JAB1 deletion in oligodendrocytes causes senescence-induced inflammation and neurodegeneration in mice

Abstract

Oligodendrocytes are the primary target of demyelinating disorders, and progressive neurodegenerative changes may evolve in the CNS. DNA damage and oxidative stress are considered key pathogenic events, but the underlying molecular mechanisms remain unclear. Moreover, animal models do not fully recapitulate human diseases, complicating the path to effective treatments. Here we report that mice with cell-autonomous deletion of the nuclear COP9 signalosome component CSN5 (JAB1) in oligodendrocytes develop DNA damage and defective DNA repair in myelinating glial cells. Interestingly, oligodendrocytes lacking JAB1 expression underwent a senescence-like phenotype that fostered chronic inflammation and oxidative stress. These mutants developed progressive CNS demyelination, microglia inflammation, and neurodegeneration, with severe motor deficits and premature death. Notably, blocking microglia inflammation did not prevent neurodegeneration, whereas the deletion of p21CIP1 but not p16INK4a pathway ameliorated the disease. We suggest that senescence is key to sustaining neurodegeneration in demyelinating disorders and may be considered a potential therapeutic target.

Keywords: Cellular senescence; Demyelinating disorders; Inflammation; Mouse models; Neuroscience.

Figure 1. Generation of oligodendrocyte-conditional Jab1-null mice.

(A) Cnp-Cre Jab1^fl/fl mouse showing reduced size as compared with littermate control. (B) Genotyping PCR using optic nerve–derived DNA providing genomic recombination only in Cnp-Cre Jab1^fl/fl mice. (C) qPCR for Cops5 (Jab1) expression in sorted oligodendrocytes (O1⁺) showing reduced expression in Cnp-Cre Jab1^fl/fl as compared with WT mice (*P < 0.05; n = 4; 2-tailed nonparametric Mann-Whitney U test). (D) Western blot analysis and quantification of optic nerve homogenate showing JAB1 reduction in Cnp-Cre Jab1^fl/fl mice (****P < 0.0001; n = 6; 1-sample 2-tailed Student’s t test; WT is reported as equal to 1). (E) Confocal images of the corpus callosum (within dashed lines) of Cnp-Cre Jab1^fl/fl Rosa26 mice stained for GFP, CC1, NG2, and DAPI. Recombined OPCs positive for GFP and NG2 (arrows) and oligodendrocytes positive for GFP and CC1 (arrowheads) are labeled in the inset magnification. (F) Kaplan-Meier curve showing reduced survival in mutant mice (n = 100 controls and 72 mutants; ****P < 0.0001, χ² Mantel-Cox test). (G) Mutant mice showing significant body weight reduction over time (*P < 0.05, ***P < 0.001; n = 10; 2-way ANOVA with Bonferroni’s post hoc correction). (H) Rotarod analysis showing significant motor deficits in mutant mice since P40 (****P < 0.0001; 2-tailed nonparametric Mann-Whitney U test). Scale bar: 20 μm; inset, 10 μm.

Figure 2. Morphological abnormalities in Jab1-mutant mice.

(A) Electron micrographs of transverse sections of the optic nerve, corpus callosum, and spinal cord from WT and Cnp-Cre Jab1^fl/fl mice at ages P20, P40, and P60, showing progressive demyelination and axonal loss. (B) Representative electron micrographs of spinal cord cross sections from P60 Cnp-Cre Jab1^fl/fl mice, showing axonal spheroid (asterisk), swelling and accumulation of organelles and ovoids (arrow), and myelin fragmentation (arrowheads). (C) Quantification of g-ratio and numbers of myelinated fibers and degenerating fibers in the optic nerves from WT and Cnp-Cre Jab1^fl/fl mice at different time points (*P < 0.05, **P < 0.01; n = 4–5; at least 2000 fibers counted per mouse; 2-tailed nonparametric Mann-Whitney U test). Scale bars: (A and B) 2 μm.

Figure 3. Longitudinal MRI and VEP evaluation.

(A) Brain MT MRI images of WT and Cnp-Cre Jab1^fl/fl mice at different ages (P30, P45, and P65). While in WT mice, the corpus callosum is clearly delineated on MT images (in dark, arrows), it is hardly distinguishable in Jab1-mutant mice, in which the contrast is similar at the first exam (P30) and becomes hyperintense at later stages (P45 and P60, arrows). (B) With DTI (color map of FA), the corpus callosum is well discriminated in all the animals, but the FA (color map) is clearly decreased at P45 and P60 in Jab1-mutant as compared with WT mice (arrows). (C) Quantitative analysis of MT and FA measured in the medial of the corpus callosum. Both MT and FA were significantly reduced with age in Jab1-mutant mice as compared with healthy controls (*P < 0.05, ***P < 0.001; n = 6; 2-way ANOVA with Bonferroni’s post hoc correction). (D) N1 latency and amplitude measured in WT and Cnp-Cre Jab1^fl/fl mice at age P30, P45, and P65 (****P < 0.0001; n = 7 WT, 6 mutant; 2-way ANOVA with Bonferroni’s post hoc correction). (E) Representative VEP waveforms recorded from WT and Cnp-Cre Jab1^fl/fl at different ages. Horizontal bar, 50 ms; vertical bar, 20 μV.

Figure 4. Inflammatory infiltration in Jab1-mutant mice.

(A) Immunolabeling of WT and Cnp-Cre Jab1^fl/fl optic nerves for IBA1 and MBP at ages P60 and P90 and relative quantification from P20 to P90, showing progressive increase in the number of IBA1⁺ cells in mutant mice (*P < 0.05, **P < 0.01; n = 3–5; 2-tailed nonparametric Mann-Whitney U test). (B) FACS analysis of P60 WT and Cnp-Cre Jab1^fl/fl brain for CD45, CD11b (representative plots), CD11c, Ly6G, and Ly6C, showing quantification of microglia, activated microglia, lymphocytes, monocytes, macrophages, and neutrophils (**P < 0.01, ***P < 0.001; n = 13 WT, n = 9 Cnp-Cre Jab1^fl/fl; 2-tailed nonparametric Mann-Whitney U test). (C) Immunolabeling for GFAP and CC1 of the optic nerve from WT and Cnp-Cre Jab1^fl/fl mice at age P60 and relative quantification, showing increased number of GFAP⁺ reactive astrocytes in mutant mice (*P < 0.05; n = 4; 2-tailed nonparametric Mann-Whitney U test). Scale bars: (A and C) 40 μm.

Figure 5. Depletion of microglia activation does not ameliorate neurodegeneration in Jab1-mutant mice.

(A) Electron micrographs of transverse sections of P50 optic nerve from WT, CCR2^–/–, Cnp-Cre Jab1^fl/fl, and Cnp-Cre Jab1^fl/fl CCR2^–/– mice and relative quantification of IBA1⁺ cells and numbers of myelinated fibers and degenerating fibers, showing no significant differences between Jab1^–/– and double-mutant Jab1^–/– CCR2^–/– mice (**P < 0.01, ***P < 0.001; n = 4–5; 1-way ANOVA with Bonferroni’s multiple-comparison test). (B) Confocal images (for IBA1 and MBP) and electron micrographs of P41 optic nerve from WT and Cnp-Cre Jab1^fl/fl mice treated (PLX +) or not (PLX –) for 21 days with PLX3397 (see schematic) and relative quantification of IBA1⁺ cells and numbers of myelinated and degenerating fibers. Although the number of IBA1⁺ cells was significantly reduced in PLX-treated mice, no differences in the numbers of myelinated and degenerating fibers were observed (*P < 0.05, **P < 0.01, ***P < 0.001; n = 4–9; 1-way ANOVA with Bonferroni’s multiple-comparison test). Scale bars: (A) 1 μm; (B) immunohistochemistry, 40 μm; electron microscopy, 2 μm.

Figure 6. DNA damage and senescence in mutant mice.

(A) Confocal immunolabeling of WT and mutant optic nerves stained for CC1 and p-H2AX; quantification shows significant increase of p-H2AX⁺ oligodendrocytes in mutant mice from P40 (*P < 0.05, **P < 0.01; n = 3–5; 2-tailed nonparametric Mann-Whitney U test). (B) Confocal images of ex vivo optic nerves x-ray–irradiated (IR) to induce DNA damage, stained for DNA-PKcs and CC1; the majority of WT oligodendrocytes (OLs) have nuclear expression of DNA-PK, whereas this percentage is reduced in Jab1^–/– optic nerves (**P < 0.01; n = 5; 2-tailed nonparametric Mann-Whitney U test). (C) β-Gal staining and quantification of the corpus callosum and optic nerve from P60 WT and Cnp-Cre Jab1^fl/fl mice double-stained with CC1, showing senescent oligodendrocytes in mutant mice (*P < 0.05; n = 5–7; 2-tailed nonparametric Mann-Whitney U test). (D) qPCR for Cdkn2a (p16^INK4a) in the optic nerve from WT and Cnp-Cre Jab1^fl/fl mice, showing significant increase from P40 (**P < 0.01; n = 5–6; 2-tailed nonparametric Mann-Whitney U test). (E) qPCR for Cdkn1a (p21^CIP1) in the optic nerve from WT and Cnp-Cre Jab1^fl/fl mice, showing significant increase from P40 (*P < 0.05, **P < 0.01; n = 5–6; 2-tailed nonparametric Mann-Whitney U test). (F) qPCR for SASP in the optic nerve from P60 WT and Cnp-Cre Jab1^fl/fl mice, showing significant elevation in mutant mice (*P < 0.05, **P < 0.01; n = 4–6; 2-tailed nonparametric Mann-Whitney U test). (G) Western blot analysis, quantification, and confocal images for HMGB1 in P60 optic nerves of WT and Cnp-Cre Jab1^fl/fl mice; HMGB1 is increased in mutant mice and is also translocated in the cytoplasm of mutant oligodendrocytes (arrows) (*P < 0.05; n = 4; 2-tailed nonparametric Mann-Whitney U test). (H) ROS quantification in optic nerve and corpus callosum from WT and Cnp-Cre Jab1^fl/fl mice, showing significant ROS elevation from P40 (*P < 0.05, **P < 0.01; n = 3–5; 2-tailed nonparametric Mann-Whitney U test). Scale bars: (A) 40 μm; (B) 2 μm; (C) light microscopy, 50 μm; fluorescence, 2 μm; (G) 20 μm.

Figure 7. CNS demyelination, inflammation, and axonal degeneration are reproduced in Plp-CreERT2 Jab1^fl/fl mice.

(A) Tamoxifen-treated Plp-CreERT2 Jab1^fl/fl mouse showing reduced size as compared with littermate control and quantification of body weight at P60 and P180 showing significant reduction at later time point (**P < 0.01; n = 18 at P60 and 6 at P180; 2-tailed nonparametric Mann-Whitney U test). dpi, days postinjection. (B) Schematic representation of tamoxifen administration and electron micrographs of P60 and P180 optic nerve and corpus callosum from tamoxifen-treated Plp-CreERT2 Jab1^fl/fl and control (vehicle-treated Plp-CreERT2 Jab1^fl/fl) mice, showing no differences at P60 but diffuse demyelination and axonal degeneration at P180. (C) Quantification showing significant increase in the number of degenerating fibers and decrease of myelinated fibers in P180 tamoxifen-treated Plp-CreERT2 Jab1^fl/fl mice (**P < 0.01; n = 3 at P60 and 6 at P180; 2-tailed nonparametric Mann-Whitney U test). (D) Immunolabeling for IBA1 and MBP of P180 optic nerves from tamoxifen-treated Plp-CreERT2 Jab1^fl/fl and control mice, and relative quantification showing significantly increased number of IBA1⁺ cells in tamoxifen-treated mice (*P < 0.05; n = 4 in controls and 5 in tamoxifen-treated mice; 2-tailed nonparametric Mann-Whitney U test). (E) β-Gal staining of the optic nerve from P180 tamoxifen-treated Plp-CreERT2 Jab1^fl/fl and control mice double-stained with CC1, showing β-gal⁺ oligodendrocytes in tamoxifen-treated mice. (F) qPCR for Cdkn1a (p21^CIP1) and Cdkn2a (p16^INK4a), showing increased expression in the optic nerve from P180 tamoxifen-treated Plp-CreERT2 Jab1^fl/fl as compared with control mice (***P < 0.001; n = 7; 2-tailed nonparametric Mann-Whitney U test). Scale bars: (B) 2 μm; (D) 40 μm; (E) 10 μm.

Figure 8. Deletion of p21^CIP1 ameliorates the phenotype in Jab1-mutant mice.

(A) Rotarod analysis showing amelioration of motor deficits in Cnp-Cre Jab1^fl/fl p21^CIP1–/– as compared with Cnp-Cre Jab1^fl/fl mice (**P < 0.01 at P40; n = 5; 2-tailed nonparametric Mann-Whitney U test). (B) Quantification of g-ratio in P60 optic nerves, showing a significant reduction in Cnp-Cre Jab1^fl/fl p21^CIP1–/– as compared with Cnp-Cre Jab1^fl/fl mice (*P < 0.05, **P < 0.01; n = 5–7; at least 2000 fibers counted per mouse; 1-way ANOVA with Bonferroni’s multiple-comparison test). (C) Confocal images and quantification of BRN3A⁺ ganglion cells in the retina at P60, showing a significant rescue in Cnp-Cre Jab1^fl/fl p21^CIP1–/– as compared with Cnp-Cre Jab1^fl/fl mice (**P < 0.01, ***P < 0.001; n = 6–8; 1-way ANOVA with Bonferroni’s multiple-comparison test). (D) Quantification of myelinated fibers and degenerating fibers in the P60 spinal cord, showing significant rescue in Cnp-Cre Jab1^fl/fl p21^CIP1–/– as compared with Cnp-Cre Jab1^fl/fl mice (*P < 0.05, ***P < 0.001; n = 7–9; 1-way ANOVA with Bonferroni’s multiple-comparison test). (E) Quantification of Western blot analysis for phosphorylated (SMI-31) and non-phosphorylated (SMI-32) neurofilaments of high molecular weight in P60 brain homogenate, showing a significant rescue in Cnp-Cre Jab1^fl/fl p21^CIP1–/– as compared with Cnp-Cre Jab1^fl/fl mice (*P < 0.05; n = 6; 1-way ANOVA with Bonferroni’s multiple-comparison test). (F) Confocal immunolabeling for CC1 and β-gal staining and quantification in P60 optic nerve, showing reduced number of senescent oligodendrocytes in Cnp-Cre Jab1^fl/fl p21^CIP1–/– as compared with Cnp-Cre Jab1^fl/fl mice (*P < 0.05; n = 4; 2-tailed nonparametric Mann-Whitney U test). (G) Heatmaps representing the expression values of key genes for cellular senescence and for SASP/inflammation differentially expressed in FACS-sorted O1⁺ oligodendrocytes. Scale bars: (C) 40 μm; (F) 40 μm.

Figure 9. Expression of JAB1 in brain of MS patients.

(A) Representative immunofluorescence images for nuclear JAB1 (red) expression in OLIG2⁺ cells (green) in control white matter (WM) or in normal-appearing white matter (NAWM) surrounding chronic inactive MS lesion. Regions of interest (ROIs) were generated on DAPI images (left panels) to select nuclei and then applied to the corresponding OLIG2 and JAB1 images. Arrows highlight JAB1⁺ oligodendrocytes. (B) Percentage of JAB1-expressing OLIG2⁺ cells in control white matter (black dots) or MS normal-appearing white matter (white dots). Each dot represents a single sample, and bars represent mean (*P < 0.05; n = 6; 2-tailed nonparametric Mann-Whitney U test). Scale bar: 10 μm.