Characterisation of the antibody response to the extracellular region of recombinant thyrotropin receptor

Residues 34-39 in the thyrotropin receptor are not the target of autoantibodies from sera of patients with Graves' disease

The thyrotropin receptor (TSHR) and alphal-antytripsin contain a fragment of sequence composed of 6 amino acids in which 5 residues are identical. Previously, we have suggested that this region of similarity [residues 34-39: (EEDFRV) in TSHR] is not the target for Graves' disease patients' autoantibodies. To verify this suggestion, we studied the reaction of patients' sera with alphal-antitrypsin. Two methods were used: TRAK assay, normally designed to estimate anti-TSHR autoantibodies in patients' sera, and immunoblotting. A modified version of the former assay was also used to study the influence of the synthetic peptide, corresponding to the region of similarity in TSHR, on Graves' patients sera or on thyrotropin (TSH) binding, and to study the influence of this peptide antiserum on TSH binding to the receptor. The TSHR stimulating and blocking activity of antisera to this peptide was studied in transfected Chinese hamster ovary cells. No influence of alphal-antitrypsin on the binding of patients' antibodies to the receptor were noticed nor were there reactions of autoantibodies with alphal-antitrypsin. We found that patients with anti-TSHR autoantibodies had a normal concentration of alphal-antitrypsin. A peptide corresponding to residues 34-39 did not influence Graves' patients sera and TSH binding and antiserum to this peptide did not influence TSH binding and adenylate cyclase activity. In summary, the results indicated that the sequence EEDFRV is not the target for patients autoantibodies.

Characterization of the thyrotropin binding pocket

A panel of monoclonal antibodies (mAbs) to the thyrotropin receptor (TSHR) was prepared using three different immunization strategies. The mAbs obtained (n = 138) reacted with linear epitopes covering most of the TSHR extracellular domain and with conformational epitopes. mAbs that bound to five different regions of the TSHR (amino acids [aa] 32-41, aa 36-42, aa 246-260, aa 277-296, and aa 381-385) were able to inhibit (125)I-labeled thyrotropin (TSH) binding to solubilized TSHR preparations. Fab and immunoglobulin G (IgG) preparations were similarly effective inhibitors for mAbs reactive with aa 246-260, aa 277-291 and aa 381-385 suggesting that these three regions of the TSHR are involved in TSH binding. In contrast mAbs reactive with aa 32-41 and aa 36-42 were not effective at inhibiting TSH binding when Fab preparations were used, suggesting that these N terminal regions of the TSHR were less critical for TSH binding. Our studies suggest that three distinct and discontinuous regions of the TSHR (aa 246-260 and 277-296 on the TSHR A subunit) and aa 381-385 (on the TSHR B subunit) fold together to form a complex TSH binding pocket. Alignment of the aa sequences of these three regions in TSHRs from different species indicates that they are highly conserved.

Limitations of the semisynthetic library approach for obtaining human monoclonal autoantibodies to the thyrotropin receptor of Graves' disease

Graves' disease (GD) is characterized by the presence of autoantibodies against the TSH-receptor (TSH-R) which are pathogenic and, upon binding to the receptor, trigger intracellular signal transduction. The autoantibodies are oligoclonal and as they are responsible for disease activity, their characterization would lead to a better understanding of the development of GD. Attempts to isolate anti-TSH-R antibodies from patients have proved to be difficult due to the exceedingly low serum levels due to rarity of these B cells, together with difficulties in obtaining purified TSH-R capable of interacting with patients autoantibodies. We employed phage antibody display technology and performed selection with a previously characterized semisynthetic antibody library on the purified extracellular ectodomain of the TSH-R. We report the isolation of six different anti-TSH-R monoclonal phage antibodies (moPhabs) from this library. All the moPhabs recognized TSH-R and its recombinant fragments by Western blotting, but failed to recognize the native TSH-R by flow cytometry. Consequently, the moPhabs did not lead to TSH-R activation. As these were the first moPhabs to TSH-R, they were analysed in terms of nucleotide and amino acid sequence and epitope specificity on the receptor. The moPhabs used immunoglobulin VH1 and VH3 germ line genes, all associated with Vlambda3 genes. Interestingly, the CDR3 regions of all moPhabs were remarkably similar, though not identical. In light of the common CDR3 usage, the epitopes recognized on TSH-R appeared to be restricted to amino acids residues 405-411 and 357-364. In summary, our results show that semisynthetic libraries may be limited in isolating human monoclonal antibodies that resemble pathogenic antithyrotropin receptor autoantibodies present in patients with GD. It is likely that until preparations of purified TSH-R that can be recognized by patients autoantibodies become available, similar to the recently described glycosylphosphatidylinositol (GPI) anchored TSH-R ectodomain, monoclonal antibodies from phage antibody display to TSH-R will be limited for isolating the rare, pathogenic antibodies of GD.

Induction of thyroiditis in mice with thyrotropin receptor lacking serologically dominant regions

Grave's disease (GD) is characterized by pathogenic autoantibodies to the human thyrotropin receptor (hTSH-R), and is frequently associated with a lymphocytic infiltrate of the thyroid gland. In attempts to establish a murine model of GD, we and others have previously shown that immunization of mice with recombinant preparations of the hTSH-R ectodomain induces high titres of specific antibodies, which, however, are not pathogenic, nor is the response accompanied by the development of thyroiditis. Since earlier reports identified the serological immunodominant determinants within the N- and C-terminal regions of hTSH-R ectodomain, we reasoned that immunization of mice with truncated fragments of ectodomain lacking these dominant regions might result in skewing of the response to other determinants of the molecule, with consequent induction of immunopathological features present in GD. We show here that multiple challenge of BALB/c mice with an amino acid fragment of residues 43-282 generates antibodies directed at hTSH-R peptides 37-56, 157-176, 217-236 and 232-251. This reactivity pattern is distinct from that induced previously with the whole ectodomain of hTSH-R in BALB/c animals. Thyroid function remained unaffected in these mice, suggesting that pathogenic antibodies were not being induced. Interestingly, some animals developed lymphocytic infiltration of the thyroid gland, clearly indicating the presence of pathogenic T cell determinants within the 43-282 fragment. Challenge with the related fragment 43-316 produced the same pattern of serological response to the synthetic peptides as fragment 43 282, but was not accompanied by thyroiditis. The results demonstrate: (i) the presence of thyroiditogenic determinants within hTSH-R, and (ii) that these pathogenic determinants are likely to be cryptic, as their effect is exhibited only when the hierarchy of immunodominance within hTSH-R is drastically altered.