Neurological Manifestations of Glutamic Acid Decarboxylase Autoimmunity in Indian Patients.

Glutamic acid decarboxylase (GAD) is the rate-limiting enzyme to produce gamma-aminobutyric acid, which is the main inhibitory neurotransmitter in the central nervous system. It has 2 isoforms GAD 65 and GAD67, of which GAD 65 is used more extensively as a clinical biomarker due to its greater autoantigenicity.[ 1 ] It is widely used as a biomarker for type 1 diabetes mellitus, but relatively lately, it is also being used as a marker of neurological autoimmunity.[ 2 ] Stiff-person syndrome, limbic encephalitis, epilepsy, cerebellar ataxia, isolated cases of palatal tremor, and paraneoplastic neurological syndromes are some of the neurological diseases in which GAD 65 positivity is described.[ 3-6 ] GAD 65 positivity has also been described in multiple system atrophy and some patients with cognitive decline.[ 7 , 8 ] Most of the literature on GAD 65 neurological autoimmunity is from the West. There are only a few studies from Asia on the neurological manifestation of the GAD 65 antibody.[ 7-11 ] Sparse data is available on neurological manifestations of GAD 65 autoimmunity in Asians and no data on Indian patients. We present a retrospective case series of GAD 65-positive patients with an intent of identifying the neurological phenotypes in the Indian population. The previous studies in the Western population show that very high values of GAD 65 were seen in neurological autoimmunity compared to endocrine autoimmunity and low values were seen in the normal population as well.[ 2 , 3 , 12-14 ]

The study was conducted in the department of neurology and biochemistry in Amrita Institute of Medical Sciences, Kochi, Kerala, which is the tertiary care post-graduate teaching hospital in South India. Inclusion criteria–Those patients who were tested positive for GAD 65 antibody were included in the study if the request was made for diagnosis of neurological disease from February 2013 to July 2019. Exclusion criteria—Patients referred from other countries were excluded. We retrospectively identified patients, who were tested for GAD 65 antibody as per the request of neurologists from February 2013 to July 2019 by reviewing electronic medical records (EMR). Clinical and treatment details of positive patients were extracted from EMR. Levels of GAD 65 antibody were detected using enzyme-linked immunosorbent assay (ELISA) using commercial kits and following manufacturer's instructions. Two ELISA diagnostics kits were used to detect levels of GAD 65 antibody during the period between February 2013 to July 2019. Medizym anti-GAD ELISA, (MEDIPAN GMBH, Berlin, Germany) was used from 2013 February to 2016 June; Anti-GAD ELISA (EUROIMMUN AG, Luebeck, Germany) was used from 2016 June to 2019 July. For the first ELISA kit >5.0 IU/ml was considered as positive and for the second ELISA kit >10 IU/ml was considered as positive. Clinical outcomes were measured with 9-question modified Rankin Scale (mRS-9Q) (reduction in score by a score of one is considered as improvement), Unified Parkinson Disease Rating Scale (UPDRS) and Glasgow Coma Scale (GCS). All consecutive GAD 65 patients were included in the study. cerebrospinal fluid (CSF) GAD 65 assay was not performed. Tissue-based assay was performed in all patients to detect neuronal antibodies. Statistical analysis was performed using IBM SPSS version 20.0 software. Categorical variables were expressed using frequency and percentage. Numerical variables were presented using mean and standard deviation (SD). To test the statistical significance of the association of two categorical variables, Chi-square test was used. A P -value of <0.05 was considered statistically significant. This retrospective study was approved by the Institute ethics committee. Ethics committee determined that participant consent was not required. Data availability - The authors confirm that the data supporting the findings of this study are available within the article [and/or] its supplementary material . GAD 65 Supplementary data

We identified 922 patients who had undergone GAD 65 testing, out of which 81 tested positive (8.78%). The mean age was 55.42 years (SD 17.39, range 9–86 years, median age 57 years). Males ( n = 47) outnumbered the females ( n = 34) [Refer Table 1 ]. All the GAD values measured were below 5000 IU/ml. Thirty-four patients had a clinical diagnosis of atypical parkinsonism (two with coexisting autoimmune encephalitis, one with coexisting myasthenia), eight had autoimmune encephalitis, three had the stiff-person syndrome, seven had neuropathy (including two with atypical GBS and one with CIDP), two had Creutzfeldt Jakob disease, two had cervical demyelinating myelopathy, two had motor neuron diseases, two had epilepsy one had aquaporin 4 IgG-positive neuromyelitis optica syndrome, one had myasthenia gravis, one lambert eaton myasthenic syndrome (LEMS), one had vasculitis affecting CNS and peripheral nervous system, one myoneuropathy, one myeloneuropathy, one cerebellar ataxia, one post-infectious extrapyramidal syndrome, one had a single episode of seizure and one frontotemporal dementia. Of the 81 cases, there were alternate diagnoses in 12 cases, which were unrelated to GAD 65 autoimmunity (Creutzfeldt-Jakob disease-2, motor neuron disease-2, vascular parkinsonism-1, tuberculous meningitis-1, metabolic encephalopathy-1, macular degeneration-1, viral encephalitis-1, suspected transient ischemic attack-1, severe anxiety with depression-1, partially treated meningitis-1). Please refer to the supplementary data for full clinical details. Follow-up period ranges from 1 month to 6 years. Demographic data Mean GAD value was 206.31 IU/ml (SD 685.45), median value 6.83 IU/ml. Most patients had low values of GAD titer. Only two had a value above 2000 IU/ml. Mean GAD 65 values in the atypical parkinsonism group was 183.73 IU/ml (SD 757.77, median value 6.07 IU/ml), autoimmune encephalitis group was 42.75 IU/ml (SD 53.66, median 17.14), and stiff person group was 1240.85 IU/ml (SD 1524.24, median 328 IU/ml) [Refer Table 2 ]. Eighteen (22.2%) had underlying autoimmunity. Three patients had type 1 diabetes mellitus. Six (7.4%) had underlying neoplasm: carcinoma breast and carcinoma ovary (1), diffuse large B cell lymphoma (1), endometrial hyperplasia with atypia (1), carcinoma ovary (1), adrenal adenoma (1), thymoma (1). Thirty-three out of 43 patients with treatment details responded to immunotherapy (76.74%). Five patients had spontaneous improvement (Suspected myelopathy-1, atypical parkinsonism-1, autoimmune encephalitis-2, GBS-1). One stiff person syndrome with a high value (3388.43 IU/ml) showed a poor response to immunotherapy. Four patients died of infection. Four patients discontinued immunosuppression due to infection and four lost follow up [Refer Table 3 ]. GAD 65 subgroups GAD 65 immunotherapy response There were 34 cases (42%) of atypical parkinsonism in our series forming the most common significant diagnostic entity with GAD 65 positivity, followed by autoimmune encephalitis (8 cases, 9.88%). Fourteen of the atypical parkinsonism (14/34, 41.2%) fulfilled the criteria for autoimmune atypical parkinsonism (AAP) proposed by Kannoth et al .[ 15 ] Regarding the sex distribution, men were more affected with atypical parkinsonism (22/34; 64.70%), stiff-person syndrome (2/3;66.66%), and neuropathy (4/7;57.1%) while the neurological phenotype of women was more consistent with autoimmune encephalitis (6/8;75%). The levels of antibodies did not cross 10000 IU/ml to qualify it as high as per the literature in our study population.[ 13 ] To determine any possible correlation of GAD 65 levels with diagnosis and treatment response, we arbitrarily classified GAD levels up to 5 times the upper limit of normal as low and those above 5 times as high. As per this classification, 68 patients (86%) had low GAD levels and 13 had high GAD levels. Low GAD 65 levels were more frequent in atypical parkinsonism (31/33; 91.2%), autoimmune encephalitis (7/9; 77.8%), and neuropathy (5/7;71.4%), while high GAD level (i.e., >5 times the upper limit of normal) was seen in stiff-person syndrome (2/3;66.7%) and myelopathy (2/3;66.66%). In the low GAD group, 11/68 (16.17%) had associated autoimmune disorders while in the high GAD group, 7/13 (53.84%) had associated autoimmune disorders ( P = 0.003). As regards response to immunotherapy, 28/35 (80%) in the low GAD group improved with immunotherapy while in the high group, 5/8 (62.50%) improved with immunotherapy ( P = 0.38). In the low GAD group, 62/68 (91.17%) had other antibodies detected in abnormal levels, while in the high GAD group, all 13 (100%) had other antibodies detectable in abnormal levels ( P = 0.59). This included neuromyelitis optica IgG, contactin-associated protein-like 2 (CASPR2) antibodies, voltage-gated calcium channel antibody, anti-Factor XIII antibody, antinuclear antibody, anti-double stranded DNA, antiphospholipid antibodies, anticardiolipin antibodies, acetylcholine receptor antibody, anti-thyroid peroxidase antibody, anti-thyroglobulin antibody, and multiple neuronal antibodies, which are yet to be characterized, termed as unclassified neuronal antibodies.[ 15 , 16 , 17 ] Low complement levels were detected in 8/25 (32%) in the low GAD group, while in the high GAD group, it was 2/3 (66.66%) ( P = 0.28).

This case series highlights the diverse spectrum of neurological diseases associated with GAD 65 autoimmunity. In most of the previously published series, women were more affected.[ 2 , 7 , 8 , 10 , 11 , 12 , 13 , 14 , 18 , 19 , 20 , 21 ] Male preponderance was reported in the paraneoplastic subgroup, low concentration subgroup, and pediatric limbic encephalitis subgroup.[ 6 , 13 , 22 ] Unlike the other series, men dominated in our series. The male preponderance in our series can be explained by the predominance of atypical parkinsonism. Autoimmune atypical parkinsonism is a male-dominant disease.[ 15 ] The other male-predominant autoimmune diseases include myasthenia in the elderly, CASPR2, and Leucine-rich glioma-inactivated 1 (LGI 1) antibody-mediated autoimmune encephalitis. Apart from this, most of the patients in our series had low GAD 65 values. In a study by Muñoz-Lopetegi et al .,[ 13 ] in the low GAD 65 antibody group, males were more commonly affected. Atypical parkinsonism was the most common single disease entity in our series. GAD 65 autoimmunity in association with extra pyramidal manifestation was reported by Pittock et al .[ 2 ] in 16% cases of which, 4 were atypical parkinsonism. Eight cases of atypical parkinsonism fulfilling the diagnosis criteria of multiple system atrophy (MSA) were found to have low GAD 65 values in the study by Kuo et al .,[ 8 ] which amounted to 29.6% of cases with neurological symptoms in their series. One of the patients with MSA phenotype in this series improved with immunotherapy. A case of low positive CSF GAD 65 was reported with atypical parkinsonism from Italy without serum positivity.[ 23 ] Glutamic acid decarboxylase 65 autoimmunity presenting as hemiparkinsonism has also been reported.[ 24 ] In our series, 14 out of the 34 atypical parkinsonism (41.17%) fulfilled the diagnostic criteria for AAP- i.e., had the presence of neuronal antibodies and responded to immunotherapy. In the rest of the cases, either immunotherapy was not attempted due to various reasons or did not respond to first-line immunotherapy making it difficult to ascertain whether they are all AAP. The high prevalence of atypical parkinsonism is attributed to the fact that our group has described the syndrome of autoimmune atypical parkinsonism. This led to careful evaluation and investigation of cases of atypical parkinsonism for autoimmune etiology. In our series, only 17.3% (14/81) presented with the core manifestations of the GAD 65 syndrome—stiff-person spectrum, limbic encephalitis, ataxia, and epilepsy. It is interesting to note that even patients with the classical syndrome had low GAD 65 values compared to the literature, i.e., <10,000 IU/ml.[ 17 , 18 ] Only five cases had multifocal manifestations/overlap syndrome. Isolated peripheral nervous system manifestations were noted in nine cases (one myasthenia, one LEMS, seven neuropathies of which two were atypical AIDP and 1 CIDP). Peripheral nervous system manifestation of GAD 65 autoimmunity is described in multiple studies.[ 25 ] The most striking finding was the low level of GAD 65 in our population. In most series from the West, GAD 65 levels, more than >10,000 IU/ml have been reported in those with neurological manifestations.[ 2 , 3 ] In contrast, a couple of previous studies from Asia have shown the presence of low titers in neurological syndromes associated with GAD autoimmunity when compared to similar diseases from the West.[ 8 , 11 ] Glutamic acid decarboxylase 65 titers were low in series reported by Kuo et al .,[ 8 ] Wang et al .,[ 11 ] and Sriwastava et al .[ 18 ] These three studies were from Taiwan, India, and China, and indicate an ethnic variation—a tendency for low titers in Chinese and Indian population. In all the three studies, tests were conducted with commercial kits, which were used widely across the globe. In another study, African Americans were found to have greater severity of the disease when compared to Caucasians and another study showed more brainstem involvement with African Americans.[ 2 , 18 ] This exemplifies the role of ethnicity in GAD 65 autoimmunity, and this remains to be explored. The immunotherapy response in our series was good—76.74%. This is in striking contrast to the previous reports where only 50% responded to immunotherapy.[ 1 ] The poor response is attributed to the intracellular location of the antigen, autoimmunity, which is usually mediated by T cells.[ 1 ] This mechanism causes more neuronal damage than that occurs with B cell or antibody-mediated disease as in the case of antibody to synaptic protein. But, immunotherapy response in our group was 76.74%, which is comparable with that of high GAD 65 (>10,000 U/ml) series by Muñoz-LopetegiA et al. (77%).[ 13 ] Since our series did not show high GAD values as found in previous literature (>10,000 U/ml), we arbitrarily classified values upto five times the normal limit as low, and more than five times as high to compare the treatment responses in both the groups. We found that many patients with low positive values also showed a good response to immunotherapy. There was no significant difference in the treatment response between the low positive (80%) and high positive (62.5%) group of patients ( P -value - 0.386). This finding is similar to a study by Wang et al . in which it was found that in neurocritical patients with GAD positivity, all patients responded well to immunotherapy irrespective of their GAD titers.[ 11 ] Analyzing the clinical diagnoses and treatment response, we found that the clinical phenotype determines the treatment outcome, e.g., Those with myasthenia, demyelinating myelopathy, immune-mediated neuropathies, autoimmune encephalitis, and AAP responded well to immunotherapy in our treatment group when compared to stiff person syndrome, though the latter had higher GAD 65 values. This goes well with the current literature where treatment response depends on the phenotype of the classical syndromes—stiff person (77%), limbic encephalitis (71%), overlap syndrome (70%), cerebellar ataxia (58%), epilepsy (50%).[ 14 ] The good immunotherapy response in low positive GAD 65 in our series can be attributed to the fact that GAD 65 may be acting only as a marker of underlying autoimmunity rather than mediating the disease. This is very evident in some of the cases with NMO spectrum disorder, LEMS, LGI1, and acetylcholine receptor antibody positivity. In others, there were other pieces of evidence for antibody-mediated disease-ten patients had low complement in our series indicating an antibody-mediated disease (nine of them had unclassified neuronal antibody to unknown target epitope, as evidenced by immunofluorescence in the monkey brain, and one had CASPR 2 antibody positivity). Antibodies other than GAD 65 were detected in 92.6% of cases and most of them were neuronal antibodies. In our series, 22.2% had an underlying autoimmune disease which again points that GAD 65 positivity at low titers itself could be just an indicator of underlying autoimmunity. Our study brings out an important finding, a low positive GAD 65 antibody in the right clinical setting can be used as a good marker for a positive response to immunotherapy. In a suspected autoimmune neurological disease, the presence of GAD 65 antibody positivity whether low or high, is a marker of potentially good immunotherapy response, which should prompt the clinician to look for the coexistence of other antibodies. If other diagnostic possibilities are excluded, then it may be worthwhile attempting a trial of immunosuppression. A similar view has been expressed by Gaspard.[ 26 ] Glutamic acid decarboxylase 65 antibody-associated diseases are often reported to have coexisting antibodies. Gamma-aminobutyric acid antibodies were reported in 70% of cases, glycine receptor antibodies were reported in 15% cases, amphiphysin antibody was reported in 5% and rarely, gephyrin in GAD-associated stiff-person syndrome.[ 27 ] Though T cell-mediated disease mechanism is experimentally proven and GAD antigen is located intracytoplasmic, IVIgG is the mainstay of treatment of GAD 65-associated disease.[ 27 ] A transient moonlighting antigen to the synaptic cleft is the paradigm that is used to explain the pathogenicity of the GAD 65 antibody.[ 28 ] In a recent study by Czempik et al .,[ 29 ] they reviewed the literature from 1989 regarding the effect of plasma exchange in stiff-person syndrome and who found that 23 out of 30 GAD-positive patients (76.67%) underwent plasma exchange showed marked improvement. Though the randomized controlled trial of rituximab on stiff-person syndrome did not reveal any positive response, it is still a therapeutic option in patients who are not responding to other therapy.[ 28 ] Multiple studies have revealed GAD titers have no correlation with the severity of disease and also that, there is no correlation between a fall in titer and improvement of disease. Even after plasma exchange, the antibody levels were not falling but patients were improving clinically (Czempic).[ 29 ] The good response to treatment modalities that work well on antibody-mediated disease like IVIgG, plasma exchange, and rituximab (rituximab effect on refractory cases), heterogenicity in clinical manifestations, lack of correlation of titers with severity and presence of multiple coexisting antibodies made many experts view the pathogenic role of GAD 65 antibody skeptically.[ 27 , 30 ] Our series also questions the pathogenic role of GAD 65 antibodies as mentioned in the previous paragraph. Six patients in our series had neoplasm in association with GAD 65 positivity and neurological syndrome. Though they did not fulfill all the criteria for paraneoplastic neurologic syndromes, this pattern is in concordance with the study by Saiz et al . from Barcelona and distinctly different from the Mayo clinic series where there was no malignancy associated with the GAD 65.[ 2 , 3 ] We would like to stress the fact that a positive GAD 65 antibody should not prevent a neurologist from looking for an underlying malignancy. Being a retrospective study, it has its own limitations. Glutamic acid decarboxylase 65 was tested in all suspected autoimmune neurological disorders and not only in the published phenotypes, so chances of false positivity cannot be ruled out. However, this kind of extensive testing is needed to reveal the expanding clinical spectrum of GAD autoimmunity as has been the case with other antibodies. Since, we identified the cases from the laboratory register, chances for clinical false positivity also are there. Another significant limitation was that CSF GAD 65 test was not performed in our laboratory. This study brings out some important observations regarding the ethnic variability of the neurological manifestations of the GAD 65 antibody. Glutamic acid decarboxylase 65 antibody values were much lower in our population among patients with neurological disorders, even in published phenotypes. Male-dominant autoimmunity was seen unlike that in Western literature. A higher frequency of malignancy was noted in our series. Most striking was the high preponderance of positive GAD 65 in atypical parkinsonism. We also found that GAD 65 positivity is a useful marker for a positive response to immunotherapy in suspected autoimmune neurological syndromes, irrespective of their titers. Even though it has been observed that low GAD positivity might not be mediating the disease, the presence of GAD antibodies, even in low titers may be considered as one of the criteria to investigate for other coexisting antibodies as well as to offer trial immunotherapy to patients in the absence of an alternate diagnosis. Nil. There are no conflicts of interest.