Human MicroRNA-33b Promotes Atherosclerosis in Apoe^−/− Mice

MicroRNAs (miRs) are small endogenous non-coding RNAs that modulate protein levels after interacting with 3´-untranslated regions of mRNA by inducing mRNA degradation or stalling protein translation ^1–3. Several reports in 2010 identified miR-33a as a regulator of ABCA1/ABCG1 and cholesterol efflux ^4–7. Simultaneously, it was observed that Mir33a gene was in the intron of Srebf2 gene that codes for the SREBP2 transcription factor, a well-known master regulator of cholesterol metabolism ⁸. Hence, it was immediately recognized that miR-33a may be co-transcribed with the host gene Srebf2 to coordinately regulate cholesterol metabolism. Srebf2 expression and activation are increased with cholesterol depletion, including that which occurs with statin inhibition of HMG CoA reductase. Thus, within the liver miR-33a could, by reducing ABCA1, decrease secretion of HDL and reduce HDL levels. Indeed, several studies involving miR-33a inhibition and knockout in mice showed that miR-33a plays a role in determining plasma HDL cholesterol levels and reverse cholesterol transport.

Sequence comparison studies showed that humans and other mammals express two miR-33 homologs, miR-33a and miR-33b, that differ by two nucleotides ⁴. These two miR-33 homologs are present in the introns of two different genes SREBF2 and SREBF1, respectively. In contrast, rodents only express miR-33a. Rodents lack miR-33b due to deletions in the homologous intronic sequences in the Srebf1 gene. In other words, rodents have a non-functional homologous Mir33b gene.

To date several studies in mice have established the role of miR-33a in cholesterol efflux, in vivo reverse cholesterol transport, plasma HDL cholesterol levels, fatty acid oxidation and atherosclerosis ^9–14. In contrast, the role of miR-33b is not well studied. Studies with anti-miR-33 in monkeys showed that inhibition of both miR-33a and miR-33b increases plasma HDL levels and lowers VLDL triglyceride ^{15, 16}. However, these studies fail to recognize specific role of miR-33b. Attempts have begun to study the role of miR-33b in mice. In the first approach, the entire mouse Srebf1 gene was replaced with the human SREBF1 gene ¹⁷. This resulted in the expression of human SREBF1 and MIR33B. However, these mice had difficulty in growth and fertility, and were difficult to study. Horie et al used a different approach ¹⁸. They generated miR-33b knock-in mice by introducing human MIR33B sequence into the intron 16 of Srebf1 gene ¹⁸. Human miR-33b is in intron 16 of human SREBF1 gene. This approach resulted in generation of knock-in (KI) mice that produced miR-33b in quantities similar to those seen in humans; however, these mice showed reduced expression of endogenous miR-33a. Introduction of miR-33b sequence had no effect on the splicing and expression of the host gene Srebf1. In these KI mice, the expression of miR-33b was co-incident with Srebf1 transcription and was regulated by LxR agonist. Thus, these KI mice co-expressed miR-33b with its host Srebf1 mRNA and their expression was co-regulated with host mRNA. The expression levels of target genes, such as ABCA1, were reduced in livers of these KI mice. Peritoneal macrophages isolated from the KI mice also expressed reduced levels of ABCA1 and ABCG1 proteins and showed reduced apo-A1 and HDL-mediated cholesterol efflux.

The effect of miR-33b expression on plasma HDL was measured in mice expressing one copy (heterozygous, MIR33B^KI+/−) and two copies (homozygous, MIR33B^KI+/+). One copy of MIR33B significantly reduced plasma HDL-cholesterol and total cholesterol compared with wild type controls. Presence of two copies did not result in any further decrease in plasma HDL cholesterol, indicating that one copy of miR-33b is sufficient for physiologic effects. These studies suggest that greater than 50% inhibition of miR-33b is probably needed to see a significant phenotypic effect.

In a paper published in this issue, Nishino et al.¹⁹ report that miR-33b is present in atherosclerotic plaques in humans and in these mice. Further, these investigators have crossed MIR33B^KI+/+ mice with Apoe^−/− mice. Expression of human miR-33b reduced serum HDL-cholesterol and cholesterol efflux, reverse cholesterol transport, inflammatory response and increased atherosclerosis plaque size in MIR33B^KI+/+Apoe^−/− mice. The atherosclerotic plaques in KI mice had larger necrotic core compared to control Mir30b^−/−Apoe^−/− mice. These plaques also contained more macrophages, less collagen content and more apoptotic cells indicating more plaque instability.

MIR33B^KI+/+Apoe^−/− mice had reduced levels of miR-33a. Despite this, hepatic mRNA and protein levels of Abca1 were reduced indicating that miR-33b was affecting expression. As expected, lower hepatic ABCA1 led to lower plasma HDL-cholesterol levels and apoB-depleted plasma from these mice showed reduced cholesterol efflux capacity. Reduced in vivo reverse cholesterol transport was associated with reduced movement of macrophage cholesterol into the serum. These studies point to functional expression of human miR-33b and its ability to modulate mouse cholesterol metabolism.

Nishino et al. also studied the effect of miR-33b on macrophage function. Induction of ABCA1/ABCG1 was suppressed in MIR33B^KI+/+Apoe^−/− peritoneal macrophages treated with acetylated LDL compared to controls. Further, MIR33B^KI+/+Apoe^−/− macrophages showed reduced cholesterol efflux to apoA1 and HDL. These peritoneal macrophages also showed increased apoptosis when ACAT activity was inhibited. Thus, it is likely that miR-33b expression reduces expression of ABCA1/ABCG1, reduces cholesterol efflux and increases sensitivity to free cholesterol assimilation.

MiR-33b expression significantly increased macrophage inflammatory response. Elicited peritoneal MIR33B^KI+/+Apoe^−/− macrophages showed increased production of pro-inflammatory IL-1β, IL-6 and CCL2 cytokines. Further studies in THP-1 cells showed that expression of miR-33b increases lipid rafts detectable using cholera toxin subunit B. These rafts can be reduced after treatment with cyclodextrins to remove cholesterol and attenuate production of pro-inflammatory cytokines. Thus, miR-33b may enhance free cholesterol accumulation in plasma membrane, lipid rafts formation, and production of pro-inflammatory cytokines.

The studies with isolated peritoneal macrophages clearly showed that miR-33b had significant effect on cholesterol efflux and inflammatory response. To determine whether bone marrow derived macrophages have similar defects bone marrow transplantation studies were performed. Mice that received bone marrow cells from MIR33B^KI+/+Apoe^−/− mice showed greater plaque areas and more macrophage content. These transplantation studies had no effect on plasma cholesterol levels. Thus, bone marrow derived cells expressing miR-33b contribute to atherosclerosis independent of regulating plasma HDL cholesterol levels.

Surprisingly, Mir33b^KI+/+Apoe^−/− mice had low plasma triglyceride levels¹⁹. To explain low plasma triglyceride levels, investigators measured heparin-releasable lipase activity and found that lipase activity was unaffected by miR-33b. Further, expression of apoC2, apoC3, Gpihbp1, Angptl3, Angptl4 and Angptl8 were unaffected. Thus, miR-33b does not appear to affect lipoprotein triglyceride hydrolysis. Lipoprotein production studies showed that miR-33b had no effect on VLDL production. Surprisingly, they observed slower lipid absorption in these mice. It is possible that reduced lipid absorption contributes to low plasma triglyceride levels. More mechanistic studies are needed to find out how and why miR-33b might be involved in lipid absorption and how it affects overall plasma triglyceride metabolism. Moreover, these studies need to be confirmed in another model system. Reductions in plasma triglycerides were not seen MIR33B^+/− and MIR33B^+/+ mice on wild type background ¹⁸. Also in contrast to these results, inhibition of miR-33a/b in monkeys lowered VLDL triglyceride ¹⁵. Therefore, it remains to be determined whether reduced plasma triglyceride levels in MIR33B^KI+/+Apoe^−/− mice is due to gene-gene interactions or they are secondary to defective lipid absorption. Studies in MIR33B^KI+/+Ldlr^−/− mice may provide some clues. Or studies can be performed in MIR33B^KI+/+ mice injected with anti-sense oligonucleotides against LDL receptors ²⁰. Further, MIR33B^KI+/+Mir33a^−/− mice may identify specific role, if any, of miR-33b in triglyceride metabolism. It is unclear whether reduced lipid absorption is secondary to miR-33b’s cellular effects on enterocytes or it is a consequence of changes in regulatory mechanisms in the brain. This is important as genetic ablation of miR-33 has been shown to increase food intake ²¹ leading to obesity and insulin resistance that can be avoided by pair feeding. Thus, miR-33 might be involved in modulating feeding behavior and modulating plasma triglyceride levels. More studies are required to explain molecular basis for the regulation of feeding behavior.

It should be noted that Srebp-1 and Srebp-2 genes are regulated differently by hormones, diets and drugs. Thus, miR-33a and miR-33b expression may be regulated differently. Hence, under different conditions these miRs may play different functions. Therefore, knowledge about their individual regulations might be useful. For this purpose, mice expressing only miR-33b (MIR33B^KI+/+Mir33a^−/−) can be generated by crossing MIR33B^KI+/+ mice with Mir33a^−/− mice. These mice may be useful in identifying individual and combined targets of miR-33a and miR-33b. As these two miRs share the same seed sequence, it is likely that these microRNAs regulate similar target mRNAs. Additionally, they may have independent functions in regulating cholesterol and triglyceride metabolism due to their different affinities for common and unique targets depending on their interactions with target mRNAs involving additional supplementary site interactions. The regulation of target mRNAs can additionally be modulated by the number of homologous target sequences for these miRs.

Even though miR-33b is co-transcribed with the Srebf1 gene, these studies did not provide convincing evidence for the regulation of triglyceride metabolism by miR-33b. It has previously been shown that miR-33a regulates triglyceride metabolism by modulating Srebf1 expression ²². It is possible that mice that express only miR-33b, and not miR-33a, may be useful in deciphering the role of miR-33b in triglyceride metabolism similar to that of the role of Srebp1c in triglyceride metabolism, hepatosteatosis and insulin resistance. However, extrapolation of these mice studies to humans may require confirmation of findings in human cells lacking either miR-33a or miR-33b. Understanding pathways and proteins modulated individually by miR-33a and miR-33b and in combination will be valuable in executing and explaining anticipated and unanticipated physiological and adverse consequences of therapeutic approaches aimed at reducing expression levels of these miRs. Further, knowledge of genes and pathways affected in different cells, especially in hepatocytes and macrophages would be more useful in successful launch of miR-33-targeted therapeutics.

It remains to be determined if unstable plaques seen in MIR33B^KI+/+Apoe^−/− mice are susceptible to rupture. Injection of angiotensin II and other modulators can be used to study the vulnerability of these plaques to rupture ²³. Further, it remains to be determined whether miR-33b has similar effect on plaque composition in Ldlr^−/−mice. Effect of miR-33b in Ldlr^−/− mice may provide some clues as to the interaction of miR-33b with LDL receptors and ApoE in mice. Further studies in mice only expressing human miR33b (MIR33B^KI+/+Mir33a^−/−Apoe^−/−) may help identify specific role of miR-33b in cholesterol metabolism and atherosclerosis. These mice can be generated by crossing MIR33B^KI+/+Apoe^−/− mice with Mir33a^−/− mice.

In short, these studies represent advancement towards teasing apart combined and individual roles and functions of miR-33a and miR-33b in the control of lipid metabolism. Next step would be to create models expressing only miR-33b. We anticipate that studies with these mice may point to evolutionary advantages of acquiring miR-33b in mammals. Additionally, they might be useful in designing therapeutic approaches to reduce either miR-33a or miR-33b as opposed to current focus of simultaneously inhibiting both the miRs.

Figure: Location of different miR-33 genes in human and mouse genome:

(A) Human genome contains MIR33B and MIR33A genes in the introns of SREBF1 and SREBF2 genes.

(B) Mouse genome contains Mir33a gene in the intron of Srebf2 gene.

(C) Human MIR33B was introduced in the intron of mouse Srebf1 gene to create a mouse that produces both miR-33a and miR-33b.