Type 3 autoimmune polyglandular syndrome (APS-3) or type 3 multiple autoimmune syndrome (MAS-3): an expanding galaxy

BACKGROUND

The number of recognised distinct autoimmune diseases (AIDs) has progressively increased over the years with more than 100 being reported today. The natural history of AIDs is characterized by progression from latent and subclinical to clinical stages and is associated with the presence of the specific circulating autoantibodies. Once presented, AIDs are generally chronic conditions. AIDs have the tendency to cluster and co-occur in a single patient. Autoimmune thyroid diseases (AITD) are the most prevalent of AIDs in the world population, and about one-third of the AITD patients also present with a non-thyroid AID during their life-span. Furthermore, patient with non-thyroid AIDs often presents with a form of AITD as a concurrent condition. Many of the clusters of AIDs are well characterized as distinctive syndromes, while some are infrequent and only described in case reports.

PURPOSE

In this review, we describe the wide spectrum of the combinations and the intricate relationships between AITD and the other AIDs, excluding Addison's disease. These combinations are collectively termed type 3 Autoimmune Polyglandular Syndrome (APS-3), also called type 3 Multiple Autoimmune Syndrome (MAS-3), and represent the most frequent APS in the world populations.

CONCLUSIONS

Numerous associations of AITD with various AIDs could be viewed as if the other AIDs were gravitating like satellites around AITD located in the center of a progressively expanding galaxy of autoimmunity.

Autoimmune polyendocrine syndrome type 1: an Italian survey on 158 patients

BACKGROUND

Autoimmune Polyglandular Syndrome type 1 (APS-1) is a rare recessive inherited disease, caused by AutoImmune Regulator (AIRE) gene mutations and characterized by three major manifestations: chronic mucocutaneous candidiasis (CMC), chronic hypoparathyroidism (CH) and Addison's disease (AD).

METHODS

Autoimmune conditions and associated autoantibodies (Abs) were analyzed in 158 Italian patients (103 females and 55 males; F/M 1.9/1) at the onset and during a follow-up of 23.7 ± 15.1 years. AIRE mutations were determined.

RESULTS

The prevalence of APS-1 was 2.6 cases/million (range 0.5-17 in different regions). At the onset 93% of patients presented with one or more components of the classical triad and 7% with other components. At the end of follow-up, 86.1% had CH, 77.2% AD, 74.7% CMC, 49.5% premature menopause, 29.7% autoimmune intestinal dysfunction, 27.8% autoimmune thyroid diseases, 25.9% autoimmune gastritis/pernicious anemia, 25.3% ectodermal dystrophy, 24% alopecia, 21.5% autoimmune hepatitis, 17% vitiligo, 13.3% cholelithiasis, 5.7% connective diseases, 4.4% asplenia, 2.5% celiac disease and 13.9% cancer. Overall, 991 diseases (6.3 diseases/patient) were found. Interferon-ω Abs (IFNωAbs) were positive in 91.1% of patients. Overall mortality was 14.6%. The AIRE mutation R139X was found in 21.3% of tested alleles, R257X in 11.8%, W78R in 11.4%, C322fsX372 in 8.8%, T16M in 6.2%, R203X in 4%, and A21V in 2.9%. Less frequent mutations were present in 12.9%, very rare in 9.6% while no mutations in 11% of the cases.

CONCLUSIONS

In Italy, APS-1 is a rare disorder presenting with the three major manifestations and associated with different AIRE gene mutations. IFNωAbs are markers of APS-1 and other organ-specific autoantibodies are markers of clinical, subclinical or potential autoimmune conditions.

Practical applications of studies on the TSH receptor and TSH receptor autoantibodies

Studies on the TSH receptor (TSHR) have numerous practical applications in vitro and in vivo. For example human monoclonal autoantibodies (MAbs) to the TSHR are useful reagents for in vitro diagnostics. Measurement of TSHR autoantibodies (TRAbs) is helpful in diagnosis and management of autoimmune thyroid disease. Currently available highly sensitive and specific assays to measure TRAbs use the human TSHR MAb M22 instead of the TSH. Furthermore, preparations of the human TSHR MAb M22 are useful as the World Health Organisation International Standard for thyroid stimulating antibody and for calibration of the assays for measuring TRAbs. Preparations of thermostabilised TSHR extracellular domain have recently become available and this is likely to have an impact on improvements in specificity testing for TRAb assays. In addition the stable TSHR preparations have practical application for specific immunoadsorption of patient serum TRAbs. Human TSHR MAbs also have promising prospects as new therapeutics. Autoantibodies with TSHR antagonistic activities are "natural" inhibitors of TSHR stimulation and are expected to be helpful in controlling TSHR activity in patients with Graves' disease, Graves' ophthalmopathy and thyroid cancer.

Epidemiology, pathogenesis, and diagnosis of Addison's disease in adults

BACKGROUND

Addison's disease (AD) is a rare disorder and among adult population in developed countries is most commonly caused by autoimmunity. In contrast, in children genetic causes are responsible for AD in the majority of patients.

PURPOSE

This review describes epidemiology, pathogenesis, genetics, natural history, clinical manifestations, immunological markers and diagnostic strategies in patients with AD. Standard care treatments including the management of patients during pregnancy and adrenal crises consistent with the recent consensus statement of the European Consortium and the Endocrine Society Clinical Practice Guideline are described. In addition, emerging therapies designed to improve the quality of life and new strategies to modify the natural history of autoimmune AD are discussed.

CONCLUSIONS

Progress in optimizing replacement therapy for patients with AD has allowed the patients to lead a normal life. However, continuous education of patients and health care professionals of ever-present danger of adrenal crisis is essential to save lives of patients with AD.

Mechanisms of Action of TSHR Autoantibodies

The availability of human monoclonal antibodies (MAbs) to the TSHR has enabled major advances in our understanding of how TSHR autoantibodies interact with the receptor. These advances include determination of the crystal structures of the TSHR LRD in complex with a stimulating autoantibody (M22) and with a blocking type autoantibody (K1-70). The high affinity of MAbs for the TSHR makes them particularly suitable for use as ligands in assays for patient serum TSHR autoantibodies. Also, M22 and K1-70 are effective at low concentrations in vivo as TSHR agonists and antagonists respectively. K1-70 has important potential in the treatment of the hyperthyroidism of Graves' disease and Graves' ophthalmopathy. Small molecule TSHR antagonists described to date do not appear to have the potency and/or specificity shown by K1-70. New models of the TSHR ECD in complex with various ligands have been built. These models suggest that initial binding of TSH to the TSHR causes a conformational change in the hormone. This opens a positively charged pocket in receptor-bound TSH which attracts the negatively charged sulphated tyrosine 385 on the hinge region of the receptor. The ensuing movement of the receptor's hinge region may then cause activation. Similar activation mechanisms seem to take place in the case of FSH and the FSHR and LH and the LHR. However, stimulating TSHR autoantibodies do not appear to activate the TSHR in the same way as TSH.

Gonadal function in males with autoimmune Addison's disease and autoantibodies to steroidogenic enzymes

Steroidogenic enzyme autoantibodies (SEAbs) are frequently present and are markers of autoimmune premature ovarian failure (POF) in females with autoimmune Addison's disease (AAD). The prevalence and significance of SEAbs in males with AAD have not yet been defined. We studied the prevalence of SEAbs in a large cohort of males with AAD and assessed the relationship between SEAbs positivity and testicular function. A total of 154 males with AAD (mean age 34 years) were studied. SEAbs included autoantibodies to steroid-producing cells (StCA), detected by immunofluorescence, and steroid 17α-hydroxylase (17α-OHAbs) and side chain cleavage enzyme (SCCAbs) measured by immunoprecipitation assays. Gonadal function was evaluated by measuring follicle-stimulating hormone (FSH), luteinizing hormone (LH), total testosterone (TT), sex hormone-binding globulin (SHGB), anti-müllerian hormone (AMH) and inhibin-B (I-B). Twenty-six males, 10 SEAbs((+)) and 16 SEAbs((-)), were followed-up for a mean period of 7·6 years to assess the behaviour of SEAbs and testicular function. SEAbs were found in 24·7% of males with AAD, with the highest frequency in patients with autoimmune polyendocrine syndrome type 1 (APS-1). The levels of reproductive hormones in 30 SEAbs((+)) males were in the normal range according to age and were not significantly different compared to 55 SEAbs((-)) males (P > 0·05). During follow-up, both SEAbs((+)) and SEAbs((-)) patients maintained normal testicular function. SEAbs were found with high frequency in males with AAD; however, they were not associated with testicular failure. This study suggests that the diagnostic value of SEAbs in males with AAD differs compared to females, and this may be related to the immunoprivileged status of the testis.

The dynamic changes of zinc transporter 8 autoantibodies in Czech children from the onset of Type 1 diabetes mellitus

AIMS

The prevalence of autoantibodies to zinc transporter 8 (ZnT8) in Czech children at the onset of Type 1 diabetes mellitus and dynamic changes in ZnT8 autoantibody levels during disease progression were studied. The value of ZnT8 autoantibody measurements in diagnosis of Type 1 diabetes was assessed.

METHODS

Serum samples from 227 children with newly diagnosed Type 1 diabetes and from 101 control children without diabetes were analysed in a retrospective cross-sectional study. One hundred and seventy-one samples from 116 of the patients with diabetes were analysed in a follow-up study at (median) intervals of 1, 3, 5 and 10 years after onset of Type 1 diabetes. ZnT8 autoantibodies were measured using a bridging enzyme-linked immunosorbent assay, while antibodies to glutamic acid decarboxylase, insulinoma antigen 2 and insulin were measured by radioimmunoassays.

RESULTS

ZnT8 autoantibodies were detected in 163/227 (72%) of children at Type 1 diabetes onset and in 1/101 (1%) of the control subjects. Sixteen out of 227 (7%) patients with Type 1 diabetes were antibody negative based on three antibodies (glutamic acid decarboxylase, insulinoma antigen 2 and insulin). This false-negative rate was reduced to 10/227 (4.4%) (P < 0.05) after inclusion of ZnT8 autoantibody measurements. Of the children, 142/227 (63%) were positive for at least three antibodies and the most common combination was insulinoma antigen 2, glutamic acid decarboxylase and ZnT8. ZnT8 autoantibody levels decreased over time after Type 1 diabetes onset and the presence and level of ZnT8 autoantibodies correlated with IA-2 autoantibodies.

CONCLUSIONS

A ZnT8 autoantibody enzyme-linked immunosorbent assay showed 72% disease sensitivity and 99% specificity at Type 1 diabetes onset. Measurements of ZnT8 autoantibodies are important for Type 1 diabetes diagnosis and should be included in the panel of autoantibodies tested at the onset of Type 1 diabetes.

Tryptophan hydroxylase autoantibodies as markers of a distinct autoimmune gastrointestinal component of autoimmune polyendocrine syndrome type 1

BACKGROUND

Autoantibodies to tryptophan hydroxylase (TPHAbs) directed against serotonin-producing enterochromaffin cells (EC) have been reported in autoimmune-polyendocrine-syndrome type 1 (APS-1) patients with gastrointestinal dysfunction (GID). Serotonin plays a critical role in enteric function and its peripheral blood levels reflect serotonin release from the gastrointestinal tract.

AIMS

We test the hypothesis that TPHAbs mark a distinct autoimmune component of APS-1 characterized by an autoimmune attack toward EC, which results in clinical GID.

METHODS

TPHAbs were measured in 64 APS-1 patients. Endoscopy with gastric (antrum/body) and duodenal biopsy was carried in 16 TPHAbs+ patients (8 with and 8 without GID) and in 2 TPHAbs- patients (without GID). Immunohistochemistry of biopsy specimens was carried out using antibodies to serotonin, chromogranin-A, CD3, CD4, CD8, and CD20. Serotonin serum levels were measured in TPHAbs+ and TPHAbs- patients who had endoscopy.

RESULTS

Thirty-seven of 64 patients were TPHAbs+ (11/12 with GID and 26/52 without GID; P < .001). Gastric and duodenal biopsies in all 8 TPHAb+ patients with GID showed lymphocytic infiltration with increased CD3+CD8+ intraepithelial lymphocytes and absence of EC. Furthermore, mean serotonin serum levels were below the normal range in TPHAb+ patients with GID (P < .01). In 8 TPHAb+ patients without GID gastric and duodenal biopsies showed different grades of inflammatory infiltration and reduced number of EC. Mean serotonin serum levels were near the lower limit of the normal range. In all TPHAbs+ patients the biopsies showed a reduced number of chromogranin-A positive cells consistent with enteroendocrine cells depletion. TPHAbs- patients without GID showed normal gastrointestinal mucosa and serotonin serum levels.

CONCLUSIONS

TPHAbs appear to be markers of a distinct autoimmune component of APS-1. Progressive involvement of the gastrointestinal EC leads to the transition from preclinical to clinical disease, characterized by GID and reduced serotonin serum levels.

Thyrotrophin receptor antibody characteristics in a woman with long-standing Hashimoto's who developed Graves' disease and pretibial myxoedema

CONTEXT

Sequential conversion of Hashimoto's thyroiditis (HT) to Graves' disease (GD) is uncommon. Distinct immune paradigms, paucity of functioning tissue in long-standing HT, and infrequent conversion of blocking (TBAb) to stimulating (TSAb) thyrotrophin receptor antibody (TRAb) may account for this. Molecular and crystal structure analysis helps delineate TSH receptor (TSHR)/TRAb interactions in detail. Such 'fingerprinting' helps determine the behaviour and characteristics of TRAb in longitudinal studies.

PATIENT

An 80-year-old woman taking thyroxine for long-standing HT became hyperthyroid. This persisted despite thyroxine withdrawal - free T3 was 7·3 pmol/l (2·6-5·7) and TSH < 0·01 mU/l (0·2-4·5) and TRAb highly positive. She had a goitre (ultrasound - HT), pretibial myxoedema, with mild inactive Graves' orbitopathy. She had RAI treatment and is on thyroxine replacement.

MEASUREMENTS AND RESULTS

Blood samples at presentation (A) and 1 year (B) showed high TSAb and TPOAb activity but no TBAb. Experiments involving TSHR mutations confirmed that (i) TRAb had stable characteristics over 1 year; (ii) TSHR mutation R255D caused complete inhibition and (iii) R109A caused marked reduction of cAMP production by M22 (TSHR-stimulating human monoclonal antibody) and A and B; (iv) mutations R80A, E107A and K129A while affecting M22 had little effect on A and B.

CONCLUSIONS

The reasons for an immunological paradigm shift in this elderly woman remain speculative. We believe that de-novo TSAb synthesis occurred converting her long-standing HT to GD although the mechanisms responsible remain unexplained. TRAb analysis confirmed stable autoantibody characteristics over 1 year and variable effects of TSHR mutations on TRAb and M22 function.

Premature ovarian failure in patients with autoimmune Addison's disease: clinical, genetic, and immunological evaluation

DESIGN

The design of the study was to investigate the prevalence of the following: 1) premature ovarian failure (POF) in patients with autoimmune Addison's disease (AD); 2) steroid-producing cell antibodies (StCA) and steroidogenic enzymes (17α-hydroxylase autoantibodies and P450 side-chain cleavage enzyme autoantibodies) in patients with or without POF; and 3) the value of these autoantibodies to predict POF.

PATIENTS

The study included 258 women: 163 with autoimmune polyendocrine syndrome type 2 (APS-2), 49 with APS-1, 18 with APS-4, and 28 with isolated AD.

METHODS

StCA were measured by an immunofluorescence technique and 17α-hydroxylase autoantibodies and P450 side-chain cleavage enzyme autoantibodies by immunoprecipitation assays.

RESULTS

Fifty-two of 258 women with AD (20.2%) had POF. POF was diagnosed in 20 of 49 (40.8%) with APS-1, six of 18 (33.3%) with APS-4, 26 of 163 (16%) with APS-2, and none of 28 with isolated AD. In patients with APS-1 and APS-4, POF developed after AD, whereas it preceded AD in patients with APS-2. StCA were detected in 31 of 43 with POF (72%) and 51 of 198 without POF (25.7%). StCA were present in 22 of 38 with APS-1 (57.9%) (11 of 13 with POF); in five of 13 with APS-4 (38.5%) (three of four with POF); in 53 of 162 with APS-2 (32.7%) (17 of 26 with POF), and in one of 28 isolated AD patients (3.6%). Twelve of 13 patients with POF with a duration less than 5 yr (92.3%) and 18 of 25 with duration longer than 5 yr (72%) were StCA positive. Twenty-eight of 31 with POF (90.3%) were positive for at least one steroidogenic antibody. Forty-one women with AD less than 40 yr were followed up for a mean period of 9 yr. Eight of 21 women (38%) positive or seroconverted for steroidogenic autoantibodies developed POF at a mean age of 23 yr (six with APS-1, one with APS-2, and one with APS-4), and none of the 20 patients negative for steroidogenic autoantibodies developed POF.

CONCLUSIONS

This study indicates that AD is frequently associated with POF and that steroidogenic antibodies are markers of patients with POF. Steroidogenic autoantibodies are predictive markers of POF in patients with AD.

TSH receptor - autoantibody interactions

TSH receptor (TSHR) autoantibodies (TRAbs)activate the TSHR cyclic AMP cascade (stimulating TRAbs) or act as TSHR antagonist (blocking TRAbs), and both types inhibit TSH binding to the TSHR. Isolation of human monoclonal TSHR autoantibodies (stimulating M22 and blocking 5C9) has been a key milestone in studies of the TSHR and TSHR autoimmunity. Comparison of M22 and TSH interactions with the TSHR at the atomic level reveal that M22 heavy and light chains mimic TSH alpha and beta chains, respectively, in the way they bind to the receptor, but the evolutionary forces which have caused this close molecular mimicry are as yet completely unknown. More recently two more human monoclonal antibodies to the TSHR (K1-18 with stimulating and K1-70 with blocking activities) have been isolated from a single blood sample collected from a patient with hypothyroidism who previously presented with hyperthyroidism. K1-18 and K1-70 were derived from different lymphocytes as shown by V region genes analysis. This provides, for the first time, clear proof that a patient can produce both blocking and stimulating TRAbs at the same time. Although it has been postulated that stimulating and blocking TRAbs bind to different regions on the TSHR, our studies showed that antibodies of both types bind well to the TSHR containing only N-terminal amino acids 22-260. Whether TRAbs make contact with other parts of the TSHR in order to produce their biological effects (stimulation or blocking) remains to be elucidated.

Thyroid stimulating autoantibody M22 mimics TSH binding to the TSH receptor leucine rich domain: a comparative structural study of protein-protein interactions

The TSH receptor (TSHR) ligands M22 (a thyroid stimulating human monoclonal antibody) and TSH, bind to the concave surface of the leucine rich repeats domain (LRD) of the TSHR and here, we show that M22 mimics closely the binding of TSH. We compared interactions produced by M22 with the TSHR in the M22-TSHR crystal structure (2.55 A resolution) and produced by TSH with the TSHR in a TSH-TSHR comparative model. The crystal structure of the TSHR and a comparative model of TSH based on the crystal structure of FSH were used as components to build the TSH-TSHR model. This model was built based on the FSH-FSH receptor structure (2.9 A) and then the structure of the TSHR in the model was replaced by the TSHR crystal structure. The analysis shows that M22 light chain mimics the TSHbeta chain in its interaction with TSHR LRD, while M22 heavy chain mimics the interactions of the TSHalpha chain. The M22-TSHR complex contains a greater number of hydrogen bonds and salt bridges and fewer hydrophobic interactions than the TSH-TSHR complex, consistent with a higher M22 binding affinity. Furthermore, the surface area formed by TSHR residues N208, Q235, R255, and N256 has been identified as a candidate target region for small molecules which might selectively block binding of autoantibodies to the TSHR.

FSH and TSH binding to their respective receptors: similarities, differences and implication for glycoprotein hormone specificity

The crystal structures of the leucine-rich repeat domain (LRD) of the FSH receptor (FSHR) in complex with FSH and the TSH receptor (TSHR) LRD in complex with the thyroid-stimulating autoantibody (M22) provide opportunities to assess the molecular basis of the specificity of glycoprotein hormone-receptor binding. A comparative model of the TSH-TSHR complex was built using the two solved crystal structures and verified using studies on receptor affinity and activation. Analysis of the FSH-FSHR and TSH-TSHR complexes allowed identification of receptor residues that may be important in hormone-binding specificity. These residues are in leucine-rich repeats at the two ends of the FSHR and the TSHR LRD structures but not in their central repeats. Interactions in the interfaces are consistent with a higher FSH-binding affinity for the FSHR compared with the binding affinity of TSH for the TSHR. The higher binding affinity of porcine (p)TSH and bovine (b)TSH for human (h)TSHR compared with hTSH appears not to be dependent on interactions with the TSHR LRD as none of the residues that differ among hTSH, pTSH or bTSH interact with the LRD. This suggests that TSHs are likely to interact with other parts of the receptors in addition to the LRD with these non-LRD interactions being responsible for affinity differences. Analysis of interactions in the FSH-FSHR and TSH-TSHR complexes suggests that the alpha-chains of both hormones tend to be involved in the receptor activation process while the beta-chains are more involved in defining binding specificity.

Islet cell, thyroid, adrenal and celiac disease related autoantibodies in patients with Type 1 diabetes from Sri Lanka

AIMS

The prevalence of islet cell, thyroid, adrenal and celiac disease related autoantibodies in patients with Type 1 diabetes mellitus (Type 1 DM) from Sri Lanka is described.

DESIGN AND METHODS

Autoantibodies to glutamic acid decarboxylase 65 (GAD65Ab), protein tyrosine phosphatase IA-2 (IA-2Ab), insulin (IAAb), thyroglobulin (TgAb), thyroid peroxidase (TPOAb), TSH receptor (TRAb), 21-hydroxylase (21-OHAb) and tissue transglutaminase (tTGAb) were measured in 122 Type 1 DM patients who had low C-peptide activity or were >20 yr old at the time of diagnosis and in 100 non-diabetic blood donors.

RESULTS

GAD65Ab and/or IA-2Ab were present in 74/122 (60.7%) Type 1 DM subjects with a significantly higher prevalence compared to non-diabetic controls (no. 100) (GAD65Ab-59 vs 4%; IA-2Ab-14 vs 0%; respectively) (p<0.001). The median (inter-quartile range) Type 1 DM duration in antibody positive subjects was 3.3 (0.99-6.9) vs 4.9 (1.7-7.5) yr in antibody negative subjects (p=0.23). IA-2Ab prevalence decreased with disease duration > or =5 yr (19 vs 4%) (p<0.001). There was no difference in the prevalence of TgAb (25 vs 33%)(p=0.21) and TPOAb (22 vs 18%) (p=0.48) in Type 1 DM and non-diabetic subjects. Also, there was no difference in TgAb and TPOAb prevalence in antibody positive Type 1 DM (34.7%) compared to antibody negative Type 1 DM (24.4%) subjects (p=0.24). tTGAb (3/119) and TRAb (1/119) were found in low prevalence and 21-OHAb were not detected.

CONCLUSIONS

Diabetes associated autoantibodies were detected in the majority of Type 1 DM subjects, suggesting a major role for autoimmunity in the pathogenesis of Type 1 DM in Sri Lankans. The prevalence of TgAb and TPOAb in Type 1 DM subjects and non-diabetic controls was relatively high and similar in both groups.

Effects of a thyroid-stimulating human monoclonal autoantibody (M22) on functional activity of LH and FSH receptors

OBJECTIVE

The glycoprotein hormones luteinizing hormone (LH), follicle-stimulating hormone (FSH), and thyrotropin (TSH) show low-level cross-reactivity between their respective receptors (R). Patient serum autoantibodies to the thyrotropin receptor (TSHR) do not appear to cross-react with the luteinizing hormone receptor (LHR) or follicle-stimulating hormone receptor (FSHR), although the concentrations of autoantibody with which it is feasible to carry out experiments of this type are limited. Consequently, we have studied the effects of high doses of the thyroid-stimulating human monoclonal autoantibody (M22) on the LHR and FSHR.

DESIGN

Chinese Hamster ovary (CHO) cells stably expressing the TSHR, LHR, and FSHR and purified M22 IgG preparations were used in the study.

METHODS

CHO-TSHR, CHO-LHR, and CHO-FSHR cells were incubated with bovine TSH (0.1-25mU/mL), human recombinant chorionic gonadotropin (hCG; 0.5-10mU/mL) or human recombinant FSH (100-5000mU/mL) or with M22 IgG (0.001-5.0 microg/mL), and the extracellular cyclic AMP was measured by radioimmunoassay.

RESULTS

Cyclic AMP levels increased in a dose-dependent manner after incubation of CHO-TSHR cells with TSH or M22 IgG, and on a molar basis the effects of TSH and M22 were similar. Cyclic AMP stimulation was not detectable in CHO-LHR and CHO-FSHR cells after incubation with M22 IgG, whereas incubation with hCG or FSH, respectively, caused dose-dependent cyclic AMP stimulation. On a molar basis, concentrations of M22 IgG approximately 100x those of FSH causing clear stimulation were ineffective with CHO-FSHR cells. Similarly, molar concentration of M22 IgG 20,000x those of hCG causing clear stimulation had no effect on CHO-LHR cells.

CONCLUSIONS

This study shows that at relatively high concentrations, M22 IgG is unable to stimulate cyclic AMP levels in CHO-LHR or CHO-FSHR cells, suggesting that TSHR autoantibodies have greater specificity for the TSHR than TSH itself.

Characteristics of a monoclonal antibody to the thyrotropin receptor that acts as a powerful thyroid-stimulating autoantibody antagonist

Analysis of nine mouse monoclonal antibodies (mAbs) to the thyrotropin receptor (TSHR) with TSH antagonist activity showed that only one of the mAbs (RSR B2) was an effective antagonist of the human thyroid stimulating autoantibody M22. Crystals of B2 Fab were analyzed by x-ray diffraction and a crystal structure at 3.3 A resolution was obtained. The surface charge and topography of the B2 antigen binding site were markedly different from those of the thyroid-stimulating mAb M22 and these differences might contribute to the different properties of the two mAbs. B2 (but not other mouse TSHR-specific mAbs) was also an effective antagonist of thyroid stimulating autoantibody activity in 14 of 14 different sera from patients with Graves' disease. 125I-labeled B2 bound to the TSHR with high affinity (2 x 10(10) L/mol) and patient serum TSHR autoantibodies inhibited labeled B2 binding to the receptor in a similar way to inhibition of labeled TSH binding (r = 0.75; n = 20). Furthermore, labeled B2 binding was inhibited by patient serum TSHR autoantibodies with TSH antagonist activity and also by mouse and human thyroid stimulating mAbs. Overall, mAb B2 is a powerful antagonist of thyroid stimulating autoantibodies (and TSH) thus resembling closely patient serum TSH antagonist TSHR autoantibodies. Furthermore, B2 might have potentially important in vivo applications when tissues containing the TSHR (including those in the orbit) need to be made unresponsive to stimulating autoantibodies.

Isolation and characterisation of a human monoclonal autoantibody to the islet cell autoantigen IA-2

A hybridoma secreting a human monoclonal autoantibody to the islet cell autoantigen IA-2 was prepared from peripheral lymphocytes of a patient with type 1 diabetes and Graves' disease using EBV infection followed by fusion with a mouse/human hybrid cell line. The monoclonal antibody (M13) is an IgG1/kappa and in an immunofluorescence test M13 at 1 microg/mL showed islet cell antibody reactivity equivalent to 40 JDF units. M13 IgG bound (35)S-labelled IA-2 (26% at 100 microg/mL) and (125)I-labelled IA-2 (34% at 100 microg/mL) in an immunoprecipitation assay and reacted well with IA-2 in western blotting analysis. Amino acids 777-808 in the PTP domain of IA-2 were found to be important for M13 binding in an analysis using modified (35)S-labelled IA-2 proteins. M13 V region genes were from VH1-3, D3-22, JH4b, VKI DPK8/Vd+ and JK3 genes and showed a high replacement/silent mutation ratio for both the heavy (11.0) and the light (6.0) chain genes. Mouse monoclonal antibodies (mMAbs) reactive with at least three different epitopes within IA-2 aa 604-686 corresponding to the juxtamembrane domain were also obtained. F(ab')(2) or Fab from the mMAbs inhibited serum IA-2 autoantibody binding to IA-2 in 20/22 diabetic sera whereas M13 F(ab')(2) caused inhibition in only 6/22 sera. M13 is representative of some patient serum IA-2 autoantibodies and as such provides a useful tool to study autoimmune responses to IA-2.

Mechanism of inhibition of cytochrome P450 C21 enzyme activity by autoantibodies from patients with Addison's disease

OBJECTIVE

To study possible mechanisms for the inhibition of cytochrome P450 C21 (steroid 21-hydroxylase) enzyme activity by P450 C21 autoantibodies (Abs) in vitro.

DESIGN

Two possible mechanisms for the inhibition of P450 C21 enzyme activity by P450 C21 Abs were studied: (a) conformational changes in the P450 C21 molecule induced by Ab binding and (b) the effects of Ab binding to P450 C21 on the electron transfer from the nicotinamide adenine dinucleotide phosphate reduced (NADPH) cytochrome P450 reductase (CPR) to P450 C21.

METHODS

The effect of P450 C21 Ab binding on the conformation of recombinant P450 C21 in yeast microsomes was studied using an analysis of the dithionite-reduced CO difference spectra. The effect of P450 C21 Abs on electron transfer was assessed by analysis of reduction of P450 C21 in the microsomes in the presence of CO after addition of NADPH.

RESULTS

Our studies confirmed the inhibiting effect of P450 C21 Abs on P450 C21 enzyme activity. Binding of the Abs did not induce significant change in the P450 C21 peak at 450nm (native form) and did not produce a detectable peak at 420 nm (denatured form) in the dithionite-reduced CO difference spectra. This indicated that conformation of P450 C21 around the heme was not altered compared with the native structure. However, incubation of the P450 C21 in yeast microsomes with P450 C21 Ab inhibited the fast phase electron transfer from the CPR to P450 C21.

CONCLUSIONS

Our observations suggested that the mechanism by which P450 C21 Abs inhibit P450 C21 enzyme activity most likely involves inhibition of the interaction between the CPR and P450 C21.

Characteristics of a human monoclonal autoantibody to the thyrotropin receptor: sequence structure and function

The properties of a human monoclonal antibody to the thyrotropin receptor (TSHR) (M22) with the characteristics of patient sera thyroid stimulating autoantibodies is described. Similar concentrations (pmol/L) of M22 Fab and porcine TSH had similar stimulating effects on cyclic adenosine monophosphate (cAMP) production in TSHR-transfected Chinese hamster ovary cells whereas higher doses of intact M22 immunoglobulin G (IgG) were required to cause the same level of stimulation. Patient sera containing TSHR autoantibodies with TSH antagonist (blocking) activity inhibited M22 Fab and IgG stimulation in a similar way to their ability to block TSH stimulation. Thyroid-stimulating monoclonal antibodies (TSmAbs) produced in mice inhibited 125I-TSH binding and 125I-M22 Fab binding to the TSHR but the mouse TSmAbs were less effective inhibitors than M22. These competition studies emphasized the close relationship between the binding sites on the TSHR for TSH, TSHR autoantibodies with TSH agonist activity, and TSHR autoantibodies with TSH antagonist activity. Recombinant M22 Fab could be produced in Escherichia coli and the recombinant and hybridoma produced Fabs were similarly active in terms of inhibition of TSH binding and cAMP stimulation. The crystal structure of M22 Fab was determined to 1.65 A resolution and is that of a standard Fab although the hypervariable region of the heavy chain protrudes further from the framework than the hypervariable region of the light chain. The M22 antigen binding site is rich in aromatic residues and its surface is dominated by acidic patches on one side and basic patches on the other in agreement with an important role for charge-charge interactions in the TSHR-autoantibody interaction.

Assessment of adrenocortical function and autoantibodies in a baby born to a mother with autoimmune polyglandular syndrome Type 2

We describe the case of a baby born to a mother with Addison's disease in the context of Autoimmune Polyendocrine Syndrome Type 2. Adrenal cortex autoantibodies and steroid 21-hydroxylase autoantibodies were detectable in the sera of both mother and baby, suggesting the transplacental passage of these autoantibodies. Adrenal autoantibodies were present in the baby's serum at delivery, at 3, 6 and till 34 months of age but no signs of clinical or subclinical adrenal insufficiency were found in the baby during the observation period. These data suggest that the presence of adrenal autoantibodies in serum alone is not a sufficient cause for the development of autoimmune adrenalitis.

Human monoclonal thyroid stimulating autoantibody

A monoclonal autoantibody (MAb) with powerful thyroid stimulating activity has been produced from lymphocytes from a patient with Graves' disease. The autoantibody and its Fab fragment bind to the thyroid stimulating hormone (TSH) receptor (TSHR) with high affinity, inhibit labelled TSH binding to the receptor and stimulate cyclic AMP production in Chinese hamster ovary cells transfected with TSHR. TSHR autoantibodies with TSH agonist or antagonist activities from patients' serum samples are effective inhibitors of labelled monoclonal autoantibody binding to TSHR. Thus, the human monoclonal autoantibody has all the characteristics of serum TSHR autoantibodies. Its availability has important implications for new studies on the pathogenesis of Graves' disease.

Autoantibodies to steroidogenic enzymes in patients with premature ovarian failure with and without Addison's disease

DESIGN

Adrenal cortex autoantibodies (ACA), steroid-producing cell autoantibodies (StCA) and autoantibodies (Abs) to steroidogenic enzymes in three groups of patients with premature ovarian failure (POF), 15 with autoimmune Addison's disease (AD), 26 with non-adrenal autoimmune diseases and 31 with isolated POF, have been assessed.

METHODS

ACA and StCA were measured using an immunofluorescence technique. Abs to 21-hydroxylase (21-OH), to 17alpha-hydroxylase (17alpha-OH) and to cytochrome P450 side-chain cleavage (P450scc) were measured using an immunoprecipitation assay.

RESULTS

Seventy-three percent of patients with POF and AD were positive for StCA, 93% for 17alpha-OH and/or P450scc Abs, 93% for ACA and 100% for 21-OH Abs. Among patients with POF and non-adrenal autoimmune diseases, 8% were positive for StCA, 12% for 17alpha-OH and/or P450scc Abs, and 8% and 12% for ACA and 21-OH Abs respectively. StCA, 17alpha-OH and/or P450scc Abs were all found in 10% of patients with isolated POF, and 13% had ACA and 21-OH Abs. All StCA-, 17alpha-OH- and/or P450scc Abs-positive patients were also positive for ACA and 21-OH Abs. Two patients with isolated POF who were ACA and 21-OH Ab positive developed AD 3 and 5 Years after the onset of POF.

CONCLUSION

This study has shown that, when POF is associated with AD, StCA, 17alpha-OH and/or P450scc Abs are present in the majority of patients, while in the other two groups these Abs are detectable in a much lower proportion of patients. Measurement of ACA/21-OH Abs in some patients with POF may be important in identifying patients at risk of developing overt AD.

Isolation and characterization of human monoclonal autoantibodies to glutamic acid decarboxylase

Production of human monoclonal autoantibodies to glutamic acid decarboxylase M(r) 65,000 (GAD65), characterization of their isotype, binding affinity, V region sequences and competition with autoantibodies in patients' sera is described. Lymphocytes from a patient with Addison's disease who had GAD65 autoantibodies without diabetes were immortalised and fused to a mouse/human hybridoma. In addition, mouse monoclonal antibodies to GAD65 were produced using standard techniques. F(ab')2S from our monoclonals and the GAD6 mouse monoclonal were used in competition with intact monoclonals and sera from diabetic patients for binding to 125I-labelled GAD65 (amino acids 46-586). Reactivities of the human monoclonals with GAD 65,000/67,000 M(r) chimeras were also studied. Variable region genes of human monoclonals were sequenced and analysed. The human monoclonals (n = 3) had affinity constants for GAD65 of 2.2 x 10(9), 5.8 x 10(9), 1.3 x 10(10) mol/l(-1); affinities of the mouse monoclonals (n = 5) ranged from 1.1 x 10(8) to 5.4 x 10(10) mol/l(-1). The binding of each of the human monoclonals was inhibited by GAD6 F(ab')2 and the binding of GAD6 antibody was inhibited by the human monoclonal F(ab')2S suggesting that the epitopes for these antibodies were overlapping. Studies with GAD65/GAD67 chimeras indicated that the human monoclonals reacted with C-terminal epitopes. The human monoclonals, GAD6 and 3/5 mouse monoclonals inhibited serum autoantibody binding to 125I-labelled GAD65. Overall, the human monoclonals were of high affinity, reacted with C-terminal epitopes and showed evidence of antigen driven maturation; they represented only a proportion of the repertoire of autoantibodies to GAD65 in the donor's serum and in the sera of patients with type-1 diabetes.

X-ray crystal structure of a monoclonal antibody that binds to a major autoantigenic epitope on thyroid peroxidase

Thyroid peroxidase (TPO) catalyzes the production of thyroid hormones and is a major autoantigen in autoimmune thyroid disease (AITD). It is believed that the majority of TPO autoantibodies bind to an immunodominant region consisting of two overlapping domains. Precise location of these domains would help our understanding of the interaction between TPO and TPO autoantibodies. 4F5 is a mouse monoclonal antibody (IgG1, kappa) that reacts with high affinity (2.6 x 10(10) mol/L(-1)) with one of the major autoantigenic regions on TPO. Heavy chain genes of 4F5 were from the VH1 germline gene family, germline genes for the D region could not be assigned and the J region was from the JH2 germline. Light chain genes were from Vkappa4/5 and Jkappa2, germline gene families. The Fab fragment of 4F5 was prepared by papain digestion, purified, crystallized, and the structure solved to 1.9 A using molecular replacement. The refined structure had an R factor of 19.5% and a free R factor of 23.9%. Deduced amino acid sequence and amino acid sequence obtained from diffraction analysis were compared and used to finalize the 4F5 Fab model. Structural analysis indicated that the structure of 4F5 is that of a standard Fab and its combining site is flat and is rich in tyrosine residues. Comparison of the structure of 4F5 with that of a TPO autoantibody Fab, TR1.9 suggests that the two antibodies are unlikely to recognise the same structures on TPO.

Reactivity of thyrotropin receptor autoantibodies with the thyrotropin receptor on western blots

Affinity purified recombinant human thyrotropin receptor (TSHR) was run on sodium dodecyl sulfate (SDS) gels and subjected to a renaturing and blotting procedure. Twenty sera from thyrotropin receptor autoantibodies (TRAb)-positive patients with a history of hyperthyroidism and 20 sera with high levels of TSH blocking activity were analyzed. Four of 20 sera with blocking-type of TRAb (i.e., TSH antagonist activity) were able to recognize the mature, fully glycosylated 120-kd form of the receptor on blots of gels run under reducing conditions. No sera recognized the 100-kd high mannose precursor form of the TSHR. Three of the four recognized a 74-kd band and 2 of the 4 recognized a 50-kd band. These bands are probably proteolytic cleavage fragments of the mature 120-kd TSHR. In the absence of reducing agent the same 4 of 20 sera described above together with a further serum sample (i.e., 5/20 in total) reacted with the 120-kd form of the receptor. No specific reaction with the TSHR was observed on Western blots with the remaining 15 sera with TSH blocking activity, nor with 20 sera from patients with a history of hyperthyroidism, nor with sera from 10 healthy blood donors, 10 Hashimoto sera (negative for TRAb) and 10 systemic lupus erythematosus sera. No clear differences were observed in the TRAb positive sera that were reactive and nonreactive on Western blots in terms of their ability to inhibit TSH binding or to immunoprecipitate 125I-labeled TSHR. Overall, our results indicate that the mature 120-kd form of the TSHR that is principally responsible for binding TSH is also responsible for binding TRAb (when this binding can be detected). These observations together with immunoprecipitation and TSH binding inhibition studies, emphasize the close relationship between the receptor's binding sites for TSH and TRAb.

Antibodies to tissue transglutaminase: comparison of ELISA and immunoprecipitation assay in the presence and in the absence of calcium ions

Endomysial antibodies are characteristic of coeliac disease and tissue transglutaminase (tTG) has been identified as a major component of the endomysial antigen. tTG autoantibodies were measured in sera from patients with coeliac disease using ELISA based on guinea pig tTG and immunoprecipitation assay (IPA) based on 35S-labelled human tTG produced in an in vitro transcription/translation system. In addition, the effect of calcium ions on the interaction between tTG autoantibodies in the two assays was studied. Under standard (i.e. Ca2+-free) conditions, 36/39 (92%) coeliac sera were positive for IgA tTG antibodies by ELISA and 34/39 (87%) sera were positive by IPA. Comparison of ELISA and IPA results showed three sera positive by ELISA but negative by IPA and one serum which was positive by IPA but negative by ELISA. Bland and Altman analysis of the correlation between the ELISA and IPA showed that the results for 37 out of 39 samples were in the agreement. The results by ELISA carried out without and with Ca2+ were in good agreement (r=0.99; n=39). IPA using Ca2+ containing buffer detected fewer samples compared to IPA using standard assay buffer however the results of the two assays also showed a good agreement (r=0.93; n=39). Our studies confirm that antibodies to tTG are good markers of coeliac disease and indicate that the autoantibody binding sites on tTG are formed in a way which is essentially independent of Ca2+.

Epitope analysis of the human thyrotropin (TSH) receptor using monoclonal antibodies

A panel of thyrotropin (TSH) receptor (TSHR) monoclonal antibodies (mAbs), produced using highly purified Chinese hamster ovary (CHO) cell-produced TSHR, has been used to study TSHR structure. All 41 mAbs recognized full-length TSHR containing complex carbohydrate (120 kDa), and 40 mAbs recognized full-length precursor-containing high mannose sugars (100 kDa). The mAbs also recognized TSHR cleavage products with three types of reactivity: type 1 mAbs reacting with bands at 70 kDa and 58 kDa, type 2 with bands at 70 kDa and 52 kDa, and type 3 with bands at 52 kDa and 40 kDa. Deglycosylation studies showed that the 70-kDa and 58-kDa bands contained complex carbohydrate, whereas the 52-kDa and 40-kDa bands were unglycosylated. These results are consistent with TSHR cleavage occurring at two sites. Cleavage at both sites gives rise to glycosylated A subunit (58 kDa) corresponding to the extracellular domain of the receptor and nonglycosylated B subunit (40 kDa) corresponding to the C-terminal transmembrane domain. Cleavage only at site 1 gives rise to the 58-kDa A subunit and a large B subunit (52 kDa). Cleavage only at site 2 gives rise to a large A subunit (70 kDa) and the B subunit (40 kDa). Four of the mAbs inhibited 125I-labeled TSH binding to solubilized full-length TSHR. TSH binding was inhibited by (a) two type 3 mAbs reactive with the N-terminal region of the B subunit (epitopes between amino acids 381 and 385 and between 380 and 418, respectively) and (b) two type 2 mAbs reactive with epitopes on the A subunit (between amino acids 246 and 260). These results together with previous studies on the direct binding of TSH to the TSHR A subunit suggest that at least two distinct regions of the TSHR sequence, including one region on the A subunit and one region on the B subunit, fold together to form part of a complex TSH binding site.

Recombinant human thyroid peroxidase produced in insect cells has similar properties to native human thyroid peroxidase

Purified native human thyroid peroxidase (nTPO) isolated from thyroid tissue and recombinant (r)TPO produced in High Five insect cells have been compared. nTPO and rTPO were purified to about 95% homogeneity and showed similar UV and visual spectra and similar 412 nm per 280 nm absorbance ratios (0.4 for nTPO and 0.4 for rTPO). The nTPO and rTPO guaiacol oxidation enzyme activities were about 1,000 guaiacol units per milligram of protein. TPO autoantibody binding characteristics of nTPO and rTPO were analyzed in an assay based on 125I-labeled nTPO and precipitation with protein A. In the assay, the effect of unlabeled nTPO or rTPO on TPO autoantibody binding from 25 patients sera was studied. Unlabeled nTPO or rTPO (from 0 to 160 ng/mL) inhibited the binding of TPO autoantibodies in a dose-dependent manner in the case of each serum studied (from 100% in the absence of unlabeled TPO to 5%-10% in the presence of 160 ng/mL of TPO). The inhibition profile for each serum was essentially identical in the case of both TPO preparations. The effect of TPO autoantibodies on enzyme activity of rTPO was analyzed after incubation of rTPO with TPO autoantibody-positive serum immunoglobulin G (IgG) (n = 12), TPO monoclonal antibodies reactive with two different epitopes on the TPO, IgG (n = 3) from glutamic acid decarboxylase autoantibody positive patient sera, and IgG (n = 3) from healthy blood donors. Effective complexing of TPO by TPO autoantibodies was tested by precipitating the complexes with solid phase protein A and measuring the TPO enzyme activity in the resulting supernatants. These studies showed that the TPO enzyme activity was not affected by incubation with TPO autoantibody-positive IgG or monoclonal antibodies despite effective complexing of the autoantibodies with TPO. Overall, our studies demonstrate that nTPO and rTPO produced in insect cells are very similar in terms of enzyme activity, UV and visible spectra, and reactivity with autoantibodies. Furthermore, in our study, TPO autoantibodies did not appear to inhibit TPO enzyme activity.

Autoantibodies to IA-2 in insulin-dependent diabetes mellitus. Measurements with a new immunoprecipitation assay

An immunoprecipitation assay for autoantibodies (Abs) to the human islet cell antigen IA-2 has been developed using 125I-labelled recombinant IA-2 expressed in E. coli. With this assay IA-2 Abs were detected in 103/217 (47%) of IDDM patients of different ages and with different disease duration. IA-2 Ab prevalence was higher in younger patients (at the age of 15 years or below) with the recent onset IDDM (64/113; 57%) compared to patients above the age of 15 years (11/25; 44%). One of 40 (2.5%) Graves' disease patients and five of 204 (2.5%) of NIDDM patients were also positive. IA-2 Abs were not detected in sera from patients with Hashimoto's thyroiditis (n=32), myasthenia gravis (n=20) or systemic lupus erythematosus (n=10). IA-2 Ab measurements based on 125I-labelled IA-2 showed a good correlation with the results of an immunoprecipitation assay based on 35S-labelled IA-2 produced in the in vitro transcription/translation system (r=0.78; n=113; p<0.001). Out of 217 IDDM sera which were tested for IA-2 Abs, 140 (65%) were positive for Abs to glutamic acid decarboxylase (GAD) and 166 (76%) were positive for Abs to IA-2 and/or Abs to GAD. In addition, Abs to IA-2, to GAD and to insulin were analysed in sera from recent onset IDDM patients who had not been treated with insulin (n=117). In all, 76/117 (65%) of these sera were positive for GAD Abs, 66/117 (56%) for IA-2 Abs, 45/117 (38%) for insulin Abs. However, 98/117 (84%) were positive for at least one of the three Abs confirming earlier observations on the complementarity of Ab testing in IDDM. Overall, the IA-2 Ab assay based on 125I-labelled recombinant IA-2 showed good sensitivity, precision and specificity which, combined with an easy and convenient protocol, makes it attractive for routine use.

Analysis of autoantibody epitopes on human thyroid peroxidase

A number of studies have indicated that the major autoantibody epitopes on human thyroid peroxidase (TPO) are conformational and are formed by two overlapping immunodominant regions on the TPO molecule. In order to investigate further autoantibody reactivity with TPO, we have studied the TPO binding characteristics of sera from patients with autoimmune thyroid disease (n = 20), autoimmune adrenal disease (Addison's disease; n = 8) and apparently healthy blood donors (n = 9) using recombinant TPO expressed with a series of truncations and internal deletions. This material was obtained using an in vitro transcription/translation system in the presence of 35S-methionine and the reactivity of TPO autoantibodies tested in an immunoprecipitation assay. In addition, we have studied the effects of denaturing purified recombinant TPO by reduction and/or sodium dodecyl sulphate on its reactivity with TPO autoantibodies by Western blotting analysis. These studies show that TPO autoantibodies can recognise TPO in Western blotting analysis when large amounts of purified TPO are run on the gels and the blotted proteins renatured prior to addition of antibody. Under these conditions TPO autoantibodies in all 20 Graves' or Hashimoto's sera tested reacted strongly with blots of non-reduced TPO but reduction of TPO had a marked effect on the ability of autoantibodies to recognise it in Western blotting analysis. Analysis of TPO autoantibody binding to 35S-labelled TPO proteins containing N-terminal, central or C-terminal deletions indicated that all modifications studied caused a statistically significant lowering of binding. In the case of some modifications, there were differences in the reactivity of TPO autoantibodies in sera from patients with Addison's disease compared to TPO autoantibodies in autoimmune thyroid disease and/or healthy blood donor sera. Overall, our results of analysis of T PO autoantibody binding in Western blotting and with modified TPO proteins in immunoprecipitation assays suggest that the main autoantibody binding sites on the TPO molecule involve extensive amino acid sequences. Our studies also suggest that TPO autoantibodies from patients with autoimmune thyroid disease, Addison's disease and apparently healthy blood donors show some differences in epitope recognition on TPO and this approach may allow differentiation between disease related and unrelated TPO autoantibodies.

The interaction of TSH receptor autoantibodies with 125I-labelled TSH receptor

Detergent-solubilized porcine TSH receptor (TSHR) has been labeled with 125I using a monoclonal antibody to the C-terminal domain of the receptor. The ability of sera containing TSHR autoantibody to immunoprecipitate the labeled receptor was then investigated. Sera negative for TSHR autoantibody (as judged by assays based on inhibition of labeled TSH binding to detergent-solubilized porcine TSHR) immunoprecipitated about 4% of the labeled receptor, whereas sera with high levels of receptor autoantibody immunoprecipitated more than 25% of the labeled receptor. The ability to immunoprecipitate labeled TSHR correlated well with ability of the sera to inhibit labeled TSH binding to the receptor (r = 0.92; n = 63), and this is consistent with TSHR autoantibodies in these samples being directed principally to a region of the receptor closely related to the TSH binding site. Preincubation of labeled TSHR with unlabeled TSH before reaction with test sera inhibited the immunoprecipitation reaction, providing further evidence for a close relationship between the TSHR autoantibody binding site(s) and the TSH binding site. This was the case whether the sera had TSH agonist (i.e., thyroid stimulating) or TSH antagonist (i.e., blocking) activities, thus, providing no clear evidence for different regions of the TSHR being involved in forming the binding site(s) for TSHR autoantibodies with stimulating and with blocking activities. The ability of TSHR autoantibodies to stimulate cyclic AMP production in isolated porcine thyroid cells was compared with their ability to immunoprecipitate labeled porcine TSHR. A significant correlation was observed (r = 0.58; n = 50; P < 0.001) and the correlation was improved when stimulation of cyclic AMP production was compared with inhibition of labeled TSH binding to porcine TSHR (r = 0.76). Overall, our results indicate that TSHR autoantibodies bind principally to a region on the TSHR closely related to the TSH binding site, and this seems to be the case whether the autoantibodies act as TSH agonists or antagonists.

Analysis of carbohydrate residues on recombinant human thyrotropin receptor

An investigation of the sugar groups on recombinant human TSH receptors (TSHR) expressed in CHO-K1 cells and solubilized with detergents is described. Western blotting studies with TSHR monoclonal antibodies showed that the receptor was present principally as two bands with approximate molecular masses of 120 and 100 kDa. Further blotting studies using lectins and/or involving treatment with different glycosidases indicated that the 100-kDa band contained about 16 kDa of high mannose-type sugars, and the 120-kDa band contained about 33 kDa of complex-type sugars. It was possible to separate the 120- and 100-kDa components of the TSHRs by lectin affinity chromatography. In particular, Galanthus nivalis lectin, which binds high mannose-type sugars, bound the 100-kDa band, but not the 120-kDa band, whereas Datura stramonium lectin, which binds complex-type sugars, bound the 120-kDa band, but not the 100-kDa band. 125I-Labeled TSH binding studies with the various lectin column fractions showed that TSH-binding activity was principally associated with the complex-type sugar containing the 120-kDa form of the receptor rather than the high mannose-containing 100-kDa form. During peptide chain glycosylation, high mannose-type sugar residues are attached first and then modified by the formation of complex type structures to form the mature glycoprotein. Our data suggest that in the case of the TSH receptor, this type of posttranslational processing has an important role in forming the TSH-binding site.

Recombinant human thyroid peroxidase expressed in insect cells is soluble at high concentrations and forms diffracting crystals

Human thyroid peroxidase (TPO), the key enzyme in thyroid hormone synthesis, can be produced in active form in the High Five insect cell line and when purified from the cell culture medium is soluble at concentrations of up to 18 mg/ml. This contrasts to a recent report in which human TPO produced in insect cells was found to be insoluble at high concentrations. Our concentrated TPO grows trigonal trapezohedral crystals of up to 0.5 mm in length in a vapour diffusion apparatus using polyethelene glycol as a precipitant. The crystals diffract X-rays to a 6 A resolution and the diffraction data from the crystals have been analysed giving unit cell dimensions. A potential molecular replacement solution has been identified using myeloperoxidase (MPO) as a phasing model.

Adrenal-cortex autoantibodies and steroid-producing cells autoantibodies in patients with Addison's disease: comparison of immunofluorescence and immunoprecipitation assays

Autoimmune Addison's disease and premature ovarian failure are characterized by the presence of organ-specific autoantibodies. The main adrenal and gonadal autoantigens have been identified and cloned, and the relationship between the autoantibodies detected by immunofluorescence techniques and those detected by the new assays using recombinant autoantigens needed to be investigated. We studied 165 patients with Addison's disease: 143 patients had different forms of autoimmune Addison's disease (13 with idiopathic premature ovarian failure) and 22 had nonautoimmune Addison's disease. Adrenal-cortex autoantibodies and steroid-producing cell autoantibodies were measured by the immunofluorescence techniques. Autoantibodies to steroid 21-hydroxylase, 17alpha-hydroxylase, and P450 side chain cleavage enzyme were measured by immunoprecipitation assay using 35S-labeled recombinant proteins. Adrenal-cortex autoantibodies and autoantibodies to 21-hydroxylase were found in 81% of the patients with autoimmune Addison's disease. None of the patients with nonautoimmune Addison's disease had adrenal-cortex autoantibodies or autoantibodies to 21-hydroxylase. A high association between these two markers in patients with different forms of autoimmune Addison's disease and in those with short- or long-standing disease was found. Steroid-producing cells autoantibodies were found in 26% of the patients with autoimmune Addison's disease, and autoantibodies to 17alpha-hydroxylase and/or P450 side chain cleavage enzyme in 36% of the patients. Steroid-producing cells autoantibodies were found in 11/13 (85%) of patients with idiopathic premature ovarian failure associated with autoimmune Addison's disease, and autoantibodies to 17alpha-hydroxylase and/or P450 side chain cleavage were found 12/13 (92%) of patients; the only case negative for all these three markers suffered from Turner's syndrome. Provided that a high standard of immunofluorescence technique is maintained, measurement of adrenal cortex autoantibodies or steroid-producing cells autoantibodies by either immunofluorescence or immunoprecipitation assay is essentially equivalent.

Analysis of autoantibody epitopes on steroid 21-hydroxylase using a panel of monoclonal antibodies

A panel of five mouse monoclonal antibodies (MAbs) to human recombinant steroid 21-hydroxylase (21-OH) were produced, characterized, and used to study the interaction of 21-OH autoantibodies (AAbs) with different epitopes on human 21-OH. AAbs in patients with isolated autoimmune Addison's disease, autoimmune polyglandular syndromes types I and II, and 21-OH antibody-positive patients without overt Addison's disease (25 patients in total) were studied. Four MAbs were IgG1 subclass, one was IgG2a, and all had kappa light chains. The affinities of four of the antibodies were in the range 2.0 x 10(8) M(-1) to 7.0 x 10(8) M(-1), and the affinity of the other was 2.3 x 10(7) M(-1) 21-OH MAbs did not cross-react with 17alpha-hydroxylase (17alpha-OH)) or P450 side chain cleavage enzyme. Studies using a series of 21-OH fragments allowed the identification of short stretches of amino acids (AA) that were involved in forming the MAb binding sites. AA 391-405, defined as epitope region (ER) 1, were found to be important for binding of M21-OH1 and M21-OH2, AA 406-411 (ER2) were important for M21-OH3 and M21-OH4 binding, and AA 335-339 (ER3) for M21-OH5 binding. In addition, MAb Fab or F(ab')2 fragments were used to study 21-OH AAb epitopes in competition experiments. These investigations demonstrated that 21-OH AAbs recognize similar epitopes to the MAbs, with ER2 and ER3 being part of two distinct major epitopes, and ER 1 being part of a minor epitope. Mixtures of M21-OH antibody Fab or F(ab')2 fragments caused almost complete inhibition (80%-95%) of AAb binding in 24 out of 25 sera, and in the case of the remaining serum, the effect was marked but incomplete (67% inhibition). There were no major differences between the binding characteristics of AAbs from patients with different forms of autoimmune adrenal disease. All five 21-OH MAbs reacted with human adrenal tissue in an immunofluorescence test, but only M21-OH1 and M21-OH2 reacted with bovine adrenal tissue in these experiments. None of the MAbs reacted with human ovarian tissue in an immunofluorescence test. Overall, these studies indicate that 21-OH AAbs bind to at least three different epitopes in the C-terminal part of 21-OH, and two of these epitopes appear to be human 21-OH specific.

Autoantibody binding to steroid 21-hydroxylase--effect of five mutations

Steroid 21-hydroxylase (21-OH) is a key haem containing steroidogenic enzyme and a major adrenal specific autoantigen. Cys 428 in 21-OH is thought to have an important role in haem binding and we now describe the effects of mutations at Cys 428 (to Ser, Arg and Phe) on 21-OH autoantibody binding. Expression of wild type and mutated 21-OH was carried out using an in vitro transcription/translation (TnT) system and reactivity of 21-OH autoantibodies with mutated 21-OH analysed by western blotting (in the case of unlabelled proteins) or immunoprecipitation assay (IPA) (in the case of 35S-labelled proteins). All 3 substitutions at Cys 428 had similar effects on 21-OH autoantibody binding and each one caused a reduction in autoantibody binding to about 50% of wild type in the case of IPA and to about 70% of wild type in the case of western blotting analysis. In addition to mutations at Cys 428, we studied 2 naturally occurring mutations at Pro 30 to Leu and Ile 172 to Asn which are associated with diminished 21-OH enzyme activity. The Pro 30 mutation had no effect, but the Ile 172 mutation caused a reduction in 21-OH autoantibody binding in the IPA to about 80% of wild type. Overall, our studies emphasise the close relationship between the 21-OH aminoacid sequences important for 21-OH enzyme activity and 21-OH autoantibody binding.

Binding characteristics of antibodies to the TSH receptor

We have used fragments of the TSH receptor (TSHR) expressed in E. coli as glutathione S-transferase fusion proteins to produce rabbit polyclonal antibodies and a panel (n=5) of monoclonal antibodies to the extracellular fragment of the TSHR. The binding characteristics of the antibodies to linear, conformational, glycosylated and unglycosylated forms of the receptor in different assay systems have been investigated. The reactivity of these antibodies with the TSHR was assessed by Western blotting with both native and recombinant human TSHR expressed in CHO cells, immunoprecipitation of 35S-labelled full-length TSHR produced in an in vitro transcription/ translation system, immunoprecipitation of 125I-TSH/TSHR complexes, inhibition of 125I-TSH binding to the TSHR and fluorescence activated cell sorter (FACS) analysis of binding to CHO-K1 cells expressing the TSHR on their cell surface. Fab fragments of monoclonal antibodies were isolated, labelled with 125I and used to determine the affinity constants of these antibodies with receptor, bound and free Fab being separated by polyethylene glycol (PEG) precipitation. Rabbit polyclonal and mouse monoclonal antibodies reacted with the TSHR in Western blotting and one monoclonal antibody (3C7) was able to inhibit 125I-TSH binding to native human TSHR (74% inhibition), recombinant human TSHR (84% inhibition) and porcine TSHR (65% inhibition). Affinity constant values for TSHR monoclonal antibody Fab fragments calculated using Scatchard analysis were about 10(7) M(-1). Four out of five monoclonal antibodies reacted in FACS analysis with TSHR expressed on the surface of CHO-K1 cells. The FACS unreactive monoclonal (3C7) bound well to detergent solubilised TSH receptors and this emphasised the importance of using a combination of FACS analysis and radioactively-labelled probes in analysis of the TSH receptor. The monoclonal antibodies produced in this study were found to be of relatively low affinity but proved useful for detection of the receptor by Western blotting and by FACS analysis.

A study of the epitopes on steroid 21-hydroxylase recognized by autoantibodies in patients with or without Addison's disease

Steroid 21-hydroxylase (21-OH) autoantibodies are found in patients with autoimmune Addison's disease (AAD), either isolated or associated with autoimmune polyglandular syndrome (APS) type I and II and in adrenal-cortex autoantibody (ACA)-positive patients without AAD. In order to assess any differences in the 21-OH autoantibodies in these different patient groups, we have studied their reactivity with different epitopes on 21-OH using full length and modified 35S-labelled 21-OH proteins produced in an in vitro transcription/translation system. There were no major differences in the pattern of autoantibody reactivity with the different modified 21-OH proteins in patients with isolated AAD or with APS types I and II, and in 21-OH autoantibody-positive patients with clinical AAD, subclinical AAD and those maintaining a normal adrenal function. Our studies also indicate that the main epitopes for 21-OH autoantibodies in patients with different forms of autoimmune adrenal disease are located in the C-terminal end and in a central region of 21-OH.

Understanding the thyrotropin receptor function-structure relationship

The thyrotropin (TSH) receptor (TSHR) is a key protein in the control of thyroid function and a major thyroid autoantigen. Recently, molecular cloning of the receptor has been carried out and we now review the impact of this work on our understanding of the physiology and pathophysiology of the TSHR. Analysis of recombinant TSHR proteins expressed in prokaryotic and eukaryotic systems has indicated that post-translational processing is important for the formation of active receptors. Studies of TSHR glycosylation have shown that a 'mature' form of the receptor containing mainly complex-type sugar residues is principally involved in TSH and TSHR autoantibody (TRAb) binding. In addition, the processing of the TSHR peptide chain into two subunits observed with native TSHR has been confirmed using recombinant TSHR. However, despite considerable efforts in many laboratories, the binding site(s) for TSH and TRAb on the TSHR have not been well characterized as yet and lessons learned from the discovery of naturally occurring amino acid mutations of the TSHR confirm the complexity of the hormone and autoantibody binding sites. Future progress in producing large amounts of pure TSHR as well as monoclonal TRAbs, followed by crystallographic analysis of TSHR-TSH complexes and TSHR-TRAb complexes, should be helpful in providing a better insight into the relationship between TSHR structure and function.

High prevalence of GAD65 (and IA-2) antibodies in Japanese IDDM patients by a new immunoprecipitation assay based on recombinant human GAD65

Marked differences have been reported in the prevalence of glutamic acid decarboxylase (GAD) antibodies between Caucasian (63-84%) and Japanese (30-50%) or Asian (5-50%) IDDM patients. Using a new immunoprecipitation assay based on 125I-labelled recombinant human GAD65 we have reassessed prevalence of GAD65 antibodies in Japanese patients. We also assessed prevalence of IA-2 antibodies. GAD65 antibodies were detected in 83.3% of sera taken within 1 year of onset, comparable to the prevalence reported in Caucasian patients. Positivity decreased to 66.7% after 2 to 3 years and to 54.3% after 3 years from onset, still higher than previously reported Asian prevalence. Except in one patient, high antibody levels persisted chronically, up to 12 years. There was no difference in the prevalence of GAD65 antibodies between Japanese IDDM patients with and without autoimmune thyroid disease (AITD). IA-2 antibodies were detected in 64.7% of sera taken within 1 year of onset. Prevalence of IA-2 antibodies was lower than that of GAD65 antibodies. The difference in positivity in Asian IDDM patients between present and previous reports arose from the sensitivity of our assay for GAD65 antibodies. Additionally, the patients we studied had classic IDDM with a well-defined onset. We conclude that prevalence of GAD65 antibodies in Japanese IDDM patients is comparable to that in Western studies. There was no relationship of GAD65 antibody positivity to coexistence of AITD. Our results suggest that autoimmunity is the most significant cause of Japanese IDDM.

Steroid 21-hydroxylase autoantibodies: measurements with a new immunoprecipitation assay

Autoantibodies (Abs) to steroid 21-hydroxylase (21-OH) are a major component of adrenal cortex Abs and are characteristic of autoimmune Addison's disease. We have developed a new method for measuring Abs to 21-OH based on 125I-labeled recombinant human 21-OH produced in yeast. With this assay, 21-OH Abs were detected in 43 of 60 (72%) sera from patients with isolated Addison's disease, 11 of 12 (92%) autoimmune polyglandular syndrome type I sera, 27 of 27 (100%) autoimmune polyglandular syndrome type II sera, and 24 of 30 (80%) sera from patients who were positive for adrenal cortex antibodies by immunofluorescence but had no overt Addison's disease. 21-OH Abs were found by 125I assay in 4 of 150 (2.7%) sera from patients with insulin-dependent diabetes mellitus, 1 of 77 (1.3%) Graves' sera, 1 of 67 (1.5%) Hashimoto's sera, and 6 of 243 (2.5%) sera from healthy blood donors. 21-OH Abs were not detected in 9 sera from patients with Addison's disease due to tuberculosis, 32 sera from patients with noninsulin-dependent diabetes mellitus, 35 sera from patients with myasthenia gravis, or 17 sera from patients with premature ovarian failure. There was good agreement between the 125I-labeled 21-OH assay and an assay based on 35S-labeled 21-OH produced in an in vitro transcription/translation system (r = 0.86; n = 129; P < 0.001). In the case of sera from patients with Addison's disease, insulin-dependent diabetes mellitus, Graves' disease, and Hashimoto's disease and from healthy blood donors that were low positive in the 125I assay, neutralization studies with unlabeled 21-OH confirmed the presence of specific 21-OH Abs. Overall, the 21-OH Ab assay based on 125I-labeled 21-OH showed good sensitivity, precision, and disease group specificity. This, combined with a simple assay protocol and the convenience of 125I handling and counting, make it attractive for routine use. Further investigations with the new assay should allow wider assessment of the prevalence and pattern of inheritance of adrenal autoimmunity. In addition, studies of the effect of treatment or possible preventative measures on 21-OH Ab levels in individuals without overt adrenal failure may suggest ways of delaying the onset of autoimmune Addison's disease.

Thyrotropin (TSH) receptor autoantibodies do not appear to bind to the TSH receptor produced in an in vitro transcription/translation system

An in vitro transcription/translation (TnT) system was used to produce 35S-labeled full-length TSH receptor (TSHR) and TSHR extracellular domain (TSHRex). The interaction of the labeled proteins with TSHR autoantibodies in Graves' sera was then studied using an immunoprecipitation assay. In the assay, 35S-labeled TSHR or TSHRex were incubated with test sera, and any immune complexes formed were precipitated with protein A-Sepharose (in the case of mouse monoclonal antibodies, antimouse IgG-agarose was used). Rabbit antibodies to the TSHR and a mouse monoclonal antibody precipitated as much as 50% of the 35S-labeled TSHR preparations compared with about 2% for normal rabbit serum and 4% for a control monoclonal antibody. However, none of 34 Graves' sera (TSHR autoantibody levels ranging from 14-95% inhibition of [125I]TSH binding) were able specifically to immunoprecipitate 35S-labeled TSHR or TSHRex. These negative findings were confirmed by analysis of the immunoprecipitates on SDS-PAGE followed by autoradiography. Our results indicate that the TnT system is not useful for producing labeled TSHR preparations that can bind TSHR autoantibodies well. This is in contrast to TnT produced 35S-labeled glutamic acid decarboxylase, thyroid peroxidase, and 21-hydroxylase, which react well with their respective autoantibodies. One main difference between these 3 autoantigens and the TSHR is that the receptor is highly glycosylated, and this extensive glycosylation may be of critical importance for correct folding of the receptor. Consequently, the inability of the TnT system to glycosylate proteins could explain in part why TnT-produced 35S-labeled TSHR and TSHRex do not bind TSHR autoantibodies.

I. Adrenal cortex and steroid 21-hydroxylase autoantibodies in adult patients with organ-specific autoimmune diseases: markers of low progression to clinical Addison's disease

Adrenal cortex antibodies (ACA) were measured by immunofluorescence in 8840 adult patients with organ-specific autoimmune diseases without overt hypoadrenalism. Sixty-seven (0.8%) patients were ACA-positive, with the highest prevalence in those with premature ovarian failure (8.9%). Forty-eight ACA-positive and 20 ACA-negative individuals were enrolled into a prospective study. Antibodies to steroid 21-hydroxylases (21-OH), steroid 17 alpha-hydroxylase (17 alpha-OH) and cytochrome P450 side chain cleavage enzyme (P450scc) were measured by immunoprecipitation assay. Human leucocyte antigens D-related (HLA-DR) genotyping was also carried out and adrenal function assessed by ACTH test. On enrollment, 75% of ACA-positive patients had a normal adrenal function, while 25% revealed a subclinical hypoadrenalism. 21-OH antibodies were positive in 91% of ACA-positive sera. Eleven patients were positive for steroid-cell antibodies by immunofluorescence, and 9 revealed a positivity for antibodies to 17 alpha-OH and/or P450scc. During the prospective study, overt Addison's disease developed in 21% and subclinical hypoadrenalism in 29% of ACA-positive patients, while 50% maintained normal adrenal function. Progression to Addison's disease was more frequent in patients with subclinical hypoadrenalism, high titers of ACA and higher levels of 21-OH antibodies, complement-fixing ACA and HLA-DR3 status. All 20 persistently ACA-negative patients were also negative for antibodies to 21-OH, 17 alpha-OH, and P450scc, and all maintained normal adrenal function during follow-up. In conclusion, the detection of ACA/21-OH antibodies in adults is a marker of low progression toward clinical Addison's disease.

II. Adrenal cortex and steroid 21-hydroxylase autoantibodies in children with organ-specific autoimmune diseases: markers of high progression to clinical Addison's disease

Adrenal cortex autoantibodies (ACA) were measured by immunofluorescence in 808 children with organ-specific autoimmune diseases without adrenal insufficiency. ACA were found in 14 children (1.7%), mostly in hypoparathyroidism (48%). Ten ACA-positive and 12 ACA-negative children were followed up for a maximum of 10 yr by evaluation of adrenocortical function (ACTH test) and autoantibody status. In all patients steroid-producing cell autoantibodies were assessed by immunofluorescence and autoantibodies to steroid 21-hydroxylase, 17 alpha-hydroxylase, and cytochrome P450 side-chain cleavage enzyme by immunoprecipitation assay. All 10 ACA-positive patients were positive for 21-hydroxylase autoantibodies. Six were positive for steroid-producing cell autoantibodies and 5 also for autoantibodies to 17 alpha-hydroxylase and/or P450 side-chain cleavage enzyme. Overt Addison's disease developed in 9 (90%) ACA/21-OH-antibody-positive children after 3-121 months, and 1 remaining child had subclinical hypoadrenalism. By contrast, all ACA/21-OH antibody-negative children maintained normal adrenal function. Adrenal failure was not related to ACA titres, sex, adrenal function, type of preexisting autoimmune disorder, or human leucocyte antigens D-related status. In conclusion, in children with autoimmune endocrine diseases, ACA/21-hydroxylase autoantibodies are important predictive markers for the development of Addison's disease.

Glutamic acid decarboxylase autoantibody assay using 125I-labelled recombinant GAD65 produced in yeast

We describe a new method for measuring autoantibodies (Ab) to the 65 kDa isoform of glutamic acid carboxylase (GAD65). In particular, GAD65 without the hydrophobic N-terminal region has been produced in yeast, purified, labelled with 125I and reacted with GAD65 Ab. Antibody bound 125I-GAD65 is then precipitated by the addition of solid phase protein A. With the assay, GAD65 Ab were detected in 59 of 71 (83%) islet cell antibody (ICA) positive IDDM patients and in 8 of 23 (35%) ICA negative IDDM patients (overall 67 of 94 (71%) of IDDM patients). Low concentrations of GAD65 Ab were also detected in 2/98 (2%) healthy blood donors and 1/27 (4%) Graves' disease patients had a high level of antibody. GAD65 Ab were not detected in any of 10 Hashimoto's thyroiditis, 20 Addison's disease or 19 myasthenia gravis sera. There was good agreement between the 125I assay and the current reference method based on 35S-labelled full-length GAD65 (produced by in vitro transcription/translation reaction) and solid phase protein A (r = 0.91, n = 108). Overall, our 125I assay showed sensitivity, precision and disease group specificity at least as good as any assay so far described. These features, combined with a simple assay protocol and the convenience of 125I counting and handling indicate that the method is suitable for routine GAD65 Ab measurements.

High-level expression of recombinant immunoreactive thyroid peroxidase in the High Five insect cell line

Expression of a major thyroid autoantigen, thyroid peroxidase (TPO) was studied using the baculovirus-insect cell expression system. Human TPO cDNA modified so as to code for the extracellular fragment of the protein was placed under the control of the strong polyhedrin promoter in baculovirus transfer vector pBlueBacIII and cotransfected with linearized AcMNPV viral DNA. Expression in two insect cell lines Spodoptera frugiperda (Sf9) and Tricoplusia ni (High Five) was investigated and levels of recombinant TPO (rTPO) monitored by RIA and SDS-PAGE followed by Western blotting. Both insect cell lines expressed rTPO, but higher levels (30 mg/l culture medium) were obtained with High Five cells. Culture medium rTPO was purified and its glycosylation and immunoreactivity analysed. Lectin-affinity blotting and treatment with glycosidases indicated that both high mannose and complex-type sugar residues were associated with the recombinant protein. Studies with an ELISA based on biotinlabelled rTPO and an immunoprecipitation assay based on 125I-labelled rTPO indicated that the rTPO and native TPO showed similar reactivity to TPO autoantibodies (r = 0.96, P < 0.001, n = 50 and r = 0.99, P < 0.001, n = 80 respectively). In addition, rTPO expressed in High Five cells showed enzyme activity comparable with that of native TPO when the heme biosynthesis precursor delta-aminolevulinic acid was included in the culture medium. Overall, our studies indicate that the High Five insect cell line provides a useful system for the expression of relatively high levels of rTPO which should be suitable for structural analysis of TPO and TPO-TPO autoantibody complexes.