Repertoire isn't everything for GVHD

Question: When is defective memory cell function a good thing? Answer: When effector memory T (T_EM) cells are administered in allogeneic hematopoietic cell transplants (HCTs). T_EM cells are defective in their capacity to induce GVHD. In this issue of Blood, Juchem et al employ a clever preclinical allogeneic HCT strategy to gain insight into why this is so.¹  After transplantation of univariate naive(T_N), central memory (T_CM), and T_EM CD4 T cells, the investigators conclude that repertoire differences cannot explain the T_EM defect. Interestingly, functional deficits intrinsic to T_EM cells appear to contribute to their impotence.² 

The transplant field continues to search for strategies to prevent GVHD after allogeneic HCT while maintaining the beneficial affects of donor T cells including engraftment, antitumor responses, and immune reconstitution. A number of groups have demonstrated that CD4 and CD8 T_EM cells differ from other subsets because unlike T_N or T_CM cells, the former fail to induce significant GVHD in several experimental model systems.^1-6  Despite such observations, evidence continues to accumulate that T_EM cells provide antitumor reactivity, the raison d'etre of most allogeneic HCTs.^7,8  Thus, understanding why T_EM cells behave differently from other subsets could lead to promoting the use of multiple T-cell populations to minimize complications and maximize the benefits of allogeneic HCT.

Several explanations have been proposed to account for the poor GVHD activity by T_EM cells. For example, their migration and priming could be deficient because these cells lack CD62L and CCR7 to gain entry into secondary lymphoid tissue. Thus, relative exclusion from sites where efficient allo-activation occurs, that is, lymph nodes, might dampen their ability to induce GVHD. However, GVHD occurs in the absence of lymph nodes as well as with T cells lacking CD62L and CCR7 indicating that trafficking alone cannot account for the failure of T_EM cells to cause GVHD.^9,10  Another difference between T_EM cells and naive T-cell populations that could influence their ability to induce GVHD may be their TCR repertoire, as memory cells have been proposed to contain significantly fewer overall alloreactive specificities.¹¹  Finally, T_EM cells may also be limited in their GVHD response as a consequence of diminished clonal expansion and/or weak pathogenic effector functions.

In an attempt to shed light on the contributions of repertoire and function, Juchem et al developed an elegant approach to eliminate repertoire differences between populations of donor T cells administered for transplantation. Using TCR transgenic CD4 T cells specific for a viral hemagglutinin (HA) antigen, the authors directly obtained naive T cells and generated memory populations using ex vivo stimulation followed by transfer, that is, parking, in syngeneic RAG-deficient recipients before cell sorting for T_EM and T_CM populations (see figure). Conditioned recipients expressing the defined viral HA epitope in all tissues were transplanted to examine T-cell activation, expansion, function, and GVHD induction. In contrast to both T_N and T_CM cells, which readily induced weight loss, lethality, and pathology characteristic of GVHD, donor T_EM cells were clearly defective in this capacity on a per cell basis. Hence, given the identical antigen specificity of all 3 populations, repertoire in this model could be eliminated as mea culpa for the observations.

An important finding by Juchem et al was that during the first month after HCT, in most target tissues including spleen, liver, and bone marrow, similar numbers of T_EM cells compared with T_N and T_CM cells could be recovered. Thus, overall poorer survival, reduced expansion, and engraftment did not appear to contribute to T_EM cells' deficiency. However, examination of one tissue, the colon, did yield intriguing results: unambiguously decreased numbers of CD4 T_EM cells versus the other populations. The reduction reached ∼ 20 times two weeks after transplant. Consistent with such observations, pulse label experiments demonstrated the frequency of colonic T_EM cells in division was also far less than T_N cells. This latter finding highlights that in addition to secondary lymphoid organs, further stimulation and division clearly occurs in target tissues weeks after transplant. Based on the known significance of the GI tract in the development of GVHD, these observations are not likely coincidental. Diminished inflammation in the GI tract, particularly early after transplant, is clearly recognized as an important venue to contain development and propagation of GVHD. In total, these findings support the hypothesis that the transgenic CD4 T_EM cells induce early GVHD activity but fail to sustain responses leading to full-fledged and severe disease. Of note, cotransfer experiments demonstrated that the presence of T_N cells together with T_EM cells, a scenario that would typically reflect the heterogenous populations of T cells applied in the clinic, resulted in further suppression of T_EM expansion in the colon as well as in the spleen. Why or how alloreactive T_N cells may regulate T_EM cells and whether Tregs are involved will await further investigation.

What is known regarding function by T_EM cells in this model? Cytokine analysis demonstrates that the progeny of these cells in the colon as well as spleen exhibit diminished production of IFN-γ evidenced by both overall frequency and the magnitude of IFN-γ produced. Although small decreases in IL-2 and IL-17 were also noted by cells in the colon, the overall low frequency of producing cells suggests a less important role here versus IFN-γ. However, reduced IL-2 production was also apparent by T_EM progeny in the spleen, where a significant frequency of IL-2–producing cells was detected. Why the progeny of T_EM cells exhibit diminished functional responsiveness (and expansion in the colon) is presently unknown. One clue may reside in the elevated PD-1 expression observed on T_EM progeny a month after transplant whose signaling may contribute to exhaustion and reduced functional capacity. Thus in the colon, diminished T_EM cell numbers together with their diminished cytokine production predicts the decreased pathology observed in this tissue.

Models of polyclonal T cell–induced GVHD have also contributed to our understanding of the roles of repertoire and intrinsic defects with respect to T_EM cells' inability to sustain GVHD. Interestingly, prior studies employing polyclonal donor T cells in GVHD models reported recovering fewer T_EM cells versus T_N cells. This contrasts with the virtually equivalent numbers detected in all tissues examined except within the colon in the transgenic model.^8,9  Juchem et al propose that the polyclonal findings may reflect the greater levels of alloreactive T cells in the T_N populations, perhaps together with inhibition of T_EM cells when they are cotransplanted with T_N cells. Thus, while repertoire could play a role should it differ between T_N cells and T_EM cells, the findings by Juchem et al illustrate that regardless of repertoire, there are fundamental differences between T_EM and T_N cells. To date, accumulating evidence supports the notion that in preclinical mouse models, T_EM cells may be pretty, pretty good for allogeneic BMT administered as therapy to treat hematopoietic malignancies. Nonetheless, we must still curb our enthusiasm regarding translation to the clinic as transplanters await further definition as well as purification of the complex human T-cell memory pool.¹²