Thyrotropin receptor cleavage at site 1 does not involve a specific amino acid motif but instead depends on the presence of the unique, 50 amino acid insertion

Thyrotropin (TSH) receptor (TSHR) A and B subunits are formed by intramolecular cleavage of the single chain receptor at two separate sites. The region involved in cleavage at Site 2 has been identified, but previous mutagenesis studies failed to identify Site 1. We now report fortuitous observations on the effect of trypsin on the TSHR that localizes a small region harboring Site 1. Thus, as detected by immunoblotting and by 125I-TSH cross-linking to TSHR expressed on the surface of intact CHO cells, trypsin clipped a small polypeptide fragment bearing a glycan moiety from the C terminus of the A subunit. Based on the TSHR primary structure, this small fragment (1-2 kDa) contains Asn-302. This information, together with estimation of the size of the deglycosylated A subunit relative to a series of C-terminal truncated TSHR ectodomain variants, places cleavage Site 1 in the vicinity of, or closely upstream to, residue 317. Remarkably, mutagenesis of every amino acid residue between residues 298-316 (present study) and 317-362 (previous data) did not prevent cleavage at Site 1. However, cleavage at this site was abrogated by deletion of a 50-amino acid segment (residues 317-366) unique to the TSHR in the glycoprotein hormone receptor family. In summary, these data provide novel insight into TSHR intramolecular cleavage. Cleavage at Site 1 does not depend on a specific amino acid motif and differs from cleavage at Site 2 by involvement of a mechanism requiring the presence of the enigmatic TSHR 50-amino acid "insertion."

An N-linked glycosylation motif from the noncleaving luteinizing hormone receptor substituted for the homologous region (Gly367 to Glu369) of the thyrotropin receptor prevents cleavage at its second, downstream site

The thyrotropin receptor (TSHR) exists in two forms (single polypeptide and two subunits), whereas the lutropin/chorionic gonadotropin receptor (LH/CGR) is a single chain. Recent data suggest that the TSHR cleaves at two sites. We mutagenized selected chimeric TSH-LH/CGR to localize the cleavage sites in the TSHR. All 23 receptors mutated in the estimated vicinity of the upstream site cleaved into two subunits as determined by 125I-TSH cross-linking to intact cells. In contrast, in a series of mutations homologous to the noncleaving LH/CGR, the downstream TSHR cleavage site localized to three amino acids (GQE367-369). Remarkably, group substitution of these residues, but not substitution of individual residues, abolished cleavage. Moreover, the mutation that prevented cleavage (GQE367-369NET) transposed a motif (NET291-293) that is glycosylated in the LH/CGR. TSHR cleavage or noncleavage after substitution of GQE367-369 with other triplets (AAA, NQE, and NQT) was consistent with a role for N-linked glycosylation at this site. In summary, our data (i) support the concept that the TSHR cleaves at two sites, (ii) relate TSHR residues GQE367-369 to cleavage at the second, downstream site, and (iii) suggest that cleavage or noncleavage at site two is related to N-linked glycosylation. These findings provide new insight into the evolutionary divergence of two closely related receptors.

Diverse Fab specific for acetylcholine receptor epitopes from a myasthenia gravis thymus combinatorial library

The muscle weakness in myasthenia gravis (MG) is caused by heterogeneous high-affinity IgG autoantibodies to the nicotinic acetylcholine receptor (AChR), a complex ion channel glycoprotein. These antibodies are clearly responsible for reducing AChR numbers at the neuromuscular junction in myasthenia; however, the origins, diversity, specificity and pathogenicity of individual antibodies have not yet been established. We have cloned and characterized four different AChR-specific Fab from an MG patient's thymus by screening an IgG1/kappa gene combinatorial lambda phage library with soluble human AChR labeled with [125I] alpha-bungarotoxin. Unlike most previously cloned human antibodies, all four Fab immunoprecipitated soluble human muscle AChR. Two Fab strongly inhibited binding of mAb to the main immunogenic region on the alpha subunits and one Fab bound to an epitope on the fetal-specific gamma subunit. In sensitivity and fine specificity, these Fab resembled the anti-AChR antibodies found in many MG patients, including the donor. The closest germline counterparts for their heavy chains were in VH families 1, 3 and 4; however, there were many differences consistent with an antigen-driven response of diverse B cell clones. The combinatorial approach holds promise for further analysis of human autoantibodies.

The epitopic "fingerprint" of thyroid peroxidase-specific Fab isolated from a patient's thyroid gland by the combinatorial library approach resembles that of autoantibodies in the donor's serum

A new thyroid peroxidase (TPO)-specific Fab (KM1) was obtained from an immunoglobulin gene combinatorial library of patient KM containing L chain genes amplified with a single "promiscuous" V kappa oligonucleotide primer. The KM1 L chain is encoded by a mutated B3 gene (V kappa IV family). Another mutated B3 L chain had been identified previously in a TPO-specific Fab (WR1.223) isolated from a different patient (WR). In contrast to patient KM, the WR L chains were amplified with a panel of V kappa family-specific primers. Both KM1 and WR1.223 bind TPO with high affinity (approximately 1 x 10(-9) M) and interact with an epitope in the B domain of the TPO immunodominant region. TPO-specific Fab previously isolated from a WR combinatorial library constructed with the promiscuous V kappa primer recognised the TPO A domain and none used a B3-like L chain. Remarkably, for both patients, Fab isolated from L chains generated with the promiscuous V kappa primer had epitopic profiles similar to autoantibodies in the donor's serum (KM-B domain; WR-A domain). Our data indicate that the promiscuous primer preferentially amplifies the dominant L chain present in vivo. However, to obtain a relatively rare Fab (such as the B domain Fab from WR), family-specific kappa primers are required. These findings provide insight into the relationship between TPO autoantibody gene usage, epitopic recognition, and the effectiveness of the combinatorial library approach.

Engineering the human thyrotropin receptor ectodomain from a non-secreted form to a secreted, highly immunoreactive glycoprotein that neutralizes autoantibodies in Graves' patients' sera

Previous attempts to generate autoantibody-reactive, secreted thyrotropin receptor (TSHR) ectodomain in mammalian cells have failed because of retention within the cell of material with immature carbohydrate. We have overcome this difficulty by performing progressive carboxyl-terminal truncations of the human TSHR ectodomain (418 amino acid residues including signal peptide). Three ectodomain variants (TSHR-261, TSHR-289, and TSHR-309) were truncated at residues 261, 289, and 309, respectively. Unlike the full ectodomain, ectodomain variants were secreted with an efficiency inversely proportional to their size. Secreted ectodomain variants contained approximately 20 kDa of complex carbohydrate. TSHR-261 was chosen for further study because it was secreted very efficiently and neutralized autoantibodies in Graves' patients' sera. This ectodomain variant was partially purified using sequential lectin and nickel-chelate chromatography, permitting the first direct visualization and quantitation of the mammalian TSHR. Most important, very small (nanogram) quantities of this material neutralized 70-100% of TSHR autoantibody activity in all 18 Graves' sera studied. In summary, carboxyl-terminal truncation of the human TSHR ectodomain generates a secreted protein with complex carbohydrate that neutralizes autoantibodies in Graves' patients' sera. Antigenically active TSHR will be valuable for future studies on the diagnosis, pathogenesis, and immunotherapy of Graves' disease.

The human thyrotropin (TSH) receptor in a TSH binding inhibition assay for TSH receptor autoantibodies

Seven years after the molecular cloning of the human TSH receptor (TSHR), the porcine TSHR remains in general use in the TSH binding inhibition (TBI) assay for autoantibodies to the TSHR. We compared porcine and recombinant human TSHR in two types of TBI assays: one using intact Chinese hamster ovary cells expressing the recombinant human TSHR on their surface, and the other using soluble receptors extracted from these cells with detergent. In the intact cell TBI assay, monolayers expressing large numbers of TSHR were less effective than cells expressing few receptors. These findings are consistent with the very low concentration of TSHR autoantibodies in serum. Binding of [125I]human TSH was about 5-fold lower than that of [125I]bovine TSH to the intact cells. Nevertheless, TBI values with the two ligands were similar for most sera. However, a few sera produced greater inhibition of human than of bovine TSH binding. In the solubilized human TSHR TBI assay, in contrast to the intact cell TBI assay, cells expressing very large number of TSHR were an excellent source for detergent extraction of soluble human TSHR, but only if the cells were extracted while still on the dish and not after scraping. A 10-cm diameter dish of cells provided TSHR for 100-200 replicate determinations when substituted for solubilized porcine TSHR in a commercial TBI kit. TBI values in serum from 30 individuals with suspected Graves' disease correlated closely when tested with solubilized human and porcine TSHR (r = 0.954; P < 0.001). However, 2 sera that were negative with the porcine TSHR were positive with the human TSHR. TBI and thyroid-stimulating activity in these sera correlated weakly regardless of whether the TBI used human or porcine TSHR. These findings open the way to a practical TBI assay using recombinant human TSHR.

Evidence that the thyrotropin receptor ectodomain contains not one, but two, cleavage sites

TSH receptor (TSHR) cleavage into two subunits (A and B) was explored using two new mammalian cell lines expressing the recombinant receptor; 1) TSHR-10,000 CHO cells overexpressing the TSHR; 2) TSHRmyc cells with a c-myc epitope inserted at residues 338-349. Immunoprecipitation or immunoblotting of TSHR-10,000 cells with mAb to either the A subunit or the B subunit revealed multiple forms of the TSHR: 1) uncleaved receptors of approximately 115 kDa and approximately 100 kDa with complex carbohydrate and high mannose carbohydrate, respectively; 2) two subunit TSHR with an approximately 62 kDa A subunit containing complex carbohydrate. The A subunit was approximately 35 kDa after enzymatic deglycosylation (predicted C-terminus near residue 330). The nonglycosylated B subunit was evident primarily as an approximately 42 kDa band (predicted N terminus near residue 380). The sum of the A and B subunit polypeptide backbones was smaller than the predicted size of the TSHR, a polypeptide backbone (84.5 kDa), raising the possibility that an approximately 5-kDa polypeptide fragment was excised during intramolecular cleavage. This hypothesis was supported by data obtained with the TSHRmyc cells. Thus, mAb to the c-myc epitope and to amino acid residues 22-35 (mAb A10) were equally effective in detecting the single chain forms of the TSHR in these cells. However, the 35 kDa, deglycosylated A subunit was clearly visible on immunoprecipitation with mAb A10 to the TSHR amino terminus, but not with the anti-myc mAb, indicating loss of the c-myc epitope at residues 338-349. Further, even though the A subunit was not detected in TSHRmyc cells with anti-myc mAb, 125I-TSH cross-linking to the cell surface showed similar A subunit expression in TSHRmyc and wild-type TSHR expressing cells. In summary, our study provides a surprising and novel finding for G protein-coupled receptors. Contrary to the prevailing concept of one cleavage site in the TSHR, we present evidence that there are, in fact, two such sites. The TSHR, like insulin, may release a C peptide during intramolecular cleavage into two subunits.

Recombinant thyroid peroxidase-specific Fab converted to immunoglobulin G (IgG) molecules: evidence for thyroid cell damage by IgG1, but not IgG4, autoantibodies

A recombinant autoantibody Fab (SP1.4) to thyroid peroxidase (TPO), cloned from intrathyroidal B cell immunoglobulin genes, interacts with an epitope on TPO recognized by all patients with autoimmune thyroid disease. To compare the biological properties of IgG1 and IgG4 TPO autoantibodies, we converted Fab SP1.4 to full-length immunoglobulins. The SP1.4 heavy and kappa light chain variable region genes, spliced by overlap PCR to a mammalian signal peptide, were transferred to expression vectors for human IgG1, IgG4, and kappa L chains. Plasmids containing the IgG1 (or IgG4) heavy chain and the kappa L chain were cotransfected into SP2/0 mouse myeloma cells. Cells secreting TPO autoantibodies were cloned, and IgG1-SP and IgG4-SP were affinity purified from medium using protein G. Their subclass specificities were confirmed by enzyme-linked immunosorbent assay and fluorometry after binding to Chinese hamster ovary cells expressing cell surface TPO. Further confirmation of SP1.4 Fab conversion to full-length molecules was the ability of protein A to precipitate IgG1-SP and IgG4-SP complexed to [125I]TPO. IgG1-SP1.4, IgG4-SP1.4, and Fab SP1.4 had similar high affinities for TPO (Kd = approximately 2 x 10(-10) mol/L). Complexes of [125I]TPO and IgG1-SP (but not IgG4-SP) bound to peripheral blood mononuclear cells (PBMC), but not to a B cell line. Flow cytometry demonstrated Fc receptors Fc gamma RI, Fc gamma RII, and Fc gamma RIII on PBMC, but only Fc gamma RII on the B cell line. Together, these data indicate that IgG1-SP/TPO complexes bind to either Fc gamma RI on monocytes or RIII on natural killer cells. In assays for antibody-dependent cytotoxicity using PBMC, 51Cr release was higher for thyroid cells preincubated with IgG1-SP (13.4%) than with IgG4-SP (2.5%) or with culture medium alone (-0.7%). No specific 51Cr release was observed when either fibroblasts or Chinese hamster ovary cells expressing cell surface TPO were used as target cells. In conclusion, a human TPO-specific Fab converted to IgG1, but not IgG-4, can mediate cytotoxic effects on human thyroid cells in vitro. These observations support the clinical relevance of TPO autoantibody subclass distribution and emphasize the likelihood that, as opposed to being simple markers of thyroid damage, TPO autoantibodies may play a role in the induction of thyroid dysfunction in vivo.

Thyrotropin receptor autoantibodies in serum are present at much lower levels than thyroid peroxidase autoantibodies: analysis by flow cytometry

Using Chinese hamster ovary (CHO) cells that express high numbers of TSH receptor (TSHR) on their surface, we studied the feasibility of detecting directly by flow cytometry the binding of autoantibodies in patients' sera to the native TSHR. After using a serum (BBI) with high potency in the TSH binding inhibition (TBI) assay to establish the protocol, we studied an additional 38 sera: 10 without TBI activity (1-4.2% inhibition), 10 with moderately high TBI values (17.3-39.4% inhibition), 10 with high TBI levels (52-95.1% inhibition), 4 from normal individuals without autoimmune thyroid disease, and 4 from patients with systemic lupus erythematosus. We observed that a number of sera, including some without thyroid autoantibodies, contain antibodies against unknown antigens on CHO cells. Preadsorption with untransfected CHO cells before addition to the TSHR-10,000 cells eliminated or greatly reduced this nonspecific background. None of the sera from normal individuals, subjects with negative TBI values, or patients with systemic autoimmunity generated a positive signal on flow cytometry with TSHR-10,000 cells relative to the signal on untransfected cells. Remarkably, only 4 of 21 TBI-positive sera (including BBI) unequivocally recognized the TSHR on flow cytometry. In contrast, when thyroid peroxidase (TPO) autoantibodies in the same sera were studied using CHO cells overexpressing TPO on their surface, all 20 sera with TPO autoantibodies clearly elicited positive net fluorescence relative to untransfected cells. Study of the potent serum, BBI, revealed similar fluorescence (approximately 250 U) for TPO autoantibodies and TSHR autoantibodies at dilutions of 1:1000 and 1:10, respectively. Thus, by flow cytometry, the titer of TPO autoantibodies in the BBI serum is about 100-fold higher than that for TSHR autoantibodies in the same serum. In conclusion, the present data provide the strongest support for the idea that TSHR autoantibodies in the sera of patients with autoimmune thyroid disease are present at much lower levels than are TPO autoantibodies. This finding has important implications for the diagnostic detection of TSHR autoantibodies and for understanding the pathogenesis of Graves' disease.

Thyroid peroxidase autoantibodies of IgE class in thyroid autoimmunity

Recently, we converted a thyroid peroxidase (TPO)-specific human autoantibody Fab (SP1.4) into an IgE molecule (IgE-SP1.4) which permits antigen capture via Fc epsilon receptors on B cells and presentation to T cells. An important question which arose was whether IgE class TPO autoantibodies are present in vivo. By ELISA, TPO autoantibodies of IgG1 and IgG4 subclasses, but not IgE, were readily detectable in patients' sera. However, such negative data were not definitive because high concentrations of IgG class TPO autoantibodies could obscure the presence of much lower concentrations of IgE class autoantibodies. We, therefore, established a specific assay based on IgE "capture" to remove other isotypes before incubation with TPO. In an initial survey, 125I-TPO binding was higher in sera from 16 patients with autoimmune thyroid disease than in 6 controls (8.4 +/- 0.8% versus 0.7 +/- 0.2%; mean +/- SEM). Unlabeled TPO (10(-8) M) inhibited 125I-TPO binding by patients' (but not controls') IgE. Further, TPO binding by IgE-SP1.4 was unaffected by IgG class TPO autoantibodies. Titers of IgE class TPO autoantibodies were low, detectable at a 1/60 dilution in 4/5 sera studied. In a larger series, IgE class TPO autoantibodies were present in 13 of 18 Graves' and in 12 of 17 Hashimoto patients (sera diluted 1/6). Sera were considered to be positive with TPO binding greater than the mean + 3 SD of values for 23 control sera (1.8%). In conclusion, we provide the first evidence for TPO autoantibodies of IgE class in patients with autoimmune thyroid disease. Because of their low concentration, these autoantibodies are unlikely to play a role in antigen presentation in vivo. However, their presence strengthens the link between autoimmune thyroid disease and immune responses involving TH2 cells.

Insight into screening immunoglobulin gene combinatorial libraries in a phage display vector: a tale of two antibodies

Combinatorial libraries of immunoglobulin genes in "phage display" vectors are a powerful tool for obtaining antigen-specific antibody fragments. To date, this approach has been used to isolate abundant, but not rare, human autoantibodies of IgG class. We have compared the relative efficiencies of panning pComb3 libraries made from intrathyroidal plasma cells for abundant human autoantibodies to thyroid peroxidase (TPO) and rare autoantibodies to the thyrotropin receptor (TSHR). TPO-specific Fab were readily obtained from a library using three different forms of recombinant antigen, (i) purified TPO, (ii) impure TPO in culture medium and, (iii) TPO expressed on the surface of CHO cells. In contrast, TSHR-specific Fab were not isolated. This was the case despite repeated pannings of six libraries from three optimal patients (IgG/kappa and IgG/lambda libraries for each patient). Both purified recombinant TSHR and CHO cells expressing TSHR on their surface were used. Library enrichment was observed on some screenings. However, Fab expressed by individual clones or from enriched libraries were not specific as determined by (i) binding to purified, radio-labeled antigen, (ii) FACS analysis of TSHR on intact CHO cells and, (iii) inhibition of radiolabeled TSH binding. Remarkably, in screening for both TPO- and TSHR-specific Fab, neither library enrichment nor the retention of cDNA inserts of the correct size correlated with obtaining Fab with the antigenic specificity sought. Indeed, excellent enrichment could be observed with conditioned medium from untransfected cells. Our data suggest that the key to isolating rare antibodies from phage display libraries is not the creation of vast libraries of greater diversity or even the development of more stable vectors. Rather, success in this endeavor appears to require reducing the "noise" of non-specific clones in a moderately sized library.

Evidence for negative cooperativity among human thyrotropin receptors overexpressed in mammalian cells

The complementary DNA for the human TSH receptor (TSHR) translated region was amplified in the genome of stably transfected Chinese hamster ovary (CHO) cells using a dihydrofolate reductase minigene. Immunoprecipitation of TSHR in whole cells precursor-labeled with [35S]methionine and [35S]cysteine revealed an approximately 10-fold increase in TSHR expression in cells stabilized in 10,000 nM methotrexate (TSHR-10,000 cells) compared to cells with the same gene not subjected to amplification (TSHR-0 cells). Similarly, [125I]TSH cross-linking to the surface of intact CHO cells revealed a progressive increase in TSH-binding sites with dihydrofolate reductase minigene amplification, with a 12.8-fold increase in TSHR in TSHR-10,000 vs. TSHR-0 cells. Based on the known number of TSHR expressed by TSHR-0 cells, TSHR-10,000 express approximately 1.9 x 10(6) TSHR on their surface. Two ligand-TSHR complexes were evident under reducing conditions, representing the single chain holoreceptor of about 115 kDa and a dissociated A subunit of about 60 kDa. In the absence of TSH, basal cAMP levels in TSHR-10,000 cells were greater than those in TSHR-0 cells (6-fold in isotonic medium and 18.5-fold in hypotonic medium), indicating that the unliganded TSHR has significant constitutive activity. We assessed the kinetics of TSH binding to CHO cells overexpressing the TSHR using [125I]TSH in the presence of increasing concentrations of unlabeled TSH as well as by attempted saturation with labeled ligand. Surprisingly, in contrast to TSHR-0 cells (Kd = approximately 5 x 10(-10) M), we observed progressively lower affinities for TSH binding by TSHR-800 cells (Kd = approximately 10(-9) M) and TSHR-10,000 cells (Kd = approximately 2 x 10(-9) M). In summary, we report a high level of expression of TSHR in CHO cells and confirm the high constitutive activity of the TSHR in the absence of ligand as well as the binding of TSH to the single subunit, uncleaved TSHR. Moreover, we found that a high level of expression is associated with apparent negative cooperativity among the TSHR in terms of their affinity for ligand.

Autoantibody-mediated capture and presentation of autoantigen to T cells via the Fc epsilon receptor by a recombinant human autoantibody Fab converted to IgE

Fc epsilon receptor (CD23)-mediated capture of IgE-antigen complexes by B cells provides a powerful antigen presenting system. Our goal was to develop a system using high affinity, human, organ-specific monoclonal autoantibodies for antigen capture by B cells. For this purpose, we converted a recombinant human autoantibody to TPO from a Fab (SP1.4) to an IgE molecule. Sera from all patients with autoimmune thyroid disease contain autoantibodies with the same epitope as SP1.4. The SP1.4 H and L chain V region genes were spliced by overlap PCR to a mammalian, non-immunoglobulin signal peptide and transferred to expression vectors for human IgG1 and kappa, respectively. After inserting the IgE constant region genes into the H chain vector, the kappa and IgE H chain vectors were expressed in SP2/0 cells. SP1.4-IgE retains its high affinity (Kd) for TPO (approximately 2 x 10(-10) M), recognizes the same epitope as Fab SP1.4 and, importantly binds to a different epitope than does Fab TR1.9. Binding of preformed complexes of SP1.4-IgE and biotinylated TPO to EB virus transformed B cells (EBVL) was weakly detectable by flow cytometry and was displaced by unlabeled TPO. SP1.4-IgE/125I-TPO complex binding to EBVL was much more clearly evident, was also inhibited by the addition of unlabeled TPO, and was greatly reduced by preincubation of the EBVL with anti-CD23. Further, autologous EBVL preincubated with SP1.4-IgE/TPO complexes stimulated proliferation of TPO-specific T cells. IgE autoantibody-mediated antigen focusing to B cells is unlikely to operate in vivo but is, instead, a powerful investigative tool. In conclusion, SP1.4-IgE is the first monoclonal human autoantibody to be developed for IgE-mediated antigen presentation to T cells by EBVL. Recombinant human autoantibodies converted to IgE, possibly in combinations if their epitopes permit simultaneous binding to the same molecule, provide a unique system to generate human T cell lines and clones specific for peptides naturally processed from internalized high affinity autoantibody/autoantigen complexes.

Is receptor cleavage into two subunits necessary for thyrotropin action?

Unlike the wild-type thyrotropin (TSH) receptor, the chimeric TSH-LH/CG receptor TSH-LHR-14 does not cleave into two subunits when cross-linked to [125I]TSH on the surface of intact cells. Immunoblotting of TSH-LHR-14 in whole cell homogenates demonstrated that only a single chain receptor was detected under reducing conditions. TSH-LHR-14, like the A subunit of another chimeric receptor (TSH-LHR-10) that does cleave into two subunits, was almost entirely resistant to endoglycosidase H, indicating that it contains predominantly complex carbohydrate. The fact that TSH-LHR-14 reaches the cell surface where it binds TSH with high affinity and transduces a signal indicates that receptor cleavage into two subunits is not a prerequisite for TSH action.

Both the 5' and 3' noncoding regions of the thyrotropin receptor messenger ribonucleic acid influence the level of receptor protein expression in transfected mamalian cells

The molecular basis for the difference in the bioresponsiveness of TSH receptor cell lines from two different laboratories has been investigated. We modified our 4-kb TSH receptor complementary DNA (cDNA) by deleting either the 5' untranslated region (UTR), the 3'UTR, or both UTRs. The 5'UTR contains two false AUG initiation codons followed by a stop codon. The cDNAs in the eukaryotic expression vector pSV2-NEO-ECE, as well as the 5'3'UTR-truncated cDNA in pSVL, were stably transfected into Chinese hamster ovary cells. Pools of more than 100 colonies were studied in order to minimize insertion site-dependent variation in the level of expression. Scatchard analysis of TSH binding indicated that the number of receptors on the surface of Chinese hamster ovary cells expressing the wild-type transcript (approximately 16,000/cell) increased approximately 2-fold with 5'UTR deletion, approximately 5-fold with 3'UTR deletion, and approximately 10-fold with both 5'UTR and 3'UTR deletion. TSH binding affinities of all constructs were in the range of 2-5 x 10(-10) M. No significant difference was evident between the 5'3'UTR truncated cDNAs in the two different vectors, pSV2-NEO-ECE and pSVL. The increase in the amplitude of the cAMP response to TSH stimulation was commensurate with the number of receptors expressed on the surface of the different cell lines. Truncation of the 5'UTR did not alter TSH receptor messenger RNA (mRNA) levels relative to the wild-type mRNA. In contrast, the level of the 3'UTR-truncated transcript, as well as the 5'3'UTR-deleted transcript, increased approximately 4-fold independent of the expression vector used. In summary, both the 5'UTR and 3'UTR of the human TSH receptor mRNA influence the level of receptor expression on transfected mammalian cells. In particular, the 3'UTR has a destabilizing influence on the MRNA. These data explain the greater level of TSH receptor expression in cell lines that are transfected with cDNA lacking these regions of the mRNA transcript.

Critical relationship between autoantibody recognition and thyrotropin receptor maturation as reflected in the acquisition of complex carbohydrate

Generation of large amounts of recombinant TSH receptor (TSHR) protein capable of recognition by TSHR autoantibodies is a goal of clinical importance. We expressed in Chinese hamster ovary cells the human TSHR ectodomain (ECD) with a carboxyl-terminus six-histidine tag. After transgene amplification, expressing clones were selected by nickel chelate chromatography in combination with [35S] methionine precursor labeling. An approximately 74-kDa protein was detected in the culture medium, and larger quantities of an approximately 68-kDa protein were found in the cell soluble fraction. Immunoblot analysis with a rabbit antiserum revealed that most of the TSHR-ECD was not secreted, but was retained within the soluble fraction of the cell. Nickel chelate chromatography was not effective in purifying significant quantities of this material. In contrast, with Concanavalin A, but not with wheat germ agglutinin, an approximately 50-fold purification of TSHR-ECD was achieved from the cell soluble fraction. Surprisingly, this affinity-enriched TSHR, containing high mannose carbohydrate, was not recognized by human TSHR autoantibodies in sera from six individuals. By ion exchange chromatography, the autoantibody-neutralizing TSHR in the cell supernatant fraction was found to be nonidentical with TSHR-ECD protein recognized by antisera from immunized animals. The present data indicate the critical relationship between autoantibody recognition and TSHR maturation as reflected in the acquisition of complex carbohydrate. Nonsecretion of the TSHR-ECD appears to be related to the specific protein rather than to the glycosylation apparatus of the host cell. Antibodies from immunized animals may be ineffective in monitoring purification of autoantigen-competent TSHR. Finally, the data explain why soluble recombinant TSHR generated in many expression systems is not recognized satisfactorily by human autoantibodies.

Structural studies of human autoantibodies. Crystal structure of a thyroid peroxidase autoantibody Fab

The three-dimensional structure of the Fab of TR1.9, a high-affinity IgG1, kappa human autoantibody to thyroid peroxidase, was determined crystallographically to a resolution of 2.0 A. The combining site was found to be relatively flat, like other antibodies to large proteins. Sequence differences from the most closely related germline genes mainly occur at positions occupied by residues with outward-pointing side chains. An increased deformability of the second and third complementarity-determining regions of the heavy chain may result from the replacement of two germline asparagines and the presence of several glycines, and may allow "induced fit" in the binding to antigen. Four exposed charged residues, resulting from the use of a particular D (diversity) and J (joining) segments in the assembly of the heavy chain, may contribute to the high affinity of antigen binding. The crystal structure of TR1.9 Fab is the first for a human IgG high-affinity autoantibody.

Thyroid peroxidase autoantibody epitopic 'fingerprints' in juvenile Hashimoto's thyroiditis: evidence for conservation over time and in families

In Hashimoto's thyroiditis, the humoral component is manifest by autoantibodies to thyroid peroxidase (TPO). Epitopic 'fingerprinting' of polyclonal serum TPO autoantibodies has been facilitated by the molecular cloning and expression as Fab of a repertoire of human TPO autoantibody genes. To investigate whether TPO autoantibody fingerprints are (i) stable over long periods of time (approximately 15 years), and (ii) inherited, we studied a cohort of nine patients with juvenile Hashimoto's thyroiditis and 21 first degree relatives of four of these patients. Fingerprints were determined by competition between four selected FAB and serum autoantibodies for binding to 125I-TPO. Regardless of titre, the TPO epitopic profile was stable in 10/12 individuals whose TPO autoantibody levels were sufficient for analysis on two or three occasions over 12-15 years. Although the TPO epitopic fingerprint profiles in two families raised the possibility of inheritance, overall the data from all four families did not reveal an obvious pattern of genetic control. In no family was the TPO epitopic fingerprint associated with HLA A, B or DR. In conclusion, TPO autoantibody epitopic fingerprints are frequently conserved over many years. Studies on additional families are necessary to establish whether or not the epitopic profiles of TPO autoantibodies are inherited.

The quest for the autoantibody immunodominant region on thyroid peroxidase: guided mutagenesis based on a hypothetical three-dimensional model

Thyroid peroxidase (TPO), a cell surface glycoprotein, is the major autoantigen in autoimmune thyroiditis in humans. The molecular cloning and expression of Ig genes from thyroid-infiltrating B cells has generated a large repertoire of human TPO Fab that have been used to map an immunodominant region on TPO. However, the topological site of this region, consisting of a cluster of highly conformational epitopes, remains unknown. Using the recently elucidated three-dimensional structure of myeloperoxidase as a model, we stably expressed on the surface of eukaryotic cells eight "guided" mutants of the TPO molecule. The sites of these mutations were strategically located to alter the surface contour of the molecule with minimal disruption to its core structure. Remarkably, in the present study (in contrast to previous unguided TPO mutagenesis studies), all eight TPO mutants retained recognition by the TPO Fab. These results support the validity of the model used for mutagenesis. Although not identifying the immunodominant region on TPO in thyroid autoimmunity, our data provide evidence against the involvement of certain topological segments and may help to narrow the search for this region. The most open region remaining as a candidate location is the antero-inferior portion of the molecule.

Influence of the light chain repertoire on immunoglobulin genes encoding thyroid autoantibody Fab from combinatorial libraries

The diversity of the immunoglobulin heavy (H) and light (L) gene libraries used to construct a combinatorial library is an important parameter in determining the characteristics of antigen-specific Fab obtained from the library. To investigate the role of library diversity, we compared Fab specific for the autoantigen thyroid peroxidase (TPO) isolated from two different combinatorial libraries. Both libraries contained the same H chain genes. The original combinatorial library (H/R) utilized kappa chains generated using a single kappa variable region oligonucleotide primer. We constructed a second combinatorial library (H/D) containing kappa chains amplified with a diverse panel of variable region primers. From the the original H/R library, only two groups of TPO-specific Fab had been obtained, involving two H chain types (V1-3B and hv1L1) but only one kappa chain type (012). In contrast, among the seven TPO Fab characterized from the second library (H/D) we observed five different VH/VL combinations, comprising three types of H chains (V1-3B, VH26 and DP7) and four types of kappa chains (O12, L12, L2/hv328H5 and B3). Besides differences in VH and VL genes, as well as VH/VL combinations, the new TPO Fab used different D regions and JH and JK elements. Nevertheless, the new kappa Fab resembled previously isolated TPO Fab in terms of their affinity for TPO (Kd approximately 10(-9)M) and preferential recognition of conformationally intact autoantigen. In summary, our studies demonstrate that the diversity of the L chain library repertoire, while having little effect on immunological properties, has a major influence on the genes encoding antigen-specific Fab selected from a combinatorial library. For the successful isolation of rare but clinically important autoantibodies (such as to the TSH receptor) by the combinatorial library approach, library diversity is likely to be a major factor.

Molecular cloning and characterization of human thyroid peroxidase autoantibodies of lambda light chain type

IgG class thyroid peroxidase (TPO) autoantibodies with kappa light (L) chains predominate in serum and the genes for a large repertoire of such autoantibodies have been characterized. The present study was performed to clone and characterize TPO autoantibodies with lambda L chains which comprise approximately 20% of serum TPO autoantibodies. From a combinatorial IgG H/lambda L chain cDNA library in the phage display vector pComb3, 24 TPO-binding clones with lambda L chains were isolated, comprising three different heavy (H) and light (L) chain combinations. These combinations utilized two genes from the Vlambda II and IIIb families (closest germline genes DPL11 and hsigg11150) and three genes from the VH1, VH3 and VH4 families (VH26, 4.34 and hv1L1). The deduced amino acid sequences of these H chains were quite different from those of kappa F(ab) isolated using the same H chain library. We expressed the proteins for these three lambda F(ab), as well as for a lambda F(ab) (Humlv318 L chain/DP10-like H chain) previously isolated from another patient. The affinities for TPO of the lambda F(ab) (Kd 8 x 10(-10) M to 10(-7) M) were lower than those of the kappa F(ab) (Kd approximately 10(-10) M). For two lambda F(ab), both H and L chain genes were close to germline configuration, but there was no straightforward relationship between the extent of somatic mutation from germline configuration and affinity for TPO. All four lambda F(ab) bound less well to denatured TPO as to native TPO. The three F(ab) for which sufficient protein could be expressed for competition studies all recognized domain B within the immunodominant region on TPO previously identified using F(ab) with kappa L chains. Aside from these TPO-specific F(ab), only a few other human IgG class, organ-specific autoantibodies with lambda L chains have been characterized at the molecular level. Our study significantly augments the small database on this category of autoantibodies in general.

Genetic and epitopic analysis of thyroid peroxidase (TPO) autoantibodies: markers of the human thyroid autoimmune response

TPO autoantibodies, the hallmark of human autoimmune thyroid disease, are of IgG class and are associated with thyroid destruction and hypothyroidism. Using the immunoglobulin gene combinatorial library approach, a panel of human monoclonal TPO autoantibodies (expressed as Fab) has been generated from thyroid tissue-infiltrating B cells. TPO-specific Fab closely resemble patients' serum autoantibodies in terms of L chain type, IgG subclass, affinities for TPO as well as epitopes recognized by > 80% of TPO autoantibodies in an individual's serum. TPO autoantibody V region genes are not unique; H chain V genes are usually mutated, while L chain V genes are sometimes in germ-line conformation. The autoantibodies recognize an immunodominant region involving conformational, overlapping epitopes in domains A and B. Finally, TPO autoantibody epitopic fingerprints are distinctive for individual sera, are not associated with hypothyroidism, but are conserved over time (indicating a lack of B cell epitope spreading). Evidence for conservation as well as inheritance of the fingerprints in some families, together with VH gene polymorphisms, may provide insight into the genetic basis of human autoimmune thyroid disease. Furthermore, monoclonal human TPO autoantibodies will be invaluable for B cell presentation of TPO to determine the T cell epitopes involved in TPO autoantibody production.

Human monoclonal autoantibodies against the immunodominant region on thyroid peroxidase: lack of cross-reactivity with related peroxidases or thyroglobulin and inability to inhibit thyroid peroxidase enzymatic activity

Thyroid peroxidase (TPO) autoantibodies are heterogeneous and have been classified in terms of whether they cross-react with myeloperoxidase (MPO), lactoperoxidase (LPO), or thyroglobulin (Tg) as well as by whether they inhibit TPO enzymatic activity. Four human monoclonal TPO autoantibodies, generated using combinatorial immunoglobulin gene libraries and expressed as F(ab), have been used to investigate these properties of TPO autoantibodies. The binding of F(ab) WR1.7, TR1.8, TR1.9, and SP1.4 to 125I-labeled recombinant TPO was inhibited 50% by approximately 10(-10) mol/L unlabeled TPO, reflecting the high affinities of these F(ab) for TPO. In contrast, F(ab) binding to TPO was unaffected by human MPO (both native and reduced), bovine LPO, or human Tg at concentrations up to 10(-8) mol/L. Further, TPO enzymatic activity, measured by guiacol oxidation, was unaffected by preincubation with the four F(ab) individually or as a pool (each at 10(-8) mol/L). In conclusion, four human TPO monoclonal autoantibodies do not cross-react with related peroxidases or Tg, nor do they inhibit TPO enzymatic activity. These monoclonal immunoglobulin G class autoantibodies define the immunodominant region on TPO and represent about 85% of TPO autoantibodies in an individual patient's serum. Consequently, our data suggest that TPO autoantibodies that cross-react with MPO, LPO, or Tg, or inhibit TPO enzymatic activity are likely to bind outside the immunodominant region.

Thyroid peroxidase autoantibody fingerprints. II. A longitudinal study in postpartum thyroiditis

It is not known whether epitopes recognized by autoantibodies in an individual remain constant or change over time, especially during perturbations of the humoral immune response. To address this question, we studied the epitopic profile ("fingerprint") of autoantibodies to thyroid peroxidase (TPO) in the sera of 19 women during the postpartum period. Fingerprints were determined in competition studies using 4 recombinant F(ab). At delivery and at 3 time intervals over the subsequent 9-12 months, the pool of F(ab) inhibited autoantibody binding to TPO by 80-100%, consistent with the definition by these F(ab) of a TPO immunodominant region (A1, A2, B1, and B2 domains). Despite a wide spectrum among individuals, the TPO epitopic fingerprints for all 19 women were relatively unchanged throughout the postpartum period. Fingerprint constancy occurred regardless of fluctuations in serum TPO autoantibody levels. When assessed numerically as a ratio of inhibition by the A domain F(ab) to inhibition by the B domain F(ab), the A/B domain ratios in individual women ranged from 0.2 (predominantly B domain) to more than 3.0 (predominantly A domain). However, for each individual woman, the A/B epitopic ratio was conserved throughout the study interval. Our TPO autoantibody epitopic fingerprint data have potential implications for understanding the humoral autoimmune response in man. First, the present study indicates a remarkable lack of spreading of B cell epitopes during a state of perturbation of the immune system over a period of 1 yr. Second, and perhaps more important, despite marked variations in TPO epitopic profiles among different individuals, their constancy over time suggests that TPO autoantibody fingerprints may be inherited.

Thyroid peroxidase autoantibody fingerprints in hypothyroid and euthyroid individuals. I. Cross-sectional study in elderly women

Human monoclonal immunoglobulin G-class autoantibodies to thyroid peroxidase (TPO), expressed as recombinant F(ab), are powerful tools for analyzing the individual components of polyclonal serum TPO autoantibodies. Four TPO-specific F(ab) interact with epitopes in two closely related domains (A and B) in the immunodominant region on TPO. In the present study, these TPO F(ab) were used to compete for serum autoantibody binding to [125I]TPO to determine the "epitopic fingerprints" in two groups of carefully controlled individuals. All individuals (14 hypothyroid and 32 euthyroid) were elderly women (60-71 yr old) with similar genetic and environmental backgrounds as well as comparable levels of serum TPO autoantibodies. Using the pool of four F(ab), serum TPO autoantibody binding was inhibited to the same extent (approximately 90%) in hypothyroid and euthyroid individuals, demonstrating that the majority of TPO autoantibodies in both groups recognize the TPO immunodominant domain. When tested individually, the F(ab) produced a spectrum of inhibition patterns, ranging from sera preferentially inhibited by domain A F(ab) to sera preferential inhibited by domain B F(ab). The ratio of inhibition by domain A F(ab) to inhibition by domain B F(ab) was similar in hypothyroid (0.11-1.39) and euthyroid (0.21-1.79) women. In conclusion, no difference in TPO autoantibody epitopes was observed in this cross-sectional study of hypothyroid and euthyroid individuals. Longitudinal studies are required to address the question of whether TPO autoantibody epitopic fingerprints are stable over time.

Combinatorial libraries: new insights into human organ-specific autoantibodies

The recent application of immunoglobulin (Ig) gene combinatorial library technology has led to a logarithmic increase in information concerning human, disease-associated, organ-specific autoantibodies of the IgG class. As reviewed here by Basil Rapoport, Stefano Portolano and Sandra McLachlan, the molecular cloning, analysis and expression of the genes for increasing numbers of these human, monoclonal autoantibodies is providing new insight into the genetic background and epitopic repertoires of such molecules.

Exclusion of two major areas on thyroid peroxidase from the immunodominant region containing the conformational epitopes recognized by human autoantibodies

We have used a chimeric molecule between thyroid peroxidase (TPO) and myeloperoxidase (MPO) as well as new information on the three-dimensional structure of MPO to refine further our understanding of the location of the TPO-immunodominant region recognized by TPO autoantibodies in patients' sera. In TPO-MPO chimera A, the amino-terminal 146 amino acids of MPO substitute for the amino-terminal 121 amino acids of TPO. We performed fluorescence-activated cell sorter analysis of Chinese hamster ovary cells expressing TPO-MPO-A on their surface using four monoclonal human autoantibody F(ab) (WR1.7, TR1.8, TR1.9, and SP1.4) that define the immunodominant region. All four F(ab) recognized the TPO-MPO-A chimeric molecule to the same extent. In a second approach to refine the location on the TPO-immunodominant region, we compared the ability of the TPO autoantibody F(ab) to inhibit the binding of serum autoantibodies to the monomeric and dimeric forms of human TPO. The F(ab) inhibited equally (approximately 80%) the binding to the TPO monomer and dimer by autoantibodies in the sera of six individual patients. The present observations exclude two major regions of TPO from the autoantibody-immunodominant region, namely the amino-terminal 121 amino acids of the TPO extracellular domain and the contact region between the two TPO monomers. These findings together with previous data on the Mab47/C21 region of TPO and the recently elucidated 3-dimensional structure of highly homologous MPO, narrow, by a process of exclusion, the site on TPO comprising the immunodominant region. The data provide further support for the thesis, still controversial, that the majority of TPO autoantibodies recognize the native molecule.

Profile of lambda light chain variable region genes in Graves' orbital tissue

Graves' ophthalmopathy, a human autoimmune disease of unknown etiology, is strongly associated with autoimmune hyperthyroidism. A major controversy is whether retro-ocular muscle or orbital fat/connective tissue is the target of the immune response. Previously, we observed preferential PCR amplification of lambda (relative to kappa) light chain DNA from cDNA of Graves' orbital tissue-infiltrating B cells/plasma cells. There is little information on V lambda gene usage in man and none in diseased tissue. To characterize the orbital lambda light chains, we constructed cDNA libraries using PCR-amplified DNA from three tissues and sequenced the variable region genes from randomly selected clones. Analysis of 27 clones from orbital fat/connective tissue libraries from two patients with acute inflammatory eye disease, and 15 clones from orbital muscle of one of these patients, revealed a diverse spectrum of lambda V region genes. The nucleotide sequences of these 42 clones were most homologous to 12 different germline genes: four family I (subfamilies I-a, -b and -c), three family II, two family III and one family VII germline genes. Each orbital tissue had a distinct profile of V lambda sequences. However, all clones used J lambda 2/3 and all three orbital tissues contained clones related to family II genes. Although some clones had V region sequences in near germline conformation, the majority differed from the closest germline gene in both framework and complementarity determining regions. Whether or not these differences result from multiple germline gene usage or somatic mutation of a smaller number of germline genes cannot be determined until information on the V lambda repertoire and its polymorphisms is complete. However, the V lambda gene diversity we observed in both orbital muscle and orbital fat/connective tissue suggests a role for lambda autoantibodies in the pathogenesis of Graves' ophthalmopathy.

Recombinant thyroid peroxidase-specific autoantibodies. I. How diverse is the pool of heavy and light chains in immunoglobulin gene libraries constructed from thyroid tissue-infiltrating plasma cells

Thyroid peroxidase (TPO) autoantibodies are a distinguishing feature of autoimmune thyroid disease. We have previously constructed immunoglobulin G heavy (H) and light (L) chain cDNA libraries from intrathyroidal B-cells. TPO-selected autoantibodies expressed by combined H and L chain libraries (combinatorial libraries) recognized a limited number of epitopes on TPO and used only a few of the many H and L chain variable region genes present in the genome (germline genes). One possible explanation for this restriction is a lack of diversity in the parental H and L chain gene libraries used to construct the combinatorial library. To address this issue, we determined the nucleotide sequences of randomly selected H and kappa L chain variable region genes from a pair of H and L chain libraries. The 12 H chain gene sequences analyzed were highly diverse, and none resembled the genes of TPO-selected autoantibodies. The sequences of 14 randomly selected kappa L chain genes were less diverse; 12 of 14 were closely related to the same germline gene (KL012) used by TPO-specific autoantibodies. However, we observed previously that only about 1 in 500 of the L chains in this library can pair with an H chain and bind TPO. We now find that, with 1 exception, the randomly selected KL012-like genes in the L chain library differ significantly from the antigen-specific KL012-like genes, particularly in the antigen-binding regions. In summary, the present data indicate that 1) the restricted number of H chain genes used by TPO-specific autoantibodies cannot be ascribed to limited H chain gene diversity in the parent library; and 2) L chains from combinatorial libraries (even when related to the same germline gene) cannot simply be regarded as plastic, or promiscuous, partners for high affinity antigen binding by a particular H chain.

Recombinant thyroid peroxidase-specific autoantibodies. II. Role of individual heavy and light chains in determining epitope recognition

Most thyroid peroxidase (TPO) autoantibodies in man recognize closely associated epitopes in two domains (A and B) on TPO. These epitopes were defined by recombinant monoclonal human autoantibodies expressed as antigen-binding fragments [F(ab)]. Only five heavy (H) and light (L) chain gene combinations encoded 34 F(ab), all of which have high affinity (Kd, approximately 10(-10) M) for TPO. We, therefore, investigated the roles of H and L chain genes in TPO domain recognition in two ways. First, we created hybrid F(ab) by forced recombination of H and L chain genes from 4 F(ab) recognizing the A or B domains. These hybrid F(ab) proteins, expressed in bacteria, bound extremely poorly (or not at all) to TPO, even at concentrations more than 100-fold higher than those required for detection of TPO binding by the original F(ab). Nucleotide sequencing of the cDNA as well as gel electrophoresis of the expressed proteins confirmed that poor hybrid F(ab) binding to TPO was not the result of cloning artifacts. Therefore, contrary to prevailing views on combinatorial libraries, we found no tolerance for H and L chain cross-combinations in high affinity TPO binding. These observations strengthen the likelihood that the H and L chain combinations from combinatorial libraries reflect those of TPO autoantibodies in vivo. In a second approach to examine the roles of H and L chains in TPO binding, we focused on three original F(ab) with similar L chains (encoded by KL012-like germline genes) and similar H chains (encoded by V1-3B-like germline genes), but different diversity (D) regions. All F(ab) bound predominantly to TPO domain A, as observed previously for a F(ab) with a KL012 L chain and a different H chain. Conversely, a F(ab) with a V1-3B-like H chain but a different L chain (A') bound to TPO domain B. These data indicate that the L chain plays a major role in defining TPO epitope recognition.

Isolation and characterization of a monoclonal human thyroid peroxidase autoantibody of lambda light chain type

Thyroid peroxidase (TPO) autoantibodies, a hallmark of human autoimmune thyroid disease, may have kappa or lambda light chains. Monoclonal human TPO autoantibodies with kappa light chains have previously been developed by cloning and expressing "combinatorial" libraries of immunoglobulin genes in bacteria. In the present study, an IgG1/lambda combinatorial library was generated from thyroid cDNA of a Graves' patient whose serum contained lambda TPO antibodies. Screening the bacteriophage library with 125I-TPO yielded one clone, TR1.41. The oligonucleotide sequence of TR1.41 was determined and the nature of its interaction with TPO was investigated. The affinity of TR1.41 for TPO is high (Kd approximately 10(-9) M), comparable to that of monoclonal kappa TPO autoantibodies derived from the same patient. The genes encoding the heavy and light chains of TR1.41 differ in a number of respects from the closest available germline genes. Such differences are consistent with somatic mutation in a high-affinity antibody. An important characteristic of TR1.41 is its interaction with the immunodominant domain on TPO recognized by approximately 80% of serum TPO autoantibodies. The frequency of TPO-specific F(ab) generated from the thyroid gland of patient TR was much lower for F(ab) with lambda light chains (1:150,000) than for F(ab) with kappa light chains (1:13,000). Despite this low frequency, the high affinity of TR1.41 and its recognition of the immunodominant region on TPO indicate that lambda autoantibodies of this type may represent an important constituent of the TPO autoantibody response in man. In conclusion, this is the first report on the molecular cloning and characterization of a thyroid autoantibody of lambda L chain type by the combinatorial library approach.

Cell-mediated or humoral immunity in Graves' ophthalmopathy? Profiles of T-cell cytokines amplified by polymerase chain reaction from orbital tissue

Whether Graves' ophthalmopathy is the results of a cell-mediated or humoral autoimmune response is controversial. T-Lymphocytes that regulate these two mechanisms, Th1 and Th2, respectively, are characterized by the profile of cytokines that they secrete. We, therefore, investigated the spectrum of T-lymphocyte cytokines expressed in Graves' orbital tissue by means of the polymerase chain reaction using reverse transcribed mRNA as template. From three Graves' thyroid tissues (positive controls), we obtained DNA products of the predicted size for the Th1 cytokines interferon-gamma (IFN gamma) and interleukin-2 (IL-2) as well as for Th2 cytokines IL-4 and IL-5. A signal for IL-10 (Th1 and Th2) was also detected. No cytokine signals were observed in peripheral blood (negative control). Despite the presence of T-cells (CD3 delta marker) in the orbital connective tissue/fat of all five patients studied as well as in orbital muscle from one of these patients, IFN gamma reverse transcribed mRNA was notably absent from these tissues. In contrast, IL-2, IL-5, and IL-10 cDNA was present in orbital tissues from one or more patients. Of particular importance, an IL-4 signal was detected in orbital connective tissue/fat of patients 6 and 7 as well as in muscle of patient 7. The balance between IL-4 and IFN gamma determines whether an immune response is predominantly humoral or cell mediated. Our finding of IL-4, but not IFN gamma, mRNA expression in orbital tissue supports a role, at least in some patients, for humoral autoimmunity in Graves' ophthalmopathy.

Immunoglobulin A in Graves' orbital tissue: deoxyribonucleic acid amplification by polymerase chain reaction

A role for IgA autoantibodies in Graves' ophthalmopathy is suggested by the presence of immunoglobulins of this class in Graves' orbital tissue, as detected by immunohistochemistry. We, therefore, investigated the possibility of using the polymerase chain reaction (PCR) to amplify IgA immunoglobulin genes from plasma cells infiltrating Graves' eye tissue. Template cDNA was reverse-transcribed from orbital muscle (M) mRNA of one patient (#7) and from orbital connective tissue/fat (F) mRNA of two patients (#1 and #7), both undergoing surgery for exophthalmos because of severe infiltrative ophthalmopathy. Preliminary studies to establish the PCR procedure were performed for kappa light chain DNA amplification. With the very small amount of orbital tissue template available, the sensitive "hot start" modification of the PCR was necessary to amplify significant amounts of kappa light chain DNA. Using this procedure, IgA heavy chain DNA was amplified from both connective tissue/fat (F7) and muscle (M7) cDNA of patient #7. The DNA yield was less for IgA than for IgG using the same template. There was no significant IgA (or IgG) DNA product using the connective tissue/fat cDNA of patient #1. While not implying that IgA-infiltrating plasma cells are specific for Graves' orbital tissue, our studies nevertheless demonstrate the feasibility of amplifying the genes coding for IgA antibodies from Graves' orbital tissue plasma cells. Expression of these immunoglobulin genes in future studies will make it possible to determine the antigen specificity of the antibodies expressed by Graves' orbital tissue plasma cells.

Recombinant thyroid peroxidase autoantibodies can be used for epitopic "fingerprinting" of thyroid peroxidase autoantibodies in the sera of individual patients

Four human monoclonal antibodies (SP1.4, WR1.7, TR1.8, and TR1.9) map the immunodominant region on thyroid peroxidase (TPO) recognized by autoantibodies in patients' sera. We used a pool of these monoclonal antibodies, expressed in bacteria as antigen-binding fragments [F(ab)], to compete for TPO autoantibody binding to radiolabeled TPO. The F(ab) inhibited TPO binding by 32 patients' sera by 82 +/- 14% (mean +/- SD), with a range from 51-100%. When each F(ab) was tested individually for its ability to compete for autoantibody binding to TPO, F(ab) TR1.8 was the most potent among the 32 sera. However, there was a wide spectrum of TPO binding inhibition when each serum was considered individually, thereby allowing an epitopic "fingerprint" to be drawn for the TPO autoantibodies in a patient's serum. There was a close association between the proportions of TPO autoantibodies to the TR1.8 and TR1.9 epitopes as well as between those to the SP1.4 and WR1.7 epitopes. These associations correspond to the previously described A and B epitopic domains in the TPO immunodominant region. No TPO epitope was observed to be associated with clinically apparent ophthalmopathy of Graves' disease, nor was there an association between TPO epitopes and patient age or sex. In summary, the present study on a large sample of sera with TPO autoantibodies indicates that by using TPO-specific F(ab) selected to cover all regions of the TPO immunodominant region, it is possible to obtain a TPO epitopic fingerprint for each serum. These data open the way to future studies directed at testing the hypothesis of disease-associated TPO epitope(s).

Molecular cloning and characterization of genes for antibodies generated by orbital tissue-infiltrating B-cells in Graves' ophthalmopathy

Graves' ophthalmopathy is a distressing autoimmune disease of unknown etiology. Analysis of the genes for antibodies secreted by orbital tissue-infiltrating plasma cells might provide insight into the pathogenesis of this disease. We, therefore, constructed an immunoglobulin heavy (H) chain and an immunoglobulin kappa light (L) chain cDNA library from the orbital tissue of a patient with active Graves' ophthalmopathy. Analysis of 15 H (IgG1) and 15 L (kappa) chains revealed a restricted spectrum of variable region genes. Fourteen of 15 variable kappa genes were about 94% homologous to the closest known germline gene, KL012. Thirteen of 15 H chain genes were 91% and 90% homologous to the closest germline genes, DP10 and hv1263, respectively. Remarkably, these germline genes also code for other autoantibodies to striated muscle (KL012) and thyroid peroxidase (KL012 and hv1263). These studies raise the possibility that particular germline genes may be associated with autoimmunity in humans. Further, the present study opens the way to identifying ocular autoantigens that may be the target of an humoral immune response.

The immunodominant region on human thyroid peroxidase recognized by autoantibodies does not contain the monoclonal antibody 47/c21 linear epitope

We performed studies to determine whether the binding sites on thyroid peroxidase (TPO) of immunoglobulin antigen binding fragments (Fabs) representing more than 80% of the human autoantibody repertoire overlap with the binding site of monoclonal antibody (Mab) 47, the only Mab whose partial epitope has been defined at the amino acid level (residues 713-721). We also investigated whether these Fabs preferentially recognize native or denatured TPO. None of the Fabs, when bound to radiolabeled TPO, interfered with the ability of Mab 47 to bind to this material. In enzyme-linked immunosorbent assay experiments, the binding of TPO autoantibody Fabs SP1.5, WR1.7, TR1.8, and TR1.9 was greatly diminished by denaturation of TPO. In contrast, binding of Mab 47 was higher to denatured TPO than to intact TPO. Our studies indicate that the Mab 47/C21 epitope lies outside the immunodominant region on TPO. Further, the data confirm that the majority of epitopes for TPO autoantibodies are highly conformational (dependent on the three-dimensional structure of the native protein). Native TPO will be needed to complete the mapping of the epitopes for TPO autoantibodies as well as to determine the amino acids at the autoantibody-antigen-binding sites.

High affinity, thyroid-specific human autoantibodies displayed on the surface of filamentous phage use V genes similar to other autoantibodies

Autoantibodies to thyroid peroxidase (TPO) are characteristic of thyroid inflammation in autoimmune thyroid disease. We have used the phage display, H and L chain combinatorial cDNA library approach to clone, from thyroid-infiltrating B cells, six new human Fab autoantibodies with high affinities (approximately 10(-10) M) for TPO. This library, in the pComb3 vector, was screened with viable, stably transfected Chinese hamster ovary cells expressing human TPO on their surface. The H and L chain genes in the six TPO-specific Fab were similar, but not identical, to those encoding Fab previously isolated from the same library by screening bacteriophage plaques in the Immunozap vector with purified TPO. The TPO-specific VK genes isolated with the phage display system are closer to germline than those obtained with Immunozap. Essentially all the V kappa isolated with pComb3 were 99 or 100% homologous with the germ-line genes KL012 and A3 that also code for low affinity systemic autoantibodies. There are two important implications of the study. First, the phage display system can be used with impure Ag to generate high affinity autoantibodies. This finding opens the way to cloning autoantibodies against other autoantigens, not previously possible with the bacteriophage lambda approach because of the lack of purified Ag. Second, germ-line L chain genes can code for very high affinity antibodies.

High affinity, thyroid-specific human autoantibodies displayed on the surface of filamentous phage use V genes similar to other autoantibodies

Autoantibodies to thyroid peroxidase (TPO) are characteristic of thyroid inflammation in autoimmune thyroid disease. We have used the phage display, H and L chain combinatorial cDNA library approach to clone, from thyroid-infiltrating B cells, six new human Fab autoantibodies with high affinities (approximately 10(-10) M) for TPO. This library, in the pComb3 vector, was screened with viable, stably transfected Chinese hamster ovary cells expressing human TPO on their surface. The H and L chain genes in the six TPO-specific Fab were similar, but not identical, to those encoding Fab previously isolated from the same library by screening bacteriophage plaques in the Immunozap vector with purified TPO. The TPO-specific VK genes isolated with the phage display system are closer to germline than those obtained with Immunozap. Essentially all the V kappa isolated with pComb3 were 99 or 100% homologous with the germ-line genes KL012 and A3 that also code for low affinity systemic autoantibodies. There are two important implications of the study. First, the phage display system can be used with impure Ag to generate high affinity autoantibodies. This finding opens the way to cloning autoantibodies against other autoantigens, not previously possible with the bacteriophage lambda approach because of the lack of purified Ag. Second, germ-line L chain genes can code for very high affinity antibodies.

Autoantibodies in the sera of patients with autoimmune thyroid disease recognize a secreted form of human thyroid peroxidase generated in a baculovirus system

We used a baculovirus vector to express the cDNA for a truncated (amino acid residues 1-848), secreted form of thyroid peroxidase (TPO) in Sf9 insect cells. Immunoreactive TPO was detected in pooled conditioned media from 10 clones using polyclonal TPO autoantibodies in the sera of patients with autoimmune thyroid disease. As a further test of TPO immunogenicity, the pooled media completely inhibited autoantibody binding to antigen. We used an ELISA to compare autoantibody reactivity to insect cell-derived TPO and TPO antigen produced by Chinese hamster ovary (CHO) cells, the standard form of antigen in present use. In a study of 22 TPO antibody-negative sera and 24 sera with different TPO autoantibody potencies, there was a highly significant correlation (r = 0.977; p < 0.001) in OD values obtained with TPO from the two different sources. The highest producing baculovirus clone generated 8.5 micrograms TPO/ml of conditioned medium, nearly 10-fold higher than previously achieved with stably transfected Chinese hamster ovary cells. Baculovirus-derived, soluble TPO therefore is an excellent source of recombinant TPO in further studies to examine the precise B cell epitopes for human autoantibodies.

Lack of promiscuity in autoantigen-specific H and L chain combinations as revealed by human H and L chain "roulette"

Individual H or L chains from a human autoantibody were used to search for other L or H chains that could form antigen-binding fragments, Fab, with the same specificity. The parent Fab (SP1.2) exhibits high affinity binding for thyroid peroxidase (TPO), a 107-kDa protein that is the major autoantigen in human autoimmune thyroiditis. This autoantibody "roulette," performed by using Ig H and L chain gene libraries expressed in bacteria, increased the frequency of TPO-binding clones in the new libraries. However, the frequency was still much lower than would be the case if promiscuous combinations with a variety of H or L chains were compatible with specific Ag binding. Nucleotide sequence analysis of the H and L chains of the new TPO-binding clones revealed even more restriction. Thus, with the SP1.2 H chain, all 11 new Fab utilized L chains from the same V kappa 1 family germline gene as SP1.2 itself. Similarly, five of six H chains "captured" by the SP1.2 L chain were very closely related to the SP1.2 H chain. However, one totally different H chain was isolated: SP4.6 has a VH region that differs substantially from that of SP1.2. SP4.6 also has a distinct D region, uses a different JH, and, unlike SP1.2, which is an IgG1, belongs to subclass IgG4. The affinities for TPO of SP4.6 (with the different H chain) and SP1.20 (which had the least mutated L chain germline gene) were similar to that of SP1.2 (approximately 10(-10) M). As expected, the SP1.2 and SP1.20 Fab, which have the same H chain and closely related L chains, bound to the same domain on TPO. However, a similar domain on TPO was recognized by both SP4.6 and SP1.2, despite the fact that their V, D, and J regions are quite different. This observation raises the possibility that the L chain is critical in defining epitope specificity, even in the presence of completely different D regions and nonidentical VH regions.

Cross-reactive idiotypes on high affinity IgG class human monoclonal thyroglobulin autoantibodies

Anti-idiotypic antibodies have been developed in rabbits against three high affinity IgG class monoclonal human autoantibodies to thyroglobulin (Tg), which resemble polyclonal Tg antibodies in patients with autoimmune thyroid disease. Antibodies to 1E10 monoclonal anti-Tg (IgG2 kappa) recognised a cross-reactive idiotype (CRI) also present on 1D3 monoclonal anti-Tg (IG1 lambda) and on VB5 monoclonal anti-Tg (IgG2 lambda). The determinant to which anti-1E10 binds appears to involve, at least in part, the binding site for Tg. In contrast, anti-idiotypic antibodies raised against VB5 failed to bind to either 1E10 or 1D3, a finding consistent with previous studies on serum polyclonal Tg antibodies, which suggested that such antibodies exhibit a mixture of private and cross-reactive idiotypes. The observed sharing of idiotypic determinants was not related to subclass, light chain type or expression of a particular VH gene family in the heavy chain. Although binding of the anti-idiotypic antibodies to Tg antibodies in a panel of patients (including the donors of the lymphocytes used to produce the monoclonal antibodies) could not be detected, the monoclonal antibodies are representative of the donor patients' serum Tg antibodies, both in terms of IgG subclass and functional affinity. Thus these idiotypes may be present in patients' sera at levels below the detection limits of the assays employed. Cross-reactive regulatory idiotypes in mice often constitute a minor component of the anti-Tg repertoire. Consequently, it is possible that low levels of a CRI, such as the 1E10 CRI, may be involved in the regulation of the autoimmune response to Tg in man.

Lack of promiscuity in autoantigen-specific H and L chain combinations as revealed by human H and L chain "roulette"

Individual H or L chains from a human autoantibody were used to search for other L or H chains that could form antigen-binding fragments, Fab, with the same specificity. The parent Fab (SP1.2) exhibits high affinity binding for thyroid peroxidase (TPO), a 107-kDa protein that is the major autoantigen in human autoimmune thyroiditis. This autoantibody "roulette," performed by using Ig H and L chain gene libraries expressed in bacteria, increased the frequency of TPO-binding clones in the new libraries. However, the frequency was still much lower than would be the case if promiscuous combinations with a variety of H or L chains were compatible with specific Ag binding. Nucleotide sequence analysis of the H and L chains of the new TPO-binding clones revealed even more restriction. Thus, with the SP1.2 H chain, all 11 new Fab utilized L chains from the same V kappa 1 family germline gene as SP1.2 itself. Similarly, five of six H chains "captured" by the SP1.2 L chain were very closely related to the SP1.2 H chain. However, one totally different H chain was isolated: SP4.6 has a VH region that differs substantially from that of SP1.2. SP4.6 also has a distinct D region, uses a different JH, and, unlike SP1.2, which is an IgG1, belongs to subclass IgG4. The affinities for TPO of SP4.6 (with the different H chain) and SP1.20 (which had the least mutated L chain germline gene) were similar to that of SP1.2 (approximately 10(-10) M). As expected, the SP1.2 and SP1.20 Fab, which have the same H chain and closely related L chains, bound to the same domain on TPO. However, a similar domain on TPO was recognized by both SP4.6 and SP1.2, despite the fact that their V, D, and J regions are quite different. This observation raises the possibility that the L chain is critical in defining epitope specificity, even in the presence of completely different D regions and nonidentical VH regions.

Thyrotropin receptor autoantibodies recognizing two different epitopes on the TSH receptor: lack of relationship to patient age, sex, and ophthalmopathy

The existence of two populations of stimulatory TSH receptor autoantibodies against different epitopes raises the possibility of a link between one type of autoantibody and the clinical manifestations of Graves' disease. To test this hypothesis, serum immunoglobulins from 48 patients with Graves' disease were assayed for TSH binding inhibition (TBI) activity with two different recombinant TSH receptor variants (TSH-LHR-6 and TSH-LHR-6-A1) expressed on Chinese hamster ovary cells. The activity of 27 of the 48 patients' immunoglobulin samples was significantly less (difference in TBI value of 9% or greater) with chimera 6-A1 than with chimera 6. No immunoglobulin sample had significantly greater TSH binding inhibitory activity with chimera 6-A1 than with chimera 6. Sensitivity to the 6-A1 epitope substitution did not correlate with patient age, sex, or the presence or absence of hyperthyroidism. Further, there was no segregation of individual patients with TSH receptor autoantibodies with 6-A1 epitope sensitivity in terms of the past or present occurrence of ophthalmopathy, including the severity (total eye score), clinical activity, duration, and type of therapy. These data indicate that recognition by autoantibodies of the 6-A1 epitope on the TSH receptor is not associated with the ophthalmopathy of Graves' disease. However, the possibility cannot be excluded that other functional (or even nonfunctional receptor autoantibodies that are not detectable by present assays) may still play a role in the pathogenesis of Graves' ophthalmopathy.