Enhanced axonal neuregulin-1 type-III signaling ameliorates neurophysiology and hypomyelination in a Charcot-Marie-Tooth type 1B mouse model

Abstract

Charcot-Marie-Tooth (CMT) neuropathies are a group of genetic disorders that affect the peripheral nervous system with heterogeneous pathogenesis and no available treatment. Axonal neuregulin 1 type III (Nrg1TIII) drives peripheral nerve myelination by activating downstream signaling pathways such as PI3K/Akt and MAPK/Erk that converge on master transcriptional regulators of myelin genes, such as Krox20. We reasoned that modulating Nrg1TIII activity may constitute a general therapeutic strategy to treat CMTs that are characterized by reduced levels of myelination. Here we show that genetic overexpression of Nrg1TIII ameliorates neurophysiological and morphological parameters in a mouse model of demyelinating CMT1B, without exacerbating the toxic gain-of-function that underlies the neuropathy. Intriguingly, the mechanism appears not to be related to Krox20 or myelin gene upregulation, but rather to a beneficial rebalancing in the stoichiometry of myelin lipids and proteins. Finally, we provide proof of principle that stimulating Nrg1TIII signaling, by pharmacological suppression of the Nrg1TIII inhibitor tumor necrosis factor-alpha-converting enzyme (TACE/ADAM17), also ameliorates the neuropathy. Thus, modulation of Nrg1TIII by TACE/ADAM17 inhibition may represent a general treatment for hypomyelinating neuropathies.

Figure 1

Overexpression of Nrg1TIII ameliorates neurophysiological parameters in P0S63del-CMT1B mice. Analyses were performed on 6-month-old mice. (A) NCV is reduced in S63del mice as compared with WT and is rescued by Nrg1TIII overexpression. (B) Representative original recording indicating the onset of the F-wave (vertical bar). (C) F-wave latency is increased in S63del mice as compared with WT and shows a significant amelioration in HANI/+//S63del as compared with S63del mice. (D) CMAP amplitude and distal latency do not show changes among the genotypes. Data represent the mean ± SEM. Number of animals per genotype is reported in (D). ^***P-value < 0.001 by one-way ANOVA with Bonferroni’s multiple comparison test.

Figure 2

Overexpression of Nrg1TIII leads to thicker myelin without increase in the expression of most myelin proteins. (A) EM of P28 sciatic nerves. HANI mice show thicker myelin (arrow) as compared with WT. Hypomyelination in S63del is rescued in small and medium caliber axons (arrow) in HANI/+//S63del nerves (size bar, 5 μm). (B) G-ratio analysis performed on sciatic nerve cross sections at P28 and (C) relative quantification (average g-ratio values WT 0.64 ± 0.006, HANI/+ 0.53 ± 0.012, S63del 0.69 ± 0.005, S63del//HANI/+ 0.62 ± 0.002). Eight to ten microscopic fields per mouse were analyzed. Three to five mice per genotype were used. (D) Western analysis on sciatic nerve lysates at P28 shows no significant increase in myelin proteins in mice overexpressing Nrg1TIII, except for PMP2. Tubulin was used as loading control. One representative blot of three is shown. (E) Protein levels as measured by densitometric analysis. (F) mRNA expression levels of myelin proteins measured via qRT-PCR. PMP2 mRNA is increased in mice overexpressing Nrg1TIII, MBP shows a small increase in HANI versus WT. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001 by one-way ANOVA with Bonferroni’s multiple comparison test; data represent the mean ± SEM. Each experiment was repeated three times on different pools of three nerves per genotype.

Figure 3

The Akt and Erk1/2 signaling pathways are upregulated by Nrg1TIII overexpression. (A) WB analysis of P28 sciatic nerves lysates. No differences in the endogenous full-length form of Ngr1 (140 kDa) were detected among the four genotypes, whereas the cleaved (65–70 kDa) active form was increased in particular in HANI/+//S63del as compared with S63del. In HANI/+ and HANI/+//S63del mice, both Akt ^(Ser473) and Erk1/2 phosphorylation are increased as compared with S63del and WT mice, respectively. Krox20 protein expression is slightly, but not significantly decreased in HANI/+//S63del as compared with S63del mice. No differences were detected in Krox20 mRNA expression by qRT-PCR (B, lower right graph). Tubulin was used as loading control. One representative experiment of three is shown. (B) Densitometric quantification of full-length and cleaved Nrg1 forms, p-Akt ^(ser473), p-Erk1/2 and Krox20 protein levels. ^*P < 0.05 and ^**P < 0.01 by one-way ANOVA with Bonferroni’s multiple comparison test.

Figure 4

P0 protein density is reduced in mice expressing the HANI transgene. (A) IEM for P0 in transverse sections of P30 sciatic nerves. (B) Quantification of the number of gold granules per myelin area shows a trend towards the reduction of P0 density in HANI/+ and HANI/+//S63del myelin as compared with WT and S63del, respectively. Twelve sciatic nerve images per genotype were counted from three mice per genotype. ^*P-value = 0.03 by Student’s t-test; error bars represent SEM.

Figure 5

Nrg1TIII overexpression increases cholesterol and fatty acids levels. Lipidomic analysis was performed on P28 sciatic nerves. (A) The level of free cholesterol increased in HANI as compared with WT, whereas no differences were detected between S63del and HANI/+//S63del nerves. (B) Levels of total saturated (palmitic acid, C16:0; stearic acid, C18:0; behenic acid, C22:0 and lignoceric acid, C24:0) and total unsaturated (oleic acid, C18:1; linoleic acid, C18:2; erucic acid, C22:1 and nervonic acid, C24:1) fatty acids and (C) the relative desaturation index (ratio 18:0/18:1) and membrane fluidity index (ratio 18:1/18:2). Data are expressed as μg of cholesterol or ng of fatty acid normalized to μg of total proteins. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001 versus WT by one-way ANOVA with Bonferroni’s multiple comparison test. Five sciatic nerves from different animals per genotype were analyzed.

Figure 6

ER stress levels do not increase in S63del/HANI/+ mice. (A) WB analysis of P28 sciatic nerves lysates shows activation of the ER stress markers Bip/Grp78, Grp94 and P-eIF2-alpha in S63del mice as compared with WT. No further increase was detected in the expression levels of these markers between S63del and HANI/+//S63del mice. Tubulin was used as loading control. One representative experiment of three is shown. (B) Densitometric quantification of Bip, Grp94, calnexin and p-eIF2-alpha relative protein levels. (C) qRT-PCR for BiP, CHOP and spliced XBP-1 mRNA in P28 sciatic nerves. All the stress makers are increased in S63del as compared with WT, but we detected no differences in their levels of expression between S63del and HANI/+//S63del nerves. Each experiment was repeated three times on different pools of three nerves per genotype. ^*P < 0.05, ^**P-value < 0.01 and ^***P < 0.001 by one-way ANOVA with Bonferroni’s multiple comparison test.

Figure 7

The expression of negative regulators of myelination is not altered by Nrg1TIII overexpression. qRT-PCR and WB analysis (with relative densitometric quantification) on P28 nerves for Sox2 (A) and Id2 (B). Both factors are increased in S63del nerves as compared with WT, but remained similarly increased in HANI/+//S63del mice. Each qRT-PCR experiment was repeated three times on different pools of three nerves per genotype. For WB, one representative experiment of three is shown. Tubulin was used as loading control, equal loading is also shown by coomassie staining; ns, not significant, by one-way ANOVA with Bonferroni’s multiple comparison test.

Figure 8

Pharmacological treatment with the TACE inhibitor BMS-561392 ameliorates myelination in DRGs explant cultures. Myelinating DRGs were treated with 1 μm BMS (BMS1) for 2 weeks. As controls, not treated and DMSO-treated DRGs were analyzed. (A) Immunostaining for MBP showing the increase in the number of MBP-positive segments in S63del-treated samples (size bar, 100 μm). Quantification of internode number in (B) WT (see also Supplementary Material, Fig. 8A) and (C) S63del DRGs, normalized to the untreated control. Note how treatment with BMS increases the number of internodes in both the genotypes with similar magnitude. Ten to fifteen DRGs per condition from three independent dissections were analyzed. ^**P < 0.01 and ^***P < 0.001 by one-way ANOVA with Bonferroni’s multiple comparison test; error bars represent SEM.