Artificial neural network-based exploration of gene-nutrient interactions in folate and xenobiotic metabolic pathways that modulate susceptibility to breast cancer.

In the current study, an artificial neural network (ANN)-based breast cancer prediction model was developed from the data of folate and xenobiotic pathway genetic polymorphisms along with the nutritional and demographic variables to investigate how micronutrients modulate susceptibility to breast cancer. The developed ANN model explained 94.2% variability in breast cancer prediction. Fixed effect models of folate (400 µg/day) and B12 (6 µg/day) showed 33.3% and 11.3% risk reduction, respectively. Multifactor dimensionality reduction analysis showed the following interactions in responders to folate: RFC1 G80A × MTHFR C677T (primary), COMT H108L × CYP1A1 m2 (secondary), MTR A2756G (tertiary). The interactions among responders to B12 were RFC1G80A × cSHMT C1420T and CYP1A1 m2 × CYP1A1 m4. ANN simulations revealed that increased folate might restore ER and PR expression and reduce the promoter CpG island methylation of extra cellular superoxide dismutase and BRCA1. Dietary intake of folate appears to confer protection against breast cancer through its modulating effects on ER and PR expression and methylation of EC-SOD and BRCA1.

Molecular insights into the association of obesity with breast cancer risk: relevance to xenobiotic metabolism and CpG island methylation of tumor suppressor genes.

Obesity, genetic polymorphisms of xenobiotic metabolic pathway, hypermethylation of tumor suppressor genes, and hypomethylation of proapoptotic genes are known to be independent risk factors for breast cancer. The objective of this study is to evaluate the combined effect of these environmental, genetic, and epigenetic risk factors on the susceptibility to breast cancer. PCR-RFLP and multiplex PCR were used for the genetic analysis of six variants of xenobiotic metabolic pathway. Methylation-specific PCR was used for the epigenetic analysis of four genetic loci. Multifactor dimensionality reduction analysis revealed a significant interaction between the body mass index (BMI) and catechol-O-methyl transferase H108L variant alone or in combination with cytochrome P450 (CYP) 1A1m1 variant. Women with "Luminal A" breast cancer phenotype had higher BMI compared to other phenotypes and healthy controls. There was no association between the BMI and tumor grade. The post-menopausal obese women exhibited lower glutathione levels. BMI showed a positive association with the methylation of extracellular superoxide dismutase (r = 0.21, p < 0.05), Ras-association (RalGDS/AF-6) domain family member 1 (RASSF1A) (r = 0.31, p < 0.001), and breast cancer type 1 susceptibility protein (r = 0.19, p < 0.05); and inverse association with methylation of BNIP3 (r = -0.48, p < 0.0001). To conclude based on these results, obesity increases the breast cancer susceptibility by two possible mechanisms: (i) by interacting with xenobiotic genetic polymorphisms in inducing increased oxidative DNA damage and (ii) by altering the methylome of several tumor suppressor genes.

Impact of hyperhomocysteinemia on breast cancer initiation and progression: epigenetic perspective.

Our recent study showing association of hyperhomocysteinemia and hypomethioninemia in breast cancer and other studies indicating association of hyperhomocysteinemia with metastasis and development of drug resistance in breast cancer cells treated with homocysteine lead us to hypothesize that homocysteine might modulate the expression of certain tumor suppressors, i.e., RASSF1, RARß1, CNND1, BRCA1, and p21, and might influence prognostic markers such as BNIP3 by inducing epigenetic alteration. To demonstrate this hypothesis, we have treated MCF-7 and MDA-MB-231 cells with different doses of homocysteine and observed dose-dependent inhibition of BRCA1 and RASSF1, respectively. In breast cancer tissues, we observed the following expression pattern: BNIP3 > BRCA1 > RARß1 > CCND1 > p21 > RASSF1. Hyperhomocysteinemia was positively associated with BRAC1 hypermethylation both in breast cancer tissue and corresponding peripheral blood. Peripheral blood CpG island methylation of BRCA1 in all types of breast cancer and methylation of RASSF1 in ER/PR-negative breast cancers showed positive correlation with total plasma homocysteine. The methylation of RASSF1 and BRCA1 was associated with breast cancer initiation as well as progression, while BRCA1 methylation was associated with DNA damage. Vitamin B12 showed inverse association with the methylation at both the loci. RFC1 G80A and cSHMT C1420T variants showed positive association with methylation at both the loci. Genetic variants influencing remethylation step were associated positively with BRCA1 methylation and inversely with RASSF1 methylation. GCPII C1561T variant showed inverse association with BRCA1 methylation. We found good correlation of BRAC1 (r = 0.90) and RASSF1 (0.92) methylation pattern between the breast cancer tissue and the corresponding peripheral blood. To conclude, elevated homocysteine influences methionine dependency phenotype of breast cancer cells and is associated with breast cancer progression by epigenetic modulation of RASSF1 and BRCA1.

Association of glutamate carboxypeptidase II (GCPII) haplotypes with breast and prostate cancer risk.

In view of the pivotal role of glutamate carboxypeptidase II (GCPII) in carcinogenesis, its expression as prostate specific membrane antigen (PSMA) and folate hydrolase (FOLH1) may be influenced by its haplotypes, contributing to the etiology of prostate and breast cancer. To test this hypothesis, breast and prostate cancer cases and controls were subjected to whole gene screening of GCPII and correlated with plasma folate levels and PSMA expression. The impact of variants on a 3-dimensional structure of GCPII was explored by in silico studies. Six novel variations i.e. V108A, P160S, Y176H, D191V, G206R and G245S; and two known variations i.e. R190W and H475Y were identified in GCPII. All-wild haplotype and a haplotype harbouring D191V showed association with breast cancer risk while haplotypes harbouring V108A and P160S reduced the risk. Haplotypes with V108A and G245S variants showed increased risk for prostate cancer due to high PSMA expression while P160S conferred protection against prostate cancer. In silico studies suggests that P160S and R190W variants result in relaxed substrate binding facilitating either rapid catalysis or exchange of substrates and products in the active site which was substantiated by high plasma folate levels associated with these variants. On the contrary, D191V was associated with very low plasma folate levels despite having a high PSMA expression. This is the first comprehensive study examining variations in GCPII in relation to breast and prostate cancer risk. Changes in the plasma folate levels and changes in PSMA expression are associated with breast and prostate cancer risk respectively.

Paradoxical role of C1561T glutamate carboxypeptidase II (GCPII) genetic polymorphism in altering disease susceptibility.

Glutamate carboxypeptidase II (GCPII) is predominantly expressed in brain, intestinal mucosa and prostate cancer in the form of three splice variants i.e. N-acetylated-α-linked acidic dipeptidase (NAALADase), folyl poly-γ-glutamate carboxypeptidase (FGCP) and prostate specific membrane antigen (PSMA) respectively. Its inhibition was found to confer protection against certain neurological disorders and cancer. Despite the pivotal role of this enzyme, the most common polymorphism i.e. H475Y has not been explored comprehensively in all its splice variants. In this study, we have determined the role of this variant in different disease conditions such as breast and prostate cancers, autism, coronary artery disease (CAD) and miscarriages (N=1561). Genotyping was done by PCR-RFLP and dideoxy sequencing. Plasma folate levels were estimated by Axysm folate kit. GCPII expression was studied by semi-quantitative RT-PCR. In silico model was developed using PYMOL. We observed the protective role of H475Y variant in cancers [breast cancer; OR (95% CI): 0.81 (0.55-1.19), prostate cancer: OR (95% CI): 0.00 (0.00-0.66)], and in autism (OR (95% CI): 0.47 (0.21-1.03), whereas inflated risk was observed in CAD (OR (95% CI): 1.69 (1.20-2.37) and miscarriages [Maternal OR (95% CI): 3.26 (2.11-5.04); Paternal OR(95% CI): 1.99 (1.23-3.21)]. Further, this variant was found to impair the intestinal folate absorption in subjects with dietary folate intake in the lowest tertile (CC vs. CT in lowest tertile; 7.56±0.85ng/ml vs. 2.73±045ng/ml, p=0.005). In silico model of GCPII showed steric hindrance with H475Y resulting in stereochemical alteration of catalytic site, thus interfering with ligand binding. Statistically significant association was not observed between dietary folate levels and GCPII expression. However, a positive correlation was seen between plasma folate levels and GCPII expression (r=0.70, p<0.05). To conclude, our data suggests that GCPII H475Y variant shows inverse association with autism and cancer while showing positive association with CAD and miscarriages.

Bcl-2/adenovirus E1B 19 kDa-interacting protein 3 (BNIP3) expression is epigenetically regulated by one-carbon metabolism in invasive duct cell carcinoma of breast.

In view of recent studies highlighting the prognostic relevance of expression and CpG island methylator phenotype (CIMP) of Bcl-2/adenovirus E1B 19 kDa-interacting protein 3 (BNIP3) in invasive duct cell carcinoma (IDC), we hypothesized in this article that impaired one-carbon metabolism might influence CIMP phenotype of BNIP3. In order to substantiate the prognostic relevance of BNIP3, we explored its association with 8-oxo-2'deoxyguanosine (8-oxodG), a marker of oxidative stress with prognostic relevance. BNIP3 expression and CIMP phenotype were studied using semi-quantitative RT-PCR and combined bisulfite restriction analysis (COBRA), respectively, in 56 IDC tumors. Eight polymorphisms in one-carbon metabolism were studied using PCR-RFLP and PCR-AFLP approaches. 8-oxodG was measured using competitive ELISA kit. BNIP3 was found to be upregulated in IDC (cases vs. controls: 0.94 ± 0.05 vs. 0.18 ± 0.08, P < 0.0001). COBRA analysis confirmed hypomethylation of BNIP3 promoter CpG island in these cases. CIMP phenotype of BNIP3 showed positive association with tubule formation (P = 0.034) and methionine synthase reductase (MTRR) A66G (P = 0.002); inverse association with cytosolic serine hydroxyl methyltransferase (cSHMT) C1420T (P < 0.005) and 8-oxodG (<10% vs. >10% methylation: 7.24 ± 2.77 ng/ml vs. 4.42 ± 2.93 ng/ml, P < 0.0005); and no association with nuclear pleomorphism or mitotic index or ER, PR, and HER statuses. Synergistic effect of MTR A2756G and MTRR A66G variants on BNIP3 hypermethylator phenotype was clearly evident (P < 0.0007). MTRR A66G and cSHMT C1420T polymorphisms influence CIMP phenotype of BNIP3, thus epigenetically regulating BNIP3 in breast cancer. The linear association between BNIP3 and 8-oxodG substantiates the role of BNIP3 as redox sensor as well as prognostic marker in breast cancer.

Association of aberrations in one-carbon metabolism with molecular phenotype and grade of breast cancer.

We have earlier demonstrated the role of aberrant one-carbon metabolism in the etiology of breast cancer. In the current study, we examine the clinical utility of these factors in predicting the subtype of breast cancer and as indicators of disease progression. Polymerase chain reaction (PCR)-restriction fragment length polymorphism (RFLP) and PCR-amplified fragment length polymorphism (AFLP) approaches were used for genetic analysis. Plasma folate and homocysteine were measured using Axsym folate kit and reverse phase HPLC, respectively. Multiple linear regression models were used to test the predictability of disease progression. Luminal A subtype was associated with late age of onset, higher body mass index and lack of family history of breast cancer. Thymidylate synthase (TYMS) 5'-UTR 28 bp tandem repeat (OR: 2.09, 95% CI: 1.05-4.16) and methylene tetrahydrofolate reductase (MTHFR) C677T (OR: 4.10, 95% CI: 1.40-11.95) were strongly associated with Luminal B. Reduced folate carrier (RFC1) G80A (OR: 2.92, 95% CI: 1.22-6.97) and methionine synthase (MTR) A2756G (OR: 4.71, 95% CI: 1.66-13.31) polymorphisms were associated with LuminA-HH subtype while MTHFR C677T showed association with HER-enriched (OR: 30.41, 95% CI: 6.47-142.91). Cytosolic serine hydroxymethyltransferase (cSHMT) conferred protection against basal-like breast cancer (OR: 0.47, 95% CI: 0.22-0.98). HER-enriched and basal-like subtypes showed positive association with familial breast cancer and inverse association with plasma folate. Hyperhomocysteinemia was observed in Luminal B and basal-like subtypes. Multiple linear regression models of aberrant one-carbon metabolism were found to be moderate predictors of breast cancer grade (area under the receiver operating characteristic curve, C = 0.72, 95% CI: 0.58-0.87, P = 0.008). To conclude, aberrations in one-carbon metabolism predict the subtype of breast cancer and disease progression.

Modulatory effect of plasma folate and polymorphisms in one-carbon metabolism on catecholamine methyltransferase (COMT) H108L associated oxidative DNA damage and breast cancer risk.

The present study was aimed to investigate the modulatory role of plasma folate and eight putatively functional polymorphisms of one-carbon metabolism on catecholamine methyltransferase (COMT)-mediated oxidative DNA damage and breast cancer risk. Plasma folate and 8-oxo-2'-deoxyguanosine (8-oxodG) were estimated by commercially available kits, while polymorphisms were screened by PCR-RFLP and PCR-AFLP methods. COMT H108L polymorphism showed independent association with breast cancer (OR: 1.73, 95% CI: 1.31-2.30). No significant interaction was observed between folate status and COMT genotype. Multifactor dimensionality reduction (MDR) analysis gave evidence for the significant epistatic (gene-gene) interactions (p<0.0001) of COMT H108L with reduced folate carrier 1 (RFC1) G80A, thymidylate synthase (TYMS) 5'-UTR 3R2R, TYMS 3'-UTR ins6/de16. Increased plasma 8-oxodG were observed in cases compared to controls (mean +/- SE: 5.59 +/- 0.60 vs. 3.50 +/- 0.40 ng/ml, p<0.004). Plasma folate deficiency alone was not a significant predictor of 8-oxodG elevation. The genotype combinations namely, RFC1 G80A/methionine synthase reductase (MTRR) A66G, RFC1 G80A/SHMT C1420T/TYMS 3R2R and serine hydroxymethyltransferase (SHMT) C1420T/TYMS 3R2R/methionine synthase (MTR) A2756G/COMT H108L were strong predictors of 8-oxodG elevation in the order of risk. To conclude, the current study provides substantial evidence for a cross talk between one-carbon metabolism and COMT catalysis that might influence oxidative DNA damage and breast cancer risk.

Cross-talk between one-carbon metabolism and xenobiotic metabolism: implications on oxidative DNA damage and susceptibility to breast cancer.

The aim of this case-control study is to explore the role of aberrations in xenobiotic metabolism in inducing oxidative DNA damage and altering the susceptibility to breast cancer. Cytochrome P4501A1 (CYP1A1) m1 (OR: 1.41, 95% CI 1.08-1.84), CYP1A1 m4 (OR: 5.13, 95% CI 2.68-9.81), Catecholamine-O-methyl transferase (COMT) H108L (OR: 1.49, 95% CI 1.16-1.92), and glutathione S-transferase (GST) T1 null (OR: 1.68, 95% CI 1.09-2.59) variants showed association with breast cancer risk. Reduced folate carrier 1 (RFC1) 80A/CYP1A1 m1/CYP1A1 m4 and RFC1 80A/thymidylate synthase (TYMS) 5'-UTR 2R/methionine synthase (MTR) 2756G/COMT 108L genetic combinations were found to inflate breast cancer risk under the conditions of low dietary folate (345 ± 110 vs. 379 ± 139 µg/day) and low plasma folate (6.81 ± 1.25 vs. 7.09 ± 1.26 ng/ml) by increasing plasma 8-oxo-2'-deoxyguanosine (8-oxodG). This increase in 8-oxodG is attributed to low methionine (49.38 ± 23.74 vs. 53.90 ± 23.85 µmol/l); low glutathione (378 ± 242 vs. 501 ± 126 µmol/l) and GSTT1 null variant; and hypermethylation of CpG island of extracellular-superoxide dismutase (EC-SOD) (92.78 ± 11.49 vs. 80.45 ± 9.86%), which impair O-methylation of catechol estrogens to methoxy estrogens, conjugation of glutathione to semiquinones/quinones and free radical scavenging respectively. Our results suggest cross-talk between one-carbon metabolism and xenobiotic metabolism influencing oxidative DNA damage and susceptibility to breast cancer.

Epistatic interactions between loci of one-carbon metabolism modulate susceptibility to breast cancer.

In view of growing body of evidence substantiating the role of aberrations in one-carbon metabolism in the pathophysiology of breast cancer and lack of studies on gene-gene interactions, we investigated the role of dietary micronutrients and eight functional polymorphisms of one-carbon metabolism in modulating the breast cancer risk in 244 case-control pairs of Indian women and explored possible gene-gene interactions using Multifactor dimensionality reduction analysis (MDR). Dietary micronutrient status was assessed using the validated Food Frequency Questionnaire. Genotyping was done for glutamate carboxypeptidase II (GCPII) C1561T, reduced folate carrier (RFC)1 G80A, cytosolic serine hydroxymethyltransferase (cSHMT) C1420T, thymidylate synthase (TYMS) 5'-UTR tandem repeat, TYMS 3'-UTR ins6/del6, methylenetetrahydrofolate reductase (MTHFR) C677T, methyltetrahydrofolate-homocysteine methyltransferase (MTR) A2756G, methyltetrahydrofolate-homocysteine methyltransferase reductase (MTRR) A66G polymorphisms by using the PCR-RFLP/AFLP methods. Low dietary folate intake (P < 0.001), RFC1 G80A (OR: 1.38, 95% CI 1.06-1.81) and MTHFR C677T (OR: 1.74 (1.11-2.73) were independently associated with the breast cancer risk whereas cSHMT C1420T conferred protection (OR: 0.72, 95% CI 0.55-0.94). MDR analysis demonstrated a significant tri-variate interaction among RFC1 80, MTHFR 677 and TYMS 5'-UTR loci (P (trend) < 0.02) with high-risk genotype combination showing inflated risk for breast cancer (OR 4.65, 95% CI 1.77-12.24). To conclude, dietary as well as genetic factors were found to influence susceptibility to breast cancer. Further, the current study highlighted the importance of multi-loci analyses over the single-locus analysis towards establishing the epistatic interactions between loci of one-carbon metabolism modulate susceptibility to the breast cancer.

Aberrations in one-carbon metabolism induce oxidative DNA damage in sporadic breast cancer.

The authors investigated the role of dietary micronutrients and eight functional polymorphisms of one-carbon metabolism in modulating oxidative stress in sporadic breast cancer. PCR-restriction fragment length polymorphism (RFLP) and PCR-amplified fragment length polymorphism (AFLP) methods were used for genetic analysis in 222 sporadic breast cancer cases and 235 controls. Standardized food frequency questionnaire was used for dietary micronutrient assessment. 8-oxo-2'-deoxyguanosine (8-oxodG), folate, and estradiol were estimated using commercial ELISA kits. Reverse-phase HPLC coupled with fluorescence detector was used for plasma homocysteine analysis. Total glutathione was estimated using Ellman's method. Reduced folate carrier 1 (RFC1) G80A and methylenetetrahydrofolate reductase (MTHFR) C677T were associated with risks of 1.34 (95% CI 1.01-1.79)- and 1.84 (95% CI 1.14-3.00)-folds, respectively, for sporadic breast cancer while cytosolic serine hydroxymethyl transferase (cSHMT) C1420T was associated with reduced risk (OR 0.71, 95% CI 0.53-0.94). Significant increase in plasma 8-oxo-2'-deoxyguanosine (P < 0.004) and homocysteine (P < 0.0001); and significant decrease in total glutathione (P < 0.01) and dietary folate (P = 0.006) was observed in cases than in controls. Oxidative DNA damage showed direct association with menopause (P = 0.02), RFC1 G80A (P < 0.05) and homocysteine (P < 0.0001); and inverse association with dietary folate (P < 0.0001), plasma folate (P < 0.0001), cSHMT C1420T (P < 0.05) and glutathione (P < 0.001). To conclude, the aberrations in one-carbon metabolism induce oxidative stress in sporadic breast cancer either by affecting the folate pool or by impairing remethylation.