Cutaneous human papillomavirus 88: Remarkable differences in viral load

Abstract

A human papillomavirus (HPV) was cloned from a patient with multiple squamous cell carcinomas (SCCs) and identified as HPV88, recently categorized into a new species within the genus Gamma. The HPV88 viral load in an SCC of the index patient exceeded 1 million copies/cell. By contrast, a survey of 447 skin lesions (79 actinic keratoses, 73 seborrhoeic keratoses, 169 basal cell carcinomas and 126 SCCs) and 362 healthy skin biopsies found detectable HPV88 DNA in only 7 specimens. All these had very low viral loads (<1 copy/10³ cells) implying extreme biological variability in viral load.

The human papillomaviruses (HPVs) comprise a large and diverse group of viruses that induce hyperproliferation in both cutaneous and genital epithelia. There are more than 100 completely characterized types and about 150 putative, partially sequenced types.^1-4 A novel HPV-type has less than 90% similarity of the L1 gene to any known HPV-type, a species has less than 70% similarity and a genus less than 60% similarity.⁵

Nonmelanoma skin cancer (NMSC), including squamous cell carcinoma (SCC) and basal cell carcinoma (BCC), is the most common cancer among Caucasians.⁶ The main risk factor for developing NMSC is exposure to ultraviolet radiation (UVR).⁷ An association of HPV infection with skin cancer was first demonstrated in patients with epidermodysplasia verruciformis (EV).⁸ These immunosuppressed patients are highly susceptible to HPV infections that often progress to SCC. Cutaneous HPV is also commonly found among the immunocompetent, both in benign and malignant skin lesions⁹ as well as in healthy skin.¹

In the case of genital HPVs, HPV-carrying cancers are always carrying at least 1 viral genome per cell and high viral load is an important risk factor for the development of cervical carcinoma.^10,11 By contrast, cutaneous HPV types are typically found in very small amounts (1 copy/100 cells or less) in both skin tumors and healthy skin.^12-14 Exceptions are, for example, HPV1 and 2,^15,16 found in skin warts, and the 2 EV-types HPV5 and 8.^17,18

The significance of low amounts of virus was questioned when it was found that the proportion of healthy skin samples that are positive for HPV can be strongly reduced by cleansing of the skin surface with tape.³

In this study, we report an invasive digital SCC that contained extremely high viral loads of HPV88. A survey of more than 800 skin biopsies found only 7 positive specimens, all with more than 10⁹ times lower viral load.

Patients and methods

Index patient

Scrapings from an invasive digital SCC from a 28-year-old immunosuppressed male with human immunodeficiency virus (HIV) infection were extracted using DNeasy blood and tissue kit (QIAGEN, Hilden, Germany). Sterile scalpels, a new one for each piece of tissue, were used to prevent contamination. The patient had several extensive SCCs on other fingers and one toe, condylomas and perianal bowenoid plaques. Previous afflictions included anal cancer and chronic infections with hepatitis B and C. PCR using degenerate primers had identified an HPV sequence not present in GenBank.¹⁹

MDA and overlapping amplicon

To preferentially amplify HPV in multiple displacement amplification (MDA) 2 thiophosphate-modified primers generic to any of the 72 HPV types in the HPV sequence database (Los Alamos, 1997) were designed.

One microliter extracted DNA from the left finger SCC and 90 μM of each primer (JK_MDA1F: 5′-TAAAACGAA^SA^SG-3′, JK_MDA1R: 5′-ATDCCATTR^ST^ST-3′ and thiophosphate-modified random hexamer (Fidelity Systems) D = A, G, T; R = A, G; S = thiophosphate – modification), 36 mM Tris–HCl, pH 8 and 9 mM MgCl₂ (Roche, Mannheim, Germany) in a volume of 5.5 μl were denatured for 3 min at 94°C, cooled on ice and mixed with a 14.5-μl solution containing 10 U Phi29 polymerase (Fermentas, Leon-Rot, Germany), 1× Phi29 polymerase reaction buffer (Fermentas) and 1.4 mM dNTP (Roche). Incubation for 16 hr at 30°C, inactivation for 10 min at 65°C, dilution 1:2 in TE-buffer (10 mM Tris, pH 8, 1 mM EDTA) and shaking at 500 rpm at +4°C for 8 hr followed.

After digestion with 40 U of BamHI and electrophoresis, we attempted to clone the full-length fragment of ~8 kb, but only managed to clone a 5.4 kb fragment. The T-A cloning used gel-excised band purified with QIAquick Gel Extraction kit (QIAGEN) and incubation with 0.6 mM dNTP and Taq polymerase for 15 min at 72°C (TOPO TA Cloning Invitrogen, Carlsbad, CA).

The remainder of the genome was amplified by Expand High Fidelity PCR (Roche) with 2.5 μl of DNA from the patient, following the manufacturer’s instructions with some modifications; 0.5 μM of each primer, LIF-F: 5′-CAGGTAGCGTTGGACTCCTTAG-3′, LIF-R: 3′-GTACCCTTTGGCAAAGTCACTG-5 (Cybergene, Huddinge, Sweden) and an annealing temperature of 57°C. The PCR mixture was prepared in a clean room, separated from the PCR. The product was electrophoresed and a band of ~3 kb was excised, purified and cloned as described earlier.

Sequencing

Sequencing used primer walking with ABI Prism Big Dye Terminator Cycle Sequencing Ready Reaction kits (Applied Biosystems, Foster City) and a 3730 sequencer (Applied Biosystems). The complete sequence (GenBank accession number: EF467176) was deposited with the International Reference Center for Papillomaviruses at the German Cancer Research Center, Heidelberg, Germany, on October 31, 2006. The sequence was identical to HPV88, which had previously been submitted to the Reference Center by Drs. Egawa and Kitasato, but not deposited in GenBank.

Patients

Skin biopsies were collected from 362 immunocompetent patients, attending Swedish and Austrian hospitals. The sample series of 316 patients from Sweden included actinic keratosis (AK) (n = 49, mean age 75 years), seborrhoic keratosis (SK) (n = 73, mean age 73 years), BCC (n = 118, mean age 76 years) or SCC (n = 76, mean age 79 years). The 46 Austrian samples included 9 patients with AK (mean age 78 years), 29 patients with BCC (mean age 72 years) and 8 patients with SCC (mean age 80 years). After tape stripping,³ a biopsy was taken from the lesion and from adjacent healthy skin of the same patient. Hence, totally, 724 samples were included. The DNA was extracted using a phenol-free method.²⁰

In addition, biopsies of lesions from 26 immunosuppressed patients with AK (n = 5) or SCC (n = 21) were collected at a dermatology clinic in Austria. The DNA was extracted using DNeasy blood and tissue kit (QIAGEN). The patients had a mean age of 62 years.

A third sample series of 59 skin biopsies were obtained from patients attending a dermatology clinic in Australia² diagnosed with AK (n = 16), BCC (n = 22) or SCC (n = 21). Thirty-eight patients were immunocompetent (mean age 72 years) and 21 were immunosuppressed (mean age 63 years). DNA was extracted using the phenol-free method.²⁰ All patients provided informed consent, and the studies were approved by the appropriate Ethical Review Boards in Sweden, Austria and Australia.

Real-time PCR

Primers and probes, for real-time PCR, specific for HPV88 and HPV26 (GenBank accession number: NC_001583) were designed using Primer Express 2.0 software program (Applied Biosystems); HPV88F: 5′-AAGCCCACAGAATCCTCTATCG-3′ and HPV88R: 5′-ACCACCCCCTGTTGCAGAA-3′, HPV26F: 5′-CACTTTGTATATTAAAGGTGCTGAATCAG-3′and HPV26R: 5′-TAGTTGTGCATCCGAAGTAACCA-3′, HPV88-probe: 5′-FAM-CCTAAATGAAAGTGGGATTCCAGACCCTACAA-TAMRA-3′, HPV26-probe: FAM-CCCTACATCTTCTATTTATTCTGCTACACCTAGTGGCTC-TAMRA-3′ (Cybergene). Standard curves used serial dilutions from 100,000 copies to one copy of purified viral DNA from plasmids containing the MDA- or the PCR-fragment of HPV88 or HPV26 with a background of 10 ng/μl placental DNA. A sensitivity of 1 copy/sample was routinely detected. PCR mixtures, prepared in a clean room, and reaction conditions were as previously described¹² and the PCR was carried out in a GeneAmp^® 5700 SDS (Applied Biosystems).

The samples from the 316 Swedish patients were tested at a 1:2 dilution in TE-buffer (10 mM Tris, pH8, 1 mM EDTA). All other samples, including 8 SCC biopsies from the index patient, were tested undiluted. Water (Sigma-Aldrich, Steinheim, Germany) was used as a nontemplate control in each run. To be considered positive, a sample had to be positive in at least 2 of 3 analyses.

All samples were also analyzed for human DNA content using real-time PCR for the β-globin gene.¹² To prevent PCR contamination, the samples were prepared in a room separated from the PCR.

Direct visualization

Fifty microliters of unamplified extracted DNA from the original tumor of the index patient, containing ~5 × 10¹⁰ copies of HPV88, was digested with 20 U of BamHI (Fermentas). After digestion, the sample was run on a 1% agarose gel stained with 2 mg ethidium bromide (EtBr).

Results

HPV88 had a genome of 7326 bp with a GC content of 40% and a genomic organization typical for papillomaviruses. HPV88 was most closely related to HPV60 with a similarity of only 61%. HPV88 represents a new species that has been designated as genus Gamma, species 5.⁵ Table I²¹ shows the similarity of the HPV88 proteins to corresponding proteins of other HPV types.

TABLE I.

SIMILARITY (%) OF HPV88 PROTEINS TO REPRESENTATIVE TYPES

HPV 88		E6	E7	E1	E2	E41	L2	L1
Related HPV type (species)	Accession number
HPV 60 (γ4)	NC_001693	47.9	51.0	59.3	50.4	32.5	41.5	61.1
HPV 95 (γ1)	AJ620210	46.2	50.5	58.0	49.5	33.3	44.2	60.0
HPV 4 (γ1)	NC_001457	46.9	47.1	57.2	48.7	33.3	46.6	59.3
HPV 48 (γ3)	NC_001690	42.1	48.0	50.8	47.8	24.1	37.5	58.8
HPV 65 (γ1)	X70829	48.3	43.6	57.6	46.3	34.0	46.2	58.6
HPV 50 (γ2)	NC_001691	45.1	43.9	50.0	47.3	23.5	39.3	57.6
HPV 38 (β2)	HPU31787	34.0	37.5	45.7	36.3	21.1	32.7	54.8
HPV 93 (β1)	AY382778	33.1	36.4	46.5	33.9	15.4	33.9	54.2
HPV 1032	DQ080078	-3	37.0	47.6	41.7	19.9	39.4	52.5
HPV 1012	NC_008189	-3	35.7	45.8	40.5	19.5	34.5	48.0
HPV 16 (α9)	NC_001526	25.3	27.6	40.3	29.9	11.6	25.5	45.4

¹Similarities calculated from complete ORF.

²HPV101 and 103 are not categorized into a genus²¹.

³HPV101 and 103 lack an E6 open reading frame²¹.

The viral load of the SCC that HPV88 was cloned from was exceptionally high, 1.3 × 10⁶ copies/cell (Table II). This amount of virus was so high that the viral genome could be directly visualized on an EtBr-stained gel (Fig 1). A dominating band of ~8 kb is clearly visible, confirmed to be HPV88 by Southern blot (data not shown).

TABLE II.

PRESENCE AND VIRAL LOAD OF HPV88 AND HPV26¹

	HPV88 viral copies/cell	HPV26 viral copies/cell
Index patient with multiple finger SCC2 (Austria)
Right finger 1	1.6	44167
Right finger 3	1.4	284553
Right finger 4	0.1	1672
Right finger 5	0.1	563
Left finger 1	33	1.2
Left finger 2	1.3 × 106	0.4
Left finger 4	15,888	1.0
Left finger 5	144	0.93
Immunocompetent patients (Sweden)
1 AK2	1.1 × 10−4
48 AK	Negative
1 SK2	7.6 × 10−4
72 SK	Negative
118 BCC2	Negative
76 SCC	Negative
316 biopsies from healthy skin	Negative
Immunosuppressed patients (Austria)
1 AK	3.1 × 10−4
4 AK	Negative
4 SCC	1.3 × 10−4
	2.3 × 10−4
	4.6 × 10−4
	3.8 × 10−5
17 SCC	Negative
Immunocompetent (n = 38) and immunosupressed (n = 21)  patients (Australia)
16 AK	Negative
22 BCC	Negative
21 SCC	Negative
Immunocompetent patients (Austria)
9 AK	Negative
29 BCC	Negative
8 SCC	Negative
46 biopsies from healthy skin	Negative

¹Mean value calculated on three values.

²SCC, squamous cell carcinoma; AK, actinic keratosis; SK, seborrhoeic keratosis; BCC, basal cell carcinoma.

³Mean value calculated on two values

Figure 1.

Direct visualization of HPV88 on an EtBr-stained gel, after digestion with BamHI, lane 1. Lane M, Gene Ruler 1 kb DNA ladder (Fermentas).

Histological evaluation revealed a well-differentiated SCC. The tumorous lesion showed a hyperkeratotic squamous epithelium with acanthosis, papillomatosis and numerously dispersed koilocytes, the latter being a typical HPV-associated phenomenon. Nuclear polymorphism with modest hyper- and hypochromatic nuclei was observed in basal and midepithelial layers. The epithelium revealed a partial loss of stratification, and tumor cell nests were occasionally found in the dermis.

All 8 SCCs from different fingers of the index patient were positive for HPV88. However, the SCCs on the left hand consistently had very high copy numbers, whereas the SCCs of the right hand fingers had copy numbers ranging from 0.1 to 1.6 copies/cell (Table II).

By contrast, HPV26 was present in high copy numbers (ranging between 56 and 44167 copies/cell) in all SCCs on the right hand, but was present in much lower copy numbers in the SCCs of the left hand (ranging from 0.4 to 1.2 copies/cell) (Table II).

Among the Swedish immunocompetent patients, HPV88 was detected in only 2/316 (0.6%) and the viral load was very low, 7.6 × 10⁻⁴ and 1.1 × 10⁻⁴, respectively (Table II). None of the healthy skin biopsies from Sweden was positive (Table II).

Five of the 26 immunosuppressed patients that were enrolled at the same dermatology clinic in Austria as the index patient were positive for HPV88 (Table II). However, the viral loads were more than 10⁹ times lower than in the index SCC (ranging from 3.8 × 10⁻⁵ to 4.6 × 10⁻⁴ copies/cell) (Table II). Both the series with 46 immunocompetent patients from Austria, and the 59 patients enrolled in Australia were entirely negative for HPV88 (Table II).

Discussion

We report the existence of extreme variability in viral load for a cutaneous HPV. Viral loads in skin tumors reported previously are rather low, usually much less than 1 copy/cell.^12-14 An exception is the index BCC from which HPV92 was cloned that had 94 copies of HPV92 per cell.¹⁴ The viral load in the index SCC of this study (1.3 × 10⁶ copies/cell) is exceptionally high even when comparing wih the viral loads found in tumors caused by genital HPVs.^10,11 The direct visualization of the viral genome on an EtBr-stained agarose gel is in line with the high copy-number demonstrated by the real-time PCR.

Before cloning, we amplified the virus using MDA, because the MDA has a lower error rate than PCR,^22,23 but considering the very high amount of virus, cloning would most likely have been possible without any prior amplification.

There was a mutually exclusive presence of high viral loads of HPV88 and HPV26 in 8 separate tumors that the patient had on his different fingers. The 4 tumors on the left hand demonstrated high loads of HPV88 and low loads of HPV26, whereas the opposite conditions were observed on the right hand.

Whether PCR detectability of cutaneous HPV DNA reflects a biologically meaningful infection or may be a viral contamination of the skin surface is difficult to resolve, but the finding that simple skin surface cleansing with tape clearly reduces the proportion of HPV-positive biopsies³ does indicate that some caution is warranted when interpreting the findings of very low viral loads. In this case, the fact that the viral loads on the left hand SCCs were up to 10⁷ times higher than the SCCs of the right hand provokes the question of whether the high HPV88 viral loads of the left hand may have resulted in a contamination of the right hand specimens and conversely whether the high HPV26 viral loads of the right hand may have contaminated the left hand specimens.

If it is possible for a patient with a high viral load lesion to contaminate other bodily sites of himself, it should also be asked whether contamination also of other subjects is possible. The viral loads of the few HPV88 positive specimens from other patients were all less than 8 × 10⁻⁴ copies/cell and 5/7 positive were immunosuppressed patients that had visited the same dermatology clinic as the index patient. Cleansing of the skin by tape stripping before taking a biopsy was not performed in the large series of samples from Australian patients that were uniformly HPV88 negative in this study. The index patient case does suggest that the HPV88 can be associated with SCC in the immunosuppressed. However, as we also had a series of uniformly HPV88-negative non-Austrian immunosuppressed patients and SCC patients, it appears that the HPV88 infection is not generally associated with immunosuppression or with SCC.

A major lesson of this study is that attempts to detect HPV in skin tumors of suspected viral etiology must be able to detect as many species of HPV as possible. HPV88 is only distantly related to known HPVs and is not readily detected by most of the general primer systems commonly in use today. Testing only with the commonly used “general” HPV detection systems would probably have concluded that both the left and right hand tumors were caused by HPV26. But with knowledge of the HPV88 quantities present, it seems more likely that HPV88 was involved in the etiology of the left hand tumors.

In conclusion, the case of HPV88 demonstrates that cutaneous HPVs may be present in extremely variable viral loads. This may have implications for design and interpretation of studies of the role of cutaneous HPV in human skin diseases.