By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
Epidemiologic evidence points to a connection between viral infection by the human papillomavirus (HPV) and a subgroup of squamous cell carcinoma of the oropharynx. To assess the impact of HPV infection on the response of these tumors toward radiotherapy, the authors retrospectively determined the presence of the virus and the integrity of the viral E2 gene in tumors of patients who have undergone curative irradiation.
Paraffin embedded biopsies from 99 patients were analyzed for HPV infection and E2 gene integrity by multiplex PCR. The experimental findings were correlated with clinical characteristics, known risk factors, and treatment outcome.
Fourteen of 99 tumors were HPV positive (11 HPV16, 1 HPV33, 1 HPV35, and 1 HPV45). Human papillomavirus positivity was closely linked to female gender (odds ratio [OR], 5.75; P = 0.004), age older than 56 years (OR, 7.42; P = 0.012), nonsmokers (OR, 21.33; P = 0.00001), and alcohol abstainers (OR, 5.35; P = 0.012). There was an inverse association with p53 nuclear immunoreactivity (OR, 0.06; P = 0.008). The Kaplan–Meier survival estimates showed a better local control (P = 0.050, log-rank) and a better overall survival (P = 0.046, log-rank) for patients with HPV positive tumors. In the multivariate analysis, HPV positivity remained to be associated with a lower risk of local failure (risk ratio [RR], 0.31; P = 0.048). Four of 11 HPV16 positive tumors had a disrupted E2 gene. Only tumors with a disrupted E2 gene manifested local treatment failure.
Human papillomaviruses (HPVs) are a family of small DNA viruses that infect epithelial cells of the skin and mucosa. Infection results in primarily benign, self-limiting warts or in epithelial tumors. Among the approximately 80 known types of HPVs, a certain number of so-called high-risk viruses, including notably HPV16 and HPV18, induce lesions with an increased risk of progression to cancer (for review, see de Villiers1). Epidemiologic studies point to these HPV infections as the major risk factor for the development of carcinoma of the uterine cervix as well as of anal, perianal, vulvar, and penile carcinomas. There is now evidence also for a connection between viral infection by HPV and the existence of a subgroup of squamous cell carcinoma of the head and neck, particularly of the oral cavity and oropharynx.2–5 However, the specific oncogenesis of these HPV-associated carcinoma entities is still largely unknown.
The HPV genome is organized into three general segments of unequal size: the long control region, the early (E) and late (L) genes (for review, see Scheffner et al.6). The E6 oncoprotein can form a complex with the host cell p53 tumor suppressor protein, thereby inducing p53 degradation and overcoming the G1/S checkpoint control in DNA-damaged cells.7 By acting as transcriptional activators or repressors, the E2 proteins regulate virus transcription and genome replication.6 Integration of viral DNA into the host cell chromosome leads to disruption of E2 gene,8 which has been reported to correlate with a poor disease prognosis in cervical carcinoma.9 The occurrence of E2 gene disruption and its prognostic impact in head and neck carcinoma has not been reported yet.
Several authors addressed the question whether HPV positivity in tumors of the head and neck is of prognostic relevance. Whereas some investigators were not able to detect any influence on survival,11, 12 others found the presence of HPV to be a marker for a favorable outcome.13–17 However, most of these studies comprised head and neck tumors of various origins that were approached with various treatment modalities. In an attempt to assess the impact of HPV infection on the response toward radiotherapy in a well defined patient cohort, we retrospectively determined the presence of the virus in tumors of individuals with oropharyngeal carcinoma who had undergone curative radiation treatment. Human papillomavirus positive oropharyngeal carcinomas were identified as radiosensitive subgroup of head and neck tumors with a better clinical outcome both in terms of local tumor control and overall survival. Moreover, they were found to be associated with a risk profile that differs from individuals with classic squamous cell carcinomas of the head and neck generally attributed to environmental exposures of tobacco and alcohol.
PATIENTS AND METHODS
Patients and Therapy
Between October 1991 and December 1997, 139 patients with biopsy-proven squamous cell carcinoma of the oropharynx were radically irradiated at the Department of Radiation Oncology, Inselspital, Bern. Forty of these subsequently were excluded from the current study owing to the following: the small size of the biopsy (17 patients), previous or synchronous malignancies (12 patients), irradiation after neck dissection (7 patients), intercurrent death (2 patients), lack of follow-up (1 patient), or the existence of distant metastasis at the onset of treatment (1 patient). The main clinical characteristics of the series of patients are shown in Table 1. All biopsies were taken before radiotherapy. Histopathologic diagnoses were reviewed by an experienced pathologist (J.L.). Drinking and smoking habits were assessed as follows: patients with consumption of tobacco or alcohol either in former times or at the time of recording the medical history were classified as smokers and/or alcohol consumers, respectively. More detailed information about the amount of exposure to tobacco and alcohol was not reliably available from the patients charts. All individuals were irradiated with curative intention. Twenty-seven patients received a concomitant chemotherapy. The patients were prepared for radiotherapy by placement of a thermoplast mask, planning computed tomography, two- or three-dimensional-based planning, and control by simulator and portal vision imaging device. Megavoltage radiotherapy was administered by linear accelerator in daily fractions, 5 times a week for 5–8 weeks. A median total dose of 74 grays (Gy; range, 54–80.5) was delivered. Baseline studies included physical examination, chest X-rays, panendoscopy of the upper aerodigestive tract, and magnetic resonance imaging, or computed tomography of the neck. The response to treatment was assessed 2–4 weeks after end of therapy and defined as follows: complete remission (complete disappearance of tumor manifestation) or incomplete remission (including partial remission, stable disease, or progressive disease). After treatment, all patients underwent clinical examinations and imaging on a regular basis.
Table 1. Bivariate Correlations with Human Papillomavirus Status
With the approval of the regional board of the Medical Ethics Commission, paraffin embedded tissue samples were obtained from the Pathology Department' archives. Oropharynx specimens (151) from the 99 patients were processed for DNA preparation. Ten-micrometer sections were taken from each specimen and covered with ATL lysis buffer (Qiagen Inc., Chatsworth, CA). After incubating for 10 minutes at 100 °C and centrifugation at 4 °C for 15 minutes, the paraffin was removed, and 20 μL of proteinase K (20 mg/mL; Fluka, Buchs, Switzerland) was added followed by an incubation at 55 °C for 12 hours. One hundred microliters of aqueous NaCl solution (5 M) and 90 μL of cetyltrimethylamoniumbromid (CTAB)/NaCl (10 g/0.7 M; Fluka) were added and incubated at 65 °C for 10 minutes. The sample was centrifuged with 800 μL of chloroform/isoamylalcohol (24:1; Fluka) for 5 minutes at 13000 rpm. The upper phase was centrifuged with 600 μL of phenol/chloroform/isoamylalcohol (24:24:1; Fluka) for another 5 minutes at 13,000 rpm. The upper phase then was mixed with 100% isopropanol at room temperature for 5 minutes followed by a centrifugation for 10 minutes at 13,000 rpm. The pellet was washed twice with 70% ethanol and resuspended in 50 μL of distilled water.
Polymerase Chain Reaction
Consensus L1-Polymerase Chain Reaction
We used the short polymerase chain reaction (PCR) fragment primer sequences developed by Kleter et al. that locate in the L1 open reading frame, allowing the amplification of the most important HPV subtypes.18 The multiplex PCR was performed in a final reaction volume of 50 μL, containing 10 μL of the DNA sample, 10 mmol/L Tris-HCl (pH 8.3), 50 mmol/L KCl, 2.5 mM MgCl2, 200 μM of each deoxynucleoside triphosphate, 100 pmol of forward and reverse primers (synthesized by Microsynth, Balgach, Switzerland), and 0.25 U of Taq-DNA-Polymerase (Roche Diagnostic, Basel, Switzerland). Conditions for PCR were as follows: preheating at 95 °C for 5 minutes was followed by 30 cycles at 95 °C for 30 seconds, 30 seconds at 52 °C, and 1 minute at 72 °C and a final extension for 7 minutes at 72 °C. Each PCR experiment was performed with a negative (water) and a positive control (DNA from an anal papilloma positive for HPV 31 [Institute of Pathology, University of Bern, Switzerland]). Amplification of sample DNA was controlled with β-actin PCR primers (Microsynth). All amplified DNA samples were sent for sequencing to the laboratory of Human Molecular Genetics, Department of Clinical Research, University of Bern, Switzerland. Identity of HPV subtypes was determined by searching the BLAST database of the National Center of Biotechnology Information, USA.
E2 specific PCR
The HPV16 positive tumors were tested for an intact E2 gene in three separate amplification reactions that allows to amplify three amplicons of various length determining the integrity of the E2 gene.9 The primer sequences used were as described in the same reference. Analysis of E2 was performed in a final reaction volume of 50 μL, containing 10 μL of sample DNA, 10 mmol/L Tris-HCl, pH 8.3, 50 mmol/L KCl 2.5 mM MgCl2, 200 μM each deoxynucleoside triphosphate, 10 pmol of forward and reverse primers (synthesized by Microsynth), and 0.25 U of Taq-DNA-Polymerase (Roche Diagnostic). Conditions for PCR were as follows: preheating for 5 minutes at 95 °C was followed by 40 cycles at 95 °C for 30 seconds, at 55 °C for 60 seconds and at 72 °C for 60 seconds, followed by a final extension at 72 °C for 7 minutes.
P53 and Intratumor Microvessel Density
P53 nuclear immunoreactivity and intratumor microvessel density (IMD) of the tumors have been determined previously in the same cohort of patients.19 Because IMD is closely related to treatment failure,19 this parameter was included in the multivariate analyses of the current study. For the current statistical analysis, the values of both p53 nuclear staining (percentages of positive nuclei per 500 cells) and IMD (CD-31 positive microvessels per high power microscopic field) were grouped according to their median.
Correlations between variables and HPV positivity were analyzed by implementing the logistic regression method in both an univariate and multivariate fashion, including a backward elimination procedure to remove variables with a P value ≥ 0.10. The qualifying criteria for inclusion in the multivariate analysis were: 1) P value less than 0.1 or 2) odds ratio (OR) less than 0.5 or greater than 2 in the univariate analysis. Variables were correlated with local failure free survival, disease free survival, and overall survival. The analysis of local failure free survival considered both local tumor progression after incomplete remission and local recurrence after complete remission as adverse events, whereas that of disease free survival took both local failure and distant metastasis into account. The analysis of overall survival included death from any cause. Survival was measured from the time when therapy was initiated to that when the first adverse event was detected or to the date of the last follow-up. Deaths due to non-tumor-related causes were censored, except in the analysis of overall survival. Survival curves were plotted according to the Kaplan–Meier method, the log-rank test being used to determine significant differences between these. A Cox regression was performed to calculate the risk ratios. Qualifying criteria for inclusion in the multivariate Cox regression analysis were as follows: 1) P value less than 0.1 or 2) risk ratio (RR) less than 0.5 or greater than 2 in the univariate analysis. A backward elimination procedure then was performed to eliminate nonsignificant variables (P ≥ 0.1). Statistical analyses were performed using the SPSS package (version 10.0.0; SPSS Inc., Chicago, IL).
Correlation between HPV Positivity and Clinicopathologic Variables
There were 14 HPV positive tumors of 99 (14%), the subtypes being HPV16 (11 cases), HPV33 (1 case), HPV35 (1 case), and HPV45 (1 case). Six HPV positive tumors were tonsillar carcinomas and eight were tumors of the base of the tongue. Human papillomavirus positivity was closely linked to female gender (OR, 5.75; P = 0.004), age older than 56 years (OR, 7.42; P = 0.012), nonsmokers (OR, 21.33; P = 0.00001), and alcohol abstainers (OR, 5.35; P = 0.012; Table 1). Moreover, HPV positivity was inversely associated with p53 nuclear immunoreactivity, grouped by the median (OR, 0.06; P = 0.008). There was no correlation between T classification, N classification, grade of differentiation, or IMD and HPV positivity. In the multivariate analysis with backward elimination of nonsignificant variables, the parameters age, smoking habits, and p53-immunoreactivity retained their significant association with HPV positivity (Table 1).
The Kaplan–Meier survival curves revealed an improved local control (P = 0.05; Fig. 1A), a trend for a better disease free survival (P = 0.13; Fig. 1B), and a significantly increased overall survival (P = 0.036; Fig. 1C) for HPV positive tumors. The respective HPV-assigned risk ratios were 0.33 (P = 0.063) for local failure free survival, 0.50 (P = 0.14) for disease free survival, and 0.35 (P = 0.046) for overall survival (Table 2). In the multivariate analysis of local failure free survival with backward elimination of nonsignificant variables, HPV positivity retained its significant correlation with improved local control (RR, 0.31; P = 0.048). However, in the multivariate analysis of overall survival, the significant association of HPV positivity with a better outcome was lost. In line with the previously reported study,19 both T classification and IMD were closely related to all three survival endpoints in both the univariate and multivariate analysis. Analysis of the two most important risk factors for the development of squamous cell carcinoma of the head and neck revealed both the absence of smoking (RR, 0.31; P = 0.031) and alcohol consumption (RR, 0.32; P = 0.031) to be significantly associated with an improved overall survival in the univariate analysis and—in the case of absence of alcohol consumption—also in the multivariate analysis (RR, 0.35; P = 0.034 [Table 2]). In terms of local failure free survival and disease free survival, the absence of smoking showed a trend for a better outcome (RR, 0.35 and 0.41, respectively).
As determined by PCR,9 4 of the 11 HPV16 positive tumors showed a disrupted E2 gene, i.e., at least one locus of the three investigated gene loci could not be amplified in those samples. All of them had a loss of the E2B gene region. The Kaplan–Meier survival curves of HPV16 positive tumors for local failure free survival, disease free survival, and overall survival revealed an advantage for tumors with a intact E2 gene (Fig. 2). Among the 11 patients with HPV16 positive tumors, only 2 suffered treatment failure at the site of the primary tumor. In these two tumors, a disrupted E2 gene was found.
In general, the development of squamous cell carcinoma of the head and neck is strongly associated with alcohol abuse and smoking.20, 21 However, there is increasing epidemiologic evidence for the existence of a subgroup of these cancers with a different oncogenesis based on HPV infection.11, 16, 22–24 The estimates of the HPV prevalence in head and neck squamous varied markedly among various reports [range, 8–100%; for review, see Steinberg and DiLorenzo3). Gillison et al. reported that among head and neck carcinomas of different origin, oropharyngeal tumors were the most likely to be HPV positive.16 Mellin et al. found an overall HPV prevalence of 43% in tonsillar carcinoma in a patient cohort that was treated either with surgery followed by radiation or with radiation alone.17 In contrast, we found a much lower prevalence of 14% in our series (Table 1). Looking at other clinicopathologic variables, Mellin et al. described a strong relation between T classification and HPV positivity: Tumors with advanced T classifications (T3/4) were less likely to be HPV infected.17 Because of the selection of patients undergoing curative radiotherapy, individuals of our study population had advanced tumor classifications in most cases (88% T3/4). This bias may have contributed to the low prevalence of HPV positive tumors in our study. In addition, geographic variation of HPV positivity cannot be excluded.25
The analysis of the correlation between HPV infection and the consumption of tobacco and alcohol may contribute to the understanding of HPV-associated oncogenesis in tumors of the head and neck. Reports addressing this issue are unequivocal. Badaracco et al.24 as well as Schwartz et al.23 were not able to detect any association between the presence of HPV and environmental exposures to tobacco and alcohol. In contrast, in the study by Gillison et al. including 253 patients with head and neck tumors from various sites of origin, a trend for an association between nonsmoking and HPV-infected tumors was discovered, but only in the subgroup of 60 patients with oropharyngeal carcinoma.16 A significant association between the absence of alcohol consumption and HPV positivity was found in the same patient subgroup. In concert with these results and others,14 we found a strong correlation between the absence of smoking and alcohol consumption and the presence of HPV-infected tumors (Table 1). The absence of these well known risk factors for head and neck carcinoma in HPV positive oropharyngeal tumors supports a causal association with HPV infection and therefore a different oncogenesis. In addition to the association with nonsmokers and alcohol abstainers, we found a higher risk for HPV positivity in female patients and individuals older than the median age of 57 years. Mellin et al. described a significant association between female gender and increased risk for HPV positive tumors as well.17 Thus, female gender and advanced age may represent components of a specific risk profile of patients with HPV-associated squamous cell carcinoma of the oropharynx.
Most studies on the prognostic impact of HPV positivity in head and neck tumors found its presence to be a marker for a favorable outcome.13–17 In the study by Mellin et al., patients with HPV positive tonsillar carcinomas were described to have an improved survival. Because all of their patients received radiation therapy—either in a postoperative setting or as curative treatment—this observation might relate to an enhanced radiosensitivity. By focussing on patients who have undergone curative radiation therapy, we now were able to directly address this question. Indeed, we found HPV positive tumors to have a higher local control rate with a risk ratio for local treatment failure of 0.33 (Fig. 1; Table 1). The statistical power was limited because of the low prevalence of HPV positive tumors in our series of patients. However, the finding that the significant association between HPV positivity and improved local control was retained in the multivariate Cox regression analysis strongly supports the validity of our findings. Moreover, patients with HPV positive tumors also had an increased overall survival, which mainly depends on local tumor control in individuals with this tumor entity. The association of HPV positivity with the absence of smoking and alcohol consumption may further have contributed to the improved overall survival.
The reasons for an increased radiocurability of HPV positive tumors are not defined yet. In our series, we found HPV positive tumors to be strongly correlated with less p53 nuclear immunostaining (Table 1). This is in line with several studies reporting a decreased rate of p53 mutations among HPV positive tumors.11, 16, 26, 27 Furthermore, it has been shown that the degradation of wild-type p53 by the E6 viral protein is not functionally equivalent to the loss of p53 function through somatic mutations of the p53 gene.28, 29 Thus, HPV positive tumors appear likely to express significant amounts of functional p53, what may contribute to the cellular radioresponsiveness.30, 31 Whether the viral oncoproteins E6 and E7 themselves may render cells more susceptible for DNA damage-induced apoptosis remains controversial.32, 33 Besides radiosensitization by viral proteins, the higher radiocurability of HPV-infected tumors also may be attributed to yet undefined differences in their specific oncogenesis. Finally, parameters related to the specific risk profile of patients with HPV positive oropharyngeal carcinoma, such as the absence of exposure to tobacco and alcohol during radiotherapy as well as patients compliance with treatment, also may contribute to the higher tumor control rates in these individuals.
Disruption of the viral E2 gene has been shown to be associated with a poor prognosis in patients with cervical carcinoma.9, 10 Because of these clinical reports and the referred interference of the E2 protein with the regulation of apoptosis and cell cycle control,34 we performed a subanalysis of the integrity of the E2 gene in HPV16 positive tumors. Because of the very limited number of patients, the results have to be interpreted with great caution. Our analyses may indicate that the local control rate as well as the probability of overall survival is the highest in cases of HPV positive tumors with an intact E2 gene (Fig. 2). Both of the two HPV16 positive tumors that failed to be controlled by curative radiotherapy harbored a disrupted E2 gene. Even though the validity of these findings is weak, we propose to include the assessment of the E2 gene integrity in future investigations on the role of HPV in cancer therapy.
In conclusion, our results indicate that HPV positivity designates a specific subgroup of oropharyngeal squamous cell carcinomas of the oropharynx that arise preferentially among individuals with no consumption of tobacco and alcohol and that have a favorable outcome attributable to an increased sensitivity toward radiotherapy. This may have an impact on future developments of individualized treatment strategies.