1D (n=566) or T2D (n=377) treated at a tertiary outpatient clinic. All participants were without gluten allergy, B12-deficiency, cancer, GI conditions and beta-blocker treatment. GI symptom burden was assessed using a previously published weighted score combining the gastroparesis composite symptom index (GCSI) and the gastrointestinal symptom rating score (GSRS), a cut-off of 2.3 was applied. DAN status was evaluated using the composite autonomic symptom 31 score (COMPASS-31) questionnaire, cardiac autonomic reflex tests (CARTs) and age-sex-adjusted electrochemical skin conductance on hands and feet (Sudoscan). Cut-off for COMPASS-31 was ≥16 points, and presence of cardiac autonomic neuropathy was defined as at least two out of three abnormal CARTs. Adjusted logistic regressions analyses were performed for the total cohort and for T1D and T2D separately. Odds ratios (OR) were adjusted for age, sex, diabetes duration, LDL cholesterol, HbA1C, systolic blood pressure and smoking status.

Results: Mean ± SD age of the total population was 58 ± 15 years and 565 (40.1%) were women. A high GI symptom burden was present in 143 (15.2%) with T1D and 142 (15.1%) with T2D. In pooled adjusted analyses a high GI symptom burden was associated with the adjusted ORs (95%CI, p-value) of having DAN evaluated by COMPASS31 (n=377): 4.20 (CI: 3.13-5.33, p<0.001), for CAN (n=116): 1.64 (CI: 1.05-2.54, p=0.003) and for Sudoscan assessments: 1.52 (1.09-2.11, p=0.013) and 1.89 (1.37-2.68. p<0.001), hands (n=244) and feet (n=444) respectively. Analyses for T1D and T2D separately showed associations between a high GI symptom burden and presence of DAN when evaluated with COMPASS-31 (p<0.001 for both) and CAN for T1D, but not for other DAN outcomes (p≥ 0.05).

Conclusion: A high GI symptom burden is associated with presence of DAN especially when DAN is assessed by COMPASS-31 score. GI symptom burden may be associated with objective measures of DAN. However, results suggest that a symptom evaluation does not correlate well with objective measures of DAN.

Clinical Trial Registration Number: H-10929796

Disclosure: H. Kufaishi: None.

OP 35 Omics and more: unbiased approaches to tackle obesity

205

Metabolic heterogeneity in obesity: the interplay between metabolome and microbiome reflects visceral adipose tissue physiology beyond BMI

R. Chakaroun, M. Pradhan, H. Brolin, M. Schöler, G. Bergström, V. Tremaroli, F. Bäckhed;

Department for Molecular Medicine, The Wallenberg Laboratory, Gothenburg, Sweden.

Background and aims: Obesity increases cardiovascular disease (CVD) risk, but BMI fails to accurately capture the clinical heterogeneity of obesity. Recent data suggest that circulating plasma metabolites can better predict obesity's underlying cardiometabolic risk, reflecting intrinsic and extrinsic factors, including the gut microbiome. We investigated whether the microbiome, independent of BMI, can detect the implicit metabolic risk of obesity.

Materials and methods: 1408 study participants were recruited from a community-based cohort and deeply phenotyped with clinical and body composition parameters, proteome, metabolome, and microbiome. Metabolome data was acquired using HPLC-MS, and proteome profiles were generated with the Olink Proteomics platform. Microbial profiles were obtained from whole-genome metagenomics using Kraken2 and Bracken (UHGG catalog-v2). The analyses performed in R computing language included ridge regression for variable prediction, distance-based redundancy analysis, differential taxonomic and functional abundance, and were corrected for medication using MetaDecoundR.

Results: The 1,408 individuals (%females = 56) aged between 50 and 65, had a wide range of BMI (18.3 - 46.3, mean = 27.1 kg/m²), and none were on antidiabetics. Both the metabolome and microbiome significantly contributed to the variance of CVD risk, metabolic, and adiposity measures, and the metabolome explained 60% of the variance in gut microbiome gene richness. BMI, body composition, insulin resistance (IR) parameters, and diet had a significant impact on microbial community structure (adjusted R² for BMI, triglyceride-glucose-index, and dietary fibers: 0.005, 0.005, and 0.003 respectively). Metabolome-predicted BMI (metBMI - out-of-sample R²= 0.52) was linked to increased IR, inflammation, dyslipidemia, and more severe visceral abdominal adipose tissue (VAT) attenuation (P<0.05). Gut microbiome features explained the levels of about 30% of blood metabolites (R² ≥ 0.3), suggesting a modulation by the microbiome. Mediation analyses confirmed a significant association between VAT attenuation, metBMI, and HOMA-B, revealing that 38% of the total effect of VAT attenuation on HOMA-B was mediated by metBMI residuals (P < 2e-16). This obesogenic metabolomic signature was characterized by a depletion of butyrate-producers and uncharacterized Lachnospiraceae, and an increase in oral bacteria and opportunistic pathogens, along with an elevated abundance of Ruminococcus gnavus (R²Abundance = 0.75). The presence of R. gnavus was strongly associated with VAT area and more severe attenuation (abs. ρ=0.2, P < 2e-16).

Conclusion: We evidence a connection between body composition, metabolites, and the microbiome. IR and glucose uptake/utilization impairment are associated with the metabolic heterogeneity of obesity, as indicated by circulating metabolites, and linked to the microbiome. This supports prior studies showing that metabolic derangement is better explained by microbiome gene richness than by BMI. Our data adds that the microbiome reflects an obesogenic metabolomic signature with an overall depletion of bacteria and enrichment of R. gnavus, linked to VAT attenuation, a non-invasive marker of adipose tissue architecture. This may affect beta cell function, highlighting the clinical importance of our findings.

Supported by: DFG - Walter Benjamin Fellowship grant

Disclosure: R. Chakaroun: None.

206

Deep phenotyping of glucose stimulation response in human adipocytes at scale

P. Kubitz ¹, H. Dashti², F.R.C. dos Santos², R. Kaalia², J. Perez-Schindler², H. Hauner¹, M. Claussnitzer²;

¹Technical University of Munich, Munich, Germany, ²BROAD Institute of MIT and Harvard, Cambridge, MA, USA.

Background and aims: Adipose tissue is critically involved in the regulation of whole body insulin sensitivity and phenotypic changes in adipocytes are linked to insulin resistance and type 2 diabetes. However, recent studies mainly focus on a small set of targeted cellular phenotypes such as adipocyte size and only a few mechanistically relate cellular morphological alterations to transcriptomic and functional signatures of metabolic disease. Thus, a principled hypothesis-free approach for modeling insulin resistance in adipocytes across high-dimensional cellular and transcriptomic data layers anchored in targeted cell-based phenotypes is currently missing.

Materials and methods: Here, we introduce a deep cellular phenotyping framework that combines multi-omics with a glucose titration screen in human white adipocytes (hWAT). First, we captured insulin-stimulated glucose uptake in adipocytes exposed to a range of glucose concentrations, covering the spectrum from physiological to pathophysiological levels, spanning the full matrix of highly insulin-sensitive to fully insulin resistant adipocytes. To link adipocyte insulin resistance to high-dimensional representations of cellular and morphological processes, we apply LipocyteProfiler, our novel metabolic disease tailored high-content imaging tool that extracts 3,005 features per single cell across four fluorescent channels. Additionally, we map differences in transcriptomic profiles between healthy and insulin resistant adipocytes by RNA-seq. Finally, we analyze and integrate the different data modalities using linear regression models, gene set enrichment analysis (GSEA) and multimarker analysis of genomic annotation (MAGMA) to integrate T2D genome-wide association study data.

Results: Using a principled deep phenotyping approach we established an in vitro model of insulin resistance in human adipocytes. We observed a dose-dependent decrease of insulin responsiveness measured by insulin-stimulated glucose uptake in cells differentiated under increasing glucose concentrations. We observed a range of 3.5-fold to no significant insulin-stimulated glucose uptake relative to basal cells (p = 4.048e-09). Linking insulin responsiveness to LipocyteProfiler feature sets revealed significant differences in DNA-, actin-, mitochondria- and lipid droplet-associated cellular profiles between healthy and insulin resistant adipocytes (n = 1,907 features, FDR < 5%). Insulin resistant cells were particularly enriched for mitochondria-associated features (feature_i ~ glucose uptake fold change, p = 2.47e-08, estimate= -0.57). Finally, linking insulin resistance associated image based features with transcriptomic changes (linear mixed model-derived significant feature-gene connections, n = 20,296, FDR<0.01%) revealed that top-scoring mitochondrial morphological features associate with known and novel contributors to insulin resistance, including ZNF768 (adj. q = 2.21e-05), RALGAPA2 (adj. q = 3.55e-07) or KLF15 (adj. q = 2.10e-05).

Conclusion: In this work, we mapped the molecular and cellular programs of insulin resistance in human adipocytes by anchoring high-dimensional deep cellular profiles to an established read-out of insulin resistance in the cell. We established the functional relationship between insulin resistance and mitochondrial hyperactivity in human adipocytes.

Supported by: NNF Center NNF21SA0072102

Disclosure: P. Kubitz: None.

207

Whole-body measurement of fatty acid uptake in fasting and postprandial conditions

I. Laitinen ¹, S. Ekström², H. Haraldsson¹, S. Pierrou¹, K. Kalliokoski³, M. Kiugel⁴, M. Scheinin⁴, S. Southekal⁵, M. Lu⁵, T. Coskun⁶, Z. Milicevic⁶, L. Johansson¹, P. Nuutila³;

¹Antaros Medical, Mölndal, Sweden, ²Antaros Medical, Uppsala, Sweden, ³Turku PET Centre, University of Turku and Turku University Hospital, Turku, Finland, ⁴Clinical Research Services Turku - CRST Oy, Turku, Finland, ⁵Eli Lilly, Philadelphia, PA, USA, ⁶Eli Lilly, Indianapolis, IN, USA.

Background and aims: Changes in fatty acid metabolism are a component of the pathogenesis of type 2 diabetes mellitus (T2DM). Fatty acid uptake (FAU) in tissues can be measured with positron emission tomography (PET) imaging, employing the fatty acid analogue radiotracer ¹⁸F-FTHA. Fatty acid fluxes in the fasted state have been investigated, but postprandial fluxes remain poorly characterized, because of the large and unpredictable variation of circulating postprandial free fatty acid (FFA) levels and their effects on the quantitation of FAU. This method development study aimed to investigate fasted and postprandial FAU using whole-body PET/MRI.

Materials and methods: 17 participants with overweight/obesity (12 with T2DM) underwent 4 repeated examinations under standardized conditions after overnight fast, twice under fasting and twice postprandially (PP). Dynamic whole-body PET/MRI using ¹⁸F-FTHA was performed (PP 107 minutes after a standard liquid meal). Tissue-specific net uptake rates (Ki) were measured in skeletal muscle (quadriceps), subcutaneous (SAT) and visceral (VAT) adipose tissues, myocardium, liver and brain. Circulating FFA levels were recorded, and FAU was calculated as Ki × median FFA level during the scan. One-way ANOVA p-values, and population mean differences (fasted - PP) and their 95% CI were calculated for group differences.

Results: The net uptake rate (Ki) of tracer ¹⁸F-FTHA was significantly greater in PP compared to the fasted state, in all investigated tissues (95% CI of difference <0), except in the brain where it was slightly lower (95% CI >0). FAU was significantly lower in PP compared to the fasted state in all investigated tissues (Table), as intake of the meal reduced circulating FFA level in the PP state. In the quadriceps muscle, the Ki was significantly lower in T2DM subjects compared to the control obese/overweight subjects (p= 0.038 for PP and 0.012 for fasted) but the difference in FAU was not significant. No differences in FAU between the subject groups were observed.

Conclusion: ¹⁸F-FTHA -PET/MRI provides information on tissue-specific FAU in the fasting state, and also postprandially, with significantly lower FAU in tissues compared to the fasted state. No significant differences were observed between participants with T2DM or without T2DM for FAU in the investigated tissues.

Clinical Trial Registration Number: EudraCT 2020-005211-48

Supported by: Antaros Medical was supported by Eli Lilly for the conduct of the study

Disclosure: I. Laitinen: Employment/Consultancy; Employed by Antaros Medical.

208

Genetic dissection of serum vaspin highlights its causal role in lipid metabolism

M. Würfel ¹, J. Breitfeld¹, K. Horn^2,3, D. de Duc^4,5, A. Velluva^5,6, C. Marzi^7,8, H. Grallert^7,8, N. Friedrich^9,10, M. Pietzner^9,10, U. Völker^10,11, H. Völzke^7,12, E. Ahlqvist¹³, A. Tönjes¹, M. Scholz^2,3, P. Kovacs¹;

¹Department of Medicine, University of Leipzig, Leipzig, Germany, ²Institute for Medical Informatics, Statistics and Epidemiology, University of Leipzig, Leipzig, Germany, ³LIFE Research Center, University of Leipzig, Leipzig, Germany, ⁴Institute of Human Genetics, University Medical Center Leipzig, Leipzig, Germany, ⁵Department of Evolutionary Genetics, Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany, ⁶Rudolf Schönheimer Institute of Biochemistry, Medical Faculty, University of Leipzig, Leipzig, Germany, ⁷Research Unit of Molecular Epidemiology, Helmholtz Center Munich, German Research Center for Environmental Health, Neuherberg, Germany, ⁸German Research Center for Environmental Health, Institute of Epidemiology II, Helmholtz Center Munich, Neuherberg, Germany, ⁹Computational Medicine, Berlin Institute of Health at Charité – Universitätsmedizin Berlin, Berlin, Germany, ¹⁰DZHK (German Center for Cardiovascular Research), partner site Berlin, Berlin, Germany, ¹¹Interfaculty Institute for Genetics and Functional Genomics, University Medicine and Ernst-Moritz-Arndt University Greifswald, Greifswald, Germany, ¹²Institute for Community Medicine, University Medicine Greifswald, Greifswald, Germany, ¹³Department of Clinical Sciences in Malmö, Clinical Research Centre, Lund University, Malmö, Sweden.

Background and aims: Vaspin (visceral adipose tissue derived serine protease inhibitor) is associated with metabolic traits related to obesity, but its causative role is still elusive. We investigated the role of genetics in the serum vaspin variability to establish its direct causative relationship with metabolically relevant traits.

Materials and methods: Meta-analysis of genome-wide association studies (GWAS) for serum vaspin from six independent cohorts (N=7,446) was conducted. Potential functional variants of the vaspin locus were included in Mendelian Randomization (MR) analyses to assess possible causative chains between vaspin and correlated metabolically relevant traits. In vivostudies in dbdb mice were conducted to validate the GWAS findings.

Results: In total, 468 single nucleotide polymorphisms (SNPs), all within the vaspin locus, were significantly associated with circulating serum vaspin (all P-values <5x10^-8). Five independent SNPs (rs7141073, rs1956709, rs4905216, rs61978267, rs73338689) indicated a considerable locus heterogeneity. Respective credible sets contained 20 variants, none of them showing strong deleteriousness estimates. MR analyses including the five index SNPs revealed causal relationships between serum vaspin concentrations and triglycerides, low-density lipoprotein and total cholesterol. Gene expression correlation analyses based on data retrieved from the GTEx project suggests that genes highly correlating with vaspin expression in the adipose tissue are involved in lipid metabolic processes. Chronic vaspin treatment reduced serum triglyceride levels in genetically obese dbdb mice.

Conclusion: Our data show that serum vaspin is strongly determined by genetic variants within vaspin and they further highlight vaspin’s causal role in lipid metabolism.

Disclosure: M. Würfel: None.

209

Genetic variants in the adiponectin receptor genes are associated with adiponectin und insulin levels

L. Kedenko, T. Kiesslich, I. Kedenko, B. Paulweber;

Regional Hospital Salzburg, Salzburg, Austria.

Background and aims: Assuming that adiponectin receptors mediate the antidiabetic effects of adiponectin, we hypothesize that genetic variation in the adiponectin receptor genes (AdipoR1 and AdipoR2) may play a role in glucose metabolism, insulin resistance and type 2 diabetes. Aims of the current study were: to genotype selected single nucleotide polymorphisms (SNPs) in the AdipoR1 and AdipoR2; to clarify the influence of genetic variability in the AdipoR1 and AdipoR2 on adiponectin levels, glucose metabolism and insulin sensitivity in the SAPHIR cohort.

Materials and methods: The SAPHIR study was initiated in 1999 as a population-based prospective study in and around the city of Salzburg (Austria). The study comprises 1770 healthy unrelated subjects (663 females and 1107 males aged 39-67 years). Total serum adiponectin concentrations were measured by an enzyme-linked immunosorbent assay kit from BioCat (Heidelberg, Germany). Glucose metabolism and insulin sensitivity were evaluated based on the levels of glucose and insulin in oral glucose tolerance test (OGTT) at 5 time points (0, 30, 60, 90,120 minutes) and HOMA-Index. Using the haplotype block structure we selected a maximally informative subset of validated SNPs in the AdipoR1 and AdipoR2 with minor allele frequency of more than 5% and pairwise r² =80% (Figure 1) and additionally two SNPs in each gene were selected based on the literature data about their association with different clinical parameters. Genotyping was done using 5´nuclease allelic discrimination TagMan genotyping method.

Results: Three SNPs in the AdipoR2 showed an association with plasma adiponectin levels (rs1223044, p=0.015; rs11612383, p=0.011 and rs1058322, p=0.022). After adjustment for age, sex and BMI these associations were still statistically significant. Unfortunately, all genotyped SNPs located in the AdipoR1 didn’t show association with plasma adiponectin levels. No associations of genetic polymorphisms in both adiponectin receptor genes were found with HOMA-index and glucose at all time points of OGTT. But rs7514221 in the AdipoR1 showed association with insulin level at 30 minutes of the OGTT (p=0.007) and rs6489326 in the AdipoR2 with insulin at 120 minutes of OGTT (p=0.041). After adjustment for age, sex and BMI the p-values for these associations were even stronger (p-value reached 0.001 for rs7514221 and 0.012 for rs6489326).

Conclusion: Association of genetic polymorphisms in the adiponectin receptor genes with adiponectin and insulin levels confirms the important role of adiponectin receptors in regulation of adiponectin levels, glucose metabolism, insulin signaling and should be further investigated.

Supported by: PMU Nr. E-11/13/066-PAU

Disclosure: L. Kedenko: Grants; Paracelsus Medical University Nr. E-11/13/066 - PAU.

210

Investigating gut epithelial adaptations to metabolic disease

M. Jacobs ¹, Y. Lei¹, N. Haq¹, M. Wilson¹, T. Pullen¹, P. Pavlidis², B. Hayee³, G. Bewick¹;

¹School of Cardiovascular and Metabolic Medicine & Sciences, King's College London, London, UK, ²2School of Immunology and Microbial Sciences, King's College London, London, UK, ³King’s College Hospital NHS Foundation Trust, London, UK.

Background and aims: Roux-en-Y gastric bypass and duodenal mucosal resurfacing are treatments for type II diabetes that prevent nutrient contact with the duodenal epithelium. These interventions promote anti-diabetic effects such as improved blood glucose and insulin levels, however, the physiological mechanisms behind this phenomenon remain largely elusive. This study aims to uncover the transcriptional signature of each small intestinal segment to determine the pathological role of obesity and glucose intolerance on the small intestine (SI).

Materials and methods: RNA-Sequencing (RNA-Seq) of the duodenum, jejunum, and ileum was conducted on mice fed a chow (CTL mice), or a 60% high-fat (DIO mice) diet for 13 weeks, to obtain the transcriptional signature. Differential gene expression analysis, gene set enrichment analysis (GSEA), and leading-edge analysis (LE), were generated in R studio, and Ingenuity Pathway Analysis (IPA) software was used to predict upstream and master regulators of the transcriptional footprint. Blood results from the mice indicate that the DIO mice had significantly higher fasting/fed glucose levels and insulin levels, compared to the CTL mice. To validate the RNA-Seq findings, duodenal organoids were cultured to model the changes in key pathways and their effect on epithelial homeostasis.

Results: The RNA-Seq analysis revealed that the duodenum had the highest amount of significantly differentially expressed genes (n = 3055) followed by the jejunum (n= 2169) and ileum (n= 543) respectively (p <0.05). Principal component analysis plots exhibited clustering according to small intestinal segment and diet, which demonstrates that each intestinal segment had a unique transcriptional footprint which was differentially affected by diet. Interferon signalling and the defence response to virus were some of the pathways most significantly upregulated in the duodenum of DIO mice using GSEA analysis (p<0.1), but they were downregulated in the jejunum and ileum. This was associated with downregulation of claudin genes controlling barrier integrity. ISG15 and IRF7 emerged as key regulators of the interferon signalling pathway (p<0.05) using LE analysis and IPA. IPA matched poly I:C transcriptional signatures to 300 of the genes altered in the duodenum, identifying it as a useful tool to model the duodenal transcriptional response in vitro. In duodenal organoids, treatments with poly I:C promoted a significant dose-dependent increase (p <0.05) in mRNA expression of several interferon-related genes: IRF7, ISG15, DHX58, IFI44. Simultaneously, barrier function-related genes: CLDN1, CLDN5, and TJP1 showed a significant dose dependent decrease (p<0.05). Barrier degradation was corroborated functionally and was associated with small increases in overall apoptosis.

Conclusion: The obese mouse duodenum has a distinct transcriptional identity as highlighted by the upregulation of interferon signalling and defence response to virus pathways in GSEA. Obesity may trigger an epithelial innate immune response which may damage epithelial integrity. Analogous to the ‘leaky gut’ hypothesis in the colon, this may offer another route for pathological signalling in the periphery which could aggravate inflammation and metabolic dysregulation.

Supported by: MRC DTP

Disclosure: M. Jacobs: Grants; Fractyl Health Incorporated.

OP 36 Cardiovascular risk, complications and fitness

211

Detailed metabolic biomarker profiling improves cardiovascular risk assessment in type two diabetes patients already receiving statin therapy

H. Julkunen, J. Barrett, P. Würtz;

Nightingale Health Plc, Helsinki, Finland.

Background and aims: Type 2 diabetes patients are at an increased risk of developing cardiovascular complications. Identifying individuals who remain at high cardiovascular risk even when using statins can guide second line options for primary prevention. Here, we assess the utility of detailed metabolic biomarker profiling in predicting the first onset of 3-point major adverse cardiovascular events (MACE) in type two diabetes patients already receiving statin therapy.

Materials and methods: Circulating biomarkers, including lipids, fatty acids, amino acids, glycolysis metabolites and inflammation markers were measured by a low-cost nuclear magnetic resonance (NMR) metabolomics assay in around 275 000 plasma samples from the UK Biobank. The study population evaluated here comprised 9 310 individuals having prevalent type 2 diabetes and using statins. During a prospective 10-year follow-up, 937 first occurrence MACE events were obtained from national hospital and death registries for the study population. Using multivariable Cox proportional hazards regression modelling, we derived a 10-year risk score for the first onset of 3-point MACE. We assessed the predictive improvements of adding the metabolic biomarkers to a model of established risk factors and clinical chemistry cholesterol measurements (age, sex, body mass index, systolic and diastolic blood pressure, smoking status, total cholesterol, LDL and HDL cholesterol).

Results: Adding the metabolic biomarkers to a comprehensive model of established risk factors and cholesterol measures improved the prediction of 3-point MACE events with area under the receiver operating characteristic curve (AUC) increasing from 0.65 to 0.70. Using only the metabolic biomarkers obtained from the single NMR assay, along with age and sex, yielded a performance close to that of the full model incorporating established risk factors, cholesterol measures and metabolic biomarkers, achieving an AUC of 0.69. Even more substantial improvements were observed when looking at the high-risk tails of the risk scores. When comparing the highest risk decile to the remaining population of diabetes patients using statins, adding the metabolic biomarkers to a comprehensive model of established risk factors and cholesterol measures increased the hazard ratio from 2.2 (95% CI 1.7-2.7) to 3.7 (95% CI 3.0-4.5), i.e. reaching an even higher risk elevation than often seen from familial hypercholesterolemia.

Conclusion: Circulating metabolic biomarkers enhanced the prediction of major adverse cardiovascular events beyond established risk factors and cholesterol measurements among type 2 diabetes patients undergoing statin therapy. From a translational perspective, this may complement the residual cardiovascular risk assessment beyond established measures, while simultaneously informing on the risk of other diabetes-related conditions.

Disclosure: H. Julkunen: Employment/Consultancy; Employee of Nightingale Health. Stock/Shareholding; Hold stock options in Nightingale Health.

212

Effects of type 2 diabetes and obesity on cardiopulmonary performance

L. Nesti, N. Pugliese, L. Santoni, M. Chiriacò, L. Sacchetta, D. Tricò, A. Natali;

Università di Pisa, Pisa, Italy.

Background and aims: Measured through oxygen uptake at peak exercise (VO_2peak), effort intolerance is highly prevalent in patients with type 2 diabetes (T2D), wherein it predicts incident heart failure. However, its quantification and pathophysiological causes remain debated.

Materials and methods: Patients at high cardiovascular risk, with or without T2D, underwent pulmonary spirometry and a thorough cardiovascular characterization, including maximal combined echocardiography-cardiopulmonary exercise test (eCPET) to exclude pulmonary and/or cardiovascular disease. Overt microvascular complications were excluded as well. To isolate the effects of T2D and to identify the underpinning mechanisms, multivariable models were built accounting for the major pathophysiological determinants of VO_2peak normalized for lean body weight (LBW), since fat mass has a neutral impact on oxygen uptake.

Results: A total of 109 patients with T2D and 97 non-diabetic controls (non-DM) were included in the analysis. The two groups had similar baseline characteristics except for higher body mass index (BMI) among T2D, with similar LBW. During exercise, T2D patients achieved lower VO_2peak than non-diabetic subjects (T2D 26.1±5.4 vs non-DM 29.5±9.6 mL/min/kg_LBW); subclinical cardiopulmonary impairments were observed, namely chronotropic incompetence, subclinical systolic dysfunction, and more marked cardiac remodeling. The main determinants of VO_2peak were tested alone and in combination to identify the one(s) undermining the independent effect of T2D. Presence of T2D was an independent negative determinant of VO_2peak in all models, maintaining statistical significance and a strong coefficient of determination (Figure 1). Overall, BMI reduced exercise capacity by 0.2 mL/min/kg_LBW per unit, an effect that is largely explained by a combination of chronotropic incompetence, reduced peripheral oxygen extraction, impaired systolic reserve, and ventilatory (in)efficiency. On the contrary, T2D independently reduces VO_2peak by 1 mL/min/kg_LBW and this effect is only marginally explained by any or all the major determinants of cardiopulmonary fitness, which are additive to the effect of T2D per se.

Conclusion: T2D displays an independent negative effect on VO_2peak even in the absence of cardiac or pulmonary disease. This effect is additive to all the other pathophysiological determinants of oxygen uptake, including obesity.

Disclosure: L. Nesti: None.

213

Increase in cardiorespiratory fitness during a lifestyle intervention associates with lower progression and higher regression of subclinical atherosclerosis

K. Kantartzis ^1,2, J. Machann^1,3, F. Schick^1,3, A. Birkenfeld^1,2, L. Fritsche¹, A. Peter^1,4, A. Fritsche^1,2, N. Stefan^1,2, A. Lehn-Stefan¹;

¹Institute of Diabetes Research and Metabolic Diseases (IDM) of the Helmholtz Centre Munich at the University of Tübingen, Tübingen, Germany, ²Department of Internal Medicine IV, University Hospital of Tübingen, Tübingen, Germany, ³Section of Experimental Radiology, Department of Diagnostic and Interventional Radiology, University Hospital of Tübingen, Tübingen, Germany, ⁴Institute for Clinical Chemistry & Pathobiochemistry, Department for Diagnostic Laboratory Medicine, University Hospital of Tübingen, Tübingen, Germany.

Background and aims: Increased carotid intima-media thickness (cIMT) is a marker of subclinical atherosclerosis and cIMT is being used to track the regression, arrest or progression of atherosclerosis. Lifestyle intervention, in general, is effective for improving cardiovascular risk factors. However, there is a large variability in these responses. As high cardiorespiratory fitness (CRF) protects from cardiovascular disease and is closely associated with mitochondrial function, an important determinant of lipid oxidation, we determined, whether change of CRF associates with the change of cIMT during a lifestyle intervention.

Materials and methods: A total of 208 subjects at risk for type 2 diabetes, who underwent 9 months of lifestyle intervention in the TULIP study (diet modification and increase in physical activity), and had measurement of CRF, cIMT, glucose and lipid metabolism and precisely measured body fat mass and fat distribution, were studied. Insulin sensitivity was estimated during a 75g oral glucose tolerance test. Total body-, visceral- and liver-fat were measured by magnetic resonance (MR) imaging and ¹H-MR spectroscopy. CRF was measured during an incremental cycle exercise (maximal aerobic capacity-VO_2max) test and cIMT was measured using high resolution ultrasound.

Results: During the intervention mean total fat mass, visceral fat mass and liver fat content decreased and insulin sensitivity, habitual physical activity (HPA) and CRF improved (all p<0.0001). Mean cIMT decreased (p=0.009), however, there was a large variability in the change of cIMT. When change of cIMT was divided into tertiles, subjects in the upper tertile had a mean (SD) increase of cIMT of 14 (9) %, while subjects in the lower tertile had a mean decrease of cIMT of 16 (7) %. In univariate analyses only changes of HPA (r=−0.16, p=0.02) and CRF (r=−0.13, p=0.06) associated with change of cIMT. In multivariate regression models, change of cIMT, adjusted for cIMT at baseline, age and sex, only associated with adjusted (for the respective baseline parameter) change of CRF (std. ß −0.18, p=0.0089), but not with change of HPA (std. ß −0.05, p=0.56). Larger adjusted increase of CRF was associated with lower odds for having a progression of cIMT (OR per SD increase [for being in the upper tertile of change of cIMT] 0.05 [95% CI 0.006−0.42], p=0.004) and with higher odds for having a regression of cIMT (OR per SD increase [for being in the lower tertile of change of cIMT] 13 [95% CI 2−105], p=0.01).

Conclusion: We provide novel data that measurement of CRF during a lifestyle intervention may help predicting the progression and regression of subclinical atherosclerosis.

Disclosure: K. Kantartzis: None.

214

Hypoglycaemia-related disorders are associated with preclinical atherosclerosis in patients with type 1 diabetes without high cardiovascular risk

A. Mesa ¹, M. Giménez¹, V. Perea², C. Viñals¹, J. Blanco¹, I. Vinagre¹, T. Serés-Noriega¹, L. Boswell^1,3, E. Esmatjes¹, I. Conget¹, A. Amor¹;

¹Endocrinology and Nutrition Department, Hospital Clínic de Barcelona, Barcelona, Spain, ²Endocrinology and Nutrition Department, Hospital Mútua de Terrassa, Terrassa, Spain, ³Althaia University Health Network, Manresa, Spain.

Background and aims: Patients with type 1 diabetes (T1D) have a high cardiovascular disease (CVD) risk, despite tight glycemic control. Recurrent hypoglycemia has been postulated as a potential CVD risk factor in T1D. We evaluated the relationship between severe hypoglycemia (SH) and/or impaired awareness of hypoglycemia (IAH) with the presence of preclinical atherosclerosis in this population.

Materials and methods: Cross-sectional study in T1D patients without CVD and with at least one of the following: ≥40 years, diabetic kidney disease or ≥10 years of T1D duration with CVD risk factors. CVD risk was estimated according to the Steno T1 Risk Engine (Steno-Risk) (<10% low, 10-20% intermediate, ≥20% high). Carotid plaque (intima-media thickness ≥1.5 mm) was evaluated by standardized ultrasonography protocol. Logistic regression models, adjusted for age, sex, T1D duration, and other CVD factors, were constructed to test the independent association between carotid plaque presence and history of SH (≥1 event in the last two years) and hypoglycemia awareness assessed by Clarke questionnaire. We further explored whether the inclusion of SH and/or CQ score in the Steno-Risk equation improved the identification of individuals with plaque (ROC curve).

Results: We included 634 patients (52.4% men, mean age 48.3±10.8 years, T1D duration 27.4±11.1 years, 5-year mean HbA1c 7.5% [7.0-8.0], 39.9% harboring plaque). A stepped increase in the presence of plaque according to Steno-Risk category was observed (13.5%, 37.7%, and 68.7%, for low, moderate, and high risk, respectively; p<0.001). Although no significant association was observed in the whole sample, both SH history (OR 4.4 [1.3-14.6], p=0.015) and CQ score (OR 1.7 [1.2-2.2], p=0.001) were associated with plaque presence in patients with low risk (n=192; both SH*Steno-Risk and Clarke*Steno-Risk interaction p<0.05; Figure). CQ score was also associated with plaque burden in low-moderate risk subjects (n=436; ≥2 plaque: OR 1.2 [1.0-1.5], p=0.031; ≥3 plaque: OR 1.4 [1.1-2.0], p=0.025). The inclusion of SH and CQ score in Steno-Risk significantly improved the identification of subjects at low-risk with atherosclerosis (area under the curve: 0.658 vs. 0.576; p=0.036).

Conclusion: In patients with T1D without high CVD risk, SH and hypoglycemia awareness assessment score were independently associated with the presence of preclinical atherosclerosis. These results suggest that, especially in those patients with few CVD risk factors, hypoglycemia-related disorders might be useful to identify those who would benefit most from an intensive approach.

Disclosure: A. Mesa: None.

215

Mortality in the Swedish Obese Subjects (SOS) study over up to 30 years in relation to 2-year diabetes remission after bariatric surgery or usual care

K. Sjöholm ¹, B. Carlsson¹, P. Jacobson¹, C. Karlsson¹, J. Andersson-Assarsson¹, F. Kristensson¹, S. Ahlin¹, P.-A. Svensson¹, M. Taube¹, I. Näslund², K. Karason¹, M. Peltonen³, L. Carlsson¹;

¹Sahlgrenska Academy at Gothenburg University, Gothenburg, Sweden, ²Faculty of Medicine and Health, Örebro University, Örebro, Sweden, ³Chronic Disease Prevention Unit, National Institute for Health and Welfare, Helsinki, Finland.

Background and aims: People with obesity and concomitant type 2 diabetes have reduced life expectancy, which may to some extent be explained by an increased risk of developing cardiovascular diseases and cancer. Here, we examined whether 2-year diabetes remission after bariatric surgery or usual care is associated with long-term mortality in participants with baseline diabetes in the Swedish Obese Subjects study.

Materials and methods: This report includes 586 participants with obesity and concomitant type 2 diabetes; 338 underwent bariatric surgery and 248 received usual obesity care. At inclusion, age was 37-60 years and BMI was ≥34 kg/m² in men and ≥38 kg/m² in women. Participants were recruited between September 1, 1987, and January 31, 2001, and median follow up was 26.2 years (interquartile range 22.7-28.7).

Results: In the full cohort, 284 participants were in remission at the 2-year examination whereas 302 were not. During follow-up, mortality rates were 16.6 deaths per 1000 person-years (95% CI:13.7-20.1) in the remission subgroup and 26.0 deaths per 1000 person-years (95% CI:22.2-30.4) in the non-remission subgroup (adjusted hazard ratio (HRadj) 0.59, 95% CI:0.46-0.77, p<0.001). Moreover, 2-year remission was associated with reduced mortality in both the surgery and usual care groups (HRadj=0.57 (95% CI:0.39-0.83), p=0.003 and HRadj=0.53 (95% CI:0.32-0.90), p=0.018, respectively). In the full cohort, cardiovascular mortality was markedly decreased in the remission subgroup (sub-HRadj 0.53, 95% CI:0.35-0.81, p=0.003).

Conclusion: Remission of type 2 diabetes after bariatric surgery or usual obesity care is associated with reduced long-term mortality, mainly due to decreased cardiovascular death.

Clinical Trial Registration Number: NCT01479452

Supported by: Swedish Research Council (LC), ALF agreement (LC, KS, PAS)

Disclosure: K. Sjöholm: Grants; Adlerbert research foundation, Swedish state under an agreement between the Swedish government and the county councils (the ALF agreement), Swedish Diabetes Foundation.

216

Exploring potential risk factors for lower limb amputation in people with diabetes: an observational cohort study of 66,565 individuals with diabetes in Sweden

S.P.O. Jansson ¹, S. Ramstrand², M. Carlberg¹, G.A. Johannesson³, A. Hiyoshi¹, G. Jarl⁴;

¹School of Medical Science, Örebro University, Örebro, Sweden, ²School of Health Sciences, Örebro University, Örebro, Sweden, ³Össur Clinics Scandinavia, Stockholm, Sweden, ⁴School of Health Science, Örebro University, Örebro, Sweden.

Background and aims: Risk factors for lower limb amputation (LLA) in individuals with diabetes have been under-studied. We examined how demographic and socioeconomic, medical and lifestyle risk factors may be associated with LLA in people with newly diagnosed diabetes.

Materials and methods: We conducted a cohort study using Swedish national register-linked data. We identified, through the Swedish national diabetes register, all individuals 18 years or older with an incident diabetes diagnosis and no previous amputation from 2007 to 2016. These individuals were followed from the date of the diagnosis to amputation, emigration, death, or the end of the study in 2017, whichever occurred first. Several national Swedish registers were used to obtain data on incident LLA and potential risk factors, including demographic and socioeconomic, medical and lifestyle variables. Variables with more than 40% missing data were excluded from the analysis. The cohort consisted of 66,569 individuals, whereof 133 had an amputation. Cox proportional hazards models were used to obtain hazard ratio (HR) with 95% confidence interval (CI) for associations between demographic and socioeconomic, medical and lifestyle variables and amputation risk. Both unadjusted and mutually adjusted models were fitted.

Results: During the median follow-up time of 4 years there were in total 133 LLA. Based on the model mutually adjusting for all variables, higher age, HR 1.08 (95% CI, 1.05 - 1.10) per year, and being divorced compared with being married, HR 1.67 (1.07 - 2.60) showed positive association. Male sex indicated higher risk, HR 1.57 (1.06 - 2.34). Individuals with an increased foot risk at baseline had increased risk for LLA compared to individuals with healthy feet (neuropathy/angiopathy, HR 4.12 (2.84 - 5.98), previous wounds, HR 8.26 (3.29 - 20.74), ongoing severe foot disease, HR 11.24 (4.82 - 26.23). Insulin treatment compared with diet-only treatment showed HR 2.03 (1.10 - 3.74). Hypertension and HbA1c were not statistically significantly associated with LLA risk. People with obesity had a statistically significant lower risk, HR 0.46 (0.29 - 0.75), compared with individuals with normal weight. Smoking was associated with an increased risk compared with no smoking, HR 1.99 (1.28 - 3.09). Finally, low physical activity (<1 time/week) was associated with an increased risk with a HR of 2.05 (1.30 - 3.23) compared with daily physical activity.

Conclusion: This study found a higher risk for LLA among people with higher age, male sex, who were divorced, who had a higher foot risk group, who were on insulin treatment, lower physical activity levels, and who were smoking. Obesity was associated with lower risk for LLA. Thus, these variables may have important roles in LLA risk among individuals with diabetes.

Supported by: The study was funded by Nyckelfonden and Forskningskommittén at Region Örebro County and by Fonden för rehabilitering och medicinsk forskning

Disclosure: S.P.O. Jansson: None.

OP 37 Inflammation: foe or friend?

217

A longitudinal clinical trajectory analysis examining the accumulation of co-morbidity in people with type 2 diabetes compared with non-type 2 diabetes

R. Qin ¹, K. McCay¹, R. Williams², W. Ollier³, Y. Peng¹, J.J. Warner-Levy⁴, J.M. Gibson⁵, A. Heald⁶;

¹Department of Computing and Mathematics, Manchester Metropolitan University, Manchester, UK, ²University of Manchester, Manchester, UK, ³Department of Life Science, Manchester Metropolitan University, Manchester, UK, ⁴The University of Manchester, Manchester, UK, ⁵School of Medicine, University of Manchester, Manchester, UK, ⁶Diabetes and Endocrinology, Salford Royal NHS Foundation Trust, Salford, UK.

Background and aims: Type 2 diabetes mellitus (T2D) is commonly associated with an increasing complexity of multi-morbidity and related treatment. While some progress has been made in identifying genetic and non-genetic risk factors for T2D, understanding the longitudinal clinical history of individuals before and after T2D diagnosis may provide additional insights into its aetiology and complex trajectory of multi-morbidity.

Materials and methods: In this study, we utilised longitudinal data from the DARE (Diabetes Alliance for Research in England) study to examine the trajectory of clinical conditions in individuals with and without T2D. Data from 1932 individuals (T2D n = 1196 vs matched non-T2D controls n = 736) were extracted and subjected to trajectory analysis over a period of up to 50 years (25 years pre-diagnosis and 25 years post-diagnosis). Conditions were combined for further analysis, specifically, we plotted the trajectories of Chronic Obstructive Pulmonary Disease (COPD), Asthma, Major Mental Disorder, Hypertension, Stroke, Chronic Kidney Disease (CKD), Respiratory Tract Infection, Retinopathy, Heart Conditions (Heart Failure, Myocardial Infarction, Angina, Coronary Angioplasty, Coronary Artery Bypass Graft (CABG), and Coronary Heart Disease (CHD)), and Eye, Nose, and Throat Infections (Pharyngitis, Sinusitis, and Conjunctivitis). The mean age of diagnosis of T2D was 52.9 years (95% CI 52.0 to 53.8).

Results: Our analysis revealed that, in the years leading up to T2D diagnosis, individuals who eventually received a T2D diagnosis consistently exhibited a considerable increase in several clinical phenotypes. Additionally, immediately prior to T2D diagnosis, a significantly greater prevalence of hypertension (35%), respiratory tract infection (34%), heart conditions (17%), eye, nose, and throat infections (19%), and asthma (12%) was observed. The corresponding trajectory of each of these conditions was much less dramatic in the matched control group. Post-T2D diagnosis, the proportions of T2D individuals exhibiting hypertension, CKD, retinopathy, and infections climbed rapidly before plateauing. Additionally, heart conditions and asthma continued to increase in this group.

Conclusion: These findings provide novel insights into the onset and natural progression of T2D, suggesting an early phase of inflammation related disease activity before any clinical diagnosis of T2D is made. Further studies on a larger cohort of patients are suggested to explore the possibility of establishing associated predictive risk scores.

Disclosure: R. Qin: None.

218

Mediation of the association between abdominal adiposity and subclinical inflammation in type 2 diabetes

S.L. Domazet ^1,2, T.B. Olesen¹, J.V. Stidsen¹, C.K. Svensson^3,2, J.S. Nielsen¹, R.W. Thomsen², N. Jessen⁴, P. Vestergaard⁵, M.K. Andersen⁶, T. Hansen⁶, C. Brøns⁷, V.H. Jensen⁷, A. Vaag^7,8, M.H. Olsen³, K. Højlund¹;

¹Steno Diabetes Center Odense, Odense, Denmark, ²Department of Clinical Epidemiology, Aarhus, Denmark, ³Department of Internal Medicine, Steno Diabetes Center Zealand, Holbæk, Denmark, ⁴Steno Diabetes Center Aarhus, Aarhus, Denmark, ⁵Steno Diabetes Center North Denmark, Aalborg, Denmark, ⁶Novo Nordisk Foundation Center for Basic Metabolic Research, Copenhagen, Denmark, ⁷Steno Diabetes Center Copenhagen, Copenhagen, Denmark, ⁸Lund University, Malmö, Sweden.

Background and aims: The association between adiposity and risk of type 2 diabetes and cardiovascular disease may be mediated by systemic subclinical inflammation. However, it is unknown whether adiposity itself or other causal and/or mediating traits are the main drivers of subclinical inflammation in type 2 diabetes. We aimed to determine the extent to which adiposity is associated with subclinical inflammation in persons with newly diagnosed type 2 diabetes, and if so, to understand the extent to which this may be mediated by physical activity, fasting hyperinsulinemia, glycemic control, plasma lipids, blood pressure, and/or other comorbidities.

Materials and methods: Using Meso Scale Discovery (MSD) assays, we measured plasma levels of the inflammatory biomarkers IL-6 (n=9,105), TNF-α (n=9,112) and hsCRP (n=9,680) in participants from the Danish Center for Strategic Research in Type 2 Diabetes (DD2) cohort. Applying a cross-sectional setup, we conducted multiple mediation analysis using structural equation modelling with post hoc estimations of total, direct and indirect effects.

Results: Waist circumference as a proxy for abdominal adiposity was positively associated with all markers of subclinical inflammation. Hence, one standard deviation (SD) increase in waist circumference (equal to 15 cm) was associated with 0.20 (95% CI 0.18; 0.22), 0.30 (95% CI 0.28; 0.32) and 0.38 (95% CI 0.36; 0.40) SD increase in TNF-α (equal to 1.5 pg/mL), IL-6 (equal to 4.4 pg/mL) and hsCRP (equal to 6.9 mg/L), respectively. Using multiple mediation analysis, we found that high fasting C-peptide, low physical activity, high triglycerides, Charlson Comorbidity Index, and high HbA₁c were significant traits involved in mediating the association between waist circumference and levels of IL-6, TNF-α and hsCRP (Table 1). Notably, fasting C-peptide was the quantitatively most important mediator, accounting for 9-25% of the association between abdominal adiposity and subclinical inflammation, followed by physical activity (5-7%) and triglyceride levels (2-6%). Although mediation by comorbidities and HbA₁c reached statistical significance, their impact was minor (1-2%).

Conclusion: In persons with new onset type 2 diabetes, abdominal adiposity is associated with subclinical inflammation, and fasting C-peptide is the quantitatively most important mediating factor suggesting a role for hyperinsulinemia in adiposity-driven inflammation.

Supported by: Danish Agency for Science. The Novo Nordisk. SST. DDA.

Disclosure: S.L. Domazet: None.

219

Interleukin 1β regulates white adipose tissue remodelling by targeting adipocyte precursors

K. Hofwimmer ¹, J. de Paula Souza², N. Subramanian¹, M. Vujičić³, I. Wernstedt Asterholm³, M. Rydén¹, M. Böni-Schnetzler², D.T. Meier², M.Y. Donath², J. Laurencikiene¹;

¹Department of Medicine Huddinge, Karolinska Institutet, Stockholm, Sweden, ²Department of Biomedicine, University of Basel and University Hospital Basel, Basel, Switzerland, ³Department of Physiology/Metabolic Physiology, The Sahlgrenska Academy at University of Gothenburg, Gothenburg, Sweden.

Background and aims: Obesity-associated metabolic diseases, such as type 2 diabetes, are closely linked to white adipose tissue (WAT) hypertrophy, whereas hyperplastic WAT expansion via differentiation of precursors into new adipocytes (adipogenesis) is protective. The role of chronic inflammation in insulin resistance is well established, but little attention has been paid to its physiological role in WAT. We observed that IL-1β is acutely upregulated postprandially most prominently in WAT-resident macrophages, suggesting it is part of a local response to energy influx. Thus, we aimed to investigate a possible physiological, metabolic role of IL-1β in WAT energy storage.

Materials and methods: We used adipocyte-specific IL-1 receptor knockout (IL1R1^Δadi) or whole-body IL1R knockout (IL1R1-KO) mice, fed chow (CD) or high-fat diet (HFD). WAT phenotyping was carried out in both models and glucose tolerance and insulin sensitivity were assessed in IL1R1^Δadi mice. IL1R1-KO mice were injected with EdU and formation of new fat cells was assessed by flow cytometry 9 weeks later. Associations of WAT morphology with expression of IL1R1, IL1B and its antagonist IL1RN were analyzed in a cohort of 56 obese and non-obese women. Mechanistic studies were performed in an in vitro model of differentiating human adipose-derived stem cells, where lipid droplet formation was quantified by high-throughput fluorescent microscopy, and gene/protein expression was measured by RNA-seq, RT-PCR and Western blotting.

Results: IL1R1^Δadi and wild-type (WT) mice showed similar WAT and insulin sensitivity phenotype, whereas IL1R1-KO mice had reduced subcutaneous (sc) and gonadal (g) WAT mass on CD (sc: -61%; p < 0.0001; g: -49%, p < 0.001) and HFD (sc: -45%; p < 0.0001; g: -18%, p < 0.01), compared to WT mice. In human WAT, progenitors were the major expressers of IL1R1, and reduced IL-1 signaling correlated to hypertrophy. In vitro, IL-1β treatment increased lipid droplet formation in differentiating human adipose-derived stem cells by 88% (p < 0.0001). IL-1β upregulated several adipogenesis- and lipid handling-related genes during early, but not late differentiation, and it promoted adipogenesis exclusively when added to early-differentiation-stage cells. The pro-adipogenic, but not pro-inflammatory effect of IL-1β was blocked by chronic pre-treatment and potentiated by acute exposure. IL-1β increased expression of the early adipogenic transcription factors CCAAT/enhancer-binding protein (C/EBP)δ and C/EBPβ, and their DNA-binding was required for its adipogenic effect. HFD-induced formation of new adipocytes was lower in scWAT (-65%, p < 0.05) and gWAT (-56%, p < 0.001) of IL1R1-KO compared to WT mice.

Conclusion: Here, we show that progenitors are the major targets of IL-1 signaling in human WAT, where IL-1β promotes hyperplastic expansion via adipogenesis. This probably has a physiological role in allowing WAT to adapt during caloric excess. We propose that IL-1β exerts this function via its acute postprandial surge, whereas the chronically high levels in obese WAT may counteract this effect and instead exacerbate inflammation.

Supported by: VR Research Grant, Novo Nordisk Foundation Research Grant, Swiss National Science Foundation

Disclosure: K. Hofwimmer: Grants; VR Research Grant, Novo Nordisk Foundation Research Grant.

220

Bariatric arterial embolisation vs bariatric surgery: comparison of effects in obese patients with type 2 diabetes

R.O. Oliveira ^1,2, F.O. Martins¹, S.V. Conde¹, T. Bilhim¹, F.V. Gomes¹, V.S. Reuters³;

¹Nova Medical School, Universidade Nova de Lisboa, Lisbon, Portugal, ²Unidade de Cirurgia Bariátrica e Metabólica CTO, Hospital Cruz Vermelha, Lisbon, Portugal, ³Unidade Cirurgia Bariátrica e Metabólica CTO, Hospital Cruz Vermelha, Lisboa, Portugal.

Background and aims: Global diabetes incidence has reached 9.3%, with over half (50.1%) of adults undiagnosed, with type 2 diabetes (T2D) accounting for about 90% of all people with diabetes (1,2). Obesity is the main risk factor for T2D, due to its relationship with insulin resistance and chronic systemic inflammatory process, leading in turn to altered glucose homeostasis. The relationship between obesity and T2D is so strong that weight loss directly impacts the improvement of glycemic control of these patients achieved with clinical and/or nutritional approaches (2). Our aim is to evaluate the long-term metabolic control, remission of T2D and weight loss after bariatric arterial embolization (BAE) in comparison with bariatric surgery (RYGB).

Materials and methods: On this prospective, randomised, non-blinded, controlled and interventional clinical single-centre study, we compared the impact of RYGB and BAE on metabolic control in patients with