Systemic inhibition of the membrane attack complex impedes neuroinflammation in chronic relapsing experimental autoimmune encephalomyelitis

Abstract

The complement system is a key driver of neuroinflammation. Activation of complement by all pathways, results in the formation of the anaphylatoxin C5a and the membrane attack complex (MAC). Both initiate pro-inflammatory responses which can contribute to neurological disease. In this study, we delineate the specific roles of C5a receptor signaling and MAC formation during the progression of experimental autoimmune encephalomyelitis (EAE)-mediated neuroinflammation. MAC inhibition was achieved by subcutaneous administration of an antisense oligonucleotide specifically targeting murine C6 mRNA (5 mg/kg). The C5a receptor 1 (C5aR1) was inhibited with the C5a receptor antagonist PMX205 (1.5 mg/kg). Both treatments were administered systemically and started after disease onset, at the symptomatic phase when lymphocytes are activated. We found that antisense-mediated knockdown of C6 expression outside the central nervous system prevented relapse of disease by impeding the activation of parenchymal neuroinflammatory responses, including the Nod-like receptor protein 3 (NLRP3) inflammasome. Furthermore, C6 antisense-mediated MAC inhibition protected from relapse-induced axonal and synaptic damage. In contrast, inhibition of C5aR1-mediated inflammation diminished expression of major pro-inflammatory mediators, but unlike C6 inhibition, it did not stop progression of neurological disability completely. Our study suggests that MAC is a key driver of neuroinflammation in this model, thereby MAC inhibition might be a relevant treatment for chronic neuroinflammatory diseases.

Keywords: Complement; Inflammasome; Neuroinflammation.

Fig. 1

MAC inhibitor prevents relapse whereas, inhibitor of C5aR1-mediated inflammation ameliorates disability. Scheme illustrating inhibition of the terminal complement pathway a. at the level of C5b-C9 (MAC) by antisense targeting of complement C6 mRNA (C6 antisense, 5 mg/kg), b. at the level of C5aR-mediated inflammation by an antagonist of C5aR1 (PMX205, 1.5 mg/kg) (a). Clinical scores of mice with chronic relapsing EAE receiving C6 antisense (n = 13, in red) or PMX205 (n = 10, in green) or no drug (n = 17, in blue). The C6 antisense-treated mice did not show relapse or neurological deterioration post-relapse phase. In contrast, PMX205 did not stop progression of neurological disability completely. Differences between groups were analyzed by using the Kruskal-Wallis test by ranks. Data represent the average clinical scores (mean) ± SEM. Statistical differences are indicated (**p < 0.01, ****p < 0.0001). RL, relapse; RM, remission; ns, not significant (b). Heatmap showing Ingenuity Pathway Analysis (IPA) canonical immune pathways induced by EAE and significantly affected by treatment with the C6 antisense or the PMX205 inhibitor. Pathways are ranked according to the z-score that predicts activation (orange) /suppression (blue). PMX205 is a less efficient inhibitor of neuroinflammation compared with the C6 antisense. Data were obtained from RNA-seq of mouse spinal cords collected at relapse (3 mice/group) (c)

Fig. 2

MAC inhibitor completely blocks NLRP3 inflammasome components while, in contrast, C5aR1 inhibition only partially decreases expression. The inflammasome pathway was modeled by Ingenuity Pathway Analysis (IPA) after analyses of gene expression data from mouse spinal cords collected at relapse (3 mice/group). Genes showing higher expression levels in the first compared to the second group of each comparison are colored red. The grey color indicates no changes of expression levels. Note that genes linked to NLRP3 inflammasome activation show no changes of expression levels in the C6 antisense group (MAC inhibitor, genes colored grey), but increased levels in the PMX205 group (inhibitor of C5aR1-mediated inflammation, genes colored red) when compared with healthy controls at relapse (a-c). Log fold change values of genes are shown in Additional file 1: Table S4

Fig. 3

MAC inhibitor completely inhibits IL-1β synthesis while, in contrast, C5aR1 inhibition only partially decreases expression. Immunostaining for interleukin 1 beta (IL-1β)/pro-IL-1β, indicator of NLRP3 inflammasome activation, shows reactivity within the spinal cord of no drug mice (n = 4, arrows in a), on cells with a glial morphology (a, b). IL-1β/pro-IL-1β was never detected in the cords of C6 antisense-treated mice (n = 4) (c), while PMX205-treated mice showed sparse reactivity (n = 3) (arrows in d, e). Tissue was collected at relapse. Scale bars (a, c, d) 25 μm, (b, e) 5 μm. Hematoxylin was used as counterstain in (a-e)

Fig. 4

C6 antisense-mediated MAC inhibition prevents demyelination and microglia/macrophage activation in chronic relapsing EAE. Histological analysis of semi-serial paraffin sections of the mouse cervical spinal cord segment showed lesions in the no drug mice (n = 4) but no signs of active disease in the C6 antisense-treated mice (n = 6). Tissue was collected post-relapse phase. Staining for Luxol fast blue (LFB) and quantification of the signal (a-c) as well as staining for proteolipid protein (PLP) and quantification of the reactivity (d-f) indicated demyelination in the no drug mice (demyelinated lesions pointed by arrows in a and d), but no significant myelin loss in the C6 antisense-treated mice (c, f). Immunostaining for ionized calcium-binding adapter molecule 1 (IBA-1) (g-j) revealed microglia/macrophages within the lesions (arrows in g) and at the peri-lesional area of the no drug mice with a morphology consistent with an activated status. In contrast, all the C6 antisense-treated mice showed amounts of IBA-1+ reactivity similar to the ones found in controls and morphology of IBA-1+ microglia consistent with a resting status (h-j). Moreover, the no drug mice had abundant reactivity for C9, a marker of MAC deposition, within the ventral horns of the spinal cord. In contrast, the C6 antisense-treated mice showed only small amounts of MAC (k-n). Scale bars: (a, b, d, e, g, h, k, m) 200 μm, (i, l) 50 μm. Cresyl violet (CV) was used as counterstain in (a, b). Hematoxylin was used as counterstain in (d, e, g-i, k-m)

Fig. 5

C6 antisense-mediated MAC inhibition protects from axonal damage. Richardson’s staining on spinal cord tissue from no drug mice showed wide areas of axonal damage corresponding to the lesioned areas (a). Electron microscopy micrograph indicates the types of axonal damage detected and quantified: 1. destruction of the myelin sheathe, 2. deformation of the myelin sheathe with gaps between tangent layers, 3. infoldings of the myelin sheathe within the axoplasmic area, and 4. onion bulbs (b). Picture of Richardson’s stained spinal cord section and electron microscopy micrograph from C6 antisense-treated mice (c-e) showing rarely detected damaged axons (arrows in c, d and arrows in e). Graph shows the percentage of axonal damage as quantified in the ventral and lateral column of the cervical and thoracic spinal cord segments from control (n = 6), no drug (n = 4) or C6 antisense-treated (n = 6) mice, collected post-relapse. Comparisons between groups were performed using the Kruskal-Wallis test by ranks. Data are expressed as the average (mean) ± SEM. Statistical differences are indicated (***p < 0.001) (f). Pictures from control mice showed intact axons only (g, h). Scale bars: (a, c, g) 25 μm, (b, e) 2 μm, (d) 500 nm (h) 200 nm

Fig. 6

C6 antisense-mediated MAC inhibition protects from synaptic alterations. Quantification of synaptophysin (SYP) positive punctae showed a significant decrease of synaptic densities (*p < 0.05) in the no drug (n = 4) compared to the C6 antisense-treated mice (n = 6), post-relapse phase. Differences between groups were analyzed by using the One-Way Analysis of Variance test. Data are expressed as the average (mean) ± SEM. Statistical differences are indicated (*p < 0.05, ***p < 0.001). Controls are healthy mice injected with adjuvant only (n = 6) (a). Pearson’s correlation coefficient showed a significant negative correlation between SYP and the C9 marker of MAC (coefficient, r = – 0.74, ***p < 0.0001), in the mouse spinal cord (n = 29 corresponding fields plotted) (b). Histological analysis of paraffin spinal cord sections from no drug mice showed low densities of SYP+ punctae (arrow pointing to a synapse), sign of synaptic alterations or loss (c), abundant C9 reactivity (arrows) (d), and ionized calcium-binding adapter molecule 1 (IBA-1) positive microglia/macrophages with a morphology consistent with an activated status (e). In contrast, histological analysis of spinal cords from C6 antisense-treated mice showed abundant SYP+ punctae (arrows) (f), sparse C9 reactivity (arrows) (g), and IBA-1+ microglia with a morphology consistent with a resting status (h). Scale bars: (c-h) 10 μm. Hematoxylin was used as counterstain in (c-h)