Raloxifene in Peripheral Nerve Regeneration

A. Specific Aims

            When a peripheral nerve is damaged by injury or disease, the symptoms may range from a tingling feeling to intolerable pain to paralysis.  In many human cases the recovery is slow, and may take as long as 24 months due to lack of efficient therapies (Samardzic et al., 2000).  Therefore, a study of neuroendocrine factors which may potentiate nerve regeneration and rehabilitation is of great clinical importance.  Studies of gender differences in susceptibility to neurological injury have implicated estrogen as one of the main hormones accountable for females’ greater resistance and recovery rates (Roof and Hall, 2000).

Estrogen has a wide range of actions in the nervous system, including neuroprotection and potentiation of nerve regeneration (Toran-Allerand, 1999; Tanzer et al., 1999; McEwen et al., 2001; Islamov et al., 2002).  In spite of the beneficial actions of estrogen on the nervous system, the opportunities for its wide therapeutic application are severely limited because of its adverse side effects in reproductive organs.  Therefore, a search for pharmacological substances with selective estrogenic action on the nervous system is of great practical significance.  One of the new drugs, Evista^® (raloxifene hydrochloride) is a selective estrogen receptor modulator (SERM).  Raloxifene is used to prevent and treat osteoporosis in women post-menopause, and it produces both estrogen-agonistic effects on bone and lipid metabolism and estrogen-antagonistic effects on uterine endometrium and breast tissue.  Recent observations have indicated that raloxifene has estrogenic activity in the nervous system as well (Cyr et al., 2002).  Our recent studies of estrogen and SERMs in peripheral nerve injury showed a 1.5 fold increase in the rate of functional recovery (Islamov et al., 2002; McMurray et al., 2002).  This means that if a patient with conventional treatment may achieve complete recovery in 24 months, treatment with SERMs alone may help this patient to recover 8 months sooner.  Therefore, there is an urgent need to characterize raloxifene regenerative action, which may prove to be the drug of choice for rehabilitation after nerve injuries.  The action of raloxifene via an estrogen receptor system may be multifaceted, affecting regenerating nerves, neuromuscular junction formation, skeletal muscles, and even behavior.  In the current application the effect of raloxifene treatment on rehabilitation after peripheral nerve injury will be analyzed at several levels employing: functional behavioral tests; electrophysiology on neuromuscular preparations and muscles; histomorphology of regenerating nerve fibers, and motor end plates.  The metabolism and phenotype of denervated and reinnervated skeletal muscles will be assessed by biochemical assays and immunohistochemistry.  In addition to studying the effects of raloxifene on recovery after nerve injury, we will also investigate raloxifene action on peripheral nerve neuropathy and regeneration in mice with experimentally induced diabetes.  Peripheral neuropathy is a common complication of diabetes.

The current is focused on initiation of a new study which will characterize the effects of raloxifene on regeneration and functional recovery after sciatic nerve crush and in an experimental model of diabetic neuropathy.  The specific aims for a three year period are:

1.  To examine raloxifene action on functional recovery after sciatic nerve crush injury.  This will be completed using behavioral tests to assess recovery of motor and sensory function and by analysis of raloxifene action on regenerating sciatic nerve, reinnervation of muscle fibers and restoration of function at neuromuscular junction.  This will be accomplished by histomorphological analysis of regenerating sciatic nerve fibers, immunohistochemical analysis of nerve terminal morphology , and physiological studies of synaptic strength at selected time points after injury.  Study of physiological changes will be done by measuring isometric and tetanic twitch tension; and synaptic safety factor on isolated neuromuscular preparations. 

2.  To evaluate whether raloxifene potentiates rehabilitation via a direct action on skeletal muscles.  This will be done by comparison of muscle phenotype and metabolism in animals with complete muscle denervation versus animals with sciatic nerve crush .  This will be completed by immunohistochemical phenotyping of soleus, plantaris and gastrocnemius muscles and by assays for selected muscle enzymes and proteins.

3.  To investigate if raloxifene may be an effective drug against peripheral neuropathy and regenerative deficit in experimental diabetes.  We will analyze sensory deficits and motor function in streptozotocin induced diabetic mice using functional behavioral tests, assessing morphology of nerve fibers, and measuring sciatic nerve conduction velocity.

B. Background and Significance

            Trauma to peripheral nerves is relatively common and often results from blunt or penetrating injuries (Noble et al., 1998).  Clinically, nerve damage results in disruption of a sensory and motor function.  Recovery of function occurs with axonal regeneration and reinnervation of the sensory receptors and muscle end plates and remyelination of the regenerated axons.  Treatment options for many of these disorders include surgical intervention, corticosteroid treatment, and physical and occupational therapy.  In general, peripheral nerves regenerate at the rate of ~1 inch a month in humans.  Therefore, a study of neuroendocrine factors which may potentiate nerve regeneration is of great practical significance.  One of the neuroendocrine factors which have neuroprotective and regenerative action in the nervous system is the sex hormone estrogen.

            Estrogen has a variety of beneficial effects on the nervous system, including modulation of mental and neurodegenerative conditions, such as schizophrenia, Parkinson’s disease and Alzheimer’s disease (for review see Cyr et al., 2000; Dhandapani and Bran , 2002); protection from trauma and stroke (Culmcee et al., 1999; Roof and Hall, 2000); stimulation of neurite growth and synaptogenesis (Toran-Allerand, 1999; Beyer, Karolczak, 2000; McEwen et al., 2001); and potentiation of regeneration (Tanzer et al., 1999; Islamov et al., 2002).  Estrogenic modulation of the nervous system has been implicated in greater resistance of females to neurological diseases and injury and in better recovery rates (for review see Roof and Hall, 2000; Cyr et al., 2002). 

            The diverse effects of estrogen on a variety of tissues including the brain are mediated via intracellular estrogen steroid receptors.  Two estrogen receptors, ERa and ERb, have been characterized to date and there is evidence of another type of estrogen receptor possibly existing (Ramirez et al., 2001; Toran-Allerand et al., 2002).  Estrogen receptors may trigger “genomic”and “nongenomic” course of events (for review see Nilsson et al., 2001).  The classical “genomic” estrogenic action involves estrogen receptor acting as a transcriptional enhancer requiring direct interaction with DNA with subsequent changes in gene expression.  The “nongenomic” mechanism includes interactions of the estrogen receptor system with different intracellular signaling pathways (Toran-Allerand et al., 1999; 2002; Kousteni et al., 2001; McEwen et al., 2001).  Some of the factors and pathways implicated in estrogen signaling include neurotrophins (Singh et al., 1999), insulin-like growth factor-1 (IGF-1) (Azcoitia et al., 1999), cAMP/protein kinase A (Beyer, Karolczak, 2000), NMDA receptors (McEwen et al., 2001), a family of stress- and mitogen- activated protein kinases (MAP), including ERK (Singh et al., 2000; Toran-Allerand, 2002) and p38 kinase (Zhang, Shapiro, 2000;  Lee and Bai, 2002), and PI3-kinase/Akt signaling cascade (Ivanova et al., 2002).  

            In spite of the numerous positive effects of estrogen on the nervous system the opportunity for its clinical applications are severely limited because of its proliferative action in mammary and endometrial tissues, which increases the risk of cancer.  Therefore, a search for pharmacological compounds with estrogenic action limited to the nervous system is of great practical importance.  Several drugs have been developed with mixed estrogen agonist/antagonist properties.  Tamoxifen is a first-generation selective estrogen receptor modulator (SERM), which acts as an estrogen antagonist in mammary tissue but mimics the effect of estrogen in other tissues (Cyr et al., 2002).  Tamoxifen, however, increases the risk of two types of cancer that can develop in the uterus: endometrial cancer, which arises in the lining of the uterus, and uterine sarcoma, which arises in the muscular wall of the uterus.  Recently, second-generation SERMs have been developed, including Raloxifene and LY117018, which belong to the benzothiophene class of compounds.  They act as estrogen receptor antagonists in breast and uterine tissue, and as agonists in bone and cholesterol metabolism (Erlandsson et al., 2000).  Lately, a third-generation drug, EM-652 has been synthesized and reported as a potent new SERM by Lambrie et al. (2001).  However, the effects of EM-652 on the nervous system have yet to be established.  

            Raloxifene is marketed in the United States as Evista^® and is approved for clinical use to prevent and treat osteoporosis in postmenopausal women.  Recent studies have indicated that raloxifene has estrogenic and neuroprotective action in the nervous system as well (Nilsen et al., 1998; Callier et al., 2001; Cyr et al., 2001).  Raloxifene induces neurite outgrowth in estrogen receptor positive PC12 cells (Nilsen et al., 1998).  The drug has estrogenic effects on modulation of NMDA receptors in the rat brain, suggesting a possible beneficial effect in neurological disorders (Cyr et al., 2001).  Moreover, raloxifene protects dopaminergic structures in mouse brain from damage induced by the neurotoxin MPTP (Callier et al., 2001).  Raloxifene binds both ERa and ERb (Jisa et al., 2001) and may induce “nongenomic” effects via ERK and Akt kinase pathways promoting cell survival (Hisamoto et al., 2001).  Therefore, there is a pressing need to investigate further the neuroprotective and regenerative actions of raloxifene in the nervous system, effects which might allow using this drug to treat neurological conditions and to accelerate the restoration of compromised function in humans. 

Our recent studies of raloxifene analog LY117018 showed significant acceleration of functional recovery after peripheral nerve injury (McMurray et al., 2002).  Therefore, there is a pressing need to investigate raloxifene regenerative action which may permit using this drug clinically.  Peripheral nerve crush initiates a cascade of events including retraction of damaged proximal axon, chromatolysis of axotomized neuron, Wallerian degeneration of distal part of the axon, degeneration of motor plates and atrophy of denervated muscles.  Connectivity between motor neuron and the target muscle is reestablished during subsequent nerve regeneration.  During rehabilitation of compromised function physiological properties and functionality of regenerated nerves, motor end plates and muscles are restored.  The action of raloxifene via estrogen receptor system may be fairly complex affecting axons, neuromuscular junctions, skeletal muscles and behavioral performance.  Consequently, the effect of raloxifene treatment on rehabilitation after peripheral nerve injury will be analyzed at several levels utilizing: behavioral tests; electrophysiology on neuromuscular preparations and muscles; histomorphology of nerve fibers, and motor end plates.  The metabolism and phenotype of denervated and reinnervated skeletal muscles will be assessed by biochemical assays and immunohistochemistry. 

Peripheral nerve neuropathy and regenerative deficit is a common and debilitating complication of diabetes (Kennedy, Zochodne, 2000; Goss et al., 2002).  In the current proposal we will also investigate if raloxifene may protect against peripheral neuropathy and regenerative deficit in streptozotocin induced diabetic mice.  We will analyze sensory deficits and motor function in diabetic mice using functional behavioral tests, histomorphology of nerve fibers, and measuring sciatic nerve conduction velocity.

C.  Preliminary Studies.

Our published observation showed that administration of 17b-estradiol significantly accelerates peripheral nerve regeneration and functional recovery in ovariectomized (ovx) female mice (Islamov et al., 2002).  We further examined expression and distribution of estrogen receptors after crush injury (Islamov et al., 2002 under revision).  RT-PCR was performed to investigate the expression of ERa and ERb in the spinal cord after sciatic nerve crush in placebo and estrogen treated ovx mice at different time points during nerve regeneration (4, 7, 14 and 21 days).  Both ERs are expressed in the spinal cord: primer pairs amplified ERa (608 bp) and ERb (351 bp) mRNAs.  To test the hypothesis that motor neurons increase expression of ERs after injury, we used real-time quantitative RT-PCR to analyze the expression of both ER genes in the ipsilateral and contralateral parts of the lumbar spinal cord after sciatic nerve crush in placebo and estrogen treated mice.  The experiments showed a higher expression of ERa and ERb genes in the ipsilateral versus contralateral sides of injury, normalized for GAPDH expression.  The level of ERa gene expression was 1.15 and 1.07 fold higher in placebo mice and in estrogen treated mice respectively. The level of ERa gene expression was 1.09 fold higher in placebo mice and 1.12 fold in estrogen treated mice.  Immunohistochemically, we found ERa and ERb localized in the cell bodies and neurites of lumbar motor neurons.  Light microscopy analysis revealed an increase in the immunoreactivity of motor neurons on the injured side of the spinal cord in both placebo and estrogen treated mice during motor neuron regeneration.  The finding that axotomized motor neurons increase expression of ERs raised the question whether ERs were transported from cell bodies into regenerating axons.  Accumulation of ERs in sciatic nerve was confirmed by Western blot.  Immunohistochemical staining of sciatic nerves from placebo and estrogen treated mice revealed the presence of both isoforms of ERs in regenerating axons.  Our data suggest that both ERa and ERb are involved in sciatic nerve regeneration.

We have further investigated whether SERMs may have a similar effect on peripheral nerve regeneration (McMurray et al., 2002).  Although we have not tested raloxifene we have examined its analog LY117018 (kindly provided by Eli Lilly and Company), which differs from raloxifene by one atom  and exerts virtually the same pharmacological action.

ICR, 8 week old, female mice were employed in this study.  All surgical procedures were performed under ketamine-xylazine (1.5mg/100g) anesthesia.  A week after bilateral ovariectomy, the right sciatic nerve was crushed for 15 seconds with a fine hemostat.  Experimental animals were injected with LY117018 subcutaneously (3mg/kg in 100ml DMSO) five days a week, for 3 weeks.  Control animals received DMSO injections.  Functional recovery was analyzed using the “Walking Corridor” test (Bain, at al., 1989; Islamov et al., 2002), which can be quantified with a sciatic functional index (SFI).  Mice (n=10 in each group) were tested at 2, 4, 7, 14 and 21 days after sciatic nerve crush (Fig. 1).

Recovery for both groups began by the 4^th day and function returned to normal levels by the 21^st post-operative day.  LY117018 treated mice had functional index values significantly higher (p<0.05) at days 7 and 14, and approached pre-operative values by the 14^th day.  For placebo mice, functional index values returned to pre-operative levels by the 21^st day of regeneration.  That means that LY117108 treated mice had the rate of functional recovery almost 1.5 times higher than placebo mice.  In clinical settings this may translate into recovery months earlier in humans.  For example, if a patient with conventional treatment may achieve complete recovery in 24 months, treatment with LY117108 may help this patient to recover in 16 months. 

For histological analysis of nerve regeneration, we removed the four-millimeter section of the sciatic nerve distal from the crush site.  Transverse semi-thin sections stained with Richard’s dye were studied by light microscopy.  Our results suggest that LY117108 promotes regeneration of the sciatic nerve after crush injury.  LY117018 enhanced functional recovery by accelerating growth of regenerating nerve fibers.  One week after nerve crush, the total number of regenerating nerve fibers in LY117018-treated mice was significantly greater than in placebo mice (Fig. 2).

Although it seems reasonable to incorporate testing of LY117018 into the current proposal, it is not possible because the drug was received from Eli Lilly and Company under a material transfer agreement which does not allow federal support for LY117018 research.  However, since the biochemical and pharmacological properties of LY117018 and raloxifene are basically the same, we anticipate that raloxifene will have an analogous effect on regeneration and functional recovery after sciatic nerve crush. 

We have further investigated whether SERMs may have a similar effect on peripheral nerve regeneration in males.  We have tested action of tamoxifen on regeneration of sciatic nerve after crush injury in normal 8 week old ICR male mice.  Tamoxifen was injected daily 1 mg/kg body weight in glycerol, intraperitoneally.  Placebo animals received vehicle.