Blocking HDAC3 in Bone Marrow Stromal Cells Has Direct Anti-Multiple Myeloma Effect and Modulates T Cell Function

Introduction

Histone deacetylases (HDAC) are therapeutic targets in Multiple Myeloma (MM) and the pan-HDAC inhibitor Panobinostat is FDA approved. However, clinical trials showed that non-selective HDAC inhibitors have a narrow therapeutic index, prompting the development of isoform-specific HDAC inhibitors such as the HDAC6 inhibitor Ricolinostat. Preclinical studies have shown that HDAC3 downregulation/inhibition exerts an anti-myeloma effect, suggesting that HDAC3 may be a novel molecular target in MM. However, the effect of targeting HDAC3 in the bone marrow (BM) microenvironment has not been investigated. The BM microenvironment, particularly stromal and immune cells, have been shown to be key in supporting MM growth, immune escape, and resistance to therapy. We sought to study whether targeting HDAC3 in BM stroma cells (BMSC) has direct cytotoxic effects against stromal cells and/or impacts MM cell and T cell function through secreted factors and/or cell-to-cell contact mechanisms.

Methods

We targeted HDAC3 in human BMSC line HS-5 and primary-derived BMSC from MM patients (MM-BMSC) via siRNA or via the HDAC3-selective inhibitor BG45. MTT assay and flow cytometry with annexin V/PI staining were used to assess viability and apoptosis post-HDAC3 KD/pharmacologic inhibition. Downregulation of HDAC3 was confirmed by immunoblotting. To assess the effect of HDAC3 silencing in MM-BMSC on MM cells, we performed siRNA-mediated HDAC3 KD in HS-5 and MM-BMSC, removed transfection media after 48 hours, and performed co-culture experiments with MM1S.Luc and H929.Luc MM cell lines. After 96 hours MM cell proliferation was measured using the Luciferase Assay. Scrambled siRNA transfected HS-5 and MM-BMSC were used as control. To assess the effect of HDAC3 silencing in BMSC on T cells, we isolated peripheral blood mononuclear cells from healthy donors by Ficoll density gradient separation. CD3⁺ T cells were isolated using EasySep™ Human T Cell Isolation Kit and activated using Dynabeads Human T-Activator CD3/CD28. Cytokine array profiling was performed on the cell-culture supernatant using the Raybiotech C-series Human Cytokine Array according to manufacturer's protocol.

Results

First, we showed that pharmacological inhibition of HDAC3 via BG45 does not have direct anti-proliferative or cytotoxic effect on the HS-5 stromal cell line at doses up to 5-fold higher than the EC50 of MM cell lines. Consistent with this data, siRNA-mediated HDAC3 KD does not have direct anti-proliferative and/or cytotoxic effects on HS-5 or MM-BMSC. However, when MM cell lines were co-cultured with HDAC3-silenced HS-5 or MM-BMSC, there was a statistically significant decrease in BMSC-induced MM proliferation compared to control (HS-5: 38% decrease in MM proliferation in HDAC3 KD, p < 0.05; MM-BMSC: 24.1-27.5% decrease in MM proliferation in HDAC3 KD, p < 0.05). To assess whether cell-to-cell contact is necessary to mediate this growth arrest, we cultured MM cell lines in conditioned media from HDAC3 or scrambled siRNA transfected BMSC and discovered that the former was sufficient to induce growth/proliferation arrest (HS-5: 31.5% decrease in MM proliferation in HDAC3 KD, p < 0.05; MM-BMSC: 24.3% decrease in MM proliferation in HDAC3 KD, p < 0.05), suggesting a paracrine effect.

We next showed that conditioned supernatant from HDAC3-KD HS-5, cultured alone or in combination with MM cell lines, significantly inhibited the proliferation of activated T-cells (HS-5 supernatant: 47.8% decrease in MM proliferation in HDAC3 KD; HS-5 + MM co-culture supernatant: 32.6% decrease in MM proliferation in HDAC3 KD). To identify the soluble factors responsible for this paracrine effect, we subjected HDAC3-KD and scrambled-transfected HS-5 supernatant to cytokine profiling. We identified a number of cytokines modulated by HDAC3 KD, which we are currently analyzing. Among these, we identified a 1.37-fold increase in soluble PD-L1 when HDAC3 was silenced in HS-5 cells.

Conclusions

Our results show that inhibiting HDAC3 in BMSC mediates anti-MM and anti-T cell effect via soluble factors. Gene expression profiling and proteomic studies are currently underway to define the molecular mechanisms supporting these effects. Our preliminary data suggest that targeting HDAC3 in BMSC may alter the checkpoint repertoire in the BM milieu, suggesting potential utility of therapeutic strategies combining HDAC3 and checkpoint inhibitors.