Mutations in the gene of the peripheral myelin protein zero (P0) give rise to the peripheral neuropathies Charcot-Marie-Tooth type 1B disease (CMT1B), Déjérine-Sottas syndrome, and congenital hypomyelinating neuropathy. To investigate the pathomechanisms of a specific point mutation in the P0 gene, we generated two independent transgenic mouse lines expressing the pathogenic CMT1B missense mutation Ile106Leu (P0sub) under the control of the P0 promoter on a wild-type background. Both P0sub-transgenic mouse lines showed shivering and ultrastructural abnormalities including retarded myelination, onion bulb formation, and dysmyelination seen as aberrantly folded myelin sheaths and tomacula in all nerve fibers. Functionally, the mutation leads to dispersed compound muscle action potentials and severely reduced conduction velocities. Our observations support the view that the Ile106Leu mutation acts by a dominant-negative gain of function and that the P0sub-transgenic mouse represents an animal model for a severe, tomaculous form of CMT1B.
The hereditary motor and sensory neuropathies are a clinically and genetically heterogeneous group of diseases of the peripheral nervous system of humans (Young and Suter, 2001, 2003). Among the different forms of these diseases is the Charcot-Marie-Tooth (CMT) type 1 neuropathy (Kamholz et al., 2000). CMT1 is characterized by distal muscle weakness and atrophy, reduced nerve conduction velocities, and de- and remyelination with onion bulb formation as observed in sural nerve biopsies (Dyck et al., 1993). Among the main forms of CMT1, CMT1B is caused by mutations in the major peripheral myelin protein zero (P0; Hayasaka et al., 1993a,c; Kulkens et al., 1993; Shy et al., 2002). Mutations in P0 also give rise to severe disorders such as Déjérine-Sottas syndrome (DSS) and congenital hypomyelinating neuropathy (Hayasaka et al., 1993b; for reviews see Warner et al., 1996; Martini et al., 1998; Kamholz et al., 2000; Bennett and Chance, 2001; Shy et al., 2002).
P0 is a transmembrane recognition molecule that belongs to the Ig superfamily. It acts as a homophilic adhesion molecule involved in the formation and maintenance of the intraperiod and major dense lines that are instrumental in the compaction of myelin lamellae (Schneider-Schaulies et al., 1990; Martini, 1994; Martini et al., 1995a). Crystallographic analysis supports the notion that the extracellular domain of the molecule is involved in interdigitation of P0 tetramers on opposing myelin lamellae (Shapiro et al., 1996; Inoue et al., 1999). Experiments on the P0-deficient mouse indicate a role for P0 not only in myelin formation, but also in maintenance of myelin integrity (Martini and Schachner, 1997).
P0 null mutant mice have been considered as a model of patients with DSS that are homozygous for functional null mutations (Giese et al., 1992; Martini et al., 1995b; Warner et al., 1996). The heterozygous P0 null mutant mouse represents a late onset, milder neuropathy, and has been considered as a model for CMT1B patients carrying a loss of function mutation in one allele (Martini et al., 1995b; Martini, 1999). The majority of human P0 mutations is heterozygous and causes a more severe phenotype than that of heterozygous P0-deficient mice. It has been suggested that these mutations act through gain of function (Hayasaka et al., 1993b; Kirschner and Saavedra, 1994). The possibility that some mutations may act through gain of function and others through loss of function could explain, at least to some extent, the findings that patients carrying mutations in the P0 gene are either affected by the mild form of CMT1B or more severe forms of the disease (Martini et al., 1995b).
To develop a mouse model of a severe and early onset form of CMT1B, we have generated a transgenic mouse line with a substitution of isoleucine at residue 106 to leucine (P0sub; Gabreëls-Festen et al., 1996). This mutation in the human is characterized by the occurrence of many folded myelin profiles, termed tomacula, appearing as focally thickened myelin passing into thinner myelin sheaths within the same internode. Fibers outside tomacula are frequently devoid of myelin or are thinly (re)myelinated. Onion bulbs are composed of thin Schwann cell layers and contain many double basement membranes. Uncompacted myelin is encountered only in few fibers. Here, we report that transgenic mice expressing this point mutation in conjunction with the two endogenous wild-type P0 alleles show a similar phenotype, including severe muscle weakness, hypomyelination, tomacula formation, and low conduction velocities, as well as retarded and arrested myelin development.
The combined observations support the view that the Ile106Leu mutation acts by dominant-negative gain of function and that the P0sub-transgenic mice represent the first authentic model for the severe and early onset form of tomaculous CMT1B in humans.
Transgenic mice expressing the P0 substitution mutation show motor abnormalities
An A→T nucleotide exchange was introduced into exon 3 of the P0 gene to express the pathogenic P0 missense mutation Ile106Leu (Gabreëls-Festen et al., 1996) under the control of the P0 promoter in addition to endogenous P0 expression. The expression construct (termed P0sub; Fig. 1) included 6 kb of 5′ flanking sequence, all exons and introns, and the natural polyadenylation signal (Feltri et al., 1999). The P0 promoter was previously shown to regulate the expression of lacZ in parallel with the endogenous P0 promoter during postnatal development and after Wallerian degeneration (Feltri et al., 1999), and to produce structurally normal myelin when P0 was expressed in P0−/− mice (Wrabetz et al., 2000).
We produced two founder P0sub mouse lines (P0sub1 and P0sub3) which transmitted the transgene to their offspring. The transgenic mice of both lines shivered dramatically and showed different signs of muscle weakness pronounced from the fourth week onwards. For example, when placed on a pole they were either not able to grab the pole with their hindlimbs or slipped down if they tried to grab the pole. In addition, the transgenic mice frequently dragged their paws while walking and clenched their paws while hanging at their tails. The transgenic males were not able to breed, most likely due to impaired motor abilities.
The transgenic P0 mRNA is overexpressed compared with the endogenous P0 transcripts in P0sub mice
To determine the ratio between transgenic and endogenous P0 message in P0sub mice, we used an RT-PCR assay based on a method described by Feltri et al. (1999), exploiting a DdeI site present in the sub transgene, but not in the endogenous P0 (Fig. 2 A). After DdeI digestion of RT-PCR products from sciatic nerves, a 228-nt band, corresponding to the transgenic message, was observed in all samples from transgenic mice, and appeared more intense than a 245-nt band representing the endogenous P0 mRNA (Fig. 2 B). The 228-nt band was specific, as omission of reverse transcription product did not produce any bands (unpublished data), and DdeI digestion of RT-PCR product from wild-type animals only yielded the 245-nt band (Fig. 2 B). Densitometry revealed that in P0sub1 and P0sub3 mice, the transgenic P0 mRNA is approximately sixfold overexpressed relative to the endogenous P0 mRNA (Table I).
P0 protein is not overexpressed in myelin of P0sub mice
The content of P0 within peripheral nerves was estimated by immunoblot analysis of total nerve homogenates (Fig. 3, A and B). The P0-immunoreactive bands of transgenic mice in the founder lines P0sub1 (Fig. 3 A) and P0sub3 (Fig. 3 B) were decreased by ∼10-fold when compared with equal protein amounts of wild-type mice (Table II). This finding might reflect the strong hypomyelination observed in peripheral nerves (see below), but does not relate to the amount of P0 protein incorporated into myelin.
To investigate if the abnormal phenotype of P0sub mice is the result of an overexpression of P0 protein in myelin, we characterized the myelin fraction of peripheral nerve homogenates by immunoblot analysis (Fig. 3, C and D). We obtained different results for the two founder lines: although the amount of P0 in the myelin fraction of P0sub1-transgenic mice (Fig. 3 C) was reduced by approximately fivefold compared with wild-type mice, the P0sub3-transgenic mice (Fig. 3 D) showed a P0 content in the range of that seen in wild-type mice (Table II). These data exclude the possibility that an over- or underexpression of P0 protein in peripheral myelin may cause the abnormal phenotype of P0sub-transgenic mice.
Light microscopic and ultrastructural abnormalities in the peripheral nervous system of the P0sub-transgenic mice
To obtain quantitative histopathological data, we investigated the two major branches of the femoral nerve of P0sub1 and P0sub3 mice, comprising the motor quadriceps and cutaneous saphenous nerve, by EM. The quadriceps nerve contains ∼550 myelinated axons that are predominantly of larger caliber, whereas the saphenous branch comprises ∼650 myelinated axons of predominantly smaller caliber (Lindberg et al., 1999). In addition, and as opposed to the quadriceps nerves, unmyelinated nerve fibers are abundant in the saphenous branch.
In 6-mo-old wild-type mice, the two major branches of the femoral nerves were completely myelinated. However, in both lines of the transgenic mice, myelination was severely impaired in both branches of the femoral nerve, but axon numbers were not significantly reduced in these nerves (568 ± 14.8 in the mutants and 554 ± 16.7 in wild types; P > 0.3). The typical pathological profiles in the cross-sectioned femoral nerves were abnormally folded myelin profiles and myelin thickenings of normal compaction (∼12 and 10% of all axonal profiles of lines P0sub1 and P0sub3, respectively; Fig. 4 and Fig. 5). In the following, we designate these abnormalities as tomacula. In addition, abnormally thinly myelinated profiles (∼9 and 10% in P0sub1 and P0sub3, respectively) or axons completely devoid of myelin (43 and 48% in P0sub1 and P0sub3, respectively) were abundant (Fig. 4 and Fig. 5 A). Furthermore, supernumerary Schwann cells in the form of onion bulbs were frequently seen (Fig. 5 A). These abnormal features were also seen in sciatic nerves (Fig. 5 C) and spinal roots (Fig. 5 D), and are highly reminiscent of those described for sural nerve biopsies from a CMT1B patient carrying the same P0 mutation as the transgenic mice (Gabreëls-Festen et al., 1996).
As a next step, we investigated single-fiber preparations from quadriceps nerves of both wild-type and mutant mice. In wild-type mice, myelin sheaths were regularly shaped with nodes of Ranvier forming the borders of apposing sheaths (Fig. 5 E). By contrast, in both transgenic lines, all fibers prepared were of abnormal appearance (50 fibers from 2 different mutant mice of each line). Each of the fibers showed tomacular swellings that were either internodal or at a paranodal position (Fig. 5 F). In addition, and in support of the electron microscopic analyses, the aspects not associated with tomacula showed much thinner myelinated profiles than those of wild-type mice, or lacked myelin completely.
Electrophysiological investigations of adult mice of the P0sub1 line revealed a robust reduction of amplitudes of compound muscle action potentials of plantar muscles from 12.8 mV (±2.9 mV) in wild-type mice to 1.1 mV (±0.19 mV) in the mutants. Mean nerve conduction velocity was reduced from 40.2 m/s (± 2.9 m/s) to 2.0 m/s (± 0.3 m/s, P < 0.001) with dispersed response. F-wave latency could not be recorded at this age (Fig. 6), but in 2-mo-old mutants, a dramatically prolonged F-wave (4.7 ± 0.4 ms in wild-type mice vs. 48 ± 5.5 ms in the mutants, P < 0.001) could be recorded. Similar values were obtained from line P0sub3 (unpublished data).
To investigate whether the severe myelinopathy of the transgenic mice developed at onset of myelination or was the result of myelin degeneration, we investigated the quadriceps branch of the femoral nerve of wild-type and transgenic mice at 4 and 10 d, and at 1 and 4 mo of age.
In 4-d-old wild-type mice, ∼50% of the fibers of the quadriceps nerve showed myelin sheaths that were all of thin appearance (Fig. 4). In 10-d-old mice, almost all fibers were myelinated, including myelin sheaths of normal and low thickness (Fig. 4). During the following stages, all myelinated fibers of the wild-type mice increased in diameter and achieved their final size at 4 mo.
In both mutant lines, there was no detectable myelin at postnatal d 4 and only very few fibers showed myelinated aspects at postnatal d 10 (Fig. 4). At 1 mo after birth, three groups of myelinated fibers were detectable comprising visually normal myelin, myelin tomacula, and abnormally thin myelin (Fig. 4). These myelin-like sheaths constantly increased in number up to 6 mo after birth, the oldest stage investigated. In spite of this increase, ∼43 and 48% of the nerve fibers of P0sub1 and P0sub3, respectively, remained unmyelinated (see above and Fig. 4). It is of note that the developing Schwann cells of the mutant showed an abnormal cytological appearance in that many of them contained abundant cytoplasmic myelin ovoids (unpublished data).
The phenotype of P0sub mutants is caused by a P0 missense mutation
To develop an authentic mouse model for one severe and early onset form of CMT1B, we have generated two independent transgenic mouse lines expressing the pathogenic CMT1B missense mutation Ile106Leu (Gabreëls-Festen et al., 1996) under the control of the P0 promoter (Feltri et al., 1999) on a wild-type genetic background. Both P0sub mouse lines showed severe impairment of their motor abilities with remarkable muscle weakness. Morphological investigation of peripheral nerves of these mice revealed a severe myelinopathy that includes features of early onset dysmyelination, such as prominent hypomyelination and tomacula formation in all nerve fibers. These characteristics are very similar to the pathological phenotype of the severely affected CMT1B patient that has initially been described as suffering from a congenital demyelinating motor and sensory neuropathy (Gabreëls-Festen et al., 1990). Later on, when genetic analyses were available, this patient was identified to carry an Ile106Leu substitution mutation in the P0 gene and was classified as a CMT1B patient (Gabreëls-Festen et al., 1996). Electrophysiologically, we found strongly reduced nerve conduction velocities of only 2 m/s that again reflect the severe phenotype of the aforementioned patient displaying conduction velocities of 7 m/s. In our P0sub mutants, we observed a strong retardation and partial arrest of the myelination process during development, which is again in line with the pathological features of the aforementioned patient, whose pathological abnormalities have been identified at early childhood (Gabreëls-Festen et al., 1990, 1996).
We investigated the relative expression of transgenic and endogenous P0 message in P0sub mice and found that in P0sub1 and P0sub3 mice, the transgenic P0 mRNA is approximately sixfold overexpressed relative to the endogenous P0 mRNA. It is of note that a comparably high overexpression of the full-length P0 transgene (approximately by factor 7 in the mutant Tg80.2) containing the P0 wild-type gene leads to severe dysmyelination characterized by impaired axon sorting and arrest of myelin formation at the promyelin stage (Wrabetz et al., 2000; Yin et al., 2000). The dysmyelinating neuropathy in our model is strikingly different from that observed in the P0-overexpressing Tg80.2 mutants because impaired sorting and persistence of axon bundles of larger calibers were never observed in our P0sub mice. In addition, P0 protein is not overexpressed in the myelin fraction of adult P0sub mutants, whereas P0 protein was overexpressed in the adult Tg80.2 mutant by a factor of ∼1.4 as evaluated by quantitative immuno-EM of the few myelin structures that were detectable in the transgenic overexpressors (Yin et al., 2000). Most importantly, both the P0sub1 and the P0sub3 lines developed tomaculous myelin structures in all peripheral nerve fibers, although the P0 protein levels differed by a factor of ∼5 in the myelin fractions from the two sublines and never exceeded those of wild-type littermates. Occurrence of tomacula cannot be explained by the overexpression of P0 transcripts because tomacula have never been observed in transgenic mice overexpressing P0 (Wrabetz et al., 2000; Yin et al., 2000). Another transgenic mutant developed in our laboratory in parallel with P0sub1 and P0sub3, which expresses a human P0 mutation causing a DSS, does not produce myelin tomacula either, but is characterized by reduced myelin thickness (unpublished data). Therefore, we conclude that tomacula formation and most probably also the observed arrested myelination is a unique feature of P0sub1 and P0sub3 mice reflecting the effects of the substitution mutation.
The P0sub mutation is consistent with a dominant-negative gain of function
The most common features of CMT1B are hypo- and demyelination, combined with onion bulb formation. However, based on additional neuropathological features, two divergent forms of the disorder have been described (Gabreëls-Festen et al., 1996). The first form is characterized by the frequent occurrence of uncompacted myelin (∼20–70% of myelinated fibers), whereas tomacula are almost absent (<1%). These features are most similar to the phenotype of heterozygous P0-deficient mice (Martini et al., 1995b), which thus express the characteristic features of these forms of CMT1B. The pathology of the other type of CMT1B, including the phenotype associated with the Ile106Leu substitution, is dominated by the abundant occurrence of tomacula and a lack of uncompacted myelin (<1%; Gabreëls-Festen et al., 1996). To date, tomaculous forms of CMT1B are described for six further point mutations: Ile33Phe, Ser49Leu, Lys67Glu, Asn93Ser, Lys101Arg, and Asn102Lys (Thomas et al., 1994; Gabreëls-Festen et al., 1996; Tachi et al., 1997; Nakagawa et al., 1999; Sindou et al., 1999; Fabrizi et al., 2000), of which the molecular pathology is unknown and which remain to be investigated in transgenic mouse models. We have shown that, in the presence of two P0 wild-type alleles, the P0sub-transgenic mice develop the same abnormal features and pathological characteristics as the human mutation Ile106Leu, which is a conservative amino acid substitution. It is worth mentioning in this context that this abnormal phenotype is not caused by a functional ablation of the mutant P0 allele in the sense that it causes a null mutation because heterozygous P0-deficient mice show normal development and maintenance of myelin until 4 mo of age, after which they display decompaction and degeneration of myelin. P0sub-transgenic mice, by contrast, show a severe and early onset dysmyelination, accompanied by the formation of tomacula in the absence of myelin decompaction. Thus, our results strongly suggest that the pathological phenotype of tomacula in P0sub-transgenic mice is the result of a dominant-negative gain of function effect of P0 protein that can impair the normal function of P0.
With regard to the developmental aspect of our work, we cannot rule out that the severe phenotype of the two P0sub-transgenic mouse lines is partially due to an overexpression of the mutant protein during development, possibly leading to an increased instability of the protein, as seen previously for P0 wild-type protein when transgenically overexpressed on a wild-type background (Wrabetz et al., 2000; Yin et al., 2000). However, we consider this possibility as sole cause for the impaired myelin development as unlikely. First, as mentioned above, the dysmyelinating phenotype in our mutants is different from and less severe than that of the Tg80.2 mutant studied by Wrabetz et al. (2000). Second, the frequent profiles of dysmyelinated fibers in the biopsies from the Ile106Leu patient are in line with the view that the mutation not only causes tomacula formation, but also impairs myelin formation as seen in the P0sub mice. Third, tomacula formation occurs in both P0sub lines independently of the amount of P0 protein expression in the myelin fraction. This argues that not the levels of the combined mutant and wild-type P0 proteins determine the characteristic features of the resultant mouse lines, but that the tomacula seen in the two independent founder P0sub-transgenic mice are the true consequence of this mutation. Thus, our P0sub-transgenic mice represent an animal model of the severe tomaculous form of the human mutation, as originally described by Gabreëls-Festen et al. (1996).
The question of whether this mutation also leads to abnormal expression of other proteins involved in myelination, such as PMP22, MAG, and periaxin, which produce tomacula (Rebai et al., 1989; Adlkofer et al., 1995, 1997; Carenini et al., 1997; Suter and Nave, 1999; Gillespie et al., 2000; Boerkoel et al., 2001; Cai et al., 2001, 2002; Guilbot et al., 2001; Takashima et al., 2002), remains presently unanswered. Tomacula are formed by homo- and heterozygous deletions of the PMP22 gene and the homozygous ablations of the myelin-associated glycoprotein MAG and periaxin. Because some of these molecules have been shown to interact with each other, such as P0 with PMP22 (D'Urso et al., 1999), and because dysregulations of a substantial set of myelin proteins in the peripheral nervous system of mutant mice deficient in a particular myelin protein have been repeatedly observed (Martini and Schachner, 1997; Menichella et al., 2001), it is possible that dysregulation in expression of the proteins so far recognized as causing tomacula may indicate a common mechanism of molecular pathology. The mechanism underlying the complex network of interdependent synthesis, degradation, and subcellular localization of the individual myelin proteins that may be affected by a mutation in one gene will remain to be investigated to allow a molecular interpretation of the abnormal phenotypes in patients affected by P0 mutations. The availability of the first mouse model carrying the human P0 mutation causing tomacula is an important step toward elucidating these mechanisms.
Materials And Methods
Generation of P0sub-transgenic mice
To generate a mouse mutant expressing the pathogenic substitution P0 Ile106Leu (P0sub; Gabreëls-Festen et al., 1996; Fig. 1) under the control of the P0 promoter (Feltri et al., 1999), the 2.6BsrGIsense vector, containing a 2.6-kb BsrGI fragment with exons 2–5 of the mouse P0 genomic DNA, was used for mutagenesis. The adenine at bp 1644 (ATG start codon in the genomic DNA as bp 1) was substituted to thymine using the Seamless PCR cloning kit (Stratagene) according to the manufacturer's protocol. In brief, the whole vector was amplified with primers flanking and containing the mutation and was re-ligated. A StuI–Eco47II fragment encoding exon 2 and mutated exon 3 of the resulting 2.6BsrGIsense-P0sub vector was cloned into the mP05.7blue vector (Lemke et al., 1988; Feltri et al., 1999), containing exons 2–6. Finally, an EcoRI fragment of the vector mP0Ewtblue (You et al., 1991; Feltri et al., 1999) comprising the P0 promoter and exon 1 was inserted into the EcoRI-digested mP05.7blue-P0sub vector.
The resulting transgenic constructs were verified by sequence analysis, and 12.3-kb fragments were excised from the vectors using SpeI and XhoI. The transgenes were microinjected into Friend virus B-type susceptibility (FVB) zygotes using standard techniques (Hogan et al., 1994). Three P0sub founder mice were identified by PCR and were crossed with wild-type FVB mice. Breeding lines of two P0sub founders (called P0sub1 and P0sub3) were crossed to FVB mice.
Sciatic nerves were dissected from 6-wk-old P0sub1- and P0sub3-transgenic mice and from the respective wild-type littermates. Total RNA was prepared using the RNeasy Lipid Tissue kit (QIAGEN) according to the manufacturer's instructions. To analyze transgene relative to endogenous P0 expression, we used a modified version of the protocol described by Feltri et al. (1999). For this purpose, we exploited the A→T point mutation responsible for the Ile106Leu amino acid exchange. In brief, 200 ng total RNA was reverse transcribed using Moloney Murine Leukemia Virus Reverse Transcriptase (Promega) and random hexanucleotide primers (Amersham Biosciences). Equal volumes of the reverse-transcribed product from sciatic nerves of transgenic and wild-type animals were amplified in the presence of α[32P]dATP, using a single primer pair recognizing P0 exon 2 (5′-GTCCAGTGAATGGGTCTCAG-3′) and exon 4 (5′-GCTCCCAACACCACCCCATA-3′) that flanks a DdeI site present in the P0sub transgenes only. PCR conditions were: initial enzyme activation at 95°C for 15 min; followed by cycles consisting of 94°C for 30 s, 63°C for 60 s, and 72°C for 60 s; and a final extension step at 72°C for 10 min, in a standard PCR reaction mix containing HotStarTaq DNA polymerase (QIAGEN). To avoid the formation of heteroduplexes between the PCR products containing the additional DdeI site and those lacking this site, only cycles in the logarithmic range were chosen. Unincorporated nucleotides were removed from the RT-PCR products using Micro Bio-Spin P-30 columns (Bio-Rad Laboratories) according to the manufacturer's instructions. 4 μl of the purified RT-PCR products was digested with DdeI (New England Biolabs) at 37°C for 90 min. DNA fragments were resolved by PAGE and visualized both by phosphorimaging and by autoradiography. Intensity of the bands was quantified by densitometry of phosphorimager signals (Fujix BAS 2000; Fuji) using TINA 2.09 software (Raytest), and the ratio between the transgene-specific 228-bp fragment and the endogene-specific 245-bp fragment was calculated. To correct for the difference in incorporated dATPs, this intensity ratio was multiplied by a correction factor (1.063), yielding the transgenic P0 message/endogenous P0 message ratio.
Preparation of sciatic nerve homogenates and immunoblot analysis
The sciatic nerves of 2-mo-old wild-type and transgenic mice were homogenized in lysis buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% NP-40, and 1× Complete® protease inhibitor cocktail [Roche]) and incubated at 4°C for 30 min. The homogenates were cleared by centrifugation three times at 12,000 g at 4°C for 20 min. The protein concentrations of homogenates were determined using the BCA Protein Assay kit (Pierce Chemical Co.). The samples were denatured in 5× sample buffer (10% glycerol, 5% β-mercaptoethanol, 5% SDS, and 0.1% bromophenol blue) at 95°C for 5 min and subjected to SDS-PAGE and immunoblot analysis using monoclonal (P07; 1:1,000 diluted; a gift of Dr. Juan Archelos; Archelos et al., 1993) or polyclonal (1:750 diluted) P0 antibodies. Primary antibodies were detected with HRP-conjugated anti–rat or anti–rabbit IgG (Dianova), and were visualized with ECL chemiluminescent substrate (Amersham Biosciences). For the quantification of films (BioMax; Kodak), images were captured by high resolution (600 × 600 dpi) 8-bit (256 gray level) microdensitometry with a flat-bed scanner (Arcus II; Agfa). The images were analyzed for optical density of bands using image analysis software (GelWorks 1D; Ultra-Violet Products). To determine the ratio between P0 expression in transgenic animals and P0 expression in wild-type animals, intensities of the main P0-immunoreactive band (25 kD) were measured in lanes loaded with different amounts of total protein (see Fig. 3) and were normalized to the total protein amount.
8-wk-old mice (5 wild-type and 15 transgenic animals for each mouse line) were anaesthetized using pentobarbital (Narcoren®) and subsequently perfused with PBS for 3–5 min in order to remove blood from peripheral nerves. Sciatic nerves, femoral nerves, and dorsal roots were prepared and frozen in liquid nitrogen. For homogenization, the nerves were mixed with ice-cold homogenization buffer (5 mM Tris-HCl, pH 7.4, 1 mM NaHCO3, and 320 mM sucrose) and homogenized in a Potter homogenizer. The homogenate was centrifuged at 100 g and 4°C for 10 min. After this centrifugation step, the pellet and the supernatant were collected. The pellet was resuspended in homogenization buffer, and supernatant and homogenized pellet were applied on top of the first step sucrose gradient (320 mM – 650 mM – 1 M – 1.2 M). The gradients were centrifuged at 100,000 g and 4°C for 1 h. The interface between the 320-mM and 650-mM sucrose solution was collected, diluted with homogenization buffer, and centrifuged again for 20 min at 100,000 g. The resulting pellets were osmotically shocked and treated with 5 mM Tris-HCl, pH 8.3, for 45 min on ice. Afterwards, the samples were centrifuged at 100,000 g and 4°C for 20 min. The pellets were suspended in 5 mM Tris-HCl, pH 8.3, and applied to the top of a second sucrose step gradient (0 mM – 650 mM – 850 mM – 1 M – 1.2 M). The gradient was centrifuged at 100,000 g and 4°C for 1 h. The myelin fraction was collected at the 0 – 650 mM sucrose interface, washed in ice-cold PBS, and centrifuged for 30 min at 100,000 g. The resulting pellet was suspended in PBS and frozen at −80°C. Protein content of the samples was determined and immunoblot analyses were performed as described (see above).
Preservation of tissue for light and electron microscopy
Femoral and sciatic nerves were processed for light and electron microscopy as reported earlier (Carenini et al., 2001). The mice were transcardially perfused using 4% PFA and 2% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4). The nerves stayed in the same fixative overnight, followed by osmification and embedding in Spurr's medium.
For light microscopic analysis, 0.5-μm-thick semi-thin sections from femoral nerves were stained with alkaline methylene blue and were investigated with a light microscope (Axiophot; Carl Zeiss MicroImaging, Inc.) using a 40× objective. For EM, ultrathin sections of 70-nm thickness were counterstained with lead citrate and investigated using an electron microscope (model EM 10B; Carl Zeiss MicroImaging, Inc.). Primary magnification was between 2,500 and 25,000.
The quantitative analysis of pathological changes was performed on ultrathin sections using a BioVision slow scan camera attached to the Zeiss EM 10B microscope and using the corresponding software analySIS 3.0 Doku (Soft Imaging Systems). The following morphological parameters were assessed: axons with myelin of normal thickness, axons with thin myelin, axons with folded myelin/tomacula, and myelin-competent axons (larger than 1 μm in diameter) devoid of myelin. For the determination of relative myelin thickness, the g-ratio (a reciprocal measure of myelin thickness; Friede, 1972), was determined as recently described (Kobsar et al., 2003).
Comparison of the pathological features of the different genotypes was done by use of a Mann-Whitney-U test and statistical significance was defined for P < 0.05. Statistical analysis of the data was performed by use of Excel (Microsoft) and SYSTAT (SPSS, Inc.). Graphs were made using SigmaPlot 2001 (SPSS, Inc.).
Preparation of single nerve fibers was performed according to Martini et al. (1995b). In brief, sciatic nerves of transcardially perfused mice (see earlier in the Materials and methods) were removed and connective tissue around the nerves was stripped off, followed by gentle “pre-teasing” of fiber bundles. After osmification of fiber bundles and dehydration with acetone, single-fiber preparation was performed in nonpolymerized Spurr's medium using no. 5 watchmaker's forceps. Single fibers were transferred into a droplet of Spurr's medium on a slide, followed by cover-slipping and polymerization at 60°C. Light microscopy was performed with an Axiophot light microscope (Carl Zeiss MicroImaging, Inc.) using a 40× objective.
Nerve conduction experiments of sciatic nerves from 4- and 6-mo-old mice were performed by established electrophysiological methods as described previously (Zielasek et al., 1996). In brief, after anesthesia the compound muscle action potential was recorded with two needle electrodes in the foot muscles after distal stimulation of the tibial nerve, one main branch of the sciatic nerve at the ankle, and proximal stimulation of the sciatic nerve at the sciatic notch. In all experiments, the investigator was not aware of the genotype. Statistical analysis was performed using a one-tailed t test for grouped data.
The authors are grateful to Dr. Michael Bösl for microinjection and transplantation of zygotes, to Dr. Marius Ader and Dr. Astrid Rollenhagen for help with preparation of nerves, to Dr. Laura Feltri for providing us with P0 vector constructs and for helpful suggestions and discussions, to Dr. Uwe Borgmeyer for help with the RT-PCR experiments and helpful comments on the manuscript, and to Dr. Juan Archelos for providing the P07 antibody. We thank Eva Kronberg for animal care and Heinrich Blazyca for excellent technical assistance (tissue processing for light and electron microscopy).
This work is supported by the Deutsche Forschungsgemeinschaft (SFB 581, Priority Program “Microglia” MA1053/3, to R. Martini) and by the Gemeinnützige Hertie-Stiftung (to M. Schachner and R. Martini).
A.E. Rünker and I. Kobsar contributed equally to this paper.
R. Martini and M. Schachner contributed equally to this paper.
A.E. Rünker's present address is Developmental Neurobiology, Department of Genetics, Smurfit Institute, Trinity College Dublin, Dublin 2, Ireland.
Abbreviations used in this paper: CMT, Charcot-Marie-Tooth disease; DSS, Déjérine-Sottas syndrome; P0, peripheral myelin protein zero.