Muscle pathology without severe nerve pathology in a new mouse model of Charcot-Marie-Tooth disease type 2E

Abstract

Mutations in neurofilament light (NF-L) have been linked to Charcot-Marie-Tooth disease type 2E (CMT2E) in humans. To provide insight into disease pathogenesis, we developed a novel line of CMT2E mice that constitutively express human NF-L (hNF-L) with a glutamic acid to lysine mutation at position 397 (hNF-L(E397K)). This new line of mice developed signs consistent with CMT2E patients. Disease signs were first observed at 4 months in hNF-L(E397K) mice, and consisted of aberrant hind limb posture, digit deformities, reduced voluntary locomotor activity, reduced motor nerve conduction velocities (MNCVs) and muscle atrophy. Reduced voluntary locomotor activity and muscle pathology occurred without significant denervation, and hNF-L(E397K) mice showed relatively mild signs of nerve pathology. Nerve pathology in hNF-L(E397K) mice was characterized by ectopic accumulations of phosphorylated NFs in motor neuron cell bodies as early as 1 month. Moreover, NF organization was altered in motor and sensory roots, with small motor axons being most affected. Peak axonal diameter was reduced for small motor axons prior to and after the onset of overt phenotypes, whereas large motor axons were affected only after onset, which correlated with reduced MNCVs. Additionally, there was a small reduction in the number of sensory axons in symptomatic hNF-L(E397K) mice. hNF-L(E397K) mice are a novel line of CMT2E mice that recapitulate many of the overt phenotypes observed in CMT2E patients and hNF-L(P22S) mice. The cellular pathology observed in hNF-L(E397K) mice differed from that recently reported in hNF-L(P22S) mice, suggesting that overt CMT2E phenotypes may arise through different cellular mechanisms.

Figure 1.

Expression of hNF-L^E397K in mice does not affect expression of other cytoskeletal elements. (A) Schematic of hNF-L protein with 18 reported mutations indicated. White indicates the N-terminal head domain, black indicates the coil-coiled rod domain and gray indicates the C-terminal tail domain. hNF-L^E397K occurs at the end of the rod domain in the most conserved sequence in the intermediate filament family (KLLEGEE). (B) Transgenic lines were identified by PCR analysis of genomic DNA derived from tail biopsies. Four lines of hNF-L and hNF-L^E397K mice were identified, three of which are shown here. hNF-L was not amplified from wild-type genomic DNA. The endogenous hNF-L promoter drove expression of wild-type and hNF-L. For both upper and lower panels: negative control was no genomic DNA; hNF-L and hNF-L^E397K negative controls were wild-type C57Bl6 mice. (C and D) To determine the effect of expressing wild-type and mutant hNF-L on endogenous mouse cytoskeletal proteins, mRNA (C) and protein (D) levels were analyzed. (C) Expression of hNF-L, hNF-L^E397K, mNF-L, NF-M, NF-H and βIII-tubulin was analyzed by reverse-transcription PCR amplification of mRNA purified from spinal cords of wild-type, hNF-L and hNF-L^E397K mice. hNF-L transgenes did not alter transcription of other cytoskeletal components. (D) Sciatic nerve lysates were fractionated on 7.5% SDS–polyacrylamide gels from 6-month-old animals and immunoblotted with antibodies recognizing hNF-L (α-NF70kDa), mNF-L (DA2), NF-M (RMO44), NF-H (CPCA-NF-H) and neuron-specific βIII-tubulin (TUJI). Low levels of hNF-L and hNF-L^E397K proteins were detected in transgenic mice. hNF-L transgenes did not affect expression of endogenous NF subunits or βIII-tubulin. (E and F) Utilizing the endogenous hNF-L promoter did not result in ectopic expression of hNF-L transgenes. Lysates from various tissues were fractionated on 7.5% SDS–polyacrylamide gels from 3–4 (E) and 14 (F) months old animals and immunoblotted with an antibody that recognizes both mouse and hNF-L (DA2). All analysis was performed on at least three hNF-L and hNF-L^E397K mice.

Figure 2.

Removal of endogenous NF-L confirms expression of hNF-L and hNF-L^E397K transgenes. hNF-L and hNF-L^E397K transgenic mice were bred with mNF-L^−/− mice to generate mice in which endogenous mouse NF-L is absent in the presence of the hNF-L or hNF-L^E397K transgenes. (A) Spinal cord lysates were fractionated on 7.5% SDS–polyacrylamide gels from 5–6-month-old animals and immunoblotted with antibodies recognizing mouse and hNF-L (DA2) and GAPDH (α-GAPDH). (B) Relative optical densities of NF-L protein normalized to GAPDH levels.

Figure 3.

Expression of hNF-L^E397K in mice results in aberrant hind limb posture and digit deformities. CMT is associated with aberrant posture of the lower limbs and structural deformities of the feet. To determine whether hNF-L^E397K mice developed similar signs of disease, hNF-L and hNF-L^E397K mice were induced to spread their hind limbs and digits. (A) When lifted by the base of the tail, hNF-L mice spread both hind limbs and digits. (B–C) However, hNF-L^E397K mice were unable to spread their hind limbs. Additionally, the digits on the paws of the hind limbs curled under the paw. This was observed both uni- (B) and bi-laterally (C). (D) Close-up of curled digits in hNF-L^E397K mice. (E) Aberrant hind limb posture and paw deformities were not due to excessive loss of musculature determined by monitoring total weight loss in both hNF-L and hNF-L^E397K mice. Weight was monitored by weighing 26 hNF-L and 35 hNF-LE397K mice once per week.

Figure 4.

hNF-L^E397K mice have reduced voluntary activity. Mice of 1–4 months of age were given free access to running wheels for a period of 7 days. Revolutions were stored on activity wheel counters and then converted into kilometers on the basis of a 5 in. diameter running wheel. (A) Daily and (B) total activity were monitored in a single line of hNF-L and hNF-L^E397K mice. Daily activity was significantly reduced at days 3–5 in line 730 hNF-L^E397K relative to line 673 hNF-L. Total activity over 7 days was also significantly reduced. To determine whether reduced activity was unique to line 730, all lines of hNF-L^E397K and hNF-L were monitored. (C) For all lines of mice, daily activity was significantly reduced relative to control at days 1, 3–5. (D) Additionally, total activity over 7 days was significantly reduced in all lines of hNF-L^E397K mice. Means for each daily activity were analyzed for overall statistical differences by two-way repeated-measures ANOVA, followed by Holm–Sidak post hoc analysis. Means for total distance run were analyzed by unpaired t-test. *P< 0.05, **P< 0.02, †P< 0.008, ‡P< 0.005. For analysis of individual lines, at least six mice were analyzed per genotype. For all lines, at least 13 mice were analyzed for each genotype.

Figure 5.

Expressing hNF-L^E397K results in ectopic phosphorylation of NFs in motor neuron cell bodies. Lumbar spinal cord sections were immunostained with an antibody to NF-L to identify motor neurons. Ectopically phosphorylated NFs were identified with an antibody that recognizes NFs in a phospho-dependent manner. (A and C) hNF-L motor neuron cell bodies do not contain ectopically phosphorylated NFs at 1 (A1–A3) and 3 (C1–C3) months of age. (B and D) However, expression of hNF-L^E397K results in ectopic phosphorylation of NFs in motor neuron cell bodies at 1 (B1–B3) and 3 (D1–D3) months. Merged images confirmed that NF accumulations were within NF-L-positive cells (B3 and D3). Scale bar, 50 μm. Spinal cord sections from four to five hNF-L and three hNF-L^E397K mice were stained for 1 and 3 month time points.

Figure 6.

Reduction of MNCV and reduced axonal diameters in hNF-L^E397K mice. (A) MNCV was measured from axons of the sciatic nerve in hNF-L and hNF-L^E397K mice. There was a statistically significant difference between conduction velocities from hNF-L and hNF-L^E397K mice. Statistical analysis was performed by unpaired t-test. *P< 0.03. (B and C) Distributions of axonal diameters from all axons of the fifth lumbar motor root prior to (B) and after (C) the onset of overt hind limb weakness in age-matched hNF-L and hNF-L^E397K mice. The peak diameter for small motor axons was reduced by 0.5 μm prior to the onset of overt hind limb weakness in hNF-L^E397K mice (B). After onset, the peak diameter was reduced by 0.5 μm for both small and large motor axons in hNF-L^E397K mice (C). Motor axon diameter distributions were analyzed for overall statistical differences utilizing the Mann–Whitney U-test. There was a significant difference in diameter distributions between hNF-L and hNF-L^E397K mice for both presymptomatic (P = 0.011) and symptomatic (P< 0.001) time points. (D) The total number of axons in the fifth lumbar motor root for age-matched hNF-L and hNF-L^E397K mice. Total axon numbers were analyzed for statistical significance by two-way ANOVA. (E) The ratio of axon diameter/fiber diameter (g-ratio) was analyzed from 10% of all motor axons of symptomatic mice. Axons were randomly selected and localized throughout the entire motor root. g-Ratios were analyzed for statistical significance by the Mann–Whitney U-test. hNF-L^E397K g-ratio was significantly larger. (F and G) Distributions of axonal diameters from all axons of the fifth lumbar sensory root prior to (F) and after (G) the onset of overt hind limb weakness in age-matched hNF-L and hNF-L^E397K mice. The peak diameter was unaltered prior to or after the onset of hind limb weakness in sensory axons of hNF-L^E397K mice. Sensory axon diameter distributions were analyzed for overall statistical differences utilizing the Mann–Whitney U-test. There was a significant difference in diameter distributions between hNF-L and hNF-L^E397K mice for both presymptomatic (P < 0.001) and symptomatic (P< 0.001) time points. (H) The total number of axons is reduced in symptomatic hNF-L^E397K mice relative to age-matched hNF-L mice. However, the difference did not reach statistical significance. Total axon numbers were analyzed for statistical significance by two-way ANOVA. (I) g-Ratios were analyzed from 10% of all sensory axons of symptomatic mice. Axons were randomly selected and localized throughout the entire motor root. g-Ratios were analyzed for statistical significance by the Mann–Whitney U-test. hNF-L^E397K g-ratio was significantly smaller. MNCVs were measured in a minimum of 15 mice per genotype. Each point represents the averaged distribution of axon diameters from the entire roots of at least three mice for each group. g-Ratios were measured on at least three mice per genotype. SEM is reported for all g-ratios. However, they are too small to be seen in the figure. SEM motor: hNF-L = 0.005; hNF-L^E397K = 0.004; SEM sensory: hNF-L and hNF-L^E397K = 0.002.

Figure 7.

NFs of proximal sensory and small motor axons are altered in mice expressing hNF-L^E397K. Axoplasmic organization was analyzed in proximal sensory and motor axons from age-matched hNF-L controls and symptomatic hNF-L^E397K mice. Distributions of NND for sensory (A) as well as small (C) and large (E) motor axons from both hNF-L and hNF-L^E397K mice. Peak NND was unaffected in both sensory (A) and large motor (E) axons, whereas peak NND in small (C) motor axons was reduced in hNF-L^E397K axons. In sensory axons, the number of NFs was reduced between 16 and 36 nm, whereas between 40 and 64 nm NF number was increased in hNF-L^E397K axons (A). However, the total number of NFs in sensory nerves was not different between control and hNF-L^E397K axons (B). For small motor axons, NF number was reduced within the range of 28–34 nm in hNF-L^E397K axons (C), and large motor axons had fewer NFs between 8 and 16 nm, whereas they had more NFs between 28 and 40 nm in hNF-L^E397K axons (E). Both small (D) and large (F) motor axons from hNF-L^E397K axons had reduced numbers of total NFs. However, these differences were not statistically significant. Total NF numbers were analyzed for statistical significance by Student's t-test (sensory axons) and Welch's t-test (small and large motor axons). NND was analyzed in at least three mice per genotype in motor axons and four mice per genotype in sensory axons.

Figure 8.

hNF-L^E397K mice display progressive muscle atrophy in hind limb muscles without denervation. To determine the morphology of lower hind limb muscles, cross-sections of the gastrocnemius and TA muscles were analyzed from wild-type, hNF-L and hNF-L^E397K mice. Muscles were isolated, sectioned and stained with hematoxylin and eosin. (A) Gastrocnemius muscle fibers of 1-month-old hNF-L and hNF-L^E397K mice were indistinguishable from wild-type mice. Muscle was analyzed at 1 month to determine if muscle pathology was observed prior to the onset of overt phenotypes. (B) Gastrocnemius fibers of 6-month-old hNF-L^E397K mice were atrophied after the onset of hind limb weakness, whereas hNF-L muscle fibers appeared similar to wild-type mice. (C) TA fibers of 14-month-old hNF-L^E397K mice were atrophied relative to wild-type and hNF-L mice, suggesting that muscle alterations in the hind limb are not limited to the gastrocnemius. Additionally, the formation of clear inclusions was increased in the muscle of aged hNF-L^E397K mice (arrow). Scale bar, 50 μm. (D and E) Gastrocnemius muscle sections were immunostained with antibodies recognizing NFs (α-NF-L) to identify the axon, and fluorescently conjugated α-bungarotoxin to identify the NMJ. There were no differences in the innervation of NMJs in hNF-L (D) and symptomatic hNF-L^E397K (E) mice. Scale bar, 10 μm. Muscle and NMJ analysis was performed on four mice per genotype.