Making treatment personal: measurement of exhaustion to target treatment in autoimmunity, infection, and vaccination

Abstract

BackgroundThe clinical course of autoimmune and infectious disease varies greatly even between individuals with the same condition. An understanding of the molecular basis for this heterogeneity could lead to substantial improvements in both monitoring and treatment. Since persistent antigen is targeted in both autoimmunity and chronic infection, similar pathways could dictate the success of that response. However, whereas a successful response to pathogen results in clearance and recovery from disease, a robust response to self-antigen might drive relapsing autoimmunity. During chronic viral infection, CD8 T cells develop progressive loss of function in a process termed exhaustion, characterised by high expression of inhibitory receptors (such as programmed cell death 1 [PD1]) and low expression of memory markers (such as interleukin-7 receptor). These can serve both as biomarkers of viral progression and as targets for therapy. We sought to determine whether an analogous exhausted state exists during autoimmune responses targeting persistent self-antigen. MethodsTo identify gene expression signatures associated with clinical outcome, we used weighted gene coexpression network analysis of concurrently sampled CD4 and CD8 T-cell transcriptomes from patients with four distinct autoimmune diseases (23 with systemic lupus erythematosus [SLE], 44 with anti-neutrophil cytoplasmic antibody-associated vasculitis [AAV], and 56 with Crohn's disease or ulcerative colitis) and with chronic viral infection (42 with HIV infection). Association with clinical outcome was independently validated in 1145 samples from 504 individuals with infectious disease (HIV, hepatitis C virus, dengue), during vaccination (malaria, yellow fever, influenza), and autoimmunity (type 1 diabetes, SLE, AAV, Crohn's disease and ulcerative colitis, and idiopathic pulmonary fibrosis). We used an artificial antigen presenting cell to modify provision of co-stimulatory and inhibitory signals (as suggested by the network analysis) to primary human T cells during in-vitro differentiation. In this way we tested whether expression signatures associated with clinical outcome might be modified by manipulating co-stimulatory signals and whether they reflected changes in T-cell function. FindingsNetwork analysis of CD8 T-cell transcriptomes identified a common signature of T-cell exhaustion during responses to chronic viral and self-antigen. However, whereas exhausted responses were associated with poor viral clearance, they predicted a less severe course in multiple autoimmune diseases. In autoimmunity, we found that where CD8 T-cell exhaustion was high a concurrent signature of CD4 T-cell co-stimulation was low. We used this signature to identify CD2 as a co-stimulatory signal preventing the development of T-cell exhaustion during controlled in-vitro assays, recreating expression patterns seen in patients with severe autoimmunity and leading to enhanced T-cell survival during primary and repeated stimulation. Finally, exogenous signals through the inhibitory receptor PD1 restricted development of this robust population. InterpretationHere we show that the process of CD8 T-cell exhaustion is important not only in facilitating viral persistence during chronic infection, but also in shaping the immune response to persistent self-antigen during autoimmunity. The observed correlation between T-cell exhaustion, co-stimulation, and subsequent prognosis increases our understanding of the pathogenesis of severe autoimmune disease and might help stratify patients for personalised therapy. Furthermore, our data suggest that future modulation of T-cell exhaustion may lead to new therapeutic opportunities in autoimmunity. FundingWellcome Trust, Biomedical Research Centre, National Institute for Health Research.