On the Pathogenesis of Autoimmune Thyroid Disease: A Unifying Hypothesis

Stelios Fountoulakis; Agathocles Tsatsoulis


Clin Endocrinol. 2004;60(4) 

In This Article

Role of Apoptosis in the Clinical Expression of AITD

Apoptotic Pathways in the Thyroid and Their Regulation

Apoptosis or programmed cell death involves the sequential activation of a series of proteolytic enzymes known as caspases that progressively digest the cell and its genetic material. The caspase-cascade is activated by two pathways; the death receptor and the mitochondrial pathway. Specific signals such as ligands that bind to death receptors or withdrawal of growth factors and trophic hormones may activate the apoptotic machinery. The best studied receptor-mediated apoptotic pathway is the Fas/Fas Ligand (FasL) system. Fas is a type I transmembrane protein and belongs to the TNF receptor superfamily. FasL is a type II transmembrane protein that on ligation with the Fas receptor induces apoptosis on Fas-expressing target cells (Fig. 2; Suda et al., 1993; Nagata & Golstein, 1995).

Figure 2.

Components of the Fas and DR4/DR5 receptor-mediated apoptotic pathways. On ligation of FasL, Fas receptor trimerizes and recruits, via its death domain, an adapter protein known as Fas-associated protein with death domain (FADD). FADD recruits and interacts with procaspase 8. Formation of the Fas-FADD-procaspase 8 complex facilitates the autocleavage and activation of procaspase 8. Caspase 8 cleaves and activates caspase 3 and other downstream effector caspases. This caspase-cascade activation finally leads to apoptosis. TRAIL legates to its receptors DR4 and DR5, which recruit an unknown adapter. Procaspase 10 is also recruited and autoactivation of caspase 10 occurs. Consequently, caspase 10 cleaves and activates caspase 8, thus initiating the apoptotic caspase-cascade. Bcl-2 family members may act as regulators of apoptosis.

Soluble forms of Fas (sFas) and FasL (sFasL) have also been detected. sFas lacking the transmembrane domain of its receptor results from alternative splicing of Fas mRNA. sFas can inhibit Fas/FasL interaction and protect target cells from Fas-induced apoptosis (Cheng et al., 1994). sFasL results from FasL cleavage by a metalloproteinase (Kayagaki et al., 1995) and it is postulated to induce apoptosis on Fas-expressing cells in vitro. However, its in vivo role remains unclear. There are indications that sFasL production, through transmembrane FasL cleavage, may downregulate FasL's killing activity and prevent suicidal death through Fas/FasL interaction among neighbouring cells (Suda et al., 1997).

The TNF-related apoptosis-inducing ligand (TRAIL) is a more recently characterized member of the TNF family of proteins that exists in both membrane bound and soluble forms and has the closest resemblance to FasL. TRAIL induces apoptosis by engaging its receptors, TRAIL-R1 (DR4) and TRAIL-R2 (DR5), via a caspase-dependent pathway (Fig. 2; Pan et al., 1997a, 1997b).

The Fas/FasL system plays an important role in normal immune regulation by removal of autoreactive T and B lymphocytes at the end of an immune response (Van Parijs et al., 1998). These molecules can also be used by immune effector cells to kill targets. A variety of cell types, including the thyroid cell, constitutively express Fas (Hammond et al., 1997; Mitsiades et al., 1998, 1999; Hiromatsu et al., 1999b). The Fas pathway is, however, inhibited under normal conditions possibly by labile protein inhibitors (Arscott et al., 1997; Mitsiades et al., 1998). This inhibition may be reversed during inflammation by the action of inflammatory cytokines. The optimal conditions for the induction of Fas-mediated apoptosis in thyroid cells require the presence of IFN-γ in combination with IL-1β or TNF-α (Bretz et al., 1999a; Wang et al., 2002).

The expression of FasL on thyroid cells is controversial. Normally, the expression of FasL is restricted, occurring only on activated lymphocytes and 'immune-privileged' sites such as the cornea, the testis and the anterior chamber of the eye (Griffith et al., 1995). However, under certain pathological conditions the thyroid cell can be induced to express FasL and to use FasL to influence its environment (Hiromatsu et al., 1999b; Giordano et al., 2001; Sera et al., 2001).

In vitro studies have also shown that TSH was able to inhibit Fas-mediated apoptosis, whereas withdrawal of TSH (Kawakami et al., 1996) resulted in the induction of apoptosis.

TRAIL death receptors DR4 and DR5 appear to be constitutively expressed on normal thyrocytes but their molecules become in vitro functional only in the presence of a protein synthesis inhibitor, which reduces labile inhibitors' activity (Bretz et al., 1999b). In addition, inflammatory cytokines can induce functional expression of TRAIL pathway in thyroid cells (Bretz et al., 1999a; Bretz & Baker, 2001).

There are also indications that expression of antiapoptotic molecules such as Bcl-2, could be regulated in either thyroid cells or lymphocytes in an attempt to avoid apoptotic death (Giordano et al., 1997, 2001; Mitsiades et al., 1998).

Apoptosis and Clinical Expression of AITD

In HT, increased apoptosis has been proposed as the mechanism for the thyroid cell destruction characteristic of the disease (Arscott & Baker, 1998). Histologically, HT is characterized by massive infiltration of lymphoid cells in or around thyroid follicular cells (Baker, 1997). In areas where intact follicular cells are observed in proximity to the infiltrating immune cells, thyroid cells appear apoptotic, suggesting immune-mediated destruction of these cells (Kotani et al., 1995). In contrast, in GD, the degree of lymphocyte infiltration in the thyroid is less and the rate of apoptosis in thyrocytes is low (Tanimoto et al., 1995).

There is some controversy regarding the mechanism of thyroid cell apoptotic death seen in HT. The prevailing view is that Fas-mediated thyroid cell apoptosis is induced by the infiltrating cytotoxic T lymphocytes expressing FasL and Bcl-2 (Fig. 3a; Arscott & Baker, 1998; Mitsiades et al., 1998; Kawakami et al., 2000; Giordano et al., 2001). According to recent evidence, activation of the Fas pathway normally inhibited in thyroid cells is induced by a certain combination of proinflammatory cytokines released locally by the infiltrating lymphocytes (Giordano et al., 1997; Hiromatsu et al., 1999b). Predominating Th1 inflammatory cytokines (IFN-γ, TNF-α, IL-1β), which are produced due to B7·1 co-stimulation of infiltrating lymphocytes by 'professional' APCs, seem to have an important role, enhancing thyroid cell susceptibility to apoptosis. Thyroid cells may also be capable of B7·1 expression, thus providing a co-stimulatory signal to T lymphocytes for Th1 cytokine production (Battifora et al., 1998). A recent study has shown a Th1 cytokine production pattern, not only in intrathyroidal but also in peripheral CD4+ and CD8+ T lymphocytes in HT patients (Mazziotti et al., 2003). Furthermore, Fas-mediated cell death was induced when thyroid cells were pretreated with IFN-γ in combination with IL-1β or TNF-α (Bretz et al., 1999a; Wang et al., 2002).

Figure 3.

Potential models for the role of Fas-mediated apoptosis in AITD. (a) Cytokines produced by activated intrathyroidal T lymphocytes induce the activation of the Fas pathway on thyroid cells leading to their apoptosis after interaction with FasL-bearing T lymphocytes or sFasL deriving from activated T lymphocytes. (b) Simultaneous expression of Fas and FasL on thyroid cells leads to their fratricidal demise. (c) Binding of thyroid cell surface FasL with Fas expressed on intrathyroidal T lymphocytes results in T lymphocyte apoptosis.

An alternative hypothesis has been put forward by Giordano et al. (1997, 2001) who suggested that the apoptosis observed in the thyroid of patients with HT is a consequence of simultaneous expression of Fas and FasL by thyrocytes, resulting in fratricide among neighbouring cells (Fig. 3b). This model is based on the observation that FasL is constitutively expressed on thyrocytes and that IL-1β produced by infiltrating immune cells induces Fas expression on thyrocytes. This hypothesis is, however, at odds with a growing number of reports suggesting that Fas, and not FasL is normally expressed on thyrocytes (Arscott et al., 1997; Mitsiades et al., 1998, 1999).

Soluble forms of Fas and FasL could also have a role in HT pathogenesis. sFasL shed off from activated lymphocytes may induce apoptosis on Fas-expressing target cells (Martinez-Lorenzo et al., 1996). On the other hand, sFas is found to be decreased in HT, so that the increased full-length membrane Fas may promote thyroid cell apoptosis (Shimaoka et al., 1998). The expression of the antiapoptotic protein Bcl-2 is also downregulated in the thyroid cells of patients with HT, creating a proapoptotic potential for thyroid cells (Andrikoula et al., 2001; Giordano et al., 2001).

Recent evidence also suggests that other ligand receptor pairs may contribute to thyroid cell apoptosis. Activated immune cells could, through the release of TNF-α and IL-1β, sensitize thyroid cells to TRAIL-mediated apoptosis, thus leading to their destruction (Bretz et al., 1999a; Bretz & Baker, 2001).

Mechanisms implicated in thyrocyte elimination in HT, apart from apoptotic death, might include exocytic granules containing perforin and granzyme B (Wu et al., 1994; Talanian et al., 1997) and complement activation. Involvement of the complement system has been described in both GD and HT even though HT and GD thyrocytes are relatively resistant to lysis by the membrane attack complex due to the expression of inhibitory proteins (Chiovato et al., 1993; Tandon et al., 1994b). In this case, complement activation may not directly kill thyroid cells but it can exacerbate the autoimmune process thus suggesting a role for the disease maintenance in human AITD (Tandon et al., 1994b). Although these mechanisms cannot be overlooked, the weight of evidence suggests that their impact may not be as important as that of apoptosis, at least as far as the clinical expression of AITDs is concerned (Okosieme et al., 2002).

Finally, the reduction or loss of intracellular communication may also contribute to thyroid cell destruction leading to hypothyroidism. In animal models of thyroid autoimmunity, a decreased assembly of connexins into gap junctions has been found and this is accompanied by deficient cell-to-cell communication (Green et al., 1996, 1997) and their demise by a type of apoptosis called anoikis (Di Matola et al., 2000).

In contrast to HT, which is characterized by thyroid cell destruction and hypothyroidism, GD is manifested by thyroid cell hyperplasia and hyperthyroidism, despite clear evidence of autoimmunity. Histologically, in GD the lymphocytic infiltration of the thyroid is milder and the number of apoptotic thyroid cells is low. This observation suggests that mechanisms for thyroid cell survival and active killing of infiltrating lymphocytes may be operative in GD. This in turn may be due to the different profile of cytokines secreted in the local microenvironment by different subsets of infiltrating lymphocytes. In GD, there appears to be a predominant production of Th2 type inflammatory cytokines (IL-4, IL-5, IL-10), in contrast to Th1 prevalence in HT, probably as a result of B7·2 co-stimulation of infiltrating lymphocytes (Mosmann & Sad, 1996; Roura-Mir et al., 1997; Bretz et al., 1999a; Stassi et al., 2000).

Th2 cytokine predominance favours humoural immunity instead of cellular immunity, enhancing autoantibody production by B lymphocytes. Indeed, patients with GD produce high titres of IgG antibodies specific for the TSH receptor. These antibodies function like TSH by activating the receptor, causing thyroid cell hyperplasia and hyperfunction. The TSH receptor-stimulating antibodies may also play an antiapoptotic role, protecting thyroid cells from apoptosis. Indeed, both TSH and IgG from patients with GD decrease Fas expression in normal thyrocytes (Kawakami et al., 1997, 2000). In addition, Th2 type cytokines may upregulate antiapoptotic molecules, including Bcl-2 protein and protect thyrocytes from apoptotic cell death (Mitsiades et al., 2001; Nagayama et al., 2003).

On the other hand, thyroid cells may also be capable of fighting off T lymphocytes via induced expression of death ligands such as FasL or TRAIL, which may kill intrathyroidal lymphocytes (Fig. 3c; Giordano et al., 1997; Bretz & Baker, 2001). Soluble Fas may also play a role in the inhibition of the Fas/FasL system in patients with GD. Serum concentration of sFas is found to be increased in patients with GD compared to control subjects. sFas was also detected in the supernatant of cultured thyrocytes from patients with GD (Shimaoka et al., 1998; Hiromatsu et al., 1999a; Feldkamp et al., 2001).

It appears that differences in the local microenvironment are crucial to the development of either GD or HT. The most important of these differences are summarized in Table 2 and include the different subsets of T lymphocytes accumulating in the thyroid and the different profile of cytokines that are secreted and that in turn influence the expression of death receptors, ligands, antiapoptotic molecules, inhibitory molecules and soluble agents. It seems highly possible that in GD all these parameters are regulated in a manner that promotes thyrocyte survival and hypertrophy, while leading to lymphocyte apoptosis. In HT, the microenvironment favors thyroid cell destruction by apoptosis induced by the invading cytotoxic immune cells.

The crucial role of the balance between Th1 vs. Th2 response in the divergent phenotypic expression of AITD is also supported by a recent clinical observation (Coles et al., 1999). According to this report a significant number (1/3) of patients with multiple sclerosis develop GD after treatment with the humanized anti-CD52 monoclonal antibody Campath 1-H, aiming to deplete T lymphocytes and suppress cellular immunity. A possible explanation for this outcome is that T lymphocytes which reemerge after depletion are biased away from a Th1 to a Th2 phenotype, leading to the development of antibody-mediated GD (Weetman, 2003).

The subset of autoimmune atrophic thyroiditis with blocking anti-TSH receptor antibobies may be regarded as the opposite end to that of GD. Here again, predominance of humoural immunity may enhance production of anti-TSH receptor antibodies. In contrast to GD, however, the prevailing antibodies are inhibitory of the TSH receptor (Rees Smith et al., 1988; Rapoport et al., 1998). A study by Kawakami et al. (1997) using IgG from patients with idiopathic myxoedema showed that thyroid cells may undergo apoptosis, due to lack of TSH trophic effect, leading to thyroid atrophy and hypothyroidism. The pathogenic role of blocking anti-TSH receptor antibodies in causing hypothyroidism in autoimmune thyroiditis is also suggested by their capacity to induce neonatal transient hypothyroidism (Brown et al., 1996).

Finally, the other forms of chronic autoimmune thyroiditis, silent (postpartum or sporadic) thyroiditis resemble that of HT but the degree of lymphocytic infiltration is milder (Pearce et al., 2003). It appears that an initial phase with destruction of thyroid cells due to apoptosis and/or necrosis takes place leading to mild thyrotoxicosis, but this destructive phase is transient and not persistent as in HT. The destructive phase is usually followed by a recovery phase during which temporary hypothyroidism may occur. However, thyroid function eventually returns to normal in the majority of patients although a small number may go on to develop HT.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: