How do T cells mediate autoimmune thyroiditis?

Folding-dependent binding of thyrotropin (TSH) and TSH receptor autoantibodies to the murine TSH receptor ectodomain

The mouse TSH receptor ectodomain (mTSHR-ecd) was amplified from murine thyroid complementary DNA and ligated into the pAcGP67B insect cell vector, and the nucleotide sequence was confirmed. Employing a baculovirus-insect cell system, the mTSHR-ecd (amino acids 22-415) was expressed as a fusion protein with the gp67 insect cell signal sequence at the NH2-terminus and a C-terminal six-histidine tag. Protein expression was assessed by Western blot using a murine monoclonal antibody (recognizing amino acids 22-35) and a rabbit antipeptide antibody (recognizing amino acids 397-415). These antibodies detected two principal species of mTSHR-ecd, one glycosylated (66 kDa) and one nonglycosylated (52 kDa), in cell lysates of infected insect cells. More than 10% of these species were present in a water-soluble (cytosolic) fraction. This fraction was then used to purify, under native conditions, 100-microg amounts of mTSHR-ecd using nickel-nitrilo-triacetic (Ni-NTA) resin chromatography. The purified cytosolic mTSHR-ecd migrated as a homogeneous 66-kDa band visible on Coomassie blue-stained gels and was confirmed by Western blotting. We also purified the mTSHR-ecd from total cell lysates under denaturing conditions, followed by "in vitro" refolding on the Ni-NTA column. Under these conditions, milligram amounts of soluble mTSHR-ecd were obtained. This material consisted primarily of the 66-kDa glycosylated form, but in addition contained four or five lower molecular mass, partially glycosylated intermediates and the 52-kDa nonglycosylated form. Deglycosylation with either endoglycosidase F or H, reduced all mTSHR-ecd glycosylated species to a 52-kDa nonglycosylated form. Both the cytosolic and refolded mTSHR-ecd preparations inhibited the binding of [125I]TSH to the full-length human TSHR expressed in Chinese hamster ovary cells in a dose-dependent manner, with similar affinities. The affinity of such interactions was 3 orders of magnitude less than observed with native porcine TSHR and was further reduced by unfolding the mTSHR-ecd preparations. The cytosolic and refolded mTSHR-ecd were also recognized by hTSHR autoantibodies in the serum of patients with hyperthyroid Graves' disease. Such autoantibody binding to mTSHR-ecd was also markedly reduced by unfolding the antigen. These results demonstrated the successful production of large quantities of well characterized, biologically active, mTSHR-ecd antigen. In addition, the data showed that although the ectodomain of the mTSHR bound TSH, intact holoreceptor may be required for high affinity ligand binding. Whether the transmembrane region is required for direct ligand binding, as seen for other G protein-linked receptors, or whether it is needed to stabilize the ligand binding to the ectodomain and maintain a correctly folded state, remains unclear.

Immunogenetic aspects of human thyroglobulin-reactive T cell lines and hybridomas

The in vitro proliferative response to T cells primed with human thyroglobulin (Tg) was compared in 11 independent haplotypes on B10 background. B10.K and B10.S mice were the most responsive, whereas, with the exception of B10.PL (H-2u), all other B10 congenics were intermediate responders. The two best responders to in vitro challenge with human Tg, of the k and s haplotype, are the same as those showing H-2-linked susceptibility to induction of experimental autoimmune thyroiditis (EAT) with mouse Tg. Since shared epitopes on human and mouse Tgs have been shown to be thyroiditogenic by adoptive transfer studies in CBA (H2k) mice, the findings indicate that shared epitopes may be studied in appropriate (i.e. EAT-susceptible) strains of mice. Therefore, we proceeded to develop methods to produce T-cell lines and hybridomas to human Tg in B10.K and B10.S mice, test their cross-reactivity to heterologous Tgs and their Ia restriction patterns. By using antigen-presenting cells from recombinant strains, we identified restriction elements encoded by the I-A subregion alone and a combinatorial molecule from the I-A/I-E subregions.

Thyroid autoantigens and human T cell responses

We investigated the ability of T cells from patients with Hashimoto's thyroiditis and with Graves' disease as well as control donors to proliferate in response to thyroid peroxidase (TPO) and thyroglobulin using (i) lymphoid cells from different lymphoid organs; (ii) unfractionated or CD8- depleted lymphoid suspensions or T cells + autologous low density cells (LDC); (iii) 200-microliters well cultures and 20-microliters hanging-drop microcultures; and (iv) intact TPO and thyroglobulin, denatured thyroglobulin and 12 synthetic peptides predicted on the basis of the amino acid sequence of TPO to be T cell epitopes. In 200-microliters well cultures, proliferative responses (assessed in terms of 3H-thymidine uptake) to intact TPO or thyroglobulin, digested thyroglobulin or synthetic TPO peptides were not significantly different in unfractionated or CD8-depleted lymphoid suspensions from blood, thyroid or lymph nodes of TPO/thyroglobulin autoantibody-positive patients, autoantibody-negative patients or control donors. In contrast, blood T cells from some high titre patients with Hashimoto's thyroiditis (but not from healthy individuals) proliferated in response to intact thyroglobulin or TPO presented by autologous LDC in hanging-drop microcultures. Heat denatured thyroglobulin (with which thyroglobulin autoantibodies do not interact) did not stimulate proliferation and this observation, together with the ability of T cells from some patients to respond to intact thyroglobulin in the absence of LDC, indicated that thyroglobulin-specific B cells may be involved in antigen presentation. As we were unable to demonstrate proliferation by blood T cells + LDC from all thyroglobulin antibody-positive patients with Hashimoto's thyroiditis, our studies suggest that the presence of sufficient precursor T cells, as well as the number and type of antigen-presenting cells, are critical for T cell proliferative responses to human TPO and thyroglobulin.

In situ analysis of T cell subset composition in experimental autoimmune thyroiditis after adoptive transfer of activated spleen cells

T cells from genetically susceptible mice developing experimental autoimmune thyroiditis (EAT) proliferate in response to restimulation with mouse thyroglobulin (MTg) in vitro. The in vitro-activated cells adoptively transfer EAT as well as differentiate into cells cytotoxic for syngeneic thyroid monolayers. To examine the kinetics of T cell subset infiltration and distribution in situ after adoptive transfer, we applied the avidin-biotin-peroxidase labeling technique to thyroid sections, utilizing rat monoclonal antibodies followed by a biotinylated rabbit anti-rat antibody. Female CBA donor mice were immunized with MTg and lipopolysaccharide. Their spleen cells were obtained 7 days later, cultured with MTg, and transferred into recipient mice. The thyroids were removed on Days 7, 10, and 14 after transfer and serially sectioned. The early phase of transferred EAT showed a higher percentage of L3T4+ cells compared to Lyt-2+ cells, yielding a ratio of 2.3 and total T cells of about 35%. By Day 10, both T cell subsets had increased to a total of about 56%. However, the relative increase was greater in the Lyt-2+ subset; the nearly doubled percentage was statistically significant, resulting in a downward shift in the subset ratio to 1.7. Little change in the in situ distribution was seen on Day 14. The percentages of F4/80+ (macrophage) population in lesions examined on Days 10 and 14 were fairly constant and B cell involvement was minimal. These findings illustrate the pathogenic role of both T cell subsets in adoptively transferred EAT and the time-dependent changes in their relative proportions leading to thyroid gland destruction.