, . 184: 4–6 (1998)

MINI-REVIEW

TGFâ—A ROLE IN SYSTEMIC SCLEROSIS?

. 1, . 2, . . 2 . 1

1Departments of Rheumatology and Pathological Sciences, University of Manchester, Manchester M13 9PT, U.K.2Centre for Rheumatic Diseases, Hope Hospital, Salford, U.K.

SUMMARY

Systemic sclerosis (SSc) is a multisystem connective tissue disorder in which there is progressive fibrosis. Transforming growth factorbeta (TGFâ) has wide-ranging cellular actions. It is a potent chemoattractant for human dermal fibroblasts, from which it may inducesynthesis of collagen, which suggests that it may have a central role to play in the pathogenesis of SSc. This is supported to some extentby in vitro studies. SSc fibroblasts produce more collagens and fibronectin than normal fibroblasts and elevated TIMP levels have beenobserved, all of which could be explained on the basis of TGFâ stimulation of fibroblasts. Some studies have suggested that fibroblastsare the source of TGFâ. However, the serum of patients with SSc is cytotoxic to endothelial cells, which could culminate in TGFâsynthesis by them, with secondary fibroblast stimulation. The role of TGFâ remains elusive, although it would seem an ideal candidateas a mediator of fibrosis in systemic sclerosis. ? 1998 John Wiley & Sons, Ltd.

J. Pathol. 184: 4–6, 1998.

KEY WORDS—transforming growth factor beta; systemic sclerosis; fibrosis; fibroblasts

INTRODUCTION

Systemic sclerosis (SSc) is a multisystem connectivetissue disorder characterized by excessive accumulationof extracellular matrix, resulting in progressive fibrosisand dysfunction of a number of organs. The skin is mostcommonly involved and here there is a progression offibrosis associated with microvascular changes. Thefibrosis is of unknown aetiology, although a number ofcauses have been implicated. These include activation ofatypical fibroblast clones, vascular phenomena, auto-immunity, and environmental toxins. At the molecularlevel, investigators have been pursuing the concept that,as in other ‘fibrotic’ disorders, transforming growthfactor beta (TGFâ) has a central role to play inpathogenesis.TGFâ is one of a family of closely related peptides,

which includes a number of TGFâ isoforms, and thebone and cartilage-derived morphogenetic proteins. Inman, the most commonly encountered isoforms ofTGFâ are TGFâ1, 2, and 3. All are homodimericmolecules, synthesized as larger precursor proteins.Each isoform may be secreted as either a latent or anactive form.1 Activation occurs by dimerization andproteolytic cleavage of the precursor peptide.1 The func-tion of a given activated isoform is largely dependent onthe presence of certain TGFâ receptor subtypes andtheir affinity with a given isoform. Differential receptor

expression allows for a precise response in differingcellular environments.TGFâ has wide-ranging cellular actions, which are

exemplified by its effects on fibroblasts and endothelialcells. It is a potent chemoattractant for human dermalfibroblasts,2 from which it may induce the release ofcytokines such as PDGF3 and the synthesis of collagen.4In contrast to its stimulatory effects on fibroblasts,TGFâ inhibits endothelial cell proliferation in vitro.5

TGFâ IN SYSTEMIC SCLEROSIS

The evidence for a role for TGFâ in SSc is somewhatcontradictory. While some of the inconsistencies in theliterature may reflect the different experimental systemsused to study TGFâ, even similar types of study haveproduced conflicting data. In general, studies of TGFâin SSc tissue fall into two broad categories: in vivostudies using immunohistochemistry (IHC) and in situhybridization (ISH) techniques, and in vitro studiesusing cell culture.

In vivo studies

The relevance of TGFâ to tissue changes occurringin vivo in SSc has largely been determined through IHCand ISH studies of skin. Some IHC studies have shownintensely fibrotic skin to be lacking TGFâ expression,6whereas others have reported increased TGFâ.7 Oneexplanation for these conflicting findings, which we andothers are pursuing, is that TGFâ expression in lesionalskin may be dependent on the stage of local progressionof the disease in the biopsied site.8,9

*Correspondence to: Shirley A. Cotton at the Departments ofRheumatology and Pathological Sciences, University of Manchester,Manchester M13 9PT, U.K.

Contract grant sponsors: Medical Research Council; Arthritis andRheumatism Council.

CCC 0022–3417/98/010004–03 $17.50 Received 2 July 1997? 1998 John Wiley & Sons, Ltd. Accepted 7 July 1997

With IHC, TGFâ expression has been localized in anumber of cell types including fibroblasts, endothelialcells, and macrophages. Endothelial expression ofTGFâ1 is found in capillaries of the dermis and subcu-taneous tissue.10 Macrophage TGFâ1 and 2 expressionhas been localized preferentially to cells around vesselsin SSc skin and other inflammatory skin disorders,6 buta macrophage infiltrate is an inconsistent finding in SScskin.One of the characteristics of TGFâ is that it can bind

to matrix proteins and it can be stored in connectivetissue. In vivo, it is uncertain whether TGFâ is secretedin an active form, or in a latent form which is seques-tered in the extracellular matrix (ECM) to be mobilizedat a later time. We have shown that TGFâ expressionincreases in the endothelium of lesional skin before andduring the early stages of fibrosis and then declines withlocal disease progression,9 implicating changed localsynthesis of TGFâ in the progression of SSc skin lesions.However, TGFâ has been localized to the ECM ofnormal skin. It has also been described in the papillary,mid-, and deep dermis in limited cutaneous SSc, but itappears to be lost in diffuse cutaneous SSc,11 indicatingthat in SSc TGFâ may be mobilized from the matrix andnot replaced. The significance of the ECM as a source ofactive TGFâ in SSc is a matter for further study.

In vitro studies

Much of the evidence suggesting that TGFâ plays arole in the pathogenesis of SSc is based on cell culturemodels. In vitro, SSc fibroblasts produce more type I andIII collagen, glycosaminoglycans, and fibronectin thannormal fibroblasts;12,13 collagenase production is similarto normal fibroblasts,11 although gelatinase levels areincreased. When applied to cultures of normal fibro-blasts, TGFâ increases expression of matrix proteinmRNA.14 At the same time, it increases the synthesis ofinhibitors of matrix metalloproteinases (TIMPs) whichinhibit collagenase activity.15 Elevated TIMP levels havealso been observed in SSc fibroblasts15 and patientsera.16 Thus, the progressive increase in matrix proteinthat characterizes SSc skin could be explained on thebasis of TGFâ stimulation of fibroblasts to producematrix proteins and inhibit their breakdown.It is unlikely that, in vivo, the regulatory role of TGFâ

on collagen gene expression is as simple as in vitroexperiments might suggest. For instance, regulation ofcollagen production by TGFâ in both normal andscleroderma fibroblasts has been shown to be altered bybasic fibroblast growth factor (bFGF), where bFGFinhibits TGFâ-induced collagen á2(I) gene expression.17In addition, it has been proposed that fibrosis could becaused by TGFâ-induced expansion of subpopulationsof fibroblasts with increased matrix-synthesizing capac-ity. In support of the latter, it has been shown thatcontinual TGFâ stimulation in SSc fibroblasts producesupregulation of PDGFá (platelet-derived growth factoralpha) receptor protein and mRNA, which correlatespositively with mitogenic responsiveness to PDGFAA.18This is another growth factor implicated in SSc patho-physiology, which could represent a route by which

TGFâ may act as a selective stimulus for expansion ofspecific subpopulations of fibroblasts.It has also been suggested from in vitro studies that

SSc fibroblasts are hyper-responsive to TGFâ stimula-tion, but demonstrating this in vivo has proved difficult.However, as stimulation by TGFâ causes normal fibro-blasts to adhere to ECM through the expression of â1integrins, it is possible that the increased expression ofintegrins by fibroblasts in SSc could be taken as evidenceof hyper-responsiveness.19Cell culture studies have also helped to elucidate the

source of TGFâ. It has been proposed that autocrinestimulation by TGFâ may occur in SSc fibroblasts.In vitro protein and mRNA expression studies haverevealed that SSc fibroblasts do not secrete more TGFâthan normal cells.20 Alternative sources responsible forTGFâ production have been sought. These studies havefocused on endothelial cells as the interface betweenblood and tissue.Studies of endothelial cells in culture have shown that

serum of patients with SSc is cytotoxic to endothelialcells.21 This has led to the proposal that this initiates asequence of events in the endothelium leading toendothelial cell production of TGFâ, which initiateslocal fibrosis. Our own studies implicate a cascade ofevents mediated by cytokines and free radicals,22 culmi-nating in TGFâ synthesis by the EC and secondaryfibroblast stimulation. This hypothesis is supported bymeasurements made in the blood in patients with SSc.Anti-endothelial antibodies have been identified as asource of endothelial cell damage and elevated levels ofTGFâ have been detected in the plasma of a proportionof patients with SSc, but not in patients with primary‘Raynaud’s or control subjects,23 although it should besaid that one study has failed to identify altered TGFâ inthe serum of patients with SSc.11 This is yet anotherexample of the frustrating contradictions in the litera-ture. There is also evidence for platelet activation inSSc,24,25 again supporting widespread endothelial celldysfunction in this disorder.

CONCLUSION

This review summarizes the body of evidence impli-cating TGFâ in SSc, but identifies inconsistencies thatstill remain to be clarified. These may reflect the hetero-geneity of the disease itself, the biological nature ofTGFâ, the techniques used to study TGFâ productionand effects, and differences in TGFâ expression in skinat different stages of the disease. The role of TGFâremains elusive, although superficially it seems an idealcandidate as a mediator of fibrosis in this disorder.Further studies are urgently required to elucidate therole of this growing family of molecules in the patho-genesis of systemic sclerosis and other fibrotic disorders.

ACKNOWLEDGEMENTS

We acknowledge the support of the Medical ResearchCouncil and the Arthritis and Rheumatism Council forthe research leading to this article.

5TGFâ—A ROLE IN SYSTEMIC SCLEROSIS?

? 1998 John Wiley & Sons, Ltd. , . 184: 4–6 (1998)

REFERENCES1. Taipale J, Miyazana K, Heldin C-H, Keski-Oja J. Latent transforming

growth factor â associates to fibroblast extracellular matrix via latenttransforming growth factor â binding protein. J Cell Biol 1994; 124:171–181.

2. Postlethwaite AE, Keski-Oja J, Moses HL, Kang AH. Stimulation of thechemotactic migration of human fibroblasts by transforming growth factorbeta. J Exp Med 1987; 165: 251–256.

3. Moses HL, Coffey RJ, Loef EB, Lyons RM, Keski-Oja J. Transforminggrowth factor beta regulation of cell proliferation. J Cell Biol 1987;5(Suppl): 1–7.

4. Varga J, Rosenbloom J, Jimanez SA. Transforming growth factor betacauses persistent increase in steady state expression of type I and type IIIcollagen and fibronectin mRNA in normal human dermal fibroblasts.Biochem J 1987; 247: 597–604.

5. Takehara K, LeRoy EC, Grotendorst GR. Transforming growth factorbeta inhibition of endothelial cell proliferation: alteration of EGF bindingand EGF induced growth regulator (competence) gene expression. Cell1987; 49: 415–422.

6. Gruschwitz M, Muller PU, Sepp N, Hofer E, Fontana A, Wick G.Transcription and expression of transforming growth factor beta in skin ofprogressive systemic sclerosis: a mediator of fibrosis? J Invest Dermatol1990; 94: 197–203.

7. Sfikakis PP, McCune B, Tsokos M, Aroni K, Vayiopoulos G, Tsokos GC.Immunohistological demonstration of transforming growth factor betaisoforms in the skin of patients with systemic sclerosis. Clin ImmunolImmunopathol 1993; 69: 199–204.

8. Sato M, Isikawa O, Miyachi Y. An immunohistochemical study on trans-forming growth factor beta in the skin of systemic sclerosis patients withreference to the histopathologic stage. Eur J Dermatol 1996; 6: 482–485.

9. Helman S, Cotton SA, Jayson MIV, Hoyland J, Herrick A, Freemont AJ.Microvascular abnormality and TGFâ in systemic sclerosis skin. Br JRheumatol 1997; 36(Suppl 1): 30(54A).

10. Gabrielli A, Di Loreto C, Taborro R, et al. Immunohistochemical localis-ation of intracellular and extracellular associated TGFâ in the skin ofpatients with systemic sclerosis (scleroderma) and primary Raynaud’sphenomenon. Clin Immunol Immunopathol 1993; 68: 340–349.

11. Higley H, Persichitte K, Chu S, Waegell W, Vancheeswaran R, Black CM.Immunocytochemical localisation and serological detection of transforminggrowth factor â1. Arthritis Rheum 1994; 17: 278–288.

12. Perlish JS, Bashey RI, Stephens RE, Fleischmajer R. Connective tissuesynthesis by cultured scleroderma fibroblasts. Arthritis Rheum 1976; 19:891–901.

13. Uitto J, Bauer EA, Eisen AZ. Scleroderma: increased biosynthesis of triplehelical type I and type III pro-collagens associated with unaltered expres-sion of collagenase by skin fibroblasts in culture. J Clin Invest 1979; 64:921–930.

14. Inagaki Y, Truter S, Ramirez F. Transforming growth factor beta stimu-lates alpha 2(I) collagen gene expression through cis-acting element thatcontains an SP1 binding site. J Cell Biol 1994; 269: 14828–14834.

15. Kikuchi K, Kadono T, Furue M, Tamaki K. Tissue inhibitor of metallo-proteinase 1 (TIMP1) may be an autocrine growth factor in sclerodermafibroblasts. J Invest Dermatol 1997; 281–284.

16. Kukuchi K, Kubo M, Sato S, Fujimoto M, Tamaki K. Serum tissueinhibitor of metalloproteinase in patients with systemic sclerosis. J Am AcadDermatol 1995; 973–978.

17. Ichiki Y, Smith EA, LeRoy EC, Trojanowska M. Basic fibroblast growthfactor inhibits basal and transforming growth factor â induced collagená2(I) gene expression in scleroderma and normal fibroblasts. J Rheumatol1997; 24: 90–95.

18. Yamakage A, Kikuchi K, Smith EA, LeRoy EC, Trojanowska M. Selectiveupregulation of PDGFá receptors by transforming growth factor beta inscleroderma fibroblasts. J Exp Med 1992; 175: 1227–1234.

19. Majewski, Huzelmann Schirren CG, Meuch C, Aumailley M, Krieg T.Increased adhesion of fibroblasts from patients with scleroderma to extra-cellular matrix components: in vitro modulation by IFNã but not by TGFâ.J Invest Dermatol 1992; 98: 86–91.

20. Clark RAF, Nielsen LD, McPherson JM. Collagen matrices depressunstimulated and TGFâ stimulated fibroblast synthesis of collagen. J CellBiol 1987; 105: 212A.

21. Kahleh MB, Sherer GK, LeRoy EC. Endothelial injury in scleroderma.J Exp Med 1979; 149: 1326–1335.

22. Cotton SA, Jayson MIV, Hoyland J, Herrick A, Freemont AJ. Alteredvascular response in systemic sclerosis skin. Br J Rheumatol 1997;36(Suppl 1): 30(53A).

23. Snowden N, Coupes B, Herrick A, Illingworth K, Jayson MIV, BrenchleyPEC. Plasma TGFâ in systemic sclerosis: a cross sectional study. AnnRheum Dis 1994; 53: 763–767.

24. Kahaleh MB, Osborn R, LeRoy EC. Elevated levels of circulating plateletaggregates and beta thromboglobulin in scleroderma. Ann Intern Med 1982;96: 610–613.

25. Herrick AL, Illingworth K, Blann A, Hay CRM, Hollis S, Jayson MIV.Von Willebrand factor, thrombomodulin, thromboxane, â thromboglobulinand markers of fibrinolysis in primary Raynaud’s phenomenon and systemicsclerosis. Ann Rheum Dis 1996; 55: 122–127.

6 S. A. COTTON ET AL.

? 1998 John Wiley & Sons, Ltd. , . 184: 4–6 (1998)