Sensory-motor deficits and neurofilament disorganization in gigaxonin-null mice

Abstract

Background: Giant Axonal Neuropathy (GAN) is a fatal neurodegenerative disorder with early onset characterized by a severe deterioration of the peripheral and central nervous system, involving both the motor and the sensory tracts and leading to ataxia, speech defect and intellectual disabilities. The broad deterioration of the nervous system is accompanied by a generalized disorganization of the intermediate filaments, including neurofilaments in neurons, but the implication of this defect in disease onset or progression remains unknown. The identification of gigaxonin, the substrate adaptor of an E3 ubiquitin ligase, as the defective protein in GAN allows us to now investigate the crucial role of the gigaxonin-E3 ligase in sustaining neuronal and intermediate filament integrity. To study the mechanisms controlled by gigaxonin in these processes and to provide a relevant model to test the therapeutic approaches under development for GAN, we generated a Gigaxonin-null mouse by gene targeting.

Results: We investigated for the first time in Gigaxonin-null mice the deterioration of the motor and sensory functions over time as well as the spatial disorganization of neurofilaments. We showed that gigaxonin depletion in mice induces mild but persistent motor deficits starting at 60 weeks of age in the 129/SvJ-genetic background, while sensory deficits were demonstrated in C57BL/6 animals. In our hands, another gigaxonin-null mouse did not display the early and severe motor deficits reported previously. No apparent neurodegeneration was observed in our knock-out mice, but dysregulation of neurofilaments in proximal and distal axons was massive. Indeed, neurofilaments were not only more abundant but they also showed the abnormal increase in diameter and misorientation that are characteristics of the human pathology.

Conclusions: Together, our results show that gigaxonin depletion in mice induces mild motor and sensory deficits but recapitulates the severe neurofilament dysregulation seen in patients. Our model will allow investigation of the role of the gigaxonin-E3 ligase in organizing neurofilaments and may prove useful in understanding the pathological processes engaged in other neurodegenerative disorders characterized by accumulation of neurofilaments and dysfunction of the Ubiquitin Proteasome System, such as Amyotrophic Lateral Sclerosis, Huntington's, Alzheimer's and Parkinson's diseases.

Figure 1

Construction of the GAN^-/- mouse. (A) Schematic representation of the disruption of the GAN gene. The targeted construct was designed to replace the endogenous exons 3-5 GAN locus with a neomycin-resistant gene (neo ^R). The diphteria toxin gene, placed at the 3' end of the targeting vector allowed positive selection for the homologous recombination at the GAN locus. Proper targeting produces an open reading frame of only 110 amino acids terminated by a premature stop codon in exon 6. The Δ GAN locus will only produce a Δ gigaxonin of 12 kDa, possibly unstable and entirely lacking the Kelch domain and a portion of the BTB domain. (B, C) Analysis of the genomic DNA from Wild Type (WT), heterozygous (Het) and GAN knock out mice (KO) by polymerase chain reaction (PCR) (B) and Southern blot after BglII digestion (C). The Embryonic Stem cell (ES) corresponds to the original GAN targeted clone. (D) Brain extracts (50 μg) from WT and GAN knock out mouse, and COS cell lysates (cosT; 0,5 μg) transfected with WT-gigaxonin were analysed by immunoblot using anti-Gigaxonin antibody.

Figure 2

Preferential expression of gigaxonin in neuronal tissues. Protein extracts (50 μg) of neuronal (A) and non neuronal (B) tissues from 24 week-old WT and GAN KO mice were immunoblotted with anti-gigaxonin (GigA) and anti-Actin antibodies. Cereb. = cerebellum; Sc. (C) = cervical section of spinal cord; Sc. (L) = lumbar section of spinal cord; SN = sciatic nerve. Specificity of detection is demonstrated by the absence of immunoreactivity in the tissues from KO mice. Note that the abundance of actin (comparable between WT and KO mice) varies in non neuronal tissues. The relative abundance of gigaxonin in multiple tissues was assessed by immunoblotting with GigA antibody (C, top panel). Coomassie blue staining indicates the equal loading of total proteins in all the tissues (C, bottom panel). (D) Temporal expression of gigaxonin, from embryonic stage E15 to adult stage (6 months) was determined in brain and lumbar sections of WT mice by immunoblotting with GigA, and normalized with anti-GAPDH antibody (n = 3; Mann-Whitney test, *, p < 0.05; bars represent standard deviation).

Figure 3

GAN mice present persistent motor deficits. Motor functions were evaluated with a Grip Strength test (A), and a Rotarod test (B). Sensory deficits were recorded with a Hot plate test (C) and Von Frey filaments test (D). Each test was performed every 4 weeks over a 72 week-period (n = 15 mice per genotype). The scores of individual mice are represented at 72 weeks with the mean score represented by a bar for each genotype. The average score was also represented over time for the Forelimb grip test, to show the statistical significance of motor impairment in the GAN mice from 60 weeks of age (two-way ANOVA with Bonferroni post-test: *, p < 0,5; **, p < 0,01 and ***, p < 0,001). Gait analysis of the GAN mice at 72 weeks of age (E). The stride length (i), the width between the right and left paws (ii) and the overlap between the forepaws and hindpaws was measured for each mouse (n = 15 per genotype) and averaged (bar). Analysis of the other GAN^ex3-5 model (GAN ^YY) did not reveal any motor deficits within the first year of age (F) (n = 10 per genotype).

Figure 4

Lumbar motor neurons and axons are preserved in GAN mice. (A) Cross sections of lumbar spinal cord from 48-week-old WT and GAN mice, stained with VaCht (n = 4 animals). Inserts show lower magnification. (B-C) Cross sections of sciatic nerve (B), L5 dorsal and ventral roots (C) from 48-week-old WT and GAN mice, stained with toluidine blue. The numbers of distal axons (B), proximal motor and sensory axons (C) are represented for each mouse (n = 4 per genotype) and averaged (bar). Bars = 50 μm (A, B, C).

Figure 5

Severe disorganization of cytoskeletal architecture in GAN mice. (A) Electron microscopic examination of the axoplasm of GAN nerves revealed a diminution in microtubule content (arrows), an abnormal orientation and an increase in the diameter of neurofilaments (individual neurofilaments indicated by an arrowhead are magnified in the inserts). (B-D) The quantification of the cytoskeletal alteration in 48 week-old GAN mice was performed in sciatic nerves, L5-ventral and dorsal roots (n = 4 mice per genotype; 4 axons per mouse; 3 random pictures of distinct regions per axon; representing a total of 12 fields per mouse). (B) The mean number of microtubules per field was significantly lower in GAN compared to WT nerves (*, p < 0.05, Mann-Whitney test). (C) The alteration of neurofilament orientation was assessed by the measurement of the circularity of individual neurofilaments (circ = 1 and circ<1 representing a perfect circle and an elongated shape, respectively). The left panel displays average circularity scores, measuring the general orientation of the neurofilaments, for individual mice in the three tissues (the mean score per genotype is represented by a bar). The right panels show the standard deviations of the average circularity scores for the same analysis, a measure that is representative of the variations in the orientation of individual neurofilaments within each tissue section (*, p < 0.05, Mann-Whitney test). (D) Neurofilament diameter is significantly increased in GAN mice (10 individual neurofilaments per field; *, p < 0.05, Mann-Whitney test).

Figure 6

Increased abundance of NF subunits in the three GAN models. The relative increase of protein content was obtained by comparing the mean abundance in each of the three GAN models (KO1 = our GAN^ex3-5; KO2 = GAN^YY; KO3 = GAN^ex1) with WT mice (n = 3 mice per genotype, except n = 2 for 48 week-old KO2). (A) Expression levels were quantified in the brain, the lumbar section of spinal cord (Sc-L) and sciatic nerves (SN) by immunoblotting using anti-NFL, NFM and NFH antibodies and normalization with GAPDH antibody. (B) The relative abundance of the gigaxonin's partners MAP1B, MAP1S and TBCB was quantified using the corresponding antibodies, with a similar approach. (Mann-Whitney test, *, p < 0.05; bars represent standard deviation). The immunoblots corresponding to the abundance of the NF subunits and the gigaxonin's partners in brain of 48 week-old GAN models are represented in (C).