The molecular basis for most non-HNPCC familial colorectal cancer cases is unknown, but there is increasing evidence that common genetic variants may play a role. We investigated the contribution of polymorphisms in two genes implicated in the pathogenesis of colorectal cancer, cyclin D1 (CCND1) and E-cadherin (CDH1), to familial and sporadic forms of the disease. The CCND1 870A/G polymorphism is thought to affect the expression of CCND1 through mRNA splicing and has been reported to modify the penetrance of HNPCC. Inactivation of E-cadherin is common in colorectal cancer, and truncating germline mutations have been reported to confer susceptibility to colorectal as well as diffuse gastric cancer. The −160A/C CDH1 polymorphism appears to affect expression of CDH1 and may therefore also confer an increased risk. We found a significantly higher frequency of CCND1 870A allele in 206 familial cases compared to 171 controls (P=0.03). Odds ratios in heterozygotes and homozygotes were 1.7 (95% CI: 1.0–2.66) and 1.8 (95% CI: 1.0–3.3) respectively. The difference was accounted for by an over-representation of A allele in non-HNPCC familial cases (P=0.007). Over-representation of the CCND1 A allele was also seen in sporadic colorectal cancer cases compared to controls but this did not attain statistical significance (P=0.08). No significant differences between the frequency of CDH1 −160A/C genotypes in familial, sporadic colorectal cancer cases and controls were seen, although a possible association between the low expressing A allele and right-sided tumours was detected in familial cases.
Colorectal cancer (CRC) is one of the commonest cancers in the western world. A recent twin study indicates that ∼35% of all colorectal cancer can be ascribed to inherited genetic susceptibility (Lichtenstein et al., 2000). However, less than 5% of cases can be ascribed to dominant syndromes such as familial adenomatous polyposis coli (FAP) and hereditary non-polyposis colon cancer syndrome (HNPCC), which are associated with mutations in the APC tumour suppressor gene or mismatch repair genes (such as MSH2 or MLH1) respectively (Lynch and de la Chapelle, 1999). Germline mutations detected in FAP and HNPCC kindreds are highly penetrant. Polymorphisms in the APC tumour suppressor gene and mismatch repair genes have been identified, but they have generally been considered to have little pathogenic potential. This may not be the case. Thus Laken et al. (1997) first drew attention to a possible relationship between the I1307K APC variant and colorectal cancer risk in the Ashkenazi population. The hypothesis that common variants of genes implicated in the pathogenesis of familial and sporadic colorectal cancer may act as low penetrance susceptibility genes led us to examine whether polymorphic variants in CDH1 or CCND1 might affect colorectal cancer risk.
Somatic inactivation of CDH1 by mutations or promoter methylation is frequent in colorectal and other cancers (Becker et al., 1994; Risinger et al., 1994; Berx et al., 1995; Graff et al., 1995; Ilyas et al., 1997) leading to increased cell mobility and possibly increased activity of the β-catenin/TCF transcription factor complex in the nucleus (reviewed in Christophori and Semb, 1999; van de Wetering et al., 2001). Germline CDH1 mutations can cause familial diffuse-type gastric cancer (Guilford et al., 1998; Gayther et al., 1998; Richards et al., 1999) sometimes in association with early-onset colorectal cancer (EOCRC). Li et al. (2000) has recently reported that a common polymorphism of the CDH1 promoter (A/C at nucleotide −160 relative to the transcription start point) markedly affects CDH1 transcription and hence may confer susceptibility.
Cyclin D1 is a key cell cycle regulatory protein. The CCND1 gene is a direct target for transactivation by the β-catenin/LEF-1 pathway through a LEF-1 binding site in the cyclin D1 promoter (Shtutman et al., 1999). Elevated expression of CCND1 is seen in about a third of colonic adenocarcinomas (Arber et al., 1996) and expression of an anti-sense CCND1 cDNA suppresses the growth of colon cancer cells in nude mice, indicating a role for cyclin D1 in colorectal tumorigenesis (Arber et al., 1997). CCND1 mRNA is alternately spliced between exon 4 and 5 to give two transcripts, which occur simultaneously in a number of tissues (Betticher et al., 1995). The spliced transcript contains exons 4 and 5; the unspliced transcript reads through the end of exon 4 into intron 4. The splicing is modulated by a A/G polymorphism at nucleotide 870 (codon 242) in the splice donor region. Transcripts arising from the A allele are less likely to be spliced than those from the G allele. Both the spliced and unspliced transcripts encode proteins that contain the functional cyclin box (amino acids 55–161). However the unspliced transcript does not contain the exon 5 sequence, encoding a PEST destruction box responsible for the rapid turnover of the protein. This may confer a longer half-life on the alternate protein (Betticher et al., 1995). At the outset of this study, the CCND1 870A/G sequence variant had been evaluated as a modifier of colorectal cancer tumorigenesis in two previous investigations: Kong et al. (2000) reported that HNPCC gene carriers with an AG or AA genotype developed colorectal cancer ∼10 years earlier than patients with the GG genotype. In addition, McKay et al. (2000) did not detect an association between the 870A/G and survival in patients with sporadic colorectal cancer.
To examine the relationship between 870A/G CCND1 and CDH1 −160A/C status and colorectal cancer risk we carried out a case-control study of these two polymorphisms in familial and non-familial disease. The CCND1 870A was significantly over-represented in cases, especially in non-HNPCC familial colorectal cancer patients, suggesting that this variant acts as a low penetrance colorectal cancer susceptibility gene.
Cyclin D1 genotypes in familial (HNPCC and non-HNPCC) colorectal cancer
We compared the cyclin D1 genotype frequencies at the CCND1 870A/G sequence variant in 206 colorectal cancer patients with a positive family history to those observed in control samples. The distribution of alleles in controls was in Hardy-Weinberg equilibrium (χ2=0.09, 1 df. P=0.76). The frequency of the A allele homo- or heterozygotes in familial cases (155/206) was higher than that in 171 controls (111/171, P=0.031). Compared to CCND1 870 GG homozygotes, the AG genotype was associated with a 1.7 increased risk (CI 1.0–2.7) of colorectal cancer and the AA genotype with a 1.8 increase in risk, (1.0–3.3).
The distribution of genotypes in cases with a family history of HNPCC (n=99) was not significantly different from controls (Table 1). However, the A allele carriers (hetero- or homozygotes) were significantly more common in non-HNPCC familial cases than in controls (86/107 versus 111/171, P=0.007) (Table 1). In non-HNPCC familial cases both AG and AA were associated with significantly increased risks, 2.2 (95% CI: 1.2–4.0) and 2.2 (95% CI: 1.1–4.6) respectively. The frequency of AA and AG genotypes was similar in non-HNPCC with early onset (<50 years) colorectal cancer (42/49) and in later onset cases (44/58).
Cyclin D1 genotypes in sporadic colorectal cancer
Comparison of the cyclin D1 genotype frequencies at the CCND1 870A/G sequence variant in 128 patients with sporadic colorectal cancer and in 171 controls did not show a significant association between the A allele and colorectal risk although there was a greater frequency of A allele hetero- and homozygotes in sporadic cases than in controls. All five sporadic cases aged <30 years had an AA or AG genotype but there was no evidence that possession of the A allele was associated with an earlier age at diagnosis in sporadic cases (mean ages (±s.d.) AA=57.3±17.9, AG=57.0±16.6 and GG=51.7±14.7 years).
CDH1 −160A/C sequence variant in familial colorectal cancer
The distribution of genotypes for the −160A/C CDH1 sequence variant was compared in 162 familial cases of colorectal cancer and in 171 controls. The distribution of CDH1 −160 alleles in controls was in Hardy-Weinberg (χ2=3.30, 1.df, P=0.06). There were no significant differences between the CDH1 −160 genotypes of controls and sporadic or familial cases (see Table 1).
As loss of E-cadherin function might be more important in the pathogenesis of right-sided tumours than left-sided tumours (see later) we compared tumour site with CDH1 −160A/C genotype in 140 familial cases in which this information was available. Four/five (80%) AA homozygotes had a right-sided tumour compared to 22/55 (40%) of AC heterozygotes and 23/80 (28.8%) in CC homozygotes. The observed association between right-sided tumours and CDH1 −160 sequence variant genotype in familial cases was statistically significant (χ2=6.43 P=0.04).
The association between the CDH1 −160A allele and right-sided tumours was present in both the HNPCC and non-HNPCC subgroups of familial cancer but was not statistically significant because of the smaller numbers. Thus 75% (14/19) HNPCC cases with an AA or AC genotype had a right-sided tumour compared to 56% (14/25) CC cases; and 29% (12/41) AA or AC genotype non-HNPCC familial cases had a right-sided tumour compared to 16% (9/55) with a CC genotype.
CDH1 −160A/C sequence variant in sporadic colorectal cancer
The distribution of genotypes for CDH1 −160A/C in 128 sporadic cases analysed was similar to that in controls (see Table 1). As the A allele appeared to be associated with right-sided tumours in familial cases we looked for an association between CDH1 genotype and tumour site. Although three of four cases (75%) with an AA genotype had right-sided tumours compared to 35 of 121 (28.9%) with AC or CC genotypes, this difference was not statistically significant (P=0.08).
We detected an overrepresentation of the CCND1 870A alleles in familial and sporadic colorectal cancer cases compared to controls. Among all cancer cases the frequency of CCND1 870A allele homo- and heterozygotes was significantly greater than in controls (OR 1.58; 95% CI: 1.06–2.37). This supports our prior hypothesis that the CCND1 870A allele may represent a low penetrance colorectal cancer susceptibility allele similar to that suggested for the APC I1307K variant in Ashkenazi Jews and E1317Q in non-Ashkenazi populations (Laken et al., 1997; Frayling et al., 1998). As with I1307K and E1317Q, we found that overrepresentation of CCND1 870A carriers was greater in familial cases than in sporadic patients. While the risk of colorectal cancer associated with this polymorphism may be modest, the possible impact on the overall burden of colorectal cancer may be significant, possibly contributing to ∼24% of all cases.
Our hypothesis predicted that low-penetrance variants such as CCND1 870A would not be expected to be associated with large kindreds containing > three individuals and a high risk of colorectal cancer, but would perhaps be predicted to be most relevant to small clusters of cases (e.g. two affected individuals) or moderate sized families with evidence of incomplete penetrance. Therefore the observation that the overrepresentation of CCND1 870A carriers in familial cases was accounted for by a markedly increased frequency in non-HNPCC cases (rather than in HNPCC cases) would be consistent with our original hypothesis. While the frequency of APC I1307K in Ashkenazi populations is <10%, the frequency of CCND1 870A homo and heterozygotes in our controls (and those reported by others) is ∼65%. Hence a common variant such as CCND1 870A may well account for a greater proportion of cases.
Transcripts originating from CCND1 870A alleles are less likely to produce a spliced mRNA containing exon 5 than those transcribed from 870G alleles. However given that both transcripts are produced in all samples including those homozygous for AA and GG it appears that both A and G alleles can encode both transcripts to varying extents. The terminal sequence of the cyclin D1 variant lacking exon 5 has no PEST rich sequence, suggesting that the half-life of this protein might be longer. Thus, CCND1 870A alleles are likely to produce more of a stable cyclin D1 protein than 870G alleles.
Four previous studies have investigated possible associations between the CCND1 870A/G variant and colorectal cancer. Kong et al. (2000) reported an earlier onset of colorectal cancer in HNPCC gene carriers with an AG or AA genotype than in patients with the GG genotype. Our study was not designed to investigate this hypothesis (e.g. we did not study unaffected gene carriers) but the concept that the CCND1 A allele is a low-penetrance susceptibility factor would be consistent with the observations of Kong et al. (2000). Recently, Bala and Peltomaki (2001) reported that although they did not find an association between age at onset of colorectal cancer in HNPCC gene carriers and CCND1 870A/G genotype, they did find an association between age at onset and CCND1 splicing variant expression. Kong et al. (2001) have recently reported a hospital-based case-control study of 156 subjects with colorectal cancer (7% of whom met the Amsterdam criteria). They found an over-representation of the AA genotype in all colorectal cancer cases (OR ∼2.6). Although our study differs from that of Kong et al. (2001) in a number of variables, both studies implicate CCND1 870A/G genotype in colorectal cancer susceptibility. While an association between CCND1 870A/G genotype and survival might complicate the interpretation of our results, in a previous report no evidence for such an association was found (McKay et al., 2000). In addition to studies of colorectal cancer, Platz et al. (2000) reported that in families with hereditary melanoma carriers of A allele had an average age at melanoma diagnosis 10 years earlier than those patients homozygous for a G allele.
We did not find any strong indication that CDH1 −160A/C variant status influenced susceptibility to colorectal cancer. To our knowledge there are no other studies of this variant and colorectal cancer susceptibility, but in a small study of 24 familial gastric cancer cases there was no difference in CDH1 −160A/C variant genotypes between familial cases and normal controls (Yoon et al., 2001). Similarly, MacLeod et al. (2001) reported no significant differences between the CDH1 −160A/C genotypes of prostate cancer cases (n=67) and controls (n=79), although there was an association between the presence of the C allele and a higher Gleason's score suggesting that the CDH1 −160A/C genotype might influence transition of prostate carcinoma to a metastatic phenotype. We were prompted to investigate possible associations between CDH1 −160A/C genotype and colorectal cancer site by previous studies which had demonstrated that the left-side of the colon is relatively protected from E-cadherin downregulation by the compensatory increase in P-cadherin which can substitute for many functions of E-cadherin (Jankowski et al., 1998; Hardy et al., 2002). In addition we have shown that P-cadherin is expressed very early in colorectal dysplasia and will also abrogate some of the biological effects of decreased or mutant E-cadherin, namely decreased adhesion and increased migration (Hardy et al., 2002). Thus we reasoned that the low expressing CDH1 −160A allele might be likely to preferentially influence the pathogenesis of right-sided tumours. Clearly the association of CDH1 −160A/C genotype with right-sided tumours in familial cases in our study is preliminary and requires confirmation, but this finding is compatible with the known functional significance of the A allele and the differences in E-cadherin biology between the proximal and distal large bowel.
We have identified a significant association between colorectal cancer susceptibility and a functionally significant variant in a candidate colorectal cancer gene. Further studies are required to confirm this association and to define the absolute risks of colorectal cancer in different CCND1 genotype subgroups. Such studies will provide a basis for determining whether CCND1 870A/G genotyping will be useful for targeting colorectal cancer screening programmes. A major advance in the management of non-HNPCC familial colorectal cancer would be the ability to identify individuals at risk of right-sided colonic tumours who require colonoscopy rather than flexible sigmoidoscopy. Hence, further studies of the apparent association of right-sided tumours with the CDH1 −160A allele in genetically susceptible individuals are warranted, because of the potential clinical implications of this finding.
Materials and methods
Peripheral blood samples for DNA extraction were obtained from 334 patients with colorectal cancer. Familial colorectal cancer cases (n=206; mean age (±s.d.) 49.0±12.42 years; males:females 1.11) were ascertained from five genetics centres within the UK. Family histories were obtained by interview and confirmed by reference to hospital records, cancer registry records or death certificates wherever possible. Sporadic colorectal cancers (n=128; mean age (±s.d.) 55.6±12.95 years; males: females 1.46) were ascertained randomly through population-based surveys of colorectal cancer in Eastern and Central England. Family history was obtained by interview. Results of investigations for microsatellite instability in a subset of sporadic and familial cases have been reported previously (Brassett et al., 1996; Verma et al., 1999). A diagnosis of HNPCC was made if (a) family history satisfied Amsterdam or modified Amsterdam criteria (Lynch and de la Chapelle, 1999) or (b) a germline mismatch repair gene mutation was identified (for details see http://www.bham.ac.uk/ICH/genetics.htm). A diagnosis of non-HNPCC familial colorectal cancer was made if the proband had one or more close relatives with colorectal cancer but did not satisfy HNPCC diagnostic criteria: The commonest reasons for not satisfying Amsterdam criteria were all affected patients aged >50 years at diagnosis or only two cases of colorectal cancer (see Brassett et al., 1996). Control blood samples were collected from 171 non-cancer patients (mean age (±s.d.) 50.7±12.95; males: females=1.11) undergoing genetic testing for unrelated disorders. The mean ages of familial cases and controls were similar (P=0.2), but controls were significantly younger than sporadic cases (P<0.01).
Detection of polymorphisms
The CDH1 −160C/A promoter polymorphism was genotyped using polymerase chain reaction restriction fragment length polymorphism analysis. A 259 bp fragment was generated using the primers PF1: AGTGAGCTGTGATCGCACCACT and PR2: CCACCCGGCCTCGCATAGACG. PCR fragments were generated in an OmnE thermal cycler (Hybaid Ltd) in a buffer of 10 mM Tris-HCl (pH 8.8), 50 mM KCl, 0.01% gelatin, 0.2 mM dNTPs, 1 mM MgCl2 with 5 pmoles each primer plus 0.3 U Taq polymerase (Life Technologies Ltd) and 100 ng DNA per 15 μl reaction at an annealing temperature of 67°C. The 259 bp PCR product was then cut with the enzyme HincII according to the manufacturer's instructions (New England Biolabs Ltd). The PCR fragment obtained containing the −160 bp C allele is not cleaved by the enzyme. The fragment containing the A allele is cleaved into pieces of 234 bp and 25 bp. The cleavage products were then differentiated on a 10% polyacrylamide gel Figure 1). DNA sequencing analysis on a subset of samples was performed on an ABI 377 to confirm the restriction digests.
Cyclin D1 genotyping
The A/G CCND1 polymorphism at nucleotide 870 (codon 242) in exon 4 was also genotyped using polymerase chain reaction restriction fragment length polymorphism analysis. A 167 bp fragment of the CCND1 was amplified by PCR as above except using the primers C26F: GTGAAGTTCATTTCCAATCCGC and C27R: GGGACATCACCCTCACTTAC (Betticher et al., 1995) and an annealing temperature of 57°C. The 167 bp PCR product was then cut with the enzyme ScrF1 according to the manufacturer's instructions (New England Biolabs). The PCR fragment obtained from the allele containing the A allele at position 870 is not cleaved by the enzyme whereas the G allele at position 870 is cleaved into pieces of 145 and 22 bp. The cleavage products were visualized on a 10% polyacrylamide gel (Figure 1) and digest products confirmed by DNA sequencing.
To test for population stratification, the distribution of genotypes in controls was tested for a departure from Hardy-Weinberg equilibrium. The relationship between CCND1 and CDH1 genotypes and colorectal cancer risk was assessed by means of the odds ratio (OR) with 95% confidence limits calculated by logistic regression adjusting for age and sex. For each polymorphism, ORs were determined for individual and grouped genotypes. The significance of the difference in frequency of genotypes between cases and controls was assessed by Fisher's exact test. All statistical manipulations were undertaken using the program STATA (Version 6.0, Stata Corporation, 702 University Drive East, College Station, Texas 77840 USA URL:http://www.stata.com).
Arber N, Hibshoosh H, Moss SF, Sutter T, Zhang Y, Begg M, Wang S, Weinstein IB, Holt PR . 1996 Gastroenterology 110: 669–674
Arber N, Doki Y, Han EK, Sgambato A, Zhou P, Kim NH, Delohery T, Klein MG, Holt PR, Weinstein IB . 1997 Cancer Res. 57: 1569–1574
Bala S, Peltomaki P . 2001 Cancer Res. 61: 6042–6045
Becker KF, Atkinson MJ, Reich U, Becker I, Nekarda H, Siewert JR, Hofler H . 1994 Cancer Res. 54: 3845–3852
Berx G, Cleton-Jansen AM, Nollet F, de Leeuw WJF, van de Vijver MJ, Cornelisse C, van Roy F . 1995 EMBO J. 14: 6107–6115
Betticher DC, Thatcher N, Altermatt HJ, Hoban P, Ryder WDJ, Heighway J . 1995 Oncogene 11: 1005–1011
Brassett C, Joyce JA, Froggatt NJ, Williams G, Furniss D, Walsh S, Miller R, Evans DG, Maher ER . 1996 J. Med. Genet. 33: 981–985
Christophori G, Semb H . 1999 Trends Biol. Sci. 24: 73–76
Frayling IM, Beck NE, Ilyas M, Dove-Edwin I, Goodman P, Pack K, Bell JA, Williams CB, Hodgson SV, Thomas HJ, Talbot IC, Bodmer WF, Tomlinson IP . 1998 Proc. Natl. Acad. Sci. USA 95: 10722–10727
Gayther SA, Gorringe KL, Ramus SJ, Huntsman D, Roviello F, Grehan N, Machado JC, Pinto E, Seruca R, Halling K, MacLeod P, Powell SM, Jackson CE, Ponder BAJ, Caldas C . 1998 Cancer Res. 58: 4086–4089
Graff JR, Herman JG, Lapidus RG, Chopra H, Xu R, Jarrard DF, Isaacs WB, Pitha PM, Davidson NE, Baylin SB . 1995 Cancer Res. 55: 5195–5199
Guilford P, Hopkins J, Harraway J, McLeod M, McLeod N, Harawira P, Taite H, Scoular R, Miller A, Reeve AE . 1998 Nature 392: 402–405
Hardy RG, Tselepis C, Hoyland J, Wallis Y, Pretlow TP, Talbot I, Sanders DSA, Matthews G, Morton D, Jankowski JA . 2002 Gut in press
Ilyas M, Tomlinson IP, Hanby A, Talbot IC, Bodmer WF . 1997 Gut 40: 654–659
Jankowski J, Bedford F, Boulton R, Cruikshank N, Hall C, Elder J, Allan R, Forbes A, Kim Y, Wright N, Sanders S . 1998 Lab. Invest. 78: 1155–1167
Kong S, Amos CI, Luthra R, Lynch PM, Levin B, Frazier ML . 2000 Cancer Res. 60: 249–252
Kong S, Wei Q, Amos CI, Lynch PM, Levin B, Zong J, Frazier ML . 2001 J. Natl. Cancer Inst. 93: 1106–1108
Laken SJ, Petersen GM, Gruber SB, Oddoux C, Ostrer H, Giardiello FM, Hamilton SR, Hampel H, Markowitz A, Klimstra D, Jhanwar S, Winawer S, Offit K, Luce MC, Kinzler KW, Vogelstein B . 1997 Nat. Genet. 17: 79–83
Li LC, Chui RM, Sasaki M, Nakajima K, Perinchery G, Au HC, Nojima D, Carroll P, Dahiya R . 2000 Cancer Res. 60: 873–876
Lichtenstein P, Holm NV, Verkasalo PK, Iliadou A, Kaprio J, Koskenvuo M, Pukkala E, Skytthe A, Hemminki K . 2000 N. Engl. J. Med. 343: 78–84
Lynch HT, de la Chapelle A . 1999 J. Med. Genet. 36: 801–818
MacLeod SL, Plaxco JR, Shinde A, Lang NP . 2001 Proc. Amer. Assoc. Cancer Res. 42: 149–
McKay JA, Douglas JJ, Ross VG, Curran S, Murray GI, Cassidy J, McLeod HL . 2000 Int. J. Cancer 88: 77–81
Platz A, Hansson J, Ringborg U . 2000 Am. J. Hum. Genet. 67: (Suppl 2) 406–
Richards FM, Mckee SA, Rajpar MH, Cole TRP, Evans DGR, Jankowski JA, McKeown C, Sanders DSA, Maher ER . 1999 Hum. Mol. Genet. 8: 607–610
Risinger JI, Berchuck A, Kohler MF, Boyd J . 1994 Nature Genet. 7: 98–102
Shtutman M, Zhurinsky J, Simcha I, Albanese C, D'Amico M, Pestell R, Ben-Ze'ev A . 1999 Proc. Natl. Acad. Sci. USA 96: 5522–5527
van de Wetering M, Barker N, Harkes IC, van der Heyden M, Dijk NJ, Hollestelle A, Klijn JG, Clevers H, Schutte M . 2001 Cancer Res. 61: 278–284
Verma L, Maher ER, Kane MF, Schmeits J, Kolodner RD, Brassett C, Evans DGR . 1999 J. Med. Genet. 36: 678–682
Yoon KA, Ku JL, Kang HC, Park JG . 2001 Proc. Amer. Assoc. Cancer Res. 42: 343–
We thank the many patients, clinicians and scientists who contributed to this study. We are grateful to the Birmingham United Hospitals Endowment Fund (TR Porter, FM Richards, ER Maher, JA Jankowski) and The Wellcome Trust (JA Jankowski) for financial support.
About this article
Cite this article
Porter, T., Richards, F., Houlston, R. et al. Contribution of cyclin d1 (CCND1) and E-cadherin (CDH1) polymorphisms to familial and sporadic colorectal cancer. Oncogene 21, 1928–1933 (2002). https://doi.org/10.1038/sj.onc.1205245
- colorectal cancer
- cyclin D1
The epithelial cadherin –160C/A polymorphism is associated with decreased risk of colorectal cancer: a case–control study
Clinical and Experimental Medicine (2020)
Identification of Prognostic Biomarker Signatures and Candidate Drugs in Colorectal Cancer: Insights from Systems Biology Analysis
MTHFR, TS and XRCC1 genetic variants may affect survival in patients with myelodysplastic syndromes treated with supportive care or azacitidine
The Pharmacogenomics Journal (2018)
Association of TP53 codon 72 and CDH1 genetic polymorphisms with colorectal cancer risk in Bangladeshi population
Cancer Epidemiology (2017)