Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Tissue-specific expression of SV40 in tumors associated with the Li–Fraumeni syndrome

Abstract

Inactivation of wild-type p53 tumor suppressor function is the primary mechanism of tumor initiation in Li–Fraumeni syndrome (LFS) individuals with germline p53 mutations. Tumors derived from LFS patients frequently retain the normal p53 allele, suggesting that alternative mechanisms in addition to gene deletion must be involved in inactivating wild-type p53 protein. DNA tumor viruses, such as SV40, target p53 for inactivation through the action of viral oncoproteins. We studied the probands from two unrelated LFS families, each of whom presented with multiple malignant neoplasms. Patient 1 developed an embryonal rhabdomyosarcoma (RMS) and a choroid plexus carcinoma (CPC), while patient 2 developed a CPC and subsequently presented with both an osteosarcoma (OS) and renal cell carcinoma (RCC). We utilized DNA sequence analysis and immunohistochemistry to determine p53 gene status in the germline and tumors, as well as evidence for SV40 T-antigen oncoprotein expression. Each patient harbored a heterozygous germline p53 mutation at codons 175 and 273, respectively. In patient 1, the normal p53 gene was lost while the mutant p53 allele was reduced to homozygosity in the RMS. Both normal and mutant genes were maintained in the CPC. In patient 2, normal and mutant p53 alleles were retained in both the CPC and RCC. Both specific PCR and immunostaining detected SV40 T-antigen in both CPCs and the RCC. In addition to chromosomal alterations, epigenetic mechanisms may disrupt p53 function during tumorigenesis. In two LFS patients, we found SV40 DNA sequences and viral T-antigen expression that could account for inactivation of the normal p53 protein. Inactivation of p53 or other tumor suppressors by viral proteins may contribute to tumor formation in specific tissues of genetically susceptible individuals.

Introduction

The Li–Fraumeni syndrome (LFS) is a paradigm of hereditary predisposition to childhood and adult-onset cancers. The syndrome is generally defined by the following criteria: (1) a bone or soft tissue sarcoma diagnosed under 45 years of age in an individual who is then designated the proband; (2) one first degree relative of the proband with cancer diagnosed before 45 years of age; and (3) one first or second degree relative of the proband in the same parental lineage with cancer under 45 years, or sarcoma diagnosed at any age (Li et al., 1988). Germline mutations of the p53 tumour suppressor gene are associated with the majority of LFS families (Malkin et al., 1990; Srivastava et al., 1990, Varley et al., 1997; Frebourg et al., 1995, Malkin, 1998). Somatic inactivation of the p53 tumor suppressor allele results in tumor formation. However, aside from gene loss, the mechanisms of somatic p53 inactivation are generally not known. We used the Li–Fraumeni syndrome as a unique model of nature to assay alternative mechanisms of inactivation of normal p53 in human tumor formation.

The most frequently observed cancers reported in LFS include soft tissue sarcoma, breast cancer, adrenocortical carcinoma, brain tumor and leukemia (Varley and Birch, 1997). Numerous other cancers have also been reported with somewhat less frequency (Kleihues et al., 1997). Rhabdomyosarcoma is the most common early onset cancer in LFS families, and is the most common sarcoma of childhood, representing 10–15% of all childhood malignancies (Kramer et al., 1983). Although the etiology of RMS is not known, several lines of epidemiologic evidence suggest that a proportion of patients is genetically susceptible to tumor development. A subset of patients with RMS harbor one constitutionally altered p53 allele, independent of a family history of cancer consistent with the LFS or its variants (Diller et al., 1995; Brugieres et al., 1993). Choroid plexus tumors account for 2–5% of pediatric brain tumors, and choroid plexus carcinomas (CPC) account for greater than 25% of all choroid plexus neoplasms of childhood (Carpenter et al., 1986; Packer et al., 1992). The etiology of CPC, like RMS, is not known, although CPCs are also commonly observed in families with the LFS (Jolly et al., 1994; Yuasa et al., 1993). Osteosarcoma is another commonly observed component tumor of LFS. Furthermore, approximately 10% of patients with sporadic osteosarcoma harbor germline mutations of either the RB1 or p53 tumor suppressor genes (Miller et al., 1996; McIntyre et al., 1994). Renal cell carcinoma is more commonly associated with von Hippel–Lindau disease than LFS; nevertheless, it too has been reported to occur in LFS families (Sedlacek et al., 1998).

Approximately 50% of tumors that develop in LFS patients retain the normal p53 allele (Sedlacek et al., 1998; Varley et al., 1997). However, the mechanism by which tumor formation occurs in these patients has not been determined. We evaluated the mechanism of p53 inactivation in LFS patients with multiple tumors. Based on previous findings of polyomavirus (e.g. SV40) DNA in CPC and OS tumors (Bergsagel et al., 1992; Carbone et al., 1996; Mendoza et al., 1998), our studies were directed at viral oncoprotein mechanisms of second hits in the loss of normal p53 function. We found that both genetic and epigenetic mechanisms of p53 inactivation occur in LFS patients. These results suggest that viral proteins that inactivate tumor suppressor genes may contribute to oncogenesis in specific tissues of genetically susceptible individuals.

Results

Clinical presentation

Patient 1

 At the age of 6 months, a previously healthy male infant was noted to have fullness of the right side of his face. At 11 months, biopsy of a right zygomatic mass revealed an embryonic rhabdomyosarcoma. Computerized tomographic and magnetic resonance imaging delineated the parameningeal extent of the RMS, as well as coincidentally identifying a second mass in the left cerebral ventricular system. The intracerebral mass was resected, and pathologically identified as a CPC. Twenty months after diagnosis, following both chemotherapy and radiotherapy, the patient died with disseminated intraspinal extension of his RMS. No postmortem examination was performed. The family history was remarkable for a paternal uncle treated for a malignant fibrohistiocytoma of the humerus at age 31, who then developed a benign parotid gland tumor at age 40. The patient's paternal grandmother died at age 30 of a malignant brain tumor. There was no history of cancer on the maternal side. The child was therefore the proband in a family with an LFS phenotype (Figure 1a).

Figure 1
figure1

(a) Pedigree of Patient 1: RMS - rhabdomyosarcoma; CPC - choroid plexus carcinoma; MFH - malignant fibrohistiocytoma; PAR - parotid tumor; CNS - brain tumor. (b) Pedigree of Patient 2: OS - osteosarcoma; RCC - renal cell carcinoma. Numbers represent ages of initial presentation of tumor. Squares denote males, circles denote females

Patient 2

 At the age of 12 months, a previously healthy male infant presented with an intracranial mass which upon surgical resection revealed a choroid plexus carcinoma. The patient was treated with multi-agent chemotherapy, and following a second resection upon tumor regrowth, has had no subsequent recurrences. He presented again at age 16 years with swelling of the left temporal area and was noted on biopsy to have an osteosarcoma. Although this tumor was non-metastatic, on staging he was found to have a bilateral renal cell carcinoma with primary disease in the left kidney and a metastasis or synchronous primary in the right kidney. Focal nodular hyperplasia of the liver was confirmed from biopsy of an isolated hepatic mass. The osteosarcoma was treated with surgical resection and multi-agent chemotherapy, and the renal cell carcinoma was resected. He is currently disease-free. The child's mother died of a ‘cerebral aneurysm’ in her 30's. There is no history of cancer on the mother's side of the family. The medical history of the paternal side of the family is not known (Figure 1b).

p53 gene analysis

To evaluate whether patient 1 harbored a germline p53 mutation, high molecular weight genomic DNA was prepared from peripheral blood leukocytes obtained from the proband and his parents. Nine sets of primers were generated to amplify DNA fragments spanning exons 2–11 of the p53 gene, encompassing the intron-exon junctions. SSCP analysis demonstrated similar altered band patterns in the p53 exon 5 fragment of both Patient 1 and his father's constitutional DNA. The SSCP pattern of Patient 1's mother was consistent with the normal gene sequence. DNA sequence analysis of both Patient 1 and his father demonstrated a C to T transition at the second position of codon 175 yielding an amino acid change from arginine to histidine (Figure 2a). In both individuals' samples, the normal p53 gene could also be detected, indicating the heterozygous nature of the mutation in the blood.

Figure 2
figure2

(a) p53 gene sequence in Patient 1's blood (germline). Products from p53 exon 5 amplification were sequenced using the direct method as described. The patient (left panel) shows a heterozygous germline mutation at R175H. A control sample (right panel) shows normal p53 sequence. (b) p53 gene sequence in patient's rhabdomyosarcoma (RMS) and choroid plexus carcinoma (CPC). The RMS shows reduction to homozygosity of the mutant R175H allele, whereas the CPC demonstrates retention of heterozygosity. (R-arginine; H-histidine) (c–e) p53 gene sequence in Patient 2's normal liver (c), choroid plexus carcinoma (d) and renal cell carcinoma (e). Products from p53 exon 8 amplification were sequenced as described. Each panel shows a heterozygous mutation (germline in the normal liver sample) at R273C. (R-arginine; C-cysteine). T-thymine; C-cytosine; G-guanine; A-adenine; arrows indicate sites of basepair substitutions

PCR amplification of p53 exon 5 was performed on each tumour from this patient, and the PCR products were sequenced. The RMS demonstrated reduction to homozygosity of the mutant R175H allele, whereas the CPC demonstrated retention of heterozygosity (Figure 2b). The entire p53 gene was sequenced to determine whether other mutations were present in the tumors. No other mutations were observed (data not shown). To determine the localization of the p53 protein expressed in both tumours, immunohistochemical analysis was performed using mouse monoclonal antibodies p53 (Ab-6), clones DO-1 and DO-7, and p53 (Ab-2), clone 1801 which recognize both mutant and wild-type conformations of p53. Representative sections stained with DO-7, which yielded the least background staining, are shown in Figure 3. Exclusive nuclear localization of p53 was demonstrated in the RMS (Figure 3A, panel b), whereas in the CPC the antibody localized to both the nucleus and the cytoplasm (‘punctate staining’, Figure 3A, panel d). Normal control choroid plexus epithelium (obtained from the Brain Bank in the Department of Pathology, Hospital for Sick Children) demonstrated no evidence of p53 immunopositivity. These results indicate that inactivation of p53 function in the RMS was accomplished by inheritance of one mutant allele followed by deletion of the second, normal, allele. In contrast, inactivation of normal p53 function in the CPC must be accomplished by a non-genetic event, as the gene and its encoded protein were retained in the tumour.

Figure 3
figure3

p53 immunohistochemical analysis. Mouse monoclonal antibody p53, clone DO-7 recognizes both mutant and wild-type conformations of p53. (A) Panel a: RMS negative control; panel b: RMS stained with DO-7 demonstrating intense immunopositivity; panel c: CPC negative control; panel d: CPC stained with DO-7 demonstrating intense nuclear and cytoplasmic immunopositivity (arrows). (B) Panel a: RCC negative control; panel b: RCC stained with DO-7 demonstrating intense immunopositivity (arrows); panel c: CPC negative control; panel d: CPC stained with DO-7 demonstrating immunopositivity (arrows); panel e: normal choroid plexus epithelium negative control; panel f: normal choroid plexus epithelium stained with DO-7 demonstrating no overexpression of p53

For Patient 2, SSCP analysis demonstrated similar altered band patterns in the p53 exon 8 fragment from normal liver tissue, as well as of the CPC and renal cell carcinoma (RCC). PCR products could not be consistently amplified from the osteosarcoma, suggesting that significant DNA degradation had occurred perhaps secondary to tissue preservation techniques. DNA sequence analysis of all samples with band shifts demonstrated a C to T transition at the first position of codon 273 yielding an amino acid change from arginine to cysteine (Figure 2, panels c–e). In each sample, normal p53 could also be detected, indicating the heterozygous nature of the mutation. The entire p53 gene was sequenced to determine whether other mutations were present in the tumors. No other mutations were observed (data not shown). To determine the localization of the p53 protein expressed in the normal and tumor tissues, immunohistochemical analysis was performed as for Patient 1. Although the overall quality of staining was poorer than for Patient 1's samples, p53 immunopositivity is demonstrated in the renal and choroid plexus carcinomas (Figure 3B, panels b,d). Normal control choroid plexus epithelium (obtained from the Brain Bank in the Department of Pathology, Hospital for Sick Children) demonstrated no evidence of p53 immunopositivity (Figure 3B, panel f). As with the CPC in Patient 1, these results indicate that inactivation of p53 in the RCC and CPC in Patient 2 was accompanied by a non-genetic event, as the normal gene and its encoded protein were retained in the tumor.

Detection of SV40 T-antigen

Based on previous reports of SV40-like DNA sequences in childhood choroid plexus tumours (Bergsagel et al., 1992), in osteosarcoma (Carbone et al., 1996; Mendoza et al., 1998), and its tropism to renal epithelium, we asked whether the presence of SV40 T-antigen could account for the ‘second-hit’ in both patients' CPCs and the RCC and osteosarcoma in Patient 2. The degraded nature of the DNA from Patient 2's osteosarcoma made it uninformative. Total genomic DNA was extracted from normal and tumour tissue sections. PCR amplification of a 574 bp SV40-specific T-antigen sequence encompassing the intron and the Rb-binding pocket was performed using the SVfor3 and SVrev primers (Figure 4a,b). The positive control, pAC373.SV, contained the cDNA for SV40 large T-antigen that amplifies an intronless (250 bp) fragment with these primers. For Patient 1, the 574 bp fragment was consistently amplified from the CPC, whereas on greater than 10 attempts, this sequence could not be amplified from the RMS. When the 574 bp PCR product was digested with BstX1, products specific to SV40 were obtained (Figure 4a). Direct DNA sequence analysis of the amplified fragment confirmed that it was derived from SV40 (in comparison to GenBank sequences, data not shown). For Patient 2, the quality of DNA amplified from the paraffin-embedded tissue was relatively poorer than for Patient 1, and long PCR products could not be amplified. However, using the SVfor3 and SVRev primers, a smaller 105 bp fragment was consistently amplified from the CPC, the RCC, whereas on greater than 10 attempts, this sequence could not be amplified from the normal liver or renal tissue (Figure 4b). As with samples from Patient 1, direct DNA sequence analysis of the amplified fragment confirmed that it was derived from SV40 (in comparison to GenBank sequences, data not shown). PCR amplifications for similar sequences from the related BK and JC viruses were negative.

Figure 4
figure4

SV40 T-antigen PCR. PCR products from amplification using primers and conditions as described are visualized on a 3% agarose gel. (a) Patient 1: The left side of the gel represents undigested PCR products. The right side shows the same products digested with BstXI. The first lane on each side contains a 100 bp marker. The second lane is a positive control, pAC373.SV (SV40), a plasmid containing the cDNA for SV40 T-antigen with a PCR product of predicted size 250 bp. The third lane is a negative control PCR reaction with no DNA template. The tumor DNA from the CPC has a single band of the expected size (574 bp) and demonstrates specific BstXI digestion in the right hand panel. The RMS-derived DNA has only non-specific bands in the left-hand panel and no restriction enzyme digestion. CPC - choroid plexus carcinoma; RMS - rhabdomyosarcoma. (b): Patient 2: The left and right-hand most lanes contain a 100 bp marker. The second lane is a positive control, using primers Svfor3 and SVRev, with a PCR product of predicted size 105 bp. The third (normal liver), fourth (osteosarcoma) and fifth (normal kidney) lanes demonstrate no amplification of SV40. The lanes containing DNA from the RCC and CPC demonstrate amplification of a 105 bp fragment. RCC-renal cell carcinoma

In order to determine whether the SV40 T-antigen protein was expressed, immunohistochemical analyses were performed using a primary antibody (PAb419) (Crawford et al., 1982) derived from ascitic fluid of SV40 transgenic mice (kindly provided by Dr Terry Van Dyke, UNC). As shown in Figure 5, there was no SV40 T-antigen staining in the RMS of Patient 1 (Figure 5A, panel b). However, diffuse punctate staining was present in Patient 1's CPC (Figure 5A, panel d), as well as Patient 2's CPC and RCC (Figure 5B, panels b and d). Although one might expect SV40 T-antigen to normally localize to the nucleus, mutant forms of the virus are reported in vitro to sequester p53 in the cytoplasm, limiting its ability for intranuclear transport (Dobbelstein and Roth, 1998). The immunohistochemistry was therefore consistent with the PCR results, and indicated that SV40 T-antigen was not only present but also expressed in these tumors.

Figure 5
figure5

SV40-T antigen immunohistochemical analysis. A primary antibody (PAb419) derived from ascitic fluid of SV40 transgenic mice was used. (A): Panel a: RMS negative control; panel b: RMS stained with PAb419 demonstrating absence of immunopositivity; panel c: CPC negative control; panel d: CPC stained with PAb419 demonstrating focal intense punctate nuclear and cytoplasmic immunopositivity (arrows); (B): panel a: RCC negative control; panel b: RCC stained with PAb419 demonstrating focal intense staining (arrow); panel c: CPC negative control; panel d: CPC stained with PAb419 demonstrating focal intense punctate cytoplasmic immunopositivity (arrow)

Discussion

The analysis of two Li–Fraumeni syndrome individuals led to an investigation of how the function of the normal p53 allele was lost in their tumors. We found that different tumors in individuals harboring a point mutation in a germline p53 allele may have the normal allele/protein inactivated by distinct mechanisms–either genetic loss of the normal allele or functional disruption of the p53 protein by its interaction with a viral oncoprotein, namely T-antigen. Our observations indicate that the ‘second hit’, as suggested by Knudson, can be distinct for different cell types in the same host. Although the mechanisms of the second hits were independent in our patients, the latency to malignant transformation and eventual tumor formation following functional inactivation of p53 was similar in embryologically unrelated tumors.

Analysis of tumors occurring in p53 heterozygous knockout mice demonstrate that a significant proportion (40%) retain the normal allele (Venkatachalam et al., 1998). These observations are recapitulated in tumors derived from LFS carriers with germline p53 mutations in which the normal allele is frequently retained in the cancer (Sedlacek et al., 1998; Varley et al., 1997). It has been suggested that a reduction in ‘functional expression’ of the normal allele leads to p53 inactivation, cell cycle dysregulation and malignant transformation (Dobbelstein and Roth, 1998). Inactivation of the wild-type allele can also be achieved by a transdominant-negative effect of the certain mutant alleles. Nevertheless, at least in some LFS patients this reduction in p53 function may be a result of a virus infection, in this case by SV40, in specific tissues to which the virus is tropic through expression of the viral T-antigen which can inactivate p53 by direct binding. This association of SV40 with LFS-associated tumors was suggested in previous analyses of tumor DNA from LFS family members where SV40 DNA sequences were detected in bone tumors and in one uncharacterized LFS patient with a CPC (Bergsagel et al., 1992; Carbone et al., 1996).

Detection of SV40 in human tissue has been previously criticized as artifact (Carbone et al., 1997; Butel and Lednicky, 1999). The observation that SV40 T-antigen DNA and immunostaining were present only in the CPC and not in parallel RMS tumor and normal lymphocyte samples from Patient 1, and that SV40 T-antigen DNA were present only in the CPC and RCC, and not in the parallel normal liver from Patient 2, effectively controls for artifacts derived from laboratory contaminants. Furthermore, SV40 strains had not previously been used in the laboratory carrying out the PCR. The lack of BK and JC virus PCR amplification is important in that T-antigens derived from these viruses are closely related to that of SV40 and, if present, could have confounded the immunohistochemical identification of SV40 T-antigen. However, these viruses have been detected by PCR in other tumor types, and their T-antigens may play a role in tumor induction similar to the p53 binding of SV40 T-antigen suggested by these patients (Flaegstad et al., 1999; Krynska et al., 1999). In particular, the recent report of punctate cytoplasmic staining of p53 in human neuroblastomas containing BK virus T-antigen (similar to that seen for p53 in the CPCs in this study), suggests that polyomavirus T-antigen binding to p53 may have a distinctive immunohistochemical pattern in human cells (Flaegstad et al., 1999).

The literature is extant with reports of SV40 and SV40-like genomic fragments in various human tumours, most frequently ependymomas, choroid plexus tumors, osteosarcomas and mesotheliomas. The role of SV40 in these tumors has remained in question partly because of the inter-observer lack of reproducibility of results and of the lack of a genetic mechanism to explain tumor formation. The immunohistochemical analysis of SV40 T-antigen in our patient's CPC may explain difficulties in reproducing PCR amplification of SV40 in these samples. As shown by our immunohistochemistry results, a relative minority of transformed cells actually expresses T-antigen protein. Thus, all or part of the viral genome may be lost upon clonal expansion and malignant transformation, i.e. hit-and-run tumor initiation. This mechanism of tumor formation with the detection of progressively decreased expression of SV40 T-antigen has been observed in both the choroid plexus and hepatocellular tumor mouse models in which not only is the level of T-antigen expression highest in the embryonic tissue, but in fact some tumor nodules are negative for T-antigen (Brinster et al., 1984; Ewald et al., 1996). If SV40 T-antigen is etiologically important in the development of the tumors, we might expect SV40 DNA (and T-antigen expression) to be present in all the tumor cells. However, because of the prediction of cellular genetic changes over time in the tumor progression model, it is possible that the accumulation of such mutations makes T-antigen functions dispensable in late-stage tumors, permitting the loss of SV40 T-antigen DNA from some tumor cells (Chen et al., 1992). Whether the sequence of primary inactivation of p53 by mutation followed by SV40-associated inactivation applies to tumors not in LFS families remains to be determined. At this point, one can only speculate on the mechanism of acquisition of SV40 in these children. Transplacental vertical transmission cannot be ruled out, as normal tissue from the mothers was not available for study. Acquired infection in the children also cannot be evaluated because serum samples were not available for antibody screening. However, the previous observation reported by Carbone et al. (1996) in which 10 of 11 osteosarcomas from LFS patients were SV40-positive (p53 status unknown), suggests that a primary p53 mutation establishes a background upon which SV40 can then provide the second event leading to complete abrogation of normal p53 function.

Although transgenic mice expressing SV40 T-antigen under the natural viral promoter uniformly develop choroid plexus tumors, evidence for a cause and effect role for polyomaviruses in human tumorigenesis has been inferential (Van Dyke et al., 1987). In light of recent studies suggesting the predominant role of the environment in the etiology of human cancers (Lichtenstein et al., 2000), the observations outlined in this study suggest the intriguing possibility of a functional balance of an epigenetic (viral) etiologic event acting in the presence of an underlying genetic alteration (germline p53 mutation) in the development of selected cancers in Li–Fraumeni syndrome. Viruses, such as human papillomaviruses and SV40, that express proteins which interact with critical tumor suppressor gene products may contribute to tumor induction in specific tissues particularly when accompanied by other genetic changes that facilitate cell growth control.

Materials and methods

DNA extraction

DNA was extracted from blood samples (obtained after informed parental consent) from Patient 1 and his parents, using a standard red blood cell lysis/nuclei lysis method with phenol/chloroform purification and ethanol precipitation. DNA was extracted from a total of 100 μ of tissue sections from the two tumors in Patient 1 and the three tumors of Patient 2 after xylene de-paraffinization and alcohol rehydration. Normal liver obtained from a biopsy at the time of metastatic evaluation was used as a source of constitutional (germline) tissue from Patient 2 for whom blood could not be obtained. A modified method of proteinase K digestion, phenol/chloroform purification, and ethanol precipitation was used for extraction from all tissue samples. Separate blades were used to cut sections from each tumor and tissue block to reduce the possibility of cross-contamination.

p53 PCR analysis

The p53 primer sequences and lengths of the amplified fragments have been previously published (Mashiyama et al., 1991; Peng et al., 1993). Samples were screened for p53 mutations using published primers for all exons with minor modifications to the primers spanning exon 10. The PCR reactions were performed using AmpliTaq Gold DNA polymerase (Perkin Elmer Cetus). Thermocycling was performed by denaturation at 95°C for 10 min, followed by cycling 30 times at 94°C for 45 s, at 55°C for 45 s, and at 72°C for 45 s. PCR samples were analysed by single strand conformational polymorphism (SSCP) analysis using a standard protocol. Each fragment was analysed using at least two electrophoretic and gel constitution conditions, with both positive and negative controls to ensure that known mutations could be detected. Mutations in exon 5 were detected in Patient 1, the primer sequences for which are 5U1 5′-TTCCTCTTCCTACAGTACTC-3′ and I5R 5′-GCAACCAGCCCTGTCGTCTC-3′. Mutations in exon 8 were detected in Patient 2, the primer sequences for which are 8U1 5′-CCTATCC TGAGTAGTGGTAA-3′ and 8L3 5′-TGAATCTGAGGCATAACTGC-3′.

SV40 PCR analysis

The primers used to amplify SV40 sequences were Svfor2 (5′-CTTTGGAGGCTTCTGGGATGCAACT-3′) from nucleotide 4945 to 4921, svfor3 (5′-TGAGGCTACTGCTGACTCTCA-3′) from nucleotide 4454 to 4477, and SvRev (5′-GCATGACTCAAAAACTTAGCAATTCTG-3′) from nucleotide 4372 to 4399 (Bergsagel et al., 1992). The PCR reactions were performed using AmpliTaq Gold DNA polymerase (Perkin Elmer Cetus). Thermocycling using primers Svfor2 and SVRev was performed by denaturation at 95°C for 10 min, followed by cycling 30 times at 94°C for 1 min, at 57°C for 1 min and at 72°C for 1 min with a final extension of 72°C for 15 min. At this point another aliquot of Taq was added to each reaction and a further 30 cycles were performed. A total of 60 cycles was performed for DNA samples extracted from paraffin. The expected PCR product size using this primer set is 574 bp. Thermocycling using primers Svfor3 and SVRev was performed by denaturation at 95°C for 10 min, followed by cycling 40 times at 94°C for 1 min, at 60°C for 1 min and at 72°C for 1 min with a final extension of 72°C for 15 min. The expected PCR product size using this primer set is 105 bp. PCR samples were analysed on a 3% (NuSieve, SeaKem) agarose gel.

Sequence analysis of PCR products

Each sample determined to have an abnormal p53 gene by band shift on SSCP analysis was amplified with the SSCP primers encompassing the respective abnormal region. Fragments were then either subcloned for sequencing or sequenced directly. A T-tailed pBSK vector was used for subcloning, and at least six independent clones were sequenced in both directions by the Sanger dideoxide method with a Sequenase 2.0 kit (US Biochemicals/Amersham Corp., Arlington Heights, IL, USA). Direct DNA sequencing of double-stranded PCR amplified products, pre-treated with exonuclease I/shrimp alkaline phosphatase to eliminate any unincorporated primers or dNTPs from the PCR product, was performed using the Thermo Sequenase sequencing kit (Amersham Life Sciences). Sequencing as above was also used for the analysis of the 105 bp SV40 product. The 574 bp SV40 PCR product was sequenced using the automated ABI Prism 377 Sequencer (Applied Biosystems, Inc.).

p53 Immunohistochemistry

Immunohistochemistry was performed using an indirect immunoperoxidase/antigen retrieval enhancement protocol. Sialinated slides were prepared using 5 μ sections of tumour tissue. No sections from the osteosarcoma were available for immunohistochemical analysis. The quality of the limited number of renal cell and choroid plexus carcinoma sections available from Patient 2 was poor. Mouse monoclonal antibodies p53 (Ab-6), clones DO-1 and DO-7 and p53 (Ab-2), clone 1801 were used for immunostaining each tissue sample. The antibodies react with both mutant and wild-type p53.

SV40 large T immunohistochemistry

Slides were prepared using 5 μ sections of tissue. The sections were dewaxed and hydrated according to standard immunohistochemistry protocol. Specific to this procedure, a trypsin pretreatment of sections was performed before immunostaining. Trypsin (1 mg/ml, PBS) was applied to each section and incubated at 37°C for 15 min. The immunostaining was performed using the Vectastain Elite ABC kit (Vector Laboratories Inc.). The primary antibody used was PAb419, originally derived from ascitic fluid generated in an SV40 T antigen-transgenic mouse (Crawford et al., 1982) (kindly provided by Dr Terry Van Dyke, University of North Carolina, NC, USA). Slides were incubated overnight at 4°C. The primary antibody was detected by the addition of a biotynilated anti-mouse IgG affinity purified secondary antibody. The biotinylated secondary antibody was then detected using the kit reagent, which is a macromolecular complex between avidin and biotinylated enzyme. The antibody complex was visualized by adding DAB (3,3-diaminobenzidine) substrate. Slides were then counter-stained with hematoxylin, dehydrated and mounted using standard immunohistochemistry protocol. For both the p53 and SV40 T antigen studies, sections of normal choroid plexus epithelium obtained from the Brain Bank of the Department of Pathology, Hospital for Sick Children were used.

References

  1. Bergsagel DJ, Finegold MJ, Butel JS, Kupsky WJ, Garcea RL . 1992 N. Engl. J. Med. 32: 988–993

  2. Brinster RL, Chen HY, Messing A . 1984 Cell 37: 367–377

  3. Brugieres L, Gardes M, Moutou C, Chompret A, Meresse V . 1993 Cancer Res. 53: 452–455

  4. Butel JS, Lednicky JA . 1999 J. Natl. Cancer Inst. 91: 119–134

  5. Carbone M, Rizzo P, Pass HI . 1997 Oncogene 15: 1877–1888

  6. Carbone M, Rizzo P, Procopio A, Giuliano M, Pass HI, Gebhardt MC, Mangham C, Hansen M, Malkin D, Bushart G, Pompetti F, Picci P, Levine A, Bergsagel JD, Garcea RL . 1996 Oncogene 13: 527–535

  7. Carpenter DB, Michelsen WG, Hays AP . 1986 J. Neurosurg. 56: 722–727

  8. Chen J, Tobin GJ, Pipas JM, Van Dyke T . 1992 Oncogene 7: 1167–1173

  9. Crawford L, Leppard K, Lane D, Harlow E . 1982 J. Virol. 42: 612–620

  10. Diller L, Sexsmith E, Gottlieb A, Li FP, Malkin D . 1995 J. Clin. Invest. 95: 1606–1611

  11. Dobbelstein M, Roth J . 1998 J. Gen. Virol. 79: 3079–3085

  12. Ewald D, Li M, Efrat S, Auer G, Wall RJ, Furth PA, Henninghausen L . 1996 Science 273: 1384–1386

  13. Flaegstad T, Andresen PA, Johnsen JI, Asomani SK, Jorgensen GE, Vignarajan S, Kjuul A, Kogner P, Traavik T . 1999 Cancer Res. 59: 1160–1163

  14. Frebourg T, Barbier N, Yan Y-X, Garber JE, Dreyfus M, Fraumeni JF, Li FP, Friend SH . 1995 Am. J. Hum. Genet. 56: 608–615

  15. Jolly KW, Malkin D, Douglass EC, Brown TF, Sinclair AE, Look AT . 1994 Oncogene 9: 97–102

  16. Kleihues P, Schauble B, Hausen AZ, Esteve J, Ohgaki H . 1997 Am. J. Pathol. 150: 1–13

  17. Kramer S, Meadows AT, Jarrett P, Evans AE . 1983 J. Natl. Cancer Inst. 70: 49–55

  18. Krynska B, Del Valle L, Croul S, Gordon J, Katsetos CD, Carbone M, Giordano A, Khalili K . 1999 Proc. Natl. Acad. Sci. USA 96: 11519–11524

  19. Li FP, Fraumeni Jr JF, Mulvihill JJ, Blattner WA, Dreyfus MG, Tucker MA, Miller RW . 1988 Cancer Res. 48: 5358–5362

  20. Lichtenstein P, Niels VH, Verkasalo PK, Iliadou A, Kaprio J, Koskenvuo M, Pukkala E, Skytthe A, Hemminki K . 2000 New Engl. J. Med. 343: 78–85

  21. Malkin D . 1998 The Li-Fraumeni Syndrome. The Genetic Basis of Human Cancer 1st edition. Vogelstein B and Kinzler KW. (eds) New York: McGraw-Hill, Inc pp. 393–408

  22. Malkin D, Li FP, Strong LC, Fraumeni Jr JF, Nelson CE, Kim DH, Kassel J, Gryka MA, Bischoff FZ, Tainsky MA, Friend SH . 1990 Science 250: 1233–1238

  23. Mashiyama S, Murakami Y, Yoshimoto T, Sekiya T, Hagiwara K . 1991 Oncogene 6: 1313–1318

  24. McIntyre JF, Smith-Sorensen B, Friend SH, Kassel J, Borresen A-L, Yan YX, Russo C, Sato J, Barbier N, Miser J, Malkin D, Gebhardt MC . 1994 J. Clin. Oncol. 12: 925–930

  25. Mendoza SM, Konishi T, Miller CW . 1998 Oncogene 17: 2457–2462

  26. Miller CW, Aslo A, Won A, Tan M, Lampkin B, Koeffler HP . 1996 J. Cancer Res. Clin. Oncol. 122: 559–565

  27. Peng H-Q, Hogg D, Malkin D, Gallie BL, Bulbul M, Jewett M, Buchanan J, Goss PE . 1993 Cancer Res. 53: 3574–3578

  28. Packer RJ, Perilongo G, Johnson D et al . 1992 Cancer 69: 580–585

  29. Sedlacek Z, Kodet R, Seemanova E, Vodvarka P, Wilgenbus P, Mares J, Poustka A, Goetz P . 1998 Br. J. Cancer 77: 1034–1039

  30. Srivastava S, Zou Z, Pirollo K, Blattner W, Chang EH . 1990 Nature 348: 747–749

  31. Van Dyke TA, Finlay C, Miller D . 1987 J. Virol. 61: 2029–2032

  32. Varley JM, Birch JM . 1997 Br. J. Cancer 76: 1–14

  33. Varley JM, McGown G, Thorncroft M, Santibanez-Koref MF, Kelsey AM, Tricker KJ, Evans DGR, Birch JM . 1997 Cancer Res. 57: 3245–3249

  34. Varley JM, Thorncroft M, McGown G, Appleby J, Kelsey AM, Tricker KJ, Evans DG, Birch JM . 1997 Oncogene 14: 865–871

  35. Venkatachalam S, Shi YP, Jones SN, Vogel H, Bradley A, Pinkel D, Donehower LA . 1998 EMBO J. 17: 4657–4667

  36. Yuasa H, Tokito S, Tokunaga M . 1993 Neurosurgurgy 32: 131–133

Download references

Acknowledgements

We thank L Becker and H Yeger for assistance in interpretation of pathologic analysis, and T Van Dyke for providing the PAb419. This work was supported by grants from the National Cancer Institute of Canada with funds of the Terry Fox Foundation, the Grant Miller Research Fund, University of Toronto, and the Pediatric Oncology Group of Ontario. D Malkin is a Research Scientist of the National Cancer Institute of Canada supported by the Canadian Cancer Society.

Author information

Affiliations

Authors

Corresponding author

Correspondence to David Malkin.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Malkin, D., Chilton-MacNeill, S., Meister, L. et al. Tissue-specific expression of SV40 in tumors associated with the Li–Fraumeni syndrome. Oncogene 20, 4441–4449 (2001). https://doi.org/10.1038/sj.onc.1204583

Download citation

Keywords

  • Li–Fraumeni syndrome
  • p53
  • SV40 polyomavirus
  • tumor initiation

Further reading

Search

Quick links