Insights into the role of fibroblasts in human autoimmune diseases

Abstract

Traditional wisdom has considered fibroblasts as contributing to the structural integrity of tissues rather than playing a dynamic role in physiological or pathological processes. It is only recently that they have been recognized as comprising diverse populations of cells exhibiting complex patterns of biosynthetic activity. They represent determinants that react to stimuli and help define tissue remodelling through the expression of molecules imposing constraints on their cellular neighbourhood. Moreover, fibroblasts can initiate the earliest molecular events leading to inflammatory responses. Thus they must now be viewed as active participants in tissue reactivity. In this short review, I will provide an overview of contemporary thought about the contribution of fibroblasts to the pathogenesis of autoimmune processes through their expression of, and responses to, mediators of inflammation and tissue remodelling.

Introduction

The growing realization that fibroblasts participate in numerous normal and pathological processes has thrust these cells into measured notoriety. The traditional view that they represented purely structural elements has subsided gradually and been replaced reluctantly with the acknowledgement that fibroblasts participate in dynamic interplay with other cells. They express a diverse array of immunomodulating factors such as cytokines, lipid mediators and growth factors [1–3]. Moreover, fibroblasts display numerous surface and intracellular receptors and the requisite molecular machinery to respond to extrinsic signals. Thus they might be considered extensions of the ‘professional’ immune system. What now emerges is the viewpoint that fibroblasts can actually initiate inflammation. They are the resident sentinels for reacting to early danger signals and coordinating recruitment of immunocompetent cells from bone marrow [4]. By contributing to the profile of cell recruitment, they play a prominent ‘pacemaker’ role for the ultimate quality and duration of tissue remodelling and wound healing. In this short review, I will attempt to put the unfolding saga of the fibroblast into the perspective of a cell-type that is gaining respect as an important member of the surveillance team that functions to direct tissue maintenance and repair. Two examples of fibroblasts involved in human autoimmune diseases will provide the focus for this discussion. Synovial fibroblasts play an integral role in the pathogenesis of rheumatoid arthritis [5] while those from orbital connective tissue become activated in Graves’ disease [6–8]. Both fibroblast types have become well-characterized and the peculiar phenotypic attributes of each seem to conform to our current notions of systemic autoimmune diseases presenting with anatomically localized manifestations.

General concepts about fibroblasts

It has become clear over the last few years that human fibroblasts exhibit substantial diversity with regard to their structure, functional role and capacity for participating in disease processes. In spite of similar ultrastructure [9] and nutritional requirements, important differences in biosynthetic repertoire, surface marker display, and metabolic activities have emerged. Evidence for regional distinction between fibroblasts can be found in studies dating back at least 40 years. Castor et al. reported that fibroblasts from different tissues, including synovium, skin, periosteum and pleura, displayed divergent patterns of proliferation and glycosaminoglycan production [10]. Schneider and colleagues described increased rates of cell proliferation in fetal lung fibroblasts compared with those from the skin of the same donor [11]. Subsequently, testosterone metabolism was found to differ in fibroblasts from male genitalia compared with cells from either deltoid or abdominal skin [12]. Gingival fibroblasts respond to diphenylhydantoin differently from other cells [13]. Many of the specialized attributes associated with the parental tissues can be recognized in derivative fibroblasts. Dynamic interplay between fibroblasts and their neighbours contributes actively to maintaining adjacent epithelium in a differentiated state. The former have now been shown to be crucial to the neoplastic transformation of the latter [14].

Besides these differences in fibroblasts deriving from distant anatomic regions, those within certain tissues also appear to be heterogeneous. Gabbiani et al. recognized fibroblast heterogeneity when they described cells exhibiting distinct phenotypes within granulation tissue [15]. Cells with characteristic peripheral cytoplasmic fibrils were associated with collagen. We now know these ‘modified’ fibroblasts as myofibroblasts, exhibiting features of both smooth muscle cells and fibroblasts [3]. Subsequently, subpopulations of fibroblasts from various tissue sites could be segregated into discreet subsets displaying cell surface markers [16–18]. Included in the discriminators has been the display of Thy-1, a surface glycoprotein [19,20]. In lung fibroblasts, where two populations can be segregated on the basis of Thy-1 expression, those not expressing the marker (Thy-1⁻) exhibit a higher level of constitutive platelet-derived growth factor α receptor than do their Thy-1⁺ counterparts [21]. Moreover, interleukin (IL)-1β can induce the expression of this receptor only in Thy-1⁻ cells as well as substantially induce IL-6 expression. Thy-1⁻ fibroblasts can also activate latent transforming growth factor (TGF)-1β in response to fibrogenic stimuli while their counterparts fail to do so [22]. Apparently the expression of Thy-1 per se is involved in this response since transfecting Thy-1⁻ cells with wild-type Thy-1 rendered the cells unresponsive [22]. Thy-1 expression also correlates with the production of IL-8, MCP-1 and IL-6 in response to proinflammatory cytokines, as does the use of IL-1 as an intermediate protein in responses to various stimuli [23,24]. These distinctions among fibroblasts have defined important functional attributes. Their recognition helps explain many features of tissue reactivity, such as the evolution from inflammatory states to those of repair, wound healing and fibrosis.

Human orbital and synovial fibroblasts: examples of frequent targets for autoimmune disease

Among human fibroblasts, certain types stand out as exhibiting particular vulnerability to autoimmune disease. Synovial and orbital fibroblasts are frequently involved in a spectrum of these disease processes and therefore should be considered prototypic of highly reactive resident cells found in anatomical regions where connective tissues display particular susceptibility to chronic inflammation.

Synovial fibroblasts participate actively in the pathogenesis of rheumatoid arthritis

Synovial fibroblasts represent a heterogeneous population of cells normally involved in maintaining the articular surfaces against which mechanical movements can occur without structural embarrassment [25]. Cells residing in the healthy, thin synovial membrane comprise two dominant populations. Included are those exhibiting the hallmark phenotypic attributes of fibroblasts (Type B synoviocytes) and macrophage-like cells (Type A synoviocytes), based on the profile of surface markers they display. The membrane is normally two to three cells in thickness and is separated from the stroma by a basement membrane. In states of inflammation, such as those associated with rheumatoid arthritis, and to a lesser degree with osteoarthritis, other cells appear more prominent. These include mast cells and T lymphocytes that become intermixed with various vascular cells and cell types exhibiting morphologies intermediate between the macrophage-like cells and fibroblasts. Synovial fibroblasts exhibit morphologies that are reminiscent of their counterparts in other tissues. They can take on a dendritic appearance, especially when exposed to cAMP-enhancing compounds [26]. Type A cells have a well-developed Golgi apparatus and are vaculated in health. Type B cells also display prominent Golgi and a more fully developed endoplasmic reticulum. They, too, can be deformed into irregular, polarized cells with extensive arborization, depending on the physiological or pathological circumstances into which they become insinuated. In states of inflammation, the numbers of cells comprising the synovial membrane, including both Types A and B fibroblasts, increase. Beside the expanded population of resident cells, bone marrow-derived cells infiltrate the involved joint. Activated T and B lymphocytes and macrophages are trafficked to sites of disease involvement. These express a number of important cytokines that allow immunocompetent cells to cross-talk extensively with the resident cells. Tissue infiltration occurs as a complex process. Synovial fibroblasts participate actively in cell recruitment by expressing powerful chemoattractants, including monocyte chemotactic protein (MCP), IL-8, RANTES and IL-16 [27,28]. Chemoattractant expression occurs as a consequence of cytokine activation or through activation of Toll-like receptors [27,28]. IgGs circulating in patients with rheumatoid arthritis recognize and can also activate synovial fibroblasts through their interactions with the insulin-like growth factor receptor-1 (IGF-1R) [29]. This leads to the expression of IL-16 and RANTES. IL-16 is a powerful chemoattractant not containing chemokine-like cysteine signatures that binds and provokes signalling through CD4 [30]. Signalling in the fibroblasts elicited by IgG involves the FRAP/mTOR/p70^s6k pathway and is blocked by the specific macrolide inhibitor, rapamycin, and by glucocorticoids [29].

Synovial fibroblasts have been studied extensively in vitro. A schematic of the complex interactions thought to occur between Type B synovial fibroblasts and recruited immunocompetent cells appears in Fig. 1. They exhibit a number of responses that may play important roles in the pathogenesis of rheumatoid arthritis. For instance, when treated with tumour necrosis factor (TNF)-α or IL-1β they express high levels of prostaglandin endoperoxide H synthase-2 (PGHS-2), the inflammatory cyclo-oxygenase (COX-2) [31]. Moreover, this cytokine treatment results in an increased level of microsomal prostaglandin E₂ synthase (mPGES), the terminal enzyme in the PGE₂ biosynthetic pathway [32]. Cytokine-treated synovial fibroblasts generate high levels of PGE₂. Because these cytokines and PGHS-2 are important components in the pathogenesis of rheumatoid arthritis, several therapies targeting them have been developed and have proven themselves to be of substantial benefit. The drugs inhibiting PGHS-2 also have been marketed. While they are currently being scrutinized for their troublesome side-effect profiles, these compounds have also demonstrated substantial therapeutic activity [33].

Fig. 1.

Schematic of the phenotypic attributes exhibited by Type B synovial fibroblasts and how they are believed to interact with recruited immunocompetent cells in rheumatoid arthritis.

Molecular underpinnings for the characteristic phenotype of fibroblasts from patients with rheumatoid arthritis have been identified. Key transcriptional factors involved in the up-regulation of synovial fibroblast genes relevant to the pathogenesis of rheumatoid arthritis include NF-κB and AP-1 [34,35]. Moreover, affected tissues exhibit substantial DNA-binding activity attributable to AP-1 [36]. Egr-1, c-jun, c-fos and c-myc appear to be over-expressed in tissues from these patients [37,38]. A number of well-travelled signalling pathways are involved in gene regulation. All three components of the MAP kinase cascade, including ERK, p38 and JNK, are relevant to the induction of several proinflammatory fibroblast genes [39]. Activation of signalling pathways culminates in the expression of matrix metalloproteinases (MMPs), their modulators, the tissue inhibitors of metalloproteinases (TIMPs), as well as several cytokines. These effectors are involved intimately in disease progression that culminates in cartilage and bone destruction. The absent expression of a powerful tumour suppressor molecule, PTEN, coupled with a p53 mutation found in disease-derived fibroblasts, might provide important clues to some of the phenotypic attributes of synovial fibroblasts [40,41]. While the issue of enhanced proliferation in these cells remains unresolved [3], alterations in apoptosis have also been identified in some reports and may involve both resistance to pro-apoptotic signals and failure to initiate apoptotic signalling. These could account at least partially for tissue hypertrophy associated with the disease. However, the details of the mechanisms involved are currently far from established.

Orbital fibroblasts are important participants in Graves' disease

Graves' disease is a syndrome arising from the multiple pathogenic complexities associated with thyroid autoimmunity [42,43]. Thyroid-associated ophthalmopathy (TAO) is an autoimmune disease component of Graves' disease. In this process, orbital connective tissues expand, causing the forward propulsion of the eye or proptosis [44]. This volumetric increase is a result of hydrophilic molecules such as hyaluronan accumulating in orbital tissue and expansion of fat tissue through an as yet unidentified mechanism [2,44–46]. In addition to expanding, tissues of the orbit, including muscle and fat/connective tissue, become infiltrated with bone marrow-derived cells, including T and B lymphocytes and mast cells [47]. It is currently believed that these immunocompetent cells, by virtue of the small molecules they express and release, activate the residential cell population and provoke disease progression. The histopathological hallmarks of TAO can be best understood by examining the unique phenotype and activity of orbital fibroblasts [7,48]. Human orbital fibroblasts have a characteristic morphology [49] that can be deformed, like synovial fibroblasts, by increasing intracellular cAMP levels [50]. When treated with agents that enhance intracellular cAMP levels, a subgroup of these fibroblasts can undergo differentiation into triglyceride-accumulating cells [51–53]. They comprise two functional subsets based on the surface display of Thy-1 [54,55]. Although the function of Thy-1 in mammalian cells remains uncertain, as does the identity of its natural ligand, it serves as a useful marker for segregating human fibroblasts. As Fig. 2 implies, each subset can respond to different extracellular stimuli and differentiate into distinct cell types. Thy-1⁺ fibroblasts can differentiate into myofibroblasts when treated with TGF-β[56]. These cells express high levels of α-smooth muscle-specific actin and are important to wound healing. In contrast, Thy-1⁻ orbital fibroblasts can differentiate into mature adipocytes when treated with PPARγ agonists as well as with cAMP-enhancing agents [55,56]. These differentiated cells possess the phenotypic attributes of mature adipocytes [55]. When treated under similar experimental conditions, dermal fibroblasts, which uniformly express Thy-1 [54,55], fail to differentiate into adipocytes. The biosynthetic capacities of Thy-1⁺ and Thy-1⁻ cell-types differ substantially with regard to their expression of small molecules, including growth factors, cytokines, gangliosides and lipid mediators [23]. Thus, even in their undifferentiated states, these two cell-types appear very different with regard to their potential for orchestrating tissue remodelling and engaging actively in immune responses.

Fig. 2.

Some fibroblasts, like those from the orbit, can differentiate into different cell types. Thy-1⁺ fibroblasts, when treated with transforming growth factor (TGF)-β, can differentiate into myofibroblasts that express high levels of α-smooth muscle specific actin. Thy-1⁻ fibroblasts treated with PPARγ agonists, accumulate triglycerides and exhibit the phenotype of adipocytes.

Orbital fibroblasts exhibit robust responses to several cytokines

Graves' disease is associated with multiple actions of several cytokines that promote thyroidal dysfunction, tissue disruption and inflammation [57]. Inflammation of the connective tissue and extra-ocular muscles is encountered frequently in the active phase of TAO, which persists typically for 12–36 months [44]. Many of the features of the inflammatory response are apparently mediated through actions on orbital fibroblasts by T cell-derived cytokines. An attribute of the orbital fibroblast phenotype, particularly in cells derived from patients with TAO, involves exaggerated responses to proinflammatory cytokines [7]. These are summarized in Fig. 3. As an example, PGHS-2 is dramatically induced by cytokines, including IL-1β, leukoregulin and CD154 [58–61]. The mechanism is complex and involves both the up-regulation of gene promoter activity and enhanced mRNA stability [58,60]. The effects are mediated through a coordinate activation of both p38 and ERK MAP kinases, and interrupting those pathways can attenuate the induction [60]. In response to these cytokines, a coordinate induction of the recently cloned mPGES results in substantial increases in PGE₂ production in orbital fibroblasts [60]. Over-production of PGE₂ has potentially important consequences for immune responses in the inflamed orbit. The prostanoid is known to bias naive T cell differentiation toward Th2 at the expense of Th1 [62,63]. In addition, PGE₂ can alter B cell behaviour, influence class-switching, and modulate mast cell activation [64,65]. Thus, tissues where particularly high levels of this prostanoid are generated might be expected to exhibit a characteristic pattern of tissue remodelling and reactivity based on the actions of this lipid mediator. Fibroblasts express receptors for PGE₂ and can respond to it in a variety of ways, including the elaboration of cAMP [66]. Prostaglandins and the signalling they provoke might therefore be considered autocrine loops in orbital fibroblasts.

Fig. 3.

Schematic of the phenotypic attributes of orbital fibroblasts and how they are thought to interact with immunocompetent cells recruited to the orbit in thyroid-associated ophthalmopathy (TAO).

Another hallmark of TAO, the accumulation of extracellular matrix-associated macromolecules, can also be understood more clearly from examining closely the activities attributed to fibroblasts. Evidence now exists for altered synthesis and degradation by fibroblasts of extracellular matrix components. Orbital fibroblasts synthesize high levels of hyaluronan when activated by IL-1β and other proinflammatory cytokines [67–69]. Levels achieved in these cells are considerably greater than those found in other fibroblast types. In addition, IgGs from patients with Graves’ disease up-regulate the synthesis of hyaluronan [70]. The critical enzymes in the hyaluronan biosynthetic pathway, including hyaluronan synthases 1, 2 and 3, as well as UDP-glucose dehydrogenase, are induced in orbital fibroblasts following cytokine treatment [69,71]. These increased levels of hyaluronan can be attenuated with glucocorticoids, but basal levels of glycosaminoglycan synthesis are uninfluenced by these steroids [49]. In contrast, basal hyaluronan synthesis in extra-orbital fibroblasts, such as those from the skin, can be depressed substantially by glucocorticoids [72]. The activity of degrading extracellular molecules involves the delicate balance between several proteolytic processes. Orbital fibroblasts express plasminogen activator inhibitor-1 (PAI-1) when activated by Th1 cytokines, such as interferon (IFN)-γ[73], as well as by TGF-β[74] and leukoregulin [75]. The levels achieved are considerably greater than those found in dermal fibroblasts [76]. Moreover, orbital cells also express high levels of TIMP-1 when activated [77]. Of particular note is the impact of both Th1 and Th2 cytokines in down-regulating the induction of TIMP-1 by IL-1β[77]. Thus, the pericellular microenvironment surrounding these cells is enriched with molecules that retard extracellular matrix degradation. In states of inflammation where proinflammatory cytokines are likely to become over-expressed, the orbital fibroblast should contribute factors favouring the accumulation of macromolecules.

Adhesion molecules play an important role in the way in which cells orientate themselves to their neighbours. They have been detected in situ in orbital connective tissue from patients with TAO [78]. Moreover, soluble adhesion molecules have been detected in the sera of patients with TAO and the levels appear to respond to glucocorticoid therapy [79]. When orbital fibroblasts are treated with cytokines including IL-1α, TNF-α and IFN-γ, they express high levels of intercellular adhesion molecule-1 (ICAM-1, CD54) [80]. This occurs in fibroblasts from patients with TAO and those without the disease. In contrast, only fibroblasts from individuals with Graves’ disease responded to protein A purified IgGs from patients with the disease [80]. Thus these cells are able to express molecules that should enhance productive interactions with other cells, including those recruited to the orbit from the bone marrow.

Searching for the common link between this set of over-determined cellular responses in orbital fibroblasts has led to a potential explanation for at least some of them. A key clue derived from the earlier findings that responses in these cells to leukoregulin, a T cell-derived cytokine [81], and CD154 are mediated through an intermediate induction of IL-1 [59]. The IL-1 family of cytokines includes a molecule, termed IL-1 receptor antagonist (IL-1RA). This polypeptide binds to the IL-1 receptor but unlike IL-α and IL-β, fails to activate it or to initiate down-stream signalling. It is thus a ‘molecular brake’ for the activating members of this cytokine family. Cao et al. reported that while extra-orbital fibroblasts express high levels of IL-1RA following treatment with IL-1, orbital fibroblasts failed to exhibit a comparable response [61]. Muhlberg et al. found discrepancies between the levels of IL-1RA in control orbital fibroblasts and those from patients with TAO [82]. When IL-1RA was introduced, either by adding the recombinant protein to the culture medium or by transfecting orbital fibroblasts with a plasmid containing IL-1RA cDNA, the exaggerated responses were greatly attenuated to levels similar to those in control dermal fibroblasts [61]. Thus it would appear that orbital fibroblasts, especially those from patients with TAO, exhibit a distinctive phenotype that might represent the basis for the extraordinary tissue reactivity found in Graves’ disease. The companion process, dermopathy, occurs with considerably less frequency than does TAO. The reasons for this disparity are uncertain. It is possible that an inflammatory phenotype among pretibial fibroblasts is relatively rare, accounting for the lower incidence of skin involvement.

Is the thyrotropin receptor involved in the development of TAO?

A relatively recent finding with obvious implications for TAO concerns the expression of the thyrotropin receptor (TSHR) in tissues other than the thyroid gland. Orbital connective tissues have been the focus of several studies over the past dozen years. Feliciello and colleagues detected TSHR mRNA in orbital tissues from healthy donors and those with Graves’ disease [83]. Subsequently a number of groups have detected the transcript in orbital tissues and derivative fibroblasts [84–87]. Moreover, TSHR mRNA and protein have been found in fibroblasts and tissues from other anatomic regions [88–90]. Thus, it would appear that the receptor is expressed widely by fibroblasts and in connective tissues, albeit at low abundance. Levels of the receptor in orbital fibroblasts have been found to be altered following treatment with cytokines such as IL-6 or exposure to differentiation protocols that yield triglyceride-accumulating adipocytes [91,92]. Insinuating TSHR into a working model for TAO and dermopathy seems both logical and appealing. Unfortunately, a paucity of solid and convincing evidence generated thus far supports the concept of functionally important interactions between the TSHR displayed on fibroblasts and its ligands. Thus, compelling roles for the receptor in the pathogenesis of extra-thyroidal Graves’ disease have yet to be identified [93].

A putative component of fibroblast activation involves the CD40/CD154 molecular bridge

A number of important signalling molecules have thus far been identified on the surface of human fibroblasts including receptors for cytokines and growth factors. In addition to traditional cytokine receptors, fibroblasts from both the synovial membrane and orbit express CD40, a member of the TNF-α receptor superfamily [94]. CD40 was described initially on the surface of B cells where it functions as a key activational molecule. The natural ligand for CD40, CD154 or CD40 ligand, is a member of the TNF-α family and is expressed at high levels by activated T lymphocytes [95]. The CD40/CD154 activational bridge is currently believed to serve as an important conduit for B cell activation by T cells. The function of CD40 displayed by fibroblasts is uncertain but may be related to a ‘short-loop’ between T cells and fibroblasts for the exchange of molecular information between the two cell types. The recent report from Kaufman et al. that fibroblasts themselves can also express CD154 [96] suggests that the stream of information could flow in both directions. CD40 display has been detected on several types of fibroblasts [97–101]. Several papers have appeared suggesting that specific fibroblast genes might be targeted for CD154 activation. Zhang et al. reported that CD40 engagement resulted in an induction of PGHS-2 and the production of high levels of PGE₂ in cultured lung fibroblasts [102]. Cao et al. also found that CD40 activation could induce PGHS-2 expression in orbital fibroblasts as a result of the intermediate induction of IL-1α[59]. Moreover, disruption of the IL-1α pathway could completely block the induction of PGHS-2 and the production of PGE₂. In addition to its impact on prostanoid production, CD40 activation in orbital fibroblasts also resulted in a substantial increase in the synthesis of hyaluronan [59]. Sempowski et al. observed that activation of the CD40 pathway can also result in orbital fibroblast production of IL-6 and IL-8 [103]. Yellen et al. found similar effects on IL-6 and CD54 expression in synovial fibroblasts [104]. Moreover, activated CD40 may play a role in the proliferation of the rheumatoid synovium [105]. Thus the CD40/CD154 molecular bridge must now be considered a potentially important mechanism through which fibroblasts can become activated at sites of inflammation and tissue remodelling.

Fibroblast-displayed insulin-like growth factor-1 receptor may play an important role as a self-antigen in the pathogenesis of autoimmune diseases

T cell infiltration of affected tissues represents a central feature of both Graves’ disease and rheumatoid arthritis. The mechanisms involved in lymphocyte recruitment relevant to either disease process are incompletely understood but may involve the fibroblasts as an important ‘clearing house’ for homing signals. The pathogenesis of autoimmune diseases involves the disordered elaboration of several molecules that are proximally involved in the recruitment of T lymphocytes and other immunocompetent cells to sites of action [106]. Recently, both orbital and synovial fibroblasts have been shown to express powerful chemoattractant molecules including RANTES, a C-C chemokine and IL-16 [28, 29, 107]. Fibroblasts from every anatomic region thus far examined, including synovial and orbital cultures, can express both molecules at relatively high levels when activated by proinflammatory cytokines such as IL-1α, IL-1β, TNF-α and leukoregulin [27]. In fibroblasts from patients with either rheumatoid arthritis or Graves’ disease, IgGs from these individuals have been found to provoke the expression of IL-16 and RANTES [29,107]. The effects involved activation of caspase-3, the translation of preformed IL-16 mRNA and rapamycin-sensitive signalling pathways [107]. IgGs from patients with Graves’ disease can displace IGF-1 from its binding sites [108]. These IgGs were found to be directed at and could activate IGF-1R, an important tyrosine kinase receptor expressed on most, if not all types of human cells [107,109]. The induction of RANTES involves the up-regulation of its steady-state mRNA while that of IL-16 is directed at the translation of preformed transcript. Up-regulation of both cytokines by IgGs can be attenuated by physiologically relevant concentrations of glucocorticoids. This set of responses to IgG can be blocked as well by monoclonal antibodies that block the activation of IGF-1R or by transfecting fibroblasts with a dominant negative mutant IGF-1R [109]. IGF-1R has been found to be over-expressed by fibroblasts from donors with Graves’ disease. This elevated level of expression appears durable over many passages and time in culture. The increased density of the receptor may underlie the ability of these fibroblasts to respond to IGF-1 and receptor-activating IgGs while fibroblasts from unaffected individuals do not.

Conclusions

Fibroblasts represent a diverse population of cells with specialized roles in health and disease. When activated, they express a wide array of factors that influence immune reactions and define the nature of tissue remodelling. They are particularly important in the development of autoimmune disease and should be considered an extension of the professional immune system. A recent insight into the role of fibroblasts in initiating tissue reactivity in autoimmune disease relates to the identification of IGF-1R as a potentially critical self-antigen, the activation of which triggers T cell chemoattraction. Because of their dynamic nature and ability to exchange molecular information with neighbouring cells, fibroblasts represent a potentially attractive target for therapeutics aimed at chronic disease.