Parvalbumin in the human anterior cingulate cortex: morphological heterogeneity of inhibitory interneurons

BRAIN RESEARCH ELSEVIER

Brain Research 729 (1996) 45-54

Research report

Parvalbumin in the human anterior cingulate cortex: morphological heterogeneity of inhibitory interneurons Peter Kalus *, Dieter Senitz Neurobiological Research Laborato~w, Department of Psychiatry. Unil,ersity of Wiir~burg. Fi~chsleinstral3e 15. D~97080 Wi~r=burg.Germany Accepted 26 March 1996

Abstract

The calcium-binding protein parvalbumin (PV) is a marker for a certain subset of GABAergic cortical interneurons. In the present study, indirect immunocytochemistry with an antibody against PV was performed on serial sections of human anterior cingulate cortex (Brodmann's area 24), an important relay centre of the limbic system. PV-positive structures are distributed in a layer- and cell type-specific manner. Based on morphological features and laminar distribution pattern, PV-immunoreactive interneurons are subdivided into eight different classes. PV immunoreactivity within the neuropil comprises dendritic and axonal processes. Area 24 contains two densely immmunolabelled neuropil bands in layers III and Vb. Axon cartridges are preferably located in layers V and V[. The results provide a 'PV immunoarchitecture' as a basis for further studies of PV immunoreactivity under pathological conditions. PV is assumed to play a role in maintaining calcium homeostasis in nerve cells, and to modulate neuronal excitability and resistance to biochemical damage. On the other hand, PV immunoreactivity has recently been shown to undergo characteristic changes during different stages of brain maturation. Therefore, examination of PV-positive structures will provide new insights into cortical circuitry in neurodegenerative as well as neurodevelopmental disorders. Keywords: lntemeuron; Calcium-binding protein; Parvalbumin; Immunocytochemistry: Cingulate cortex; Neuronal circuitry

1. Introduction

The calcium-binding protein parvalbumin (PV) is a marker protein for certain subgroups of GABAergic inhibitory interneurons [3,4,8]. PV-immunoreactive cells have been identified in numerous brain regions of different species [5-7,9,18,20,22]. The physiological functions of PV in nerve cells are as yet unknown, although PV is strongly suggested to play a role in calcium homeostasis, thus influencing the neuronal properties of excitability and resistance to biochemical damage in a protective manner [4]. Recent immunocytochemical studies demonstrated characteristic changes of cortical PV distribution during different stages of brain maturation [1,2]. These aspects stimulated speculations of a possible significance of PV in neurodegenerative processes, in which disturbances of calcium metabolism may be of importance, and, thus, motivated several studies in human neuropsychiatric disorders including Alzheimer's disease, Pick's disease, epilepsy and others (for reviews see [3] and

* Corresponding author. Fax: (49) (931) 203425.

[16]). However, data from these examinations are inconsistent: For example, for Alzheimer's disease, significantly decreased numbers of cortical PV-immunoreactive neurons [24] as well as normal cell counts [14,17] were reported. Apart from methodological and material-dependent differences, the absence of an exact qualitative characterization of the PV-immunoreactive cells and fibres may be one cause for the divergent results of these studies. Until now, only few studies examined the differentiated morphology of PV-positive profiles in the human cerebral cortex. The group of Braak described the morphological variety of PV-immunoreactive structures in the human hippocampus [6] and entorhinal region [26]. Bliimcke and colleagues compared PV-positive cell types in the visual cortex of monkeys and humans [5]. Exact knowledge of the morphological features of the PV-positive material in healthy brains is an essential prerequisite for the examination in diseased states. For this purpose, the aim of the present study was a qualitative examination of the distribution of PV-positive structures in order to provide a "normal PV immunoarchitecture' of Brodmann's area 24 of the human anterior cingulate cortex (ACC).

0006-8993/96/$15.00 Copyright ~ 1996 Elsevier Science B.V. All rights reserved. PH S0006-8993(96)00415 5

46

P. Kalus, D. Sen#z/Brain Research 729 (1996)45 54

The ACC represents an important relay centre of the limbic system, being intercalated between the anterior thalamic nuclei and the presubicular region in the 'Papez circuit' [27,29,30]. Functional studies revealed the ACC to be involved in a number of complex cognitive tasks such as memory and learning processes, attention, social interaction, pain reception, emotional aspect of perception and somatic and visceral motor activities [13]. Dysfunctions of ACC areas have been observed in numerous neuropsychiatric diseases including obsessive-compulsive disorders, pathological anxiety, epilepsy, Alzheimer's disease, Gilles de la Tourette syndrome and schizophrenia [13].

2. Materials and methods The study was performed on the brains of five individuals who did not suffer from any neurological or psychiatric disorder (Table 1). In all cases, degenerative, ischemic or infectious changes were excluded by careful neuropathological evaluation. Blocks comprising the ACC of the left hemisphere were fixed in a buffered mixture of 4% paraformaldehyde and 0.2% picric acid for 48 h. After washing out the fixative, blocks were treated with a cryoprotective solution of sucrose and serially cut into 60-p~m-thick coronal sections using a freezing microtome. Indirect immunocytochemistry was conducted on freefloating sections, which were carefully kept in motion on a rotating table to allow a homogeneous penetration by the immunoreagents. Preincubation procedures with normal horse serum (Vector, Burlingame CA, USA), 3% hydrogen peroxide and 0.3% Triton X-100 (Sigma, Deisenhofen) were carried out in order to obviate non-specific staining and to enhance membrane permeability. They were followed by primary antibody incubation with a monoclonal antibody against parvalbumin (clone no. 235; Sigma, Deisenhofen, Germany) [8] for 48 h at a dilution of 1:500 at 4°C. Secondary antibody reaction was performed with biotinylated anti-mouse-IgG raised in horse (Vector) for 2 h at room temperature. Subsequent incubation with avidin-biotin-peroxidase complex (Vector) for 3 h was followed by the visualization reaction using a 0.05% solution of 3',3'-diaminobenzidine (Sigma) containing 0.01% hydrogen peroxide. Adjacent sections were stained with toluidine blue for cytoarchitectonical delimitation. Some sections were counterstained with either gallocyanin chrom-alum or toluidine blue after immunostaining. All the sections from one brain used for quantification of PV-positive cells were incubated at the same time and in identical solutions. Control sections, in which the primary or secondary antibody was omitted or the primary

Table 1 Patient data No.

Sex

Age (years)

Cause of death

Post-mortem delay (h)

1 2 3 4 5

m in f m m

40 67 78 86 97

Acute drug intoxication Pulmonaryembolism Acute myocardiacinfarction Pulmonaryembolism Aspirationpneumonia

10 12 20 24 18

antibody was preabsorbed with excess quantities of the antigen, did not show specific immunoreactivity. For quantitative determination of the relative laminar distribution of PV-immunoreactive neuronal profiles, lefthemispheric ACC from three brains was available. The profile numbers in test fields of serial sections spanning the total thickness of the tissue blocks, gained by systematical random sampling, were determined [31]. After identifying the exact laminar position of a test field at low magnification, an unbiased counting frame of 160 ~ m X 160 I~m was optically superimposed over the visual field and all soma profiles of PV-positive neurons containing a distinct nucleus were counted using a 63 X oil-immersion objective [15,22]. Counting was carried out in 100 test fields per lamina within area 24c, in which laminar borders can be determined most precisely. No corrections were made for cell size and tissue shrinkage.

3. Results 3.1. Cytoarchitecture

Area 24 is an agranular proisocortex characterized by a marked, cell-dense layer Va and a relatively cell-sparse layer Vb [30]. On the basis of differences in laminar constitution, it can be divided into three main subdivisions: Area 24a is located within the callosal sulcus and has a thin, but nevertheless prominent lamina Va. Layer Vb contains numerous spindle cells. In this subarea, layers II and III are difficult to distinguish. The border between layer Vb and VI is partly ill-defined. Area 24b lies rostral and dorsal to area 24a and has the most voluminous layer Va in the ACC. The density of spindle cells in layer Vb is decreased compared to area 24a. Layers II and III are well-defined in this location. Area 24c covers the rostral depth of the cingulate sulcus. In this subarea, layer Va is again thinner and contains more small pyramidal neurons (Fig. l d). Layer Vb appears to be especially cell-sparse. 3.2. Immunocytochemistrv

PV immunoreactivity is found in neuronal nuclei, perikarya, dendrites and axonal structures of non-pyra-

Fig. 1. Cortical traverses of area 24a (a), 24b (b) and area 24c (c and d). PV-immunoreactivepreparations (a c) and Nissl preparation (d). CC, corpus callosum. Bar = 300 Ixm (valid for a-d).

P. Kalus, D. Sen#z/Brain Research 729 (1996) 45-54

47

P. Kalus, D. Sen#z/Brain Research 729 (1996)45 54

48

midal neurons, lending a Golgi-like aspect to these cells. However, although axons are generally well-labelled by anti-PV, the individual axon belonging to a particular cell can only occasionally be traced in adult human material. In the ACC, PV-immunoreactive ceils show characteristic features with respect to their perikarya, dendritic arborization, and location within the cortex, thus providing the basis for a classification into eight easily distinguishable morphological categories, indicated by the capital letters A to H (Figs. 3-7, Fig. 9). The immunoreactive material distributed throughout the neuropil is composed of dendritic and axonal fibres and of puncta, point-like structures, which were shown by electron microscopy to represent axonal terminal boutons [10,11]. Furthermore, axon cartridges, the characteristic arrangements of axonal termination structures of chandelier cells, can be found in ACC (see below). Corresponding to the growing differentiation of cortical architecture from area 24a towards area 24c, the distribution of PV-positive profiles acquired a more and more differentiated laminar specificity throughout the subareas of area 24 (Fig. 1a-c).

~

A

i/

C

III

3.2.1. PV-lmmunoreactive cells

Layer I was totally devoid of PV-positive somata within all subdivisions of area 24. Layer II exhibited a relatively low density (14.73%) of PV-immunoreactive cell profiles in area 24c (Fig. 2). Type A neurons show a perpendicularly oriented, fusiform soma with a maximum diameter of about 18 txm x 8 Ixm and possess a bipolar or bitufted dendritic arborization pattern with thin and short dendrites (Figs. 3 and 4). Type B neurons are small globoid, sometimes granulelike multipolar cells of a maximum soma diameter of 15 I~m with short dendrites arising from all parts of the cell body (Figs. 3 and 4). Type A and B cells represented the clear majority of immunoreactive neurons in area 24.

Fig. 3. Camera lucida drawings of PV-immunoreactive classes of interneurons in layer II. Bar = 50 ~m.

b

\ []

IC C

16 10

s 0

I

II

III

Vo

Vb

VI

LAYER

Fig. 2. Relative laminar distribution of PV-positive neuronal profiles in area 24c (mean values and standard deviation).

Fig. 4. Camera lucida drawings of PV-positive interneurons in layer III Bar = 50 /xm.

49

P. Kalus, D. Senitz / Brain Research 729(1996)45 54

Large, vertically oriented bitufted cells (soma diameter of about 25 ~ m × 25 Ixm) with ovoid-shaped somata are named O'pe C cells (Fig. 3, Fig. 4, Fig. 7a). Their dendrites sprout from a stout trunk to form perpendicularly arranged narrow apical and basal dendritic trees. Type D neurons (Figs. 3 and 7a) are defined as large, dark and roundish cells with a soma diameter of about 25 txm. Their thick dendrites arise from the soma in a multipolar pattern, but then course vertically upwards and downwards to form columnar apical and basal arborization zones of fibres which branch no further and are arranged parallel to one another. Occasionally, single dendrites of type D cells were observed to traverse layer II horizontally at the level of their parent cell (Fig. 3). Layer IIl exhibited the largest relative neuronal profile density of all layers (33.30%). As in layer II, the main contingent of somata was provided by type A and type B cells (Fig. 4). In addition, the lower part of layer III harboured a variant of the large bitufted type C cells equipped with extraordinarily long basal dendrites, which could be traced up to 600 Ixm, traversing large parts of layer V (Fig. 4). Conversely, the upper part of layer III contained bitufted type C neurons with especially long apical dendrites reaching layer I1, while their basal dendritic tree was only sparsely developed. Very large multipolar 'giant' cells with a soma diameter of more than 30 Ixm could occasionally be found at the

C

Vb

Fig. 5. Camera lucida drawings of PV-positivenonpyramidal neurons in layer Va/b. Bar = 50 txm.

\

7b

VI

Fig. 6. Camera lucida drawings of PV-immunoreactiveinterneurons in layers Vb and VI. Bar = 50 Ixm.

lower border of layer III. These t3'pe E cells (Fig. 4) possess an irregularly elliptical to roundish soma with thick, varicose dendrites preferentially entering layer Va. Layer Va exhibited a medium density of immunoreactive neuronal profiles (17.01%). The prevailing elements were again type A and type B cells. Moreover, large type C cells with numerous long dendrites were identified, which, in this location, often possess 'recurrent' dendrites joining the dendritic field of the opposite pole of the cell (Fig. 5). Another cell type occurring in lamina Va (t3'pe F cells) possesses medium-sized (diameter 20 to 25; ~xm) round multipolar somata giving rise to thick primary dendrites, which spread radially from the perikaryon, and often proceed slightly curved before bifurcating after a course of about 20 pxm (Fig. 5). The long secondary dendrites may branch again and radiate for more than 300 p~m in all directions. At the lower border of layer Va, type E intemeurons were occasionally found. In layer Vb, neuronal profile density was found to increase again (22.03%). The layer contained a few small type A and type B cells. Multipolar type F cells and large bitufted type C cells with 'recurrent' dendrites could be identified in this location.

50

P. Kalus, D. Sen#z/Brain Research 729 (1996)45-54

Type G cells possess a m e d i u m - to large-sized (diameter 25 to 30 ~ m ) , m u l t i a n g u l a r cell body with stout stem dendrites arising from the edges of the soma and usually bifurcating close to the soma (Figs. 6 and 7b). The stout secondary dendrites take a long (up to 500 txm), oblique course through the cortex. Type H cells have a large, horizontally arranged fusiform

soma with a diameter of about 30 p~m X 10 ~ m with bipolar or bitufted dendritic d o m a i n s (Figs. 6 and 7c). Dendrites originating from the same tuft may take either a long horizontal or a vertical course, sometimes reaching a length of 500 txm. Layer VI contained relatively few i m m u n o r e a c t i v e cell profiles (12.75%). Rarely, multiangular type G and hori-

Fig. 7. Photomicrographs of several interneuron types of area 24c. a: type C and D neurons in the lower part of layer I1. Note the increasing density ot" the PV-immunoreactive fibre network in upper layer III. b: multiangular type G neuron of layer Vb. Note the patchy structure of the band-like neuropil immunoreactivity within layer Vb. c: two horizontal type H neurons of layer V1. Bar = 50 ~m (a c).

51

P. Kalus. D. Sen#=/Brain Research 729 (I 996) 45-54

Fig. 8. Axon cartridges in layers Vb and VI of area 24c. a: some cartridges are indicated by arrows: the large arrow points to the cartridge shown in b at a higher magnification. Bar = 100 Ixm. b: a single axon cartridge, consisting of two parallel oriented rows of PV-immunoreactive ax(m terminals. Bar = 5 ~&nl.

zontal type H neurons as well as some multipolar type F cells were seen. Furthermore, a few type B and type A cells were scattered throughout the layer.

III, where they were sometimes difficult to recognize because of the dense fibre network (Fig. 8). PV-Immunoreactive baskets, another specific type of axon terminal arrangement, were not found in our material.

3.2.2. P V - I m m u n o r e a c t i c e material within the neuropil

The intensity of PV-immunoreactive neuropil structures such as dendritic and axonal fibres and axon terminals (puncta and axon cartridges) varied considerably between the different layers and areal subdivisions (Fig. l a - c ) . Regarding neuropil immunoreactivity, areas 24b and 24c were characterized by two densely immunoreactive fibre plexuses in layers III and Vb, which are macroscopically visible as dark bands (Fig. lb, c). Adjacent frontal areas did not show this typical two-band pattern and can thus easily be distinguished from area 24. In area 24a, only one immunoreactive band in layer III was recognized (Fig. l a). In all subdivisions of area 24, layer I harboured only a few PV-immunoreactive apical dendrites from layer Il neurons and short axonal fibres, while the neuropil of layer II exhibited a somewhat higher level of PV immunoreactivity. Layer III contained a conspicuous dense band of neuropil immunoreactivity composed of large numbers of PV-positive axonal fibres, dendrites and puncta (Fig. l a - c , Fig. 7a). The density of immunoreactive material in lamina Va was markedly lower than in layer III, while layer Vb again exhibited a large amount of PV-positive material (Fig. 7b). The intensity of the labelling of layer VI was rather lower than that of layer Va. Axon cartridges were especially evident in layers Vb and VI, and. to a markedly lesser extent, in layers Va and

4. Discussion The present study provides for the first time a morphological classification of PV-immunoreactive interneurons in the human ACC. According to the shape and size of cell somata, their location within the cortex and the dendritic arborization pattern, PV-positive cells can be divided into eight different types (Fig. 9).

A nn

HI

B C

D

E

F

G

H

,

.~%

Vb

Fig. 9. Summary diagram with schematic drawings of the different PV-immunoreactive interneuron classes A H and Iheir laminar distribution in area 24.

52

P. Kalus, D. Sen#z/Brain Research 729 (1996) 45 54

4.1. Comparison of PV-positir, e neurons in different brain regions and species Previous studies on PV immunocytochemistry in the ACC of the rat, monkey and human [18] did not describe the differential morphology of PV-positive neurons in detail. However, immunocytochemical examinations of other brain areas provide enough information to reveal striking similarities with several interneuron types defined in this present study. In the visual cortex of the monkey, Van Brederode and colleagues [28] found numerous classes of PV-positive nerve cells; however, only two cell types were described by them in detail: their 'multipolar neurons with predominantly vertically arrayed dendrites' (their Fig. 2D), which were located in layers I I / I I I A and were considered to represent chandelier cells, closely resemble our type D cells, which were found only in layer II of area 24. The second cell type of Van Brederode et al., described as 'multipolar stellate neurons' (their Fig. 9E) may correspond to our type G cells, which are similarly located in layers V and VI. Bl~imcke and associates [5] compared the monkey and human visual cortices and reported numerous similar PVpositive nonpyramidal neuron types in both species, leading them to assume a comprehensive structural homology amongst interneurons. The main contingent of immunoreactive cells observed by them consisted of 'small nonspiny stellate cells' (their Fig. 7), located in layers II to VI, the morphology, frequency and location of which are very similar to our type B cells. The cell type depicted in their Fig. 6d, which was found only in human cortex and was denoted as the 'double bouquet cell', may correspond to our type D neuron. The cell type shown in Fig. 8i of Bliimcke et al. and designated the 'chandelier cell' resembles our multipolar type F neuron. Class H neurons in our material are very similar to the horizontally arranged ovoid-shaped cells depicted in their Figs. 8c and 8j. The 'basket cells' drawn in their Figs. 8g and 8k can not readily be compared to cells in our material. Lurid and Lewis [21] provided a precise description of the local circuit neurons of monkey prefrontal cortex by means of Golgi impregnation, thus distinguishing 13 classes of interneurons. A comparison with different immunocytochemical markers showed PV to be contained in several types of prefrontal interneurons, two of which were classified as 'chandelier neurons' and 'wide arbor neurons'. The chandelier neuron of layer II, the Golgi aspect of which is shown in their Fig. 3, resembles our type C cell. The 'multipolar wide arbor neuron' of Lund and Lewis, which is shown in their Fig. 6, may correspond to our type D neuron, but, unlike type D neurons, it was located in layers III and V. Interestingly, there are some similarities between other of our cell types and the interneurons of Lund and Lewis, which did not contain PV but other immunocytochemical markers in the monkey prefrontal cortex; the

CCK-positive 'medium arbor neuron' depicted in their Fig. 21 closely resembles our multiangular type G cell, the calretinin-containing 'vertical cascade neuron' (their Fig. 25) resembles our type C cell, and the calbindin-positive 'neurogliaform neuron' (their Fig. 24) may be compared 1o our type B cell. If these interneuron classes are in fact morphologically homologous in monkey and human cortex, they might possess different functional properties in the two species, as indicated by the expression of different calcium-binding proteins or neuropeptides. The Braak group examined PV immunoreactivity in the allocortical entorhinal region and discriminated diverse types of multipolar and bipolar neurons [26]. The dendritic arborization pattern of their large multipolar cells (their Fig. 4b and c) closely resembles that of our type G neurons. Some of their large bipolar neurons (their Fig. 5a and d) may readily be compared to our type H cells, which, in contrast to the entorhinal neurons, are restricted to the deep layers. Contrary to our findings, the group of Braak also described the occurrence of PV-positive cells within the molecular layer. Although comparisons of nerve cell types based on their shape always remain somewhat speculative because of the subjectivity involved in the rating of qualitative morphological features, striking similarities between most of the interneuron classes observed in human ACC and cell types described in the above cited studies are obvious. This may be an indication of the usefulness of the presented classification scheme. Furthermore, the results support the assumption of functional homologies for interneuron types in several brain regions in different primate species.

4.2. Chandelier cells and axon cartridges Chandelier neurons and their specific terminal structures, axon cartridges, are generally considered to be PVimmunoreactive [1,9,21,26]. However, as has been shown in Golgi preparations, the somata and dendrites of chandelier cells may have a quite variable appearance [23]. Therefore, in PV preparations the identification of chandelier cells is difficult because the axon of a specific cell can only rarely be continuously traced within the neuropil. In the human entorhinal cortex, Schmidt and colleagues [26] described a characteristic cell type with dendritic varicosities and spine-like appendages showing continuous axonal connections to axon cartridges. In our material, this neuron type could not be identified in area 24, although the presence of axon cartridges points to the existence of chandelier cells. On the basis of Golgi morphology [23], our type C or type D neurons may represent chandelier neurons, but this could not be confirmed with PV staining. Although the cingulate cortex was the first site where chandelier cells were observed [23], their specific termination structures, axon cartridges, have not been demonstrated in this location by PV immunocytochemistry in previous studies [18]. However, in our material, PV-im-

P. Kalus, D. Senitz / Brain Research 729 (I 996) 45-54

munoreactive cartridges could be readily found in area 24, located mainly in the deep layers V and VI. In monkey prefrontal cortex, PV-positive axon cartridges were reported to be most numerous in layers II and III, while the infragranular layers contained only a few [20,32]. A similar laminar distribution of cartridges was reported for the monkey pre- and postcentral gyri by DeFelipe and coworkers [11]. [n contrast, in human temporopolar cortex, axon cartridges were concentrated in layer IV and, to a lesser degree, in layers V and VI, while layers II and IIl included substantially fewer [12]. In the human entorhinal cortex, cartridges were located mainly in the deep layers pri-cx and pri-~, [26]. The different laminar distribution patterns of axon cartridges in diverse locations is likely to be of significance for the differential output activity of cortical regions be-cause cartridges are presumed to play an important role in the modulation of the axonal information flow of pyramidal neurons by forming inhibitory axo-axonal synapses on the initial axon segments of these cells [1,2,10].

4.3. PV-immunoreactive fibre plexuses A characteristic feature of human area 24 is the intense band-like labelling of two PV-positive fibre plexuses in layers III and Vb, which was similarly shown for monkey ACC [18]. Previous studies demonstrated analogous phenomena in structurally and functionally differing types of cerebral cortex: In monkey prefrontal association cortex, Cond~ and colleagues [9] observed a broad PV-immunoreactive band located in deep layer III up to layer V. In human temporal association cortex obtained neurosurgically from patients with intractable temporal lobe epilepsy, DeFelipe and co-workers [12] found the band-like labelling to be restricted to the deeper half of layer III. The Braak group established that band-like PV labelling was contained in layers pre-c~ and pre-',/ of the human entorhinal cortex [26]. In primary sensory regions of the monkey (somatosensory and visual cortex), dense, patchy PV immunoreactivity was found to be located in layer IV of area 3b and in layer IVc of area 17 as well as in the upper part of layer Vl of both cortices [11]. In total, diverse types of cortex have been demonstraled to exhibit different distribution patterns of PV-positive fibre plexuses. In monkey visual and somatosensory cortex, electron microscopic examination revealed that the zones of high density immunoreactivity in the neuropil contain a great number of PV-positive axon terminals from excitatory, probably extrinsic, projection neurons [11]. Therefore, PV-immunoreactive bands were considered to represent thalamocortical termination fields. This hypothesis is supported by the observation that the thalamic nuclei projecting to primary visual and somatosensory cortex contain numerous PV-positive projection neurons [19]. The question arises as to whether PV-immunoreactive bands in other brain regions also correspond to termination zones

53

from thalamic relay neurons. For the entorhinal region, Schmidt and colleagues [26] speculate that the axon terminals within the bands may not represent extrinsic neurons because there is no known strong afferent PV-immunoreactive projection to the upper layers of the entorhinal cortex. Furthermore, the thalamic nuclei projecting to the entorhinal region contain only few or no PV-positive projection neurons [19]. There is still not much information about the laminar distribution of the thalamocortical termination fields in area 24, but electrophysiological findings support the assumption that in this region PV-immunoreactive fibre plexuses might represent thalamocortical termination zones: in the rabbit, Sikes and Vogt [25] observed that area 24 neurons displaying nociceptive responses similar to those of neurons of thalamic nuclei projecting to the ACC were mainly located in layers III and V, the location of the PV-positive bands.

4.4. Conclusions In the ACC, PV immunocytochemistry may provide detailed information about (a) the distribution patterns of morphologically heterogeneous subgroups of' GABAergic local circuit neurons by Golgi-like staining of cell somata and dendrites, (b) the input features of cortical areas by labelling terminal structures of thalamocortical afferents, and (c) the output properties of cortical areas by detecting axon cartridges, the presynaptic sites of inhibitory synapses modulating the activity of projection neurons. Therefore, PV immunocytochemistry may serve as a means for further improving our knowledge of detailed cortical circuitry in the healthy state as well as in neurodegenerative and neurodevelopmental disorders involving the cingulate cortex.

Acknowledgements The authors are indebted to Dr. M. Bauer (Dept. of Neuropathology, Wiesloch) and Dr. W. Gsell (Dept. of Neurochemistry, Wtirzburg) for providing the', autopsy material. The skilful technical assistance of Mrs. M. Winnig and Mrs. R. Senitz is gratefully acknowledged. The authors would like to thank Dr. S. Klinke and Mr. P. Foley for critical reading of the manuscript.

References [1] Akil, M. and Lewis, D.A., Postnatal development of parvalbumin immunoreactivity in axon terminals of basket and chandelier neurons in monkey neocortex, Prog. Neuro-Psvchopharrnacol. Biol. Psychiat., 16 (1992) 329-337. [2] Anderson, S.A., Classey, J.D., Condt~, F., Lurid, J.S. and Lewis D.A., Synchronous development of pyramidal neuron dendritic spines and parvalbumin-immunoreactive chandelier neuron axon terminals in layer III of monkey prefrontal cortex, Neurosciem'e, 67 (1995) 7-22.

54

P. Kalus, D. Senitz./ Brain Research 729 (1996) 45-54

[3] Andressen, C., Bliimcke, I. and Celio, M.R., Calcium-binding proteins: selective markers of nerve cells, Cell Tissue Res., 71 (1993) 181-208. [4] Baimbridge, K.G., Celio, M.R. and Rogers, H., Calcium-binding proteins in the nervous system, Trends Neurosci., 15 (1992) 303308. [5] Bliimcke, I., Hof, P.R., Morrison, J.H. and Celio, M.R., Distribution of parvalbumin immunoreactivity in the visual cortex of old world monkeys and humans, J. Comp. Neurol., 301 (1990) 417-432. [6] Braak, E., Strotkamp, B. and Braak, H., Parvalbumin-immunoreactive structures in the hippocampus of the human adult, Cell Tissue Res., 264 (1991) 33-48. [7] Celio, M.R., Parvalbumin in most gamma-amino-butyric acid-containing neurons of the rat cerebral cortex, Science, 231 (1986) 995-997. [8] Celio, M.R., Baier, W., Sch~rer, L., DeViragh, P.A. and Gerday, C., Monoclonal antibodies directed against the calcium binding protein parvalbumin, Cell Calcium, 9 (1988) 81-86. [9] Cond&, F., Lund, J.S., Jacobowitz, D.M., Baimbridge, K.G. and Lewis, D.A., Local circuit neurons immunoreactive for calretinin, calbindin D-28k or parvalbumin in monkey prefrontal cortex: distribution and morphology, J. Comp. Neurol., 341 (1994)95-116. [t0] DeFelipe, J., Hendry, S.H.C. and Jones, E.G., Visualization of chandelier cell axons by parvalbumin immunoreactivity in monkey cerebral cortex, Proc, Natl. Acad. Sci. USA, 86 (1989) 2093-2097. [11] DeFelipe, J. and Jones, E.G., Parvalbumin immunoreactivity reveals layer IV of monkey cerebral cortex as a mosaic of microzones of thalamic afferent terminations, Brain Res., 562 (1991) 39-47. [12] DeFelipe, J., Sola, R.G., Marco, P., DelRio, M.R., Pulido, P. and Ramon y Cajal, S., Selective changes in the microorganization of the human epileptogenic neocortex revealed by parvalbumin immunoreactivity, Cereb. Cortex, 3 (1993) 39-48. [13] Devinsky, O., Morrell, M.J. and Vogt, B.A., Contributions of anterior cingulate cortex to behaviour, Brain, 118 (1995) 279-306. [14] Ferrer, I., Zujar, M.J., Rivera, R., Soria, M., Vidal, A. and Casas, R., Parvalbumin-immunoreactive dystrophic neurites and aberrant sprouts in the cerebral cortex of patients with Alzheimer's disease, Neurosci. Lett., 158 (1993) 163-166. [15] Gundersen, H.J.G., Bendtsen, T.F., Korbo, L., Marcussen, N., Moller, A., Nielsen, K., Hyengaard, J.R., Pakkenberg, B., Sorensen, F.B., Vesterby, A. and West, M.J., Some new, simple and efficient stereological methods and their use in pathological research and diagnosis, Acta Pathol. Microbiol. Immunol. Scand., 96 (1988) 379-394. [16] Heizmann, C.W. and Braun, K., Changes in Ca2+-binding proteins in human neurodegenerative disorders, Trends Neurosci., 15 (1992) 259-264. [17] Hof, P.R., Cox, K., Young, W.G., Celio, M.R., Rogers, J. and Morrison, J.H., Parvalbumin-immunoreactive neurons in the neocorrex are resistant to degeneration in Alzheimer's disease, J. Neuropathol. Exp. Neurol., 50 (1991) 451 462.

[18] Hof, P.R., Ltith, H.-J., Rogers, J.H. and Celio, M.R., Calcium-binding proteins define subpopulations of interneurons in cmgulate cortex. In B.A. Vogt and M. Gabriel (Eds.), Nenrot~iolow qf Cingulate Cortex and Limbie Thalamus." A Comprehensi~:e Hamlbook, Birkh~iuser, Basel, 1993, pp. 181-205. [t9] Jones, E.G. and Hen&y, S.H.C., Differential calcium-binding protein immunoreactivity distinguishes classes of relay neurons in monkey thalamic nuclei, Eur. J. Neurosci., 1 (1989) 222-246. [21)] Lewis, D.A. and Lund, J.S., Heterogeneity of chandelier neurons in monkey neocortex: corticotropin-releasing factor and parvalbumin immunoreactive populations, J. Comp. Neurol., 293 (1991)) 599-615. [21] Lurid, J.S. and Lewis, D.A., Local circuit neurons of developing and mature macaque prefrontal cortex: Golgi and immunocytochemical characteristics, J. Comp. Neurol., 328 (1993) 282-312. [22] McMullen, N.T., Smelser, C.B. and de Venecia, R.K., A quantitative analysis of parvalbumin neurons in rabbit auditory neocortex, J. Comp, Neurol.. 349 (1994) 493-511. [23] Peters, A. and Jones, E.G. (Eds.), Cerebral Cortex, Vol. 1, Cellular Components ~[" the Cerebral Cortex, Plenum Press, New York, 1984. [24] Satoh, J.. Tabira, T., Sano, M., Nakayama, H. and Tateishi, J., Parvalbumin-immunoreactive neurons in the human central nervous system are decreased in Alzheimer's disease, Acta Neuropathol., 81 (1991) 388 395. [25] Sikes, R.W. and Vogt, B.A., Nociceptive neurons in area 24 of rabbit cingulate cortex, J. Neurophysiol., 68 (1992) 1720-1732. [26] Schmidt, S., Braak, E. and Braak, H., Parvalbumin-immunoreactive structures of the adult human entorhinal and transentorhinal region, Hippocampus, 3 (1993) 459-470. [27] Stephan, H., Allocortex. In W. Bargmann (Ed.), Handbuch der miktvskopischen Anatomic, Vol. 4/9, Springer, Berlin, 1975. [28] Van Brederode, J.F.M., Mulligan, K.A. and Hendrickson, A.E., Calcium-binding proteins as markers for subpopulations of GABAergic neurons in monkey striate cortex, J. Comp. Neurol.. 298 (1990) 1 22. [29] Vogt, B.A., Structural organization of cingulate cortex: areas, neurons, and somatodendritic transmitter receptors. In B.A. Vogt and M. Gabriel (Eds.), Neurobiology of Cingulate Cortex arid Limbie Thalamus: A Comprehensit e Handbook, Birkh~iuser, Basel. 1993, pp. 19-70. [30] Vogt, B.A.. Nimchinsky, E.A., Vogt, L.J. and HoE P.R., Human cingulate cortex: Surface features, flat maps, and cytoarchitecture, J. Comp. Neurol., 359 (1995) 490-506. [31] West, M.J. and Gundersen, H.J.G., Unbiased stereological estimation of the number of neurons in the human hippocampus, J. Comp. Neurol., 296 (1990) 1-22. [32] Williams, S.M., Goldman-Rakic, P.S. and Leranth, C., The synaptology of parvalbumin-immunoreactive neurons in the primate prefrontal cortex, J. Comp. Neulwl., 320 (1992) 353-369.