Department of Endocrinology Academic Medical Center University of Amsterdam Amsterdam 1105 AZ, The Netherlands
Address correspondence and requests for reprints to: Wilmar M. Wiersinga, M.D., Department of Endocrinology, Academic Medical Center, University of Amsterdam, F5-171, Meibergdreef 9, P.O. Box 22700, Amsterdam 1105 AZ, The Netherlands.
![]() |
Introduction |
---|
![]() ![]() ![]() |
---|
Knowledge of the early events in the immunopathogenesis of GO is, however, required if we want to use more specific immunological interventions. It is against this background that studies on the immunopathogenesis of GO continues to call the attention of both clinicians and scientists. The last few years have witnessed remarkable advances in our understanding of the immunopathogenesis of GO, which nevertheless is still one of the remaining enigmas in the thyroid field.
The swelling of retrobulbar tissues is largely attributable to
excessive secretion of glycosaminoglycans (GAGs) by orbital fibroblasts
(OFs). Biopsies of affected extraocular muscles show a marked expansion
of the endomysial space, caused by an increased number of collagen
fibers interspersed with a granular amorphous material containing
hyaluronic acid (1). Orbital connective tissue GAG amounts
to 254 µg/g wet tissue in GO patients and 150 µg/g in controls; the
higher GAG content in GO is largely due to an excess of chondroitin
sulfate and hyaluronic acid (2). Owing to their
polyanionic charge, GAGs osmotically attract and bind large amounts of
water, thus contributing to the increased volume of affected tissues.
In vitro studies have shown that OFs are capable to produce
GAGs in response to various cytokines. These cytokines are probably
released by infiltrating T-lymphocytes in the orbit. The lymphocytic
infiltrate is composed predominantly of CD4+ and
CD8+ T cells with a few B cells. Retrobulbar T
cells from GO patients recognize autologous OFs but not eye muscle
extracts in a MHC-class I-restricted manner. Homing of T cells to the
orbit is facilitated by the expression of HLA-DR and adhesion molecules
in endothelial cells and OF, which is enhanced by interleukin
(IL)-1, TNF
, interferon (IFN)-
, and Graves IgG. Extraocular
muscle fibers themselves are spared any immunoreactivity for adhesion
molecules (3). These findings have led to the widely
accepted view that the OF is the primary target of the autoimmune
attack.
The next obvious question is the nature of the autoantigen. Consensus has now been reached that the full-length TSH receptor (TSH.R) is expressed at the messenger RNA and protein level in orbital adipose/connective specimens of GO patients but scarcely in that of controls. The TSH.R is functional, as evident from an increase of cAMP in response to TSH. Additional studies have indicated that in vitro differentiation of preadipocyte OF (of GO and normal tissues) into adipocytes is associated with a markedly enhanced expression of functional TSH.R (4). The concept emerges that a subpopulation of OFs may be the target cells in GO, namely the preadipocytes that, if stimulated to differentiate into mature adipocytes, express increased levels of TSH.R.
So is the TSH.R the long sought-after shared antigen between the orbit and the thyroid? After all, TSH.R-stimulating antibodies (TSAbs) are the immediate cause of hyperthyroidism in Graves disease, and the view of Graves disease as an autoimmune disease directed to the TSH.R in the orbit and thyroid gland provides an attractive hypothesis to explain its various phenotypic appearances. A number of recent clinical and experimental studies seem to favor this hypothesis. In patients with untreated Graves hyperthyroidism, the prevalence of GO increases with increasing TSAb levels (5). In another study among GO patients who had been euthyroid for at least 2 months, a direct relationship was observed between serum TSAbs and quantitative measures of the eye disease like proptosis and the clinical activity score (6). The results can be interpreted as circumstantial evidence for a pathogenetic role of TSAbs in GO, but the higher TSAb levels in more severe GO may, on the other hand, just reflect a more severe autoimmune attack, in general. And how do we explain the occurrence of euthyroid or even hypothyroid GO? In such cases, the presence of TSH.R blocking antibodies (TBAbs) can be postulated preventing the development of hyperthyroidism (7), but will they not also block the effect of TSAbs in the orbit? Furthermore, not all patients with demonstrable TSAb levels have clinical manifestations of GO, although it is appreciated that the vast majority, if not all, patients with Graves hyperthyroidism harbor GO as evident from subclinical orbital involvement in the patients without apparent eye disease. Genetic or environmental factors may determine the phenotypic appearance of GO. So far, no striking genetic differences have been found between Graves hyperthyroid patients with or without clinically apparent GO. HLA-DPB1*201 is slightly more frequent in patients with manifest GO (8). A recent study indicates that a polymorphism in the CTLA4 gene also favors the clinical expression of GO (9). An association between TSH.R polymorphism and GO has not been confirmed. The strongest risk factor for GO seems to originate from the environment: the odds ratio for GO in smokers is 7.7 (95% CI, 4.313.7) (10). Indeed, 65% of GO patients smoke compared with 43% of Graves hyperthyroid patients without apparent GO and 34% of controls. Smoking is also associated with a less favorable outcome of immunosuppressive treatment of GO (11). The biological explanation of the association between smoking and GO remains to be elucidated. Possible clues are increased HLA-DR expression of OFs exposed to nicotine or tar (12), and increased GAG production by OFs cultured under hypoxic conditions.
Better evidence for a pathogenetic role of the TSH.R in GO comes from animal experiments. The first attempts to establish an animal model of GO were done with naive BALB/c and NOD mice treated with TSH.R-primed T cells, generated using a TSH.R fusion protein or by genetic immunization. None of the NOD mice but the majority of the BALB/c mice displayed orbital changes consisting of infiltration by lymphocytes and mast cells, edema, dissociation of muscle fibers, accumulation of adipose tissue, and TSH.R immunoreactivity; hyperthyroidism and TSAbs were not induced, but TBII and TBAbs were (13). The strong IL-4 and IL-10 immunoreactivity in the thyroids of animals with orbital involvement suggests that a Th2 autoimmune response to the TSH.R is required for the development of GO. Novel experiments using NMR outbred mice (more comparable with the outbred nature of humans) treated by genetic immunization with TSH.R complementary DNA, demonstrated increased T4 levels, TSAb as well as TBII and TBAb, thyroiditis and orbital pathology in a small minority of the female mice (14). The animal experiments so far indicate that orbital pathology seems to depend on sex, genetic background, and a Th2 response to the TSH.R, but does not require TSAbs.
How far does the animal data reflect what is going on in the human orbit? Is there a Th1 or Th2 response in the orbit? In patients with untreated GO both Th1- and Th2-derived cytokines in serum are elevated (15), but serum concentrations may not accurately reflect local release of cytokines in the orbit. Measurement of cytokine expression in retrobulbar tissues of GO patients is, therefore, of paramount importance. Interpretation of the few available studies is, however, hampered by treatment with steroids or irradiation before tissue sampling, and by uncertainty whether sampling was done in the early active or in the late inactive stage of the eye disease: both duration of the disease and prior immunosuppression may alter cytokine release. Analysis of orbital T cells in GO hasmaybe not quite unexpectedproduced conflicting results with respect to their phenotype and ambient cytokine profile (16, 17, 18, 19), but a recent study clearly suggests predominant cell-mediated (Th1 type) immune reactions in early GO whereas humoral immunity (Th2 type) might play the greater role in later stages (20).
The human data (implying primarily a Th1-like response to the initial
autoimmune attack in the orbit) are, thus, at variance with animal data
(indicating a Th2-like response to the TSH.R). The discrepancy casts
doubts if the orbital TSH.R is, indeed, the primary target of the
autoimmune attack in GO, although it may still be true if there is a
Th1 response to the TSH.R in the orbit. On the other hand, the
available data are compatible with the view that evolving humoral
immunity and TSAb may aggravate the course of GO, because the
expression of the TSH.R on OF increases on differentiation of
preadipocytes into adipocytes. The stimulus for differentiation is
unknown, and it is in this respect that the study by Valyasevi et
al. (21), in this issue of the journal, is of much
interest. They report that the addition of TNF or IFN
to OF in
culture inhibit differentiation into adipocytes and TSH.R expression.
As the authors state, the outcome came as a surprise: based on the
presented results one may assume that these cytokines decrease the
expression of the putative orbital autoantigen in GO and counter tissue
expansion by inhibiting adipogenesis, whereas earlier in
vitro studies demonstrate a stimulatory effect of the same
cytokines on GAG production by OF. Treatment with TNF
monoclonal
antibodies or receptor antagonists might have adverse effects as
extrapolated from the current study, but could be beneficial according
to the earlier studies. The ambiguity, once again, underscores the need
for more studies on TSH.R expression in relation to the cytokine
content of retrobulbar tissues (predominance of a Th1-like profile of
IL-2, TNFß, IFN
or of a Th2-like profile of IL-4, IL-5, IL-6 and
IL-10) of untreated GO patients in their early active or late fibrotic
stage of the disease. For, we do not really know when TNF
or IFN
appear in the orbit. Nor do we know if reported in vitro
effects are counteracted by other cytkines that may be present in
vivo. Such knowledge would provide a more scientific base for
proposed immunological interventions. T cells are much more involved in
the early than the later stages of the disease (22). Early
intervention with immunosuppressive therapy to down-regulate the
production or effects of cytokines released by T cells may alter the
natural course of GO and reduce its severity. In other fields,
effective therapies have been developed that directly interfere with
immune attacks. These include drugs interfering with TNF
, IL-1,
IL-2, and IL-10, but also with the CTLA-4 pathway. It, therefore, now
becomes clinically important to know if and when these cytokines are
present in GO orbits. Clinical trials can then be designed with
anticytokine treatment, another example being nicotinamide, which
decreases cytokine-induced activation of OF (23). The
search for the initial events in the immunopathogenesis is going on,
but we are likely on the right track. That we have got it all wrong,
becomes less probable with the emerging animal models of GO (24, 25), which may also be helpful in testing new
immunotherapies.
Received December 19, 2000.
Accepted December 19, 2000.
![]() |
References |
---|
![]() ![]() ![]() |
---|