Mapping of Human Autoantibody Epitopes on Aromatic L-Amino Acid

0021-972X/07/$15.00/0 Printed in U.S.A.

The Journal of Clinical Endocrinology & Metabolism 92(3):1096 –1105 Copyright © 2007 by The Endocrine Society doi: 10.1210/jc.2006-2319

Mapping of Human Autoantibody Epitopes on Aromatic L-Amino Acid Decarboxylase Paola Candeloro, Carla Borri Voltattorni, Roberto Perniola, Mariarita Bertoldi, Corrado Betterle, Massimo Mannelli, Roberta Giordano, Annamaria De Bellis, Claudio Tiberti, Stefano Laureti, Fausto Santeusanio, and Alberto Falorni, on behalf of the Italian Addison Network* Department of Internal Medicine (P.C., S.L., F.S., A.F.), Section of Internal Medicine and Endocrine and Metabolic Sciences, University of Perugia, 06126 Perugia, Italy; Department of Morphological and Biochemical Sciences (C.B.V., M.B.), University of Verona, 37134 Verona, Italy; Pediatrics Unit (R.P.), Regional Hospital “Vito Fazzi”, 73100 Lecce, Italy; Department of Medical and Surgical Sciences (C.B.), University of Padova, 35131 Padova, Italy; Department of Clinical Pathophysiology (M.M.), Section of Endocrinology, University of Florence, 50139 Florence, Italy; Division of Endocrinology and Metabolism (R.G.), Department of Internal Medicine, University of Turin, 10126 Turin, Italy; Department of Clinical and Experimental Medicine and Surgery “F. Magrassi, A. Lanzara” (A.D.B.), Second University of Naples, 80131 Naples, Italy; and Department of Clinical Sciences (C.T.), University of Rome “La Sapienza”, 00161 Rome, Italy Context: Aromatic L-amino acid decarboxylase (AADC) is target of autoantibodies in autoimmune polyendocrine syndrome I (APS I), especially in patients with autoimmune hepatitis. Little information is currently available on AADC autoantibody epitopes and on the interrelation between autoantibody-mediated inhibition of enzymatic activity and epitope specificity. Design: We tested the immunoreactivity of full-length porcine AADC and of eight fragments of the enzyme with human serum from 18 patients with APS I, 199 with non-APS I autoimmune Addison’s disease, 124 with type 1 diabetes mellitus, 36 with Graves’ disease, and 141 healthy control subjects, and we evaluated the autoantibodymediated enzymatic inhibition. Results: AADC antibodies (Ab) were detected in 12 of 18 (67%) APS I patients and in six of 199 (3%) autoimmune Addison’s disease pa-

A

UTOIMMUNE ADDISON⬘S DISEASE (AAD) is the most prevalent form of primary adrenal insufficiency in Western countries and Japan (1, 2) and is made evident by the appearance of circulating adrenal autoantibodies (3, 4). It may present as an isolated disorder, but is often part of an autoimmune polyendocrine syndrome (APS). AAD represents a major component of both APS type I and APS type II (1, 5). APS I is a rare autosomal recessive disease defined by the

First Published Online January 2, 2007 * See Acknowledgments for additional members of the Italian Addison Network. Abbreviations: aa., Amino acid; AAD, autoimmune Addison’s disease; AADC, aromatic l-amino acid decarboxylase; Ab, antibodies; ACA, adrenal cortex autoantibodies; APS, autoimmune polyendocrine syndrome; DOPA, 3,4-dihydroxyphenylalanine; GAD, glutamic acid decarboxylase; GADA, GAD autoantibody; IA-2A, protein tyrosinephosphatase-like pancreatic islet cell antigen-2 autoantibody; IAN, Italian Addison Network; LKM, liver kidney microsome; 17␣OHAb, 17␣hydroxylase autoantibody; 21OHAb, 21-hydroxylase autoantibody; P450sccAb, P450 side-chain cleavage autoantibody; PLP, pyridoxal 5⬘phosphate; T1DM, type 1 diabetes mellitus. JCEM is published monthly by The Endocrine Society (http://www. endo-society.org), the foremost professional society serving the endocrine community.

tients. Four patients with autoimmune hepatitis were all positive for AADCAb. None of the 141 healthy control subjects, 82 patients with nonautoimmune adrenal insufficiency, 124 with type 1 diabetes mellitus, and 36 with Graves’ disease were found positive. Two epitope regions, corresponding to amino acids 274 –299 (E1) and 380 – 471 (E2) were identified. Localization of E1 was confirmed by displacement studies with synthetic peptides corresponding to peptides of porcine AADC. All 12 AADCAb-positive APS I sera reacted with E1, and seven of 12 (58%) reacted also with E2. E2-specific, but not E1-specific, autoantibodies were associated with a significant inhibition of in vitro AADC enzymatic activity. Conclusions: We mapped the human AADCAb epitopes to the middle and COOH-terminal regions of the enzyme. Autoantibodies to the COOH-terminal region induce a significant inhibition of enzymatic activity. (J Clin Endocrinol Metab 92: 1096 –1105, 2007)

concomitant presence of at least two of three main disease components: hypoparathyroidism, adrenal insufficiency, and chronic mucocutaneous candidiasis (5). Additional components, such as autoimmune hepatitis, hypergonadotropic hypogonadism, chronic thyroiditis, type 1 diabetes mellitus (T1DM), alopecia, vitiligo, chronic atrophic gastritis, pernicious anemia, malabsorption, neoplasias, keratoconjunctivitis, and nail dystrophy may also be present. APS I is caused by mutations of the autoimmune regulator gene on chromosome 21q22.3, which encodes for a transcription factor expressed in cells of the immune system (5, 6). APS II is more typically diagnosed in adult subjects with AAD, in the presence of a concomitant thyroid autoimmune disease or T1DM (1). APS II is a complex genetic syndrome with a major contribution of the HLA class II region (7–9). In patients with AAD, the enzyme steroid-21-hydroxylase (21OH) is the main adrenal autoantigen, irrespective of the type of APS (10 –14). Other steroidogenic enzymes, such as 17␣-hydroxylase (17␣OH) and P450 side-chain cleavage enzyme (P450scc) are autoantibody targets in APS I patients, and less frequently in APS II patients (1, 3, 4, 15–18). In patients with APS I, several other autoantibody specificities have been recognized, including autoantibodies against glu-

1096

Candeloro et al. • Mapping AADC Autoantibody Epitopes

tamic acid decarboxylase (GAD65 and GAD67), protein tyrosine-phosphatase-like pancreatic islet cell antigen-2 (IA-2), thyroid peroxidase, cytochrome P450 IA2, aromatic l-amino acid decarboxylase (AADC), tryptophane hydroxylase, and histidine decarboxylase (1, 19 –24). The enzyme 3,4-dihydroxyphenylalanine (DOPA) decarboxylase is a homodimeric pyridoxal 5⬘-phosphate (PLP)dependent enzyme that catalyzes the conversion of l-DOPA, l-5-hydroxytryptophan, and other cathecol- or indolerelated l-amino acids to their corresponding amines and is more generally described as AADC (25). AADC autoantibodies (AADCAb) have been detected in a proportion of patients with APS I and related to the development of autoimmune hepatitis and vitiligo (20, 21). AADCAb have also been found in a small fraction of patients with isolated AAD or APS II (26, 27). Although it has been proposed that human autoantibodies might affect the enzymatic activity of AADC (28), little information is currently available on the interrelation between AADCAb epitope specificity and autoantibody-mediated enzymatic inhibition. With the aim to investigate the prevalence and epitope specificity of AADCAb in patients with APS I and in those with non-APS I AAD, we took advantage of a large group of samples collected by the Italian Addison Network (IAN) (8, 9, 29). We mapped the major human AADC autoantibody epitope regions and we addressed the question of whether human autoantibodies affect the enzymatic activity of AADC in vitro, by also investigating the association between autoantibody-induced inhibition of enzymatic activity and autoantibody epitope specificity. Patients and Methods Patients and healthy control subjects A total of 299 patients with a diagnosis of primary adrenal insufficiency, recruited by the IAN between 1998 and 2006, were enrolled in this study. Both diagnosis and etiological classification of primary adrenal insufficiency were made according to criteria described in a previous work of the IAN (29). Three distinct groups of IAN patients were studied. The first group was represented by 18 APS I patients (median age 23 yr, range 5– 40 yr, eight males and 10 females). Three of these 18 APS I patients were suffering from autoimmune hepatitis and were all positive for anti-liver kidney microsome (LKM) Ab. Two other APS I patients were positive for anti-LKM Ab with no or transient clinical signs of autoimmune hepatitis. Two APS I patients (including one with autoimmune hepatitis) were suffering from vitiligo. A total of 199 non-APS I AAD patients (median age 42 yr, range 6 –70 yr, 69 males and 130 females, time from diagnosis: median 6 yr, range 0 – 45 yr) constituted the second group; of these, 89 had isolated AAD, 95 had APS II, and 15 had APS IV (according to the classification proposed by Betterle et al. in Ref. 1). Another 82 patients (median age 64 yr, range 46 – 87 yr, 43 males and 39 females, time from diagnosis: median 9 yr, range 2–53 yr) were enrolled in the group of subjects suffering from nonautoimmune adrenal insufficiency; this group included patients with X-linked adrenoleukodystrophy, adrenal hypoplasia congenita, posttubercular, postsurgical, sepsis-related, or neonatal adrenal insufficiency. We also analyzed serum samples from 124 patients with recent onset T1DM (median age 13 yr, range 1– 45 yr, 70 males and 54 females) and from 36 patients with Graves’ disease (median age 31 yr, range 18 –52 yr, nine males and 27 females) with no clinical or biochemical signs of either adrenal autoimmunity or adrenal insufficiency. Diagnosis of T1DM was made according to National Diabetes Data Group criteria (30). Islet cell autoantibodies were detected in over 95% of T1DM patients. Diagnosis of Graves’ disease was confirmed by thyroid

J Clin Endocrinol Metab, March 2007, 92(3):1096 –1105

1097

ultrasound, thyroid scintiscan, and/or presence of TSH receptor autoantibodies in subjects with biochemical and clinical signs of hyperthyroidism. A group of 141 healthy subjects (median age 39 yr, range 10 –55 yr, 65 males and 76 females), with no chronic or autoimmune diseases, served as control group. All patients and healthy subjects gave their informed consent for the study, and the study was approved by the local ethic committees.

Porcine AADC plasmid preparation and development of a radiobinding assay for human autoantibody detection Full-length porcine AADC cDNA (31) was subcloned into the HindIII/XbaI site of a modified pcDNAII vector (Invitrogen, Groningen, The Netherlands), containing an AATTCACC sequence upstream of the ATG start codon to enhance the efficiency of the in vitro translation reaction, under the control of the SP6 promoter. Porcine AADC shares 88.7% homology with human AADC and 85.6% homology with rat AADC with 100% homology at the level of the PLP-binding site (32). Full-length 35S-labeled AADC protein was produced in an in vitro coupled transcription/translation system with SP6 RNA polymerase using TnT coupled reticulocyte lysate system (Promega Corp., Madison, WI) and 35S-methionine (NEN Life Science Products, Inc., Boston, MA) and used for immunoprecipitation of patient sera following a previously described method for detection of human GAD65 autoantibodies (33), with only slight modifications. In each autoantibody assay, one positive standard serum and two negative standard sera were used. The positive standard serum was from an APS I patient and the two negative sera from two healthy control subjects. AADCAb levels were expressed as a relative index using the formula: (cpm sample ⫺ mean cpm negative controls)/(cpm positive control ⫺ mean cpm negative controls). The positive standard serum from an APS I patient immunoprecipitated 32% of total trichloroacetic acid-precipitable 35S-AADC, against a percentage of 3.5–3.7 performed by the two negative standard sera from healthy subjects of the control group. The upper level of normal range of AADCAb index was set at 0.10, as calculated as the mean value ⫹ 3 sd of the AADC index observed in a total of 141 healthy control subjects.

Epitope mapping of AADCAb To characterize the epitope specificity of human AADCAb, we generated fragments of porcine AADC, corresponding to the NH2-terminal region [amino acids (aa.), 1–130, fragment NH], the middle region (aa. 131–300, fragment MID), the COOH-terminal region (aa. 301– 486, fragment COOH), and five additional fragments overlapping the first three, one corresponding to aa. 69 –228 (fragment A), one to aa. 229 –394 (fragment B), one to aa. 69 –261 (fragment D), one to aa. 301–379 (fragment E) and one to aa. 380 – 486 region (fragment F) (Fig. 1). To generate the eight fragments, we designed primers containing overhangs that included restriction sites for either HindIII or XbaI endonucleases and an AATTCACC sequence upstream of the ATG start codon to enhance the efficiency of the in vitro translation reaction (Table 1). We amplified the cDNA coding for each fragment using AccuPrime Pfx DNA polymerase (Invitrogen). Amplified DNA products were then purified from agarose gel using a commercially available kit (Wizard SV Gel Clean-up System from Promega Corp.), and each cDNA was subcloned into the HindIII/XbaI site of the plasmid vector pcDNA3.0 (Invitrogen), under the control of the T7 promoter. After each cloned cDNA had been verified by direct sequencing, in vitro transcription/translation of the eight different AADC fragments was performed using T7 reticulocyte lysate system (Promega Corp.), and immunoprecipitation assays were developed according to procedures similar to those used for the full-length AADC. Because of the different numbers of methionine residues in the different fragments (taking also into consideration the presence of an additional methionine residue added to each fragment to include an ATG start codon), radiolabeling was performed with different radiolabeled aa., namely: 35S-cysteine (NEN Life Science Products, Inc.) for fragment AADC-MID (two methionines and four cysteines), fragment AADC-COOH (four methionines and five cysteines), fragment AADC-B (four methionines and five cysteines), and fragment AADC-F (one methionine and four cysteines), and 35S-methionine for fragment AADC-NH (ten methionines and three cysteines), fragment AADC-A

1098



FIG. 1. Schematic representation of the AADC fragments used in the autoantibody epitope assays. Numbers indicate aa. positions in porcine AADC cDNA.

(seven methionines and three cysteines), fragment AADC-D (seven methionines and six cysteines), and fragment AADC-E (four methionines and one cysteine). Efficiency of in vitro translation of each AADC fragment was verified by trichloroacetic acid precipitation. For each epitopespecific Ab assay, the upper level of normal was estimated using the same 141 sera from healthy control subjects used to define the cut-off level of the full-length AADCAb assay. However, because no AADCAbpositive sera appeared to clearly react with AADC-NH, AADC-A, AADC-D, or AADC-E, results were expressed as percent of immunoprecipitation for the four epitope assays using these fragments. For the four remaining epitope-specific assays results were expressed as AADCMID index, AADC-COOH index, AADC-B index, and AADC-F index, respectively, using one positive standard serum from an APS I patient and two negative standard sera from healthy control subjects. Upper levels of normal of the eight epitope-specific autoantibody assays were 0.27 for AADC-MID index, 0.23 for AADC-COOH index, 0.05 for AADC-B index, 0.20 for AADC-F index, 4.5% immunoprecipitation of AADC-NH, 4.5% immunoprecipitation of AADC-A, 4.7% im-

munoprecipitation of AADC-D, and 4.8% immunoprecipitation of AADC-E. For competition experiments, cold AADC-NH, AADC-MID, and AADC-COOH were produced by in vitro transcription/translation using nonradiolabeled aa.

Displacement of AADCAb with an excess of nonradiolabeled recombinant porcine AADC or with synthetic peptides To address the question whether the immunoprecipitated radioactivity was specifically related to the presence of AADC autoantibodies in human serum, we performed separate experiments with an excess of cold, nonradiolabeled purified recombinant porcine AADC or cold, nonradiolabeled purified human GAD65, another PLP-binding decarboxylase. In each displacement experiment, a total of 1 ␮g of cold AADC or 1 ␮g of cold GAD65 was added to the immunoprecipitation reaction containing in vitro-translated 35S-AADC and human serum, in separate tubes. AADC cDNA was inserted into plasmid pKKDDCD3D4, which

TABLE 1. Primers used for the cloning of AADC fragments Primer sequences

AADC-NH (aa. 1–130) AADC-MID (aa. 131–300) AADC-COOH (aa. 301– 486) AADC-A (aa. 69 –228) AADC-B (aa. 229 –394) AADC-D (aa. 69 –261) AADC-E (aa. 301–379) AADC-F (aa. 380 – 486)

5⬘-CCCAAGCTTAATTCACCATGAATGCCAGCGATTTCCG-3⬘ 5⬘-GCTCTAGATTATTATTCCGGAAGCTGGAGCATTTTG-3⬘ 5⬘-CCCAAGCTTAATTCACCATGTTCCTGGCGGGCGAAGCCGGT-3⬘ 5⬘-GCTCTAGATTATTAATTAAAGTTAAATGAATCTGC-3⬘ 5⬘-CCCAAGCTTAATTCACCATGCCCCACAAATGGCTCTTGGTGA-3⬘ 5⬘-GCTCTAGATCATCAACTTTTGATCTCTGCC-3⬘ 5⬘-CCCAAGCTTAATTCACCATGCACTGGCACAGCCCGTACTT-3⬘ 5⬘-GCTCTAGATTATTATCTCTCCAGGGCCTCCTGCA-3⬘ 5⬘-CCCAAGCTTAATTCACCATGGACAAGGCTGCCGGCCTGATT-3⬘ 5⬘-GCTCTAGATTATTACCTGAAGCACAAATGCCTCAA-3⬘ 5⬘-CCCAAGCTTAATTCACCATGCACTGGCACAGCCCGTACTT-3⬘ 5⬘-GCTCTAGATTATTAACAGATGGGTCCCACTTCTAAGAG-3⬘ 5⬘-CCCAAGCTTAATTCACCATGCACAAATGGCTCTTGGTGAATT-3⬘ 5⬘-GCTCTAGATTATTAGATATAGGCCTGCAGTCCCTT-3⬘ 5⬘-CCCAAGCTTAATTCACCATGAAGCACGTGCAGCTGTCTCATG-3⬘ 5⬘-GCTCTAGATCATCAACTTTTGATCTCTGCC-3⬘

The underlined sequences are complementary to porcine AADC cDNA.


was expressed in SVS370 Escherichia coli cells using a procedure described elsewhere (31). AADC was purified to homogeneity (31, 34), and the enzyme concentration was determined by using the extinction coefficient at 280 nm of 1.3 ⫻ 105 m⫺1䡠cm⫺1 (35). Nonradiolabeled recombinant human GAD65, expressed into and purified from a baculovirus expression system, was a kind gift of Dr. John Robertson (Diamyd Co., Stockholm, Sweden). Synthetic peptides corresponding to aa. 227–241, 242–256, 249 –263, 272–286, 279 –293, and 286 –300 of porcine AADC were obtained from Primm S.r.l. (Milan, Italy). The peptides were resuspended at a concentration of 1 mg/ml in 10 or 20% acetonitrile. The immunoprecipitation buffer used in the AADCAb assay was supplemented with 0.1 mg/ml of either peptide to investigate the effect on AADCAb binding to in vitro-translated AADC-MID, the fragment that contains the tested peptides. In separate series of experiments, the AADCAb assay was performed in the presence of cold, nonradiolabeled AADC-NH, AADC-MID, or AADC-COOH fragment produced by in vitro transcription/translation. In each well, 4 ␮l (or 8 ␮l) of cold AADC fragment (or the in vitro transcription/translation mix with no cDNA, for the control reaction with no AADC fragment) was added to the immunoprecipitation mix that contained approximately 0.1 ␮l of in vitro-transcribed, in vitrotranslated 35S-AADC.

Effect of human AADCAb on in vitro AADC enzymatic activity The standard reaction for AADC enzymatic activity contained, in a final volume of 250 ␮l, 100 ␮l of 0.1 m potassium phosphate buffer, pH 6.8, 0.1 ␮m recombinant AADC (31), 10 ␮m PLP, and 0.5 mm l-DOPA in the absence or presence of serum samples (10 ␮l). After 5 min of incubation at 25 C, the reaction was stopped by heating at 100 C for 1 min, and the AADC activity was measured as described by Sherald et al. (36), as modified by Charteris and John (37). The percentage of inhibition for each sample was the mean for three independent experiments; in each case, the se was less than 5%. To test the effect of APS I AADCAb on in vitro AADC enzymatic activity serum samples were available from 15 of the 18 APS I patients.

Other autoantibody assays Autoantibodies to 21OH (21OHAb), 17␣OH (17␣OHAb), and P450scc (P450sccAb) were detected by using radiobinding assays with in vitrotranslated recombinant human autoantigens radiolabeled with 35S (12, 18), similar to that for detection of AADCAb. The upper levels of normal were 0.06, 0.10, and 0.06 for 21OHAb, 17␣OHAb, and P450sccAb, respectively. The full-length cDNAs for human 17␣OH and for human P450scc were kindly donated by Dr. Walter L. Miller (Department of Pediatrics and Metabolic Research Unit, University of California, San Francisco, CA). Autoantibodies to GAD65 (GADA) and protein tyrosine-phosphatase-like pancreatic islet cell antigen-2 (IA-2A) were detected using previously described quantitative radioimmunoprecipitation assays (38). The full-length cDNA for human GAD65 was kindly donated by Dr. Åke Lernmark (Department of Medicine, University of Washington, Seattle, WA). The (a.a.256 –556:630 –979)BDC cDNA for human IA-2 was kindly donated by Dr. George S. Eisenbarth (Barbara Davis Center for Childhood Diabetes, University of Colorado, Denver, CO). Adrenal cortex autoantibodies (ACA) were determined in two separate laboratories using indirect immunofluorescence on cryostatic sections of either human (39) or monkey (40) adrenal glands. Levels of ACA were expressed as the reciprocal of the end point dilution titer.


1099

Results Prevalence of AADCAb in patients with APS I, AAD, or other endocrine autoimmune diseases

Autoantibodies against AADC were detected in 12 of 18 patients with APS I (67%) and in six of 199 (3%) non-APS I AAD patients. In AADCAb-positive patients, AADCAb titer was significantly higher in APS I patients than in non-APS I AAD patients (P ⫽ 0.006). Of the six APS I patients found negative for AADCAb, none had autoimmune hepatitis or anti-LKM Ab, whereas one had vitiligo. On the contrary, of the 12 AADCAb-positive APS I patients, three had autoimmune hepatitis (including one with vitiligo) and two others were positive for anti-LKM Ab, with no or transient clinical signs of hepatitis. Of the six non-APS I AAD patients who tested positive for AADCAb, four (one male and three females) had isolated AAD, one male patient had AAD and Hashimoto’s thyroiditis, and one female patient had AAD and chronic active hepatitis. Time from diagnosis of AAD of the six non-APS I patients positive for AADCAb (median: 10 yr, range: 1–33 yr) was not significantly different from that of the 193 AADCAb-negative AAD patients (median: 6 yr, range 0 – 45 yr). None of 18 AAD subjects who had also vitiligo was found positive for AADCAb. On the other hand, none of the AADCAb-negative Addison patients had chronic autoimmune hepatitis. None of the patients with other forms of primary adrenal insufficiency, as well as none of T1DM or Graves’ patients, was found positive for AADCAb (Fig. 2). Similarly, none of 141 healthy subjects had an AADC index higher than the upper level of normal range. The ␹2 test showed a significant difference in the prevalence of AADCAb between the six groups tested (P ⬍ 0.001); when the prevalence of AADCAb-positive sera of the APS I patient group was compared with that of the other groups, a marked statistical significance was found (in all cases P ⬍ 0.001). No further statistical difference was found in the remaining comparisons among groups. When the AADCAb-positive sera from both APS I and non-APS I AAD groups were tested with an excess of cold, nonradiolabeled porcine AADC, a complete displacement of

Statistical analysis Differences in autoantibody frequencies among the groups were tested with the ␹2 method. Yates’ correction or the Fisher exact test were used when appropriate. Differences in autoantibody titer were tested by the nonparametric Mann-Whitney test or the Wilcoxon test for paired samples. A two-tailed P ⬍ 0.05 was considered significant in all tests.

FIG. 2. AADCAb titer (AADC index) in patients with APS I, non-APS I AAD, nonautoimmune adrenal insufficiency (NAAI), Graves’ disease (GD), T1DM, and in healthy control subjects (HC). Dotted line shows the upper level of normal range of the assay. P ⬍ 0.001 APS I vs. all other groups.

1100


radiolabeled AADC from autoantibody binding was observed, and the AADCAb titer decreased from 0.12– 0.96 to 0 – 0.10 (Fig. 3A). On the other hand, no significant displacement of radiolabeled AADC was observed when the immunoprecipitation reaction was carried out in the presence of an excess of cold, nonradiolabeled, recombinant human GAD65 (Fig. 3B), thus indicating a specific autoantibody binding to AADC. We then tested whether the presence of AADCAb is associated with that of other adrenal and nonadrenal autoantibodies. As shown in Table 2, no significant difference in 21OHAb, 17␣OHAb, P450sccAb, ACA, GADA, or IA-2A prevalence was observed between AADCAb-positive and -negative APS I patients. Characterization of AADCAb epitope regions

To localize the autoantibody epitope regions on AADC, we used radiolabeled fragments of the enzyme (Fig. 1). We first tested the immunoreactivity of the AADC-NH fragment, corresponding to aa. region 1–130, using the sera from 12 AADCAb-positive APS I patients and six AADCAb-positive non-APS I Addison patients (Table 3). None of the AADCAb-positive sera reacted with the AADC-NH fragment. We then tested the immunoreactivity of the AADC-MID fragment and observed that all the 12 AADCAb-positive APS I sera and two of six (33%) AADCAb-positive sera from non-APS I Addison patients were positive for AADC-MID Ab. When the radiolabeled AADC-COOH fragment was used in the assay, seven of 12 (58%) AADCAb-positive APS I sera and four of six (67%) AADCAb-positive non-APS I sera


TABLE 2. Frequency of endocrine autoantibodies in APS I patients according to the presence or absence of AADC autoantibodies Autoantibody specificity

AADCAb-positive patients (n ⫽ 12)

AADCAb-negative patients (n ⫽ 6)

21OHAb ACA 17␣OHAb P450sccAb GADA IA-2A

10 (83) 10 (83) 8 (67) 10 (83) 6 (50) 2 (17)

6 (100) 6 (100) 4 (67) 4 (67) 3 (50) 0/6

Data are presented as number (percentage).

were found also positive for AADC-COOH-Ab. No association was found between the presence of AADC-COOH-Ab and the risk for autoimmune hepatitis as three of the five (60%) APS I patients positive for anti-LKM and four of the seven (57%) negative for anti-LKM were found positive for Ab directed to the COOH-terminal region of the enzyme. The same seven of 12 (58%) APS I sera and four of six (67%) non-APS I sera found positive for AADC-COOH-Ab were also found positive for AADC-F-Ab (Table 3). Subsequently, we tested four additional fragments: AADC-A, corresponding to aa. 69 –228, AADC-B, corresponding to aa. 229 –394, AADC-D, corresponding to aa. 69 –261, and AADC-E, corresponding to aa. 301–379. By using the first one (AADC-A), none of the AADCAb-positive sera did react, this showing that the first 228 aa. of the enzyme do not contain major autoantibody epitope regions. On the contrary, all the AADCAb-positive sera reacted significantly with the second of the two fragments (AADC-B), suggesting that the major autoantibody epitopes are localized in the aa. 229 –394 region of the enzyme. To further restrict the autoantibody epitopes located in the middle region of the enzyme, we tested the reactivity of the AADC-D fragment. None of the AADCAb-positive sera reacted with this fragment, showing that the first 261 aa. of the enzyme do not contain major autoantibody epitope regions and that the middle epitopes identified by the AADC-MID fragment (aa. 130 –300) are restricted to the aa. 261–300 region. Finally, we found that none of the AADCAb-positive sera reacted with the AADC-E fragment, which shows that the COOH-terminal epitope(s) is(are) restricted to the aa. 380 – 486 region. Displacement of human AADCAb by synthetic peptides or AADC fragments

FIG. 3. AADCAb titer (AADC index) in the absence/presence of an excess of nonradiolabeled, recombinant, porcine AADC (A) or in the absence/presence of an excess of nonradiolabeled, recombinant, human GAD65 (B). Dotted line shows the upper level of normal range of the assay. APS I patients are represented by squares, and non-APS I AAD patients are represented by circles.

The use of several AADC fragments enabled the identification of two major epitope regions corresponding to the aa. 261–300 (E1) and the aa. 380 – 486 (E2) regions. To confirm the localization of the major immunodominant E1 epitope present in all AADCAb-positive APS I sera, we tested the 12 APS I sera positive for AADC-MID-Ab in the presence or absence of synthetic peptides corresponding to aa. 227–241, 242–256, 249 –263, 272–286, 279 –293, and 286 –300 of porcine AADC. When the results were expressed as percent of the observed immunoprecipitation in the absence of synthetic peptides (Fig. 4), a significant displacement of autoantibody binding was observed in the presence of either of the three synthetic peptides corresponding to aa. 272–286, 279 –293, or



1101

TABLE 3. Autoantibody epitope specificities in 12 APS I and 6 non-APS I autoimmune Addison patients positive for AADCAb Serum samples

APS I 1 2 3 4 5 6 7 8 9 10 11 12 Non-APS I AAD 1 2 3 4 5 6

AADCAb index (full length)

AADC-NH-Ab (aa. 1–130)

AADC-A-Ab (aa. 69 –228)

AADC-D-Ab (aa. 69 –261)

AADC-MID-Ab (aa. 131–300)

AADC-B-Ab (aa. 229 –394)

AADC-COOH-Ab (aa. 301– 486)

AADC-E-Ab (aa. 301–379)

AADC-F-Ab (aa. 380 – 486)

1.07 0.11 1.11 0.64 0.82 1.0 0.87 1.02 0.87 1.1 0.41 0.9

⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺

⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺

⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺

⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹

⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹

⫹ ⫺ ⫹ ⫹ ⫹ ⫺ ⫹ ⫺ ⫹ ⫺ ⫺ ⫹

⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺

⫹ ⫺ ⫹ ⫹ ⫹ ⫺ ⫹ ⫺ ⫹ ⫺ ⫺ ⫹

0.15 0.97 0.13 0.17 0.19 0.42

⫺ ⫺ ⫺ ⫺ ⫺ ⫺

⫺ ⫺ ⫺ ⫺ ⫺ ⫺

⫺ ⫺ ⫺ ⫺ ⫺ ⫺

⫺ ⫺ ⫹ ⫹ ⫺ ⫺

⫺ ⫹ ⫹ ⫹ ⫺ ⫹

⫹ ⫹ ⫺ ⫺ ⫹ ⫹

⫺ ⫺ ⫺ ⫺ ⫺ ⫺

⫹ ⫹ ⫺ ⫺ ⫹ ⫹

286 –300. Conversely, no significant displacement of autoantibody binding was observed in the presence of the synthetic peptides corresponding to aa. 227–241, 242–256, or 249 –263 (Fig. 4). To estimate the contribution of the middle epitope(s) to the overall AADCAb reactivity, seven serum samples, found positive only for E1-specific AADCAb (from five patients with APS I—nos. 2, 6, 8, 10 and 11 of Table 3—and two with non-APS I AAD—nos. 3 and 4 of Table 3), were tested in the presence of an excess of cold, nonradiolabeled AADC-NH, AADC-MID, or AADC-COOH (4 ␮l of in vitro-translated fragments per well) (Fig. 5). A significant displacement of AADC-Ab was demonstrated in the presence of an excess of AADC-MID (P ⫽ 0.013), but not of AADC-NH or AADCCOOH. Similar results were observed when 8 ␮l/well of nonradiolabeled in vitro-translated fragments were tested. Effect of human autoantibodies on in vitro AADC enzymatic activity

APS I sera determined an inhibition of the enzymatic activity that ranged from 2.2– 41% (median value: 14%), compared with an inhibition that ranged from 1–14% (median value: 7%) when five AADCAb-negative APS I sera were used in the assay (P not significant). The inhibition of the enzymatic activity obtained with the six AADCAb-positive sera from non-APS I AAD patients ranged from 6 –18% (median value: 7%), compared with an inhibition that ranged from 4 –21% (median value: 9%) observed with 10 sera from AADCAbnegative autoimmune Addison patients (P not significant). Although there was no statistically significant difference in inhibition of enzymatic activity between AADCAb-positive and AADCAb-negative sera, the degree of inhibition (median value 32%, range 3– 41%) caused by six AADCAbpositive APS I sera reacting against both fragments MID and COOH was significantly higher than that observed with four AADCAb-positive APS I sera reacting against the fragment MID, but not against the fragment COOH (median value 3.5,

When human serum was included in the in vitro assay to determine AADC enzymatic activity, 10 AADCAb-positive

FIG. 4. Displacement of AADC-MID-Ab binding by incubation with synthetic peptides corresponding to aa. 227–241, 242–256, 249 –263, 272–286, 279 –293, and 286 –300 of porcine AADC. Per each peptide, results are expressed as mean ⫾ SD of percent binding observed in the absence of synthetic peptides, in 12 APS I autoantibody-positive sera. *, P ⬍ 0.02; **, P ⬍ 0.002.

FIG. 5. AADCAb titer (AADC index) in seven serum samples positive only for E1-specific autoantibodies, in the absence/presence of cold, nonradiolabeled AADC fragments, produced by in vitro transcription/ translation (4 ␮l/well). Each line connects the results of a patient. Dotted line shows the upper level of normal range of the assay. APS I patients are represented by squares, and non-APS I AAD patients are represented by circles. *, P ⫽ 0.013.

1102


range 2.2–10; P ⫽ 0.019) (Fig. 6). Such a finding is not explained by the difference in AADCAb titer, because autoantibody titer was similar in both APS I patients exhibiting serum reactivity against the fragment COOH (AADC index median value 0.87, range 0.64 –1.11) and those not exhibiting it (AADC index median value 1.01, range 0.11–1.10). Discussion

In this study, we analyzed serum samples from 18 patients with APS I, 199 with non-APS I AAD, 82 with nonautoimmune adrenal insufficiency, 124 with T1DM, and 36 with Graves’ disease, for the presence of AADCAb. We also dissected the autoantibody epitope regions on AADC and studied the correlation between in vitro inhibition of enzymatic activity and Ab epitope specificity. We detected AADCAb in two thirds of APS I patients and in 3% of non-APS I patients with AAD, but in no patients with nonautoimmune adrenal insufficiency or T1DM or Graves’ disease. The use of AADC fragments enabled us to identify two major autoantibody epitope specificities, one in the middle region (E1), present in all AADCAb-positive APS I patients, and one in the COOH-terminal region (E2), present in 58% of them. When the AADC enzyme assay was performed in the presence of human serum, a significant inhibition of enzymatic activity was detected in the presence of AADCAb directed to the COOH-terminal region of the enzyme. Our study confirms and expands the findings of previous studies on AADCAb specificity (20, 21, 26, 27) and provides original data on the autoantibody epitope regions and on the effect of human AADCAb on in vitro enzymatic activity. The prevalence of AADCAb in our APS I patients (67%) was similar to that found in other previous studies (20, 21, 26, 27). In addition, by analyzing a large population of patients with non-APS I AAD, we confirm the 3% frequency of AADCAb positivity registered in a recent study performed on a smaller population (27). Accordingly, the presence of AADCAb appears to be a sensitive marker for APS I, but not for isolated or APS II/IV-related AAD. The diagnostic sensitivity of AADCAb for APS I is similar to that of 17␣OHAb,

FIG. 6. In vitro AADC enzymatic activity in the presence of human serum from APS I patients found negative for AADCAb, positive only for AADC-MID-Ab or positive for both AADC-MID-Ab and AADC-COOH-Ab. Results are expressed as percent of the activity observed in the absence of human serum.


but slightly lower than those of both P450sccAb and 21OHAb. Furthermore, the presence of AADCAb was not dependent on that of other autoantibodies, such as 21OHAb, 17␣OHAb, P450sccAb, GADA, or IA-2A, thus indicating that AADC targets a distinct immunologic response. The specificity of the autoantibody binding was demonstrated in our study by competition experiments with an excess of nonradiolabeled AADC. Conversely, no significant displacement was observed in the presence of an excess of recombinant human GAD65, another PLP-binding decarboxylase. Human GAD65 was selected as control protein because of the low homology with porcine AADC and the presence of GADA in nine of our 18 APS I patients. Our results can be interpreted to indicate that our assay is able to selectively detect AADCAb and that AADCAb represents a distinct autoantibody specificity from GADA. It had been suggested initially that AADC would be a pancreatic ␤-cell autoantigen (41). However, subsequent studies have failed to detect a significant reactivity to AADC in the sera from T1DM patients not having APS I (20, 27). In the present study, we detected AADCAb in none of 124 patients with recent onset T1DM. Similarly, 36 patients with Graves’ disease scored negative for AADCAb, thus indicating that this autoantibody specificity can mainly be detected in APS I patients, but not in patients with other endocrine autoimmune diseases or syndromes, as also shown by its absence in 110 patients with complete or incomplete APS II/APS IV, in our study. It has been reported that AADCAb would be present in 75–92% of APS I patients with hepatitis (20, 21, 27). Although our study cannot provide definitive data on this specific topic (because of the limited number of patients with autoimmune hepatitis), our results support the hypothesis that AADCAb is a sensitive marker for autoimmune hepatitis in patients with primary adrenal insufficiency. Indeed, the three APS I patients and the non-APS I AAD patient with autoimmune hepatitis were all found positive for AADCAb. In addition, two other APS I subjects who were positive for anti-LKM, with no or transient clinical signs of hepatitis, were also


found positive for AADCAb. The high sensitivity of AADCAb for autoimmune hepatitis is not paralleled by a similarly high diagnostic specificity because seven APS I patients and five non-APS I AAD patients were positive for AADCAb in the absence of both anti-LKM Ab and clinical signs of hepatitis. Our finding of AADCAb in seven of 13 (54%) APS I patients without hepatitis or anti-LKM Ab is in line with the results of previous studies; Husebye et al. (20) found AADCAb in 24 of 57 (42%) and Dal Pra et al. (27) found AADCAb in 10 of 14 (71%) APS I patients without hepatitis. Future follow-up studies of these AADCAb-positive subjects will be instrumental in testing the predictive value of this marker for autoimmune hepatitis. No association between AADCAb and vitiligo was observed in our population. An important novelty of our study is the use of porcine AADC, instead of rat AADC that was used in previous studies (20, 21, 26, 27). Rat and porcine AADC have an overall homology of 85.6% (32). Thus, it is not surprising that we detected AADCAb at a similar frequency to that of the studies using rat autoantigen. However, it must be noted that the rat and porcine AADC show significantly different COOHterminal tails, with only three identical aa. residues in the region corresponding to the 472– 486 peptide of porcine AADC. Consequently, it is highly unlikely that the last 15 aa. residues of the COOH-terminal tail of porcine AADC may contain an autoantibody epitope. In addition, it must be noted that human AADC ends at position 480. To better characterize the autoantibody epitope regions, we constructed eight different fragments of porcine AADC and we tested their reactivity with the sera initially found positive for AADCAb. In the first series of experiments, we observed that all the APS I AADCAb-positive sera reacted with a middle fragment (AADC-MID) corresponding to the aa. 131–300 region, evidently containing the major epitope(s), and to which we refer to as E1. In addition, separate epitope(s) was(were) detected in the COOH-terminal fragment corresponding to the aa. 301– 486 region. This second group of epitopes are referred to as E2. Differently from E1, which were reactive in all AADCAb-positive patients, E2 reacted only with seven of the 12 APS I patients found positive for AADCAb. Slightly different results were observed with the six AADCAb-positive sera from patients with isolated or APS II/IV AAD, because only two sera were positive for AADC-MID-Ab and four for AADC-COOH-Ab; moreover, such positives were mutually exclusive and no serum reacted with both E1 and E2. The difference in behavior between APS I and non-APS I sera may be related to the lower autoantibody titer observed in the latter patients, which may make difficult the accurate dissection of the epitope specificities or to a more pronounced epitope spreading in APS I. Interestingly, none of the 12 APS I or the six non-APS I AADCAb-positive sera reacted with AADC-NH, AADC-A, or AADC-D, which leads us to consider that no major autoantibody epitope is located in the region corresponding to the first 261 aa. This finding is apparently at variance with the results of a previous study (27) that suggested that additional epitopes may be located in the NH2-terminal region of the enzyme. However, our present study and the previous one (27) are based on different methodological approaches


1103

and there was a lack of statistical significance in the results of the previous study (27) regarding the existence of NH2terminal epitopes. Indeed, our strategy of using large AADC fragments does not exclude the possibility of the existence of additional, conformational epitopes in the NH2-terminal region of the enzyme. Because of the chronic nature of the autoimmune process, it is likely that some of the autoantibody epitopes may be conformational. Nevertheless, the significant displacement of AADCAb by an excess of cold AADC-MID, although not complete in some of the tested samples, confirms that one or more major E1-epitope(s) is(are) located in the middle domain of the autoantigen. The incomplete displacement of AADCAb by an excess of AADC-MID, observed in three of seven APS I-tested patients, may be the consequence of technical problems related to the use of nonpurified AADC fragment produced by in vitro transcription/translation, but can also be interpreted to indicate that additional, conformational epitopes of AADCAb are present. The use of a fragment corresponding to aa. 229 –394 (AADC-B) confirmed the immunoreactivity of the middle region of the enzyme as all the ADDCAb-positive sera scored positive for this fragment. By combining our results and the results of a previous study (27), we can hypothesize that E1 is(are) located in the region corresponding to aa. 262–300. A detailed analysis of the aa. 261–300 region of AADC reveals that aa. 262–270 is characterized by a low homology between rat and porcine with only four identical aa. residues. In addition, it is interesting to note that a cluster of five conserved aa. residues (HVDAA) can be detected upstream of the PLP binding lysine in both the 269 –273 region of porcine AADC and the 362–366 region of human GAD65, which shows that this peptide is unlikely to be involved in AADCAb reactivity. We then conclude that E1 is(are) most likely located in the aa. region 274 –299, because the NPHK peptide corresponding to aa. residues 300 –303 of porcine AADC is identical to the PLP-binding site of human GAD65. It is noteworthy that the 274 –299 region of porcine AADC is identical to the corresponding region of rat and human AADC (32). The mapping of the middle epitope region to aa. region 274 –299 was confirmed by the signification inhibition of autoantibody binding by coincubation with synthetic peptides corresponding to aa. 272–286, 279 –293, and 286 –300 (Fig. 4). Husebye et al. (28) have shown that human AADCAb may variably inhibit in vitro enzymatic activity of AADC. In our study, we show that incubation with human serum containing AADCAb induced a 2.2– 41% inhibition when APS I sera were analyzed, or a 6 –18% inhibition when non-APS I sera were studied. Although the observed inhibition was nonstatistically different from that obtained with AADCAb-negative sera, the subgroup of APS I sera reacting against the COOH-terminal epitopes induced a significantly pronounced inhibition of enzymatic activity, thus providing a clear interrelation between autoantibody epitope specificity and autoantibody-mediated enzymatic inhibition. The AADC-COOH fragment contains a peptide corresponding to aa. 328 –339 that creates a mobile loop that plays an important role in the catalytic mechanism (42). However, as for E2 location on AADC-COOH, we hypothesize that the region

1104


corresponding to aa. residues 380 – 471 (present in the AADC-F fragment) is a likely candidate to contain major autoantibody epitopes on the basis of the negative results obtained with the AADC-E fragment. Interestingly, the analysis of the crystal structure of porcine AADC (42) revealed that the COOH-terminal domain contains a four-stranded antiparallel ␤-sheet with three helices packed against the face opposite of a large central domain in which the PLP bindingsite is located, thus indirectly supporting the hypothesis that autoantibodies directed against the COOH-terminal tail may interfere with enzymatic activity. A similar pattern of autoantibody reactivity has been demonstrated for other organspecific autoantibodies, and more specifically for GAD65Ab that, similarly to AADCAb, are directed mainly (although not exclusively) to epitopes located in the middle and COOH-terminal regions (43, 44), with COOH-terminal specific autoantibodies correlating with an in vitro inhibition of GAD65 enzymatic activity in patients with Stiff-Person syndrome (45). In conclusion, we have demonstrated the frequent occurrence of AADCAb in APS I patients and the ability of autoantibodies specific to the COOH-terminal region of the enzyme to induce in vitro enzymatic inhibition. A mapping of the major AADCAb epitope regions was performed. At present, it is still unclear whether E1 and E2 identify two distinct subgroups of patients with different mechanisms of AADCAb production or are the expression of different phases of the natural history of autoantibody production which is characterized by epitope spreading. Further analyses are needed to address this specific question.


2. 3. 4. 5. 6. 7. 8.

9.

10. 11.

12.

13.

14.

Acknowledgments We thank Dr. John Robertson (Diamyd Co.) for the kind gift of purified recombinant human GAD65, Dr. Åke Lernmark (Department of Medicine, University of Washington) for the kind gift of human GAD65 cDNA, Dr. Walter L. Miller (Department of Pediatrics and Metabolic Research Unit, University of California) for the kind gift of human 17␣OH cDNA and human P450scc cDNA, and Dr. George S. Eisenbarth (Barbara Davis Center for Childhood Diabetes, University of Colorado) for the kind gift of human IA-2 (a.a.256 –556:630-979)BDC cDNA. In addition to the authors, the following members of the Italian Addison Network contributed to the collection of data and blood samples from patients with primary adrenal insufficiency: B. Ambrosi (Milan), A. Angeli (Orbassano), G. Arnaldi (Ancona), E. Arvat (Turin), A. Baccarelli (Milan), L. Barbetta (Milan), P. Beck-Peccoz (Milan), A. Bellastella (Naples), A. Bizzarro (Naples), M. Boscaro (Ancona), F. Cavagnini (Milan), F. Dotta (Siena), E. Ghigo (Turin), S. Laureti (Perugia), R. Libe` (Milan), F. Lore´ (Siena), F. Mantero (Padova), G. Mantovani (Milan), P. Paccotti (Orbassano), F. Pecori-Giraldi (Milan), P. Toja (Milan), M. Torlontano (S.Giovanni Rotondo), V. Toscano (Rome), V. Trischitta (S.Giovanni Rotondo), and R. Zanchetta (Padova). Received October 24, 2006. Accepted December 27, 2006. Address all correspondence and requests for reprints to: Alberto Falorni, M.D., Ph.D., Department of Internal Medicine, Section of Internal Medicine and Endocrine and Metabolic Sciences, Via E. Dal Pozzo, 06126 Perugia, Italy. E-mail: [email protected]. This study was supported in part by the EURAPS: Autoimmune polyendocrine syndrome type I, a rare disorder of childhood as a model of autoimmunity (Grant LSHM-CT-20005-005223, to C.B.). Disclosure Statement: The authors have nothing to disclose.

References 1. Betterle C, Dal Pra C, Mantero F, Zanchetta R 2002 Autoimmune adrenal insufficiency and autoimmune polyendocrine syndromes: autoantibodies, au-

15.

16. 17.

18.

19.

20.

21.

22. 23.

24.

toantigens, and their applicability in diagnosis and disease prediction. Endocr Rev 23:327–364 Nomura K, Depura H, Saruta T 1994 Addison’s disease in Japan: characteristics and changes revealed in a nationwide survey. Intern Med 33:602– 606 Falorni A, Laureti S, Santeusanio F 2002 Autoantibodies in autoimmune polyendocrine syndrome type II. Endocrinol Metab Clin North Am 31:369 –389 Falorni A, Laureti S 2000 Adrenal autoimmunity and correlation with adrenal dysfunction. Endocrinologist 10:145–154 Betterle C, Greggio NA, Volpato M 1998 Autoimmune polyglandular syndrome type 1. J Clin Endocrinol Metab 83:1049 –1055 Heino M, Peterson P, Kudoh J, Shimizu N, Antonarakis SE, Scott HS, Krohn K 2001 APECED mutations in the autoimmune regulator (AIRE) gene. Hum Mutat 18:205–211 Robles DT, Fain PR, Gottlieb PA, Eisenbarth GS 2002 The genetics of autoimmune polyendocrine syndrome type II. Endocrinol Metab Clin North Am 31:353–368 Gambelunghe G, Kockum I, Bini V, De Giorni G, Celi F, Betterle C, Giordano R, Libe` R, Falorni A 2005 Retrovirus-like long terminal repeat DQ-LTR13 and genetic susceptibility to type 1 diabetes mellitus and autoimmune Addison’s disease. Diabetes 54:900 –905 Ghaderi M, Gambelunghe G, Tortoioli C, Brozzetti A, Jatta K, Gharizadeh B, De Bellis A, Pecori Giraldi F, Terzolo M, Betterle C, Falorni A 2006 MHC2TA single nucleotide polymorphism and genetic risk for autoimmune adrenal insufficiency. J Clin Endocrinol Metab 91:4107– 4111 Winqvist O, Karlsson FA, Ka¨mpe O 1992 21-Hydroxylase, a major autoantigen in idiopathic Addison’s disease. Lancet 339:1559 –1562 Bednarek J, Furmaniak J, Wedlock N, Kiso Y, Baumann-Antczak A, Fowler S, Krishnan H, Craft JA, Rees Smith B 1992 Steroid 21-hydroxylase is a major autoantigen involved in adult onset autoimmune Addison’s disease. FEBS Lett 309:51–55 Falorni A, Nikoshkov A, Laureti S, Grenba¨ck E, Hulting AL, Casacci G, Santeusanio F, Brunetti P, Luthman H, Lernmark Å 1995 High diagnostic accuracy for idiopathic Addison’s disease with a sensitive radiobinding assay for autoantibodies against recombinant human 21-hydroxylase. J Clin Endocrinol Metab 80:2752–2755 Tanaka H, Perez MS, Powell M, Sanders JF, Sawicka J, Chen S, Prentice L, Asawa T, Betterle C, Volpato M, Rees Smith B, Furmaniak J 1997 Steroid 21-hydroxylase autoantibodies: measurements with a new immunoprecipitation assay. J Clin Endocrinol Metab 82:1440 –1446 Falorni A, Laureti S, Nikoshkov A, Picchio ML, Hallengren B, Vandewalle CL, Gorus FK, Belgian Diabetes Registry, Tortoioli C, Luthman H, Brunetti P, Santeusanio F 1997 21-Hydroxylase autoantibodies in adult patients with endocrine autoimmune diseases are highly specific for Addison’s disease. Clin Exp Immunol 107:341–346 Perniola R, Falorni A, Clemente MG, Forini F, Accogli E, Lobreglio G 2000 Organ-specific and non-organ-specific autoantibodies in children and young adults with autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy (APECED). Eur J Endocrinol 143:497–503 Krohn K, Uibo R, Aavik E, Peterson P, Savilahti K 1992 Identification by molecular cloning of an autoantigen associated with Addison’s disease as steroid 17␣-hydroxylase. Lancet 339:770 –773 Chen S, Sawicka J, Betterle C, Powell M, Prentice L, Volpato M, Rees Smith B, Furmaniak J 1997 Autoantibodies to steroidogenic enzymes in autoimmune polyglandular syndrome, Addison’s disease, and premature ovarian failure. J Clin Endocrinol Metab 81:1871–1876 do Carmo Silva R, Kater CE, Dib SA, Laureti S, Forini F, Cosentino A, Falorni A 2000 Autoantibodies against recombinant human steroidogenic enzyme 21-hydroxylase, side-chain cleavage and 17␣-hydroxylase in Addison’s disease and autoimmune polyendocrine syndrome type III. Eur J Endocrinol 142:187–194 Tuomi T, Bjo¨rses P, Falorni A, Partanen J, Perheentupa J, Lernmark Å, Miettinen A 1996 Antibodies to glutamic acid decarboxylase and insulindependent diabetes in patients with autoimmune polyendocrine syndrome type I. J Clin Endocrinol Metab 81:1488 –1494 Husebye ES, Gebre-Medhin G, Tuomi T, Perheentupa J, Landin-Olsson M, Gustafson J, Rorsman F, Ka¨mpe O 1997 Autoantibodies against aromatic l-amino acid decarboxylase in autoimmune polyendocrine syndrome type I. J Clin Endocrinol Metab 82:147–150 Gebre-Medhin G, Husebye ES, Gustafson J, Winqvist O, Goksøyr A, Rorsman F, Ka¨mpe O 1997 Cytochrome P450IA2 and aromatic l-amino acid decarboxylase are hepatic autoantigens in autoimmune polyendocrine syndrome type I. FEBS Lett 412:439 – 445 Ekwall O, Sjo¨berg K, Mirakian R, Rorsman F, Ka¨mpe O 1999 Tryptophan hydroxylase autoantibodies and intestinal disease in autoimmune polyendocrine syndrome type I. Lancet 354:568 Ekwall O, Hedstrand H, Haavik J, Perheentupa J, Betterle C, Gustafsson J, Husebye E, Rorsman F, Ka¨mpe O 2000 Pteridin-dependent hydroxylases as autoantigens in autoimmune polyendocrine syndrome type I. J Clin Endocrinol Metab 85:2944 –2950 Skoldberg F, Portela-Gomes GM, Grimelius L, Nilsson G, Perheentupa J, Betterle C, Husebye ES, Gustafsson J, Ronnblom A, Rorsman F, Ka¨mpe O 2004 Histidine decarboxylase, a pyridoxal phosphate-dependent enzyme, is an


25. 26.

27. 28. 29.

30. 31. 32. 33. 34.

autoantigen of gastric enterochromaffin-like cells. J Clin Endocrinol Metab 88:1445–1452 Borri Voltattorni C, Minelli A, Dominici P 1983 Interaction of aromatic amino acids in d and l forms with 3,4-dihydroxyphenylalalnine decarboxylase from pig kidney. Biochemistry 22:2249 –2254 So¨derbergh A, Rorsman F, Halonen M, Ekwall O, Bjo¨rses P, Ka¨mpe O, Husebye ES 2000 Autoantibodies against aromatic l-amino acid decarboxylase identifies a subgroup of patients with Addison’s disease. J Clin Endocrinol Metab 85:460 – 463 Dal Pra C, Chen S, Betterle C, Zanchetta R, McGrath V, Furmaniak J, Rees Smith B 2004 Autoantibodies to human tryptophan hydroxylase and aromatic l-amino acid decarboxylase. Eur J Endocrinol 150:313–321 Husebye ES, Bøe AS, Rorsman F, Ka¨mpe O, Aakvaag A, Rygh T, Flatmark T, Haavik J 2000 Inhibition of aromatic l-amino acid decarboxylase activity by human autoantibodies. Clin Exp Immunol 120:420 – 423 Falorni A, Laureti S, De Bellis A, Zanchetta R, Tiberti C, Arnaldi G, Bini V, Beck-Peccoz P, Bizzarro A, Dotta F, Mantero F, Bellastella A, Betterle C, Santeusanio F 2004 Italian Addison Network study: update of diagnostic criteria for the etiological classification of primary adrenal insufficiency. J Clin Endocrinol Metab 89:1598 –1604 National Diabetes Data Group 1979 National Diabetes Data Group classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 28:1039 –1057 Moore PS, Dominici P, Borri Voltattorni C 1996 Cloning and expression of pig kidney dopa decarboxylase: comparison of the naturally occurring and recombinant enzymes. Biochem J 315:249 –256 Maras B, Dominici P, Barra D, Bossa F, Voltattorni CB 1991 Pig kidney 3,4-dihydroxyphenylalanine (DOPA) decarboxylase. Primary structure and relationships to other amino acid decarboxylases. Eur J Biochem 201:385–391 ¨ rtqvist E, Persson B, Lernmark Å 1995 Radioimmunoassays Falorni A, O for glutamic acid decarboxylase (GAD65) and GAD65 autoantibodies using 35 3 S or H recombinant human ligands. J Immunol Methods 186:89 –99 Bertoldi M, Moore PS, Maras B, Dominici P, Borri Voltattorni C 1996 Mechanism-based inactivation of DOPA decarboxylase by serotonin. J Biol Chem 271:23954 –23959


1105

35. Dominici P, Moore PS, Borri Voltattorni C 1993 Dissociation, unfolding and refolding trias of pig kidney 3,4-dihydroxyphenylalanine (DOPA) decarboxylase. Biochem J 295(Pt 2):493–500 36. Sherald AF, Sparrow JC, Wright TRF 1973 A spectrophotometric assay for Drosophila DOPA decarboxylase. Anal Biochem 56:300 –305 37. Charteris A, John RA 1975 An investigation of the assay of dopamine using trinitrobenzenesulphonic acid. Anal Biochem 66:365–371 38. Cabrera Rode E, Sarmento L, Tiberti C, Molina G, Barrios J, Hernandez D, Diaz-Horta O, Di Mario U 2003 Type 1 diabetes islet associated antibodies in subjects infected by echovirus 16. Diabetologia 46:1348 –1353 39. Betterle C, Scalici C, Precotto F, Pedini B, Moro L, Rigon F, Mantero F 1988 The natural history of adrenal function in autoimmune patients with adrenal autoantibodies. J Endocrinol 117:467– 475 40. De Bellis A, Bizzarro A, Rossi R, Amoresano Paglionico V, Criscuolo T, Lombardi G, Bellastella A 1993 Remission of subclinical adrenocortical failure in subjects with adrenal autoantibodies. J Clin Endocrinol Metab 76:1002–1007 41. Rorsman F, Husebye ES, Winqvist O, Bjo¨rk E, Karlsson FA, Ka¨mpe O 1995 Aromatic l-amino acid decarboxylase, a pyridoxal phosphate-dependent enzyme, is a novel ␤-cell autoantigen. Proc Natl Acad Sci USA 92:8626 – 8629 42. Burkhard P, Dominici P, Borri Voltattorni C, Jansonius JN, Malashkevich VN 2001 Structural insight into Parkinson’s disease treatment from druginhibited DOPA decarboxylase. Nat Struct Biol 8:963–967 43. Falorni A, Ackefors M, Carlberg C, Daniels T, Persson B, Robertson J, Lernmark Å 1996 Diagnostic sensitivity of immunodominant epitopes of glutamic acid decarboxylase (GAD65) autoantibodies in childhood IDDM. Diabetologia 39:1091–1098 44. Falorni A, Gambelunghe G, Forini F, Kassi G, Cosentino A, Candeloro P, Bolli GB, Brunetti P, Calcinaro F 2000 Autoantibody recognition of COOHterminal epitopes of GAD65 marks the risk for insulin requirement in adultonset diabetes mellitus. J Clin Endocrinol Metab 85:309 –316 ¨ rtqvist E, Hampe 45. Raju R, Foote J, Banga JP, Hall TR, Padoa CJ, Dalakas MC, O CS 2005 Analysis of GAD65 autoantibodies in Stiff-Person syndrome patients. J Immunol 175:7755–7762

JCEM is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.