Entinostat Decreases Immune Suppression to Promote Antitumor Responses in a HER2+ Breast Tumor Microenvironment

Abstract

Therapeutic combinations to alter immunosuppressive, solid tumor microenvironments (TME), such as in breast cancer, are essential to improve responses to immune checkpoint inhibitors (ICI). Entinostat, an oral histone deacetylase inhibitor, has been shown to improve responses to ICIs in various tumor models with immunosuppressive TMEs. The precise and comprehensive alterations to the TME induced by entinostat remain unknown. Here, we employed single-cell RNA sequencing on HER2-overexpressing breast tumors from mice treated with entinostat and ICIs to fully characterize changes across multiple cell types within the TME. This analysis demonstrates that treatment with entinostat induced a shift from a protumor to an antitumor TME signature, characterized predominantly by changes in myeloid cells. We confirmed myeloid-derived suppressor cells (MDSC) within entinostat-treated tumors associated with a less suppressive granulocytic (G)-MDSC phenotype and exhibited altered suppressive signaling that involved the NFκB and STAT3 pathways. In addition to MDSCs, tumor-associated macrophages were epigenetically reprogrammed from a protumor M2-like phenotype toward an antitumor M1-like phenotype, which may be contributing to a more sensitized TME. Overall, our in-depth analysis suggests that entinostat-induced changes on multiple myeloid cell types reduce immunosuppression and increase antitumor responses, which, in turn, improve sensitivity to ICIs. Sensitization of the TME by entinostat could ultimately broaden the population of patients with breast cancer who could benefit from ICIs.

Myeloid-derived suppressor cell dynamics control outcomes in the metastatic niche

“…MDSCs can be sub-divided into one of two states: monocytic M-MDSCs (typically assumed to be more immunosuppressive) and granulocytic/poly-mononuclear (G- or PMN-MDSCs) (1, 3, 6). The relative proportion of G-to M-MDSCs can alter the immunosuppressive properties of the tumor microenvironment (15, 82). For example, if the relative proportion of G-to M-MDSCs skews toward M-MDSCs, we would expect larger effects of MDSC delays (as seen in Figure 5), whereas the opposite would be expected if G-MDSCs dominate.…”

“…Modeling how tumors interact with the immune system is critical for understanding treatment responses and predicting the best possible therapeutic strategies in response to metastasis. Myeloid-derived suppressor cells (MDSCs) have been identified in various tumor microenvironments (8, 9, 15), where they can exert strong immunosuppressive effects leading to worse outcomes (12, 17, 37), yet a rigorous theoretical characterization of MDSC dynamics in the tumor microenvironment has remained lacking. Here, through the introduction of a stochastic delay differential equation (SDDE) model with which to study tumor-MDSC dynamics, we have provided means to characterize the plasticity of MDSCs and their effects on tumor progression and outcome.…”

“…Although we have assumed as such here – for lack of data with which to quantify subpopulation-specific MDSC rate parameters – future models ought to consider MDSC heterogeneity. MDSCs are typically classified into one of two possible cell types, monocytic (M-MDSCs) and granulocytic/polymononuclear (G-/PMN-MDSCs), which exhibit different levels of immunosuppression (1, 3, 15). M-MDSCs in metastatic breast cancer patients resemble monocytes isolated from patients with sepsis, indicating fascinating similarities between the immunosuppression capability of the MDSCs present in metastatic (but perhaps not primary) breast cancer patients and those involved in the immunosuppressive sepsis response (90).…”

“…Moreover, these additional data will permit the fitting of more detailed mathematical models that are able to describe patient-specific (or even tumor site-specific) dynamics, and quantify the possible benefits of treatments targeting MDSCs. Current knowledge suggests that shifting MDSC phenotypes towards G-MDSCs is beneficial as this state is less immunosuppresive (1, 15), however further characterization of these states is needed. The models we have developed of MDSCs in the tumor microenvironment do not consider space, although of course spatial architectures play an important role in tumor progression (28, 32), in primary growth as well as for circulating tumor cells that seed metastases (34, 91).…”

“…We include the anti-tumor immune populations of cytotoxic T (CTL) cells and natural killer (NK) cells. MDSC-CTL interactions are important given the primary function of intratumoral MDSCs is suppression of CTLs (1, 6, 15, 16). MDSC-NK interactions are also important (20–22, 24, 25), and NK cells are increasingly being studied as an immune population specifically affected by tumor cells to promote metastasis (46, 47).…”