The PD-L1 overexpression is an important event of immune escape and metastasis in triple-negative breast cancer (TNBC), but the molecular mechanism remains to be determined. Interferon gamma (IFNγ) represents a major driving force behind PD-L1 expression in tumor microenvironment, and histone deacetylase 2 (HDAC2) is required for IFN signaling. Here, we investigated the regulation of HDAC2 on the IFNγ-induced PD-L1 expression in TNBC cells. We found the HDAC2 and PD-L1 expression in TNBC was significantly higher than that in non-TNBC, and HDAC2 was positively correlated with PD-L1 expression. HDAC2 promoted PD-L1 induction by upregulating the phosphorylation of JAK1, JAK2, and STAT1, as well as the translocation of STAT1 to the nucleus and the recruitment of STAT1 to the PD-L1 promoter. Meanwhile, HDAC2 was recruited to the PD-L1 promoter by STAT1, and HDAC2 knockout compromised IFNγ-induced upregulation of H3K27, H3K9 acetylation, and the BRD4 recruitment in PD-L1 promoter. In addition, significant inhibition of proliferation, colony formation, migration, and cell cycle of TNBC cells were observed following knockout of HDAC2 in vitro. Furthermore, HDAC2 knockout reduced IFNγ-induced PD-L1 expression, lymphocyte infiltration, and retarded tumor growth and metastasis in the breast cancer mouse models. This study may provide evidence that HDAC2 promotes IFNγ-induced PD-L1 expression, suggesting a way for enhanced antitumor immunity when targeting the HDAC2 in TNBC.

HIF-1α has a broad impact on tumors, including enhanced utilization of glucose, tumor cell stemness, migration, metastasis and so on. In pilot study, we found that the expression of HIF-1α significantly increased in breast cancer cell lines and tissue samples with higher malignant behaviors and decreased in luminal subtype breast cancer cells and tissue samples. We analyzed and found there is one large CpG island in HIF-1α promoter around transcription start site, and the hypermethylation occurred at these CpGs and their surrounding non-CpGs sites. Epigenetic events driving tumorigenesis has been characterized. However, knowledge is lacking on the non-CpGs methylation of HIF-1α promoter in breast cancer cells. We validated that non-CpGs methylation can directly regulate HIF-1α expression by luciferase activity assay. We also found DNMT3a and Mecp2 play vital role in methylation at non-CpGs and CpGs sites. In addition, we noticed that H3K9ac modification could promote the transcription of HIF-1α in MDA-MB-231 cells by binding to the region contained hypomethylated non-CpG and CpG sites. Taken together, the hypomethylation status at non-CpG and CpG loci in HIF-1α promoter and H3K9ac modification together contribute to maintain higher HIF-1αactivity in invasive breast cancer cells when compared with the non-invasive breast cancer cells, which may establish a tissue-specific epigenetic modification pattern and point to the new directions for future understanding breast cancer therapy.

Advanced breast cancer (BC), especially basal like triple-negative BC (TNBC), is a highly malignant tumor without viable treatment option, highlighting the urgent need to seek novel therapeutic targets. Arylsulfatase D (ARSD), localized at Xp22.3, is a female-biased gene due to its escaping from X chromosome inactivation (XCI). Unfortunately, no systematic investigation of ARSD on BC has been reported. In this study, we observed that ARSD expression was positively related to ERα status either in BC cells or tissue specimens, which were associated with good prognosis. Furthermore, we found a set of hormone-responsive lineage-specific transcription factors, FOXA1, GATA3, ERα, directly drove high expression of ARSD through chromatin looping in luminal subtype BC cells. Opposingly, ARSD still subjected to XCI in TNBC cells mediated by Xist, CpG islands methylation, and inhibitory histone modification. Unexpectedly, we also found that ectopic ARSD overexpression could inhibit proliferation and migration of TNBC cells by activating Hippo/YAP pathway, indicating that ARSD may be a molecule brake on ERα signaling pathway, which restricted ERα to be an uncontrolled active status. Combined with other peoples’ researches that Hippo signaling maintained ER expression and ER + BC growth, we believed that there should exist a regulative feedback loop formation among ERα, ARSD, and Hippo/YAP pathway. Collectively, our findings will help filling the knowledge gap about the influence of ARSD on BC and providing evidence that ARSD may serve as a potential marker to predict prognosis and as a therapeutic target.