of fat has recently been reported to be more abundant in adult humans than previously anticipated, especially in smokers [37], and, as discussed below, smoking is a major risk factor for the development of GO.
Other factors that could contribute to the association of thyroid and orbit pathologies:
•The orbit and the thyroid share draining lymph nodes, so sensitization to thyroid antigens could theoretically be extended to the orbit by trafficking dendritic cells originating from the thyroid [38].
•Autoreactive B cells to the TSH-R could act locally as antigen-presenting cells and support the initiation or the development of local autoimmunity. A clue to the role of B cells as antigen-presenting cells is provided by the unexpected favourable effect of anti-CD20 (rituximab) monoclonal antibody on GO, an agent which deletes pre-B and B cells but not IgG-secreting plasmocytes [39, 40]. The therapeutic effect of rituximab has been observed after a single dose (500 or 1,000 mg), and in almost all patients reported in the literature active GO has not relapsed after 1 cycle of treatment [41, 42]. After rituximab, depletion of CD3 cells, besides CD20 cells, has been observed [43] suggesting that such therapy may switch off the mechanisms of GO progression and modify the natural course of disease.
•The fibrocyte has been suggested as a novel source of TSH-R. These CD34+ haematopoietic stem cells have been found to be more readily generated in vitro from GD peripheral blood mononuclear cells than normal [44]. Furthermore, they were found to be abundant in GO orbital tissue. The cells express high levels of both IGF-1R and TSH-R. This fascinating concept is certainly worthy of further investigation, but it is curious that other investigators have not previously demonstrated large numbers of TSH-R-positive cells (including the authors themselves) in GO orbital tissues.
•A recent study suggests that intraorbital adipogenesis could be dependent on the local increase in cortisol production through the regulation of the 11β-hydroxysteroid dehydrogenase (11β-HSD) type 1, the enzyme which activates cortisone to cortisol in the subcutaneous and omental adipose tissue [45]. Indeed, 11β-HSD1 activity was increased in both undifferentiated adipose stromal cells and in differentiated mature adipocytes from orbits taken from patients with GO. Also, the enzyme was induced by cytokines, notably TNF-α. In addition, 11β-HSD1 regulated the production of proinflammatory cytokines by GO adipose stromal cells, but not control orbital adipose stromal cells, indicating that local increased cortisol generation might also have a local anti-inflammatory action. These very interesting data extend to the orbit mechanisms known to occur in omental tissue and observed either in transgenic animals or in animals treated with 11β-HSD1 inhibitors and could lead to new therapeutic approaches.
•More recent data suggest that adipose tissue enlargement can be elevated by hypoxia caused by inflammation and swelling of orbital tissues. Hypoxia elicits hypoxia-inducible factor-1 (HIF-1) pathways in orbital fibroblasts promoting both vascularization and adipogenesis. In consequence, HIF-1-dependent pathways can strongly contribute to deterioration of GO [46]. Although the unfortunate marriage between autoimmunity and hypoxia needs to be further clarified, targeting HIF-1 and/ or VEGF may be a therapeutic option to prevent deterioration of GO.
What Kind of Immune Reactions Take Place within the Orbit?
Upon infiltrating the orbit, inflammatory cells, T and B lymphocytes, macrophages and mast cells interact with orbital fibroblasts through a whole array of cytokines. This interplay amplifies and perpetuates inflammatory/autoimmune reactions and activation of fibroblasts. However, as shown by the Rundle curve, evolution of GO is monophasic and appears as self-limited, and fibrosis ultimately develops, notably in the extraocular muscles.
IL-1, IL-4 and IFN-γ have been detected in orbital connective tissue of patients with GO. T cells obtained from GO orbital tissue appear to elicit a mixed Th1/Th2 pattern. While the Th1 pattern (IL-2, IFN-γ, TNF-α) predominates in recent-onset GO, the Th2 pattern (IL-4, IL-5, IL-10) might be associated with remission [47]. IL-6 is found in the majority of GO T-cell clones.
In vitro, cytokines have many stimulatory effects on orbital fibroblasts [36] (Fig. 3–5). They:
•increase the expression of HLA class 2 molecules, heat shock protein 72 and ICAM-1;
•stimulate the production of prostaglandin E2, a modulator of the immune response;
•stimulate the production of chemoattractants (IL-16, RANTES) as well as IL-6;
•enhance the synthesis of GAGs;
•induce extracellular matrix remodelling activity through modulation of the pericellular proteolytic environment [48];
•stimulate adipocyte differentiation (TGF-β, IFN-γ and IL-1, but not IFN-α);
•stimulate adipogenesis (IFN-α is rather inhibitory) [49].
Fig. 3. The orbital fibroblast takes part in the activation and perpetuation of the inflammatory process through the expression of HLA-DR and adhesion molecules as well as the production of chemoattractants and cytokines.
Fig. 4. Exacerbation of the production of glycosaminoglycans, notably of hyaluronan, by the fibroblasts is central to the swelling of retro-orbital tissues. Since orbital fibroblasts do not express hyaluronidase, hyaluronan accumulates. Also, the turnover of the extracellular matrix is modified by the increased expression of proteinase inhibitors which leads to changes in the structure of the adipoconnective tissue.
Fig. 5. An important feature of the pathogenesis of Graves’ orbitopathy is the differentiation of preadipocyte fibroblasts into adipocytes. This evolution is under balanced control by cytokines and activation of peroxisome proliferator-activated receptor-γ (PPAR-γ). Central to the pathogenesis of the disease is the possible link between the adipogenic differentiation and the expression at the cell surface of a functional thyroid-stimulating hormone receptor (TSH-R). TRAb, antibody against TSH-R.
Moreover, orbital fibroblasts are upregulated by IFN-γ to express CD40 allowing direct interaction with activated T cells through the CD40 ligand (CD154), which results in fibroblast activation as evidenced by the induction of IL-6 and IL-8 [50].