Inhibition of myocyte-specific enhancer factor 2A improved diabetic cardiac fibrosis partially by regulating endothelial-to-mesenchymal transition

Cardiac fibrosis is an important pathological process of diabetic cardiomyopathy, the underlying mechanism remains elusive. This study sought to identify whether inhibition of Myocyte enhancer factor 2A (MEF2A) alleviates cardiac fibrosis by partially regulating Endothelial-to-mesenchymal transition (EndMT). We induced type 1 diabetes mellitus using the toxin streptozotocin (STZ) in mice and injected with lentivirus-mediated short-hairpin RNA (shRNA) in myocardium to inhibit MEF2A expression. Protein expression, histological and functional parameters were examined twenty-one weeks post-STZ injection. We found that Diabetes mellitus increased cardiac MEF2A expression, aggravated cardiac dysfunction and myocardial fibrosis through the accumulation of fibroblasts via EndMT. All of these features were abolished by MEF2A inhibition. MEF2A gene silencing by shRNA in cultured human umbilical vein endothelial cells (HUVECs) ameliorated high glucose–induced phenotypic transition and acquisition of mesenchymal markers through interaction with p38MAPK and Smad2. We conclude that inhibition of endothelial cell-derived MEF2A might be beneficial in the prevention of diabetes mellitus-induced cardiac fibrosis by partially inhibiting EndMT through interaction with p38MAPK and Smad2.


INTRODUCTION
Diabetes mellitus (DM) can affect cardiac structure and function and lead to heart failure in the absence of atherosclerosis and hypertension, which is called diabetic cardiomyopathy (DCM) [1].Myocardial fibrosis is often present in end-stage heart failure caused by DCM.Although fibrosis, which is attributed to an excess deposition of extracellular matrix (ECM) components, is one of the most common pathological changes found in various organs, including the heart, the detailed mechanism responsible for this change remains unclear.Given the increased risk of heart failure in diabetic patients, a better understanding of the underlying mechanisms and additional therapeutic strategies would be of considerable value.
Studies have shown that endothelial-to-mesenchymal transition (EndMT) plays an important role in myocardial fibrosis [2].EndMT is considered to be a driving process resulting in the trans-differentiation of endothelial cells into mesenchymal cell types, characterized by a loss of cell-cell adhesion and a change in cell polarity and accompanied by a reduction in endothelial marker expression, such as vascular endothelial cadherin (VE-cadherin) and CD31, and an increase in mesenchymal marker expression, including fibroblast-specific protein-1 (FSP-1), α-smooth muscle actin (α-SMA), and vimentin [3] and so on.EndMT is stimulated by TGF-β2 through Smad, MEK (MAPK [mitogen-activated protein kinase]/ERK [extracellular signal-regulated kinase]), PI3K (phosphoinositide3-kinase), and p38 MAPK signaling pathways.Inhibitors of these pathways prevent TGF-β2induced EndMT [4].Increasing evidence has shown that high glucose levels can induce EndMT [5][6][7].However, the factors regulating EndMT under pathologic conditions of high glucose are not clear and remain to be elucidated.
Myocyte enhancer factor 2A (MEF2A) belongs to a family of four closely related transcription factors (MEF2A, -B, -C, and -D) that are conserved from yeast to humans [8].MEF2A functions during fetal development of the cardiovascular system and controls cell proliferation, differentiation, and death in both the developing fetus and the adult [3].Recent reports [9,10] have provided evidence of a role for hepatic stellate cell and myocardial cell MEF2A in fibrosis, although the precise functions of this transcription factor in endotheliocyte are still unclear.MEF2A is expressed in endothelial cells and is closely associate with angiogenesis [11].The overall expression pattern of MEF2A is similar to vascular endothelial growth factor receptor 2(VEGFR2) and Von Willebrand factor [12].
In addition, some studies indicated that bone morphogenetic protein2 (BMP-2), Smad2 [13], MAPKs p38 and ERK5 [14] interacted with MEF2A.Meanwhile, these proteins are important signal molecules of signaling pathways, which regulate EndMT.Therefore, we hypothesized that administration of MEF2A might achieve cardioprotective effects against fibrosis in diabetic hearts partially by regulating EndMT.To support our hypothesis, we performed a series of experiments both in vivo and in vitro.
Diabetes enhanced the expression of fibrotic markers collagen I and III as compared with the control group, whereas shRNA-MEF2A transfection in diabetic mice significantly reduced collagen levels as compared with vehicle treatment (Figure 3A-3E).The immunohistochemistry data were confirmed by western blot analysis results (p< 0.05; Figure 3F and 3G).

MEF2A knockdown inhibits myocardial fibrosis partially by suppressing EndMT in vitro and in vivo
To determine whether MEF2A inhibits high glucose (HG)-induced EndMT, we performed in vitro and in vivo experiments.
In vitro, HUVECs were treated with 33 mmol/L d-glucose with or without ShRNA-MEF2A for 5 days.Fluorescence microscopy revealed that control HUVECs showed a typical rounded or cobblestone shapes, and MEF2A silencing inhibited the change from a cobblestone-like to spindle-shaped feature with HG induction (Figure 4A).Immunofluorescence and western blot analysis demonstrated that HG-treatment significantly increased the levels of mesenchymal markers FSP-1, α-SMA, and vimentin and caused a reduction in the levels of endothelial markers CD31 and VE-Cadherin as compared with controls.However, the features above were alleviated when MEF2A was knocked down.GFP was used to evaluate transfection efficiency, which reached values of up to 90% (Supplementary Figure S1B).Immunofluorescence-colocalized staining showed that MEF2A silencing reduced expression of S100A4/FSP1 and α-SMA and increased expression of CD31.In the MEF2A knockdown group, the cells transfected by GFP had weak α-SMA expression, while those not transfected by GFP had strong α-SMA expression (Figure 4B and 4C).The effect of MEF2A on HG-induced EndMT was also detected by western blot.Mesenchymal markers (FSP-1, α-SMA, and vimentin) increased in the HG-group compared with the control group, and knockdown of MEF2A caused them to decrease (Figure 4D-4F).However, the opposite results were seen for the endothelial markers (CD31 and VE-Cadherin; Figure 4G-4H).

HG-induced EndMT is mediated by the translocation of MEF2A to the cytoplasm and interactions with p38MAPK and Smads in HUVECs
Immunofluorescence and western blot analysis revealed HG induced relocation of MEF2A in the cytoplasm of HUVECs, but the phosphorylation state of MEF2A located in the nucleus consistently and HG induced phosphorylation of MEF2A increasing (Figure 6A-6C).Conversely, Western blot analysis showed inhibition of phosphorylation of p38 by inhibitor SB203580 was sufficient to prevent the increase of MEF2A, p-MEF2A, FSP-1,α-SMA, and the decrease of CD31, VE-cadherin induced by HG (Figure 6D).To study the interaction between MEF2A and Smad2, CO-IP and PLA was performed to image proteinprotein interaction between MEF2A and Smad2.Co-Immunoprecipitation (CO-IP) results showed there was relationship between the MEF2A and Smad2 (Figure 6E).Interestingly, PLA showed that the dots representing MEF2A/Smad2-bingding proteins tended to localize in the nucleus in control group, and localized in the cytoplasm in HG group (Figure 6F).Western blot analysis showed that HG increase Smad2 phosphorylation in HUVECs as compared with control group, while MEF2A knockout made Smad2 phosphorylation declined (Figure 6G).Conversely, MEF2A expression decreased after smad2 knockdown by si-RNA.FSP-1 and α-SMA decreased, and CD31 and VE-Cadherin increased after smad2 knockdown before HG treatment (Figure 6H).

DISCUSSION
DM can deteriorate cardiac structure and function, which may lead to heart failure in the absence of coronary atherosclerosis and hypertension.However, the multifactorial nature of the disease remains incompletely understood.The present investigation showed that inhibition of MEF2A had a protective role in cardiac function associated with the mechanisms of EndMT and improved hyperglycemic-induced cardiac fibrosis.This is the first report demonstrating that MEF2A inhibition alleviates hyperglycemic-induced EndMT and diabetic myocardial dysfunction.
Myocyte enhancer factor 2 (MEF2) represents the second class of transcriptional factors that regulate expression of many muscle-specific, growth factorinduced, and stress-induced genes [15].MEF2A is a key nuclear mediator that may participate in the pathological remodeling and accumulation of focal fibrosis in hypertrophic cardiomyopathy [16,17].Indeed, a recent study demonstrated that characterization of MEF2A knock-out mice revealed severe myofibrillar defects in cardiac muscle, indicating a requirement for MEF2A in cytoarchitectural integrity [18].However, another study showed that MEF2A overexpression is sufficient to induce cardiac hypertrophy, and dominant negative inhibition of MEF2A signaling blocked cardiomyocyte hypertrophy [19,20].In our present study, MEF2A expression was upregulated in the diabetic myocardium.MEF2A inhibition improved cardiac function and remodeling in diabetic mice.Thus, prolonged activation of MEF2Adependent genes in myocytes may become maladaptive, contributing to pathological remodeling and accumulation of focal fibrosis in diabetes-induced cardiomyopathy.However, the role of MEF2A in diabetic myocardial fibrosis in endothelial cells is still unknown.
The most important pathological feature of DCM is hyperglycemia-induced excess production of ECM, mainly collagen types I and III, which can alter the structure and function of the heart [21].The immunohistochemistry and western blot analysis results showed that knocking down MEF2A made a significant reduction in collagen deposition and expression of collagen I and III in the perivascular and intramyocardial regions of the diabetic mouse myocardium.Thus, hyperglycemia leading to MEF2A activation is an essential mechanism that may contribute to myocardial remodeling and fibrosis in DCM, and silencing MEF2A may have therapeutic potential in ameliorating these processes.
Chronic hyperglycemia is an important initiator of EndMT [5].Recently, the contributions of EndMT to cardiac fibrosis have been reported [2,22].EndMT can contribute to the progression of multiple diseases, including diabetic cardiomyopathy, diabetic nephropathy, and hypertensive cardiomyopathy [5,23,24].In the present study, we observed decreased levels of the endothelial marker CD31 and VE-cadherin with HG treatment and increased expression of the fibroblast markers FSP-1 and α-SMA.We then observed colocalization of the endothelial marker CD31 and fibroblast markers FSP-1 and α-SMA in DM arteries.The disaggregated endothelial cells start to alter their structure, exhibit a decrease in endothelial marker expression, and acquire mesenchymal characteristics.This cascade of events reveals the importance of preserving endothelial cell integrity by suppressing MEF2A activity, thereby preventing the initiation of DM-induced EndMT.We further observed that MEF2A plays an important role in these endothelial cells that are undergoing phenotypic transition and beginning to acquire fibroblast characteristics, a process called EndMT in the late stages of DM.Our results indicated that elevated levels of MEF2A may be a trigger of hyperglycemia-induced EndMT and as a profibrotic factor in diabetes-induced cardiomyopathy.
Similar to other published reports [2,5], we observed that only ~20% to 30% of fibroblasts coexpressed both endothelial markers and fibroblast markers in the hearts of diabetic mice, and we did not evaluate the exact contribution of proliferating resident mesenchymal cells and circulating bone marrow-derived fibroblasts to cardiac fibrosis, which is a limitation in our study.MEF2A is primarily considered to be a transcription factor having a common amino-terminal DNA binding domain and playing pivotal roles in cardiac, muscle, and neuron gene expression [25].However, previous studies have demonstrated that MEF2A exhibits regionally specific cytoplasmic expression in rodent forebrain [26].
To further evaluate the functions of MEF2A under hyperglycemic conditions in vitro, we examined the subcellular localization of MEF2A in HUVECs and clearly showed that HG induced higher expression levels of MEF2A and translocation into the cytoplasm from the nucleus.Moreover, HG only increased the activity of MEF2A (phosphorylation) in the nucleus.In line with this observation, the results of the western blot analysis showed a similar result.HG may enhance MEF2A transcriptional activation and regulate diabetic symptoms.The major MAPK signaling cascades ERK1/2, JNK, and p38 MAPK are strongly activated by hyperglycemia and MEF2Ais a nuclear target for the p38MAPK signaling pathway [27,28].In our study, we found that p38 MAPK inhibition decreased MEF2A transcriptional activity in the nucleus of HUVECs with high glucose medium (HG) than normal medium (NG) treatment.Thus, the HG-MAPK-MEF2A pathway is an essential mechanism that may contribute to myocardial remodeling and fibrosis in DCM.MEF2A may mediate cardiac remodeling via multiple mechanisms depending on the underlying pathological condition.The TGF-β superfamily and the Smads with their downstream receptors and signal transducers are the major regulators of EndMT, and they are the main reasons for fibroblast activation, which sequentially leads to myocardial fibrosis and diastolic dysfunction [29][30][31].We found that MEF2A silencing could markedly inhibit HGinduced Smad2 expression and activity and also attenuated HG-induced EMT in HUVECs.These results contrast with a previous study demonstrating the function of Smad2 as a co-modulator for MEF2 transcriptional regulatory proteins [13].
In summary, we illustrate that cardiac MEF2A silencing may protect against cardiac fibrosis and improve myocardial function in diabetic mice by regulating EndMT.The mechanism may be involved with the translocation of MEF2A to the cytoplasm and interactions with p38MAPK and Smad2.Given the cardioprotective effects of MEF2A silencing, MEF2A may be a potential therapeutic target for diabetic heart diseases.

Animal model and experimental protocol
C57BL/6J wild-type (WT) male 8-week-old mice (23-28 g) were used for the in vivo experiments.Type 1 diabetes mellitus was induced by intraperitoneally injections of streptozotocin toxin (STZ; Sigma, St. Louis, MO) dissolved in citrate buffer (pH 4.5) at 60 mg/ kg body weight for 5 consecutive days.Control mice (n=25) were injected with citrate buffer only.Mice with randomly measured glucose levels of 20mmol/L 7 days after STZ injection were considered diabetic.Blood glucose was measured using an Accu-Check Active glucometer (Roche).The diabetic mice did not receive any insulin treatment.Thirteen weeks post-STZ injection, the diabetic mice were randomly divided into 3 groups: diabetes mellitus (DM) (n=30), lentivirus-mediated green fluorescent protein of MEF2A interference NC (LV-GFP[-]) (n=30), and lentivirus-mediated MEF2A interference (LV-MEF2A[-]) (n=30).Lentivirus was administered directly to the heart by intramyocardial injection.
The salient steps of delivering lentivirus into the left ventricular wall of the mouse involve administration of anesthesia, intratracheal intubation, incision to open the chest and expose the heart, and delivery of lentivirus by a sterile 30-gauge needle and a precision microliter syringe.For treatment, an amount of 1×107 UT / 30μl of lentivector with MEF2A shRNA or the same volume of lenti-vehicle were injected into 3 sites of the left ventricle.The recombinant lentivirus vector containing a green fluorescent protein (GFP) reporter for measuring transfection efficiency (Supplementary Figure S1A) All mice were given free access to a normal diet and water.Mice were humanely euthanized and evaluated after twenty-one weeks post-STZ injection.All experiments conformed to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health and Shandong University.

Cell culture and RNA interference
Human umbilical vein endothelial cells (HUVECs) were purchased from American Type Culture Collection.Cells were grown to confluence in endothelial cell medium (ECM) supplemented with 5% fetal bovine serum and 1% endothelial cell growth supplement.Cells were cultured in a humidified 5% CO 2 incubator at 37°C and used between the fourth and sixth passages.Cells were treated with 5 or 33 m mol/L d-glucose.The medium was changed every 48 h for 5 consecutive days.Before glucose treatment, the HUVECs were infected with lentivirus at a multiplicity of infection (MOI) of 10 for 24h.For Smad2 inhibition, cells were transfected with small interfering RNA (siRNA) of Smad2 or a nontarget gene using Lipofectamine 2000 reagent (Invitrogen) according to the manufacturer's instructions.Optimal knockdown of Smad2 was obtained by 4h incubation with siRNA.

Cardiac function measurement
Cardiac diameter and function was measured under 2.0 % isoflurane anesthesia by transthoracic echocardiography using Vevo770 imaging system (Visual www.impactjournals.com/oncotargetSonics, Toronto, Canada) with a 10-MHz probe.M-mode tracing was recorded at the level of the papillary muscles.Lift ventricular end-diastolic dimension (LVEDD), left ventricular end-systolic dimension (LVESD), and left ventricular diastolic posterior wall thickness (LVPWd) were measured.Percentage of left ventricular ejection fraction (LVEF) was calculated as 100×[(LVvol d -LVvold s )/LVvol d ] and percentage left ventricular fractional shortening (LVFS) was calculated as 100×[(LVEDD -LVESD)/LVEDD].Pulsed-wave Doppler echocardiography was used to measure the ratio of peak early diastolic ventricular filling velocity to peak atria filling velocity (E/A).And the ratio of diastolic mitral annulus velocities (E'/A') were measured in tissue Doppler imaging.E/E' ratio was calculated.

Endothelial cells isolated from the heart
Hearts were explanted twenty-one weeks post-STZ injection and cardiac cells were dissociated using a gentle MACS Dissociator (Miltenyi Biotec) as described by the manufacturer.Endothelial cells were isolated using positive selection by magnetic affinity cell sorting using a CD146 antibody (Miltenyi Biotec) and were then used for RNA isolation and RT-PCR.

Histological analysis
After the physiological analysis, mice were sacrificed for sections.To examine cardiac fibrosis, heart sections were stained with Masson's trichrome (MTC) and Picrosirius red.The intramyocardial and perivascular region fibrotic area was measured from all groups in every 5 randomly chosen views of each sample and analyzed by the Image-Pro Plus 6.0 program.Perivascular fibrosis was calculated as the ratio of the fibrotic area surrounding the vessels to the total vessel area [32].

Co-Immunoprecipitation (CO-IP)
For immunoprecipitation of MEF2A and Smad2, HUVECs were lyzed in Lysis Buffer (P0013) containing PMSF.Supernatant was incubated with MEF2A, Smad2 and IgG antibody for 4°C overnight, and incubated with Protein A/G sepharose beads for 6h.Beads were washed with lysis buffer three times, incubated for 5 min at 95 °C with 1X loading buffer and subjected to SDS-PAGE and Western Blot analysis.

Duolink ® in situ-proximity ligation assay
HUVECs were fixed in 4% formaldehyde and incubated with antibodies (anti-MEF2A [mouse origin], Smad2 [rabbit origin]) at 4°C overnight.The cells were subsequently incubated with the PLUS and MINUS PLA probes (Duolink® In Situ PLA® Probe Anti-Mouse PLUS Affinity-purified Donkey anti-Mouse IgG (H+L), Sigma, DUO92001; Duolink® In Situ PLA® Probe Anti-Rabbit PLUS Affinity-purified Donkey anti-Rabbit IgG (H+L),Sigma DUO92002) at 37°C for 1h.Following incubation, the cells were washed and incubated in the Ligation-Ligase solution for 30 min at 37°C.Following an additional wash, the cells were incubated in Amplification-Polymerase solution for 100 min at 37°C.The samples were mounted using Duolink In Situ Mounting Medium with DAPI (Sigma, DUO82040), allowed to adhere for 15 min, and analyzed using a confocal microscope.

RT-PCR
Primer details are shown in Supplementary Table S1.SYBR Green RT-PCR and quantitative assays involved the use of a sequence detector system (IQ5 Real-Time PCR cycler; Bio-Rad Laboratories, CA, USA).Quantitative values were obtained using the threshold cycle (CT) values.Relative mean fold changes in expression were calculated by the 2 -△△CT method.

Statistical analysis
Results are presented as mean ± SEM.Statistical analyses were performed using the Student's t-test for direct 2-group comparisons and the Tukey-Kramer test following significant one-way ANOVA test for multiplegroup comparisons.Data analysis was carried out by SPSS v16.0 software (SPSS Inc., Chicago, IL) on the results of at least 3 independent experiments.A value of p <0.05 was considered statistically significant.

Figure 1 :
Figure 1: MEF2A expression and Echocardiography measurements of mouse hearts.A. Quantitative RT-PCR analysis of MEF2A mRNA expression; B. Western blot analysis of MEF2A in hearts from all groups; C. Typical echocardiographic images of 2D echocardiograms, M-mode echocardiograms, pulsed-wave Doppler echocardiograms and tissue Doppler echocardiograms in all groups; D-J.Quantitative analysis of echocardiographic measurements parameters.(D) Left ventricular ejection fraction (LVEF).(E) Fractional shortening (FS).(F) Early to late mitral flow (E/A).(G) Early to late ratio of diastolic mitral annulus velocities (E'/A').(H) Early mitral flow ratio to tissue Doppler-derived early mitral annulus velocity (E/E').(I) Left ventricular posterior wall thickness at diastole (LVPWd).(J) Left ventricular end-diastolic dimension (LVEDD).Data are mean±SEM.*p<0.05 compared with control; # p<0.05 compared with vehicle treatment group.n= 8 to 10 mice for each group.

Figure 2 :
Figure 2: MEF2A inhibition limits diabetes-induced myocardial fibrosis.A. Interstitial and perivascular fibrosis were analyzed by Masson's trichrome (blue staining) and Picrosirius Red staining (red and yellow staining) after 20 weeks of diabetes mellitus in all groups mouse.Scale bars=40 μm.B. Quantification of interstitial fibrosis after 20 weeks of diabetes mellitus.C. Quantification of perivascular fibrosis after 20 weeks of diabetes mellitus.Data are mean±SEM.*p<0.05 compared with control; # p<0.05 compared with vehicle treatment group.n= 8 mice for each group.

Figure 3 :
Figure 3: MEF2A inhibition limits diabetes-induced collagen deposition in a diabetic mouse model.A. Immunostaining of collagen I and III in the interstitial and perivascular region, respectively (scale bars=40 μm); B. and C. Quantification analysis of collagen I in the interstitial and perivascular region; D. and E. Quantification analysis of collagen III in the interstitial and perivascular region, respectively; F. and G.Western blot analysis of collagen I and collagen III; Data are mean ± SEM. *p< 0.05 compared with control; #p< 0.05 compared with vehicle treatment group; n= 8 mice for each group.

Figure 5 :
Figure 5: MEF2A inhibition limits cardiac EndMT in a diabetic mouse model.A. Double immunofluorescence staining of antibodies to CD31 (Red) with antibodies to S100A4/FSP-1(Green) and α-SMA (green) in coronary arterioles of all groups mouse.Colocalization of CD31 with S100A4/FSP-1 and α-SMA expression in coronary arterioles is shown in yellow.DAPI (Blue) was used to stain nucleus.Scale bars=40 μm.B. and C. Representative z-stack image analysis shows specific overlay of double immunostaining, CD31+/S100A4+ cells in specific ordinate were analyzed in z stack with optimal interval range of 0.8μm.D. and E. The percentage of S100A4+ CD31+ cells and α-SMA+ CD31+ cells in diabetic hearts.F-I.RT-PCR analysis shows CD31, VE-Cadherin, FSP-1 and α-SMA in sorted cardiac endothelial cells by magnetic affinity cell sorting using a CD146 antibody.Data are mean ± SEM. *p<0.05 compared with control; #p<0.05 compared with vehicle treatment group.n=12 mice for each group.

Figure 6 :
Figure 6: HG-induced EndMT is mediated by the translocation of MEF2A to the cytoplasm and interactions with p38 MAPK and smads in HUVECs.A. and B. Immunofluorescence analysis of MEF2A and phosphorylated-MEF2A localization in HUVECs.MEF2A is stained green, nuclei are stained blue with DAPI (scale bar =3 μm); C. Western blot analysis of MEF2A and phosphorylated-MEF2A expression levels in cytoplasm and nuclear proteins of HUVECs; D. Western blot analysis of phosphorylated-p38, p38, phosphorylated-MEF2A, MEF2A, CD31, VE-Cadherin, FSP-1 and α-SMA in HUVECs after HG treatment with or without SB203580; E. CO-IP analysis of the protein-protein interaction between MEF2A and Smad2.F. PLA analysis of the protein-protein interaction between MEF2A and Smad2.The purple-stained dots represent MEF2A/Smad2-binding proteins.Nuclei are stained blue with DAPI (scale bar =3μm); G.Western blot analysis of MEF2A, phosphorylated-Smad2 and Smad2in HUVECs after HG treatment with or without MEF2A knockdown; H.Western blot analysis of Smad2, MEF2A, and EndMT markers in HUVECs after HG treatment with or without Smad2 knockdown by siRNA.