In situ hybridization and immunohistochemical studies in oropharyngeal squamous cell carcinomas. A to D, A human papillomavirus (HPV)–positive tumor. A, The HPV-positive tumor demonstrates nuclear staining for the HPV-16 probe by in situ hybridization. B, Immunohistochemical studies in this tumor show nuclear staining for p16. C, Negative findings are demonstrated for p53. D, Negative findings are also demonstrated for epidermal growth factor receptor (EGFR). E to H, An HPV-negative tumor. E, No staining for the HPV-16 probe is seen by in situ hybridization. F, Immunohistochemical studies are negative for p16. G, Nuclear staining is positive for p53. H, Membranous staining is positive for EGFR (original magnification, ×200).
The cumulative overall survival by Kaplan-Meier analysis for patients with oropharyngeal cancer. A, The patients with a human papillomavirus (HPV)–positive tumor have significantly better survival than those with an HPV-negative tumor. B, The patients with a tumor showing high p16 expression have significantly better survival than those with low p16 expression. C, The patients with a tumor showing low p53 expression have significantly better survival than those with high p53 expression. D, The patients with an epidermal growth factor receptor (EGFR)–negative tumor have significantly better survival than those with an EGFR-positive tumor.
Al-Swiahb JN, Huang C, Fang F, Chuang H, Huang H, Luo S, Chen C, Chen C, Chien C. Prognostic Impact of p16, p53, Epidermal Growth Factor Receptor, and Human Papillomavirus in Oropharyngeal Cancer in a Betel Nut–Chewing Area. Arch Otolaryngol Head Neck Surg. 2010;136(5):502-508. doi:10.1001/archoto.2010.47
To evaluate the prevalence of human papillomavirus (HPV) and the prognostic significance of epidermal growth factor receptor (EGFR), p53, and p16 among patients with oropharyngeal carcinoma.
Academic Institute of Otolaryngology, Kaohsiung, Taiwan.
Two hundred seventy-four patients who were diagnosed as having oropharyngeal carcinoma underwent testing for the presence of the HPV genome in the nuclei of their tumor cells from January 1, 1992, through March 31, 2008.
The HPV genome was detected by performing polymerase chain reaction–based assays and in situ hybridization on tumor tissue from paraffin blocks. Immunohistochemistry staining for p16, p53, and EGFR was also performed.
Main Outcome Measures
We used the Fisher exact test to evaluate the correlation between the clinicopathological variables and the presence of HPV in tumor cells. Survival analysis was based on the Kaplan-Meier method.
We detected HPV in 45 of the 274 patients (16.4%); of these, HPV-16 and -18 were identified in 42 (93.3%) of the HPV-positive tumors. The HPV-positive oropharyngeal cancers were more likely to occur in females, nonsmoking individuals, and those who did not chew betel quid. The HPV-positive tumors significantly expressed p16 and were inversely associated with EGFR and p53 expression (all, P < .001). In addition, patients with tumor tissue that was positive for HPV (P = .008) and had negative expression of EGFR (P = .01), low expression of p53 (P = .01), and high expression of p16 (P = .04) had a better prognosis.
Our results suggest that HPV, EGFR, p53, and p16 are useful biomarkers in predicting the clinical outcomes of oropharyngeal cancer.
The human papillomavirus (HPV) is associated with a large spectrum of epithelial lesions. Most of these are benign hyperplasia (warts) that rarely progress to cancers. A subgroup of HPVs, however, are associated with lesions that have a tendency to undergo malignant transformation.1 Currently, there are sufficient molecular and epidemiological data to suggest a pathological significance of HPV in head and neck squamous cell carcinomas (HNSCCs). The role of HPV in the pathogenesis of HNSCCs was first described in 1983,1 when histopathological features consistent with HPV infection were identified in oral cancers.2 Two years later, viral DNA from high-risk HPV-16 was detected in an oral carcinoma by Southern blot hybridization.2 By the early 1990s, discoveries of HPV in association with oropharyngeal carcinomas became consistent.1
Currently, we have more molecular evidence that indicates a relationship between high-risk HPV and the pathogenesis of oropharyngeal cancers. Expression of HPV e6/e7 oncogenes are the current criterion standard for determining a causal role of oncogenic HPV in human tumors.3 The E6 and E7 genes produce oncoproteins and are responsible for the degradation of p53 and Rb, respectively,3 which are important events that lead to dysregulation of the cell cycle and thereby immortalization of the cells.
The silencing of p16 by promoter methylation is commonly observed in HNSCC.4 In contrast, HPV is readily detected in HPV-positive tumors and is highly correlated with p16 expression, which is most likely the result of transcriptionally active HPV infection.5 Epidermal growth factor receptor (EGFR) is a tyrosine kinase receptor that is expressed in low amounts in normal squamous cells. Clinically, overexpression of EGFR in HNSCC is correlated with more aggressive tumor behavior and poorer prognosis.6 Therefore, the aims of our study were to explore the prevalence of HPV and the expression of p53, p16, and EGFR in oropharyngeal cancer and their relationship to clinicopathological variables in a betel nut–chewing area.
Our study, which was approved by the Human Research Ethics Committee of Chang Gung Memorial Hospital, included 274 patients with oropharyngeal carcinoma who underwent diagnosis and treatment with surgical resection followed by postoperative adjuvant therapy, primary radiotherapy, or concurrent chemoradiotherapy between January 1, 1992, and March 31, 2008, at the Academic Institute of Otolaryngology, Kaohsiung, Taiwan. All patients who were treated during this period were enrolled according to the databases of the departments of Radiation Oncology and Otolaryngology. The patients underwent clinical staging of their cancer according to the 1997 American Joint Committee on Cancer system.7 Clinicopathological data including age, sex, smoking, betel quid chewing, alcohol intake history, nodal status, tumor site, and outcome data were obtained retrospectively.
Group 1, which was used for survival and clinicopathological analyses, consisted of 220 patients who received a diagnosis between January 1, 1992, and December 31, 2005, and had a mean age of 51.2 (range, 20-89) years. Group 1 included 111 patients with tonsil cancer, 71 patients with tongue base cancer, and 38 patients with soft palate cancer. Group 2 consisted of 54 patients who were diagnosed as having oropharyngeal squamous cell carcinoma from January 1, 2006, through March 31, 2008, with a mean age of 51.3 (range, 31- 84) years. The group 2 patients were only used for analyzing the prevalence rate of HPV in oropharyngeal cancer. These patients were not included in the survival analysis because they had only a short-term follow-up for survival.
To study HPV DNA, the most representative paraffin block of the tumor for each case was selected by examining the hematoxylin-eosin–stained sections. For the cases with small tumor nests on the section, macrodissection was performed to ensure that the tumor cell purity was greater than 50%. Ten tissue sections, 10 μm thick each, were obtained from the selected tissue block of each specimen and collected in a 1.5-mL Eppendorf tube for DNA extraction. After deparaffinization, the genomic DNA was extracted using a tissue kit (QIAamp; QIAGEN GmbH, Hilden, Germany) according to the manufacturer's recommendation and finally dissolved in 20 μL of distilled water for later use.
All samples were checked for DNA integrity by amplifying β-globin as a housekeeping gene. Primer sequences and the size of polymerase chain reaction (PCR) products for β-globin and HPV-DNA analyses are listed in Table 1. The DNA samples with positive test results for β-globin DNA by PCR analysis were subsequently studied using consensus primers MY09 and MY11. The primer sequences were the same as those previously reported by Resnick et al,8 flanking the conserved region L1 open reading frame that was found in a broad spectrum of HPV. The PCR analysis was performed by denaturation of the samples at 94°C for 5 minutes before the addition of 2.5 U of thermostable Taq polymerase, which was then followed by 35 cycles of amplification. A final extension of 7 minutes completed the last cycle. The PCR products were analyzed by electrophoresis and visualized on a UV light transilluminator. To further confirm the PCR products amplified by MY09 and MY11 primers and to overcome the possible false-negative result for HPV detection potentially caused by DNA degradation in aging archival paraffin-embedded tissues, we also used a second PCR-based strategy with another pair of primers, GP05 and GP06, franking a shorter fragment (150 base pairs [bp]) within the L1 open reading frame in supplementary analyses for all cases and control specimens (Table 1).
Catalyzed reporter deposition in situ hybridization was then used to determine the type of HPV and to localize the tissue sections of the samples that were positive for HPV L1 by PCR. If the type of HPV was beyond the detection of in situ hybridization using the HPV-6, -11, -16, and -18 probes, the HPV type was then determined by results of a linear array HPV genotyping test (Roche Diagnostics GmbH, Mannheim, Germany).
Whole genomes of HPV-6, -11, -16, and -18 cloned into a 2743-bp pGEM II plasmid (Promega Corporation, Fitchburg, Wisconsin) were used as DNA templates. The sequences of the consensus primers were identified according to the similarity of DNA sequences flanking the segments of the noncoding regions described by Huang et al.9 A 50-μL PCR reaction mixture was prepared using 25 ng of DNA template, 1.5mM magnesium chloride, 0.5μM of each primer, 5 U of Taq DNA polymerase, and deoxyribonucleotide triphosphates. The concentrations of deoxyadenosine triphosphate, deoxythymidine triphosphate, deoxycytidine triphosphate, and deoxyguanosine triphosphate were 200μM each in the reaction mixture, but 10% of the deoxythymidine triphosphate was substituted with DIG-16-deoxyuridine triphosphate (Roche Diagnostics GmbH). The reaction mixture was denatured at 94°C for 5 minutes, followed by 35 amplification cycles. The final product was cooled down to 4°C after the reaction, followed by purification using a PCR purification kit (QIAGEN GmbH), and was diluted to 3.0 ng/μL with DNA in situ hybridization solution (DAKO, Carpinteria, California).
Paraffin sections of formalin-fixed CaSki cell line (known to contain 500-600 copies of HPV-16 DNA) were included and detected in parallel with the sample tissue sections as positive controls. The SiHa cell line (known to contain only 1 to 2 copies of HPV-16) was used to ensure high detection sensitivity for samples with very few copies of HPV.9,10 The C33A cell line served as the negative control. The cell blocks used for control samples and the tissue blocks from the specimens that were positive for HPV L1 by PCR were cut into 5-μm-thick sections onto poly-L-lysine–coated slides.
The slides were placed in a 60°C incubator for 30 minutes, followed by deparaffinization. The target DNA was unmasked by digestion with protease K (DAKO), followed by quenching of the endogenous peroxidase with 3% hydrogen peroxide at room temperature for 20 minutes. After the DNA probe was added, the sections were denatured by heating at 95°C for 5 minutes and were incubated at 45°C overnight in a humid chamber. After hybridization, the slides were rinsed with TBST (TRIS-buffered saline polysorbate 20, containing 50mM TRIS hydrochloride [pH, 7.6], 300mM sodium chloride, and 0.1% Tween 20, provided in the GenPoint kit [DAKO]), followed by a stringency wash with 0.1 × 11 saline–sodium citrate and detergent (provided in the GenPoint kit) at 48°C for 20 minutes. Slides were then rinsed with TBST for 5 minutes and incubated with primary anti-DIG–horseradish peroxidase antibody (DAKO), which was diluted 1:500 in phosphate-buffered saline for 30 minutes and finally followed by a TBST wash.
The GenPoint kit was used for signal amplification. Briefly, the slides were incubated with biotinyl tyramide for 15 minutes and then washed and agitated in TBST 3 times for 5 minutes each time. After the amplification step, streptavidin–horseradish peroxidase was applied to the slides, which were then incubated at room temperature for 15 minutes, followed by three 5-minute washes in TBST. Finally, the signals were developed by diaminobenzidine.
The representative blocks of the formalin-fixed, paraffin-embedded biopsy tissue samples were retrieved and sectioned for the immunohistochemical study. The p16, p53, and EGFR monoclonal antibodies were purchased from Neomarkers Inc (Fremont, California), DAKO, and Santa Cruz Biotechnology (Santa Cruz, California), respectively. These monoclonal antibodies were all diluted 1:100 in phosphate-buffered saline according to the manufacturer's instructions. The sections were deparaffinized, incubated for 1 hour at room temperature, and then treated with 3% hydrogen peroxide for 10 minutes to inhibit endogenous peroxidase activity. They were then microwaved in 10mM citrate buffer (pH, 6.0) to unmask the epitopes. After antigen retrieval, the sections were incubated with diluted primary antibodies for 1 hour, followed by a phosphate-buffered saline wash. Horseradish peroxidase–Fab polymer conjugate (PicTure-Plus kit; Zymed Laboratories, South San Francisco, California) was then applied to the sections for 30 minutes. After extensive washing, the sections were incubated with peroxidase substrate diaminobenzidine for 5 minutes. Thereafter, the sections were counterstained with Gill hematoxylin and mounted with mounting medium. All the sections were interpreted by a pathologist (C.-C.H.) who did not know the clinical data.
Strong nuclear staining as well as strong cytoplasmic staining was considered positive for p16 expression. Immunostaining for p16 was regarded as high expression if it was strong and diffuse and if more than 60% of the tumor cells were positive for p16.5 When evaluating the intensity of nuclear staining for p53, high expression was defined as expression in greater than 50% of the tumor cells.11 Expression of EGFR was regarded as positive when greater than 50% of cell membranes in the tumor cells were completely stained.5
We used the Fisher exact test to evaluate the correlation between the clinicopathological variables and the status of HPV in tumor cells. In addition, the correlation among the expressions of EGFR, p16, p53, and HPV was also analyzed. P < .05 was considered significant in all the statistical analyses. The variables, including age, sex, T and N stages, TNM stage, the presence of HPV in tumor cells, and expression of p16, p53, and EGFR, were accounted for in survival analysis based on the Kaplan-Meier method. Statistical significance was assessed by a log-rank test. To determine the distinct prognostic factors for survival, a multivariate analysis was performed according to the Cox proportional hazards regression model.
There were 206 men and 14 women in group 1 and 54 men and no women in group 2. Sixteen patients were classified as having T1 disease; 88, T2; 60, T3; and 110, T4. Sixty patients were classified as having N0 disease; 42, N1; 129, N2; and 43, N3. There were 4 stage I, 25 stage II, 43 stage III, and 202 stage IV carcinomas. The presence of HPV in the tumor nuclei was more likely to be found among females (P < .001), nonsmoking patients (P = .009), patients with positive nodal metastasis (P = .04), and those who did not chew betel nuts (P = .054) (Table 2). The HPV-positive tumors were significantly more likely to express p16 and were inversely associated with EGFR and p53 expression (all, P < .001) (Figure 1). In addition, patients with the presence of HPV (P = .008), negative expression of EGFR (P = .01), low expression of p53 (P = .01), and high expression of p16 (P = .04) in tumor cell nuclei had better prognoses of cumulative 5-year overall survival (Table 3 and Figure 2). However, the Cox proportional hazards regression analysis demonstrated that HPV status (P = .02) and T stage (P < .001) were the only independent factors for cumulative 5-year overall survival in this cohort (Table 4).
In total, 45 of the 274 patients (16.4%) were identified with HPV within the nuclei of their tumor cells in both groups. Thirty-five patients (77.8%) had positive findings for HPV-16; 4 (8.9%), for HPV-18; 3 (6.7%), for HPV-16 and -18; 1 (2.2%), for HPV-31; 1 (2.2%), for HPV-16 and -31; and 1 (2.2%), for HPV-6 and -11. In group 1, the prevalence rates of HPV in tonsil cancer, tongue base cancer, and soft palate cancer were 12.6%, 18.3%, and 15.8%, respectively. The prevalence rate of HPV in group 1 patients who were enrolled from 1992 to 2005 was 15%. The prevalence rate of HPV in group 2 patients who were enrolled from 2006 to 2008 showed an increasing trend to 22.2%. In group 2, the prevalence rates of HPV in tonsil cancer, tongue base cancer, and soft palate cancer were 17.9%, 25.0%, and 27.3%, respectively.
Because the patients were enrolled in a wide time frame (1992-2008), there were great variations in the therapeutic strategy. The survival might be affected by different methods of treatment, especially for the patients with stages III and IV cancer. Accordingly, we also compared the survivals of the patients with stages III and IV cancer who underwent treatment before and after 1999 in this study cohort based on the Kaplan-Meier method. The patients treated after 1999 had better survival rates compared with those treated before (P = .007). This finding might be attributed to the evolution of surgical and radiation techniques, a better imaging study in the selection of patients, and, most important, the intervention of multidisciplinary team work for the treatment of patients in recent years.
It is widely accepted that the most common risk factors for HNSCC are tobacco use, heavy alcohol consumption, betel quid chewing, and high-risk HPV infection. High-risk HPVs are currently well accepted as causative agents for most cervical cancers and have recently been associated with a subset of head and neck cancers.12- 14 Oropharyngeal cancer has emerged as another mucosal cancer that is associated with HPV infection.13 Klussmann et al15 reported HPV DNA to be present in 20% of all HNSSCs and in nearly 60% of tonsillar cancers. In contrast, our study showed that only 12.6% of tonsillar cancers were HPV positive in a betel nut–chewing area.
Patients with HPV tend to be younger than those without the virus.16 This finding may be the result of high-risk sexual behaviors that are more common among younger patients. However, our study was inconsistent in this observation because our patients showed a higher percentage of betel nut chewing and represented a mean age of approximately 50 years. In addition, the most commonly detected HPV in HNSCC is HPV-16, which represents 90% to 95% of all HPV-positive HNSCC cases.17,18 In the present series, the prevalence rate of HPV was low in oropharyngeal cancer, but HPV-16 was still shown to be the dominant type. In addition, the presence of HPV in the tumor nuclei was more likely to be found among patients with metastasis to the neck in this study. The reason is unclear, but more studies are necessary to explore this answer.
Patients with HPV-positive oropharyngeal cancer responded well to treatment and showed good prognoses.18,19 Absence of p16 mutation and high expression of p16 are also distinguishing molecular markers of HPV-positive HNSCC.20 Therefore, it has been suggested that high expression of p16 may be used as a surrogate marker to identify HPV-positive HNSCC. In this study, strong expression of p16 was significantly associated with HPV-positive tumors and with better clinical outcomes. These findings are consistent with other studies.5 In HNSCC, p53 is frequently mutated,21 overexpressed,22 and associated with poorer prognosis.23 In contrast, in the present series, low expression of p53 and the presence of HPV in tumors were associated with better prognosis. Generally, p53 is inactivated by HPV E6 protein in an HPV-positive tumor, which may be the reason for low expression of p53 in the tumor cells.
Growth receptors that are normally expressed at lower rates in most epithelial cells, in particular EGFR, have attracted attention owing to their inherent ability to stimulate the proliferation of epithelial cells.24 In HNSCC, EGFR is usually overexpressed and is related to poor prognosis.5,25 Our data showed results consistent with this. In the present series, the molecular explanation for the inverse correlation between p16 and EGFR remains unclear, and more studies are necessary to explore this observation.
In conclusion, the prevalence rate of HPV-positive tumors in oropharyngeal cancer was low in our cohort and, in most cases of HPV positivity, the subtype was HPV-16. There was a higher rate of HPV incidence in females, nonsmoking individuals, and those who did not chew betel nuts among patients with oropharyngeal cancer. The HPV-positive tumors showed a rising trend among these patients in the last 2 years of the study. This might have been caused by a change in lifestyle, particularly in sexual behavior. To our knowledge, this study is the first combining the promising predictive markers, including HPV DNA and p16, p53, and EGFR expression in a large series of oropharyngeal cancer in an area of betel nut chewing such as Taiwan. For clinicians, a combined immunohistochemical study for p16, p53, and EGFR expression and detection of HPV DNA are easily applicable in a routine pathology laboratory, and it can provide invaluable prognostic information for individual patients with more relevance than staging alone.
Correspondence: Chih-Yen Chien, MD, Department of Otolaryngology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, 123 Ta-Pei Road, Niao-Song Hsiang, Kaohsiung County, 833 Taiwan (email@example.com).
Submitted for Publication: May 26, 2009; final revision received October 23, 2009; accepted November 25, 2009.
Author Contributions: Drs Al-Swiahb and C.-C. Huang contributed equally to this study. Drs Al-Swiahb, C.-C. Huang, and Chien had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Al-Swiahb, C.-C. Huang, and Chien. Acquisition of data: Al-Swiahb, C.-C. Huang, Fang, Chuang, and Luo. Analysis and interpretation of data: C.-C. Huang, Chuang, H.-Y. Huang, Luo, C.-H. Chen, and C.-M. Chen. Drafting of the manuscript: Al-Swiahb and Chien. Critical revision of the manuscript for important intellectual content: C.-C. Huang, Fang, Chuang, H.-Y. Huang, Luo, C.-H. Chen, C.-M. Chen, and Chien. Statistical analysis: Fang. Administrative, technical, and material support: Al-Swiahb, C.-C. Huang, Chuang, H.-Y. Huang, Luo, C.-H. Chen, C.-M. Chen, and Chien. Study supervision: Chien.
Financial Disclosure: None reported.
Funding/Support: This study was supported by grant CMRPG860512 from the Chang Gung Memorial Hospital.
Previous Presentation: This study was presented at the Asia-Pacific Congress on Oral Cavity Cancer in Conjunction with the 12th Annual Meeting of the Taiwan Cooperative Oncology Group; December 6, 2008; Taipei, Taiwan.