Neutrophil-microglia interaction drives motor dysfunction in a neuromyelitis optica model induced by subarachnoid AQP4-IgG
Fangfang Qi, Vanda A. Lennon, Shunyi Zhao, Yong Guo, Husheng Ding, Caiyun Liu, Whitney M. Bartley, Tingjun Chen, Claudia F. Lucchinetti, Long-Jun Wu
Fangfang Qi, Vanda A. Lennon, Shunyi Zhao, Yong Guo, Husheng Ding, Caiyun Liu, Whitney M. Bartley, Tingjun Chen, Claudia F. Lucchinetti, Long-Jun Wu
View: Text | PDF
Research Article Autoimmunity Neuroscience

Neutrophil-microglia interaction drives motor dysfunction in a neuromyelitis optica model induced by subarachnoid AQP4-IgG

Abstract

Neutrophils and neutrophil extracellular traps (NETs) contribute to early neuromyelitis optica (NMO) histopathology initiated by IgG targeting astrocytic aquaporin-4 (AQP4) water channels. Yet, the mechanisms underlying neutrophil recruitment and their pathogenic roles in disease progression remain unclear. To investigate molecular-cellular events preceding classical complement cascade activation in a mouse NMO model, we continuously infused, via spinal subarachnoid route, a non-complement-activating mouse monoclonal AQP4-IgG. Parenchymal infiltration of netting neutrophils containing C5a ensued with microglial activation and motor impairment but no blood-brain barrier leakage. Motor impairment and neuronal dysfunction both reversed when AQP4-IgG infusion stopped. Two-photon microscopy and electron microscopy–based reconstructions revealed physical interaction of infiltrating neutrophils with microglia. Ablation of either peripheral neutrophils or microglia attenuated the motor deficit, highlighting their synergistic pathogenic roles. Of note, mice lacking complement receptor C5aR1 exhibited reduction in neutrophil infiltration, microglial lysosomal activation, neuronal lipid droplet burden, and motor impairment. Pharmacological inhibition of C5aR1 recapitulated this protection. Immunohistochemical analysis of an NMO patient’s spinal cord revealed disease-associated microglia surrounding motor neurons in nondestructive lesions. Our study identifies neutrophil-derived C5a signaling through microglial C5aR1 as a key early driver of reversible motor neuron dysfunction in the precytolytic phase of NMO.

Authors

Fangfang Qi, Vanda A. Lennon, Shunyi Zhao, Yong Guo, Husheng Ding, Caiyun Liu, Whitney M. Bartley, Tingjun Chen, Claudia F. Lucchinetti, Long-Jun Wu

×

Figure 4

Neutrophil-derived C5a enhances NET production and activates microglia via C5aR1 signaling.

Options: View larger image (or click on image) Download as PowerPoint
Neutrophil-derived C5a enhances NET production and activates microglia v...
(A and B) Gray matter of AQP4-IgG recipient mice, C5a immunoreactivity was not seen in GFAP+ astrocytes (A) and rarely seen in IBA1+ microglia/macrophages (B). (C) Triple immunostaining of lumbar cord of AQP4-IgG recipient mice revealed C5a in neutrophils (MPO+, left) and monocytes (CCR2-GFP+, enhanced by anti-GFP-IgG, right). (D) The percentage of spinal cord cells expressing C5a in AQP4-IgG-infused mice (n = 4 mice per group). (E) Colocalization analysis of C5a and MPO signals in C using Zen software. (F and G) Confocal images of neutrophil (Ly6G, red) containing C5a (green in F), and C5 (green in G) after TNF-α incubation in vitro. Granule protein, MPO+, magenta; nuclear segmentation, DAPI+, blue. BF, bright-field. (H) Lumbar cords from IgG-infused mice were dissociated enzymatically and subjected to high parametric flow cytometric analysis. (I) t-SNE maps identifying CD11b+, Ly6G+, Cx3cr1+, and C5aR1+ cells and their expression levels among CD45+ immune cells. (J) Public database (Brain RNA-Seq) documents that C5ar1 mRNA in normal mouse brain is predominantly expressed in microglia/macrophages. (K) Flow cytometric plot shows that neutrophils infiltrate the lumbar cord in WT mice and in C5ar1-deficient mice at day 3. (L) Quantification of the percentage of neutrophils (CD45+CD11b+Ly6G+Ly6C– cells among CD45+CD11b+ cells) in K (n = 4 mice per group). (M) Motor function, assessed by rotarod performance (latency to fall), in WT and C5ar1–/– mice infused with AQP4-IgG or (only C5ar1–/– mice) normal mouse IgG (0.1 μg/μL, n = 6 mice per group). (N) Motor function of mice assessed as rotarod fall latency; treatment: F(1, 6) = 9.961, P = 0.0197; n = 6–7 mice per group. Statistics: D used Tukey’s post hoc multiple comparisons test in (1-way ANOVA); t test in J; 2-way repeated measures ANOVA with Holm-Šídák post hoc test in M and N.