Insulin trigger for diabetes

Type I diabetes occurs when the immune system destroys crucial cells in the pancreas. But what prompts the body to turn against itself so disastrously? It seems that insulin is the key.

Autoimmune disorders are caused by aberrant immune responses directed towards the body's own proteins (or ‘autoantigens’). In type I diabetes, for example, immune cells called lymphocytes react against and destroy the β-cells in the pancreas of genetically susceptible individuals (Fig. 1). As these cells produce insulin — the hormone that helps to regulate glucose metabolism — their loss leads to the chaotic blood glucose levels that characterize diabetes and that have serious consequences. How the lymphocytes target β-cells has been unclear, but work from Nakayama et al.1 and Kent et al.2 in this issue (on pages 220 and 224) shows that insulin is itself an autoantigen that initiates the immune response leading to diabetes.

Figure 1: Development of type I diabetes.

Genetic predisposition determines almost entirely whether a person will develop immune reactivity against insulin-producing β-cells in the pancreas. However, environmental factors such as nutrition and infections can have a major impact on whether type I diabetes manifests itself clinically; this occurs after 80–90% of the β-cells have been destroyed. The remnants of the β-cells are transported to the pancreatic draining lymph node (PDLN), where the ensuing autoimmune process is thought to be coordinated. Debris from the β-cells is picked up by antigen-presenting cells (APC) and ‘displayed’ to immune cells called lymphocytes (L), prompting the lymphocytes either to kill more β-cells or to signal further immune responses. Nakayama et al.1 and Kent et al.2 show that the initial immune trigger is insulin, although other immune targets could be involved later in the disease.

Although many autoantigens are thought to be involved in type I diabetes, conclusive evidence that any of these instigate or propagate autoimmune responses has remained elusive. Historically, however, insulin has been a good candidate. For example, lymphocytes isolated from ‘non-obese diabetic’ (NOD) mice — an animal model of type I diabetes — can recognize segments of the insulin molecule3,4. In addition, autoantibodies against insulin are found in the blood of NOD mice and human diabetic patients, well before the clinical symptoms of the disease appear. The onset of clinical diabetes occurs after more than 90% of β-cells have been destroyed, but the insulin antibodies arise after only a small fraction of β-cells have been attacked, providing a useful early marker for the occurrence of autoimmunity.

Nakayama and colleagues1 now provide convincing evidence that insulin is an essential autoimmune target in the initiation of diabetes. They stopped all natural insulin production in NOD mice by eliminating the genes that encode the two chains of insulin. So that the mice would not die, they reinstated a modified gene to produce insulin that is hormonally active but is not recognized by lymphocytes. Normally, NOD mice show an immune response against the β-cells — there are antibodies against the β-cells in their blood, and lymphocytes infiltrate into the pancreas, surrounding the β-cells before killing them. But the mice with the modified insulin showed no signs of any immune response against the β-cells, and they did not develop diabetes.

One question that remains is whether insulin is the crucial autoantigen only in the initiation of the disease, or whether other β-cell proteins become targets of the autoimmune response once the process has been set in motion. Lymphocytes do recognize several other targets as β-cell destruction progresses (by a process termed ‘antigenic spreading’)5. Notably, damped down expression of these antigens in β-cells, or elimination of the lymphocytes that recognize them, can lead to a significant reduction in the incidence of diabetes, but never to a complete absence of β-cell-specific autoimmune attacks6,7. This suggests that although insulin is crucial for the initiation of β-cell destruction, other antigens might be involved in subsequent stages of the disease.

The relative importance of insulin over other candidate target antigens is underscored by Jaeckel and colleagues8,9. They expressed insulin and glutamic acid decarboxylase (another protein proposed to be an autoantigen in type I diabetes) in the thymus of NOD mice. This meant that the immune system no longer reacted to these proteins. Tolerance to insulin led to a remarkable reduction in the occurrence of diabetes, but tolerance to glutamic acid decarboxylase did not.

How relevant are these findings for human diabetes? Kent and colleagues' results2 support the concept that insulin is also an essential autoantigen in people. They isolated lymphocytes from the pancreatic draining lymph node of patients with type I diabetes, propagated them, and then analysed the proteins that the cells recognized. Remarkably, about 50% of the lymphocytes recognized a piece of the insulin A chain. By contrast, no healthy control subjects showed a similar accumulation of lymphocytes that recognize the insulin A fragment.

Lymphocytes come into contact with the insulin peptide as it is ‘displayed’ on the surface of so-called antigen-presenting cells; these cells collect protein fragments from dying β-cells before convening with the lymphocytes in the pancreatic draining lymph node. Kent et al. found that the cell-surface protein that binds to and displays the insulin A fragment on the antigen-presenting cells is encoded by a gene known to confer genetic susceptibility to diabetes.

Findings from Arif et al.10 also suggest that insulin could be the target of the autoimmune response in humans, providing a potential candidate for treatment. Arif and colleagues found that insulin-reactive lymphocytes from diabetics appeared to have a destructive nature; they release molecules that are harmful to β-cells. Insulin-reactive lymphocytes from healthy individuals, however, seemed mostly to have a regulatory function, given the signalling molecules that they secrete. The idea that self-reactive lymphocytes can have different functions is therapeutically interesting in that it suggests two principal pathways to mitigate the effects of autoimmune lymphocytes: wipe out the aggressive cells, or ensure that more of the cells end up with a regulatory function11,12. Indeed, recent work indicates that inducing such a change in the function of autoreactive lymphocytes from aggressive to regulatory might be possible, not only in mice, but also in some diabetic patients13.

So, we will not only have to learn which proteins and peptides are recognized by autoreactive lymphocytes, but we will also have to see whether this recognition elicits a regulatory or a destructive state. Dysregulation of such responses might ultimately lead to autoimmune diseases such as type I diabetes, for example, if lymphocytes switch from a regulatory to an aggressive state. The identification of a high proportion of insulin-A-specific lymphocytes in human pancreatic lymph nodes gives us a valuable population of autoimmune lymphocytes to study, and highlights the importance of recovering cells directly from human target organs.


  1. 1

    Nakayama, M. et al. Nature 435, 220–223 (2005).

    ADS  CAS  Article  PubMed  PubMed Central  Google Scholar 

  2. 2

    Kent, S. et al. Nature 435, 224–228 (2005).

    ADS  CAS  Article  PubMed  Google Scholar 

  3. 3

    Wegmann, D. R., Norbury-Glaser, M. & Daniel, D. Eur. J. Immunol. 24, 1853–1857 (1994).

    CAS  Article  PubMed  Google Scholar 

  4. 4

    Wong, F. S. et al. Nature Med. 5, 1026–1031 (1999).

    CAS  Article  PubMed  Google Scholar 

  5. 5

    Tian, J. et al. Immunol. Rev. 164, 119–127 (1998).

    CAS  Article  PubMed  Google Scholar 

  6. 6

    Lieberman, S. M. et al. Proc. Natl Acad. Sci. USA 100, 8384–8388 (2003).

    ADS  CAS  Article  PubMed  Google Scholar 

  7. 7

    Yoon, J. W. et al. Science 284, 1183–1187 (1999).

    ADS  CAS  Article  Google Scholar 

  8. 8

    Jaeckel, E., Lipes, M. A. & von Boehmer, H. Nature Immunol. 5, 1028–1035 (2004). Errata: 5, 1190 (2004); 6, 219 (2005).

    CAS  Article  Google Scholar 

  9. 9

    Jaeckel, E., Klein, L., Martin-Orozco, N. & von Boehmer, H. J. Exp. Med. 197, 1635–1644 (2003).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  10. 10

    Arif, S. et al. J. Clin. Invest. 113, 451–463 (2004).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  11. 11

    Homann, D. et al. Immunity 11, 463–472 (1999).

    CAS  Article  PubMed  Google Scholar 

  12. 12

    von Herrath, M. G. & Harrison, L. C. Nature Rev. Immunol. 3, 223–232 (2003).

    CAS  Article  Google Scholar 

  13. 13

    The Diabetes Prevention Trial — Type I Study Group Diabetes Care doi:28.05.05.dc04–1693 (2005).

Download references

Author information



Rights and permissions

Reprints and Permissions

About this article

Cite this article

von Herrath, M. Insulin trigger for diabetes. Nature 435, 151–152 (2005).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.