Abstract
Systemic sclerosis (SSc) is a complex disease characterized by early microvascular abnormalities, immune dysregulation and chronic inflammation, and subsequent fibrosis of the skin and internal organs. Excessive fibrosis, distinguishing hallmark of SSc, is the end result of a complex series of interlinked vascular injury and immune activation, and represents a maladaptive repair process. Activated vascular, epithelial, and immune cells generate pro-fibrotic cytokines, chemokines, growth factors, lipid mediators, autoantibodies, and reactive oxygen species. These paracrine and autocrine cues in turn induce activation, differentiation, and survival of mesenchymal cells, ensuing tissue fibrosis through increased collagen synthesis, matrix deposition, tissue rigidity and remodeling, and vascular rarefaction. This review features recent insights of the pathogenic process of SSc, highlighting three major characteristics of SSc, microvasculopathy, excessive fibrosis, and immune dysregulation, and sheds new light on the understanding of molecular and cellular mechanisms contributing to the pathogenesis of SSc and providing novel avenues for targeted therapies.
Introduction
Systemic sclerosis (SSc) or scleroderma is a complex autoimmune disease characterized by widespread vasculopathy, multi-systemic excessive fibrosis, and autoantibody production. This combination of seemingly disparate features differentiates SSc from any other diseases. Intractable progression of vascular and fibrotic organ damage accounts for the chronic morbidity and high mortality. Multiple genes contribute to disease susceptibility, but environmental exposures are likely to play a major role in causing and progressing the disease. Pathogenesis is dominated by vascular changes; evidence of autoimmunity with distinct autoantibodies and activation of both innate and adaptive immunity as well as sustained low-grade inflammation; and fibrosis of the skin and visceral organs that results in irreversible scarring and organ failure. In other word, fibrosis is the end-result of a cascade of vascular and immune responses to environmental injury, and represents a failed tissue repair process. There have been advances in understanding of pathogenic processes that reflect the interplay between immune-inflammatory processes and vasculopathy and fibrosis. This review focuses on recent insights of functional consequences of SSc, highlighting three major characteristics of SSc; microvasculopathy, excessive fibrosis, and immune dysregulation.
Microvasculopathy
Although, widespread structural changes of the microvasculature are well documented in SSc patients, many aspects of SSc vasculopathy, including the nature of the injury and the fate of injured endothelial cells (ECs), remain poorly understood. Nevertheless, studies in recent years have brought a wealth of new information on the molecular and cellular mechanisms contributing to SSc vasculopathy. Diverse triggers have been implicated in inducing EC damage, including infection, immune-mediated cytotoxicity, anti-endothelial autoantibodies (AECAs), and ischemia-reperfusion injury. The resulting pathogenic changes in ECs may also vary and may include EC apoptosis, endothelial to mesenchymal transition (EndoMT), and other forms of EC activation. A bi-directional cross-talk between ECs and fibroblasts has been described in SSc (Fig. 1). These variable manifestations may depend on the nature of the trigger and/or the patient's ability to respond to those triggers. While this review article focused on ECs, other vascular cells, such as endothelial progenitor cells, pericytes/smooth muscle cells, adventitial fibroblasts also contribute to the pathogenic changes in SSc microvasculature. Several comprehensive reviews on SSc vasculopathy have recently been published and the readers are referred to those articles for additional information (1–2–3–4–5–6).
Relationship between endothelial cell (EC) injury and fibrosis in systemic sclerosis (SSc). Injured ECs contribute to the development of fibrosis through several mechanisms, including apoptosis, EndoMT and activation and resultant secretion of pro-fibrotic mediators capable of recruitment and activation of dermal fibroblasts toward the myofibroblast phenotype. EndoMT = endothelial mesenchymal transition.
EC apoptosis
SSc microvasculature is characterized by progressive structural damage of capillaries leading to capillary rarefaction (7–8–9). EC apoptosis has been proposed to contribute to capillary dropout; however, only a single published report directly demonstrated endothelial cell death in SSc skin and the findings have not been reproduced by other investigators (10, 11). The main support for this concept comes from the animal models of SSc and from the in vitro studies. In a pioneering study using UCD-200/206 chicken, Sgonc and colleagues found microvascular damage with EC apoptosis and subsequent infiltration of lymphocytes and tissue fibrosis that closely resembled human SSc (11). Similar observations were made in another model of SSc, the Fra-2 transgenic mice (12). The Fra-2 transgenic mice showed a marked EC apoptosis, followed by a progressive loss of small capillaries and tissue fibrosis in the skin. Dermal EC apoptosis and other histopathological features of SSc, including loss of micro-vessels and increased dermal fibrosis, have also been reported in a different genetic model of SSc, the urokinase-type plasminogen activator receptor (uPAR) knockout mice (13).
The loss of capillaries in SSc prompted the search for the circulating cytotoxic factors in SSc. In vitro studies have shown that SSc sera have a capacity to induce programed cell death in ECs. Kahaleh et al have attributed this cytotoxicity to a protease-like factor, later identified as granzyme-A (14). Several groups have reported the presence of AECAs, which were found in a subset of SSc patients and were capable of inducing EC apoptosis (15–16–17). In addition, AECAs were also shown to induce cytokines, chemokines and adhesion molecules on ECs, thus contributing to the overall inflammatory state of the endothelium (17). In SSc patients, the AECAs are more prevalent in patients with pulmonary arterial hypertension (PAH) (18, 19).
Importantly, immunosuppressive therapies have shown some promise in reversing capillary rarefaction in SSc patients. Fleming et al have reported that high-dose immunosuppressive therapy and autologous stem-cell transplant increased capillary density in five of the seven subjects (10). However, these findings were not reproduced in a different group of patients undergoing similar therapy (20).
EndoMT
EndoMT is a process whereby ECs undergo molecular changes that result in a gradual loss of EC characteristics and emergence of mesenchymal cell markers. A number of studies using various experimental models of fibrosis have implicated EndoMT as a potential source of fibroblasts/myofibroblasts (21). Importantly, recent studies provided evidence for EndoMT in human pathologies, including SSc. For example, Good et al (22) demonstrated the presence of cells co-expressing von Willebrand factor (vWF) and α-smooth muscle actin (α-SMA) in pulmonary vessels of SSc patients with PAH. Similar observations were also made in patients with idiopathic PAH (23). EndoMT may also contribute to the fibrotic process in SSc-associated pulmonary fibrosis (24). Mendoza and colleagues (24) showed co-localization of endothelial and mesenchymal cell markers in small pulmonary vessels in SSc patients, suggesting that these cells may be responsible for the production and deposition of extracellular matrix in the luminal space leading to vessel obliteration. This conclusion was corroborated by the comparison of gene expression in immunopurified SSc and healthy control lung ECs. SSc ECs expressed markedly elevated expression of collagen, fibronectin, connective tissue growth factor (CTGF), transforming growth factor-β (TGF-β), and other profibrotic genes. Expression of the EndoMT markers, SNAIL2 and TWIST1, was also increased (24).
Further evidence supporting the process of EndoMT in SSc was provided by a recent report by Manetti and colleagues (13). Consistent with SSc lung studies, co-expression of EC markers CD31 or VE-cadherin and α-SMA was frequently observed in SSc dermal capillaries and arterioles (13). Further in vitro comparison of SSc and healthy control microvascular ECs showed higher expression of mesenchymal markers, such as collagen, α-SMA, and S100A4/FSP1 and lower expression of CD31 and VE-cadherin in SSc ECs. Interestingly, sera from SSc patients, but not healthy controls, induced EndoMT in control ECs with potency similar to that of TGF-β, suggesting that a factor(s) in SSc sera is contributing to the EndoMT in vivo. The potential candidates may include angiotensin II (Ang II) and endothelin (ET)-1 (25); however, further investigations may reveal additional EndoMT stimuli in SSc serum. Importantly, the presence of EndoMT has also been reported in several other fibrotic disorders, including cardiac and renal fibrosis further underscoring the contribution of EndoMT to the fibrogenic process in different organs (21).
The studies in cultured endothelial cells have begun to characterize signaling pathways governing EndoMT and the nature of the factors that induce this process. TGF-β is considered the primary inducer of EndoMT, but the involvement of other agonists has also been reported (Tab. I) (21, 26–27–28–29–30–31–32–33). For example, during cardiac infarction, activation of the Wnt/β-catenin pathway was shown to induce transition of ECs to the α-SMA-positive cells (26). The Wnt/β-catenin pathway was also implicated in inducing EndoMT by the complement fragments C3a/C5a in a model of diabetic kidney disease (27). Elevated levels of nuclear β-catenin have been shown in SSc skin tissues, but the involvement of the Wnt/β-catenin pathway in SSc vasculopathy has not yet been formally investigated (34). Ang II has also been implicated in SSc vascular complications. SSc patients, in the early stage of the disease, have elevated serum levels of Ang II (35). Moreover, functional, pathogenic antibodies directed against the angiotensin II type 1 receptor (AT1R) and endothelin type A receptor (ETAR) were identified in patients with SSc and linked to increased prevalence of SSc-related vascular and fibrotic complications and higher risk for SSc-related mortality (25, 36). Ang II was shown to induce EndoMT in the dermis of experimental mice and in cultured human microvascular ECs (28).
Inducer of EndoMT potentially implicated in SSc pathogenesis
TGF-β was required for EndoMT in embryonic, but not in adult valve ECs.
EndoMT = endothelial mesenchymal transition; TGF-β = transforming growth factor-β; IL = interleukin; TNF = tumor necrosis factor; IFN = interferon; TLR = toll-like receptor; ECs = endothelial cells.
A number of studies drew attention to the inflammatory cytokines as inducers of EndoMT. For example, a short-term exposure of valve endothelial cells to interleukin (IL)-1β or tumor necrosis factor (TNF)-α induced transient EndoMT, while a longer exposure induced permanent transformation to myofibroblasts (37). Similarly, IL-1β greatly potentiated TGF-β-induced EndoMT in human umbilical vein endothelial cells (HUVECs) (29). Furthermore, in cardiac valve endothelial cells TNF-α or IL-6-induced EndoMT was mediated through activation of the Akt/NFκB pathway (30). However, IL-6 was not effective in inducing EndoMT in human dermal microvascular ECs, while a related cytokine, oncostatin M, was a potent inducer of EndoMT in those cells (Stawski and Trojanowska, unpublished observations), suggesting that ECs from the different vascular beds may differ in responses to inflammatory cytokines. Echeverría and colleagues investigated a role of toll-like receptor (TLR)-4 agonist, endotoxin, in induction of EndoMT in human umbilical venous ECs (31). Their studies showed that endotoxin-induced EndoMT was mediated through the endogenous TGF-β/ALK5 and Ca2+-permeable channel TRPM7 signaling. In addition, reactive oxygen species (ROS) generated by endotoxin contributed to the process of EndoMT and subsequent vascular fibrosis. Endogenous TGF-β has also mediated interferon (IFN)-γ induced EndoMT in human microvascular ECs (32).
Interestingly, EndoMT occurred spontaneously in mice with caveolin-1 deficiency (38). Caveolin-1 is involved in the TGF-β receptor internalization and signaling, thus further supporting the key role of TGF-β in the process of EndoMT. TGF-β promotes EndoMT through a complex mechanism that involves both Smad-dependent and Smad-independent pathways (39). Notably, Li and Jimenez have shown that c-Abl and PKCδ mediate EndoMT in murine pulmonary ECs (40). This profibrotic signaling pathway is known to down-regulate transcription factor Fli1 in dermal fibroblasts and ECs (41, 42). Accordingly, Fli1 deficiency is sufficient to induce EndoMT in human microvascular ECs by directly down-regulating VE-cadherin (43) and up-regulating FSP1, and Snail1 (44). Mice with Fli1 haploinsufficiency spontaneously develop EndoMT in the dermis (44).
Given the potentially important role of EndoMT in SSc vasculopathy, therapies that could mitigate this process would be beneficial in improving vascular function in SSc patients. Indeed, a dual endothelin receptor antagonist, bosentan, has been shown to prevent the development of new digital ulcers without having a beneficial effect on pre-existing digital ulcers in patients with SSc (45, 46). Furthermore, bosentan was effective in reversing severe loss of capillaries and improving dermal microvascular stability in SSc-PAH patients (47). Consistent with these findings, bosentan prevented EndoMT in patient-derived microvascular endothelial cells (48). Mechanistic studies by Akamata et al have shown that bosentan reversed ET-1-mediated activation of the c-Abl/PCKδ signaling pathway and normalized Fli1 protein levels in cultured ECs (41). Bosentan also normalized Fli1 protein levels and improved vascular fragility in vivo, in Fli1-conditional knockout (CKO) mice (41). In another study, Yamashita et al showed that glycyrrhizin, a saponin used in Japan to treat chronic hepatic disease, had similar beneficial effects on dermal vasculature. Glycyrrhizin ameliorated EndoMT after bleomycin injections and normalized vascular fragility in Fli1-CKO mice (49). Given the promising effects of these drugs in reversing EndoMT in experimental models of SSc vasculopathy, it would be of interest to determine whether these findings would be applicable to patients with SSc.
Molecular characteristics of microvascular ECs in SSc
Molecular underpinnings of SSc vasculopathy were largely unexplored until the methodology to isolate and propagate ECs from SSc skin biopsies was introduced by del Rosso and colleagues. It then became possible to begin systematic studies of the molecular characteristics of these cells and define the signaling pathways contributing to defective angiogenesis. Comparisons of transcriptomes of ECs isolated from SSc and control subjects revealed a large number of genes that were differentially expressed in SSc ECs, suggesting dysregulation of multiple pathways (50). Among such defective pathways, dysregulation of the urokinase-type plasminogen activator (uPA)/uPAR system was prominent. It was shown that overproduction of matrix metalloproteinase (MMP)-12 by SSc ECs caused the cleavage of uPAR resulting in truncated receptor unable to bind its ligand uPA and to properly interact with integrins (51). The follow-up studies showed that truncated uPAR was uncoupled from the αMβ2 and αXβ2 integrins in SSc ECs resulting in impaired cell motility (52). Other dysregulated genes in SSc skin ECs included the members of the kallikrein family, KLK9, KLK11, and KLK12. Decreased protein levels of the kallikreins were observed in SSc skin biopsies (50). Moreover, functional in vitro studies have suggested that reduced levels of KLK 9, 11, and 12 could contribute to reduced angiogenesis in SSc (53).
There is strong evidence supporting dysregulation of one of the central proangiogenic pathways, vascular endothelial growth factor (VEGF)-A and its receptor VEGFR-2, in SSc patients (54). It has been shown that in SSc patients, the ratio between pro- and anti-angiogenic splice variants of VEGF-A is tilted towards the anti-angiogenic variant VEGF165b. In addition, the angiogenic ability, including response to VEGF is impaired in SSc dermal ECs despite elevated expression of VEGFR-2. This deficiency was addressed in a recent study by Romano et al (55). They found that neuropilin-1, which functions as a co-receptor for VEGFR-2, was significantly decreased in SSc ECs. Interestingly, neuropilin-1 is transcriptionally controlled by Fli1, and its down-regulation was owed, at least in part, to the Fli1 deficiency in SSc ECs. Other Fli1 target genes that are dysregulated in SSc vasculature and may contribute to SSc vasculopathy, include CCN1/CYR61 and cathepsin L (56, 57). A recent study has also shown that Fli1 and its close homolog Erg1, transcriptionally control IFN-β gene in pulmonary ECs; thus, endothelial Fli1 and Erg1 deficiency leads to activation of IFN-β and its target genes in ECs (58). IFN-β is known to have potent anti-angiogenic effects (59).
A connection between impaired VEGF signaling and oxidative stress in SSc ECs was investigated by Tsou et al (60). They reported that 8-isoprostane, which is elevated in patients with diffuse cutaneous SSc, exerted anti-VEGF activity through the hyper-activation of thromboxane A2 receptor (TXAR). They also found that the inhibitory effects of TXAR signaling were linked to the abnormally elevated levels and activity of ROCK1/2 in SSc ECs. Together, these studies suggested that different molecular mechanisms may contribute to the impaired response to VEGF in SSc ECs.
Cultured SSc ECs maintain their abnormal characteristics, suggesting that the epigenetic mechanisms may be partly responsible for the phenotypic alterations. Indeed, epigenetic repression of bone morphogenetic protein receptor (BMPR)-IIgene was reported in SSc ECs (61). Studies by Tsou et al have also shown that the elevated expression of HDAC-5 in SSc ECs leads to the epigenetic repression of a subset of proangiogenic genes, including CCN1/CYR61, FSTL1, and PVRL2 (62). It is worth noting that studies in SSc fibroblasts demonstrated epigenetic suppression of Fli1 gene (63, 64). Although not formally tested, given the critical role of Fli1 in regulation of EC homeostasis, it is likely that epigenetic mechanisms also contribute to its down-regulation in SSc vasculature.
Excessive fibrosis
Fibrosis, which can occur in any organ, is characterized by replacement of normal tissue architecture with rigid connective tissue rich in collagen and other extracellular matrix (ECM) macromolecules. ECM is composed of collagens, particularly type I collagen, along with proteoglycans, fibrillin, fibronectin, tenascin-C, cartilage oligomeric matrix protein (COMP) and other matricellular and adhesion molecules. The ECM plays a vital instructive role regulating the differentiation, proliferation, migration, adhesion, biosynthetic capacity and survival of tissue stromal cells. Moreover, it also serves as a reservoir for latent growth factors. Fibroblasts are tissue-resident stromal cells of mesenchymal origin that are capable of both synthesis and degradation of ECM. There is increasing recognition that fibroblasts even within a single tissue, such as the skin, actually represent a heterogeneous population of cells with diverse origins and functions. In fibrotic tissues, the pool of collagen-producing activated mesenchymal cells is expanded. This involves not only proliferation of resident fibroblasts, which is generally rather modest, but also trans-differentiation from other tissue resident cell lineages. Recent advances in cell fate mapping point to perivascular mesenchymal stem cells called pericytes as precursors of collagen-producing myofibroblasts in fibrosis (65). In addition, under certain conditions, epithelial cells and endothelial cells can undergo transformation to fibroblasts, a process referred to as epithelial-mesenchymal transition (EMT) and EndoMT, respectively. Finally, circulating bone marrow-derived CD34+ mesenchymal progenitor cells called “fibrocytes” can secrete proinflammatory and profibrotic mediators, and also synthesize collagen (66).
In SSc patients, overproduction of collagen and other ECM molecules by activated fibroblasts occurs in the affected organs. The synthesis of ECM in fibroblasts during tissue remodeling is tightly regulated by soluble paracrine/autocrine mediators, cell-cell contact, hypoxia, and ROS, as well as through the biomechanical properties of the surrounding ECM (Fig. 2). Impaired collagen degradation and turnover, and increased numbers of biosynthetically active ECM-producing mesenchymal cells also appear to contribute to the process. In addition, rigid ECM further amplifies fibrosis through mechanotransduction (67).
Molecular mechanisms controlling fibroblast activation and regulation of relevant gene expression in SSc patients. Gene expression of collagens and other ECM molecules by fibroblasts is dysregulated in SSc patients, through soluble paracrine/autocrine mediators, ROS, and biomechanical properties of the surrounding ECM.
Positive regulators of ECM accumulation
Multiple mediators have potent effects on the maintenance of ECM homeostasis (68). Relevant signals implicated in fibroblast activation and resultant ECM overproduction in SSc patients are listed in Table II.
Selected extracellular signals implicated in SSc fibrosis
TGF-β = transforming growth factor-β; ECM = extracellular matrix; PDGF = platelet-derived growth factor; CTGF = connective tissue growth factor; IL = interleukin; TLR = toll-like receptor.
TGF-β
TGF-βs are secreted by platelets, monocytes/macrophages, T cells, dendritic cells and fibroblasts, and most cell types express surface receptors for TGF-β. A pool of biologically inactive TGF-β is sequestered in a latent form within the ECM bound to fibronectin and fibrillin through latent TGF-β binding proteins (LTBPs). In mesenchymal cells, TGF-β is a potent inducer of fibrillar collagen gene transcription and collagen secretion; it also stimulates migration, adhesion, ROS generation and trans-differentiation into myofibroblasts (Tab. III). In endothelial and epithelial cells, TGF-β elicits EMT and EndoMT, and transgenic mice overexpressing TGF-β in the endothelium develop spontaneous fibrosis in the skin and multiple organs via an EndMT-dependent mechanism (69). A recent study indicates that adipocytic cells resident in the white adipose tissue (WAT) of the skin (dermal WAT) also undergo a similar myofibroblast transition, a process now called adipocyte-mesenchyme transition, in response to TGF-β (70). TGF-β is exceedingly pleiotropic in its function, and plays essential roles in normal tissue repair, angiogenesis, and immune regulation (71). Activated TGF-β binds to ALK5, triggering an evolutionarily conserved signal transduction pathway involving cytosolic signal transducers called Smads (72, 73).
Fibrogenic TGF-β activities relevant to SSc pathogenesis
TGF-β = transforming growth factor-β; CTGF = connective tissue growth factor; IFN = interferon; PDGF = platelet-derived growth factor; EMT = epithelial-mesenchymal transition; EndoMT = endothelial mesenchymal transition.
Non-canonical TGF-β signaling
Alternative non-Smad pathways activated by TGF-β include non-receptor protein tyrosine kinases, p38 and JNK, integrin-associated focal adhesion kinase (FAK), and TGF-β-activated kinase TAK1, TRAF6, PI3 kinase and its downstream target Akt, and calcium-dependent phosphatase calcineurin. c-Abl, implicated in chronic myelogenous leukemia (CML), mediates profibrotic signals induced by TGF-β as well as PDGF (74). c-Abl is constitutively activated in SSc fibroblasts (75). Tyrosine kinase inhibitors prevented the development of skin fibrosis in scleroderma mouse models (76). JAK-2 is also activated in the skin of patients with SSc, particularly in fibroblasts, in a TGF-β-dependent manner (77). JAK2 inhibition not only prevented bleomycin-induced fibrosis but also effectively reduced skin fibrosis in TSK-1 mice (77). Smad7 and NR4A1 are cellular mediators that negatively regulate TGF-β signaling, but are also direct targets of TGF-β in negative feedback loops.
CTGF
CTGF or CCN2 is implicated in all forms of pathological fibrosis. Normal tissues have undetectable levels of CTGF, but its expression is markedly elevated in fibrotic conditions. Serum levels of CTGF correlate with the extent of skin and pulmonary fibrosis. Transgenic mice overexpressing CTGF develop skin fibrosis and microvascular alterations (78), whereas fibroblast-targeted ablation of CCN2 attenuated bleomycin-induced fibrosis in mice (79). In vitro, CTGF exerts profibrotic effects similar to those induced by TGF-β, suggesting that TGF-β responses are mediated through endogenous CTGF. However, the cellular CTGF receptors and mechanism mediating CTGF responses remain incompletely characterized.
Platelet-derived growth factor (PDGF)
Originally isolated from platelets, the PDGF isoforms are also produced by macrophages, endothelial cells and fibroblasts. In fibroblasts, PDGF induces ROS generation, ECM biosynthesis, and the secretion of TGF-β1, MCP-1, and IL-6 (80). Transgenic expression of constitutively active mutant PDGFα receptor (PDGFRα) elicits skin and organ fibrosis (81). SSc skin fibroblasts show elevated expression of both PDGF and its beta receptor (82, 83). Furthermore, PDGFRα mRNA correlated with CTGF and other fibrotic markers in the skin of SSc patients (84). A recent study reported the presence of functional circulating antibodies targeting the PDGF receptor in SSc patients (85). These autoantibodies were found to induce fibroblast activation and ROS generation.
Reactivation of developmental pathways
Developmental pathways play fundamental roles in organogenesis, and their deregulation is implicated in various disorders. Wnts comprise a family of glycoproteins with dual roles in cell–cell adhesion and transcriptional regulation, and essential roles in morphogenesis, stem cell homeostasis, and cell fate determination. Abnormal Wnt signaling is implicated in cancer, rheumatoid and osteoarthritis, osteoporosis, pulmonary arterial hypertension and even aging (86). Intracellular Wnt signaling is mediated via β-catenin, YAP/TAZ and other pathways, with extensive cross-talk with the TGF-β pathways. Transgenic mice overexpressing Wnt10b, or a constitutively active mutant form of β-catenin, develop exuberant wound healing, dermal fibrosis, atrophy of intradermal adipose tissue and collagen accumulation in the skin (87–88–89). Skin and lungs from SSc patients show marked nuclear β-catenin accumulation (90, 91). Transcriptome profiling of SSc skin biopsies shows elevated expression of multiple Wnt ligands, Wnt receptors, and Wnt targets.
The Hippo pathway, an evolutionarily conserved intracellular cascade involved in regulation of cell proliferation, apoptosis and differentiation (92–93–94), consists of serine/threonine protein kinases MST1/2 and LATS1/2, YAP and TAZ. In the nucleus TAZ/YAP function as transcriptional co-activators partnering with TEAD, Smads, Egr1 and Runx (95). TAZ and YAP are closely integrated with the Wnt/β-catenin pathway as components of the β-catenin destruction complex, and activation of Wnt signaling leads to concurrent induction of β-catenin and TAZ/YAP-mediated responses (96, 97). Notch is a transmembrane receptor for its ligand Jagged, and plays a fundamental role in embryonic development, as well as wound healing. Signaling via Snail, Notch regulates endothelial and fibroblast responses including myofibroblast differentiation. ADAM-17, a proteinase that initiates Notch signal transduction, was elevated in SSc skin biopsies (98).
The Hedgehog (Hh) pathway is another key developmental pathway involved in tissue repair and regeneration (99). Sonic hedgehog (Shh), the most widely expressed and best characterized mammalian Hh ligand, binds to Patched 1 and triggers intracellular signaling cascades via Gli1, 2 and 3. Accumulating evidence links aberrant hedgehog signaling to fibrosis in the liver, kidney, and lungs (100). Components of the Hh signaling pathway, including Shh, Gli-1 and Gli-2, are over-expressed in SSc skin (101). Fibroblast treatment with Shh stimulated collagen gene expression and myofibroblast trans-differentiation. In vivo studies further support a profibrotic role of the Hh pathway (101). Although developmental pathways are attractive as targets for anti-fibrotic therapies, due to their roles in stem-cell function, long-term blockade may be associated with substantial toxicity. Nevertheless, therapeutic efficacy with relatively low toxicity was demonstrated by low doses Hh/Wnt or Hh/Notch inhibitors. Combination therapies demonstrated additive anti-fibrotic effects in preventive, as well as therapeutic regiments (102).
Interleukins
IL-1β is an inflammatory cytokine produced by multiple cell types via intracellular inflammasome. Epidermal keratinocytes from SSc patients constitutively secreted IL-1α (103). Mice with deleted T-bet, a signature transcription factor that specifies Th1 differentiation of T cells, exhibit exaggerated fibrosis upon bleomycin challenge (104). Indeed, a predominance of Th2 cytokines, promotes fibrogenesis, and is a consistent observation in SSc and multiple fibrotic conditions (105).
IL-6, a multifunctional cytokine belonging to a family that also includes oncostatin M, and leukemia inhibitory factor, is produced by monocytes, activated B cells, fibroblasts, fibrocytes and endothelial cells. The biological activities of IL-6 are mediated through the gp130 receptors (also known as CD130), which form a heterodimer with IL-6 receptor to induce Jak-Stat signaling shared with other cytokines. Intracellular IL-6 signaling is tightly regulated by PIAS and SOCS, and the tyrosine phosphatase SHP-2, which dephosphorylates activated JAK and STAT. Transgenic mice carrying a mutated gp130 receptor exhibited worse lung fibrosis when challenged with bleomycin, whereas knockdown of IL-6 ameliorated bleomycin-induced fibrosis (106). Serum levels of IL-6 are elevated in SSc patients with interstitial lung disease, and correlated with the severity and progression of skin and lung involvement. A recent phase II clinical trial in SSc reported modest efficacy of tocilizumab, which binds to the IL-6 receptor and blocks IL-6 signaling (107). JAK is a major downstream pathway for many growth factors and cytokines, including IL-6. Blockade of JAK signaling is shown to be effective in preventing fibrosis (77), and thus is a potential alternative approach for IL-6 blockade.
The Th2 cytokine IL-13 is primarily implicated in asthma and fibrotic conditions. The profibrotic effects of IL-13 involve stimulation of endogenous TGF-β production by macrophages, as well as direct effects on fibroblast proliferation and collagen synthesis. Circulating monocytes, as well as CD8+ T cells, from SSc patients produce increased IL-13, and serum levels of IL-13 are elevated (108, 109).
Chemokines
Chemokines have a broad range of cellular targets and biological activities. The CC chemokine MCP-1 stimulates collagen production directly, as well as through induction of endogenous TGF-β production. Serum levels of MCP-1, along with those of MIP-1α, IL-8, CXCL8, and CCL18, are elevated in SSc and correlate with the extent of skin fibrosis. Explanted SSc fibroblasts show constitutive up-regulation of the MCP-1 receptor CCR2, while MCP-1 null mice are resistant to fibrosis (110). Additional chemokines overexpressed and potentially important in SSc patients include MIP-1α, CXCL8, and CCL18, as well as RANTES and PARC (CC chemokines), and IL-8, MIP-2and fractalkine (CXC chemokines). Insulin-like growth factor binding protein (IGFBP)-1 stimulates collagen synthesis and fibroblast proliferation and induces TGF-β. Adenoviral delivery of IGFBP-5 induced scleroderma-like fibrosis in mice (111).
Fibrogenic and vasoactive peptides
Ang II, part of the renin-angiotensin system (112), is also a potent profibrotic molecule that induces kidney, liver, heart, and skin fibrosis in animal models (113). Ang II signals through AT1R locally stimulates tyrosine kinases, c-Src, FAK, and PI3K, as well as RhoA/Rho kinase and MAP kinase families (112). The fibrogenic effects of Ang II are also mediated through TGF-β signaling, as well as through the up-regulation of NOX-4 and ROS. The role of angiotensin II type 2 receptor (AT2R) is less known, and some studies suggest that it may have a protective function by counteracting AT1R-mediated effects. The “protective arm” of renin-angiotensin system also includes the newly described Ang-derived small peptide. Ang(1–2–3–4–5–6–7) signals through the G protein-coupled receptor Mas, as well as angiotensin-converting enzyme (ACE)-2 that selectively degrades Ang II, generating Ang(1-7). Ang II/AT1R blockade (114, 115), or up-regulating the ACE2/Ang(1-7)/Mas axis, mitigated experimentally induced lung fibrosis and pulmonary hypertension (116, 117). ETs including ET1, ET2, and ET3, that signal through two distinct receptors, ETAR and endothelin type B receptor (ETBR). ET1, the best characterized isoform, has a well-documented role in fibrosis (118). Potent profibrotic effects of ET-1 were demonstrated in transgenic mice overexpressing ET-1 in skin or lungs, or by blocking ET-1 signaling in bleomycin induced models (119, 120). ET-1 and its receptors are elevated in SSc skin (121).
Bioactive lipids
Several prostanoids inhibit fibrotic responses through a variety of mechanisms. In contrast, prostaglandin (PG)-F2α, which is elevated in SSc patients with pulmonary fibrosis, stimulates collagen production and fibroblast proliferation (122). Mice with targeted deletion of the PG-F receptor are protected from bleomycin-induced pulmonary fibrosis. Lysophosphatidic acid (LPA) is generated locally via the hydrolysis of membrane phospholipids. LPA was shown to induce fibroblast chemotaxis and CTGF production (123, 124). Significantly, levels of LPA are elevated in patients with pulmonary fibrosis (125, 126). LPA1 knockout mice are protected from bleomycin-induced skin and lung fibrosis. A recent study indicates that LPA induces αvβ6 integrin-mediated activation of latent TGF-β in epithelial cells, contributing to sustained autocrine and paracrine TGF-β signaling (124). On-going clinical trials are focusing on the safety and efficacy of therapeutic LPA antagonism in SSc patients.
Innate immune signaling and the inflammasome
The robust association of SS with genes related to innate immunity implicates innate immunity in SSc (127). TLRs are key components of the innate immune system. The TLRs serve as the first line of defense against microbial pathogens. As primary pattern recognition receptors, TLRs sense and react rapidly to proteins, lipids and nucleic acids from infectious pathogens, or from the injured host (self). TLRs can be localized to the cell surface (TLR2 and TLR4), or endosomal membranes (TLR3, TLR7 and TLR9). Activation of TLR4 by lipopolysaccharide plays a causal role in liver fibrosis, with sensitization of hepatic stellate cells to TGF-β as the underlying mechanism (128). In addition, TLR4 activation also induced the expression of the profibrotic Egr1/2 transcription factors, enhanced focal adhesion kinase (FAK) activation, and suppressed anti-fibrotic microRNAs (miR29). Activation of fibroblast TLRs in SSc might be triggered by endogenous TLR ligands. These so-called damage-associated molecular patterns (DAMPs) are generated at sites of tissue injury in response to mechanical damage, inflammation, autoimmunity and oxidative stress. Three general classes of DAMP endogenous TLR ligands are recognized: matricellular molecules such as hyaluronan, alternatively spliced form of fibronectin (extra domain A) and tenascin C, and biglycan; intracellular stress proteins (alarmins) such as HMGB1 and Hsp60; and immune complexes, or nucleic acids released from damaged or necrotic cells.
The expression of several TLRs (TLR3, TLR4 and TLR9) and the TLR4-specific co-receptor MD2, is elevated in SSc skin and lung biopsies, and levels generally correlate with disease severity and progression. Elevated TLR4 expression is accompanied by extracellular accumulation of endogenous TLR ligands. Mice lacking either TLR4, or endogenous TLR4 ligands fibronectin-extra domain A or Tenascin-C, exhibits reduced skin, lung and cardiac fibrosis when challenged with bleomycin or Ang II. Similarly, mice lacking interferon regulatory factor (IRF) 5, a downstream mediator for TLR4 signaling, are protected from fibrosis. These observations suggest that fibroblasts exposed to endogenous TLR4 ligands generated during tissue injury switch to an activated phenotype. In this way, fibroblast signaling initiated and sustained by damage-associated endogenous TLR ligands in SSc might be responsible for transformation of self-limited regenerative tissue repair into intractable fibrotic scar formation. TLR9, expressed on endosomes, recognizes nucleic acids, primarily unmethylated CpG-containing bacterial and viral DNA. By also recognizing endogenous damage-associated DNA as well as mitochondrial DNA released during cell injury, TLR9 plays a major pathogenic role in autoimmune diseases such as systemic lupus erythematosus. TLR9 stimulation on fibroblasts elicits fibrotic responses in vitro, and TLR9 expression and activity are elevated in SSc skin biopsies. Nonetheless, the precise role of TLR9 in SSc remains to be clarified. The TLR3 ligand poly(I:C) causes dramatic induction of type I IFN, along with IL-6 and other inflammatory cytokines in normal fibroblasts, and blocks TGF-β-induced fibrotic responses (129). TLR3 therefore plays a complex role in SSc pathogenesis, with both pro- and anti-fibrotic activities.
In addition to the TLRs, both immune and non-immune cells contain cytosolic innate immune sensors including nucleotide binding oligomerization domain-like receptor (NLR), RIG-I, and Nalp3 that recognize and respond to nucleic acids, damage-associated endogenous molecules, as well as environmental signals such as silica, bleomycin, and gadolinium. Once activated, these cytosolic pattern recognition receptors facilitate inflammasome assembly with activation of caspase-1 and secretion of proIL-1β and IL-18. Nlrp1, a key scaffolding protein for inflammasome assembly, is a susceptibility gene for SSc and associated pulmonary fibrosis. Moreover, expression of the NLRP3 inflammasome is elevated in SSc skin biopsies. The TNFAIP3 gene is a candidate gene associated with SSc. TNFAIP encodes the ubiquitin-editing enzyme A20, a critical intracellular negative regulator of both TLR signaling and inflammasome activation. The expression and function of A20 has been extensively investigated in hematopoietic cells, but its biology in the context of stromal cells and fibrosis is currently unknown. Impaired A20 expression or activity in SSc might contribute to unchecked innate immune signaling, fibroblast activation and consequent tissue fibrosis. Inflammasome activation and IL-1β secretion play pathogenic roles in experimental lung fibrosis in the mouse, and are increasingly recognized as important factors and potential therapeutic targets in SSc and other fibrosing conditions (130, 131).
Negative regulators of ECM accumulation
To optimize tissue repair and forestall uncontrolled matrix accumulation and scarring in response to injury, a plethora of cell-autonomous and paracrine anti-fibrotic regulatory mechanisms have evolved. For example, Smad7 is a cell-autonomous negative feedback regulator of fibrotic TGF-βsignaling. Smad7 is a TGF-β-inducible protein that blocks canonical TGF-β signaling by binding to the TGF-β receptor and accelerating its degradation. Functional impairment of Smad7 was demonstrated in SSc fibroblasts, which might result in unopposed canonical Smad signaling in these cells. Other cell-autonomous repressors of collagen synthesis and myofibroblast transformation include the transcription factors Sp3, Fli-1, p53, Ras, cofactors such as Nab2, HDACs including the anti-aging enzymes SIRT1 and SIRT3, microRNAs, the deubiquitinase A20 and nuclear receptors such as NR4A1 and peroxisome proliferator-activated receptor (PPAR)-γ. Particularly interesting in this regard is the transcription factor Nrf2, which is induced by a variety of stresses, and in turn activates the expression of hundreds of anti-oxidant and cyto-protective genes (131). Accumulating data indicate impaired expression, regulation or function of these endogenous inhibitors in SSc, which may contribute to failure to limit fibroblast activation.
IFN-γ
The inflammatory Th1 cytokine interferon-γ (IFN-γ), produced primarily by lymphocytes, is a major paracrine negative regulator of fibroblast activation and myofibroblast transformation. IFN-γ represses collagen gene expression and abrogates stimulation induced by TGF-β (132). However, IFN-γ may have an indirect profibrotic effect by stimulating TGF-β and ET1 synthesis in microvascular ECs, inducing EndoMT, and promoting vascular injury (133). Therapeutic trials of IFN-γ for SSc patients have demonstrated modest and inconsistent improvement in skin fibrosis, but a randomized trial in idiopathic pulmonary fibrosis showed no benefit, and enthusiasm for this form of therapy has waned in recent years (134, 135).
Adipogenesis, PPAR-γ and adipokines
PPAR-γ is a type 2 nuclear receptor in adipocytes, but also expressed in macrophages, ECs and fibroblasts. Initially identified as an essential regulator of adipogenesis, PPAR-γ is a pleiotropic dual function molecule acting as both a nuclear receptor and ligand-inducible transcription factor. A number of lipid moieties and electrophilic prostanoids such 15d-PGJ2 can function as endogenous (natural) PPAR-γ ligands. Remarkably, the insulin-sensitizing glitazone drugs such as rosiglitazone or pioglitazone were found to be potent pharmacological agonists for PPAR-γ. Many target genes regulated by PPAR-γ are implicated in adipogenesis.
PPAR-γ also regulates a broad range of vascular and immune processes. Abnormal PPAR-γ function is implicated in lipodystrophies, atherosclerosis, PAH, and inflammatory diseases. Recent studies have explored the regulation and role of PPAR-γ in connective tissue homeostasis and fibrotic diseases. In skin and lung fibroblasts, 15d-PGJ2 or pharmacological PPAR-γ ligands caused virtual abrogation of TGF-β-induced collagen production, myofibroblast trans-differentiation, EMT, and other Smad3-dependent profibrotic responses (136). Some of these anti-fibrotic effects are mediated through the adipokine adiponectin, a direct transcriptional target of PPAR-γ. In turn, adiponectin by binding to its two surface receptors induces AMP-activated protein kinase (AMPK), and disrupts focal adhesion assembly and FAK activation. The skin expression and activity of PPAR-γ are impaired in patients with SSc and scarring alopecia. In SSc, both candidate gene and genome-wide approaches have identified SNPs in PPARG, but the functional consequences of these genetic polymorphisms and their precise contribution to disease manifestations such as fibrosis and PAH remain to be elucidated.
Recent studies have implicated impaired adipogenesis (formation of adipocytes in white adipose tissue) within the dermal adipose layer as a potentially pathogenic factor in SSc (137). The expression of adipogenic genes, and peroxisomal biogenesis, are compromised in patients with SSc, as well as pulmonary fibrosis. Additionally, circulating levels of adiponectin, a direct transcriptional target of PPAR-γ and a marker for its activity, are reduced in SSc patients and correlate with disease severity (138). Furthermore, PPAR-γ expression in lesional tissue shows an inverse relationship with TGF-β signaling. It is worthy of note that many factors implicated in fibrosis including TGF-β, Wnt ligands, hypoxia, LPA, and CTGF, potently inhibit PPAR-γ expression or function (139). Compromised PPAR-γ expression and function in SSc and other fibrotic conditions may be an important pathogenic factor that, perhaps by disrupting adipogenesis and the generation of anti-fibrotic and vasculoprotective adipokines such as adiponectin, contribute to the persistence and progression of fibrosis as well as vascular damage (140).
Sirtuins
The anti-aging sirtuins represent a family of mammalian deacetylases that catalyze a number of cellular proteins, contributing to extending lifespan and health-span in many organisms. Emerging evidence implicates defective sirtuin expression and function in SSc patients (141–142–143). In particular, reduced SIRT1 and SIRT3 in lesional tissues, potentially resulting from altered nicotinamide adenine dinucleotide (NAD) balance and metabolic dysregulation, contributes to unopposed TGF-β signaling and fibrosis. Sirtuins exert an anti-fibrotic effect, in part by antagonizing canonical TGF-β signaling (143) and attenuating cellular mitochondrial oxidative stress. This impaired function in SSc patients phenocopies the changes seen in various age-associated pathologies, resulting in constitutive fibroblast activation, ROS generation and myofibroblast transformation. In this regard, SSc-associated pathological fibrosis shares biological mechanism common with aging. Restoring normal SIRT activity by augmenting NAD levels might be a promising novel therapeutic strategy for chronic fibrosis.
Immune dysregulation and autoimmunity
SSc has been traditionally recognized as an autoimmune disease, but lack of evidence showing roles of autoimmunity in generating the clinical and pathologic phenotypes and relative inefficacy of immunosuppressive therapies have led many to doubt the importance of autoimmunity in the pathogenesis. Many lines of evidence mentioned above have shown that immune mediators, including innate immune signals, immune cells, and cytokines promote vascular damage and pro-fibrotic environment, but these responses are rather non-specific and there is no evidence supporting involvement of autoreactivity in these processes. In this regard, epithelial cell-specific Fli1 knockout mice spontaneously developed dermal and esophageal fibrosis, along with remarkable autoimmunity derived from defects of AIRE, an autoimmune regulator, in the thymic epithelial cells (144). In contrast, distinct specificities of antinuclear antibodies (ANAs) are selectively detected in sera from SSc patients and are associated with unique disease manifestations, although they do not have a proven pathogenic role. A new group of autoantibodies reactive with cell surface receptors or extracellular matrix proteins have been identified in SSc patients, and potentially directly activate pathways that may contribute to tissue and vascular damage.
Roles of non-specific immune and inflammatory mediators in SSc pathogenesis
A primary role of immune mechanisms in the pathogenesis of SSc involves the profound effect of immune cells and their secreted factors on other cells and tissues, including ECs and fibroblasts. In fact, recent studies using genome-wide transcriptome analysis of SSc skin and lung reveal a robust and consistent inflammatory signature, indicating activation of immune signaling within the affected tissues (145). Innate immune signals operated in response to activation of TLRs and other pattern recognition receptors contribute to pro-inflammatory and pro-fibrotic cellular milieu in SSc. In addition, a number of studies documented local and systemic changes in cytokine, chemokine, and growth factor levels and activation, migration, and differentiation of T cells, B cells, monocytes, macrophages, natural killer cells, and dendritic cells (146). T cells show skewing toward Th2 or Th17, which creates a fibrosis-prone immune environment (147). Important roles for B cells in SSc patients are highlighted in generating autoantibodies as well as secreting IL-6, TGF-β, and other pro-fibrogenic cytokines, which promote fibroblast activation (148).
Autoantibodies
One of the distinctive hallmarks of the immune dysregulation in SSc patients is the presence of circulating autoantibodies reactive with various cellular components (149). The majority of disease-associated autoantibodies in SSc patients are ANAs that target proteins that play essential roles in transcription, splicing, and cell division. Actually, ANAs detected by indirect immunofluorescence (IIF) technique are found in >95% of the patients. Currently, at least 10 ANA specificities specific to SSc patients have been reported and well characterized (Tab. IV). These SSc-specific ANAs are detectable in >80% of SSc patients, and are associated with unique disease manifestations. Interestingly, SSc-specific ANAs are usually present at diagnosis of SSc or even precede appearance of SSc-related clinical manifestations such as skin thickness. In addition, patients rarely have two or more SSc-specific ANAs together, indicating mutual exclusiveness. Therefore, individual ANAs are attractive biomarkers in routine rheumatology practice, although their direct role (if any) in the disease pathogenesis still remains to be documented.
Main cellular function of target and clinical correlation of SSc-specific ANAs
DM = dermatomyositis; DU = digital ulcer; GAVE = gastric antral vascular ectasia; GI = gastrointestinal tract; ILD = interstitial lung disease; MCTD = mixed connective tissue disease; PAH = pulmonary arterial hypertension; SRC = scleroderma renal crisis.
Recently, a new group of autoantibodies reactive with extracellular matrix proteins or cell surface receptors have been reported (Tab. V). Hypothetically, these autoantibodies are potentially involved in pathogenesis of SSc, by promoting immune and inflammatory processes through the Fc region of the immunoglobulin and/or directly activating pathways that may contribute to tissue and vascular damage. Autoantibodies with capacity to bind ECs or fibroblasts have been reported in patients with SSc, but are not specific to SSc patients. Several groups have reported pathogenic roles of AECAs, which are capable of inducing EC apoptosis, and inducing cytokines, chemokines and adhesion molecules on ECs (15–16–17). Anti-fibroblast antibodies have been shown to be capable of inducing the expression of adhesion molecules and the preferential transcription of chemokines with profibrotic and proangiogenic potential in a proteasome- and TLR4-dependent manner and Fc-independent mechanism that allows internalization of the antibodies (150). Using a proteomic approach combining two-dimensional electrophoreses and immunoblotting, Terrier et al have identified 13 target antigens of anti-fibroblast antibodies from SSc patients with α-enolase being the main target antigen (151). Although these autoantibodies have been reported by several groups, more studies are needed to define specific pathogenic effects and to characterize the cell surface antigens that are targeted by those antibodies. Antibodies against cell surface receptors are also reported in patients with SSc, and are potentially functional through binding to extracellular receptors and triggering the activation of signal transducing pathways, resulting in a stimulatory or suppressive effect (152). For example, stimulatory autoantibodies toward PDGF receptor or antibodies targeting G protein–coupled receptors (e.g. AT1R and ETAR) are potentially pathogenic roles in the pathogenesis of SSc. High levels of these functional autoantibodies might dysregulate the response of ECs and fibroblasts as well as innate and adaptive immune cells, including myeloid cells and lymphocytes, leading to extensive fibrosis, vasculopathy, and abnormal immune responses. However, results are often conflicting, probably because of lack of standardization of the assays for detection of those antibodies.
Putative pathogenetic molecules targeted by autoantibodies in SSc patients
ND = not determined; GI = gastrointestinal involvement; PAH = pulmonary arterial hypertension; SRC = scleroderma renal crisis; DU = digital ulcer; dcSSc = diffuse cutaneous systemic sclerosis; ILD = interstitial lung disease; EScSG = European Scleroderma Study Group.
Autoantibodies to fibrillin-1 have been shown to be able to activate normal human fibroblasts in vitro (153). In fact, normal human fibroblasts exposed in vitro to affinity-purified anti-fibrillin-1 antibodies have shown over-expression of several extracellular matrix components and nuclear translocation of phosphorylated Smad3, suggesting that immunity against fibrillin-1 causes the release of sequestered TGF-β1 from fibrillin-1-containing microfibrils in the extracellular matrix, with subsequent fibroblast activation. Finally, antibodies against extracellular matrix metalloproteinases (MMP) may promote fibrosis hampering extracellular matrix degradation (154, 155). Antibodies against enzymes of the oxidative stress cascade, including methionine sulfoxide reductase A and peroxiredoxin I may enhance the oxidative stress and contribute to the disease severity of SSc by inhibiting the enzyme activity (156, 157).
Summary
The pathogenesis of SSc is complex, multifactorial, and incompletely understood. Recent efforts successfully identified molecular and cellular underpinnings involved in pathogenic process of SSc; however, effective therapies are not yet available. The discovery of key pathways and mediators identified within the skin and blood vessels, and of a growing number of druggable cellular and molecular targets, is a fast-moving research field with enormous translational potential.
Footnotes
Financial support: No grants or funding have been received for this study.
Conflict of interest: None of the authors has financial interest related to this study to disclose.