Role of progesterone in peripheral nerve repair

Reviews of Reproduction (2000) 5, 189 199 Role of progesterone in peripheral nerve repair H. L. Koenig* W. H. Gong* and P. Pelissier Laboratoire de Neurobiologie du Développement, Université Bordeaux 1, 33405 Talence, France Progesterone is synthesized in the peripheral nervous system in glial cells. The functions of progesterone are indicated by the findings that it stimulates neurite outgrowth from dorsal root ganglia sensory neurones in explant cultures, accelerates the maturation of the regenerating axons in cryolesioned sciatic nerve, and enhances the remyelination of regenerated nerve fibres. The formation of myelin sheaths around axons is a sexually dimorphic process, as the sheaths are thicker in female than in male regenerating nerves. The progesterone-induced myelination is probably mediated by progesterone receptors, as it is impaired by mifepristone (RU486), a progesterone antagonist. The stimulation of neurite growth in the peripheral nervous system may be mediated by a progesterone metabolite, 5α-tetrahydroprogesterone, through GABA A receptors. Progesterone is conventionally thought of as a female gonadal steroid hormone synthesized in the ovary. Its classical target tissues are the uterus, mammary glands and brain. There is now abundant evidence that progesterone is more than just a female sex hormone. It is also a neurosteroid synthesized in the central and peripheral nervous systems, where its nonreproductive functions are only beginning to be understood (Koenig et al., 1995; Schumacher et al., 1996; Baulieu, 1997; Jung- Testas et al., 1999). This review summarizes the evidence that progesterone has neurotrophic roles in the peripheral nervous system (PNS), that it activates the growth and maturation of axons and stimulates the repair and replacement of myelin sheaths in regenerating nerve fibres. Progesterone, as has been suggested for androgens, may be a potential therapeutic agent in enhancing the reparative response of neurones to injury (Jones, 1993). The possibility that steroid hormones are also involved in nerve regeneration was first indicated by the finding that inhibition of the synthesis of cholesterol, the precursor of steroid hormones, inhibits neurite outgrowth in retinal explants (Heacock et al., 1984), delays myelination in the sciatic nerve of postnatal mice (Rawlins and Uzman, 1970) and that the treatment of crushed rabbit nerves with steroid hormones accelerates muscle re-innervation (Vita et al., 1983). This review reports studies of previously unrecognized neurotrophic roles of progesterone in the peripheral nervous system: (1) in activating the growth of axons and their maturation, and (2) in stimulating the repair and replacement of the myelin sheaths in regenerating nerve fibres. Progesterone in the peripheral nervous system Classically, two experimental systems are used in the study of the synthesis or function of molecules such as progesterone in *Present address: Laboratoire de Génétique Moléculaire de la Neurotransmission et des Processus Neurodégénératifs, CNRS UMR C9923, Bât. CERVI, Hôpital de la Pitié-Salpétrière, 75013 Paris, France. 2000 Journals of Reproduction and Fertility 1359-6004/2000 the PNS. First, in vitro, dorsal root (DRG) and sympathetic ganglia explants or dissociated cells, incubated in the presence of trophic factors, are used to assay neurite-promoting activity. Second, in vivo, the influence of neurotrophins and neurotrophic factors, including progesterone, is analysed mainly in the lesioned peripheral nerves of adult animals After a nerve is wounded (cut, crushed or locally frozen), the distal part of the axons degenerate rapidly (Fig. 1). Simultaneously, the myelin sheaths break down within the Schwann cells, which are then phagocytosed by the macrophages that invade the degenerating nerve. New Schwann cells migrate to the lesioned zone and multiply rapidly inside the basal lamina tubes. The basal lamina tubes are maintained after freezing, as they are after crushing of the nerve but are not maintained after its section (Mira, 1979, 1988). In the distal stump, the unfrozen Schwann cells divide actively while degenerating axons and myelin sheaths are removed by macrophages. These events are referred to as Wallerian degeneration (Fig. 1). Axonal sprouts, which result from the branching of the proximal unsevered axons, appear rapidly in the frozen zone, and are enclosed in the basal lamina tubes at the external surface of the degenerating Schwann cells. The axonal sprouts surrounded by new Schwann cells constitute the regenerating axonal bundles. Next, the diameter of one or two sprouts become larger (approximately 2 µm). These sprouts segregate from the bundles and form separate single axons, surrounded by Schwann cells in the usual 1:1 relationship. During this process, myelination is initiated, either inside the bundles or in the separated single axons (Fig. 1). This sequence of events, which mimics the developmental process, may be called axonal maturation (Scherer and Salzer, 1996). The first evidence that progesterone is synthesized in the PNS comes from the finding that pregnenolone, the precursor of progesterone, is present in large quantities in the sciatic nerves in rats (Akwa et al., 1993) and humans (Morfin et al., 1992). In mice, both pregnenolone and progesterone accumulate in the peripheral nerve (Koenig et al., 1995). In vitro, [ 3 H]pregnenolone is converted to [ 3 H]progesterone in Schwann cells harvested from rat embryonic dorsal root ganglia

190 H. L. Koenig et al. (a) (b) Axon Motoneurone body Schwann cell body (c) (c) Myelin Cryode Nucleus Node of Ranvier Internode Invading SC Macrophage Unmyelinated SC Promyelin fibre Muscle fibre Basal lamina Axonal sprouts Remyelinating SCs Reinnervation of muscle fibre Motor endplate Growth cone Fig. 1. The processes involved in the regeneration of peripheral nerve fibres after damage by local freezing (cryolesion). (a) An unlesioned nerve. The myelinated Schwann cells are separated by nodes of Ranvier, but the basal lamina is continuous up to the motor end plate. Schwann cells are responsible for myelin formation. (b) A few days after the nerve is damaged, the Schwann cells, myelin sheaths and axons degenerate distally to the site of injury. Macrophages invade the lesioned area and the distal part of the nerve to phagocytose the degenerating debris. (c) New Schwann cells appear in the cryolesioned zone, and multiply in this zone and in the distal stump of the nerve. The proximal unfrozen regenerating axons end with growth cones and axonal sprouts issued from these axons, which grow inside the basal lamina tubes. (d) After approximately 1 week, the regenerated axons are re-ensheathed and remyelinated by Schwann cells in the cryolesioned nerve. The target muscle cells are also reinnervated. explants (Koenig et al., 1995). In DRG cultures, both Schwann cells (satellite cells) and neurones contain 3β-hydroxysteroid dehydrogenase (3β-HSD), which converts pregnenolone to progesterone (Guennoun et al., 1997). Neurones of DRG incubated in [ 3 H]pregnenolone produce significant amounts of [ 3 H]progesterone and [ 3 H]5α-dihydroprogesterone. In similar conditions, Schwann cells isolated from DRG produce radioactive progesterone, 5α-dihydroprogesterone and 3α,5αtetrahydroprogesterone (Guennoun et al., 1997). Rat sciatic nerve fragments were shown to contain 5α-reductase, the enzyme that converts progesterone to 5α-dihydroprogesterone (Celotti et al., 1992). Effect of progesterone on axonal growth and maturation in the peripheral nervous system Studies in vitro The addition of progesterone to cultured DRG from 16-dayold rat embryos results in the growth of a halo of neurites surrounding the explants and increased myelination (Fig. 2). Both the surface area and the apparent density of the neurites increase (compare Fig. 2a with b). The surface area of the growing neurite increases by 60 130% after 3 days of progesterone treatment (Fig. 2e). Progesterone may stimulate neurite growth through progesterone receptors present in DRG (Jung-Testas et al., 1999). Low amounts of progesterone receptor have been detected in rat peripheral nerves (Jung- Testas et al., 1996) but not in mouse sciatic nerves (I. Jung- Testas and H. Koenig, unpublished). The addition of RU486 to control cultures of DRG containing no progesterone did not decrease the growth of neurites (W. H. Gong and H. Koenig, unpublished). Therefore, it is uncertain whether the endogenous progesterone produced within DRG plays a role in neurite outgrowth. The addition of 5α-dihydroprogesterone and 3α,5αtetrahydroprogesterone to cultured DRG also increased neurite outgrowth (W. H. Gong and H. Koenig, unpublished). Since 3α,5α-tetrahydroprogesterone does not bind to nuclear progesterone receptors, but interacts with gamma-aminobutyric acid (GABA) A receptors (Celotti et al., 1992; Melcangi et al., 1999), it is possible that 3α,5α-tetrahydroprogesterone stimulates neurite outgrowth via this receptor subtype. This view is supported by the finding that the stimulation of neurite outgrowth from DRG explants by 3α,5α-tetrahydroprogesterone, 5α-dihydroprogesterone and progesterone is inhibited by the GABA A antagonists: bicuculline and picrotoxin (W. H. Gong and H. Koenig, unpublished). Furthermore, progesteroneinduced neurite outgrowth from DRG explants is blocked by finasteride, an inhibitor of 5α-reductase. Therefore, it appears that neurite outgrowth is stimulated by 5α-dihydroprogesterone or 3α,5α-tetrahydroprogesterone, or both, rather than by progesterone (W. H. Gong and H. Koenig, unpublished). Sensory neurones in DRG are heterogeneous cells with different cell body sizes, functions and receptors. Small cells mediate pain and temperature, express trk A receptors and are sensitive to nerve growth factor (NGF). Medium-size cells mediate pressure sensations, express trk B receptors and respond to brain-derived neurotrophic factor (BDNF) and neurotrophin 4 5. Large cells mediate proprioception, express trk C and respond to neurotrophin 3. The question arises if all or only subpopulations of DRG neurones are progesteroneand 3α,5α-tetrahydroprogesterone-sensitive and repond to steroids.

Role of progesterone in peripheral nerve repair 191 Fig. 2. Effects of progesterone on neurite growth and myelination in explanted dorsal root ganglion (DRG) sensory neurones. (a b) Neurite outgrowth in embryonic rat DRG explants taken at day 16 of embryo development. After a 2 day treatment with cytosine arabinoside (Ara-c), progesterone was added to the control culture medium, which permanently contained 10 ng NGF ml 1 to allow survival and a minimal outgrowth of neurites. In control medium, 24 h after Ara-C treatment, the neurite outgrowth is poor (a); after addition of 100 nmol progesterone l 1 for 24 h, the neurites occupy a significantly larger surface area which grows further (b). (c d) Myelination in rat DRG explants taken at day 17 of embryo development. Myelin was obtained according to the method of Eldridge et al. (1987), modified by DoThi (1992), as described in Koenig et al. (1995). Myelin was stained with Sudan black. After 4 weeks in Sato s defined medium, the ganglia were cultured in a myelin-promoting medium containing ascorbic acid for 2 more weeks. In the presence of 20 nmol progesterone l 1 (d) the total length and density (mm 2 ) of the myelinated fibres were increased approximately sixfold compared with control dishes (c). Several nodes of Ranvier are visible in the cultures (arrow). G: dorsal root ganglion explant. (e) Kinetic of neurite outgrowth (expressed in mm 2 ) in control and progesterone-treated (100 nm) DRG explants. Both types of cultures contained 10 ng NGF ml 1 for 5 days (2 days Ara-C treatment plus 3 days of experiment). (f) Density of myelin (expressed in µm mm 2 ) measured 15 days after progesterone (20 nm) was added to the myelinating culture medium. **P < 0.001 (ANOVA Student s t test). Studies in vivo Studies in vivo on the effects of progesterone on axonal maturation have been carried out in the sciatic nerve from the mutant Trembler mouse (Koenig et al., 1991). The Trembler has an autosomal dominant mutation of the myelin protein PMP22 (Suter et al., 1992). Mutant mice develop tremor, quadriparesis and transient seizures during early development (Falconer,

192 H. L. Koenig et al. Fig. 3. Effect of progesterone on the maturation of regenerating axons in the hypomyelinated peripheral nerve of a Trembler male mutant mouse. Electron micrographs from a series of ultrathin transverse sections through the sciatic nerve. The sciatic nerves were cryolesioned by repeated freezing (6 8 cycles of freezing thawing) with a 2 mm copper cryode previously frozen in liquid nitrogen. Progesterone (100 mg kg 1 ), dissolved in sesame oil, was immediately applied onto the nerves. After 2 4 further applications, the nerves were dissected out and fixed for electron microscopy after 7, 12 and 20 days, respectively. (a) Control sham-operated nerve, 7 days after cryolesion and application of the vehicle (sesame oil). Most regenerated axons (ax) and the sprouts (s) they give rise to are grouped in dark Schwann cells. This structure forms the axonal bundles (thick arrows), which are surrounded by multilayered basal laminae, which characterize the Trembler peripheral nerve fibre (thin arrows). There were no myelinated axons in the nerves examined. (b) Progesterone-treated nerve, 7 days after cryolesion. The number of single axons in a 1:1 relationship with Schwann cells increases as a consequence of the retraction of the sprouts. Some axons (ma) are already surrounded by a thin myelin sheath (arrowhead). (c) Progesterone-treated nerve, 12 days after cryolesion. The number of axonal bundles is markedly reduced, whereas myelination is increased. (d e) Sciatic nerves, 20 days after cryolesion: (d) control (non-treated) nerve; (e) progesterone-treated nerve. The mean thickness of Trembler regenerated myelin sheaths is approximately 15% lower in control nerves than in progesterone-treated nerves.

Role of progesterone in peripheral nerve repair 193 (a) 60 (b) 4.0 Clusters of regenerated fibres 50 40 30 20 10 *** ** Axons (number per cluster) 3.5 3.0 2.5 2.0 1.5 1.0 0.5 ** ++ 0 7 12 Days after cryolesioning 0 7 12 Days after cryolesioning (c) Myelinated fibres (%) 30 25 20 15 10 5 *** * (d) 10 8 6 4 2 Mean number of lamellae ++ * 0 0 7 12 20 7 12 20 Unoperated Days after cryolesioning Days after cryolesioning control Fig. 4. Effect of progesterone applications to cryolesioned sciatic nerve on axonal regeneration and myelination in Trembler mutant mice. The sciatic nerve was exposed and a portion subjected to local freezing. Progesterone and control vehicle (sesame oil) were then applied several times for 7, 12 and 20 days. Measurements were made directly on the electron microscope screen or on the electron micrographs. The Trembler mice used in these experiments were heterozygous adult mice (homozygous mutants do not survive in our strain). (a) Number of axonal bundles present after 7 and 12 days after cryolesion in progesterone-treated ( ; n = 8) and sesame oil-treated (control; ; n = 5) regenerating nerves. (b) Number of axonal sprouts per axon bundle. (c) Number of myelinated fibres. (d) Number of myelin lamellae. *** P < 0.001; ** P < 0.01; * P < 0.05; ++ P > 0.01 (ANOVA Student s t test). (a d) The results were obtained in the same groups of animals (7, 12 and 20 day old mice. 1951). The mutation produces severe hypomyelination and demyelination of the peripheral nerve fibres. The number of Schwann cells in peripheral nerve fibres is 8 10 times higher than in normal mice. Each Schwann cell is surrounded by multilayered basal laminae (Ayers and Anderson, 1973; Koenig et al., 1991; Do Thi et al., 1993; Fig. 3a d). The Trembler mouse has two features that make it useful for experimentation. First, 20% of the sciatic nerve axons only are thinly myelinated in cross-sections of the nerve. These axons degenerate rapidly, within 24 48 h in the areas subjected to cryolesioning (Koenig et al., 1991). Thus, the regenerating axonal sprouts and subsequent myelination are easy to see. The rate of axonal regeneration after cryolesion of

194 H. L. Koenig et al. Progesterone Progesterone Schwann cells nuclear receptor of progesterone Neurotrophic factors PR DHP,THP Neurone soma THP Neurone GABA A -R Schwann cell 5α-DHP 3α,5α-THP Signal to axon (membrane or axoplasm) SC GABA A -R Axon GABA A -R Retrograde signal to neurone soma Gene activity Axonal growth Axonal maturation Neuronal gene activity Signal to axon Axonal growth Hyper polarization (axon membrane) Retrograde signal to neurone soma Gene activity Axonal maturation Myelin initiation Myelin wrapping Fig. 5. Hypothetical mechanisms explaining the direct stimulatory effect of progesterone on myelination in the Schwann cells of regenerating peripheral nerve fibres. PR: progesterone receptor; DHP: dihydroprogesterone; THP: tetrahydroprogesterone; GABA A -R: GABA A receptor. the sciatic nerve in Trembler nerves is double that in normal mice (Ferzaz et al., 1989). Second, cryolesioned axons in Trembler mice have robust regenerative capabilities (Ferzaz et al., 1989). Axonal sprouts first appear 6 12 h after cryolesion and remain enclosed in the Schwann cell sheath, surrounded by the basal lamina tube, for more than 12 days. Local application of exogenous progesterone to the cryolesioned sciatic nerve of the Trembler mouse accelerates the process of regeneration. The number of clusters of regenerating fibres (Fig. 4a) and the number of axons per cluster decreases in the bundles of treated nerves (Fig. 4b). Twelve days after cryolesion, the axonal bundles and the number of their sprouts are still lower in progesterone-treated nerves than in the untreated control nerves (Fig. 4a,b). As a consequence of the loss of redundant sprouts and bundles, the number of individualized single axons doubles in 7 days in progesterone-treated nerves (Fig. 3c). It is not known whether the progesterone produced in the Trembler sciatic nerve is involved in axonal growth and maturation, since the concentrations of progesterone and its precursor, pregnenolone, are very low (Koenig et al., 1995). Axonal maturation promoted by exogenous progesterone is probably the result of an acceleration of the natural regenerative processes. It is not known whether 5α-dihydroprogesterone and 3α,5α-tetrahydroprogesterone accelerate axonal maturation in vivo. The mechanism by which axonal growth is stimulated by progesterone or its 5α-derivatives probably involves several neuronal genes expressed in neurofilaments, microtubules and in the axolemma. The expression of these genes may be regulated either at transcription or at translation level by progesterone or its 5α-metabolites. The progestagen target cells may be neurones or Schwann cells. In regenerating nerves, Schwann cells synthesize and secrete several neurotrophic factors that contribute to axonal growth (Terenghi, 1999). Progestagens may enhance the synthesis of neurotrophins or other factors that induce an axonal signal for growth. The signal molecule may act either locally on the axonal membrane or the growth cones or be transported in a retrograde direction to the neurone cell bodies to stimulate the genes of proteins associated with and necessary to the growth of the axons (Fig. 5).

Role of progesterone in peripheral nerve repair 195 Fig. 6. Specificity of the action of progesterone on myelination in lesioned peripheral nerves of wild type mice. The nerves were processed for ultrastructural analysis 15 days after cryolesion (as described in Fig. 3 legend) after a series of four local applications of either the vehicle (sesame oil), progesterone or progesterone blocker. (a) Male control nerve treated with the vehicle only. (b) Male nerve after trilostane treatment (a progesterone synthesis inhibitor): the myelin sheaths are thinner than in controls. (c) Reversal of the effect of trilostane on myelin sheaths by simultaneous administration of progesterone in male nerve, indicating that the decreased myelination is not due to the toxicity of trilostane. (d) Nerve from a female ovariectomized mouse 10 days before cryolesion, treated with progesterone after the lesion. The myelin sheaths are thicker than in male mice treated with progesterone. (e) Quantification of the effects of pregnenolone, progesterone and inhibitors (mifepristone, RU486; trilostane; and trilostane plus progesterone) on thickness of myelin sheaths in male nerves, classified in increments of five lamellae. The inset shows the average width of the myelin sheaths (that is, the number of lamellae). The data are obtained from the same animals as used for data in Fig. 4, and are expressed as percentages of control (baseline). From 300 to 500 myelin sheaths were counted per nerve. Progesterone-treated nerves were surrounded by thicker myelin sheaths than nerves from vehicle-treated mice. The action of progesterone on myelination was inhibited by drugs that inhibit the action of progesterone (modified from Koenig et al., 1995). During the process of peripheral nerve regeneration, the thickness of the myelin sheaths is not related to the diameter of the axons. This absence of relationship is maintained for life in Trembler hypomyelinated mice, both in regenerated and non-operated nerves. Effects of progesterone on myelination of injured axons in the peripheral nervous system Myelin originates from the Schwann cell surface membrane, which forms a mesaxon that elongates and wraps spirally around the axon. Further growth leads to a compact sheath composed of concentrically arranged lamellae. The stimulatory effect of progesterone on the remyelination (1) of regenerating sciatic nerves in wild-type male mice (Fig. 6 a c) and (2) of outgrown neurites in DRG explants in culture (Fig. 2c,d,f) was first demonstrated by Koenig et al. (1995). Progesterone increases the rate of myelin synthesis and accelerates the initiation of myelin formation in DRG neurone Schwann cell cultures (Chan et al., 1998). The application of progesterone to cryolesioned sciatic nerves in mice stimulates thickening of the myelin sheaths (Figs 6 and 7). Treatment with pregnenolone produces similar results (Koenig et al., 1995). Remyelination of cryolesioned sciatic nerves is inhibited by trilostane, which blocks the conversion of pregnenolone to progesterone or mifepristone (RU486), a progestagen receptor blocker, by nearly 25 and 44%, respectively (Figs 6 and 7). The inhibitory effect of trilostane is not due to toxicity, since its effect on remyelination is reversed by the simultaneous application of progesterone (Koenig et al., 1995 and Fig. 6c). At this stage of the remyelinating process, the nerves recover their progesterone concentrations. In vitro, RU486 does not influence myelin formation in DRG cultures, but inhibits the stimulatory effect of the exogenous progesterone (Jung-Testas et al., 1999). Both exogenous pregnenolone and progesterone increase the neoformation of myelin sheaths in lesioned peripheral nerves. A crucial question is: which is the active molecule? When the conversion of cholesterol to pregnenolone is blocked by aminoglutethimide, the formation of myelin decreases (H. Koenig and B. Ferzaz, unpublished). Progesterone and its 5α-reduced metabolites, dihydroprogesterone and tetrahydroprogesterone, stimulate the expression of genes encoding peripheral myelin proteins Po

196 H. L. Koenig et al. (a) Mean number of myelin lamellae (b) 500 Treatment Control Progesterone Male Intact mice 12.9 ± 0.1 15.7 ± 0.1 Castrated mice 15.7 ± 0.5 17.4 ± 0.2 Female Intact mice 15.8 ± 0.4 19.4 ± 1.0 Castrated mice 18.8 ± 0.1 21.4 ± 0.8 Myelin sheaths thickness (%) (as compared to the controls = 100) 400 300 200 100 0 100 200 RU486 TRIL PROG Castrated female 3 5 6 10 11 15 16 20 21 25 26 30 31 35 >35 Number of lamellae Fig. 7. Sexual dimorphism in the effects of progesterone on peripheral nerve remyelination. (a) Thickness of sciatic nerve myelin sheaths (that is, number of lamellae) from intact and gonadectomized male and female mice, processed simultaneously for electron microscopy after cryolesion and local treatment with the vehicle (control) or progesterone. (b) Thickness of sciatic nerve myelin sheaths from ovariectomized females treated with progesterone, an inhibitor of progesterone synthesis (trilostane) or a progesterone receptor blocker (mifepristone, RU486). Six to ten mice were analysed 15 days after they received four local applications of progesterone or progesterone blockers. and PMP22; 3α,5α-tetrahydroprogesterone is more potent in this respect than 5α-dihydroprogesterone and progesterone (Melcangi et al., 1998). Therefore, the progesterone synthesized in Schwann cells or the exogenous progesterone synthesized in the sciatic nerve may serve as the parent molecule for several potentially active steroids and steroid derivatives. Exogenous progesterone also enhances the concentrations of Po and PMP22 proteins in co-cultures of DRG neurones and Schwann cells (Notterpeck et al., 1999) and stimulates the expression of their gene promoters in cultured rat Schwann cells (Desarnaud et al., 1998). These observations imply that progesterone binds to progesterone receptors in the Schwann cell cytoplasm or nucleus, although further investigation is required since progesterone receptor was not detected in mouse sciatic nerves (I. Jung-Testas and H. Koenig, unpublished). Furthermore, in progesterone receptor knock-out adult mice, the myelin sheaths around sciatic nerve axons are the same as in wild-type mice (Jung-Testas et al., 1999). Endogenous progesterone may activate the process of myelination through autocrine paracrine actions involving a limited number of progesterone receptors (Koenig et al., 1995; Baulieu et al., 1996; Baulieu and Schumacher, 1997). An alternative possibility is that the effect of progesterone on myelination is mediated through GABA A receptors (Melcangi et al., 1999). Finally, progesterone may increase the transfer of axonally orthograde transported phospholipids from axons to myelin (Droz et al., 1981; Toews et al., 1988). Such a transfer would increase the rate of spiral wrapping of the Schwann cell membrane, which constitutes the myelin lamellae. This interpretation is sustained by two observations. First, remyelination begins 1 week or more after nerve damage, when regenerating axons have already matured. Second, in nerves treated for 7 days during the second week after cryolesion, the thickness of myelin sheaths is similar to that in nerves treated with progesterone for 15 days immediately after the lesion (P. Pelissier and H. Koenig, unpublished). These observations indicate that progesterone is less, or not, effective in stimulating myelination in axons while they are regenerating. It remains to be determined whether progesterone acts directly on Schwann cells or if its myelin-stimulating effect is mediated through an axonal signal. An axonal signal is known to be necessary for the initiation of the myelination process (for review, see Salzer, 1995). The hypothetical mechanisms of the stimulatory effect of progesterone on remyelination in lesioned nerves are shown (Fig. 5). Progesterone in sexual dimorphism in peripheral nerve myelination Sexual dimorphism of reproductive steroid hormones is generally associated with sexual dimorphism in behaviour and reproductive functions. Therefore, it is relevant to ask whether the production of progesterone in peripheral nerves is also sexually dimorphic. In rodents, the progesterone content of the peripheral nerve is higher in females than that found in males. In female mice, the concentrations of endogenous progesterone and pregnenolone differ among groups of nerves, but are higher than in male sciatic nerves (M. Schumacher and H. Koenig, unpublished). In female rats, the concentration of progesterone in the sciatic nerves is higher in pro-oestrus ( 10) than it is in oestrus ( 5), as compared with that in males (Pelissier et al., 1996; M. Schumacher, unpublished). These observations prompt several questions. Does the higher progesterone content in female nerves result in a faster rate of formation of myelin in damaged nerve fibres in females than in males? Does the administration of progesterone increase myelination in lesioned nerves in males but not in females? As progesterone

Role of progesterone in peripheral nerve repair 197 Progesterone Neurotrophic factors (NTF) DHP THP Po, MBP, PMP22 PR NTF-R Po, PMP2... GABA A -R Hyperpolarization Myelin wrapping Gene activity Myelin wrapping Gene activity Spinal cord motoneurone or dorsal root ganglion neurone Axonal signal Axonal signal for myelination Axonal transport of phospholipids Increased transfer Axonal signal for myelination Target cell Increase in myelin wrapping Increase in myelin wrapping Po, PMP22... Po, PMP22... Schwann cell Schwann cell Fig. 8. Possible mechanisms explaining the effects of progesterone on growth and maturation of the regenerating axons and consecutive myelination in lesioned peripheral nerve. In vitro, progesterone may also directly stimulate neuronal soma, in addition to Schwann cells as it does in in vivo experiments. PR: progesterone receptor; DHP: 5α-dihydroprogesterone; MBP: myelin basic protein; NTF-R: neurotrophic factor receptor; PMP22: peripheral myelin protein 22; Po: protein Po; THP: 3α,5α-tetrahydroprogesterone; GABA A -R: GABA A receptor. is mainly synthesized in the gonads, does castration influence the myelination process of damaged peripheral nerve fibres in a different way in males and in females? In intact female mice, the mean number of myelin lamellae formed around regenerating nerve fibres after cryolesion is significantly higher (> 22%) than in regenerating nerves from intact males (Fig. 7a). A similar difference is observed in ovariectomized females (> 20%) as compared with castrated males (Fig. 7a). In both genders, progesterone treatment enhanced myelin width by approximately 22% and, in gonadectomized animals, by > 11% in males and > 14% in females. These findings indicate that the presence of gonadal hormones in the peripheral nerve and in the peripheral circulation stimulates myelination in damaged nerves. In intact and castrated progesteronetreated animals of both sexes, the increase in mean thickness of myelin is due to the increased number of thick myelin sheaths made up of more than 20 lamellae. This increase of thick regenerated myelin sheaths was identical in intact males and females (> 125%), whereas it was different in gonadectomized males (> 60%) and females (> 38%). However, in both intact and ovariectomized females, the number of thick myelin sheaths is higher than in males (intact and castrated). The faster rate of remyelination in females is probably associated with the high content of progesterone in female nerves (Pelissier et al., 1996) and the higher concentrations of residual progesterone remaining after cryolesion (M. Schumacher and H. Koenig, unpublished). A similar difference is described in the rat hypoglossal nerve (Yu, 1982) and in the male hamster facial nerve, in which the axonal regeneration is 20% slower than in females (Kujawa et al., 1991). The question of the role of 5α-reduced metabolites of progesterone in remyelination of damaged peripheral nerves remains to be analysed. In these nerves, activities of 5α-reductase (an enzyme involved in the conversion of progesterone to 5αdihydroprogesterone) are high (Celotti et al., 1992). Furthermore, small quantities of 3α,5α-tetrahydroprogesterone increase the myelin protein Po expression in rat Schwann cell cultures (Melcangi et al., 1998). Initiation of myelination by progesterone in Trembler mutant mice In Trembler mice, the treatment of regenerating nerves with progesterone results in the appearance of myelin lamellae at

198 H. L. Koenig et al. day 7, after axonal maturation is nearly completed (Figs 3b and 4c). In Trembler cryolesioned untreated (control) nerves, myelin is formed after day 10. Thus, it is likely that the accelerated formation of the myelin sheaths in nerves of Trembler mice induced by progesterone is a consequence of the faster maturation of the lesioned axons. At later time intervals (after 12 and 20 days), the increased number of myelinated fibres (Fig. 4c) and the increased myelin width (Fig. 4d) in progesterone-treated nerves may be the result of an increased rate of myelin synthesis or an increase in the rate at which lamellae wrap around the axons. Concluding remarks and future perspectives The dual function of progesterone and its metabolites in the regeneration of PNS nerves is well established. One function of progestagens is to stimulate axonal growth, which entails, sequentially, the maturation of the axons, their covering by Schwann cells, the initiation of myelin formation from mesaxons, and the elongation and spiral wrapping of Schwann cell membranes inside the cytoplasm. A second function of progestagens is to accelerate the myelinating process directly. Several intracellular pathways may be involved in the effects of progesterone or its 5α-metabolites on axonal growth and maturation and on the remyelination processes (Fig. 8). In fact, nothing is known of the mechanisms involved in progestagen-dependent PNS nerve regeneration. The mechanisms must involve progestagen-dependent functions in both the damaged neurones and their associated Schwann cells, possibly including changes in the gene transcription or protein translation of proteins required for repair of the whole injured system. Progesterone, or its derivatives, may act either directly on the cells responsible for axonal regeneration or contribute to the release of growth or myelin regulatory molecules at the site of peripheral nerve injury. For both axonal growth and myelination, two questions arise: does progesterone act through its receptor or through the GABA A receptors (Figs 5 and 8); and what are the subsequent signalling pathways? In addition to a study of myelin proteins and steroid enzymes stimulated during nerve regeneration, research is needed on the effect of progesterone on the enzymes involved in lipid synthesis in both axon and Schwann cells. Axon membranes and myelin contain high concentrations of lipids, the synthesis of which may be stimulated by progesterone. Cholesterol transport mechanisms are probably also important, since apolipoprotein E is involved in axonal growth (Handelmann et al., 1992; Nathan et al., 1994; Mahley et al., 1996) and apolipoprotein A I gene expression is increased in myelinating nerves (LeBlanc et al., 1989). Other cells, such as macrophages and fibroblasts, also secrete trophic substances, which might contribute to PNS nerve regeneration, but to a lesser extent than steroids. If motoneurones and their axons are shown to be sensitive to progesterone or to its 5α-reduced metabolites, better understanding of the molecular control of steroid modulation on axonal growth and myelination is likely to lead to the development of therapies for PNS diseases and spinal cord diseases. This review is dedicated to Professor René Couteaux, for his 90th birthday. The authors thank N. A Do Thi for providing new photographs of a common work (Koenig et al., 1995) and for expertise in DRG cultures; B. Ferzaz for initiating the in vivo regeneration model in our laboratory; M. Schumacher for contributing to the castration of the mice; Prof. E. E. Baulieu for the first fellowship to W. H. Gong (Société de Secours des Amis des Sciences) and J. Prémont for his advice on the GABA experiments. This work was supported by the Université Bordeaux I and the Association Francaise contre les Myopathies, who also provided grants to P. Pélissier and W. H. Gong. At present, W. H. Gong is the recipient of a post-doctoral fellowship from Institut pour la Recherche sur la Moelle Epinière. The authors also thank J. Mallet for allowing H. Koenig and W. H. Gong to pursue their research in his laboratory. H. Koenig is deeply grateful to Prof. R. Couteaux for his permanent help and support. His enthusiasm and creativity in discussions were an exemple highly appreciated and source of inspiration. References Key references are indicated by asterisks. Akwa Y, Schumacher M, Jung-Testas I and Baulieu EE (1993) Neurosteroids in rat sciatic nerves and Schwann cells Comptes-rendus de l Académie des Sciences Série III Paris 316 410 414 Ayers MM and Anderson RMcD (1973) Development of onion bulb neuropathy in the Trembler mouse. Comparison with normal nerve maturation Acta Neuropathologica 32 43 59 Baulieu EE (1997) Neurosteroids of the nervous system, by the nervous system, for the nervous system. In Recent Progress in Hormone Research Vol. 52 pp 1 32 Ed. PM Conn. Endocrine Society Press, Bethesda, MD Baulieu EE and Schumacher M (1997) Neurosteroids, with special reference to the effect of progesterone on myelination in peripheral nerves Multiple Sclerosis 3 105 112 Baulieu EE, Schumacher M, Koenig HL, Jung-Testas I and Akwa Y (1996) Progesterone as a neurosteroid actions within the nervous system Cellular and Molecular Neurobiology 16 143 154 Celotti F, Melcangi RC and Martini L (1992) The 5alpha-reductase in the brain: molecular aspects and relation to brain function Frontiers in Neuroendocrinology 13 163 215 Chan JR, Phillips LJ and Glaser M (1998) Glucocorticoids and progestins signal the initiation and enhance the rate of myelin formation Proceedings National Academy of Sciences USA 95 10 459 10 464 Desarnaud F, Do Thi NA, Brown AM, Lemke G, Suter U, Baulieu EE and Schumacher M (1998) Progesterone stimulates the activity of the promoters of peripheral myelin protein-22 and protein zero genes in Schwann cells Journal of Neurochemistry 71 1765 1768 Do Thi NA (1992) Ontogenese des anomalies phénotypiques des cellules de Schwann et ses conséquences chez la souris dysmyélinique Trembler Etude in vivo et in vitro. Thèse de Doctorat de l Université Bordeaux I Do Thi NA, Koenig HL, Vigny M, Fournier M and Ressouches A (1993) In vivo proliferative pattern of Trembler hypomyelinating Schwann cells is modified in culture: an experimental analysis Developmental Neuroscience 15 10 21 *Droz B, Giamberardino JD and Koenig HL (1981) Contribution of axonal transport to the renewal of myelin phospholipids in peripheral nerves I. Quantitative radioautographic study Brain Research 219 57 71 Eldridge CF, Bartlett M, Bunge RP and Wood PM (1987) Differentiation of axon-related Schwann cell in vitro I. Ascorbic acid regulates basal lamina assembly and myelin formation Journal of Cell Biology 105 1023 1034 Falconer DS (1951) Two new mutants, Trembler and Reeler, with neurological actions in the house mouse (Mus musculus L) Journal of Genetics 50 192 201 Ferzaz B, Koenig HL and Ressouches A (1989) Axonal regeneration in Trembler mouse, a Schwann cell mutant Comptes-rendus de l Académie des Sciences Série III Paris 309 377 382 *Guennoun R, Schumacher M, Robert F, Delespierre B, Guézou M, Eychenne B, Akura Y, Robes P and Baulieu EE (1997) Neurosteroids: Expression of functional 3β-hydroxy-steroid dehydrogenase by rat sensory neurons and Schwann cells European Journal of Neuroscience 9 2236 2247 Handelmann GE, Boyles JK, Weisgraber KH, Mahley RW and Pitas RE (1992) Effect of apolipoprotein E,β-very low density lipoproteins, and cholesterol on the extension of neurites by rabbit DRG neurons in vitro. Journal of Lipid Research 33 1677 1688 Heacock AM, Klinger PD, Seguin EB and Agranoff BW (1984) Cholesterol synthesis and nerve regeneration Journal of Neurochemistry 42 987 993

Role of progesterone in peripheral nerve repair 199 Jones KJ (1993) Gonadal steroids and neuronal regeneration. A therapeutic role Advances in Neurology 59 227 240 Jung-Testas I, Schumacher M, Robel P and Baulieu EE (1996) Demonstration of progesterone receptors in Schwann cells Journal of Steroid Biochemistry and Molecular Biology 58 77 82 *Jung-Testas I, Do Thi A, Koenig H, Desarnaud F, Shazand K, Schumacher M and Baulieu EE (1999) Progesterone as a neurosteroid: synthesis and action in rat glial cells Journal of Steroid Biochemistry and Molecular Biology 69 97 107 *Koenig HL, Do Thi A, Ferzaz B and Ressouches A (1991) Schwann cell proliferation during postnatal development, Wallerian degeneration and axon regeneration in Trembler dysmyelinating mutant. In Plasticity and Regeneration of the Nervous System pp 227 238 Eds PS Timiras et al. Plenum Press, New York Koenig HL, Schumacher M, Ferzaz B, Do Thi NA, Ressouches A, Guennoun R, Jung-Testas I, Robel P, Akwa Y and Baulieu EE (1995) Progesterone synthesis and myelin formation by Schwann cells Science 268 1500 1503 Kujawa KA, Emeric E and Jones KJ (1991) Testosterone differentially regulates the regenerative properties of injured hamster facial motor neurons Journal of Neurology 11 3898 3906 LeBlanc AC, Földvari M, Spencer DF, Breckenridge WC, Fenwick RG, Williams DL and Mezei C (1989) The apolipoprotein A-I gene is actively expressed in the rapidly myelinating avian peripheral nerve Journal of Cell Biology 109 1245 1256 Mahley RW, Nathan BP, Bellosta S and Pitas RE (1996) Apolipoprotein E: Structure, function, and possible roles in modulating neurite extension and cytoskeletal activity. In Apolipoprotein E and Alzheimer s Disease pp 49 58 Eds Roses, Weisgraber and Christen. Fondation Ipsen, Springer Verlag, Berlin Melcangi RC, Magnaghi V, Cavarretta I, Martini L and Piva F (1998) Ageinduced decrease of glycoprotein Po and myelin basic protein gene expression in the rat sciatic nerve. Repair by steroid derivates Neurosience 85 569 578 *Melcangi RC, Magnaghi V, Cavarretta I, Zucchi I, Bovolin P, D Urso D and Martini L (1999) Progesterone derivatives are able to influence peripheral myelin protein 22 and Po gene expression: possible mechanisms of action Journal of Neuroscience Research 56 349 357 Mira JC (1979) Quantitative studies of the regeneration of rat myelinated fibres: variations in the number and size of regenerating nerve fibres after repeated localized freezings Journal of Anatomy 129 77 93 Mira JC (1988) The hand. In The Biology of Regeneration in Peripheral Nerves pp 383 404 Ed. R Tubiana. WB Saunders, Philadelphia Morfin R, Young J, Corpechot C, Egestad B, Sjövall J and Baulieu EE (1992) Neurosteroids: pregnelonone in human sciatic nerves Proceedings National Academy of Sciences USA 89 6790 6793 Nathan BP, Bellosta S, Sanan DA, Weisgraber KH, Mahley RU and Pitas RE (1994) Differential effects of apolipoproteins E3 and E4 on neuronal growth in vitro. Science 264 850 852 Notterpek L, Snipes GJ and Shooter EM (1999) Temporal expression pattern of peripheral myelin protein 22 during in vivo and in vitro myelination Glia 25 358 369 Pelissier P, Robert F, Corpechot C, Ressouches A, Baulieu EE, Schumacher M and Koenig HL (1996) Sexual dimorphism in progesterone-induced myelination of regenerating peripheral nerves 2nd European Meeting on Glial Cell Function in Health and Disease B36 p 70 Rawlins FA and Uzman BG (1970) Retardation of peripheral nerve myelination in mice treated with inhibitors of cholesterol biosynthesis A quantitative electron microscopic study Journal of Cell Biology 46 505 517 Salzer JL (1995) Mechanisms of adhesion between axons and glia cells. In The Axon pp 164 184 Eds S Waxman, J Kocsis and P Stys. Oxford University Press, New York Scherer SS and Salzer JL (1996) Axon Schwann cell interactions during peripheral nerve degeneration and regeneration. In Glial cell development: basic principles and clinical relevance pp 165 185 Eds KR Jessen and WD Richardson. Bio. Scientific Publications, Oxford *Schumacher M, Robel P and Baulieu EE (1996) Development and regeneration of the nervous system: a role for neurosteroids Developmental Neuroscience 18 6 21 Suter U, Welcher AA, Ozceik T, Snipes GJ, Kosaras B, Francke U, Billings- Gagliardi S, Sidman RL and Shooter EM (1992) Trembler mouse carries a point mutation in a myelin gene Nature 356 241 244 Terenghi G (1999) Peripheral nerve regeneration and neurotrophic factors Journal of Anatomy 194 1 14 Toews AD, Armstrong R, Ray R, Gould RM and Morell P (1988) Deposition and transfer of axonally transported phospholipids in rat sciatic nerve Journal of Neuroscience 8 593 601 Vita G, Dattola R, Girlanda P, Oteri G, Lo Presti F and Messina C (1983) Effects of steroid hormones on muscle reinnervation after nerve crush in rabbit Experimental Neurology 80 279 287 Yu WH (1982) Sex difference in the regeneration of the hypoglossal nerve in rats Brain Research 238 404 406