Characterization of Monoclonal Thyroid-Stimulating and Thyrotropin Binding-Inhibiting Autoantibodies from a Hashimoto’s Patient Whose Children Had Intrauterine and Neonatal Thyroid Disease

A multiplicity of TSH receptor autoantibodies (TSHRAbs) have beencharacterizedaftersubcloningheterohybridomasproducedfromthelymphocytesofapatientwhohasHashimoto’sthyroiditisandhadthreechildrenwithintrauterineorneonatalhyperthyroidism.TwelveclonesproducedstimulatingTSHRAbsthatincreasedcAMPlevelsandiodideuptakeinratFRTL-5thyroidcellsandincreasedcAMPlevelsinChinesehamsterovary(CHO)cellstransfectedwiththehumanTSHR;like95%ofGraves’stimulatingTSHRAbs,all12havetheirfunctionalepitopeontheN-terminusoftheTSHRextracellulardomain,requiringresidues90–165foractivity.All12bindtohumanthyroidmembranesintheabsence,butnotthepresence,ofTSH,butareonlyweakinhibitorsofTSHbindinginassaysmeasuringTSHbinding-inhibitingIgs(TBIIs).Incontrast,8differentclonesproducedTSHRAbsthatdidnotincreasecAMPlevels,but,instead,exhibitedsignificantTBIIactivity.FourinhibitedtheabilityofTSHorastim-ulating IgG, ,0.8mmol/L 2 centrifugation IgG 100- m L aliquots 2 IgG to affinity chromatography with protein A-Sepharose CL-4B columns as described The method of preparation no significant difference in any results obtained; yield varied from 0.45–1.4 mg purified IgG/10 flasks. The IgG concentration was measured in an enzyme-linked immu-nosorbant assay as previously described (39), using a b -galactosidase- coupled affinity-purified antihuman IgG (Zymed Laboratories, Burlingame, Reactions were stopped with 0.05 mL 8 mol/L H 2 SO 4 and quantitated in a Titer-Tek Multiskan Reader (Flow Laboratories, McLean, VA). This program provides an overview of the scope and methods employed in infectious disease epidemiology and research, the unique aspects of hospital epidemiology and infection control, and the epidemiology and prevention of specific infectious diseases. The program is designed for all practitioners in the disciplines of epidemiology, public health, health administration, medicine, nursing, and related professions. The format features lectures and discussion with faculty. Topics to be covered include: Trends in AIDS and HIV Prophylaxis; Herpes Virus Infections—Association with Kaposi’s Sarcoma; Mycoplasmas; Chlamydiae; Emerging Infectious Diseases; Cholera; Ehrlichiosis; Coccidioidomycosis; Rickettsiosis; Autochthonous Mar-laria Transmission; Helicobacter pylori; Antimicrobial Drug Resistance. program

ulating TSHRAb to increase cAMP levels and had their functional epitope on the C-terminal portion of the TSHR external domain, residues 261-370, mimicking the properties of blocking TSHRAbs that cause hypothyroidism in patients with idiopathic myxedema. The 4 other TBIIs inhibited the ability of TSH, but not that of a stimulating TSHRAb, to increase cAMP levels, like TBIIs in Graves' patients. The functional epitope for 3 of these Graves'-like TBIIs was residues 90 -165; the functional epitope for the fourth was residues 24 -89. The fourth also increased arachidonic acid release and inositol phosphate levels in FRTL-5 thyroid cells and exhibited conversion activity, i.e. the ability to increase cAMP levels in the presence of an anti-human IgG. Thus, this TBII exhibited signal transduction activity, unlike the other 3 Graves'-like TBIIs. The patient, therefore, has stimulating TSHRAbs and 3 different types of TBIIs, each with different functional properties and different epitopes on the TSHR. (J Clin Endocrinol Metab 82: 3998 -4009, 1997) O UR KNOWLEDGE of transplacentally acquired thyroid disease has expanded considerably, based in part on the discovery that maternal Igs can act as TSH receptor (TSHR) antagonists (1)(2)(3)(4) as well as agonists (5), that maternal TSHR antagonists can cause neonatal hypothyroidism (1)(2)(3)(4) as well as neonatal Graves' disease (6), and that a mother's thyroid status is not necessarily concordant with that of her neonate (7)(8)(9)(10). Contributing to this knowledge were reports that described an unusual patient with Hashimoto's thyroiditis who gave birth to three children with intrauterine and/or neonatal hyperthyroidism, each with different clinical courses (7,(11)(12)(13). This patient's serum contained both a thyroid-stimulating autoantibody (TSAb) and an antibody that was able to inhibit the activity of either her own or a Graves' TSAb (11)(12)(13). The presumption was that the clinical syndrome in the children reflected the different activities of two different TSHR antibodies (TSHRAbs) during the gestational and neonatal periods (11)(12)(13). It was recognized, however, that definitive proof required their isolation and characterization, i.e. determination of their mechanism of action and epitopes on the TSHR (11)(12)(13).
Knowledge of TSHR structure and the epitopes of TSHRAbs that act as antagonists or agonists has improved (14 -16). Studies from multiple laboratories (17)(18)(19)(20)(21)(22)(23)(24)(25) have demonstrated that stimulating TSHRAbs differ from TSHRAbs that inhibit TSH binding (TBIIs) or block TSH activity; that Graves' TBIIs do not inhibit stimulating TSHRAb activity (26,27), contrary to the case for TBIIs from patients with hypothyroidism and idiopathic myxedema or Hashimoto's disease (28); and that some, but not all, TBIIs are functional and can activate the inositol phosphate signal transduction system (18,19). Studies evaluating patient IgG preparations using Chinese hamster ovary (CHO) cells transfected with the human TSHR or chimeras of the TSHR and LH/CG receptor (LH/CGR) have identified functional epitopes for stimulating TSHRAbs and two different types of Graves' TBIIs on the N-terminus of the TSHR external domain (29 -32). TBIIs associated with hypothyroidism were shown to be different from Graves' TBIIs (29 -32) and were shown to have their functional epitope on the C-rather than the N-terminus of the TSHR extracellular domain, confirming site-directed mutagenesis or substitution studies (33)(34)(35). Experiments using both approaches together, monoclonals and chimera assays to identify epitopes, have not, however, been performed to cross-validate conclusions concerning TSHRAb diversity, nor have the diversity and its clinical consequences been fully evidenced in a single patient.
This study was, accordingly, directed at isolating the multiplicity of TSHRAbs in the serum of the mother with Hashimoto's thyroiditis who had given birth to children with intrauterine and/or neonatal hyperthyroidism (7,(11)(12)(13), then defining their properties by combining a monoclonal approach with assays not only in functioning FRTL-5 thyroid cells, but also in CHO cells transfected with the human TSHR or TSHR chimeras with residues 9 -165 (Mc1ϩ2), 90 -165 (Mc2), or 261-370 (Mc4) substituted by equivalent residues of the LH/CGR. We used assays measuring cAMP, arachidonic acid release, inositol phosphate production, and TBII activity. We identify clones producing Graves' stimulating TSHRAbs, two different types of Graves' TBIIs (one functional and the other not), and a TBII-blocking type TSHRAb characteristic of TSHRAbs causing hypothyroidism. The data establish the potential diversity of TSHRAbs in Hashimoto's mothers and the potential variability of anti-TSHR autoimmune disease in their progeny, consistent with conclusions from studies using TSHR-LH/CGR chimeras to characterize heterogeneous populations of TSHRAbs in patient IgG preparations (31,32). They support the conclusion that disease diversity and clinical course are caused by the differences in functional epitopes and properties of TSHRAbs.

Clinical state of the patient at the time of these studies
The patient was 31 yr of age when both serum and cells were obtained. She had a history of Hashimoto's thyroiditis, with hypothyroidism diagnosed at age 17 yr and had been taking 0.2 mg/day levothyroxine ever since. While receiving replacement therapy, she delivered three children with intrauterine and/or neonatal hyperthyroidism, each with different clinical courses (7,(11)(12)(13). She exhibited no clinical features of hypo-or hyperthyroidism at the time of the study, which was 19 months after the delivery of her third child. Her T 4 level was 6.1 g/dL (normal range, 5.5-13 g/dL), T 3 was 98 ng/dL (normal range, 80 -220 ng/dL), and TSH was 2.8 U/mL (normal range, Ͻ10 U/mL).

Graves' patients, patients with hypothyroid-idiopathic myxedema, and normal subjects from whom serum was obtained
The diagnosis of hyperthyroidism due to Graves' disease was based on clinical findings, including diffuse goiter, diffuse 99m Tc uptake on thyroid scans, elevated serum levels of free and/or total iodothyronines, and an absent serum TSH response to TRH. Each had stimulating TSHRAb activity in their serum measured in FRTL-5 cell assays. Patients with primary myxedema had clinical features of hypothyroidism, low serum levels of free and/or total iodothyronines, and no palpable goiter. Each had blocking TSHRAb activity measured with both TSH and a Graves' stimulating TSHRAb using FRTL-5 cell assays or assays in human (h) TSHR-transfected CHO cells. The Graves' and myxedema patients all exhibited positive TBII activity in a commercial assay. Normal subjects had no goiter, no personal or family history of thyroid disease and/or ophthalmopathy, normal circulating iodothyronine and TSH levels, and absent TSHR, microsomal, or thyroglobulin antibodies. In this study, we used sera from 20 normal individuals, 7 Graves' patients, and 4 subjects with myxedema; there was no significant difference in results when IgG from individual Graves' or myxedema sera were substituted within individual assays. Sera from these individuals were stored at Ϫ20 C until IgG preparations were made. All sera were obtained under institution guidelines, with appropriate approval and consent.
IgGs were extracted from the sera of patients by affinity chromatography with protein A-Sepharose CL-4B columns; the resulting eluent was dialyzed against distilled water, lyophilized, and stored at Ϫ20 C as previously described (31). Normal IgG was a pool of sera from 20 normal individuals. Purified IgG was dissolved in assay buffer to the desired concentration immediately before use.

Production of lymphocyte heterohybridoma clones
Peripheral lymphocytes from the mother were harvested by Ficoll-Hypaque gradient centrifugation of 100 mL blood within 30 min of venipuncture and were washed twice with Hanks' Balanced Salt Solution (HBSS; Biofluids, Rockville, MD). Half of the cells were taken for immediate fusion; half were fused 1 day later after being maintained overnight in glutamine-supplemented DMEM (Biofluids) containing 10% FCS. Before use in fusion experiments, cell number was adjusted to 1 ϫ 10 7 /mL in DMEM. Polyethylene glycol fusion with 5 ϫ 10 7 P3-NSI/1-AG4 -1 mouse myeloma cells was performed as previously described (17,18,36). The fusion products were dispersed in hypoxanthine-aminopterin-thymidine selection medium; additional medium was added every 4 days until visible hybridoma colonies had reached the 100-to 200-cell stage (17,18,36).

Screening assays
Heterohybridoma clones producing human IgG were detected by incubating the medium from each well in a solid phase RIA system previously described (37). Hybridoma antibodies binding to thyroid membranes were assayed using human thyroid membranes in suspension in both the presence and absence of 1 ϫ 10 Ϫ7 mol/L/L purified bovine (b) TSH (17,18). In brief, culture medium was incubated with a suspension of human thyroid membranes in the wells of 96-well microtiter plates for 1 h at 4 C. The plates were centrifuged at 4 C for 10 min, the supernatant solutions were discarded, and the pellets were washed with phosphate-buffered saline, pH 7.4, containing 0.1% gelatin and 0.02% sodium azide. 125 I-Labeled antihuman F(abЈ) in phosphatebuffered saline containing 1% gelatin and 1% crystalline BSA was added and incubated for 45 min at 4 C; unbound radioactivity was removed by centrifugation, pellets were washed as described above, plates were dried under an infrared lamp, and individual wells were cut and assayed for bound radioactivity.

Heterohybridoma subcloning and antibody purification
The heterohybridomas were subcloned by three rounds of limiting dilution over a 6-month period using 96-well plates. In each round, cells were plated to 3 and 1 cells/well (17,18). This procedure is similar to that used recently (25), which is accepted to produce cells that are clonal in origin. Single clusters of cells were expanded in 24-well, then in 100-mm tissue culture dishes using the selective medium described above. IgG was prepared from the medium of overgrown cultures from 10 750-mL flasks that had been harvested by centrifugation at 500 ϫ g for 10 min. After ammonium sulfate precipitation (0 -45% saturation), the IgG were dialyzed for 24 h with three changes of 0.1 mol/L Tris-chloride, pH 7.5, subjected to high pressure liquid chromatography (38) to prepare purified IgG, and dialyzed in 50 mmol/L Tris-chloride, pH 7.4, or NaCl-free HBSS (5.4 mmol/L KCl, 1.3 mmol/L CaCl 2 , 0.8 mmol/L MgSO 4 , 0.3 mmol/L Na 2 HPO 4 , 0.4 mmol/L KH 2 PO 4 , and 0.1% glucose) containing 20 mmol/L HEPES, pH 7.4, and 222 mmol/L sucrose. After centrifugation to remove any precipitate, IgG was distributed in 100-L aliquots and stored at Ϫ70 C. Alternatively, IgG was subjected to affinity chromatography with protein A-Sepharose CL-4B columns as described above. The method of preparation had no significant difference in any results obtained; yield varied from 0.45-1.4 mg purified IgG/10 flasks.

Determination of heavy and light chains
Subclasses of IgG antibodies were determined using an enzymelinked immunosorbent assay as previously described (40). Mouse monoclonal subclass-specific antibodies were obtained from Boehringer Mannheim (Indianapolis, IN) as was the peroxidase-conjugated antihuman IgG antibody; standards were purchased from Nordic (Bochum, Germany). Light chains were determined by an immunodot procedure (41); antihuman light chain-specific and antihuman light chainspecific mouse monoclonal antibodies were obtained from Dianova (Hamburg, Germany).
The ability of TSH or antibodies to increase cAMP levels in FRTL-5 cells was measured as previously described (37,43) in NaCl-free HBSS containing 20 mmol/L HEPES (pH 7.4), 1% BSA, 0.5 mmol/L 3-isobutyl-1-methylxanthine (IBMX), and 222 mmol/L sucrose. In assays measuring blocking TSHRAb activity, 1 ϫ 10 Ϫ10 mol/L bTSH or a standard Graves' IgG with stimulating TSHRAb activity was included in the assay. In some experiments, total inositol phosphate as well as cAMP production were measured in the same wells after incubation for 2 h at 37 C with 0.2 mL NaCl-free HBSS assay buffer containing 20 mmol/L HEPES (pH 7.4), 1% BSA, 0.5 mmol/L IBMX, 222 mmol/L sucrose, 10 mmol/L LiCl, and bTSH or IgG at the concentrations noted (44); values in these experiments were corrected for total tritiated inositol incorporated as well as for cell protein.
TSH-and stimulating TSHRAb-increased iodide uptake were measured according to the previously described procedure (45), as was the ability of an IgG or TSH to release arachidonic acid from cells preincubated with [ 3 H]arachidonic acid overnight (46). In brief, cells for the latter assays were incubated with 50 nmol/L [ 3 H]arachidonic acid (20 Ci/10-cm diameter dish). After washing with NaCl-free HBSS containing 20 mmol/L HEPES (pH 7.4) and 0.05% fatty acid-free BSA, cells were incubated in the same buffer for 30 min with TSH, IgG, or 10 mol/L A23187.
Results are expressed per microgram DNA (43). All experiments included normal human IgG as a negative control and both a known Graves' stimulating TSHRAb and purified bTSH as positive controls.

Stable transfectants containing CHO-hTSHR and CHO-hTSHR-LH/CGR chimeras and measurement of stimulating or blocking TSHRAb activity in hTSHRtransfected CHO cells
hTSHR cloning, amplification of rat LH/CGR complementary DNA fragments by PCR, construction of chimeric receptors, subcloning into pSG5 expression vectors, and plasmid purification have been described previously (29 -32). The three chimeric receptors used in this report are designated to indicate the substituted segment of the TSHR, numbered from the methionine start site. In Mc1ϩ2, residues 9 -165 of the TSHR were replaced by residues 10 -166 of the LH/CGR; in Mc2, residues 90 -165 of the TSHR were replaced by residues 91-166 of the LH/CGR; in Mc4, TSHR residues 261-370 were replaced by residues 261-329 of the LH/CGR.
Stable CHO cell transfectants were made with lipofectin (Life Technologies) by cotransfection of the wild-type hTSHR or the hTSHR-LH/ CGR chimeras in pSG5 with a pMAMneo selection marker (31,32). Cells were fed every other day and split every 4 -5 days with trypsin-ethylenediamine tetraacetate. After harvesting, cells were plated in 24-well plates (3-4 ϫ 10 4 cells/well) and fed fresh medium 48 h later. They were used in assays when they were at 100% confluency, usually 60 -72 h after plating. In a typical assay, there were 5-6 ϫ 10 5 cells/well. The cAMP response to a standard amount of bTSH or a standard pool of Graves' IgG was stable for over 3 months of continuous culture (ϳ30 passages).
Studies of the Mc4 chimera used transient transfection procedures, as previously described (33)(34)(35)44), wherein COS-7 cells were transfected by electroporation with 25 g purified plasmid DNA. Cells transfected with the pSG5 vector alone were negative controls. Cells transfected with wild-type TSHR were positive controls.
Stimulating TSHRAb activity was measured as previously detailed (31,32,37,44). Incubations with IgG alone or with either 1 ϫ 10 Ϫ10 mol/L TSH or a standard Graves' IgG was performed for 3 h at 37 C in with NaCl-free HBSS containing 20 mmol/L HEPES (pH 7.4), 222 mmol/L sucrose, 1% BSA, and 0.5 mmol/L IBMX. cAMP was measured in the supernatants by RIA. Conversion assays were performed using Mc2transfected CHO cells by the procedure described previously (32). In brief, cells in 96-well plates were washed with NaCl-free HBSS containing 20 mmol/L HEPES (pH 7.4), and 222 mmol/L sucrose and 100 L IgG in the same buffer were added. After 30 min at 37 C, cells were washed and incubated with 100 L goat antihuman IgG (1:50 diluted Fab fragment specific; from Cooper Biomedical, Malvern, VT) for 30 min at 4 C, then for 3 h at 37 C. cAMP levels were again measured in the supernatants by RIA. All experiments included normal human IgG as a negative control.

[ 125 I]TSH binding to thyroid membrane preparations and measurement of TBII activity
Highly purified bTSH (NIDDK bTSH I-1; 30 U/mg) was radioiodinated as previously described (47). The ability of an IgG to inhibit 125 I-labeled TSH binding to human thyroid membranes was measured in a solid phase assay, previously described (17,18). In brief, [ 125 I]TSH binding was measured in 10 mmol/L Tris-HCl, pH 7.4, containing 50 mmol/L NaCl and 0.5% BSA after incubations at 37 C for 2 h. Wells were doubly washed, and residual bound radioactivity was measured. Maximal binding averaged 2-15% of the total counts added; control values are subtracted from the data presented and never exceeded 1% of the radioactivity added. Controls included incubations with all components except membranes or incubations where normal human IgG substituted for the experimental IgGs.
TBII assays were performed using a commercially available kit (TRAK kit, Henning, Berlin, Germany) that uses solubilized porcine thyroid membranes. Additionally, TBII values were measured in an assay using solubilized membranes from the CHO cells transfected with the wild-type TSHR, the Mc2, or the Mc1ϩ2 chimera (32). In brief, 50 L each of serum and solubilized membranes were mixed and incubated for 15 min at room temperature. [ 125 I]TSH (30,000 cpm) was added and incubated for 16 h at 4 C before 300 L 10 mmol/L Tris-chloride, pH 7.4, containing 50 mmol/L NaCl and 0.1% BSA were added together with 500 L 30% polyethylene glycol in 1 mol/L NaCl. Tubes were centrifuged, pellets were washed with 20% polyethylene glycol in an identical buffer, and radioactivity was measured. TBII activity is expressed as the percent inhibition of [ 125 I]TSH binding to the TSH receptor by comparison to pooled normal IgG; a TBII value exceeding 15%, which is greater than 2 sd above the mean value from 20 normal samples, was considered positive.

Other assays and statistics
The protein concentration was determined by the method of Lowry et al. (48) or using Bio-Rad reagents (Bio-Rad, Richmond, CA); bovine IgG was the standard. When protein in membrane preparations was measured, the membranes were first solubilized by treatment with 1 n NaOH for 1 h.
In all cases, experiments were repeated on at least three occasions using different batches of cells. All incubations were performed at least in duplicate, and all wells were assayed in duplicate. Results are expressed as the mean Ϯ sd of the results of an individual experiment unless otherwise noted. Statistical analysis was performed by one-way ANOVA to determine the P value between different groups and by Spearman's rank correlation coefficient to validate the correlation between two series of data.

Materials
TSH was a purified bovine preparation (49)

Activity of the IgG in the patient's serum at the time of this study
In FRTL-5 rat thyroid cells, the IgG from the patient exhibited a biphasic ability to increase, then decrease, cAMP levels as a function of IgG concentration (Fig. 1A). The IgG was also able to inhibit the ability of an IgG from a Graves' patient to increase cAMP levels (Fig. 1B). These results established that the patient's lymphocytes were still likely to be producing at least two types of TSHRAbs: stimulating and inhibiting. Additionally, they showed that the FRTL-5 rat thyroid cell assay system could be used as an assay tool during cloning, contrary to previous results (12); the basis for the discrepancy is not clear.
The sequence of addition affected the ability of the patient's IgG to inhibit Graves' IgG. Thus, better inhibition was evident if the patient IgG was preincubated with FRTL-5 thyroid cells before addition of Graves' IgG during the screening studies (data not shown). . In A, the activity of the patient IgG (f) is compared with that of purified bovine TSH (F) or a Graves' IgG (å) at the noted concentrations of each. In B, Graves' IgG was added at the indicated concentration together with no additional IgG, with IgG from a pool of 20 normal individuals, with the Graves' IgG itself, or with the patient's IgG. In both panels, the data are the mean Ϯ SD of quadruplicate values from one of three experiments, each yielding similar results. Incubations with one or both IgGs were performed for 3 h; TSH assays were performed 30 min after TSH addition. In both cases preliminary experiments indicated these were the optimal times of assay. In B, a single asterisk represents a significant increase in activity (P Ͻ 0.02), whereas two asterisks denote a significant decrease (P Ͻ 0.01). The Graves' IgG used in these experiments was from patient 7. Data using 4 other Graves' IgGs (no. 1, 2, 4, and 5) were similar except for the absolute values as a function of concentration. Normal IgG was a pool from 20 normal individuals.

Isolation of individual autoantibodies from the lymphocytes of the patient
Fusions with NIS-1-AG4 -1 mouse myeloma cells resulted in 3516 seeded wells. Human IgG was detected in 592; cells in 310 of these survived for a 6-week period and were subcloned or frozen. Supernatants from 3 colonies were reactive with a monoclonal mouse stimulating TSHRAb (17,18) or TSH, 5 with thyroglobulin, 5 with thyroid peroxidase, and 54 with thyroid membranes. Clones producing antibodies reactive with the thyroid membranes were screened to identify clonal populations producing IgG able to bind to the membranes in the absence, but not presence, of TSH; values ranged from 470-1030 cpm bound without TSH and 200 -300 cpm with TSH (data for IgG from some subclones are presented in Table 2). IgG from 20 of these clones were further characterized; each of these produced a G1 IgG with a light chain.

Thyroid-Stimulating TSHRAbs
IgG produced by 12 clones were stimulating TSHRAbs ( Each IgG increased cAMP levels as a function of IgG concentration, as illustrated for three clones in Fig. 2. In no case was there a biphasic increase, then decrease, evident, even when concentrations as high as 1000 g/mL were tested (data not shown), nor did they inhibit TSH-or Graves' IgGstimulated cAMP levels (see Fig. 6). They did not, therefore, exhibit the inhibiting activity evident in the patient's IgG (Fig. 1B).
Like the IgG in 95% of Graves' patients (31), the stimulating TSHRAb activity of these 12 clones was lost in CHO cells transfected with the TSHR/LH-CGR chimeras, Mc2 and Mc1ϩ2, which substitute residues 90 -165 and 9 -165, respectively, of the TSHR with equivalent residues of the LH/CGR (Table 1, columns 5 and 6). Thus, the functional epitope of each of these stimulating TSHRAbs exists on the N-terminal portion of the extracellular domain of the TSHR and requires residues 90 -165.
The binding of each IgG to FRTL-5 (data not shown) or human thyroid membranes ( Table 2) was inhibited by 1 ϫ 10 Ϫ7 mol/L TSH, but not by an equivalent concentration of LH or hCG. Despite this, none of the clonal stimulating TSHRAbs exhibited significant ability to inhibit TSH binding measured in assays using solubilized porcine membranes ( Table 3, column 2), solubilized membranes from CHO cells containing the wild-type human TSHR (Table 3, column 3), or intact human thyroid membranes, even as a function of the IgG concentration (Fig. 3).

TBIIs
Eight clones produced TBIIs that inhibited TSH binding to either solubilized porcine thyroid membranes (the commercial TRAK assay) or solubilized membranes from human TSHR-transfected CHO cells (Table 3, columns 2 and 3, re- Bold values are statistically increased over control values, P Ͻ 0.02 or greater. cAMP levels and iodide uptake were measured as described in Materials and Methods. Data are the mean Ϯ SD of 4 experiments, each performed in quadruplicate. Bold clones are considered stimulating TSHRAbs, because they increase cAMP levels in TSHR-transfected cells as well as in FRTL-5 thyroid cells. ND represents an experiment that was not performed, because IgGs from these clones were not able to increase cAMP levels in FRTL-5 cells or CHO cells transfected with the TSHR. Data using IgGs from all 4 other Graves' and all 3 other myxedema patients were similar. Normal IgG was a pool from 20 normal individuals. Bold, italicized clones are TBIIs (see Text and Tables 3 and 4). spectively). None of these increased cAMP levels or iodide uptake in FRTL-5 thyroid cells (Table 1, columns 2 and 3) or cAMP levels in TSHR-transfected CHO cells (Table 1, column 4). The TBIIs could, however, be grouped into TBIIs that mimicked the activity of blocking TSHRAbs prevalent in hypothyroid patients with idiopathic myxedema or Hashimoto's disease or TBIIs similar to those present in Graves' patients.
TBIIs that are blocking TSHRAbs associated with hypothyroidism. Four TBIIs in this group inhibited TSH binding to human thyroid membranes as a function of the IgG concentration, without exhibiting the initial increase in binding caused by the patient IgG (Fig. 4). All four inhibited both TSH-and Graves' IgG-stimulated cAMP levels in FRTL-5 cells (Table  4B, columns 2 and 3) or TSHR-transfected CHO cells (Table  4B, columns 4 and 5). These four activities mimicked those of a blocking TSHRAb from a patient with hypothyroidism and idiopathic myxedema (Tables 3 and 4B, rows 6 and 1, respectively).
The Mc2 chimera responds to TSH, but not to a stimulating TSHRAb, whereas the converse is true in Mc4 cells (Table 4A) (29 -32). The ability of these TBIIs to inhibit TSH binding or to inhibit TSH activity was not lost in assays using the Mc2 chimera (Table 3, column 4, and Table 4B, column 6, respectively), whereas the ability to inhibit stimulating TSHRAb activity was lost in the Mc4 chimera (Table 4B, column 7). Again, this mimicked the actions of an IgG from a patient with hypothyroidism and idiopathic myxedema (Tables 3  and 4B, rows 6 and 3, respectively).
In summary, these data indicated that TBIIs from four independent clones, DT11E1, LK3D21, CM6F4, and DT4D2, had the characteristics of TBII-blocking TSHRAbs found in some patients with idiopathic myxedema or Hashimoto's disease and hypothyroidism (1-4, 26 -28, 32, 50). Thus, they inhibited the activity of both TSH as well as a stimulating TSHRAb (27), and their functional epitope was on the Crather than the N-terminal portion of the TSHR extracellular domain, as evidenced by their ability to retain activity in the Mc2, but not the Mc4, chimeras.
The ability of CM5C3 to inhibit TSH binding was lost in the Mc1ϩ2 chimera (Table 3, column 5) as was its ability to inhibit TSH increased activity. Thus, TSH-stimulated cAMP levels in the Mc1ϩ2 chimera were 156 Ϯ 12-fold higher than control activity in both the presence and absence of the CM5C3 IgG.
In summary, these data indicated that TBIIs, LK7G1, WH7F2, DT2F1, and CM5C3, had their functional epitope on the N-rather than the C-terminal portion of the extracellular domain and did not behave like an IgG from a patient with hypothyroidism and idiopathic myxedema in all assays (Tables 3 and 4B, rows 6 and 1, respectively), but, rather, behaved like Graves' TBIIs. The functional epitope of three, LK7G1, WH7F2, and DT2F1, was between residues 90 -165, and the functional epitope of the fourth, CM5C3, was between residues 24 -89, presuming the absence of the signal peptide.
Early studies of monoclonal TSHRAbs reported that some TBIIs increased thyroid cell growth in the absence of increased cAMP levels (18,19,51) and that this activity was inhibited by indomethacin (19). Subsequently, IgGs were described in Graves' patients that increased arachidonic acid release and growth, but not cAMP levels (52,53). The IgG produced by clone CM5C3, but not that produced by clone LK7G1, WH7F2, or DT2F1, increased arachidonic acid release and inositol phosphate levels in FRTL-5 cells (Fig. 6).
In a recent report, we noted that some TBIIs in Graves' patients completely lost TBII activity using membranes from the Mc2 chimera, whereas others did not (32). We associated the Graves' TBIIs that exhibited residual TBII activity in the Mc2 chimera with a phenomenon known as the conversion assay and termed these type II TBIIs (32). Conversion activity is the phenomenon by which an antihuman IgG converts a blocking-type TSHRAb to a stimulating TSHRAb (54). The IgG produced by clone CM5C3, but not that produced by clones LK7G1, WH7F2, and DT2F1, exhibited conversion activity ( Table 5). As predicted (32,54), the clones producing blocking-type TSHRAbs typical of those in patients with hypothyroidism and idiopathic myxedema also exhibited conversion activity (Table 5).
In summary, the IgG produced by clone CM5C3 was different from the other TBIIs with their functional epitope on the N-terminal portion of the TSHR. Its functional epitope appeared to be within residues 24 -89, rather than 90 -165; it activated the inositol phosphate signal transduction system, whereas the other TBIIs did not; and it could be converted to a stimulating TSHRAb that increased cAMP levels in the presence of antihuman IgG.

Discussion
The present report adapted monoclonal antibody techniques and TSHR-LH/CGR chimera assays (42) to characterize the TSHRAbs in a patient with Hashimoto's disease and hypothyroidism who had occult stimulating autoantibodies in her serum that caused intrauterine and/or neonatal hyperthyroidism, compatible with the diagnosis of Graves' disease, in all her children. Early studies of the patient indicated that her IgG had TSAb activity but that this activity was masked at high IgG concentrations by the presence of an inhibitor of TSAb activity (11)(12)(13). The hypothesis was that she had a mixture of stimulating and inhibiting TSHRAbs in her serum and that these had both caused and influenced, respectively, the expression of hyperthyroidism in the children.
The nature of the TSHRAbs in the patient was uncertain. For example, the TSAb activity appeared to be unlike the TSAb activity in Graves' patients; the activity was not retained by the F(ab) fragment, was not evident in FRTL-5 cells, and was absorbed by thyroid, but not fat, cell membranes. They did find that the TSAb inhibitory activity was not directly related to TBII activity, consistent with the accumulation of evidence in monoclonal TSHRAb studies that a Graves' TBII and a Graves' stimulating TSHRAb were distinct antibody populations (15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27). However, as Graves' TBIIs had been reported not to inhibit stimulating TSHRAb activity in monoclonal TSHRAb studies (26,27), despite their ability to inhibit TSH-increased cAMP levels, it was unclear how a Graves' TBII and a Graves' stimulating TSHRAb could account for alternating hyper-and hypothyroid states during the neonatal period. Rather, one had to presume that these patients might have a TBII seen in patients with primary myxedema or Hashimoto's disease with hypothyroidism that inhibits stimulating TSHRAbs as well as TSH (28). It was, therefore, unclear whether two types of TBIIs existed in the patient, whether the TSAb inhibitor in the patient of this study was only one of these, and why they functioned differently. The monoclonal approach was used to isolate the individual TSHRAbs; our recently described chimera assay system (29 -32) categorized the different TSHRAbs and identified their epitopes. Using TSHR-LH/CGR chimera assays, we had made the observation that there might be three types of TBIIs as well as multiple types of stimulating TSHRAbs (29 -32). Two types of TBIIs were in Graves' patients, one that completely lost TBII activity when assayed using chimera receptors in which the N-terminus of the TSHR, residues 90 -165, was replaced by the LH/CGR, and the other that did not (32). This second Graves' TBII, which retained activity with the Mc2 chimera, was associated with conversion activity, i.e. the ability of antihuman IgG to convert a TBII to a stimulating TSHRAb (32,54). The third TBII was present in the serum of hypothyroid patients with idiopathic myxedema or Hashimoto's disease and was associated with blocking TSHRAb activity, but, like the second type of Graves' TBII, retained binding activity using Mc2 chimeras and exhibited conversion activity (32). We suggested that this was different from the Graves' TBII based on site-directed mutagenesis studies that identified the functional epitope for TBII-blocking TSHRAbs on the C-terminal portion of the TSHR extracellular domain (34 -36, 50). Definitive proof of this, however, required the characterization of monoclonal TBIIs from patients, as in the present studies.
The data in the present report show unequivocally that the patient has lymphocytes capable of producing stimulating TSHRAbs present in 95% of Graves' patients. They are active in FRTL-5 cells and in cells transfected with the human wild-type TSHR; their functional epitope is on the N-terminal portion of the TSHR extracellular domain, and they require residues 90 -165 for activity (31). They bind to thyroid membranes specifically, being displaced by TSH much better than hCG or LH; however, they have weak or minimal TBII activity. The patient also has a TBII with blocking activity that inhibits both stimulating TSHRAb and TSH activity. Its functional epitope is on the C-not the N-terminus of the TSHR extracellular domain, as neither activity is not lost in Mc2 or Mc1ϩ2 chimeras with TSHR residues 90 -165 or 8 -165 sub-stituted, but both are lost in Mc4 chimeras with residues 261-370 substituted. This antibody very likely accounted for the ability of the maternal IgG to inhibit TSAb activity (11)(12)(13). Transplacental passage of this and the stimulating TSHRAb at different times, as a result perhaps of different maternal populations of antibodies, cannot unreasonably account for variable degrees in the delayed onset of hyperthyroidism and the different clinical courses of the children in the absence of therapy. From the point of view of the management of mothers such as the patient herein, it is clear that the CHO-chimera assay system, as evidenced in recent reports (31,32), can provide advance recognition of these two antibody populations in the IgG of a mother, thereby better enabling physicians to anticipate the clinical course of the neonate.
The present studies also establish that the prediction of two different TBII populations in Graves' patients (32) is correct, and the patient has both. One TBII, tentatively termed type I, is directed against residues 90 -165 and can block TSH, but not stimulating TSHRAb, activity. It is termed a type I TBII in our previous report (32). This antibody would not be likely to affect the ability of a stimulating TSHRAb to induce hyperthyroidism, but could well influence the course of the disease during therapy. Thus, as stimulating TSHRAbs are decreased in the sera by methimazole and as TSH levels return to normal, this type of TBII could cause hypothyroidism by blocking TSH action (31). Its presence, when detected by the chimera assay, could influence the need for thyroid hormone supplementation and the duration of this supplementation in a patient. The second type of Graves' TBII, termed tentatively type II, has a different functional epitope from the type I TBII, residues between 24 -89, not residues 90 -165. Thus, its binding is lost to membranes from the Mc1ϩ2, but not the Mc2, chimera. The type II TBII exhibits conversion activity, as we previously suggested (32), and activates the inositol phosphate signal transduction system, as predicted in for some TBIIs in early monoclonal TSHRAb studies (19). This Graves' TBII is functional and cannot be construed simply as an inhibitor of TSH activity. These antibodies could have important functional consequences in increasing iodide efflux, increasing iodination of TG by activating the peroxidase system, and increasing thyroid growth in concert with antibodies increasing cAMP levels (15,16,19,(51)(52)(53).
Several miscellaneous points to consider are as follows. In a previous report (55), the TSHRAb population in the patient was predicted to have a light chain; the present report confirms this prediction. In previous reports (11)(12)(13), the patient's IgG enhanced TSH binding at low concentrations, but inhibited it at high concentrations. This phenomenon is not accounted for by the TSHRAb populations described herein; however, clones with enhancing properties were identified in early studies of monoclonal TSHRAbs (17). The patient does have clones producing antibodies that can enhance TSH binding and activity; the existence of such clones was noted during the screening process, but they were not further characterized. Finally, the results suggest that the term blocking TSHRAb might be best reserved for the TBII autoantibodies that inhibit both stimulating TSHRAb and TSH activities and that a Graves' TBII should not be equated with it, because it has a different functional epitope and is not a blocker of stimulating TSHRAb activity.
In summary, the patient's lymphocytes produce stimulating TSHRAbs, two TBIIs characteristic of Graves' TBIIs, and a TBII that is a blocking TSHRAb associated with hypothyroidism in some patients with idiopathic myxedema and Hashimoto's disease. One of the purposes of the characterization of monoclonal receptor antibodies is to define the epitopes and mechanisms of TSHRAbs; this report achieves this, but requires the chimera assays to accomplish the task. The data validate predictions in studies using the TSHR-LH/CGR chimeras to characterize the heterogeneous population of TSHRAbs in patient IgG preparations (31,32,56) that disease diversity is caused by the differences in functional epitopes and properties of TSHRAbs and that the chimera assay system can be used to identify individual types of TSHRAbs in the serum, and this information can be used to predict the clinical course of autoimmine thyroid disease caused by TSHRAbs. Additionally, a reasonable conclusion from the data is that Graves' and Hashimoto's patients have a diversity of IgGs, reinforcing conclusions from other studies (17-19, 22, 25-35, 44, 50 -54, 56).  6. Ability of IgGs from heterohybridoma clones CM5C3, DT2F1, LK7G1, and WH7F2 to increase [ 3 H]arachidonic acid release (A) or inositol phosphate production (B) in FRTL-5 thyroid cells. The activity in A is compared with the activity of two positive (10 mol/L A23187 and 1 ϫ 10 Ϫ8 mol/L purified bovine TSH) as well as three negative controls (ethanol at the concentration in the A23187 solution, the DT11E1 blocking TSHRAb, and IgG from a pool of 20 normal individuals). In B, the activities of CM5C3 (Ç), WH7F2 (Ⅺ), LK7G1 (E), and DT2F1 (É) are compared as a function of IgG concentration, and results are expressed relative to the action of the same concentration of a pool of IgG from 20 normal individuals, whose activity is set at 1 and approximates 750 Ϯ 100 cpm in all cases. Data are the mean Ϯ SD of quadruplicate values from one of three experiments, each yielding similar results. Bold values represent a statistically significant increase in cAMP levels, P Ͻ 0.01. The ability of 200 g/ml of each IgG to increase cAMP levels in the presence or absence of goat antihuman IgG (1:50 diluted Fab fragment) was measured as described in Materials and Methods. Data using IgGs from all 3 other myxedema patients were similar. Normal IgG was a pool from 20 normal individuals.