Identification and validation of microRNAs directly regulating the UDP-glucuronosyltransferase 1A subfamily enzymes by a functional genomics approach

Abstract

Posttranscriptional repression of UDP-glucuronosyltransferase (UGT) 1A expression by microRNAs (miRNAs) may be an important mechanism underlying interindividual variability in drug glucuronidation. Furthermore, the UGT1A 3'-UTR shared by all UGT1A enzymes is polymorphic, containing three linked SNPs (rs10929303, rs1042640, and rs8330) that could influence miRNA binding. The aim of this study was to identify the complete complement of miRNAs that could regulate UGT1A expression through binding to the reference and/or common variant UGT1A 3'-UTR. Luciferase reporter plasmids containing either the reference or variant UGT1A 3'-UTR were screened against a 2048 human miRNA library to identify those miRNAs that decrease luciferase activity by at least 30% when co-transfected into HEK293 cells. Four novel miRNAs (miR-103b, miR-141-3p, miR-200a-3p, and miR-376b-3p) were identified that repressed both reference and variant UGT1A 3'-UTR, while two other miRNAs selectively repressed the reference (miR-1286) or variant (miR-21-3p) 3'-UTR. Deletion and mutagenesis studies confirmed the binding site location for each miRNA. rs8330 disrupted miR-1286 binding to the reference UGT1A 3'-UTR, while rs10929303 enhanced miR-21-3p binding to the variant 3'-UTR. Transfection of miR-21-3p, miR-103b, miR-141-3p, miR-200a-3p, and miR-376b-3p mimics into LS180 human intestinal cells showed repression of UGT1A1 and UGT1A6 mediated glucuronidation and mRNA without affecting UGT2B7 activity or mRNA. Furthermore, transfection of miR-21-3p, miR-141-3p, and miR-200a-3p into primary human hepatocytes, repressed UGT1A1 activity and mRNA without affecting CYP3A activity. Finally, miR-21-3p and miR-200a-3p expression were negatively correlated with UGT1A6 activity and mRNA in human liver samples. Thus, UGT1A is regulated by multiple miRNAs with some showing allele-dependent effects.

Keywords: Functional genomics; Glucuronidation; UDP-glucuronosyltransferase; UGT1A; microRNA.

Copyright © 2017 Elsevier Inc. All rights reserved.

Figures

Figure 1.

miRNAs identified through a functional genomics library screen that reduce UGT1A 3’-UTR luciferase reporter (pMIR-UGT1A) activity by at least 30% without affecting activity of the empty vector (pMIR). Shown are relative luciferase activities (firefly/Renilla ratio normalized to the pMIR control ratio) of HEK293 cells co-transfected with the pMIR-UGT1A or pMIR and with each miRNA mimic, a positive control (siUGT1A: short interfering RNA directed against the UGT1A 3’-UTR) or the mimic negative control. Bars represent mean ± one S.D. from three independent experiments performed in quadruplicate. * p < 0.05, ** p < 0.01, *** p < 0.001.

Figure 2.

MicroRNA response elements identified in the UGT1A 3’-UTR (reference sequence) by in silico analysis for 5 miRNAs (A-F) shown to reduce UGT1A 3’-UTR luciferase reporter activity by at least 30% without affecting an empty luciferase reporter. Nucleotide positions in the mRNA are indicated relative to the last nucleotide of the stop codon.

Figure 3.

Functional validation of novel miRNA response elements (MREs; shown in Figure 2) in the UGT1A 3’-UTR (reference sequence). For each miRNA evaluated (panels A – E) luciferase reporter constructs were generated containing either the full length UGT1A 3’-UTR, a deletion (Δ) of the predicted MRE, or the triplicated MRE in the forward or reverse (negative control) orientation. Shown are relative luciferase activities (firefly/Renilla ratio normalized to the miRNA control) of HEK293 cells co-transfected with the luciferase reporter and either the miRNA mimic or the mimic negative control. Bars represent mean ± one S.D. from three independent experiments performed in quadruplicate. * p < 0.05, ** p < 0.01, *** p < 0.001.

Figure 4.

Effect of the most common UGT1A 3’-UTR variant allele (comprised of three linked SNPs: rs10929303, rs8330, and rs1042640) on the binding of miRNAs to the reference UGT1A 3’-UTR. A) The rs8330 C>G SNP (bold and underlined) is located in the seed sequence of the miR-1286 miRNA response element (MRE) and predicted to disrupt binding of miR-1286 to the variant UGT1A 3’-UTR. B) miR-1286 decreases luciferase activity of the reference UGT1A 3’-UTR not the variant UGT1A 3’-UTR (all 3 SNPs) or the rs8330 UGT1A 3’-UTR (single SNP). C) miR-1286 decreases luciferase activity of the forward (F) triplicate reference (wt) miR-1286 MRE reporter but not of the triplicate variant (var) miR-1286 MRE reporter, or of the reverse orientation (R) reference or variant miR-1286 MRE reporters. D) All of the other miRNAs evaluated (miR-103b, miR-141–3p, miR-200a-3p and miR-376b-3p) decrease luciferase activities of both the reference and variant UGT1A 3’-UTR constructs. Shown are relative luciferase activities (firefly/Renilla ratio normalized to the miRNA control) of HEK293 cells co-transfected with the luciferase reporter and either the miRNA mimic or the mimic negative control. Bars represent mean ± one S.D. from three independent experiments performed in quadruplicate. * p < 0.05, ** p < 0.01, *** p < 0.001.

Figure 5.

Effect of the rs10929303 SNP on the binding of miR-21–3p to the UGT1A 3’-UTR. A) The rs10929303 C>T SNP (bold and underlined) is located in the seed sequence of the miR-21–3p miRNA response element (MRE) and predicted to enhance binding of miR-21–3p to the variant UGT1A 3’-UTR (right) compared to the reference UGT1A 3’-UTR (left). B) miR-21–3p decreases luciferase activity of the variant UGT1A 3’-UTR (all 3 SNPs) and the rs10929303 UGT1A 3’-UTR (single SNP) but not the reference UGT1A 3’-UTR. C) miR-21–3p decreases luciferase activity of the triplicate variant (var) miR-21–3p MRE reporter more than the reference (wt) miR-21–3p MRE reporter in the forward (F) orientation, but has no effect on the reverse orientation (R) reporters. Shown are relative luciferase activities (firefly/Renilla ratio normalized to the miRNA control) of HEK293 cells co-transfected with the luciferase reporter and either the miRNA mimic or the mimic negative control. Bars represent mean ± one S.D. from three independent experiments performed in quadruplicate. * p < 0.05, ** p < 0.01.

Figure 6.

Effect of miRNA mimic overexpression on UGT1A1 (A), UGT1A6 (B), and UGT2B7 (C) selective activities and mRNA levels in LS180 human intestinal cells. miRNA mimics or the mimic negative controls were transfected into LS180 cells at 30 nM concentration. 48 hr after transfection, glucuronidation activities (UGT1A1: ezetimibe glucuronidation, Ez-Glu; UGT1A6: 5-hydroxytryptamine glucuronidation, 5HT-Glu; UGT2B7: morphine 3-glucuronidation, M3-Glu) were measured and total RNA isolated for the measurement of expression levels of the respective genes. Glucuronidation activity values and mRNA levels were compared with the average values of cells transfected with the mimic negative controls and non-transfected cells (control). Bars represent mean ± one S.D. from three independent experiments performed in quadruplicate. * p < 0.05, ** p < 0.01, *** p < 0.001.

Figure 7.

Effect of miRNA mimic overexpression on UGT1A1 and CYP3A enzyme activities, and UGT1A1, UGT1A4, UGT1A6, UGT1A9, and CYP3A4 mRNA levels in sandwich-cultured primary human hepatocytes. miRNA mimics or the mimic negative controls were transfected into primary human hepatocytes at 30nM concentration. 48hr after transfection enzyme activities (UGT1A1: ezetimibe glucuronidation, Ez-Glu; CYP3A: midazolam 1-hydroxylation) were measured and total RNA isolated for the measurement of expression levels of the respective genes. Enzyme activity values and mRNA levels are compared with the average values of cells transfected with the mimic negative controls and non-transfected cells (control). Bars represent mean ± one S.D. from three independent experiments performed in quadruplicate. * p < 0.05, *** p < 0.001.

Figure 8.

Locations of known miRNA response elements (MREs) in the reference (top) and variant (bottom) UGT1A 3’-UTR. Indicated are the positions (from nucleotides +1 to +740 downstream of the UGT1A stop codon) of the six MREs validated in this study (red triangles) and the MRE for miR-491–3p, which had been reported previously [16]. Also indicated are the positions of the three linked SNPs (rs10929303, rs1042640, and rs8330) that comprise the variant UGT1A 3’-UTR allele. The rs10929303 SNP was found to substantially increased binding of miR-21–3p to the variant allele, while the rs8330 SNP abolished binding of miR-1286 to the variant 3’-UTR.