COMMENTARY on: HDL-transferred microRNA-223 regulates ICAM-1 expression in endothelial cells

The accumulation of cholesterol in the arterial wall initiates the progression of atherosclerosis, which is one of the major causes of death in Western societies^{1, 2}. Excess cholesterol must be removed and transported from the peripheral tissues to the liver for its reutilization or its excretion into feces in a physiological process traditionally known as reverse cholesterol transport (RCT)³. During RCT, plasma high-density lipoprotein (HDL) is thought to function as a sterol transporter that facilitates the movement of sterols from peripheral cells to the liver. In addition to its role in regulating RCT, a number of studies have shown that HDL may also have anti-atherogenic properties^{4, 5}. Indeed, HDL decreases endothelium inflammation and oxidative stress, and increases nitric oxide production (NO) and endothelial cell (EC) survival, thus preventing atherogenesis^6–8. Even though these observations have been reported in several studies, the molecular mechanisms underlying these effects are still unclear.

In a recent report published in the February 28, 2014, issue of Nature Communications, Tabet and colleges showed that HDL can transfer microRNAs (miRNAs) to ECs influencing gene expression in the recipient cell⁹. miRNAs are small non-coding RNAs that regulate gene expression at the post-transcriptional level by inhibiting the translation or decreasing the stability of mRNA target genes. The authors found that ECs treated with native HDL (nHDL) showed increased levels of miR-223. This miRNA was reducing EC inflammation by directly targeting intercellular adhesion molecule 1 (ICAM1). The enrichment of miR-223 in ECs was specifically due to incubation of ECs with nHDL since their incubation with other HDL components, such as ApoA1 or recombinant HDL (rHDL), did not influence the endothelial levels of miR-223. The authors employed a number of elegant experimental approaches to demonstrate that the miRNA transfer occurs between nHDL and ECs in vitro. For instance, to avoid the confounding effect of endogenous miR-223 in ECs, the authors treated ECs with actinomycin D (to inhibit de novo transcription) or silenced Dicer expression using a siRNA (to inhibit endogenous miR-223 maturation) in the presence of nHDL. In both experiments, in the presence of nHDL, miR-223 levels remained similar to untreated controls (absence of actinomycin D or scrambled siRNA), demonstrating that nHDL efficiently transfers miR-223 to ECs.

To assess the functional relevance of miR-223 in ECs, the authors analyzed miRNA-predicted targets using bioinformatic algorithms (TargetScan). Interestingly, they found ICAM-1, a glycoprotein that regulates vascular inflammation by facilitating leukocyte recruitment, and colony stimulating factor 2 (CSF2), a cytokine that controls the production, differentiation and function of macrophages, as predicted miR-223-target genes. To demonstrate that miR-223 regulates the ICAM-1 and CSF2 expression at the post-transcriptional level, the authors cloned the 3′UTR region of both genes in a luciferase reporter vector and assessed luciferase activity after overexpressing miR-223. The results indicated that miR-223 down-regulated ICAM-1 and CSF-2 protein expression levels. More interestingly, miR-223 diminished ICAM-1 protein expression in pro-inflammatory conditions [ECs treated with pro-atherogenic cytokines, such as tumor necrosis factor alpha (TNFα)].

Finally the authors tested the role of HDL-derived miR-223 in regulating EC activation by comparing the anti-inflammatory effect of HDL isolated from wild-type (WT) and miR-223 deficient mice. Notably, ECs treated with HDL isolated from WT mice diminished ICAM-1 and CSF-2 levels. However, this anti-inflammatory effect was lost in ECs treated with HDL isolated from miR-223^−/− mice, suggesting that HDL-derived miR-223 plays an important role in the well-described anti-inflammatory properties of HDL.

One important question that needs to be addressed is the mechanism by which the miRNAs are transfered between HDL and ECs. Previous work from the Ramaley lab demonstrated that the scavenger receptor B1 (SRB1) was critical for the uptake of miRNAs in human hepatic cell lines (Huh7)¹⁰. Since SRB1 is also expressed in ECs, it could be possible that the same receptor may mediate the HDL-derived miRNA transfer to ECs.

Other groups have also studied the potential transfer of HDL-containing miRNAs to ECs. Dimmeler and colleagues found that miR-223 was the most abundant miRNA in HDL but they were unable to demonstrate the transfer of miRNAs between HDL and ECs¹¹. Moreover, they did not find differences in the miRNA content of HDL isolated from healthy control subjects and patients with stable coronary artery disease (CAD) or acute coronary syndrome (ACS)¹¹. The discrepancies between the results obtained by both groups might be explained by the different origin of ECs used in their respective studies. While Tabet et al. used primary human coronary aortic endothelial cells (HCAECs), Wagner and colleagues performed their studies in human umbilical venous endothelial cells (HUVECs). The different expression levels of SRB1, as well as other receptors that mediate miRNA transfer between HDL and ECs, in HCAECs and HUVECs might address this discrepancy. It is also important to note that the study of cellular transport in ECs in vitro is very challenging for a number of reasons including: 1) the loss of endothelial glycocalyx that controls lipoprotein retention and mechanotransduction, 2) the absence of caveolae observed in primary ECs cultured in vitro and 3) the loss of EC polarization that may influence membrane receptor localization. Therefore, to definitely demonstrate the biological significance of these findings, the transfer of HDL-derived miRNAs should be tested using an in vivo model or in cannulated vessels.

In summary, this interesting study shows the potential transfer of HDL-associated miRNAs to ECs and provides a novel mechanism by which HDL might regulate EC activation. Further studies of how HDL-derived miRNAs might influence gene expression in other cells associated with atherosclerotic vascular disease, such as macrophages and vascular smooth muscle cells, might be of interest.