In this report, we predicted several putative TF binding sites in UGT1A1 proximal promoter using a bioinformatic tool and demonstrated by EMSA that HNF1-alpha, USF1/2, and NF-Y would bind to UGT1A1 proximal promoter. The influence of these TFs upon UGT1A1 transcriptional activity was then demonstrated by transient transfection in colon adenocarcinoma cell line HT29, and solely HNF1-alpha and USF1/2 have been shown to have significant impact.
Mutations in the HNF1-alpha motif resulted in a substantial reduction of UGT1A1 promoter activity in HT29 cells. This HRE located in the human UGT1A1 promoter consists of a very well conserved half-site and a more divergent one with respect to the consensus sequence, which is very similar to the one from mouse UGT1A1 promoter and rat albumin promoter [26, 27]. Consistent with previous studies in mice, HNF1-alpha proteins failed to bind (data not shown) and trans-activate the human UGT1A1 promoter when the conserved half-site is altered (Table 1). Besides, our data indicate that the one half of the HRE palindromic sequence is sufficient for its recognition of the UGT1A1 promoter, and that HNF1-alpha is critical for UGT1A1 expression.
HNF1-alpha is well-known to be involved in regulation of several UGTs, including human UGT2B7, UGT2B17, UGT1A1, UGT1A3, UGT1A4, UGT1A8, UGT1A9, UGT1A10 and rat UGT1A7 [20, 21, 26, 28–30]. Although the role of HNF1-alpha in the regulation of UGT1A1 has already been studied , the data were limited to transient transfections of the -617/+15 UGT1A1 promoter and its HNF1-deleted construct into UGT1A1-negative HEK293 kidney cells. In here, we emphasized the importance of HNF1-alpha in the regulation of UGT1A1, and in contrast to previous observations, data were demonstrated in cells with a known glucuronidation capacity.
As observed for HNF1-alpha, mutations in URE also resulted in a drastic reduction of the promoter activity in HT29 cells, supporting for the first time, a role for this TF in the regulation of UGT1A1 promoter. Upstream stimulatory factors, USF1 and 2 are late TFs able to interact as homo- and/or heterodimers on E boxes of CACGTG sequence [31–34]. USFs are ubiquitously expressed proteins that have been described as positive or negative regulators of numerous genes, including cyclin-cdk encoding genes, tumour suppressor genes, and growth factor networks [35, 36]. To our knowledge, no interaction of USF1 or USF2 with phase II enzymes such as the UGT family members has been documented thus far.
While EMSA indicated that NF-Y might also bind UGT1A1 promoter, mutations in its binding motif did not significantly reduce the luciferase activity compared with the wild-type construct in HT29 cells, suggesting that basal promoter activity does not require direct interaction of this TF. Although informative, promoter-reporter constructs inadequately mimic the chromosomal context. It is now appreciated that chromatin-associated factors are various key determinants for specific gene expression . Accordingly, we may not rule out that NF-Y would contribute to UGT1A1 gene expression in native cells.
The observation that URE includes a CpG dinucleotide contact point, which is critical for recognition by the USF proteins, prompted us to hypothesize that a USF E-box element that contained 5-methylcytosine in the CACGTG core might affect the binding for USF1/2. EMSA using unmethylated probe resulted in the formation of an USF-UGT1A1 complex. When methylated, URE-containing oligonucleotide competed poorly for USF1/2 binding, showing that specific methylation of CpG-4 dinucleotide decrease the affinity for USF1/2. It was previously shown that methylation at the CpG site, centrally located in the E-box motif (CACpGTG), strongly inhibits formation of DNA-protein complex [38, 39] and negatively regulates gene expression. Single nucleotide polymorphisms, within the E-box core motif, also modulate gene regulation. Notably, a single G/C base transition within the USF E-box consensus of the thymidylate synthase gene, implicated in folate metabolism, prevents USF proteins from binding to their cognate sequence .
As we observed previously for UGT1A1 , data indicate that DNA methylation is one mechanism likely involved in the down-regulation of HNF1A gene expression in colon cells. DNA methyltransferase inhibitor treatment of UGT1A1-negative HCT116 colon cells restored HNF1A gene expression. 5-aza-dC-induced gene reactivation has two distinct requirements: 1) the reversal of promoter DNA hypermethylation, and 2) the presence of transcriptional activators competent for activation of the target promoter. Considering that HNF1-alpha is essential for UGT1A1 gene expression, the methylation of HNF1A gene promoter represents a second level of DNA methylation-mediated regulation, which highlights the complexity of epigenetic gene regulation. The modulation of HNF1A expression might also have impact on the regulation of other genes, notably on additional phase II enzymes including other UGTs [20, 21, 26, 28–30], glutathione transferase , and sulfotransferase .
Interestingly, the UGT1A1-associated HRE, which is free of CpG dinucleotide, is located between CpG-3 and -4, and we demonstrated that methylation of proximal CpG dinucleotides is not sufficient to significantly alter HNF1-alpha binding in vitro. However, we may not rule out the importance of DNA methylation in the binding of HNF1-alpha in vivo, because such a DNA modification induces a repressive chromatin structure, and might restrain the accessibility of HNF1-alpha to its recognition sequence in UGT1A1 promoter. However, we suggest that UGT1A1 proximal promoter methylation may directly affect transcriptional activity by suppressing the interaction of USF1/2 with its cognate sequence.
Taken together, our results reveal that both HNF1-alpha and USF1/2 could play an important role in activating the transcription from UGT1A1 promoter. The interplay between HNF1-alpha and USF1/2 has been previously shown to be implicated in the liver-specific expression of the pyruvate kinase gene, in the regulation of three human class I alcohol dehydrogenase genes and in the constitutive expression of CYP1A2 [43–45]. Considering that UGT1A1-mediated glucuronidation is the primary route of irinotecan inactivation, it was suggested that the level of UGT1A1 expression might contribute to the differential chemosensitivity of colon tumors [46–48]. In a previous report, we showed that methylation of UGT1A1 promoter may conduct to reduction of gene expression level, leading to a lower UGT1A1 glucuronidation activity. Accordingly, positive UGT1A1 methylation in tumors, and subsequent repression of UGT1A1-associated metabolic pathways would be involved in retention of active SN-38 within colon cancer cells. This could lead to higher sensitivity to irinotecan. In contrast, the presence of high levels of UGT activity and expression was identified as a characteristic associated with a resistance phenotype to SN-38 in colon cancer cells, as supported by a previous report .