Both Naïve and Memory T-Cell Subsets Participate in Alloresponses to HLA-Mismatched Targets – Implications for Adoptive Transfer of Viral Antigen-Specific T Cells.

Abstract 2439

Poster Board II-416

It is generally assumed that T cell responses to HLA disparate targets arise from the naïve pool. However, the risk of graft-versus-host disease (GvHD) also increases if the donor has circulating T cells recognizing multiple persistent DNA viruses, suggesting that memory T cells may contribute to the overall alloresponse. We used a flow cytometric CFSE-based proliferation & activation (CD38 expression acquisition) assay to systematically assess the contribution of naïve and memory T cells to alloreactivity. In some experiments donor peripheral blood mononuclear cells (PBMC) were primed for 8 days with allogeneic, HLA mismatched PBMC (alloPBMC), after which the intracellular cytokine production (ICS) was examined upon rechallenge with autologous or allogeneic PBMC (alloPBMC). CFSE-labeled umbilical cord mononuclear cells (UCMC) from five donors were stimulated with an irradiated pool of HLA-disparate donor peripheral blood mononuclear cells (alloPBMC). By day 8, over 80% of the CD4+ and CD8+ T cells were CFSE[dim], i.e. had proliferated, and expressed the activation marker CD38, indicating that naïve T cells can mount an alloresponse. Next, PBMC from adult donors were separated into naïve (CD57-CD45RO-), memory (CD57-CD45RA- and CD57-CD62L- T cells), and effector (CD27-CD45RO-) T cells. We found that adult donor naïve T cells can also mount an alloresponse; importantly, both memory T cell subsets – and to a lesser extent, effector cells – were CFSE[dim]CD38+ at least to the same extent as naïve T cells, indicating their potential to respond to alloantigens in addition to their priming foreign antigen. These data were confirmed using a different approach: By first priming responder T cells with HLA-disparate donor PBMC for 8 days, followed by the enumeration of alloreactivity by ICS after stimulation with the original stimulator PBMC. To exclude reactivity of responder T cells with viral antigens present in PBMC (EBV; CMV), the experiment was repeated using activated T cells (T-APC) as antigen presenting cells, and yielded identical results. The direct ex vivo alloreactivity of T cell subsets was next tested by stimulating donor PBMC with HLA mismatched donor PBMC or activated T cells as APC in an 18 hour ICS. This experiment confirmed the rapid kinetics of alloreactivity and dominance of memory T cells in the alloresponse. Since a prominent role of memory T cells was apparent from our experiments we next tested T cells specific for DNA virus-derived antigens. EBV- and CMV-specific T cells, tested against a panel of 30 T-APC with a broad coverage of the most prominent HLA, displayed exquisite specificity for certain mismatched HLA alleles, indicating that DNA virus antigen-specific T cells may indeed cross-react with cellular antigens presented in the context of mismatched HLA class I and II proteins. Collectively our data demonstrate that both naïve and memory T cells mount an alloresponse, but that memory T cells are more rapidly and strongly recruited by alloantigen stimulation. These findings indicate that in man alloresponses are not confined to particular subsets of post-thymic T cells and that donor-derived viral antigen-specific T cells in an unrelated recipient may cross-react with peptide-MHC complexes against which they were not negatively selected.