- Email: [email protected]
Possible association of p53 codon 72 polymorphism with susceptibility to adult and pediatric high-grade astrocytomas
Molecular Brain Research 137 (2005) 98 – 103 www.elsevier.com/locate/molbrainres
Research report
Possible association of p53 codon 72 polymorphism with susceptibility to adult and pediatric high-grade astrocytomas Preeti Parhara, Rona Ezera, Yongzhao Shaob,c, Jeffrey C. Allend, Douglas C. Millera,b, Elizabeth W. Newcomba,b,* a
Department of Pathology, New York University School of Medicine, 550 First Avenue, MSB531, New York, NY 10016, USA b New York University Cancer Institute, New York, NY 10016, USA c Department of Environmental Medicine, New York University School of Medicine, New York, NY 10016, USA d Department of Pediatrics, Division of Pediatric Hematology/Oncology, New York University School of Medicine, New York, NY 10016, USA Accepted 13 February 2005 Available online 21 March 2005
Abstract Polymorphisms in codon 72 of the p53 tumor suppressor gene have been associated with susceptibility to human cancer. We wished to evaluate whether variant allelic forms of the p53 protein were associated with brain tumors. In this study, we scored 135 brain tumor samples (92 adult and 43 pediatric cases consisting of 64 high-grade astrocytomas and 71 non-astrocytomas) for the P53 Arg72Pro polymorphisms. Our data show that the genotype frequencies of P53 Arg72Pro vary not only between patients with brain tumors and controls, but also between different histological subtypes of brain tumors. Specifically, we found (i) that the genotype distributions of the P53 Arg72Pro between all brain tumors and controls were statistically significant (P b 0.001) as well as their variant allele frequencies between cases and controls (P b 0.001); (ii) that there was a significant increase in the Arg/Pro heterozygous genotype among high-grade astrocytomas compared with non-astrocytomas (P = 0.002); and (iii) that there was a significant increase in the Arg/Pro heterozygous genotype among high-grade astrocytomas containing transdominant as well as recessive p53 mutations compared with controls (P = 0.002). Our results suggest a possible association between P53 Arg72Pro polymorphisms and susceptibility to brain tumors, particularly high-grade astrocytomas. D 2005 Elsevier B.V. All rights reserved. Theme: Disorders of the nervous system Topic: Brain tumors Keywords: p53 codon 72 polymorphisms; High-grade astrocytomas
1. Introduction The most common mutation found in human cancers occurs in the p53 tumor suppressor gene [8,14]. The p53 gene encodes a nuclear phosphoprotein involved in cell cycle arrest, apoptosis, and genetic stability [1,14]. The gene consists of three functional domains: an N-terminal transactivating domain, a central DNA binding domain, and a C* Corresponding author. Department of Pathology, New York University School of Medicine, 550 First Avenue, MSB531, New York, NY 10016, USA. Fax: +1 212 263 8211. E-mail address: [email protected] (E.W. Newcomb). 0169-328X/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.molbrainres.2005.02.016
terminal oligomerization domain [1,14]. The N-terminal acidic domain induces transcription of genes. Examples of p53 inducible genes upregulated in response to DNA damage include GADD45, which is involved in DNA nucleotide excision repair, and p21, which inhibits several cyclin-dependent kinases to stop the G1/S phase transition in the cell cycle and thus halt cellular division [14]. Apoptosis can be triggered by p53-mediated transcriptional upregulation of bax and insulin-like growth factor-binding protein-3 (IGF-BP3), two genes involved in the apoptotic pathway [14]. The central DNA binding domain has contact points for consensus DNA binding sequences. Finally, the Cterminal domain is important for oligomerization. To dsenseT
P. Parhar et al. / Molecular Brain Research 137 (2005) 98–103
DNA damage the p53 protein must be in a conformationally stable tetramer. Some mutations decrease the ability to form a stable tetramer rendering the molecule non-functional. These mutations are referred to as dtransdominantT or ddominant-negativeT mutations since they are able to inactivate the protein by forming a heterotetramer complex with the wild-type p53 protein encoded by the remaining copy of the normal allele. In contrast, drecessiveT mutations are unable to alter the conformation of the p53 protein in the presence of wild-type p53 protein and the protein retains its function. Predisposition to many adult cancers has now been associated with the inheritance of polymorphisms in genes, singly or in combination, that may make an individual more susceptible to malignant transformation [11]. For the p53 tumor suppressor gene a polymorphism has been identified at codon 72 within exon 4: Arg, encoded by CGC, and Pro, encoded by CCC [17]. The amino acid substitution at codon 72 affects the primary structure of p53 that results in functional differences between the different alleles [4,23]. The Pro allele is associated with increased transcriptional activity compared to the Arg allele because it binds more tightly to the transcription factors TAFII32 and TAFII70 [23]. More recently it was found that the Arg allele induced apoptosis fivefold more than the Pro allele, due to its ability to localize to the mitochondria and bind with the MDM2 protein more avidly [4]. In addition, it has been shown in squamous cell carcinomas that the p53 codon 72 polymorphism influences clinical response to DNA-damaging chemo- and radiation-therapy [2]. These differences between the biological functions of the different p53 variants may have implications not only for increasing the risk for cancer in certain individuals, but also for influencing overall survival in patients with cancer [2,11]. Although the frequency and spectrum of mutations in the p53 gene has been extensively studied in brain tumors, the p53 codon 72 polymorphism has only been examined in a few cases [8,22]. The purpose of this study was to genotype brain tumors for the different p53 codon 72 polymorphic variants and determine the distribution of genotype frequencies. In order to evaluate the question of whether variant allelic forms of the p53 gene may be associated with brain tumors, we compared the genotype frequencies and the ratio of the variant alleles in our patients with brain tumors of different histological subtypes to that reported for a control population.
2. Materials and methods 2.1. Tumor samples Tumor samples were obtained from patients who had surgery at New York University Medical Center (NYUMC) or Beth Israel Medical Center (BIMC) between 1986 and
99
2001. DNA was extracted and banked. These samples have been used in several other genetic studies previously reported [5,7,12,18,21]. Histopathologic grading of astrocytomas was done in accordance with the WHO classification scheme except that necrosis was required for a diagnosis of glioblastoma [10]. For this study of codon 72 p53 polymorphisms, limited to Caucasian patients, we retrieved 135 cases from our tumor bank. DNA was extracted from frozen or paraffin-embedded tissues as described [20]. 2.2. Patient population Table 1 summarizes the brain tumor samples used and some of the demographics of the study population. Of the 135 cases, 43 were from pediatric patients and 92 were from adult patients. There were 64 high-grade astrocytomas (consisting of 9 anaplastic astrocytoma and 54 glioblastomas), and 71 were non-astrocytomas (consisting of oligodendrogliomas, oligoastrocytomas, gangliogliomas, ganglioneurocytomas, and glioneurocytomas). Of the 43 pediatric patients, an equal number of samples was analyzed from patients 0–9 and 10–19 years of age. Of the 92 adult patients, the peak age for astrocytomas was 50–79, while the peak age for non-astrocytomas was 20–49. The distribution of gender among patients with astrocytomas and nonastrocytomas was identical, consisting of 66% males and 34% females. 2.3. Genotype detection PCR followed by restriction enzyme digestion was used to assess p53 codon 72 polymorphisms with protocols described previously [5,13]. The PCR product (10 Al) was digested overnight with BstUI (NE Biolabs) to determine the amino acid at codon 72 according to the manufacturer’s directions. A restriction enzyme premix (10 Al) containing 10 U of enzyme and 2 NEBuffer 2, made up of 100 mM Tris–HCl and 1 mM dithiothreitol (pH 7.0), was added to the PCR product for a final volume of 20 Al. For all assays, the amplified PCR fragments were
Table 1 Brain tumor samples and demographics of study population Variable
No.
Astrocytomas
Non-astrocytomas
Tumor samples Age at diagnosis Pediatric cases 0–9 10–19 Adult cases 20–49 50–79 Sex Male Female
135
64 (47%)
71 (53%)
26 15 11 38 13 25
17 6 11 54 39 15
43
92
89 46
(58%) (42%) (34%) (66%)
42 (66%) 22 (34%)
(35%) (65%) (72%) (28%)
47 (66%) 24 (34%)
100
P. Parhar et al. / Molecular Brain Research 137 (2005) 98–103
separated on 3.5% agarose gels and visualized by staining with GelStar nucleic acid gel stain (BioWhittaker Molecular Applications). Gels were photographed using a Mitsubishi P91 video copy processor. Genotypes were scored as homozygous for the Pro allele if the 199-bp PCR fragment after digestion remained uncut or homozygous for the Arg allele if after enzyme digestion two bands (113 and 86 bp) were observed. Gels with three DNA fragments (199, 113, and 86 bp) were scored as heterozygotes. 2.4. Statistical analysis Observed frequencies of genotypes in brain tumors were compared to controls using the chi-square or Fisher’s exact test when expected frequencies were small. The statistical computer software S-Plus 2000 was used to perform all statistical analyses (Insightful Corporation). Statistical significance was set at P b 0.05.
3. Results 3.1. Frequency of P53 codon Arg72Pro polymorphisms in brain tumor samples Brain tumor samples were evaluated for the P53 codon Arg72Pro polymorphism. The substitution of Arg for Pro at codon 72 creates a unique BstUI restriction site. Fig. 1 shows DNAs from representative samples incubated without ( ) and with (+) the restriction enzyme. Genotypes were scored as homozygous for the Arg allele if the 199bp DNA fragment after enzyme digestion gave bands of 113 and 86 bp (lane 2) or homozygous for the Pro allele if the PCR fragment after digestion remained uncut (lane 6). Presence of all three DNA fragments indicated a heterozygote (lane 4).
Fig. 1. Detection of P53 Arg72Pro polymorphisms. DNAs from representative brain tumor samples were scored for P53 Arg72Pro polymorphisms by PCR-RFLP analysis. PCR products were incubated without ( ) or with (+) the BstUI restriction enzyme and run side-by-side to better evaluate the extent of enzyme digestion. Samples that had the Arg (CGC) at codon 72 gained a unique BstUI cleavage site. Lane 2 contains DNA from an Arg/ Arg homozygote that is completely digested into two fragments of 113 and 86 bp. Lane 4 contains DNA from an Arg/Pro heterozygote that after digestion shows three bands of 199, 113, and 86 bp. Lane 6 contains DNA from a Pro/Pro homozygote that remains completely undigested to give a single band of 199 bp.
Table 2 Frequency of P53 Arg72Pro polymorphisms in brain tumor samples Genotype
Group 1 No.
Arg/Arg Arg/Pro Pro/Pro Pro allele frequency Arg/Arg Arg/Pro Pro/Pro Pro allele frequency Arg/Arg Arg/Pro Pro/Pro Pro allele frequency Arg/Arg Arg/Pro Pro/Pro Pro allele frequency
Group 2 (%)
Controlsa 72 (62) 42 (36) 3 (3) 0.21 Controls 72 (62) 42 (36) 3 (3) 0.21 Controls 72 (62) 42 (36) 3 (3) 0.21 All Astrocytomas 8 (12.5) 55 (86) 1 (1.5) 0.45
No.
P value* (%)
All brain tumors 38 (28) 94 (69) 3 (2) 0.37 All astrocytomas 8 (12.5) 55 (86) 1 (1.5) 0.45 All non-astrocytomas 30 (42) 39 (55) 2 (3) 0.30 All non-astrocytomas 30 (42) 39 (55) 2 (3) 0.30
b0.001
b0.001 b0.001
b0.001 0.03
0.04 b0.001
0.02
a
Control genotypes and allele frequencies as described by Weston et al. [25]. * P values were obtained using Fisher’s exact test or chi-square test.
As shown in Table 2, the frequencies of the Arg/Arg, Arg/Pro, and Pro/Pro genotypes among control cases were 62%, 36% and 3%, respectively, resulting in a Pro allele frequency of 0.21 as described [25]. The distribution of these genotype frequencies was in agreement with those calculated from the Hardy–Weinberg equilibrium model for controls (P N 0.25). We next analyzed the distribution of P53 Arg72Pro genotypes in all brain tumors compared to the controls. The frequencies of the Arg/Arg, Arg/Pro, and Pro/Pro genotypes among all brain tumor cases were 28%, 69%, and 2%, respectively, which was significantly different from the genotypes observed in the controls (P b 0.001). Likewise, the difference in the distribution of the variant Pro allele between controls and brain tumor cases was also statistically significant (P b 0.001). We next stratified the brain tumor samples by histological subtype into astrocytomas and non-astrocytomas and evaluated their genotype frequencies compared to the controls. The frequencies of the Arg/Arg, Arg/Pro, and Pro/Pro genotypes among all astrocytomas were 12.5%, 86%, and 1.5%, respectively, which was significantly different from the genotypes observed in the controls (P b 0.001). Similarly, the difference in the distribution of the variant Pro allele between controls and all astrocytomas was also statistically significant (P b 0.001). Since age can be a confounding variable, we also stratified the high-grade astrocytomas by age and compared the frequencies of the genotypes between pediatric and adult patients. The frequencies of the genotypes and the distribution of the variant Pro allele were similar in both age groups (data not shown). For non-astrocytomas, the
P. Parhar et al. / Molecular Brain Research 137 (2005) 98–103
frequencies of the Arg/Arg, Arg/Pro, and Pro/Pro genotypes was 42%, 55% and 3%, respectively, which was marginally significant from the genotypes observed in the controls (P = 0.03). Finally, we compared the distribution of genotypes between astrocytomas and non-astrocytomas. The frequencies of the genotypes between astrocytomas was significantly different from the genotypes observed in non-astrocytomas (P b 0.001), whereas the difference in the distribution of the variant Pro allele was marginally significant (P = 0.02). 3.2. P53 codon Arg72Pro polymorphism in astrocytomas in relation to transdominant p53 mutations Some studies have looked at different types of p53 mutations, transdominant versus recessive mutation, in relation to the p53 codon 72 polymorphisms [16,22]. A transdominant mutation is one in which an amino acid change in the p53 protein alters the conformation of the p53 heterotetramer rendering it non-functional, despite the presence of a wild-type allele. In contrast, a recessive mutation does not alter the function of the p53 protein in the presence of the wild-type allele; only cells which subsequently undergo loss of the remaining wild-type allele, loss of heterozygosity (LOH), demonstrate a p53 protein with an altered function. Previously 105 astrocytomas had been screened for the presence of mutations in the p53 gene [7,12,18,21] and 34 cases had sufficient DNA for p53 codon Arg72Pro polymorphism analysis. Of the 34 cases, 13 had recessive mutations and 21 had transdominant p53 mutations (Table 3). The transdominant p53 mutations occurred in the well-known bhot-spotQ codons 175, 248, and 273 [16]. We observed that the frequency of the heterozygous Arg/Pro genotype was 68% (23 of 34) in these tumors and increased to 77% (10 of 13) in tumors with recessive p53 mutations. The frequency of the Arg/Pro genotype in the controls was 36% (42 of 117) which was significantly different from that observed 68% in this group of astrocytomas with p53 mutations (P = 0.002).
4. Discussion The P53 Arg72Pro polymorphic variants have been shown to produce proteins that differ in functional activity in several ways [2,4,15,23]. First, the biological function of the two variants was tested for their ability to suppress cellular transformation by oncogenic ras [23]. The Arg variant was twice as efficient as the Pro variant in suppressing transformation and this activity was associated with increased p53-mediated apoptosis [23]. Second, p73, a homologue of the p53 tumor suppressor gene was identified that has marked homology with p53 in the central core DNA binding domain suggesting that p53 and p73 may have similar functions [26]. The polymorphic variants in p53
101
Table 3 P53 Arg72Pro polymorphism in relation to transdominance of p53 mutations Sample
Age/Sex
Codon
Mutation
Transdominant
Codon 72
29 239 81 14 50 155 280 307 113 119 7 139 3 200 12 15 169 10 108 83 126 19 112 225 81 255 62 110 131 132 133 65 66 111
13/F 37/M ND 60/M 61/M 35/F 73/F 18/M 65/M 56/F 16/M 24/F 5/M 70/M 48/M 66/M 7/M 11/F 35/F 57/F 23/M 12/F 55/F 43/M 80/F 67/F 36/M 50/M 62/F 65/M 67/M 32/M 35/M 53/M
47 158 159 161 162 168 173 191 235 235 265 282 285 156 175 175 175 175 236 237 237 240 240 245 248 248 248 248 248 248 248 273 273 273
Pro N Ser Arg N Gly Ala N Val Ala N Ser Ile N Thr His N Arg Val N Ser Pro N Leu Asn N Lys Val N Leu Leu N Met Arg N Trp Glu N Lys Arg N Leu Arg N His Arg N His Arg N His Arg N His Tyr N Cys Met N Ile Met N Ile Ser N Thr Ser N Thr Gly N Ser Arg N Gln Arg N Gln Arg N Gln Arg N Gln Arg N Gln Arg N Gln Arg N Gln Arg N Cys Arg N Cys Arg N His
No No No No No No No No No No No No No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
arg-pro arg-arg arg-pro arg-pro arg-pro arg-pro arg-pro arg-arg pro-pro arg-pro arg-pro arg-pro arg-pro arg-pro arg-pro arg-arg arg-arg arg-pro arg-pro arg-pro arg-pro arg-pro arg-pro pro-pro arg-arg arg-arg arg-pro arg-pro arg-pro pro-pro arg-pro pro-pro arg-pro pro-pro
codon 72 affect the ability of the p73 protein to coimmunoprecipitate with mutant p53. The Arg variant appears to increase the ability of mutant p53 to bind p73, thus neutralizing p73-induced apoptosis and allowing cellular transformation by oncogenic ras. More recently, p73 has been shown to determine the response of cells to anticancer drugs and the polymorphisms and mutations at p53 codon 72 influence the inhibition of p73-dependent apoptosis [2]. Analysis of loss of heterozygosity in several different types of tumors with p53 gene mutations arising in Arg/Pro heterozygotes demonstrated preferential retention of the Arg allele [15,19,26]. Recently, it was demonstrated that individuals who are germline Arg/Pro heterozygotes and develop tumors with mutant p53 display a significant enrichment for the Arg allele [22]. Given the biologic activity of the Arg variant, this would favor inactivation of p73 function in a subset of tumors with mutant p53. Lastly, it was recently shown that p53 codon 72 polymorphic variants differ markedly in their potential to induce apoptosis [4]. The Arg variant was fivefold better at inducing apoptosis due to the fact that it was better able to localize to the mitochondria and this was associated with
102
P. Parhar et al. / Molecular Brain Research 137 (2005) 98–103
greater binding activity with MDM2. The greater apoptotic potential of the Arg variant may impact not only on the course of tumor development but also on the intrinsic sensitivity of tumors to radiation and chemotherapeutic treatments. Therefore, these differences between the biological activities of the different p53 codon 72 polymorphic variants may have implications for increasing the risk for cancer in certain individuals and for response to therapy in patients with cancer [2,11]. Several studies have found an association between the p53 Pro polymorphism and increased risk for tumor development. Studies of lung cancer, esophageal squamous cell carcinoma, and urothelial cancer report an increased risk for tumor development in individuals inheriting the Pro allele [6,13,19]. A large case–control study found that inheriting a least one Pro allele increased risk for lung adenocarcinoma 1.45-fold among smokers (95% confidence interval 1.02–2.06) [6]. This risk was greatest among heavy smokers, defined as individuals who had a greater than a 56 pack year smoking history. This subgroup experienced an odds ratio (OR) of 6.1 (95% CI 3.1–12.2). Another case–control study of esophageal squamous cell carcinoma in Taiwan reported that the Pro allele was enriched in patients with this tumor [13]. The OR for the increased risk in patients homozygous for the Pro allele was 2.56 (95% CI 1.29–5.08) and the OR for the heterozygous genotype Arg/Pro was 1.86 (95% CI 1.04– 3.35) [4]. In addition a case–control study of urothelial cancer looked at the frequency of the p53 codon 72 polymorphisms among male patients with urothelial tumors [19]. While the study did not find a statistically significant difference in the frequency distribution of the p53 codon 72 polymorphisms between the patient group and the controls, the study did identify a significantly higher frequency of the homozygous Pro/Pro genotype among patients who were smokers. Smokers with the homozygous genotype had an OR of 2.28 for urothelial cancer risk when compared to never-smokers with the same genotype (95% CI 1.12–4.66). In this study, we scored 135 brain tumor samples from Caucasians for the P53 Arg72Pro polymorphism. The genotype distributions of the P53 Arg72Pro between brain tumors and controls were statistically significant (P b 0.001) as well as their variant allele frequencies between cases and controls (P b 0.001). We found a significant enrichment of the Arg/Pro genotype in astrocytomas 86% (55 of 64) compared to the frequency of the Arg/Pro genotype 55% (39 of 72) in non-astrocytomas (P = 0.002). The significance of this association in brain tumors is comparable to similar associations reported in other human cancers [3,9,19,24,25,27]. Our results suggest a possible association between P53 Arg72Pro polymorphisms and susceptibility to brain tumors, particularly high-grade astrocytomas. Some mutations of p53 are able to inactivate the wildtype p53 protein in heterotetramer complexes and these are
known as dtransdominantT or ddominant-negativeT mutations [14,16,22]. One study examined whether the presence of transdominant p53 mutations in brain tumors was associated with any difference in biological or clinical characteristics compared with tumors harboring recessive p53 mutants. It was shown that transdominant mutations of the p53 gene were associated with early onset of glioblastomas [16]. Of 40 patients studied with sporadic gliomas, those with transdominant p53 mutations were significantly younger in age at diagnosis compared with those with recessive mutations (P b 0.012) [16]. More recently, with the discovery of the p73 homologue and the fact that the p53 codon 72 polymorphic variants can alter the biological activity of p73, one study examined the allelic ratio of the Arg/Pro alleles in relation to transdominant versus recessive p53 mutants [22]. Tada et al. [22] examined 39 high-grade astrocytomas with mutant p53 for p53 codon 72 polymorphisms. Among 14 tumors with transdominant p53 mutations, they found no enrichment of either the Pro or Arg alleles. In contrast, there was a significant enrichment of the Arg allele in the 25 tumors containing the recessive p53 mutations. They postulate that there may be a selective advantage for retention of the Arg allele in tumors with recessive p53 mutations because of increased inhibition of p73 function, i.e., this would be equivalent to a transdominant mutation in the p53 gene. In this study we analyzed 34 astrocytomas with p53 mutations (13 recessive and 21 transdominant) (Table 3). The Arg/Pro allelic ratio was 55:45 in tumors with recessive mutations and 50:50 in tumors with transdominant mutations. Tada et al. [22] observed a similar allelic ratio of 50:50 in tumors with p53 transdominant mutations. We did not have information regarding LOH of p53 in these tumors. However, similar to the finding by Tada et al. [22] that the Arg allele is selected in cancers with recessive p53 mutations, we noted that the frequency of Arg/Pro heterozygotes was significantly enriched to 77% (10 of 13) in tumors with recessive p53 mutations compared with the observed frequency of 36% in control (P = 0.006). Preferential retention of the Arg allele has also been reported in lung and breast cancers and associated with poor prognosis [3,19]. In summary, we have shown (i) that the genotype distributions of the P53 Arg72Pro between all brain tumors and controls were statistically significant (P b 0.001) as well as their variant allele frequencies between cases and controls (P b 0.001); (ii) that there was a significant increase in the Arg/Pro heterozygous genotype among high-grade astrocytomas compared with non-astrocytomas (P = 0.002); and (iii) that there was a significant increase in the Arg/Pro heterozygous genotype among high-grade astrocytomas containing transdominant as well as recessive p53 mutations compared with controls (P = 0.002). Our results suggest a possible association between P53 Arg72Pro polymorphisms and susceptibility to brain tumors, particularly high-grade astrocytomas.
P. Parhar et al. / Molecular Brain Research 137 (2005) 98–103
Acknowledgments We would like to thank Yevgeniy Lukyanov for helpful assistance in the preparation of the manuscript. This work was supported in part by the Children’s Brain Tumor Foundation, The National Childhood Cancer Foundation Grant No. 5 U10 CA13539-29 and National Institutes of Health Grant No. CA90290 to EWN. References [1] E. Balint, K.H. Vousden, Activation and activities of the p53 tumour suppressor protein, Br. J. Cancer 85 (2001) 1813 – 1823. [2] D. Bergamaschi, M. Gasco, L. Hiller, A. Sullivan, N. Syed, G. Trigiante, I. Yulug, M. Merlano, G. Numico, A. Comino, M. Attard, O. Reelfs, B. Gusterson, A.K. Bell, V. Heath, M. Tavassoli, P.J. Farrell, P. Smith, X. Lu, T. Crook, p53 polymorphism influences response in cancer chemotherapy via modulation of p73-dependent apoptosis, Cancer Cell 3 (2003) 387 – 402. [3] M. Bonafe, C. Ceccarelli, F. Farabegoli, D. Santini, M. Taffurelli, C. Barbi, E. Marzi, C. Trapassi, G. Storci, F. Olivieri, C. Franceschi, Retention of the p53 codon 72 arginine allele is associated with a reduction in disease-free and overall survival in arginine/proline heterozygous breast cancer patients, Clin. Cancer Res. 9 (2003) 4860 – 4864. [4] P. Dumont, J.I. Leu, A.C. Della Pietra III, D.L. George, M. Murphy, The codon 72 polymorphic variants of p53 have markedly different apoptotic potential, Nat. Genet. 33 (2003) 357 – 365. [5] R. Ezer, M. Alonso, E. Pereira, M. Kim, J.C. Allen, D.C. Miller, E.W. Newcomb, Identification of glutathione S-transferase (GST) polymorphisms in brain tumors and association with susceptibility to pediatric astrocytomas, J. Neruro-Oncol. 59 (2002) 123 – 134. [6] R. Fan, M.T. Wu, D. Miller, J. Wain, K.T. Kelsey, J.K. Wiencke, D.C. Christiani, The p53 codon 72 polymorphism and lung cancer risk, Cancer Epidemiol., Biomarkers Prev. 9 (2000) 1037 – 1042. [7] R.H. Frankel, W. Bayona, M. Koslow, Newcomb EW, p53 mutations in human malignant gliomas: comparison of loss of heterozygosity with mutation frequency, Cancer Res. 52 (1992) 1427 – 1433. [8] P. Hainaut, M. Hollstein, p53 and human cancer: the first ten thousand mutations, Adv. Cancer Res. 77 (2000) 81 – 137. [9] X. Jin, X. Wu, J.A. Roth, C.I. Amos, T.M. King, C. Branch, S.E. Honn, M.R. Spitz, Higher lung cancer risk for younger AfricanAmericans with the Pro/Pro p53 genotype, Carcinogenesis 16 (1995) 2205 – 2208. [10] P. Kleihues, W.K. Cavenee (Eds.), World Health Organization Classification of Tumours of the Nervous System, IARC Press, Lyon, 2000. [11] L.E. Knudsen, S.H. Loft, H. Autrup, Risk assessment: the importance of genetic polymorphisms in man, Mutat. Res. 482 (2001) 83 – 88. [12] F.F. Lang, D.C. Miller, S. Pisharody, M. Koslow, Newcomb EW, High frequency of p53 protein accumulation without p53 gene mutation in human juvenile pilocytic, low grade and anaplastic astrocytomas, Oncogene 9 (1994) 949 – 954.
103
[13] J.M. Lee, Y.C. Lee, S.Y. Yang, W.L. Shi, C.J. Lee, S.P. Luh, C.J. Chen, C.Y. Hsieh, M.Y. Wu, Genetic polymorphisms of p53 and GSTP1, but not NAT2, are associated with susceptibility to squamouscell carcinoma of the esophagus, Int. J. Cancer 89 (2000) 458 – 464. [14] A.J. Levine, p53, the cellular gatekeeper for growth and division, Cell 88 (1997) 323 – 331. [15] M.C. Marin, C.A. Jost, L.A. Brooks, M.S. Irwin, J. O’Nions, J.A. Tidy, N. James, J.M. McGregor, C.A. Harwood, I.G. Yulug, K.H. Vousden, M.J. Allday, B. Gusterson, S. Ikawa, P.W. Hinds, T. Crook, W.G. Kaelin Jr., A common polymorphism acts as an intragenic modifier of mutant p53 behavior, Nat. Genet. 25 (2000) 47 – 54. [16] M. Marutani, H. Tonoki, M. Tada, M. Takahashi, H. Kashiwazaki, Y. Hida, J. Hamada, M. Asaka, T. Moriuchi, Dominant-negative mutations of the tumor suppressor p53 relating to early onset of glioblastoma multiforme, Cancer Res. 59 (1999) 4765 – 4769. [17] G.J. Matlashewski, S. Tuck, D. Pim, P. Lamb, J. Schneider, L.V. Crawford, Primary structure polymorphism at amino acid residue 72 of human p53, Mol. Cell. Biol. 7 (1987) 961 – 963. [18] E.W. Newcomb, H. Cohen, S.R. Lee, S.K. Bhalla, J. Bloom, R.L. Hayes, D.C. Miller, Survival of patients with glioblastoma multiforme is not influenced by altered expression of p16, p53, EGF-R, MDM2, or Bcl-2 genes, Brain Pathol. 8 (1998) 655 – 667. [19] E. Papadakis, N. Soulitzis, D.A. Spandidos, Association of p53 codon 72 polymorphism with advanced lung cancer: the Arg allele is preferentially retained in tumours arising in Arg/Pro germline heterozygotes, Br. J. Cancer 87 (2002) 1013 – 1018. [20] R. Soong, B.J. Iacopetta, A rapid and nonisotopic method for screening and sequencing p53 gene mutations in formalin-fixed paraffin-embedded tumors, Mod. Pathol. 10 (1997) 252 – 258. [21] T. Sung, D.C. Miller, R.L. Hayes, M. Alonso, H. Yee, E.W. Newcomb, Preferential inactivation of the p53 tumor suppression pathway and lack of EGF-R amplification distinguish de novo high grade pediatric astrocytomas from de novo adult astrocytomas, Brain Pathol. 10 (2000) 249 – 259. [22] M. Tada, K. Furuuchi, M. Kaneda, J. Matsumoto, M. Takahashi, A. Hirai, Y. Mitsumoto, R.I. Iggo, T. Moriuchi, Inactivate the remaining p53 allele or alternative p73? Preferential selection of the Arg72 polymorphism in cancers with recessive p53 mutants but not transdominant mutants, Carcinogenesis 22 (2001) 515 – 517. [23] M. Thomas, A. Kalita, S. Labrecque, D. Pim, L. Banks, G. Matlashewski, Two polymorphic variants of wild-type p53 differ biochemically and biologically, Mol. Cell. Biol. 19 (1999) 1092 – 1100. [24] Y.C. Wang, C.Y. Chen, A.K. Chen, Y.Y. Chang, P. Lin, p53 codon 72 polymorphism in Taiwanese lung cancer patients: association with lung cancer susceptibility and prognosis, Clin. Cancer Res. 5 (1999) 129 – 134. [25] A. Weston, C. Pan, B.H. Ksieski, S. Wallenstein, G.S. Berkowitz, P.I. Tartter, I.J. Bleiweiss, S.T. Brower, R.T. Senie, M.S. Wolff, p53 haplotype determination in breast cancer, Cancer Epidemiol., Biomarkers Prev. 6 (1997) 105 – 112. [26] A. Yang, M. Kaghad, D. Caput, F. McKeon, On the shoulders of giants: p63, p73 and the rise of p53, Trends Genet. 18 (2002) 90 – 95. [27] Z.W. Zhang, P. Newcomb, A. Hollowood, R. Feakins, M. Moorghen, A. Storey, M.J.G. Farthing, D. Alderson, J. Holly, Age-associated increase of codon 72 arginine p53 frequency in gastric cardia and noncardia adenocarcinoma, Clin. Cancer Res. 9 (2003) 2151 – 2156.
Research report
Possible association of p53 codon 72 polymorphism with susceptibility to adult and pediatric high-grade astrocytomas Preeti Parhara, Rona Ezera, Yongzhao Shaob,c, Jeffrey C. Allend, Douglas C. Millera,b, Elizabeth W. Newcomba,b,* a
Department of Pathology, New York University School of Medicine, 550 First Avenue, MSB531, New York, NY 10016, USA b New York University Cancer Institute, New York, NY 10016, USA c Department of Environmental Medicine, New York University School of Medicine, New York, NY 10016, USA d Department of Pediatrics, Division of Pediatric Hematology/Oncology, New York University School of Medicine, New York, NY 10016, USA Accepted 13 February 2005 Available online 21 March 2005
Abstract Polymorphisms in codon 72 of the p53 tumor suppressor gene have been associated with susceptibility to human cancer. We wished to evaluate whether variant allelic forms of the p53 protein were associated with brain tumors. In this study, we scored 135 brain tumor samples (92 adult and 43 pediatric cases consisting of 64 high-grade astrocytomas and 71 non-astrocytomas) for the P53 Arg72Pro polymorphisms. Our data show that the genotype frequencies of P53 Arg72Pro vary not only between patients with brain tumors and controls, but also between different histological subtypes of brain tumors. Specifically, we found (i) that the genotype distributions of the P53 Arg72Pro between all brain tumors and controls were statistically significant (P b 0.001) as well as their variant allele frequencies between cases and controls (P b 0.001); (ii) that there was a significant increase in the Arg/Pro heterozygous genotype among high-grade astrocytomas compared with non-astrocytomas (P = 0.002); and (iii) that there was a significant increase in the Arg/Pro heterozygous genotype among high-grade astrocytomas containing transdominant as well as recessive p53 mutations compared with controls (P = 0.002). Our results suggest a possible association between P53 Arg72Pro polymorphisms and susceptibility to brain tumors, particularly high-grade astrocytomas. D 2005 Elsevier B.V. All rights reserved. Theme: Disorders of the nervous system Topic: Brain tumors Keywords: p53 codon 72 polymorphisms; High-grade astrocytomas
1. Introduction The most common mutation found in human cancers occurs in the p53 tumor suppressor gene [8,14]. The p53 gene encodes a nuclear phosphoprotein involved in cell cycle arrest, apoptosis, and genetic stability [1,14]. The gene consists of three functional domains: an N-terminal transactivating domain, a central DNA binding domain, and a C* Corresponding author. Department of Pathology, New York University School of Medicine, 550 First Avenue, MSB531, New York, NY 10016, USA. Fax: +1 212 263 8211. E-mail address: [email protected] (E.W. Newcomb). 0169-328X/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.molbrainres.2005.02.016
terminal oligomerization domain [1,14]. The N-terminal acidic domain induces transcription of genes. Examples of p53 inducible genes upregulated in response to DNA damage include GADD45, which is involved in DNA nucleotide excision repair, and p21, which inhibits several cyclin-dependent kinases to stop the G1/S phase transition in the cell cycle and thus halt cellular division [14]. Apoptosis can be triggered by p53-mediated transcriptional upregulation of bax and insulin-like growth factor-binding protein-3 (IGF-BP3), two genes involved in the apoptotic pathway [14]. The central DNA binding domain has contact points for consensus DNA binding sequences. Finally, the Cterminal domain is important for oligomerization. To dsenseT
P. Parhar et al. / Molecular Brain Research 137 (2005) 98–103
DNA damage the p53 protein must be in a conformationally stable tetramer. Some mutations decrease the ability to form a stable tetramer rendering the molecule non-functional. These mutations are referred to as dtransdominantT or ddominant-negativeT mutations since they are able to inactivate the protein by forming a heterotetramer complex with the wild-type p53 protein encoded by the remaining copy of the normal allele. In contrast, drecessiveT mutations are unable to alter the conformation of the p53 protein in the presence of wild-type p53 protein and the protein retains its function. Predisposition to many adult cancers has now been associated with the inheritance of polymorphisms in genes, singly or in combination, that may make an individual more susceptible to malignant transformation [11]. For the p53 tumor suppressor gene a polymorphism has been identified at codon 72 within exon 4: Arg, encoded by CGC, and Pro, encoded by CCC [17]. The amino acid substitution at codon 72 affects the primary structure of p53 that results in functional differences between the different alleles [4,23]. The Pro allele is associated with increased transcriptional activity compared to the Arg allele because it binds more tightly to the transcription factors TAFII32 and TAFII70 [23]. More recently it was found that the Arg allele induced apoptosis fivefold more than the Pro allele, due to its ability to localize to the mitochondria and bind with the MDM2 protein more avidly [4]. In addition, it has been shown in squamous cell carcinomas that the p53 codon 72 polymorphism influences clinical response to DNA-damaging chemo- and radiation-therapy [2]. These differences between the biological functions of the different p53 variants may have implications not only for increasing the risk for cancer in certain individuals, but also for influencing overall survival in patients with cancer [2,11]. Although the frequency and spectrum of mutations in the p53 gene has been extensively studied in brain tumors, the p53 codon 72 polymorphism has only been examined in a few cases [8,22]. The purpose of this study was to genotype brain tumors for the different p53 codon 72 polymorphic variants and determine the distribution of genotype frequencies. In order to evaluate the question of whether variant allelic forms of the p53 gene may be associated with brain tumors, we compared the genotype frequencies and the ratio of the variant alleles in our patients with brain tumors of different histological subtypes to that reported for a control population.
2. Materials and methods 2.1. Tumor samples Tumor samples were obtained from patients who had surgery at New York University Medical Center (NYUMC) or Beth Israel Medical Center (BIMC) between 1986 and
99
2001. DNA was extracted and banked. These samples have been used in several other genetic studies previously reported [5,7,12,18,21]. Histopathologic grading of astrocytomas was done in accordance with the WHO classification scheme except that necrosis was required for a diagnosis of glioblastoma [10]. For this study of codon 72 p53 polymorphisms, limited to Caucasian patients, we retrieved 135 cases from our tumor bank. DNA was extracted from frozen or paraffin-embedded tissues as described [20]. 2.2. Patient population Table 1 summarizes the brain tumor samples used and some of the demographics of the study population. Of the 135 cases, 43 were from pediatric patients and 92 were from adult patients. There were 64 high-grade astrocytomas (consisting of 9 anaplastic astrocytoma and 54 glioblastomas), and 71 were non-astrocytomas (consisting of oligodendrogliomas, oligoastrocytomas, gangliogliomas, ganglioneurocytomas, and glioneurocytomas). Of the 43 pediatric patients, an equal number of samples was analyzed from patients 0–9 and 10–19 years of age. Of the 92 adult patients, the peak age for astrocytomas was 50–79, while the peak age for non-astrocytomas was 20–49. The distribution of gender among patients with astrocytomas and nonastrocytomas was identical, consisting of 66% males and 34% females. 2.3. Genotype detection PCR followed by restriction enzyme digestion was used to assess p53 codon 72 polymorphisms with protocols described previously [5,13]. The PCR product (10 Al) was digested overnight with BstUI (NE Biolabs) to determine the amino acid at codon 72 according to the manufacturer’s directions. A restriction enzyme premix (10 Al) containing 10 U of enzyme and 2 NEBuffer 2, made up of 100 mM Tris–HCl and 1 mM dithiothreitol (pH 7.0), was added to the PCR product for a final volume of 20 Al. For all assays, the amplified PCR fragments were
Table 1 Brain tumor samples and demographics of study population Variable
No.
Astrocytomas
Non-astrocytomas
Tumor samples Age at diagnosis Pediatric cases 0–9 10–19 Adult cases 20–49 50–79 Sex Male Female
135
64 (47%)
71 (53%)
26 15 11 38 13 25
17 6 11 54 39 15
43
92
89 46
(58%) (42%) (34%) (66%)
42 (66%) 22 (34%)
(35%) (65%) (72%) (28%)
47 (66%) 24 (34%)
100
P. Parhar et al. / Molecular Brain Research 137 (2005) 98–103
separated on 3.5% agarose gels and visualized by staining with GelStar nucleic acid gel stain (BioWhittaker Molecular Applications). Gels were photographed using a Mitsubishi P91 video copy processor. Genotypes were scored as homozygous for the Pro allele if the 199-bp PCR fragment after digestion remained uncut or homozygous for the Arg allele if after enzyme digestion two bands (113 and 86 bp) were observed. Gels with three DNA fragments (199, 113, and 86 bp) were scored as heterozygotes. 2.4. Statistical analysis Observed frequencies of genotypes in brain tumors were compared to controls using the chi-square or Fisher’s exact test when expected frequencies were small. The statistical computer software S-Plus 2000 was used to perform all statistical analyses (Insightful Corporation). Statistical significance was set at P b 0.05.
3. Results 3.1. Frequency of P53 codon Arg72Pro polymorphisms in brain tumor samples Brain tumor samples were evaluated for the P53 codon Arg72Pro polymorphism. The substitution of Arg for Pro at codon 72 creates a unique BstUI restriction site. Fig. 1 shows DNAs from representative samples incubated without ( ) and with (+) the restriction enzyme. Genotypes were scored as homozygous for the Arg allele if the 199bp DNA fragment after enzyme digestion gave bands of 113 and 86 bp (lane 2) or homozygous for the Pro allele if the PCR fragment after digestion remained uncut (lane 6). Presence of all three DNA fragments indicated a heterozygote (lane 4).
Fig. 1. Detection of P53 Arg72Pro polymorphisms. DNAs from representative brain tumor samples were scored for P53 Arg72Pro polymorphisms by PCR-RFLP analysis. PCR products were incubated without ( ) or with (+) the BstUI restriction enzyme and run side-by-side to better evaluate the extent of enzyme digestion. Samples that had the Arg (CGC) at codon 72 gained a unique BstUI cleavage site. Lane 2 contains DNA from an Arg/ Arg homozygote that is completely digested into two fragments of 113 and 86 bp. Lane 4 contains DNA from an Arg/Pro heterozygote that after digestion shows three bands of 199, 113, and 86 bp. Lane 6 contains DNA from a Pro/Pro homozygote that remains completely undigested to give a single band of 199 bp.
Table 2 Frequency of P53 Arg72Pro polymorphisms in brain tumor samples Genotype
Group 1 No.
Arg/Arg Arg/Pro Pro/Pro Pro allele frequency Arg/Arg Arg/Pro Pro/Pro Pro allele frequency Arg/Arg Arg/Pro Pro/Pro Pro allele frequency Arg/Arg Arg/Pro Pro/Pro Pro allele frequency
Group 2 (%)
Controlsa 72 (62) 42 (36) 3 (3) 0.21 Controls 72 (62) 42 (36) 3 (3) 0.21 Controls 72 (62) 42 (36) 3 (3) 0.21 All Astrocytomas 8 (12.5) 55 (86) 1 (1.5) 0.45
No.
P value* (%)
All brain tumors 38 (28) 94 (69) 3 (2) 0.37 All astrocytomas 8 (12.5) 55 (86) 1 (1.5) 0.45 All non-astrocytomas 30 (42) 39 (55) 2 (3) 0.30 All non-astrocytomas 30 (42) 39 (55) 2 (3) 0.30
b0.001
b0.001 b0.001
b0.001 0.03
0.04 b0.001
0.02
a
Control genotypes and allele frequencies as described by Weston et al. [25]. * P values were obtained using Fisher’s exact test or chi-square test.
As shown in Table 2, the frequencies of the Arg/Arg, Arg/Pro, and Pro/Pro genotypes among control cases were 62%, 36% and 3%, respectively, resulting in a Pro allele frequency of 0.21 as described [25]. The distribution of these genotype frequencies was in agreement with those calculated from the Hardy–Weinberg equilibrium model for controls (P N 0.25). We next analyzed the distribution of P53 Arg72Pro genotypes in all brain tumors compared to the controls. The frequencies of the Arg/Arg, Arg/Pro, and Pro/Pro genotypes among all brain tumor cases were 28%, 69%, and 2%, respectively, which was significantly different from the genotypes observed in the controls (P b 0.001). Likewise, the difference in the distribution of the variant Pro allele between controls and brain tumor cases was also statistically significant (P b 0.001). We next stratified the brain tumor samples by histological subtype into astrocytomas and non-astrocytomas and evaluated their genotype frequencies compared to the controls. The frequencies of the Arg/Arg, Arg/Pro, and Pro/Pro genotypes among all astrocytomas were 12.5%, 86%, and 1.5%, respectively, which was significantly different from the genotypes observed in the controls (P b 0.001). Similarly, the difference in the distribution of the variant Pro allele between controls and all astrocytomas was also statistically significant (P b 0.001). Since age can be a confounding variable, we also stratified the high-grade astrocytomas by age and compared the frequencies of the genotypes between pediatric and adult patients. The frequencies of the genotypes and the distribution of the variant Pro allele were similar in both age groups (data not shown). For non-astrocytomas, the
P. Parhar et al. / Molecular Brain Research 137 (2005) 98–103
frequencies of the Arg/Arg, Arg/Pro, and Pro/Pro genotypes was 42%, 55% and 3%, respectively, which was marginally significant from the genotypes observed in the controls (P = 0.03). Finally, we compared the distribution of genotypes between astrocytomas and non-astrocytomas. The frequencies of the genotypes between astrocytomas was significantly different from the genotypes observed in non-astrocytomas (P b 0.001), whereas the difference in the distribution of the variant Pro allele was marginally significant (P = 0.02). 3.2. P53 codon Arg72Pro polymorphism in astrocytomas in relation to transdominant p53 mutations Some studies have looked at different types of p53 mutations, transdominant versus recessive mutation, in relation to the p53 codon 72 polymorphisms [16,22]. A transdominant mutation is one in which an amino acid change in the p53 protein alters the conformation of the p53 heterotetramer rendering it non-functional, despite the presence of a wild-type allele. In contrast, a recessive mutation does not alter the function of the p53 protein in the presence of the wild-type allele; only cells which subsequently undergo loss of the remaining wild-type allele, loss of heterozygosity (LOH), demonstrate a p53 protein with an altered function. Previously 105 astrocytomas had been screened for the presence of mutations in the p53 gene [7,12,18,21] and 34 cases had sufficient DNA for p53 codon Arg72Pro polymorphism analysis. Of the 34 cases, 13 had recessive mutations and 21 had transdominant p53 mutations (Table 3). The transdominant p53 mutations occurred in the well-known bhot-spotQ codons 175, 248, and 273 [16]. We observed that the frequency of the heterozygous Arg/Pro genotype was 68% (23 of 34) in these tumors and increased to 77% (10 of 13) in tumors with recessive p53 mutations. The frequency of the Arg/Pro genotype in the controls was 36% (42 of 117) which was significantly different from that observed 68% in this group of astrocytomas with p53 mutations (P = 0.002).
4. Discussion The P53 Arg72Pro polymorphic variants have been shown to produce proteins that differ in functional activity in several ways [2,4,15,23]. First, the biological function of the two variants was tested for their ability to suppress cellular transformation by oncogenic ras [23]. The Arg variant was twice as efficient as the Pro variant in suppressing transformation and this activity was associated with increased p53-mediated apoptosis [23]. Second, p73, a homologue of the p53 tumor suppressor gene was identified that has marked homology with p53 in the central core DNA binding domain suggesting that p53 and p73 may have similar functions [26]. The polymorphic variants in p53
101
Table 3 P53 Arg72Pro polymorphism in relation to transdominance of p53 mutations Sample
Age/Sex
Codon
Mutation
Transdominant
Codon 72
29 239 81 14 50 155 280 307 113 119 7 139 3 200 12 15 169 10 108 83 126 19 112 225 81 255 62 110 131 132 133 65 66 111
13/F 37/M ND 60/M 61/M 35/F 73/F 18/M 65/M 56/F 16/M 24/F 5/M 70/M 48/M 66/M 7/M 11/F 35/F 57/F 23/M 12/F 55/F 43/M 80/F 67/F 36/M 50/M 62/F 65/M 67/M 32/M 35/M 53/M
47 158 159 161 162 168 173 191 235 235 265 282 285 156 175 175 175 175 236 237 237 240 240 245 248 248 248 248 248 248 248 273 273 273
Pro N Ser Arg N Gly Ala N Val Ala N Ser Ile N Thr His N Arg Val N Ser Pro N Leu Asn N Lys Val N Leu Leu N Met Arg N Trp Glu N Lys Arg N Leu Arg N His Arg N His Arg N His Arg N His Tyr N Cys Met N Ile Met N Ile Ser N Thr Ser N Thr Gly N Ser Arg N Gln Arg N Gln Arg N Gln Arg N Gln Arg N Gln Arg N Gln Arg N Gln Arg N Cys Arg N Cys Arg N His
No No No No No No No No No No No No No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
arg-pro arg-arg arg-pro arg-pro arg-pro arg-pro arg-pro arg-arg pro-pro arg-pro arg-pro arg-pro arg-pro arg-pro arg-pro arg-arg arg-arg arg-pro arg-pro arg-pro arg-pro arg-pro arg-pro pro-pro arg-arg arg-arg arg-pro arg-pro arg-pro pro-pro arg-pro pro-pro arg-pro pro-pro
codon 72 affect the ability of the p73 protein to coimmunoprecipitate with mutant p53. The Arg variant appears to increase the ability of mutant p53 to bind p73, thus neutralizing p73-induced apoptosis and allowing cellular transformation by oncogenic ras. More recently, p73 has been shown to determine the response of cells to anticancer drugs and the polymorphisms and mutations at p53 codon 72 influence the inhibition of p73-dependent apoptosis [2]. Analysis of loss of heterozygosity in several different types of tumors with p53 gene mutations arising in Arg/Pro heterozygotes demonstrated preferential retention of the Arg allele [15,19,26]. Recently, it was demonstrated that individuals who are germline Arg/Pro heterozygotes and develop tumors with mutant p53 display a significant enrichment for the Arg allele [22]. Given the biologic activity of the Arg variant, this would favor inactivation of p73 function in a subset of tumors with mutant p53. Lastly, it was recently shown that p53 codon 72 polymorphic variants differ markedly in their potential to induce apoptosis [4]. The Arg variant was fivefold better at inducing apoptosis due to the fact that it was better able to localize to the mitochondria and this was associated with
102
P. Parhar et al. / Molecular Brain Research 137 (2005) 98–103
greater binding activity with MDM2. The greater apoptotic potential of the Arg variant may impact not only on the course of tumor development but also on the intrinsic sensitivity of tumors to radiation and chemotherapeutic treatments. Therefore, these differences between the biological activities of the different p53 codon 72 polymorphic variants may have implications for increasing the risk for cancer in certain individuals and for response to therapy in patients with cancer [2,11]. Several studies have found an association between the p53 Pro polymorphism and increased risk for tumor development. Studies of lung cancer, esophageal squamous cell carcinoma, and urothelial cancer report an increased risk for tumor development in individuals inheriting the Pro allele [6,13,19]. A large case–control study found that inheriting a least one Pro allele increased risk for lung adenocarcinoma 1.45-fold among smokers (95% confidence interval 1.02–2.06) [6]. This risk was greatest among heavy smokers, defined as individuals who had a greater than a 56 pack year smoking history. This subgroup experienced an odds ratio (OR) of 6.1 (95% CI 3.1–12.2). Another case–control study of esophageal squamous cell carcinoma in Taiwan reported that the Pro allele was enriched in patients with this tumor [13]. The OR for the increased risk in patients homozygous for the Pro allele was 2.56 (95% CI 1.29–5.08) and the OR for the heterozygous genotype Arg/Pro was 1.86 (95% CI 1.04– 3.35) [4]. In addition a case–control study of urothelial cancer looked at the frequency of the p53 codon 72 polymorphisms among male patients with urothelial tumors [19]. While the study did not find a statistically significant difference in the frequency distribution of the p53 codon 72 polymorphisms between the patient group and the controls, the study did identify a significantly higher frequency of the homozygous Pro/Pro genotype among patients who were smokers. Smokers with the homozygous genotype had an OR of 2.28 for urothelial cancer risk when compared to never-smokers with the same genotype (95% CI 1.12–4.66). In this study, we scored 135 brain tumor samples from Caucasians for the P53 Arg72Pro polymorphism. The genotype distributions of the P53 Arg72Pro between brain tumors and controls were statistically significant (P b 0.001) as well as their variant allele frequencies between cases and controls (P b 0.001). We found a significant enrichment of the Arg/Pro genotype in astrocytomas 86% (55 of 64) compared to the frequency of the Arg/Pro genotype 55% (39 of 72) in non-astrocytomas (P = 0.002). The significance of this association in brain tumors is comparable to similar associations reported in other human cancers [3,9,19,24,25,27]. Our results suggest a possible association between P53 Arg72Pro polymorphisms and susceptibility to brain tumors, particularly high-grade astrocytomas. Some mutations of p53 are able to inactivate the wildtype p53 protein in heterotetramer complexes and these are
known as dtransdominantT or ddominant-negativeT mutations [14,16,22]. One study examined whether the presence of transdominant p53 mutations in brain tumors was associated with any difference in biological or clinical characteristics compared with tumors harboring recessive p53 mutants. It was shown that transdominant mutations of the p53 gene were associated with early onset of glioblastomas [16]. Of 40 patients studied with sporadic gliomas, those with transdominant p53 mutations were significantly younger in age at diagnosis compared with those with recessive mutations (P b 0.012) [16]. More recently, with the discovery of the p73 homologue and the fact that the p53 codon 72 polymorphic variants can alter the biological activity of p73, one study examined the allelic ratio of the Arg/Pro alleles in relation to transdominant versus recessive p53 mutants [22]. Tada et al. [22] examined 39 high-grade astrocytomas with mutant p53 for p53 codon 72 polymorphisms. Among 14 tumors with transdominant p53 mutations, they found no enrichment of either the Pro or Arg alleles. In contrast, there was a significant enrichment of the Arg allele in the 25 tumors containing the recessive p53 mutations. They postulate that there may be a selective advantage for retention of the Arg allele in tumors with recessive p53 mutations because of increased inhibition of p73 function, i.e., this would be equivalent to a transdominant mutation in the p53 gene. In this study we analyzed 34 astrocytomas with p53 mutations (13 recessive and 21 transdominant) (Table 3). The Arg/Pro allelic ratio was 55:45 in tumors with recessive mutations and 50:50 in tumors with transdominant mutations. Tada et al. [22] observed a similar allelic ratio of 50:50 in tumors with p53 transdominant mutations. We did not have information regarding LOH of p53 in these tumors. However, similar to the finding by Tada et al. [22] that the Arg allele is selected in cancers with recessive p53 mutations, we noted that the frequency of Arg/Pro heterozygotes was significantly enriched to 77% (10 of 13) in tumors with recessive p53 mutations compared with the observed frequency of 36% in control (P = 0.006). Preferential retention of the Arg allele has also been reported in lung and breast cancers and associated with poor prognosis [3,19]. In summary, we have shown (i) that the genotype distributions of the P53 Arg72Pro between all brain tumors and controls were statistically significant (P b 0.001) as well as their variant allele frequencies between cases and controls (P b 0.001); (ii) that there was a significant increase in the Arg/Pro heterozygous genotype among high-grade astrocytomas compared with non-astrocytomas (P = 0.002); and (iii) that there was a significant increase in the Arg/Pro heterozygous genotype among high-grade astrocytomas containing transdominant as well as recessive p53 mutations compared with controls (P = 0.002). Our results suggest a possible association between P53 Arg72Pro polymorphisms and susceptibility to brain tumors, particularly high-grade astrocytomas.
P. Parhar et al. / Molecular Brain Research 137 (2005) 98–103
Acknowledgments We would like to thank Yevgeniy Lukyanov for helpful assistance in the preparation of the manuscript. This work was supported in part by the Children’s Brain Tumor Foundation, The National Childhood Cancer Foundation Grant No. 5 U10 CA13539-29 and National Institutes of Health Grant No. CA90290 to EWN. References [1] E. Balint, K.H. Vousden, Activation and activities of the p53 tumour suppressor protein, Br. J. Cancer 85 (2001) 1813 – 1823. [2] D. Bergamaschi, M. Gasco, L. Hiller, A. Sullivan, N. Syed, G. Trigiante, I. Yulug, M. Merlano, G. Numico, A. Comino, M. Attard, O. Reelfs, B. Gusterson, A.K. Bell, V. Heath, M. Tavassoli, P.J. Farrell, P. Smith, X. Lu, T. Crook, p53 polymorphism influences response in cancer chemotherapy via modulation of p73-dependent apoptosis, Cancer Cell 3 (2003) 387 – 402. [3] M. Bonafe, C. Ceccarelli, F. Farabegoli, D. Santini, M. Taffurelli, C. Barbi, E. Marzi, C. Trapassi, G. Storci, F. Olivieri, C. Franceschi, Retention of the p53 codon 72 arginine allele is associated with a reduction in disease-free and overall survival in arginine/proline heterozygous breast cancer patients, Clin. Cancer Res. 9 (2003) 4860 – 4864. [4] P. Dumont, J.I. Leu, A.C. Della Pietra III, D.L. George, M. Murphy, The codon 72 polymorphic variants of p53 have markedly different apoptotic potential, Nat. Genet. 33 (2003) 357 – 365. [5] R. Ezer, M. Alonso, E. Pereira, M. Kim, J.C. Allen, D.C. Miller, E.W. Newcomb, Identification of glutathione S-transferase (GST) polymorphisms in brain tumors and association with susceptibility to pediatric astrocytomas, J. Neruro-Oncol. 59 (2002) 123 – 134. [6] R. Fan, M.T. Wu, D. Miller, J. Wain, K.T. Kelsey, J.K. Wiencke, D.C. Christiani, The p53 codon 72 polymorphism and lung cancer risk, Cancer Epidemiol., Biomarkers Prev. 9 (2000) 1037 – 1042. [7] R.H. Frankel, W. Bayona, M. Koslow, Newcomb EW, p53 mutations in human malignant gliomas: comparison of loss of heterozygosity with mutation frequency, Cancer Res. 52 (1992) 1427 – 1433. [8] P. Hainaut, M. Hollstein, p53 and human cancer: the first ten thousand mutations, Adv. Cancer Res. 77 (2000) 81 – 137. [9] X. Jin, X. Wu, J.A. Roth, C.I. Amos, T.M. King, C. Branch, S.E. Honn, M.R. Spitz, Higher lung cancer risk for younger AfricanAmericans with the Pro/Pro p53 genotype, Carcinogenesis 16 (1995) 2205 – 2208. [10] P. Kleihues, W.K. Cavenee (Eds.), World Health Organization Classification of Tumours of the Nervous System, IARC Press, Lyon, 2000. [11] L.E. Knudsen, S.H. Loft, H. Autrup, Risk assessment: the importance of genetic polymorphisms in man, Mutat. Res. 482 (2001) 83 – 88. [12] F.F. Lang, D.C. Miller, S. Pisharody, M. Koslow, Newcomb EW, High frequency of p53 protein accumulation without p53 gene mutation in human juvenile pilocytic, low grade and anaplastic astrocytomas, Oncogene 9 (1994) 949 – 954.
103
[13] J.M. Lee, Y.C. Lee, S.Y. Yang, W.L. Shi, C.J. Lee, S.P. Luh, C.J. Chen, C.Y. Hsieh, M.Y. Wu, Genetic polymorphisms of p53 and GSTP1, but not NAT2, are associated with susceptibility to squamouscell carcinoma of the esophagus, Int. J. Cancer 89 (2000) 458 – 464. [14] A.J. Levine, p53, the cellular gatekeeper for growth and division, Cell 88 (1997) 323 – 331. [15] M.C. Marin, C.A. Jost, L.A. Brooks, M.S. Irwin, J. O’Nions, J.A. Tidy, N. James, J.M. McGregor, C.A. Harwood, I.G. Yulug, K.H. Vousden, M.J. Allday, B. Gusterson, S. Ikawa, P.W. Hinds, T. Crook, W.G. Kaelin Jr., A common polymorphism acts as an intragenic modifier of mutant p53 behavior, Nat. Genet. 25 (2000) 47 – 54. [16] M. Marutani, H. Tonoki, M. Tada, M. Takahashi, H. Kashiwazaki, Y. Hida, J. Hamada, M. Asaka, T. Moriuchi, Dominant-negative mutations of the tumor suppressor p53 relating to early onset of glioblastoma multiforme, Cancer Res. 59 (1999) 4765 – 4769. [17] G.J. Matlashewski, S. Tuck, D. Pim, P. Lamb, J. Schneider, L.V. Crawford, Primary structure polymorphism at amino acid residue 72 of human p53, Mol. Cell. Biol. 7 (1987) 961 – 963. [18] E.W. Newcomb, H. Cohen, S.R. Lee, S.K. Bhalla, J. Bloom, R.L. Hayes, D.C. Miller, Survival of patients with glioblastoma multiforme is not influenced by altered expression of p16, p53, EGF-R, MDM2, or Bcl-2 genes, Brain Pathol. 8 (1998) 655 – 667. [19] E. Papadakis, N. Soulitzis, D.A. Spandidos, Association of p53 codon 72 polymorphism with advanced lung cancer: the Arg allele is preferentially retained in tumours arising in Arg/Pro germline heterozygotes, Br. J. Cancer 87 (2002) 1013 – 1018. [20] R. Soong, B.J. Iacopetta, A rapid and nonisotopic method for screening and sequencing p53 gene mutations in formalin-fixed paraffin-embedded tumors, Mod. Pathol. 10 (1997) 252 – 258. [21] T. Sung, D.C. Miller, R.L. Hayes, M. Alonso, H. Yee, E.W. Newcomb, Preferential inactivation of the p53 tumor suppression pathway and lack of EGF-R amplification distinguish de novo high grade pediatric astrocytomas from de novo adult astrocytomas, Brain Pathol. 10 (2000) 249 – 259. [22] M. Tada, K. Furuuchi, M. Kaneda, J. Matsumoto, M. Takahashi, A. Hirai, Y. Mitsumoto, R.I. Iggo, T. Moriuchi, Inactivate the remaining p53 allele or alternative p73? Preferential selection of the Arg72 polymorphism in cancers with recessive p53 mutants but not transdominant mutants, Carcinogenesis 22 (2001) 515 – 517. [23] M. Thomas, A. Kalita, S. Labrecque, D. Pim, L. Banks, G. Matlashewski, Two polymorphic variants of wild-type p53 differ biochemically and biologically, Mol. Cell. Biol. 19 (1999) 1092 – 1100. [24] Y.C. Wang, C.Y. Chen, A.K. Chen, Y.Y. Chang, P. Lin, p53 codon 72 polymorphism in Taiwanese lung cancer patients: association with lung cancer susceptibility and prognosis, Clin. Cancer Res. 5 (1999) 129 – 134. [25] A. Weston, C. Pan, B.H. Ksieski, S. Wallenstein, G.S. Berkowitz, P.I. Tartter, I.J. Bleiweiss, S.T. Brower, R.T. Senie, M.S. Wolff, p53 haplotype determination in breast cancer, Cancer Epidemiol., Biomarkers Prev. 6 (1997) 105 – 112. [26] A. Yang, M. Kaghad, D. Caput, F. McKeon, On the shoulders of giants: p63, p73 and the rise of p53, Trends Genet. 18 (2002) 90 – 95. [27] Z.W. Zhang, P. Newcomb, A. Hollowood, R. Feakins, M. Moorghen, A. Storey, M.J.G. Farthing, D. Alderson, J. Holly, Age-associated increase of codon 72 arginine p53 frequency in gastric cardia and noncardia adenocarcinoma, Clin. Cancer Res. 9 (2003) 2151 – 2156.