Fig. 2: White matter astrocyte CCN1 restricts the spatiotemporal dynamics of microglia-mediated white matter inflammation after SCI.
From: Lesion-remote astrocytes govern microglia-mediated white matter repair
a, Non-negative matrix factorization (NMF) on deconvolved spatial transcriptomics data was used to identify co-occurring cell states across space and time. NMF loading weights plotted across healthy and post-injury time points predicts injury-reactive alterations in regionally co-occurring cell states. b, Spatial plotting of NMF 4 reveals enrichment in lesion ipsilateral white matter. NMF 4 is primarily composed of white matter LRAs (WM3/4) and Mg2/Mg5 microglia. c, Spatial transcriptomic localization of Mg2/Mg5 microglia and WM3/4 astrocytes illustrate overlapping unilateral white matter enrichment. d, Volcano plot of DEGs in Mg2/5 microglia versus homeostatic microglia (Mg1 and Mg3 (Mg1/3) (FDR P ≤ 0.05). FC, fold change. e,f, Dot plot of mean normalized expression of metabolism and function (e) and previously published microglia state-associated molecular signatures18,19,20,21,22,23,24,25,26 (f) in iSCI lesion-remote microglia clusters. g, WDM assemble into multicellular nodules that phagocytose myelin debris. Scale bars, 10 µm. h, NicheNet analysis of reactive WM3/4 LRAs (senders) and WDM (Mg2/5) (receivers) identified several putative ligands secreted by white matter astrocytes. i, Ccn1 is selectively expressed by LRAs neighbouring WDM nodules in degenerating white matter. Scale bar, 10 µm. j, Left, aligned average density plots of Ccn1-expressing astrocytes, illustrating regional and intraspinal relationships between Ccn1+ LRAs and anatomically defined zones of Wallerian degeneration. Right, quantification of Ccn1+ astrocytes per time point in each region. Cells were counted from cervical healthy (n = 5), thoracic healthy (n = 6), lumbar healthy (n = 4), cervical 3 dpi (n = 3), thoracic 3 dpi (n = 5), lumbar 3 dpi (n = 4), cervical 7 dpi (n = 5), thoracic 7 dpi (n = 4), lumbar 7 dpi (n = 5), cervical 14 dpi (n = 4), thoracic 14 dpi (n = 4), lumbar 14 dpi (n = 4), cervical 28 dpi (n = 4), thoracic 28 dpi (n = 8), lumbar 28 dpi (n = 6), cervical 90 dpi (n = 4), thoracic 90 dpi (n = 5) and lumbar 90 dpi (n = 4); 2–11 sections per mouse, region or time point. Scale bars, 250 µm. k, Schematic of iSCI lesion-remote spinal cord regions and expected patterns of Wallerian degeneration from which Ccn1+ astrocytes were evaluated. l, WDM nodules in Wallerian degenerating dorsal white matter (mid-thoracic) from wild-type (WT) and Ccn1-cKO spinal cord following iSCI. Yellow arrowheads indicate WDM in Wallerian degenerating regions; white arrowheads indicate spatially aberrant WDM nodules. Scale bars, 250 µm. m, Left, aligned average density plots of WDM nodules in Ccn1-cKO spinal cord after iSCI. Right, quantification of WDM nodules from lumbar, thoracic and cervical spinal cord of wild-type and Ccn1-cKO mice following iSCI. Statistical comparison between wild-type and Ccn1-cKO nodule counts by Conway–Maxwell–Poisson method and Tukey’s post hoc test. Nodules were counted from WT cervical 7 dpi (n = 5), Ccn1-cKO cervical 7 dpi (n = 5), WT thoracic 7 dpi (n = 4), Ccn1-cKO thoracic 7 dpi (n = 4), WT lumbar 7 dpi (n = 4), Ccn1-cKO lumbar 7 dpi (n = 4), WT cervical 28 dpi (n = 4), Ccn1-cKO cervical 28 dpi (n = 4), WT thoracic 28 dpi (n = 6), Ccn1-cKO thoracic 28 dpi (n = 5), WT lumbar 28 dpi (n = 6), Ccn1-cKO lumbar 28 dpi (n = 4), WT cervical 90 dpi (n = 5), Ccn1-cKO cervical 90 dpi (n = 4), WT thoracic 90 dpi (n = 5), Ccn1-cKO thoracic 90 dpi (n = 5), WT lumbar 90 dpi (n = 4), Ccn1-cKO lumbar 90 dpi (n = 5), 2–11 sections per mouse, genotype region or time point. Graphs show mean ± s.e.m. In graphs of histological counts or continuous data, coloured data points represent the mean value for each biological replicate (individual mouse); grey data points indicate replicate measurements from individual tissue sections. Unless stated otherwise, statistical analyses were performed using two-way ANOVA on mean values from biological replicates with Holm–Sidak’s post hoc test. Scale bars, 250 µm. *P ≤ 0.05, **P ≤ 0.002, ***P ≤ 0.0002, ****P ≤ 0.0001; NS, not significant. Exact P values are provided in the Source Data.
