Characterization of New Transgenic Mouse Models for Two Charcot-Marie-Tooth-Causing HspB1 Mutations using the Rosa26 Locus

Abstract

Background: Charcot-Marie-Tooth (CMT) and associated neuropathies, the most common inherited diseases of the peripheral nervous system, remain so far incurable. Three existing murine models of Charcot-Marie-Tooth type 2F (CMT2F) and/or distal hereditary motor neuropathy type IIb (dHMNIIb), caused by mutations in the small heat shock protein B1 gene (HSPB1/HSP27), partially recapitulate the hallmarks of peripheral neuropathy. Because these models overexpress the HSPB1 mutant proteins they differ from the patients’ situation.

Objective: To overcome the possible bias induced by overexpression, we generated and characterized a transgenic model in which the wild type or mutant HSPB1 protein was expressed at a moderate, more physiologically relevant level.

Methods: We generated a new transgenic mouse model in which a human wild type (hHSPB1^WT) or mutant (hHSPB1^R127W; hHSPB1^P182L) HSPB1 transgene was integrated in the mouse ROSA26 locus. The motor and sensory functions of the mice was assessed at 3, 6, 9, 12 and 18 month.

Results: However, the mice expressing the mutant hHSPB1 do not develop motor or sensory deficits and do not show any sign of axonal degeneration, even at late age. Quantitative PCR analyses reveal contrasting tissue-specific expression pattern for the endogenous mouse and exogenous human HSPB1 and show that the ratio of human HSPB1 to the endogenous mouse HspB1 is lower in the sciatic nerve and spinal cord compared to the brain.

Conclusion: These results suggest that expressing the transgene at a physiological level using the ROSA26 locus may not be sufficient to model inherited peripheral neuropathies caused by mutation in HSPB1.

Keywords

Peripheral nervous system disease HSPB1 transgenic mice ROSA26 locus

INTRODUCTION

With a prevalence of 1 in 2500 individuals, Charcot-Marie-Tooth (CMT) disease and related peripheral neuropathies are the most frequent inherited neuromuscular disorders [1]. They are characterized by a length-dependent degeneration of peripheral nerves in a distal to proximal pattern. The neurodegeneration induces progressive muscle wasting and atrophy in the distal limbs. As a consequence, CMT patients suffer from hand and foot deformities, gait disturbance, sensory loss, impaired locomotion and wheelchair dependence in the most severe cases [2] and, so far, the disease remains incurable. CMT is clinically and genetically extremely heterogeneous [1, 3]. It can be classified into mixed motor and sensory neuropathy (HMSN), predominantly motor neuropathy (distal hereditary motor neuropathy: dHMN) or hereditary sensory and/or autonomous neuropathy (HSAN) [4]. Additionally, morphological and electrophysiological criteria differentiate CMT type 1 from CMT type 2 by primary demyelination and primary axonal degeneration, respectively [4]. While some of the 81 CMT-associated genes encode for proteins specifically expressed in Schwann cells or neurons, others are pleiotropic which makes it more challenging to understand how their mutation leads to specific degeneration of the peripheral nerves [3]. One of these genes is HSP27 which codes for the small heat shock protein B1 (HSPB1), a ubiquitously expressed molecular chaperone which canonical function is to preserve cellular proteostasis in stress conditions [5]. Additionally, HSPB1 is known to regulate actin dynamics and cell differentiation and to have anti-apoptotic and antioxidant functions [6 –8]. Up to now, 18 mutations in HSP27 have been linked to CMT [9]. Interestingly, mutations located inside (e.g. HSPB1^R127W or HSPB1^S135F) or outside (HSPB1^P182L) its highly conserved α-crystallin domain have different consequences on the protein function. While the former ones are associated to CMT2F and cause the protein to increase its chaperone activity and overstabilize microtubules, the latter is associated with dHMN and does not affect the chaperone activity of HSPB1 [10, 11]. On the other hand, mutations inside and outside the α-crystallin domain can both induce the hyperphosphorylation of neurofilament and increase acetylation of tubulin [12, 13]. However, the pathogenicity of these deficits is unclear. To better understand the pathomechanisms, a first transgenic mouse model of CMT-causing HSPB1 mutants was generated with the human wild type or mutant (S135F or P182L) HSPB1 cDNA cloned in the nervous system specific Thy1.2 over-expression cassette [13]. At 6 month of age, the transgenic HSPB1^S135F and HSPB1^P182L mutant mice exhibited impaired locomotor performance and muscle strength in the hind limbs. The axonopathy was further corroborated by a decrease in compound motor action potential (CMAP) amplitudes and the loss of large myelinated fibres in the sciatic nerve. In a second model, the wild type or mutant human HSPB1 gene was expressed under the nervous system specific prion protein (PrP) promoter and resulted in a tenfold overexpression of the HSPB1^R136W mutant transgene [14]. However these mice never developed a motor deficit despite electrophysiological and histopathological evidences of axonal loss. In a recently published model, over-expression of the human HSPB1^S135F through the cytomegalovirus (CMV) promoter allowed the development of motor deficits with axonal loss in mutant mice [15]. The common feature of these three models is the overexpression of the transgene. This could have several consequences. First, overexpression of HSPB1 could affect its ability to form hetero-oligomers and, hence, the biological functions that these hetero-oligomers mediate [16]. Second, in human and in model organisms, overexpression of a wild type or a mutant gene often leads to a hyper- or neomorphic phenotype [17, 18]. Copy number variation or cancer-related gene dysregulation are well-known examples where overexpression of a wild type gene leads to a new pathological phenotype [19, 20]. Overexpression of mutant proteins can also cause a gain of an abnormal function such as aggregation, altered binding specificity or mislocalization [21 –23]. When it comes to disease modelling, overexpression may lead to an artificial phenotype [24, 25]. To avoid these potential drawbacks, we generated novel genetically engineered mice expressing the transgene at a moderate level, a situation that better mimics the mono-allelic expression of the mutated gene in autosomal dominant CMT2F/dHMNIIb. The human wild type and mutant transgenes (here abbreviated as hHSPB1) were inserted in the ROSA26 locus which encodes three non-coding transcripts of unknown function and was originally identified in a gene-trap screen in murine ES cells [26, 27]. The ROSA26 locus drives a ubiquitous expression in embryonic and adult mice. We generated a wild type transgenic line (hHSPB1^WT) as well as two mutant transgenic lines, one expressing a mutation (hHSPB1^R127W) located inside the α-crystallin and another mutation (hHSPB1^P182L) outside the α-crystallin domain. Furthermore, we made use of the Cre-loxP system in order to express the transgenes either specifically in the nervous system or ubiquitously. We phenotyped the different hHSPB1 lines but failed to observe functional or histopathological defects. An expression study demonstrated that the level of the exogenous hHSPB1, when compared to the endogenous mouse HspB1 (here abbreviated as mHSPB1), was relatively low in the sciatic nerve and spinal cord of mice expressing the transgene exclusively in nervous cells. However, in the mice expressing the transgene ubiquitously, the mRNA level of the exogenous mutant hHSPB1 and the endogenous wild type mHSPB1 are comparable. Our study suggests that expressing the transgene at a subtle and physiologically relevant level may not be sufficient to model inherited peripheral neuropathies caused by mutation in HSPB1.

MATERIALS AND METHODS

Animal care

All mouse experiments were carried out with approval of the Ethical Committee for Laboratory Animals (University of Antwerp). Mice were housed under the care of the Animal Facility Interfaculty Unit, which is accredited by the Association for Assessment and Accreditation of Laboratory Animals. All experiments were performed on female hHSPB1^WT/+, hHSPB1^R127W/+ and HSPB1^P182L/+ lines and wild type littermates, except when differentially specified.

Generation of the hHSPB1 targeting constructs

Constructs were generated using Gateway recombination (Invitrogen, Thermo Fisher Scientific Inc., Waltham, MA, USA). Briefly, human HSPB1 wild type, R127W and P182L mutations were amplified from constructs described elsewhere [11] and cloned in a pDONR to generate Entry vectors. The final constructs were then generated by recombining the Entry vectors into the Gateway-compatible pROSA26 Destination Vector (pROSA26-DV1 LMBP 6350) [28]. Final constructs contained the full length genes without any tags or additional amino acids.

Chimeric mouse generation and breeding

The linearized plasmid DNA were inserted in G4 embryonic stem (G4 ES, Mount Sinai, Toronto, CA) cells by electroporation [29] and targeted to the ROSA26 locus by homologous recombination as previously described [28, 30]. G418-resistant clones were screened for correct targeting by PCR and Southern blot. Correctly targeted cells were aggregated with outbred Swiss morula, which were then implanted into pseudo pregnant Swiss mice. All ES cell manipulations and transgenic mouse development were performed by T.H., S.G. and J.J.H.. This lead to the generation of three transgenic lines having the human HSPB1 (hHSPB1) integrated in the ROSA26 locus: One carrying the wild type hHSPB1 gene (hHSPB1^WT), and two carrying mutant hHSPB1 (hHSPB1^R127W and hHSPB1^P182L). Transgenic mice from the F2-generation were backcrossed on a C57BL/6J genetic background and bred by D.B.. Transgenic mice expressing the wild type or mutant hHSPB1 in nervous tissues or ubiquitously were obtained by crossing the hHSPB1^WT, hHSPB1^R127W and hHSPB1^P182L lines with Nestin-Cre and Sox2-Cre lines respectively. The Nestin-Cre and Sox2-Cre mouse lines were provided by T.H, S.G. and J.J.H.

PCR analyses

Genotyping of transgenic animals was performed by PCR on a Veriti 96 well Thermal Cycler (Applied Biosystems, Thermo Fisher Scientific Inc., Waltham, MA, USA) using two primer sets. The PCR contained standard 10x PCR buffer, 50 mM MgCl₂, 10 mM dNTPs, 0.10 μM of each primer, 1 unit of platinium Taq polymerase and 100 ng genomic DNA isolated from mouse ear or tail biopsies. The primers for detection of the targeted ROSA26 allele containing the STOP cassette were as follows: ROSA5, AAAGTCGCTCTGAGTTGTTAT; and ROSA_mut3, GCGAAGAGTTTGTCCTCAACC. The reaction resulted in a 215 bp PCR fragment. The wild-type ROSA26 allele was detected by an amplicon of 322 bp using primers ROSA5 and ROSA WT3: GAGCGGGAGAAATGGATATG. The primers for the detection of the Cre recombinase transgene were CRE5: ATGTCCAATTTACTGACCG and CRE3: CGCCGCATAACCAGTGAA. The PCR program was as follows: Initial denaturation for 2 min at 95°C followed by 40 cycles of 30 sec at 95°C, 1 min at 64°C (for the ROSA26 allele detection) or 58°C (for the Cre recombinase detection) and 1 min at 72°C and a final extension of 10 min at 72°C. PCR products were detected on 2% agarose gel.

The genotype was validated by sequencing the human HSPB1 gene after amplification by PCR. This PCR contains 60 ng DNA, 10X TiTaq Buffer, 10 mM dNTP’s, 3M Betaine and TiTaq polymerase, using the following primers: 5′-GGAGATCCGGCACACTGC-3′ and 5′-CAGCTGGCTGACCTGTAGC-3′. The following PCR program was used; initial denaturation for 1 min at 95°C followed by 35 cycles of 30 sec at 95°C, 45 sec at 65°C, 45 sec at 68°C and a final extension of 3 min at 68°C. Sequencing was performed on purified DNA using the BigDye^® Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Thermo Fisher Scientific Inc., Waltham, MA, USA) and separated on an ABI3730xl DNA Analyzer (Applied Biosystems, Thermo Fisher Scientific Inc., Waltham, MA, USA). Resulting DNA sequences were aligned and analyzed with the SeqManTM II software.

Behavioural phenotype of the mice

All behavioural experiments were carried out according to the recommendation of the Ethical Committee for Laboratory Animals of the University of Antwerp. The motor and sensory functions were assessed at 3, 6, 9, 12 and 18 months of age in 10 female mice per genotype. The behavioural assessment was performed by an experimenter (D.B.) blind to the mice genotype.

Tail suspension test

Each mouse was lifted up by the tail at a height of approximately 20 cm and the hind limb spreading reflex versus clasping behaviour were assessed.

Accelerating rotarod

The locomotor performance was assessed with a Five station Rota-Rod Treadmill for mouse (ENV-575M, Med associates Inc., St Albans, VT, USA) according to the manufacturer’s instructions. Briefly, the mice were put in separated sections on an accelerating rotating rod (4 to 40 rpm over a 300-second-period) and the latency to fall was recorded. The time spent clinging on the rotating rod without walking was subtracted from the final score. Each mouse was trained for 5 consecutive days before each test.

Grip strength test

Four-limb grip strength was measured using the Bioseb grip strength tester BIO-G3S (Bioseb, Vitrolles, France). Mice were placed on a grid accessory and pulled firmly backwards by the tail, provoking a grip response. The maximum force exerted on the grid was recorded on the apparatus. The final score was determined by averaging the strength of three trials.

Footprint analysis

The analysis of the mice hind paw footprints was performed as previously described [31]. Briefly, the hind-feet of the mice were coated with non-toxic black ink so that the mice leave a trail of footprints as they walk or run along a corridor to a goal box. The animals were allowed to walk along a 50 cm-long, 10 cm-wide runway. A blank sheet of white paper was placed on the floor of the runway for each run. The toe spreading (distance between the first and last toe) and the plantar length (distance between the tip toe and the heel) of three consecutive steps were measured and averaged.

Hot plate test

The sensitivity to heat was measured using the IITC Life Science Hot plate analgesia meter (IITC Inc. Woodlands Hills, CA, USA) according to the manufacturer instructions. Briefly, each mouse was placed in a glass bottom-less container placed on a platform heated at 52°C. The temperature was chosen according to the mouse strain in order to produce a slight discomfort without inducing pain or injury. The latency before showing a sign of discomfort in the hind paws (licking, fast removal or jump) was recorded. Mice that did not show any sign of discomfort after 20 sec were removed in order to prevent any injury. The final score was determined by averaging the reaction time of three trials.

Nerve conduction studies

We performed nerve conduction studies using the Neuro-EMG-Micro system from Neurosoft (Neurosoft, Ivanovo, Russia). Subdermal 0.4 mm electrodes were used for stimulation and recording on anesthetized mice (Ketamine and Xylazine anaesthetic 10 ml/kg). CMAPs were measured by placing the stimulating electrodes at the sciatic notch and the recording electrodes on the gastrocnemius muscle. We measured CMAPs at supramaximal stimulation. Three consecutive recordings were performed and the highest recorded amplitude was used as the final score.

Histology

Immunohistochemistry on frozen sections

After receiving a lethal dose of Xylazine and Ketamine mixture, 15-month-old mice (3 mice per genotype) were transcardially perfused with 4% paraformaldehyde (PFA). The spinal cord, sciatic nerve and gastrocnemius muscle were dissected, post-fixed in 4% PFA overnight and stored in 30% sucrose PBS solution. The spinal cord lumbar enlargement (containing the neurons innervating the lower limbs) was sliced in 14 μm cross-sections and immuno-stained for HSPB1 (sc-1048, Santa Cruz Biotechnology Inc., Dallas, TX, USA). The gastrocnemius muscle was sliced in 40 μm longitudinal sections and stained with Alexa Fluor 594-conjugated α-bungarotoxin (Molecular Probes, Thermo Fisher Scientific Inc., Waltham, MA, USA) and SV2a (DSHB, Iowa City, IA, USA). Immuno-histochemistry was performed as previously described [32]. Briefly, after permeabilization and blocking of the unspecific epitopes, sections were incubated overnight at 4°C with the primary antibody and 1 hour at room temperature with a fluorophore-conjugated secondary antibody (Jackson Immuno Research Laboratories Inc., West Grove, PA, USA).

Morphometric and EM analysis on semi-thin sections

Sections of the tibial nerve (distal to the sciatic nerve trifurcation) from 15-month-old mice (3 mice for each genotype) were dissected and fixed in a 3.9% Glutaraldehyde for 48 h at room temperature. Samples were stained with unbuffered aqueous 1% Osmium Tetroxide, dehydrated and embedded in araldite epoxy resin. Semi-thin sections were cut on an ultra-microtome, stained with 1% Toluidine Blue and examined by light microscopy. The total number of axons was counted manually and the diameter of the myelinated axon was measured with ImageJ [33]. Transmission electron microscopy (TEM) of the glutaraldehyde-fixed, resin embedded sciatic nerve tissue was performed as previously described [34, 35].

RNA isolation and quantitative PCR

Two-month-old animals were euthanized by CO₂ inhalation and the brain, spinal cord, sciatic nerve and muscle were dissected out, snap frozen in liquid nitrogen and stored at –80°C. The frozen tissue was homogenized in Trizol and total RNA was extracted using the RNeasy Lipid Tissue kit (Qiagen, Venlo, The Netherlands) according to the manufacturer’s protocol. Possible resident DNA contamination was degraded by DNase treatment (Ambion, Thermo Fisher Scientific Inc., Waltham, MA, USA). RNA was then converted into cDNA by using the SuperScript^® III reverse transcriptase (Life Technologies, Carlsbad, CA, USA). Real-time quantitative polymerase chain (RT-qPCR) reactions were done with 10 ng cDNA diluted in SYBR Green I mix (Life Technologies, Carlsbad, CA, USA) and run on a ViiA 7 Real-Time PCR System (Applied Biosystems, Thermo Fisher Scientific Inc., Waltham, MA, USA) with gene-specific primes (mHSPB1:5′-TGGACCCCACCCTAGTGTC-3′/ 5′-GTGACTGCTTTGGGCAACG-3′; hHSPB1:5′-ACGGTCAAGACCAAGGATGG-3/ 5′-AGCGTGTATTTCCGCGTGA-3′, EGFP: 5′-ACGTAAACGGCCACAAGTTC-3′/ 5′-AAGTCGTGCTGCTTCATGTG-3′, ROSA26AS: 5′-AACTGGCTGGAAAACTCCCAT-3′/ 5^′-TTACCTGCACTCTGATTTGCC-3′). All PCR rea-ctions were performed in triplicate. Primers weredesigned making use of Primerbank (www.pga.mgh.harvard.edu/primerbank). Relative gene expression was calculated using the Δ ΔC_T method [36] with four or five housekeeping genes. For the relative quantities of hHSPB1 mRNA, data were normalized to levels of the endogenous mHSPB1 (2^Δct) and expressed as a ratio. The primer sets for hHSPB1 and mHSPB1 had comparable efficiencies (R² = 0.996 and 0.99 respectively). All qPCR analyses were run with biological triplicate (n = 3).

Protein isolation and Western blot analysis

Animals were CO₂-euthanized. Flash-frozen tissue samples were maintained at – 80°C until further processing. For Western Blotting, samples were homogenized in RIPA buffer (50 mM Tris, 150 mM NaCl, 1% NP40, 0.5% Sodium deoxycholate, 0.1% Sodium Dodecyl Sulphate) complemented with protease inhibitors (Complete, Roche Diagnostics, Basel, Switzerland) and phosphatase inhibitors (PhosStop, Roche, Basel, Switzerland). Equal concentrations of proteins, as determined by the BCA protein assay (Thermo Fisher Scientific, Waltham, MA, USA), were mixed with NuPAGE LDS sample buffer, heated to 95°C for 10 min, and then separated by NuPAGE Bis-Tris gel (4–12% polyacrylamide), electro-transferred (XCell SureLock, Invitrogen, Thermo Fisher Scientific Inc., Waltham, MA, USA) onto Nitrocellulose membrane and immunoblotted. Briefly, membranes were blocked in 5% non-fat dry milk in PBS-Tween (0.1%) for 1 h at room temperature and incubated overnight at 4°C with a primary antibody. The membranes were incubated with HRP-linked secondary antibody treated with Enhanced Chemiluminescence ECL Plus kit reagents (Thermo Fisher Scientific Inc., Waltham, MA, USA) and imaged with ImageQuant imager (GE Healthcare, Wauwatosa, WI, USA). GAPDH or alpha tubulin were used as loading controls. All Western blot analyses were run with biological triplicate (n = 3).

We used the following primary antibodies: hHSPB1/Hsp27 (GTX112964, Gene Tex Inc., Irvine, CA, USA), GAPDH (GTX627408, Gene Tex Inc., Irvine, CA, USA), α-tubulin (ab7291, Abcam, Cambridge, UK) and the following HRP-conjugated secondary antibodies: Anti-mouse IgG1 (1070-05, Southern Biotech, Birmingham, AL, USA), anti-mouse IgG2b (1090-05, Southern Biotech, Birmingham, AL, USA) and anti-rabbit (W401B, Promega, Madison, WI, USA).

Statistical analysis

Statistical analyses were performed using GraphPad Prism version 6 (GraphPad, La Jolla, CA, USA). Differences between genotypes over different ages were tested using Two-way repeated measures ANOVA and post hoc Tukey’s analysis. Comparisons of three datasets were analysed by ANOVA or the Kruskal-Wallis test according to the normal distribution of the data. Values are expressed as mean ± s.e.m.. Statistical significance was set at p < 0.05. All analyses were performed blind to mice genotype.

RESULTS

Generation of the hHSPB1 lines

The transgenic HSPB1 mouse lines were generated using a conditional strategy, with the HSPB1 genes integrated specifically in the ROSA26 locus [37]. We developed three constructs containing either the human wild type gene (hHSPB1^WT), the R127W mutation located inside the α-crystallin domain (hHSPB1^R127W) or the P182L mutation located outside the α-crystallin domain (hHSPB1^P182L). The constructs contained a Neomycin resistance cassette (NEO) flanked by two loxP sites situated upstream of the hHSPB1 which can be excised to activate its expression, and an IRES-EGFP reporter (Fig. 1A). Stable transgenic mouse colonies carrying the hHSPB1^WT, the hHSPB1^R127W or the hHSPB1^P182L allele were generated. These founder lines were activated by crossing hHSPB1 heterozygous with either Nestin-Cre (NC-specific) targeting the neurons and some glial cells of the central nervous system [38] or Sox2-Cre (ubiquitous) [39] heterozygous mice. The mice genotype was identified by PCR using Cre internal primers and specific primers to differentiate the wild type ROSA26 allele (322 bp amplicon) from the transgenic ROSA26 locus (215 bp amplicon) (Fig. 1B). The presence of the wild type or mutant hHSPB1 and/or Cre transgene was further validated by Sanger sequencing (Fig. 1C). Expression of the hHSPB1 at the mRNA and protein level was assessed by qPCR and Western blot respectively using human specific HSPB1 antibody and primers. Our results demonstrate that hHSPB1 is expressed in the sciatic nerve of 15-month-old NC-specific hHSPB1^WT/+, hHSPB1^R127W/+ and hHSPB1^P182L/+ but not in the non-activated control mice (Fig. 1D-E). A similar pattern is observed for the expression of the reporter gene EGFP, at the mRNA level (Fig. 1F). The cell-specific expression profile of HSPB1 was evaluated by immuno-histochemistry. Cross-section of the lumbar spinal cord were immuno-stained with an antibody recognising both mouse and human HSPB1. Representative micrographs show a strong immunoreactivity in the large neurons of the ventral horns in non-transgenic mice (Supplementary Figure 1A) as well as in transgenic mice expressing the hHSPB1^P182L/+ mutant in neural cells (NC-specific) (Supplementary Figure 1B) or ubiquitously (Sox2-Cre specific) (Supplementary Figure 1C). The immunoreactivity seems stronger in the ventral horn large neurons of the Sox2-Cre transgenic animals compared to the non-transgenic mice. Because the antibody identify mouse and human HSPB1, the increase in immunoreactivity can likely be attributed to the transgene. A similar pattern can be observed for the transgenic mice expressing the hHSPB1^R127W/+ transgene (data not shown). These results suggest that in the spinal cord endogenous mHSPB1 is predominantly expressed in motor neurons, in agreement with the literature [40], and that exogenous hHSPB1 is also predominantly expressed in motor neurons although the transgene expression is not regulated by the HSPB1 promoter or its regulatoryregions.

Functional assessment of the hHSPB1 lines

The NC specific hHSPB1^WT/+, hHSPB1^R127W/+ and hHSPB1^P182L/+ lines were screened for motor or sensory deficits comparable to a CMT patient phenotype. Mice were tested at 3, 6, 9, 12 and 18 months of age. From 3 to 18 months of age, none of the mice displayed a clasping behaviour when suspended by the tail. The walking pattern, assessed by toe spreading and plantar length on a foot print analysis, did not differ between mutant mice and wild type littermates (Fig. 2A). The mutant mice did not develop weakness in distal muscles (Fig. 2B) as assessed by the grip strength meter. These mice show no deficit in motor abilities (Fig. 2C) or sensitivity to heat (Fig. 2D) when evaluated with the Rotarod and Hot Plate testrespectively. In CMT2F and dHMNIIb patients the CMAPs amplitudes in the peripheral nerves is reduced attesting the axonal loss [41]. To examine if this was the case in our mice, we performed nerve conduction velocity (NCV) studies but failed to find significant differences in CMAP amplitudes (Fig. 2E) or latency (Fig. 2F).

Histopathological analysis

CMT2F and dHMNIIb patients display prominent loss of larger axons in distal sections of the peripheral nerve [4 , 42]. Transgenic mouse models for other CMT2 subtypes can also display axon atrophy in the peripheral nerves [43, 44]. Hence, we quantified the axon number and average diameter in the tibial nerve, distal branch of the sciatic nerve, of 15-month-old mutant and wild-type mice. However the total number of axons (Fig. 3A) and their average diameter (Fig. 3B) in the mutant and wild type mice were comparable. Because the mutation may induce subtle abnormalities in the distal nerve, we screened the semi-thin sections for axon or myelin defects. However the NC-specific hHSPB1^R127W/+ and hHSPB1^P182L/+ mice were indistinguishable from the wild types with no evidence of demyelination/remyelination nor other axon or myelin defects (Fig. 3C–F). Analysis of the distal sciatic nerve fiber ultrastructure by electron microscopy did not reveal any significant difference between mutant and wild type mice (data not shown). Representative fluorescent immunostainings of the neuromuscular junction (NMJ) in the gastrocnemius muscle of 15-month-old mice showed the juxtaposition of the motor neuron axon terminal (SV2a staining in fluorescent green) and the motor endplate (α-bungarotoxin, in fluorescent red) demonstrating that the NMJs are intact and innervated in the non-transgenic (Fig. 3G) as well as the mutant NC-specific hHSPB1^R127W/+ (Fig. 3H) and hHSPB1^P182L/+ (Fig. 3I) mice.

Relative expression of exogenous and endogenous HspB1

The HSPB1 mutations causing CMT2F and dHMNIIb are autosomal dominant, meaning that the expression of both mutant and normal HSPB1 is supported by one allele in patients. In order to know if an imbalance between the mutant HSPB1 (hHSPB1) and the endogenous HSPB1 (mHSPB1) could explain the absence of a neuropathy phenotype in our model, we compared the hHSPB1 and mHSPB1 mRNA levels in muscle, brain, spinal cord and sciatic nerve of the SN-specific hHSPB1^WT/+, hHSPB1^R127W/+ and hHSPB1^P182L/+ mice and their littermate controls. As expected, there was no mRNA expression of the exogenous hHSPB1 in the muscle or in control (non-activated) mice tissue (data not shown). For the activated mice where both endogenous mHSPB1 and exogenous hHSPB1 were expressed, the level of the transgene (hHSPB1) was normalized to the level of the endogenous mHSPB1 in order to obtain an hHSPB1/mHSPB1 ratio, or hHSPB1 relative expression. The hHSPB1/mHSPB1 ratio varies among peripheral and central nervous tissues for the three genotypes (HSPB1^WT/+, hHSPB1^R127W/+ and hHSPB1^P182L/+ shown in Fig. 4A–C respectively). In peripheral tissue, the exogenous hHSPB1 is five to ten times lower than the endogenous mHSPB1. In brain tissue, the mRNA level of hHSPB1 is higher than the level of endogenous mHSPB1 in HSPB1^WT/+, half the level of endogenous mHSPB1 in hHSPB1^R127W/+ mice and slightly lower than the level of mHSPB1 in hHSPB1^P182L/+ mice. Interestingly, the three transgenic lines have an hHSPB1/mHSPB1 ratio significantly lower (p < 0.05) in the sciatic nerve compared to the brain (Fig. 4A–C). The hHSPB1/mHSPB1 ratio was also significantly lower (p < 0.05) in the spinal cord than in the brain for the hHSPB1^WT/+ and hHSPB1^P182L/+ mice. This could partially be attributed to the differential expression of the endogenous mHSPB1. Indeed, a qPCR analysis showed that the mRNA level of the endogenous mHSPB1 varies among tissues and is significantly higher in the sciatic nerve and spinal cord (n = 3; p < 0.01) as compared to the brain (Fig. 4D). The lack of a specific antibody for mHSPB1 prevented us to check these differences at the protein level. We then compared the protein level of hHSPB1 in the sciatic nerve, spinal cord and brain of hHSPB1^WT/+ 15-month-old mice (n = 3). As seen in Fig. 4E, the level of the exogenous protein is lower in the sciatic nerve as compared to spinal cord and brain lysates. This pattern of expression may be a direct consequence of the Nestin-Cre promoter. While Nestin is transiently expressed in neurons, and glial cells which represent the main cell-types in the brain and spinal cord, it is not expressed in Schwann cell which is, with the axons of motor and sensory neurons, the main constituent of the sciatic nerve. However, neuron-restricted expression of the transgene has proven to be sufficient to induce a CMT-like phenotype in CMT2F mouse models overexpressing the transgene [13 –15]. Furthermore, in vivo studies suggest that Schwann cells do not express endogenous HSPB1 [40, 45]. Nonetheless, we decided to explore the effect of ubiquitous expression of the transgene using the Sox2-Cre promotor.

Characterization of the lines expressing the transgene ubiquitously

We assessed the behavioural phenotype of the ubiquitous (Sox2-Cre) hHSPB1^WT/+, hHSPB1^R127W/+ and hHSPB1^P182L/+ mice. However, the mutant ubiquitous hHSPB1^R127W/+ and hHSPB1^P182^L/+ mice did not demonstrate any motor (Fig. 5A–C) or sensory (Fig. 5D) deficit or decreased CMAP amplitudes (Fig. 5E–F), suggesting that they do not develop axonopathy. The absence of axonopathy was further supported by morphometric analysis of the tibial nerve in 15-month-old mutant and wild type mice showing that the total number of axons (Fig. 6A) and the average axon diameter (Fig. 6B) in the mutant and wild type mice were comparable. Semi-thin sections were screened but ubiquitous hHSPB1^R127W/+ (Fig. 6D) and hHSPB1^P182L/+ (Fig. 6E) mice were indistinguishable from non-transgenic animals (Fig. 6C) with no evidence of demyelination/remyelination nor other axon or myelin defects. The integrity of the gastrocnemius NMJ was assessed by immunofluorescent staining showing that NMJs are intact and innervated in 15-month-old non-transgenic (Fig. 6F) and ubiquitous mutant hHSPB1^R127W/+ (Fig. 6G) and hHSPB1^P182L/+ (Fig. 6H) mice. In order to know if the absence of motor deficits and axonopathy was associated with a low expression of the transgene in the peripheral nervous system, we calculated the hHSPB1/mHSPB1 ratio. Although the three Sox2-Cre transgenic lines show a tissue-dependent pattern of expression similar to the Nestin-Cre lines, with an hHSPB1/mHSPB1 ratio higher in the brain compared to the sciatic nerve and spinal cord, the ratio itself differs. In the hHSPB1^P182L/+ (Fig. 7C) and hHSPB1^R127W/+ brain, the level of the exogenous hHSPB1 is two to four times higher than the level of endogenous mHSPB1. In the spinal cord, the level of hHSPB1 is a little bit more than half the level of mHSPB1 (0.53 and 0.59 for hHSPB1^R127W/+ (Fig. 7B) and hHSPB1^P182L/+ (Fig. 7C) respectively). In sciatic nerve tissue, the level of hHSPB1 is near (0.82) or above (1.21) the level of mHSPB1 in the hHSPB1^P182L/+ and hHSPB1^R127W/+ mutant mice respectively. These results demonstrate that we reached a transgene expression level in the sciatic nerve comparable to the endogenous mHSPB1 and therefor physiologically relevant. Similar to what was seen with the Nestin-Cre lines, part of the tissue-associated variability in the hHSPB1/mHSPB1 ratio can be explained by the pattern of mHSPB1 mRNA level, higher in sciatic nerve than the brain (Fig. 7D), and the inverse pattern of hHSPB1 protein level, higher in the brain than the sciatic nerve (Fig. 7E). The similar tissue-dependent expression profile seen in the Nestin-Cre and Sox2-Cre lines suggests that it is not linked to the Cre promotor. To test if it was related to the Rosa26 locus, we compared the mRNA level of the Rosa26 transcript AS (ROSA26AS), encoded by the Rosa26 region [27]. As seen in Fig. 7F, the ROSA26 mRNA levels are similar in sciatic nerve, spinal cord and brain. The calculation of Pearson correlation coefficient using the raw C_T values of the hHSPB1 and ROSA26AS mRNA reveals a weak correlation (ρ= – 0.3168). We also compared the mRNA level of EGFP, the reporter gene whose expression is regulated by an IRES box. Interestingly, EGFP shows a higher expression level in the brain than sciatic nerve (Fig. 7G), a pattern reminiscent of hHSPB1 protein level and hHSPB1/mHSPB1 mRNAratio.

All together, these results demonstrate that a subtle expression of the transgene can be achieved in the central and peripheral nervous system using the Rosa26 locus. However, this physiologically relevant level of expression of the transgene does not allow the development of a CMT-like phenotype. Furthermore, our data show that both hHSPB1 and EGFP transgenes display a tissue-dependent expressionprofile which contrasts with the even expression of the ROSA26AS transcript.

DISCUSSION

The development of transgenic mice that reliably model inherited diseases is a necessary step in the evaluation of therapeutic strategies. The main goal for the generation of the ROSA26 HSPB1 mice was to assess the effect of a moderate expression of HSPB1 mutations which is closer to the patient situation. This model is the first that does not induce a massive over-expression of HSPB1 mutants. It is also the first to model the HSPB1_R127W mutation, located in the α-crystallin domain and leading to CMT2F [41, 46]. Although we achieved a physiologically relevant expression of the transgene, sometimes close to the level of endogenous mHSPB1, our results indicate that moderate expression of HSPB1_R127W or HSPB1_P182L mutant is insufficient to produce a CMT-like phenotype. Distal HMN and CMT2F patients develop motor or sensorimotor deficits in the distal limbs [41]. Hence, the analysis of the sensory and motor functions of our transgenic mice failed to show differences between the mutant mice and their littermate controls. Interestingly, two out of the three published models of CMT-causing HSPB1 mutants failed to observe motor deficits in any [14] or in some [15] of the mutant mice, suggesting that overexpression of the mutation does not always lead to a behavioural phenotype. However, the three studies describe age-dependent axonopathy, as assessed by electrophysiological (NCV studies) and histopathological/morphometric analyses. The sciatic NCV studies that we performed demonstrate that the amplitude and latency of the CMAP are not affected in mutant mice. Axon quantification in the distal sciatic nerve further confirms that there is no axonal loss and histopathological analysis failed to detect axon or myelin abnormalities in the mutant mice. The discrepancy between our results and the published data cannot be attributed to the strain or age of the mice. Indeed, we used a similar mouse strain to Lee’s model [15] and our mice were tested until they reached 18 months compared to ten or twelve in the previously described HSPB1 models. The absence of axonopathy cannot be explained by a failure to express the mutation since the qPCR and Western blot analyses demonstrate the presence of the transgene at the mRNA and protein level in the central and peripheral nervous system, including the sciatic nerve. Additionally, the sequencing analyses confirm that the respective point mutations are present in the transgene.

The activation of exogenous hHSPB1, in our transgenic mice, is conditional to Cre-loxP recombination, allowed by transient expression of Cre recombinase driven by a nervous tissue-specific (Nestin) and ubiquitous (Sox2) promotor. The level of hHSPB1 mRNA relative to mHSPB1 is two times (in brain tissue) to sixty times (in sciatic nerve) lower in the Nestin-Cre hHSPB1 transgenic mouse lines compared to the Sox2-Cre. The restricted number of cell-types expressing Nestin-Cre may partially explain this difference. Additionally, the Cre recombination efficiency associated with each promotor may also account for some discrepancy between the Nestin-Cre and the Sox2-Cre hHSPB1 mice [47, 48]. In our model, the expression of the transgene is driven by the ROSA26 locus. This locus can easily be targeted for homologous recombination, supports ubiquitous and strong but physiological expression of inserted sequences and is not prone to gene silencing effects [27, 49]. Its use successfully led to the generation of transgenic mice modelling gain-of-function mutations, pathogenic mutations and inherited disease [50 –53]. The ROSA26 locus is known to support ubiquitous expression and ROSA26-driven transgene expression has been described in brain, spinal cord and peripheral neurons, although a subpopulation of cells in the brain remains negative [27, 54]. Our qPCR results demonstrate that the ROSA26 locus drives a uniform expression of the ROSA26AS transcript in central and peripheral nervous tissues and induces equal, low or high hHSPB1/mHSPB1 ratio in the sciatic nerve, spinal cord and brain, respectively. With a ubiquitous expression driven by the Sox2-Cre promoter, the mRNA level of the hHSPB1 transgene aligned the level of the endogenous mHSPB1 in the sciatic nerve of the hHSPB1^P182L/+ and hHSPB1^R127W/+ mutant mice. Although our model is closer to the situation seen in CMT2F patients, compared to overexpressing models, the transgene expression proves insufficient to induce a neuropathy. That a moderate but physiologically relevant expression level driven by ROSA26 locus may not always be sufficient to produce a measurable effect is somehow supported by the study of Ruan & al. who combined the ROSA26 promoter with a composite cyto-megalo virus (CMV)/chicken β-actin promoter to enhance the expression of the transgene [50]. The fact that all the previously described HSPB1 transgenic mouse models are overexpression models, also suggests that HSPB1 mutants may require overexpression to lead to a CMT-like phenotype in mice. Because mutant hHSPB1 show comparable levels (in the sciatic nerve) or lower levels (in the spinal cord) than the mHSPB1, the latter could rescue or compensate the potential deficits caused by the former, jeopardizing the development of a phenotype. Using homozygous mice would have raised the level of expression of the transgene and may have increased our chance to observe a phenotype. However, our goal was to precisely study the effect of a moderate and physiologically relevant expression of the transgene.

The results of our qPCR analysis show an unexpected tissue-specific profile for hHSPB1 and EGFP that seems independent from the Rosa26 locus regulation. The lower expression of both transgenes in the sciatic nerve and spinal cord compared to the brain could be explained by possible tissue-specific Cre excision efficiency [48], epigenetic regulation [55] and transgene stability.

Because our rational was to generate a transgenic mouse model that would express the transgene at a physiologically relevant level and because we think that this moderate expression is the very cause of the absence of CMT-like phenotype, this raises the question of the definition of a good animal model when it comes to inherited diseases. The insertion of the mutation in the orthologous mouse gene would ideally model the autosomal dominant condition of the CMT2F patients. Genome editing is now possible via the knock-in or Crispr-cas9 strategy [44, 56] and has recently led to the generation of a transgenic mouse model of CMT2E [44]. Because the value of a complex model organism, such as the mouse, is to provide a reliable tool to study the pathomechanisms at the histopathological and behavioural levels and to test therapeutic strategies, the models need to be associated with some disease-related phenotype. However, the development of a phenotype should not be the unique criterion to define a relevant model. The model should also reproduce, as closely as possible, the disease conditions seen in patients. The use of an inadequate disease model and/or the underestimation of its limits leads to result misinterpretation, misleading extrapolation and failure to validate pre-clinical results in clinical trials. Hence, we think that describing how and why some models fail to produce the expected phenotype may contribute to improve the generation of faithful disease models.

In summary, this study demonstrates that moderate expression of CMT-causing HSPB1 mutations proves insufficient to cause distal axonopathy and associated motor and sensory deficits in mice. This study also reveals an hHSPB1 and EGFP expression profile which seems independent of the Rosa26 locus and suggests tissue specific regulation of the transgene.

COMPETING INTERESTS STATEMENT

The authors declare that they do not have any competing or financial interests.

FUNDING

This work was supported in part by the University of Antwerp, the Fund for Scientific Research (FWO-Flanders), the “Association Belge contre les Maladies Neuromusculaires” (ABMM), the EU FP7/2007_2013 under grant agreement number 2012-305121 (NEUROMICS) and the American Muscular Dystrophy Association (MDA). D.B., S.J., S.G, and T.G. obtained postdoctoral or PhD fellowships from FWO. L.A.S. and S.J. initiated this work and are currently affiliated to respectively the MRC Laboratory of Molecular Biology, Cambridge, UK and the VIB Inflammation Research Center, Ghent University, Belgium. J.H is currently affiliated to the Australian Centre for Blood Diseases, Monash University, Melbourne, Australia. We acknowledge Prof. Dr. C. Ceuterick from the Institute Born Bunge for allowing us to make use of her cryo-section equipment for immunohistochemistry experiments.

Footnotes

Appendix

References

Baets

, De Jonghe

, Timmerman

. Recent advances in Charcot-Marie-Tooth disease. Curr Opin Neurol. 2014; 27(5):532–40.

Barisic

, Claeys

, Sirotkovic-Skerlev

, Lofgren

, Nelis

, De Jonghe

, et al . Charcot-Marie-Tooth disease: A clinico-genetic confrontation. Ann Hum Genet. 2008; 72(Pt 3):416–41.

Timmerman

, Strickland

, Zuchner

. Genetics of Charcot-Marie-Tooth (CMT) Disease within the Frame of the Human Genome Project Success. Genes (Basel). 2014; 5(1):13–32.

Dyck

, Thomas

, Griffin

, Low

, Poduslo

. Peripheral Neuropathy. 4th ed. Philadelphia (PA): WB Saunders; 2005.

Haslbeck

, Vierling

. A first line of stress defense: Small heat shock proteins and their function in protein homeostasis. J Mol Biol. 2015; 427(7):1537–48.

Schneider

, Hamano

, Cooper

. In vivo evaluation of hsp27 as an inhibitor of actin polymerization: hsp27 limits actin stress fiber and focal adhesion formation after heat shock. J Cell Physiol. 1998; 177(4):575–84.

Mehlen

, Mehlen

, Godet

, Arrigo

. Hsp27 as a switch between differentiation and apoptosis in murine embryonic stem cells. J Biol Chem. 1997; 272(50):31657–65.

Arrigo

, Firdaus

, Mellier

, Moulin

, Paul

, Diaz-latoud

, et al . Cytotoxic effects induced by oxidative stress in cultured mammalian cells and protection provided by Hsp27 expression. Methods. 2005; 35(2):126–38.

Benndorf

, Martin

, Kosakovsky Pond

, Wertheim

. Neuropathy- and myopathy-associated mutations in human small heat shock proteins: Characteristics and evolutionary history of the mutation sites. Mutat Res Rev Mutat Res. 2014; 761:15–30.

10.

Almeida-Souza

, Asselbergh

, d’Ydewalle

, Moonens

, Goethals

, De Winter

, et al . Small heat-shock protein HSPB1 mutants stabilize microtubules in Charcot-Marie-Tooth neuropathy. J Neurosci. 2011; 31(43):15320–28.

11.

Almeida-Souza

, Goethals

, De Winter

, Dierick

, Gallardo

, Van Durme

, et al . Increased monomerization of mutant HSPB1 leads to protein hyperactivity in Charcot-Marie-Tooth neuropathy. J Biol Chem. 2010; 285(17):12778–86.

12.

Holmgren

, Bouhy

, Timmerman

. Neurofilament phosphorylation and their proline-directed kinases in health and disease. Journal of the peripheral nervous system: JPNS. 2012; 17(4):365–76.

13.

d’Ydewalle

, Krishnan

, Chiheb

, Van Damme

, Irobi

, Kozikowski

, et al . HDAC6 inhibitors reverse axonal loss in a mouse model of mutant HSPB1-induced Charcot-Marie-Tooth disease. Nat Med. 2011; 17(8):968–74.

14.

Srivastava

, Renusch

, Naiman

, Gu

, Sneh

, Arnold

, et al . Mutant HSPB1 overexpression in neurons is sufficient to cause age-related motor neuronopathy in mice. Neurobiol Dis. 2012; 47(2):163–73.

15.

Lee

, Jung

, Joo

, Choi

, Moon

, Kwak

, et al . Overexpression of mutant HSP27 causes axonal neuropathy in mice. J Biomed Sci. 2015; 22:43.

16.

Arrigo

. Human small heat shock proteins: Protein interactomes of homo- and hetero-oligomeric complexes: An update. FEBS Lett. 2013; 587(13):1959–69.

17.

Miroux

, Walker

. Over-production of proteins in Escherichia coli: Mutant hosts that allow synthesis of some membrane proteins and globular proteins at high levels. J Mol Biol. 1996; 260(3):289–98.

18.

Prelich

. Gene overexpression: Uses, mechanisms, and interpretation. Genetics. 2012; 190(3):841–54.

19.

Shastry

. Overexpression of genes in health and sickness. A bird’s eye view. Comp Biochem Physiol B Biochem Mol Biol. 1995; 112(1):1–13.

20.

Timmerman

, Nelis

, Van Hul

, Nieuwenhuijsen

, Chen

, Wang

, et al . The peripheral myelin protein gene PMP-22 is contained within the Charcot-Marie-Tooth disease type 1A duplication. Nat Genet. 1992; 1(3):171–75.

21.

Herskowitz

. Functional inactivation of genes by dominant negative mutations. Nature. 1987;3296136 219–22.

22.

Khachatoorian

, Ramirez

, Hernandez

, Serna

, Kwok

. Overexpressed Arabidopsis Annexin4 accumulates in inclusion body-like structures. Acta Histochem. 2015; 117(3):279–87.

23.

Bezard

, Yue

, Kirik

, Spillantini

. Animal models of Parkinson’s disease: Limits and relevance to neuroprotection studies. Movement disorders: Official journal of the Movement Disorder Society. 2013; 28(1):61–70.

24.

Saito

, Matsuba

, Mihira

, Takano

, Nilsson

, Itohara

, et al . Single App knock-in mouse models of Alzheimer’s disease. Nat Neurosci. 2014; 17(5):661–63.

25.

Gibson

, Seiler

, Veitia

. The transience of transient overexpression. Nature methods. 2013; 10(8):715–21.

26.

Friedrich

, Soriano

. Promoter traps in embryonic stem cells: A genetic screen to identify and mutate developmental genes in mice. Genes Dev. 1991; 5(9):1513–23.

27.

Zambrowicz

, Imamoto

, Fiering

, Herzenberg

, Kerr

, Soriano

. Disruption of overlapping transcripts in the ROSA beta geo 26 gene trap strain leads to widespread expression of beta-galactosidase in mouse embryos and hematopoietic cells. Proc Natl Acad Sci U S A. 1997; 94(8):3789–94.

28.

Nyabi

, Naessens

, Haigh

, Gembarska

, Goossens

, Maetens

, et al . Efficient mouse transgenesis using Gateway-compatible ROSA26 locus targeting vectors and F1 hybrid ES cells. Nucleic Acids Res. 2009; 37(7):e55.

29.

George

, Gertsenstein

, Vintersten

, Korets-Smith

, Murphy

, Stevens

, et al . Developmental and adult phenotyping directly from mutant embryonic stem cells. Proc Natl Acad Sci U S A. 2007; 104(11):4455–60.

30.

Haenebalcke

, Goossens

, Naessens

, Kruse

, Farhang Ghahremani

, Bartunkova

, et al . Efficient ROSA26-based conditional and/or inducible transgenesis using RMCE-compatible F1 hybrid mouse embryonic stem cells. Stem Cell Rev. 2013; 9(6):774–85.

31.

Ydens

, Demon

, Lornet

, De Winter

, Timmerman

, Lamkanfi

, et al . Nlrp6 promotes recovery after peripheral nerve injury independently of inflammasomes. Journal of neuroinflammation. 2015; 12:143.

32.

Bouhy

, Malgrange

, Multon

, Poirrier

, Scholtes

, Schoenen

, et al . Delayed GM-CSF treatment stimulates axonal regeneration and functional recovery in paraplegic rats via an increased BDNF expression by endogenous macrophages. FASEB J. 2006; 20(8):1239–41.

33.

Schneider

, Rasband

, Eliceiri

. NIH Image to ImageJ: 25 years of image analysis. Nat Methods. 2012; 9(7):671–75.

34.

Groh

, Weis

, Zieger

, Stanley

, Heuer

, Martini

. Colony-stimulating factor-1 mediates macrophage-related neural damage in a model for Charcot-Marie-Tooth disease type 1X. Brain. 2012; 135(Pt 1):88–104.

35.

Verhoeven

, Claeys

, Zuchner

, Schroder

, Weis

, Ceuterick

, et al . MFN2 mutation distribution and genotype/phenotype correlation in Charcot-Marie-Tooth type 2. Brain. 2006; 129(Pt 8):2093–102.

36.

Hellemans

, Mortier

, De Paepe

, Speleman

, Vandesompele

. qBase relative quantification framework and software for management and automated analysis of real-time quantitative PCR data. Genome Biol. 2007; 8(2):19.

37.

Hohenstein

, Slight

, Ozdemir

, Burn

, Berry

, Hastie

. High-efficiency Rosa26 knock-in vector construction for Cre-regulated overexpression and RNAi. Pathogenetics. 2008; 1(1):3.

38.

Haigh

, Morelli

, Gerhardt

, Haigh

, Tsien

, Damert

, et al . Cortical and retinal defects caused by dosage-dependent reductions in VEGF-A paracrine signaling. Dev Biol. 2003; 262(2):225–41.

39.

Lakso

, Sauer

, Mosinger

Jr , Lee

, Manning

, Yu

et al . Targeted oncogene activation by site-specific recombination in transgenic mice. Proc Natl Acad Sci U S A. 1992; 89(14):6232–36.

40.

Ohshima

, Nakakura-Ohshima

, Yamamoto

, Maeda

. Alteration in the expression of heat shock protein (Hsp) 25-immunoreactivity in the dental pulp of rat molars following tooth replantation. Arch Histol Cytol. 2001; 64(4):425–37.

41.

Evgrafov

, Mersiyanova

, Irobi

, Van Den Bosch

, Dierick

, Leung

, et al . Mutant small heat-shock protein 27 causes axonal Charcot-Marie-Tooth disease and distal hereditary motor neuropathy. Nat Genet. 2004; 36(6):602–06.

42.

Harding

, Thomas

. The clinical features of hereditary motor and sensory neuropathy types I and II. Brain. 1980; 103(2):259–80.

43.

Achilli

, Bros-Facer

, Williams

, Banks

, AlQatari

, Chia

, et al . An ENU-induced mutation in mouse glycyl-tRNA synthetase (GARS) causes peripheral sensory and motor phenotypes creating a model of Charcot-Marie-Tooth type 2D peripheral neuropathy. Dis Model Mech. 2009; 2(7-8):359–73.

44.

Adebola

, Di Castri

, He

, Salvatierra

, Zhao

, Brown

, et al . Neurofilament light polypeptide gene N98S mutation in mice leads to neurofilament network abnormalities and a Charcot-Marie-Tooth Type 2E phenotype. Hum Mol Genet. 2015; 24(8):2163–74.

45.

Yamamoto

, Fan

, Wakayama

, Amano

, Iseki

. Constitutive expression of the 27-kDa heat-shock protein in neurons and satellite cells in the peripheral nervous system of the rat. Anat Rec. 2001; 262(2):213–20.

46.

Solla

, Vannelli

, Bolino

, Marrosu

, Coviello

, Murru

, et al . Heat shock protein 27 R127W mutation: Evidence of a continuum between axonal Charcot-Marie-Tooth and distal hereditary motor neuropathy. J Neurol Neurosurg Psychiatry. 2010; 81(9):958–62.

47.

Araki

, Imaizumi

, Okuyama

, Oike

, Yamamura

. Efficiency of recombination by Cre transient expression in embryonic stem cells: Comparison of various promoters. J Biochem. 1997; 122(5):977–82.

48.

Bao

, Ma

, Schuster

, Lin

, Yan

. Incomplete cre-mediated excision leads to phenotypic differences between Stra8-iCre; Mov1011 (lox/lox) and Stra8-iCre; Mov10l1(lox/Delta) mice. Genesis. 2013; 51(7):481–90.

49.

Irion

, Luche

, Gadue

, Fehling

, Kennedy

, Keller

. Identification and targeting of the ROSA26 locus in human embryonic stem cells. Nat Biotechnol. 2007; 25(12):1477–82.

50.

Ruan

, Zhang

, Mu

, Burnley

, Wang

, Coder

, et al . Biological significance of FoxN1 gain-of-function mutations during T and B lymphopoiesis in juvenile mice. Cell Death Dis. 2014; 5:e1457.

51.

Tsika

, Kannan

, Foo

, Dikeman

, Glauser

, Gellhaar

, et al . Conditional expression of Parkinson’s disease-related R1441C LRRK2 in midbrain dopaminergic neurons of mice causes nuclear abnormalities without neurodegeneration. Neurobiol Dis. 2014; 71:345–58.

52.

Johnson

, Hemnes

, Perrien

, Schuster

, Robinson

, Gladson

, et al . Cytoskeletal defects in Bmpr2-associated pulmonary arterial hypertension. Am J Physiol Lung Cell Mol Physiol. 2012;302(5):L474–L484.

53.

Hemnes

, Brittain

, Trammell

, Fessel

, Austin

, Penner

, et al . Evidence for right ventricular lipotoxicity in heritable pulmonary arterial hypertension. Am J Respir Crit Care Med. 2014;189(3):325–34.

54.

Mar

, Yang

, Ma

. Genetic marking and characterization of Tac2-expressing neurons in the central and peripheral nervous system. Mol Brain. 2012; 5:3.

55.

Zhou

, Zhang

, Jiang

, Xu

, Zhang

, et al . Unstable expression of transgene is associated with the methylation of CAG promoter in the offspring from the same litter of homozygous transgenic mice. Mol Biol Re. 2014; 41(8):5177–86.

56.

Yin

, Xue

, Chen

, Bogorad

, Benedetti

, Grompe

, et al . Genome editing with Cas9 in adult mice corrects a disease mutation and phenotype. Nat Biotechnol. 2014; 32(6):551–53.