Upregulation of CEP55 Predicts Dismal Prognosis in Patients with Liver Cancer.

Purpose This study was performed to investigate the association of CEP55 expression with liver cancer and explore potential underlying mechanisms. Materials and Methods. Data obtained from The Cancer Genome Atlas (TCGA) was used to investigate CEP55 expression, its prognostic value, the potential mechanisms of its upregulation, CEP55-related pathways, and its biological functions in liver cancer. Data from Gene Expression Omnibus (GEO) and International Cancer Genome Consortium (ICGC) was used to validate survival analysis. The correlation between CEP55 and tumor-infiltrating immune cells (TIICs) in liver cancer was determined by using Tumor Immune Estimation Resource (TIMER). Results CEP55 was significantly overexpressed in the liver tumor sample compared to the adjacent normal liver sample. High CEP55 expression was significantly associated with histological grade, advanced stages, histological type, high T classification, and survival status. High CEP55 expression was significantly related to dismal prognosis compared with low CEP55 expression, which was validated by the GSE54236 dataset and ICGC database. Meanwhile, CEP55 was identified as the risk factor to independently predict overall survival (OS) for patients with liver cancer upon multivariate analysis. Enrichment analysis indicated that cell cycle, DNA replication, pathways in cancer, mTOR signaling pathway, and VEGF signaling pathway were significantly enriched in the high CEP55 expression group. In addition, the CEP55 expression was significantly related to the infiltration level of B cells, CD4+ T cells, CD8+ T cells, macrophages, neutrophils, and dendritic cells in hepatocellular carcinoma (HCC). CEP55 methylation level was negatively correlated to its mRNA expression. And patients with CEP55 hypermethylation and low expression can achieve a better prognosis than those with CEP55 hypomethylation and high expression. Conclusion CEP55 may serve as a candidate treatment target for it is a determinant of prognosis and immune infiltration in liver cancer patients. DNA hypomethylation might contribute to the overexpression of CEP55 in liver cancer.


Introduction
Liver cancer ranks the 6 th place in terms of global tumor incidence, and it is the 4 th leading cause of cancer-related deaths [1]. Hepatocellular carcinoma (HCC), one of the frequently seen primary liver tumors, occupies about 80% of liver cancers, followed by cholangiocarcinoma (CCA), accounting for approximately 15%. Worldwide, the highest liver cancer morbidity is reported in Asia and Africa. Approximately 75% of liver cancer occurs in Asia, and China accounts for more than half of the total cases in the world. Aflatoxin exposure and chronic hepatitis B virus (HBV) infection are two major risk factors reported across most high-incidence countries in Asia and Africa [2]. Although great progresses have been achieved in diagnosing and treating liver tumor over the past few decades, prognosis for liver tumor remains dismal. Liver cancer has become the second most fatal tumor after pancreatic cancer with a 5-year survival rate of 18% [3]. Finding reliable predictors and potential therapeutic targets involved in the occurrence and development of liver cancer is urgently needed.
Centrosome proteins have long been considered to be scaffold proteins that regulate mitotic spindle and microtubule tissue, so they are essential to the cell cycle process [4].
Centrosome protein 55 (CEP55), also referred as FLJ10540 and C10orf3, was initially recognized as a key component of abscission, which is the final stage of cytoplasmic division that is responsible for regulating the physical disjunction of two daughter cells [5]. CEP55 is located in the centrosome during the whole cell cycle, in the mitotic spindle during mitosis, and in the midbody during the process of cytokinesis [6]. Cytokinesis is strictly controlled in the process of cell division and requires the recruitment of multicomponent subunits to the midbody in a CEP55-dependent manner [4,7,8]. CEP55 has been identified both as a cancer testis antigen and a tumor-associated antigen [9,10]. Cancer testis antigens are proteins normally expressed predominantly in the testes but which become more widely expressed in cancer [11]. Recently, some literature reported that CEP55 participated in promoting tumorigenesis and regulating the PI3K/AKT signaling pathway [12,13]. Growing evidence indicates the association between CEP55 upregulation and the development and progression of a variety of malignant tumors, including breast tumors, gastric tumors, and lung tumors [13][14][15]. The knockdown of CEP55 can significantly inhibit the viability and proliferation of a tumor cell and even lead to tumor cell death [16,17].
Though Li et al. [18] previously reported that the overexpression of CEP55 can result in poor prognosis of liver patients, their study is limited to HCC. In this study, we included both HCC and CCA patients from The Cancer Genome Atlas (TCGA) to investigate the prognostic significance of CEP55 expression and DNA methylation of CEP55 in liver cancer. The biological functions and signal pathways associated with CEP55 regulatory mechanism were also explored. In addition, we investigated the correlation between CEP55 and tumor-infiltrating immune cells (TIICs) in liver cancer using Tumor Immune Estimation Resource (TIMER).

Data Mining and Collection.
A flow chart of our study design is showed in Figure 1. The gene expression profile data, DNA methylation data, and the clinical information of liver cancer patients had been collected from TCGA (https:// cancergenome.nih.gov/). The clinical data included age, sex, survival time, survival status, clinical stage, histological grade, TNM classification, and histological type. Additionally, gene expression data of liver cancer patients with survival information was downloaded from the International Cancer Genome Consortium (ICGC) (http://icgc.org/). And GSE54236 dataset was obtained from the Gene Expression Omnibus (GEO) (https://www.ncbi.nlm.nih.gov/geo/). The expression levels of CEP55 in several common cancers were reviewed by using the TIMER database (https://cistrome.shinyapps.io/timer/). The CEP55 protein expression in normal liver samples and liver cancer samples was examined using immunohistochemistry (IHC) data from the Human Protein Atlas (HPA) database (http://www.proteinatlas.org/).

Functional Enrichment Analysis.
Genes significantly related to CEP55 expression were extracted using the Pearson correlation analysis (|R| ≥ 0:4,p < 0:001), and these genes were sent for Gene Ontology (GO) and Kyoto Encyclopedia  of Genes and Genomes (KEGG) enrichment analyses using the clusterProfiler package (version 3.10.0). Gene terms with p value < 0.05 and false discovery rate ðFDRÞ < 0:05 were considered significantly.
2.3. Gene Set Enrichment Analysis (GSEA). GSEA had been carried out for determining whether an a priori defined set of genes showed significant differential expression between the high and low CEP55 expression groups and to identify the potential mechanisms of CEP55 expression in liver cancer prognosis. Each analysis was run 1000 times gene set permutation. CEP55 expression level had been utilized to be a phenotype label. Gene sets with normalized (NOM) p value < 0.05 and FDR < 0:05 were considered significantly enriched.

Immune Infiltrates Correlation via TIMER.
Associations between CEP55 expression and TIIC infiltration levels were analyzed via TIMER platform, a web tool for gene-specific correlational analysis with TIICs [19]. TIICs included B cells, CD4+ T cells, CD8+ T cells, macrophages, neutrophils, and dendritic cells.
2.5. Analysis of DNA Methylation of CEP55. The relationship between CEP55 methylation and its mRNA expression was determined using the Pearson correlation analysis. Liver cancer patients were divided into the hypermethylation and hypomethylation groups according to the median CEP55 methylation level. The prognostic value of CEP55 methylation was evaluated by the Kaplan-Meier analysis.
2.6. Statistical Analysis. The R software 3.5.0 was utilized for statistical analysis. The expression of CEP55 in liver tumor samples and adjacent normal samples was compared using the Wilcoxon rank sum test. Patients with liver tumors were classified as the high or low expression group based on the median value of the CEP55 expression. The relationship between clinicopathological characteristics and CEP55 expression had been examined by logistic regression and Wilcoxon signed-rank test or Kruskal-Wallis test. Overall survival (OS) in the high CEP55 expression group was compared with that in the low CEP55 expression group by the Kaplan-Meier method. The p values were computed using a log-rank test. The receiver operating characteristics (ROC) curve was used to evaluate the prognostic value of CEP55 in 1, 3, and 5 years by using the survivalROC package. Univariate as well as multivariate Cox analysis was conducted for determining the relationships of CEP55 expression with OS as well as other clinical features (such as age, sex, clinical stage, histological grade, and TNM classification). A difference of p < 0:05 was deemed to be of statistical significance.   (Figure 2(a)). CEP55 expression in liver tumor samples and adjacent normal samples was also compared using data directly obtained from TCGA, and the result demonstrated that CEP55 expression was markedly increased within liver cancer samples compared to normal liver samples (p = 1:397e − 29) (Figure 2(b)). Furthermore, the expression of CEP55 in 58 paired liver cancer samples and adjacent normal samples was analyzed, and our findings suggested a marked overexpression of CEP55 for liver tumor (p = 2:649e − 17) (Figure 2(c)). To further examine the protein expression of CEP55, we retrieved the IHC staining data from the HPA (Figure 2(d)). In three normal liver samples, all hepatocytes had medium CEP55 staining, while all bile duct cells had low CEP55 staining. In tumor samples, most of the HCC samples had medium (5/7) CEP55 staining, and most of the CCA samples similarly had medium (3/5) CEP55 staining. The results showed that CEP55 had a significantly higher protein expression in the CCA samples than in the normal liver samples. Although CEP55 had a significantly higher mRNA expression in the HCC samples than in the normal liver samples, there was no significant difference in the protein expression between the HCC samples and normal liver samples.   (Table 3). Figure 7 presents the results of univariate and multivariate analysis illustrated by the forest plot.    (Figure 10(b)). However,   no significant relationship was detected between CEP55 expression and the infiltration level of any TIICs in CCA (Figure 10(a)).

Prognostic
Value of DNA Methylation of CEP55. The result of the Pearson correlation analysis showed that CEP55 methylation level was negatively correlated to its mRNA expression (R = −0:233, p = 4:814e − 7) (Figure 11(a)). However, there was no significant difference of OS rates between the CEP55 hypermethylation and hypomethylation groups (p = 0:818) (Figure 11(b)). Therefore, we combined the methylation and the expression of CEP55 and classified liver cancer patients into the CEP55 hypermethylation and low expression group and the CEP55 hypomethylation and high expression group based on the median values. The Kaplan-Meier analysis was performed to compare the OS rates between the two groups, and the result indicated that the CEP55 hypermethylation and low expression group can achieve a better prognosis than the CEP55 hypomethylation and high expression group (p = 0:011) (Figure 11(c)).

Discussion
In this study, we conducted bioinformatic analysis using high-throughput RNA-sequencing data obtained from TCGA. The results showed that the expression of CEP55 in liver tumor tissues was significantly higher than that in adjacent normal tissues. CEP55 overexpression showed correlation with the dismal prognosis in liver cancer patients. In addition, high CEP55 expression in liver cancer was positively related to advanced stage, high histological grade, and high T classification. CEP55 expression had been recognized to be an independent risk factor in predicting liver cancer patient survival upon univariate and multivariate analyses. CEP55, initially described as a midbody-related protein, has the size of 55 kDa, and it consists of 464 amino acids. It plays a critical role in regulating the physical cytokinesis [20]. CEP55 upregulation was previously reported to promote the migration and invasion of tumor cells and is related to the dismal prognosis for lung cancer [16], breast cancer [21], oral squamous cell carcinoma [22], cervical cancer [23], and osteosarcoma [24]. Such finding was consistent with our findings in liver cancer. The knockdown of CEP55 was found to inhibit tumor cell viability and proliferation [15,25]. The overexpression of CEP55 can lead to disordered cytokinesis and an increase in multinucleated cells, which is    an oncogenic feature in tumorigenesis [26]. According to the results in this study, CEP55 expression showed association with differentiation within liver tumor tissues. The higher the expression of CEP55 was, the poorer the differentiation of tumor cells. Similar results were found in prostate cancer and nasopharyngeal carcinoma [27,28]. These findings clearly suggest the significance of CEP55 in tumorigenesis.
To further explore the role of CEP55 within liver tumors, enrichment analysis was performed and the results showed that cell cycle, DNA replication, pathways in cancer, mTOR signaling pathway, and VEGF signaling pathway were significantly enriched in the high CEP55 expression group. As a cell cycle and proliferation gene, the evidence of CEP55 overexpression promoting tumor progression in many cancers is reasonable. A recent study reported that mTOR can stimulate glycerophospholipid synthesis and that increased lipogenesis is associated with increased mTOR activity and the occurrence of HCC in humans [29]. Angiogenesis has been recognized as the essential condition for cancer growth, while VEGF has been identified as its critical regulator. VEGF was overexpressed within liver tumor, and VEGF level showed significant relationship with the clinical classification as well as lymph node and lung metastases [30]. CEP55 may promote liver tumor cell growth and metastasis through the VEGF signaling pathway. Currently, surgical resection is the best method for the treatment of early liver cancer patients. Due to the difficulty in early diagnosis, most liver cancer patients, especially in China, are diagnosed at an advanced stage. Liver cancer is highly malignant, the therapeutic options for patients with advanced liver cancer are limited, and the therapeutic effect is poor. The 5-year liver cancer survival rate normalized by age is only 10.1% in China [31]. Therefore, we explored the association between CEP55 expression and OS. We found that CEP55 may be helpful in guiding treatment selection for liver cancer patients. Subgroup analysis results suggested that high CEP55 expression predominantly affected the prognosis of patients with clinical stage I/II, T1/T2, N0, and M0 disease, thus implying the special prognostic value of CEP55 and its potential contribution to the accurate treatment of liver cancer. However, few cases at advanced stage and high TNM classification may have led to this result, and future studies with large samples are needed to verify this conclusion.
TIICs are important parts of the tumor microenvironment, which are related to patient outcome and tumor behavior [32]. For the first time, we used the TIMER database to uncover the association between CEP55 expression and TIICs in liver cancer. Our results indicated that there is a significantly positive relationship between the CEP55 expression level and the infiltration level of B cells, CD4+ T cells, CD8+ T cells, macrophages, neutrophils, and dendritic cells in HCC. These correlations suggest that CEP55 may play an important role in the recruitment and regulation of immune cells in HCC. CEP55 is a tumor-associated antigen due to its aberrant expression in cancer. This property makes CEP55 an ideal candidate for cancer vaccine therapies. Some studies have reported the effectiveness of CEP55-based immunotherapy vaccines in the treatment of chemotherapy-resistant colon cancer stem cells and tumor-initiating cells [33,34]. However, in vitro/vivo experiment are required to determine whether CEP55based immunotherapy vaccines can be effectively used in HCC. In terms of CCA, no significant relationship was detected between CEP55 expression and the infiltration level of any TIICs. We speculate that this result may be due to the relatively small sample size of CCA.
DNA methylation is a common reason for gene expression change in tumors [35,36]. In order to explore the potential mechanism of overexpression in liver cancer, we first used DNA methylation and gene expression data of liver cancer patients in the TCGA database to analyze the correlation between DNA methylation and gene expression of CEP55. We found that CEP55 expression is negatively correlated with its DNA methylation level, suggesting that DNA hypomethylation is one of the underlying causes for the overexpression of CEP55 in liver cancer. In addition, we performed survival analysis to investigate the prognostic value of the CEP55 methylation level in liver cancer patients. However, the results showed that the CEP55 methylation level did not affect the prognosis of liver cancer patients. Then, joint prognostic survival analysis was performed and we found that the combination of CEP55 methylation and its expression had a significant correlation with the prognosis of liver cancer patients. Patients with CEP55 hypermethylation and low expression can achieve a better prognosis than those with CEP55 hypomethylation and high expression. Therefore, we suppose that the CEP55 methylation level does not directly affect the prognosis of liver cancer patients, but by regulating the expression of CEP55 and then affecting the prognosis of liver cancer patients. Further study is needed to confirm this conjecture in the future.
The present study improves our understanding of the association between CEP55 and liver cancer, but some limitations still exist. First, this study was conducted based on the data obtained from a public database. Although we have made a validation of survival analysis and protein expression of CEP55 based on the data from the ICGC, GEO, and HPA databases, further experiments are needed to explore the molecular mechanisms associated with CEP55 in liver cancer progression and prognosis. Second, our evidence indicating the role of CEP55 as a prognostic predictor is limited to patients with early-stage disease and low TNM classification. It is essential to identify more effective predictors in advanced liver cancer patients in the future.

Conclusions
In summary, CEP55 is overexpressed in liver tumor tissues compared to normal liver tissues and CEP55 overexpression is related to dismal prognosis and increased immune infiltration levels of B cells, CD4+ T cells, CD8+ T cells, macrophages, neutrophils, and dendritic cells in liver cancer. DNA hypomethylation of CEP55 may contribute to its overexpression in liver cancer. CEP55 might serve as a candidate treatment target for liver cancer. Further basic or clinical experiments are needed to prove the biological impact of CEP55 in liver cancer. Overall survival