The Distribution and Cellular Localization of CRF-R1 in the Vermis of the Postnatal Mouse Cerebellum

Experimental Neurology 178, 175–185 (2002) doi:10.1006/exnr.2002.8052

The Distribution and Cellular Localization of CRF-R1 in the Vermis of the Postnatal Mouse Cerebellum James S. King and Georgia A. Bishop Department of Neuroscience, The Ohio State University, 333 West 10th Avenue, Columbus, Ohio 43210 Received July 7, 2002; accepted September 16, 2002

The distribution of corticotropin-releasing factor (CRF), the development of CRF-binding sites, and the age at which application of CRF elicits a physiological response have been described previously in the postnatal mouse cerebellum. The intent of the present study was to determine the cellular and subcellular distribution of the CRF type 1 receptor (CRF-R1) in the vermis of the postnatal mouse cerebellum and to correlate these data with those presented in previous studies. On P0, CRF-R1 is present in the apical processes of migrating Purkinje cells. Between P0 and P8, CRF-R1 immunostaining is confined to a supranuclear position in Purkinje cell bodies. Between P9 and P14, the receptor immunolabeling circumscribes Purkinje cell nuclei and extends into their primary dendrites. An adult-like distribution is achieved between P16 and P21. Between P0 and P14, the CRF-R1 antibody also labels processes of migrating GABAergic interneurons that are directed toward the pial surface. By P12, labeling begins to circumscribe the nucleus of GABAergic cells in the internal granule cell layer. Finally, astrocytic processes in the white matter, as well as radial glial processes, show focal labeling with the CRF-R1 antibody beginning at P3 and throughout postnatal development. A previous study demonstrated that CRF does not elicit a physiological response in Purkinje cells until P9. This observation, together with the data presented in this study, suggests that the binding of CRF to the type 1 receptor may be involved in regulating the development of cerebellar neurons and glia immediately after birth, before CRF assumes its function as a neuromodulator later in postnatal development and in the adult. ©

2002 Elsevier Science (USA)

Key Words: Purkinje cell; basket cell; astrocyte; Bergmann glia; cerebellar interneuron; neuromodulator; neurotrophic factor; climbing fiber; mossy fiber.

INTRODUCTION

Corticotropin-releasing factor (CRF) is a 41-aminoacid peptide that has been identified as a neuromodulator in adult cerebellar circuits (2, 10, 19). In addition

to its presence in cerebellar afferents of adult animals, recent studies have shown that CRF is evident in punctate profiles in the embryonic cerebellum (4). Between birth and postnatal day (P) 10, the peptide distributes throughout all histological layers and lobules of the cerebellum (23). Further, a previous study in the mouse demonstrated that CRF-binding sites are present at birth (15), although they do not have an adult-like distribution (13); the identity of the cells expressing these binding sites has yet to be determined. Although CRF and its binding sites are present at birth, the peptide does not begin to alter the firing rate of developing Purkinje cells until P9 (3). Further, the response has immature physiological characteristics until P12–P16 when more adult-like responses are recorded (3). It is possible that CRF may have targets other than Purkinje cells at early postnatal ages. One experimental paradigm to confirm this would be to determine which cerebellar neurons express CRF receptors during development since, to elicit a response, CRF must first bind to either the CRF type 1 receptor (CRF-R1), for which it has a high affinity, or the CRF type 2 receptor (CRF-R2), for which it has a lower affinity (1, 6, 7). Based on the lack of a physiological effect on Purkinje cells before P9, it is possible that these receptors are not present on the cellular membranes of these neurons prior to this time or, if present, they have a conformation that does not allow them to bind to their ligand. Finally, although the second messenger pathway used by CRF, adenylate cyclase, is present during early postnatal life (12), its activation, induced by binding of the ligand to its receptor, may subserve a different function during early development compared to a role in regulating ion channels in adult circuits (10). Alternatively, the receptor may be coupled to a different effector protein that triggers a different cell signaling pathway. The purpose of the present study was to identify the targets of CRF-labeled terminals from birth to P14 by determining the cellular and subcellular distribution of the CRF type 1 receptor in the postnatal cerebellum.

175 ©

0014-4886/02 $35.00 2002 Elsevier Science (USA) All rights reserved.

176

KING AND BISHOP

Several studies have shown that the CRF-1 receptor is essential for mediating the biological effects of this peptide (1, 7) in the central nervous system. In the adult cerebellum, an antibody to CRF-R1 labels the somata and primary dendrites of nearly all Purkinje cells, whereas CRF-R2 is localized primarily on the initial segments of restricted populations of Purkinje cells and parallel fibers (5). In the present study, we used immunohistochemical techniques to determine the distribution of the CRF-1 receptor from P0 to P14. In addition to using the antibody to the receptor, cell-specific markers were applied to identify the cerebellar cells that express the receptor and to localize the receptor to specific compartments of the labeled cells (i.e., soma, dendrite) at different stages of development. The data indicate that CRF-R1 has unique cellular localizations that vary with the developmental age of different types of cerebellar neurons, as well as astrocytes. MATERIALS AND METHODS

Animals C57Bl/6J mice at P0, P3, P5, P8, P10, P12, and P14 were anesthetized with Avertin (2.5%, 0.2 ml/10 g, given intraperitoneally) and perfused through the heart with saline followed by 4% paraformaldehyde. The brains were removed and placed in the fixative for 6 – 8 h at 4°C. They were then transferred to Sorenson’s phosphate buffer that contained 20% sucrose. They were stored at 4°C for 24 –30 h or until they sank. Cellular Identification Immunohistochemical processing. The brains were cut either at 60 ␮m in the sagittal plane on a freezing microtome and processed as “free-floating” sections or at 30 ␮m on a cryostat (Zeiss), collected, and processed on chrome alum-subbed slides. CRF-R1 was immunolabeled with an antibody obtained from Santa Cruz Biochemicals (Santa Cruz, CA) that was generated in a goat against CRF-R1 at a concentration of 5 ␮g/␮l in phosphate-buffered saline (PBS) plus 0.3% Triton X-100 (PBT). A double or triple labeling paradigm was used to identify cells that express the receptor protein. In this paradigm, cell-specific antibodies, generated in species other than goat, were combined with the receptor antibody in a single solution and applied to the same sections. An antibody generated: (i) in a mouse against calbindin (Sigma, St. Louis, MO), used at concentration of 1:500, was used to identify Purkinje cells; (ii) in a mouse against glial fibrillary acidic protein (GFAP; Sigma), diluted 1:500, was used to label glial cells; (iii) in a rabbit against ␥-aminobutyric acid (GABA;

Sigma), diluted 1:500, was used to identify basket, stellate, and Golgi cells; and (iv) in a rabbit against PAX-2 (Zymed, San Francisco, CA), diluted 1:500, was used to identify migrating and immature GABAergic interneurons (16). Sections were incubated in a “cocktail” containing a primary antibody for the CRF-1 receptor and a primary antibody for one or two of the cell-specific markers; the tissue was incubated for 30 – 48 h at 4°C with constant agitation. Sections were then rinsed in PBS and placed in a solution that contained a mixture of two or three different secondary antibodies. CY-3 (goat IgG; red fluorescence; Jackson ImmunoResearch Laboratories, West Grove, PA) was used to label the receptor sites and Cy-2 (mouse or rabbit IgG; green fluorescence; Jackson ImmunoResearch Laboratories), Alexafluor 488 (mouse or rabbit IgG; green fluorescence; Molecular Probes, Eugene, OR), or AMCA (mouse or rabbit IgG; blue fluorescence; Jackson ImmunoResearch Laboratories, West Grove, PA) to label specific cerebellar cells. Different combinations of primary antibodies were used to determine the precise cellular localization of the CRF receptor. All of the experiments were carried out in accordance with procedures defined in the National Institutes of Health Guide for the Care and Use of Laboratory Animals. A protocol for these studies ensuring humane use of animals has been reviewed and approved by an Institutional Laboratory Animal Care and Use Committee. Data Analysis In the double and triple label studies, sections, primarily from the vermis, were examined with a Zeiss fluorescent microscope. An excitation filter of 575 nm was used to visualize Cy-3 labeled structures; 490 nm for CY-2 or Alexafluor 488; and 345 nm for AMCA. An overlay of the same images, exposed with two different filters, was computer generated to verify colocalization. Specific fields were selected and either a single focal plane or multiple, consecutive focal planes, spaced at 1-␮m intervals (z stacks), were collected with appropriate filters using a black and white digital camera and Axiovision software. Sequential focal planes of selected microscopic fields were obtained using the z stack portion of the software. The z-stacked images of each fluorochrome were further processed using a deconvolution program in the Axiovision software. They were transferred to Photoshop 7, pseudocolorized, superimposed, and merged. In all images the receptor immunoreactivity was pseudocolorized red, and the cell-specific immunofluorescence was pseudocolorized green or blue. CRF Receptor Antibodies CRF-R1 is an affinity-purified polyclonal antibody raised against the peptide corresponding to amino acid

CRF-R1 IN THE POSTNATAL MOUSE

sequence 396 – 415 mapping at the carboxy terminus of the CRF-R1 precursor of human origin which is identical to the corresponding mouse and rat sequences. The region used to generate the antibody (siptsptrvsfhsikqstav) has 85% homology with the CRF type 2 receptor, containing three substitutions (siptsptrisfhsikqtaav). Thus, this particular antibody has the potential to react with both types of CRF receptors. However, data from a previous study in the adult cerebellum (5) suggest little if any recognition of the CRF-R2 site by the antibody used in the present study. Unpublished observations on the distribution of CRF-R2 in postnatal animals using an antibody specific for CRF-R2 confirm that it is present, but in a different cellular and subcellular location. RESULTS

Cellular Distribution of CRF-R1 Cell-specific antibodies were combined with the CRF-R1 antibody to determine the cellular distribution of CRF-R1 at different stages of postnatal development. An antibody to calbindin was used to selectively immunolabel Purkinje cells and antibodies to GABA or PAX-immunolabeled GABAergic interneurons such as basket, stellate, and Golgi cells, and an antibody to GFAP was used to immunostain astrocytes and Bergmann glia. The following observations were made from the vermis of the postnatal cerebellum.

177

arrows). By P5 (Figs. 1C and 1D), Purkinje cells have formed an irregular monolayer. Although some dendritic extensions still emerge from the lateral aspects of these neurons, the primary outgrowth of dendrites is now present at the apical pole of the neurons. CRF-R1immunoreactive profiles have an irregular configuration that remains confined to the apical pole of the soma (Figs. 1C and 1D, block arrows). In a few cells, small patches of CRF-R1 immunoreactivity are evident at the base of the emerging dendrites (not illustrated). At P8 (Fig. 1E), the Purkinje cell bodies have a distinct primary dendrite that emerges from the apex of the soma. This primary dendrite gives rise to multiple secondary branches. Concomitant with this change in morphology, there also is a shift in the distribution of the receptor. Labeled patches of CRF-R1 are present on the primary dendrites (Fig. 1E, arrows), as well as within the cytoplasm of the soma (Fig. 1E, block arrows); CRF-R1 immunolabeling was not detected in dendritic branches beyond the primary dendrite. A distribution comparable to that seen in the adult (5) is achieved by P12 (Fig. 1F). At this age, the adult somadendritic characteristics of Purkinje cells are apparent as the dendrites nearly reach the pial surface. The immunolabeled profiles also have changed their configuration. They now appear as elongated profiles that completely circumscribe the Purkinje cell nucleus (Fig. 1F, block arrows). In the dendrites, patches of CRF-R1 labeling extend along the primary dendrites (Fig. 1F, arrows); immunolabeling cannot be visualized in the secondary branches.

Purkinje Cells Based on calbindin staining, at birth (P0), Purkinje cells have nearly completed their migration from the ventricular zone to the cortical mantle. At this stage of development they form a multitiered layer that is five to seven cells deep (Fig. 1A). Cell bodies near the outermost edge of the Purkinje cell layer (i.e., closer to the pial surface) appear to have smooth, round somata and extend rudimentary proximal dendrites. In these Purkinje cells, CRF-R1 immunoreactivity appears as irregular clusters in the cytoplasm at the apical pole of the cell body. This is clearly seen at P3 (Fig. 1B, block arrows). Cells deeper in the layer at P0 (i.e., closer to the white matter core) have more elliptical cell bodies and an elongated apical dendrite that is oriented toward the pial surface. In these deeper Purkinje cells, the receptor is localized in the initial portion of this process (Fig. 1A, arrows). By P3 (Fig. 1B), Purkinje cells are in a multitiered layer that is two to three cells deep. The apical dendrite observed at P0 is no longer evident, and short, irregular dendritic processes emerge from all aspects of the cell body. As the single elongated process retracts, the CRF-R1 immunoreactivity is evident in the apical pole of most, if not all, Purkinje cell bodies (Fig. 1B, block

GABAergic Interneurons GABA and PAX-2 (16) are markers that can be used to label immature basket, stellate, and Golgi cells. Antibodies to GABA label these neurons as it is the primary transmitter used by these inhibitory neurons in the adult. PAX-2, which is a paired box transcription factor expressed in several regions of the developing mammalian central nervous system, has been shown to label GABAergic interneurons (16) prior to the time they express their amino acid neurotransmitter, thus defining migrating inhibitory interneurons that are in a more immature state. It should be noted that whereas GABA labels the cytoplasm of interneurons, PAX-2 labeling is confined to the nucleus. At birth, CRF-R1 immunoreactivity is present in PAX-2- (Fig. 2A, arrows) and GABA- (Fig. 2B, block arrows) immunopositive neurons as they migrate through the white matter and into the IGL. The CRF-R1 immunolabeling is most often confined to the apical pole of the cell, which is oriented toward the pial surface or is present in a confined area, immediately adjacent to the nucleus. Between P3 and P5, a similar pattern of distribution is seen. In most GABAergic neurons CRF-R1 immuno-

FIG. 1. Distribution of CRF-R1 in calbindin-immunolabeled Purkinje cells. Tissue sections were simultaneously incubated with antibodies to calbindin (pseudocolorized green), which selectively labels Purkinje cells in the cerebellum and to CRF-R1 (pseudocolorized red). In A–F, each image is from a single focal plane of a “z-stack.” The molecular layer (ML) is labeled for orientation. At P0 (A), Purkinje cells form a multitiered layer. Neurons located deeper in the multitiered Purkinje cell layer contain CRF-R1 immunoreactivity (arrows) in processes that extend from the apical pole of the cell bodies and are directed toward the pial surface. At P3 (B), Purkinje cells are still in a multitiered layer that is two to three cells deep. At this age, CRF-R1 immunolabeling (block arrows) is confined to the apical cytoplasm of immature Purkinje cell bodies. By P5 (C, D), Purkinje cells are in an irregular monolayer and immature dendrites are beginning extend further into the ML from the apical pole of the cell bodies. D, higher magnification of Purkinje cells located within the box in C. CRF-R1 immunoreactivity (block arrow) remains confined to the apical pole of the Purkinje cells (D), at times extending into the base of the immature dendrite. By P8 (E), CRF-R1 immunoreactivity is present within the primary dendrites (arrows), as well as within the somata of Purkinje cells (block arrows). At this age, the reaction product is located not only at the apex of the cells, but it often assumes a circumnuclear position (E, block arrows). By P12 (F), the reaction product has an adult-like distribution as it completely surrounds the nucleus (block arrows) and extends into the primary dendrites (arrows). Bars in A, B, and E, 40 ␮m; bars in C, D, 20 ␮m; bar in F, 50 ␮m.

179

CRF-R1 IN THE POSTNATAL MOUSE

reactivity is present in a dense cluster that is confined to a circumscribed area of the cell body, immediately adjacent to the nucleus (Figs. 2C–2H, open block arrows). If a dendritic process is evident in GABA-immunolabeled neurons, it typically is oriented toward the pial surface (Figs. 2C–2G). The CRF-R1 receptor immunolabeling tends to localize in the cytoplasm between the nucleus and the base of dendritic processes (Figs. 2D, 2E, 2G, and 2H, open block arrows). This pattern is most apparent as the cells migrate from the presumptive white matter into and through the IGL. The cell bodies of some GABAergic interneurons within or close to the Purkinje cell layer are spherical in shape; CRF-R1 immunolabeling was not detected in these neurons (Figs. 2C and 2D, curved arrows). By P8 (Figs. 2I and 2J), the CRF-R1 immunolabeling remains at the apical pole of GABAergic interneurons in the IGL. In some GABAergic neurons, presumably Golgi cells, the immunolabeling becomes more dispersed throughout the cytoplasm of the soma and initial portion of the dendrites (Fig. 2J, block arrows). These cells retain their CRF-R1 immunoreactivity throughout all stages of development, as well as in the adult. In contrast, as observed at P5, immunolabeling cannot be detected in neurons destined to become basket or stellate cells as they begin to pass through the Purkinje cell layer to enter the molecular layer. By P14, an adult pattern of distribution is observed in GABAergic neurons in the IGL (Fig. 2K, open block arrows) in which the receptor immunostaining typically circumscribes the nucleus. Astrocytes An antibody to GFAP was used to label astrocytes. GFAP is first detected in the mouse cerebellum at P3, at which time it is present in cells located in the white matter immediately anterior and superior to the cerebellar nuclei (Figs. 3A and 3B, arrowheads). By P5, astrocytic processes that are colabeled with the CRF-R1 antibody are present throughout the white matter of the cerebellum, including the white matter core that extends into individual folia (Fig. 3C, arrowheads). Triple labeling of sections illustrates the differential distribution of CRF-R1 in GABAergic interneurons (Fig. 3C, open block arrow) and astrocytes (Fig. 3C, arrowheads). In addition, the clusters of labeling, characteristic of CRF-R1 immunolabeling in Purkinje cells, is also evident (Fig. 3C, PK). Glial processes in the white matter and the somata of astrocytes also express CRF-R1 immunostaining in their cytoplasm (Fig. 3D, arrowheads). By P8, CRF-R1 is observed in the somata of Bergmann glia (Figs. 3E and 3F, open arrowheads), as well as in the radial processes emerging from the Bergmann glial cell body and extending into the molecular layer (Figs. 3E–3G, arrowheads). At P14, astrocytes in the IGL and white matter continue

to show CRF-R1 labeling in their cell bodies and processes (Figs. 3H and 3I). DISCUSSION

During postnatal development of the cerebellum, the laminar distribution of the CRF type 1 receptor coincides with the laminar distribution of CRF (23) and CRF-binding sites (15), both of which are present at birth. At birth and through P10, CRF immunolabeling is present in puncta that are present in the immature Purkinje cell layer, IGL, and lower molecular layer; they are most prominent in the Purkinje cell layer. After P10, the CRF immunolabeling is present in morphologically identified climbing and mossy fiber terminals that have a restricted distribution in specific lobules of the cerebellum. Data from the present study indicate that the CRF-R1 subtype is present at birth in Purkinje cells, as well as in populations of GABAergic interneurons (basket cells, stellate cells, and/or Golgi cells). In addition, astrocytes including those migrating through the white matter and those located within the IGL, as well as Bergmann glia, also express CRF-R1 during postnatal development of the cerebellum. These results, taken together with those derived from two previous studies that described the distribution of the peptide CRF (23) and the presence of binding sites for CRF (15) in the early postnatal cerebellum, suggest a functional role for CRF during early stages of postnatal development. However, its function is likely not that of a neuromodulator, as described for CRF in the adult (2, 10), because CRF does not alter Purkinje cell activity until P9 (3). These data support our hypothesis that CRF has a different role during early stages of cerebellar maturation and that this effect may be temporally correlated with the developmental state of different populations of cerebellar neurons and astrocytes. CRF-R1 in Developing Purkinje Cells The vast majority of Purkinje cells in the mouse undergo their final mitotic division by embryonic day (E) 12 (28, 29). Between E12 and birth, the majority of Purkinje cells migrate to the cortex where they organize into a layer approximately five to seven cells in depth. Between P3 and P5, immature Purkinje cells begin to form the monolayer characteristic of the adult cerebellum. At birth, Purkinje cells located deep in the layer (i.e., closer to the white matter) have an elongated process extending from the apical pole of their somata, whereas those located more superficially, that is closer to the pial surface, no longer display this process. CRF-R1 was localized at the base of and within the elongated process of the deeper cells and was localized at the apical pole of the somata of the more superficially located neurons. The receptor remains in this supranuclear position until the Purkinje

180

KING AND BISHOP

CRF-R1 IN THE POSTNATAL MOUSE

cell elaborates the dendritic tree characteristic of the adult, between P3 and P8. At this time, the receptor shifts from an exclusively supranuclear position to a perinuclear distribution. In addition, CRF-R1 also appears along the primary shaft of the Purkinje cell dendrite. This shift in cellular distribution may be correlated with changes in the distribution of climbing fiber synapses from the apical process at birth, to the somata, and finally to the primary dendrites during Purkinje cell development (8, 17, 21). Relationship of CRF-R1 to Climbing Fiber Development Several studies have confirmed the presence of CRF in climbing fibers arising from the inferior olive (9, 13, 18, 22, 24). Olivocerebellar terminals have a ubiquitous distribution throughout the cerebellum and are present in the Purkinje cell layer prior to birth (8, 17). Immature climbing fibers go through several stages of synaptic development. The earliest stage, defined as the “creeper” stage (9), is characterized by transient contacts between climbing fibers and the apical process of Purkinje cells (8, 21). In the present study, the distribution of CRF-R1 at the apical pole and in the elongated processes of immature Purkinje cells coincides with the dendritic location of these transient climbing fiber synapses. As the apical process regresses and the transient synapses are lost by P5 in the rat (8), the CRF type 1 receptors are evident in a supranuclear position in the cytoplasm. Based on their distribution in the somatic cytoplasm, it is possible that the receptors are restricted to the Golgi apparatus and represent new receptor protein production that is yet to be inserted into the plasma membrane or the redistribution of existing receptors. This correlation is based on electron micrographs of Purkinje cells in the mouse at P7 and P14 (14) that demonstrate that the Golgi complex is confined to an area of the cytoplasm immediately above the nucleus during early stages of development, and then this cellular organelle assumes a perinuclear localization as the cell matures.

181

Between P3 and P8, Purkinje cells begin to undergo dendrogenesis and numerous perisomatic processes begin to emerge from the somata. During this stage of development, climbing fibers are in a stage of development defined as the pericellular nest phase (25), in which climbing fiber synaptic boutons are concentrated around the cell body (8, 21). At this time, CRF-R1 is confined to the apical cytoplasm of Purkinje cells. At the apical pole of the cell, a clearly defined primary dendrite begins to emerge and extend into the overlying molecular layer. A few focal patches of the CRF-R1 are evident at the base of the emerging primary dendrite by P5. As Purkinje cells begin to form the dendritic tree characteristic of the adult, more mature synaptic connections are made between developing climbing fibers and Purkinje cells (8). At this time, climbing fibers synapses are lost from the somata as they establish mature synaptic junctions on the dendrites (21). At this same time, CRF-R1 immunoreactivity begins to expand from a supranuclear to a perinuclear position and to the primary dendrites which is the pattern described in the adult (5). Mature synaptic contacts between climbing fibers and Purkinje cell dendrites also begin to form (8, 21). Clearly defined foci of CRF-R1 immunoreactivity are evident along the shaft of the primary dendrite, coincident with this translocation of climbing fiber synapses from the soma to the dendrite. Based on a previous study (23), P10 is the earliest age at which distinct CRF-immunolabeled climbing fibers could be detected. Prior to this age, CRF is present in a diffuse punctate plexus that has no specific phenotypic characteristics. However, this does not preclude that CRF is present in immature climbing fibers. Based on the distribution of the CRF-immunolabeled puncta within the Purkinje cell and molecular layers between birth and P10, it is likely that some if not all of the CRF immunoreactivity present during early postnatal cerebellar development is in climbing fibers. Recent observations indicate that CRF has no electrophysiological effect on Purkinje cells during the first

FIG. 2. Distribution of CRF-R1 immunoreactivity in GABAergic interneurons. Tissue sections were simultaneously incubated with antibodies to GABA or PAX-2 (pseudocoloized green) and CRF-R1 (pseudocolorized red). In A–K, each image is from a single focal plane of a “Z stack.” Immunolabeling at P0 is shown in A and B. In A, the PAX-2 antibody labels the nuclei of migrating interneurons. In B, an antibody to GABA immunolabels interneurons (open block arrows). At P0, CRF-R1 immunoreactivity is present in PAX-2- (A, arrows) and GABA- (B, block arrows) immunopositive neurons as they migrate through the white matter and into the IGL. At P3 (C–E), CRF-R1 immunolabeling is present in many (open block arrows) but not all (curved arrows) migrating GABAergic interneurons The GABAergic interneurons enclosed by the rectangles in C (labeled D and E), immediately subjacent to the Purkinje cell layer (PK), are shown at higher magnification in D and E. CRF-R1 immunoreactivity is confined to a focal area of the cell body and/or in dendrites oriented perpendicular to the Purkinje cell layer (C–E, block arrows). As some neurons migrate through the Purkinje cell layer, they become more spherical in shape (C, D, curved arrows); CRF-R1 immunolabeling is not evident in this population of interneurons. CRF-R1 immunolabeling is extensive in GABAergic interneurons at P5 (F–H). The areas enclosed by the rectangles labeled G and H in F are shown at higher magnifications in G and H. At P5 (F–H), CRF-R1 immunoreactivity in GABAergic interneurons is found in the cytoplasm and the primary dendrites (F–H, block arrows). At P8, both PAX-2 (I, arrows) and GABA (J, open block arrows) label interneurons in the IGL. CRF-R1 immunoreactivity is present in the soma and initial portions of the emerging dendrites (J, block arrows). By P14 (K), CRF-R1 immunoreactivity is more extensive in the cytoplasm of GABAergic interneurons located within the IGL (open block arrows). Bars in A–D, F, and I–K, 20 ␮m; bars in E, G, and H, 10 ␮m.

182

KING AND BISHOP

CRF-R1 IN THE POSTNATAL MOUSE

postnatal week of cerebellar development (3). Combined with the present data, these observations suggest that the adult function of CRF, which is to alter ion channels leading to increased excitability of Purkinje cells (10), may not be present until the mature synaptic relationship between climbing fibers and Purkinje cell dendrites is established. The observations of the present study are not unique for the CRF receptor. The receptor for calcitonin generelated peptide (CGRP), a peptide transiently expressed by climbing fibers, appears to follow a similar pattern of development (20). The CGRP receptor has a cytoplasmic location in Purkinje cells at P6 that gradually shifts to a membrane associated position in older animals. However, the timing of the receptor expression and the relocation is later than that described in the present study for CRF-R1. Whereas CRF-R1 is present in the Purkinje cell at birth, the CGRP receptor is not present in the cytoplasm of this neuron until P6. The CGRP receptor does not appear to be membrane associated until the second week of life, whereas CRF-R1 assumes an adult-like distribution between P8 and P12 (5). The ontogeny of the CGRP receptor suggests that it may not mediate a physiological effect on Purkinje cells until P15. In contrast, CRF was shown to exert an effect on Purkinje cells as a neuromodulator by P9, although even at this age the response was immature compared to that seen in the adult (3). CRF-R1 in Developing Interneurons Unlike Purkinje cells, which are born during embryonic stages of cerebellar development, GABAergic interneurons such as basket cells, stellate cells, and Golgi cells are born later, around E15 (16, 30). Thus, they undergo the vast majority of their migration and differentiation postnatally. Data in the present study demonstrate that both GABAergic and PAX 2-positive neurons express CRF-R1 as they migrate through the white matter and the IGL. As observed in developing Purkinje cells, the receptor is confined to a specific area of the somata, most often toward the apical pole from which a process emerged. Two populations of interneurons are identified, those that remain within the IGL, Golgi neurons, and those that continue into the molec-

183

ular layer, basket and stellate cells. The receptor is present in Golgi cells throughout postnatal development and in adulthood (5), suggesting that it may have a developmental and/or physiological effect on this type of interneuron, which is yet to be defined. In contrast, interneurons destined for the molecular layer coursed through the Purkinje cell layer between P3 and P12. As they traverse the Purkinje cell layer, the level of the receptor appears to decrease and once the cells are within the molecular layer proper, the receptor is no longer detected. These observations suggest that CRF could play a developmental role during the migration of cerebellar interneurons to their appropriate lamina. The present data further suggest that as basket and stellate cells reach their final laminar destination, the receptor is downregulated and is no longer detected in tissue sections processed for immunohistochemistry. CRF-R1 in Developing Astrocytes CRF-R1 is detected in astrocytes, identified with an antibody to GFAP, in the white matter core of the cerebellum beginning at P3. Several studies (11, 27) have shown that CRF has multiple effects on astrocytes in tissue cultures. First, CRF was shown to stimulate calcium influx into astrocytes via a cAMP-independent mechanism (27). This in turn could activate many different glial functions including release and/or uptake of neurotransmitters, as well as the release of neurotrophic factors, both of which act on neurons (26). In another study (11), CRF was shown to induce proliferation of cerebellar astrocytes in dispersed cell cultures. Based on the extensive distribution of CRF during early stages of development, the ligand may exert this proliferative effect on glial cells in vivo as they migrate through the white matter. In another study (26), it was found that one of the functions of peptide receptors on glial cells is to regulate synthesis and release of a variety of neuropeptides and growth factor peptides which can then act on other neurons or glia. The fact that glial cells and their processes intermingle with migrating and maturing GABAergic interneurons suggests that binding of CRF to CRF-R1 receptors on glial cells could subserve a similar function in the cerebellum.

FIG. 3. Distribution of CRF-R1 immunoreactivity in astrocytes. Tissue sections were simultaneously incubated with antibodies to GFAP (pseudocolorized green or blue) and CRF-R1 (pseudocolorized red). In A–I, each image is from a single focal plane of a “Z stack.” At P3, CRF-R1 immunoreactivity is present within the cell bodies of astrocytes (A, B, arrowheads) located in the white matter (WM) adjacent to the cerebellar nuclei (A, B, CN). C, an image of a section from a P5 mouse labeled for GFAP (blue), GABA (green), and CRF-R1 (red). CRF-R1 immunolabeling is present within glial processes in the white matter (arrowheads), and in GABAergic cells, located in the IGL (open block arrow). In addition, the clusters of CRF-R1 immunoreactivity, characteristic of labeling in Purkinje cells, are evident (PK). In addition to astrocytic processes. CRF-R1 immunoreactivity is also present in the cell bodies and processes of astrocytes at P5 (D, arrowheads). At P8 (E–G), CRF-R1 labels Bergmann glial cell bodies located within the Purkinje cell layer (E, F, open arrowhead), as well as their processes (E–G, arrowheads). The processes enclosed by the rectangles in E labeled F and G are shown at a different at higher magnifications in F and G; both images were taken from different focal planes in the same Z stack. At P14, CRF-R1 immunolabeling continues to be present in the cell bodies of astrocytes located within the white matter (rectangular box in H, I). The cells enclosed by the rectangle in H, labeled I, are shown at higher magnification in I. Bars in A–C, F, G, and I, 10 ␮m; and bars in D, E, and H, 20 ␮m.

184

KING AND BISHOP

Morara and colleagues (20) also demonstrated CGRP receptors in cerebellar glial cells although the timing of their appearance was different from that shown for CRF receptors. The data from the present study, as well as those derived from an analysis of the development of CGRP receptors, suggest that neuropeptides may exert a neurotrophic effect on cerebellar neurons indirectly through activation of signaling pathways involving glial cells. Future studies using cell cultures will allow us to better define what effect CRF may have on developing neurons. Once defined, these can be incorporated into theories of cerebellar development in general and the role of CRF in particular.

ACKNOWLEDGMENTS This study was supported by NINDS Grant NS08798 to J.S.K. We are grateful for the editorial comments provided by Dr. Paul Madtes, Jr. We also appreciate the excellent technical assistance of Barbara Diener-Phelan.

REFERENCES 1.

2.

CONCLUSION 3.

Data from the present and previous studies indicate that between birth and P10 major morphological, biochemical, and physiological changes occur in the cerebellum with respect to CRF and its receptor. As described in the present study, CRF receptor location changes concomitant with the extension, retraction, and remodeling of the dendritic tree of developing Purkinje cells. At the same time, the distribution of CRF changes from a diffuse pattern that cannot be localized to any particular afferent system to one that is clearly located within the axons and synaptic terminals of immature climbing and mossy fibers (23). Likewise, at birth, CRF-binding sites are present and by P10 they are uniformly distributed throughout the cerebellum (15). Functionally, application of CRF does not elicit a physiological response from Purkinje cells until P9 (3), when the CRF-R1 is present in the primary dendrites of Purkinje cells as shown in the present results. CRF-R1 also is expressed in GABAergic interneurons within the white matter and internal granule cell layer as early as P0. However, CRF-R1 expression is not evident in those neurons destined to become basket and stellate cells. This apparent reduction of receptor expression occurs as the cell bodies change from an elongated to a spherical shape and course from the internal granule cell layer to the molecular layer. The timing of these changes in the distribution of CRF and its type 1 receptor with respect to Purkinje, Golgi, and stellate/basket cells provides support for the proposal that CRF has distinct roles for different neuronal populations during postnatal cerebellar development. Prior to P10, it is suggested that binding of CRF to its receptor may primarily be involved in regulating developmental events related to Purkinje cell migration, dendrogenesis, and/or synaptogenesis. Once major developmental events have been completed around P10, it appears that CRF begins to function as a neuromodulator of Purkinje cell activity, as in mature cerebellar circuits.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

Behan, D. P., D. E. Grigoriadis, T. W. Lovenberg, D. T. Chalmers, S. Heinrichs, C. W. Liaw, and E. B. De Souza. 1996. Neurobiology of corticotropin releasing factor (CRF) receptors and CRF-binding protein: Implications for the treatment of CNS disorders. Mol. Psychiatry 1: 265–277. Bishop, G. A. 1990. Neuromodulatory effects of corticotropin releasing factor on cerebellar Purkinje cells: An in vivo study in the cat. Neuroscience 39: 251–257. Bishop, G. A. 2002. Development of a corticotropin releasing factor-mediated effect on the firing rate of Purkinje cells in the postnatal mouse cerebellum. Exp. Neurol 178, 165–174. Bishop, G. A., and J. S. King. 1999. Corticotropin releasing factor in the embryonic mouse cerebellum. Exp. Neurol. 160: 489 – 499. Bishop, G. A., C. M. Seelandt, and J. S. King. 2000. Cellular localization of corticotropin releasing factor receptors in the adult mouse cerebellum. Neuroscience 101: 1083–1092. Chalmers, D. T., T. W. Lovenberg, and E. B. De Souza. 1995. Localization of novel corticotropin-releasing factor receptor (CRF2) mRNA expression to specific subcortical nuclei in rat brain: Comparison with CRF1 receptor mRNA expression. J. Neurosci. 15: 6340 – 6350. Chalmers, D. T., T. W. Lovenberg, D. E. Grigoriadis, D. P. Behan, and E. B. DeSouza. 1996. Corticotropin-releasing factor receptors: From molecular biology to drug design. TIPS 17: 166 –172. Chedotal, A., and C. Sotelo. 1993. The ‘creeper stage’ in cerebellar climbing fiber synaptogenesis precedes the ‘pericellular nest’—Ultrastructural evidence with parvalbumin immunocytochemistry. Dev. Brain Res. 76: 207–220. Cummings, S., B. Sharp, and R. Elde. 1988. Corticotropinreleasing factor in cerebellar afferent systems: A combined immunohistochemistry and retrograde transport study. J. Neurosci. 8: 543–554. Fox, E. A., and D. L. Gruol. 1993. Corticotropin-releasing factor suppresses the after hyperpolarization in cerebellar Purkinje neurons. Neurosci. Lett. 149: 103–107. Ha, B. K., G. A. Bishop, J. S. King, and R. W. Burry. 2000. Corticotropin releasing factor induces proliferation of cerebellar astrocytes. J. Neurosci. Res. 62: 789 –798. Insel, T. R., G. Battaglia, D. W. Fairbanks, and E. B. De Souza. 1988. The ontogeny of brain receptors for corticotropin-releasing factor and the development of their functional association with adenylate cyclase. J. Neurosci. 8: 4151– 4158. King, J. S., Jr., P. C. Madtes, G. A. Bishop, and T. L. Overbeck. 1997. The distribution of corticotropin-releasing factor (CRF), CRF binding sites and CRF receptor mRNA in the mouse cerebellum. Prog. Brain Res. 114: 55– 66. Larramendi, L. M. H. 1969. Analysis of synaptogenesis in the cerebellum of the mouse. In Neurobiology of Cerebellar Evolution and Development (R. Linas, Ed.), pp. 803– 843. Am. Med. Assoc. Press, Chicago. Madtes, P. C., Jr., and J. S. King. 1999. The temporal and spatial development of CRF binding sites in the postnatal mouse cerebellum. Neurosci. Res. 34: 45–50.

CRF-R1 IN THE POSTNATAL MOUSE 16.

17.

18.

19.

20.

21.

22.

Maricich, S. M., and K. Herrup. 1999. Pax-2 expression defines a subset of GABAergic interneurons and their precursors in the developing murine cerebellum. J. Neurobiol. 41: 281–294. Mason, C. A., S. Christakos, and S. M. Catalano. 1990. Early climbing fiber interactions with Purkinje cells in the postnatal mouse cerebellum. J. Comp. Neurol. 297: 77–90. Merchenthaler, I. 1984. Corticotropin releasing factor (CRF)like immunoreactivity in the rat central nervous system: Extrahypothalamic distribution. Peptides 5: 27–32. Miyata, M., D. Okada, K. Hashimoto, M. Kano, and M. Ito. 1999. Corticotropin-releasing factor plays a permissive role in cerebellar long-term depression. Neuron 22: 763–775. Morara, S., A. Rosina, L. Provini, G. Forloni, A. Caretti, and S. J. Wimalawansa. 2000. Calcitonin gene-related peptide receptor expression in the neurons and glia of developing rat cerebellum: An autoradiographic and immunohistochemical analysis. Neuroscience 100: 381–391. Morara, S., J. J. van der Want, H. de Weerd, L. Provini, and A. Rosina. 2001. Ultrastructural analysis of climbing fiber-Purkinje cell synaptogenesis in the rat cerebellum. Neuroscience 108: 655– 671. Mugnaini, E., and B. J. Nelson. 1989. Corticotropin-releasing factor (CRF) in the olivo-cerebellar system and the feline olivary hypertrophy. In The Olivocerebellar System in Motor Control (P. Strata, Ed.), pp. 187–197. Springer-Verlag, Berlin/Heidelberg/New York.

23.

24.

25. 26.

27.

28.

29.

30.

185

Overbeck, T. L., and J. S. King. 1999. Developmental expression of corticotropin-releasing factor in the postnatal murine cerebellum. Brain Res. Dev. Brain Res. 115: 145–159. Powers, R. E., E. B. DeSouza, L. C. Walker, D. L. Price, W. W. Vale, and W. S. I. Young. 1987. Corticotropin-releasing factor as a transmitter in the human olivocerebellar pathway. Brain Res. 415: 347–352. Ramon y Cajal, S. 1911. Histologie du system nerveux de I’homme et des vertebres. Maloine, Paris. Schwartz, J. P., N. Nishiyama, D. Wilson, and T. Taniwaki. 1994. Receptor-mediated regulation of neuropeptide gene expression in astrocytes. Glia 11: 185–190. Takuma, K., T. Matsuda, T. Yoshikawa, J. Kitanaka, M. Gotoh, K. Hayata, and A. Baba. 1994. Corticotropin-releasing factor stimulates Ca 2⫹ influx in cultured rat astrocytes. Biochem. Biophys. Res. Commun. 199: 1103–1107. Uzman, L. L. 1960. The histogenesis of the mouse cerebellum as studied by its tritiated thymidine uptake. J. Comp. Neurol. 114: 137–159. Yuasa, S., K. Kawamura, K. Ono, T. Yamakuni, and Y. Takahashi. 1991. Development and migration of Purkinje cells in the mouse cerebellar primordium. Anat. Embryol. (Berlin) 184: 195–212. Zhang, L., and J. E. Goldman. 1996. Generation of cerebellar interneurons from dividing progenitors in white mater. Neuron 16: 47–54.