Antioxidants such as SFN have
Antioxidants such as SFN have been proposed to be chemopreventive agents to inhibit, delay or reverse the development of cancer (Bayat Mokhtari et al., 2018). SFN was shown in breast epithelial cell culture to suppress oxidative metabolism of E2/E1 and thus protect against estrogen-mediated DNA damage through alteration of E2 metabolizing enzymes such as COMT, NQO1 and GSTA1 (Yang et al., 2013). We showed that E2 or GEN suppressed NQO1 and GSTA1 (Supplementary Fig. 5) suggesting the existence of oxidative stress in response to estrogens. In contrast, SFN induced NQO1 (Nrf2 responsive gene) in MCF-7 96365 and reversed the E2- or GEN-induced NQO1 suppression (Supplementary Fig. 5A). Yang et al. suggested that the effect of SFN on COMT activity may not be directly regulated by Nrf2 signaling because SFN did not affect the mRNA level of COMT. Similar to Yang et al., our results indicated that SFN alone did not significantly change COMT mRNA levels or CpG methylation of COMT promoters in MCF-7 cells. However, SFN reversed the epigenetic changes induced by E2 and GEN at the MB-COMT promoter resulting in the increased COMT expression. Our data suggest that, in addition to Nrf2-modulated antioxidant defense, SFN might modulate E2- or GEN-induced binding of epigenetic modulators (DNMT3B, MBD2 and HDAC1) to COMT promoter which may eventually result in modifications of COMT-mediated estrogen metabolism. It has been reported that SFN can modify the activities of DNMTs and HDACs in cancer cells (Tortorella et al., 2015). Thus, SFN may provide opportunities for cancer prevention by regulating the components of epigenetic machinery, especially DNMTs and MBDs (Kar et al., 2014) and epigenetic modifications of suppressor/oncogene transcription (Hsu et al., 2011). Our finding indicates that SFN may serve as epigenetic modifier to reverse the E2 or GEN-induced DNA methylation and histone acetylation at the MB-COMT promoter, may ultimately modulate the E2/E1 metabolism pathway in MCF-7 cells. We suggest that epigenetic variation at COMT may play an equally important role as genetic variation in assessing the risk of human BCa. An abundance of epidemiology studies reported the association between genetic variations at COMT and BCa risk (Singh et al., 2005) but some produced conflicting results (Qin et al., 2012; Spurdle et al., 2017; Horvath, 2017; Movassagh et al., 2017). In addition, several studies suggesting environmental factors such as environmental chemicals with hormonal activity (Kocabaş et al., 2002; Cerne et al., 2011; Kallionpää et al., 2017), smoking (Saintot et al., 2003) and diets or consumption of nutritional supplements (Wang et al., 2011; Low et al., 2005) produce mixed results on the relation of genetic variations of COMT to human BCa risk. This raises possibilities for other mechanisms underlying the interaction between gene and environment. Swift-Scanlan et al. were the first to report site-specific DNA methylation changes at promoter regions for S-COMT and MB-COMT and their correlation to COMT gene expression in human BCa cells (Swift-Scanlan et al., 2014). In addition, they reported the association of environmental factors (e.g. socioeconomic status and alcohol use), with loci-specific DNA methylation and a functional SNP (Val158 Met of MB-COMT) in saliva-derived DNA using a cohort of 48 healthy participants. In the present study, we provide additional evidence that estrogens and phytoestrogens modulate COMT methylation in MCF-7 cells in a CpG site-specific manner. We found estrogens did not induce differential methylation at the CpG sites (located in S-COMT promoter and MB-COMT proximal promoter, Fig. 1) reported in Swit-Scanlan's study but at another CpG cluster flanking the TSS of MB-COMT distal promoter (MB-COMT-PR2, illustrated in Fig. 1A). Our data suggests that estrogens or antioxidants could alter COMT methylation at particular loci that may affect the expression of COMT transcripts. Future studies on epigenetic variations within the COMT gene may be beneficial for risk assessment for human BCa.