FOXA1 Leads to Aberrant Expression of SIX4 Affecting Cervical Cancer Cell Growth and Chemoresistance

Cervical cancer (CC) is among the most prevalent cancers among female populations with high recurrence rates all over the world. Cisplatin (DDP) is the first-line treatment for multiple cancers, including CC. The main problem associated with its clinical application is drug resistance. This study is aimed at investigating the function and downstream regulation mechanism of forkhead-box A1 (FOXA1) in CC, which was verified as an oncogene in several cancers. Using GEO database and bioinformatics analysis, we identified FOXA1 as a possible oncogene in CC. Silencing of FOXA1 inhibited CC cell growth, invasion, and chemoresistance. Afterwards, the downstream gene of FOXA1 was predicted using a bioinformatics website and validated using ChIP and dual-luciferase assays. SIX4, a possible target of FOXA1, promoted CC cell malignant aggressiveness and chemoresistance. In addition, overexpression of SIX4 promoted phosphorylation of PI3K and AKT proteins and activated the PI3K/AKT signaling pathway. Further overexpression of SIX4 reversed the repressive effects of FOXA1 knockdown on CC cell growth, invasion, and chemoresistance in DDP-resistant cells. FOXA1-induced SIX4 facilitates CC progression and chemoresistance, highlighting a strong potential for FOXA1 to serve as a promising therapeutic target in CC.


Introduction
Despite as one of the most preventable malignancies through screening, cervical cancer (CC) claimed the lives of 4138 women in 2018 in the United States, and one-half of whom were aged ≤58 years at death [1]. Treatment for CC patients depends on the disease severity at diagnosis and locally available resources and may include radical hysterectomy or chemoradiation or a combination of both [2]. The chemotherapeutic agent cisplatin (DDP), a smallmolecule platinum compound that was firstly identified to inhibit bacterial growth and later identified as an anticancer drug, has been applied to effectively treat advanced/recurrent CC [3]. The acquirement of DDP resistance in human cancer cells in vivo and in vitro stems from complex genetic and epigenetic alterations in gene expression [4].
Forkhead-box A1 (FOXA1) is expressed by preinvasion lesions of the uterine cervix and almost half invasive CC, supporting its implication in human papilloma virus pathogenesis [5]. Moreover, FOXA1 knockdown in DDP-resistant nasopharyngeal carcinoma cells increased the chemosensitivity to DDP [6]. Therefore, we wondered there is also an association between FOXA1 overexpression and DDP resistance in CC cells. Interestingly, FOXA1 was confirmed to induce the expression of circRNA derived from oxysterol binding protein-like 10 in CC cells, thereby regulating CC cell proliferation and migration [7]. As a consequence, we believed that the possible effects of FOXA1 on DDP resistance in CC are elicited through its transcription factor role. The sine oculis homeobox (SIX) family homeobox genes have been displayed to participate in the tumor initiation and progression [8]. The functions of SIX genes are predominantly based on their structure and their regulatory roles in response to external or internal stimuli, which suggests that members of the SIX superfamily can be considered as new target molecules to inhibit tumor growth and progression [9]. For one of them, SIX1 was induced by the E7 oncoprotein of human papillomaviruses in CC, and increased SIX1 expression contributed to the upregulation of genes related to the initiation of DNA replication [10]. In the present study, SIX4 was identified as a downstream target of FOXA1 using a bioinformatics tool hTFtarget (http://bioinfo.life .hust.edu.cn/hTFtarget#!/). SIX4 has been lately revealed to indicate dismal clinical outcome of patients with esophageal squamous cell carcinoma and to potentiate tumor growth and cell metastasis [11]. However, its role in CC, especially in regulating chemoresistance, remains largely unclear. Hence, our study investigated the potential roles of FOXA1 in development of CC and its functions in connection with SIX4.

Complementary DNA (cDNA) Synthesis and RT-qPCR.
Total RNA was isolated from cells using TRIzol (15596018, Thermo Fisher), and the concentration and quality of total RNA were measured using a Nanodrop spectrophotometer (ND-LITE-PR, Thermo Fisher). For mRNA expression analysis, cDNA was synthesized using PrimeScript RT master mix (RR036B, Takara Holdings Inc., Kyoto, Japan). Realtime PCR was performed using SYBR Select Master Mix (4472919, Thermo Fisher Scientific) on a CFX96 real-time PCR detection system (Bio-Rad, Inc., Hercules, CA, USA). Relative expression was calculated utilizing 2 −ΔΔCt method with normalization to glyceraldehyde-3-phosphate dehydro-genase (GAPDH). The sequences of primers are presented in Table 1.

Colony Formation Assay.
HT-3/DDP and HeLa/DDP cell suspensions were seeded into 6-well plates at approximately 5 × 10 2 cells/well. Subsequently, 2 mL RPMI-1640 medium was supplemented to each well, and the solution was changed at an interval of 2 d. After 10 d, the cells were fixed with formaldehyde and treated with crystal violet, and the number of colonies formed in each well of the plate was counted.
2.6. 5-Ethynyl-2′-Deoxyuridine (EdU) Assay. The EdU Staining Proliferation kit (ab222421, Abcam) was applied to assess the proliferation of HT-3/DDP and HeLa/DDP cells, and all operations were performed strictly as per the instructions. Stably transfected cells (5 × 10 3 cells/well) were seeded into 96-well culture dishes and grown for 24 h. Then, 100 μL medium containing 20 μM EdU was supplemented into each well, and the cells were incubated at 37°C for 2 h. The 2.9. Dual-Luciferase Assay. A potential binding site for the SIX4 promoter with a conserved FOXA1 binding sequence was obtained from Jaspar (http://jaspar.genereg.net/), and the binding region was PCR amplified and cloned into the pGL3 vector (Promega, Madison, WI, USA) to obtain a Promoter luciferase reporter vector. The above reporter vectors were cotransfected with sh-NC or sh-FOXA1 into HT-3/ DDP and HeLa/DDP cells, respectively, using Lipofectamine 3000 (Invitrogen). After 48 h of transfection, luciferase activity was tested by a dual-luciferase reporter system (Promega).

Chromatin Immunoprecipitation (ChIP).
ChIP was conducted using the EZ-Magna ChIP kit (17-10461, Sigma-Aldrich). Briefly, HT-3 or HeLa cells were crosslinked by using a 1% formaldehyde solution for 10 min at room temperature and then quenched by glycine. DNA fragments were obtained by ultrasound treatment. The lysate was immunoprecipitated with anti-FOXA1 (1 : 50, ab170933, Abcam) or IgG (1 : 100, ab172730, Abcam)coupled magnetic beads. Finally, the enrichment ability of the precipitates for the SIX4 promoter was analyzed by PCR.
2.11. Xenograft Tumor Model. The whole procedures of experiments were approved by the committee of Jilin Cancer Hospital. Twelve BALB/c nude mice (female, 4-6 weeks, 21:2 ± 2:3 g, Vital River, Beijing, China) were housed and cared according to the institutional guidelines for animal care. The mice were randomly divided into 2 groups (sh-NC and sh-FOXA1). A 12-12 h lights on-off cycle was maintained in a specific-pathogen-free environmentally controlled room (25°C, 45% humidity). The mice were given adequate food and water. HeLa/DDP cells transfected with sh-NC or sh-FOXA1 suspended in serum-free RPMI-1640 medium were injected subcutaneously into nude mice after 1 week of acclimation feeding. After that, the tumor volume of the mice was measured every 7 days. Tumor volume = 0:5 × length × width 2 . After 4 weeks, the mice were euthanized by intraperitoneal injection of 1% pentobarbital sodium, and tumors were collected for weighing.
2.12. Immunohistochemistry. The tumor tissues of mice were fixed in formalin for 4 h, routinely embedded in paraffin, and sectioned (5 μm). The dewaxed and hydrated tissues were fixed on slides and incubated for 15 minutes at room temperature with 3% H 2 O 2 , followed by the incubation with normal goat serum (SL038, Solarbio) for 15 min at room temperature. The sections were stained with primary antibody KI67 (1 : 100, 14-5699-82, Thermo Fisher scientific) overnight at 4°C and with secondary antibody (1 : 2000, ab205719, Abcam) for 15 min at 37°C. Afterwards, 40 μL HRP-labeled streptavidin working solution (SE068, Solarbio) was added dropwise and incubated for 15 min. After a 30 s hematoxylin counter-staining, dehydration, and fixation, the number of KI67-positive cells in the sections was counted under a light microscope (Olympus), and five areas were randomly selected for each section separately.
2.13. Statistical Analysis. The SPSS 22.0 statistic software (SPSS, Inc, Chicago, IL, USA) was used for statistical analyses. The data were presented as mean ± SD. Unpaired t-tests were applied to describe the differences between different groups. Data among multiple groups were compared by one-way or two-way ANOVA followed by a Tukey's post hoc test. We considered p < 0:05 as significant in all cases.

FOXA1 Is Highly Expressed in CC and Is Associated with
Drug Resistance. First, we found the gene expression profiles GSE64217 (incisional biopsies from cervical squamous cell carcinoma and normal cervical mucosa) and GSE63514 (24 normal and 28 cancer specimens) regarding CC were downloaded from the GEO database (https://www.ncbi.nlm.nih .gov/geo/). We screened 179 and 192 differentially expressed transcription factors with p < 0:05 and log 2FC > 1 or log 2 FC > 2 as screening conditions, respectively. Among the intersecting genes, we screened out that FOXA1 expression was closely associated with CC (Figure 1(a)). By querying the GEPIA database (http://gepia.cancer-pku.cn/), we found that FOXA1 was highly expressed in various cancers (Figure 1(b)) and that FOXA1 was significantly differentially expressed in CC tissues versus normal cervical tissues (Figure 1(c)). Similarly, analysis of TCGA database on FOXA1 expression in CC through the UALCAN website 3 Analytical Cellular Pathology    revealed that FOXA1 was significantly highly expressed in CC (Figure 1(d)). Consistently, FOXA1 expression was much higher in CC cell lines (HT-3, CaSki, HeLa, and SiHa) than that in normal cervical epithelial cells H8 (Figures 1(e) and 1(f)). Subsequently, we selected HT-3 and HeLa cells with significant differences of FOXA1 expression to induce DDP-resistant HT-3 cell lines (HT-3/DDP) and HeLa cell lines (HeLa/DDP). The successful development of drugresistant cell lines was validated using the MTT assay (Figure 1(g)). The results of RT-qPCR and Western blot revealed that FOXA1 was upregulated in drug-resistant cells relative to parental cells (Figures 1(h) and 1(i)). The above results suggest that FOXA1 is highly expressed in CC and is associated with drug resistance.

Silencing of FOXA1 Inhibits CC Cell Growth and
Chemoresistance. To define the effect of FOXA1 on the growth and drug resistance of HT-3 and HeLa cells, we transfected sh-NC or sh-FOXA1 into HT-3/DDP and HeLa/DDP cells, and the expression of FOXA1 in the cells was detected 48 h posttransfection to confirm the successful transfection (Figure 2(a)). Changes in the activity of drugresistant cells were detected by colony formation assay and EdU staining, and it was found that the activity of drugresistant cells was significantly inhibited after knocking down FOXA1 (Figures 2(b) and 2(c)). According to the Transwell assay, we further identified that the invasive ability of drug-resistant cells was significantly reduced after silencing of FOXA1 (Figure 2(d)). MTT assay results showed that chemoresistance to DDP decreased more significantly in the sh-FOXA1 group with increasing concentrations of DDP (Figure 2(e)). Subsequently, in in vivo experiments, we selected HeLa/DDP cells with stable low expression of FOXA1 for tumor xenograft experiments. It was demonstrated that loss of FOXA1 significantly hindered the growth

Knockdown of FOXA1 Downregulates the Expression of SIX4 in CC Cells.
To explore the pathways through which FOXA1 acts on CC, the hTFtarget database was queried. Among all the downstream genes of FOXA1, we found that SIX4 had the largest log2FC value (log 2FC = 7:11) in the GSE63514 database, so we selected SIX4 as the downstream target of FOXA1 in this study (Figure 3(a)). SIX4 was significantly highly expressed in CC as queried by TCGA database (Figure 3(b)). We then obtained the conserved binding sequence of FOXA1 in Jaspar (Figure 3(c)). The promoter sequence of SIX4 was also predicted to have a potential binding site to FOXA1 (Figure 3(d)). The GEPIA database predicted the overexpression of SIX4 in CC (Figure 3(e)). RT-qPCR result was consistent with the prediction (Figure 3(f)). The SIX4 expression in HT-3/DDP and HeLa/DDP cells with FOXA1 knockdown was detected by Western blot, and its protein expression was found to be significantly constrained (Figure 3(g)). We constructed the Promoter luciferase reporter vector by inserting the SIX4 promoter sequence with the highest binding score of poten-tial to FOXA1 into the pGL3 luciferase vector. The vectors were cotransfected with sh-FOXA1 or its control into HT-3/DDP and HeLa/DDP cells, respectively. The luciferase activity was measured after 48 h. Inhibition of FOXA1 was found to significantly suppress the luciferase activity of Promoter (Figure 3(h)). In ChIP assays, we observed that anti-FOXA1 significantly enriched the promoter sequence of SIX4 compared to IgG (Figure 3(i)). The above experiments demonstrated that FOXA1 can bind to the promoter sequence of SIX4 and affect its expression.

Overexpression of SIX4 Activates the PI3K/AKT Signaling Pathway in CC.
To further determine the effect of SIX4 on the growth and drug resistance of CC cells, we overexpressed SIX4 in drug-resistant cells HT-3/DDP and HeLa/DDP and found that its mRNA expression was significantly elevated after overexpression of SIX4 using RT-qPCR (Figure 4(a)). Colony formation assays with EdU staining revealed that the activity of drugresistant cells was significantly higher after overexpression of SIX4 (Figures 4(b) and 4(c)). The results of the Transwell assay showed a significant higher invasive capacity of drug-resistant cells in the oe-SIX4 group (Figure 4(d)). The chemoresistance of resistant cells to DDP was

10
Analytical Cellular Pathology  on the Promoter luciferase activity of SIX4 measured using dual-luciferase assay. (i) The ability of anti-FOXA1 to enrich SIX4 promoter sequence examined using ChIP assay. Data among multiple groups were analyzed by one-way or two-way ANOVA followed by Tukey's post hoc test ( * p < 0:05, * * p < 0:01, * * * p < 0:001). The cell experiment was independently repeated three times.  13 Analytical Cellular Pathology significantly enhanced after overexpression of SIX4, as evidenced by MTT assay (Figure 4(e)). It has been reported that SIX4 expedited metastasis in colorectal cancer by activating the PI3K/AKT signaling pathway [12]. We performed Western blot assays to measure the PI3K/AKT signaling pathway activation in CC cells overexpressing SIX4 cells. It showed that the extent of PI3K and AKT phosphorylation was significantly increased after promoting the expression of SIX4 (Figure 4(f)). The above results indicate that SIX4 can activate PI3K/ AKT signaling, thus promoting CC cell growth and enhancing chemotherapy resistance.

Overexpression of SIX4 Partially Reverses the Repressive
Effects of sh-FOXA1 on CC Cells. We overexpressed SIX4 in HT-3/DDP and HeLa/DDP drug-resistant cells with FOXA1 knockdown for rescue experiments. The successful transfection was confirmed using RT-qPCR ( Figure 5(a)). We found that further overexpression of SIX4 in drugresistant cells with FOXA1 knockdown significantly increased the activity of drug-resistant cells by colony formation assay and EdU staining (Figures 5(b) and 5(c)). Transwell invasion assay results showed that overexpression of SIX4 reversed the inhibitory effect of sh-FOXA1 on the ability of drug-resistant cells to invade ( Figure 5(d)). We examined the changes in IC50 values of drug-resistant cells to DDP using MTT and observed that overexpression of SIX4 significantly enhanced the chemoresistance of drugresistant cells to DDP in the presence of sh-FOXA1 ( Figure 5(e)). Western blot assay showed that the protein expression of p-PI3K and p-AKT was elevated, and the PI3K/AKT signaling pathway was activated after overexpression of SIX4 ( Figure 5(f)). Taken together, overexpression of SIX4 partially reversed the repressing effect of sh-FOXA1 on CC cell resistance to chemoresistance by activating the PI3K/AKT signaling pathway.

Discussion
DDP has been considered as the primary chemotherapy for locally advanced CC or recurrent cancers, while it should be taken into consideration that the majority of patients treated for persistent and/or recurrent CC represent previous chemo-or chemoradiation failures [13]. Thus, for some patients, the establishment of certain molecular mechanisms impairing the response to DDP might have a significant role in improving poor prognosis. DDP exerts anticancer activity via manifold mechanisms, but its most conventional mechanism remains production of DNA lesions by interacting with purine bases on DNA, followed by activation of several signal transduction pathways [14]. Functional experiments in our study revealed that FOXA1 promoted CC cell resistance to DDP in vitro and facilitated tumor growth in vivo by upregulating SIX4 expression and activating the PI3K/AKT signaling.
Initially, FOXA1 was characterized to be elevated in CC tissues and cells using data mining, RT-qPCR, and Western blot. Furthermore, silencing of FOXA1 using shRNA inhibited CC cell proliferation, colony formation, and invasion. FOXA1 expression was also significantly higher in colorectal cancer tissues derived from TCGA datasets and was linked to worse prognosis in the R2 database [15]. Hight et al. revealed that reduction of FOXA1 abolished the ability of non-small-cell lung cancer cells to form clonogenicity in vitro (about 4-5 folds of decrease) and tumors in mice [16]. More relevantly, FOXA1 has been reported to inhibit apoptosis and cause abnormal cell proliferation in CC [17]. In addition, increased FOXA1 expression has been recently identified to promote cancer chemotherapy resistance in mice and patients with prostate cancer and breast cancer [18]. FOXA1 upregulation induced genome-wide enhancer reprogramming and activated the HIF-2α expression and a prometastatic transcriptional program in endocrineresistant breast cancer [19]. Upon FOXA1 knockdown in    [20]. Meanwhile, Labbé and Brown identified the association between FOXA1 and castration-resistant prostate cancer [21]. FOXA1 contributed to acquisition of chemoresistance in human lung adenocarcinoma via transactivation of SOX5 [22]. In the present study, we presented evidence showing the sensitizer role of sh-FOXA1 on CC cells to DDP. FOXA1 has been documented to bind to the promoter of KDM6A and PLOD2, thus promoting their transcription in bladder cancer and non-small-cell lung cancer, respectively [23,24]. Similarly, in CC, FOXA1 was verified to activate the transcription of LINC00662 and PDK4 [25]. Therefore, we postulated that the regulatory effects of FOXA1 on CC resistance to DDP were elicited through its transcription factor role. As expected, SIX4 was revealed to be a downstream candidate of FOXA1 in CC. SIX4 has been identified as a target of multiple microRNAs in different cancers, including gastric cancer, glioblastoma multiforme, and bladder cancer [26][27][28]. Intriguingly, Camolotto et al. reported that hepatocyte nuclear factor 4α directly repressed SIX4, and SIX4 drove proliferation and differentiation in hepatocyte nuclear factor 4α-negative pancreatic ductal adenocarcinoma cells [29]. Here, we revealed that the expression of SIX4 was governed by FOXA1. Moreover, overexpression of SIX4 not only expedited the chemoresistance of CC cells but also reversed the chemosensitivity promoting effects of sh-FOXA1. The tumor-supporting effects of SIX4 were also substantiated in hepatocellular carcinoma, non-small-cell lung cancer, and breast cancer [30][31][32]. However, its involvement in CC and in chemoresistance is scarcely understood, which highlights the novelty of our study. It has been reported that SIX4 expedited metastasis in colorectal cancer by activating the PI3K/AKT signaling pathway [12]. We similarly found that the extent of PI3K and AKT phosphorylation, which was reduced by silencing of FOXA1, was activated by SIX4 upregulation in CC cells. The interaction between FOXA1 and the PI3K/AKT pathway has also been underscored in hormone sensitivity of breast cancer [33].

Conclusion
All in all, we identified a novel molecular and functional network in CC cell resistance to DDP that coordinates FOXA1, SIX4, and the PI3K/AKT pathway. The result hinted that FOXA1 knockdown might represent a potential candidate for overcoming CC resistance to DDP. Thus, future work focusing on precise mechanisms that underlie the chemoresistance of DDP could be useful in guiding clinical therapeutic decisions. Besides, the validation of this axis on patient samples might be of great clinical value to develop efficient therapeutic strategy for CC.

Data Availability
All the data generated or analyzed during this study are included in this published article.