Abstract
Background/Aim: Patients with human papillomavirus (HPV)-positive tonsillar and base of tongue cancer have a better outcome than those with corresponding HPV-negative tumors (80% vs. 40% 5-year disease free survival with conventional radiotherapy). They should not all need chemoradiotherapy, but before tapering treatment, more markers are needed to predict treatment response. In the present study, human leukocyte antigen (HLA) - HLA-A*02 was analyzed with HPV as a prognostic factor for tonsillar and base of tongue cancer. Patients and Methods: Pre-treatment biopsies, previously tested for HPV DNA, from 425 patients diagnosed with tonsillar and base of tongue cancer between 2000-2009 at the Karolinska University Hospital were examined for HLA-A*02. Results: HLA-A*02 was present in 144/305 (47.2%) of the HPV-positive and 63/120 (52.8%) of the HPV-negative tumours. Among 383 patients treated with curative intent, absence of HLA-A*02 was correlated with increased disease-free survival in the HPV-positive (p=0.016), but not in the HPV-negative group. Conclusion: Absence of HLA-A*02 correlated with better disease-free survival for patients with HPV-positive tonsillar and base of tongue cancer.
- HLA-A*02 genotype
- human papillomavirus (HPV) oropharyngeal squamous cell carcinoma
- tonsillar cancer
- base of tongue cancer
- clinical outcome
In 2007, the association of human papillomavirus (HPV) with oropharyngeal squamous cell carcinoma (SCC), where tonsillar and base of tongue cancer (TTSCC) dominate, was recognized by the International Agency for Cancer Research-based on research by others and us (1-3). Noticeably during the past decades, the incidence of HPV-positive oropharyngeal SCC has increased in many Western countries (4-15). Furthermore, HPV-positive, clearly have a much better clinical outcome than the respective HPV-negative cancer with a roughly 80% vs. 40-60% 5-year disease free survival (1-3, 16-23). However, during the past decade, treatment for head and neck squamous cell carcinoma, including TTSCC, has been intensified with chemoradiotherapy, and in some cases with cetuximab, with increased severe side-effects (24). It is likely that most patients with HPV-positive TTSCC do not require intensified therapy. However, since not all patients with HPV-positive TTSCC survive, additional markers are needed to select patients with HPV-positive TTSCC with a high probability of a good response to therapy (25-31). Some attempts to find biomarkers in this direction have already been made and it has been shown that lack of or low expression of human leukocyte antigen (HLA) class I, cell surface glycoprotein CD44 or high CD8+ tumour-infiltrating lymphocyte counts in HPV-positive TTSCC all are favourable prognostic factors and indicate a 95-100% probability of 5-year disease-specific survival (25, 26, 28, 30, 31). However, these markers only select a proportion of all patients with good survival prospects, hence additional markers are necessary.
The highly polymorphic HLA-A locus is located on chromosome 6 in humans and encodes peptide-presenting proteins that perform the first step in adaptive immunity and respond to foreign pathogens or detected abnormalities (32, 33). Previously, we showed that the HLA-A*02 allele, a common allele in the Scandinavian population, is not only over-represented, but also presents a strong negative prognostic factor in patients with stages III-IV ovarian cancer (34-36). In addition, a correlation with prognosis and HLA-A*02 has also been demonstrated for prostate cancer and malignant melanoma (36, 37). Finally, HLA-A*02 was associated with an increased incidence of vulvar cancer, while data for cervical cancer are still inconsistent, still possibly indicating a higher susceptibility to some types of HPV-induced cancer (38, 39).
Based on the data above and our method of assay for HLA-A*02 in formalin fixed paraffin-embedded (FFPE) tissues (40), we herein analyzed for the presence of the HLA-A*02 gene in 425 patients with TTSCC in relation to HPV status and clinical outcome.
Patients and Methods
Patients and biopsies. Between 2000 and 2009, 445 patients were diagnosed with tonsillar SCC or base of tongue SCC (ICD-10 codes: C09 and C01.9, respectively) at the Karolinska University Hospital (26). Of these, 425 (96%) had diagnostic pre-treatment biopsies previously tested for HPV DNA, with sufficient material to test for HLA-A*02, since blood was not available (26). Of these 425 patients, all 383 treated with curative intent (but not the 42 receiving palliative treatment) were included in the survival analysis. For 260/383 (68%), therapy included conventional radiotherapy (2.0 Gy/day, for 6.5-7 weeks, total dose: 68 Gy) or accelerated radiotherapy (1.1+2.0 Gy/day for 4.5 weeks, total dose: 68 Gy), while some 123/383 (32%) also received induction chemotherapy and concomitant radiotherapy. Moreover, some patients also had interstitial radiation (brachytherapy) (total dose of 78 Gy), while patients with nodal disease underwent neck dissection 6-8 weeks after they had completed radiotherapy. Patients were then followed-up by clinical examination every three months for the first two years, and every six months the third year.
Patients' characteristics and survival data were obtained from their records. The study was carried out according to permissions 2003/507, 2005/431-31/4, 2005/1330-32 and 2009/1278-31/4 from the Regional Ethical Committee at Karolinska Institute, Stockholm, Sweden.
DNA extraction. DNA from 425 FFPE samples was previously extracted using High Pure RNA extraction kit (Roche, Molecular Biochemicals, Mannheim, Germany), according to the manufacturer, excluding the DNase treatment as described elsewhere (19, 26). Blank controls previously used to exclude cross-contamination of HPV DNA were also included to exclude cross-contamination of HLA-A*02 between samples.
HPV analysis. The presence of HPV DNA of 27 HPV types and human beta-globin was analysed by a bead-based multiplex assay evaluated on a MagPix instrument (Luminex Corporation, Austin, TX, USA), as described previously (26).
HLA-A*02 analysis, primer design and amplification. PCR-amplifiable DNA from 425 FFPE samples, previously tested for the presence of HPV and the housekeeping gene beta-globin, was included in the current analysis (26). HLA-A*02 primers (forward A*577LL and reverse A*503invLL) were made to amplify HLA-A*02 exon 2, with exception of the A*020109, 0248, 0250, 0255 alleles, which are uncommon among Swedes, and have been described in detail previously (37, 40).
The reaction mix consisted of: 5 μl 10× PCR Buffer PE, 8 μl DNTP (1.25 mM/dNTP), 3 μl MgCl2 (25 mM), 2 μl forward primer (10 pmol/μl), 2 μl reverse primer (10 pmol/μl), 27.1 μl distilled water, 0.4 μl Taq polymerase, and a 2.5 μl sample of DNA (50 ng), making a total volume of 50 μl.
Amplification was performed in an automated thermocycler (GeneAmp PCR system 9700; Applied Biosystems, Foster City, CA, USA). The cycles consisted of an initial denaturation of 5 min at 94°C, followed by 40 cycles of 94°C for 30 s, 56°C for 30 s, 72°C for 45 s and finally a step of 72°C for 5 min. Water was used as a negative control. DNA extracted from an HLA-A*02 gene-positive fresh blood sample was used as a positive control (obtained according to ethical permission 2003/507 from the Regional Ethical Committee at Karolinska Institute).
The obtained PCR products were evaluated on a 2.0% agarose gel stained with Gel Green (Nucleic acid Gel Stain, diluted ×10,000 in water; Biotium, Hayward, CA, USA) and visualized under UV light. The expected product was 149 bp for exon 2 of HLA-A*02.
Statistical analysis. The Chi-square test was used for categorical data analysis and independent samples t-test to compare mean values. Patient survival was measured in days from the date of diagnosis until the occurrence of an event or until three years (1,095 days) after diagnosis when censoring of the patients took place. An event defined as death of the patients due to any cause (overall survival, OS) or recurrence of the disease (disease-free survival, DFS). Patients without any prior recurrence that died were censored at day 0 when calculating DFS. Kaplan–Meier estimator of survival was used for the analysis in order to estimate DFS, and OS and differences in survival were tested using the log-rank test. Cox-regression analysis (Cox proportional hazards model) was used for the calculation of the unadjusted and adjusted hazard ratios (HRs). All of the tests used were performed at a two-sided 5% level of significance. All calculations and analyses were performed using IBM, SPSS Statistics (version 21, New York, NY, USA).
Results
Presence of HLA-A*02 estimated by exon 2 PCR of FFPE samples. The characteristics of the 425 patients and their tumors (307 tonsillar SCCs and 118 base of tongue SCCs) are shown in Table I. Patients with HPV-positive tumours (N=305) were younger (mean and median age of 60 and 59 years, respectively) than those with HPV-negative tumours (N=120) (mean and median age 65 years) (p<0.001) (Table I). In total, 207/425 (48.7%) samples were HLA-A*02-positive (shown for 3/6 patients in Figure 1), with 144/305 (47.2%) in the HPV-positive and 63/120 (52.8%) in the HPV-negative group. There were no differences in HLA-A*02 frequency irrespective of HPV status and sex (p=0.326 and p=0.744, respectively; data not shown). There was no difference in HLA-A*02 frequency in relation to tumour stage, nodal disease, or distant metastasis for the whole group (p=0.867, p=0.299, and p=0.527 respectively) irrespective of HPV status (data not shown).
Presence or absence of HLA-A*02 and HPV status in correlation to clinical outcome. Out of the 425 patients, 383 were treated with intention to cure. Overall, 290/305 (95%) of the patients (mean and median age of 59 years) with HPV-positive tumours were treated with curative intent in comparison to 93/120 (77.5%) of the patients (mean and median age, 63 and 62 years, respectively) with HPV-negative tumours (data not shown).
Absence of HLA-A*02 was correlated to better 3-year DFS (log-rank test, p=0.046) in the whole group of patients, and especially for the HPV-positive group (log-rank test, p=0.016), but not for the HPV-negative group (log-rank test, p=0.968) illustrated as Kaplan–Meier curves shown in Figure 2 (A-C). When partitioned by sex, absence of HLA-A*02 was correlated to a better clinical outcome only in males for the whole group (log-rank test, p=0.004), and only for males in the HPV-positive group (log-rank test, p=0.003) (data not shown).
Characteristics of patients with tonsillar squamous cell carcinoma and base of tongue squamous cell carcinoma and their tumors.
Performing a univariate Cox regression analysis, HLA-A*02, age and HPV status were independently significantly correlated to a better disease-free survival at p=0.048, p<0.001 and p<0.001, respectively. In a multivariate analysis for the whole group including sex, age, tumor stage, HPV and HLA-A*02, a lower age and positive HPV status correlated to a better DFS (p=0.003 and p<0.001, respectively) (data not shown). When using a Cox regression model separately for the HPV-positive and negative groups, in the HPV-positive group, absence of HLA-A*02 and lower age correlated to a better DFS (p=0.036 and p= 0.003, respectively) (Table II). A similar multivariate analysis for the HPV-negative group did not result in statistical significance for any of the included parameters (Table II).
Example of results obtained after loading samples on a 2% agarose gel. Positive samples present a band of approximately 150 bp (P3-P5, positive samples). MW: Molecular weight; −: negative control; +: positive control; P1-P6: patient samples.
There was no correlation between 3-year OS and presence or absence of HLA-A*02 in the whole group, nor when stratified according to the HPV status of the tumors, as illustrated by the Kaplan–Meier curves shown in Figure 2 (D-F). Similarly, no differences were observed in the univariate and the multivariate analyses (data not shown).
Kaplan–Meier curves for disease-free survival (DFS) (A-C) and overall survival (OS) (D-F) of patients with tonsilIar and base of tongue squamous cell carcinomas treated with intention to cure and tested for HLA-A*02. Kaplan–Meier curves for the whole patient cohort according to HLA-A02* status irrespective of the HPV status of the tumor (A and D). Kaplan–Meier curves for DFS and OS according to HLA-A*02 status for HPV-positive (B and E) HPV-negative (C and F) cases.
Univariate and multivariate analyses for disease-free survival of patients with tonsillar squamous cell carcinoma and base of tongue squamous cell carcinoma.
Discussion
In the present study, the presence of the HLA-A*02 gene was investigated in 307 tonsillar SCCs and 118 base of tongue SCCs and correlated to the presence of HPV and clinical outcome. HLA-A*02 was found in approximately 50% of patients, irrespective of HPV status, and absence of HLA-A*02 resulted in a significantly better 3-year DFS when considering the whole group (p=0.046), and for patients with HPV-positive tumors (p=0.016). However, there was no correlation between absence of HLA-A*02 and OS.
The fact that the presence of HLA-A*02 is similar in patients with HPV-positive and HPV-negative TTSCC suggests that there is most likely no specific sensitivity to the development of HPV-positive tumours correlated to the presence of this antigen. Neither is it possible to argue that the frequency of HLA-A*02 in the examined cohort is higher than in the general Swedish population, reported to be 58% (for the phenotype) in 2000 (36).
Notably, the absence of HLA-A*02 correlated with a significant better 3-year disease-free survival for the whole group (p=0.046), especially in patients with HPV-positive tumors (p=0.016), and in males with HPV-positive tumours (p=0.003). This is similar to our previous findings with regard to patients with ovarian and prostate cancer and malignant melanoma (34-37). Notably, absence of HLA-A*02, did not correlate to survival in patients with HPV-negative tumours or females irrespective of HPV status. It could be that absence of HLA-A*02 and the association with a better 3-year DFS is mainly attributed to the presence of the HPV-positive group, and the fact that this reaches statistical significance also in the whole cohort is due to the fact that the majority of the patients treated with curative intent (290/383 78%) had HPV-positive tumors and were males. This is also supported by the multivariate analysis, where absence of the HLA-A*02 remained a prognostic factor. On the other hand, one could also argue that the number of patients with HPV-negative tumors and the number of females with HPV-positive and HPV-negative tumors were fewer and we therefore did not reach statistical significance in these groups. The latter possibility is supported by the fact that not only absence of HLA-A*02, but also age disappears as a positive prognostic factor in the HPV-negative group upon the multivariate analysis.
There were no differences with regard to clinical outcome in the absence of HLA-A*02 and 3-year OS, neither for the whole group, nor when stratified by HPV status. We do not know why this is the case. On the other hand, it is likely that OS is influenced by factors other than the disease itself, as is the case for other tumour entities, and thus the presence or absence of HLA-A*02 may not influence this outcome in the same way as it does for DFS.
There are some limitations to our study. First of all, the study was retrospective, but this might not necessarily influence the results. Secondly, there were relatively fewer patients with HPV-negative tumors and also fewer females. In addition, somewhat fewer patients in the HPV-negative group were treated with curative intent, and this could possibly be because they were generally older. This is unfortunate, but is a reflection of the numbers of patients with HPV-negative tumours, the older age of this group, and the lower numbers of females with these types of cancer during this period. Consequently, due to the limited number of patients with HPV-negative TTSCC and of female patients, it would be of interest to conduct additional studies with larger patient numbers. Only in this way will it be possible to elucidate whether absence of HLA-A*02 is a favourable prognostic marker only for patients with HPV-positive TTSCC, especially males, or whether this is true also for patients with HPV-negative tumours and females with HPV-positive or HPV-negative tumours. Finally, one can also argue that this study does not include all types of oropharyngeal SCC cases. The reason for this is in fact that TTSCC are the only types of oropharyngeal SCC arising from Waldeyer's lymphoid ring, where HPV is most commonly found and where a positive HPV status is correlated to a positive clinical outcome (14,15). In contrast to TTSCC, oropharyngeal SCC at other locations has a lower HPV prevalence, and the presence of HPV does not have a similar impact on clinical outcome (41).
In conclusion, this study shows that the presence of HLA-A*02 is similar in patients with HPV-positive and HPV-negative TTSCC. In addition, our data suggest that absence of HLA-A*02 could be a favourable prognostic factor for patients with HPV-positive tumours. However, additional studies with larger numbers of patients is important in order to confirm these findings, and to elucidate whether absence of HLA-A*02 is also a favourable factor for patients with HPV-negative TTSCC and for women with such HPV-positive tumours.
Acknowledgements
This work was supported by The Swedish Cancer Society, The Swedish Medical Research Council, The Stockholm Cancer Society, Henning and Ida Perssons Research Foundation, Karolinska Institute, the Stockholm City Council, Swedish Cancer and Allergy Foundation, and the Stockholm County ALF
- Received February 18, 2014.
- Revision received March 9, 2014.
- Accepted March 10, 2014.
- Copyright© 2014 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved