Neuroscience Letters, 135 (1992) 33-36 © 1992 Elsevier Scientific Publishers Ireland Ltd. All rights reserved 0304-3940/92/$ 05.00
33
NSL 08327
Reorganization of input synapses of parvalbumin-containing neurons in the rat fascia dentata following entorhinal lesion Robert Nitsch, Sabine Bader and Michael Frotscher Institute of Anatomy, University of Freiburg, Freiburg (F.R.G.)
(Received 22 August 1991; Revised version received 9 October 1991; Accepted 11 October 1991) Key words: GABAergic inhibitory neuron; Reactive synaptogenesis; Terminal proliferation; Neuronal plasticity; Transneuronal atrophy; Feedforward inhibition
Removal of ipsilateral entorhinal afferents to the fascia dentata results in a retraction of postsynaptic parvalbumin (PARV)-containingdendrites of GABAergicneurons from the outer molecular layer. This study analyzesthe reorganization of input synapsesof these identified neurons following deafferentation. The density of synaptic input (total length of synaptic membrane specializations) of PARV-immunostaineddendrites increased by 34% in the outer molecular layer of the fascia dentata 8 days, and by 21% 55 days followinglesion when compared with unoperated controls. Eight days postlesion this increase was mainly due to an enlargement of synaptic membrane specializations of single terminals whereas 55 days after the lesion there was an increase in the number of synapses on the identified dendrites. Our results suggest expansion of terminals of remaining afferent systems (i.e. commissural fibers) in the early postlesional period and reactive synaptogenesis (i.e. de novo formation of synaptic contacts) on PARV-positive dendrites after long survival time. This increased innervation may be of functional importance as it might compensate for the reduction of the receptive field of the PARV-positive,supposedly inhibitory neurons in the postlesional dentate gyrus.
Studies about the process of reinnervation in the dentate gyrus of the adult rat have demonstrated terminal proliferation and reactive synaptogenesis in the ipsilateral fascia dentata following entorhinal lesion [4, 10, 15]. These studies focused on the reinnervation of the dominating cell type in the fascia dentata, the granule cells. Little is known about the reinnervation of 7~-aminobutyric acid (GABA)ergic inhibitory neurons. Recently, we have shown persisting retraction of the dendrites of GABAergic parvalbumin (PARV)-containing neurons from the termination zones of the perforant pathway following entorhinal lesion [13]. This is in contrast to the granule cells which do not retract their dendrites from the outer molecular layer [2]. The PARV-containing neurons represent a subpopulation of hippocampal and dentate GABAergic inhibitory neurons, a m o n g them the basket and axo-axonic cells [7, 12, 14]. The by far more numerous granule cells are the excitatory projection neurons of the fascia dentata. The entorhinal glutamatergic afferents terminate on both types o f neurons [1, 11, 16, 17]. Differences between excitatory granule cells and GABAergic inhibitory neurons in their reaction to entorhinal lesion may influence the balance of excitation and Correspondence: R. Nitsch, Institute of Anatomy, University of Freiburg, AlbertstraBe 17, D-7800 Freiburg, F.R.G.
inhibition [8] in the postlesional fascia dentata. In this study the synaptic reorganization on the retracted dendrites of inhibitory GABAergic neurons following entorhinal lesion was analyzed by employing immunocytochemistry for parvalbumin and electron microscopy. Ten adult male Sprague-Dawley rats (250 g) raised under standard laboratory conditions were used for this study. Animals were subjected to unilateral entorhinal cortex lesion [17] and were perfusion-fixed 8 (n=4) or 55 days (n=3) following operation. Three unoperated animals served as controls. Immunocytochemistry was performed on 40/tm-thick vibratome sections from the mid-septo-temporal portion of the ipsilateral hippocampus. Sections were incubated in a monoclonal mouse-anti parvalbumin antiserum (diluted 1:5000 for 48 h at 4°C; [3]) and processed according to the ABCmethod [6] using diaminobenzidine as a chromogen. Control sections were incubated the same way except that the primary antibody was omitted. N o staining was observed under these conditions. Sections were embedded in Araldite and the molecular layer of fascia dentata, sandwiched between the granule cell layer and the hippocampal fissure, was trimmed out. Photographs were taken at low power magnification (x 2900) in a Zeiss EM 109 electron microscope. Photomontages of the entire width of the molecular layer were made at a
34 final magnification of these prints of x 8700. The width of the molecular layer was scaled from 0.0 (upper border of granule cell layer) to 1.0 (hippocampal fissure), and the position of a dendritic segment could be defined according to this scale. Using an interactive computerized morphometric system (Ibas 1000, Kontron), the following measurements and calculations were made from photomicrographs at a final magnification of x 23200: (1) total length of synaptic membrane specialization along the membrane of immunoreactive dendrites, (2) mean length of synaptic membrane specializations, and (3) number of terminals establishing synaptic contact per length of membrane of immunopositive dendrites (see Fig. 1). The molecular layer was divided into an inner and outer part at the mid-point (0.5 on the scale described above). No correction for shrinkage [9] of the postlesional molecular layer was made since our quantification referred to identified and localized postsynaptic elements. Data of single identified dendrites as well as individual animals were tested for statistical significance using the t-test and the Duncan's multiple range test. We studied the length of dendritic membrane occupied by synaptic membrane specialization as a morphological parameter for the density of synaptic input. Our results show an increase in the density of synaptic input on PARV-positive dendrites in the outer molecular layer. Relative to control values, the density was increased by 34% 8 days and by 21% 55 days following lesion (Fig. 2a). The postlesional increase in the density of synaptic
input occurred to varying degrees throughout the width of the molecular layer (Fig. 2b). The average length of a single synaptic membrane specialization in the inner molecular layer increased by 28% 8 days following lesion and in the outer molecular layer by 32% (Fig. 2c). Fiftyfive days following lesion there was no increase in the average length of single synaptic contacts when compared with controls but the number of terminals establishing synaptic contact with PARV-positive dendrites in the outer molecular layer increased by 23% relative to control values (Fig. 2d). These data indicate that there is an expansion of terminals contacting dendrites of inhibitory neurons throughout the entire width of the molecular layer 8 days following entorhinal lesion. In fact, it has been shown that sprouting of remaining hippocampal afferents (e.g. commissural fibers) is at its peak at about this postlesional stage [10, 15] and that there is synaptic reorganization throughout the entire molecular layer [5]. As no significant increase in the number of terminals was monitored at this postlesional stage, we suggest that the increase in the density of synaptic input is mainly due to an enlargement of the specialized membranes of remaining afferent terminals which thereby extend their termination field into the former perforant path termination zone (Fig. 3). Conversely, the increase in the density of synaptic input 55 days post lesion is apparently due to an increase in the number of terminals since there is no significant difference to controls in the mean length of spe-
Fig. 1. Electron micrographs of parvalbumin (PARV)-immunoreactivedendrites in the molecular layer of fascia dentata of a control (a) and a lesioned(8 days postlesion)animal(b). Synapticmembranespecializationsweremeasuredas indicatedby the distancebetweensmallarrows, x 39300.
35 0.20 % 140u~
~.-~
120
.---"
EE
......... .
.............................................
-,
lO0
o'-
80
>.~
60
~.~
3~
40
\
20
0,10 0
5;d
~ 8d 5~d
iml
a %
0.0
~%
0.4
0.6
0.8
1.0
I
~:ou+
8o"~ "
60-
~
200 C
c
0.2
gl
b
**
il°-
.~¢
- -
oml
8(I
iml
I 55d
C
8d 55d oml
1~
C
d
8d 55d iml
C
8d 55d oml
Fig. 2. a: relative length of synaptic membrane specializations on PARV-immunoreactive dendrites in the inner molecular layer (iml) and outer molecular layer (oml) of fascia dentata 8 days (d) and 55 days following ipsilateral entorhinal lesion. Values of control animals (C) were taken as 100%. ** P
33
NSL 08327
Reorganization of input synapses of parvalbumin-containing neurons in the rat fascia dentata following entorhinal lesion Robert Nitsch, Sabine Bader and Michael Frotscher Institute of Anatomy, University of Freiburg, Freiburg (F.R.G.)
(Received 22 August 1991; Revised version received 9 October 1991; Accepted 11 October 1991) Key words: GABAergic inhibitory neuron; Reactive synaptogenesis; Terminal proliferation; Neuronal plasticity; Transneuronal atrophy; Feedforward inhibition
Removal of ipsilateral entorhinal afferents to the fascia dentata results in a retraction of postsynaptic parvalbumin (PARV)-containingdendrites of GABAergicneurons from the outer molecular layer. This study analyzesthe reorganization of input synapsesof these identified neurons following deafferentation. The density of synaptic input (total length of synaptic membrane specializations) of PARV-immunostaineddendrites increased by 34% in the outer molecular layer of the fascia dentata 8 days, and by 21% 55 days followinglesion when compared with unoperated controls. Eight days postlesion this increase was mainly due to an enlargement of synaptic membrane specializations of single terminals whereas 55 days after the lesion there was an increase in the number of synapses on the identified dendrites. Our results suggest expansion of terminals of remaining afferent systems (i.e. commissural fibers) in the early postlesional period and reactive synaptogenesis (i.e. de novo formation of synaptic contacts) on PARV-positive dendrites after long survival time. This increased innervation may be of functional importance as it might compensate for the reduction of the receptive field of the PARV-positive,supposedly inhibitory neurons in the postlesional dentate gyrus.
Studies about the process of reinnervation in the dentate gyrus of the adult rat have demonstrated terminal proliferation and reactive synaptogenesis in the ipsilateral fascia dentata following entorhinal lesion [4, 10, 15]. These studies focused on the reinnervation of the dominating cell type in the fascia dentata, the granule cells. Little is known about the reinnervation of 7~-aminobutyric acid (GABA)ergic inhibitory neurons. Recently, we have shown persisting retraction of the dendrites of GABAergic parvalbumin (PARV)-containing neurons from the termination zones of the perforant pathway following entorhinal lesion [13]. This is in contrast to the granule cells which do not retract their dendrites from the outer molecular layer [2]. The PARV-containing neurons represent a subpopulation of hippocampal and dentate GABAergic inhibitory neurons, a m o n g them the basket and axo-axonic cells [7, 12, 14]. The by far more numerous granule cells are the excitatory projection neurons of the fascia dentata. The entorhinal glutamatergic afferents terminate on both types o f neurons [1, 11, 16, 17]. Differences between excitatory granule cells and GABAergic inhibitory neurons in their reaction to entorhinal lesion may influence the balance of excitation and Correspondence: R. Nitsch, Institute of Anatomy, University of Freiburg, AlbertstraBe 17, D-7800 Freiburg, F.R.G.
inhibition [8] in the postlesional fascia dentata. In this study the synaptic reorganization on the retracted dendrites of inhibitory GABAergic neurons following entorhinal lesion was analyzed by employing immunocytochemistry for parvalbumin and electron microscopy. Ten adult male Sprague-Dawley rats (250 g) raised under standard laboratory conditions were used for this study. Animals were subjected to unilateral entorhinal cortex lesion [17] and were perfusion-fixed 8 (n=4) or 55 days (n=3) following operation. Three unoperated animals served as controls. Immunocytochemistry was performed on 40/tm-thick vibratome sections from the mid-septo-temporal portion of the ipsilateral hippocampus. Sections were incubated in a monoclonal mouse-anti parvalbumin antiserum (diluted 1:5000 for 48 h at 4°C; [3]) and processed according to the ABCmethod [6] using diaminobenzidine as a chromogen. Control sections were incubated the same way except that the primary antibody was omitted. N o staining was observed under these conditions. Sections were embedded in Araldite and the molecular layer of fascia dentata, sandwiched between the granule cell layer and the hippocampal fissure, was trimmed out. Photographs were taken at low power magnification (x 2900) in a Zeiss EM 109 electron microscope. Photomontages of the entire width of the molecular layer were made at a
34 final magnification of these prints of x 8700. The width of the molecular layer was scaled from 0.0 (upper border of granule cell layer) to 1.0 (hippocampal fissure), and the position of a dendritic segment could be defined according to this scale. Using an interactive computerized morphometric system (Ibas 1000, Kontron), the following measurements and calculations were made from photomicrographs at a final magnification of x 23200: (1) total length of synaptic membrane specialization along the membrane of immunoreactive dendrites, (2) mean length of synaptic membrane specializations, and (3) number of terminals establishing synaptic contact per length of membrane of immunopositive dendrites (see Fig. 1). The molecular layer was divided into an inner and outer part at the mid-point (0.5 on the scale described above). No correction for shrinkage [9] of the postlesional molecular layer was made since our quantification referred to identified and localized postsynaptic elements. Data of single identified dendrites as well as individual animals were tested for statistical significance using the t-test and the Duncan's multiple range test. We studied the length of dendritic membrane occupied by synaptic membrane specialization as a morphological parameter for the density of synaptic input. Our results show an increase in the density of synaptic input on PARV-positive dendrites in the outer molecular layer. Relative to control values, the density was increased by 34% 8 days and by 21% 55 days following lesion (Fig. 2a). The postlesional increase in the density of synaptic
input occurred to varying degrees throughout the width of the molecular layer (Fig. 2b). The average length of a single synaptic membrane specialization in the inner molecular layer increased by 28% 8 days following lesion and in the outer molecular layer by 32% (Fig. 2c). Fiftyfive days following lesion there was no increase in the average length of single synaptic contacts when compared with controls but the number of terminals establishing synaptic contact with PARV-positive dendrites in the outer molecular layer increased by 23% relative to control values (Fig. 2d). These data indicate that there is an expansion of terminals contacting dendrites of inhibitory neurons throughout the entire width of the molecular layer 8 days following entorhinal lesion. In fact, it has been shown that sprouting of remaining hippocampal afferents (e.g. commissural fibers) is at its peak at about this postlesional stage [10, 15] and that there is synaptic reorganization throughout the entire molecular layer [5]. As no significant increase in the number of terminals was monitored at this postlesional stage, we suggest that the increase in the density of synaptic input is mainly due to an enlargement of the specialized membranes of remaining afferent terminals which thereby extend their termination field into the former perforant path termination zone (Fig. 3). Conversely, the increase in the density of synaptic input 55 days post lesion is apparently due to an increase in the number of terminals since there is no significant difference to controls in the mean length of spe-
Fig. 1. Electron micrographs of parvalbumin (PARV)-immunoreactivedendrites in the molecular layer of fascia dentata of a control (a) and a lesioned(8 days postlesion)animal(b). Synapticmembranespecializationsweremeasuredas indicatedby the distancebetweensmallarrows, x 39300.
35 0.20 % 140u~
~.-~
120
.---"
EE
......... .
.............................................
-,
lO0
o'-
80
>.~
60
~.~
3~
40
\
20
0,10 0
5;d
~ 8d 5~d
iml
a %
0.0
~%
0.4
0.6
0.8
1.0
I
~:ou+
8o"~ "
60-
~
200 C
c
0.2
gl
b
**
il°-
.~¢
- -
oml
8(I
iml
I 55d
C
8d 55d oml
1~
C
d
8d 55d iml
C
8d 55d oml
Fig. 2. a: relative length of synaptic membrane specializations on PARV-immunoreactive dendrites in the inner molecular layer (iml) and outer molecular layer (oml) of fascia dentata 8 days (d) and 55 days following ipsilateral entorhinal lesion. Values of control animals (C) were taken as 100%. ** P
