p53 mutations implicate sunlight in post-transplant skin ... - Nature

Oncogene (1997) 15, 1737 ± 1740  1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00

SHORT REPORT

p53 mutations implicate sunlight in post-transplant skin cancer irrespective of human papillomarivus status Jane M McGregor1, Ron JM Berkhout2, Magdalena Rozycka3, Jan ter Schegget2, Jan N Bouwes Bavinck4, Louise Brooks5 and Tim Crook3 1

Department of Photobiology, St John's Insitute of Dermatology, St. Thomas' Hospital, Lambeth Palace Road, London SE1 7EH, UK; 2Department of Virology, Academic Medical Centre, University of Amsterdam and 4Department of Dermatology, University Hospital Leiden, The Netherlands; 3Section of Cell Biology and Experimental Pathology, Institute of Cancer Research, Haddow Laboratories, Sutton, Surrey SM2 5NG, UK; 5Department of Clinical Sciences, London School of Hygiene and Tropical Medicine, London WC1, UK

Mutations in p53 were detected in 11/23 (48%) of non melanoma skin cancers in renal allograft recipients and in 5/8 (63%) of sporadic tumours from immune competent patients. 9/12 (75%) of mutations in transplant patients and all 5 mutations in non transplant tumours were consistent with damage caused by ultraviolet (u.v.) irradiation. DNA sequences, predominantly of the epidermodysplasia verruciformis (EV) subgroup, were detected in 9/23 (39%) of transplant tumours and in 2/8 (25%) of eight non-transplant tumours. There was no relationship between HPV status and p53 mutation, HPV DNA being present in 5/16 (31%) of tumours with p53 mutation and 6/15 (40%) of tumours lacking p53 mutation. These data are consistent with an important role for sunlight in the development of post-transplant skin cancer, and with limited functional data suggesting that E6 proteins of the cutaneous and EV-related papillomaviruses do not target p53 for ubiquitin-mediated degradation. Keywords: non-melanoma skin; transplantation; p53 gene; human papillomavirus

Chronically immunosuppressed renal transplant recipients are at considerably higher risk of non-melanoma skin cancer (NMSC) than the general population. A number of European studies report a 27 ± 40% cumulative risk of developing either a basal cell (BCC) or squamous cell cancer (SCC) within the ®rst 20 years following transplantation (Hartevelte et al., 1990; London et al., 1995; McGregor and Proby, 1995). Several mechanisms have been suggested to account for this increase, including chemical carcinogenesis (Lennard et al., 1985; Swann et al., 1996), altered DNA repair (Kelly et al., 1987; Swann et al., 1996) and loss of immune surveillance (Doll and Kinlen, 19970). Human papillomaviruses (HPV) have also been implicated, since other tumours which occur at high frequency in transplant recipients are those associated with a viral aetiology (Morris et al., 1996), including lymphoma, Kaposi's sarcoma and anogential SCC (Birkeland et al., 1995).

Correspondence: T Crook Received 10 March 1997; revised 11 June 1997; accepted 11 June 1997

A viral aetiology for skin cancer was ®rst proposed in 1922 with the description of a rare inherited condition known as epidermodysplasia verruciformis (EV) (Lewandowsky and Lutz, 1922). Patients with EV develop widespread cutaneous wart infection and multiple cutaneous SCCs on sun-exposed sites. In this respect they resemble transplant patients. Analysis of EV-associated skin cancers has revealed that over 90% contain HPV types 5 or 8, an association which equals that of the so called `high risk' genital HPV types with anogenital SCC (IARC monograph, 1995). Thus an aetiological link between HPV types 5 and 8 and the development of skin cancer seems likely. With the recent availability of degenerate PCR primers designed to detect a wide range of mucosal, cutaneous and EVrelated types, several groups have now reported diverse, often multiple HPV sequences in transplant NMSC (McGregor and Proby, 1996). However, `high risk' genital types and types 5 and 8 are not often found in transplant NMSC, suggesting that neither anogenital nor EV-associated SCC represents a good model for cutaneous malignancy in this patient group. Whatever the role of HPV in post-transplant skin cancer, other factors are also likely to be involved, including most importantly, solar ultraviolet radiation (UVR). Epidemiological data have long implicated sun exposure as a risk factor for sporadic NMSC (Blum, 1948) and the recent demonstration of p53 mutations consistent with u.v. -induced damage in these tumours provides compelling molecular evidence to support this (Brash et al., 1991; Zeigler et al., 1993; Nataraj et al., 1995). Furthermore, the majority of skin cancers in immune suppressed transplant patients also occur on sun-exposed sites, but the role, if any of UVR in this group of patients has not been determined. In this study we have investigated the HPV and p53 status of a series of NMSC from renal allograft recipients, together with a smaller group of sporadic NMSC. HPV DNA sequences were sought in genomic DNA from 23 post-transplant and 8 non-transplant skin cancers, using nested PCR as described by Berkhout et al. (1995). The two degenerate primers, CP65 and CP70 are located in the L1 open reading frame and amplify a 455 ± 467 bp product dependent on the target HPV type. The second nested PCR reaction utilises the degenerate primers CP66 and CP69 and ampli®es a 377 ± 389 bp product. Using this PCR, all known EVassociated HPV genotypes can be ampli®ed. Sensitivity of the PCR, as evaluated for four dierent HPV types

p53 mutations in post-transplant skin cancer JM McGregor et al

1738

Table 1 Patient

(tumour no.)

Renal transplant tumours JC (1) (3) (4) (11) (23) (24) (25) (26) (2) (14) (15) (28) (29) (30) (31) KD (32) RH (10) RB (12) (8) MS (9) (7) LH (6) JH (5) DW Immunecompetent tumours ML (16) (17) MBra (19) MBro (20) AT (21) GA (22) LL (18) FR (27) JB RD AK SM JM

p53 gene mutations and HPV typing in NMSC from renal transplant and IC patients Codon

Amino acid change

± ± ± 5 5 4 ± 8 ± 7 ± ± ± 7 7 ± ± ± 5 4 ± 4 7 4

± ± ± 170 161 80 ± 295 ± 243 ± ± ± 245 248 ± ± ± 135 85 ± 104-5 250 117-8

± ± ± Thr Ala no change Pro Leu ± Pro Leu ± Met Arg ± ± ± Trp Val Arg Gln ± ± ± Cys Ter Pro Leu ± frameshift pro Leu frameshift

7 8 ± ± ± 6 5 5

236 287 ± ± ± 224 178 137

Tumour type

Site

HPV (type)

Mutation

Exon

SCC SCC SCC SCC

hand head head head

+ (novel) + (20) ± ±

SCC SCC SCC SCC SCC SCC SCC SCC SCC SCC SCC SCC SCC SCC SCC SCC SCC BCC BCC

head hand head hand head head head head head head hand hand ± ± ± ± head head back

± + (?) ± + (?) + (24-related) ± ± ± ± ± ± ± + (23) + (23) + (25) ± + (21) ± ±

± ± ± GA:GG* CC:CT* CC:CT* ± CC:CT* ± AT:AG ± ± ± GG:GT* GG:AG* ± ± ± CC:AC* CC:CT* ± GG:GGG CC:CT* GG:GGG

SCC SCC SCC SCC SCC SCC BCC BCC

head forearm head head head head neck head

+ (novel) ± ± ± + (?) ± ± ±

CT:CA* GA:AA* ± ± ± GA:TA* CC:CT* CC:AC*

Tyr Asn Glu Lys ± ± ± Glu Ter His Tyr Ala Asp

Tumour specimen no. 13 lost. *Mutations consistent with DNA damage caused by UVR (occuring at dipyrimidine sites or base alterations consistent with damage caused by u.v.-induced reactive oxygen species)

(8, 15, 20 and 23) ranges from 1 ± 10 fg (Berkhout et al. (1995). In addition, the genital HPV types 6, 11, 16, 18, 31 and 33, together with the cutaneous HPV 1, can also be detected, but other cutaneous HPV types either cannot readily be ampli®ed or can only be ampli®ed by primer set CP65/CP70 (types 2, 26 and 27). HPV DNA typing was performed by direct sequence analysis of PCR amplimers. HPV DNA sequences were detected in nine of 23 post-transplant tumours (39%) and in two of eight non-transplant tumours (25%) (Table 1). Known EV-associated HPV types 20, 21, 23 and 25 were found in post-transplant SCCs and putative novel HPV types were found in further tumours, amoung both transplant and non-transplant patients, but no tumour contained DNA from HPV 5/8 or genital HPV types. To detect p53 mutations genomic DNA was subjected to Single Strand Conformation Polymorphism (SSCP) analysis over codons 2 ± 11. Representative analyses are shown in Figure 1. Apparently abnormal DNAs were analysed by sequence analysis of multiple plasmid clones containing the exon with the proposed mutation. Twelve mutations were identi®ed in 11 of 23 transplant tumours (48%), comprising 2 BCCs and 9 SCCs (1 SCC contained two mutations). Nine of 12 (75%) mutations were consistent with DNA damage caused by solar UVR (Nataraj et al., 1995), consisting of base changes at dipyrimidine sites, predominantly C4T transitions. G4T mutations were also identi®ed, lesions which may result from the oxidation of guanine by singlet oxygen, a cellular product of u.v.-induced

reactive oxygen species (ROS) (Cheng et al., 1992; Piette, 1991). Frame shift mutations were identi®ed in two additional transplant tumours, both single G in reiterated sequences of G residues (Figure 2). All ®ve mutations in eight non-transplant tumours (2 BCC and 3 SCC) were consistent with u.v.-induced damage, consisting of base changes at dipyrimidine sites and a single G4T transversion. There was no relationship between the presence of HPV DNA sequences and p53 mutation. Overall, HPV DNA was detected in ®ve of 16 (31%) of tumours with p53 mutation, and in 6 of the remaining 15 (40%) tumours without p53 mutation. Despite a widely held view that post-transplant NMSC may be pathogenically distinct from those tumours occurring in the general population, accumulating data actually reveal many similarities. Risk factors for skin cancer in both transplant and nontransplant patients are similar (Bouwes Bavinck et al., 1993) and cutaneous, mucosal and EV-related HPV sequences are found in a high porportion of both BCC and SCC in both patient groups (McGregor and Proby, 1996). We have examined the spectrum of p53 mutations in post-transplant compared with sporadic skin cancers, and show that p53 mutations, predominantly of a u.v.-related nature, are present irrespective of HPV status. We also show that both the frequency and type of p53 mutations in transplant NMSC is similar to that observed in sporadic NMSC. The striking excess of NMSC in transplant patients has prompted the hypothesis that this may be explained by


1739

1

2

3

4

5

6

7

8

MT

9

G Figure 1 SSCP analysis of p53 in skin cancer. lanes 1 ± 3: exon 6. showing abnormal mobility band (arrowed) in lane 1. Lanes 4 ± 9: exon 7, showing abnormal mobility bands (arrowed) in lane 6. Genomic DNA was obtained from bipsy specimens of BCC (n=2) and SCC (n=21) from 13 renal transplant patients. These patients had received a renal allograft between 1965 and 1987, and skin tumours developed at least 2 years after transplantation in each case. Biopsy specimens BCC (n=2) and SCC (n=6) were obtained from 8 non-transplant patients. Genomic DNA was isolated from snap-frozen tumour tissue by proteinase K digestion and subjected to SSCP analysis of exons 2 ± 11 of P53 using the Human p53 Amplimer panel (Clontech, Palo Alto, USA) according to the manufacturer's instructions. Following PCR, reactions were boiled in formamide, then resolved on 5% nature polyacrylamide gels with or without 10% glycerol

an enhanced susceptibility to a potentially oncogenic virus (Morris et al., 1996). In addition, however, a considerable body of epidemiological evidence is consistent with an association between sun exposure and skin cancer in transplant recipients. The cumulative incidence of skin cancer following transplantation in sub-tropical regions of Australia is considerably higher than in patients transplanted and resident in Europe and furthermore, over 90% of transplant skin cancers occur on sun-exposed sites (Hartevelt et al., 1990; London et al., 1995; McGregor and Proby, 1995; Bouwes Bavinck et al., 1996). The data we present in this study provide molecular evidence consistent with a causal role for sun exposure in the development of post-transplant skin cancer. The majority of p53 mutations (9/12) were consistent with DNA damage caused by UVR, the majority located at positions likely to compromise p53 function i.e. within the highly conserved regions (II ± IV) of the sequence-speci®c DNA binding domain (Prives et al., 1994). The speci®city of u.v.-induced mutation re¯ects the absorption spectrum of DNA at UVB (280 ± 310 nm) wavelengths, with the majority of photo products involving bonds between adjacent pyrimidines on the same strand. Attempts at repair lead to base substitutions at the sites, predominantly C4T and more rarely CC4TT tandem base transitions. Such genetic alterations appear unique to u.v. and NMSC (Nataraj et al., 1995). Other mutations have occasionally been described in experimental systems with u.v., including G4T, G4C and G4A point mutations at non-dipyrimidine sites and are thought to result from

A

WT

T

C

G

A

T

C

Figure 2 DNA sequence analysis of exon 4 of p53. The mutant (MT) sequence is shown with an inserted G (arrowed). The wildtype (WT) sequence is also shown for comparison. Exon 4 was reampli®ed from genomic DNA, then ligated into pGEMT (Promega). The sequence of multiple plasmid clones was determined using T7 DNA polymerase (Pharmacia). Each mutation was veri®ed by sequencing of both DNA strands, and con®rmed in clones derived from at least two indpendent PCR reactions

oxidation of guanine residues by singlet oxidation and other u.v.-generated radicals (Piette, 1991; Cheng et al., 1992; Nataraj et al., 1995). This mechanism may account for the presence of mutations at nondipyrimidine sites reported in previous studies of sporadic BCC and SCC and also for the mutation present in tumour 22 in the present series. Of the three mutations in transplant tumours which were not typical of u.v., two were frame shift in which a single G residue was inserted into a sequence of reiterated guanine residues. The mechanisms underlying such mutations are less well understood than other types of mutation, but the great majority occur at monotonic runs of bases (Greenblatt et al., 1996), and it has been proposed that they represent errors in DNA replication (or possibly repair) occurring during cell division (Jego et al., 1993; Greenblatt et al., 1996). Such mutations are very rarely observed in sporadic NMSC and the signi®cance of these mutations in the context of post-translant skin cancer is not clear. However, recent suggestions that chronic exposure to azathioprine, and perhaps also to other immunosuppressive drugs used in transplantation, may contribute to carcinogenesis by selecting for clones defective in post-replicative DNA repair requires further study (Kelly et al., 1987; Swann et al., 1996). Our observation that p53 mutations are present irrespective of the HPV status of NMSC is in contrast to the situation observed in anogenital SCC where the concurrent presence of p53 mutation and expression of `high risk' genital type E6 proteins is uncommon (Crook et al., 1992). These data imply that the p53inactivating function attributable to the `high risk' genital types and thought to play a critical role in the development of anogenital SCC, is not important in the pathogenesis of cutaneous SCC, data which are consistent with mechanistic studies of EV-related types. E6 proteins of `high risk' genital HPV types


1740

exhibit transforming properties (Woodworth et al., 1989), and these activities are also present, albeit weakly, in EV-related viruses (Iftner et al., 1988; Kiyono et al., 1989). However, the ability of EV type HPV E6 proteins to directly abrogate p53 function is clearly dierent from the genital `high risk' viruses; whereas HPV 16-encoded E6 and E7 proteins bind eciently and therefore functionally inactivate p53 and RB respectively, HPV 8E7 has only weak ability to associate with pRb, and E6 lacks any detectable activity to either associate with or promote the proteolysis of p53 (Steger and P®ster, 1992). In this context, the requirement for p53 inactivation in the development of skin cancer must occur either by

mutation, as suggested by our data, or alteration of the function of other proteins elsewhere in the p53 pathway, or, perhaps, by virally-encoded functions independent of direct protein-protein interaction.

Acknowledgements These studies were supported by the Cancer Research Campaign. The authors would like to thank Dr DM MacDonald, Dr MH Rustin and the renal physicians at both Guy's Hospital and the Royal Free Hospital, London, for providing patients and for their consideration and support during this study.

References Berkhout JM, Tieben LM, Smits HL, Bouwes Bavinck JN, Vermeer BJ and ter Schegget J. (1995). J. Clin. Microbiol., 33, 690 ± 695. Birkeland SA, Storm HH and Lamm LU. (1995). Int. J. Cancer, 60, 183 ± 189. Blum HF. (1948). J. Natl. Inst. Cancer, 9, 247 ± 258. Bouwes Bavinck JN, de Boer A and Vermeer BJ. (1993). Br. J. Dermatol., 129, 242 ± 249. Brash DE, Rudolph JA, Simon JA, Lin A, McKenna GJ, Baden HP, Halperin AJ and Ponten J. (1991). Proc. Natl. Acad. Sci. USA, 88, 10124 ± 10128. Cheng KCDS, Cahill H and Kasai S. (1992). J. Biol. Chem., 267, 166 ± 172. Crook T, Wrede D, Tidy JA, Mason WP, Evans DJ and Vousden KH. (1992). Lancet, 339, 1070 ± 1073. Doll R and Kinlen L. (1970). Br. Med. J., 4, 420 ± 422. Greenblatt MS, Grollman AP and Harris CC. (1996). Cancer Res., 56, 2130 ± 2136. Hartevelt MM, Bouwes Bavinck JN, Kootte AMM, Vermeer BJ and Vandenbroucke JP. (1990). Transplantation, 49, 506 ± 509. IARC. (1995). IARC Monographs on the Evaluation of the Carcinogenic Risk to Humans, vol 64, Lyon. Iftner T, Bierfelder S, Csapo Z and P®ster H. (1988). J. Virol, 62, 3655 ± 3661. Jego N, Thomas G and Hamelin R. (1993). Oncogene, 8, 209 ± 213. Kelly GE, Meikle W and Sheil AGR. (1987). Br. J. Dermatol., 117, 429 ± 440.

Kiyono T, Nagashima K and Ishibashi M. (1989). Virology, 173, 551 ± 565. Lennard L, Thomas S, Harrington CI and Maddocks JL. (1985). Br. J. Dermatol., 113, 723 ± 729. Lewandowsky F and Lutz W. (1922). Arch. Dermatol. and Syphil., 141, 193 ± 203. London NJ, Farmery SM and Will EJ. (1995). Lancet, 346, 403 ± 406. McGregor JM and Proby CM. (1995). Lancet, 346, 964 ± 965. McGregor JM and Proby CM. (1996). Cancer Surveys, 26, 219 ± 236. Morris JD, Eddleston AL and Crook T. (1995). Lancet, 346, 754 ± 758. Nataraj AJ, Trent JC and Honnavara N. (1995). Photochem. Photobiol., 62, 218 ± 230. Piette J. (1991). J. Photochem. Photobiol., 11, 241 ± 260. Prives C. (1994). Cell, 78, 543 ± 546. Steger G and P®ster H. (1992). Arch. Virol., 125, 355 ± 360. Swann PF, Waters TR, Moulton DC, Xu Y-Z, Zheng Q, Edwards M and Mace R. (1996). Science, 273, 1109 ± 1111. Woodworth CD, Doniger J and DiPaolo JA. (1989). J. Virol., 63, 159 ± 164. Ziegler A, Leel DJ, Kunala S, Sharma HW, Gailani M, Simon JA, Halperin AJ, Baden HP, Shapiro PE, Bale AE and Brash DE. (1993). Proc. Natl. Acad. Sci. USA, 90, 4216 ± 4220.