DACT1 Involvement in the Cytoskeletal Arrangement of Cardiomyocytes in Atrial Fibrillation by Regulating Cx43

Objective To determine the role of the dishevelled binding antagonist of beta catenin 1 (DACT1) in the cytoskeletal arrangement of cardiomyocytes in atrial fibrillation (AF). Methods The DACT1 expression and its associations with the degree of fibrosis and β-catenin in valvular disease patients were analyzed by immunohistochemistry and Masson’s staining. DACT1 was overexpressed in the atrial myocyte cell line (HL-1) and the cardiac cell line (H9C2) by adenoviral vectors. Alterations in the fibrous actin (F-actin) content and organization and the expression of β-catenin were detected by flow cytometry, immunofluorescence, and Western blotting. Additionally, the association of DACT1 with gap junctions connexin 43 (Cx43) was detected by immunohistochemistry, immunofluorescence, and Western blotting. Results Decreased cytoplasmic DACT1 expression in the myocardium was associated with AF (P=0.037) and a high degree of fibrosis (weak vs. strong, P=0.028; weak vs. very strong, P=0.029). A positive association was observed between DACT1 and β-catenin expression in clinical samples (P=0.028, Spearman’s rho=0.408). Furthermore, overexpression of DACT1 in HL-1 and H9C2 cells induced an increase in β-catenin and subsequent partial colocalization of DACT1 and β-catenin. In addition, F-actin content and organization were enhanced. Interestingly, DACT1 was positively correlated with the Cx43 expression in clinical samples (P=0.048, Spearman’s rho=0.370) and changed the Cx43 distribution in cardiac cell lines. Conclusion DACT1 proved to be a novel AF-related gene by regulating Cx43 via cytoskeletal organization induced by β-catenin accumulation in cardiomyocytes. DACT1 could thus serve as a potential therapeutic marker for AF.


INTRODUCTION
Atrial fibrillation (AF) is a ubiquitous arrhythmia and causes considerable morbidity and mortality; however, its mechanism remains unclear. AF is known to be caused by significant atrial electrical and structural remodeling [1] . The cytoskeleton plays an important role in this process, because of its involvement in the channel opening mediated by integrin in the process of the force transmitting force indirectly to stretch-activated channels (SACs) [2] , as well as the regulation of L-type Ca 2+ current (ICa,L) and signaling pathway transduction [3] . Moreover, the force transmission can be directly or indirectly mediated by the focal proteins on the cytoskeleton [4] . Acute mechanical changes produce electrophysiological alterations and arrhythmias, mediated by both the cytoskeleton and the extracellular matrix [4][5][6] . Once AF is induced, rapidly and inhomogeneously contracting and interacting atrial segments would tend to perpetuate electrophysiological dispersion [4] , which could facilitate the geometry (fibrosis and scarring of atrium) formation [7] . Thus, the cytoskeleton could be an essential systematic regulator and a potential therapy marker for AF. β-catenin is not just an important transcriptional effector of the canonical Wnt signaling pathway in the nucleus, but a crucial regulator of the cytoskeleton and signal transduction outside the nucleus as well [8] . β-catenin can affect both the cellular microtubule arrays by interacting with the complex of EB1/P150/dynein/dynactin and P120-catenin and affect the organization of the cellular actin filament by interacting with actin-bundling protein [9,10] . In cardiac cells, β-catenin directly binds to the C-terminal region of cadherin, whereas α-catenins link the cadherin/catenin complex to the actin cytoskeleton [11] . Thus, β-catenin is essential for the maintenance of cardiac cellular function. Abnormal β-catenin in vivo could lead to hypertrophic cardiomyopathy, heart failure, and arrhythmias [12,13] . However, the mechanism of cytoskeletal arrangement via β-catenin involvement in AF remains unclear.
A previous study showed that the dishevelled binding antagonist of beta catenin 1 (DACT1), a novel bidirectional regulator of β-catenin, was associated with cytoplasmic β-catenin accumulation, suggesting its potential role in cytoskeletal arrangement [14] . Yuan et al. [15] suggested that DACT1 induced the β-catenin accumulation in the cytoplasm by inhibiting the glycogen synthase kinase beta (GSK3-β) activity and interacting directly with β-catenin. The actin cytoskeleton, which may be regulated by β-catenin, serves as a link to the adherens junctions, gap junctions, and desmosomes, which constitute the intercalated disc [16] . The remodeling of gap junctions, which are assembled from connexins, has been acknowledged to be an important contributor to AF [17] . Thus, we hypothesized that a new mechanism mediated by DACT1 might exist by which the stability of gap junctions is regulated in AF. However, little is known about this mechanism.
In the present study, we demonstrated that DACT1 is involved in AF. Then, we demonstrated that DACT1 was involved in AF by regulating the actin cytoskeleton. Furthermore, we found out that connexin 43 (Cx43) remodeling might be the result of the effect of DACT1 on the cytoskeleton in AF.

Ethics
This study was approved by the Human Ethics Committee of the First Affiliated Hospital of Sun Yat-Sen University and complied with the principles governing the use of human tissues that are outlined in the Declaration of Helsinki. Informed consent was given by patients before participating in the study.

Human Tissue Preparation
Tissue samples from the right atrial (RA) appendage were obtained from 29 patients with valvular heart disease who had undergone valve replacement surgery. Ten patients, who constituted the sinus rhythm (SR) group, did not have a history of AF, and 19 patients, who constituted the AF group, had documented arrhythmias from which they had suffered for more than six months before undergoing surgery. These tissue samples were obtained during the surgeries, before the aortic cross-clamping, and were immediately fixed in 4% paraformaldehyde (PFA). The diagnosis of AF was made based   The hearts obtained at autopsy were devoid of any abnormal findings and the causes of death were not heart related.

Immunohistochemical Staining
All the samples were fixed in 4% PFA, embedded in paraffin, and stained with hematoxylin and eosin for routine histological examination. Immunohistochemical staining was performed on 4-μm-thick tissue sections. After deparaffinization and rehydration, all the sections were microwaved (10 min) in 0.01 mol/L sodium citrate buffer (pH 6.0) for antigen retrieval. To block endogenous peroxidase activity, we incubated the sections with 10% normal goat serum in phosphate-buffered saline (PBS) for 15 min at room temperature. Then, all the sections were incubated with a rabbit polyclonal antibody against DACT1 (1:100; Abcam), β-catenin (1:100; Cell Signaling Technology), or Cx43 (1:100; Sigma) overnight at 4°C. The slides were subsequently treated with a SuperPic Ture Polymer Detection Kit and Liquid DAB Substrate Kit (Zymed/Invitrogen, San Francisco, USA) and counterstained with hematoxylin, dehydrated, and mounted.

Masson's Trichrome Staining
The sections were stained with Masson's trichrome for fibrosis quantification. For Masson's trichrome staining, the sections were dewaxed with xylol (two dewaxing steps lasting 2 min each, followed by soaks in a series of graded alcohols, with concentrations ranging from 95% to 99%). All the slices were then washed in distilled water and placed in a hematoxylin solution for 3 min, after which a color change was induced with lithium carbonate. The slices were subsequently washed in pure water and colored with Ponceau red staining (in a 45 kW oven at 30°C for 20 sec). Next, the slices were placed in distilled water and phosphomolybdic acid for 1 min before being labeled with a green fluorescent marker and washed again with distilled water. The severity of fibrosis was then assessed in each of the sections upon their collection.

Immunostaining Evaluation
Immunohistochemical expression was evaluated using the Image-Pro Plus 6.0 software (Media Cybernetics, Silver Spring, Maryland, USA). Briefly, at least three positive expression fields in a section of myocardial tissue were randomly selected, and then these positive regions were analyzed with the Image-Pro Plus 6.0 software to determine their integral optical density and area. The average of the optical density values, which represented the expression intensity in the section, was subsequently calculated.

Fibrosis Evaluation
Fibrosis severity was evaluated using the Image-Pro Plus 6.0 software. At least three fields in a section of myocardial tissue were randomly selected after which the ratio of the fibrotic area to the total area of each selected field was calculated to assess fibrosis severity. The average ratio, which represented the severity of the fibrosis in the section of myocardial tissue, was subsequently determined.

Replication-Defective Recombinant Adenoviral Vectors
Briefly, the complementary deoxyribonucleic acid (cDNA) of two transcript variant isoforms of DACT1 (named D1V1 and D1V2) were purchased from GeneCopoeia (cat: HOC23467 and GC-H1635, respectively), and inserted into pGA1-2A-GFP through enzymatic digestion and ligation to obtain pGA1-D1V1-GFP and pGA1-D1V2-GFP. The autocleavage 2A linker between DACT1 and green fluorescent protein (GFP) facilitates the formation of separated DACT1 and GFP reporter. Subsequently, pGA1-D1V1-GFP and pGA1-D1V2-GFP were linearized by enzymatic digestion and subjected to homologous recombination with the linearized pAd5∆E1∆E3 backbone in competent E. coli BJ5183 cells to obtain pAd5-D1V1-GFP and pAd5-D1V2-GFP, respectively. pAd5-D1V1-GFP and pAd5-D1V2-GFP were amplified in competent E. coli XLI-Blue cells and prepared with a Qiagen Plasmid Midi kit (QIAGEN), according to the manufacturer's protocol. pAd5-D1V1-GFP and pAd5-D1V2-GFP were then linearized and transfected into HEK293 cells to rescue and propagate the respective recombinant adenoviruses. The expression levels of DACT1 were analyzed further by Western blotting analysis, and GFP-expressing cells were observed using a fluorescence microscope (Axio Observer Z1).

Cell Culture and Adenovirus Infection
The H9C2 and HL-1 cell lines were used in this study. The HL-1 cell line was obtained from EMD Millipore Corporation (Cat. SCC065). Cells were cultured using Claycomb medium supplemented with 10% fetal bovine serum (FBS), 2 mM L-glutamine, 100 U/mL penicillin, and 100 mg/L streptomycin in flasks precoated with fibronectin and gelatin. The H9C2 cell line was cultured in Dulbecco's Modified Eagle Medium (DMEM)/ high-glucose medium (HyClone) with 15% FBS and antibodies (100 U/mL penicillin and 100 mg/L streptomycin). All the cells were grown under a 5% CO 2 atmosphere at 37°C.
Cells were plated 24 hours before transduction, incubated with fresh media containing the required multiplicity of infection (MOI) per cell of the virus, left for 18 hours, washed, and maintained until harvesting. Forty-eight hours after transduction, GFP-expressing cells were observed using a fluorescence microscope (Axio Observer Z1).

Statistical Analysis
Correlations between the immunohistochemical results and patient clinical variables were analyzed by χ 2 -tests. Continuous variables are presented as the mean ± SEM. Comparisons of continuous variables between groups were performed with Student's t-test, and the correlations between DACT1 expression levels and β-catenin or Cx43 were assessed with the nonparametric Spearman's rank correlation test. P<0.05 was considered to indicate statistical significance. All statistical analyses were performed with Statistical Package for the Social Sciences (SPSS) software, version 19.0.

Association between DACT1 Expression and AF in Valvular
Heart Disease DACT1 was located in both the cytoplasm and nucleus of cardiac cells in human myocardial tissue ( Figures 1A, 1B, and 1C). Very strong staining for DACT1 was detected in the normal human heart tissue ( Figure 1A), and decreased expression was observed in the myocardial tissue of the patients with valvular heart diseases ( Figures 1B and 1C). The basic characteristics of patients with valvular diseases are summarized in Supplement Table 1. A significant difference was observed in left atrial (LA) diameter, RA diameter, and NYHA classification, while no significant difference was observed in CHADS2 score and CHA2DVASs between the AF and SR groups (Supplement Table 1). Furthermore, the relationship between DACT1 expression and clinical features was analyzed. In the SR group, no patient was detected to have weak staining for cytoplasmic DACT1 level, while in the AF group, the percentage electrophoresis (SDS-PAGE) and transferred to polyvinylidene difluoride (PVDF) membranes (Millipore, Billerica, Massachusetts, USA). Primary antibodies against DACT1 (1:1000; Origene), β-catenin (1:1000; Cell Signaling Technology), or Cx43 (1:1000; Sigma) were used, and antigen-antibody complexes were detected by a Western blotting luminol reagent (Santa Cruz Biotechnology). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (Proteintech) or β-actin (Wuhan GoodBio Tech. CO. LTD) served as an internal reference, and at least two independent experiments were performed for each cell line. Image J software was used to analyze the mean light density of each band. The expression of target genes was normalized to that of GAPDH or β-actin.

Quantitation of Fibrous Actin (F-Actin) Levels
The technique used was described previously [18] . The F-actin analysis was conducted via flow cytometry, and cells were fixed with 70% ethanol overnight at 4°C. Cell pellets were incubated in PBS containing 0.1% Triton-X 100 for 15 min at room temperature and then they were treated with phalloidin-coumarin (2 mg/mL; Thermo Fisher) for 30 min. The F-actin content was analyzed by flow cytometry (Beckman, cytoFLEX). FlowJo 7.6 software (FlowJo, Ashland, Oregon, USA) was used to analyze the F-actin content in each group.

Immunofluorescence Assay
The technique used was described previously [18] . Cells were fixed with 4% PFA for 15 min, blocked, and permeabilized with 1% FBS and 0.2% Triton X-100 in PBS at room temperature for one hour. Then, the cells were incubated with primary antibody overnight at 4°C. Primary antibodies against DACT1 (1:200; Origene), β-catenin (1:200; Cell Signaling Technology), or Cx43 (1:200; Sigma) were used at the indicated dilutions. After the cells were washed three times with PBS, secondary antibody (Alexa Fluor 488, Alexa Fluor 568, or Alexa Fluor 647) was diluted at 1:500 and applied to the cells for one hour at room temperature. For F-actin staining, the  [19] ; thus, we generated an adenovirus with two isoforms of DACT1 (named V1 or V2 in this study) for further investigation, and the overexpression efficiency was detected (Supplement Figures 1B and 1C). The concentration of β-catenin protein was just slightly increased after the two isoforms of DACT1 restored expression compared with that of the control group (Figure 2A). Immunofluorescence results showed that DACT1 could enhance the β-catenin staining in both HL-1 and H9C2 cells ( Figure 2B), and further analysis in patient samples confirmed this finding. No significant association was observed between β-catenin expression and clinicopathological features (Supplement Table 3); however, we found a positive correlation between expression levels of DACT1 and β-catenin in patients with valvular heart disease (P=0.028, Spearman's rho=0.408) (Supplement Table 4, Figure 2C). Moreover, both DACT1-V1 and DACT1-V2 displayed the partial colocalization was significantly increased, suggesting that a decrease in DACT1 was associated with AF (P=0.037, Supplement Table 2). The degree of fibrosis observed was higher in the weak cytoplasmic DACT1 expression group than in other groups (weak=0.226±0.106, strong=0.120±0.060, very strong=0.144±0.067; weak vs. strong, P=0.028; weak vs. very strong, P=0.029) (Figures 1D to 1G). These data suggested that decreased cytoplasmic DACT1 levels participated in AF and might serve as a risk factor.

DACT1 Induced Cytoskeletal Rearrangement in Myocardial Cells
In the present study, we investigated the regulatory role of DACT1 in β-catenin and cytoskeletal rearrangement. First, the endogenous expression of DACT1 was detected in the atrial myocyte cell line HL-1 and cardiac cell line H9C2, but neither of the cell lines was detected positive for DACT1 (Supplement Figure 1A). Thus, the HL-1 and H9C2 cell lines were selected for high exogenous Supplement Table 3. Mean values comparing the clinical parameters of patients with valvular heart disease in different groups of β-catenin expression in the myocardium. element of the cytoskeleton, were then detected. Phalloidincoumarin staining indicated that DACT1 overexpression increased the F-actin concentration ( Figure 3A). Furthermore, fluorescent images of vector groups showed less bundled actin in the vector groups than in the DACT1-V1 and DACT1-V2 groups in both H9C2 and HL-1 cells ( Figure 3B). Taken together, with β-catenin in the HL-1 and H9C2 cells ( Figure 2D). No DACT1 staining was observed in the vector group of HL-1 and H9C2 (data not shown). These results indicated that DACT1 could induce an alteration in β-catenin in the myocardial cell and might have functions in cytoskeletal rearrangement. The concentration and distribution of F-actin, an essential Supplement

The Regulation of the Cardiac Gap Junction Cx43 by DACT1
Given that cytoskeletal reorganization could lead to an alteration of expression and cellular distribution of Cx43 and gap junction remodeling [20], we hypothesized that the Cx43 might be involved in the regulation of the cytoskeleton by DACT1 in myocardial cells. High Cx43 expression was associated with low fibrosis degree, right ventricular (RV), and RA values (Supplement Table 5), suggesting the structural and fibrotic relevance of Cx43. In the present study, overexpression of neither DACT1 isoform could alter Cx43 concentration in H9C2 and HL-1 cells ( Figure 4A); however, an increase in the distributed heterogeneity was observed ( Figure 4B), suggesting that a sole DACT1 isoform could affect only the Cx43 location in the myocardial cell. Interestingly, we found a significant association between DACT1 expression and Cx43 in the patients with valvular heart disease (P=0.048, Spearman's rho=0.370) ( Figure  4C, Supplement Table 6). We speculated that both the DACT1 isoforms existed in the clinical samples and that their synergistic effect might exist to regulate Cx43. Thus, their association could only be observed in the clinical samples.

DISCUSSION
AF, the most common cardiac arrhythmia in clinical practice, is a major cause of embolic stroke [21] . Despite the recent significant advances in the understanding of the mechanisms associated with AF, complexities in the etiology of atrial electrical dysfunction have prevented definitive elucidation [22] . DACT1, which was first found as a Dvl-interacting protein, plays a crucial role in the normal vertebrate development [23] and human diseases [14] by affecting β-catenin. It was proposed that DACT1 was very important in human diseases at least because of the contribution for determining the role of β-catenin, a structural protein and/or a signal transduction protein. Our investigation in the present study confirmed this hypothesis. In the myocardium of the patients, β-catenin was not expressed in the nucleus, but was present in the cytoplasm and/or membrane. Furthermore, its expression was positively correlated with DACT1 cytoplasmic expression, suggesting that DACT1 may facilitate the extranuclear accumulation of β-catenin, which might serve as a structural protein in the myocardial cells of the valvular heart disease patients. Further analysis of the myocardial structure confirmed this speculation. DACT1 induced F-actin filament accumulation in the myocardial cells. Actin filaments reorganization can lead to alteration of the gap junction protein Cx43 [24] . Thus, unsurprisingly, we found a correlation between the expression of DACT1 and Cx43 in clinical samples and found that DACT1 altered the cellular distribution of Cx43 in the cardiac cell line. Cx43 is a very important susceptibility gene in AF, and  the reduced number and organization of Cx43-containing gap junction plaques likely play a fundamental role in the increased incidence of arrhythmias and degree of fibrosis [25,26] . We also found the structural and fibrotic relevance of Cx43 (Supplement Table 5), and therefore, Cx43 remodeling may be the reason for the association of the weak DACT1 expression with AF and the degree of fibrosis found in our study. Thus, it is proposed that DACT1 is a novel AF-related gene by regulating Cx43 via β-catenin in the myocardial cells. To our knowledge, it was the investigation of the involvement of DACT1 with human AF.

CONCLUSION
In summary, we demonstrated for the first time that a lack of DACT1 was associated with AF and the degree of fibrosis in valvular heart disease. Specifically, DACT1 regulated the gap junction protein Cx43 by accelerating the cytoskeleton rearrangement via accumulation of β-catenin in the myocardial cells, indicating that new mechanisms mediated by DACT1 exist in AF. DACT1 is a novel potential therapeutic target for AF.