Background
Nuclear receptor families play a pivotal role in regulating genes involved in drug metabolism and disposition. Pregnane X receptor (PXR, also termed SXR, PAR, and
NR1I2 as its gene name) is a crucial regulator of various phase I and phase II drug metabolizing enzymes and drug transporters. PXR is expressed in liver, small intestine and other organs. PXR, with a number of therapeutic drugs and other xenobiotics as its ligands, dimerizes with retinoid X receptor α (RXRα). The ligand-PXR-RXRα complex binds to promoter and enhancer elements located upstream of cytochrome P450s (CYPs) 3A and 2C family members, UDP-glucuronosyltransferases (UGTs), sulfotransferases (SULTs), glutathione
S-transferases (GSTs), and ATP binding cassette (ABC) drug transporters (reviewed in [
1‐
3]).
Wide inter-individual variability has been documented in the expression of hepatic CYP3A4 with respect to basal and PXR-inducible activities. The genetic variability of PXR and CYP3A4 is not sufficiently frequent to explain the apparent inter-individual variability [
4]. The inter-individual variability in basal hepatic CYP3A4 expression may include variability in PXR expression, as PXR is activated by endogenous steroid hormones and bile acids. Inter-individual variability of CYP3A4 expression is also observed in human intestine. Interestingly, a report using paired tissue samples of livers and small intestines indicated no observed correlation between the hepatic and small intestine CYP3A4 expression levels [
5]. Another research group reported that a majority of CYP3A4 resided in the proximal region of the small intestine, and that the CYP3A4 protein levels decreased dramatically in the distal small intestine [
6]. These observations suggest the CYP3A4 is regulated through tissue specific epigenetic regulation in normal tissue. The aim of the present study is to find possible and not yet fully elucidated mechanisms which regulate heterogeneous basal PXR and CYP3A4 expression and activity in cancerous tissues as well. Indeed, several studies previously found a fraction of genes that exhibited inter-individual differences in transcript levels associated with DNA methylation status [
7,
8].
DNA methylation of the CpG-rich sequence around exon 3 of the
PXR gene is involved in the epigenetic regulation of PXR in human neuroblastoma [
9]. However, epigenetic regulation of PXR and CYP3A4 in human gut is poorly understood. In order to determine whether epigenetic mechanisms function in PXR and CYP3A4 regulation and intestinal metabolism, we examined DNA methylation and mRNA expression of several candidate genes on the PXR/CYP3A4 regulatory pathway in human colon cancer cell lines and tissues.
Discussion
In the present study, 6 colon cancer cell lines showed heterogeneous mRNA expression profiles and were able to be classified into two groups with respect to their basal levels of the PXR/CYP3A4 transcripts (high expression cells, LS180 and LoVo; low expression cells Caco-2, HT29, HCT116, and SW48). These results are consistent with previous studies, in which LS180 and Caco-2 cells were characterized as PXR-sufficient and PXR-deficient cells, respectively [
16,
17].
Genetic polymorphisms in the regions that regulate transcription are often a major cause of inter-individual variability in the levels of transcripts. However, such polymorphisms have not been frequently observed in the human
PXR or
CYP3A4 genes, implying that certain epigenetic mechanisms are involved in the regulation of PXR and CYP3A4 expression. We found that the CpG-rich sequence within the
PXR promoter region is methylated to different levels in high and low expression cells. Importantly, the magnitude of this promoter methylation was inversely associated with the levels of PXR and CYP3A4 expression. Furthermore, the levels of the PXR and CYP3A4 transcripts in low expression cells were mostly restored when DNA methylation was reversed by treatment with 5-aza-dC. Although this CpG-rich sequence did not strictly satisfy the criteria for a CpG island, the most affected CpG sites were located in a highly restricted region (segments 1 and 2) and these CpG sites were proximal to several putative transcription factor binding sites (such as Sp1 and hepatocyte nuclear factor 4 alpha) [
18‐
20]. Therefore,
PXR gene expression is most likely transcriptionally regulated by methylation of these promoter CpG sites.
CYP3A4 is transactivated by functional interplays with VDR-RXRα or PXR-PRMT1 [
20‐
22]. CpG-island methylation of the
VDR or
PRMT1 promoter was not detected in these cell lines and the mRNA expression of VDR was not affected by 5-aza-dC treatment. These observations imply that DNA methylation of
PXR, but not
VDR or
PRMT1, resulted in downregulation of the CYP3A4 mRNA in these colon cancer cells.
It is still uncertain whether re-expression of the CYP3A4 by 5-aza-dC treatment was due to the re-expression of some other genes than
PXR. However,
PXR must be a candidate for methylation and reduced expression of the PXR by promoter methylation, even if partially, contributes to downregulation of the CYP3A4. Indeed, several studies demonstrated that selective downregulation of the PXR by siRNA reduces the basal level of the CYP3A4 transcripts in a dose-dependent manner [
23].
CpG islands in the exon 3 region were fully methylated throughout the cancer cell lines and most cancer tissues. Even after treatment of the cell lines with 5-aza-dC, no increase in the PXR mRNA levels was observed in the high-PXR expressing cell lines, LS180 and LoVo. This strongly suggests that in human colon cancer cells, the methylated CpG islands in the exon 3 play a much less role in the epigenetic regulation of PXR, instead, promoter methylation plays a pivotal role in its regulation. In contrast, Misawa
et al. previously demonstrated a distinct methylation profile of neuroblastoma cells, in which mRNA expression of the PXR splicing variant (exon 1a-2) was specifically regulated by the methylation of the exon 3 region rather than promoter methylation [
9]. We found no marked difference in the levels of the PXR (exon 5-6) and PXR (exon 1a-2) transcripts in the colon cancer cells. Therefore, a tissue-specific DNA methylation profile is most likely involved in the transcriptional regulation of the
PXR gene.
DNA methylation of the
PXR promoter was detected in only 1 of the 18 colorectal cancer tissue samples. The results reflect the genuine DNA methylation status, because we examined pure cancerous and normal epithelia using crypt isolation and directly compared the DNA methylation status between paired epithelia. Therefore, a low level of
PXR promoter methylation, which was observed in the high expression cells, appears to be a common feature of colorectal cancers. We also demonstrated that the level of
PXR promoter methylation is decreased during carcinogenesis, since paired adjacent normal tissues mostly showed higher levels of
PXR promoter methylation. We could not directly compare DNA methylation status with the PXR mRNA expression, because crypt isolation provided ethanol-fixed epithelia and it was difficult to obtain fresh mRNA samples. However, most cancer tissues exhibited a pattern of promoter methylation quite similar to that observed in cultured cells with high expression (LS180 and LoVo) (Figures
5 and
6c). Therefore, the association between promoter methylation and transcriptional silencing of the
PXR gene is most likely applicable to primary colorectal cancers. As observed in the colon cancer cell lines, the decreased level of
PXR promoter methylation most likely led to increased expression of PXR mRNA in the colorectal cancer tissues. These results are consistent with a recent study that showed strong expression of PXR mRNA in colon cancers, with great variability [
24]. In contrast, Ouyang
et al. found that PXR expression was lost or greatly diminished in many colon cancers using histochemical analysis [
25]. Although the role of the altered PXR expression in colorectal carcinogenesis remains to be clarified, Zhou
et al. demonstrated that PXR plays an antiapoptotic role in colon carcinogenesis by induction of multiple antiapoptotic genes [
26].
We cannot rule out the possibility that alterations of the PXR methylation levels play direct roles in tumorigenesis, because certain oncogenes or tumor suppressor genes may be trascriptionally regulated by PXR. Partial methylation of the PXR observed in adjacent normal mucosa may be associated with "field defect" for carcinogenesis. However, numerous studies have demonstrated that ligand-binding activation or siRNA-mediated silencing of the PXR can affect the activity of metabolic enzymes including CYP3A4, without changes in the cell proliferation capacity. Therefore, we think that altered level of the PXR methylation does not provide a selective growth advantage during colorectal cancer progression.
Interestingly, overexpression of PXR in the colorectal cancer tissue samples was correlated with an increase in UDP glucuronosyl transferases UGT1A1, UGT1A9 and UGT1A10, and led to a marked chemoresistance to the active metabolite of irinotecan (CPT-11) [
24]. In addition, CYP3A4 and p-glycoprotein, which are transcriptionally activated by PXR, play important roles in intestinal first-pass metabolism and determine a drug's bioavailability. We hypothesized that PXR may play a key role in the colon cancer cell response to anticancer drugs by modulating expression of drug metabolizing enzymes and transporters including UGT1A, CYP3A4 and p-glycoprotein. Therefore, DNA methylation of the
PXR promoter might be a good predictor of chemotherapy outcome and toxicity in colorectal cancers.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
WH designed this study and carried out the cell culture, molecular genetic studies and drafted the manuscript. GT and JT participated in the data analysis. TS performed crypt isolation and pathological diagnosis. KO and GW performed the surgeries and obtained informed consent from the patients. SO participated in the design of this study and helped to draft the manuscript. All authors read and approved the final manuscript.