Fig. 6: TGFβ signaling instructs macrophage localization but not their mobility within spinal EAE lesions.

a, Unbiased gene-set analysis of GO biological process gene sets downregulated by cytokine receptor KOs (see Methods for details on the calculation of the KO-specific differentially regulated genes). GO terms related to cell migration or adhesion are highlighted in bold. For statistical analysis, gprofiler analysis was used (Methods). b, Left, expression of the transcriptional signature ‘Adhesion & Migration’ in the Hoxb8FL-derived control cells; right, phenotype of the KOs in control Hoxb8FL-derived cells for this signature across clusters (Ma_Cx_1 to Ma_8) and overall (All) compared to the control cells. The shading/lines represent the distribution of log₂(fold change) of the KO compared to the control for all the genes that make up the signature, with the thick line being the median (quantile 50) log₂(fold change), and the thinner overlaid shadows representing the lines of the adjacent quantiles: 40th and 60th quantiles of the log₂(fold change) delimit the darkest shadow flanking the median line, in steps of 10 until the 10th and 90th quantiles delimiting the lightest shadow at the edges. For statistical analysis, GSEA was used (Methods). c, Scheme of the experimental approach to study the migration phenotype of the Tgfbr1-KO myeloid cells. A mix of tdTomato⁺ control cells and eGFP⁺ KO Hoxb8FL cells was transferred into an immunized mouse before disease onset; at the peak of disease the animal was perfused, and then the spinal column was decalcified and cut coronally. d, Representative confocal microscopy image of a cross-section of the whole spinal column of an EAE animal co-transferred with control (red) and Tgfbr1-KO (green) cells. Higher-magnification image on the right shows myeloid cell infiltration in the pia delimited by laminin staining, and in adjacent parenchymal lesions. Scale bars, 50 μm. e, Quantification of the distribution of the Hoxb8FL-derived cytokine receptor KO myeloid cells relative to the control cells in the same animal across compartments. N = 4 animals for Ifngr1-KO, n = 3 animals for Tnfrsf1a-KO and n = 5 animals for Csf2ra-KO and Tgfbr1-KO. f, Scheme of the experiment for tracking the movement of Hoxb8FL-derived myeloid cell in neuroinflammatory lesions (left); time-lapse images showing cell movement along the midline vein in the dorsal spinal cord at peak of EAE of control-edited and Tgfbr1-KO Hoxb8FL-derived myeloid cells (middle); and quantification of movement parameters (right). Filled white arrowheads indicate control cells, outlined white arrowheads indicate Tgfbr1-KO cells, and lines denote cell tracks. Scale bars, 50 μm (overview image) and 10 μm (insets). Pathways shown in a include only GO_BP and GO_MF terms, with an adjusted P value < 0.01, an intersection size ≥ 5 and term size < 2,000; b, KO names marked by asterisks at the top of the plots and black ‘All’ labels indicate that the phenotype of the signature for that KO versus control in the global Hoxb8FL-derived cells is significant. Blue indicates downregulated in the KO compared to the control. Asterisks and black cluster name labels indicate the phenotype is significant for that cluster for the Hoxb8FL-derived KO versus the control. Significance was determined with the GSEA pathway analysis algorithm for a NOM P value < 0.05 or an FDR q value < 0.25 and absolute NES > 1.5 (Methods); e, Two-tailed paired t-test or Wilcoxon test for KO versus control for pia and for parenchyma separately; data are shown as a ratio for convenience. f, Two-tailed paired t-test or Wilcoxon test. NS, P value > 0.05; *P value < 0.05, **P value < 0.01, ***P value < 0.001, ****P value < 0.0001; bars show the mean.

Source data