HK2 Is a Crucial Downstream Regulator of miR-148a for the Maintenance of Sphere-Forming Property and Cisplatin Resistance in Cervical Cancer Cells

The acquisition of cancer stem-like properties is believed to be responsible for cancer metastasis and therapeutic resistance in cervical cancer (CC). CC tissues display a high expression level of hexokinase 2 (HK2), which is critical for the proliferation and migration of CC cells. However, little is known about the functional role of HK2 in the maintenance of cancer stem cell-like ability and cisplatin resistance of CC cells. Here, we showed that the expression of HK2 is significantly elevated in CC tissues, and high HK2 expression correlates with poor prognosis. HK2 overexpression (or knockdown) can promote (or inhibit) the sphere-forming ability and cisplatin resistance in CC cells. In addition, HK2-overexpressing CC cells show enhanced expression of cancer stem cell-associated genes (including SOX2 and OCT4) and drug resistance-related gene MDR1. The expression of HK2 is mediated by miR-145, miR-148a, and miR-497 in CC cells. Overexpression of miR-148a is sufficient to reduce sphere formation and cisplatin resistance in CC cells. Our results elucidate a novel mechanism through which miR-148a regulates CC stem cell-like properties and chemoresistance by interfering with the oncogene HK2, providing the first evidence that dysregulation of the miR-148a/HK2 signaling plays a critical role in the maintenance of sphere formation and cisplatin resistance of CC cells. Our findings may guide future studies on therapeutic strategies that reverse cisplatin resistance by targeting this pathway.


INTRODUCTION
Cervical cancer (CC) is ranked the fourth most commonly diagnosed cancer in women and is the second-leading cause of cancer-related death among women in developing countries (1). Surgery followed by a combination of platinum/paclitaxel-based chemotherapy is currently considered the standard of care for CC (2). Although the majority of CC patients have an initial response to chemotherapy, such as cisplatin (CDDP), a substantial proportion of patients eventually develop chemoresistance and relapse (2). Therefore, a better understanding of the mechanisms underlying CDDP resistance is required to retrieve the chemosensitivity in CC patients.
Most cancers contain a small subpopulation of cells (known as "cancer stem cells") that drives the persistence of malignant tumors by producing new cancer cells (3). The cancer stem cell theory explains numerous clinical observations, such as the recurrence of tumors after initially successful chemotherapy and/or radiation therapy, and metastasis (3). Cancer stem cells have been found in many cancer types, including CC (4). In CC, cancer stem cells have been associated with resistance to commonly used anti-cancer drugs such as CDDP (4). Several potential stem cell markers, including MSI1, ALDH1, SOX2, and OCT4, have been used to identify putative cancer stem cells of CC (5). However, the molecular mechanisms underlying the induction and maintenance of cancer stem cells in CC remain to be explored.
In addition to several key genetic alterations in oncogenes and tumor suppressor genes, several epigenetic mechanisms, including microRNAs (miRNAs), have also been shown to regulate the progression and stemness of CC (6). Hexokinase 2 (HK2) is an enzyme that catalyzes the first committed step in glucose metabolism and converts glucose to glucose-6-phosphate (7). After knocking down HK2 expression, CC cells demonstrated significantly attenuated proliferation ability and glycolysis (8). Deletion of HK2 inhibits the growth and migration of CC cells (9). The same study has also demonstrated that knockdown of HK2 decreases the phosphorylation of AKT and mTOR, and increases the expression of p53 in CC cells (9). Another report indicated that HK2 promotes the proliferation of CC cells in vitro and tumor formation in vivo by regulating the Raf/MEK/ERK signaling pathway (10). Moreover, a previous study demonstrated that HK2 expression was suppressed by miR-9-5p by directly binding its 3'-untranslated region (3′-UTR) in CC cells (9). To the best of our knowledge, there have been no reports on the function and the underlying mechanism of HK2 in regulating CC stemness and CDDP resistance.
Here, we have explored for the first time that HK2 is overexpressed in CC, and its overexpression promotes sphere formation and CDDP resistance of CC cells. Moreover, we have demonstrated that the induction of HK2 expression in CC is partly due to the downregulation of miR-145, miR-148a, and miR-497. In particular, miR-148a acts as a key tumor suppressor that controls sphere formation and CDDP resistance by targeting HK2. Together, we have unveiled a novel epigenetic mechanism that controls sphere formation and CDDP resistance via the miR-148a/HK2 axis in human CC. Thus, therapeutic targeting this axis may reverse CDDP resistance in CC patients.

Human Tissue Specimens
This study was approved by the Research Ethics Committee of Affiliated Hospital of Inner Mongolia Medical University. Written informed consent was obtained from all patients. A total of 30 CC patients with no preoperative chemotherapy, immunotherapy, or radiotherapy were enrolled in this study. Samples were collected from patients who were diagnosed with CC and underwent surgical resection in the Affiliated Hospital of Inner Mongolia Medical University. After surgery, CC tissues and corresponding adjacent normal tissues were immediately frozen by liquid nitrogen and stored at -80°C for further analysis.

Cell Lines and Cell Culture
Two human CC cell lines (HeLa and SiHa) were obtained from the American Type Culture Collection (ATCC, Manassas, VA). The HPV16 E6/E7-immortalized human cervical epithelial cell line Ect1/E6E7 (ATCC) was considered to be a non-cancerous ectocervical epithelial cell line (11,12). All cells were cultured in DMEM/F12 medium (Sigma-Aldrich, St. Louis, MO) supplemented with 10% fetal bovine serum (FBS, Invitrogen, Carlsbad, CA) at 37°C with a humidified air containing 5% CO2. These cells were tested for mycoplasma contamination using the DAPI staining method.

Western Blotting Analysis
Total proteins were isolated from CC cells, which were lysed on ice with a RIPA lysis buffer (Cell Signaling Technology, MA). The protein concentration was determined using a BCA Protein Assay Kit (Beyotime Biotechnology, Shanghai, China). Extracted protein (30 µg) was separated by 12% SDS-PAGE and transferred onto a PVDF membrane (GE Healthcare Life Sciences, Piscataway, NJ). The membranes were blocked with 5% non-fat milk for 1 h and then incubated with anti-HK2 antibody (1:1000, #2106, Cell Signaling) and anti-GAPDH antibody (1:5000, #2118, Cell Signaling) at 4°C overnight. Then, the membranes were subsequently incubated with secondary antibodies for 2 h at room temperature. The protein bands were detected using an ECL detection kit (Amersham Pharmacia Biotech, UK).

Sphere Formation Assay
Sphere formation assay was performed as previously described (6). CC cells (3000) cells were seeded in 6-well ultra-low attachment plates (Corning Incorporated, Corning, NY), where contain serum-free medium DMEM/F12 (Sigma-Aldrich) supplemented with N2 plus media supplement (Invitrogen), 20 ng/ml epidermal growth factor (Invitrogen), 20 ng/ml basic fibroblast growth factor (Invitrogen), and 4 mg/ml heparin (Sigma-Aldrich) for 14 days. The images were captured after 2 weeks of cell culture, and the number of spheres larger than 50 mm was counted.

Cell Viability Assay
Cell viability was determined using Cell Counting Kit-8 (CCK-8) assay (Dojindo, Japan). In brief, CC cells (5000 cells per well) were seeded into 96-well plates and cultured for 24 h. The medium was refreshed and cells were exposed to varying concentrations of CDDP. Twenty-four hours after the addition of CDDP, CCK-8 solution (10 ml per well, Dojindo, Japan) was added to each well. Cells were incubated at 37°C for 3 h. The absorbance value was measured at a wavelength of 450 nm.

Xenograft Mouse Model
The study on animal subjects was approved by the Ethics Committee of Affiliated Hospital of Inner Mongolia Medical University. In vivo xenograft mouse experiments were performed as previously described (14,15). Female mice (BALB/c nude, 4-6 weeks old) were purchased from Vital River Laboratory Animal Technology (Beijing, China). CC cells transfected with HK2 overexpression vector (OriGene, Rockville, MD) or HK2 shRNA (Santa Cruz Biotechnology; Santa Cruz, CA), were subcutaneously injected into the flank of nude mice under aseptic conditions. After a 7-day administration, mice were treated with CDDP (30 mg/kg) intraperitoneally. The tumors were measured in two dimensions by using manual calipers every 3 days. The tumor volume was calculated using the following formula: volume = 0.5 × length × width 2 . After 30 days, all mice were sacrificed and tumor samples were weighted.

Dual-Luciferase Reporter Assay
The human HK2 wild-type 3′-UTR luciferase reporter vector was constructed from Genechem (Shanghai, China). Mutations of the miRNA binding sites in the HK2 3′-UTR sequence were made using site-directed mutagenesis using the QuickChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). CC cells were seeded into 24-well plates, and then co-transfected with HK2 3′-UTR wild-type (or mutated) luciferase reporter vector, the pRL-CMV vector (Promega, Madison, WI), along with 30 nM of miRNA mimic, control mimic, miRNA inhibitor, or control inhibitor, using Lipofectamine 3000 (Invitrogen). After 48 h, luciferase activity was examined using the Dual-Luciferase Reporter Assay System (Promega). Firefly luciferase activity was normalized to that of the Renilla luciferase activity.

Statistical Analysis
Statistical analysis was performed using SPSS 20.0 (SPSS, Chicago). Results were presented as the mean ± standard error (SD) from three independent experiments. Statistical significance was determined using an unpaired two-tailed Student's t-test, one-way ANOVA test, or Mann-Whitney U test. Pearson's correlation test was used to measure the correlation coefficient between two variables. Differences were considered statistically significant if the P-value was < 0.05.

Overexpression of HK2 Correlates With Poor Prognosis in CC Patients
The Oncomine database (https://www.oncomine.org/) was utilized to investigate the expression of HK2 between tumors and normal tissues. The results revealed that HK2 expression was elevated in most cancer tissues (including CC) compared with normal tissues ( Figure 1A). To verify these results, we explored the expression of HK2 in TCGA CC patients using the UALCAN database (http://ualcan.path.uab.edu/index.html). The levels of HK2 were significantly increased in CC tissues than in normal tissues ( Figure 1B). We also analyzed the correlation between HK2 expression and tumor stage in CC patients using the UALCAN database. The findings showed that those patients with advanced-stage CCs tended to exhibit higher levels of HK2 ( Figure 1C). We evaluated the prognostic significance of HK2 in CC patients using the KM plotter database (http://kmplot.com/ analysis/). Increased expression of HK2 was strongly associated with poor overall survival ( Figure 1D). Using qRT-PCR analysis, we found an upregulation of HK2 mRNA expression in CC cell lines (HeLa and SiHa), compared to the normal cervical epithelial cell line Ect1/E6E7 ( Figure 1E). These results suggested that HK2 is upregulated in CC and might have a tumor-promoting function.

HK2 Promotes Sphere Formation and CDDP Resistance in CC Cells
To elucidate the biological function of HK2 in CC cells, we generated HeLa cell lines with stable HK2 overexpression and established SiHa cell lines with silencing of HK2 using HK2 shRNA. The overexpression and knockdown of HK2 in CC cells were confirmed by western blotting analysis (Figure 2A). The overexpression of HK2 increased the sphere formation in HeLa cells, whereas the knockdown of HK2 decreased the sphere formation in SiHa cells ( Figure 2B). Cell viability assay showed that the overexpression (or knockdown) of HK2 significantly increased (or decreased) the resistance of CC cells to CDDP in dose-dependent assays ( Figure 2C). The expression of MDR1, SOX2, and OCT4 were increased after the overexpression of HK2 in HeLa cells, and an opposite effect was observed after HK2 shRNA transfection in SiHa cells ( Figure 2D).
To analyze whether the inhibition of HK2 could sensitize CC cells to CDDP treatment in vivo, HK2-overexpressing HeLa cells, SiHa cells with HK2 silencing, as well as the respective control cells, were injected subcutaneously into nude mice and were treated with CDDP. Mice implanted with cells transfected with HK2 overexpression vector developed bigger tumor volumes than the control group ( Figure 2E). Additionally, the weight of the tumors derived from the HK2-overexpressing HeLa cells was much heavier than those from the control cells ( Figure 2E). Conversely, compared to the control mice, the mice implanted with the HK2-silenced SiHa cells had smaller tumor volumes and lighter tumor weight ( Figure 2E). Collectively, these results suggested that HK2 overexpression facilitates the formation of spheroids and the development of CDDP resistance in CC cells.

MiR-145, MiR-148a, and MiR-497 Are Upstream Regulators of HK2 in CC Cells
To investigate the upstream regulatory mechanism of HK2 in CC, we conducted in silico analysis to identify the potential of miRNAs to bind to HK2 mRNA. Using TargetScan, miRDB, and miRSystem databases, a set of miRNAs (including miR-145, miR-148a, and miR-497) were found to have binding sites in the 3′-UTR of the HK2 transcript ( Figures 3A, B). Since the downregulation of miR-145, miR-148a, and miR-497, but not the remaining miRNAs, were correlated with poor overall survival in CC patients ( Figure 3C), we decided to focus on these three miRNAs in our subsequent experiments. Our qRT-PCR experiments verified that the levels of miR-145, miR-148a, and miR-497 were significantly lower in SiHa cells than in HeLa cells (data not shown). To explore the regulatory relationship between these miRNAs and HK2, we performed the qRT-PCR assay to examine the mRNA expression of HK2 in SiHa cells transfected with miRNA mimics and in HeLa cells transfected with miRNA inhibitors. The expression of HK2 was significantly suppressed in miR-145/148a/497 mimicexpressing SiHa cells, while the inhibition of miR-145/148a/497 upregulated HK2 expression in HeLa cells ( Figure 3D). These results demonstrated that HK2 is negatively regulated by miR-145/148a/497.

MiR-148a Inhibits Sphere Formation and CDDP Resistance in CC Cells
MiR-145 and miR-497 are known tumor suppressors in CC (16)(17)(18)(19). However, little is known about the role of miR-148a in CC  CDDP resistance. Thus, we focused on miR-148a in this study.
To analyze the effect of miR-148a overexpression or silencing on sphere formation and CDDP resistance in CC cells, we transfected miR-148a mimic into SiHa cells and introduced miR-148a inhibitor into HeLa cells. Sphere formation assays and cell viability assays showed that miR-148a overexpression significantly reduced the sphere-forming ability and attenuated CDDP resistance of SiHa cells (Figures 5A, B). In contrast, miR-148a inhibition significantly promoted sphere formation and increased CDDP resistance in HeLa cells (Figures 5A, B). The results of qRT-PCR assays confirmed the downregulation of MDR1, SOX2, and OCT4 in miR-148a-overexpressing SiHa cells, as well as their upregulation in miR-148a-silenced HeLa cells ( Figure 5C). To determine whether the above results were reproducible, miR-148a-overexpressing SiHa cells or miR-148a-silenced HeLa cells were injected into nude mice and were treated with CDDP. We found that the tumors formed by miR-148a-silenced HeLa cells were larger and had more weight than control tumors ( Figure 5D). However, tumors formed by miR-148a-overexpressing SiHa cells were smaller in both size and weight than the tumors formed by control cells ( Figure 5D). Thus, we demonstrated that miR-148a serves as a tumor suppressor by antagonizing sphere formation and increasing the CDDP sensitivity of CC cells.

The MiR-148a/HK2 Axis Mediates Sphere Formation and CDDP Resistance in CC Cells
To investigate whether miR-148a suppresses sphere-forming ability and CDDP resistance of CC cells by targeting HK2, we transfected SiHa cells with miR-148a mimic (or control mimic),  along with HK2 overexpression vector (or control vector), and also transfected HeLa cells with miR-148a inhibitor (or control inhibitor), together with HK2 siRNA (or control siRNA). The protein expression of HK2 was examined by western blotting analysis. The expression of HK2 was decreased with the transfection with miR-148a mimic, and this reduction was reversed by the introduction of the HK2 expression vector ( Figure 6A). In addition, the protein expression of HK2 was induced by miR-148a inhibition, and this increase was reversed by transfection with HK2 siRNA ( Figure 6A). The results of sphere formation and cell viability assays showed that sphere formation and CDDP resistance were suppressed by transfection of miR-148a mimic, whereas this repression could be eliminated by forced expression of HK2 ( Figures 6B, C). We also verified that miR-148a inhibition induced sphere formation and CDDP resistance, while knockdown of HK2 attenuated these malignant properties of CC cells (Figures 6B, C). Consequently, these results demonstrated that the miR-148a/HK2 axis functionally regulates sphere formation and CDDP resistance in CC cells.

Correlation Between MiR-145/148a/497 and HK2 Expression in CC Tissues
To further define the role of miR-145, miR-148a, miR-497, and HK2 in human CC, we measured their expression in 30 adjacent normal tissues and 30 CC tissues using the qRT-PCR assay. We found that CC tissues exhibited lower expression of miR-145, miR-148a, and miR-497 ( Figure 7A). Our results also showed increased expression of HK2 in human CC tissues (Figure 7). Overall, these results support the notion that HK2 is crucial for the maintenance of the sphere-forming property and CDDP resistance in CC cells, and upregulation of HK2 observed in CC tissues might be caused by the repression of multiple tumor suppressor miRNAs (including miR-145, miR-148a, and miR-497).

DISCUSSION
Even though CDDP has been certified as a front-line drug for CC treatment, primary and acquired CDDP resistance leads to the failure of CDDP-based therapy and accounts for the recurrence of CC (2). Growing evidence has indicated that cancer stem cells could contribute to chemoresistance through various mechanisms, such as induction of metabolic enzyme aldehyde dehydrogenase (20), overexpression of ABC transporters (21), activation of Notch, Hedgehog, and Wnt pathways (22). In this work, we have demonstrated that HK2, a critical enzyme in glucose metabolism, is commonly  overexpressed in human CC, and knocking down HK2 sensitizes CC cells to CDDP. Besides, overexpression of HK2 could be partially due to the downregulation of miRNAs (especially miR-148a). To our knowledge, this study is the first to elucidate the oncogenic role of HK2 in promoting CDDP resistance in CC, and to reveal the biological relevance of the miR-14a/HK2 axis in the sphere formation ability and CDDP sensitivity of CC cells. Aberrant expression of HK2 in CC has been previously reported (8,9). The results from our meta-analysis and qRT-PCR assays have confirmed the elevated expression of HK2 in CC tissues and cell lines, and the upregulation of HK2 is closely associated with worse survival rates of CC patients. Although the silencing of HK2 was known to inhibit the growth and migration of CC cells (8,9), whether HK2 could influence the sphereforming properties and the development of CDDP resistance of CC cells remains unknown. Our gain-of-function and loss-offunction assays have uncovered the promoting role of HK2 in maintaining cancer stem cell-like ability and the acquisition of CDDP resistance. Importantly, knockdown of HK2 combined with CDDP treatment reduces CC proliferation in vivo. These results together suggested that targeting HK2 may abrogate cancer stem cell populations, as demonstrated by reduced sphere formation capacity, and regain the sensitivity of CC patients to CDDP.
In addition to regulating glucose metabolism, HK2 promotes cancer stem cell self-renewal in CC cells, elucidating that HK2 is an upstream activator of cancer stem cell expansion (23). In ovarian cancer, overexpression of HK2 enhances stemness properties by upregulating the expression of cancer stemness-related genes, such as OCT4 (24). Moreover, HK2 is required for the maintenance of esophageal cancer stem cell phenotypes (25). The chemoresistance to 5-FU and CDDP and the tumorigenesis of esophageal cancer stem cells are reduced after HK2 knockdown (25). Overexpression of HK2 effectively promoted epithelial-mesenchymal transition (EMT) phenotypes and enhanced aerobic glycolysis in EC cells (26). Consistent with these previous studies, the current demonstrated that silencing HK2 decreases the expression of cancer stem cell markers OCT4 and SOX2, suggesting that therapeutic approaches targeting HK2 might be an effective strategy for suppressing the expansion and maintenance of the cancer stem cell population in CC cells.
Targeting HK2 using HK2 inhibitors has been a hopeful strategy for cancer treatment. To date, different types of HK2 inhibitors have been developed (27). For instance, the selective HK2 inhibitor Benitrobenrazide binds directly to HK2, blocks glycolysis, and induces apoptosis in HK2-overexpressing cancer cells (28). Glycolysis is activated in radioresistant CC cells, and inhibiting glycolysis with HK2 inhibitor 2-DG improves the sensitivity of radioresistant CC cells to irradiation (29). These results support the possibility that inhibition of glycolysis via specific HK2 inhibitors, in combination with other therapies, would be an effective strategy for the treatment of CC.
Having established the role of HK2 in sphere formation and CDDP resistance of CC cells, we sought mechanistic evidence for activation of HK2. According to our in silico analysis, qRT-PCR assays, and luciferase reporter assays, three miRNAs (miR-145, miR-148a, and miR-497) were determined as upstream regulators of HK2 in CC cells. MiR-145 and miR-497 suppress the aggressive phenotypes of CC cells (16)(17)(18)(19). MiR-145 negatively modulates proliferation, EMT, migration, and invasion of CC cells (16). Interestingly, miR-145 directly binds to HK2 mRNA and reduces its expression in renal cell carcinoma (30). In our work, we have found that miR-145, miR-148a, and miR-497 inhibit endogenous HK2 expression by binding to its 3′-UTR. Consequently, silencing of these miRNAs in CC tissues stimulates the expression of HK2, facilitating the development of cancer stem cell-like properties and CDDP resistance. Collectively, we propose a complex regulatory relationship between these miRNAs and HK2 in CC. Of note, HK2 levels are regulated at the transcriptional, post-translational, and translational levels (31). Different genetic and epigenetic mechanisms may be involved in the regulation of HK2 in CC, the exact mechanisms for the upregulation of HK2 are deserving of future studies.
Aberrant expression of miR-148a is a common event in human tumors (32). For instance, the downregulation of miR-148a has been detected in various cancers, including gastric, colon, pancreatic, liver, esophageal, breast, and lung cancer (32). It is known that miR-148a could suppress the proliferation, migration, invasion, and metastasis of cancer cells by targeting a large number of downstream genes (32). Of note, in esophageal cancer cells, re-expression of miR-148a significantly improves the sensitivity of cells to CDDP and 5-FU (33). In cervical cancer, a previous study using miRNA array and qRT-PCR assay has suggested that miR-148a levels were significantly lower in CC tissues compared to normal tissues (34). Overexpression of miR-148a could markedly suppress the proliferation of CC cells (35). Moreover, lncRNA SNHG4 promotes the proliferation of CC cells through regulating c-Met via targeting miR-148a (36).
LncRNA SNHG12 sponges miR-148a to increase CDK1 expression, thereby attenuating the radiation-induced apoptosis in CC cells (37). In line with this research, we have shown that miR-148a attenuates CDDP resistance in CC cells, at least in part, by targeting HK2. Therefore, our results elucidated a new mechanism for miR-148a in chemoresistance. Cancer cells frequently overexpress some proteins implicated in glucose metabolism, such as GLUT1, the main glucose transporter (38). Since a recent study has reported that GLUT1 is a direct target of miR-148a in human intrahepatic cholangiocarcinoma (39), miR-148a may also regulate glucose metabolism by inhibiting GLUT1 in CC. Future studies to identify the interactions between miR-148a and its targets would strengthen our understanding of the function of miR-148a, and yield key insights into the metabolic landscape of CC.

CONCLUSION
In summary, our findings demonstrate the involvement of HK2 in the maintenance of the sphere-forming property and CDDP resistance in CC cells. In addition, we have identified miR-148a as an upstream suppressor of HK2 to inhibit the ability of sphere formation and increase the sensitivity of CDDP in CC cells ( Figure 8). We have uncovered a previously unknown miR-  148a/HK2 axis that modulates cancer stem cell-like features and CDDP resistance in CC.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding authors.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Research Ethics Committee of Affiliated Hospital of Inner Mongolia Medical University. The patients/participants provided their written informed consent to participate in this study. The animal study was reviewed and approved by Ethics Committee of Affiliated Hospital of Inner Mongolia Medical University.