Intra-Leukemic IFNγ Signaling Mediates Cell Cycle Suppression and Chemoresistance in AML

Abstract

Introduction: The acute myeloid leukemia stem cell (LSC) model proposes that LSCs are a quiescent and functionally distinct cellular fraction that can escape chemotherapy and produce disease during relapse. Here, we reevaluated the relationship between quiescence and LSCs, and explored cell non-autonomous mechanisms contributing to chemotherapy escape. Methods: Using the definition that LSCs can engraft in immunodeficient mice, we evaluated the serial engraftment capability and cell cycle status of xenotransplanted AML patient samples and AML cells derived from AML-iPSC (iAML). To functionally interrogate regulators of cell cycle and chemotherapy escape, we used molecular profiling and perturbation studies to identify and validate mechanisms of cell cycle suppression. Results: To test the paradigm that quiescent LSCs are responsible for relapse after chemotherapy, we xenotransplanted both G0 and cycling cells from fresh primary AML cells or iAML. Contrary to previous reports, both G0 and cycling cells were transplantable, indicating that LSCs can exist in any cell cycle status. Having demonstrated that cell cycle status is heterogeneous in primary AML and not linked to cell-autonomous LSC identity, we speculated that cell cycle can be regulated by cell-extrinsic mechanisms. In particular, we hypothesized that leukemia cells themselves can interact to affect cell cycle, engraftment, and response to chemotherapy. To investigate this hypothesis, we established primary patient co-injection xenograft models in which the co-injected samples could be identified using two different HLA antibodies (Figure). Each sample expanded when transplanted by itself. Surprisingly, in all three patient pair models, one of the two co-injected patient samples (“Suppressed”) displayed very limited and decreased engraftment relative to the other sample (“Dominant”). Importantly, the Suppressed samples displayed less active cycling in Co-injection mice than in Alone mice, suggesting that the presence of Dominant samples inhibited the cycling of Suppressed samples. However, when isolated from primary recipients for secondary transplantation, Suppressed samples from Co-injection and Alone mice demonstrated similar engraftment capacity and caused similar mortality. Similar results were obtained in an iAML syngeneic co-injection model, demonstrating that suppression also occurs in cells derived from the same patient. These findings support our hypothesis that AML cell cycle can be suppressed by other AML cells without affecting the Suppressed cells' ability to initiate disease. Next, to identify mechanisms regulating cell cycle suppression, we evaluated RNA-seq of Dominant and Suppressed samples from our AML co-injection xenograft models and single-cell RNA-seq data from 12 AML patient specimens described in the literature. Suppressed samples exhibited enrichment of IFNγ signaling, which was also observed in G1 cells compared to S/G2/M cells in patient samples, suggesting IFNγ involvement in regulating cell cycle. To validate that IFNγ signaling mediates suppression, we disrupted expression of the IFNγ receptor IFNGR1 using CRIPSR/Cas9 gene editing in the Suppressed samples in the co-injection model. IFNGR1 loss increased cell cycle in the Suppressed samples compared to controls. Thus, IFNγ signaling contributes to AML cell cycle suppression. Finally, we investigated the role of IFNγ signaling in modulating cell cycle and chemosensitivity by treating our co-injection model with chemotherapy with and without anti-IFNGR1 blocking antibody. RNA-seq of both Suppressed and Dominant samples before and after chemotherapy revealed that Dominant samples and samples that escape suppression after chemotherapy downregulate IFNγ receptor expression and signaling. Furthermore, in samples enriched for IFNγ signaling, blocking IFNGR1 before and during chemotherapy treatment increased cell cycling before chemotherapy and decreased leukemia burden after chemotherapy. Conclusion: Here, we show that AML cell cycle can be suppressed in a cell non-autonomous manner that is mediated by IFNγ signaling. Moreover, blocking IFNγ signaling can increase chemosensitivity through activating the cell cycle. These findings provide novel insights into the mechanisms of chemotherapy escape and highlight IFNγ signaling as a potential therapeutic target to counter chemoresistance.