The involvement of TEMRA CD8 is evident in a large array of immunological conditions ranging from auto- to allo-immunity. Nevertheless, the factors leading to their accumulation and activation remain ill-defined and, efficient therapeutics to control their inflammatory response is lacking. Here, we show that IL-15-stimulated TEMRA from kidney-transplant (KT) recipients promote inflammation by inducing the expression of CX3CL1 by endothelial cells in an IFN-γ- and TNF-α-dependent manner. The responsiveness of TEMRA to IL-15 is not restricted to chronic stimulation, as TEMRA from healthy volunteers respond earlier and faster when compared to effector memory (EM). IL-15 induces antiapoptotic signals and promotes proliferation dependent of PI3K/Akt, MAPK, and ERK pathways. Without ex vivo stimulation, TEMRA cells are metabolically more active than naive and EM, as shown by their high ATP reservoir and a high expression of genes involved in glycolysis, glutaminolysis, and the Pentose Phosphate Pathway. Upon stimulation, TEMRA adapt their metabolism by sustaining an increased mitochondrial respiration and glycolysis. Finally, we show that the inhibition of glycolysis is highly effective in preventing endothelial inflammation induced by TEMRA from KT recipients. Together, our findings highlight the metabolic fitness that tightly regulates the immune function of TEMRA in physiological and pathogenic situations.

Objective. To compare the effects of rituximab (RTX) and conventional immunosuppressants (CIs) on CD4+ T cells, Treg cells, and CD8+ T cells in antineutrophil cytoplasmic antibody-associated vasculitis (AAV).Methods. A thorough immunophenotype analysis of CD4+, Treg, and CD8+ cells from 51 patients with AAV was performed. The production of cytokines and chemokines by CD8+ T cells stimulated in vitro was assessed using a multiplex immunoassay. The impact of AAV B cells on CD8+ T cell response was assessed using autologous and heterologous cocultures.Results. CD4+ and Treg cell subsets were comparable among RTX-treated and CI-treated patients. In contrast, within the CD8+ T cell compartment, RTX, but not CIS, reduced CD45RA+CCR7-(TEMRA) cell frequency (from a median of 39% before RTX treatment to 10% after RTX treatment [P < 0.01]) and efficiently dampened cytokine/chemokine production (e.g., the median macrophage inflammatory protein 1α level was 815 pg/ml in patients treated with RTX versus 985 pg/ml in patients treated with CIs versus 970 pg/ml in those with active untreated AAV [P < 0.01]). CD8+ T cell subsets cocultured with autologous B cells produced more proinflammatory cytokines in AAV patients than in controls (e.g., for tumor necrosis factor-producing effector memory CD8+ T cells: 14% in AAV patients versus 9.2% in controls [P < 0.05]). In vitro disruption of AAV B cell-CD8+ T cell cross-talk reduced CD8+ T cell cytokine production, mirroring the reduced CD8+ response observed ex vivo after RTX treatment.Conclusion. The disruption of a pathogenic B cell/CD8+ T cell axis may contribute to the efficacy of RTX in AAV. Further studies are needed to determine the value of CD8+ T cell immunomonitoring in B cell-targeted therapies.