Subacute nodopathy with conduction blocks and anti-neurofascin 140/186 antibodies: an ultrastructural study

Sir, The concept of nodo-paranodopathy was first applied to axonal Guillain-Barré syndrome subtypes with antibodies to gangliosides, then extended to some patients classified as chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) carrying IgG4 antibodies against paranodal axo-glial proteins, to focus on the site of nerve injury, to reminding specific pathophysiological mechanisms and reconcile contrasting electrophysiological and pathological findings (Uncini et al., 2013; Uncini and Vallat, 2017). In 2017, in an article published in Brain by Delmont et al., the authors observed that anti-Nfasc (neurofascin) 140/186 IgG3 are associated with a subset of CIDP patients. They suspected that these antibodies induce functional alterations at nodes, without any pathological study (Delmont et al., 2017). We herein report the first electrophysiological and pathological description of a patient with nodopathy and anti-Nfasc140/186 IgG3 antibodies. A male, in his 70’s, was admitted because of numbness of hands followed by weakness in the four limbs, which began 8 days before. No previous infection was detected. Examination showed predominantly proximal motor weakness of the four limbs, generalized areflexia, distal loss of pinprick and vibration sense at the four limbs, and mild respiratory failure. Serology for HIV, hepatitis B and C viruses, cytomegalovirus, Epstein-Barr virus, Campylobacter jejuni and Borrelia burgdorferi was negative. No antibodies to glycolipids were detected. Spine MRI, CSF, VEGF and serum complement (C3-C4-CH50) were normal. Serum immunofixation revealed IgA-lambda monoclonal gammopathy and bone marrow aspiration a low-grade IgA-lambda myeloma. No bone lesions were found on the whole-body radiography. Intravenous immunoglobulin (IVIg) was given (0.4 g/kg/day × 5 days) and the patient improved, such that he could walk again. Almost 3 weeks after the end of the first course of IVIg, he developed tetraplegia in a few days. A second course of IVIg was administered (without improvement), followed by five plasma exchanges (PLEX). At the same time the patient was treated with intravenous corticosteroids (dexamethasone: 50 mg/day × 5 days). Two months after the first symptoms and despite the treatments, the patient required mechanical ventilation by percutaneous tracheostomy and nasogastric tube, until Day 105. Thereafter he began to improve quickly. Three weeks after weaning artificial ventilation, he was admitted to a rehabilitation centre; after 4 weeks, he was able to drive his car. Nine months after the onset, he did not present any weakness, was able to walk, and complained only of distal paraesthesias of lower limbs. Four electrophysiological studies were performed (Supplementary material). At Day 7 after onset, partial conduction blocks [>30% decrease of proximal compound muscle action potential (CMAP) amplitude with <30% increased duration] (Uncini et al., 2017) were detected in the axilla-elbow and Erb’s point-axilla segments of the left median nerve (Fig. 1) and in the Erb’s point-axilla segment of the right ulnar nerve. Distal motor latencies (DML) were normal or slightly prolonged in the left ulnar and left peroneal nerves. Motor conduction velocities (MCV) were normal or slightly slow. Sensory action potential (SAP) amplitudes were all normal. At Day 30, there was a significant reduction of distal CMAP amplitudes of both median and ulnar nerves. Partial conduction block without excessive temporal dispersion emerged in the elbow-wrist segment of both median nerves. SAP amplitudes of both median and ulnar nerves were reduced whereas both sural SAPs were relatively preserved. Needle EMG revealed mild fibrillation and a reduced activation pattern with high frequency discharging motor units in distal limb muscles. The third test, performed on Day 98, at the clinical nadir, showed no recordable peroneal CMAPs and further reduction of both median and ulnar CMAP amplitudes with 142–146% increased DMLs. A complete conduction block was present in the elbow-wrist segment of both median nerves and an almost complete conduction block was detected in the below elbow-wrist segment of ulnar nerves with 47–50% slowed MCVs. EMG showed abundant fibrillation and severely reduced or absent activation in distal limb muscles. At Day 270, when the patient had almost completely recovered, distal CMAP amplitudes of median and ulnar nerves were 159–446% increased compared to the third study, but distal right peroneal CMAP was barely detectable. Conduction blocks were no longer present. DMLs and MCV were normal or only slightly slowed. Electrophysiological and immunological findings. (A) Left median nerve. Superimposed CMAPs recorded from abductor pollicis brevis muscle after stimulation at wrist, elbow, axilla and Erb’s point. On Day 7, distal CMAP amplitude was normal (8.2 mV) but the amplitude of proximal CMAP was 38% reduced in the segment axilla-elbow and additionally 31% reduced in the Erb’s point-axilla segment without temporal dispersion of all CMAPs. On Day 30, distal CMAP amplitude was 38% reduced, although still in the normal range, compared to Day 7 and there was 43% reduction of CMAP amplitude from elbow stimulation without temporal dispersion of CMAPs. MCV was 40 m/s. On Day 98 distal CMAP amplitude was 2.2 mV (73% decreased compared to Day 7) and there was a complete conduction block in the elbow-wrist segment. Distal motor latency was 7.5 ms. On Day 270, distal CMAP amplitude (5.1 mV) reached the normal range again, DML (3.7 ms) and MCV (48 m/s) were normal. (B) These are mouse sciatic nerve fibres stained with the patient’s IgG3 (green) and for CNTN1 (red) to label paranodes. Human IgG3 bound specifically to the node of Ranvier. Scale bars = 5 μm. (C) The isotype of anti-Nfasc140/186 IgG was measured by ELISA. The patient’s autoantibodies were mostly of the IgG3 isotype. Error bars represent mean and standard deviation. (D) Clinical remission (evaluated with R-ODS: Rasch-built Overall Disability Scale) matched with the depletion of anti-Nfasc140/186 IgG. (E) The anti-Nfasc140/186 IgG of the patient (NodoP1) potently activated complement in vitro. As control, the serum from another patient with IgG3 to Nfasc140/186 was tested (NodoP2); error bars represent mean and standard deviation. OD = optical density. Serum samples taken at 3, 8 and 9 months from the onset were tested on teased sciatic nerve to examine antibody reactivity against nodal/paranodal compartment. At clinical nadir (3 months), the sample presented a strong IgG fixation at the nodes of Ranvier that decreased by 8 months and completely disappeared at 9 months (Fig. 1). We then evaluated antibody reactivity against Nfasc155, Nfasc140/186, contactin-1, and Caspr-1. No samples presented antibody reactivity toward contactin-1 or Caspr-1. However, the serum sample at nadir presented a strong reactivity toward the nodal isoforms of neurofascin: Nfasc140/186. These antibodies also recognized Nfasc155. The titres of anti-Nfasc140/186 antibodies were 1:8000 at nadir, then 1:200 at 8 months, and became undetectable at 9 months. The isotype of the autoantibodies was predominantly IgG3. A complement activation assay also revealed that the anti-Nfasc140/186 IgG3 from this patient and another previously described patient potently induced C1q deposition in vitro (Fig. 1). A sural nerve biopsy was made 75 days after onset and was examined as described elsewhere (Vallat et al., 2011). Several longitudinal sections were examined by electron microscopy to study the nodes of Ranvier, which were compared to five normal controls, five acquired neuropathies (two vasculitic neuropathies, one nutritional neuropathy, two lymphomatous neuropathies), and five CIDP patients without serum reactivity to paranodal and nodal proteins. Moderately large myelinated fibre loss and few ovoids, in favour of a discrete acute axonal process, were seen, as well as a few demyelinating changes without any cellular infiltration, or ‘onion-bulb’ formations. Rare clusters of regeneration were observed. Microvilli of all examined nodes of Ranvier were almost completely absent or sparse; the outermost elongated cytoplasm parts of two adjacent Schwann cells tended to overgrow along the nodal axolemma (Fig. 2), to completely occlude the nodal gap. Rarely, a layer of this cytoplasm process joined the axon and constituted a ‘triple band’-like structure. All the paranodal regions appeared to be normal. Alterations of nodal structure in patients with anti-Nfasc140/186 IgG3 antibodies. (A) Normal control subject. Electron microscopy, longitudinal section, sural nerve biopsy: at the level of the node area, presence of several microvilli (arrows on each side of this nodal region). Myelin loops (ml) characterize the paranodal area. (B) Patient. Electron microscopy, longitudinal section, sural nerve biopsy: disappearance of microvilli, which are replaced by elongated extensions of Schwann cell cytoplasm so that the nodal gap is occluded. Ax = axon; C = collagen; MS = myelin sheath of another axon. The patient we report developed an acute onset, progressive sensory motor polyneuropathy with a different course from the classical Guillain-Barré syndrome. He had IgG3 anti-Nfasc140/186 antibodies and developed persistent conduction blocks that recovered without the development of excessive temporal dispersion of CMAPs, an electrophysiological hallmark of autoimmune nodopathy (Uncini and Kuwabara, 2015; Uncini and Vallat, 2017). Anti-Nfasc140/186 IgG3 bound the mouse sciatic nerve node and strongly activated complement in vitro. The evolution of conduction blocks and of anti-Nfasc140/186 antibody titres mirrored the patient’s clinical conditions. Electron microscopy showed disappearance of microvilli that were replaced by elongated extensions of Schwann cell cytoplasm occluding the node of Ranvier. Nfasc186 is located at the nodal axolemma and its ablation in mice results in nodal disorganization, including loss of Nav channels and ankyrin-G, and, as in the patient we report, invasion of the nodal gap by paranodal domains (Thaxton et al., 2011; Taylor et al., 2017). Fast, saltatory conduction in myelinated nerves requires the clustering of voltage-gated sodium channels (Nav) at nodes of Ranvier and persistent conduction block in the patient we report could be due to loss of Nav channel and ‘insulation’ of the nodal region by the Schwann cell cytoplasm extensions impeding the regeneration of action potentials. At nadir, the patient presented also slow MCV and prolonged DML that could be misinterpreted as an expression of demyelination-remyelination. However, inactivation of Nav channels by intravenous lidocaine or tetrodotoxin intoxication reduces MCV, even reaching the demyelinating range, possibly by increasing the rise time of the action potential and the internodal conduction time (Uncini and Vallat, 2017). Moreover in the mice in which Nfasc186 was not completely ablated, and complete conduction failure was not expected, MCV was reduced by 35% (Thaxton et al., 2011). The patient also presented evidence of distal axonal degeneration that, in analogy with acute nodopathies with antiganglioside antibodies, could be caused by complement-mediated damage of the nodal axolemma resulting in excessive calcium influx into the axon, protease activation and Wallerian-like degeneration (Uncini et al., 2013). In addition, the postnatal loss of Nfasc186 in transgenic mice was found to lead to abnormal myelin structure and axonal degeneration (Taylor et al., 2017). This suggests that the loss of Nfasc186/gliomedin interaction may induce axonal degeneration, in a still unknown way (Thaxton et al., 2011). Only one case with an IgG3 anti-Nfasc140/186 patient has been reported (Delmont et al., 2017): this patient had a subacute onset with a progressive severe neuropathy with cranial nerve involvement and respiratory failure. Conduction studies showed an axonal pattern but, even if not fulfilling the electrodiagnostic criteria of the Van den Bergh et al. (2010), the patient was classified as ‘probable CIDP’ (Delmont et al., 2017). In keeping with the case presented here, this patient did not respond to IVIg and was treated with repeated PLEX and steroids: after 7 months, he was able to walk without aids and distal CMAP had recovered in two out of three nerves. The anti-Nfasc140/186 IgG3 from this patient also activated complement in vitro (Delmont et al., 2017). The patient we report is the first description of the ultrastructural abnormalities of the node of Ranvier induced by antibodies against Nfasc186 and of their electrophysiological correlates. This case expands the clinical and immunological spectrum of autoimmune nodopathies, emphasizing the role of repeated electrophysiological examination and search of antibodies against constituents of the nodal region to individualize patients with a distinct pathophysiology and eventually predict treatment responsiveness. J.D. acknowledges grant support from the ‘Association Française contre les Myopathies’ (grant #21532). Supplementary material is available at Brain online.