Introduction: We have shown that an increase of TEMRA CD8 (CD45RA⁺CCR7^-) evidenced 1 year post-transplantation is associated with a higher risk of kidney graft failure. The functionality of TEMRA CD8 and especially their reactivity upon donor-specific stimulation have never been investigated. In this study, we characterize before and one year after transplantation the functional response of TEMRA CD8 from Kidney Transplantation (KT) recipients after donor-specific stimulation and we setup a model of humanized NSG mice to investigate in vivo their pro-inflammatory response.
Methods: PBMC from 24 living-donor kidney-transplant recipients have been prospectively collected before and 12 month post-transplantation. Extensive phenotype of CD8 T cell subsets have been performed using multi-parameter flow cytometry. Naïve, TEMRA and EM CD8 from KT recipients have been cultured with irradiated donor-specific CD3-depleted PBMC. Expression of activation marker (CD25, CD69), cytotoxic marker (CD107a) and secretion of pro-inflammatory cytokines (Luminex Assay) have been analyzed after 24h and 48h of stimulation. Proliferation (CPD dilution) has been investigated after 5 days of culture. One to 5.10⁶ TEMRA CD8 purified from KT recipients were co-injected i.v. with 10⁷ T- and NK-depleted autologous PBMC in NSG HLA-A*0201 transgeneic mice. Xenogeneic GVHD was monitored by daily evaluation of the body weight and CD8 T cell infiltration was investigated using fluorescent microscopy in different organs.
Results: Kidney transplantation results in an increase of TEMRA CD8 (25 vs 39%), characterized by a high expression of GZMb and effector-associated transcription factor T-bet, and a concomitant decrease of NAÏVE CD8 whereas frequency of EM CD8 remains unchanged. Donor-specific stimulation results in the early expression of CD25⁺CD69⁺CD107a⁺ by TEMRA CD8 whereas NAÏVE CD8 failed to upregulate CD25 or CD107a. Expression of cytotoxic molecule CD107a by TEMRA was even enhanced after kidney transplantation. Donor specific stimulation results in the proliferation of TEMRA CD8. Finally, the infusion of TEMRA CD8 from KT recipients induces GVHD in 7 to 10 days with an infiltration of TEMRA in colon.
Conclusion: We demonstrate that TEMRA CD8 exhibit potent pathogenic function both in vitro when challenged with donor cells and in vivo resulting in an acute GVHD in immunodeficient mice. Our study highlights the need to design TEMRA specific therapy to improve kidney graft outcome.