HDAC Inhibition Rejuvenates Antitumor Immunity in Acute Myeloid Leukemia Via CD9 Transcriptional Reactivation

Background: Epigenetic therapies present modest and transient single-agent activities in acute myeloid leukemia (AML), but emerging evidence suggest that they may enhance antitumor immunity. Our recent findings pinpoint epigenetic silencing of CD9 could arrest differentiation and mediate immune escape, hereby conferring adverse clinical outcomes in pediatric AML (Xu et al, ASH2023). Based on these mechanistic insights, we pursued to identify potential epigenetic agent(s) to restore CD9 expression and provoke immune response against AML.

Methods: CD9 reactivator profiling was conducted with a collection of 896 epigenetic agents in 4 CD9- AML cell lines (MOLM13, THP-1, MV4-11 and OCI-AML3) adopting fold change in cell surface CD9 fluorescence intensity as readout, followed by ChIP-qPCR to validate the post-treatment reversal of focal histone acetylation. Ex vivo drug response of primary myeloblasts was measured in an optimized cytokine-supported culture system. In vivo immunoregulatory activity of the histone deacetylase (HDAC) inhibitor panobinostat was captured in NSG mice co-infused with the AML cell line MOLM13 and HLA-mismatched human peripheral blood mononuclear cells (PBMCs), with or without perturbation by CD9 neutralizing antibody. The underlying mechanisms were dissected by high-dimensional immunophenotyping of bone marrow infiltrated or circulating T cells using mass cytometry with a panel of 33 characteristic markers.

Results: An epigenetic compound library screen identified three major classes of chemicals with consistent CD9 reactivation potential: HDAC inhibitors (17 compounds, 2.05-2.96-fold), Aurora kinase inhibitors (3 compounds) and protein kinase C activators (5 compounds). Exposure of CD9- AML cell lines (n=8) or samples (n=9) to panobinostat significantly elevated CD9 mRNA and protein expression (3.1-32.2-fold, P<0.05), restored activating histone acetylation marks H3K9Ac and H3K27Ac at the CD9 loci (4.1-41.6-fold, P<0.05), and potently suppressed primary myeloblast proliferation ex vivo (median IC50: 21.4 nM), with preferential susceptibility of CD9- over CD9+ AML (IC50: 13.9 vs. 31.9 nM, P=0.022). Treatment of AML-grafted NSG mice with panobinostat showed notable elevation of CD9 expression (13.8-fold increase, P<0.001) in residual leukemic cells. In humanized NSG mice, panobinostat priming followed by human PBMC infusion exhibited a drastic reduction of systemic leukemic load (47.9%-81.3% decrease, P<0.01), a significantly prolonged survival (2-fold extension, P<0.001) concomitant with an enhanced bone marrow infiltration of CD3+ T cells (3.19-fold increase) compared with animals receiving panobinostat or PBMC alone. Such synergistic benefits were largely abrogated by co-administration of CD9 neutralizing antibody (P<0.05). Deep immune profiling of bone marrow-infiltrated or circulating T cells revealed strengthened release of cytotoxic moieties (granzyme B: 2.54-fold increase; perforin: 4.19-fold increase) and alleviated exhaustion (Tim-3: 4.22-fold decrease; OX40: 4.11-fold decrease) in mice receiving panobinostat treatment, and was reversible by CD9 neutralization.

Conclusions: HDAC inhibition could remarkably boost allogeneic immunity through epigenetic rewiring of CD9, inspiring a direct translation into clinical testing to evaluate the benefits of panobinostat priming prior to allogeneic hematopoietic stem cell transplantation in poor-risk AML patients.

No relevant conflicts of interest to declare.