Radiol Oncol 1998; 32(1): 35-39
Genetic polymorphisms of xenobiotic metabolizing enzymes in human colorectal cancer
Vita Dolžan1, Metka Ravnik-Glavač1'2, Katja Breskvar1
1Institute of Biochemistry, Medical Faculty, 2Laboratory of Molecular Pathology, Institute of Pathology, Medical Faculty, Ljubljana, Slovenia
It was proposed that both hereditary and environmental factors contribute to the development of colorectal cancer (CRC). Carcinogenic polycyclic aromatic hydrocarbons (PAHs) from food ar tobacco smoke can form DNA adducts and thus initiate carcinogenesis after metabolic activation via cytochrome P4501A1 (CYP1A1). Intermediate metabolites are detoxified by conjugation with glutathione 5-trans-ferases. Our aim was to look far inherited metabolic susceptibility to CRC. We used PCR- based genotyp-ing approach to determine the frequencies of polymorphic alleles of two cytochromes P450 (CYP2D6 and CYP1A1) and two glutathione 5-transferases (GSTMl and GSTTl) in DNA samples from 31 sporadic, 25 familial CRC cases and 73 healthy controls. The difference in frequencies of poor metabolisers due to CYP2D6 gene polymorphism was close to the limit of statistical significance between sporadic CRC and healthy control group (x2 = 5.52, m=2, p=0.06) despite the small sample size. The frequencies of either CYP1A1 Mspl, GST Ml ar GST Tl genotypes were not significantly different in both CRC cases and in controls. Although our study suggests some difference in metabolic susceptibility between sporadic and familial CRC, further studies are needed to investigate the combined effect of polymorphic genes involved in carcinogen metabolism in a larger group of patients with defined exposure to dietary carcinogens and smoking.
Key words: colorectal neoplasms-genetics; polymorphism (genetics); cytochrome P-450CYP1A1; cytochrome P-450 CYP2D6; gluthatione transferases
Introduction
It is generally accepted that both hereditary and environmental factors contribute to the development of colorectal cancer (CRC). A genetic model far colorectal tumorigenesis
Correspondence to: Dr. Vita Dolžan, MD, Institute of Biochemistry, Medical Faculty, Vrazov trg 2, 1000 Ljubljana, Slovenia. Phone: +386 61 1320 019; Fax: +386 61 1320 016; E-mail: dolzan@ibmi.mf.uni-lj.si
suggested by Faeron and Vogelstein proposed that colorectal tumors arise due to mutations in oncogenes and tumor suppressor genes.1 The sequences of adenoma to carcinoma transition are well established.2 However, the earliest events in human colorectal tumor formation are not well defined. As far as environmental factors are concerned a large proportion of human cancers is known to be caused by chemicals from the environment.3
36
Dolžan Vet al
A strong association has been observed between colorectal cancer and consumption of broiled and grilled meat. Exposure of food to pyrolysis can lead to the formation of compounds such as polycyclic aromatic hydrocarbons (PAHs), heterocyclic amines and others.4 In addition to diet, tobacco smoking was also associated with colorectal cancer.5 It is possible that PAHs in food exposed to pyrolysis and in tobacco smoke can play a role in some of the genetic alterations leading to colorectal tumor development.5
Most environmental chemicals have to be metabolically activated to their ultimate carcinogenic form and individual differences in metabolism of precarcinogenic substances exist due to polymorphisms in many enzymes involved in this process. So, after sufficient exposure to carcinogen individual differences in its metabolism may result in different genetic susceptibility to the development of cancer.3
Carcinogens as well as other xenobiotics are usually metabolized in two consecutive steps. Phase I is an activation step and phase II is the detoxification step. Enzymes from the cytochrome P450 families 1, 2 and 3 (CYPl, CYP2 and CYP3 respectively) are the main enzymes involved in the oxidative activation of chemical compounds into reactive electrophilic form.3 Among others, CYP1A1 is involved in activation of variety of chemical carcinogens to their ultimate carcinogenic forms, particularly PAHs from tobacco smoke and PAHs and aromatic amines from food exposed to pyrolysis. CYP2D6 is involved in metabolism of many drugs and yet undefined carcinogens from tobacco smoke.6 In the detoxification reaction the activated substance is conjugated with some organic acid or intracellular glutathione to form inactive water-soluble metabolites. This reactions are mainly catalyzed by N-acetyl transferases (NATs) and glutathione S-trans-ferases (GSTs). There are four known mammalian classes of soluble GSTs: a, n and -i:,
which allow far a broad overlapping substrate specificity. Among them, GST x and -i: are polymorphic and as such of interest in molecular epidemiological studies. GST x is expressed in the liver, leukocytes, colon and other tissues, but it is absent in approximately half of the population due to a deletion (GSTM1 *null allele). It catalyses conjugation of epoxides with glutathione and may thus protect individuals against chemical mutagens or carcinogens, among others also carcinogens from tobacco smoke. Although the predominant role of GST is in detoxification, certain glutathione conjugation reactions, particularly those involving halogenated alka-nes and alkenes can result in the formation of reactive electrophiles that are toxic and carcinogenic.6
Initiation of tumorigenesis by a chemical compound in vivo requires at least three successive reactions: First, metabolic conversion of chemically inert compounds to a reactive electrophilic form has to occur. This is an obligatory initiation step in chemical carcino-genesis. Next, adducts have to form between these reactive metabolites and DNA, resulting in base changes or DNA rearrangements. Finally, these DNA alterations must be fixed and must lead to oncogene activation. Enhanced activity of phase I enzymes or decreased activity of phase II enzymes result in higher amount of activated carcinogen and increased formation of DNA adducts.3 This mechanism implies that the presence and activity of relevant xenobiotic metabolizing enzymes contribute to the risk of cancer development.3 However, genetic susceptibility factor only becomes relevant if sufficient exposure to carcinogen occurs. Risk factors far tumor development include the extent of environmental exposure to a procarcinogen as well as the host susceptibility factor due to the polymorphism of drug metabolizing enzymes involved in the formation of activated carcinogen.
In our study we wished to find out if
Genetic polymorphisms in colorectal cancer
37
genetic polymorphism of xenobiotic metabolism represents a risk factor for CRC. The frequency of polymorphic alleles of two cytochrome P450 dependent monooxygenas-es (CYP1A1 and CYP2D6) and two glutathione S-transferases (GST Ml and GST Tl), supposed to be associated with higher cancer risk, were determined in DNA samples from Slovenian sporadic and familial CRC patients and healthy controls.
Materials and methods
DNA samples from 31 sporadic and 25 familial CRC patients were analyzed. Seventy-three DNA samples from 107 healthy controls previously analyzed for CYP2D67 were included in the study. No clinical data regarding patient characteristics, the localization or stage of tumor or epidemiological data considering exposure were obtained at this stage of the study.
PCR- based genotyping approach was used to determine the frequencies of polymorphic alleles of two cytochrome P450-dependent monooxygenases (CYP2D6 and CYP1A1) and two glutathione S-transferases (GSTMl and GSTTl).
PCR followed by restriction with BstNI enzyme was used to distinguish between the wild type allele (CYP2D6wt) and deficiency allele CYP2D6B.8 Allele specific PCR with two pairs of primers was used to distinguish between the wild type allele and deficiency allele CYP2D6A.9
Table 1. CYP2D6 phenotype frequencies in CRC cases and controls
	Cases	CYP2D6	CYP2D6	CYP2D6
	(number)	EM	HEM	PM
CRC total	56	0.75	0.21	0.04
- sporadic	31	0.84	0.16	0.00
-familial	25	0.64	0.28	0.08
Controls	107	0.63	0.31	0.07
The polymorphism in 3' flanking region of CYP1A1 gene was analyzed by PCR and restriction with MspI.10
Triplex PCR was used to analyze for GST Ml and T1 null alleles. Conserved beta globin primers were used as internal control.11
Results and discussion
The frequencies of CYP2D6 wt, A and B alleles were 90.9 %, 7.6 % and 1.5 % in sporadic CRC cases and 80.8 %, 19.2 % and O% in familial CRC cases. The frequencies of CYP2D6 alleles in the healthy control population studied previously were as follows: CYP2D6A 1 %, CYP2D6B 21 %, and CYP2D6wt 78 % of all alleles.7 CYP2D6 phe-notype frequencies in CRC patients and in controls are presented in Table 1. The pheno-type was predicted according to the genotype }2 Individuals homozygous for two deficiency alleles were considered poor (slow) metabolizers (PM), heterozygotes for one deficiency allele were considered heterozygous extensive metabolizers (HEM) and homozygotes for two wild type alleles were considered extensive metabolizers (EM). As presented in Table 1, no PM of debrisoquine was identified in sporadic CRC cases, while 8 % of familial CRC and 7 % of healthy controls were identified as PM. The difference in PM frequencies between sporadic CRC and healthy control group is close to the limit of statistical significance (x2 = 5.52, m=2, p= 0.06) despite the small sample size. There
Table 2. CYPlAl MspI genotype frequencies in CRC cases and controls
	Cases	CYPlAl	CYPlAl	CYPlAl
	(number)	A	B	C
CRC total	56	0.86	0.12	0.02
- sporadic	31	0.84	0.13	0.03
-familial	25	0.88	0.12	0.00
Controls	73	0.81	0.18	0.01
38
Dolžan Vet al
was no statistically significant difference in CYP2D6 phenotype distribution between familial CRC cases and healthy controls.
CYP1A1 Mspl genotype frequencies in CRC patients and in controls are presented in Table 2. The predominant genotype in both CRC cases as well as healthy controls is genotype A which indicates homozygotes for two wild type alleles in which the Mspl restriction site is absent. Genotype B indicates heterozygotes for one wild type allele and one allele with mutation creating Mspl restriction site. Genotype C which indicates homozygotes for two alleles with Mspl restriction site present is very rare in both CRC cases and in controls. It was shown that CYP1A1 in-ducibility and enzymatic activity is increased in individuals with genotype C.10 Genotype C frequency seemed to be slightly higher in sporadic CRC cases than in familial CRC cases or controls, however, the difference in frequencies of CYP1A1 Mspl genotypes was not statistically significant (x2 = 1.67, m=4, p=O.8O).
GST Ml and GST T1 genotype frequencies in CRC patients and in controls are presented in Table 3. GST Ml and GST T1 positive genotype indicates individuals homozy-gous or heterozygous for the presence of GST Ml or GST T1 gene respectively, while the null genotype indicates individuals homozy-gous for the deletion of GST Ml or GST T1 gene respectively. The frequencies of GST Ml as well as GST T1 genotypes were not significantly different in both CRC cases and controls.
On the basis of our results we can condude that inherited metabolic susceptibility to carcinogens from the environment seems to be higher in sporadic than in familial CRC. Polymorphic CYP2D6 gene probably has some role in colorectal carcinogenesis in our population. Enhanced metabolic activation of PAHs was observed in CYP1A1 Mspl C genotype10 but its frequency in Slovenian population is probably to low to contribute significantly to colorectal tumor formation. It is interesting that this polymorphism was related to an increased risk for in situ CRC in the Japanese and Hawaiians, but not in the Caucasians. 13 The discrepancy between the above and our study may be due to a low frequency of CYP1A1 susceptibility allele in our as well as in other Caucasian populations, which limits the statistical power of analysis in small groups of patients. Another possible explanation is that PAHs may influence the earliest genetic alterations leading to colorec-tal tumorigenesis and their role may be masked in advanced disease. Our results also indicate that the polymorphic GST genes probably do not represent a significant risk factor for colorectal carcinogenesis.
However, we must keep in mind that cancer is a polygenic disease and the penetrance of any single gene is not sufficient to produce an observable effect. Colorectal cancer results from multiple mutations and it may be difficult to demonstrate a direct involvement of P45Os and GSTs in colorectal cancer, especially in advanced disease. However, genetic differences in metabolism of carcino-
Table 3. GST Ml and GST T1 genotype frequencies in CRC cases and controls
Cases	GST Ml	GSTMl	GSTTl	GSTTl
(number)	positive	nul!	positive	null
CRC total	56	0.50	0.50	0.79	0.21
-sporadic	31	0.52	0.48	0.81	0.19
-familial	25	0.48	0.52	0.76	0.24
Controls	73	0.48	0.52	0.78	0.22
Genefic polymorphisms in colorectal cancer
39
gens may influence the earliest genetic alterations leading to colorectal tumorigenesis.
Although our study suggests a difference in metabolic susceptibility in sporadic and familial CRC, further studies are needed to investigate the combined effect of polymorphic genes involved in carcinogen metabolism in a larger group of patients with defined exposure to dietary carcinogens and smoking.
Acknowledgments
The authors would like to acknowledge dr. Koželj and Prof. Križman from the Univ. Dept. of Gastroenterology for their contribution of blood samples from familial colorectal cancer patients, and Prof. Golouh from the Institute of Oncology for his contribution of normal tissue samples of sporadic colorectal cancer patients. We would also like to acknowledge Prof. Bohinjec and Blanka Vidan - Jeras from the Tissue Typing Center, Blood Transfusion Center of Slovenia, for their contribution of DNA samples from healthy controls. This work was financially supported by the Ministry of Science and Technology of Slovenia.
6.	Wolf CR. Metabolic factors in cancer susceptibility. Cancer Surv 1990; 9: 437-74
7.	Dolzan V, Rudolf Z, Breskvar K. Human CYP2D6 gene polymorphism in Slovene cancer patients and healthy controls. Carcinogenesis 1995; 16: 2675-8.
8.	Gough AC, Miles JS, Spurr NK, Moss JE, Gaedik A, Eichelbaum M, Wolf CR. Identification of primary gene defect at the cytochrome P450 CYP2D6 locus. Nature 1990; 347: 773-5.
9.	Heim M, Meyer UA. Genotyping of poor metabo-lizers of debrisoquine by allele-specific PCR amplification. Lancet 1990; 336: 529-32.
10.	Hayashi S, Watanabe J, Nakachi K, Kawajiri K. Genetic linkage of lung cancer associated Mspl polymorphisms with amino acid replacement in the heme binding region of the human cytochrome P450 1A1 gene. J Biochem 1991; 110: 407-11.
11.	Chen C-L, Liu Q, Relling MV. Simultaneus characterization of GST Ml and T1 polymorphisms by polymerase chain reaction in American whites and blacks. Pharmacogenetics 1996; 6: 187-91.
12.	Wolf CR, Smith CAD, Gough AC. Relationship between the debrisoquine hydroxylase polymorphism and cancer susceptibility. Carcinogenesis 1992; 13: 1035-8.
13.	Sivaraman L, Leatham MP, Yee J, Wilkens LR, Lau AF, LeMarechand L. CYP1A1 genetic polymorphisms and in situ colorectal cancer. Cancer Res 1994; 54: 3692-5.
References
1.	Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990; 61:759-67.
2.	Vogelstein B, Kinzler KW. The multistep nature of cancer: Trends Genet 1993; 9: 138-41.
3.	Kawajiri K, Fujii-Kuriyama Y. P450s and human cancer. ]pn J Cancer Res 1991; 82: 1325-35.
4.	Sugimura T, Sato S. Mutagens - carcinogens in foods. Cancer Res 1985; 34: 2415S-21S.
5.	Giovannucci E, RimEB, Stampfer MJ, Colditz GA, Ascherio A, Kearney J et al. A prospective study of cigarette smoking and risk of colorectal adenoma and colorectal cancer in U.S. men. J Natl Cancer Inst 1994; 86: 183-91.