BACKGROUND: Immune checkpoint blockadewith anti-PD-1/PD-L1 therapyhas demonstrated remarkable efficacy in multiple tumor types. Biomarker candidates for predicting likelihood of response to targeted immunotherapy are being actively investigated including inhibitory or activating receptors on CD8+ lymphocytes, corresponding ligands on tumor or antigen-presenting cells (APCs), T-cell functionality, and the T-cell receptor (TCR) repertoire found within a tumor microenvironment. Myelofibrosis (MF) and Chronic Myeloid Leukemia (CML) are tumors responsive to immunotherapy, most notably allogeneic transplantation (alloSCT), and donor lymphocyte infusion. Although tyrosine kinase inhibitors can improve patient outcomes, a potentially curative therapeutic option other than alloSCT is needed.

PURPOSE: To determine the immune profile of the bone marrow tumor microenvironment in patients with CML and MF compared to healthy donors in order to assess the rationale and potential efficacy of novel immune checkpoint therapies.

METHODS: Cryopreserved bone marrow aspirate mononuclear cells (MNCs) from healthy donors (HDs) (n=11), untreated CML (n=9) or MF (n= 12) were analyzed by flow cytometry. CD3+ CD8+ lymphocytes were divided into naïve, central memory (CM), effector memory (EM), and terminal effector (TEMRA) subsets for analysis. Expression of immune checkpoint receptors including PD-1, 4-1BB, TIM3, LAG3, and TIGIT were evaluated on each population. Known corresponding ligands including PD-L1 and PD-L2 were assessed in CML samples on blasts, plasmacytoid dendritic cells (pDCs), myeloid dendritic cells (mDCs), and monocytes. T-cell function was evaluated by cytokine production, cytotoxicity, and proliferation in CD3+ CD8+ PD1+ or PD1- populations. To assess the TCR repertoire found within the tumor microenvironment, non-naive CD8+ T-cells were sorted into PD-1+ and PD-1- populations, and then CDR3 region of the TCRB gene, together with sufficient flanking sequence to identify most V, D, and J genes was sequenced using the immunoSEQ platform from Adaptive Biotechnologies.

RESULTS: There was a significant difference in the CML CD3+ CD8+ subset distribution compared to HDs with EM% increased at 60.01% vs. 41.25% (p =0.0137), and TEMRA 44.51% vs. 20.64% (p=0.0004). CM% trended downwards (32.15% to 21.58%, p=0.118) while naïve% was equivalent in CML and HDs (22.13% vs. 20.87%). The percentage of PD-1+ non-naïve CD8+ T-cells (EM, TEMRA, CM combined) was significantly increased in CML samples at 55.14% (range 31-69%) compared to HDs at 38.98% (range 34.8% to 55.5%; p=0.0050). PD-1 expression was consistently increased across all subgroups in CML (CM: 67.06% vs 53.22%, EM: 60.01% vs. 41.25%, TEMRA: 44.51% vs 20.64% p <0.05 for all). There was no statistically significant difference in CML compared to HDs for secondary receptors including TIGIT, TIM3, LAG3, or 4-1BB. Fewer than 5% of CML blasts were positive for the PD-L1 or PD-L2 ligands, however PD-L1 expression was increased on mDCs compared to HD samples (53.08% vs 24.63%; p=0.0015). In contrast to these findings in CML there was no significant proportional difference in CD8+ subsets, PD-1 status, or other receptors between MF and HDs. Anti-CD3/28 stimulation did not induce differential IFN-γ/TNF-alpha production, granzyme production, or proliferation (Ki67+) among the CD8+ PD-1+ or PD-1- T-cells from CML samples. To begin to estimate T cell clonality in the bone marrow tumor microenvironment, TCRβ sequencing of sorted non-naïve CD8- T-cells showed several clones markedly overrepresented in the diseased PD-1+ compartment.

Conclusions: The CML tumor microenvironment is enriched in CD8+ T-cells expressing the inhibitory receptor PD-1 while APC subsets express increased PD-L1. This represents a potential axis of tumor driven immunosuppression amenable to immune checkpoint blockade. This is in contrast to MF, where the immunoprofile was not detectably different from healthy donors. These findings may reflect differences in tumor immunogenicity, cytokine mileu, or the APC types present. In-vivo testing using murine models for both diseases is underway to gain a better understanding of the role of immune checkpoint therapies.

Disclosures

Mangan:Incyte Corporation: Membership on an entity's Board of Directors or advisory committees. Vignali:Adaptive Biotechnologies: Employment, Equity Ownership. Emerson:Adaptive Biotechnologies: Employment, Equity Ownership. Robins:Adaptive Biotechnologies: Consultancy, Equity Ownership, Patents & Royalties. Yusko:Adaptive Biotechnologies: Employment, Equity Ownership.

Author notes

*

Asterisk with author names denotes non-ASH members.

This icon denotes a clinically relevant abstract

Sign in via your Institution