- Email: [email protected]
The cytochrome P450 2D6 (CYP2D6) gene polymorphism among breast and head and neck cancer patients
Clinica Chimica Acta 296 (2000) 101–109 www.elsevier.com / locate / clinchim
The cytochrome P450 2D6 (CYP2D6) gene polymorphism among breast and head and neck cancer patients ˇ ´ a , Ana Maria Ivanisevic ˇ ´ a, Elizabeta Topic´ a , *, Mario Stefanovic b c ˇ ˇ´ Rajka Petrinovic´ , Ivica Curcic a
Clinical Institute of Chemistry, School of Medicine, University of Zagreb and Sestre Milosrdnice University Hospital, Vinogradska c. 29, 10000 Zagreb, Croatia b Cancer Institute of Croatia, Zagreb, Croatia c ENT Department, Dubrava University Hospital, Zagreb, Croatia Received 17 May 1999; received in revised form 10 January 2000; accepted 24 January 2000
Abstract The prevalence of CYP2D6*3 and CYP2D6*4 alleles in normal controls and cancer patients was studied using the reliable PCR-SSCP method. In the control group (n 5 144), four subjects (2.8%) were found to carry CYP2D6*3 allele (heterozygote), while 30 (20.8%) subjects carried CYP2D6*4 allele (18.8% heterozygotes, 2.1% homozygotes). One (1.3%) of the breast cancer (BC) patients (n 5 76) carried CYP2D6*3 allele, but 24 (31.6%) carried CYP2D6*4 allele (26.3% heterozygotes, 5.3% homozygotes). In the head and neck cancer (HNC) group (n 5 56), two (3.6%) patients were heterozygous for CYP2D6*3 mutation and 15 (26.8%) for CYP2D6*4 mutation. Fourteen of 56 (25%) and one of 56 (1.8%) of these patients carried heterozygous and homozygous mutations, respectively. In controls, 2.1% were identified as poor metabolizers (PM), 76.4% as extensive metabolizers (EM), and 21.5% as intermediate heterozygotes (IEM). In BC group, 5.3, 27.6 and 67.1% were classified as PM, IEM and EM, respectively. In HNC group, the incidence of PM was 1.8, but as many as 28.6% were identified as IEM phenotypes. 2000 Elsevier Science B.V. All rights reserved. Keywords: CYP2D6 polymorphism; PCR-SSCP; Breast cancer; Head and neck cancer
*Corresponding author. Tel. / fax: 1 385-1-3768-280. ´ E-mail address: [email protected] (E. Topic) 0009-8981 / 00 / $ – see front matter 2000 Elsevier Science B.V. All rights reserved. PII: S0009-8981( 00 )00221-7
102
E. Topic´ et al. / Clinica Chimica Acta 296 (2000) 101 – 109
1. Introduction Most environmental toxic chemicals and carcinogens need to be metabolically activated to exert their toxic or carcinogenic effects. The most important group of activating enzymes is the large cytochrome P450 (CYP) ‘superfamily’. Of 15 human CYP isoenzymes characterized so far, eight forms of CYP (CYP1A1, 1A2, 2A6 2C9, 2D6, 2E1 and 3A4) have been shown to be polymorphic at the phenotype or genotype level, or both. Because an individual’s capability to metabolize these toxicants can be altered by carrying variant alleles, the genetic polymorphism of CYP enzymes has been proposed as a biomarker of susceptibility to environmental carcinogenesis and toxicity [1]. The association of CYP genetic polymorphism and human cancer risk has received increasing attention. Examples include CYP1A1 with lung and breast cancers [2–4], CYP2D6 with different types of cancer [5], and CYP2E1 with lung, liver and nasopharyngeal cancers [2,6–9]. However, many reports are controversial. One important factor in interpreting these results is that there are significant ethnic differences in frequency distribution of the CYP genetic polymorphism. The aim of this study was to assess the prevalence of CYP2D6*3 and *4 gene mutations in a group of healthy subjects and in patients with cancer etiologies by a reliable method for their identification. The CYP2D6 isoform metabolizing more than 25% of most common drugs has more than 20 mutant alleles identified to date. The mutant CYP2D6*3 allele with A 2637 deletion in exon 5 and mutant CYP2D6*4 allele with G 1934 → A splice site defect are among the most common ones, with a frequency of 2.7 and 28.6%, respectively. These mutations result in a decreased or no activity of CYP2D6 isoenzyme, leading to the poor metabolizer phenotype (PM). These individuals are at an increased risk of adverse side effects or therapeutic failure following drug treatment [10]. The CYP2D6 locus in humans has been identified as an array of at least three highly homologous genes, on the long arm of chromosome 22. In addition to the active CYP2D6 gene consisting of nine exons and eight introns, an inactive pseudogene, CYP2D8P, and an inactive homologue, CYP2D7, are present in a tandem array [11].
2. Material and methods
2.1. Subjects A total of 276 subjects, 132 cancer patients and 144 healthy volunteers, were included in the study. The cancer group consisted of 76 breast cancer (BC)
E. Topic´ et al. / Clinica Chimica Acta 296 (2000) 101 – 109
103
patients and 56 head and neck cancer (HNC) patients. Genotyping of the CYP2D6*3 and *4 alleles was performed on DNA isolated from EDTA-blood by the PCR-SSCP method. Whole blood samples (3 ml) were collected in sodium EDTA vacutainers, assigned a unique code, and stored at 2 208C. Leukocyte DNA was isolated by standard phenol / chloroform extraction and ethanol precipitation protocols using a slightly modified method by Topic and Gluhak [12].
2.2. PCR-SSCP procedure Target DNA (170 ng genomic DNA in 50 ml of PCR mixture) was amplified by polymerase chain reaction (PCR) [11] using PCR Core Kit (Roche Diagnostics, Mannheim, Germany) and specific amplification primers (MWG-Biotech, Ebersberg, Germany) in the Progene Thermal Cycler (Techne, Cambridge, England). The 2D6*3 mutation was detected by amplification of the region of interest with 2D6*3 allele primers (base pairs 4147–4416): forward, 59-gatgagctgctaactgagccc-39; reverse, 59-ccgagagcatactcgggac-39. The 2D6*4 mutation was detected with 2D6*4 allele primers (base pairs 3359–3713): forward, 59-aaatcctgctcttccgaggc-39; reverse, 59-gccttcgccaaccactcc-39; as described by Wolf et al. [13] and Vermes et al. [14]. Table 1 shows optimized conditions for CYP2D6*3 and CYP2D6*4 PCR according to chemicals and cycling [15]. To detect mutations in the CYP2D6 gene we used a strategy based on single-strand conformation polymorphism (SSCP) [16]. After the amplification, 3 ml of the PRC product were added to 7 ml of loading solution (formamide containing 10 mmol / l NaOH and 0.5% bromphenol blue), denatured at 958C for 5 min, quickly chilled on ice and applied (8 ml / lane) to GMAE, Mini Gels for SSCP (Elchrom Scientific, Switzerland). Electrophoresis was performed at a constant temperature of 88C in 30 mmol / l TAE, pH 7.5. The gels were stained with Table 1 Optimized conditions for CYP2D6*3 and CYP2D6*4 PCR Allele 2D6*3
Allele 2D6*4
Chemicals: Magnesium concentration (mmol / l) Primer concentration (mmol / l) DNTP concentration (mmol / l) Taq DNA polymerase concentration (U / ml)
0.80 0.15 0.20 0.025
1.20 0.20 0.20 0.025
Cycling: Cycle denaturation (temperature and time) Primer annealing (temperature and time) Primer extension (temperature and time) Cycle number
928C / 45 s 56.58C / 30 s 728C / 600 s 38
928C / 45 s 578C / 30 s 728C / 600 s 38
104
E. Topic´ et al. / Clinica Chimica Acta 296 (2000) 101 – 109
SYBR Green II (1:10 000 in 10 mmol / l TAE) for 1 h and photographed at 254 nm with a Polaroid 667 film. Negative quality controls generated by inclusion of all reagents except DNA and positive quality control containing a known genotype (obtained by the courtesy of Michelle Bon, Enschede, NL) were run with every run. Subjects homozygous for the wild-type allele were classified as extensive metabolizers, subjects carrying one wild-type and one mutant allele as intermediate extensive metabolizers, and subjects homozygous for mutant alleles as poor metabolizers [17].
3. Results The results of PCR-SSCP presented in Fig. 1 show reproducible and uniform band patterns for mutant and wild-type variants of CYP2D6*3 and *4 alleles. Five and six bands for mutant and wild-type variants of 2D6*3 and 2D6*4, respectively, were observed. One of these bands appeared to be distinctive for *3 allele of the heterozygous phenotype (no homozygous phenotype was found) and two for *4 allele, denoting a heterozygous and homozygous genotype. In the
Fig. 1. PCR-SSCP analysis of the CYP2D6 gene. (A) CYP2D6*3: lane 1, homozygote wt / wt (positive control); lanes 2 and 3, patients wt / wt; lane 4, positive control of *3 / wt mutant; lane 5, patient *3 / wt; and lane 6, CYP2D6*3 PCR non-denatured product. (B) CYP2D6*4: lane 1, homozygote wt / wt (positive control); lane 2, patient wt / wt; lane 3, *4 / wt mutant heterozygote (positive control); lane 4, patient *4 / wt; lane 5, *4 / *4 mutant homozytoge (positive control); lane 6, patient *4 / *4; and lane 7, CYP2D6*4 PCR non-denatured product. Arrows indicate bands that discriminate 2D6*3 allele of the heterozygous phenotype (A) and 2D6*4 allele of the heterozygous and homozygous phenotype (B).
E. Topic´ et al. / Clinica Chimica Acta 296 (2000) 101 – 109
105
PCR-SSCP approach, the genotypes previously detected by PCR-RFLP [15] were used as positive quality controls as follows: homozygous wild-type and mutant heterozygous for 2D6*3 allele, and wild- and mutant homozygote as well as mutant heterozygote for 2D6*4 allele. Comparison of the electrophoresis patterns of PCR-SSCP findings demonstrated that the SSCP analysis identified the previously described mutations of CYP2D6 gene and discriminated between the homozygous and heterozygous individuals. Study results showed that 23.6% (n 5 34) of 144 healthy volunteers carried mutant 2D6*3 and / or 2D6*4 alleles. The frequency of CYP2D6 alleles among healthy subjects and cancer patients is presented in Table 2. A total of 1.4% (4 / 288) of alleles were 2D6*3 alleles with a frame shift mutation in exon 5 and 11.4% (33 / 288) were 2D6*4 alleles with G to A mutation at intron 3–exon 4 junction. The genotypic frequencies of all studied groups were consistent with Hardy–Weinberg equilibrium. In BC (n 5 76) patients, 32.9% (n 5 25) of those genotyped for 2D6*3 and 2D6*4 alleles were found to carry mutant alleles. One of them (0.7%) was found to bear 2D6*3 allele, but 18.4% (28 / 152) had alleles with G to A mutation at intron 3–exon 4 junction. In HNC patients (n 5 56), as many as 26.8.2% (n 5 15) were found to carry mutant alleles. Among them, 1.8% (2 / 112) had 2D6*3 allele and 14.3% (16 / 112) 2D6 4* allele. Results of genotype distribution (Table 3), showed four of 144 (2.8%) and 27 / 144 (18.8%) control group subjects to be heterozygotes according to the frame shift mutation in exon 5 and G to A mutation at intron 3–exon 4 junction, respectively. In this group, there were no homozygous 2D6*3 mutants, however 2.1% (n 5 3) of homozygous 2D6*4 mutants were detected. In the BC group (n 5 76), 1.3% (n 5 1) heterozygous and no homozygous mutant 2D6*3 genotype was detected. Out of 24 patients bearing 2D6*4 allele, 26.3% (n 5 20) were heterozygotes and 5.3% (n 5 4) homozygotes. In the group of HNC patients (n 5 56), results of genotype distribution Table 2 Frequency of CYP2D6 alleles in healthy volunteers and patient groups
Healthy Breast cancer Head and neck cancer a
No. of alleles
Allele *3 a no. (%)
288 152 112
4 (1.4) 1 (0.7) 2 (1.8)
P value c
0.948 0.934
Allele *4 b no. (%) 33 (11.4) 28 (18.4 16 (14.3)
Allele *3, CYP2D6*3 allele with frame shift mutation in exon 5. Allele *4, CYP2D6*4 allele with G to A mutation at intron 3–exon 4 junction. c Statistical significance (P) was calculated using z-test. b
P value
0.098 0.448
E. Topic´ et al. / Clinica Chimica Acta 296 (2000) 101 – 109
106
Table 3 Genotype distribution in healthy individuals and cancer patients Groups
Healthy (n 5 144)
Allele:
Breast cancer (n 5 76)
2D6*3
Mut / wt N 4 / 144 (%) 2.8 Mut / mut N — (%) — Statistical significance (P)a a
Head and neck cancer (n 5 56)
2D6*4
2D6*3
2D6*4
2D6*3
2D6*4
27 / 144 18.8 3 / 144 2.1
1 / 76 1.3 — —
20 / 76 26.3 4 / 76 5.3
2 / 56 3.6 — —
14 / 56 25 1 1.8
0.338
0.186
2
P was calculated using x -test.
showed 3.6% (2 / 56) of the patients to be heterozygous for 2D6*3 mutation and 25% (14 / 56) for 2D6*4 mutation. Among them only one (1.8%) was found to be homozygous and all other were heterozygous for 2D6*4 mutation. The predicted phenotype distribution among healthy individuals and cancer patients is presented in Table 4. The phenotype was predicted according to the genotypes as follows: extensive metabolizers were individuals carrying two CYP2D6 wild-type alleles; intermediate extensive metabolizers (IEM) carried one deficiency and one wild-type allele; and poor metabolizers (PM) were homozygous for two deficiency alleles. In the healthy group, 2.1% of individuals were identified as poor metabolizers, 76.4% as extensive metabolizers, and 21.5% as intermediate heterozygotes. In the group of BC, 5.3 and 27.6% of individuals were classified as PM and IEM, respectively, while 67.1% were identified as EM. In the group of HNC, the incidence of PM individuals was 1.8, but as many as 28.6% were identified as IEM phenotypes.
Table 4 Predicted phenotype distribution in healthy individuals and cancer patients Groups
Control BC HNC a
No. of subjects 144 76 56
Genotype a
a
a
EM (%)
IEM (%)
PM (%)
76.4 67.1 69.6
21.5 27.6 28.6
2.1 5.3 1.8
Statistical significance (P)b – 0.229 0.573
EM, extensive metabolizer carrying two wild-type CYP2D6 alleles; IEM, intermediate heterozygous metabolizer carrying one deficiency allele and one wild-type allele; and PM, poor metabolizer carrying two deficiency alleles. b P was calculated using x 2 -test.
E. Topic´ et al. / Clinica Chimica Acta 296 (2000) 101 – 109
107
4. Discussion The preliminary results indicated a difference in the polymorphism and zygosity of CYP 2D6*3 and 2D6*4 alleles between healthy volunteers and cancer patients. Although not statistically significant, there were lower wild-type frequencies in the cancer groups compared to the controls. The frequency of PMs in the control group (2.1%) was slightly below the range reported for other Caucasian populations (2.5–5.3%) [17–19]. Comparison of genotype distribution between the control and breast cancer groups yielded no statistically significant difference (P 5 0.164). Some authors have presented similar results in larger cohorts. In their cohort of 437 breast cancer patients, Smith et al. [19] found 66.8% of EM, 29.3% of IEM and 3.9% of PM genotype. In a cohort of 720 subjects in the UK, Wolf et al. [5] detected 66.8% of EM, 29.4% of IEM and 3.8% of PM genotypes. The proportion of genotyped PMs in these cohorts was slightly lower than in our breast cancer group. In our cohort of breast cancer patients, there was a relatively large increase in the proportion of PMs (5.3%), although at a significance level of P 5 0.349 this increase may present a risk associated with the PM genotype, as the apparent mutant allele frequency is substantially increased. Anthony et al. have postulated that the PM phenotype may be a risk factor for postmenopausal breast cancer [20]. Many more patients should be studied to provide a clearer picture of the extent of correlation between breast cancer and PM genotype. Considering the distribution of CYP2D6 genotypes in the head and neck cancer patients, no statistically significant shift was observed in the proportion of EMs and IEMs compared with the control population (P 5 0.146). In a group of 75 patients affected by head and neck cancer, Gonzalez et al. [21] found 4% of PM and 14% of IEM genotypes, suggesting the absence of association of the CYP2D6, CYP2E1 and GSTM1 genotypes with the risk of head and neck cancer. Head and neck cancer is a multifactorial disease, and predisposition to the development of such diseases is likely to be the result of several genetic polymorphisms. However, the fact that as many as 28.6% of head and neck cancer patients had heterozygous mutation on 2D6*4 allele should not be neglected. In conclusion, the proposed PCR-SSCP protocol using the genotyping approach can be recommended as a reliable and fast method for detecting the polymorphism and zygosity of the most frequent mutations of CYP2D6 gene, i.e., CYP2D6*3 and CYP2D6*4 alleles. The preliminary results indicate a difference in the polymorphism and zygosity of CYP2D6*3 and CYP2D6*4 alleles between controls and cancer patients. In the overall cohort of 276 subjects studied for the two common mutations of CYP2D6 gene, the data presented in this report showed an increase in the proportion of PMs in breast cancer and an increase of IEM in head and neck cancer patients compared to
108
E. Topic´ et al. / Clinica Chimica Acta 296 (2000) 101 – 109
controls. On the basis of our results, we conclude that the polymorphic CYP2D6 gene does play a role, probably indirectly, in carcinogenesis. However, additional studies in greater cohorts should be done to provide a clearer picture of the extent and correlation of mutant genotypes in breast and head and neck cancer patients versus healthy controls.
Acknowledgements The authors wish to thank Miss Michelle Bon (Enschede, NL) for her valuable suggestions during the work, and Mrs. A. Redovnikovic for assistance in preparing the manuscript. The Ministry of Science and Technology of the Republic of Croatia financially supported this work.
References [1] Hong JY, Yang CS. Genetic polymorphism of cytochrome P450 as a biomarker of susceptibility to environmental toxicity. Environ Health Perspect 1997;105(Suppl 4):759–62. [2] Kawajiri K, Nakachi K, Imai K, Yoshii A, Shinoda N, Watanabe J. Identification of genetically high risk individuals to lung cancer by DNA polymorphism of the cytochrome P4501A1 gene. FEBS Lett 1990;263:131–3. [3] Taioli E, Trachman J, Chen X, Toniolo P, Garte SJ. A CYP1A1 restriction fragment length polymorphism is associated with breast cancer in African-American women. Cancer Res 1995;55:3757–68. [4] Tefre T, Ryberg D, Haugen A et al. Human CYP1A1 (cytochrome P450 P1450) gene: lack of association between the MspI restriction fragment length polymorphism and the incidence of lung cancer in a Norwegian population. Pharmacogenetics 1991;1:20–5. [5] Wolf CR, Smith CAD, Gough AC et al. Relationship between the debrisoquine hydroxylase polymorphism and cancer susceptibility. Carcinogenesis 1993;13:1035–8. [6] Hidesheim A, Chen C-J, Caporaso NE et al. Cytochrome P4502E1 genetic polymorphism and risk of nasopharyngeal carcinoma: results from a case-control study conducted in Taiwan. Cancer Epidemiol Biomarkers Prev 1995;4:607–10. [7] Hirvonen A, Husgafvel-Pursiainen K, Anttila S, Karjalainen A, Vainio H. The human CYP2E1 gene and lung cancer: DraI and RsaI restriction fragment length polymorphism in a Finnish study population. Carcinogenesis 1993;14:85–8. [8] Kato S, Shields PG, Caporaso NE et al. Analysis of cytochrome P450 2E1 genetic polymorphism in relation to human lung cancer. Cancer Epidemiol Biomarkers Prev 1995;4:515–8. [9] Yu M-W, Gladek-Yarborough A, Chiamprasert S, Santella RM, Liaw Y-F, Chen C-J. Cytochrome P450 2E1 and glutathione S-transferase M1 polymorphism and susceptibility to hepatocellular carcinoma. Gastroenterology 1995;109:1266–73. ¨ [10] Stuven T, Griese EU, Kroemer HK, Eichelbaum M, Zanger UM. Rapid detection of CYP2D6 null alleles by long distance- and multiplex-polymerase chain reaction. Pharmacogenetics 1996;6:417–21.
E. Topic´ et al. / Clinica Chimica Acta 296 (2000) 101 – 109
109
[11] Kimura S, Umeno M, Skoda RC, Meyer UA, Gonzales F. The human debrisoquine 4-hydroxylase (CYP2D) locus: sequence and identification of the polymorphic CYP2D6 gene, a related gene, and a pseudogene. Am J Hum Genet 1989;45:889–904. [12] Topic E, Gluhak J. Isolation of restrictible DNA. Eur J Clin Chem Clin Biochem 1991;29:327–30. [13] Wolf CR, Moss JE, Miles JS, Gough AC, Spurr NK. Deletion of debrisoquine hydroxylation phenotypes. Lancet 1990;335:1452–3. [14] Vermes I, Bon M, van den Bergh FATMJ et al. Pharmacogenetics: a tool for therapeutic drug monitoring pretherapeutic genotyping of CYP2D6. In: 12th IFCC European Congress of Clinical Chemistry, MEDLAB 97, Switzerland, 1997, p. 104, [abstract]. ˇ ˇ ´ M, Nikolic´ V, Zoricic ˇ ´ I, Ivanisevic ˇ ´ A-M, Zuntar [15] Topic´ E, Stefanovic I. Detection of CYP2D6*3 and 2D6*4 allelic variants by PCR-restriction fragment length polymorphism. Clin Chem Lab Med 1998;36:655–8. [16] Broly F, Marez D, Sabbagh N et al. An efficient strategy for detection of known and new mutations of the CYP2D6 gene using single strand conformation polymorphism analysis. Pharmacogenetics 1995;5:373–84. [17] Gough AC, Miles JS, Spurr NK et al. Identification of primary gene defect at the cytochrome P450 CYP2D locus. Nature 1990;347:773–6. [18] Agundez JAG, Martinez C, Ladero JM et al. Debrisoquine oxidation genotype and susceptibility to lung cancer. Clin Pharmacol Ther 1994;55:10–4. [19] Smith CA, Moss JE, Gough AC, Spurr NK, Wolf CR. Molecular genetic analysis of the cytochrome P450-debrisoquine hydroxylase locus and association with cancer susceptibility. Environ Health Perspect 1992;98:107–12. [20] Anthony LB, Thomas JB, Hande KR. Cytochrome P-450IID6 phenotyping in cancer patients: debrisoquine and dextromethorphan as probes. Cancer Chemother Pharmacol 1995;36:125– 8. [21] Gonzales MV, Alvarez V, Pello MF, Menendez MJ, Suarez C, Coto E. Genetic polymorphism of N-acetyltransferase-2, glutathione S-transferase-M1, and cytochromes P450IIE1 and P450IID6 in susceptibility to head and neck cancer. J Clin Pathol 1998;51:294–8.
The cytochrome P450 2D6 (CYP2D6) gene polymorphism among breast and head and neck cancer patients ˇ ´ a , Ana Maria Ivanisevic ˇ ´ a, Elizabeta Topic´ a , *, Mario Stefanovic b c ˇ ˇ´ Rajka Petrinovic´ , Ivica Curcic a
Clinical Institute of Chemistry, School of Medicine, University of Zagreb and Sestre Milosrdnice University Hospital, Vinogradska c. 29, 10000 Zagreb, Croatia b Cancer Institute of Croatia, Zagreb, Croatia c ENT Department, Dubrava University Hospital, Zagreb, Croatia Received 17 May 1999; received in revised form 10 January 2000; accepted 24 January 2000
Abstract The prevalence of CYP2D6*3 and CYP2D6*4 alleles in normal controls and cancer patients was studied using the reliable PCR-SSCP method. In the control group (n 5 144), four subjects (2.8%) were found to carry CYP2D6*3 allele (heterozygote), while 30 (20.8%) subjects carried CYP2D6*4 allele (18.8% heterozygotes, 2.1% homozygotes). One (1.3%) of the breast cancer (BC) patients (n 5 76) carried CYP2D6*3 allele, but 24 (31.6%) carried CYP2D6*4 allele (26.3% heterozygotes, 5.3% homozygotes). In the head and neck cancer (HNC) group (n 5 56), two (3.6%) patients were heterozygous for CYP2D6*3 mutation and 15 (26.8%) for CYP2D6*4 mutation. Fourteen of 56 (25%) and one of 56 (1.8%) of these patients carried heterozygous and homozygous mutations, respectively. In controls, 2.1% were identified as poor metabolizers (PM), 76.4% as extensive metabolizers (EM), and 21.5% as intermediate heterozygotes (IEM). In BC group, 5.3, 27.6 and 67.1% were classified as PM, IEM and EM, respectively. In HNC group, the incidence of PM was 1.8, but as many as 28.6% were identified as IEM phenotypes. 2000 Elsevier Science B.V. All rights reserved. Keywords: CYP2D6 polymorphism; PCR-SSCP; Breast cancer; Head and neck cancer
*Corresponding author. Tel. / fax: 1 385-1-3768-280. ´ E-mail address: [email protected] (E. Topic) 0009-8981 / 00 / $ – see front matter 2000 Elsevier Science B.V. All rights reserved. PII: S0009-8981( 00 )00221-7
102
E. Topic´ et al. / Clinica Chimica Acta 296 (2000) 101 – 109
1. Introduction Most environmental toxic chemicals and carcinogens need to be metabolically activated to exert their toxic or carcinogenic effects. The most important group of activating enzymes is the large cytochrome P450 (CYP) ‘superfamily’. Of 15 human CYP isoenzymes characterized so far, eight forms of CYP (CYP1A1, 1A2, 2A6 2C9, 2D6, 2E1 and 3A4) have been shown to be polymorphic at the phenotype or genotype level, or both. Because an individual’s capability to metabolize these toxicants can be altered by carrying variant alleles, the genetic polymorphism of CYP enzymes has been proposed as a biomarker of susceptibility to environmental carcinogenesis and toxicity [1]. The association of CYP genetic polymorphism and human cancer risk has received increasing attention. Examples include CYP1A1 with lung and breast cancers [2–4], CYP2D6 with different types of cancer [5], and CYP2E1 with lung, liver and nasopharyngeal cancers [2,6–9]. However, many reports are controversial. One important factor in interpreting these results is that there are significant ethnic differences in frequency distribution of the CYP genetic polymorphism. The aim of this study was to assess the prevalence of CYP2D6*3 and *4 gene mutations in a group of healthy subjects and in patients with cancer etiologies by a reliable method for their identification. The CYP2D6 isoform metabolizing more than 25% of most common drugs has more than 20 mutant alleles identified to date. The mutant CYP2D6*3 allele with A 2637 deletion in exon 5 and mutant CYP2D6*4 allele with G 1934 → A splice site defect are among the most common ones, with a frequency of 2.7 and 28.6%, respectively. These mutations result in a decreased or no activity of CYP2D6 isoenzyme, leading to the poor metabolizer phenotype (PM). These individuals are at an increased risk of adverse side effects or therapeutic failure following drug treatment [10]. The CYP2D6 locus in humans has been identified as an array of at least three highly homologous genes, on the long arm of chromosome 22. In addition to the active CYP2D6 gene consisting of nine exons and eight introns, an inactive pseudogene, CYP2D8P, and an inactive homologue, CYP2D7, are present in a tandem array [11].
2. Material and methods
2.1. Subjects A total of 276 subjects, 132 cancer patients and 144 healthy volunteers, were included in the study. The cancer group consisted of 76 breast cancer (BC)
E. Topic´ et al. / Clinica Chimica Acta 296 (2000) 101 – 109
103
patients and 56 head and neck cancer (HNC) patients. Genotyping of the CYP2D6*3 and *4 alleles was performed on DNA isolated from EDTA-blood by the PCR-SSCP method. Whole blood samples (3 ml) were collected in sodium EDTA vacutainers, assigned a unique code, and stored at 2 208C. Leukocyte DNA was isolated by standard phenol / chloroform extraction and ethanol precipitation protocols using a slightly modified method by Topic and Gluhak [12].
2.2. PCR-SSCP procedure Target DNA (170 ng genomic DNA in 50 ml of PCR mixture) was amplified by polymerase chain reaction (PCR) [11] using PCR Core Kit (Roche Diagnostics, Mannheim, Germany) and specific amplification primers (MWG-Biotech, Ebersberg, Germany) in the Progene Thermal Cycler (Techne, Cambridge, England). The 2D6*3 mutation was detected by amplification of the region of interest with 2D6*3 allele primers (base pairs 4147–4416): forward, 59-gatgagctgctaactgagccc-39; reverse, 59-ccgagagcatactcgggac-39. The 2D6*4 mutation was detected with 2D6*4 allele primers (base pairs 3359–3713): forward, 59-aaatcctgctcttccgaggc-39; reverse, 59-gccttcgccaaccactcc-39; as described by Wolf et al. [13] and Vermes et al. [14]. Table 1 shows optimized conditions for CYP2D6*3 and CYP2D6*4 PCR according to chemicals and cycling [15]. To detect mutations in the CYP2D6 gene we used a strategy based on single-strand conformation polymorphism (SSCP) [16]. After the amplification, 3 ml of the PRC product were added to 7 ml of loading solution (formamide containing 10 mmol / l NaOH and 0.5% bromphenol blue), denatured at 958C for 5 min, quickly chilled on ice and applied (8 ml / lane) to GMAE, Mini Gels for SSCP (Elchrom Scientific, Switzerland). Electrophoresis was performed at a constant temperature of 88C in 30 mmol / l TAE, pH 7.5. The gels were stained with Table 1 Optimized conditions for CYP2D6*3 and CYP2D6*4 PCR Allele 2D6*3
Allele 2D6*4
Chemicals: Magnesium concentration (mmol / l) Primer concentration (mmol / l) DNTP concentration (mmol / l) Taq DNA polymerase concentration (U / ml)
0.80 0.15 0.20 0.025
1.20 0.20 0.20 0.025
Cycling: Cycle denaturation (temperature and time) Primer annealing (temperature and time) Primer extension (temperature and time) Cycle number
928C / 45 s 56.58C / 30 s 728C / 600 s 38
928C / 45 s 578C / 30 s 728C / 600 s 38
104
E. Topic´ et al. / Clinica Chimica Acta 296 (2000) 101 – 109
SYBR Green II (1:10 000 in 10 mmol / l TAE) for 1 h and photographed at 254 nm with a Polaroid 667 film. Negative quality controls generated by inclusion of all reagents except DNA and positive quality control containing a known genotype (obtained by the courtesy of Michelle Bon, Enschede, NL) were run with every run. Subjects homozygous for the wild-type allele were classified as extensive metabolizers, subjects carrying one wild-type and one mutant allele as intermediate extensive metabolizers, and subjects homozygous for mutant alleles as poor metabolizers [17].
3. Results The results of PCR-SSCP presented in Fig. 1 show reproducible and uniform band patterns for mutant and wild-type variants of CYP2D6*3 and *4 alleles. Five and six bands for mutant and wild-type variants of 2D6*3 and 2D6*4, respectively, were observed. One of these bands appeared to be distinctive for *3 allele of the heterozygous phenotype (no homozygous phenotype was found) and two for *4 allele, denoting a heterozygous and homozygous genotype. In the
Fig. 1. PCR-SSCP analysis of the CYP2D6 gene. (A) CYP2D6*3: lane 1, homozygote wt / wt (positive control); lanes 2 and 3, patients wt / wt; lane 4, positive control of *3 / wt mutant; lane 5, patient *3 / wt; and lane 6, CYP2D6*3 PCR non-denatured product. (B) CYP2D6*4: lane 1, homozygote wt / wt (positive control); lane 2, patient wt / wt; lane 3, *4 / wt mutant heterozygote (positive control); lane 4, patient *4 / wt; lane 5, *4 / *4 mutant homozytoge (positive control); lane 6, patient *4 / *4; and lane 7, CYP2D6*4 PCR non-denatured product. Arrows indicate bands that discriminate 2D6*3 allele of the heterozygous phenotype (A) and 2D6*4 allele of the heterozygous and homozygous phenotype (B).
E. Topic´ et al. / Clinica Chimica Acta 296 (2000) 101 – 109
105
PCR-SSCP approach, the genotypes previously detected by PCR-RFLP [15] were used as positive quality controls as follows: homozygous wild-type and mutant heterozygous for 2D6*3 allele, and wild- and mutant homozygote as well as mutant heterozygote for 2D6*4 allele. Comparison of the electrophoresis patterns of PCR-SSCP findings demonstrated that the SSCP analysis identified the previously described mutations of CYP2D6 gene and discriminated between the homozygous and heterozygous individuals. Study results showed that 23.6% (n 5 34) of 144 healthy volunteers carried mutant 2D6*3 and / or 2D6*4 alleles. The frequency of CYP2D6 alleles among healthy subjects and cancer patients is presented in Table 2. A total of 1.4% (4 / 288) of alleles were 2D6*3 alleles with a frame shift mutation in exon 5 and 11.4% (33 / 288) were 2D6*4 alleles with G to A mutation at intron 3–exon 4 junction. The genotypic frequencies of all studied groups were consistent with Hardy–Weinberg equilibrium. In BC (n 5 76) patients, 32.9% (n 5 25) of those genotyped for 2D6*3 and 2D6*4 alleles were found to carry mutant alleles. One of them (0.7%) was found to bear 2D6*3 allele, but 18.4% (28 / 152) had alleles with G to A mutation at intron 3–exon 4 junction. In HNC patients (n 5 56), as many as 26.8.2% (n 5 15) were found to carry mutant alleles. Among them, 1.8% (2 / 112) had 2D6*3 allele and 14.3% (16 / 112) 2D6 4* allele. Results of genotype distribution (Table 3), showed four of 144 (2.8%) and 27 / 144 (18.8%) control group subjects to be heterozygotes according to the frame shift mutation in exon 5 and G to A mutation at intron 3–exon 4 junction, respectively. In this group, there were no homozygous 2D6*3 mutants, however 2.1% (n 5 3) of homozygous 2D6*4 mutants were detected. In the BC group (n 5 76), 1.3% (n 5 1) heterozygous and no homozygous mutant 2D6*3 genotype was detected. Out of 24 patients bearing 2D6*4 allele, 26.3% (n 5 20) were heterozygotes and 5.3% (n 5 4) homozygotes. In the group of HNC patients (n 5 56), results of genotype distribution Table 2 Frequency of CYP2D6 alleles in healthy volunteers and patient groups
Healthy Breast cancer Head and neck cancer a
No. of alleles
Allele *3 a no. (%)
288 152 112
4 (1.4) 1 (0.7) 2 (1.8)
P value c
0.948 0.934
Allele *4 b no. (%) 33 (11.4) 28 (18.4 16 (14.3)
Allele *3, CYP2D6*3 allele with frame shift mutation in exon 5. Allele *4, CYP2D6*4 allele with G to A mutation at intron 3–exon 4 junction. c Statistical significance (P) was calculated using z-test. b
P value
0.098 0.448
E. Topic´ et al. / Clinica Chimica Acta 296 (2000) 101 – 109
106
Table 3 Genotype distribution in healthy individuals and cancer patients Groups
Healthy (n 5 144)
Allele:
Breast cancer (n 5 76)
2D6*3
Mut / wt N 4 / 144 (%) 2.8 Mut / mut N — (%) — Statistical significance (P)a a
Head and neck cancer (n 5 56)
2D6*4
2D6*3
2D6*4
2D6*3
2D6*4
27 / 144 18.8 3 / 144 2.1
1 / 76 1.3 — —
20 / 76 26.3 4 / 76 5.3
2 / 56 3.6 — —
14 / 56 25 1 1.8
0.338
0.186
2
P was calculated using x -test.
showed 3.6% (2 / 56) of the patients to be heterozygous for 2D6*3 mutation and 25% (14 / 56) for 2D6*4 mutation. Among them only one (1.8%) was found to be homozygous and all other were heterozygous for 2D6*4 mutation. The predicted phenotype distribution among healthy individuals and cancer patients is presented in Table 4. The phenotype was predicted according to the genotypes as follows: extensive metabolizers were individuals carrying two CYP2D6 wild-type alleles; intermediate extensive metabolizers (IEM) carried one deficiency and one wild-type allele; and poor metabolizers (PM) were homozygous for two deficiency alleles. In the healthy group, 2.1% of individuals were identified as poor metabolizers, 76.4% as extensive metabolizers, and 21.5% as intermediate heterozygotes. In the group of BC, 5.3 and 27.6% of individuals were classified as PM and IEM, respectively, while 67.1% were identified as EM. In the group of HNC, the incidence of PM individuals was 1.8, but as many as 28.6% were identified as IEM phenotypes.
Table 4 Predicted phenotype distribution in healthy individuals and cancer patients Groups
Control BC HNC a
No. of subjects 144 76 56
Genotype a
a
a
EM (%)
IEM (%)
PM (%)
76.4 67.1 69.6
21.5 27.6 28.6
2.1 5.3 1.8
Statistical significance (P)b – 0.229 0.573
EM, extensive metabolizer carrying two wild-type CYP2D6 alleles; IEM, intermediate heterozygous metabolizer carrying one deficiency allele and one wild-type allele; and PM, poor metabolizer carrying two deficiency alleles. b P was calculated using x 2 -test.
E. Topic´ et al. / Clinica Chimica Acta 296 (2000) 101 – 109
107
4. Discussion The preliminary results indicated a difference in the polymorphism and zygosity of CYP 2D6*3 and 2D6*4 alleles between healthy volunteers and cancer patients. Although not statistically significant, there were lower wild-type frequencies in the cancer groups compared to the controls. The frequency of PMs in the control group (2.1%) was slightly below the range reported for other Caucasian populations (2.5–5.3%) [17–19]. Comparison of genotype distribution between the control and breast cancer groups yielded no statistically significant difference (P 5 0.164). Some authors have presented similar results in larger cohorts. In their cohort of 437 breast cancer patients, Smith et al. [19] found 66.8% of EM, 29.3% of IEM and 3.9% of PM genotype. In a cohort of 720 subjects in the UK, Wolf et al. [5] detected 66.8% of EM, 29.4% of IEM and 3.8% of PM genotypes. The proportion of genotyped PMs in these cohorts was slightly lower than in our breast cancer group. In our cohort of breast cancer patients, there was a relatively large increase in the proportion of PMs (5.3%), although at a significance level of P 5 0.349 this increase may present a risk associated with the PM genotype, as the apparent mutant allele frequency is substantially increased. Anthony et al. have postulated that the PM phenotype may be a risk factor for postmenopausal breast cancer [20]. Many more patients should be studied to provide a clearer picture of the extent of correlation between breast cancer and PM genotype. Considering the distribution of CYP2D6 genotypes in the head and neck cancer patients, no statistically significant shift was observed in the proportion of EMs and IEMs compared with the control population (P 5 0.146). In a group of 75 patients affected by head and neck cancer, Gonzalez et al. [21] found 4% of PM and 14% of IEM genotypes, suggesting the absence of association of the CYP2D6, CYP2E1 and GSTM1 genotypes with the risk of head and neck cancer. Head and neck cancer is a multifactorial disease, and predisposition to the development of such diseases is likely to be the result of several genetic polymorphisms. However, the fact that as many as 28.6% of head and neck cancer patients had heterozygous mutation on 2D6*4 allele should not be neglected. In conclusion, the proposed PCR-SSCP protocol using the genotyping approach can be recommended as a reliable and fast method for detecting the polymorphism and zygosity of the most frequent mutations of CYP2D6 gene, i.e., CYP2D6*3 and CYP2D6*4 alleles. The preliminary results indicate a difference in the polymorphism and zygosity of CYP2D6*3 and CYP2D6*4 alleles between controls and cancer patients. In the overall cohort of 276 subjects studied for the two common mutations of CYP2D6 gene, the data presented in this report showed an increase in the proportion of PMs in breast cancer and an increase of IEM in head and neck cancer patients compared to
108
E. Topic´ et al. / Clinica Chimica Acta 296 (2000) 101 – 109
controls. On the basis of our results, we conclude that the polymorphic CYP2D6 gene does play a role, probably indirectly, in carcinogenesis. However, additional studies in greater cohorts should be done to provide a clearer picture of the extent and correlation of mutant genotypes in breast and head and neck cancer patients versus healthy controls.
Acknowledgements The authors wish to thank Miss Michelle Bon (Enschede, NL) for her valuable suggestions during the work, and Mrs. A. Redovnikovic for assistance in preparing the manuscript. The Ministry of Science and Technology of the Republic of Croatia financially supported this work.
References [1] Hong JY, Yang CS. Genetic polymorphism of cytochrome P450 as a biomarker of susceptibility to environmental toxicity. Environ Health Perspect 1997;105(Suppl 4):759–62. [2] Kawajiri K, Nakachi K, Imai K, Yoshii A, Shinoda N, Watanabe J. Identification of genetically high risk individuals to lung cancer by DNA polymorphism of the cytochrome P4501A1 gene. FEBS Lett 1990;263:131–3. [3] Taioli E, Trachman J, Chen X, Toniolo P, Garte SJ. A CYP1A1 restriction fragment length polymorphism is associated with breast cancer in African-American women. Cancer Res 1995;55:3757–68. [4] Tefre T, Ryberg D, Haugen A et al. Human CYP1A1 (cytochrome P450 P1450) gene: lack of association between the MspI restriction fragment length polymorphism and the incidence of lung cancer in a Norwegian population. Pharmacogenetics 1991;1:20–5. [5] Wolf CR, Smith CAD, Gough AC et al. Relationship between the debrisoquine hydroxylase polymorphism and cancer susceptibility. Carcinogenesis 1993;13:1035–8. [6] Hidesheim A, Chen C-J, Caporaso NE et al. Cytochrome P4502E1 genetic polymorphism and risk of nasopharyngeal carcinoma: results from a case-control study conducted in Taiwan. Cancer Epidemiol Biomarkers Prev 1995;4:607–10. [7] Hirvonen A, Husgafvel-Pursiainen K, Anttila S, Karjalainen A, Vainio H. The human CYP2E1 gene and lung cancer: DraI and RsaI restriction fragment length polymorphism in a Finnish study population. Carcinogenesis 1993;14:85–8. [8] Kato S, Shields PG, Caporaso NE et al. Analysis of cytochrome P450 2E1 genetic polymorphism in relation to human lung cancer. Cancer Epidemiol Biomarkers Prev 1995;4:515–8. [9] Yu M-W, Gladek-Yarborough A, Chiamprasert S, Santella RM, Liaw Y-F, Chen C-J. Cytochrome P450 2E1 and glutathione S-transferase M1 polymorphism and susceptibility to hepatocellular carcinoma. Gastroenterology 1995;109:1266–73. ¨ [10] Stuven T, Griese EU, Kroemer HK, Eichelbaum M, Zanger UM. Rapid detection of CYP2D6 null alleles by long distance- and multiplex-polymerase chain reaction. Pharmacogenetics 1996;6:417–21.
E. Topic´ et al. / Clinica Chimica Acta 296 (2000) 101 – 109
109
[11] Kimura S, Umeno M, Skoda RC, Meyer UA, Gonzales F. The human debrisoquine 4-hydroxylase (CYP2D) locus: sequence and identification of the polymorphic CYP2D6 gene, a related gene, and a pseudogene. Am J Hum Genet 1989;45:889–904. [12] Topic E, Gluhak J. Isolation of restrictible DNA. Eur J Clin Chem Clin Biochem 1991;29:327–30. [13] Wolf CR, Moss JE, Miles JS, Gough AC, Spurr NK. Deletion of debrisoquine hydroxylation phenotypes. Lancet 1990;335:1452–3. [14] Vermes I, Bon M, van den Bergh FATMJ et al. Pharmacogenetics: a tool for therapeutic drug monitoring pretherapeutic genotyping of CYP2D6. In: 12th IFCC European Congress of Clinical Chemistry, MEDLAB 97, Switzerland, 1997, p. 104, [abstract]. ˇ ˇ ´ M, Nikolic´ V, Zoricic ˇ ´ I, Ivanisevic ˇ ´ A-M, Zuntar [15] Topic´ E, Stefanovic I. Detection of CYP2D6*3 and 2D6*4 allelic variants by PCR-restriction fragment length polymorphism. Clin Chem Lab Med 1998;36:655–8. [16] Broly F, Marez D, Sabbagh N et al. An efficient strategy for detection of known and new mutations of the CYP2D6 gene using single strand conformation polymorphism analysis. Pharmacogenetics 1995;5:373–84. [17] Gough AC, Miles JS, Spurr NK et al. Identification of primary gene defect at the cytochrome P450 CYP2D locus. Nature 1990;347:773–6. [18] Agundez JAG, Martinez C, Ladero JM et al. Debrisoquine oxidation genotype and susceptibility to lung cancer. Clin Pharmacol Ther 1994;55:10–4. [19] Smith CA, Moss JE, Gough AC, Spurr NK, Wolf CR. Molecular genetic analysis of the cytochrome P450-debrisoquine hydroxylase locus and association with cancer susceptibility. Environ Health Perspect 1992;98:107–12. [20] Anthony LB, Thomas JB, Hande KR. Cytochrome P-450IID6 phenotyping in cancer patients: debrisoquine and dextromethorphan as probes. Cancer Chemother Pharmacol 1995;36:125– 8. [21] Gonzales MV, Alvarez V, Pello MF, Menendez MJ, Suarez C, Coto E. Genetic polymorphism of N-acetyltransferase-2, glutathione S-transferase-M1, and cytochromes P450IIE1 and P450IID6 in susceptibility to head and neck cancer. J Clin Pathol 1998;51:294–8.