Current doctrine is that microvascular inflammation (MVI) triggered by a transplant -recipient antibody response against alloantigens (antibody-mediated rejection) is the main cause of graft failure. Here, we show that histological lesions are not mediated by antibodies in approximately half the participants in a cohort of 129 renal recipients with MVI on graft biopsy. Genetic analysis of these patients shows a higher prevalence of mismatches between donor HLA I and recipient inhibitory killer cell immunoglobulin-like receptors (KIRs). Human in vitro models and transplantation of β2-microglobulin-deficient hearts into wild-type mice demonstrates that the inability of graft endothelial cells to provide HLA I-mediated inhibitory signals to recipient circulating NK cells triggers their activation, which in turn promotes endothelial damage. Missing self-induced NK cell activation is mTORC1-dependent and the mTOR inhibitor rapamycin can prevent the development of this type of chronic vascular rejection.

Antibody-mediated rejection is currently the leading cause of transplant failure. Prevailing dogma predicts that B cells differentiate into anti-donor-specific antibody (DSA)-producing plasma cells only with the help of CD4+ T cells. Yet, previous studies have shown that dependence on helper T cells decreases when high amounts of protein antigen are recruited to the spleen, two conditions potentially met by organ transplantation. This could explain why a significant proportion of transplant recipients develop DSA despite therapeutic immunosuppression. Using murine models, we confirmed that heart transplantation, but not skin grafting, is associated with accumulation of a high quantity of alloantigens in recipients’ spleen. Nevertheless, neither naive nor memory DSA responses could be observed after transplantation of an allogeneic heart into recipients genetically deficient for CD4+ T cells. These findings suggest that DSA generation rather result from insufficient blockade of the helper function of CD4+ T cells by therapeutic immunosuppression. To test this second theory, different subsets of circulating T cells: CD8+, CD4+, and T follicular helper [CD4+CXCDR5+, T follicular helper cells (Tfh)], were analyzed in 9 healthy controls and 22 renal recipients. In line with our hypothesis, we observed that triple maintenance immunosuppression (CNI + MMF + steroids) efficiently blocked activation-induced upregulation of CD25 on CD8+, but not on CD4+ T cells. Although the level of expression of CD40L and ICOS was lower on activated Tfh of immunosuppressed patients, the percentage of CD40L-expressing Tfh was the same than control patients, as was Tfh production of IL21. Induction therapy with antithymocyte globulin (ATG) resulted in prolonged depletion of Tfh and reduction of CD4+ T cells number with depleting monoclonal antibody in murine model resulted in exponential decrease in DSA titers. Furthermore, induction with ATG also had long-term beneficial influence on Tfh function after immune reconstitution. We conclude that CD4+ T cell help is mandatory for naive and memory DSA responses, making Tfh cells attractive targets for improving the prevention of DSA generation and to prolong allograft survival. Waiting for innovative treatments to be translated into the clinical field ATG induction seems to currently offer the best clinical prospect to achieve this goal.