Islet Autoreactive CD8 T-cells in Type 1 Diabetes: Licensed to Kill?

While differentiating T-cell autoreactivity against pancreatic islets between type 1 diabetic patients and nondiabetic control subjects has, over time, proven to be a difficult task, the ultimate challenge for cellular markers of this disease will involve the definition of immunological measures associated with disease progression and intervention. With the notion that type 1 diabetes is a T-cell–mediated autoimmune disease, it is therefore logical for investigators to direct research efforts toward studies of T-cells. However, relatively few studies associating T-cell autoreactivity with disease progression have been attempted (i.e., in comparison with efforts monitoring autoantibodies), and those that have been performed are largely limited to CD4 T-cells (1).

Ever since the demonstration of their presence in insulitic lesions in the pancreas of patients at clinical onset of type 1 diabetes, CD8-positive cytotoxic T-cells have been implicated in the disease process (2). This association and its potential importance for disease was corroborated by the genetic predisposition of HLA class I loci for type 1 diabetes that is independent of linkage disequilibrium with HLA class II (3); a setting that implied the role of HLA class I with type 1 diabetes is more than guilt by genetic association alone (4). Theoretically, HLA class I–restricted CD8 T-cells could be the sole candidate effector pathway for islet autoantigen specific β-cell destruction since autoreactive cytotoxic T-cells recognize peptide epitopes displayed on the β-cell surface in the context of HLA class I molecules. In recent years, a surge of studies has identified and validated such epitopes, derived primarily from β-cell proteins (5–10). The consistencies among these studies have been remarkable as well as encouraging (11).

However, it is still unresolved as to what extent circulating CD8 T-cells may act as correlates of disease progression. One pertinent study addressing this question revealed a strong correlation between increases of autoreactive CD8 T-cells and recurrent autoimmunity, yet this was following islet transplantation and graft failure in type 1 diabetic patients (6). In this issue of Diabetes, Martinuzzi et al. (12) addressed this knowledge void and observed a change in CD8 autoreactivity to islet epitopes in the opposite direction; namely, they measured a drop in numbers of islet epitope–specific CD8 T-cells in the first year following disease diagnosis. If confirmed by subsequent studies, this could imply that measuring CD8 T-cell autoreactivity in peripheral blood mirrors insulitis or changes in β-cell mass. This would have important ramifications for studies of T-cells as markers of autoimmunity.

However, it is also important to note a number of items. First, the authors did not compare changes in islet autoreactivity with β-cell function. In addition, the authors used an ELISPOT assay that they previously reported to detect elevated CD8 T-cell autoreactivity to islets (8). Even though their assay suffers from large variations in background cytokine production that may affect their interpretation and impair enumerating changes in response rates, it is clear that the reactivity they detect is associated with disease activity. Yet, cytokine production does not equal cytolytic activity in that the number of cytokine spots may underrepresent the precursor frequency of islet specific CD8 T-cells. Indeed, even with clonal populations of CD8 T-cells, <10% of cells elicit cytokine production in response to their cognate antigen (13). HLA tetramer studies may offer added value in this context. Another issue involved selection of β-cell epitopes (14). While the authors tested a series of validated epitopes, these were selected on their presumed high binding affinity to HLA-A2. Perhaps surprisingly, Ouyang et al. (9) recently observed an inverse correlation between the binding affinity of β-cell peptides to HLA-A2 and CD8 T-cell responses against those peptides in recent-onset type 1 diabetic patients. Ascertainment by high binding affinity, and in particular on the basis of matrix-assisted algorithms such as SYFPEITHI, may be flawed. Indeed, “immunosophically” it would make sense that T-cells recognizing peptides with intermediate or low binding affinity may escape central tolerance (i.e., clonal deletion) during thymic education and therefore turn pathogenic upon peripheral stimulation. With the unambiguous identification of several HLA class I alleles predisposing to type 1 diabetes, the next generation of epitopes may be derived from peptides binding to these HLA variants (3). The new candidate target of β-cells, the cation efflux transporter ZnT8, deserves consideration for this purpose (15). These issues offer great opportunities for future T-cell workshops, preferably in a blinded context, of the Immunology of Diabetes Society, coordinated by a new generation of researchers.

Although an increasing number of studies, including that of Martinuzzi et al., have demonstrated the presence of circulating islet autoreactive CD8 T-cells around the clinical onset of disease, clinical evidence for a role for these T-cells in β-cell destruction is still lacking. The new report corroborates that CD8 T-cell populations may decrease in periphery with β-cell mass waning, and this observation supports a direct correlation with β-cell destruction active insulitis. This study also supports the use of peripheral blood as source of biomarkers of disease activity (Fig. 1). Would CD8 T-cells be suitable targets for immune intervention therapy in type 1 diabetes? Immunotherapy with modified monoclonal antibodies against the common T-cell marker CD3 resulted in the induction of CD8 lymphocytosis and inversion of the CD4-to-CD8 ratio that was perhaps partly reflecting reactivation of Epstein-Barr virus–specific CD8 T-cells (16). Interestingly, the anti-CD3 mAb hOKT3γ1(Ala-Ala) caused activation of CD8+ T-cells and induced regulatory CD8+CD25+ T-cells expressing CTLA4 and Foxp3 (17,18). These cells inhibited the responses of CD4+ cells to the mAb itself and to the antigen. Foxp3 was also induced on CD8+ T-cells in patients during mAb treatment, which suggests a potential mechanism of the anti-CD3 mAb immune modulatory effects involving induction of a subset of regulatory CD8+ T-cells. It would therefore be important to consider and appraise pathogenic as well as regulatory features of CD8 T-cells in type 1 diabetes.

With the identification of circulating immune correlates of disease activity, the opportunity arises to use such validated markers for monitoring progression of type 1 diabetes in the context of immune intervention trials in order to assess the immunological, as well as clinical efficacy, of candidate therapies in type 1 diabetes.