Functional POR A503V is associated with the risk of bladder cancer in a Chinese population

Human cytochrome P450 oxidoreductase (POR) plays important roles in the metabolism of exogenous carcinogens and endogenous sterol hormones. However, few studies have explored the association between POR variants and the risk of bladder cancer. In this study, we first sequenced all 16 POR exons among 50 randomly selected controls, and found three variants, rs1135612, rs1057868 (A503V) and rs2228104, which were then assessed the relation to risk of bladder cancer in a case-control study of 1,050 bladder cancer cases and 1,404 cancer-free controls in a Chinese population. People with A503V TT genotype have a decreased risk of bladder cancer in a recessive model (TT vs. CC/CT, OR = 0.73, 95% CI = 0.57–0.93), which was more pronounced among elderly male, non-smoking, subjects. Especially, A503V TT genotype showed a protective effect in the invasive tumor stage. Functional analysis revealed that A503V activity decreased in cytochrome c reduction (50.5 units/mg vs. 135.4 units/mg), mitomycin C clearance (38.3% vs. 96.8%), and mitomycin C-induced colony formation (78.0 vs 34.3 colonies per dish). The results suggested that POR A503V might decrease the risk of bladder cancer by reducing its metabolic activity, and should be a potential biomarker for predicting the susceptibility to human bladder cancer.

Bladder cancer is the 10th most common cancer in the world. It was reported to have the highest incidence in Europe, North America, and Australia but a relatively low prevalence in Asia 1 . However, the etiology and pathogenesis of bladder cancer remain poorly understood. Accumulating epidemiological studies suggested that age, sex, smoking, and occupational exposure are risk factors of bladder cancer 2,3 . However, only a fraction of individuals exposed to specific risk factors develop bladder cancer in their lifetime, suggesting that genetic susceptibility plays a role in bladder cancer.
During the past few years, genome-wide association studies (GWAS) have identified a large number of robust associations between specific chromosomal loci and complex human diseases, including bladder cancer. Several significant GWAS of bladder cancer have identified over 10 independent loci and single-nucleotide polymorphisms (SNPs) [4][5][6][7][8][9][10][11] . In general, metabolic enzyme genes and DNA damage repair genes (i.e., hOGG1 and XRCC) are usually involved in protecting against toxicant-induced cell damage 12 . The balance between the activation and detoxification of carcinogens is responsible for the amount of DNA damage repair. Accordingly, a growing number of genes that encode metabolic enzymes have been identified as tumor susceptibility genes [13][14][15] . For example, metabolic (detoxic) enzyme cytochrome P450 2E1 (CYP2E1) and glutathione S-transferase P1 (GST P1) have been found to be modulators of bladder cancer risk 14,15 . Choi et al. found that the CYP2E1 variants might act an important role in the development of smoking-related bladder cancer among Korean men 14 . Besides, N-acetyltransferase 2 (NAT2) and GSTM1 are associated with bladder cancer risk 16,17 . The effect of NAT2 slow acetylation was stronger for smokers of black tobacco, and GSTM1 may reduce the risk of bladder cancer through mechanisms that are not specific to the detoxification of polycyclic aromatic hydrocarbons (PAHs) in tobacco smoke 16,18 . Rothman et al. also identified the interactions between the GSTM1 deletion and a tag SNP for NAT2 acetylation status and bladder cancer risk 4 , which suggested that bladder cancer susceptibility loci participated in tumorigenesis by complex combination of genetic and environmental factors.
Cytochrome P450 oxidoreductase (POR; EC 1.6.2.4) was initially identified by Horecker in 1950 as a cytochrome c reductase 19 . It is an obligate donor that mediates electron transfer from nicotinamide adenine dinucleotide phosphate (NADPH) to cytochrome P450 (CYP) enzymes, which is involved in catalyzing the metabolism of drugs, xenobiotics, and steroids 20,21 . POR has two separate lobes: flavin adenine dinucleotide (FAD) and flavin mononucleotide (FMN). During the electron transfer process, electrons pass from NADPH through FAD to FMN and finally to P450 enzymes 21 . POR is not only involved in electron transfer; it is also an active enzyme that directly participates in the metabolism of exogenous chemicals and drugs, such as mitomycin C (MMC), paraquat, midazolam, and quinidine 22,23 . Knockout of POR gene was shown to be lethal in the embryonic stage in mice 24 , and liver-specific deletion of the POR gene led to a severe disruption of hepatic drug metabolism 25 . Therefore, POR may involve in the metabolism of xenobiotics, particularly the metabolic activation of environmental carcinogens that leads to cancer. POR variations that influence the FAD or FMN domain may also affect the electron transfer process and metabolic activity of CYP enzymes.
Human POR gene is located on chromosome 7 and consists of 15 protein-coding exons and an untranslated exon (exon 1U) that lies 38.8 kb upstream and initiates transcription 26 . It is expressed in various normal tissues and tumor cells. The gene that encodes human POR is genetically polymorphic. Six POR variants (Y181D, A287P, R457H, V492E, C569Y, and V608F) were previously identified 27 , and more than 50 variations have been subsequently reported. Most of those defects in the POR gene can cause POR dysfunction 28 and change the affinity for CYP enzymes, thereby reducing CYP activity 29 . Interindividual variations show at least 4-to 5-fold differences in POR activity in human liver microsomes 30 . Genetic polymorphisms in the POR gene were associated with many autosomal recessive genetic diseases, such as ambiguous genitalia, congenital adrenal hyperplasia, Antley-Bixler syndrome, and polycystic ovary syndrome 27,31 . The common mechanism in each of these diseases is based on POR deficiency. The association between POR polymorphisms and human cancer, however, has been rarely reported. Haiman et al. first reported that the synonymous POR SNP G5G might increase breast cancer risk in African Americans, but the variant A503V was not significantly associated with breast cancer risk in any racial or ethnic group 32 . To date, no study between POR variants and bladder cancer risk has been reported.
Considering the importance of POR in the CYP-mediated metabolism of exogenous and endogenous substrates, it is rational to assume that POR variants, particularly missense variants, could cause metabolic disorders directly or indirectly via the CYP metabolic pathway, thereby contribute to the risk of bladder cancer. To test this hypothesis, we genotyped POR gene polymorphisms in the coding region and explored their relation to the risk of bladder cancer in our ongoing, hospital-based, case-control study in a Chinese population. Moreover, to illustrate the possible mechanism, mutant and wild type POR protein was expressed in a baculovirus/Sf9 insect system, and cytochrome c reduction activity and the metabolic activity of MMC, a direct substrate of POR, were used to evaluate POR metabolic activity.

Results
Demographic characteristics. The demographic characteristics and risk factor distributions between cases and controls are summarized in Supplementary Table S2. Cases and controls were well-matched by age (P = 0.316) and sex (P = 0.573). However, the cases were overrepresented by individuals who were former smokers (P < 0.001), especially heavy smokers (P < 0.001). The distribution of tumor grade for G1, G2, and G3 was 49.2%, 35.2%, and 15.6%, respectively, among the cases. For tumor stage, 65.5% of the patients were in the superficial stage, and 35.5% were in the invasive stage.
Identification of POR variants. Following the general principle that the SNP minor allele frequency (MAF) is greater than 5%, we conducted DNA sequencing in 50 randomly selected controls to ensure that the variations occurred in the coding region of genomic DNA in our study population. As shown in Fig. 1, three POR gene polymorphisms were observed, located in exon 4 (rs1135612, A > G) and exon 12 (rs2228104, C > T; rs1057868, C > T). The MAF of rs1135612, rs2228104 and rs1057868 was 0.509, 0.140 and 0.382, respectively, which were similar to the MAF in the HapMap database (0.458, 0.133, and 0.396, respectively; see Supplementary Table S3). Among these, rs1135612 and rs2228104 are synonymous mutations (Pro130Pro and Ala486Ala), but rs1057868 is a missense mutation (Ala503Val, A503V).
Considering the hypothesis based on functional changes induced by the POR missense mutation, we chosed rs1057868 as a candidate SNP. The MAF of rs1057868 in the cases and controls was 0.367 and 0.374, respectively, which is consistent with HapMap database,0.396. The frequency distribution for this SNP in the controls did not deviate from HWE (p = 0.242).
Distribution and main effects of POR polymorphism. Seven samples (six cases and one control) failed to genotype for rs1057868 (A503V), therefore, total of 1,044 cases and 1,403 controls were included in the final analysis. Table 1 shows the genotype distribution of A503V in both cases and controls. Compared with the CC genotype, people with A503V CT genotype (OR = 1.095, 95% CI = 0.985-1.322, Scientific RepoRts | 5:11751 | DOi: 10.1038/srep11751 P = 0.054) and TT genotype (OR = 0.802, 95% CI = 0.617-1.042, P = 0.099) have not significant association with the risk of bladder cancer. Furthermore, we analyzed the data in both dominant and recessive models to identify the effects of the genotypes on bladder cancer. In the recessive genetic model, risk of bladder cancer significantly decreased by 27.3% with the A503V TT genotype compared with the CC/ CT genotype (OR = 0.727, 95% CI = 0.570-0.927, P = 0.008). However, no significant association was found in the dominant model (OR = 0.911, 95% CI = 0.772-1.074, P = 0.265), suggesting that the A503V polymorphism tended to be a recessive factor.

Comparison of the activity of wildtype and mutant POR proteins.
To investigate the changes in A503V POR, we expressed both wildtype and A503V POR proteins in a baculovirus/Sf9 system. The sequencing of POR cDNA indicated a nuclei acid change from GCC to GTC at codon 1508 ( Fig. 2A). Compared with the positive control from rat liver microsomes, the immunoblotting assay data showed   a single 78 kDa band in microsomes prepared from either wildtype Sf9 cells or Sf9 cells infected with A503V POR cDNA, whereas the vector protein showed no band (Fig. 2B). The cytochrome c reduction assay is a popular method to evaluate POR reduction ability. Thus, we recorded reduced cytochrome c absorbance every 10 s for 3 min. The results showed a decreased reduction rate in A503V POR compared with wildtype POR, whereas the vector protein showed very little absorbance (Fig. 3A). The reduction activity of A503V (50.5 units/mg) was significantly lower than wildtype POR (135.4 units/mg, p < 0.01), which was only 36.4% of wildtype POR. The vector protein showed nearly no activity (6.0 units/mg) (Fig. 3B).
To investigate POR-mediated metabolic activity, the metabolic clearance efficiency of MMC, a direct substrate of POR, was evaluated using an in vitro metabolism system. The results clearly showed that A503V POR metabolized MMC at a significantly lower rate than wildtype POR (Fig. 4). After 10 min, only 38.3% of MMC was cleared by A503V POR, whereas 96.8% of MMC was cleared by wildtype POR (P < 0.01). As expected, the vector protein showed nearly no MMC metabolic activity (9.9%). In addition, MMC induced significant decrease in colony formation in A503V cells than that in vector cells (P < 0.01) (Fig. 5). Expression of POR protein using an immunoblotting assay. Cell lysate proteins (30 μ g) were prepared from Sf9 cells that expressed wildtype POR and A503V POR. Cell lysate protein from the Sf9 cells transfected with the vector alone was used as a negative control (Vector). Microsomal proteins from rat liver were used as a positive control (P), and tubulin was used as an internal reference (Tubulin).

Discussion
The gene that encodes human POR is genetically polymorphic. To date, more than 50 POR variants have been identified. As an obligate electron donor for CYP enzymes, POR plays a critical role in the metabolism of exogenous and endogenous substrates. Although the associations between POR variants and human disease have been occasionally reported, very few studies have investigated the relationship between POR variants and human bladder cancer. In present study, we genotyped three POR  polymorphisms among 50 controls and found a missense mutation, A503V, that changed the amino acid from alanine to valine at codon 503. The A503V polymorphism has been reported to be the most common POR variant, which is a conservative change in an unstructured loop of the FAD binding domain 26 , an important element for electron transfer. The frequency of this variant (36.8%) in our study was similar to Huang et al. 33 who found a frequency of 36.7% in Chinese American alleles, and a recent study in healthy Chinese men (43.2%) 34 .
We then found that people with A503V TT genotype have a decreased bladder cancer risk compared one with the CC/CT. Furthermore, the association was obviously stronger in older subjects (> 65 years old), male subjects, and never-smoking subjects. In bladder cancer cases, the A503V TT genotype showed a protective effect on tumor grade, especially in the invasive tumor stage. Further functional studies showed that the A503V variant not only decreased cytochrome c reduction activity, which was similar to previous reports 28,33 , but also reduced MMC-metabolizing capability, a direct substrate of POR 23 . Furthermore, A503V variant decreased the ability of MMC-induced the colony formation in Flp-In CHO cells. The present study demonstrated that A503V POR might decrease bladder cancer risk by reducing its electron transfer activity or exogenous and endogenous substrate metabolism.
Cigarette smoking is a primary risk factor for bladder cancer, accounting for approximately 50% of all cases in the United States 2,35 . Cigarette smoke contains thousands of toxic, hazardous, and complex substances, among which several chemicals have been confirmed to be carcinogenic, such as nicotine, benzo-a-pyrene (BaP), and 4-aminobiphenyl. The metabolic activation of CYP450 enzymes, especially CYP1A2, has been shown to be responsible for the carcinogenesis of almost all of these carcinogens 36,37 . The first step in this activation pathway is N-oxidation. Several studies found that smokers have higher CYP1A2 activity than nonsmokers 38 , which was thought to be a genetic factor in the susceptibility to bladder cancer caused by arylamine and PAH 37 . To date, 21 alleles that contain more than 30 SNPs have been identified in the CYP1A2 gene 39 . However, a limited number of these polymorphisms and haplotypes in CYP1A2 are reported to reflect interindividual differences in enzyme activity. Interestingly, POR A503V was reported to cause a 15% decrease in CYP1A2 activity in vitro 40 . A recent in vivo study further found significantly higher CYP1A2 activity in carriers of the A503V TT haplotype after smoking cessation 41 , suggesting that POR A503V might be involved in smoking-mediated bladder cancer by decreasing CYP1A2 activity. Additionally, some studies found the POR A503V reduced CYP3A4 activity by 61-97% compared with wildtype POR with variable substrates, such as testosterone, midazolam, quinidine, and erythromycin 29,42 . N'-nitrosonornicotine (NNN), a highly oncogenic tobacco-specific nitrosamine, is primarily metabolically activated by CYP3A4 to exert its carcinogenic effects 43 . The POR A503V variant-mediated decrease in metabolic CYP3A4 activation would generate decreased 2'-hydroxylated NNN and possibly reduce bladder cancer risk. Tobacco smoking, along with occupational exposure to aromatic amines, was first noted as the most notable risk factor for the development of bladder cancer in England over 100 years ago 44,45 . The compounds 2-naphthylamine, 4-aminobiphenyl, and benzidine can be widely found in rubber products, hair dyes, paints, cigarette smoke, motor vehicle exhaust, and industrial pollutants 46 . All arylamines require CYP enzymes (i.e., CYP1A1, CYP1A2, and CYP3A4) to exert their genotoxic effects 37,43 . Therefore, A503V POR might decrease the risk of bladder cancer through reducing the metabolic activities of CYP enzymes in exerting ultimate carcinogens from parent compounds.
Men were reported to experience 3-to 4-times higher rates of bladder cancer than women in the United States and Europe 47 . Several studies showed that androgen and the androgen receptor (AR) play important roles in the development and progression of bladder cancer 48,49 . Treatment with N-butyl-N-(4-hydroxybutyl) nitrosamine caused the development of bladder cancer in more than 92% of wildtype male mice and 42% of wildtype female mice, whereas no AR knockout mice developed bladder cancer 48 . Additionally, clinical anti-androgen treatment and AR downregulation abolished androgens 50 . Thus, androgen likely promotes bladder carcinogenesis. The human body has three steroidogenic enzymes, P450c17 (17,20-lyase), P450c21 (21-hydroxylase), and P450aro (aromatase), all of which are involved in adrenal steroidogenesis and receive electrons from NADPH via POR. P450c17 plays an important role in the conversion of pregnenolone and progesterone to androstenedione and testosterone, respectively 51 . The activity of P450c17 is especially sensitive to disturbances in electron transport; thus, POR-deficient males are always hypoandrogenic 20 . Numerous studies reported a decrease in the activity of POR A503V on P450c17. Huang's study showed that POR A503V retained 60% of 17,20-lyase activity and 56% of 17a-hydroxylase activity using human P450c17 52 , confirming that POR A503V led to the degradation of P450c17 activity compared with wildtype POR 29 . Therefore, the POR A503V might contribute to a decrease in the catalytic activity of P450c17 and subsequent androgen formation, which may ultimately attenuate bladder cancer risk.
In the present study, it should be noted that people with A503V TT genotype have a decreased risk of bladder cancer in a recessive model (TT vs. CC/CT, OR = 0.727, P = 0.008) rather than in the dominant model (CT/TT vs. CC, OR = 0.911, P = 0.265). In general, all the GWAS-identified cancer susceptibility loci have been either additional or dominant models, and recessive model tends to have potential false positive results. A false positive is an error in data reporting in which a test result improperly indicates presence of a condition 4,5 . The sample size is the important factor that influences the results, and small sample often increases rates of false-positive and false-negative findings 6,7 . The present study included 1044 cases and 1403controls, and there was 118 and 207 TT genotype among the case and control, respectively, which suggested that the statistical power should be enough to detect the statistical significance. Moreover, the functional analysis revealed that A503V activity decreased in cytochrome c reduction, mitomycin C clearance, and mitomycin C-induced colony formation, which should be biologically plausible that the people carrying TT genotype had a lower predisposition of bladder cancer as discussed above, and still provide the reliability of results in the recessive model.
In conclusion, we found that POR A503V has a significantly association with bladder cancer risk among the Chinese population. Because of the decreased activity of POR A503V in cytochrome c reduction, MMC metabolism, and MMC-induced colony formation, two possible pathways could be predicted: (1) POR A503V variants decreases the metabolic activation of CYP enzymes, subsequently reducing carcinogens formed from tobacco smoking and occupational exposure, and (2) POR A503V variants reduces the catalytic ability of P450c17 to slow androgen formation. Both of these pathways would eventually contribute to a decreased bladder cancer risk. The present study may contribute to a better mechanistic understanding of environmental carcinogen-induced bladder cancer and provide a potential biomarker for predicting the susceptibility to human bladder cancer.

Methods
Study subjects. This case-control study was performed in 1,050 bladder cancer patients and 1,404 cancer-free controls. Bladder cancer patients who had a histopathological diagnosis were consecutively recruited in The First Affiliated Hospital of Nanjing Medical University beginning in 2003, and the persons who sought the general physical examinations at the outpatient were recruited as the controls in the same hospital. All of the controls were genetically unrelated to the cases and had neither a cancer history nor clinical symptoms of bladder cancer. Additionally, cancer-free controls were frequency-matched to cases based on age (± 5 years) and sex. Before being interviewed via questionnaire to obtain the demographic and lifestyle characteristics of the subjects, all of the participants were asked to sign an informed consent form. Subsequently, 5 ml of peripheral venous blood was obtained from each subject for genomic DNA extraction. The response rates for the cases and controls in this study were > 85%. The study was approved by the Institutional Review Board of Nanjing Medical University, and conducted according to the principles expressed in the Declaration of Helsinki.
Sequencing and genotyping. Genomic DNA was isolated from whole blood using a traditional phenol-chloroform method. To screen the hypothesized functional variants, we designed primers for 16 Scientific RepoRts | 5:11751 | DOi: 10.1038/srep11751 exons throughout the coding region of the POR gene (Generay Biotech, Shanghai, China). The primers and reaction conditions are listed in Supplementary Table S1. Fifty controls were randomly selected from the controls for sequencing on an automated sequencer-ABI model 377 DNA analyzer (Applied Biosystems, Foster City, CA, USA). The sequencing results were subjected to a Basic Local Alignment Search Tool with the wild type POR gene (POR WT; NM_000941). Genotyping for the selected SNPs was performed using the TaqMan assay and ABI 7900HT Real Time PCR System (Applied Biosystems, Foster City, CA, USA). Approximately 10% of the randomly selected samples were analyzed in duplicate, and the results showed 100% consistency.
Expression of wildtype and mutant POR proteins in baculovirus/sf9 system. Recombinant plasmids of pFastBac ™ 1-POR (wildtype) and mutant POR cDNA were prepared in our laboratory previously. POR proteins were expressed in the Bac-to-Bac baculovirus expression system, and the microsomes were prepared according to the manual's instructions. The proteins were then identified by an immunoblotting assay using a specific antibody to human POR. The protein from the Sf9 insect cells transfected with vector alone was used as a negative control.
Determination of the activity of wildtype and mutant POR protein. The cytochrome c assay was used to determine the activity of POR. Briefly, microsomal protein (diluted approximately 30 times) was placed in a 1 cm optical path cuvette, to which 80 nM cytochrome c was added, and baseline was recorded. The initial rate of the cytochrome c reduction was monitored at 550 nm for 3 min after 0.2 mM NADPH was added and immediately mixed. Activity is expressed as units per milligram protein. One unit of reductase activity is defined as 1 μ mol cytochrome c reduced per minute.

Mitomycin C metabolism and HPLC analysis. Mitomycin C (MMC) metabolism in the presence
of wildtype or mutant POR protein was performed in an oxygen-free environment in vitro. Briefly, sodium phosphate buffer, protein microsomes, and MMC were first added to the sealed test tubes, and then NADPH was added to the tubes as a cofactor for the metabolic reaction. The total volume of the incubation solution was 240 μ l. Nitrogen gas was used to generate an oxygen-free environment from beginning to end. The reaction was performed at 30 °C for 0-20 min and terminated by the addition of 240 μ l ice-cold acetonitrile to precipitate the proteins and extract MMC. The reaction mixtures were stored at − 20 °C overnight and then centrifuged at 10,000 × g for 15 min prior to high-performance liquid chromatography (HPLC) described previously 53 . Authentic standards of MMC were used to establish the chromatographic retention times. The corresponding integrated peaks were quantified with the regression formulae obtained from standard curves.
Colony formation assay in Flp-In CHO cells expressing wildtype and mutant POR. The procedures to establish the Flp-In CHO cells stably expressing wild-type and mutant human POR were described previously 22 . The Flp-In system could make the similar expression efficiency among the cells transfected with different plasmids. Then the cells were plated into a six-well plate (5 × 10 5 cells per well) and grown for 24 h. After incubation with MMC (2 μ M) for 2 h, the cells were grown in the replaced fresh medium for 24 h. The cells were then replated at a density of 400 cells/100-mm dish and grown for 7 to 8 days until the discrete colonies could be visualized. After being washed with phosphate-buffered saline, the colonies were stained with 0.5% crystal violet in ethanol and counted. In each group, the cells incubated with vehicle only were set as control.
Statistical analysis. All of the statistical analyses were performed using Statistical Analysis System software (version 9.1; SAS Institute, Cary, NC, USA) based on a two-sided α level of 0.05. Demographic characteristics, risk factors, and genotype frequency distributions between the case and control groups were analyzed using the χ 2 test. The Hardy-Weinberg Equilibrium (HWE) of the genotype distribution in the control group was analyzed using the goodness-of-fit χ 2 test. Associations between the mutant genotype and bladder cancer were evaluated using a recessive genetic model. The effects of genotype on bladder cancer risk were measured based on the adjusted ORs and corresponding 95% CIs in unconditional logistic regression models. The Mann-Whitney test and analysis of variance (ANOVA) were used to compare differences in activity between wildtype and mutant POR protein.