- Email: [email protected]
Coexistence of parvalbumin and GABA in nonpyramidal neurons of the rat entorhinal cortex
BRAIN RESEARCH ELSEVIER
Brain Research 706 (1996) 113-122
Research report
Coexistence of parvalbumin and GABA in nonpyramidal neurons of the rat entorhinal cortex Maria Miettinen, Esa Koivisto, Paavo Riekkinen Sr., Riitta Miettinen * Department of Neurology and A.L Virtanen Institute, University of Kuopio, PO Box 1627, FIN-70211 Kuopio, Finland Accepted 12 September 1995
Abstract The possible coexistence of the calcium-binding protein, parvalbumin, with the major inhibitory neurotransmitter, gamma-aminobutyric acid (GABA), and its synthesizing enzyme, glutamate decarboxylase (GAD), was studied in nonpyramidal cells of the rat medial and lateral entorhinal cortex. The material was analyzed by two different methods, the first of which was a mirror technique where the possible coexistence of two different antigens was analyzed from cells cut in half at the surface of the adjacent section. The other method consisted of analyzing double immunofluorescent-stained sections with a confocal microscope. The colocalization analysis revealed that all parvalbumin-immunoreactive neurons (mirror technique n = 688 and confocal microscopy n = 644) in all layers of the medial and lateral entorhinal cortex were also immunopositive for GABA or GAD. Parvalbumin-cells made up 52% of the GABA cells in most of the layers in the medial and lateral entorhinal cortex. In layer III of the entorhinal cortex, the proportion was about 40%. Thus, parvalbumin-containing neurons in the entorhinal cortex represent a large GABAergic cell population, which is likely to play an important role in controlling both the input and the output of the entorhinal cortex. Keywords: Calcium-binding protein; Confocal microscopy; Glutamate decarboxylase; Hippocampal formation; Immunocytochemistry; Colocalization
1. Introduction The entorhinal cortex occupies a key position in the neuronal circuitry between the hippocampus and various cortical and subcortical areas. Giving rise to the perforant path, which is the main cortical pathway to the hippocampus, the entorhinal cortex provides the hippocampal formation with most of the cortical sensory information [33]. The spiny stellate and pyramidal cells in layers II and III of the entorhinal cortex are the main source of input into the dentate gyms. In addition, there are projections from layer II to the CA1 subfield; but this layer may also innervate CA2 and CA3 regions and the subiculum. Layer III also projects to the CA1 and CA3 subfields. The deeper layers (IV-VI) are targets for input originating from the CA1 field of the hippocampus and from the subiculum and project back to the neocortex and subcortical areas. Thus, the entorhinal cortex is an important structure not only for getting information to the hippocampus but also for receiv-
* Corresponding author. Dept. of Neurology, Univ. of Kuopio, PO Box 1627, FIN-70211 Kuopio, Finland. Fax: (358) (71) 162048. Email: [email protected] 0006-8993/96/$15.00 © 1996 Elsevier Science B.V. All rights reserved SSDI 0 0 0 6 - 8 9 9 3 ( 9 5 ) 0 1 2 0 3 - 6
ing information from the hippocampus [1,16]. Recent studies have suggested that the entorhinal cortex plays a role in learning by integrating information and memories which have previously been processed by the hippocampus, amygdala or medial septum [14]. In several disorders of the central nervous system, the entorhinal cortex is known to be vulnerable. These disorders include for example, Alzheimer's disease, schizophrenia and epilepsy [2,4,8], all of which are accompanied by cognitive impairment. Alterations in cholinergic, noradrenergic, serotonergic and dopaminergic transmission as well as in the major inhibitory neurotransmission, gammaaminobutyric acid (GABA), may participate in impairment of learning and memory [10,27]. Recently, our group found that the number of neurons containing a calcium-binding protein, parvalbumin, in the entorhinal cortex of aged rats was smaller than in young rats [24]. This decline correlated with the learning and memory deficits observed in aged rats. Since parvalbumin is present in different types of GABAergic nonpyramidal neurons in several brain areas [5-7,9,12,13,17,18,25,30], our finding suggested that degeneration of the inhibitory system is involved in cognitive dysfunctions associated with aging. However, since there was no direct evidence that the
114
M. Miettinen et al. / Brain Research 706 (1996) 113-122
parvalbumin-neurons in the entorhinal cortex are GABAergic, our finding remained partly equivocal. Therefore, in the present study we investigated whether cells in the entorhinal cortex that contain parvalbumin do, in fact, contain GABA or its synthesizing enzyme, GAD.
using 3,3'diaminobenzidine as a chromogen. All the washes and antibody dilutions were done in 0.05M Tris-buffered saline, pH 7.4, which contained 1% normal serum. When the sections were to be used for cell-plotting, Triton X-100 was included in the solutions. The sections for the cell-
2. Materials and methods
2.1. Tissue preparation Wistar rats of 9 to 10 weeks old, (National Animal Center at the University of Kuopio, Finland; wt: 200-300 g) were used in this study. Deeply anaesthetized (chlornembutal, 0.3 ml/100 g, i.p.) rats were perfused transcardially with physiological saline (3 min) and then with 300 ml of fixative for 30 min. For GAD-immunohistochemistry, the fixative contained 4% paraformaldehyde, 0.05% glutaraldehyde, 0.2% picric acid in 0.1 M phosphate buffer (PB), pH 7.4. For GABA-immunohistochemistry, the fixative contained 2.5% paraformaldehyde, 1% glutaraldehyde and 0.2% picric acid in PB. Both fixatives were used for parvalbumin-immunohistochemistry. Before immunostaining, the material from the rats perfused with 1% glutaraldehyde was treated with 1% sodium borohydride for 30 min in order to reduce free aldehyde groups and double bonds, thus improving the immunoreactivity of protein antigens [20]. The brains were sectioned horizontally on a Vibratome at 60 /xm (1:5 series) and the sequential order of the sections was preserved. Cytoarchitectonic borders for precise localization of cortical areas and laminar borders in the entorhinal cortex were obtained from the adjacent Nissl- and acetylcholinesterase-stained (ACHE) sections. A modification of the method of Karnovsky-Roots [11] was used for AChE-staining.
B
2.2. Immunocytochemistry 2.2.1. Immunoperoxidase staining For analysis of the coexistence of parvalbumin with GABA or GAD analyzed with the mirror technique and for cell-plotting, free-floating sections were immunostained with the avidin-biotin method. After extensive washings in PB, the sections were incubated first in 10% normal goat or horse serum (depending on the origin of the secondary antiserum) and then at 4°C for 2 days in the primary antiserum: rabbit anti-parvalbumin (1:8000, PV-28, Swant) or mouse anti-parvalbumin (1:8000, PV-Mab, Swant). Adjacent sections were incubated either in rabbit anti-GAD (1:200, Chemicon) or mouse anti-GABA (1:100) [31]. These sections were incubated in the secondary antibody (1:50, biotinylated goat anti-rabbit IgG for rabbit primaries and biotinylated horse anti-mouse IgG for mouse primaries, Vector) at 4°C for 16 h. The third layer was avidin-biotinylated horseradish peroxidase complex (1:100, Vector). The immunoperoxidase reaction was developed
C Fig. 1. Light micrographs of horizontal sections from the entorhinal cortex of the rat. Nissl-staining (A) and acetylcholinesterase-staining (B) show the cytoarchitectonics of the ¢ntorhinal cortex. Distribution of parvalbumin-immunoreactivity is demonstrated in C. The border of the medial (MEA) and lateral (LEA) entorhinal cortex is indicated by an arrow. The layers of the entorhinal cortex are marked with Roman numerals (I-VI) and the lamina dissecans is indicated by asterisks. Scale bar = 500/zm; ab, angular bundle; PaS, parasubiculum; PERI, perirhinai area; PreS, presubiculum.
M. Miettinenet aL/ BrainResearch 706 (1996) 113-122 plotting were mounted on slides with Depex (BDH Laboratory Supplies, Gurr R) and covered with a coverslip. The sections for the mirror technique were treated with 0.5% OsO 4 for 30 min, dehydrated in ethanol and propylene oxide, and embedded in Durcupan (ACM, Fluka) on slides.
115
cent Nissl-stained sections the boundaries of the entorhinal cortex and its layers were drawn with camera lucida. Thereafter, the outlines were superimposed on scanned plots by using Canvas software on a Macintosh computer.
2.4. Analysis of colocalization 2.2.2. Immunofluorescence staining Double immunofluorescent staining for parvalbumin and GABA was used to analyze the possible coexistence of these two antigens with the confocal microscope. Freefloating sections were incubated in rabbit anti-parvalbumin (1:4000, PV-28, Swant) together with mouse anti-GABA (1:50) at 4°C for 2 days. The sections were then incubated in a mixture of Texas Red-labeled goat anti-rabbit (1:50, Vector) and biotinylated horse anti-mouse (1:50, Vector) at 4°C for 16 h. After that the biotinylated goat anti-rabbit antiserum was detected by fluorescein isothiocyanatelabeled avidin (FITC-avidin, 1:100, Vector, at room temperature for 4 h). In order to evaluate non-specific fluorescence arising either from autofluorescence or from nonspecific binding of antibodies or fluorochromes, some sections were processed as described above, but the respective primary antibodies were omitted. In some sections GABA was detected with Texas Red and parvalbumin with FITC. Other sections were processed for either parvalbumin or GABA alone. Tris-buffered saline, pH 7.4, which contained 1% normal horse serum and 0.5% Triton X-100, was used for washings and antiserum dilutions. After immunostaining, sections were washed with Trisbuffer, pH 7.4, mounted on slides with Gel/Mount TM (Bicmeda) and covered with a coverslip. 2.3. Analysis of the distribution of parvalbumin- and GABA-immunoreactive neurons The distribution of parvalbumin- and GABA-immunoreactive (IR) cell bodies in the entorhinal cortex was plotted with a computer-aided digitizing system (Minnesota Datametrics) with a 25 × objective. From the adja-
PV
2.4.1. Mirror technique The mirror technique [19] was used to study the colocalization of two different antigens, parvalbumin and GABA or parvalbumin and GAD, in the same cell. Camera lucida drawings, in which a 40 X objective was used, were made from the parvalbumin-IR cell bodies cut in half at the surface of the section. The surrounding capillaries were used as landmarks for locating the other half of the IR somata from the matching surface of the adjacent section (incubated for GABA or GAD). The drawn cells were re-examined with a 100 × oil immersion objective to ensure that they were at the surface of the section. In order to avoid false negative results, only those cells were included in the evaluation which had been bisected during sectioning and for which the other half, whether IR or not, could be identified unequivocally in the adjacent section. 2.4.2. Confocal microscopy The immunostained neurons of the entorhinal cortex were scanned on a Confocal Laser Scanning Microscope (Leica Lasertechnik, Germany) equipped with an argonkrypton laser mounted on a upright microscope (Leitz Diaplan). Confocal imaging was performed with a 25 × NPL oil immersion Fluotar lens with a numerical aperture of 1.30. Texas Red was excited by laser light (568 nm). An excitation beam splitter (580 nm short pass), 580 nm high pass detection beam splitter and a barrier high pass filter (590 nm) were used to detect emission. When the FITClabeled material was analyzed, laser light (488 nm), an excitation beam splitter (510 nm short pass), a 580 nm short pass detection beam splitter and a barrier band pass filter (535 nm) were used.
GABA
PERI
' ~" "1
.
J ~
•
i !-"X.
(\.~:..:\,
:'.;-./ .,
•
.
,
. ', ....~....~
..
.
•
A---
. . . . .
x
B
x
Fig. 2. Computer-generatedplots of sections showing the distribution of parvalbumin-containingneurons (A) and GABA-containingneurons (B). The border of the medial (MEA) and lateral (LEA) entorhinal cortex is indicated by an arrow. The layers of the entorhinal cortex are marked with Roman numerals (I-VI) and the laminadissecans is indicated by two dashed lines. Scale bar = 400 p,m; ab, angular bundle; PaS, parasubiculum;PERI, perirhinal area; PreS, presubiculum.
116
M. Miettinen et al. / Brain Research 706 (1996) 113-122
Table 1 Numbers of GABA- and GAD-positive parvalbumin (PV) immunoreactive neurons in the different layers of the medial and lateral entorhinal cortex (EC) Area
Confocal microscopy
Mirror technique
PV + GABA +
PV + GABA +
PV+ GAD+
0/0 75/75 77/77 50/50 12/12 214/214
0/0 69/69 103/103 91/91 16/16 279/279
0/0 5/5 15/15 12/12 1/1 33/33
Layer I Layer II Layer III Layers IV and V Layer VI Total
0/0 171/171 177/177 71/71 11/11 430/430
0/0 75/75 191/191 124/124 19/19 409/409
0/0 7/7 22/22 16/16 7/7 52/52
Total
644/644
688/688
85/85
Medial EC
Layer I Layer II Layer III Layers IV and V Layer VI Total Lateral EC
With the confocal microscope, the distribution of each labeled neuron could be assessed within a single confocal optical plane, which was approximately 0.5 /xm thick. Optical images were made in the XY plane at 0 . 5 - 1 . 0 / z m intervals for a total of 4-10 optical sections per scanning sequence using an image size of 512 X 256 or 256 X 256 pixels. Furthermore, to ensure scanning of identical confocal planes for FITC- and Texas Red-labeled cells, the double-stained sections were analyzed either simultaneously or sequentially so that emission from two different fluorescent labels were collected into two different channels at the same time or one after another, respectively. Photographs were obtained through a video printer (Sony Mavigraph, Japan).
3. Results 3.1. Nomenclature and subdivisions of the entorhinal cortex
In this study, the entorhinal cortex was subdivided into lateral and medial subdivisions according to cytoarchitectonic criteria. Each subdivision consists of six layers including cell-poor layer IV, lamina dissecans, separating superficial (I-III) and deep layers (V-VI) [1]. The cytoarchitecture of the entorhinal cortex, as seen in the Nissland AChE-stained sections, is presented in Fig. 1A and B, respectively.
3.2. Distribution of parvalbumin-immunoreactive neurons in the entorhinal cortex
The distribution of parvalbumin-IR cells in the entorhinal cortex was analyzed from the sections treated with Triton X-100. Antibodies against parvalbumin gave intensive staining of the soma and dendrites of nonpyramidal cells (Fig. 1C). The intensity of immunostaining of parvalbumin-IR cell bodies showed little variation in the different layers of the entorhinal cortex. In some areas, however, there was clear layer specific distribution with respect to staining for parvalbumin-IR. For example, there were more parvalbumin-IR puncta in the border of the perirhinal area and in layers II and III of the entorhinal cortex than elsewhere in the entorhinal cortex. Fig. 2A shows reconstructions of the parvalbumin-IR cells plotted from sections prepared for the mirror technique. In the present study, the distribution and morphology of parvalbumin-IR neurons were similar to those reported in earlier studies [24,34]. Therefore, only a brief description is given here. Parvalbumin-IR cell bodies were found in all layers of the entorhinal cortex, except in layer I. The majority of the parvalbumin-IR neurons were observed in layers II, III and VI. Parvalbumin-IR neurons were morphologically heterogeneous, including small and medium-size multipolar and fusiform cells. Neurons containing parvalbumin were oriented irregularly and extended their dendrites in various directions, especially into layers II and III of the entorhinal
Fig. 3. Light micrographs of matching surfaces of adjacent sections of the entorhinal cortex stained for parvalbumin (A and C) and for GABA (B and D). Parvalbumin-immunoreactive neurons bisected during sectioning (large arrows) were identified in the adjacent GABA-immunostained sections using nearby capillaries (C1_ 3) as landmarks. Scale bar = 10 /xm.
M. Miettinen et al. / Brain Research 706 (1996) 113-122
9 ¸ ~
117
118
M. Miettinenet aL/Brain Research 706 (1996) 113-122
cortex. However, parvalbumin-IR cell bodies in layers II and III usually extended their dendrites into layer I, towards the pial surface. In addition, especially in layer II, parvalbumin-IR varicosities outlined the unstained cell bodies. 3.3. Distribution of GABA- and GAD-immunoreactive neurons in the entorhinal cortex
The distribution of GABA- and GAD-IR neurons observed in the entorhinal cortex was identical to that described in previous studies [21]. Immunoreactivity to GAD and G A B A was found throughout the soma and the dendrites. Although GABAand GAD-IR cell bodies were found in all layers of the entorhinal cortex (Fig. 2B), the majority of the IR cells occurred in layer II and the superficial part of layer IlL Only a few GABA- and GAD-IR cells were present in layer I. The intensity of immunostaining of the GABA-IR cell bodies varied little in the different entorhinal layers. There were also cells whose staining intensity, although
weak, was still above the background level; therefore these cells could be characterized unequivocally as GABA-IR cells. 3.4. Analysis of colocalization of parvalbumin with GABA and GAD 3.4.1. Mirror technique Parvalbumin-containing cells were analyzed in nine horizontal sections (from two rats), which were located ventrally from the middle dorsoventral level of the entorhinal cortex ( - 5.8 to - 7 . 5 ventral to the pial surface). Analysis of 688 parvalbumin-IR soma showed that all parvalbuminIR neurons were GABA-IR (Table 1 and Figs. 3 and 5). The possible coexistence of parvalbumin with GAD was analyzed from one animal. In this analysis, we found that all 85 of the parvalbumin-IR perikarya analyzed were GAD-positive (Table 1 and Fig. 4). In sections analyzed by the mirror technique, 65% of all parvalbumin-IR neurons (n = 688) were located in layers II and III (Fig. 1C and 2A, Table 1). There were more parvalbumin-IR cells
Fig. 4. Light micrographs of matching surfaces of adjacent sections of the entorhinal cortex stained for parvalbumin (A and B) and for GAD (C and D). Nearby capiUaries (C1_3) were used as landmarks when the consecutive cell pairs marked with large arrows were identified. Scale bar = 10 p.m.
M. Miettinen et al. / Brain Research 706 (1996) 113-122
in the lateral (59%, n = 688) than in the medial (41%, n = 688) entorhinal cortex. 3.4.2. Confocal microscopic analysis
The confocal microscopic analysis consisted of colocalization of parvalbumin with GABA. Analysis of six sections revealed that all the parvalbumin-IR cell bodies contained GABA-immunoreactivity (Table 1, Fig. 5). With the confocal microscope we could take into account more parvalbumin-IR cells from an individual section of the entorhinal cortex than with the mirror technique. As a result, the number of parvalbumin-IR cells was found to be
119
higher when analyzed with the confocal microscopy than with the mirror technique. Layers II and III contained the majority (78% from 644 analyzed cells) of the parvalbumin-IR cells. Due to the limited penetration of GABA-antiserum, the fluorescence became weaker deeper in the section than at the surface of the specimen. Therefore, to reduce the number of false-negative parvalbumin cells, it was necessary to limit scanning to the surface of the specimen (up to 10 /~m from the surface). When the primary antiserum was omitted (control sections), there was no detectable staining. To estimate the proportion of GABAergic neurons con-
GABA
PV
A
B
Fig. 5. Confocalmicroscopeimagesof double-immunofluorescencestaining (A-C). GABA(A) and parvalbumin(B) were stained using antibodies labeled with FITC and Texas Red, respectively.When two datasets collected into different channels are combined,colocalizationis observedas yellow(C). Scale bar = 10 /~m.
120
M. Miettinen et al. / B r a i n Research 706 (1996) 113-122
taining parvalbumin in the entorhinal cortex, the GABAergic cells that did not contain parvalbumin-immunoreactivity were also counted during the colocalization experiment. According to these analysis, in both the medial and lateral entorhinal cortex 52% of the GABA-cells were also IR for parvalbumin. There was no marked difference between layers, except in layer III of both the medial and lateral entorhinal cortex, where 40% of the GABAergic cells contained parvalbumin-immunoreactivity.
4. Discussion In the present study, we examined the coexistence of parvalbumin-immunoreactivity and GABA in the entorhinal cortex by using the mirror technique at the light microscopic level and double immunofluorescence staining at the confocal microscopic level. The main findings can be summarized as follows: (i) results obtained by the two techniques were to a large extent identical, (ii) all the parvalbumin-containing neurons in the medial and lateral entorhinal cortex are GABAergic, (iii) parvalbumin-containing cells constitute about 50% of all GABAergic neurons in the entorhinal cortex.
4.1. Comments on the methods used The mirror technique is widely used in colocalization studies [19]. To study the possible colocalization of two antigens in the same cell, paired surfaces of neighboring sections immunostained with different antigens are compared with the aid of camera lucida drawings. Nearby capillaries are used as landmarks for an accurate localization of the IR cells of interest. After the cell pairs are evaluated, both immunopositive and immunonegative cell bodies are carefully verified with a 100 X oil immersion objective. In colocalization studies, this procedure is highly specific; but it is also excessively tedious, technically complex and time-consuming. Alternatively, the coexistence of two antigens can be analyzed from double-stained sections with a confocal microscope. The advantage of this method is that the possible coexistence can be analyzed from the same individual section and the same cell body by collecting images from two dyes into two different channels and combining the two data sets into one data package, where coexistence can easily be recognized as a distinct color in the same field of view. Therefore, analysis at the confocal microscopic level is easy and accurate and provides a rapid and precise means of evaluating the colocalization. In addition, since confocal microscope images have high resolution, even the cells that are very near to each other can easily be verified. This is probably the reason why in this study we found more parvalbumin-IR cells from one individual section, especially from layers II and III, with the confocal microscope than with the mirror technique. In fact, with
the mirror technique several cells had to be rejected since we could analyze only neurons that were bisected during sectioning; and cell pairs were included in our results only when the other half could be reliably identified from the adjacent section.
4.2. GABAergic cells containing parvalbumin Neurons containing parvalbumin represent a subpopulation of GABAergic neurons in many areas of the rat brain, such as the cerebral cortex [5,12], olfactory bulb [18] and hippocampus [17,30]. The proportion of GABA neurons that contain parvalbumin varies in different areas of the neocortex. In the rat somatosensory cortex, about 70% of the GABAergic neurons are IR for parvalbumin [5]. In the monkey striate cortex, 70-75% of the GABA-neurons contain parvalbumin. The relative proportion of these neurons, however, varies in different layers [32]. In layer 2, 30% of the GABAergic neurons contain parvalbumin and the proportion of such neurons gradually increases to almost 100% in layer 4C. In some cortical areas as well as in the dentate gyms, however, only a small proportion of GABAergic neurons contain parvalbumin [7,17,30]. For example, in the hippocampal formation only about 20% of the GABA neurons contain parvalbumin [17]. Like in the cerebral cortex, also in the hippocampus the degree of coexistence depends on the layer of the hippocampus. In the stratum pyramidale about 50% of the GABAergic neurons contain parvalbumin; and in the dentate granule cell layer and in the stratum oriens of the CA3 and CA1 subfields, 20-30% of the GABAergic cells contain parvalbumin, whereas less than 5% of the GABAergic neurons in the stratum radiatum and stratum lacunosum-moleculare are also IR for parvalbumin [17]. In the present study, we found that about 50% of all the GABAergic neurons in the entorhinal cortex contain parvalbumin. This is less than the amount reported for the somatosensory cortex [5] but higher than in the hippocampal formation [17]. Thus, neurons containing parvalbumin in the entorhinal cortex constitute a significant population of GABAergic inhibitory neurons. Parvalbumin is known to be localized mainly within different types of nonpyramidal cells in the cerebral cortex [3,5,6,13,22]. In the entorhinal cortex, parvalbumin is also present in nonpyramidal cells that are morphologically heterogeneous. The prominent feature of these neurons in the entorhinal cortex is that they extend their dendrites in various directions. The dendrites of the parvalbumin-containing neurons in the deeper layers sometimes reach superficial layers II and III, and parvalbumin-containing neurons in layers II and III often extend their dendrites into layer I [24,34]. However, the dendrites of the parvalbumin-containing cells in layer III are largely confined within the same layer. Parvalbumin-containing cells are thus in a position to receive information from many layers of the entorhinal cortex as well as from different brain
M. Miettinen et al. / Brain Research 706 (1996) 113-122
areas which innervate different areas of the entorhinal cortex. On the other hand, since the parvalbumin-containing neurons in the entorhinal cortex have morphological features of the chandelier (a subpopulation of axo-axonic cells) and basket cells [23,2,6,28,29], they are likely to be the GABAergic cell population that effectively controls the output of the entorhinal cortex. This agrees with the findings that parvalbumin-positive axon terminals make synaptic contacts with the dendrites and soma of both parvalbumin-immunonegative and parvalbumin-immunopositive neurons in layers II, III and IV (layer V according to our criteria) of the entorhinal cortex [34]. In addition, electrophysiological studies [15,16] have shown that basket-like interneurons in layer II of the entorhinal cortex exert widespread and strong inhibitory control over neurons that project into or from the hippocampus. Taking these lines of evidence together with our finding that 50% of the GABA cells in the entorhinal cortex contain parvalbumin, parvalbumin-cells in the entorhinal cortex may represent an important cell population that can modulate both the input and the output of the pyramidal cells, which relay information between the entorhinal cortex and other cortical areas as well as between the entorhinal cortex and the hippocampus. Therefore, any changes in parvalbumin-containing cells in the entorhinal cortex, like those we observed in aged memory-impaired rats [24], can cause disturbances in GABA-mediated inhibition and may thus interfere with normal functioning of the entorhinal cortex.
Acknowledgements We thank Prof. I. Virtanen (University of Helsinki, Helsinki, Finland) for the gift of antiserum against GABA and Dr. A. Pitk~inen (University of Kuopio, Kuopio, Finland) for her helpful comments on the manuscript. This research was supported by the Finnish Academy of Sciences.
References [1] Amaral, D.G. and Insausti, R., Hippocampal formation. In G. Paxinos (Ed.), The Human Nervous System, Academic Press, San Diego, 1990, pp. 711-755. [2] Arnold, S.E., Hyman, B.T., Van Hoesen, G.W. and Damasio, A.R., Some cytoarchitectural abnormalities of the entorhinal cortex in schizophrenia, Arch. Gen. Psychiatry, 48 (1991) 625-632. [3] Bliimcke, I., Hof, P.R., Morrison, J.H. and Cello, M.R., Distribution of parvalbumin immunoreactivity in the visual cortex of old world monkeys and humans, J. Comp. Neurol., 301 (1990) 417-432. [4] Braak, H. and Braak, E., Nettropathological stageing of Alzheimerrelated changes, Acta Neuropathol., 82 (1991) 239-259. [5] Cello, M.R., Parvalbumin in most 3,-aminobutyric acid-containing neurons of the rat cerebral cortex, Science, 231 (1986) 995-997. [6] Cello, M.R., Calbindin D-28k and parvalbumin in the rat nervous system, Neuroscience, 35 (1990) 375-475.
121
[7] 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. [8] Du, F., Whetsell Jr., W.O., Abou-Khalil, B., Blumenkopf, B., Lothman, E.W. and Schwarcz, R., Preferential neuronal loss in layer III of the entorhinal cortex in patients with temporal lobe epilepsy, Epilepsy Res., 16 (1993) 223-233. [9] Freund, T.F., GABAergic septohippocampal neurons contain parvalbumin, Brain Res., 478 (1989) 375-381. [10] Gottfries, C.G., Neurochemical aspects on aging and diseases with cognitive impairment, J. Neurosci. Res., 27 (1990) 541-547. [11] Hedreen, J.C., Bacon, S.J. and Price, D.L., A modified histochemical technique to visualize acetylcholinesterase-containing axons, J. Histochem. Cytochem., 33 (1985) 134-140. [12] Heizmann, C.W. and Cello, M.R., Immunolocalization of parvalbumin, Meth. EnzymoL, 139 (1987) 552-570. [13] Hendry, S.H.C., Jones, E.G., Emson, P.C., Lawson, D.E.M., Heizmann, C.W. and Streit, P., Two classes of cortical GABA neurons defined by differential calcium binding protein immunoreactivities, Exp. Brain Res., 76 (1989) 467-472. [14] Izquierdo, I. and Medina, J.H., Correlation between the pharmacology of long-term potentiation and the pharmacology of memory, NeurobioL Learn. Memory, 63 (1995) 19-32. [15] Jones, R.S.G., Entorhinal-hippocampal connections: a speculative view of their function, Trends Neurosci., 16 (1993) 58-64. [16] Jones, R.S.G. and Biihl, E.H., Basket-like interneurones in layer II of the entorhinal cortex exhibit a powerful NMDA-mediated synaptic excitation, Neurosci. Lett., 149 (1993) 35-39. [17] Kosaka, T., Katsumaru, H., Hama, K., Wu, J.-Y. and Heizmann, C.W., GABAergic neurons containing the Ca2÷-binding protein parvalbumin in the rat hippocampus and dentate gyms, Brain Res., 419 (1987) 119-130. [18] Kosaka, T., Kosaka, K., Heizmann, C.W., Nagatsu, I., Wu, J.-Y., Yanaihara, N. and Hama, K., An aspect of the organization of the GABAergic system in the rat main olfactory bulb: laminar distribution of immunohistochemically defined subpopulations of GABAergic neurons, Brain Res., 411 (1987) 373-378. [19] Kosaka, T., Kosaka, K., Tateishi, K., Hamaoka, Y., Yanaihara, N., Wu, J.-Y. and Hama, K., GABAergic neurons containing CCK-8-1ike and/or VIP-like immunoreactivities in the rat hippocampus and dentate gyms, J. Comp. Neurol., 239 (1985) 420-430. [20] Kosaka, T., Nagatsu, I., Wu, J.-Y. and Hama, K., Use of high concentrations of glutaraldehyde for immunocytochemistry of transmitter-synthesizing enzymes in the central nervous system, Neuroscience, 18 (1986) 975-990. [21] K6hler, C., Wu, J.-Y. and Chan-Palay, V., Neurons and terminals in the retrohippocampal region in the rat's brain identified by anti-3,aminobutyric acid and anti-giutamic acid decarboxylase immunocytochemistry, Anat. Embryol., 173 (1985) 35-44. [22] Lewis, D.A. and Lund, J.S., Heterogeneity of chandelier neurons in monkey neocortex: corticotropin-releasing factor- and parvalbuminimmunoreactive populations, J. Comp. Neurol., 293 (1990) 599-615. [23] Lorente de N6, R., Studies on the structure of the cerebral cortex II. Continuation of the study of the Ammonic system, J. Psychol. Neurol., 46 (1934) 113-177. [24] Miettinen, R., Sirvi6, J., Riekkinen Sr., P., Laakso, M. P., Riekkinen, M. and Riekkinen Jr., P., Neocortical, hippocampal and septal parvalbumin- and somatostatin-containing neurons in young and aged rats: correlation with passive avoidance and water maze performance, Neuroscience, 53 (1993) 367-378. [25] Nitsch, R., Soriano, E. and Frotscher, M., The parvalbumin-containing nonpyramidal neurons in the rat hippocampus, Anat. Embryol., 181 (1990) 413-425. [26] Ribak, C.E. and Seress, L., Five types of basket cell in the hippocampal dentate gyms: a combined Golgi and electron microscopic study, J. Neurocytol., 12 (1983) 577-597. [27] Reinikainen, K.J., Palj~irvi, L., Huuskonen, M., Soininen, H., Laakso,
122
M. Miettinen et al. / Brain Research 706 (1996) 113-122
M. and Riekkinen, P.J., A post-mortem study of noradrenergic, serotonergic and GABAergic neurons in Alzheimer's disease, J. Neurol. Sci., 84 (1988) 101-116. [28] Somogyi, P., Smith, A.D., Nunzi, M.G., Gorio, A., Takagi, H. and Wu, J.Y., Glutamate decarboxylase immunoreactivity in the hippocampus of the cat: distribution of immunoreactive synaptic terminals with special reference to the axon initial segment of pyramidal neurons, J. Neurosci., 3 (1983) 1450-1468. [29] Soriano, E. and Frotscher, M., A GABAergic axo-axonic cell in the fascia dentata controls the main excitatory hippocampal pathway, Brain Res., 503 (1989) 170-174. [30] Soriano, E., Nitsch, R. and Frotscher, M., Axo-axonic chandelier cells in the rat fascia dentata: Golgi-electron microscopy and immunocytochemical studies, J. Comp. Neurol., 293 (1990) 1-25.
[31] Szabat, E., Soinila, S., H~ipp61~i, O., Linnala, A. and Virtanen, I., A new monoclonal antibody against the GABA-protein conjugate shows immunoreactivity in sensory neurons of the rat, Neuroscience, 47 (1992) 409-420. [32] 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. [33] Witter, M.P., Organization of the entorhinal-hippocampat system: a review of current anatomical data, Hippocampus, 3 (1993) 33-44. [34] Wouterlood, F.G., H~irtig, W., Briickner, G. and WiRer, M.P., Parvalbumin-immunoreactive neurons in the entorhinal cortex of the rat: localization, morphology, connectivity and ultrastructure, J. Nearocytol., 24 (1995) 135-153.
Brain Research 706 (1996) 113-122
Research report
Coexistence of parvalbumin and GABA in nonpyramidal neurons of the rat entorhinal cortex Maria Miettinen, Esa Koivisto, Paavo Riekkinen Sr., Riitta Miettinen * Department of Neurology and A.L Virtanen Institute, University of Kuopio, PO Box 1627, FIN-70211 Kuopio, Finland Accepted 12 September 1995
Abstract The possible coexistence of the calcium-binding protein, parvalbumin, with the major inhibitory neurotransmitter, gamma-aminobutyric acid (GABA), and its synthesizing enzyme, glutamate decarboxylase (GAD), was studied in nonpyramidal cells of the rat medial and lateral entorhinal cortex. The material was analyzed by two different methods, the first of which was a mirror technique where the possible coexistence of two different antigens was analyzed from cells cut in half at the surface of the adjacent section. The other method consisted of analyzing double immunofluorescent-stained sections with a confocal microscope. The colocalization analysis revealed that all parvalbumin-immunoreactive neurons (mirror technique n = 688 and confocal microscopy n = 644) in all layers of the medial and lateral entorhinal cortex were also immunopositive for GABA or GAD. Parvalbumin-cells made up 52% of the GABA cells in most of the layers in the medial and lateral entorhinal cortex. In layer III of the entorhinal cortex, the proportion was about 40%. Thus, parvalbumin-containing neurons in the entorhinal cortex represent a large GABAergic cell population, which is likely to play an important role in controlling both the input and the output of the entorhinal cortex. Keywords: Calcium-binding protein; Confocal microscopy; Glutamate decarboxylase; Hippocampal formation; Immunocytochemistry; Colocalization
1. Introduction The entorhinal cortex occupies a key position in the neuronal circuitry between the hippocampus and various cortical and subcortical areas. Giving rise to the perforant path, which is the main cortical pathway to the hippocampus, the entorhinal cortex provides the hippocampal formation with most of the cortical sensory information [33]. The spiny stellate and pyramidal cells in layers II and III of the entorhinal cortex are the main source of input into the dentate gyms. In addition, there are projections from layer II to the CA1 subfield; but this layer may also innervate CA2 and CA3 regions and the subiculum. Layer III also projects to the CA1 and CA3 subfields. The deeper layers (IV-VI) are targets for input originating from the CA1 field of the hippocampus and from the subiculum and project back to the neocortex and subcortical areas. Thus, the entorhinal cortex is an important structure not only for getting information to the hippocampus but also for receiv-
* Corresponding author. Dept. of Neurology, Univ. of Kuopio, PO Box 1627, FIN-70211 Kuopio, Finland. Fax: (358) (71) 162048. Email: [email protected] 0006-8993/96/$15.00 © 1996 Elsevier Science B.V. All rights reserved SSDI 0 0 0 6 - 8 9 9 3 ( 9 5 ) 0 1 2 0 3 - 6
ing information from the hippocampus [1,16]. Recent studies have suggested that the entorhinal cortex plays a role in learning by integrating information and memories which have previously been processed by the hippocampus, amygdala or medial septum [14]. In several disorders of the central nervous system, the entorhinal cortex is known to be vulnerable. These disorders include for example, Alzheimer's disease, schizophrenia and epilepsy [2,4,8], all of which are accompanied by cognitive impairment. Alterations in cholinergic, noradrenergic, serotonergic and dopaminergic transmission as well as in the major inhibitory neurotransmission, gammaaminobutyric acid (GABA), may participate in impairment of learning and memory [10,27]. Recently, our group found that the number of neurons containing a calcium-binding protein, parvalbumin, in the entorhinal cortex of aged rats was smaller than in young rats [24]. This decline correlated with the learning and memory deficits observed in aged rats. Since parvalbumin is present in different types of GABAergic nonpyramidal neurons in several brain areas [5-7,9,12,13,17,18,25,30], our finding suggested that degeneration of the inhibitory system is involved in cognitive dysfunctions associated with aging. However, since there was no direct evidence that the
114
M. Miettinen et al. / Brain Research 706 (1996) 113-122
parvalbumin-neurons in the entorhinal cortex are GABAergic, our finding remained partly equivocal. Therefore, in the present study we investigated whether cells in the entorhinal cortex that contain parvalbumin do, in fact, contain GABA or its synthesizing enzyme, GAD.
using 3,3'diaminobenzidine as a chromogen. All the washes and antibody dilutions were done in 0.05M Tris-buffered saline, pH 7.4, which contained 1% normal serum. When the sections were to be used for cell-plotting, Triton X-100 was included in the solutions. The sections for the cell-
2. Materials and methods
2.1. Tissue preparation Wistar rats of 9 to 10 weeks old, (National Animal Center at the University of Kuopio, Finland; wt: 200-300 g) were used in this study. Deeply anaesthetized (chlornembutal, 0.3 ml/100 g, i.p.) rats were perfused transcardially with physiological saline (3 min) and then with 300 ml of fixative for 30 min. For GAD-immunohistochemistry, the fixative contained 4% paraformaldehyde, 0.05% glutaraldehyde, 0.2% picric acid in 0.1 M phosphate buffer (PB), pH 7.4. For GABA-immunohistochemistry, the fixative contained 2.5% paraformaldehyde, 1% glutaraldehyde and 0.2% picric acid in PB. Both fixatives were used for parvalbumin-immunohistochemistry. Before immunostaining, the material from the rats perfused with 1% glutaraldehyde was treated with 1% sodium borohydride for 30 min in order to reduce free aldehyde groups and double bonds, thus improving the immunoreactivity of protein antigens [20]. The brains were sectioned horizontally on a Vibratome at 60 /xm (1:5 series) and the sequential order of the sections was preserved. Cytoarchitectonic borders for precise localization of cortical areas and laminar borders in the entorhinal cortex were obtained from the adjacent Nissl- and acetylcholinesterase-stained (ACHE) sections. A modification of the method of Karnovsky-Roots [11] was used for AChE-staining.
B
2.2. Immunocytochemistry 2.2.1. Immunoperoxidase staining For analysis of the coexistence of parvalbumin with GABA or GAD analyzed with the mirror technique and for cell-plotting, free-floating sections were immunostained with the avidin-biotin method. After extensive washings in PB, the sections were incubated first in 10% normal goat or horse serum (depending on the origin of the secondary antiserum) and then at 4°C for 2 days in the primary antiserum: rabbit anti-parvalbumin (1:8000, PV-28, Swant) or mouse anti-parvalbumin (1:8000, PV-Mab, Swant). Adjacent sections were incubated either in rabbit anti-GAD (1:200, Chemicon) or mouse anti-GABA (1:100) [31]. These sections were incubated in the secondary antibody (1:50, biotinylated goat anti-rabbit IgG for rabbit primaries and biotinylated horse anti-mouse IgG for mouse primaries, Vector) at 4°C for 16 h. The third layer was avidin-biotinylated horseradish peroxidase complex (1:100, Vector). The immunoperoxidase reaction was developed
C Fig. 1. Light micrographs of horizontal sections from the entorhinal cortex of the rat. Nissl-staining (A) and acetylcholinesterase-staining (B) show the cytoarchitectonics of the ¢ntorhinal cortex. Distribution of parvalbumin-immunoreactivity is demonstrated in C. The border of the medial (MEA) and lateral (LEA) entorhinal cortex is indicated by an arrow. The layers of the entorhinal cortex are marked with Roman numerals (I-VI) and the lamina dissecans is indicated by asterisks. Scale bar = 500/zm; ab, angular bundle; PaS, parasubiculum; PERI, perirhinai area; PreS, presubiculum.
M. Miettinenet aL/ BrainResearch 706 (1996) 113-122 plotting were mounted on slides with Depex (BDH Laboratory Supplies, Gurr R) and covered with a coverslip. The sections for the mirror technique were treated with 0.5% OsO 4 for 30 min, dehydrated in ethanol and propylene oxide, and embedded in Durcupan (ACM, Fluka) on slides.
115
cent Nissl-stained sections the boundaries of the entorhinal cortex and its layers were drawn with camera lucida. Thereafter, the outlines were superimposed on scanned plots by using Canvas software on a Macintosh computer.
2.4. Analysis of colocalization 2.2.2. Immunofluorescence staining Double immunofluorescent staining for parvalbumin and GABA was used to analyze the possible coexistence of these two antigens with the confocal microscope. Freefloating sections were incubated in rabbit anti-parvalbumin (1:4000, PV-28, Swant) together with mouse anti-GABA (1:50) at 4°C for 2 days. The sections were then incubated in a mixture of Texas Red-labeled goat anti-rabbit (1:50, Vector) and biotinylated horse anti-mouse (1:50, Vector) at 4°C for 16 h. After that the biotinylated goat anti-rabbit antiserum was detected by fluorescein isothiocyanatelabeled avidin (FITC-avidin, 1:100, Vector, at room temperature for 4 h). In order to evaluate non-specific fluorescence arising either from autofluorescence or from nonspecific binding of antibodies or fluorochromes, some sections were processed as described above, but the respective primary antibodies were omitted. In some sections GABA was detected with Texas Red and parvalbumin with FITC. Other sections were processed for either parvalbumin or GABA alone. Tris-buffered saline, pH 7.4, which contained 1% normal horse serum and 0.5% Triton X-100, was used for washings and antiserum dilutions. After immunostaining, sections were washed with Trisbuffer, pH 7.4, mounted on slides with Gel/Mount TM (Bicmeda) and covered with a coverslip. 2.3. Analysis of the distribution of parvalbumin- and GABA-immunoreactive neurons The distribution of parvalbumin- and GABA-immunoreactive (IR) cell bodies in the entorhinal cortex was plotted with a computer-aided digitizing system (Minnesota Datametrics) with a 25 × objective. From the adja-
PV
2.4.1. Mirror technique The mirror technique [19] was used to study the colocalization of two different antigens, parvalbumin and GABA or parvalbumin and GAD, in the same cell. Camera lucida drawings, in which a 40 X objective was used, were made from the parvalbumin-IR cell bodies cut in half at the surface of the section. The surrounding capillaries were used as landmarks for locating the other half of the IR somata from the matching surface of the adjacent section (incubated for GABA or GAD). The drawn cells were re-examined with a 100 × oil immersion objective to ensure that they were at the surface of the section. In order to avoid false negative results, only those cells were included in the evaluation which had been bisected during sectioning and for which the other half, whether IR or not, could be identified unequivocally in the adjacent section. 2.4.2. Confocal microscopy The immunostained neurons of the entorhinal cortex were scanned on a Confocal Laser Scanning Microscope (Leica Lasertechnik, Germany) equipped with an argonkrypton laser mounted on a upright microscope (Leitz Diaplan). Confocal imaging was performed with a 25 × NPL oil immersion Fluotar lens with a numerical aperture of 1.30. Texas Red was excited by laser light (568 nm). An excitation beam splitter (580 nm short pass), 580 nm high pass detection beam splitter and a barrier high pass filter (590 nm) were used to detect emission. When the FITClabeled material was analyzed, laser light (488 nm), an excitation beam splitter (510 nm short pass), a 580 nm short pass detection beam splitter and a barrier band pass filter (535 nm) were used.
GABA
PERI
' ~" "1
.
J ~
•
i !-"X.
(\.~:..:\,
:'.;-./ .,
•
.
,
. ', ....~....~
..
.
•
A---
. . . . .
x
B
x
Fig. 2. Computer-generatedplots of sections showing the distribution of parvalbumin-containingneurons (A) and GABA-containingneurons (B). The border of the medial (MEA) and lateral (LEA) entorhinal cortex is indicated by an arrow. The layers of the entorhinal cortex are marked with Roman numerals (I-VI) and the laminadissecans is indicated by two dashed lines. Scale bar = 400 p,m; ab, angular bundle; PaS, parasubiculum;PERI, perirhinal area; PreS, presubiculum.
116
M. Miettinen et al. / Brain Research 706 (1996) 113-122
Table 1 Numbers of GABA- and GAD-positive parvalbumin (PV) immunoreactive neurons in the different layers of the medial and lateral entorhinal cortex (EC) Area
Confocal microscopy
Mirror technique
PV + GABA +
PV + GABA +
PV+ GAD+
0/0 75/75 77/77 50/50 12/12 214/214
0/0 69/69 103/103 91/91 16/16 279/279
0/0 5/5 15/15 12/12 1/1 33/33
Layer I Layer II Layer III Layers IV and V Layer VI Total
0/0 171/171 177/177 71/71 11/11 430/430
0/0 75/75 191/191 124/124 19/19 409/409
0/0 7/7 22/22 16/16 7/7 52/52
Total
644/644
688/688
85/85
Medial EC
Layer I Layer II Layer III Layers IV and V Layer VI Total Lateral EC
With the confocal microscope, the distribution of each labeled neuron could be assessed within a single confocal optical plane, which was approximately 0.5 /xm thick. Optical images were made in the XY plane at 0 . 5 - 1 . 0 / z m intervals for a total of 4-10 optical sections per scanning sequence using an image size of 512 X 256 or 256 X 256 pixels. Furthermore, to ensure scanning of identical confocal planes for FITC- and Texas Red-labeled cells, the double-stained sections were analyzed either simultaneously or sequentially so that emission from two different fluorescent labels were collected into two different channels at the same time or one after another, respectively. Photographs were obtained through a video printer (Sony Mavigraph, Japan).
3. Results 3.1. Nomenclature and subdivisions of the entorhinal cortex
In this study, the entorhinal cortex was subdivided into lateral and medial subdivisions according to cytoarchitectonic criteria. Each subdivision consists of six layers including cell-poor layer IV, lamina dissecans, separating superficial (I-III) and deep layers (V-VI) [1]. The cytoarchitecture of the entorhinal cortex, as seen in the Nissland AChE-stained sections, is presented in Fig. 1A and B, respectively.
3.2. Distribution of parvalbumin-immunoreactive neurons in the entorhinal cortex
The distribution of parvalbumin-IR cells in the entorhinal cortex was analyzed from the sections treated with Triton X-100. Antibodies against parvalbumin gave intensive staining of the soma and dendrites of nonpyramidal cells (Fig. 1C). The intensity of immunostaining of parvalbumin-IR cell bodies showed little variation in the different layers of the entorhinal cortex. In some areas, however, there was clear layer specific distribution with respect to staining for parvalbumin-IR. For example, there were more parvalbumin-IR puncta in the border of the perirhinal area and in layers II and III of the entorhinal cortex than elsewhere in the entorhinal cortex. Fig. 2A shows reconstructions of the parvalbumin-IR cells plotted from sections prepared for the mirror technique. In the present study, the distribution and morphology of parvalbumin-IR neurons were similar to those reported in earlier studies [24,34]. Therefore, only a brief description is given here. Parvalbumin-IR cell bodies were found in all layers of the entorhinal cortex, except in layer I. The majority of the parvalbumin-IR neurons were observed in layers II, III and VI. Parvalbumin-IR neurons were morphologically heterogeneous, including small and medium-size multipolar and fusiform cells. Neurons containing parvalbumin were oriented irregularly and extended their dendrites in various directions, especially into layers II and III of the entorhinal
Fig. 3. Light micrographs of matching surfaces of adjacent sections of the entorhinal cortex stained for parvalbumin (A and C) and for GABA (B and D). Parvalbumin-immunoreactive neurons bisected during sectioning (large arrows) were identified in the adjacent GABA-immunostained sections using nearby capillaries (C1_ 3) as landmarks. Scale bar = 10 /xm.
M. Miettinen et al. / Brain Research 706 (1996) 113-122
9 ¸ ~
117
118
M. Miettinenet aL/Brain Research 706 (1996) 113-122
cortex. However, parvalbumin-IR cell bodies in layers II and III usually extended their dendrites into layer I, towards the pial surface. In addition, especially in layer II, parvalbumin-IR varicosities outlined the unstained cell bodies. 3.3. Distribution of GABA- and GAD-immunoreactive neurons in the entorhinal cortex
The distribution of GABA- and GAD-IR neurons observed in the entorhinal cortex was identical to that described in previous studies [21]. Immunoreactivity to GAD and G A B A was found throughout the soma and the dendrites. Although GABAand GAD-IR cell bodies were found in all layers of the entorhinal cortex (Fig. 2B), the majority of the IR cells occurred in layer II and the superficial part of layer IlL Only a few GABA- and GAD-IR cells were present in layer I. The intensity of immunostaining of the GABA-IR cell bodies varied little in the different entorhinal layers. There were also cells whose staining intensity, although
weak, was still above the background level; therefore these cells could be characterized unequivocally as GABA-IR cells. 3.4. Analysis of colocalization of parvalbumin with GABA and GAD 3.4.1. Mirror technique Parvalbumin-containing cells were analyzed in nine horizontal sections (from two rats), which were located ventrally from the middle dorsoventral level of the entorhinal cortex ( - 5.8 to - 7 . 5 ventral to the pial surface). Analysis of 688 parvalbumin-IR soma showed that all parvalbuminIR neurons were GABA-IR (Table 1 and Figs. 3 and 5). The possible coexistence of parvalbumin with GAD was analyzed from one animal. In this analysis, we found that all 85 of the parvalbumin-IR perikarya analyzed were GAD-positive (Table 1 and Fig. 4). In sections analyzed by the mirror technique, 65% of all parvalbumin-IR neurons (n = 688) were located in layers II and III (Fig. 1C and 2A, Table 1). There were more parvalbumin-IR cells
Fig. 4. Light micrographs of matching surfaces of adjacent sections of the entorhinal cortex stained for parvalbumin (A and B) and for GAD (C and D). Nearby capiUaries (C1_3) were used as landmarks when the consecutive cell pairs marked with large arrows were identified. Scale bar = 10 p.m.
M. Miettinen et al. / Brain Research 706 (1996) 113-122
in the lateral (59%, n = 688) than in the medial (41%, n = 688) entorhinal cortex. 3.4.2. Confocal microscopic analysis
The confocal microscopic analysis consisted of colocalization of parvalbumin with GABA. Analysis of six sections revealed that all the parvalbumin-IR cell bodies contained GABA-immunoreactivity (Table 1, Fig. 5). With the confocal microscope we could take into account more parvalbumin-IR cells from an individual section of the entorhinal cortex than with the mirror technique. As a result, the number of parvalbumin-IR cells was found to be
119
higher when analyzed with the confocal microscopy than with the mirror technique. Layers II and III contained the majority (78% from 644 analyzed cells) of the parvalbumin-IR cells. Due to the limited penetration of GABA-antiserum, the fluorescence became weaker deeper in the section than at the surface of the specimen. Therefore, to reduce the number of false-negative parvalbumin cells, it was necessary to limit scanning to the surface of the specimen (up to 10 /~m from the surface). When the primary antiserum was omitted (control sections), there was no detectable staining. To estimate the proportion of GABAergic neurons con-
GABA
PV
A
B
Fig. 5. Confocalmicroscopeimagesof double-immunofluorescencestaining (A-C). GABA(A) and parvalbumin(B) were stained using antibodies labeled with FITC and Texas Red, respectively.When two datasets collected into different channels are combined,colocalizationis observedas yellow(C). Scale bar = 10 /~m.
120
M. Miettinen et al. / B r a i n Research 706 (1996) 113-122
taining parvalbumin in the entorhinal cortex, the GABAergic cells that did not contain parvalbumin-immunoreactivity were also counted during the colocalization experiment. According to these analysis, in both the medial and lateral entorhinal cortex 52% of the GABA-cells were also IR for parvalbumin. There was no marked difference between layers, except in layer III of both the medial and lateral entorhinal cortex, where 40% of the GABAergic cells contained parvalbumin-immunoreactivity.
4. Discussion In the present study, we examined the coexistence of parvalbumin-immunoreactivity and GABA in the entorhinal cortex by using the mirror technique at the light microscopic level and double immunofluorescence staining at the confocal microscopic level. The main findings can be summarized as follows: (i) results obtained by the two techniques were to a large extent identical, (ii) all the parvalbumin-containing neurons in the medial and lateral entorhinal cortex are GABAergic, (iii) parvalbumin-containing cells constitute about 50% of all GABAergic neurons in the entorhinal cortex.
4.1. Comments on the methods used The mirror technique is widely used in colocalization studies [19]. To study the possible colocalization of two antigens in the same cell, paired surfaces of neighboring sections immunostained with different antigens are compared with the aid of camera lucida drawings. Nearby capillaries are used as landmarks for an accurate localization of the IR cells of interest. After the cell pairs are evaluated, both immunopositive and immunonegative cell bodies are carefully verified with a 100 X oil immersion objective. In colocalization studies, this procedure is highly specific; but it is also excessively tedious, technically complex and time-consuming. Alternatively, the coexistence of two antigens can be analyzed from double-stained sections with a confocal microscope. The advantage of this method is that the possible coexistence can be analyzed from the same individual section and the same cell body by collecting images from two dyes into two different channels and combining the two data sets into one data package, where coexistence can easily be recognized as a distinct color in the same field of view. Therefore, analysis at the confocal microscopic level is easy and accurate and provides a rapid and precise means of evaluating the colocalization. In addition, since confocal microscope images have high resolution, even the cells that are very near to each other can easily be verified. This is probably the reason why in this study we found more parvalbumin-IR cells from one individual section, especially from layers II and III, with the confocal microscope than with the mirror technique. In fact, with
the mirror technique several cells had to be rejected since we could analyze only neurons that were bisected during sectioning; and cell pairs were included in our results only when the other half could be reliably identified from the adjacent section.
4.2. GABAergic cells containing parvalbumin Neurons containing parvalbumin represent a subpopulation of GABAergic neurons in many areas of the rat brain, such as the cerebral cortex [5,12], olfactory bulb [18] and hippocampus [17,30]. The proportion of GABA neurons that contain parvalbumin varies in different areas of the neocortex. In the rat somatosensory cortex, about 70% of the GABAergic neurons are IR for parvalbumin [5]. In the monkey striate cortex, 70-75% of the GABA-neurons contain parvalbumin. The relative proportion of these neurons, however, varies in different layers [32]. In layer 2, 30% of the GABAergic neurons contain parvalbumin and the proportion of such neurons gradually increases to almost 100% in layer 4C. In some cortical areas as well as in the dentate gyms, however, only a small proportion of GABAergic neurons contain parvalbumin [7,17,30]. For example, in the hippocampal formation only about 20% of the GABA neurons contain parvalbumin [17]. Like in the cerebral cortex, also in the hippocampus the degree of coexistence depends on the layer of the hippocampus. In the stratum pyramidale about 50% of the GABAergic neurons contain parvalbumin; and in the dentate granule cell layer and in the stratum oriens of the CA3 and CA1 subfields, 20-30% of the GABAergic cells contain parvalbumin, whereas less than 5% of the GABAergic neurons in the stratum radiatum and stratum lacunosum-moleculare are also IR for parvalbumin [17]. In the present study, we found that about 50% of all the GABAergic neurons in the entorhinal cortex contain parvalbumin. This is less than the amount reported for the somatosensory cortex [5] but higher than in the hippocampal formation [17]. Thus, neurons containing parvalbumin in the entorhinal cortex constitute a significant population of GABAergic inhibitory neurons. Parvalbumin is known to be localized mainly within different types of nonpyramidal cells in the cerebral cortex [3,5,6,13,22]. In the entorhinal cortex, parvalbumin is also present in nonpyramidal cells that are morphologically heterogeneous. The prominent feature of these neurons in the entorhinal cortex is that they extend their dendrites in various directions. The dendrites of the parvalbumin-containing neurons in the deeper layers sometimes reach superficial layers II and III, and parvalbumin-containing neurons in layers II and III often extend their dendrites into layer I [24,34]. However, the dendrites of the parvalbumin-containing cells in layer III are largely confined within the same layer. Parvalbumin-containing cells are thus in a position to receive information from many layers of the entorhinal cortex as well as from different brain
M. Miettinen et al. / Brain Research 706 (1996) 113-122
areas which innervate different areas of the entorhinal cortex. On the other hand, since the parvalbumin-containing neurons in the entorhinal cortex have morphological features of the chandelier (a subpopulation of axo-axonic cells) and basket cells [23,2,6,28,29], they are likely to be the GABAergic cell population that effectively controls the output of the entorhinal cortex. This agrees with the findings that parvalbumin-positive axon terminals make synaptic contacts with the dendrites and soma of both parvalbumin-immunonegative and parvalbumin-immunopositive neurons in layers II, III and IV (layer V according to our criteria) of the entorhinal cortex [34]. In addition, electrophysiological studies [15,16] have shown that basket-like interneurons in layer II of the entorhinal cortex exert widespread and strong inhibitory control over neurons that project into or from the hippocampus. Taking these lines of evidence together with our finding that 50% of the GABA cells in the entorhinal cortex contain parvalbumin, parvalbumin-cells in the entorhinal cortex may represent an important cell population that can modulate both the input and the output of the pyramidal cells, which relay information between the entorhinal cortex and other cortical areas as well as between the entorhinal cortex and the hippocampus. Therefore, any changes in parvalbumin-containing cells in the entorhinal cortex, like those we observed in aged memory-impaired rats [24], can cause disturbances in GABA-mediated inhibition and may thus interfere with normal functioning of the entorhinal cortex.
Acknowledgements We thank Prof. I. Virtanen (University of Helsinki, Helsinki, Finland) for the gift of antiserum against GABA and Dr. A. Pitk~inen (University of Kuopio, Kuopio, Finland) for her helpful comments on the manuscript. This research was supported by the Finnish Academy of Sciences.
References [1] Amaral, D.G. and Insausti, R., Hippocampal formation. In G. Paxinos (Ed.), The Human Nervous System, Academic Press, San Diego, 1990, pp. 711-755. [2] Arnold, S.E., Hyman, B.T., Van Hoesen, G.W. and Damasio, A.R., Some cytoarchitectural abnormalities of the entorhinal cortex in schizophrenia, Arch. Gen. Psychiatry, 48 (1991) 625-632. [3] Bliimcke, I., Hof, P.R., Morrison, J.H. and Cello, M.R., Distribution of parvalbumin immunoreactivity in the visual cortex of old world monkeys and humans, J. Comp. Neurol., 301 (1990) 417-432. [4] Braak, H. and Braak, E., Nettropathological stageing of Alzheimerrelated changes, Acta Neuropathol., 82 (1991) 239-259. [5] Cello, M.R., Parvalbumin in most 3,-aminobutyric acid-containing neurons of the rat cerebral cortex, Science, 231 (1986) 995-997. [6] Cello, M.R., Calbindin D-28k and parvalbumin in the rat nervous system, Neuroscience, 35 (1990) 375-475.
121
[7] 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. [8] Du, F., Whetsell Jr., W.O., Abou-Khalil, B., Blumenkopf, B., Lothman, E.W. and Schwarcz, R., Preferential neuronal loss in layer III of the entorhinal cortex in patients with temporal lobe epilepsy, Epilepsy Res., 16 (1993) 223-233. [9] Freund, T.F., GABAergic septohippocampal neurons contain parvalbumin, Brain Res., 478 (1989) 375-381. [10] Gottfries, C.G., Neurochemical aspects on aging and diseases with cognitive impairment, J. Neurosci. Res., 27 (1990) 541-547. [11] Hedreen, J.C., Bacon, S.J. and Price, D.L., A modified histochemical technique to visualize acetylcholinesterase-containing axons, J. Histochem. Cytochem., 33 (1985) 134-140. [12] Heizmann, C.W. and Cello, M.R., Immunolocalization of parvalbumin, Meth. EnzymoL, 139 (1987) 552-570. [13] Hendry, S.H.C., Jones, E.G., Emson, P.C., Lawson, D.E.M., Heizmann, C.W. and Streit, P., Two classes of cortical GABA neurons defined by differential calcium binding protein immunoreactivities, Exp. Brain Res., 76 (1989) 467-472. [14] Izquierdo, I. and Medina, J.H., Correlation between the pharmacology of long-term potentiation and the pharmacology of memory, NeurobioL Learn. Memory, 63 (1995) 19-32. [15] Jones, R.S.G., Entorhinal-hippocampal connections: a speculative view of their function, Trends Neurosci., 16 (1993) 58-64. [16] Jones, R.S.G. and Biihl, E.H., Basket-like interneurones in layer II of the entorhinal cortex exhibit a powerful NMDA-mediated synaptic excitation, Neurosci. Lett., 149 (1993) 35-39. [17] Kosaka, T., Katsumaru, H., Hama, K., Wu, J.-Y. and Heizmann, C.W., GABAergic neurons containing the Ca2÷-binding protein parvalbumin in the rat hippocampus and dentate gyms, Brain Res., 419 (1987) 119-130. [18] Kosaka, T., Kosaka, K., Heizmann, C.W., Nagatsu, I., Wu, J.-Y., Yanaihara, N. and Hama, K., An aspect of the organization of the GABAergic system in the rat main olfactory bulb: laminar distribution of immunohistochemically defined subpopulations of GABAergic neurons, Brain Res., 411 (1987) 373-378. [19] Kosaka, T., Kosaka, K., Tateishi, K., Hamaoka, Y., Yanaihara, N., Wu, J.-Y. and Hama, K., GABAergic neurons containing CCK-8-1ike and/or VIP-like immunoreactivities in the rat hippocampus and dentate gyms, J. Comp. Neurol., 239 (1985) 420-430. [20] Kosaka, T., Nagatsu, I., Wu, J.-Y. and Hama, K., Use of high concentrations of glutaraldehyde for immunocytochemistry of transmitter-synthesizing enzymes in the central nervous system, Neuroscience, 18 (1986) 975-990. [21] K6hler, C., Wu, J.-Y. and Chan-Palay, V., Neurons and terminals in the retrohippocampal region in the rat's brain identified by anti-3,aminobutyric acid and anti-giutamic acid decarboxylase immunocytochemistry, Anat. Embryol., 173 (1985) 35-44. [22] Lewis, D.A. and Lund, J.S., Heterogeneity of chandelier neurons in monkey neocortex: corticotropin-releasing factor- and parvalbuminimmunoreactive populations, J. Comp. Neurol., 293 (1990) 599-615. [23] Lorente de N6, R., Studies on the structure of the cerebral cortex II. Continuation of the study of the Ammonic system, J. Psychol. Neurol., 46 (1934) 113-177. [24] Miettinen, R., Sirvi6, J., Riekkinen Sr., P., Laakso, M. P., Riekkinen, M. and Riekkinen Jr., P., Neocortical, hippocampal and septal parvalbumin- and somatostatin-containing neurons in young and aged rats: correlation with passive avoidance and water maze performance, Neuroscience, 53 (1993) 367-378. [25] Nitsch, R., Soriano, E. and Frotscher, M., The parvalbumin-containing nonpyramidal neurons in the rat hippocampus, Anat. Embryol., 181 (1990) 413-425. [26] Ribak, C.E. and Seress, L., Five types of basket cell in the hippocampal dentate gyms: a combined Golgi and electron microscopic study, J. Neurocytol., 12 (1983) 577-597. [27] Reinikainen, K.J., Palj~irvi, L., Huuskonen, M., Soininen, H., Laakso,
122
M. Miettinen et al. / Brain Research 706 (1996) 113-122
M. and Riekkinen, P.J., A post-mortem study of noradrenergic, serotonergic and GABAergic neurons in Alzheimer's disease, J. Neurol. Sci., 84 (1988) 101-116. [28] Somogyi, P., Smith, A.D., Nunzi, M.G., Gorio, A., Takagi, H. and Wu, J.Y., Glutamate decarboxylase immunoreactivity in the hippocampus of the cat: distribution of immunoreactive synaptic terminals with special reference to the axon initial segment of pyramidal neurons, J. Neurosci., 3 (1983) 1450-1468. [29] Soriano, E. and Frotscher, M., A GABAergic axo-axonic cell in the fascia dentata controls the main excitatory hippocampal pathway, Brain Res., 503 (1989) 170-174. [30] Soriano, E., Nitsch, R. and Frotscher, M., Axo-axonic chandelier cells in the rat fascia dentata: Golgi-electron microscopy and immunocytochemical studies, J. Comp. Neurol., 293 (1990) 1-25.
[31] Szabat, E., Soinila, S., H~ipp61~i, O., Linnala, A. and Virtanen, I., A new monoclonal antibody against the GABA-protein conjugate shows immunoreactivity in sensory neurons of the rat, Neuroscience, 47 (1992) 409-420. [32] 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. [33] Witter, M.P., Organization of the entorhinal-hippocampat system: a review of current anatomical data, Hippocampus, 3 (1993) 33-44. [34] Wouterlood, F.G., H~irtig, W., Briickner, G. and WiRer, M.P., Parvalbumin-immunoreactive neurons in the entorhinal cortex of the rat: localization, morphology, connectivity and ultrastructure, J. Nearocytol., 24 (1995) 135-153.