AUTHORS
Tiankai Yin, BA, Michael D. Bern, MD, PhD2, Rachel L. Baits3, Nicole A. Dallas, BS4, Mark M. Li5, Hunter Cassidy Cochran, MD6, Sharon Heath7, Michael Harris Kramer, MD PhD8, Nichole Helton, MS9, Daniel R. George, BA10, Sai Mukund Ramakrishnan, MS11, Christopher A. Miller, PhD12, Timothy Ley, MD13
AFFILIATIONS
1 Washington University School of Medicine, Saint Louis, Missouri, United States
2 Washington University School of Medicine, Saint Louis, Missouri, United States
3 Washington University School of Medicine, Saint Louis, Missouri, United States
4 Washington University School of Medicine, Saint Louis, Missouri, United States
5 Washington University School of Medicine, Saint Louis, Missouri, United States
6 Washington University School of Medicine, Saint Louis, Missouri, United States
7 Washington University School of Medicine, Saint Louis, Missouri, United States
8 Washington University School of Medicine, Saint Louis, Missouri, United States
9 Washington University School of Medicine, Saint Louis, Missouri, United States
10 Washington University School of Medicine, Saint Louis, Missouri, United States
11 Washington University School of Medicine, Saint Louis, Missouri, United States
12 Washington University School of Medicine, Saint Louis, Missouri, United States
13 Washington University School of Medicine, Saint Louis, Missouri, United States
ABSTRACT
Acute Myeloid Leukemia (AML) is a malignancy of hematopoietic stem/progenitor cells, arising in the bone marrow. Although immune checkpoint blockade has been effective for some cancers, this approach has had limited efficacy in AML patients, for unknown reasons. A better understanding of the mechanisms contributing to the immune escape of AML at presentation may lead to new insights for identifying mechanism-based immunotherapies. There have been previous attempts at characterizing the immune response in AML, but differences in experimental techniques and disease heterogeneity in AML patients have precluded a consensus in the field about the dominant mechanisms of immune escape. To that end, we have characterized the immune response to AML by evaluating the immune cell phenotypes in the bone marrow (BM) immune microenvironment, and the AML-driven mechanisms that may correspond to these responses.
We performed 42-marker Cytometry by Time of Flight (CyTOF) to assess the immune cells in 59 representative, adult, de novo AML BM samples obtained at presentation, with 12 healthy donor bone marrow samples as comparators. The AML samples were chosen to be representative of common AML mutations, morphological (FAB) phenotypes, and responses to standard-of-care treatment regimens. This cohort had complete clinical and mutational annotation available. Bulk RNA sequencing was performed on all AML BM samples to define the transcriptomic features of the dominant AML clone in each.
Compared to healthy donors, AML bone marrow samples were significantly enriched for differentiated and effector lymphocyte phenotypes (B, T, and NK cells). This finding was most prominent in T cell subsets, where AML samples had significantly more memory effector CD4+ and CD8+ cells. Despite these overall trends, we observed substantial heterogeneity of T cell phenotypes between patients: one group was significantly enriched for cytotoxic effector memory T cells (“TEMRA-hi”), and another had predominantly naïve, non-activated T cells (“TEMRA-lo”). We hypothesized that these phenotypes may represent different mechanisms of immune escape: for example, AML samples with naïve lymphocytes may lack neoantigens, or may be immunologically silent due to other mechanisms (e.g. active immunosuppression mediated by AML cells). In contrast, the T cells of AML bone marrow samples with large numbers of activated, terminal effector T cells may be recognizing and attempting to eliminate the AML cells, but AML cells in this context may be resistant to T cell mediated killing, and are therefore able to escape.
To test this hypothesis, we compared total marrow RNA-seq data (representing expression from the dominant AML clone), from patients in the "TEMRA-lo" and “TEMRA-hi" groups. We found that AML cells from patients with predominantly naïve BM T cells (“TEMRA-lo”) often exhibited overexpression of VSIR, which encodes VISTA, a checkpoint that inhibits T cell activation. We inactivated Vsir in a murine AML model initiated by Dnmt3aR878H and Npm1cA (which is enriched for high VISTA expression), and showed that loss of VISTA significantly reduces immunosuppression in vitro. Likewise, Vsir inactivation potentiated in vivo immune clearance of these AMLs that were secondarily transplanted in immunocompetent mice.
In contrast, AML cells from BM enriched for TEMRA cells (“TEMRA-hi”) expressed significantly lower levels of interferon gamma receptor 2 (IFNGR2), an essential subunit of the IFNG receptor complex, and this low expression was confirmed by flow cytometry of AML blasts. Functional studies in AML cell lines suggested that the loss of IFNGR2 expression can lead to resistance to T cell-mediated killing by IFNG. Likewise, overexpression of IFNGR2 alone was sufficient to confer IFNG sensitivity to an AML cell line (Kasumi-1) that was otherwise resistant to IFNG mediated killing.
Collectively, these data suggest that VISTA overexpression on the AML cells of some patients may lead to suppression of T cell activation. Conversely, AML cells with low VSIR expression are associated with a TEMRA-rich (and therefore IFNG-rich) BM, and may utilize low IFNGR2 expression as a mechanism to escape IFNG-mediated killing. These mechanisms may be mutually exclusive, based on RNA-seq data from multiple AML cohorts. VISTA and IFNGR2 may be targets for novel immunotherapies, defined by the characteristics of the BM immune microenvironment at disease presentation.
DISCLOSURES
Tiankai Yin, BA: No relevant conflicts of interest to declare.