Expression of calretinin in diverse neuronal populations during development of rat hippocampus

Pergamon

PII:

Neuroscience Vol. 81, No. 4, pp. 1137–1154, 1997 Copyright ? 1997 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306–4522/97 $17.00+0.00 S0306-4522(97)00231-5

EXPRESSION OF CALRETININ IN DIVERSE NEURONAL POPULATIONS DURING DEVELOPMENT OF RAT HIPPOCAMPUS M. JIANG* and J. W. SWANN*†‡ *The Cain Foundation Laboratories, Department of Pediatrics, Baylor College of Medicine, Houston, Texas 77030, U.S.A. †Division of Neuroscience, Baylor College of Medicine, Houston, Texas 77030, U.S.A. Abstract––The prenatal and postnatal expression of calretinin was studied in hippocampus of the rat using immunohistochemical procedures. Calretinin was detected as early as embryonic day 15 in the primordial hippocampus where calretinin-containing neurons and fibres were localized to the primitive plexiform layer. Upon emergence of the hippocampal plate (the prospective stratum pyramidale), large numbers of immunopositive multipolar cells were observed in the marginal zone. Fewer cells with fusiform cell bodies were observed bordering the hippocampal plate and subplate. During the perinatal period (embryonic day 20 to postnatal day 0), large numbers of immunoreactive pyramidal-like neurons were observed at the margin of the hippocampal plate with the subplate. At this same time, many calretinin-containing neurons with irregularly shaped dendrites were observed in stratum radiatum. Soon after birth (postnatal day 3), the calretinin immunoreactivity of both these later cell types rapidly declined and a new population of calretinin-immunopositive cells emerged, the Cajal–Retzius cells of stratum lacunosum-moleculare and the dentate gyrus. The Cajal–Retzius cells rapidly matured but disappeared by the second postnatal week. During the second postnatal week, calretinin interneurons of the adult hippocampal formation began to appear. Their immunoreactivity increased by postnatal day 15, when the number of calretininimmunopositive interneurons in area CA1 and stratum radiatum of CA3 exceeded that of the adult. At this time, the soma and proximal dendrites of many calretinin interneurons were found to contact each other. The frequency of such cellular appositions decreased in adulthood. The results presented here show that calretinin immunohistochemistry can be very useful in recording the development of subpopulations of hippocampal neurons that are present during distinct embryonic and postnatal periods. Although some neuronal types may exist only briefly during hippocampal development, others appear to express calretinin transiently during restricted phases of neuronal differentiation. Surprisingly, this includes some hippocampal pyramidal cells. However, even as the adult pattern of immunostaining emerges in week 2, morphological refinement of interneurons continues to take place, which eventually leads to the population of calretinin-containing interneurons of the mature hippocampus. ? 1997 IBRO. Published by Elsevier Science Ltd. Key words: hippocampus, calcium binding protein, interneurons, calretinin, Cajal–Retzius cells.

Intracellular calcium is thought to play a key role in a wide variety of neurodevelopmental processes. For instance, recent studies have implicated changes in cytoplasmic calcium activity in various forms of neuronal differentiation, including neurite outgrowth,18 expression of neurotransmitter receptors,46 neuronal migration,27,28 growth cone motility,26,38 use-dependent developmental plasticity,10,41 and gene expression.46,47 Regulation of the concentration of calcium in cytoplasm is thought to be critical for proper development. Indeed, in some systems a dynamic range of optimal intracellular calcium activity exists. If levels drop below or rise above this range, dramatic changes can occur in neurodevelopmental processes.2,26 Numerous molecular systems are employed by developing neurons to generate intracellular calcium transients. Other processes regulate the levels of ‡To whom correspondence should be addressed. Abbreviations: E, embryonic day; P, postnatal day.

calcium achieved in the cytoplasm. Examples of the latter are thought to be the calcium-binding proteins, calbindin, parvalbumin, and calretinin.5 Although direct evidence demonstrating their ability to buffer intracellular calcium is not available, the notion that calcium-binding proteins play a role in buffering calcium persist. In part, this is because they are found in high levels in populations of central neurons that are resistant to calcium-mediated excitotoxic damage.20,39,42,49 Calcium-binding proteins have been reported to have distinct patterns of expression during central nervous system development. The ontogenesis of calbindin-D28k and parvalbumin expression has been the subject of numerous studies. Less is known about the development of calretinin containing neurons and this is the focus of the present investigation. Previously, immunohistochemical studies have shown that parvalbumin is expressed relatively late during central nervous system maturation.3,44

1137

1138

M. Jiang and J. W. Swann

For instance, in rat hippocampus, interneurons are immunopositive for parvalbumin starting around postnatal day (P)8 and increase in number and staining intensity during the next week of life.8,44 In contrast, calbindin-D28k has been shown to be expressed transiently during brain development in numerous neuronal systems and animal species.17,24,30 In mouse hippocampus, for example, faint calbindinimmunoreactive neurons have been reported as early as embryonic day (E)15–E17 in marginal zone and subplate neurons. These diminish in staining intensity at later prenatal ages. Calbindin-D28k has been reported to reappear on P2 in clusters of neurons in stratum radiatum and stratum oriens.45 The ontogeny of calretinin-immunoreactive cells has been recently reported in a few areas of brain.6,11,16,45,48,50 Most of these studies have been of neocortex and uniformly show that calretinin is expressed very early in development. For instance, upon formation of the cortical plate, neurons of the subplate and marginal zone form continuous bands of calretinin immunoreactivity in the mouse, rat, and monkey.11,16,50 The Cajal–Retzius cells of the marginal zone of neocortex are particularly well-visualized by antibodies developed against calretinin.29 In the rat, some pyramidal cells of neocortex also appear to express calretinin transiently during early postnatal life.16 Later in postnatal life, the adult pattern of immunoreactivity begins to emerge as interneurons of neocortex begin to express calretinin immunoreactivity. In hippocampus, one study of embryonic and early postnatal mouse reported intense calretininimmunoreactive neurons of stratum lacunosummoleculare that were reminiscent of neocortical Cajal–Retzius cells.45 In contrast, calretininimmunopositive neurons in the adult hippocampus have been described in detail. There are two major subclasses of calretinin-immunoreactive neurons: spiny cells of stratum lucidum and nonspiny interneurons of various morphologies scattered in all layers and subfields.22,35 The spiny cells form a distinct population of non-pyramidal cells that have complex dendritic arbors that are restricted to stratum lucidum and form a feltwork of processes in this laminae. The non-spiny calretinin-immunoreactive

neurons have smooth and often varicose dendrites that run radially, often ascending or descending through multiple layers of the hippocampus. In area CA1, these vertical dendrites have been reported to make frequent contact with other calretininimmunoreactive dendrites or cell bodies. These and other observations have lead Gulyas et al.21 to suggest that calretinin-immunoreactive neurons form networks of cells that are electronically coupled and control the activity of other subpopulations of GABA-containing inhibitory interneurons. In this study, the ontogeny of calretininimmunoreactive neurons was examined in the rat from midgestation to adulthood. EXPERIMENTAL PROCEDURES

Calretinin immunohistochemistry Timed pregnant and families of Wistar rats were obtained from Harlan Sprague–Dawley Corporation (Indianapolis, IN). The day of mating was considered E0 and the day of birth P0. Rat pups were studied on E15, E17, E20, P0, P1, P3, P5, P7, P10, and P15. Comparisons were made to adult rats examined on P50. All experiments were carried out in accordance with the National Institute of Health guidelines for the care and use of laboratory animals. Fetuses from 10 separate litters were collected by caesarean section after anaesthesia of the dam with an intraperitoneal injection of a ketamine/xylazine mixture (33 mg ketamine and 1.5 mg xylazine per kg). At postnatal ages, rat pups were obtained from 19 separate litters. At least six rats were used at each developmental age. After anaesthesia, all rats were perfused transcardially with 0.1 mol/l phosphate-buffered salines, followed by 4% paraformaldehyde in 0.1 mol/l phosphate buffer (pH 7.4). Rat brains were then removed and immersed overnight in the same fixative (4)C), which was replaced with chilled 10% and then 20% sucrose. Brains were coronally sectioned (50 µm thickness) on a freezing microtome (Zeiss, HM 400). Free-floating sections were first incubated is 3% hydrogen peroxide to inhibit endogenous peroxidases. They were blocked with 5% goat serum. Sections were subsequently incubated overnight in a cold room (4)C) with a rabbit polyclonal antibody against calretinin (1:15,000 dilution; Chemicon International, Temecula, CA). The tissue-bound primary antibodies were then detected by incubating for 1 h with biotinylated goat antirabbit antibodies (1:200 dilution, Jackson ImmunoResearch) and strepavidin-conjugated horseradish peroxidase (1:200 dilution, Jackson ImmunoResearch). All immunoreagents were diluted in 0.05 mol/l Tris-buffered saline containing 0.3% Triton X-100. Antibodies bound to antigens in the tissue were visualized with 0.05% 3,3*-diaminobenzidine

Abbreviations used in figures CP DG DMZ DP E GL H H HP SP IMZ IZ

cortical plate dentate gyrus dentate marginal zone dentate plate embryonic day granule cell body layer hilus of the dentate gyrus hippocampus hippocampal plate hippocampal subplate inner marginal zone intermediate zone

MZ ML OMZ P PPL SP SR SLM SR SO SP VZ

marginal zone molecular layer of the dentate gyrus outer marginal zone postnatal day primitive plexiform layer stratum pyramidale stratum radiatum stratum lacunosum-moleculare stratum radiatum stratum oriens subplate ventricular zone

Calretinin in developing hippocampus tetrahydrochloride and 0.015% hydrogen peroxide. Often slices were counterstained with Cresyl Violet in order to define the boundaries between hippocampal laminae more clearly. This was especially important when examining embryonic tissue. Usually such counterstained sections were chosen for preparation of photomicrographs. For immunocytochemical controls, sections from animals at both embryonic and postnatal ages were incubated in the absence of the primary antibody or by substitution of normal serum. Neuron reconstructions, camera lucida drawings, and cell counts Neuron reconstructions were performed using the Eutectic Neuron Reconstruction System (Sun Technologies, Durham, NC). Camera lucida drawings were done using a 100# oil immersion objective. The number of calretininimmunoreactive neurons in each of the hippocampal laminae (Fig. 9E) was counted in rats on P15 and at adulthood. Coronal sections through the middorsal hippocampus were selected 0–1 mm caudal to the first appearance of the dorsal lateral geniculate. Well-stained cell bodies with clearly demarked boundaries against the background were counted. Cell bodies found along the full extent of each laminae were counted in sections of 10 hippocampi from five rats on both P15 and adulthood. This procedure avoided cell dilution due to the mediolateral and dorsoventral growth of the hippocampus. Cell counts were then corrected according to the Abercrombie formula.1 Correction of dilutions due to hippocampal growth in the rostrocaudal axis were made by multiplying Abercrombie’s corrected values for adults sample by 1.13. This latter value was derived by estimating the length of the hippocampus in the rostrocaudal axis on P15 and in adulthood in five rats at each age. This was accomplished by dissecting the hippocampi free of adhering brain tissue following paraformaldehyde fixation and measuring the length of each hippocampi with a hand-held calibre. P15 and P50 hippocampi were found to be 8.9&0.14 mm and 10.1&0.27 mm in length, respectively. For statistical anaylsis of cell counts, population data were first tested for normal distribution. Thereafter, the mean number of immunoreactive cells in each laminae on P15 and P50 were compared using a Student’s t-test. Nomenclature of developing hippocampal laminae Layers in the developing hippocampus were defined as in previous studies (e.g., Soriano et al.45) At E15, only the ventricular zone and primitive plexiform layer were delimited. From E17 to E20, the following layers were recognized in the hippocampus: ventricular zone, intermediate zone, subplate (embryonic stratum oriens), hippocampal plate (embryonic pyramidal cell body layer), and the inner and outer marginal zones. The inner marginal zone will eventually become stratum radiatum, whereas the outer marginal zone corresponds to stratum lacunosum-moleculare. In embryos, the dentate gyrus was divided into the hilus, dentate plate (the eventual granule cell body layer), and the dentate marginal zone (the eventual molecular layer). After birth, the nomenclature employed for adult hippocampus was used. RESULTS

Cells of the primitive plexiform layer and the emerging subplate and marginal zone By E15 in the rat, antibodies against calretinin revealed a single layer of immunolabelled cells in the primordial hippocampus (Fig. 1A). At this time, the hippocampal plate (forerunner of the pyramidal cell

1139

body layer) had not yet formed. The layer of dense immunostaining was continuous with a band of cells seen in the outermost layer of the adjacent cortex (arrows, Fig. 1A). Thus, it was referred to as the ‘‘primitive plexiform layer’’ (Fig. 1B). In sections of dorsal hippocampus, this layer of the two hippocampi formed two bands that ran nearly parallel as they descended into the lateral ventricle (Fig. 1A, B). Both cells and processes were immunostained at this stage of development (Fig. 1C). The cells varied in their intensity of staining and were difficult to describe morphologically due to their high density. When discernible, their soma appeared to be of medium size (4–10 µm in diameter). Unusually large calretinin-positive processes were also present and restricted to the primitive plexiform layer (open arrows, Fig. 1C). By E17, the hippocampal plate was clearly visible (Fig. 2) and continuous with the cortical plate of adjacent cortex (Fig. 2A). In cortex, the emergence of the cortical plate has been shown to split the calretinin-positive cells of the primitive plexiform layer into two distinct populations: calretinincontaining neurons above and below the cortical plate in the marginal zone and subplate, respectively.11,16,50 In hippocampus, this also appeared to occur. However, immunolabelled cells in the hippocampal marginal zone outnumbered those of the subplate. The photomicrographs and neuronal reconstructions in Fig. 2 illustrate the location and morphology of these cells in E17 preparations. The cells were more loosely arranged than on E15, and the morphology of neurons was more clearly apparent. The soma of the cells in the marginal zone varied in size (6–10 µm in diameter). The cells’ perikarya were a variety of shapes. Most cells were multipolar and possessed one or more very fine processes that radiated parallel to the hippocampal plate (arrows, Fig. 2C and D inset 1). Other cells were located in the hippocampal plate itself and resembled immunopositive cells that bordered the hippocampal plate and subplate (open arrows Fig. 2C and D inset 2). The latter usually had fusiform-shaped soma and a single process that projected perpendicular to the hippocampal plate and often into the cell layer (open arrows, Fig. 2C and D inset 2). Neurons bordering the hippocampal plate and subplate and nonpyramidal cells of the marginal zone By E20, major changes had occurred in hippocampal morphology. At this age, the lamination of the hippocampus had taken on a form resembling that of the adult in that the hippocampal plate was shaped so that areas destined to become the CA1 and CA3 subfields could be recognized. The dentate plate (Fig. 3A) was also presents and cells formed a suprapyramidal blade of the prospective granule cell body layer. The infrapyramidal blade had yet to form. The most dramatic and intense immunostaining observed

Fig. 1. Calretinin immunoreactivity on embryonic day 15 (E15). (A) Low-magnification photomicrograph of a coronal section of the brain demonstrating bilateral calretinin immunoreactivity in the hippocampal anlage (arrowheads and H). The layer of immunoreactivity is continuous with one in neocortex (arrows). (B) At higher magnification, the primitive plexiform layer (PPL) of the hippocampus is immunopositive for calretinin. (C) Immunopositive cells and fibres (open arrows) of the primitive plexiform layer. Scale bars=500 µm in A, 100 µm in B, and 10 µm in C.

Calretinin in developing hippocampus

1141

Fig. 2. Calretinin immunoreactivity on embryonic day 17 (E17). (A) Low-magnification photomicrograph of hippocampus (H) and adjacent neocortex. Immunopositive cells are positioned on both sides of the hippocampal plate (HP) and cortical plate (CP). (B) In hippocampus, immunopositive cells are numerous in the marginal zone (MZ). (C) Higher magnification of immunoreactive neurons in the marginal zone (MZ), hippocampal plate (HP), and hippocampal subplate (SP). Arrows – multipolar neurons of marginal zone. Open arrows – neurons in hippocampal plate. (D) Neuron reconstruction of calretinin-containing neurons depicting their morphology and location. Lines denote boundary of hippocampal plate. Inset 1: Camera lucida drawings of two cells of the marginal zone. Inset 2: Camera lucida drawings of two neurons bordering the hippocampal plate and subplate. IZ, intermediate zone. VZ, ventricular zone. Scale bars=200 µm in A, 100 µm in B, 20 µm in C, 100 µm in D, and 65 µm in insets 1 and 2.

at this time was a nearly continuous band of neurons located at the border of the hippocampal plate and subplate (Fig. 3A). Immunopositive cells could also be seen within the hippocampal plate and appeared to be interspersed among non-immunoreactive cells. Based on their morphology, laminar location, and packing density, these cells were interpreted to be hippocampal pyramidal cells. In some ways, these cells of the hippocampal plate had the appearance of migrating neurons. Their perikarya could be pyramidal but usually fusiform in shape. They also had a single, rather thick process that extended into and on rarer occasions through the hippocampal plate to the border with the marginal zone. This process was reminiscent of the leading process described for migrating cortical neurons.27,37 A second class of calretinin-containing neurons populated the inner marginal zone. These cells were similar in size to the subplate cells observed at E20.

However, their perikarya were multipolar or quite irregularly shaped and had numerous short, thin, irregular-shaped processes that projected in a seemingly random fashion within this zone. As is apparent in the photomicrographs of Fig. 3, these cells were quite numerous and uniformly distributed in the inner marginal zone of prospective areas CA1 and CA3. Similar cells were apparent, but less numerous, in the intermediate zone and subplate. Calretinin immunoreactivity was less conspicuous in the outer marginal zone of area CA1 or the molecular zone of the dentate gyrus at E20. The pattern of immunostaining at P0 was similar to that at E20. However, cells located at the border of the hippocampal plate and subplate were usually more heavily stained than on E20, and a higher percentage of these cells appeared to be located in the pyramidal cell body layer (Fig. 4A, B, D) as opposed to the border between stratum pyramidale and

1142

M. Jiang and J. W. Swann

Fig. 3. Calretinin immunoreactivity on embryonic day 20 (E20). (A) Low-magnification photomicrograph of hippocampus, The suprapyramidal blade of the dentate plate (DP) is present. Large numbers of immunopositive cells can be seen in the hippocampal subplate (SP). A second population of cells is in the inner marginal zone (IMZ). (B) Higher magnification of cells of the subplate. These are monopolar with a process projecting towards or into the hippocampal plate (HP). (C) Immunoreactive cells of the inner marginal zone (IMZ). (D) Neuron reconstruction illustrating location and morphology of immunoreactive cell types. VZ, ventricular zone. IZ, intermediate zone. OMZ, outer marginal zone. DMZ, dentate marginal zone. H, dentate hilus. Scale bars=200 µm in A, 20 µm in B, 20 µm in C, 100 µm in D.

stratum oriens. As on E20, usually a single process emanated from these cells. Many neurons were immunoreactive in stratum radiatum on P0 (Fig. 4C, D). Like their counterparts of the inner marginal zone on E20, these cells were multipolar or highly irregular in shape. Their processes appeared to be longer and more complex than those observed on E20. On P0, calretininpositive cells were present in stratum lacunosummoleculare but were fewer in number than observed on P3. Cajal–Retzius cells of stratum lacunosum-moleculare and the molecular layer of the dentate gyrus Between P0 and P3, dramatic changes occurred again in the calretinin immunoreactivity of the developing hippocampus. This is clearly evident when sections of P3 rat hippocampus (Fig. 5A) are compared with those taken from P0 rats (Fig. 4A). Many of the neurons that were intensely immunoreactive

for calretinin on P0 displayed a decrease or a complete lack of immunoreactivity. These changes were most dramatic in the CA1 subfield. In some preparations, immunoreactive neurons present on P0 in area CA3 were still evident. However, in others (e.g., Fig. 5A), a dramatic decrease in the number of immunoreactive cells was seen. At this time, a new population of neurons was easily distinguished. These cells were at the boundary of stratum lacunosum-moleculare of area CA1 and the molecular layer of the dentate gyrus. Immunoreactive neurons were present in both of these laminae and thus bordered the hippocampal fissure (arrows, Fig. 5B). The soma of these cells were small (4–8 µm in diameter) and sometimes irregular in shape (Fig. 5C), with dendritic processes. The dendrites were most often short with few branches (reconstruction and camera lucida drawing, Fig. 5D) and ran within the laminae in which their soma were located. Occasionally, longer small calibre fibres were seen, which were presumed to be axons (arrow, Fig. 5C). These cells were considered to be hippocampal

Calretinin in developing hippocampus

1143

Fig. 4. Distribution of calretinin-containing cells from postnatal day 0 (P0). (A) Low-magnification photomicrograph show cells in proximal stratum oriens (SO) and in stratum radiatum (SR). (B) Higher magnification of cells in stratum oriens. These cells are likely the same population seen in the hippocampal subplate on E20, although they are more intensity immunoreactive and more often located in stratum pyramidale (SP). (C) Immunopositive cells in stratum radiatum (SR) have irregularly shaped dendrites. (D) Neuron reconstruction shows morphology and location of calretinin-immunopositive cells. SLM, stratum lacunosum-moleculare. ML, molecular layer of the dentate gyrus. GL, granule cell body layer. H, hilus of the dentate gyrus. Scale bars=200 µm in A, 20 µm in B, 20 µm in C, and 100 µm in D.

Cajal–Retzius cells that were described recently in the mouse.45 In addition, the hilus of the dentate gyrus was populated with immunoreactive neurons that were similar in appearance to those in the molecular layer (Fig. 5A, D). These cells were distributed randomly within the hilus and did not appear to share dendritic orientation like that described earlier. Some of these cells were also observed in the granule cell body layer. By P7, the Cajal–Retzius cells of stratum lacunosum-moleculare and the molecular layer of dentate gyrus were still evident (Fig. 6A, B). However, they were fewer in number, and their morphology had changed dramatically. The cell bodies were no longer irregular in shape. Instead, they were ovoid and often had a swollen appearance (arrows, Fig. 6C). Identifiable dendrites were absent, but fine, beaded processes (open arrow, Fig. 6C) arose from the soma. By P10, the Cajal–Retzius cells were no longer evident. On both P7 and P10, new populations of calretinin-immunoreactive neurons were observed

in the dentate gyrus and the CA1 and CA3 subfields (Fig. 6A). The perikarya of these cells varied in staining intensity. Often, they were lightly stained. When dendrites were observed, they were short. These neurons appeared to be forerunners of calretinin-containing interneurons that were intensely stained and more easily visualized on P15. Interneurons of the hippocampus and dentate gyrus By P15, layers of all regions of the hippocampus and dentate gyrus contained calretininimmunoreactive neurons. Most striking was the number of cell bodies in and near the pyramidal cell body layer and the hilus of the dentate gyrus. Indeed, the number of immunoreactive cells appeared to be much higher on P15 than in the adult hippocampus (compare Fig. 7A with Fig. 9A). To examine this possibility more closely, we undertook cell counts of immunopositive soma in each hippocampal laminae (see Fig. 9E for delineation of laminae) on P15 and at

1144

M. Jiang and J. W. Swann

Fig. 5. Immunopositive Cajal–Retzuis cells of stratum lacunosum-moleculare (SLM) and the dentate gyrus on postnatal day 3 (P3). (A) Low-magnification photomicrograph showing immunopositive cells that are particularly evident on either side of the hippocampal fissure. (B) Higher magnification photomicrograph showing the distribution of these cells. Arrows denote the approximate location of the hippocampal fissure. (C) Photomicrograph showing immunopositive cells and processes in stratum lacunosum-moleculare (SLM) and the molecular layer of the dentate gyrus (ML). Arrow denotes a long thin process, which is likely an axon arising from this Cajal–Retzius cell. (D) Upper portion shows camera lucida drawings of two Cajal–Retzius cells that were located in stratum lacunosum-moleculare (SLM). Lower portion of panel shows the distribution and morphology of cells. SO, stratum orien. SP, stratum pyramidale. SR, stratum radiatum. GL, granule cell body layer. H, hilus of the dentate gyrus. Scale bars=200 µm in A, 50 µm in B, 20 µm in C, and 100 µm in D (25 µm for upper camera lucida drawings).

adulthood. Results in Table 1 show that while cell counts were higher on P15 than in adulthood, these differences were confined to certain laminae. In area CA1, there was a 35–60% decrease in number of immunoreactive neurons in strata pyramidale, oriens and radiatum between P15 and adulthood. In area CA3, there was a 21% decrease in cell number in stratum pyramidale, which did not reach statistical signficance. On the other hand, there was a 43%

decrease in cell number in stratum radiatum of CA3. Other laminae were similar in cell number at the two ages. Immunoreactive cells on P15 had the appearance of hippocampal interneurons and comprised a diversity of interneuronal subtypes. The soma of these cells had a variety of shapes, including multipolar or irregularly shaped, bipolar, fusiform, and pyramidal. They varied in size from 10 to 18 µm in diameter.

Fig. 6. Distribution of calretinin-immunopositive cells on postnatal day 7 (P7). (A) At this time, forerunners of the interneurons of adulthood begin to express calretinin in areas CA1 and CA3 and in the dentate hilus (H). Cajal–Retzius cells of stratum lacunosum-moleculare (SLM) and the molecular layer of the dentate gyrus (ML) are still visible. (B) At higher power, the location and distribution of the Cajal–Retzius cells are shown. Arrows point to blood vessels that indicate the location of the hippocampal fissure. (C) A higher power photomicrograph shows the morphological alterations that have taken place since P3. Cells appear swollen (arrows) and bear very thin, beaded processes (open arrows). SO, stratum oriens, SP, stratum pyramidale, SR, stratum radiatum, GL, granule cell body layer. Scale bars=200 µm in A, 50 µm in B, 20 µm in C.

Calretinin in developing hippocampus

1145

1146

M. Jiang and J. W. Swann Table 1. Calretinin-immunoreactive neurons in hippocampal laminae on postnatal day 15 and in adulthood

Lamina

Cell number

Cell number

Area CA1 Pyramidale (SP) Oriens (SO) Radiatum (SR) Lacunosum-moleculare (SLM) Area CA3 Pyramidale (SP) Oriens (SO) Radiatum (SR) Lacunosum-moleculare (SLM) Lucidum (SL) Dentate gyrus Granule layer (GL) Hilus (H) Molecular layer (ML)

P15 70.8&5.5*** 78.0&2.4*** 61.4&5.6** 25.9&2.3

Adult 28.6&1.4 51.2&3.5 35.8&1.6 27.6&2.1

84.6&10.8 31.9&4.6 51.3&4.4** 8.27&2.0 31.8&4.2

66.4&9.8 29.2&4.0 28.8&3.5 4.65&0.3 33.5&3.5

4.3&0.6 58.2&7.4 14.6&1.3

3.6&0.6 53.9&3.0 14.1&2.2

Reported are the number of immunoreactive neurons in 50 µm-thick sections. Values are mean&S.E.M. and were adjusted for hippocampal growth. n=5. Abbreviations (in parentheses) for each laminae refer to areas demarcated in the drawing in Fig. 9E. P, Postnatal day. **P<0.01, ***P<0.001 when cell counts were compared to adult samples.

Calretinin-containing dendrites were clearly evident by this age, but their immunoreactivity appeared to be less than in the adult (compare Fig. 7B with Fig. 9B). Dendrites tended to run radially in the CA1 subfield. In contrast, the dendrites of cells in the dentate hilus arborized extensively just beneath the granule cell body layer (Fig. 7A, C; see also Fig. 10). Often, dendrites had a smooth or beaded appearance (open arrows, Fig. 7B). Another striking feature of the calretinincontaining interneurons seen on P15 was the numerous contacts cell bodies and dendrites made with each other in the hippocampal cell body layer. Often, cell bodies of immunoreactive neurons were found to be in direct apposition. Examples of such contacts are shown in Fig. 8. In addition, the proximal dendrites of calretinin-immunoreactive interneurons were found to contact the cell bodies or dendrites of neighbouring calretinin-containing cells (Fig. 8A, B). In five representative sections from P15 and adult rats, the number of such contacts was counted in the CA1 cell body layer. On P15, 6.8&1.4 (S.E.M.) contacts were observed. In adults, contacts were far less frequent and numbered 1.8&0.15. In area CA1 of P15 rats, the cell bodies of the majority of interneurons were located in or near the pyramidal cell body layer. These cells had long, tapering dendrites and branched rarely as they coursed through stratum radiatum and oriens. Examples of this cell type are the neurons labelled 1 and 2 in Fig. 7D. Immunoreactive cell bodies were also found at the interface of stratum radiatum and stratum lacunosum-moleculare. These neurons sent long dendrites through the hippocampal laminae and into stratum oriens (Fig. 7A, C and reconstructed neurons 4 and 5 in D). There were also immunoreactive cell bodies in the more distal aspects of stratum oriens (reconstructed neurons 3 in Fig. 7D), stratum lacunosum-moleculare and near the hippocampus

fissure. These cells had short dendrites that arborized locally (Fig. 7A, C). In contrast to area CA1, dendrites of calretininimmunoreactive cells in area CA3 had relatively short and more highly branched dendrites (Fig. 7E). These cells appeared to be distributed throughout the subfield but with some preference for the cell body layer. Most often, the dendrites branched locally in this subfield (Fig. 7A, E). A major population of calretinin-immunoreactive cells of the adult area CA3 are the spiny cells of stratum lucidum.22 Cells of this type were numerous in our adult preparation (Fig. 9D), but not identifiable by P15. Although cells were present in this laminae at this age (PIS), they always had short aspiny dendrites (Fig. 7E, cells 1, 2, and 3 and Fig. 8B) and thus could not be clearly identified as immature spiny cells of stratum lucidum. On P15, numerous calretinin-immunoreactive interneurons of the dentate hilus were observed just beneath the granule cell body layer (Fig. 10A, B). Immunoreactive dendrites and axons were seen coursing through the subgranular zone. Some of the cells sent dendrites through the granule cell body layer and into stratum moleculare of the dentate (open arrow, Fig. 10B). Immunoreactive cells were also located in the polymorphic layer and the cell body layer of CA3c. These cells also possessed long, tapering dendrites that on occasion would project through the granule cell body layer into stratum moleculare (arrow, Fig. 10B). As in all of the hippocampal subfields, the dendrites of immunoreactive cells of the dentate hilus appeared less elaborate on P15 than in the adult. Photomicrography and reconstructions of calretininimmunoreactive neurons of the adult rat (Fig. 10C, D) show that the patterns of dendritic branching were more extensive than observed on P15 (Fig. 10A, B).

Calretinin in developing hippocampus

Fig. 7. Distribution and morphological features of calretinin-containing hippocampal interneurons on postnatal day 15 (P15). (A) A lower power photomicrograph showing large numbers of calretininimmunopositive cells, which are numerous in the hippocampal CA1 and CA3 cell body layer and hilus of the dentate gyrus (H). (B) Higher power photomicrograph showing the distribution of soma and dendritic processes in area CA1. Open arrows denote the beaded nature of some dendrites. (C) A panoramic view of area CA1 and the dentate gyrus in higher magnification showing the distribution and morphology of interneurons and their processes. (D) Neuron reconstruction of some interneuronal cell types seen in area CA1. Some cells (1 and 2) have their soma in or near stratum pyramidale (SP) and have long radial dendrites penetrating both stratum radiatum (SR) and stratum oriens (SO). Other cells (such as 4 and 5) have cell bodies at the border between stratum radiatum (SR) and stratum lacunosum-moleculare (SLM) and send dendrites through the hippocampal laminae to stratum oriens (SO). Cells labelled 6 and 7 actually consist of two cells in apposition. (E) Interneurons of area CA3 have dendrites that are arranged horizontal to stratum pyramidale (SP). These are short and highly branched. Cells labelled 1, 2, and 3 may be forerunners of the spiny cells of stratum lucidum (SL) seen in the adult. Cells labelled 1 are actually two cells in apposition. ML, molecular layer of the dentate gyrus, GL, granule cell body layer. Scale bars=200 µm in A, 50 µm in B, 50 µm in C, 100 µm in D, and 100 µm in E.

1147

1148

M. Jiang and J. W. Swann

Fig. 8. The soma and proximal dendrites of calretinin interneurons are often observed in direct apposition on P15. Photomicrographs show examples of the types of cell contacts observed: A, area CA1; B, area CA3; C, the dentate hilus; D, area CA3. Arrows denote contact of two soma. Arrowheads mark sites of putative dendrodendritic contact. Open arrows show sites of potential contact of proximal dendrites with a soma. SO, stratum oriens, SP, stratum pyramidale, SR, stratum radiatum, SL, stratum lucidum. Scale bars=20 µm in A, 20 µm in B, and 20 µm in C.

DISCUSSION

The major findings of this study are as follows: 1) calretinin is expressed at the earliest times studied (E15) in cells and processes of the primordial plexiform layer. 2) Later in gestation (E20) and at birth, two major neuronal populations are observed. The first are neurons bordering the hippocampal plate and subplate. Their morphology, laminar location, and packing density suggest they are migrating pyramidal cells. The second are multipolar neurons of the inner marginal zone. Both cell types express calretinin only transiently and are not observed in samples taken during week 2. 3) By P3, Cajal–Retzius cells are observed in stratum lacunosum-moleculare of area CA1 and the dentate gyrus. On P7, these cells have ovoid somas and very fine, beaded processes. Thereafter, they disappear. 4) Between P7 and P10, interneurons of the hippocampus and dentate gyrus emerge. By P15, they are more numerous in some hippocampal laminae than in the adult. Although their patterns of dendritic branching are less elabo-

rate than those of their adult counterparts, their soma and proximal dendrites are often in direct contact with each other. Transient calretinin immunoreactivity in pyramidallike cells and multipolar neurons of the inner marginal zone One of the more unexpectant findings of this study was the transient expression of calretinin in a large number of hippocampal pyramidal-like cells in late gestational and neonatal specimens (Figs 3, 4). In the adult and developing cortical structures, calretinin is thought to be expressed preferentially in subpopulations of inhibitory interneurons that use GABA as their neurotransmitter.35 Indeed, several studies have co-localized GABA in the majority of neocortical neurons that express calretinin.16,35,50 One exception is the Cajal–Retzius cells of the neocortical marginal zone and layer 1 of the neonate.16,25,50 These cells are likely to use glutamate as their neurotransmitter.11 Another possible exception stems from a recent study

Calretinin in developing hippocampus

1149

Fig. 9. Distribution and anatomical features of calretinin-containing interneurons in the adult (P50) rat. (A) A lower power photomicrograph. When compared with results on P15 (Fig. 7A), interneurons appear less numerous, but dendrites appear more elaborate and immunoreactive for calretinin. (B) Higher power photomicrograph showing calretinin interneurons in area CA1. (C) Cells have long radial dendrites that often display beaded or varicose features (open arrow). (D) Spiny cells of stratum lucidum (SL). Inset, D1 shows a dendrite with numerous spines (arrows). (E) Drawing depicts laminae in which cell counts were made. Dotted lines separate area CA1 from CA3 and CA3 from the dentate gyrus (DG). SO, stratum oriens, SP, stratum pyramidale, SR, stratum radiatum, SLM, stratum lacunosum-moleculare, ML, molecular layer of the dentate gyrus, GL, granule cell body layer, H, hilus of the dentate gyrus, SL, stratum lucidum. Scale bars=200 µm in A, 60 µm in B, 20 µm in C, 20 µm in D and 5 µm in D1.

of rat neocortex in which a transient expression of calretinin was reported in pyramidal-like neurons in layer VIa and V.16 Expression was particularly notable between P3 and P12. These cells did not co-localize GABA and had a pyramidal-shaped soma and thick apical dendrites. These features coupled with their laminar location and high packing density led the authors to conclude that at least some neocortical pyramidal cells express calretinin during early postnatal life. Thus, our findings of rat hippocampal pyramidal cell immunoreactivity during the perinatal period are similar to these authors’ results.16 However, calretinin expression in pyramidal cells is clearly species-specific, because similar patterns of pyramidal cell immunoreactivity have not been seen in primates.50 Likewise, in a study of mouse hippocampus, antibodies to calretinin were not reported to stain pyramidal cells. However, this study was limited because it examined specimens only

until P5.45 Whether pyramidal cells in species other than the rat show a transient developmental expression of this calcium-binding protein remains to be determined. The appearance of the pyramidal-like cells lead us to wonder if these cells could be late-born pyramidal cells that are in their final stages of migration to the hippocampal plate. Yan et al.50 have described ‘‘tadpole-like’’ calretinin-immunoreactive neurons in primate neocortex that were thought to be interneurons in the process of radial migration. Recent studies have described neurons that were visualized confocally during migration in in vitro neocortical slices.27,37 These neurons were bipolar and had large cell bodies with thick leading processes and a smaller, thinner trailing process. The cells bordering the hippocampal plate and subplate shown in Figs 3, 4 share many of these features. Considering the birth dates of rat hippocampal pyramidal cells and their rates of

1150

M. Jiang and J. W. Swann

Fig. 10. Comparison of calretinin-immunopositive interneurons in the hilus of the dentate gyrus (H) and area CA3c. Photomicrograph in A (P15) and C (adult) show some features of interneurons positioned just below the granule cell body layer (GL) of the dentate gyrus. At both ages, a plexus of immunopositive dendrites and axons are seen. Neuron reconstruction in B (P15) shows that the dendrites in early life can be long and in some instances penetrate the granule cell layer to reach the molecular layer (ML) of the dentate (open arrow). This can also be true for dendrites (arrow) whose soma lie in stratum pyramidale (SP) of CA3c. D) However, the branching pattern of dendrites in the adult are more complex, showing that interneurons in week 2 are anatomically immature and continue to ramify processes. Scale bars=20 µm in A and C, 100 µm in B and D.

migration, one could easily envision these cells as the last generation of pyramidal cells that are in their final stage of migration to and through the hippocampal plate. In the rat hippocampus, pyramidal cells are born on E16–E19.7,40 In area CA1, a large fraction of pyramidal cells are born on E19.40 Studies of rates of migration suggest that it can take up to five days for hippocampal neurons to complete their migration to the cell body layer.36 Thus, it is easy to imagine that the immunoreactive cells of the subplate described here on E20 and P0 are in fact a population of late born cells that for unknown reasons transiently express calretinin as they approach their final destination in stratum pyramidale. Additional experiments will be required to address this issue. Concurrent with the immunostaining of pyramidal-like cells, a large number of neurons are stained in the inner molecular layer on E20 and stratum radiatum on P0. To our knowledge, cells

with similar morphological features have not been described previously in the developing hippocampus, although one immunohistochemical study of GABA in neocortex described neurons with similar features.12 Their short and highly irregular dendritic arbors may reflect ongoing processes of growth and ramification. These are likely the forerunners of interneurons located in this laminae at later postnatal ages. Emergence and disappearance immunoreactive Cajal–Retzius cells

of

calretinin-

From the outset, one of the goals of this study was to describe the ontogeny of Cajal–Retzius cells of rat hippocampus. Previous studies had demonstrated that calretinin antibodies were particularly useful in visualizing this neuronal population during early brain development. Thus, we were surprised when we were unable to identify these cells convincingly until

Calretinin in developing hippocampus

P1–P3. In some ways, these results may be similar to those reported by Fonseca et al.16 in rat neocortex. Although these authors were able to describe calretinin-immunoreactive cells in the marginal zone as early as E14, they noted a decrease in calretinin immunoreactivity in the marginal zone at E17–E20 and a reappearance of calretinin at later postnatal ages. In hippocampus, a population of intensely immunoreactive cells were present in the marginal zone on both E15 and E17. It is likely that some of these marginal zone cells (Fig. 2) are Cajal–Retzius cells. However, an apparent decrease in immunoreactivity in the outer marginal zone, and the presence of innumerable calretinin-immunoreactive neurons in the inner marginal zone and stratum radiatum on E20 and P0 made identification of Cajal–Retzius cells difficult at these ages. This problem was resolved when the former cells dramatically decreased in number by P3 (Fig. 5). These developmental patterns of Cajal–Retzius cell immunoreactivity in the rat appear to be different from that of mouse. For instance, del Rio et al.11 described intense immunoreactivity in the marginal zone and layer 1 of neocortex at all late gestational and early postnatal ages (E14–P8). In this and one other study, calretinin-immunoreactive cells, identified as Cajal–Retzius cells, were clearly evident in mouse hippocampus at these ages.11,45 These results clearly differ from the results presented here in the rat and again point out significant interspecies differences in calretinin expression during development. Results from numerous studies suggest that Cajal–Retzius cells are a transient cortical population9,13,29,48 that likely undergo apoptotic degeneration during the second postnatal week in rodents.11 Our observations that immunoreactive cell bodies were swollen on P7 and had thin, beaded processes (Fig. 5) do not directly address this issue but seem to be in keeping with those earlier observations in neocortex. Perhaps the strongest evidence that Cajal–Retzius cells actually degenerate comes from pulse injections of bromodeoxyuridine on E10– E11, which selectively labelled the majority of these cells on E18–P5. When animals were killed on P10 or later the number of bromodeoxyuridine-positive neurons in neocortex dropped abruptly. In some instances, pyknotic nuclei were observed that were bromodeoxyuridine positive. These events were concurrent with the disappearance of calretininimmunoreactive cells.11 There has been much discussion on the role Cajal– Retzius cells play in early neocortical and hippocampal development. In neocortex, speculation has implicated these cells in neuronal migration.11,31 However, in hippocampus the presence of these cells in stratum lacunosum-moleculare and the molecular layer of the dentate gyrus in early postnatal life has prompted others to suggest that these cells may play a role in the guidance of axons of the perforant path.45

1151

Maturation of calretinin-containing interneurons in postnatal life Beginning on P7, a pattern of cellular staining emerged that was similar to that of the adult rat. However, there were notable differences. First, between P7 and P15, the dendrites of calretininimmunoreactive cells were far less extensive than those of the adult. This could be because the dendrites are still growing and have yet to elaborate long and complex dendritic arbors or because calretinin is not efficiently transported into the more distal segments of dendrites in early life. We favour the former explanation because immunoreactive dendrites could often be followed long distances to their termination points some distance from their soma (see the reconstructions of hilar interneurons in Fig. 10). The immaturity of the dendrites of interneurons is perhaps most dramatically illustrated by the absence of clearly identifiable spiny cells in stratum lucidum on P15 (Figs 7, 8B). These cells in the adult rat (Fig. 9D) form a relatively homogeneous and easily identifiable subpopulation of calretinin-containing interneurons.22 Although calretinin-positive cells with soma in stratum lucidum are present on P15, their dendrites are short and lack spines. On the other hand, stratum lucidum of the adult is populated by a meshwork of spiny dendrites arising from these cells. This meshwork is not present on P15. It is interesting to speculate that the dendrites of these cells might be the last to mature in hippocampus because their growth is influenced in some way by the maturation of the mossy fibre pathway, which is well-known to form late in hippocampal development due to the postnatal birth of dentate granule cells. Another striking feature of calretinin immunostaining on P15 is the large number of immunoreactive interneurons that are present. In area CA1 and stratum radiatum of CA3, the density far exceeds that observed in the adult (Figs 7, 9, Table 1). A similar peak in the density of calretinin-containing interneurons has been reported in rat neocortex.16 It seems unlikely that this loss of immunoreactive cells is consequent to the death of a select subpopulation of developing interneurons. Instead, it is more likely that certain hippocampal interneurons transiently express calretinin during early postnatal life while others maintain this calcium-binding protein into adulthood. A related observation is that the number of calretinin-containing interneurons that appear to make contact with other calretinin-immunoreactive cells is unusually high on P15 (Fig. 8). Thus, it would appear that if some hippocampal interneurons do down-regulate calretinin expression, it would be a subpopulation of cells that are in physical contact with each other, since such contacts are rarely observed in adulthood. The high incidence of this type of cell apposition early in life is interesting in light of recent studies in adult hippocampus that report the ability of

1152

M. Jiang and J. W. Swann

interneurons to interact directly with each other in ways that lead to sychronized discharging of networks of inhibitory interneurons.33 Electrophysiological studies suggest that direct interneuron-tointerneuron interactions are likely mediated by both chemical and electrical synapses.34 Recently, Gulyas et al.21 showed that radially running dendrites of calretinin-containing interneurons in adult area CA1 bundle in long braid-like structures. The dendrites of one calretinin interneuron can be in direct contact with many other dendrites within such bundles. These regions of contact are suspected sites of electrotonic coupling via gap junctions. Our results suggest that on P15 the soma and proximal dendrites of calretinin interneurons may be in contact to a greater extent than in adulthood. This could mean that the types of interaction between subpopulations of interneurons change with age. However, at this time it is impossible to draw such a conclusion since demonstration of such interactions must await results of ultrastructural studies. Electrophysiological studies appear to support the idea that networks of inhibitory interneurons exist in early life and may be more highly interactive than in adulthood.32,43 Another study has demonstrated a high frequency of dye coupling between electrophysiologically characterized interneurons filled with biocytin during the second postnatal week in the rat.19 Dye coupling has long been thought to be a marker of electrotonic coupling via gap junctions. Does calretinin have a role in hippocampal development? Calretinin has been used as an anatomical marker for subpopulations of hippocampal interneurons in the adult.21,22,35 However, only one study of hippocampal development has been conducted, and it was in the mouse.45 Here, we report that calretinin is a good marker for a variety of developing rat hippocampal neurons. However, this calcium-binding protein is expressed only transiently and at different times in each of these neuronal populations. This would suggest that calretinin plays a role in developmental processes that take place over a limited period. At earlier and later times, neurons are either

less dependent upon the regulatory processes controlled by calretinin or this calcium-binding protein is replaced by other proteins, such as calbindin-D28k or parvalbumin. The role any calcium-binding protein plays in neuronal function remains unclear.5 These proteins have been proposed to function as buffers for intracellular calcium and as such potentially play important roles in neurodevelopment. Because each of the calcium-binding proteins has a different pattern of developmental expression, they likely play quite different roles in neuronal maturation. Calretinin has been proposed to participate, at least indirectly, in synaptogenesis in retina.14 Other authors have suggested roles for calcium-binding proteins in developmental processes as diverse as axonal elongation, dendritic remodelling, and the generation of cytoskeletal elements.4,15,23 The delineation of their functional roles in different neuronal populations at different times in hippocampal development must await further investigation. CONCLUSION

The present findings demonstrate that the ontogeny of calretinin-containing neuronal elements in rat hippocampus is complex and that many neurons express this calcium-binding protein for only brief periods in embryonic or postnatal life. Some neuronal populations, such as the Cajal–Retzius cells, likely die early in life. This would in part explain the appearance and then disappearance of at least some immunopositive cells. However, other cells express this calcium-binding protein during distinct phases of maturation. This suggests that calretinin plays a role in regulating some aspect of neuronal differentiation. What this role may be and whether it involves regulation of intracellular calcium activity is unknown at this time. None the less, our results demonstrate that calretinin immunohistochemistry can be a valuable tool in identifying and studying select subpopulations of developing hippocampal neurons at distinct times in brain development. Acknowledgements—This work was supported by NIH grants NS11535 and NS18309.

REFERENCES

1. 2. 3. 4. 5. 6. 7.

Abercrombie M. (1946) Estimation of nuclear population from microtome sections. Anat. Rec. 94, 239–247. Al-Mohanna F. A., Cave J. and Bolsover S. R. (1992) A narrow window of intracellular calcium concentration is optimal for neurite outgrowth in rat sensory neurones. Devl Brain Res. 70, 287–290. Alca´ntara S. and Ferrer I. (1994) Postnatal development of parvalbumin immunoreactivity in the cerebral cortex of the cat. J. comp. Neurol. 348, 133–149. Andressen C., Blumche I. and Celio M. R. (1993) Calcium-binding proteins, selective markers of nerve cells. Cell Tiss. Res. 271, 181–208. Baimbridge K. G., Celio M. R. and Rogers J. H. (1996) Calcium-binding proteins in the nervous system. Trends Neurosci. 15, 303–308. Bastianelli E. and Pochet R. (1993) Transient expression of calretinin during development of chick cerebellum. Neurosci. Res. 17, 53–61. Bayer S. A. (1980) Development of the hippocampal region in the rat. I. Neurogenesis examined with 3H-thymidine autoradiography. J. comp. Neurol. 190, 87–114.

Calretinin in developing hippocampus 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41.

1153

Bergmann I., Nitsch R. and Frotscher M. (1991) Area-specific morphological and neurochemical maturation of non-pyramidal neurons in the rat hippocampus as revealed by parvalbumin immunocytochemistry. Anat. Embryol. 184, 403–409. Bradford R., Parnavelas G. J. and Lieberman A. R. (1996) Neurons in layer 1 of the developing occipital cortex of the rat. J. comp. Neurol 176, 121–132. Constantine-Paton M., Cline H. T. and Debski E. (1990) Patterned activity, synaptic convergence, and the NMDA receptor in developing visual pathways. A. Rev. Neurosci. 13, 129–154. Del Rio J. A., Martinez A., Fonseca M., Auladell C. and Soriano E. (1995) Glutamate-like immunoreactivity and fate of Cajal–Retzius cells in the murine cortex as identified with calretinin antibody. Cerebral Cortex 1, 13–21. Del Rio J. A., Soriano E. and Ferrer I. (1992) Development of GABA-immunoreactivity in the neocortex of the mouse. J. comp. Neurol. 326, 501–526. Derer P. and Derer M. (1990) Cajal–Retzius cell ontogenesis and death in mouse brain visualized with horseradish peroxidase and electron microscopy. Neuroscience 36, 839–856. Ellis J. H., Richards D. E. and Rogers J. H. (1991) Calretinin and calbindin in the retina of the developing chick. Cell Tiss. Res. 64, 197–208. Enderlin S., Norman A. W. and Celio M. R. (1987) Ontogeny of the calcium binding protein Calbindin D-28k in the rat nervous system. Anat. Embryol. 177, 15–28. Fonseca M., Del Rı´o J. A., Martı´nez A., Go´mez S. and Soriano E. (1995) Development of calretinin immunoreactivity in the neocortex of the rat. J. comp. Neurol. 361, 177–192. Friauf E. (1994) Distribution of calcium-binding protein calbindin-D28k in the auditory system of adult and developing rats. J. comp. Neurol. 349, 193–211. Gomez T. M., Snow D. M. and Letourneau P. C. (1995) Characterization of spontaneous calcium transients in nerve growth cones and their effect on growth cone migration. Neuron 14, 1233–1246. Gomez-DiCesare C. M., Smith K. L., Rice F. L. and Swann J. W. (1996) Anatomical properties of fast spiking cells that initiate synchronized population discharges in immature hippocampus. Neuroscience, 75, 83–97. Goodman J. H., Wasterlain C. G., Massarweh W. F., Dean E., Sollas A. L. and Sloviter R. S. (1993) Calbindin-D28k immunoreactivity and selective vulnerability to ischemia in the dentate gyrus of the developing rat. Brain Res. 606, 309–314. Gulyas A. I., Hajos N. and Freund T. F. (1996) Interneurons containing calretinin are specialized to control other interneurons in the rat hippocampus. J. Neurosci. 16, 3397–3411. Gulyas A. I., Miettinen R., Jacobowitz D. M. and Freund T. F. (1992) Calretinin is present in non-pyramidal cells of the rat hippocampus. I. A new type of neuron specifically associated with the mossy fibre system. Neuroscience 48, 1–27. Hendrickson A. E., Van Brederode J. F. M., Mulligan K. A. and Celio M. R. (1991) Development of calcium-binding protein parvalbumin and calbindin in monkey striate cortex. J. comp. Neurol. 307, 626–646. Hogan D. and Berman N. E. J. (1993) Transient expression of calbindin-D28k immunoreactivity in layer V pyramidal neurons during postnatal development of kitten cortical areas. Devl Brain Res. 74, 177–192. Huntley G. W. and Jones E. G. (1990) Cajal–Retzius neurons in developing monkey neocortex show immunoreactivity for calcium binding proteins. J. Neurocytol. 19, 200–212. Kater S. B., Mattson M. P., Cohan C. S. and Connor J. A. (1988) Calcium regulation of the neuronal growth cone. Trends Neurosci. 11, 315–321. Komuro H. and Rakic P. (1992) Selective role of N-type calcium channels in neuronal migration. Science 257, 806–809. Komuro H. and Rakic P. (1996) Intracellular Ca2+ fluctuations modulate the rate of neuronal migration. Neuron 17, 275–285. Konig N., Hornung J.-P. and Van Der Loos H. (1981) Identification of Cajal–Retzius cells in immature rodent cerebral cortex: a combined Golgi–EM study. Neurosci. Lett. 27, 225–229. Liu F.-C. and Graybiel A. M. (1992) Transient calbindin-D28K-positive systems in the telencephalon: ganglionic eminence, developing striatum and cerebral cortex. J. Neurosci. 12, 674–690. Marin-Padilla M. (1990) Three-dimensional structural organization of layer I of the human cerebral cortex: a Golgi study. J. comp. Neurol. 299, 89–105. Michelson H. B. (1994) Synchronization of inhibitory neurons by glutamate-independent processes in the developing rat. Neurosci. Abstr. 20, 640. Michelson H. B. and Wong R. K. S. (1991) Excitatory synaptic responses mediated by GABAA receptors in the hippocampus. Science 253, 1420–1423. Michelson H. B. and Wong R. K. S. (1994) Synchronization of inhibitory neurones in the guinea-pig hippocampus in vitro. J. Physiol., Lond. 477, 35–45. Miettinen R., Gulyas A. I., Baimbridge K. G., Jacobowitz D. M. and Freund T. F. (1992) Calretinin is present in non-pyramidal cells of the rat hippocampus. II. Co-existence with other calcium binding proteins and GABA. Neuroscience 48, 29–43. Nowakowski R. S. (1985) Neuronal migration in hippocampal lamination defect (Hld) mutant mouse. In Cellular and Molecular Control of Direct Cell Interactions (ed. Marthy H.-J.), pp. 133–154. Plenum, New York. O’Rourke N. A., Dailey M. E., Smith S. J. and McConnell S. K. (1992) Diverse migratory pathways in the developing cerebral cortex. Science 258, 299–302. Rehder V. and Kater S. B. (1992) Regulation of neuronal growth cone filopodia by intracellular calcium. J. Neurosci. 12, 3175–3186. Scharfman H. E. and Schwartzkroin P. A. (1989) Protection of dentate hilar cells from prolonged stimulation by intracellular calcium chelation. Science 246, 257–260. Schlessinger A. R., Cowan W. M. and Swanson L. W. (1978) The time of origin of neurons in Ammon’s horn and the associated retrohippocampal fields. Anat. Embryol. 154, 153–173. Singer W. (1995) Development and plasticity of cortical processing architectures. Science 270, 758–764.

1154 42. 43. 44. 45. 46. 47. 48. 49. 50.

M. Jiang and J. W. Swann Sloviter R. S. (1989) Calcium-binding protein (calbindin-D28K) and parvalbumin immunocytochemistry: localization in the rat hippocampus with specific references to the selective vulnerability of hippocampal neurons to seizure activity. J. comp. Neurol. 280, 183–196. Smith K. and Swann J. (1994) Evidence for interneuronal networks in the developing hippocampal area CA3. Epilepsia 35, 60. Solbach S. and Celio M. R. (1991) Ontogeny of the calcium binding protein parvalbumin in the rat nervous system. Anat. Embryol. 184, 103–124. Soriano E., Del Rio J. A., Martinez A. and Super H. (1994) Organization of the embryonic and early postnatal murine hippocampus. I. Immunocytochemical characterization of neuronal populations in the subplate and marginal zone. J. comp. Neurol. 342, 571–595. Spitzer N. C. (1994) Spontaneous Ca2+ spikes and waves in embryonic neurons: signaling systems for differentiation. Trends Neurosci. 17, 115–118. Spitzer N. C., Debaca R. C., Allen K. A. and Holliday J. (1993) Calcium dependence of differentiation of GABA immunoreactivity in spinal neurons. J. comp. Neurol. 337, 168–175. Vogt Weisenhorn D. M. and Weruaga Prieto E. (1994) Localization of calretinin in cells of layer I (Cajal–Retzius cells) of the developing cortex of the rat. Devl Brain Res. 82, 293–297. Weiss J. H., Koh J. H., Baimbridge K. G. and Choi D. W. (1990) Cortical neurons containing somatostatin- or parvalbumin-like immunoreactivity are atypically vulnerable to excitotoxic injury in vitro. Neurology 40, 1288–1292. Yan Y.-H., Van Brederode J. F. M. and Hendrickson A. E. (1995) Developmental changes in calretinin expression in GABAergic and nonGABAergic neurons in monkey striate cortex. J. comp. Neurol. 363, 78–92. (Accepted 28 April 1997)