Patients and controls
Serum samples from 100 GBS patients included in the Spanish cohort of the International GBS Outcome study (IGOS) [21] were used in this screening. The IGOS is a multicentre, prospective, observational cohort study that investigates factors that determine and predict the clinical course, subtype and outcome of GBS [22]; including patients fulfilling diagnostic criteria of the National Institute of Neurological Disorders and Stroke (NINDS) [1] or Miller Fisher syndrome (MFS) and other variants of GBS [3, 23]. Patients from the Spanish cohort were enrolled between February 2013 and January 2020. All patients fulfilled diagnostic criteria for GBS and were included within 2 weeks from onset of weakness. Serum samples were aliquoted and stored at − 80 °C until needed. In this study, we used serum samples extracted at baseline. Sixty-two (62%) of the baseline samples analysed were collected before starting treatment.
Clinical variants were defined as sensorimotor, pure motor, pure sensory, Miller Fisher syndrome (MFS) and ataxic. Nerve conduction studies results were classified as acute inflammatory demyelinating polyneuropathy (AIDP), acute motor axonal neuropathy (AMAN), acute motor-sensory axonal neuropathy (AMSAN), equivocal or normal. The outcome of all patients with GBS at 6 months and 1 year from disease onset were assessed using the GBS disability score (GDS), a widely accepted system for evaluating the functional ability of patients [24]. Patients unable to walk independently (≥ 3) at 6 months were defined as having a poor outcome in this study.
Additionally, serum samples from a control group (n = 90) including 45 healthy controls and 45 patients with other neuromuscular disorders [23 amyotrophic lateral sclerosis (ALS), 22 Charcot–Marie–Tooth (CMT)] were included.
Autoantibody screening protocol
Autoantibody screening experiments included anti-ganglioside antibody detection with ELISA, nodo/paranodal (NF155, NF140, NF186, CNTN1 and CNTN1/CASPR1 complex) antibody detection by ELISA and cell-based assays, immunocytochemistry using patient sera on murine dorsal root ganglia (DRG) neurons and neuroblastoma-derived human motor neurons (IgG and IgM) and reactivity pattern assessment by immunohistochemistry on monkey sciatic nerve sections (IgG and IgM).
Testing for nodo/paranodal antibodies
Autoantibodies against NF140, NF186, NF155, CNTN1 and CASPR1 were tested by ELISA.
Maxisorb 96-well ELISA plates (Thermo Fisher Scientific, NUNC, Denmark) were coated with 1 μg/ml human recombinant CNTN1 protein (Sino Biological Inc., Georgia, USA), 1 μg/ml NF155 protein (Origene, Maryland, USA), 1 μg/ml NF140 protein (Sino Biological), 1 μg/ml NF186 protein (Origene) or 5 μg/ml CASPR1 protein (R&Dsystems, MI, USA) overnight at 4 °C. Wells were blocked with 5% non-fat milk in PBS 0.1% Tween 20 for 1 h, incubated with sera diluted 1/100 in blocking buffer for 1 h, and then incubated with peroxidase conjugated rabbit anti-human IgG secondary antibody (Invitrogen, CA, USA) for 1 h at room temperature. ELISA was developed with tetramethylbenzidine solution (Biolegend, California, USA), and the reaction was stopped with 25% sulfuric acid. Optical density (OD) was measured at 450 nm in a Multiscan ELISA reader. Samples were considered positive by ELISA when they had a ΔOD higher than mean healthy control ΔOD plus two standard deviations.
Cell-based assays were used as previously described [25] as a second confirmatory technique for questionable cases. Briefly, mammalian expression vectors encoding human NF140, NF186, NF155, CNTN1 or CASPR1 were transfected into HEK293 cells using Lipofectamine 2000 (Invitrogen). Cells were then fixed with 4% paraformaldehyde and blocked. ICC experiments were performed using patient’s sera and appropriate primary and secondary antibodies.
Testing for anti-ganglioside antibodies
Patients’ sera were screened for the presence of anti-ganglioside antibodies using a previously validated ELISA protocol [26] as the general detection method, and thin layer chromatography [27] for confirmatory experiments. Anti-ganglioside antibodies were considered positive at a 1/1000 titre.
Rat dorsal root ganglia neurons immunocytochemistry
DRG were dissected from E16 rat embryos, dissociated and plated in glass coverslips coated with laminin (Invitrogen) and poly-d-lysine (Sigma, MO, USA). Cells were grown in neurobasal medium (Gibco BRL, NY, USA) supplemented with B27 (Gibco), Glutamax (Gibco) and nerve growth factor (NGF) (Invitrogen). After 24 h, cytosine arabinoside (ARA-C) (Sigma) and fluorouracil (5-FU) (Sigma) were added to the medium to remove fibroblasts. Then, medium was replaced every other day until reaching complete growth and differentiation of DRG neurons.
Live DRG neurons were incubated for 1 h with patients’ sera diluted 1/100 (for IgG experiments) or 1/40 (for IgM experiments) in culture medium at 37 °C. Cells were then fixed for 10 min with 4% PFA and incubated with secondary antibodies. Goat anti-human IgG or IgM AF488 (Molecular probes, Oregon, USA) were used as secondary antibodies at 1/1000 concentration.
Coverslips were mounted with Vectashield with DAPI and fluorescence signal intensity was scored in a 0–3 scale by two independent researchers. Images were obtained with an Olympus BX51 Fluorescence Microscope (Olympus Corporation, Tokyo, Japan).
Human neuroblastoma-derived neurons immunocytochemistry
SH-SY5Y cells were plated in glass coverslips coated with laminin at 2.5 µg/ml (Invitrogen). Cells were grown in proliferation medium containing DMEM/F12 (1:1), fetal bovine serum (10%), l-glutamine (1%) and Sodium pyruvate (1%). After 24 h, proliferation medium was replaced by differentiation medium containing Neurobasal (Gibco) supplemented with B27 (Gibco), Glutamax (Gibco), nerve growth factor (Invitrogen) and retinoic acid at 10 µM (Sigma). Then, medium was replaced every other day until full differentiation was achieved. On days 5 or 6 of differentiation, cells were fixed for 15 min with paraformaldehyde 4%; and blocked with 5% normal goat serum in PBS; followed by incubation with patients’ sera at 1/40 (for IgM) or 1/100 (for IgG). To observe the correct differentiation of the cells we also incubated them with chicken anti-panNeurofascin mAb (R&Dsystems) at 1/200. Goat anti-chicken IgG AF594 and goat anti-human IgG AF488 or goat anti-human IgM AF488 (Molecular Probes) were used as secondary antibodies at 1/1000 concentration.
Coverslips were mounted with Vectashield with DAPI and fluorescence signal intensity was scored in a 0–3 scale by two independent researchers. Images were obtained with an Olympus BX51 Fluorescence Microscope.
Peripheral nerve immunohistochemistry
Macaque peripheral nerve tissue slides (Inova Diagnostics, Inc., San Diego, CA) were blocked with 5% normal goat serum in PBS; followed by incubation with patients’ sera at 1:10 (for IgM) or 1:20 (for IgG). Monkey-adsorbed goat anti-human IgG AF488 (Southern Biotech, Alabama, US) or goat anti-human IgM AF488 (Molecular Probes) were used as secondary antibodies at 1/500 concentration. Finally, slides were mounted with Fluoromount medium (Sigma) and examined by two independent observers. Immunostaining patterns were analysed scoring fluorescence signal intensity of each nerve structure in a 0–3 scale. The nerve structures analysed were: nodes or paranodes, myelin from small myelinated fibres, myelin from large myelinated fibres, Schwann cells from unmyelinated fibres, large-fibre axons, and small-fibre axons. Reactivity against other non-nerve structures (fibroblasts, connective tissue, vessels) was not considered in the analysis.
To further study the staining patterns, peripheral nerve tissue slides were coated with mouse anti-human CD56 antibody (Becton Dickinson, New Jersey, USA) at 1:50 to stain non-myelinating Schwann cells (Remak bundles); or with rabbit anti-human S100 antibody (Abcam, Cambridge, UK) at 1:50 to stain myelinating Schwann cells. Goat anti-mouse IgG AF594 (for CD56), and goat anti-rabbit IgG AF594 (for S100) were used as secondary antibodies at 1/500 concentration.
Images were acquired using Leica TSC SP5 confocal microscope.
Statistical analysis
Results were analysed by GraphPad Prism v8.0 (GraphPad Software). Statistical comparison of proportions among groups was performed using contingency analysis with the application of Chi-square and a two-tailed Fisher’s exact test, accepting an alpha-level < 0.05 for statistical significance. To represent the results and perform hierarchical clustering of the results heatmap diagrams using the Clustvis web tool were performed [28].
To investigate the association between anti-ganglioside antibodies and prognosis a multivariable logistic regression analysis to predict the inability to walk at 6 months and at 1 year of follow-up (GDS ≥ 3) was performed using the STATA software. A stepwise backward regression modelling to select variables independently associated with the outcome was performed first. The variables introduced in our initial multivariable models were selected based on known prognostic factors: age, initial GDS, diarrhoea, AMAN, serum NfL levels (analysed in a previous study with the same cohort) [29], serum anti-GM1 IgG antibodies and serum anti-GD1a IgG antibodies [29,30,31,32]. To perform the multivariable analysis patients with MFS were excluded, because our aim was to predict GBS prognosis and MFS is considered a different disease, including different pathophysiology, clinical presentation (it does not present with weakness), treatment (often untreated) and outcome (considered self-limiting and benign). Finally, the ability of the variable “presence of anti-GM1 IgG antibodies” to predict the inability to run at 1 year of follow-up (GDS ≥ 2) was evaluated in our previously reported multivariable logistic regression analysis [29].
Odds-ratios (OR) for the logistic regression analysis were reported with 95% confidence intervals and p values.