Search

Myasthenic syndromes are characterised by weakness and increased fatigability due to an impairment of neuromuscular transmission. Autoimmune or genetic factors compromise the function of muscle-type nicotinic acetylcholine receptors (AChRs), which have two isoforms composed of five subunits in muscle - fetal AChRs have a γ- and adult AChRs have an ε-subunit. Rapsyn is another key component of the postsynaptic molecular machinery required for neuromuscular transmission. It directly interacts with AChRs and is required for AChR clustering at the neuromuscular junction (NMJ). The NMJ is one of the most widely studied synapses but there are still aspects of normal NMJ development and function that need to be better understood. In this work, patch clamp experiments on muscle (TE671/CN21) and non-muscle cells (HEK293) were performed to elucidate AChR function under different physiological and pathological conditions. The adult AChR isoform recovered faster from desensitisation compared to the fetal AChR. Rapsyn was shown to induce AChR aggregation in TE671/CN21 and HEK293 cells and facilitated recovery kinetics in both fetal and adult AChRs in the muscle cell line. Rapsyn also increased the sensitivity for myasthenia gravis (MG) sera to inhibit adult AChR currents, and confined current inhibition to sera with AChR autoantibodies only. Finally, the novel αL251R mutation found in a patient with congenital myasthenic syndrome (CMS) was shown to be associated with a voltage-dependent conversion of AChR ion selectivity from cationic to anionic, transforming the receptor into an inhibitory channel. This work provides an evaluation of the functional impact of rapsyn on AChRs and helps to understand better the significance of the developmental switch from the fetal to the adult isoform. The data suggest that rapsyn is important for the analysis of the functional effect of MG sera on AChRs, and show novel functional effects of a new CMS mutation that sheds light on AChR function.

Myasthenia gravis (MG) is an autoimmune disorder of the neuromuscular junction (NMJ). Most patients have antibodies (Abs) against the acetylcholine receptor (AChR) or muscle specific kinase (MuSK) detected by radioimmunoassays (RIAs). Cell-based assays can more sensitively detect these and other Abs (e.g. to low density lipoprotein receptor-related protein 4 (LRP4)) but no large cohorts of MG patients have been systematically studied. RIA positive MuSK-Abs are mainly IgG4 subclass and block the interaction between LRP4 and MuSK, resulting in reduced MuSK phosphorylation and agrin-induced AChR clustering at the NMJ. MuSK phosphorylation could provide a potential therapeutic target. An international cohort of 231 MG patients negative for AChR and MuSK Abs by RIA, were tested with CBAs for all three known antigens. Negative sera were screened for novel antigenic targets by immunoprecipitation and mass spectrometry (IP/MS). MuSK phosphorylation was assessed in cultured mouse C2C12 myotubes treated with MuSK-MG IgG subclass fractions. The formation, dispersal, and re-formation of AChR clusters by MuSK-Abs in the presence or absence of a tyrosine phosphatase SHP2 inhibitor (NSC-87877) were measured quantitatively. First, the detection of MuSK-IgG Abs by CBA was optimised, and the potential relevance of the antibodies shown by their ability to reduce agrin-induced AChR clusters in the myotubes. Overall, AChR-Abs (10%), MuSK-Abs (7%) and LRP4-Abs (<1%) antibodies were newly identified. Five negative sera bound to human muscle cell lines, but no plausible novel antigens were identified by IP/MS. MuSK-MG IgG4 fractions inhibited MuSK phosphorylation, which was restored by SHP2 inhibition. SHP2 inhibition also increased AChR cluster formation, blocked dispersal and increased re-formation of AChR clusters by MuSK-Abs. As discussed, finding new antibody targets in MG was confounded by testing patients who were no longer severely affected following immunotherapies. MuSK-MG, by contrast, is frequently severe and modifying phosphorylation of MuSK and its downstream signalling pathways may provide a novel treatment approach for this and other forms of MG.

The recent history of autoimmune neurology is marked by the discovery of many central nervous system (CNS) antibody-mediated diseases. These disorders are caused by antibodies that target important proteins expressed in the neuronal surface, which are believed to be directly pathogenic. These antibodies are immunoglobulin G (IgG) isotype and, as such, have the potential to cross the placenta during gestation. Foetal exposure to CNS-targeting antibodies could alter developing neuronal circuits, leading to disease. However, the consequences of exposure to these antibodies during neurodevelopment has hardly been considered. To study the relationship between maternal antibodies towards neuronal surface proteins and neurodevelopmental disorders in the foetus a dual approach was undertaken. First, pregnancy serum samples from mothers of children later diagnosed with a neurodevelopmental disorder and from mothers of children with typical development were screened for the presence of neuronal surface antibodies. Next, the effects of pathogenic neuronal surface antibodies in the offspring were assessed in a maternal-to-foetal transfer mouse model. Antibodies to neuronal surface proteins in the gestational serum, particularly CASPR2 antibodies, were found to associate with an increased risk of mental retardation and disorders of psychological development in the progeny. The animal model showed that mice exposed in utero to CASPR2 antibodies have long term behavioural sequelae and histological findings suggestive of abnormalities in brain development. These findings support a model in which maternal antibodies towards foetal neuronal proteins cause long-term behavioural deficits and permanent abnormalities at the cellular and synaptic level in a subset of children with neurodevelopmental disorders.

Autoimmunity is a failure to tolerate circulating or cell surface expressed self antigens,leading to activation of the immune system and attack of self tissues. Muscle-specific kinase (MuSK) myasthenia gravis (MG) is a disease caused by antibodies to MuSK and hallmarked by fatigable muscle weakness. MuSK is a tyrosine kinase that interacts with low-density lipoprotein receptor-related protein 4 (LRP4), resulting in maintenance of the high density of acetylcholine receptors (AChRs) at the neuromuscular junction; this high density is essential for efficient transmission of signals from nerve to muscle, and MuSK antibodies impair this transmission. MuSK antibodies are predominantly IgG4, a subclass that does not induce immunological damage. Thus how these antibodies cause neuromuscular junction dysfunction is not clear. Potential mechanisms of the MuSK antibodies were explored in in vitro experiments. Plasmas from fourteen MuSK-MG patients were studied. IgG antibodies and IgG subclass profiles were measured with flow cytometry. Total IgG, Fabs, IgG4 and IgG1-3 subclass antibodies were prepared and purified; these were used to investigate the effects on MuSK surface expression, binding of LRP4 to MuSK, and agrin-LRP4-MuSK-induced AChR clustering in C2C12 mouse myotubes. No evidence for MuSK endocytosis by MuSK IgG, IgG1-3 or IgG4 antibodies was found. The predominant IgG4 subclass, and the monovalent IgG Fabs, blocked binding between LRP4 and MuSK but both IgG4 and IgG1-3 subclass antibodies were equally able to disperse pre-formed and newly-induced AChR clusters in C2C12 myotubes. The block of LRP4-MuSK interaction by IgG4 antibodies is likely to be a major pathogenic mechanism in MuSK-MG, which may lead to disrupted signal transduction, reduced AChR aggregation and neuromuscular transmission failure at the neuromuscular junction. In addition, MuSK IgG1-3, until now described as nonpathogenic, may also contribute to the reduced AChR density and neuromuscular dysfunction in MuSK-MG. These results provide new evidence concerning the pathogenic antibodies and their mechanisms in MuSK-MG.

Neuromyelitis optica (NMO) and NMO spectrum disorders (NMOSD) are associated with the disease-specific autoantibody aquaporin-4 (AQP4-Ab), which is thought to be pathogenic. Some NMO/NMOSD patients do not have this antibody and may have different clinical and immuno-pathological disease characteristics, but previous clinical NMO/NMOSD studies have been confounded by inclusion of such patients. To define better the characteristics of AQP4-Ab disease, disease course, outcomes and predictors of disability were investigated in 106 AQP4-Ab positive NMO/NMOSD patients from the UK and Japan. AQP4-Ab positivity conferred a high risk of relapsing disease and substantial disability; age at disease onset and ethnicity were important predictors of disability type. Visual disability was more common in younger patients and those of Afro-Caribbean ethnicity whilst older patients and Caucasian patients were more at risk of motor disability. To determine whether disease characteristics were influenced by AQP4-Ab binding specificities, the differential binding of patient AQP4-Ab against the two main AQP4 isoforms was investigated. Although the relative binding to the two isoforms differed between patients, there was no association between these differences and clinical features such as relapse type, severity, onset age and ethnicity. The clinical and in vitro characteristics of AQP4-Ab negative NMO/NMOSD patients were studied. It was shown that these patients represent an aetiologically heterogeneous group. Some have other inflammatory and infectious disorders, some have low levels of AQP4-Ab and a significant proportion have antibodies to myelin oligodendrocyte glycoprotein (MOG-Ab). Others have antibodies that bind to neurons or oligodendrocytes in primary cultures. Attempts to identify novel antigens by immunocapture techniques were made, but were not successful. Patients with MOG-Ab showed differences when compared to those with AQP4-Ab including higher proportion of males, younger age at disease onset and greater likelihood of conus involvement on imaging. Additionally, patients with MOG-Ab appear to have more favourable outcomes, with better improvement from the onset attack and a lower probability of subsequent relapses. In conclusion, the work in this thesis has shown that AQP4-Ab disease is not synonymous with the term NMO and that seronegative NMO/NMOSD patients represent a clinically and aetiologically heterogeneous group and should therefore be classed separately from those with AQP4-Ab.

Autoimmune encephalopathies, mainly described in adults, are neurological syndromes characterized predominantly by impaired alertness usually with seizures, movement disorders, psychiatric features or autonomic instability. Well-characterised serum or cerebrospinal fluid (CSF) antibodies to neuronal surface proteins have been identified in these disorders. The aim was to define and extend the role of autoantibodies in paediatric encephalopathies. The clinical, paraclinical and immunological features of patients with autoimmune encephalopathies, acute encephalitis, N-methyl-D-aspartate receptor (NMDAR)-antibody and voltage gated potassium channel (VGKC)-complex antibodies, and in children with basal ganglia encephalitis were studied in detail. In 48 children with a diagnosis of probable autoimmune encephalopathy, 21 (43%) had CNS autoantibodies, principally to NMDAR and VGKC-complexes, but children without these antibodies could also respond well to immunotherapies. Immune-mediated/autoimmune encephalitis was also the largest subgroup (34%) of encephalitis in an Australian paediatric encephalitis cohort. NMDAR-antibodies were found in children with well-defined encephalitis, but also in patients presenting without an encephalopathy, with distinct clinico-radiological white matter syndromes, or with a neurological relapse following herpes-simplex encephalitis. Children with VGKC-complex antibodies did not have autoantibodies to the complex associated proteins leucine-rich glioma inactivated 1 (LGI1) and contactin-associated protein 2 (CASPR2) as frequently seen in adults, but were more likely to have a primary neuroinflammatory disease than in patients without these antibodies. Finally, despite a recent report, dopamine receptor antibodies or an alternative autoantibody were not found in children with basal ganglia encephalitis. Novel clinical syndromes were identified with NMDAR-antibodies, raising important questions about the role of NMDARs in white matter, and the relationship with preceding infections. Patients with basal ganglia encephalitis may have an alternative pathobiology mediating their disease. Moreover, antibody positivity is not essential for the diagnosis of autoimmune encephalitis and children with or without detectable neuronal antibodies may respond to immunotherapy.

Glycine receptor antibodies have been identified in a few patients with progressive encephalomyelitis with rigidity and myoclonus (PERM), a highly disabling disorder characterised by rigidity, spasm and brainstem symptomatology. The clinical characteristics of patients with glycine receptor antibodies have not yet been fully described and it is not clear whether GlyR-Abs are pathogenic or just an epiphenomenon. This study examined the clinical features and immunotherapy responses of 45 patients; characterised the GlyR-Ab pathogenicity, subunit specificity and binding to different brain region in vitro, and examined mice injected with GlyR-Abs to model the disease in vivo. Most of the patients were classified as PERM but some patients had symptomatology beyond the classical motor manifestations and there were four patients with tumours (thymomas and lymphomas). GlyR-Ab titres were varied in serum and CSF, but there was intrathecal synthesis in the six patients with suitable samples. Most patients were very disabled but almost all showed excellent responses to immunotherapies. The antibodies were mainly IgG1 and IgG3 subclasses, activated complement on glycine receptor-transfected HEK cells at room temperature, and caused internalisation and lysosomal degradation of the glycine receptors at 37°C. GlyR-Abs bound to rodent spinal cord and brainstem co-localising with monoclonal antibodies to GlyRα1 on the surface of neurons. GlyR-IgG injected intra-peritoneally led to impairment in forced walking ability, sensorimotor function and coordination. Analysis of the brain showed that animals injected with patients' IgG, but not control IgG, had antibodies bound to the brainstem, spinal cord, cerebellum and caudate, co-localising with GlyRα1 monoclonal antibody. Intra-cerebroventricular injection of GlyR-IgG caused an anxiety-like behaviour in mice but no evident motor disturbances. These results provide the first evidence of in vitro and in vivo pathogenicity of the GlyR-Abs, supporting the use of long term immunosuppression in these patients to provide them with a good prognosis.

Immune mechanisms are thought to be involved in the pathological disease process in a number of childhood epileptic syndromes and encephalitis. Of particular interest is the occurrence of autoantibodies to essential neuronal proteins, for example the N-methyl-D-aspartate receptor (NMDAR), in the blood and spinal fluid in some of these patients. The aims of this study were: to examine the sera of newly diagnosed paediatric epilepsy patients for specific neuronal autoantibodies, correlate to epilepsy phenotype and disease outcomes; to investigate the pathogenicity and epileptogenicity of central nervous system autoantibodies (CNS) in vivo; and to test new therapies in vitro and in vivo based on the potential pathogenic mechanisms. In 290 paediatric patients with new-onset epilepsy and seizures tested for CNS autoantibodies, 11.4% were positive (33/290 versus 8/112 in controls; p=0.01, Fisher's exact test). Previously unreported contactin-2 antibody positive and contactin-associated-protein 2 (CASPR2) antibody positive epilepsy patients were described. Patients with 'focal epilepsy of unknown cause' were more likely to be antibody positive. To test the pathogenicity and epileptogenicity of these antibodies, a novel in vivo telemetry system was used to continuously record electroencephalogram (EEG) in mice injected into the cerebral lateral ventricle with NMDAR antibody (NMDAR-Ab) positive immunoglobulin (IgG). Although no spontaneous seizures were seen, mice challenged with the pro-convulsant pentylenetetrazole (PTZ) had increased seizure susceptibility, and more epileptiform "spikes" in the EEG after PTZ compared to healthy control (HC) IgG injected mice. Seizure susceptibility strongly correlated with binding intensity of NMDAR-Ab IgG analysed in post-mortem tissue. Given the hypothesis this epileptogenic effect was mediated by NMDAR-Abs internalising cell surface NMDARs, and to try and rescue this deficit, a neurosteroid, pregnenolone sulphate (PregS) known to increase NMDAR cell surface expression, was therapeutically used. This approach worked in vitro, and although in vivo effects were not yet established, treatment with neurosteroids may be beneficial for autoantibody mediated neurological disease.

Autism spectrum disorders (ASD) are neurodevelopmental disorders with multiple neurobiological aetiologies, which could be genetic, structural, metabolic or immune-mediated. ASDs are diagnosed with deficits in social communication and restricted and repetitive behaviours, and are associated with sensorial atypicalities. 30% of cases have co-existing epilepsy. A series of in vitro, in vivo and post-mortem investigations were undertaken to examine sensory atypicalities in ASD. In vitro characterisation of hippocampal neuronal cultures using immunofluorescence demonstrated the presence of multiple cell types including neurons, astrocytes and microglia. The distribution of ion channels of the Shaker family and tumour necrosis factor α receptors in astrocytes and neurons were identified but not explored further. Neuroanatomical and neuropathological investigations of primary olfactory cortex, using post-mortem stereology, demonstrated a specific increase in glial cell densities in layer II, which was negatively associated with age in ASD. Increases in glia were also associated with symptom severity and often co-localised with the presence of corpora amylacea in layer I. Qualitative analysis of the olfactory tubercle demonstrated that corpora amylacea did not extend to this neighbouring region of the primary olfactory cortex in ASD. These changes were independent of co-existing epilepsy and not observed in epilepsy without ASD. Preliminary pilot studies of the hippocampus provided a stereological sampling strategy to quantify cell densities in future investigations of this area in ASD. Neurophysiological investigations using collected magnetoencephalography data demonstrated diminished occipital gamma oscillatory synchrony in ASD in a visual time perception task. This did not always predict behavioural outcome but was specific to ASD and could not be explained simply in terms of changes in task performance. Moreover, changes in oscillatory synchrony were associated with symptom severity. These observations in primary sensory domains in post-mortem tissue and in patients suggest possible novel mechanisms in ASD and extend knowledge of the neurobiological bases of these disorders.

Over the last few years, antibody-mediated disorders of the central nervous system (CNS) have been a rapidly expanding field in neurology. The antibodies bind to extracellular domains of proteins expressed in the CNS and impair neuronal function. Patients with limbic encephalitis, Morvan’s syndrome have antibodies that bind to the voltage gated potassium channel (VGKC)-complex proteins, LGI1 and CASPR2, while patients with antibodies to the NMDAR have more complex encephalopathies. Clinical evidence suggests that these antibodies are pathogenic as immunotherapy reduces antibody levels, which correlate closely with clinical improvement. Antibodies can be detected in patient sera using human embryonic kidney (HEK) cells transfected with CASPR2, LGI1 and NMDAR subunits, or by binding to rodent brain sections and primary neuronal cultures. In vitro NMDAR-antibodies have been shown to cross-link synaptic NMDARs and cause their subsequent internalisation from the cell membrane. It is not known if passive transfer of NMDAR-antibodies can recapitulate features of disease or if these pathogenic mechanisms occur with VGKC-complex antibodies. The characteristics of LGI1 and CASPR2 antibodies were explored in vitro. IgG1 and IgG4 were identified as the predominant antibody subclasses, with all CASPR2 sera containing IgG1 antibodies and all LGI1 sera containing IgG4. On transfected cells both LGI1 and CASPR2 antibodies activated the classical pathway of the complement cascade. Furthermore CASPR2 antibody binding to transfected HEK cells showed a reduction in intensity over time which was most likely a result of internalisation of the CASPR2-antibody bound complex. Both CASPR2 and LGI1 antibodies bound strongly to the surface of live hippocampal neurons in culture, and this staining was reduced 24 hours after the removal of IgG, corresponding to a loss of antigen expressed at the cell surface. Patients with NMDAR-encephalitis present with a psychotic disturbance, memory loss and seizures, usually progressing to a reduced level of consciousness and a marked movement disorder. A single injection of IgG, purified from patients with NMDAR-encephalitis was administered into the lateral ventricle of mice. Animals were observed for 40 days and showed significant behavioural changes, including a deficit in spontaneous alternation, dyskinetic hind-limb clasping and visual signs of spontaneous seizures, which are reminiscent to features of the disease. Behavioural changes were not observed in animals injected with IgG from healthy individuals. These behavioural alterations appeared to be reversible and were no longer present at day 40, which was supported by protein analysis showing equivalent levels of the NMDAR expressed in the hippocampus in animals receiving patient and control IgG. In summary, these findings provide further evidence that supports a direct role of autoantibodies in NMDAR-encephalitis, limbic encephalitis and Morvan’s syndrome.

N-methyl-D-aspartate receptor (NMDAR) antibody encephalitis is a recently described autoimmune encephalopathy defined by the presence of serum antibodies that bind NMDARs (NMDAR-Abs). NMDAR-Ab encephalitis is a severe, but treatmentresponsive encephalitis with subacute onset. It can be associated with tumours and affects mainly young adults. Patients present with cognitive dysfunction, seizures, psychiatric and sleep disorders and most develop dyskinesias, autonomic instability and reduced consciousness. To explore further the NMDAR-Abs and their potential pathogenicity, a series of in vitro investigations were performed and preliminary attempts at passive transfer of disease. Human embryonic kidney (HEK) cells transfected with the NR1 and NR2B subunits, and live cultured neurons, were used first to detect NMDAR-Ab binding. Immunocytochemistry and ow cytometry demonstrated that binding to transfected HEK cells could be improved when NMDAR were presented in clusters by cotransfection with the postsynaptic density protein PSD-95. The NR1 subunit was identified as the target of NMDAR-Abs, and a novel quantitative assay based on immunoprecipitation of NR1 tagged by fusion with green uorescent protein was developed. Measurement of NMDAR-Ab levels showed that antibody levels corresponded to the clinical disease score within individual patients. Although the purification of full length NR1 was not successful, a secreted N-terminal construct was created and expressed in HEK cells. The binding of NMDAR-Abs was confirmed and this construct will be used for active immunisation in future. To explore pathogenic mechanisms in vitro, the main antibody subclasses were shown to be IgG1 and IgG3. Moreover the patients' autoantibodies, but not healthy control antibodies, were able to activate the complement cascade in vitro in cell lines and primary cultures. Finally, the NMDAR-Abs were shown to bind to primary microglial cultures and to cause morphological changes corresponding to early activation processes after prolonged exposure. The research has developed new assays that could be used for diagnosis and serial studies and revealed new potential mechanisms in NMDAR-Ab encephalitis.

Voltage-gated potassium channel antibodies (VGKC-Abs) have been described in a spectrum of immunotherapy-responsive neurological diseases. These include neuromyotonia (NMT), a peripheral nervous system disease, limbic encephalitis (LE) a central nervous system disease, and Morvan's syndrome which affects both the peripheral and central nervous systems. Occasionally, these diseases can be associated with tumours. VGKC-Abs are defined by their ability to precipitate VGKCs from 2% digitonin-solubilised rabbit brain membranes labelled with 125I-α-dendrotoxin. α-dendrotoxin is known to bind Kv1.1, 1.2 and 1.6. We asked whether the clinical heterogeneity could be explained by the specificity of the patient antibodies for the different Kv1 channels. We received clinical data regarding 113 VGKC-antibody positive cases: 78 had LE/Epilepsy, 17 had NMT, 13 had MoS and five had other syndromes. In addition, we collected clinical data about 26 patients with a unique form of dystonic epilepsy in association with VGKC-Abs. Surprisingly, a series of biochemical and immunohistochemical experiments in both brain tissue and in Kv1-transfected cells showed that potassium channels were the target of only 3% of the 113 'VGKC'-Ab positive sera. We proposed that the 'VGKC'-antibodies were not directed against Kv1s but against other proteins in the 125I-α-dendrotoxin labelled rabbit brain complexes. Lgi1 (leucine-rich glioma inactivated 1), CASPR2 (contactin-associated protein 2) and contactin-2 were found within these complexes and, when individually expressed in a heterologous mammalian cell line, were found to be the direct target of the 'VGKC'-Abs. Lgi1-antibodies were found in 58% of the cases with LE/Epilepsy, 71% of the dystonic epilepsy patients and 15% of the MoS cases. By contrast, CASPR2 antibodies were found in 85% of the MoS, 61% of the NMT and 12% of the LE/Epilepsy cases. CASPR2 antibodies had a 99% negative predictive value for the presence of an underlying tumour. Contactin-2 antibodies were found uniquely in 2 of 113 cases. These findings describe novel antigenic targets which help to explain the clinical heterogeneity seen in 'VGKC-complex'-antibody associated diseases. In addition, they offer a method to predict the cases with underlying tumours. These findings have important implications for understanding the biology of these diseases.