Endothelial‐to‐mesenchymal transition in systemic sclerosis

Summary

Systemic sclerosis (SSc) is an autoimmune disease characterized by significant vascular alterations and multi‐organ fibrosis. Microvascular alterations are the first event of SSc and injured endothelial cells (ECs) may transdifferentiate towards myofibroblasts, the cells responsible for fibrosis and collagen deposition. This process is identified as endothelial‐to‐mesenchymal transition (EndMT), and understanding of its development is pivotal to identify early pathogenetic events and new therapeutic targets for SSc. In this review, we have highlighted the molecular mechanisms of EndMT and summarize the evidence of the role played by EndMT during the development of progressive fibrosis in SSc, also exploring the possible therapeutic role of its inhibition.

During Systemic Sclerosis (SSc), the alteration of endothelial cells (ECs) functions plays a pivotal role during vascular remodelling, playing as trigger for fibroproliferative disorder. After pathological stimuli, such as TGF‐β, ET‐1, TNF‐α, IL‐1, INF, hypoxia, ROS and miRs, ECs may trans‐differentiate toward myofibroblasts, through the endothelial‐to‐mesenchymal transition (EndMT), losing vascular ECs markers, gaining mesenchymal cell markers and invading the surrounding tissue. A better understanding of EndMT mechanisms is pivotal, to identify clinically useful biomarkers, in order to develop effective anti‐fibrotic therapies during SSc, a condition in which an appropriated therapy is still lacking.

Introduction

Systemic sclerosis (SSc) is an autoimmune disease, characterized by vascular disorder and progressive tissue fibrosis [1]. Recently, a systematic review emphasized the evidence that, during SSc, endothelial cell (EC) dysfunction is the pivotal event contributing to SSc vasculopathy [2]. Much work has confirmed that injured microvascular cells may transdifferentiate towards myofibroblasts, the cells responsible for fibrosis and collagen deposition in the affected tissues [3, 4, 5, 6, 7, 8, 9, 10]. In‐vivo animal models, in which collagen‐producing cells have been tracked, showed that pericytes may differentiate to myofibroblasts, as shown during the development of kidney fibrosis [11, 12, 13, 14, 15]. With regard to ECs, several reports [4, 16, 17] have confirmed the ability of these cells to transdifferentiate towards myofibroblasts through the endothelial‐to‐mesenchymal transition (EndMT), which is a non‐malignant phenomenon of cellular transdifferentiation by which ECs undergo a phonotypical differentiation, losing vascular ECs markers and gaining mesenchymal cell markers [2, 18, 19, 20, 21, 22]. EndMT was first reported in studies on the physiological development of the heart [18, 23] but, to date, this process is considered as a possible pathogenetic mechanism for different conditions, including cardiac fibrosis, kidney fibrosis, diabetic nephropathy, pulmonary hypertension and SSc [24, 25, 26, 27, 28, 29]. Understanding the mechanism and functions of the EndMT process is pivotal to individuate early pathogenetic events and possibly new therapeutic targets for SSc in a very early phase, before the fibrotic process, which may be considered an end‐stage condition [10, 30, 31]. In this review, we highlight the molecular mechanisms of EndMT and summarize the evidence on the role of EndMT in SSc fibrosis. Although EndMT is now considered to be the first step leading to the fibrotic processes of many diseases [18], its role in the pathogenesis of SSc vasculopathy still needs further clarification. The possibility to more clearly identify the mechanistic steps and the related molecules of EndMT in SSc will allow us to develop new therapeutic perspectives for fibrosis in SSc, a condition for which an appropriate therapy is still lacking.

Methods

We designed a comprehensive search of literature on EndMT in SSc because this process plays an important role in the pathogenesis of SSc, linking the two main aspects of the disease, such as vasculopathy and fibrosis, by a review of reports published in indexed international journals until 31 December 2020. We followed the proposed guidelines for preparing a biomedical narrative review [32]. MedLine (via PubMed) and Embase databases were searched. The bibliography of relevant articles was also hand‐searched for identification of other potentially suitable studies.

The EndMT process

The EndMT process is an embryonic physiological method observed in 1975, during the development of heart valves in vertebrate embryos. In this pioneering study, the authors reported that, at day 9 of embryo life, the endocardial cells from the atrioventricular canal and the efferent tract showed cellular hypertrophy, lateralization of the Golgi apparatus, formation of cellular appendages and loss of cell polarity [23]. This observation was further confirmed in chicken embryos, where the ECs derived from the heart underwent morphological transdifferentiation, with activation of migratory phenotype and expression of α‐smooth muscle actin (α‐SMA) [33].

Molecularly, the EC ‘mesenchymal’ phenotype is characterized by the loss of both junctions among cells and EC markers, such as von Willebrand Factor (vWF), CD31, Tie‐1 and Tie‐2 receptors and vascular endothelial‐cadherin (VE‐cadherin), as well as acquisition of the expression of mesenchymal markers such as α‐SMA, smooth muscle 22 (Sm22), CD44, neuronal‐(N‐)cadherin, vimentin, fibroblast‐specific protein‐1 (FSP 1) (S100A4) and collagen [18, 34, 35, 36, 37, 38]. Furthermore, ECs lose their typical cobblestone morphology, acquiring the phenotype and proliferative ability typical of mesenchymal cells, but diminishing barrier skill [39, 40, 41]. During EndMT, the vessel lining is impaired due to disaggregation of ECs from the vessel layer and EC invasion of the surrounding tissue [42, 43, 44, 45] (Fig. 1).

Fig. 1.

Outline of the endothelial‐to‐mesenchymal transition (EndMT) process. (a) In physiological conditions, endothelial cells (ECs) express von Willebrand factor (vWF), Tie‐1 and Tie‐2 receptors, CD31 and vascular endothelial‐cadherin (VE‐cadherin), displaying cobblestone morphology and surrounded by pericytes. (b) After pathological stimuli, such as transforming growth factor (TGF)‐β, endothelin‐1 (ET‐1), tumour necrosis factor (TNF)‐α, interleukin (IL)‐1, interferon (IFN), hypoxia, reactive oxygen species (ROS) and microRNAs (miRs), ECs acquire invasive properties and express mesenchymal markers such as α‐smooth muscle actin (α‐SMA), smooth muscle 22 (Sm22), CD44, N‐cadherin, vimentin and fibroblast‐specific protein‐1 (FSP‐1) (S100A4) and collagen. In these conditions, ECs show the typical elongated shape of mesenchymal cells and, together with surrounding pericytes, may transdifferentiate toward myofibroblasts. (c) ECs, differentiated towards mesenchymal cells, increase their migratory and proliferative capacity, losing their barrier skill. During EndMT, resident ECs disaggregate from the organized cells layer of the vessel walls and invade the surrounding tissue. In this phase, the transdifferentiated cells may release collagen in the tissue, contributing to fibrosis.

The EndMT process was observed not only during physiological development [24, 46] and wound healing [47], but also during pathological processes, characterized by fibrosis, vascular injury and inflammation [24, 25, 26, 27, 28]. The use of lineage tracing models, assessing the EndMT in vivo, confirmed the EC lineage conversion [48]. This strategy has been successfully employed to show the EndMT process in several diseases, including cardiac [18] and kidney [49, 50] fibrosis, vein graft remodelling [51] and cancer [18].

Microvascular damage and remodelling in SSc

Vascular alterations and remodelling are a pivotal event of SSc, observed in different affected organs, including lung, heart, skin and kidney [34, 52, 53, 54, 55, 56, 57]. At present, the SSc vasculopathy, including pulmonary arterial hypertension (PAH), heart involvement and scleroderma renal crisis, is the main cause of disease‐related mortality [58]. Small‐ and medium‐sized arteries may undergo intimal hyperplasia, medial thickening, obliteration of the lumen, perivascular inflammation and microthrombi [58, 59]. Additionally, in the early stages of SSc, the capillary are often enlarged and thinned during the later phases. Despite the persistent hypoxia observed after the loss of the microvasculature during SSc, compensatory angiogenesis did not occur [60]. Hypoxia promotes the release of vascular endothelial growth factor (VEGF) and transforming growth factor (TGF‐β), both promoting the EC activation and their transdiffferentiation towards mesenchymal cells, thus leading to fibrosis [58]. It must be noted that, despite the high levels of VEGF in skin and serum of SSc patients, an impaired angiogenic response persists [61, 62]. This impaired compensatory angiogenesis may be due to an anti‐angiogenic VEGF165b isoform, derived from alternative splicing of VEGF‐A pre‐mRNA, which may be over‐expressed in dermal ECs, fibroblasts and inflammatory cells isolated from SSc patients [63, 64]. Furthermore, it has been shown that platelets may release VEGF165b after activation with the damaged SSc endothelia, probably the main source of this molecule in bloodstream [65, 66].

Recently, our group showed that the deficit in compensatory angiogenesis observed in SSc patients may be related to impaired endothelium–perivascular cell cross‐talk. We have reported that the maintenance and the stabilization of the new vessel is regulated by the cross‐talk between ECs and pericytes via the release of both cytokines and growth factors [13]. This homeostasis is altered during SSc: after injury, the SSc‐ECs release TGF‐β and platelet‐derived growth factor (PDGF‐BB) and these molecules promote pericyte proliferation and transdiffferentiation towards myofibroblasts, thus leading to fibrosis. Additionally, the increased expression of VEGFR receptor‐2 (VEGFR‐2) on pericytes may inhibit PDGF receptor (PDGFR) signalling through the induction of VEGFR‐2/PDGFR complexes, interfering with vessel stabilization and leading to vascular regression [13].

Approach to operations in patients with SSc

SSc patient management during perioperative phases is an important challenge for the clinicians, due to its rare incidence, little evidence‐based guidance and multi‐organ involvement. The extension and the severity of disease should be evaluated during preoperative and anaesthesiological evaluation [67, 68, 69, 70].

Cardiovascular involvement, especially in SSc patients at risk for cardiomyopathy, should be carefully assessed during preoperative phases [71, 72]. During post‐operative vascular implants, including stents, vascular grafts and heart valves, the patients should be monitored because it has been reported that EndMT may be involved in endothelia disfunction leading to in‐stent restenosis. Previous work has shown that substrate stiffness may influence EC behaviour; stronger EndMT was observed when ECs were cultured on stiffer film [73, 74]. Furthermore, conflicting results have been reported concerning the use of silicone, a synthetic polymer retained as an inert substance present in several medical products, including breast implants and insurgence of SSc, especially after implant rupture, which has been supposed to activate cell transdiffferentiation [75, 76].

The facial manifestations of SSc are strongly disabling and severely impair the patients’ self‐image, compromising their quality of life [77]. The face of SSc patients is frequently involved by fibrosis leading to aesthetic and functional concerns [78]. Due to skin thickening, fibrosis may lead to microstomia, restricted mouth opening and poor cervical extension [79, 80]. Additionally, during intra‐operative management of the patients, skin thickening may generate a barrier for intravenous access, leading to a low threshold for ultrasound guidance for vascular catheter insertion [81]. Autologous fat tissue grafting, also known as lipofilling, is one of the most common procedures in the area of plastic surgery used to restore the defect of soft tissue. It has recently been reported that autologous fat grafting improves facial scleroderma from both aesthetic and functional aspects, playing a therapeutic role on fibrosis [82, 83]. The effect of fat tissue grafting depends not only upon its volumizing effect, but also upon its regenerative/reparative effect, probably for the adipose‐derived stem cell (ASC) content [9, 84, 85]. The regenerative effect of ASC, contained in fat tissue, is due the ability of these cells to transdiffferentiate in situ [86], and probably to control inflammation and angiogenesis, as reported during the early phase of tissue repair in experimental peritoneal fibrosis [87]. At present, further investigation is needed to confirm the possible active role of fat tissue grafting on EndMT in SSc patients.

Mediators involved in EndMT

TGF‐β

TGF‐β is a cytokine involved during development of the embryo, cellular differentiation and inflammation, playing a role in fibrotic disease, promoting the release of collagens and preventing the expression of metalloproteinases [58, 88]. The TGF‐β family is currently considered to be the pivotal inducer of EndMT during development of the heart [89], cancer [90] and SSc. In the latter, TGF‐β is considered to be the main actor in fibrotic process [88, 91]. There are three TGF‐β isoforms which bind to the TGF‐β receptor type II, promoting the activation of TGF receptor type I (the so‐called activin receptor‐like kinases, ALKs), thus activating the cell signals via the SMA‐ and mad‐related (Smad) signal pathway family [34]. Although TGF‐β is the pivotal trigger for EndMT, the type I receptors’ contribution to TGF‐β‐induced EndMT is still unknown. Recently, it has been reported that TGF‐β1 and TGF‐β2 promoted EndMT mediated by an increase of ALK5 expression, associated with a decrease of ALK1, in the presence of N‐cadherin over‐expression and inhibition of endothelial nitric oxide synthase (eNOS) [92]. Although all TGF‐β isoforms may induce EndMT, it has been reported that TGF‐β1 is mainly associated with the pathological fibrotic process associated with EndMT, while TGF‐β2 seems to be mainly associated with the physiological EndMT condition, such as embryonic heart development [93]. TGF‐β promotes its profibrotic role, inducing morphological changes in ECs, committing them to a fibrogenic fate and stimulating the transient expression of PDGFR‐α mRNA, resulting in increased expression of the mesenchymal markers associated with a reduced expression of endothelial markers [94, 95, 96].

Endothelin‐1 (ET‐1)

ET‐1 is a potent vasoconstrictor and controls the vascular tone, binding both endothelin receptors A (ETRA) and B (ETRB), thus promoting fibrosis, and possibly contributing to the small‐vessel rarefaction observed in SSc patients [97, 98]. Interestingly, it has been reported that, in skin fibroblasts, TGF‐β may promote the transcription of the ET‐1 gene [52, 99], and ET‐1 may increase the de‐novo synthesis and secretion of TGF‐β [34, 100], suggesting that the TGF‐β/ET‐1 axis in these cells may modulate fibrogenic responses [34, 101] (Fig. 2). In fibroblasts, ET‐1 may induce transdiffferentiation towards myofibroblasts, promoting the release of extracellular matrix proteins such as types I and III collagen and fibronectin and the inhibition of metalloproteinase expression [102, 103]. Conversely, in human ECs, ET‐1 may promote increased expression of α‐SMA, as well as increased collagen production, thus modulating the EndoMT [4, 16, 17, 34].

Fig. 2.

Role of endothelin‐1 (ET‐1) and transforming growth factor (TGF)‐β during endothelial‐to‐mesenchymal transition (EndMT). Both TGF‐β and ET‐1, which are largely over‐expressed in systemic sclerosis (SSc), may promote the decrease of endothelial markers, such as CD31, von Willebrand factor (vWF) and vascular endothelial‐cadherin (VE‐cadherin) and the increase of mesenchymal markers, such as α‐smooth muscle actin (α‐SMA), smooth muscle 22 (Sm22) and collagen (Col1A1). ET‐1 may promote the TGF‐β mRNA transcription which, in turn, further promote ET‐1 mRNA, in a vicious loop. Both these molecules may modulate fibrogenic responses.

Wingless/integrated (Wnt) proteins

The Wnt proteins include a family of glycoproteins, playing a pivotal role in the pathogenic process occurring in fibrotic diseases, including SSc, via canonical and non‐canonical intracellular signalling pathways [104, 105, 106]. TGF‐β activates the canonical Wnt pathway and several genes which are transcriptional targets of Wnt/β‐catenin [107], leading to tissue fibrosis. It has been reported that, in human ECs, Wnt3a may modulate EndMT promoting the expression of cadherin and inhibiting the expression of vimentin [108].

Interleukin (IL)‐1

IL‐1 family consists of 11 pro‐ and anti‐inflammatory cytokines. Expression of the principal IL‐1 family cytokines, such as IL‐1α, IL‐1β, IL‐18 and IL‐33, is impaired in several autoimmune diseases, including SSc. Additionally, gene polymorphisms of IL‐1α, IL‐1β, IL‐18 and IL‐33 correlate with SSc susceptibility [109] The involvement of these cytokines during EndMT was first demonstrated via in‐vitro experiments, where human umbilical vein endothelial cells (HUVECs) were cultured with supernatants derived from activated peripheral blood mononuclear leucocytes. In this setting, morphological and phenotypical changes were observed in HUVECs, mirroring the EndMT process and confirming the important role of IL‐1β during this process [110]. IL‐1β acts synergistically with TGF‐β2 to promote the EndMT in HUVECs. ECs, stimulated with both IL‐1β and TGF‐β2, exhibited morphological and phenotypical changes and significantly increased Sm22 expression when compared to cells separately stimulated with IL‐1β or TGF‐β2 [91]. This observation confirmed that IL‐1β and other inflammatory cytokines may be potent stimuli for EndMT [111, 112]. Recently, it has been reported that IL‐1β promotes EndMT in HUVECs via the natural killer (NF)‐κB/Snail pathway and protein tyrosine phosphatase L1 (PTPL1), a non‐receptor‐type protein tyrosine phosphatase, implicated in several signal pathways. The inhibition of PTPL1/NF‐κB signalling may prevent the IL‐1β‐induced EndMT process in HUVECs [113]

Tumour necrosis factor (TNF)‐α

TNF‐α is a cytokine mainly released by macrophages, modulating immune response and driving the main cellular response [112]. It has been reported that serum concentrations of TNF‐α are increased in SSc patients and correlate with the onset of pulmonary fibrosis and PAH [114]. Additionally, over‐expression of TNF‐α has been shown to result in severe pulmonary hypertension and emphysema in mice, suggesting that TNF‐α plays a pivotal role in pulmonary vascular physiology, stimulating the progression of the endothelial dysfunction [115]. TNF‐α has been implicated in EC activation following inflammatory events, promoting the expression of adhesion molecules such as vascular cell adhesion molecule 1 (VCAM‐1) and intercellular adhesion molecule 1 (ICAM‐1) [19, 116]. Furthermore, TNF‐α has been implicated in EndMT activation in lymphatic ECs; EndMT‐related miRNAs were over‐expressed in rat mesenteric lymphatic ECs [117]. During the last year it has been reported that EC treatment with TNF‐α stimulates Smad2/3 signals, which are probably promoted by the increased expression of TGF‐β type I receptor, TGF‐β2, activin A and integrin α V. These data suggest that TNF‐α enhances TGF‐β‐induced EndMT by stimulating the TGF‐β signalling pathway [118].

Interferon (IFN)

Another possible mediator of EndMT is the IFN family. It has been reported that IFN‐γ induces the expression of TGF‐β2, ET‐1 and α‐SMA in human microvascular ECs. Furthermore, IFN‐γ may also induce the expression of several genes related to EndMT, including regulator of G protein signalling 2 (RGS2), fibronectin 1, plasminogen activator inhibitor 1 (PAI1), TWIST‐related protein 1 (TWIST1), signal transducer and activator of transcription 3 (STAT‐3), Snail‐1 and genes implicated in Wnt signalling [119]. Saigusa et al. showed that the IRF5 gene, a member of the IFN regulatory factor (IRF) family, may play a critical role in the bleomycin (BLM) SSc murine model, acting as a SSc susceptibility gene. Dermal and pulmonary fibrosis induced by BLM are inhibited in IRF5‐deficient (Irf5^−/−) mice. Furthermore, also the classical hallmarks of SSc, such as fibroblast activation, inflammation, EndMT, vascular regression, impaired T helper type 2 (Th2)/Th17 immune polarization and B cell stimulation, are inhibited in this model. On this basis, IRF5 may be involved in EndMT, and suppressing the IRF5 pathways may lead to the inhibition of SSc development [120].

MicroRNA (miR)

New advances for EndMT mediators are represented by miRs. These molecules are small non‐coding RNAs, which are post‐transcriptional repressors of genes [121, 122]. Previous studies have shown the involvement of several miRs (miR‐29s, miR‐125b and 126, miR‐130) in the development of EndMT [42] and the interaction between TGF‐β and several miRs (miR‐21, miR‐31, miR‐155) in modulating EndMT [42]. Of note, the increased expression of miR‐148b stimulates, in ECs, the ability to form vessels, migrate and proliferate, whereas its silencing increases the EndMT after TGF‐β stimulus. [123]. Additionally, the constitutive expression of miR‐31 induces TGF‐β‐induced EndMT in ECs [124] and the up‐regulation of miR‐130a, which is increased in the PAH experimental model, may promote the expression of α‐SMA, a key molecule in the EndMT process [125]. Considering the regulatory effects exerted by miRs on several signalling pathways, these molecules have been extensively studied as potential therapies for a large number of diseases, including cancers and cardiovascular diseases [126, 127, 128]. Concerning cancer diseases, intratumoral injections of miR drugs, such as miR mimics or repressors, may regulate the expression of specific genes decoding for molecules involved in the cancer development, thus enhancing target specificity, efficacy and minimizing side effects [129, 130, 131]. It has been shown that miR‐204 is inhibited by hypoxia in both rat pulmonary arterial intima and human pulmonary artery ECs. When miR‐204 is suppressed using a specific inhibitor the lack of this miR, in the context of hypoxia, leads to enhanced autophagy, a cellular homeostatic process that occurs both under basal and stressful conditions, with subsequent death of ECs and inhibition of hypoxia‐induced EndMT [132]. Recently, it has been reported that miR‐200c‐3p may play a critical role in EndMT. In a model of aortic grafting, miR‐200c‐3p expression is strongly increased, and grafted arteries undergo to neointimal hyperplasia via the EndMT process. Conversely, the inhibition of this miRin grafted arteries is associated with a reduction of neointimal formation [133]. Another important miR is MiR‐181b, which is down‐expressed in rat models of pulmonary arterial hypertension (PAH). Its over‐expression, in this model, attenuated the pulmonary vascular hypertrophy, right ventricular remodelling and the EndMT process. Additionally, in primary rat pulmonary arterial ECs, the induction of miR‐181b reversed the EndMT [134]. Also, miR‐92a may be a pivotal regulator in promoting EndMT during vein graft remodelling, Interestingly, in vivo, the inhibition of miR‐92a, by adeno‐associated virus‐mediated gene therapy, decreases EndMT, reducing vein graft neointimal formation and improving graft patency [135]. We may speculate that a clearer understanding of the role of miR during the EndMT observed in SSc and in other diseases may open new therapeutic advances, targeting key molecules of fibrotic disorders [121].

Reactive oxygen species (ROS)

Another mediator of EndMT may be ROS [25, 42]. It has been shown that ROS play an important role in controlling the TGF‐β profibrotic effects via Smad2/Smad3 activation [42, 136, 137, 138]. It has been demonstrated that [25] chronic oxidative stress and abnormal fibrillin‐1 expression may mediate EndMT in Tsk^−/+ mice and that reducing the oxidative stress, by administration of oxidized phospholipids scavengers, may decrease both endothelial apoptosis and mesenchymal transition [25]. Furthermore, in models of SSc‐like BLM‐induced fibrosis, this agent induces collagen (I and III) synthesis, mediated by ROS [139]. Anti‐oxidants such as N‐acetyl‐cysteine may attenuate collagen expression in experimental models of BLM‐induced lung fibrosis [140, 141]. Although the latter studies do not explore the effect on EndMT, the reduction of ROS levels is able to inhibit EndMT process, both in vitro and in the lung of rats with pulmonary hypertension [142].

Hypoxia inducible factor (HIF)

Hypoxia is another pivotal characteristic of SSc, playing a key role during angiogenesis. Furthermore, hypoxia is a potent promoter of EndMT, mainly in PAH [143] and in radiation‐induced fibrosis [144], via the HIF‐dependent pathways [145]. Furthermore, hypoxia may also regulate the expression of TGF‐β, the main promoter of EndMT [146]. EndMT keys transcription factors, such as Snail [147] and Twist‐1 [148], are also identified as targets of hypoxia. The role of hypoxia and HIF signalling are particularly relevant in SSc angiogenesis [149]. Recently it has been reported that autophagy may mediate both hypoxia‐induced fibroblast collagen synthesis and EndMT in SSc. Hypoxia exposure up‐regulates the expression of collagen I and connective tissue growth factor (CTGF) in SSc fibroblasts and the expression of EndMT markers in HUVECs. In the same conditions, fibroblasts and HUVECs form more autophagosomes and autolysosomes, suggesting the induction of autophagy [150].

Evidence of EndMT during human SSc

Following the first evidence of the lack of endothelial‐specific marker (VE‐cadherin) in SSc skin [151] our group was able to show that, after ET‐1 and TGF‐β stimulation, SSc‐ECs display a reduction of endothelial markers, including vWF, CD31 and VE‐cadherin, and an up‐regulation of profibrotic markers such as sSm22, α‐SMA and collagen, confirming the induction of the EndMT programme [34].

ECs lining injuries are the early events occurring in patients with SSc [152, 153]. In these patients, arteries are generally impaired and characterized by intimal hyperplasia, medial thickening, obliteration of the lumen and inflammation [58, 154], thus contributing to the vascular remodelling occurring during PAH [155]. It has been reported that EndMT may also play a pivotal role during the PAH pathogenesis, promoting dysfunction of the ECs [156]. Histological assessment of patients with SSc‐associated PAH identifies the presence vWF/α‐SMA‐positive ECs in up to 5% of pulmonary vessels [156]. Furthermore, the exposure of ECs, obtained from the pulmonary artery, to proinflammatory cytokines, including IL‐1β, TNF‐α and TGF‐β, changes their typical cobblestone phenotype and induces the expression of mesenchymal markers [156]. CD31⁺/CD102⁺ ECs isolated from SSc lungs expressed at the same time as mesenchymal/EC markers support the evidence of EndMT in lung tissues from SSc patients with ILD [157, 158]. Successively, EndMT is documented in SSc skin [22], showing the increased expression of mesenchymal markers in SSc‐ECs when compared with healthy controls (HC). Taken together, these data confirm the evidence of the EndMT programme in all the affected tissues and indicate the important pathogenic role of this mechanism during this disease.

Recently, in an experimental model of SSc, the double heterozygous mice for Klf5 and Fli1 (Klf5^+/−; Fli1^+/−), showing the three main pathological characteristics of SSc [159], the dermal microvascular ECs isolated from these mice, show defective angiogenesis and reduced expression of VE‐cadherin and CD31, confirming the evidence of EndMT in this SSc model [160].

As previously discussed, the exposition to several local biological mediators, such as TGF‐β, PDGF, VEGF and ET‐1 [42], activates the ECs and modulates the expression of adhesion molecules on the EC surface [161, 162], thus promoting the recruitment activated T and B lymphocytes and profibrotic macrophages. Following recruitment, inflammatory cells release profibrotic growth factors such as TGF‐β and CTGF in the tissue. Under the effect of these growth factors, ECs release ET‐1, activating EndMT [163, 164]. Furthermore, increased levels of the IL‐6 family, such as oncostatin M (OSM) and IL‐6, have been shown in many pathological diseases, characterized by inflammation, vasculopathy and fibrosis, including SSc [165]. Recently, it has been reported that, in dermal SSc‐ECs, the treatments with OSM or IL‐6⁺sIL‐6R stimulate proinflammatory genes, together with genes related to EndMT [166]. The serum levels of the leucotriene (LT) B4, an inflammatory lipid mediator, are increased in patients with diffuse cutaneous SSc and ILD and promote EndMT via the phosphatidylinositol 3‐kinase (PI3K)/protein kinase B (AKT)/mammalian target of the rapamycin (mTOR) pathway, independently of TGF‐β release [167].

Macrophages are both a player and a target in SSc [147]. These cells play an important role in the innate response [168, 169, 170, 171], although conflicting results are present in the available literature. Alternatively, activated macrophages (M2) prompt precursor transdiffferentiation towards myofibroblasts. These macrophages are present in the perivascular area of the tissues of patients with SSc, supporting fibrosis [172, 173, 174]. Moreover, some authors have shown that these macrophages may influence the function and molecular repertoire of ECs, inhibiting EndMT and preventing fibrosis, in a model of muscle repair after sterile injury [175]. Successively, their inhibiting role has been confirmed by lineage‐tracing transgenic mice models, in which endothelial‐derived cells (EdCs) transdiffferentiate towards mesenchymal cells after treatment with BLM. EdCs retrieved from the lung show a significant inhibition of endothelial markers and an increase of mesenchymal markers. This effect is counterbalanced by macrophages, which preserve the endothelial morphology of EdCs. Finally, the EndMT was activated after macrophage depletion [176].

Targeting EndMT

Due to the importance of EndMT in the pathogenic steps of many diseases, several drugs have been assessed in experimental models as potential EndMT inhibitors.

Conflicting results have been reported concerning rapamycin, an inhibitor of the mTOR signalling pathway, which plays a key role in TGF‐β‐mediated EndMT [159, 177, 178, 179]. It has been shown that this drug may prevent EndMT by suppressing the EC ability to migrate and to degrade the extracellular matrix [180]. It has been reported that in mice undergoing peritoneal dialysis the treatment with rapamycin reduces the peritoneal membrane thickness and EndMT process, suggesting that rapamycin has a protective effect on peritoneal membrane during peritoneal dialysis through an anti‐fibrotic and anti‐proliferative effect [181]. Moreover, it has been recently reported that, although rapamycin is able to suppress the senescence‐associated phenotype in human coronary artery ECs, treatment with rapamycin in this model may promote EndMT through the activation of autophagy [182].

Tanshinone IIA (Tan IIA), a phytochemical drug, plays an anti‐fibrotic effect, ameliorating skin thickness and collagen deposition, in the BLM‐treated SSc mouse model. Tan IIA contrasts the inhibition of angiogenesis promoted by BLM probably interfering with the induction of the Akt/mTOR/p70S6K pathway, which seems to be implicated in vascular damage of the BLM‐treated SSc mouse model [159, 177, 183, 184].

Geniposide, an iridoid glycoside, has been shown to inhibit the EndMT in the BLM‐induced scleroderma experimental model via the suppression of the mTOR signalling. On this basis, it is possible to speculate that geniposide may be a new potential therapeutic candidate to prevent vascular damage in SSc patients [185].

Imatinib, a tyrosine kinase inhibitor of the PDGF receptor, which is the treatment of chronic myeloid leukemia [186], has been shown to relieve EndMT in experimental models of PAH, induced by hypoxia, suggesting its potential therapeutic use [187, 188]. Recently, another tyrosine kinase inhibitor of PDGF and VEGF, nintedanib, is able to improve PAH by inhibiting EndMT. Nintedanib attenuates the expression of mesenchymal markers in human pulmonary microvascular ECs and the proliferation of pulmonary arterial smooth muscle cells, suggesting that this molecule might be another new therapeutic treatment for PAH, preventing vascular remodelling [189].

Among the EndMT mediators, TGF‐β is the principal molecule involved, and inhibiting TGF‐β may be a potential therapeutic approach. Spironolactone, an aldosterone receptor‐blocker, significantly inhibits EndMT induced via TGF‐β, down‐regulating vimentin, up‐regulating CD31 in HUVECs as well as inhibiting cell migration during EndMT, via the block of TGF‐β and Notch signalling [190]. Another TGF‐β inhibitor, the dipeptidyl peptidase‐4 (DPP‐4) inhibitor linagliptin, may block TGF‐β2‐induced EndMT, interfering with TGF‐β/integrin‐β1 interaction [191]. Arginylglycylaspartic acid (RGD) is an Arg‐Gly‐Asp tripeptide motif, implicated in cell adhesion to the extracellular matrix. It has been reported that the RGD antagonist may revert TGF‐β1‐induced EndMT and consequently may be potentially used as an anti‐fibrotic therapeutic approach [192]. Glycyrrhizin, clinically used for chronic hepatic diseases and itching dermatitis, improves dermal fibrosis and EndMT in BLM‐treated mice, interfering with TGF‐β signalling in dermal fibroblasts via thrombospondin‐1 down‐regulation [193]. Another TGF‐β signalling inhibitor is AdipoRon, a novel orally active small molecule selective for adiponectin receptors. It has been shown that adiponectin exerts its protective role in preclinical models of cardiac, pulmonary and hepatic fibrosis, interfering with TGF‐β signalling. It has been shown that chronic administration of AdipoRon significantly improves BLM‐induced dermal fibrosis in mice, attenuating fibroblast proliferation, adipocyte‐to‐myofbroblast transdifferentiation, Th2/Th17 polarization, vascular regression and EndMT within the affected skin. These results suggest that AdipoRon plays a protective role in microvascular damage induced by the treatment with BLM, preventing EndMT and vessel regression [194]. Additionally, treatment with hepatocyte growth factor (HGF) significantly prevents the occurrence of TGF‐β1‐induced EndMT in HUVECs via blocking Notch signalling [37, 195, 196].

Tamibarotene (Am80) is a synthetic retinoid, able to control the pathological events of several autoimmune and inflammatory diseases. It has been shown that this molecule significantly attenuates dermal and hypodermal fibrosis as well as EndMT, ICAM‐1 expression in ECs, infiltration of macrophages, mast cells and lymphocytes and M2 macrophage differentiation in BLM‐treated mice [197].

Interestingly, in previous work [52] we reported that macitentan, an ET‐1 receptor antagonist in vitro, inhibits both ET‐1‐ and TGF‐β‐induced EndMT in microvascular ECs isolated from HC and SSc patients [34, 52]. This effect is mediated by the TGF‐β receptors/ET receptor heterodimer complex expressed on the cell surface, which may be activated by both ET‐1 on the TGF‐β, thus leading to the profibrotic programme and probably inhibited by the ET‐1 antagonist [52]. Successively, these results were confirmed by Corallo et al. [198] using both macitentan and bosentan, another ET‐1 receptor antagonist, in the fibroblast and ECs co‐culture model [198]. These findings suggest that EndMT may be a reversible mechanism, opening a new therapeutic perspective for fibrotic diseases [199, 200, 201].

A decreased expression of the junctions between ECs seems to be a pivotal aspect of EndMT. In blood vessels, ECs form a lining that separates blood from the surrounding tissue. During EndMT ECs lose their cell adhesion, lose their ability to form a barrier and detach from the endothelial monolayer [202]. It has been reported that the adherens junctions, such VE‐cadherin and β‐catenin, are disconnected from the cell membranes, contributing to vasculopathy and EndMT. Interestingly, iloprost, a synthetic analogue of prostacyclin (PGI2), may promote VE‐cadherin clustering at adherens junctions, preventing their loss from the ECs membrane [203, 204]. Recently, Tsou et al. showed that iloprost promotes the stabilization of adherens junctions, promotes angiogenesis and prevents EndMT, all these activities having important therapeutic benefits in SSc [205].

Recently, it has been reported that paeoniflori (PF), a monoterpene glycoside with endothelial protection, vasodilation, anti‐fibrotic, anti‐inflammatory and anti‐oxidative properties, significantly prevents chronic hypoxia/SU5416‐induced PAH in rats and its therapeutic effect seems to be related to its ability to inhibit EndMT in pulmonary arterial ECs [206].

Concluding remarks and future perspectives

During SSc, dysregulation of EC functions plays a pivotal role during vascular remodelling, acting as trigger for fibroproliferative disorder [2, 207, 208]. The transdiffferentiation of injured ECs towards a myofibroblastic phenotype plays an important role in SSc pathogenesis, this process being the link between vascular activation and fibrotic process [4, 34, 52]. A clearer understanding of the processes responsible for EndMT is of primary importance to recognize clinically useful biomarkers in order to predict fibrotic remodelling, and/or to develop effective anti‐fibrotic therapies in different fibrotic conditions. To date, many EndMT inhibitors are available, and exploring their efficacy in SSc may be an important challenge, considering that, to date, no disease‐modifying agents are still available.

Disclosures

The authors declare no conflicts of interest.

Author contributions

All authors made substantial contributions to the conception or design of the work, the acquisition and interpretation of literature. All authors contributed to the critical review and revision of the manuscript and approved the final version. All authors agreed to be accountable for all aspects of the work. P. D. B. and P. R.: study conception and design, literature search, figure creation, writing, paper revision and acceptance; O. B., M. V., L. N. and V. D.: literature search, writing, paper revision and acceptance; P. C., R. G.: study conception and design, writing, paper revision and acceptance.

Data Availability Statement

Data sharing is not applicable to this article, as no new data were created or analyzed in this study.