differential expression of calbindin and calretinin in the human fetal amygdala

Differential Expression of Calbindin and Calretinin in theHuman Fetal AmygdalaMATTHIAS SETZER1,2 AND NORBERT ULFIG1*1 Neuroembryonic Research Laboratory, Department of Anatomy, University of Rostock, D-18055 Rostock, Germany2 Department of Neurosurgery, University of Frankfurt, D-60528 Frankfurt/M., Germany

KEY WORDS development; neuronal types; calcium-binding proteins; interneurons; pyramidalcells

ABSTRACT The distribution patterns of the calcium-binding proteins calbindin and calretinin,both expressed early during development within the various amygdaloid nuclei and areas, havebeen investigated. Anti-calbindin as well as anti-calretinin mark immature, partly migratingneurons in the 5th gestational month; the number of calretinin-immunoreactive neurons isdistinctly higher. In the 8th month, calbindin and calretinin are found in a small proportion ofpresumed pyramidal cells and in various types of non-pyramidal neurons. Small and large bipolarand small and large multipolar neurons are shown to express calbindin and calretinin. Double-labellings show that calbindin and calretinin are largely contained in different subsets of theseneuronal types, which are considered to represent interneurons. These nerve cell classes arewidespread within the amygdala with mainly moderate to high packing densities. Diffuseimmunoreactive structures, which are found in different intensities in the various amygdaloidnuclei, display distinct redistribution during fetal development. The results show that during earlyfetal development calbindin and particularly calretinin may be involved in the regulation ofneuronal migration. In later development, definite subsets of interneurons, which are likely to befunctionally different, are marked by anti-calbindin and -calretinin. Different diffuse immunolabel-ling at various developmental stages probably indicates the sequential arrival of afferent input frombrain areas containing calbindin- or calretinin-immunoreactive nerve cells. With the exception thatcalretinin may be transiently expressed in pyramidal neurons, the distribution of calbindin- andcalretinin-immunoreactive structures to a large degree corresponds to that in the adult. Thus, littlereorganisation is to be expected during proceeding development. Microsc. Res. Tech. 46:1–17, 1999.r 1999 Wiley-Liss, Inc.

INTRODUCTIONThe amygdaloid complex is a prominent structure

within the temporal lobe. It is a key component of thelimbic system and plays an important role in neurobio-logical processes like emotion, motivation, learning,and memory (Aggleton, 1993) Furthermore, it could beshown that it is involved in neurological diseases likedementia, amnesia, epilepsy, schizophrenia and Parkin-son’s disease (Aggleton, 1993; Braak et al., 1994; Hi-rano and Zimmermann, 1962; Hooper and Vogel, 1976;Jamada and Mehraein, 1968; Stevens, 1973; Torrey andPeterson, 1974). Some of these pathological processesare thought to be developmental in origin. Despite theincreasing interest in this brain structure, relativelylittle is known about the normal fetal development ofthe amygdala in the human brain. The detailed knowl-edge of the normal structure and its ontogenesis,however, is an important prerequisite to describe andinterpret even subtle pathological changes of the amyg-dala.

Antibodies against the calcium-binding proteins cal-bindin (CB) and calretinin (CR) have been used exten-sively as neuronal markers throughout the centralnervous system (Baimbridge et al., 1992; Braun, 1990;Jacobowitz and Winsky, 1991). CB and CR are cytosolicproteins that are expressed in distinct subpopulations

of nerve cells. Both belong to a larger group of calcium-binding proteins characterized by the EF-hand struc-ture, which is involved in buffering intracellular cal-cium ions (Baimbridge and Miller, 1984; Celio, 1986,1990). Parvalbumin, which is another main calcium-binding protein, occurs distinctly later during develop-ment. (Hendrickson et al., 1991; Nitsch et al., 1990;Solbach and Celio, 1991). It has, therefore, been associ-ated with mature neuronal activity. CB and CR, bothexpressed during early development, may control onto-genetic events (Ellis et al., 1991; Enderlin et al., 1987).Therefore, the distribution patterns of CB and CR inthe various amygdaloid nuclei and areas at different

*Correspondence to: Prof. Norbert Ulfig, Neuroembryonic Research Labora-tory, Department of Anatomy, University of Rostock, Gertrudenstr. 9, D-18055Rostock. E-mail: [email protected]

Received 14 October 1999; accepted in revised form 3 January 1999*Abbreviations used: AAA 5 anterior amygdaloid area; AB 5 accessory basal

nucleus of the amygdala; AHA 5 amygdalohippocampal area; B 5 basal nucleusof the amygdala; CA 5 anterior commissure; CB 5 calbindin; CE 5 centralnucleus of the amygdala; Cl 5 claustrum; COa 5 anterior cortical nucleus of theamygdala; COp 5 posterior cortical nucleus of the amygdala; CR 5 calretinin;GE 5 ganglionic eminence; GP 5 globus pallidus; H 5 hippocampus; IC 5intercalated nuclei of the amygdala; IR 5 immunoreactivity, immunoreactive;L 5 lateral nucleus of the amygdala; M 5 medial nucleus of the amygdala;NbM 5 basal nucleus of Meynert; P 5 putamen; PAC 5 periamygdaloid cortex;PIR 5 piriform cortex; PL 5 paralaminar nucleus of the amygdala; TO 5 optictract; V 5 ventricle.

MICROSCOPY RESEARCH AND TECHNIQUE 46:1–17 (1999)

r 1999 WILEY-LISS, INC.

stages of fetal development have been investigated inthis study.

Immunohistochemical staining with antisera di-rected against CB and CR yields a Golgi-like appear-ance of nerve cells allowing neuroanatomists to carryout neuronal classification (Baimbridge et al., 1992;Rogers et al., 1990). CB and CR share about 60% oftheir amino-acid sequence (Parmentier et al., 1987;Pasteels et al., 1987, 1990; Rogers, 1989). Genes codingfor CB and CR are located on different chromosomes inhumans. Both genes are situated in close vicinity toregions of carbonic anhydrase isoenzyme genes. There-fore, a common duplication of CB/CR and carbonic anhy-drase genes appears likely (Parmentier et al., 1991).

Although the exact functions of CB and CR areunknown, it is believed that these proteins may beinvolved in triggering calcium-dependent processes.Important developmental events, such as cell move-ment or process outgrowth, are closely related to theintracellular calcium-concentrations (Spitzer, 1994).Thus, it appears plausible that calcium-binding proteinsmediate such calcium-dependent ontogenetic processes.

MATERIALS AND METHODSThis study was carried out on 18 human fetal brains

that ranged in age from the 16th to the 38th week ofgestation. The brains were obtained from legal electiveand spontaneous abortions according to German lawswith the prior consent of an ethical committee; they didnot exhibit any hints of neuropathological alterations.The brains were obtained 1–10 hours after death. Fromthe whole brain, the brain stem was detached at thelevel of the mesencephalon and the cerebral hemi-spheres were divided by a medio-sagittal cut. After-wards, the brains were fixed by immersion in 3.7%paraformaldehyde (pH 7.4, for around 2 days, on arotator), and cut into frontal blocks and cryoprotectedin 0.05 M TRIS-buffered saline (TBS, pH 7.4) contain-ing 30% sucrose on a rotator for about 2 days. Then theblocks were frozen and cut into 120-µm-thick frontalsections on a cryostat. Free-floating sections were rinsedin TBS after each step of the incubation procedure (10minutes) (if not stated otherwise, see double-labelling).

Non-specific staining was reduced by two preincuba-tion steps: (1) 10% methanol and 7% hydrogen peroxidein TBS; (2) 1.5% lysine, 0.25% triton X-100 and 10%bovine serum albumin (BSA) (all from Sigma, St. Louis,MO) in TBS (60 minutes). Then the sections wereincubated in the primary antibody at 4°C for 2 days(anti-CB, rabbit, Swant, Bellinzona, Switzerland, di-luted 1:5,000, and anti-CR, rabbit, Swant, diluted1:5,000). Afterwards, the sections were incubated withthe biotinylated secondary antibody (secondary anti-rabbit IgG, Vector Laboratories, Burlingame, CA; di-luted 1:200 in TBS with 2% BSA for 2 hours at roomtemperature and thereafter transferred to the avidin-biotin-peroxidase complex (ABC, Vector Laboratories,diluted 1:25 in TBS) for another 2 hours. To visualizethe immunoreaction product, the chromogene 3,3’-diaminobenzidine-tetra-hydrochloride (DAB, Sigma,0.07% in TBS with 0.003% hydrogen peroxide) wasused. Finally the sections were mounted on gelatin-coated slides, dehydrated in a graded series of alcohol,cleared in xylene and coverslipped with DePeX (Boeh-ringer Ingelheim Bioproducts, Heidelberg, Germany).

Further details concerning immunohistochemistry arefound in Ulfig et al. (1998a).

For the double-labelling procedure, preincubationsteps were identical to those described above. Sectionswere incubated with two antibodies of different species(CB[mouse] 1:5,000 1 CR[rabbit] (Swant) 1:5,000 inTBS 1 2% BSA) for 2 days. Then peroxidase-(1:200)and alkaline phosphatase- (1:100) labelled secondaryantibodies (Vector Laboratories) were applied for 3hours. First, the peroxidase was visualized using 0,007%DAB/0,003% H2O2. After rinsing three times for 10minutes in TBS containing 0,025% levamisole (Sigma),the alkaline phosphatase was visualized with the aid of5-bromo-4-chloro-indolyl-phophatase/nitro-blue-tetra-zolium (BCIP/NBT)(Vector Laboratories) 1:50 in TBS.Finally sections were rinsed, mounted on gelatin-coatedslides, and coverslipped with glycerol gelatin (Sigma).

Every seventh cryostat section was not immuno-stained but stained with cresyl-violet (Nissl-stain).These sections were analyzed at a magnification of 1:12using a stereomicroscope. Thus, the borders of theamygdaloid nuclei and areas were determined. Cameralucida drawings of the borders were made, then immu-noreactive structures were plotted on these drawingsfrom the adjacent immunopreparations.

Photographic documentation was done at low andhigh magnification. Moreover, camera lucida drawingsof the various neuronal types were made at high magnifica-tion. (Magnifications are specified in each figure.)

RESULTSThe nomenclature proposed by Amaral et al. (1992)

for the non-human primate amygdala has been used inthis study. The amygdaloid complex can roughly bedivided into basolateral (deep) and corticomedial (super-ficial) nuclei (Fig. 1). The basolateral group consists ofthe lateral, basal, accessory basal, and paralaminarnuclei. The corticomedial group encompasses the ante-rior cortical, posterior cortical and central nuclei, andthe periamygdaloid cortex. The remaining nuclei andareas that cannot be easily assigned to the amygdaloidportions are the anterior amygdaloid and the amygdalo-hippocampal areas and the intercalated nuclei.

In the following descriptions, cellular immunolabel-ling (i.e., immunoreactive [ir] somata with their den-dritic trees) and punctate or diffuse immunolabelling(i.e., cross sections of axons and dendrites withoutcontinuity to perikarya) are distinguished. Fibrousimmunoreactivity (IR) is not seen within the amygda-loid complex at the developmental stages investigatedhere.

Generally, ir puncta are evenly distributed within anucleus (see Figs. 3, 8). No distinctive differences in thedensity of ir puncta are detected along the antero-posterior extension of a nucleus. Moreover, no signifi-cant differences could be observed between the varioussubdivisions of a nucleus.

CB-ir Structures in the Fetal AmygdalaCB-ir Neuronal types.5th and 6th gestational month. In the corticomedial

nuclei medium-sized roundish perikarya giving rise totwo till four dendrites are encountered (see Fig. 4a).The primary dendrites frequently branch in some dis-tance to the soma. The basolateral nuclei harbor smallovoid somata from which one or two short unbranched

2 SETZER AND ULFIG

processes emerge. A medium number of bipolar cellsappear oriented perpendicularly to the ganglionic emi-nence. The precursor cells of the amygdala are gener-ated in the ganglionic eminence, which represents acircumscribed enlargement of the ventricular and sub-ventricular zone. It is particularly prominent betweenthe 5th and 7th gestational month, and it is devoid ofCB- or CR-ir structures (Figs. 2,3; see also Fig. 8).

8th and 9th gestational month. At this develop-mental stage, different neuronal types can be distin-guished in the CB-immunopreparations:

1. Presumed pyramidal nerve cells (Figs. 4c,5b). Asmall number of nerve cells revealing the typicalcharacteristics of pyramidal cells are found in the

basal and the lateral nucleus. From the base of thepyramidal-shaped soma, dendrites revealing me-dium diameters emerge (Figs. 4c,5b). At the oppositepole, the soma merges into the stout main dendrite,which gives off several side branches. The basaldendrites do not always display the same calibers.Occasionally, one of the dendrites show a particularstoutness and generates several branches in proxim-ity of the cell body. From the base of the pyramids,the axon emerges and often bends close to the soma.The axon can regularly be traced up to around200µm. In general, the presumed pyramidal cells donot show a preferential orientation and their den-dritic trees do not extend beyond the nuclear bor-ders.

Fig. 1. Frontal section (100-µm-thick, stained with the Nissl dye cresyl violet) of the amygdaloid nucleiin the 5th gestational month.

3CALBINDIN AND CALRETININ IN THE HUMAN FETAL AMYGDALA

Fig. 2. Distribution of CB-ir structures in the amygdala of the 5th gestational month, frontal section of100-µm thickness.

4 SETZER AND ULFIG

Fig. 3. Distribution of CB-ir structures in the 5th/6th (I) and 8th/9th (II) gestational month. Theamygdaloid complex is schematically drawn in frontal sections from anterior (a–d) to posterior (e–h). Thenumber of larger dots respresents the number of immunoreactive neurons (a–d). The intensity of smalldots corresponds (e–h) to the intensity of diffuse neuropil immunostaining.


The bulk of CB-ir neurons represents three types ofnon-pyramidal neurons:

2. Bipolar nerve cells (Figs. 4c,6). These small ormedium-sized spindle-shaped neurons are found ineven number in all amygdaloid nuclei. The dendritesemanating from opposite sites of the soma areobserved to branch at various distances from thesoma. Some of the bipolar nerve cells exhibit aconspicuously small cell body (Fig. 6).

3. Large multipolar nerve cells (Figs. 4c,5a). The polygo-nal cell body generates slender as well as stout

dendrites that spread out in all directions. Onaverage, five or six dendrites emerge; ramificationsare observed near the soma. The dendrites and theirbranches extend straightly. Thus, these nerve cellsbear resemblance to cortical stellate neurons. Occa-sionally the axons, often revealing a curved course,are immunostained.

4. Small multipolar nerve cells (Fig. 4c,5a). Theseneurons reveal a small roundish soma from whichseveral dendrites spreading in all directions arise.Ramifications of dendrites are only infrequently

Fig. 4. Camera lucida drawings of CB- (a,c) and CR- (b,d) ir neurons of the 5th/6th (a,b) and 8th/9th(c,d) gestational month. Cells next to large dots: presumed pyramidal neurons (c) and triangularpyramidal-like neurons (d); cells next to large asterisks: large multipolar neurons; cells next to smallasterisks: small multipolar neurons; remaining cells: bipolar neurons;

6 SETZER AND ULFIG

Fig. 5. Microphotograph of CB-ir neurons of the 8th/9th gestational month. a: multipolar neurons,large and a small one (lateral nucleus) (marked by a star); b: presumed pyramidal neurons (basalnucleus).

observed. The diameters of dendritic stems varyonly to a limited extent. Only sometimes the stem ofthe axon is immunolabelled.

Distribution of CB-ir Nerve Cells5th/6th gestational month (Fig. 3 Ia–d). The highest

number of CB-ir nerve cells is found in the anterior andposterior cortical nucleus, the medial nucleus, and theperiamygdaloid cortex. The central nucleus containsonly very few ir nerve cells. Moderate numbers of irneurons are seen in the remaining nuclei.

8th/9th gestational month (Fig. 3 IIa–d). Moderateto high numbers of CB-ir neurons are observed in thebasal, lateral, cortical, and medial nucleus. The centraland paralaminar nucleus and the periamygdaloid cor-tex contain only low numbers of CB-ir nerve cells.

Diffuse CB-Staining5th/6th gestational month (Figs. 2,3 Ie–h). The

central nucleus exhibits the highest density of diffuseimmunostaining in the amygdaloid complex. Withinthe basolateral nuclei, a moderate intensity of immuno-labelling is observed. Distinct differences between thebasolateral nuclei do not become obvious. The interca-late nucleus habours an average amount of immuno-stained structures. In the medial and cortical nuclei, aswell as in the periamygdaloid cortex and the amygdalo-hippocampal area, only a weak punctate IR is visible.Densely packed fiber bundles are found next to thelateral margin of the amygdala. They run between thecell islands of the pars striatalis (Ulfig et al., 1998b) ofthe lateral nucleus.

8th/9th gestational month (Fig. 3 IIe–h). Whencomparing the immunopreparations of the 5th and 8thgestational month some distinct differences are evi-

dent. The medial and cortical nucleus and the periamy-gdaloid cortex contain the highest density of ir-puncta.The density in the central nucleus ranges from moder-ate to high. Medium levels of labelling intensity areobserved in the lateral, paralaminar, and posteriorcortical nucleus. The basal and accessory basal nucleusshow sparse immunolabelling.

CR-ir Structures in the Human FetalAmygdala

CR-ir Nerve Cells

5th/6th gestational month (Fig. 4b; see also Fig. 9).In the 5th gestational month, the characteristics ofCR-ir neurons largely correspond to that described forCB-ir cells at this developmental stage. A distinctdifference between the CB- and CR-immunoprepara-tions at this developmental age concerns the packingdensity of ir-neurons. It is strikingly higher in CR-immunopreparations (see Fig. 8 Ia–d).

8th/9th gestational month (Fig. 4d; see also Figs.12,13). In the 8th gestational month, large multipolar,small multipolar, and bipolar nerve cell types areencountered. Their characteristics closely correspondto that described for CB-ir neurons. Small multipolarnerve cells express CR more frequently than CB. More-over, medium-sized and large triangular neurons areCR-ir. These nerve cells show striking similarity topyramidal cells. The soma shape and the arrangementof dendritic stems correspond to that of pyramidal cells.In contrast to the presumed CB-ir pyramidal cells,dendritic ramifications are distinctly less frequent. Theaxon is observed to emerge from the stem of a basaldendrite. Although these nerve cells cannot be defi-

Fig. 6. Microphotograph of CB-ir bipolar nerve cells of the 8th/9th gestational month (accessory basalnucleus, medial nucleus). Note the different sizes of the somata.

8 SETZER AND ULFIG

nitely classified as pyramidal nerve cells, they arereferred to as pyramids in the Discussion.

Distribution of CR-ir Nerve Cells5th/6th gestational month (see Fig. 8 Ia–d). The

basolateral nuclei contain high numbers of mostlybipolar CR-ir neurons, in particular the basal andaccessory basal nucleus. A large percentage of thesenerve cells appear oriented with their long axis runningperpendicularly to the ganglionic eminence. The cortico-medial nuclei display only a few ir nerve cells.

8th/9th gestational month (see Fig. 8 IIa–d). Thecortical, medial, and lateral nucleus (see Fig. 14) ex-hibit a large number of ir nerve cells. In the remainingnuclei, moderate numbers of ir neurons are seen. Onthe whole, the number of CR-ir nerve cells is distinctlyhigher than that of CB-ir neurons.

Diffuse CR-Immunostaining5th and 6th gestational month (Figs. 7,8 Ie–h) The

intensity of diffuse immunostaining distinctly variesbetween the amygdaloid areas. The central (Fig. 10)and medial nucleus as well as the anterior and poste-rior cortical nucleus and the amygdalohippocampalarea stand out conspicuously due to their high densityof ir puncta. Moderate densities of puncta are foundwithin the lateral nucleus and the anterior amygdaloidarea. The intensity of punctate immunolabelling in theintercalate nuclei ranges between low and medium.The basal, accessory basal nucleus, and the periamygda-loid cortex reveal only low diffuse IR.

8th and 9th gestational month (Fig. 8 IIe–h). Somesignificant differences in the distribution pattern ofdiffuse IR between the 5th and 8th gestational monthbecome obvious. Intense punctate IR is observed in theanterior cortical nucleus, the medial nucleus, and theamygdalohippocampal area. The lateral, central, para-laminar, and posterior cortical nucleus exhibit moder-ate diffuse IR. Low IR is seen in the basal and accessorybasal nucleus and the periamygdaloid cortex.

Coexpression of CB and CR (see Fig. 15)The results concerning coexpression stem form two

brains of the 8th gestational month. On the colourplate, double-labelled neurons appear black (CB- andCR-ir), single-labelled either brown (CR-ir) or blue(CB-ir).

The great majority of neurons contain either CB orCR. Only infrequently double-labelled small neurons,which mainly reveal a bipolar shape, are detected.Double-labelled large multipolar neurons are ex-tremely rare. Nerve cells expressing both CB and CRare more often found in the basolateral nuclei than thecorticomedial nuclei.

DISCUSSIONCB and CR show a widespread distribution within

the human fetal amygdala. They occur in distinctsubpopulations of neurons that may, therefore, be distin-guished by specific calcium-dependent processes. Thenuclear specific diffuse immunolabelling is indicative ofdifferential afferent input of the amygdala.

Nerve Cell Types Immunostained With Anti-CBand Anti-CR in the Human Fetal Amygdala. Oneessential criterion for distinguishing between principalneuronal classes in the amygdaloid complex is thepresence, sparseness, or absence of dendritic spines(Braak and Braak, 1983; McDonald, 1992; Sorvari etal., 1996a,b). As spines mainly develop postnatally(Kostovic et al., 1992), a neuronal classification duringfetal development can only be based on the size andshape of somata and characteristics of the dendritictrees, such as number, caliber and extension of den-drites, their sites of origin from the somata, and loca-tion of dendritic branching sites.

In general, the CB- and CR-ir neurons show a highvariability in the shape and size of their somata. Thenon-pyramidal nerve cells strongly resemble the class

Fig. 7. Distribution of CR-ir structures in the amygdala of the 6th gestational month, frontal section of100-µm thickness.


of aspinous and sparsely spined neurons described byBraak and Braak (1983) as class II and class III nervecells. Such a correlation between CB- and CR-ir neu-

rons and class II and III nerve cells has also beensuggested by Pitkanen and Amaral (1993) and Sorvariet al. (1996a,b). This group of nerve cells resemble

Fig. 8. Distribution of CR-ir structures in the 5th/6th (I) and 8th/9th (II) gestational month. Theamygdaloid complex is schematically drawn in frontal sections from anterior (a–d) to posterior (e–h). Thenumber of larger dots respresents the number of immunoreactive neurons (a–d). The intensity of smalldots corresponds (e–h) to the intensity of diffuse neuropil immunostaining.

10 SETZER AND ULFIG

cortical stellate neurons and it has been assumed thatCB- and CR-ir nerve cells (Sorvari et al., 1996a,b), liketheir putative cortical counterpart, represent local cir-cuit neurons. Within the hippocampal formation it hasbeen definitely shown that CB and CR occur in largelycomplementary subpopulations of stellate GABAergicinterneurons that express the inhibitory transmitterGABA (Miettinen et al., 1992).

The colocalization of CB or CR and GABA is difficultto investigate in the human fetal brain because appro-priate antibodies reliably marking GABAergic cells inhuman fetal brain tissue are so far not available(personal observations).

The results of this study show differential distribu-tion patterns of CB- and CR-ir neurons as well as a rarecoexpression of CB and CR in double-labellings. Thus,evidence is provided that CB and CR are mainly foundin largely complementary neuronal subpopulations.The distinct subsets of interneurons most probably are

functionally different because they are likely to havedifferent input-output characteristics. Thus, they mayplay specialized roles in the amygdaloid circuits (seePhysiological Role of CB and CR in Nerve Cells of theMature Amygdala).

As concerns the morphology of CB- and CR-ir non-pyramidal neurons, it is obvious that both antibodiesmark the same neuronal types; small multipolar, largemultipolar neurons, small and large bipolar (spindle-shaped) neurons are present in the immunosections.These nerve cell types demonstrated in this studyclosely correspond to those described in the humanadult amygdala by Sorvari et al. (1996a,b). In theirstudies, these authors clearly correlated CB- and CR-irnerve cells with the neuronal types described by Braakand Braak (1983).

On the whole, it becomes obvious that non-pyramidalneurons are morphologically and neurochemically het-erogeneous. This well-established characteristic of theadult amygdala can already be detected in the 8thgestational month. This observation suggests that themature neuronal circuitries are to a large extent devel-oped in the last trimenon of pregnancy. With regard topreterm delivery, this finding appears of significantimportance for neuropediatricians.Afrequent neurologi-cal problem is the occurrence of intracranial hemor-rhage (Volpe, 1994). The most common site localizationof such hemorrhage is the ganglionic eminence, whichprovides precursor cells for, among others, the amyg-dala. As small neurons, such as local circuit neurons,are known to be produced later than large neurons(Jacobson, 1991), the generation of interneurons couldbe affected by the hemorrhage. It can be deduced fromthe findings of this study that in the 8th gestationalmonth, the production of interneurons is mostly notimpaired because no significant numbers of immatureCB- and CR-ir nerve cells are encountered. At earlierdevelopmental stages, however, quite a number ofsmall nerve cells displaying the shape of migratingneurons are visible in the immunopreparations. Theincidence of hemorrhage is directly correlated with thedegree of prematurity (Volpe, 1994). Therefore, onemajor conclusion of this study is that anti-CB andanti-CR are appropriate tools to investigate the distri-bution of interneurons in pathological specimens after

Fig. 9. Microphotograph of CR-ir nerve cells of the 5th gestational month in the basolateral nuclei.

Fig. 10. Dense diffuse CR-ir neuropil in the central nucleus of the6th gestational month.


the occurrence of hemorrhage in the ganglionic emi-nence of very young premature infants.

In addition to the interneurons, it could be demon-strated that pyramidal cells are presumably also CB-and CR- ir. These presumed pyramidal cells are sup-posed to be the projection neurons of the amygdala(Braak and Braak, 1983; Pitkanen and Amaral, 1993).They constitute the predominant nerve cell type in thebasolateral amygdala (McDonald, 1992). In CB- andCR-immunopreparations, however, only a small to mod-erate number of pyramidal cells is detected. So only asmall percentage of presumed pyramidal cells presentin a section is CB- or CR-ir.

In the rat and monkey amygdala, pyramidal neuronshave been described to be CR-ir (McDonald, 1994). Incontrast, in the human adult amygdala pyramidalneurons were observed to be CB-ir, but not CR-ir(Sorvari et al., 1996a,b). In our material, large CR-irneurons were seen that to a large extent resemblepyramidal neurons. However, one main feature of pyra-midal cells, i.e., spiny dendrites, can unfortunately notbe judged in the fetal specimens. Therefore, it stillremains to be determined definitely if these cells in thefetal amygdala are pyramidal cells or a subpopulationof large non-pyramidal nerve cells. If they truly repre-sent pyramidal neurons, the difference between thefetal and adult amygdala cannot easily be related todifferent antibodies because Sorvari et al. (1996a) usedthe same antibodies as applied in this study. Moreover,postmortem delay is unlikely to account for the differ-ence, as in both studies brains with short as well as

quite long delays were included. Thus, it appears mostlikely that pyramidal neurons may transiently expressCR. Similarly, CR has been demonstrated to be transito-rily expressed in the rat cerebral cortex (Hendrickson etal., 1991; Schierle et al., 1997).

Distinct differences in the packing densities of nervecells can be observed when comparing the amygdala ofthe 5th and 8th gestational month. For instance, thebasolateral nuclei of the 5th month harbor a largenumber of CR-ir nerve cells whereas the corticomedialnuclei contain only very few ir nerve cells. In the 8thgestational month, corticomedial nuclei also exhibitmoderate to high number of CR-ir neurons. This obser-vation may be probably related to the process of migra-tion. In the 5th month, migrating neurons that pass thebasolateral nuclei have not yet reached the corticome-dial nuclei.

Diffuse Immunostaining in the Various Nucleiof the Human Fetal Amygdala. Two general fea-tures concerning diffuse CB- and CR-immunolabellingin the human fetal amygdala become obvious:

1. Dense diffuse immunolabelling is mainly encoun-tered in nuclei that habor only few immunostainednerve cell bodies.

2. Distinct changes in the spatio-temporal distributionof diffuse CB-IR can be observed.

The first finding suggests that CB and CR may becontained in projection neurons of other brain areasthat provide afferent input to the amygdala. The second

Fig. 11. Microphotograph of CR-ir immunopreparation of the 9th gestational month. Note the sharpborder (marked by arrowheads) between the amygdaloid basal nucleus and the basal nucleus of Meynert.

12 SETZER AND ULFIG

Fig. 12. Microphotograph of CR-ir neurons of the 8th/9th gestational month. Cells marked by stars:triangular (pyramidal) neurons (basal nucleus); cells not marked: bipolar neurons (lateral, basal,accessory basal and medial nucleus).


finding indicates the mature immunolabelling patternis reached only after a prolonged period of development,which is characterized by increases as well as decreasesof diffuse immunolabelling in the various amygdaloidnuclei.

The significant decrease of immunoreactive fibersmay, of course, indicate that neurons projecting to theamygdala express CB or CR only transiently. Decreasesin punctate IR may, more probably, be due to fiberdiluting effects of ontogenetic events, such as differen-tiation of dendritic trees or occurrence and maturationof glial cells or arrival of other afferents.

Various subcortical nuclei that contain CB- or CR-irnerve cells project to the amygdaloid nuclei. For in-stance, the periaqueductal grey of the midbrain thatcontains a significant number of CB-ir nerve cells(DeLeon et al., 1994) is reciprocally connected with thecentral nucleus (Rizvi et al., 1991). The latter is charac-terized by a quite dense diffuse CB-immunolabelling inthe 8th gestational month. The magnocellular complexof the basal forebrain revealing a high number of CB-irnerve cells (Celio, 1990) projects in particular to thebasal nucleus (Amaral et al., 1992). In this nucleus,CB-ir puncta are especially pronounced in the 5th/6thgestational month. This finding is in accordance withthe observation that the basal forebrain complex andits efferents develop early (Kostovic, 1986). The basalforebrain complex belongs to a system of non-thalamicnuclei that project to the cerebral cortex in a non-specific manner (Ulfig, 1989). Another member of thissystem is the hypothalamic tuberomamillary nucleus,

which also exerts influence on various nuclei of theamygdala (Panula et al., 1989). Large parts of thetuberomamillary nucleus reveal a high number of CR-irneurons (‘‘supramammillary area,’’ Borhegyi and Ler-anth, 1997; Kiss et al., 1997; personal observation). Thedifference in diffuse immunolabelling between the 5thand 8th gestational month could also be related to thesequential arrival of afferents from the basal nucleusand the tuberomamillary nucleus, which are known todifferentiate at different stages of development (Pauland Ulfig, 1998).

When comparing the results concerning diffuse CB-and CR- immunostaining in the adult amygdala asdescribed by Sorvari et al., (1996a,b) and the results inthe 8th gestational month presented here, only minordifferences become obvious. Such minor discrepanciesmay be attributed to technical details, such as sectionthickness, which is considerably higher in this study.

On the whole, it is apparent that the amygdala of the8th gestational month has reached a high degree ofmaturity judging from the diffuse as well as cellularimmunostaining. Thus, only subtle reorganization, offor instance afferent projection, are to be expectedduring the proceeding development.

Functional Role of CB and CR in MigratingNerve Cells of the Amygdala. CR as well as CB-irneurons can be observed during early fetal develop-ment. The immature neurons immunolabelled in theamygdala of the 5th gestational month reveal distinctmorphological characteristics that are typical of migrat-ing nerve cells. Moreover, the orientation of these

Fig. 13. Microphotopgraph of small and large multipolar CR-ir neurons, 8th/9th gestational month(lateral and basal nucleus).

14 SETZER AND ULFIG

immature nerve cells with regard to the ganglionicemminence underlines the assumption that these aremigrating cells.

Calcium-binding proteins are most likely to play arole in neuronal migration because calcium ions areinvolved in regulating cell movement (Frassoni et al.,1998). As has been shown in vitro, amplitude andfrequency of intracellular calcium fluctuations are asso-ciated with cell movement (Komuro and Rakic, 1996).The elevation of intracellular calcium in migratingneurons is an essential parameter determining theirrate of migration.

Developmental Roles of CB and CR in ImmatureNeurons After Settling. After having reached theirfinal destinations, immature neurons differentiate andsynaptic connections are established. Experimentaldata demonstrate that calcium fluctuations play apivotal role in the maturation of ion channels and theappearance of neurotransmitters (Spitzer, 1994). Fur-thermore, levels of intracellular calcium are essentialfor neurite outgrowth, growth cone motility, and expres-sion of neurotransmitter receptors (Gomez et al., 1995;Kater et al., 1988; Spitzer, 1994). Calcium levels have tobe kept within a definite range within the cytoplasm.Drops or rises beyond this range lead to severe alter-ations of developmental events. Calcium binding pro-

teins have been suggested to regulate the levels ofcalcium via their ability to buffer intracellular calcium.

So far, it is not clear why different calcium bindingproteins are found in different subsets of neuronaltypes. It is conceivable that a neuronal subpopulationcharacterized by the expression of a certain calciumbinding protein can be distinguished from others by adistinct functional feature. It is tempting to assumethat the expression of CB or CR in different interneuro-nal classes may be related to differential connections,which these two classes establish during development(see the following section).

Physiological Role of CB and CR in Nerve Cellsof the Mature Amygdala. Data obtained from stud-ies on animal and adult human amygdala and theresults presented here (see above) indicate that CB andCR are expressed in different subsets of interneurons.It can be assumed that these amygdaloid interneuronsare also inhibitory in nature although for confirmationdouble-labelling studies are required (see above). Usingdetailed data of neuronal circuitry in hippocampalformation (Jiang and Swann, 1997), the following wayof controlling amygdaloid pyramidal cells has beenpostulated (Sorvari et al., 1998): CR-ir interneuronsmake inhibitory synaptic contacts on CB-ir nerve cells.These CB-ir interneurons, in turn, form inhibitory

Fig. 14. CR-ir immunoreactive nerve cells in the lateral nucleus of the 9th gestational month.


synapses on pyramidal cells. Thus, CR-ir interneuronscould modulate pyramidal cell activity via a disinhibi-tory interneuronal pathway. The intermediate neuronof this circuitry, i.e., the CB-ir neurons, may, moreover,be targeted by CB-ir afferents from the basal forebraincomplex (Freund and Gulyas, 1991).

ACKNOWLEDGMENTSDr. Jurgen Bohl, University of Mainz, is gratefully

acknowledged for providing the autopsy material. Theauthors also thank Gunther Ritschel for the skillfulpreparation of the figures.

REFERENCESAggleton JP. 1993. The contribution of the amygdala to normal and

abnormal emotional states. Trends Neurosci 16:328–333.

Amaral DG, Price JL, Pitkanen A Carmichael ST. 1992. Anatomicalorganization of the primate amygdaloid complex. In: Aggleton JP,editor. The Amygdala: neurobiological aspects of emotion, memory,and mental dysfunction. New York: Wiley-Liss. p 1–66.

Baimbridge KG, Miller JJ. 1984. Hippocampal calcium-binding pro-tein during commisural kindling-induced epileptogenesis: progres-sive decline and effects of anticonvulsants. Brain Res, 324:85–90.

Baimbridge KG, Celio MR, Rogers JH. 1992. Calcium-binding proteinsin the nervous system. Trends Neurosci 15:303–308.

Borhegyi Z, Leranth C. 1997. Distinct substance P- and calretinin-containing projections from the supramammillary area to thehippocampus in rats; a species difference between rats and mon-keys. Exp Brain Res 115:369–374.

Braak H, Braak E. 1983. Neuronal types in the basolateral amygda-loid nuclei of man. Brain Res Bull 11:349–365.

Braak H, Braak E, Yilmazer D, DeVos RA, Jansen EN, Bohl J,Jellinger K. 1994. Amygdala pathology in Parkinson’s disease. ActaNeuropathol 88:493–500.

Braun K. 1990. Calcium-binding proteins in avian and mammalian

Fig. 15. Small bipolar (accessory basal nucleus) and small multipo-lar ir-neurons (lateral nucleus) and a large multipolar ir neuron (basalnucleus) (marked by a star) taken from section double-labelled withanti-CB and anti-CR, 8th gestational month. a: single labelled CR-ir

neurons (brown) (basal and lateral nucleus); b: single labelled CB-irneurons (blue) (basal and lateral nucleus); c: double-labelled CB- andCR-ir neurons (black).

16 SETZER AND ULFIG

central nervous system: Localization, development and possiblefunctions. Prog Histochem Cytochem 21:1–64.

Celio MR. 1986. Parvalbumin in most gamma-amminobutyric acid-containing neurons of the rat cerebral cortex. Science 231:995–997.

Celio MR. 1990. Calbindin D28k and parvalbumin in the rat nervoussystem. Neuroscience, 35:375–475.

DeLeon M, Covenas R, Narvaez JA, Aguirre JA, Gonzalez-Baron S.1994. Distribution of calbindin D–28k-immunoreactivity in the catbrain stem. Arch Ital Biol 132:229–241.

Ellis JH, Richards DE, Rogers JH. 1991. Calretinin and calbindin inthe retina of the developing chick. Cell Tissue Res 264:197–208.

Enderlin S, Norman AW, Celio MR. 1987. Ontogeny of the calcium-binding protein calbindin D28k in the rat nervous system. AnatEmbryol 177:15–28.

Frassoni C, Arcelli P, Selvaggio M, Spreafico R. 1998. Calretininimmunoreactivity in the developing thalamus of the rat: A marker ofearly generated thalamic cells. Neuroscience 83:1203–1214.

Freund TF, Gulyas AI. 1991. GABAergic interneurons containingcalbindin D28k or somatostatin are major targets of GABAergicbasal forebrain afferents in the rat neocortex. J Comp Neurol314:187–199.

Gomez TM, Snow DM, Letournau PC. 1995. Characterization ofspontaneous calcium transients in nerve growth cones and theireffect on growth cone migration. Neuron 14:1233–1246.

Hendrickson AE, van Brederode JF.M, Mulligan KA, Celio MR. 1991.Development of the calcium-binding proteins parvalbumin andcalbindin in monkey striate cortex. J Comp Neurol 307:626–646.

HiranoA, Zimmermann HM. 1962.Alzheimer’s neurofibrillary changes:A topographic study. Arch Neurol 7:227–242.

Hooper MW, Vogel FS. 1976. The limbic system in Alzheimer’s disease.Am J Pathol 85:1–20.

Jacobowitz DM, Winsky L. 1991. Immunocytochemical localization ofcalretinin in the forebrain of the rat. J Comp Neurol 304:198–218.

Jacobson M. 1991. Developmental neurobiology. Plenum Press NewYork.

Jamada M, Mehraein P. 1968. Verteilungsmuster der senilen Verand-erungen im Gehirn. Arch Psychiat Neurol 211:308–324.

Jiang M, Swann JW. 1997. Expression of calretinin in diverse neuro-nal populations during development of rat hippocampus. Neurosci-ence 81:1137–1154.

Kater SB, Mattson MP, Cohan CS, Connor JA. 1988. Calcium regula-tion of the neuronal growth cone. Trends Neurosci 11:315–321.

Kiss J, Csaki A, Bokor H, Kocsis K, Szeiffert G. 1997. Topographiclocalization of calretinin, calbindin VIP, substance P CCK andmetabotropic glutamate receptor ir neurons in the supramammil-lary and related areas of the rat. Neurobiology (Bp) 5:361–388.

Komuro H, Rakic P. 1996. Intracellular Ca21 fluctuations modulate therate of neuronal migration. Neuron 17:275–285.

Kostovic I. 1986. Prenatal development of the nucleus basalis complexand related fiber systems in man: a histochemical study. Neurosci-ence 17:1047–1077.

Kostovic I, Petanjek Z, Delalle I, Judas M. 1992. Developmentalreorganization of the human association cortex during perinatal andpostnatal life. In: Kostovic I, Knezevic S, Wisniewski HM, andSpilich GJ, editors. Neurodevelopment, aging, and cognition. Bos-ton: Birkhauser. p 3–17.

McDonald AJ. 1992. Cell types and intrinsic connections of theamygdala. In: Aggleton JP, editor. The amygdala: neurobiologicalaspects of emotion memory and mental dysfunction. New York:Wiley-Liss. p 67–96.

McDonald AJ. 1994. Calretinin immunoreactive neurons in the baso-lateral amygdala of the rat and monkey. Brain Res 667:238–242.

Miettinen R, Gulyas AI, Baimbridge KG, Jacobowitz DM, Freund T.1992. Calretinin is present in non-pyramidal cells of the rathippocampus - II. Co-existence with other calcium binding proteinsand GABA. Neuroscience 48:29–43.

Nitsch R, Soriano E, Frotscher M. 1990. The parvalbumin-containing

nonpyramidal neurons in the rat hippocampus. Anat Embryol181:413–425.

Panula P, Pirvola U, Auvinen S, Airaksinen MS. 1989. Histamine-immunoreactive nerve fibers in the rat brain. Neuroscience 28:585–610.

Parmentier M, Ghysens M, Rypens F, Lawson DEM, Pasteels JL,Pochet R. 1987. Calbindin in vertebrate classes: Immunohistochemi-cal localization and Western blot analysis. Gen Comp Endocrinol65:399–407.

Parmentier M, Passage E, Vassart G, Mattei MG. 1991. The humancalbindin D28k (calb1. and calretinin (calb2. genes are located at8q21.3----q22.1 and 16q22----q23, respectively, suggesting a commonduplication with the carbonic anhydrase isozyme loci. CytogenetCell Genet 57:41–43.

Pasteels B, Miki BN, Hatakenaka S, Pochet R. 1987. Immunobiochemi-cal cross-reactivity and electrophoretic comigration between calbin-din D–27k and visinin. Brain Res 412:107–113.

Pasteels B, Rogers J, Blachier F, Pochet R. 1990. Calbindin andcalretinin localization in retina from different species. Vis Neurosci5:1–16.

Paul A, Ulfig N. 1998. Lectin staining in the basal nucleus (Meynert)and the hypothalamic tuberomamillary nucleus of the developinghuman prosencephalon. Anat Rec 252:149–158.

Pitkanen A, Amaral DG. 1993. Distribution of calbindin-D28k immu-noreactivity in the monkey temporal lobe: The amygdaloid complex.J Comp Neurol 331:199–224.

Rizvi TA, Ennis M, Behbehani MM, Shipley MT. 1991. Connectionsbetween the central nucleus of the amygdala and the midbrainperiaqueductal gray: topography and reciprocity. J Comp Neurol303:121–131.

Rogers JH. 1989. Two calcium-binding proteins mark many chicksensory neurons. Neuroscience 31:697–709.

Rogers J, Khan M, Ellis J. 1990. Calretinin and other CaBPs in thenervous system. Adv Exp Med Biol 269:195–203.

Schierle GS, Gander JC, D’Orlando C, Celio MR, Vogt-WeisenhornDM. 1997. Calretinin-immunoreactivity during postnatal develop-ment of the rat isocortex: a qualitative and quantitative study. CerebCortex 7:130–142.

Solbach D, Celio MR. 1991. Ontogeny of the calcium-binding proteinparvalbumin in the rat nervous system. Anat Embryol 184:103–124.

Sorvari H, Soininen H, Pitkanen A. 1996a. Calretinin-immunoreactivecells and fibers in the human amygdaloid complex. J Comp Neurol369:188–208.

Sorvari H, Soinen H, Pitkanen A. 1996b. Calbindin-D28k-immunore-active cells and fibres in the human amygdaloid complex. Neurosci-ence 75:421–443.

Sorvari H, Miettinen R, Soininen H, Paljarvi L, Karkola K, PitkanenA. 1998. Calretinin-immunoreactive terminals make synapses oncalbindin D28k-immunoreactive neurons in the lateral nucleus ofthe human amygdala. Brain Res 783:355–358.

Spitzer NC. 1994. Spontaneous Ca21 spikes and waves in embryonicneurons: signaling systems for differentiation. Trends Neurosci17:115–118.

Stevens JR. 1973. An anatomy of schizophrenia? Arch Gen Psychiat29:177–189.

Torrey EF, Peterson MR. 1974. Schizophrenia and the limbic system.Lancet 2:942–946.

Ulfig N. 1989. Configuration of the magnocellular nuclei in the basalforebrain of the human adult. Acta Anat 134:100–105.

Ulfig N, Nickel J, Bohl J. 1998a. Monoclonal antibodies SMI 311 andSMI 312 as tools to investigate the maturation of nerve cells andaxonal patterns in human fetal brain. Cell Tissue Res 291:433–443.

Ulfig N, Setzer M, Bohl J. 1998b. Transient architectonic features inthe basolateral amygdala of the human fetal brain. Acta Anat163:99–112.

Volpe JJ. 1994. Neurology of the newborn. Philadelphia: Saunders.


differential expression of calbindin and calretinin in the human fetal amygdala

Documents