Where in the (lncRNA) world is CARMN? Safeguarding against vascular dysfunction

The development of massively parallel sequencing enhanced our appreciation of the regulatory role of RNA. Amongst the breakthroughs that challenged traditionally held view that RNA acts solely as an “intermediary”, is the identification of long noncoding RNAs (lncRNAs). Defined as transcripts greater than 200 bp that do not code for proteins ¹, lncRNAs represent the largest class of noncoding RNAs. The vast majority of lncRNAs structurally resemble mRNA, in that they are transcribed by RNA polymerase II, undergo polyadenylation, cap methylation and splicing. Most lncRNAs are expressed in a very specific cellular context, implying their potential role as signaling mediators integrating developmental cues or exogenous stimuli. Over the last decade, there has been a tsunami of studies investigating lncRNAs in various biologic contexts. What we have learned from these studies is that lncRNAs work through diverse mechanisms. Some function through interactions with epigenetic modifiers or acting as scaffolds for transcriptional machinery ¹. Others, competitively inhibit microRNAs². A few, surprisingly, have been shown to harbor micropeptides ³. More importantly, we have learned that at least a subset of lncRNAs are functionally relevant and that their dysregulation has significant implications in health and disease states⁴.

The vast majority of lncRNA studies are in the area of organismal development and cancer biology, but despite a slow start, rapidly accumulating evidence supports important roles for lncRNAs in cardiovascular disease⁵. In this issue, Vacante et al expand our understanding of the role lncRNA in atherosclerosis pathways ⁶. The authors investigated the contributions of the lncRNA CARMN in vascular smooth muscle cell (VSMC) responses to lipid overload (Figure 1). Loss of CARMN in human VSMCs is associated with enhanced proliferation, migration and cellular plasticity. CARMN depletion enhanced the expression of pro-inflammatory cytokines, leukocyte migration factors and reduced the expression of the smooth muscle identity markers Acta2 and Myh11. The authors show that the effects of CARMN are operational in vivo since global deletion of CARMN was associated with enhanced VSMC proliferation and atherosclerosis development. It should be noted that this is not the first study to report an important role for CARMN in cell fate determination. Previous work has shown that CARMN (AKA CARMEN) is required for cardiovascular lineage specification and differentiation⁷. Thus, multiple lines of evidence reinforce an ascribed role for CARMN in cell plasticity.

Figure:

Contribution of CARMN in atherosclerosis. CAMRN, miR143, miR145 are expressed in vascular smooth muscle cells (VSCMs) and are down regulated in advanced plaques in both mice and humans. Loss of CARMN reduces miR143/miR145 and enhances VSMCs migration, inflammation, de-differentiation and proliferation in response to pro-atherogenic stimuli. CARMN effects on proliferation may be independent of miR143/miR145. CARMN knockout mice develop accelerated atherosclerosis and advanced plaques.

Numerous studies have shown that lncRNAs may impact the expression of their adjacent genes^{1, 8}. Hence, the location of a lncRNA, arguably, provides the most critical clues regarding its function. The CARMN gene locus is in the vicinity of a noteworthy gene cluster known to impact cardiovascular cell-fate determination, which includes miR-143, miR-145 and Braveheart (Bvht). Bvht is one of the first lncRNAs shown to impact cardiovascular lineage commitment ⁹. miR-143 and miR-145 regulate smooth muscle plasticity in response to proatherogenic stimuli and enhancing their expression reduces lesion size and leads to more favorable plaque composition ¹⁰. To explore whether CARMN regulates its bordering genes, the authors first examined the independence of CARMN gene products from neighboring genes. This is an essential step since lncRNA annotations have poorly defined boundaries that may not capture context-specific isoforms or longer transcripts that overlap with neighboring genes ¹¹. Using a combination of isoform specific primers, RACE experiments and transcript discovery algorithms the authors defined over 10 CARMN isoforms and at least one transcript, surprisingly, overlapped with miR-143 and miR-145. GapmeR antisense oligonucleotide (ASO) knockdown of CARMN reduced the expression of miR-143 and miR-145 but not Bvht. Since miR-143 and miR-145 may be part of a parent CARMN transcript, it becomes difficult to uncouple direct effects of the knockdown strategy on miR143/miR145 from downstream consequences of loss of CARMN. Crucially, however, the authors show that restoration of miR-143 and miR-145 expression is not sufficient to fully rescue the loss of CARMN phenotype. Thus, at least some of the effects of CARMN appear to be independent of miR143 and miR145. Although there are many unanswered questions about the role and mechanisms of CARMN, the study highlights the critical value of meticulous dissection of lncRNA expression patterns. Initial “housekeeping” studies that characterize a given lncRNA are often tedious and time consuming, but they are essential in unraveling mechanisms at play and should predate in depth functional interrogation.

A logical premise in biology is that evolutionary conservation serves as a surrogate marker for significance, however, comparing base pairs across lncRNAs from different species provides a narrow viewpoint of importance ¹². LncRNAs exhibit low sequence conservation, in line with the idea that nature is perhaps not under evolutionary pressure to preserve lncRNA open reading frames, rather secondary structure conservation or act of lncRNA production may be more important. Indeed, few lncRNAs show sequence conservation even when liberalizing the degree of sequence homology. An intriguing finding here is that CARMN appears to be one lncRNA that shows syntenic and partial sequence conservation, comparing the mouse and human orthologues. In their study, Vacante et al show that CARMN expression is downregulated in advanced atheroma lesions from patients undergoing carotid endarterectomy (CEA )as well as advanced atherosclerosis lesions from Ldlr^−/− mice. These results are consistent with in vitro analysis in hCASMCs showing that CARMN is downregulated in response to proatherogenic stimulation including oxidized LDL loading or PDGF-BB treatment. In addition, loss of CARMN led to proliferation defects in both human cell lines and murine atherosclerosis model, suggesting that functional conservation of this lncRNA may be preserved. Since we lack clear paradigms that define the importance of lncRNA relationships across species ¹², the identification of CARMN may provide welcome opportunities to investigate a molecular basis for lncRNA conservation.

Arguably, the most important conceptual advance from this study is the generation a murine knockout of a lncRNA. Although there is no shortage of studies claiming functional relevance of lncRNAs in cardiovascular disease, only a limited number showed functional effects in knockout models particularly in the context of atherosclerosis ⁴. To examine the contributions of CARMN in atherogenesis, the authors used a gain of function AAV-PCSK9 approach that allows rapid and efficient induction of hypercholesterolemia. Administration of AAV-PCSK9 to CARMN^−/− mice enhanced cellular proliferation within lesions compared with controls. These findings are important since a given lncRNA may regulate specific pathways in only one cell type but not the other or function in vitro but not in vivo. For example, approximately 90% of lncRNAs that modify cellular growth do so exclusively in one cell type ¹³. In addition, numerous studies showed minimal phenotypic changes in chronic loss of function models of lncRNAs, perhaps due to compensatory effects ¹⁴. It should be noted that the use of ASO or other RNAi approaches is well-justified since it is convenient tool and does not perturb embedded DNA but one must consider off target effects particularly when interpreting proliferative or inflammatory phenotypes ¹⁵. Akin to generating a knockout for a protein-coding gene, a genetic perturbation approach, when feasible, is invaluable in determining the physiologic contribution of a given lncRNA.

In the popular 90s video game “Where in the World is Carmen Sandiego?” a player must use a series of cues to infer Carmen’s destination. Navigating the lncRNA discovery pipeline is a similar endeavor in that one must build a case for lncRNA significance starting with simple clues, such as lncRNA location, regulation, abundance, localization, and protein-coding potential, that ultimately reveal mechanism(s) and physiologic role(s). The characterization of CARMN adds a new player to the complex world of cellular proliferative responses within lesions. In the future, it would be interesting to further disentangle the effects of CARMN from neighboring genes and the molecular basis of feedback circuits between CARMN and miR143/miR145. Testing the function of CARMN in other cell types and defining its direct target genes and key interacting partners will also be prudent to underpinning its significance in atherosclerosis.