|Home | About | Journals | Submit | Contact Us | Français|
MicroRNAs (miRNAs/miRs) are small conserved RNA molecules of 22 nucleotides, which negatively modulate gene expression primarily through base paring to the 3′ untranslated region (UTR) of target mRNAs. The muscle-specific miR-1 has been implicated in cardiac hypertrophy, heart development, cardiac stem cell differentiation, and arrhythmias through targeting of regulatory proteins. In this study, we investigated the molecular mechanisms through which miR-1 intervenes in regulation of muscle cell growth and differentiation.
Based on bioinformatics tools, biochemical assays and in vivo models, we demonstrate that 1) IGF-I and insulin growth factor 1 receptor (IGF-1R) are targets of miR-1; 2) miR-1 and IGF-1 protein levels are inversely correlated in models of cardiac hypertrophy and failure as well as in the C2C12 skeletal muscle cell model of differentiation; 3) the activation state of the IGF-1 signal transduction cascade reciprocally regulates miR-1 expression through the Foxo3a transcription factor; and 4) miR-1 expression inversely correlates with cardiac mass and thickness in myocardial biopsies of acromegalic patients, in which IGF-1 is overproduced following aberrant synthesis of growth hormone.
Our results reveal a critical role of miR-1 in mediating the effects of the IGF-1 pathway and demonstrate a feedback loop between miR-1 expression and the IGF-1 signal transduction cascade.
MicroRNAs (miRNAs/miRs) are small conserved RNA molecules of ~22 nt, which negatively modulate gene expression in eukaryotic organisms. The base pairing of a specific miR to the 3′UTR of its mRNAs targets leads to mRNA cleavage and/or translation repression1. Bioinformatic analysis predicts that each miR may regulate hundreds of targets, indicating their important role in most biological processes such as cell proliferation, apoptosis, and stress responses2–5. Recently, these small RNA molecules have also been demonstrated to be involved in myocardial and cardiac stem cells development6, cardiac hypertrophy and remodeling, arrhythmias7–11, as is the case of miR-16, 12, 13, 14, the expression of which inversely correlates with cardiac hypertrophy11, 15. Notably, mice harboring deletion of miR-1 showed cardiac defects, including misregulation of cardiac morphogenesis, electrical conduction, and cell proliferation10, 14. Among putative miR-1 targets, proteins such as growth factors (Toll-like), transcription factors, and signal transduction kinases have been validated as true targets12, 13. Insulin-like growth factor-1 (IGF-1) is a key regulator of growth, survival, and differentiation in most cell types16 and its importance is demonstrated by the conservation throughout the evolutionary scale16. In myocardial biology, IGF-1 and its signal transduction cascade is involved in the control of virtually every critical biological process, including development, cardiomyoctye size and survival, action potential, and excitation-contraction coupling17, 18. In this manuscript, we demonstrate that IGF-1 is not only a target of miR-119 but remarkably that miR-1 expression per se depends on the activation state of the IGF-1 signal transduction cascade. Indeed, IGF-1 upregulation correlates with miR-1 depression and vice versa. All together, these findings unravel a unique reciprocal molecular circuit between miR-1 and IGF-1.
All procedures involving animals were performed in accordance with institutional guidelines for the care and use of laboratory animals.
Transverse aortic constriction (TAC) experiments were performed on male C57/Bl6 mice ranging in age from 10 to 12 weeks, with 6 animals per group. AKT transgenic mice with cardiac specific overexpression of constitutively active AKT were male mice of age included between 10 and 12 weeks.
mouse neonatal cardiomyocytes were prepared as previously described11. Mouse C2C12 and 293T cells were obtained from ATCC (Manassas, VA) and maintained in DMEM (Invitrogen, Carlsbad, CA) supplemented with 4.5 g/L glucose, 4 mM L-glutamine, 10% fetal bovine serum, and penicillin/streptomycin at 37 °C in a 5% CO2 atmosphere. C2C12 were differentiated in DMEM supplemented with 4% Horse serum. MiR-1 and miR-133 adenoviral vectors (Ad-miR-1 and Ad-miR133, respectively) as well as the Ad-Empty control have been described previously11.
miR-1 target prediction was performed using the following algorithms: miRanda (http://microrna.sanger.ac.uk), TargetScan (http://www.targetscan.org), and PicTar (http://pic-tar.bio.nyu.edu). 3′UTR ΔG was calculated using the software mFOLD (http://frontend.bioinfo.rpi.edu/applications/mfold/). MiR-1 promoter analysis was done using the software MatInspector (http://www.genomatix.de/products/MatInspector).
The miR-1 promoter, which contains two potential binding sites for Foxo3a, was amplified from mouse chromosome 2 using KOD Taq polymerase (Novagen, San Diego, CA) and the primers: forward, 5′-CTCTCAGTATCCTAATCCTC-3′, and reverse, 5′-GTAGGCACTCCTGCGCCGGC-3′. The promoter was cloned into the reporter plasmid, pGL3.basic (Promega, Madison, WI). For the promoter experiment, 293T cells were seeded in 24-well plates (Nunc, Rochester, NY) and infected with adenoviruses (Ad) carrying a dominant negative (DN) or dominant active (DA) form of AKT20, 21 or Foxo3a22 (kindly provided by Dr. Domenico Accili, Columbia University) and subsequently co-transfected with the miR-1 promoter-luciferase reporter (100ng) construct and the Renilla luciferase plasmid (10ng) using Lipofectamine 2000 (Invitrogen, Carlsbad, CA). For IGF-1 and IGF-1R 3′UTR reporter assays, experiments were performed on human 293T and murine C2C12 cell lines. 3′ UTR segments of IGF-1 and IGF-1R 3′UTR mRNAs were subcloned by standard procedures into psiCHECK-2 (Promega, Madison, WI) immediately downstream of the stop codon of the luciferase gene. MiR-1 seed mutagenesis was performed as described by the manufacturer (Stratagene, La Jolla, CA). 293T cells were transfected with the reporter plasmid (10, 50 and 100ng) and miR mimics (20nM) (Dharmacon, Lafayette, CO). The IGF-1 reporter plasmid (50ng) was transfected into C2C12 cells with Lipofectamine following a modified manufacturer’s protocol23. Cells were lysed and assayed for luciferase activity either 24 or 48 h after transfection. All luciferase assays were performed using the Dual Luciferase kit (Promega, Madison, WI) as previously described11,12.
Western blot analyses: neonatal cardiomyocytes (nCMC), C2C12, and 293T cells were collected at different time points (24, 48h) and lysated using RIPA buffer. The following antibodies were used: IGF-1 (Lab Vision Corporation, Fremont, CA), α-MHC (ABcam, Cambridge, MA), TroponinT (Thermo Scientific, Waltham, CA), and GAPDH (Mouse anti-Rabbit GAPDH, Cell Signaling, Danvers, MA), Total-AKT (Cell Signaling, Danvers, MA), Ser473 P-AKT (Cell Signaling, Danvers, MA), Total Foxo3a (Cell Signaling, Danvers, MA) and Tyr32 P-Foxo3a (Cell Signaling, Danvers, MA). Densitometry analyses were performed using ImageJ (NIH).
For IGF1 RNA quantification, total RNA was extracted using the TRIzol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s protocol. cDNA was prepared using SuperScript® Reverse Transcriptase cDNA Kit (Invitrogen, Carlsbad, CA). Sybr green qPCR was performed using the following primers: IGF1-5′ (5-CACCTCAGACAGGCATTGTG-3), IGF1-3′ (5-TCTGAGTCTTGGGCATGTCA-3); GAPDH-5′ (5–GACGGCCGCATCTTCTTGT-3), GAPDH-3′ (5–CACACCGACCTTCACCATTTT-3). For miRNA qRT-PCR total RNA was extracted using TRIzol; primers and probes specific for human miR-1 and the internal controls Sno202 RNA and U6 were purchased from Applied Biosystems. Amplification and detection was performed using the 7300 Sequence Detection System (ABI), using 40 cycles of denaturation at 95 °C (15 s) and annealing/extension at 60 °C (60 s). For the miRNA quantification this was preceded by reverse transcription at 42 °C for 30 min and denaturation at 85 °C for 5 min.
Northern blotting was performed to confirm the expression levels of miR-1. Probes and anti-sense oligoes against mature miR-1 and U6 were Locked_nucleic_acid (LNA) based (Exiqon, Vedbaek, Denmark). Densitometry analyses were performed using ImageJ (NIH).
C2C12 myoblasts were plated in Lab-Tek 2-well chamber slides (2.105 cells per well). IGF-1 staining was performed using primary antibody (Lab Vision Corporation, Fremont, CA) together with Alexa Fluor 594 phalloidin (Molecular Probes, Carlsbad, CA), diluted in 5% BA/0.1% Triton X-100/PBS. Goat anti-rabbit antibody conjugated to Alexa Fluor 488 (Molecular Probes, Carlsbad, CA) was used as secondary antibody.
Eight patients (6 male, 2 female, mean age 46.8±13.6 years) with acromegaly were evaluated with non-invasive (ECG, 2D-echocardiography) and invasive cardiac studies (cardiac catheterization, coronary and left ventricular angiography, and left ventricular endomyocardial biopsy) before neurosurgery and/or somatostatin analogue therapy. All invasive studies were performed after informed consent and approval by the ethic committee of Department of Cardiovascular and Respiratory Science, University La Sapienza, Rome. Endomyocardial biopsies were drawn, stored, and processed as previously described24.
Statistical and frequency distribution analysis was performed by GraphPad Prism 4.0 (GraphPad Software, Inc, La Jolla, CA). Data in figure 1 were analyzed using 2-way ANOVA adjusted pairwise test. Data in figures 2 and and33 were analyzed with two way ANOVA adjusted for multiple comparison and one way ANOVA test, respectively. Data in figures 4 and and55 were analyzed with two way ANOVA adjusted for multiple comparison. Data in figures 6 and 8 were analyzed using one way ANOVA test and two way ANOVA adjusted for multiple comparison, respectively. The linear correlation on acromegaly patient parameters was calculated using the Pearson product moment correlation coefficient. A value of p<0.05 or less was considered to be statistically significant.
The identification of true miR target genes is a great challenge and computational algorithms have been the major driving force in predicting miR targets to date25–27. Here, we used the following target prediction algorithms: miRanda (http://microrna.sanger.ac.uk)28, TargetScan (http://www.targetscan.org)29, and PicTar (http://pic-tar.bio.nyu.edu)5, 30. These approaches are based on the identification of elements in the 3′UTR of target genes complementary to the seed sequence of the miR of interest, calculation of thermodynamic properties of 3′UTRs, and phylogenetic conservation of complementary sequences in the 3′UTRs of orthologous genes. To reduce the number of false positive targets, we also analyzed mRNA secondary structure flanking the seed sequence. In fact, several line of evidence has been shown that complex mRNA tertiary structures might prevent miR/mRNA interactions that limit the number of single-stranded regions accessible for binding to miRs10, 13. Using this approach, we identified several putative targets of cardiac relevance, including, as very recently shown, IGF-119. Its 3′UTR contains only one seed sequence, which has a ΔG energy value of −4.5 and −11.9 kcal/mol for mRNA sequences 5′ and 3′ of the seed sequence, respectively. Interestingly, this sequence is evolutionally conserved between species (Figure 1A). An other putative target included IGF-1R, which is involved in the IGF-1 pathway (Supplementary Figure 1A). IGF-1R belongs to the large class of tyrosine kinase transmembrane receptors and is directly activated by IGF-1 as well as the related growth factor IGF-231. The ΔG energy value of the 5′ and 3′ RNA sequences around their seed sequences are −18 and −14 kcal/mol for IGF-1R.
In nCMC, overexpression of miR-1 but not its physiologically co-expressed miR-133, resulted in a significant decrease in IGF-1 protein levels (Figure 1B and C) without affecting its mRNA levels (Figure 1D). To confirm IGF-1 as a real miR-1 target, the wild-type 3′UTR of IGF-1 was cloned downstream of the luciferase gene and assayed in 293T cells, which do not express endogenous miR-1 (data not shown). When cotrasfected with miR-1 mimics, the assay resulted in a significant decrease (>45%) in luciferase activity (Figure 1E). In contrast, no effect was observed when an unrelated miR (miR-143) with no complementarity to any seed sequence in the 3′UTR of IGF-1 was used (Figure 1E). Consistent with this data, cotransfection of miR-1 with constructs containing mutated 3′UTR seed sequences did not result in any decrease in luciferase activity (Figure 1E). Similar results were obtained for luciferase constructs containing 3′UTRs of IGF-1R, where cotransfection with miR-1 mimics resulted in a significant decrease in luciferase activity (Supplementary Figure 1B). Consistent down-regulation of IGF-1R protein level was found in differentiated C2C12 cells (Supplementary Figure 1C).
It is well established that IGF-1 regulates skeletal muscle cells differentiation32 and that the miR-1 level changes during development in skeletal muscle cells12. Here, we used C2C12 cells as an in vitro cell model, which recapitulates skeletal muscle differentiation. Upon serum withdrawal, C2C12 myoblasts stop proliferating, irreversibly exit from the cell cycle, differentiate, and fuse into multinucleated skeletal myotubes after the concomitant coexpression of muscle differentiation markers33. Consistent with that, upregulation of cell differentiation markers 2 days after serum withdrawal, confirmed C2C12 differentiation (Figure 2A). Precursor and mature miR-1 level, while absent at the myoblast stage, were remarkably increased at day 2 of differentiation (Figure 2B), confirming previous observations12. The expression of miR-1 increased significantly following the switch to differentiation conditions, and remained at a high level up to day 10, when myotube formation was completed and contractile activity could be observed (data not shown).
To monitor changes in the miR-1 level upon C2C12 differentiation, we performed a luciferase assay by transfecting the IGF-1 luciferase construct in proliferating and differentiating C2C12 cells. Results show a significant reduction (>35%) in luciferase expression in differentiated cells compared to mutant IGF-1 3′UTR vector and to myoblast state (Figure 2C), indicating that inhibition of the reporter gene was a consequence of the increased endogenous level of miR-1. No changes were obtained when a luciferase reporter not carrying any 3′UTR was transfected (data not shown). In addition, immunofluorescence (IF) staining of C2C12 cells kept in differentiation medium demonstrated a striking decrease in the IGF-1 protein level starting from day 2 of differentiation (Figure 2D). In contrast, the IGF-1 mRNA level was unaltered (Figure 2E).
We11 and others15 have previously shown an inverse correlation between miR-1 expression and cardiac hypertrophy. We thus hypothesized that repression of miR-1 in the TAC model (Supplementary Table 1), would be accompanied by an increase in the IGF-1 protein level. In agreement with this, western blot analysis demonstrated an increased IGF-1 protein level after TAC, together with a decreased miR-1 level (figure 3A and B). Similarly, a reduction in miR-1 expression was found in another mouse model of cardiac hypertrophy, in which the downstream target of IGF-1, AKT, is constitutively overexpressed20. Consistent with our observations, decrease miR-1 expression was accompanied by an increased IGF-1 protein level (figure 3C and D).
Molecular circuits in which miRs control the level of a factor which in turn regulate the same miRs have been described. For instance, Fontana and al. described a circuitry involving sequentially miR-17-5p-20a-106a and AML1, where miR-17-5p-20a-106a functions as a master gene complex interlinked with AML1 in a mutual negative feedback loop34. We hypothesized that a circuit between miR-1 and IGF-1 might exist in striated muscle. To test this assumption, we measured miR-1 levels in different in vitro cell models in which IGF-1 activity was either diminished or increased. In nCMC treated with IGF-1, miR-1 expression was significantly decreased starting from 3 hours after treatment and remained depressed until 6 hours post treatment (Figure 4A). Notably, in C2C12 differentiated cells in which miR-1 is highly expressed, treatment with increasing IGF-1 doses induced a concomitant decrease in miR-1 expression with a significant reduction at a dose of 100nM IGF-1 (Figure 4B). A similar downregulation in miR-1 levels was observed when nCMC, were infected with an Ad carrying a dominant active form of AKT (DA-AKT)20 but not a dominant negative mutant (DN-AKT) (Figure 4C).
To study miR-transcriptional regulation, we analyzed the promoter region of miR-1 and identified two potential binding sites for Foxo3a, a transcription factor with cellular localization and transcriptional activity depending on its phosphorylation status, which in turn is regulated by AKT (Figure 5A). The miR-1 level was measured in nCMCs infected with Ad carrying a dominant active (DA-foxo) or dominant negative form (DN-foxo) of Foxo3a. In DA-foxo, all three potential AKT phosphorylation sites have been replaced by non-phosphorylatable amino acids, preventing it from being excluded from the nucleus in response to insulin, making it constitutive active22. In DN-foxo the transactivation domain (Δ256), that is important for co-activator recruitment, has been deleted. Our results demonstrate that while DA-foxo could upregulate miR-1 levels, the DN-foxo mutant could not (Figure 5C). These data strongly suggest that Foxo3a is able to influence miR-1 expression levels. Moreover, to test whether Foxo3a can regulate miR-1 promoter activity, we cloned a fragment containing the two potential binding sites mapped on mouse chromosome 2 into the luciferase vector. As shown in Figure 5B, DA-foxo but not DN-foxo, induced a marked elevation of miR-1 promoter activity.
Mir-1 overexpression should induce a decrease in the activation state of Akt and Foxo-3, due to the decreased levels of IGF-1 and IGF-1 receptors and thus a generalized decrease of IGF-1 signal transduction pathway. We therefore determined phospho-Foxo3a and phospho-AKT in differentiated C2C12 and nCMC infected with Ad-miR-1; data show a significant down-regulation of both phospho-Foxo3a and phospho-AKT (Figure 5D).
Taken together, these data suggest that Foxo3a can transcriptionally regulate miR-1 expression.
Acromegaly is a syndrome that results from overproduction of Growth Hormone (GH), which in turn increases IGF-1 synthesis in peripheral tissues and leads to a significant elevation in the IGF-1 plasma level35. Cardiac hypertrophy is a dangerous consequence of GH overproduction and consistently echocardiographic analysis of cardiac function in patients with acromegaly showed significantly increased left ventricular mass as well as end-diastolic left ventricular wall and septum thickness (Table 1)36. Furthermore, histological examination of myocardial biopsies showed hypertrophied cardiomyocytes with interstitial, perivascular, and focal replacement fibrosis (Figure 6A). Since blood levels of IGF-1 are elevated in patients with acromegaly, we measured miR-1 levels in myocardial biopsies by qRT-PCR (Figure 6B). Interestingly, all patients showed a significant reduction in cardiac miR-1 levels compared to healthy donors and there was a linear inverse correlation between miR-1 levels and left ventricular mass and end-diastolic left ventricular wall thickness (Figure 6C).
MiR-mediated post-transcriptional gene regulation is now considered a fundamental player of many genetic programs. However, despite our ability to identify miRs, few regulatory targets have been established for vertebrate miRs. In this study, we demonstrated the potential feedback loop in which miR-1 and its target IGF-1 are reciprocally regulated.
Binding of IGF-1 to its receptor activates its intrinsic receptor tyrosine kinase, which phosphorylates several intracellular substrates such as the insulin receptor substrate-1 and Shc37, 38, leading to the activation of signaling pathways, including phosphatidylinositol 3-kinase (PI3K)/AKT pathways39. Active AKT in turn phosphorylates and inhibits the winged-helix family of transcription factors, Foxo3a. Here we demonstrated not only through IGF-1 stimulation, but also with transfection of AKT or its downstream transcription factor Foxo3a that the IGF-1 pathway controls miR-1 expression (Figure 7). In fact, overexpression of AKT in nCMC reduces miR-1 expression, while overexpression of Foxo3a increases the miR level. These data are further supported by the identification of IGF-1R, the IGF-1 receptor31, as miR-1 targets.
Several miR-1 cardiac targets involved in the control of hypertrophy, cell cycle, excitation-contraction coupling, and membrane excitability have been identified, including HDAC412, Hand213, the K+ channel subunit, Kir2.1 (KCNJ2), connexin 43 (GJA1)14, Ras GTPase–activating protein (RasGAP), cyclin-dependent kinase 9 (Cdk9), fibronectin, Ras homolog enriched in brain (Rheb)15, and, as very recently shown, IGF-119. This indicates that miR-1 is involved in many different biological processes.
The importance of the IGF-1 pathway in cardiac function is well established40. Indeed, the effects of IGF-1 on adult muscle are pleiotropic, ranging from anti-apoptotic18, 41–43, hypertrophic, and regenerative effects on cardiac muscle31, 44, 45. Thus, both miR-1 and IGF-1 are involved in most biological processes controlling cardiac function. Notably, our results strongly indicate miR-1 as a modifier and a regulator of the multiple effects of IGF-1 in cardiac muscle. The possibility that IGF-1 is a target of miR-1 has previously been suggested19, 46. However, we show here the biological relevance of this functional interaction and demonstrate that it is part of autoregulatory circuit in cardiac and skeletal muscle (Figure 7). Our results also indicate that miR-1 may simultaneously downregulate IGF-1R together with IGF-1 protein, leading to a significant down-modulation of the whole downstream signal transduction cascade. The importance of our observation is strengthened by data from human samples, which strongly indicate the involvement of the miR-1 and IGF-1 regulatory loop in controlling human cardiac hypertrophy. Indeed, the striking inverse correlation between miR-1 levels and the degree of cardiac hypertrophy indicates that miR-1 is key player in this process in humans.
Funding Sources: This work is supported by the NIH (HL078797-01A1), Fondazione CARIPLO, Fundation LeDucq, Italian Ministry of Health, Italian Ministry of Research.
MicroRNAs (miRNAs/miRs) are small conserved RNA molecules of 22 nucleotides, which negatively modulate gene expression primarily through base paring to the 3' untranslated region (UTR) of target mRNAs. The muscle-specific miR-1 has been implicated in cardiac hypertrophy, heart development, cardiac stem cell differentiation, and arrhythmias through targeting of regulatory proteins. In this study, we investigated the molecular mechanisms through which miR-1 intervenes in regulation of cardiac and skeletal muscle cell growth and differentiation. We demonstrate that miR-1 controls the expression of IGF-1 and IGF-1 Receptor protein levels by translation. The IGF-1 pathway is critically involved in many aspects of cardiac development and cardiac function. We found that in conditions in which miR-1 is decreased such as cardiac hypertrophy, IGF-1 is increased; the regulation is reciprocal, since IGF-1 stimulation leads to upregulation of miR-1 levels; this effect is dependent upon IGF-1 activation of the Foxo3 transcription factor, which regulates mir-1 promoter activity. The clinical relevance of these observations is demonstrated by analyzing the expression of miR-1 in cardiac biopsies of patients with acromegaly, a condition in which Growth Hormone and IGF-1 are overproduced and cardiac myocyte size is dramatically increased: miR-1 levels inversely correlate with echocardiographic parameters of cardiac mass in these patients.
Our results add information on how the IGF-1 pathway regulates cardiac function demonstrating the role of miR-1 in response to cardiac stress and hypertrophy.