TCF-1 is Required By CD8+ T Cells for the Maintenance of Alloimmune Responses in Graft-Vs-Host Disease

Graft-vs-host disease (GVHD) is a common complication of allogeneic stem cell transplant (alloSCT) wherein donor T cells target alloantigens on recipient tissues. It is unclear how alloimmune responses are maintained in GVHD despite abundant antigen, which causes T cell anergy, deletion and exhaustion. Previously, we identified alloreactive TCF-1^high T cells arising post-transplant that resemble exhausted progenitors (T_EXP) capable of propagating immune responses in other chronic antigen models. Here, we sought to further characterize these cells in the B6→129 MHC-matched GVHD mouse model, in which 129 recipients express the immunodominant H-2K^b-restricted minor histocompatibility antigen (miHA) H60. At day +7 post-transplant, alloreactive CD8⁺ cells specific to H60 (as determined by MHC-I-tetramer staining; Tet^H60+) were nearly uniformly PD-1^hiTox^hi whereas Tet^H60- cells displayed a bimodal distribution into discrete PD-1^hiTox^hi and PD-1^loTox^lo populations, indicative of more diverse antigen experiences. Among these both Tet^H60+ and Tet^H60- cells were TCF-1^hi cells. TCF-1^hi Tet^H60+ cells were uniformly CD39^loTox^hiPD-1^hi, which is a canonical T_EXP phenotype. In contrast, among activated Tet^H60- cells there were TCF-1^hi cells that were CD39^loTox^hiPD-1^hi and Tox^loPD-1^lo. At later times in spleen and lymph node, and in GVHD target tissues, these populations of TCF-1⁺ Tet^H60+ and Tet^H60- were found. To test if these CD39^loTCF-1^hi T_EXP had proliferative advantages in GVHD, we sorted congenic TCF-1^hiCD39^lo and TCF-1^loCD39^hi CD8⁺ cells from recipient spleens 14-days post-transplant and adoptively transferred them in competition in a 1:1 ratio (of Tet^H60+ cells) into newly transplanted recipients. Among Tet^H60+ cells in all tissues at day 14 post-transfer, TCF-1^hiCD39^lo-sorted progeny greatly outperformed TCF-1^loCD39^hi-sorted progeny. In line with their role as a source of GVHD effectors, progeny of TCF-1^hiCD39^lo cells were mostly TCF-1^loCD39^hi; however, a fraction remained TCF-1^hi consistent with their being able to undergo self-renewal. Conversely, we observed few if any TCF-1⁺ progeny of CD39^hi cells. We next tested whether TCF-1 was an important mediator of T cell fitness or whether it was only a marker for functionality. To do so we competed congenic wild-type (WT) and Tcf7^p45-/- (p45^-/-) donor CD8 cells, which lack the N-terminal β-catenin binding domain of TCF-1, in allogeneic (129) and syngeneic (B6) recipients. Strikingly, p45^-/- CD8 cells were greatly outcompeted by WT CD8 cells in 129 recipients in all tissues and at all times post-transplant, among both Tet^H60+ and Tet^H60- cells. In contrast, in B6 recipients, WT and p45^-/- cells remained evenly matched, suggesting that full-length TCF-1 isoforms are dispensable for lymphopenia-induced T cell expansion. Further, p45^-/- cells were also not disadvantaged when adoptively transferred into B6 mice and acutely challenged with H60 antigen by vaccination. Together these data suggest a model wherein TCF-1^hi progenitor like T cells are seeded in GVHD target organs where they may serve as a key local source for GVHD effectors, and moreover, full-length TCF-1 is itself critical for alloreactive T cell fitness in GVH responses.