Skip to main content

Congenital beta cell defects are not associated with markers of islet autoimmunity, even in the context of high genetic risk for type 1 diabetes



A key unanswered question in type 1 diabetes is whether beta cells initiate their own destruction or are victims of an aberrant immune response (beta cell suicide or homicide?). To investigate this, we assessed islet autoantibodies in individuals with congenital beta cell defects causing neonatal diabetes mellitus (NDM).


We measured autoantibodies to GAD (GADA), islet antigen-2 (IA-2A) and zinc transporter 8 (ZnT8A) in 242 individuals with NDM (median age diagnosed 1.8 months [IQR 0.39–2.9 months]; median age collected 4.6 months [IQR 1.8–27.6 months]; median diabetes duration 2 months [IQR 0.6–23 months]), including 75 whose NDM resulted from severe beta cell endoplasmic reticulum (ER) stress. As a control cohort we also tested samples from 69 diabetes-free individuals (median age collected 9.9 months [IQR 9.0–48.6 months]) for autoantibodies.


We found low prevalence of islet autoantibodies in individuals with monogenic NDM; 13/242 (5.4% [95% CI 2.9, 9.0%]) had detectable GADA, IA-2A and/or ZnT8A. This was similar to the proportion in the control participants who did not have diabetes (1/69 positive [1.4%, 95% CI 0.03, 7.8%], p=0.3). Importantly, monogenic individuals with beta cell ER stress had a similar rate of GADA/IA-2A/ZnT8A positivity to non-ER stress aetiologies (2.7% [95% CI 0.3, 9.3%] vs 6.6% [95% CI 3.3, 11.5%] p=0.4). We observed no association between islet autoimmunity and genetic risk, age at testing (including 30 individuals >10 years at testing) or diabetes duration (p>0.4 for all).


Our data support the hypothesis that beta cell stress/dysfunction alone does not lead to the production of islet autoantibodies, even in the context of high-risk HLA types. This suggests that additional factors are required to trigger an autoimmune response towards beta cells.

Graphical abstract

figure b


An ongoing debate in type 1 diabetes research is the question of beta cell homicide vs suicide; is beta cell death solely immune-cell driven, or do beta cells trigger their own demise [1, 2]? Several lines of evidence suggest a role for beta cell abnormalities in type 1 diabetes pathogenesis, including HLA class I hyperexpression [3] and irregular expression of immune genes under inflammatory conditions [4]. Reports of nonconventional, immunogenic polypeptides (e.g., defective ribosomal products [DRiPs]) acting as self-antigens in type 1 diabetes also suggests beta cells under endoplasmic reticulum (ER) stress may potentiate autoimmunity [5, 6].

Islet autoantibodies are present in up to 90% of individuals with recent-onset type 1 diabetes [7]. They are not considered pathogenic, but are markers of beta cell autoimmunity, evidenced by their use as markers of active autoimmunity in trials [8]. To examine whether congenital beta cell abnormalities are associated with islet autoantibodies, we measured autoantibodies to GAD (GADA), islet antigen-2 (IA-2A) and zinc transporter 8 (ZnT8A) in 242 individuals with monogenic neonatal diabetes mellitus (NDM). We assessed their association with NDM mechanistic subtype, HLA/polygenic risk for type 1 diabetes and age at sampling.


We studied 242 individuals referred to our laboratory for genetic testing for NDM, in whom a pathogenic variant was identified and sufficient plasma was available for autoantibody testing (electronic supplementary material [ESM] Table 1: female participants: 136 [56%]; median age diagnosed 1.8 months [IQR 0.39–2.9 months]; median age collected 4.6 months [IQR 1.8–27.6 months]; median diabetes duration 2 months [IQR 0.6–23 months]) and 69 diabetes-free individuals (ESM Table 2: female participants: 38 [55%]; median age collected 9.9 months [IQR 9.0–48.6 months]). These samples came from unaffected relatives of individuals with monogenic diabetes who were sent for predictive testing and were found not to have inherited the pathogenic variant. They did not have a family history of type 1 diabetes and their risk of islet autoimmunity was therefore equivalent to the background population risk.

Autoantibodies to GAD, IA-2 and ZnT8 were measured by RIA at the Pacific Northwest Diabetes Research Institute, USA, who are participants of the Islet Autoantibody Standardisation Program. Positivity thresholds were defined by the 99th centile of index scores in 200 healthy control participants.

Individuals were categorised into mechanistic subtypes as follows: beta cell ER stress (biallelic EIF2AK3 or heterozygous INS missense mutations [n=75], which have been assessed by mechanistic studies [9]); a functional defect (gain-of-function ABCC8 or KCNJ11, biallelic GCK and biallelic INS variants [n=110]); pancreatic developmental defects (PTF1A, PDX1 and GATA6 variants [n=7]); or other (GLIS3, IER3IP1 and SLC19A2 variants or 6q24 methylation defect [n=50]) (see ESM Table 3 for detailed descriptions of genetic subtypes). We also grouped participants by sampling age: <4 months (n=106); 4 months–1 year (n=53); 1–5 years (n=28); 5–10 years (n=16); and ≥10 years (n=30) (ESM Fig. 1).

Where sufficient DNA was available (230/242) participants were also assessed for their type 1 diabetes genetic risk score (T1D-GRS) as described previously [10], and HLA-DR status.

We used the Mann–Whitney U test for continuous variables and Fisher’s exact test for categorical variables. Our study was powered to detect a 10% difference in antibody prevalence of the entire cohort compared with the control population and 30% difference in antibody prevalence between subcategories at 80% power and 5% significance.

All study participants gave informed consent or assent was obtained where children were too young and parental consent was provided, in accordance with the declaration of Helsinki. This study was approved by the Genetic Beta Cell Research Bank, Exeter, UK. Ethical approval was provided by the North Wales Research Ethics Committee, UK (IRAS project ID 231760).


We found low prevalence of islet autoantibodies in all mechanistic categories of NDM (Fig. 1a). Out of 242 participants, 13 (5.4% [95% CI 2.9, 9.0%]) had GADA (n=4 [1.7%]), IA-2A (n=6 [2.5%]) and/or ZnT8A (n=4 [1.7%]), compared with one individual who was positive for IA-2A out of 69 diabetes-free individuals (1.4% [95% CI 0.03, 7.8%] p=0.3).

Fig. 1
figure 1

Proportion of individuals positive for islet autoantibodies in those with a congenital monogenic beta cell defect leading to ER stress (n=75), impacting beta cell development (n=7), function (n=110) or other (n=50) and in diabetes-free control participants (n=69) (a). Proportion of individuals with a congenital monogenic beta cell defect positive for islet autoantibodies in those with (n=82) or without (n=146) an HLA risk allele (b). Antibody positive individuals (n=13) were defined as exceeding the 99th centile of index scores in 200 healthy control participants for GADA, IA-2A and/or ZnT8A. Ab neg, antibody negative; Ab pos, antibody positive

To account for duration effects which could result in lower seropositivity over time, we also conducted a subanalysis of participants with <24 months duration. We found a similar prevalence of autoantibodies: 10/175 (5.7% [95% CI 2.8, 10.3%]) had GADA, IA-2A and/or ZnT8A.

We did not observe any associations between seropositivity and mechanistic subtype, T1D-GRS, age at diagnosis, birthweight or sex (p>0.4 for all comparisons). Individuals with beta cell ER stress had a similar rate of positivity to other aetiologies (2/75 [2.7%, 95% CI 0.3, 9.3%] vs 11/167 [6.6%, 95% CI 3.3, 11.5%] p=0.4).

Age at sampling also did not influence autoantibody positivity in this cohort. Of 106 very young patients (<4 months), 7 (6.6% [95% CI 2.7, 13.1%]) were positive for GADA, IA-2A and/or ZnT8A compared with 1/53 (1.9% [95% CI 0.04, 10.0%] p=0.3), 4/28 (14.3% [95% CI 4.0, 32.7%] p=0.2), 0/16 (p=0.6) and 1/30 (3.3% [95% CI 0.08, 17.2%] p=0.7) of those aged 4 months–1 year, 1–5 years, 5–10 years and ≥10 years, respectively (ESM Fig. 1b).

Positivity was not linked with HLA risk alleles; 3/82 (3.7% [95% CI 0.7, 10.3%]) participants with HLA-DR3 and/or DR4 had detectable islet autoantibodies compared with 10/146 (6.8% [95% CI 3.3, 12.2%]) without DR3 or DR4 (p=0.4). Of these, 3/3 and 7/10 individuals were aged <5 years at sampling (Fig. 1b). Of the 75 participants with a beta cell ER stress causing variant, 26 (34.7% [95% CI 24.0, 46.5%]) had an HLA risk allele, 1 (3.8% [95% CI 0.1, 19.6%]) of which was positive for IA-2A.

One participant (age at diagnosis 1.6 months, duration 1.3 months) was GADA and IA-2A positive, but was heterozygous for the known KCNJ11 pathogenic variant, p.(Arg201His) [11]. They had low T1D-GRS (5th centile of type 1 diabetes control group), low birthweight (−2.2 SDs) and successfully transferred from insulin to sulfonylurea treatment after diagnosis.


We did not see a strong association between monogenic beta cell defects, including those causing beta cell death resulting from ER stress, and an islet-specific humoral response as measured against three islet autoantigens. Autoantibody prevalence was similar in our diabetes-free control participants and monogenic NDM cohorts. Our data do not negate evidence indicating that beta cell abnormalities, such as incorrect insulin processing [12], HLA class I hyperexpression [3] and irregular expression of immune genes [4] contribute to the pathogenesis of type 1 diabetes. Nevertheless, we show that severe beta cell stress/dysfunction, even in the context of high-risk HLA types, is unlikely to be sufficient to cause islet autoimmunity as measured by autoantibodies. This suggests that additional factors are necessary to initiate islet autoimmunity.

We defined autoimmunity by the presence of any of three accepted type 1 diabetes islet autoantibodies (GADA, IA-2A and ZnT8A). We cannot rule out novel autoantibodies specific to alternative autoantigens [5]. Indeed, defective ER function (e.g., caused by coding INS variants) can lead to accumulation of aberrantly processed molecules (such as insulin or DRiPs), which could function as neo-autoantigens. These specific misfolded INS proteins in individuals with missense variants could lead to specific autoantibodies which we are unable to detect. We are also unable to assess insulin autoimmunity in our cohort as all participants were insulin treated, meaning we would be unable to distinguish between insulin autoantibodies (IAA) recognising exogenous insulin (immunity) and IAA to endogenous insulin (autoimmunity) [13, 14].

Our study is cross-sectional; we cannot rule out the development of autoantibodies after participants were sampled. Islet autoantibodies often precede diagnosis with seroconversion peaking in early childhood (<5 years) [15]. The low prevalence in very young individuals may reflect the inefficiency of antibody production by the developing immune system. Although we do not have longitudinal samples, age at sampling ranged from 1 week to 58 years (ESM Table 1), and we saw no association between positivity and age at sampling.

Despite our best attempts at follow-up, we are also unable to exclude coincidental islet autoimmunity and type 1 diabetes in one individual with multiple autoantibodies. Their clinical presentation was in keeping with monogenic NDM and their diabetes remained controlled by sulfonylurea treatment 3 months after insulin was discontinued, supporting that they did not have severe insulin deficiency at the time of last follow-up (7 months), but it is possible that they have subsequently progressed to this.

To our knowledge, this is the largest assessment of evidence of autoimmunity in individuals with congenital beta cell defects causing NDM. A 2007 study of 11 individuals with NDM caused by pathogenic KCNJ11 variants found that of nine individuals with long duration (>10 years), 5 (56%) had at least one islet antibody [16]. In our cohort, 1 of 30 participants with ≥10 years duration (who had a pathogenic INS missense variant) was positive for an islet autoantibody (IA-2A). It is possible that this disparity in prevalence is attributable to the different autoantibodies tested in each study, positivity thresholds, testing methodologies used or the numbers of individuals assessed.

Our study focused on investigating the humoral autoimmune response in NDM. Islet autoantibodies are markers of autoimmunity rather than pathogenic, and future work could look for evidence of T cell mediated beta cell autoimmunity. Alternatively, it may be possible to identify novel autoantibodies to beta cell antigens which are specific to NDM, such as misfolded insulin epitopes found in individuals with monoallelic dominant INS variants.

We found 5.4% of participants with monogenic NDM were positive for at least one of either GADA, IA-2A or ZnT8A. Additionally, individuals with NDM due to monogenic autoimmunity commonly have islet autoantibodies [17]. The presence of islet autoantibodies should therefore not prevent comprehensive genetic testing for NDM in patients diagnosed aged <6 months.

In conclusion, we found low prevalence of islet autoantibodies in participants with a congenital beta cell defect, including those with severe beta cell ER stress. The number of these individuals was not significantly higher than seen in age and geographically matched diabetes-free control participants, implying that beta cell stress/dysfunction in isolation is unlikely to trigger beta cell autoimmunity.

Data availability

Access to data is open only through collaboration. Requests for collaboration will be considered following an application to the Genetic Beta Cell Research Bank ( Contact by email should be directed to the Lead Nurse, Bridget Knight (



Defective ribosomal products


Endoplasmic reticulum


GAD antibody


Islet antigen-2 antibody


Insulin autoantibodies


Neonatal diabetes mellitus


Type 1 diabetes genetic risk score


Zinc transporter 8 antibody


  1. Bottazzo GF (1986) Death of a Beta-cell - homicide or suicide? Diabetic Med 3(2):119–130.

    CAS  Article  PubMed  Google Scholar 

  2. Roep BO, Thomaidou S, van Tienhoven R, Zaldumbide A (2021) Type 1 diabetes mellitus as a disease of the β-cell (do not blame the immune system?). Nat Rev Endocrinol 17(3):150–161.

    CAS  Article  PubMed  Google Scholar 

  3. Marroqui L, Dos Santos RS, de Beeck AO et al (2017) Interferon-alpha mediates human beta cell HLA class I overexpression, endoplasmic reticulum stress and apoptosis, three hallmarks of early human type 1 diabetes. Diabetologia 60(4):656–667.

    CAS  Article  PubMed  Google Scholar 

  4. Eizirik DL, Sammeth M, Bouckenooghe T et al (2012) The human pancreatic islet transcriptome: expression of candidate genes for type 1 diabetes and the impact of pro-inflammatory cytokines. PLoS Genet 8(3):17.

    CAS  Article  Google Scholar 

  5. Kracht MJL, van Lummel M, Nikolic T et al (2017) Autoimmunity against a defective ribosomal insulin gene product in type 1 diabetes. Nat Med 23(4):501.

    CAS  Article  PubMed  Google Scholar 

  6. Baker RL, Jamison BL, Haskins K (2019) Hybrid insulin peptides are neo-epitopes for CD4 T cells in autoimmune diabetes. Curr Opin Endocrinol Diabetes Obes 26(4):195–200.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  7. Hameed S, Ellard S, Woodhead HJ et al (2011) Persistently autoantibody negative (PAN) type 1 diabetes mellitus in children. Pediatr Diabetes 12(3):142–149.

    CAS  Article  PubMed  Google Scholar 

  8. So M, Speake C, Steck AK et al (2021) Advances in type 1 diabetes prediction using islet autoantibodies: beyond a simple count. Endocr Rev.

  9. Cnop M, Toivonen S, Lgoillo-Esteve M, Saipea P (2017) Endoplasmic reticulum stress and eIF2 alpha phosphorylation: the Achilles heel of pancreatic beta cells. Mol Metab 6(9):1024–1039.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  10. Oram RA, Patel K, Hill A et al (2016) A type 1 diabetes genetic risk score can aid discrimination between type 1 and type 2 diabetes in young adults. Diabetes Care 39(3):337–344.

    CAS  Article  PubMed  Google Scholar 

  11. De Franco E, Saint-Martin C, Brusgaard K et al (2020) Update of variants identified in the pancreatic β-cell K(ATP) channel genes KCNJ11 and ABCC8 in individuals with congenital hyperinsulinism and diabetes. Hum Mutat 41(5):884–905.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  12. Leete P, Oram RA, McDonald TJ et al (2020) Studies of insulin and proinsulin in pancreas and serum support the existence of aetiopathological endotypes of type 1 diabetes associated with age at diagnosis. Diabetologia.

  13. Greenbaum CJ, Palmer JP (1991) Insulin-antibodies and insulin autoantibodies. Diabetic Med 8(2):97–105.

    CAS  Article  PubMed  Google Scholar 

  14. Davis SN, Thompson CJ, Peak M, Brown MD, Alberti KG (1992) Effects of human insulin on insulin binding antibody production in nondiabetic subjects. Diabetes Care 15(1):124–126.

    CAS  Article  PubMed  Google Scholar 

  15. Ziegler AG, Hummel M, Schenker M, Bonifacio E (1999) Autoantibody appearance and risk for development of childhood diabetes in offspring of parents with type 1 diabetes: the 2-year analysis of the German BABYDIAB study. Diabetes 48(3):460–468

    CAS  Article  PubMed  Google Scholar 

  16. Gach A, Wyka K, Malecki MT et al (2007) Islet-specific antibody seroconversion in patients with long duration of permanent neonatal diabetes caused by mutations in the KCNJ11 gene. Diabetes Care 30(8):2080–2082.

    CAS  Article  PubMed  Google Scholar 

  17. Johnson MB, Patel KA, De Franco E et al (2018) A type 1 diabetes genetic risk score can discriminate monogenic autoimmunity with diabetes from early-onset clustering of polygenic autoimmunity with diabetes. Diabetologia 61(4):862–869.

    Article  PubMed  PubMed Central  Google Scholar 

Download references


The authors thank the referring clinicians, individuals and their families. We also thank L. Weymouth and J. Kirkwood (Exeter National Institute for Health Research Clinical Research Facility, Royal Devon and Exeter NHS Foundation Trust, Exeter, UK).

Authors’ relationships and activities

The authors declare that there are no relationships or activities that might bias, or be perceived to bias, their work.

Contribution statement

MBJ, RAO and ATH conceived the study. RCW collated and analysed the data and wrote the first draft of the manuscript. WAH performed autoantibody measurement and analysis. BOR contributed to interpretation of data. KAP analysed type 1 diabetes genetic risk scores and called HLA variants. BR, RD and MH contributed to acquisition of clinical and biochemical data, and analysis and interpretation of data. EDF, SE and SEF were involved in acquisition, analysis and interpretation of genetic data. All authors reviewed and approved the final version of the manuscript. MBJ is the guarantor of this study and takes full responsibility for the contents of this manuscript.

This study is supported by the NIHR Exeter Clinical Research Facility. The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care.


This work was supported by a Wellcome Trust Senior Investigator Award to SE and ATH (grant number 098395/Z/12/Z). ATH is a National Institute for Health Research (NIHR) senior investigator. WAH is supported by National Institutes of Health National Institute of Diabetes and Digestive and Kidney Diseases (NIH NIDDK) grant numbers DK063829 and DK017047 and JDRF grant numbers 3-SRA-2019-827-S-B and 2-SRA-2020-964-S-B. BOR is supported by grants from the Dutch Diabetes Research Foundation, Stichting DON, the European Commission, JDRF and the Wanek Family Project for Type 1 diabetes. KAP is funded by a Wellcome Trust Fellowship (219606/Z/19/Z). EDF is a Diabetes UK RD Lawrence Fellow (19/005971). SEF has a Sir Henry Dale Fellowship jointly funded by the Wellcome Trust and the Royal Society (105636/Z/14/Z). RAO is a Diabetes UK Harry Keen Fellow (16/0005529). MBJ is the recipient of an Exeter Diabetes Centre of Excellence Independent Fellowship funded by Research England’s Expanding Excellence in England (E3) fund. Additional support came from the Helmsley Foundation’s Breakthrough Initiative, Diabetes Research and Wellness Foundation, University of Exeter and the NIHR Exeter Clinical Research Facility.

Author information

Authors and Affiliations



Corresponding author

Correspondence to Matthew B. Johnson.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Please see electronic supplementary material for details of the EXE-T1D Consortium.

Supplementary information


(PDF 415 kb)

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Wyatt, R.C., Hagopian, W.A., Roep, B.O. et al. Congenital beta cell defects are not associated with markers of islet autoimmunity, even in the context of high genetic risk for type 1 diabetes. Diabetologia 65, 1179–1184 (2022).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:


  • ER stress
  • HLA
  • Islet autoantibodies
  • Monogenic
  • Neonatal diabetes