Upper Motor Neuron-Predominant Motor Neuron Disease: A Novel Immunotherapy-Responsive Association of GAD65 Autoimmunity

Abstract

Background:

Autoimmune disorders can present as motor neuronopathies and need to be excluded prior to the diagnosis of Amyotrophic Lateral Sclerosis (ALS). We aimed to characterize the clinical phenotypes of patients with motor neuron disease (MND) in the context of high titer serum/CSF GAD65 antibodies (radioimmunoassay).

Methods:

A retrospective review of all Mayo patients (between 1/1/2003 and 12/31/2023) with motor neuronopathy and co-existing high titer GAD65 antibodies (≥20nmol/L in serum [equivalent to >10,000 IU, ELISA] or detection in CSF) was performed. Clinical phenotypes and outcomes were compared with ALS patients diagnosed in the last 5 years (1/1/2019–12/31/2023) who tested negative for GAD65 IgG.

Results:

We identified 12 patients with high-titer GAD65 IgG and motor neuronopathy, who often had lower back spasms, history of an exaggerated startle response with immunotherapy responsiveness as compared to ALS patients. On further analysis, a subgroup of these patients with neurogenic changes on EMG, had an upper motor neuron (UMN) predominant syndrome (58%), with history of exaggerated startle (57%), lower back spasms (43%), tandem gait impairment (86%) and UMN bladder symptoms (71%) that were significantly different from the ALS controls. The UMN predominant GAD65 MN responded favorably to immunotherapy with stable electromyography; significantly lesser worsening in mRS and mortality on long term follow-up.

Discussion:

An upper motor neuron predominant motor neuronopathy is a distinct manifestation of GAD65 autoimmunity. Co-existing symptoms like exaggerated startle response, lower back spasms, impaired tandem gait, and UMN bladder signs might warrant consideration of an immunotherapy trial, which could yield favorable results.

Introduction

Amyotrophic lateral sclerosis (ALS) is a debilitating and fatal disorder with progressive degeneration of upper and lower motor neurons.[1] Immune mediated diseases such as multifocal motor neuropathy (MMN), motor predominant chronic inflammatory demyelinating polyneuropathy, inflammatory myositis, anti-IgLON5 disease, paraneoplastic motor neuron disease are some ALS mimics that are often treatable and hence should be excluded prior to the diagnosis of ALS.[2–6] Glutamic acid decarboxylase-65 (GAD65) neurological autoimmunity often presents with well recognized syndromes such as stiff person syndrome, cerebellar ataxia, epilepsy and limbic encephalitis.[7, 8] However, no study has comprehensively explored motor neuronopathy (MN) presentations specifically in patients with GAD65 autoimmunity except for few case reports.[9, 10] We aimed to explore the clinical phenotypes and outcomes of patients with MN phenotype and concurrent high serum GAD65 antibody titers in comparison to patients with sporadic ALS without GAD65 antibodies.

Methods

Using the Mayo Data Explorer, we retrospectively reviewed all patients diagnosed with motor neuron disease (MND), between 1/1/2003 and 12/31/2023, meeting the revised El-Escorial criteria for ALS[11] and cross-referenced them with the Mayo Clinic Neuroimmunology Laboratory database for co-existing high titer GAD65 antibodies by radioimmunoprecipitation assay (GAD65 ≥20nmol/L in serum [equivalent to >10,000 IU on enzyme linked immunosorbent assay], with or without CSF GAD65 antibody positivity).[7, 12] The comparison group consisted of all consecutive patients diagnosed with sporadic ALS in the last 5 years (between 1/1/2019 and 12/31/2023), who tested negative for serum GAD65 antibodies. An inflammatory cerebrospinal fluid was defined as pleocytosis (>5cells/mcl) and/or protein elevation >100mg/dL. Patients with modified Rankin Scale (mRS)≥4 at last follow up were deemed to have poor outcomes. Categorical and continuous variables were reported as frequencies and median with range, respectively. Kaplan-Meier curves were used to compare the survival between the compared groups. The study was approved by the Mayo Clinic Institutional Review Board. Informed consent for research participation was obtained from all participants.

Results

Between 2003 and 2023, we identified 843 patients with high titer GAD65 antibodies who were seen at Mayo Clinic. Among these patients, we identified 12 patients (1.4%) with MND diagnosis who had high titer serum/CSF GAD65 antibodies and called them GAD65 motor neuronopathies (MN). These patients had evidence of both upper and lower motor neuron involvement raising concern for ALS. Non-neurological diseases associated with GAD65 autoimmunity such as Hashimoto’s thyroiditis was present in only 17% of the patients, while none had type 1 diabetes mellitus. The median serum GAD65 IgG titer in the GAD65 MN group was 221.5 nmol/L (range: 38.7–3077 nmol/L). CSF GAD65 IgG testing was done in five of the 12 patients, all of whom tested positive with a median of 2.17nmol/L (range: 0.15–20.9 nmol/L).

GAD65 MN vs ALS:

We compared these GAD65 MN patients to those with ALS to identify phenotypic differences (Table 1). Patients with GAD65 motor neuronopathies were more likely to have lower back spasms (25% vs 0%, P=0.04), history of an exaggerated startle response leading to suspicion for stiff person syndrome (33% vs 0%, P=0.014) and less likely to have fasciculations (42% vs 85%, P=0.018) as compared to patients with ALS. A history of exaggerated startle led to a subsequent movement disorder laboratory study in four GAD65 MN patients, although only one of them had electrophysiological evidence for an abnormal startle response.

Table 1:

Comparison between patients with GAD65 motor neuronopathy patients and Amyotrophic Lateral Sclerosis

Characteristic	GAD65+ motor neuronopathy, N=12	Sporadic ALS, N=20	P value
Median age at onset of symptoms, in years (range)	65 (36–79)	60.5 (37–79)	0.954
Males (%)	6 (50)	12 (60)	0.718
Diabetes mellitus type 2	2 (16.7)	4 (20)	1.000
Hashimoto’s thyroiditis	2 (16.7)	3 (15)	1.000
UMN predominant syndrome	7 (58.3)	7 (35)	0.277
Initial symptom onset			0.018
Limb	7 (58.3)	19 (95)
Bulbar	5 (41.7)	1 (5)
Cognitive decline	2 (16.7)	5 (25)	0.683
Spastic dysarthria	7 (58.3)	4 (20)	0.053
Flaccid dysarthria	1 (8.3)	0 (0)	0.375
Spastic-flaccid dysarthria	1 (8.3)	2 (10)	1.000
Dysphagia	8 (66.7)	10 (50)	0.471
Dyspnea	7 (58.3)	10 (50)	0.726
Weakness	11 (91.7)	18 (90)	1.000
Wasting	7 (58.3)	13 (65)	0.724
Fasciculation	5 (41.7)	17 (85)	0.018
Spasticity	9 (45)	9 (75)	0.147
Hyperreflexia	11 (91.7)	18 (90)	1.000
Pseudobulbar affect	5 (41.7)	2 (10)	0.073
Bladder urgency, frequency	5 (41.7)	4 (20)	0.240
Tandem gait impaired	7 (58.3)	4 (20)	0.053
Falls	8 (66.7)	10 (50)	0.471
Probable stiff person syndrome as per criteria	1 (8.3)	0 (0)	0.375
Axial rigidity	1 (8.3)	1 (5)	1.000
Lower back spasm	3 (25)	0 (0)	0.044
History of exaggerated startle	4 (33.3)	0 (0)	0.014
Movement disorder lab study suggestive of SPS	1 (8.3)	0 (0)	0.375
Epilepsy	1 (8.3)	0 (0)	0.375
Cerebellar ataxia	1 (8.3)	0 (0)	0.375
mRS <4 at diagnosis	7 (58.3)	12 (60)	1.000
mRS <4 at last follow up	2 (16.7)	3 (15)	1.000
mRS delta worsening, median (range)	1 (0–4)	2 (0–4)	0.021
Revised El Escorial criteria			0.030
Possible/Probable	8 (66.7)	5 (25)
Definite	4 (33.3)	15 (75)
Inflammatory CSF (protein >100mg/dL or cells >5/mcl)	3 (25)	2 (10)	0.423

UMN-upper motor neuron, MN-motor neuronopathy, SPS-stiff person syndrome, EMG-electromyography

GAD65 MN patients were immunotherapy responsive:

Immunotherapy was initiated in six (50%) of the GAD65 MN cases as compared to four (25%) in the ALS control group due to concern for an underlying autoimmunity. An upper motor neuron predominant syndrome with features atypical for ALS and high titer GAD65 antibodies were the main reasons for suspecting autoimmunity in the GAD65 MN group, while in the ALS patients false positive antibodies and confounding ancillary investigations led to suspicion of an autoimmune etiology. The salient characteristics of these patients who received immunotherapy in both groups are summarized in Supplementary Table 1.

Majority of the GAD65 MN cases had an improvement post immunotherapy (83% vs 0%, P=0.048) with a stable electromyography (EMG) on follow up (100% vs 16.7%, P=0.015).

We further investigated the cohort of GAD65 MN cases to identify distinct clinical phenotypes that were predictive of immunotherapy responsiveness. We classified seven patients who had an upper motor neuron (UMN) predominant syndrome and some electrodiagnostic signs of anterior horn cell involvement as UMN predominant GAD65 MN. The remaining five patients had a mixed UMN and LMN or LMN predominant phenotype and were classified as GAD65 motor neuron disease (MND).

A representative case of an UMN predominant GAD65 MN:

A 75-year-old gentleman presented with a subacute onset of dysarthria, dysphagia about 10 days following his fourth COVID vaccination. On examination he had spastic dysarthria with no tongue atrophy or fasciculations. He had hyperreflexia with extensor plantar responses and no weakness of limbs. MRI of his brain and spine were normal. He was diagnosed to have ALS elsewhere and was referred for a second opinion. The EMG revealed increased insertional activity in the lower limb muscles (gastrocnemius and peroneus longus) and large motor unit potentials in certain C6–7 innervated muscles (biceps and triceps), with no signs of structural or compressive radiculopathies on MRI spine. He had a serum GAD65 IgG of 554 nmol/L and a CSF GAD65 IgG of 3.80 nmol/L along with CSF protein of 103mg/dL and 4 nucleated cells. He was treated with IV methyl prednisolone for 12 weeks (1g for 3 days, then once a week for 5 weeks, then once in two weeks for 6 weeks) with which he had 60% improvement in his speech and swallowing functions. A repeat EMG done after 3 months of immunotherapy was normal with no evidence of increased insertional activity or large motor unit potentials in the same muscles. About 1 year after his initial visit with us, he continued to do well with mild residual dysarthria.

Patients with UMN predominant GAD65 MN exhibit unique characteristics:

We compared the three groups of UMN predominant GAD65 MN, GAD65 MND and ALS patients to look for differences in clinical phenotype and outcomes (Table 2). The median age of symptom onset was not significantly different between the three groups.

Table 2.

Comparison of characteristics between UMN predominant GAD65 MN, GAD65 MND and ALS patients

Characteristic	A. UMN predominant GAD65 motor neuronopathy, N=7 (%)	B. GAD65+MND, N=5 (%)	C. Sporadic ALS, N=20 (%)	P value (A vs B)	P value (A vs C)	P value (B vs C)
Median age at onset of symptoms, in years (range)	54 (36–74)	70 (43–79)	60.5 (37–79)	0.268	0.370	0.336
Males (%)	3 (42.9)	3 (60)	12 (60)	1.000	0.662	1.000
Initial symptom onset				0.558	0.156	0.016
Limb	5 (71.4)	2 (40)	19 (95)
Bulbar	2 (28.6)	3 (60)	1 (5)
Cognitive decline	0 (0)	2 (40)	7 (35)	0.152	0.283	0.274
Spastic dysarthria	5 (71.4)	2 (40)	4 (20)	0.558	0.023	0.562
Flaccid dysarthria	0 (0)	1 (20)	0 (0)	0.417	*	0.200
Spastic-flaccid dysarthria	0 (0)	1 (20)	2 (10)	0.417	1.000	0.504
Dysphagia	4 (57.1)	4 (80)	10 (50)	0.576	1.000	0.341
Dyspnea	2 (28.6)	5 (100)	10 (50)	0.028	0.408	0.061
Weakness	6 (85.7)	5 (100)	18 (90)	1.000	1.000	1.000
Wasting	1 (14.3)	5 (100)	13 (65)	0.015	0.033	0.274
Fasciculation	1 (14.3)	3 (60)	17 (85)	0.222	0.002	0.252
Spasticity	6 (85.7)	3 (60)	9 (45)	0.523	0.091	0.645
Hyperreflexia	7 (100)	4 (80)	18 (90)	0.417	1.000	0.504
Pseudobulbar affect	5 (71.4)	0 (0)	2 (10)	0.028	0.005	1.000
Extensor plantar	6 (85.7)	1 (20)	8 (40)	0.072	0.077	0.621
Bladder urgency, frequency	5 (71.4)	0 (0)	4 (20)	0.028	0.023	0.549
Tandem gait impaired	6 (85.7)	1 (20)	4 (20)	0.072	0.004	1.000
Falls	6 (85.7)	2 (40)	10 (50)	0.222	0.183	1.000
Probable stiff person syndrome as per criteria	1 (14.3)	0 (0)	0 (0)	1.000	0.259	*
Axial rigidity	1 (14.3)	0 (0)	1 (5)	1.000	0.459	1.000
Lower back spasm	3 (42.9)	0 (0)	0 (0)	0.205	0.012	*
History of exaggerated startle	4 (57.1)	0 (0)	0 (0)	0.081	0.002	*
Movement disorder lab study suggestive of SPS	1 (14.3)	0 (0)	0 (0)	1.000	0.259	*
Epilepsy	0 (0)	1 (20)	0 (0)	0.417	*	0.200
Cerebellar ataxia	1 (14.3)	0 (0)	0 (0)	1.000	0.259	*
mRS <4 at diagnosis	3 (42.9)	4 (80)	12 (60)	0.293	0.662	0.621
mRS <4 at last follow up	1 (14.3)	1 (20)	3 (15)	1.000	1.000	1.000
mRS delta worsening, median (range)	0 (0–2)	1 (0–4)	2 (0–4)	0.202	0.003	0.621
Revised El Escorial criteria				0.222	0.009	0.597
Possible/Probable	6 (85.7)	2 (40)	5 (25)
Definite	1 (14.3)	3 (60)	15 (75)
Inflammatory CSF (protein >100mg/dL or cells >5/mcl)	2 (28.6)	1 (20)	2 (10)	0.066	0.099	0.447
Improvement post immunotherapy	5/5 (100)	0/1 (0)	0/4 (0)	0.167	0.008	*
Follow up EMG				*	0.015	*
Stable	5 (71.4)	0 (0)	1 (5)
Worsened	0 (0)	0 (0)	5 (25)
Not done	2 (28.6)	5 (100)	14 (70)
Mortality at 5 years	1 (14.3)	3 (60)	14 (70)	0.222	0.024	1.000
Requirement of Non-invasive ventilation	3 (42.9)	4 (80)	6 (30)	0.293	0.653	0.121
Median duration since disease onset, months (range)	78 (19–182)	38 (9–47)	32.5 (9–117)	0.073	0.026	0.621

^*Statistics not computable; UMN-upper motor neuron, MN-motor neuronopathy, SPS-stiff person syndrome, EMG-electromyography

The UMN predominant GAD65 MN patients had a history of exaggerated startle (57%), lower back spasms (43%), tandem gait impairment (86%) and UMN bladder symptoms of urgency and frequency (71%) that were significantly different from the sporadic ALS group. Only one patient satisfied the criteria for definite ALS as per the revised El Escorial criteria compared to the 75% in the ALS group (P=0.009). One patient in the UMN predominant GAD65 MN group exhibited cerebellar ataxia and lower back spasms with an exaggerated startle reflex that was confirmed by electrophysiology, indicating central hyperexcitability. None of the other patients had lumbar lordosis or hypertonia of the lumbar and paraspinal muscles upon examination and did not meet the diagnostic criteria for stiff person syndrome (the criterion of limb hypertonia was excluded due to its potential overlap with ALS).[13] Inflammatory CSF (CSF protein ≥100 mg/dl, CSF nucleated cell count >5 cells/mcl) was observed in 29% of UMN predominant GAD65 MN, 20% of GAD65 MND, and 10% of ALS patients. Importantly, there was no distinct differences in the clinical phenotypes or outcomes between the GAD MND group and the ALS group.

Treatment outcomes at follow up:

Apart from the five patients with UMN predominant GAD65 MN, one patient in GAD65 MND group and four patients in the ALS group had also received immunotherapy. Patients in the UMN predominant GAD65 MN had improvement in speech, swallow, gait and stiffness while none in the ALS group showed an improvement (100% vs 0%, P=0.008). Remarkably, UMN predominant GAD65 MN patients also had an improved or stable EMG on follow up. Apart from the patient whose EMG normalized on follow up, another patient had an initial improvement in EMG a year after immunotherapy with subsequent worsening on follow up studies.

The median follow up duration was significantly longer in the UMN predominant GAD65 MN compared to the ALS group (Table 2). The median change in mRS from diagnosis to final follow up (0 vs 2, P=0.004) and mortality at 5 years (14% vs 70%, P=0.024) were significantly less in the UMN predominant GAD65 MN as compared to the ALS group (Figure 1). Importantly, this change in mRS or mortality was not significant between the GAD65 MND and the ALS groups (Table 2). The survival curves of the three groups at 5 years follow up are shown in Figure 1.

Figure 1.

Kaplan-Meier curve showing the differences in survival between the upper motor neuron predominant GAD65 motor neuronopathy, GAD65 motor neuron disease and amyotrophic lateral sclerosis patients.

Discussion

We describe a cohort of UMN predominant MN patients with high titer serum/CSF GAD65 antibodies who show a response to immunotherapy with better outcomes and lower mortality on long term follow up. Along with electrodiagnostic evidence of MN, a history of exaggerated startle, lower back spasms, impaired tandem gait is often noticed in these patients although they often do not satisfy the electrophysiologic criteria for stiff person syndrome. Only one patient had fulfilled the proposed diagnostic criteria for probable stiff person syndrome.[13] Interestingly, most of these patients did not satisfy the requirement for “definite ALS” when using the revised El Escorial criteria and they often had improved/stable EMG on follow up which is not characteristic of ALS.

A high titer of GAD65 antibodies or positive CSF GAD65 IgG is highly specific for neurological autoimmunity, with around 2% of the general population showing GAD65 antibody positivity. Given ALS’s prevalence of about 6 cases per 100000, the occurrence of both ALS and higher titer GAD65 IgG in multiple patients is unlikely.[1, 14] None of our patients with GAD65 neurological manifestations had type 1 diabetes mellitus, suggesting that elevated GAD65 antibody titers were not due to systemic autoimmunity.

The classical syndromes of GAD65 autoimmunity include stiff person syndrome, cerebellar ataxia, epilepsy and limbic encephalitis while other rare neurological manifestations include cognitive impairment, brainstem dysfunction and myelopathy.[7] Presentation as an isolated motor neuronopathy without any of these well recognized manifestations is hitherto unreported. A paraneoplastic etiology was reported in a motor neuronopathy patient with SOX-1 and GAD65 antibodies in the serum, but the GAD65 titer was quite low (1:32) and no testing method was provided.[10]

The reasons for suspecting autoimmunity were varied in our patients, but most of the patients with GAD65 motor neuronopathy had a fairly consistent clinical phenotype. On the contrary, all the four ALS patients received immunotherapy due to confounding ancillary investigations. All five GAD65 motor neuronopathy patients who improved post immunotherapy had a UMN predominant presentation although they had evidence of LMN involvement by EMG. There exists a possibility that the patients who survived longer with immunotherapy might have a prolonged course of illness which has been infrequently observed in ALS. However, the stability of EMGs on follow up and improvement in EMG findings in two patients argues against this assumption.

Improvement post immunotherapy may not be dramatic in GAD65 autoimmunity, and most patients have incomplete recovery.[8] This may reflect the underlying pathophysiology as GAD is an intracellular antigen and response to these autoimmune disorders generally do not mirror the great benefit noted with antibodies targeted against extracellular antigens. There is a lack of consensus on whether GAD65 autoantibodies are truly pathogenic or biomarkers of T-cell mediated autoimmunity as antibody titers do not correlate with the severity of the disease and immunotherapy does not bring about consistent improvement across all clinical manifestations.[15–17]In support of these, a T cell mediated pathology consisting of CD8+ T cells and microglial proliferation was demonstrated in the spinal cord of a patient with anti GAD65 stiff person syndrome who also had lower motor neuron involvement with neurogenic changes on EMG.[9]

The GAD65 MND group in our study, which showed a clinical course similar to sporadic ALS cases, raises the concern that GAD65 IgG positivity might be a secondary phenomenon in a subset of cases. These patients could have neurodegenerative ALS with a humoral response to GAD65 due to neuronal loss, leading to the release of GAD65 epitopes. On the contrary, the UMN predominant GAD65 MN group had distinct features with better outcomes compared to the ALS patients, which suggests true underlying autoimmunity and warrants clinical recognition.

Few limitations include the retrospective nature and small number of cases in our study. Given the retrospective nature, we were unable to utilize more objective measures for improvement in domains such as speech and swallow. By using the modified Rankin Scale (mRS), we were able to capture the disability changes for all patients, though this measurement may not be specific to motor neuron disease (MND). There also exists a potential ascertainment bias as patients with UMN predominant syndrome might have had a longer follow up due to diagnostic uncertainties compared to classical ALS patients.

Our study points to the rare occurrence of GAD65 antibodies in patients with motor neuronopathy who may resemble the phenotype of sporadic ALS. Identifying those with a UMN predominant presentations with features such as exaggerated startle, lower back spasms and overactive bladder may lead to consideration of an immunotherapy trial with possible beneficial outcomes. Recognizing this rare manifestation of GAD65 autoimmunity may avoid premature closure of an ALS diagnosis in patients with a motor neuronopathy presentation. Larger studies are necessary to confirm our findings.