Cyclic stretching boosts microRNA‐499 to regulate Bcl‐2 via microRNA‐208a in atrial fibroblasts

Abstract MicroRNAs that modulate transcription can regulate other microRNAs and are also up‐regulated under pathological stress. MicroRNA‐499 (miR‐499), microRNA‐208a (miR‐208a) and B‐cell lymphoma 2 (Bcl‐2) play roles in cardiovascular diseases, such as direct reprogramming of cardiac fibroblast into cardiomyocyte and cardiomyocyte apoptosis. Whether miR208a, miR499 and Bcl‐2 were critical regulators in cardiac fibroblast apoptosis under mechanical stretching conditions in human cardiac fibroblasts‐adult atrial (HCF‐aa) was investigated. Using negative pressure, HCF‐aa grown on a flexible membrane base were cyclically stretched to 20% of their maximum elongation. In adult rats, an aortocaval shunt was used to create an in vivo model of volume overload. MiR208a was up‐regulated early by stretching and returned to normal levels with longer stretching cycles, whereas the expression of miR499 and Bcl‐2 was up‐regulated by longer stretching times. Pre‐treatment with antagomir‐499 reversed the miR‐208a down‐regulation, whereas Bcl‐2 expression could be suppressed by miR‐208a overexpression. In the HCF‐aa under stretching for 1 h, miR‐499 overexpression decreased pri‐miR‐208a luciferase activity; this inhibition of pri‐miR‐208a luciferase activity with stretching was reversed when the miR‐499‐5p binding site in pri‐miR‐208a was mutated. The addition of antagomir‐208a reversed the Bcl‐2‐3′UTR suppression from stretching for 1 h. Flow cytometric analysis revealed that pre‐treatment with miR‐499 or antagomir‐208a inhibited cellular apoptosis in stretched HCF‐aa. In hearts with volume overload, miR‐499 overexpression inhibited myocardial miR‐208a expression, whereas Bcl‐2 expression could be suppressed by the addition of miR‐208a. In conclusion, miR‐208a mediated the regulation of miR‐499 on Bcl‐2 expression in stretched HCF‐aa and hearts with volume overload.


| INTRODUC TI ON
A microRNA is a small, non-coding strand of RNA with approximately 22 nucleotides, which functions in RNA silencing by interacting with the 3′-untranslated regions of the target mRNA that result in the destabilizing effects of target mRNA (a process referred to as gene silencing). 1 This results in the target RNA being silenced through one or more of the following processes: cleavage of the mRNA strand, destabilization of the mRNA by reducing the length of its poly(A) tail, or less efficient translation of the mRNA into proteins by ribosomes. 2 MicroRNAs (miRNA) are known to be widely involved in gene regulation in the pathogenesis of cardiovascular diseases, including cellular hypertrophy, fibrosis and apoptosis of cardiomyocytes. 3 Global miRNA expression profiling has identified miRNA-499 (miR-499), miRNA-208a (miR-208a) in the heart, and these miRNAs play important roles in cardiomyocyte hypertrophy and fibrosis in response to stress. 4,5 Like other RNAs, miRNA is also a product of transcription; therefore, one miRNA may regulate the expression of another miRNA. 6,7 For miR499, approximately 30% of its indirect cardiac mRNA target regulation is due to higher-order effects from secondarily regulated microRNAs. 6 The miR-208a regulation by miR-499 expanded the number of miR-499-modulated cardiac microRNAs. 6 However, van Rooij et al observe miR499 down-regulation upon miR208a knock down in their study. 5 Such microRNA-mediated regulation of mi-croRNA can be detected when properly assayed.
B-cell lymphoma 2 (Bcl-2), encoded in humans by the BCL2 gene, is the founding member of the Bcl-2 family of regulator proteins, which are key regulatory components of the mitochondrial cell death pathways. 8 Bcl-2 and its family members, such as Bcl-xl, Bcl-w and Mcl-1, are cytoprotective; other family members, such as Bad, Bak, Bax, Bid, Bim and Bmf, promote apoptosis. 9 Bcl-2 is localized in the outer membrane of mitochondria where it plays an important role in promoting cellular survival and inhibiting the actions of proapoptotic proteins. 9,10 The proliferation of atrial fibroblasts under pathological conditions plays an important role in atrial myocardial electrical and morphological remodelling. Several experimental studies have reported that atrial cardiac fibroblasts couple electrically with myocytes through gap junctions, exerting an electrotonic influence and causing atrial fibrillation. 11,12 Cyclic mechanical stretching of cultured cells (ie subjecting the cells to repeated stretching and relaxation) at all levels comparable with dynamic stretch overload in vivo has been well studied, including studies that investigated the molecular mechanisms of gene expression and signal transduction in cardiomyocytes, vascular smooth muscle cells and neonatal rat cardiomyocytes. [13][14][15] However, the effect of mechanical stretching on atrial cardiac fibroblasts has rarely been reported.
The relationship between miR-499, miR-208a and Bcl-2 in stretched atrial cardiac fibroblast and the haemodynamically overloaded heart has not been clearly established. Additionally, the molecular regulatory mechanisms underlying miR-499-and miR-208a-induced apoptosis of stretched atrial cardiac fibroblasts remain poorly understood. Therefore, the present study aimed to investigate the molecular mechanism of regulating miR-499 and miR-208a on Bcl-2 protein expression in stretched atrial fibroblasts and in a rat model of volume overload-induced heart failure.

| Primary culture of HCF-aa
HCF-aa, obtained from ScienCell Research Laboratories (San Diego, CA, USA), were cultured in a fibroblast growth medium (Catalog #2301, ScienCell Research Laboratories, CA, USA) containing essential and non-essential amino acids, vitamins, organic and inorganic compounds, hormones, growth factors, trace minerals and a low concentration (2%) of foetal bovine serum. The medium was buffered with HEPES and bicarbonate and had a pH of 7.4 when equilibrated in a humidified incubator containing 5% CO 2 at 37°C.

| In vitro cyclic stretching of cultured HCF-aa
The Flexcell FX-2000 strain unit (Flexcell International) consists of a vacuum unit linked to a valve controlled by a computer program.
HCF-aa cultured on a flexible membrane base was subjected to cyclic stretching by applying a sinusoidal negative pressure, and the peak negative pressure was 15 kPa. A 20% elongation cyclic stretch was performed at a frequency of 1 Hz (60 cycles/min) for various periods of time as previously description. 15

| Rat model aortocaval shunt
As mentioned previously, aortocaval (AV) shunts were placed in adult Wistar rats to induce volume overload. 16 Polyethylene catheters and a Grass tachograph preamplifier were used for haemodynamic monitoring. At the end of the shunt period (which lasted for a various number of weeks, according to the protocol), the rats were killed with an excess of isoflurane. Tissues were obtained from the left ventricle for Western blot analysis and miRNAs quantification as previously described. 17 All animal procedures were conducted according to our Institutional Committee of Animal Care and Use (protocol no. 0990816001) and conformed to the guidelines in the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (publication no. 86-23, revised 1996).
As for mRNA quantification, total RNA was extracted by

| Construction and delivery of miR-499 and miR-208a antagomirs and mutant microRNA expression vectors into cultured HCF-aa and ventricular myocardium
The 85-bp hsa-miR-499 precursor construct was generated as follows: First, genomic DNA was amplified with the for- The results indicated that the transfection efficiency is 30%.

| RNA interference
The HCF-aa were transfected with 800 ng Bcl-2 siRNA (Sigma, Singapore). According to a computer program provided by
For the mutant, the conserved site ACAAGATGGTAATCCGACTTAT was mutated into TGTTCAAGGAAATCGCAGAATA and constructed using the same method. All the cloned plasmids were confirmed through DNA sequencing (Seeing Bioscience Co. Ltd. Taipei, Taiwan).
The test plasmid (2 μg) was transfected with the gene gun in each well and then replaced with a normal culture medium. Following various periods of mechanical stretch exposure, cell extracts were prepared using the Nano-Glo dual-luciferase reporter assay system (Promega) and measured for luciferase activity by using a luminometer (Glomax Multi Detection System, Promega, Madison, WI, USA).

| Protein isolation
The HCF-aa subjected to stretching were harvested by scraping and then centrifuged at 300 g for 10 min at 4°C. The pellet was resuspended and homogenized in modified RIPA buffer, including 1× PBS, 0.3% SDS, 0.5% sodium deoxycholate and 1% NP-40, by centrifuging at 10,600 g (centrifugal force) for 20 min. A Bio-Rad protein assay was used to quantify the protein content of supernatant. Equal amounts of protein (30 μg) were loaded into a 10% sodium dodecyl sulphate (SDS)-polyacrylamide minigel and then electrophoresed.

| Western blot analysis
As described previously, the HCF-aa or ventricular cardiomyocytes were homogenized in a modified RIPA buffer.
The addition of mutant miR-208a had no effect on Bcl-2-3′UTR luciferase activity. These results suggest that the miR-208a binding site in the Bcl-2-3′UTR is essential for the transcriptional regulation induced by mechanical stretching.
F I G U R E 2 miR-499 overexpression inhibits pri-miR-208a luciferase activity under mechanical stress, whereas pre-treatment with antagomir-208a enhanced Bcl-2-3′UTR luciferase activity. A, Putative binding site for miR-499-5p on human pri-miR-208a transcript as predicted by luciferase assay. As shown in the schematic diagram, a pri-miR-208a matching site (orange rectangle) for miR-499-5p (blue rectangle) was located 57bp downstream of the pri-miR-208a gene loci. B, Quantitative analysis of wild-type and mutant pri-miR-208a luciferase activity under mechanical stress with miR-499 and mutant miR-499. C, Sequence of the Bcl-2-3′UTR target site for miR-208a binding, located at Bcl-2-3′UTR (nucleotide from 3542 to 3563). D, Quantitative analysis of wild-type and mutant Bcl-2-3′UTR luciferase activity under mechanical stretching with antagomir-208a and mutant miR-208a. Luciferase activity in the cell lysates was measured and normalized by Renilla activity using a dual-luciferase assay system (n = 5 per group). *P < .001 vs. control

| MiR-499 inhibited miR-208a expression in stretched HCF-aa
Compared with stretching alone, the miR-499 overexpression significantly inhibited the miR-208a expression in the HCF-aa that had been subjected to mechanical stretching for 1 h ( Figure 3A).
However, the miR-208a expression was not inhibited by the addition of mutant miR-499 or antagomir-499. The addition of antagomir-499 attenuated the miR-208a down-regulation with stretching for 4 h ( Figure 3B). These findings suggested that the down-regulation of miR-208a expression may have resulted from the miR-499 upregulation during mechanical stretching.

| Mechanical stretching affected the expression of Bcl-2 mRNA and protein via miR-208a
Compared with stretching alone, the addition of miR-208a significantly inhibited the expression of Bcl-2 mRNA and protein induced by mechanical stretching for 4 h ( Figure 3C-E), whereas the addition of mutant miR-208a or antagomir-208a had no effect on this Bcl-2 expression. Compared with stretching alone, Bcl-2 small interfering RNA (siRNA) inhibited the expression of Bcl-2 mRNA and protein induced by mechanical stretching for 4 h, whereas the addition of scrambled siRNA had no effect on this inhibitory effect.
Pre-treatment with wild-type miR-208a has no effect on the Bcl-2 expression in the HCF-aa without stretching.
Immunohistochemical staining confirmed the increased Bcl-2 expression in the HCF-aa after mechanical stretching for 4 h ( Figure 4A,B). The addition of wild-type miR-208a inhibited this Bcl-2 expression, whereas pre-treatment with mutant miR-208a or antagomir-208a had no effect ( Figure 4C-E). Wild-type miR-208a has no effect on the Bcl-2 expression in the HCF-aa without stretching ( Figure 4F). The addition of Bcl-2 siRNA inhibited the Bcl-2 expression in the HCF-aa after stretching for 4 h, whereas the addition of scrambled siRNA had no effect on this inhibition ( Figure 4G,H).
These results suggest that mechanical stretching for 4 h affects the Bcl-2 expression via miR-208a.

| Mechanical stretching-induced apoptosis was mediated by miR-499 and miR-208a in HCF-aa
Propidium iodide-annexin V double-staining and FACS analysis were used to evaluate apoptosis. The percentage of cells stained with both propidium iodide and annexin V after stretching for 1 h was significantly higher than in the control group ( Figure 5A,B).
The observed increases in propidium iodide and annexin V staining were significantly reversed by the miR-499 overexpression, whereas mutant miR-499 had no effect on apoptosis ( Figure 5C,D).

| MiR-499 regulated myocardial Bcl-2 expression via miR-208a in rats with AV shunts
The influence of miR-499 and miR-208a on Bcl-2 expression in vivo was examined in adult rats with aortocaval (AV) shunts to induce volume overload. The haemodynamic and echocardiographic parameters of the failing heart induced by AV shunt were summarized in Table S1. The AV shunt resulted in a significant increase in myocardial miR-208a expression from 3 to 5 days but decreased from 7 to 14 days ( Figure 6A). Additionally, miR-499 expression increased from 7 to 14 days ( Figure 6B). The wild-type miR-499 overexpression significantly reduced the miR-208a expression at 5 days, but the addition of mutant miR-499 and antagomir-499 had no effect on miR-208a expression ( Figure 6C).
Rats with AV shunts exhibited significant induction of myocardial Bcl-2 mRNA and protein expression from 7 to 14 days compared with sham rats ( Figure 6D-F). After 14 days with an AV shunt, the rats exhibited a significant increase in myocardial Bcl-2 mRNA and protein expression ( Figure 6G-I). Pre-treatment with miR-208a significantly inhibited this increase, whereas the addition of mutant miR-208a and antagomir-208a had no effect. Pre-treatment with Bcl-2 siRNA inhibited the expression of myocardial Bcl-2 mRNA and protein compared with the rats with AV shunts, whereas the addition of scrambled siRNA had no effect on this inhibition. These findings suggested that the miR-208a expression could be regulated by miR-499 and may regulate Bcl-2 expression in rat hearts with volume overload.

| D ISCUSS I ON
The major findings of this study were as follows: (a) mechanical stretching or an AV shunt placement first up-regulated and later down-regulated miR-208a expression in the HCF-aa or the rat Mechanical stretching promotes cardiac remodelling by increasing apoptosis. 18 Recent studies suggest that mechanical stretching simultaneously causes cardiomyocyte apoptosis and cardiac fibroblast proliferation, resulting in reduced myocardial contractility and increased fibrous tissue. [18][19][20] Cardiomyocyte fibrosis is a critical event in cardiac remodelling and may evolve into heart failure.
It has been reported that miRNA 1, 29b, 34a, 101, 122, 208a and 320 play a role in cardiomyocyte apoptosis, whereas miRNA 21,  The Bcl-2 protein principally determines whether a cell commits to apoptosis and hence has a crucial role in cellular development.
Consequently, the role of Bcl-2 protein has the potential role to affect multiple mechanism of cardiomyocyte apoptosis in response to various stress, including hypoxia, calcium dysregulation and oxidative stress. 27,28 The relationship between Bcl-2 and miR-208a was rarely reported in the past. In the present study, the Bcl-2 mRNA In conclusion, we found that miR-208a was a direct target of miR-499 and that the suppression of miR-208a relieved the suppression of Bcl-2 in mechanically stretched HCF-aa (Figure 7).
We also observed that antagomir-208a overexpression enhanced Bcl-2 expression in a model of heart failure induced by volume overload. The results of microRNA-mediated microRNA regulation are a potential new application of microRNA-directed cardiovascular therapy.

ACK N OWLED G EM ENTS
Not applicable.

CO N FLI C T O F I NTE R E S T S
The authors declare no conflict of interests.

AUTH O R CO NTR I B UTI O N S
SKC: Study design and manuscript draft. BWW, YJY and WJF: Most of the experiments. CML and KGS: Study design and manuscript revision. All authors provided clarification and guidance on the manuscript, were involved in editing the manuscript and approved the final manuscript.

E TH I C S A PPROVA L A N D CO N S E NT TO PA RTI CI PATE
All animal procedures were performed in accordance with the

CO N S E NT FO R PU B LI C ATI O N
Not applicable.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated or analysed during this study are included in this published article.