Comprehensive analysis of the expression and prognosis for CDCAs in head and neck squamous cell carcinoma

Head and neck squamous cell carcinoma (HNSCC), a tumor included oral cavity, lips, larynx, oropharynx, and the nasopharynx et al. The cell division cycle-associated (CDCA) protein family (CDCA1-8) critical for normal cell function and cancer cell proliferation. We explored the mutation signatures and expression levels of various CDCAs in detail in HNSCC. A comprehensive bioinformatics analysis pipeline based on copy number and gene expressions data from patients with HNSCC in order to given new insights into the possible functions and distinct prognostics that underlie CDCAs regulation. We compared the transcriptional expression of CDCAs in HNSCC and found significantly elevated mRNA expression of CDCA1-8 in HNSCC tissues across multiple datasets. We also found CDCA5/6/8 are over-expressed both transcriptionally and translationally in patients with HNSCC. Our results suggested that that mRNA levels of CDCA1/2/4/7 related to the prognosis and can be used as a new useful biomarker for predicting the survival of HNSCC patients. The top 5 CDCAs neighboring gene alterations in HNSCCs were found in MYC, STAG1, RAD21, KLHL9 and NDC80. Multivariable Cox proportional hazard model also showed that CD8+ T cells were higher (P<0.05) in HNSCC-HPV-pos patients and that this was related to CDCA1/2/3/4/5/7. This study utilizes online tools to conduct specific gene analyses from free open databases, but our study requires more large-scale genomics research and basic research.


Introduction
Head and neck squamous cell carcinoma (HNSCC), a tumor included oral cavity, lips, larynx, oropharynx, and the nasopharynx et al [1]. The tumor with a yearly incidence of over 650,000 new diagnosis and 90,000 decease worldwide [2]. The risk factors for HNSCC involve in smoking, alcohol drinking and virus infection, such as human papilloma virus (HPV) [3]. Unfortunately, there is insufficiency of symptoms at the early stage of the cancer, causing most patients with HNSCC to be diagnosed at the progressive stages. Consequently, the survival rate of 5-year is below 50% and patients that suffer from local recurrence and metastasis have an even a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 lower survival rate of 35% [4]. When in the advanced stage, therapeutics can affect organ structures function that related to swallowing and speaking, leading to a decline in the patient's quality of life [5,6]. The occurrence of HNSCC is a complicated mechanism that involves multiple molecules. Ni et al [7] found that HPV and HPV16 DNA was detected in 26.4% and 71% of the 303 HNSCCs, respectively. Thus, prophylaxis against HPV infection may help reduce the incidence of this disease. A recent study proposed that zeste homolog 2 (EZH2) regulates epithelial-to-mesenchymal transition (EMT), metastasis and tumor invasion in HNSCC by regulating the STAT3/VEGFR2 axis [8]. Valenti et al [9] reported that miR-205-5p's can impact genomic instability in HNSCC by selectively targeting the DNA damage response (DDR) genes RAD17 and BRCA1. In spite of the advances that have been made in the past decades, which include combining chemotherapy, radiation, and surgery, many patients still experience tumors recurrences and metastasis even received treatment, which leads to therapeutic failure [10].
The cell division cycle-associated (CDCA) protein family (CDCA1-8) not only necessary for normal cell function, but also plays a key role in cancer cell proliferation. Some studies have highlighted that abnormal expression of cell cycle regulatory proteins may cause cancer. Phan et al [11] found that CDCA3/5/8 are significantly higher in breast cancer tissue than control tissue, leading to a dramatic reduction in patient survival among breast cancer patients. A clinical trial that was now performed with castration resistant prostate cancer (CRPC) by a CDCA1 peptide vaccination was found to effectively induce peptide-specific CTLs for CRPC patients [12]. In addition, siRNA-mediated knockdown of CDCA1 in oral cavity carcinoma (OCC) tumor cells was found to induce a significant apoptotic response [13]. The CDCA1 protein family is often co-expressed with many other cell cycle regulators, involving CDC23/CDC7/CDC2/ MCAK/MKI67 and topoisomerase II, to regulate tumor cell growth [14]. To date, the mechanism by which CDCAs are activated or deactivated in the development and progression of HNSCC still remains unclear. We explored the mutation signatures and expression levels of various CDCAs in detail using a comprehensive bioinformatics analysis pipeline based on copy number and gene expressions data from patients with HNSCC in order to offer more knowledge into the potential functions and distinct prognostics that underlie CDCAs regulation. We also discuss the opportunities and challenges in using these to derive clinical benefit for HNSCC patients.

ONCOMINE database and Human Protein Atlas
The HNSCC mRNA expression data of CDCAs were obtained from the Oncomine [15], which is a database that involve 86,733 samples and 715 gene expression data sets. Oncomine as well the largest oncogene chip database as well as incorporated data mining database. This analysis was based on a number of prior HNSCC researches. The level of CDCAs was evaluated in HNSCC tissue and in control tissue. P<0.05 considered statistically significant. All the Data from Genomic Data Commons Data Portal. The Human Protein Atlas (HPA) is an online tool that included immunohistochemistry expression data for distribution and expression of proteins across 20 cancer tissues, 48 human normal tissues, 47 cell lines, and 12 blood cells [16]. We used immunohistochemistry images to directly compare protein expression of CDCAs among normal and cancer tissues. GEPIA[17] is an interactive online database which allowed users to found RNA seq expression data or samples based on the Genotype Tissue Expression projects (GTEx) and The Cancer Genome Atlas (TCGA). Meanwhile, GEPIA also offers customizable functions such as profiling based on pathological stage of cancer, type of cancer, survival analysis, correlation analysis and similar gene identification. UALCAN [18] is a website that helps analyze, integrate and discover cancer transcriptomic data and deep analyses of TCGA gene expression information. One of the portal's highlight characteristic is that it can determined biomarkers or to perform in silico analysis of potential candidate genes of interest to assess expression in various subgroups, such as age, gender, race, and grade.

Kaplan-Meier plotter and cBioPortal
Kaplan-Meier plotter [19] was used to predicted the prognostic significance of different CDCAs in HNSCC. The database includes RNA-seq information based on TCGA and GEO. By setting different parameters, different subgroups can explore including patients with various pathologies, treatment ways, and data sets. The cBioPortal [20] is a free asset that can download large-scale cancer genomics data sets encompassing 245 cancer researches. Using cBioPortal to explored CDCAs genetic alterations in CDCAs. An interaction network of the CDCAs and the co-expressed genes were also analyzed. GO and KEGG functions of CDCAs mutations and top 50 genes that were obviously linked to CDCAs mutations were performed via DAVID online tool.

TIMER analysis
TIMER [21] is a useful tool for systematic found of immune infiltrates across different cancer types. Gene module can explore correlation among CDCAs and the abundance of immune infiltrates in HNSCC. The survival module was used to draw Kaplan-Meier plots for immune infiltrates and CDCAs for visualization of survival differences.
We next analyzed the protein expression of CDCAs and the result indicated low protein expression of CDCA5/6/8 in normal tissues and high protein expression in tumor tissues. In addition, results also indicate medium expression of CDCA2 in normal tissues and high

Clinical subgroup analysis
We first using the GEPIA dataset to compared the expression of CDCAs among cancer and normal tissues. Our results indicate that the CDCA1/2/3/4/5/6/8 are significantly higher in HNSCC tissues (Fig 3). Next, we performed subgroup analysis of multiple clinical pathological features using the TCGA database. Subgroup analysis by age, indicated that transcriptional levels of CDCAs were higher in HNSCC patients when compared to healthy individuals.

PLOS ONE
Additionally, subgroup analysis by HPV status analysis; gender subgroup, and tumor grade demonstrated that CDCAs were significantly higher in HNSCC patients across all subgroups (Fig 4).

Prognostic analysis
Next, we tried to explore the prognostic significance of CDCAs in HNSCC patients, data for which was obtained from publicly available online datasets. The results are shown in

Function analysis of CDCAs in HNSCC
We explored CDCAs alterations and networks using the cBioPortal. 50 neighboring genes that were found to be significantly linked to CDCAs mutations. Among the 528 HNSCC tumor samples that were sequenced, genetic alterations were found in 90 samples with a mutation rate of 18%. CDCA5 was ranked as the most mutated gene among CDCAs with mutation rates of 5%. We also showed the network for CDCAs and the 50 most frequently altered neighboring genes (Fig 6). The top 5 CDCAs neighboring gene alterations in HNSCCs were found in MYC, STAG1, RAD21, KLHL9 and NDC80 (Table 2). Next, we analyzed the functions of CDCAs and these 50 genes using GO and KEGG (S1 Appendix). GO analysis indicate that changes in biological processes included enrichment in sister chromatid cohesion, cell division, mitotic nuclear division, gene silencing by RNA, and protein sumoylation among others. Molecular function was mainly enriched in protein heterodimerization activity, microtubule plus-end binding, protein phosphatase type 2A regulator activity, nucleocytoplasmic transporter activity, and protein binding. Changes in cell component were largely enriched in condensed chromosome kinetochore, chromosome, centromeric region, kinetochore, cytosol, nucleosome and others. Pathway enrichment analysis according to KEGG was mainly enriched in PI3K-Akt and AMPK signaling pathway, endometrial cancer, acute myeloid leukemia, colorectal cancer, central carbon metabolism in cancer, transcriptional misregulation in cancer, and chronic myeloid leukemia.

Immune infiltrates of CDCAs in HNSCC
There is a statistically significant correlation between CDCAs expression in HNSCC and abundance of immune infiltrates (P<0.05, Fig 7). We explored the difference in cumulative survival between HNSCC, HNSCC-HPV-pos and HNSCC-HPV-neg and found that the HNSCC-HPV-pos subgroup showed significantly higher B cells, CD8+ T cells and neutrophil immune infiltrates, (P<0.05) which was related to CDCAs levels. This indicates that these immune cell infiltrations significantly affect prognosis. Therefore, it is worth further

PLOS ONE
researching and exploring this association (Fig 8). Multivariable Cox proportional hazard model also showed that CD8+ T cells immune infiltrates were significant higher (P<0.05) in HNSCC-HPV-pos patients and that this was related to CDCA1/2/3/4/5/7 Table 3.

Discussion
Though certain CDCAs have been shown to play a critical role in tumor, the specific roles of CDCAs in HNSCC remains unclear. Thus, we first explored the mutational, gene expression, and prognostic landscape of various CDCAs in patients with HNSCC. We found higher mRNA expression across all CDCAs, and the expression of CDCAs was significantly linked to patients' individual cancer stages. Moreover, we explored the immune status of HNSCC patients which can potentially help guide the development of novel therapies and to improve response to immunotherapy.
A growing number of studies have shown that CDCAs are highly expressed in tumors and have a role in regulating tumor cell cycle, promoting tumor cell proliferation, and reducing tumor cell apoptosis, which results in poor prognosis. CDCA1, also known as NUF2, codes for a protein that is essential for nuclear division and microtubule stabilization [27]. Tokuzum et al [27] reported that CDCA1-specific siRNA inhibits the cell proliferation of WM115 and SKMEL2 cells, but does not reduce the invasion activity or migration in malignant melanoma patients. Tomita et al [28] demonstrated that the existence of CDCA1-specific Th cell responses

PLOS ONE
in HNSCC patients underline the potential utility of CDCA1-LPs for propagation of both CDCA1-specific CTLs and Th cells. Similarly, Kaneko et al [29] found that knockdown of CDCA1 and KNTC2 genes in colorectal cancer cells better inhibits tumor cell growth. Our results show that CDCA1 is highly expressed in HNSCC tissues, and CDCA1 is significantly correlated to patients' survival and abundance of immune infiltrates. Moreover, our cumulative survival analyses show that CD8+ T cell immune infiltrates significantly affect the prognosis of these patients. Thus, it is worth further exploring this association. CDCA2 is a nuclear protein that binds to protein phosphatase 1γ (PP1γ) and participated in DNA damage during cell cycle [30]. Moreover, CDCA2 modulates phosphorylation of the primary mitotic histone H3 in a PP1-dependent manner [31]. Some studies indicated that CDCA2 act for a very powerful prognostic marker for poor patient survival and malignancy in cancers such as neuroblastoma, lung adenocarcinoma, and oral squamous cell carcinoma tissue [32][33][34]. A recent study found that overexpression of CDCA2 may target CCND1 to promote colorectal cancer cell proliferation and tumorigenesis via activation of the PI3K/AKT pathway [35]. Interestingly, in our analysis of 50 neighbor genes that were significantly related to CDCAs mutations, the KEGG results showed a high enrichment of genes involved in the PI3K-Akt and AMPK signaling pathway. Thus, our study provides critical information that can be utilized for future studies.
CDCA3 is part of the SKP1-Cullin-RING-F-box (SCF) ubiquitin ligase (E3) complex, which degrades the endogenous cell cycle inhibitor WEE1, thereby regulating cell cycle [36]. CDCA3, through regulation by specificity protein 1 (SP1) and hypomethylation of its gene body, affects gastric cancer (GC) cell proliferation and invasion [37]. In addition, CDCA3 activated the Ras signaling pathway to facilitate cell proliferation in vitro and in vivo in GC cells [38]. Another study also found that HoxB3 can bind to the CDCA3 promoter region and transactivate CDCA3 expression to induce prostate cancer progression [39]. Our results show that HNSCC tissue highly express CDCA3. To date, however, no studies have investigated the connection between HNSCC and CDCA3 and more research is needed.
CDCA4, also known as HEPP/SEI-3/TRIP-Br3 is a target gene of transcription factor E2F, was discovered in 2001 and has shown to be related to the regulation of genes regulating the growth and differentiation of hematopoietic stem and progenitor cells [40]. Xu et al [41] found that CDCA4 enhanced proliferation and reduced apoptosis in the MCF-7/ADM breast cancer cells in vitro. A recent study also suggested that CDCA4 may be involved in regulating triple negative breast cancer (TNBC) progression [42]. Results from our study indicate that CDCA1/ 2/4/7 may serve as novel biomarkers for prediction of HNSCC patients' survival. CDCA5 is a critical regulator of sister chromatid condensation and separation during cell division [43]. CDCA5 could promote proliferation, migration, invasion, apoptosis resistance and decrease chemosensitivity to cisplatin in esophageal squamous cell carcinoma (ESCC) cells [44]. Moreover, CDCA5 was shown to be upregulated in hepatocellular carcinoma (HCC) tissues compared to paracancerous tissues, is negatively correlated with patient survival and associated with cell abnormalities via upregulation of the AKT pathway [45]. CDCA6 also known as CBX2, encodes a component of the polycomb multiprotein complex. CDCA6 depletion abrogated cell viability and induced caspase 3-mediated apoptosis in metastatic prostate cancer cell lines [46]. One study also found that CDCA6 upregulation and amplification was significantly related to lower overall survival and metastatic progression across many cancer types [47]. While our study shows high expression of CDCA6 in HNSCC tissues, though there is a paucity of studies in literature that have investigated this connection. Thus, there is a need to conduct research on the role of CDCA6 in HNSCC. CDCA7, also known as JPO1, is considered to be a c-Myc target gene that is involved in c-Myc-mediated cell transformation [48]. One study found that depletion of CDCA7 extremely minimize the tumorigenicity and colonization capacities of TNBC cells in vivo [49]. Jenness et al reported that the HELLS-CDCA7 complex possesses nucleosome remodeling activity [50]. Another study discovered a role for CDCA7 in Centromeric Instability and Facial Anomalies syndrome, a life-threatening immunodeficiency [51]. In addition, AKT signaling to CDCA7 could alter MYC-dependent growth and transformation, contributing to tumorigenesis [52]. CDCA8, also known as Borealin/DasraB, encodes a component of the chromosomal passenger complex and is essential for chromatin-induced microtubule stabilization and spindle formation [53]. One study also reported that CDCA8 was significantly linked to poor prognosis in patients with cutaneous melanoma [54], breast cancer [55], colorectal cancers [56] and lung cancer [57]. Our results suggest that CDCA5/6/8 are higher expressed in patients with HNSCC, both transcriptionally and translationally. Overall, the function and pathways of CDCAs and their 50 frequently altered neighboring genes showed that these genes were mainly enriched in changes in cell division, mitotic nuclear division, protein binding and other cell functions. KEGG pathway analysis showed an enrichment in PI3K-Akt and AMPK signaling pathway, as well as some cancers and cancer-related signaling pathway. Thus, modifications to CDCAs is associated with post-transcriptional regulation, which is largely linked to protein translation.
To date, no studies have investigated the role of CDCAs and the connection between tumor infiltrating immune cells and HNSCC. We first explored the difference between cumulative survival between HNSCC, HNSCC-HPV-pos and HNSCC-HPV-neg tumors and found that HNSCC-HPV-pos group had a significantly higher infiltration of B cells, CD8+ T cells and neutrophil cells (P<0.05), which was positively related to CDCAs expression. This indicates that immune cells may have a significant effect on the prognosis of this disease. Therefore, it is worth further investigation in subsequent studies. There were several limitations, one being that all the data in our study was based on online free databases. Additionally, our study does not provide precise clinical information. Hence, more studies are needed to prove our findings. Another limitation is that we did not assess the possible therapeutic and diagnostic roles of CDCAs as the histological types of HNSCC as well as the multiple anatomical sites of the cancer varies widely. Thus, future studies are needed. Finally, we were incapable to contrast the differences in function of CDCAs among HPV-positive and HPV-negative in HNSCC due to insufficient data, though we plan to investigate this in the future.

Conclusion
Our results indicate that CDCAs play a key role in the HPV-pos HNSCC patients. This study made use of online free tools to perform target gene analyses on HNSCC from open databases, which enables more genomics research and subsequent functional exploration.