Memory CD8. ⁺ T Cells Specific for Minor Histocompatibility Antigen H60 Can Mediate the Graft-Versus-Leukemia Effect.

Abstract 1332

Poster Board I-354

In allogeneic hematopoietic stem cell transplantation (alloSCT), donor T cells mediate graft-vs-leukemia (GVL) but also cause graft-vs.-host disease (GVHD). Our previous studies indicate that memory T cells (T_M) induce less GVHD than do naïve T cells (T_N). Therefore the selective infusion of memory T cells could improve immune reconstitution with less GVHD. However, human memory T cells have a more restricted TCR repertoire than do T_N and therefore may be less effective in mediating GVL. One approach for addressing this concern would be to vaccinate donors against a single miHA expressed by host leukemic cells. Then the selective transfer of memory T cells could improve both GVL and immune reconstitution with less GVHD. Questions remain about how best to apply this strategy. For example, differences in target antigen distribution could affect the re-expansion of transferred memory cells and determine whether vaccination augments GVHD. Also, different types of memory cells could behave differently. To begin to optimize this approach in mouse models we chose as our target antigen the H-2K^b-restricted minor histocompatibility antigen (miHA) H60. We cloned the H60 peptide sequence (LTFNYRNL) into a genetic construct encoding the heavy chain of a monoclonal antibody against DEC205. Donor C3H.SW (H-2^b, H60-) mice were vaccinated with a single injection of anti-DEC205-H60 plus an agonist antibody against CD40 (FGK45). By two months post vaccination, H60-reactive memory cells were a stable population that comprised approximately 4-8% of splenic and peripheral blood CD8 cells. Approximately 50% of H60-tetramer+ cells had central memory phenotype, which could be ideal as central memory T cells mount strong anti-tumor responses. Similar results were obtained by using a prime/boost approach with H60-pulsed DCs. To test their function, CD44⁺ memory CD8 cells from H60-vaccinated mice were sort-purified and transferred into recipient B6 mice congenic for H60 (B6.H60; expression hematopoietically restricted ). By day 7 post transplant, H60-specific CD8⁺ T_M from vaccinated mice comprised 70-90% of total splenic and blood CD8 cells, as compared to 1-5% in recipients of T_M from unvaccinated mice. In contrast to anti-H60 responses by naïve CD8 cells, expansion of H60-tetramer+ cells from H60-vaccinated mice did not require CD4 help. To test whether target antigen distribution affects the re-expansion of H60-reactive cells from H60-vaccinated mice, expansion was compared in B6.H60 and B6.H60<right arrow>B6.actH60 (ubiquitous expression of H60 driven by an actin promoter) bone marrow chimeras. Approximately 60% of splenic and lymph node CD8 cells in both recipient groups were H60-tetramer+, though there was a trend towards increased overall numbers of tetramer+ cells in B6.H60 recipients. We are currently testing the GVL potency of T_M from H60-vaccinated against mouse models of chronic phase (CP-CML) and blast crisis chronic myelogenous leukemia (BC-CML). In an ongoing experiment, very low numbers of CD8⁺ T_M from H60-vaccinated mice mediate potent GVL against CP-CML relative to CD8⁺ T_M from unvaccinated donors. Small numbers of CD8⁺ T_M cells from vaccinated mice also mediated GVL against BC-CML, which is typically GVL resistant. Future studies will better define the potency of T_M from H60-vaccinated mice and test these cells against CP-CML and BC-CML that do or do not express H60 but are otherwise identical. GVHD effect will also be tested.