KDM3A is a Positive Regulator of Cardiac Fibroblasts Conversion that Increases Smad3 Phosphorylation Following TGFβ1 Stimulation


 The epigenetic molecule KDM3A has been shown to be involved in improving cardiovascular diseases, but its effect on cardiac fibroblasts (CFs) remains unclear. Thus, we designed the gain- and loss-of-function experiments to investigate the biological functions of KDM3A in cardiac fibroblasts (CFs). Moreover, we added a SIS3-HCL (a specific inhibitor of p-Smad3) to explore the underlying mechanism. The cells viability and migration were verified by CCK-8 and cell migration experiments, respectively, and the degree of fibrosis was measured by Western blot analysis. Our data reveal that KDM3A enhance the proliferation and migration of cardiac fibroblasts, meanwhile, increasing the fibroblast-to-myofibroblast transition, while enabling Smad3 phosphorylation response to TGFβ1 stimuli. However, these results could be abolished by SIS3-HCL, an inhibitor of the p-Smad3. Furthermore, KDM3A inhibition obviously protect cardiac fibroblasts conversion against TGFβ1 stimuli. These results identify that KDM3A may be a novel regulator of the cardiac fibroblasts conversion, through its ability to modulate phosphorylation of Smad3 following TGFβ1 stimuli.


Introduction
Cardiac brosis is involved in the pathogenesis of most myocardial diseases [1] . Despite tissue homeostasis requires the maintenance of collagen bers, sustained chronic injury and microenvironmental perturbation will reverse the benign tissue repair into unrestrained or dysregulated brotic scar, which can eventually lead to myocardial structural and pathophysiological alterations and heart failure [2,3] . Cardiac broblasts (CFs) are responsible for extracellular matrix (ECM) homeostasis.
Excessive ECM deposition can further active resident cardiac broblasts differentiate into myo broblast, which in turn exacerbate secretion of ECM [4] . There are multiple mechanistic pathways involved in the conversion of myocardial broblasts into myo broblasts (CMF). However, e cacious therapies to prevent excessive cardiac brosis are lacking. Therefore, identifying novel options are needed.
KDM3A, also known as JMJD2A, plays a role as a histone methyltransferase by demethylating mono-or dimethylated H3K9, has a close relationship of myocardial tissue repair and remodeling, partially through maintaining in ammation, Foxidative stress, and neovascularization [5,6] . Previous study shows that KDM3A stimulation could exacerbate pathological brosis in cardiac hypertrophy [7] , however, the role of KDM3A and its underlying mechanism in CFs has never been addressed thoroughly. This present study was undertaken to explore the cardiac broblast differentiation induced with TGFβ1 by KDM3A administration or inhibition. Further, we determined investigated the molecular mechanism.

Adenovirus construction and transfection
The adenovirus that encoding the full-length rat KDM3A sequence (AdKDM3A) and KDM3A-siRNA (AdshKDM3A) were validated to signi cantly regulate the protein expression of KDM3A in previous studies [5,6,8] . And western blots for protein expression analysis were implemented to determine whether the adenovirus was successfully transfected in these experiments.

The culture and identi cation of CFs
Primary cardiac broblasts were isolated from neonatal Sprague-Dawley (SD) (n=10, 1-3 days old) rats according to the differential centrifugation method as previously described [9] . CFs were routinely at passages 3-4 for subsequent experiments.
The cell purity was analyzed by anti-vimentin (Abcam) and anti-α-SMA((MilliporeSigma) antibodies in an immuno uorescence containing experiment. The ratios of α-SMA and Vimentin staining were observed under a uorescence microscope to identify the purity of CFs, and cells that negative for α-SMA and 95% positive for Vimentin were used for the subsequent experiments [10,11] .

Experimental Design
For the loss-of-function experiments, CFs were randomly classi ed into 4 groups (n=3/group) in the following order: 1) Control + AdshRNA group: transfected by siRNA group; 2) Control +AdshKDM3A group: transfected by siRNA-KDM3A group; 3) TGF-β1+ AdshRNA group: transfected by siRNA in the response to TGF-β1; 4) TGF-β1+ AdshKDM3A group: transfected by siRNA-KDM3A in the response to TGF-β1. We further investigated KDM3A gain-of-function experiments, the experimental group were the same as the loss-of-function experiments, but in order to explore the underlying mechanism, we added a SIS3-HCL (a speci c inhibitor of p-Smad3)( Selleck) as previous described [12] in TGF-β1+ AdshKDM3A group as another group. Next, the cells viability and migration were veri ed by CCK-8 and cell migration experiments, respectively, and the degree of brosis was measured by Western blot analysis.

Immunofluorescence staining
To evaluate the degree of cardiac brosis, we performed α-SMA and Vimentin-speci c immuno uorescence staining of CFs. According to a previous study [13] , CFs were harvested, xed with 4% paraformaldehyde, and blocked with 3% goat serum for 1 h. Subsequently, anti-α-SMA and anti-vimentin primary antibodies were applied to the cells, followed by Alexa Fluor-conjugated secondary antibodies. Nuclear staining was achieved using DAPI. The cells were visualized and photographed under a uorescence microscope.

CCK-8 assay
Cardiac broblast proliferation was analyzed using the CCK-8 assay as described in detail previously [10] .
Brie y, CFs were digested and reseeded in a 96-well plate at a density of 1´10 3 cells per well, 100 µl of medium was added, and the cells were cultured for 48 h. Ten microliters of CCK-8 reagent (Dojindo Molecular Technologies) was added to each well and incubated for 4 hours at 37°C. Finally, the optical density (OD) values were measured at 450 nm using a microplate reader.

Cell migration assay
Cell migration was determined using a 24-well Transwell plate (Corning Costar). Brie y, CFs were resuspended in the upper chamber of the plate with serum-starved. And culture medium containing 10% FBS was added to the lower chamber as a chemoattractant, then incubation at 37 °C for 24 h. The migrated cells were counted in ve randomly chosen elds at magni cation of 200× by light microscopy (Olympus Corporation, Tokyo, Japan), and analysis by Image J 1.51w software.

Western blot analysis
After CFs were harvested, total protein was extracted, and the protein concentration was measured using a BCA kit according to the manufacturer's instructions. Equal amounts of protein were subjected to SDS-PAGE and transferred to polyvinylidene di uoride membranes. The membranes were incubated with the appropriate primary antibodies and then probed with peroxidase-conjugated secondary antibodies for 2 h at room temperature. Finally, the results were analyzed with an ECL detection system, the bands were visualized with ImageLab 5.2 software, and the intensities were normalized to GAPDH. Antibodies against Collage I, α-SMA, CTGF vimentin, DDR2, KDM3A and GAPDH were obtained from Abcam. Antibodies against Smad3 and p-Smad3 were obtained from CST.

Statistical analysis
Statistical analyses were performed using GraphPad Prism 7.0 software (GraphPad Software Inc, USA).
All data were expressed as the mean ± standard deviation (SD). Intergroup differences were compared using ANOVA. Statistical signi cance was set at p < 0.05.

KDM3A expression in CFs.
Firstly, we identi ed the phenotype and purity of CFs after 2-3 generations. Previous studies demonstrated that vimentin was surface-speci c antigen of cardiac broblasts [14] , and α-SMA antibodies was surface marker of myo broblasts [15] . As shown in Figure 1a, we successfully isolated highly puri ed primary rat CFs by immunofluorescence analysis. Ninety-ve percent of cells were stained with vimentin (red), almost no cells were stained with α-SMA (green), indicating that the isolated CFs could be utilized for subsequent experiments.
Next, we constructed relevant adenovirus to administered KDM3A expression. CFs were infected with KDM3A(AdKDM3A) or siRNA-KDM3A(AdshKDM3A). The results showed CFs transfected with AdKDM3A or AdshKDM3A (Figure 1b) emitted strong green uorescence (GFP), indicating that target virus vectors were successfully transfected into the CFs. Then, we further measured the protein expression of KDM3A by Western blot analysis, and the results indicated that KDM3A expression was signi cantly enhanced after transfection with AdKDM3A, while AdshKDM3A transfection blocked KDM3A expression (Figure 1cd). These data indicated that the target adenovirus we constructed effectively regulated the protein expression of KDM3A in CFs.
As we all know, TGFβ/Smad signaling as a classic pro brotic pathway that increase broblast abundance and cardiac broblast differentiation is a key event in cardiac remodeling [16,17] . Moreover, TGFβ1 is the most common subtype of TGFβ. Thus, rstly, we veri ed the effects of TGFβ1 on cardiac broblast differentiation, and expression levels of vimentin, DDR2, α-SMA and CTGF were used as standard indexes of cardiac broblast differentiation [18] . Figure 2a-b showed that CTGF and α-SMA were signi cantly upregulated, while Vimentin and DDR2 were signi cantly downregulated response to TGFβ1 treatment for 24 h. As expected, we demonstrated that TGFβ1 could induce cardiac broblast differentiation. Importantly, consistent with the previous researches, KDM3A might participated in cardiac brosis [19] . In present study, we found that compared to control group, KDM3A level was signi cantly elevated (Figure 2c-d). Taken together, these results suggest that TGF-β1 promotes cardiac broblast differentiation and signi cantly enriches KDM3A expression. Based on the above observations, we speculated that KDM3A might be a key regulation factor during the course of cardiac broblast differentiation. To manifest this hypothesis, KDM3A loss-and gain-of-function assays were preformed in CFs.

KDM3A disruption signi cantly in uences CFs proliferation and migration.
Based on above observations, we further explored whether KDM3A can intervene in speci c biological functions of broblasts. First, we designed CCK-8 and migration experiments to evaluate broblast proliferation and migration function, respectively. The results showed that loss of KDM3A markedly suppressed CFs proliferation and migration under the stimulation of TGF-β1 for 24 h. (Figure 3a, c, d); however, when KDM3A was upregulated, the opposite tendency was observed, CFs proliferation and migration were signi cantly increased (3b, c, d). Collectively, suppression of KDM3A in CFs could abrogate the proliferation and migration, but overexpression of KDM3A had the opposite consequences.
3.4 KDM3A is a positive regulator of cardiac broblast differentiation, but KDM3A inhibition can reverse myo broblast conversion.
To determine whether KDM3A is involved in cardiac broblast differentiation, we analyzed the changes in α-SMA, collagen I, and CTGF expression in response to changes in KDM3A expression. CFs were transfected with siRNA-KDM3A to knock down its expression, and transfected with AdKDM3A to active its expression. The results showed that CFs conversion was decreased by KDM3A downregulation that the expression of collagen I, CTGF and α-SMA in broblast was signi cantly lower than that in the control group under the stimulation of TGF-β1 for 24 h (Figure 4a). Consistent with the previous studies, after KDM3A upregulation, the conversion of CFs was obviously aggravated (Figure 4b). These data indicate that KDM3A inhibition could protect CFs against the conversion, but overexpression of KDM3A could further exacerbate CFs differentiation.
3.5 KDM3A regulates the biological function of CFs through the Smad3-dependent manner.
It is obviously noted that cardiac broblast differentiation is pivotal for the pathogenesis of cardiac brosis and hypertrophy [11,20] . However, the underlying molecular mechanisms of cardiac broblast differentiation are still confused. Since CFs were featured with an enhanced expression of KDM3A induced by TGF-β1 for 24 h (Figure 2c), we further illustrated KDM3A participate in regulating functionalities of CFs ( Figure 3). Additionally, we found that high expression of KDM3A could also simultaneously increase Smad3 phosphorylation, and Smad3 level had no signi cant change (Figure 4d). Nevertheless, after knocked down KDM3A by transfected siRNA-KDM3A, Smad3 phosphorylation was suppressed evidently by treated with TGF-β1 (Figure 4c). These ndings prompted us to assume that KDM3A may be induced by TGF-β1 signaling, and then regulating functionalities of CFs by Smad3 signaling.
To con rm this hypothesis, CFs were treated with SIS3-HCL (a speci c inhibitor of p-Smad3), followed by TGF-β1 induction. It was remarkably noted that p-Smad3 de ciency by SIS3-HCL could both signi cantly abolish KDM3A positive effect, which was accompanied with weakened cardiac broblast differentiation related proteins (collagen I, CTGF, α-SMA) induced by TGF-β1 (Figure 4b, d), indicating that Smad3 were indispensable factors for KDM3A alter cardiac broblast differentiation in response to TGF-β1. Moreover, consistent with the results of cardiac broblast differentiation, we found that p-Smad3 inhibition by SIS3-HCL was indeed impair CFs proliferation and migration ability (Figure 3 b,c,d). Altogether, these data corroborated that overexpression of KDM3A aggravated cardiac broblast differentiation induced by TGF-β1 was a Smad3-dependent manner.

Discussion
In the present study, we demonstrated a novel function for the rst time that KDM3A increases cardiac broblast differentiation, improves CFs migration and proliferation in response to TGF-β1 stimulation. We also observed that these positive effects were suppressed by the p-Smad3 inhibitor, which may be a compensatory response against cardiac brosis. By contrast, KDM3A de ciency could reverse these effects, protect CFs from TGF-β1 induced cardiac broblast differentiation. These results further suggest that KDM3A exhibits its pro-brotic effects at least partly be relieved by suppressing Smad3 signaling pathway in response to TGF-β1 stimulation. Consistently, KDM3A de ciency may contribute to its ability to prevent collagen deposition and cardiac brosis. These ndings may open a new horizon in cardiac brosis disease therapy.
The accumulation of collagen bers is widespread in the myocardial interstitium of most chronic heart diseases [21] . Cardiac remodeling caused by myocardial brosis will eventually become a decisive factor for heart failure and its clinical progression, importantly, the more interstitial brosis deposited, the more likely it is to increase the mortality of patients [22] . Thus, it is urgent need to explore strategies to inhibit cardiac broblasts proliferate and phenotype conversion which were critical evolution of cardiac brosis.
In the current study, we observed KDM3A elevate TGF-β1 induced CFs migration, proliferation and cardiac broblast differentiation, but suppressed by p-Smad3 inhibitor, indicating that KDM3A may generate probrotic effects by Smad3 signaling in TGF-β1 stimuli. Thus, the involvement of the TGF-β1/KDM3A/Smad3 axis may partly perform the differentiation of broblasts into myo broblasts. However, the more complex and warrants in these pathological processes need further investigation.
While the recent studies of KDM3A provide a powerful evidence for the investigation of the complicated pathogenesis of cardiovascular diseases and molecular mechanisms involved [5,8,19] , the role of KDM3A in the regulation of cardiac brosis and functions remain elusive. The Gene Expression Omnibus (GEO) database revealed that KDM3A can target TGF-β1, thereby regulating the TGF-β1/Smad signaling pathway and affecting the progression of brosis [23] . Previous studies targeting cardiac resident cells, such as cardiomyocyte, endothelial cells and macrophages, KDM3A have generated several controversial results. Depletion of KDM3A in endothelial cells dampened vascular in ammation by inhibiting CSF1 transcription and macrophage recruitment [24] . However, by contrast, our previous study demonstrated that KDM3A improved cardiac performance and alleviated cardiac brosis in MI-induced myocardial injury by mitigating myocardial apoptosis [25] . Perhaps, we speculate that the different effects of KDM3A in cardiovascular diseases may vary depending on diverse pathological conditions. In general, KDM3A has potential myocardial damage effects following cardiac hypertrophy through both pro-brosis and proin ammation mechanisms [26][27][28] .
In addition, previous studies have shown that crosstalk between Brg1 and KDM3A via co-regulate procholesterogenic transcription to contribute to SREBP2-dependent cholesterol synthesis in hepatocyte [29] . Interestingly, excessive synthesis of cholesterol may become the pathological basis of cardiac hypertrophy and atherosclerosis. Thus, these results not only raise novel insights into the understanding of the role of KDM3A underlying cardiac hypertrophy, but also provide a potent amenable therapeutic approach to improve myocardial brosis induced cardiac hypertrophy. Importantly, KDM3A, a ubiquitous histone lysine demethylase, functions as epigenetic modi cations in regulating tissue brosis [19,30] . Indeed, emerging evidence has demonstrated that KDM3A silencing equivalently attenuated HG-induced CTGF induction in renal tubular epithelial cells by decreasing dimethylated H3K9 recruit to the CTGF promoter to suppress transcription [30] . Unfortunately, following above studies only con rmed KDM3A participate in modulating myocardial brosis in animal models. Therefore, we explored whether KDM3A exerts a similar pro-brotic effect directly in TGFβ1-stimulated CFs in vitro. Consistently, the current study indicated that KDM3A signi cantly induced CFs migration, proliferation and cardiac broblast differentiation, which was attenuated by p-Smad3 inhibitor treatment.
TGFβ1 is a recognized brogenic cytokine, which bind to TGFβR1 receptor before phosphorylated by TGFβR2, and then induce canonical pathway (mainly Smad2/3 proteins), leading to collagen production [31,32] . Interestingly, in this study, KDM3A knockdown in primary CFs resulted in signi cant down-regulation of pro-brotic genes and predominantly decreased Smad3 phosphorylation in response to TGFβ1. Moreover, KDM3A was upregulated by TGF-β1 stimuli. Collectively, KDM3A may act as a mediator between TGF-β1 and Smad3, and jointly mediate the pathological process of myocardial brosis. Supporting these ndings, KDM3A was robustly delayed renal brosis degree via elevate Smad2/3 phosphorylation [33] . Accumulating evidence indicates that Smad3 is a key intracellular regulatory protein in the TGF-β superfamily signal cascade [17] . On the one hand, Smad3 is involved in regulation of cardiac broblast phenotype and function [34] . On the other hand, Smad3 has been suggested to promote the transformation of cardiac broblasts into myo broblasts, then inhibiting protease-mediated ECM degradation to maintain a stable microenvironment, thus, affecting the formation of tissue scars [35,36] . Besides, attenuation of Smad3 phosphorylation and disturbing SMAD3/SMAD4 complex formation obviously abrogated expression of hepatic XPO4, which associates with severity of brosis in metabolic associated fatty liver disease [37] . In keeping with an expanding body of evidence, we found that cardiac broblast phenotype and function were signi cantly modulated in the KDM3A activation condition, accompanied with Smad3 phosphorylation increases. Additionally, the available results suggested that Overexpression KDM3A in cardiomyocytes obviously promoted the transcription of the brosis-related factor Timp1, and further exacerbating the pathological progression of cardiac brosis and ventricular remodeling [7] . This view coincides with the therapeutic mechanism of KDM3A, which corporately shows that it plays a multi-channel regulatory role in cardiac brosis, but targets the Smad3 signaling pathway in the present study.

Conclusion
In summary, we characterized KDM3A as a potential bifurcation point in the phenotypic transition of broblasts to myo broblasts in response to TGFβ1 stimulation and, to our knowledge, were the rst focus on cardiac broblasts. Pathophysiologically, we highlight the complexity of the molecular mechanisms of KDM3A in cardiac brosis, providing data to illustrate that KDM3A might in uence cardiac broblasts proliferation, migration and conversion by manipulating the Smad3-dependent pathway. Figure 1 Identi cation of cardiac broblasts and transfection of AdKDM3A or AdshKDM3A adenoviruses. a Immuno uorescence staining of vimentin (red) and α-SMA (green) (×200).The cell nuclei were stained with a DAPI kit; b: Representative uorescent images of adenovirus transfection of AdKDM3A or KDM3A shRNA (×100); c-d: KDM3A expression levels were quanti ed by Western blotting; (n=3). *p< 0.05 vs AdGFP or AdshRNA.