WNT/β-Catenin Signaling Promotes TGF-β-Mediated Activation of Human Cardiac Fibroblasts by Enhancing IL-11 Production

Cardiac fibrosis is a pathological process associated with the development of heart failure. TGF-β and WNT signaling have been implicated in pathogenesis of cardiac fibrosis, however, little is known about molecular cross-talk between these two pathways. The aim of this study was to examine the effect of exogenous canonical WNT3a and non-canonical WNT5a in TGF-β-activated human cardiac fibroblasts. We found that WNT3a and TGF-β induced a β-catenin-dependent response, whereas WNT5a prompted AP-1 activity. TGF-β triggered profibrotic signatures in cardiac fibroblasts, and co-stimulation with WNT3a or co-activation of the β-catenin pathway with the GSK3β inhibitor CHIR99021 enhanced collagen I and fibronectin production and development of active contractile stress fibers. In the absence of TGF-β, neither WNT3a nor CHIR99021 exerted profibrotic responses. On a molecular level, in TGF-β-activated fibroblasts, WNT3a enhanced phosphorylation of TAK1 and production and secretion of IL-11 but showed no effect on the Smad pathway. Neutralization of IL-11 activity with the blocking anti-IL-11 antibody effectively reduced the profibrotic response of cardiac fibroblasts activated with TGF-β and WNT3a. In contrast to canonical WNT3a, co-activation with non-canonical WNT5a suppressed TGF-β-induced production of collagen I. In conclusion, WNT/β-catenin signaling promotes TGF-β-mediated fibroblast-to-myofibroblast transition by enhancing IL-11 production. Thus, the uncovered mechanism broadens our knowledge on a molecular basis of cardiac fibrogenesis and defines novel therapeutic targets for fibrotic heart diseases.


Introduction
Cardiac fibrosis refers to the excessive accumulation of stromal cells and extracellular matrix (ECM) components in the myocardium and is often associated with cardiac pathologies. On the one hand, fibrotic scar develops following myocardial infarction to support tissue integrity and regeneration. On the other hand, hypertension, cardiomyopathy, local and systemic inflammation, co-morbidities, certain medications and aging may cause pathological tissue remodeling, affecting cardiac function [1]. Increased stiffness of cardiac tissue and disturbed propagation of electrical impulses are direct outcomes of fibrosis, and these changes may lead to heart failure and arrhythmias [2]. Consequently, patients with persistent cardiac fibrosis show longer hospitalization and increased mortality [3]. The prevalence of myocardial fibrosis varies depending on age and the occurrence of cardiovascular diseases. Late gadolinium-enhancement cardiac magnetic resonance identified fibrotic scars in up to one fourth of infarcted hearts [4], and in 31-38% of non-ischemic cardiomyopathy patients [5,6]. Furthermore, post-mortem histopathology studies reported fibrosis in every fifth heart of sudden cardiac death [7] and COVID-19 victims [8].

Canonical and Non-Canonical WNT Responses in Human Cardiac Fibroblasts
Before investigating the effect of WNT signaling on the differentiation of human cardiac fibroblasts, we addressed their responsiveness to exogenous WNT3a and WNT5a. In this approach, we used the conditional medium of L-WNT3a, L-WNT5a (referred later as WNT3a and WNT5a, respectively) and of control L cell lines (used as controls for WNT3a and WNT5a), as well as TGF-β, which is known to induce endogenous WNT production. Nuclear translocation of β-catenin represents the most characteristic feature of the canonical WNT signaling. As expected, both WNT3a and TGF-β, but not WNT5a, induced nuclear translocation of β-catenin in cardiac fibroblasts ( Figure 1A). In line with this finding, treatment with WNT3a or TGF-β induced expression of the β-catenin-responsive genes AXIN2 and TCF7 ( Figure 1B), confirming activation of canonical WNT pathway. For the activation of the non-canonical WNT pathway, we analyzed the expression of ATF2 and JUN and measured activity of the AP-1 transcription factor by using human cardiac fibroblasts transfected with the AP-1 reporter system (AP-1-hCF). WNT5a induced expression of ATF2 and JUN in cardiac fibroblasts ( Figure 1B). Furthermore, we observed significantly upregulated AP-1 activity in AP-1-hCFs following treatment with WNT5a or TGF-β ( Figure 1C). Collectively, our data confirmed that human cardiac fibroblasts differentially respond to WNT3a and WNT5a, and that TGF-β could effectively activate canonical and non-canonical WNT pathways in these cells.
WNTs are typically produced by activated cells and act as secondary triggers. Therefore, the effect of exogenous WNTs was studied in TGF-β-activated fibroblasts additionally stimulated with WNT3a or WNT5a. Stimulation with WNT3a resulted in deregulation of 124 genes in TGF-β-treated fibroblasts (Supplementary Figure S2). Pathway analysis confirmed activation of WNT signaling (GO:0016055, 12 genes, p value ≤ 0.0001) and pointed to the induction of cell differentiation (GO:0030154, 26 genes, p value ≤ 0.0001) and cellular component organization (GO:0051128, 13 genes, p value ≤ 0.001). WNT3a specifically up-regulated genes involved in stress fibers formation, such as ACTA2 (encoding alpha smooth muscle actin; αSMA), ACTG2 (encoding gamma smooth muscle actin; αSMA) and VCL (encoding vinculin), as well as LTBP1 (encoding latent-TGF-β-binding protein 1), which regulates TGF-β bioavailability (Figure 2A). In contrast to the profound effect of WNT3a, treatment with WNT5 upregulated expression of only 2 and downregulated 21 genes in TGF-β-activated cells ( Figure 2B). Altogether, these results pointed to the relevance of the WNT canonical arm in enhancing TGF-β-mediated profibrotic responses in cardiac fibroblasts.

WNT/β-Catenin Signaling Promotes TGF-β-Dependent Fibroblast-to-Myofibroblast Transition
Our previous data showed that WNT3a stimulated production of αSMA and type I collagen in TGF-β-activated fibroblasts [29]. Following these results and the described above transcriptomics data, we measured the formation of αSMA-positive stress fibers by immunocytochemistry. WNT3a in the absence of TGF-β failed to increase the number of αSMA-positive cardiac fibroblasts, whereas in TGF-β-activated cells, WNT3a effectively boosted formation of αSMA-positive fibers ( Figure 3A), which is a characteristic feature of myofibroblasts. Stress fibers are the major contractile structures in the cell; therefore, in the next step, we measured functional contraction of cardiac fibroblasts in a collagen gel. Obtained data confirmed that WNT3a enhanced contractile properties of TGF-β-activated fibroblasts ( Figure 3B).
Next, we asked whether a specific pharmacological activation of the β-catenin-dependent pathway with the GSK3β inhibitor CHIR99021 could promote myofibroblast phenotype in TGF-β-activated cells. Treatment with CHIR99021 indeed up-regulated expression of ACTA2, FN1 and COL1A1 ( Figure 3C), as well as increased the corresponding intracellular protein levels of αSMA, fibronectin ( Figure 3D) and the amount of secreted pro-collagen I α1 ( Figure 3E) in TGF-β-activated fibroblasts. Of note, CHIR99021 in non-activated cells failed to trigger the myofibroblast signature.
Transcriptomics data suggested that WNT5a was not associated with profibrotic response. In fact, treatment with WNT5a suppressed TGF-β-induced expression of ACTA2 and COL1A1 ( Figure 4A). On a protein level, however, we found no difference in αSMA levels between cells treated with TGF-β and with TGF-β and WNT5a ( Figure 4B). Instead, WNT5a suppressed secretion of pro-collagen I α1 in TGF-β-activated fibroblasts ( Figure 4C). Collectively, these data indicate that specific activation of the WNT/β-catenin pathway promotes fibroblast-to-myofibroblast transition induced by TGF-β. αSMA) and VCL (encoding vinculin), as well as LTBP1 (encoding latent-TGF-β-binding protein 1), which regulates TGF-β bioavailability (Figure 2A). In contrast to the profound effect of WNT3a, treatment with WNT5 upregulated expression of only 2 and downregulated 21 genes in TGF-β-activated cells ( Figure 2B). Altogether, these results pointed to the relevance of the WNT canonical arm in enhancing TGF-β-mediated profibrotic responses in cardiac fibroblasts.

WNT3a Boosts TAK1 Signaling and IL-11 Production in Activated Cardiac Fibroblasts
In the next step, we aimed to identify the molecular mechanism by which WNT3a enhances the profibrotic TGF-β response. In the canonical response, TGF-β activates Smad-dependent signaling. In contrast to TGF-β, WNT3a did not induce phosphorylation of Smad2 and co-stimulation with WNT3a and TGF-β showed similar activation of Smad2 to stimulation with TGF-β alone ( Figure 5A). Treatment with TGF-β activates not only Smad-dependent, but also non-canonical Smad-independent pathways, such as TAK1 signaling. Even though WNT3a alone failed to activate TAK1, we observed that in the presence of TGF-β, WNT3a effectively enhanced phosphorylation of TAK1 ( Figure 5B). These results suggest that WNT3a is implicated in strengthening the non-canonical arm of the TGF-β signaling in cardiac fibroblasts.
Recently, IL-11 has been implicated in profibrotic TGF-β responses in cardiac fibroblasts [17]; therefore, we addressed whether WNT3a could regulate IL-11. We found that although WNT3a alone did not regulate production and secretion of IL-11, it enhanced IL-11 production in TGF-β-activated cells ( Figure 5C). As TGF-β induces WNT secretion in profibrotic responses, we studied how the blockade of WNT secretion with the porcupine inhibitor WNT-C59 affected TGF-β-induced IL-11 production. We found that treatment with WNT-C59 reduced IL-11 levels in TGF-β-activated cardiac fibroblasts ( Figure  5D). These results indicate that canonical WNT signaling controls TGF-β-dependent IL-11 production. shows pro-collagen I α1 levels measured in supernatants of cardiac fibroblasts treated as indicated for 5 days (n = 5). Bars and whiskers represent means ±SEM. Main effect of WNT5a, TGF-β and their interaction were tested using two-way ANOVA followed by Tukey's multiple comparisons test, when the p-value for interaction was < 0.05, * p < 0.05 (post-hoc TGF-β vs. TGF-β+WNT5a).

WNT3a Boosts TAK1 Signaling and IL-11 Production in Activated Cardiac Fibroblasts
In the next step, we aimed to identify the molecular mechanism by which WNT3a enhances the profibrotic TGF-β response. In the canonical response, TGF-β activates Smaddependent signaling. In contrast to TGF-β, WNT3a did not induce phosphorylation of Smad2 and co-stimulation with WNT3a and TGF-β showed similar activation of Smad2 to stimulation with TGF-β alone ( Figure 5A). Treatment with TGF-β activates not only Smad-dependent, but also non-canonical Smad-independent pathways, such as TAK1 signaling. Even though WNT3a alone failed to activate TAK1, we observed that in the presence of TGF-β, WNT3a effectively enhanced phosphorylation of TAK1 ( Figure 5B). These results suggest that WNT3a is implicated in strengthening the non-canonical arm of the TGF-β signaling in cardiac fibroblasts.

IL-11 Mediates Profibrotic WNT3a-Mediated Response in Activated Cardiac Fibroblasts
Our data suggested that WNT3a-dependent production of IL-11 might contribute to profibrotic responses in cardiac fibroblasts. To address this hypothesis, we blocked IL-11 Recently, IL-11 has been implicated in profibrotic TGF-β responses in cardiac fibroblasts [17]; therefore, we addressed whether WNT3a could regulate IL-11. We found that although WNT3a alone did not regulate production and secretion of IL-11, it enhanced IL-11 production in TGF-β-activated cells ( Figure 5C). As TGF-β induces WNT secretion in profibrotic responses, we studied how the blockade of WNT secretion with the porcupine inhibitor WNT-C59 affected TGF-β-induced IL-11 production. We found that treatment with WNT-C59 reduced IL-11 levels in TGF-β-activated cardiac fibroblasts ( Figure 5D). These results indicate that canonical WNT signaling controls TGF-β-dependent IL-11 production.

IL-11 Mediates Profibrotic WNT3a-Mediated Response in Activated Cardiac Fibroblasts
Our data suggested that WNT3a-dependent production of IL-11 might contribute to profibrotic responses in cardiac fibroblasts. To address this hypothesis, we blocked IL-11 activity with an anti-IL-11 neutralizing antibody in cardiac fibroblasts activated with TGF-β and WNT3a and analyzed their fibrotic profile. Indeed, the neutralizing anti-IL-11 antibody effectively suppressed transcription of profibrotic genes ACTA2, COL1A1, FN1 and ACTG2 in cells co-stimulated with TGF-β and WNT3a ( Figure 6A). In line with regulation of ACTA2 and FN1 transcripts, inhibition of IL-11 reduced protein levels of αSMA and fibronectin ( Figure 6B). Furthermore, inhibition of IL-11 in TGF-β-activated and WNT3a-boosted cardiac fibroblasts reduced secretion of pro-collagen I α1 ( Figure 6C) and suppressed their contractile properties ( Figure 6D). Collectively, obtained data confirmed the profibrotic activity of IL-11 produced by cardiac fibroblasts activated with TGF-β and WNT3a. activity with an anti-IL-11 neutralizing antibody in cardiac fibroblasts activated with TGFβ and WNT3a and analyzed their fibrotic profile. Indeed, the neutralizing anti-IL-11 antibody effectively suppressed transcription of profibrotic genes ACTA2, COL1A1, FN1 and ACTG2 in cells co-stimulated with TGF-β and WNT3a ( Figure 6A). In line with regulation of ACTA2 and FN1 transcripts, inhibition of IL-11 reduced protein levels of αSMA and fibronectin ( Figure 6B). Furthermore, inhibition of IL-11 in TGF-β-activated and WNT3aboosted cardiac fibroblasts reduced secretion of pro-collagen I α1 ( Figure 6C) and suppressed their contractile properties ( Figure 6D). Collectively, obtained data confirmed the profibrotic activity of IL-11 produced by cardiac fibroblasts activated with TGF-β and WNT3a.  . Panel (D) shows quantification of gel contraction (n = 3) containing cardiac fibroblasts pre-treated as indicated. Bars and whiskers represent means ±SEM. The main effect of WNT3a-TGF-β, the anti-IL11 antibody and their interaction were tested using two-way ANOVA followed by Tukey's multiple comparisons test, when the p-value for interaction was ≤ 0.1; * p < 0.05 (post-hoc TGF-β+WNT3a vs. TGF-β+WNT3a+anti-IL11).

Discussion
Activation of cardiac fibroblasts and their transition into the myofibroblast phenotype plays a key role in cardiac fibrosis. TGF-β and WNT ligands have been identified as important profibrotic factors involved in this process, however, little is known about molecular cross-talk between these two pathways in cardiac fibrogenesis. Our data showed that specific activation of WNT/β-catenin pathways (either with WNT3a or with GSK3β inhibitor CHIR99021) promoted the fibroblast-to-myofibroblast transition, mainly by enhancing TGFβ-induced IL-11 production. In particular, WNT/β-catenin signaling specifically promoted formation of functional contractile stress fibers, an important feature of mature myofibroblasts. Previously, we showed that cardiac fibroblasts produced profibrotic WNTs upon treatment with TGF-β [29]. Here, we confirmed that also exogenous canonical WNT3a could promote fibrotic processes in cardiac fibroblasts. This is not surprising because WNT ligands are molecules that exert their signaling functions through the extracellular space. Of note, WNT secretion is a tightly regulated process resulting in controlled secretion either of soluble WNT proteins or WNTs incorporated into extracellular vesicles [29,30].
Previous studies highlighting the importance of WNT/β-catenin pathway in cardiac fibrosis were commonly based on loss-of-function approaches [16,24]. Our findings not only complement published results by gain-of-function data, but also point to the involvement of multiple profibrotic pathways in the fibroblast-to-myofibroblast transition. In fact, the inability of WNT3a to trigger fibrotic responses in non-activated fibroblasts (i.e., in the absence of TGF-β) was an important finding of this study. On a molecular level, WNT3a, unlike TGF-β, was unable to activate the Smad pathway. In contrast to our findings, treatment of mouse cardiac fibroblasts with recombinant WNT3a (in the absence of any other stimulus) activated the myofibroblast phenotype by stimulating TGF-β production and activating Smads [31]. These differences might result from differential activation status of the fibroblasts due to culture conditions, isolation procedures and/or cellular origins. Less likely, it might originate from the difference in WNT/β-catenin signaling between mouse and human cells. Nevertheless, results of both studies point to cooperative activity of TGF-β and WNT3a in the cardiac fibroblast-to-myofibroblast transition. We think that the efficient phenotypic switch to myofibroblasts requires co-activation of multiple molecular pathways including Smad, β-catenin and TAK1 in cardiac fibroblasts. This multiple pathway co-activation hypothesis is further strengthened by the data showing that blockade of one of these pathways protects cardiac fibroblasts from activation and the heart from fibrosis [12,23,32].
In our experiments, WNT3a showed no effect on activation of the Smad pathway, but enhanced TAK1 signaling in the presence but not in the absence of TGF-β. These findings may suggest that once Smad signaling is on, the output of the β-catenin pathway regulates profibrotic responses in cardiac fibroblasts. It should be noted, however, that in addition to Smads, TGF-β activates non-classical responses dependent on mitogen-activated protein kinases, such as Erk1/2, the JNKs and the p38 isoforms [14]. Regulation of these nonclassical TGF-β pathways by WNT3a in cardiac fibroblasts remains, however, unknown. Regardless of the need for activation of the classical (Smad-dependent) or non-classical arm of TGF-β response, the β-catenin pathway seems to play a central role in regulating the degree of myofibroblast differentiation and thereby fibrosis. This study identifies TAK1 pathway as a molecular target of β-catenin in activated fibroblasts. Interestingly, previous data showed that inhibition of TAK1 signaling blocked the activation of the β-catenin pathway in TGF-β-activated cells [23]. In the light of both findings, a β-catenin-TAK1 feedback loop mechanism could be suggested. At this point, one needs to mention that TGFβ activates β-catenin and our recent study indicated critical involvement of endogenous angiotensin II signaling in this process. Interestingly, cells lacking the angiotensin II receptor 1 showed impaired TAK1 activation following TGF-β treatment [33]. All these data point to the β-catenin-TAK1 positive feedback loop as a novel mechanism in TGF-β-activated cells.
Our study identified canonical WNT as a new regulator of profibrotic IL-11 in TGFβ-activated cardiac fibroblasts. Previous studies showed that IL-11 is a downstream effector of TGF-β-mediated fibrosis in the heart, lungs, aorta and kidneys [17,26,28]. In line with previous findings, we observed that the blockade of IL-11 suppressed profibrotic signature in activated myofibroblasts. As in the case of profibrotic response, WNT3a in the absence of TGF-β was insufficient to regulate IL-11. We consider that IL-11 is directly regulated by β-catenin signaling at the transcriptional level, but its expression requires co-activation of other TGF-β-dependent pathways. Of note, IL-11 regulates Erk1/2 and thereby accelerates one of the non-classical arms of TGF-β signaling. Collectively, these data underline a superior role of TGF-β in cardiac fibrosis and the complexity of the downstream mechanisms. The β-catenin-TAK1 axis triggering IL-11 production seems to represent a major molecular mechanism of TGF-β-dependent fibrosis. Regulation of IL-11 by the β-catenin pathway has not been reported in fibrosis, however, cooperation of TGF-β and WNT in regulating IL-11 was observed in intestinal tumor models [34].
In opposition to profibrotic activity of the canonical WNT3a, a non-canonical WNT5a suppressed the fibroblast-to-myofibroblast transition induced by TGF-β. It should be noticed that WNT5a does not activate the β-catenin pathway and therefore is unable to promote the myofibroblast phenotype. In line with our findings, anti-fibrotic but also anti-angiogenic effects of WNT5a have been also reported in a model of TGF-β-dependent peritoneal injury [35]. Furthermore, published data demonstrated WNT5a controlling IL-6 production in human cardiac fibroblasts [29,36]. It seems that WNT5a is switching the phenotype of cardiac fibroblasts from profibrotic towards proinflammatory. Interestingly, despite IL-11 and WNT5a showing opposite effects on the activation of cardiac fibroblast, the Erk1/2 pathway was identified as a key downstream signal for both of them [17,37]. It suggests that Erk1/2 might act in a pro-or anti-fibrotic way depending on the presence of specific co-activators or the status of activated cells, but any other alternative mechanism cannot be excluded.
Cardiac fibrosis often develops as a response to pathogenic conditions in hearts, such as hypoxia, mechanical or oxidative stress and inflammation. In this light, cell culture models are helpful to study specific molecular mechanisms but do not reflect complexity of fibrogenesis, which in the human heart is not only multifactorial, but is also a long-lasting process. On molecular level, TGF-β signaling is considered as a master profibrotic trigger in the heart, however, there are other secreted biomolecules, such as angiotensin II, damage associated molecular pattern molecules, endothelin-1, connective tissue growth factor and platelet-derived growth factor, cytokines IL-1 or IL-17, that can induce or exacerbate cardiac fibrosis. It should be acknowledged that our data were obtained with fetal cardiac fibroblasts showing premature differentiation status. However, cardiac fibroblasts isolated from adult human hearts once cultured in vitro usually become activated, show the myofibroblast phenotype and respond poorly to TGF-β, and therefore they are not a useful model to study the TGF-β-dependent fibroblast-to-myofibroblast transition. Furthermore, the relevance of in vitro-identified molecular mechanisms needs to be verified in animal models and in humans.
A growing body of evidence indicates elevated levels of multiple WNT ligands in the effected heart [16,32]. Thus, a balance between canonical and non-canonical WNT signaling might be decisive for the fate of activated cardiac fibroblasts in the injured heart and eventually for development of cardiac fibrosis. Accumulating data strongly indicate that the canonical β-catenin pathway is critically involved in pathogenesis of cardiac fibrosis. Herein, by using a model of human cardiac fibroblasts, we elucidated an interplay between TGF-β, WNT/β-catenin and IL-11 and proposed a novel mechanism controlling cardiac fibrosis. This study adds molecular knowledge to our understanding of profibrotic mechanisms in the heart, which is essential for the development of effective cardioprotective therapies targeting cardiac fibrosis.

Luciferase Assay
Luciferase expression in the AP-1 reporter fibroblasts was assessed using the Bright-Glo Luciferase Assay System (Promega, Madison, WI, USA) according to the manufacturer's protocol. Bioluminescence was measured by a SyneryHT plate reader (BioTek, Winooski, VT, USA) or Wallac Victor2 1420 Multilabel counter (Perkin-Elmer, Waltham, MA, USA).

RNA-Seq Analysis
Total RNA from three biological independent experiments of cardiac fibroblasts were sent to Genewiz for preparation in RNA-Seq libraries. Libraries were run with Illumina, strand-specific RNA-Seq with PolyA selection and Illumina NovaSeq 2 × 150 bp sequencing, 10M read pairs. Raw reads were mapped to the Ensembl human GRch38 reference genome using the STAR aligner v.2.5.2b. Unique gene hit counts were calculated by using feature Counts from the Subread package v.1.5.2. After extraction of gene hit counts, the gene hit counts table was used for downstream differential expression analysis. Comparison of the gene expression between the groups was performed using DESeq2. The Wald test was used to generate p values and log2 fold changes. Genes with an adjusted p value < 0.05 and absolute log2 fold change > 1 were called as differentially expressed genes for each comparison. Differentially expressed probes from respective groups are highlighted in the heatmap plot. The heatmap was generated using online tool: https://software.broadinstitute.org/morpheus/. The raw values were transformed, scaled with red representing upregulated genes and blue representing downregulated genes. The GO analyses were performed using online tool: https://tools.dice-database.org/GOnet/ for human species. The raw data file is uploaded to Gene Expression Omnibus (GEO) database with number: GSE181575.

ELISA
Type I collagen was measured in cell culture supernatants using Human Procollagen I alpha 1 DuoSet ELISA (R&D Systems) and IL-11 was measured in cell culture supernatants using IL-11 Human ELISA kit (Thermo Fisher Scientific) according to the manufacturer's protocol.

Collagen Contraction Assay
After stimulations, cells were trypsinized, counted and mixed with the Collagen Gel Working Solution prepared according to manufacturer's protocol (Cell Biolabs Inc., San Diego, CA, USA). The solution was added to 24 well plates and allowed to polymerize for 1 h at 37 • C. Fresh growth media was added to the solidified collagen gels and plates were returned to the incubator. After 24 h, the surface of contraction was released and contraction was monitored over the next 24 h. Next the surface area of contracted gels was imaged using a ChemiDoc instrument (Bio-Rad) and measured using ImageJ software (NIH).

Statistical Analysis
Experimental results are presented as mean± SEM. Raw data normally distributed and data with homogenous variance (Brown-Forsythe test) were compared using a oneway ANOVA followed by Dunn's post-hoc test or two-way ANOVA followed by Tukey's post-hoc test for the indicated comparisons. All analyses were computed using GraphPad Prism 9.1.1 software. Differences were considered as statistically significant for p < 0.05.