Oncogene (2004) 23, 1954–1956
& 2004 Nature Publishing Group All rights reserved 0950-9232/04 $25.00 www.nature.com/onc
A TP53 polymorphism is associated with increased risk of colorectal cancer and with reduced levels of TP53 mRNA Federica Gemignani1,2,5, Victor Moreno3,5, Stefano Landi1,2,5, Norman Moullan1, Ame´lie Chabrier1, Sara Gutie´rrez-Enrı´ quez1, Janet Hall1, Elisabeth Guino3, Miguel Angel Peinado4, Gabriel Capella3 and Federico Canzian*,1 1
International Agency for Research on Cancer (IARC), 150, Cours Albert Thomas, Lyon 69372, France; 2Dipartimento di Scienze dell’Uomo e Ambiente, University of Pisa, 56100, Italy; 3Institut Catala d’Oncologia, Hospitalet, Barcelona, Spain; 4Institut de Recerca Oncologica, Hospitalet, Barcelona, Spain
We undertook a case–control study to examine the possible associations of the TP53 variants Arg4Pro at codon 72 and p53PIN3, a 16 bp insertion/duplication in intron 3, with the risk of colorectal cancer (CRC). The p53PIN3 A2 allele (16 bp duplication) was associated with an increased risk (OR 1.55, 95% CI 1.102.18, P ¼ 0.012), of the same order of magnitude as that observed in previous studies for other types of cancer. The Pro72 allele was weakly associated with CRC (OR ¼ 1.34, 95% CI 0.981.84, P ¼ 0.066). The possible functional role of p53PIN3 was investigated by examining the TP53 mRNA transcripts in 15 lymphoblastoid cell lines with different genotypes. The possibility that the insertion/deletion could lead to alternatively spliced mRNAs was excluded. However, we found reduced levels of TP53 mRNA associated with the A2 allele. In conclusion, the epidemiological study suggests a role for p53PIN3 in tumorigenesis, supported by the in vitro characterization of this variant. Oncogene (2004) 23, 1954–1956. doi:10.1038/sj.onc.1207305 Published online 1 December 2003 Keywords: TP53; colorectal cancer; polymorphism; mRNA levels
A total of 14 polymorphisms have been identified in the TP53 gene (http://www.iarc.fr/p53/Index.html), the most studied being Arg72Pro, p53PIN3 (where the A2 allele carries a duplication of 16 bp in intron 3), and MspI (A4G, intron 6). Clues to their biological function have only been reported for the Arg72Pro variant. The wild-type form was found to be more susceptible than the variant Pro72 to degradation by the E6 onco-protein following human papillomavirus infection (Storey et al., 1998), and more efficient in triggering cellular apoptosis due to a greater ability to localize to the mitochondria (Dumont et al., 2003). *Correspondence: F Canzian,Genome Analysis Group, IARC; E-mail:
[email protected] 5 These authors contributed equally to the manuscript. Received 18 June 2003; revised 21 July 2003; accepted 22 October 2003
We performed a case–control study on 374 cases of colorectal cancer (CRC) and 322 controls, in order to examine possible associations between the Arg72Pro or the p53PIN3 A1/A2 alleles and the risk of cancer. Incident CRC cases (histologically confirmed) attending the Bellvitge Hospital (Barcelona, Spain) over the period 1996–1998 made up the case group. Controls were patients admitted to several departments of this hospital over the same period. Full details of the cohort are published elsewhere (Landi et al., 2003). The study was approved by the local Ethical Committees. All subjects gave written informed consent according to the Helsinki declaration. The p53PIN3 variant was genotyped as previously reported (Lazar et al., 1993), while the Arg72Pro variant was analysed using a primer extension approach (To˜nisson et al., 2002). The MspI intron 6 polymorphism was not investigated because it has been reported to be in complete linkage disequilibrium (LD) with the p53PIN3 variant (Weston et al., 1997). We found that the p53PIN3 genotypes A1/A2 and A2/A2 were statistically associated with an increased risk of CRC. The association of the Pro72 allele with CRC was weaker and statistically borderline (Table 1). Haplotype frequencies were reconstructed by use of the EM algorithm (Excoffier et al., 1995) and the Pro72 and A2 alleles were found in strong but not complete LD (r2 ¼ 0.63, Po0.001), according to the Genotype Transposer software (Cox and Canzian, 2001). When the analysis was mutually adjusted for each polymorphism, the association of p53PIN3 remained significant (Ptrend ¼ 0.042), while the Pro72 variant was clearly unrelated to CRC (Ptrend ¼ 0.96). Previous studies have documented associations between these variants and cancer. In particular, the A2 allele has been found to be associated with increased risk of cancers of the lung (Wu et al., 2002), breast (Weston et al., 1997; Powell et al., 2002; Wang-Gohrke et al., 2002), and ovary (Runnebaum et al., 1995; Wang-Gohrke et al., 1999). Exon 3 and intron 3 of TP53 are only 21 and 112 bp, respectively, and it is possible that an increase of 15% in the length of the intron could alter mRNA splicing, thus affecting p53 function. However, publicly available databases (http://www.ncbi.nih.gov/IEB/Research/ Acembly/) do not detail splicing variants affecting exon
Association of TP53 polymorphism with colorectal cancer F Gemignani et al
1955 Table 1 Odds ratios and confidence intervals for carriers of the 16 bp insertion in the intron 3 of the TP53 gene (p53PIN3) and for Arg72Pro: data adjusted for sex and age Controls n
Cases n
Odds ratios (95% CI)
p53PIN3 A1/A1 A1/A2 A2/A2 Trend test P-value
248 68 6
256 108 10
1.00 1.54 (1.082.20) 1.62 (0.574.62) 0.014
P53PIN3 (grouped) A1/A1 A1/A2+A2/A2 P-value
248 74
256 118
1.00 1.55 (1.102.18) 0.012
Arg72Pro Arg/Arg Arg/Pro Pro/Pro Trend test P-value
202 95 19
201 133 18
1.00 1.41 (1.021.97) 0.98 (0.501.93) 0.18
Arg72Pro (grouped) Arg/Arg Arg/Pro+Pro/Pro P-value
202 114
201 151
1.00 1.34 (0.981.84) 0.066
A1 ¼ wild-type allele, no insertion; A2 ¼ variant allele carrying the p53PIN3 allele. ORs are calculated setting the wild-type homozygote genotype as the reference group. CI ¼ confidence interval, n ¼ number
3. Moreover, analysis of the sequence by using a couple of algorithms that predict splice sites (Brunak et al., 1991; Hebsgaard et al., 1996; Rogan et al., 1998) (http://genome.cbs.dtu.dk/services/NetGene2/; http:// www.lecb.ncifcrf.gov/Btoms/delilaserver.html) did not support the hypothesis of differential splicing in this region. In order to experimentally confirm the data from in silico analyses, we analysed the mRNA of 15 Epstein– Barr virus (EBV) immortalized lymphoblastoid cell lines carrying either the A1/A1, A1/A2, or A2/A2 variants (5 of each genotype), and for which the codon 72 status was also known. A fragment containing exon 3 was amplified using primers located in exon 2 (50 -CCCCTCTGAGTCAGGAAACA-30 ) and exon 4 (50 -GGGACAGCATCAAATCATCC-30 ). Analysis by agarose gel electrophoresis of this RT–PCR product showed only one band of the expected size (107 bp) in all the cell lines (data not presented). Thus, we concluded that the polymorphism is not associated with altered splicing of the pre-mRNA. To investigate further possible influences of this polymorphism on TP53 expression, we measured the basal level of mRNA in the same 15 lymphoblastoid cell lines analysed for splice variants. We employed the reverse transcriptase method followed by a real-time PCR (on an Applied Biosystems 7900HT instrument), using the 50 nuclease assay, following the protocol recommended by the manufacturer. The level of the TP53 cDNA was normalized to that of b-actin (TP53 primers and probes obtained from Applied Biosystems, ‘Assays-On-Demand’; b-actin primers and probes designed in-house using the Primer Express software from Applied Biosystems). The set of 15 mRNAs were
extracted twice using standard methods, and, for each sample, six replicates of the cDNAs were analysed on the same plate. The threshold cycle (Ct), defined as the cycle where the amplification of the PCR product enters the exponential phase, was determined for each gene by plotting the fluorescence level vs the cycle number on a logarithmic scale (Figure 1). The relative expression level of the TP53 gene in the different cell lines was then estimated by calculating the DCt value, defined as the difference in the Ct value for the target (TP53) and the reference gene (b-actin). The DCt is inversely proportional to the TP53 mRNA level, with a high DCt value corresponding to a lower mRNA level. The average and standard deviation of the six measurements for each sample from the two independent RNA extractions are reported in Table 2. The analysis of variance with nested effects showed that the amount of TP53 mRNA expressed in the cell lines with A1/A1 genotype is higher than that in the cell lines with either the A1/A2 or the A2/A2 genotypes. No statistical difference was found between the A1/A2 and A2/A2 genotypes. Similar differences in mRNA levels were found when the analysis was carried out based on the codon 72 genotype: the strong LD between the p53PIN3 and codon 72 variants can explain this result. In conclusion, we have shown that the A2 allele is associated with an increased risk of CRC and reduced basal levels of TP53 mRNA in EBV immortalized lymphoblastoid cell lines. One possible mechanism explaining this finding could reside in a reduced stability of the pre-mRNA carrying the 16 bp insertion. However, the possibility that the polymorphism is in LD with
Figure 1 Real-time RT–PCR amplification plot. The logarithmic scale of Rn, a measure of the fluorescence intensity, is reported on the Y-axis, the number of PCR cycles on the X-axis. b-Actin and TP53 mRNAs for the A1/A1 and A2/A2 genotypes of TP53 intron 3 polymorphism are shown as dark and light lines, respectively. The exponential phase is reached at a higher cycle number in the samples with A2/A2 genotype compared to the A1/A1 genotype Oncogene
Association of TP53 polymorphism with colorectal cancer F Gemignani et al
1956 Table 2 Summary statistics of the DCt values obtained after six measurements for each cell line in two independent experiments (expt) Cell line
p53PIN3 genotype
Codon72 genotype
DCt 1st expt. (means7s.d.)
DCt 2nd expt. (means7s.d.)
1 2 3 4 5 Pooled
A1/A1 A1/A1 A1/A1 A1/A1 A1/A1
Arg/Arg Arg/Arg Arg/Arg Arg/Arg Arg/Pro
4.9570.13 5.3570.12 5.0670.20 5.1870.12 5.0170.15 5.11a70.20
5.0670.21 5.7370.16 5.4470.18 5.3870.13 5.5170.18 5.4270.27
6 7 8 9 10 Pooled
A1/A2 A1/A2 A1/A2 A1/A2 A1/A2
Arg/Pro Pro/Pro Arg/Arg Arg/Pro Arg/Pro
5.2970.10 5.8270.10 5.7770.13 5.3870.14 5.6070.11 5.57b70.24
6.1370.31 6.8270.16 6.0470.11 6.2370.11 6.1870.27 6.2870.34
11 12 13 14 15 Pooled
A2/A2 A2/A2 A2/A2 A2/A2 A2/A2
Pro/Pro Arg/Pro Pro/Pro Pro/Pro Pro/Pro
4.8970.31 5.9770.17 5.7370.11 5.5670.07 5.5970.17 5.55c70.40
5.9770.14 6.2470.21 5.8870.18 6.5870.22 5.8170.19 6.0970.34
Contrasts: a vs b, Po0.001; a vs c, Po0.001; b vs c, P ¼ 0.49. The results of the second experiment confirmed the first experiment. The DCt value is inversely proportional to the TP53 mRNA level. s.d. ¼ standard deviation
others, hereto unknown, having a more direct role in regulating p53 expression, cannot be ruled out. Neither can it be formally excluded that this reduced basal mRNA level is associated with the presence of EBV-
encoded proteins, which specifically interact with the A2 allele, reducing the stability of this transcript. However, a series of measurements carried out on mRNA extracted from whole blood of individuals with the A1/A2 and A1/A1 genotypes showed a trend similar to that observed in cell lines (data not shown), suggesting that EBV immortalization does not influence the variation of TP53 mRNA levels observed. Though the epidemiological data link the risk to the A2 allele of p53PIN3, the data from the available cell lines do not allow us to establish whether the A2 variant alone influences the TP53 mRNA stability, or whether this effect requires the presence of the Pro72 variant. Based on the data presented here, the A2 variant allele is associated with a high population attributable risk, influencing as much as 11.2% of CRC in the general population. Considering that CRC is one of the most common cancers, this risk factor could have a sizable impact on public health. Acknowledgements Federica Gemignani is a recipient of a fellowship from the International Association for the Study on Lung Cancer (IASLC), part of the Cancer Research Foundation of America (CRFA). Stefano Landi is a recipient of a Marie Curie fellowship (HPMFCT-2000-00483) from the European Commission. S Gutie`rrez-Enriquez and N Moullan are recipients of Special Training Awards from the International Agency for Research on Cancer. This work was partially funded by grants from the Maraton of TV3 48/95, Fondo de Investigaciones Sanitarias FIS 96/0797, FIS 00/0027, FIS 01/1264 and SAF 00/ 81, and Association pour la Recherche sur le Cancer #7478.
References Brunak S, Engelbrecht J and Knudsen S. (1991). J. Mol. Biol., 220, 49–65. Cox DG and Canzian F. (2001). Bioinformatics, 17, 738–739. Dumont P, Leu JI, Della Pietra AC, George DL and Murphy M. (2003). Nat. Genet., 33, 357–365. Excoffier L and Slatkin M. (1995). Mol. Biol. Evol., 12, 921–927. Hebsgaard SM, Korning PG, Tolstrup N, Engelbrecht J, Rouze P and Brunak S. (1996). Nucleic Acids Res., 24, 3439–3452. Landi S, Moreno V, Gioia-Patricola L, Guino E, Navarro M, de Oca J, Capella G and Canzian F. (2003). Cancer Res., 63, 3560–3566. Lazar V, Hazard F, Bertin F, Janin N, Bellet D and Bressac B. (1993). Oncogene, 8, 1703–1705. Powell BL, van Staveren IL, Roosken P, Grieu F, Berns EM and Iacopetta B. (2002). Carcinogenesis, 23, 311–315. Rogan PK, Faux BM and Schneider TD. (1998). Hum. Mutat., 12, 153–171. Runnebaum IB, Tong XW, Konig R, Zhao H, Korner K, Atkinson EN, Kreienberg R, Kieback DG, Hong Z and Zhao H. (1995). Lancet, 345, 994.
Oncogene
Storey A, Thomas M, Kalita A, Harwood C, Gardiol D, Mantovani F, Breuer J, Leigh IM, Matlashewski G and Banks L. (1998). Nature, 393, 229–234. To˜nisson N, Zernant J, Kurg A, Pavel H, Slavin G, Roomere H, Meiel A, Hainaut P and Metspalu A. (2002). Proc. Natl. Acad. Sci. USA, 99, 5503–5508. Wang-Gohrke S, Becher H, Kreienberg R, Runnebaum IB and Chang-Claude J. (2002). Pharmacogenetics, 12, 269–272. Wang-Gohrke S, Weikel W, Risch H, Vesprini D, Abrahamson J, Lerman C, Godwin A, Moslehi R, Olipade O, Brunet JS, Stickeler E, Kieback DG, Kreienberg R, Weber B, Narod SA and Runnebaum IB. (1999). Br. J. Cancer, 81, 179–183. Weston A, Pan CF, Ksieski HB, Wallenstein S, Berkowitz GS, Tartter PI, Bleiweiss IJ, Brower ST, Senie RT and Wolff MS. (1997). Cancer Epidemiol. Biomarkers Prev., 6, 105–112. Wu X, Zhao H, Amos CI, Shete S, Makan N, Hong WK, Kadlubar FF and Spitz MR. (2002). J. Natl. Cancer Inst., 94, 681–690.
