IA-2 antibodies

Islet antigen-2 (IA-2), previously known also as ICA-512, is a major target of islet cell autoantibodies. The protein is found in neural tissue and cells of the pancreatic islets, and its gene has been localized to chromosome 2q35. The cDNA encodes a 979 amino acid transmembrane protein which is enzymatically inactive, and a related PTP-like molecule termed IA-2β or phogrin is also a major islet autoantigen whose location and intracellular domain are 74% identical to IA-2. Autoantibodies to IA-2 are present in up to 80% of children and adolescents at diagnosis of type 1 diabetes. Testing for IA-2A complements GADA measurement, since more than 90% of children have antibodies to at least one of these proteins at diabetes onset. IA-2A generally develops later in the process leading to type 1 diabetes and is therefore associated with more rapid progression. These antibodies do not persist as long following diagnosis as GADA, and are less common in patients who are diagnosed with type 1 diabetes over the age of 30 years. They are therefore less useful than GADA for characterising diabetes in longer term or older patients.

The discovery of IA-2

IA-2 was independently but simultaneously isolated by several groups: first, by screening a human islet expression library (where it was termed ICA512) and second, by cloning and screening a human insulinoma subtraction library where it was identified as the insulinoma-associated protein-2 (IA-2), both being virtually identical except for the lack of 388 amino acids at the N-terminus and 65 at the C-terminus[1][2].

This antigen appeared to be a protein tyrosine phosphatase (PTP)-like molecule that lacked enzymatic activity due to several substitutions in the highly conserved sites and is expressed primarily in neuroendocrine cells such as the central nervous system as well as in the islets of Langerhans. In both cell types, it is anchored in the membrane of dense core (insulin) secretory granules and antibodies are predominantly directed against the intracellular domain (AA 601-979) which protrudes into the cytoplasm and is only exposed in case of cell damage.

Its function remains unclear, but in vitro studies in mouse models have provided evidence for an enhancing role of IA-2 in insulin secretion. Islet cells also express differentially spliced mRNA for ICA512, with one form lacking the trans- and juxtamembrane region (AA 557-629) of the molecule. Together with standard recombinant IA-2ic, this aberrant antigen - the ICA512bdc construct - was also used to develop antibody assays in the prediction of type 1 diabetes [3].

Soon after, several investigators discovered a related PTP-like molecule termed IA-2β [4] or phogrin. This is similar in many respects to IA-2, especially in its location and intracellular domain which is 74% identical. This protein is also enzymatically inactive.

Development of the IA-2 assay

The importance of autoantibodies to this family of molecules for prediction of the development of diabetes was clearly demonstrated before the sequences of IA-2 and IA-2β were defined. Studies in the early 1990s showed that immunoprecipitation of insulinoma cell lysates with type 1 diabetes sera yielded a 64kDa protein which upon treatment with trypsin resulted in 37 and 40kDa fragments [5].

Although the identity of these fragments was at first obscure, additional studies revealed that IA-2 is the precursor of the 40kDa and IA-2β of the 37kDa molecule. Antibodies have now been found in the majority of newly diagnosed type 1 diabetic patients as compared to <2% of normal controls. The assay that so far has given the best results in terms of specificity, sensitivity and reproducibility, is the immunoprecipitation (using Protein A Sepharose) of radiolabelled (35S) recombinant IA-2ic prepared in an in vitro transcription-translation system.

Several international workshops, such as the Diabetes Autoantibody Standardization Program (DASP), have been conducted to evaluate and standardize different assays worldwide. In this program, IA-2A and IA-2β assays were found to have a high sensitivity (resp. 66% and 53%) and specificity (99% and 98%)[6]. In contrast, ELISA performed less well. It is hypothesized that during this test, the conformational-dependent epitopes required for antibody binding are destroyed, and lose their antigenic structure by reduction in disulphide bonds.

It was initially shown that antibodies in type 1 sera reacted exclusively with the intracellular - and unexpectedly not with the extracellular - domain of IA-2 and IA-2β. Within the IA-2ic as well as the IA-2βic, the major antigenic determinants reside at the C-terminus while only 40% of sera reacts with the amino-terminus in the case of IA-2. Some patients also demonstrate antibodies to the juxtamembrane region. More recently, IA-2A have been detected against the extracellular region, and are possibly associated with a higher risk for type 1 diabetes.

In combination with other islet autoantibodies, detection of IA-2 is widely used to facilitate classification of diabetic patients, to study its natural history and to identify and select subjects at risk of developing diabetes for participation in intervention trials.

Role in pathogenesis and classification

IA-2 could be a primary target of the immune process which is believed to destroy the insulin-secreting islet cells. Alternatively, IA-2A might develop as a consequence of this destructive process, releasing the sequestered antigen IA-2 and thus inducing the immune response.

IA-2A are almost always present in combination with other islet autoantibodies in autoimmune diabetes. The frequency of autoantibodies against IA-2 is similar in the USA and Europe, ranges between 60 to 80% in newly diagnosed type 1 diabetic patients and, like islet cell antibodies (ICA), decreases to around 45% in those diagnosed after the age of 20.

The prevalence of IA-2A varies not only with age, but also with HLA, and is greatest in patients carrying HLA DR4 and the HLA DQA10301-DQB10302 (DQ8) genotype[7][8]. There is also evidence that IA-2A positivity persists up to a few years after diagnosis but decreases rapidly thereafter, unlike GADA, which also appears to increase with age. Therefore, the latter can be used to identify autoimmune diabetes in adults masquerading as type 2 diabetes (LADA), while screening for IA-2A in older patients is less useful since the diagnostic sensitivity for LADA (2-4%) is much lower.

Factors significantly associated with the duration of clinical partial remission are age, gender, ICA positivity and a high initial HbA1C, whereas the presence of IA-2A appears to have no influence.

In the case of IA-2β, 35-50% of the newly diagnosed patients have autoantibodies as compared to <1% of normal controls. Over 95% of the diabetic sera that react with IA-2β also react with IA-2, but only 50-80% react the other way round, indicating that both molecules possess unique epitopes and that IA-2βA most likely result from subsequent epitope spreading.

IA-2 in prediction of type 1 diabetes

Attempts to predict type 1 diabetes, with the aim of preventing the disease, have focused on autoantibodies as disease markers; studying T-cell changes is technically difficult and less specific. Numerous studies have demonstrated the predictive value of ICA (9), although it involves use of human pancreatic tissue and has important techical limitations.

Testing for IA-2A complements GADA measurement, as more than 90% of children have antibodies to at least one of these proteins at diabetes onset. IA-2 and GAD make the predominant - but not exclusive - contribution to ICA-reactivity 3 [9][10]. Since they are in consequence highly prevalent in pre-diabetic people, and can be identified with sensitive recombinant assays, IA-2A and GADA are now widely used for population screening and diabetes prediction together with IAA, IA-2βA and the more recent detected ZnT8A, supplemented by metabolic testing.

Several studies on the predictive power of IA-2A have shown a high specificity, sensitivity and positive predictive value (PPV) for developing diabetes in monozygotic twins (all >90%) and in first-degree relatives where the 5-yr PPV ranged from 65 to 85%3,7. Although IA-2A has a high positive predictive value, the risk of diabetes progression increases with the number of autoantibodies present, and antibody combinations are therefore generally considered to provide the best predictive marker.

Many studies support this generally held view but a few indicate that the presence of IA-2A in particular - which is often associated with other antibodies - confers a higher risk of rapid progression toward clinical onset than multiple antibodies per se [11]. It is suggested that the production of IA-2A coincides with a critical switch in disease progression, where the intracellular domain of IA-2 may only become visible to the immune system at the outer cell surface in the case of beta cell damage or dysfunction.

These observations support the hypothesis that IA-2A are markers of active β-cell destruction and suggest that autoimmunity against IA-2 participates in accelerating β-cell destruction prior to clinical disease or has characteristics that promote the process. For that reason, IA-2A often appear relatively late in the pre-clinical phase leading to type 1 diabetes and are associated with rapid progression toward disease onset [12]. Moreover, at this stage IA-2A tend to cluster with IA-2βA and ZnT8A and confirm, alone or in combination, the association with rapid progression to diabetes, with potential increase in diagnostic sensitivity.

Further detailed investigation of the humoral response showed that antibodies in the earlier phase of the disease process are predominantly directed toward the juxtamembrane region of IA-2ic, whereas by the time of diagnosis antibodies to multiple epitopes in the PTP regions of IA-2 and IA-2β predominate, especially in younger people. Epitope spreading to both multiple epitopes and juxtamembrane region antibodies is linked with an increased risk of progression to overt diabetes.

Differences in predictive models might therefore be explained by the use of assays based on the ICA512bdc construct in which the juxtamembrane region is selectively deleted. Conversely, in progression to type 1 diabetes, data describing autoantibody isotypes (switch) remain highly controversial, although it has been suggested that diabetes risk can be further stratified if other characteristics including IgG subclass as well as antibody titer are also taken into consideration 12.


  1. ^ Rabin DU, et al. Islet cell antigen 512 is a diabetes-specific islet autoantigen related to protein tyrosine phosphatases. J Immunol 1994; 152: 3183-3188

  2. ^ Lan MS, et al. Molecular cloning and identification of a receptor-type protein tyrosine phosphatase, IA-2, from human insulinoma. DNA Cell Biol 1994; 13:505-514

  3. ^ Verge CF, et al. Prediction of type I diabetes in first-degree relatives using a combination of insulin, GAD, and ICA512bdc/IA-2 autoantibodies. Diabetes 1996; 45:926-933

  4. ^ Lu J, et al. Identification of a second transmembrane protein tyrosine phosphatase, IA-2β, as an autoantigen in insulin-dependent diabetes mellitus: precursor of the 37kDa tryptic fragment. Proc Natl Acad Sci USA 1996; 93:2307-2311

  5. ^ Christie MR et al. Antibodies to GAD and tryptic fragments of islet 64K antigen as distinct markers for development of IDDM. Studies with identical twins. Diabetes 1992; 41:782-787

  6. ^ Bingley PJ, et al. Diabetes Antibody Standardization Program: first assay proficiency evaluation. Diabetes.2003; 52:1128-1136

  7. ^ Hawa M, et al. Value of antibodies to islet protein tyrosine phosphatase-like molecule in predicting type 1 diabetes. Diabetes 1997; 48:1270-1275

  8. ^ Gorus F, et al. IA-2-autoantibodies complement GAD65-autoantibodies in new-onset IDDM patients and help predict impending diabetes in their siblings. The Belgian Diabetes Registry. Diabetologia 1997; 40:95-99

  9. ^ Gorus F, et al. IA-2-autoantibodies complement GAD65-autoantibodies in new-onset IDDM patients and help predict impending diabetes in their siblings. The Belgian Diabetes Registry. Diabetologia 1997; 40:95-99

  10. ^ Kulmala P, et al. Prediction of insulin-dependent diabetes mellitus in siblings of children with diabetes. A population-based study. The Childhood Diabetes in Finland Study Group. J Clin Invest 1998; 101:327-336

  11. ^ Decochez K, et al. IA-2 autoantibodies predict impending type I diabetes in siblings of patients. Diabetologia 2002; 45:1658-1666

  12. ^ Achenbach P, et al. Stratification of type 1 diabetes risk on the basis of islet autoantibody characteristics. Diabetes 2004; 53:384-392


Nobody has commented on this article

Commenting is only available for registered Diapedia users. Please log in or register first.