Human papillomavirus-16 DNA methylation patterns support a causal association of the virus with oral squamous cell carcinomas


  • Adriana Balderas-Loaeza,

    1. Facultad de Quimica, Universidad Nacional Autónoma de México, México D.F., México
    2. División de Investigación, Instituto Nacional de Cancerologia-SS, México D.F., México
    Search for more papers by this author
  • Gabriela Anaya-Saavedra,

    1. División de Investigación, Instituto Nacional de Cancerologia-SS, México D.F., México
    2. Universidad Autonoma Metropolitana, México D.F., México
    Search for more papers by this author
  • Velia A. Ramirez-Amador,

    1. Universidad Autonoma Metropolitana, México D.F., México
    Search for more papers by this author
  • Miriam C. Guido-Jimenez,

    1. Departamento de Biología Molecular y Biotecnología, Instituto de Investigaciones Biomédicas, Universidad Nacional Autónoma de México, México, D.F., México
    Search for more papers by this author
  • Mina Kalantari,

    1. Department of Molecular Biology and Biochemistry, University of California Irvine, Irvine, CA
    Search for more papers by this author
  • Itzel E. Calleja-Macias,

    1. Department of Molecular Biology and Biochemistry, University of California Irvine, Irvine, CA
    Search for more papers by this author
  • Hans-Ulrich Bernard,

    Corresponding author
    1. Department of Molecular Biology and Biochemistry, University of California Irvine, Irvine, CA
    • Department of Molecular Biology and Biochemistry, University of California Irvine, Irvine, CA 92697, USA
    Search for more papers by this author
    • Fax: +1-949-824-8551.

  • Alejandro Garcia-Carranca

    1. División de Investigación, Instituto Nacional de Cancerologia-SS, México D.F., México
    2. Departamento de Biología Molecular y Biotecnología, Instituto de Investigaciones Biomédicas, Universidad Nacional Autónoma de México, México, D.F., México
    Search for more papers by this author


Infection with human papillomavirus-16 (HPV-16) is the cause of most anogenital carcinomas. This virus is also detected in about 20% of all head and neck squamous cell carcinomas. While there is strong evidence for a causal etiological role in the case of tonsillar carcinomas, causal association with malignant lesions of the oral cavity is not yet conclusive. Our previous investigations of HPV-16 DNA methylation in anogenital sites have identified hypermethylation of the L1 gene and part of the long control region in many malignant lesions, but rarely in asymptomatic infections and low-grade precancerous lesions. Here, we report hypermethylation of this diagnostically important segment of the viral DNA in 10 out of 12 HPV-16 positive oral carcinomas from Mexican patients. These data indicate epigenetic changes of HPV-16 in oral carcinomas similar to those in anogenital carcinomas, suggesting carcinogenic processes under the influence of HPV-16 in most if not all of these oral malignant lesions. © 2007 Wiley-Liss, Inc.

Infection with human papillomaviruses (HPVs), notably with high-risk HPV types such as HPV-16 and 18, is a necessary step in the etiology of anogenital cancers, specifically carcinoma of the cervix uteri. This notion is based on epidemiological research that identified a high relative risk of carcinogenic progression associated with high-risk HPV infection, DNA diagnosis that confirmed the presence and transcription of HPV genomes in all cervical carcinomas, and the study of HPV oncoproteins that catalyze a variety of molecular mechanisms that convert normal into malignant cells.1, 2, 3, 4, 5, 6 In case of head and neck squamous cell carcinomas (HNSCCs), an HPV dependent etiology is not as consistently confirmed as in anogenital cancers. Approximately 80% of all HNSCCs do not contain HPV genomes, and must therefore originate from HPV independent etiological processes, likely including mutational events triggered by tobacco and alcohol consumption. However, there is extensive evidence that a proportion of all HNSCCs contain DNA of high-risk HPV types.7, 8, 9, 10, 11, 12, 13, 14 Among head and neck sites, squamous cell carcinomas of the tonsils have the strongest statistical support for an HPV dependent etiology.9, 15, 16, 17 There is less strength and consistency for a linkage between HPV infection and carcinogenesis at sites of the oral mucosa such as tongue, palate, floor of mouth and gingiva. Nevertheless, epidemiological studies have established statistical evidence for a causal association of HPVs and malignancies even for these oral sites.9, 16 In addition, analysis of some HPV containing oral carcinomas revealed recombination between HPV genomes and cellular DNA as well as HPV oncogene expression,10, 11, 12, 13, 18, 19, 20 properties that are generally viewed as support of carcinogenic processes under the influence of HPVs. On the basis of these observations one must conclude that infection with high-risk HPVs plays an etiological role in at least a proportion of malignancies of the oral cavity.

The number of mechanistic studies that support the causal association with malignant lesions of the oral cavity is still small, and the type of evidence, for example the distinction between episomal and chromosomally recombined HPV genomes, and the stoichiometric relationship between the number of cells in a biopsy and the amount of detectable HPV DNA and RNA has technical limitations and is often difficult to interpret. We started the project reported here with the goal to supplement and strengthen the mechanistic evidence for the role of HPV-16 in oral carcinomas. Our study follows up on our previous investigation of DNA methylation of HPV-16 and 18 genomes during the regular viral life cycle and carcinogenic progression in anogenital sites.

The term DNA methylation refers to the transfer of a methyl group to cytosine residues, which are stably maintained at CpG dinucleotides (meCpGs). MeCpGs lead to the binding of transcriptional repressors, part of a network of epigenetic regulatory pathways.21, 22, 23 MeCpGs can be detected by a combination of bisulfite modification, PCR amplification, cloning and DNA sequencing. When we targeted this approach in past studies to HPV-16 and 18 genomes in cervical and anal lesions, we observed a complex diversity of methylation patterns of HPV genomes both in the comparison of different clinical samples and even within individual clinical samples.24, 25, 26, 27, 28 Among these different patterns, hypermethylation of the HPV-16 and 18 L1 gene and of the HPV-16 long control region (LCR) was common in carcinomas and high-grade lesions, but rare or absent in asymptomatic infections or low-grade lesions, likely a consequence of genomic recombination and aberrant transcription. Here we report that we observed the same meCpG patterns in HPV-16 DNA in oral carcinomas. Our findings show that epigenetic changes of HPV-16 genomes resemble one another in epithelia as divergent as the cervix, the anus, and different sites of the oral cavity, and strongly suggest an epigenetic alteration of these viral genomes typical of carcinogenic processes.

Material and methods

Clinical specimens

This study was based on 62 consecutive patients with newly diagnosed oral cancer and without any previous treatment that attended the Head and Neck Department of the Instituto Nacional de Cancerología-SS, México, D. F. All patients were diagnosed and treated following standard clinical practice. All patients signed written, informed Institutional Review Board approved consent. No patient was sampled for the purpose of the specific research described here. No patient identifiers were available to all researchers involved in this study. Samples were obtained through punch biopsy (6 mm), and divided into 2 parts: 1 section was fixed for histopathological diagnosis, and the other 1 placed in a Falcon tube containing PreserveCyt solution (Cytyc, Marlborough, USA), and stored at −20°C. DNA was extracted from the samples by standard procedures, and HPV-typing was performed with MY09/11 and GP5+/6+ consensus primers as described.29 Fifteen out of 62 patients (24%) were found to contain HPV-16. Twelve of the 15 samples had sufficient DNA for the experiments reported in this study.

Bisulfite modification

For bisulfite treatment,30 50–1,000 ng of sample DNA supplemented with 1 μg of salmon sperm DNA in a total volume of 18 μl in water were denatured with 2 μl of 3 M NaOH and incubated at 37°C. After denaturation, 278 μl of 4.8 M sodium bisulfite and 2 μl of 100 mM hydroquinone were added with the mixture being incubated in a thermal cycler for 20 cycles of 55°C for 15 min and 95°C for 30 sec. The modified DNA was desalted with the QIAquick PCR purification protocol and desulfonated thereafter by adding 5.5 μl of 3 M NaOH and 5 μg glycogen prior to 15 min incubation at 37°C. The DNA was precipitated with 5.6 μl of 3 M sodium acetate and 150 μl of 100% ethanol, followed by centrifugation. The pellet was washed with 70% ethanol and dissolved in 30–50 μl of TE buffer (10 mM Tris-HCl pH 8, 1 mM EDTA).

Polymerase chain reactions, primers, T/A cloning and DNA sequencing

We report here the detailed analysis of a segment of the HPV-16 genome between the genomic positions 7049 and 115, which includes the 3′ part of the L1 gene and the LCR. Since bisulfite treated DNA is partially degraded, large amplicons cannot be generated, and we dissected this segment into 3 amplicons. The sequences of the 3 primer pairs were designed according to the genomic sequence of HPV-1631 assuming conversion of all cytosine residues into uracils. They are identical to those described in a previous study.26 We amplified part of the L1 gene and the 5′LCR with the primers 16msp3F (position 7049–7078, AAGTAGGATTGAAGGTTAAATTAAAATTTA) and 16msp3r (position 7590–7560, AACAAACAATACAAATCAAAAAAACAAAAA); the HPV-16 enhancer with the primers 16msp4F (position 7465–7493, TATGTTTTTTGGTATAAAATGTGTTTTT) and 16msp7R (position 7732-7703, TAAATTAATTAAAACAAACCAAAAATATAT); and the HPV-16 promoter with the primers 16msp5F (position 7748–7777, TAAGGTTTAAATTTTTAAGGTTAATTAAAT) and 16msp8R (position 115–86, ATCCTAAAACATTACAATTCTCTTTTAATA).

PCR was carried out in a 25 μl volume containing 0.2 mM of each of the 4 dNTPs, 10 pmol of each primer, 2 mM MgCl2 and 1 unit of AmpliTaqGold (Perkin-Elmer). The PCR conditions were 94°C for 1 min, followed by 40 cycles of 94°C for 10 sec, 54°C for 30 sec and 68°C for 1 min with a final extension at 68°C for 7 min. The presence of PCR products was verified by agarose gel electrophoresis, and confirmed amplicons were cloned with the TOPO TA cloning kit for sequencing (Invitrogen). Cloned DNAs were sequenced by Big Dye terminator v3.1 Cycle Sequencing (Applied Biosystems).


HPV-16 genomes in 15 out of 62 oral carcinomas from Mexican patients

Our research was based on DNA extracted from carcinomas of the oral cavity of 62 consecutive patients that required medical treatment at the Instituto Nacional de Cancerología-SS, México. Our experiments were unrelated to the medical services rendered to the patients and were performed with biopsies discarded after completion of all diagnoses and treatments. Twenty-six of the 62 samples contained any HPV type as measured by PCR amplification with MY09/11 and GP5+/6+ primers and direct DNA sequencing. A detailed pathological study of these 62 samples will be published elsewhere (Anaya-Saavedra and Garcia-Carranca, unpubl. observations). Among the 26 samples with HPV DNA, 15 contained HPV-16. Twelve of these 15 DNA preparations contained sufficient DNA for the analysis of DNA methylation. The demographic and clinical characteristics of these 12 patients are reported in Table I. Specifically, it should be noted that samples from 4 different sites were included in this study, 7 from the tongue, 3 from the gingiva, and 1 each from the palate and the floor of the mouth. There was no obvious linkage of the HPV-16 DNA methylation patterns discussed in the following paragraphs with anatomical site, gender, pathological state, smoking or alcohol consumption.

Table I. Demographic and Clinical Characteristics of 12 Patients with HPV-16 Associated Carcinomas of the Oral Cavity1
SampleGenderAgeTobacco consumptionAlcohol consumptionAnatomical site/OSCC localizationClinical stageHistological grade
  • 1

    The classification of the clinical stage and histological grade followed AJCC recommendations.32

C1F52YesNoGingivaIVModerately differentiated
C2M62YesNoTongueIIIWell differentiated
C3M31NoNoPalateIIIWell differentiated
C4M53NoNoTongueIIIWell differentiated
C5F75NoNoGingivaIIIWell differentiated
C6M55YesYesTongueIVModerately differentiated
C7F59NoNoTongueIWell differentiated
C8F56NoNoTongueIVWell differentiated
C9M71YesYesFloor mouthIVModerately differentiated
C10F66YesNoTongueIIIWell differentiated
C11M57NoNoGingivaIWell differentiated
C12F31NoNoTongueIIIModerately differentiated

Design of the HPV-16 DNA methylation analysis

In published studies we had learned that DNA preparations from each individual cervical or anal HPV infection frequently contain mixtures of HPV-16 and 18 genomes with different methylation patterns, i.e. in the same position of different HPV genomes either CpGs or meCpGs.26, 27, 28 To approach a satisfactory analysis of such mixtures of molecules, we have developed a standardized protocol that we applied here to DNA preparations from oral carcinomas. We treat DNA samples with bisulfite, which converts cytosines into uracils but does not affect methyl cytosine. A subsequent PCR amplification converts methyl cytosines into cytosines and uracils into thymines.30 The amplicons are cloned into E. coli plasmids and individual plasmid clones are sequenced. Cytosine residues indicate nucleotides that had been methyl cytosines in vivo. In this standardized protocol we analyze a 913 bp genomic segment between the positions 7079 and 85, which contains part of the L1 gene and the LCR.33 The segment has altogether 19 CpGs, 3 derived from the 3′ end of the L1 gene (7091, 7136 and 7145), 5 from the 5′-segment of the LCR (7270 to 7461), 5 from the transcriptional enhancer (7535 to 7695), one from the replication origin (7862) and 5 from the E6 promoter (31–58). Unfortunately, bisulfite modification introduces nicks in DNA. As a consequence, it is impossible to PCR amplify efficiently genomic segments with sizes exceeding a few hundred base pairs. Because of this limitation, we dissect the 913 bp segment into 3 amplicons with ends at the genome positions 7091 to 7461, 7535 to 7695 and 7862 to 58. As a consequence, while we aim to gather methylation information from 10 HPV-16 DNA molecules from each sample, data for each of these 3 segments represent the methylation status of different, non-contiguous molecules. Figure 1 shows each amplicon as a line of horizontal segments, the three amplicons separated by two white vertical bars. The numbers C1 to C12 identify the 12 carcinomas, and each of the 10 lines of rectangles a cloned PCR amplicon. White rectangles indicate CpGs, black rectangles meCpGs. Altogether, our investigation led to a database of the methylation state of 2244 CpG dinucleotides in 354 amplicons. In contrast to our past studies of HPV infections in anogenital sites, we present here only an analysis of malignant lesions, and not of asymptomatic infections and precursor lesions. The clinical investigations that led to our research identified 10 HPV-16 positive samples from asymptomatic sites with viral DNA concentrations at the threshold of detectability by PCR. Bisulfite treatment reduced 9 of the 10 HPV-16 DNA preparations to levels that did not generate a measurable PCR signal. It is interesting; however, in consideration of the methylation patterns in carcinomas reported below, that we could generate from 1 asymptomatic sample with sufficient DNA fifteen amplicons, which had all 19 CpG residues completely unmethylated (data not included in Fig. 1).

Figure 1.

CpG methylation of HPV-16 genomes derived from carcinomas of the oral cavity. The figure visualizes the methylation of 19 CpG dinucleotides in the HPV-16 genomes of 10 independent clones each derived from 1 of 12 clinical samples, a database of the methylation status of 2,244 CpG dinucleotides. C1–C12: Designation of carcinomas from the patients listed in Table I. White horizontal rectangles represent unmethylated CpGs, black rectangles meCpGs. Each group of 10 horizontal lines represents clones derived from 1 patient. A horizontal group of rectangles does not indicate a contiguous 913 bp amplicon, but is derived from 3 different HPV-16 genomes, as indicated by 2 white vertical bars. These discontinuities were a technical necessity, since the destruction of bisulfite modified DNA did not allow us to generate the analyzed 913 bp fragment in 1 contiguous amplicon. Two grey areas indicate a lack of information, as we could not generate those missing 6 plasmid clones due to a shortage of the respective sample DNA.

Methylation of the L1 gene and the 5′ LCR of HPV-16

The amplicon representing the 5′ part of the target segment includes 8 CpGs, 3 within the L1 gene (position 7091, 7136 and 7145), and 5 in the 5′ part of the LCR. All of these 8 CpGs had been found hypomethylated in asymptomatic infections and precancerous lesions of the cervix, but hypermethylated (2- to 4-fold increased) in cervical carcinomas, probably since L1 is not transcribed in carcinomas.26 In the 12 oral lesions examined here, we observed methylation frequencies of the 3 L1 CpGs of 25, 38 and 39%, approaching the high levels found in cervical lesions (43–54%). While two samples contained many unmethylated molecules (C1 and C8), there was substantial methylation of most molecules in all other 10 samples. The average methylation of all CpGs in this segment was maximal in the samples C6, C8, C11 and C12 (50, 55, 58 and 65%, respectively), exceeding the methylation levels found in most cervical carcinomas.

Methylation of the enhancer and promoter of HPV-16

The enhancer and promoter of at least 1 HPV-16 genome in precancerous lesions and carcinomas have to remain unmethylated to actively transcribe the E6 and E7 oncogenes. This is exemplified in the cell line SiHa, which contains only 1 HPV-16 genome. However, most carcinomas contain many, sometimes hundreds of HPV-16 genomes, exemplified by the cell line CaSki. Most of these genomes are methylated, possibly resulting from tandem arrangements and chromosomal recombination. As a consequence, they are transcriptionally silent, while continued oncogene expression occurs from a single HPV-16 genome.25, 34 As a consequence of this constellation, and at first glance against intuition, HPV-16 enhancer and promoter methylation is in many clinical samples an indicator of carcinogenic processes in the infected cell.

The central amplicon targeted in this study contains 5 CpGs between the genomic positions 7535 and 7695, overlapping with the transcriptional enhancer. In the examined oral carcinomas, this amplicon was found hypermethylated in C7, C8, C11 and C12, and nearly unmethylated in the remaining 8 samples.

The 3′ amplicon includes 6 CpGs between the genomic positions 7862 and 58. The CpG at position 7862 is part of the E2 binding site at the viral replication origin, the other 5 CpG between the genomic positions 31 and 58 of the HPV-16 genome are part of the Sp1 and E2 binding sites of the E6 promoter. Figure 1 shows that promoter sequences are nearly completely methylated in 3 of the 12 cancers (C8, C11 and C12), a pattern that we had only observed in anogenital high-grade lesions and carcinomas. Interestingly, position 7862, which is part of the replication origin, was nearly always unmethylated, a phenomenon that we noted in all previous studies, but which is yet unexplained.26, 28

Overall, we measured a methylation rate of 28.2% for the 5 CpGs in the promoter segment in oral lesions, exceeding the rate found in cervical cancer and identical to that in high-grade anal lesions (18 and 28.2%, respectively), while methylation rates in asymptomatic infections and low-grade anogenital lesions had been 1 to 2 orders of magnitude lower.26, 28


The investigation of DNA methylation of cellular and viral genes is of great importance, as this modification affects the conformation of nucleosomes in a complex network of epigenetic changes that include histone methylation, histone acetylation and histone deacetylation.21, 22, 23 The resulting chromatin changes constitute a major molecular pathway that regulates cellular gene expression during embryogenesis and differentiation. In the biology of most viruses DNA methylation does not appear to play a role, but previous studies from our group24, 25, 26, 27, 28 and others34, 35, 36 have shown that HPV-16 and 18 are among those few viruses that are strongly affected by this mechanism.

Some preliminary studies suggest that DNA methylation plays a role in the normal lifecycle of HPV-16, possibly by silencing transcription in undifferentiated cells and/or by inducing latency.26, 36 This phenomenon, which mostly affects the viral enhancer and occurs at very low levels of HPV DNA concentration, is unfortunately not yet well understood because of the difficulties to study these viruses in cell culture. The research presented here does not attempt to address DNA methylation during the normal viral biology.

In contrast, there is now detailed information about the methylation of HPV DNA during carcinogenesis based on the study of HPV genomes in clinical samples. As HPV-16 and 18 genomes often recombine with cellular DNA during tumor progression, they are linearized and their normal gene sequences become interrupted, which terminates transcription of genes that are now positioned 5′ of the LCR. This leads apparently to methylation of these DNA sequences, since lack of transcription is known to trigger de novo DNA methylation of any DNA.22 Similarly, extensive evidence suggests that the recombination of any DNA with chromosomes triggers methylation, probably since these sequences frequently target heterochromatin, transcriptionally silent parts of the genomes.37 As a consequence, a combination of high HPV DNA concentrations in a clinical sample combined with hypermethylation of L1, the enhancer and promoter has become a useful test to detect recombination, and thereby characterize established malignancies and identify mixed cell populations with subpopulations that progress toward malignancy.26, 27, 28, 35, 36

Here we documented that HPV-16 genomes in 10 out of 12 carcinomas of the oral cavity were altered by DNA methylation in a manner typical of carcinomas and high-grade lesions of the cervix and the anus, suggesting that HPV-16 genomes are chromosomally integrated in oral malignancies and affect the host cells in a similar manner as in anogenital malignancies. It should also be noted that certain patterns, e.g. the nearly complete lack of methylation at position 7862 in all samples, and the selective hypermethylation of promoter sequences in some samples, have become a recurrent theme of HPV-16 methylation patterns in any anatomic site, and are clearly properties of the viral DNA irrespective of the anatomical site of the affected host cell. It has been pointed out by others35 that promoter methylation may favor carcinogenesis, as it prohibits binding of the E2 protein,38 which is a repressor of the E6 promoter, and assures continued oncogene expression even in the subpopulation of lesions that did not lose E2 as a consequence of genomic interruption.39

The data of this study and our previous studies provide strong support that the measurement of DNA methylation is a useful biomarker to characterize malignant lesions. Ten of the 12 oral lesions that we examined showed hypermethylation typical for malignancies. The characteristics of SiHa cells and some anogenital lesions that we examined previously26 show that these methylation patterns can be sufficient but are not necessary for the malignant state, as cancers that contain a single viral genome must keep this in unmethylated form for continued oncogene transcription. Therefore, HPV-16 may have been the etiological cause of all 12 oral carcinomas of this study. The experimental strategy that we selected here documented refined details of the heterogenous patterns HPV-16 DNA populations encountered in oral carcinomas. To develop HPV DNA methylation as a sensitive and robust diagnosis we recently examined methylation specific primers and methylation specific fluorescence probes in real time amplifications, targeting HPV-18 DNA.40 While these two approaches do not reveal all subtleties of methylation patterns, they cut the laboratory work required for qualitative and quantitative analysis of larger sets of clinical samples by 1 to 2 order of magnitude, and will allow high-throughput analyses required for larger medical studies once similar technology is established for HPV-16.


Our research was supported by NIH grant ROI CA-91964 to H.U.B., by a UC-MEXUS-CONACYT to A.G.C. to fund sabbatical research in the H.U.B. lab, by the grant CONACYT (Consejo Nacional de Ciencia y Tecnologia) number 45987 to V.R.A., and by funds from the Chao Family Comprehensive Cancer Center of the University of California Irvine to H.U.B.