TGFβ1 and HGF regulate CTGF expression in human atrial fibroblasts and are involved in atrial remodelling in patients with rheumatic heart disease

Abstract Objective This study aimed to investigate the effects of transforming growth factor β1 (TGF β1) and hepatocyte growth factor (HGF) on the expression of connective tissue growth factor (CTGF) in human atrial fibroblasts, and to explore the relationship of these factors in atrial fibrosis and atrial anatomical remodelling (AAR) of patients with atrial fibrillation (AF). Methods Fresh right auricular appendix tissue of 20 patients with rheumatic heart disease undergoing valve replacement surgery was collected during surgeries, 10 patients had sinus rhythm(SR), and 10 patients had chronic atrial fibrillation (CAF). Atrial fibroblasts were then cultured from the tissues with differential attachment technique and treated with either TGFβ1 (10 ng/mL) or HGF (100 ng/mL). CTGF mRNA levels were measured by RT‐PCR, and CTGF protein content was determined using immunofluorescence and Western blotting assays. Results CAF group had higher left atrial diameters (LADs) and higher CTGF mRNA expression in atrial fibroblasts compared with SR group. The CTGF protein content in CAF group was higher than that of SR group and positively correlated with LAD and AF duration. After CAF group was treated with TGFβ1, CTGF mRNA and protein expression were significantly down‐regulated, whereas when treated with HGF, expression was up‐regulated compared with SR group. Conclusions Increased CTGF expression was associated with enlarged LAD, atrial fibrosis and AAR in patients with AF. TGFβ1 and HGF regulate CTGF expression in human atrial fibroblasts with up‐regulation of mRNA and down‐regulation of protein, therefore, either promote or inhibit atrial fibrosis, which could be related to the incidence and persistence of AF.


| INTRODUC TI ON
Atrial fibrillation (AF) is the most common arrhythmia 1 and has high morbidity and mortality. The main complications are thromboembolic events, stroke and heart failure. Rheumatic heart disease (RHD) is one cause of AF, for which a mechanism is not fully understood. AF occurrence and persistence are related to atrial anatomical remodelling (AAR). Myocardial fibrosis (MF) is one of the main pathological changes of the myocardium, which reflects diseases prognosis. AF promotes electrical and structural remodelling of the atrial tissue, forms 'matrices', increases AF relapse probabilities and prolongs AF durations. 2 Atrial fibrosis is the hallmark of AF-dependent structural remodelling. 3 Studies have shown that atrial fibrosis is significantly increased in patients with AF. 4,5 MF is a huge global burden, and new treatment options are greatly needed. 6 MF progression is controlled by various factors, including the renin-angiotensin-aldosterone system, chemokines, cytokines and growth factors. 7,8 Fibrosis is mainly characterized by the deposition of extracellular matrix (ECM), and fibrosis-associated biomarker detection provides insight into cardiac ECM remodelling.
Transforming growth factor-β1 (TGFβ1) is one of the most potent cytokines and is involved in many biological processes including fibrosis, tissue repair, anti-inflammatory responses, hypertrophy and atherosclerosis. 9 In the initial stages of myocardial remodelling, TGFβ1 promotes fibroblast activation and induces ECM protein production including collagen and fibronectin. Connective tissue growth factor (CTGF) is a multi-domain stromal cell protein that mediates fibrosis in various tissues and is a key downstream mediator of the TGFβ1 signalling pathway in fibroblasts. 10,11 Studies have shown that TGFβ1 enhances the activity and expression of the CTGF promoter in cardiac fibroblasts. 12 CTGF is an effective pro-fibrotic factor under various pathophysiological conditions including AF. 13 Using a mouse model of myocardial infarction, it was confirmed that TGFβ and CTGF were expressed in scar tissue fibroblasts. 14 The increased content of CTGF in the left atrial tissue was associated with enhanced fibrosis in AF patients. However, hepatocyte growth factor (HGF) has an anti-TGFβ1 effect and has been shown to reverse the fibrotic process in animal models. HGF significantly reduces TGFβ1-induced CTGF production in tubulointerstitial fibroblasts 15 and antagonizes MF by inhibiting the endothelial-mesenchymal transition. 16 TGFβ1 and HGF promote and inhibit MF, respectively, and these regulatory effects may be related to the molecular mechanisms of MF. As TGFβ1 has multiple functions in human physiology, blocking its effects could result in adverse outcome. 17 CTGF, which is a mutual TGFβ1 and HGF downstream signalling factor, acts more specifically on connective tissue, 18 promoting fibrosis and ECM production. Data suggest that although enhanced CTGF expression is closely linked to fibrosis, CTGF alone is insufficient to promote the progression of fibrosis in pathophysiological environments. It is assumed that several factors need to be combined to turn CTGF into a pro-fibrotic and 'pro-arrhythmic' factor. And, in another study, CTGF was proposed as a new target against renal fibrosis. 19 This study investigated the 'network' of TGFβ1, HGF and CTGF, with the expectation that CTGF will be a new therapeutic target of MF.
To study the occurrence and persistence of AF at the cellular and molecular level, our laboratory developed methods to isolate and culture human atrial fibroblasts. 20 In this study, atrial fibroblasts from RHD patients were selected as the target cells. The effects of HGF and TGFβ1 on CTGF secretion in human atrial fibroblasts were investigated, and the relationship between CTGF and left atrial diameter (LAD), and duration of AF episodes was evaluated. We emphasized potential therapeutic targets and the significance of these targets in the clinical treatment of AF. The results of this study will be beneficial for those who want to understand the key molecular mechanisms of AAR and atrial fibrosis in patients with AF and provide evidence for the prevention and treatment of AF.

| General data
Twenty patients with RHD who underwent valve replacement surgery in Fujian Provincial Hospital from January 2008 to July 2012 were enrolled, including nine males and 11 females. Patients were divided into sinus rhythm (SR) and chronic atrial fibrillation (CAF) groups, with 10 patients in each group. According to the 2016 European Society of Cardiology Guidelines for the Management of Atrial Fibrillation, 21 AF episodes lasting longer than 7 days or 1year were referred to as CAF or long-term CAF, respectively. In this study, the CAF group was defined as patients with AF episodes lasting longer than 6 months. AF was diagnosed with an electrocardiogram or a dynamic electrocardiogram. Clinical data from the electrocardiogram and echocardiography were collected. Informed consent was provided by the patients of this study and approved by the Ethics Committee of the Fujian Provincial Hospital.

| Exclusion criteria
Patients with severe heart failure, hypertension, coronary atherosclerotic heart disease, chronic pulmonary heart disease, ischemic or non-ischemic cardiomyopathy, hyperthyroidism, and recent (five atrial fibrillation, connective tissue growth factor, fibroblasts, hepatocyte growth factor, transforming growth factor β1 drug half-lives) intake of angiotensin-converting enzyme and angiotensin II receptor inhibitors, aldosterone receptor antagonists and statin lipid-lowering drugs were excluded.

| Isolation, culture and identification of human atrial fibroblasts and experimental grouping
Fresh aseptic right atrial appendix tissue (20 mg) was collected during valve replacement surgeries of the RHD patients. Samples were kept in large EP tubes containing aseptic PBS (placed on ice) and transported to the laboratory within 15 minutes of tissue acquisition. Atrial fibroblasts were isolated using enzyme digestion and differential adherence methods previously developed by our laboratory. 20 Cells were cultured in DMEM containing 20% foetal bovine serum at 37°C in a 5% CO2 incubator. Cell growth was observed with an inverted microscope, and atrial fibroblasts were passaged after reaching 80% confluence. At the third passage, fibroblasts were identified by α-SMA and vimentin immunofluorescent antibody staining. 20 The second or third passages of cells were used in the following studies. Fibroblasts were divided into five groups: the ①SR group; the ②CAF group; ③ the TGFβ1 intervention group (10 ng/mL TGFβ1 was added to the CAF group cells); ④the HGF intervention group (100 ng/mL HGF was added to the CAF group cells); and ⑤ the TGFβ1+HGF intervention group (10 ng/mL TGFβ1 and 100 ng/mL HGF were added to the CAF group cells).
Total RNA was extracted using the single-step method with the TRIzol reagent (Invitrogen, USA). The first-strand cDNA was synthesized by the two-step method of the reverse transcription kit using 2 µL total RNA (with a final concentration of 1 μg/μL) and was used as the PCR template. The PCR products were synthesized and separated using 2% agarose gel electrophoresis, stained with the nucleic acid dye, GELVIEW, and scanned using a gel imaging system. The relative content of CTGF mRNA was determined by calculating the integral ratio of gray-scale values of the CTGF gene and the internal reference GAPDH fragments. Primer sequences and amplified fragments were GAPDH: upstream 5'-aga agg ctg ggg ctc att tg-3' and downstream 5'-Agg ggc cat cca cag tct tc-3' cDNA, the fragment length was 258 bp; and CTGF: upstream 5' -CCAACTATGATTAGAGCCAACTG-3', downstream 5'-AGGCACAGGTCTTGATGA AC-3' cDNA, the fragment length was 400 bp. The PCR reaction procedure was 35 cycles of pre-denaturation at 94°C for 5 minutes, followed by denaturation at 94°C for 30 seconds, annealing at 60°C for 45 seconds, and extensions at 72°C for 1 minute, and extensions at 72°C for 10 minutes.

| Immunofluorescence and Western blotting analysis
CTGF protein expression was detected using immunofluorescence and Western blot assays. One mL of a 2 × 10 6 atrial fibroblast suspension (in DMEM medium containing 20% foetal bovine serum) was added to each well of a 6-well plate and cultured for 24 hours. The media were then replaced with serum-free DMEM, and the cells were incubated for another 24 hours. At this point, the culture media were discarded, and appropriate cytokines were added to the wells of the intervention groups (in DMEM culture medium containing 2% foetal bovine serum).
Cells were cultured for 48 hours, and then the culture medium was discarded. Fibroblasts were fixed with 1 mL 4% paraformaldehyde for 30 minutes, washed with PBS buffer three times, and blocked with 1 mL blocking solution on a shaker for 60 minutes. The blocking solution was removed, and the primary goat anti-human CTGF antibody (1:200) was added. The reaction system was incubated at 4°C overnight and washed three times (3-5 minutes each time). Then, the fluorescence-labelled secondary anti-goat FITC antibody (1:100) was added, and samples were incubated for 1 hour on a shaker, washed three times, observed with a fluorescent microscope after sealing, and developed with FITC. Imagepro Plus software and the SigmaScan 4.0 system were used for analysing the immunofluorescence images to determine the cumulative integrated optical density (IOD).
Western blot extraction and SDS-PAGE gel electrophoresis were performed according to the previously described method. 22 Briefly, 2 mL of the cell suspensions were added to each well. After performing 10% SDS-PAGE, the CTGF protein was electrophoretically trans-

| Statistical analysis
Excel 2010 and SPSS 21.0 were used to process and analyse data. The study data were expressed as mean ± standard deviation (x ± s), and the mean between the SR and CAF groups was compared using a bilateral t-test. Means of the two groups were compared with the one-way ANOVA and the least significant difference (LSD) test. The correlation between indicators was analysed with correlation and regression analysis. The results were considered statistically significant when P < 0.05.  Figure 1A shows fibroblast morphology observed with a microscope.

| Culture and identification of human atrial fibroblasts
Fibroblasts were identified as myocardial fibroblasts (MFbs) using cytoplasmic vimentin and α-SMA antibody immunofluorescent stains ( Figures 1B,C). The cells were passaged and divided into the five groups.

| The effects of TGFβ1 and HGF on CTGF mRNA expression in atrial fibroblasts
The positions of the amplified CTGF and GAPDH electrophoretic bands were 400 bp and 258 bp, which was consistent with the set lengths of the fragments (Figure 2A). The relative CTGF mRNA expression level was significantly increased in the CAF group compared with the SR group (P < 0.01). CTGF mRNA expression in the SR, CAF, TGFβ1, HGF and TGFβ1+HGF groups were 1.16 ± 0.85, 2.24 ± 0.97, 3.16 ± 1.03, 0.67 ± 0.59 and 1.44 ± 0.77, respectively ( Figure 2B). Compared with the CAF group, CTGF mRNA expression was higher in the TGFβ1 group (P < 0.01) and lower in the TGFβ1+HGF group (P < 0.05) and the HGF group (P < 0.01).
Compared with the TGFβ1+HGF group, CTGF mRNA expression was lower in the HGF group (P < 0.05) and higher in the TGFβ1 group (P < 0.01) ( Figure 2B).

| CTGF protein expression in myocardial atrial fibroblasts
MFb cytoplasmic CTGF fluoresces green using immunofluorescence technology ( Figure 3 in the TGFβ1+HGF (P < 0.05) and HGF groups (P < 0.01). Compared with the TGFβ1+HGF group, CTGF protein expression was lower in the HGF group (P < 0.05) and higher in the TGFβ1 group (P < 0.01).
The results show consistency between the two methods for CTGF protein expression (Figures 3, 4).
Correlations between the relative CTGF protein levels and LAD or duration of AF episodes in the CAF group were also studied.   These findings support that CTGF plays a key role in the AAR of AF patients.

| D ISCUSS I ON
Our previous studies demonstrated that CTGF and α-SMA expression in the atrial tissue of AF patients were significantly increased and positively correlated with type I collagen. TGFβ1 expression activated atrial fibroblasts to become MFbs, which then expressed α-SMA and secreted type I collagen in large amounts, promoting atrial fibrosis in AF patients. 22 In this work, RT-PCR results showed that the relative CTGF mRNA levels in MFbs of CAF patients  Several studies using experimental models have shown that TGFβ1 is a potential anti-fibrotic target. Caveolin-1, mitogen-activated protein kinase 4, and relaxin are negative TGFβ1 regulators that can reverse atrial fibrosis. [38][39][40] TGFβ1 is an upstream and initiating factor with a wide range of physiological functions; and therefore, its ability to treat AF effectively is limited by its lack of specificity and side effects. Studies have shown that TGFβ1 downstream signalling molecules, such as CTGF, can provide safer and more effective targets for the treatment of these fibrotic diseases. 41 Therapeutic interventions that inhibit CTGF biological activity have been tested in clinical trials for fibrosis of other organs. 42 Antibodies against CTGF were shown to partially inhibit TGFβ1-induced collagen synthesis in fibroblasts.
Future research directions will determine whether atrial anti-CTGF gene transfer can be applied to AF animal models. In these animal models, AF with decreased LVEF is associated with atrial structural remodelling, which is characterized by increased atrial fibrosis and collagen secretion. Profibrotic signalling includes the activation of the TGF-β1-Smad-CTGF pathway and the transdifferentiation of myofibroblasts. 43 This could provide a basis for future remodelling-based antiarrhythmic treatments for atrial fibrosis. Recent reports show that MiR-455 and MiR-132 reduced myocardial fibrosis by directly inhibiting CTGF, 44,45 and could be used as new targets for drug therapy.
Based on the literature and experimental results from this study, TGFβ1 and CTGF antibodies could exert anti-fibrotic effects by blocking TGFβ1 and CTGF. The use of CTGF antagonism, which could be more tissue-specific than TGFβ1 antagonism, and human recombinant HGF to treat atrial fibrosis merits further research.

ACK N OWLED G EM ENTS
Zhang JC and Zhu PL designed the research study, Chen JQ and Cheng XL performed the research and analysed the data, Chen Q and Wu HL wrote the paper, Xiang GJ and Huang QL performed the research, Guo YS and Chen MY analysed the data.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.