Microglial neuropilin-1 promotes oligodendrocyte expansion during development and remyelination by trans-activating platelet-derived growth factor receptor

Abstract

Nerve-glia (NG2) glia or oligodendrocyte precursor cells (OPCs) are distributed throughout the gray and white matter and generate myelinating cells. OPCs in white matter proliferate more than those in gray matter in response to platelet-derived growth factor AA (PDGF AA), despite similar levels of its alpha receptor (PDGFRα) on their surface. Here we show that the type 1 integral membrane protein neuropilin-1 (Nrp1) is expressed not on OPCs but on amoeboid and activated microglia in white but not gray matter in an age- and activity-dependent manner. Microglia-specific deletion of Nrp1 compromised developmental OPC proliferation in white matter as well as OPC expansion and subsequent myelin repair after acute demyelination. Exogenous Nrp1 increased PDGF AA-induced OPC proliferation and PDGFRα phosphorylation on dissociated OPCs, most prominently in the presence of suboptimum concentrations of PDGF AA. These findings uncover a mechanism of regulating oligodendrocyte lineage cell density that involves trans-activation of PDGFRα on OPCs via Nrp1 expressed by adjacent microglia.

Fig. 1. Anti-Nrp1 antibody blocks PDGF AA-mediated OPC proliferation in white but not gray matter.

A Schematic showing slice cultures from P8 NG2cre;Z/EG mice to assay for OPC proliferation in response to 50 ng/mL PDGF AA and different concentrations of anti-Nrp1 antibody. B–E Slice cultures from P8 NG2cre;Z/EG mice that were fixed and labeled for EGFP (green) and EdU (magenta) in the cortex (B and C) or corpus callosum (D and E) in the presence of 50 ng/mL PDGF AA in the absence (B and D) or presence (C and E) of 1 µg/mL anti-Nrp1 antibody. Arrows, examples of EGFP+ EdU+ cells. Scale bars, 50 µm. F Quantification of OPC proliferation in slice cultures after 5 h of EdU labeling. y-axis, proportion of EGFP + cells that were EdU+. Tukey’s multiple comparisons test, n = 4, F(4, 30) = 9.966, means ± standard deviations. The differences between OPC proliferation in gray and white matter were significant at anti-Nrp1 antibody concentrations of 0 (p < 0.0001) and 0.01 µg/mL (p = 0.0065), but not at higher concentrations of the antibody. Black, gray matter; gray, white matter. Two-way ANOVA, Tukey’s multiple comparisons test, F(1, 30) = 89.71, n = 4. G, H Slice cultures treated with 1 µg/mL control goat IgG (G) or goat anti-Nrp1 antibody (H) and labeled for EGFP (green) and TUNEL (magenta). Arrows indicate a TUNEL+ cell. Scale bars, 50 µm. Source data are provided as a Source Data file.

Fig. 2. Nrp1 expression in the developing postnatal brain.

A–C Low-magnification view of a coronal forebrain section labeled for Nrp1 (red) and F4/80 (green). Arrowheads, examples of Nrp1+ blood vessels. Arrowheads, blood vessels. Boxed area with a cluster of Nrp1+ cells is shown in D–F. Scale, 100 µm. ctx, cortex; cc, corpus callosum; LV lateral ventricle. D–F High-magnification view showing Nrp1+ cells in the corpus callosum express F4/80. Arrows in F indicate blood vessels that are F4/80-negative. Scale, 50 µm. G–I Double labeling for Nrp1 (red) and PDGFRα in P5 corpus callosum. Nrp1 is not expressed on PDGFRα + OPCs, but Nrp1+ processes are found close to OPC cell bodies and processes (arrows in I). J–L Double labeling for Nrp1 (red) and YFP (green) in Cx3CR1^{creERT2-ires-EYFP} mice from P5 through P14. Scale = 50 μm. M The proportion of EYFP + microglia that expressed Nrp1 in the corpus callosum. One-way ANOVA, Tukey’s multiple comparisons test, n = 3, F(2, 6) = 438.8. Source data are provided as a Source Data file.

Fig. 3. Effects of mg-Nrp1 cko on OPC proliferation in the white matter.

A Schematic of generating microglia-specific Nrp1 knock out (mg-Nrp1-cko) and the heterozygous control (mg-Nrp1-cont). B Schematic of tissue analysis at P5 and P14 after cre induction at P2-3. C, D. 3D images of P5 corpus callosum labeled for Nrp1 (red), EYFP (green), PDGFRα (light blue), and EdU (pink). Scale, 20 μm. Arrows, examples of EdU+ PDGFRα+ cells in contact with EYFP + microglia. E Quantification of proliferation of OPCs in mg-Nrp1-cont (black) and cko (blue). E1 Total EdU+ OPC density in P5 corpus callosum. Unpaired Student’s t-test, n = 5, t = 7.171, df = 8. E2 The density of EdU+ OPCs in cont or cko with (circles) or without (triangles) contact with microglia. Two-way ANOVA, Sidak’s multiple comparisons test, F(1, 16) = 53.41, p < 0.0001, n = 5. F Proportion of PDGFRα+ OPCs that were in contact with EYFP + microglia. Unpaired Student’s t-test, n = 5, t = 3.940, df = 8. G Proportion of OPCs that were EdU+ in P5 cerebellum, P5 corpus callosum, and P14 corpus callosum. Two-way ANOVA, Sidak’s multiple comparisons test. F(1, 13) = 32.44, p < 0.0001, n = 3–4. H Quantification of PDGFRα+ OPC density in P5 and P14 corpus callosum. Two-way ANOVA, Sidak’s multiple comparisons test. F(1, 12) = 19.07, p = 0.0009, n = 5 (P5), n = 3 (P14). I Quantification of MBP immunofluorescence intensity in P14 corpus callosum of cont and cko mice. n = 3, t = 0.1896, df = 4. Black symbols, mg-Nrp1-cont; blue symbols, mg-Nrp1-cko in E–I. Source data are provided as a Source Data file.

Fig. 4. Effects of mg-Nrp1-cko on OPC proliferation and myelin repair after demyelination.

A Schematic of demyelination experiments. B, C Nrp1 (magenta) upregulation on EYFP+ microglia/macrophages 3 days after LPC injection but not after PBS injection. Scale, 50 µm. D, E Similar extent of CD68 (blue) expression and infiltration of activated microglia/macrophages into demyelinated lesion at 7 dpl in mg-Nrp1-cont (D) and mg-Nrp1-cko (E) mice. Red, Nrp1; green, EYFP. Arrows, Nrp1+ CD68+ macrophages; arrowheads, Nrp1-negative CD68+ cells. Scale, 50 µm. F Quantification of the density of EYFP+ microglia in the lesion at 3 dpl. Unpaired Student’s t-test, t = 0.1678, df = 8, n = 5. G–J Labeling for EdU (blue), NG2 (red), and EYFP (green) in mg-Nrp1-cont and cko mice sacrificed at 3 dpl after EdU pulse labeling. Arrowheads, NG2+ EdU+ OPCs. Scale, 25 µm. K Quantification of proliferating OPCs at 3 and 7 dpl in cont (black) and cko (blue) mice showing significantly lower extent of EdU incorporation into OPCs in mg-Nrp1-cko mice. OPC proliferation was higher at 3 dpl compared to 7 dpl for both genotypes (p = 0.0007 for cont and p = 0.0351 for cko). Two-way ANOVA, Tukey’s multiple comparisons test, F(1, 16) = 54.70 for comparison between fl/+ and fl/fl, F(1, 16) = 32.52 for comparison between 3 and 7 dpl, n = 5. Black circles, mg-Nrp1-cont; blue squares, mg-Nrp1-cko in F and K. Source data are provided as a Source Data file.

Fig. 5. The extent of oligodendrocyte regeneration and myelin repair in mg-Nrp1-cko.

A, B Labeling mg-Nrp1-cont (L) and cko (M) for CC1 (red), DAPI (blue), and EYFP (green) at 14 dpl. Arrowheads, CC1+ cells. Scale, 25 µm. C Quantification of OL density in cont and cko lesions at 14 dpl. Student’s t-test, unpaired, two-tailed, n = 5, t = 4.753, df = 8. D, E Labeling mg-Nrp1-cont (O) and cko (P) for MBP (blue) and non-phosphorylated neurofilaments (non-P-NF, red) using smi-32 antibody to identify axons that have not yet undergone complete remyelination. Dotted lines indicate the extent of partial or complete demyelination. Scale, 50 μm. LV, lateral ventricle. F Quantification of demyelinated area in cont and cko lesions at 14 dpl. Student’s t-test, unpaired, two-tailed, n = 5, t = 3.490, df = 8. G, H Labeling for MBP (blue) and non-phosphorylated neurofilaments (non-P NF, red) in the lesioned corpus callosum of mg-Nrp1-cont (G) and cko (H) mice at 28 dpl. I Quantification of MBP immunofluorescence in the lesioned corpus callosum of mg-Nrp1-cont and cko mice at 28 dpl. Student’s t-test, unpaired, two-tailed, n = 5, t = 4.031, df = 8. Black circles, mg-Nrp1-cont; blue squares, mg-Nrp1-cko in C, F, and I. J, K Electron microscopic images of cross-sections of LPC-lesioned corpus callosum at 28 dpl. J mg-Nrp1-cont (fl/+) showing sheaths of various thickness (arrows), but most axons appear fully myelinated. K mg-Nr1p-cko (fl/fl) showing more axons without myelin sheaths (arrowheads), axons with thin myelin (arrow), and abundant swollen glial cell processes (*). Scale bars 1 μm. Source data are provided as a Source Data file.

Fig. 6. Effects of Nrp1-Fc on OPC proliferation in slice cultures from P8 NG2cre;Z/EG mice.

A–F Labeling for EGFP (green) and EdU (magenta) in the cortex (A–C) and corpus callosum (D–F) of slices treated with 50 ng/mL PDGF AA only (A, D), 50 ng/mL PDGF and 2 μg/mL Nrp1-Fc (B, E), or 50 ng/mL PDGF AA, 2 μg/mL Nrp1-Fc, and 1 µg/mL goat anti-mouse PDGFRα function-blocking antibody (C, F). Scale, 50 µm. G Dose-response of OPC proliferation in the cortex and corpus callosum to exogenous Nrp1-Fc in the presence of 50 ng/mL PDGF AA. Values are the percentages of EGFP+ cells that were EdU+. Two-way ANOVA, Tukey’s multiple comparisons test, n = 3, F(4, 20) = 5.622 for comparisons among different Nrp1-Fc concentrations and F(4, 20) = 317.6 for gray versus white matter comparison. Black, cortex; gray, corpus callosum. H The effects of anti-PDGFRα blocking antibody on PDGF AA-mediated OPC proliferation in the cortex (gray bars) and corpus callosum (white bars) in the presence (checkered bars) or absence (solid bars) of exogenous Nrp1-Fc. All samples were treated with 50 ng/mL PDGF AA. Two-way ANOVA, Sidak’s multiple comparisons test, n = 4, F(3, 16) = 266.6. Circles, no blocking antibody; squares, PDGFRα blocking antibody. Black and gray, no Nrp1-Fc; blue, 2 μg/mL Nrp1-Fc. Source data are provided as a Source Data file.

Fig. 7. Effects of Nrp1-Fc on dissociated cultures of OPCs.

A–D Immunolabeling for Olig2 (green), NG2 (red), and EdU (gray) of cells grown in no PDGF AA (A, B) or 15 µg/mL PDGF AA in the absence (A, C) or presence of 2 µg/mL Nrp1-Fc (B, D). Scale in A, 50 µm. Arrows: EdU+ Olig2+ NG2+ cells. E Quantification of the effects of combination of PDGF AA and Nrp1-Fc on OPC proliferation. Two-way ANOVA, Sidak’s multiple comparisons test, n = 4, F(1, 30) = 36.29. F Immunoblots of OPCs treated with 15 ng/mL of PDGF AA alone or 15 ng/mL of PDGF AA and 2 µg/mL Nrp1-Fc immunostained with antibody to total PDGFRα (left, red) or phosphorylated PDGFRα (right, green). Arrows indicate the expected bands for PDGFRα and phosphorylated PDGFRα. G Quantification of the intensity of the phosphorylated PDGFRα band relative to total PDGFRα bands. **p = 0.0081, Student’s t-test, paired, two-tailed n = 3, t = 2.453, df = 2. Black, absence of Nrp1-Fc; blue, presence of 2 μg/mL Nrp1-Fc. Source data are provided as a Source Data file.

Fig. 8. Proposed mechanism of action of microglial Nrp1 on OPC proliferation.

Resting ramified microglia (homeostatic microglia) that do not express Nrp1 (left, light pink) become activated and upregulate Nrp1. Nrp1 on activated microglia stimulates phosphorylation of PDGFRα on adjacent OPCs (gray) and promotes PDGF AA-dependent OPC proliferation. Question mark above the gray box indicates that the mechanism that causes Nrp1 upregulation on activated microglia is unknown.