The human papillomavirus (HPV) has been implicated as an etiological factor in a subset of head and neck squamous cell carcinoma (HNSCC). Because circulating tumor DNA has previously been detected in the sera of patients with advanced HNSCC (stage III or IV), we hypothesized that HPV DNA might be present in the sera of HPV-positive HNSCC patients. Serum DNA extracts from 70 patients with HNSCC were screened for HPV using conventional PCR and a real-time quantitative assay. All samples subjected to conventional PCR were further tested by dot blot hybridization, and positives were confirmed by Southern blotting. Paired tumor DNA from archived tissues was then similarly screened for HPV genomic material (n = 51) or tested by in situ hybridization (n = 19). HPV-16 DNA was detected with L1 primers in 0 of 65 sera and in 15 of 70 (21%) tumors. Conventional PCR with E7 primers and Southern blot hybridization detected HPV-16 DNA in four (6%) sera. Using real-time quantitative PCR, six samples were found to contain various levels of circulating HPV DNA (mean, 12 copies/ml; range,<1–35.) All six serum-positive patients had corresponding tumors positive for E7. Four of these patients with HPV-positive tumors later developed distant metastases, suggesting that HPV DNA in serum may represent occult hematogenous spread of cancer cells in this subset of patients. Although a much larger prospective trial is required, the presence of HPV genomic material in serum DNA of HPV-positive HNSCC patients may serve as a useful marker of early metastatic disease.
HPV2 has been implicated in the etiology of squamous cell carcinoma involving both the anogenital and upper aerodigestive tract. Despite the lack of a complete description of the mechanism of carcinogenesis, mounting epidemiological and molecular evidence strongly supports this important association. Nearly 20 years have passed since the first report linking certain HPV types to benign and malignant head and neck neoplasms (1) , and current data suggest that a sizable minority (approximately 15–25%) of HNSCC is linked to HPV infection (2, 3, 4, 5, 6, 7, 8, 9) . Recent observations suggest that patients with HPV-associated HNSCC may display a clinical course different from those of patients with HNSCC whose etiology is linked to smoking and drinking. Moreover, these HPV-positive patients may be helped by additional therapeutic modalities, including vaccine immunotherapy. Because of these considerations, the detection of HPV in patients with head and neck squamous neoplasia is of clinical importance.
Circulating tumor DNA has been identified in the serum or plasma of cancer patients using a variety of approaches. We previously described the presence of HNSCC tumor DNA in the serum by detecting microsatellite DNA alterations identical to those in the tumors of 6 of 21 patients with advanced HNSCC (10) . Similarly, a variety of molecular techniques (including in situ hybridization, Southern and Northern blotting, PCR, reverse transcription-PCR, and DNA sequence analysis) have been used to detect HPV genomic material in HNSCC tumor tissue (2, 3, 4, 5, 6, 7, 8, 9) . Recent studies have also shown that viral sequences may be detectable in serum. Mutirangura et al. (11) described the presence of EBV DNA in the sera of 13 of 42 patients with NPC. EBV typing between the tumors and the sera showed identical results, suggesting that the serum EBV DNA represented disseminated tumor DNA. Additionally, Lo et al. (12) quantified EBV DNA in the plasma of NPC patients and correlated these findings with NPC stage. In light of these findings, we hypothesized that HPV DNA might be detectable in the serum of HPV-positive HNSCC patients and therefore may represent a novel marker for disseminated disease.
Materials and Methods
Sample Collection and DNA Isolation.
Tumors from 72 HNSCC patients at the Johns Hopkins Medical Institutions were obtained fresh from surgical resection with prior consent, along with corresponding venipuncture blood samples. Clinical information including tumor location, stage, and nodal status was recorded. To the best of our knowledge, all patients were fully immunocompetent. Tumor tissue was frozen and microdissected as described previously (13) . Clotted blood specimens were centrifuged at low speed for 5 min, and the serum was stored at− 80°C before DNA extraction. Serum samples (400 μl) were used for DNA extraction. Serum and tumor tissue samples were digested in SDS and proteinase K at 48°C overnight, followed by phenol/chloroform extraction and ethanol precipitation of DNA. After resuspension in 50μ l of distilled water, the mean working DNA concentrations were 110 ± 50 ng/μl per serum sample, and 2 μl were usually sufficient for robust amplification. Two paired samples were discarded because of poor DNA quality.
L1 degenerate consensus primers MY09/MY11/HMB01, targeting a 450-bp region in the L1 open reading frame of HPVs, were used in single amplification reactions of 50 μl volume. Each reaction contained 5 μl of 10× buffer, 200 μm deoxynucleotide triphosphates, 3 mm MgCl2, 0.1 μm each β-globin primer, 0.5 μm each HPV-16 E7 primer, 2 μl of DNA sample, and 1.25 units of AmpliTaq Gold. All amplifications were performed in a laboratory physically separated from the primary DNA extraction site. The PCR protocol used 40 cycles, with denaturation at 95°C for 20 s, annealing at 55°C for 30 s, and extension at 72°C for 30 s. The last cycle used an additional 5-min extension at 72°C. Amplification was followed by transfer to nylon membranes and hybridization of the PCR products with 33 biotin-labeled HPV probes (32 HPV type-specific probes and 1 HPV generic probe) in a dot blot format. Enhanced chemiluminescence (ECL system; Amersham) was used for signal detection. β-Globin primers that target a 268-bp region of the β-globin gene were included in each reaction at a concentration lower than that of the L1 primers. An oligonucleotide primer specific for β-globin was also used for hybridization of the PCR products. Any specimen negative for both a HPV amplicon and a β-globin amplicon was considered unsatisfactory.
Each dot blot hybridization included the following controls: (a) specimen integrity controls (β-globin amplification); (b) sensitivity controls [SiHa cells (one genomic copy of HPV-16 per cell) and CD4II cells (one genomic copy of HPV-18 per cell) diluted to provide 10 and 100 copies of HPV-16 and HPV-18]; (c) contamination controls [K562 cells (HPV-negative cells) were placed in every twelfth position on the plate]; and (d) hybridization controls (PCR products from previously amplified reference specimens were placed in the top 12 wells after the amplification step prior to filter transfer). Four sera samples lacked amplification with βglobin secondary to the presence of degraded or insufficient DNA.
Because the L1 primers may be associated with a 2–3% false negative rate due to integration events in the L1 open reading frame, E7 HPV type-specific primers were used as well. The E7 gene is thought to be necessary for maintenance of the malignant state and is therefore unlikely to be disrupted by integration events. Recent results from a large trial revealed HPV-16 to be the predominant type found in HNSCC; therefore, E7 HPV-16 primers were designed, targeting a 132-bp region of the E7 HPV-16 gene (14) .
Southern Blot Analysis.
All HPV-positive serum samples were subjected to Southern blot confirmation. PCR amplification of these positive serum samples was repeated with fresh stock DNA and loaded into a 2% agarose gel with a biotinylated HindI digest DNA ladder. Southern transfers were performed as outlined previously and hybridized with internal oligonucleotide HPV probes (15) .
In Situ Hybridization.
Because 19 serum samples did not have adequate fresh frozen paired tumor tissues available, in situ hybridization using catalyzed signal amplification was performed using archived paraffin-embedded tumor tissues. A biotin-labeled wide spectrum HPV probe and HPV-16/18 probe were obtained, along with a negative control biotinylated plasmid probe (DAKO). Control cell blocks were prepared with SiHa, HeLa, and C33A cell lines (American Type Culture Collection, Manassas, VA). The paraffin blocks of the specimens and the control cell lines were sectioned at 5 mm. The sections were placed on Chem-Mate slides (BioTek Solutions) and deparaffinized. Target DNA was retrieved by placing the slides in 95°C 10 mm citrate buffer (pH 6.0) in a steamer for 20 min. After target retrieval, the tissue sections were digested with 4 mg/ml proteinase K in 50 mm Tris-HCl (pH 7.6) at room temperature for 5 min, and the proteinase K was inactivated by washing three times in distilled water with agitation. The digestion procedure was followed by quenching of the endogenous peroxidase with 3% hydrogen peroxide at room temperature for 10 min. After the DNA probe was applied onto the sections, the target DNA and the probe were denatured by heating at 95°C for 5 min on a heat block. The sections were hybridized at 45°C overnight in a moisture chamber. After hybridization, the coverslips were removed by soaking in a TBST solution [50 mm Tris-HCl (pH 7.6), 300 mm NaCl, and 0.1% Tween 20]. Stringent wash was performed by incubating the slides in 0.1× SSC (containing 15 mm NaCl and 1.5 mm sodium citrate) at 55°C for 20 min. The DAKO GenPoint kit was used for the catalyzed signal amplification cycle, which consisted of the application of primary streptavidin-horseradish (1:500 dilution in the diluent) for 15 min and the application of biotinyl-tyramide solution for 15 min. After the amplification cycle, secondary streptavidin-horseradish was applied for 15 min. Five-min TBST washes with agitation were performed three times after the application of each reagent. The signal was developed by adding 3,3′-diaminobenzidine for 5 min, followed by a distilled water wash for another 5 min. Meyer’s hematoxylin was used as a counterstain before mounting.
Real-Time Quantitative HPV DNA PCR.
Serum HPV DNA concentrations were measured using a real-time quantitative PCR system. Real-time PCR reactions were set up in a reaction volume of 25 μl using the TaqMan Universal PCR Master Mix (Perkin-Elmer Corp., Foster City, CA). HPV-16 E7 primers and probes were designed using Primer Express software (Perkin-Elmer Corp.). Fluorogenic probes were custom synthesized by Synthetic Genetics (San Diego, CA). PCR primers were synthesized by Life Technologies, Inc. (Gaithersburg, MD). As an internal positive control, real-time PCR analysis was performed on the β-globin gene in parallel. The β-globin gene primers/probes used were as described previously (16) . DNA amplifications were carried out in a 96-well reaction plate format in a PE Applied Biosystems 7700 Sequence Detector (Perkin-Elmer). Both the HPV and β-globin PCR reactions were carried out in duplicate. Multiple negative water blanks were included in every analysis.
A standard curve was run in parallel with each analysis using DNA extracted from a HPV-positive cell line, CaSki (CRL-1550; American Type Culture Collection). CaSki was previously reported to contain 600 integrated viral genomes/cell. Serial dilutions of CaSki DNA were made. To express the quantitative results in genome equivalents for the unknown samples, a conversion factor of 6.6 pg DNA/diploid cell was used.
Concentrations of circulating cell-free HPV DNA were expressed as copies of HPV genome/ml serum and calculated using the following equation (12) :
in which C represents the target concentration in serum expressed as copies/ml, Q represents the copy number as determined by the sequence detector, Vdna represents the total volume of DNA obtained after DNA extraction (50 μl), Vpcr represents the volume of DNA used for the PCR reaction (5 μl), and Vext represents the volume of serum used to extract the DNA (400 μl).
We tested 70 HNSCC tumor and paired serum samples for the presence of HPV DNA with PCR followed by dot blot hybridization. HPV genomic material using L1 primers was detected in 15 of 51 tumors (14 HPV-16, 1 HPV-33) but in 0 of 67 paired sera (the β-globin control did not amplify in three samples). The remaining 19 tumor samples were negative for HPV by in situ hybridization. We then proceeded to test the samples with E7 type 16-specific primers. All but one of the tumor samples tested with the E7 primers were concordant with the L1 primer data. Combining the PCR results with the in situ hybridization results (n = 70), 9 of 17 (53%) oropharyngeal tumors, 3 of 31 (10%) oral cavity tumors, 2 of 17 (12%) larynx tumors, and 1 of 4 (25%) hypopharyngeal tumors were HPV positive (one tumor was from an unknown site). Of the 13 tumors that were HPV positive by both L1 and E7 PCR assays, 11 (85%) had advanced disease (stage III or IV). Of these, nine had nodal disease, but none had clinically evident distant metastases.
Interestingly, HPV DNA was detected in 11 of 66 (17%) sera using the E7 primers (Table 1⇓ ; the β-globin control did not amplify in four samples). However, only 4 of the 11 seropositive samples by dot blot hybridization were confirmed to be HPV positive by subsequent Southern blot analysis (Fig. 1)⇓ . All four of these samples were from patients with oropharyngeal tumors. Four had nodal disease, but none displayed evidence of metastases at the time of phlebotomy.
In addition, quantitative analysis of HPV DNA was performed on all 70 HNSCC sera. Using real-time PCR, varied levels of circulating HPV DNA were detected in the four samples confirmed to be HPV positive by Southern blot analysis. Moreover, using this assay, two additional sera samples had measurable levels of circulating HPV DNA not previously detected by conventional PCR and Southern blot analysis.
In this study, HPV DNA was detected in the serum of a small subset of patients with HPV-positive HNSCC. Thirteen of 70 patients had HPV-16 DNA verified in their tumor tissues by PCR amplification with both L1 and E7 primers, and 6 of these demonstrated HPV-16 DNA in their sera, as confirmed by Southern blot and/or real-time quantitative PCR. Because the life cycle of HPV occurs entirely within epithelial tissues, the virus is not usually found in the bloodstream. Pao et al. (17) described HPV mRNA in the bloodstream of 12 patients with grossly metastatic stage IVb carcinoma of the uterine cervix. The presence of viral mRNA in the bloodstream was attributed to shed cervical cancer cells. In our current study, however, none of the patients had clinical evidence of distant metastases. This implies that for HPV DNA to be present, the tumors must have been locally aggressive and actively shedding cells or cellular debris into the bloodstream at the time of venipuncture. When cells become necrotic or die because of apoptosis, they may become bereft of their membrane integrity, causing spillage of their cellular contents into the surrounding milieu. Thus, even in the absence of active cell shedding, DNA and proteins can be translocated from the neoplastic cells to a distant site through the blood stream.
Whereas L1 HPV DNA was not detected in the sera of any patients, the E7 type 16-specific primers identified circulating DNA in a specific subset. The progression from premalignant to invasive cancer is associated with the conversion of the viral genome from an episomal form to an integrated form with concomitant deletion of some viral genes. Because L1 is a viral capsid protein and is not required for the progression of viral tumorigenesis, it is possible that integration of HPV results in loss of the L1 template. In contrast, there is selective pressure to maintain E7 expression in this tumor progression model. HPV-16 E7 is able to bind and inactivate the retinoblastoma protein (Rb), thereby enhancing cellular proliferation.
In this study, we have developed an assay to quantitate HPV DNA using real-time PCR. Similar assays for EBV DNA have proven to be quite reliable in detecting and monitoring NPC (12) . The data obtained using the real-time assays were completely concordant with conventional PCR and Southern blot analysis. Moreover, we found that real-time PCR has several advantages over these conventional methods. The real-time systems were more rapid, sensitive, and specific. They are based on the continuous optical monitoring of the progress of a fluorogenic PCR. In addition to the two amplification primers (as used in conventional PCR), a dual labeled fluorogenic hybridization probe (as in conventional Southern blotting) is included. This system thus combines the sensitivity of a PCR reaction with the specificity of a Southern blot. In addition to quantitating the levels of circulating HPV DNA in the four samples confirmed to be HPV positive by Southern blot, we detected and quantitated two additional sera samples missed by the conventional methods.
Previous studies have shown that patients with HPVpositive tumors actually benefit from a better overall disease-specific survival than patients with HPV-negative tumors (14) . The real-time assay detected HPV DNA in the sera of 6 of in 13 patients with HPV-16-positive tumors. Moreover, patients with HPV-positive serum presented with advanced-stage III/IV disease, and four of thee six patients developed distant metastasis. At the time this report was written, three of the six (50%) HNSCC patients with HPV-positive serum have died (two from unrelated causes and one from distant metastases.) Of the seven remaining patients with HPV-positive tumors but no evidence of circulating HPV DNA, six (86%) have no evidence of disease, and one has been lost to follow-up. Despite the presence of cellular debris in the bloodstream, it is possible that metastatic disease may be delayed or prevented in some patients by host immunity. Alternatively, as in the case of protein markers, shed DNA may simply be a tumor marker that has accessed the circulation.
The overall survival of patients with squamous cell carcinoma of the head and neck has not changed significantly in 30 years. One of the difficulties encountered is the high incidence of locoregional recurrence and distant metastasis. A total of 30–50% of patients diagnosed with HNSCC this year will die of locoregional recurrence in 5 years, 20–40% will have clinical metastases, and up to 60% may have occult metastases at autopsy (18) . Because 99% of cervical cancer and approximately 25% of HNSCC is associated with HPV infection, cervical cancer can serve as a model system to understand HPV infection and its potential mechanisms of head and neck carcinogenesis. In addition, novel therapeutic approaches directed toward HPV-associated cervical lesions can easily be applied to HPV-associated HNSCC. Currently, HPV vaccines are in clinical trials for premalignant cervical lesions as well as cervical carcinoma, and such additional therapies may be benefit patients with HPV-associated HNSCC.
In this study, we report the detection of HPV DNA in the sera of patients with HPV-associated HNSCC. In addition, we propose that quantitative real-time PCR is a more sensitive and specific assay to monitor serum cell-free HPV DNA. Because the sera samples studied represent a single time point during the patient’s care, additional longitudinal studies are required to study the relationship between serum HPV DNA and tumor load. It will be interesting to see whether the levels of circulating HPV DNA will predict time to recurrence. If so, a quantitative measure of circulating HPV DNA may become an important tool for the early detection of metastasis and tumor recurrence.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 To whom requests for reprints should be addressed, at Department of Otolaryngology–Head and Neck Surgery, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205.
↵2 The abbreviations used are: HPV, human papillomavirus; HNSCC, head and neck squamous cell carcinoma; NPC, nasopharyngeal carcinoma.
- Received June 16, 2000.
- Revision received August 25, 2000.
- Accepted August 30, 2000.