Expression of hormone receptors is associated with specific immunological profiles of the breast cancer microenvironment

Breast cancer is the most commonly occurring cancer in women. Despite recent advances in multimodal treatment, advanced and recurrent cases are challenging to cure [1], and the urgent need to develop innovative treatment strategies has to be addressed. Owing to the development of immune checkpoint inhibitors (ICIs) that reinvigorate the adaptive immune response in the tumor microenvironment and the successful application of these ICIs in various neoplasms, tumor immunology has recently gained renewed interest across multiple cancers [2, 3]. ICIs have also been applied for breast cancer treatment, but their effectiveness is limited, likely due to the immunosuppressive tumor microenvironment [4‐9]. Elucidating the unique immunomodulatory mechanisms of the breast cancer microenvironment will provide significant insights into the development of new therapeutic strategies.

Hormone dependency is one of the prominent biological features of breast cancer. Approximately 70% of breast cancer cells express the estrogen receptor (ER), resulting in ER-dependent growth of breast cancer [1]. Therefore, treatment strategies that inhibit ER function are frequently used as postoperative adjuvant therapy for patients with early-stage breast cancer and as systemic therapy for metastatic cases [1]. The progesterone receptor (PgR) is clinically considered a complementary marker of hormone dependency in breast cancer [1, 10] because it is driven partly, but not exclusively, by ER-mediated transcriptional events [11, 12]. PgR is a binding partner and major modifier of ER-mediated processes, suggesting its additional role in breast cancer other than its identification as an ER activity marker [13, 14]. The androgen receptor (AR) is a nuclear transcription factor with a diverse range of biological actions, mainly in the development and maintenance of the male reproductive system [15]. It is widely expressed in all breast cancer subtypes to varying extents, with approximately 60–80% of the cases being AR-positive [16‐18]. Furthermore, according to The Human Protein Atlas, AR expression, at both gene and protein levels, is second-highest in breast cancer, after prostate cancer, among various malignancies [19]. Although the function of AR in breast cancer depends on the tumor subtype, treatment, and other factors, it is suggested to have a tumor-promoting role [20‐22], thereby attracting attention as a new therapeutic target for breast cancer treatment [23, 24].

Some recent studies have suggested that sex steroid hormones and their receptor signaling have immune regulatory functions. In vitro studies have shown that estrogen can expand the regulatory T-cell fraction and reduce the function of antigen-presenting cells [25, 26]. In addition, estrogen can promote immune tolerance by interfering with human leukocyte antigen-II expression in ER-positive breast cancer cell lines [27]. These findings suggest that estrogen signaling in the tumor microenvironment regulates anti-tumor immunity [2]. In agreement with this, hormone receptor-positive breast cancer is characterized by low infiltration of tumor-infiltrating lymphocytes (TILs) and minimal response to ICIs [5, 9, 28‐32]. To the best of our knowledge, the immunological function of PgR in breast cancer has not yet been reported. However, limited studies have shown a relationship between PgR expression and tumor immunity in breast cancer; PgR expression is inversely associated with programmed death-ligand 1 (PD-L1) expression in epithelial cells or the stroma and the infiltration of CD8+ T and CD20+ B cells [9, 33, 34]. Further, immune regulatory functions of AR signals have been demonstrated via in vivo models of various autoimmune diseases and some malignancies [35]. Moreover, in breast cancer, AR expression is inversely correlated with immune cell infiltration and cytotoxic immune activity, suggesting an immunosuppressive effect of AR signals [36‐38].

Despite fragmentary evidence on sex steroid hormone signals and tumor immunity, the interactions between hormone receptors and immune cells are not well understood because of the complexity of the immune milieu in the breast cancer microenvironment and limited reports on systematic evaluation of immune cell composition in breast cancer tissue [39]. In this study, we systematically analyzed the relationship between the expression of sex steroid hormone receptors such as ER, PgR, and AR and the immunological profiles of breast cancer tissues. Our results demonstrated that hormone receptor expression, at both gene and protein levels, correlates with specific immunological profiles of the breast cancer microenvironment, strongly suggesting their direct or indirect immunomodulatory role.

This study systematically analyzed the relationship between hormone receptor expression (at both gene and protein levels) and the immunological profile of breast cancer tissues, including multiple immune cell fractions. Expression of ER and AR in breast cancer tissues was associated with a decreased infiltration of immune cells into the tumor microenvironment. More specifically, ER expression was associated with decreased infiltration of macrophages and CD4+ T cells, while AR expression with reduced macrophage infiltration. These results were consistent at the gene and protein expression levels.

The present analysis showed that gene expression of hormone receptors was inversely correlated to total immune cell infiltration into the tumor microenvironment (Fig. 1a–f). This was further verified by performing GSEA on the METABRIC dataset, which analyzed the relationship between the expression levels of hormone receptors and hallmark gene set collection (50 gene sets) representing specific well-defined biological states or processes. Expression levels of ESR1, PGR, and AR showed a significant positive correlation with gene sets representing estrogen response, such as estrogen response early and estrogen response late (Additional file 1: Figure S1a–c). On the other hand, ESR1 and AR expression levels showed a significant negative correlation with multiple gene sets representing immunological processes such as inflammatory response, allograft rejection, complement, and interferon-gamma response (Additional file 1: Figure S1d, f). Likewise, a significant negative correlation was observed between PgR expression level and gene sets of inflammatory response, allograft rejection, and interferon-gamma response (Additional file 1: Figure S1e). These findings suggest that the gene expression levels of hormone receptors, such as ESR1, PGR, and AR, correlate with immunosuppressive phenotypes of breast cancer.

Several preclinical studies have suggested that ER signaling may suppress the immune response of breast cancer [2, 25‐27], and ER-positive breast cancers have shown reduced TIL infiltration and minimal response to ICI [5, 9, 28‐32]. This is consistent with the present findings, suggesting an immunomodulatory function of ER. Although numerous studies have analyzed the relationship between a single immune cell lineage and hormone receptor expression [32, 54‐57], few reports have systematically investigated the multiple immune cell composition of breast cancer tissues [51, 58‐60]. The present study reports the first systematic analysis demonstrating that ER expression is preferentially associated with reduced infiltration of macrophages and CD4+ T cells (Fig. 1g, 3c, 3f). This result is consistent with earlier reports that show a negative correlation between ER expression and intratumoral infiltration of macrophages and CD4+ T cell [54‐57]. Further, an inverse correlation between ER expression and CD8+ T cells has been reported previously [32, 54]; however, in our analysis, this correlation was weaker than that for CD4+ T cells, both at the gene and protein levels (Figs. 1g and 3). In addition, PGR expression showed a weaker correlation with total immune content than ESR1 and AR, as estimated by CIBERSORTx. Moreover, the correlation between PgR status and hTIL and hPD-L1 was insignificant at the protein level, probably due to the small number of PgR-positive samples. To validate these findings, a greater number of samples are required to be analyzed.

AR signaling is suggested to have immunomodulatory functions in some malignancies, including breast cancer [35‐38]; however, its immunological role in breast cancer has not been fully validated. To the best of our knowledge, a single study has investigated the relationship between immune cell composition and AR expression in breast cancer using the CIBERSORT algorithm [38]. Tumors with high AR expression were reported to be associated with pro-cancer regulatory T cells, and those with low AR expression were associated with anti-cancer immune cells, such as CD4, CD8, gamma delta T cells, and memory B cells in ER-positive breast cancer. In our study, CIBERSORTx was run in absolute mode and included all subtypes in the analysis; therefore, a simple comparison was not possible. High AR-expressing tumors showed reduced macrophage infiltration into the tumor microenvironment; this finding was consistent with CIBERSORTx and FCM analysis results.

In the present study, expressions of ER and AR were inversely correlated with hPD-L1 expression (Fig. 2b, f). We analyzed the PD-L1 expression in each immune cell fraction by FCM analysis, similar to our previous study [39]. According to hormone receptor status, we found that ER positivity was associated with decreased PD-L1 expression in Mo/Mφ and mDC (Additional file 1: Figure S2d, h), PgR positivity with increased PD-L1 expression in MDSCs and decreased PD-L1 expression in NK cells (Additional file 1: Figure S3f, j), and AR positivity with increased PD-L1 expression in CD8+ T and decreased PD-L1 expression in Mo/Mφ cells (Additional file 1: Figure S4b, d). PD-L1 expression reflects ongoing (or active) immune responses in addition to immunosuppression via the PD-1/PD-L1 pathway [60‐62]. Thus, ER positivity and AR positivity may reflect specific immune response in the breast cancer microenvironment.

Findings of this study suggest that the hormone receptor signals may primarily affect specific immune cell lineages. In particular, the expression of ER and AR showed a negative correlation not only with total immune content in the tumor microenvironment but also with certain immune cell subsets such as macrophages and CD4+ T cells. This indicates that specific immune cell lineages may be primary targets of the immunoregulatory function of the hormone receptor signals. Our next goal is to verify this hypothesis using an in vitro or in vivo analysis model.

In our previous study [39], a relatively small number of samples were used for FCM analysis, and compared to those of the general breast cancer cohort, the clinicopathological characteristics of tumors were biased with larger invasive tumor sizes, more ER-negative cases, and higher Ki67 cases. Similarly, in this study, the small sample size prevented us from performing subgroup analyses based on the tumor subtype. Therefore, the inclusion of a greater number of samples and more detailed analyses are required in future studies. In this exploratory analysis, the cut-off value for ER and PgR positivity was set at ≥ 10% according to our previous study [39]. However, in a recent clinical guideline update [63], 1% is recommended as a cutoff value for ER- and PgR-positive cells because of limited but present data on endocrine therapy benefit for cancers with 1% to 10% of cells staining ER positive. Therefore, the analysis of Figs. 2, 3, and 4 was repeated using the cutoff points 1% for ER and PgR, as shown in Additional file 1: Figure S5a–h. When the cutoff value is set to 1%, for Fig. 2, the expression of hPD-L1 showed a negative correlation with the protein expression of PgR (Fig S5d). Similarly, for Fig. 3, the significant association between ER positivity and some immune cell compositions were lost (i.e., leukocyte density, infiltration of total T, CD4+ T, Mo/Mφ) (Fig S5e). For Fig. 4, PgR positivity gained significant association with decreased leukocyte density and reduced infiltration of total T, CD4+ T, MDSC, DC (Fig S5f). No meaningful changes were observed in other results. Only 2 cases each had changes in ER and PgR status due to changes in the cutoff values. Therefore, we speculate that the discrepancies in the analysis results are a consequence of the small sample size.

In the present study, the gene expression levels of hormone receptors correlated with immunosuppressive phenotypes of breast cancer. The expression level of ER and AR proteins was associated with decreased tumor-infiltrating immune cells and decreased PD-L1 expression. These data suggest that hormone receptor signaling may suppress tumor immunity through a specific mechanism. Additionally, our data showed that certain immune cell lineages might get more strongly affected by hormonal signals than others, providing useful suggestions for further analysis of the effects of hormonal signals on tumor immunity.