Adipose tissue–derived stromal cells’ conditioned medium modulates endothelial‐mesenchymal transition induced by IL‐1β/TGF‐β2 but does not restore endothelial function

Abstract Objectives Endothelial cells undergo TGF‐β–driven endothelial‐mesenchymal transition (EndMT), representing up to 25% of cardiac myofibroblasts in ischaemic hearts. Previous research showed that conditioned medium of adipose tissue–derived stromal cells (ASC‐CMed) blocks the activation of fibroblasts into fibrotic myofibroblasts. We tested the hypothesis that ASC‐CMed abrogates EndMT and prevents the formation of adverse myofibroblasts. Materials and methods Human umbilical vein endothelial cells (HUVEC) were treated with IL‐1β and TGF‐β2 to induce EndMT, and the influence of ASC‐CMed was assessed. As controls, non‐treated HUVEC or HUVEC treated only with IL‐1β in the absence or presence of ASC‐CMed were used. Gene expression of inflammatory, endothelial, mesenchymal and extracellular matrix markers, transcription factors and cell receptors was analysed by RT‐qPCR. The protein expression of endothelial and mesenchymal markers was evaluated by immunofluorescence microscopy and immunoblotting. Endothelial cell function was measured by sprouting assay. Results IL‐1β/TGF‐β2 treatment induced EndMT, as evidenced by the change in HUVEC morphology and an increase in mesenchymal markers. ASC‐CMed blocked the EndMT‐related fibrotic processes, as observed by reduced expression of mesenchymal markers TAGLN (P = 0.0008) and CNN1 (P = 0.0573), as well as SM22α (P = 0.0501). The angiogenesis potential was impaired in HUVEC undergoing EndMT and could not be restored by ASC‐CMed. Conclusions We demonstrated that ASC‐CMed reduces IL‐1β/TGF‐β2‐induced EndMT as observed by the loss of mesenchymal markers. The present study supports the anti‐fibrotic effects of ASC‐CMed through the modulation of the EndMT process.


| INTRODUC TI ON
Heart failure (HF) is an irreversible and potentially lethal clinical condition that affects nearly 23 million people worldwide. 1 The five-year survival is approximately 50%. Obviously, HF impacts significantly on the quality of life and is an increasing burden on society and health care. Heart failure presents as various forms of idiopathic or heritable cardiomyopathy and as the consequence of adverse cardiac tissue remodelling after acute myocardial infarction.
In normal physiology, cardiac tissue is in homeostasis that is maintained by a well-regulated biochemical and biomechanical crosstalk between the parenchyma and the supportive tissue stroma. Cardiac parenchyma comprises cardiomyocytes, while the stroma consists of vasculature, fibroblasts and their product, the extracellular matrix (ECM). The ECM provides structural support and architecture and instructs adhered tissue cells. 2 HF disrupts the cardiac tissue homeostasis. A prominent feature is the proliferation of myofibroblasts and their excessive deposition and accumulation of fibrotic ECM. Thus, HF is a process of cardiac fibrosis.
Differentiation of cardiac fibroblasts to myofibroblasts is a major contribution to HF. 3,4 Several other cell types, both endogenous in the heart and exogenous, also contribute to cardiac fibrosis. 5 The heart is particularly rich in capillaries and thus endothelial cells. Under pathological conditions, endothelial cells contribute to adverse wound healing and tissue remodelling via endothelialmesenchymal transition (EndMT) and contribute significantly to cardiac fibrosis and development of heart failure. 6,7 After acute myocardial infarction in mice, up to 25% of cardiac myofibroblasts are the consequence of EndMT. 8 Irrespective of the source, for example, fibroblasts or endothelial cells, the resulting myofibroblasts are indistinguishable with respect to proliferation and ECM remodelling. Interestingly, EndMT is pivotal during cardiogenesis, when EndMT underlies the development of heart valves. 9 In contrast, in adult life, EndMT is related to pathophysiological phenomena such as cardiac fibrosis, [10][11][12] after myocardial infarction, 8,13 diabetic cardiomyopathy 14,15 and hypertensive heart disease. 11,16 Therefore, inhibition or reversal of cardiac EndMT is a therapeutic option to interfere with heart failure.
Endothelial-to-mesenchymal transition is a relatively slow dedifferentiation process (days to weeks) that is driven by pro-fibrotic growth factors of the TGF-ß superfamily, 17,18 such as TGF-β2. Several processes coincide: endothelial cells loose cell-to-cell contacts and the downregulated endothelial markers. This causes the cells to transit from their characteristic cobblestone morphology to a spindle-like shape.
Simultaneously, a progressive upregulation of mesenchymal markers occurs, such as smooth muscle protein 22 alpha (SM22α), calponin and alpha-smooth muscle actin (αSMA). Similar to myofibroblasts, the EndMT process renders cells highly migratory and proliferative, while these become resistant to apoptosis too. The process of EndMT also coincides with the increased production and deposition of extracellular matrix, which contributes to the development and progression of cardiac fibrosis including the increased stiffness of the failing heart. 19,20 The TGF-β superfamily is important during embryogenesis, but also for wound healing, and thus influences cell growth, proliferation, differentiation and migration. 21 In addition, TGF-β members are strong regulators of ECM remodelling, in particular, through upregulation of constructive proteins such as collagens. All three TGF-β isoforms stimulate EndMT. [22][23][24][25][26][27][28] In cardiovascular wound healing, fibrosis coincides with inflammation. In fact, EndMT is synergized by TGF-β2 and IL-1β. 29,30 Heart failure is also associated with pro-fibrotic stimuli by members of the TGF-β superfamily, inflammation and reactive oxygen species (ROS).
These three triggers are tightly interrelated because TGF-ß promotes inflammatory activation via TAK1, similar to ROS, and with it, EndMT. 31 We have shown that pro-fibrotic stimuli and pro-inflammatory stimuli synergize EndMT, 29,30 and other studies showed that ROS mediates the EndMT process through the TGF-β superfamily. 32,33 Cardiac stem cell therapy with mesenchymal stem/stromal cells (MSC) has shown to improve remodelling after acute myocardial infarction. This suggests that MSC affect myofibroblast formation and function. The intramyocardial administration of mesenchymal stromal cells (MSC), which include adipose tissue-derived stromal cells (ASC), has benefit for cardiac function and remodelling in a variety of cardiac diseases. [34][35][36][37][38][39][40][41][42] As a matter of fact, injection of conditioned medium of MSC (CMed) also improved cardiac function. 43,44 Previous research in our laboratory showed that ASC secrete paracrine factors that abrogate TGF-β-induced differentiation of dermal fibroblasts to myofibroblasts which is a mesenchymal transition too. 45 In general, ASC and their secreted bioactive factors harbour pro-regenerative [46][47][48] and anti-inflammatory potential. 49,50 In addition, ASC promote angiogenesis. Therefore, we hypothesized that the formation of myofibroblasts from endothelial cells via EndMT, which also is a TGF-βdriven process and synergized by IL-1β, is down-modulated by the paracrine action of ASC while it would rescue their endothelial phenotype. We tested our hypothesis in vitro, by assessing the influence of adipose tissue-derived stromal cells' conditioned medium (ASC-CMed) on pro-fibrotic and pro-inflammatory-induced EndMT.  Table 1, and cultured for 5 days. Human recombinant interleukin-1β (IL-1β; #200-01B, PeproTech) and human transforming growth factor beta 2 (TGF-β2; #100-35B, PeproTech) were used to stimulate the EndMT process at a concentration of 10ng/mL in all experiments.

| Cell sources, cell culture, conditioned medium and experimental groups
Human ASC were isolated as described previously. 51 Briefly,

| Immunofluorescence
HUVEC were cultured in 96-well tissue culture plates in ECMed.

| Gene expression analysis
HUVEC were cultured in 75 cm 2 flasks. After 5 days of induction, total RNA was isolated using TRIzol reagent (#15596018, Invitrogen Corp) according to the manufacturer's protocol. RNA concentration and purity were determined using a NanoDrop Spectrophotometer (Thermo Scientific). Between 300 ng and 5000 ng of total RNA was used for cDNA synthesis, which was performed using RevertAid TM First Strand cDNA Synthesis Kit

| Statistical analysis
All data were obtained from at least three independent experiments performed in duplicate. Data are presented as the mean ± SE of the mean (SEM). Graphs and statistical analysis were done using GraphPad Prism (version 6.01; GraphPad Software, Inc). Differences among multiple groups were analysed by one-way ANOVA with Sidak's multiple comparison test for the two groups of interested in each scenario.

| ASC secrete fibroblast growth factor 1 (FGF-1) and vascular endothelial growth factor (VEGF)
The growth factor release from ASC was determined by the measurement of FGF-1 and VEGF, in the medium collected from the cells, using the Magnetic Luminex Human Premixed Multi-Analyte Kit.
The concentration of growth factors was 21.7 ± 0.7 pg/mL for FGF-1 and 95.6 ± 3.1 pg/mL for VEGF. DMEM only showed growth factor concentrations close to zero (Figure 1).

| HUVEC undergoing EndMT present conformational changes
All cells started the experiment as a cobblestone morphology (Figure 2).

| Inflammatory gene expression in activated HUVEC is refractory to ASC-secreted factors
Pro-inflammatory stimulation of HUVEC with IL-1β upregulated expression of IL8, ICAM1 and VCAM1, which encode respectively chemoattractant and adhesion molecules required for endothelial transmigration of activated leucocytes (Figure 3). This upregulation was refractory to simultaneous treatment with ASC-CMed ( Figure 3).
Also, IL-1β stimulation upregulated expression of two pro-inflammatory cytokine genes, IL1B and IL6, which was unaffected by ASC-CMed, except for IL6 that was slightly upregulated by ASC-CMed (one-way ANOVA, P = 0.0075; Sidak's multiple comparison test, P = 0.0683). The influence of TGF-β2 on IL-1β-stimulated HUVEC was negligible with respect to the expression of these inflammatory activation-related genes, neither did co-stimulation with ASC-CMed affect these genes. However, the expression of IL6 was normalized compared to stimulation of HUVEC with IL-1ß and ASC-CMed.

| Mesenchymal gene expression in EndMTinduced HUVEC is suppressed by ASC-secreted factors while extracellular matrix genes are not
Pro-inflammatory stimulation of HUVEC with IL-1β did not change the expression of PECAM1 and CDH5 ( Figure 4A

| Factors secreted by ASC fail to restore impaired sprouting capacity of HUVEC undergoing EndMT
Endothelial cell function was assessed by short-term sprouting on Matrigel® and quantified through determination of nodes, segments, branches, total length, meshes and mean mesh size (Figure 7).

| D ISCUSS I ON
The aim of our investigation was to assess the impact of factors EndMT has been shown as an important process for the generation of myofibroblasts and, thus, fibrosis. 3 The potential of ASC to inhibit EndMT may be one of the mechanisms involved in myocardial regeneration following cell therapies based on ASC. [34][35][36][37][38]41 Literature supports that growth factors known to be secreted by ASC-such as FGF and VEGF 53,54 -could block EndMT. [55][56][57] Besides growth factors, the ASC secretome comprises microRNAs (often secluded in exosomes), among which are miR-155, miR-31, and miR-21, all known regulators of EndMT. 16,58,59 Another mechanism that may influence the expression of SM22ɑ is an epigenetic modification, for instance, the trimethylation of histone three (H3K27me3) by enhancer of zeste homolog 2 (EZH2) 30 Previously, mesenchymal cells derived from menstrual blood (MMC) were shown to ameliorate cardiac fibrosis via inhibition of EndMT in myocardial infarction. 60 The authors showed that the total number of cells co-expressing CD31 and ɑSMA in the infarcted heart was reduced from 30% in the control group to 20% in the group treated with MMC. In our in vitro study, we also showed that the The detailed underlying molecular mechanism of EndMT blockage was not dissected in the present study. We expected a decrease in SNAI1, SNAI2 and TWIST1 expression after use of ASC-CMed because these are transcription factors involved in TGF-β-induced EndMT. 25,65 In contrast, we found that SNAI2 was overexpressed when HUVEC co-stimulated with IL-1β/TGF-β2 were cultured in ASC-CMed, while no differences were found for SNAI1 or TWIST1. Still, it was described in the literature that although EndMT is associated with an increased expression of SNAI2, the overexpression of SNAI2 alone is not enough to promote EndMT, being also required the inhibition of the SNAI2 inhibitor GSK-3β. 23 The GSK-3β, in turn, is inhibited by Smad2/3, 66 which is recruited by TGF-β. 67 Thus, in the hypothesis that ASC-CMed F I G U R E 7 A, Sprouting assay (8 h) of HUVEC under stimulation with IL-1β or co-stimulation with IL-1β/TGF-β2, both in ECMed and in ASC-CMed, for five days. Brightfield microscopy, augmentation 2.5X. Quantification of B, number of nodes, C, number of segments, D, number of branches, E, number of meshes, F, total length and G, mean mesh size. Values represent mean ± SEM of three independent experiments in duplicate would interrupt the canonical TGF-β pathway, Smad2/3 would be decreased and GSK-3β would not be inhibited, consequently blocking the SNAI2. Still, besides the predominant TGF-β canonical pathway, the non-canonical pathway was also described as mediating EndMT. 68,69 Other mechanisms involve the AKT signalling pathway, via the FOXO3 transcription factor, [70][71][72] and the MAPK/ ERK pathway, via the ELK1 transcription factor. 28 Besides these pathways, the study of exosomes and miRNAs has emerged in the past few years, showing the presence of several entities involved in the EndMT process, such as mi21, mi146, let7 12,72,73

| CON CLUS ION
The present study supports the anti-fibrotic effects of ASC-CMed through the modulation of the endothelial-mesenchymal transition process. We demonstrated that ASC-CMed reduces EndMT induced by co-stimulation with IL-1β and TGF-β2 as evidenced by the reduction in expression of mesenchymal markers. Still, further investigations are needed to elucidate the exact underlying mechanisms.

ACK N OWLED G EM ENTS
The authors would like to express their very great appreciation for the assistance provided by Henk Moorlag for the isolation and culture of human umbilical vein endothelial cells (HUVEC).

CO N FLI C T O F I NTE R E S T
The authors declare no conflicts of interest.

AUTH O R CO NTR I B UTI O N
TTAL has contributed to the conception and design of the work, data collection, data analysis and interpretation, drafting the article, critical revision of the article and final approval of the manuscript text.
GRL has contributed to the conception and design of the work, data collection, data analysis and interpretation, drafting the article, critical revision of the article and final approval of the manuscript text.
LFPM has contributed to the data analysis and interpretation, critical revision of the article and final approval of the manuscript text.
MCH has contributed to the conception and design of the work, data analysis and interpretation, critical revision of the article and final approval of the manuscript text.