Over the past decade, FOXA1 expression has been examined in several human cancers, and oncogenic and tumor-suppressive roles have been proposed for FOXA1 depending on the cancer type and, in some cases, the subtype. In acute myelocytic leukemia, esophageal squamous cell carcinomas, lung adenocarcinomas, thyroid carcinoma, prostate cancer, and AR-positive molecular apocrine breast cancer [12, 23–25], FOXA1 acts as an oncogene. However, in hepatocellular carcinoma, pancreatic, and estrogen receptor (ER)-positive breast cancer, FOXA1 has been reported to have a tumor-suppressive function [26–28]. On one hand, FOXA1 acts as a tumor oncogene. In oesophageal squamous cell carcinoma, FOXA1 expression is correlated with lymph node metastases in immunohistochemical specimens and FOXA1 expression inhibition decreases cellular invasion and migration . Also, FOXA1 is over-expressed in aggressive thyroid cancers (ATC) and involved in cell cycle progression via down-regulation of p27Kip1 in an ATC cell line . On the other hand, FOXA1 has been reported to act as a tumor suppressor. It has been reported that FOXA1 positively regulates miRNA-122, which is correlated with favourable prognosis in human hepatocellular carcinoma . In addition, FOXA1 acts as an important antagonist of the epithelial-to-mesenchymal transition (EMT) in pancreatic ductal adenocarcinoma through its positive regulation of E-cadherin and maintenance of the epithelial phenotype . It is critical to note that the role of FOXA1, as a tumor oncogene or a tumor suppressor gene, has been reported to vary in prostate and breast cancers depending on multiple cancer subtypes and states of hormone dependence or independence [11, 12, 28].
A previous study has addressed the expression and function of FOXA1 in EC; immunohistochemical analysis by Abe et al. indicated that FOXA1 was negatively associated with lymph node status in EC immunohistochemical specimens in Japanese, and FOXA1 repressed proliferation and migration in one type of EC cells (Ishikawa) . However, our study found that the FOXA1 level in ECs was significantly higher than that in atypical hyperplasia and normal tissues (p < 0.05) in immunohistochemical specimens and that FOXA1 promoted tumor cell proliferation in EC, which differs from the previous results. The difference might be attributed to the immunohistochemical samples in different countries used. Alternatively, the cancer subtype may affect the results: the function of FOXA1 as a tumor suppressor in the Abe et al. study was investigated in the Ishikawa cell line, which is ER-positive , whereas we used MFE-296 (high levels of FOXA1 and AR) and AN3CA (low levels of FOXA1 and AR), which are both ER-negative cell lines [33, 34]. This idea consists with breast cancer studies that have shown that FOXA1 functions as a tumor suppressor in ER-positive breast cancer cells (MCF-7)  but as a tumor activator in ER-negative breast cancer cells (MDA-MB-453) . Furthermore, this idea of the effects of forkhead family members depending on ER expression is also consistent with the study that have shown the Forkhead box class o 3a transcription factor (FoxO3a) has inhibitory effects on motility and invasiveness of ER-positive breast cancer cells but inducing effects on motility and invasiveness of ER-negative breast cancer cells . More comprehensive studies covering several EC cell lines in different cancer subtypes will be necessary to define the role of FOXA1 in EC development.
Most researches on hormone receptors in EC have focused on ER and progesterone receptor (PR). However, the expression of AR in the human normal endometrium and its disorders is not well understood. Though higher serum androgen levels have been certified to exist in the utero-ovarian vein blood samples from women with EC , the details of AR expression and its actions in EC are a topic of dispute. Longer CAG repeats in AR promote carcinogenesis of uterine endometrial cells . Androgens and AR may be involved in endometrial cell proliferation by regulating the expression of insulin growth factor I (IGF-I) in the uterus . Our results suggest that AR expression is significantly higher in EC than in normal endometrium and that AR activated by FOXA1 might promote the Notch pathway, which may be another mechanism involving AR in EC.
Most FOXA1 studies have focused on its role as a pioneer factor that binds to DNA packaged in chromatin and opens the chromatin for binding of additional transcription factors including AR [39, 40]. According to our results from qRT-PCR and western blotting, FOXA1 regulates AR target genes by up-regulation of AR expression. Interestingly, our co-immunoprecipitation results (Figure 4A and 4B) showed that FOXA1 interacted with AR at the protein level. Apart from that, our ChIP-PCR results suggested that FOXA1 and AR were directly bound to the same regions upstream of MYC (Figure 4C and 4D). Based on the above results, we suggest that FOXA1 may also directly regulate AR target genes (at least MYC) by binding to AR in EC. Our results regarding an interaction between AR and FOXA1 may be related to the finding that the AR and FOXA1 binding sites are adjacent on multiple promoters of AR target genes in prostatic cells [9, 41]. Thus, FOXA1 may regulate the AR target genes through at least two means: AR over-expression or physical interaction with AR in order to induce easy AR accessibility to binding to its target genes. MYC is an immediate early response gene downstream from AR pathway and is tightly regulated through AR cis-regulatory elements identified within its proximal promoters and distal enhancer regions , which is consistent with our ChIP-PCR results (Figure 4C and 4D). Interestingly, we showed that FOXA1 and AR more evidently bound to the MYC enhancer regions as compared to MYC promoter regions. These results could be attributed to other co-regulators involved in this binding process. Since TCF7L2, a protein mediating DNA looping for long-distance interactions of distal enhancers and proximal promoters, physically interacts with FOXA1 and AR and mediates the transcription of MYC in breast cancer , future investigation will be needed to clarify which co-regulators are involved in FOXA1/AR binding to the enhancer regions upstream of MYC in EC cells.
Although the underlying mechanisms governing the FOXA1-AR correlation in tumor progression are not fully understood, a pathway analysis showed that 187 FOXA1/AR dual target genes were involved in the cellular growth/proliferation pathway in liver cancer . The Notch pathway is implicated in the development of various cancers, and the Notch pathway blockade appears to affect cell proliferation in multiple types of cancers. Notch pathway inhibition in breast cancer cells induces cell cycle arrest and apoptosis . Similarly, downregulation of Notch1 contributes to cell growth inhibition in pancreatic cancer . Our results suggest that downregulation of AR attenuated FOXA1-induced upregulation of the Notch pathway in EC cells. These findings indicate that FOXA1 might promote AR-mediated transcription and ultimately activate the Notch pathway. Here, we describe, for the first time, the association between FOXA1 expression and the Notch pathway in cancer.
The specific mechanism of cell proliferation in EC reported so far has been limited, although several classical transcription factors related to proliferation have been identified, including cyclin D1, p53, IGFBP-1, PTEN, and p27Kip1[44–48]. In this study, we suggest that FOXA1 promotes cell proliferation in EC by interaction with AR, possibly via the Notch pathway, which may be a newly identified regulatory mechanism of cell proliferation in EC.
We further investigated the effects of FOXA1 and AR on migration and invasion of EC cells, and found that neutralization of AR activity did not inhibit FOXA1-enhanced cancer cell migration or invasion. These observations indicate that the promoting effect of FOXA1 on migration and invasion is not dependent on AR. Our findings in migration and invasion assays are consistent with our findings in immunohistochemical staining, which showed that higher expression of FOXA1 but not AR is found in tumors that displayed a greater depth of myometrial invasion. These results suggest that AR is not the only downstream target of FOXA1 in EC. Future studies will be necessary to define which transcription factors or pathways are involved in FOXA1-enhanced cell migration and invasion in EC.
The traditional endocrine treatment (mainly targeting ER and PR) is ineffective in most ER-negative and PR-negative ECs, and even in some ER-positive and PR-positive ECs . In our investigation, 9 of the15 ER-negative EC cases (60.0%) and 41 of the 61 ER-positive EC cases (67.2%) were AR positive, and the majority of ECs were also FOXA1 positive (Table 1). Thus, AR and FOXA1 might be alternative targets in ECs insensitive to traditional endocrine treatment or could be targets for adjuvant treatment following surgery and traditional endocrine treatment. There has been speculation about the use of anti-androgens for the treatment of ECs ; this hypothesis warrants clinical investigation in light of our findings.