miR-33 links SREBP-2 induction to repression of sterol transporters

Abstract

The sterol regulatory element binding protein 2 (SREBP-2) and the liver X receptor (LXR) control antagonistic transcriptional programs that stimulate cellular cholesterol uptake and synthesis, and cholesterol efflux, respectively. The clinical importance of SREBP-2 is revealed in patients with hypercholesterolemia treated with statins, which reduce low-density lipoprotein (LDL) cholesterol levels by increasing hepatic expression of SREBP-2 and its target, the LDL receptor. Here we show that miR-33 is encoded within SREBP-2 and that both mRNAs are coexpressed. We also identify sequences in the 3' UTR of ABCA1 and ABCG1, sterol transporter genes both previously shown to be regulated by LXR, as targets for miR-33-mediated silencing. Our data show that LXR-dependent cholesterol efflux to both ApoAI and serum is ameliorated by miR-33 overexpression and, conversely, stimulated by miR-33 silencing. Finally, we show that ABCA1 mRNA and protein and plasma HDL levels decline after hepatic overexpression of miR-33, whereas they increase after hepatic miR-33 silencing. These results suggest novel ways to manage hypercholesterolemic patients.

Fig. 1.

miR-33 and SREBP-2 are coexpressed. (A and B) The intragenic miR-33 is encoded within intron 16 of SREBP-2 (A), and both its sequence and genomic position are conserved across evolution (hsa, human; ptr, chimpanzee; mmu, mouse; rno, rat; bta, cow; gga, chicken; dme, fruit fly) (B). (C and D) Expression of miR-33 and selected SREBP-2 and LXR target genes in mouse (C) and human (D) primary macrophages following 48 h incubation in media containing low (closed bars) or high (open bars) levels of sterols (see Materials and Methods). **P < 0.01. Data are mean ± SD of three (human) or two (mouse) independent experiments in triplicate.

Fig. 2.

The expression of human and mouse ABCA1, and of mouse ABCG1 are regulated by miR-33. (A–C) Evolutionary conserved sequences in the 3′ UTR of ABCA1 and ABCG1 are partially complementary to miR-33. Annealings of miR-33 to some of the sequences are shown. (D) Luciferase activity in HEK293 cells after cotransfection of different constructs containing these putative response elements for miR-33 cloned downstream of the reporter stop codon, cotransfected with or without a miR-33 expression plasmid. Repression of luciferase activity suggests that these sequences are physiological targets for miR-33. Deviation from miR-33 complementarity results in loss of regulation by miR-33 (ABCA1 box 2; human ABCG1 sequence). (E) Expression of selected genes in Hep3B human hepatoma cells 48 h after transduction with an empty or a miR-33 adenovirus. (F) ABCA1 protein levels also are decreased after transduction of Hep3B cells with the miR-33 adenovirus. Where indicated, cells were incubated for 16 h with LXR:RXR agonists [1 μmol/L T0901317 (T) and 1 μmol/L 9-cis retinoid acid (9cRA)]. **P < 0.01. Data are mean ± SD of three independent experiments in duplicate.

Fig. 3.

Manipulation of miR-33 expression alters HDL lipidation. (A and B) Cholesterol efflux assay in HEK293 cells transfected with an empty plasmid or a plasmid encoding miR-33 (A), or with scrambled or miR33-specific hairpin inhibitors (B) (see Materials and Methods for details). After 36 h, the cells were washed and incubated for 16 h in media supplemented with [³H]-cholesterol (1 μCi/mL) in the presence or absence of LXR:RXR ligands, as described in Fig. 2. After 16 h, fresh media supplemented with BSA (0.2%), ApoAI (15 μg/mL), or FBS (20%) was added to the cells. Radioactivity in the media and in cell lysates was measured 6 h later. The percentage of efflux is expressed as dpm in the media versus total dpm (media + cells). *P < 0.01. Data are mean ± SD of eight wells (two independent experiments in quadruplicate). (C–F) For in vivo experiments, 8- to 10-wk-old male C57BL/6 mice were infused i.v. with 2 × 10⁹ empty or miR-33 adenovirus (C and D) or with 5 mg/kg/d scrambled or anti–miR-33 antisense oligonucleotides for 3 consecutive d (E and F) (n = 6–8 for Ad-GFP/Ad-miR33; n = 5/group for antisense oligonucleotides). The levels of hepatic ABCA1 mRNA and protein were measured 5 d (adenovirus) or 12 d (antisense oligonucleotides) after the infusion (C and E). Total cholesterol and HDL-cholesterol were measured in plasma of these same mice, using colorimetric enzymatic kits (D), and cholesterol distribution across plasma fractions were analyzed by FPLC (F). *P < 0.01. Data are mean ± SD.

Fig. 4.

A unified paradigm for cholesterol homeostasis. SREBP-2 and LXR coordinately regulate the positive (green) and negative (red) balances of intracellular cholesterol metabolism. Both pathways are not independent, but intersect through IDOL-1 and miR-33. Statin drugs, or conditions of low intracellular cholesterol, induce both SREBP-2 and the intragenic miR-33, leading to increased synthesis and uptake of sterols, as well as minimizing sterol loss through repression of ABC transporters.