NF‐YA promotes the cell proliferation and tumorigenic properties by transcriptional activation of SOX2 in cervical cancer

Abstract NF‐YA is considered as a crucial regulator for the maintenance of cancer stem cell (CSC) and involved in various types of malignant tumours. However, the exact function and molecular mechanisms of NF‐YA in the progression of cervical cancer remains poorly understood. Here, the expression of NF‐YA detected by immunohistochemistry was gradually increased from normal cervical tissues, to the high‐grade squamous intraepithelial lesions, and then to cervical cancer tissues. NF‐YA promoted the cell proliferation and tumorigenic properties of cervical cancer cells as well as tumorsphere formation and chemoresistance in vitro. The luciferase reporter assay combined with mutagenesis analyses and Western blotting showed that NF‐YA trans‐activated the expression of SOX2 in cervical cancer. Furthermore, quantitative chromatin immunoprecipitation (qChIP) and electrophoretic mobility shift assay (EMSA) confirmed that NF‐YA protein directly bound to the CCAAT box region located upstream of the SOX2 promoter. Together, our data demonstrated that NF‐YA was highly expressed in cervical cancer and promoted the cell proliferation, tumorigenicity and CSC characteristic by trans‐activating the expression of SOX2.

others. 12 However, the surface markers are of instability and scarcity to isolate CSC in solid tumours. Thus, other markers, such as nuclear transcription factors, 13 side populations, 14 sphere formation and aldehyde dehydrogenase (ALDH) activity, 15 have been widely explored. In our previous study, the cervical CSC has been isolated and identified by the nuclear transcription factor SOX2 16 and cytoplasm ALDH1. 17 However, the exact mechanism and key genes involved in the maintenance and regulation of cervical CSC have not been clearly revealed.
NF-Y (nuclear transcription factor-Y, also known as CBF, CCAATbinding factor), a heterodimeric protein complex, is a ubiquitously expressed trimetric transcription factor comprising the subunits (NF-YA, NF-YB and NF-YC). 18 NF-YA is required for the complex assembly and sequence-specific DNA binding to CCAAT box. It is reported that NF-YA functions as an oncogene or suppressor through the mechanism of cell proliferation, metastasis and tumorigenicity. [19][20][21][22] The recent study has identified NF-YA as a CSC marker in hepatocellular cancer, oral cancer and embryonic cancer. 20,[23][24][25] It was shown that NF, as a transcription factor, regulates the expression of several human SOX genes, including SOX3, 26 SOX9 27 and SOX18, 28 by direct binding to CCAAT boxes within promoters of target genes and by making complex interplay with other factors involved in transcription regulation of human SOX genes.
Here, our study demonstrated that NF-YA was up-regulated during the progression of cervical cancer, which was essential for the promotion of cell proliferation, tumorigenicity and stemness properties by transcriptional activation of SOX2 protein in cervical cancer.
The overall score of ≤ 3 was defined as negative, >3 but ≤ 6 as weak positive, and> 6 as strong positive. Two different pathologists evaluated all the specimens in a blinded manner.
The expression of NF-YA protein was also detected by immunocytochemistry experiment in cervical cancer cells. In brief, cells cultured on coverslips were fixed with 4% paraformaldehyde for 30 minutes, permeabilized with 0.2% Triton X-100 for 15 minutes at room temperature and then incubated with the primary antibodies described above.

| Cervical cancer cell lines
The human cervical cancer cell lines, including SiHa, HeLa, C33A, CaSki and HT-3, were obtained from the American Type Culture Collection (ATCC). SiHa, HeLa and C33A cells were cultured in DMEM (Dulbecco's modified Eagle medium-high glucose, Sigma-Aldrich), CaSki cells were cultured in RPMI1640 (Sigma-Aldrich), and HT-3 cells were cultured in McCoy's 5A Medium 10,27 all supplemented with 10% FBS (foetal bovine serum, Invitrogen, Carlsbad, CA). All the cells were maintained at 37°C in an atmosphere containing 5% carbon dioxide.
Protein concentrations were determined using the bicinchoninic acid assay kit (Pierce Chemical Corporation). A total of 30μg protein was performed as previously described by Western blotting assays. 10 The primary antibodies mouse anti-human NF-YA (1:1000, #sc-17753, Santa Cruz), goat polyclonal anti-human SOX2 (1:500, #sc-17320, Santa Cruz) and mouse anti-human GAPDH (1:1000, #sc-32233, Santa Cruz) were performed at 4℃ overnight. A horseradish peroxidase-conjugated anti-mouse or anti-goat IgG (Thermo Fisher Scientific) as the secondary antibodies was incubated at room temperature for 1 hour. The signals were then detected by enhanced chemiluminescence reagent (Millipore).

| Cell growth and viability assays
Cell growth assay was performed for 7 days. In brief, 2 × 10 4 cells were cultured in 35-mm culture dishes, harvested every day and then the living cells dyed by trypan were counted using a haemocytometer under the light microscope.
Additionally, cell viability assay was assessed according to a standard protocol using MTT dye (3-(4,5-dimethylthiazole-yl)-2,5-diphenyltetrazolium bromide, Sigma-Aldrich). Following the manufacturer's instructions, 1 × 10 3 cells were planted in 96-well plate and 20 µL of MTT solution with the concentration of 5 mg/mL was added to 200 µL of the culture media every day for one week.
The plates were then incubated for 4 hours at 37°C, and the optical density was measured at 490 nm.

| The detection of chemoresistance
For the chemotherapy drug resistance assays, cells were cultured in 96-well plates at a density of 10 4 cells/well and allowed to recover overnight before initiating drug treatments. The cell viability with MTT assay was measured when the cells were exposed in various concentrations of cisplatin (0, 3, 6, 12 or 24 µg/mL) and 5-Fu (5-Fluorouracil, 50, 100 or 200 µg/mL) for 24 hours. Then, the value of IC50 for drugs was calculated. Additionally, the cells were exposed to a constant concentration of cisplatin (6 µg/mL) and 5-Fu (100 µg/ mL) for 24, 48 or 72 hours, and the cell viability was measured.
Also, cells (1 × 10 5 ) cultured in 35-mm dish were treated with cisplatin (6 µg/mL) and 5-Fu (5-Fluorouracil, 100 µg/mL) for 72 hours, and then, the cell viability was determined by Giemsa staining. For Giemsa staining, briefly, cells were washed with PBS and fixed with formaldehyde for 30 minutes and washed again before incubation with Giemsa staining solution (Sigma). After 30 minutes of staining, cells were washed and allowed to dry. The viability cells were counted in 10 random high-power field.

| Tumorsphere formation assay
Cells with the density of 200 cells per well in 24-well ultra-low attachment plates or the density of 1 cell per well in 96-well plates in triplicate were maintained in the serum-free medium with DMEM/ F12, containing N2 and B27 supplements (Invitrogen), 20 ng/mL human recombinant epidermal growth factor (EGF) and 20 ng/mL basic fibroblastic growth factor (bFGF; PeproTech Inc., Rocky Hill, NJ). The number of tumorspheres generated within 2 weeks was counted and calculate the percentage of sphere-forming. For serial tumorsphere formation assays, the spheres were harvested, disaggregated with 0.25% trypsin/EDTA, filtered through a 40-μm mesh and re-plated as described above.

| SOX2 promoter reporters
To characterize the transcriptional effects of mutations in the  Table S1.

| Chromatin immunoprecipitation (ChIP)
ChIP assay was performed according to the manufacturer's proto-

| TCGA data acquisition
RNAseq data were acquired using TCGA (The Cancer Genome Atlas) database by cervical squamous cell CESC patients (N = 306) matched with the TCGA normal data (N = 13). According to TCGA publication guidelines, these mRNA sequencing data have no restrictions on publication, and no additional approval by an ethics committee was required (http://cance rgeno me.nih.gov/publi catio ns/publi catio nguid elines).

| Statistical analysis
Statistical analyses were performed based on the software GraphPad Prism 5.01. In detail, when compared between two groups, twotailed Student's t test was applied. To examine differences among 3 groups, an ANOVA was performed. A P value of < .05 was regarded as statistically significant.

| NF-YA-positive cervical cancer cells shared the higher tumorsphere formation and cell growth in vitro
Previous studies showed that NF-YA protein was involved in the maintenance of stemness of stem cells by activating multiple other stem cell-related genes. We focused on the role of NF-YA in driving cervical CSC characteristics. To assess the ability of self-renewal, a critical characteristic of CSC in vitro, tumorsphere formation ability was valued with cells cultured in the serum-free medium. As shown in Figure 3A

| NF-YA-positive cervical cancer cells shared ability of drug resistance
The resistance of CSC to current chemotherapeutics is thought to be responsible for cancer recurrence and metastasis. Cells were exposed to the cisplatin, one of the most commonly chemotherapeutic  Figure 4C and D P < 0.05). Additionally, the number of viability cells was counted in 10 random high-power fields by Giemsa staining after exposure to a constant concentration of 6 μg/mL cisplatin and 100 μg/mL 5-Fu for 72 hours, suggesting much higher viability both in NF-YA overexpressed SiHa and C33A cells than that in control cells ( Figure 4E and F, P < 0.05). All these results above suggested that NF-YA contributed to the chemoresistance of cervical cancer.

| NF-YA promoted the cell proliferation and tumorigenicity by up-regulating SOX2 in cervical cancer
All these results above suggested that NF-YA could maintain the characteristic of cervical cancer CSC by increased the properties F I G U R E 4 NF-YA-positive cells shared the chemical resistance characteristics of CSC. (A and B) Cell viability was measured using an MTT assay after treatment with different concentrations of cisplatin for 24 h and calculated the IC 50 of different cell groups. (B) Cell viability was measured using an MTT assay after treatment with a constant dose (6 μg/mL) of cisplatin for 1, 3, 5 and 7 days. (C) Cell viability was measured using an MTT assay after treatment with different concentrations of 5-Fu for 48 h and calculated the IC 50 . (D) Cell viability was measured using an MTT assay after treatment with a constant dose (100 μg/mL) of 5-Fu for 36, 48, 60 and 72 h. (E and F) Giemsa staining after exposure to constant concentration of 6 μg/mL cisplatin or 100 μg/mL 5-Fu for 72 h in SiHa-NF-YA and C33A-NF-YA cells. Data are presented as the mean ± SD of experiments in triplicate and statistically analysed with Student's t test. The symbols represent the following: *P < 0.05, **P < 0.01 of tumorigenicity, cell growth, self-renewal and chemoresistance. In our previous study, SOX2 protein has been identified as a marker of cervical CSC. Here, we found that NF-YA up-regulated the expression of SOX2 protein detected by Western blotting and IHC ( Figure 5A and B). Combining with the trans-activation function of NF-YA to the SOX factors, we analysed the SOX2 promoter, containing 4650bp and 1828bp through two sides of CDS region, and constructed several deletions fused to pGL3 basic luciferase reporter plasmid ( Figure 5C). The luciferase activity detected by dual-luciferase reporter assay in NF-YA overexpressed SiHa and C33A cells was significantly higher than that in control cells when containing −1185bp region regardless of whether there was −1828 downstream or not ( Figure 5C, P < 0.05).
Additionally, two binding sites (SOX2 and NF-Y) were found in the candidate region of SOX2 promoter based on the Web Promoter Scan Service and TF SEARCH online software. Then, 2 mutations in SOX2 binding site (GAACAATG to TCCACCGT, pink site) and in NF-Y binding site (TGATTGGTC to GTCGGTTGA, green site) of plasmid pSOX2-1185-luc were constructed ( Figure 5D). The luciferase activity of both the 2 mutations was significantly inhibited compared with pSOX2-1185-Luc reporter in NF-YA overexpressed cells. Of note the mutation of NF-Y binding site resulted in a significant decrease in transcriptional activity in NF-YA overexpressed cells to the level of that in control cells ( Figure 5D, P < 0.05). These results suggested that CCAAT/ATTGG box upstream of SOX2 promoter was indispensable for the transcription of SOX2 by NF-YA.  Figure 6A. ChIP assay demonstrated that SOX2 and NF-YA protein physiologically binds directly to the cis-element in SiHa-NF-YA and C33A-NF-YA cells, respectively ( Figure 6B, P < 0.05). Additionally, EMSA was then employed to assess whether the nuclear protein lysate from the SiHa-NF-YA and C33A-NF-YA cells binds to the probe sequence containing the CCAAT/ATTGG box. As shown in Figure 6C, we detected a strong band in the group of probe and protein in SiHa-NF-YA and C33A-NF-YA cells. Additionally, biotin-EBNA DNA and EBNA extract supplied by the reagent kit were carried out as the positive control.
In order to further confirm SOX2 was up-regulated by NF-YA, the xenograft assay showed that the tumour formation ability was inhibited. This suggests that silenced SOX2 could reverse the tumour growth promoted by NF-YA ( Figure 6D, P < 0.05). At last, the IHC assay of cervical cancer tissues from patients (N = 10) determined the positive correlation between NF-YA and Ki-67/SOX2 ( Figure 6E, P < 0.05). Also, the TCGA database showed that NF-YA mRNA was positive related both to the cell proliferation marker Ki-67 and pluripotency marker SOX2 ( Figure S2, P < 0.05). These findings suggested that NF-YA up-regulated the expression of SOX2 in cervical cancer cells by directly binding to the NF-Y binding site (CCAAT/ATTGG box) upstream of the SOX2 promoter ( Figure 6F).

| D ISCUSS I ON
NF-Y, also known as the CCAAT-binding factor, is an evolutionarily conserved transcription factor composed of three subunits: NF-YA, NF-YB and NF-YC. It binds to CCAAT motif in the proximal promoter region to induce gene expression. 29 NF-YA which is considered as the limited regulatory subunit is required for the complex assembly and DNA binding. 30 NF-Y has been reported to be involved in the embryo-genesis and tumour progression. 31 Recent observations highlight that NF-Y involved in the maintenance of CSC characteristics of several type of cancers such as oral cancer, hematopoietic stem cell and embryonic carcinoma. 20,25,32 NF-YA also involved in the cell proliferation, metastasis and other malignant biological in different types of carcinomas. In breast cancer, NF-YA is associated with a proliferative signature, signal loss of epithelial features, acquisition of EMT and more aggressive behaviour and also has worst clinical outcomes. 33 Also, NF-YA contributes to tumour invasion and angiogenesis through EZH2-STAT3 signalling in human melanoma cells. 34 In hepatocellular carcinoma, NF-YA repressed by ZHX2 inhibited the activation of MDR1 transcription and, in doing so, enhances the effects of chemotherapeutics. 35 However, in human embryonal carcinoma cells, NF-YA significantly reduced the cell growth and cell pluripotency by the decrease in the level of the pluripotency marker SOX2. 24 Whether and how it participates in the process of cervical carcinogenesis remains obscure. Here, IHC analysis has revealed the gradually high expression of NF-YA from NC to HISL and then to SCC, indicating that NF-YA might promote the progression of cervical cancer. Additionally, NF-YA promoted the cell proliferation in vivo and in vitro by tumour xenograft, cell growth and cell viability assays. The CSC characteristics of self-renewal and chemoresistance abilities were also increased. Here, we firstly identified that NF-YA maintains the stemness and tumorigenic properties of cervical cancer cells.
NF-Y has been reported to transcriptional activate the expression of SOX genes, including SOX3, SOX9 and SOX18 mediated, at least in part, by direct binding to CCAAT boxes. [26][27][28] As the target gene, SOX2 has been confirmed being repressed by NF-YA to decrease the proliferation and pluripotency in embryonic carcinoma. 24 Here, we found a new binding site for NF-YA in the SOX2 promoter, resulting to the transcription in cervical cancer cells. Meanwhile, SOX2 has been previously identified as the key factor in several types of CSC, 36,37 including cervical cancer. 16 We previously have isolated the SOX2-positive cervical CSC from SiHa and C33A, which exhibited the major characteristics of CSC, including self-renewal, differentiation and tumour progression properties. 17 To further define the cis-regulatory elements of SOX2 promoter bound by NF-YA, the TFSEARCH database was used to identify a putative CCAAT motif, which was the binding site of NF-YA.
Mutagenesis of this motif demonstrated that CCAAT box plays a crucial functional role in the regulation of the SOX2 promoter. Gel mobility shift analysis and ChIP assay demonstrated that the CCAAT box motif could bind the transcription factor NF-Y in vitro and in vivo. As the activator of SOX2, high expression of NF-YA in cervical cancer was positive related to it and promoted the stemness of cervical cancer cells, suggesting that NF-YA might be another biomarker or regulator of cervical CSC. However, whether there are other target genes of pathway regulated by NF-YA to promote the cell proliferation and tumorigenicity should to be further explored.
In conclusion, NF-YA highly expressed in cervical cancer, promoted the cell growth in vitro and in vivo and maintained the cervical CSC characteristics by driving SOX2 expression. This study might provide an important insight into the biology of CSC and identified a potential target for intervention of cervical cancer. Analysis) repository, http://gepia.cance r-pku.cn/.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data used to support the findings of this study are available from the corresponding author upon request.