Abstract
Background
A spectrum of cutaneous human papillomaviruses (HPVs) is detectable in nonmelanoma skin cancers, as well as in healthy skin, but the significance that the presence of these types of HPV DNA has for the pathogenesis of skin cancer remains unclear.
Methods
We studied 349 nonimmunosuppressed patients with skin lesions (82 with squamous cell carcinomas, 126 with basal cell carcinomas, 49 with actinic keratoses, and 92 with benign lesions). After superficial skin had been removed by tape, paired biopsy samples—from the lesion and from healthy skin from the same patient—were tested for HPV DNA. Risk factors for HPV DNA were analyzed in multivariate models.
Results
Overall, 12% of healthy skin samples were positive for HPV DNA, compared with 26% of benign lesions, 22% of actinic keratoses, 18% of basal cell carcinomas, and 26% of squamous cell carcinomas. HPV DNA was associated with sites extensively exposed to the sun, both for the lesions (odds ratio [OR], 4.45 [95% confidence interval {CI}, 2.44–8.11]) and for the healthy skin samples (OR, 3.65 [95% CI 1.79–7.44]). HPV types of Beta-papillomavirus species 2 predominate in squamous cell carcinomas (OR, 4.40 [95% CI, 1.92–10.06]), whereas HPV types of Beta-papillomavirus species 1 are primarily found in benign lesions (OR, 3.47 [95% CI, 1.72–6.99]).
Conclusions
Cutaneous HPV types are primarily detected at sites extensively exposed to the sun. HPV types of Beta-papillomavirus species 2, but not of species 1, are associated with squamous cell carcinoma.
Nonmelanoma skin cancers (NMSCs) are the most prevalent malignancies among white populations [1], with increasing prevalence in many European countries [2]. UV radiation is the major etiological factor in skin cancer [3], and genomes of the cutaneous human papillomaviruses (HPVs) are commonly detected in the tumors [4]. However, cutaneous HPV is also common in healthy human skin [5-12], and the pathological significance of these viruses is unclear.
Among genital HPV types, different genera and HPV types have clearly various carcinogenic potential [13]. Specific cutaneous HPV types have been associated with squamous cell carcinoma (SCC) of the skin of individuals with the rare hereditary disease Epidermodysplasia verruciformis (EV) [14], but such association has not been reported for SCC in the general population.
The cutaneous HPVs are classified into different genera, of which the genus Beta-papillomavirus contains 23 different fully characterized HPV types (previously designated “EV” types), the genus Gamma-papillomavirus contains 7 fully characterized HPV types (HPV 4, 48, 50, 60, 65, 88, and 95), the genus Mu-papillomavirus contains 2 HPV types (HPV 1 and 63), and the genus Nu-papillomavirus contains HPV 41 [15]. The genus Beta-papillomavirus is further divided into 5 distinct species containing related HPV types (Beta-1, containing HPV 5, 8, 12, 14, 19, 20, 21, 24, 25, 36, 47, and 93; Beta-2, containing HPV 9, 15, 17, 22, 23, 37, 38, and 80; Beta-3, containing HPV 49, 75, and 76; Beta-4, containing HPV92; and Beta-5, containing HPV 96) [15].
The possibility exists that viruses shed from infected skin might be trapped on the surface of skin tumors, resulting in HPV-positive tumor samples, although the viral DNA is not present in the tumor itself. Indeed, simple stripping of the skin surface by tape greatly reduces the proportion of tumors that test positive for HPV DNA [16]. If HPV infection is of direct etiological relevance, the virus should presumably be present not only on top of the tumor but also within the tumor. No case-control study using tape-stripped lesions has thus far been reported that would allow assessment of risk factors for HPV DNA inside the lesions. To obtain consistent evidence regarding risk factors for HPV infection in benign and malignant tape-stripped lesions, we performed a hospital-based case-control study in which 3 independent laboratories identified the presence of HPV DNA by cloning and sequencing of polymerase chain reaction (PCR)–generated amplicons.
PATIENTS, MATERIALS, AND METHODS
Patients
The present study was designed as a hospital-based case-control study of NMSCs and premalignant and benign skin lesions. The study base was defined as patients seeking medical care at participating dermatology clinics, where surgical removal of a skin lesion was medically indicated and the patients gave informed consent to participate. In total, 365 patients were enrolled; 16 of them subsequently were excluded on the basis of a follow-up review of the study file. The exclusion criteria were decided by the entire study group and without prior knowledge of any HPV DNA results. Restriction of the study to immunocompetent patients accounted for 3 of the 16 patients excluded, and restriction to nongenital lesions accounted for 1 of the 16 patients excluded; 5 patients were excluded because they had disease that did not meet case definitions (3 had melanomas, 1 had fungal infection, and 1 had 2 different tumors that belonged to different case groups); and 7 patients were excluded because their tumors had inadvertently not been tape-stripped before a biopsy was performed.
The remaining 349 immunocompetent patients had attended dermatology clinics in either Sweden (Stockholm, 148 patients; Gothenburg, 91 patients; and Malmö, 102 patients) or Austria (8 patients). Cases of SCC (n = 82; mean age, 80 years [range, 50-94 years]) and of basal cell carcinoma (BCC) (n = 126; mean age, 75 years [range, 34–93 years]) were confirmed histologically. There were 49 patients with actinic keratoses (AKs) (mean age, 74 years [range, 53–95 years]) and 92 with benign lesions (mean age, 71 years [range, 29–94 years]) consisting of seborrhoeic keratoses (SKs) (n = 48; mean age, 75 years [range, 48–89 years]) and other benign lesions (n = 44; mean age, 70 years [range, 29–94 years]) (benign epidermal dysplasia, 1 case; unspecified benign lesion, 1 case; benign squamous papilloma, 1 case; epidermoid cyst, 1 case; fibroma, 2 cases; inflammatory lesion, 1 case; keratoacanthoma, 7 cases; neurofibroma, 1 case; nevus, 7 cases; no tumor/healthy skin, 3 cases; pilaracanthoma, 1 case; scar tissue, 6 cases; skin tag, 1 case; verruca, 1 case; verruca seborrhoica, 4 cases; verruca vulgaris, 1 case; trichoepithelioma, 1 case; and acrotrichoma, 1 case); in the benign-lesion group we also included 1 case each of cornu cutaneum, squamous atypia, and squamous dysplasia.
The study adhered to the Declaration of Helsinki and was approved by the Ethical Review Committees of Karolinska Institute and of Lund University (Sweden) and of Medical University Vienna (Austria).
Collection of Samples and Data
Each patient donated both a biopsy sample from the lesion and a control biopsy sample from healthy skin, the latter of which typically was collected 10–15 cm from the lesion, at sites such as the neck, the shoulder, the back, or proximally to the ear. Before 2-mm punch-biopsy samples were taken, the skin was anesthetized (without ethanol cleaning) and stripped with tape (article 1527-1, Transpore; 3M Health Care) that was attached 5 times, and a new piece of tape was attached for another 5 times [16]. The remaining tumor was excised and sent for histopathological diagnosis. In most cases, AKs, SKs, and other benign lesions were diagnosed only clinically. Patients were interviewed by use of a standardized questionnaire inquiring about skin type (I–IV, according to the Fitzpatrick classification [17]), history of sunburns, hair color, history of profession, eye color, smoking habits, history of skin cancer, and family history of skin cancer. The level of sun exposure at the site of biopsy was classified, by a single dermatologist (B.L.), as being in 1 of 3 categories based on anatomical site: extensive (i.e., head and neck, dorsal side of the hands), moderate (i.e., trunk and extremities), or low (i.e., buttocks and genital area). For 18% (63/349) of the patients, the paired biopsy samples were taken from sites with different sun-exposure classification.
Analysis of HPV DNA
Adequacy of sample
DNA was extracted from all samples by a simple phenol-free method, in a single laboratory [18]. The adequacy of the sample was tested by PCR amplification of the human β-globin gene [16]. If the result was negative (3.7% [26/698] of the samples), the β-globin PCR was repeated after ethanol precipitation; if the sample remained negative (0.5% [4/698] of the samples), the DNA was column purified (QIAmp Min Elute Media Kit; Qiagene) and retested, resulting in all samples being β-globin positive.
Testing
For maximum reliability and sensitivity of analysis, in the effort to detect as many infections as possible, 3 different experienced laboratories applied their routinely used HPV-detection method. The data were pooled for maximum sensitivity, and all positive samples were verified by cloning and sequencing, for maximum specificity.
In the 3 laboratories, aliquots of each sample were tested for the presence of HPV DNA by use of PCR amplification with degenerate primers described as being able to amplify all known and new putative HPV types. Laboratory 1 used both a single-round PCR with FAP primers [18] and single tube–nested “hanging droplet” FAP PCR [19], whereas laboratory 3 used single tube–nested “hanging droplet” FAP PCR only [19]; laboratory 2 used a single-round FAP-PCR [18], a 2-step version of the nested FAP-PCR [19], a single-round PCR with CP primers, and a nested CP-PCR [20] with modified PCR conditions [21]. PCR amplicons were cloned and sequenced. A least 10 clones per sample were sequenced by laboratory 2, and 3 clones per sample were sequenced by laboratories 1 and 3.
A sample was scored as being positive for HPV DNA if (1) the sequences adequately aligned (i.e., harbored the conserved regions defining HPV) to known HPV sequences in the Gen-Bank database when Blast searches were performed (http://www.ncbi.nlm.nih.gov/BLAST/) and (2) the identical HPV type/putative type was detected in at least 2 of the 3 laboratories; by these criteria, 31% (219/698) of the samples were positive for any HPV type, and 17% (120/698) of them were positive for identical HPV type(s).
An isolate was classified as being a known HPV type if its sequence homology to the known HPV type was >90% and as being a putative type if such similarity was <90%. For categorization of the new putative types into genera and species of the family Papillomaviridae, phylogenetic trees were constructed by use of ClustalW, in BioEdit [22], and Molecular Evolutionary Genetics Analysis (MEGA), version 2.1 [23].
The necessary precautions to avoid contamination between samples were observed during all levels of sample handling, and, in each batch of HPV tests, water samples without DNA were included as negative controls.
Statistical Analysis
Odds ratios (ORs) and 95% confidence intervals (CIs) were estimated by use of multivariate logistic regression models in LogXact (version 6; Cytel Software Corporation). Results for which the 95% CI excluded unity were considered to be significant.
RESULTS
A total of 42 different HPV types/putative types were identified, of which 37 were classified as belonging to the genus Beta-papillomavirus, 3 to the genus Gamma-papillomavirus, and 2 to the genus Alpha-papillomavirus, harboring mainly the genital HPV types (table 1). The most frequently detected types were HPV20 (18 samples) and HPV21 (11 samples) in genus Beta-papillomavirus species 1 and HPV38 (10 samples) in genus Beta-papillomavirus species 2 (table 1). Overall, 12.0% of healthy skin samples were positive for HPV DNA, compared with 26.1% of benign lesions, 22.5% of AKs, 17.5% of BCCs, and 25.6% of SCCs (table 2).
Table 1. Frequency of 42 human papillomavirus (HPV) types/putative types in lesions and in healthy skin biopsy samples.
| Genus /species, type/ putative type |
Closest related type (%)a |
Samples, no. | |||||
|---|---|---|---|---|---|---|---|
| Lesions | Healthy skin (n = 349) |
Total (n = 698) |
|||||
| SCC (n = 82) |
BCC (n = 126) |
AK (n= 49) |
Benign (n = 92) |
||||
| Beta-1 | |||||||
| HPV5 | 1 | 1 | |||||
| FA23b | HPV5 (75) | 1 | 1 | 1 | 1 | 1 | 5 |
| FA37 | HPV5 (77) | 2 | 1 | 3 | |||
| HPV8 | 1 | 1 | 2 | 4 | |||
| HPV12 | 1 | 1 | 2 | ||||
| HPV14 | 2 | 2 | 1 | 3 | 8 | ||
| HPV19 | 1 | 1 | 2 | ||||
| FA14 | HPV19 (75) | 1 | 2 | 3 | |||
| HPV20 | 3 | 2 | 7 | 6 | 18 | ||
| vs20-4c | HPV20 (76) | 1 | 2 | 3 | |||
| HPV21 | 1 | 3 | 2 | 3 | 2 | 11 | |
| FA25 | HPV21 (76) | 1 | 1 | 2 | 4 | ||
| HPV24 | 1 | 1 | 1 | 5 | 8 | ||
| FA18 | HPV24 (76) | 1 | 1 | ||||
| FAIMVS11 | HPV24 (77) | 1 | 1 | ||||
| vs75-3 | HPV24 (82) | 1 | 1 | 1 | 3 | ||
| HPV93 | 1 | 1 | 2 | ||||
| Beta-2 | |||||||
| HPV9 | 1 | 1 | 1 | 3 | |||
| HPV15 | 1 | 1 | 2 | ||||
| FA84 | HPV15 (76) | 1 | 1 | ||||
| sck59/FAN19d | HPV15 (80) | 1 | 1 | ||||
| HPV23 | 1 | 1 | 2 | 1 | 5 | ||
| FA16 | HPV23 (81) | 1 | 1 | 2 | |||
| HPV37 | 2 | 1 | 1 | 4 | |||
| HPV38 | 3 | 2 | 2 | 3 | 10 | ||
| FA51 | HPV38 (76) | 1 | 1 | 2 | |||
| HPV80 | 2 | 2 | 4 | ||||
| FA5 | HPV80 (78) | 1 | 1 | 2 | |||
| FA75 | HPV80 (79) | 1 | 1 | 1 | 3 | ||
| FA85 | HPV80 (75) | 2 | 1 | 3 | 1 | 7 | |
| vs92-1 | HPV80 (79) | 1 | 1 | ||||
| vs102-4 | HPV80 (83) | 1 | 1 | ||||
| HPV100 | 1 | 1 | |||||
| FA40 | HPV17 (76) | 1 | 1 | ||||
| Beta-3 | |||||||
| HPV49 | 1 | 1 | 1 | 3 | |||
| HPV76 | 1 | 1 | |||||
| Beta-5 | |||||||
| HPV96 | 1 | 1 | |||||
| Gamma | |||||||
| FA137 | HPV65 (69) | 1 | 1 | ||||
| FA117 | HPV50 (70) | 1 | 1 | ||||
| sck29/FAN5e | HPV50 (73) | 1 | 1 | ||||
| Alpha-4 | |||||||
| HPV27 | 1 | 1 | |||||
| Alpha-9 | |||||||
| HPV16 | 1 | 1 | |||||
NOTE. AK, actinic keratosis; BCC, basal cell carcinoma; SCC, squamous cell carcinoma.
Sequence homology between putative HPV type and HPV type, of parts of the L1 open reading frame, by Blast.
FA: ~430 nt.
vs: ~680 nt.
191 nt.
297 nt.
Table 2. Risk factors for cutaneous human papillomavirus (HPV).
| Category | HPV status, no. (%) of subjectsa |
Adjusted OR (95% CI)b |
|
|---|---|---|---|
| Positive (n = 120) |
Negative (n = 578) |
||
| Diagnosis | |||
| Healthy skin | 42 (12.0) | 307 (88.0) | 1.0 (referent) |
| Benign | 24 (26.1) | 68 (73.9) | 3.12 (1.69–5.75) |
| AK | 11 (22.5) | 38 (77.6) | 1.41 (0.64–3.07) |
| BCC | 22 (17.5) | 104 (82.5) | 1.44 (0.80–2.60) |
| SCC | 21 (25.6) | 61 (74.4) | 2.13 (1.13–4.03) |
| Age | |||
| <60 years | 10 (10.6) | 84 (89.4) | 1.0 (referent) |
| 61–75 years | 42 (18.0) | 192 (82.0) | 1.64 (0.75–3.60) |
| >76 years | 68 (18.4) | 302 (81.6) | 1.58 (0.73–3.42) |
| Sex | |||
| Female | 45 (15.0) | 255 (85.0) | 1.0 (referent) |
| Male | 75 (18.8) | 323 (81.2) | 1.57 (1.00–2.47) |
| Skin type | |||
| I | 12 (16.2) | 62 (83.8) | 1.0 (referent) |
| II | 29 (17.3) | 139 (82.77) | 1.26 (0.57–2.78) |
| III | 71 (18.2) | 319 (81.8) | 1.56 (0.74–3.27) |
| IV | 8 (12.1) | 58 (87.9) | 0.80 (0.27–2.40) |
| Self-reported previous sunburns | |||
| Never | 19 (22.1) | 67 (77.9) | 1.0 (referent) |
| A few times | 51 (17.1) | 247 (82.9) | 0.52 (0.26–1.04) |
| Sometimes not every year | 35 (14.5) | 207 (85.5) | 0.37 (0.18–0.77) |
| Every year | 15 (20.8) | 57 (79.2) | 0.82 (0.34–1.95) |
| Eye color | |||
| Blue/grey | 83 (16.6) | 417 (83.4) | 1.0 (referent) |
| Green/mixed | 27 (21.1) | 101 (78.9) | 1.26 (0.74–2.13) |
| Brown | 9 (13.6) | 57 (86.4) | 0.64 (0.29–1.42) |
| Sun exposure at biopsy site | |||
| Low/moderate | 36 (9.21) | 355 (90.8) | 1.0 (referent) |
| Extensive | 84 (27.4) | 223 (72.6) | 4.28 (2.67–6.85) |
NOTE. AK, actinic keratosis; BCC, basal cell carcinoma; CI, confidence interval; OR, odds ratio; SCC, squamous cell carcinoma.
Because some subjects did not provide information, the sum of the total numbers of subjects in the subgroups may be less than the total number of subjects.
Adjusted for all variables in the table.
In multivariate models, the presence of HPV DNA was associated with skin sites classified as having been extensively exposed to the sun (OR 4.28 [95% CI, 2.67–6.85]) (table 2). This applied to both the site of the lesion (OR, 4.45 [95% CI, 2.44–8.11]) and the site of the healthy skin sample (OR, 3.65 [95% CI, 1.79–7.44]), after adjustment for age, sex, skin type, sunburns, and eye color. SCCs and benign lesions were more commonly positive for HPV DNA than were healthy skin samples (OR, 2.13 [95% CI, 1.13–4.03] for SCCs; OR, 3.12 [95% CI, 1.69–5.75] for benign lesions) (table 2). HPV types of genus Beta-papillomavirus species 2 were more common in SCCs than in healthy skin samples (OR, 4.40 [95% CI, 1.92–10.06]), whereas HPV types of genus Beta-papillomavirus species 1 were more common in benign lesions than in healthy skin samples (OR, 3.47 [95% CI, 1.72–6.99]) (table 3).
Table 3. Diagnosis associated with human papillomavirus (HPV) of Beta-papillomavirus species 1 or 2.
| Diagnosis | Subjects, no. (%) | Adjusted OR (95% CI)a |
|
|---|---|---|---|
|
Beta-papillomavirus species 1 | |||
| Positive (n = 77) |
Negative (n = 621) |
||
| Healthy skin | 27 (7.7) | 322 (92.3) | 1.0 (referent) |
| Benign | 18 (19.6) | 74 (80.4) | 3.47 (1.72–6.99) |
| AK | 7 (14.2) | 42 (85.7) | 1.26 (0.50–3.19) |
| BCC | 13 (10.3) | 113 (84.6) | 1.20 (0.58–2.48) |
| SCC | 12 (14.6) | 70 (85.3) | 1.56 (0.72–3.39) |
|
Beta-papillomavirus species 2 | |||
| Positive (n = 50) |
Negative (n = 648) |
||
| Healthy skin | 15 (4.3) | 334 (95.7) | 1.0 (referent) |
| Benign | 7 (7.6) | 85 (92.4) | 2.13 (0.81–5.58) |
| AK | 5 (8.2) | 44 (89.8) | 1.75 (0.58–5.27) |
| BCC | 9 (7.1) | 117 (92.9) | 1.65 (0.68–3.97) |
| SCC | 14 (17.1) | 68 (82.9) | 4.40 (1.92–10.06) |
NOTE. AK, actinic keratosis; BCC, basal cell carcinoma; CI, confidence interval; OR, odds ratio; SCC, squamous cell carcinoma.
Adjusted for all variables listed in table 2.
Pairwise comparisons of lesions and healthy skin samples from the same patients found that HPV types, in particular of genus Beta-papillomavirus species 1, were more frequently detected in benign lesions than in the matched healthy skin samples (OR, 18.5 [95% CI, 3.1–∞]) (table 4). There was a tendency for HPV types of genus Beta-papillomavirus species 2 to associate with SCC and AK (table 4). Identical HPV types in the lesion and the healthy skin sample from the same patient were identified in only 2.3% (8/349) of the patients.
Table 4. Pairwise comparisons of human papillomavirus (HPV) DNA in lesions and in healthy skin biopsy samples.
| Category | Samples, no. (%) | Adjusted OR (95% CI)a |
|
|---|---|---|---|
| Lesions | Healthy skin | ||
| Benign (n = 92) | |||
| HPV | 24 (26.1) | 7 (7.6) | 17.2 (2.7–171) |
| HPV Beta-1 | 18 (19.6) | 4 (4.5) | 18.51 (3.10–∞) |
| HPV Beta-2 | 7 (7.6) | 4 (4.4) | 2.5 (0.41–26.25) |
| AK (n = 49) | |||
| HPV | 11 (22.5) | 7 (14.3) | 1.61 (0.40–7.59) |
| HPV Beta-1 | 7 (14.3) | 5 (10.2) | 1.16 (0.19–8.10) |
| HPV Beta-2 | 5 (10.2) | 1 (2.0) | 5.29 (0.66–∞) |
| BCC (n = 126) | |||
| HPV | 22 (17.5) | 14 (11.1) | 2.11 (0.85–5.75) |
| HPV Beta-1 | 13 (10.3) | 9 (7.1) | 2.0 (0.54–9.08) |
| HPV Beta-2 | 9 (7.1) | 6 (4.8) | 1.76 (0.50–6.97) |
| SCC (n = 82) | |||
| HPV | 21 (25.6) | 14 (17.1) | 1.32 (0.45–4.07) |
| HPV Beta-1 | 12 (14.6) | 9 (11.0) | 1.16 (0.38–3.67) |
| HPV Beta-2 | 13 (15.9) | 4 (4.9) | 5.84 (0.91–80.79) |
NOTE. AK, actinic keratosis; BCC, basal cell carcinoma; CI, confidence interval; OR, odds ratio; SCC, squamous cell carcinoma.
Adjusted for sun exposure at biopsy location. Because the lesion samples and healthy skin samples were taken from the same subjects, all other covariates are identical for the lesion samples and healthy skin samples and therefore are not included in the model.
DISCUSSION
The present study indicates that extensive sun exposure at the site of biopsy is the strongest risk factor for cutaneous HPV infection. Previous studies of the association between UV exposure and HPV had reported variable results, from finding no association [12] to increased prevalence of HPV on the forehead and other exposed sites than at sites that were less exposed to the sun [6, 24]. A possible explanation for the present study’s finding of a strong association between HPV and UV exposure may lie in the fact that this is the first casecontrol study of cutaneous HPV that has used “stripping” to reduce surface contamination of HPV. Also, in the present study the paired biopsy samples from each patient were taken from the same region of the body, which enhanced the possibility to detect whether a local HPV infection was specifically associated with skin lesions. This regional matching did not impair the ability to detect association with UV exposure, because, for a substantial proportion of the patients, the paired biopsy samples were taken from sites with different sun-exposure classifications. In addition, minimum misclassification was assured by the requirement that, for a sample to be scored as being positive for HPV DNA, the same HPV type had to be detected by at least 2 of the 3 laboratories.
The increased prevalence of HPV at sun-exposed skin sites may result from amplification of the viral genome via UV light, as has been reported for bovine papillomavirus (BPV) DNA in BPV-transformed cells [25]. Furthermore, in vitro studies have demonstrated increased viral promoter activity after UV irradiation, for a number of cutaneous HPV types [26, 27], as well as decreased apoptosis after UV treatment of host cells containing single cutaneous HPV types [28]. Such decreased apoptosis might lead to the accumulation of UV-induced mutations—for example, p53 mutations, which have been found in >90% of SCCs [29].
It has been established that UV exposure can cause reactivation of herpes simplex virus [30] and that local UV-immunosuppressive effects appear to be involved in such reactivation [31]. Our preferential finding of HPV DNA in sun-exposed skin sites might also be related to local UV-immunosuppressive effects [32], which may have promoted somewhat increased copy numbers of HPV. It is also tempting to speculate that the high risk of SCC among renal transplant patients [33] might be due to increased activity of HPV at sun-exposed sites, in combination with local UV immunosuppression and immunosuppressive therapy.
However, our finding of a strong association between HPV and UV exposure, the major risk factor for SCC, raises the possibility that the association between HPV and SCC could be attributable to confounding via UV exposure. Because both the extent of UV exposure and HPV infection are difficult to determine, significant effects remaining after adjustment might be caused by residual confounding. Another possibility is that a carcinogenic effect of UV exposure is in part mediated via its effect on HPV: if HPV and UV exposure act on the same causal pathway, the inclusion of UV exposure in the model could risk removal of an etiological association. Interestingly, a recent serological study found a tendency for UV exposure to be a risk factor for SCC primarily in HPV-seropositive subjects [34].
Contrary to the results of a previous study of NMSCs [35], the present study detected only a single case of genital highrisk HPV (HPV16), which was present in an SCC from the perianal region, suggesting that genital HPV types are rare in skin disorders, which were the subject of the present study. The earlier study had used nonstripped skin samples, which might cause an increased rate of detection of genital HPV types if superficially trapped on lesions [35].
In the present study, no cutaneous high-risk HPV type could be identified. However, the group of phylogenetically related HPV types of genus Beta-papillomavirus species 2 was more associated with SCC than with healthy skin, in all patients (OR, 4.40 [95% CI, 1.92–10.06]), and a similar, albeit not significant, point estimate of association was observed for SCC samples compared with matched healthy skin samples from the same patient (OR, 5.84 [95% CI, 0.91–80.79]). Other studies have not stripped the skin surface before sampling, which may explain the differences in results [12, 35-37]; for example, HPV38, of genus Beta-papillomavirus species 2, has been reported to be much more prevalent in NMSCs (50%) than in healthy skin samples (10%) [38]; however, when biopsy samples from stripped skin were analyzed by highly sensitive type-specific PCR, HPV38 was found in only 3% of NMSCs and in 1% of healthy skin samples [39].
Overall, the present study’s finding of a strong association between HPV and sun exposure, in combination with the fact that the point estimate of the association between genus Beta-papillomavirus species 2 and SCC was reduced after adjustment for sun exposure, implies that this association must be interpreted with caution. Much larger studies with strong epidemiological study designs, such as prospective cohort studies, would be necessary to allow inferences as to any possible causal role. However, the results of the present study suggest that a specific group of cutaneous HPV types (i.e., Beta-2) are an important risk factor in the etiology of SCC; in vitro studies are necessary to validate this observation.
The increased prevalence, among the benign lesions (most of which were SKs), of HPV of genus Beta-papillomavirus species 1 might be due to morphological features resulting in less efficient tape-stripping of superficially occurring HPV. The data of the present study are in line with those of a previous study that reported that HPV20 of genus Beta-papillomavirus species 1 was the predominating type in both SKs and healthy skin samples [40].
Quantitative real-time PCR, performed for a limited number of HPV types (i.e., HPV38, 92, 93, and 96) in the present study, found only very low viral copy numbers (range, 1 copy/50 cells-1 copy/70,000 cells) [39, 41]. Low copy numbers of HPV DNA in AKs and malignant skin lesions is a consistent finding in other studies also [24, 42]. The present study has not explored the issue of whether the presence of low amounts of HPV DNA is attributable to remaining superficial HPV or implies that HPV DNA is not required for maintenance of the malignant phenotype. The latter possibility is supported by the fact that HPV DNA is not detectable after in vitro passages of cutaneous SCCs, which suggests that HPV DNA is not required for maintenance of the malignant phenotype in every carcinoma cell [43, 44].
In conclusion, we find that reproducibly detectable cutaneous HPV is strongly associated with extensive sun exposure at the biopsy site and that different species of genus Beta-papillomaviruses tend to have different associations with malignant versus benign skin lesions. Therefore, it will be essential to consider both sun exposure and HPV species in future epidemiological studies of the etiology of SCC.
Acknowledgments
We thank Anna Kristoffersson for laboratory assistance and Pontus Naucler for statistical-analysis assistance.
Financial support: European Commission (grant “Viraskin” QLK2-CT-2002-01500); Swedish Research Council; Swedish Cancer Society.
Appendix
VIRASKIN STUDY GROUP
Members of the Viraskin Study Group are as follows: Katharina Slupetzky (Laboratory of Viral Oncology, Division of Immunology, Allergy and Infectious Diseases, Medical University of Vienna, Vienna, Austria), Angelika Iftner (Medical Virology, Section Experimental Virology, University Hospital of Tübingen, Tübingen, Germany), Sonja Stephan and Corinna Whitley (Division of Tumourvirus Characterization, Deutsches Krebsforschungszentrum, Heidelberg, Germany), and Christina Gerrouda (Department of Medical Microbiology, Malmö University Hospital, Lund University, Malmö, Sweden).
Footnotes
Potential conflicts of interest: none reported.
Presented in part: 23rd International Papillomavirus Conference, Prague, Czech Republic, 6 September 2006 [abstract PS 28-6].
References
- 1.Stern RS. The mysteries of geographic variability in nonmelanoma skin cancer incidence. Arch Dermatol. 1999;135:843–4. doi: 10.1001/archderm.135.7.843. [DOI] [PubMed] [Google Scholar]
- 2.MacKie RM. Awareness, knowledge and attitudes to basal cell carcinoma and actinic keratoses among the general public within Europe. J Eur Acad Dermatol Venereol. 2004;18:552–5. doi: 10.1111/j.1468-3083.2004.00993.x. [DOI] [PubMed] [Google Scholar]
- 3.Black HS, deGruijl FR, Forbes PD, et al. Photocarcinogenesis: an overview. J Photochem Photobiol B. 1997;40:29–47. doi: 10.1016/s1011-1344(97)00021-3. [DOI] [PubMed] [Google Scholar]
- 4.Akgul B, Cooke JC, Storey A. HPV-associated skin disease. J Pathol. 2006;208:165–75. doi: 10.1002/path.1893. [DOI] [PubMed] [Google Scholar]
- 5.Antonsson A, Karanfilovska S, Lindqvist PG, Hansson BG. General acquisition of human papillomavirus infections of skin occurs in early infancy. J Clin Microbiol. 2003;41:2509–14. doi: 10.1128/JCM.41.6.2509-2514.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Antonsson A, Forslund O, Ekberg H, Sterner G, Hansson BG. The ubiquity and impressive genomic diversity of human skin papillomaviruses suggest a commensalic nature of these viruses. J Virol. 2000;74:11636–41. doi: 10.1128/jvi.74.24.11636-11641.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Antonsson A, Erfurt C, Hazard H, et al. Prevalence and type spectrum of human papillomaviruses in healthy skin samples collected in three continents. J Gen Virol. 2003;84:1881–6. doi: 10.1099/vir.0.18836-0. [DOI] [PubMed] [Google Scholar]
- 8.Boxman IL, Berkhout RJ, Mulder LH, et al. Detection of human papillomavirus DNA in plucked hairs from renal transplant recipients and healthy volunteers. J Invest Dermatol. 1997;108:712–5. doi: 10.1111/1523-1747.ep12292090. [DOI] [PubMed] [Google Scholar]
- 9.Astori G, Lavergne D, Benton C, et al. Human papillomaviruses are commonly found in normal skin of immunocompetent hosts. J Invest Dermatol. 1998;110:752–5. doi: 10.1046/j.1523-1747.1998.00191.x. [DOI] [PubMed] [Google Scholar]
- 10.Wieland U, Ritzkowsky A, Stoltidis M, et al. Papillomavirus DNA in basal cell carcinomas of immunocompetent patients: an accidental association? J Invest Dermatol. 2000;115:124–8. doi: 10.1046/j.1523-1747.2000.00015.x. [DOI] [PubMed] [Google Scholar]
- 11.Forslund O, Ly H, Reid C, Higgins G. A broad spectrum of human papillomavirus types is present in the skin of Australian patients with non-melanoma skin cancers and solar keratosis. Br J Dermatol. 2003;149:64–73. doi: 10.1046/j.1365-2133.2003.05376.x. [DOI] [PubMed] [Google Scholar]
- 12.Harwood CA, Surentheran T, Sasieni P, et al. Increased risk of skin cancer associated with the presence of epidermodysplasia verruciformis human papillomavirus types in normal skin. Br J Dermatol. 2004;150:949–57. doi: 10.1111/j.1365-2133.2004.05847.x. [DOI] [PubMed] [Google Scholar]
- 13.Lörincz AT, Reid R, Jenson AB, Greenberg MD, Lancaster W, Kurman RJ. Human papillomavirus infection of the cervix: relative risk associations of 15 common anogenital types. Obstet Gynecol. 1992;79:328–37. doi: 10.1097/00006250-199203000-00002. [DOI] [PubMed] [Google Scholar]
- 14.Jablonska S, Majewski S. Epidermodysplasia verruciformis: immunological and clinical aspects. Curr Top Microbiol Immunol. 1994;186:157–75. doi: 10.1007/978-3-642-78487-3_9. [DOI] [PubMed] [Google Scholar]
- 15.de Villiers E-M, Fauquet C, Broker TR, Bernard H-U, zur Hausen H. Classification of papillomaviruses. Virology. 2004;324:17–27. doi: 10.1016/j.virol.2004.03.033. [DOI] [PubMed] [Google Scholar]
- 16.Forslund O, Lindelof B, Hradil E, et al. High prevalence of cutaneous human papillomavirus DNA on the top of skin tumors but not in “stripped” biopsies from the same tumors. J Invest Dermatol. 2004;123:388–94. doi: 10.1111/j.0022-202X.2004.23205.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Fitzpatrick TB. The validity and practicality of sun-reactive skin types I through VI. Arch Dermatol. 1988;124:869–71. doi: 10.1001/archderm.124.6.869. [DOI] [PubMed] [Google Scholar]
- 18.Forslund O, Antonsson A, Nordin P, Stenquist B, Hansson BG. A broad range of human papillomavirus types detected with a general PCR method suitable for analysis of cutaneous tumours and normal skin. J Gen Virol. 1999;80:2437–43. doi: 10.1099/0022-1317-80-9-2437. [DOI] [PubMed] [Google Scholar]
- 19.Forslund O, Ly H, Higgins G. Improved detection of cutaneous human papillomavirus DNA by single tube nested ‘hanging droplet’ PCR. J Virol Methods. 2003;110:129–36. doi: 10.1016/s0166-0934(03)00109-5. [DOI] [PubMed] [Google Scholar]
- 20.Berkhout RJ, Tieben LM, Smits HL, Bavinck JN, Vermeer BJ, ter Schegget J. Nested PCR approach for detection and typing of epidermodysplasia verruciformis-associated human papillomavirus types in cutaneous cancers from renal transplant recipients. J Clin Microbiol. 1995;33:690–5. doi: 10.1128/jcm.33.3.690-695.1995. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.de Villiers EM, Lavergne D, McLaren K, Benton EC. Prevailing papillomavirus types in non-melanoma carcinomas of the skin in renal allograft recipients. Int J Cancer. 1997;73:356–61. doi: 10.1002/(sici)1097-0215(19971104)73:3<356::aid-ijc9>3.0.co;2-z. [DOI] [PubMed] [Google Scholar]
- 22.Hall TA. BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symp Ser. 1999;41:95–8. [Google Scholar]
- 23.Kumar S, Tamura K, Jakobsen IB, Nei M. MEGA2: molecular evolutionary genetics analysis software. Bioinformatics. 2001;17:1244–5. doi: 10.1093/bioinformatics/17.12.1244. [DOI] [PubMed] [Google Scholar]
- 24.Meyer T, Arndt R, Christophers E, Nindl I, Stockfleth E. Importance of human papillomaviruses for the development of skin cancer. Cancer Detect Prev. 2001;25:533–47. [PubMed] [Google Scholar]
- 25.Schmitt J, Schlehofer JR, Mergener K, Gissmann L, zur Hausen H. Amplification of bovine papillomavirus DNA by N-methyl-N′-nitro-N-nitrosoguanidine, ultraviolet irradiation, or infection with herpes simplex virus. Virology. 1989;172:73–81. doi: 10.1016/0042-6822(89)90108-6. [DOI] [PubMed] [Google Scholar]
- 26.Purdie KJ, Pennington J, Proby CM, et al. The promoter of a novel human papillomavirus (HPV77) associated with skin cancer displays UV responsiveness, which is mediated through a consensus p53 binding sequence. EMBO J. 1999;18:5359–69. doi: 10.1093/emboj/18.19.5359. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Ruhland A, de Villiers EM. Opposite regulation of the HPV 20-URR and HPV 27-URR promoters by ultraviolet irradiation and cytokines. Int J Cancer. 2001;91:828–34. doi: 10.1002/1097-0215(200002)9999:9999<::aid-ijc1129>3.0.co;2-b. [DOI] [PubMed] [Google Scholar]
- 28.Jackson S, Storey A. E6 proteins from diverse cutaneous HPV types inhibit apoptosis in response to UV damage. Oncogene. 2000;19:592–8. doi: 10.1038/sj.onc.1203339. [DOI] [PubMed] [Google Scholar]
- 29.Leffell DJ. The scientific basis of skin cancer. J Am Acad Dermatol. 2000;42:18–22. doi: 10.1067/mjd.2000.103340. [DOI] [PubMed] [Google Scholar]
- 30.Ichihashi M, Nagai H, Matsunaga K. Sunlight is an important causative factor of recurrent herpes simplex. Cutis. 2004;74:14–8. [PubMed] [Google Scholar]
- 31.Goade DE, Nofchissey RA, Kusewitt DF, et al. Ultraviolet light induces reactivation in a murine model of cutaneous herpes simplex virus-1 infection. Photochem Photobiol. 2001;74:108–14. doi: 10.1562/0031-8655(2001)074<0108:uliria>2.0.co;2. [DOI] [PubMed] [Google Scholar]
- 32.Kelly DA, Young AR, McGregor JM, Seed PT, Potten CS, Walker SL. Sensitivity to sunburn is associated with susceptibility to ultraviolet radiation-induced suppression of cutaneous cell-mediated immunity. J Exp Med. 2000;191:561–6. doi: 10.1084/jem.191.3.561. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Lindelof B, Sigurgeirsson B, Gabel H, Stern RS. Incidence of skin cancer in 5356 patients following organ transplantation. Br J Dermatol. 2000;143:513–9. [PubMed] [Google Scholar]
- 34.Karagas MR, Nelson HH, Sehr P, et al. Human papillomavirus infection and incidence of squamous cell and basal cell carcinomas of the skin. J Natl Cancer Inst. 2006;98:389–95. doi: 10.1093/jnci/djj092. [DOI] [PubMed] [Google Scholar]
- 35.Iftner A, Klug SJ, Garbe C, et al. The prevalence of human papillomavirus genotypes in nonmelanoma skin cancers of nonimmunosuppressed individuals identifies high-risk genital types as possible risk factors. Cancer Res. 2003;63:7515–9. [PubMed] [Google Scholar]
- 36.Shamanin V, zur Hausen H, Lavergne D, et al. Human papillomavirus infections in nonmelanoma skin cancers from renal transplant recipients and nonimmunosuppressed patients. J Natl Cancer Inst. 1996;88:802–11. doi: 10.1093/jnci/88.12.802. [DOI] [PubMed] [Google Scholar]
- 37.Harwood CA, Surentheran T, McGregor JM, et al. Human papillomavirus infection and non-melanoma skin cancer in immunosuppressed and immunocompetent individuals. J Med Virol. 2000;61:289–97. doi: 10.1002/1096-9071(200007)61:3<289::aid-jmv2>3.0.co;2-z. [DOI] [PubMed] [Google Scholar]
- 38.Caldeira S, Zehbe I, Accardi R, et al. The E6 and E7 proteins of the cutaneous human papillomavirus type 38 display transforming properties. J Virol. 2003;77:2195–206. doi: 10.1128/JVI.77.3.2195-2206.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Hazard K, Eliasson L, Dillner J, Forslund O. Subtype HPV38b[FA125] demonstrates heterogeneity of human papillomavirus type 38. Int J Cancer. 2006;119:1073–7. doi: 10.1002/ijc.21920. [DOI] [PubMed] [Google Scholar]
- 40.Li YH, Chen G, Dong XP, Chen HD. Detection of epidermodysplasia verruciformis-associated human papillomavirus DNA in nongenital seborrhoeic keratosis. Br J Dermatol. 2004;151:1060–5. doi: 10.1111/j.1365-2133.2004.06244.x. [DOI] [PubMed] [Google Scholar]
- 41.Vasiljević N, Hazard K, Eliasson L, et al. Characterization of two novel cutaneous human papillomaviruses, HPV93 and HPV96. J Gen Virol. 2007;88:1479–83. doi: 10.1099/vir.0.82679-0. [DOI] [PubMed] [Google Scholar]
- 42.Weissenborn SJ, Nindl I, Purdie K, et al. Human papillomavirus-DNA loads in actinic keratoses exceed those in non-melanoma skin cancers. J Invest Dermatol. 2005;125:93–7. doi: 10.1111/j.0022-202X.2005.23733.x. [DOI] [PubMed] [Google Scholar]
- 43.Boxman IL, Mulder LH, Vermeer BJ, Bavinck JN, ter Schegget J, Ponec M. HPV-DNA is not detectable in outgrowing cells from explant cultures of skin lesions established at the air-liquid-interface. J Med Virol. 2000;61:281–8. doi: 10.1002/1096-9071(200007)61:3<281::aid-jmv1>3.0.co;2-p. [DOI] [PubMed] [Google Scholar]
- 44.Purdie KJ, Sexton CJ, Proby CM, et al. Malignant transformation of cutaneous lesions in renal allograft patients: a role for human papillomavirus. Cancer Res. 1993;53:5328–33. [PubMed] [Google Scholar]