Extracellular Matrix-derived Damage Associated Molecular Patterns: Implications in Systemic Sclerosis and Fibrosis

Abstract

Damage-associated molecular patterns (DAMPs) are intracellular molecules released under cellular stress or recurring tissue injury that serve as endogenous ligands for toll-like receptors (TLRs). Such DAMPs are either actively secreted by immune cells or passively released into the extracellular environment from damaged cells, or generated as alternatively spliced mRNA variants of extracellular matrix (ECM) glycoproteins. When recognized by pattern recognition receptor (PRR)s such as TLRs, DAMPs trigger innate immune responses. Currently, the best characterized PRRs include, in addition to TLRs, nucleotide-binding oligomerization domain (NOD)-like receptors (NLRs), retinoic acid-inducible gene-I (RIG-I)-like RNA helicases, C-type lectin receptors (CLRs), and many more. Systemic sclerosis (SSc) is a chronic autoimmune condition characterized by inflammation and progressive fibrosis in multiple organs. Using unbiased survey for SSc-associated DAMPs, we have identified the ECM glycoproteins fibronectin-EDA (Fn-EDA) and tenascin-C (TNC) as the most highly up-regulated in SSc skin and lung biopsies. These DAMPs activate TLR4 on resident stromal cells to elicit profibrotic responses and sustained myofibroblasts activation resulting in progressive fibrosis. This review summarizes current understanding of the complex functional roles of DAMPs in progression and failure of resolution of fibrosis in general, with a particular focus on SSc, and considers viable therapeutic approaches targeting DAMPs.

Introduction

The innate immune system comprising a variety of pattern recognition signals (PRR) recognizes both pathogen-associated molecular pattern (PAMP) derived from bacteria or viruses, and endogenous sterile damage-associated molecular patterns (DAMPs) derived from extracellular matrix (ECM) or from dying cells as a result of chronic or recurrent tissue damage [1]. DAMPs then trigger cellular innate immune responses via toll like receptors (TLR) and non-TLR sensors and inflammasomes expressed in both immune as well as non-immune cells [2-5]. Recent findings demonstrate important pathogenic roles for DAMPs in TLR-dependent activation of fibroblasts to induce fibrotic gene expression and myofibroblast differentiation in multiple fibrotic disorders [6-11]. In particular, we showed that the DAMPs Fn-EDA and tenascin-C are elevated in skin and lung biopsies from patients with SSc. By initiating innate immune signaling in resident fibroblasts through TLR4, Fn-EDA and tenascin-C cause persistent myofibroblast activation, preventing fibrosis resolution. In this review, we will focus on recent progress in understanding the pathogenic networks that underlie DAMP-induced self-sustaining fibroblast activation, DAMP interaction with TLRs and other pattern recognition receptors, downstream cellular signaling pathways, and innovative therapies targeting DAMPs to block fibrosis.

ECM derived DAMPs in fibrosis

The differentiated fibroblast and its ECM microenvironment are crucial in the pathogenesis of fibrosis [12]. Persistence of ECM-associated DAMPs contributes to many inflammatory and autoimmune diseases. DAMPs encompass an extensive repertoire of molecules released by dying or damaged cells, such as high mobility box 1 (HMGB1) and heat shock proteins (HSPs), self-nucleic acids including single and double-stranded DNA, RNA, mitochondrial DNA and nucleic acid-antibody complexes; Serum Amyloid A (SAA), fragments of ECM macromolecules, and “oncofetal” ECM macromolecule isoforms that are generated through alternative mRNA splicing [3, 13]. The class of ECM-derived DAMPs comprises decorin, biglycan, versican, low molecular weight hyaluronan, heparan sulfate, along with Fn-EDA and tenascin-C. Decorin and biglycan are small leucine-rich proteoglycan components of ECM that decorate the surface of collagen fibrils regulating fibril assembly. A direct role of decorin interaction with TGF-β from a former study proposed that decorin type proteoglycans regulate TGF-β activities by sequestering TGF-β into extracellular matrix, and removal of the chondroitin sulphate chains using chondroitinase treatment increased TGF-β binding activity of decorin [14]. In cardiac fibroblasts, biglycan levels were elevated after left ventricular pressure overload, and blocking biglycan reduced left ventricular hypertrophy and fibrosis [8]. Moreover, a clinical study of 120 patients with biopsy-proven hepatitis B demonstrated elevated biglycan levels in serum, which might serve as a potential fibrosis marker [15].

Versican, an alternatively-spliced chondroitin sulfate proteoglycan, was shown to activate hepatic stellate cells in liver fibrosis [16]. Interestingly, versican knockdown in hepatic stellate cells attenuated fibrosis. The levels of glycosaminoglycan hyaluronan, another DAMP, were shown to be elevated in multiple fibrotic diseases [10]. For instance, during the progression of hypertrophic cardiomyopathy hyaluronan metabolism is studied in cardiac tissue from patients. Glucose is metabolized rather than regular fatty acids as an energy source and results in increased production of the extracellular glycosaminoglycan hyaluronan and larger cardiomyocytes in cardiac tissue from patients studied using gas phase electrophoresis [17]. The alternatively spliced variant of fibronectin called Fn-EDA has been implicated in skin, bone marrow and hepatic fibrosis [7, 18, 19]. Alternatively spliced variant tenascin-C has been implicated in fibrosis in lung, heart and skin [6, 9, 20]. By immunohistology, we identified Fn-EDA and tenascin-C as most highly up-regulated DAMPs in SSc skin and lung biopsies [19, 20] which we will discuss in detail in the following sections. The involvement of these ECM-derived DAMPs in fibrosis via innate immune signaling is schematically presented in Fig. 1, and the various ECM-DAMPs in fibrosis and related disease conditions are summarized in table 1.

Figure 1: Macromolecules endogenously generated from the matrix serve as TLR4 ligands to activate innate immune responses in resident stromal cells and promote fibrosis.

A) Chronic or recurrent tissue injury generates DAMPs via alternative mRNA splicing, or post-translational modifications, of matrix macromolecules are accumulate in the ECM microenvironment, activate innate immune signaling converting self-limited tissue repair into sustained and unresolving fibrosis. Sustained DAMP accumulation further triggers TLR-mediated stromal cell activation in a vicious fibrosis amplification loop. The mRNA splicing factor SRSF6, upregulated in fibrotic fibroblasts, contributes to generation of alternatively-spliced tenascin-C isoform. B) Schematic representation of fibronectin Type III module of tenascin-C (upper panel) and fibronectin (lower panel). C) Vicious cycle of fibrosis involving DAMPs, innate immune signaling via TLR4, TGF-beta and Wnt signaling. Furthermore, intrinsic protective signals that usually regulates as brakes on fibroblast activation might be nonfunctional in SSc. Both the figures 1B and 1C were adapted from previous review (Figures 1 and 2) (Bhattacharyya et al) [13].

Table 1:

DAMPs derived from ECM with emphasis on fibrosis

DAMP	Sample/cell type studied	Disease	Findings	Reference
Biglycan	Fibroblasts	Cardiac hypertrophy, Fibrosis	Fibroblasts as the major source of biglycan. Elevated levels of biglycan in the pressure-overloaded heart resulting in pro-hypertrophic gene program in cardiomyocytes.	Beetz et al [8]
Biglycan	Serum	Liver fibrosis	Biglycan levels of chronic hepatitis B patients are significantly higher than healthy controls in serum. Significant positive correlation exists between serum biglycan level and fibrosis stage (P=.004; r=.213).	Ciftciler et al [15]
Versican	Primary mouse hepatic stellate cells, CCl4-treated mice	Liver fibrosis	Transient knockdown of versican in Hepatic Stellate Cells reduced markers of fibrogenesis. Versican expression is increased in the mouse model of liver fibrosis.	Bukong et al [16]
Hyaluronan	Human cardiac tissue	Hypertrophic cardiomyopathy	NMR and qRT-PCR analysis showed dysregulated glucose and hyaluronan metabolism in the patients. Gas phase electrophoresis showed increased low molecular mass hyaluronan and larger cardiomyocytes in patients.	Loren et al [17]
FN-EDA	Human fibrosis liver samples, CCl4-treated mice, FNEDA KO mice	Hepatic fibrosis	Fn-EDA was positively correlated with angiogenesis (CD 31) in hepatic fibrosis FN-EDA KO mice observed with reduced neovessel density, CD31 expression and fibrosis compared to control group	Su et al [7]
FN-EDA	Plasma and bone marrow biopsies	Bone marrow fibrosis	Fn-EDA levels are increased in plasma and bone marrow biopsies of primary myelofibrosis patients as compared with healthy controls	Malara et al [18]
FN-EDA	Human serum and dermal fibroblasts, FNEDA KO mice	Systemic sclerosis	5-fold increase in serum FnEDA levels in scleroderma patients compared to control subjects. TGF-β induced a dose-dependent increase in Fn-EDA levels (both mRNA and protein) in fibroblasts. Fn-EDA KO mice treated with bleomycin demonstrated reduced dermal thickness and collagen compared	Bhattacharyya et al [19]
Tenascin-C	Skin biopsies, dermal fibroblasts and mouse models of	Systemic sclerosis	Increased levels of tenascin-C mRNA in SSc skin biopsies compared with healthy control. Tenascin-C induced profibrotic responses in fibroblasts. Tenascin-C mediated fibrotic responses are TLR4-dependent.	Bhattacharyya et al [20]

Tenascin-C

Tenascin C is a large multidomain multifunctional ECM glycoprotein that undergoes alternative mRNA splicing to yield multiple isoforms [21]. Tenascin-C is a hexamer, each monomer consisting of the tenascin assembly (TA) domains at the N-terminus, responsible for the assembly of oligomeric proteins, epidermal growth factor-like (EGFL) repeats, fibronectin type III (FN III) domains, and a globular fibrinogen domain at the C-terminus (Fig.1B). Under physiological condition in the adult, cells produce small tenascin C isoform (MW180–190 kDa) [22] (Fig. 1B). Pathological conditions generate tenascin-C full-length ‘large’ TNC isoforms (MW 210, 220, 230, 250, 260, 280, 320 and 330 kDa) that are spliced alternatively between fibronectin-like domain 5 and 6 (TNCfn5 and TNCfn6) (Fig. 1B). These large tenascin C isoforms accumulated in the ECM of a variety of human tumors including breast, colon, bladder, ovaries, prostate, pancreas, kidney, liver, uterus, brain, mouth, lung, skin, cartilage, connective tissues and peripheral nervous system with post-transcriptional modifications [23]. The human tenascin-C molecule contains nine alternatively-spliced FNIII repeats. Earlier studies suggested involvement of fibronectin domain of tenascin-C in chronic hepatitis and several cancers [24-26].

In a meta-analysis of three skin transcriptome datasets (GSE56038, GSE59785), we found elevated tenascin-C mRNA expression (p<0.00010) in 80 SSc biopsies compared to 26 healthy controls [20]. Tenascin-C mRNA levels showed correlation with both TLR4 as well as its direct target IL-6. Moreover, tenascin-C mRNA expression was correlated with increased expression of fibrotic genes, suggesting a role for tenascin-C in driving fibrosis via TLR4. Interestingly, levels of tenascin-C showed strong correlation with the modified Rodnan Skin Score (MRSS; r=0.73) in the inflammatory intrinsic SSc biopsy subset. Circulating tenascin-C levels were elevated in serum from SSc patients with interstitial lung disease (SSc-ILD) compared to SSc patients with no ILD or healthy controls. By immunofluorescence, we showed significantly increased accumulation of tenascin-C in SSc skin biopsies, largely localized in the papillary dermis near basement membrane, while negligible tenascin-C expression was noted in healthy controls [20].

Normal skin fibroblasts incubated with tenascin-C showed enhanced collagen gene expression and myofibroblast transformation mediated via TLR4 [20]. Immunoprecipitation assays revealed direct interaction of cellular TLR4 with tenascin-C. Treatment with a TLR4 inhibitor, or with MyD88 blocking peptides, attenuated tenascin-C-induced stimulation of collagen and alpha smooth muscle actin (α-SMA), while TLR4-deficient skin fibroblasts failed to mount fibrotic responses upon tenascin-C stimulation [20]. Mice lacking tenascin-C showed attenuation of bleomycin-induced skin and lung fibrosis, and accelerated fibrosis resolution, accompanied by reduced TGF-β signaling and expression of TLR4 target genes such as IL-6 and MCP-1. Altogether, these results identify tenascin-C as an endogenous TLR4 ligand that is elevated in SSc, drives TLR4-dependent fibroblast activation, and by its persistence inhibits fibrosis resolution. We recently summarized the role of tenascin-C in fibrosis along with its mechanism and potential therapeutic approaches to ameliorate fibrosis by targeting TNC [27].

Tenascin-C is also implicated in renal and cardiac fibrosis. For instance, tenascin-C was found to be highly expressed in kidneys in the unilateral ureteral obstruction model, while its knockdown prevented fibrotic responses [28]. The effects of TNC appeared to be mediated through integrin/focal adhesion kinase/ERK1/2 [28]. Another study on kidney fibrosis demonstrated that tenascin-C is expressed by NG2+PDGFRβ+ cells around injured tubules and was responsible for profibrotic responses via STAT3 signaling [29]. Moreover, tenascin-C is found to be constitutively expressed by the renal medullary interstitial cells but not in endothelial cells and progenitor cells [29]. Further, tenascin-C was found to activate cardiac fibroblasts via p65/NF-κB promoter demethylation resulting in proinflammatory responses and fibrosis, while tenascin-C KO mice demonstrated attenuated responses in a mouse model of muscular dystrophy [30].

Tenascin-C alternative pre-mRNA splicing is mediated by the enzyme serine/arginine-rich splicing factor 6 (SRSF6)

Tenascin-C pre-mRNA notably undergoers alternative splicing to generate tenascin-C-FL, a process catalyzed by serine/arginine-rich splicing factor 6 (SRSF6). Serine/arginine-rich splicing factor 6 is a multifunctional protein that regulates alternative mRNA splicing of hundreds of target genes. Additional roles for SRSF6 include mRNA export and translation. Several cancers including colon and ovarian, show upregulated expression of SRSF6, which generally serves as an oncogene, and is thus a potential therapeutic target. The expression of SRSF6 is also transiently elevated during wound healing. Forced expression of SRSF6 in stromal cells was shown to generate full-length tenascin-C-FL isoform in mouse [31]. In this regard, Jansen et al generated a mouse transgenic for human SRSF6 cDNA and internal ribosome-entry site (IRES)-EGFP under the control of a tetracycline-responsive promoter (TREtight) at the Col1a1 locus (ColA1-SRSF6). The mouse skin and small intestine showed severe epithelial hyperplasia with increased keratinocyte number in the interfollicular epithelium [31].

Overexpressing SRSF6 in transgenic mice further promoted aberrant alternative splicing (AS). To determine how SRSF6 controls tenascin C AS, the authors searched for putative SRSF6 motifs within exons at or near the alternatively spliced region of tenascin C by using genome-wide mapping of splicing factor binding sites. Alternatively spliced exons of tenascin C scored among the highest changes in their analysis, demonstrating the potential role of tenascin C as a functionally relevant target of SRSF6. The authors further proved that SRSF6 specifically promotes alternative splicing including exons E10–15 tenascin C isoforms [31]. Additional literature supports the global switches in splicing associated with the epithelial to mesenchymal transition [32], encouraging a need to understand whether the splicing in epithelial phenotype can have an impact on the fibroblasts.

We have demonstrated that SRSF6 siRNA knockdown in normal skin fibroblasts reduced the production of tenascin-C-FL (data not shown). Skin biopsies from patients with SSc display elevated SRSF6 mRNA expressions (GSE56038 and GSE59785). Importantly, tenascin-C expression showed significant correlation with SRSF6 expression (r=0.358, p=0.0002) in the same dataset (Fig. 2A). We further confirmed elevated SRSF6 mRNA and protein expression in SSc skin biopsies compared to healthy controls (Fig. 2B). Numerous SRSF6-positive spindle-shaped fibroblastic cells as well as endothelial cells in both papillary and reticular dermis were detected in SSc skin, but not in healthy controls (Fig. 2C). The number of SRSF6-positive cells showed strong correlation (r=0.76, p=0.036) with tenascin-C-FL levels (Figs. 2D and 2E). By immunofluorescence, the most prominent tenascin-C-FL in SSc skin biopsies was detected in the papillary dermis subjacent to the basement membrane [20]. These findings together reflect a notable connection between alternative tenascin-C mRNA splicing and SRSF6 in SSc. Additional recent studies provide fresh evidence for a potential role of SRSF6 in pathological fibrosis. For instance, bleomycin and TGF-β1 both stimulated SRSF6 levels in pleural mesothelial cells (PMCs), and SRSF6 enhanced their proliferation and synthesis of collagen. In vivo, inhibition of SRSF6 by shRNA knockdown mitigated pleural fibrosis in mice [33]. Of note, SRSF6 was previously implicated in the pathogenesis of SSc microvasculopathy via in alternative splicing of VEGF, yielding an anti-angiogenic isoform [34]. Additionally, this study further demonstrated that the genetic SRSF6 (SRp55 gene) variant rs2235611 A minor allele and AA genotype was significantly associated with SSc-ILD (A allele: p=0.046; AA genotype: p=0.007). Further significant association of SSc with the AA genotype of SRSF6 included association with the late nailfold capillaroscopy (NVC) pattern changes. These observations implicated SRSF6 in the pathogenesis of SSc, and suggested that the SRSF6 rs2235611 AA genotype significantly influenced susceptibility to SSc, and specifically the presence of SSc-related ILD and late NVC pattern [35]. Further in-depth studies on the SRSF6 gene locus will hopefully provide better understanding of genetic predisposition to major SSc-related manifestations such as pulmonary fibrosis and peripheral microvasculopathy. Future work is necessary to confirm the role of tenascin-C alternative mRNA splicing in progression of organ fibrosis in SSc and other pathological conditions. It will also be crucial to understand how SRSF6 is regulated during fibrosis and whether other mRNA splicing factors contribute to these processes as well. These observations also raise the possibility of targeting of SRSF6 to reduce the generation of alternately spliced “pathological” TNC isoforms that contribute to fibrosis as a potential therapeutic strategy in SSc. Indeed, small molecule inhibitors of SRSF6 function are currently under development for the treatment of cancers.

Figure 2:

The mRNA splicing factor SRSF6 is upregulated in SSc. A. Elevated SRSF6 mRNA in SSc skin biopsies compared to healthy controls (from public transcriptome datasets GSE56038 and GSE59785). B. Elevated SRSF6 mRNA in SSc (n=11) and healthy control (n=4) skin biopsies (qRT-PCR results, relative to GAPDH, are the means ± SD of duplicate determinations). C. Elevated SRSF6 in SSc (n=8) and healthy control (n=5) skin biopsies. Representative immunohistochemistry images. Brown nuclear staining indicates SRSF6-positive cells. Hematoxylin counterstain. a,c-healthy control; b,d,e-SSc biopsies. D. Quantitation of SRSF6. Each triangle represents the number (mean) or percent of immunopositive cells from four separate areas/hpf per biopsy. Bars = 20 and 10 μm. Mann-Whitney U test. E. Correlation of tenascin-C-FL levels [20] with the number of SRSF6+ cells in the papillary dermis. Spearman’s correlation.

Fibronectin-EDA

Fibronectin (FN) is a ubiquitously expressed high-molecular weight glycoprotein that can exist in circulation as soluble plasma form, and in tissues as an insoluble ECM component [36]. Fn-EDA is alternatively spliced isoform of FN that is involved in wound healing and fibrosis [37]. Initial findings demonstrated that the alternately spliced fibronectin variant Fn-EDA contributes to, and indeed is required for, inducing myofibroblast differentiation by TGF-β [38]. In studies to investigate the role of Fn-EDA in idiopathic pulmonary fibrosis (IPF), elevated Fn-EDA and α-SMA expression was noted in IPF lung fibroblasts compared to control fibroblasts. Furthermore, Fn-EDA null mice failed to develop pulmonary fibrosis upon bleomycin administration [39].

Following up on these earlier studies, we have demonstrated significant increase in Fn-EDA levels in SSc. Specifically, we showed that circulating, and tissue mRNA and protein levels of Fn-EDA were elevated in SSc compared to control subjects. Moreover, Fn-EDA null mice challenged with subcutaneous bleomycin injections showed reduced dermal thickness and collagen deposition, attenuated expression of α-SMA and phospho-Smad2 compared to wild type mice [19]. Furthermore, in skin fibroblasts, TLR4-specific inhibitor or small interfering RNA (siRNA), and TLR4-mutant skin fibroblasts attenuated Fn-EDA-induced stimulation of collagen deposition and myofibroblast differentiation [19]. Another group of researchers found that Fn-EDA, and its partially unfolded type III domain (FnIII-1c), induced TLR4-dependent inflammatory signaling in fibroblasts [40]. This domain becomes uncovered to rigid fibrotic microenvironments when exposed to tensional forces [40]. We demonstrated that there was a close association between expression of FN-EDA and TLR4 in SSc skin fibroblasts (Fig 3). Our hypothetical model was further corroborated by the discovery via in-silico molecular docking that interacting sequence of Fn-EDA “SPEDGIRELF” formed a stable complex with TLR4-MD2 heterodimer at central and C-terminal domain region of TLR4, where Fn-EDB formed weak complex consistent with our finding [41]. We therefore speculate that in SSc, tensional forces within the stiff tissue milieu of the fibrotic skin drive exposure of the fibronectin EDA and FnIII-1c domains, which, combined with increased EDA isoform generation via alternative mRNA splicing in resident fibroblasts, results in increased Fn-EDA bioavailability and TLR4-mediated profibrotic activity, contributing to progressive fibrosis [19].

Figure 3:

Increased interaction of Fn-EDA with TLR4 interaction in SSc fibroblasts. Explanted skin fibroblasts from SSc patients (n=3) and healthy controls (n=2) were immunostained with antibodies to TLR4, and isoform-specific Fn-EDA. Representative images.

The role of Fn-EDA in fibrosis was further studied in a spinal cord injury mouse model. In injured spinal cords, the levels of Fn-EDA were elevated. Importantly, we showed that Fn-EDA was required to stabilize the fibrotic scars in spinal cord, and Fn-EDA-null mice were protected from chronic scarring [42]. Fn-EDA secreted from hepatic stellate cells was positively correlated with intrahepatic angiogenesis during the progression of hepatic fibrosis by activating VEGFR2 with co-receptors integrin and CD63. Further, inhibition of Fn-EDA expression in mice was correlated with reduced intrahepatic angiogenesis and fibrosis [43].

Substrate stiffness has intense impact on fibrogenesis [44]. In this context, only Fn-EDA efficiently accelerates recruitment of latent TGF-β-binding protein-1 (LTBP-1) within the matrix but not FN-EDB, consistent with our previous finding [19]. In skin fibroblast cultures, matrix stiffness resulted in elevated expression and co-localization of LTBP-1 and FN-EDA, thus immobilizes LTBP-1 and retains TGF-β1 in the ECM, thus promoting fibrosis [45]. Increased Fn-EDA levels were detected in plasma and bone marrow biopsies of primary myelofibrosis patients compared to healthy subjects [18]. Furthermore, EDA overexpressed mouse developed fibrosis while the KO mice failed to develop fibrosis [18]. Fn-EDA also regulates inflammation in stromal cells. In dermal fibroblast, type III repeat of fibronectin induced the expression of cytokines indicating activation of inflammatory pathways [46]. Another study by Kelsh et al demonstrated that Fn-EDA domain and the partially unfolded III-1 domain, FnIII-1c, exhibits synergism by activating pro-inflammatory cytokines in human dermal fibroblasts [47]. These findings implicate a potential pathogenic role of Fn-EDA in fibrogenesis.

Splicing factors involved in the generation of Fn-EDA

Fibronectin is alternatively spliced in multiple disease conditions by having increased alternative exons, EIIIA and EIIIB. One such splicing variant, Fn-EDA is found to be over expressed in psoriatic non-lesional epidermis and triggers keratinocytes to downstream mitogenic effects. A cDNA microarray experiment performed by Szlavicz et al revealed three serine/arginine rich splicing factors including splicing factor arginine/serine-rich 18 (SFRS18), peptidyl-prolyl cis–trans isomerase G (PPIG), and luc-7 like protein 3 (LUC7L3) might regulate fibronectin mRNA maturation in psoriatic non-involved skin. The team also identified double silencing of LUC7L3 and SFRS18 efficiently reduced the production of the psoriasis-associated FN-EDA isoform [48]. The authors further investigated the mechanism regulated by double silencing of LUC7L3 and SFRS18 in HPV-immortalized keratinocyte cells using paired-end RNA sequencing. Interestingly, they found promising change in IFI6 gene, an interferon- inducible gene and shown that the expression of associated partners of IFI6 along with other critical psoriasis-associated transcripts were significantly higher upon double silencing [49].

Another study showed serine/arginine-rich protein serine/arginine-rich splicing factor 1 (SRSF1) as one of the regulatory factors of Fn-EDA isoform. SRSF1 expression is high in endometrium and SRSF1 regulated higher inclusion rate of exon Fn-EDA in endometrium as evidenced from RNA interference, studies. They proposed that endometrium expressed increased Fn-EDA splicing isoform mediated by SRSF1, that resulted in stronger matrix incorporation capacity [50]. By contrast, Fn-EDB, not Fn-EDA, overexpression was associated with breast cancer invasion that was regulated by SRSF6 [51]. In a study based on tumor invasive potential using TGF-β-treated and drug-resistant MCF7 and MDA-MB-468 cells, the knockdown of SRSF6 resulted in Fn-EDB downregulation and reduction in the invasive potential of cancer cells, confirming the role of SRSF6 in Fn-EDB regulation. The role of SRSF6 on Fn-EDA expression needs to be investigated [51].

Other ECM-derived endogenous TLR ligand DAMPs and TLR activation

TLRs are the best known PRR involved in driving inflammation and fibrosis, and as majority of the DAMPs signals through TLRs, DAMPS are thus considered as endogenous TLR ligands in pathogenesis of various immune and inflammatory disorders [52]. Some of the proteoglycans also function via TLR to regulate inflammation and ECM remodeling besides the glycoproteins we discussed [18-20, 53]. For example, biglycan activates inflammation via TLR 2/4 along with their co-receptor CD14 in macrophages. Further inhibition of CD14 reduced biglycan-mediated inflammatory responses [54]. Moreover, biglycan elicits inflammatory responses in human chondrocytes through TLR4 signaling [55]. It has been proposed that decorin and biglycan in their native forms stimulated TLR4 responses in burn wounds suggesting involvement of DAMPs-TLR4 interaction in hypertrophic scaring [56]. Interestingly, altered decorin from burn wounds can cause sustained stimulation of TLR4 thus contributing to hypertrophic scarring [11]. On the other hand, decorin showed anti-fibrotic effect by attenuating TGF-β-mediated fibronectin deposition [57]. Another ECM proteoglycan, versican can activate macrophages via TLR2 and TLR6 signaling [58]. Low molecular weight hyaluronic acid, generated by hyaluronidases, can also function as pathogenic DAMP [59]. In a study with obese individuals, low molecular weight hyaluronic acid elicited inflammatory leukocyte responses and cytokine release via TLR2 signaling [60]. Another DAMP, citrullinated vimentin, has been linked to the pathogenesis of CCl4-induced liver fibrosis in mice [61] and interstitial lung disease associated with RA (RA-ILD) [62]. It was demonstrated that carbon black/cadmium-induced citrullinated vimentin acted as a DAMP to activate fibroblasts through TLR4 signaling to induce lung fibrosis [63].

Future Directions

As summarized above, growing evidence supports a previously unanticipated important role of DAMP-TLR4 signaling in the development, progression and persistence of fibrosis in SSc and other fibrotic conditions (Fig. 1A). Further, the vicious cycle of DAMPS, immune signaling, TGF and Wnt in fibrosis activation is shown in Fig 1C adapted from a review with permission (Fig 1C). However, a thorough understanding of pathogenic DAMPs, cell-type-specific signaling pathways and mechanisms regulated by various DAMPS in fibrosis is still lacking. It will be crucial to sort out if targeting selective TLR4 activation by a particular DAMP is effective in SSc, or if it is more appropriate to target shared intracellular pathways among multiple DAMPs.

As several studies demonstrate the efficacy of small molecule TLR4 inhibitors in ameliorating fibrosis, there is a strong case to be made that these molecules could be considered for clinical trials to determine if they can disrupt the vicious cycle in fibrosis amplification [27, 52]. However, blocking TLR may lead to inappropriate immune responses in specific cell types or in response to certain injuries, as well as compromising anti-microbial host defenses. One potential therapeutic approach to treat fibrosis might be to selectively target disease-associated DAMPs by blocking alternatively-spliced domains of fibronectin and TNC using specific antibodies. For instance, the 81C6 antibodies selectively recognize the D domain of TNC, whereas the F16 antibody targets the A1 domain; the Fn-EDA domain is the target of F8 antibody [64]. Another interesting potential strategy might be to selectively target the FBG domain of TNC, which is responsible for binding to and eliciting TLR4 signaling [65]. Additionally, one could target the activity of mRNA splicing factors, such as SRSF6, that are responsible for generating pathogenic DAMP isoforms, an approach already being investigated in the treatment of cancer. Further preclinical and clinical studies examining the therapeutic potential of targeting of the TLR4-DAMPs are clearly warranted. At the same time, comparative analysis of transcription factors, co-receptors and downstream kinases implicated in DAMP signaling could also have an advantage to novel therapeutic approaches by selectively suppressing DAMP-TLR4 signaling while preserving intact anti-microbial host responses. Furthermore, DAMPs are required for protein folding in TLR trafficking to elucidate cellular responses. As TLR folding in turn is regulated by chaperone proteins including gp96, CNPY3 and CNPY4, it will be important to understand how ECM DAMPs regulates these chaperone proteins, so that pathogenic roles of chaperones in SSc can be discovered, which progressively aids in discovering novel therapeutic strategies to mitigate fibrosis in SSc. In brief, this review highlighted emerging information implicating TLR4-DAMP signaling in sustained fibroblast activation and progressive fibrosis in SSc, and appealing aspects for targeting DAMPs and their cellular receptors for therapeutic intervention to treat SSc and related disorders.