A Novel HSPB1^S139F Mouse Model of Charcot-Marie-Tooth Disease

Abstract

Charcot-Marie-Tooth Disease (CMT) is a commonly inherited peripheral polyneuropathy. Clinical manifestations for this disease include symmetrical distal polyneuropathy, altered deep tendon reflexes, distal sensory loss, foot deformities, and gait abnormalities. Genetic mutations in heat shock proteins have been linked to CMT2. Specifically, mutations in the heat shock protein B1 (HSPB1) gene encoding for heat shock protein 27 (Hsp27) have been linked to CMT2F and distal hereditary motor and sensory neuropathy type 2B (dHMSN2B) subtype. The goal of the study was to examine the role of an endogenous mutation in HSPB1 in vivo and to define the effects of this mutation on motor function and pathology in a novel animal model. As sphingolipids have been implicated in hereditary and sensory neuropathies, we examined sphingolipid metabolism in central and peripheral nervous tissues in 3-month-old Hsp^S139F mice. Though sphingolipid levels were not altered in sciatic nerves from Hsp^S139F mice, ceramides and deoxyceramides, as well as sphingomyelins (SMs) were elevated in brain tissues from Hsp^S139F mice. Histology was utilized to further characterize Hsp^S139F mice. Hsp^S139F mice exhibited no alterations to the expression and phosphorylation of neurofilaments, or in the expression of acetylated α-tubulin in the brain or sciatic nerve. Interestingly, Hsp^S139F mice demonstrated cerebellar demyelination. Locomotor function, grip strength and gait were examined to define the role of Hsp^S139F in the clinical phenotypes associated with CMT2F. Gait analysis revealed no differences between Hsp^WT and Hsp^S139F mice. However, both coordination and grip strength were decreased in 3-month-old Hsp^S139F mice. Together these data suggest that the endogenous S139F mutation in HSPB1 may serve as a mouse model for hereditary and sensory neuropathies such as CMT2F.

INTRODUCTION

Charcot-Marie-Tooth Disease (CMT) is an inherited peripheral polyneuropathy estimated to affect every 1 in 2500 people (1). Clinical phenotypes for this disease comprise symmetrical distal polyneuropathy (2, 3), decreased or absent deep tendon reflexes, distal sensory loss, foot deformities such as pes cavus (hammer toes), and gait abnormalities (4-6). CMT is analogously named hereditary motor and sensory neuropathy (7), corresponding to the initial description of the disease. More than 70 genes have been implicated in CMT (8), with the majority of these mutations falling into one of two major subtypes: CMT1 and CMT2. CMT1 is most prevalent, encompassing 80% of cases, and has been classified by the demyelination of neurons; this results in reduced motor nerve conduction velocity of the motor neuron (8). CMT2 does not affect nerve conduction velocity, but has been defined by axonal damage, resulting in reduced compound muscle action potential (8).

Genetic mutations in heat shock proteins have been linked to CMT2. Specifically, mutations in the heat-shock protein B1 (HSPB1) gene, which encodes for heat shock protein 27 (Hsp27), have been linked to the CMT2F subtype (9). Hsp27 is part of the family of small Hsps which are chaperone proteins important for cellular protein homeostasis (10). Hsp27 has also been shown to provide neuroprotective effects as well as protection against oxidative stress, aging, and apoptosis (10-12). To better understand the underlying mechanisms that cause the pathophysiology seen in CMT2F, many cell lines and mouse models have been generated; however, the conclusions from these studies have not been definitive (13).

Sphingolipids, which have been implicated in hereditary and sensory neuropathies, are generated de novo by the condensation of serine and palmitoyl-CoA by serine palmitoyl transferase (SPT) which following several enzymatic steps leads to the generation of ceramide. Mutations in SPT are implicated in hereditary sensory and autonomic neuropathy type 1 (HSAN1) via the accumulation of deoxysphingolipids (14), resulting in similar phenotypes to CMT2F. Ceramide serves as the central lipid in sphingolipid metabolism and can regulate cell growth, senescence, proliferation, differentiation, and apoptosis (15). The generation of ceramide is catalyzed by one of six ceramide synthases (CerS) each with fatty acid specificity, resulting in the generation of specific species of ceramide (16). CerS1, which catalyzes the formation of C18 and C18:1 ceramide(s) (17, 18), is highly expressed in the nervous system, skeletal muscle, and testes (19, 20). Deficiency in CerS1 result in decreased ceramide levels, degeneration of cerebellar Purkinje cells and lipofuscin accumulation (21). CerS2, which incorporates very-long chain (VLC; C22-C24) fatty acids into ceramides, has been suggested to regulate myelination (22). In mouse hippocampal neuronal (HT-22) cells, mutant Hsp27 S135F has been shown to decrease localization of CerS1 to the mitochondria (23). These studies implicate the potential for sphingolipids and their generating enzymes in other HSANs such as CMT.

Despite the prevalence of CMT disease there are limited treatments and early diagnostics available. This current study sought to characterize a novel mouse model for CMT2F. These mice contain a S139F mutation in mouse Hsp25, which corresponds to the S135F point mutation in human Hsp27; a common mutation in CMT2F. The data presented here suggest HSP^S139F mice may be an in vivo model to study CMT2F or dHMSN2B.

METHODS AND MATERIALS

Generation of HSP^S139F mice

To generate the HSP^S139F mice, female C57BL/6J mice (JAX stock# 000664) were superovulated and mated. Zygotes were collected and microinjected with ~2–5 pl of a mixture as previously described (57). The microinjection mixture contained: Cas9 mRNA (Trilink, 100 ng/μl), Cas9 protein (PNABio, 50 ng/μl), a single-guide RNA (sgRNA#2165, 50 ng/μl) and an 83nt oligonucleotide donor (Integrated DNA Technologies, oligo#2169, 3 ng/μl). Designed as a truncated guide to reduce off-target potential (60), we used the online software tool “Breaking-Cas” (https://bioinfogp.cnb.csic.es/tools/breakingcas; 58), to identify sgRNA#2165, which targets the sequence: 5’-GAACATGGCTACATCTCT. The template for homology-directed repair (oligo#2169) was designed as an asymmetric oligonucleotide donor (59): 5’-GCCCTCCATGCAAGGAATCAGAACTCACGTGTATTTCCGGGTGAAGCACCGA(A)AGATGTAGCCATGTTCGTCCTGCCTTTCTT. This mixture was designed to generate a cytosine to thymine point mutation (TCT to TTT) in the codon for serine at position 139 in the murine HSPB1 gene. Of 18 offspring, at least 6 founders appeared to have the intended mutation in HSPB1 as detected by PCR and sequencing. Founders were backcrossed to wild-type (WT) C57BL/6J mice, and N1 heterozygotes (HSP^S139F/+) with only the point mutation at S139F were selected to establish the HSP^S139F C57BL/6J strain, designated C57BL/6J-Hspb1^em1Mvw/Mvw (stock # 030923; The Jackson Laboratory) and abbreviated here to HSP^S139F mice. Note, with 100% homology, the sgRNA#2165 was predicted to also cut in a pseudogene on Chromosome 13 (Hsp25-ps1), and indeed all six of the founder animals contained edits at this site when screened. However, stock# 030923 was established using only mice that carried the HSP^S139F allele on Chromosome 5 and the WT form of Hsp25-ps1 on Chromosome 13 (i.e., after segregation of the unintended targeting event).

Genotypes of the offspring were confirmed via genetic sequencing and male and female mice were used for experiments. All animal experiments were approved by the University of Arizona Institutional Animal Care and Use Committee and followed US National Institutes of Health “Using Animals in Intramural Research” guidelines for animal use and the guidelines of the American Veterinary Medical Association.

Rotarod testing

Coordination of mice was tested using Rotarod Rotamex 5 (Columbus Instruments, Columbus, Ohio, USA). Mice were placed individually in one of four lanes of the instrument and recorded simultaneously. Rod speed started at 4 RPM and the rod speed increased 0.5 RPM every 5s until the mouse could no longer maintain the speed. Testing was conducted three separate times with twelve minutes rest between and averaged for analyses.

Grip strength testing

Muscular grip strength of mice was measured using a Grip Strength Meter (Columbus Instruments, Columbus, Ohio, USA). A square metal grid was attached to the meter and used for both forepaw and all paw clasping. Mice were handled by the base of their tails and allowed to grip the metal grid. The mice were then gently and consistently pulled, and grip strength was recorded in grams force. A total of five recordings were taken with five minutes rest between and the average used for analyses.

Mass spectrometry for quantitation of sphingolipids

Sphingolipids were measured as done previously (24-26). Briefly, tissues were homogenized with glass beads using MP Biomedicals FastPrep-24 in 1 mL cell extraction solvent (2:3, 70% isopropanol/ethyl acetate). Sample extract was transferred to glass tubes containing appropriate internal standards and dried down under nitrogen, and then resuspended in mobile phase (1 μM ammonium formate, 2% formic acid in MEOH). Analyses were conducted at Stony Brook University on an HESI probe enabled TSQ Quantum Ultra triple quad mass spectrometer tandem Accela 1250 HPLC system. Further analysis of collected data was performed by the University of Arizona Cancer Center Analytical Chemistry Shared Resource Core. Data were normalized to total lipid phosphate (Pi) (27, 28) and represented as pmol/nmol of lipid Pi.

Immunofluorescence

Mice were anesthetized under isoflurane anesthesia and perfused with 10% formalin (Masterflex Pump Cole-Parmer^® Instruments, Vernon Hills, IL, USA). Brain and sciatic nerve tissues were harvested, kept in 10% formalin for 24 hours and then were transferred into 70% ETOH. Samples were processed, embedded, and sectioned by the University of Arizona Cancer Center (UACC) Tissue Acquisition & Cellular/Molecular Analysis Shared Resource (TACMASR). Tissue slides were deparaffinized at 60°C for 1 hour in a STRATAGENE Hybridizer 600 (Decisive Diagnostics, Jackson, WY, USA) and rehydrated stepwise in 100% Xylene, 100%, 95%, and 70% ETOH, followed by ddH20 and 1x phosphate buffered saline PBS. Antigen retrieval was achieved in a decloaking chamber (BioCare Medical, Pacheo, CA, USA) in 1x citrate buffer heated to 95°C for 45 min. Tissues were blocked in 10% goat serum + 5% BSA in 1x tris phosphate buffered saline with 0.1% Tween 20 (TBS-T). Acetylated anti-α tubulin (ab179484) and anti-neurofilament light antibodies (ab223343) (Abcam, Fremont, CA, USA) were used at dilutions 1:800, and phospho-neurofilament heavy (sc-32730, Santa Cruz Biotechnology, Dallas, TX, USA) at 1:100. Texas Red goat anti rabbit secondary antibody (ab6719, Abcam, Fremont, CA, USA) was used at dilutions 1:1000. No antibody and secondary only antibody controls were used. For myelin staining, slides were incubated at 56°C overnight in Luxol Fast Blue (S3382-25G, MilliporeSigma, KGaA, Dremstadt, Germany). Slides were rinsed in 95% ethanol, differentiated in 0.05% lithiam carbonate for 30 seconds, 70% ethanol for 30 seconds, and rinsed in distilled water. Tissues were counterstained in 0.1% Cresyl Violet solution (ab246816, Abcam, Fremont, CA, USA) for 30 seconds, then dehydrated in 95%-100% ethanol and xylenes prior to mounting. All slides were mounted using Vecta Shield mounting medium (Vector Laboratories, Burlingame, CA, USA).

Statistical Analysis

Analysis Software GraphPad Prism 9 (Boston, MA, USA) was used. Data were analyzed using 2-Way ANOVA with Tukey Multiple Testing Corrections.

RESULTS

Generation of the novel HSP^S139F mouse

Previous studies investigating CMT have utilized transgenic mouse models harboring mutations in human proteins or HSPB1 knockout (KO) mice (29); however, these mouse models may not accurately reflect the effects of mutations in the endogenous Hsp25 protein and don’t account for endogenous Hsp25 activity. Therefore, to generate a novel mouse model containing a mutation in the endogenous mouse HSPB1 gene, CRISPR-Cas9 technology was used, creating a potential new model for the study of CMT2F. In collaboration with Jackson Laboratory, a novel genetically modified mouse model containing a point mutation in HSPB1 was generated, creating a S139F mutation in the mouse Hsp25 that corresponds to the S135F mutation found in the human Hsp27 protein (Table 1). Genetic sequencing confirmed a homozygous cytosine to thymine point mutation resulting in a shift from serine to phenylalanine at this position. These mice harboring the S139F mutation in the endogenous Hsp25 protein were then characterized as a potential novel mouse model for CMT2F.

Table 1:

Sequence alignment of human and mouse HSPB1 gene.

Human	VVEITGKHEERQDEHGYISRCFTR 135
HSPWT Mouse	VVEITGKHEERQDEHGYISRCFTR 139
HSPS139F Mouse	VVEITGKHEERQDEHGYIFRCFTR 139

CRISPR-Cas9 technology was used to generate a novel mouse model containing a mutation in the endogenous mouse HSPB1 gene. A homozygous cytosine to thymine point mutation was introduced at S139F in the mouse HSPB1 corresponding to the S135F mutation in humans.

HSP^S139F mice accumulate ceramides, sphingomyelins, and deoxyceramides

Studies by Schwartz et. al. demonstrated altered ceramide metabolism in sciatic nerves, but not brain tissues from HSPB1 KO mice (23). In addition to decreases in total ceramides, VLC ceramides (C20, C22, C24 and C26) were decreased in sciatic nerves from HSPB1 KO mice (23). Alterations in ceramide and deoxyceramide levels have been implicated in HSAN, neurodegeneration and Purkinje cell death (22, 30, 31). Therefore, mass spectrometry was utilized to measure lipid species in central and peripheral nervous tissues. Sciatic nerves from HSP^S139F mice exhibited no alterations in ceramide levels in comparison to HSP^WT mice (Table S1); however, C14, C18 and C18:1 SM levels were significantly higher in HSP^S139F mice (Table S2). Next, sphingolipid levels in brain tissues were measured from HSP^WT and HSP^S139F mice. Total ceramide levels as well as sphingosine-1-phosphate (S1P) were not different between genotypes (Figure 1A & B). Sphingosine, C16 ceramide, and several species of VLC ceramides (C22, C24:1, C24, and C26) were elevated in the brains of HSP^S139F mice as compared to HSP^WT at 3 months of age (Figure 1C-H). Deoxysphinganine and deoxydihydroceramides were not altered (data not shown); however, deoxyceramides were significantly elevated in brain tissues from HSP^S139F mice (Figure 1I). Total SMs were also elevated in the brain tissues form HSP^S139F mice (Figure 2A). Specific long-chain SM were significantly higher in HSP^S139F mice, including C18 and C18:1, but not C16 (Figure 2B-D). Similar to ceramide levels in the brain, several species of VLC SMs (C22, C24 and C26:1) were significantly increased in HSP^S139F mice (Figure 2E-H). These data demonstrate that while sphingolipid metabolism is not significantly altered in sciatic nerves of HSP^S139F mice, specific species of ceramide, SM, and total deoxyceramides are increased in brain tissues of HSP^S139F mice.

Figure 1. Ceramides and deoxyceramides are altered in brain tissues HSP^S139F mice.

Cerebellar samples were analyzed from HSP^WT and HSP^S139F mice at 3-months of age for A) total ceramide, B) S1P, C) sphingosine, D-) individual ceramide species, as well as I) deoxyceramides using LC-MS/MS and normalized to total lipid phosphate (Pi). Data represent mean ± SEM, n=8 mice. *p<0.05, **p<0.01.

Figure 2. Sphingomyelins are altered in brain tissues HSP^S139F mice.

Cerebellar samples were analyzed from HSP^WT and HSP^S139F mice at 3-months of age for A) total SM, and B-H) individual SM species by LC-MS/MS and normalized to total lipid phosphate (Pi). Data represent mean ± SEM, n=8 mice. *p<0.05, **p<0.01.

HSP^S139F mice do not exhibit an axonal phenotype, but present with cerebellar demyelination

CMT2F presents with a length-dependent axonal phenotype, having increased severity dependent on the length of the nerve (32), and defects in axonal transport and cytoskeletal elements (33-35). In SH-SY5Y human neuroblastoma cells, HSPB1^S135F interfered with anterograde transportation and caused Cdk5-mediated elevations in hyperphosphorylation of neurofilaments (NFs) but did not alter the expression of neurofilament light (NFL) (32). Transgenic mice expressing the human HSP^S135F mutation also exhibited hyperphosphorylated NF-heavy (NFH). Therefore, expression and phosphorylation of NFs in HSP^WT and HSP^S139F mice were examined. Expression of NFL and phosphorylation of NFH remained unaltered in sciatic nerves (Figure S1A & B) and cerebellar tissues of HSP^S139F mice (Figure S2A & B).

Transgenic HSP^S135F mice, harboring mutant human Hsp27, also exhibited reduced acetylation of α-tubulin in sciatic nerves but retained myelination (36). Therefore, levels of acetylated α-tubulin and myelination in central nervous tissues were assessed. Similar to NF, expression of acetylated α-tubulin nor myelination were altered in sciatic nerves (Figures S1C & D) or cerebellar tissues from either genotype (Figure S2C). Interestingly, HSP^S139F mice demonstrated stark cerebellar demyelination (Figure 3). These data suggest that though HSP^S139F mice do not exhibit an axonal phenotype at 3 months of age, the S139F mutation in the endogenous Hsp protein may alter myelination in the cerebellum.

Figure 3. HSP^S139F mice demonstrate cerebellar demyelination.

Brains of 3-month-old HSP^WT and HSP^S139F mice were collected and stained for myelination using Luxol fast blue. A) Representative images of cerebellum and B) quantification of Luxol staining using ImageJ. E-H). Magnification 4x, scale bar = 600μm. Data represent mean ± SEM, n≥7. *p<0.05.

HSP^S139F impairs muscle strength and motor coordination

CMT2F results in gait abnormalities, as well as atrophy and weakness in distal limbs (37). Therefore, HSP^WT and HSP^S139F mice were analyzed for locomotor function, grip strength and gait. Gait analysis revealed no significant changes in the HSP^S139F mice compared to the WT (Figure S3). Rotarod performance was assessed to determine motor coordination. Though maximum tolerated speed was similar between genotypes, total run time was significantly decreased in HSP^S139F mice (Figure 4A & B). Grip strength was measured with a grip strength meter. All paw grip strength was significantly decreased in HSP^S139F mice (Figure 4D). This decrease was primarily due to a significant decrease in forepaw grip strength (Figure 4E &F). These data illustrate distal muscle weakness and motor coordination loss in mice harboring the S139F mutation in HSPB1.

Figure 4. Motor coordination and grip strength are reduced in HSP^S139F mice.

Coordination of mice was tested using a Rotarod Rotamex-5 at 3-months of age. A) Maximum tolerated speed and B) total run time and were assessed. Muscular strength in the D) all paws, E) hind paws and F) forepaws of HSP^WT and HSP^S139F mice was assessed at 3-months of age using a grip-strength meter. Data represent the mean ± SEM for 5 separate tests; n=8 mice; *p< 0.05, ***p < 0.0001.

DISCUSSION

Point mutations in human HSPB1 that encode for Hsp27 have been implicated in CMT2F. This study aimed to characterize a novel mouse model containing a point mutation in the endogenous Hsp25 protein in mice, which corresponds to Hsp27 in humans. VLC ceramides and SM species were increased in the brains of HSP^S139F mice. These sphingolipids have been implicated in many neurodegenerative diseases and linked to HSP^S135F in cells. Alterations in these specific sphingolipids also suggest altered activity of specific ceramide synthases (CerS). In addition, HSP^S139F mice exhibited cerebellar demyelination and loss of locomotor coordination. Similar to patients with CMT2F, HSP^S139F mice also exhibited decreased muscular strength. Overall, the endogenous S139F mutation in HSPB1 resulted in similar clinical phenotypes demonstrated in other models of CMT and human patients.

Previous work from our group demonstrated a decrease in ceramides in sciatic nerves, but not brains, of HSPB1 KO mice as early as one month of age (23). Conversely, in HSP^S139F mice ceramides and deoxysphingolipids were not altered in sciatic nerves, while long chain SMs were significantly increased. Examination of cerebellar tissues revealed significantly increased VLC ceramides and SMs. CerS2, which generates VLC ceramides (C22-26), is vital for proper neuronal functioning and is most active in the first several weeks of post-natal mouse development (22, 30). CerS2 deficient mice lack VLC ceramides in the brain and exhibit a progressive loss of myelin stability in trigeminal nerve axons. Contrary to these findings, HSP^S139F mice exhibited significantly increased VLC ceramides in cerebellar tissues with a concomitant loss of myelin. These data suggest the potential that levels of VLC sphingolipids, either high or low, may regulate myelin stability. Spassieva et. al. demonstrated that ectopic expression of CerS2 in CerS1 deficient mice reversed Purkinje cell loss and neurodegeneration. This reversal of phenotype was due in part to decreased accumulation of sphingoid bases (38). Purkinje cells were not altered and only sphingosine was elevated in HSP^S139Fmice. Together these studies suggest that the balance of sphingolipids, specifically VLC sphingolipids, in cerebellar tissues may influence cerebellar pathology.

Accumulation of SM in cerebellar tissues has been implicated in neurodegeneration associated with lysosomal storage disorders. SMs levels are significantly elevated in several visceral organs and the brain in Nieman-Pick Disease (NPD) patients, resulting from mutations or deficiency in acid sphingomyelinase (aSMase). This increased SM resulted in defects in motor coordination and degeneration of cerebellar Purkinje cells (39). Similarly, deficiency or mutations in acid ceramidase result in Farber disease due to the accumulation of ceramide species within the spinal cord and brain (40, 41). These increased ceramides also lead to Purkinje cell degeneration, as well as impaired motor coordination and balance in mice (42). Though HSP^S139F mice did not exhibit loss of Purkinje cells (data not shown), they did exhibit decreased motor coordination as measured by rotarod performance. Taken together, these studies suggest that the increased SMs in HSP^S139F mice may contribute to altered motor coordination.

Mutations in the initial enzyme for de novo sphingolipid generation, SPT, have been implicated in HSN1 via the accumulation of neurotoxic deoxysphingolipids (14). Treatment with exogenous deoxysphinganine, the precursor for deoxyceramides, resulted in mitochondrial disfunction in motor and sensory neurons (43). Similarly, deoxysphinganine treatment decreased neurite formation in cultured sensory neurons and altered neurofilament structure (14). Patients with HSN1C exhibited slowed conduction velocity in motor nerves of the arms and legs, suggesting damage to both myelin and axons in peripheral nerves. This was accompanied by increased deoxysphinganine, deoxysphingosine and deoxyceramides (44). Deoxyceramide, but not deoxysphinganine, was elevated in brain tissues of HSP^S139F mice, suggesting that these sphingolipid species may contribute to demyelination, loss of motor control, and decreased strength.

Disruption in myelin homeostasis is typically associated with CMT1; however, recent work by Hwang et. al. also demonstrated patients with mutations in NEFL exhibited decreased volume of white matter, suggesting the possibility of myelin defects in patients with mutations associated with both CMT1 and CMT2 phenotypes (45). Our data documented significant cerebellar demyelination in HSP^S135F mice, which we hypothesized would model CMT2F. Interestingly loss of CerS2, which generates VLC ceramides, has been implicated in the disruption of myelination in the brain, including the cerebellum (46). Additionally, patients with NPD accumulate SM and exhibit impaired myelination in the brain (47). Conversely, inhibition or genetic deletion of neutral sphingomyelinase 2 in oligodendrocytes reduced ceramide levels and promoted remyelination in a mouse model (48). Further, mice harboring CerS2 deletion specifically in oligodendrocytes demonstrated reduced myelination in peripheral and central nervous systems, impaired motor coordination, and weakness in paw strength (49). Despite the major role oligodendrocytes play in proper myelination of neurons, there is a gap in knowledge as to the role of sphingolipid metabolism in these cells in the context of CMT2F. While these studies suggest that dysregulation of sphingolipids may alter myelin homeostasis, it is also possible that the HSP^S135F mouse may not accurately model CMT2F. Mutations in HSPB1 are also associated with amyotrophic lateral sclerosis (ALS) which can present with axonal loss and demyelination (50, 51). Mutations in connexon proteins in oligodendrocytes lead to myelination defects and are more commonly associated with CMT1X, a subset which interestingly displays both demyelinating and axonal phenotypes (52, 53). Additional investigation into motor nerve conduction velocity (MNCV) and compound muscle action potential (CMAP) could further validate the HSP^S139F mouse as a novel model for CMT and/or HSN.

Phosphorylated neurofilament and acetylated tubulin have been described as molecular pathological markers for mutant HSPB1 mediated axonopathy. Neurofilaments are an important molecule in axons responsible for radial growth and phosphorylation of these molecules have been established to be a determinant of axon fiber caliber, plasticity and stability (54). Previous models for CMT have demonstrated alterations in neurofilament in spinal cords and sciatic nerves. Cultured motor neurons, derived from E13 mouse spinal cord, transfected with HSBP1^S135F showed disruption in NF networks and aggregation of NFL (33). Sciatic nerves from transgenic HSP^S135F mice also exhibited increased phosphorylated NFH at 10 months of age (29). SH-SY5Y neuroblastomas expressing the human HSP^S135F mutation demonstrated hyperphosphorylated NFs; however, expression remained unaltered. Acetylated α-tubulin is a cytoskeletal component in neurons that regulate transport along microtubules (55), and has also been described as molecular markers for mutant HSPB1 mediated axonopathy (36). Transgenic HSP^S135F mice exhibited a significant decrease in acetylated α-tubulin within sciatic nerves (29, 36). However, transfected HSP^S135F HT-22 cells displayed unaltered acetylated α-tubulin (23). Mice harboring the S139F mutation in the endogenous Hsp25 protein, did not exhibit altered NF expression or phosphorylation, nor altered expression of acetylated α-tubulin, in sciatic nerves or cerebellar tissues at 3 months of age. Interestingly, at this time point motor coordination and grip strength were already decreased. It is possible that the S139F mutation in the endogenous protein did not significantly alter NF or acetylated α-tubulin at 3 months of age, but that this may develop with increasing age, as patients with CMT exhibit highly variable progression of disease, with early onset ranging from 21 to 67 years of age (32).

CMT presents clinically with abnormal gait, loss of distal sensation and reflexes. Recent studies investigating common mutations in CMT2 in C. elegans demonstrated that mutations in Hsp25 did not alter locomotion, but significantly altered muscle morphology (56). In mice expressing the S135F mutant form of Hsp27, locomotor coordination and all paw grip strength decreased after 5 months of age (29). In the present study HSP^S139F mice demonstrated phenotypic changes at 3 months of age. Muscular strength and locomotor coordination were both decreased in the HSP^S139F mice, providing potential insight into the effects of the S139F mutation on clinical phenotypes associated with CMT2F. These data indicate loss of both coordination and muscle strength in HSP^S139F mice, demonstrating similar phenotypes to other models of CMT, as well as clinical presentations of CMT.

The S139F endogenous mutation of HSPB1 in mice, rather than the overexpression of the human mutation or knockout of HSPB1, may prove to be a viable model for CMT. HSP^S135F mice displayed CMT2F-like loss of motor control and reduction in distal strength. Sphingolipid metabolism panels and histological investigation determined that HSP^S135F mice do not exhibit sciatic nerve pathology but do exhibit cerebellar pathology, potentially due to CerS2 activity. Utilization of novel mouse models will allow for investigations into the mechanisms by which these mutations result in disease and may identify novel therapeutic targets. Based on our studies implicating altered sphingolipid metabolism, cerebral demyelination, and decreased muscle strength and locomotor coordination, HSP^S139F mice may serve as an appropriate model for CMT2F or other HSANs.

Highlights.

• Novel murine HSPB1 mutations model human manifestations of CMT.

• Sphingolipid profiles are altered in the brain of HSPB1 mutant mice.

• Novel murine HSPB1 mutation results in cerebellar demyelination.

ABBREVIATIONS

CMT

Charcot-Marie-Tooth

HSPB1

heat-shock protein B1

Hsp27

heat shock protein 27

dHMSNB2

distal hereditary motor and sensory neuropathy type 2B

SM

sphingomyelin

SPT

serine palmitoyl transferase

HSAN

hereditary sensory and autonomic neuropathy

CerS

ceramide synthases

VLC

very-long chain

NF

neurofilament

NFL

neurofilament light

NFH

neurofilament heavy

NPD

Nieman-Pick Disease

aSMase

acid sphingomyelinase

CMAP

compound muscle action potential

MNCV

motor nerve conduction velocity