Brain Research, 529 (1990) 165-177 Elsevier
165
BRES 15903
The postsubicular cortex in the rat: characterization of the fourth region of the subicular cortex and its connections Thomas van Groen and J. Michael Wyss Department of Cell Biology and Anatomy, University of Alabama at Birmingham, University Station, Birmingham, AL 35294 (U.S.A.) (Accepted 3 April 1990) Key words: Anterior thalamic nucleus; Cingulate cortex; Hippocampus; Limbic system; Subicular cortex
The hippocampal formation contributes importantly to many cognitive functions, and therefore has been a focus of intense anatomical and physiological research. Most of this research has focused on the hippocampus proper and the fascia dentata, and much less attention has been given to the subicular cortex, the origin of most extrinsic projections from the hippocampal formation. The present experiments demonstrate that the postsubiculum is a distinct area of the subicular cortex. The major projections to the postsubiculum originate in the hippocampal formation, the cingulate cortex, and the thalamus (primarily from the anterodorsal (AD) nucleus and to a lesser extent from the anteroventral (AV) and lateral dorsal (LD) nuclei). These projections differ from the thalamic projections to presubiculum and parasubiculum. Efferent projections from the postsubiculum terminate in both cortical and subcortical areas. The cortical projections terminate in the subicular and retrosplenial cortices and in the caudal lateral entorhinal and perirhinal cortices. Subcortical projections primarily end in the AD and the LD nuclei of the thalamus. These thalamic projections end in areas that are distinct from those to which the presubiculum and parasubiculum project. For instance, the postsubiculum has a dense terminal field in the AD nucleus, but presubicular axons terminate predominantly in the AV nucleus. The cortical projections also distinguish postsubiculum. All subicular areas project to the entorhinal cortex, but the postsubicular projection ends in the deep layers (i.e. IV-VI), whereas presubiculum projects to layers I and III, and parasubiculum projects to layer II. Postsubiculum projects to retrosplenial granular b cortex and only incidentally to retrosplenial granular a cortex. In contrast presubiculum projects to the retrosplenial granular a cortex but not to the retrosplenial granular b cortex. These differences clearly mark the postsubiculum, the presubiculum, and the parasubiculum as distinct regions within the subicular cortex and suggest that they subserve different roles in the processing and integration of limbic system information. INTRODUCTION The hippocampus contributes importantly to many cognitive functions, such as learning and memory (e.g. refs. 20, 27, 28). The anatomy and the physiology of the hippocampal formation have been a focus of intense anatomical and physiological research over the past 100 years; however, the distinct role of each component of the hippocampal formation has remained enigmatic (but see refs. 13, 14, 24). Many studies have focused on the hippocampus (i.e. the fascia dentata and the hippocampus proper (i.e. areas C A 3 and CA1) ) and the entorhinal cortex, but few studies have rigorously examined the subicular cortex despite the fact that the subicular cortex (together with the entorhinal cortex) play important roles in the transfer of information to and from the hippocampus. Past researchers have been consistent in designating the subiculum, parasubiculum and presubiculum as the m a j o r divisions of the subicular cortex, but they are less in agreement on the prosubiculum 19 and the postsubiculum 2'32'5° being distinct components of this region.
The difficulty in defining the number of regions that compose the subicular cortex is confounded by the fact that not all divisions of the subicular cortex are equally represented in all species. For example, the prosubiculum is not recognized in rat, whereas its presence is obvious in cat and primate. Our recent study of the presubicular and parasubicular cortices 49 indicates that these regions of the subicular complex have unique anatomical and connectional characteristics and thus, play a distinct role in the integration of information by the hippocampal formation. The thalamic and cortical connections of each subicular region are strikingly different 49. For instance, the presubiculum is innervated by the AV and LD thalamic nuclei, whereas the projection to the parasubiculum originates in A D and LD 48'49. Further, presubiculum, parasubiculum, and postsubiculum of the hippocampal formation have distinct projections to the thalamic nuclei 35'4°'49'59 and each region of the subicular cortex has dissimilar projections to cortical regions. In the present study we characterize the postsubicular cortex, and we describe its efferent and
Correspondence: J.M. Wyss, Department of Cell Biology and Anatomy, University of Alabama at Birmingham, University Station, Box 302, Birmingham, AL 35294, U.S.A. 0006-8993/90/$03.50 t~) 1990 Elsevier Science Publishers B.V. (Biomedical Division)
166 a f f e r e n t c o n n e c t i o n s . T h e results of these e x p e r i m e n t s d e m o n s t r a t e that the p o s t s u b i c u l u m is anatomically and c o n n e c t i o n a l l y distinct from the divisions o f the subicular cortex.
plane on a freezing microtome (30/tm thick sections) and collected in trays. Two series were mounted on gelatin coated slides immediately; one was stained with methylene blue and coverslipped, the other was coverslipped only. The third series was stained for acetylcholinesterase (ACHE), mounted and coverslipped as previously described3s. All fluorescent material was inspected under bright field and fluorescent illumination.
MATERIALS AND METHODS Twenty-eight, male Sprague-Dawley rats (250-350 g, Charles River, Inc. Wilmington, MA) were deeply anesthetized with sodium pentobarbital prior to surgery (35 mg/kg i.p.). A glass pipette (10 /am tip diameter) or Hamilton 0.5/~1 microsyringe needle with a 90 /~m tip was lowered to the appropriate region by stereotaxic guidance. For autoradiographic, anterograde transport experiments, a mixture of [3H]amino acids (20-100/~Ci//d; equal parts of proline, lysine and leucine or proline only; New England Nuclear, Boston, MA) was injected (n = 10) either by iontophoresis (5-10 rain, 1.0 /~A positive pulsed current) or by pressure injection (20-100 nl) as previously described57'58. After a 2-7 day survival period the animals were reanesthetized, perfused with 10% formalin, the brains were removed and processed for autoradiography as described by Cowan et al. 3. After paraffin embedding, the brains were sectioned in the frontal plane (15/~m thick sections), and a 1 in 10 series of sections from each brain was mounted on slides and nuclear emulsion was exposed to them at 4 °C for 2-10 weeks. The resulting autoradiographs were subsequently inspected in both bright and dark field microscopy. In 20 animals the anterograde transport of Phaseolus vulgaris leucoagglutinin (PHA-L; Vector, Burlingame) was employed to study the pattern of axonal terminals I°. Glass micropipettes (10-20 /~m tip) were filled with a 2.5% solution of PHA-L in 0.05 M riffs buffer, and stereotaxically positioned in the brain. A positive pulsed (7 s on, 7 s off) DC current (4-6/~A) was applied for 20-40 min to the pipette, using a Midgard CS-3 constant current source. After a survival time of 7-9 days, the rats were reanesthetized and transcardially perfused with 100 ml phosphate buffered (0.1 M) saline (pH 7.4) followed by 150 ml of a Na-acetate (0.1 M) buffered, 4% paraformaldehyde solution (pH 6.5), and 250 ml of a Na-borate (0.1 M) buffered, 4% paraformaldehyde, 0.1% glutaraldehyde solution (pH 9.5). Following overnight postfixation in the final fixative, the brains were stored 2 days at 4 °C in 0.1 M phosphate buffer (pH 7.4) containing 25% sucrose. Two or 3 series of frozen sections (30 #m) were cut on a freezing microtome, and collected in phosphate buffer (0.1 M, pH 7.4). The first series was mounted immediately on gelatin coated slides and stained with Cresyl violet. The second series was rinsed overnight in a solution of 0.05 M Tris buffer (pH 8.6), 0.5 M NaCI and 0.5% Triton X-100 (TBS-T). The next day the sections were transferred to a TBS-T solution containing a 1:t000 concentration of the primary goat-antbPhaseolus (Vector) antibody. The tissue was incubated on a rotation table for 18 h at 20 °C in the dark. The sections were rinsed 3 x 5 rain in TBS-T and incubated with rabbit-anti-goat whole serum (Sigma) in TBS-T (1:400) for 2 h. After washes (3 × 5 min) in TBS-T, the sections were transferred to goat peroxidase-antiperoxidase (PAP; Sigma) in TBS-T (1:400) for 4 h. After another rinse (3 x 5 min in TBS-T) the sections were incubated for 1 h with a solution containing 40 mg diaminobenzidine (DAB) in 100 ml TBS-T to which 0.9 ml H20 2 (1.5%) was added. The sections were washed thoroughly, mounted on subbed slides and coverslipped. For retrograde transport experiments a small (10-30 nl) amount of one of two fluorescent dyes (Fast blue, FB; Iiling, ER.G.; or fluorogold, FG; Fluorochrome, Inc. Englewood, CO; 4% in DH20 ) was injected by pressure into a defined region of the cortex. After a 5-7 day survival period, the rats were reanesthetized, transcardially porfused with 150 ml of buffered saline followed by 250 ml of 4% paraformaldehyde in phosphate buffer. The brain was removed from the skull and stored for 12 h in the fixative (4 °C), after which it was placed in 0.1 M phosphate buffer containing 30% sucrose (4 °C). Three (or two) 1 in 6 series of sections were cut in the frontal
RESULTS
Nomenclature The subicular cortex was subdivided into the subiculum p r o p e r , the parasubiculum, the presubiculum and the postsubiculum 2'32'33'5°. The parasubiculum, presubiculum and postsubiculum were characterized by their two broad cellular laminae (the external and internal principal lamina), and the clear lamina dissecans that separated the two laminae. T h e external lamina was subdivided into layers I - I I I ; the internal lamina was subdivided into layers IV-V118'19 (Fig. 1). The external lamina of the postsubiculum consisted of a characteristic cellular arr a n g e m e n t of small layer III cells that were covered by a thin layer of s o m e w h a t larger, darkly stained neurons (layer II) g r o u p e d into clusters or islands (Fig. 1A,C). These islands did not stain for ACHE, in contrast to the a r e a surrounding the islands that was densely stained (Fig. 1B,D). Ventrally and laterally, the postsubiculum was bord e r e d by the presubiculum, the b o r d e r was characterized by an abrupt change in the cyto- and histochemical staining. T h e layer II cell islands of the postsubiculum were not present in the presubiculum. Further. in postsubiculum, layer III neurons were organized in rows parallel to the pial surface; conversely, the presubiculum displayed no such organization (Fig. I A , C ) . In the A C h E material, the d e e p layers of postsubiculum were darkly stained for ACHE, whereas the deep layers of the presubiculum were m o r e lightly stained (Fig. IB,D). F u r t h e r m o r e , in Vogt-silver stained material 51, the postsubiculum was characterized by a dense fiber plexus in layer Ia and extremely few fibers in layer II. In contrast. the presubiculum had an even distribution of stained fibers in layers 1-11149 . Dorsally and medially, the postsubiculum was b o u n d e d by the retrosptenial granular a cortex (Rga; Fig. 1). A t the junction of these cortical areas, the lamina dissecans of the postsubiculum d i s a p p e a r e d and was r e p l a c e d by the granular cells of layer IV of Rga. The b o r d e r b e t w e e n the postsubiculum and Rga also was characterized by an abrupt change in A C h E staining. L a y e r I of the postsubiculum was evenly stained, but in Rga, layer I displayed patches of light and dark staining (Fig. 1B). F u r t h e r , layer III was stained in postsubiculum but not in Rga (Fig. 1B,D), and layer IV in R g a was
IM
168 darkly stained (Fig. 1B (arrowhead), D), whereas this layer of postsubiculum was less densely stained.
Tracing experiments Efferent connections. Injections of [3H]amino acids into the postsubiculum of the rat (e.g. CIR 29 and CIR 114) resulted in anterogradely transported label in the ipsilateral Rga, Rgb, presubiculum, parasubiculum, and caudal entorhinal and perirhinal cortices. Contralaterally, cortical label was predominantly confined to the caudal entorhinal and perirhinal cortices with only a few labeled silver grains in the postsubiculum. In addition to the cortical labeling, the postsubicular injections labeled the anterodorsal and laterodorsal nuclei of the thalamus and lightly labeled the lateral mamillary nucleus.
To study the projections of the postsubiculum in more detail, injections of the anterograde tracer PHA-L were made in the postsubiculum. Following a small injection in the middle of the postsubiculum (PHA 28), the pattern of labeled fibers (Fig. 2) was similar to that observed in the [3H]amino acid experiments. Labeled axons extended from the postsubiculum into the retrosplenial granular cortex (Fig. 2A-C). In Rga a few fibers could be seen running dorsally through layer I. arborizing occasionally, apparently synapsing in layer I. and extending into Rgb where the fibers coursed rostrally in layers I and II. synapsing in those layers. In contrast to the above route. most labeled fibers destined for the retrospleniai cortex coursed rostrally through the cingulum to the Rgb cortex. where fibers left the cingulum to innervate layers I
A B
,,at
~b
° Fig. 2. Four line drawings to demonstrate the position of axons and terminals labeled by an injection of PHA-L into the postsubicularcortex (crosshatched area). Scale bar = 1000/tin.
Fig. 3. Bright field photomicrographs of postsubiculum (A-C), thalamus (D), and entorhinal and perirhinal cortices (E,F) to demonstrate the pattern of PHA-L labeling in individual experiments. A and B show the pattern of labeled axons in postsubiculum following an injection into Rgb/IR cortex (A) and subiculum (B). C demonstrates an injection site in postsubiculum and D the resulting labeling in the LD nucleus of the thalamus. E (demonstrating PHA-L labeled fibers) and F (corresponding Nissl stained section) demonstrate the label in the deep layers of the entorhinal and perirhinal cortices following an injection in the postsubiculum. Note in B that the subicular projection labels layer II in Rga (above the arrowhead) but primarily the deep layers in postsubiculum. The arrowhead in D denotes labeled fibers in the reticular nucleus of the thalamus. Scale bar in A = 100/zm, scale bars in B - E = 250/~m.
170 (primarily layer Ib and Ic) and II (Fig. 2A,B). The projections from the postsubiculum to the retrosplenial cortex were organized topographically. Injections in the rostrai part of postsubiculum gave rise to a terminal field in the rostral part of the Rgb cortex, while more caudal and ventral parts of the postsubieulum projected to more caudal regions of Rgb. Many labeled axons coursed ventrally from the postsubiculum to the presubiculum, parasubiculum and caudal lateral entorhinal and perirhinal cortices. In the
caudal presubiculum, the axons arborized extensively in a small terminal patch in the superficial layers (II and III) and a larger terminal field in the deeper (primarily V-IV) layers (Fig. 2D). The deeper patch of labeled fibers extended into the deep layers of the parasubiculttrn. Most of the ventrally directed fibers labeled by the postsubiculum injection extended into the caudal perirhinal and caudal lateral entorhinal cortices, where they ended in layers IV-VI (Figs. 2D and 3E,F). A dense (approx. 50% of ipsilateral labeling), terminal field was labeled in
A
B
5O0 p Fig. 4. Line drawings to demonstrate the position of labeled axons in the thalamus following PHA-L injections into r.he rostral (A) postsubicular and the caudal (B) postsubicular cortex.
171
( Fig. 5. Five line drawings to demonstrate the positionof neuronal cell bodies (diamonds) labeled by an injection of FG into the postsubiculum (crosshatching). One symbol equals approximately 5 labeled neurons. Scale bar = 1000/~m.
the contralateral entorhinal and perirhinal cortices in the same place (Fig. 2D). In addition to the cortical projections, the postsubiculum injection labeled a few axons that passed through the fimbria/fornix to innervate the mamillary bodies. In the mamillary bodies the labeled axons terminated specifically in the lateral mamillary nucleus ipsilateral to the injection (Fig. 2B). Other labeled axons coursed around the .stria terminalis to the rostral thalamic nuclei and terminated in two dense terminal fields in the lateral dorsal nucleus (LD; Figs. 2A, 3D and 4), and a dense patch in the anterodorsal nucleus (AD; Fig. 4). Further, in the anteroventral (AV) nucleus a very small terminal field was present anterior dorsally (Fig. 4), and a patch of labeling consistently was present in the portion of the reticular nuclei of the thalamus that was adjacent to LD. A few labeled fibers were also present in the nucleus reuniens (Fig. 4). The projections from the postsubiculum to the thalamus are organized topographically, although some overlap exists. Injections in the rostral part of the postsubiculum give rise to terminal fields in the ventromedial part of the AD nucleus and in two adjacent patches in the middle (lateral-medial) part of the LD nucleus (Fig. 4A). In contrast, injections in the caudal postsubiculum labeled a terminal field in rostro-
dorsal AD and two terminal fields in the LD nucleus; a large lateral patch and a smaller medial patch (Fig. 4B). Retrograde tracing experiments were used to confirm the anterograde projections and to reveal the laminar organization of neurons giving rise to each projection. Injections confined to the postsubiculum (CF 310) labeled layer II-V neurons ipsilaterally in postsubiculum, but labeled only a few neuronal cell bodies in layers II and V in the postsubiculum contralateral to the injection. Following injections into the presubiculum (e.g. CF 328) or parasubiculum (e.g. CF 332), cell bodies were labeled in layers II, V and VI of postsubiculum; however, the majority (approx. 70%) of the labeled cell bodies were in layer V. Injections into the lateral entorhinal cortex (e.g. CF 322) labeled neurons in layers II, III and V of the ipsilateral and contralateral postsubiculum. Injections of fluorescent tracers into Rga (e.g. CF 310) or Rgb (e.g. CF 274) labeled neurons in layer V of postsubiculum. Following anterior thalamic nuclei injections (e.g. CF 280, 286) layer VI neurons were labeled in postsubiculum. Afferent connections. Injections of retrogradely transported tracers in the postsubiculum (e.g. CF 385; Figs. 5D and 6A) labeled both cortical and subcortical neuronal cell bodies (Fig. 5). Ipsilateral cortical label was
172
Fig. 6. Fluorescent ( A - D ) and dark field (E) photomicrographs of the limbic cortex to demonstrate the position of labeled cells in A - D , or axons and terminals in E. A shows an FG injection into postsubiculum; B demonstrates labeled neurons in the AD and AV nuclei of the anterior thalamus; C shows labeled neurons in the nucleus reuniens and D demonstrates labeled pyramidal layer neurons in the septal third of area of CA1 following an FG injection into ipsilateral postsubiculum. E shows the label in the postsubiculum following an injection of [3H}amino acids in the A D nucleus of the thalamus. Scale bars = 250/~m.
173 present in area CA1, in subiculum, presubiculum, parasubiculum, in the entorhinal perirhinal, retrosplenial and infraradiata cortices; contralaterally the postsubicular, presubicular, perirhinal, retrosplenial, and infraradiata cortices were labeled. Subcortically, neurons were labeled in the horizontal and vertical limbs of the diagonal band of Broca, the medial septal nucleus, the rostral claustrum, in the AD, AV, LD, and the reuniens nuclei of the thalamus, the supramamillary region, the dorsal and ventral raphe nuclei, and the locus coeruleus. In the subiculum, cell bodies were labeled in the dorsal, septal third of the subiculum ipsilateral to the injection (Fig. 5D). The postsubiculum injections also ipsilaterally labeled pyramidal layer neurons (Figs. 5C and 6D) and a few non-pyramidal neurons in the septal third of area CA1, and there were few (< 5%) contralaterally (compared to ipsilaterally) labeled cells. In the presubiculum ipsilateral and contralateral to the injection site, neurons were labeled in the external lamina (i.e. layers II and III; Fig. 5E), but in the parasubiculum only a few neurons were labeled in the external lamina. Labeled neurons were confined to caudal segments of the presubiculum and parasubiculum. In the caudal entorhinal cortex, neuronal cell bodies were labeled in layers II and III (Fig. 5E), and in the caudal perirhinal cortex layers II, III and V contained labeled neurons bilaterally A
F
Fig. 7. Line drawings to demonstrate the position of labeled neuronal cell bodies in the thalamus following an injection of FB in the middle part of the postsubiculum.
[SUB ICA t I
II III IV V
":':'::=7:
-
..
165
BRES 15903
The postsubicular cortex in the rat: characterization of the fourth region of the subicular cortex and its connections Thomas van Groen and J. Michael Wyss Department of Cell Biology and Anatomy, University of Alabama at Birmingham, University Station, Birmingham, AL 35294 (U.S.A.) (Accepted 3 April 1990) Key words: Anterior thalamic nucleus; Cingulate cortex; Hippocampus; Limbic system; Subicular cortex
The hippocampal formation contributes importantly to many cognitive functions, and therefore has been a focus of intense anatomical and physiological research. Most of this research has focused on the hippocampus proper and the fascia dentata, and much less attention has been given to the subicular cortex, the origin of most extrinsic projections from the hippocampal formation. The present experiments demonstrate that the postsubiculum is a distinct area of the subicular cortex. The major projections to the postsubiculum originate in the hippocampal formation, the cingulate cortex, and the thalamus (primarily from the anterodorsal (AD) nucleus and to a lesser extent from the anteroventral (AV) and lateral dorsal (LD) nuclei). These projections differ from the thalamic projections to presubiculum and parasubiculum. Efferent projections from the postsubiculum terminate in both cortical and subcortical areas. The cortical projections terminate in the subicular and retrosplenial cortices and in the caudal lateral entorhinal and perirhinal cortices. Subcortical projections primarily end in the AD and the LD nuclei of the thalamus. These thalamic projections end in areas that are distinct from those to which the presubiculum and parasubiculum project. For instance, the postsubiculum has a dense terminal field in the AD nucleus, but presubicular axons terminate predominantly in the AV nucleus. The cortical projections also distinguish postsubiculum. All subicular areas project to the entorhinal cortex, but the postsubicular projection ends in the deep layers (i.e. IV-VI), whereas presubiculum projects to layers I and III, and parasubiculum projects to layer II. Postsubiculum projects to retrosplenial granular b cortex and only incidentally to retrosplenial granular a cortex. In contrast presubiculum projects to the retrosplenial granular a cortex but not to the retrosplenial granular b cortex. These differences clearly mark the postsubiculum, the presubiculum, and the parasubiculum as distinct regions within the subicular cortex and suggest that they subserve different roles in the processing and integration of limbic system information. INTRODUCTION The hippocampus contributes importantly to many cognitive functions, such as learning and memory (e.g. refs. 20, 27, 28). The anatomy and the physiology of the hippocampal formation have been a focus of intense anatomical and physiological research over the past 100 years; however, the distinct role of each component of the hippocampal formation has remained enigmatic (but see refs. 13, 14, 24). Many studies have focused on the hippocampus (i.e. the fascia dentata and the hippocampus proper (i.e. areas C A 3 and CA1) ) and the entorhinal cortex, but few studies have rigorously examined the subicular cortex despite the fact that the subicular cortex (together with the entorhinal cortex) play important roles in the transfer of information to and from the hippocampus. Past researchers have been consistent in designating the subiculum, parasubiculum and presubiculum as the m a j o r divisions of the subicular cortex, but they are less in agreement on the prosubiculum 19 and the postsubiculum 2'32'5° being distinct components of this region.
The difficulty in defining the number of regions that compose the subicular cortex is confounded by the fact that not all divisions of the subicular cortex are equally represented in all species. For example, the prosubiculum is not recognized in rat, whereas its presence is obvious in cat and primate. Our recent study of the presubicular and parasubicular cortices 49 indicates that these regions of the subicular complex have unique anatomical and connectional characteristics and thus, play a distinct role in the integration of information by the hippocampal formation. The thalamic and cortical connections of each subicular region are strikingly different 49. For instance, the presubiculum is innervated by the AV and LD thalamic nuclei, whereas the projection to the parasubiculum originates in A D and LD 48'49. Further, presubiculum, parasubiculum, and postsubiculum of the hippocampal formation have distinct projections to the thalamic nuclei 35'4°'49'59 and each region of the subicular cortex has dissimilar projections to cortical regions. In the present study we characterize the postsubicular cortex, and we describe its efferent and
Correspondence: J.M. Wyss, Department of Cell Biology and Anatomy, University of Alabama at Birmingham, University Station, Box 302, Birmingham, AL 35294, U.S.A. 0006-8993/90/$03.50 t~) 1990 Elsevier Science Publishers B.V. (Biomedical Division)
166 a f f e r e n t c o n n e c t i o n s . T h e results of these e x p e r i m e n t s d e m o n s t r a t e that the p o s t s u b i c u l u m is anatomically and c o n n e c t i o n a l l y distinct from the divisions o f the subicular cortex.
plane on a freezing microtome (30/tm thick sections) and collected in trays. Two series were mounted on gelatin coated slides immediately; one was stained with methylene blue and coverslipped, the other was coverslipped only. The third series was stained for acetylcholinesterase (ACHE), mounted and coverslipped as previously described3s. All fluorescent material was inspected under bright field and fluorescent illumination.
MATERIALS AND METHODS Twenty-eight, male Sprague-Dawley rats (250-350 g, Charles River, Inc. Wilmington, MA) were deeply anesthetized with sodium pentobarbital prior to surgery (35 mg/kg i.p.). A glass pipette (10 /am tip diameter) or Hamilton 0.5/~1 microsyringe needle with a 90 /~m tip was lowered to the appropriate region by stereotaxic guidance. For autoradiographic, anterograde transport experiments, a mixture of [3H]amino acids (20-100/~Ci//d; equal parts of proline, lysine and leucine or proline only; New England Nuclear, Boston, MA) was injected (n = 10) either by iontophoresis (5-10 rain, 1.0 /~A positive pulsed current) or by pressure injection (20-100 nl) as previously described57'58. After a 2-7 day survival period the animals were reanesthetized, perfused with 10% formalin, the brains were removed and processed for autoradiography as described by Cowan et al. 3. After paraffin embedding, the brains were sectioned in the frontal plane (15/~m thick sections), and a 1 in 10 series of sections from each brain was mounted on slides and nuclear emulsion was exposed to them at 4 °C for 2-10 weeks. The resulting autoradiographs were subsequently inspected in both bright and dark field microscopy. In 20 animals the anterograde transport of Phaseolus vulgaris leucoagglutinin (PHA-L; Vector, Burlingame) was employed to study the pattern of axonal terminals I°. Glass micropipettes (10-20 /~m tip) were filled with a 2.5% solution of PHA-L in 0.05 M riffs buffer, and stereotaxically positioned in the brain. A positive pulsed (7 s on, 7 s off) DC current (4-6/~A) was applied for 20-40 min to the pipette, using a Midgard CS-3 constant current source. After a survival time of 7-9 days, the rats were reanesthetized and transcardially perfused with 100 ml phosphate buffered (0.1 M) saline (pH 7.4) followed by 150 ml of a Na-acetate (0.1 M) buffered, 4% paraformaldehyde solution (pH 6.5), and 250 ml of a Na-borate (0.1 M) buffered, 4% paraformaldehyde, 0.1% glutaraldehyde solution (pH 9.5). Following overnight postfixation in the final fixative, the brains were stored 2 days at 4 °C in 0.1 M phosphate buffer (pH 7.4) containing 25% sucrose. Two or 3 series of frozen sections (30 #m) were cut on a freezing microtome, and collected in phosphate buffer (0.1 M, pH 7.4). The first series was mounted immediately on gelatin coated slides and stained with Cresyl violet. The second series was rinsed overnight in a solution of 0.05 M Tris buffer (pH 8.6), 0.5 M NaCI and 0.5% Triton X-100 (TBS-T). The next day the sections were transferred to a TBS-T solution containing a 1:t000 concentration of the primary goat-antbPhaseolus (Vector) antibody. The tissue was incubated on a rotation table for 18 h at 20 °C in the dark. The sections were rinsed 3 x 5 rain in TBS-T and incubated with rabbit-anti-goat whole serum (Sigma) in TBS-T (1:400) for 2 h. After washes (3 × 5 min) in TBS-T, the sections were transferred to goat peroxidase-antiperoxidase (PAP; Sigma) in TBS-T (1:400) for 4 h. After another rinse (3 x 5 min in TBS-T) the sections were incubated for 1 h with a solution containing 40 mg diaminobenzidine (DAB) in 100 ml TBS-T to which 0.9 ml H20 2 (1.5%) was added. The sections were washed thoroughly, mounted on subbed slides and coverslipped. For retrograde transport experiments a small (10-30 nl) amount of one of two fluorescent dyes (Fast blue, FB; Iiling, ER.G.; or fluorogold, FG; Fluorochrome, Inc. Englewood, CO; 4% in DH20 ) was injected by pressure into a defined region of the cortex. After a 5-7 day survival period, the rats were reanesthetized, transcardially porfused with 150 ml of buffered saline followed by 250 ml of 4% paraformaldehyde in phosphate buffer. The brain was removed from the skull and stored for 12 h in the fixative (4 °C), after which it was placed in 0.1 M phosphate buffer containing 30% sucrose (4 °C). Three (or two) 1 in 6 series of sections were cut in the frontal
RESULTS
Nomenclature The subicular cortex was subdivided into the subiculum p r o p e r , the parasubiculum, the presubiculum and the postsubiculum 2'32'33'5°. The parasubiculum, presubiculum and postsubiculum were characterized by their two broad cellular laminae (the external and internal principal lamina), and the clear lamina dissecans that separated the two laminae. T h e external lamina was subdivided into layers I - I I I ; the internal lamina was subdivided into layers IV-V118'19 (Fig. 1). The external lamina of the postsubiculum consisted of a characteristic cellular arr a n g e m e n t of small layer III cells that were covered by a thin layer of s o m e w h a t larger, darkly stained neurons (layer II) g r o u p e d into clusters or islands (Fig. 1A,C). These islands did not stain for ACHE, in contrast to the a r e a surrounding the islands that was densely stained (Fig. 1B,D). Ventrally and laterally, the postsubiculum was bord e r e d by the presubiculum, the b o r d e r was characterized by an abrupt change in the cyto- and histochemical staining. T h e layer II cell islands of the postsubiculum were not present in the presubiculum. Further. in postsubiculum, layer III neurons were organized in rows parallel to the pial surface; conversely, the presubiculum displayed no such organization (Fig. I A , C ) . In the A C h E material, the d e e p layers of postsubiculum were darkly stained for ACHE, whereas the deep layers of the presubiculum were m o r e lightly stained (Fig. IB,D). F u r t h e r m o r e , in Vogt-silver stained material 51, the postsubiculum was characterized by a dense fiber plexus in layer Ia and extremely few fibers in layer II. In contrast. the presubiculum had an even distribution of stained fibers in layers 1-11149 . Dorsally and medially, the postsubiculum was b o u n d e d by the retrosptenial granular a cortex (Rga; Fig. 1). A t the junction of these cortical areas, the lamina dissecans of the postsubiculum d i s a p p e a r e d and was r e p l a c e d by the granular cells of layer IV of Rga. The b o r d e r b e t w e e n the postsubiculum and Rga also was characterized by an abrupt change in A C h E staining. L a y e r I of the postsubiculum was evenly stained, but in Rga, layer I displayed patches of light and dark staining (Fig. 1B). F u r t h e r , layer III was stained in postsubiculum but not in Rga (Fig. 1B,D), and layer IV in R g a was
IM
168 darkly stained (Fig. 1B (arrowhead), D), whereas this layer of postsubiculum was less densely stained.
Tracing experiments Efferent connections. Injections of [3H]amino acids into the postsubiculum of the rat (e.g. CIR 29 and CIR 114) resulted in anterogradely transported label in the ipsilateral Rga, Rgb, presubiculum, parasubiculum, and caudal entorhinal and perirhinal cortices. Contralaterally, cortical label was predominantly confined to the caudal entorhinal and perirhinal cortices with only a few labeled silver grains in the postsubiculum. In addition to the cortical labeling, the postsubicular injections labeled the anterodorsal and laterodorsal nuclei of the thalamus and lightly labeled the lateral mamillary nucleus.
To study the projections of the postsubiculum in more detail, injections of the anterograde tracer PHA-L were made in the postsubiculum. Following a small injection in the middle of the postsubiculum (PHA 28), the pattern of labeled fibers (Fig. 2) was similar to that observed in the [3H]amino acid experiments. Labeled axons extended from the postsubiculum into the retrosplenial granular cortex (Fig. 2A-C). In Rga a few fibers could be seen running dorsally through layer I. arborizing occasionally, apparently synapsing in layer I. and extending into Rgb where the fibers coursed rostrally in layers I and II. synapsing in those layers. In contrast to the above route. most labeled fibers destined for the retrospleniai cortex coursed rostrally through the cingulum to the Rgb cortex. where fibers left the cingulum to innervate layers I
A B
,,at
~b
° Fig. 2. Four line drawings to demonstrate the position of axons and terminals labeled by an injection of PHA-L into the postsubicularcortex (crosshatched area). Scale bar = 1000/tin.
Fig. 3. Bright field photomicrographs of postsubiculum (A-C), thalamus (D), and entorhinal and perirhinal cortices (E,F) to demonstrate the pattern of PHA-L labeling in individual experiments. A and B show the pattern of labeled axons in postsubiculum following an injection into Rgb/IR cortex (A) and subiculum (B). C demonstrates an injection site in postsubiculum and D the resulting labeling in the LD nucleus of the thalamus. E (demonstrating PHA-L labeled fibers) and F (corresponding Nissl stained section) demonstrate the label in the deep layers of the entorhinal and perirhinal cortices following an injection in the postsubiculum. Note in B that the subicular projection labels layer II in Rga (above the arrowhead) but primarily the deep layers in postsubiculum. The arrowhead in D denotes labeled fibers in the reticular nucleus of the thalamus. Scale bar in A = 100/zm, scale bars in B - E = 250/~m.
170 (primarily layer Ib and Ic) and II (Fig. 2A,B). The projections from the postsubiculum to the retrosplenial cortex were organized topographically. Injections in the rostrai part of postsubiculum gave rise to a terminal field in the rostral part of the Rgb cortex, while more caudal and ventral parts of the postsubieulum projected to more caudal regions of Rgb. Many labeled axons coursed ventrally from the postsubiculum to the presubiculum, parasubiculum and caudal lateral entorhinal and perirhinal cortices. In the
caudal presubiculum, the axons arborized extensively in a small terminal patch in the superficial layers (II and III) and a larger terminal field in the deeper (primarily V-IV) layers (Fig. 2D). The deeper patch of labeled fibers extended into the deep layers of the parasubiculttrn. Most of the ventrally directed fibers labeled by the postsubiculum injection extended into the caudal perirhinal and caudal lateral entorhinal cortices, where they ended in layers IV-VI (Figs. 2D and 3E,F). A dense (approx. 50% of ipsilateral labeling), terminal field was labeled in
A
B
5O0 p Fig. 4. Line drawings to demonstrate the position of labeled axons in the thalamus following PHA-L injections into r.he rostral (A) postsubicular and the caudal (B) postsubicular cortex.
171
( Fig. 5. Five line drawings to demonstrate the positionof neuronal cell bodies (diamonds) labeled by an injection of FG into the postsubiculum (crosshatching). One symbol equals approximately 5 labeled neurons. Scale bar = 1000/~m.
the contralateral entorhinal and perirhinal cortices in the same place (Fig. 2D). In addition to the cortical projections, the postsubiculum injection labeled a few axons that passed through the fimbria/fornix to innervate the mamillary bodies. In the mamillary bodies the labeled axons terminated specifically in the lateral mamillary nucleus ipsilateral to the injection (Fig. 2B). Other labeled axons coursed around the .stria terminalis to the rostral thalamic nuclei and terminated in two dense terminal fields in the lateral dorsal nucleus (LD; Figs. 2A, 3D and 4), and a dense patch in the anterodorsal nucleus (AD; Fig. 4). Further, in the anteroventral (AV) nucleus a very small terminal field was present anterior dorsally (Fig. 4), and a patch of labeling consistently was present in the portion of the reticular nuclei of the thalamus that was adjacent to LD. A few labeled fibers were also present in the nucleus reuniens (Fig. 4). The projections from the postsubiculum to the thalamus are organized topographically, although some overlap exists. Injections in the rostral part of the postsubiculum give rise to terminal fields in the ventromedial part of the AD nucleus and in two adjacent patches in the middle (lateral-medial) part of the LD nucleus (Fig. 4A). In contrast, injections in the caudal postsubiculum labeled a terminal field in rostro-
dorsal AD and two terminal fields in the LD nucleus; a large lateral patch and a smaller medial patch (Fig. 4B). Retrograde tracing experiments were used to confirm the anterograde projections and to reveal the laminar organization of neurons giving rise to each projection. Injections confined to the postsubiculum (CF 310) labeled layer II-V neurons ipsilaterally in postsubiculum, but labeled only a few neuronal cell bodies in layers II and V in the postsubiculum contralateral to the injection. Following injections into the presubiculum (e.g. CF 328) or parasubiculum (e.g. CF 332), cell bodies were labeled in layers II, V and VI of postsubiculum; however, the majority (approx. 70%) of the labeled cell bodies were in layer V. Injections into the lateral entorhinal cortex (e.g. CF 322) labeled neurons in layers II, III and V of the ipsilateral and contralateral postsubiculum. Injections of fluorescent tracers into Rga (e.g. CF 310) or Rgb (e.g. CF 274) labeled neurons in layer V of postsubiculum. Following anterior thalamic nuclei injections (e.g. CF 280, 286) layer VI neurons were labeled in postsubiculum. Afferent connections. Injections of retrogradely transported tracers in the postsubiculum (e.g. CF 385; Figs. 5D and 6A) labeled both cortical and subcortical neuronal cell bodies (Fig. 5). Ipsilateral cortical label was
172
Fig. 6. Fluorescent ( A - D ) and dark field (E) photomicrographs of the limbic cortex to demonstrate the position of labeled cells in A - D , or axons and terminals in E. A shows an FG injection into postsubiculum; B demonstrates labeled neurons in the AD and AV nuclei of the anterior thalamus; C shows labeled neurons in the nucleus reuniens and D demonstrates labeled pyramidal layer neurons in the septal third of area of CA1 following an FG injection into ipsilateral postsubiculum. E shows the label in the postsubiculum following an injection of [3H}amino acids in the A D nucleus of the thalamus. Scale bars = 250/~m.
173 present in area CA1, in subiculum, presubiculum, parasubiculum, in the entorhinal perirhinal, retrosplenial and infraradiata cortices; contralaterally the postsubicular, presubicular, perirhinal, retrosplenial, and infraradiata cortices were labeled. Subcortically, neurons were labeled in the horizontal and vertical limbs of the diagonal band of Broca, the medial septal nucleus, the rostral claustrum, in the AD, AV, LD, and the reuniens nuclei of the thalamus, the supramamillary region, the dorsal and ventral raphe nuclei, and the locus coeruleus. In the subiculum, cell bodies were labeled in the dorsal, septal third of the subiculum ipsilateral to the injection (Fig. 5D). The postsubiculum injections also ipsilaterally labeled pyramidal layer neurons (Figs. 5C and 6D) and a few non-pyramidal neurons in the septal third of area CA1, and there were few (< 5%) contralaterally (compared to ipsilaterally) labeled cells. In the presubiculum ipsilateral and contralateral to the injection site, neurons were labeled in the external lamina (i.e. layers II and III; Fig. 5E), but in the parasubiculum only a few neurons were labeled in the external lamina. Labeled neurons were confined to caudal segments of the presubiculum and parasubiculum. In the caudal entorhinal cortex, neuronal cell bodies were labeled in layers II and III (Fig. 5E), and in the caudal perirhinal cortex layers II, III and V contained labeled neurons bilaterally A
F
Fig. 7. Line drawings to demonstrate the position of labeled neuronal cell bodies in the thalamus following an injection of FB in the middle part of the postsubiculum.
[SUB ICA t I
II III IV V
":':'::=7:
-
..
