C1q-targeted inhibition of the classical complement pathway prevents injury in a novel mouse model of acute motor axonal neuropathy

Abstract

Introduction: Guillain-Barré syndrome (GBS) is an autoimmune disease that results in acute paralysis through inflammatory attack on peripheral nerves, and currently has limited, non-specific treatment options. The pathogenesis of the acute motor axonal neuropathy (AMAN) variant is mediated by complement-fixing anti-ganglioside antibodies that directly bind and injure the axon at sites of vulnerability such as nodes of Ranvier and nerve terminals. Consequently, the complement cascade is an attractive target to reduce disease severity. Recently, C5 complement component inhibitors that block the formation of the membrane attack complex and subsequent downstream injury have been shown to be efficacious in an in vivo anti-GQ1b antibody-mediated mouse model of the GBS variant Miller Fisher syndrome (MFS). However, since gangliosides are widely expressed in neurons and glial cells, injury in this model was not targeted exclusively to the axon and there are currently no pure mouse models for AMAN. Additionally, C5 inhibition does not prevent the production of early complement fragments such as C3a and C3b that can be deleterious via their known role in immune cell and macrophage recruitment to sites of neuronal damage.

Results and conclusions: In this study, we first developed a new in vivo transgenic mouse model of AMAN using mice that express complex gangliosides exclusively in neurons, thereby enabling specific targeting of axons with anti-ganglioside antibodies. Secondly, we have evaluated the efficacy of a novel anti-C1q antibody (M1) that blocks initiation of the classical complement cascade, in both the newly developed anti-GM1 antibody-mediated AMAN model and our established MFS model in vivo. Anti-C1q monoclonal antibody treatment attenuated complement cascade activation and deposition, reduced immune cell recruitment and axonal injury, in both mouse models of GBS, along with improvement in respiratory function. These results demonstrate that neutralising C1q function attenuates injury with a consequent neuroprotective effect in acute GBS models and promises to be a useful new target for human therapy.

Fig. 1

Comparison of ganglioside expression and localisation in wild type and novel GalNAcT^−/−(Tg-neuronal) mice that exclusively express complex gangliosides in neurons. a Enzyme activity assays showed a restoration of GalNAcT activity to ~50 % in 2 different GalNAcT ^−/− -Tg(neuronal) lines compared to wild type (n = 3/genotype). The dashed line represents the threshold level of activity for the assay therefore GalNAcT ^−/− falls below this and is essentially zero despite minimal value. b Ganglioside biosynthetic pathway. The GalNAcT enzyme is necessary for generation of complex gangliosides (surrounded by the green box). c Brain extracts from wild type, GalNAcT ^−/− and 2 strains of GalNAcT ^−/− -Tg(neuronal) were probed with anti-GM1 IgG antibody, DG2. This antibody bound all genotype extracts except for GalNAcT ^−/− which lack complex gangliosides including GM1. GM1 lipid was printed on the left and provided a positive control for anti-GM1 antibody binding. d Axonal binding can be observed in the GalNAcT ^−/− -Tg(neuronal) mice treated with anti-GM1 antibody (arrows), while it bound to the terminal kranocyte (asterisk) in wild type mice and was absent in GalNAcT^−/− tissue. Anti-GQ1b/GD3 antibody bound similarly in both strains along the axons and on the perisynaptic Schwann cell membranes (arrowheads) that are simple ganglioside GD3 positive. Scale bar = 10 μm

Fig. 2

C1q neutralisation attenuates injury in a mouse MFS model. a A significant reduction in tidal volume is shown in mice treated with isotype control mAb (n = 3) compared to M1 (n = 4) antibody (p < 0.05). Representative flow-charts from the plethysmography recordings are shown for each treatment group at 6 h post-NHS treatment. Bars represent mean ± SEM. b Top panels: Illustrative images show MAC (orange) and C3c (green) deposited at nerve terminals (identified by α-BTx, red; CFP-positive axons, blue) from mice treated with control mAb, while this staining is absent from those treated with anti-C1q antibody. Lower panels: Anti-ganglioside antibody (orange) is present at terminals from both treatment groups, but neurofilament immunostaining (green) is only present at those terminals from mice treated with anti-C1q antibody. c The early and end-stage complement products C3c and MAC, respectively, showed significantly greater deposits at control mAb treated mice (p < 0.05, p < 0.05, respectively) than anti-C1q antibody treated mice (n = 3 control mAb, n = 4 anti-C1q antibody) nerve terminals. Axonal integrity was a measure of neurofilament immunostaining overlying the endplate. Axonal integrity was significantly more intact at anti-C1q antibody protected mice nerve terminals compared to the control mAb group (p < 0.05). Box and whisker plots represent the spread of all data points per condition and significance was based on Mann–Whitney statistical analysis of the median from each animal per treatment. * p < 0.05, unpaired student t-test (a), Mann–Whitney test (b). Scale bar = 20 μm. AGAb = anti-ganglioside antibody, nAChR = nicotinic acetylcholine receptor

Fig. 3

C1q neutralisation attenuates injury in a novel mouse model of AMAN. a Tidal volume and respiratory rate decreased compared to baseline measurements in both treatment groups. Tidal volume and respiratory rate reached a significant reduction in mice treated with isotype control mAb (p < 0.01, p < 0.05, respectively) compared to anti-C1q antibody (n = 5/group). Representative flow-charts from the plethysmography recordings are shown for each treatment group at baseline and 6 h post-NHS treatment. b Behavioural tests showed isotype control mAb treated mice (n = 5) spent significantly less time on the accelerating rotarod than did mice treated with anti-C1q antibody (p < 0.001). There was no significant difference in grip strength between groups. c Top panels: Illustrative images show MAC (orange) and C3c (green) deposited at nerve terminals (nAChR’s identified by α-BTx, red; CFP-positive axons, blue) from mice treated with control mAb, while this staining is absent from those treated with anti-C1q antibody. Lower panels: Anti-ganglioside antibody (orange) is present at terminals from both treatment groups, but neurofilament immunostaining (green) is only present at those terminals from mice treated with anti-C1q antibody. d The diaphragm nerve terminals were identified by fluorescently labelled α-BTx. Nerve terminals were immunohistochemically assessed for complement deposits and axonal integrity. The early and end-stage complement products C3c and MAC, respectively, showed significantly greater deposits at control mAb treated mice than anti-C1q antibody treated mice nerve terminals (p < 0.01). Axonal integrity was a measure of percentage “normal” neurofilament immunostaining overlying the endplate. Axonal integrity was significantly more intact at anti-C1q antibody protected mice nerve terminals compared to the isotype control mAb group. Box and whisker plots represent the spread of all data points and significance was based on Mann–Whitney statistical analysis of the median from each animal per treatment. * p < 0.05; ** p < 0.01; *** p < 0.001, unpaired student t-test (a,b), Mann–Whitney test (c). Scale bar = 20 μm. AGAb = anti-ganglioside antibody, nAChR = nicotinic acetylcholine receptor

Fig. 4

Immune cell infiltration in acute MFS and AMAN models. a In the MFS model of disease, there was significantly less CD11b positive cells per field of view (FOV) within the proximity of the NMJ in animals who were treated with anti-C1q antibody. b No significance in NIMP-R positive cells existed between the two groups, despite a trend towards lower numbers in the anti-C1q antibody treated group. c In the AMAN model of disease, there was no significant difference between anti-C1q antibody treated and control mAb-treated animals in either CD11b or (d) NIMP-R14 positive cells. Neutrophils appeared to be absent within the proximity of the NMJ in this model. Bars represent mean ± SEM, *p < 0.05, unpaired student t-test. Scale bars = 20 μm. nAChR = nicotinic acetylcholine receptor