The Thyrocyte-Fibrocyte Link: Closing the Loop in the Pathogenesis of Graves’ Disease?

Graves’ disease (GD) is an antibody-driven autoimmune process affecting the thyroid gland and extrathyroidal target tissues in roughly 90% of the cases (1). Generation of activating autoantibodies against the TSH receptor (TSHR) leads to excessive thyroid hormone synthesis uniquely in GD (2). However, the mechanisms underlying extrathyroidal tissue inflammation and remodeling remain uncertain, as does their relationship with the processes occurring in the thyroid. Manifestations of GD include thyrotoxicosis and several extrathyroidal signs including Graves’ orbitopathy (GO), dermopathy, and acropachy (3). GO and dermopathy are connective tissue manifestations of GD. Tissue remodeling is a prominent feature of both and is apparently controlled by adaptive T cells. Lymphocytes and other bone marrow-derived cells recruited into the orbit appear to be the major effectors of autoagressive tissue destruction in GO. The most likely sequence of events is that T cells infiltrate the orbital and dermal connective tissue and respond to a depot of orbital/dermal autoantigens that display an identical structure to have epitopes that cross-react with a thyroid autoantigen (1–3). Orbital fibroblasts comprise a heterogeneous population of cells possessing divergent phenotypes and potential for differentiation. They represent a bimodal population of cells with regard to the surface display of the glycoprotein, Thy1. Moreover, they exhibit attributes differing from their nonorbital counterparts that may underlie anatomic-selective involvement of the orbit in GD (4). Thus, orbital fibroblasts exhibit a unique phenotype, including exaggerated responses to proinflammatory cytokines. Cytokines play a key role in the development of the extrathyroidal manifestations of GD. These molecules are produced in the orbit of GO patients by infiltrating inflammatory cells as well as orbital fibroblasts. Cytokines can modulate the immune reaction in GO by increasing major histocompatibility complex class II, adhesion molecules, CD40 (a member of the TNF-α receptor superfamily), prostaglandin, and heat shock protein expression in the orbit, thereby having a role in localizing and augmenting the inflammatory response (5). Locally released cytokines, particularly IL-1, TNF-α, and interferon γ, stimulate fibroblast proliferation and also activate fibroblasts to secrete the hydrophilic glycosaminoglycans (GAG) that cause edema with consequent proptosis. CD40 engagement leads to substantial increases in GAG synthesis in orbital fibroblasts. Also, CD40 ligand up-regulates the synthesis of IL-1α, and blocking this cytokine with exogenous IL-1 receptor antagonist or with IL-1α neutralizing antibodies partially attenuates the induction of prostaglandin endoperoxide H synthase-2. Thus, the CD40/CD40 ligand bridge represents a potentially important activational pathway for orbital fibroblasts that may underlie their cross-talk within the gateway of B-T-cell communication. Finally, smoking may be linked to GO through the generation of partial hypoxia, which increases GAG synthesis by fibroblasts, and by direct effects of cigarette smoke constituents on fibroblast activation (3). The disordered accumulation of the GAG in orbital and dermal connective tissues leads to the distinctive histological features associated with GD-related GO and dermopathy (6). This often occurs in the context of dramatic inflammation and tissue remodeling, which is a prominent feature of the extrathyroidal manifestations of GD and is apparently driven by activated T cells. GAG take part in extracellular matrix organization, the regulation of cell proliferation and differentiation, and the regulation of diffusion of macromolecules through tissues, growth factor attachment, cell-to-cell interaction, and regulation of IL-1 production in local inflammatory response. Hyaluronic acid (HA) is the largest GAG molecule containing several thousand repeated disaccharides. In contrast to the other GAG components, HA is unsulfated. Interactions of proteins and other organic molecules with HA generally involve specific, noncovalent binding to sites defined by several disaccharide units on the HA monomer. One HA backbone molecule can associate with up to 200 proteoglycans via stabilizing link proteins that form high molecular weight herringbone-like aggregates of 5 × 108 Da. Although GAG disaccharides maintain the essential water and electrolyte content of the extracellular matrix, their accumulation en masse with negatively charged sulfate residues results in enhanced water-binding capacity (1000-fold higher than albumin on a molar basis). The altered structure and nature of sulfated GAG units not only promotes inflammatory signaling but also is the likely major contributor to increased volume of the orbital and pretibial connective tissue, leading to an elevated pressure in the bony orbit and to displacement of the globe anteriorly (3, 6). The nature of the putative cross-reactive antigen connecting the thyroid gland and the extrathyroidal manifestations of GD remains contentious; the current best bet is the TSHR (1–3, 7). Subpopulations of orbital and dermal fibroblasts that express full-length TSH-R can be programmed to differentiate into fat cells. TSH-R-stimulating antibodies may therefore trigger the excess orbital fat in GO through their differentiating effects on these preadipocyte fibroblasts. TSHR, a dominant autoantigen in GD, is underscored by its conspicuous expression on both fibroblasts and preadipocytes in affected target connective tissue depots. Recently, a second autoantigen, IGF-1R, has been proposed (8–12) based on the following compelling evidence: 1) the ability of IgG isolated from the sera of patients with GD (GD-IgG) to provoke the synthesis of HA in IGF-1R+ orbital fibroblasts (8); 2) the replicating effect of GD-IgG by IGF-I appeared to be mediated through the IGF-1R and was abolished with glucocorticoid treatment; and 3) TSH failed to influence the synthesis of HA. In contrast to GD fibroblasts, cultures derived from donors without thyroid disease were nonresponsive to GD-IgG or IGF-I. The observation that GD-IgG induced HA production in fibroblasts suggests that the IGF-1R and its activating autoantibodies represent a key pathway through which important pathogenic events in GO are mediated. Fibroblasts from patients with GD expressed IL-16, a CD-specific chemoattractant, and RANTES, a CCL5 memory T cell chemokine, in response to GD-specific IgG (9). These responses resulted from a functional interaction between GD-IgG and the IGF-1R. Neither GD-IgG nor IGF-I activated chemoattractant expression in control fibroblasts from donors without GD. Interrupting IGF-1R function with specific receptor-blocking antibodies or by transiently transfecting fibroblasts with a dominant-negative mutant IGF-1R completely attenuated signaling provoked by GD-IgG. Moreover, GD-IgG displaced specific radioiodine-labeled IGF-I binding to fibroblasts and attenuated IGF-1R detection by flow cytometry (10). These findings identified a novel disease mechanism involving a functional GD-IgG/IGF-1R bridge, which potentially explains T cell infiltration in GD. Interrupting this pathway may constitute a specific therapeutic strategy. Furthermore, a disproportionately large fraction of GD-derived peripheral blood T cells expressed durable, constitutive IGF-1R expression in culture. The additions of IGF-I to anti-CD3 activated T cells of GD but not normal donors led to preservation of an apoptotic-resistant T cell subset (11). These findings suggest a potential role for IGF-1R displayed by lymphocytes in supporting the expansion of memory T cells in GD. Finally, IGF-1Rß and TSHR colocalized to the perinuclear and cytoplasmic compartments in fibroblasts and thyrocytes by confocal microscopy (12). Examination of orbital tissue from GO patients revealed similar colocalization to cell membranes. To further enlighten the “tangled” relationship between the thyroid and the extrathyroidal manifestations, in this issue of JCEM, Douglas et al. (13) introduced for the first time the notion of fibrocytes in subjects with GD. Fibrocytes constitute a population of cells possessing phenotypic attributes that both resemble and diverge from those of fibroblasts. They derive from monocyte or B cell bone marrow lineages and infiltrate connective tissues in response to injury, where they provide antigen-specific T cell stimulation (14). They participate in inflammation and tissue remodeling and also drive fibrosis. Fibrocytes synthesize collagen I, display cell-surface CD34 and CXCR4, and traffic to peripheral tissue sites in response to CXCL12, the cognate ligand for CXCR4 (15, 16). Of interest, fibrocytes can differentiate into myofibroblasts and adipocytes when treated with TGF-β and peroxisome proliferator-activated receptor γ ligands, respectively (17). In their innovative paper, Douglas et al. (13) aimed to determine whether CD34+ fibrocytes are generated more frequently in GD, whether they infiltrate extrathyroidal connective tissues, and whether they express the TSHR. To answer these questions, a generation of fibrocytes from peripheral blood mononuclear cells (PBMCs) was examined in samples from 70 patients with GD and 25 healthy control subjects. Fibrocytes were characterized by flow cytometry, and orbital tissues and fibroblast culture strains were examined for their presence. The main findings of this important work may be summarized as follows: Fibrocytes, CD34+CXCR4+ collagen 1+ CD11b+ cells with well-documented monocytoid lineage, were far more frequently generated from cultured PBMCs of donors with GD compared with healthy subjects. This increase in generation is not specific to GD but is also found in other diseases and under physiological stress. Fibrocytes expressed TSHR at high levels that rivaled those of thyroid epithelium. In addition, these fibrocytes displayed IGF-1R; thus, two of the autoantigens implicated in GD were expressed by these cells. These two antigens shared a close functional and physical relationship. The TSHR expressed by fibrocytes was functional. When fibrocytes were treated with TSH, they generated extremely high levels of IL-6 and TNF-α. Numerous fibrocytes accumulated in the orbital connective tissue from patients with GO but were not found in control orbital tissue. Tissue-infiltrating fibrocytes expressed TSHR in situ and comprised a subpopulation of GO-derived orbital fibroblasts. Moreover, these CD34+ cells appeared to account for the cells within the GO orbital tissue that expressed TSHR. CD34+ fibroblasts, those orbital fibroblasts whose putative origins are from fibrocytes, were found abundantly among the fibroblasts cultivated from GO orbital tissues. They were absent in fibroblast cultures from control tissues. This may account for the divergent phenotypes of control vs. GO fibroblasts. Moreover, it might help explain the cellular heterogeneity associated with GO orbital fibroblasts. Although Fig. 5 in the Douglas et al. paper (13) shows CD34+ collagen 1+ orbital fibroblasts (isolated from GO patients having orbital decompression surgery) expressing TSH-R after differentiation into adipocytes, one technical limitation of this paper may be the long “in vitro adaptation period” to recover CD34+ fibroblasts from 14-d adherent cell culture rather than analysis of these cells at baseline. These findings raise a number of important possibilities concerning the TSH/TSHR pathway as immune regulators (18). They build on the earlier findings of Klein (19) who has reported finding TSHR- and TSH-positive cells in the bone marrow. The findings also identify a potential mechanism for the loss of peripheral tolerance to TSHR found in GD. They suggest that the fibrocyte, by virtue of a high level TSHR expression, might “immunize” individuals at risk with TSHR, thereby resulting in the generation of immune reactions against the receptor. Thus, the findings might represent insight into a previously unidentified mechanism underlying the pathogenesis of GD and GO. Now, how do we put all these original data together? And at what stage (bone marrow precursors or in blood circulation) do CD34+ GO fibrocytes become corrupted toward TSH-R expression? It is tempting to go a step further and to speculate on the clinical relevance of this “thyrocyte-fibrocyte link”; on the close relationship between TSHR, TSHR autoantibodies, and the fibroblast putative precursor fibrocyte; and therefore on the innovative notion of a systemic GD linked to a bone marrow stem. Based on the findings stated above and on previous reports (18, 19), the TSHR may play a major role in initiating an immune response in the bone marrow, much earlier before the circulating PBMCs reach the extrathyroidal target tissues in GD. On the other hand, TSHR bearing circulating fibroblasts and fibrocytes may directly be activated by the TSHR autoantibodies, leading to the secretion of high amounts of collagen 1, HA, and other hydrophilic GAG. This leads to edema and swelling of the extrathyroidal target tissues and would explain the previously reported increased levels of circulating GAG in various body fluids of patients suffering from active GD and/or GO (20). Although much more work in this area is still warranted, the knowledge that the complex immunopathogenesis of GD is linked to bone marrow stem cells brings us closer to closing the loop between the thyroid gland and the immune system. Disclosure Summary: The author has nothing to declare. Glycosaminoglycans Graves’ disease IgG isolated from the sera of patients with GD Graves’ orbitopathy hyaluronic acid peripheral blood mononuclear cells TSH receptor