Centromere Protein H Is a Novel Prognostic Marker for Nasopharyngeal Carcinoma Progression and Overall Patient Survival

Nasopharyngeal carcinoma (NPC) is an EBV-associated cancer (1), which is highly prevalent in Southeast Asia. The incidence of NPC in South China, especially in the Cantonese region around Guangzhou, is ∼100-fold higher compared with populations in Europe and North America (2). Numerous studies indicate that the etiology of NPC includes viral, genetic, and environmental factors. However, the molecular mechanism of the development and progression of NPC is still poorly understood. It is of great clinical value to further understand the molecular mechanism of this cancer and find valuable early diagnostic markers as well as novel therapeutic strategies.

A large number of human cancers have an abnormal number of chromosomes, a feature referred to as aneuploidy, which often develops during the early stage of tumorigenesis (3, 4). Several studies have shown that aneuploidy is necessary in neoplastic transformation (5, 6). Recently, it was shown that chromosome missegregation during mitosis is one of the main causes of aneuploidy and contributes to tumorigenesis. Potential mitotic defects such as sister chromatid cohesion, kinetochore assembly, centrosome duplication, microtubule dynamics, and checkpoints of the cell cycle have been shown to play important roles in the induction of aneuploidy and carcinogenesis (7–9). Previous research has shown that p53 and Rb mutations were not as common in NPC as in other solid tumors (10, 11). However NPC progressively gains widespread genomic imbalances which possibly occur before and during carcinogenesis (12). Evidence of consistent chromosomal gains and losses during the progression of this disease has been shown by different strategies of comparative genomic hybridization (13–16).

The kinetochore is a large multiprotein complex that is assembled on centromeric DNA. It serves as the attachment site for spindle microtubules, which play an essential role for proper chromosome segregation during mitosis (17). Several kinetochore proteins have been identified in humans, and the list of kinetochore associated proteins continues to grow, including centromere protein (CENP)-A, CENP-B, CENP-C, CENP-E, CENP-F, CENP-H, CENP-I, and INCENP (17–21). Studies have shown that kinetochore malfunction is a major cause of aneuploidy and is closely associated with carcinogenesis. Normal expression of core kinetochore components is essential to prevent chromosome instability and carcinogenesis. Recently, it was reported that CENP-A, which was one of the first kinetochore components identified in humans, is overexpressed and mistargeted in colorectal cancer tissues (22). CENP-H was initially identified as a component of the mouse centromere. Human CENP-H protein was recently isolated and shown to localize in the inner plate together with CENP-A and CENP-C (19, 23) and is a fundamental component of the active centromere complex (24). Studies using budding yeast have shown that a molecular core consisting of CENP-A, CENP-C, CENP-H, and Ndc80/HEC plays a central role in linking centromeres to the spindle microtubule (25). Recent research has shown that CENP-H is up-regulated in most colorectal cancers, and ectopic overexpression of CENP-H induces chromosome missegregation and aneuploidy in a diploid cell lines (26).

Although CENP-H may play an important role in chromosome instability and carcinogenesis, there are no reports on its role in tumorigenesis and progression of NPC. In this study, we characterize CENP-H expression in human NPC.

Cell lines. A normal nasopharyngeal epithelial cell line (NPEC) was established in our lab and was cultured in Keratinocyte-SFM (Invitrogen, Carlsbad, CA) supplemented with antibiotics (120 μg/mL streptomycin and 120 μg/mL penicillin) and bovine pituitary extract. The NP69 cell line was obtained from the University of Hong Kong, PR China (27), and was also cultured in Keratinocyte-SFM. Two immortalized nasopharyngeal epithelial cell lines were induced by the oncogene Bmi-1 and the human telomerase gene (hTERT), respectively, and maintained in Keratinocyte-SFM in our laboratory (28). The NPC cell lines SUNE-1, 6-10B, and 5-8F were established in our lab and maintained in RPMI 1640 (Invitrogen) supplemented with 10% FCS (29). C666, CNE1, CNE2, and Hone1 were maintained in RPMI 1640 supplemented with 10% FCS.

Patients and tissue specimens. This study was conducted on a total of 160 paraffin-embedded NPC samples, which were histologically and clinically diagnosed from the Cancer Center, Sun Yat-sen University, between 1999 and 2002. For the use of these clinical materials for research purposes, prior patient's consent and approval from the Institute Research Ethics Committee were obtained. The disease stages of all the patients were classified or reclassified according to the 1992 NPC staging system of China as mentioned before (30), which was described in Table 1 (31). Clinical information of the samples is described in detail in Table 2. Patients included 117 males and 43 females, of ages ranging from 14 to 72 years (mean, 47.2 years). All cases were with no metastasis at its presence at original presentation with NPC. The figures on metastasis pertain to its presence at any time in follow-up. The median follow-up time for overall survival was 61.7 months for patients still alive at the time of analysis, and ranged from 11 to 79 months. A total of 64 (40%) patients died during follow-up and 40 (25%) patients experienced metastasis.

RNA extraction and reverse transcription-PCR. Total RNA from cells was extracted using Trizol reagent (Invitrogen) according to the manufacturer's instructions. The RNA was pretreated with DNase and used for cDNA synthesis with random hexamers. The full-length open reading frame of CENP-H was PCR amplified from cDNA samples of non-NPC and NPC cell lines. The following primers were used for amplification of CENP-H: sense primer, 5-TGCAAGAAAAGCAAATCGAA-3; antisense primer, 5-ATCCCAAGATTCCTGCTGTG-3. Glyceraldehyde-3-phosphate dehydrogenase was amplified as an internal control using sense primer, 5′-AATCCCATCACCATCTTCCA-3′, and antisense primer, 5′-CCTGCTTCACCACCTTCTTG-3′. The appropriate size of PCR products was confirmed by agarose gel electrophoresis.

Western blotting. Sample preparation for immunoblotting was done as previously described (30). Briefly, cells were harvested in 1× SDS sample buffer [62.5 mmol/L Tris-HCl (pH 6.8), 2% SDS, 10% glycerol, and 5% 2-mercaptoethanol] and were heated for 5 min at 100°C. Protein concentration was determined by the Bradford assay (Bio-Rad Laboratories, Hercules, CA). Equal amounts of proteins were separated electrophoretically on 12% SDS/polyacrylamide gels and transferred onto polyvinylidene difluoride membranes (Amersham Pharmacia Biotech, Piscataway, NJ). The membrane was probed with an anti-CENP-H rabbit polyclonal antibody (1:1,000; Bethyl Laboratories, Montgomery, TX). Expression of CENP-H was determined with horseradish peroxidase–conjugated antirabbit immunoglobulin G (1:3,000; Amersham Pharmacia Biotech) and enhanced chemiluminescence (Amersham Pharmacia Biotech) according to the manufacturer's suggested protocols. An anti–α-tubulin mouse monoclonal antibody (1:1,000; Santa Cruz Biotechnology, Santa Cruz, CA) was used to confirm equal loading.

Immunohistochemistry. Immunohistochemistry was done to study altered protein expression in 160 human NPC tissues. The procedures were done similarly to previously described methods (30). In brief, paraffin-embedded specimens were cut into 4-μm sections and baked at 65°C for 30 min. The sections were deparaffinized with xylenes and rehydrated. Sections were submerged into EDTA antigenic retrieval buffer and microwaved for antigenic retrieval. The sections were treated with 3% hydrogen peroxide in methanol to quench the endogenous peroxidase activity, followed by incubation with 1% bovine serum albumin to block the nonspecific binding. Rabbit polyclonal anti-CENP-H (1:500; Bethyl Laboratories) was incubated with the sections overnight at 4°C. For negative controls, the primary antibody was replaced by normal goat serum. After washing, the tissue sections were treated with biotinylated antirabbit secondary antibody (Zymed, San Francisco, CA), followed by further incubation with streptavidin-horseradish peroxidase complex (Zymed). The tissue sections were immersed in 3-amino-9-ethyl carbazole and counterstained with 10% Mayer's hematoxylin, dehydrated, and mounted in Crystal Mount. The degree of immunostaining of formalin-fixed, paraffin-embedded sections was reviewed and scored by two independent observers. The proportion of CENP-H-expressing cells varied from 0% to 100%, and the intensity of nuclear staining varied from weak to strong. The cells at each intensity of staining were recorded on a scale of 0 (no staining), 1 (weak staining, light yellow), 2 (moderate staining, yellowish brown), and 3 (strong staining, brown). Using this method of assessment, we evaluated the expression of CENP-H in benign nasopharyngeal epithelium and malignant lesions using a method that is similar with the Allred 8-unit system (32), which is also used by Coppola et al. (33). Cutoff values for CENP-H were chosen on the basis of a measure of heterogeneity with the log-rank test statistic with respect to overall survival. An optimal cutoff value was identified. An intensity score of ≥2 with at least 50% of malignant cells with positive CENP-H staining was used to classify tumors with high expression, and <50% of malignant cells with nuclear staining or <2 intensity score classified tumors with low expression of CENP-H antigen.

Statistical analysis. All statistical analyses were carried out using the SPSS 10.0 statistical software package. Mann-Whitney U test was used to analyze the relationship between CENP-H expression and clinicopathologic characteristics. Survival curves were plotted by the Kaplan-Meier method and compared by the log-rank test. The significance of various variables for survival was analyzed by the Cox proportional hazards model in the multivariate analysis. P < 0.05 in all cases was considered statistically significant.

Expression analysis of CENP-H by reverse transcription-PCR and Western blot. To investigate the expression levels of CENP-H transcripts and protein in nasopharyngeal epithelial cell lines, semiquantitative reverse transcription-PCR analysis and Western blotting analysis were done in the following cell lines: primary nasopharyngeal epithelial cells (NPEC), NPC cells (6-10B, C666, CNE1, CNE2, SUNE-1, 5-8F, and Hone1), and three immortalized primary nasopharyngeal epithelial cells (Bmi-1/NPEC, hTERT/NPEC, and NP69). All seven NPC cell lines and the immortalized nasopharyngeal epithelial cell lines showed higher level expression of CENP-H mRNA in comparison with the normal nasopharyngeal epithelial cell line NPEC (Fig. 1A and B). Western blotting analysis shows that CENP-H protein was highly expressed in all NPC cell lines as well as the immortalized cell lines, whereas it was weakly detected in NPEC (Fig. 1C and D).

Immunohistochemical analysis. Expression and subcellular localization of CENP-H protein was determined by immunohistochemistry in 160 paraffin-embedded, archival NPC tissues. CENP-H protein was detected in 153 of 160 (95.6%) cases. Specific CENP-H staining was mostly found in the nuclei of carcinoma cells, in some areas of metaplasia with atypical hyperplasia, and in a few scattered infiltrating lymphocytes in the form of yellow-brown granules (Fig. 2B-D). No specific CENP-H staining was observed in normal nasopharyngeal epithelial cells (Fig. 2A) and in the surrounding stroma cells.

Correlation between CENP-H protein expression and clinicopathologic features.Table 3 shows the relationship between the expression of CENP-H protein and clinical characteristics. There was no significant correlation between the expression level of CENP-H protein and age, gender, histologic classification, N classification, or distant metastasis of NPC patients. However, the expression of CENP-H is closely associated with clinical stage of NPC patients (P = 0.024). In addition, there was a significant difference of CENP-H expression in patients categorized according to T classification (P = 0.027). The expression of CENP-H protein was positively correlated with clinical staging and T classification (Table 3). Higher clinical staging and T classification correlated with higher CENP-H expression.

Survival analysis. The expression level of CENP-H protein in NPC was significantly correlated with patients' survival time (P < 0.001); the correlation coefficient was −0.27, indicating that higher levels of CENP-H expression was correlated with shorter survival time. Kaplan-Meier analysis and the log-rank test were used to calculate the effect of classic clinicopathologic characteristics (including age, gender, histologic classification, clinical stage, T classification, N classification, and distant metastasis) and CENP-H expression on survival. The log-rank test showed that the survival time was significantly different between these two groups (P = 0.0197). The low CENP-H expression group had better survival, whereas the high CENP-H expression group had shorter survival (Fig. 3). The cumulative 5-year survival rate was 76.02% (95% confidence interval, 0.668-0.852) in the low CENP-H expression group, whereas it was only 58.78% (95% confidence interval, 0.476-0.699) in the high CENP-H expression group.

In addition, T classification, N classification, clinical stage, and histologic classification were also significantly correlated with survival in Kaplan-Meier analysis and log-rank test (for T classification, P = 0.002; for N classification, clinical stage, and histologic classification, P < 0.0001). We did multivariate survival analysis, which included CENP-H expression level, histologic classification, clinical stage, and N classification, to determine if CENP-H expression level is an independent prognostic factor of outcomes. In this analysis, clinical stage, N classification, histologic classification, and CENP-H expression were recognized as independent prognostic factors (Table 4). Thus, our findings indicate that CENP-H protein expression level has a significant correlation with prognosis of NPC.

We also analyzed the prognostic value of CENP-H expression in selective patient subgroups stratified according to the tumor clinical stage, T classification, N classification, and distant metastasis, respectively. A trend toward shorter overall survival times of patients with high CENP-H expression was revealed in early clinical stage subgroups (stage I and II) or small T (T₁ and T₂) subgroups, but there were no significant differences (data not shown). However, patients with tumors exhibiting high CENP-H expression had significantly shorter overall survival compared with patients with low expression of CENP-H in the N₀-N₁ subgroup (n = 106; log-rank, P = 0.0047; Fig. 4A). A similar analysis of the N₂-N₃ subgroups (n = 54; log-rank, P = 0.3740; Fig. 4B) did not show statistically significant differences between patients with low or high levels of CENP-H expression. Thus, CENP-H seems to be a valuable prognostic marker for early-stage NPC patients.

In this report, we present the first evidence that a kinetochore protein, CENP-H, is overexpressed at both transcriptional and translational levels in NPC cell lines. CENP-H protein was observed in 95.6% of NPC specimens, and the expression level of CENP-H protein was found to be significantly correlated with the invasion of NPC primary tumor and the prognosis of NPC patients.

Our study suggests an important role for centromere proteins in the development and progression of NPC. In this study, we have shown that the expression level of CENP-H is much higher in NPC cell lines than in a normal nasopharyngeal epithelial cell line at both transcriptional and translational levels. It is interesting to note that all the three immortalized cell lines have a high level of CENP-H expression at both transcriptional and translational levels, although the cell lines were induced to immortalize by different genes. We then investigated the status of CENP-H expression in NPC tissue specimens. As determined by immunohistochemical analysis, 153 of 160 (95.6%) paraffin-embedded archival NPC biopsies showed moderate to strong nuclear staining of CENP-H in tumor cells, scattered infiltrated lymphocytes, and areas of squamous metaplasia. No positive nuclear staining of CENP-H was detected in the adjacent noncancerous epithelial cells.

We further analyzed the relationship between the expression of CENP-H and clinical characteristics of the patients. There was no significant correlation between the expression of CENP-H and age, gender, histologic classification, N classification, or M classification of NPC patients. However, there was a significant relationship of CENP-H expression in patients categorized according to clinical stage (P = 0.024) and T classification (P = 0.027), strongly suggesting that CENP-H can be used as a marker to identify subsets of NPC cancer patients with more aggressive disease. In addition, these results indicate that CENP-H might play an important role in the progression and invasion of NPC.

These observations provide new insight for understanding chromosome instability in NPC and highlight the important role of centromere proteins in the development and progression of NPC. Chromosomal abnormalities, including abnormal chromosome numbers, chromosome deletion, and amplification, are commonly found in NPC (14, 34, 35). Recently, it was reported that defects in mitotic checkpoints occur frequently (∼40%) in NPC cells (36). Evidence has shown that chromosomal instability plays an important role in the development and progression of cancer because aneuploidy is frequently found in the earliest stages of tumorigenesis (37). More and more kinetochore proteins were proved to be associated with carcinogenesis. CENP-F and INCENP are up-regulated in human cancer cells (38, 39). CENP-A has been shown to be overexpressed in colorectal cancer cells and is mistargeted to noncentromeric regions (22). CENP-F has been implicated in malignancy (40, 41), and its expression is correlated with tumor size in node-negative breast cancer (42). The CENP-F gene is amplified and overexpressed in head and neck squamous cell carcinoma (38). A recent study has shown that increased CENP-F protein levels influence tumorigenesis at early stages of tumor development (43). The CENP-H protein is expressed at high levels in human colorectal cancer and is overexpressed in tumor epithelium and not in the surrounding mesenchymal cells (26). This evidence suggests that kinetochore proteins may play an important role in carcinogenesis and that CENP-H may be a valuable diagnostic marker for tumor development and progression.

In our study, we found that CENP-H was overexpressed in NPC cell lines as well as in NPC tissues both at transcriptional and translational levels. The overexpression of CENP-H might deplete other centromere-kinetochore components and disrupt the kinetochore complex, or prevent normal kinetochore assembly and consequently cause aneuploidy and induce the development of cancer. CENP-H is a component of active centromere-kinetochore complexes in mammals, colocalizing with both CENP-A and CENP-C, which are found in the inner kinetochore plate throughout the cell cycle (19, 23). Inappropriate expression of CENP-H could induce abnormal kinetochore function and chromosome missegregation. In addition, CENP-H is necessary for localization of CENP-C to the centromere (24). A study in colorectal cancer cell lines showed that aneuploidy occurred most frequently in cells that overexpress CENP-H, and in these cells CENP-H was absent from the centromeres of metaphase chromosomes (26). Other reports show that deletion of CENP-H results in an accumulation of cells in metaphase and subsequent cell death as a result of chromosome aberrations and missegregation (24). Research in yeast has shown that the up-regulation of kinetochore protein MIF2, the yeast homologue of CENP-C, increased the frequency of missegregation of chromosomes during mitosis (44). These results suggest that the relationship among kinetochore proteins may be crucial for the appropriate localization and the proper function of the kinetochore. Further investigation is required to clarify the mechanism of CENP-H-induced chromosome instability in nasopharyngeal epithelial cells and its role in the development of NPC.

The development and progression of NPC stages, including single hyperplasia and single squamous metaplasia, atypical hyperplasia and allotype squamous metaplasia, in situ carcinoma, infiltrating carcinoma, and metastatic carcinoma, may involve the accumulation of multiple genetic alterations over a long period of time (45). In the present study, we have found that CENP-H was overexpressed in some areas of squamous metaplasia, which occurs at the early stages of NPC development. These results suggest that the dysfunction of CENP-H may play an important role in early-stage NPC carcinogenesis. Cell immortalization is the ability of normal cells to grow through an in definite number of divisions in culture. Because immortalized cells are capable of unlimited proliferation and represent the early stage of transformation before malignant transformation, we examined CENP-H expression in immortalized nasopharyngeal epithelial cell lines. We found that CENP-H protein was up-regulated in NP69 cells, as well as in two immortalized nasopharyngeal epithelial cell lines induced by the telomerase gene hTERT and oncogene Bmi-1 (Fig. 1B). These results suggest that CENP-H may be an early transformation factor of nasopharyngeal epithelial cells.

Furthermore, we have shown in univariate and multivariate analyses that high expression of CENP-H is a significant predictor of poor prognosis for NPC patients. Importantly, we found that CENP-H might function as a new prognostic marker in NPC for the N₀-N₁ patient subgroups. There is also a trend toward shorter overall survival times of patients with high expression of CENP-H, as revealed in early clinical stage subgroups (stage I and II) or small T (T₁ and T₂) subgroups, but these differences were not significant (data not shown). This trend also suggests that CENP-H may be an important prognostic marker for early-stage NPC patients. Further studies are clearly needed to verify these findings to establish CENP-H as a prognostic marker in NPC and to clarify its role in carcinogenesis by functional analysis.

In conclusion, this is the first study showing the expression of CENP-H in NPC cell lines as well as tissues, highlighting the clinical significance of CENP-H in NPC. Higher CENP-H expression is associated with advanced tumor staging and T classification, and high CENP-H expression is also a significant prognostic marker of poor survival in NPC patients. CENP-H may serve as a useful molecular marker for NPC and an indicator for tumor progression and invasion. In combination with other biomarkers of NPC, CENP-H expression status may be useful to stratify patients for novel therapeutic strategies, such as those of adjuvant chemotherapy, radiosensitization, or the establishment of rational treatment selection criteria for patients with this disease. However, these findings still need to be replicated, and further investigation in another patient population is required to verify these hypotheses. Additionally, further studies are needed to clarify the mechanism by which CENP-H is involved in the development and progression of NPC and its exact role in the regulation of chromosome instability in NPC.