Hypermethylation of PRKCZ Regulated by E6 Inhibits Invasion and EMT via Cdc42 in HPV-Related Head and Neck Squamous Cell Carcinoma

Simple Summary HPV+ and HPV- HNSCC share distinct epigenetic characteristics and clinicopathological features. The aim of our study was to assess whether DNA methylation plays a role in the progression of HPV+ HNSCC. We used a HumanMethylation450 BeadChip array (Illumina) to identify PRKCZ genes exhibiting different levels of DNA methylation between HPV+ and HPV- HNSCC. PRKCZ acts as a potent tumor promoter in HPV+ HNSCC. These findings may provide a possible explanation for the differences in the clinicopathological characteristics between HPV+ and HPV- HNSCC and promising ideas for the treatment of HPV+ HNSCC. Abstract Purpose: To study the role of target genes with aberrant DNA methylation in HPV+ HNSCC. Methods: A HumanMethylation450 BeadChip array (Illumina) was used to identify differentially methylated genes. CCK-8, flow cytometry, wound healing, and cell invasion assays were conducted to analyze the biological roles of PRKCZ. Western blot, qRT-PCR, immunohistochemistry, and animal studies were performed to explore the mechanisms underlying the functions of PRKCZ. Results: We selected PRKCZ, which is associated with HPV infection, as our target gene. PRKCZ was hypermethylated in HPV+ HNSCC patients, and PRKCZ methylation status was negatively related to the pathological grading of HNSCC patients. Silencing PRKCZ inhibited the malignant capacity of HPV+ HNSCC cells. Mechanistically, HPV might promote DNMT1 expression via E6 to increase PRKCZ methylation. Cdc42 was required for the PRKCZ-mediated mechanism of action, contributing to the occurrence of epithelial-mesenchymal transition (EMT) in HPV+ HNSCC cells. In addition, blocking PRKCZ delayed tumor growth in HPV16-E6/E7 transgenic mice. Cdc42 expression was decreased, whereas E-cadherin levels increased. Conclusion: We suggest that PRKCZ hypermethylation induces EMT via Cdc42 to act as a potent tumor promoter in HPV+ HNSCC.


Introduction
Head and neck squamous cell carcinoma (HNSCC) is the most common malignant tumor stemming from the oral cavity, with approximately 890,000 new cases and 450,000 deaths occurring worldwide in 2018 [1,2]. Despite constant improvements in treatments, the 5-year survival rate of HNSCC patients remains unsatisfactory at only approximately 50% [3,4]. In addition to classic carcinogenic factors such as smoking, drinking,

Materials and Methods
Patients and clinical specimens 40 primary HNSCC tissue specimens, including 12 paraffin-embedded tissue specimens and 28 fresh tissue specimens, were derived from patients without history of previous treatments or autoimmune diseases who had accepted radical surgery at West China Hospital of Stomatology, Sichuan University. All patients gave informed consent. These specimens have been pathologically identified as HNSCC and divided into different grades on the basis of the International Union for Cancer Control's latest standards. The clinicopathological characteristics of these patients are provided in Supplementary Table S3. All samples were fixed with 10% formalin, paraffin-embedded, or frozen at −80 • C for RNA and DNA extraction, or for subsequent tests.

HumanMethylation450 BeadChip Array (Illumina)
An HumanMethylation450 BeadChip array (Illumina, Design ID: OE2015Q1029) of clinical specimens from HNSCC patients who had accepted radical surgery at West China Hospital of Stomatology, Sichuan University, including 3 HPV-positive cases and 3 HPVnegative cases, was provided by oeBiotech Limited Company. The clinicopathological characteristics of these 6 patients were provided in Supplementary Table S4.

Immunohistochemistry (IHC)
Formalin-fixed paraffin-embedded sections were firstly dewaxed in xylene and rehydrated in graded alcohol solutions. Then antigen retrieval was performed by boiling for 3 min in 0.01 M citrate buffer (pH = 6.0), and sections were immersed in 3% hydrogen peroxide for 10 min and in PBS for 5 min twice. After incubating with 5% goat serum for antigen blocking, the sections were incubated with the primary antibody at 4 • C overnight in a moist chamber. Additionally, the following primary antibodies were listed: anti-PKC-ζ (1:300, rabbit anti-human, Proteintech), anti-P16 (1:500, rabbit anti-human, Bioss, Boston, MA, USA), anti-Cdc42 (1:200, rabbit anti-human, Proteintech), anti-E-cadherin (1:500, mouse anti-human, Cell Signaling Technology Inc.), anti-N-cadherin (1:100, rabbit anti-human, Proteintech), anti-Vimentin (1:100, rabbit anti-human, Proteintech). After washing with PBS twice, the sections were incubated with biotinylated anti-mouse/rabbit IgG and sequent with streptavidin-biotin peroxidase both for 15 min. DAB was added to detect the primary antibody, and hematoxylin was used to stain the nucleus. As negative controls, sections were analyzed in parallel except incubating with isotype-specific immunoglobulin (IgG) instead of the primary antibody. Quantification was evaluated by three independent investigators who were blinded to patient characteristics. According to the degree of staining in ten randomly selected fields at 400× magnification, the level of target protein was assigned as 0 (negative), 1 (weak), 2 (moderate), and 3 (strong).

Transfections
GenePharma Co., Ltd., Suzhou, China provided the design and composition services of small-interfering RNAs (siRNAs) targeting E6, E7, and PRKCZ. EndoFectin TM (GeneCopoeia) was used for transfections according to the manufacturer's instructions. Further experiments were conducted 48 h after transfection. The plasmids for HPV16 E6 and E7 overexpression and negative control plasmid were designed by GenePharma Co., Ltd., China. Stably transfected Cal27 with E6 and E7 overexpression were derived from the parental cells by a fluorescence microscope (Olympus BX51, Olympus Corporation, Shinjuku City, Tokyo, Japan) detection and puromycin (Sigma-Aldrich, St. Louis, MO, USA) selection. The sequences of siRNAs are listed in Supplementary Table S6.

Cell Proliferation Assay and Flow Cytometry-Based Apoptosis Analysis
CCK-8 assay was used to test cell proliferation capacity. Treated cells were seeded in 96-well plates at a density of 1 × 10 3 (cells/well) and cultured at 37 • C in a humidified atmosphere with 5% CO 2 for 1-5 days. Subsequently, cells were treated with 10 µL CCK-8 reaction solution (Hanbio Inc., Shanghai, China) and incubated for 1 h in cell constant temperature incubator. Then, the light absorption value (OD value) at 450 nm was measured. Each assay was carried out in triplicate.
Flow cytometry using the Annexin V-FITC Apoptosis Detection Kit (Invitrogen) was for the test of cell apoptosis. Treated cells were seeded in 6-well plates and digested after 48 h. Then, cells were resuspended with 1X Binding Buffer. Stained with 5 µL of Annexin V-FITC Conjugate and 5 µL of Propidium Iodide Solution, cells were incubated for 30 min at room temperature under dark conditions. After washing and resuspending, the percentage of stained cells was analyzed using a FACScalibur (Becton-Dickinson). Experiments were conducted in triplicate.

Wound Healing Assay and Cell Invasion Assay
Appropriate cells were seeded in 6-well plates to obtain 100% confluence next day. After 24 h of starvation in serum-free medium, a pipette tip was used to scratch the cells and form the clearance space with suitable width. After washing with PSC to clear out the dead cells, photomicrographs of clearance space were obtained at the timepoint of 0 and 48 h. Closed scratch areas were calculated by ImageJ software. Each experiment was conducted in triplicate. 5 × 10 4 cells were seeded into the upper chamber of 24-well transwell plates (pore size 8 µm, Millipore) with Matrigel-coated membrane (BD Bioscience). After incubation for 24 h in cell constant temperature incubator, the medium in upper chamber was changed into serum-free, while that in lower chamber contained 10% FBS. Then, after 24 h incubation, cells that invaded through the Matrigel-coated membrane were fixed in 4% paraformaldehyde for 30 min and stained with 0.1% crystal violet and counted under a microscope. Five randomly selected fields were used to count the number of stained cells. Assays were conducted in triplicate.

Statistical Analysis
Means comparisons were performed with Student t-tests or one-way ANOVA. The association between PRKCZ methylation and mRNA expression was assessed via 2-tailed Pearson's statistics. All cellular experiments were conducted independently at least three times and in triplicate each time. GraphPad Prism 7.0 (GraphPad Software) was used for processing all data, and values were presented as means ± SD. p < 0.05 was considered to be statistically significant.

PRKCZ Was Hypermethylated in HPV+ HNSCC Compared with HPV-HNSCC
To identify the differentially methylated genes between HPV+ and HPV-HNSCC, a HumanMethylation450K BeadChip array (Illumina) was performed. According to the analyzed data, 1344 genes were hypermethylated and 754 genes were hypomethylated in HPV+ HNSCC, compared with HPV-HNSCC ( Figure 1A). We collected 25 significantly hypermethylated and 25 obviously hypomethylated genes in HPV+ HNSCC based on array data ( Figure 1B, Tables S1 and S2). Furthermore, we performed KEGG analysis on these selected genes. PKRCZ was found to be the only gene associated with HPV infection ( Figures 1C and S1). Therefore, PRKCZ was selected as the focus of our study.
HumanMethylation450K BeadChip array (Illumina) was performed. According to the analyzed data, 1344 genes were hypermethylated and 754 genes were hypomethylated in HPV+ HNSCC, compared with HPV-HNSCC ( Figure 1A). We collected 25 significantly hypermethylated and 25 obviously hypomethylated genes in HPV+ HNSCC based on array data ( Figure 1B, Tables S1 and S2). Furthermore, we performed KEGG analysis on these selected genes. PKRCZ was found to be the only gene associated with HPV infection ( Figures 1C and S1). Therefore, PRKCZ was selected as the focus of our study. All assays were carried out in triplicate. Results were presented as means ± SD. * p < 0.05, ** p < 0.01, *** p < 0.001.
We used results from the TCGA database, paraffin-embedded tissues, and fresh tissues to verify the accuracy of the data obtained from the HumanMethylation450 BeadChip array. According to results from the TCGA database [28,29], PRKCZ mRNA expression was higher in HPV+ HNSCC compared with HPV-cases and the degree of PRKCZ methylation was negatively associated with its mRNA level ( Figure 1D). However, no significant difference in the methylation degree of PRKCZ was noted between HPV+ HNSCC and HPV-cases. In addition, we found that PRKCZ was hypermethylated in HPV+ HNSCC paraffin-embedded tissues and fresh tissues, compared with HPV-samples via qMSP ( Figure 1E). These findings were consistent with the results from the array analysis. In addition, PRKCZ mRNA expression in HPV+ HNSCC fresh tissues was significantly lower than that in HPV-tissues ( Figure 1F). A negative correlation between PRKCZ methylation degree and the mRNA expression level was also observed in fresh tissues (r = −0.3788, p = 0.0469) ( Figure 1G). This difference in the level of PRKCZ methylation in the TCGA database and our data may be due to the use of different detection methods to assess PRKCZ methylation degree and HPV status. Furthermore, the rate of positive PRKCZ staining, which was mainly localized in the cytoplasm, was higher in HPV+ HNSCC. In HPV+ HNSCC, weak, moderate, and strong positive rates were 61.5% Furthermore, we detected the differences in PRKCZ methylation and mRNA levels between HPV+ HNSCC cells (SCC47 and SCC090) and HPV-HNSCC cells (Cal27, SCC25 and SCC9). Consistent with the results above, HPV+ HNSCC cells had increased methylation degrees and decreased mRNA and protein expressions (Figure 2A,B). To further verify the correlation between PRKCZ methylation degree and mRNA, and protein expression, we treated cells with 5-aza-2'-deoxycytidine (5-aza), a potent DNA methylation-specific inhibitor [30,31]. As the concentration of 5-aza increased, the degree of PRKCZ methylation decreased, whereas PRKCZ mRNA and PKC-ζ protein levels increased in SCC47 ( Figure 2C,D) and Cal27 ( Figure 2E,F) cells. Taken together, we found that PRKCZ was hypermethylated in HPV+ HNSCC, which was related to decreased mRNA and protein levels. All assays were carried out in triplicate. Results were presented as means ± SD. ns > 0.05, * p < 0.05, ** p < 0.01, *** p < 0.001.

Relationship between PRKCZ Methylation and Clinicopathological Parameters of HNSCC Patients
We tested the predictive value of PRKCZ methylation status for relevant clinical and pathological parameters of collected HNSCC specimens, including 12 paraffin-embedded samples and 28 fresh samples. The PRKCZ methylation levels of each specimen were determined via qMSP. The associations between PRKCZ methylation status and patients' sex, age, T stage, lymph node metastasis, and tumor recurrence were not significant (p = 0.3138, p = 0.4679, p = 0.2365, p = 0.5322, and p = 0.5827, respectively), whereas PRKCZ methylation status was significantly related to the pathological grading of HNSCC patients (p = 0.0386). Patients with PRKCZ hypermethylation seemed to have worse pathological manifestations ( Figure 2G). The fresh samples we collected were obtained within 3 years; thus, we did not assess the prognostic value of PRKCZ methylation status. Thus, we used bioinformatics to assess the prognostic value of PRKCZ methylation and mRNA from the TCGA and GEO databases. These databases showed that low PRKCZ expression in HPV+ HNSCC patients seems to have a better prognosis; however, HPV+ HNSCC patients with PRKCZ hypermethylation and low expression have much better survival. In addition, for HPV-HNCSS patients, PRKCZ expression and methylation level have no significant effects on the prognosis ( Figure 2H,I). The relations between PRKCZ methylation status and relevant clinical and pathological parameters of HPV+ and HPV-patients were presented in Figure S2. No differences between PRKCZ methylation status and relevant clinical and pathological parameters were noted in HPV+ patients and HPV-patients.

Blocking PRKCZ Inhibited HPV+ HNSCC Cells Proliferation, Invasion, and Migration, Promoted Apoptosis
To provide more powerful proof of the HPV effect on PRKCZ methylation and the PRKCZ effect on HPV+ HNSCC, we established a new HPV+ HNSCC-like cell line termed Cal+E6E7 cell by transfecting Cal27 cells with E6 and E7 lentivirus plasmids. The Western blot results revealed E6 and E7 expression in the new cell line, illustrating successful construction ( Figure 3A). Next, two pairs of siRNAs targeting different sites of PRKCZ mRNA were used to silence expression. Compared with the control siRNA, siRNA1 and siRNA2 reduced PRKCZ mRNA relative expression levels in SCC47 and Cal27+E6E7 cells by 25% and 38%, and 32% and 45%, respectively. Western blot experiments also demonstrated the silencing effects of both pairs of siRNAs on PKC-ζ protein expression ( Figure 3B). Thus, siRNA1 was selected for follow-up experiments.  All assays were carried out in triplicate. Results were presented as means ± SD. ns > 0.05, * p < 0.05, ** p < 0.01, *** p < 0.001.
After transfecting two types of HPV+ HNSCC cells (SCC47 and Cal27+E6E7) with siRNA1, we found that the OD values on Days 3, 4, and 5 were significantly lower in the transfected cells compared with two cells transfected with control siRNA. This finding suggests that silencing PRKCZ impairs the proliferation abilities of HPV+ HNSCC cells ( Figure 3C). We applied wound healing (Figures 3D and S3) and transwell invasion ( Figures 3E and S4) assays to investigate the effect of PRKCZ on the migration and invasion of SCC47 and Cal27+E6E7 cells. The results showed that the wound healing areas of HPV+ HNSCC cells transfected with siRNA1 and control siRNA were 15.150% and 23.216% (SCC47), as well as 17.801% and 25.050% (Cal27+E6E7), respectively. The numbers of invaded cells were 498 and 756 (SCC47), as well as 287 and 473 (Cal27+E6E7), respectively. These data suggest that decreased PRKCZ levels dramatically reduce the migratory and invasive capacities of SCC47 and Cal27+E6E7 cells. Regarding the role of PRKCZ in cell apoptosis, flow cytometry analysis showed that after transfection with siRNA1 for 48 h, the apoptosis rates of SCC47 and Cal27+E6E7 cells were 27.10% and 22.40%, respectively, whereas those of these two cells transfected with control siRNA were 12.04% and 13.21%, respectively ( Figure 3F).
In addition, we used a specific inhibitor of PRKCZ, a PKC-ζ pseudosubstrate inhibitor, which decreased PRKCZ mRNA level in SCC47 cells to 10% ( Figure 3G). Similar to the results above, the OD values on Days 3, 4, and 5 of SCC47 cells treated with the pseudosubstrate inhibitor were significantly reduced compared with those of normal SCC47 ( Figure 3H). In addition, the percentage of wound healing areas ( Figures 3I and S5) and the number of invading SCC47 cells (Figures 3J and S5) treated with the pseudosubstrate inhibitor were 14.571% and 271, respectively, while those of normal SCC47 cells were 24.174% and 478, respectively. These results suggest that the pseudosubstrate inhibitor reduces the proliferation, invasion, and migration behaviors of SCC47 cells. In addition, after treatment with the pseudosubstrate inhibitor for 48 h, the apoptosis rate of SCC47 cells was increased (24.30%), compared with that of control cells (10.77%) ( Figure 3K). All the results suggested that PRKCZ could promote HPV+ HNSCC cell proliferation, invasion, and migration potential and inhibit apoptosis.

Methylation Degree of PRKCZ Was Increased by Partly DNMT1 via HPV E6
E6 and E7, two main well-known carcinogens of HPV, target host tumor suppressor genes or proteins to drive carcinogenesis and progression [32]. Thus, we investigated whether HPV affected the methylation degree of PRKCZ via E6 or E7. We compared the methylation degree of PRKCZ between Cal27 and Cal27+E6E7 cells. It was suggested that after transfection with E6 and E7 lentivirus plasmids, PRKCZ methylation levels were increased, whereas PRKCZ mRNA levels decreased ( Figure 4A). To further determine whether E6 or E7 was more powerful, E6 and E7 levels in SCC47 were inhibited using respective siRNAs. Additionally, the most effective siRNAs for E6 and E7 were selected by qPCR and WB experiments ( Figure 4B). We found that PRKCZ methylation levels decreased more obviously when E6 was inhibited. Accordingly, the mRNA level increased even more ( Figure 4C).
The DNMT family has been demonstrated to act as irreplaceable and crucial enzymes involved in DNA methylation [33,34]. Thus, we hypothesized that HPV mediates PKRCZ methylation by regulating the expression of the DNMT family. The expression levels of several main members of the DNMT family (DNMT1, DNMT3a, DNMT3b) were detected after E6 or E7 was inhibited in SCC47 cells. The results showed that when E6 was inhibited, DNMT1 levels decreased, whereas changes in the levels of DNMT3a and DNMT3b were not obvious. The same phenomenon was observed when E7 was inhibited. In addition, the role of E6 on DNMT1 was more potent than that of E7 ( Figure 4D,E). Overall, it was demonstrated that in SCC47 cells, HPV at least partly regulated DNMT1 expression through E6, leading to an increase in PRKCZ methylation and a decrease in PRKCZ mRNA expression.   Results were presented as means ± SD. ns > 0.05, * p < 0.05, ** p < 0.01, *** p < 0.001.

PRKCZ Regulated Rap Signaling Pathway to Mediate EMT of HPV+ HNSCC Cells
To investigate the internal mechanisms by which PRKCZ influences HPV+ HNSCC cell biological behaviors, we performed a KEGG enrichment analysis of PRKCZ. The results suggested that among the pathways in which PRKCZ is involved, the Rap signaling pathway was ranked 7th among the top 20 KEGG pathways for abnormal gene methylation enrichment (Figures 5A and S6). In addition, according to the results described above, we found that HPV+ HNSCC cell migration and invasion were significantly inhibited after silencing PRKCZ expression. It is well known that EMT is a vital process in tumor invasion and metastasis [35,36]. Thus, we hypothesized that PRKCZ regulated the Rap signaling pathway to mediate EMT, ultimately influencing HPV+ HNSCC cell migration and invasion.
First, we detected the expression changes of several key genes involved in the Rap signaling pathway (Epac1, Rac1, Rap1A, Rap1B, Cdc42, RhoA, and RhoB) in HPV+ HN-SCC cells after PRKCZ expression was blocked. The qPCR results suggested that after PRKCZ silencing, Epac1, Rac1, Rap1A, Rap1B, Cdc42, and RhoB mRNA expression decreased, whereas the change in RhoA was not significantly different. In addition, Cdc42 mRNA levels changed most dramatically ( Figure 5B). Cdc42 protein levels were also inhibited ( Figure 5C). Next, the expression levels of several key genes involved in EMT were detected. After inhibiting PRKCZ, E-cadherin mRNA and protein expression levels increased, whereas N-cadherin and Vimentin mRNA and protein expression levels decreased ( Figure 5D). This result suggested that silencing PRKCZ could inhibit the occurrence of EMT in HPV+ HNSCC cells, thus inhibiting cell migration and invasion.

Blocking PRKCZ Delayed the Tumor Growth of HPV16-E6/E7 Transgenic Mice
The 4-NQO supplied in drinking water is widely used as a carcinogen to induce cancers of the oral cavity (HPV-) in rodents due to the similar histological and molecular changes observed in human oral carcinogenesis [37]. In our experiment, we used 4NQOinduced HPV16-E6/E7 transgenic mice tongue tumor model, which was obtained from Rosa26-E6-E7 constitutive knock-in homozygous C57BL/6 mice. The subsequent breeding and identification of transgenic mice were completed by our research group in the early stages [38,39]. After drinking water with 4-NQO for 10 weeks followed by distilled water for an additional 10 weeks, PKC-ζ pseudosubstrate inhibitor and PRKCZ siRNA were injected near the tumor at 3-day intervals, and the mice were assessed after 3 weeks ( Figure 6A). The control group was treated with DEPC water. HE staining results demonstrated the successful establishment of tongue squamous carcinoma via 4-NQO ( Figure 6B). Tumor lesions in the tongue were more obvious in the DEPC-treated group compared with the inhibitor-treated and siRNA-treated groups. In addition, tongue lesions were slightly more apparent in the inhibitor-treated group compared with the siRNA-treated group, but the difference was not statistically significant. The tumor size was (1.852 ± 0.43) mm 3 on average in inhibitor-treated mice and (1.799 ± 0.630) mm 3 in siRNA-treated mice, whereas that in DEPC-treated mice was (5.108 ± 1.962) mm 3 ( Figure 6C). These results revealed that blocking PRKCZ could weaken the formation of tongue lesions induced by 4-NQO in HPV16-E6/E7 transgenic mice. The mRNA and protein expression (include densitometry analysis of WB results) changes of E-cadherin, N-cadherin, and Vimentin in HPV+ HNSCC cells after PRKCZ silencing. All assays were carried out in triplicate. Results were presented as means ± SD. ns > 0.05, * p < 0.05, ** p < 0.01. First, we detected the expression changes of several key genes involved in the Rap signaling pathway (Epac1, Rac1, Rap1A, Rap1B, Cdc42, RhoA, and RhoB) in HPV+ HNSCC cells after PRKCZ expression was blocked. The qPCR results suggested that after PRKCZ silencing, Epac1, Rac1, Rap1A, Rap1B, Cdc42, and RhoB mRNA expression decreased, whereas the change in RhoA was not significantly different. In addition, Cdc42 mRNA levels changed most dramatically ( Figure 5B). Cdc42 protein levels were also inhibited ( Figure 5C). Next, the expression levels of several key genes involved in EMT were detected. After inhibiting PRKCZ, E-cadherin mRNA and protein expression levels increased, whereas N-cadherin and Vimentin mRNA and protein expression levels decreased ( Figure 5D). This result suggested that silencing PRKCZ could inhibit the occurrence of EMT in HPV+ HNSCC cells, thus inhibiting cell migration and invasion.

Blocking PRKCZ Delayed the Tumor Growth of HPV16-E6/E7 Transgenic Mice
The 4-NQO supplied in drinking water is widely used as a carcinogen to induce cancers of the oral cavity (HPV-) in rodents due to the similar histological and molecular changes observed in human oral carcinogenesis [37]. In our experiment, we used 4NQO- (D) The mRNA and protein expression (include densitometry analysis of WB results) changes of E-cadherin, N-cadherin, and Vimentin in HPV+ HNSCC cells after PRKCZ silencing. All assays were carried out in triplicate. Results were presented as means ± SD. ns > 0.05, * p < 0.05, ** p < 0.01, *** p < 0.001.
Then, IHC staining demonstrated that Cdc42 expression decreased in tumors from the inhibitor-treated and siRNA-treated groups, whereas E-cadherin levels increased ( Figure 6D). These results indicated that suppressing PRKCZ could delay 4NQO-induced tongue tumor growth by promoting Cdc42 expression and down-regulating E-cadherin expression.
the inhibitor-treated and siRNA-treated groups. In addition, tongue lesions were slightly more apparent in the inhibitor-treated group compared with the siRNA-treated group, but the difference was not statistically significant. The tumor size was (1.852 ± 0.43) mm 3 on average in inhibitor-treated mice and (1.799 ± 0.630) mm 3 in siRNA-treated mice, whereas that in DEPC-treated mice was (5.108 ± 1.962) mm 3 ( Figure 6C). These results revealed that blocking PRKCZ could weaken the formation of tongue lesions induced by 4-NQO in HPV16-E6/E7 transgenic mice.

Discussion
The importance of DNA methylation in host-viral interactions and cancer development is self-evident [40,41]. The methylation landscapes between HPV+ and HPV-HNSCC are quite different. In this study, we identified PRKCZ as a differentially methylated gene between HPV+ and HPV-HNSCC using a HumanMethylation450K BeadChip array (Illumina). We addressed, for the first time, the existence of an HPV-mediated PRKCZ hypermethylation signature in HPV+ HNSCC, which was largely due to the role of E6 in stimulating DNMT1 expression. Moreover, we suggest that PRKCZ plays a tumorpromoting role in HPV+ HNSCC via Cdc42, inhibiting E-cad expression, promoting N-cad and Vim expression, and finally contributing to EMT.
HPV has a specific effect on shaping the DNA methylome of HNSCC [42]. In addition, the accumulation of abnormal epigenetic alterations may result, at least partly, in the differences between HPV+ HNSCC and HPV-HNSCC [43]. HPV+ HNSCC and HPV-HNSCC show different DNA methylation profiles. HPV-HNSCC is primarily characterized by genome-wide hypomethylation, whereas HPV+ counterparts seem to have increased maintenance of global methylation and exhibit a greater connection with promoter hypermethylation [44]. In our study, we found that compared with HPV-HNSCC, HPV+ HNSCC exhibited PRKCZ hypermethylation, which was associated with decreased PRKCZ mRNA and PKC-ζ protein expression. Additionally, decreased PRKCZ expression was responsible for the impaired HPV+ HNSCC cell proliferation, migration, and invasion. Thus, we suggest that the impaired malignant behaviors of HPV+ HNSCC cells due to PRKCZ hypermethylation might explain why HPV+ HNSCC has a better prognosis. In addition, PRKCZ hypermethylation was associated with the pathological grade of HNSCC patients, but not with sex, age, T stage, lymph node metastasis, or tumor recurrence. The results from cell experiments and clinical specimens were slightly contradictory. There are two possible explanations. First, the insufficient number of clinical specimens in our study might explain the difference. The number of patients enrolled in our study was limited (40 patients), and the follow-up date was short (the longest was 78 months, the shortest was 13 months, and the average was 35.6 months). Second, the tumor microenvironment is very complex. In vitro cell experiments cannot accurately reflect changes in the body. Whether the effect of PRKCZ on some important tumor stromal cells, such as cancer-associated fibroblasts, can affect or even counteract its effect on tumor cells, which needs further research.
PRKCZ was the only gene associated with HPV infection among 50 selected genes with the most obvious methylation differences between HPV+ HNSCC and HPV-HN-SCC according to array data. This finding was further supported by data obtained from TCGA, clinical specimens, and cell lines, revealing that PRKCZ was hypermethylated in HPV+ HNSCC. However, the difference in the level of methylation observed in TCGA data was not significant, which may be due to a discrepancy in the PRKCZ methylation degree and HPV status. PRKCZ hypermethylation plays a vital role in several diseases, such as type 2 diabetes mellitus [45], CTCF deletion syndrome [46], and postmenopausal osteoporosis [47]. However, to the best of our knowledge, only one study has reported the role of PRKCZ in cancers. In COX-2-induced hepatocellular carcinoma, PRKCZ was hypermethylated and associated with reduced gene expression. However, the relationship between PRKCZ methylation and the biological behaviors of hepatocellular carcinoma cells has not been thoroughly studied [48]. Thus, it is the first study to focus on the role of PRKCZ methylation in cancers, and a detailed mechanistic study is provided.
HPV-mediated cancer development is inseparable from the continuous expression of the viral E6 and E7 oncogenes [49]. In addition to the role of E6 in proteolytic degradation of tumor suppressor p53 and E7 in interference with the activity of tumor suppressor pRb [50], the deregulation of E6 and E7 in other carcinogenic mechanisms is also important. To verify the effect of E6 and E7 on PRKCZ methylation, we used E6 and E7 lentivirus plasmids to transfect the HPV-HNSCC cell line and E6 or E7 siRNA to transfect the HPV+ HNSCC cell line. Our results revealed that the expression of HPV E6 and E7, especially E6, resulted in PRKCZ promoter hypermethylation by stimulating DNMT1 expression, which led to decreased expression of PRKCZ in HPV+ HNSCC cells. Additionally, consistent with our results, HPV E6 was found to increase DNMT activity and expression, ultimately resulting in impaired E-cadherin expression in cervical cancer [51].
In our study, we used two methods to inhibit PRKCZ expression, including siRNA and a pseudosubstrate inhibitor, which is a specific and efficient inhibitor of PKC-ζ [24]. However, some studies have proposed that its specificity may not be attributed to the high degree of sequence homology between PKC-ζ and PKC-ι [52]. Thus, in our study, the application of a pseudosubstrate inhibitor was employed. Our results suggested that PRKCZ acted as a potential tumor promoter for HPV+ HNSCC. Inhibiting PRKCZ expression led to impaired proliferation, migration, and invasion abilities, and increased the apoptosis rate of HPV+ HNSCC cells. In addition, blocking PRKCZ expression delayed the tumor growth of HPV16-E6/E7 transgenic mice. These results were consistent with previous studies in prostate cancer, lung cancer, and HNSCC, where PRKCZ expression was higher than that in respective normal tissues. After knocking down PRKCZ expression via siRNA, prostate cancer cells presented impaired proliferation, migration, and tumorigenesis, which resulted from changed voltage-gated K+ channel activity [53]. The application of a PKC-ζ specific myristoylated pseudosubstrate inhibitor impaired the chemotactic capacity of non-small cell lung cancer cells [54]. PRKCZ acted as a potent tumor promoter in HNSCC by influencing DNA synthesis [55], holoprotein integrity and telomerase activity [56], and metastatic activity [57]. However, the role of PKC-ζ in oncogenesis remains controversial. Some studies have proposed the exact opposite viewpoints. They suggested that PRKCZ acts as a potential tumor suppressor in intestinal tumors, prostate cancer, and lung cancer via involvement in metabolism reprogramming [58], suppressing the polarization of macrophages to the M2 phenotype [59], and inhibiting EMT [60]. The reasons for the differences among these results remain unknown and must be solved.
Cdc42 is a Rho GTPase responsible for the dynamic reorganization of the actin filament system [61]. Cdc42 plays a pivotal role in migration, invasion, and EMT via the regulation of cytoskeletal and microtubule dynamics [62]. For example, Cdc42 is overexpressed and activated in hepatocellular carcinoma, correlating with reduced E-cadherin expression and an enhanced EMT phenotype [63]. In addition, Cdc42 functions together with Par-3, atypical PKC, and Par-6 in cell polarity and asymmetric cell division [64]. In nasopharyngeal cancer cells, Cdc42 modulates PKC-ζ expression to control telomerase activity. Here, our results suggest that PRKCZ mediates the level of Cdc42 to regulate EMT in HPV+ HNSCC cells. In addition, HPV18 E6 activates the small GTPase Rac1 and subsequently contributes to the E6 stimulation of NFκB in cervical cancer [65]. Depletion of PKC-ζ inhibits nuclear localization of NFκB-p65, inhibiting EMT, invasion, and metastatic progression [66]. Hence, NFκB might be necessary for PRKCZ mediating Cdc42 to regulate EMT in HPV+ HNSCC cells. More should be performed in the future.
In our HPV16-E6/E7 transgenic mouse model of tongue tumors, we found that tumor lesions in the tongue were much smaller in the inhibitor-treated and siRNA-treated groups compared with the DEPC-treated group, which might be due to decreased Cdc42 expression and increased E-cad expression. In our animal experiments, we did not demonstrate whether demethylation therapy with PRKCZ was effective in suppressing tumor growth because this function had been reported by Biktasova A et al. These researchers proposed that the application of 5-aza could inhibit HPV+ HNSCC tumor growth and mouse blood vessel invasion by repressing the expression and activity of matrix metalloproteinases [67]. Therefore, we focused on the effect of blocking PRKCZ treatment rather than repetitively testing the effect of demethylation therapy. However, it will be interesting to explore the effect of combination therapy between blocking PRKCZ and demethylation therapy.
Overall, we illustrated for the first time that PRKCZ was hypermethylated in HPV+ HNSCC. In addition, PRKCZ acts as a tumor promotor that induces HPV+ HNSCC cell proliferation, invasion, and migration and prevents apoptosis. Mechanistically, HPV E6 might promote the expression of DNMT1 to induce PRKCZ hypermethylation, and PRKCZ increased Cdc42 levels to promote EMT in HPV+ HNSCC cells. These results indicated that PRKCZ hypermethylation might represent one possible reason why HPV+ HNSCC has a better prognosis and that blocking PRKCZ might be a new strategy for HPV+ HNSCC.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/cancers14174151/s1, Figure S1: The KEGG pathways of Human papillomavirus infection, Figure S2: The relations between PRKCZ methylation status and relevant clinical and pathological parameters of HPV+ and HPV-patients, Figure S3: The results of wound healing assays between PRKCZ siRNA and control siRNA groups, Figure S4: The results of transwell invasion assays between PRKCZ siRNA and control siRNA groups, Figure S5: The results of wound healing and transwell invasion assays between pseudosubstrate inhibitor and normal groups, Figure S6: The KEGG pathways of RAP1 signaling pathway, Table S1: The top 25 abnormally hypermethylated genes in HPV+ HNSCC tissue compared with HPV-HNSCC tissue, Table S2: The top 25 abnormally hypomethylated genes in HPV+ HNSCC tissue compared with HPV-HNSCC tissue, Table S3: Clinicopathological characteristics of patients whose tissue specimens were used for qMSP, Table S4: Clinicopathological characteristics of patients whose tissue specimens were used for