Nf1 loss and Ras hyperactivation in oligodendrocytes induce NOS-driven defects in myelin and vasculature

Abstract

Patients with neurofibromatosis type 1 (NF1) and Costello syndrome Rasopathy have behavioral deficits. In NF1 patients, these may correlate with white matter enlargement and aberrant myelin. To model these features, we induced Nf1 loss or HRas hyperactivation in mouse oligodendrocytes. Enlarged brain white matter tracts correlated with myelin decompaction, downregulation of claudin-11, and mislocalization of connexin-32. Surprisingly, non-cell-autonomous defects in perivascular astrocytes and the blood-brain barrier (BBB) developed, implicating a soluble mediator. Nitric oxide (NO) can disrupt tight junctions and gap junctions, and NO and NO synthases (NOS1-NOS3) were upregulated in mutant white matter. Treating mice with the NOS inhibitor NG-nitro-L-arginine methyl ester or the antioxidant N-acetyl cysteine corrected cellular phenotypes. CNP-HRasG12V mice also displayed locomotor hyperactivity, which could be rescued by antioxidant treatment. We conclude that Nf1/Ras regulates oligodendrocyte NOS and that dysregulated NO signaling in oligodendrocytes can alter the surrounding vasculature. The data suggest that antioxidants may improve some behavioral deficits in Rasopathy patients.

Copyright © 2013 The Authors. Published by Elsevier Inc. All rights reserved.

Figures

Figure 1. Nf1 Loss or HRas Activation in Oligodendrocytes Causes Optic-Nerve Enlargement, Myelin Decompaction, Loss of Claudin-11, and Altered Cx32 Localization

(A) Gross micrographs of optic nerves at the level of the chiasm in 12-month-old animals. Ruler shows 1 mm markings. (B) Toluidine-blue-stained semithin optic-nerve cross-sections, 1 mm rostral to the chiasm. (C) Optic-nerve g-ratio scatterplots (~5,000 axons/graph; measurements 1 mm rostral to the chiasm in three to five electron micrographs at 10,000×). Red box: 99% of WT g-ratios superimposed upon graphs of other genotypes. ANOVA, Tukey post hoc, p = 2.2 × 10−16. (D) Diagram of myelin compaction. Brown circle: axon cross-section; gray: major dense lines (MDLs); orange: Cx32 GJs between compact myelin layers; blue lines: claudin-11+ TJs in compact myelin. (E) Electron micrographs (scale bar, 10 nm) of WT, PLPCre;Nf1 fl/fl, and CNP-HRas optic-nerve myelin. White arrows: intraperiod lines (ILs); black arrows: MDLs. (F–I) Western blots of total optic-nerve lysate or whole-brain cytosolic, endosomal, or myelin fractions showing myelin proteins (F), claudin-11 (G), or Cx32 cellular localization (H and I). n = 3–5 animals/genotype/experiment. See also Figures S1, S2, S3, and S4.

Figure 2. GJs Are Lost in Enlarged Optic-Nerve Astrocyte Endfeet after Nf1 Loss or HRas Activation

(A) Electron micrograph cross-sections of capillaries and perivascular space in optic nerve 1 mm from the chiasm (10,000×; scale bar, 2 μm). (B) Left: Diagram of capillary. Yellow: GJs. As, astrocyte endfoot; BV, blood vessel; Endo, endothelial cell. (B) Electron micrographs of capillaries 1 mm from the chiasm (50,000×; scale bar, 250 nm). Red arrows: GJs on astrocyte endfeet; red arrowheads: astrocyte endfeet lacking GJs. (C) Quantified perivascular area (% total area divided by vascular area). (D) Percentage of astrocyte endfeet with visible GJs. (E and F) Western blots using total optic-nerve lysate (E) or whole-brain endosomal fractions (F) showing Cx43. n = 3–5 animals/genotype. *p

Figure 3. Nf1 Loss or HRas Activation in the Optic Nerve Causes Changes in Claudins and BBB Permeability

(A) Electron micrographs of optic-nerve cross-sections of capillaries 1 mm from the chiasm (50,000×; scale bar, 200nm) show electron-dense TJs between endothelial cells. Blue arrowheads: areas of TJ disruption. (B) Quantification of the total TJs with gaps, with 200–300 TJs counted per genotype. (C) Evans blue stain of longitudinal sections of PLPCre; Nf1fl/fl and CNP-HRas optic nerve. Insets: cross-sections; arrows: blood vessel. (D) Western blots from total optic-nerve lysates showing claudin-1 and claudin-5. All experiments included three to five animals per genotype. *p

Figure 4. Increased ROS after Nf1 Loss or HRas Activation

NOS sufficiency for phenotype presentation. (A) Western blot analysis of NOS (NOS1–NOS3) in optic-nerve protein lysates. Blot intensity was measured with ImageJ software and caps at upper intensity levels; “max” indicates intensity beyond the upper intensity limitations. (B) DCF-DA stain in unfixed optic-nerve cross-sections (5×; inset: 40×) from WT and CNP-HRas mice with and without the NOS inhibitor L-NAME. White line shows optic-nerve boundaries. (C) g-Ratio scatterplots show measurements of >1,000 axons/animal. (D) Quantification. (E–K) Electron micrographs of optic-nerve cross-sections 1 mm rostral to the chiasm (E and G: 10,000× and scale bar, 2 μm; I and K: 50,000× and scale bar, 100 nm). (F) Optic area quantified in semithin cross-sections 1 mm rostral to the chiasm. (H, J, and L) Quantification of perivascular area (H), percentage of astrocyte endfeet with visible GJs (J,~300 GJs/animal), and percentage of endothelial cells with TJ gaps (L, ~50–60 TJs/animal) measured using electron microscopy; n = 3–5 animals/genotype. See also Figures S6 and S7.

Figure 5. Heterozygous Nf1 Loss in Oligodendrocytes Causes Myelin and Vascular Phenotypes

(A) Gross micrographs of optic nerves at the level of the chiasm in 12-month-old animals. Ruler shows 1 mm markings. (B–F) Optic nerve 1 mm rostral to the chiasm. (B) Toluidine-blue-stained semithin cross-sections. (C) Scatterplots of g-ratios (~5,000 axons/graph) from three to five electron micrographs (10,000×) of three to five animals/genotype. Red box: 99% of WT g-ratios superimposed upon graphs of other genotypes. ANOVA, Tukey post hoc, p = 2.2 × 10−16. (D) Electron micrographs of capillaries and perivascular space (10,000×; scale bar, 2 μm). (E) Electron micrographs of optic-nerve cross-sections of capillaries (50,000×; scale bar, 200 nm). Red arrows: GJs on astrocyte endfeet; red arrowheads: astrocyte endfeet without GJs. (F) Electron micrographs of TJs between endothelial cells (50,000×, scale bar, 200 nm). Blue arrowheads: areas of TJ disruption. (G) Western blot analysis of NOS (NOS1–NOS3) in total optic-nerve lysates. (H) Western blots in total optic-nerve lysate showing Cx43.

Figure 6. Increased Reactive Oxygen Is Detected Only in Oligodendrocytes

(A) DCF-DA fluorescence intensity measured by flow cytometry in dissociated forebrain + optic nerves of WT and PLP;Nf1fl+ mice 6 months after tamoxifen injection. Gates for DCF-DA with cell percentages indicated. (B) Gates were used to discriminate cell types. Top row: WT animals are shown with red gates. Middle row: PLP;Nf1fl+ animals are shown with blue gates. Bottom row: DCF-DA fluorescence intensities per cell type with WT (red) and PLP;Nf1fl+ (blue) lines. (C) Cell Rox Orange fluorescence in GFP− and GFP+ cells.

Figure 7. The Antioxidant NAC Rescues Myelin and Vascular Phenotypes in PLP;Nf1floxed and CNP-HRas Mice

Quantification of phenotypes in WT, PLP;Nf1fl+, PLP;Nf1fl/fl, and CNP-HRas animals after 6 weeks of vehicle or NAC treatment, all in optic nerve 1 mm rostral to the chiasm. (A) Optic-nerve diameter quantified from semithin cross-sections. (B and C) Quantification of g-ratio (B) and axon diameter (C) of >1,000 axons/genotype measured by electron microscopy. (D) Quantification of % perivascular area normalized to blood vessel area. (E) Percentage of astrocyte endfeet with visible GJs (~300 GJs/animal). (F) Percentage of endothelial TJs with gaps (~50–60 TJs/animal); n = 3–5 animals/genotype.

Figure 8. CNP-HRas Animals Have Hyperactive Locomotion and Hypersensitivity to Startle Behaviors

Quantification of behavior in CNP-HRas mice after 6 weeks of vehicle or NAC treatment. (A) Peripheral beam interruptions on locomotor behavioral test. *Vehicle p 0.08. (C) Average startle Vavg. † Vehicle p = 0.07; NAC p