Conceived and designed the experiments: CG KVT MH CS. Performed the experiments: CG AW BG TW MG RD DP. Analyzed the data: CG MS MH BG KVT CS AR RD DP. Contributed reagents/materials/analysis tools: FF DP MG. Wrote the paper: CG KVT FF CS.
The authors have declared that no competing interests exist.
Anxiety is a heterogeneous behavioral domain playing a role in a variety of neuropsychiatric diseases. While anxiety is the cardinal symptom in disorders such as panic disorder, co-morbid anxious behavior can occur in a variety of diseases. Stiff person syndrome (SPS) is a CNS disorder characterized by increased muscle tone and prominent agoraphobia and anxiety. Most patients have high-titer antibodies against glutamate decarboxylase (GAD) 65. The pathogenic role of these autoantibodies is unclear.
We re-investigated a 53 year old woman with SPS and profound anxiety for GABA-A receptor binding in the amygdala with (11)C-flumazenil PET scan and studied the potential pathogenic role of purified IgG from her plasma filtrates containing high-titer antibodies against GAD 65. We passively transferred the IgG fraction intrathecally into rats and analyzed the effects using behavioral and
The observations in rats after passive transfer lead us to propose that anxiety-like behavior can be induced in rats by passive transfer of IgG from a SPS patient positive for anti-GAD 65 antibodies. Anxiety, in this case, thus may be an antibody-mediated phenomenon with consecutive disturbance of GABAergic signaling in the amygdala region.
Anxiety and fear are the leading symptoms in anxiety disorders such as panic disorder and phobias, which are thought to feature complex neurobiological underpinnings with both genetic as well as environmental factors. Cortico-limbic pathways and GABAergic signaling are thought to be key components of anxiety disorders, yet the precise molecular mechanisms are still unknown. In addition to anxiety disorders in a narrower sense, anxious behavior can frequently be found in a wide range of neuropsychiatric diseases. Although the pathomechanisms are even less clear in these disorders, it can be supposed that disease mechanisms overlap and a common final pathway may exist. Stiff person syndrome (SPS), a rare and multi-facetted disorder of the central nervous system, is one of the neuropsychiatric disorders where anxious symptoms are found most frequently
It is a matter of debate whether anxiety and agoraphobia may be secondary to the motor instability caused by the enhanced startle response associated with uncontrolled drop attacks, or if these are additional and autonomous symptoms reflecting central GABAergic dysfunction
The PET scan of a 53-year-old woman with SPS reported earlier as a clinical case
This figure shows the results of the statistical parametric mapping (SPM) with voxel-based analysis and thus the comparison of GABA-A receptor binding potential in the patient when compared to normal controls. Note that the (11)C-FMZ binding potential is significantly reduced bilaterally in the patient's amygdala region.
test | test result SPS patient | interpretation |
HARS | 16 | (0–56)significant anxiety (>14) |
HAM-D | 0 | (0–66)mild depression (>18) |
MADRS | 1 | (0–60)mild depression (>9–18) |
SAS | 51 | (0–80)moderate anxiety level (45–59) |
SCL-90 | anxiety: 28 | (item 5: 0–40)significant anxiety: (>20) |
MMSE | 30 | (0–30)no cognitive impairment |
Neuropsychological evaluation of the SPS patient
Because the patient was clinically affected by profound anxiety, agoraphobia, and panic attacks, we focused on anxiety-like behavior in experimental rats receiving repetitive i.th. injections of purified patient IgG. Indeed, in the elevated plus maze (EPM), rats treated with GAD 65 antibody positive IgG showed pronounced anxiety-like behavior. The animals preferred to stay in the closed arms of the maze, spent less then 1% of the testing time in the open arms and made significantly less entries in either arm, when compared to controls (
(A and B) Animals after intrathecal passive transfer of IgG were tested on an elevated plus maze (EPM) to analyze anxiety-like behavior. Rats treated with SPS patient IgG (SPS IgG; n = 6) moved normally but spent significantly less time in the open arms and explored significantly less arms of the EPM in testing period compared to controls (control IgG: n = 9; saline n = 7), indicating increased anxiety-like behavior (* p<0.05; Mann-Whitney Test). (C and D) Locomotor activity, assessed as time spent in the periphery of the open field (OF) and total distance moved during the observation period, was not different between the experimental groups and did not contribute to the behavioral observations in the EPM. (E) Representative tracks of a rat treated with saline (left panel) and SPS IgG (right panel) in the EPM. Whereas the control animal explored 3 arms of the EPM including one open arm with bright illumination, the rat treated with SPS IgG stayed in the closed, darker arms and avoided entries into the open arms.
In contrast to observations made in an acute
Western blotting with homogenized rat CNS tissue (
(A) Incubation of naïve rat brain sections with purified patient SPS IgG resulted in distinct labeling of different brain areas, such as the rostral and temporobasal part of the cortex (CTXro, CTXtb), hippocampus (HC, CA1), nucleus paraventricularis thalami (PVT), lateral and basolateral amygdala (LA, BLA), whereas in other areas, e.g. most of the thalamus, parts of the cortex and midbrain regions, staining was absent or less pronounced (scale bar: 2.0 mm). (B) At higher magnification, the staining pattern differed within the labeled brain regions. In the HC strongest immunoreactivity was detected in the CA1 region just below the pyramidal cell layer with clear labeling of single neurons resembling GABAergic basket neurons. In the immunoreactive areas of the cortex (temporobasal part: CTXtb), labeling of single neurons with extensive staining of the dendrites was observed. The strongest immunoreactivity was found in the region of the amygdala nuclei, most pronounced in the lateral and basolateral parts (basolateral part: BLA) with dense reticular staining and strongly immunoreactive small cell bodies showing a dense network within the amygdala nuclei. No specific staining was detectable after incubation with control IgG (scale bar: 25 µm). (C) Western blotting of patients purified IgG on rat CNS tissue revealed a single band at 65 kDa (lane a) while the characteristic double band at 65 and 67 kDa was seen when a commercial polyclonal GAD 65/67 antibody was used (lane b). (D) Western blotting of patient purified IgG (lane a) and a rabbit polyclonal GAD 65 antibody (lane b) on a rat GAD 65 fusion protein displayed specific binding with a single band at the expected molecular weight of 90 kDa.
Tissue of normal human cerebellum, amygdala, frontal cortex, and spinal cord was immunoreacted with purified control IgG or IgG preparations containing high titer of GAD 65 antibodies (scale bar 200 µm). Incubation with control IgG (A) resulted in unspecific perivascular staining (arrowheads), whereas SPS IgG (B) gave intense labeling of the molecular (arrows) and granular layer (arrowheads) in the cerebellum, of the amygdala core region, and the grey matter of the frontal cortex (arrows). SPS IgG immunoreactivity in the ventral horn (VH) of the spinal cord (arrowheads) was less pronounced as compared to the highly immunoreactive brain regions, particularly the amygdala and frontal cortex. (C) Higher magnification revealed (1) strongly positive staining of GABAergic basket cell fibers around Purkinje cells in the cerebellum (thin arrows), (2) a very intense staining of the densely packed reticular network in the amygdala and some small sized interneurons (arrowheads), (3) positive immunoreactivity of single interneurons (thick arrows) in the deeper layers of the frontal cortex, and (4) less strong staining of perineuronal dendrites and proximal dendrites of a motor neuron (open arrowheads) in the ventral horn of the spinal cord (scale bar: 25 µm).
When brain sections of the intrathecally treated rats were stained for human IgG, strongly immunoreactive neurons were found in the basolateral amygdala, suggesting neuronal uptake of patient IgG in the amygdala complex. These findings also demonstrate that the i.th. injected IgG had indeed access to the CNS hemispheres. No immunoreactivity was found in the animals treated with control IgG (
In rats treated intrathecally with SPS IgG (A) but not in those treated with control IgG (B), immunoreaction against human IgG showed positive staining of neurons in the amygdala complex (scale bar: 10 µm). (C) Dissociated hippocampal neurons from E18 mouse embryos were incubated with SPS or control IgG. The increase of GABA release into the supernatants induced by stimulation with 90 mmol KCl was absent in SPS IgG treated neuronal cell cultures as demonstrated by HPLC analysis (* p<0.05, Student's t-test).
To assess a direct effect of patient IgG on GABAergic transmission, we measured GABA release in the culture supernatant of cultured hippocampal neurons with high performance liquid chromatography (HPLC). Incubation with patient SPS IgG resulted in a significant decrease of GABA release after stimulation with 90 mmol KCl compared to incubation with control IgG (
The principal finding of our study is that one of the core symptoms – namely, anxiety – of the afflicted SPS patient could be reproduced in the recipient rat by passively transferring her GAD 65 antibody-containing IgG into the subarachnoid space. This points to a probable pathogenic action of IgG autoantibodies. Furthermore, application of IgG at the caudal spinal cord lead to distant IgG binding at the amygdala that is likely linked to the observed behavioral alterations.
How do our findings relate to the neuropsychiatric observations in the patient? The extension of the PET study of the SPS patient, showing a reduced 11C-FMZ binding potential in limbic structures, supports the concept that the GABAergic system in the amygdala region is affected by the disease, and this is consistent with a significant reduction in GABA-A receptor binding observed in patients with panic disorder and other anxiety-related disorders
Based on our passive transfer experiments we like to propose that the neuropsychiatric changes may be a direct consequence of circulating IgG autoantibodies to GAD 65 although it cannot be excluded that autoantibodies with other as yet unknown specificities are instrumental in causing the behavioral abnormalities and IgG binding in recipient rats (see below).
The GABAergic involvement of the limbic cortex shown here might be due to the high demand of modulatory control in these regions, based on their finely tuned GABAergic transmission
In contrast to a recent report on the effects of GAD 65 antibody containing IgG in an acute ex-vivo preparation
One question that still needs to be solved is whether the anti-GAD 65 antibodies in the patient's purified IgG fraction are directly responsible for the observed effects, or whether antibodies directed against yet unidentified surface antigens of GABAergic neurons might play a role. Using western blotting, we could not detect any reactivity towards antigens other than GAD65. However, as in the example of GABA-A-receptor-associated protein (GABARAP), identified as an additional antigen in patients with anti-GAD-positive SPS
Recently a new family of paraneoplastic and idiopathic CNS disorders with neuropsychiatric symptoms, namely limbic encephalitis with psychiatric features, SPS, hyperexplexia, epilepsy and various other optional features has been identified to be linked to autoantibodies directed at CNS autoantigens including NMDA-receptors, AMPA-receptors, Caspr2 and LGI-1; some of these were formerly classified as voltage-gated potassium channelopathies
Further studies will be necessary to confirm our findings in a larger number of patients. Another aim for follow-up studies is to identify the exact target of the SPS antibodies responsible for these observations. Taken further, our findings may imply that autoantibodies should be searched for in other anxiety related disorders, and, if detected, this may open new diagnostic and therapeutic options for patients with anxiety syndromes.
The experimental study was performed with purified IgG containing antibodies to GAD 65 obtained from a 53-year-old woman with SPS, who first presented with signs of muscle rigidity and spasms at age 52. Besides the typical motor symptoms, she suffered from severe anxiety, panic attacks, and agoraphobia
Patient IgG and IgG from a control patient with chronic inflammatory polyneuropathy was purified from plasma filtrate obtained at therapeutic plasma exchange as a part of standard patient care by separation on exchange chromatography as described
The GAD 65 gene was sub-cloned into a pGEX-6P expression vector system (GE Healthcare, Munich, Germany) to generate a glutathione S-transferase (GST) fusion protein, which was transformed into DH5alpha cells. For Western blot analysis, 10–50 µg GAD 65 fusion protein was resuspended in 30 µl Laemmli sample buffer, fractionated by 10% SDS-PAGE and electroblotted onto a nitrocellulose membrane. Nitrocellulose was incubated with a polyclonal rabbit anti-GAD 65 antibody (1∶500, Acris, Hiddenhausen, Germany), a polyclonal anti-GAD 65/67 antibody (1∶500, Chemicon, Millipore, MA, USA) and patient IgG (100mg/ml, 1∶500). The proteins were detected by chemiluminescence with Supersignal West Pico Substrate (PIERCE Biotechnology, Rockford, IL, USA) using a LAS-3000 Bioimaging System (Fuji, Duesseldorf, Germany).
The patients reported in this study have provided written consent for the publication of the clinical details and for utilization of their plasma exchange material for experimental studies. The study was approved by the ethics commission of the Medical Faculty of the University of Würzburg (02/06; January, 19th 2006).
Animal use and care were in accordance with the institutional guidelines. All animal experiments were approved by the Bavarian State authorities (Regierung von Unterfranken, # 55.5-2531.01-78/05 and # 55.5-2531.01-12/10).
Twenty-two female Lewis rats were used (6–8 weeks old, purchased from Harlan-Winkelmann, Borchen, Germany). Intrathecal catheters (0.28-mm inner diameter; 0.61-mm outer diameter; intrathecal length: 7.0 cm), were placed in the subarachnoid space following the method of Yaksh and Rudy
Behavioral analyses were done by the same group of investigators with longstanding experience in rat behavioral studies; these were kept blinded as to treatment allocation. Animals were observed daily in their cages and on a plane surface with tunnels and obstacles. Rats were trained on an accelerating RotaRod (TSE Systems, Bad Homburg, Germany) and quantitative testing was performed after recovery from surgery and before the 1st injection (baseline) and on day 4 and 19 after starting IgG injections. Forelimb grip strength was tested with a digital grip force meter (Chatillon, Greensboro, NC, USA). Gait analysis
H-reflex testing: A total of 22 rats were used for these experiments. Under anesthesia with i.p. injections of ketamin and xylazin (80–100 mg/kg and 5 mg/kg, respectively), H-Reflex recording was performed as described previously
Dorsal root potential (DRP) recordings: lumbar laminectomy was performed to expose spinal segments TH12-L5, and DRP were recorded from dorsal roots L4 and 5 on both sides with an ELC-03X amplifier (npi electronic, Tamm, Germany) after stimulation of the tibial nerve using a Grass S88 stimulator (duration 0.2 ms, single stimulation and train of 3 stimuli at 100 Hz; Grass technologies, RI, USA) as described previously
At the end of the experiments, rats were deeply anesthetized with intraperitoneal injections of ketamin and xylazin (80–100 mg/kg and 5 mg/kg, respectively) and blood was withdrawn. Thereafter, the rats were sacrificed, and the lumbar spinal cord and the brain were mounted in Tissue-Tec OTC embedding compound, and deep frozen. Twenty-µm cryosections of the brain hemispheres were cut and serial hematoxylin and eosin staining was performed to identify the brain regions with the structures of interest according to a rat brain atlas
To test the binding properties of the patient plasma filtrate and purified IgG, we used 20-µm frozen sections of brain and lumbar spinal cord of naive rats and, in addition, 20-µm frozen sections from frontal cortex, amygdala, and spinal cord of human control autopsy material without CNS disease. Sections were incubated with patient plasma exchange material in concentrations ranging from 1∶100 to 1∶10.000 and with purified patient IgG in concentrations of 100 µg/ml and 10 µg/ml over night as the primary antibody, followed by rabbit anti-human IgG (Dako) at 1∶200 for 30 minutes and visualized with diaminobenzidine.
Besides binding we were interested in potential cell destruction, complement activation and immune cell infiltration. We incubated 20-µm frozen brain sections at 4°C over night with a rabbit monoclonal anti-capase-3 antibody (BD Biosciences, San Jose, USA, 1∶200), a rabbit polyclonal anti-C5b-9 antibody (abcam, Cambridge, UK, 1∶1000), a mouse monoclonal anti-rat CD68 antibody (Linaris, Wertheim-Bettingen, Germany, 1∶000) and a mouse monoclonal anti-CD3 antibody (BD Biosciences, San Jose, USA, 1∶200). Immunoreactions were then visualized with diaminobenzidine.
Hippocampal cell cultures were prepared from E18 mouse embryos as described previously
Statistical analyses were done using SPSS software (SSPS Inc.). The data represent means +/− SEM. The significance of differences in means was tested using students't-test or Mann-Whitney-U-test depending on the distribution of data.
(TIF)
We thank L. Biko, S. Hellmig, B. Dekant, H. Bruenner, R. Burger and H. Wetzstein for expert technical assistance in animal experiments, histology, neurochemistry, and IgG preparations.