16643136

(1)

Expression of calcium-binding proteins in the mouse claustrum

M

a

A

´ ngeles Real, Jose´ Carlos Da´

v

ila, Sal

v

ador Guirado

*

Departamento de Biologı´a Celular, Gene´tica y Fisiologı´a, Universidad de Ma´laga, Campus de Teatinos, 29071 Ma´laga, Spain

Received 12 July 2002; received in revised form 29 October 2002; accepted 1 December 2002

Abstract

The present paper describes the distribution of three calcium-binding proteins (calbindin D28k, calretinin, and parvalbumin) in the mouse dorsal claustrum and endopiriform nucleus. The three calcium-binding proteins were distinctly expressed in structures of both the claustrum and the endopiriform nucleus. Calbindin was the calcium-binding protein showing the highest expression in the claustrum and the endopiriform nucleus. In contrast, calretinin-immunoreactive structures, particularly cell bodies, werevery scarce in these regions. Both calbindin-immunoreactive and parvalbumin-immunoreactive neurons were more abundant in the claustrum than in the endopiriform nucleus, and more in rostral than in caudal levels. Nevertheless, calcium-binding protein immunoreactive neurons constitute a minority population of claustral neurons. The colocalization study of calbindin and parvalbumin immunoreactivities has demonstrated that both calcium-binding proteins are mostly expressed by separate claustral neurons in the mouse. On the other hand, our results on parvalbumin and calretinin immunoreactivity match a novel subdivision of the mouse claustrum mostly based on the pattern of cadherin expression [Neuroscience 106 (2001) 505]. In this sense, we propose that a specific zone of the dorsal claustrum with cell bodies that strongly express Rcad and cadherin-8 would be the selective target for parvalbumin-expressing fibers, and that they would be mostly avoided by calretinin-expressing axons.

Keywords: Calbindin; Parvalbumin; Calretinin; Lateral pallium; Ventral pallium

1. Introduction

The claustrum is a pallial subcortical region present in all mammals. Two parts of the claustrum can be distinguished: the dorsal part is usually called claustrum proper and it is located deep to the insular cortex (therefore it is also named dorsal or insular claustrum); the ventral part is called endopiriform nucleus and is located deep to the piriform cortex (Druga, 1966; Sherk,

1988; Dinopoulos et al., 1992).Krettek and Price (1977)

considered a ventral division of the endopiriform nucleus located deep to the periamygdaloid cortex and adjoiningventral part of the middle region of piriform cortex. For the sake of clarity, from now on the dorsal claustrum will be referred to as claustrum whereas the ventral claustrum will be referred to as endopiriform nucleus.

The claustrum has extensive connections with the neocortex (Pearson et al., 1982; Macchi et al., 1983; Markowitsch et al., 1984; Li et al., 1986; Sloniewski et al., 1986; Sherk, 1988; Sadowski et al., 1997; Kowianski

et al., 1998; Majak et al., 2000), whereas the

endopiri-form nucleus is hodologicaly related with the prepiri-form and entorhinal cortices (Druga, 1971;

Markowitsch et al., 1984; Witter et al., 1988).

Despite the evidence for hodologicaly different zones of the claustrum, previous cyto and chemoarchitectonic studies show an overall uniform structure (Sherk, 1988;

Reynhout and Baizer, 1999). Recently, however, three

subdivisions within the claustrum have been distin-guished on the basis of specific cadherin expression patterns (Obst-Pernberg et al., 2001). It is not known whether these claustral subdivisions have different connections or play different functions.

In this context, the expression of calcium-binding proteins has probed to be very useful in revealing subdivisions in different regions of the central nervous system, notably the thalamus (Jones and Hendry, 1989;

Da´vila et al., 2000). Therefore, here we studied the

* Corresponding author. Tel.:/34-952-13-1961; fax:/

34-952-13-2000.

E-mail address: [email protected](S. Guirado).

www.elsevier.com/locate/jchemneu

(2)

distribution of three calcium-binding proteins (calbindin D28k, calretinin, and parvalbumin) in the mouse claustrum searching for specific distribution patterns putatively related either to hodologicaly-based subdiv i-sions or cadherin expression patterns.

2. Material and methods

Fifteen adult OF1 mice (42/46 g body weight) were used in the present study. Throughout the experimental work animals were treated according to the European Communities Council directive (86/609/EEC) on treat-ment of experitreat-mental animals.

Mice were deeply anesthetized with sodium pentothal (65 mg/kg; Abbott Laboratories) and transcardially perfused with 0.1 M phosphate-buffered saline (PBS), pH 7.4, followed by 4% paraformaldehyde, 0.1% glutaraldehyde and 0.2% picric acid in PBS at room temperature for 30 min. The brains were then removed and stored in 4% paraformaldehyde and 0.2% picric acid in PBS at 4 8C overnight; afterwards they were embedded in 4% agar and cut into 50-mm-thick frontal sections, using avibratome. The sections were washed extensively in PBS prior to immunocytochemical stain-ing with the peroxidase/antiperoxidase method.

Free-floating sections were first incubated in 2% normal goat serum and 0.3% Triton X-100 in PBS at room temperature for 1 h, to block nonspecific binding of the antibodies and permeate the tissues, respectively, and then were transferred to the primary antibody. The three polyclonal antibodies, anti-parvalbumin, anti-calretinin, and anti-calbindin, were raised in rabbits (SWant, Bellinzona, Switzerland) and used at a dilution of 1:2000 for 18 h. After three washes in PBS for 45 min, the sections were incubated in goat anti-rabbit IgG diluted 1:35 for 1 h, washed again in PBS for 45 min, and incubated in peroxidase/antiperoxidase diluted 1:100 for 1 h. The immunolabeling was revealed with 0.05% diaminobenzidine (DAB; Sigma Chemical Co, St. Louis, MO), 0.05% nickel ammonium sulfate and 0.03%

hydrogen peroxide (H2O2) in PBS. All steps were carried out at room temperature with gentle agitation. After a thorough wash in PBS, the sections were mounted on gelatinized slides, air dried, dehydrated in ethanol, cleared in xylene and coverslipped with DPX (BDH, Poole, UK).

In order to provide some quantitative data about the populations of calcium-binding protein immunoreactive neurons as well as the putative colocalization of these proteins in the same neuron, we carried out post-embedding immunocytochemistry on semithin sections of the claustrum in one mouse. After fixation, brain slices 500-mm-thick, obtained on a vibratome, were collected in PBS, dehydrated in acetone, and flat embedded in Araldite 502 (Sigma). Resin-embedded slices containing the claustrum were glued onto poly-merized blocks, trimmed and then sectioned on an ultramicrotome. Series of three sections of 1.5 mm were made. The first series was stained with toluidine blue, whereas the second and third series were stained with anti-calbindin and anti-parvalbumin, respectively. After removing the resin with sodium ethoxide, sections were processed according to an immunohistochemical proce-dure similar to that described above, except for that the primary antibodies were used at a higher concentration (anti-CB: 1/1000; anti-PV: 1/500). To determine coex-istence, we took into consideration only those cells clearly positive to calbindin or parvalbumin and dis-playing a wide nuclear profile.

Controls: as controls of the immunohistochemical method used in the present study, sections were pro-cessed as indicated but the corresponding primary antiserum was replaced by rabbit nonimmune serum (1:500). No immunostaining could be detected under these conditions. In addition, as a control of the specificity of the different primary antisera under the experimental conditions used in this work, we incubated control sections in the primary antibody preadsorbed with the corresponding protein (1 mg/ml of the diluted antibody). As a result, specific immunostaining was completely abolished.

Fig. 1. Calbindin-immunoreactivity. (A) Panoramicview of the claustrum (Cl) and dorsal endopiriform nucleus (DEn) at a rostral level (see Fig. 17 ofPaxinos and Franklin, 2001). Transverse section. Scale bar: 1 mm. (B) Detail of the claustrum and dorsal endopiriform nucleus at an intermediate rostrocaudal level. The claustrum is characterized by the presence of a number of immunoreactive cells within a moderately immunostained neuropil, similarly to the adjacent cortical areas and dorsal endopiriform nucleus. Scale bar: 500mm. (C) Most neuronal cell bodies in the claustrum are devoid of calbindin immunoreactivity and appear as white profiles surrounded by an immunostained neuropil. Scale bar: 100mm. (D) Morphological features of calbindin immunoreactive neurons. Most stained neurons are multipolar cells with aspiny dendrites. Scale bar: 100mm. (E) Pairs of calbindin immunostained neurons with apposed cell bodies are sometimes observed in the claustrum (arrow). A number of lightly stained somata can also be observed. Scale bar: 100mm. (F) Panoramicview of the endopiriform region at a caudal level (see Fig. 41 ofPaxinos and Franklin, 2001). Both the number of immunoreactive neurons and the intensity of neuropil staining decrease at these levels. Scale bar: 1 mm. aca, Anterior commissure, anterior; AID, agranular insular cortex, dorsal; AIP, agranular insular cortex, posterior; AIV, agranular insular cortex,ventral; BLA, basolateral amygdaloid nucleus, anterior; Cl, claustrum; CPu, caudate putamen; DEn, dorsal endopiriform nucleus; DI, dysgranular insular cortex; ec, external capsule; La, lateral amygdaloid nucleus; lo, lateral olfactory tract; Pir, piriform cortex; Ven,ventral endopiriform nucleus.

M.A´ . Real et al. / Journal of Chemical Neuroanatomy 25 (2003) 151/160 152

(3)

(4)

Light microscopic images were photographed by using a Leica microscope equipped with a Nikon DXM1200 digital camera. Digital images were loaded into ADOBE PHOTOSHOP software and converted to grayscale images. Brightness and contrast were adjusted for the final images. No additional filtering or manip-ulation of the images was performed. The final figures were composed and labeled with ADOBE PAGEMAKER software and printed with an Epson Photo 750 printer.

3. Results

The three calcium-binding proteins studied in this work were distinctly expressed in structures of both the claustrum and the endopiriform nucleus. We will next describe separately calbindin-immunoreactivity in the claustrum and in the endopiriform nucleus, and then calretinin- and parvalbumin-immunoreactivities, respec-tively.

3.1. Calbindin-immunoreactivity

3.1.1. Claustrum

Sections immunostained for calbindin showed an overall stronger staining of both neurons and neuropil than those stained for either of the other antibodies; accordingly numerous calbindin-immunoreactive (ir) neurons embedded in a moderately immunostained neuropil were found from rostral to caudal levels of the claustrum (Fig. 1A and B). Nevertheless, the number of immunostained cells decreased slightly at caudal levels. As in other parts of the telencephalon, two cell-staining patterns for calbindin were observed through-out the claustrum: darkly stained neurons with a Golgi-like appearance, and lightly stained cell bodies with barelyvisible processes (Fig. 1D and E). Cell morphol-ogiesvaried from small, round or elongated neurons to medium-sized multipolar neurons (Fig. 1C and D). These distinct cell types appeared intermingled through-out the claustrum. Multipolar calbindin-ir neurons displayed typically three to five processes splitting into a few secondary aspiny dendrites. Sometimes, pairs of calbindin-ir multipolar neurons with apposed somata were observed within the claustrum (Fig. 1E).

The uniform background staining of the claustrum and adjacent cortical areas made it difficult to delineate the borders of the claustrum in calbindin immunos-tained sections (Fig. 1B). A moderately stained neuropil extended over the whole anterior/posterior extent of the claustrum. Neuropil staining consisted ofvaricose axons oriented in all directions and puncta surrounding numerous immunonegative cell profiles (Fig. 1C). Although many perisomatic calbindin-positive terminals were observed, they did not form pericellular baskets.

3.1.2. Endopiriform nucleus

As in the claustrum, calbindin-ir neurons in the endopiriform nucleus exhibited a variety of staining intensities, sizes, and morphologies. Small or medium-sized multipolar cells were the most common types, although pyramidal-shaped cell bodies were also found at intermediate levels of the dorsal endopiriform nu-cleus. In the dorsal endopiriform nucleus, the number of calbindin immunostained neurons was highest at rostral levels (Fig. 1A), decreasing at intermediate and caudal levels (Fig. 1F). On the other hand, the number of calbindin immunoreactive neurons in the ventral en-dopiriform nucleus was similar to that found in the caudal part of the dorsal endopiriform nucleus. Pairs of calbindin immunostained cells with apposed somata were also found in the endopiriform nucleus.

A uniform, moderately stained, neuropil extended throughout both the dorsal and ventral regions of the endopiriform nucleus (Fig. 1F). Calbindin-positive ax-ons bearingvaricosities were observed in this neuropil.

3.2. Calretinin-immunoreactivity

3.2.1. Claustrum

Calretinin immunostaining in the claustrum consisted of a few scattered positive cells against a background of stained fibers and boutons. Most calretinin-ir neurons were densely stained and exhibited small, round or elongated cell bodies (Fig. 2C), from which two or three aspiny dendrites arose. Some immunostained bipolar neurons located near the external capsule had their cell bodies and dendrites oriented parallel to it whereas other calretinin-ir neurons extended their dendrites, perpendicular to the external capsule and to the cortical surface.

In contrast to the piriform cortex and to the subjacent striatum, the claustrum is characterized by the presence of a dense calretinin-ir neuropil consisting of both stained fibers and puncta (Fig. 2A and B). Nevertheless, the neuropil immunostaining was unevenly distributed within the claustrum. A calretinin-negative neuropil zone, virtually devoid of immunoreactive fibers or puncta, appeared as an oval region in the core of the claustrum (Fig. 2B and C). This oval region was also largely devoid of calretinin-ir neurons, which, if present, were found at the periphery of the region (Fig. 2C). The calretinin-negative region was clearlyvisible from inter-mediate to caudal levels of the claustrum since it was surrounded by immunopositive structures, including a thin deep layer of neuropil that separates it from the fibers of the external capsule (Fig. 2C), and a superficial layer of neuropil continuous with the deep layers of the neighboring insular cortex.

(5)

The overall calretinin immunostaining in the endo-piriform nucleus was similar to that of the claustrum and the insular cortex. Few scattered neurons embedded in a moderate-to-dense immunoreactive neuropil were found in the endopiriform nucleus. The most abundant calretinin-ir cell type consisted of small bipolar cells with

the elongated cell body and processes oriented parallel to the external capsule. Bipolar neurons were observed in both the dorsal and ventral endopiriform nuclei.

Neuropil staining was highest at intermediate levels in the region of the dorsal endopiriform nucleus adjacent to the claustrum (Fig. 2B). For the rest of the dorsal endopiriform nucleus and the whole ventral endopiri-Fig. 2. Calretinin-immunoreactivity. (A) Transverse section at a rostral level of the mouse telencephalon (see Fig. 16 ofPaxinos and Franklin, 2001) showing the claustrum (Cl) and the dorsal endopiriform nucleus (DEn). Scale bar: 1 mm. (B) At intermediate rostrocaudal levels (see Fig. 31 of

Paxinos and Franklin, 2001), the claustrum, the dorsal endopiriform nucleus, and the insular cortex are characterized by a moderately immunoreactive neuropil. An oval immunonegative area can be recognized within the claustrum. Scale bar: 1 mm. (C) Detail of the boxed area in B. This oval area isvirtually devoid of immunoreactive fibers. A small stained neuron can be observed in the periphery of the negative region (arrow). Scale bar: 100mm. (D) Panoramicview of the endopiriform region at a caudal level (see Fig. 41 ofPaxinos and Franklin, 2001). Scale bar: 1 mm. aca, anterior commissure, anterior; acp, anterior commissure, posterior; AID, agranular insular cortex, dorsal; AIP, agranular insular cortex, posterior; AIV, agranular insular cortex, ventral; BLA, basolateral amygdaloid nucleus, anterior; Cl, claustrum; CPu, caudate putamen; DEn, dorsal endopiriform nucleus; DI, dysgranular insular cortex; ec, external capsule; GI, granular insular cortex; La, lateral amygdaloid nucleus; Pir, piriform cortex; VEn,ventral endopiriform nucleus.

(6)

form nucleus, the neuropil was moderately immunos-tained for calretinin (Fig. 2D). Calretinin immunoreac-tive axons were fine and displayed smallvaricosities.

3.3. Parvalbumin-immunoreactivity

3.3.1. Claustrum

Parvalbumin immunostaining in the claustrum con-sisted of a number of immunoreactive neurons sur-rounded by a dense meshwork of stained fibers and puncta (Fig. 3C and D). Both darkly and lightly stained neurons were found throughout the claustrum but they were less abundant at the caudal part of the claustrum. Most parvalbumin-ir cells were medium in size and had multipolar cell bodies (Fig. 3C) with several thin, beaded dendrites that extended for long distances.

Parvalbumin-ir varicose axons were oriented in all directions within the claustrum (Fig. 3C and D). However, it was unclear whether these axons formed perisomatic contacts.

In sections immunostained for parvalbumin, the claustrum was characterized by the presence of a sharply defined, moderately stained patch of neuropil particu-larly evident at intermediate and caudal levels of the claustrum (Fig. 3B). The deep layers of the adjacent agranular insular cortex presented also a patch of moderately stained neuropil, and between these two neuropil patches there was an intervening immunone-gative cell-poor zone that was most apparent at inter-mediate levels of the claustrum (Fig. 3B).

The patch of parvalbumin immunopositive neuropil in the claustrum roughly corresponds to the calretinin-negative oval area described above.

A number of parvalbumin-ir neurons embedded in a plexus of immunostainedvaricose axons and processes were found at rostral levels of the dorsal endopiriform nucleus (Fig. 3A). At intermediate and caudal levels of the dorsal endopiriform nucleus, the number of positive neurons and axons decreased dramatically: very few immunoreactive neurons were observed, and the

neuro-pil was almost negative at these levels (Fig. 3E). The parvalbumin immunoreactivity pattern in the ventral endopiriform nucleus was similar to that of the caudal part of the dorsal endopiriform nucleus.

Most parvalbumin stained neurons located in the rostral part of the dorsal endopiriform nucleus were medium in size and exhibited multipolar morphologies, without a specific orientation of their cell processes. In the ventral endopiriform nucleus, some multipolar immunoreactive neurons with the cell body and main dendrites oriented parallel to the pial surface were observed (Fig. 3F).

3.4. Colocalization study

We carried out a colocalization study for the two major calcium-binding proteins present in neurons of the mouse claustrum, i.e. calbindin and parvalbumin, using post-embedding immunocytochemistry on adja-cent semithin sections. The analysis was made in the dorsal claustrum, since it was the claustral region with more positive neurons for either calcium-binding pro-tein.

In addition, by comparing immunostained semithin sections with the adjacent Nissl-stained semithin section, we obtained quantitative data about the relative density of calbindin- and parvalbumin-immunoreactive popula-tions.

Calbindin- and parvalbumin-immunoreactive neu-rons constituted a minority of the claustral neuneu-rons, accounting for only 5.4% (n/122) and 7.9% (n/178) of the total number of counted cells (n/2260), respec-tively. Neurons expressing either of the two proteins represent 12.3% of the claustral cells. In addition, calbindin- and parvalbumin-immunoreactive cells are mostly segregated populations (Fig. 4): more than 87% of parvalbumin-ir cells do not express calbindin (156 out of 178) and, conversely, 82% of calbindin-ir cells do not express parvalbumin (100 out of 122), whereas only 22 out of 2260 cells coexpress both proteins.

Fig. 3. Parvalbumin-immunoreactivity. (A) Low power photomicrograph of a transverse section of the rostral telencephalon (see Fig. 17 ofPaxinos and Franklin, 2001). At this level, the dorsal endopiriform nucleus (DEn) presents its highest number of immunoreactive neurons. Scale bar: 1 mm. (B) Detail of the claustrum at an intermediate level. A central region with a moderately stained neuropil, just deep to theventral part of the agranular insular cortex can be easily recognized (arrows). Scale bar: 500mm. (C) Two multipolar medium-size neurons can be observed in the central region of the claustrum. Scale bar: 100mm. (D) Neuropil immunostaining consists of a dense meshwork of axons and terminals, although no pericellular baskets can be distinguished. Scale bar: 50mm. (E) Detail of the endopiriform region at a caudal level (see Fig. 38 ofPaxinos and Franklin, 2001). Both the dorsal andventral parts of the endopiriform nucleus are characterized by an immunonegative neuropil. Scale bar: 1 mm. (F) Multipolar stained neurons with their elongated cell bodies and dendrites oriented parallel to the pial surface are observed in theventral endopiriform nucleus (VEn). Scale bar: 100mm. aca, Anterior commissure, anterior; AID, agranular insular cortex, dorsal; AIP, agranular insular cortex, posterior; AIV, agranular insular cortex,ventral; BLA, basolateral amygdaloid nucleus, anterior; Cl, claustrum; CPu, caudate putamen; DEn, dorsal endopiriform nucleus; DI, dysgranular insular cortex; ec, external capsule; La, lateral amygdaloid nucleus; lo, lateral olfactory tract; Pir, piriform cortex; VEn,

ventral endopiriform nucleus.

(7)

(8)

4. Discussion

Although little is known about the exact functions of the calcium-binding proteins parvalbumin, calbindin and calretinin, they have provided useful markers of specific neuronal subpopulations in studies of the neuronal circuitry of the cerebral cortex and other brain regions, and they reveal chemoarchitectonic subdiv i-sions in different regions of the central nervous system. Results of the present study indicate the existence of distinct immunoreactivity patterns for each of the three calcium-binding proteins in both the claustrum and the endopiriform nucleus as it has been reported in other mammals, and also reveal details not described pre-viously. We will next discuss our results in comparison with other mammals and then analyze the calcium-binding protein expression patterns in the light of recently described subdivisions within the mouse claus-trum.

Calbindin and parvalbumin immunoreactivities have been reported in the claustrum and endopiriform nucleus in other mammals but inter-specific differences exist regarding the expression pattern of each protein. In the rat, parvalbumin and calbindin exhibit a largely complementary distribution pattern: parv albumin-im-munoreactivity is concentrated in the claustrum, whereas calbindin-immunoreactivity prevails in the en-dopiriform nucleus (Celio, 1990; Druga et al., 1993). From our results in mouse, we cannot demonstrate this complementary distribution pattern between calbindin and parvalbumin; instead, we showed that for either of these calcium-binding proteins, more immunostained neurons were observed in the claustrum than in the endopiriform nucleus, and more in rostral than in caudal levels.

Our colocalization analysis has demonstrated that, in spite of a similar distribution pattern, parvalbumin and calbindin are mostly expressed by separate claustral neurons in the mouse. In this context, is to be noted that in the rat cortex parvalbumin- and calbindin-ir cells represent two classes of GABAergic interneurons dis-playing morphological and neurochemical specificity

(Celio, 1986; Hendry et al., 1989; Kosaka and

Heiz-mann, 1989), and indirect evidence suggests that

cal-cium-binding protein immunoreactive neurons may also represent, at least in part, inhibitory local circuit neurons in the rat claustrum (Druga et al., 1993).

With regards to calretinin, it has been claimed that the rat claustrum is demarcated by the absence of calretinin staining (Paxinos et al., 1998). However, calretinin immunoreactive structures have been reported in the monkey claustrum (Reynhout and Baizer, 1999) and in the endopiriform nucleus of the guinea pig (Frassoni et

al., 1998). Our results in the mouse showed that

calretinin staining consisted of a few positive cell bodies embedded in a moderate neuropil both in the claustrum

and the endopiriform nucleus. Calretinin is present in the claustral regions of monkeys, guinea pigs, and mice. Thus, the reported absence of calretinin in the rat claustrum could be due to inter-specific differences or to a differential sensitivity of the immunocytochemical method.

Calcium-binding protein immunoreactive neurons constitute together a small subset (12.3%) of the claustral neurons in the mouse, as demonstrated by our study on semithin sections. This low incidence of calcium-binding protein immunoreactive neurons in the claustrum was to be expected if they represent different subpopulations of GABAergic interneurons, as sug-gested above, since GABAergic neurons in the rabbit claustrum accounted only for 12% of all neurons in the same claustral samples (Go´mez-Urquijo et al., 2000).

These data on the low incidence of inhibitory claustral neurons are also in agreement with the presence in the rat claustrum of numerous latexin-immunoreactive neurons (Arimatsu et al., 1992), most of which are glutamate positive (and GABA negative), suggesting that they are excitatory projection neurons (Arimatsu et

al., 1999).

4.1. Different zones within the dorsal claustrum

Recently, three novel subdivisions within the mouse claustrum based on the pattern of cadherin expression and cytoarchitecture have been distinguished (

Obst-Pernberg et al., 2001): a superior, an intermediate and

an inferior part. The superior and intermediate parts are located deep to the insular cortex whereas the inferior part is located deep to the dorsal part of the piriform cortex. Cytoarchitectonically the intermediate part is characterized by the presence of cell aggregates. The superior and intermediate parts contain densely packed cell bodies that strongly express R-cadherin (Rcad;

Obst-Pernberg et al., 2001) and cadherin-8 mRNA

(Korematsu and Redies, 1997), whereas the inferior

part shows weak expression of Rcad, and only contain scattered cell bodies expressing cadherin-8 mRNA. In addition, the superior part shows a very strongly immunoreactive neuropil for Rcad and moderately labeled for cadherin-N (Ncad); the intermediate part is strongly immunoreactive for Rcad and moderately to strongly labeled for Ncad; and the inferior part is weakly labeled for Rcad and moderately labeled for Ncad (Obst-Pernberg et al., 2001).

Our results on parvalbumin and calretinin immunor-eactivity match these novel subdivisions of the mouse claustrum. Thus, in parvalbumin-immunostained sec-tions a patch of moderately immunoreactive neuropil occupies a position just deep to the ventral part of the agranular insular cortex. This parvalbumin positive zone is bordered dorsal andventrally by claustral zones virtually devoid of immunostained neuropil. The dorsal M.A´ . Real et al. / Journal of Chemical Neuroanatomy 25 (2003) 151/160

(9)

negative zone lies deep to the dorsal part of the agranular insular cortex and to the dysgranular cortex, whereas the ventral zone negative for parvalbumin is continuous with the endopiriform nucleus. In calretinin immunostained sections, on the other hand, a comple-mentary pattern was observed: the intermediate region wasvirtually devoid of calretinin immunoreactive fibers, and dorsal and ventral calretinin-positive zones bor-dered it. It is tempting to relate these three zones of the dorsal claustrum showing a distinct calcium-binding protein expression pattern to the three subdivisions described by Obst-Pernberg and coworkers on the basis of cadherin expression patterns (Obst-Pernberg et al., 2001). These authors suggest that the selective adhesion of neural structures that express the same types of cadherin contribute to the formation of gray matter areas, neural circuits and functional connections in the postnatal forebrain of the mouse. In this sense, we propose that the cell aggregates of the intermediate zone of the dorsal claustrum with cell bodies that strongly express Rcad and cadherin-8 would be the selective target for parvalbumin-expressing fibers, and that they would be mostly avoided by calretinin-expressing axons. We cannot conclude whether these fibers have an intrinsic or extrinsic origin.

While these novel subdivisions of the mouse dorsal claustrum appear to be well documented from a cyto and neurochemical point ofview, it is difficult to relate them to the different functional zones that have been described in the mammalian claustrum on the basis of its connections. In mammals with a well developed

claus-trum the claustro-neocortical connections are topogra-phically organized, with an anterior part of the claustrum linked mainly with motor and prefrontal cortices, a central part linked with somatosensory cortex, a posterior claustrum related to visual cortex, and a ventral zone connected with auditory cortex

(Pearson et al., 1982; Macchi et al., 1983; Sherk, 1988;

Morys et al., 1996). In the rat claustrum, however, two

main cortico-related zones have been described, an anterodorsal sensorimotor and a posteroventral v i-suoauditory zones (Sadowski et al., 1997).

Our results indicate that the incidence of calcium-binding protein-expressing neurons is higher in rostral than caudal levels of the mouse dorsal claustrum. Since all hodological studies confirm at least two different functional zones in the claustrum, an anterior and a posterior region, it is likely that calcium-binding pro-tein-expressing neurons, putatively GABAergic cells, are more represented in claustral intrinsic circuits related with the limbic and motor cortices rather than with the visuoauditory regions. As suggested for somatostatin-, neuropeptide Y-, and vasoactive intestinal peptide-ir neurons in the rat claustrum (Kowianski et al., 2001), calcium-binding protein-expressing neurons do not appear to play a significant role in the claustro-cortical projection but are most probably involved in modula-tion and informamodula-tion transfer in the claustrum.

It is to be noted that the complementary expression of calretinin and parvalbumin in patches of neuropil in the mouse claustrum is best seen at intermediate/posterior levels in the rostrocaudal axis of the claustrum. It Fig. 4. Photomicrographs of two adjacent semithin sections immunostained for calbindin (A) and parvalbumin (B). Only one neuron in this field is immunoreactive for both calcium-binding proteins (arrow). Capillaries (c) were used as landmarks. Scale bar: 30mm.

(10)

cannot be determined, however, whether these claustral zones correspond with claustral regions with specific cortical connections. In this sense, it would be of interest to perform combined immunohistochemical-fiber tra-cing experiments in order to study the specific connec-tions of the chemoarchitectonic subdivisions of the dorsal claustrum.

Acknowledgements

We would like to thank Luis Olmos for his excellent technical assistance. This work was supported by Spanish DGI grant BFI2000-1359-C02-01, as well as by Spanish FIS grant 01-0057-01.

References

Arimatsu, Y., Miyamoto, M., Nihonmatsu, I., Hirata, K., Uratani, Y., Hatanaka, Y., Takiguchi-Hayashi, K., 1992. Early regional speci-fication for a molecular neuronal phenotype in the rat neocortex. Proc. Natl. Acad. Sci. USA 89, 8879/8883.

Arimatsu, Y., Ishida, M., Sato, M., Kojima, M., 1999. Corticocortical associative neurons expressing latexin: specific cortical connectivity formed invivo and invitro. Cereb. Cortex 9, 569/576.

Celio, M.R., 1986. Parvalbumin in most gamma-aminobutyric acid-containing neurons of the rat cerebral cortex. Science 231, 995/

997.

Celio, M.R., 1990. Calbindin D-28k and parvalbumin in the rat nervous system. Neuroscience 35, 375/475.

Da´vila, J.C., Guirado, S., Puelles, L., 2000. Expression of calcium-binding proteins in the diencephalon of the lizardPsammodromus algirus. J. Comp. Neurol. 427, 67/92.

Dinopoulos, A., Papadopoulos, G.C., Michaloudi, H., Parnavelas, J.G., Uylings, H.B., Karamanlidis, A.N., 1992. Claustrum in the hedgehog (Erinaceus europaeus) brain: cytoarchitecture and con-nections with cortical and subcortical structures. J. Comp. Neurol. 316, 187/205.

Druga, R., 1966. The claustrum of the cat (Felis domestica). Folia Morphol. (Praha) 14, 7/16.

Druga, R., 1971. Projection of prepyriform cortex into claustrum. Folia Morphol. (Praha) 19, 405/410.

Druga, R., Chen, S., Bentivoglio, M., 1993. Parvalbumin and calbindin in the rat claustrum: an immunocytochemical study combined with retrograde tracing frontoparietal cortex. J. Chem. Neuroanat. 6, 399/406.

Frassoni, C., Radici, C., Spreafico, R., de Curtis, M., 1998. Calcium-binding protein immunoreactivity in the piriform cortex of the guinea-pig: selective staining of subsets of non-GABAergic neurons by calretinin. Neuroscience 83, 229/237.

Go´mez-Urquijo, S.M., Gutie´rrez-Ibarluzea, I., Bueno-Lopez, J.L., Reblet, C., 2000. Percentage incidence of gamma-aminobutyric acid neurons in the claustrum of the rabbit and comparison with the cortex and putamen. Neurosci. Lett. 282, 177/180.

Hendry, S.H., Jones, E.G., Emson, P.C., Lawson, D.E., Heizmann, C.W., Streit, P., 1989. Two classes of cortical GABA neurons defined by differential calcium binding protein immunoreactivities. Exp. Brain Res. 76, 467/472.

Jones, E.G., Hendry, S.H.C., 1989. Differential calcium-binding protein immunoreactivity distinguishes classes of relay neurons in monkey thalamic nuclei. Eur. J. Neurosci. 1, 222/245.

Korematsu, K., Redies, C., 1997. Restricted expression of cadherin-8 in segmental and functional subdivisions of the embryonic mouse brain. Dev. Dynam. 208, 178/189.

Kosaka, T., Heizmann, C.W., 1989. Selective staining of a population of parvalbumin-containing GABAergic neurons in the rat cerebral cortex by lectins with specific affinity for terminalN -acetylgalac-tosamine. Brain Res. 483, 158/163.

Kowianski, P., Morys, J., Karwacki, Z., Dziewiatkowski, J., Narkie-wicz, O., 1998. The cortico-related zones of the rabbit claustrum-study of the claustrocortical connections based on the retrograde axonal transport of fluorescent tracers. Brain Res. 784, 199/209. Kowianski, P., Timmermans, J.P., Morys, J., 2001. Differentiation in

the immunocytochemical features of intrinsic and cortically projecting neurons in the rat claustrum*/combined immunocyto-chemical and axonal transport study. Brain Res. 905, 63/71. Krettek, J.E., Price, J.L., 1977. Projections from the amygdaloid

complex to the cerebral cortex and thalamus in the rat and cat. J. Comp. Neurol. 172, 687/722.

Li, Z.K., Takada, M., Hattori, T., 1986. Topographic organization and collateralization of claustrocortical projections in the rat. Brain Res. Bull. 17, 529/532.

Macchi, G., Bentivoglio, M., Minciacchi, D., Molinari, M., 1983. Claustroneocortical projections studied in the cat by means of multiple retrograde fluorescent tracing. J. Comp. Neurol. 215, 121/134.

Majak, K., Kowianski, P., Morys, J., Spodnik, J., Karwacki, Z., Wisniewski, H.M., 2000. The limbic zone of the rabbit and rat claustrum: a study of the claustrocingulate connections based on the retrograde axonal transport of fluorescent tracers. Anat. Embryol. 201, 15/25.

Markowitsch, H.J., Irle, E., Bang-Olsen, R., Flindt-Egebak, P., 1984. Claustral efferents to the cat’s limbic cortex studied with retrograde and anterograde tracing techniques. Neuroscience 12, 409/425. Morys, J., Berdel, B., Maciejewska, B., Sadowski, M., Sidorowicz, M.,

Kowianska, J., Narkiewicz, O., 1996. Division of the human claustrum according to its architectonics, morphometric para-meters and cortical connections. Folia Morphol. (Warsz) 55, 69/

82.

Obst-Pernberg, K., Medina, L., Redies, C., 2001. Expression of R -cadherin and N-cadherin by cell groups and fiber tracts in the developing mouse forebrain: relation to the formation of functional circuits. Neuroscience 106, 505/533.

Paxinos, G., Franklin, K.B.J., 2001. The Mouse Brain in Stereotaxic Coordinates. Academic Press, New York.

Paxinos, G., Kus, L., Ashwell, K.W.S., Watson, C., 1998. Chemoarch-itectonic Atlas of the Rat Forebrain. Academic Press, New York. Pearson, R.C., Brodal, P., Gatter, K.C., Powell, T.P., 1982. The organization of the connections between the cortex and the claustrum in the monkey. Brain Res. 234, 435/441.

Reynhout, K., Baizer, J.S., 1999. Immunoreactivity for calcium-binding proteins in the claustrum of the monkey. Anat. Embryol. 199, 75/83.

Sadowski, M., Morys, J., Jakubowska-Sadowska, K., Narkiewicz, O., 1997. Rat’s claustrum shows two main cortico-related zones. Brain Res. 756, 147/152.

Sherk, H., 1988. The claustrum and the cerebral cortex. In: Jones, E.G., Peters, A. (Eds.), Sensory-Motor Areas and Aspects of Cortical Connectivity, vol. 5. Plenum Press, New York, London, pp. 467/499.

Sloniewski, P., Usunoff, K.G., Pilgrim, C., 1986. Retrograde transport of fluorescent tracers reveals extensive ipsi- and contralateral claustrocortical connections in the rat. J. Comp. Neurol. 246, 467/477.

Witter, M.P., Room, P., Groenewegen, H.J., Lohman, A.H., 1988. Reciprocal connections of the insular and piriform claustrum with limbic cortex: an anatomical study in the cat. Neuroscience 24, 519/539.