Introduction
The complex pathogenesis of systemic sclerosis (SSc) is dominated by progressive fibrotic replacement of normal tissue architecture. Research in the past few years has elucidated many of the important cellular and molecular mechanisms and mediators of pathological fibrogenesis and identified a fundamental role for transforming growth factor 
(TGF-) in the process.1 This cytokine promotes fibroblast proliferation, differentiation, migration, adhesion, and survival, induces cytokine secretion, and, most importantly, upregulates the synthesis of collagen and extracellular matrix.2 In light of its key role in the pathogenesis of SSc, TGF- has emerged as an attractive therapeutic target. Multiple strategies for blocking the TGF- pathways exist (Table 1) and are currently under investigation.3 Biologic therapies that use antibodies to neutralize a pathogenetic ligand have proven to be highly effective in inflammatory conditions, such as rheumatoid arthritis. Small molecules that can be administered orally, however, might interrupt selected TGF- responses without affecting the important physiological functions of this multifunctional cytokine. Although none of these anti-TGF- therapies has yet reached the clinic, many clinical trials for various indications are ongoing. This Review summarizes the biology of TGF- in the context of fibrosis and the strategies for its inhibition, and highlights progress towards the development of anti-TGF- therapies for the treatment of SSc.
![Table 1 - Proposed therapeutic strategies for blocking the TGF-|[beta]| pathway](/common/images/table_thumb.gif)
TGF-
in health and disease
TGF- has important homeostatic roles in the control of wound healing and tissue repair, epithelial integrity, and innate and adaptive immune responses.4 Aberrant TGF- regulation is associated with inherited conditions, such as hereditary hemorrhagic telangiectasia, Loeys–Dietz syndrome, familial pulmonary hypertension, Camurati–Engelmann disease, Marfan syndrome and fibrodysplasia ossificans progressiva, cancers, both hereditary (e.g. juvenile polyposis and Cowden syndrome) and sporadic (breast, colon, lung and pancreas), and fibrosing disorders such as post-angioplasty restenosis, pulmonary fibrosis, glomerulosclerosis and SSc.5 Reduced TGF- signaling resulting from decreased expression of the type I TGF- receptor confers a substantially increased risk of colorectal cancer, first in mice and then in humans.6, 7 The functional duality of TGF- was illustrated in a mouse model of autoimmunity, where it was shown to be necessary for maintaining immune tolerance while promoting tissue fibrosis.8 These conditions demonstrate that either insufficient or excessive TGF- activity is harmful; therefore, therapeutic targeting of TGF- must consider the impact of TGF- blockade on physiological as well as pathological processes.
TGF- and SSc
Excessive TGF- activity is a common feature of a number of fibrotic conditions of diverse etiologies, making this group of disorders potential candidates for anti-TGF- therapies.9 The link between aberrant TGF- signaling and pathological fibrosis is particularly compelling in SSc. Mice with gain-of-function mutations in the TGF- pathway develop progressive fibrosis in multiple organs.10, 11 A subset of patients with diffuse cutaneous SSc display a 'TGF- responsive gene signature' on gene expression profiling of the lesional skin.12, 13 These patients have been suggested to have more severe disease (higher Rodnan skin scores and greater risk of lung involvement) than those without this pattern of gene expression (J Sargent et al., unpublished data). Gene expression profiling with microarray analysis could be used in the future to identify patients with a TGF- signature, who would be more likely to respond to therapy than those without this pattern.
Précis: TGF- signaling axis
TGF- is generally secreted from monocytes, lymphocytes and fibroblasts as a biologically inactive precursor protein. Activation of latent TGF-, a critical regulatory step in TGF- signaling, is catalyzed by serine proteases and thrombospondin, as well as cell-surface integrins.14 The extracellular matrix, a major TGF- depot, sequesters the cytokine in a latent form. Intracellular signal transduction is initiated through sequential activation of the TGF- receptor complex and downstream intermediates (Figure 1). The canonical SMAD pathway is uniquely associated with TGF- signaling and is deregulated in SSc.15 Several non-SMAD signaling pathways that are shared by multiple cytokines and growth factors are also implicated in translating TGF- signal into relevant biological responses.16
Figure 1 | Major components of the TGF-
signaling pathway.
![Figure 1 : Major components of the TGF-|[beta]| signaling pathway. Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, or to obtain a text description, please contact npg@nature.com](/nrrheum/journal/v5/n4/thumbs/nrrheum.2009.26-f1.jpg)
TGF- secreted from monocytes, macrophages, lymphocytes and fibroblasts is sequestered in the extracellular matrix in a biologically inactive, latent form. Latent TGF- activation is catalyzed by 
v6 integrin on epithelial cell membranes. Active TGF- binds to serine/threonine kinase cell surface receptors that phosphorylate downstream SMAD2/3. Phosphorylated SMAD2/3 complexes with SMAD4 and accumulates within the nucleus, where it collaborates with other transcription factors, and recruits cofactors, such as p300, to target genes, resulting in transcription. SMAD7 is a TGF--inducible endogenous SMAD inhibitor that negatively regulates TGF- signaling. TGF- can also induce cellular responses, such as EMT, via SMAD-independent pathways involving the kinases JNK, p38, PI3K, c-Abl and TAK1. The duration of the TGF- signal is regulated by the uptake of the TGF- receptor–ligand complex into caveolin-lined endosomes that promote degradation. Excessive TGF- activity, or deregulated receptor trafficking or intracellular SMAD signaling result in collagen overproduction and fibrosis. Abbreviations: c-Abl, c-Abelson; EMT, epithelial–mesenchymal transition; JNK, Jun N-terminal kinase; PI3K, phosphoinositide 3 kinase; TAK1, TGF--activated kinase 1; TGF-, transforming growth factor .
The intensity of a TGF- response is modulated by innate control mechanisms.17 Endogenous negative regulators of TGF- include SMAD7, the nuclear phosphatase PPM1A (protein phosphatase 1A [formerly 2C], magnesium-dependent, isoform ),18 and MAN1 (also known as LEMD3), an inner nuclear membrane protein that inhibits TGF- signaling by sequestering SMAD2 and SMAD3 (receptor-regulated SMADs) inside the nucleus.19 Internalization of the activated TGF- receptor complex by caveolin-1-associated membrane lipid rafts leads to intracellular degradation of the receptor–ligand complex and cessation of TGF- signaling.20 Changes in caveolin-1 expression or function result in perturbed TGF- activity; receptor internalization, therefore, represents an important regulatory mechanism. Studies in patients with SSc have revealed a marked decrease in caveolin-1 expression in the skin and lungs, suggesting that reduced caveolin-1 could be responsible for amplification of TGF- signaling contributing to progressive fibrosis in SSc.21, 22
Blockade of the ligand
Strategies to block the production and biological activity of TGF- include neutralizing antibodies, soluble receptors, antisense oligonucleotides and RNA interference (Figure 2). Neutralizing antibodies to TGF- have been used in animal models to prevent organ fibrosis.23, 24, 25 In the first clinical trial of neutralizing antibodies against TGF-, the human monoclonal antibody metelimumab (also known as CAT-192) was compared with placebo in 45 patients with early SSc.26 The antibody was given by intravenous infusion at baseline and at weeks 6, 12 and 18, and patients were evaluated at 24 weeks. The trial showed no improvements in skin scores and other disease manifestations in patients treated with CAT-192. Adverse effects were more frequent in patients receiving CAT-192, but did not reflect increased autoimmunity or other spontaneous immune activation. Limitations of the study included the restricted isotype specificity of the antibody and its low binding affinity, the short treatment duration and small number of patients.
Figure 2 | Strategies for blocking TGF- pathways.
![Figure 2 : Strategies for blocking TGF-|[beta]| pathways. Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, or to obtain a text description, please contact npg@nature.com](/nrrheum/journal/v5/n4/thumbs/nrrheum.2009.26-f2.jpg)
Biologic therapies are shown in pink boxes; small-molecule therapies are shown in yellow boxes. Neutralizing antibodies to v6 integrin inhibit latent TGF- activation only at sites of injury where v6 integrin is expressed. Neutralizing antibodies to TGF- sequester the active ligand, and soluble receptor peptides prevent its binding to cell surface receptors. Small-molecule kinase inhibitors of type I TGF- receptor block SMAD2/3 activation and abrogate SMAD-dependent profibrotic responses, such as collagen synthesis and myofibroblast transformation. Tyrosine kinase inhibitors, such as imatinib mesylate, block SMAD-independent intracellular TGF- signaling. Abbreviations: ALK5, activin-like kinase 5; EMT, epithelial–mesenchymal transition; TGF-, transforming growth factor .
A monoclonal neutralizing antibody to TGF-2 (lerdelimumab, or CAT-152) has been tested for the prevention of scarring following glaucoma surgery.27 A pan-specific monoclonal antibody (GC1008) targeting all three TGF- isoforms is currently in a phase I trial for treating idiopathic pulmonary fibrosis.28 Disrupting TGF- receptor activation by sequestering the ligand has been shown to prevent fibrosis in animal models. For instance, a soluble type III TGF- receptor (TGFBR3) prevented diabetic glomerulosclerosis,29 and a topically administered small peptide fragment of TGFBR3 prevented bleomycin-induced SSc.30
Targets for blockade of TGF- signaling
Integrin function
A novel strategy for therapeutic disruption of TGF- involves blocking the activation of matrix-bound latent TGF- by targeting the v6 integrin, which is expressed on epithelial cells. This membrane integrin catalyzes the activation of latent TGF- in the local microenvironment.31 Mice with targeted deletion of v6 integrin developed spontaneous lung inflammation, but were protected from bleomycin-induced fibrosis.32 Studies have demonstrated that antibodies to v6 integrin blocked latent TGF- activation and prevented the development of lung fibrosis induced by intratracheal bleomycin or radiation.33, 34 A theoretical advantage of targeting v6 integrin would be that such an approach would not interfere with homeostatic TGF- functions, as TGF- activation is blocked only at sites of injury where v6 integrin is induced.35, 36
TGF-
receptor activity
Small molecules that bind to the ATP-binding domain of the serine/threonine kinase TR1 prevent ligand-induced SMAD2/3 phosphorylation and consequent fibrotic responses in vitro,37, 38 ameliorate experimental fibrosis in the kidneys, liver, blood vessels, and lungs, and prevent diabetic nephropathy in the db/db mouse model.39, 40, 41, 42, 43 Whether orally available TR1 antagonists will ultimately be investigated in clinical trials for the treatment of SSc remains uncertain. The cross-reactivity characteristically associated with kinase inhibitors raises concern that TGFBR1 inhibition could disrupt multiple TGF- pathways, in addition to off-target effects unrelated to TGF- signaling, resulting in undesirable effects on immune regulation, cancer surveillance and wound healing.
SMAD intracellular signal transduction
Strategies to disrupt intracellular SMAD signaling include endogenous SMAD inhibitors, SMAD sequestration or targeting degradation (Figure 2).17, 44 Gene transfer of inhibitory SMAD7 ameliorated pulmonary, renal and peritoneal fibrosis and vitreous retinopathy in animal models.45, 46, 47, 48 Hepatocyte growth factor, also known as scatter factor, is a naturally occurring anti-fibrotic cytokine that works in part by inducing inhibitory SMAD7.49 A synthetic analog of hepatocyte growth factor, BB3, is undergoing preclinical evaluation for the treatment of hepatic fibrosis.50 Some of the antifibrotic activities of interferon 
might also be attributed to endogenous SMAD7.51 Paclitaxel, a cancer drug that works by stabilizing the microtubules, attenuated constitutive SMAD2 activation in skin grafts of SSc patients xenotransplanted onto mice with severe combined immunodeficiency (SCID).52 An intriguing novel anti-TGF- strategy uses peptide aptamers to selectively block intracellular signal transduction. Introduction of thioredoxin-A SARA aptamers into mammalian cells blocked epithelial–mesenchymal transformation and related TGF- responses without generally inhibiting SMAD-dependent signaling.53
Non-SMAD intracellular signal transduction
The panoply of non-SMAD signal transduction pathways downstream of TGF- (Figure 1) provides multiple opportunities for TGF--targeting therapies. An activating mutation of the protein tyrosine kinase c-Abelson (c-Abl), a member of the Src family, underlies the pathogenesis of chronic myelogenous leukemia (CML). In recent studies, c-Abl was shown to be activated by TGF- in fibroblasts, and to mediate some of the profibrotic effects independent of SMAD signaling.54, 55 Moreover, c-Abl was found to be constitutively phosphorylated in the lesional skin of patients with SSc.56, 57
Imatinib mesylate, a selective protein tyrosine kinase inhibitor active against oncogenic Bcr-Abl, as well as platelet-derived growth factor (PDGF) receptor and c-kit, is now a widely used and highly effective oral therapy for CML.58 Imatinib has been shown to block the induction of c-Abl activity and fibrotic gene responses elicited by TGF-, and normalized collagen overproduction in explanted SSc fibroblasts.54, 59, 60 The anti-TGF- effects of imatinib are associated with blockade of the activation of SMAD1 and early growth response protein 1.56, 61 Although imatinib prevented fibrosis in the lung, kidney and skin of mice,59, 62, 63 it did not arrest the progression of established lung fibrosis in animal models.62, 64 The antifibrotic effects of imatinib might, therefore, be more prophylactic than therapeutic.59, 61
Anecdotal reports suggest that imatinib is effective in ameliorating skin fibrosis in some patients with SSc, and in fibrosing conditions, such as chronic graft-versus-host disease, nephrogenic fibrosing syndrome and localized forms of SSc.65, 66, 67, 68, 69, 70 An intriguing study in mice demonstrated that bleomycin-induced lung injury was associated with a marked increase in serum and tissue levels of the acute phase reactant 1-acid glycoprotein (AGP), a protein known to neutralize imatinib.71 Furthermore, levels of AGP were shown to be elevated in patients with pulmonary fibrosis,71 and in the serum and bronchoalveolar lavage fluid of patients with SSc.72, 73 These observations suggest that AGP mediates drug resistance, and its induction by tissue injury could account for the failure of imatinib to ameliorate fibrosis.
The therapeutic efficacy of imatinib as an antifibrotic agent is currently under evaluation in clinical trials. Preliminary results from a placebo-controlled multicenter clinical trial of imatinib in idiopathic pulmonary fibrosis do not indicate a notable treatment advantage (C Daniels, personal communication). The lack of detectable clinical response to imatinib in this trial could have been a result of either the incomplete blockade of TGF- signaling achieved with the drug doses employed, or the accumulation of AGP, which neutralizes the effects of imatinib, as discussed above. Observations in patients with CML and other forms of malignancies have shown that imatinib is generally well tolerated, although mild adverse effects are common.74 Of potential importance for SSc, imatinib was shown to ameliorate pulmonary arterial hypertension, a frequent complication of SSc, in mouse models of pulmonary hypertension, individual case reports, and a small phase II clinical trial.75, 76, 77, 78 Dasatanib, a novel tyrosine kinase that is active against multiple members of the Src family of kinases, as well as c-Abl and the PDGF receptor, is in early-stage clinical trials for SSc.79
In light of their relative tolerability, ease of administration, favorable pharmacokinetics, and antifibrotic activity predicted from in vitro experiments and animal studies, protein tyrosine kinase inhibitors are promising candidates for the treatment of various forms of fibrosis and SSc.65 Clinical experience with their use in fibrotic conditions, however, is limited, and their safety profile in this setting remains unknown. Currently, therefore, we do not recommend that patients with SSc be treated with imatinib and other protein tyrosine kinase inhibitors off label, but would encourage such patients to enroll in randomized clinical trials.
TGF--induced stimulation of collagen synthesis involves chromatin remodeling, which is mediated through the recruitment of histone acetyltransferases, such as p300. Accumulation of p300 on a specific gene is associated with locus-specific hyperacetylation of histone H4, resulting in enhanced gene transcription (AK Ghosh et al., personal communication). The histone deacetylase inhibitor trichostatin A, which is used for the treatment of prostate cancer, blocked in vitro the stimulatory effects of TGF- on collagen gene expression in cultured normal skin fibroblasts, and normalized the activated phenotype of SSc fibroblasts.81, 82, 83 These findings suggest that pharmacological modulation of histone activity could be a novel strategy in the treatment of fibrosis.
Concerns about interfering with TGF- biology
Given the exceptionally broad range of biological activities ascribed to TGF- and its fundamental physiological roles, nonselective TGF- blockade could have undesired consequences. Complete abrogation of TGF- signaling could lead to loss of immune tolerance with uncontrolled activation of T and B cells and inhibition of regulatory T-cell (CD4+CD25+) function, resulting in inflammation and spontaneous autoimmunity. Indeed, upregulated immunity induced by TGF- blockade could be desirable in cancer therapy.84 Interestingly, spontaneous autoimmunity has not been observed in preclinical studies of anti-TGF- antibodies or soluble receptors.85, 86 Even in lupus-prone NZB 
NZW mice, anti-TGF- antibody did not exacerbate autoimmunity.8 As neutralizing antibodies, soluble receptors and natural antagonists achieve only partial TGF- deficiency, they might interfere with excess TGF- activity without altering homeostatic TGF- signaling or abrogating pathological TGF- responses, such as fibrosis, while preserving homeostatic functions. Long-term observation of TGF- blockade in clinical trials will be required to validate this concept.
Perspectives on anti-TGF- therapies for SSc
Perturbed TGF- expression and function is a fundamental abnormality underlying the pathogenesis of distinct fibrosing disorders, and TGF- is the molecular target of choice for antifibrotic therapy. Multiple platforms for blocking TGF- exist in the pipeline. The pleiotropic biological activities of TGF- regulate both physiological and pathological processes, and can be beneficial or detrimental. Selectivity of TGF- inhibition could be achieved spatially by, for instance, blockade of TGF- activation only at sites where integrin v6 is expressed; or by selective inhibition of deleterious effects by blockade of target gene-specific coactivator interactions using aptamers. Both biologic therapies and orally administered small molecules that block TGF- are currently undergoing investigation.
In addition to statins and kinase inhibitors, a number of other drugs that are currently in clinical use have shown anti-TGF- activity (Table 2). The angiotensin II receptor type 1 blocker losartan antagonizes TGF- signaling through inhibition of the renin–angiotensin axis. In a mouse model of mutant fibrillin-1 that resembles Marfan syndrome, losartan abrogated the activation of TGF- signaling and restored normal tissue architecture.87, 88 The anticancer drug paclitaxel attenuated SMAD activation in human SSc-affected skin grafts transplanted onto SCID mice in vivo,52 but has been linked to SSc-like reactions in some cancer patients, and remains to be evaluated in patients with SSc.89, 90 Insulin-sensitizing drugs, such as rosiglitazone and pioglitazone, which are widely used in the treatment of type 2 diabetes, activate peroxisome proliferator-activated receptor (PPAR). These drugs inhibit TGF--induced fibrotic responses and SMAD-mediated transcription, in part by targeting early growth response protein 1 and blocking the recruitment of the p300 histone acetyltransferase coactivator.91, 92 Furthermore, treatment of mice with PPAR ligands, such as rosiglitazone, prevents or attenuates lung fibrosis and bleomycin-induced SSc.93, 94, 95
The endothelin-1-receptor antagonist bosentan, which is used to treat idiopathic and SSc-associated pulmonary hypertension, blocks TGF--mediated fibrotic responses in vitro.96 A randomized clinical trial of bosentan, however, failed to show notable benefit in patients with SSc.97 Tranilast, a synthetic tryptophan metabolite that has long been used in Japan for the treatment of allergic conditions, keloids and hypertrophic scars, was found to inhibit collagen production by cultured fibroblasts and prevent fibrosis in animal models.98 The putative mechanism of action involves reduced expression of the TGF- receptors and blockade of SMAD2 activation.99 The established safety profile of tranilast, together with its combined antifibrotic and anti-inflammatory activities, make this an appealing drug for further clinical development in SSc.
Conclusions
The pleiotropic roles of TGF- in tissue homeostasis suggest that blocking TGF- activity is associated with important toxic effects. To date, however, this outcome has not been supported by data from preclinical studies, possibly because only partial abrogation of TGF- activity can be achieved. The mild toxic effects associated with anti-TGF- interventions provide confidence for pursuing their clinical development for SSc. The selection of appropriate patients who would be most likely to respond, such as those with the 'TGF- signature', elevated levels of circulating TGF- or genetically determined elevations in basal levels of endogenous TGF- signaling, is a critical consideration, as is the optimal dose and route of administration. In addition, the intervention is likely to be most effective when initiated in early-stage disease, when fibrosis is TGF--driven and potentially reversible. Carefully designed long-term clinical trials that incorporate biomarkers of biological response and clinical efficacy will be required for evaluating novel therapies that target the TGF- pathway in SSc. These studies will require optimized study design and intensive collaboration among investigators from multiple centers.
Review criteria
English-language, full-text articles published between 1990 and 2009, were sourced for inclusion in this Review from a PubMed search with the terms "scleroderma", "systemic sclerosis" and "therapy", and from NIH and biotech websites.
