n-6 Linoleic Acid Induces Epigenetics Alterations Associated with Colonic Inflammation and Cancer

The farnesoid-X-receptor (FXR) protects against inflammation and cancer of the colon through maintenance of intestinal bile acid (BA) homeostasis. Conversely, higher levels of BA and cyclooxygenase-2 (COX-2) are risk factors for inflammation and cancer of the colon. In the United States, n-6 linoleic acid (LA) is the most commonly used dietary vegetable fat. Metabolism of n-6 fatty acids has been linked to a higher risk of intestinal cancer. The objectives of this study were to investigate in colonic mucosa the effects of a high-fat diet rich in LA (n-6HFD) on CpG methylation of Fxr and prostaglandin-endoperoxide synthase-2 (Ptsg-2) genes, and the impact on the expression of tumor suppressor adenomatous polyposis Coli (Apc) and proliferative cyclin D1 (Ccnd1) genes. Weaned C57BL/6J male mice were fed for 6 weeks either an n-6HFD containing 44% energy (44%E) from 22% safflower oil (SO, 76% LA by weight) or a 13% energy (13%E) control diet (Control) from SO (5% by weight). Mice fed the n-6HFD had reduced (60%) Fxr promoter CpG methylation and increased (~50%) Fxr mRNA. The expression of FXR-target ileal bile acid-binding protein (Ibabp), small heterodimer protein (Shp), and anti-inflammatory peroxisome proliferator-activated-γ1 genes was increased. The n-6HFD reduced Ptgs-2 CpG methylation, increased the expression of Cox-2, and increased Apc CpG methylation in colonic mucosa. Accordingly, reduced expression of Apc was coupled to accumulation of c-JUN and Ccnd1, respectively cofactor and gene targets for the β-catenin/Wnt signaling pathway. Finally, the n-6HFD reduced the expression of histone deacetylase-1 while favoring the accumulation of acetylated histone 3. We conclude that an n-6HFD epigenetically modifies Fxr, leading to the activation of downstream factors that participate in BA homeostasis. However, epigenetic activation of Ptsg-2 coupled with silencing of Apc and accumulation of C-JUN and Ccnd1 may increase the risk of inflammation and cancer of the colon.


Introduction
A diet high in n-6 fatty acids (n-6HFD) is a recognized risk factor for inflammatory bowel diseases (IBDs) [1,2]. Individuals with diets high in linoleic acid (LA), an n-6 fatty acid found predominantly in plant oils (e.g., soybean, corn, safflower), are at increased risk for one common form of IBD, ulcerative colitis [3,4]. In animal studies, the feeding of a vegetable oil (corn oil) rich in LA increases cell proliferation in the colon [5,6]. An n-6HFD induces the biliary excretion of bile acids (BA), increases fecal levels of the secondary BA deoxycholic acid (DCA), and chemically induced colon cancer a Values are calculated from ingredient analysis or manufacturer data (Harlan Laboratories). n-6HFD = diet high in n-6 fatty acids; TBHQ = Tertiary butylhydroquinone.

Real-Time PCR
Colonic mucosa were scraped from the proximal colon as described previously [27], and total RNA was prepared using the Quick-RNA Miniprep Kit (Zymo Research, Irvine, CA, USA). The RNA concentration was assessed with the Nanodrop1000 Spectrophotomoter (Thermo Scientific, Waltham, MA, USA). Total RNA (500 ng) was used to prepare cDNA using qScript cDNA SuperMix (Quanta Biosciences, Gaithersburg, MD, USA). For PCR amplification of cDNA, we used the PerfeCTa SYBR Green Fast Mix, Rox (Quanta Biosciences). PCR reactions were performed at a final volume of 10 µL comprising 5 µL of SYBR Green Mix, forward and reverse primers (1 µL each, 10 nM), 2 µL nuclease-free water, and 1 µL of cDNA. Glyceraldehyde 3-phosphate dehydrogenase (Gapdh) amplification was used for normalization of mRNA expression. The mouse primers (Sigma-Aldrich) used for RT-PCR are shown in Table 2.

Genomic DNA Methylation
The procedure for measurement of promoter methylation was described previously [26]. In short, genomic DNA was prepared from 10-15 mg of proximal colon mucosa using the DNeasy Blood and Tissue Kit (Qiagen, Valencia, CA, USA). The DNA (1 µg) was modified via bisulfite treatment using the EpiTect Bisulfite Modification Kit (Qiagen) followed by PCR amplification using 1 cycle at 94 • C (1 min); 33-35 cycles at 94 • C (30 s), 59 • C (30 s), and 72 • C (1 min); and 1 cycle at 72 • C (5 min). Amplification was performed at a volume of 25 µL comprising bisulfite modified genomic DNA (50 ng), 0.4 µL JumpStart Taq DNA polymerase (Sigma-Aldrich), 2.5 µL 10X PCR buffer, 3.5 µL 25 mM MgCl 2 (final concentration 3.5 mM), 0.5 µL 10 mM dNTP mix (final concentration 200 µM), forward and reverse primers (1 µL each), and water to a final volume of 25 µL. The PCR products were analyzed on 2% agarose gels and examined by ethidium bromide staining. The size and authenticity of the PCR products were confirmed by a molecular weight analysis and DNA sequencing. The primers (Sigma-Aldrich) used for DNA methylation studies are shown in Table 2.

Statistical Analysis
Densitometry analyses of PCR products were carried out using the Kodak ID Image Analysis Software (Eastman Kodak Company, Rochester, NY, USA). Expression and promoter methylation Nutrients 2019, 11, 171 5 of 14 data were analyzed by ANOVA. Post-hoc multiple comparisons among all means were performed using Tukey's test after the main effects and interactions were confirmed to be significant at p ≤ 0.05. Data are presented as means ± standard error of the mean (SEM) and statistical differences highlighted with asterisks or different letters for multiple comparisons.

n-6HFD Induces Fxr Gene Promoter Hypomethylation and Expression in Proximal Colonic Mucosa
Compared to animals assigned to the control diet (5%E), mice fed the n-6HFD (44%E) had higher weight gain (~10%) starting at week 4 through week 6 of the experiment ( Figure 1A). of the PCR products were confirmed by a molecular weight analysis and DNA sequencing. The primers (Sigma-Aldrich) used for DNA methylation studies are shown in Table 2.

Statistical Analysis
Densitometry analyses of PCR products were carried out using the Kodak ID Image Analysis Software (Eastman Kodak Company, Rochester, NY, USA). Expression and promoter methylation data were analyzed by ANOVA. Post-hoc multiple comparisons among all means were performed using Tukey's test after the main effects and interactions were confirmed to be significant at p ≤ 0.05. Data are presented as means ± standard error of the mean (SEM) and statistical differences highlighted with asterisks or different letters for multiple comparisons.
We examined the effects of the n-6HFD on Fxr mRNA expression in proximal colonic mucosa and found that Fxr transcripts were increased by~40% compared to animals fed the control diet ( Figure 1B). As a control for changes in regulation of lipid metabolism, we monitored the expression of Pparγ1, which was increased by~80%. The upregulation of Fxr and Pparγ1 were consistent with findings of previous reports documenting higher weight gain and increased levels of FXR in the distal small intestine in response to a high-fat diet (HFD) [28], and FXR-mediated transcriptional activation of Pparγ [29]. We then turned our analysis to the intestinal gene targets for FXR, Ibabp, and Shp, whose expression was increased respectively 1.5-and 3.2-fold compared to the control ( Figure 1C,D). Based on these results, we examined the impact of the n-6HFD on the CpG methylation status of Fxr. We focused on a 470 bp promoter region (−54/+416) comprising 13 CpG sites and flanking the transcription start site (+1) of exon-3. This promoter generates FXRα3/4 transcripts (Figure 2A), which are expressed at higher levels than the FXRα1/2 mRNA variants in the intestine [20].
of Pparγ [29]. We then turned our analysis to the intestinal gene targets for FXR, Ibabp, and Shp, whose expression was increased respectively 1.5-and 3.2-fold compared to the control ( Figure 1C and 1D). Based on these results, we examined the impact of the n-6HFD on the CpG methylation status of Fxr. We focused on a 470 bp promoter region (-54/+416) comprising 13 CpG sites and flanking the transcription start site (+1) of exon-3. This promoter generates FXRα3/4 transcripts (Figure 2A), which are expressed at higher levels than the FXRα1/2 mRNA variants in the intestine [20]. Amplification of CpG-methylated Fxr promoter amplicons from bisulfonated genomic colonic mucosal DNA was conducted in the linear range as reported previously [26]. The Fxr gene CpG methylation was reduced by ~60% ( Figure 2B and 2C) in mice fed the n-6HFD, in agreement with a previous experimental report documenting increased intestinal FXR expression and BA levels as a result of feeding an HFD [30].

n-6HFD Induces Ptsg-2 Gene Promoter Hypomethylation and Expression in Proximal Colonic Mucosa
Cyclooxygenase-2 is a recognized risk factor in intestinal inflammation and CRC [31,32]. Levels of proximal colonic mucosal Cox-2 mRNA were increased ~1.5-fold in mice fed the n-6HFD ( Figure  3A). These results were consistent with other studies showing increased COX-2 expression associated with intestinal inflammation induced by an n-6HFD [33] and carcinogenesis of the colon [34,35]. Turning to the epigenetic regulation of the Ptsg-2 promoter ( Figure 3B), feeding of the n-6HFD to C57BL/6J mice reduced CpG methylation on average by ~35% ( Figure 3B and 3C). Amplification of CpG-methylated Fxr promoter amplicons from bisulfonated genomic colonic mucosal DNA was conducted in the linear range as reported previously [26]. The Fxr gene CpG methylation was reduced by~60% ( Figure 2B,C) in mice fed the n-6HFD, in agreement with a previous experimental report documenting increased intestinal FXR expression and BA levels as a result of feeding an HFD [30].

n-6HFD Induces Ptsg-2 Gene Promoter Hypomethylation and Expression in Proximal Colonic Mucosa
Cyclooxygenase-2 is a recognized risk factor in intestinal inflammation and CRC [31,32]. Levels of proximal colonic mucosal Cox-2 mRNA were increased~1.5-fold in mice fed the n-6HFD ( Figure 3A). These results were consistent with other studies showing increased COX-2 expression associated with intestinal inflammation induced by an n-6HFD [33] and carcinogenesis of the colon [34,35]. Turning to the epigenetic regulation of the Ptsg-2 promoter ( Figure 3B), feeding of the n-6HFD to C57BL/6J mice reduced CpG methylation on average by~35% ( Figure 3B,C).
The gain of Cox-2 mRNA levels and reduction in Ptsg-2 CpG methylation were paralleled by an accumulation of COX-2 protein ( Figure 4A). As a positive control for activation of COX-2 expression by the n-6HFD, we performed Western blotting of cell lysates from nontumor human FHC cells treated in culture with LA ( Figure 4B), which induced COX-2 expression compared to control DMEM.
Taken together, these data suggest that exposure to an n-6HFD rich in LA elicits an inflammatory response characterized by CpG hypomethylation of Ptsg-2 and increased expression of COX-2. The gain of Cox-2 mRNA levels and reduction in Ptsg-2 CpG methylation were paralleled by an accumulation of COX-2 protein ( Figure 4A). As a positive control for activation of COX-2 expression by the n-6HFD, we performed Western blotting of cell lysates from nontumor human FHC cells treated in culture with LA ( Figure 4B), which induced COX-2 expression compared to control DMEM. Taken together, these data suggest that exposure to an n-6HFD rich in LA elicits an inflammatory response characterized by CpG hypomethylation of Ptsg-2 and increased expression of COX-2.  The gain of Cox-2 mRNA levels and reduction in Ptsg-2 CpG methylation were paralleled by an accumulation of COX-2 protein ( Figure 4A). As a positive control for activation of COX-2 expression by the n-6HFD, we performed Western blotting of cell lysates from nontumor human FHC cells treated in culture with LA ( Figure 4B), which induced COX-2 expression compared to control DMEM. Taken together, these data suggest that exposure to an n-6HFD rich in LA elicits an inflammatory response characterized by CpG hypomethylation of Ptsg-2 and increased expression of COX-2.

n-6HFD Lowers the Expression of Apc and Activates Downstream Targets for the β-Catenin/Wnt Pathway
The loss of APC along with the activation of the KRAS oncogene contribute to nuclear localization of β-catenin, a component of the Wnt signaling pathway, which promotes the expression of genes involved in proliferation [36][37][38]. An association between increased nuclear levels of β-catenin and COX-2 expression is seen in human and murine colon cancer cells with defective Apc [24,39]. We noted that the n-6HFD increased Apc CpG methylation ( Figure 5A) while reducing the Apc transcript levels ( Figure 5B) in colonic mucosa. Because the Ccnd1 gene is a direct transcriptional target for activation by β-catenin transcription complexes [40,41], we measured the levels of Ccnd1 transcripts and found increased expression associated with the n-6HFD ( Figure 5C). The upregulation of c-JUN, whose gene is also a direct target for transcriptional activation by β-catenin [42] and contributes to the activation of COX-2 [43], provided a positive control for activation of the Wnt signaling pathway under conditions of diminished Apc expression ( Figure 5D). Finally, we noted that the n-6HFD reduced the colonic levels of histone deacetylase-1 (HDAC-1) while increasing the expression of acetylated histone 3 (ACH3), suggesting effects on epigenetic regulation.
The loss of APC along with the activation of the KRAS oncogene contribute to nuclear localization of β-catenin, a component of the Wnt signaling pathway, which promotes the expression of genes involved in proliferation [36][37][38]. An association between increased nuclear levels of βcatenin and COX-2 expression is seen in human and murine colon cancer cells with defective Apc [24,39]. We noted that the n-6HFD increased Apc CpG methylation ( Figure 5A) while reducing the Apc transcript levels ( Figure 5B) in colonic mucosa. Because the Ccnd1 gene is a direct transcriptional target for activation by β-catenin transcription complexes [40,41], we measured the levels of Ccnd1 transcripts and found increased expression associated with the n-6HFD ( Figure 5C). The upregulation of c-JUN, whose gene is also a direct target for transcriptional activation by β-catenin [42] and contributes to the activation of COX-2 [43], provided a positive control for activation of the Wnt signaling pathway under conditions of diminished Apc expression ( Figure 5D). Finally, we noted that the n-6HFD reduced the colonic levels of histone deacetylase-1 (HDAC-1) while increasing the expression of acetylated histone 3 (ACH3), suggesting effects on epigenetic regulation. In summary, these results suggest that an n-6HFD rich in LA epigenetically alters gene expression in colonic mucosa, leading to an accumulation of proinflammatory and proliferative factors associated with a higher risk of CRC.

Discussion
In this study, we first address whether or not changes in Fxr promoter CpG methylation contribute to the regulation of Fxr expression in the colonic mucosa in response to an n-6HFD, which mimics dietary fat exposure known to increase the risk of intestinal inflammation in humans [3,4]. The total energy as fat (44%E) of the n-6HFD approaches the one used in previous mouse models that linked inflammation to the development of CRC [44,45]. We show that feeding of the n-6HFD to In summary, these results suggest that an n-6HFD rich in LA epigenetically alters gene expression in colonic mucosa, leading to an accumulation of proinflammatory and proliferative factors associated with a higher risk of CRC.

Discussion
In this study, we first address whether or not changes in Fxr promoter CpG methylation contribute to the regulation of Fxr expression in the colonic mucosa in response to an n-6HFD, which mimics dietary fat exposure known to increase the risk of intestinal inflammation in humans [3,4]. The total energy as fat (44%E) of the n-6HFD approaches the one used in previous mouse models that linked inflammation to the development of CRC [44,45]. We show that feeding of the n-6HFD to C57BL/6 mice reduces Fxr promoter CpG methylation while increasing the expression of Fxr and that of the FXR-target genes Ibabp and Shp. These results suggest that conditions that have been demonstrated to increase the intestinal levels of BA [9,10,30] direct the removal of repressive methylation marks at the Fxr gene to augment FXR expression and maintain BA homeostasis. We also show that the n-6HFD induces Pparγ1 expression in the colonic mucosa. The upregulation of Pparγ1 by the n-6HFD is consistent with the results of previous investigations documenting transactivation of the Pparγ gene by FXR through its physical interaction with an FXR response element [29], and genetic evidence that PPARγ expression is compromised in FXR-deficient (Fxr −/− ) models [46]. The stimulatory action of FXR on the PPARγ gene is believed to reduce the steady-state levels of β-catenin in intestinal cells with normal Apc [47,48]. In addition, we report that the n-6HFD increases COX-2 expression associated with CpG hypomethylation at the Ptsg-2 gene. In esophageal and gastrointestinal cells, BA increase the expression of COX-2 through various signals, including the p38-mitogen-activated protein kinase (p38MAPK) pathway [49] and the transcription factors nuclear factor kappa-light-chain-enhancer of activated B (NFkB) [50] and activator protein-1 (AP-1) [49,51]. Interestingly, SHP, a transcription factor whose expression is induced by FXR, interacts physically with and functions as a positive coactivator of NFkB [52], and is required for caudal-related homeobox-1 gene (CDX1)-mediated activation of COX-2 [53]. Therefore, the activation of FXR associated with chronic exposure to higher levels of BA due to the n-6HFD may have the adverse effect of triggering an inflammatory response supported by SHP, which in turn may override the anti-inflammatory activities of FXR on NFkB [54]. Moreover, the increased COX-2 expression is expected to support the enzymatic conversion of arachidonic acid (AA) to prostaglandin E2 (PGE2), which, through its EP2 and EP4 receptors, may activate the phosphoinositide 3-kinase (PI3K)/protein kinase B (Akt)-dependent upregulation of β-catenin. This may lead to further activation of COX-2 expression [55,56]. Thus, the long-term exposure to an n-6HFD rich in LA may trigger an intestinal feed-forward loop through which BA amplifies COX-2 transcription mediated by the transcription factors β-catenin, SHP, and AP-1. In support of this hypothesis, we also show that the n-6HFD increases the expression of c-JUN, a key component of the AP-1 transcription factor [57]. Based on this cumulative evidence, we propose that accumulation of SHP and c-JUN coupled to CpG hypomethylation at the Ptsg-2 gene may serve as a biomarker of increased risk of intestinal inflammation and tumorigenesis related to chronic exposure to an n-6HFD [58,59]. In accord with this suggestion, changes in DNA methylation profiles of genes involved in lipid metabolism and inflammation have been proposed as candidate biomarkers of CRC [14].
A second question addressed by this study pertains to whether or not feeding an n-6HFD epigenetically alters the expression of Apc and β-catenin signaling in the colonic mucosa. The rationale for this objective stems from earlier experimental evidence showing that colonic inflammation due to an HFD is paralleled by an increase in the gastro-intestinal levels of β-catenin [60][61][62] and the activation by the β-catenin/transcription factor (TCF)/lymphoid enhancer binding factor (LEF) transcription complex of genes (e.g., Ccnd1) involved in intestinal tumorigenesis [45,63,64]. We show that the CpG hypermethylation of the Apc gene in colonic mucosa of mice fed the n-6HFD correlates with reduced Apc expression, and an accumulation of Ccnd1 and c-JUN, which are established transcription targets for the β-catenin/TC/LEF transcription complex [40,41]. The latter is constitutively active in Apc-deficient colon carcinoma [65][66][67]. One question raised by our data relates to the mechanisms contributing to the silencing of Apc in colonic mucosa of mice fed the n-6HFD. The enrichment in c-JUN and downregulation of Apc expression correlate in this study with a reduction in HDAC-1 and a gain in AcH3. These changes underscore the impact of dietary conditions that promote obesity, such as the weight gain observed in this study, on epigenetic regulation in the colonic mucosa [68]. Obesity induces DNA hypermethylation of Apc in the small intestine and changes the epigenetic landscape in the colonic epithelium [69], thus enhancing proliferation mediated by the Wnt/β-catenin pathway [70][71][72]. Studies in patients with gastric adenomas [73] and CRC cell lines [74] have confirmed the existence of a positive association between an accumulation of β-catenin and hypermethylation of the Apc gene. Recently,~69% of CRCs were reported to harbor hypermethylated Apc [75], which correlated positively with tumor size and lymph node metastasis. To our knowledge, this is the first report showing a direct effect of an n-6HFD on epigenetic disregulation of Apc and Pstg-2 via CpG methylation in the colonic mucosa. These epigenetic modifications may be useful in monitoring the susceptibility to CRC associated with the adoption of an n-6HFD and overweight conditions.

Conclusions
In summary, the current study provides evidence that an n-6HFD contributes epigenetically to the activation of FXR expression via CpG demethylation to support the expression of genes whose products (i.e., SHP, IBABP) participate in the regulation of BA homeostasis through the enterohepatic circulation ( Figure 6). Our data also suggest that chronic exposure to an n-6HFD downregulates APC expression via CpG hypermethylation and this associates with increased expression of COX-2 via PTSG-2 CpG hypomethylation and accumulation of C-JUN and CCND1, thus increasing the risk of inflammation and cancer of the colon. disregulation of Apc and Pstg-2 via CpG methylation in the colonic mucosa. These epigenetic modifications may be useful in monitoring the susceptibility to CRC associated with the adoption of an n-6HFD and overweight conditions.

Conclusions
In summary, the current study provides evidence that an n-6HFD contributes epigenetically to the activation of FXR expression via CpG demethylation to support the expression of genes whose products (i.e., SHP, IBABP) participate in the regulation of BA homeostasis through the enterohepatic circulation ( Figure 6). Our data also suggest that chronic exposure to an n-6HFD downregulates APC expression via CpG hypermethylation and this associates with increased expression of COX-2 via PTSG-2 CpG hypomethylation and accumulation of C-JUN and CCND1, thus increasing the risk of inflammation and cancer of the colon. Studies are ongoing in our laboratory to clarify the impact of diets that vary in their fatty acid profile (i.e., n-6 versus n-3) on CpG methylation of the Fxr and Cox-2 genes and effects on the expression of FXR-and β-catenin target genes. Future investigators should elucidate how an n-6HFD modifies the interplay between microbiota and epigenetics of inflammation and colon cancer. For example, a diet mimicking the human Mediterranean diet, which is rich in olive oil, fruits, and vegetables, favors a microbiota composition associated with reduced carcinogenesis in intestinal cells with defective Apc [76]. Progress in these areas may help to address the question of whether the adoption of an n-6HFD associates with a higher risk of IBD and CRC and provide new strategies for epigenetic targeting through dietary interventions. Studies are ongoing in our laboratory to clarify the impact of diets that vary in their fatty acid profile (i.e., n-6 versus n-3) on CpG methylation of the Fxr and Cox-2 genes and effects on the expression of FXR-and β-catenin target genes. Future investigators should elucidate how an n-6HFD modifies the interplay between microbiota and epigenetics of inflammation and colon cancer. For example, a diet mimicking the human Mediterranean diet, which is rich in olive oil, fruits, and vegetables, favors a microbiota composition associated with reduced carcinogenesis in intestinal cells with defective Apc [76]. Progress in these areas may help to address the question of whether the adoption of an n-6HFD associates with a higher risk of IBD and CRC and provide new strategies for epigenetic targeting through dietary interventions.