The Influence of Migration, Alloreactive Repertoire and Memory Subset on the Differential Ability of Naive and Memory T Cells To Induce GVHD.

Allogeneic stem cell transplantation (alloSCT) can cure many hematologic malignancies and hematopoietic stem cell disorders but is frequently complicated by graft vs. host disease (GVHD). We and others have published that memory phenotype (CD62LloCD44hi) T cells do not cause GVHD but can engraft and mount immune responses, including graft-vs.-tumor (GVT) effects. Importantly, these findings apply to GVHD induced by CD4 or CD8 T cells, and across major MHC differences. Thus, the inability to induce GVHD seems a fundamental property of memory phenotype (M) cells. We are investigating several hypotheses to explain why naïve cells (N) cause potent GVHD but M cells do not, including: 1. M cells lack CD62L and fail to traffic to LN and PP, two sites that may be essential for initial priming to allogeneic antigens and 2. M cells have a restricted alloreactive repertoire. There are at least two types of memory cells: central memory (CM) and effector memory (EM). EM cells (CD62LloCCR7neg) quickly express effector functions upon restimulation, preferentially migrate to tissues and spleen (bypassing LNs) and have relatively lower proliferative potential. CM cells (CD62LhiCCR7pos) have hybrid properties of both N and effector cells. Like N cells, their adhesion and chemokine receptors promote migration to LNs; however, their effector functions are more vigorous than N cells, yet slower than EM cells. It is of clinical interest for the design of M cell transfusion to determine the contributions of CM vs. EM cells in GVHD, but prior studies have not clearly investigated CM cells. We therefore compared GVHD initiated by N, EM and CM CD4 cells in the B6 (H-2b) ->BALB/c (H-2d) model. CM cells caused severe GVHD, similar to that induced by N cells. This suggests that LN homing enables M cells to initiate GVHD and/or that CM cells have unique functional properties required to initiate GVHD which are lacking in EM cells. We next asked whether LN entry was required for donor cells to initiate GVHD. N cells induced GVHD, albeit less severe than in WT recipients, in recipients lacking all secondary lymphoid organs (LN, PP and spleen), suggesting that priming in tissues is sufficient. Both of these findings support those of Beilhack, et al. (Blood2005;106:1113) that migration affects T cell initiation of GVHD. However, we also found that N cells from CD62L-deficient donors caused GVHD. Thus, because CD62L function and even secondary lymphoid tissue are dispensable for GVHD induction, we conclude that homing differences alone do not control GVHD and that critical functional properties other than homing must limit EM cells’ ability to induce GVHD. To address the role of repertoire, we increased the alloreactive precursor frequency of M cells. We isolated Thy1.1 (B6.C) (H-2d) effector T cells from mice with active GVHD (BALB/c recipients), and therefore with expanded anti-BALB/c repertoire, and parked them in syngeneic RAG1-/- (B6.C) recipients to allow them to differentiate into M cells. After 4 weeks we re-isolated the alloreactive M cells and transplanted them into new BALB/c recipients. The alloreactive M cells engrafted and caused transient skin GVHD without gut GVHD, unlike the more severe chronic GVHD induced by fresh N cells. This muted GVHD argues that although M cells can cause limited GVHD if their repertoire is artificially increased, they are still lacking in other functional properties required for optimal GVHD induction.