We have previously demonstrated that immunotherapy combining agonistic anti-CD40 and IL-2 (IT) results in synergistic anti-tumor effects. IT induces expansion of highly cytolytic, antigen-independent "bystander-activated" (CD8(+)CD44high) T cells displaying a CD25(-)NKG2D(+) phenotype in a cytokine dependent manner, which were responsible for the anti-tumor effects. While much attention has focused on CD4(+) T cell help for antigen-specific CD8(+) T cell expansion, little is known regarding the role of CD4(+) T cells in antigen-nonspecific bystander-memory CD8(+) T cell expansion. Utilizing CD4 deficient mouse models, we observed a significant expansion of bystander-memory T cells following IT which was similar to the non-CD4 depleted mice. Expanded bystander-memory CD8(+) T cells upregulated PD-1 in the absence of CD4(+) T cells which has been published as a hallmark of exhaustion and dysfunction in helpless CD8(+) T cells. Interestingly, compared to CD8(+) T cells from CD4 replete hosts, these bystander expanded cells displayed comparable (or enhanced) cytokine production, lytic ability, and in vivo anti-tumor effects suggesting no functional impairment or exhaustion and were enriched in an effector phenotype. There was no acceleration of the post-IT contraction phase of the bystander memory CD8(+) response in CD4-depleted mice. The response was independent of IL-21 signaling. These results suggest that, in contrast to antigen-specific CD8(+) T cell expansion, CD4(+) T cell help is not necessary for expansion and activation of antigen-nonspecific bystander-memory CD8(+) T cells following IT, but may play a role in regulating conversion of these cells from a central memory to effector phenotype. Additionally, the expression of PD-1 in this model appears to be a marker of effector function and not exhaustion.

Recent studies have underscored the critical role of retinoic acid (RA) in the development of lineage-committed CD4 and CD8 T cells in vivo. We have shown that under acute graft-versus-host disease (GVHD) inflammatory conditions, RA is upregulated in the intestine and is proinflammatory, as GVHD lethality was attenuated when donor allogeneic T cells selectively expressed a dominant negative RA receptor α that blunted RA signaling. RA can function in an autocrine and paracrine fashion, and as such, the host cell lineage responsible for the production of RA metabolism and the specific RA-metabolizing enzymes that potentiate GVHD severity are unknown. In this study, we demonstrate that enhancing RA degradation in the host and to a lesser extent donor hematopoietic cells by overexpressing the RA-catabolizing enzyme CYP26A1 reduced GVHD. RA production is facilitated by retinaldehyde isoform-2 (RALDH2) preferentially expressed in dendritic cells (DCs). Conditionally deleted RA-synthesizing enzyme RALDH2 in host or to a lesser extent donor DCs reduced GVHD lethality. Improved survival in recipients with RALDH2-deleted DCs was associated with increased T cell death, impaired T effector function, increased regulatory T cell frequency, and augmented coinhibitory molecule expression on donor CD4⁺ T cells. In contrast, retinaldehydrogenase isoform-1 (RALDH1) is dominantly expressed in intestinal epithelial cells. Unexpectedly, conditional host intestinal epithelial cells RALDH1 deletion failed to reduce GVHD. These data demonstrate the critical role of both donor and especially host RALDH2⁺ DCs in driving murine GVHD and suggest RALDH2 inhibition or CYP26A1 induction as novel therapeutic strategies to prevent GVHD.