Impaired Cav-1 expression in SSc mesenchymal cells upregulates VEGF signaling: a link between vascular involvement and fibrosis

Systemic sclerosis (SSc) is a chronic autoimmune connective tissue disease affecting the skin and a variety of internal organs. The hallmark of the early stages of SSc is endothelial involvement, with a perivascular inflammatory infiltrates and a decreased capillary density[1], whereas later stages are characterized by an excessive accumulation of extracellular matrix (ECM), resulting in fibrosis[2, 3]. The vascular alterations might be considered an important trigger for fibrosis, because of endothelial cells and pericytes, after injuries, might trans-differentiate toward myofibroblast, a cell producing increased amounts of collagen types I, III, VI, and VII, fibronectin, and glycosaminoglycans[4].

Originally, myofibroblasts were thought to derive from resident fibroblasts which, after the interaction with effector molecules in injured tissues, might be activated, thus leading to reparative fibrotic response. More recently, it has been suggested that these cells may be generated by both epithelial-mesenchymal and endothelial-mesenchymal transformation[5‐7]. In the last years, the use of lineage tracing, helped us to better define the myofibroblasts origin and evolution, confirming that pericytes and resident fibroblasts are the major, if not the only, source of myofibroblasts in at least one animal model of kidney fibrotic diseases[8‐10]. On these bases, it might be important to assess if the molecular mechanisms involved in the differentiation of pericytes toward myofibroblasts may play a role in the pathological process leading to fibrosis during SSc.

Recent works showed that perivascular cells share surface markers and differentiative ability with bone marrow mesenchymal stem cells (MSCs), and MSCs express pericyte markers and cooperate with endothelial cells to form a vascular network, supporting the concept that pericytes are members of the adult multipotent MSCs family[11‐15]. Because of these similarities and to overcome the difficulties to isolate primary pericytes, as reported by several authors[16, 17], in our work we used MSCs as pericytes. Furthermore, we already reported that SSc-MSCs have been shown to be an alternative system for studying the contribution of pericytes in the pathogenesis of SSc[18].

Vascular endothelial growth factor A (VEGF) is a key growth factor, involved in the reparative angiogenesis, after injuries[19, 20]. In addition to this well known effect, this molecule may induce perivascular cells to produce connective tissue growth factor (CTGF), which stimulates extracellular matrix production and fibrosis[21]. In fact, in presence of CTGF, pericytes differentiate toward a migratory and myofibroblast phenotype[22]. VEGF has been suggested to be involved in the pathogenesis of SSc and its expression is markedly increased in different cell types, both in the epidermis and dermis of patients with SSc and in the bloodstream, correlating with organ manifestations[23, 24]. At present, many different isoforms of VEGF family are known, such as VEGF121, VEGF189, and VEGF165, and recently, a new VEGF165 isoform (VEGF165b), generated through alternative splicing, with a possible antiangiogenic effect, has been described[25]. Although a switch from proangiogenic to antiangiogenic isoforms has been recently described in SSc patients[26], available literature showed conflicting results about the potential inhibitory role of this isoform[25, 27]. VEGF exerts its functions by binding to the tyrosine kinase receptors VEGF receptor 1 (VEGFR1) and VEGF receptor 2 (VEGFR2). The tridimensional VEGFR organization, on cell surface, modulates the transmission of its specific signal. It is well known that the cell surface proteins trafficking is regulated by the lipid rafts microdomains which are present in the plasma membranes. Caveolae are special structures of lipid raft, which may be detected by electron microscopy as flask shaped invaginations in the cell surface membrane[28, 29]. These structures are enriched in cholesterol and may be distinguished from other structures of the lipid raft by the presence of a specific molecular family: the caveolin scaffolding proteins, which are essential for caveolae formation and cholesterol binding[30‐32]. Caveolin-1 (Cav-1) is a member of this family and it is involved in the regulation of signaling activity, via a direct interaction of its scaffolding domain with a consensus sequence present in several signaling proteins, including VEGFR2[33]. Of note, the interaction of Cav-1 with the signal transducing proteins, has been shown to inhibit the activity of the latters. On these bases, Cav-1 may play an important role in VEGF-induced signaling, promoting VEGFR2 degradation by caveolin-mediated transport to proteasome, under specific conditions.

Conflicting results have been published on available literature concerning Cav-1 in SSc. Lower levels of Cav-1 were found in SSc lung fibroblasts and linked to constitutive activation of JNK, ERK, and Akt signaling, leading to overexpression of profibrotic markers such as collagen and alpha smooth muscle actin[34]. Del Galdo et al.[35] showed, in dermal fibroblasts, a downregulation of Cav-1 in SSc skin and suggested that this finding may contribute to the increased collagen deposition via the activation of canonical transforming growth factor (TGF-b) pathway. Haines et al.[36] showed that Cav-1 is a positive regulator of both TGF-b and CTGF genes expression and signaling in human dermal fibroblasts and they found increased levels of Cav-1 in SSc fibroblasts.

The goal of this work was to evaluate whether an alterations in Cav-1 expression, affecting perivascular cells, may contribute to the pathogenic events leading to fibrosis during SSc, via the persistence of the VEGF signaling in the these cells. In this work, we provide evidence that a downregulation of Cav-1 in MSCs, isolated from SSc patients, strongly supports the pro-fibrotic effect of VEGF165, whose role in SSc has been largely explored[23‐26]. In fact, an impairment of Cav-1 mediated internalization and degradation of VEGFR2 leads to a vicious loop, with redundant effects of VEGF signaling and consequent increased CTGF production. Furthermore, to better clarify the mechanisms linking the loss of Cav-1 to the abnormal persistent VEGF signaling in SSc, we silenced Cav-1 expression in HC-MSCs by specific siRNA, and observed that the decreased Cav-1 expression in HC-MSC is associated to a persisting VEGF signaling and CTGF production, mirroring the results observed in SSc-MSCs. These data confirm the lack of Cav-1 in perivascular cells may lead to an aberrant production of CTGF, which may contribute to the progression of fibrosis in SSc patients.

Fibrosis, characterized by excessive extracellular matrix accumulation, is the hallmark of SSc. Experimental studies suggest that a complex network including endothelial cells, pericytes, fibroblasts, and immune cells may lead to the pathogenic events of the disease. Many cytokines such as TGF-b, VEGF, CTGF, and endothelin-1 are considered as key mediators in this process, because of their skill to induce the differentiation of fibroblasts toward a myofibroblastoid phenotype[21, 37].

A better understanding of the mechanisms that form an initial vascular alteration leading to a fibrotic response might suggest new potential targets for anti-fibrotic therapies of SSc. In this light, recent studies focused on the differentiantive ability of perivascular cells toward myofibroblast in fibrotic disorders. Dulauroy et al.[38], showed that the perivascular cells derived from mesenchymal lineage differentiated toward myofibroblast in an experimental model of kidney fibrosis. Furthermore, pericytes were shown to be the major source of myofibroblasts, thus opening new perspective about the role of perivascular cells in fibrosis.

In this work, using MSCs isolated from bone marrow of SSc patients, which are largely accepted as an alternative source of functional pericytes[11‐15, 18], we provide evidence that these cells display a failure in Cav-1 trafficking inside the cell, which lead to a significant decrease of VEGFR2 proteosomal degradation, upregulation of the VEGF downstream pathways, and finally, a consequent increased expression of CTGF. This abnormal persistent VEGF signaling in SSc perivascular mesenchymal cells, associated with an increased expression of CTGF, which is a pivotal cytokine involved in the development of tissue fibrosis, suggests that pericytes might be considered a new potential therapeutic target in SSc.

The VEGF signaling activation involves its binding to VEGFR2 and consequent internalization into the cytoplasm. The endocytosis of the receptor tyrosine kinases (RTKs) family may be divided into two different steps: the first step is internalization by translocation from the cell surface into intracellular vesicles. Internalization mainly controls the number of cell surface receptors and the sensitivity of the cell to incoming signal. In the second step, internalized receptors are sorted through a complex system of intracellular vesicle compartments, known as endosomes[39‐41]. From early endosomes, which are characterized morphologically by small size and proximity to the plasma membrane[42], incoming ligand-receptor complexes are sorted into two different fates: recycling or degradation compartments. At present, it is well known that a dysfunctional regulation of RTKs endocytosis may promote uncontrolled activation of signaling[43].

Two different internalization pathways have been postulated for VEGFR2: the Caveolae-mediated endocytosis and the Clathrin-mediated internalization. The central question of whether or not VEGFR2 uses caveolar route for internalization cannot be unambiguously answered based on available data[44].

Caveolae are plasmalemmal caveolin-coated invaginations with trafficking functions, whose activity depends by the main protein Cav-1. Cav-1 oligomers constitute the filaments covering the cytosolic surface of caveolae[45]. These oligomers form a scaffold for the assembly of many signaling molecules, including receptors, their ligands, and signal transducers, thus regulating the activation and degradation state of these complexes[46].

Conflicting results have been published in available literature, concerning the possibility that VEGF may modulate the Cav-1 expression, probably due to the different cell types (endothelial, mesangial, fetoplacental artery endothelial cells) used in the experiments[47‐49]. Furthermore, no data are available about the VEGFR2/Cav-1 interaction and its functions in MSCs. Normally, in the unstimulated state, caveolae are static structures, but after surface receptor stimulation, caveolae form active intracellular vesicles. Phosphorylation of Cav-1 (Y14) by Src kinase are required for caveolar fission[44]. The localization of both receptors and signaling partners to caveolae, has been considered a pivotal mechanism to control the levels of both receptors and signaling proteins, their availability, and activation.

In our work, we showed that in MSCs VEGF upregulates the levels of Cav-1 and its phosphorylation, suggesting that this mechanism may be involved in the following VEGFR2 activation. Conflicting results have been recently published concerning the optimal methods to evaluate the caveolae-associated Cav-1[47, 49, 50]. Thus, in this paper, we lysed the cells by using neutral detergent, which completely dissolved the Cav-1 associated lipid micro-domains, or alternatively SDS, which did not detach the Cav-1 molecule from the caveolae. Independently of the constituents of buffer in SSc-MSCs stimulated and unstimulated by VEGF, the Cav-1 levels and its phosphorylation were always significantly lower when compared to HC-MSCs. Moreover, in untreated HC-MSCs, Cav-1 was primarily localized at the plasma membrane, and, after VEGF treatment, Cav-1 was internalized into the cytoplasm around the perinuclear area. On the contrary, this vesicular trafficking was significantly impaired in SSc-MSCs.

To assess the pathogenetic implication of the reduced Cav-1 expression in the specific trafficking and signaling of VEGFR2 during SSc, we evaluated the Cav-1/VEGFR2 co-localization, after VEGF stimulus. Our data showed that in SSc-MSCs this molecular co-localization was significantly impaired, when compared to HC-MSC and these data were confirmed by IF and immunoprecipitation.

The signaling of VEGF starts after VEGFR2 internalization and the fate of activated receptors depends on its transport to late endosome for degradation, or alternatively for recycling. Several studies, carried out on different RTKs suggest that the transport to late endosome should be mediated by Cav-1[51, 52] and a decrease of Cav1 may result in enhanced growth factor signaling[51]. The receptor degradation occurs in the proteasome. It recognizes and rapidly degrades ubiquitinated proteins involved in the regulation of transcription, cell cycle progression, apoptosis, signal transduction, and angiogenesis[53, 54]. Ubiquitin is a conserved 76-amino-acid polypeptide, activated by an ubiquitin-activating enzyme (E1) in an ATP-dependent manner and it is transferred to an ubiquitin-conjugating enzyme (E2). Furthermore, an ubiquitin-protein ligase (E3) specifically attaches ubiquitin to a target protein by an amino group of a lysine residue. E3 ubiquitin ligases are a large family of proteins that are known to be involved in the regulation of turnover and activity of many target proteins. Although it has been established that VEGFR2 is ubiquitinated in a ligand-dependent manner in response to VEGF[53], the role of ubiquitination in VEGF signaling is still unclear. Several ubiquitin ligases are known to be involved in VEGFR2 degradation and attenuation of signaling[55]. Earlier studies established that an E3 ubiquitin ligase, Nedd4, modulates the proteasomal degradation of VEGFR2, and that this degradation might be inhibited by the sequestration of Nedd4[54]. It may be considered that the amount of ubiquinated VEGFR2, co-localizing with Cav-1[33, 55‐57] may be carried to degradation, following the same fate of other RTKs[44]. We observed that the levels of ubiquitinated VEGFR2, after VEGF treatment, were significantly reduced in SSc-MSC and this decreased level of ubiquitination might contribute to upregulation of VEGF signaling. In fact, different studies showed that ubiquitin-mediated proteolysis plays an important role in the regulation of many cellular processes, including receptors activity, by facilitating the timely destruction of key proteins[58‐60]. It is well known that VEGF may induce CTGF production[59], and in this setting we observed that following the increased VEGF signaling in SSc-MSCs, a consequent increased expression of CTGF may be shown. This molecule, also known as CCN2, is a member of the CCN (CCN1-6) family of matricellular proteins. Many studies focused on the role of CTGF in increasing ECM production, during fibrosis[61]. CTGF is generally overexpressed in all fibrotic conditions, and induces collagen type I deposition[62] and it has been suggested to play a crucial role in SSc tissue fibrosis[61‐64]. It has been shown to induce fibroblast proliferation, cell adhesion, and stimulation of extracellular matrix production[65]. it has been reported that VEGF is able to modulate CTGF production[21], in fact, VEGF may affect both the proliferation of fibrocellular components and the wound healing process, via CTGF induction[19, 20]. Interestingly, the SSc-MSCs production of CTGF was significantly higher when compared with healthy cells and the increased levels of VEGF observed in SSc patients[23] may contribute to the over expression of CTGF, which plays different physiological roles, not only in fibrotic process but also in angiogenesis, such as the maintenance of capillary strength via the extracellular matrix production[66, 67].

To confirm the functional role of loss of Cav-1 in VEGF signaling and CTGF modulation, we silenced the Cav-1 expression in HC-MSCs. We showed that the downregulation of Cav-1 induced a significant increase in VEGFR2 phosphorylation and its interaction with PI3-kinase p85 subunit, which is associated to an increase of VEGF signaling, in knockdown cells, when compared to NT HC-MSCs, thus mirroring the same results observed in SSc-MSCs.

This Cav-1 downregulation further strengthens our data in SSc patients, and supports the hypothesis that the loss of Cav-1 in SSc-MSCs may be involved in the molecular pathogenetic steps of the disease, and linking the earlier vascular damages to the subsequent fibrosis.

This is the first paper, in our knowledge evaluating the function of Cav-1 on the VEGF signaling in SSc, and particularly on MSC/pericyte lineage[1], showing that these pathways may strongly influence fibrotic process during scleroderma. These data underline the potential role of Cav-1 in fibrotic diseases and suggest this molecule as a possible therapeutic target in the disease. Furthermore, the evidence of an impaired expression and function of Cav-1 in perivascular SSc cells strongly confirms the potential role of the MSCs/pericytes in this pathologic process probably in the earlier phases of the disease following the initial endothelial injury. Taken together our data suggest that MSCs/pericytes lineage, localized in the perivascular niches, may be a potential cellular target in SSc and open new perspectives in the field of MSC transplantation and regenerative medicine in scleroderma.