GAD65 autoantibodies and its role as biomarker of Type 1 diabetes and Latent Autoimmune Diabetes in Adults (LADA)

Summary

One of the hallmarks of autoimmune diabetes is the presence of adaptive responses directed to neuroendocrine proteins. One of these proteins is glutamic acid decarboxylase (GAD). While GAD is widely distributed in neuroendocrine tissues, its specific significance in diabetes has paralleled the advances in understanding humoral and cellular immunity in Type 1 diabetes (T1D) and in a subset of Type 2 diabetes (T2D), going from the seminal discoveries of islet autoantibodies to the development and standardization of bioassays as diagnostic tools, to studies on the structure of GAD and its antigenic determinants. GAD65 autoantibodies can accurately predict T1D development in combination with other surrogate humoral biomarkers and they are considered the most sensitive and specific biomarker which identifies a subset of clinically diagnosed T2D termed Latent Autoimmune Diabetes in Adults (LADA). We and others provided evidence indicating that GAD65 autoantibody detection should be part of the diagnostic assessment for clinically diagnosed T2DM mainly because it predicts the rate of progression to insulin requirement in patients affected by LADA. More recently GAD has been used as a “tolerogenic vaccine” to preserve beta cell function in autoimmune diabetes. While the results of Phase III clinical trials did not substantiate the earlier promise of Phase I and II trials, there are still many unanswered questions and approaches that need to be investigated in the applications of GAD in the therapy of T1D and LADA.

A. Introduction

The plasticity of the interest in glutamic acid decarboxylase (GAD) across time is only matched by its relevance in human health and by the interest it continues to elicit as a tool to predict or possibly treat diabetes. It is difficult to conceive of another compound that has such critical biochemical relevance in a major pathway of neural physiology, whose genes are necessary for normal development, whose tissue distribution encompasses regulation of major parts of physiology in nerve function and metabolic regulation and also whose role as a self-antigen in T1D allows for the detection of at-risk populations and for the development of potential therapeutic approaches.

1. What is GAD?

In basic biochemical terms, L-Glutamic acid decarboxylase is the major enzyme in the synthesis of g-amino butyric acid (GABA) which is a potent inhibitory neurotransmitter and a critical component of neurophysiologic function. It requires a co-factor, pyridoxal 5’-phosphate (i.e. (PLP or activated vitamin B6] to catalyze this reaction. GAD and GABA are mainly present in ‘GABA-ergic’ nerve cells but, interestingly, GAD/GABA are also detected in certain non-neural cells and organs such as the pancreas [1–3], where GABA is stored in synaptic-like vesicles in islet beta cells [4]. Although its functional relevance is presently unclear, it may be related to paracrine effects in the modulation of glucagon secretion in alpha cells [5].

It is noteworthy that the term GAD basically denotes an enzymatic activity that is part of a critical synthetic pathway. In terms of gene or protein structure and function, GAD involves two major protein isoforms that catalyze GABA synthesis. One isoform has a molecular size of 65kDa and is termed GAD65, while the second one, of 67kDa size, is termed GAD67. These proteins are also the product of separate genes [6] which also reside in different chromosomes. GAD65 is the product of a gene located on Chromosome 10 [7] and GAD67 on Chromosome 2 [8]. However, in spite of these differences, the two isoforms share 65% homology in their amino acid sequence, with GAD65 containing 585 aminoacids [2;3]. The main sequence variation occurs in the N-terminal region, which also appears to account for their intracellular distribution [9]. In the pancreas, GAD65 is mostly detected in synaptic-like vesicles (SNLVs) while GAD67 exhibits a cytosolic localization in the beta cells. Interestingly, both of these localizations are separate from the large dense-core insulin secretory granules and other T1D islet cell autoantigens [2;4]. Their synthetic path in neural tissue and β-cells shows similarities; both GAD65 and GAD67 are initially synthesized in the cytosol rather than the ER and importantly, GAD proteins exhibit different translational regulation than proinsulin and other insulin secretory granule autoantigens [10–12]. In a thought-provoking report, relevant to the hyperglycemic conditions of diabetes, medium to longer-term exposure of isolated islets to glucose, can transcriptionally increase specifically GAD65 synthesis above the overall level of protein synthesis in the cells [13]. However, there are still areas in need of elucidation in terms of the physiological effects of GAD65. A key example is that in spite of the importance of the tissue localizations of GAD and the relevance of the synthetic path it modulates, research in animal models has yielded puzzling results such as with the null GAD mouse model, in which islet cell function does not appear to be impaired [2;3].

2. Historical Perspective in Diabetes

The earlier indications of the relevance of GAD as an important biomarker for type 1 diabetes (T1D) arose from immunoprecipitation studies [14] involving incubation of rat islets with radiolabeled methionine, followed by exposure of solubilized membranes to sera from newly diagnosed T1D patients or controls. The results of these experiments showed that diabetic sera precipitated a ligand with a molecular weight of 64 kDa. Moreover, as a critical indication of the potential of GAD as a predictor of risk for T1D, the immunoreactivity to the 64 kDa antigen was observed both in about 80% of new onset T1D patients and in also in pre-diabetic subjects. The nature of the 64 kDa antigen was later elucidated by the work of Solimena et al, showing autoantibodies to GABA-ergic neurons and pancreatic beta cells in a rare condition termed stiff man syndrome [15]. This finding facilitated the identification of GAD as the 64kDa autoantigen in T1D by Baekkeskov et al [16]. Subsequent research on GAD65 humoral autoimmunity in diabetes has led to the development of assays using reticulocyte transcribed/translated GAD65 cDNA to incorporate radiolabeled methionine, which is followed by exposure to sera, and the subsequent immunoprecipitation of the labeled antigen, to determine whether the sera contains anti-GAD65 specific autoantibodies [17]. This is now the standard procedure used throughout the world. The ability to assess humoral anti-GAD65 autoimmunity and the relative simplicity of the assay, have established GAD65 as the primary diagnostic end point to assess T1D-associated autoimmunity and T1D risk in a wide diversity of research goals and conditions, such as in different age-defined populations [18], in diverse ethnic groups [19], in the characterization of novel subpopulations of diabetes [20–23] and in the development of potential immunotherapies to prevent or treat T1D to name but a few of the more relevant instances. As illustrated in Fig 1 the story of GAD and its applications in diabetes is a textbook example of evolution and progress in research, with landmark early studies, an abundant contemporary body of research and still, an exciting future with significant potential for future avenues of basic and clinical studies.

Figure 1.

Highlights of GAD65 in diabetes, past, Present and an exciting future.

B. Antigenicity of GAD

In T1D autoimmunity, only GAD65 is antigenic despite the higher expression of GAD67 in islet beta cells [2;3;24]. Interestingly, in stiff man syndrome, the other well documented case of GAD antigenicity, GAD65 is also the major autoantigen [15]. The differences in the antigenicity of the GAD isoforms merits scrutiny, particularly in view of the potential of GAD 65 to be used in immunotherapies. Both isoforms have similar activities in catalyzing GABA synthesis but exhibit clear differences in parts of their aminoacid sequences and in their deduced tertiary conformation. GAD65 is assumed to be more flexible in the C-terminal region [24], which is a potentially consequential feature since these unique mobility regions in the GAD65 molecule coincide with the region that exhibits key antigenic epitopes [24]. This however, does not mean that the N terminus of the molecule is not relevant. The N terminal region is the site of palmitoylation in GAD65, which is a feature not found in other T1D islet antigens and which in turn, appears to determine the subcellular localization of this GAD isoform, as was highlighted by the report of Solimena et al. [25] showing that substitution of the first 29 amino acids of GAD 67 with the first 27 amino acid residues of GAD 65 re-directs the normally cytosolic 67 kDa isoform to the Golgi, which is the usual localization of GAD65; thus indicating that the first 27 amino acids of GAD65 contain an intracellular localization signal. Other studies have also shown that post-translational modifications of the molecule in this region are highly relevant to its function. GAD65 is palmitoylated on two cysteine residues in the N-terminal region and that feature determines the subcellular orientation of the GAD isoform [26]. While GAD67 remains persistently cytosolic, palmitoylated GAD65 is incorporated into SNLV membranes, which results in the outward orientation of SNLV-associated GAD65 facing the cytosol. It is in this position that GAD65 forms a small protein complex containing the vesicular GABA transporter (VGAT). Meaningfully, this allows the local production of GABA by GAD65 and the rapid transport of GABA by VGAT into the SNLV storage compartment in the beta cell [2].

Other relevant research has postulated that the NH2-terminal domain of GAD65 (mainly expressed in pancreatic islets) involves a two-step modification, which leads to membrane anchoring involving the post-translational hydroxylamine sensitive palmitoylation [26]. The anchored GAD65 can be released from the membrane through an apparent enzyme activity in islets, which indicates that the membrane anchoring and release steps are subject to regulation. The hydrophobic modifications and subsequent membrane anchoring and orientation of GAD65 to microvesicles may be significant for its role as an islet autoantigen.

It is noteworthy that human vascular endothelial cells (Ecs) can process and present immune-relevant GAD65 epitopes to autoreactive T-cells [27]. The in vitro transmigration by autoreactive T-cells across an EC monolayer is stimulated by presentation of cognate peptide/HLA complexes on the EC surface and is also LFA-1 (Lymphocyte function-associated antigen 1) dependent. In consequence, it is possible that presentation of GAD65, by islet endothelium in vivo could promote transmigration of circulating islet autoantigen-specific T-cells that have been primed in regional lymph nodes against the pertinent GAD 65 epitopes.

There is also a well documented body of research showing that the early humoral autoimmunity in T1D is directed against epitopes primarily in the middle region of the molecule and but also include the C-terminus [28;29]. Additionally, recent thought-provoking structural crystallography studies aided by monoclonal antibody testing of antigenic determinants [30] have indicated that the more flexible C-terminal region of GAD65 exhibits a grouping of autoantibody and T cell antigenic determinants.

Moreover, these studies indicated that there is close proximity of the B- and T-cell-related epitopes in the GAD65 molecule, with the important implication that that the antigen–antibody complexes could contribute to GAD65-induced T-cell reactivity. While the importance of the middle region and the C-terminus as the major sites of immunoreactive epitopes have become well-established. Immunoreactive epitope localization is frequently not stable over time and can spread later to epitopes in the N-terminal region of GAD65 [31;32], although not all studies have substantiated time-related intramolecular epitope spreading [33;34]. The dynamic nature of GAD65 humoral autoimmunity in T1D is likely to continue to grow as an area of research in order to develop more precise and effective diagnostic assays and immunotherapeutic approaches.

C. Clinical Relevance for T1D Diagnosis

1. Estimation of Risk

Autoantibodies against GAD are a valuable predictor of risk and progression to overt autoimmune diabetes. The combined positivity to GAD65 and other islet autoantigens in the determination of T1D autoantibodies has a reliable and accurate predictive application [35–37]. Interestingly, the intramolecular increase in antigenic determinants described for GAD in the previous section, also has an intermolecular counterpart in the spreading of islet autoantigenic proteins during development of autoimmune diabetes, where the number of autoantibodies against different islet autoantigens generally increases sequentially [38]. While anti-insulin autoantibodies have been frequently reported among the first ones to appear in both animal models of type 1 diabetes and humans, and appear to be the more relevant single autoantibody in the prediction of risk and progression to clinical type 1 diabetes [39], it is also true that anti-GAD autoantibodies are frequently found at early preclinical stages [40] this early presence and the relative ease of assaying for anti-GAD autoantibodies have made them the most commonly used first screen to assess risk or progression to the insulin-requiring stage of the disease. Additionally, an increase in intermolecular islet antigen spreading has enormous predictive value, as the number of autoantibodies against diverse islet autoantigens has been shown to correlate with risk and rate of progression to overt diabetes [41–43].

Therefore, in the body of knowledge that has developed from the original seminal studies on single and multiple humoral autoimmunity and prediction of T1D risk, a consensus has emerged to the effect that the detection of a single autoantibody against islet proteins has a moderate predictive power. However, screening for a single autoantibody such as GAD65 autoantibodies provides a valuable warning sign should the screen show positivity and this single positivity is an indication that additional islet autoantibody assays are indicated in order to accurately assess risk and predict onset of T1D. The pertinent literature is plentiful and a variety of prediction models have been developed. Some models, for instance, show that triple autoantibody positivity against different islet proteins implies a risk of developing diabetes that ranges between 48 to 86% over five years and 64 to 86% over 10 years [35;43], while other models have estimated that the risk of T1D onset, with triple autoantibody positivity approaches 100% after 5 years [41].

Specific strategies to assay for islet autoantibodies have become part of clinical practice and depend on both the predictive value and relative methodological accessibility of the assays. However, aside from this trilogy of assays as determinants of risk, the need to use easier-to-implement and easier-to-monitor assays has led to the standardization of T1D autoantibody assays in current clinical practice. This has resulted in an alternative approach in which the most used strategy to screen at-risk populations for T1DM is to assay for autoantibodies against the islet antigens insulin, GAD65 and IA-2 [35]. More recently, antibody assays against the zinc transporter ZnT8 have been considered as an added screen [44]. In either case, it is evident that GAD is a “common denominator” assay in either strategy for multiple autoantibody testing in order to estimate risk and progression rate to the insulin-requiring stage in T1D. As a matter of fact, GAD autoantibody determinations constitute the first assay in the examination of autoantibody reactivity in T1D, in usual clinical health-care practice.

The need to coordinate methodological protocols has resulted in standardization strategies such as the adoption of a serum reference standard for GAD and IA-2 antibodies by the World Health Organization (WHO) and importantly, the creation in 2000, of the Diabetes Antibody Standardization Program (DASP) by a combined effort of the Immunology of Diabetes Society (IDS) and the Centers for Disease control and Prevention (CDC). DASP involves an international network or consortium of 47 key laboratories in 18 countries, which provide technical support and training, proficiency testing and supports development of methodologies, reference methods and materials. DASP activities also have led to the current recognition of four major islet autoantigens, namely GAD, Insulin, IA-2, and ZnT8 in the clinical laboratory assessment of humoral T1D-associated autoimmunity, which again underscores the continuing relevance of GAD assays as a tool in the estimation of diabetes risk. The need to provide a basis for comparison of results has also continued to evolve into the harmonization efforts to produce unified protocols that can take into account differences in use of standards such as the WHO standard relative to the use of common standards used by the National Institute of diabetes and Digestive and Kidney diseases (NIDDK) consortia laboratories [45].

While GAD autoantibodies are a first-line tool in assessment of T1D-risk and the combined positivity of GAD and other islet autoantibodies yields a highly accurate predictor of risk, the determination of GAD autoantibodies has also unique features in clinical application, due to its age-related incidence in susceptible populations. It is well established that other islet autoantibodies such as insulin autoantibodies and phogrin are highly accurate T1D risk predictors in childhood [46] and the age of detection of insulin autoantibodies in particular, exhibits close associations with age of T1D diagnosis in children [39]. In contrast, the presence of autoantibodies against GAD65 appears to be related to adult ages and also it has been reported that high GAD65 antibody titers exhibit a correlation with longer-term diabetic complications such as retinopathy, 15 years after T1D onset [47].

2. Therapeutic Applications

The maintenance of production of self-antigens by the thymus and similarly specialized immune cells is a major factor in the maintenance of immune tolerance to these self antigens [48;49]. More specifically and in regard to the topic of this review, it has been shown that the administration of GAD65 to non-obese diabetic (NOD) mice can prevent the autoimmune loss of pancreatic beta cells, possibly by maintenance of exposure to the immune system, thereby promoting the recognition of “self” and immune tolerance. As a result of this body of investigations, a great deal of interest has motivated the examination of the therapeutic value of administering GAD in humans. As a result, clinical studies have been conducted to assess if the administration of GAD65, as a “tolerogenic vaccine” in recently diagnosed T1D patients can prevent or ameliorate the loss of functional beta cells. This interest in turn led to the implementation of Phase I and II clinical trials in which recent onset T1D patients were administered subcutaneously GAD65 formulated with aluminum hydroxide (GAD-alum). In the Phase I and Phase II clinical studies, the results showed promising effects of the “GAD-alum” vaccine, in the prevention of decline of insulin secretion without adverse effects. Additionally, the GAD specificity of the response was substantiated by endpoints indicating that there was GAD induced humoral and cellular autoimmunity [23]. Subsequently, Phase III clinical trials were conducted to further assess the efficacy and dose-responsiveness of the GAD-alum treatment protocol in preserving beta cell function and the results have been recently made public. Unfortunately, the analysis of the endpoints related to insulin secretion and preservation of beta cell function, as shown in Fig 2, did not substantiate the earlier positive findings and led to the conclusion that the translation of the approach used successfully in animal models presents challenges in obtaining similar effectiveness in human patients for applications in clinical practice [50].

Figure 2.

Proportion of patients with 2-h peak C-peptide of 0·2 nmol/L or higher Data are adjusted for age, sex, and baseline value of C-peptide. GAD=glutamic acid decarboxylase. (Reproduced with permission from Wherrett DK et al and the Type 1 Diabetes TrialNet GAD Study Group. Lancet (2001) 378:319–327)

D. Role in LADA

The incidence of both 1 and 2 diabetes has continued to expand world-wide and the understanding of the role of autoimmunity in diabetes has evolved along with the detection of new populations and subgroups. While the presence of autoantibodies against islet proteins continues to be the hallmark of “classical” T1D, a number of studies have shown that diabetes-related autoantibodies can be detected in 2–12% of patients with types of diabetes, such as type 2 diabetes, that have not usually been associated with autoimmunity in their pathoetiology [51]. The presence of T1D autoantibodies has led to the identification of unique subpopulations of diabetes that exhibit overlap or similarities with both type 1 and type 2 diabetes. As the detection and biochemical characterization of these novel subgroups presents conceptual, epidemiological and clinical challenges, there is a need to assess the value of specific markers that permit the accurate detection and treatment of these sui generis cohorts. A growing body of research has documented the case of patients who are generally adults, non-obese, who otherwise present a type-2 diabetes phenotype and who puzzlingly also have circulating anti-islet autoantibodies. These characteristics have led to the term Latent Autoimmune Diabetes of Adults (LADA), sometimes are also termed type1.5 diabetes [51]. Currently, the diagnosis of LADA relies primarily on the detection of autoantibodies against GAD65 in the serum of clinically diagnosed T2DM patients. In this regard, GAD65 testing provides the critical first identifier to detect this unique diabetic disease phenotype that in addition of the presence of these GAD65 autontibodies includes impaired insulin secretion and in a significant number of cases, treatment with insulin as part of the therapy [52;53]. Additionally, it has also been demonstrated that, as is the case with “canonical” T1D patients, LADA patients also possess T-cells reactive to islet proteins, which is not the case in “classic” T2D without humoral islet autoimmunity. While the specific characteristics of LADA are the subject of ongoing examination and the possibility that both classical type 2 diabetes and autoimmune diabetes can be present in the same patient cannot be currently conclusively discounted, the fact remains that LADA remains the consensus term to describe older populations that exhibit GAD humoral autoimmunity.

The detection of GAD65, as well as IA-2 autoantibodies in diagnosed T2D patients by conventional criteria (ADA or WHO) commonly ranges between 5–10% and may be higher if insulin is used as a therapy. The epidemiological consequences of this demographic assessment are many. First, there may be similar numbers of GAD65 positive older diabetics with GAD65 autoantibodies as there are children with “canonical” T1D. Second, there is a need to investigate if more thoroughly characterized assays for relevant reactive epitopes of antigens involved in islet autoimmunity, might detect a larger proportion of T2DM patients with islet autoimmunity. Therefore, testing for GAD65 autoantibodies and the use of biochemically defined antigens in the assays might permit the identification of immunodominant LADA-associated epitopes. Hopefully, the relative ease of GADA assays might make the use of these assays part of the clinical characterization of diagnosed T2D patients, as it might aid in predicting the rate of progression to insulin requirement in these generally adult populations. GAD65 antibody-positive patients may be suitable for early insulin therapy or tighter and more closely monitored identifiers of glycemic control. In this regard, GAD65 continues to be a unique primary screen in this group to clearly distinguish and suitably treat GAD65 antibody-positive older diabetic patients.

The application of GAD autoantibody testing in LADA and LADA-related groups has continued to grow in interest and scope. For instance, studies have shown that in adult populations, the prevalence of GAD65 humoral autoimmunity and its correlation with beta cell function exhibits specific associations with racial and ethnic characteristics of the study population, as illustrated in Fig 3 and is also a relevant biomarker in elderly patients in the assessment of cardiovascular status and T1D autoimmunity [18]. The relationship between epitope immunoreactivity and LADA features is an important area of investigation that has been shown in studies covering LADA-associated subpopulations. There appears to be a correlation between severity of development to insulin requirement and the presence of autoantibodies against both the middle region and the C-terminus in new onset T2D patients [54]. Additionally, more recent observations suggest that some of the more widely known concepts about development of GAD epitope autoimmunity may need to be further elucidated and perhaps reassessed. While epitope spreading is a well accepted and documented phenomenon overall in autoimmune diabetes, the specific role of reactive epitopes in autoimmunity in LADA is likely to become an even more active area of investigation in view of a diversity of findings. A recent LADA-related comprehensive study in the UK termed United Kingdom Prospective Diabetes Study (UKPDS) [55] included the assessment of GAD65 epitopes in 242 patients who had tested positive at the time of diagnosis, and whose epitope immunoreactivity was assessed over a 6 year period, which appears to be in contrast to the concept of epitope spreading over time. The results from the study indicated that GAD65 autoantibodies were directed mostly (in over 70% of the subjects) against middle region or C-terminal epitopes but the immunoreactivity of the epitopes did not exhibit changes over the 6 year period. Other recent LADA studies have also shown novel differences GAD65 epitope associations that relate to ethnicity. In a report by Jin P et al [56] with Chinese populations, the major epitopes of humoral autoimmune reactivity were located in the middle region and in the C-terminus of GAD65, in line to the current consensus. However, another LADA-related study with adult (mean age 48y) Japanese patients exhibiting slowly progressive insulin dependent diabetes mellitus [57]; describes a unique epitope in the N-terminus of GAD65, which was found to be prevalent in this subgroup and also shown to be inversely related to length of time before insulin is required therapeutically, suggesting the relevance of the N-terminus of GAD65 in LADA, in specific cohorts. GAD65 epitope specificity in LADA populations may also show associations with co-morbidities such as thyroid autoimmune disease (AITD). In the study by Jin P and coworkers [58] LADA Chinese patients exhibiting LADA and AITD also exhibited a higher frequency of antibodies against both the middle region and C-terminal regions of GAD65, compared to patients without thyroid autoimmunity, suggesting that epitope-specific detection of autoantibodies may aid in the prediction of thyroid autoimmune co-morbidity in LADA. In addition to the higher likelihood of thyroid autoimmunity, it was also found that LADA patients with antibodies against both middle and C-terminus reactivity had lower levels of C-peptide and exhibited a requirement for insulin treatment. Although prima facie, the main difference in the clinical presentation between the studies is the presence of autoimmune thyroid disease in the Chinese report, the data must be interpreted with caution as they cannot firmly be construed to represent overall specific epitope reactivities in the corresponding cohorts. Additionally, the described genetic associations, such as HLA haplotypes related to diabetes risk, provided interesting albeit limited insight in these studies. The Japanese study, describing GAD N-terminus reactivity, indicates that in their LADA-related group, the HLA DQA1*0303DQB1*0401 haplotype was more common than in control subjects, while in the related Chinese study [56], the differences between the immunoreactive GAD65 C- and M-terminus epitopes among DQA1*-DQB1* haplotypes in LADA patients were largely inconclusive, possibly because of insufficient population sample sizes. In consequence, although these reports provide interesting insights, it is important not to over-interpret the role of HLA genotypes in assessing their effect on humoral autoimmunity in limited studies. These considerations await much larger studies and the necessary confirmations in additional publications. Nonetheless, they also highlight the need to further the depth and breadth of research on two key areas, namely GAD epitope immune-recognition as it relates to co-morbidities that show associations with type 2 diabetes such as thyroid autoimmunity and also, the specific roles of genetic and immune interactions between ethnic and racial subgroups in regions that are increasingly the focus of diabetes research.

Figure 3.

Prevalence of GAD65 AAs by C-peptide levels (pmol/ml) and race/ethnicity among all participants with diabetes, NHANES III (n = 1,059). *Overall race/ethnicity-specific comparisons (Reproduced from Barinas-Mitchell E. et al. Diabetes (2004) 53:1293–1302).

E. Concluding Remarks

Although the expression of GAD65 is not exclusive to the beta cell of the pancreas, the relevance of GAD65 as a specific biomarker in diabetes risk prediction, as well as a potential avenue for immunoregulatory therapies has continued to expand as an area of investigation. The application of assays do detect autoantibodies against GAD65 continues to be a primary screen for the detection of autoimmune diabetes. The knowledge gained from our understanding of the structural features of the GAD65 molecule and their implications in the development of immunoreactive autoantigenic epitopes, has provided useful tools to better identify risk and prediction to insulin-requiring stages in susceptible populations, not only in “classic” Type 1 Diabetes but also in more recently characterized forms of the disease, as is the case of LADA. These areas of research are likely to remain the focus of a continuing large body of research for the foreseeable future.

The use of GAD65 as an immunomodulatory therapy also merits discussion. The current evidence describes promising effects on the preservation of beta cell function with the use of GAD65 as a “tolerogenic” vaccine, at least in Phase I and II studies. While a large and recent Phase III clinical trial did not substantiate the earlier and promising findings, the fact remains that there still remain questions and potential alternative avenues of therapy with GAD65. These include different presentations of the GAD65 molecule complexes as well as alternative frequencies, dosages and routes of administration of the GAD65 “vaccine”. These considerations need to be closely scrutinized and deserve further research. Even if the larger Phase III trial provided disappointing results, the overall body of data is by no means discouraging and deserves further efforts. Future diabetes-related research on GAD65 is likely to continue to broaden and deepen the knowledge and potential clinical applications of this unique versatile molecule.