Informed, written consent was obtained from the patient by Dr. Eric Lancaster to allow for donation of apheresed plasma and subsequent purification of immunoglobulins.
Therapeutic plasma exchange and immunosuppressive therapy in a patient with anti-GAD antibody-related epilepsy: Quantification of the antibody response
Article first published online: 25 JUN 2014
© 2014 Wiley Periodicals, Inc.
Journal of Clinical Apheresis
Volume 30, Issue 1, pages 8–14, February 2015
How to Cite
Farooqi, M. S., Lai, Y., Lancaster, E., Schmitt, S. E. and Sachais, B. S. (2015), Therapeutic plasma exchange and immunosuppressive therapy in a patient with anti-GAD antibody-related epilepsy: Quantification of the antibody response. J. Clin. Apheresis, 30: 8–14. doi: 10.1002/jca.21342
- Issue published online: 10 FEB 2015
- Article first published online: 25 JUN 2014
- Manuscript Accepted: 5 JUN 2014
- Manuscript Revised: 26 MAY 2014
- Manuscript Received: 18 SEP 2013
- NIH/NINDS. Grant Number: K08 NS075142
- a Dana Foundation Neuroimmunology Award and National Organization for Rare Disorders (NORD)
Antibodies to glutamic acid decarboxylase (GAD) have been associated with a host of neurological disorders including stiff person syndrome, cerebellar ataxia, limbic encephalitis, and epilepsy. Whether anti-GAD antibodies have an etiological role in these neurological disorders or simply serve as disease markers is unclear. Here, we report a case of a patient with recurrent seizures, poorly responsive to conventional treatment, associated with anti-GAD antibodies. The patient was experiencing near daily seizures at the time of presentation and had marked improvement while receiving immunosuppressive therapy and therapeutic plasma exchange (TPE). We go on to show that the patient had a substantial reduction of her GAD autoantibody burden following this therapy. Using immunostaining, we further demonstrate a progressive loss of GAD reactivity in the patient's sera to neurons and GAD-expressing HELA cells with successive TPEs. Hence, these data support the concept of an immune-mediated pathogenic component to these autoantibody-associated neurological syndromes. J. Clin. Apheresis 30:8–14, 2015. © 2014 Wiley Periodicals, Inc.
Glutamic acid decarboxylase (GAD) is an enzyme which decarboxylates the excitatory neurotransmitter glutamate to form the inhibitory neurotransmitter gamma-aminobutyric acid (GABA) . In mammals, there are two isoforms of GAD, each encoded by its own gene: gad1 and gad2 . These encode proteins of slightly different molecular weights—a 67 kilodalton (kDa) protein and a 65 kDa protein, respectively—termed GAD67 and GAD65 . Both proteins are expressed in neural tissues, though GAD65 is also strongly expressed in the pancreas . While both isoforms catalyze the decarboxylation of glutamate to produce GABA, they differ in their cellular location and kinetics. GAD67 is expressed uniformly throughout a cell's cytoplasm and is constitutively active, whereas GAD65 is found near nerve terminals and forms GABA rapidly as needed for synaptic activity and neurotransmission .
Autoantibodies to GAD, mainly GAD65, have been found in 70–80% of patients with Type I diabetes and are thought to be markers of pancreatic cell destruction since GAD is not found on the cell surface [1, 2]. Anti-GAD antibodies have also been associated with a range of neurological disorders, most notably Stiff Person Syndrome (SPS): a disease typified by progressive muscle stiffness and spasm of axial muscles (usually, proximal limb and lumbar muscles), as well as hyperreflexia and extreme sensitivity to external stimuli . Approximately 60% of patients with this syndrome have circulating anti-GAD antibodies . GAD autoantibodies have also been found in patients with cerebellar ataxia, limbic encephalitis, and epilepsy, especially in cases of therapy-resistant epilepsy [4-7]. Whether anti-GAD antibodies have an etiological role in these neurological disorders or simply serve as disease markers is unclear .
The concept of anti-GAD autoantibodies as disease markers is supported by the fact that generally pathogenic autoantibodies target surface epitopes of transmembrane proteins . GAD, an intracellular enzyme, would be inaccessible to antibodies in vivo, and a mechanism allowing anti-GAD antibodies to enter and affect neurons has not been proven. Furthermore, in a mouse model of GAD autoimmunity, it was shown that GAD-specific CD(4)+ T cells, in the absence of any other T cells or B cells, produced a lethal encephalomyelitis-like disease . These GAD-specific CD(4)+ T cells were present throughout the central nervous system (CNS), activated microglia, and produced interferon gamma . The addition of B cells, and subsequent generation of high levels of anti-GAD antibodies, had no effect on the incidence or severity of disease .
However, there are also some lines of evidence to suggest that anti-GAD antibodies are involved in the pathogenesis of neurological disease. First, relative to patients with type I diabetes, patients with SPS have far higher titers of anti-GAD antibodies in their serum, sometimes by as much as 100- to 500-fold [9-12]. Second, purified immunoglobulin G (IgG) from patients with neurological symptoms and anti-GAD antibodies affects the excitability of rodent motor cortex and spinal cord when injected into the CNS of living rats . Third and most importantly, patients report symptomatic improvement with immunotherapy [11, 14-20]. Specifically, intravenous immunoglobulin (IVIg) and therapeutic plasma exchange (TPE), which should work primarily for antibody-mediated disorders, have been shown to improve symptoms in some patients with SPS [14-17]. IVIg and TPE have also been used to treat patients with other anti-GAD-associated neurologic syndromes, such as limbic encephalitis, cerebellar ataxia, and epilepsy, and have led to symptomatic improvement [11, 18-20].
Here we report a case of a patient with recurrent seizures, poorly responsive to conventional treatment, associated with anti-GAD antibodies. The patient was experiencing near daily seizures at the time of presentation and had marked improvement while receiving immunosuppressive therapy and TPE. We also quantify the acute effects of TPE on the GAD antibody response both by direct measurement and immunohistochemistry.
MATERIALS AND METHODS
Clinical history and laboratory data were obtained by retrospective review of the electronic medical record. Informed, written consent was obtained from the patient by Dr. Eric Lancaster to allow for donation of apheresed plasma and subsequent purification of immunoglobulins.
Detection of Antibodies to GAD65
HELA cells were grown to near confluence on 12 mm glass coverslips. Cells were transiently transfected to express GAD65 using a plasmid that expressed human GAD65 under control of a CMV promoter (generous gift of Dr. Damien Bresson, LaJolla Institute of Allergy and Immunology, La Jolla, CA) . After allowing 24 h for expression, cells were fixed with 4% paraformaldehyde (5 min), washed three times with phosphate-buffered saline (PBS), permeabilized with 0.3% Triton X-100 (5 min) in PBS, washed three times with PBS, and blocked for 1 h in 5% normal goat serum in PBS. Patient sera or plasmapheresate (diluted 1:200 to 1:500 in blocking solution) and a mouse monoclonal antibody to GAD65 (ABCAM GAD6; ab26113; Cambridge, England) diluted 1:50 were applied for 1 h at room temperature. Coverslips were washed three times with PBS then stained with appropriate secondary antibodies (488 fluorescent anti-human and TRITC-conjugated anti-mouse; Molecular Probes; Eugene, OR) for 1 h at room temperature. Coverslips were stained with 4′,6-diamidino-2-phenylindole (DAPI) to visualize cell nuclei, mounted on slides, then imaged using a Leica DMR fluorescent microscope (Leica Microsystems, Buffalo Grove, IL).
Immunohistochemistry of Cultured Neurons
Cultured rat embryonic hippocampal neurons were generated as previously described  and grown in culture for 2–3 weeks. Cells were fixed with 4% paraformaldehyde (5 min), washed three times with PBS, permeabilized with 0.3% Triton X-100 (5 min), washed three times with PBS, and then placed in blocking solution (5% normal goat serum in PBS). Human sera or plasmapheresate (1:200 to 1:500 in blocking solution) was applied overnight at 4°C. For some experiments, the mouse monoclonal antibody to GAD65 (ABCAM GAD6, above) was also applied. Coverslips were washed three times with PBS, then treated with appropriate fluorescent secondary antibodies (above) for 1 h at room temperature. Coverslips were washed three times with deionized water, counterstained with DAPI, mounted on slides, and imaged using Leica DMR fluorescent microscope.
A 23-year-old woman presented to our hospital due to increasing seizure frequency at home over the past few weeks. Her seizures were characterized as generalized tonic-clonic and she said that they were typically preceded by an aura of anxiety, sometimes déjà vu. The seizures usually occurred nocturnally and lasted for approximately 1 to 2 min, followed by about 5 min of confusion. In the month prior to presentation, they had increased in frequency from approximately 1 per week to almost 1 per day. Her last seizure was 2 days before presentation.
The patient was first diagnosed with epilepsy about 4 years ago. Electroencephalographic (EEG) monitoring done at the time showed frequent left temporal epileptiform discharges with intermittent left temporal slowing. A subsequent MRI demonstrated T2 hyperintensity in her left mesial temporal lobe as well as a small area of T2 hyperintensity in her right temporal lobe. Basic laboratory studies of her blood and cerebral spinal fluid were all within normal limits, except for the finding of elevated anti-GAD antibodies in her serum. The patient did not have any other signs or symptoms associated with anti-GAD antibodies such as diplopia, dizziness, ataxia, muscle stiffness, encephalopathy/memory changes, or cognitive impairment.
Since her diagnosis, the patient had been treated with various doses and combinations of levetiracetam, oxcarbazepine, and lamotrigine, as well as methylprednisolone and IVIg. Her seizures initially responded to antiepileptic drugs and immunosuppressive therapy, and she had extended intervals where she remained seizure-free—the longest being roughly one and a half years. However, over the past several months, her seizure-free periods had been decreasing: where a course of IVIg would once provide a 2-month seizure-free interval, her last round of IVIg (given a month prior to presentation) benefitted her for only 2 weeks. Also, the patient's anti-GAD antibody levels, measured multiple times since her diagnosis, had remained persistently elevated and above the measurable reference range. As the patient was now having near daily seizures, she was admitted, at the recommendation of her outpatient neurologist, to our institution for initiation of immunosuppressive therapy and TPE.
In the hospital, the patient was continued on her home antiepileptic medications (200 mg of lamotrigine twice daily and 900 mg of oxcarbazepine twice daily) and placed on seizure precautions as well as 24-h EEG monitoring. No deficits were noted on neurological exam, and the patient scored a 30/30 on a mini-mental state examination. Her vital signs and the remainder of her physical examination were similarly unremarkable.
An anti-GAD antibody level was drawn, which was positive with titers outside the upper limit of detection (>250.0 international units/milliliter (IU/ml); a normal value is ≤5.0 IU/ml; measured by quantitative enzyme-linked immunosorbent assay (ELISA) (test #: 2001771) at ARUP Laboratories, Salt Lake City, UT). Laboratory tests demonstrated leukopenia (white blood cell count of 2,900/µl, likely secondary to recent IVIg therapy) and anemia (hemoglobin of 9.4 g/dl and hematocrit of 29%; reference range: 12.0–16.0 g/dl for hemoglobin and 36–46% for hematocrit). She was subsequently found to have low ferritin (4 ng/ml; reference range: 13–150 ng/ml) and low serum iron (26 µg/dl; reference range: 28–170 µg/dl), consistent with iron-deficiency anemia, and was begun on iron replacement therapy. Folate and vitamin B12 levels were within normal limits, and evaluation for celiac disease via tissue transglutaminase antibody testing was negative. Apart from a slightly elevated aspartate aminotransferase level (46 U/l; reference range: 15–41 U/l), her liver panel was within normal range. Basic metabolic and coagulation panels were also unremarkable. HIV and pregnancy tests were negative.
Consistent with prior EEGs, the in-house 24-h EEG monitoring study was abnormal at baseline, showing left anterior temporal focal slowing with abundant epileptiform discharges that became more frequent in sleep. No seizure activity was noted. The patient was begun on immunosuppressive therapy with mycophenolate mofetil 1 g by mouth per day for 2 days followed by oral prednisone (60 mg daily).
The Transfusion Medicine service was consulted regarding initiation of TPE. Due to the patient's significant disease, previous positive responses to immunosuppressive therapies, markedly elevated anti-GAD antibodies, and published case reports of the beneficial effects of TPE in comparable circumstances [19, 20], our service decided to initiate TPE. A total of five exchanges were performed over 9 days. For each treatment, 1.5 plasma volumes (4.4 L) were replaced with 5% albumin using the Spectra Optia® Apheresis System (TerumoBCT, Lakewood, CO). Calcium supplementation was given with each procedure (a 2% calcium gluconate solution infused at 80 ml/h) as the patient noted mild citrate toxicity symptoms (paresthesias) minutes into the first treatment. Although the effect of TPE on the patient's antiseizure medications was judged to be small, the administration times of the patient's antiepileptic drugs were adjusted in relation to TPE to minimize any possible medication removal.
The patient underwent her first three TPEs as an inpatient and was then discharged, as she did not have any seizures during this initial treatment period. She was continued on her home anticonvulsant medications (as above) as an outpatient and also received prednisone (60 mg by mouth daily) as well as two more TPE treatments. Over this time, she continued to remain seizure-free.
Anti-GAD antibody levels were measured via quantitative ELISA (ARUP laboratories) at various points over the 9-day period that the patient received TPE, but all returned as positive and outside the upper limit of detection (>250.0 IU/ml). Thus, to determine the effectiveness of antibody removal by our course of TPE, anti-GAD antibody levels were measured after serial dilution of the patient's serum samples (both dilution of stored samples and quantitative ELISA were performed at ARUP laboratories). The results of these studies are shown in Figure 1.
These tests demonstrate that the patient's initial level of GAD autoantibody was very high (115,900 IU/ml), more than 460-fold above the upper limit of detection of the assay. Even after completing five sessions of TPE, her serum anti-GAD antibody level (3970 IU/ml) was still about 16-fold above reference range. Overall, however, the course of TPE along with immunosuppressive therapy lowered her serum anti-GAD antibody levels by approximately 97%. We should note, however, that the second and final samples were drawn after TPE procedures, whereas the first and third samples were drawn prior to TPE. As seen in Figure 1, there is a spike in GAD autoantibody titer between the second and third samples. This change is likely due to the timing of the draws in that the third sample (drawn before TPE) reflects more time for reequilibration of IgG from the extravascular compartment as opposed to the second (and final) samples. Despite this timing difference, it is clear that the overall trend is toward a marked reduction of the anti-GAD antibody level after three and five TPE treatments.
Since anti-GAD antibodies may coexist with other autoantibodies, for example those to the GABA-B receptor , we wanted to determine whether GAD was the most important target of this patient's autoimmune response. We therefore further investigated the reactivity of patient sera with cultured embryonic rat hippocampal neurons. As shown in Figure 2, the patient's serum immunostained these neurons in a pattern that colocalized with a mouse monoclonal antibody to GAD, suggesting that the latter is the dominant neuronal autoantigen for this patient. Furthermore, GAD reactivity to neurons progressively decreased as the patient underwent TPE (Fig. 3), once again highlighting the effectiveness of TPE in lowering anti-GAD antibody levels. As seen by measurement with the immunoassay (Fig. 1), neuronal staining is markedly decreased after the first three treatments (Fig. 3). This pattern is seen again in Figure 4: HELA cells transfected to express GAD65 exhibit declining reactivity to the patient's sera as she undergoes sessions of TPE (Fig. 4).
This decrease in her GAD autoantibody burden and decline in serum reactivity to GAD antigens also correlated with clinical recovery. The patient noted marked improvement in her seizures, going from near-daily seizures to being seizure-free over the next month. After this period, however, her seizures recurred once more.
Anti-GAD antibodies have been linked to a variety of neurological syndromes including SPS, cerebellar ataxia, limbic encephalitis, and epilepsy . Patients with these syndromes tend to symptomatically improve with immunosuppressive therapy, suggesting a pathogenic role for these antibodies [11, 14-20]. Here, we have reported a case of a patient with very high serum levels of anti-GAD antibodies suffering from recurrent seizures poorly responsive to conventional treatment. In the past, the patient has had significant improvement, albeit temporary, with immunosuppressive therapy and, once again, had marked improvement while receiving mycophenolate and prednisone in combination with TPE.
With this treatment, the patient had a substantial reduction of her GAD autoantibody burden and had a seizure-free period of about 1 month. The correlation of lowered anti-GAD antibody levels with clinical improvement (and vice versa) has also been noted by other studies using TPE to treat anti-GAD-associated seizures [18-20]. While this does not establish pathogenicity, such data, combined with the results presented here, support the notion of an immune-mediated component to these syndromes, and of anti-GAD antibodies as clinically useful disease markers.
Another interesting aspect of the case was the possibility of TPE affecting the therapeutic levels of the patient's antiepileptic and immunosuppressive drugs. Though data regarding drug removal by TPE for specific drugs are not always available, one can estimate the effects of TPE on a medication by examining its volume of distribution and degree of protein binding . Generally, drugs with a high rate of protein binding (e.g., greater than 80%) and/or a low (e.g., roughly less than 0.2 L/kg) volume of distribution (Vd) are most likely to be removed during TPE . Neither oxcarbazepine nor lamotrigine fit these criteria, with both being moderately protein bound (40 and 55%, respectively) and both having moderate Vd s (0.75 L/kg and 0.9 L/kg, respectively) [24-26]. Accordingly, it was found that very little oxcarbazepine was removed by TPE when this question was studied in a teenage girl with epilepsy . Specifically, six single-volume TPEs separated by resting days removed approximately 4% of the dose and about 6% of total body stores. Similarly, in patients receiving maintenance prednisone (50–60 mg/day), one TPE removed only 1% of the daily dose. So while the timings of her medications were adjusted, it was for these reasons that the patient's medication dosages were not increased while she was receiving TPE.
In summary, here we have reported a case of a patient who was experiencing near daily seizures at the time of presentation and became seizure-free for 1 month after receiving immunosuppressive therapy and TPE. We showed that the patient had a considerable reduction of her GAD autoantibody burden following this treatment. Using immunostaining, we went on to characterize GAD as the likely dominant neuronal autoantigen for this patient and demonstrated a progressive loss of GAD reactivity in the patient's sera to neurons and GAD-expressing HELA cells as she underwent successive TPEs. Based on her last follow-up visit, the patient is currently being trialed on zonisamide, a sulfonamide anticonvulsant (300 mg by mouth daily), in addition to her other antiseizure medications. If her seizures continue to increase in frequency, however, TPE remains a viable therapeutic option. Whether it will once again provide symptomatic improvement, and whether it can be used as maintenance treatment as part of long-term therapy, remains to be seen.
The authors thank Christa Eisenmann, RN, Melissa Murter, RN, and Jennifer Green, RN at the Hospital of the University of Pennsylvania—Apheresis unit for their assistance in patient care.
- 17Successful treatment in a patient with a focal form of stiff-person syndrome using plasma exchange and intravenous immunoglobulin therapy. Rinsho Shinkeigaku 2002;42:766–770., , , , , , , .
- 25Novartis. Oxcarbazepine [package insert]. Novartis, East Hanover, NJ, 2011.
- 26GlaxoSmithKline. Lamotrigine [package insert]. GlaxoSmithKline, Research Triangle Park, NC, 2009.