Deciphering Minor Histocompatibility Antigen Immunoreactivity through Optimized T Cell Expansion Strategies

Background: For patients with high-risk blood cancers, cure following allogeneic hematopoietic stem cell transplantation (AHSCT) is mostly attributable to its Graft-vs-Leukemia (GVL) effect that is mediated by donor T lymphocytes. As both GvL activity and graft-vs-host disease (GvHD) are mediated by donor T cells present in the donor graft, AHSCT would greatly benefit from the ability to select donor T cells capable of generating GvL activity without causing GVHD. We have previously identified 98 minor histocompatibility antigens (MiHAs) preferentially expressed on hematopoietic cells (Granados et al, Leukemia 2016). T cells recognizing these MiHAs have the potential to eliminate hematologic cancer cells while sparing healthy tissues. MiHAs are human-leukocyte antigen (HLA)-restricted peptides generally originating from germline polymorphisms and therefore broadly expressed. A single amino-acid difference in MiHA sequences between the patient and its donor is sufficient to elicit strong GvL activity in HSCT. However, the low frequency of MiHA-specific T cells in donor grafts may not allow for optimal GvL. Thus, ex vivo expansion of anti-leukemia cells represents an attractive strategy to enhance GVL activity. We have recently demonstrated the feasibility and safety of infusing ex vivo expanded T cells against a single MiHA as well as observed signs of anti-cancer activity in a phase I first in human clinical trial (Roy et al, ASH 2022).

Objectives: In order to improve on the ability of such T cells to fight cancer, we investigated i) whether the generation of T cells targeting multiple MiHAs is feasible; ii) the impact of MiHA-specific CD8 T cell precursor frequency on T cell expansion; and iii) anti-MiHA T cell immunoreactivity.

Methods: We have evaluated 38 different MiHAs either separately or in combination with other MiHAs according to the donor MiHA profile, targeting 7 different HLAs from 22 healthy volunteers. We co-cultured dendritic cells (DC) pulsed with single or multiple MiHAs with peripheral blood mononuclear cells (PBMC) for 21 days. DC and T cell immune phenotypes were measured using flow cytometry, and immune reactivity using IFNγ, TNFα, IL-2 and CD107a expression upon MiHA peptide restimulation.

Results: Dendritic cells pulsed with multiple MiHAs showed differences in expression of co-stimulatory molecules and activation markers compared to single-MiHA pulsed DCs. Multiple-MiHA pulsed DCs led to a significant increase in the frequency of MiHA-specific T cells expansion, with 37.8% of cultures showing MiHA-specific T cells compared with 12.0% in single-MiHA pulsed DC conditions (p=0.0046). Interestingly, MiHA competition for the same HLA in multiple-MiHAs pulsed DC conditions did not alter the frequency of MiHA-specific T cell expansion nor their reactivity. Next, we sought to investigate whether anti-MiHA T cell generation was limited by the initial number of CD8+ T cells. We found that increasing the CD8+/CD4+ T cell ratio 2.9-fold in the initial culture resulted in 58.6% of cultures showing MiHA-specific T cells. Importantly, CD8+ T cell enrichment was associated with greater multi-MiHA reactivities compared to non-enriched cultures, with up to three simultaneous MiHA reactivities upon multiple MiHA DC exposure (p<0.0001). In an attempt to develop a broad immunizing strategy, we also exposed T cells to the entire set of 98 MiHAs. While no significant difference was observed in DC phenotype upon MiHA pulsing, the 98-MiHA-mix resulted in superior expansion of MiHA-specific T cells with 88.5% of cultures showing MiHA reactivity. However, as the 98-MiHA-Mix contained both foreign and self MiHAs for a given donor, it also resulted in the expansion of a few anti-self MiHA specific T cells. Finally, we found that ex vivo stimulation with different MiHAs demonstrated variable levels of anti-MiHA responses. We therefore uncovered a hierarchy of MiHA most likely to be recognized and expanded from multiple donor repertoires and used in MiHA-directed T-cell therapy. This MiHA hierarchy was independent of predicted MiHA-HLA binding affinity.

Conclusions: MiHA-specific T cell expansion resulting from the use of multiple peptides and CD8 enrichment, along with the identification of a hierarchy of MiHA immunoreactivity should translate into significant improvement in adoptive cell therapy and will be tested soon in a Phase I clinical trial for patients with high-risk blood cancers.