Cumulative sunlight exposure and the risk of developing skin cancer in Florida

Genetic factors determining cutaneous basal cell carcinoma phenotype

BACKGROUND

Basal cell carcinoma (BCC) patients demonstrate considerable phenotypic diversity. The basis of this heterogeneity is poorly understood. We have shown that presentational phenotypes are associated with BCC numbers. Thus, patients with a cluster of new BCC at any presentation comprise a subgroup, termed MPP, that is at increased risk of developing numerous lesions. Patients with more than one cluster (multiple cluster MPP) are at particular risk.

PROCEDURE

We determined in a cohort of BCC cases, whether: (i) tumor accrual was altered after clustering; and (ii) multiple cluster MPP is associated with characteristics linked with sensitivity to UV or, GSTT1, GSTM1, GSTM3, GSTP1, MC1R, CYP2D6, TNF-alpha, and VDR genotypes previously associated with BCC presentational phenotypes.

RESULTS

(i) After clustering BCC accrual increased; and (ii) exposure to UV in single and multiple cluster MPP cases were similar. In multiple cluster cases, mean age at first presentation with a single tumor occurred earlier and, the frequencies of CYP2D6 EM (94.4%) and GSTT1 null (41.2%) were significantly greater (P = 0.028 and P = 0.004) than in single cluster cases (67.1 and 14.3%). The odds ratios for these associations with the multiple cluster MPP were large; 15.5 and 7.39, respectively.

CONCLUSIONS

The finding of clusters of new, primary BCC is a critical event that is followed by markedly increased accrual of further tumors. Clustering occurs at a relatively late age and may be associated with a failure in immune surveillance. We propose the MPP is not the consequence of excessive UV exposure but reflects the presence of a distinct BCC subgroup defined by a combination of risk genes.

Basal cell carcinomas: association of allelic variants with a high-risk subgroup of patients with the multiple presentation phenotype

Previous studies have shown that patients who present at first or a later presentation with a cluster of new basal cell carcinoma (BCC) comprise a subgroup, termed multiple presentation phenotype (MPP), that is at increased risk of developing further lesions. In this study, we examined the hypothesis that patients who develop multiple clusters are a high-risk subgroup. We found, in a total group of 926 BCC patients, 32 patients with 2-5 BCC clusters (multiple cluster MPP) and 113 cases with only one cluster (single cluster MPP). Multiple cluster MPP cases had mean of 11.3 BCC compared with 3.7 in single cluster MPP cases during similar follow-up. Ultraviolet (UV) exposure in these groups was similar. We determined whether the multiple cluster MPP was associated with characteristics associated with sensitivity to UV or glutathione S-transferase (GST) GSTT1, GSTM1, cytochrome P450 (CYP) CYP2D6, tumour necrosis factor (TNF)-alpha and vitamin D receptor (VDR) genotypes previously associated with BCC presentational phenotypes. While the frequencies of blue eyes and male gender were greater in multiple cluster than single cluster cases, these differences were not significant. In multiple cluster cases, mean age at first presentation with single tumours occurred earlier and the frequencies of CYP2D6 extensive metabolizer (EM) (94.4%) and GSTT1 null (41.2%) were significantly greater (P = 0.028 and P = 0.004) than in single cluster cases (67.1% and 14.3%, respectively). The odds ratios for the individual associations of CYP2D6 EM and GSTT1 null with the multiple cluster MPP were relatively larger; 15.5 and 7.39, respectively. TNF-alpha and VDR genotypes were not associated with multiple cluster MPP. We propose that the MPP is not the consequence of excessive UV exposure but rather reflects the presence of a distinct BCC subgroup which is defined by combinations of risk genes.

Basal cell carcinoma

BACKGROUND

Basal cell carcinoma (BCC) is characterized by marked interpatient variation in tumor accrual. The authors previously reported that presentation with a cluster of BCC is associated with an inherited predisposition to develop many additional lesions suggesting clustering is a critical event. A cluster is defined as the presence of two or more new, primary BCCs, at initial or later presentation.

METHODS

The authors recruited 927 cases and determined whether 1) clustering was an early or late event and 2) tumor accrual was altered after clustering.

RESULTS

In the cases, 669 patients developed only 1 lesion, 112 patients presented more than once but with single lesions (single presentation phenotype[SPP]-more), 94 cases had a cluster at first presentation (multiple presentation phenotype [MPP]-cluster initial), and 52 cases first presented with 1 lesion but later had a cluster (MPP-cluster later). The authors found that 1) clustering occurred relatively late. The mean ages at first presentation with 1 BCC of the SPP-more (61.5 years) and MPP-cluster later patients (60.4 years) were similar although presentations with clusters in the MPP-cluster initial (67.6 years, P = 0.0002) and -cluster later cases (68.1 years, P = 0.002) occurred significantly later. 2) Clustering was associated with increased accrual. Thus, 26 patients (MPP-cluster later/a) in the MPP-cluster later group had a additional BCC postcluster. Mean accrual post-cluster (1.99 BCC/year) in these cases was significantly increased (P = 0.0001) compared with precluster accrual (0.39 BCC/year).

CONCLUSIONS

The authors found that the formation of BCC clusters represents a critical event such that after a cluster presentation, tumor accrual is significantly increased. Cluster presentation is a relatively late event suggesting reduced effectiveness in immune surveillance.

Preliminary evidence of an association of tumour necrosis factor microsatellites with increased risk of multiple basal cell carcinomas

Tumour necrosis factor alpha (TNF-alpha) appears important in ultraviolet-induced immunosuppression, suggesting that it is a susceptibility candidate for cutaneous basal cell carcinoma (BCC). We now describe data on the association between TNF microsatellite polymorphisms, first on susceptibility in 202 controls and 133 cases each having two to 30 BCCs, and secondly, within the cases, on BCC numbers. The data show that the proportions of individuals with TNF a1- and a7-containing genotypes were significantly different (P = 0. 0271, P = 0.0393, respectively) between cases and controls. Secondly, within the cases, TNF alleles d4 (P = 0.023) and d6 (P = 0.006) alone, and the TNF a2-b4-d5 haplotype (P = 0.007), were significantly associated with the number of BCC lesions. These preliminary data provide the first evidence that TNF microsatellite polymorphism may influence the pathogenesis of multiple BCC.

Glutathione S-transferase polymorphisms: influence on susceptibility to cancer

The glutathione S-transferase supergene family includes several loci that demonstrate well characterised polymorphisms. The apparently critical role of these enzymes in cellular protection from the cytotoxic and mutagenic effects of electrophiles suggest that alleles associated with impaired detoxification will confer an increased susceptibility to a wide range of diseases. This hypothesis has been examined in case control studies and while data in some diseases such as lung cancer are conflicting, an increasing body of evidence suggests the importance of several glutathione S-transferase polymorphisms. In particular, GST genotypes have been associated with an increased susceptibility or worse outcome in diseases associated with oxidative stress. For example, both GSTM1 and GSTT1 genotypes are associated with susceptibility and outcome in cutaneous basal cell carcinoma. It still remains unclear however, why particular glutathione S-transferase loci are associated with altered risk in some diseases but not others. Further, the true in vivo substrates of these enzymes is unknown, consequently their mechanism of action remains unclear.

Polymorphism in glutathione S-transferase loci as a risk factor for common cancers

Though a developing body of data indicates polymorphism at GST genes influences cancer susceptibility, it is unclear why a genotype is associated with one cancer but not another. We believe the GST exert a critical role in normal cell house-keeping activities. GSTM1, GSTM3 and GSTT1 influence tumorigenesis because these enzymes utilise the products of UV-induced oxidative stress. Further support for the importance of these genes in the protection of skin from UV comes from studies in systemic lupus erythematosus (Ollier et al, 1996). Thus, GSTM1 null is associated with increased anti-Ro (but not anti-La) antibodies, a phenotype associated with photosensitivity. At present there is no basis for predicting which cancers will be influenced by GST polymorphisms though other studies do indicate that the GSTs are critical in the metabolism of environmental carcinogens. For example, GSTT1 null confers an increased risk of astrocytoma (Hand et al, 1996). While brain tumours are not clearly associated with environmental pollutants, N-methyl-N-nitrosourea, processed meats and occupation have been implicated. Why GSTT1 but not GSTM1 or GSTM3 influences the risk of astrocytoma is unclear. GSTM3 appears a good susceptibility candidate, as some astrocytes demonstrate strong expression (Hand et al, 1996). Susceptibility to squamous cell cancer of the larynx, a pathology associated with chronic consumption of tobacco and alcohol, is also influenced by allelism at GSTM3 (Jahnke et al, 1996). The roles of CYP2D6 and CYP1A1 are even more unclear, though the finding that systemic agents such as arsenic predispose to multiple BCC, suggests that CYP2D6-mediated hepatic detoxification of photosensitizing agents may be important. Importantly, the extent of altered risk conferred by genotypes is generally 2-3 fold and it is necessary to identify which other genes interact with the GST so that haplotypes associated with 10-20 fold increases in risk can be defined.

Genetic variation in low-dose UV-induced suppression of contact hypersensitivity and in the skin photocarcinogenesis response

Two of the major cutaneous consequences of ultraviolet (UV) radiation exposure are immunosuppression and the development of skin cancer. This study examined whether these effects are genetically determined. Suppression of contact hypersensitivity by local, low-dose UV radiation was examined in what have been termed "UV-susceptible" and "UV-resistant" strains of mice. C3H/HeJ mice ("UV resistant") were resistant to the adverse effects of low-dose UV radiation when normal doses of hapten were applied to UV-irradiated skin; however, they were sensitive when the amount of hapten used for sensitization was reduced. A similar effect was observed in BALB/c mice ("UV resistant") and when the hapten was dimethylbenz(a)anthracene, thus indicating that the genetic variation was not strain or hapten specific. Despite the fact that some strains were sensitive and some were resistant to low-dose UV radiation when high doses of hapten were employed, all strains initially sensitized to hapten through UV-irradiated skin were found to be unresponsive when rechallenged on normal skin, no matter what the initial sensitizing dose of hapten was. To determine whether other biologic effects of UV also exhibited genetic variation, C3H/HeN and C3H/HeJ mice were compared for susceptibility to UVB-induced skin cancer formation. C3H/HeJ mice developed significantly more tumors than C3H/HeN mice when subjected to a single dose of UV radiation followed by repeated exposure to the tumor promoter 12-O-tetradecanoyl-phorbol-13-acetate. These studies provide strong evidence that genetic factors influence individual susceptibility to the biologic effects of UV radiation.

Truncal tumor site is associated with high risk of multiple basal cell carcinoma and is influenced by glutathione S-transferase, GSTT1, and cytochrome P450, CYP1A1 genotypes, and their interaction

Basal cell carcinoma (BCC) places increasing burdens on clinicians; incidence is rising and patients may develop multiple primary tumors. Although UV exposure is critical, many patients develop tumors at less-exposed sites, such as the trunk, suggesting a genetic predisposition. We previously showed that polymorphism in loci encoding the detoxifying enzymes, glutathione S-transferase (GSTM1, GSTM3, GSTT1) and cytochrome P450 (CYP2D6, CYP1A1) influences susceptibility to BCC. We now describe a case-control approach in 345 patients with BCC that examines the role of these polymorphisms and patient characteristics (age, gender, skin type, hair color, eye color, smoking, occupation) in determining susceptibility to truncal tumors. GST and CYP genotypes were identified using polymerase chain reaction-based methods. Patients with one or more truncal tumors were significantly younger (p = 0.0170) than those with no truncal tumors. Male gender also appeared more common in the truncal tumor group, although this did not achieve significance (p = 0.0925). Patients whose first tumor was truncal had significantly more tumors (p = 0.0297). GSTT1 null (p = 0.0245, odds ratio 2.24) and CYP1A1 Ile/Ile (p = 0.0386, odds ratio 2.86) were associated with truncal site after correction for age and gender. The combination, GSTT1 null and CYP1A1 Ile/Ile, was particularly significant (p = 0.0059, odds ratio = 2.95). These effects were present after correction for tumor numbers. These data show first, patients with truncal tumors constitute a high-risk group for BCC, second, a significant genetic influence on BCC site, and third, a significant interaction between GSTT1 and CYP1A1 genotypes.