Methods

PATIENTS AND VOLUNTEERS

Peripheral blood for the p53 codon 72 polymorphism analysis was taken from 55 patients with skin cancer attending the dermatology clinic at Beaumont Hospital and from 115 matched ethnic Irish control volunteers. Forty nine benign and malignant skin lesions comprising viral warts (19), squamous cell carcinomas (SCCs; 21), and normal skin (nine) from 34 of the 55 patients with skin cancer were examined to determine the HPV types present. Twenty normal skin samples from 20 of the control volunteers were also examined for HPV. We obtained ethics committee approval for our study.

PCR ANALYSIS OF P53 codon 72 polymorphism

The amplification of p53 proline and arginine sequences was performed as previously described by Storey et al.¹ The products of each of the two polymerase chain reaction (PCR) amplifications were combined and analysed by 12% (wt/vol) polyacrylamide gel electrophoresis (sample gel shown in fig 1).

STATISTICAL ANALYSIS

Chi squared analysis was used to examine the differences in proportions of the p53 codon 72 genotypes between the patients with skin cancer and the control group. The odds ratio (OR) for having the arginine variant and the development of skin cancer was also calculated. Exact binomial 95% confidence intervals (CI) were calculated for the proportions of the genotypes in each group.

HPV TYPING

DNA was extracted from viral warts, SCCs, and normal skin by conventional methods. Two PCR strategies were used for the detection of HPV: a degenerate PCR using MY09 and MY11 degenerate primers, as described previously by Manos et al,⁹ and an EV specific nested PCR, using CP65–70-EV specific HPV primers, as described previously by Berkhout et al.¹⁰ Using both the degenerate and EV specific primer sets for the detection of HPV has recently been shown to be much more effective than using either in isolation.⁶ All PCR protocols contained positive, negative, and contamination controls, and were performed in a DNA Engine™ thermal cycler (MJ Research, Massachusetts, USA), initially using Taq polymerase (Promega, Madison, Wisconsin, USA), followed by reamplification with Pfu polymerase (Stratagene, Cambridge, UK). All Pfu generated PCR products were purified by TAE agarose (2% wt/vol) gel electrophoresis and the amplified DNA recovered with the Wizard™ PCR prep kit (Promega), according to the manufacturer's instructions. All purified fragments were then cloned into pCR Script AMP SK(+) (Stratagene) and cycle sequenced using CY5 labelled M13 universal primer and the Thermosequenase fluorescent labelled primer cycle sequencing kit (Amersham, Little Chalfont, Buckinghamshire, UK). Cycle sequenced products were then analysed on an ALF express genetic analyser (Pharmacia, Uppsalla, Sweden). All sequences generated were aligned with known HPV sequences in the Genome Sequence Database using the Basic Local Alignment Search Tool (BLAST).¹¹

Results

ANALYSIS OF THE DISTRIBUTION OF P53 codon 72 genotypes among patients and controls

The proportions of p53 codon 72 genotypes found were 78% arginine homozygous, 2% proline homozygous, and 20% heterozygous among patients with skin cancer and 79% arginine homozygous, 3.5% proline homozygous, and 17.5% heterozygous among the control population (fig 2). Statistical analysis showed that there were no significant differences in the distribution of the various alleles between the control group and the skin cancer group (χ² = 0.49; degrees of freedom (df) = 2; p = 0.78). The OR for the association of the arginine variant with the development of skin cancer was calculated at 1.06 (95% CI, 0.49 to 2.3).

HPV TYPING

EV associated HPV types predominate in SCCs (both from renal transplant recipients and immunocompetent patients with skin cancer) and both EV associated HPV types and common cutaneous HPV types are detected in viral warts, again from both skin cancer groups (table 1). Normal skin from the patients with skin cancer and the control volunteers was also found to harbour EV associated HPV types; however, the frequency of detection was much lower in the control volunteers than in patients with SCCs or viral warts (table 2). Table 3 summarises our findings.

Discussion

The p53 codon 72 polymorphism has been implicated as a risk factor in several human cancers.^12–14 individuals who are proline homozygotes have been shown to have a greater risk of developing lung cancer.^{15, 16} A recent study¹ suggests that the arginine isoform of p53 facilitates degradation by the E6 protein of low risk HPV-11. Because cutaneous carcinomas frequently harbour multiple low risk HPV types and not high risk HPV-16/18 it would seem that the presence of an arginine instead of a proline could be an important risk factor in the development of these tumours. All of the patients with skin cancer analysed in the study by Storey et al were either arginine homozygotes (75%) or heterozygotes (25%),¹ implying that the arginine isoform of p53 confers a susceptibility to tumorigenesis, although only a small number of patients were examined. A second study conducted on 29 immunocompetent patients who had developed skin cancer also suggests that possession of the arginine isoform of p53 is associated with an increased risk of developing HPV associated skin cancer.¹⁷

Our results do not confirm this hypothesis. We have shown that, at least in a Celtic population, there is no apparent link between the presence of HPV, the p53 codon 72 genotype, and the development of cutaneous carcinomas in both renal transplant recipients and in immunocompetent patients with skin cancer. This conclusion is supported by a recent study that also found no apparent link between a susceptibility to skin cancer and the arginine variant of p53 after renal transplantation.¹⁸ It is possible, however, that in populations less prone to skin cancer, where the arginine isoform of p53 is less prevalent,¹⁹ that it could still confer increased susceptibility to malignancy because both our study and that of Marshall and colleagues¹⁸ were conducted on northern European populations. However, it has been shown that in some populations where the proline isoform predominates, individuals homozygous for the arginine isoform are at no greater risk of developing cervical carcinoma.^20–22 Whether or not this remains the case for cutaneous malignancies is yet to be established.

The HPV types demonstrated in the patients with skin cancer were also found in the control group, albeit at a much lower frequency. However, EV associated HPV types have been detected in the hair follicles of 46% of healthy volunteers,²³ so that a reservoir of EV HPV might exist in a latent or subclinical state in the population as a whole to a greater degree than that found in our study. Recently, intratype variation in the E6 gene of HPV-16 has been suggested as a possible additional factor in determining the susceptibility to cervical cancer associated with the p53 arginine isoform.²⁴ HPV types with a G instead of a T at position 350 of the E6 gene were found to be significantly over-represented in arginine homozygous women with cervical cancer.

We conclude that in a Celtic population there is no correlation between the presence of HPV and the development of skin cancer and that p53 codon 72 polymorphism is not associated with cutaneous malignancies.