Long Noncoding RNAs in Cardiovascular Disease
Long noncoding RNAs (lncRNAs) consist of a new group of noncoding RNAs emerging as genetic modifiers in cardiovascular disease. Whereas miRNAs belong to the short RNAs (i.e., <200 nucleotides), lncRNAs harbor more than 200 nucleotides. In addition, other than miRNAs, lncRNA interaction is not (almost) exclusively dependent on mRNA translational repression or mRNA degradation but rather activates or silences gene transcription through chromatin regulation and transcriptional modulation (Geisler and Coller 2013; Wahlestedt 2013). In addition, transcriptional modulation of lncRNAs acts by cis (at the site of synthesis) or trans (at many different sites) fashion (Wahlestedt 2013; Mercer and Mattick 2013). LncRNA are not well preserved between species, and around 30% are specifically detected in primates (Derrien et al. 2012). Initial studies were performed in the oncology field, where investigated lncRNA was linked to cancer diseases (Gutschner et al. 2013; Yap et al. 2010). Some of these lncRNAs now gain also attention in the cardiovascular field, such as MALAT1 and ANRIL (Vausort et al. 2014; Burd et al.
2010). As gene transcription regulators, lncRNAs, like miRNAs, are involved in the regulation of cardiac development. One such lncRNA is Fendrr. Fendrr knockout models have been shown to impair embryonic cardiomyocyte proliferation, leading to hypoplasia and impaired cardiac function, resulting in embryonic death (Grote et al. 2013). Mechanistic studies observed that Fendrr modifies chromatin regulation through the binding of PRC2 and TrxG/MLL complexes (Grote et al. 2013). Deletion of Fendrr in another knockout model (Sauvageau et al. 2013) resulted in perinatal lethality. Ventricular septal defects and unstructured vessels were observed. In addition, reduced expression of Fendrr was observed in mutants of an endothelial-specific knockout model of Forkhead Box transcription factor F1 (FOXF1), a critical factor for vascular development (Ren et al. 2014). Another lncRNA involved in heart development is Braveheart, which is expressed in mice but lacks an ortho- logue in other species (Klattenhoff et al. 2013). Braveheart was found to interact in a gene network upstream of Mesp1 and is needed for activation of cardiac transcription factors, which drive mesodermal cells toward a cardiovascular phenotype (Klattenhoff et al. 2013). Braveheart was also detected as a cardiac-enriched lncRNA in a study investigating lncRNA expression in different mouse tissues (hearts, livers, and skin cells) using RNA sequencing (Matkovich et al. 2014). In adult mouse hearts, 152 lncRNAs showed high expression levels. Out of these lncRNAs, 48 lncRNAs are enriched in the heart as compared to liver and skin cells. Furthermore, RNA sequencing in cardiomyocytes and fibroblasts from adult mouse hearts indicates that most of these lncRNAs are enriched in the cardiomyocyte cell fraction (Matkovich et al. 2014). RNA sequencing allows also to search for differentially regulated lncRNA in cardiovascular disease models. Pedrazzini’s research group investigated alterations in cardiac lncRNA profiles after myocardial infarction in mice (Ounzain et al. 2015). Analysis of the RNA sequencing revealed 988 annotated lncRNAs but also identified 1521 novel lncRNAs, of which 60% are heart specific according to computational analysis. Importantly, human orthologues were found in 73% of novel lncRNAs detected in mouse. Downregulation of an unannotated novel lncRNA, NovInc6, was further shown in patients with dilated cardiomyopathy, in concert with suppression of a predicted target Nkx2-5, a key transcription factor of cardiac development and cardiac gene program. Using a similar approach, Zangrando et al. (2014) screened for differentially expressed lncRNAs 24 h after induction of MI in C57/BL6 mice using an Agilent microarray with 55,681 probes. Ten and twenty lncRNAs were down- and upregulated more than twofold, with NR_028427 (named myocardial infarction-associated transcript 1 (MIRT1)) and ENSMUST000001005122 (named MIRT2) showing the highest fold changes between groups. As a trend toward a correlation with LV remodeling parameters was observed, computational analysis of genes involved in remodeling processes identified strong correlation with 18 (for MIRT 1) and 17 (for MIRT 2) remodeling genes. However, no orthologues of MIRT1 and MIRT2 exist in humans.
As experimental models significantly affect lncRNA levels, it is of interest whether therapeutic interventions result in changes of lncRNA expression in humans. Yang et al. (2014) investigated differential expression of lncRNA in ischemic and nonischemic cardiomyopathy before and after LVAD support using RNA sequencing. Interestingly, cluster analysis revealed lncRNA signatures discriminating between ischemic and nonischemic cardiomyopathy. After LVAD support, a higher percentage of lncRNA show improved or normalized levels as compared to miRNA and mRNA expression profiles. LncRNA expression profiles furthermore were able to distinguish between before and after LVAD treatment, which indicates that lncRNA is involved in signaling pathways leading to reverse remodeling after LVAD support (Yang et al. 2014). Changes in transcriptome, including lncRNA expression, not only occur in the heart. A recent report by Deveaux’s group (Vausort et al. 2014) assumed that MI alters lncRNA levels in the blood drawn from these patients. In a large sample group, five pre-specified lncRNAs associated with cardiovascular disease processes (hypoxia-inducible factor 1A antisense RNA 2 (aHIF), ANRIL, potassium voltage-gated channel, KQT-like subfamily, member 1 opposite strand/antisense transcript 1 (KCNQ1OT1), MIAT, and MALAT1) were analyzed in patients with acute MI and presumably healthy subjects. Levels of aHIF, MALAT1, and KCNQ1OT1 were higher, and expression of ANRIL is lower in patients with MI as compared to healthy subjects. Expression analysis of the five lncRNAs in subpopulations of mononuclear cells (in healthy subjects) showed that the distribution pattern of lncRNAs differs in the subpopulations. In addition, ANRIL and KCNQ1OT1 added prognostic information to a clinical model for LV dysfunction (LVEF <40%) at 4-month follow-up.
Genetic variants have been observed as predictors of cardiovascular diseases. Variations on chromosome 9p21 (Samani et al. 2007; Ye et al. 2008) increase the susceptibility of cardiovascular disease. Single nucleotide polymorphisms (SNPs) within this genomic region are associated with coronary artery disease and premature myocardial infarction (Abdullah et al. 2008; Samani et al. 2007). Interestingly chromosome 9p21 harbors the lncRNA ANRIL (antisense noncoding RNA at the ink4 locus or CDKN2BAS (antisense to CDKN2B)). Recently, ANRIL expression was correlated with variants associated with a higher risk for coronary artery disease, suggesting that ANRIL regulates chromatin modulation of coronary artery disease susceptibility genes like the INK/ARF locus (Burd et al. 2010). Overexpression of ANRIL in monocytic cell lines increased proliferation, cell adhesion, and blunts apoptosis (Holdt et al. 2013), potential mechanisms that trigger atherosclerosis. This raises the possibility that at least some lncRNAs are the missing link between SNP and risk of MI and CAD. Another example of SNPs causing a risk of myocardial infarction is the discovery of myocardial infarction associated transcript (MIAT). Subjects with a SNP in exon 5 of MIAT show a higher susceptibility for myocardial infarction in a large-scale case-control association study.
Numerous reports observed that MIAT is involved in splicing efficiency, which may explain the findings of the aforementioned study (Aprea et al. 2013; Tsuiji et al.
2011). Metastasis-associated lung adenocarcinoma transcript 1 (MALAT1)/NEAT2 is another nuclear lncRNA with splicing ability (Hutchinson et al. 2007) and was found to enhance proliferation in human diploid fibroblasts and HeLa cells via this mechanism. However, in endothelial cells, where MALAT1 silencing impairs endothelial cell proliferation, expression of splicing-related genes is not altered, but cyclins and kinases were downregulated (Michalik et al. 2014). Together, these studies expand the knowledge of silencing and activation of gene networks in cardiovascular research and introduce lncRNA as new regulators in the complex molecular understanding. As lncRNAs are crucially involved in key features of cardiac injury, such as apoptosis, inflammation, impaired angiogenesis, and device treatment that lead to a change in lncRNA expression, lncRNAs may provide a future diagnostic and therapeutic clinical tool.
Acknowledgements Philipp Jakob and Ulf Landmesser are supported by the German Center of Cardiovascular Research (DZHK, Germany) and Berlin Institute of Health (BIH). Philipp Jakob is a participant in the BIH-Charite Clinical Scientist Program funded by the Charite - Universitatsmedizin Berlin and the Berlin Institute of Health (BIH).
Conflict of Interest Philipp Jakob has received research grants from Bayer Healthcare (Grants4Targets).
Compliance with Ethical Standards
All applicable international, national, and/or institutional guidelines for the care and use of animals were followed.
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.