T-Cell Transcriptome Analysis Reveals the Mechanisms Controlling Synergy Between Costimulation Blockade and mTOR Inhibition during Graft Versus Host Disease Prevention

While novel therapeutics for graft versus host disease (GvHD) are desperately needed, new prevention and treatment strategies for this disease have been extremely slow to reach the clinic. One of the major barriers has been the historic lack of a powerful translational system capable of both (1) interrogating the clinical efficacy of novel approaches and (2) discovering the underlying mechanisms of GvHD prevention. Our lab has addressed this unmet need with the development of a non-human primate (NHP) model of GvHD, and the first therapeutic approach developed with the NHP model, costimulation blockade with CTLA4-Ig, is currently in clinical trials for GvHD prevention (Clinical Trials.gov #NCT01743131). While the first-in-disease trials of GvHD prevention add CTLA4-Ig to standard tacrolimus/methotrexate, previous work has suggested that mTOR inhibition with sirolimus is more pro-tolerogenic than calcineurin inhibition when paired with costimulation blockade. We have now addressed the possibility of pairing CTLA4-Ig with sirolimus to prevent GvHD in the NHP model, and have interrogated the outcomes using a systems-based approach. Our new experiments provide strong clinical, immunologic and trancriptomic evidence for potent synergy when CTLA4-Ig and sirolimus are combined for GvHD prevention. In this study, we determined the impact of the following treatment groups on GvHD after high-risk haplo-identical HSCT using T cell-replete peripheral blood stem cell transplantation: 1) no therapy (n=4) 2) CTLA4-Ig monotherapy (using the second generation CTLA4-Ig formulation, belatacept, n=3) 3) sirolimus monotherapy (n=4) 4) combination belatacept and sirolimus (n=3). To determine the relative impact of each therapeutic approach, we monitored clinical GvHD, GvHD-free survival, and flow cytometric signs of immune activation post-transplant. Moreover, to create a comprehensive molecular map of their impact on GvHD, we performed transcriptomic analysis, on CD3+/CD20- T cells that were purified on day +14 post-transplant and analyzed using the Affymetrix microarray platform. The synergistic impact of combined belatacept +sirolimus was evidenced through all analyses techniques: Thus, GvHD-free survival with belatacept + sirolimus was prolonged compared to all other groups (MST belatacept + sirolimus = 33d, p < 0.02 compared to MST for untreated controls (7.5d), belatacept monotherapy (9d, p < 0.03) and sirolimus monotherapy (12d) (Figure 1a). GvHD clinical scores mirrored the clinical survival (Figure 1b). In addition, canonical flow cytometric signs of CD8+ T cell activation (proliferation, measured by Ki-67, and excessive cytotoxicity measured by granzyme B overexpression) also mirrored the clinical synergy we measured with combined belatacept +sirolimus compared to all other groups (Figure 1c). Comparing the expression profile of T cells during acute GvHD in the four cohorts allowed examination of treatment synergy at an unprecedented level of molecular detail. Unsupervised analysis of transcriptomic signatures as a whole revealed clustering of principal component projections from belatacept + sirolimus to be strikingly similar to a large comparative healthy control cohort (n=28), underscoring the high degree of control alloreactivity with this regimen. Moreover, the comparison of differentially expressed genes from animals receiving belatacept + sirolimus revealed significant divergence from monotherapy with either belatacept or sirolimus, again underscoring the profound control of allo-activation that was observed with belatacept + sirolimus (Figure 2a). The mirror-image analysis provided additional support or this synergy, with transcriptional analysis of T cells from belatacept + sirolimus-treated animals most dissimilar from the transcriptome of unprophylaxed GvHD (Figure 2b). Those genes for which expression was significantly normalized by combination therapy showed pathway enrichment prominently in T cell effector function (prominently including granzyme signaling – Figure 2c), cytokine networks (prominently IL2, IL12 and IL-18 – Figure 2d), as well as in proliferation and cell cycle pathways (Figure 2e). These data reveal a treatment synergy between T cell costimulation blockade with CTLA-4-Ig and mTOR inhibition and suggest that this combination of therapy will be useful for acute GvHD immunoprophylaxis in humans. Disclosures No relevant conflicts of interest to declare.