Functional recovery after cerebellar damage is related to GAP-43-mediated reactive responses of pre-cerebellar and deep cerebellar nuclei

Experimental Neurology 233 (2012) 273–282

Contents lists available at SciVerse ScienceDirect

Experimental Neurology journal homepage: www.elsevier.com/locate/yexnr

Functional recovery after cerebellar damage is related to GAP-43-mediated reactive responses of pre-cerebellar and deep cerebellar nuclei Lorena Burello a, b, Paola De Bartolo a, b, Francesca Gelfo a, Francesca Foti a, c, Francesco Angelucci a, Laura Petrosini a, b,⁎ a b c

Fondazione Santa Lucia, Via Ardeatina 306, 00179, Rome, Italy Dipartimento di Psicologia, Università “Sapienza” di Roma, Via dei Marsi 78, 00185, Rome, Italy Dipartimento di Psicologia dei Processi di Sviluppo e Socializzazione, Università "Sapienza" di Roma, Via dei Marsi 78, 00185, Rome, Italy

a r t i c l e

i n f o

Article history: Received 25 July 2011 Revised 28 September 2011 Accepted 18 October 2011 Available online 28 October 2011 Keywords: GAP-43 Cerebellar lesion Motor symptomatology Pre-cerebellar nuclei Deep cerebellar nuclei Lesion-induced plasticity Retrograde death Post-lesional rearrangement

a b s t r a c t Since brain injuries in adulthood are a leading cause of long-term disabilities, the development of rehabilitative strategies able to impact on functional outcomes requires detailing adaptive neurobiological responses. Functional recovery following brain insult is mainly ascribed to brain neuroplastic properties although the close linkage between neuronal plasticity and functional recovery is not yet fully clarified. The present study analyzed the reactive responses of pre-cerebellar (inferior olive, lateral reticular nucleus and pontine nuclei) and deep cerebellar nuclei after a hemicerebellectomy, considering the great plastic potential of the cerebellar system in physiological and pathological conditions. The time course of the plastic reorganization following cerebellar lesion was investigated by monitoring the Growth Associated Protein-43 (GAP-43) immunoreactivity. The time course of recovery from cerebellar symptoms was also assessed to parallel behavioral and neurobiological parameters. A key role of GAP-43 in neuronal reactive responses was evidenced. Neurons that underwent an axotomy as consequence of the right hemicerebellectomy (neurons of left inferior olive, right lateral reticular nucleus and left pontine nuclei) exhibited enhanced GAP-43 immunoreactivity and cell death. As for the not-axotomized neurons, we found enhanced GAP-43 immunoreactivity only in right pontine nuclei projecting to the spared (left) hemicerebellum. GAP-43 levels augmented also in the three deep cerebellar nuclei of the spared hemicerebellum, indicating the ponto-cerebellar circuit as crucially involved in functional recovery. Interestingly, each nucleus showed a distinct time course in GAP-43 immunoreactivity. GAP-43 levels peaked during the first post-operative week in the fastigial and interposed nuclei and after one month in the dentate nucleus. These results suggest that the earlier plastic events of the fastigial and interposed nuclei were driving compensation of the elementary features of posture and locomotion, while the later plastic events of the dentate nucleus were mediating the recovered ability to flexibly adjust the locomotor plan. © 2011 Elsevier Inc. All rights reserved.

Introduction In light of the amount of adult subjects affected by brain injuries provoking severe sensory, motor or cognitive impairments, constant efforts are carried out to detail brain reactions to damage and hence to develop the most effective treatments. It is quite established that unlike the developing brain that has striking regenerative properties, the mammalian adult brain displays a micro-environmental composition that inhibits axon regeneration (Properzi and Fawcett, 2004). Nevertheless, the multiple possibilities for functional recovery following brain insult in adulthood highlight that some adaptive neurobiological responses take place even when CNS ontogeny is over. Neuroplasticity properties are considered the driving force of the ⁎ Corresponding author at: Dipartimento di Psicologia, University “Sapienza” of Rome, Via dei Marsi 78, 00185 Rome, Italy. Fax: + 39 0649917522. E-mail address: [email protected] (L. Petrosini). 0014-4886/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.expneurol.2011.10.016

functional recovery occurring either spontaneously and during rehabilitation (Duffau, 2006; Nudo et al., 2001; Wieloch and Nikolich, 2006). Thus, a thorough knowledge of the neuroplastic reactions to brain insult is need to elaborate rationally guided rehabilitative interventions able to steer behavioral improvements. The cerebellar system may represent a reliable structure to gain insight into post-lesional rearrangement that gives rise to functional recovery. In fact, cerebellar circuits are endowed with great plastic potentialities supporting a variety of use-dependent learning mechanisms in physiological conditions (Hansel et al., 2001). Moreover, axons of neurons belonging to deep cerebellar and pre-cerebellar nuclei express plastic potentials so large that they are able to regenerate when axotomized, provided a permissive environment is met (Buffo et al., 1998; Chaisuksunt et al., 2000; Dusart et al., 2005; Wehrlé et al., 2001). In this respect, the compensation that follows a cerebellar injury may represent an excellent model of lesioninduced plasticity. In fact, most of the motor symptoms elicited by

274

L. Burello et al. / Experimental Neurology 233 (2012) 273–282

cerebellar damage gradually and efficiently compensate over time, as observed in patients as well as in animals following surgical ablation or stroke injury of the cerebellar structures (Holmes, 1917; Luciani, 1891). Namely, the development of the compensated state in both humans and animals is characterized by a gradual decrease in the severity of static symptoms, including eye and head nystagmus and head and body tilt, while the dynamic symptoms, including complex and coordinated behaviors, compensate less consistently over a longer time course and to a lesser extent. Although the cellular mechanisms are still a matter of debate, it is believed that the spontaneous recovery from postural and motor disturbances of cerebellar origin as well as the maintenance of the compensated state is due to profound synaptic and circuital modifications in many brain regions involved in postural and motor behaviors. The present research was aimed at investigating the neurobiological substrate of the spontaneous recovery following a hemicerebellectomy (HCb) performed in adult rats, focusing in particular the post-lesional responses of some pre-cerebellar structures, namely the inferior olive, lateral reticular nucleus and pontine nuclei. Reactive changes in the deep cerebellar nuclei of the spared hemicerebellum were also examined. Levels of the Growth Associated Protein-43 (GAP-43) were used to monitor plastic post-lesional reorganization. Indeed, GAP-43 is a useful marker of axonal growth during development as well as in axonal remodeling and regeneration in the adult. This is a neuron specific, calmodulin-binding phosphoprotein and a major substrate of protein kinase C (PKC) (Aigner et al., 1995; Schaechter and Benowitz, 1993). It is produced at high levels in every nerve cell during neurite outgrowth and early stages of synaptogenesis (Gispen et al., 1991; Skene, 1989; Skene and Willard, 1981) and represents a main constituent of the axonal growth cone (de Graan et al., 1985). Its importance for developing nerve terminal is demonstrated by the observation that mice bearing a deletion of GAP-43 gene show defect in axonal pathfinding and most of them die shortly after birth (Strittmatter et al., 1995). Moreover, it has been reported that axonal growth cones depleted in GAP-43 show deficiency in persistent spreading, branching and adhesion (Aigner and Caroni, 1995). When developing neurons reach appropriate synaptic target, while GAP-43 expression is downregulated in most of them (Skene, 1989), high levels of GAP-43 persist in restricted brain regions, as hippocampus and associative cortices that continue to exhibit activity-dependent plasticity (Benowitz and Routtenberg, 1987, 1997). Indeed, it has been demonstrated that GAP43 is critically involved in long-term potentiation and memory trace formation (Routtenberg et al., 2000). In addition, up-regulation of this protein occurs in several adult neuronal populations after axotomy even though GAP-43 can promote not only axonal re-growth but also degenerative phenomena, depending on environmental cues (Dusart et al., 2005; Gagliardini et al., 2000; Wehrlé et al., 2001). By using a densitometric methodology, GAP-43 immunoreactivity (IR) was assessed at short, middle and long post-operative time-points after HCb to detect the relationship between GAP-43 IR and recovery time course. Considering that neurons of pre-cerebellar structures were unilaterally axotomized following HCb and that retrograde neuronal death could affect them, to elucidate the meaning of GAP-43 IR also the presence of neuronal death was assessed by cell counting. Importantly, to relate functional recovery with plastic neuro-anatomical rearrangement, the time course of the compensation of postural and locomotor symptoms as well as of complex coordinated abilities following the HCb was analyzed. Material and methods Subjects and experimental groups In the present research 34 adult male Wistar rats (7–8 weeks of age, 200–250 g; Harlan Laboratories, Udine, Italy) paired housed in standard cages were used. They were maintained and handled in

accordance with European Communities Council Directive (2010/ 63/EU revising Directive 86/609/EEC). Animals were randomly assigned to the different experimental groups. In details, a group of 7 animals was hemicerebellectomized (HCbed) to evaluate the time course of recovery from postural and motor symptoms evoked by a right HCb (Group name: H). At the end of postural evaluation (42nd post-operative day) these animals were sacrificed. Out of the 27 remaining animals, 21 animals were submitted to a right HCb to evaluate neuro-anatomical plastic rearrangements and 6 animals were used as controls. The 21 HCbed animals were sacrificed at the 1st (Group name: H1), 3rd (H3), 6th (H6), 14th (H14), 21st (H21), 30th (H30) and 36th (H36) post-operative day (pod) (n= 3/pod). The six controls were sham-operated animals sacrificed at sampled intervals as the 1st, 14th and 36th post-operative day (n=2/pod). Data homogeneity among sham-operated animals was verified by comparing the findings obtained in these six shamoperated rats on the nine neuro-anatomical parameters considered (see in the following) by means of one-way MANOVA. This analysis revealed no significant differences among sham-operated animals (Pillai's Trace = 0.36, F18, 28 = 0.35 p = n.s.). Then, we randomly selected 3 of the 6 animals forming a group referred to as Control group (C). Thus, the experimental design planned to monitor neuroplastic responses to the cerebellar injury was comprised of 8 experimental groups, each of which included 3 animals. Surgical procedure The animals were deeply anesthetized with an i.p. injection of a Zoletil 100 (Tiletamine and Zolazepam: 50 mg/kg — Virbac s.r.l., Milan, Italy) and Rompun (Xylazine: 10 mg/Kg — Bayer s.p.a., Milan, Italy) solution. A craniotomy was performed over the right hemicerebellum. In rats belonging to HCbed groups the dura was excised and the right cerebellar hemisphere and hemivermis with the fastigial, interpositus and dentate nuclei were ablated by suction. Care was taken to avoid any lesion of the extracerebellar structures. At the end of suction procedure the cavity was filled with sterile haemostatic absorbable gelatin sponge (Spongostan, Johnson & Johnson Medical Ltd., USA). In rats belonging to the sham-operated group only the craniotomy was performed without neither excision of meningeal membranes nor cerebellar ablation. In all animals, the wound edges were then sutured and the animals were allowed to recover from anesthesia and surgical stress on a thermal blanket before being replaced in their cages. Neurological assessment of cerebellar symptomatology Postural asymmetries and motor behavior were assessed by means of a behavioral rating scale (Table 1) in all HCbed and sham-operated rats at variable time intervals starting from 24 h after cerebellar lesion or sham-surgery (Centonze et al., 2008; Cutuli et al., 2011; Federico et al., 2006; Foti et al., 2011; Gelfo et al., 2011). Video records were taken throughout the entire testing cycle and were used to supplement direct behavioral observations. A trained experimenter blind to the surgical procedure (HCb or sham-surgery) the rats were submitted to assigned a score from 0 to 2 to each symptom according to its degree of severity (Table 1). The thresholds to distinguish the severity of the cerebellar symptoms are described in details in the Supplementary data. Symptom lateralization was taken into account to verify the correctness of cerebellar lesion, because a right HCb evokes most symptoms ipsilaterally to the lesion side. Behavioral evaluations were performed in the stabulation room and repeated each day from 1 to 42 days after the HCb. The duration of the testing session was slightly variable based on the performance of the animal and could last till 15 min for each animal. As 22 symptoms were considered, the total score ranged from 0 (complete absence of any deficits) to 44

L. Burello et al. / Experimental Neurology 233 (2012) 273–282

275

Table 1 Behavioral rating scale for postural and motor symptoms following right hemicerebellectomy. Score

Head posture

0 1 2 Score

Eye symptoms

Exploratory behavior

Head tilt (R)

Oscillations (bobbing)

Nystagmus (R)

Rearing

Absent Slight Marked

Absent Occasionally present Repeatedly present

Absent b 20 beats/min >20 beats/min

Repeatedly present Occasionally present Absent

Trunk and limb posture Body tilt (R)

Fore-/Hind-limb hyperflexion (R)

Fore-/Hind-limb hyperextension (L)

Ankle extrarotation (R)

Hypotonia (R)

Side Falls (R)

Muscle asthenia (holding on a wire)

0 1

Absent Slight

Absent Slight

Absent Slight

Absent Slight

Absent Slight

>10 s. b10 s.

2

Marked

Marked

Marked

Marked

Marked

Absent Occasionally present Repeatedly present

Score

0 1 2

Absent

Locomotion Wide base

Collapse on the belly

Steering (R)

Circling (R)

Pivoting (R)

Tremor

Absent Slight tendency Markedly present

Absent Slight tendency Markedly present

Absent Occasionally present Compulsively present

Absent Occasionally present Compulsively present

Absent Rarely present Present

Absent Slight Marked

Score

0 1 2

Complex motor behaviors Ascending a ladder

Descending a ladder

Vestibular drop

Successful Only a few steps Failed

Successful Only a few steps Failed

Present without directionality Present with side prevalence Absent

R = to the right side;L = to the left side.

(presence of all symptoms to the highest degree). Since the behavioral rating scale evaluated symptoms and asymmetries, sham-operated animals obtained always scores of 0. This kind of evaluation prevented the inclusion of the scores obtained by sham-operated animals in the ANOVAs. Tissue processing After the settled post-operative days, animals were deeply anesthetized and transcardially perfused with saline followed by 4% paraformaldehyde in PBS (4°, pH 7.5). Brains were removed and cryoprotected in 30% buffered sucrose and frozen to be cut on a freezing microtome. The posterior part of the brains embracing cerebellum and brainstem was cut into 30-μm coronal sections. Sections were collected at the level of cerebellum (− 8.80/− 14.60 mm from bregma); lateral reticular nucleus (− 13.24/− 14.60 mm from bregma); inferior olive (− 11.80/− 14.60 mm from bregma); pontine nuclei (− 6.72/− 8.00 mm from bregma). Namely, two series of adjacent sections were collected for thionin staining and GAP-43 immunohistochemistry. GAP-43 immunohistochemical staining To visualize GAP-43 immunopositive regions, sections were successively pre-incubated in: — 0.05% H2O2 (Sigma-Aldrich Chemie, Steinheim, Germany) in PBS for 45 min; — 1% milk powder (SigmaAldrich Chemie, Steinheim, Germany) in PBS for 30 min; — 0.05% Triton X-100 (Sigma-Aldrich Chemie, Steinheim, Germany) and 0.5% Bovine Serum Albumine (Sigma-Aldrich Chemie, Steinheim, Germany) in PBS for 30 min; — 0.05% Triton X-100, 0.5% Bovine serum Albumine and 0.005% Normal Horse Serum (Vector Laboratories, Burlingame, CA) in PBS for 1 h. Pre-incubation was performed at room temperature. Sections then received a 36 hours incubation at 4 °C in a solution composed of 0.05% Triton X-100, 0.5% Bovine

Serum Albumine, 0.005% Normal Horse Serum and the primary anti-GAP-43 antibody (monoclonal, 1:2000; Sigma-Aldrich Chemie, Steinheim, Germany) in PBS. Later, sections were incubated in 0.05% Triton X-100, 0.75% Bovine Serum Albumine, 0.005% Normal Horse serum and the secondary biotinylated anti-mouse antibody (made in horse, 1:150; Vector Laboratories, Burlingame, CA). Staining was performed according to the avidin-biotin-peroxidase method (Vectastain, ABC Elite kit, Vector Laboratories, Burlingame, CA) and revealed using a solution of 0.05% 3,3′ diaminobenzidine, 0.1% H2O2 and 0.3% nickel ammoniun sulfate in Tris buffer. To exclude artifacts, some random sections for each case were processed as described except for the elimination of the primary antibody. Stained sections were mounted on slides, air-dried, dehydrated and coverslipped. Verification of cerebellar lesion In all HCbed animals the extent of cerebellar lesion was determined in thionin stained sections (Fig. 1). In all animals included in the present research, the left side of the cerebellum and all extracerebellar structures were spared, except for the dorsal cap of Deiters' nuclei, which in some cases were slightly affected. Variability in the extent of lesion in the vermian and floccular regions was not taken into account, since in all cases the ablation of cerebellar peduncles and deep nuclei of the right side caused the functional disconnection of the residual vermian and floccular tissue. To control variability of cerebellar lesions among animals, thionin stained sections at representative levels throughout the cerebellum (− 10.04; −11.60; −13.24 mm from the bregma) were selected for each animals. For each level, lesion boundaries were drawn on schematic drawings of corresponding atlas tables (Paxinos and Watson, 1998) and the ablated areas computed by the software ImageJ (version 1.42q, NIH) (Abramoff et al., 2004) to obtain the mean percentage of ablation (Fig. 1). The overall percentage of ablation across the three neuroanatomical levels was then determined.

276

L. Burello et al. / Experimental Neurology 233 (2012) 273–282

In deep cerebellar nuclei of the spared (left) hemicerebellum, the imbalance in GAP-43 IR between sides was obviously not computable. Thus, to avoid the influence of staining variability, another correction method was applied. Since GAP-43 is not localized in white matter (Kapfhammer and Schwab, 1994a,b), the staining background was determined for each section by measuring 4 mean gray samples from myelinated fibers. Mean background value was then subtracted from the mean gray values measured for each nucleus in the same section. Quantification of neuronal death

Fig. 1. Cerebellar lesion. In A, a thionine-stained coronal section through cerebellum and brain stem in a HCbed rat. Note the total absence of the right hemicerebellum and the sparing of any extracerebellar structure. Scale bar: 2 mm. In B, a schematic representation of the connections between cerebellum (CCx: Cerebellar Cortex; DCn: Deep Cerebellar nuclei) and pre-cerebellar nuclei (IO: Inferior Olive, solid gray line; LRn: Lateral reticular nucleus, solid black line; Pn: Pontine nuclei, dashed black line). The dotted gray lines represent the ablated right hemicerebellum and the axotomized tracts. In C, schematic reconstructions of the cerebellum and brain stem at − 10.04, − 11.60 and − 3.24 mm from the bregma, illustrating minimal (dark gray) and maximal (light gray) extension of the hemicerebellar lesion in all HCbed animals.

Quantification of GAP-43 IR In C, H1, H3, H6, H14, H21, H30 and H36 groups, GAP-43 IR was analyzed by means of densitometry. Photomicrographs of histological sections were acquired using an optical microscope (Zeiss AxioLab; magnification: 40 ×) fitted with a camera (Optronics DEI-750). Quantification was made using the software ImageJ (Abramoff et al., 2004). Photomicrographs were converted in 8-bit images in which pixel gray values could vary from 0 (black) to 255 (white). For each image a measure of pixel mean gray was performed in a part of the slide containing no tissue to quantify light intensity. This value was then adjusted to a predetermined reference value in order to obtain a standardized light intensity among images (Bremner et al., 1994). After calibrating the illumination, mean gray in sample areas of the selected regions was measured. Mean gray values were converted in Optical Density (OD) values according to the function OD= log10 (255/pixel value) implemented by ImageJ (Abramoff et al., 2004). Analyses were performed on 7 sections for the inferior olive, 4 sections for both the lateral reticular nucleus and the pontine nuclei and 2 sections for the deep cerebellar nuclei. For each structure, 6 samples (3/side) were taken from each section. In pre-cerebellar structures, analysis of GAP-43 IR resulted in series of OD values for each side. Since several random factors may influence immunostaining intensity, we did not determine absolute staining levels. Following the method adopted by Kraus and Illing (2004) and Meidinger et al. (2006), we computed the OD ratio between right (ipsilateral to HCb) and left side for each structure, considering that an OD ratio value equal to 1 indicated a balanced IR between sides, OD ratio values greater than 1 indicated an IR superior on the right in comparison to the left side, and OD ratio values less than 1 indicated an IR superior on the left in comparison to the right side.

To evaluate the HCb effect on neuronal death in the pre-cerebellar structures, cell density was assessed in C and H36 groups. Thionin stained sections were visualized by using an optical microscopy (magnification: 200 ×) fitted with a camera. Measurements were made by means of the software Neurolucida (Microbrigthfield, Williston, VT). For each structure, analyses were performed on 7 sections for the inferior olive and 4 sections for lateral reticular nucleus and pontine nuclei. Counting of the neuronal soma was performed separately for each side. Cell density in each side was then calculated as ratio between cell number and area value. Measures of cell density have been used to compute the percentage difference of cell density on the axotomized side (projecting to the ablated hemicerebellum), taking the not axotomized side (projecting to the spared hemicerebellum) as reference. Statistical analysis All data were presented as the mean ± SEM and were tested for normality (Will-Shapiro's test) and homoscedasticity (Levene's test). One-way ANOVAs with HCbed groups as fixed factor were performed on percentages of cerebellar ablation (or behavioral scores exhibited at various post-operative days). Nested ANOVAs with “animal” as random and “group” as fixed factors have been performed on OD ratio values and percentage of neuronal loss for each pre-cerebellar structure as well as on OD values for each cerebellar nucleus. Additionally, cell density on each side in the C group was compared by means of paired-sample t tests. Furthermore, OD ratio values in the C group were compared with the value of 1 (indicating a perfectly balanced IR between sides) by means of a one-sample t test. A two-way ANOVA with “group” as between- and “side” as within-factor has been performed on OD values for pontine nuclei. Percentage values were always corrected by angular transformation prior to statistical analyses. When it was appropriated, post-hoc comparisons were made using HSD Tukey's tests. Statistical analyses were made by means of the software SPSS 8. Results Verification of cerebellar lesion The overall percentage of ablation of cerebellar tissue among all HCbed groups (H and H1–H36) was of 50.03% ± 0.1. One-way ANOVA on mean percentages of ablation revealed no differences among groups (F7, 76 = 0.12; p = n.s.) (Fig. 1). Neurological assessment of cerebellar symptomatology Twenty-four hours after the right HCb, the animals exhibited severe postural and motor deficits. Namely, HCb induced strong extensor hypotonia ipsilateral to the lesion side causing tilted posture to the right. Locomotion was markedly ataxic and stunted, with oscillatory movements and staggering. Dynamic tremor was also present. While stepping, the most evident symptoms were right hindlimb hyperflexion during the swing phase and joint angle fluctuations.

L. Burello et al. / Experimental Neurology 233 (2012) 273–282

277

Fig. 2. Neurological assessment of cerebellar symptomatology. Behavioral scores of all cerebellar symptoms (total) as well as of single symptom clusters (head posture, trunk and limb posture, locomotion and complex motor skills) at variable time intervals. Dotted line represents sham animals performance. Abscissa: time in days from HCb. Ordinate: degree of severity of cerebellar symptoms evaluated according to the rating scale described in Table 1. Data are presented as means ± SEM. Symbols indicating total or single clusters of symptoms also indicate significant differences between subsequent time intervals. (Post-hoc comparisons: Δ p > .05; ✖✖ p > .005; ♦♦♦/ /✖✖✖ p > .0005).

To change direction, the right hindlimb was used as a pivot, remaining flexed under their body weight. Wide-based locomotion and collapsing on the bellies were present. Locomotion tended also to veer to the lesion side provoking asymmetrical gait and falls. To avoid falling, animals tended to lean against support surfaces. Eye nystagmus beating to the right side was also present. These severe postural and locomotor deficits tended to progressively diminish over time. At the end of postural evaluation (42nd post-operative day), the animals still exhibited some slight cerebellar symptoms as right extensor hypotonia, right hindlimb flexion, and wide-based locomotor posture. Nevertheless, they succeeded in walking quite efficiently without support. One-way ANOVA with repeated measures on the scores of all cerebellar symptoms taken into account obtained by H group revealed a significant time effect (F11, 66 = 51.31; p b .00005), given symptom severity significantly diminished as days went by. One-way ANOVAs on scores of the symptoms relative to head (F11, 66 = 5.68; p b .00005), trunk and limb (F11, 66 =57.17; pb .00005), locomotion (F11, 66 =18.00; pb .00005) and complex motor skills (F11, 66 =54.13; pb .00005) revealed significant time effects. Interestingly, post-hoc comparisons indicated the maximal decrease of symptom severity between the 14th and 21st day as for trunk and limb symptoms and locomotor handicaps, while the maximal decrease of severity of the complex motor skills was reached between the 21st and 28th day (Fig. 2). GAP-43 IR Inferior olive In C group, the inferior olive showed a faint and symmetric GAP-43 IR, as confirmed by the one-sample t test performed on OD ratio values with 1 as test value (t = 1.46; p = n.s.). In the presence of the right HCb, a strong increase in GAP-43 IR was observed in all individual sub-nuclei (principal, medial accessory and dorsal accessory) of the left inferior olive. ANOVA on OD ratio values calculated in the seven HCbed (H1–H36) groups and in the C group revealed a significant group effect (F7, 16 = 128.44; p b .00005). Post-hoc comparisons revealed that OD ratio values were significantly lowered (always p b .00005) in H1, H3 and H6 groups in comparison to C group. Namely, OD ratio values reached its negative peak in the H6 group. Such a finding indicates a GAP-43 IR superior on the left inferior olive in comparison to the right one during the first post-operative week. Alternatively, in the last post-operative groups (H14–H36) OD ratio values did not differ from those of the C group. Such a finding indicates that from the second post-operative week onward GAP-43 IR became balanced between sides (Figs. 3, 4). The slight prevalence in the right IO observed in the H36 group has to be attributed to the progressive cell death affecting the left IO (as described in Neuronal death Results section and illustrated in Fig. 7).

Lateral reticular nucleus In the C group, lateral reticular nucleus displayed a weak and balanced GAP-43 IR (t = 1.2; p = n.s.). In the presence of the right HCb, GAP-43 IR showed a striking increase in the ipsilateral lateral reticular nucleus. ANOVA on OD ratio values evidenced a significant group effect (F7, 16 = 9.66; p b .0005). Post-hoc comparisons demonstrated that in the H1–H30 groups OD ratio values were significantly higher (always p b .0005) than in the C group. Such data indicate that the superior GAP-43 IR on the right side persisted up to a month after injury. Notably, this side imbalance reached its peak in H3 and H6 groups. Only in the H36 group OD ratio values became comparable to those found in the C group, indicating a balanced GAP-43 IR between sides (Figs. 3, 4). Pontine nuclei In the C group, pontine nuclei appear poorly immunoreactive and balanced between sides (t= 0.7; p = n.s.). After the right HCb, GAP-43 IR showed an early imbalance between sides. ANOVA on OD ratio values demonstrated a significant group effect (F7, 16 = 185.44; p b .00005). Post-hoc comparisons displayed that in all HCbed groups (H1–H36) OD ratio values were significantly different (at least p b .05) as regards those detected in the C group. Namely, in the H1 group OD ratio values were significantly lowered (pb .05) in comparison to the C group, while in all remaining groups OD ratio values were significantly higher (always p b .00005). Notably, this imbalance peaked in the H6 group. These findings indicate that GAP-43 IR prevailed in the left side at the first post-operative day, afterwards a marked shifting toward the right side occurred and lasted for over a month (Figs. 3, 4). Given GAP-43 IR shifted between sides, an analysis on OD values was retained useful to further examine the actual time course of GAP-43 IR in each side. In fact, in addition to OD ratio values the analysis of OD values allowed specifically controlling GAP-43 IR in the left side, verifying thus the eventual presence of masking effects linked to the high GAP-43 IR in the right side. Two-way ANOVA (group × side) on OD values evidenced a significant interaction (F7, 184 = 124.46; p b .00005). Post-hoc comparisons indicated a GAP43 IR time course substantially similar to that of OD ratio values. Particularly, when the left side was analyzed, GAP-43 IR significantly increased (pb .005) only in the H1 group, while in the remaining groups (H3–H36) it did not differ from that detected in the C group. Thus, this analysis indicated no masking effect of the right side and an actual shifting from the left to the right side. Deep cerebellar nuclei In the C group, deep cerebellar nuclei appeared weakly and homogeneously immunoreactive. In the presence of the right HCb, the three nuclei of the spared side showed different immunoreactive responses.

278

L. Burello et al. / Experimental Neurology 233 (2012) 273–282

These findings indicate that the fastigial and interposed nuclei have immunoreactive responses that follow time courses different from those of the dentate nucleus (Figs. 5, 6).

Neuronal death Inferior olive In the C group, a cell density balanced between sides was obviously observed, as confirmed by the paired-sample t test (t=1; p=n.s.). In the H36 group, the left inferior olive (projecting to the ablated hemicerebellum and then whose neurons were axotomized) exhibited a marked decrease of cell density in comparison to the right inferior olive (Fig. 7). ANOVA on percentage difference of cell density between sides in lesioned (H36) and un-lesioned (C) groups evidenced a significant lesion effect (F1, 4 = 216.71; p b .0005) related to the decrease of cell density exhibited by the H36 group in the left inferior olive (Fig. 7). Furthermore, neuronal degeneration unevenly affected the olivary sub-nuclei. In fact, the neurons of the principal olive showed higher sensitivity (−66.2% ± 3.4) to axotomy caused by HCb in comparison to the neurons of medial accessory (−46.9% ± 2.4) and dorsal accessory olive (−54.3% ± 5.7), as it was previously shown in a peduncolotomy model in the rat (Buffo et al., 1998).

Lateral reticular nucleus In the C group a cell density balanced between sides was present (t = 1; p = n.s.). In the H36 group, the right lateral reticular nucleus (projecting to the ablated hemicerebellum and then whose neurons were axotomized) presented a striking reduction of cell density in comparison to the left side (Fig. 7). ANOVA on percentage difference of cell density between sides in lesioned (H36) and un-lesioned (C) groups revealed a significant lesion effect (F1, 4 = 2503.86; p b .00005) due to the unilateral loss in cell density displayed by the H36 group (Fig. 7).

Fig. 3. Optical density (OD) ratio in the pre-cerebellar nuclei. OD ratio in the precerebellar nuclei of the sham Control (C) animals and HCbed groups at the various post-operative days. OD ratio is computed between right (ipsilateral to HCb) and left side (OD ratio = 1: balanced immunoreactivity (IR) between sides; OD ratio > 1: IR superior on the right side; OD ratio b 1: IR superior on the left side). In this and in the following figures, H1 indicates the HCbed group sacrificed at the 1st post-operative day, H3, H6, H14, H21, H30, H36 the HCbed groups sacrificed at the 3rd, 6th, 14th, 21st, 30th and 36th post-operative day, respectively (n= 3/group); data are presented as means± SEM; while the circles indicate significant differences between single HCbed groups and the C group, the asterisks indicate significant differences between HCbed groups (°/* p b .05; ** p b .005; °°°/*** p b .00005).

Pontine nuclei In the C group a cell density balanced between sides was present (t = 3.46; p = n.s.). In the H36 group, the left pontine nuclei (projecting to the ablated hemicerebellum and then whose neurons were axotomized) exhibited a strong decrease of cell density in comparison to the contralateral side (Fig. 7). ANOVA on percentage difference of cell density between sides in lesioned (H36) and un-lesioned (C) groups demonstrated a significant lesion effect (F1, 4 = 389.72; p b .00005) as a consequence of the unilateral loss in cell density showed by the H36 group (Fig. 7).

Discussion In the fastigial nucleus, ANOVA on OD values revealed a significant group effect (F7, 16 = 248.44; p b .00005). Post-hoc comparisons showed that in all post-operative groups (H1–H36) OD values were significantly higher (at least p b .05) than those of the C group. Particularly, OD values displayed their peak in the H6 group. In the interposed nucleus, ANOVA on OD values evidenced a significant group effect (F7, 16 = 11.65; p b .00005). Post-hoc comparisons displayed that OD values in all post-operative groups were significantly higher (always p b .00005) in comparison to those observed in the C group. Specifically, GAP-43 IR peaked in the H3 and H6 groups. In the dentate nucleus, ANOVA on OD values demonstrated a significant group effect (F7, 16 = 265.85; p b .00005). Also in this nucleus, post-hoc comparisons indicated that in all post-operative groups OD values were significantly higher (always pb .00005) than in the C group. Nevertheless, the peak of immunoreactive responses was markedly delayed. Namely, GAP-43 IR reached its highest point in the H30 group.

Gaining a deepened knowledge of the neuronal responses to brain damage allows more tuned interventions to improve functional recovery. At this aim, because of the remarkable compensation of cerebellar symptoms after damage as well as the great plastic potentialities of cerebellum, the cerebellar system represents a structure particularly effective to address the not yet fully clarified linkage between plastic post-lesional rearrangement and amelioration of functional outcomes. The present research addressed the plastic mechanisms at the basis of the cerebellar compensation by employing an experimental model of hemicerebellectomy. Post-lesional remodeling has been analyzed by assessing the time course of GAP-43 immunoreactivity in brainstem pre-cerebellar structures and in deep cerebellar nuclei of the hemicerebellum spared. The time course of the neuro-anatomical modifications has been compared with the time course of the compensation of a variety of cerebellar symptoms.

L. Burello et al. / Experimental Neurology 233 (2012) 273–282

279

Fig. 4. GAP-43 immunoreactivity in the pre-cerebellar nuclei. Photomicrographs representing GAP-43 IR in the pre-cerebellar nuclei (IO: Inferior Olive; LRn: Lateral reticular nucleus; Pn: Pontine nuclei) of the Control (C) and representative HCbed groups (H6 and H30). Scale bar: 200 μm.

GAP-43 levels in pre-cerebellar axotomized neurons The neurons that underwent an axotomy as consequence of the right HCb, that is, the neurons of left inferior olive, right lateral reticular nucleus and left pontine nuclei, exhibited increased GAP-43 levels and enhanced cell death. The association of both reactive phenomena demonstrates that GAP-43 does take part in neuronal reactions to damage (Dusart et al., 2005). Namely, GAP-43 is retained to be involved in the molecular cascade providing axotomized neurons with information about the presence (or absence) of a favorable microenvironment directing then the essential choice towards either neuronal death or axonal re-growth (Wehrlé et al., 2001). Since the adult CNS microenvironment has a molecular composition that inhibits axonal re-growth, the lesion-induced increase of GAP-43 levels observed in the current research should be related to the role of driving

Fig. 5. Optical density (OD) in the deep cerebellar nuclei. OD in the deep cerebellar nuclei of the spared hemicerebellum of the Control (C) and all HCbed groups (n = 3/ group) (°/* p b .05; ** p b .005; °°°/*** p b .00005).

axotomized neurons towards the retrograde death (Herdegen et al., 1997; Wehrlé et al., 2001). Noteworthy, the increase of GAP-43 levels displayed a definite time course in each pre-cerebellar nucleus. In fact, after the right HCb GAP-43 prevalence lasted for a month in the right lateral reticular nucleus, for a week in the left inferior olive and only for a day in the left pontine nuclei. Such a variability in timing seems to indicate that the mere GAP-43 presence is enough to operate as intrinsic sensor of local microenvironmental features. Timing variability among pre-cerebellar nuclei is also consistent with literature data demonstrating that GAP-43 basal levels as well as ability to express it after axotomy significantly vary among different neuronal phenotypes (Benowitz and Routtenberg, 1997; Denny, 2006; Dusart et al., 2005). GAP-43 levels in pre-cerebellar not-axotomized neurons Interestingly, after the initial GAP-43 increase in the left pontine nuclei, from 3rd post-operative day onward an abrupt increase of GAP-43 levels was observed in the not-axotomized neurons of right pontine nuclei projecting to the spared (left) hemicerebellum. This shifting in immunoreactivity might indicate the development of plastic compensatory responses probably linked to pivotal role of pontine nuclei within the cortico-cerebellar connectivity. Tracing studies have shown that in rats virtually all and in primates about two-thirds of cortical projections directed to the cerebellum synapse the pontine nuclei to then contact the cerebellar granular layer as mossy fibers (Glickstein et al., 1985; Legg et al., 1989). Conversely, cortical projections to inferior olive and lateral reticular nucleus are scarce and restricted to sensori-motor cortical areas (Ruigrok and Cella, 1995; Schwarz and Thier, 1999). Accordingly, pontine nuclei are considered not a simple relay station but a highly integrative processing center to orchestrate the composition of neocortical inputs to the demands of cerebellar signal processing (Möck et al., 2006; Schwarz and Thier, 1999; Schwarz and Schmitz, 1997). It is interesting to note that the involvement of the pontine nuclei in compensatory plastic phenomena is supported by studies reporting that GAP-43 expression in intact neurons is an intrinsic determinant mediating not only axonal growth but also activity-dependent local reorganization (Benowitz and Routtenberg, 1987, 1997; Strittmatter et al., 1992). Thus, by analyzing the time course of recovery from cerebellar symptoms and the time course of GAP-43 IR in the not-axotomized neurons of right pontine nuclei it is possible to evidence that the increase of GAP-43 levels peaked before cerebellar deficits started to significantly diminish. This finding suggests that plastic rearrangements in the spared (cortico)ponto-cerebellar network may precede and drive compensation of

280

L. Burello et al. / Experimental Neurology 233 (2012) 273–282

Fig. 6. GAP-43 immunoreactivity in the deep cerebellar nuclei. Photomicrographs representing GAP-43 IR in the deep cerebellar nuclei of the Control (C) and representative HCbed groups (H6 and H30). Scale bar: 200 μm.

postural and motor cerebellar asymmetries and also indicate GAP-43 as a key mediator of spontaneous plastic remodeling. In this regard, it is interesting to consider that pharmacological drugs as glucocorticoids have been repeatedly proposed as

controllers of neuro-inflammation and neuronal death after CNS injury (Heiduschka and Thanos, 2006; Kanellopoulos et al., 1997; Kaptanoglu et al., 2000; Leypold et al., 2007; Vaquero et al., 2006). However, since it is known that glucocorticoid treatment strongly

Fig. 7. Cell death in the pre-cerebellar nuclei. In A, the percentage difference of cell density on the axotomized side (projecting to the ablated hemicerebellum), taking the not axotomized side (projecting to the spared hemicerebellum) as reference in C and H36 groups (** p b .005; *** p b .00005). In B, photomicrographs of the not axotomized (not-axt) and axotomized (axt) pre-cerebellar nuclei. Note the marked retrograde cell death only in the axotomized nuclei. Scale bar: 200 μm.

L. Burello et al. / Experimental Neurology 233 (2012) 273–282

represses expression of GAP-43 (Federoff et al., 1988), detrimental rather than beneficial effects on the functional recovery after a brain lesion could be obtained following a glucocorticoid treatment. This concern emerged from present data supplements recent studies reporting a questionable effectiveness of glucocorticoids on CNS inflammatory and neurodegenerative responses (Dinkel et al., 2003; MacPherson et al., 2005) even if no reference to GAP-43 expression is made.

281

Disclosure statement The authors declare that they have not competing interests. Acknowledgments This investigation was supported by MIUR grants to LP.

GAP-43 levels in neurons of the spared deep cerebellar nuclei Appendix A. Supplementary data GAP-43 IR increased also in the three deep cerebellar nuclei of the spared hemicerebellum. Taking into account that cerebellar nuclei receive the pontine projections via collaterals of mossy fibers and in turn send out feedback excitatory projections to pontine nuclei (Schwarz and Thier, 1999; Voogd, 1995), the current data demonstrate that reactive reorganization entails each single component of the ponto-cerebellar network. Interestingly, each cerebellar nucleus showed a distinct time course of GAP-43 IR. In the fastigial and interposed nuclei GAP-43 IR peaked during the first post-operative week, whereas in the dentate nucleus only after one month. Fastigial and interposed nuclei receive inputs directly from the spinal cord and influence the motor ventro-medial and dorso-lateral systems, respectively (Morton and Bastian, 2004). On the contrary, the dentate nucleus loops neocortical areas, as the prefrontal, premotor, primary motor and parietal cortices (Manto and Taib, 2010; Morton and Bastian, 2004; Voogd, 1995). Hence, each cerebellar nucleus is engaged in functionally different modules. In particular, fastigial nucleus and its connections play a primary role in regulating extensor tone and maintaining dynamic balance control, modulating the rhythmic flexor and extensor muscle activity within the locomotor pattern (Chambers and Sprague, 1955; Thach et al., 1992). Instead, the interposed nucleus and its network play a more important role in directing limb placement and regulating agonist–antagonist muscle pairs to control timing, amplitude and trajectory of locomotor movements (Chambers and Sprague, 1955; Thach et al., 1992). Finally, the dentate nucleus and its reciprocal neocortical connections play a significant role in motor control when complex adaptations of the locomotor pattern are required, when polymodal cue-guided and cue-timed tasks are performed and when the attentional load is higher (Manto and Taib, 2010). Not by chance, clustering the different symptoms of cerebellar origin according to postural, locomotor or complex functions, we have found that the symptoms related to limb and trunk posture as well as the deficits related to the locomotor pattern showed a significant compensation between the 2nd and 3rd postoperative week and exhibited recovery time courses closely similar each other. Instead, the complex motor behaviors demanding muscle coordination and attention burden reached their maximal recovery between the 3rd and the 4th post-operative week. Therefore, it is reasonable to advance that the earlier plastic events of the fastigial and interposed nuclei were driving compensation of the elementary features of posture and locomotion, while the later plastic events of the dentate nucleus were mediating the recovered ability to flexibly adjust the locomotor plan in response to more requiring contexts. Conclusions The present data advance that the remodeling of the cerebellar networks underlies functional compensation of faceted motor functions and indicate that the most effective rehabilitative interventions should exploit brain plastic potential to guide and enhance recovery. In this respect, some of us recently described that the enhancement of brain plastic properties induced by the exposure to enriched environment does attenuate the manifestations of cerebellar damage and that beneficial effects of enrichment on cerebellar symptoms are exerted through plastic adaptations of various neuronal circuits connected with cerebellar networks (Cutuli et al., 2011).

Supplementary data to this article can be found online at doi:10. 1016/j.expneurol.2011.10.016. References Abramoff, M.D., Magalhaes, P.J., Ram, S.J., 2004. Image processing with imageJ. Biophotonics Int. 11, 36–42. Aigner, L., Caroni, P., 1995. Absence of persistent spreading, branching, and adhesion in GAP-43-depleted growth cones. J. Cell Biol. 128, 647–660. Aigner, L., Arber, S., Kapfhammer, J.P., Laux, T., Schneider, C., Botteri, F., Brenner, H.R., Caroni, P., 1995. Over-expression of the neural growth-associated protein GAP-43 induces nerve sprouting in the adult nervous system of transgenic mice. Cell 83, 269–278. Benowitz, L.I., Routtenberg, A., 1987. A membrane phosphoprotein associated with neural development, axonal regeneration, phospholipid metabolism, and synaptic plasticity. Trends Neurosci. 10, 527–532. Benowitz, L.I., Routtenberg, A., 1997. GAP-43: an intrinsic determinant of neuronal development and plasticity. Trends Neurosci. 20, 84–91. Bremner, W.J., Millar, M.R., Sharpe, R.M., Saunders, P.T.K., 1994. Immunoistochemical localization of androgen receptors in the rat testis: evidence for stage-dependent expression and regulation by androgens. Endocrinology 135, 1227–1234. Buffo, A., Fronte, M., Oestreicher, A.B., 1998. Degenerative phenomena and reactive modifications of the adult rat inferior olivary neurons following axotomy and disconnection from their targets. Neuroscience 85, 587–604. Centonze, D., Rossi, S., De Bartolo, P., De Chiara, V., Foti, F., Musella, A., Mataluni, G., Rossi, S., Bernardi, G., Koch, G., Petrosini, L., 2008. Adaptations of glutamatergic synapses in the striatum contribute to recovery from cerebellar damage. Eur. J. Neurosci. 27, 2188–2196. Chaisuksunt, V., Zhang, Y., Anderson, P.N., Campbell, G., Vaudano, E., Schacher, M., Lieberman, A.R., 2000. Axonal regeneration from CNS neurons in the cerebellum and brainstem of adult rats: correlation with the patterns of expression and distribution of messenger RNAs for L1, CHL1, c-jun and growth-associated protein-43. Neuroscience 100, 87–108. Chambers, W.W., Sprague, J.M., 1955. Functional localization in the cerebellum. I. Organization in longitudinal cortico-nuclear zones and their contribution to the control of posture, both extrapyramidal and pyramidal. J. Comp. Neurol. 103, 105–129. Cutuli, D., Rossi, S., Burello, L., Laricchiuta, D., De Chiara, V., Foti, F., De Bartolo, P., Musella, A., Gelfo, F., Centonze, D., Petrosini, L., 2011. Before or after does it matter? Different protocols of environmental enrichment differently influence motor, synaptic and structural deficits of cerebellar origin. Neurobiol. Dis. 42, 9–20. De Graan, P.N., van Hooff, C.O., Tilly, B.C., Oestreicher, A.B., Schotman, P., Gispen, W.H., 1985. Phosphoprotein B-50 in nerve growth cones from fetal rat brain. Neurosci. Lett. 61, 235–241. Denny, J.B., 2006. Molecular mechanisms, biological actions, and neuropharmacology of the growth-associated protein GAP-43. Curr. Neuropharmacol. 4, 293–304. Dinkel, K., MacPherson, A., Sapolsky, R., 2003. Novel glucocorticoid effects on acute inflammation in the CNS. J. Neurochem. 84, 705–716. Duffau, H., 2006. Brain plasticity: from pathophysiological mechanisms to therapeutic applications. J. Clin. Neurosci. 13, 885–897. Dusart, I., Ghoumari, R., Wehrlé, R., Morel, M.P., Bouslama-Oueghlani, L., Camand, E., Sotelo, C., 2005. Cell death and axon regeneration of Purkinje cells after axotomy: challenges of classical hypotheses of axon regeneration. Brain Res. Rev. 49, 300–316. Federico, F., Leggio, M.G., Mandolesi, L., Petrosini, L., 2006. The NMDA receptor antagonist CGS 19755 disrupts recovery following cerebellar lesions. Restor. Neurol. Neurosci. 24, 1–7. Federoff, H.J., Grabczyk, E., Fishman, M.C., 1988. Dual regulation of GAP-43 gene expression by nerve growth factor and glucocorticoids. J. Biol. Chem. 263, 19290–19295. Foti, F., Laricchiuta, D., Cutuli, D., De Bartolo, P., Gelfo, F., Angelucci, F., Petrosini, L., 2011. Exposure to an enriched environment accelerates recovery from cerebellar lesion. Cerebellum 10, 104–119. Gagliardini, V., Dusart, I., Frankhausen, C., 2000. Absence of GAP-43 can protect neurons from death. Mol. Cell. Neurosci. 16, 27–33. Gelfo, F., Cutuli, D., Foti, F., Laricchiuta, D., De Bartolo, P., Caltagirone, C., Petrosini, L., Angelucci, F., 2011. Enriched environment improves motor function and increases neurotrophins in hemicerebellar lesioned rats. Neurorehabil. Neural Repair 25, 243–252. Gispen, W.H., Nielander, H.B., De Graan, P.N., Oestreicher, A.B., Schrama, L.H., Schotman, P., 1991. Role of the growth-associated protein B50/GAP-43 in neuronal plasticity. Mol. Neurobiol. 5, 61–85.

282

L. Burello et al. / Experimental Neurology 233 (2012) 273–282

Glickstein, M., May III, J.G., Mercier, B.E., 1985. Corticopontine projection in the macaque: the distribution of labelled cortical cells after large injections of horseradish peroxidase in the pontine nuclei. J. Comp. Neurol. 235, 343–359. Hansel, C., Linden, D.J., D'Angelo, E., 2001. Beyond parallel fiber LTD: the diversity of synaptic and non-synaptic plasticity in the cerebellum. Nat. Neurosci. 4, 467–475. Heiduschka, P., Thanos, S., 2006. Cortisol promotes survival and regeneration of axotomised retinal ganglion cells and enhances effects of aurintricarboxylic acid. Graefes Arch. Clin. Exp. Ophthalmol. 244, 1512–1521. Herdegen, T., Skene, P., Bähr, M., 1997. The c-Jun transcription factor — bipotential mediator of neuronal death, survival and regeneration. Trends Neurosci. 20, 227–231. Holmes, G., 1917. The symptoms of acute cerebellar injuries due to gunshot injuries. Brain 40, 461–535. Kanellopoulos, G.K., Kato, H., Wu, Y., Dougenis, D., Mackey, M., Hsu, C.Y., Kouchoukos, N.T., 1997. Neuronal cell death in the ischemic spinal cord: the effect of methylprednisolone. Ann. Thorac. Surg. 64, 1279–1285. Kapfhammer, J.P., Schwab, M.E., 1994a. Inverse patterns of myelination and GAP-43 expression in the adult CNS: neurite growth inhibitors as regulators of neuronal plasticity? J. Comp. Neurol. 340, 194–206. Kapfhammer, J.P., Schwab, M.E., 1994b. Increased expression of growthassociated protein GAP-43 in myelin-free rat spinal cord. Eur. J. Neurosci. 6, 403–411. Kaptanoglu, E., Tuncel, M., Palaoglu, S., Konan, A., Demirpence, E., Kilinc, K., 2000. Comparison of the effects of melatonin and methylprednisolone in experimental spinal cord injury. J. Neurosurg. 93, 77–84. Kraus, K.S., Illing, R.-B., 2004. Superior olivary contributions to auditory system plasticity: medial but not lateral olivocochlear neurons are the source of cochleotomyinduced GAP-43 expression in the ventral cochlear nucleus. J. Comp. Neurol. 475, 374–390. Legg, C.R., Mercier, B., Glickstein, M., 1989. Corticopontine projection in the rat: the distribution of labelled cortical cells after large injections of horseradish peroxidase in the pontine nuclei. J. Comp. Neurol. 286, 427–441. Leypold, B.G., Flanders, A.E., Schwartz, E.D., Burns, A.S., 2007. The impact of methylprednisolone on lesion severity following spinal cord injury. Spine (Phila. Pa. 1976) 32, 373–378. Luciani, L., 1891. Il cervelletto. Nuovi studi di fisiologia normale e patologica. Le Monnier, Firenze. MacPherson, A., Dinkel, K., Sapolsky, R., 2005. Glucocorticoids worsen excitotoxininduced a expression of pro-inflammatory cytokines in hippocampal cultures. Exp. Neurol. 194, 376–383. Manto, M., Taib, N.O.B., 2010. Cerebellar nuclei: key roles for strategically located structures. Cerebellum 9, 17–21. Meidinger, M.A., Hildebrandt-Schoenfeld, H., Illing, R.-B., 2006. Cochlear damage induces GAP-43 expression in cholinergic synapses of the cochlear nucleus in the adult rat: a light and electron microscopic study. Eur. J. Neurosci. 23, 3187–3199.

Möck, M., Butovas, S., Schwarz, C., 2006. Functional unity of the ponto-cerebellum: evidence that intrapontine communication is mediated by a reciprocal loop with the cerebellar nuclei. J. Neurophysiol. 95, 3414–3425. Morton, S.M., Bastian, A.J., 2004. Cerebellar control of balance and locomotion. Neuroscientist 10, 247–259. Nudo, R.J., Plautz, E.J., Frost, S.B., 2001. Role of adaptive plasticity in recovery of function after damage to motor cortex. Muscle Nerve 24, 1000–1019. Paxinos, G., Watson, C., 1998. The Rat Brain in Stereotaxic Coordinates, 4th ed. Academic Press, San Diego. Properzi, F., Fawcett, J.W., 2004. Proteoglycans and brain repair. News Physiol. Sci. 19, 33–38. Routtenberg, A., Cantallops, I., Zaffuto, S., Serrano, P., Namgung, U., 2000. Enhanced learning after genetic overexpression of a brain growth protein. Proc. Natl. Acad. Sci. U. S. A. 97, 7657–7662. Ruigrok, T.J.H., Cella, F., 1995. Precerebellar nuclei and red nucleus, In: Paxinos, G. (Ed.), The Rat Nervous System, 2nd ed. Academic press, San Diego, pp. 277–308. Schaechter, J.D., Benowitz, L.I., 1993. Activation of protein kinase C by arachidonic acid selectively enhances the phosphorilation of GAP-43 in nerve terminal membranes. J. Neurosci. 13, 4361–4371. Schwarz, C., Schmitz, Y., 1997. Projection from the cerebellar lateral nucleus to precerebellar nuclei in the mossy fiber pathway is glutamatergic: a study combining anterograde tracing with immunogold labeling in the rat. J. Comp. Neurol. 381, 320–334. Schwarz, C., Thier, P., 1999. Binding of signals relevant for action: towards a hypothesis of the functional role of the pontine nuclei. Trends Neurosci. 22, 443–451. Skene, J.H., 1989. Axonal growth-associated proteins. Annu. Rev. Neurosci. 12, 127–156. Skene, J.H., Willard, M., 1981. Changes in axonally transported proteins during axon regeneration in toad retinal ganglion cells. J. Cell Biol. 1989, 86–95. Strittmatter, S.M., Vartanian, T., Fishman, M.C., 1992. GAP-43 as a plasticity protein in neuronal form and repair. J. Neurobiol. 23, 507–520. Strittmatter, S.M., Fankhauser, C., Huang, P.L., Mashimo, H., Fishman, M.C., 1995. Neuronal pathfinding is abnormal in mice lacking the neuronal growth cone protein GAP-43. Cell 80, 445–452. Thach, W.T., Goodkin, H.P., Keating, J.G., 1992. The cerebellum and the adaptive coordination of movement. Annu. Rev. Neurosci. 15, 403–442. Vaquero, J., Zurita, M., Oya, S., Aguayo, C., Bonilla, C., 2006. Early administration of methylprednisolone decreases apoptotic cell death after spinal cord injury. Histol. Histopathol. 21, 1091–1102. Voogd, J., 1995. Cerebellum, In: Paxinos, G. (Ed.), The Rat Nervous System, 2nd ed. Academic press, San Diego, pp. 277–308. Wehrlé, R., Caroni, P., Sotelo, C., Dusart, I., 2001. Role of GAP-43 in mediating the responsiveness of cerebellar and precerebellar neurons to axotomy. Eur. J. Neurosci. 13, 857–870. Wieloch, T., Nikolich, K., 2006. Mechanisms of neural plasticity following brain injury. Curr. Opin. Neurobiol. 16, 258–264.