Pan-Cancer Prognostic Role and Targeting Potential of the Estrogen-Progesterone Axis

Introduction Estrogen receptors (ESRs) and progesterone receptors (PGRs) are associated with the development and progression of various tumors. The feasibility of ESRs and PGRs as prognostic markers and therapeutic targets for multiple cancers was evaluated via pan-cancer analysis. Methods The pan-cancer mRNA expression levels, genetic variations, and prognostic values of ESR1, ESR2, and PGR were analyzed using the Gene Expression Profiling Interactive Analysis 2 (GEPIA2) and cBioPortal. The expression levels of ERa, ERb, and PGR proteins were detected by immunohistochemical staining using paraffin-embedded tissue specimens of ovarian serous cystadenocarcinoma (OV) and uterine endometrioid adenocarcinoma (UTEA). Correlation between immunomodulators and immune cells was determined based on the Tumor and Immune System Interaction Database (TISIDB). Results ESR1, ESR2, and PGR mRNAs were found to be differentially expressed in different cancer types, and were associated with tumor progression and clinical prognosis. ERa, ERb, and PGR proteins were further determined to be significantly differentially expressed in OV and UTEA via immunohistochemical staining. The expression of ERa protein was positively correlated with a high tumor stage, whereas the expression of PGR protein was conversely associated with a high tumor stage in patients with OV. In patients with UTEA, the expression levels of both ERa and PGR proteins were conversely associated with tumor grade and stage. In addition, the expression levels of ESR1, ESR2, and PGR mRNAs were significantly associated with the expression of immunomodulators and immune cells. Conclusion ESR1, ESR2, and PGR are potential prognostic markers and therapeutic targets, as well as important factors for the prediction, evaluation, and individualized treatment in several cancer types.


INTRODUCTION
Nowadays, cancer has become a leading cause of death worldwide, with continuously increasing rates of morbidity and mortality (1). In 2017, approximately 2.6 million Chinese individuals died of various types of cancer, accounting for 26.07% of the total deaths (2,3). Multiple therapeutic strategies, including but not limited to surgery, radiotherapy, chemotherapy, and immunotherapy, have been developed for the comprehensive and individualized treatment of malignant tumors. However, overall clinical outcomes in patients with advanced cancers are still dissatisfactory, especially given the concomitant adverse effects. Therefore, there is an urgent need to identify potentially valuable molecular targets for the improvement of therapeutic efficacy and specificity.
Estrogen receptors (ESRs) belong to nuclear receptor superfamily of hormone-inducible transcription factors, which comprise ERa and ERb, encoded by ESR1 and ESR2, respectively (4,5). PGR encodes a member of the steroid receptor superfamily, named progesterone receptors (PGRs) (6). In physiological state, the activation of ESRs and PGRs by the binding of their ligands are associated with a series of normal physical activities. However, under pathological conditions, ESR1, ESR2, and PGR have been demonstrated to be associated with tumorigenesis and tumor progression (7,8). For instance, ESR1 is well characterized as a factor that promotes cell proliferation in breast cancer (9). In contrast, ESR2 seems to be a tumor suppressor gene (10), which is not expressed in early stages of breast cancer (11). Further, PGR is associated with the development of breast cancer (12). In addition to breast cancer, ESR1, ESR2, and PGR also mediate the progression of prostate cancer (13)(14)(15), colon cancer (16)(17)(18), ovarian cancer (19)(20)(21), and lung cancer (22)(23)(24). Accordingly, ESR1, ESR2, and PGR may be prognostic biomarkers as well as potential therapeutic targets for a variety of cancer types, necessitating further evaluation.
In this study, we conducted a comprehensive pan-cancer analysis of ESR1, ESR2, and PGR on the basis of online databases. The expression levels of ESR1, ESR2, and PGR, and the correlation of ESR1, ESR2, and PGR with overall survival (OS) and disease-free survival (RFS) in patients were assessed using Gene Expression Profiling Interactive Analysis 2 (GEPIA2). The expression levels of ERa, ERb, and PGR proteins in ovarian serous cystadenocarcinoma (OV) and uterine endometrioid adenocarcinoma (UTEA) were validated using in-house tissue specimens, and the relationship between protein levels of ERa, ERb, and PGR and clinicopathological characteristics of OV or UTEA patients was explored. Genetic alterations and immunological effects of ESR1, ESR2, and PGR were analyzed using the cBioPortal and Tumor and Immune System Interaction Database (TISIDB), respectively.

Patient Tissue Sample Collection
Forty-two paraffin-embedded OV and 51 UTEA tissue specimens were collected from patients who underwent surgery at the Hightech district of the First Affiliated Hospital of Anhui Medical University (Hefei, Anhui, China) between 2017 and 2019. We also collected 11 specimens of normal ovarian tissue from 42 patients with OV (31 specimens of tumors involving bilateral ovarian tissue were excluded) and 34 specimens of normal endometrial tissue adjacent to the cancer in 51 patients with UTEA (17 specimens of tumors involving the entire endometrial tissue were excluded). No patient had a history of other malignant tumors and no patient had undergone preoperative interventions such as radiotherapy or chemotherapy. Each patient provided written informed consent, and the study was approved by the institutional review board.

GEPIA2 Dataset Analysis
The expression levels of ESR1, ESR2, and PGR mRNAs in tumor and matched normal samples were compared using the GEPIA2 database, which is a webserver that provides cancer genomics data based on TCGA, and the GTEx database (25). In this study, differentially expressed gene analysis of tumor and matched normal samples, isoform profiling, and clinicopathological stage analysis were performed using the GEPIA2 dataset. Differentially expressed gene analysis and clinicopathological stage analysis were conducted by one-way ANOVA. Genes with |log2FC| > 1 and Q-value < 0.01 were considered to be differentially expressed. We used log 2 (TPM+1) for log-scaling differential expression in different clinicopathological stages, and regarded Pr(>F) < 0.05 to be statistically significant. In addition, correlative prognostic analysis of ESR1, ESR2, and PGR, including OS and RFS, was conducted to evaluate the prognostic significance using log-rank test for hypothesis evaluation at the median cutoff with 50% for either low-or high-expression cohorts.

cBioPortal Analysis
The cBioPortal for Cancer Genomics is a widely used openaccess website, providing a visualization and analysis tool for multidimensional cancer genomics data (26,27). The cBioPortal was employed to analyze the OncoPrint, mutual exclusivity, alteration frequency in multiple cancer types, and amino acid changes in proteins and for the Clinical Attribute Test. Mutual exclusivity analysis among ESR1, ESR2, and PGR was conducted using Log 2 odds ratio, P-value, and Q-value, and P-value < 0.001 and Q-value < 0.001 were regarded as statistically significant.

TISIDB Analysis
The TISIDB is a user-friendly web portal containing 988 immune-related anti-tumor genes derived from 4,176 records in 2,530 publications. This database enables users to analyze the function of selected genes in the tumor-immune interplay through high-throughput data analysis or literature mining (28). In this study, we used TISIDB to construct heat maps for analyzing the spearman correlations between the expression levels of ESR1, ESR2, and PGR and immunomodulators and immune cells in multiple cancer types. A p value < 0.05 was regarded as statistically significant.

Statistical Analysis
SPSS22.0 was used for data analysis. Chi-square test was used for variable comparison, with p < 0.05 regarded as statistically significant. Spearman's method was used to assess the correlation between factors. p < 0.05 was regarded as statistically significant.

ESR1, ESR2, and PGR mRNAs Are Differentially Expressed in Various Cancers
To explore the expression of ESR1, ESR2 and PGR in pan-cancer, we analyzed their mRNA levels via GEPIA2. We found that the ESR1 mRNA was highly expressed in breast invasive carcinoma (BRCA) and OV samples compared with their matched normal samples ( Figure 1A). In contrast, low expression levels of ESR1 mRNA were found in bladder urothelial carcinoma (BLCA), cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC), liver hepatocellular carcinoma (LIHC), testicular germ cell tumors (TGCT), and uterine carcinosarcoma (UCS) samples. ESR2 mRNA was observed to be highly expressed only in the lymphoid neoplasm diffuse large B-cell lymphoma (DLBCL) samples, whereas a low expression of ESR2 mRNA was found in adrenocortical carcinoma (ACC), OV, and TGCT samples ( Figure 1B). In addition, a low expression of PGR mRNA was found in CESC, colon adenocarcinoma (COAD), OV, prostate adenocarcinoma (PRAD), rectal adenocarcinoma (READ), TGCT, UCEC, and UCS samples ( Figure 1C). Together, ESR1, ESR2, and PGR are differentially expressed in multiple cancer types.

Correlation Between the Expression of ESR1, ESR2, and PGR mRNAs and Tumor Stage Across Multiple Cancers
To examine the clinical relevance of ESR1, ESR2, and PGR in pan-cancer, we analyzed the correlations of their expression with tumor stage. As shown in Figures 4A-C, ESR1 and PGR transcription levels were correlated with the tumor stage (p < 0.05), and the higher expression of ESR1 and ESR2 are associated with higher tumor stage. In contrast, no significant association between ESR2 and tumor stage was observed (p > 0.05). Taken together, the expression of ESR1 and PGR was significantly associated with pan-cancer tumor stage.

Expression Levels of ERa, ERb, and PGR Proteins in OV and UTEA
To detect the expression levels of ERa, ERb, and PGR proteins in OV and UTEA, which were not reported in any previous studies, we next performed immunohistochemical staining using paraffin-embedded tissue specimens. The results showed that the expression level of ERa was significantly higher, while ERb and PGR were significantly lower in OV compared to these in their matched normal samples (p < 0.05) (Supplementary Material Figure S1A). The expression levels of ERa, ERb, and PGR proteins were significantly lower in UTEA samples   Figure S1B). Moreover, ERa, ERb, and PGR proteins were highly expressed in 42.9%, 50%, and 21.4% of OV tumor samples and in 64.7%, 35.3%, and 60.8% of UTEA tumor samples, respectively (Table 1). Together, the expression trends of ERa, ERb, and PGR proteins in OV and UTEA as well as the matched normal tissue are in consistent with the findings obtained from GEPIA2 ( Figure 5).

Association Between the Expressions of ERa, ERb, and PGR Proteins and the Clinicopathological Characteristics of Patients With OV or UTEA
To explore the clinical significance of ERa, ERb, and PGR expression in OV and UTEA, we further correlated their expression to clinicopathological characteristics of patients with OV or UTEA. Interestingly, the expression level of ERa protein was positively correlated to a high tumor stage, whereas the expression level of PGR protein was inversely correlated to a high tumor stage in patients with OV (Both p < 0.05) ( Table 2).
In UTEA, the expression levels of both ERa and PGR proteins were inversely correlated with high tumor grade and stage (all p < 0.05); this trend was not found in ERb (both p > 0.05) ( Table 3). Further, the expression levels of ERb and PGR proteins were significantly correlated with patient age (both p < 0.05) ( Table 3).
Collectively, the expression of ERa, ERb, and PGR might be associated with the progression of OV and UTEA.

Genetic Alterations and Clinical Relevance of ESR1, ESR2, and PGR in Different Cancers
To inquiry genetic alterations of ESR1, ESR2, and PGR that may be associated with tumorigenesis, we analyzed these in pancancer involving a total of 10,189 patients. Genetic alterations (including amplification, fusion, deep deletion, missense mutation, and truncating mutation) were detected in 2.7%, 1.3%, and 3% of ESR1, ESR2, and PGR genes, respectively ( Figure 6A). Moreover, a mutual exclusivity analysis showed the selected genes tended toward co-occurrence rather than mutual exclusivity (p < 0.05) ( Figure 6B). Mutations in ESR1, ESR2, and PGR genes were the most frequent alterations in multiple cancer types, followed by amplifications and deep deletions ( Figure 6C). The patients were further divided into ESR1, ESR2 and PGR altered and unaltered groups to conduct the clinical attribute test ( Figure 7A). OncoTree Code was selected to indicate the ratio of cancer patients with/without genetic alterations in ESR1, ESR2, and PGR ( Figure 7B), and results suggested the potentially critical roles of ESR1, ESR2 and PGR in onset of multiple cancers. Together, ESR1, ESR2, and PGR are likely closely correlated and have a role in multiple tumor genesis.

Mutation Site Analysis of ESR1, ESR2, and PGR in Multiple Cancer Types
To identify mutation sites in the ESR1, ESR2, and PGR genes, we assessed 10,189 samples from multiple cancer types. The mutation sites were most commonly located within the Oest_recep, zf-C4, Hormone_recep, and ESR1_C domains ( Figure 8A). Specifically, 169 mutations of ESR1 were detected, consisting of 131 missense mutations, 19 truncating mutations, 3 inframe mutations, and 16 other types of mutations. Seven of these mutations were E247K/D, a hotspot for protein activation.  FIGURE 5 | Immunohistochemical analysis of ERa, ERb, and PGR protein expression levels in ovarian serous cystadenocarcinoma and uterine endometrioid adenocarcinoma. Representative immunohistochemical images showed that ERa, ERb, and PGR were differentially expressed in ovarian serous carcinoma and endometrioid adenocarcinoma tissue. Left panels, low expression levels of ERa, ERb, and PGR proteins in serous ovarian carcinoma. Right panels, high expression levels of ERa, ERb, and PGR proteins in endometrioid adenocarcinoma. All micrographs were captured at ×400 magnification. In addition, 112 ESR2 and 214 PGR nonsynonymous mutation sites were detected in different cancers, with the highest frequency mutations in R227H/C/L and R740Q/* (Figures 8B, C). Together, there are an abundance of mutation sites of ESR1, ESR2 and PGR, suggesting the complexity of their mutations.

Immunological Correlation Between ESR1, ESR2, and PGR and Immune Modulatory Factors Across Multiple Cancer Types
To assess the relevance of ESR1, ESR2, and PGR with immune system that plays critical roles in cancer progression (28), we first  compared their co-expression with the abundance of immunomodulators. As shown in Figures 9A-C, there is a positive correlation of the expression levels of ESR1, ESR2, and PGR genes with multiple immune-inhibitors (such as CD274, TIGIT, and CTLA4). Moreover, the expression levels of ESR1, ESR2, and PGR were observed to have a positive correlation with several immune-stimulators (such as CD27, CD28, and CXCL12) (Figures 9D-F). These findings suggest ESR1, ESR2 and PGR might be associate with both immune stimulation and inhibition. Together, the potential roles of ESR1, ESR2 and PGR in cancer are likely in an immune system-dependent manner.

Blood Cell Type-Specific Expression Profiles of ESR1, ESR2, and PGR Across Multiple Cancer Types
To further explore the correlation of ESR1, ESR2 and PGR with immunity, we analyzed their expression in peripheral blood cell types. ESR1 was found to be expressed in classical monocyte, MAIT T-cell, naive CD4 T cells, memory CD4 T cells, memory CD8 T-cell, naïve CD8 T cell, memory B cell and myeloid dendritic cells (DC) ( Figure 10A). Similarly, the expression of ESR2 was observed in multiple peripheral blood cells, with the highest expression level in plasma cell-like DC ( Figure 10B). In contrast, PGR expression was not observed in peripheral blood cells ( Figure 10C). Together, ESR1 and ESR2 might potentially influence multiple immune cells.

Contribution of ESR1, ESR2, and PGR to Survival Across Multiple Cancer Types
To investigate the potential roles of ESR1, ESR2, and PGR in prognosis, we analyzed the correlation of their expression with pan-cancer survival. As excepted, the expression of these genes was significantly associated with the OS and RFS in several cancers ( Figures 11A, B). For instance, the higher expression of ESR1 is significantly related with superior OS in head and neck squamous cell carcinoma (HNSC), kidney renal clear cell carcinoma (KIRC), liver hepatocellular carcinoma (LIHC) and skin cutaneous melanoma (SKCM). In contrast, the downregulation of ESR1 is suggestive of better OS in acute myeloid leukemia (LAML), brain lower grade glioma (LGG), lung squamous cell carcinoma (LUSC) and stomach adenocarcinoma (STAD). Together, these findings suggest that ESR1, ESR2 and PGR are potential prognostic factor in multiple cancers.

DISCUSSION
ESRs and PGR promote cell proliferation in breast cancer (29). Further, ESRs and PGR, which are associated with tumorigenesis and progression under pathological conditions, have become ideal molecular treatment targets (30)(31)(32)(33)(34). Accordingly, previous studies have demonstrated that drugs targeting ESR1, ESR2, and PGR are effective in the treatment of breast cancer and improve promote clinical outcomes (Supplementary Material Tables S1-S3). In addition, it has already been described the role of ESR1, ESR2, and PGR in promoting ovarian, lung, and prostate tumorigenesis (35)(36)(37). However, the roles of ESR1, ESR2, and PGR in other cancer types have rarely been studied and further investigations are needed to reach a consensus. In our study, we selected ESR1, ESR2, and PGR for an in-depth analysis of mRNA expression, genetic alternations, and clinical outcomes as well as the co-expression of these genes with immunomodulatory factors in a variety of cancer types. We also validated the expression levels of ERa, ERb, and PGR proteins in OV and UTEA using paraffin-embedded tissue specimens, and explored the relationship between ERa, ERb, and PGR proteins and clinicopathological characteristics of patients. To the best of our knowledge, this is the first such study based on integrated bioinformatics analysis. Through this comprehensive pan-cancer analysis, the feasibility of using ESR1, ESR2, and PGR as prognostic markers and therapeutic targets for multiple cancers was evaluated.
The results based on GEPIA2 dataset analysis were partly consistent with those reported previously (38)(39)(40). Hishida et al. showed that ESR1 gene transcripts were absent or decreased in more than 90% of liver cancer (n = 24) samples compared with their matched normal liver tissue counterparts. These results highlighted ESR1 as a tumor suppressor gene in liver cancer, and indicated that lower cellular estrogen levels stimulated liver cancer cell growth (41). The results of the present study revealed that the expression of ESR1 and PGR correlated with the tumor stage, whereas the expression of ESR2 did not. Additionally, prognostic analysis suggested that ESR1, ESR2, and PGR were significantly correlated with OS and RFS in patients with specific cancer types. However, due to the heterogeneity, subtypes, and sample size of cancers, or limited length of follow-up in TCGA datasets, there were discrepancies between the results of this study and those of previously published studies. For example, analysis using TCGA database showed that patients had shorter OS and RFS with higher ERb expression levels in renal cell carcinoma (42). This is in contrast to the results of the present study, which showed that ESR2 expression was not a risk factor for kidney cancer. Generally, the results of this study indicated that ESR1, ESR2, and PGR can be regarded as predictive and prognostic biomarkers across different cancer types.
The progression of cancer is influenced by multiple factors including, but not limited to, somatically acquired genetic, epigenetic, transcriptomic and proteomic alterations (43). Some alterations in particular genomic regions exhibit potential pro-and anti-tumor effects (44). Therefore, we used the cBioPortal web tool for further analysis of the genetic mutations in ESR1, ESR2, and PGR in multiple cancer types. Our results revealed genetic alterations in ESR1, ESR2, and PGR in multiple cancer types, including amplification, fusion, deep deletion, missense mutation, and truncating mutation. In addition, we identified a trend for co-occurrence of genetic alterations in ESR1, ESR2, and PGR. Based on these results, combining the expressions of ESR1, ESR2, and PGR may provide a better prognostic value in cancer patients. Yi et al. proposed that higher ESR1 expression and a higher ESR ratio (ESR1/ESR2) were associated with worse overall survival in female papillary thyroid carcinoma patients (45). There were also differences in  It has been reported that genomic diversity increases with the rate of genetic alterations result in cancers, resulting in an increased frequency of neoantigens and greater immune cell infiltration (46). Understanding the effects of immune cells on cancers will lead to a new era in oncotherapy. Therefore, the effectiveness and efficiency of immune checkpoint-target agents, which direct the host immune system to target cancer cells, has become a focus of research. However, results showed relatively low response rates of immune checkpoint-target agents in some tumors (47)(48)(49). To overcome this challenge, further understanding of immunotherapy is needed to select the patients who will benefit most from this type of therapy. Studies have found that immune checkpoint proteins (PD-L1, VISTA) are more frequently expressed in certain ESR-negative breast cancers (50,51). Liu et al. demonstrated an inverse correlation between ESRs and PD-L1 in breast cancer cells, indicating that PD-L1 gene transcription is negatively regulated by ESRs (52), which is consistent with the results of the current study. Hence, in the present study, we investigated the potential of ESR1, ESR2, and PGR as predictive and prognostic biomarkers in multiple cancer types from an immuno-oncological perspective based on bioinformatics analysis to provide a reference for future studies and the application of immunotherapies. In this study, we explored the relationship between ESR1, ESR2, and PGR and immunomodulators or immune cells using the TISIDB database. The results demonstrated that ESR1 had the greatest correlation with immunoinhibitors (such as CD274, CD96, CFS1R, and CTLA-4) and immunostimulators (such as CD27, CD28, and CXCL12). It is noteworthy that the role of ESR1 in the function of immunomodulators is context-dependent. ESR2 and PGR showed similar results. In addition, we found that there was a certain relationship between the expression levels of ESR1, ESR2, and PGR and peripheral blood cells. Therefore, this preliminary analysis of the association between ESR1, ESR2, and PGR and immune function highlights the importance of future research to elucidate the potential roles of ESR1, ESR2, and PGR as predictive and prognostic biomarkers as well as therapeutic targets for immunotherapy across multiple cancer types.
Our study also has some limitations. The results derived from different online databases are inevitably accompanied by background heterogeneity. Moreover, our immunohistochemical verification experiment was conducted only in OV and UTEA, with inadequate prognostic studies of patient cohorts. More cancer types need to be included in subsequent verification experiments, which can be further verified by adding cytological function studies and patient cohort studies.

CONCLUSIONS
In summary, we identified significant differences in the expression levels of ESR1, ESR2, and PGR mRNAs in different cancer types, which associated with tumor progression and clinical prognosis. Our study provides comprehensive evidence that ESR1, ESR2, and PGR are feasible prognostic markers and therapeutic targets for multiple cancers and that they could be a factor for disease prediction, disease evaluation, and individualized treatment in various types of cancer.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Anhui Medical University, Hefei 230032, Anhui, China. The patients/participants provided their written informed consent to participate in this study.