Research Article

CRMP2-binding compound, edonerpic maleate, accelerates motor function recovery from brain damage

See allHide authors and affiliations

Science  06 Apr 2018:
Vol. 360, Issue 6384, pp. 50-57
DOI: 10.1126/science.aao2300

A small molecule for stroke therapy

Better therapies for motor impairments after stroke are greatly needed. In mice and nonhuman primates, Abe et al. found that edonerpic maleate enhanced synaptic plasticity and functional recovery after a traumatic insult to the brain (see the Perspective by Rumpel). This recovery of motor function was accompanied by functional reorganization of the cortex.

Science, this issue p. 50; see also p. 30

Abstract

Brain damage such as stroke is a devastating neurological condition that may severely compromise patient quality of life. No effective medication-mediated intervention to accelerate rehabilitation has been established. We found that a small compound, edonerpic maleate, facilitated experience-driven synaptic glutamate AMPA (α-amino-3-hydroxy-5-methyl-4-isoxazole-propionic-acid) receptor delivery and resulted in the acceleration of motor function recovery after motor cortex cryoinjury in mice in a training-dependent manner through cortical reorganization. Edonerpic bound to collapsin-response-mediator-protein 2 (CRMP2) and failed to augment recovery in CRMP2-deficient mice. Edonerpic maleate enhanced motor function recovery from internal capsule hemorrhage in nonhuman primates. Thus, edonerpic maleate, a neural plasticity enhancer, could be a clinically potent small compound with which to accelerate rehabilitation after brain damage.

Brain damage mainly caused by stroke is a severe neurological condition that may lead to paralysis and compromise work capacity and self-care. No pharmacological intervention that could foster recovery and complement current rehabilitation has yet been established as effective. Restoration of motor impairment after brain damage is considered to be the result of compensative neural plasticity in intact brain regions, mediated by the reorganization of cortical motor maps (17). Experience-dependent synaptic AMPA (α-amino-3-hydroxy-5-methyl-4-isoxazole-propionic-acid) receptor (AMPAR) delivery underlies behaviors that require neural plasticity such as learning (818). We have previously shown that synaptic AMPAR trafficking plays crucial roles in the compensative cortical reorganization of the sensory cortex (15). Thus, the facilitation of experience-dependent synaptic AMPAR delivery could result in rehabilitative training-dependent motor cortical reorganization and the acceleration of motor function recovery with rehabilitation after brain damage. Collapsin-response-mediator-protein 2 (CRMP2) is a downstream molecule of semaphorin (19, 20) and is thought to be related to synaptic plasticity and learning (21, 22).

Results

Edonerpic maleate facilitates experience-dependent synaptic AMPAR delivery in the adult mice barrel cortex

Edonerpic maleate (T-817MA; 1-{3-[2-(1-benzothiophen-5-yl)ethoxy]propyl}azetidin-3-ol maleate) (Fig. 1A) protects neuronal cells and modifies their morphology (23). Edonerpic maleate has been the most characterized compound among this series of compounds. Further, phase I of a clinical trial of edonerpic maleate was successfully terminated, and the safety of edonerpic maleate was proven. However, its clinical application has not been determined. In order to explore actions of edonerpic maleate on neuronal function, we focused on the roles of edonerpic maleate in experience-dependent synaptic plasticity, which has not been studied. To examine whether edonerpic maleate affects experience-dependent synaptic AMPAR delivery, we focused on layer 4–2/3 pyramidal synapses of adult mice barrel cortex, where natural whisker experience–dependent AMPAR delivery is not observed (Fig. 1, B and C). We administrated edonerpic maleate orally, twice a day (30 mg/kg) for 3 weeks to 2-month-old adult mice. Then, we prepared acute brain slices and examined synaptic responses at layer 4–2/3 pyramidal synapses of the barrel cortex. We first recorded the ratio of evoked AMPAR- to NMDA (N-methyl-d-aspartate) receptor (NMDAR)–mediated synaptic currents (AMPA/NMDA ratio). We found an increased AMPA/NMDA ratio in edonerpic maleate–administered compared with vehicle-administered mice (Fig. 1B). This effect was whisker experience–dependent because we detected no increase of the AMPA/NMDA ratio in edonerpic maleate–administered mice in the absence of whiskers (deprived for 2 or 3 days) (Fig. 1B). There was no difference in kinetics of NMDAR-mediated currents among these groups (fig. S1). We then replaced extracellular Ca2+ with Sr2+ in order to induce asynchronous transmitter release and analyzed the quantal EPSCs (excitatory postsynaptic currents) at layer 4–2/3 pyramidal synapses. We found increased amplitude of evoked miniature EPSCs (mEPSCs) in edonerpic maleate–administered mice, compared with vehicle-administered mice, in the presence but not in the absence of whiskers (Fig. 1C). Three days of treatment with edonerpic maleate exhibited the same effect (fig. S2). Consistent with this, the induction of long-term potentiation (LTP) at layer 4–2/3 barrel cortical synapses onto pyramidal neurons was facilitated by the presence of edonerpic maleate (fig. S3).

Fig. 1 Edonerpic maleate–induced facilitation of synaptic AMPAR delivery.

(A) Chemical structure of edonerpic maleate. (B) (Left) Representative EPSCs in the barrel cortex treated with vehicle with intact whiskers, edonerpic with intact whiskers, edonerpic with whiskers deprived (WD), or vehicle with WD. (Right) Average AMPA/NMDA ratio. (C) (Left) Evoked mEPSC as in (B). (Right) Average amplitudes of evoked mEPSCs. (D) Silver-stained gel showing edonerpic-binding proteins. The black arrowhead indicates the protein selected for analysis with mass spectrometry. (E) Tandem mass spectra from tryptic digests of CRMP2. Fragment ions of the a, b, and y series identified in the tandem mass spectra from each peptide are shown. (F) Immunoblot of cortical lysate pulled down by indicated beads. (G) ITC-based measurements of edonerpic fumarate binding to CRMP2. (Top) Raw thermogram. (Bottom) Integrated titration curve. (H) (Top) Immunoblot of purified CRMP2 reacted with edonerpic. (Bottom) Monomer-to–total CRMP2 ratio. Data were normalized to control. (I) (Left) Representative EPSCs in the barrel cortex in CRMP2 knockout mice. Data of WT mice treated with vehicle with intact whiskers were derived from Fig. 1B. (Right) Average AMPA/NMDA ratio. (J) (Left) Evoked mEPSC in the barrel cortex in CRMP2 knockout mice. Data of WT mice treated with vehicle with intact whiskers were derived from Fig. 1C. (Right) Average amplitudes of evoked mEPSCs. *P < 0.05. Data were analyzed with one-way ANOVA, with Dunnett’s post hoc tests [(B), (C), (H), (I), and (J)]. The number of animals used for each experiment is indicated in the figure.

Edonerpic maleate binds to CRMP2

We prepared affinity columns of edonerpic-conjugated resin (edonerpic beads) (fig. S4A), only linker-conjugated resin (linker beads), or inactivated control resin (control beads). After the protein purification, we analyzed purified proteins and detected a band near 60 kDa specific to purified proteins by edonerpic beads (Fig. 1D). Mass spectrometry revealed that this band corresponded to CRMP2 (Fig. 1E). Immunoblotting of mice brain lysate pulled down by edonerpic beads (but not linker beads or control beads) also detected the CRMP2-specific band (Fig. 1F). To further examine the binding of edonerpic to CRMP2, we pulled down the brain lysate obtained from CRMP2-deficient mice with edonerpic beads (24). We did not observe detectable CRMP2-positive bands with CRMP2-deficient mice (fig. S4B). We performed isothermal titration calorimetry (ITC) with cell-free conditions and found that the dissociation constant (Kd) value of the edonerpic-binding to CRMP2 at ~7.35 × 10−4 M (Fig. 1G). To further investigate the effect of edonerpic on CRMP2, we mixed purified CRMP2 with edonerpic maleate in a cell-free condition and analyzed with native polyacrylamide gel electrophoresis. Edonerpic maleate significantly (P < 0.05) decreased the amount of monomeric CRMP2 (Fig. 1H), indicating that edonerpic maleate regulates multimerization of CRMP2 through direct interaction.

To examine whether CRMP2 mediates the edonerpic maleate–induced facilitation of synaptic AMPAR delivery in the adult mice barrel cortex, we used CRMP2-deficient mice (24). We administered edonerpic maleate as described above to 3-month-old CRMP2-deficient mice, measured the AMPA/NMDA ratio, and evoked mEPSC at layer 4–2/3 synapses onto pyramidal neurons of the barrel cortex. We observed a decreased AMPA/NMDA ratio and the amplitude of evoked mEPSC in edonerpic maleate–administered CRMP2-deficient mice, compared with edonerpic maleate–administered wild-type (WT) mice, in the presence of whiskers [Fig. 1, I and J (and Fig. 1, B and C)]. The AMPA/NMDA ratio and the amplitude of evoked mEPSC in edonerpic maleate– or vehicle-administered CRMP2-deficient mice was comparable with that in vehicle-administered WT mice with whiskers and edonerpic maleate–administered WT mice in the absence of whiskers [Fig. 1, I and J) (and Fig. 1, B and C)]. There was no difference in kinetics of NMDAR-mediated currents among these groups (fig. S4C). We next knocked down the expression of CRMP2 of pyramidal neurons at layer 2/3 of the adult barrel cortex by means of lentivirus-mediated in vivo gene transfer of short hairpin RNA (shRNA) targeted to CRMP2 (25), examined the AMPA/NMDA ratio, and evoked mEPSC at layer 4–2/3 pyramidal synapses in the adult barrel cortex with whole-cell recordings. CRMP2 knockdown blocked edonerpic maleate–induced increase of the AMPA/NMDA ratio (fig. S4D) and the amplitude of evoked mEPSC at synapses formed from layer 4 to layer 2/3 pyramidal neurons (fig. S4E).

Edonerpic maleate accelerates motor function recovery from injury of the motor cortex

Recovery of motor function with rehabilitation after brain damage is considered to be a training-dependent plastic event in the nervous system. Edonerpic maleate, an enhancer of experience-dependent synaptic AMPAR delivery, could accelerate the effect of rehabilitation after brain damage of motor function recovery in a training-dependent manner. To assess this effect on forelimb movements, we trained mice to reach a forelimb for food pellets. Reaching forelimb movements were analyzed by the success rate of taking the food pellets (supplementary materials, materials and methods). As the mice were trained, the success rate improved (Fig. 2A). We first investigated whether the acquisition of the reaching task required synaptic AMPAR delivery. We overexpressed green fluorescent protein (GFP)–tagged cytoplasmic portion of GluA1 (a subunit of AMPA receptors) or GFP only in layer 5 of the motor cortex and trained animals in the reaching task. This peptide (GluA1-ct) prevents synaptic GluA1 delivery (12). Expression of GFP-GluA1-ct prevented acquisition of the reaching task, whereas GFP expression did not (fig. S5).

Fig. 2 Accelerated recovery of the motor function after functional cortical reorganization.

(A) Average success rate (SR) in the reaching task. (B) (Left) Schema of cryoinjury. (Right) Hematoxylin-eosin–stained with severe cryoinjury. (C) Average SR after cryoinjury. (D) (Left) Experimental design of treatments. (Right) Average performance score (PS). (E) (Top) Experimental design for the evaluation with second lesion at the peri-injured cortex. (Bottom left) Schema of second lesion. (Bottom right) Average SR in day 28 (before second lesion) or 35 (after second lesion). (F) (Top) Representative mEPSC at layer 5 pyramidal neurons in the peri-injured region. (Bottom left) Representative mEPSC at layer 5 pyramidal neurons in the peri-injured cortex after day 56 [as in (D)]. Edonerpic with rehabilitative training or vehicle with rehabilitative training. (Bottom right) Average amplitudes of mEPSCs. (G) (Left) Experimental design for the evaluation with GFP-GluA1-ct or GFP expression by lentivirus in the peri-injured cortex. (Middle) Representative photomicrograph of the virus injection site. (Right) GFP-expressing cells in the peri-injured cortex. The dotted line represents the cryo-injured region. LV, lateral ventricle; CC, corpus callosum. (H) Average PS in mice with GFP-GluA1-c-tail or GFP expression. Data were analyzed with two-way ANOVA, followed by Bonferroni’s post hoc tests in (D) and (H), or unpaired t test in (E) and (F). *P < 0.05. The number of animals used in each experiment is indicated in the figure.

In humans, brain damage such as stroke disrupts once-acquired motor skills and leads to severe impairments. For the induction of brain injury, we used the cryoinjury method (supplementary materials, materials and methods). We trained animals in the reaching task and then induced motor cortical cryoinjury. Cryoinjury in the motor cortex of trained animals impaired their success rate (Fig. 2, B and C). After mild cortical cryoinjury, the decreased success rate in the reaching task could be recovered through training. This recovery was synaptic AMPAR delivery–dependent because expression of GFP-GluA1-ct in layer 5 of the intact motor cortex prevented recovery after training (fig. S6).

Next, we produced a more severe motor cortical cryoinjury (Fig. 2B) in trained animals. In this condition, training was not sufficient for recovery (Fig. 2C). One day after the injury, we initiated oral administration of edonerpic maleate (30 mg/kg, twice a day) or vehicle. Three weeks later, we treated mice with or without training. Training was initiated 1 hour after oral administration of edonerpic maleate, based on pharmacokinetic results, in which the maximum concentration of edonerpic was observed in the plasma and the brain in rodents at 1 hour after the oral administration (fig. S7). Concomitant training with edonerpic maleate administration dramatically recovered the impaired success rate in the reaching task. There was no obvious recovery in either edonerpic maleate–administered animals without training or vehicle-administered animals with or without training (Fig. 2D and fig. S8A). We also examined the dose-dependent effect of edonerpic maleate. One day after the injury, we started oral administration of edonerpic maleate in lower doses (20, 5, and 1 mg/kg, once a day). Three days later, we initiated training. Double-blind examination revealed that edonerpic maleate–administered mice at the dose of 20 or 5 mg/kg, but not 1 mg/kg, exhibited prominent recovery in reaching task performance in a rehabilitative training–dependent fashion (fig. S8B). We detected no significant difference (P = 0.79) of the injury size between vehicle-treated and edonerpic maleate–treated (20 mg/kg) mice during experiments (fig. S8C). Further, no behavioral abnormalities by the administration of edonerpic maleate have been observed (fig. S9). Neither SA4503 nor paroxetine, previously reported (26, 27) as potential accelerators of rehabilitation, showed effects on motor function recovery in this experimental design (fig. S10).

Edonerpic maleate drives AMPARs into synapses of peri-injured regions

We next examined whether functional cortical reorganization is accompanied by edonerpic maleate–induced recovery after cryoinjury of the motor cortex. We produced the motor cortical cryoinjury and administered edonerpic maleate (30 mg/kg, twice a day) for three weeks, as described above, and trained animals for a week when we detected the motor function recovery. Then, we introduced a second lesion at the peri-injured region (just rostral to the first-injured region) (Fig. 2E). A week later, the animals with the second lesion exhibited deterioration of once-reacquired motor function compared with sham-operated animals (Fig. 2E). This indicates that edonerpic maleate–induced motor function recovery after cryoinjury results from functional reorganization of the cortex.

We examined whether synaptic AMPAR contents are altered by edonerpic maleate in the motor cortex of injured animals with recovered motor function. We produced the motor cortical cryoinjury and administered edonerpic maleate (30 mg/kg, twice a day) or vehicle for 3 weeks. Then, we treated animals with or without training and, 6 weeks later, prepared acute brain slices and recorded from layer 5 pyramidal neurons in the above detected peri-injured region of the motor cortex, which could compensate for lost cortical function (Fig. 2F). Edonerpic maleate–administered recovered animals exhibited increased amplitude of mEPSCs, compared with vehicle-administered unrecovered mice (Fig. 2F). Consistent with the dose-dependent effects of edonerpic maleate on motor function recovery, edonerpic maleate–administered mice at 20 or 5 mg/kg (recovered), but not 1 mg/kg (nonrecovered), exhibited a prominent increase of the amplitude of mEPSC in the compensatory peri-injured region. In this experiment, 1 day after the injury, we started oral administration of edonerpic maleate at a reduced dose: 20, 5, and 1 mg/kg, once a day. Three days later, we initiated training. Four weeks after the start of training, we prepared acute brain slices and recorded from pyramidal neurons of layer 5 (fig. S11).

Next, we examined whether edonerpic maleate–induced recovery of motor function after cryoinjury of the motor cortex requires synaptic delivery of AMPARs in the cortical region, which was not primarily responsible for reaching task performance before the cryoinjury. We trained animals in the reaching task and expressed GFP-tagged GluA1-ct or GFP in the motor cortex by means of lentivirus-mediated in vivo gene transfer (Fig. 2G). Injected areas were wider than injured areas and covered potential compensatory areas detected in the previous experiments (Fig. 2E). We then introduced cryoinjury in the motor cortex as described above. One day after the injury, we started oral administration of edonerpic maleate (30 mg/kg, twice a day). Three weeks after the initiation of edonerpic maleate administration, we initiated the training. One week after the beginning of the training, we began evaluating the recovery of motor function in the reaching task. Expression of GluA1-ct prevented edonerpic maleate–induced recovery of motor function (Fig. 2H).

CRMP2 mediates edonerpic maleate–induced acceleration of motor function recovery through rehabilitative training

We found that edonerpic bound to CRMP2 (Kd = ~7.35 × 10−4 M). To examine whether edonerpic maleate–induced functional recovery is mediated by CRMP2, we produced the motor cortical cryoinjury in trained CRMP2-deficient mice in the reaching task, as described above. After the injury, we orally administered edonerpic maleate (30 mg/kg, twice a day) or vehicle. Three weeks later, we treated mice with or without training. During the training period, we evaluated the reaching task of these mice once a week as in the experiments described above. Training failed to recover reaching task performance in the mutant mice treated with edonerpic maleate (Fig. 3A and fig. S8A). Consistent with this behavioral experiment, we detected no increase in the amplitude of mEPSCs of layer 5 pyramidal neurons of intact compensatory peri-injured motor cortical region of edonerpic maleate–administration CRMP2-deficient mice 6 weeks after the beginning of the training (Fig. 3B).

Fig. 3 CRMP2 mediates edonerpic–induced functional recovery via ADF/cofilin activation.

(A) Average performance score. WT mice data were derived from Fig. 2D. CRMP2 knockout data were added. (B) (Left) Representative mEPSC at layer 5 pyramidal neurons in the peri-injured region of CRMP2 knockout mice (Fig. 2F). Edonerpic + rehabilitative training (Tr.)/knockout or vehicle + Tr. / knockout. (Right) Average amplitudes of mEPSCs. (C) (Left) Immunoblots of Cofilin, phosphorylated Cofilin, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) obtained from cLTP-induced cortical slices of WT or CRMP2 knockout mice. (Right) The level of p-Cofilin. GAPDH was used as the reference. The data were normalized to the vehicle-treated group. (D) (Left) Representative EPSCs from Venus-expressing neurons or Cofilin S3E in the barrel cortex of mice administered with edonerpic or vehicle. (Right) Average AMPA/NMDA ratio. (E) (Left) Evoked mEPSC as in (D). (Right) Average amplitudes of evoked mEPSCs. (F) (Left) Immunoblots of Cofilin, p-Cofilin, and GAPDH from peri-injured cortical region of mice at day 28 (Fig. 2D). (Middle) Phosphorylation level of Cofilin. GAPDH was used as the reference. The data were normalized to the vehicle-treated group. (Right) Phosphorylation level of Cofilin in the peri-injured regions of edonerpic maleate–administered mice (comparison between with and without training). GAPDH was used as the reference. Data were analyzed with two-way ANOVA, followed by Bonferroni’s post hoc tests [(A); edonerpic + Tr. / WT versus other groups in day 35, 42, and 49/56], or unpaired t test [(B), (C), (D), (E), and (F)]. *P < 0.05. n.s. indicates not significant. The number of animals used in each experiment is indicated in the figure.

Because abnormal neurological conditions could modulate CRMP2 (20, 28), we investigated the phosphorylation status of CRMP2 in the compensatory peri-injured region from edonerpic maleate–treated injured mice that recovered with rehabilitative training. We detected decreased amount of phosphorylated CRMP2 in edoneripic maleate–treated recovered mice than vehicle-treated unrecovered mice with training (fig. S12A).

CRMP2 mediates edonerpic maleate–induced activation of ADF/cofilin

To further elucidate the mechanisms of how edonerpic-CRMP2 interaction modifies synaptic function under the plasticity-inducing condition (29), we performed biochemical studies with chemical LTP-induced cortical slices. We prepared cortical slices of the motor cortex and chemically induced LTP (cLTP) by briefly exposing the slices to the potassium channel blocker tetraethylammonium (TEA) (30). The actin-depolymerizing factor (ADF)/cofilin mediates AMPAR trafficking during synaptic plasticity (3032); thus, we focused on this molecule as a potential downstream effector of edonerpic-CRMP2 interaction. ADF/cofilin is inactivated by phosphorylation and activated by dephosphorylation at the serine-3 (Ser3) residue. To determine whether edonerpic-maleate activates ADF/cofilin in slices of the motor cortex under cLTP, we prepared the synaptoneurosome fraction from cLTP-treated acute motor cortical slices. We detected decreased phosphorylation of ADF/cofilin at Ser3 in the edonerpic maleate–treated slices, compared with vehicle-treated ones, under cLTP but not under the normal condition, suggesting that ADF/cofilin is activated by edonerpic maleate under cLTP (Fig. 3C and fig. S12B). Edonerpic maleate–induced activation of ADF/cofilin under cLTP was abolished in slices from CRMP2-deficient mice (Fig. 3C and fig. S12B). Further, the phosphorylation of ADF/cofilin at Ser3 was decreased in slices of WT mice than CRMP2-deficient mice in the presence of edonerpic maleate (fig. S12C). These results indicate that edonerpic maleate–induced activation of ADF/cofilin under cLTP is mediated by CRMP2.

To test whether the activation of ADF/cofilin mediates edonerpic-CRMP2–induced facilitation of synaptic AMPAR delivery, we overexpressed the dominant negative form of ADF/cofilin (S3E) together with Venus or Venus alone in the adult barrel cortex of edonerpic maleate–administered or vehicle-treated mice with lentivirus. Three weeks after the initiation of edonerpic maleate administration, we prepared acute brain slices and examined the AMPA/NMDA ratio and evoked mEPSC at layer 4–2/3 pyramidal synapses. We detected decreased AMPA/NMDA ratio and the amplitude of evoked mEPSC by the expression of ADF/cofilin S3E, compared with the expression of Venus alone, in edonerpic maleate–administered animals (Fig. 3, D and E). We detected no significant (P > 0.20) decrease of AMPA/NMDA ratio and the amplitude of mEPSC at layer 4–2/3 pyramidal synapses by the expression of ADF/cofilin S3E compared with the expression of Venus alone in vehicle-treated animals (Fig. 3, D and E).

Consistent with this, we found activation of ADF/cofilin (decreased phosphorylation of ADF/cofilin at Ser3) in the synaptoneurosome fraction obtained from the compensatory peri-injured region of edonerpic maleate–administered recovered mice compared with vehicle-treated nonrecovered mice (Fig. 3F). We also found significant (P < 0.01) decrease of the phosphorylation levels of ADF/cofilin in edonerpic maleate–treated animals with training than in those without training (Fig. 3F).

Edonerpic maleate facilitates motor function recovery after ICH in nonhuman primates

Stroke such as hemorrhage and embolism in the internal capsule leads to severe paralysis of motor functions. The severity and outcome of motor impairments depend on the degree of damage to this region (3336). To further show that edonerpic maleate facilitates training-dependent recovery from brain damage, we used an internal capsule hemorrhage (ICH) model in nonhuman primates. We trained macaque monkeys in two different tasks. A simple reach-to-grasp task aimed at evaluating the performance in both reaching and gross grasping (Fig. 4A). In nonhuman primates and humans, development of the corticospinal tract correlates with improvement in the index of dexterity, particularly in the ability to perform precision grip, holding a small object between the thumb and index finger tips (37). In the vertical-slit task, the performance of dexterous hand movements, typical of primates, were evaluated (Fig. 4B).

Fig. 4 Edonerpic maleate accelerates motor function recovery in nonhuman primates.

(A) (Left) Simple reach-to-grasp task. (Right) Representative reaching and grasping for each location. (B) (Left middle) The vertical slit task. (Right) The finger-thumb grip in the vertical slit task before collagenase injection. (C) The fluid-attenuated inversion recovery images of MRI scanning (monkey D) 3 days after the collagenase injection into the right internal capsule. (D) Stroke volume of edonerpic maleate– or vehicle-administered monkeys. Each symbol indicates data from one monkey. There was no significant difference between the two groups (Mann-Whitney U-test, P = 0.90). (E) Experimental design. (F and G) Time course of the PS in the simple reach-to-grasp task. (H and I) Average PS in the simple reach-to-grasp in the early (days 2 to 11) and in the late (days 30 to 39) phase of the training period. (J) Time course of the PS for successful retrievals in the vertical slit task. (K) The average PS in the vertical slit task in the early and in the late phase of the training period. (L) The average PS for time to retrievals in the late phase of the training period of the vertical slit task. (M) Sequential captures of the late phase (day 32) of the training period. Edonerpic maleate–administered monkeys could perform the task smoothly without dropping the piece of apple (black arrowheads indicate the apple’s positions.), whereas the vehicle-administered monkeys could not. Data were analyzed with two-way ANOVA [edonerpic-maleate, P < 0.0001; (F), (G), and (J)] and Mann-Whitney U-test [(H), (I), (K), and (L)]. *P < 0.05.

After monkeys learned to perform the tasks, we injected collagenase to the hemisphere contralateral to the preferred hand, under magnetic resonance imaging (MRI)–stereotaxic guidance. MRI scanning confirmed hemorrhage in the internal capsule (Fig. 4C). The area of the hyperintense signal expanded at days 3 to 7 after injection and then decreased (fig. S13A). Thereafter, the residual hypointense area was almost stable until the end of the experiment, 6 months after injection (fig. S13B). The lesion in the internal capsule was also histologically confirmed after the behavioral experiment ended (fig. S13C). Although there was a tendency that the decrease of lesion area in edonerpic maleate–treated animals during experiment is smaller than vehicle-treated monkeys, we did not detect statistical significance (fig. S13D).

Before ICH, all animals smoothly performed both tasks, and they used precision grip in the vertical-slit task (movie S1). Immediately after ICH, flaccid paralysis of the contralateral forelimb, almost complete paralysis of the hand digits, and incomplete but severe paralysis of the wrist, elbow, and shoulder were observed. Forelimb motor functions then gradually recovered. Some of the monkeys showed mild paralysis of the contralateral hindlimb immediately after ICH, but the paralysis disappeared within a few days. Although the average stroke volume of edonerpic maleate–administered monkeys was higher than that of vehicle-administered monkeys, the difference between the two groups was not statistically significant (P = 0.90, Mann-Whitney U test) (Fig. 4D). All monkeys were able to move the elbow and shoulder joints 1 to 2 weeks after ICH. On the day when the monkeys first reached for the piece of apple, presented in each task, rehabilitative training began (Fig. 4E and Table 1). The rehabilitative training was initiated 15 min after the administration of edonerpic maleate, based on the results of the pharmacokinetic study (fig. S14).

Table 1 Monkeys used in the present study.

Three monkeys were used in each treatment. There are no significant differences (P = 0.2 for weight, P = 0.6 for days of first reach) in body weight, age, or days of first reach.

View this table:

During rehabilitative training, the performance of the simple reach-to-grasp task gradually recovered in both the edonerpic maleate– and vehicle-administered monkeys (Fig. 4, F, G, H, and I; and figs. S15 and S16). Two-way analysis of variance (ANOVA) revealed significant (P < 0.001) effects of edonerpic maleate administration on the recovery of time to retrievals, compared with the vehicle-administered group, for both near and far locations (Fig. 4, F and G, and movies S2 and S3). Although the effect of edonerpic maleate was significant (P < 0.005) for both locations in the early phase, the effect was significant (P < 0.001) only for far location in the late phase of the rehabilitative training period (Fig. 4, H and I). Retrieval from the far location required more coordination of forelimb muscles, including the proximal and distal parts, as compared with the near location. In the vertical slit task, a significant (P < 0.001) effect of edonerpic maleate administration was found for recovery of both the success rate and the time to retrieval, throughout the rehabilitative training period (Fig. 4, J, K, and L; and figs. S15 and S16). The edonerpic maleate–administered monkeys frequently showed dexterous hand movements, including precision grip after the rehabilitative training period (Fig. 4M and movie S4). On the other hand, precision grip was rarely observed in the vehicle-administered monkeys after ICH (Fig. 4M and movie S4).

Discussion

Although medication during the acute phase of brain damage exhibits some effectiveness, there is no small compound–mediated intervention to enhance the effect of later rehabilitation after functional loss due to brain damage (38). Previous experiments have shown that motor cortical reorganization in the intact regions of a damaged brain is crucial for functional recovery (4, 6, 7, 29, 39, 40). Here, we found that edonerpic maleate, a CRMP2-binding compound, accelerates functional recovery after brain damage in a rehabilitative training-dependent manner, which induces functional motor cortical reorganization. Thus, edonerpic maleate could provide a pharmacological solution for unmet medical needs. Although many compounds exhibit some effectiveness on motor function recovery in rodents, most fail to prove their efficacy in primates. In this study, we proved the prominent effect of edonerpic maleate on training-dependent motor function recovery in primates. Thus, edonerpic maleate may be a strong candidate for a small compound to accelerate rehabilitative training–dependent motor function recovery after brain damage, such as stroke, in humans.

CRMP2 can bind to actin, and its regulator proteins, which is crucial for synaptic AMPAR delivery (20, 31, 41, 42). Among them, the activation of ADF/cofilin drives trafficking of AMPAR into the spine surface under plasticity-inducing conditions (31). The activation of ADF/cofilin is involved in synaptic AMPAR trafficking in various genetic and environmental conditions (30, 32). We detected CRMP2-dependent activation of ADF/cofilin by edonerpic maleate in the plasticity-inducing condition. We also found that edonerpic CRMP2–induced activation of ADF/cofilin mediates the facilitation of synaptic AMPAR trafficking. CRMP2 is required for the trafficking of N-type voltage-sensitive Ca2+ channels (43). Thus, edonerpic-CRMP2 interaction could facilitate synaptic AMPAR delivery through the regulation of actin dynamics. In addition to actin dynamics, neuromodulatory systems such as dopaminergic and serotonergic inputs could regulate synaptic AMPAR trafficking (15, 44). It will be crucial to examine whether edonerpic-CRMP2 complex affects dopaminergic and serotonergic signaling.

Although the decrease of lesion area in edonerpic maleate–treated monkeys during experiment tended to be smaller than that in vehicle-treated monkeys, we did not detect statistical significance (fig. S13D). We also did not find any difference of the change of lesion volume during experiments between edonerpic maleate–treated and vehicle-treated mice (fig. S8C). Further, the recovery of edonerpic maleate–treated mice with cryoinjury was blocked by the expression of GluA1-ct in the compensatory cortical area (Fig. 2H). Taken together, edonerpic maleate–mediated facilitation of motor function recovery after brain injury is primarily mediated by the augmented synaptic AMPA receptor delivery. It will also be crucial to further study whether edonerpic maleate promotes neuroregeneration.

Although the affinity of edonerpic binding to CRMP2 was moderate (Kd = ~7.35 × 10−4 M), we found that (i) a major band specific to the sample that was pulled down with edonerpic-conjugated beads from the lysate of mice cortical primary culture corresponded to CRMP2 (there existed a faint band near CRMP2-positive bands with the edonerpic beadspulled-down preparation from CRMP2-deficient mice; this could be due to other isoforms of CRMP family), (ii) edonerpic maleate decreased the monomer of CRMP2 in the cell-free condition, (iii) edonerpic maleate–induced activation of ADF/cofilin in the plasticity-induced condition was blocked in the absence of CRMP2, (iv) edonerpic maleate–induced facilitation of synaptic AMPAR delivery was abolished in CRMP2-deficient mice, (v) edonerpic maleate–induced facilitation of synaptic AMPAR delivery was blocked by knocking down the expression of CRMP2 with shRNA, and (vi) edonerpic maleate–induced acceleration of motor function recovery was prevented in CRMP2-deficient mice. CRMP2 is an intracellular protein, and it is difficult to estimate the concentration of edonerpic maleate in the cortical neurons of treated animals. However, the evidence presented here suggests that CRMP2 is a primary target of edonerpic maleate for the rehabilitative training–dependent acceleration of motor function recovery from traumatic brain injury.

The efficacy of edonerpic maleate in humans should be evaluated in clinical trials because safety profiles of this compound have already been well established in clinical phase I studies. For stroke recovery, engineering technologies for rehabilitation, such as brain machine interface and robotics, are expected to be promising tools (45, 46). Other biological technologies, such as cell transplantation, may also be potential therapeutic alternatives, with distinct mechanisms from edonerpic maleate (47). Thus, the combination of these tools with the application of edonerpic maleate could induce synergistic effects and greatly increase the number of treatable patients with pathological brain damage.

Supplementary Materials

www.sciencemag.org/content/360/6384/50/suppl/DC1

Materials and Methods

Figs. S1 to S16

References (4867)

Movies S1 to S4

References and Notes

Acknowledgments: We thank Y. Kimura and T. Akiyama (Advanced Medical Research Center, Yokohama City University) for technical assistance in analysis of mass spectrometry. We thank M. Taguri (Department of Biostatistics, Yokohama City University) for useful advice in statistical analysis. We also thank Y. Katakai for polite assistance in nonhuman primate experience. Funding: This project was supported by Special Coordination Funds for Promoting Science and Technology (T.T.) and partially supported by the Strategic Research Program for Brain Sciences from Japan Agency for Medical Research and Development (AMED) (T.T.) and the Brain Mapping by Integrated Neurotechnologies for Disease Studies (Brain/MINDS) from AMED (T.T.). This project is also partially supported by Grants-in-Aid for Scientific Research in a Priority Area (grant 17082006) and Creation of Innovation Centers for Advanced Interdisciplinary Research Areas Program in the Project for Developing Innovation Systems (grant 42890001) from the Ministry of Education, Science, Sports and Culture (Y.G.). Author contributions: T.T. designed the project and experiments, interpreted and analyzed data, and wrote the manuscript. H.A. conducted behavioral experiments (mice and monkey) and biochemical experiments and wrote the manuscript. S.J. conducted electrophysiological and biochemical experiments and wrote the mansucript. W.N. and Y.M. conducted behavioral and histological experiments of monkey. A.J.-T. conducted electrophysiological experiments. Y.K. and N.M. conducted behavioral experiments (mice). H.T., K.S., H.M., and T.K. conducted biochemical experiments. A.S. conducted viral preparation. T.O., Y.G., and N.H. contributed to interpret the data. We shared the data with all authors who contributed to the analysis and interpretation of the data and to the design of the experiments. Competing interests: The authors of this publication include employees of Toyama Chemical Co., which holds intellectual property rights of edonerpic maleate (T-817MA) and funded Yokohama City University for the study using macaque monkeys. T.T. is one of the inventors of a patent application claiming the use of edonerpic maleate to enhance the functional recovery after brain damage [substance patent (WO03/035647, Alkyl ether derivatives or salts thereof) and use patent (WO2015/115582, Post nerve injury rehabilitation effect-enhancing agent comprising alkyl ether derivative or salt thereof)]. All the other authors declare that they have no competing interests. Data and materials availability: Edonerpic maleate is a compound under the clinical development stage. Thus, research use of edonerpic maleate requires the approval from Toyama Chemical Co. and Fujifilm Corporation. There is no fixed format of those requests. All data are available in the manuscript or the supplementary materials. Original data are kept on local hard drives (with backups). Processed data for this study are additionally kept and curated on local hard drives of the Yokohama City University (http://neurosci.med.yokohama-cu.ac.jp/aao2300_processed_data.html) and servers of Toyama Chemical Co.
View Abstract

Stay Connected to Science

Subjects

Navigate This Article