Neuronal regeneration in the cerebellum of adult teleost fish, Apteronotus leptorhynchus: guidance of migrating young cells by radial glia

Developmental Brain Research 130 (2001) 15–23 www.elsevier.com / locate / bres

Research report

Neuronal regeneration in the cerebellum of adult teleost fish, Apteronotus leptorhynchus: guidance of migrating young cells by radial glia ¨ Sorcha C. Clint, Gunther K.H. Zupanc* School of Biological Sciences, University of Manchester, 3.614 Stopford Building, Oxford Road, Manchester M13 9 PT, UK Accepted 29 May 2001

Abstract In contrast to mammals, adult fish exhibit an enormous potential to replace injured brain neurons by newly generated ones. In the present study, the role of radial glia, identified by immunostaining against fibrillary acidic protein (GFAP), was examined in this process of neuronal regeneration. Approximately 8 days after application of a mechanical lesion to the corpus cerebelli in the teleost fish Apteronotus leptorhynchus, the areal density of radial glial fibers increased markedly in the ipsilateral dorsal molecular layer compared to shorter survival times, or to the densities found in the intact brain or in the hemisphere contralateral to the lesion. This density remained elevated throughout the time period of up to 100 days examined. The increase in fiber density was followed approximately 2 days later by a rise in the areal density of young cells, characterized by labeling with the nuclear dye DAPI, in the ipsilateral dorsal molecular layer. Based on this remarkable spatio-temporal correlation, and the frequently observed close apposition of elongated young cells to radial glial fibers, we hypothesize that radial glia play an important role in the guidance of migrating young cells from their proliferation zones to the site of lesion where regeneration takes place.  2001 Elsevier Science B.V. All rights reserved. Theme: Development and regeneration Topic: Regeneration Keywords: Neuronal regeneration; Adult neurogenesis; Migration; Radial glia; Cerebellum; Teleost fish

1. Introduction While the existence of adult neurogenesis in some brain regions of non-primate mammalian species has been known for several decades [1–5,8,10,23–25,31,32,34], it was the recent discovery of newborn cells in the hippocampus of adult humans [12] and related species [17,18] that has sparked a tremendous interest in this phenomenon. Although establishing an important phenomenon, these studies have also revealed the limitations of the mammalian brain to generate new neurons during adulthood. As a consequence, replacement of neurons lost in the course of neurodegenerative diseases or injuries is usually impossible in mammals.

*Corresponding author. Tel.: 144-161-275-7278; fax: 144-161-7757126. E-mail address: [email protected] (G.K.H. Zupanc).

This observation has prompted a search for other vertebrate species capable of overcoming the limits imposed upon mammals. Among non-mammalian vertebrates, the ability to produce new neurons, as well as the potential to regenerate neural tissue after injuries, is especially pronounced in fish where it persists over the entire adult life (for reviews, see Refs. [28,53,59]). A detailed quantitative analysis in one of the best-examined model systems, the gymnotiform fish Apteronotus leptorhynchus, has revealed the production of new cells in nearly 100 different brain regions, with the number of cells produced within any 2-h period totaling, on average, 100 000 [61,66]. This corresponds to roughly 0.2% of the total number of cells in the adult brain of this species. Within the first few weeks of their life, these cells migrate from their proliferation zones to specific target areas where many of them differentiate into neurons [63]. Following their arrival, approximately half of the young cells are eliminated through apoptosis [45,46]. The other half

0165-3806 / 01 / $ – see front matter  2001 Elsevier Science B.V. All rights reserved. PII: S0165-3806( 01 )00193-6

16

S.C. Clint, G.K.H. Zupanc / Developmental Brain Research 130 (2001) 15 – 23

integrate into pre-existing neural networks and survive for the rest of the fish’s life [38,50,63]. Apoptotic cell death and continuous neurogenesis are also two key factors that enable fish to repair brain tissue rapidly and with high efficiency, even after extensive injuries. Apoptosis is used to eliminate cells damaged in the course of injury [64]. This is very different to the situation in the mammalian brain where most damaged cells undergo necrosis [27]. The cardinal feature of apoptosis, a ‘clean’ type of cell death, is that it avoids the side effects associated with necrosis, such as tissue inflammation and cavity formation at the lesion site ([57]; for review, see Ref. [26]). Neurogenesis, the other key factor, involves the replacement of the eliminated neurons by newly generated ones. These new neurons are recruited from two sources: cells specifically produced in response to the injury in specific proliferation zones [65]; and cells from a pool of continuously generated undifferentiated cells [61,63,65]. Unlike the involvement of apoptosis and neurogenesis in neuronal regeneration, nothing is known about how the young neurons migrate from their birthplace in the proliferation zones to their final destination at the injury site, which may be up to several millimeters away. It has been suggested that, in several regions of the embryonic mammalian brain, neuronal precursors migrate along the guiding processes of radial glia ([40–42,44,55]; for review, see Ref. [43]). A similar function of radial glia has been proposed for certain regions in the intact brain of adult non-mammalian vertebrates, such as birds [6,7,16,21,49], reptiles [36], amphibians [56], and fish [20,33,48] (Gheteu and Zupanc, in preparation). Therefore, in the present study, we examined the potential role of radial glia, identified by immunolabeling against glial fibrillary acidic protein (GFAP), in the process of neuron regeneration in the adult teleost fish, Apteronotus leptorhynchus.

2. Materials and methods

2.1. Animals Brown ghosts (Apteronotus leptorhynchus; Gymnotiformes, Teleostei) were supplied by tropical fish importers (Neil Hardy Aquatica and G.L. Aquatics). The fish were kept in 30- to 150-l tanks at temperatures of approximately 268C and pH values around 7, with water conductivity of 100–150 mS / cm under a 12-h light, 12-h dark photoperiod. They were fed daily on frozen mosquito larvae. A total of 39 fish were used in this study. Twenty of these fish were male, 16 were female, and three individuals could not be sexed, as determined by post mortem gonadal inspection. Body size ranged between 100 and 152 mm total length (mean: 125 mm; median: 122 mm) and 1.4 and 6.4 g body weight (mean: 3.7 g; median: 3.5 g); thus, all

the fish used can be regarded as adult. The relative gonadal weight, defined as fresh weight of gonads divided by body weight, of 30 fish was determined. It ranged between 0.0003 and 0.0075 in males (mean: 0.0022; median: 0.0014) and 0.0026 and 0.0180 in females (mean: 0.0010; median: 0.0111). This suggests that most of the individuals of this seasonally breeding species were sexually immature.

2.2. Application of cerebellar lesions All chemicals were purchased from Sigma (St Louis, MO), unless otherwise stated. Mechanical lesions were applied to the dorsal-most subdivision of the cerebellum, the corpus cerebelli (CCb), as described previously [60,64,65]. Briefly, individual fish were subjected to general anesthesia by immersion into aquarium water containing approximately 2% ethyl carbamate (‘urethane’) and to local anesthesia with 2% lidocaine solution (Caesar and Loretz, Hilden). Guided by landmarks on the fish’s head, a 3-mm deep stab wound was created with a sterile scalpel (no. 11, Feather, Osaka), such that an approximately 1-mm deep cut traveling in parasagittal direction within the CCb resulted. The wound on the head was covered by applying Histoacryl (B. Braun, Melsungen), and the fish were returned to oxygenated aquarium water to recover. They were, then, transferred to isolation tanks. All experiments were carried out in accordance with the regulations of the Animals (Scientific Procedures) Act 1986.

2.3. Isolation of brain tissue After a post-injury survival period varying from 3 h to 100 days, the fish were killed by immersion into a 1.5% solution of 3-aminobenzoic acid ethyl ester (‘MS-222’) in aquarium water. They were, then, intracardially perfused with a flush solution consisting of 0.1 M phosphate buffer (PB) supplemented with 0.9% NaCl, 67 mg / l heparin sodium salt (186 000 IU / g; Serva Feinbiochemica, Heidelberg), and 10 ml / l of 2% lidocaine. After all blood had been washed out, the perfusion was continued with 2% paraformaldehyde (E. Merck, Darmstadt) in 0.1 M PB. The brains were removed and post-fixed in the same fixative solution for 1 h at 48C and cryoprotected in 1 M sucrose in phosphate-buffered saline (PBS) for 24–48 h at 48C. After embedding in a 1:1 mixture of Aqua-Mount (Lerner Laboratories, Pittsburgh) and Tissue Tek O.C.T. Compound (Sakura, Finetek), transverse 15-mm sections were cut on a cryostat and thaw-mounted on to gelatin / chrome-alum coated slides.

2.4. Glial fibrillary acidic protein ( GFAP) immunohistochemistry Detection of GFAP-positive structures was carried out by employing an immunohistochemical protocol. The

S.C. Clint, G.K.H. Zupanc / Developmental Brain Research 130 (2001) 15 – 23

frozen sections were dried in a desiccator for 90 min at room temperature (RT) and rehydrated in three changes of a wash buffer consisting of 0.1 M PB with 0.5 M NaCl. In order to permeabilize the tissue and block unspecific binding sites, the sections were treated for 30 min at RT with 10% normal goat serum (Gibco BRL), 1% bovine serum albumin (Serva Feinbiochemica), 1% teleostean gelatine, and 0.3% Triton X-100 in wash buffer. They were incubated with a 1:400 dilution of mouse-anti-GFAP, clone G-A-5, in antibody diluent, overnight at 48C. The antibody diluent consisted of 1% bovine serum albumin and 0.3% Triton X-100 in wash buffer. Unbound antibody was removed by three rinses for 3 min each in wash buffer. Antigenic sites were visualized by incubating the sections with a 1:200 dilution in antibody diluent of a secondary goat-anti-mouse IgG (H1L) antibody conjugated to the ¨ biochemifluorescent dye Cy3 (Dianova Gesellschaft fur sche, immunologische und mikrobiologische Diagnostik, Hamburg) for 90 min at RT. After three rinses in wash buffer for 3 min each, the sections were counterstained with the nuclear dye 49,6diamidino-2-phenylindoledihydrochloride (DAPI) by incubating them in 2 mg / ml of this nuclear dye in PBS for 3 min at RT (for details, see Ref. [58]). Finally, the sections were washed three times for 3 min each in PBS, embedded in polyvinyl alcohol, average molecular weight 30 000– 70 000, containing n-propyl gallate (for details, see Ref. [58]), coverslipped, and stored at 2208C. Control experiments to test the specificity of the antiGFAP antibody included: (a) pre-adsorption of the antiGFAP antiserum with GFAP, (b) omission of the primary antibody, (c) incubation of the sections with IgG instead of application of the anti-GFAP antibody, and (d) incubation of the sections with normal mouse serum instead of application of the anti-GFAP antibody. Each of these treatments resulted in absence of any specific labeling, while positive control sections processed in parallel showed normal immunostaining.

2.5. Microscopic analysis and data analysis The GFAP-stained sections were examined under epifluorescence with a fluorescence microscope (Zeiss Axioskop) using the appropriate filter sets. Photographs were taken using a Zeiss MC80 camera with Kodak T-Max 400 film. Images were digitized from the negatives employing a Digital Film scanner (UY-577 Sony) with Foto Station 3.5.1 software. Further image processing was done using Adobe Photoshop, release 5.0 (Adobe Systems Inc.). For quantitative analysis of the GFAP-labeled fibers, a total of 180 sections were used in the 39 fish killed after various post lesion survival times. The number of sections analyzed per fish ranged from 2 to 10 (mean: 5; median: 4). Labeled fibers were examined in three parts of the dorsal molecular layer (mol) of CCb, namely, the site of lesion (defined as a zone stretching 50 mm laterally and 50

17

mm medially from the knife cut); the remaining ipsilateral region of CCb-mol (defined as the ipsilateral hemisphere minus the site of the lesion); and the hemisphere contralateral to the site of the lesion within CCb-mol. Labeled fibers and the outlines of the respective areas were drawn at a magnification of 3524 and 3131, respectively, by means of a camera lucida and digitized using a Flatbed Scanner (Epson GT-9500) and Adobe Photoshop. In addition, DAPI-labeled nuclei were counted in the same three parts of the CCb-mol, at a magnification of 3400. Areal and fiber length measurements were performed using Scion Image for Windows (Beta 3b, 1998, Scion Corporation). For determination of the areal densities of both fibers and DAPI-stained nuclei, the total length of GFAP positive fibers and the total number of DAPI-stained nuclei found in each part of the CCb-mol were divided by the area of the respective part. Statistical tests were performed using SPSS, release 7.5.1 (SPSS Inc.) using a Pentium-II computer.

3. Results

3.1. Morphology GFAP-labeled fibers were most conspicuous at the midline, and in the vicinity of the midline, in CCb. They formed a densely packed fiber bundle running dorsally along, and parallel to, the midline. Shortly before reaching the dorsal surface of the brain, these fascicles curved laterally and continued to travel for up to several hundred micrometers in either hemisphere (Fig. 1). These midline fibers were thick (approximately 1–1.5 mm in diameter), unbranched, and relatively straight. In the dorsal half of the CCb-gra, ramified, short (approximately 10 mm in length) fiber fragments oriented in all directions were observed. Clusters of short, ramified fibers were also present in the dorsomedial and ventromedial regions, where CCb-gra protrudes into CCb-mol to form ‘tips’ at the midline [63]. Near the site of the lesion, at the various time points examined (see below), the entire gamut of fiber types was found (Fig. 1). Some fibers were relatively thick, straight, and unbranched, whereas others exhibited a relatively thin, short, and ramified morphology. Comparison of the distribution of the GFAP-labeled radial glial fibers with that of the DAPI-stained nuclei frequently revealed a close apposition of cells with the fibers within the dorsal CCb-mol (Fig. 2). These nuclei, typically, exhibited an elongated shape, with the major axis often being three to four times longer than the minor axis.

3.2. GFAP-labeling as a function of survival time The density and distribution of GFAP-positive fibers were strongly dependent upon survival time following application of the lesion to CCb (Fig. 3). At the site of the

18

S.C. Clint, G.K.H. Zupanc / Developmental Brain Research 130 (2001) 15 – 23

Fig. 1. Transverse section through the corpus cerebelli 10 days after application of a mechanical lesion. The path of the lesion is indicated by arrowheads. The density of GFAP-labeled fibers is markedly higher at the site of the lesion and in the remaining ipsilateral dorsal molecular layer than in the contralateral dorsal molecular layer. The GFAP-stained radial glial fibers are also prominent at the midline (arrow), where they extend over considerable distances laterally in each hemisphere; d, dorsal molecular layer; gra, granule cell layer; v, ventral molecular layer.

lesion, the mean areal density of GFAP-labeled fibers, expressed as total fiber length per unit area, was relatively low both in intact fish (arbitrarily assigned the post-lesioning survival time ‘0 h’) and in lesioned fish after survival

Fig. 2. Double exposure of GFAP-labeled radial glial fibers (red) and DAPI-stained nuclei (blue) in the dorsal molecular layer of the corpus cerebelli. Several of these nuclei (indicated by arrowheads) are closely apposed to the fiber bundle consisting of a number of individual radial glial fibers. The arrow points to a cell that obviously has just become attached to, or detached from, the fiber.

periods of 3, 12, 24, 48, 72, and 120 h (overall mean of these two groups of 3.9 mm / mm 2 ) (Fig. 3A). However, 8 days after application of the lesion, the density showed a notable increase (mean of 34.9 mm / mm 2 ). The fiber density remained elevated and increased further throughout the remaining post-injury time points (overall mean of 31.8 mm / mm 2 at 10 and 15 days; overall mean of 56.0 mm / mm 2 at 20, 21, and 30 days; overall mean of 97.4 mm / mm 2 at 45, 60, and 100 days). The difference in fiber density at 0–120 h versus the density at all the following survival times was statistically significant at P,0.001 (Mann–Whitney U-test, two-tailed; U58; n 1 517 vs. n 2 5 22 fish). In the remaining ipsilateral region of the CCb, the temporal pattern of GFAP-stained fibers was similar to that seen at the site of the lesion (Fig. 3B). After survival periods of 0–120 h, the areal density of fibers was low (overall mean of 4.8 mm / mm 2 ). The density was increased at 8–100 days (overall mean of 16.7 mm / mm 2 ). However, the magnitude of the increase was not as pronounced as at the site of the lesion. The difference in fiber density at 0–120 h versus the density at all the following survival times was statistically significant at P,0.001 (Mann– Whitney U-test, two-tailed; U514; n 1 517 vs. n 2 522 fish). In the hemisphere contralateral to the site of the lesion, after survival periods of 0–120 h, the areal density of labeled fibers (overall mean of 5.1 mm / mm 2 ) was comparable to that seen in the ipsilateral hemisphere (overall mean

S.C. Clint, G.K.H. Zupanc / Developmental Brain Research 130 (2001) 15 – 23

19

Fig. 3. Areal densities of lengths of GFAP-labeled fibers (A–C) and DAPI-stained nuclei (A9–C9) in the corpus cerebelli after various post-lesioning survival times. (A, A9) Site of lesion. (B, B9) Remaining ipsilateral side. (C, C9) Side contralateral to the lesion. The values plotted are mean areal densities of GFAP-labeled fibers / DAPI-labeled nuclei of individual fish. The survival time of ‘0 h’ indicates the results of unlesioned fish. The vertical bars represent S.E. Note differences in scaling in the three sub-figures.

of 4.8 mm / mm 2 ) (Fig. 3C). The density of fibers was increased slightly at 8–100 days (overall mean of 9.5 mm / mm 2 ). However, this increase in the density of fibers at 0–120 h versus the density at all the following survival times was statistically not significant (P.0.05, Mann– Whitney U-test, two-tailed; U5119; n 1 517 vs. n 2 522 fish).

3.3. Quantitative analysis of DAPI-labeled nuclei The density and distribution of DAPI-labeled nuclei were also strongly dependent upon post-injury survival time (Fig. 39). At the site of the lesion, the density of labeled nuclei was low after survival times of 0, 3, 12, 24, 48, 72, 120, and 192 h (overall mean of 9.7 / 1000 mm 2 )

20

S.C. Clint, G.K.H. Zupanc / Developmental Brain Research 130 (2001) 15 – 23

(Fig. 3A9). Ten days after application of the injury, a notable rise in the density of labeled nuclei was observed. The density remained elevated until 30 days post lesion (overall mean of 40.0 / 1000 mm 2 at 10–30 days). Subsequently, the density of DAPI-stained nuclei was reduced at 45, 60, and 100 days (overall mean of 19.5 / 1000 mm 2 ). The difference in density of DAPI-labeled nuclei at 0–192 h and 45–100 days versus the density at 10–30 days was statistically significant at P,0.001 (Mann–Whitney U-test, two-tailed; U526; n 1 523 vs. n 2 516 fish). In the remaining ipsilateral hemisphere of the CCb, the density of labeled nuclei was comparatively low throughout the time-points examined (Fig. 3B9). The stained nuclei density was particularly low at the survival times of 0–192 h (overall mean of 5.6 / 1000 mm 2 ). The density was increased in the period of 10–100 days (overall mean of 10.2 / 1000 mm 2 ). This increase in density in DAPI-labeled nuclei density at 0–192 h versus the density at all the following survival times was statistically significant at P,0.001 (Mann–Whitney U-test, two-tailed; U576; n 1 5 19 vs. n 2 520 fish). In the unlesioned, contralateral, hemisphere of the CCb, and in the unlesioned brain, the DAPI-labeled nuclei densities were similar to those found in the remaining ipsilateral part of the CCb (Fig. 3C9). On average, the density was 7.4 / 1000 mm 2 throughout all the time points examined. There appeared to be no systematic change in areal density of DAPI-labeled nuclei with survival time.

4. Discussion

4.1. Effect of cerebellar lesions on radial glia In the cerebellum of the teleost fish Apteronotus leptorhynchus, radial glial fibers, identified by immunostaining against GFAP, persist beyond embryonic stages of development (Gheteu and Zupanc, in preparation). The present study was aimed at examining whether the density and distribution of these fibers changes in response to mechanical lesions applied to the corpus cerebelli, a readily accessible cerebellar subdivision. Our analysis demonstrated a localized increase in the density of radial glial fibers at and near the site of the lesion in the ipsilateral dorsal molecular layer, starting at around 8 days after application of the injury. The densities indicated at these later time points are likely to be rather conservative, since especially at very high fiber densities it was difficult to resolve all individual fibers. There was no indication of the fiber density returning to background levels within 100 days post lesion. Also, no significant change in fiber density after lesioning was observed in the contralateral hemisphere. Similar to the density of radial glial fibers, the density of DAPI-stained nuclei was increased at and near the site of the lesion in the dorsal molecular layer in response to the

cerebellar lesions. However, this increase occurred slightly later (approximately 10 days post lesion) than the increase in the density of radial glial fibers (approximately 8 days post lesion). Moreover, a return to background levels was observed at about 45 days post lesion. In contrast, in the remaining ipsilateral dorsal molecular layer the density of DAPI-stained nuclei remained elevated, and even after post-lesion survival times as long as 100 days no marked decline was visible. No significant change in the density of DAPI-stained nuclei occurred in the dorsal molecular layer contralateral to the site of the lesion.

4.2. Identification of radial glia Radial glia were identified both by immunostaining against GFAP and by morphological characteristics. GFAP is an intermediate filament protein characteristically found in the cytoplasm of astrocytes [9,11]. In mammals, GFAP appears to be characteristically expressed by mature radial glia, whereas immature radial glia label for vimentin only [52]. At later stages of embryonic development, or early stages of postnatal development, the mature radial glia are transformed into astrocytes [15,19,29,39,44,51,52]. Conversely, in non-mammalian vertebrates, such as fish [20] (Gheteu and Zupanc in preparation), amphibians [37,56], reptiles [30,35,54], and birds [6] radial glia persist well into adulthood. The notion that the GFAP-positive fibers observed in the present study are radial glial fibers, is also supported by their morphological characteristics. Radial glia are specialized non-neuronal cells which have a cell body located in the ventricular zone, and one or more long processes extending, in a radial fashion, to the outer surface of the brain. Similarly, the glial processes labeled in our investigation extended over several hundreds of micrometers. Moreover, they often originated in the proliferation zones of CCb-mol and radiated outwards towards the pial surface of the brain.

4.3. Identification of young cells In the brain of adult gymnotiform fish, generation of new cells has been demonstrated through incorporation of radioactively labeled thymidine or the thymidine analogue 5-bromo-29-deoxyuridine (BrdU) into replicating DNA during the S phase of mitosis [38,50,61–63,65,66]. As shown by this approach, in the corpus cerebelli of Apteronotus leptorhynchus, new cells are produced continuously and at high rate in the dorsal and ventral molecular layers at and near the midline [61,63]. From there, the young cells migrate, within a few weeks following their generation, into the granule cell layers. However, as BrdU is available for incorporation into newly synthesized DNA for just about 4 h following a single intraperitoneal injection of the thymidine analogue [61], this method reveals only a fraction of the young cells actually present

S.C. Clint, G.K.H. Zupanc / Developmental Brain Research 130 (2001) 15 – 23

in the molecular layer. Considering the number of cells labeled after a single injection of BrdU and the time course of their migration into the granule cell layer, as well as the total number of cells present in the molecular layer as revealed by Nissl staining, it appears reasonable to assume that the vast majority of DAPI-stained nuclei in the molecular layer of the corpus cerebelli belong to young cells not older than a few days. In the present study, we have, therefore, used the DAPI staining of nuclei to characterize the density and distribution of young cells in the molecular layer.

21

molecular layer of the remaining ipsilateral hemisphere continues to remain elevated for all following time points, including 100 days post lesion. Since the path of lesion created by our lesioning paradigm is usually difficult to detect at survival times longer than approximately 30 days post lesion, a considerable degree of regeneration appears to occur in the first few weeks following an injury. Nevertheless, it is possible that an increased number of young cells is still directed to the site of injury for further wound healing at these later time points.

4.5. Possible origin of radial glia 4.4. Role of radial glia in neuronal migration After application of mechanical lesions to CCb of adult Apteronotus leptorhynchus, new cells are generated both in discrete proliferation zones at the midline of CCb-mol and in the dorsal CCb-mol adjacent to the lesion site [65]. As we have demonstrated in the present study, it is these areas where the largest increase in the density of radial glial fibers was observed in response to the injury. The course of these fibers appears to delineate the path taken by the young cells when they migrate from the proliferation zones to the injury site. This is in agreement with our finding that the DAPI-labeled nuclei — supposedly belonging to the young cells — occur at much higher densities in the ipsilateral hemisphere, particularly at and near the site of the lesion, than in the contralateral hemisphere of the dorsal part of CCb-mol. We, therefore, propose that the radial glial fibers guide the young cells from their site of origin to their target site. This hypothesis is supported by the observation that elongated DAPI-labeled nuclei could, frequently, be found in close apposition to radial glial fibers. Neuroblasts with a similar elongated morphology, also directly apposed to radial glial fibers, have been described in various brain regions during embryonic development [13,14,40–42,55]. Further support for a guidance function of the radial glial fibers in the regenerating adult cerebellum of Apteronotus leptorhynchus is provided by the time course of the emergence of the radial glial fibers and the young cells. The density of radial glia fibers starts to increase approximately 8 days post injury, while the density of DAPIlabeled cells follows around 2 days later. As a consequence, radial glial fibers define the migrational path before the young cells arrive at the site of the injury. Obviously, the presence of the radial glial fibers before the arrival of the young cells is a prerequisite for the validity of our hypothesis that the radial glia exert a guidance function in the lesioned cerebellum. The presumptive guidance of young cells appears to persist for long periods of time after the injury. Although the density of DAPI-labeled nuclei in the dorsal CCb-mol at the site of the lesion returns to background level after approximately 45 days post injury, both the density of radial glial fibers and DAPI-labeled nuclei in the dorsal

The mechanism by which the increase in the density of radial glial fibers is mediated in response to cerebellar lesions remains elusive. Mitotic radial glia have been identified during embryonic and postnatal development in the rat brain [22] and during optic nerve regeneration in the adult optic tectum of goldfish [47]. Alternatively to a similar mitotic proliferation, a ‘quiescent’ population of radial glial may be present also in the intact brain of Apteronotus leptorhynchus. These cells may become detectable only after they start up-regulating GFAP expression in response to injury. These two alternative hypotheses will have to be tested in future experiments.

Acknowledgements We are grateful to Gerhard Engler for his help using the computer software. Alina Gheteu kindly provided immunostained brain sections of the two control fish analyzed in the present study. Jonathan R. Banks, Derek Dunn, Marianne M. Zupanc, and the two anonymous referees made helpful comments on the manuscript. This investigation received financial support by grants from the Royal Society and the Leverhulme Trust to GKHZ, as well as by a Postgraduate Studentship from the Biotechnology and Biological Sciences Research Council to SCC.

Note added in proof In addition to providing guidance cues for the migrating young cells, radial glia may also exert other functions in the development of the neurons born during adulthood. A recent study employing the same lesioning paradigm in Apteronotus leptorhynchus as used in the present investigation has demonstrated a dramatic up-regulation of the neuropeptide somatostatin 12–24 h post lesion [67]. Since this transient expression of somatostatin ceases before the young neurons are guided to the site of the lesion, it is unlikely that this phenomenon is related to the scaffolding function of radial glia. Rather, the up-regulation of somatostatin may relate to the increase in cell proliferation and the differentiation of these cerlls into granule cell

22

S.C. Clint, G.K.H. Zupanc / Developmental Brain Research 130 (2001) 15 – 23

neurons, as these events take place within the first 10 days following the injury.

References [1] J. Altman, Are new neurons formed in the brains of adult mammals?, Science 135 (1962) 1127–1128. [2] J. Altman, Autoradiographic investigation of cell proliferation in the brains of rats and cats, Anat. Rec. 145 (1963) 573–591. [3] J. Altman, Autoradiographic and histological studies of postnatal neurogenesis. III. Dating the time of production and onset of differentiation of cerebellar microneurons in rats, J. Comp. Neurol. 136 (1969) 269–294. [4] J. Altman, Autoradiographic and histological studies of postnatal neurogenesis. IV. Cell proliferation and migration in the anterior forebrain, with special reference to persisting neurogenesis in the olfactory bulb, J. Comp. Neurol. 137 (1969) 433–458. [5] J. Altman, G.D. Das, Autoradiographic and histological evidence of postnatal hippocampal neurogenesis in rats, J. Comp. Neurol. 124 (1965) 319–336. [6] A. Alvarez-Buylla, D.R. Buskirk, F. Nottebohm, Monoclonal antibody reveals radial glia in adult avian brain, J. Comp. Neurol. 264 (1987) 159–170. [7] A. Alvarez-Buylla, F. Nottenbohm, Migration of young neurons in adult avian brain, Nature 335 (1988) 353–354. [8] S.A. Bayer, J.W. Yackel, P.S. Puri, Neurons in the rat dentate gyrus granular layer substantially increase during juvenile and adult life, Science 216 (1982) 890–892. [9] A. Bignami, L.F. Eng, D. Dahl, C.T. Uyeda, Localization of the glial fibrillary acidic protein in astrocytes by immunofluorescence, Brain Res. 43 (1972) 429–435. [10] F.S. Corotto, J.A. Henegar, J.A. Maruniak, Neurogenesis persists in the subependymal layer of the adult mouse brain, Neurosci. Lett. 149 (1993) 111–114. [11] L.F. Eng, J.J. Vanderhaeghen, A. Bignami, B. Gerstl, An acidic protein isolated from fibrous astrocytes, Brain Res. 28 (1971) 351–354. ¨ [12] P.S. Eriksson, E. Perfilieva, T. Bjork-Eriksson, A.M. Alborn, C. Nordborg, D.A. Peterson, F.H. Gage, Neurogenesis in the adult human hippocampus, Nat. Med. 4 (1998) 1313–1317. [13] U.E. Gasser, M.E. Hatten, Central nervous system neurons migrate on astroglial fibers from heterotypic brain regions in vitro, Proc. Natl. Acad. Sci. USA 87 (1990) 4543–4547. [14] U.E. Gasser, M.E. Hatten, Neuron–glia interactions of rat hippocampal cells in vitro: glial-guided neuronal migration and neuronal regulation of glial differentiation, J. Neurosci. 10 (1990) 1276– 1285. [15] G.T. Ghooray, G.F. Martin, Development of radial glia and astrocytes in the spinal cord of the North American opossum (Didelphis virginiana): an immunohistochemical study using anti-vimentin and anti-glial fibrillary acidic protein, Glia 9 (1993) 1–9. [16] S.A. Goldman, V. Lemmon, S.S.-M. Chin, Migration of newly generated neurons upon ependymally derived radial guide cells in explant cultures of the adult songbird forebrain, Glia 8 (1993) 150–160. [17] E. Gould, B.S. McEwen, P. Tanapat, L.A.M. Galea, E. Fuchs, Neurogenesis in the dentate gyrus of the adult tree shrew is regulated by psychosocial stress and NMDA receptor activation, J. Neurosci. 17 (1997) 2492–2498. ¨ [18] E. Gould, P. Tanapat, B.S. McEwen, G. Flugge, E. Fuchs, Proliferation of granule cell neurons in the dentate gyrus of adult monkeys is diminished by stress, Proc. Natl. Acad. Sci. USA 95 (1998) 3168–3171. [19] M. Hirano, J.E. Goldman, Gliogenesis in rat spinal cord: evidence

[20]

[21]

[22]

[23] [24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32] [33]

[34]

[35]

[36]

[37]

[38]

[39]

for origin of astrocytes and oligodendrocytes from radial precursors, J. Neurosci. Res. 21 (1988) 155–167. ´ ´ Astroglial architecture of the carp (Cyprinus carpio) M. Kalman, brain as revealed by immunohistochemical staining against glial fibrillary acidic protein (GFAP), Anat. Embryol. (Berl.) 198 (1998) 409–433. ´ ´ A.D. Szekely, ´ M. Kalman, A. Csillag, Distribution of glial fibrillary acidic protein and vimentin-immunopositive elements in the developing chicken brain from hatch to adulthood, Anat. Embryol. (Berl.) 198 (1998) 213–235. Y. Kamel, N. Inagaki, M. Nishizawa, O. Tsutsumi, Y. Taketani, M. Inagaki, Visualization of mitotic radial glia lineage cells in the developing rat brain by Cdc2 kinase-phosphorylated vimentin, Glia 23 (1998) 191–199. M.S. Kaplan, Neurogenesis in the 3-month-old rat visual cortex, J. Comp. Neurol. 195 (1981) 323–338. M.S. Kaplan, D.H. Bell, Neuronal proliferation in the 9-month-old rodent: radioautographic study of granule cells in the hippocampus, Exp. Brain Res. 52 (1983) 1–5. M.S. Kaplan, J.W. Hinds, Neurogenesis in the adult rat: electron microscopic analysis of light radioautographs, Science 197 (1977) 1092–1094. ´ C.M. Winterford, B.V. Harmon, Anatomical J.F.R. Kerr, G.C. Gobe, methods in cell death, in: L.M. Schwartz, B.A. Osborne (Eds.), Cell Death, Academic Press, San Diego, CA, 1995, pp. 1–27. J.F.R. Kerr, J. Searle, B.V. Harmon, C.J. Bishop, Apoptosis, in: C.S. Potten (Ed.), Perspectives on Mammalian Cell Death, Oxford University Press, Oxford, 1987, pp. 93–128. A.J. Larner, A.R. Johnson, R. Keynes, Regeneration in the vertebrate central nervous system: phylogeny, ontogeny, and mechanisms, Biol. Rev. 70 (1995) 597–619. P. Levitt, P. Rakic, Immunoperoxidase localization of glial fibrillary acidic protein in radial glial cells and astrocytes of the developing rhesus monkey brain, J. Comp. Neurol. 193 (1980) 815–840. M. Lizzari, V. Franceschini, F. Ciani, Glial fibrillary acidic protein and vimentin in radial glia of Ambystoma mexicanum and Triturus carnifex: an immunocytochemical study, J. Hirnforsch. 38 (1997) 187–194. C. Lois, A. Alvarez-Buylla, Proliferating subventricular zone cells in the adult mammalian forebrain can differentiate into neurons and glia, Proc. Natl. Acad. Sci. USA 90 (1993) 2074–2077. C. Lois, A. Alvarez-Buylla, Long-distance neuronal migration in the adult mammalian brain, Science 264 (1994) 1145–1148. ´ Expression of a M.J. Manso, M. Becerra, M. Becerra, R. Anadon, low-molecular-weight (10 kDa) calcium binding protein in glial cells of the brain of the trout (Teleostei), Anat. Embryol. (Berl.) 196 (1997) 403–416. ˇ Z. Lodin, An autoradiographic study of DNA synthesis in V. Mares, adolescent and adult mouse forebrain, Brain Res. 76 (1974) 557– 561. ´ M. Monzon-Mayor, C. Yanes, J. De Barry, C. Capdevilla-Carbonell, J. Renau-Piqueras, G. Tholey, G. Gombos, Heterogeneous immunoreactivity of glial cells in the mesencephalon of a lizard: a double labeling immunohistochemical study, J. Morphol. 235 (1998) 109– 119. ´ J. Nacher, C. Ramırez, J.J. Palop, A. Molowny, J.A. Luis de la ´ ´ Radial glia and cell debris removal during Iglesia, C. Lopez-Garcıa, lesion-regeneration of the lizard medial cortex, Histol. Histopathol. 14 (1999) 89–101. C. Naujoks-Manteuffel, G. Roth, Astroglial cells in a salamander brain (Salamandra salamandra) as compared to mammals: a glial fibrillary acidic protein immunohistochemistry study, Brain Res. 487 (1989) 397–401. R. Ott, G.K.H. Zupanc, I. Horschke, Long-term survival of postembryonically born cells in the cerebellum of gymnotiform fish, Apteronotus leptorhynchus, Neurosci. Lett. 221 (1997) 185–188. S.K. Pixley, J. de Vellis, Transition between immature radial glia and

S.C. Clint, G.K.H. Zupanc / Developmental Brain Research 130 (2001) 15 – 23

[40] [41]

[42] [43]

[44]

[45]

[46]

[47] [48]

[49] [50]

[51]

[52]

[53]

mature astrocytes studied with a monoclonal antibody to vimentin, Brain Res. 317 (1984) 201–209. P. Rakic, Guidance of neurons migration to the fetal monkey neocortex, Brain Res. 33 (1971) 471–476. P. Rakic, Neuron–glia relationship during granule cell migration in developing cerebellar cortex: a Golgi and electronmicroscopic study in Macacus rhesus, J. Comp. Neurol. 141 (1971) 283–312. P. Rakic, Mode of cell migration to the superficial layers of fetal monkey neocortex, J. Comp. Neurol. 145 (1972) 61–83. P. Rakic, Radial glial cells: scaffolding for brain construction, in: H. Kettenmann, B.R. Ransom (Eds.), Neuroglia, Oxford University Press, New York, 1995, pp. 746–762. D.E. Schmechel, P. Rakic, A Golgi study of radial glia cells in the developing monkey telencephalon: morphogenesis and transformation into astrocytes, Anat. Embryol. (Berl.) 156 (1979) 115–152. J. Soutschek, G.K.H. Zupanc, Apoptosis as a regulator of cell proliferation in the central posterior / prepacemaker nucleus of adult gymnotiform fish, Apteronotus leptorhynchus, Neurosci. Lett. 202 (1995) 133–136. J. Soutschek, G.K.H. Zupanc, Apoptosis in the cerebellum of adult teleost fish, Apteronotus leptorhynchus, Dev. Brain Res. 97 (1996) 279–286. J.A. Stevenson, M.G. Yoon, Mitosis of radial glial cells in the optic tectum of adult goldfish, J. Neurosci. 1 (1981) 862–875. J.A. Stevenson, M.G. Yoon, Morphology of radial glia, ependymal cells, and periventricular neurons in the optic tectum of goldfish (Carassius auratus), J. Comp. Neurol. 205 (1982) 128–138. G.F. Striedter, S. Beydler, Distribution of radial glia in the developing telencephalon of chicks, J. Comp. Neurol. 387 (1997) 399–420. T. Stroh, G.K.H. Zupanc, The postembryonic development of somatostatin immunoreactivity in the central posterior / prepacemaker nucleus of weakly electric fish, Apteronotus leptorhynchus: a double-labelling study, Dev. Brain Res. 93 (1996) 76–87. ´ ´ ´ A rapid replacement of vimentinA. Tuba, L. Kallai, M. Kalman, containing radial glia by glial fibrillary acidic protein-containing astrocytes in transplanted telencephalon, J. Neural Transplant. Plast. 6 (1997) 21–29. T. Voigt, Development of glial cells in the cerebral wall of ferrets: direct tracing of their transformation from radial glia into astrocytes, J. Comp. Neurol. 289 (1989) 74–88. S.G. Waxman, M.J. Anderson, Regeneration of central nervous system structures: Apteronotus spinal cord as a model system, in: T.H. Bullock, W. Heiligenberg (Eds.), Electroreception, Wiley, New York, 1986, pp. 183–208.

23

´ [54] C. Yanes, M. Monzon-Mayor, M.S. Ghandour, J. de Barry, G. Gombos, Radial glia and astrocytes in developing and adult telencephalon of the lizard Gallotia galloti as revealed by immunohistochemistry with anti-GFAP and anti vimentin antibodies, J. Comp. Neurol. 295 (1990) 559–568. [55] S. Yuasa, K. Kawamura, R. Kuwano, K. Ono, Neuron–glia interrelations during migration of Purkinje cells in the mouse embryonic cerebellum, Int. J. Dev. Neurosci. 14 (1996) 429–438. [56] A.J. Zamora, M. Mutin, Vimentin and glial fibrillary acidic protein filaments in radial glia of the adult urodele spinal cord, Neuroscience 27 (1988) 279–288. [57] Z. Zhang, C.J. Krebs, L. Guth, Experimental analysis of progressive necrosis after spinal cord trauma in the rat: etiological role of the inflammatory response, Exp. Neurol. 143 (1997) 141–152. [58] G.K.H. Zupanc, An in vitro technique for tracing neuronal connections in the teleost brain, Brain Res. Brain Res. Protoc. 3 (1998) 37–51. [59] G.K.H. Zupanc, Neurogenesis, cell death and regeneration in the adult gymnotiform brain, J. Exp. Biol. 202 (1999) 1435–1446. [60] G.K.H. Zupanc, Up-regulation of somatostatin after lesions in the cerebellum of the teleost fish Apteronotus leptorhynchus, Neurosci. Lett. 268 (1999) 135–138. [61] G.K.H. Zupanc, I. Horschke, Proliferation zones in the brain of adult gymnotiform fish: a quantitative mapping study, J. Comp. Neurol. 353 (1995) 213–233. [62] G.K.H. Zupanc, I. Horschke, Salvage pathway of pyrimidine synthesis: divergence of substrate specificity in two related species of teleostean fish, Comp. Biochem. Physiol. B Biochem. Mol. Biol. 114 (1996) 269–274. [63] G.K.H. Zupanc, I. Horschke, R. Ott, G.B. Rascher, Postembryonic development of the cerebellum in gymnotiform fish, J. Comp. Neurol. 370 (1996) 443–464. [64] G.K.H. Zupanc, K.S. Kompass, I. Horschke, R. Ott, H. Schwarz, Apoptosis after injuries in the cerebellum of adult teleost fish, Exp. Neurol. 152 (1998) 221–230. [65] G.K.H. Zupanc, R. Ott, Cell proliferation after lesions in the cerebellum of adult teleost fish: time course, origin, and type of new cells produced, Exp. Neurol. 160 (1999) 78–87. [66] G.K.H. Zupanc, M.M. Zupanc, Birth and migration of neurons in the central posterior / prepacemaker nucleus during adulthood in weakly electric knifefish (Eigenmannia sp.), Proc. Natl. Acad. Sci. USA 89 (1992) 9539–9543. [67] G.K.H. Zupanc, S.C. Clint, Radial glia-mediated up-regulation of somatostatin in the regenerating adult fish brain, Neurosci. Lett. (in press).