Antiprogesterone therapy uncouples axonal loss from demyelination in a transgenic rat model of CMT1A neuropathy

Authors

  • Gerd Meyer zu Horste MD,

    1. Department of Neurogenetics, Max-Planck-Institute of Experimental Medicine, University of Göttingen, Göttingen, Germany
    Search for more papers by this author
    • G.M.z.H. and T.P. contributed equally to this article.

  • Thomas Prukop,

    1. Department of Neurogenetics, Max-Planck-Institute of Experimental Medicine, University of Göttingen, Göttingen, Germany
    Search for more papers by this author
    • G.M.z.H. and T.P. contributed equally to this article.

  • David Liebetanz MD,

    1. Department of Clinical Neurophysiology, University of Göttingen, Göttingen, Germany
    Search for more papers by this author
  • Wiebke Mobius PhD,

    1. Department of Neurogenetics, Max-Planck-Institute of Experimental Medicine, University of Göttingen, Göttingen, Germany
    Search for more papers by this author
  • Klaus-Armin Nave PhD,

    1. Department of Neurogenetics, Max-Planck-Institute of Experimental Medicine, University of Göttingen, Göttingen, Germany
    Search for more papers by this author
  • Michael W. Sereda MD

    Corresponding author
    1. Department of Neurogenetics, Max-Planck-Institute of Experimental Medicine, University of Göttingen, Göttingen, Germany
    2. Center of Neurology, University of Göttingen, Göttingen, Germany
    • Max-Planck-Institute of Experimental Medicine, Department of Neurogenetics, Hermann-Rein-Strasse 3, D-37075 Göttingen, Germany
    Search for more papers by this author

Abstract

Objective

Charcot–Marie–Tooth disease (CMT) is the most common inherited neuropathy, and a duplication of the Pmp22 gene causes the most frequent subform CMT1A. Using a transgenic rat model of CMT1A, we tested the hypothesis that long-term treatment with anti-progesterone (Onapristone) reduces Pmp22 overexpression and improves CMT disease phenotype of older animals, thereby extending a previous proof-of-concept observation in a more clinically relevant setting.

Methods

We applied placebo-controlled progesterone-antagonist therapy to CMT rats for 5 months and performed grip-strength analysis to assess the motor phenotype. Quantitative Pmp22 RT-PCR and complete histological analysis of peripheral nerves and skin biopsies were performed.

Results

Anti-progesterone therapy significantly increased muscle strength and muscle mass of CMT rats and reduced the performance difference to wildtype rats by about 50%. Physical improvements can be explained by the prevention of axon loss. Surprisingly, the effects of anti-progesterone were not reflected by improved myelin sheath thickness. Electrophysiology confirmed unaltered NCV, but less reduced CMAP recordings in the treatment group. Moreover, the reduction of Pmp22 mRNA, as quantified in cutaneous nerves, correlated with the clinical phenotype at later stages.

Interpretation

Progesterone-antagonist treatment. Pmp22 overexpression to a degree at which the axonal support function of Schwann cells is better maintained than myelination. This suggests that axonal loss in CMT1A is not caused by demyelination, but rather by a Schwann cell defect that has been partially uncoupled by anti-progesterone treatment. Pmp22 expression analysis in skin may provide a prognostic marker for disease severity and for monitoring future clinical trials. Ann Neurol 2007;61:61–72

Charcot–Marie–Tooth disease (CMT) is the most common inherited neuropathy. It causes slowly progressive, distally pronounced, symmetric muscle atrophy in patients with age of onset at young adolescence.1 Originally, CMT was classified using clinical and electrophysiological criteria into demyelinating (type 1), axonal (type 2), and other less frequent forms.1

More recently, mutations in numerous different genes were found to cause CMT; therefore, the classification was extended by genetic means (reviewed in Niemann and colleagues2). In more than 50% of all cases, the genetic defect underlying CMT is an intrachromosomal duplication spanning 1.5Mb on human chromosome 17p12, the locus defining CMT subtype 1A (CMT1A).3, 4 The gene encoding the peripheral myelin protein of 22kDa (Pmp22) is located in the duplicated region and was identified as the responsible disease gene.5–8 Animal models transgenically overexpressing Pmp22 have, indeed, provided formal proof that mere overexpression of Pmp22 in Schwann cells suffices to cause CMT1A9–11 (reviewed in Sereda and Nave12). Pmp22 messenger RNA (mRNA) overexpression leads to dysmyelination and demyelination, which is thought to cause axonal loss and neurogenic muscle atrophy. However, the mechanisms underlying myelin-dependent axonal defects are not well understood.13, 14

The CMT rat overexpresses Pmp22 and mimics human CMT1A symptoms of peripheral demyelination, axonal loss, and neurogenic muscle atrophy particularly well.11 In this model, proof of principle was provided15 that the nuclear progesterone receptor is a potential pharmacological target in CMT1A. Progesterone is a known regulator of myelin gene expression, and progesterone administration to Schwann cells increases Pmp22 expression.15–17 In the previous proof-of-principle study, the progesterone receptor antagonist onapristone18 was given to early postnatal CMT rats, and it successfully ameliorated the neuropathic phenotype.15 However, it was unknown whether this therapeutic approach would be effective when initiated in older rats and when extended over months. We therefore undertook a new therapeutic study with antiprogesterone in CMT rats that better reflects the clinical situation in CMT1A patients. We can demonstrate, by quantifying axonal loss and muscle atrophy after 5 months of therapy, that antiprogesterone effects are stable. Surprisingly, axonal loss and demyelination have been pharmacologically uncoupled, demonstrating that axonal support is a Schwann cell function that is independent of myelin itself.

Finally, CMT1A patients display a marked interindividual variability of disease severity, even in monozygotic twins,19 that has been similarly documented for CMT rats.15 In these studies, we found that the reduction in Pmp22 mRNA can be quantified in cutaneous nerves not only as a diagnostic,20 but also as a prognostic marker. This suggests that Pmp22 expression in small skin nerve fibers can be used to monitor future human clinical trials.

Materials and Methods

Transgenic Rats

The generation of Pmp22 transgenic CMT rats has been described previously.11 Routine genotyping was performed by polymerase chain reaction (PCR), using genomic DNA from tail biopsies and mouse transgene-specific primers under standard conditions as described previously.11 All experiments were performed according to the German regulations of Lower Saxony for animal experimentation.

Steroid Treatment

Randomly chosen female transgenic rats were treated with either onapristone (n = 16) (ZK 98299 from Schering, Berlin, Germany) or placebo (n = 16). Each compound was incorporated into time-release pellets (Innovative Research of America, Sarasota, FL) at 180mg/pellet for the first and 192mg/pellet for the second implantation. Two pellets per animal were subcutaneously implanted at 5 and 18 weeks of age, respectively, resulting in an average onapristone dosage of 20mg/kg/day. Wild-type littermates (n = 5) and transgenic control rats (n = 16) received placebo pellet treatment for comparison. For implantation, rats were anesthetized with a mixture of ketamine (100mg/kg) and xylazine (20mg/kg). Five of initial 45 animals died during narcosis, and 3 animals were excluded from further analyses because of pellet rejection. Thus, 16 CMT rats remained in the onapristone and placebo group, respectively, and 5 wild types received comparable placebo treatment.

Phenotyping

All phenotype analyses were performed by the same investigator who was blinded to genotype and treatment arm. Motor performance of CMT rats was assessed in standardized grip strength tests for forelimbs (Fig 1B, top) and hind limbs (see Fig 1B, bottom) separately, as described previously.21 With their forelimbs, the animals gripped a horizontal T-bar (width 14cm, diameter 3.2mm) connected to a gauge whereas the investigator pulled their proximal tail from the bar with increasing force (see Fig 1B, top). Hind-limb grip strength was measured by supporting the forelimbs and pulling the animal's tail toward the horizontal bar (see Fig 1B, bottom). In both test variants, the maximum force (measured in Newtons) exerted onto the T-bar before the animals lost grip was recorded. The circumference (measured in millimeters) of the forelimb muscles was determined, in a blinded fashion, with a thin nylon string at the median level between elbow and wrist after removing skin and adipose tissue.

Figure 1.

Long-term onapristone treatment continuously improves the Charcot–Marie–Tooth disease (CMT) phenotype, as demonstrated by grip-strength analysis and forelimb muscle circumference measurements. (A) Our previous proof-of-principle study had used onapristone in young postnatal CMT rats (age: postnatal day 5 until 7 weeks).15 To better mimic the clinical situation, this study focussed on older CMT rats (age, 5 weeks) and was extended over 5 months. The progesterone antagonist and placebo were continuously released from subcutaneously implanted pellets. (B) Regular grip-strength analyses of forelimbs (top) and hind limbs (bottom) were performed every 2 and (after 1 month) every 3 weeks. Onapristone-treated CMT rats (n = 16) (green) displayed greater forelimb (D) and hind-limb (F) grip strength compared with placebo-treated control rats (n = 16) (red). Differences between treatment groups increased with age and became significant at 12 and 18 weeks of age for hind limbs and forelimbs, respectively. Onapristone-treated CMT rats remained stronger until the study end, which is illustrated in magnifications of the two final time points of forelimbs (E) and hind limbs (G). Wild-type values (n = 5) are depicted for comparison (blue). Forelimb muscle circumference was significantly larger in onapristone-treated CMT rats than in placebo-treated control rats (C). *p < 0.05, onapristone compared with placebo. ns = not significant; SEM = standard error of the mean.

Electrophysiology

Nerve conduction velocities (NCVs) and compound muscle action potentials (CMAPs) were determined by an independent examiner who was blinded for the genotype and treatment groups, as described previously.22 Rats were anesthetized with a mixture of ketamine (100mg/kg) and xylazine (20 mg/kg) and immediately placed into an oil bath (37°C) to stabilize temperature for the 15-minute procedure. Tail NCV proved to be more reproducible than recordings from the sciatic nerve, which had to be exposed surgically.11 CMAP recordings from tail muscles to single electric stimuli of 0.1-millisecond duration to the tail nerves were recorded with fine subcutaneous needle electrodes, using a Jaeger-Toennies Neuroscreen (Würzburg, Germany) instrument. NCVs were calculated from the latency difference between the CMAPs after successive proximal stimulation at two sites 20mm apart. CMAP amplitudes were calculated peak to peak.

Tissue Preparation

Before treatment (age, 5 weeks) and after 4 weeks of onapristone treatment (age, 9 weeks), a skin biopsy from the tail was collected from each animal after short CO2 narcosis for mRNA analysis. After 21 weeks of treatment (age, 26 weeks), all rats were killed by CO2 narcosis. For mRNA quantification, sciatic nerves and one tail skin biopsy were dissected from each animal. For histological analysis, the animals were perfused with Hank's buffered salt solution, followed by a modified fixation with 2.5% glutaraldehyde and 4% paraformaldehyde in phosphate buffer.11, 23

Histology of Peripheral Nerve and Skin Biopsies

After perfusion, sciatic nerves, tibial nerves, and skin biopsies from tail were dissected and embedded in Epoxy resin.24 Semithin sections (0.5μm) of sciatic nerves were obtained 5mm distal to the greater trochanter before any branching of the nerve had occurred.25 Semithin sections of tibial nerves were cut at level with the medial ankle. Sections were stained with Azur II-methylene blue and photographed using a standard video frame grabber (ProgRes C14; Jenoptic, Jena, Germany) installed at a Zeiss Axiophot microscope (Zeiss, Oberkochen, Germany). Overlapping photographs of the entire nerves were taken and merged using PanoToolsAssembler software (Tawbaware, www.tawbaware.com.). After blinding, the entire number of total, myelinated, and unphysiologically unmyelinated axons per sciatic nerve and per tibial nerve were counted manually by the same investigator using the CellCounter plug-in of ImageJ (v1.36, National Institutes of Health, Bethesda, MD). Each individual axon was manually marked and automatically counted. Physiologically unmyelinated axons (diameter <1μm) and Remak-bundle fibers were not included. ImageJ was also used to determine myelin thickness and axonal diameter by outlining and measuring the axonal and myelin sheath perimeter and area. The g-ratio was calculated by dividing the axonal circumference by the circumference of the respective myelin sheath. More than 100 randomly chosen fibers per animal were analyzed. The axonal area of 200 axons/animal was measured, and the axonal diameter was calculated.

Ultrathin sections of skin biopsies were cut using a Leica Ultracut S ultramicrotome (Leica, Vienna, Austria) and stained with an aqueous solution of 4% uranyl acetate followed by lead citrate.26 The sections were viewed in a LEO EM 912AB electron microscope (Zeiss), and pictures were taken with an on-axis 2048 × 2048 charge-coupled device camera (Proscan, Schering, Germany).

Gene Expression Analysis

Total RNA from tissue samples was extracted using the RNeasy Lipid Tissue Mini Kit according to the manufacturer's protocol (Quiagen, Hilden, Germany). The integrity of purified RNA was confirmed using the Agilent 2100 Bioanalyser (Agilent Technologies, Böblingen, Germany). For quantitative reverse transcriptase PCR (RT-PCR) analysis, cDNA (complementary DNA) was synthesized from total RNA using random nonamer primers and Superscript III reverse transcriptase (Invitrogen, Karlsruhe, Germany).

Quantitative RT-PCR reactions (Taqman PCR assay) were conducted using the ABI Prism 7700 Sequence Detection System (Applied Biosystems, Darmstadt, Germany). A TaqMan PCR master mix was prepared to a final reaction volume of 20μl. The TaqMan PCR followed the standard two-step protocol, and quantitation of PCR product was performed using the comparative ΔΔCt method as recommended by the manufacturer. All RT-PCRs were conducted in duplicate and averaged. All TaqMan primers and probe sets were designed using the primer express software V.1.65 (Applied Biosystems). Pmp22 TaqMan primers were designed to coamplify endogenous (rat) Pmp22 cDNA and transgene-derived (mouse) Pmp22 cDNA. For quantification of Pmp22 expression, total Pmp22 cDNA amplification was normalized against a specific cDNA fragment containing the nonsteroid regulated and ubiquitously expressed exon 1B of the Pmp22 gene.

Wild-type Pmp22 expression was defined as 1.0-fold relative expression, and Pmp22 expression in placebo-treated rats was defined as 1.6-fold. The sequences of PCR primers and fluorescent probe sequences (TaqMan) are as follows: total Pmp22 forward primer: 5′-TGT ACC ACA TCC GCC TTG G-3′; total Pmp22 reverse primer: 5 ′-GAG CTG GCA GAA GAA CAG GAA C-3′; total Pmp22 TaqMan probe: 5′-CCA TGA TCC TGT CTG TCA TCT TCA GCG TC-3′; Pmp22 exon 1B forward primer: 5′-GCT GTC CCT TTG AAC TGA AA-3′; Pmp22 exon 1B reverse primer: 5′-GAA CAG GAT CCC CAA CAA GAG TAG-3′; Pmp22 exon 1B TaqMan probe: 5′-AGC CCA ACT CCC AGC CAC CAT G-3′.

Statistical Analysis

All values are expressed as mean ± standard error of the mean. We have tested all our data on Gaussian distribution and where applicable have used parametric (Student's t test for unrelated groups) or nonparametric (Wilcoxon–Mann–Whitney U test) testing. Correlation analyses were performed using the Spearman's rank correlation test. Statistica 6.0 (StatSoft, Hamburg, Germany) was used for statistical analyses.

Results

Antiprogesterone Reduces Progressive Muscle Atrophy in Charcot–Marie–Tooth Disease Rats

To better mimic the clinical situation of CMT1A in the rat model, we initiated the antiprogesterone trial in a cohort of 16 CMT rats at the age of 5 weeks. Onapristone (kindly provided by Schering, Berlin, Germany) was continuously given for a period of 5 months, when rats reached adulthood (schematically depicted in Fig 1A). No signs of toxicity were noted. We used custom-made drug pellets that continuously released a defined dosage of either onapristone (average daily dosage, 20mg/kg) or placebo. These were subcutaneously implanted at the age of 5 weeks, and again at 18 weeks of age. Whereas the initial proof-of-principle study used male rats, this long-term study investigated female rats.

The grip-strength analysis (see Fig 1B) of the hind limbs differentiated between wild-type and (placebo) treated transgenic rats from baseline (age, 5 weeks) throughout the whole study (see Fig 1F). Forelimb grip-strength analysis differentiated between wild-type rats and transgenic control rats only until the age of 12 weeks (see Fig 1D). To monitor treatment effects, we performed phenotyping every 2 weeks, and after 1 month every 3 weeks (see Fig 1A). As expected, after treatment began (age, 5 weeks), the onapristone group (n = 16) showed no obvious difference in grip strength compared with placebo-treated CMT rats (n = 16). After 7 weeks of treatment (age, 12 weeks), however, hind limb grip strength was significantly greater (p < 0.05, Student's t test) in the onapristone-treated rats (see Fig 1F). At the age of 18 weeks also the forelimb grip strength was significantly greater than that of the placebo-treated group (p < 0.05, Student's t test) (see Fig 1D). This beneficial treatment effect is more obvious when the two final time points are directly compared, as depicted in Figures 1E and 1G. When compared with wild-type rats, the defects of muscle strength that define CMT rats as measured by hind-limb grip-strength analysis were reduced to about 50% by the antiprogesterone treatment. Forelimb muscle circumference served as an alternative measure of muscle atrophy and was found to be significantly (p < 0.05, Student's t test) greater in onapristone compared with placebo-treated CMT rats (see Fig 1C). This confirmed preservation of muscle mass by onapristone treatment.

Electrophysiological Examinations Demonstrate Unaltered Nerve Conduction Velocity But Ameliorated Compound Muscle Action Potential after Antiprogesterone Treatment

NCV and CMAPs were determined at the tail motor nerve at the end of the study (age, 26 weeks) (Fig 2A). Compared with placebo control CMT rats (n = 5; NCV, 14.78 ± 0.39m/sec), onapristone-treated animals (n = 5; NCV 15.74 ± 0.65m/sec) showed unaltered NCV, whereas wild-type control animals reached reference values (n = 4; NCV, 44.7 ± 1.97m/sec) (p < 0.05 Wilcoxon–Mann–Whitney U test) (see Fig 2B). After progesterone antagonist therapy, the CMAPs of the tail muscle were significantly higher in treated animals (2.48 ± 0.53mV) than placebo control animals (0.88 ± 0.24mV) (p < 0.05, Wilcoxon–Mann–Whitney U test), although wild-type values were not reached (6.48 ± 1.23mV) (see Fig 2C).

Figure 2.

Long-term onapristone treatment attenuates loss of compound muscle action potential (CMAP) amplitudes but does not influence nerve conduction velocities (NCVs). Representative NCV and CMAP recordings after proximal and distal stimulation are depicted in (A) as determined for the tail motor nerve at the end of the study (age, 26 weeks). Note the different horizontal and vertical scaling. Compared with placebo control animals, onapristone-treated animals showed unaltered NCV (B). After progesterone antagonist therapy, the CMAP amplitudes were significantly higher in treated animals than placebo control animals, although wild-type values were not reached (C). CMT = Charcot–Marie–Tooth disease; SEM = standard error of the mean.

Onapristone Prevents Loss of Axons without Altering Myelin Sheath Thickness

Histologically, peripheral nerves in CMT rats show a characteristic dysmyelination and demyelination11 (Figs 3C, E; arrow) that is combined with focal hypermyelination of small-caliber axons (see Figs 3C, E; open arrowhead) and progressive axonal loss when compared with wild-type control animals.11 Myelination defects were equally observed in both treatment groups (see Figs 3C, E; arrow). However, large-caliber axons appeared to be more frequent in onapristone-treated than in placebo-treated CMT rats (see Fig 3C, filled arrowhead). We thus performed more detailed morphometric analyses at the end of the trial. First, g-ratios (diameter of axon divided by diameter of myelinated axon) were calculated for individual axon-myelin units in semithin cross sections of the tibial nerve. When these ratios were plotted as a function of the axonal diameter, a hypermyelination (ie, lower g-ratio) of small-caliber axons became obvious that contrasted with hypomyelination (ie, higher g-ratio) of large-caliber axons in CMT rats (see Figs 3D, F), in comparison with age-matched wild-type control animals (see Fig 3B). Importantly, myelin sheath thickness did not differ between CMT rats treated with onapristone (see Fig 3D) and those treated with placebo (see Fig 3F). Thus, the clinical improvements could not be secondary to a higher degree of myelin preservation. Furthermore, no regenerative clusters were seen, ruling out functionally irrelevant axonal sprouting.

Figure 3.

Onapristone treatment does not stably alter the myelin pathology of Charcot–Marie–Tooth disease (CMT) rats. Morphologically, peripheral nerves of CMT rats (C, E) exhibit concurrent hypermyelination (C, E; open arrowhead) and hypomyelination (C, E; arrow). Whereas myelination defects (C, E; arrow) were observed comparably, the frequency of large-caliber axons appeared higher in onapristone (C; filled arrowhead) than in placebo-treated (E) CMT rats. G-ratio measurements in CMT rats showed hypermyelination of smaller and hypomyelination of larger axons (D, F) compared with wild-type rats (B). However, g-ratio and myelin thickness do not differ between onapristone- (D) and placebo-treated (F) CMT rats after 5 months of treatment.

In untreated CMT rats, the axonal size distribution was shifted toward smaller axons when compared with normal rats (Fig 4A). In contrast, CMT rats treated with onapristone exhibited a greater percentage of mid-to-larger caliber myelinated axons (3–4μm) than the placebo group (p < 0.05, Student's t test), when quantified in the tibial nerve, whereas the percentage of small myelinated axons (1–2μm) was correspondingly lower (p < 0.05, Student's t test) (see Fig 4A). The percentage of axons larger than 4μm did not significantly differ between placebo and onapristone groups (see Fig 4A).

Figure 4.

Increased axonal diameters and reduced total axonal loss in long-term onapristone-treated Charcot–Marie–Tooth disease (CMT) rats (age, 26 weeks). (A) Quantification demonstrated a significantly reduced number of large-caliber axons in CMT rats compared with wild-type control animals (n = 5; white bars). In onapristone-treated CMT rats, however, axonal size distribution is shifted toward larger caliber axons. Tibial nerves from onapristone-treated CMT rats (n = 14; light gray bars) contained significantly more mid-to-larger (3–4μm) and less smaller (1–2μm) axons than placebo-treated CMT rats (n = 12; dark gray bars) (A). The number of myelinated and pathologically unmyelinated axons per sciatic (B) and tibial nerve (C) was determined. Physiologically unmyelinated axons with a diameter smaller than 1μm and Remak-bundle fibers were not counted. Sciatic nerves from onapristone-treated CMT rats contained a significantly higher number of axons (+6.5%) than nerves from placebo-treated animals, almost reaching wild-type values (B). Comparably, tibial nerves of onapristone-treated CMT rats contained significantly more axons (+6.3%) (C). ns = not significant; SEM = standard error of the mean.

To further clarify the therapeutic effect of onapristone, we quantified the absolute number of axons in the sciatic and tibial nerves, at a defined level (see Materials and Methods), and determined the axonal loss for CMT rats. After 5 months of onapristone treatment, sciatic nerves of CMT rats contained about 7% more axons (p < 0.05, Student's t test) than those of placebo-treated CMT rats (see Fig 4B). In contrast, there was no significant difference in axon number between onapristone-treated CMT rats and wild-type control rats (see Fig 4B). Also, in cross sections of the distal tibial nerve, onapristone-treated CMT rats contained about 6% more axons (p < 0.001, Student's t test) than placebo-treated rats (see Fig 4C).

Taken together, the long-term treatment with antiprogesterone was sufficient to significantly reduce progressive axonal loss in CMT rats, but did not alter the visible myelination pathology. We conclude that axonal support and myelin assembly are distinct physiological functions of Schwann cells, both of which fail in CMT rats and have been partially uncoupled in mutant Schwann cells by antiprogesterone treatment. Therefore, because the drug family of selective progesterone receptor antagonists provides a therapeutic benefit and appears to be well tolerated, clinical trials should be considered.

Quantitation of Pmp22 in Cutaneous Nerves

Any translation of these findings into clinical applications requires a quantitative monitoring of Pmp22 expression, despite significant interindividual differences in clinical phenotype19 and Pmp22 mRNA levels among CMT1A patients.27 We sought to test small skin biopsies, with embedded cutaneous nerves, as a source of Schwann cells for histopathology and quantitative gene expression analysis. In the rat, bundles of myelinated axons could be readily identified in the subcutaneous layer of a skin biopsy, both in wild-type (Fig 5A) and CMT rats (see Fig 5B). Single axons and smaller groups of myelinated axons, passing into superficial cutaneous layers (not shown), resemble findings in humans.20 Cutaneous nerves from CMT rats exhibited pathological hallmarks of CMT1A, including hypomyelination and hypermyelination (see Figs 5C, D). These features clearly distinguished CMT from wild-type nerves. By quantitative RT-PCR, transgenic rats at the age of 9 weeks overexpressed Pmp22 mRNA in cutaneous nerves (p < 0.05, Student's t test) at levels comparable with sciatic nerves11 (see Fig 5E). When quantified before treatment with antiprogesterone (age, 5 weeks), Pmp22 mRNA levels in skin were equal in both treatment arms (see Fig 5F). Four weeks later, onapristone treatment resulted in significantly reduced Pmp22 mRNA overexpression (approximately −20%; p < 0.01, Student's t test) in cutaneous nerves of CMT rats (see Fig 5G). Interestingly, the cutaneous Pmp22 mRNA level, as determined by RT-PCR in young rats (age, 9 weeks), showed a significant correlation (r = −0.42; r2 = 0.176; Spearman's p = 0.0027) with the clinical performance of the same rat, determined by grip-strength measurements at 26 weeks of age (see Fig 5H).

Figure 5.

Cutaneous nerves exhibit Charcot–Marie–Tooth disease (CMT) hallmarks and allow to quantify the therapeutic effect of onapristone on Pmp22 expression, a possible prognostic marker in CMT subtype 1A (CMT1A). Myelinated axons can be easily identified in the dermal and subcutaneous layers of skin biopsies taken from 26-week-old wild-type (A) and CMT rats (B–D). These small nerve fibers exhibit typical pathological features of CMT1A (B) including hypomyelination (C) and parallel hypermyelination of small-caliber axons (D). Transgenic Pmp22 overexpression, when measured by quantitative reverse transcriptase polymerase chain reaction in cutaneous nerves of 9-week-old rats (E), is comparable with that in sciatic nerves.11 Quantifying Pmp22 expression in cutaneous nerves enables monitoring of the treatment effect. Onapristone reduced Pmp22 overexpression in cutaneous nerves after 4 weeks of treatment (age, 9 weeks) (G), whereas both groups started at the same expression levels before treatment began (F). The amount of Pmp22 overexpression in cutaneous nerves at 9 weeks of age correlated with the hind-limb grip strength at 26 weeks of age (H). Thus, CMT rats with higher cutaneous Pmp22 expression at young age were predisposed to a more severe course of disease. mRNA = messenger RNA; SEM = standard error of the mean; ax = axon; m = myelin; nu = nucleus; sc = Schwann cell.

Discussion

In a previous proof-of-principle study,15 we have shown that onapristone, a selective inhibitor of the nuclear progesterone receptor,18 downregulates Pmp22 overexpression in a Pmp22 transgenic CMT rat model. The primary objective of this study was to investigate whether onapristone is equally effective under clinically more realistic conditions. These conditions included: (1) a treatment initiated in older CMT rats; (2) a long-term treatment over 5 months into adulthood; (3) the involvement of female animals; (4) a safe long-term administration of drugs (avoiding subcutaneous injections every 24 hours); and (5) the improved clinical phenotyping of CMT rats, by recording grip strength, muscle mass, and compound muscle action potentials as clinically relevant measures of neurogenic muscle atrophy. After the optimization of all these parameters (not shown), we performed a blinded two-arm study with onapristone and placebo. Using self-release pellets, we began treatment of sixteen 5-week-old female CMT rats, an age when animals already displayed a distinct neuropathic phenotype resembling young adults.11 Treatment was continued for 5 months, extending into adulthood (age, 6 months).

We delivered onapristone by implanting (twice) custom-made time-release pellets that have been previously used in rats to administer antiprogestins28 and progesterone.29 Because of the animals' increasing body weight, the calculated daily onapristone dosage decreased during the course of this study but remained sufficiently high to have a therapeutic effect. In a control experiment, we verified, using an electrochemiluminescence assay, that implanting progesterone-releasing pellets led to the expected elevation of progesterone levels in serum (data not shown).

The clinical phenotype of onapristone-treated CMT rats showed clear differences from placebo-treated control rats. Although the onapristone group started out with a slightly lower grip strength (due to unstratified randomization before treatment beginning), grip strength in this group surpassed the placebo group and became significantly greater after 7 weeks of treatment for hind limbs (see Fig 1F) and after 13 weeks of treatment for forelimbs (see Fig 1D). Thus, onapristone ameliorates CMT markers in transgenic rats (ie, axonal numbers and limb-strength assays) even when treatment begins after disease onset, suggesting the promise of antiprogestins, in principle, for a long-term treatment of CMT1A. In this study, we noticed significant therapeutic effects only after 7 weeks of treatment, that is, 2 weeks later than in the pilot study.15 The most likely explanation is the older age of rats at treatment onset (5 weeks vs 5 days postnatally). Differences between wild-type and placebo-treated CMT rats became more obvious in hind-limb than in forelimb grip-strength measurements. This may resemble that muscle atrophy is more pronounced in the distal parts of especially the lower extremities in CMT1A patients.1

The maximum force a rat exerts onto the T-bar during grip-strength analysis depends on its muscle strength. However, in our observations, the forelimb bar also mechanically stresses the paws and causes more sensitive animals to release early, meaning that, to a small extent, sensitivity to pain also is tested. The onapristone treatment more effectively protects mid-to-larger axons with an axonal diameter of 3 to 4μm, which may partly represent motor axons (although the larger axons >4μm are unchanged), from degeneration than smaller (sensory) ones. Therefore, a high forelimb grip strength may derive from a combination of reduced muscle atrophy (also demonstrated by larger muscle circumference measurements) and increased sensory deficits. Also, the wild-type group (n = 5) is markedly smaller than both treatment groups of CMT rats (n = 16 each), and this test limitation may partly account for the surprising trend toward higher forelimb grip strength in onapristone-treated CMT rats compared with wild-type rats (see Figs 1C, E).

Particularly relevant clinically is that the decreased axonal loss was reflected by ameliorated CMAP measurements after progesterone antagonist therapy, whereas the NCV was unaltered in both treatment groups. This may be important for monitoring future clinical trials. It is generally thought that, in human CMT1A, in analogy to other forms of demyelinating inherited neuropathies, overexpression of Pmp22 in Schwann cells causes dysmyelination, segmental demyelination, and onion bulb formation, which collectively lead to a secondary axonal dysfunction and progressive axon loss. Axon loss, unlike demyelination, is thus the direct cause of neurogenic muscle atrophy in CMT1A. The molecular mechanisms by which myelin defects cause axonal loss are poorly understood, however.2, 12 For the myelinated central nervous system, it is known that mutations in oligodendrocyte-specific genes can lead to widespread loss of myelinated axons, independent of obvious dysmyelination or demyelination.30, 31 Our data suggest that Schwann cells also normally preserve axonal integrity, and that this is a primary glial cell function independent of the amount of myelin present.

We found that the average axon diameter of the tibial nerve is reduced in CMT rats (see Fig 4A), as previously described in biopsies from CMT1A patients32 and other animal models,33 suggesting either a loss of large fibers or abnormal axonal shrinkage. We counted the absolute number of axons larger than 1.0μm and found a significant reduction of axon numbers in the proximal sciatic (see Fig 4B) and the distal tibial nerves (see Fig 4C) of CMT rats when compared with wild-type nerves. Unexpectedly, this axonal loss was not distally pronounced when comparing sciatic with tibial nerves (see Figs 4B, C). This more uniform distribution of muscle weakness in the CMT rat (in contrast to the length-dependent fiber loss in CMT1A patients) may reflect limitations of the rodent animal model.

Also unexpectedly, we found no long-term effect of onapristone treatment on increased myelin sheath thickness (see Figs 3D, F), and the percentage of myelinated axons in both peripheral nerves was the same (see Figs 4B, C). At the same time, the axonal size distributions in the tibial nerve were shifted toward a higher (wild-type–like) pattern after onapristone treatment (see Fig 4A). Furthermore, total axon counts were strikingly higher in onapristone- versus placebo-treated CMT rats (see Figs 4B, C). Thus, onapristone treatment has partially maintained the axonal integrity and reduced axonal losses, particularly of mid- to large-caliber axons (3–4μm). We therefore conclude that the CMT phenotype was not ameliorated by improved myelination. Rather, onapristone reduced the pathological load of Pmp22 overexpression on CMT Schwann cells and their ability to maintain axonal integrity, a function of glia that is independent of myelination. This interpretation is supported by recent case reports of an axonal neuropathy (CMT2) identified as a mutation in the Mpz gene encoding a myelin-specific protein.34, 35 Thus, both central nervous system oligodendrocytes31 and Schwann cells have the primary function to support long-term axonal integrity, in addition to myelination (not by myelination). Indeed, nonmyelinating Schwann cells also are essential for axonal integrity.36 In specific conditions, the dual function of myelinating glia can be uncoupled, either genetically31 or by pharmacological intervention, as shown here.

Nerve biopsies constitute an important but invasive diagnostic tool in patients with peripheral neuropathies. The commonly used sural nerve biopsy of patients cannot be performed repeatedly and causes irreversible sensory loss. Therefore, we explored in the CMT rat model whether the more easily available cutaneous tissue imitates the pathology and abnormal gene expression of large peripheral nerves. In accordance with previous findings,20 myelinated and pathologically demyelinated axons were easily identified in cutaneous biopsies (see Figs 5A–D). Pmp22 overexpression was quantifiable (see Fig 5E), and the level of mRNA was comparable with that of sciatic nerves.11 This indicates that cutaneous nerves can be analyzed instead of other less readily available peripheral nerves.

We used cutaneous biopsies to measure the effect of the onapristone long-term treatment, demonstrating reduced Pmp22 overexpression in treated rats. Pathological signs of hereditary neuropathies in cutaneous nerves had been described previously.20 This study, however, was able to quantify both the CMT phenotype and its therapeutical modulation using cutaneous biopsies. This may be translated to human CMT1A patients who exhibit a marked interindividual variability of disease severity and age at onset.1, 37 Thus, cutaneous nerves could not only serve to quantify Pmp22 gene expression but also to identify individuals prone to a more severe course of disease. Indeed, Pmp22 overexpression determined in rat skin biopsies at early time points correlated with the same animal's phenotype at the end of the study. Perhaps cutaneous Pmp22 expression may provide a prognostic marker in CMT1A patients that can predict the course of disease and identify individuals in urgent need of treatment.

Acknowledgements

This work was supported by the European Union (LSHM-C72004-502987, K.A.N.), the Myelin Project by the European Leukodystrophy Association (ELA, K.A.N.) and the German Research Foundation (DFG, OGM0643, K.A.N.), M.W.S. was supported by the Fonds Anne Catherine del Marmol.

We thank Drs J. Hoffmann, H. Hiemisch, and J. Lampe (Schering, Berlin, Germany) for providing us with onapristone. We thank A. Abdrabou for excellent technical assistance and M. Wehe for excellent animal care. We are grateful to members of the Nave laboratory for discussion.

Ancillary