Oral Oncology 47 (2011) 698–701

Contents lists available at ScienceDirect

Oral Oncology

journal homepage: www.elsevier .com/locate /ora loncology

Squamous cell carcinoma of the oral cavity rarely harbours oncogenichuman papillomavirus

Victor Lopes a,⇑, Paul Murray b, Hazel Williams b, Ciaran Woodman b, John Watkinson b, Max Robinson c

a Head and Neck Directorate NHS Lothian and University of Edinburgh, UKb School of Cancer Sciences, University of Birmingham, UKc School of Dental Sciences, Newcastle University, UK

a r t i c l e i n f o s u m m a r y

Article history:Received 17 December 2010Received in revised form 21 April 2011Accepted 30 April 2011

Keywords:Oral squamous cell carcinomaHPVQuantitative PCRMolecular diagnosis

1368-8375/$ - see front matter � 2011 Elsevier Ltd. Adoi:10.1016/j.oraloncology.2011.04.022

⇑ Corresponding author. Address: 4th Floor LauristEdinburgh EH3 9HA, UK. Tel./fax: +44 01315363926.

E-mail address: vlopes@staffmail.ed.ac.uk (V. Lope

Although it is now well established that a significant proportion of oropharyngeal squamous cell carcino-mas (SCC) harbour oncogenic human papillomavirus (HPV) sequences, the frequency with which thesesequences are detected in oral SCC (excluding oropharyngeal subsites) is highly variable. In an attemptto establish the true prevalence of HPV-16 and HPV-18 subtypes in oral SCC, we screened 142 consecu-tive cases from a UK cohort using both conventional PCR with consensus primers and type-specific quan-titative PCR (Q-PCR), while at the same time employing a rigorous protocol to avoid samplecontamination. Q-PCR revealed HPV sequences in five cases; two contained HPV-16 alone, two HPV-18alone, and one sample carried both genotypes. However, only two of these cases (both HPV-16-positive)had moderate viral loads (51 and 91 viral copies per 100 cells respectively) and were positive for HPVDNA by conventional PCR. Both cases contained HPV DNA in tumour cells as shown by Q-PCR analysisof micro-dissected tissue and by in situ hybridisation. The remaining three cases had only very low viralloads (between 3 and 7 viral copies per 100 cells), were negative by conventional PCR and lacked HPVDNA in tumour cells. Our data provide strong evidence that oncogenic HPV is uncommon in oral SCCand that routine HPV testing of these tumours cannot be advocated.

� 2011 Elsevier Ltd. All rights reserved.

Introduction Unlike oropharyngeal cancer, there is uncertainty whether high

It is well established that a proportion of head and neck squa-mous cell carcinomas (SCC) harbour oncogenic human papilloma-virus (HPV). There is compelling evidence that high risk HPV,particularly HPV-16, has an important role in the pathogenesis oforopharyngeal SCC (which comprises tumours of pharyngeal ton-sils, posterior one third of tongue and soft palate).1,2 A meta-anal-ysis estimated that around a third of oropharyngeal cancersharbour HPV (35.6% of 969 cases; CI, 32.6–38.7),3 however, morerecent studies indicate that around 65% of oropharyngeal cancersare HPV positive.4,5 This is an important observation because pa-tients with ‘HPV related’ oropharyngeal SCC have a better progno-sis than individuals with ‘conventional’ HPV-negativeoropharyngeal SCC (overall survival meta-HR: 0.72 95% CI: 0.5–1.0).4–6 Recently, the US National Comprehensive Cancer Networkhave recommended HPV testing for oropharyngeal SCC.7 Followingon from this, there have been calls for standardisation of HPV test-ing for head and neck cancers.8,9

ll rights reserved.

on Building, Lauriston Place,

s).

risk HPV has a significant role in the pathogenesis of oral cavitySCC (which comprises tumours of the anterior two thirds of ton-gue, floor of mouth, alveolus, gingiva, buccal and labial mucosaand hard palate). Although, around a quarter of oral SCCs are re-ported to show evidence of HPV infection (23.5% of 2642 cases;CI 21.9–26.3) there is considerable variation between individualstudies (range 4–80%).3 Surprisingly, a meta-analysis showed nodifference in overall survival between patients with HPV positiveand HPV negative SCCs at ‘non-oropharyngeal’ sites.6

The current study investigates the prevalence of high risk HPVin a UK cohort of patients with oral cavity SCC. The study is impor-tant because it analyses a large number of patients: the classifica-tion of the tumour site is accurate (oropharynx cases have beenexcluded): several methods were used to detect HPV DNA andstringent measures have been used to prevent sample contamina-tion ensuring reliability of the data.

Materials and methods

Frozen biopsy specimens

24 consecutive patients with oral cavity SCC treated at the Uni-versity Hospitals Birmingham NHS Trust were recruited for this

V. Lopes et al. / Oral Oncology 47 (2011) 698–701 699

part of the study. Following patient consent, tumour specimenswere harvested at the time of surgical resection and immediatelyfrozen in liquid nitrogen. There were 13 males and 11 females witha mean age at diagnosis of 61 years (range 37–79 years; standarddeviation 11 years).

Formalin fixed paraffin embedded specimens

145 consecutive cases of oral cavity SCC were identified usingthe West Midlands Oral Cancer Registry spanning a three year per-iod. Formalin-fixed paraffin-embedded (FFPE) tumour blocks wereretrieved from the pathology archive at University Hospitals Bir-mingham NHS Trust. There were 84 males and 61 females with amean age at diagnosis of 64 years (range 24–89 years; standarddeviation 14 years). In a small subset of samples (n = 5) laser cap-ture micro-dissection was used to harvest the malignant epithelialcells (Pixcell IIe, Arcturus, USA). The study was approved by theWest Midlands Research Ethics Committee.

Preparation of DNA

Genomic DNA was prepared from 24 frozen specimens and 145FFPE tumour specimens using a standard laboratory DNA extrac-tion kit (QIAamp DNA mini kit, QIAGEN Ltd, UK). Four sampleswere collected from each specimen in separate Eppendorf tubes,each comprising three 20 lm sections. Rigorous efforts were madeto avoid cross-contamination at every stage of the process: follow-ing the preparation of each sample the microtome was cleanedwith xylene and the blade discarded. After the preparation of eachset of 10 specimens an ‘interval control’ comprising an FFPE cellblock of a human cell-line known to contain Epstein-Barr virus(EBV; X50-7 cell line), but not HPV, was used to monitor inadver-tent sample contamination.

Target amplification using polymerase chain reaction

Control PCR reactionsThe presence of intact human genomic DNA was established by

amplifying the housekeeping gene glyceraldehyde 3-phosphatedehydrogenase (GAPDH) using PCR. DNA from interval controlswas amplified for EBV latent membrane protein-1 (LMP-1) geneby PCR.10 PCR products were resolved on a 1% agarose gel contain-ing ethidium bromide and DNA bands visualised with UV light.

PCR for HPV DNA using GP5+/GP6+ primers

HPV DNA was analysed by PCR using the improved general pri-mer set GP5+/6+ which amplifies part of the HPV L1 gene encodingthe HPV major capsid protein, detecting low risk genotypes -6, -11,-40, -42, -43, -44 and high risk genotypes -16, -18, -31,-33, -35, -39,-45, -51, -52, -56, -58, -59, -66, -68 as previously described.11

Quantitative PCR for HPV-16 and HPV-18

Genomic DNA samples were examined for the presence ofHPV-16 or HPV-18 by Q-PCR. Accumulation of HPV specific geno-type sequences was compared with concurrent amplification ofthe b-2-microglobulin gene using a standard mastermix and anABI Prism 7700 sequence detection system (Perkin Elmer, USA).10

DNA from three human cell lines were used as controls; X50-7 cells(HPV negative), SiHa cells (HPV-16 positive) and HeLa cells(HPV-18 positive).

Primers and probe sequences

HPV-16Forward CTT GTC CAG CTG GAC CAT CTA TTT.Reverse TTG CAG ATC ATC AAG AAC ACG TAG A.Probe (FAM) AAT CAT GCA TGG AGA TAC ACC TAC ATT GCA TGA(TAMARA).

HPV-18Forward CTC GTC GGG CTG GTA AAT GTT.Reverse CAA CCG AGC ACG ACA GGA A.Probe (VIC) ATT ATT AAG TAT GCA TGG ACC TAA GGC AAC ATTGCA A.

b-2 Microglobulin geneForward GGA ATT GAT TTG GGA GAG CAT C.Reverse CAG GTC CTG GCT CTA CAA TTT ACT AA.Probe (VIC) AGT GTG ACT GGG CAG ATC ATC CAG CTT C(TAMARA).

High risk HPV in situ hybridisation

In situ hybridisation for high risk HPV types was carried outusing a proprietary kit (Inform HPV III, Ventana Medical SystemsInc, USA). The Inform HPV III family 16 Probe (B) was used to detecta range of high risk genotypes 16, 18, 31, 33, 35, 39, 51, 52, 56, 58and 66. FFPE cell blocks of CaSki cells (HPV-16 positive), HeLa cells(HPV-18 positive) and C-33A cells (HPV negative) were used ascontrols (Ventana Medical Systems Inc, USA).

Results

Control PCR reactions

DNA from all 24 frozen specimens produced GAPDH PCR ampli-cons of appropriate size as did the majority of the FFPE specimens(118 of 145; 81%); negative specimens were excluded from thesubsequent HPV DNA analysis. All ‘interval control’ samples gener-ated appropriate sized amplicons for GAPDH PCR and EBV LMP-1PCR, but were negative for HPV by consensus PCR.

Detection of high risk HPV DNA

All 24 fresh-frozen specimens from patients with oral cancertested negative for HPV-16 and for HPV-18 using both consensusPCR and Q-PCR. Of the 118 patients who provided FFPE tumourblocks with PCR amplifiable DNA, two tested positive using bothconsensus PCR and Q-PCR and three using Q-PCR alone (Fig. 1,Table 1). The two samples which tested positive using both con-sensus PCR and Q-PCR had higher viral copy numbers (51 and 91copies/100 cells) than the three samples which tested positiveusing Q-PCR alone (3–7 copies/100 cells). Only those sampleswhich tested positive using both consensus PCR and Q-PCR werefound to test positive using in situ hybridisation (Fig. 2). Whenthe Q-PCR experiments were repeated, but this time using DNAprepared from laser capture micro-dissected material, only thosesamples which had tested positive in whole tumours using bothconsensus PCR and Q-PCR continued to test positive. Tumour en-riched samples from the remaining three specimens which hadtested positive in whole tumour using Q-PCR alone, were consis-tently negative.

Discussion

This study demonstrates that the prevalence of high risk HPVtypes in oral cavity SCC is very low; less than 2%. We believe that

Figure 1 The reaction products of 20 samples following PCR using Gp5+/Gp6+ consensus primers resolved on a 1% agarose gel containing ethidium bromide and visualisedwith UV light. SiHa cells (HPV-16 positive) were used as a positive control. Two of the samples (Lanes 5 & 16; Cases 50 & 68) produced a distinct 150 base pair band. Theremaining samples did not show any detectable PCR amplicons.

Table 1Five of 142 oral squamous cell carcinomas showed variable positivity for a portfolio of HPV tests; only two cases were consistently positive for HPV-16 infection.

Caseno.

ConsensusPCRa

Tumour & stroma Tumour only (LCM) HR-HPVCISH

HPV-16 QPCR Copy no./100cells

HPV-18 QPCR Copy no./100cells

HPV-16 QPCR Copy no./cell

HPV-18 QPCR Copy no./cell

29 Negative 3 0 0 0 Negative41 Negative 0 6 0 0 Negative50 Positive 51 4 4 0 Positive53 Negative 0 7 0 0 Negative68 Positive 91 0 22 0 Positive

HR-HPV CISH = high risk HPV chromogenic in situ hybridisation. LCM = laser capture microdissection. QPCR = quantitative polymerase chain reaction.a Consensus polymerase chain reaction (PCR) using GP5+/GP6+ primers.

Figure 2 Photomicrographs of oral squamous cell carcinomas stained withhaematoxylin and eosin (H&E) and chromogenic in situ hybridisation for high riskHPV (HPV CISH). Case 68 shows a strong diffuse blue reaction product for high riskHPV. Case 50 shows a punctate blue signal that co-localises with the nuclei ofmalignant cells. Case 53 show no evidence of high risk HPV DNA.

700 V. Lopes et al. / Oral Oncology 47 (2011) 698–701

this is a robust estimate for the following reasons. This studyincludes both fresh-frozen and FFPE samples from a large numberof consecutive patients with oral SCC. The tumours in our cohortwere all located in the oral cavity. Precise assignment of the ana-tomical site is important because ‘contamination’ of the cohortwith oropharyngeal cancers would bias the results towardsincreasing numbers of HPV positive cases. The locations of casesthat contained HPV sequences in the current study were validatedby checking Cancer Registry information against the medicalnotes; the two cases that were consistently HPV positive were lo-cated in the buccal mucosa and the floor of the mouth respectively.All cases were screened for HPV-16 and HPV-18 using both consen-sus PCR and Q-PCR. Laser capture micro-dissection and in situhybridisation were used to confirm the HPV status of malignantkeratinocytes. Stringent measures were taken to prevent samplecontamination and ‘interval controls’ used to monitor theireffectiveness.

Although the meta-analysis by Kreimer et al.3 demonstratedthat around a quarter of oral SCCs show evidence of HPV infection,the majority of studies included in the study employed non-quan-titative PCR methods to detect HPV DNA. The latter, when usedalone, lacks specificity for oncogenic infection.12 In the largest mul-ticentre study yet reported, HPV DNA was only identified in 4% of766 oral cancers using the consensus PCR primers Gp5+/Gp6+.13

Significantly, Braakhuis et al.14 demonstrated that 6 of 106 (6%)oral SCC contained HPV-16 DNA and E6/E7 mRNA transcripts;the ‘gold standard’ test for active oncogenic virus. Q-PCR amplifiesHPV genotype specific sequences in a quantitative manner againsta concurrently amplified endogenous gene. The technique gives anindication of the viral load; viral gene copy number per number ofcells analysed.15–17 Several studies have used Q-PCR techniques toidentify HPV DNA in oral SCC. Ha et al.18 examined a series of 34oral SCC for HPV-16 by Q-PCR and found that only one tumoursample (3%) contained detectable HPV DNA. Koskinen et al.19

showed similar results; oral cavity SCC had undetectable or lowviral load, whereas, tonsil cancers had very high viral loads. Boy

et al.20 examined 59 oral SCC and demonstrated HPV-18 DNA in7 cases (11%). Whilst these data indicate that oral cancers

V. Lopes et al. / Oral Oncology 47 (2011) 698–701 701

infrequently harbour high viral loads, the studies are limited by thesmall number of tumours analysed (pooled samples, n = 106 cases).By contrast, a recent Hungarian study demonstrated that 31 of 65(48%) cases of oral cancer had evidence of HPV using Q-PCR meth-ods.21 The latter study may reflect geographic variation, however,it is possible that a proportion of positive cases represent transientHPV infection not associated with the elaboration of oncoproteinsor sample contamination with extrinsic HPV DNA.

The current study extends the information from the two studiesthat examined large cohorts of patients with oral cancer and re-ported a HPV prevalence of 2–6% using non-quantitative PCRmethods.13,14 Our results concur with several other studies thathave used Q-PCR methods to amplify oncogenic HPV in oral cavitySCC samples and found low prevalence of oncogenic HPV.18–20 Thepooled data from studies that used Q-PCR, combined with the cur-rent data, gives a prevalence rate of 4% (10 of 248 cases).

Recently, Laborde et al.22 analysed a panel of cellular viral re-sponse genes in a series of oral and oropharyngeal SCCs. The panelincluded genes encoding a variety of antimicrobial peptides, mole-cules involved in antigen presentation and cytokines, all of whichare typically elevated in the face of persistent viral infection. Geneexpression was assessed by Q-PCR. Whilst viral response geneswere strongly expressed in HPV related oropharyngeal SCC, therewas no evidence of a ‘viral signature’ in either HPV negative oro-pharyngeal SCC or oral SCC. This study provides further evidencethat viruses are unlikely to have an important role in the aetiologyof oral cancer.

The strong association of high risk HPV with a proportion oforopharyngeal SCCs is thought to reflect the unique microenviron-ment of the specialised mucosa and mucosal associated lymphoidtissue that comprises the pharyngeal tonsils and the lingual tonsilslocated at the base of the tongue. The lymphoepithelial tissue isconsidered to facilitate high risk HPV infection of the basal kerati-nocytes of the reticulated crypt epithelium, which leads to persis-tent infection, stable oncoprotein expression and transformation tocarcinoma; although the precise mechanisms have not yet beenelucidated.2 The absence of specialised lymphoepithelial tissue inthe oral cavity provides an explanation for the lack of HPV infectionat these subsites.

Interestingly, our study shows that consensus PCR for high riskHPV correlated with the Q-PCR results; only cases with the higherviral loads were positive by consensus PCR. Furthermore, thosecases that showed very low copy number by Q-PCR on whole tu-mour DNA (tumour and stroma) were negative when the Q-PCRwas repeated on DNA extracted from laser capture micro-dissectedmalignant keratinocytes. Therefore, the study raises the possibilitythat non-quantitative consensus PCR methods, when subject tostringent quality control measures, are an effective method ofidentifying HPV DNA in FFPE biopsy material. Smeets et al.12 advo-cated a diagnostic algorithm for oncogenic HPV infection that wasbased on screening tumours for p16 expression, a surrogate markerof HPV oncoprotein expression, and consensus PCR using the Gp5+/Gp6+ primer system. In their study, the combination of these twotests was 100% sensitive and 100% specific when compared todetection of E6/E7 mRNA by reverse transcriptase PCR in fresh tis-sue; the ‘gold standard’ test for transcriptionally active oncogenicHPV. By contrast, detection of high risk HPV DNA by in situ hybrid-isation lacks the sensitivity of target amplification based tests andits utility is limited by false negative rates in the region of 15%.12,23

In summary, there is little evidence to suggest that oral cavitySCC is associated with the detection of high risk HPV types androutine testing of oral cancers for HPV cannot be justified. Whilevaccination against oncogenic HPV may offer some protectionagainst oropharyngeal cancers, it is unlikely to influence the

incidence of oral cancers which account for the majority of headand neck cancers.23

Conflict of interest statement

None declared.

References

1. IACR Monographs on the Evaluation of Carcinogenic Risks to Humans. HumanPapillomavirus: IACR Monographs Volume 90. Geneva, Switzerland: WHO Press;2007: 1-636.

2. Marur S, D’Souza G, Westra WH, Forastiere AA. HPV-associated head and neckcancer: a virus-related cancer epidemic. Lancet Oncol 2010;11:781–9.

3. Kreimer AR, Clifford GM, Boyle P, Franceschi S. Human papillomavirus types inhead and neck squamous cell carcinomas worldwide: a systematic review.Cancer Epidemiol Biomarkers Prev 2005;14:467–75.

4. Ang KK, Harris J, Wheeler R, Weber R, Rosenthal DI, Nguyen-Tân PF, et al.Human papillomavirus and survival of patients with oropharyngeal cancer. NEngl J Med 2010;363:24–35.

5. Lewis JS, Thorstad WL, Chernock RD, Haughey BH, Yip JH, Zhang Q, et al. p16Positive oropharyngeal squamous cell carcinoma:an entity with a favorableprognosis regardless of tumor HPV status. Am J Surg Pathol 2010;34:1088–96.

6. Ragin CC, Taioli E. Survival of squamous cell carcinoma of the head and neck inrelation to human papillomavirus infection: review and meta-analysis. Int JCancer 2007;121:1813–20.

7. National Comprehensive Cancer Network Clinical Practice Guidelines inOncology. Head and Neck Cancers. V.1.2009.

8. Braakhuis BJ, Brakenhoff RH, Meijer CJ, Snijders PJ, Leemans CR. Humanpapillomavirus in head and neck cancer: The need for a standardised assay toassess the full clinical importance. Eur J Cancer 2009;45:2935–9.

9. Robinson M, Sloan P, Shaw R. Refining the diagnosis of oropharyngeal cancerusing human papillomavirus testing. Oral Oncol 2010;46:492–6.

10. Junying J, Herrmann K, Davies G, Lissauer D, Bell A, Timms J, et al. Absence ofEpstein-Barr virus DNA in the tumor cells of European hepatocellularcarcinoma. Virology 2003;306:236–43.

11. Jacobs MV, Snijders PJ, van den Brule AJ, Helmerhorst TJ, Meijer CJ, WalboomersJM. A general primer GP5+/GP6+-mediated PCR-enzyme immunoassay methodfor rapid detection of 14 high-risk and 6 low-risk human papillomavirusgenotypes in cervical scrapings. J Clin Microbiol 1997;35:791–5.

12. Smeets SJ, Hesselink AT, Speel EJ, Haesevoets A, Snijders PJ, Pawlita M, et al. Anovel algorithm for reliable detection of human papilomavirus in paraffinembedded head and neck cancer specimen. Int J Cancer 2007;121:2465–72.

13. Herrero R, Castellsagué X, Pawlita M, Lissowska J, Kee F, Balaram P, et al.Human papillomavirus and oral cancer: the International Agency for Researchon Cancer multicenter study. J Natl Cancer Inst 2003;95:1772–83.

14. Braakhuis BJ, Snijders PJ, Keune WJ, Meijer CJ, Ruijter-Schippers HJ, LeemansCR, et al. Genetic patterns in head and neck cancers that contain or lacktranscriptionally active human papillomavirus. J Natl Cancer Inst2004;96:998–1006.

15. Mellin H, Dahlgren L, Munck-Wikland E, Lindholm J, Rabbani H, Kalantari M,et al. Human papillomavirus type 16 is episomal and a high viral load may becorrelated to better prognosis in tonsillar cancer. Int J Cancer 2002;102:152–8.

16. Klussmann JP, Gültekin E, Weissenborn SJ, Wieland U, Dries V, Dienes HP, et al.Expression of p16 protein identifies a distinct entity of tonsillar carcinomasassociated with human papillomavirus. Am J Pathol 2003;162:747–53.

17. Weinberger PM, Yu Z, Haffty BG, Kowalski D, Harigopal M, Brandsma J, et al.Molecular classification identifies a subset of human papillomavirus-associatedoropharyngeal cancers with favourable prognosis. J Clin Oncol 2006;24:736–47.

18. Ha PK, Pai SI, Westra WH, Gillison ML, Tong BC, Sidransky D, et al. Real-timequantitative PCR demonstrates low prevalence of human papillomavirus type16 in premalignant and malignant lesions of the oral cavity. Clin Cancer Res2002;8:1203–9.

19. Koskinen WJ, Chen RW, Leivo I, Mäkitie A, Bäck L, Kontio R, et al. Prevalence andphysical status of human papillomavirus in squamous cell carcinomas of thehead and neck. Int J Cancer 2003;107:401–6.

20. Boy S, Van Rensburg EJ, Engelbrecht S, Dreyer L, van Heerden M, van HeerdenW. HPV detection in primary intra-oral squamous cell carcinomas–commensal,aetiological agent or contamination? J Oral Pathol Med 2006;35:86–90.

21. Szarka K, Tar I, Fehér E, Gáll T, Kis A, Tóth ED, et al. Progressive increase ofhuman papillomavirus carriage rates in potentially malignant and malignantoral disorders with increasing malignant potential. Oral Microbiol Immunol2009;24:314–8.

22. Laborde RR, Novakova V, Olsen KD, Kasperbauer JL, Moore EJ, Smith DI.Expression profiles of viral responsive genes in oral and oropharyngeal cancers.Eur J Cancer 2010;46:1153–8.

23. Shi W, Kato H, Perez-Ordonez B, Pintilie M, Huang S, Hui A, et al. Comparativeprognostic value of HPV16 E6 mRNA compared with in situ hybridization forhuman oropharyngeal squamous carcinoma. J Clin Oncol 2009;27:6213–21.