Neddylation orchestrates the complex transcriptional and posttranscriptional program that drives Schwann cell myelination

Abstract

Myelination is essential for neuronal function and health. In peripheral nerves, >100 causative mutations have been identified that cause Charcot-Marie-Tooth disease, a disorder that can affect myelin sheaths. Among these, a number of mutations are related to essential targets of the posttranslational modification neddylation, although how these lead to myelin defects is unclear. Here, we demonstrate that inhibiting neddylation leads to a notable absence of peripheral myelin and axonal loss both in developing and regenerating mouse nerves. Our data indicate that neddylation exerts a global influence on the complex transcriptional and posttranscriptional program by simultaneously regulating the expression and function of multiple essential myelination signals, including the master transcription factor EGR2 and the negative regulators c-Jun and Sox2, and inducing global secondary changes in downstream pathways, including the mTOR and YAP/TAZ signaling pathways. This places neddylation as a critical regulator of myelination and delineates the potential pathogenic mechanisms involved in CMT mutations related to neddylation.

Fig. 1.. Pharmacological inhibition of neddylation blocks Schwann cell myelination.

(A) Graphical representation of the neddylation pathway and its main target, Cullin proteins. Mutations in several genes that encode critical components of the neddylation-regulated CRL complexes (pink boxes) are associated with peripheral neuropathies. (B) Immunoblot analyses of NAE1 and neddylated proteins, myelin proteins (MPZ and MBP), and positive (EGR2 and ZEB2) and negative (c-Jun and Sox2) regulators of myelination in total sciatic nerve extracts at indicated ages. (C and D) Nae1 expression in resident cells from P1 sciatic nerves in single-cell RNA sequencing (scRNA-seq) dataset GSE138577 (17) [pericytes and vascular smooth muscle cells (Per/VSMC) and fibroblast-related cluster (FbRel*)]. (C) UMAP expression of Nae1. Color corresponds to log-normalized expression values scaled to the maximum of the gene. UMAP, Uniform Manifold Approximation and Projection. (D) Violin plot showing the log-normalized Nae1 transcript levels across the three Schwann cell clusters. Wilcoxon rank sum test between groups. Prol. SC, proliferating Schwann cells; iSC, immature Schwann cells; pmSC, pro-myelinating Schwann cells. (E) Bar plot showing Nae1 transcript levels in populations enriched in myelinating and non-myelinating cells (mSC and nmSC, respectively) from P5 sciatic nerves from public dataset GSE138577 (17). Two-tailed unpaired Student’s t test. (F) Representative electron microscopy (EM) pictures of P5 sciatic nerve cross sections from mice treated with vehicle and MLN4924. Scale bar, 2 μm. Graph shows quantification of myelinated axons per sciatic nerve. Data are presented as means ± SEM; n = 3 to 4. Two-tailed unpaired Student’s t test. (G) Immunoblot and densitometric analyses of MPZ, MBP, and NEDD8 in sciatic nerve extracts from vehicle or MLN4924-treated mice. Data are presented as means ± SEM; n = 5 vehicle and 6 MLN4924 treatment. Two-tailed unpaired Student’s t test. a.u., arbitrary units. (H) Immunoblot analyses of MPZ, NAE1, and NEDD8 levels in primary rat Schwann cells cultured under basal and myelinogenic conditions [dibutyryl cyclic adenosine 3′,5′-monophosphate (db cAMP) treatment], in the presence or absence of MLN4924. β-Actin is used as loading control for immunoblots in (B), (G), and (H).

Fig. 2.. Nae1 mutant mice show severe nerve deficits.

(A) Immunoblot and densitometric quantification of NAE1 and neddylated proteins in sciatic nerve extracts from WT, heterozygotes (HET), and Nae1 cKO mice. Low and high exposures are shown for NEDD8 immunoblot to visualize differences in high (asterisk) and low MW proteins (arrowhead) respectively. Data are presented as means ± SEM. One-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test. Residual expression of NAE1 likely corresponds to other cell types present in nerves (Fig. 1C). β-Actin and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) is used as loading control for immunoblots. (B) Survival curves of WT and Nae1 cKO mice (n = 40 WT and 48 Nae1 cKO). P values are from a log-rank test between groups. (C) Tail suspension test showing abnormal hindlimb clasping in Nae1 cKO mice but not in WT littermate control at P25, a common sign of peripheral neuropathy in mice. (D to F) P25 Nae1 cKO show poor responses compared to control mice when (D) latency to fall off the accelerating rotarod, (E) grip strength of all four limbs or forelimbs, and (F) nociceptive responses using the hot-plate test were measured. Data are presented as means ± SEM. Each data point represents an individual animal. Two-tailed unpaired Student’s t test. (G to K) Nae1 mutant mice display a marked reduction in conduction velocity. (G) Electrophysiological recording of CMAPs from sciatic nerves of Nae1 cKO and control mice at P25. Representative traces are shown. S, stimulus; R, initiation of CMAP response (red arrows). Graphs show (H) nerve conduction velocities, (I) latency, (J) mean peak amplitudes of CMAPs, and (K) average durations of CMAPs in sciatic nerves of control, heterozygotes, and Nae1 cKO mice at P25. Data are presented as means ± SEM. WT (n = 8), HT (n = 6), and cKO (n = 7). One-way ANOVA with Tukey’s multiple comparisons test.

Fig. 3.. NAE1 is essential for Schwann cell myelination.

(A) Representative EM pictures showing ultrastructure of control and Nae1 cKO sciatic nerves, at indicated ages. Arrowheads indicate promyelin Schwann cells. (B to D) Graphs show quantification of (B) myelinated axons, (C) amyelinated axons (1:1) per nerve, and (D) proportion of Schwann cells in a 1:1 relationship with an axon that are myelinated or remain unmyelinated, in control and Nae1 cKO mice at indicated ages. Data are presented as means ± SEM; n = 3 to 6. Two-way ANOVA with Sidak’s multiple comparisons test. (E) Immunoblot analyses of myelin proteins (MPZ, CNP, and MBP) in total sciatic nerve lysates from control, heterozygote, and Nae1 cKO mice at P28. (F) Representative EM pictures showing axonal swellings with organelle accumulations, a typical sign of axonal pathology (22). Scale bar, 1 μm. Graphs shows the quantification of these abnormal axon profiles per nerve in control and Nae1 cKO mice at indicated ages. (G and H) Graph shows quantification of (G) axons in a 1:1 relationship with a Schwann cell, and (H) total axons per nerve in control and Nae1 cKO mice at indicated ages. Data are presented as means ± SEM; n = 3 to 6. Two-way ANOVA with Sidak’s multiple comparisons test. (I) Immunoblot analyses of neurofilament (NF) in total sciatic nerve lysates from control and Nae1 cKO mice at P30. Graph shows densitometric quantification. Data are presented as means ± SEM; n = 4 to 5. Two-tailed unpaired Student’s t test.

Fig. 4.. Neddylation regulates the Schwann cell differentiation program.

(A) Volcano plot of transcriptome profiles between control and Nae1 cKO P7 sciatic nerves (n = 4). Red and blue dots represent genes significantly down-regulated and up-regulated, respectively, in Nae1 cKO mice [fold change (FC) > 2, adjusted P value < 0.05]. (B) Gene Ontology (GO) analysis of up-regulated genes in Nae1 cKO sciatic nerves compared to control nerves. Each dot in the connecting lines represents the gene count of the corresponding biological function categories. (C) Immunofluorescence labeling for proliferative EdU⁺ cells (red) in sciatic nerves from control and Nae1 cKO at P7. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) (blue). Scale bar, 100 μm. Graphs shows quantification of EdU⁺ cells in sciatic nerves of control and Nae1 cKO mice at P1, P7, P15, and P30 (n = 3 animals per genotype). Data are presented as means ± SEM. Two-way ANOVA with Sidak’s multiple comparisons test (between groups). (D) GO analysis of down-regulated genes in Nae1 cKO sciatic nerves compared to control nerves. Each dot in the connecting lines represents the gene count of the corresponding biological function categories. (E) ToppCluster plot showing the functional networks among the genes down-regulated in Nae1 cKO sciatic nerves.

Fig. 5.. Neddylation inhibition leads to mTOR hyperactivation in Schwann cells.

(A) GSEA plot showing enrichment of TSC1-regulated and PTEN-regulated Schwann cell gene signature in Nae1 cKO mice. NES, normalized enrichment score. (B) Immunoblot analyses and densitometric quantification of mTOR pathway components in sciatic nerve lysates from control and Nae1 cKO mice at P28. Data are presented as means ± SEM; n = 4 to 7. Two-tailed unpaired Student’s t test, *P < 0.05, **P < 0.01, ****P < 0.0001. (C and D) Immunoblot analyses (C) and densitometric quantification (D) of MPZ and p-4E-BP1 levels in sciatic nerve lysates from control and Nae1 cKO mice, treated with vehicle or rapamycin. Data are presented as means ± SEM. Two-way ANOVA with Sidak’s multiple comparisons test. (E and F) Representative EM micrographs (E) and quantification of myelinated and amyelinated (1:1) axons per nerve (F) of P10 sciatic nerves from control and Nae1 cKO mice treated with vehicle or rapamycin. Scale bar, 2 μm. Data are presented as means ± SEM; n = 3 to 4. Two-way ANOVA with Sidak’s multiple comparisons test. (G) Immunoblot analyses of mTOR pathway components in primary rat Schwann cells cultured under basal or myelinogenic conditions (db cAMP treatment), in the presence or absence of MLN4924. (H) Immunoblot analyses showing that inhibiting the hyperactivation of mTOR pathway with rapamycin does not rescue the MLN4924-induced suppression of MPZ in primary rat Schwann cells cultured under myelinogenic conditions (db cAMP treatment). β-Actin is used as loading control for immunoblots in (B), (C), (G), and (H).

Fig. 6.. Neddylation inhibition leads to suppression of the YAP/TAZ pathway in Schwann cells.

(A) Immunoblot analyses and densitometric quantification of Hippo-Yap signaling pathway components in sciatic nerve lysates from control and Nae1 cKO mice at P28. Data are presented as means ± SEM; n = 3 to 8, *P < 0.05, **P < 0.01, ****P < 0.0001. Two-tailed unpaired Student’s t test. (B) RT-qPCR showing regulation of YAP/TAZ target genes in sciatic nerves from Nae1 cKO mice at P28. Data are presented as means ± SEM; n = 4 to 12. Two-tailed unpaired Student’s t test. (C and D) Immunoblot analyses (C) and densitometric quantification (D) of MPZ and p-YAP levels in sciatic nerve lysates from control and Nae1 cKO mice, treated with vehicle or XMU-MP-1. Data are presented as means ± SEM. Two-way ANOVA with Sidak’s multiple comparisons test. (E and F) Representative EM micrographs (E) and quantification of myelinated and amyelinated axons (1:1) per nerve (F) of P10 sciatic nerves from control and Nae1 cKO mice treated with vehicle or XMU-MP-1. Scale bar, 2 μm. Data are presented as means ± SEM; n = 3 to 4. Two-way ANOVA with Sidak’s multiple comparisons test (between groups). (G) Immunoblot analyses of YAP/TAZ pathway components in primary rat Schwann cells cultured under basal or myelinogenic conditions (db cAMP treatment), in the presence or absence of MLN4924. β-Actin is used as loading control for immunoblots. (H) Immunoblot analyses showing that treatment with the MST1/2 inhibitor XMU-MP-1 does not rescue the MLN4924-induced suppression of MPZ in primary rat Schwann cells cultured under myelinogenic conditions (db cAMP treatment). β-Actin is used as loading control for immunoblots in (A), (C), (G), and (H).

Fig. 7.. Proteomics analyses show compromised ubiquitin-mediated proteolysis in Nae1 cKO mice.

(A) Volcano plot of proteomic analysis in control and Nae1 cKO P7 sciatic nerves (n = 5). Red and blue dots represent significantly down-regulated and up-regulated proteins, respectively, in Nae1 cKO mice (FC > 2, adjusted P value < 0.05). (B) Top enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) terms of dysregulated proteins. PI3K, phosphatidylinositol 3-kinase. (C) Z-score polar plot of deregulated proteins associated with ubiquitination and neddylation. Green represents down-regulated proteins, and red represents up-regulated proteins. (D) Immunoblot analyses of c-Jun and Sox2 levels in sciatic nerve lysates from P28 Nae1 cKO mice. (E) RT-qPCR of c-Jun–regulated Schwann cell repair genes in sciatic nerves from P28 Nae1 cKO mice. Data are presented as means ± SEM; n = 4 to 6. Two-tailed unpaired Student’s t test. (F) Immunoblot analyses of c-Jun and Sox2 levels in primary rat Schwann cells. (G and H) RT-qPCR showing expression of (G) c-Jun and (H) Sox2 mRNA levels in primary Schwann cells. Data are presented as means ± SEM; n = 3. Two-way ANOVA with Sidak’s multiple comparisons test. (I and J) Immunoblot and densitometric quantification of (I) c-Jun and (J) Sox2 levels in primary Schwann cells. Data are presented as means ± SEM; n = 3. Two-way ANOVA with Sidak’s multiple comparisons test. Dotted lines represent nonlinear regression analysis (one-phase decay) of c-Jun and Sox2 degradation over time after db cAMP treatment. *P < 0.05; **P < 0.01; ****P < 0.0001. (K and L) Immunoblot and densitometric quantification of (K) c-Jun and (L) Sox2 levels in primary Schwann cells, treated with cycloheximide (CHX) for the indicated times. Data are presented as means ± SEM; n = 8. Two-way ANOVA with Sidak’s multiple comparisons test. Dotted lines represent nonlinear regression analysis (one-phase decay) of c-Jun and Sox2 degradation over time after CHX treatment. (M and N) Immunoblot analyses of MPZ levels after (M) c-Jun and (N) Sox2 silencing. β-Actin is used as loading control for immunoblots in (D), (F), (H), (I), and (K) to (N).

Fig. 8.. Neddylation stabilizes EGR2 protein.

(A and B) GSEA plots showing suppression of (A) EGR2-regulated and (B) Zeb2-regulated Schwann cell gene signature (38) in Nae1 cKO mice. (C and D) Immunoblots (C) and densitometric quantification (D) of EGR2 and ZEB2 in sciatic nerve lysates from P28 WT, HET, and Nae1 cKO mice. Data are presented as means ± SEM; n = 5 to 11. One-way ANOVA with Tukey’s multiple comparisons test. (E and F) Immunoblot analyses of EGR2 and ZEB2 levels in primary rat Schwann cells in (E) whole-cell lysates, or (F) cytosolic and nuclear extracts. (G) Graph showing positive correlation between EGR2 and NEDD8 protein levels in WT, heterozygote, and Nae1 mutant mice (two-sided Pearson’s correlation). The dashed line shows the regression line, and the shaded area represents the 95% confidence interval. (H) Immunoblot analysis of EGR2 and ZEB2 levels in HALO-NEDP1 pulldowns and whole-cell lysates (input) from primary Schwann cells cultured under myelinogenic conditions (db cAMP), in the absence or presence of MLN4924. (I and J) Co-immunoprecipitation of (I) endogenous NEDD8 and (J) endogenous EGR2 from primary Schwann cells. Immunoglobulin G (IgG) heavy chain is denoted by an asterisk. (K and L) Immunoblot and densitometric analyses of (K) EGR2 and (L) ZEB2 levels in primary Schwann cells, cultured in the presence or absence of MLN4924, and treated with CHX (5 μm) for the indicated times. Data are presented as means ± SEM; n = 3. Two-way ANOVA with Sidak’s multiple comparisons test. Dotted lines represent nonlinear regression analysis (one-phase decay) of EGR2 and ZEB2 degradation over time after CHX treatment. *P < 0.05; ****P < 0.0001. (M) Co-immunoprecipitation of endogenous EGR2 from primary Schwann cells, followed by immunoblot analysis of whole-cell lysates and immunoprecipitates with anti-ubiquitin antibody. (N) Immunoblot analyses of MPZ levels in db cAMP and MLN4924-treated primary Schwann cells after treatment with MG132. β-Actin is used as loading control for immunoblots in (C), (E), (F), and (H) to (N).

Fig. 9.. Neddylation inhibition blocks peripheral nerve generation after injury.

(A) Schematic diagram depicting the strategy for nerve regeneration paradigm. Sixty-day-old Nae1 icKO mice were treated with tamoxifen, followed by crush injury 40 days later, and sacrifice at P140. (B) Quantification of sensory-motor function using sciatic functional index (SFI) in WT and Nae1 icKO mice. Mutants show a substantial block in regeneration, although mutant and WT SFIs are similar before and immediately after injury. Two-way ANOVA with Sidak’s multiple comparisons test (*P < 0.05; **P < 0.01; ****P < 0.0001). (C and D) Nae1 mutant mice display a marked reduction in conduction velocity. (C) Electrophysiological recording of CMAPs from regenerated sciatic nerves 40 days after crush. Representative traces are shown. S, stimulus; R, initiation of CMAP response (red arrows). (D) Graphs show nerve conduction velocities, latency, mean peak amplitudes of CMAPs, and average durations of CMAPs. Data are presented as means ± SEM. WT (n = 7) and icKO (n = 6). Two-tailed unpaired Student’s t test. (E) Representative EM pictures showing ultrastructure of regenerated sciatic nerves 40 days after crush. Scale bar, 5 μm. (F) Graphs show quantification of myelinated axons, amyelinated axons (1:1), Schwann cell nuclei and total axons per nerve. Two-tailed unpaired Student’s t test. (G) Immunoblot analyses of myelin proteins (MPZ and MBP), and axonal marker NF in regenerated sciatic nerves 40 days after crush. (H) Graphical abstract showing role of neddylation in regulating levels of key molecular signals during myelination.