Rapid Publications Antibodies to Glutamic Acid Decarboxylase Discriminate Major Types of Diabetes Mellitus MERRILL J. ROWLEY, IAN R. MACKAY, QIAO-YI CHEN, WILLIAM J. KNOWLES, AND PAULZ. ZIMMET

Insulin-dependent diabetes mellitus (IDDM) is marked by circulating antibodies to a 64,000-Mr islet cell antigen identified as glutamic acid decarboxylase (GAD). We describe a radioimmunoprecipitation assay with GAD isolated from pig brain. The sera tested were from 80 patients with IDDM including 26 with disease of recent onset and 54 with disease of longer duration (3-42 yr), 20 with non-insulin-dependent diabetes mellitus (NIDDM), and 55 nondiabetic subjects. Conventional assays for islet cell cytoplasmic antibodies were performed concurrently. The level of antibody in serum was expressed in units based on percentage reactivity of a standard reference serum. The frequency of antibody to GAD in IDDM was 69% in short-duration cases and 59% in longduration cases. The latter was substantially higher than the frequency of islet cell cytoplasmic antibody. Antibodies to GAD were elevated (means±3 SD) in 5% NIDDM cases and in none of the nondiabetic subjects. A simple laboratory test with a defined autoantigen has substantial implications for population screening and early diagnosis of IDDM and for better understanding of its pathogenesis. Diabetes 41:548-51, 1992

D

iabetes mellitus is a heterogeneous disorder. Insulin-dependent (IDDM) and non-insulin-dependent diabetes mellitus (NIDDM) are subtyped physiologically according to patients' dependence for survival on treatment with insulin (1). IDDM is attributed to autoimmunity (2) by reason of

From The Centre for Molecular Biology and Medicine, Monash University, Clayton; the International Diabetes Institute, Caulfield General Medical Centre, Caulfield South, Australia; and Molecular Diagnostics Inc., West Haven, Connecticut. Address correspondence and reprint requests to Professor Paul Z. Zimmet, International Diabetes Institute, Caulfield General Medical Centre, 260 Kooyong Road, Caulfield South, 3162, Australia. Received for publication 2 December 1991 and accepted 2 January 1992.

548

disease associations, autoantibodies to pancreatic islet cell antigens, similarities with animal models of IDDM and HLA linkages. The disease-relevant autoantigen or antigens in IDDM include a 64,000-Mr protein demonstrable by immunoprecipitation from pancreatic islets (3) but hitherto unidentified because of low abundance and assay difficulties. Immunoassay for antibody to the 64,000-Mr antigen indicated that it was a specific marker for IDDM and preceded symptomatic onset of disease (4). The recognition in cases of the Stiff Man syndrome of coexisting IDDM and a neural antigen of 64,000-65,000 Mr identified as glutamic acid decarboxylase (GAD) led Baekkeskov et al. (5) to establish the coidentity of their 64,000-Mr antigen and GAD. In this study, we describe a high frequency of positive results by radioimmunoprecipitation assay with IDDM sera and a GAD preparation from pig brain and propose that detection of antibodies to GAD may be a useful diagnostic test as an autoimmune marker associated with IDDM. RESEARCH DESIGN AND METHODS

Sera were tested from 80 patients with IDDM, 20 patients with NIDDM, 3 with unclassified diabetes, and 55 nondiabetic blood donors with no known first-degree relatives with diabetes. The diagnosis of IDDM or NIDDM was made according to the National Diabetes Data Group criteria for the classification of diabetes mellitus (6). Thus, our patients with IDDM were nonobese and presented symptoms such as polydipsia, polyuria, and weight loss; most patients became diabetic before 40 yr of age, and all required daily insulin injections to control blood glucose concentrations. Our patients with NIDDM were mostly obese and age at onset was >40 yr, and in almost all patients blood glucose concentrations were adequately controlled by diet or diet and oral hypoglycemic drugs. The 3 unclassified patients presented as NIDDM with blood glucose concentrations adequately controlled by diet and subsequent oral hypoglycemic drugs for 4-7 yr before becoming insulin dependent with low C-pep-

DIABETES, VOL. 41, APRIL 1992

M.J. ROWLEY AND ASSOCIATES

TABLE 1

Anti-glutamic acid decarboxylase (GAD) n(M/F) * Insulin-dependent diabetes mellitus Short term (5 JDF U. Preparation of pork GAD. Fresh pig brain was homogenized in ice-cold water containing 0.1 mM 2-aminoethylisothiouronium bromide (AET), 5 mM EDTA, 20 |xM pyridoxal-5'-phosphate (PLP), 0.1 mM phenylmethylsulfonylfluoride (PMSF), and 20 |xM leupeptin. The supernatant was clarified by successive centrifugations for 45 min at 4°C at 20,000 and 100,000 g (8). To purify GAD, the supernatant was passed over an affinity column in which antibody to GAD-1 (9) was coupled to Sepharose CL-4B (Pharmacia, Uppsala, Sweden) with cyanogen bromide according to the supplier's instructions. This antibody binds to both isoforms of GAD. GAD was eluted with 50 mM KH2PO4, 10 mM diethylamine, 20 mM glutamic acid, 1 mM AET, 20 JJLM PLP, pH 10.5, collected as 2-ml fractions in 100 |xl 1 M NaH2PO4 (pH 6.9) containing 20 (xM PLP. Fractions were assayed by gel electrophoresis and enzyme analysis, dialyzed into 20 mM KH2PO4 (pH 7.2) containing 20 jxM PLP, and stored at -20°C in 50% glycerol. GAD enzyme assay. GAD catalyzes the production of T-aminobutyric acid (GABA) from glutamate. The activity of the enzyme was measured by a method modified from Chude and Wu (10) that depends on the binding of glutamate but not GABA to the ion-exchange resin AG1-X8 (Bio-Rad). The reaction mixture contained GAD in 50 mM KH2PO4 with 1 mM 2-mercaptoethanol, 1 mM PMSF, 0.2 mM PLP, 1 mM AET, and 10 mM glutamate containing 200,000 cpm L-[T-3H]-glutamine (Amersham, Aylesbury, UK) in a total vol of 200 |xl. After 30 min at

DIABETES, VOL. 41, APRIL 1992

37°C, the reaction was stopped with 20 |xl of 0.25 M H2SO4. To separate GABA from glutamate, the mixture was applied to a 0.5-ml resin column prepared in a microfuge tube and centrifuged into a tube. The column was washed with 200 |xl ice-cold water. The radioactivity in the filtrate was compared with that in a sample without enzyme. Immunoprecipitation assays. Purified and enzymatically active pork GAD showed both isoforms (67,000 and 65,000 Mr) by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). This was confirmed by Western- blot analysis with a GA D- 6 antibody (11) and by direct protein sequence analysis. The GAD was iodinated with 125I to 90 ^Ci/ixg with chloramine T (12). Nonspecific reactivity with normal sera was reduced by preincubating the [125I]GAD with pooled normal sera as follows. Control human serum (250 JJLI) was added to 2 x 106 cpm of labeled GAD in 400 IJLI wash buffer (20 mM Tris [pH 7.4], 150 mM NaCI, 0.5% Triton X-100 wt/vol). After 1 h on ice, 500 jxl of 50% protein A-Sepharose (Pharmacia) was added, the mixture was left on ice for 1 h, centrifuged, and the supernatant was removed. The protein A Sepharose was washed once with 600 jil wash buffer, and the supernatants containing [125I]GAD were pooled. For immunoprecipitation, 40 |xl [125I]GAD (50,000 cpm) was added to 25 (xl test serum diluted 1:2 in cold wash buffer. The samples were left at 4°C overnight, and 50 |xl of 50% protein A Sepharose was added for 1 h at 4°C. The samples were centrifuged, the pellets were washed three times in 750 |xl wash buffer, and counted for radioactivity. RESULTS Immunoprecipitation of GAD. The immunoprecipitation reaction was quantitated by including in each assay a positive reference serum from a patient with early diabetes, which was defined to contain 100 U reactivity. The activity of test sera was expressed as a percentage of the counts precipitated by the reference serum. Sera from nondiabetic blood donors were included with every assay. The interassay coefficient of variation (CV.) was 11% (n = 6), and the intra-assay CV. was 13% (n = 9).

549

GAD AND DIABETES

dance between anti-GAD and ICA (Table 2). Note that ICA were less frequently demonstrable, particularly in long-duration cases of IDDM.

140 120 CO EH

I(

100 80 -

#»

DISCUSSION

•

The onset of IDDM is preceded by a clinically silent < destruction of the pancreatic p-cells usually by an auO I 60 toimmune process (2,13) serologically associated with autoantibodies to ICAs. These include the 64,000-Mr 55 40 protein, now recognized as GAD (3,4), the cytoplasmic ; •s* # ICA demonstrable by immunofluorescence (14) and in20 •• sulin autoantibodies (15). Antibody to the 64,000-/Wr .Is [42] ••*! • [is] molecule has been extremely difficult to assay because 0 NORMAL IDDM-A IDDM-B NIDDM of the need to prepare antigen from large quantities of islet cell preparations. However, with radioiodine-labeled FIG. 1. Levels of antibody to glutamic acid decarboxylase (GAD) in GAD prepared from pig brain, the antibody could be normal subjects (n = 55), patients with insulin-dependent diabetes mellitus readily and specifically detected by an immunoprecipiof duration 3 yr (IDDM-B; n = 54), and non-insulin-dependent diabetes mellitus (NIDDM; n = 20). Dotted line, 3SD tation assay. It is of interest that the 64,000-/Wr antigen is above mean for normal subjects. an enzyme molecule pertaining to various disease-relevant autoantigens, most notably thyroid peroxidase in + + Although protein A-Sepharose was used routinely for autoimmune thyroiditis (16), H K ATPase in autoimprecipitation of Igs, 10 sera were analyzed in duplicate mune gastritis (17), and pyruvate dehydrogenase in with protein A Sepharose and protein G Sepharose with primary biliary cirrhosis (18). GAD is not species specific and, in fact, cDNAs for brain GAD from humans (19), rat similar results. The mean ± SD antibody for the 55 nondiabetic blood (20), and cat (21) have such close homology that antidonors was 7 ± 4 U. Sera were considered positive if body reactivity could be expected with any mammalian they contained 20 U of antibody, which was 3SD above GAD preparation. Also, the autoantigen is clearly not the mean of the control group. By this criterion, 50 (63%) organ specific because our source was brain. However, of 80 patients with IDDM had antibodies to GAD; 69% of we can note that the sequence of GAD from brain and 26 patients with diabetes of short duration (0-12 mo) and pancreatic islets, the major sites in which GAD is demon59% of 54 patients with diabetes of long duration (3-42 strable, appears to differ only by allelic variations (22,23). yr) had antibodies to GAD (Table 1). The mean level of GAD exists as at least two isoforms (24), and our anti-GAD in the patients who gave a positive result was preparation contained both representatives, as indicated 74 ± 31 U, and the distribution of antibodies was similar by our data by PAGE. in each group (Table 1). Of the 20 NIDDM sera, one (5%) The identification of the 64,000-/Wrantigen as GAD and was positive at a borderline level (26 U). All 3 patients the development of simple assays for anti-GAD such as with unclassified diabetes had antibodies to GAD (81- we have described should provide a valuable tool for 102JDFU). diabetes research. Studies in progress in our laboratory Islet cell antibodies. ICA were present in 17 (21 %) of 80 on effects of serum volumes and incubation times could patients with IDDM, 38% of 26 patients with early IDDM of further optimize this assay. Note that anti-GAD is not a known duration, and 13% of 54 patients with long- transient reactivity, unlike the sensitivity to the cytoplasduration IDDM; levels ranged from 5 to 320 JDF U. One mic antigen, because 60% of long-standing IDDM cases patient with NIDDM had 5 JDFU ICA (Table 1). ICA was retained positivity. Similar data were reported for antipresent in 2 of the 3 patients with unclassified diabetes (5 body to the 64,000-Mr antigen (4). We note that 23% of and 40 JDF U, respectively). There was some concor- sera from newly diagnosed cases of IDDM were negative for anti-GAD and cytoplasmic ICA. It will be of interest to make detailed comparisons between cases of acute TABLE 2 IDDM that are positive or negative for anti-GAD to gain Frequencies of anti-glutamic acid decarboxylase (GAD) and/or islet cell antibody (ICA) in insulin-dependent evidence of etiologic and/or serological heterogeneity. diabetes mellitus (IDDM) of short and long duration or The function of GAD in the pancreatic islet may relate non-insulin-dependent diabetes mellitus (NIDDM) to signaling between a-cells via the transmitter GABA (25), the catalytic product of GAD. Just how GAD IDDM becomes implicated in the islet cell autoimmune reaction Short term Long term NIDDM is unknown, as does the relationship of GAD to any (n = 54) (n = 26) (n = 20) Autoantibodies putative islet cell cytoplasmic autoantigen. Interest will also be taken in the mapping of B- and T-cell epitopes in GAD + ICA+ 8(31) 6(11) 0 10(38) 26 (48) GAD + ICA" 1(5) the GAD molecule as has been done for other recently 2(8) 1 (2) GAD - I C A + 1 (5) identified enzyme autoantigens, with specific immuno6(23) 21 (39) GAD - ICA" 18(90) therapy a reasonable expectation. Finally, the availability of simple assays for anti-GAD with purified brain prepaValues are n positive with percentages in parentheses.

550

••

DIABETES, VOL 41, APRIL 1992

M.J. ROWLEY AND ASSOCIATES

rations or recombinant antigens will make screening of large populations a practical endeavor, allowing opportunities for treatment before all islet tissue is destroyed. ACKNOWLEDGMENTS This study was supported by a grant from the Leon Mow Research Trust. M.J.R. is a Bicentennial Fellow of the Arthritis Foundation of Australia. We thank Dr. P. Hertzog for iodinations, Dr. M. Phillips for help in establishing the enzyme assay, Bart Haigh and Donna Guralski for expert technical assistance, Dr. Leon Chapman for collection of sera and data from patients' medical records, Dr. Sue Serjeantson (Canberra), Dr. Gordon Senator (Tasmania), physicians at the International Diabetes Institute for supplying patients, and Sue Fournel for preparation of the manuscript. GAD-1 and GAD-6 were obtained from the Development Studies Hybridoma Bank maintained by the Dept. of Pharmacology and Molecular Sciences, John Hopkins University School of Medicine, Baltimore, MD, and the Department of Biology, University of Iowa, Iowa City, IA (National Institute of Child Health and Human Development Contract NO1-HD-6-2915). REFERENCES 1. World Health Organization: Diabetes Mellitus: Report of a WHO Study Group. Geneva, World Health Org., 1985 (Tech. Rep. Ser., no. 727) 2. Ziegler AG, Herskowitz RD, Jackson RA, Soeldner JS, Eisenbarth GS: Predicting type I diabetes. Diabetes Care 13:762-75, 1990 3. Baekkeskov S, Nielsen JH, Marner B, Bilde T, Ludvigsson J, Lernmark A: Autoantibodies in newly diagnosed diabetic children immunoprecipitate human pancreatic islet cell proteins. Nature (Lond) 298:167-69, 1982 4. Atkinson MA, Maclaren NK, Scharp DW, Lacy PE, Riley WJ: 64000 Mr autoantibodies as predictors of insulin-dependent diabetes. Lancet 335:1357-60, 1990 5. Baekkeskov S, Aanstoot H-J, Christgau S, Reetz A, Solimena M, Cascalhom, Folli F, Ritcher Olesen H, De Camilli P: Identification of the 64K autoantigen in insulin-diabetes as the GABA-synthesizing enzyme glutamic acid decarboxylase. Nature (Lond) 347:151-56, 1990 6. National Diabetes Data Group: Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 28:1039-57, 1979 7. Bakos S, Mackay IR, Rowley MJ, Knowles W, Zimmet P: Islet cell antibodies and other markers of autoimmunity and diabetes mellitus in Nauruans. Diabetologia. 34:796-800, 1991.

DIABETES, VOL. 41, APRIL 1992

8. Wu J-Y, Denner L, Lin C-T, Song G: L-Glutamate decarboxylase from brain. Methods Enzymol 113:3-10, 1985 9. Chang YC, Gottlieb Dl: Characterization of the proteins purified with monoclonal antibodies to glutamic acid decarboxylase. J Neurosci 8:2123-30, 1988 10. Chude O, Wu J-Y: A rapid method for assaying enzymes whose substrates and products differ by charge: application to brainLglutamate decarboxylase. J Neurochem 27:83— 86, 1976 11. Gottlieb Dl, Chang YC, Schwob JE: Monoclonal antibodies to glutamic acid decarboxylase. Proc Natl Acad Sci USA 83:8808-12, 1986 12. O'Rourke EC, Drummond RJ, Creasey AA: Binding of 125l-labeled recombinant interferon (IFN- Ser17) to human cells. Mol Cell Biol 4:2745-49, 1984 13. Maclaren NK, Schatz DS, Drash A, Graves G: Initial pathogenic events in IDDM. Diabetes 38:534-38, 1988 14. Bottazzo GF, Florin-Christensen A, Doniach D: Islet cell antibodies in diabetes mellitus with autoimmune polyendocrine deficiencies. Lancet 2:1279-83, 1974 15. Palmer JP, Asplin CM, Clemons P, Lyen K, Tatpati O, Raghu PK, Paguette TL: Insulin antibodies in insulin-dependent diabetes before insulin treatment. Science 222:1337-39, 1983 16. Czarnocka B, Ruf J, Ferrand M, Carayon P, Lissitzky S: Purification of the human thyroid peroxidase and its identification as the microsomal antigen involved in autoimmune thyroid disease. FEBS LefM 90:147-52, 1985 17. Gleeson PA, Toh B-H: Molecular targets in pernicious anaemia. Immunol Today 12:233—39, 1991 18. Gershwin ME, Mackay IR: Primary biliary cirrhosis: paradigm or paradox for autoimmunity. Gastroenterology 100:822-33, 1991 19. Cram DS, Barnett LD, Joseph JL, Harrison LC: Cloning and partial nucleotide sequence of human glutamic acid decarboxylase (GAD) cDNA from brain and pancreatic islets. Biochem Biophys Res Commun 176:1239-44, 1991 20. Wyborski RJ, Bond RW, Gottlieb Dl: Characterization of a cDNA coding for rat glutamic acid decarboxylase. Mol Brain Res 8:19398, 1990 21. Kobayashi Y, Kaufman DL, Tobin AJ: Glutamic acid decarboxylase cDNA: nucleotide sequence encoding an enzymatically active fusion protein. J Neurosci 7:2768-72, 1987 22. Karlsen AE, Hagoplan WA, Grubin CE, Dube S, Disteche CM, Adler DA, Barmeier H, Mathewes S, Grant FJ, Foster D, Lernmark A: Cloning and primary structure of a human isoform of glutamic acid decarboxylase from chromosome 10. Proc Natl Acad Sci USA 88:8337-41, 1991 23. Michelson BK, Petersen JS, Boel E, Moldrup A, Dryberg T, Madsen OD: Cloning, characterization, and autoimmune recognition of rat islet glutamic acid decarboxylase in insulin-dependent diabetes mellitus. Proc Natl Acad Sci USA 88:8754-58, 1991 24. Kaufman DL, McGinnis JF, Krieger NR, Tobin AJ: Brain glutamate decarboxylase cloned in T-aminobutyric acid. Science 232:113840, 1986 25. Reetz A, Solimena MM, Matteoli M, Folli F, Takei K, De Camilli P: GABA and pancreatic B—cells: colocalization of glutamic acid decarboxylase (GAD) and GABA with synaptic-like microvesicles suggests their role in GABA storage and secretion. EMBO J 10:1275-84, 1991

551