Combinatorial detection of autoreactive CD8+ T cells with HLA-A2 multimers: a multi-centre study by the Immunology of Diabetes Society T Cell Workshop

Abstract

Aims/hypothesis: Validated biomarkers are needed to monitor the effects of immune intervention in individuals with type 1 diabetes. Despite their importance, few options exist for monitoring antigen-specific T cells. Previous reports described a combinatorial approach that enables the simultaneous detection and quantification of multiple islet-specific CD8+ T cell populations. Here, we set out to evaluate the performance of a combinatorial HLA-A2 multimer assay in a multi-centre setting. Methods: The combinatorial HLA-A2 multimer assay was applied in five participating centres using centralised reagents and blinded replicate samples. In preliminary experiments, samples from healthy donors were analysed using recall antigen multimers. In subsequent experiments, samples from healthy donors and individuals with type 1 diabetes were analysed using beta cell antigen and recall antigen multimers. Results: The combinatorial assay was successfully implemented in each participating centre, with CVs between replicate samples that indicated good reproducibility for viral epitopes (mean %CV = 33.8). For beta cell epitopes, the assay was very effective in a single-centre setting (mean %CV = 18.4), but showed sixfold greater variability across multi-centre replicates (mean %CV = 119). In general, beta cell antigen-specific CD8+ T cells were detected more commonly in individuals with type 1 diabetes than in healthy donors. Furthermore, CD8+ T cells recognising HLA-A2-restricted insulin and glutamate decarboxylase epitopes were found to occur at higher frequencies in individuals with type 1 diabetes than in healthy donors. Conclusions/interpretation: Our results suggest that, although combinatorial multimer assays are challenging, they can be implemented in multiple laboratories, providing relevant T cell frequency measurements. Assay reproducibility was notably higher in the single-centre setting, suggesting that biomarker analysis of clinical trial samples would be most successful when assays are performed in a single laboratory. Technical improvements, including further standardisation of cytometry platforms, will likely be necessary to reduce assay variability in the multi-centre setting.