Abstract
Zoniporide, an inhibitor of the Na+-H+ exchanger-1, was administered by continuous intravenous infusion to rats and dogs for up to 1 month. In 1-month studies, histological and functional changes were observed in select portions of the peripheral nervous system; however, these findings were not detected in 2-week studies at similar or higher doses. In the 1-month rat study, there was dose-dependent, minimal, focal, or multifocal nerve fiber (axonal) degeneration in the spinal cord and/or sciatic nerve. In a follow-up rat study, findings included slowing of caudal nerve conduction velocity and axonal degeneration in the spinal cord (dorsal funiculus), dorsal roots, dorsal root ganglia (DRG), radial, sciatic, and tibial nerves. In the 1-month dog study, there was impairment of the patellar reflex and associated postural reaction changes, minimal to marked proximal nerve fiber degeneration in the DRG, and minimal nerve fiber degeneration in the dorsal roots and funiculi of the spinal cord. Minimal nerve fiber degeneration of equivocal significance was noted in various peripheral nerves. Taken together, these findings were consistent with a specific effect on peripheral sensory nerve fibers. These studies demonstrated that zoniporide produces clinical, electrophysiologic, and microscopic evidence of peripheral sensory axonopathy and establishes the importance of careful preclinical evaluation of neurological function.
Keywords
Introduction
The sodium-hydrogen exchanger (NHE) is a major membrane transporter responsible for the regulation of cytosolic pH and cell volume and taking up sodium ions into the cell (Masereel et al., 2003; Tracey et al., 2003). Following intracellular acidosis, NHE restores physiological pH by H+ extrusion. Excessive stimulation results in an increase of intracellular Na+ concentration and a subsequent activation of Na+/K+ ATPase. The high intracellular Na+ level contributes to activate the sarcolemmal Na+/Ca++ antiporter, which leads to increased intracellular Ca++ (Masereel et al., 2003). To date, 9 isoforms (NHE-1 to NHE-9) have been identified within the NHE family (Slepkov et al., 2007). The NHE-1 isoform is ubiquitously expressed including the peripheral and central nervous system (Bond et al., 1998) and is activated by growth factors expressed in several cell types including cardiomyocytes, platelets, and on the basolateral membrane on renal tubules (Masereel et al., 2003). In preclinical models, inhibition of this isoform has been shown to produce marked reductions in myocardial infarct size and arrhythmias (Knight et al., 2001). Zoniporide (Figure 1) is a highly selective inhibitor (157-fold and 15700-fold vs. human NHE-2 and rat NHE-3, respectively; Marala et al., 2002) of the Na+ -H+ exchanger-1 (NHE-1). CP-703,160 (Figure 1) is an active metabolite of zoniporide but is slightly less potent (Tracey et al., 2003). This metabolite is formed in rats, dogs, and humans. It is also selective for NHE-1 versus NHE-2 and NHE-3. Zoniporide is intended for the reduction of myocardial ischemic injury in acute coronary syndromes, in the high-risk surgical setting (perioperative myocardial injury), and in secondary prevention in patients with ischemic disease (Tracey et al., 2003).
Preclinical toxicology studies, including 2-week continuous intravenous infusion studies in rats and dogs, were conducted to support clinical trials with zoniporide (Tracey et al., 2003). These studies showed that zoniporide was well tolerated with no gross or microscopic findings. In clinical studies of 3 to 5 days dosing duration, zoniporide produced periocular pain and paresthesia in some subjects (unpublished observations). Longer-term (1 month) preclinical studies were conducted in rats and dogs to support further clinical development of zoniporide. In contrast to the 2-week studies, a standard 1-month continuous intravenous (iv) infusion study in rats showed degeneration of the sciatic nerve and spinal cord. Follow-up studies in rats and dogs were designed and conducted to further characterize this finding and determine if there was a functional correlate. These follow-up studies included regular clinical and electrophysiological evaluations of neurological function, and extensive collection of nervous tissues for microscopic examination. In this report, we described the clinical, electrophysiologic, and microscopic evidence of peripheral sensory axonopathy and established the importance of careful preclinical evaluation of neurological function.
Material and Methods
Animals and Husbandry
Male Sprague-Dawley CD (Crl:CD®(SD) IGS BR) rats (8 to 9 weeks of age) were obtained from Charles River Canada, St. Constant, Quebec, Canada. Male and female beagle dogs (8 to 12 months of age) were obtained from Covance Research Products, Inc., Kalamazoo, MI, USA. Animals were individually housed in stainless-steel mesh-bottomed cages (rats) or bar-type floor (dogs) equipped with an automatic watering valve. The animal room environment was controlled (22 ± 3°C, humidity 50 ± 20%, 12-h light/dark). Animals received a certified commercial pelleted laboratory diet (PMI Certified Rodent Chow: 5002 or Certified Dog Chow: 5007: PMI Nutrition International, Richmond, IN). Rats had ad libitum access to diet. Dogs initially had access to diet for 2 to 3 h; however, due to poor appetence associated with high room activity levels, they were provided unrestricted access to food beginning on day 8. Animals were fasted overnight prior to blood sampling for clinical pathology and prior to necropsy. Drinking water was provided ad libitum.
The protocol and any amendment(s) or procedures involving the care and use of animals in these studies were reviewed and approved by Charles River Laboratory’s Institutional Animal Care and Use Committee prior to study conduct. During the course of the studies, the care and use of animals were conducted in accordance with the regulations of the United States National Research Council and Canadian Council on Animal Care.
Test Substance, Experimental Design, Observations, and Measurements
Zoniporide was supplied by Pfizer Global Research and Development. Five percent dextrose for injection was supplied by Baxter Canada. Solutions of zoniporide were prepared in 5% dextrose for injection, and the formulations were sterile filtered through 0.2 μm polyvinylidene difluoride filters prior to transfer to syringes or iv bags for administration. Samples were periodically analyzed for pH, density, osmolality, and zoniporide concentration.
For all studies, zoniporide or vehicle control (5% dextrose for injection) were administered by continuous iv infusion. The high dose level for each study was the maximum tolerated dose as determined by previous dose range finding and 2-week studies. The middle and low doses were selected in consideration of the anticipated clinically efficacious dose and plasma concentration. The first day of dosing was defined as study day 1. For each study, all animals were observed at least once daily for mortality and signs of illness and behavioral changes. Body weights and food consumption (daily for dogs) were determined weekly. Ophthalmologic examinations were conducted by a board-certified veterinary ophthalmologist (prestudy and end of study for rats; prestudy, day 8, and end of study for dogs). Routine clinical laboratory evaluations were determined prestudy and end of study for rats and prestudy, day 8, and end of study for dogs. Toxicokinetic evaluations were conducted on day 1 and end of study. Complete necropsies, organ weights (initial 1-month rat and 1-month dog), and a comprehensive (initial 1-month rat and 1-month dog) or focused (1-month rat neurotoxicity) list of tissues was collected. Standard tissues were embedded in paraffin wax, sectioned, stained with hematoxylin and eosin, and examined. Special tissue fixation and processing techniques employed for the dog and rat neurotoxicity studies are outlined below (“Postmortem Procedures”).
In the initial 1-month rat study, rats (15/sex/treatment) received vehicle control, 15, 50, or 150 mg/kg/day zoniporide. Based on findings of axonal degeneration in the sciatic nerve in the initial study, a follow-up 1-month neurotoxicity study was conducted at 150 mg/kg/day zoniporide (15 males/treatment). For the rat neurotoxicity study, a functional observational battery (FOB) with an analgesic test (tail flick) was performed on each animal once prior to surgery, once postsurgery, and again on days 14 and 28. FOBs were based on a modification of the procedures of Moser et al. (1988) and Haggerty (1989). FOB tests were performed by trained technicians with knowledge of the animal’s treatment. The FOB is a series of noninvasive observational and interactive measures that assess the neurobehavioral and functional integrity of the animal. The FOB encompasses measures of home cage and open field activity, stimulus reactivity, as well as assessment of physiologic and neuromuscular function (Haggerty, 1991). Functional observational battery tests included assessments of body position, presence of convulsions and bizarre or stereotypic behavior, ease of removal from cage, vocalizations, rearing activity, gait assessment, palpebral closure, presence of piloerection, respiratory rate/pattern, locomotor activity level, arousal level, grooming behavior, frequency and character of defecation and urination, presence of lacrimation, pupil size, salivation, urinary staining, body tone, extensor thrust, corneal reflex, pinna reflex, toe pinch, tail pinch, visual placing, positional passivity, auricular startle, air righting reflex, olfactory response, fore and hind limb grip strength, and tail flick.
A battery of electrophysiological measures were assessed postsurgery but prior to dosing and on days 15 and 29. All measures were recorded with the rats on isoflurane anesthesia (15/treatment) using a portable electrophysiologic instrument (BIOPAC MP150, Acknowledge version 3.5.7 software, BioPac Systems, Santa Barbara, CA). Endpoints included maximal nerve conduction velocity and peak amplitude of a purely sensory nerve (distal sensory sciatic) or mixed nerve (caudal). This endpoint assessed the integrity of the myelin sheath, mean cross-sectional diameter of the responding neuron fibers, and the integrity of the myelin-axon interface. Slowing is an accurate and valid measure of toxic neuropathy. Peak amplitude of a response driven by supramaximal stimulation reflects the number and timing properties of responding axons. Compound muscle action potential (CMAP)–distal motor nerve conduction and amplitude of the induced muscle contraction associated with supramaximal stimulation of a motor nerve was assessed in the tibial nerve. The peak maximal CMAP reflects the number of responding motor fibers and the synchrony of innervation. This measure is sensitive to neuropathy affecting conduction in the distal peripheral motor nerve, alteration of neuromuscular transmission, and myopathy affecting depolarization of the muscle. In addition, somatosensory-evoked potentials (SEP) were obtained from the cervical spinal cord following stimulation of the median nerve. SEP is a measure of the initial depolarization overlying the cervical spinal cord following stimulation of a distal peripheral nerve. The latency of the response reflects the integrity of transmission along the sensory nerve from the distal site of stimulation, through the proximal nerve (including the dorsal root) as well as the activation of dorsal horn sensory neurons and afferent fiber tracts. Slowing of SEP latency with normal peripheral nerve measures provides compelling evidence for a central nervous system deficit. During all recording sessions, rats were placed in a prone position, respiration was maintained within normal limits, and temperature was maintained between 35.0 and 39.0°C. All data were obtained using subdermal needle electrodes (platinum) positioned with reference to bony landmarks. Supramaximal stimulation was achieved using a constant voltage square pulse (0.05 to 0.2 msec duration) isolated from the ground.
In the 1-month dog study, dogs (5/sex/treatment) received vehicle, 2, 10, or 20 mg/kg/day. Zoniporide and vehicle were administered by continuous intravenous infusion (dose rate was 2.0 mL/kg/h). Neurological examinations were performed once prior to surgery, once postsurgery, and again on days 7, 14, 21, and 28. Neurological examinations were conducted as described by Schaeppi (1982) and De Lahunta (1983). Neurological assessments were conducted by a trained technician with oversight by a clinical veterinarian. Assessments were conducted with knowledge of the animal’s treatment. The assessment evaluated general attitude and behavior, gait, postural reactions (assessments of proprioceptive positioning, hemihopping/hemistanding, wheel barrowing, hopping, visual and tactile placing reactions, righting reaction), cranial nerve function (assessment of head movement/symmetry, head muscle tone, eye reactions, eye symmetry, vestibular nystagmus, eye position, corneal reflex, papillary light reflex, nasal septum pinch, resistance to opening mouth, tongue is pulled and observed for size or symmetry, movement and strength, and pharynx [swallow or gag reflex test]), and spinal nerve function (assessments of muscle tone, patellar reflex, flexor reflex, panniculus reflex, and perineal reflex). Electrocardiography was performed on all dogs twice during pretreatment (once prior to surgery and once postsurgery and again on days 8 and 29).
Surgical Procedures
A minimum of 1 week was allowed between receipt of the animals and surgical implantation of the catheters. Animals received an intramuscular (im) dose of benzathine penicillin G and procaine penicillin cocktail on the day of surgery and again 2 days following surgery. Rats were anesthetized with isoflurane, and dogs were preanesthetized with an im injection of a mixture of acepromazine, butorphanol, and glycopyrrolate followed by an iv injection of thiopentone sodium to induce anesthesia. The dogs were then intubated and anesthesia was maintained with isoflurane. The right femoral vein was cannulated with a medical-grade silicone-based catheter that was placed in the vena cava at approximately the level of the kidneys. The catheter was connected to a calibrated infusion pump (Baxter AS 40 for rats, Sigma 8000 for dogs) placed outside the cage, and all animals were continuously infused with 0.9% sodium chloride for injection at a rate of 0.4 mL/kg/h (rats) or 4 mL/h (dogs) until day 3. From day 3, the rate was increased to 2 mL/kg/h and the saline infusion continued until day 1, at which time the infusate was switched to 5% dextrose for injection until initiation of treatment. On day 1, vehicle and zoniporide were administered by infusion at 2 mL/kg/h. Infusions were temporarily interrupted during conduct of the neurological assessments.
Toxicokinetics
Toxicokinetic analyses were conducted to confirm systemic exposure to zoniporide and its active metabolite (CP-703,160). Blood samples were collected from a jugular vein. Lithium heparin was used as an anticoagulant. Rat blood samples were collected in a population sampling manner on days 1 (24 and 48 h after initiation of infusion), 14 or 15 (336 h after initiation of infusion), and 28 or 29 (0.25, 4, and 24 h after end of infusion). For dogs, blood samples were collected from each animal at approximately 24 and 48 h after initiation of infusion, on day 15, and again at 0.25, 4, and 24 h after the end of infusion on day 29. Plasma concentrations of zoniporide and CP-703,160 were determined by solid-phase sample extraction and LC/MS/MS detection. Toxicokinetic parameters were calculated using WinNonlin v. 3.2.
Postmortem Procedures
On day 30, rats were subject to necropsy following an overnight period of food deprivation. For the neurotoxicity study, one subset of 10 rats/group (A) was exsanguinated from the abdominal aorta following isoflurane anesthesia. Tissues were collected, retained, and fixed in 10% neutral buffered formalin. The second subset of 5 rats/group (B) was subject to whole-body perfusion fixation via the left ventricle at necropsy. Rats were deeply anesthetized by ip injection of sodium pentobarbital and perfused with 2.5% glutaraldehyde, 0.5% paraformaldehyde in 0.1 M sodium phosphate buffer. The bottles with the solutions for perfusion were kept approximately 1 m above the level of the animal (1 m = approximately 75 mm Hg pressure). The perfusion was performed with a 14-gauge catheter inserted into the left ventricle. A perfusate composed of a sodium chloride solution containing heparin and sodium nitrite was infused for approximately 2 min prior to the perfusion of the fixative solution (mixture of 2.5% glutaraldehyde, 0.5% paraformaldehyde in 0.1 M sodium phosphate buffer). The fixative was circulated until a volume equal to the body weight had been used. On completion of the perfusion, nervous tissues were harvested and placed in 10% neutral buffered formalin. Collected tissues from subsets A and B included brain, spinal cord, gastrocnemius, cauda equina, radial nerve, sciatic nerve, sural nerve, tibial nerve, gasserian ganglion, lumbar (L4), thoracic (T7) and cervical (C5) dorsal root ganglion (DRG), dorsal root and ventral root, and grossly abnormal central or peripheral nervous system tissues. All tissues from subset A as well as the brain, spinal cord, cauda equina, and muscles from subset B rats were embedded in paraffin, sectioned at 4 μm, stained with hematoxylin and eosin, and examined with light microscopy. Gasserian ganglia, nerves, dorsal root, and DRG (subset B) were embedded in epoxy, cut at 0.5 μm, stained with toluidine blue, and examined by light microscopy.
Histological grading was performed in a semiquantitative manner. The approximate numbers of degenerative axons were recorded as a modifier in the individual microscopic data, representing the highest number seen in either the cross or longitudinal sections (not a summation). For nerve fiber degeneration, a grade 1 (minimal) was recorded when approximately 1 to 20 affected fibers were observed in the tissue section; a grade 2 (slight) was recorded when more that 20 fibers were affected; and a grade 3 (moderate) was recorded when more than 30 degenerating fibers were observed (there were no grade 4 and 5 changes).
Representative radial (mid forelimb region), sciatic (mid thigh region), sural (knee) and tibial (knee) nerves (cross and longitudinal sections), ventral root, dorsal root, and DRG samples from 2 control and zoniporide-treated rats of subset B were selected for ultrastructural examination. Tissues were rinsed in 0.1 M sodium cacodylate buffer and placed in 2% osmium tetroxide for 2 h, and then each piece of tissue was rinsed in buffer and stained in a 1% aqueous solution of uranyl acetate for 2 h. The tissues were thereafter rinsed in distilled water, dehydrated in ascending concentrations of ethyl alcohol, and embedded in a mixture of Jembed and Araldite. Epoxy sections (0.5 μm) were obtained with a glass knife, stained with borate buffered 1% toludine blue, cover slipped, and examined by light microscopy to choose the best regions to examine under electron transmission microscopy. Based on this, appropriate ultrathin sections were mounted on copper grids, contrasted with uranyl acetate and lead citrate, and evaluated with an electron transmission microscope.
On day 30, overnight fasted dogs were sedated with im Ketamine HCl and Xylazine. Animals were anesthetized by an iv dose of sodium pentobarbital and exsanguinated following an incision of the axillary arteries. A detailed necropsy was performed, selected organs were weighed, and a comprehensive set of tissues was collected and fixed in 10% neutral buffered formalin. The tissue collection list also included the brain, the extraocular muscles (M. rectus dorsalis, M. rectus ventralis, M. rectus lateralis, and M. rectus medius), peripheral nerves (cranial nerve V [trigeminal], femoral, median, saphenous, sciatic, peroneal, sural, tibial, and ulnar), and spinal cord (cervical, mid thoracic and lumbar, dorsal and ventral roots, and corresponding DRG for each cord segment sampled). Tissues were embedded in paraffin, cut at 4 μm, stained with hematoxylin and eosin, and examined with light microscopy.
Data Analyses
Body weight, food consumption, and quantitative FOB data (count data were transformed before analysis) were subjected to calculation of group mean values and standard deviations. Group variances for the appropriate parameters were then compared using the F test (rats) or Levene’s test (dogs). When differences between group variances were not significant (p > 0.05), the student’s t test (rat) or Dunnett’s t test (dog) were used to determine differences between the control and treated group. Significance was declared at the 0.05, 0.01, or 0.001 alpha levels. For rats, when the F test indicated a significant difference between group variances for a given parameter, the significance of the differences between the control and treated group was assessed using Cochran and Cox’s modified t test (Cochran and Cox, 1950). Significance was declared at the 0.05, 0.01, or 0.001 alpha levels. Where appropriate, qualitative FOB data were analyzed by comparing the control group to the treated group using Fisher’s exact probability test. Significance was declared at the 0.05, 0.01, or 0.001 alpha levels, where appropriate. For the dog study, if Levene’s test indicated significance of the differences between the control group and each test group was assessed using Cochran and Cox’s modified t test. Significance was declared at the 0.05, 0.01, or 0.001 alpha levels.
Results
1-Month Rat Study
In the initial 1-month rat study, zoniporide was clinically well tolerated at all doses. There was no mortality and no clearly defined effects upon clinical condition attributable to zoniporide. Body weight, food consumption, hematology, clinical chemistry, urinalysis, ophthalmology, and organ weights were unaffected. The principal toxicity associated with zoniporide was axonal degeneration (1–25 nerve fibers per tissue section), degeneration in the sciatic nerve and/or spinal cord at ≥ 15 mg/kg. Dose levels of 15 and 50 mg/kg were characterized by minimal focal axonal degeneration (1 nerve fiber per tissue section), while a higher incidence of axonal pathology was detected at 150 mg/kg (1–25 nerve fibers per tissue section). These effects appear to be compound- and dose-related; however, minimal axonal degeneration (1 nerve fiber per tissue section) was also observed in one male and one female control rat. There were no clinical signs that correlated with the sciatic nerve and spinal cord findings; however, clinical observations performed within the confines of the cage environment are likely to have low sensitivity for the detection of minor neurological lesions. Systemic exposure (AUC 0-tlast and Css) of zoniporide and its major metabolite, CP-703,160, increased in a dose-related manner, and there were no apparent gender-related differences in exposure (Table 1).
1-Month Rat Neurotoxicity Study
The 1-month neurotoxicity study evaluated the effects of 150 mg/kg zoniporide on the central and peripheral nervous system. In this study, zoniporide was clinically well tolerated, although yellow urogenital fur staining was commonly observed. Body weight and food consumption were usually comparable or slightly lower than those of the control animals. Weeks 1, 2, and 4 body weight gains were reduced to 63% to 77% of control. Toxicokinetic data were similar to that obtained in the initial 1-month study (Table 2; Figures 2– 7).
Prestudy mean hindlimb grip strength values were similar in control (600.3 g) and zoniporide-treated (599.9 g) rats. On study day 14, mean hindlimb grip strength values were slightly lower in the zoniporide (642.9 g) versus the control (689.2 g) rats (not statistically significant). However, a small, statistically significant (p < 0.01) decrease in mean hindlimb grip strength was noted in zoniporide-treated rats on study day 28 (control 752.6 g, zoniporide 635.7 g). Forelimb grip strength values were unaffected by zoniporide. There were no other treatment-related FOB changes.
A battery of electrophysiologic measures was used to assess the possible presence of functional deficits consistent with an induced peripheral axonopathy and/or altered transmission in spinal sensory pathways, each of which was suggested by histopathology in the initial study. As expected, maximal caudal nerve conduction velocity (NCV) progressively increased in the control group over the course of the study (Table 2; Figure 2), consistent with maturation and increased myelination (ANOVA = 0.0001). Exposure to zoniporide at 150 mg/kg/day resulted in a small (approximately 10%), but significant, slowing of NCV in the caudal nerve at day 15. At this time, the velocity in the treated group was slowed by 4.4 meters/sec compared to age-matched controls (p = 0.0001). The deficit in caudal nerve conduction was slight, representing an approximate 10% change from the control group. Stated another way, caudal NCV increased by approximately 10% from baseline values in the control group and remained unchanged in the treated group over the initial 15-day treatment period. There were no significant changes in the amplitude of the caudal nerve response or in any of the other physiologic measures across treatment groups at day 15. The difference in caudal nerve NCV between the treated and control groups continued to be present at the final recording point (i.e., day 29) (p = 0.0003); however, there was little evidence of progressive deficit. Mean values were reduced by 4.4 meters/sec relative to control animals at day 15 and by 5.5 meters/sec at day 29. As was the case at day 15, the slowing of caudal nerve NCV at day 29 was not associated with significant change in caudal nerve amplitude or with a significant deficit in any other physiologic measure. The SEP recorded from the cervical spinal cord, which provides a sensitive measure of conduction in the dorsal columns of the spinal cord, was especially stable over the treatment periods and showed no difference or trend across treatment groups.
Treatment-related microscopic findings included degeneration of nerve fibers in the gasserian ganglia; dorsal funiculus of the spinal cord (cervical, thoracic, and lumbar); the cervical, thoracic, and lumbar dorsal nerve roots and DRG; the cauda equina; and the radial, sciatic, and tibial nerves. No effects were noted in the ventral funiculus of the spinal cord (cervical, thoracic, and lumbar); the cervical, thoracic, and lumbar ventral roots; and the sural nerve. There was no evidence of necrosis or degeneration of nerve cell bodies in any site including spinal cord and DRG.
Examination of selected tissues by electron microscopy revealed giant axonal swelling typically located proximally to the neuronal cell body in the DRG. The changes were predominantly affecting the axonal hillock and the initial segment and proximal portion of the myelinated axons. Markedly enlarged axons with noticeable accumulation of neurofilaments, thinning, retraction or rupture of the myelin sheath, peri-axonal void space, and infiltration of debris-laden macrophages typified the ultrastructural changes.
1-Month Dog Study
Zoniporide was clinically well tolerated. There were no treatment-related effects on clinical condition, body weight, food consumption, blood pressure, body temperature, or respiratory rate. Moreover, there were no toxicologically significant effects upon routine hematology, immunochemistry and clinical chemistry parameters, and organ weights, and there were no treatment-related macroscopic changes (Tables 3 and 4; Figure 8).
The principal finding from the neurological examination was an absence of the patellar reflex, which was first noted at the day 7 neurological examinations in 2/10 and 9/10 dogs at 10 and 20 mg/kg, respectively. Additional dogs that received 10 mg/kg were affected at subsequent examinations on days 14 (6/10) through 28 (8/10). Slight changes in postural reactions including hemihopping/hemistanding, wheel barrowing (hindlimbs), hopping (hindlimbs), placing reaction (hindlimbs), righting reaction, and gait were noted at 10 and 20 mg/kg. The incidence and time course for these changes was variable, and all effects were limited to the hindlimbs. These postural and gait changes were considered to be secondary to the patellar reflex change. There was no evidence of paresis or muscle atrophy. The clinical presentation and minimal progression or expansion of neurological findings is consistent with a sensory axonopathy. On an incidental note, on day 28 intravenous infusion of zoniporide was temporarily interrupted in a 20 mg/kg dog in order to repair a leaking catheter site. Subsequent examination of this dog on day 29 showed that the patellar reflex had returned to normal, inferring that this effect was rapidly reversible upon cessation of dosing. Microscopic findings include dose-dependent nerve fiber degeneration in the DRG (cervical, thoracic, and lumbar) at ≥ 2 mg/kg/day and minimal nerve fiber degeneration in the dorsal roots (cervical, thoracic, and lumbar) and dorsal funiculi of the spinal cord at ≥ 2 mg/kg/day. Minimal focal or multifocal nerve fiber degeneration was noted in the peripheral nerves of a few dogs of the control and zoniporide-treated animals, indicating no clear effect of zoniporide. The findings in the dorsal funiculus of the spinal cord are attributable to central projections of axons from the DRG. The predilection for degeneration of dorsal funiculus and roots and DRG as compared to ventral funiculus and roots is also consistent with a mild peripheral sensory axonopathy. There was no evidence of paresis, muscle atrophy, and no histological evidence of demyelination, neuronal necrosis, or effects on Schwann cells. Extraocular muscles were evaluated because of the aforementioned observation of periocular pain in the human clinical trial studies. There was no microscopic abnormalities detected in the examined extraocular muscles.
Discussion
In clinical trials of zoniporide, some treated human subjects experienced reversible periocular pain and mild, transient paresthesia that included tingling of the tongue and lips (unpublished observations), although 2-week preclinical studies showed no effect on nervous system tissues or clinical signs consistent with neurological impairment (Tracey et al., 2003). In the 1-month rat and dog studies, zoniporide treatment resulted in changes consistent with a peripheral sensory axonopathy. These findings included impairment of the patellar reflex, slowing of the caudal nerve conduction velocity, degenerative changes in the dorsal nerve roots and spinal ganglia, spinal cord changes (DRG projections), and axonal degeneration in large myelinated peripheral nerves. It is noteworthy that there was minimal progression or expansion of the functional changes. For instance, there was minimal progression in the percent change relative to control in caudal nerve conduction velocity from days 14 to 28. There was also minimal progression in the severity of the results from the neurological examination in dogs. The small decrease in hindlimb grip strength noted in the zoniporide-treated rats possibly resulted from the peripheral sensory axonopathy; however, a small decrease in body weight gain was also noted, and this may have contributed to the decrease in hindlimb grip strength.
The patellar tendon reflex is commonly used in the clinical diagnosis of lower motor neuron conditions in dogs. The reflex is mediated by the femoral nerve through the fourth to sixth lumbar spinal cord segments (De Lahunta, 1983). Decreases in this reflex correlate well with electrodiagnostic tests (Levine et al., 2002). Postural reactions require that all components of the peripheral and central nervous system be intact. In the dog study, the postural reactions were decreased but not asymmetrical. These results are consistent with the microscopic findings of a mild peripheral sensory axonopathy.
The caudal nerve contains large myelinated sensory fibers. The lack of effect on tibial NCV is consistent with the minimal degeneration seen microscopically. Furthermore, there was not a clear distal to proximal gradient. The histologic findings in the dorsal funiculus of the spinal cord are attributable to central projections of axons from the DRG. However, the minimal severity of these spinal cord changes was consistent with the lack of effect on the cervical somatosensory electrophysiology response. The predilection for degeneration of dorsal funiculus and roots and DRG as compared to ventral funiculus and roots is also consistent with a mild peripheral sensory axonopathy. In addition, there was no evidence of muscle atrophy; paresis; effects on tibial nerve conduction velocity or cervical somatosensory response; effects on motor neurons, ventral roots, and funiculus, distal to proximal gradient; or evidence of any neuronal cell body degeneration or necrosis, demyelination, or effects on Schwann cells.
The apparent vulnerability of the sensory axons to zoniporide may be related to the relative fenestration of the blood-nerve barrier at the DRG (Spencer, 2000). Many of the capillaries in the vicinity of the DRG lack specialized endothelial tight junctions, and thus the blood-nerve barrier is relatively ineffective in this region. This allows circulating compounds to freely communicate with the interstitial fluids in the ganglia and to have relatively unimpeded access to the soma of bipolar sensory neurons. The exposed neurons send distal projections to form the sensory nerves and central projections that enter the spinal cord (Arezzo et al., 1982), and they are clearly sensitive to a number of neurotoxic insults, including those related to a variety of chemotherapy agents, antiviral compounds, and industrial toxins (Schaumburg, 1992).
In the 1-month rat and dog studies, the functional changes preceded the morphological changes by weeks. In the rat, the functional neurological deficits were detected by decreased hindlimb grip strength and electrophysiology evaluations, and in the dog by neurological evaluation. Routine evaluation of clinical signs failed to detect any evidence of neurological changes. Taken together, these studies reinforce the importance of including detailed functional evaluation of the nervous system in preclinical studies for potential neurotoxicants. Since these pre-clinical evaluations have correlative tests in humans, they provide important guidance for monitoring of drug candidates for peripheral sensory axonopathy in clinical studies. This was particularly useful for this drug candidate since it is intended for short-term administration for prevention of a serious surgical complication (myocardial ischemic injury). Since the functional neurological deficits preceded the morphological changes, the risk to patients would be manageable, as zoniporide is only intended for administration over a limited timeframe thus minimizing the risk for developing structural neurological changes, which would be expected to reverse more slowly. Reversal of the patellar reflex impairment in the animal that had dosing temporarily discontinued while awaiting surgical repair suggests that the functional deficits may be readily reversible. This would need to be studied in a follow-up study.
Preclinical studies (unpublished studies) demonstrated that zoniporide does not penetrate the blood-brain barrier. This is consistent with the distribution of the morphological changes, as the affected areas were limited to those outside the blood-brain barrier. This is also consistent with the clinical findings of periocular pain and paresthesia. Since the gasserian ganglion (trigeminal nerve) has both motor and sensory components, is outside the blood-brain barrier, and innervates the ophthalmic, maxillary, and madibular nerves, the preclinical findings are consistent with the clinical findings of periocular pain and paresthesia of the tongue and lips. The anatomical distribution of these findings suggests that the primary site of insult of NHE-1 inhibition is at the dorsal root ganglia. This is consistent with the observation that the sural nerve is unaffected by zoniporide, as it is protected by the peripheral blood-nerve barrier.
The mechanism for these neurological findings with zoniporide is unknown. NHE is a ubiquitously expressed protein and has been studied in a number of hypoxia and ischemia/reperfusion models in cardiac and central nervous system models (Avkiran and Marber, 2002; Douglas et al., 2003; Park et al., 2005; Yao et al., 2001). The importance of NHE-1 in function of the central nervous system is demonstrated by the observation that homozygous NHE-1 mutant mice exhibit decreased rates of postnatal growth, ataxia, and epileptic seizures (Bell et al., 1999). This seizure disorder may be related to an abnormally high hippocampal neuronal excitability due to an up-regulation of the sodium current and alterations in the channel characteristics (Gu et al., 2001). NHE-1 functions as a membrane anchor for the actin-based cytoskeleton and has been reported to play a role in regulating the cytoskeleton and cell shape independently of hydrogen ion translocation (Denker et al., 2000). It is possible that prolonged inhibition of NHE-1 could lead to chronic acidification or impaired volume control in neurons and alter other membrane transport mechanisms leading to neuronal hyperexcitability and/or cell death. Although zoniporide selectivity was not assessed against NHE-6, which is expressed in the brain (Orlowski and Grinstein, 1997), zoniporide does not cross the blood-brain barrier. Therefore, inhibition of NHE-6 is unlikely to have contributed to the axonopathy.
It is possible that the unusual morphologic manifestation of toxicity on the axon extending into the axonal hillock, while apparently sparing the neuronal cell body, may be a specific consequence of NHE-1 antagonism. The axonal hillock is the transition point between the axon proper and the cell body and shares some functions special to each. Whether the swelling of the axon is a primary event that progresses proximally to, but not beyond, the axonal hillock, or is in fact the primary site of damage to these sensory neurons, is open to speculation. A time course study may provide some insight to this question.
In conclusion, zoniporide produces preclinical findings of a mild peripheral sensory axonopathy that is consistent with clinical findings of mild, transient paresthesia and periocular pain. These findings can be monitored in clinical studies by the inclusion of routine neurological examinations. The onset of functional neurological changes preceded microscopic changes, suggesting that these findings can be managed by limiting the duration of dose administration. These results support the importance of careful preclinical evaluation on neurological function.