Characterization of the proteins purified with monoclonal antibodies to glutamic acid decarboxylase

Immunoaffinity columns are prepared from the monoclonal antibody (MAb) GAD-1. These columns are used to enrich glutamic acid decarboxylase (GAD) from the cytosolic fraction of rat brain homogenates and from Triton X-100 extracts of the brain membrane fraction. In each case enzyme activity is enriched over 400-fold. The immunopurified fractions were analyzed by SDS-PAGE. Fractions purified from the cytosol consisted of a quantitatively major band of 59 kDa, and one band of 63 kDa, as well as a group centered around 55 kDa. Fractions purified from membranes consisted primarily of the 59 and 63 kDa components; only traces of the lower-molecular-weight components were present. The entire set of proteins purified on GAD-1 immunoaffinity columns is strongly recognized by 2 widely used antisera to GAD, those described in Saito et al. (1974) and Oertel et al. (1981). The 59 kDa protein from the cytosolic fraction was purified to homogeneity by preparative SDS-PAGE; a partial amino acid sequence of this protein was obtained. The 59 kDa protein has a high degree of sequence homology with the deduced amino acid sequence of the protein that was coded for by a cDNA for feline GAD (Kaufman et al., 1986; Kobayashi et al., 1987). Thus, these proteins are either products of a single gene that diverged during the evolution of rat and cat from a common ancestor, or are members of a closely related set of genes found in both species. The MAb GAD-6 recognizes the 59 kDa band and the group of bands centered around 55 kDa on Western blots. Therefore, these proteins are immunochemically related. GAD-6 does not recognize the 63 kDa band. In Western blots of unfractionated homogenates of the whole brain, the only band recognized by GAD-6 is a 59 kDa band.(ABSTRACT TRUNCATED AT 250 WORDS)     

Colocalization of fixative-modified glutamate and glutaminase but not GAD in rubrospinal neurons

In an attempt to identify putative neurotransmitters of rubrospinal neurons, immunocytochemical procedures were utilized in combination with retrograde tracing techniques in 15 adult male rats. Following injections of horseradish peroxidase (HRP) or wheat germ agglutinin conjugated to HRP (WGA-HRP) into the spinal cord, midbrain sections were processed with a combined procedure that allowed visualization of both the retrograde tracer and one or more antigens including glutamate, glutaminase, and glutamatic acid decarboxylase (GAD). Initial colocalization studies demonstrated that glutamatelike and glutaminaselike immunoreactivities were cocontained within the same neurons. Following injections of HRP or WGA-HRP into the spinal cord approximately 53% of retrogradely labeled neurons contained glutamate immunoreactivity. Triple-labeling experiments indicated that glutamatelike immunoreactivity was colocalized with glutaminase immunoreactivity in retrogradely labeled rubrospinal neurons. Retrogradely labeled neurons did not contain GAD immunoreactivity. Moreover, triple labeling experiments verified that glutamatelike immunoreactive retrogradely labeled cells did not cocontain GAD immunoreactivity. These studies demonstrate that glutamate and its synthesizing enzyme, glutaminase, are present in some rubrospinal neurons and raise the possibility that a component of the rubrospinal projection may be glutamatergic. GAD, on the other hand, is not present in rubrospinal neurons. This finding supports the hypothesis that GABAergic neurons play a role as interneurons in the red nucleus.     

Glutamate decarboxylase (GAD67 and GAD65) gene expression is increased in a subpopulation of neurons in the putamen of parkinsonian monkeys

The cellular distribution of the mRNAs encoding for the two isoforms of glutamate decarboxylase, GAD67 and GAD65, was analyzed by in situ hybridization histochemistry in the caudate nucleus and putamen of control and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-treated parkinsonian squirrel monkeys. On brain sections processed with a radioactive and a digoxigenin-labeled cRNA probe, the GAD67 and GAD65 mRNAs were colocalized in virtually all labeled neurons of the caudate nucleus and putamen, in both control and MPTP-treated monkeys. Furthermore, neurons labeled with the GAD cRNAs constituted at least 90% of all striatal neurons, as estimated on adjacent Nissl-stained sections. In the two groups of monkeys, double-labeling experiments using a combination of radioactive GAD67 or GAD65 and digoxigenin-labeled preproenkephalin (PPE) cRNA probes showed that roughly half of all neurons labeled with the GAD cRNAs were also labeled with the PPE cRNA probe. When compared to controls, GAD67 and GAD65 mRNA levels were higher in the putamen, and to a lesser extent in the caudate nucleus, of MPTP-treated monkeys. Further analysis of labeling at the cellular level in a dorsolateral sector of the putamen revealed that GAD67 and GAD65 mRNA levels in MPTP-treated monkeys were increased in PPE-labeled (presumed striato-pallidal) neurons but not in PPE-unlabeled (presumed striato-nigral) neurons. Our results demonstrate that most neurons in the caudate nucleus and putamen of squirrel monkeys contain the mRNAs encoding for the two GAD isoforms. In addition, the selective increase in GAD mRNA levels in PPE-labeled neurons provides further evidence that striato-pallidal GABAergic neurons are hyperactive in MPTP-treated parkinsonian monkeys. Synapse 27:122–132, 1997. © 1997 Wiley-Liss, Inc.     

Quantitative maps of GAbAergic and glutamatergic neuronal systems in the human brain

GABAergic and glutamatergic neuronal systems in adult normal human brains were shown quantitatively and in detail through the distributions of glutamate decarboxylase (GAD) and glutamate dehydrogenase (GDH), respectively. Consecutive coronal sections containing part of the striatum and the substantia nigra were obtained from the right hemisphere of three deceased persons with no history of neurological or psychiatric diseases and were stained immunohistochemically for GAD and GDH. Each stained section was divided into approximately 3 million microareas and the immunohistochemical fluorescence intensity in each area was measured by a human brain mapping analyzer, which is a microphotometry system for analysis of the distribution of neurochemicals in a large tissue slice. In the analyzed brain regions, conspicuously intense GAD-like immunoreactivity was observed in the substantia nigra, globus pallidus, and hypothalamus. GDH was widely and rather evenly distributed in the gray matter compared to GAD, although intense GDH-like immunoreactivity was observed in the lateral geniculate nucleus and substantia nigra. Within the substantia nigra, the globus pallidus, and other regions, characteristic distributions of GAD- and GDH-like immunoreactivity were found. We believe that the analysis of the human brain by this novel technique can help to understand the functional distribution of neuronal systems in the normal human brain and may be able to identify abnormal changes in the diseased human brain. It can also provide basic data to help in the interpretation of functional magnetic resonance imaging or positron emission tomography.     

Retinal Ganglion Cell Depletion Alters the Phenotypic Expression of GABA and GAD in the Rat Retina

We have looked at the phenotypic expression of γ-aminobutyric acid (GABA) and the two isoforms of its synthetic enzyme [glutamic acid decarboxylase (GAD)-65 and -671 in adult rat retinas that had the superior colliculus, pretectum and optic tract lesioned unilaterally at birth. It has been shown previously that this type of manipulation induces retrograde degeneration of retinal ganglion cells presumably without affecting other intraretinal neurons. We present evidence that GABAergic amacrine cells are affected by such manipulation. The number of cells immunoreactive for GABA, GAD-65 and GAD-67 decreased in the inner nuclear layer. In the retinal ganglion cell layer, however, the number of GABA- and GAD-65–labelled cells increased, while the number of GAD-67–labelled cells did not change. Biochemical assay showed that overall GAD activity was not altered in retinas of lesioned animals. Our results support the notion that, while neonatal lesion reorganizes the expression of GABA and GAD in the retina, enzyme activity is maintained within normal levels.     

A novel monoclonal antibody specific for the N-terminal end of GAD65

We describe a novel monoclonal antibody raised towards the N-terminus of the 65 kDa isoform of glutamate decarboxylase (GAD65). This N-GAD65 mAb is highly specific for GAD65 and can be used in a wide range of applications such as Western blot analysis, immunoprecipitation and immunocytochemistry. Full-length cDNAs coding for N-GAD65 mAb heavy and light chains were cloned and characterized by nucleotide sequence analysis. Both heavy and light chain-specific cDNAs are functional and are members of mouse heavy chain subgroup IgG1 and kappa chain group 2, respectively. Comparing the N-GAD65 mAb gene sequences with GenBank shows that the cDNAs were not reported previously.     

GABAergic input to the synaptic terminals of mb1 bipolar cells in the goldfish retina

An EM-autoradiographical/immunocytochemical technique was used to study amacrine cell synapses onto mb1 bipolar cell terminals in goldfish retina. Tissue was double labeled for [3H]GABA uptake and glutamate decarboxylase (GAD) immunolocalization. Nearly 90% of the amacrine cell synaptic processes onto both proximal and distal halves of mb1 terminals were labeled with either [3H]GABA or GAD-immunoreactivity (IR). Proximal half: 73% of the amacrine synapses were labeled with [3H]GABA uptake and 82% with GAD-IR; 88% of [3H]GABA labeled contacts were double labeled. Distal half: 17% of the amacrine synapses were labeled with [3H]GABA uptake and 67% with GAD-IR; 63% of [3H]GABA labeled contacts were double labeled. After consideration of the possible sources of [3H]GABA labeled synapses onto mb1 terminals, we concluded that the synaptic terminals of pyriform Ab amacrine cells double label for [3H]GABA and GAD-IR despite our previous report that Ab cell bodies do not stain for anti-catfish brain GAD antiserum. We suggest that Ab cells contain isoenzymes of GAD which differ in subcellular distribution, thereby accounting for the differential staining of the cell bodies and dendrites obtained with the GAD antiserum we used.     

GAD67 and GAD65 mRNA and protein expression in cerebrocortical regions of elderly patients with schizophrenia

Gamma-Aminobutyric acid (GABA), the principal inhibitory neurotransmitter of CNS, has been consistently implicated in the pathophysiology of schizophrenia. GABA is synthesized from glutamate by the enzyme glutamic acid decarboxylase (GAD). Two isoforms of GAD have been identified and have been named GAD65 and GAD67 based on their apparent molecular weights. In this study, GAD65 and GAD67 mRNA and protein levels were measured by using real-time RT-PCR and immunoblotting, respectively, in post-mortem brain tissue from the dorsolateral prefrontal cortex (DLPFC) and the occipital cortex of the elderly persons with schizophrenia and matched normal controls. In addition, the mRNA expression of GAT-1, one of the principal transporters of GABA, was also studied in the same subjects. Expression of GAD65 and GAD67 mRNA in the DLPFC and in the occipital cortex was significantly elevated in patients with schizophrenia, whereas the expression of the corresponding proteins and GAT-1 mRNA was unchanged. Although the levels of GAD65 and GAD67 messages were increased in schizophrenia subjects, the proportion of the two GAD isoforms remained constant in controls and schizophrenics. In the human DLPFC, GAD65 mRNA was found to be expressed significantly less than the message for GAD67, approximately 16% of that observed for GAD67. On the contrary, the abundance of GAD65 protein in the DLPFC was about 350% of that observed for GAD67. The results suggest a substantial dysregulation of GAD mRNA expression in schizophrenia and, taken together with the results of protein expression studies, raise the possibility that both cortical and subcortical GABA function may be compromised in the disease.     

Proline, glutamate and glutamine metabolism in mouse brain synaptosomes

In nerve terminals, glutamate (Glu) may serve as precursor of the inhibitory neurotransmitter, GABA, and the putative excitatory transmitter, aspartate (Asp), in addition to exerting its own excitatory neurotransmitter role in brain. Glu carbon can originate from glucose through glycolysis and the Krebs cycle, from glutamine (Gln) subsequent to uptake, and from proline (Pro) and ornithine (Orn). Orn, but not Glu, is an effective precursor in nerve terminals of Pro, a putative inhibitory neurotransmitter. [3H]Arg can be converted in mouse brain nerve terminals to Orn, which in turn gives rise to Glu, Pro and GABA. In the present study, the conversion subsequent to uptake of labeled Glu, Gln and Pro to other amino acids was studied in unfractionated and subfractionated synaptosomal particles which layered, respectively, on 1.0 M, 1.2 M, 1.3 M and 1.5 M sucrose after centrifugation in a discontinuous gradient (fractions 1-4, respectively). Fraction 1 contained small synaptosomal fragments with vesicles and almost no mitochondria. Fractions 2 and 3 showed numerous normal-appearing mitochondria-containing synaptosomes, and fraction 4 contained large synaptosomes and more free mitochondria than the other fractions. Glu was readily taken up in all fractions and converted to Asp, Gln and GABA, the greatest formation of Asp from Glu occurring in fractions 2 and 3 and of Gln in fraction 4. In contrast, Gln was taken up poorly in fraction 1 and not metabolized, converted extensively to Glu and GABA in fractions 2-4, giving rise only to very small amounts of Asp in fractions 2 and 3. Although Pro was taken up to the greatest extent in fraction 2, it was by far most readily converted to Glu, Gln and GABA in fraction 1, showing only small amounts of Asp formation in fractions 1-3 and none in 4. There was no significant production of Pro from Glu or Gln or of Arg and Orn from any of the 3 precursors studied. The above results suggest that Glu, Gln and Pro may be taken up largely in different classes of synaptosomes which are distributed among the centrifugally separated fractions and which possess differing transport and metabolic characteristics. Determination of glutamate decarboxylase activity (GAD) indicated that GABA-forming nerve terminals were present in all synaptosomal fractions studied. Amino acid determinations by HPLC in the subfractionated synaptosomes showed a similar distribution for Glu, Asp and GABA contents, peaking in fraction 2, and an inverse relationship of the latter 3 with Arg contents.(ABSTRACT TRUNCATED AT 400 WORDS)     

Comparison of the expression of two forms of glutamic acid decarboxylase (GAD67 and GAD65) in the visual cortex of normal and dark-reared cats

In normal development, there are dramatic changes in both the level and the laminar pattern of expression of the two forms of glutamate decarboxylase (GAD67, GAD65), the synthetic enzyme for gamma-aminobutyric acid (GABA). We have used antibodies to determine whether these normal postnatal changes in the expression of the two GADs depend on visual input by comparing normal and dark-reared cat visual cortex. Western blot analysis showed no significant differences in the levels of expression of the two enzymes between rearing conditions at either 5 or 20 weeks. Immunohistochemistry was used to compare the laminar distribution of the GADs in the two rearing conditions. At 1 week of age, both GAD67 and GAD65 immunoreactivity is concentrated in deep layers of visual cortex. At 5 and 20 weeks in both rearing conditions, GAD67-stained cells bodies were distributed rather uniformly across all cortical layers. GAD65 primarily labeled puncta (synaptic terminals) and these were also distributed rather uniformly across all visual cortical layers in both rearing conditions. Counts of GAD67-positive cell bodies and GAD65-positive puncta also revealed no differences between the rearing conditions. Thus, both GAD67, which produces the basal pool of GABA, and GAD65, which is specialized to respond to short-term increases in demand in synaptic terminals, developed normal levels of expression and normal intracellular and laminar distributions in the absence of visual input. Physiological studies suggest immaturity in the GABA system of dark-reared visual cortex. The present results indicate that such abnormalities are not due to presynaptic alterations in GABA synthetic enzymes.     

Elevation of brain GABA levels with vigabatrin (γ-vinylGABA) differentially affects GAD65 and GAD67 expression in various regions of rat brain

Previous studies have demonstrated that the expression of one of the isoforms of glutamate decarboxylase, GAD₆₇, is selectively reduced in cultured cortical neurons and in rat cerebral cortex when the concentration of GABA is elevated. We asked whether the expression of GAD₆₇ was similarly affected by elevated GABA throughout the brain. The concentration of GABA in rat brain was increased by inhibiting GABA transaminase (GABA-T) with vigabatrin (γ-vinylGABA, GVG), an antiepileptic drug and selective inhibitor of GABA-T. Rats were injected with saline or vigabatrin (150 mg/kg) daily for 5 days, and the effects of accumulated GABA on total GAD activity and the expression of GAD₆₅ and GAD₆₇ proteins were determined in twelve brain regions. The GABA concentration was significantly elevated in all regions except amygdala and olfactory bulb after vigabatrin treatment. Total GAD activity was significantly lower than controls in six regions: cerebellum, frontal cortex, thalamus, substantia nigra, ventral tegmentum, and the remaining midbrain. The decrease in GAD activity was largest in cerebellum and thalamus (33% and 29%), while the changes in the other four areas were 15–18%. Vigabatrin treatment significantly reduced GAD₆₇ protein in all regions except olfactory bulb, whereas GAD₆₅ protein decreased significantly only in cerebellum. The failure to detect significant changes in GAD activity in regions having a significant change in GAD₆₇ levels is attributable to the small contribution of GAD₆₇ to total GAD in those regions. It is evident that there are marked regional differences in the effects of tissue GABA levels on the expression of GAD₆₇. J. Neurosci. Res. 52:736–741, 1998. © 1998 Wiley-Liss, Inc.     

Glutamate alteration of glutamic acid decarboxylase (GAD) in GABAergic neurons: the role of cysteine proteases

Brain cell vulnerability to neurologic insults varies greatly, depending on their neuronal subpopulation. Among cells that survive a pathological insult such as ischemia or brain trauma, some may undergo morphological and/or biochemical changes that could compromise brain function. We previously reported that surviving cortical GABAergic neurons exposed to glutamate in vitro displayed an NMDA receptor (NMDAR)-mediated alteration in the levels of the GABA synthesizing enzyme glutamic acid decarboxylase (GAD65/67) [Monnerie, H., Le Roux, P., 2007. Reduced dendrite growth and altered glutamic acid decarboxylase (GAD) 65- and 67-kDa isoform protein expression from mouse cortical GABAergic neurons following excitotoxic injury in vitro. Exp. Neurol. 205, 367–382]. In this study, we examined the mechanisms by which glutamate excitotoxicity caused a change in cortical GABAergic neurons' GAD protein levels. Removing extracellular calcium prevented the NMDAR-mediated decrease in GAD protein levels, measured using Western blot techniques, whereas inhibiting calcium entry through voltage-gated calcium channels had no effect. Glutamate's effect on GAD protein isoforms was significantly attenuated by preincubation with the cysteine protease inhibitor N-Acetyl-l-Leucyl-l-Leucyl-l-norleucinal (ALLN). Using class-specific protease inhibitors, we observed that ALLN's effect resulted from the blockade of calpain and cathepsin protease activities. Cell-free proteolysis assay confirmed that both proteases were involved in glutamate-induced alteration in GAD protein levels. Together these results suggest that glutamate-induced excitotoxic stimulation of NMDAR in cultured cortical neurons leads to altered GAD protein levels from GABAergic neurons through intracellular calcium increase and protease activation including calpain and cathepsin. Biochemical alterations in surviving cortical GABAergic neurons in various disease states may contribute to the altered balance between excitation and inhibition that is often observed after injury.     

Expression pattern of glutamate decarboxylase (GAD) in the developing cortex of the embryonic chick brain

The development of the GABAergic system in the chick embryo telencephalon has been studied. Special emphasis was placed on the development of glutamate decarboxylase (GAD) between embryonic day 8 (E8) and E17. The GABA immunoreactivity and neuron-specific enolase expression was detected simultaneously in glutardialdehyde fixed sections, which confirmed that GABAergic cells exhibit neuronal phenotype. The GAD expression was studied by means of immunohistochemistry on cryo-sectioned material both at the light and electron microscopic levels. Furthermore, the presence and localization of GAD65 and GAD67 mRNAs were studied with an in situ hybridization technique with digoxigenin-labeled RNA probes. Protein expression as well as mRNA appearance mostly coincided both temporally and spatially.In the parahippocampal area, as well as in other regions of the developing cortex, GAD staining was seen from E8 onwards. The number of positive cells increased as did the intensity of staining up to E14. As observed in the electron microscope, the GAD protein was co-localized with GABA in most cases, although some GAD-positive cells devoid of GABA-staining also were observed. The pattern of GAD mRNA expression was in general similar to that of GAD immunostaining. Both GAD65 and GAD67 mRNA were detected during the entire period. Furthermore, GAD67 mRNA localization spatially was more correlated with GAD protein expression. The study provides evidence for the notion that development of the GABAergic system occurs rapidly during embryogenesis and, as suggested from mRNA data, that two forms of GAD with a slight difference in distribution can contribute to this. © 1997 ISDN     

GAD65 antibody‐associated autoimmune epilepsy with unique independent bitemporal‐onset ictal asystole

Antibodies against the 65‐kDa isoform of the intracellular enzyme, glutamate decarboxylase (GAD65), have been found in patients with limbic encephalitis and drug‐resistant autoimmune epilepsy. We report a 22‐year‐old female who presented with new‐onset seizures and neuropsychiatric symptoms. Video‐EEG captured unique, independent bitemporal‐onset focal seizures with impaired awareness and ictal asystole. An autoimmune epilepsy panel revealed elevated GAD65 antibodies in the serum (225 nmol/l) and CSF (2.78 nmol/l), while [¹⁸F]‐fluoro‐deoxy‐glucose positron emission tomography showed bitemporal hypometabolism (left > right). The patient was diagnosed with GAD65 antibody‐associated autoimmune epilepsy. Our observation adds to the spectrum of neurocardiac syndromes associated with autoimmune epilepsy.     

GABAergic dysfunction in schizophrenia and mood disorders as reflected by decreased levels of glutamic acid decarboxylase 65 and 67 kDa and Reelin proteins in cerebellum

BACKGROUND: Glutamic acid decarboxylase (GAD) is the rate limiting enzyme responsible for conversion of glutamate to gamma-aminobutyric acid (GABA) regulating levels of glutamate and GABA in the mammalian brain. Reelin is an extracellular matrix protein that helps in normal lamination of the embryonic brain and subserves synaptic plasticity in adult brain. Both GAD and Reelin are colocalized to the same GABAergic interneurons in several brain sites. We hypothesized that levels of GAD and Reelin would be altered in cerebellum of subjects with schizophrenia and mood disorders differentially vs. controls. METHODS: We employed SDS-PAGE and Western blotting to measure levels of GAD isomers 65 and 67 kDa and Reelin isoforms 410-, 330- and 180-kDa proteins as well as beta-actin in cerebellum of subjects with schizophrenia, bipolar disorder and major depression vs. controls (N = 15 per group). RESULTS: GAD 65- and 67-kDa levels were decreased significantly in bipolar, depressed and schizophrenic subjects (p < 0.05) vs. controls. Reelin 410- and 180-kDa proteins decreased significantly (p < 0.05) in bipolar subjects vs. controls. Reelin 180 kDa was decreased significantly (p < 0.05) in schizophrenics vs. controls. beta-Actin levels did not vary significantly between groups. There were no significant effects of confounding variables on levels of various proteins. CONCLUSION: This study demonstrates for the first time significant deficits in GABAergic markers Reelin and GAD 65 and 67 proteins in bipolar subjects and global deficits in the latter proteins in schizophrenia and mood disorders, accounting for the reported alterations in CSF/plasma levels of glutamate and GABA in these disorders.     

Knockout of GAD65 has major impact on synaptic GABA synthesized from astrocyte-derived glutamine

γ-Aminobutyric acid (GABA) synthesis from glutamate is catalyzed by glutamate decarboxylase (GAD) of which two isoforms, GAD65 and GAD67, have been identified. The GAD65 has repeatedly been shown to be important during intensified synaptic activity. To specifically elucidate the significance of GAD65 for maintenance of the highly compartmentalized intracellular and intercellular GABA homeostasis, GAD65 knockout and corresponding wild-type mice were injected with [1-¹³C]glucose and the astrocyte-specific substrate [1,2-¹³C]acetate. Synthesis of GABA from glutamine in the GABAergic synapses was further investigated in GAD65 knockout and wild-type mice using [1,2-¹³C]acetate and in some cases γ-vinylGABA (GVG, Vigabatrin), an inhibitor of GABA degradation. A detailed metabolic mapping was obtained by nuclear magnetic resonance (NMR) spectroscopic analysis of tissue extracts of cerebral cortex and hippocampus. The GABA content in both brain regions was reduced by ∼20%. Moreover, it was revealed that GAD65 is crucial for maintenance of biosynthesis of synaptic GABA particularly by direct synthesis from astrocytic glutamine via glutamate. The GAD67 was found to be important for synthesis of GABA from glutamine both via direct synthesis and via a pathway involving mitochondrial metabolism. Furthermore, a severe neuronal hypometabolism, involving glycolysis and tricarboxylic acid (TCA) cycle activity, was observed in cerebral cortex of GAD65 knockout mice.     

Expression of glutamic acid decarboxylase in nervous tissue structures targeted by autoantibodies in patients with diabetic autonomic neuropathy

We have previously identified an association between symptomatic diabetic autonomic neuropathy (DAN) and autoantibodies to sympathetic and parasympathetic nervous structures. The antigens identified by these autoantibodies are not known, but glutamic acid decarboxylase (GAD) has been suggested as a candidate target, since anti-GAD autoantibodies are present in patients with long-term diabetes and GAD is expressed in a variety of cell types and structures in the nervous system. The aim of this study was to examine GAD expression in sympathetic ganglia and vagus nerve and to compare the distribution of GAD within these tissues with that of anti-sympathetic ganglia and anti-vagus nerve autoantibodies from patients with DAN, using single and double indirect immunofluorescence on tissue sections. The monoclonal antibody GAD-6, specific for GAD₆₅, gave a granular, peripheral, cytoplasmic staining pattern in sympathetic ganglion cells. Dual immunofluorescence demonstrated that serum from a patient with anti-sympathetic ganglion autoantibodies stained the same cells, but homogeneously throughout the cytoplasm. In the vagus nerve, patient's serum stained the fibres only; GAD-6 stained the cytoplasm of parasympathetic ganglion cells but only occasional fibres. In addition, GAD enzymatic activity was detectable in both sympathetic ganglia and vagus nerve. Incubation of sera or GAD-6 overnight with a crude homogenate of human brain as an antigen source abolished staining of the nervous tissues by GAD-6, but not by patients' sera. The different localisation of GAD and the autoantigens targeted by patients' sera indicates that GAD is not the target of the autoantibodies characteristic of DAN. Moreover, absorption studies using human brain homogenate suggest that the targets of anti-sympathetic ganglion and anti-vagus nerve autoantibodies are absent or represented only at low levels in the central nervous system and may be confined to the periphery.