Brain Research, 565 (1991) 290-300 © 1991 Elsevier Science Publishers B.V. All rights reserved. 0006-8993/91/$03.50 ADONIS 0006899391
290
BRES 17191
Morphological heterogeneity within the cingulate cortex in rat: a horseradish peroxidase transport study Deyan Zeng and Sherry L. Stuesse Neurobiology Department, N.E. Ohio Universities College of Medicine, Rootstown, OH 44272 (U.S.A.) (Accepted 9 July 1991)
Key words: Cingulate cortex; Limbic system: Autonomic system: Medial frontal cortex: Prefrontal cortex; Rat
We compared the connections of two areas within rat cingulate cortex, the Cgl/Cg2 area vs the Cg3 area, by iontophoresing small quantities of wheatgerm agglutinin-horseradish peroxidase (WGA-HRP) into either of these two divisions and identifying afferent and efferent connections. Cortical projections were more widespread for the cingulate cortex (Cg3) area than for the Cgl/Cg2 area and included the dysgranular and agranular insular cortex, and perirhinal cortex. The Cg3 area received input from the CA1 layer of the hippocampus while the Cgl/Cg2 area was interconnected primarily with retrosplenial cortex, in the brainstem, both received input from Barrington's nucleus, however, many of the subcortical connections of the two areas differed and supported the hypothesis that the Cg3 area is part of the limbic and visceral motor system while the Cgl/Cg2 area is more closely allied with somatic motor control. The Cg3 area received input from the basolateral nucleus of the amygdala, the supramammillary hypothalamie nucleus, the laterodorsal tegmental nucleus, and the lateral parabrachial nucleus. The Cgl/Cg2 area received input from the substantia nigra and targeted deep layers of the superior colliculus. Thus, rat cingulate cortex is a heterogeneous area that can be further subdivided into separate limbic/autonomic (Cg3) and somatic motor areas (Cgl/ Cg2). INTRODUCTION
The medial frontal cortex of rat encompasses a large portion of frontal cortex and has been subdivided in various ways. Based on cytoarchitectonics, Zilles 4H divided the rat medial frontal cortex into 3 basic areas: frontal cortex, area 2 (Fr2), cingulate cortex (Cg), and infra. limbic cortex (IL). Cingulate t;ortex was further subdivided into 3 areas, Cgl through Cg3. Also based on cytoarchitectonics, Krettek and Price ts divided the cingulate cortex of rats into two regions: the anterior cingulate area and the prelimbic area. Krettek and Price further subdivided the anterior cingulate area into a dorsal division (ACd) and a ventral division (ACv) which correspond well with the Cgl and Cg2 areas delineated by Zilles4s. The prelimbic area coresponds to Zilles' Cg3 area and a small portion of the very rostral end of the Cgl area. The Cgl/Cg2 and Cg3 areas in rat probably correspond to Brodmann's area 24 and area 32 in humans, respectively 18.44,4~. The medial frontal cortex of rat may subserve a number of functions such as control of eye movement 6'9 and eye-head coordination 2s, attention and m e m o r y t4-t6'24. 30 and learning 23'46. Part of the medial frontal cortex
may be invol;'cd in the control of autonomic functions such as regulation of gastric motility tt and circulation 2' .~,3v,43. There likely are subdivisions within this part of medial frontal cortex which subserve this autonomic control, Microstimulation of the infralimbic (IL) and Cg3 areas of medial frontal cortex in Ketamine-anesthetized rats significantly inhibits gastric motility 1~. Burns and Wyss~ demonstrated that electrical stimulation in the Cg2 and Cg3 areas alters blood pressure in awake or anesthetized rats. In pentobarbital anesthetized rats, stimulation significantly decreased blood pressure. In awake rats, stimulation of the rostral third of the eingulate cortex increased blood pressure, while stimulation in the caudal third of the cingulate cortex decreased it, Stimulation of the middle third of the cingulate cortex produces pressor responses followed by depressor responses 5, In both anesthetized and unanesthetized rats, stimulation of the Cg3 area inhibits cardiovascular components of the defence reaction produced by stimulation of the basolateral amygdaloid nucleus or the hypothalamus 2. Bilateral excitotoxic lesions of the Cg3 and infralimbic areas in rat reduce the sensitivity of the baroreceptor heart rate reflex43. Bradycardia can be elicited by stimulating the infralimbic area in both awake and anes-
Con'espondence: D. Zeng, Ncurobiology Department, N.E. Ohio Universities College of Medicine, Rootstown, Ohio 44272-0095, U.S.A. Fax: (I)(216) 325-0522.
291
(
e
Frl
l /
CPu
Fig. 1. Location and size of Cgl/Cg2 injection sites are shown. The photo shows an actual injection site which is approximately 0.2 mm rostral to Bregma. The arrow shows the labeling in the corresponding area of the opposite hemisphere. The drawing is modified from Paxinos and Watson's brain atlas29. thetized rats 39. The variability of autonomic responses with electrode position suggests that within the cingulate portion of the medial frontal cortex, the pathways for autonomic control must vary. Results from anatomical studies have corroborated the role of the cingulate cortex in autonomic control. In rats, the medial frontal cortex projects directly to the solitary nucleus. This projection is mainly from the infralimbic area, although to a lesser extent, it originates from the Cg3 area 3s'4°-42. This pathway may play an important role in controlling autonomic activity because the solitary nucleus receives visceral input from the periphery and is an integrating center for autonomic reflexes. The Cg3 areas of the medial frontal cortex also project to the laterodorsai tegmental nucleus (LDTg) 32, a pontine micturition center involved in a micturition
reflex 3~. Thus initial anatomical studies and physiological studies support the contention that there may be subdivisions within the cingulate cortex with distinct anatomical connections. We chose two regions of medial frontal cortex (Cgl/ Cg2 and Cg3, Zilles' subdivisions) 4s and compared the connections of these areas. The recent study by Sesack et al. 32 utilized an anterograde tracer and described projections from the Cg3 area (prelimbic) in detail. However, she did not specifically compare projections from the Cgl/Cg2 vs the Cg3 area, and she did not describe afferent projections. Terreberry and Neafsey utilized wheatgerm agglutinin-horseradish peroxidase (WGAHRP) as a tracer, but their injections covered both infralimbic and the Cg3 areas 4°. Our study is similar to a portion of the excellent study by Vogt and Miller 44 ex-
Fr2
I Frl
Fr3 \
\
Fig. 2. Location and size of Cg3 injection sites are shown. The photo shows an actual injection site which is approximately 3.7 mm rostral to Bregma. The arrow indicates the labeling in the corresponding area of the opposite hemisphere. The drawing is modified from Paxinos and Watson's atlas2~.
292
o~
~•o
.~o~,
" •
.
m
.>db~ ¸ ¸°
Fig. 3. Cortical label resulting from injections into the Cgl/Cg2 area. Left photograph shows a Nissl stain of retrosplenial granular cortex and retrosplenial agranular cortex (separated by the dotted line) in bright field microscopy. Middle photograph shows labeled cells in retrosplenial granular cortex in polarized field microscopy. Right photograph shows labeled cells in retrosplenial agranular cortex in polarized field microscopy. Bars = 150 .urn.
cept that our injection sites are smaller, and we have included an analysis of subcortical connections. Beckstead 4 carried out an earlier autoradiographic study of efferent projections from this area. We made small injections of WGA-HRP, a tracer which is transported in both retrograde and ant :rograde directions 24, in each of these two
by use of Mesulam's 24 tetramethylbenzidene procedure. One to two series were used for this study. The remaining series were not used. The sections were counterstained with thionin t or dehydrated and left unstained. The sections were examined under both bright field and polarized light microscopy. To standardize comparisons among animals, labeled fibers and cells were drawn by hand on a series of transverse sections faken from the Paxinos and Watson atlas ~, Selected photographs were made of the results.
areas. MATERIALS AND METHODS We made 6 successful single injections in the CgI/Cg2 area and 5 successful injections in the Cg3 area of 13, male Spragun-DawIcy rats (200-225 g). The rats were anesthetized with Nembutal (45 mg/kg, i.p.) and supplemental doses were given when necessary. The rats were placed in a Kopf stereotaxic head holder, and the skull was exposed. A small hole was drilled in the right side of the skull at approximately 0.2 mm or 3.7 mm rostral to Bregma and a glass micropipette (tip diameter 8-10 ~um) filled with 4% WGAHRP (Sigma, buffered to pH 8.0 with sodium phosphaie solution) was carefully lowered into the brain. WGA-HRP was iontophoresed into either the Cgl/Cg2 area (0.2 mm rostral to Bregma, 0,5 mm lateral to the midline and 2 mm depth from the surface of the brain) or the Cg3 area (3.7 mm rostral to Bregma, 0,5 mm lateral to the midline and 3.0 mm depth) based on Zillis' coordinates"* for these regions. After 48-72 h survival, the rats were reanesthetized with overdoses of Nembutal, perfused intracardially, and their brains removed, The brains were cut transversely in 5 series at 50 ,urn thickness and processed for horseradish ~roxidase localization
RESULTS
Injection sites Small injection sites were localized to the Cgl/Cg2 or Cg3 areas. The Cgl/Cg2 area injection sites were approximately 0.8 × 1.4 mm in diameter, and they were just dorsal to the corpus callosum (Fig. 1). Collectively the injections covered the entire extent of the Cgl/Cg2 area. Within the injection sites, WGA-HRP labeled fibers formed networks, and labeled neurons were scattered between the fibers. In the superficial layers, the label was mainly within small pyramidal cells, while in the deep layers, it was mainly in larger pyramidal cells. At the ventrolateral corner of the injection site, the fibers exiting from the Cgl/Cg2 area bifurcated into two
-.), Fig. 4. Brain labeling resulting from CgllCg2 injections, depicted in a series of 50/~m transverse sections, The open circles represent labeled cells, Small dots and lines represent labeled fibers, Sections are modified from the Paxinos and Watson atlas 29 The level of the section relative to Bregma is given in the lower right by each section, See list for abbreviations,
293
-5.20
F
- 7.80
G
-8.30
H -9.30
-2.30
294
W
w
~
CPu
o/i
CI
%w
0.20
G
"O.U4
-8,80
H
-9.16 -2.30
295 bundles (Fig. 1). One bundle of fibers traveled through the dorsal portion of the corpus callosum from the right hemisphere to the left hemisphere, and the second bundle traversed the corpus cailosum to enter the caudateputamen from the dorsal border. Each of the Cg3 area injections (Fig. 2) covered the entire Cg3 area and spread to the dorsal edge of the infralimbic area. Within the injection sites, small pyramidal cells were labeled in layer 3 and large pyramidal cells were in layer 5. The labeled fibers formed a network surrounding the labeled cells. The fibers left the Cg3 area and part of them coursed along the ventral part of the corpus callosum from the right hemisphere into the left hemisphere (Fig. 5A). The other part of the fibers entered the caudate-putamen at its medial border and traversed in the ventromedial part of the caudate-putamen (Fig. 5A,B) to join the internal capsule.
Cerebral cortex In the contralateral Cgl/Cg2 area, extensively labeled neurons and fibers were seen mainly in layers 1 through 3 (Fig. 1). Labeled fibers and cells were also seen in the ipsilateral retrosplenial granular cortex and the retrospleniai agranular cortex (Fig. 3). The fibers terminated predominantly in layers 1 and 5 while the cells were mostly in layers 3 and 5. The Cg3 area had more extensive cortical interconnections than the Cgl/Cg2 area. In the contralateral Cg3 area, labeled cells and fibers were mainly in layers 1-3 (Fig. 2). Some labeled cells were also present in dysgranular insular cortex, agranular insular cortex (Fig. 5B), and perirhinal cortex (Fig. 5E), predominantly ipsUaterally. Pyramidal cells in the ipsilateral CA1 layer of the hippocampus were clearly labeled (Fig. 6B). Other forebrain connections Fibers from the Cgl/Cg2 area formed several bundles that bilaterally traversed the caudate-putamen. Most of them passed through the striatum without terminating, but a few axons terminated in the ipsilateral, dorsal part of the caudate-putamen (Fig. 4A). The nucleus of the horizontal limb of the diagonal band contained labeled cells ipsilaterai to the injection site (Fig. 4A). Labeled cells in the claustrum were predominantly ipsilateral while labeled fibers were bilateral (Fig. 4A). After injections into the Cg3 area, the ipsilateral basolateral amygdaloid nucleus contained dense labeled cells and terminal fibers (Fig. 6A), while only a few la-
beled fibers were present on the contralateral side (Fig. 5D). The Cg3 area was reciprocally connected to the claustrum (Fig. 5A). The claustrum label due to injections into Cg3 was indistinguishable from that observed from Cgl/Cg2 injections. The Cg3 area received projections from the ipsilateral ventral pallidum and the ipsilateral nucleus of the horizontal limb of the diagonal band (Fig. 5B).
Diencephalon Fibers in the thalamus which originated from Cgl/Cg2 or Cg3 injection sites descended bilaterally but were denser on the ipsilateral side. After injections into the Cgl/Cg2 area, a dense cluster of labeled cells and terminal fibers was seen in the anteromedial thalamic nucleus (Fig. 4B). Less dense but still substantially labeled cells and terminal fibers were observed in ventrolateral and ventromedial thalamic nuclei (Fig. 4C). A few labeled cells and terminal fibers were also present in the laterodorsal thalamic nucleus, the lateral part of the mediodorsal thalamic nuclei (Fg. 4D), and the reticular thalamic nucleus (Fig. 4B). Labeled fibers terminated heavily in the zona incerta (Fig. 4D). The sole connections to the hypothalamus consisted of a few cells in the lateral hypothalamic area. In contrast to the Cgl/Cg2 area, injections into the Cg3 area yielded dense fiber labeling and many cells in the medial part of the mediodorsal thalamic nucleus (Fig. 5D) and in the parateniai thalamic nucleus (Fig. 5C). The centromediai and anteromedial thalamic nuclei were reciprocally connected to the Cg3 area (Fig. 5C), but the nuclei were not as heavily labeled as the mediodorsal and paratenial thalamic nuclei. The reuniens and ventral reuniens thalamic nuclei contained lightly labeled cells and sparse fibers (Fig. 5C,D). The hypothalamus contained a few, lightly labeled cells in the supramammillary nucleus and few labeled cells in the lateral hypothalamic area. Midbrain
Fibers from the Cgl/Cg2 injectionsitedescended caudally within the medial portion of the internal capsule, into the cerebral peduncle. At the midbrain level,some of the fibers projected to deep layers of the superior colliculus. This projection is ipsilateral and forms patches mainly within layers 3 and 5 (Fig. 4F). Fibers as well as cells were present within the median and dorsal raphe nuclei (Fig. 4G). The substantia nigra contained several
Fig. 5. Brain labeling resulting from Cg3 injections, depicted in a series of 50/~m transverse sections. See the legend of Fig. 4 for f, rther details.
297 Fig. 6. A-D: examples of label in the brain which resulted after injections into the Cg3 area. A: labeled cells and terminal fibers in the basolateral amygdaloid nucleus. B: labeled cells in the CA1 layer of the hippocampus (arrow). C: labeled cells and terminal fibers in the laterodorsal tegmental nucleus ventral to the fourth ventricle. D: labeled cells in the lateral parabrachial nucleus (arrows) dorsolateral to the superior cerebellar peduncle, in lower right of picture. Polarized light microscopy. Bars -- 500/~m.
clearly labeled cells (Fig. 4E). Fibers from the Cg3 area projected to the central gray and the dorsal raphe nucleus. In turn, the Cg3 area received projections from the dorsal raphe nucleus, the ventral tegmental area, and the rostral linear nucleus of the raphe (Fig. 5F, G). Occasional labeled cells were present in all of the latter 3 locations, Potl$
Barrington's nucleus projected ipsilaterally to both Cgl/Cg2 and Cg3 regions; a few labeled cells were present in Barrington's nucleus after injections in either area (Figs. 4H and 5H). In addition, injections into the Cg3 area yielded many bilaterally labeled cells in the laterodorsal tegmental nucleus (Fig. 6C) and a few cells in the lateral parabrachial nucleus (Fig, 6D). Medulla
After either CgllCg2 or Cg3 injections, no labeled cells or fibers could be seen in the medulla.
DISCUSSION In the experiments reported herein, we studied the rat cingulate cortex, located on the medial wall of the cerebral hemisphere rostral to the genu of the corpus caliosum. We used WGA-HRP so that we could trace both anterograde and retrograde connections while confining the injections to one of the two cytoarchitectonically defined regions of the cingulate cortex. The disadvantage of using WGA-HRP is that for small injections, afferent and efferent connections that are sparse may result in light label that might be overlooked in the analysis. The advantage is that we will be focusing on those connections that are the most substantial in terms of numbers of connections. A second disadvantage is that we do not know the size of the effective injection site from which transport occurred. It is known that processing with tetramethyibenzidene overstates the injection site size so the site of transport is problably only from the densest, center portion of the injection site 25. This second disadvantage does not negate the descriptions and conclusions drawn in this paper, but the reader should be aware that the injection sites are probably smaller than those actually illustrated.
Somatic motor connections
The CgllCg2 area had reciprocal connections with retrosplenial granular and retrosplenial agranular coi~ex. These connections are in agreement with previous studies4'44. The retrosplenial granular cortex and the retrosplenial agranular cortexmay be involved in the control of eye movements because these areas have reciprocal connections with primary and secondary visual cortex 44, and with medial agranular cortex (Fr2) 35. The medial agranular cortex has extensive connections to visuomotor areas and may contain the rat homologue of the primate frontal eye fields6'9'35'44. In contrast to the Cgl/Cg2 area, injections into the Cg3 area did not yield label in the retrosplenial granular and retrospleniai agranular cortex. This negative finding confirms that of Vogt and Miller~. However, in her study with a lectin, Sesack et al. 3~ describe some projections from the Cg3 area to retrosplenial cortex but say that the label is lighter in this area of cortex than in more ro~tral cortical areas. In the brainstem, the Cgl/Cg2 area receives projections from substantia nigra pars compacta while the Cg3 area does not. This observation is in agreement with Beckstead's ~ results who found that rostrai prefrontal cortical areas did not receive projections from the substantia nigra but that the area infraradiate (corresponding in location to the Cgl/Cg2 area) did. The substantia nigra contains dopaminergic ncarons which project to the basal ganglia and plays an important role ill Iocon~otor control. Another connection of the Cgl/Cg2 area which could be classified as somatic motor is the projection to deep layers of the superior colliculus. Deep layers of the superior colliculus integrate sensory stimuli from visual, auditory and somatosensory areas and translate the sensory input into motor commands for orienting eye movement 34. The Cg3 area does not connect to the deep layers of superior colliculus. Using autoradiography, Beckstead 4 reported that all parts of the medial part of the mediodorsal cortex projected to the superior colliculus, but that the label was heavier from more caudal injections. Our Cg3 injection is 3.7 mm rostral to the Cgl/Cg2 injection site, and this rostral location may partially explain our observation. Sesack et al. 32 found that injections into the Cg3 area yielded very few terminal fibers in the superior colliculus while injections which overlapped the Cgl/Fr2 region projected heavily to the superior colliculus. Our results are in
298 agreement with Sesack's. One cannot differentiate between terminal fibers and fibers of passage by use of autoradiography, so the different techniques used may also lead to discrepancies when label is light. These extensive somatic motor connections of the Cgl/Cg2 area suggest that the Cgl/Cg2 area may be involved in motor control. We were unable to confirm the connections between visual areas 17 and 18b described previously 4~. Because medial agranular cortex, just dorsal to Cgl, has extensive connections with visual cortex 44, this difference between our results and Vogt and Miller's could be explained if their HRP injection sites included a portion of the medial agranular cortex. Limbic-autonomic axis connections
The Cg3 area has connections with several structures which are included in the limbic system. It has reciprocal connections with the basolateral amygdaloid nucleus as previously reported by others 4'32. Unilateral lesions of the basolateral amygdaloid nucleus in rat produce a disturbance of certain conditioned-reflex behavior 33. The Cg3 region receives afferent input from the CAI layer of the hippocampus, as described previously ~''2'3~. The Cg3 region also received projections from the supramammillary nucleus. This connection has not been reported before, in anterograde tracing studies, Haglund et al. found that the supramammillary nucleus projects to all parts of the hippocampal formation s, We also found that the Cg3 area received afferent inpt, t from perirhinal cortex, a connection which has not been previously reported. The efferent connection between perirhinal cortex and Cg3 has b,~en described 4, The perirhinal cortex has a direct projection to the CAI layer in the hippocampus in monkeys~r', and bilateral lesions of the perirhinal cortex in monkey produce severe memory ira. pairment s°, Thus perirhinal cortex is included in the limbic system. The Cgl/Cg2 area was not connected to any of these limbic structures. The Cg3 region is also connected to regions of the brain which process visceral sensory information, For example, the Cg3 region received projections from dysgranular insular cortex and agranular insular cortex. The latter connection has been previously reported z'~. The insular cortex has a direct connection with the nucleus of the solitary tract 4z,42and is involved in the processing of taste information z~. In rats, a strongly aversive taste stimuli that animals had been trained to avoid increased the level of the calcitonin gene-related peptide-like immunoreactivity within the agranular and dysgranular insular cortices 4~. Insular cortex may serve as visceral cortex involved in the integration of gustatory and limbic information since it has extensive connections with other gustatory areas such as the parabrachial nucleus 2~, the
ventromedial thalamic nucleus (which is visceral sensory), and the mediodorsal thalamic nucleus (limbic) 19. In the brainstem, we found that the Cg3 area received bilateral afferent projections from the lateral parabrachial nucleus. This connection has not been described before. The lateral parabrachial nucleus receivds baroreceptor information ~3 and may be a center for cardiovascular processing in rat ~°'4s. Electric stimulation of the lateral parabrachial nucleus increases arterial blood pressure while stimulation by the excitatory amino acid, glutamate, causes either a pressor-depressor or a depressor response 26. We also found that the Cg3 area received projections from the laterodorsal tegmental nucleus. The Cg3 area projected to the iaterodorsal tegmentai nucleus, in agreement with Sesack et al. 32. The laterodorsai tegmental nucleus has direct descending projections to the nucleus tractus solitarius and to the sacral intermediolateral cell column 22, and is involved in a micturition reflex ~1. We were unable to confirm the projection from cingulatc cortex to the nucleus of the solitary tract. Terreberry and Neafsey 3s'4° and Krushel and van der Kooy 19 have described this projection in rats. We found no label in the solitary complex after either Cgl/Cg2 or Cg3 area injections. However, most of the input to the solitary complex originates in the inffalimbic area ventral to the cingulate cortex, and the projections from the Cg3 area to the nucleus of the solitary tract may be sparse '9''~s'40'4'. Thus, we may have missed it. The extensive anatomical connections of the Cg3 area with the other visceral centers, such as dysgranular and agranular insular cortices, the lateral parabrachial nucleus, the laterodorsal tegmental nucleus, as well as nucleus tractus solitarius, indicate that the Cg3 area is probably involved in autonomic function. A few of these visceral sensory connections, such as to Barrington's nucleus, were also seen after injections into either the Cgl/Cg2 area or the Cg3 area, thus there may be a slight overlap in function between the two cingulate cortical regions. Other connections Although we have so far emphasized differences in connections in the two areas, there are a number of connections of the two cingulate area~ which are similar. For example, both areas project to the mediodorsal thalamic nucleus, although to separate regions of it. This connection h~s been described by many other investigators starting with Leonard 2°,2', and Krettek and Price TM who thoroughly studied the mediodorsal thalamic nucleus' cortical projection field, which includes cingulate cortex. Leonard2° subdivided the mediodorsal thalamic nucleus into 3 segments: the anterior, internal, and external segments which correspond to the medial, central, and lab eral segments of Krettek and Pricela. Krettek and Price
299 demonstrated that the Cgl area is mainly interconnected with the central and lateral segments of the mediodorsal thalamic nucleus while the Cg3 area is mainly connected with the medial segment, Our connections with the mediodorsal thalamic nucleus corroborate these earlier studies, In addition, both cingulate areas were reciprocally connected with the ventromedial thalamic nucleus 3"2] and to the anteromedial thalamic nucleus 3"4. In summary, the rat cingulate cortex has been previously subdivided based on cytoarchitectonics. These differences in cytoarchitectonics probably underlie differences in function which have been previously reported for the various regions 2'5'11"14-16'39'43'44. Sesack et al.
have emphasized the functional heterogeneity within the medial prefrontal cortex 32. Our results extend theirs by concentrating on a smaller area of medial prefrontal cortex. Within the cingulate cortex, the Cg3 area should be considered as a visceral motor and limbic cortex. The Cgl/Cg2 area has connections that are more reminiscent of those of motor cortices. Although there is probably some overlap in function, these two cingulate areas had only a few projections in common.
ABBREVIATIONS
ic
2n
4V ac
AI AM Bar BLA CA1 Cgl-3 CO CI eLi CM cp CPu D! DR f 1~'1-3 LDTg LPB HDB
optic nerve 4th ventricle anterior commissure agranular insular cortex anteromediai thalamie nucleus Barrington's nucleus basolateral amygdaloid nucleus hippocampus CAI corpus callosum cingulate cortex, areas 1-3 central gray claustrum caudal linear nucleus of the raphe contromedial thalami¢ nucleus cerebral peduncle caudate putamen (striatum) dysgranular insular cortex dorsal raphe fornix frontal cortex, areas 1-3 laterodorsal tegmental nucleus lateral parabrachlal nucleus nucleus of the horizontal limb of the diagonal band
REFERENCES
1 Adams, J.C., Stabilizing and rapid thionin staining of TMBbased HRP reaction product, Neurosci. Lett., 17 (1980) 7-9. 2 AI Maskati, H. and Zbrozyna, A.W., Stimulation in prefrontal cortex area inhibits cardiovascular and motor components of the clarence reaction in rats, J. Auton. Nerv. Syst., 28 (1989) 117126. 3 Beckstead, R.M., Convergent thalamic and mesencephalic projections to the anterior medial cortex in the rat, J. Comp. New rol.. 166 (1976) 403-416. 4 Beekstead, R M., An autoradiographie examination of corticocortical and subcortical projections of the mediodorsal-projection (prefrontal) cortex of the rat, J. Comp. Neurol., 184 (1979) 43-62. 5 Burns, S. and Wyss, J.M., The involvement of the anterior cingulate cortex in blood pressure control, Brain Research, 340 (1985) 71-77. 6 Donoghue, J.F. and Wise, S.E, The motor cortex of the rat: cytoarchitecture and microstimulation mapping, J. Comp. Neurol., 212 (1982) 76-88. 7 Ferino, F., Thierry, A.M. and Glowinski, J., Anatomical and electrophysiological evidence for a direct projection from Ammen's horn to the medial prefrontal cortex in the rat, Exp.
Acknowledgements. Supported by Grant MH43363 (to S.S.) from the National Institutes of Health.
LDDM LV MDL MDM ml PRh Pn PT PY Re RSA RSG Rt SC scp SNC SNR SuM VL VM VP VRe WGA-HRP Zi
internal capsule laterodorsal thalamic nucleus, dorsomedial part lateral ventricle lateral part of the mediodorsal thalamic nucleus medial part of the mediodorsal thalamic nucleus medial lemniscus perirhinal cortex pontine nuclei paratenial thalamic nucleus pyramidal tract reuniens thalamic nucleus retrosplenial agranular cortex retrosplenial granular cortex reticular thalamic nucleus superior colliculus superior cerebellar peduncle substantia nigra, compact part substantia nigra, reticular part supramammillary nucleus ventrolateral thalamic nucleus ventromedial thalamic nucleus ventral pallidum ventral reuniens thalamic nucleus wheatgerm agglutinin-horseradish peroxidase zona incerta
Brain Res., 65 (1987) 421-426. 8 Haglund, L., Swanson, L.W. and K6hler, C., The projection of the supramammillary nucleus to the hippocampal formatio,: an immunohistochemical and anterograde transport study with the lectin PHA-L in the rat, J. Comp. Neurol., 229 (1984) 171185. 9 Hall, R.D. and Lindholm, E.E, Organization of motor and somatosensory neocortex in the albino rat, Brain Research, 66 (1974) 23-38. 10 Hubbard, J.W., Buchholz, R.A., Keeton, T.K. and Nathan, M.A., Parabrachial lesions increase plasma norepinephrine concentration, plasma renin activity and enhance baroreflex sensitivity in the conscious rat, Brain Research, 421 (1987) 226-234. 11 Hurley-Gius, K.M. a~d Neafsey, E.J., The medial frontal cortex and gastric motility: microstimulation results and their possible significance for the overall pattern of organization of rat frontal and parietal cortex, Brain Research, 365 (1986) 241-248. 12 Jay, T.M., Glowinski, J. and Thierry, A.M., Selectivity of the hippocampal projection to the prelimbic area of the prefrontal cortex in the rat, Brain Research, 505 (1989) 337-340. 13 Jhamandas, J.H., Aippersbach, S.E. and Harris, K.H., Cardiovascular influences on rat parabrachial nucleus: an electrophysiological study, Am. J. Physiol., 260 (~991) R225-R231. 14 Kesner, R.P. and Holbrook, T., Dissociation of item and older
300 spatial memory in rat following medial prefrontal cortex lesions, Neuropsychology. 25 (1987) 653-664. 15 Kesner, R.P., Retrospective and prospective coding of information: role of the medial prefrontal cortex, Exp. Brain Res., 74 (1989) 163-167. 16 Kesner, R.P., Memory for frequency in rats: role of the hippocampus and medial prefrontal cortex, Behav. NeuroL Biol., 53 (1990) 402-410. 17 Kitamura, R., Localization and gustatory responsiveness of cortical taste area in the hamster, J. Osaka Univ. Dent. See., 34 (1989) 213-231. 18 Krettek, J.E. and Price, J.L., The cortical projections of the mediodorsal nucleus and adjacent thalamic nuclei in the rat,/. Comp. Neurol., 171 (1977) 157-192. 19 Krnshel, L.A. and Van Der Kooy, D., Visceral cortex: integration of the mucosal senses with limbic information in the rat agranular insular cortex,/. Comp. Neurol., 270 (1988) 39-54. 20 Leonard, C.M., The prefrontal cortex of the rat. I. Cortical projection of the mediodorsal nucleus. !!. Efferent connections, Brain Research, 12 (1969) 321-343. 21 Leonard, C.M., The connections of the dorsomedial nuclei, Brain Behav. Evoi., 6 (1972) 524-541. 22 Loewy, A.D., Saper, C.B. and Baker, R.P., Descending projections from the pontine micturition center, Brain Research, 172 (1979) 533-538. 23 Markowitsch, H.J. and Pritzel, M., Comparative analysis of prefrontal learning functions in rats, cats, and monkeys, Psychol. Bull., 84 (1977) 817-837. 24 Mesulam, M.-M., Tetramethylbenzidine for horseradish peroxidase neurochcmistry: a noncareinogenic blue reaction product with superior sensitivity for visualizing neuronal afferents and efferents,/. Histochem. Cytochem., 26 (1978) 106-117. 25 Mesulam, M.-M., Tracing Neural Connections with Horseradish Peroxidase, Ch. 1, John Wiley, New York, 1982, pp. 1-151. 26 Miura, M, and Takayama, K., Circulatory and respiratory responses to glutamate stimulation of the lateral parabrachial nucleus of the e,at, J. Auton. Nerv. Syst., 32 (1991) 121-134. 27 Moga, M.M., Herbert, H., Hurley, K.M., Yasui, Y., Gray, T.S. and Saper, C.B., Organization of cortical, basal forebrain, and hypothalami¢ ,fferents to the parabrachial nucleus in the rat, J. Comp. Neurol.. 295 (1990) 624-661. 28 Neafsey, ILJ., Bold, E,L., Haas, G., Hurley.Guis, K,M., Quirk, G., Sievert, C.F, and Terreberry, R.R., The organization of the rat motor cortex: a microstimulation mapping study, Brain Res, Rev., 11 (1986) 77-96, 29 Paxinos, (3. and Watson, C., The Rat Brain in Stereotaxic Co. ordinates, 2nd edn., Academic Press, New York, 1986. 30 Reap, R.L., Relationship between prefrontal and limbic cortex: a comparative anatomical review, Brain Behav. EvoL, 25 (1984) 5-80. 31 Satoh, K., Shimizu, N., Tohyama, M. and Maeda, T., Localization of the micturition reflex center at the dorsolateral pontine tegmentum of the rat, Neurosci, Lett,, 8 (1978) "27-33, 32 Sesack, S.R., Scutch, A.Y,, Roth, R.H, and Bunney, B.S., Topographical organization of the efferent projections of the medial prefrontal cortex in the rat: an antcrograde tract-tracing study with Phaseolus vuigaris leucoagglutinin, J. Comp, New rol., 290 (1989) 213-242.
33 Sidomv, B.M. and Sosina, V.D., Eiectrophysiological and behavioral changes after destruction of the basolateral amygdaloid nucleus in rats, Neurosci. Behav. Physiol., 17 (1987) 375--379. 34 Sparks, D.L. and Hartwich-Young, R., The deep layers of the superior colliculus, Rev. Oculomotor Res., 3 (1989) 213-255. 35 Stuesse, S.L. and Newman, D.B., Projections from the medial agranular cortex to brain stem visuomotor centers in rats, Exp. Brain Res., 80 (1990) 532-544. 36 Suzuki, W.A. and Amaral, D.G., Cortical inputs to the CAI field of the monkey hippocampus originate from the perirhinal and parahippocampal cortex but not from area TE, Neurosci. Lett., 115 (1990) 43--48. 37 Swanson, L,W., A direct projection from Ammon's horn to prefrontal cortex in tbe rat, Brain Research, 217 (1981) 150154. 38 Terreberry, R.R. and Neafsey, E.J., Rat medial frontal cortex: a visceral motor region with a direct projection to the solitary nucleus, Brain Research, 278 (1983) 245-249. 39 Terreberry, R.R. and Neafsey, E.J., The effects of medial prefrontal cortex stimulation on heart rate in the awake rat, Soc. Neurosci. Abstr., 10 (1984) 614. 40 Terreberry, R.R. and Neafsey, EJ., The rat medial frontal cortex projects directly to autonomic regions of the brainstem, Brain Res. Bull., 19 (1987) 639-649. 41 Van Der Kooy, D., McGinty, J.F. and Koda, L.Y., Visceral cortex: a direct connection from prefrontal cortex to the solitary nucleus in rat, Neurosci. Left, 33 (1982) 123-127. 42 Van der Kooy, D., Koda, L.Y., McGinty, J.E, Gerfen, C.R. and Bloom, F.E., The organization of projections from the cortex, amygdala, and hypothalamt,s to the nucleus of the solitary tract in rat, J. Comp. Neurol., 224 (1984) 1-24. 43 Verberne, A.J., Lewis, S.J., Wetland, P.J., Beart, P.M., Jarrott, B., Christie, M.J. and Louis, WJ., Medial prefrontal cortical lesions modulate baroreflex sensitivity in the rat, Brain Research. 426 (1987) 243-249. 44 Vogt, B.A, and Miller, M.W., Cortical connections between rat cingulate cortex and visual, motor, and postsubicular cortices, J. Comp. Neurol., 216 (1983) 192-210. 45 Ward, D.G., Stimulation of the parabrachial with monosodium glutamate increases arterial pressure, Brain Research, 462 (1988) 383-390, 46 Wikmark, R.G.I~,, Divae, I, and Weiss, R,, Retention of spatial delayed alternation in rats with lesion in the frontal lobes, Brain Behav, F~vol., 8 (1973)329-339. 47 Yamamoto, 1"., Matsuo, R., lchikawa, H., Wakisaka, S., Akai, M,, Imai, Y,, Yonehara, N, and Inoki, R,, Aversive taste stimuli increase CGRP levels in the gustatory insular cortex of the rat, Neurosci. Lett., 112 (1990) 167-172. 48 Zilles, K., The Cortex o,f the Rat. A Stereotaxic Atlas, SpringerVoting, Berlin, 1985, 49 Zilles, K, and Wree, A,, Cortex: areal and laminar structure, in G, Paxinos (Ed,), The Rat Nervous System, Vol, I, Academic Press, New York, 1985, pp. 375-416, 50 Zola-Morgan, S,, Squire, L,R,, Amaral, D.G. and Suzuki, W,A,, Lesions of perirhinal and parahippocampal cortex that spare the amygdala and hippocampal formation produce severe memory impairment, I. Neurosci., 9 (1989) 4355-4370,
290
BRES 17191
Morphological heterogeneity within the cingulate cortex in rat: a horseradish peroxidase transport study Deyan Zeng and Sherry L. Stuesse Neurobiology Department, N.E. Ohio Universities College of Medicine, Rootstown, OH 44272 (U.S.A.) (Accepted 9 July 1991)
Key words: Cingulate cortex; Limbic system: Autonomic system: Medial frontal cortex: Prefrontal cortex; Rat
We compared the connections of two areas within rat cingulate cortex, the Cgl/Cg2 area vs the Cg3 area, by iontophoresing small quantities of wheatgerm agglutinin-horseradish peroxidase (WGA-HRP) into either of these two divisions and identifying afferent and efferent connections. Cortical projections were more widespread for the cingulate cortex (Cg3) area than for the Cgl/Cg2 area and included the dysgranular and agranular insular cortex, and perirhinal cortex. The Cg3 area received input from the CA1 layer of the hippocampus while the Cgl/Cg2 area was interconnected primarily with retrosplenial cortex, in the brainstem, both received input from Barrington's nucleus, however, many of the subcortical connections of the two areas differed and supported the hypothesis that the Cg3 area is part of the limbic and visceral motor system while the Cgl/Cg2 area is more closely allied with somatic motor control. The Cg3 area received input from the basolateral nucleus of the amygdala, the supramammillary hypothalamie nucleus, the laterodorsal tegmental nucleus, and the lateral parabrachial nucleus. The Cgl/Cg2 area received input from the substantia nigra and targeted deep layers of the superior colliculus. Thus, rat cingulate cortex is a heterogeneous area that can be further subdivided into separate limbic/autonomic (Cg3) and somatic motor areas (Cgl/ Cg2). INTRODUCTION
The medial frontal cortex of rat encompasses a large portion of frontal cortex and has been subdivided in various ways. Based on cytoarchitectonics, Zilles 4H divided the rat medial frontal cortex into 3 basic areas: frontal cortex, area 2 (Fr2), cingulate cortex (Cg), and infra. limbic cortex (IL). Cingulate t;ortex was further subdivided into 3 areas, Cgl through Cg3. Also based on cytoarchitectonics, Krettek and Price ts divided the cingulate cortex of rats into two regions: the anterior cingulate area and the prelimbic area. Krettek and Price further subdivided the anterior cingulate area into a dorsal division (ACd) and a ventral division (ACv) which correspond well with the Cgl and Cg2 areas delineated by Zilles4s. The prelimbic area coresponds to Zilles' Cg3 area and a small portion of the very rostral end of the Cgl area. The Cgl/Cg2 and Cg3 areas in rat probably correspond to Brodmann's area 24 and area 32 in humans, respectively 18.44,4~. The medial frontal cortex of rat may subserve a number of functions such as control of eye movement 6'9 and eye-head coordination 2s, attention and m e m o r y t4-t6'24. 30 and learning 23'46. Part of the medial frontal cortex
may be invol;'cd in the control of autonomic functions such as regulation of gastric motility tt and circulation 2' .~,3v,43. There likely are subdivisions within this part of medial frontal cortex which subserve this autonomic control, Microstimulation of the infralimbic (IL) and Cg3 areas of medial frontal cortex in Ketamine-anesthetized rats significantly inhibits gastric motility 1~. Burns and Wyss~ demonstrated that electrical stimulation in the Cg2 and Cg3 areas alters blood pressure in awake or anesthetized rats. In pentobarbital anesthetized rats, stimulation significantly decreased blood pressure. In awake rats, stimulation of the rostral third of the eingulate cortex increased blood pressure, while stimulation in the caudal third of the cingulate cortex decreased it, Stimulation of the middle third of the cingulate cortex produces pressor responses followed by depressor responses 5, In both anesthetized and unanesthetized rats, stimulation of the Cg3 area inhibits cardiovascular components of the defence reaction produced by stimulation of the basolateral amygdaloid nucleus or the hypothalamus 2. Bilateral excitotoxic lesions of the Cg3 and infralimbic areas in rat reduce the sensitivity of the baroreceptor heart rate reflex43. Bradycardia can be elicited by stimulating the infralimbic area in both awake and anes-
Con'espondence: D. Zeng, Ncurobiology Department, N.E. Ohio Universities College of Medicine, Rootstown, Ohio 44272-0095, U.S.A. Fax: (I)(216) 325-0522.
291
(
e
Frl
l /
CPu
Fig. 1. Location and size of Cgl/Cg2 injection sites are shown. The photo shows an actual injection site which is approximately 0.2 mm rostral to Bregma. The arrow shows the labeling in the corresponding area of the opposite hemisphere. The drawing is modified from Paxinos and Watson's brain atlas29. thetized rats 39. The variability of autonomic responses with electrode position suggests that within the cingulate portion of the medial frontal cortex, the pathways for autonomic control must vary. Results from anatomical studies have corroborated the role of the cingulate cortex in autonomic control. In rats, the medial frontal cortex projects directly to the solitary nucleus. This projection is mainly from the infralimbic area, although to a lesser extent, it originates from the Cg3 area 3s'4°-42. This pathway may play an important role in controlling autonomic activity because the solitary nucleus receives visceral input from the periphery and is an integrating center for autonomic reflexes. The Cg3 areas of the medial frontal cortex also project to the laterodorsai tegmental nucleus (LDTg) 32, a pontine micturition center involved in a micturition
reflex 3~. Thus initial anatomical studies and physiological studies support the contention that there may be subdivisions within the cingulate cortex with distinct anatomical connections. We chose two regions of medial frontal cortex (Cgl/ Cg2 and Cg3, Zilles' subdivisions) 4s and compared the connections of these areas. The recent study by Sesack et al. 32 utilized an anterograde tracer and described projections from the Cg3 area (prelimbic) in detail. However, she did not specifically compare projections from the Cgl/Cg2 vs the Cg3 area, and she did not describe afferent projections. Terreberry and Neafsey utilized wheatgerm agglutinin-horseradish peroxidase (WGAHRP) as a tracer, but their injections covered both infralimbic and the Cg3 areas 4°. Our study is similar to a portion of the excellent study by Vogt and Miller 44 ex-
Fr2
I Frl
Fr3 \
\
Fig. 2. Location and size of Cg3 injection sites are shown. The photo shows an actual injection site which is approximately 3.7 mm rostral to Bregma. The arrow indicates the labeling in the corresponding area of the opposite hemisphere. The drawing is modified from Paxinos and Watson's atlas2~.
292
o~
~•o
.~o~,
" •
.
m
.>db~ ¸ ¸°
Fig. 3. Cortical label resulting from injections into the Cgl/Cg2 area. Left photograph shows a Nissl stain of retrosplenial granular cortex and retrosplenial agranular cortex (separated by the dotted line) in bright field microscopy. Middle photograph shows labeled cells in retrosplenial granular cortex in polarized field microscopy. Right photograph shows labeled cells in retrosplenial agranular cortex in polarized field microscopy. Bars = 150 .urn.
cept that our injection sites are smaller, and we have included an analysis of subcortical connections. Beckstead 4 carried out an earlier autoradiographic study of efferent projections from this area. We made small injections of WGA-HRP, a tracer which is transported in both retrograde and ant :rograde directions 24, in each of these two
by use of Mesulam's 24 tetramethylbenzidene procedure. One to two series were used for this study. The remaining series were not used. The sections were counterstained with thionin t or dehydrated and left unstained. The sections were examined under both bright field and polarized light microscopy. To standardize comparisons among animals, labeled fibers and cells were drawn by hand on a series of transverse sections faken from the Paxinos and Watson atlas ~, Selected photographs were made of the results.
areas. MATERIALS AND METHODS We made 6 successful single injections in the CgI/Cg2 area and 5 successful injections in the Cg3 area of 13, male Spragun-DawIcy rats (200-225 g). The rats were anesthetized with Nembutal (45 mg/kg, i.p.) and supplemental doses were given when necessary. The rats were placed in a Kopf stereotaxic head holder, and the skull was exposed. A small hole was drilled in the right side of the skull at approximately 0.2 mm or 3.7 mm rostral to Bregma and a glass micropipette (tip diameter 8-10 ~um) filled with 4% WGAHRP (Sigma, buffered to pH 8.0 with sodium phosphaie solution) was carefully lowered into the brain. WGA-HRP was iontophoresed into either the Cgl/Cg2 area (0.2 mm rostral to Bregma, 0,5 mm lateral to the midline and 2 mm depth from the surface of the brain) or the Cg3 area (3.7 mm rostral to Bregma, 0,5 mm lateral to the midline and 3.0 mm depth) based on Zillis' coordinates"* for these regions. After 48-72 h survival, the rats were reanesthetized with overdoses of Nembutal, perfused intracardially, and their brains removed, The brains were cut transversely in 5 series at 50 ,urn thickness and processed for horseradish ~roxidase localization
RESULTS
Injection sites Small injection sites were localized to the Cgl/Cg2 or Cg3 areas. The Cgl/Cg2 area injection sites were approximately 0.8 × 1.4 mm in diameter, and they were just dorsal to the corpus callosum (Fig. 1). Collectively the injections covered the entire extent of the Cgl/Cg2 area. Within the injection sites, WGA-HRP labeled fibers formed networks, and labeled neurons were scattered between the fibers. In the superficial layers, the label was mainly within small pyramidal cells, while in the deep layers, it was mainly in larger pyramidal cells. At the ventrolateral corner of the injection site, the fibers exiting from the Cgl/Cg2 area bifurcated into two
-.), Fig. 4. Brain labeling resulting from CgllCg2 injections, depicted in a series of 50/~m transverse sections, The open circles represent labeled cells, Small dots and lines represent labeled fibers, Sections are modified from the Paxinos and Watson atlas 29 The level of the section relative to Bregma is given in the lower right by each section, See list for abbreviations,
293
-5.20
F
- 7.80
G
-8.30
H -9.30
-2.30
294
W
w
~
CPu
o/i
CI
%w
0.20
G
"O.U4
-8,80
H
-9.16 -2.30
295 bundles (Fig. 1). One bundle of fibers traveled through the dorsal portion of the corpus callosum from the right hemisphere to the left hemisphere, and the second bundle traversed the corpus cailosum to enter the caudateputamen from the dorsal border. Each of the Cg3 area injections (Fig. 2) covered the entire Cg3 area and spread to the dorsal edge of the infralimbic area. Within the injection sites, small pyramidal cells were labeled in layer 3 and large pyramidal cells were in layer 5. The labeled fibers formed a network surrounding the labeled cells. The fibers left the Cg3 area and part of them coursed along the ventral part of the corpus callosum from the right hemisphere into the left hemisphere (Fig. 5A). The other part of the fibers entered the caudate-putamen at its medial border and traversed in the ventromedial part of the caudate-putamen (Fig. 5A,B) to join the internal capsule.
Cerebral cortex In the contralateral Cgl/Cg2 area, extensively labeled neurons and fibers were seen mainly in layers 1 through 3 (Fig. 1). Labeled fibers and cells were also seen in the ipsilateral retrosplenial granular cortex and the retrospleniai agranular cortex (Fig. 3). The fibers terminated predominantly in layers 1 and 5 while the cells were mostly in layers 3 and 5. The Cg3 area had more extensive cortical interconnections than the Cgl/Cg2 area. In the contralateral Cg3 area, labeled cells and fibers were mainly in layers 1-3 (Fig. 2). Some labeled cells were also present in dysgranular insular cortex, agranular insular cortex (Fig. 5B), and perirhinal cortex (Fig. 5E), predominantly ipsUaterally. Pyramidal cells in the ipsilateral CA1 layer of the hippocampus were clearly labeled (Fig. 6B). Other forebrain connections Fibers from the Cgl/Cg2 area formed several bundles that bilaterally traversed the caudate-putamen. Most of them passed through the striatum without terminating, but a few axons terminated in the ipsilateral, dorsal part of the caudate-putamen (Fig. 4A). The nucleus of the horizontal limb of the diagonal band contained labeled cells ipsilaterai to the injection site (Fig. 4A). Labeled cells in the claustrum were predominantly ipsilateral while labeled fibers were bilateral (Fig. 4A). After injections into the Cg3 area, the ipsilateral basolateral amygdaloid nucleus contained dense labeled cells and terminal fibers (Fig. 6A), while only a few la-
beled fibers were present on the contralateral side (Fig. 5D). The Cg3 area was reciprocally connected to the claustrum (Fig. 5A). The claustrum label due to injections into Cg3 was indistinguishable from that observed from Cgl/Cg2 injections. The Cg3 area received projections from the ipsilateral ventral pallidum and the ipsilateral nucleus of the horizontal limb of the diagonal band (Fig. 5B).
Diencephalon Fibers in the thalamus which originated from Cgl/Cg2 or Cg3 injection sites descended bilaterally but were denser on the ipsilateral side. After injections into the Cgl/Cg2 area, a dense cluster of labeled cells and terminal fibers was seen in the anteromedial thalamic nucleus (Fig. 4B). Less dense but still substantially labeled cells and terminal fibers were observed in ventrolateral and ventromedial thalamic nuclei (Fig. 4C). A few labeled cells and terminal fibers were also present in the laterodorsal thalamic nucleus, the lateral part of the mediodorsal thalamic nuclei (Fg. 4D), and the reticular thalamic nucleus (Fig. 4B). Labeled fibers terminated heavily in the zona incerta (Fig. 4D). The sole connections to the hypothalamus consisted of a few cells in the lateral hypothalamic area. In contrast to the Cgl/Cg2 area, injections into the Cg3 area yielded dense fiber labeling and many cells in the medial part of the mediodorsal thalamic nucleus (Fig. 5D) and in the parateniai thalamic nucleus (Fig. 5C). The centromediai and anteromedial thalamic nuclei were reciprocally connected to the Cg3 area (Fig. 5C), but the nuclei were not as heavily labeled as the mediodorsal and paratenial thalamic nuclei. The reuniens and ventral reuniens thalamic nuclei contained lightly labeled cells and sparse fibers (Fig. 5C,D). The hypothalamus contained a few, lightly labeled cells in the supramammillary nucleus and few labeled cells in the lateral hypothalamic area. Midbrain
Fibers from the Cgl/Cg2 injectionsitedescended caudally within the medial portion of the internal capsule, into the cerebral peduncle. At the midbrain level,some of the fibers projected to deep layers of the superior colliculus. This projection is ipsilateral and forms patches mainly within layers 3 and 5 (Fig. 4F). Fibers as well as cells were present within the median and dorsal raphe nuclei (Fig. 4G). The substantia nigra contained several
Fig. 5. Brain labeling resulting from Cg3 injections, depicted in a series of 50/~m transverse sections. See the legend of Fig. 4 for f, rther details.
297 Fig. 6. A-D: examples of label in the brain which resulted after injections into the Cg3 area. A: labeled cells and terminal fibers in the basolateral amygdaloid nucleus. B: labeled cells in the CA1 layer of the hippocampus (arrow). C: labeled cells and terminal fibers in the laterodorsal tegmental nucleus ventral to the fourth ventricle. D: labeled cells in the lateral parabrachial nucleus (arrows) dorsolateral to the superior cerebellar peduncle, in lower right of picture. Polarized light microscopy. Bars -- 500/~m.
clearly labeled cells (Fig. 4E). Fibers from the Cg3 area projected to the central gray and the dorsal raphe nucleus. In turn, the Cg3 area received projections from the dorsal raphe nucleus, the ventral tegmental area, and the rostral linear nucleus of the raphe (Fig. 5F, G). Occasional labeled cells were present in all of the latter 3 locations, Potl$
Barrington's nucleus projected ipsilaterally to both Cgl/Cg2 and Cg3 regions; a few labeled cells were present in Barrington's nucleus after injections in either area (Figs. 4H and 5H). In addition, injections into the Cg3 area yielded many bilaterally labeled cells in the laterodorsal tegmental nucleus (Fig. 6C) and a few cells in the lateral parabrachial nucleus (Fig, 6D). Medulla
After either CgllCg2 or Cg3 injections, no labeled cells or fibers could be seen in the medulla.
DISCUSSION In the experiments reported herein, we studied the rat cingulate cortex, located on the medial wall of the cerebral hemisphere rostral to the genu of the corpus caliosum. We used WGA-HRP so that we could trace both anterograde and retrograde connections while confining the injections to one of the two cytoarchitectonically defined regions of the cingulate cortex. The disadvantage of using WGA-HRP is that for small injections, afferent and efferent connections that are sparse may result in light label that might be overlooked in the analysis. The advantage is that we will be focusing on those connections that are the most substantial in terms of numbers of connections. A second disadvantage is that we do not know the size of the effective injection site from which transport occurred. It is known that processing with tetramethyibenzidene overstates the injection site size so the site of transport is problably only from the densest, center portion of the injection site 25. This second disadvantage does not negate the descriptions and conclusions drawn in this paper, but the reader should be aware that the injection sites are probably smaller than those actually illustrated.
Somatic motor connections
The CgllCg2 area had reciprocal connections with retrosplenial granular and retrosplenial agranular coi~ex. These connections are in agreement with previous studies4'44. The retrosplenial granular cortex and the retrosplenial agranular cortexmay be involved in the control of eye movements because these areas have reciprocal connections with primary and secondary visual cortex 44, and with medial agranular cortex (Fr2) 35. The medial agranular cortex has extensive connections to visuomotor areas and may contain the rat homologue of the primate frontal eye fields6'9'35'44. In contrast to the Cgl/Cg2 area, injections into the Cg3 area did not yield label in the retrosplenial granular and retrospleniai agranular cortex. This negative finding confirms that of Vogt and Miller~. However, in her study with a lectin, Sesack et al. 3~ describe some projections from the Cg3 area to retrosplenial cortex but say that the label is lighter in this area of cortex than in more ro~tral cortical areas. In the brainstem, the Cgl/Cg2 area receives projections from substantia nigra pars compacta while the Cg3 area does not. This observation is in agreement with Beckstead's ~ results who found that rostrai prefrontal cortical areas did not receive projections from the substantia nigra but that the area infraradiate (corresponding in location to the Cgl/Cg2 area) did. The substantia nigra contains dopaminergic ncarons which project to the basal ganglia and plays an important role ill Iocon~otor control. Another connection of the Cgl/Cg2 area which could be classified as somatic motor is the projection to deep layers of the superior colliculus. Deep layers of the superior colliculus integrate sensory stimuli from visual, auditory and somatosensory areas and translate the sensory input into motor commands for orienting eye movement 34. The Cg3 area does not connect to the deep layers of superior colliculus. Using autoradiography, Beckstead 4 reported that all parts of the medial part of the mediodorsal cortex projected to the superior colliculus, but that the label was heavier from more caudal injections. Our Cg3 injection is 3.7 mm rostral to the Cgl/Cg2 injection site, and this rostral location may partially explain our observation. Sesack et al. 32 found that injections into the Cg3 area yielded very few terminal fibers in the superior colliculus while injections which overlapped the Cgl/Fr2 region projected heavily to the superior colliculus. Our results are in
298 agreement with Sesack's. One cannot differentiate between terminal fibers and fibers of passage by use of autoradiography, so the different techniques used may also lead to discrepancies when label is light. These extensive somatic motor connections of the Cgl/Cg2 area suggest that the Cgl/Cg2 area may be involved in motor control. We were unable to confirm the connections between visual areas 17 and 18b described previously 4~. Because medial agranular cortex, just dorsal to Cgl, has extensive connections with visual cortex 44, this difference between our results and Vogt and Miller's could be explained if their HRP injection sites included a portion of the medial agranular cortex. Limbic-autonomic axis connections
The Cg3 area has connections with several structures which are included in the limbic system. It has reciprocal connections with the basolateral amygdaloid nucleus as previously reported by others 4'32. Unilateral lesions of the basolateral amygdaloid nucleus in rat produce a disturbance of certain conditioned-reflex behavior 33. The Cg3 region receives afferent input from the CAI layer of the hippocampus, as described previously ~''2'3~. The Cg3 region also received projections from the supramammillary nucleus. This connection has not been reported before, in anterograde tracing studies, Haglund et al. found that the supramammillary nucleus projects to all parts of the hippocampal formation s, We also found that the Cg3 area received afferent inpt, t from perirhinal cortex, a connection which has not been previously reported. The efferent connection between perirhinal cortex and Cg3 has b,~en described 4, The perirhinal cortex has a direct projection to the CAI layer in the hippocampus in monkeys~r', and bilateral lesions of the perirhinal cortex in monkey produce severe memory ira. pairment s°, Thus perirhinal cortex is included in the limbic system. The Cgl/Cg2 area was not connected to any of these limbic structures. The Cg3 region is also connected to regions of the brain which process visceral sensory information, For example, the Cg3 region received projections from dysgranular insular cortex and agranular insular cortex. The latter connection has been previously reported z'~. The insular cortex has a direct connection with the nucleus of the solitary tract 4z,42and is involved in the processing of taste information z~. In rats, a strongly aversive taste stimuli that animals had been trained to avoid increased the level of the calcitonin gene-related peptide-like immunoreactivity within the agranular and dysgranular insular cortices 4~. Insular cortex may serve as visceral cortex involved in the integration of gustatory and limbic information since it has extensive connections with other gustatory areas such as the parabrachial nucleus 2~, the
ventromedial thalamic nucleus (which is visceral sensory), and the mediodorsal thalamic nucleus (limbic) 19. In the brainstem, we found that the Cg3 area received bilateral afferent projections from the lateral parabrachial nucleus. This connection has not been described before. The lateral parabrachial nucleus receivds baroreceptor information ~3 and may be a center for cardiovascular processing in rat ~°'4s. Electric stimulation of the lateral parabrachial nucleus increases arterial blood pressure while stimulation by the excitatory amino acid, glutamate, causes either a pressor-depressor or a depressor response 26. We also found that the Cg3 area received projections from the laterodorsal tegmental nucleus. The Cg3 area projected to the iaterodorsal tegmentai nucleus, in agreement with Sesack et al. 32. The laterodorsai tegmental nucleus has direct descending projections to the nucleus tractus solitarius and to the sacral intermediolateral cell column 22, and is involved in a micturition reflex ~1. We were unable to confirm the projection from cingulatc cortex to the nucleus of the solitary tract. Terreberry and Neafsey 3s'4° and Krushel and van der Kooy 19 have described this projection in rats. We found no label in the solitary complex after either Cgl/Cg2 or Cg3 area injections. However, most of the input to the solitary complex originates in the inffalimbic area ventral to the cingulate cortex, and the projections from the Cg3 area to the nucleus of the solitary tract may be sparse '9''~s'40'4'. Thus, we may have missed it. The extensive anatomical connections of the Cg3 area with the other visceral centers, such as dysgranular and agranular insular cortices, the lateral parabrachial nucleus, the laterodorsal tegmental nucleus, as well as nucleus tractus solitarius, indicate that the Cg3 area is probably involved in autonomic function. A few of these visceral sensory connections, such as to Barrington's nucleus, were also seen after injections into either the Cgl/Cg2 area or the Cg3 area, thus there may be a slight overlap in function between the two cingulate cortical regions. Other connections Although we have so far emphasized differences in connections in the two areas, there are a number of connections of the two cingulate area~ which are similar. For example, both areas project to the mediodorsal thalamic nucleus, although to separate regions of it. This connection h~s been described by many other investigators starting with Leonard 2°,2', and Krettek and Price TM who thoroughly studied the mediodorsal thalamic nucleus' cortical projection field, which includes cingulate cortex. Leonard2° subdivided the mediodorsal thalamic nucleus into 3 segments: the anterior, internal, and external segments which correspond to the medial, central, and lab eral segments of Krettek and Pricela. Krettek and Price
299 demonstrated that the Cgl area is mainly interconnected with the central and lateral segments of the mediodorsal thalamic nucleus while the Cg3 area is mainly connected with the medial segment, Our connections with the mediodorsal thalamic nucleus corroborate these earlier studies, In addition, both cingulate areas were reciprocally connected with the ventromedial thalamic nucleus 3"2] and to the anteromedial thalamic nucleus 3"4. In summary, the rat cingulate cortex has been previously subdivided based on cytoarchitectonics. These differences in cytoarchitectonics probably underlie differences in function which have been previously reported for the various regions 2'5'11"14-16'39'43'44. Sesack et al.
have emphasized the functional heterogeneity within the medial prefrontal cortex 32. Our results extend theirs by concentrating on a smaller area of medial prefrontal cortex. Within the cingulate cortex, the Cg3 area should be considered as a visceral motor and limbic cortex. The Cgl/Cg2 area has connections that are more reminiscent of those of motor cortices. Although there is probably some overlap in function, these two cingulate areas had only a few projections in common.
ABBREVIATIONS
ic
2n
4V ac
AI AM Bar BLA CA1 Cgl-3 CO CI eLi CM cp CPu D! DR f 1~'1-3 LDTg LPB HDB
optic nerve 4th ventricle anterior commissure agranular insular cortex anteromediai thalamie nucleus Barrington's nucleus basolateral amygdaloid nucleus hippocampus CAI corpus callosum cingulate cortex, areas 1-3 central gray claustrum caudal linear nucleus of the raphe contromedial thalami¢ nucleus cerebral peduncle caudate putamen (striatum) dysgranular insular cortex dorsal raphe fornix frontal cortex, areas 1-3 laterodorsal tegmental nucleus lateral parabrachlal nucleus nucleus of the horizontal limb of the diagonal band
REFERENCES
1 Adams, J.C., Stabilizing and rapid thionin staining of TMBbased HRP reaction product, Neurosci. Lett., 17 (1980) 7-9. 2 AI Maskati, H. and Zbrozyna, A.W., Stimulation in prefrontal cortex area inhibits cardiovascular and motor components of the clarence reaction in rats, J. Auton. Nerv. Syst., 28 (1989) 117126. 3 Beckstead, R.M., Convergent thalamic and mesencephalic projections to the anterior medial cortex in the rat, J. Comp. New rol.. 166 (1976) 403-416. 4 Beekstead, R M., An autoradiographie examination of corticocortical and subcortical projections of the mediodorsal-projection (prefrontal) cortex of the rat, J. Comp. Neurol., 184 (1979) 43-62. 5 Burns, S. and Wyss, J.M., The involvement of the anterior cingulate cortex in blood pressure control, Brain Research, 340 (1985) 71-77. 6 Donoghue, J.F. and Wise, S.E, The motor cortex of the rat: cytoarchitecture and microstimulation mapping, J. Comp. Neurol., 212 (1982) 76-88. 7 Ferino, F., Thierry, A.M. and Glowinski, J., Anatomical and electrophysiological evidence for a direct projection from Ammen's horn to the medial prefrontal cortex in the rat, Exp.
Acknowledgements. Supported by Grant MH43363 (to S.S.) from the National Institutes of Health.
LDDM LV MDL MDM ml PRh Pn PT PY Re RSA RSG Rt SC scp SNC SNR SuM VL VM VP VRe WGA-HRP Zi
internal capsule laterodorsal thalamic nucleus, dorsomedial part lateral ventricle lateral part of the mediodorsal thalamic nucleus medial part of the mediodorsal thalamic nucleus medial lemniscus perirhinal cortex pontine nuclei paratenial thalamic nucleus pyramidal tract reuniens thalamic nucleus retrosplenial agranular cortex retrosplenial granular cortex reticular thalamic nucleus superior colliculus superior cerebellar peduncle substantia nigra, compact part substantia nigra, reticular part supramammillary nucleus ventrolateral thalamic nucleus ventromedial thalamic nucleus ventral pallidum ventral reuniens thalamic nucleus wheatgerm agglutinin-horseradish peroxidase zona incerta
Brain Res., 65 (1987) 421-426. 8 Haglund, L., Swanson, L.W. and K6hler, C., The projection of the supramammillary nucleus to the hippocampal formatio,: an immunohistochemical and anterograde transport study with the lectin PHA-L in the rat, J. Comp. Neurol., 229 (1984) 171185. 9 Hall, R.D. and Lindholm, E.E, Organization of motor and somatosensory neocortex in the albino rat, Brain Research, 66 (1974) 23-38. 10 Hubbard, J.W., Buchholz, R.A., Keeton, T.K. and Nathan, M.A., Parabrachial lesions increase plasma norepinephrine concentration, plasma renin activity and enhance baroreflex sensitivity in the conscious rat, Brain Research, 421 (1987) 226-234. 11 Hurley-Gius, K.M. a~d Neafsey, E.J., The medial frontal cortex and gastric motility: microstimulation results and their possible significance for the overall pattern of organization of rat frontal and parietal cortex, Brain Research, 365 (1986) 241-248. 12 Jay, T.M., Glowinski, J. and Thierry, A.M., Selectivity of the hippocampal projection to the prelimbic area of the prefrontal cortex in the rat, Brain Research, 505 (1989) 337-340. 13 Jhamandas, J.H., Aippersbach, S.E. and Harris, K.H., Cardiovascular influences on rat parabrachial nucleus: an electrophysiological study, Am. J. Physiol., 260 (~991) R225-R231. 14 Kesner, R.P. and Holbrook, T., Dissociation of item and older
300 spatial memory in rat following medial prefrontal cortex lesions, Neuropsychology. 25 (1987) 653-664. 15 Kesner, R.P., Retrospective and prospective coding of information: role of the medial prefrontal cortex, Exp. Brain Res., 74 (1989) 163-167. 16 Kesner, R.P., Memory for frequency in rats: role of the hippocampus and medial prefrontal cortex, Behav. NeuroL Biol., 53 (1990) 402-410. 17 Kitamura, R., Localization and gustatory responsiveness of cortical taste area in the hamster, J. Osaka Univ. Dent. See., 34 (1989) 213-231. 18 Krettek, J.E. and Price, J.L., The cortical projections of the mediodorsal nucleus and adjacent thalamic nuclei in the rat,/. Comp. Neurol., 171 (1977) 157-192. 19 Krnshel, L.A. and Van Der Kooy, D., Visceral cortex: integration of the mucosal senses with limbic information in the rat agranular insular cortex,/. Comp. Neurol., 270 (1988) 39-54. 20 Leonard, C.M., The prefrontal cortex of the rat. I. Cortical projection of the mediodorsal nucleus. !!. Efferent connections, Brain Research, 12 (1969) 321-343. 21 Leonard, C.M., The connections of the dorsomedial nuclei, Brain Behav. Evoi., 6 (1972) 524-541. 22 Loewy, A.D., Saper, C.B. and Baker, R.P., Descending projections from the pontine micturition center, Brain Research, 172 (1979) 533-538. 23 Markowitsch, H.J. and Pritzel, M., Comparative analysis of prefrontal learning functions in rats, cats, and monkeys, Psychol. Bull., 84 (1977) 817-837. 24 Mesulam, M.-M., Tetramethylbenzidine for horseradish peroxidase neurochcmistry: a noncareinogenic blue reaction product with superior sensitivity for visualizing neuronal afferents and efferents,/. Histochem. Cytochem., 26 (1978) 106-117. 25 Mesulam, M.-M., Tracing Neural Connections with Horseradish Peroxidase, Ch. 1, John Wiley, New York, 1982, pp. 1-151. 26 Miura, M, and Takayama, K., Circulatory and respiratory responses to glutamate stimulation of the lateral parabrachial nucleus of the e,at, J. Auton. Nerv. Syst., 32 (1991) 121-134. 27 Moga, M.M., Herbert, H., Hurley, K.M., Yasui, Y., Gray, T.S. and Saper, C.B., Organization of cortical, basal forebrain, and hypothalami¢ ,fferents to the parabrachial nucleus in the rat, J. Comp. Neurol.. 295 (1990) 624-661. 28 Neafsey, ILJ., Bold, E,L., Haas, G., Hurley.Guis, K,M., Quirk, G., Sievert, C.F, and Terreberry, R.R., The organization of the rat motor cortex: a microstimulation mapping study, Brain Res, Rev., 11 (1986) 77-96, 29 Paxinos, (3. and Watson, C., The Rat Brain in Stereotaxic Co. ordinates, 2nd edn., Academic Press, New York, 1986. 30 Reap, R.L., Relationship between prefrontal and limbic cortex: a comparative anatomical review, Brain Behav. EvoL, 25 (1984) 5-80. 31 Satoh, K., Shimizu, N., Tohyama, M. and Maeda, T., Localization of the micturition reflex center at the dorsolateral pontine tegmentum of the rat, Neurosci, Lett,, 8 (1978) "27-33, 32 Sesack, S.R., Scutch, A.Y,, Roth, R.H, and Bunney, B.S., Topographical organization of the efferent projections of the medial prefrontal cortex in the rat: an antcrograde tract-tracing study with Phaseolus vuigaris leucoagglutinin, J. Comp, New rol., 290 (1989) 213-242.
33 Sidomv, B.M. and Sosina, V.D., Eiectrophysiological and behavioral changes after destruction of the basolateral amygdaloid nucleus in rats, Neurosci. Behav. Physiol., 17 (1987) 375--379. 34 Sparks, D.L. and Hartwich-Young, R., The deep layers of the superior colliculus, Rev. Oculomotor Res., 3 (1989) 213-255. 35 Stuesse, S.L. and Newman, D.B., Projections from the medial agranular cortex to brain stem visuomotor centers in rats, Exp. Brain Res., 80 (1990) 532-544. 36 Suzuki, W.A. and Amaral, D.G., Cortical inputs to the CAI field of the monkey hippocampus originate from the perirhinal and parahippocampal cortex but not from area TE, Neurosci. Lett., 115 (1990) 43--48. 37 Swanson, L,W., A direct projection from Ammon's horn to prefrontal cortex in tbe rat, Brain Research, 217 (1981) 150154. 38 Terreberry, R.R. and Neafsey, E.J., Rat medial frontal cortex: a visceral motor region with a direct projection to the solitary nucleus, Brain Research, 278 (1983) 245-249. 39 Terreberry, R.R. and Neafsey, E.J., The effects of medial prefrontal cortex stimulation on heart rate in the awake rat, Soc. Neurosci. Abstr., 10 (1984) 614. 40 Terreberry, R.R. and Neafsey, EJ., The rat medial frontal cortex projects directly to autonomic regions of the brainstem, Brain Res. Bull., 19 (1987) 639-649. 41 Van Der Kooy, D., McGinty, J.F. and Koda, L.Y., Visceral cortex: a direct connection from prefrontal cortex to the solitary nucleus in rat, Neurosci. Left, 33 (1982) 123-127. 42 Van der Kooy, D., Koda, L.Y., McGinty, J.E, Gerfen, C.R. and Bloom, F.E., The organization of projections from the cortex, amygdala, and hypothalamt,s to the nucleus of the solitary tract in rat, J. Comp. Neurol., 224 (1984) 1-24. 43 Verberne, A.J., Lewis, S.J., Wetland, P.J., Beart, P.M., Jarrott, B., Christie, M.J. and Louis, WJ., Medial prefrontal cortical lesions modulate baroreflex sensitivity in the rat, Brain Research. 426 (1987) 243-249. 44 Vogt, B.A, and Miller, M.W., Cortical connections between rat cingulate cortex and visual, motor, and postsubicular cortices, J. Comp. Neurol., 216 (1983) 192-210. 45 Ward, D.G., Stimulation of the parabrachial with monosodium glutamate increases arterial pressure, Brain Research, 462 (1988) 383-390, 46 Wikmark, R.G.I~,, Divae, I, and Weiss, R,, Retention of spatial delayed alternation in rats with lesion in the frontal lobes, Brain Behav, F~vol., 8 (1973)329-339. 47 Yamamoto, 1"., Matsuo, R., lchikawa, H., Wakisaka, S., Akai, M,, Imai, Y,, Yonehara, N, and Inoki, R,, Aversive taste stimuli increase CGRP levels in the gustatory insular cortex of the rat, Neurosci. Lett., 112 (1990) 167-172. 48 Zilles, K., The Cortex o,f the Rat. A Stereotaxic Atlas, SpringerVoting, Berlin, 1985, 49 Zilles, K, and Wree, A,, Cortex: areal and laminar structure, in G, Paxinos (Ed,), The Rat Nervous System, Vol, I, Academic Press, New York, 1985, pp. 375-416, 50 Zola-Morgan, S,, Squire, L,R,, Amaral, D.G. and Suzuki, W,A,, Lesions of perirhinal and parahippocampal cortex that spare the amygdala and hippocampal formation produce severe memory impairment, I. Neurosci., 9 (1989) 4355-4370,
