The role of human papillomavirus in head and neck squamous cell carcinoma: A case control study on a southern Chinese population
Abstract
HPV plays a role in the development of a portion of head and neck squamous cell carcinoma (HNSCC), but only limited information on its role in southern Chinese population is available. A multicenter case-control study was conducted. HPV type, viral integration, E6/7 mRNA expression status, and TP53 mutation were determined. A total of 228 HNSCC were recruited including 137 (60.1%) oral SCC, 34 (14.9%) oropharyngeal SCC, 31 (13.6%) laryngeal SCC, 21 (9.2%) hypopharyngeal SCC, and 5 (2.2%) lip and paranasal sinus SCC. High-risk HPV infection was found in 7.5% (17/228) of HNSCC, but only a small proportion of samples had evidence of viral integration (5.3%, 12/228) or E6/7 mRNA expression (4.4%, 10/228). HPV infection with oncogenic phenotype (integration and E6/7 mRNA expression) was significantly more common in oropharyngeal SCC than controls (9/34, 26.5% vs. 0/42, 0.0%, P < 0.001). Smoking showed a significant association with HNSCC, oropharyngeal SCC, and laryngeal SCC. TP53 mutation was associated with HNSCC (P < 0.001). Older age, TP53 mutation, and HPV16 infection with oncogenic phenotypes were independently associated factors for HNSCC with odds ratios of 1.03 (1.02–1.05), 3.38 (1.71–6.66), and 9.19 (1.13–74.68), respectively. High-risk HPV infection of head and neck mucosa is not uncommon in the Hong Kong population. This study found that 26–30% of oropharyngeal carcinoma was associated with HPV infection, mostly HPV16, and that smoking which predisposes to TP53 mutations was another important risk factor. J. Med. Virol. 88:877–887, 2016. © 2015 Wiley Periodicals, Inc.
INTRODUCTION
Head and neck squamous cell carcinoma (HNSCC) comprises a diverse group of cancers that often share similar risk factors such as smoking and alcohol consumption. Despite a decline in the overall incidence of HNSCC over the past decades as a result of anti-smoking legislation, the burden of disease remains substantial in most parts of the world. Globally, more than 550,000 cases of HNSCC are reported annually [Jemal et al., 2011]. In recent years, the role that infection with high-risk types of human papillomavirus (HPV) plays in the etiology of at least a portion of HNSCC has become recognized [Marur et al., 2010]. While a strong body of evidence supports an etiological role of HPV in the development of oropharyngeal cancer, especially of the palatine and lingual tonsils [Feller et al., 2010], the role of HPV in the development of carcinoma in other head and neck mucosal sites remains unclear [Termine et al., 2008].
The oncogenic pathway of HPV is believed to be multistep involving various host, genetic, and environmental factors. Exposure to certain environmental factors such as carcinogenic pollutants may well be more frequent in some parts of the world than others. For instance, a recent meta-analysis revealed a high prevalence of HPV in Chinese patients with laryngeal cancer [Li et al., 2013; Zhang et al., 2014]. Furthermore, the prevalence of HPV in esophageal cancer was found to vary widely across the world, with a high prevalence found in certain areas of China where esophageal cancer was common [Li et al., 2014]. In this study, we explored the association between HPV infection and the development of HNSCC in an ethnically southern Chinese population, for whom data are scarcely available.
METHODS
Patient Recruitment
A multicenter cross-sectional case-control study was conducted from January 2012 to December 2014. Patients with a suspected benign or malignant lesion of the head and neck mucosa were invited to participate. Nasopharyngeal carcinoma was excluded as a known viral etiology, Epstein–Barr virus, has been established. Recurrent cancers and metastases from a primary site outside of the head and neck region were excluded. Tissue samples were collected from the diagnostic biopsy or definitive resection. When fresh tissue was not available, formalin-fixed paraffin-embedded tissues were used. A written informed consent was obtained from all participants. The study was approved by the Joint CUHK-NTEC Clinical Research Ethics Committee and the respective ethics committee of the participating hospitals.
HPV DNA Detection and Genotyping
Total DNA was extracted from fresh frozen tissue samples using the QIAamp DNA Mini Kit (Qiagen), while formalin-fixed paraffin-embedded tissue samples were first dewaxed with two washes of xylene and ethanol prior to DNA extraction using the QIAamp DNA Mini Kit as described previously [Chan et al., 2001, 2009]. The length and quality of extracted DNA preparations were assessed by PCR targeting the housekeeping gene, beta-globin [Chan et al., 2001]. The presence of HPV DNA was detected by nested consensus PCR using primers PGMY09/11 and GP5+/6+_HK52 modified from the published methodology [Chan et al., 2001]. Positive amplicons were sequenced and the HPV type was identified based on more than 95% homology with the reference sequence available in GenBank (http://www.ncbi.nlm.nih.gov/genbank/).
HPV E6/E7 mRNA Expression
To assist with the analysis of the oncogenic role of high-risk HPVs found, the expression of mRNA encoded viral oncoproteins E6/E7 was examined. As the E6 and E7 open reading frames (ORFs) are next to each other, a single strand of pre-mRNA molecule was transcribed from the E6 and E7 ORFs, which then underwent different patterns of splicing. The most dominant and important spliced E6/E7 mRNA was the E6*I mRNA, which was measured in this study. Reverse transcription real-time PCR (RT-qPCR) used were based on primers and probes that were specifically designed to cross the splicing site of E6*I mRNA as shown in Table S1. The reaction contained 2 µl of purified RNA, 0.25 µl of SuperScript® III/Platinum® Taq Mix (Life Technologies), 12.5 µl of 2× Reaction Mix in a final reaction volume of 25 µl. The cycling conditions were 50°C for 15 min, 95°C for 2 min, followed by 40 cycles of 95°C for 15 sec, and 55°C (for HPV16) or 58°C (for HPV31 and HPV52) for 30 sec using the SuperScript® III Platinum® One-Step qRT-PCR Kit with ROX (Life Technologies).
Fresh frozen tissue samples were preserved with RNAlater (Life Technologies), and subjected to RNA extraction by using the QIAGEN RNeasy kit (Qiagen). RNA was extracted from formalin-fixed paraffin-embedded tissues using the QIAGEN RNeasy FFPE kit (Qiagen). The quality of extracted RNA was assessed by RT-qPCR with primers and probe specifically targeting the splicing region of mRNA encoded by the housekeeping gene RPS18: (forward primer RPS18-F125)—CCT TTG CCA TCA CTG CCA TT, (reverse primer RPS18-R254)—ACT GGC GTG GAT TCT GCA TAA, (probe RPS18-224P)—5′FAM CTG AGG ATG AGG TGG AAC 3′MGBNFQ. The RT-qPCR reaction was the same as described above. The primer and probe concentrations were 0.3 µM and 0.2 µM, respectively. The cycling conditions were 50°C for 30 min, 95°C for 2 min, followed by 40 cycles of 95°C for 15 sec and 60°C for 30 sec using the SuperScript® III Platinum® One-Step qRT-PCR Kit with ROX (Life Technologies).
HPV Integration
In addition to E6/E7 mRNA detection, the viral integration status of high-risk HPV types was examined to assist with the assessment of the oncogenic role of HPV. Briefly, viral integration that disrupts the E2 ORF of HPV genome results in loss of control on the expression of viral oncogenes E6 and E7. By comparing the number of copies of E2 and E7 genes in a given sample, one can estimate the physical status of the viral genome. In this study, qPCR methods adopted from previous studies were applied to quantify the levels of E2 and E7 genes [Cheung et al., 2006]. The primer sequences used are shown in Table S2. For the purpose of analysis in this study, a 10-fold (≥3 threshold cycle, CT) difference in genome copies between E2 and E7 was regarded as an indication of HPV integration.
TP53 Mutation
DNA extracted from tumor or control tissues were used to analyze mutations in the TP53 gene. Four primer sets were used to amplify the genomic DNA sequences spanning exons 4–9 of TP53 where the majority of mutations were found (http://p53.iarc.fr) (Table S3). The PCR was conducted in a 25-μl reaction mix containing 2.5 μl of DNA, 1.25 U of HotstarTaq Plus polymerase (Qiagen), 0.2 mM of dNTPs, and 0.4 μM of primers. The thermal cycling conditions were an initial denaturation at 95°C for 5 min, followed by 40 cycles of denaturation at 94°C for 1 min, primer annealing at 60°C (Exon 4 and 5-6 PCR), or 64°C (Exon 7 PCR) or 55°C (Exon 8-9 PCR) for 1 min, and extension at 72°C for 1 min, and then a final extension at 72°C for 10 min. The PCR products were analyzed by electrophoresis using a 1.5% agarose gel and visualized using SYBR-SAFE pre-staining method. PCR products were purified with QIAquick PCR Purification kit (Qiagen), and submitted to a commercial laboratory (Macrogen) for sequencing. Polymorphisms and mutations were determined based on the reference sequences available from the International Agency for Research on Cancer (IARC) TP53 database (http://p53.iarc.fr).
Statistical Analysis
Subjects with malignant tumors were grouped according to histological type. For the purpose of analysis, squamous cell carcinoma (SCC) included poorly differentiated carcinoma. Subgroup analyses of malignant cases were performed according to the site of tumor. Salivary gland tumors were separated from HNSCC as they are different in terms of histology and etiology. Subjects with non-malignant lesions including normal or inflammatory lesions were regarded as the “control group”, and further divided according to the site of specimen collection. Differences in proportion of categorical variables between cancer groups and the corresponding anatomical site-matched control groups were assessed by χ-squared or Fisher's exact test as appropriate, using the StatCalc program of Epi Info (Version 7.0.8.3, Centers for Disease Control and Prevention, Georgia). Differences in age distribution were assessed by t-test, and independent associations were assessed by logistic regression using SPSS (version 20, IBM).
RESULTS
Tissue samples were obtained from 539 subjects. Of the 256 subjects who had histology-confirmed malignant disease, the majority (228/256, 89.1%) were HNSCC, whereas 13 of 256 (5.1%) were salivary gland malignant tumors and 15 of 256 (5.9%) were cancers of other histological types (Table I). Among the 228 patients with HNSCC, the majority (137/228, 60.1%) were in the oral cavity, followed by the oropharynx (34/228, 14.9%) including base of tongue and palatine tonsil, and the rest included larynx (31/228, 13.6%), hypopharynx (21/228, 9.2%), lip (3/228, 1.3%) and paranasal sinuses (2/228, 0.9%) (Table I). Of those 283 recruited subjects with non-malignant disease, the lesions were benign or inflammatory or normal tissue collected from various sites in the head and neck. These subjects were divided into different control groups according to the anatomical site of the tissue sample (Table I). The 228 subjects with HNSCC were aged 27–92 years (mean 62.6, standard deviation [SD]: 13.3), whereas the 283 subjects in the control group were aged 20–91 years (mean: 57.0, SD: 14.4). A male predominance was observed for subjects with either malignant or benign lesions across all anatomical sites except for the lip and paranasal sinuses (Table I).
| Disease status | N | Male:female | Age (mean, range, standard deviation) years | HPV DNA positive N (%) | TP53 mutation N (%) | Ever smoked N (%) |
|---|---|---|---|---|---|---|
| All malignant cases from the head and neck region | 256 | 2.1:1 | 62.7, 21–92, 13.5 | 27 (10.5) | 52 (20.3) | 124 (48.4) |
| Head and neck squamous cell carcinoma (HNSCC) | 228 | 2.1:1 | 62.6, 27–92, 13.3 | 22 (9.6) | 47 (20.6) | 112 (49.1) |
| Oral cavity SCC | 137 | 1.2:1 | 63.5, 27–92, 14.3 | 3 (2.2) | 26 (19.0) | 47 (34.3) |
| Oropharyngeal SCC | 34 | 7.5:1 | 56.4, 30–88, 13.6 | 10 (29.4) | 8 (23.5) | 23 (67.6) |
| Laryngeal SCC | 31 | All male | 64.8, 48–84, 9.0 | 5 (16.1) | 7 (22.6) | 25 (80.6) |
| Hypopharyngeal SCC | 21 | 4.3:1 | 63.0 49–83, 9.4 | 3 (14.3) | 4 (19.0) | 16 (76.2) |
| Lip and paranasal sinus SCC | 5 | 0.3:1 | 65.2, 46–82, 14.5 | 1 (20.0) | 2 (40.0) | 1 (20.0) |
| Salivary gland malignant tumours | 13 | 3.3:1 | 61.2, 38–84, 13.2 | 2 (15.4) | 2 (15.4) | 6 (46.2) |
| Cancer of other histological types | 15 | 2.0:1 | 64.3, 21–86, 15.9 | 3 (20.0) | 3 (20.0) | 6 (40.0) |
| Non-malignant subjects from the head and neck region | 283 | 1.1:1 | 57.0, 20–91, 14.4 | 30 (10.6) | 16 (5.7) | 89 (31.4) |
| Oral cavity | 139 | 4.0:1 | 59.7, 21–91, 13.9 | 16 (11.5) | 3 (2.2) | 35 (25.2) |
| Oropharynx | 42 | 1.5:1 | 50.4, 24–81, 15.5 | 2 (4.8) | 4 (9.5) | 17 (40.5) |
| Larynx | 34 | 5.8:1 | 54.6, 28–76, 13.7 | 6 (17.6) | 4 (11.8) | 18 (52.9) |
| Lip and paranasal sinus | 20 | 0.7:1 | 62.4, 46–84, 10.1 | 3 (15.0) | 3 (15.0) | 5 (25.0) |
| Salivary gland | 48 | 1.2:1 | 54.8, 20–81, 14.7 | 3 (6.3) | 2 (4.2) | 14 (29.2) |
HPV Prevalence and Characteristics
A subgroup of 61 patients had both fresh tissue and formalin-fixed paraffin-embedded tissue available. These 61 pairs of samples were tested in parallel to compare the yield for HPV by PCR. As a result of identical findings obtained from both tissue samples from the same patient, either type of tissue specimen was considered applicable to this study.
HPV DNA (all types combined) was found in 56 of 539 (10.4%) samples, with the majority (42/56, 75.0%) being HPV16, followed by HPV6 (7/56, 12.5%), HPV11 (3/56, 5.4%), HPV31, and HPV52 (each 2/56, 3.6%). Only 16 (38.1%) of the 42 HPV16-positive samples showed integration based on the comparison between gene copies of E2 and E7. The two HPV31-positive and two HPV52-postive samples were negative for integration. Furthermore, the majority (12/16, 75.0%) of the samples showing HPV16 integration were positive for HPV E6*I mRNA (Table II).
| Disease status | No. of case | HPV DNA (any type) positive, N (%) | HPV DNA (high-risk type) positivea, N (%) | HPV16 DNA positive, N (%) | HPV integration positiveb, N (%) | HPV E6*I mRNA positiveb, N (%) |
|---|---|---|---|---|---|---|
| Head and neck cancer–overall | 256 | 27 (10.5) | 22 (8.6) | 20 (7.8) | 14 (5.5) | 11 (4.3) |
| Control-1 (all non-malignant subjects) | 283 | 29 (10.2) | 24 (8.5) | 22 (7.8) | 2 (0.7) | 1 (0.4) |
| Head and neck squamous cell carcinoma (HNSCC) | 228 | 22 (9.6) | 17 (7.5) | 16 (7.0) | 12 (5.3) | 10 (4.4) |
| Control-2 (all non-malignant subjects excluding salivary gland) | 235 | 26 (11.1) | 21 (8.9) | 19 (8.1) | 2 (0.9) | 1 (0.4) |
| Oral cavity squamous cell carcinoma (OCSCC) | 137 | 3 (2.2) | 1 (0.7) | 1 (0.7) | 1 (0.7) | 0 (0.0) |
| Control-3 (non-malignant subjects from oral cavity) | 139 | 16 (11.5) | 15 (10.8) | 13 (9.4) | 1 (0.7) | 1 (0.7) |
| Oropharyngeal squamous cell carcinoma (OPSCC) | 34 | 10 (29.4) | 10 (29.4) | 10 (29.4) | 10 (29.4) | 9 (26.5) |
| Control-4 (non-malignant subjects from oropharynx) | 42 | 2 (4.8) | 2 (4.8) | 2 (4.8) | 0 (0.0) | 0 (0.0) |
| Laryngeal squamous cell carcinoma (LSCC) | 31 | 5 (16.1) | 3 (9.7) | 3 (9.7) | 1 (3.2) | 1 (3.2) |
| Control-5 (non-malignant subjects from larynx) | 34 | 5 (14.7) | 1 (2.9) | 1 (2.9) | 0 (0.0) | 0 (0.0) |
| Overallc | 539 | 56 (10.4) | 46 (8.5) | 42 (7.8) | 16 (3.0) | 12 (2.2) |
- a High-risk HPV types found were HPV16, 31, and 52.
- b HPV E6*I mRNA detection was only performed for high-risk HPV types. All samples positive for HPV E6*I mRNA were positive for integration.
- c Head and neck cancer-overall + Control-1.
Oropharyngeal SCC showed the highest positive rate for high-risk HPV DNA and were the only cancer group with a significantly higher positive rate compared to the corresponding control group (10/34, 29.4% vs. 2/42, 4.8%, P = 0.003) (Table II). All HPV-positive oropharyngeal SCC samples were HPV16 and showed integration, and the majority (9/10) also had HPV16 E6*I mRNA detected.
A higher proportion of laryngeal SCC was positive for high-risk HPV compared to the corresponding control group, although the difference did not reach statistical significance (3/31, 9.7% vs. 1/34, 2.9%, P = 0.272) (Table II). Furthermore, one of the five high-risk HPV identified from laryngeal SCC specimens was positive for viral integration and E6*I mRNA expression.
Among the control groups, the overall positive rate for HPV DNA was 29/283, 10.2%, and was as high as 5/34, 14.7% in laryngeal samples (Table II). Of note, one specimen from the oral cavity control group was positive for HPV16 integration and E6*I mRNA.
Smoking History
Overall, a self-reported history of smoking was significantly more common among subjects with malignant lesions than those without (124/256, 48.4% vs. 89/283, 31.4%, P < 0.001) (Table I). Subjects with SCC of the oropharynx, larynx, or hypopharynx appeared to be associated with a higher prevalence of smoking than those with SCC of the oral cavity, lip, paranasal sinuses, salivary gland malignant tumors, and cancers of other histological types (67.6–80.6% vs. 20.0–46.2%) (Table I). When compared with the corresponding control group, subjects with oropharyngeal and laryngeal SCC had a significantly higher prevalence of a self-reported history of smoking (P = 0.018 for both).
The prevalence of smoking was significantly higher among HNSCC patients who were infected with HPV compared to those without infection (19/22, 86.4% vs. 93/206, 45.1%, P < 0.001). The prevalence of smoking was also higher among patients with oropharyngeal SCC who had HPV infection (9/10, 90.0% vs. 14/24, 58.3%), but probably because of small sample size, the difference did not reached statistical significance (P = 0.072).
TP53 Mutations
Altogether, 40 different mutations of TP53 were found in this study, of which nine have been reported as hot-spot mutations in previous studies (Table III) [Petitjean et al., 2007; Lindenbergh-van der Plas et al., 2011]. The majority (36/40, 90.0%) of the mutations were non-synonymous which resulted in amino acid substitution, and 5 of 40 (12.5%) resulted in stop codon. The location, amino acid changes and predicted functional implications are shown in Table III [Petitjean et al., 2007; Zanaruddin et al., 2013]. The most frequently found mutation, detected in nine subjects, was located at exon four resulting in a non-synonymous change of codon 37 from Serine to Threonine. Eight subjects had a mutation at codon 249 (arginine-to-methionine) and six subjects had a mutation at codon 176 (cysteine-to-phenylalanine). The majority (66/68, 97.1%) of subjects with TP53 mutations had a mutation at only one spot, while two subjects had mutations at two spots. Both of them were ex-smokers, one had laryngeal SCC and the other had oral cavity SCC.
| Frequency of detection | Exon/codon | Base substitution | Amino acid change | Transition/transversion | CpG island | Hotspot | Within DNA-binding Domain | Within L2/L3/LSH | Dis-ruptive | Trun-cating |
|---|---|---|---|---|---|---|---|---|---|---|
| 9 | 4/37 | TCC to ACC | S37T | Transversion | No | No | No | No | No | No |
| 1 | 4/68 | GAG to TAG | E68 Stop codon | Transversion | No | No | No | No | Yes | Yes |
| 1 | 4/104 | CGT to CTT | R104L | Transversion | No | No | Yes | No | No | No |
| 1 | 4/124 | TGC to GGC | C124G | Transition | No | No | Yes | Yes | No | No |
| 1 | 4/125 | ACG to ACT | no change | Transversion | Yes | No | Yes | Yes | No | No |
| 1 | 5/131 | AAG to AGG | K131R | Transition | No | No | Yes | Yes | No | No |
| 2 | 5/138 | GCC to GTC | A138V | Transversion | No | No | Yes | No | No | No |
| 1 | 5/153 | CCC to CCT | no change | Transition | Yes | No | Yes | No | No | No |
| 1 | 5/156 | CGC to CGT | no change | Transition | Yes | No | Yes | No | No | No |
| 1 | 5/157 | GTC to TTC | V157F | Transversion | Yes | No | Yes | No | No | No |
| 1 | 5/158 | CGC to CCC | R158P | Transversion | Yes | No | Yes | No | No | No |
| 1 | 5/173 | GTG to GGG | V173G | Transversion | No | No | Yes | Yes | Yes | No |
| 2 | 5/175 | CGC to CAC | R175H | Transition | Yes | Head and neck cancer hotspot | Yes | Yes | No | No |
| 6 | 5/176 | TGC to TTC | C176F | Transversion | No | Head and neck cancer hotspot | Yes | Yes | Yes | No |
| 2 | 5/179 | CAT to CGT | H179R | Transition | No | No | Yes | Yes | No | No |
| 1 | 6/192 | CAG to TAG | Q192 Stop codon | Transition | No | No | Yes | Yes | Yes | Yes |
| 1 | 6/193 | CAT to TAT | H193Y | Transversion | No | No | Yes | Yes | Yes | No |
| 1 | 6/196 | CGA to TGA | R196 Stop codon | Transversion | Yes | No | Yes | No | Yes | Yes |
| 1 | 6/199 | GGA to TGA | G199 Stop codon | Transversion | No | No | Yes | No | Yes | Yes |
| 1 | 6/203 | GTG to TTG | V203L | Transversion | No | No | Yes | No | No | No |
| 1 | 6/205 | TAT to TGT | Y205C | Transition | No | No | Yes | No | No | No |
| 1 | 6/215 | AGT to ACT | S215T | Transversion | No | No | Yes | No | No | No |
| 1 | 6/216 | GTG to ATG | V216M | Transition | No | No | Yes | No | No | No |
| 3 | 6/220 | TAT to TGT | Y220C | Transition | No | Head and neck cancer hotspot | Yes | No | No | No |
| 1 | 7/237 | ATG to AAG | M237K | Transversion | No | No | Yes | Yes | Yes | No |
| 2 | 7/245 | GGC to TGC | G245C | Transversion | Yes | General hotspot | Yes | Yes | No | No |
| 2 | 7/248 | CGG to TGG | R248W | Transversion | Yes | General hotspot | Yes | Yes | Yes | No |
| 8 | 7/249 | AGG to ATG | R249M | Transversion | No | General hotspot | Yes | Yes | Yes | No |
| 1 | 7/249 | AGG to ATG | R249M | Transversion | No | General hotspot | Yes | Yes | Yes | No |
| 1 | 7/251 | ATC to AAC | I251N | Transversion | No | No | Yes | No | No | No |
| 1 | 7/259 | GAC to TAC | D259Y | Transversion | No | No | Yes | No | No | No |
| 1 | 8/266 | GGA to GTA | G266V | Transversion | No | No | Yes | No | No | No |
| 1 | 8/267 | CGG to TGG | R267W | Transversion | Yes | No | Yes | No | No | No |
| 3 | 8/272 | GTG to TTG | V272L | Transversion | No | No | Yes | Yes | No | No |
| 2 | 8/273 | CGT to CCT | R273P | Transversion | Yes | General hotspot | Yes | Yes | No | No |
| 1 | 8/280 | AGA to ACA | R280T | Transversion | No | No | Yes | Yes | No | No |
| 1 | 8/281 | GAC to TAC | D281Y | Transversion | No | No | Yes | Yes | No | No |
| 1 | 8/282 | CGG to TGG | R282W | Transversion | Yes | Head and neck cancer hotspot | Yes | Yes | No | No |
| 2 | 8/306 | CGA to TGA | R306 Stop codon | Transition | Yes | No | No | No | Yes | Yes |
| 1 | 9/319 | AAG to AAA | no change | Transition | No | No | No | No | No | No |
A significantly higher prevalence of TP53 mutation was found in subjects with any head and neck cancer (52/256, 20.3% vs. 16/283, 5.7%, P < 0.001), HNSCC (47/228, 20.6% vs. 14/235, 6.0%, P < 0.001), and oral cavity SCC (26/137, 19.0% vs. 3/139, 2.2%, P < 0.001) when compared to the corresponding control groups (Table IV). Although it did not reach statistical significance, TP53 mutation was more common among oropharyngeal SCC than the corresponding control group (8/34, 23.5% vs. 4/42, 9.5%, P = 0.177). Transversion mutations were about as twice as common as transition mutations in all head and neck cancers (35/52, 67.3% vs. 17/52, 32.7%), HNSCC (32/47, 68.1% vs. 15/47, 31.9%), and oral cavity SCC (16/26, 61.5% vs. 10/26, 38.5%), while there were too few cases in the other categories to make a comparison with. Furthermore, 77.3% of transversion mutations occurred in non-CpG sites, and G:C > T:A was the most frequent pattern of substitution observed across all subject groups.
| N (% among subjects with any TP53 mutation) | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Disease status | No. of cases | Any mutation,N (%) | P-valuea | Any non-synonyous mutation, N (%) | P-valuea | Any trans-version | Any tran-sition | Non-CpG trans-version | A:T >T:A | A:T >C:G | A:T >G:C | G:C >A:T | G:C >C:G | G:C >T:A |
| Head and neck cancer-overall | 256 | 52 (20.3) | <0.001 | 43 (16.8) | <0.001 | 35 (67.3) | 17 (32.7) | 25 (48.1) | 7 (13.5) | 1 (1.9) | 6 (11.5) | 11 (21.2) | 4 (7.7) | 23 (44.2) |
| Control-1 (all non-malignant subjects) | 283 | 16 (5.7) | 15 (5.3) | 9 (56.3) | 7 (43.8) | 9 (56.3) | 4 (25.0) | 0 (0.0) | 2 (12.5) | 5 (31.3) | 0 (0.0) | 5 (31.3) | ||
| Head and neck squamous cell carcinoma | 228 | 47 (20.6) | <0.001 | 40 (17.5) | <0.001 | 32 (68.1) | 15 (31.9) | 22 (46.8) | 5 (10.6) | 1 (2.1) | 6 (12.8) | 9 (19.1) | 4 (8.5) | 22 (46.8) |
| Control-2 (all non-malignant subjects excluding salivary gland) | 235 | 14 (6.0) | 13 (5.5) | 8 (57.1) | 6 (42.9) | 8 (57.1) | 3 (21.4) | 0 (0.0) | 1 (7.1) | 5 (35.7) | 0 (0.0) | 5 (35.7) | ||
| Oral cavity squamous cell carcinoma | 137 | 26 (19.0) | <0.001 | 22 (16.1) | 0.001 | 16 (61.5) | 10 (38.5) | 12 (46.2) | 1 (3.8) | 1 (3.8) | 3 (11.5) | 7 (26.9) | 2 (7.7) | 12 (46.2) |
| Control-3 (non-malignant subjects from oral cavity) | 139 | 3 (2.2) | 3 (2.2) | 1 (33.3) | 2 (66.7) | 1 (33.3) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 2 (66.7) | 0 (0.0) | 0 (0.0) | ||
| Oropharyngeal squamous cell carcinoma | 34 | 8 (23.5) | 0.177 | 7 (20.6) | 0.204 | 4 (50.0) | 4 (50.0) | 3 (37.5) | 1 (12.5) | 0 (0.0) | 3 (37.5) | 1 (12.5) | 0 (0.0) | 3 (37.5) |
| Control-4 (non-malignant subjects from oropharynx) | 42 | 4 (9.5) | 4 (9.5) | 3 (75.0) | 1 (25.0) | 3 (75.0) | 1 (25.0) | 0 (0.0) | 0 (0.0) | 1 (25.0) | 0 (0.0) | 2 (50.0) | ||
| Laryngeal squamous cell carcinoma | 31 | 7 (22.6) | 0.406 | 5 (16.1) | 0.463 | 6 (85.7) | 1 (14.3) | 3 (42.9) | 1 (14.3) | 0 (0.0) | 0 (0.0) | 1 (14.3) | 1 (14.3) | 4 (57.1) |
| Control-5 (non-malignant subjects from larynx) | 34 | 4 (11.8) | 3 (8.8) | 3 (75.0) | 1 (25.0) | 3 (75.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 1 (25.0) | 0 (0.0) | 3 (75.0) | ||
| Overallb | 539 | 68 (12.6) | 58 (10.8) | 44 (64.7) | 24 (35.3) | 34 (50.0) | 11 (16.2) | 1 (1.5) | 8 (11.8) | 16 (23.5) | 4 (5.9) | 28 (41.2) | ||
- a By χ-squared test or Fisher's exact test.
- b Head and neck cancer-overall+Control-1.
The proportion of patients with TP53 mutation was similar among HNSCC cases with or without HPV infection (5/22, 22.7% vs. 42/206, 20.4%, P = 0.070). Similar observation was made for OPSCC patients (TP53 mutation in HPV-positive vs. HPV-negative: 2/10, 20.0% vs. 6/24, 25.0%, P = 0.754).
Sexual and Drinking History
Altogether, 6.7% of subjects reported a history of sexually transmitted diseases (Table V). When compared to the corresponding control group, the history of sexually transmitted diseases was significantly more common among any head and neck cancers (25/256, 9.8% vs. 11/283, 3.9%, P = 0.006), HNSCC (24/228, 10.5% vs. 11/235, 4.7%, P = 0.017), and oropharyngeal SCC (7/34, 20.6% vs. 0/42, 0.0%, P = 0.002) (Table V). Only 15 out of 539 (2.8%) subjects reported a history of HPV disease, mainly (12/15, 80%) genital warts, and no significant differences between subject groups were observed. Patients with oropharyngeal SCC were less likely to report having had just a single sexual partner in their lifetime (12/34, 35.3% vs. 28/42, 66.7%, P = 0.006), whereas oral sex was less commonly reported by patients with laryngeal SCC (4/31, 12.9% vs. 12/34, 35.3%, P = 0.046) compared to their respective control group (Table V).
| Self-reported history, N (%) | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Disease status, N | Sexually transmitted disease | P-value | HPV disease | P-value | Never had sexual partner | P-value | Lifetime sexual partner = 1 | P-value | Lifetime sexual partner ≥ 11 | P-value | Ever had oral sex | P-value | Regular drinkera | P-value | Heavy drinkerb | P-value |
| H&N cancer–overall, 256 | 25 (9.8) | 0.006 | 7 (2.7) | 0.948 | 13 (5.1) | 0.523 | 144 (56.3) | 0.287 | 15 (5.9) | 0.390 | 54 (21.1) | 0.743 | 55 (21.5) | <0.001 | 11 (4.3) | 0.018 |
| Control-1, 283 | 11 (3.9) | 8 (2.8) | 18 (6.4) | 172 (60.8) | 12 (4.2) | 63 (22.3) | 26 (9.2) | 3 (1.1) | ||||||||
| HNSCC, 228 | 24 (10.5) | 0.017 | 7 (3.1) | 0.839 | 12 (5.3) | 0.215 | 128 (56.1) | 0.060 | 15 (6.6) | 0.269 | 48 (21.1) | 0.779 | 52 (22.8) | <0.001 | 10 (4.4) | 0.043 |
| Control-2, 235 | 11 (4.7) | 8 (3.4) | 7 (3.0) | 152 (64.7) | 10 (4.3) | 52 (22.1) | 22 (9.4) | 3 (1.3) | ||||||||
| OCSCC, 137 | 10 (7.3) | 0.434 | 4 (2.9) | 0.722 | 7 (5.1) | 0.538 | 91 (66.4) | 0.867 | 7 (5.1) | 0.538 | 28 (20.4) | 0.852 | 25 (18.2) | 0.019 | 4 (2.9) | 0.059 |
| Control-3, 139 | 7 (5.0) | 3 (2.2) | 5 (3.6) | 91 (65.5) | 5 (3.6) | 28 (20.1) | 12 (8.6) | 0 (0.0) | ||||||||
| OPSCC, 34 | 7 (20.6) | 0.002 | 0 (0.0) | 0.499 | 1 (2.9) | 0.216 | 12 (35.3) | 0.006 | 1 (2.9) | 0.447 | 12 (35.3) | 0.062 | 7 (20.6) | 0.302 | 4 (11.8) | 0.167 |
| Control-4, 42 | 0 (0.0) | 2 (4.8) | 5 (11.9) | 28 (66.7) | 0 (0.0) | 7 (16.7) | 5 (11.9) | 1 (2.4) | ||||||||
| LSCC, 31 | 4 (12.9) | 0.701 | 2 (6.5) | 1.000 | 1 (3.2) | 1.000 | 11 (35.5) | 0.818 | 5 (16.1) | 0.726 | 4 (12.9) | 0.046 | 8 (25.8) | 0.076 | 1 (3.2) | 1.000 |
| Control-5, 34 | 3 (8.8) | 2 (5.9) | 1 (2.9) | 13 (38.2) | 4 (11.8) | 12 (35.3) | 16 (47.1) | 1 (2.9) | ||||||||
| Overallc, 539 | 36 (6.7) | 15 (2.8) | 31 (13.0) | 316 (58.6) | 27 (5.0) | 117 (21.7) | 81 (15.0) | 14 (2.6) | ||||||||
- H&N, head and neck; HNSCC, head and neck squamous cell carcinoma; OCSCC, oral cavity squamous cell carcinoma; OPSCC, oropharyngeal squamous cell carcinoma; LSCC, laryngeal squamous cell carcinoma.
- Control-1, all non-malignant subjects; Control-2, all non-malignant subjects excluding salivary gland; Control-3, non-malignant subjects from oral cavity; Control-4, non-malignant subjects from oropharynx; Control-5, non-malignant subjects from larynx.
- P-value by χ-squared test or Fisher's exact test as appropriate, statistical significant results (P ≤ 0.05) are bolded.
- aSelf-reported to have regular drinking habit regardless of amount of consumption.
- bRegularly drunk ≥5 glasses of beer or wine, or ≥3 glasses of cocktail per day.
- c Head and neck cancer-overall+Control-1.
HNSCC patients with HPV infection were more likely to report a history of sexually transmitted disease than those without HPV infection (5/22, 22.7% vs. 19/206, 9.2%, P = 0.05). Whereas, for the OPSCC group, HPV-positive patients were more likely to report a history of sexually transmitted disease (1/10, 10.0% vs. 6/24, 25.0%), though the difference was not statistically significant (P = 0.324). OPSCC patients infected with HPV were less likely to report having single partner in their lifetime compared to those not infected with HPV (2/12, 16.7% vs. 10/24, 41.7%), though the difference was not statistically significant (P = 0.134).
Patients with any head and neck cancer, HNSCC, and oral cavity SCC were more likely to report a history of regular alcohol drinking than the corresponding control group (Table V). Similar associations were obtained when the analysis focused on heavy drinking (≥5 glasses of beer/wine, >3 glasses of cocktail per day).
HNSCC patients infected with HPV were more likely to report a history of regular drinking than those who were not infected (15/22, 68.2% vs. 45/206, 21.8%, P < 0.001).
Multivariate Analysis
Variables showing significant or close to significant associations on univariate analysis for HNSCC were included in a multivariate analysis. Older age, TP53 mutation, and HPV16 infection with oncogenic phenotype (integration and E6/7 mRNA expression) were independently associated factors for HNSCC with an odds ratio of 1.03 (1.02–1.05) for older age, 3.38 for a TP53 mutation (1.71–6.66), and 9.19 (1.13–74.68) for HPV16 infection, respectively. The number of cases for each subgroup of cancer was too few to allow a robust multivariate analysis.
DISCUSSION
About 15–20% of human cancers are associated with viral infections [zur Hausen, 1991; Parkin, 2006]. Recognizing the etiological role of viral infections in carcinogenesis provides a foundation for the development of a screening assay and a vaccination program. These important implications on long term control of virus-associated cancers have been successfully demonstrated in hepatocellular carcinoma and cervical cancer. Since the availability of two highly effective prophylactic vaccines for HPV infection [FUTURE II Study Group, 2007; Paavonen et al., 2009], a tremendous effort has been focused on searching for other cancers that may be associated with HPV [Fei et al., 2006; Goto et al., 2011; Wang et al., 2012]. While conclusions for lung, esophageal and breast cancers are still uncertain or cannot be made at the present moment, the role of HPV in the development of oropharyngeal SCC is clear and widely accepted [Leemans et al., 2011; Rautava et al., 2012; Rampias et al., 2014].
Although HPV is known to have an etiological role in oropharyngeal SCC, especially those of the palatine and lingual tonsils, the proportion of these tumors attributed to HPV varies widely in different parts of the world or is unknown [Evans et al., 2013; Hama et al., 2014; Ndiaye et al., 2014; Nomura et al., 2014; No et al., 2015], and this holds true for the ethnic Southern Chinese for whom this information is limited. High-risk HPV infections of other mucosal areas of the head and neck region have been reported [Syrjänen et al., 2011; Kreimer et al., 2013]. However, an etiological association with carcinoma developing in sites other than the oropharynx has not been fully established [Blioumi et al., 2014; Krüger et al., 2014]. In this study, a large group of subjects with suspected head and neck lesions were examined to determine the association of HPV with head and neck cancers in Hong Kong.
The current study shows that HPV infection of the head and neck mucosa is not uncommon. About 10% of our study samples were found to be HPV-positive, and more than 80% of them were of high-risk types. Of note, less than one-third of these high-risk HPV types demonstrated evidence of viral integration or oncogene E6/7 mRNA expression. Although our assessment criteria using viral integration and E6/7 mRNA expression may be slightly lower in sensitivity, we did find that the majority (90%) of HPV found in oropharyngeal SCC cases fulfilled our criteria, whereas only a very small proportion of cancers from other sites did so.
By taking either E6/7 mRNA expression or viral integration as an indication of an oncogenic phenotype of high-risk HPV infection, we estimate that close to 30% of oropharyngeal SCC in Hong Kong could be associated with HPV. This attributed fraction is lower than those reported from Japan (50%) [Hama et al., 2014] and Beijing (40%) [Wei et al., 2012], but is close to that from Korea (32%) [No et al., 2015].
A recent meta-analysis concluded that there was a high prevalence (30%) of HPV16/18 in laryngeal cancer specimens collected from Chinese patients and also with a strong association (odds ratio: 8.07) [Zhang et al., 2014]. In contrast, we found that only 9.7% of patients with laryngeal SCC had HPV16 infection. Furthermore, only one-third of them showed evidence of integration or E6/7 mRNA expression. Since most previous studies did not verify the oncogenic phenotype of HPV found, the role of HPV could have been overestimated. More in-depth studies are needed to scrutinize the role of HPV in the development of laryngeal cancer.
TP53 is a key tumor suppressor gene. Alcohol and smoking have been recognized as risk factors for TP53 mutation which then predisposes to the development of cancer. High-risk HPV types encode E6 protein which disrupts the normal function of p53 and thus the cell escapes the need for TP53 mutation in cancer development. In line with this, most studies found that HPV-associated head and neck cancers have a lower prevalence of TP53 mutation, which is also a key reason for their better prognosis [Allen et al., 2010; Best et al., 2012]. In this study, we found that 20.6% of subjects with HNSCC were positive for TP53 mutation. While our observed mutation rate was lower than those reported from Europe (45%) and America (50%) [Dai et al., 2004; Braakhuis et al., 2014], it is closer to the reported figure in oral SCC in an Asian population (27.7%) [Zanaruddin et al., 2013]. The most common pattern of mutation found in this study was G:C > T:A transversion which is known to be associated with tobacco exposure [Toyooka et al., 2003]. In line with this, we also observed that the prevalence of TP53 mutation among smokers were about twice as high (19.0 vs. 10.4%) as that of non-smokers.
Although the current study represents the largest study on HPV infection and head and neck cancer conducted in Hong Kong to date, the heterogeneous nature of associations across different anatomical sites and a limited number of patients prevented a robust site-specific analysis. The cross-sectional nature of the current study was also unable to answer questions on the outcome of infection, in particular for those subjects showing evidence of high-risk HPV oncogene expression and yet with normal tissue histology of the biopsy samples. Self-reported epidemiological characteristics, especially those on more sensitive topics like sexual habits might not be accurate. Nevertheless, this study has generated certain key findings. Since infection of high-risk HPV was found to be not uncommon in the oral mucosa of our population, there is a need to monitor the incidence of HPV-associated head and neck cancers, of which there is a raising trend in many countries [Näsman et al., 2009; Marur et al., 2010; Chaturvedi et al., 2011]. More in-depth and specifically designed studies should focus on laryngeal cancer, in which a suspicion of association has been raised by this study.
Although our study has identified HPV as a risk factor for a substantial proportion of oropharyngeal SCC in Hong Kong, there are still a number of unanswered questions such as the epidemiological profile and risk factors associated oral infection with high-risk HPV types. Further studies using tools in addition to self-reported questionnaires should be explored.
CONCLUSIONS
In summary, infection with high-risk HPV of the head and neck mucosa is not uncommon in the population in Hong Kong. We showed in this study that high-risk HPV infection and TP53 mutation are independent risk factors for HNSCC. Smoking, which predisposes to TP53 mutation, is an important risk factor that needs to be managed through awareness. The association with high-risk HPV infection was site-specific and mainly confined to the oropharynx where palatine and lingual tonsil carcinoma develop. Our findings suggest that in Hong Kong, about 26–30% of oropharyngeal carcinoma is associated with high-risk HPV infection, mostly HPV16.
