Circulating miRNAs as Potential Biomarkers Associated with Cardiac Remodeling and Fibrosis in Chagas Disease Cardiomyopathy

Chagas disease (CD) affects approximately 6–7 million people worldwide, from which 30% develop chronic Chagas cardiomyopathy (CCC), usually after being asymptomatic for years. Currently available diagnostic methods are capable of adequately identifying infected patients, but do not provide information regarding the individual risk of developing the most severe form of the disease. The identification of biomarkers that predict the progression from asymptomatic or indeterminate form to CCC, may guide early implementation of pharmacological therapy. Here, six circulating microRNAs (miR-19a-3p, miR-21-5p, miR-29b-3p, miR-30a-5p, miR-199b-5p and miR-208a-3p) were evaluated and compared among patients with CCC (n = 28), CD indeterminate form (n = 10) and healthy controls (n = 10). MiR-19a-3p, miR-21-5p, and miR-29b-3p were differentially expressed in CCC patients when compared to indeterminate form, showing a positive correlation with cardiac dysfunction, functional class, and fibrosis, and a negative correlation with ejection fraction and left ventricular strain. Cardiac tissue analysis confirmed increased expression of microRNAs in CCC patients. In vitro studies using human cells indicated the involvement of these microRNAs in the processes of cardiac hypertrophy and fibrosis. Our study suggests that miRNAs are involved in the process of cardiac fibrosis and remodeling presented in CD and indicate a group of miRNAs as potential biomarkers of disease progression in CCC.


Introduction
Chagas disease, caused by the protozoan parasite Trypanosoma cruzi, affects approximately 6-7 million people worldwide mainly in Latin American countries, and is spreading to other continents due to population migration [1]. The majority of infected patients will remain asymptomatic during the chronic phase, in what is known as the indeterminate form of Chagas disease, while approximately 30% will develop the cardiac form, also known as chronic Chagas disease Cardiomyopathy (CCC), with a high rate of morbidity and mortality [2]. Residual parasitism is found in different organs in chronic Chagas disease, despite the clinical presentation [3]. In the cardiac form, however, myocardial dysfunction occurs due to parasite persistence, chronic inflammation [4,5], and cardiac fibrosis [6]. Additionally, cardiac remodeling can lead to heart chamber dilation, wall thinning, arrhythmia, and cardiac hypertrophy [4,[7][8][9].
Although currently available diagnostic tests are capable of adequately identifying infected patients, there is no biomarker that can predict the individual risk for a patient's potential to progress from the indeterminate form to CCC [8]. Therefore, the identification of novel biomarkers allowing for early identification of myocardial damage is desirable for interventions aiming at delaying heart dysfunction development and progression.
In Chagas disease, the expression levels of miR-208a and miR-208b were altered in the heart from CCC subjects [24]. Additionally, the potential of miR-208a as a biomarker for Chagas disease has been explored [25]. In the present study, we evaluated the expression profile of circulating microRNAs in three groups: subjects with CCC, subjects with indeterminate form of Chagas disease, and healthy controls, in order to identify possible correlations with clinical parameters associated with the disease prognosis. Additionally, we compared the expression of selected microRNAs in CCC hearts, and explored their roles in vitro using human cardiac fibroblasts and cardiomyocytes.

Expression Profile of Circulating microRNAs in Subjects with Chronic Chagas Disease
Since the detection of circulating miRNAs can vary within each type of analyzed sample [26], we compared the levels of the selected miRNAs in serum and plasma samples obtained from 16 chronic Chagas subjects. Increased expression levels of miR-19a-3p, miR-29b-3p, and miR-30a-5p were observed in serum samples when compared to plasma from the same subjects ( Figure S1A,C,D). No differences were observed between miR-21-5p or miR-199b-5p when comparing serum and plasma samples ( Figure S1B,E). Amplification of miR-208a-3p was successful in only four samples, with higher levels detected in plasma samples ( Figure S1F).
The following validation analysis was performed in serum samples from 42 subjects, stratified into three groups (healthy, indeterminate form, and CCC). Patients with Chagas disease underwent evaluations by echography and MRI. Clinical and demographic characteristics of the Chagas disease patients are shown in Table 1.

MicroRNAs in Cardiac Tissue Samples of Subjects with End-Stage CCC
Left ventricular sections obtained from explanted hearts were evaluated after staining with hematoxylin and eosin, showing the presence of inflammatory infiltrates composed of mononuclear cells, areas of cardiac fibrosis and hypertrophic myocytes ( Figure 3A-D). In order to evaluate whether the microRNA expression in the cardiac tissues is consistent with the findings detected in the serum, explanted cardiac tissues from chronic Chagas disease subjects (n = 9) and unaffected controls (n = 8) were evaluated by RT-qPCR. Increased expression levels of all microRNAs tested (miR-19a-3p, miR-21-5p, miR-29b-3p, miR-30a-5p, miR-199b-5p and miR-208a-3p) were found in the heart samples from subjects with Chagas disease when compared to controls ( Figure 3E). Next, we evaluated the correlation between clinical parameters of Chagas disease (indeterminate vs CCC) and the expression levels of the selected miRNAs. MiR-19a-3p and miR-21-5p expression levels showed a positive correlation with cardiac dysfunction, functional class (NYHA II/IV), and percentage of fibrosis measured by cardiac MRI. Negative correlations were found between the levels of miR-19a-3p and miR-21-5p ejection fraction and left ventricular strain, measured by echocardiography. Furthermore, miR-29b-3p was negatively correlated with ejection fraction and positively correlated with cardiac dysfunction, functional class, and percentage of fibrosis (Table 2).

MicroRNAs Expression Levels in Fibrosis Using Human Cardiac Fibroblasts
To evaluate the possible involvement of miRNAs in the cardiac fibrosis process that occurs in Chagas disease, we used an in vitro model of human cardiac fibroblasts stimulated with the profibrogenic factor TGF-β1. This cytokine found in highly elevated levels in CCC [27] and induces collagen production ( Figure 4A). First, markers of fibroblasts, such as DDR2, αSMA, and vimentin, were detected in cells obtained from heart explants ( Figure 4B-D). TGF-β1 stimulation was associated with a significant increase in COL1A1 expression, as shown by high-content imaging and RT-qPCR analyses ( Figure 4E-G). Overexpression of miR-21-5p was found in TGF-β1-stimulated cardiac The fibers are random and there is focal interstitial inflammation (top left corner); (D) multifocal inflammation characterized by accumulation of plasma cells and mature lymphocytes. The nuclei of myocytes are mildly enlarged and pleomorphic (moderate myocyte hypertrophy), and there is mild interstitial fibrosis. Bars = 500 µm; (E) miR-19a-3p, miR-21-5p, miR-29b-3p, miR-30a-5p, miR-199b-5p and miR-208a-3p evaluation by RT-qPCR, using 2 −∆∆Ct method and normalized with miR-423-3p. Data represent the mean ± SEM. Controls (CTL), n = 8, and chronic Chagas cardiomyopathy (CCC), n = 9, analyzed using Student's t-test, *** p < 0.001, ** p < 0.01.

MicroRNAs Expression Levels in Fibrosis Using Human Cardiac Fibroblasts
To evaluate the possible involvement of miRNAs in the cardiac fibrosis process that occurs in Chagas disease, we used an in vitro model of human cardiac fibroblasts stimulated with the profibrogenic factor TGF-β1. This cytokine found in highly elevated levels in CCC [27] and induces collagen production ( Figure 4A). First, markers of fibroblasts, such as DDR2, αSMA, and vimentin, were detected in cells obtained from heart explants ( Figure 4B-D). TGF-β1 stimulation was associated with a significant increase in COL1A1 expression, as shown by high-content imaging and RT-qPCR analyses ( Figure 4E-G). Overexpression of miR-21-5p was found in TGF-β1-stimulated cardiac fibroblasts, compared to unstimulated cells ( Figure 4H). Other miRNAs tested (miR-19a-3p, miR-29b-3p, miR-30a-5p, miR-199b-5p and miR-208a-3p) were not altered by the addition of TGF-β1 in cardiac fibroblast cultures.

Expression of MicroRNAs in a Hypertrophy Model Using hiPSC-Derived Cardiomyocytes
Next, to investigate a possible role of the selected miRNAs in cardiomyocyte hypertrophy, a feature of CCC [28], we used an in vitro model based on human induced pluripotent stem cellsderived cardiomyocytes (hiPSC-CMs), stimulated with 10 nM ET-1 ( Figure 5A). HiPSC-CMs

Expression of MicroRNAs in a Hypertrophy Model Using hiPSC-Derived Cardiomyocytes
Next, to investigate a possible role of the selected miRNAs in cardiomyocyte hypertrophy, a feature of CCC [28], we used an in vitro model based on human induced pluripotent stem cells-derived cardiomyocytes (hiPSC-CMs), stimulated with 10 nM ET-1 ( Figure 5A). HiPSC-CMs presented spontaneous contractility and cardiac troponin T expression ( Figure 5B-D). In addition, hiPSC-CMs were positively stained for the cardiomyocyte markers alpha-actinin, cardiac troponin T, GATA-4, and MF-20 ( Figure 5E-G). Cardiomyocyte hypertrophy induction by ET-1 stimulation was confirmed by increased NPPB expression and NT-pro-BNP levels ( Figure 5H-I). Moreover, ET-1-stimulated cardiomyocytes presented a significant increase in cell size, confirming a hypertrophic phenotype ( Figure 5J). stimulated cardiomyocytes presented a significant increase in cell size, confirming a hypertrophic phenotype ( Figure 5J). After validating the in vitro hypertrophy model, we evaluated the selected miRNAs. A significant increase in the expression of miR-19a-3p, miR-21-5p, miR-29b-3p and miR-199b-5p (p < 0.05) was observed in cardiomyocytes stimulated with ET-1 compared to untreated controls ( Figure  6A-D). After validating the in vitro hypertrophy model, we evaluated the selected miRNAs. A significant increase in the expression of miR-19a-3p, miR-21-5p, miR-29b-3p and miR-199b-5p (p < 0.05) was observed in cardiomyocytes stimulated with ET-1 compared to untreated controls ( Figure 6A-D).

Discussion
The potential role of miRNAs as novel biomarkers and therapeutic targets have been extensively investigated in the cardiovascular field [17]. Chagas disease, however, is a neglected disease, and, therefore, there are only two reports of miRNA expression in human samples, both focusing on cardiac tissue analyses [24,29]. The present study is the first to investigate the expression of circulating miRNAs -miR-19a-3p, miR-21-5p, miR-29b-3p, and miR-199b-5p in subjects with Chagas disease, and to perform correlation analyses with relevant clinical parameters, aiming at evaluating their potential role as biomarkers for cardiac dysfunction.
Our results revealed that miR-19a-3p expression was higher in subjects with CCC compared to those with the indeterminate form and showed correlations with different clinical features of disease severity. Previously, miR-19a was associated with cardiomyocyte hypertrophy in vitro [30] and demonstrated increased cardiac expression in a mouse model of hypertrophic cardiomyopathy induced by aortic constriction [31]. These data corroborate the findings from our in vitro hypertrophy model with hiPSC-CM, which showed increased miR-19a-3p expression, along with BNP, and increased cardiomyocyte area after ET-1 exposure. MiR-19a was also found to be overexpressed in plasma of patients with heart failure [18] and was indicated as a potential novel biomarker for early diagnosis of acute myocardial infarction [32]. In the present study, miR-19a-3p expression revealed a positive correlation with the percentage of fibrosis and NYHA functional class, and a negative correlation with left ventricular strain and ejection fraction. Moreover, we found that miR-19a-3p was upregulated in the heart tissue of subjects with end-stage CCC, which presented typical histological characteristics of CCC: cardiomyocyte hypertrophy, chronic inflammation, and fibrosis [33]. These data suggest that miR-19a-3p could be further explored as a cardiac dysfunction progression marker in Chagas disease in a longitudinal study, especially focused on subjects in the indeterminate form, who may progress to a cardiac form.
In the present study, we found a negative correlation among serum miR-29b-3p expression and ejection fraction, and a positive correlation with cardiac fibrosis, NYHA functional, and cardiac dysfunction. We also demonstrated increased miR-29b-3p expression in heart samples from patients with end-stage CCC. Additionally, using an in vitro cardiomyocyte hypertrophy model, increased levels of miR-29b-3p were also observed. Similarly, a recent study reported that miR-29 promotes hypertrophic growth of cardiomyocytes and suggested that in vivo anti-miR-29 delivery might have therapeutic value for pathological cardiac remodeling and fibrosis [34].
The association between miR-21 and cardiac fibrosis in subjects with heart failure has been previously reported [35][36][37]. For the first time, we showed that miR-21-5p is also upregulated in the heart tissue of subjects with end-stage CCC. Additionally, our study confirmed that miR-21-5p is overexpressed in human cardiac fibroblasts following stimulation with TGFβ, a cytokine involved with heart fibrogenesis in different disease settings, including Chagas disease [27,38]. Furthermore,

Discussion
The potential role of miRNAs as novel biomarkers and therapeutic targets have been extensively investigated in the cardiovascular field [17]. Chagas disease, however, is a neglected disease, and, therefore, there are only two reports of miRNA expression in human samples, both focusing on cardiac tissue analyses [24,29]. The present study is the first to investigate the expression of circulating miRNAs -miR-19a-3p, miR-21-5p, miR-29b-3p, and miR-199b-5p in subjects with Chagas disease, and to perform correlation analyses with relevant clinical parameters, aiming at evaluating their potential role as biomarkers for cardiac dysfunction.
Our results revealed that miR-19a-3p expression was higher in subjects with CCC compared to those with the indeterminate form and showed correlations with different clinical features of disease severity. Previously, miR-19a was associated with cardiomyocyte hypertrophy in vitro [30] and demonstrated increased cardiac expression in a mouse model of hypertrophic cardiomyopathy induced by aortic constriction [31]. These data corroborate the findings from our in vitro hypertrophy model with hiPSC-CM, which showed increased miR-19a-3p expression, along with BNP, and increased cardiomyocyte area after ET-1 exposure. MiR-19a was also found to be overexpressed in plasma of patients with heart failure [18] and was indicated as a potential novel biomarker for early diagnosis of acute myocardial infarction [32]. In the present study, miR-19a-3p expression revealed a positive correlation with the percentage of fibrosis and NYHA functional class, and a negative correlation with left ventricular strain and ejection fraction. Moreover, we found that miR-19a-3p was upregulated in the heart tissue of subjects with end-stage CCC, which presented typical histological characteristics of CCC: cardiomyocyte hypertrophy, chronic inflammation, and fibrosis [33]. These data suggest that miR-19a-3p could be further explored as a cardiac dysfunction progression marker in Chagas disease in a longitudinal study, especially focused on subjects in the indeterminate form, who may progress to a cardiac form.
In the present study, we found a negative correlation among serum miR-29b-3p expression and ejection fraction, and a positive correlation with cardiac fibrosis, NYHA functional, and cardiac dysfunction. We also demonstrated increased miR-29b-3p expression in heart samples from patients with end-stage CCC. Additionally, using an in vitro cardiomyocyte hypertrophy model, increased levels of miR-29b-3p were also observed. Similarly, a recent study reported that miR-29 promotes hypertrophic growth of cardiomyocytes and suggested that in vivo anti-miR-29 delivery might have therapeutic value for pathological cardiac remodeling and fibrosis [34].
The association between miR-21 and cardiac fibrosis in subjects with heart failure has been previously reported [35][36][37]. For the first time, we showed that miR-21-5p is also upregulated in the heart tissue of subjects with end-stage CCC. Additionally, our study confirmed that miR-21-5p is overexpressed in human cardiac fibroblasts following stimulation with TGFβ, a cytokine involved with heart fibrogenesis in different disease settings, including Chagas disease [27,38]. Furthermore, TGFβ signaling pathway has been demonstrated to be associated with miR-21 expression in fibrogenesis [39][40][41]. Interestingly, we also observed increased miR-21-5p in CCC subjects, when compared to indeterminate form subjects, positive correlation with cardiac dysfunction, NYHA class, and cardiac fibrosis, and negative correlation with ejection fraction and left ventricular. A previous report showed the association between miR-21 and hypertrophy in murine and human models during cardiac remodeling and heart failure [42]. These data corroborate our findings using an in vitro model of cardiomyocyte hypertrophy using hiPSC-CM. High miR-21 expression promotes cardiac hypertrophy and fibrosis and may be a key regulator in heart failure [36]. Evaluation of miRNA expression levels in hearts from T. cruzi-infected mice also demonstrated miR-21 upregulation in the acute phase of Chagas disease model [43].
Calcineurin/NFAT signaling is apparently regulated by the expression levels of miR-199b-5p in cardiac remodeling and dysfunction [23]. In our study, circulating miR-199b-5p levels were not increased in Chagas disease patients, when compared to healthy controls. However, we found that miR-199b-5p was upregulated in the heart tissue of subjects with end-stage CCC, and in our in vitro model, suggesting its involvement with cardiomyocyte hypertrophy. Further studies with increased sample size are necessary to clarify these results.
MiR-208a, a heart-specific miRNA, was correlated with the severity of coronary heart disease [44]. In our study, miR-208a-3p was not detected in most of the serum samples. A previous study reported upregulation of miR-208a-3p in plasma samples from patients with the indeterminate form of Chagas disease [25]. Based on these findings, a larger cohort comparing the expression levels of miR-208a-3p in plasma and serum would be relevant to better define the role of this miR and its value as a circulating biomarker. Previously, miR-208a was found to be upregulated in cardiac tissues obtained from patients with CCC [24]. In accordance, our study confirmed the expression of miR-208a previously described in heart samples from subjects with CCC.
Reinforcing our findings in serum samples, we observed increased expression of miRNAs in our in vitro models of cardiac fibrosis and hypertrophy. We found elevated expression of miR-19a-3p, miR-21-5p, miR-29b-3p, and miR-199b-5p in hiPSC-CM, whereas miR-21-5p was upregulated in cardiac fibroblasts. These results are in accordance with previous data showing altered miRNA expression in the same in vitro model of cardiac fibroblast [45] and cardiac hypertrophy using hiPSC-CM [46].
In our study, high expression levels of miR-29b-3p and miR-30a-5p were observed in explanted hearts of end-stage CCC patients. The involvement of a member of the miR-30 family in cardiomyocyte hypertrophy induction has been previously demonstrated [47], and miR-29 and miR-30 presented anti-fibrotic effects [48]. Additional studies are needed to better understand the roles of these miRNAs, since different miRNA expression patterns may reflect specific aspects of the complex pathogenesis and signaling pathways involved in the CCC.
It is well established that microRNAs can be detected in the blood [49], plasma [25], and serum [15]. However, a significant variation of the results could be observed, which may be due to different detection sensitivities of the circulating microRNAs measured in serum and plasma [11,26]. Nevertheless, in our study, the serum samples showed a higher miRNA concentration compared to plasma samples for most of the miRNAs evaluated. This difference in circulating miRNA concentrations in biofluids may be explained by the microRNA trafficking system between cellular compartments, extracellular environment, and coagulation process [26,50].
In conclusion, our results demonstrate alterations in the expression levels of circulating miRNAs in Chagas disease, some of which correlated with cardiac tissue expression profile and were shown to be regulated in cardiac fibroblasts and human cardiomyocytes, under pro-fibrotic and pro-hypertrophic conditions in vitro. Importantly, the increased circulating levels of miR-19a-3p, miR-21-5p, and miR-29b-3p correlated with cardiac injury and disease severity in Chagas disease. It is important to recognize that the present study is exploratory, with a small sample size and a cross-sectional design, in which a single time point measurement of the microRNAs was performed. These miRNAs could be further explored in a large cohort of cases as novel biomarkers or validated as molecular targets for therapeutic intervention.

Ethics and Study Design
This is an exploratory study consisting of the analysis of circulating microRNAs followed by heart tissue analysis and functional in vitro assays. The procedures complied with the Declaration of Helsinki and received prior approval by the Ethics Committee of the São Rafael Hospital (Protocol # 20023913.6.0000.0048, 07/30/2013), and has been registered at ClinicalTrials.gov under the identifier, NCT01842880. All subjects signed the written informed consent before inclusion in the study. Heart samples were obtained from the Heart Transplant Service at Messejana Hospital, Fortaleza, Brazil, with approval by the São Rafael Hospital's Ethics committee (Protocol # 51025115.3.0000.0048).

Study Population
The circulating microRNAs were evaluated in 42 subjects divided into three groups: i. Chagas disease, cardiac form with left ventricular dysfunction, CCC (n = 28), ii. Chagas disease, indeterminate form (n = 10), and iii. healthy controls (n = 10). The participants were selected from a database of outpatient clinics from São Rafael Hospital (Table 1).
Subjects with CCC were included according to the following criteria: diagnosis of Chagas disease confirmed by two methods: indirect hemagglutination and indirect immunofluorescence, symptomatic heart failure (NYHA classes II, III and IV), left ventricular ejection fraction ≤55%, measured by Doppler echocardiogram -Simpson method, presence of myocardial fibrosis visualized as delayed enhancement in cardiac magnetic resonance imaging (MRI). Subjects with the indeterminate form of Chagas disease were selected using the following inclusion criteria: diagnosis of Chagas disease confirmed by indirect hemagglutination and indirect immunofluorescence; absence of clinical diagnosis of heart failure; absence of abnormalities in echography, Holter and MRI. The inclusion criteria for healthy controls were: age > 40 years, absence of clinical heart failure diagnosis or other known medical conditions, and negative result of a serological test for Chagas disease.
Heart tissue samples were obtained from 17 patients submitted to heart transplantation at the Messejana Hospital Heart Transplant Unit in Fortaleza, Brazil. Explanted cardiac tissue samples obtained from ventricles of subjects with end-stage CCC (n = 9) and normal control heart tissue samples (n = 8) were evaluated for the expression of miRNAs by RT-qPCR. Control heart tissues were chosen from biopsies of donor hearts without any abnormal histological findings.

Total RNA, MicroRNA Extraction and cDNA Synthesis
Exosomes were isolated from serum and plasma samples using the miRCURY™ Exosome Isolation Kit (Exiqon) and 0.5 mL serum/plasma from Chagas disease subjects and healthy controls, following the manufacturer's instructions. MicroRNA was isolated from exosome preparations, cells, or heart tissue using the miRCURY ™ RNA Isolation Kit (Exiqon). Synthesis of cDNA was performed using the Universal cDNA Synthesis kit (Exiqon). Total RNA was extracted using Trizol®(ThermoFisher Scientific, Waltham, MA, USA) and cDNA was synthesized using High-Capacity cDNA Reverse Transcription Kit (ThermoFisher Scientific).

Quantitative Real-time PCR (RT-qPCR)
The expression levels of microRNAs were evaluated in serum and plasma samples. Quantitative real-time PCR assays were performed with miRCURY LNA™ Universal RT microRNA PCR with SYBR®and LNA™ primers set (Exiqon). Five micro RNAs (miR-16, miR-93, miR-423, cel-miR-39-3p, and SNORD44) were selected from the literature [58,59] and used for normalization upon demonstration of stable expression among the samples. Spike-in RNAs were used to calibrate the results (Exiqon).
Cardiac fibroblasts were incubated with 10 ng/mL recombinant human TGF-β1 (Peprotech, Rocky Hill, NJ, USA) for 24 h at 37 • C incubation and 5% CO 2 for fibroblast activation. Type I collagen (COL1A1) gene expression was evaluated by RT-qPCR, as described above. Protein expression was analyzed with the Operetta High Content Screening, using the Harmony high-content analysis software (Perkin Elmer, Waltham, MA, USA).

Modeling Human Cardiac Hypertrophy in hiPSC-CM
Two human induced pluripotent stem cell (hiPSC) lines obtained from healthy male donors were used in this study. The cells were previously generated from the reprogramming of somatic cells (skin fibroblasts and peripheral blood mononuclear cells) with a non-integrating cell reprogramming method, as previously described [61]. HiPSCs were maintained on plates coated with Matrigel (Corning, New York, NY, USA) in Essential 8 medium (ThermoFisher Scientific), in a humidified atmosphere containing 5% CO 2 and at 37 • C. Cardiomyocyte differentiation was performed using PSC Cardiomyocyte Differentiation Kit (ThermoFisher Scientific).
For functional analysis, a hypertrophic response was induced in hiPSC-CM by incubation with 10 nM endothelin-1 (ET-1, Sigma-Aldrich), as previously described [30]. Cell-free culture supernatant was collected 24 h later for N-terminal prohormone of brain natriuretic peptide (NT-pro-BNP) measurement (Vidas®NT-pro-BNP, Biomerieux, Marcy, France). Additional wells were stained with anti-cTNT for cell size measurements in the Operetta High Content Screening (Perkin Elmer), using the Harmony High-Content Analysis software (Perkin Elmer).

Statistical Analyses
Categorical data were presented as numbers (percentages), and continuous data were expressed as mean (SD) or median (interquartile range). Comparisons of continuous variables among groups were performed with analysis of variance (ANOVA) test or Kruskal-Wallis, depending on normality assessed by the Shapiro-Wilk test. Chi-Square or Fisher tests were applied for proportion comparisons. Correlations between continuous variables were evaluated by Pearson or Spearman coefficients. Analyses of candidate microRNAs and clinical parameters were performed using Statistic Data Analysis STATA ® and p < 0.05 (two-tailed) was considered statistically significant.