gms | German Medical Science

Several risk factors exist on the gene level for the development of Head and Neck Squamous Cell Carcinoma (HNSCC). Besides amplification of epidermal growth factor receptor (EGFR) and loss of heterozygosity on 9p or 3p infection with HPV16 and promoter methylation are important for carcinogenesis in a subset of HNSCC. Little knowledge exists about the potential interaction of HPV16 and promoter methylation in HNSCC. Therefore, we analyzed tissue from 55 HNSCC patients and 31 controls for promoter methylation of 12 carcinogenesis-relevant genes (TIMP3, CDH1, CDKN2A, DAPK1, TCF21, CD44, MLH1, MGMT, RASSF1, CCNA1, LARS2 and CEBPA) by Methylation-specific PCR with respect to their HPV16-status. The highest prevalence of gene methylation in HNSCC was seen for TCF21 (62.2%), CDH1 (35.1%) and CCNA1/Cyclin A1 (34.7%). Compared to controls methylation of DAPK1 (p=0.043), Cyclin A1 (p=0.016) and TCF21 (p=0.0005) was significantly more present in HNSCC, with the weakest correlation for DAPK1.

There are many studies about methylation rates of different genes in HNSCC. With regard to the genes investigated herein, those methylation rates show great variation. Some of our results are comparable to those of other studies; while the values of other examined genes (e.g. CDKN2A and RASSF1) differ somewhat from those of former studies. Yet, most of the studies on promoter methylation in HNSCC done so far used saliva rinses for examination and methylation-specific PCR based on bisulfite-conversion [1], [2], [3]. In addition, our results are predominantly from tumors of oropharyngeal site (tonsil and base of the tongue). These details could explain why the results shown herein do not necessarily agree with other published data.

Only a few other studies mention the HPV-status as a dependable variable [4]. In our study the genes TIMP3 (p=0.018) and Cyclin A1 (p=0.015) showed higher methylation frequency in HPV16 positive HNSCC compared to HPV16 negative tumors. Since more than 90 percent of HPV induced HNSCC are caused by HPV16 we only tested for HPV16 by PCR and p16 overexpression [5]. A possible involvement of other HPV types in five cases with p16 overexpression but negative HPV16-DNA PCR was ruled out by in-situ hybridization. This is the first study, to our knowledge, that shows a possible linkage between HPV16 infection and Cyclin A1 methylation. Dong et al. showed an inverse correlation between HPV16/18-positivity and methylation of RASSF1 [4]. This correlation could reflect a functional role of both variables, RASSF1 and HPV-E6/7, in the G1 cell cycle check point.

Cyclin A1 is also an important regulator of the cell cycle and is required for S phase and passage through G2 [6]. HPV16-E7 can transform rodent fibroblasts through induction of Cyclin A and E. The transcription factor E2F forms a complex with p107, Cyclin A1, and the CDK2 kinase (E2F-cyclinA complex) during S phase. HPV16-E7 associates very efficiently with the E2F-CyclinA complex which is crucial for its transforming activity [7]. In this context Santopietro et al could identify protein expression of Cyclin A as independent predictor of high-risk HPV and high-risk HPV associated high-grade lesions in cervical cancer [8]. In contrast Rivera et al. could show that Cyclin A1 is a downstream target of p53 and can induce apoptosis and G2M-arrest if up-regulated [9]. To gain further insight into the molecular mechanisms of the Cyclin A1 pathway we did immunohistochemical staining of Cyclin A1, p53 and p53 mutation analysis. The protein expression of Cyclin A1 was independent of Cyclin A1 methylation and HPV16-status. All HPV16-positive HNSCC were lacking p53 mutation and p53 overexpression, while some HPV16-negative HNSCC showed overexpression of p53 in the absence of p53 mutation. This could already be demonstrated by others [10]. In addition, Cyclin A1 methylation showed no correlation with p53 mutation or p53 overexpression. An inverse correlation between Cyclin A1 methylation and p53 mutation in HNSCC could be shown by Tokumaru et al. [11]. However, Tokumaru et al. did not determine the HPV-status in their collective. Several studies on cervical cancer, which is mostly HPV16 positive, could point out that Cyclin A1 methylation is strikingly common and, additionally, it is also specific to the invasive phenotype in comparison with other histological stages during multistep carcinogenesis [12]. Of note, HPV-induced cervical cancer is characterized through wild-type p53 which is inactivated through E6. We found a strong correlation between HPV16-positivity and Cyclin A1 methylation independent of p53 mutation or p53 overexpression and without relevance for Cyclin A1 protein expression. One could hypothesize that HPV16 infection of mucosa cells of the oropharynx leads to Cyclin A1 methylation through an almost unknown pathway leaving the protein expression intact, which is crucial for E7-dependent carcinogenesis. The methylation of the Cyclin A1 promoter region might only influence the specific function of Cyclin A1 responsible for apoptosis and cell cycle arrest [9]. Incoherence between promoter methylation and protein expression of Cyclin A1, MLH1, MGMT and TIMP3 could be established by Xu et al. [13]. On the other hand it is possible that HPV16 can override the effect of decreased Cyclin A1 protein expression, due to a coincidentally present Cyclin A1 promoter methylation and E6-dependent inactivation of p53, by its viral protein E7.

We further compared the methylation status of each gene to the clinicopathological characteristics age, sex, tumor size (T), cervical lymph node status (N), distant metastases (M), UICC stage, primary site, histological grade, alcohol and nicotine consumption. Methylation of TCF21 was associated with higher age (p=0.044) and nicotine abuse (p=0.035). Since higher age and smoking are well accepted risk factors for tumor development, especially in HNSCC, and methylation of TCF21 is a frequent event in HNSCC, TCF21 may be the corresponding molecular mediator for age- and/or nicotine-induced carcinogenesis [14]. In our collective methylation of Cyclin A1 was significantly more present in females (p=0.003). This is in accordance to our previous findings that the percentage of females is significantly higher in HPV16-positive cases than in HPV16-negative cases [15]. Shaw et al. could find an inverse correlation between Cyclin A1 methylation and histological grade [16]. A lower histological grade in HPV-positive HNSCC could be demonstrated several times in previous studies [17]. Unfortunately, we were not able to correlate methylation status with survival data because of the short follow-up period. We would be very interested in survival results of patients with HPV16-positive, Cyclin A1 methylated tumors compared with those of HPV16-positive, and Cyclin A1 unmethylated tumors. This requires a much greater patient collective and longer follow-up period.

Taken together, our findings suggest that methylation of cell cycle control protein Cyclin A1 is induced by HPV16. Although the methylation of tumor suppressor genes (like Cyclin A1) is thought to inactivate its function as an important step in carcinogenesis, the methylation of Cyclin A1 in HPV16-dependent HNSCC does not lead to decreased protein expression or functional inactivation of the Cyclin A1 protein since the malignant transformation induced by the E7 protein of HPV16 depends on active Cyclin A1. Further investigations are needed to proof the relevance of Cyclin A1 methylation in HPV16-dependent carcinogenesis.