CYP1B1 Gene Polymorphisms Have Higher Risk for Endometrial Cancer, and Positive Correlations with Estrogen Receptor α and Estrogen Receptor β Expressions

Endometrial carcinoma is a common gynecological malignancy of the female urogenital tract, and its incidence is increasing significantly (1). However, the genetic basis of this disease is not well understood. The metabolic conversion of estrogens to 4-hydroxy estrogens has been postulated to be a major factor in carcinogenesis (2, 3). CYP1B1 converts estrogens to 4-hydroxy estrogens that induce DNA damage (4, 5). 4-Hydroxy estrogens are shown to induce endometrial adenocarcinoma in mice (4). CYP1B1 displays the highest level of expression in endometrial tissue (3). These reports strongly suggest that the CYP1B1 gene is a key player in endometrial carcinogenesis (3, 4, 5). Several studies have shown that polymorphisms in several genes such as CYP1B1 can induce hyperactivation of proteins and increase the incidence of several cancers (6, 7, 8). Several polymorphisms of the CYP1B1 gene have been described of which four result in amino acid substitutions: intron 1–13C→T, codon 48 C→G, codon 119 G→T, codon 432 C→G, codon 449 T→C, and codon 453 A→G (9, 10, 11). Intron 1 contains a polymorphic site at nucleotide −13, exon 2 contains two polymorphic sites at codons 48 and 119, and exon 3 contains three polymorphic sites at codons 432, 449, and 453. Of the six polymorphisms in the CYP1B1 gene, amino acid replacements occur at codons 48, 119, 432, and 453 leading to the replacement of Arg → Gly, Ala → Ser, Leu → Val, and Asn → Ser, respectively. The polymorphisms on exons 2 and 3 have significant effects on the catalytic function of CYP1B1 (8, 9). The 4-hydroxylation activities of the variant enzymes, especially polymorphisms on codons 119 and 432, were 2–4-fold higher than the wild-type enzyme (8, 9). Because the CYP1B1 polymorphisms are inherited, they will dictate exposure levels of these estrogen metabolites during the lifetime of an individual (8, 9). Thus, inherited alterations in the activity of CYP1B1 leads to differences in estrogen metabolism and thereby, may possibly explain interindividual differences in endometrial cancer risk associated with estrogen-mediated carcinogenesis (8, 9). A recent study has shown significant correlations between 432G/G and ERα² and PR positivity in breast cancer (9). We hypothesize that polymorphisms of the CYP1B1 gene and activation of steroid receptors are important in pathogenesis of endometrial cancer. The present study was designed to investigate six polymorphisms on the CYP1B1 gene in endometrial cancer. We additionally investigated whether expression of steroid receptors, ERα, ERβ, PR, and AR are influenced by CYP1B1 polymorphisms.

DNA from 113 sporadic endometrial cancer patients was obtained from the Department of Gynecology at the hospital of Hokkaido University (Hokkaido, Japan) between 1992 and 2000 (age range, 29–87 years; mean age, 61.7). The histopathological types of these cancers were as follows: 86 of endometrioid cancer, 6 of adenoacantoma, 2 of adenosquamous cancer, 2 of clear cell cancer, and 17 of unknown type. Cancer-free control samples were obtained from 202 unrelated Japanese healthy volunteers in the same prefecture (100 women and 102 men, age range, 35–69 years; mean age, 59.3) at the same time period as endometrial cancer patient samples. Because no patients and healthy controls of other ethnic groups were recruited, this study was limited to a Japanese population. There were no differences between patients and control groups in regard to age, race, family history of cancer, indices of body size (height, weight, and body mass index), or income. Appropriate informed consent was obtained from the patients in accordance with local ethical committee guidelines of the hospital.

For the analysis of CYP1B1 polymorphisms, a two-step PCR procedure was designed. A region covering the polymorphic sites in each exon was amplified in the first PCR followed by single nucleotide polymorphism-specific second PCR. In the first PCR, DNA was amplified using 10 ng DNA, 1.5 mm MgCl₂, 0.8 mm deoxynucleotide triphosphate, 0.5 units of Taq polymerase (Applied Biosystems Inc., Foster City, CA), and specific first PCR primer sets (Table 1). The first PCR product was subsequently used as a template in the second PCR. Each polymorphic fragment from first PCR was amplified with 1.5 mm MgCl₂, 0.8 mm deoxynucleotide triphosphate, 0.5 units of Taq polymerase, and specific second PCR primer sets. Each second PCR product (8 μl) was mixed with 1 μl of 10× loading dye and then run on 3% agarose gels. The bands on the gels were visualized by ethidium bromide staining. DNA samples with homozygous wild-type and mutated-type genotypes were used as controls to ensure nonspecific amplifications or PCR failures. Blank DNA controls were used to rule out PCR contamination, and ambiguous results were repeated until a clear result was obtained. For confirmation of genotyping, the PCR products were subjected to direct sequence. The PCR products were first purified by QIAquick PCR Purification kit (Qiagen, Valencia, CA). Then, a double-strand sequence analysis of the PCR products was performed using first PCR primer and ABI 377 Sequencer and Dye Terminator Cycle sequencing kit (Applied Biosystems Inc.; Ref. 12).

Ninety-three endometrial cancer tissues from 113 endometrial cancer patients were also used for the immunohistochemical analysis by blinded procedure. Anti-ERα, -ERβ, -PR, and -AR antibodies (F10, N19, C-19, and 441, respectively; Santa Cruz Biotech, Santa Cruz, CA) were used to identify the expression of these receptors in cancerous endometrium by standard immunohistochemical techniques (12). Paraffin-embedded, endometrial cancer blocks were cut into 4-μ sections and dried at room temperature. After deparaffinization and rehydration, sections were treated with 2% hydrogen peroxidase in methanol for 20 min to inactivate endogenous peroxidase. Antigen retrieval was done by autoclaving (18 PSI) for 5 min in an antigen unmasking solution (Vector, Burlingame, CA). After blocking with 3% normal goat serum for 15 min, sections were incubated with each primary antibody at a 1:100 dilution with 5% goat serum in PBS overnight at 4°C under a humid chamber. Sections were washed with a solution containing 20 mm Tris, 150 mm NaCl, and 0.025% Tween (pH 7.8), and then incubated with secondary antibodies for 20 min. Immunostaining was done by using an avidin-biotin-peroxidase method (Lab Vision, Fremont, CA) with diaminobenzidine as the chromagen, followed by counterstaining with hematoxylin.

χ² analysis was used to test deviation of genotype distributions from Hardy-Weinberg equilibrium (1, 12). The relative risk of endometrial cancer associated with each genotype was analyzed compared with female control group by calculating odds ratios with 95% CIs. Subsequent analysis included logistic regression analyses, adjusting for the potential confounding factors (age, height, weight, and body mass index). Homogeneity was tested as described previously (1, 12). χ² and t tests, all two-sided, were used to analyze data for statistical significance (1, 12).

The structure of CYP1B1 gene and locations of six different polymorphic loci are shown in Fig. 1,A. Fig. 1,B shows representative samples of genotyping for intron 1 and codons 48, 119, 432, 449, and 453 of CYP1B1 gene in endometrial cancer patients. The frequency of distribution of the six genetic polymorphisms of CYP1B1 in endometrial cancer patients and the control group are shown in Table 2. No difference was observed for the distributions of any polymorphic loci between men and women in normal control (Table 2 A). The polymorphism on codon 453 was not detected in either the control group or the patient group, i.e., all of the individuals tested in this study were homozygous A for codon 453.

The distributions of genotypes on codons 119 and 432 were significantly different between endometrial cancer patients and control (P < 0.001; Table 2,B). 14.2% of endometrial cancer patients showed the 119T/T, although 6.0% of female controls showed this genotype. Of endometrial cancer patients, 42.5% showed 119G/T compared with 19.0% of female controls. The relative risks of 119T/T and 119G/T were calculated as 3.32 and 2.45 as compared with wild-type, respectively. On codon 432, 15.0% of endometrial cancer patients showed the 432G/G, whereas 7.0% of female controls showed this genotype. Of cancer patients, 34.5% showed genotype C/G, although 24.0% of female controls showed this genotype. The relative risks of 432G/G and 432C/G were calculated as 2.49 and 1.57 compared with wild-type, respectively. Other loci, intron 1, codon 48, and codon 449, were not different between endometrial cancer patients and normal healthy control. Table 2,C shows combined analysis of 119–432 genotypes for endometrial cancer (Table 2,C). There are several samples with genotype of 119G/G-432C/G, 119T/T-432C/G, 119G/T-432C/C, and 119G/T-432G/G. Thus, codons119 and 432 are not in linkage disequilibrium (Table 2,C). The relative risks for samples that have both 119T and 432G were much higher compared with samples with 119G/G-432C/C (Table 2 C).

Table 2,D shows the correlation of the six polymorphisms of CYP1B1 with clinical stage of endometrial cancers. The 119T/T showed a significant correlation with higher stages (>stage 3) of endometrial cancer (P < 0.05; Table 2 C). Of samples, 25.7% had 119T/T at higher stages (>stage 3), whereas 8.3% of samples were at lower stages (<stage 2). No significant correlations were found in any of the polymorphisms of CYP1B1 with pathological types of cancer, age, family history of cancer, height, weight, and body mass index, or income of the patients (data not shown).

The correlation of each CYP1B1 genotype with steroid receptor expression in endometrial cancer tissues is shown in Table 3. The 119T/T showed a significant correlation with ERα and ERβ positivities. Of samples, 93.3% with 119T/T were ERα-positive, whereas 55.3% of samples with 119G/G showed ERα positivity (P < 0.001; Table 3). Of samples, 40.0% with 119T/T were ERβ-positive compared with 2.6% of samples with 119G/G (P < 0.001). The 432G/G also showed correlations for ERα positivities (P < 0.025; Table 3). Fig. 1,C shows typical results of genotyping at six polymorphic regions of CYP1B1 gene, direct sequencing, and immunostaining data for ERα and ERβ in endometrial cancer patients with 119T/T and 119G/G (Fig. 1 C). The ERα and ERβ proteins were strongly positive in a patient with 119T/T, whereas they were negative in a patient with 119G/G.

The human CYP1B1 gene, located at the 2p21–22 region, consists of three exons (one of which is noncoding) and two introns (3, 11, 12, 13). The expression of CYP1B1 and the formation of 4-hydroxy estrogens have been associated with estrogen-related tumors in multiple tissues and species (4, 5, 6, 7, 14, 15, 16). Elevated 4-hydroxy estrogen production has been associated with carcinogenesis of various cancers including endometrial cancer (1, 3, 5, 9, 14, 17).

In the present study, significant differences in the allelic distributions of the codon 119 and codon 432 of CYP1B1 were found between endometrial cancer patients and healthy controls. The relative risks of 119T/T and 432G/G were significantly higher when compared with the wild-type. Moreover, we found a significant correlation between the 119T/T and higher stages of endometrial cancer. These data clearly demonstrate an association between the susceptible genotypes on exons 2 and 3 of CYP1B1 and endometrial cancer. In this regard, Hanna et al. (8) observed that polymorphic variants on exons 2 and 3 of CYP1B1 display a much stronger estrogen hydroxylation activity than wild-type in breast cancer (18). These results indicate that polymorphisms on exons 2 and 3 modify enzymatic activity, thus increasing risk of developing endometrial and breast cancer because CYP1B1 expression is particularly high in these cancers (4, 8, 14, 17). When treating MCF-7 breast cancer cells with 4-hydroxy estrogens, the rate of cell proliferation and the expression of estrogen-inducible genes were increased (4, 8, 14, 17). In addition, 4-hydroxy estrogens can undergo redox cycling that result in the formation of free radicals such as superoxide and in the generation of reactive semiquinones/quinone intermediates that have been shown to damage biological target molecules such as DNA (14, 15). Exon 1 of the human CYP1B1 gene is noncoding, and no polymorphism has been identified in it. The polymorphism in intron 1 is located 13 nucleotides upstream of the 5′-end of exon 2 and is, therefore, unlikely to affect enzyme activity. The polymorphism at codon 449 is silent, because the amino acid sequence is not affected by this mutation (9, 10). It seems unlikely that polymorphisms other than those in codons 48, 119, 432, and 453 are present in the coding region (9, 10, 18).

The 119T/T also showed a significant correlation for ERα and ERβ positivities. Previous studies have shown that genotype 432G/G has a positive correlation with ER and PR expression in breast cancer, although some studies have shown no association between this polymorphism and breast cancer risk (8, 9, 19). In other studies, treatment of MCF-7 breast cancer cells with 4-hydroxy estrogens increased the rate of cell proliferation and the expression of estrogen-inducible genes such as ER and PR (20).

This is the first report to show that codons 119 and 432 polymorphisms on the CYP1B1 gene may have higher risk of endometrial cancer. The results of these experiments are important in understanding the role of CYP1B1 polymorphisms and activation of ERs in the pathogenesis of endometrial cancer.