Nitric oxide synthase in the frog cerebellum: Response of Purkinje neurons to unilateral eighth nerve transection

Authors

  • Maria Bonaria Pisu,

    1. Dipartimento di Biologia Animale, Università di Pavia, Pavia, Italy
    2. Centro di Studio per l'Istochimica del CNR, Pavia, Italy
    Search for more papers by this author
  • Elena Conforti,

    1. Dipartimento di Biologia Animale, Università di Pavia, Pavia, Italy
    2. Centro di Studio per l'Istochimica del CNR, Pavia, Italy
    Search for more papers by this author
  • Laura Botta,

    1. Dipartimento di Scienze Fisiologiche, Farmacologiche, Cellulari-Molecolari, Università di Pavia, Pavia, Italy
    Search for more papers by this author
  • Paolo Valli,

    1. Dipartimento di Scienze Fisiologiche, Farmacologiche, Cellulari-Molecolari, Università di Pavia, Pavia, Italy
    Search for more papers by this author
  • Graziella Bernocchi

    Corresponding author
    1. Dipartimento di Biologia Animale, Università di Pavia, Pavia, Italy
    2. Centro di Studio per l'Istochimica del CNR, Pavia, Italy
    • Dipartimento di Biologia Animale, Università di Pavia, Piazza Botta 10, I-27100, Pavia, Italy
    Search for more papers by this author
    • Fax: +39-0382-506325


Abstract

When vestibular damage occurs, nitric oxide synthase (NOS) expression in rat cerebellar flocculus is affected. Since compensation for postural symptoms occurs and Purkinje cells play an important role in movement coordination and motor learning, we analyzed in situ the induction of NOS in the Purkinje cell population of the cerebellum (corpus cerebelli) of frog after unilateral transection of the eighth statoacoustic nerve to gain insight into the role of NO in neural plasticity after injury. Three days after neurectomy, the early effects induced NADPH diaphorase reactivity in most of the Purkinje cells on the ipsilateral side, while on the contralateral side the highest labeling was observed at 15 days. This finding can give information on the dynamics of vestibular compensation, in which NOS involvement was investigated. At 30 days, NADPH diaphorase reactivity was present in a large number of Purkinje cells of the whole cerebellum, while at 60 days a down-regulation for NADPH diaphorase reactivity was evident. A similar trend was observed for NOS-immunoreactivity, which was still present at 60 days in a high percentage of Purkinje cells, mainly on the ipsilateral side. On the basis of cell density evaluations, it was proposed that the early induction of NOS after neurectomy was linked to the degeneration of a part of the Purkinje neurons, while the permanence of NOS labeling might be due to a neuroprotective role of NO in the restoration phase of the vestibular compensation process. Anat Rec 268:73–83, 2002. © 2002 Wiley-Liss, Inc.

In the central nervous system (CNS), nitric oxide (NO) has many functions. It acts as an unconventional neuronal messenger and is synthesized on demand (for review, see Steinbusch et al., 2000). It is generated by NO synthase (NOS), an enzyme expressed by discrete neuronal populations throughout the nervous system (Schuman and Madison, 1994). Aberrant synthesis and interactions of NO with other molecules are thought to underlie a number of neurodegenerative processes and CNS disorders (for reviews, see Schulz et al., 1995; Wolf, 1997). It is also reported that NO has a role in neuroprotection. It has been demonstrated that the moderately enhanced production of NO at the onset of ischemia has neuroprotective effects, by limiting ischemic brain damage from vascular effects (Verrecchia et al., 1995).

In situ reliable markers of NO-producing neurons are a histochemical reaction for NADPH-diaphorase (NADPHd) activity, and an immunocytochemical reaction using anti-NOS antibodies. There have been several reports presenting evidence that the levels of NADPHd/NOS are dynamic and that NADPHd/NOS can be induced in neurons that normally express low or no detectable levels of these enzymes. The induction of NADPHd/NOS has been found in a number of CNS regions following mechanical tissue damage (Saxon and Beitz, 1996).

Transection of peripheral nerves generates complex and long-lasting neurobiological events, ranging from neuronal cell death to successful regeneration with restored cellular functions. It has been proposed that NO plays an important role in the synaptic plasticity that follows certain neuronal injuries (Brüning, 1993; Wu, 1993; Wu and Scott, 1993; Herdegen et al., 1994). It has also been suggested that the blockage of neuronal NOS expression (Novikov et al., 1997; Yu, 1997) is responsible for axon regeneration of axotomized spinal motor neurons. In the regenerative response of motoneurons in the lizard tail spinal cord after autotomy, NADPHd reactivity was induced during tail regrowth, and NOS immunoreactivity increased (Cristino et al., 2000).

Vestibular compensation has been used as a model of post-lesional plasticity in the CNS (Llinas and Walton, 1979; Lacour and Xerri, 1981; Flugel et al., 1994; Darlington and Smith, 2000). The phenomenon represents a process of behavioral recovery following damage to the vestibular system that results in ocular motor and postural symptoms. In the case of unilateral vestibular deafferentiation, the symptoms are particularly dramatic, as a result of the severe imbalance in neuronal activity between the ipsilateral and contralateral vestibular nuclear complexes. This experimental condition indicates that inhibition of NOS results in a retardation of the vestibular compensation process, even if low levels of NOS activity are present in the uncompensated stage (Kitahara et al., 1997).

It has been suggested that the cellular mechanisms underlying vestibular adaptation are at first localized in the mammalian cerebellar cortex (Ito, 1982)—specifically in the flocculus, which receives vestibular and visual afferents. Damage to the vestibular system after labyrinthectomy affects NOS expression of unipolar brush cells (Kitahara et al., 1997) in the rat cerebellar flocculus, and changes in NOS activity in the whole cerebellum of guinea pig have also been reported (Paterson et al., 2000).

After vestibular peripheral damage, although the vestibulocerebellum is primarily affected (Nagao and Ito, 1991; Flugel et al., 1994; Li et al., 1995), we cannot exclude the involvement of the corpus cerebelli, wherein Purkinje cells play an important role in movement coordination and motor learning (Ito, 1982).

In the extensive literature on vestibular compensation (for review, see Dieringer, 1995), the frog has also been considered as a neurobiological model. The involvement of Purkinje cells following the stimulation of the eighth nerve has also been reported (Dieringer and Precht, 1979a, b).

Within the Purkinje cell layer of the corpus cerebelli, as regards the central projections of primary vestibular fibers in bullfrog, a small number of thin fibers (mossy fibers and climbing fibers) terminate on the neurons (Kuruvilla et al., 1985). Moreover, terminals of primary vestibulocerebellar afferents in the molecular layer appear to contact the main dendrites of the Purkinje cells in a manner similar to that of climbing fibers (Nieuwenhuys and Opdam, 1976).

To our knowledge, however, there have been no findings regarding the neurochemical changes occurring in situ in Purkinje cells of the corpus cerebelli following injury to the vestibular system.

Purkinje cells of adult animals do not normally exhibit NOS or NADPHd activity (Bredt and Snyder, 1989; Vincent and Kimura, 1992; Muñoz et al., 1996; Alonso et al., 2000; Steinbusch et al., 2000), but after traumatic or chemical injuries NOS expression can be induced (Wu, 2000).

To elucidate in situ the role of NO after unilateral eighth-nerve transection in the Purkinje cells of the corpus cerebelli, we chose to examine the frog cerebellum. There are two main reasons why this is a good model. First, it is small in size, and the analysis of the whole organ is simple, consequently, the comparison of the different areas, e.g., right and left or vestibulocerebellum and corpus cerebelli, is quite easy. The second and more important reason is that the regenerative capacity of nonmammalian vertebrates is higher than that of mammals (for review, see Larner et al., 1995). In the present research we considered both the degenerative and regenerative stages, through the study of early and late phases after neurectomy. Starting from 3 days after unilateral nerve transection, some morphological findings on the degeneration patterns in the granular layer of the frog corpus cerebelli are reported (Hillman, 1969; Sotelo, 1976).

MATERIALS AND METHODS

Three groups of male adult frogs (Rana esculenta) were considered: 1) operated frogs, 2) sham-operated controls, and 3) non-operated controls. The frogs in the first group were deeply anesthetized by immersion in 0.1% MS222 solution (Sandoz, Basel, Switzerland) and the right eighth cranial nerve was excised between Scarpa's ganglion and the brain stem. In the frogs of the second group, under anesthesia, the nerve was exposed but not cut.

After 3, 15, 30, and 60 days, operated animals (five per stage), non-operated controls (16; four per stage), and sham-operated (two per stage) were again anesthetized and perfused intracardially with Ringer solution followed by 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4, at room temperature. The entire brain was removed from all animals immediately and postfixed in the same fixative for 90 min at 4°C, cryoprotected in 30% sucrose, and frozen in liquid nitrogen. The cerebella were then cut with a cryostat into 12-μm-thick sections in the transversal plane in order to examine simultaneously the ipsilateral and contralateral sides to the lesion.

In addition, three operated frogs per stage and three non-operated controls were anesthetized and perfused intracardially with Ringer solution followed by 1% paraformaldehyde and 1% glutaraldehyde in 0.1 M PB, pH 7.2, at room temperature. The entire brains were immediately excised, immersed in the same fixative for 4 hr and then kept in PB overnight at 4°C. After postfixation in 1% OsO4 in PB for 1 hr, the brains were dehydrated in ethanol and embedded in Durcupan ACM (Fluka, Buchs, Switzerland). Then 1.5 μm semithin sections of the ipsilateral and contralateral cerebellum were stained with toluidine blue to estimate the density of Purkinje cells in the parasagittal planes, which show a more homogeneous thickness of the Purkinje cell layer.

NADPH Diaphorase Histochemistry

Adjacent cerebellar sections of all stages were incubated simultaneously for 1 hr in the dark in the following medium (according to Van Noorden and Frederiks, 1992): 0.1 M phosphate buffer, pH 7.4, containing 18% (W/V) polyvinyl alcohol, 0.5 mM β-NADPH (Sigma Chemical Co., St. Louis, MO), 5 mM nitro blue tetrazolium (NBT; Sigma) and 0.2% Triton X-100. After incubation, the sections were rinsed in 0.1 M phosphate buffer (pH 7.4) and then mounted in glycerin.

Control reaction for NADPHd histochemistry was performed by incubating sections after: 1) omission of NADPH; 2) omission of NBT; and 3) replacement of NADPH with NAD. In addition, some sections were preincubated with 1 mM L-NNA or 1 mM L-NAME dissolved in glycine-NaOH buffer (pH 8.5). Both L-NNA and L-NAME are specific inhibitors of NOS activity and, consequently, of NO production. No reactivity was present in these sections.

The intensity of staining of Purkinje neurons was determined by scanning the photographs of sections with a color scanner (AGFA ARCUS II) and then using Photoshop 5.0 for computerized image analysis. The color photographs were first converted into black-and-white photographs; the intensity of staining was expressed as the percentage of black on a gray scale (255 gray levels). All the intensity values obtained after control reaction omitting NADPH were subtracted from the corresponding intensity values obtained after specific NADPHd reaction. Data are mean ± S.D. The evaluations were made on each neuron of five sections for each individual, and 30 measurements were obtained for each section.

Immunocytochemistry

Cerebellar sections of all stages were incubated simultaneously in 3% H2O2 in 10% methanol in phosphate-buffered saline (PBS) for 7 min, 10% normal goat serum in 0.1 M Tris/HCl buffer, pH 7.6, for 15 min, followed by 1:500 rabbit polyclonal antibody raised against rat cerebellar NOS-I (Chemicon, Temecula, CA) in PBS overnight. A biotin-streptavidin-HRP complex (Histomark, Gaithersburg, MD) was used to reveal the sites of antigen/antibody reaction. The peroxidase reaction was performed with 0.05% 3,3′-diaminobenzidine tetrahydrochloride in 0.05M Tris-HCl buffer, pH 7.6, containing 0.01% H2O2, for 20 min. All incubations were at room temperature. The sections were dehydrated in ethanol, cleared in xylene, and mounted in Eukitt (Kindler GmbH, Freiburg, Germany).

For the control staining, some sections were incubated with normal rabbit serum or PBS instead of the primary antibody. No immunoreactivity was present in these sections. Details of the specificity of the primary antibody were provided by the commercial supplier.

Purkinje Cell Counts and Density

Sections were observed and photographed with a Zeiss Axioskop microscope. Counts were made on the corpus cerebelli of five sections per individual per stage. Negative Purkinje cells were counted separately from Purkinje cells that showed weak to intense reactivity. The labeled cells were expressed as a percentage of the total neuronal population. For each stage, we added all the values, since no significant differences were seen among animals.

The cellular linear density (Aherne and Dunnill, 1982) of the Purkinje cells was estimated in at least 20 semithin sections, for every animal and stage, by measuring the length of the boundary between the molecular layer and the internal granular layer by means of a graphic table on line with an Apple II C personal computer. The number of cells was then counted and the mean linear cell density per 1,000 μm was calculated.

Statistical Analysis

For measurements of NADPH diaphorase staining intensity, and for counts of NADPH diaphorase-labeled and NOS-immunoreactive Purkinje neurons, the time-dependent changes were tested on ipsilateral and contralateral areas with a one-way ANOVA procedure followed by Student's t-test.

Behavioral Observations

The effectiveness of the vestibular lesion after unilateral nerve transection was indicated by observations of the classical postural and locomotor deficits shown by the animals.

RESULTS

In the frog cerebellum, the Purkinje cell layer is not formed by a single row of cells; instead they are grouped in clusters of two or three rows between the molecular and granular layers (Sotelo, 1976).

NADPHd activity (Fig. 1) was detected in both the neuronal cytoplasm and fibers. NOS immunoreactivity was found to be concentrated in the neuronal soma (Fig. 2).

Figure 1.

NADPH diaphorase activity. A: 3-day-operated frogs. In the Purkinje cell layer on the ipsilateral side, several neurons are labeled (small arrows). In the granule cell layer (gcl), some very strongly NADPHd-stained, thickened fibers are present; also, varicosities are NADPHd-labeled (large arrows). B: 3-day-operated frog. On the contralateral side, few Purkinje neurons are NADPHd-labeled (small arrows); in the granule cell layer (gcl) NADPHd-stained fibers and several varicosities are present (large arrows). C: 15-day-operated frog. On the ipsilateral side, numerous NADPHd-stained Purkinje cells show an increased intensity of labeling (small arrows); in the granule cell layer (gcl) thickened and tortuous fibers appear strongly labeled (large arrows). D: 30-day-operated frog. A large part of the Purkinje neurons are NADPHd-stained; in the granule cell layer (gcl) varicosities and fine-beaded fibers are labeled (arrows). E: 60-day-operated frog. Purkinje neurons show weak NADPHd staining; in the granule cell layer (gcl), short-thickened (large arrows) and fine-beaded fibers (small arrows) are NADPHd-labeled. F: Control. A few Purkinje neurons are weakly NADPHd-labeled (arrows); in the granule cell layer (gcl) NADPHd staining is present in fibers of different calibers and varicosities. All scale bars: 50 μm.

Figure 2.

NOS immunoreactivity. A: 3-day-operated frog. On the ipsilateral side, numerous Purkinje neurons appear immunoreactive (arrows). B: 30-day-operated frog. Strong NOS immunoreaction is present in almost all Purkinje neurons. C: 60-day-operated frog. On the ipsilateral side, some of the Purkinje neurons show NOS-immunoreactivity (arrows). D: Control. In the Purkinje cell layer, only some neurons are immunoreactive (arrows). All scale bars: 50 μm.

The most interesting results concern the patterns of the Purkinje neurons. Figure 3 summarizes the trend of the two markers that labeled the Purkinje cell population, Figure 4 shows the percentage of Purkinje neurons labeled after the reactions, and Figure 5 shows the intensity values of NADPHd staining of the Purkinje cells. We considered non-operated animals as controls, since no differences were found between sham-operated and non-operated animals. Moreover, no differences were observed among controls of different stages, then the 60 days controls were considered.

Figure 3.

Schematic representation of the distribution of NADPHd-reactive (NADPHdr), NOSi Purkinje neurons in the frog cerebellum in the controls (C) and at 3 days (3d), 15 days (15d), 30 days (30d), 60 days (60d) after unilateral neurectomy. Black dots = labeled neurons; empty dots = negative neurons.

Figure 4.

Histogram of percentages of NADPHdr, NOSi Purkinje neurons at different times after unilateral neurectomy. Values (numerical values and counts are specified in the text) are expressed as average ± S.D. For NADPHd reactivity differences were significant (Student's t-test) between control–3 days (P < 0.001), 30–60 days (P < 0.001), 3–15 days (P < 0.01), 15–30 days (P < 0.05) on the ipsilateral side; and between control–3 days (P < 0.01), 3–15 days (P < 0.001), 30–60 days (P < 0.001), 60 days–control (P < 0.05) on the contralateral side. For NOS-immunoreactivity, differences were significant between control–3 days (P < 0.001), 60 days–control (P < 0.001), 30–60 days (P < 0.001), 3–15 days (P < 0.05) on the ipsilateral side; and between 3–15 days (P < 0.001), 30–60 days (P < 0.001), control–3 days (P < 0.01), 60 days–control (P < 0.05) on the contralateral side. Gray column = ipsilateral side area; black column = contralateral side area.

Figure 5.

Histogram of the intensity of NADPHd staining in Purkinje neurons at different times after unilateral neurectomy, expressed as the percentage of black in the gray scale. Differences were significant (Student's t-test) between control–3 days (P < 0.001), 30–60 days (P < 0.001), 3–15 days (P < 0.01) on the ipsilateral side; and between 3–15 days (P < 0.001) and 30–60 days (P < 0.001) on the contralateral side. Gray column = ipsilateral side area; black column = contralateral side area.

There were also some interesting results with regard to fiber reactivity to NADPHd in the granule cell layer (Fig. 1). On the ipsilateral side at 3 days after nerve transection some fibers were tortuous, thickened, and very strongly stained. These were also observed on the contralateral side at 15 days, when axons that appeared to be thin and beaded were also present on both sides. Short and thick, and fine-beaded fibers increased from 30 to 60 days.

NADPH Diaphorase Activity

At 3 days, on the ipsilateral side, the majority of Purkinje neurons (82%) were NADPHd-stained (Fig. 1A), compared to 29% on the contralateral side (Fig. 1B).

In 15-day-operated frogs, in the Purkinje cell layer on the ipsilateral side, a significant increase was observed not only in the number of NADPHd-labeled neurons (92%, significance P < 0.01 vs. 3 days), but also in the intensity of labeling (Fig. 1C). On the contralateral side, the number of NADPHd-labeled Purkinje cells increased remarkably (76%, significance P < 0.001 vs. 3 days), as did the intensity of their reaction.

At 30 days, on both the ipsilateral and contralateral sides, the pattern of NADPHd activity (number of positive cells and reaction intensity) was in general unvaried (85%, significance P < 0.05 vs. 15 days, and 80% of Purkinje neurons were labeled, respectively) (Fig. 1D) in comparison to 15-day-operated frogs.

At 60 days, on both the ipsilateral and contralateral sides, the number of NADPHd-labeled Purkinje neurons was drastically reduced (17%, significance P < 0.001 vs. 30 days; and 10%, significance P < 0.001 vs. 30 days, respectively) (Fig. 1E).

The 60-day-operated frogs showed a pattern similar to the controls (significance P < 0.05 vs. 60 days, on the contralateral side). In general, Purkinje cells showed no labeling, and weak NADPHd activity was observed in a few (about 20%) of these neurons (Fig. 1F). The significance between control–3 days was P < 0.001 on the ipsilateral side and P < 0.01 on the contralateral side.

Figure 5 compares the changes of intensity of NADPHd staining in Purkinje neurons on both the ipsilateral and contralateral sides. On the ipsilateral side, an increased staining was present 3 days after neurectomy (significance between control–3 days: P < 0.001), a higher labeling was observed at 15 days (significance between 3–15 days: P < 0.01) and 30 days, while on the contralateral side the increase of NADPHd staining was evident at 15 days (significance between 3–15 days: P < 0.001) and again at 30 days. At 60 days, on both the ipsilateral and contralateral sides, the intensity of NADPHd staining decreased (significance between 30–60 days: P < 0.001).

NOS Immunoreactivity

In 3-day-operated frogs, on the ipsilateral (Fig. 2A) and contralateral sides, the number of NOS-immunoreactive (NOSi) Purkinje neurons was 63% and 50%, respectively.

At 15 days, the number of labeled neurons was higher: about 80% (ipsilateral 78%, significance P < 0.05 vs. 3 days; contralateral 79%, significance P < 0.001 vs. 3 days). A similar percentage was found in 30-day-operated frogs (ipsilateral 82%, contralateral 83%) (Fig. 2B).

Surprisingly, at 60 days, on the ipsilateral side (Fig. 2C), a part of the Purkinje cells still present were NOSi (60%, significance P < 0.001 vs. 30 days); on the contralateral side, the percentage of NOSi Purkinje neurons was lower (50%, significance P < 0.001 vs. 30 days).

In the control (Fig. 2D), in the Purkinje cell layer, the immunoreactivity was weak in some neurons (about 40%, significance between 60 days–control: P < 0.001 on the ipsilateral side and P < 0.05 on the contralateral side; significance between control–3 days: P < 0.001 on the ipsilateral side and P < 0.01 on the contralateral side).

Purkinje Cell Density

Table 1 shows that the density of Purkinje cells, starting 3–15 days after neurectomy, had a decrease that was more evident at 30 days. A further diminished density was observed at 60 days on both sides. Interestingly, at 3 days the decreased cell density was observed on the ipsilateral side, while on the contralateral one it occurred at 15 days, when similar values were observed on both sides.

Table 1. Cell density in the Purkinje cell layer expressed as mean value ± S.D.
Days after neurectomyIpsilateral side area (x ± S.D.)Contralateral side area (x ± S.D.)
Control155 ± 10150 ± 10
3 days126 ± 34150 ± 29
15 days130 ± 16129 ± 17
30 days90 ± 9102 ± 16
60 days76 ± 882 ± 12

Behavioral Observations

After neurectomy of the eighth nerve, frogs assumed an asymmetrical posture in which the head was tilted toward the side of the lesion and the limbs were extended on the opposite side. This asymmetrical pose was maintained for about 3 days, after which the frogs recovered an almost normal posture—but only in static conditions. During locomotion, the frogs tended to turn again toward the damaged side. Neurectomized frogs recovered normal behavior, in both static and dynamic conditions, about 15 days after the unilateral vestibular neurectomy.

DISCUSSION

In the present study we show that unilateral nerve transection induces changes of NOS expression and activity in frog Purkinje neurons, with differences in the time-course responses of the ipsilateral and contralateral sides of the corpus cerebelli. In general, the markers indicate a critical period of induction of NOS from 3 to 30 days after neurectomy, followed by down-regulation.

NOS Induction After Axotomy

Transection of a peripheral nerve branch results in a great degeneration of fibers and neurons (Lieberman, 1974; Donoff, 1995). However, some damaged neurons can survive if the injured nerve is able to regrow and successfully reinnervate the appropriate target, with the formation of functionally efficient synapses (Aguayo et al., 1991; Welcher et al., 1991). The regeneration capacity of CNS neurons is based on differences in their molecular responses to injury, and particularly on their ability to upregulate specific proteins for neuron survival. This ability may have potential as a therapeutic agent in the treatment of human neurodegenerative disorders (Lindsay, 1994; Vaudano et al., 1998).

There are molecules like NO which are particularly complex from a functional point of view (Prast and Philippu, 2001). In fact, NO is a biological signaling molecule that is important in the physiology and pathophysiology of many different tissues (Schulz et al., 1995). NADPHd and NOS protein are inducible in term of de novo synthesis in traumatically injured neurons (Wu, 2000). After peripheral axotomy, NOS protein and NOS mRNA increase in some neuronal populations (Verge et al., 1992; Vizzard et al., 1995); moreover, NOS is up-regulated in axotomized motoneurons (Yu, 1994; Yu, 1997; Müller and Stoll, 1998; Ma et al., 2000).

Both vulnerability and type of lesion have to be considered when induction of NADPHd or NOS is studied in vivo. In the mammal cerebellum, Purkinje cells normally do not contain NOS (Vincent, 1994; Blottner et al., 1995; Vincent, 1996), but a lesion can induce NADPHd in the perikaria (Blottner et al., 1995; Saxon and Beitz, 1996). Increased NOSi and NADPHd activity were described in dorsal root ganglia on the ipsilateral side to peripheral axotomy (Vizzard et al., 1995), from 3 days to 4 weeks. Similar results were obtained after electrolyte lesions of the nucleus basalis magnocellularis (Sabbatini et al., 1999). Unilateral ligation of L5 and L6 spinal nerves results in an ipsilateral increase of NOS-like immunoreactivity (Meller and Gebhart, 1993).

In our research we have taken into account that Purkinje cells are the only output of the cerebellar cortex, and therefore analysis of this type of neuron can represent the cerebellum's response to injury.

In the frog cerebellum (Rana perezi, Xenopus laevis) in normal conditions, absence of NOS activity in the Purkinje cell population, or weak positivity in some of them, has been observed (Brüning and Mayer, 1996; Muñoz et al., 1996; Alonso et al., 2000).

In our study, unilateral nerve transection induced NADPHd activity and NOS protein in Purkinje cells of Rana esculenta. While the control animals had less than 20% of weakly NADPHd-positive cells, 3 days after neurectomy 82% of cells became labeled on the ipsilateral side. An increased reactivity was evident at 15 days, when up-regulation of NOS enzymatic activity also occurred on the contralateral side. A similar trend was observed for NOS-immunoreactivity.

These findings demonstrate that the first important result in situ is that the Purkinje cells, i.e., the cerebellar neurons that play a role in motor coordination and learning (Ito, 1982), of the corpus cerebelli are involved in vestibular compensation. Most of the reports in the literature concern the involvement of the vestibular nucleus or the vestibulocerebellum, such as the flocculus (Kitahara et al., 1997, 1999; Zheng et al., 2001), in the compensation process. As regards more specific studies on the effects of vestibular neurectomy in the frog, since the early 1970s Dieringer has studied the vestibular nucleus, spinal cord, and cerebellum. Dieringer (1995) reported the sole finding on the response of Purkinje neurons to stimulation of the eighth nerve; however, no mention was made of the Purkinje cells of the corpus cerebelli or vestibulocerebellum.

The cytochemical changes we observed are explained by the presence of vestibular fiber projections in the Purkinje cell layer of the corpus cerebelli (Kuruvilla et al., 1985). The ascending branch of the ventral root of the eighth nerve sends primary fibers (mossy fibers) to the granule cell layer of the auricular lobe of the ipsilateral and contralateral sides and to the ipsilateral acoustic area of the brain stem, which, in turn, is connected to the contralateral acoustic area. The small number of thin fibers which terminate within the Purkinje cell layer of the corpus cerebelli could represent either collaterals of mossy fibers or a population of climbing fibers. Moreover, secondary vestibular fibers (climbing fibers) from vestibular nuclei are distributed in the granule cell layer of the lobus vestibularis and corpus cerebelli, and in the molecular layer of both sides, terminating on the Purkinje cell dendrites. In turn, Purkinje cells axons terminate on the cerebellar and statoacoustic nuclei (Gregory, 1974; Nieuwenhuys and Opdam, 1976; Kuruvilla et al., 1985; Montgomery, 1988; ten Donkelaar, 1998).

In the field of research on the effects of axotomy, the increased NO production in our model (at different times on the ipsilateral and contralateral sides) is also a new finding. This can be explained by the anatomical distribution of fiber projections in the frog cerebellum (see above), and in particular by the interconnections between the two vestibular nuclei. These projections suggest that the ipsilateral side to the lesion is altered earlier (3 days after neurectomy) and that only secondarily, via vestibular nuclei, does the contralateral area show changes (at 15 days). Our findings point out that changes of NOS expression in Purkinje cells of the frog cerebellum are indicative of the time-course and dynamics of vestibular compensation (Markham and Yagi, 1984).

Moreover, the pattern of markers of the contralateral side suggest that it is advisable to maintain caution in the use of this area as a control in experiments dealing with cellular processes of degeneration and regeneration.

NOS/NADPHd induction was observed at different times on the ipsilateral and contralateral sides in other models, such as in the rat lateral geniculate nucleus, after cauterization of limbus-draining veins (Wang et al., 2000).

NOS involvement in vestibular compensation has been widely investigated (for review, see Darlington and Smith, 2000). In a study of the frog, Flugel et al. (1994) were the first to find that inhibition of NOS prevents functional restoration after vestibular lesion. Kitahara et al. (1997) showed that changes in NOS expression occur in the rat cerebellar flocculus after unilateral vestibular deafferentiation, and that injection of L-NAME into the flocculus inhibits the compensation process. Kitahara et al. (1999) also reported that NOS expression in the flocculus modulates the expression of c-Fos in the vestibular nucleus complex following unilateral vestibular deafferentiation.

The importance of NOS in vestibular compensation was demonstrated in the current study by a comparison of the time-courses of behavioral changes and the cytochemical changes of the Purkinje cells. After neurectomy of the eighth nerve, the frogs, as expected (Goltz, 1870; Beritoff, 1928), assumed an asymmetrical posture that was maintained for about 3 days, after which the frogs recovered an almost normal posture (but only in static conditions). Normal behavior, in both static and dynamic conditions, was observed about 15 days after the unilateral vestibular neurectomy.

Finally, with regard to the recovery of cell functionality, 30-day-transected frogs showed a steady pattern in both the percentage and reactivity of Purkinje cells. A re-establishment of the normal pattern was achieved at 60 days, as shown by the intensity of NADPHd reaction but not NOSi, which remained in a high percentage of neurons (mainly on the ipsilateral side).

A down-regulation of NOS expression at late stages of axotomy or following mechanical lesions was recently reported in the spinal motoneurons after complete regeneration of lizard tail, although NADPHd and NOS reactivities persisted in the majority of the spinal motoneurons (Cristino et al., 2000).

However, our findings indicate a discrepancy between the two histochemical reactions. NOSi persisted in a large part of the Purkinje neurons 60 days after neurectomy, whereas NADPHd labeling did not. A possible interpretation of this result is that NOS protein is still higher than in controls, but it is inactive being NADPHd reactivity absent. The excess of NOS suggests that 60 days after transection the recovery is still in a critical phase that can be better evaluated together with other features of frog cerebellar architecture. Discrepancies between NOS expression and NOS activity/NADPHd labeling have been detected in the vestibular nucleus following unilateral labyrinthectomy (Zheng et al., 2001), as well as in other neurobiological models (Gabbott and Bacon, 1993; Cooke et al., 1994; Pisu et al., 1999). Possible reasons for these discrepancies are: 1) there are different isoforms of NOS responsible for NADPHd activity, 2) the NOS molecule contains an NADPH binding site that is inactive after fixation, or 3) the NOS molecule does not have an NADPH binding site. Furthermore, conformational changes of the molecule affecting the immunoreactive responses cannot be excluded.

What Is the Role of NO After Axotomy?

The gaseous free radical NO, which is generated in biological tissues, is an important regulator of a broad range of functions (Moncada et al., 1991; Alonso et al., 2000; Vincent, 2000) and is considered a factor in stimulating neurotransmitter release in the CNS (Ohkuma and Katsura, 2001). Regarding the CNS, research on different neurobiological models from vertebrates and invertebrates has produced fundamental findings on the role of NO in the normal physiology of several areas. It is known that NO is involved in the behavior of animals, i.e., feeding processes, visual functionality, memory, and learning (Chapman et al., 1992; Moroz et al., 1993; Elphick et al., 1995, 1996; Müller, 1996, 1997; Seidel and Bicker, 1997; Conforti et al., 1999; Pisu et al., 1999).

In the cerebellum NO is a marker of synaptic plasticity (Vincent, 1996; Alonso et al., 2000). However, NO has a critical and complex role in pathophysiology and response to injury. The most important question concerns the cytotoxic or neuroprotective role of NO when NOS induction occurs. It is likely that both roles for NO coexist in the same CNS region, even in the same neuronal population, at different times following injury. This could involve other chemical properties of the CNS region (Grassi and Pettorossi, 2001).

The induction of NOS a few days after nerve transection coincides with the first signs of the system's plasticity (i.e., degeneration process). After unilateral neurectomy, cytoskeletal phosphorylated neurofilaments in some Purkinje neurons' soma were observed (Bernocchi et al., 1994), in agreement with other researchers' findings concerning axonal injury or nerve crush (Sinicropi and McIlwain, 1983; Rosenfeld et al., 1987; Gold et al., 1991). In some cells degenerative changes may be irreversible and may result in cell death (Vignola et al., 1992), as shown by the lower density of the Purkinje cell layer starting from 15 days. This means that the appearance of NOS in Purkinje cells at early stages after neurectomy may be linked to a neurotoxic effect of the molecule.

Increased NOS activity was also detected in vestibular nuclei following labyrinthectomy (Zheng et al., 2001); nevertheless, degeneration of neurons in these nuclei was not observed. However, several neurochemical changes have been observed in these nuclei, such as modifications of GABA immunoreactivity in the cat (Tighilet and Lacour, 2001) and transient increase of NMDA receptor subunit expression in the guinea pig (Sans et al., 1997). Neurochemical changes were found not only in neurons but also in glial cells. During the recovery from injury to the peripheral vestibular system, GFAP-immunoreactivity changes in the astroglia of the lateral vestibular nuclei in the cat were shown (Cass and Goshgarian, 1990). Increased levels of astroglial S100 immunoreactivity were also found in guinea pig after labyrinthectomy (Rickmann et al., 1995). Data indicate a role of glial cells in neuroplasticity during vestibular compensation.

At present, the different degrees of response to injury of Purkinje cells and vestibular nuclei neurons are quite intriguing, and further experimental investigations are necessary.

In addition to the changes of Purkinje neurons, starting from 3 days after neurectomy, large-caliber fibers, which are very strongly NADPHd-reactive, were present in the granule cell layer. These look like degenerating fibers, as reported previously in the frog cerebellum (Bernocchi et al., 1994) and in other areas of the vertebrate CNS (Kaplan and Clemente, 1981; Marbey and Browner, 1987; Oestreicher et al., 1988).

The fiber distribution in the granule cell layer of 60-day-operated frogs was restored, and thin and beaded NADPHd-reactive axons were present. These regenerative signs support the idea that the persistence of NOS immunolabeling in Purkinje neurons at 2 months after neurectomy provides evidence for the neurotrophic or neuroprotective function of this molecule.

The factors that govern the effects of NO on cell physiology are complex and not completely understood. They appear to be related to cell type and NO concentration, and to the production of other free radicals or molecules that are also expressions of the functional involvement of the neural cells (Murphy et al., 1993; Schulz et al., 1995; Whittle, 1995; Vincent, 1996; Peuchen et al., 1997; Steinbusch et al., 2000).

Acknowledgements

All experiments reported in this article were carried out in accordance with the guidelines of the law (116/92) regarding the care and use of laboratory animals. We thank Dr. Harsha Basudev for the English revision of the manuscript, and Mrs. P. Veneroni for technical assistance.

Ancillary