Variants of the melanocyte-stimulating hormone receptor gene are associated with red hair and fair skin in humans

Melanin pigmentation protects the skin from the damaging effects of ultraviolet radiation (UVR). There are two types of melanin, the red phaeomelanin and the black eumelanin, both of which are present in human skin. Eumelanin is photoprotective whereas phaeomelanin, because of its potential to generate free radicals in response to UVR, may contribute to UV-induced skin damage. Individuals with red hair have a predominance of phaeomelain in hair and skin and/or a reduced ability to produce eumelanin, which may explain why they fail to tan and are at risk from UVR. In mammals the relative proportions of phaeomelanin and eumelanin are regulated by melanocyte stimulating hormone (MSH), which acts via its receptor (MC1R), on melanocytes, to increase the synthesis of eumelanin and the product of the agouti locus which antagonises this action. In mice, mutations at either the MC1R gene or agouti affect the pattern of melanogenesis resulting in changes in coat colour. We now report the presence of MC1R gene sequence variants in humans. These were found in over 80% of individuals with red hair and/or fair skin that tans poorly but in fewer than 20% of individuals with brown or black hair and in less than 4% of those who showed a good tanning response. Our findings suggest that in humans, as in other mammals, the MC1R is a control point in the regulation of pigmentation phenotype and, more importantly, that variations in this protein are associated with a poor tanning response.

Prognostic value of Ki67 antigen expression in basal cell carcinomas

Recurrence of basal cell carcinoma (BCC) following treatment is a common event and long-term follow-up of all patients presenting with a primary BCC has been recommended. Proliferation indices have been recognized as important prognostic factors in several tumour types in a variety of cancer systems, being significantly elevated in more aggressive lesions. We have examined 51 BCCs (17 non-recurrent tumours [group 1], 17 original tumours which later recurred [group 2-O], and the corresponding 17 recurrent specimens [group 2-R]) for Ki67 antigen expression, a proliferation-associated antigen using immunohistochemistry with the monoclonal antibody MIB1. There was a significant increase in the percentage positive for MIB1 in the Group 2-O as compared with the group 1 BCCs (P < 0.05). p53 protein expression, as assessed by immunohistochemistry with the monoclonal antibody DO7, was similar in each group. These results show that Ki67 antigen expression differs between BCCs which later recur and BCCs that do not recur.

Up-regulation of p21WAF1/CIP1 in psoriasis and after the application of irritants and tape stripping

p21WAF1/CIP1 is a nucleoprotein that was initially characterized by its ability to be regulated transcriptionally by p53 and by its ability to mediate growth arrest by binding to cyclin-dependent kinases. Although p21WAF1/CIP1 is thought to mediate the effects of p53 in causing growth arrest, p21WAF1/CIP1 is also regulated in a p53-independent manner, e.g., during terminal differentiation of some cell lines. Growth factors including epidermal growth factor also induce p21WAF1/CIP1 through p53-independent pathways. Because the epidermal growth factor signaling pathway is abnormal in psoriatic epidermis, we studied p21WAF1/CIP1 expression, using in situ hybridization and immunohistochemistry, in psoriasis. Both p21WAF1/CIP1 mRNA and protein were significantly elevated in untreated psoriatic plaques compared with uninvolved psoriatic skin (p < 0.0001), with the up-regulation of p21WAF1/CIP1 being predominantly suprabasal. This increase was accompanied by a small increase in p53 protein expression of uncertain significance. Furthermore, p21WAF1/CIP1 expression was induced in skin after sellotape stripping and by the application of agents, such as dithranol, that are capable of inducing hyperproliferation. The pattern of p21WAF1/CIP1 expression observed is consistent with a role in induction and maintenance of differentiation. Our experiments, however, cannot determine whether the abnormalities of p21WAF1/CIP1 epidermal expression in psoriasis and after insult are independent of changes in p53 expression.

Topological control of p21WAF1/CIP1 expression in normal and neoplastic tissues

The p53-regulated gene product p21WAF1/CIP1 is the prototype of a family of small proteins that negatively regulate the cell cycle. To learn more about p21WAF1/CIP1 regulation in vivo, monoclonal antibodies were developed for immunohistochemistry. These revealed that p21WAF1/CIP1 expression followed radiation-induced DNA damage in human skin in a pattern consistent with its regulation by p53. A detailed comparison of the human, rat, and mouse p21WAF1/CIP1 promoter sequences revealed that this induction was probably mediated by conserved p53-binding sites upstream of the transcription start site. In unirradiated tissues, p21WAF1/CIP1 expression was apparently independent of p53 and was observed in a variety of cell types. Moreover, there was a striking compartmentalization of p21WAF1/CIP1 expression throughout the gastrointestinal tract that correlated with proliferation rather than differentiation. As epithelial cells migrated up the crypts, the Ki67-expressing proliferating compartment near the crypt base ended abruptly, with the coincident appearance of a nonproliferating compartment expressing p21WAF1/CIP1. In colonic neoplasms, this distinct compartmentalization was largely abrogated. Cell cycle inhibitors are thus subject to precise topological control, and escape from this regulation may be a critical feature of neoplastic transformation.

Use of in situ detection of histone mRNA in the assessment of epidermal proliferation: comparison with the Ki67 antigen and BrdU incorporation

The labelling index is commonly used as a measure of proliferation. However, the use of tritiated thymidine or BrdU labelling of S-phase cells is limited to prospective samples. We have employed an oligonucleotide cocktail complementary to the mRNA species encoding the replication-dependent histones H2B, H3 and H4 for non-isotopic in situ hybridization (NISH), and have compared the resultant proliferation indices in normal skin with those obtained by bromodeoxyuridine (BrdU) incorporation and by Ki67 immunohistochemistry (IHC) using the monoclonal antibody MIB1. In addition, we compared the staining characteristics of histone NISH and Ki67 IHC in a further 25 samples from a variety of inflammatory dermatoses and neoplastic conditions, as well as from normal skin. In normal skin, S-phase (histone NISH and BrdU) and cycling (Ki67) cells were confined to the basal and low suprabasal layers. The labelling indices determined by histone NISH and BrdU incorporation were similar, whereas that of Ki67 IHC was four times greater. In biopsies from hyperproliferative dermatoses and dysplastic or malignant lesions, the number of histone NISH- and Ki67 IHC-positive cells was generally elevated; in accordance with the differential expression of these two markers during the cell cycle, MIB1 consistently gave higher results. The advantage of histone NISH over Ki67 IHC is that it is a marker of the same part of the cell cycle as BrdU incorporation. However, the combined use of both histone NISH and Ki67 IHC to measure two cell cycle parameters, namely S-phase and the number of cycling cells, allows more detailed retrospective study of epidermal proliferation than has been possible previously.

Microsatellite instability in human non-melanoma and melanoma skin cancer

Microsatellite instability secondary to replication errors (RER), characterized by length changes at repetitive loci scattered throughout the genome, is a recently recognized genetic mechanism important in the development of some human cancers. Although RER has been reported in sebaceous gland tumors from patients with the Muir-Torre syndrome, the frequency of RER in human non-melanoma and melanoma skin cancers is not known. In this study, we investigated the importance of RER in human skin carcinogenesis. RER was identified in three of four actinic keratoses from a patient belonging to a kindred with documented Muir-Torre syndrome, which indicates that defective DNA replication may contribute to skin cancer development in such patients. Examination of a series of tumors from patients without Muir-Torre, including 137 skin cancers (47 basal cell carcinomas, 49 squamous cell carcinomas, and 41 primary malignant melanomas), 19 actinic keratoses, and 20 cases of Bowen's disease, using 10 or more microsatellite markers, identified repeat-sequence instability in less than 5% of the tumors studied. In six of the eight tumors, the sole change was an alteration 2 base pairs in length at a single locus. One patient with a squamous cell carcinoma showed changes at multiple loci suggesting defective mismatch repair. Although the low frequency of RER found in this study of a large series of human skin tumors suggests that this phenomenon is uncommon in patients with skin cancer, the identification of RER at multiple loci in two patients suggests that error-prone replication may be important in skin cancer development in some individuals.

Normal sweat secretion rate in patients with alopecia areata

We have examined sweat secretion rates in 22 patients with alopecia areata, and 22 age- and sex-matched controls. Mean sweat rate on the forearm in patients with alopecia areata was 20 mg/cm2 per h (95% confidence limits 15-25 mg/cm2 per h), and in controls was 24.1 mg/cm2 per h (95% confidence limits 19.1-29.1 mg/cm2 per h). Sweat secretion was higher in males than females in both the disease and control groups (27.8 mg/cm2 per h [95% confidence limits 21.3-34.3 mg/cm2 per h], compared with 18.08 mg/cm2 per h [95% confidence limits 14.63-21.6 mg/cm2]; P > 0.01). Our results confirm the previously reported sex difference in sweat secretion rate, and demonstrate that there is no statistically significant difference between patients with alopecia areata and controls. We discuss our results in the light of a previous report claiming that patients with alopecia areata have reduced rates of cholinergic-induced sweating.

Reduced experimental contact sensitivity in squamous cell but not basal cell carcinomas of skin

Basal cell carcinomas (BCC) and squamous cell carcinomas (SCC) show clinical and epidemiological differences not accounted for by different ultraviolet radiation exposure. We have studied experimentally induced contact sensitivity to dinitrochlorobenzene by measuring increases in skin-fold thickness. Patients (n = 37) with squamous tumours had impaired responses compared with controls (33) and patients with BCCs (31) (mean increase 4.5 vs 7.8 and 8.6 mm, respectively; p = 0.002). This diminished immunological response may be causally related to the development of SCC. Because glutathione S-transferase (GST) metabolises dinitrochlorobenzene and polymorphisms of GST are associated with multiple skin tumours, variations in GST may underlie these differences.

Loss of heterozygosity in sporadic primary cutaneous melanoma

Difficulties in obtaining clinical samples from primary melanomas have meant that most genetic analyses of melanoma have concentrated on cell lines and metastases. Because the Breslow thickness of the primary tumour is the single best prognostic indicator, it is important to identify genetic abnormalities in primary melanomas and relate these changes to the thickness of the lesion. We have investigated 47 sporadic melanomas, of which 41 were primary lesions, for loss of heterozygosity (LOH) on several chromosomal arms, including areas where genes involved in familial melanoma and other relevant hereditary syndromes map, and where LOH has previously been reported in cell lines, or metastatic lesions. LOH was identified at 66 (18%) of 358 informative loci in primary melanomas, and there was a significant relationship between the overall frequency of LOH and Breslow thickness (P < 0.0005). Loss of chromosome arm 9p was most frequent, occurring in 15 (47%) of 32 informative primary tumours, and was observed in 3 of 11 informative lesions < or = 1.5 mm in depth. LOH on chromosome arms 3p, 6q, 10q, 11q, and 17p was also relatively frequent, with loss of 3p and 10q heterozygosity in lesions < or = 1.5 mm in depth, while LOH on 6q, 11q, and 17p was only detected in more invasive tumours. The results suggest that loss of these chromosome regions are important in sporadic cutaneous melanoma, and are consistent with chromosome arm 9p loss occurring before loss of other chromosome arms.

Changes in mean telomere length in basal cell carcinomas of the skin

Human telomeres consist of arrays of the sequence TTAGGG up to 15-20 kb in length, which are essential for the maintenance of normal chromosomal stability. It has been suggested that genomic instability observed in tumours may be due to loss of telomere sequences. Somatic cells that are dividing continuously appear to progressively lose telomere sequences, and it would therefore be anticipated that cell type specific differences in mean telomere length may exist within an individual. Previous reports have suggested that mean telomere length may be different in human neoplasia when compared to control. Basal cell carcinomas are epidermal derived tumours and in order therefore to make valid cell type specific comparisons we have measured mean telomere length in 20 basal cell carcinomas as well as in both adjacent epidermis and dermis. Mean telomere length was significantly reduced in epidermis in comparison with dermis, from clinically normal skin immediately adjacent to the tumours (mean difference 2.5 kb). This result is not related to the presence of the tumour as similar results were obtained from skin samples of healthy volunteers. Basal cell carcinomas showed increased mean telomere length in 13/20 samples in comparison with matched epidermis (mean difference 3.1 kb), whereas in 7/20 mean telomere length was reduced (mean difference 2.2 kb). These results showing that mean telomere length varies from cell type to cell type underpin the importance of performing cell type specific controls when assessing changes in tumour telomeres.

Delineation of two distinct deleted regions on chromosome 9 in human non-melanoma skin cancers

The mapping of the naevoid basal cell carcinoma syndrome (NBCCS) and the Ferguson-Smith syndrome to the same region on chromosome arm 9q has led to speculation that the two conditions may reflect different mutations within the same gene. Loss of heterozygosity of 9q alleles in both familial and sporadic basal cell carcinomas (BCCs) suggests that the NBCCS gene on 9q is acting as a tumour suppressor gene. Although LOH of 9q markers has not been studied in squamous cell neoplasms from patients with the Ferguson-Smith syndrome, chromosome 9 allele loss has been reported in sporadic squamous cell carcinomas (SCCs) of the skin. In order to characterise further the deleted region on chromosome 9 in BCCs and SCCs of the skin we have examined a series of non-melanoma skin cancers using a panel of highly informative microsatellite markers. Forty-four BCCs and 49 SCCs were studied. Loss of heterozygosity of one or more 9q markers was seen in 33 of the 44 BCCs. Only 4 of the 33 BCCs with 9q loss showed loss of 9p markers. Twenty-two BCCs showed loss of all informative 9q markers. Partial or interstitial 9q deletions were seen in 5 BCCs, and in 3 of these 5 BCCs the breakpoint occurred within the currently defined NBCCS locus. Chromosome 9 loss was seen in 16 of 49 SCCs. In contrast to the low frequency of 9p loss in BCCs, LOH of 9p markers was a common finding in SCCs, occurring in 15 of the 16 SCCs with chromosome 9 loss. In 5 SCCs 9p loss occurred with retention of 9q alleles.(ABSTRACT TRUNCATED AT 250 WORDS)

Linkage analyses in British pedigrees suggest a single locus for Darier disease and narrow the location to the interval between D12S105 and D12S129

Darier disease is a dominantly inherited skin disorder in which there appears to be abnormal adhesion between keratinocytes. We and others have shown that the disease in some British pedigrees is closely linked to markers mapping to 12q23-q24.1. In the present study we have defined crossovers that enable us to narrow the location of the disease gene to the interval between the D12S105 and the D12S129 markers. This interval may be expected to be on the order of about 4 cM on the basis of linkage data obtained using the primary CEPH reference families. Our data provide further evidence for locus homogeneity: each of four large British pedigrees, two of which have previously been subjected to preliminary characterization, shows statistically significant evidence for linkage to markers mapping to 12q23-q24.1.

Dissociation of erythema and p53 protein expression in human skin following UVB irradiation, and induction of p53 protein and mRNA following application of skin irritants

The mechanisms mediating the varied effects of ultraviolet radiation (UVR) on human skin are unclear, although a relationship between erythema and DNA damage is suggested by photosensitivity in xeroderma pigmentosum. Increased p53 expression in response to UVR is thought to reflect direct DNA damage, but recent evidence indicates that UVR also activates membrane and cytosolic signal transduction pathways. In this study, we have investigated the relationship between erythema and p53 induction following UVB and whether this p53 response is specific to UVR. p53 protein expression was determined by immunocytochemistry using the monoclonal antibody DO7, and p53 mRNA expression was examined by non-isotopic in situ hybridization. Incremental doses of UVB were administered to the lower back of eight subjects. Immunostaining revealed that p53 positive nuclei were significantly increased 8 h after suberythemogenic doses of UVB (79 +/- 12), compared to normal unirradiated skin (8 +/- 6, p < 0.0005), but no change in p53 mRNA was seen. Higher UVB doses, which resulted in moderate erythema, resulted in a similar or greater induction of p53 protein. Indomethacin (1% w/v), applied immediately after UVB irradiation, significantly inhibited UVB erythema at 8 h in six subjects (p < 0.005), but did not reduce p53 immunostaining. Dithranol (1 microgram/microliter, n = 8), sodium dodecylsulphate (5%, n = 4), and retinoic acid (0.5%, n = 4), applied for 48 h, caused erythema, significantly increased p53 protein levels (p < 0.05), and also increased p53 mRNA. Our results show that in human skin, UVB-induced p53 elevation can be dissociated from erythema and skin irritants can also induce p53 protein. The induction of p53 mRNA by irritants but not UVR suggests different mechanisms of action.

High frequency of loss of heterozygosity in actinic keratoses, a usually benign disease

Actinic keratoses (AKs) are focal areas of dysplasia with low risk of progression to squamous cell cancer; many regress spontaneously. Using polymerase-chain-reaction microsatellite analysis, we found that loss of heterozygosity on several chromosome arms, including 17p, 17q, 9p, 9q, and 13q, was common in AKs. More than half the AKs examined showed loss of heterozygosity at four or more loci. The apparent genetic instability of these lesions contrasts with their benign clinical course.

A gene for pachyonychia congenita is closely linked to the keratin gene cluster on 17q12-q21

Pachyonychia congenita (PC) is a group of hereditary syndromes which have in common a hypertrophic dystrophy of the distal nail, and are associated with a variety of additional features, notably various dyskeratoses of skin and mucous membranes. The pathology is unknown but the array of clinical features suggests the possibility of a keratin abnormality. In the present report we describe linkage analyses in a large PC pedigree of the Jackson-Lawler type, a subtype which is characterised by multiple epidermal cysts, hair abnormalities, and natal teeth. The disease locus in this family was found to be tightly linked to markers mapping within, or very close to, the keratin type I cluster at 17q12-q21; maximum lod scores for linkage of the disease to a KRT10 polymorphism and to D17S800, a marker known to be very tightly linked to KRT10, were respectively +4.51 and +7.73, both at theta = 0.00. Although always likely, our findings provide strong evidence of a keratin gene anomaly underlying an inherited disorder affecting epidermis, nail, hair, and mucosa. These findings permit testing to see if pachyonychia congenita shows any locus heterogeneity and suggest specific candidate keratin genes for mutation searching studies. In addition, they suggest a role for keratins in the phenomenon of natal dentition.

Basal cell carcinomas and squamous cell carcinomas of human skin show distinct patterns of chromosome loss

Although basal cell carcinomas and squamous cell carcinomas are clinically and pathologically distinct, the molecular basis for these differences is not clear. We have used polymorphic microsatellite markers to determine the pattern and extent of chromosome losses in a series of 44 basal cell carcinomas and 47 squamous cell neoplasms of the skin. Basal cell carcinomas showed a distinctive allelotype with chromosome loss largely confined to a single chromosome arm, 9q (26 of 44 informative tumors). In contrast to the predominance of loss on a single chromosome arm in basal cell carcinomas, squamous cell neoplasms showed more widespread loss with loss of heterozygosity of markers from 35 of 39 chromosome arms in one or more of the tumors studied. The pattern of loss was also different from basal cell carcinomas with frequent loss of heterozygosity of markers from 9p (41%), 13q (46%), 17p (33%), 17q (33%), and 3p (23%) in squamous cell neoplasms. The frequent loss of markers from these chromosome arms relative to other chromosome losses suggests that these arms may contain genes important in the development of cutaneous squamous cell carcinomas.

Wavelength-specific upregulation of keratin mRNA expression in response to ultraviolet radiation

Keratin intermediate filaments are heteropolymers of coexpressed type I and type II protein chains, whose expression is tightly linked to the differentiation status of the keratinocyte. Epidermal basal keratinocytes coexpress keratins K5 and K14, whereas suprabasal keratinocytes downregulate K5 and K14 and begin to coexpress keratins K1 and K10. Using both isotopic and non-isotopic in situ hybridization, we have investigated the changes in expression of the messenger RNA species encoding the K5/K14 and K1/K10 keratin pairs in response to ultraviolet radiation. Here we report that following irradiation, the mRNA species encoding both keratin pairs is upregulated in a wavelength-specific manner, and that the link between the pattern of keratin mRNA expression and the differentiation status of the keratinocyte is disrupted. Forty-eight hours following ultraviolet B exposure, the amount of detectable mRNA encoding all four keratins studied had increased. Following UVA irradiation, the K1 and K10 signal increased to a much lesser extent than following ultraviolet B, whereas no change in the amount of mRNA encoding the K5/K14 pair was observed. Only two samples were examined following ultraviolet C exposure, but in both, increased K5/K14 signal, but not suprabasal K1/K10 signal, was observed. We suggest that the observations reported here may reflect important qualitative changes involved in photoadaptation of the epidermis, and provide further molecular markers of the different biological effects of ultraviolet radiation of different wavelengths.

Chromosome 9 allele loss occurs in both basal and squamous cell carcinomas of the skin

Linkage studies of kindreds with the nevoid basal cell carcinoma syndrome and the high frequency of chromosome 9 allele loss in sporadic basal cell carcinomas indicate that chromosome 9 may contain tumor suppressor genes important in the development of sporadic and familial basal cell carcinomas. The recent mapping of the Ferguson-Smith syndrome, which predisposes affected individuals to the development of multiple lesions histologically indistinguishable from squamous cell carcinomas, suggests that tumor suppressor genes on 9q may also be important in the development of squamous cell neoplasms of the skin. Fifty-four non-melanoma skin cancers (24 basal cell carcinomas, 14 squamous cell carcinomas, and 16 cases of Bowen's disease) were examined for loss of heterozygosity on chromosome 9. Allelic loss at one or more loci on chromosome 9 was observed in 14 of 24 basal cell carcinomas, four of 14 squamous cell carcinomas, and three of 16 cases of Bowen's disease. Allelic deletion of one or more 9q markers was seen in 14 basal cell carcinomas, three squamous cell carcinomas, and three cases of Bowen's disease. Five basal cell carcinomas had interstitial deletions and in one the breakpoint mapped within the nevoid basal cell carcinoma syndrome locus. 9p loss occurred in three of nine informative squamous cell carcinomas. Allelic deletion of 9p markers was not seen in 19 basal cell carcinomas and seven cases of Bowen's disease. These findings suggest that chromosome 9 contains one or more tumor suppressor genes important in the development of both basal and squamous cell carcinomas of the skin.

The relation between p53 mutation and p53 immunostaining in non-melanoma skin cancer

Extensive study of the p53 gene has established its role as a tumour-suppressor gene, and the involvement of mutant p53 in a wide spectrum of human malignancy. Many mutations of p53 result in a protein product that is abnormally stable, so that it becomes readily detectable by immunocytochemistry. In contrast, under normal conditions, it has been considered that levels of wild-type p53 were too low to be detectable. Although positive immunocytochemistry has been used as a marker of mutation, recent evidence suggests that this assumption may not always be valid. We have carried out both PCR-sequencing of exons 5-8 of the p53 gene in 20 basal cell carcinomas (BCC), and immunocytochemistry of these tumours with the anti-p53 antibody DO7. Twenty cases of Bowen's disease, in which we had previously documented mutations, were also immunostained. We report a low rate of p53 mutation in the BCCs we examined (2/20), and a discrepancy between tumours with positive immunostaining and those with mutation in both Bowen's disease and BCC. Of eight tumours in which we detected mutation, only four were immunopositive: of 19 immunopositive samples, only four showed detectable mutation. We discuss the implications of our results for the use of positive immunostaining in clinical diagnosis, and the involvement of p53 in skin carcinogenesis.

Wavelength specific patterns of p53 induction in human skin following exposure to UV radiation

We report that, in human skin, exposure to equally erythemogenic doses of UVA, UVB, and UVC increases immunocytochemically detected p53 in a wavelength-specific pattern. UVC produced immunostaining confined to the upper epidermis. With UVB, staining was seen throughout the epidermis, whereas with UVA staining predominated in the basal layer. The results with UVB and UVC are understandable on the basis of their known differences in penetration, whereas those with UVA are not. This suggests that within one cell type the pattern of p53 response to UV radiation is wavelength dependent.

p53 mutations are common and early events that precede tumor invasion in squamous cell neoplasia of the skin

Mutations of the p53 gene are the most common genetic abnormality described in human cancer; p53 mutations have recently been reported in more than half of the cases of squamous cell carcinomas of the skin. We have previously reported positive p53 immunostaining in Bowen's disease and actinic keratosis. To determine if this abnormal immunostaining reflects p53 mutation or alternative pathways of p53 protein inactivation we have performed direct sequencing of p53 in 20 further cases of Bowen's disease. We found eight mutations in 20 cases, seven of which would produce alterations in the p53 protein product. Our results suggest that p53 mutation is an early event in malignant conversion, frequently preceding invasion in squamous cell neoplasia of the skin. The type and site of the observed mutations reflect known mutational hotspots and support the role of ultraviolet radiation in the pathogenesis of these tumors.

p53 protein expression in benign and malignant skin tumours

The skin affords an excellent model of human carcinogenesis because a variety of lesions from benign tumours to invasive malignancy, with or without metastatic potential, are commonly found, and are accessible to biopsy. To date, few genetic alterations have been observed in skin neoplasia. In this study we have used a recently developed monoclonal antibody (DO7) to examine p53 protein expression in a wide variety of benign and malignant skin lesions. Benign skin lesions were negative, but a significant number of malignant epithelial lesions showed detectable p53; 56% of squamous carcinomas and 42% of basal cell carcinomas were positive. A smaller proportion of dysplastic epithelial lesions were positive (27%), and only 3.6% of malignant melanomas were positive. Thus, although detectable p53 protein is a common occurrence in malignant epithelial lesions, it does not correlate with the malignant phenotype or with metastatic potential. The finding of a lower proportion of positivity in dysplastic lesions, and absence of staining in benign tumours, suggests that p53 mutation may be involved in the progression towards invasive malignancy in human squamous skin lesions.

Codon 12 Harvey-ras mutations are rare events in non-melanoma human skin cancer

ras mutations have been reported as an early event in some human malignancies and in the mouse skin model of multistep carcinogenesis; early studies in human non-melanoma skin cancers have reported variable rates of ras mutations. A recent study, however, has reported a high frequency of activating mutations of the Harvey-ras proto-oncogene in non-melanoma skin cancers, and the site specificity of the mutation at the second position of codon 12 prompted us to re-examine the importance of Ha-ras codon 12 mutations as an early event in the development of these tumours, using a combination of PCR and restriction fragment polymorphism of codon 12 of the Ha-ras gene. Dilution experiments confirmed that the method was sensitive and capable of detecting mutations at this codon when only 4% of the total alleles are mutated. We were surprised to find no mutations in the 40 basal cell carcinomas, 12 squamous cell carcinomas and 12 cases of Bowen's disease studied. We conclude that Ha-ras codon 12 mutations are rare events in human non-melanoma skin cancer in the U.K. The marked differences in the frequency of codon 12 Ha-ras mutations in published studies may relate to either technical artefacts, or differences in the molecular epidemiology between areas of low and high sun exposure.