Protein tyrosine phosphatase-like IA-2 autoantigen is one of the major targets of humoral autoimmunity in patients with insulin-dependant diabetes mellitus (IDDM). In an effort to define the epitopes recognized by autoantibodies against IA-2, we generated five human mAbs (hAbs) from peripheral B lymphocytes isolated from patients most of whom had been recently diagnosed for IDDM. Determination and fine mapping of the critical regions for autoantibody binding was performed by RIA using mutant and chimeric constructs of IA-2- and IA-2β-regions. Four of the five IgG autoantibodies recognized distinct epitopes within the protein tyrosine phosphatase (PTP)-like domain of IA-2. The minimal region required for binding by three of the PTP-like domain-specific hAbs could be located to aa 777–979. Two of these hAbs cross-reacted with the related IA-2β PTP-like domain (IA-2β aa 741-1033). A further PTP-like domain specific hAb required the entire PTP-like domain (aa 687–979) for binding, but critical amino acids clustered in the N-terminal region 687–777. An additional epitope could be localized within the juxtamembrane domain (aa 603–779). In competition experiments, the epitope recognized by one of the hAbs was shown to be targeted by 10 of 14 anti-IA-2-positive sera. Nucleotide sequence analysis of this hAb revealed that it used a VH germline gene (DP-71) preferably expressed in autoantibodies associated with IDDM. The presence of somatic mutations in both heavy and light chain genes and the high affinity or this Ab suggest that the immune response to IA-2 is Ag driven.

Autoimmunity directed against pancreatic islet cells causes progressive β-cell destruction and, as a consequence, insulin-dependent diabetes mellitus (IDDM).3 Although IDDM appears to be mostly a T cell-mediated autoimmune disease, the autoimmune process is accompanied by a humoral response against islet-specific Ags (1). Long before the onset of IDDM, a variety of islet-specific autoantibodies can be detected in the serum of affected individuals. Abs directed against islet cell Ags, such as glutamic acid decarboxylase (GAD65) (2), insulin (3), and the protein tyrosine phosphatase (PTP)-like proteins IA-2 (islet cell Ag 512) (4, 5) and IA-2β (phogrin/islet cell Ag-related PTP) (6, 7) are frequently present in the serum of prediabetic individuals and therefore represent valuable markers to predict IDDM in first degree relatives of IDDM patients. Anti-IA-2 autoantibodies are highly IDDM specific, and in combination with additional immunological markers such as anti-GAD65 and anti-insulin autoantibodies they are to date considered reliable markers for the diagnosis and prediction of IDDM (8).

IA-2 as a major target of humoral autoimmunity in IDDM was initially identified by screening of an islet cDNA expression library with patients’ sera (4). The full length coding sequence of the initial fragment, designated ICA512, was subsequently characterized and named IA-2 (9). IA-2 is closely related to phogrin/IA-2β, both being transmembrane proteins within the secretory granule membrane of neuroendocrine cells (7, 12). Their intracellular regions feature protein tyrosine phosphatase (PTP)-like domains of ∼300 aa with no (IA-2) and weak (IA-2β) phosphatase activity detected (5, 13). IA-2 and IA-2β share 88% amino acid sequence homology within this region. The juxtamembrane domain of less than 100 aa links the PTP-like domain with the transmembrane domain and shows 50% homology between IA-2 and IA-2β. The luminal ectodomains of both proteins consist of ∼600 aa and share less than 10% homology. Ab binding occurs within the intracellular cytoplasmic domain (11). IA-2 and IA-2β probably share common epitopes but also show distinct epitopes; 50–80% of IDDM sera that react with IA-2 also recognize IA-2β. Alternatively, even 95% of IDDM sera reacting with IA-2β also recognize IA-2 (10).

The role of autoantibodies in the IDDM autoimmune process remains elusive. The availability of disease-related human mAbs such as IA-2/IA-2β-specific Abs for immunological studies should help to elucidate the pathophysiological relevance of humoral components in the IDDM autoimmune process. Moreover, hAbs against IA-2 or IA2-β may help to overcome the difficulties of standardization of conventional IA-2 Ab assay systems and will allow comparison of different studies more reliably. However, all efforts to produce hAbs against IA-2 have failed thus far. In this study, we describe for the first time the development and characterization of IA-2-specific hAbs of IgG isotype derived from patients with newly diagnosed IDDM. With the use of mutant forms of IA-2 and chimeric IA-2/IA-2β proteins, the epitopes targeted by these hAbs have been mapped in detail and Ig V-gene regions have been characterized.

Blood samples were obtained from the Institute of Diabetic Research, Munich, Germany. PBMCs were isolated from heparinized peripheral blood of IDDM patients exhibiting high IA-2-specific Ab levels by Ficoll (Amersham Pharmacia Biotech, Uppsala, Sweden) density gradient centrifugation (for details, see Table I). To separate B lymphocytes by magnetic cell sorting, PBMCs were first labeled with mouse anti-human IgG Ab (Dianova, Hamburg, Germany) and subsequently with magnetic microbeads binding mouse IgG (Dynabeads M-280, Dianova). For immortalization, isolated IgG-positive B lymphocytes were incubated under regular gentle shaking for 2 h at 37°C with EBV-containing supernatant from the B-95-8 marmoset lymphoma cell line (American Type Culture Collection, Manassas, VA). B lymphocytes were seeded on microtiter plates at densities of 40–200 cells/well plus 20,000–50,000 irradiated PBMC feeder cells/well. Immortalized B lymphocytes were cultured in IMDM supplemented with 10% FCS, 2 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, 400 μM insulin, 1 mM pyruvate, nonessential amino acids, 1 μM oxaloacetate, and 100 U/ml IL-6 at 37°C and 7% CO2. Cells of this primary culture were fed once a wk and cultured over 2–3 wk before screening for IA-2-specific Abs.

Table I.

Immortalization of B lymphocytes

PatientAge (yr)Duration of IDDMSerum Anti-IA-2 Concentration (U)IA-2-Positive Primary WellsMonoclonal B Cell Lines
30 New onset 86  
36 New onset 58   
21 New onset 80   
32 New onset 87   
33 New onset 86   
New onset 184 76/12 
19 New onset 81  
33 2 yr 78   
20 2 yr 110 96/3, 96/4, 96/5 
10 49 New onset 101 103/5 
11 26 New onset 101   
PatientAge (yr)Duration of IDDMSerum Anti-IA-2 Concentration (U)IA-2-Positive Primary WellsMonoclonal B Cell Lines
30 New onset 86  
36 New onset 58   
21 New onset 80   
32 New onset 87   
33 New onset 86   
New onset 184 76/12 
19 New onset 81  
33 2 yr 78   
20 2 yr 110 96/3, 96/4, 96/5 
10 49 New onset 101 103/5 
11 26 New onset 101   

Supernatants of the EBV-transformed B lymphocyte cell lines were screened by a recently established anti-IA-2 ELISA (Roche Diagnostics, Penzberg, Germany). In brief, 50 μl supernatant of each well were diluted 1:2 with culture medium, transferred into streptavidin-coated microtiter plates (Microcoat, Bernried, Germany), which had been coated with biotin-conjugated recombinant IA-2ic at a concentration of 150 ng/ml (≡2, 6 nM) and incubated for 1 h at room temperature while shaking (14). After extensive washing (three times) with PBS-0.05% Tween 20, 100 μl of mouse anti-human IgG peroxidase conjugate (0.2 U/ml, Roche) were added to the wells. After incubation (1 h at room temperature), plates were washed with PBS-0.05% Tween 20 (three times), and bound Abs were detected by reaction with the peroxidase substrate 2,2′-azinobis(3-ethylbenzthiazoline-6-sulfonic acid) (ABTS, Roche). Extinction was measured after 1 h at 405 nm with a reference wavelength of 492 nm. As a standard, patients’ sera proved to be positive for anti-IA-2 Ab in conventional RIA were used. Anti-IA-2 Ab-producing cell lines were expanded to 0.5 × 105 cells and cloned repeatedly on a single-cell level in microtiter wells by limiting dilution. Culturing conditions were the same as used for primary culture.

Subclasses of human IgG Abs were determined as follows. Anti IA-2 Abs bound to streptavidin-biotin-IA-2ic-coated microtiter plates (see above) were coupled to mouse anti-hAbs (10 μg/ml) specific for 1) the different human IgG subclasses and 2) the two human light chains. The bound mouse Ab was detected by HRP-conjugated anti-mouse IgG (62.5 mU/ml) (Roche) and ABTS reaction as described above.

Human IgG was determined by a sandwich ELISA, using goat anti-human IgG-Fc-specific Ab (Dianova) for coating of microtiter plates at a concentration of 10 μg/ml coating buffer (Roche), and bound human IgG was detected by 50 ng/ml peroxidase-coupled anti-human IgG-F(ab′)2-specific goat Ab (Dianova) and ABTS.

Ab affinity was determined by real-time interaction analysis with a BIAcore system (Pharmacia, Uppsala, Sweden) according to the manufacturer’s instructions. Biotinylated IA-2ic was immobilized to streptavidin-coated sensor chips.

To preform immune complexes with low background staining, IgG mAbs were complexed with HRP-conjugated goat anti-human IgG Abs as described (15). Cryostat sections of human and animal pancreas and other human tissues were incubated with these preformed immune complexes for 2 h at 4°C in a humidified chamber. After intense washing of each section with cold PBS, the bound Ab was detected by staining with aminoethylcarbazole and hydrogen peroxide.

Double immunofluorescence staining was performed as follows. The isolated IA-2-specific hAbs were labeled with digoxigenin (DIG, Roche) according to the manufacturer’s instructions using a molar reaction mixture, Ab:DIG (1:15). Cryostat sections of human pancreas were incubated for 45 min with these DIG-labeled Abs (5 μg/ml) and the α cell-specific Ab BISL-32 (1:2000 diluted, Roche) (16). After three washings with PBS, the bound Abs were detected either by a FITC-conjugated mouse anti-DIG-Ab (diluted 1:600, Roche) and a Cy-3-conjugated goat anti-mouse IgG (diluted 1:200). The double-stained sections were examined under a fluorescence microscope.

IA-2/IA-2β constructs used in this study are shown in Fig. 2. IA-2687–979 and IA-2β741–1033 were used to differentiate between the PTP-like domain of IA-2 and IA-2β. The IA-2389–779 served to identify Abs binding to the juxtamembrane region of IA-2. These constructs were prepared using the pGEM-T cloning vector (Promega, Madison, WI) under control of the SP6 promoter as described previously (11, 17). To define critical amino acids within epitope regions of the PTP domain, seven constructs with point mutations changing IA-2-specific amino acids to IA-β-specific amino acids were prepared using the pSP64 cloning vector (Promega) in combination with the QuickChange (Stratagene, La Jolla, CA) method (17). Additionally, two chimeric IA-2/IA-2β PTP constructs were prepared by in-frame joining of selected portions of IA-2 and IA-2β and expressed using pGEM-T vector as described recently (17). Purified plasmid DNA was transcribed, translated, and labeled using the TnT SP6-coupled rabbit reticulocyte lysate system (Promega) in the presence of [35S]methionine (Amersham, Aylesbury, U.K.). Unincorporated radioactivity was removed by gel chromatography on NAP5 columns (Pharmacia).

FIGURE 2.

IA-2 and IA-2β cDNA constructs used in this study. Numbers indicate the amino acid positions within the predicted protein sequences of IA-2 (GenBank accession number L18983) and IA-2β (GenBank accession number Y08569). JM, Juxtramembrane region; TM, transmembrane region.

FIGURE 2.

IA-2 and IA-2β cDNA constructs used in this study. Numbers indicate the amino acid positions within the predicted protein sequences of IA-2 (GenBank accession number L18983) and IA-2β (GenBank accession number Y08569). JM, Juxtramembrane region; TM, transmembrane region.

Close modal

The assay was essentially performed as described previously (17). Briefly, the in vitro translated protein fragments with 15,000–20,000 cpm incorporated radioactivity were diluted in 25 μl TBST (50 mM Tris, 150 mM NaCl (pH 7.2), 1% Tween 20), mixed with 2–10 μl hAb-containing cell supernatants, and incubated overnight on ice. Subsequently, immune complexes were recovered by adding 1 mg protein A-Sepharose (Pharmacia), preswollen and resuspended in 50 μl TBST, and incubated for 1 h at 4°C under shaking. After four washings with 2 ml TBST, the protein A-Sepharose-coupled immune complexes were resuspended in 100 μl TBST, transferred to a 96-well Optiplate (Packard, Groningen, The Netherlands), mixed with 150 μl Microscint 40, and cooled for 30 min under shaking. The precipitated radioactivity was measured in a TopCount scintillation counter (Packard) for 5 min. Reactivity toward the mutagenized IA-2 proteins was evaluated by using a mouse anti-IA-2 mAb (mAb 76F) as internal standard, which binds to an epitope within the juxtamembrane region of IA-2 (17).

Different human anti-IA-2-reactive sera from recently diagnosed IDDM patients (kindly provided by P. Pozzilli, University Campus Biomedico, Rome, Italy) were used for blocking studies in an IA-2-specific ELISA performed as described above with some modifications. After coupling of biotinylated IA-2ic (150 ng/ml) to streptavidin-coated microtiter plates, 100 μl of IA-2-reactive sera (diluted to 125 ng/ml IA-2-specific Ab equivalents) were transferred to the wells and incubated for 1 h at room temperature to block epitope-specific binding sites. After intense washing with PBS-0.05% Tween 20, the remaining free binding sites were detected by monoclonal DIG-labeled anti-IA-2 Ab (2.5 ng/ml). Bound DIG-labeled hAb was detected by HRP-conjugated mouse anti-DIG Ab (1:1000 diluted, Roche) and ABTS reaction. ODs were measured at 405 nm with a reference wavelength of 492 nm.

RNA isolation, cDNA synthesis, and RT-PCR amplification (primary PCR).

Total RNA from ∼106 lymphoblastoid B cells producing IA-2-specific hAb was isolated using the Trisolv method (Biotecx, Houston, TX) according to the manufacturer’s instructions and subjected to cDNA synthesis using an AMV Reverse Transcriptase Kit (Roche). Primary PCR was performed as described by McCafferty et al. (18) using six BACK VH and four FOR JH gene family-specific primers for the variable region of heavy IgG chains, seven BACK Vλ- and three FOR Jλ-specific primers for the variable region of λ light chains, and six BACK Vκ- and five FOR Jκ- specific primers for the variable region of κ light chains. For amplification of heavy and light chain cDNA fragments, each BACK primer was used in a separate reaction mixture. A 50 μl reaction mixture was prepared containing 1× PCR buffer with 20 mM Mg2+, 5 mM PCR nucleotide mix, 5 U Pwo polymerase (Roche), 0.5 μM primer, and 0.1–0.75 μg template DNA. The reaction mixture was subjected to 30 cycles of amplification, 1 cycle consisting of 1 min denaturation at 94°C, 1 min touchdown-annealing at 70–40°C, and 1 min extension at 72°C. Amplification products were analyzed on a 1.4% agarose gel and purified by gel extraction (Qiagen, Hilden, Germany).

Linker and assembly PCR to generate single-chain variable fragments (scFv).

A commercially available mouse DNA fragment (Pharmacia) linking the 3′-end of the heavy chain variable domain to the 5′-end of the light chain variable region had to be “humanized” by PCR using a mixture of six BACK primers and seven FOR primers (18). PCR amplification was performed in 25 cycles, 1 cycle consisting of denaturation at 94°C for 1 min, annealing at 65°C for 1 min, and extension at 72°C for 2 min. The humanized linker fragments were subsequently used for assembly of scFv consisting of VH and VL fragments connected by the linker. After introducing SfiI and NotI restriction sites, the scFv fragments were ligated into pCANTAB-5E-phagemid (Pharmacia) for sequence determination and expression of soluble scFv-Ab fragments (for details, see manufacturer’s instruction manual).

Amplified cDNA fragments and recombinant phagemid DNA were sequenced by the dideoxy chain termination procedure (19). The established VH/L sequences, DH gene diversity sequences, and JH/L sequences were compared with sequences present in the V BASE sequence directory (www.mrc-spe.ca.ac.uk/imt-doc/public/INTRO.html). Subdivision of sequences into framework (FR) and complementarity-determining regions (CDRs) were performed according to the method of Kabat (20).

B lymphocytes from 11 IDDM patients exhibiting high anti-IA-2 Ab levels in their sera were subjected to EBV immortalization. By selecting IgG-producing B lymphocytes, we eliminated the IgM-producing B lymphocytes, which predominate in the pool of PBLs and usually produce Abs of low affinity and specificity (21). Screening of ∼40,000 B lymphocyte culture supernatants by an anti-IA-2-specific ELISA identified 16 anti-IA-2 Ab-producing B lymphocyte cell lines (Table I). Five of them could be stabilized at single-cell level by repeated limiting dilution. Due to the instability and low cloning efficiency of EBV-transformed B cell lines (22) 11 of the primarily anti-IA-2-positive cultures did not survive this procedure or stopped Ab production during the cloning steps. Addition of IL-6 (100 U/ml) as a B cell growth factor (23) ameliorated the survival rate but could not rescue most of the unstable cell lines.

Five cell lines producing hAb anti-IA-2 were derived from three different patients who showed very high anti-IA-2 Ab sera levels (Table I): cell line 76/12 (patient 6); cell lines 96/3, 96/4, and 96/5 (patient 9); and cell line 103/5 (patient 10). Nevertheless, only one cell line (96/3) showed long term stability for >12 mo and could be used as constant source of anti-IA-2 Ab with a production rate of 5–8 μg/106cells in 24 h. This hAb was purified and its affinity was determined by the BIAcore method. It revealed high affinity to recombinant IA-2ic which was coupled to the sensor chip (KD = 0.13 nM). Using the hAb96/3 as standard in the anti IA-2-ELISA proved that the assay was linear for Ab concentrations within 0.1–5 ng/ml.

Analysis of the IgG subtypes classified all five isolated hAbs as IgG1, which is in accordance with the observation that in IDDM patients pancreas-reactive autoantibodies are predominantly IgG1 (24, 25). Three of the light chains expressed were of κ subtype and two of the light chains were of λ subtype.

The immunohistochemistry on pancreata of different species revealed no species specificity for human, mouse, and rat islets. All tested hAb anti-IA-2 stained the pancreatic islets, whereas exocrine tissue showed no reactivity (Fig. 1,A). Background staining by endogenous Ig in tissues was avoided by using preformed (anti-IA-2/peroxidase-anti-human IgG) complexes. Double staining of pancreatic islets with DIG-labeled hAb96/3 and the α cell-specific mouse mAb BISL-32 showed that this IA-2-specific hAb stained β cells and a majority of the α cells (Fig. 1 B). Among other human tissues tested, IA-2-specific hAbs revealed no reactivity toward thyroid, liver, lung, stomach, renal, or intestine tissue. However, a faint reactivity could be seen with neurons of the cerebellar cortex.

FIGURE 1.

A, Immunohistology of human pancreas stained with anti-IA-2 hAbs. Cryostat sections of human pancreas were incubated with preformed immune complexes composed of peroxidase-conjugated anti-human IgG and hAb96/3. Detection was performed with aminoethylcarbazole and hydrogen peroxide. B, Double staining of human pancreas. α cells were stained with the α cell-specific BISL-32 mouse mAb and Cy-3-conjugated anti-mouse IgG. DIG-labeled IA-2-specific hAb96/3 was detected by FITC-conjugated anti-DIG Ab.

FIGURE 1.

A, Immunohistology of human pancreas stained with anti-IA-2 hAbs. Cryostat sections of human pancreas were incubated with preformed immune complexes composed of peroxidase-conjugated anti-human IgG and hAb96/3. Detection was performed with aminoethylcarbazole and hydrogen peroxide. B, Double staining of human pancreas. α cells were stained with the α cell-specific BISL-32 mouse mAb and Cy-3-conjugated anti-mouse IgG. DIG-labeled IA-2-specific hAb96/3 was detected by FITC-conjugated anti-DIG Ab.

Close modal

For epitope studies, we used 13 different constructs of IA-2/IA-2β (Fig. 2) in a conventional anti-IA-2 RIA. These constructs comprised various truncated forms of IA-2 (PTP domain only, PTP domain plus juxtramembrane region, and a construct including also the transmembrane region). In addition, different mutated forms of IA-2 as well as chimeras between IA-2 and IA-2β were tested.

Four of five anti-IA-2 autoantibodies, hAb76/12, hAb96/3, hAb96/4, and hAb96/5, reacted with IA-2687–979, the PTP-like domain of IA-2, and did not react with IA-2389–779 covering the juxtamembrane region of IA-2 (Table II). All four hAbs thus recognized epitopes within a region between aa 687 and aa 979. In addition, hAb96/3 revealed a strong reactivity toward construct IA-2β741–1033, the IA-2β-PTP-like domain. Neither hAb96/4 nor hAb96/5 reacted toward the IA-2β PTP-like domain, whereas hAb76/12 showed weak reactivity. A particular epitope could be specified for hAb103/5. It recognized IA-2389–779 but not the partially overlapping IA-2687–979 region and therefore must be directed against a determinant located N-terminal to the PTP-like domain. Because all mAbs were initially screened with the truncated construct IA-2ic603–979, we suppose that hAb103/5 is directed against a region within or very close to the juxtamembrane region of IA-2. In summary, we have isolated two hAbs recognizing one or two determinants that are shared within the PTP-like domain of IA-2 and IA-2β, two hAbs with restricted specificity toward the PTP-like domain of IA-2, and one further hAb reactive against the juxtamembrane region of IA-2 or a C-terminally closely neighboring region. As expected, none of the isolated Abs was exclusively IA-2β specific because the ELISA screening system was performed with IA-2ic.

Table II.

Reactivity of IA-2-specific hAbs toward different IA-2/IA-2β constructsa

hAb76/12hAb96/3hAb96/4hAb96/5hAb103/5
IA-2603–979 +++ +++ +++ +++ +++ 
      
IA-2687–979 +++ +++ +++ +++ − 
IA-2777–979 +++ +++ − +++ − 
IA-2389–779 (juxtamembrane) − − − − +++ 
      
IA-2-β741–1033 ++ − − − 
      
Chimeric constructs      
IA-2-β741–848/IA-2794–889/IA-2- β943–1033 +++ +++ − − − 
IA-2-β741–848/IA-2794–845/IA-2- β899–1033 +++ +++ − − − 
      
Mutagenized constructs      
Listed in Fig. 2  All+++ All++ All+++ except T804V All+++ except construct with positions P876Y; A877D; E878R; T880V exchanged simultaneously All− 
      
Minimal region identified for Ab binding IA-2777–979 IA-2777–979 IA-2687–979 IA-2777–979 IA-2603–779 
 or or    
 IA-2-β741–1033 IA-2-β741–1033    
 PTP domain PTP domain PTP domain PTP domain Juxtamembrane domain 
      
Critical amino acids identified IA-2794–845 Position 815 and/or 818, 829, 830, 834 None identified IA-2687–777 Position 804 IA-2777–793 and/or IA-2889–979 Positions 876, 877, 878, 880 None identifed 
hAb76/12hAb96/3hAb96/4hAb96/5hAb103/5
IA-2603–979 +++ +++ +++ +++ +++ 
      
IA-2687–979 +++ +++ +++ +++ − 
IA-2777–979 +++ +++ − +++ − 
IA-2389–779 (juxtamembrane) − − − − +++ 
      
IA-2-β741–1033 ++ − − − 
      
Chimeric constructs      
IA-2-β741–848/IA-2794–889/IA-2- β943–1033 +++ +++ − − − 
IA-2-β741–848/IA-2794–845/IA-2- β899–1033 +++ +++ − − − 
      
Mutagenized constructs      
Listed in Fig. 2  All+++ All++ All+++ except T804V All+++ except construct with positions P876Y; A877D; E878R; T880V exchanged simultaneously All− 
      
Minimal region identified for Ab binding IA-2777–979 IA-2777–979 IA-2687–979 IA-2777–979 IA-2603–779 
 or or    
 IA-2-β741–1033 IA-2-β741–1033    
 PTP domain PTP domain PTP domain PTP domain Juxtamembrane domain 
      
Critical amino acids identified IA-2794–845 Position 815 and/or 818, 829, 830, 834 None identified IA-2687–777 Position 804 IA-2777–793 and/or IA-2889–979 Positions 876, 877, 878, 880 None identifed 
 
  
   
\
a

Reactivity is expressed relative to a 1:10 dilution of a pool of human normal serum. Threshold reactivity, expressed as “+,” is at least 5-fold above the reactivity observed with the 1:10 diluted human serum, which measured in the range of 30–80 cpm. Reactivity against the mutagenized IA-2 proteins is expressed relative to that observed with a mouse anti-IA-2 mAb (76F) that binds to an epitope contained within the juxtamembrane region of IA-2.

 
 
  
   
\
a

Reactivity is expressed relative to a 1:10 dilution of a pool of human normal serum. Threshold reactivity, expressed as “+,” is at least 5-fold above the reactivity observed with the 1:10 diluted human serum, which measured in the range of 30–80 cpm. Reactivity against the mutagenized IA-2 proteins is expressed relative to that observed with a mouse anti-IA-2 mAb (76F) that binds to an epitope contained within the juxtamembrane region of IA-2.

 

To define epitope regions within the PTP-like domain in more detail, different mutant and chimeric constructs, as well as the construct IA-2777–979 were used (Fig. 2 and Table II). One of the IA-2 PTP domain-specific Abs, hAb96/4, did not bind to construct IA-2777–979 but showed reactivity with construct IA-2687–979. Obviously, the Ab recognized an epitope within aa 687–979 with residues 687–777 being critical for Ab binding. The Ab reacted with all the mutant constructs except construct T804V. Therefore, residue 804 is likely to be required for the tertiary structure of the epitope or is a critical Ab contact site. There was no reactivity of hAb96/4 toward any of the chimeric constructs containing IA-2 regions 794–845 and 794–889 framed by regions derived from the IA-2β isoform, indicating that IA-2-specific amino acids outside the region 794–889 were involved in Ab binding.

The second IA-2 PTP domain-specific Ab, hAb96/5, revealed reactivity to construct IA-2687–979 and also bound to construct IA-2777–979; therefore, its epitope was located within the region between residues 777 and 979. This Ab did not react with the chimeric constructs; therefore, hAb96/5 binding required residues 777–794 and/or residues 889–979. All mutagenized constructs reacted with hAb96/5 except the one in which positions 876, 877, 878, and 880 were simultaneously exchanged. Obviously, the region between aa 876 and 880 contributed to epitope formation while the exchange of IA-2-specific aa 804, 813, 821, 822, 862, and 886 did not affect Ab binding.

The hAbs cross-reactive with IA-2β (hAb96/3 and hAb76/12) revealed a similar reaction pattern toward all chimeric and mutagenized constructs (Table I). Both hAbs reacted with constructs IA-2687–979 and IA-2-β741–1033, with hAb76/12 exhibiting much weaker reactivity (320 cpm vs 3600 cpm with hAb96/3). Because both hAbs did not react with construct IA-2389–779, residues within region 779–979 were compulsory to form the epitopes. Because there was strong binding of both hAbs toward construct IA-2777–979, a conformational contribution of amino acids N-terminal to this region could be excluded. Determination of the exact binding region of hAb76/12 was possible due to its weak reaction toward IA-2β. Because hAb76/12 revealed high reactivity toward both chimeric proteins IA-2-β741–848/IA-2794–889/IA-2β943–1033 and IA-2β741–848/IA-2794–845/IA-2β899–1033 Ab binding required IA-2 region 794–845 with potential participation of IA-2/IA-2β-homologous amino acids outside this region. There was no significant loss of reactivity toward any of the mutant constructs; therefore, IA-2-specific aa 804, 813, 821, and 822 located within this region seemed not to be critical for Ab binding of hAb76/12. The epitope of hAb96/3 could not be restricted to a smaller region than aa 777–979 because hAb96/3 was highly cross-reactive and all chimeric and mutagenized constructs revealed strong reactivity with this Ab.

The in vivo relevance of the epitope recognized by hAb96/3 was tested by ELISA blocking studies using 14 different anti-IA-2-reactive sera of newly diagnosed IDDM patients. In a competition assay, 100% inhibition was achieved after blocking 2.5 ng/ml DIG-labeled IA-2-specific hAb96/3 with 125 ng/ml unlabeled hAb96/3. Therefore, IDDM sera for blocking studies were diluted to a concentration of 125 ng/ml IA-2-specific Ab equivalents as determined in the conventional anti-IA-2 ELISA. Ten of 14 tested sera competed with hAb96/3, indicating that they contained anti-IA-2-specific Ab(s) recognizing the same or an adjacent epitope as hAb96/3. Percent inhibition ranged between 20 and 85% (Table III).

Table III.

Binding competition of DIG-labeled hAb96/3 by sera of IDDM patientsa

Serum Used for BlockingmE Obtained with DIG-Labeled hAb96/3% Inhibition
Normal control 1.350 ± 0.01 
Patient 1 1.226 ± 0.003 No significant inhibition 
Patient 2 0.452 ± 0.027 67 
Patient 3 0.774 ± 0.009 43 
Patient 4 0.847 ± 0.004 37 
Patient 5 1.074 ± 0.015 20 
Patient 6 0.254 ± 0.003 81 
Patient 7 0.592 ± 0.01 56 
Patient 8 0.782 ± 0.006 42 
Patient 9 1.187 ± 0.008 No significant inhibition 
Patient 10 0.210 ± 0.009 85 
Patient 11 1.275 ± 0.007 No significant inhibition 
Patient 12 1.245 ± 0.009 No significant inhibition 
Patient 13 0.628 ± 0.007 54 
Patient 14 0.748 ± 0.006 44 
Serum Used for BlockingmE Obtained with DIG-Labeled hAb96/3% Inhibition
Normal control 1.350 ± 0.01 
Patient 1 1.226 ± 0.003 No significant inhibition 
Patient 2 0.452 ± 0.027 67 
Patient 3 0.774 ± 0.009 43 
Patient 4 0.847 ± 0.004 37 
Patient 5 1.074 ± 0.015 20 
Patient 6 0.254 ± 0.003 81 
Patient 7 0.592 ± 0.01 56 
Patient 8 0.782 ± 0.006 42 
Patient 9 1.187 ± 0.008 No significant inhibition 
Patient 10 0.210 ± 0.009 85 
Patient 11 1.275 ± 0.007 No significant inhibition 
Patient 12 1.245 ± 0.009 No significant inhibition 
Patient 13 0.628 ± 0.007 54 
Patient 14 0.748 ± 0.006 44 
a

Patients’ sera (diluted to a concentration of 125 ng/ml IA-2-specific Ab equivalents) were used to block binding of DIG-labeled hAb96/3 toward IA-2 in an ELISA.

We determined the nucleotide sequence of hAb96/3 (Fig. 3) by cloning of the VL and VH DNA fragments as scFv Abs. Soluble scFv Abs expressed in Escherichia coli bound specifically to IA-2ic, which was confirmed by ELISA (data not shown). Gene sequencing was performed on two independently generated PCR products using primers annealing within different gene segment regions to exclude that the observed mutations were not generated by the gene amplification and sequencing process.

FIGURE 3.

Alignment of the nucleotide sequences (A) and the amino acid sequences (B) of germline V-genes DP-71 and IGLV3S2 (top line) and hAb96/3 V genes (bottom line). Dashes represent identity between both sequences. Solid lines above each cluster encompass FR and CDR regions. PCR primer sites are underlined. The presented V sequences are deposited in the V BASE directory under GenBank accession numbers Z12371 (VH), X97051 (DH), X86355 (JH), X71966 (VL), and M15641 (JL).

FIGURE 3.

Alignment of the nucleotide sequences (A) and the amino acid sequences (B) of germline V-genes DP-71 and IGLV3S2 (top line) and hAb96/3 V genes (bottom line). Dashes represent identity between both sequences. Solid lines above each cluster encompass FR and CDR regions. PCR primer sites are underlined. The presented V sequences are deposited in the V BASE directory under GenBank accession numbers Z12371 (VH), X97051 (DH), X86355 (JH), X71966 (VL), and M15641 (JL).

Close modal

Alignment of the VH gene sequence of hAb96/3 against the V base yielded VH4 as germline counterpart. HAb96/3 displayed the highest degree of identity (86%) with germline VH4 DP-71 gene (Fig. 3, Table IV). The 39 nucleotide differences were scattered throughout FRs and CDRs and yielded amino acid replacement-silent mutation ratios (R:S ratios) of 0.5 and 0.9, respectively. The D segment sequence of hAb96/3 revealed a high homology to gene segment D6–25 over a stretch of 19 positions plus 8 additional nucleotides at the 5′-end (Fig. 3 A). Comparison of the expressed JH gene sequence with those of the known germline JH genes showed that hAb96/3 used a slightly mutated form of JH4b. The CDR3 region of the heavy chain comprises 9 additional nucleotides within the D segment and 11 nucleotide exchanges within the D and JH gene segment leading to an insertion of 3 aa and 4 aa exchanges.

Table IV.

Gene analysis of hAb96/3

VH/L FamilyClosest VH/L Gene% Nucleotide IdentityNumber of Silent and Replacement SubstitutionsaFR R:S RatioCDR R:S RatioClosest DH GeneClosest JH Gene
FR1CDR1FR2CDR2FR3CDR3
Heavy chain VHDP-71 86 (S) 10  7:15 8:9 D6-25 JH4b 
   84 (R)  (0.5) (0.9)   
Light chain VλIGLV3 93 (S) 5:5 7:3  Jλ
  s2 87 (R) (1) (2.3)  Jλ3a 
VH/L FamilyClosest VH/L Gene% Nucleotide IdentityNumber of Silent and Replacement SubstitutionsaFR R:S RatioCDR R:S RatioClosest DH GeneClosest JH Gene
FR1CDR1FR2CDR2FR3CDR3
Heavy chain VHDP-71 86 (S) 10  7:15 8:9 D6-25 JH4b 
   84 (R)  (0.5) (0.9)   
Light chain VλIGLV3 93 (S) 5:5 7:3  Jλ
  s2 87 (R) (1) (2.3)  Jλ3a 
a

Nucleotide substitutions within the primer regions are not included in calculation.

Alignment of hAb96/3 Vλ nucleotide sequence displayed the highest degree of identity (93%) with the germline IGLV3S2 gene of the Vλ3 family (Fig. 3, Table IV). The nucleotide differences resulted in higher putative R:S ratios within the CDR than the FR, being 2.3 and 1.0, respectively. More than one-half of the nucleotide changes were replacement changes. The JL segment of hAb96/3 exhibited two silent mutations and a stretch N-terminal additions coding for four amino acids within the CDR3 region (Fig. 3).

The processes of humoral immune response leading to production of islet cell autoantibodies in IDDM are largely unknown. The disease is believed to be mainly due to a cytotoxic immune response against islet cell Ags, resulting in pancreatic β cell destruction. Recent evidence, however, indicates that B lymphocytes play a critical role within the autoimmune process given that B lymphocyte-deficient mice are free of insulitis and diabetes (26). A variety of islet cell autoantigens is recognized by autoreactive T and B lymphocytes. Autoantibodies specific for self-Ags arise long before the clinical onset of IDDM and provide valuable markers for disease prediction. They may serve to investigate the autoimmune process to define autoantigens being attacked and to determine epitope structures.

Our studies focused on generating human mAbs directed against the diabetes-associated autoantigen PTP-like IA-2 to perform a detailed mapping of IA-2 epitopes recognized by autoreactive B lymphocytes present in IDDM patients. In studies performed with sera from IDDM patients, it has been previously described that the cytoplasmic portion of IA-2 (IA-2ic) is the major target of humoral autoimmunity. Sera react against either the juxtamembrane region or the PTP-like domain of IA-2ic, and there is no binding of autoantibodies to the IA-2 ecto (luminal)- and transmembrane domain (11, 27). Therefore, the complete cytoplasmic IA-2ic part was sufficient for screening of anti-IA-2 Ab producing B cell lines.

By combining EBV transformation with a high throughput ELISA screening system, we succeeded in generating five human B cell lines producing anti-IA-2-specific mAbs (hAb76/12, hAb96/3, hAb96/4, hAb96/5, and hAb103/5) from peripheral blood of three IDDM patients exhibiting very high anti-IA-2 reactivity. To generate only IgG-producing B cell lines, we preselected IgG-positive B cells before EBV transformation as described by Richter et al. (28). All isolated B cell lines produced anti-IA-2 Abs of the IgG class. Four of the human B cell lines maintained stable Ab production for several months; only one line (96/3) showed stability for >1 year with a production rate of 5–8 μg/106cells in 24 h. The generation of stable B cell lines for production and analysis of human autoantibodies is still not a routine procedure such as the generation of murine hybridomas. In the context of autoimmune diabetes to date, only human B cell lines secreting autoantibodies against insulin (29) and glutamate decarboxylase (28) have been described. To our knowledge, this is the first report describing hAbs against IA-2. These hAbs should greatly facilitate a detailed mapping of the autoantigenic epitopes within the IA-2ic sequence. We performed epitope mapping studies using in a first step a series of truncated forms of IA-2 and IA-2β. These studies revealed that the five isolated mAbs recognized distinct epitopes. Two hAbs were directed against the PTP-like domain of IA-2 and IA-2β, two hAbs were restricted toward the PTP-like domain of IA-2 and one hAb was reactive against the juxtamembrane region, which is unique to IA-2 (11). Fine mapping was performed by using mutant and chimeric constructs of IA-2 and IA-2β. The epitopes of the cross-reactive anti-IA-2/IA-2β autoantibodies could be limited to a region between residues 777 and 979, which includes the most conserved region among various PTPs between residues 777 and 937 (30). Cross-reactivity is likely to occur within this conserved region. Critical residues for the epitope of the weakly cross-reactive hAb76/12 were localized in the region between residues 794 and 845. Mutagenization of the IA-2-specific residues 804, 813, 821, and 822 to the corresponding IA-2β-specific residues did not change reactivity of this mAb. Presumably, the remaining IA-2-specific residues 815, 818, 829, 830, and 834 were responsible for the higher reactivity toward IA-2 than toward IA-2β. The epitope recognized by hAb96/3 could not be identified with specific residues within the region 794–845, because there was equally strong reactivity between the two isoforms and none of the mutated positions influenced Ab binding. A cross-reactive epitope region located in the PTP domain between residues 687 and 979 has already been described (17). Our results demonstrate that at least two different epitopes exist within the PTP-like domain of IA-2 which can induce Abs cross-reactive with the IA-2β isoform.

The two IA-2-specific Abs also recognized different epitopes. One was located within the PTP-like domain requiring aa 687–979. Although Lampasona et al. (11) already described this region being targeted by IDDM sera autoantibodies, fine mapping with hAb96/4 revealed that region 687–777 and the IA-2-specific residue 804 play a critical role for epitope formation. Because none of the chimeric constructs reacted with this Ab, IA-2-specific amino acids outside the IA-2-region 794–889 were also important for Ab binding. For the epitope recognized by the second IA-2-specific Ab (hAb96/5), aa 889–979 at the C terminus of IA-2 were required. According to Lampasona et al. (11), most IDDM patients have Abs directed against the region between 777 and 937 of the PTP-like domain. Four of our five mAbs recognized epitopes within or very close to this described epitope region. In addition, competition studies performed with the PTP-like domain specific hAb96/3 showed that it could be very efficiently competed by patients’ sera, demonstrating that an identical or a closely adjacent epitope was targeted by serum Abs induced in a majority of patients.

Immunohistochemistry studies demonstrated that none of the IA-2-specific autoantibodies reacted with thyroid, liver, lung, stomach, renal, or intestine tissues, but all five reacted with pancreatic islets of different mammalian species and with neurons of the cerebellar cortex. IA-2 and also IA-2β are predominantly expressed in cells of neuroendocrine origin, particularly in pancreatic islets and brain (10). Double staining by the PTP domain-specific hAb96/3 showed that this Ab reacted with pancreatic β cells and also with α cells. Lu et al. (31) found a differential expression of IA-2 and IA-2β in pancreatic cell lines. IA-2 was preferentially expressed in the α cell line, whereas IA-2β was preferentially expressed in a β cell line. The fact that hAb96/3 is cross-reactive with IA-2β would explain why it stained both α cells and β cells.

Analysis of the nucleotide sequences of the anti-IA-2-specific hAb96/3 revealed that the Ab used a gene of the VH4 family in association with the Vλ3 gene. The highest homology for the VH gene was found with germline gene DP-71, which has already been reported to be preferably rearranged in GAD-specific hAbs. Richter et al. (32) reported the sequence analysis of seven anti-GAD65 hAbs and found that three of them used a member of the VH4 family. Two of them derived from the same germline gene as our IA-2-specific hAb 96/3. Moreover, another two anti-GAD-specific hAbs described by Madec et al. (33) used a member of the VH4 family, one of them being germline gene DP-71. Thus, the VH4 family and especially the DP-71 germline gene seem to be overrepresented in autoantibodies associated with autoimmune diabetes. Comparison of the VH gene sequence between DP-71 and hAb96/3 revealed that hAb96/3 accumulated a series of nucleotide exchanges. These mutations were scattered throughout the CDR (22 mutations) and FR (17 mutations). Eight mutations within the CDR and seven mutations within the FR resulted in amino acid replacements. The resulting R:S ratio within CDR was 0.9 and 0.5 within the FR, which is relatively low compared with V genes of other high affinity Abs and autoantibodies (34). Also the VL segment of Ab96/3 displayed a moderate R:S ratio (2.3) throughout the CDR and, as expected, a low R:S ratio (1.0) in the respective FR. The relatively low R:S ratios found in the VH and VL segment may be explained in at least two ways: 1) what we considered as silent mutations could in fact derive from the existence of allelic VH or VL gene variants used by this patient; and 2) a number of nonrelevant silent mutations may accumulate in the V gene regions during a chronic autoimmune disease like type I diabetes, and as a consequence R:S values become blurred. However, the presence of mutations observed in both VH and VL gene segments and the high affinity of this Ab (KD 0.13 nM) are clearly indicative for an Ag-driven selection process.

In summary, we generated five human monoclonal anti-IA-2 specific IgG Abs directed against five distinct epitope regions of IA-2ic. All of these were conformational epitopes that required larger protein domains for binding, suggesting that native folded IA-2 is the immunogen for autoreactive B-cells. Whereas most of the epitopes clustered in the PTP-like domain, one epitope could be localized to the juxtamembrane region. From one of the PTP domain-specific hAbs the V-genes could be characterized. It used a VH germline gene already described to be overrepresented in autoantibodies associated with IDDM. Using this hAb for competition studies, it was shown that most IDDM patient sera contained high amounts of autoantibodies recognizing an identical or a closely adjacent epitope. Because this hAb can be provided in unlimited amounts, it should be of great value for development and standardization of anti IA-2 screening assays for prediction of IDDM.

We thank Dr. M. Grol (Roche Diagnostics) for performing the BIAcore analysis and Dr. L. Mantovani (Roche Diagnostics) for discussion and for critically reading the manuscript.

2

Address correspondence and reprint requests to Dr. J. Endl, Roche Diagnostics GmbH, Nonnenwald 2, D-82377 Penzberg, Germany.

3

Abbreviations used in this paper: IDDM, insulin-dependent diabetes mellitus; GAD, glutamic acid decarboxylase; PTP, protein tyrosine phosphatase; hAb, human mAb; IA-2ic, intracellular IA-2; ABTS, 2,2′-azinobis(3-ethylbenzthiazoline-6-sulfonic acid); DIG, digoxigenin; scFv, single-chain variable fragment; FR, framework region; CDRs, complementarity-determining regions; R:S ratio, replacement-silent mutation ratio.

1
Botazzo, G. F., A. Florin-Christensen, D. Doniach.
1974
. Islet cell antibodies in diabetes mellitus with autoimmune polyendocrine deficiency.
Lancet
2
:
1279
2
Baekkeskov, S., H. J. Aanstoot, S. Christgau, A. Reetz, M. Solimena, M. Cascalho, F. Folli, H. Richter Olesen, P. De Camilli.
1990
. Identification of the 64K autoantigen in insulin-dependent diabetes as the GABA-synthesizing enzyme glutamic dacarboxylase.
Nature
347
:
151
3
Palmer, J. P., C. M. Asplin, P. Clemons, K. Lyen, O. Tapati, P. Raghu, T. L. Pauquette.
1984
. Insulin autoantibodies in insulin-dependant diabetes before insulin treatment.
Science
222
:
13337
4
Rabin, D. U., S. M. Pleasic, J. A. Shapiro, H. Yoo Warren, J. Oles, J. M. Hicks, D. E. Goldstein, P. M. Rae.
1994
. Islet cell antigen 512 is a diabetes-specific islet autoantigen related to protein tyrosine phosphatases.
J. Immunol.
152
:
3183
5
Cui, L., W. P. Yu, H. J. DeAizpurua, R. S. Schmidli, C. J. Pallen.
1996
. Cloning and characterization of islet cell antigen-related protein tyrosine phosphatase (PTP), a novel receptor like PTP and autoantigen in insulin-dependent diabetes.
J. Biol. Chem.
271
:
24817
6
Hatfield, E. C. I., C. J. Hawkes, M. A. Payton, M. R. Christie.
1997
. Crossreactivity between IA-2 and phogrin/IA-2-β in binding to autoantibodies in IDDM.
Diabetologia
40
:
1327
7
Wasmeier, C., J. C. Hutton.
1996
. Molecular cloning of phogrin, a protein phosphatase homologue localized to insulin secretory granule membranes.
J. Biol. Chem.
271
:
18161
8
Verge, C. F., R. Gianani, E. Kawasaki, L. Yu, M. Pietropaolo, R. A. Jackson, P. H. Chase, G. S. Eisenbarth.
1996
. Prediction of type I diabetes in first degree relatives using a combination of insulin, GAD, and ICA512bdc/IA-2 autoantibodies.
Diabetes
45
:
926
9
Lan, M. S., J. Lu, Y. Goto, A. L. Notkins.
1994
. Molecular cloning and identification of a receptor-type protein tyrosine phosphatase, IA-2, from human insulinoma.
DNA Cell Biol.
13
:
505
10
Notkins, A. L., M. S. Lan, R. D. G. Leslie.
1998
. IA-2 and IA-2β: the immune response in IDDM.
Diabetes Metab. Rev.
14
:
85
11
Lampasona, V., M. Bearzatto, S. Genovese, E. Bosi, M. Ferrari, E. Bonifacio.
1995
. Autoantibodies in insulin-dependent diabetes recognize distinct cytoplasmic domains of the protein tyrosine phosphatase-like IA-2 autoantigen.
J. Immunol.
157
:
2707
12
Solimena, M., R. Drikx, J. M. Hermel, S. Pleasie-Williams, J. A. Shapiro, L. Caron, D.U. Rabin.
1996
. ICA512, an autoantigen of type I diabetes is an intrinsic membrane protein of neurosecretory granules.
EMBO J.
15
:
2102
13
Lu, J., A. L. Notkins, M. S. Lan.
1994
. Isolation, sequence and expression of a novel mouse brain cDNA, mIA-2, and its relatedness to members of the protein tyrosine phophatase family.
Biochem. Biophys. Res. Commun.
204
:
930
14
Payton, M. A., C. J. Hawkes, M. R. Christie.
1995
. Relationship of the 37.000- and 40.000-M(r) tryptic fragments of islet antigens in insulin-dependant diabetes to the protein tyrosine phosphatase-like molecule IA-2 (ICA 512).
J. Clin. Invest.
96
:
1506
15
Tuson, J. R., E. W. Pascoe, D. A. Jacob.
1990
. A novel immunohistochemical technique for demonstration of specific binding of human monoclonal antibodies to human cryostat tissue sections.
J. Histochem. Cytochem.
38
:
923
16
Srikanta, S., A. Rabizadeh, M.A. K Omar, S. Eisenbarth.
1984
. Assay for islet cell antibodies: protein-A-monoclonal antibody method.
Diabetes
34
:
300
17
Bonifacio, E., V. Lampasona, P. J. Bingley.
1998
. IA-2 (islet cell antigen 512) is the primary target of humoral autoimmunity against type 1 diabetes-associated tyrosine phosphatase autoantigens.
J. Immunol.
161
:
2648
18
McCafferty J., H. Hoogenboom, and D. Chriswell. 1996. In Antibody Engineering: a Practical Approach. Oxford University Press, New York, p.1.
19
Maniatis, T., E. F. Fritsch, J. Sambrook.
1989
.
Molecular Cloning: a Laboratory Manual
2nd Ed. Cold Spring Harbor Laboratory Press, Plainview, NY.
20
Kabat E., 1991. In Sequences of Proteins of Immunological Interest, 5th Ed. National Institutes of Health Publication, U.S. Department of Health and Human Services, Bethesda, MD.
21
Casali, P., A. L. Notkins.
1983
. Probing the human B-cell repertoire with EBV: polyreactive antibodies and CD5-+ B-lymphocytes.
Annu. Rev. Immunol.
7
:
513
22
Crawford, D. H..
1985
. E. G. Engelmann, and K. H. Foung, and J. Larrick, and A. Raubitschek, eds.
Human Hybridomas and Monoclonal Antibodies
37
Plenum, New York.
23
Tosato, P., J. Tanner, K. D. Jones, M. Revel, S. E. Pike.
1990
. Identification of interleukin-6 as an autocrine growth factor for Epstein-Barr virus-immortalized B-cells.
J. Virol.
64
:
3033
24
Dean, B. M., G. F. Bottazzo, A. G. Cudworth.
1983
. IgG subclass distribution in organ specific autoantibodies: the relationship to complement fixing activity.
Clin. Exp. Immunol.
52
:
61
25
Schatz, D. A., D. J. Barrett, N. K. MacLaren, W. J. Riley.
1983
. Polyclonal nature of islet cell autoantibodies in insulin-dependent diabetes.
Autoimmunity
1
:
45
26
Serreze, D. V., S. A. Fleming, H. D. Chapman, S. D. Richard, E. H. Leiter, R. M. Tisch.
1998
. B lymphocytes are critical antigen-presenting cell for the initiation of T cell-mediated autoimmune diabetes in nonobese diabetic mice.
J. Immunol.
161
:
3912
27
Zhang, B., M. S. Lan, A. L. Notkins.
1997
. Autoantibodies to IA-2 in IDDM: location of major antigenic determinants.
Diabetes
46
:
40
28
Richter, W., J. Endl., T. H. Eiermann, M. Brandt, R. Kiensch-Engel, C. Thivolet, H. Jungfer, W. A. Scherbaum.
1992
. Human monoclonal islet cell antibodies from a patient with insulin-dependent diabetes mellitus reveal glutamate decarboxylase as the target antigen.
Proc. Natl. Acad. Sci. USA
89
:
8467
29
Ikematsu, H., J. Ichiyoshi., E. W. Schettino, M. Nakamura, P. Casali.
1994
. VH and Vκ segment structure of anti-insulin IgG autoantibodies in patients with insulin-dependent diabetes mellitus.
J. Immunol.
152
:
1430
30
Fischer, E. H., H. Charbonneau, N. K. Tonks.
1991
. Protein tyrosine phosphatases: a diverse family of intracellular and transmembrane enzymes.
Science
253
:
401
31
Lu, J., Q. Li, H. Xie, Z. Chen, A. E. Borovitskaya, N. K. MacLaren, A. L. Notkins, M. S. Lan.
1996
. Identification of a second transmembrane tyrosine phosphatase, IA-2β, as an autoantigen in insulin-dependent diabetes mellitus: precursor of the 37-kDa tryptic fragment.
Proc. Natl. Acad. Sci. USA
93
:
2307
32
Richter, W., K. M. Jury, D. Loeffler, B. J. Manfras, T. H. Eiermann, B. O. Boehm.
1995
. Immunoglobulin variable gene analysis of human autoantibodies reveals antigen-driven immune response to glutamate decarboxylase in type I diabetes mellitus.
Eur. J. Immunol.
25
:
1703
33
Madec, A., F. Rousset, S. Ho, F. Robert, C. Thivolet, Y. Orgiazzi, S. Lebecque.
1996
. Four IgG anti islet human monoclonal antibodies isolated from a type 1 diabetes patient recognize distinct epitopes of glutamic acid decarboxylase 65 and are somatically mutated.
J. Immunol.
156
:
3541
34
Mantovani, L., R. L. Wilder, P. Casali.
1993
. Human rheumatoid B-1a (CD5* B) cells make somatically hypermutated high affinity IgM rheumatoid factors.
J. Immunol.
151
:
473