Exploring Regulatory Mechanisms of Atrial Myocyte Hypertrophy of Mitral Regurgitation through Gene Expression Profiling Analysis: Role of NFAT in Cardiac Hypertrophy

Background Left atrial enlargement in mitral regurgitation (MR) predicts a poor prognosis. The regulatory mechanisms of atrial myocyte hypertrophy of MR patients remain unknown. Methods and Results This study comprised 14 patients with MR, 7 patients with aortic valve disease (AVD), and 6 purchased samples from normal subjects (NC). We used microarrays, enrichment analysis and quantitative RT-PCR to study the gene expression profiles in the left atria. Microarray results showed that 112 genes were differentially up-regulated and 132 genes were differentially down-regulated in the left atria between MR patients and NC. Enrichment analysis of differentially expressed genes demonstrated that “NFAT in cardiac hypertrophy” pathway was not only one of the significant associated canonical pathways, but also the only one predicted with a non-zero score of 1.34 (i.e. activated) through Ingenuity Pathway Analysis molecule activity predictor. Ingenuity Pathway Analysis Global Molecular Network analysis exhibited that the highest score network also showed high association with cardiac related pathways and functions. Therefore, 5 NFAT associated genes (PPP3R1, PPP3CB, CAMK1, MEF2C, PLCE1) were studies for validation. The mRNA expressions of PPP3CB and MEF2C were significantly up-regulated, and CAMK1 and PPP3R1 were significantly down-regulated in MR patients compared to NC. Moreover, MR patients had significantly increased mRNA levels of PPP3CB, MEF2C and PLCE1 compared to AVD patients. The atrial myocyte size of MR patients significantly exceeded that of the AVD patients and NC. Conclusions Differentially expressed genes in the “NFAT in cardiac hypertrophy” pathway may play a critical role in the atrial myocyte hypertrophy of MR patients.


Introduction
Mitral regurgitation (MR) is an important cause of heart failure related to valvular heart disease [1]. Left atrial enlargement has prognostic significance in MR patients undergoing mitral valve surgery [2]. Structural remodeling associated with atrial enlargement, especially pathological hypertrophy of myocytes, developed in the left atrial myocardium of patients with MR [3,4]. However, the molecular regulatory mechanisms and functional biological pathways related to the left atrial myocyte hypertrophy of MR patients remain unclear.
In this study, we aimed to systemically explore the crucial differences in the RNA expression pattern between the left atrial myocardium of MR patients and normal subjects, and the molecular regulatory mechanisms and functional biological pathways related to the atrial myocyte hypertrophy using high-density oligonucleotide microarrays and enrichment analysis. The left atrial myocardium of patients with severe aortic valve disease was also used as a reference for gene analysis of the significant pathways as the left atrial size was smaller in patients with aortic valve disease compared to MR patients. The results of this study may recognize some of the differentially expressed genes and related pathways that contribute to the left atrial myocyte hypertrophy in patients with MR.

Patient Population
This study enrolled 14 patients with symptomatic severe non-ischemic MR in sinus rhythm (age: 58±9 years), and 7 age-matched patients with symptomatic severe aortic valve disease in sinus rhythm (age: 63±7 years; aortic stenosis in 1, aortic regurgitation in 4, combined aortic stenoregurgitation in 2). Exclusion factors include previous myocardial infarction, febrile disorder, infectious or inflammatory disease, autoimmune disease, malignancy, acute or chronic viral hepatitis or use of immunosuppressive drugs. Written informed consent was obtained from each study patient, and the study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in a priori approval by the Institutional Review Board of Chang Gung Memorial Hospital (100-0067C).
Six normal adult left atrial tissue samples (24-year-old Caucasian male, 27-year-old Caucasian male, 30-year-old Asian male, 60-year-old Caucasian female, 76-year-old Caucasian female and 77-year-old Caucasian male) were purchased from BioChain, Newark, CA, USA, and these 6 normal atrial tissues were used as the normal controls for gene analysis.

Quantitative Determination of RNAs by Real-Time RT-PCR
The RNA samples were quantified using a spectrophotometer. First-strand cDNAs were synthesized with reverse transcriptase and oligo (dT) primers. Real-time quantitative PCR was performed on the ABI Prism 7500 FAST sequence detection system (Applied Biosystems, CA, USA), using SYBR Green PCR Master Mix (Applied Biosystems, CA, USA). The results of RNAs were normalized against 18S gene expression (the endogenous control). The selected genes and primer sequences are presented in Table 1. The microRNAs (miRs) were extracted from the tissues by using a RNA MiniPrep kit (Zymo Research, CA, USA) according to the manufacturer's protocol. Reverse transcription of miRs was performed using the TaqMan™ microRNA reverse transcription kit (Applied Biosystems, CA, USA) according to manufacturer's recommendations. Briefly, 5 ng of miR was combined with deoxyribonucleoside triphosphates, MultiScribe™ reverse transcriptase, and the primer specific for the target miR (Applied Biosystems, CA, USA). The cDNA was combined with the TaqMan™ assay specific for the target miR. The results of miRs were normalized against U6 snRNA (Applied Biosystems, CA, USA). Quantitative RT-PCR values were presented in 4Cq units.

Western Blotting
The protein extracts of human atrial tissues were examined by Western blot analysis. 20μg protein extracts were electrophoresed on 10% acrylamide SDS-PAGE gels and immunoblotted onto polyvinylidene difluoride membranes. The membranes were preblocked for 1 h in TBST (10 mM Tris-HCl pH 7.6, 150 mM NaCl, 0.1% Tween-20) containing 5% w/v nonfat dry milk and then incubated at 4˚C overnight with anti-α-sarcomeric actin (Sigma Aldrich, Louis, MO, USA). The result of protein was normalized against GAPDH.

Histological Analysis
Atrial tissue sections were deparaffinized in xylene and rehydrated in decreasing concentrations of alcohol. Slides were then stained with hematoxylin/eosin. Tissue sections were observed under an Olympus BX51 microscope with the analysis including at least 100 randomly selected cells under 400 X magnification. All images of each specimen were captured using an Olympus DP70 camera. Atrial cardiomyocytes were analyzed (UTHSCSA, Image tool, Version 3.0).

Cell Culture and Mechanical Stretching of HL-1 Atrial Myocytes
HL-1 atrial myocytes were plated on silicone rubber culture dishes. HL-1 atrial myocytes were cultured for 24 hours in claycomb medium containing 10% FBS, penicillin and streptomycin. Thereafter, culture medium was changed to serum free claycomb medium and HL-1 atrial myocytes were stretched for an additional 8 hours in the same medium. HL-1 atrial myocytes received 15% uniaxial cyclic stretch at 1 Hz for 8 hours by NST-140 cell stretching system (NEPA GENE, Japan) in the cell stretched study groups. Stretched and control (non-stretched) experiments were carried out simultaneously with the same pool of cells in each experiment to match temperature, CO 2 content, and pH of the medium for the stretched and control HL-1 atrial myocytes.

Immunofluorescence Staining
HL-1 atrial myocytes were fixed for 10 min with 4% paraformaldehyde, then exposed to 0.1% triton-X100, and stained with CytoPainter Phalloidin-iFluor 488 reagent (Abcam, Cambridge, USA), according to the manufacturer's protocol. Nuclei were stained with Hoechst 33258 (1:1000 dilution; Sigma, MO, USA). Four randomly chosen fields per section corresponding to at least fifty cells were examined at high magnification (400X). All images of each specimen were captured using a Leica DMI3000 microscope. Atrial cardiomyocytes were analyzed (UTHSCSA, Image tool, Version 3.0).

Statistical Analysis
Data are presented as mean ± SD (baseline characteristics) or SEM (gene and protein expressions). Categorical variables were compared using chi-square test or Fisher exact test as appropriate. Continuous variables among 3 groups were analyzed by the Kruskal-Wallis Test, and continuous variables between 2 groups were analyzed by the Mann-Whitney Test. Statistical analysis was performed using commercial statistical software (IBM SPSS Statistics 22). All P values were two-sided, and the level of statistical significance was set at 0.05. Table 2 lists the clinical characteristics of the study patients with MR and patients with aortic valve disease. The two groups did not significantly differ in age, or heart failure status. The two groups also did not significantly differ in the preoperative left atrial ejection fraction, left ventricular size and left ventricular ejection fraction. However, the left atrial size was significantly larger in the MR patients than patients with aortic valve disease. Seventy-eight percent of MR patients and forty-two percent of patients with aortic valve disease received renin-angiotensin system blockers (P = 0.102).

Identification and Enrichment Analysis of Differential Expression Genes between MR Patients and Normal Subjects
To determine the effect of MR on gene expression, we compared the expression profile in the left atrial free walls of the 7 MR patients to 3 normal subjects (76-year-old Caucasian female, 24-year-old Caucasian male and 27-year-old Caucasian male). A total of 244 differentially expressed genes were discovered by using genefilter R package [6] with the P value < 0.01 (ttest) and a fold-change cut-offs of > 1.5. A total of 112 genes were identified to be differentially up-regulated between MR patients and normal subjects (Table 3), and a total of 132 genes were identified to be differentially down-regulated between MR patients and normal subjects (Table 4), with the heat map graph being depicted in Fig 1. As with the unsupervised hierarchical clustering, the samples and genes were sorted corresponding to their respective groups.
To elucidate the molecular mechanisms of MR on left atrial gene expression, we used Ingenuity Pathway Analysis to search for enrichment in predicted function. A network with highest score (P-score = 50, i.e. P value < 10 −50 ) was generated from 244 differentially expressed genes using Ingenuity Pathway Analysis Global Molecular Network algorithm as depicted in Fig 2, and 26 focused genes were identified to be involved in the network, including PLCE1, PPP3R1, PPP3CB, MEF2C and etc. Top involved canonical pathways in this network included role of nuclear factor of activated T cells (NFAT) in cardiac hypertrophy, cardiac hypertrophy signaling, and calcium signaling ( Table 5). Top diseases and functions in this network included cardiovascular system development and function, organ morphology, and organismal development (Table 5). These results demonstrated that the network was significantly associated with cardiac related pathways and functions, such as role of NFAT in cardiac hypertrophy and cardiovascular system development and function.
Furthermore, we applied the activation z-score analysis method, which was proposed by Andreas Krämer et al [5] in 2014, to measure activation states (increased or decreased) of the pathways affected by differentially expressed genes. We take a statistical approach by defining a quantity (z-score) that determines whether a biological function has significantly more "increased" predictions than "decreased" predictions (z>0) or vice versa (z<0). In practice, zscores greater than 2 or smaller than -2 can be considered significant. Only "Role of NFAT in cardiac hypertrophy" pathway ( Fig 3) had a z-score of 1.34 and P < 0.02. Thus, according to log2 fold-change values and the predictive activities of the differentially expressed genes significantly involved in NFAT pathway (Table 6), we derived an activation z-score equal to 1.34 which suggests that these differentially expressed genes moderately activate the Role of NFAT in cardiac hypertrophy pathway (Fig 3). The detailed information of differentially expressed genes involved in NFAT pathway was shown in Table 6, implicating that the expression patterns of PPP3CB (Calcineurin A beta), PPP3R1 (Calcineurin B), PLCE1, MEF2C, and CAMK1 played a role in the activation of hypertrophy of atrial myocytes in MR patients compared to normal subjects. The results of the real-time quantitative RT-PCR of these 5 genes were consistent with the RNA microarray data (Table 7).
Additionally, Significant Analysis of Microarrays method [7] was also applied for significant gene analysis. One thousand and eighty-three significant genes were identified with false discovery rate of 0.05, and four out of five focused genes (PLCE1, CAMK1, PPP3R1, and PPP3CB) were also identified using Significant Analysis of Microarrays method. The canonical pathway of Role of NFAT in Cardiac Hypertrophy was still significantly identified using Ingenuity Pathway Analysis with P-value of 0.039 and z-score of 0.83. Therefore, we focused on deciphering and discussing their regulatory roles in cardiac hypertrophy in the following sections and 5 focused NFAT associated genes (PLCE1, PPP3R1, PPP3CB, CAMK1, MEF2C) were studies for experimental validation.

Hypertrophy of Atrial Myocytes in MR Patients Compared to Normal Subjects and Patients with Aortic Valve Disease
The average cell surface area of myocytes in the left atrial tissue of the MR patients (n = 14) significantly exceeded the average cell surface area of myocytes in the left atrial tissue of the              (Fig 4). However, the average cell surface area and nucleus size of myocytes in the left atrial tissue did not significantly differ between patients with aortic valve disease and normal subjects (P = 0.456 and P = 0.881, respectively).

Hypertrophy of HL-1 Atrial Myocytes by Mechanical Stretching
The causal relationship between atrial hypertrophy and atrial dilatation due to volume overload of MR was partly mimicked by mechanical stretching of HL-1 atrial myocytes. The

Discussion
This study identifies and reports the alteration of the RNA expression pattern, molecular mechanisms and biological processes involving in atrial myocyte hypertrophy between the left atrial myocardium of MR patients and normal subjects using high-density oligonucleotide microarrays and enrichment analysis. A total of 112 genes were identified to be differentially up-regulated and 132 genes were identified to differentially down-regulated in the left atria between MR patients and normal subjects. Notably, the expression patterns of PPP3CB, PPP3R1, PLCE1, MEF2C and CAMK1 in the "NFAT in cardiac hypertrophy" pathway played a significant role in the activation of hypertrophy of atrial myocytes in MR patients compared to normal subjects.

Activation of "Role of NFAT in Cardiac Hypertrophy" Pathway
Calcineurin/NFAT coupling has been reported to participate in pathological, but not physiological, cardiac hypertrophy [8]. Cardiac hypertrophy is a compensatory response to pathological states and hemodynamic overload. However, pathological hypertrophy leads to atrial myocardial disarrangement and consequently, atrial enlargement that is correlated with poor prognosis in MR patients [2].
The initial phase in the development of myocardial hypertrophy involves factors, such as endothelin-1, angiotensin-II, and adrenergic agonists at the cell membrane, binding to the Gprotein coupled receptors (Fig 3), and several interdependent signaling cascades that include G-proteins, GTPases such as Ras, RhoA and Rac, and kinases such as ERK/MAPK and PKC  [9]. Prior study showed enhanced expression of Rho-associated kinase in the left atrial myocytes of MR patients [10]. In all of the hypertrophic pathways, NFAT plays a critical role in the development of cardiac hypertrophy. Several studies have shown the importance of Ca 2+ sensitive signaling molecules, including calcineurins (i.e. PPP3CB, PPP3R1) and CAMK, a calcium/ calmodulin-dependent protein kinase, in hypertrophic pathways [11]. Activation of protein kinase C leads to increased Ca 2+ levels that activate calcineurins. Calcineurin activation leads to the dephosphorylation of NFATc4, allowing its nuclear localization where it cooperates with other transcription factors to participate in the cardiac hypertrophy. In this study, the expression of PPP3CB, the catalytic subunit, was significantly up-regulated in the MR patients compared to normal subjects. However, the expression of PPP3R1, the regulatory subunit B, was significantly down-regulated in the MR patients compared to normal subjects. Overexpression of PPP3R1, also known as modulatory calcineurin-interacting protein-1, has been reported to attenuate left ventricular hypertrophy after myocardial infarction [12]. As shown in Table 6, the predictive activities of the differential trend between PPP3CB and PPP3R1 to the Role of NFAT in cardiac hypertrophy pathway implicated activation of this pathway. The CAMK1 was also found to be significantly down-regulated in MR patients compared to normal subjects in this study. However, the molecular function of CAMK1 in human heart is rarely reported. Further studies are warranted to investigate the role of CAMK1 in the development of atrial hypertrophy in patients with MR.
MEF2 (myocyte enhancer factor 2), especially MEF2C, is an important transcription factor regulating the cardiac gene program during myocardial cell hypertrophy [13]. The activation of MEF2 by CAMK is mediated mainly through the phosphorylation of transcriptional repressors, the class II histone deacetylases, resulting in disrupting the association between MEF2 and histone deacetylases in the nucleus and transcriptional activation of cardiac hypertrophy [14,15]. In this study, MEF2C was found to be significantly up-regulated in MR patients compared to normal subjects.
PLCE (phospholipase C epsilon), an effector of Ras, has been shown to involve in stressinduced hypertrophy and PLCE, scaffolded to muscle-specific A kinase-anchoring protein in cardiac myocytes, responds to hypertrophic stimuli to generate diacylglycerol from phosphatidylinositol 4-phosphate in the Golgi apparatus, in close proximity to the nuclear envelope, to regulate activation of nuclear protein kinase D and hypertrophic signaling pathways [16]. In this study, PLCE1 was found to be up-regulated in MR patients compared to patients with aortic valve disease and normal subjects.

Hypertrophy-Related MicroRNAs
The expressions of antihypertrophic miRs, miR-1 and miR-133, in the left atrial tissue were significantly down-regulated in the MR patients compared to normal controls, while the expression of agonist of the hypertrophic response, miR-208, in the left atrial tissue were significantly up-regulated in the MR patients compared to normal controls [17].

Hypertrophy-Related Decorin and Calponin
Decorin has been reported to promote myoblast proliferation mediated by an endoplasmic reticulum stress-related pathway [18]. Decorin was identified to be differentially down-regulated in MR patients compared to normal subjects in this study. Therefore, decorin might not be involved in the hypertrophy of atrial myocytes in MR patients, which was mainly related to the Role of NFAT in cardiac hypertrophy pathway.
Calponin has been reported to be involved in the hypertrophy of smooth muscle cells [19,20]. Calponin 2 was identified to be differentially down-regulated in MR patients compared to normal subjects in this study. Therefore, calponin 2 might not be involved in the hypertrophy of atrial myocytes in MR patients.

Pathological Hypertrophy of Atrial Myocytes in MR
Atria, like the ventricles, can undergo hypertrophy in response to increased volume and pressure overload. In MR, the volume and pressure in the left atrium are greatly increased. The left atrium of MR patients responds by undergoing chronic dilation, which enables it to accommodate the increased volume without a large increase in pressure because of its increased compliance. However, extreme hypertrophy and dilatation is deleterious because it increases the oxygen demand of the heart and decreases mechanical efficiency. Furthermore, atrial fibrillation, an important risk factor of stroke and systemic embolization, may develop as a consequence of atrial enlargement, and vice versa [21].

Study Limitations
There are several limitations of this study. Firstly, most of the patients with MR received renin-angiotensin system blockers. Therefore, the expression of some genes might have been modified by renin-angiotensin system blockers [22]. However, there was no significant difference in the expressions of PPP3CB (14.66±0.38 vs. 13.28±0.02, P = 0.167), PPP3R1 (14.29 ±0.41 vs. 16.16±0.91, P = 0.052), PLCE1 (15.95±0.46 vs. 14.22±0.02, P = 0.167), MEF2C (15.99 ±0.58 vs. 14.46±1.47, P = 0.392) and CAMK1 (18.49±0.29 vs. 18.86±0.38, P = 0.484) between MR patients with renin-angiotensin system blockers (n = 11) vs. MR patients without reninangiotensin system blockers (n = 3). Secondly, the sample size was relatively small. However, the gene expression pattern by microarray analysis was quite consistent in the same group ( Fig  1). Thirdly, the age of the normal subjects (n = 6) was younger than that of MR patients (n = 14) (49±25 vs. 58±9 years, P = 0.620), however the difference did not reach statistical significance. Finally, the microarrays were conducted on frozen, unsorted tissue samples. It is hence impossible to ascertain the sub-tissue or cellular origin of generated data. However, histological analysis did show atrial myocyte hypertrophy of MR patients compared to patients with aortic valve disease and normal subjects, implicating at least some involvement of cellular origin.

Conclusions
Significant hypertrophy developed in the left atrial myocytes of MR patients compared to normal subjects and patients with aortic valve disease. The differentially expressed genes in the "Role of NFAT in cardiac hypertrophy" pathway may play a critical role in the atrial myocyte hypertrophy of MR patients and these differentially expressed genes may serve as potential targets for human MR to prevent the progression of left atrial enlargement and its related complications, such as atrial fibrillation, and heart failure.

Accession Codes
The data discussed in this manuscript have been deposited in NCBI's Gene Expression Omnibus (GEO) and are accessible through GEO Series accession number GSE63045 (http://www. ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE63045).