Heterozygosity for p53 promotes microsatellite instability and tumorigenesis on a Msh2 deficient background

Article metrics


In colorectal tumorigenesis, loss of function of the mismatch repair genes is closely associated with genomic instability at the nucleotide level whereas p53 deficiency has been linked with gross chromosomal instability. We have addressed the contribution of these two forms of genetic instability to tumorigenesis using mice mutant for Msh2 and p53. As previously reported, deficiency of both genes leads to rapid lymphomagenesis Here we show that heterozygosity for p53 also markedly reduces survival on an Msh2 null background. We characterized the patterns of genomic instability in a small set of tumours and showed that, as predicted p53 deficiency predisposes to aneuploidy and Msh2 deficiency leads to microsatellite instability (MSI). However, heterozygosity for p53 in the absence of Msh2 resulted in increased MSI and not aneuploidy. This implied role for p53 in modulating MSI was confirmed using a large cohort of primary fibroblast clones. The differences observed were highly significant (P<0.01) in both the fibroblast clones (which all retained p53 functionality) and the tumours, a proportion of which retained p53 functionality. Our results therefore demonstrate a dose sensitive role for p53 in the maintenance of genomic integrity at the nucleotide level.


Cancer is the endpoint of a process whereby normal cells are altered by the cumulative acquisition of genetic changes which confer proliferative, invasive and metastatic properties upon the cell (Vogelstein and Kinzler, 1993). This process can be rapidly accelerated by the loss of systems that safeguard genomic stability resulting in an increased rate of mutagenesis and chromosomal rearrangements (Hartwell, 1992; Kolodner, 1995). Acquisition of genomic instability in colorectal cancers has been associated with two separate pathways (Cottu et al., 1996; Lengauer et al., 1997). The first of these is characterized by defects in components of the DNA mismatch repair pathway, for example by Msh2 deficiency. The primary manifestation of such loss is an increase in MSI (Liu et al., 1995), presumed to arise as a direct consequence of failed DNA repair. The observation of a mismatch repair-dependent G2 cell cycle checkpoint (Hawn et al., 1995) is consistent with such a role. However, mismatch repair proteins may also exert a protective effect by mediating the deletion of cells which bear DNA damage, as functional Msh2 has been shown to confer sensitivity to alkylating agents both in vitro (de Wind et al., 1995) and in vivo (Toft et al., 1999; Sansom et al., 2001).

The majority of colorectal cancers show no evidence of MSI. However, these cancers frequently show gross genomic instability (Reichmann et al., 1981). Several lines of evidence suggest that defects in p53 are associated with this form of instability. First, alterations in p53 are correlated with the divergence of aneuploid sub-clones in colorectal cancers (Carder et al., 1993). Second, cells become permissive for DNA amplification when p53 is lost (Yin et al., 1992). Third, cells derived from p53 null mice display elevated levels of chromosomal instability (e.g. Harvey et al., 1993). The role of p53 is however not limited to preventing chromosomal instability, as p53 transcriptionally regulates a number of genes associated with cell cycle arrest, apoptosis and DNA repair. Furthermore, the p53 protein has been shown to directly interact with DNA, both through a 3′5′ exonuclease activity (Mummenbrauer et al., 1996) and through direct binding of the C-terminal domain to DNA mismatches (Lee et al., 1995; Dudenhoffer et al., 1998, Degtyareva et al., 2001). P53 has therefore been characterized to mediate a number of functions which may have direct tumour suppressor activity. Some of these activities have been shown to be p53 gene dose dependent (Clarke et al., 1993, 1994; Rafferty et al., 1996). It is therefore perhaps not surprising that evidence has been produced which suggests heterozygosity for p53 may predispose to neoplasia (Venkatachalam et al., 1998).

Recently evidence has emerged of a close relationship between p53 and Msh2 following DNA damage, most notably that MMR-mediated apoptosis is mediated through p53 (Toft et al., 1999) and that p53 becomes phosphorylated in an MMR-dependent fashion after alkylation damage (Duckett et al., 1999). To study the effects of p53 and Msh2 deficiency upon neoplasia, several murine strains have been produced which bear mutant alleles of Msh2 (de Wind et al., 1995; Reitmair et al., 1995, 1996) and p53 (Donehower et al., 1992; Clarke et al., 1993; Lowe et al., 1993). Mice null for p53 are viable, although a significant proportion of female p53−/− mice die during embryogenesis (Armstrong et al., 1995). Surviving animals rapidly develop tumours, predominantly thymic lymphomas with a smaller proportion succumbing to sarcomas (Donehower et al., 1992; Purdie et al., 1994). p53+/− heterozygote mice develop both lymphomas and sarcomas in approximately equal ratios, but with a longer latency period when compared to homozygote mice (Purdie et al., 1994). Mice null for Msh2 develop lymphomas with a peak incidence at 2–3 months of age (de Wind et al., 1995; Reitmair et al., 1995, 1996). Of those animals surviving past 6 months of age, 70% develop intestinal neoplasms (Reitmair et al., 1996) and 7% develop skin neoplasms analogous to those of the Muir–Torre syndrome (Reitmair et al., 1996).

To examine any possible synergism between Msh2 and p53 in neoplasia, Cranston et al. (1997) generated (Msh2−/−, p53−/−) mice. Combined deficiency of Msh2 and p53 resulted in developmental arrest of all female embryos at 9.5 days (Cranston et al., 1997). In contrast, male (Msh2−/−, p53−/−) mice were viable but succumbed to lymphoma more rapidly than either Msh2−/− or p53−/− alone, demonstrating co-operativity between the two genes. Such co-operativity did not appear to arise as a consequence of increased MSI, as tumours arising in (Msh2−/−, p53−/−) mice did not show significantly increased levels of MSI compared to tumours from Msh2−/− mice. We have also intercrossed mice bearing mutant p53 and Msh2 alleles mice, and confirm here the phenotype of (Msh2−/−, p53−/−) mice. We have additionally characterized the effect of heterozygosity at each locus with respect to murine survival and tumour development. Furthermore, we have investigated the influence of genotype upon the pattern of genetic instability within tumours. Taken together, these studies demonstrate a novel dose sensitive role for p53 in suppressing MSI.


Msh2−/− mice (de Wind et al., 1995) were crossed to p53−/− mice (Clarke et al., 1993) to generate a cohort of animals segregating for mutant alleles. Mice with the following genotypes were studied: (Msh2−/− p53+/+), (Msh2+/+ p53−/−), (Msh2−/− p53+/−), (Msh2+/− p53−/−) and (Msh2−/− p53−/−). Viable animals were obtained for each genotype, and both males and females in all five genotype groups were fertile beyond 6 weeks of age. In contrast to the findings of Cranston et al. (1997) we did not observe increased rates of female embryonic lethality in (Msh2−/−, p53−/−) mice (Toft et al., 1998).

Cohorts of mice were studied to investigate survival rates and tumorigenesis. Colonies were monitored for a period of 240 days, and mice were killed when they appeared visibly ill. Kaplan–Meier survival plots were generated for all genotypes (Figure 1). Survival data for (Msh2+/+, p53+/−) mice was generated from historical controls (Purdie et al., 1994). (Msh2−/−, p53−/−) mice became ill significantly earlier than either Msh2−/− or p53−/− mice alone (P<0.0006, log rank test). Heterozygosity for Msh2 on a p53−/− background did not significantly alter survival from p53 deficiency alone (P>0.09, log rank test). In contrast, Msh2 deficiency on a p53 heterozygote background (Msh2−/−, p53+/−) reduced survival compared with both (Msh2+/+, p53+/−) mice (P<0.00001, log rank test) and (Msh2−/−, p53+/+) mice (P<0.00001 log rank test).

Figure 1

Kaplan–Meier survival plots. Age of animals is given in days. Survival of p53+/− mice was reproduced from historical data (Purdie et al., 1994). All cohorts of mice are labelled accordingly. Cohort sizes were as follows: (Msh2−/−, p53−/−) n= 20; (Msh2+/−, p53−/−) n=18; (Msh2−/−, p53+/−) n= 47; Msh2−/− n=28; p53−/− n=57

Histological examination showed that all Msh2−/− mice were characterized by lymphoma, with 20% additionally developing intestinal neoplasia (adenomas). Eighty-eight per cent of p53−/− mice developed lymphoma and 12% developed poorly differentiated sarcomas. The tumour spectrum of (Msh2−/−, p53+/−) mice paralleled that of Msh2−/− mice (95% lymphoma, 11% intestinal neoplasia). The tumour spectrum of (Msh2+/−, p53−/−) mice paralleled that of p53−/− mice (83% lymphoma, 22% sarcoma), although two of the four sarcomas occurring in the (Msh2+/− p53−/−) mice were identifiable as leiomyosarcomas of the caecal wall, a tumour type not observed in the p53−/− cohort. All (Msh2−/−, p53−/−) mice developed lymphoma with no other tumour type observed.

We also analysed the DNA content of tumour cells by flow cytometry (Table 1). The DNA content was classified as aneuploid if an additional G1 or G2 peak with a different DNA index was present. Lymphomas from Msh2−/−, (Msh2−/−, p53+/−) and (Msh2−/−, p53−/−) mice all displayed a diploid karyotype. In contrast, two out of seven (29%) of the lymphomas arising in p53−/− mice showed evidence of aneuploidy. Flow cytometric analysis permits a relatively crude assessment of chromosomal instability. We therefore also analysed stability a separate set of tumours using comparative genomic hybridization (CGH). CGH was performed on a minimum of four to eight metaphase spreads per tumour and an average green/red profile was generated for each chromosome. A green/red ratio outside the limits of 1.2 or 0.8 was scored as an increase or decrease in DNA content for a single chromosome. Tumours arising in Msh2−/−, (Msh2−/−, p53+/ −), (Msh2+/−, p53−/−) and, perhaps most significantly (Msh2−/−, p53−/−) mice displayed stable genomes with few chromosomal gains or losses identified (Table 2). In contrast, tumours arising in p53−/− mice showed a number of amplified or deleted chromosomal regions.

Table 1 Analysis of thymic lymphomas in MSH2/p53 intercrossed mice by flow cytometry
Table 2 Analysis of thymic lymphomas in MSH2/p53 intercrossed mice by CGH

To investigate whether the accelerated development of tumours in (Msh2−/−, p53−/−) mice was associated with a p53-dependent increase in MSI, tumours were assessed for instability at four separate microsatellite loci. Data is summarized in Table 3. Compared to tumours arising in Msh2−/− mice, MSI was more frequently observed in tumours arising in (Msh2−/−, p53+/−) mice (P=0.008 Fisher exact test). In the smaller cohort of (Msh2−/−, p53−/−) tumours analysed a similar trend was observed, but this difference was not significant (P=0.38 Fisher exact test). These findings indicate that heterozygosity for p53 increases genomic instability at the nucleotide level on an Msh2 null background.

Table 3 MSI in thymic lymphomas from MSH2/p53 intercrossed mice

In the above study we did not demonstrate an increase in MSI in the complete absence of p53 (i.e. in tumours arising on a Msh2−/−, p53−/− background), possibly because the sample size analysed was too small. We therefore extended our analysis to primary fibroblast clones representative of each genotype combination. Primary fibroblast cultures were derived from E14 embryos and cells plated at cloning density. Two weeks later clones were picked and instability scored as before in each clone (Table 4). As predicted, the level of instability was low in wild type clones and higher in Msh2 null clones. In agreement with our analysis of tumour samples, heterozygosity for p53 significantly increased the level of instability (P>0.001, chi-squared test). Remarkably, however although the same trend was observed in the double null samples, the increase in the level of instability was only significant at the 10% level (P=0.068, chi-squared test see Table 2).

Table 4 MSI in embryonic fibroblast clones from MSH2/p53 intercrossed mice

These results raised the possibility that heterozygosity for p53 was sufficient to increase the levels of microsatellite instability, and indeed that a reduction in p53 levels has a more dramatic affect than the complete absence of p53. This was of particular interest as heterozygosity for p53 has been demonstrated to predispose to malignancy (Venkatachalam et al., 1998). We therefore assessed the status of the wild-type p53 allele in tumours arising in (Msh2−/−, p53+/−) mice. Loss of the remaining allele was not observed in any of nine thymic lymphomas examined by PCR (data not shown), demonstrating that gross chromosomal deletion of the remaining wild-type p53 allele had not occurred. Three tumours were then analysed by sequencing the ‘p53’ hotspot, exons 5–9 and a single G→A transition was observed at codon 234 in one of these tumours, predicted to give an amino acid change from methionine to isoleucine. These preliminary findings suggested that a percentage of tumours may have retained functional p53. To determine if this were the case we assessed p53 activity by analysing the levels of p53 and p21 by Western analysis 6 h following exposure to 5 Gy gamma-irradiation (Figures 2c and 3). In wild-type tissue p53 became stabilized and p21 levels were increased. No p53 was detectable in tumours arising on a p53 null background, nor was induction of p21 observed. However, induction of p21 was observed in four out of 18 tumours arising in mice heterozygous for p53, all of which showed stabilization of p53. One further tumour showed stabilization of p53 but no transactivation of p21, suggesting mutation of p53 consistent with our sequencing data. To confirm that this reflected retention of p53 activity we performed Electrophoretic Mobility Shift Assay (EMSA) analysis upon the same material, and established that p53 specific DNA binding does occur within tumours arising on a heterozygous background (Figure 2b). Taken together, these results strongly argue for retained p53 function in a significant subset of tumours arising on an Msh2 null, p53 heterozygous background.

Figure 2

Analysis of p53 status (a) p53 immunohistochemistry of fibroblast clones. Top panel, (MSH2−/−, p53+/−) clones 6 h following 10 J UV irradiation. Middle panel, mock treated (MSH2−/−, p53+/−) fibroblast clones. Bottom panel, (MSH2−/−, p53−/−) clones 6 h following 10 J UV irradiation. (b) EMSA analysis of p53 DNA binding activity in a tumour sample which showed p53 stabilization and p21 induction following DNA damage. Lanes 1–3, wild type ES cells following 10 J UV damage. Lanes 4–6, Msh2−/− p53+/− tumour sample following 5 Gy gamma irradiation. Lanes 7–9, p53 null tumour following 5 Gy gamma irradiation. Lanes 1, 2, 4, 5, 7 and 8 reflect incubation with a p53 specific DNA oligonucleotide which resulted in p53-DNA complex formation in lanes 2 and 5. Lanes 3, 6 and 9 reflect incubation with a DNA oligonucleotide to which p53 does not bind. Lanes 1, 4 and 7 reflect additional incubation with a p53 antibody which resulted in supershifts in lanes 1 and 4. (c) p53 Western analysis 6 h following 5 Gy gamma irradiation. Lanes 1, 3, 4 and 5, irradiated (Msh2−/−, p53+/−) tumours; lane 2, mock treated (Msh2−/−, p53+/−) tumour; lane 6 irradiated p53−/− tumour. Lane 7, irradiated (Msh2−/− p53+/−) pooled fibroblast clones. Lane 8, mock treated pooled fibroblast clones. In total five tumours showed p53 stabilization following DNA damage, three of which are shown in this figure (lanes 1, 3 and 4)

Figure 3

Analysis of p21 status. Western analysis 6 h following irradiation with 5 Gy gamma irradiation. Top panel; lanes 1 and 3–8 irradiated Msh2−/− p53+/− tumours. Lane 9, mock treated Msh2−/− p53+/− tumour. Lane 2, irradiated Msh2−/− p53−/− tumour. Bottom panel. Lane 1, mock treated pooled Msh2−/− p53+/− fibroblast clones. Lanes 2 and 3, pooled Msh2−/− p53+/− irradiated fibroblast clones. Lanes 4 and 5, irradiated spleen and thymus from a wild type mouse. Lane 6, irradiated p53 null tumour. Lanes 7, irradiated spleen and tumour from a Msh2−/− p53+/− mouse. Lane 8, irradiated tumour from a Msh2−/− p53+/− mouse. Lane 9 is a positive control obtained from Kim2 cells overexpressing p21

We used similar approaches to determine if p53 function had been retained in the primary fibroblast cultures obtained from p53 heterozygous, Msh2 mutant embryos. We used immunohistochemical and Western analysis to determine p53 functionality in mass cultures of primary fibroblasts. These were maintained at low density for 2 weeks prior to harvest such that they had undergone a similar number of divisions to those cells analysed as single clones for MSI. Six hours following exposure to DNA damage (10 J UVB irradiation for immunohistochemistry, 5 Gy for Western analysis) p53 null cultures did not show upregulation of p21. In contrast, wild-type and p53 heterozygous cultures showed p53 stabilization (Figure 2) and upregulation of p21 (Figure 3). Furthermore, individual immunohistochemical analysis of individual heterozygous clones showed that p53 was upregulated following DNA damage in Msh2−/−, p53+/− cultures (Figure 2a). Taken together these data argue very strongly that p53 functionality was retained in p53 heterozygous cells over the period of cloning, and therefore that the increase in microsatellite instability reported above occurs in a functional p53 heterozygous background.


Cranston et al. (1997) have previously described the phenotype of mice functionally null for both p53 and Msh2. They found that females died in utero, but that males developed normally to birth, although they subsequently succumbed to lymphoma earlier than either p53 or Msh2 singly null mice (Cranston et al., 1997). We have generated a similar intercross, using different null alleles (Clarke et al., 1993; de Wind et al., 1995). As we have previously reported, we did not observe increased rates of female embryonic lethality in p53−/−, Msh2−/− mice (Toft et al., 1998). This difference probably arises from differences in genetic background, as has recently been strongly argued by Cranston and Fishel (1999). Data summarizing survival and tumour spectrum in all the genotype groups agree well with previously published work (de Wind et al., 1995; Reitmair et al., 1995, 1996a; Purdie et al., 1994; Cranston et al., 1997). This suggests that no fundamental differences exist between the targeted alleles used in this study and those used by Cranston et al. (1997).

Accelerated tumorigenesis in (Msh2−/−, p53+/−) mice

Heterozygosity for p53 reduced survival on an Msh2−/− background, the most obvious mechanism for this is through inactivation of the remaining wild type p53 allele. However we found no evidence for this by PCR and furthermore sequencing of the ‘mutation hotspot’ (Vogelstein and Kinzler, 1992) failed to identify mutations in two out of three tumours. These preliminary observations raised the possibility that p53 function had been retained in at least a subset of tumours as has previously been reported (Venkatachalam et al., 1998) and we therefore chose to test if p21 could still be transcriptionally activated following DNA damage. This is a particularly potent test of p53 activity as upregulation of p21 in these circumstances is absolutely dependent upon p53 and because it has recently been confirmed that the transcriptional activity of p53 is essential for many p53-dependent process including apoptosis, cell cycle arrest and tumour suppression (Jimenez et al., 2000; Chao et al., 2000). Western analysis of both p53 and p21 within our samples confirmed retention of p53 activity in a proportion (25%) of tumours, demonstrating that a reduction in p53 gene dosage can be sufficient to accelerate tumorigenesis. Although 75% of Msh2−/−, p53+/− tumours showed loss of p53 functionality, LOH at the p53 allele was not observed in any of tumours analysed of this genotype. Although this analyses were performed using different tumour sets, this suggests that mutation of the remaining allele is occurring in a MMR dependent fashion. This mirrors observations made at the Apc locus in (ApcMin/+, Msh2−/−) mice, where although only 20% of tumours exhibited LOH all were characterized by inactivating mutations of Apc (Shoemaker et al., 2000). This suggests protocols relying upon screening of LOH or common mutation hotspots can considerably underestimate the presence of mutations in certain MMR deficient genetic backgrounds.

Tumorigenesis is not acclerated in (Msh2+/−, p53−/−) mice

Heterozygosity for Msh2 in a p53 null environment did not accelerate spontaneous tumorigenesis. This is perhaps not surprising since Msh2+/− cells do not show MSI, and are reported to be proficient in DNA mismatch repair (de Wind et al., 1995; Kolodner, 1995). However, this finding also suggests that the remaining wild-type Msh2 allele does not act as a target for mutation even though p53 deficiency has been shown to cause genomic instability (Bouffler et al., 1995) and DNA amplification (Yin et al., 1992). These observations can perhaps be reconciled by the fact that p53 deficiency does not necessarily lead to an increase in the in vivo spontaneous mutation frequency (Clarke et al., 1997). Heterozygosity for Msh2 did weakly influence the site of sarcoma development. This must reflect complex tissue specific gene dependency in the prevention of neoplasia, a feature previously reported in several different intercrosses (e.g. Clarke et al., 1995; Sansom and Clarke, 2000). Taken together, our survival studies indicate that heterozygosity for p53 has a profound effect upon Msh2 dependent tumorigenesis, but that the inverse is not true. This directly suggests that p53 may be involved in modulation the levels of MSI within an MMR deficient background; a hypothesis we tested by analysing levels of MSI in tumour samples.

Increased MSI conferred by heterozygosity for p53

Colorectal cancers exhibit two distinct forms of genomic instability, which have been classified by Lengauer et al. (1997) as: MIN (microsatellite instability, MSI), predicted to be associated with mutations in the MMR genes; and CIN (chromosomal instability), predicted to associated with loss of p53 and the development of aneuploidy. The data presented here directly contradicts this view as we demonstrate that, in the absence of Msh2, p53 can function to suppress MSI and that a reduction in p53 gene dosage decreases tumour latency specifically through increasing MSI.

Suppression of MSI in the absence of Msh2 appears to be a general function of p53, as it occurs not only within tumour derived samples, but also in primary fibroblasts derived from the mutant strains. Our results apparently contradict previous data obtained by Cranston et al. (1997). However, their study did not include mice heterozygous for p53, for which we see the clearest phenotype. Furthermore, where both studies compared Msh2−/− and (Msh2−/−, p53−/−) mice a similar trend to increased instability was observed in the absence of p53, although in neither study did this attain significance at the 5% level.

Two possible explanations may account for the p53 dependent increase in MSI. First, deficiency of p53 may reduce genomic instability at the nucleotide level by mediating repair, either directly or indirectly. This is supported by the observation that p53 is capable of binding insertion/deletion loops, the DNA lesion associated with MSI (Lee et al., 1995). A second possibility is that p53 may normally function by clearing (through apoptosis) cells characterized by instability. Both p53 and Msh2 have well-characterized pro-apoptotic roles following damage of mismatch type (Clarke et al., 1993; Lowe et al., 1993; Toft et al., 1999; Sansom and Clarke, 2000), and again the ability of p53 to bind insertion/deletion loops supports a direct role for p53 in lesion recognition.

Currently it is not possible to discriminate between these possibilities, but it is clear from these results that p53 status only becomes relevant to MSI in the absence of Msh2, as singly mutant p53 deficient mice are not characterized by increased MSI. This strongly suggests that p53 is redundant to the normal monitoring and clearance of MSI associated lesions, but that following loss of Msh2 the presence of normal levels of p53 becomes essential to this process.

These findings characterize an interaction between Msh2 and p53 mutations in increasing MSI and tumour development. Remarkably the largest increase in MSI occurs in p53 heterozygotes, suggesting that complete ablation of p53 leads to the upregulation of relatively efficient compensatory mechanisms, perhaps mediated by one of the growing family of p53 homologues. It seems unlikely that this compensation is mediated by p73, as doubly mutant cells have previously been reported to have lost the ability to upregulate p73 and initiate p53 independent apoptosis following challenge with cisplatin (Gong et al., 1999). In human colorectal cancer coincident mutations in both Msh2 and p53 have rarely been reported (Cottu et al., 1996), although examples of combined microsatellite and chromosomal instability do exist (Lengauer et al., 1997). One simple explanation for the absence of combined mutation is that retained heterozygosity for p53 is difficult to establish and is not routinely screened for. Indeed, the results presented above do not predict combined microsatellite and chromosomal instability, as double mutants are characterized by enhanced MSI and diminished aneuploidy.

In conclusion, these results demonstrate a novel pathway to tumorigenesis by establishing a role for p53 in maintaining integrity at the nucleotide level and they also raise the strong possibility that combined mutation of both p53 and the DNA mismatch repair proteins are currently under reported in human colorectal tumorigenesis.

Materials and methods

Generation and maintenance of Msh2/p53 mouse colonies

All mice were maintained under non-barrier conditions. The intercross was generated on an outbred background segregating for 129/Ola, Balb-c, and SWR genomes. All experimental mice were derived from a single intercross on this mixed genetic background. Mice were genotyped according to Toft et al. (1999) for both Msh2 and p53 Status. Loss of heterozygosity at the p53 locus was also scored using these primers.

Analysis of wild-type p53 allele in MSH2−/− p53+/− mice

DNA was extracted from small pieces (3×3 mm) of tumour using phenol/chloroform extraction and resuspended in 500 μl TE and p53 status determined by PCR Total RNA was extracted using 1 ml Trizol (Gibco–BRL). One μl of RNA was reverse transcribed into cDNA using oligo(dT)15 primers. PCR primers 33A and 21B (Ozbun et al., 1993) were used to amplify a 686 bp fragment of p53 cDNA which included exons 5–9. For each tumour cDNA three separate PCR reactions were performed. Each PCR product was cloned and sequenced from both directions using a Fluorescent Automated DNA Sequencer (Licor).

DNA flow cytometry

DNA content of tumours was analysed according to Vindelov et al. (1983). DNA content was analysed on a EPICS XL Flow Cytometer using Multicycle Software (Coulter).

Comparative genomic hybridization (CGH)

CGH was carried out on a series of 42 lymphomas according to the method of Kallioniemi et al. (1992), slightly modified. Briefly, DNA was labelled by incorporation of digoxygenin-11-dUTP (normal DNA) or biotin-16-dUTP (test DNA) by nick translation. Metaphase spreads were prepared from karyotypically normal E14 mouse embryonic stem cells. Slides were pretreated with 10 μg/ml RNase A at 37°C for 1 h and 100 ng/ml proteinase K solution at 37°C for 2.5 min, denatured in 70% formamide/2×SSC solution at 70°C for 3 min and dehydrated through an ethanol series. Five hundred nanograms of each labelled probe and 20 μg mouse Cot-1 DNA were denatured at 70°C for 5 min and allowed to reanneal at 37°C for 1 h before hybridization onto slides for 2 days. Detection was carried out with FITC-avidin and antidigoxygenin-rhodamine and chromosomes counterstained with DAPI as described (Kallioniemi et al., 1992). Hybridizations were analysed using Quantitative Image Processing System (QUIPS) software (Vysis Ltd, Richmond, UK). For each tumour, four to eight metaphase spreads were karyotyped manually and green/red ratio cut-off points of 1.2 and 0.8 were chosen (following analysis of normal DNA hybridized to normal karyotypes) for scoring of chromosome copy number changes.

Microsatellite analysis of mouse tumours

Four microsatellite loci were chosen for analysis: D1Mit4, D7Mit17, D10Mit2 and D14Mit15. Primers sequences and reaction conditions were obtained from the Whitehead Institute for Biomedical Research/Massachusetts Institute of Technology Centre for Genome Research website at http://www.genome.wi.mit.edu. MSI was assessed by comparison of normal and tumour PCR products on a 6% denaturing polyacrylamide gel with silver staining, as described elsewhere (Bubb et al., 1996).

Functional p53 analysis

p53 immunohistochemistry was performed on fibroblast clones grown on chamber slides plated at a cloning density. Slides were stained using the CM5 p53 antibody (Novacastra) at 1 : 250 dilution for 2 h. For p53 and p21 Western analysis whole animals or cultures were irradiated and material harvested. Protein was extracted and run on SDS-polyacrylamide gels. Equal loading of gels was confirmed by staining the blotted gel with Coomassie blue (0.1%). Membranes were incubated in blocking buffer (5% Marvel in TBS with 0.1% Tween 20) for 1 h. Antibodies were obtained as follows: p21, Santa Cruz; p53 CM5 antibody, Novacastra. Specifically bound antibody was detected with horseradish peroxidase-conjugated secondary antibodies and ECL (Amersham). For the p53 EMSA analyses the methodology was as described in Venkatachalam (1998). The p53 supershift antibody used was Ab-1(PAB421) from Santa Cruz.


  1. Armstrong JF, Kaufman MH, Harrison DJ, Clarke AR . 1995 Curr. Biol. 5: 931–936

  2. Bouffler SD, Kemp CJ, Balmain A, Cox R . 1995 Cancer Res. 55: 3883–3889

  3. Bubb VJ, Curtis LJ, Cunningham C, Dunlop MG, Carothers AD, Morris RG, White S, Bird CC, Wyllie AH . 1996 Oncogene 12: 2641–2649

  4. Carder PJ, Wyllie AH, Purdie CA, Morris RG, White S, Piris J, Bird CC . 1993 Oncogene 8: 1397–1401

  5. Chao C, Saito S, Kang J, Anderson CW, Appella E, Xu Y . 2000 EMBO J. 19: 4967–4975

  6. Clarke AR, Purdie CA, Harrison DJ, Morris RG, Bird CC, Hooper ML, Wyllie AH . 1993 Nature 362: 849–852

  7. Clarke AR, Gledhill S, Hooper ML, Bird CC, Wyllie AH . 1994 Oncogene 9: 1767–1773

  8. Clarke AR, Cummings MC, Harrison DJ . 1995 Oncogene 11: 1913–1920

  9. Clarke AR, Howard LA, Harrison DJ, Winton DJ . 1997 Oncogene 14: 2015–2018

  10. Cottu PH, Muzeau F, Estreicher A, Flejou J-F, Iggo R, Thomas G, Hamelin R . 1996 Oncogene 13: 2727–2730

  11. Cranston A, Bocker T, Reitmair A, Palazzo J, Wilson T, Mak T, Fishel R . 1997 Nature Genet 17: 114–118

  12. Cranston A, Fishel R . 1999 Mamm. Genome 10: 1020–1022

  13. Degtyareva N, Subramanian D, Griffith JD . 2001 J. Biol. Chem. 276: 8778–8784

  14. de Wind N, Dekker M, Berns A, Radman M, te Riele H . 1995 Cell 82: 321–330

  15. Duckett DR, Bronstein SM, Taya Y, Modrich P . 1999 Proc. Natl. Acad. Sci. USA 96: 12384–12388

  16. Donehower LA, Harvey M, Slagle BL, McArthur MJ, Montgomery CA, Butel JS, Bradley A . 1992 Nature 356: 215–221

  17. Dudenhoffer C, Rohaly G, Will K, Deppert W, Wiesmuller L . 1998 Mol. Cell. Biol. 18: 5332–5342

  18. Gong JG, Costanzo A, Yang HQ, Melino G, Kaelin WG, Levrero M, Wang JYJ . 1999 Nature 399: 806–809

  19. Hartwell L . 1992 Cell 71: 543–546

  20. Harvey M, Sands AT, Weiss RS, Hegi ME, Wiseman RW, Pantazis P, Giovanella BC, Tainsky MA, Bradley A, Donehower LA . 1993 Oncogene 8: 2457–2467

  21. Hawn MT, Umar A, Carethers JM, Marra G, Kunkel TA, Boland CR, Koi M . 1995 Cancer Res. 55: 3721–3725

  22. Jimenez GS, Nister M, Stommel JM, Beeche M, Barcase ER, Zhang X-Q, O'Gorman S, Wahl G . 2000 Nature Gen. 26: 37–43

  23. Kallioniemi OP, Kallioniemi A, Sudar D, Rutovitz D, Gray JW, Waldman F, Pinkel D . 1992 Semin. Cancer Biol. 4: 41–46

  24. Kolodner RD . 1995 Trends Biochem. Sci. 20: 397–401

  25. Lee S, Elenbaas B, Levine A, Griffith J . 1995 Cell 81: 1013–1020

  26. Lengauer C, Kinzler KW, Vogelstein B . 1997 Nature 386: 623–627

  27. Liu B, Nicolaides NC, Markowitz S, Willson JK, Parsons RE, Jen J, Papadopoulos N, Peltomaki P, de la Chapelle A, Hamilton SR, Vogelstein B, Kinzler KW . 1995 Nature Genet. 9: 48–55

  28. Lowe SW, Schmitt EM, Smith SW, Osborne BA, Jacks T . 1993 Nature 362: 847–849

  29. Mummenbrauer T, Janus F, Muller B, Wiesmuller L, Deppert W, Grosse F . 1996 Cell 85: 1089–1099

  30. Ozbun MA, Jerry J, Kittrell FS, Medina D, Butel JS . 1993 Cancer Res. 53: 1646–1652

  31. Purdie CA, Harrison DJ, Peter A, Dobbie L, White S, Howie SE, Salter DM, Bird CC, Wyllie AH, Hooper ML . 1994 Oncogene 9: 603–609

  32. Rafferty JA, Clarke AR, Sellapan D, Koref MS, Frayling IM, Margison GP . 1996 Oncogene 12: 693–697

  33. Reichmann A, Martin P, Levin B . 1981 Int. J. Cancer 28: 431–440

  34. Reitmair AH, Schmits R, Ewel A, Bapat B, Redston M, Mitri A, Waterhouse P, Mittrucker HW, Wakeham A, Liu B, Thomason A, Griesser H, Gallinger S, Ballhausen WG, Fishel R, Mak TW . 1995 Nature Genet 11: 64–70

  35. Reitmair AH, Redston M, Cai JC, Chuang TC, Bjerknes M, Cheng H, Hay K, Gallinger S, Bapat B, Mak TW . 1996 Cancer Res. 56: 3842–3849

  36. Sansom OJ, Clarke AR . 2000 Mut. Res. Fund. Mol. Mechan. Mutagen. 452: 149–162

  37. Sansom OJ, Toft NJ, Winton DJ, Clarke AR . 2001 Oncogene 20: 3580–3584

  38. Shoemaker AR, Haigis KM, Baker SM, Dudley S, Liskay RM, Dove WF . 2000 Oncogene 19: 2774–2779

  39. Toft NJ, Arends MJ, Wyllie AH, Clarke AR . 1998 Nature Genet. 18: 17

  40. Toft NJ, Winton DJ, Kelly J, Howard LA, Dekker M, Riele HT, Arends MJ, Wyllie AH, Margison GP, Clarke AR . 1999 Proc. Natl. Acad. USA 96: 3911–3915

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

  42. Vindelov LL, Christensen IJ, Nissen NI . 1983 Cytometry 3: 323–327

  43. Vogelstein B, Kinzler KW . 1992 Cell 70: 523–526

  44. Vogelstein B, Kinzler KW . 1993 Trends Genet. 9: 138–141

  45. Yin Y, Tainsky MA, Bischoff FZ, Strong LC, Wahl GM . 1992 Cell 70: 937–948

Download references


We thank John Verth and his staff for animal care, Jennifer Doig for technical assistance and Nathalie Sphyris for help with the p53 immunohistochemistry and EMSA. This work was supported by grants from the Cancer Research Campaign. NJ Toft is a Leckie-Mactier Research Fellow supported by the Faculty of Medicine, University of Edinburgh. AR Clarke is a Royal Society Research Fellow.

Author information

Correspondence to Alan R Clarke.

Rights and permissions

Reprints and Permissions

About this article


  • mismatch repair
  • p53
  • tumorigenesis
  • microsatellite instability
  • apoptosis

Further reading