Cardiac injury is accompanied by dynamic changes in the expression of microRNAs (miRs). For example, miR-150 is down-regulated in patients with acute myocardial infarction, atrial fibrillation, dilated and ischaemic cardiomyopathy as well as in various mouse heart failure (HF) models. Circulating miR-150 has been recently proposed as a better biomarker of HF than traditional clinical markers such as brain natriuretic peptide. We recently showed using the β-arrestin-biased β-blocker, carvedilol that β-arrestin1-biased β1-adrenergic receptor cardioprotective signalling stimulates the processing of miR-150 in the heart. However, the potential role of miR-150 in ischaemic injury and HF is unknown.
Here, we show that genetic deletion of miR-150 in mice causes abnormalities in cardiac structural and functional remodelling after MI. The cardioprotective roles of miR-150 during ischaemic injury were in part attributed to direct repression of the pro-apoptotic genes egr2 (zinc-binding transcription factor induced by ischaemia) and p2x7r (pro-inflammatory ATP receptor) in cardiomyocytes.
These findings reveal a pivotal role for miR-150 as a regulator of cardiomyocyte survival during cardiac injury.
MicroRNAs (miRNAs or miRs) have become increasingly recognized as major regulators of various physiological processes.1 Recently, modulation of miR activity in the heart has been suggested as an important mechanism that underlies the pathogenesis of cardiac diseases.2,3 MiR-150 was shown to be down-regulated in patients with acute myocardial infarction (AMI), atrial fibrillation, dilated cardiomyopathy, and ischaemic cardiomyopathy4–6 as well as in various mouse heart failure (HF) models (TAC, MI, and I/R injury).6,7 Intriguingly, analysis of the global expression of miRs in an experimental model of physiological left-ventricular (LV) hypertrophy in mice showed an increase in miR-150 levels after 35 days of voluntary exercise.8 Studies of patients also showed that circulating miR-150 may be a better biomarker of HF than clinically used markers such as brain natriuretic peptide.4–6,9,10 Despite the increasing data from both human and rodent studies that implicate miR-150 as an important factor in the regulation of cardiac remodelling, direct evidence demonstrating a role for miR-150 in cardiac stress responses is lacking.
The pro-apoptotic genes egr2 (zinc-binding transcription factor induced by ischaemia) and p2x7r (pro-inflammatory ATP receptor) have been shown to be regulated by miR-150 in cancer and pulmonary cells.11–13 Notably, egr2 was significantly down-regulated during preconditioning in mouse hearts,14 and loss-of-function variants of p2x7r were associated with reduced risk of ischaemic heart disease.15 It is unknown, however, whether these two genes are regulated by miR-150 in the heart.
We show here that genetic deletion of miR-150 alters the pathological responses of the heart to MI and that miR-150 acts as a gatekeeper of cardiomyocyte survival by repressing pro-apoptotic egr2 and p2x7r. Therefore, miR-150 may represent a novel therapeutic target for combating cardiac injury.
MiR-150 knock-out (KO) mice, mouse model of MI, and post-MI mortality
MiR-150 KO mice16 were obtained from Jackson Laboratory and 8–12-week-old miR-150 KO or wild type (WT) controls were subjected to MI as previously published.17 Briefly, the mice were anaesthetized using 1–4% inhalant isoflurane and placed on a heating pad. Animals were intubated and ventilated with room air using a MiniVent mouse ventilator. The left-anterior descending coronary artery (LAD) was visualized under a microscope and ligated by using a 6-0 prolene suture. Regional ischaemia was confirmed by visual inspection under a dissecting microscope by discolouration of the occluded distal myocardium. Sham-operated animals underwent the same procedure without occlusion of the LAD. One dose of buprenorphine (0.05 mg/kg) was given subcutaneously immediately after the surgery. We used response to toe/skin pinch, heart rate and blood pressure for anaesthesia, and post-operative monitoring plan. We also monitored the survival of miR-150 KO mice and their WT control littermates following MI.
Left-ventricular performance was assessed by two-dimensional echocardiography using a Visual Sonics Vevo 2100 Ultrasound at baseline (pre-surgery) and post-MI (1 day, 1, 2, 4, and 8 weeks) as previously used.18,19 M-mode tracings were used to measure anterior and posterior wall thicknesses at end diastole and end systole. Left-ventricular internal diameter (LVID) was measured in either diastole (LVIDd) or systole (LVIDs). A single observer blinded to mouse genotypes performed echocardiography and data analysis. Fractional shortening (FS) was calculated according to the following formula: FS (%) = [(LVIDd − LVIDs)/LVIDd] × 100. Ejection fraction (EF) was calculated by: EF (%) = [end-diastolic volume (EDV) − end-systolic volume (ESV)/EDV] × 100.
Histology and immunohistochemistry
The hearts were harvested and weighed before undergoing gross anatomical examinations. Morphometric analysis of heart size was performed as published.20–22 We then examined histological changes in the cardiac tissues such as fibrosis (Masson trichrome staining) and inflammation (neutrophil and T cell staining) using standard procedures as previously used.23–26 For gross histological examination, sections were stained with haematoxylin and eosin (H and E). Myocardial sections were also stained for TUNEL to measure apoptosis using In Situ Cell Death Detection Kit (Roche) according to the manufacturer's instructions. The following antibodies were used: CD3, rabbit polyclonal (ab5690, Abcam); CD31, rabbit polyclonal (ab28364, Abcam); and Troponin I, rabbit polyclonal (sc-15368, Santa Cruz).
Cell culture and transfection
Mouse adult atrial cardiomyocyte HL-1 cells obtained from Dr Claycomb and rat embryonic ventricular cardiomyocyte H9c2 cells were maintained as previously described.23 Immortalized mouse cardiac endothelial cell (MCEC) line was purchased from Cedarlane and maintained according to the company's recommendation. Primary vascular smooth muscle cells (VSMCs) from adult mouse aorta were isolated and maintained as we described previously.27 Primary neonatal rat ventricular cardiomyocytes (NRVCs) were isolated by dissociation of 1- to 2-day-old Sprague–Dawley rats and were maintained as we published.28 Cardiomyocytes were transfected with a siRNA control (Santa Cruz, cat#: sc-37007), or siRNAs targeting egr2 (OriGene, cat#: SR505739) or p2x7r (Santa Cruz, cat#: sc-108056) with Lipofectamine™ 2000 reagent (Invitrogen) as previously described.29 In order to inhibit microRNA-150 expression in cardiomyocytes, we also transfected Ambion Anti-miR™ miRNA Inhibitors (Life Technologies) specific to miR-150 (MH10070) and a miR inhibitor negative control (4464076) using Lipofectamine™ 2000 reagent (Invitrogen) as described previously.24
For gain-of-function studies, we transfected CMV expression plasmids for miR-150 (Origene, SC400788) or miR-150 mimics (Life Technologies, MC10070). All in vitro assays were performed 60–72 h after transfection when maximum knockdown efficiency was reached.
In vitro simulated ischaemia–reperfusion assays
Cells plated on coverslips or six-well plates were transfected with miR inhibitors, miR mimics, or siRNAs as aforementioned, washed, and placed in an ischaemia buffer that contained 118 mM NaCl, 24 mM NaH2CO3, 1 mM NaHPO4, 2.5 mM CaCl2, 1.2 mM MgCl2, 20 mM sodium lactate, 16 mM KCl, and 10 mM 2-deoxyglucose (pH 6.2). Cells were then incubated in the anoxic chamber (5% CO2, 0% O2) for 1 h followed by 4 h of reperfusion-mimicking conditions (by replacing the ischaemic buffer with normal cell medium under normoxia conditions) as described.30 Coverslips or plates were processed for qRT–PCR, immunoblotting, and TUNEL staining as mentioned in Supplementary material online, Supplemental methods.
Animal study approval
Eight- to 12-week-old C57BL/6 WT and miR-150 KO mice, and 1- to 2-day-old Sprague–Dawley rats were used for this study. Research with animals carried out for this study was performed according to approved protocols and animal welfare regulations of Georgia Regents University's Institutional IACUC Committees. All animal procedures were performed to conform the NIH guidelines (Guide for the care and use of laboratory animals). The mice were euthanized by thoracotomy with 1–4% inhalant isoflurane for further heart analysis and the neonatal rats were euthanized by decapitation for cardiomyocyte isolation.
Data are expressed as mean ± SEM from at least four independent experiments with different biological samples or mice per group. Statistical significance was determined by using two-way comparisons for echocardiographic data, one-way ANOVA with Bonferroni correction for multiple comparisons, or Student unpaired t-tests (GraphPad Prism version 5). A P-value <0.05 was considered statistically significant.
Other methods are provided in Supplementary material online.
MiR-150 mutant mice have normal cardiac structure and function at baseline
To investigate in vivo roles of miR-150 in cardiac injury, we obtained miR-150 KO mice16 and we confirmed that miR-150 was not detected in the KO heart (Figure 1A). We then showed that the hearts of miR-150 KO mice at baseline were functionally (Supplementary material online, Table S1 and Figure 1B) and structurally (Figure 1C–G) normal, did not exhibit enhanced neutrophil infiltration, and had normal expression patterns of genes involved in cardiac stress, fibrosis, apoptosis, and inflammation (Figure 1D–G).
MiR-150 null mice display significantly impaired cardiac function and structure after MI
Despite the normal phenotypes at baseline, the mutant mice responded differently to ischaemic cardiac injury. MI induced by permanent ligation of LAD resulted in a significant increase in mortality in miR-150 KO mice compared with WT controls. The detrimental effect of miR-150 deficiency on survival became obvious after 2 days (Figure 2A). Because miR-150 is involved in regulation of the innate immunity16 and there is evidence that an intense inflammatory reaction contributes to the development of ventricular rupture (VR) after MI,31 we next tested if VR is the predominant cause of death in miR-150 KO mice after MI. Upon postmortem examination, mutant mice were indeed more susceptible to cardiac rupture (KO, 26.5% compared with WT, 11.1%, P < 0.01), as previously reported that the expression of miR-150 in human autopsy samples of infarcted heart tissues with VR is lower compared with MI patients without VR.32 We next examined the expression of miR-150 in different myocardial cells and aortic vascular smooth muscle cells (VSMCs). The expression of miR-150 in VSMCs was significantly lower than in myocardial cells (Supplementary material online, Figure S1). By qRT–PCR analysis, we observed down-regulation of miR-150 in the hearts of WT mice subjected to 8 weeks of MI (Figure 2B), which is consistent with a recent study in the hearts at post-MI 5 days and 4 weeks.33 We next tested the cardiac response of miR-150 KO mice to the same MI procedure. Indeed, the miR-150 KO mice displayed severely impaired cardiac function at 1 day (Figure 2C and Supplementary material online, Table S2), 4 weeks (Supplementary material online, Table S5), and 8 weeks (Figure 2D and Supplementary material online, Table S6) after MI, but not at post-MI 1 week and 2 weeks (Supplementary material online, Tables S3 and S4). On the contrary, WT controls showed a less functional deficit following MI (Figure 2C–D and Supplementary material online, Tables S2, S5, and S6).
We also found that miR-150 KO hearts have higher numbers of TUNEL-positive cells (Figure 3A), increased neutrophil infiltration (Figure 3B), as well as increased necrosis and disorganized structure (Figure 3C) after 1 day of MI when compared with WT MI hearts. However, we did not observe any difference in T cell infiltration between WT and miR-150 KO mice (Supplementary material online, Figure S2). To determine which cell types are undergoing apoptosis in miR-150 KO hearts, we combined TUNEL labelling with cardiomyocyte (CM) marker Troponin I (TnI) or cardiac endothelial cell (EC) marker CD31 on infarcted mouse hearts from miR-150 KO and WT controls because miR-150 is abundantly expressed in these two cell types in the heart (Supplementary material online, Figure S1A). We found that miR-150 KO hearts had higher numbers of TUNEL-positive CMs, not ECs after 1 day of MI. Interestingly, we observed that apoptotic CMs disassemble their sarcomeres because TnI staining on most of the TUNEL-positive area was relatively dim (Supplementary material online, Figure S3).
To further assess the response of miR-150 KO mice to MI, we examined fibrosis. Masson's trichrome staining of hearts at 8 weeks post-MI revealed small areas of fibrosis in WT hearts, while miR-150 KO hearts contained larger fibrotic regions (Figure 3D). Because of relatively high miR-150 levels in cardiac ECs (Supplementary material online, Figure S1A), we next tested the hypothesis that impaired EC response may contribute to increased fibrosis and cell death seen in miR-150 KO MI hearts. Notably, capillary density or the expression of endothelial cell markers was comparable between WT and miR-150 KO groups (Supplementary material online, Figure S4), suggesting that miR-150 protects the myocardium through the inhibition of apoptosis, not neovascularization after MI. Our biochemical data also showed that miR-150 KO MI hearts had increased mRNA levels of fetal ANP, fibrotic Col3a1, apoptotic Bax, and inflammatory IL-6 compared with WT controls (Figure 3E–H), which is consistent with our immunohistochemical data (Figure 3A–D). These results suggest that deletion of miR-150 resulted in diverse abnormalities during post-ischaemic cardiac structural/functional remodelling.
MiR-150 regulates pro-apoptotic egr2 and p2x7r
In order to identify candidate miR-150 targets, we used several prediction algorithms including miRDB,34 PicTar,35 and Targetscan.36In silico ingenuity pathway analysis37 showed that one of the top associated network functions of the predicted targets of miR-150 is anti-proliferation, cell-cycle arrest, or apoptosis. Accordingly, we focused on apoptosis-related genes and found that egr2, p2x7r, and p53, which were predicted by all three target algorithms, were potential targets of miR-150. The binding sites of miR-150 were well conserved among egr2 mRNAs from mouse, rat, and human, which elevates the significance of miR-150 on the regulation of egr2. However, rat p2x7r 3′-untranslated region (3′-UTR) does not have the miR-150 binding site although mouse and human 3′-UTRs possess the miR-150 binding site.
Because programmed CM death has been suggested to underlie progressive ventricular remodelling and ischaemic cardiac failure38–41 as supported by our data (Supplementary material online, Figure S3), we next postulated that elucidating the roles of miR-150/target pairs in CMs will provide important insight into cardioprotective signalling. To identify the functional targets of miR-150 in CMs, we performed loss- and gain-of- function studies in HL-1 cells (adult mouse atrial CMs where p2x7r was shown to be expressed42) and NRVCs. Two of predicted targets (egr2 and p2x7), but not p53, were up-regulated with miR-150 inhibition and down-regulated with miR-150 overexpression (Figure 4A–E and Supplementary material online, Figure S5A).
The CM results were confirmed in vivo by WB analysis revealing significantly increased levels of egr2 and p2x7r (not p53) in miR-150 KO mouse hearts at both baseline and post-MI at 8 weeks (Figure 4F AND Supplementary material online, Figure S5B). To test whether egr2 and p2x7r are direct targets of miR-150 repression, we co-transfected HL- 1 cells with constitutively active luciferase reporter constructs containing 3′-UTRs of egr2 or p2x7r (Figure 4G) and miR-150 expression plasmid. We observed repression of luciferase activity by miR-150 for the egr2- or p2x7r-3′-UTR reporter (Figure 4H). Mutation of seed binding sites for miR-150 made the reporter insensitive to miR-150 transfection (Figure 4H), indicating the specific dependence of target 3′-UTRs on miR- 150.
To test whether the two targets of miR-150 were up-regulated in either CMs exposed to low oxygen conditions or hearts of WT mice subjected to 8 weeks of MI, we measured the expression levels in both CMs and hearts. The two targets were up-regulated in simulated I/R (Supplementary material online, Figure S5C–E) and in the hearts at post-MI 8 weeks (Supplementary material online, Figure S5F–G), which is consistent with previous studies at post-MI 2–8 h, 1 day, and 4 weeks.43,44 These results strongly suggest that the two genes are functional CM targets of miR-150 because miR-150 is down-regulated in I/R6,7 and MI (Figure 2B).
We next determined if the two targets of miR-150 regulate CM apoptosis. Loss-of-function approaches uncovered that the knockdown of egr2 or p2x7r decreased NRVC or HL-1 cell apoptosis in response to simulated I/R (Supplementary material online, Figures S6 and S7).
Taken together, our data indicate that down-regulation of pro-apoptotic egr2 and p2x7r in part contributes to the beneficial actions of miR-150 in the heart.
MiR-150 functions as a protective miR by repressing pro-apoptotic egr2 and p2x7r in CMs
Because our data suggest that pro-apoptotic egr2 and p2x7r are direct targets of miR-150, we further hypothesized that miR-150 may function as a cell-survival miR. I/R-induced miR-150 down-regulation in CMs was first confirmed by qRT–PCR (Figure 5A and B). To determine the importance of miR-150 for CM function in low oxygen conditions, we used in vitro models of I/R to show that miR-150 protects CMs from cell death. Loss-of-function of miR-150 in HL-1 cells and NRVCs increased CM apoptosis (Figure 5C–E). We next tested whether miR-150 activates survival signalling in adult CMs and found that miR-150 overexpression increases p-AKT levels in both basal and simulated I/R conditions (Figure 5F). Because the expression of miR-150 is also relatively high in cardiac ECs (Supplementary material online, Figure S1A), we investigated the role of miR-150 in cardiac ECs and showed that miR-150 overexpression blocks MCEC apoptosis in simulated I/R condition (Supplementary material online, Figure S8), suggesting that miR-150 functions as a survival miR in both CMs and cardiac ECs.
Despite our ability to validate binding of miR-150 to the respective 3′-UTRs of egr2 and p2x7r, the question remained whether a functional relationship between miR-150 and these targets existed with respect to CM apoptosis. On the basis of the assumption that anti-miR-150 exerts its pro-apoptotic effect through the derepression of specific targets, we designed an siRNA-based strategy (Supplementary material online, Figures S9 and S10) to validate the functional relevance of these targets. If the targets we identified were responsible for the effect of anti-miR-150 on CMs, RNA interference with their expression should specifically counteract anti-miR-150-induced apoptosis. This concept was tested by transfecting CMs with siRNAs to silence the expression of targets of miR-150. Consistent with our earlier observations (Figure 5C–E), anti-miR-150 alone enhanced CM apoptosis, yet the siRNA against egr2 or p2x7r efficiently prevented the pro-apoptotic effect of anti-miR-150 (Supplementary material online, Figures S9 and S10). Taken together, our CM data support the in vivo evidence that miR-150 exerts cardioprotective effects in part through direct repression of pro-apoptotic egr2 and p2x7r.
The results of this study reveal miR-150 as a central, ischaemic stress-responsive protector against CM apoptosis both in vivo and in vitro. Mice deficient for miR-150 are sensitized to MI, evidenced by increased cardiac apoptosis and fibrosis as well as loss of pump function. MiR-150 directly inhibits pro-apoptotic egr2 and p2x7r such that elevated expression of egr2 and p2x7r in miR-150 KO mice causes a greater degree of CM death sustained during ischaemic injury. CMs lacking miR-150 have an increased sensitivity to I/R-induced apoptosis, while CMs overexpressing miR-150 have increased pro-survival signalling.
We recently showed that miR-150 is a β1-adrenergic receptor (β1AR)/β-arrestin1-regulatable miR which is post-transcriptionally activated by a biased β-blocker, carvedilol (Figure 6A–C). Together with the results presented here (Figure 6D), we postulate that β- arrestin1-biased β1AR regulatory mechanism of miR processing in CMs (the only cardiac cell type in which β1ARs are expressed) may result in beneficial adaptive remodelling in failing hearts. This concept is further supported by the observation that three other β1AR/β-arrestin1-regulatable miRs (miR-125b-5p, miR-199a-3p, and miR-214) activated by carvedilol45 are cardioprotective in vivo during MI and I/R injury.30,46,47 Interestingly, two other studies linked carvedilol treatment to cardioprotective miR up-regulation have been also reported in rat models of MI.48,49 A recent study showed that the expression of the cardioprotective miR-13350,51 in myocardial tissue was significantly up-regulated in the carvedilol pre-treatment group before MI surgery compared with the MI group and that up-regulation of miR-133 mediates the anti-apoptotic action of carvedilol in the isolated cardiomyocytes.48 The up-regulation of MiR-29b, which is another cardioprotectve miR,52 was also shown to mediate the effect of carvedilol on attenuating MI-induced fibrosis.49 These previous studies provided the evidence that the cardioprotective actions of carvedilol against HF are associated with increased levels of cardioprotective miRs. Future studies are warranted to fully understand the possible overlapping/compensatory effects of these miRs on carvedilol-mediated cardioprotection and the detailed underlying mechanisms of their actions.
Egr2 and p2x7r were shown to be regulated by miR-150 in cancer and pulmonary cells.11–13 In the context of heart, egr2 was significantly down-regulated during opioidergic sustained ligand-activated preconditioning in mouse hearts.14 P2x7r is expressed in atrial CMs (not ventricular CMs) and other cardiac cells such as cardiac microvascular endothelial cells.42 P2x7r is also expressed in immune cells and p2x7r in cardiac master cells was shown to regulate cardiac function.54 P2x7r was specifically up-regulated in graft-infiltrating lymphocytes in cardiac-transplanted humans and mice, and targeting p2x7r has shown initial promise in prolonging cardiac transplant survival.55 Loss-of-function variants of p2x7r in humans have been associated with reduced risk of ischaemic heart disease.15 Our data further support that inhibition of these genes in myocardial cells and/or immune cells, where miR-150 is also expressed,56 could be considered as therapeutic options for cardiac disease. In fact, a recent study demonstrated that (i) microglial p2x7r activates pro-inflammatory cytokines in the hypothalamic paraventricular nucleus, thereby contributing to the augmented sympathetic nerve activity in acute myocardial ischaemic rats and (ii) systemic administration of p2x7r antagonist BBG in AMI rats improves cardiac function.44
Notably, clinical-grade p2x7r antagonists have been already used in inflammatory bowel disease, rheumatoid arthritis, and chronic obstructive airway disease.57 Given our data that these two apoptotic genes are direct targets of miR-150, MI patients with reduced levels of circulating miR-150 could be particularly considered for current and future targeted treatments based on p2x7r and egr2.
In conclusion, our results suggest that miR-150 protects the heart against ischaemic stress by blunting cell death in response to injury in part through its repression of egr2 and p2x7r. Interestingly, a recent study in a mouse model of AMI reported that (i) overexpression of miR-150 protects the heart from ischaemic injury by inhibiting inflammatory monocyte migration, and that (ii) CXCR4 is a target of miR-150 in monocyte.58 This study suggests a prominent role for miR-150 in post-MI monocyte recruitment (i.e. extra-cardiac effect of miR-150), which may contribute to the reduced CM apoptosis seen in our current study. However, the detailed mechanisms by which miR-150 affects myocardium apoptosis are unknown and so tissue-specific KO studies will help to define cardiac vs. extra-cardiac actions of miR-150. Given that down-regulation of miR-150 also underlies cardiac hypertrophy and other forms of heart disease4–7 and that miR-150 may play an important role in reactive oxygen species- mediated gene regulation in CMs,59 the cardioprotective role of miR-150 is likely broadly applicable to a variety of stress settings and thus boosting miR-150 levels to attenuate cardiac cell death may provide therapeutic benefits. However, additional in vivo injury models (cardiac hypertrophy and I/R) and detailed studies on other functions of miR-150 in CMs such as calcium signals are needed before considering this miR as a therapeutic option.
Y.T., Y.W., K.P., Q.U., and I.K. designed experiments, directed the study, and wrote the paper. Y.T., Y.W., K.P., and Q.U. performed most of the experiments, analysed the data, and prepared the figures. I.K. supervised the study and provided financial support. J.T., Z.B., P.R., C.J., J.L., H.S., Y.T., and G.R. performed some experiments, helped to analyse the data, and helped to write the paper.
This work was supported by American Heart Association (AHA) Greater Southeast Affiliate Postdoctoral Fellowship 13POST16840074 to P.R., National Institutes of Health (NIH) R01 DK083379 to G.R., and AHA Grant-in-Aid 12GRNT12100048, AHA National Scientist Development Grant 14SDG18970040, and National Institutes of Health R01 HL124251 to I.K.
We thank Drs Neal Weintraub and John Johnson for critically reviewing the manuscript.
Conflict of interest: none declared.