Etiology and pathogenesis
| |
• PERM is an autoimmune disorder caused in a substantial number of cases by glycine receptor antibodies. |
| |
• The autoantibodies to glycine receptors found in patients are predominantly IgG1 subclass. |
| |
• IgG purified from patients with PERM and glycine receptor autoantibodies causes a reduction of glycinergic neurotransmission in cultured motoneurons. |
With the discovery of glycine receptor autoantibodies, PERM is now considered to have an autoimmune pathology in many, if not all, cases. In a small proportion (less than 20%) the disease appears to have a paraneoplastic etiology with thymomas or lymphomas found in some cases. However, the relationship between these tumors and the specific glycine receptor antibodies is unknown. In the majority of patients an underlying malignancy is not found even after years of follow-up. Some patients describe a preceding infectious illness suggesting that mechanisms such as molecular mimicry or a temporary breech in the blood-brain barrier may play a role.
The first pieces of evidence that glycine receptor autoantibodies are pathogenic in PERM were the clinical observations that some patients improve with immunotherapies that lower circulating antibodies and that antibody titers correlate with clinical improvement and relapses in some cases (20). Patients with PERM who are “seronegative” have a similar response to treatment as those patients with glycine receptor autoantibodies, suggesting that they too have pathogenic autoantibodies to as yet unknown CNS targets. On the other hand, patients with GAD65 antibodies seem to have a worse prognosis, though the numbers of patients in outcome studies is small (25).
Glycine receptor autoantibodies in PERM. Glycine receptors are pentameric ligand-gated chloride channels that exist as alpha homomers and alpha-beta heteromers, and their expression is spatially and developmentally regulated. In adults, alpha 1 heteromeric glycine receptors are the most abundant and are found on motor neurons in the spinal cord where glycine acts as an inhibitory neurotransmitter. Poisoning with strychnine, a potent antagonist of glycine receptors, causes profound (usually fatal) muscle rigidity as a direct result of loss of this spinal inhibition. Alpha 1 heteromeric glycine receptors are also found in the brainstem where glycinergic neurotransmission modulates eye movements; respiratory rhythms; and auditory, vestibular, and autonomic functions. The alpha 3 subunit is also expressed in the spinal cord, particularly in the dorsal horns, and glycinergic neurotransmission here is involved in the processing of pain and itch. Glycine receptors are extensively expressed in the retina where alpha subunits are restricted to specific retinal layers. Glycine receptors are found in the basal ganglia and forebrain, albeit at much lower levels, and probably located extrasynaptically. The physiological roles of glycine receptors in these brain areas remains poorly understood but in some regions they mediate tonic currents.
Glycine receptor autoantibodies from patients bind to the exposed extracellular domains of the 1-subunits in homomeric or alpha/beta heteromeric glycine receptors. Some patients’ sera additionally bind alpha 2 or alpha 3 glycine receptor subunits (04) or beta receptor subunits (44). Patients’ sera, but not that of healthy controls, contain antibodies that bind to neurons in sections of rat brainstem and ventral and dorsal horns of the spinal cord, where glycine receptors are most abundant. Because glycine receptor autoantibodies have been detected in cerebrospinal fluid as well as in serum, with substantial intrathecal synthesis in at least some cases, there is the potential for neurologic disease to arise as a direct result of modulation of the glycine receptor by autoantibodies. Antibody-mediated reduction or loss of glycinergic neurotransmission within the spinal cord and brainstem would be expected to result in hyperexcitability of motorneurons, hyperekplexia, cranial nerve dysfunction, disturbance of respiratory rhythms, and autonomic dysfunction accounting for many of the clinical features observed in patients with PERM (10).
The autoantibodies to glycine receptors found in patients are predominantly IgG1 subclass, with some IgG3 subclass. Divalent IgG1 can crosslink ionotropic receptors triggering their internalization and removal from the postsynaptic membrane in myasthenia gravis and anti-NMDAR encephalitis. Similarly, in HEK293 cells expressing glycine receptors, glycine receptor autoantibodies from patients are internalized, suggesting that these antibodies could also act by internalizing glycine receptors and thereby reducing the efficacy of inhibitory glycinergic neurotransmission (04). IgG1 and IgG3 antibodies are also capable of fixing complement and patient antibodies to glycine receptors have been shown to do this in vitro. However, it is not known whether complement activation in patients is responsible for any of the clinical manifestations. It seems unlikely that early disease is characterized by widespread neuronal destruction because MRI appearances are often normal and there is frequently a substantial response to immunotherapy.
A study directly showed a profound reduction of glycinergic neurotransmission in cultured motorneurons following incubation in IgG purified from patients with PERM and glycine receptor autoantibodies, providing the strongest evidence so far that these autoantibodies are pathogenic (11). The reduction in glycinergic currents was seen even after short incubations and when monovalent Fab fragments were used, suggesting that there are direct antagonistic actions of glycine receptor autoantibodies on glycine receptors in at least some patients. These conclusions are supported by studies demonstrating decreased potency of glycine on glycine receptors in the presence of patient autoantibodies (39). In similar studies, beta glycine receptor subunit specific autoantibodies were shown to reduce maximal glycine-gated currents providing evidence of a further mechanism that could contribute to disease pathology in a minority of patients whose antibodies also bind beta subunits (44).
The role of GAD65 autoantibodies in PERM. Autoantibodies to GAD65 are present in most patients with stiff-person syndrome and are found in a minority of patients with PERM (29; 04). GAD catalyzes the conversion of glutamic acid to GABA in GABAergic neurons. The GAD65 isoform is found at the cytoplasmic surface of synaptic vesicles and as such it is inaccessible to circulating antibodies in intact healthy neurons. Therefore, autoantibodies to GAD65 are unlikely to be the primary driver of stiff-person spectrum disorders. However, experiments using IgG from GAD65-antibody positive patients with stiff-person syndrome suggest the presence of pathogenic autoantibodies to other unknown CNS targets, which may affect inhibitory GABAergic neurotransmission (08).
Interestingly, there is some evidence that the patients with GAD65 autoantibodies have a worse prognosis than those with glycine receptor autoantibodies or “seronegative” cases (25). It is possible that the presence of these antibodies is indicative of substantial neuronal damage with exposure of intracellular antigens to the immune system, although there are alternative explanations such as earlier diagnosis and more aggressive treatment of patients found to have glycine receptor autoantibodies.
DPPX autoantibodies. Antibodies from patients with anti-DPPX encephalitis bind to extracellular epitopes of DPPX, a modulator of the potassium channel, Kv4.2. Kv4.2 and DPPX are widely expressed in the brain and enteric nervous system. A small study of anti-DPPX encephalitis reported lower DPPX and Kv4.2 expression in cultured neurons following 3-days incubation in a DPPX antibody-positive patient IgG compared to control IgG, suggesting that antibody-dependent crosslinking and internalization of the entire Kv4.2 complex may occur. However, an increase in gut excitability within seconds of exposure to anti-DPPX serum has also been described, hinting at rapid pathogenic mechanisms not consistent with endocytosis of the Kv4.2 potassium channel (35). Although it seems likely that DPPX antibodies are pathogenic, given the strength of the clinical association and improvements seen with plasmapheresis and immunosuppressive treatments, the molecular mechanisms of this disorder have not yet been fully explored.
The role of glycine receptor autoantibodies in neurologic presentations other than PERM. Although the pathogenic actions of glycine receptor autoantibodies in PERM are well established, the role of glycine receptor autoantibodies in other neurologic disorders is less clear. Glycine receptor autoantibodies from individual patients all bind alpha 1 subunits but they have varying specificities for alpha 2 and alpha 3 subunits. However, this does not appear to account for different clinical presentations because there is no correlation between the subunit specificities and the clinical features in patients. Studies have documented differences in the binding patterns of patient immunoglobulins to various brain regions expressing glycine receptors, as well as differences in levels of binding to pre- and post-synapses, both of which may offer possible explanations for variability in clinical syndromes (37; 45). However, these studies are technically challenging, have used only relatively small numbers of patient samples, and remain to be substantiated.
In stiff-person spectrum disorders and brainstem encephalitis, where the clinical presentation could be an anatomically limited variant of PERM, it seems likely that glycine receptor autoantibodies are pathogenic. Supporting this hypothesis, the frequency of clinical improvement with immunotherapy is high.
There are a growing number of reports and small case series of adult and pediatric patients with epilepsy occurring in association with glycine receptor autoantibodies including mesial temporal lobe, other focal epilepsies, and epileptic encephalopathies. A review of all glycine receptor antibody positive cases in the literature in 2018 (187 cases) identified 47.6% with PERM/stiff-person syndrome and 22.4% with epilepsy (42). However, the absolute number of cases reported remains small and those who have high titers of glycine receptor autoantibodies comprise a tiny fraction of all patients who present with seizures. Among these are patients who respond to standard antiepileptic treatment and others who have refractory seizures. A proportion of patients with refractory seizures respond partially or completely to immunotherapy (15). Interestingly, some immunotherapy responsive patients have additional neurologic features such as dyskinesia or ataxia, and perhaps these additional signs are an important clue for an underlying autoimmune epilepsy syndrome (07). The pathogenic relevance of glycine receptor autoantibodies in epilepsy remains unclear at the time of writing – they could be a bystander phenomenon, indicative of an autoimmune process or perhaps themselves play a causative role in disease. However, it is worth reminding the reader that epileptic seizures are not typically observed in strychnine poisoning.
In other neurologic presentations, the response to immunotherapy has only been reported in a small number of cases but there appears to be a partial or complete response in at least some individuals.
Neuropathology. There are several reports documenting the findings in postmortem studies of patients with a clinical diagnosis of PERM. These describe perivascular lymphocytic cuffing and infiltration and a variable degree and distribution of neuronal loss in the brainstem and spinal cord, suggesting a cytotoxic T-cell mediated component to the disease. However, there are few pathological cases in the literature in which the autoantibody status of the patient was known. It is possible that the majority of the cases were anti-GAD65-positive, an antibody which is now associated with a less complete response to immunotherapy than glycine receptor autoantibodies, and may reflect less reversible pathology such as T-cell mediated cytotoxicity. Furthermore, the pathological changes in “end-stage” disease may not be representative of the pathogenic mechanisms earlier in the disease course.
A single documented case with coexisting glycine receptor and GAD65 autoantibodies showed T-cell mediated extensive neuronal loss in a pallido-nigro-dento-bulbar-spinal pattern, as well as reduced glycine receptor and glycine transporter 2 (GlyT2) expression on surviving spinal neurons (46). These findings would be consistent with both cytotoxic T-cell-mediated damage and antibody-mediated internalization. Evidence of cytotoxic T-cell-mediated neuronal injury was also observed at postmortem in a young man with a clinical syndrome of PERM and retrospective findings of both glycine receptor and NMDAR autoantibodies in his serum (43). However, in this case, the interpretation is complicated by the presence of two likely pathogenic neuronal autoantibodies and the very rapidly fatal disease course, which is not typical.
Single case reports of autopsies from patients with glycine receptor antibody-associated PERM document relatively mild and anatomically limited inflammatory changes, neuronal loss, and gliosis (32; 47). Further postmortem studies would be helpful to establish the pathological characteristics of glycine receptor antibody-positive PERM and provide evidence for the relative contributions of T-cell-mediated neuronal injury and more reversible antibody-mediated effects to the clinical phenotype.