Involvement of AMPA/kainate-excitotoxicity in MK801-induced neuronal death in the retrosplenial cortex.

MK801 is a prototypical non-competitive NMDA receptor-antagonist that induces behavioural changes and reversible toxicity at low doses, while at higher doses triggers neuronal death that mainly affects the retrosplenial cortex (RSC) and to a lesser extent other structures such as the posterolateral cortical amygdaloid nucleus (PLCo). The mechanism of MK801-induced neurodegeneration remains poorly understood. In this study we analysed the participation of GABA-ergic and glutamatergic neurotransmission in MK801-induced neuronal death. We used a single i.p. injection of MK801 (2.5 mg/kg) that induced moderate neuronal death in the RSC and PLCo of female rats, and combined this treatment with the i.p., i.c.v., or intra-RSC infusion of drugs that are selective agonists or antagonists of the GABA-ergic or glutamatergic neurotransmission. We found that neuronal death in the RSC, but not the PLCo, was significantly reduced by the i.p. injection of thiopental, and the i.c.v. application of muscimol, both GABA-A agonists. MK801-toxicity in RSC was abrogated by intra-RSC infusion of muscimol, but the GABA antagonist picrotoxin had no effect. HPLC-analysis showed that levels of glutamate, but not GABA, in the RSC decreased after i.p. treatment with MK801. Intra-RSC infusion of MK801 did not enhance toxicity triggered by the i.p. injection of MK801, indicating that toxicity is not due to direct blockade of NMDA receptors in RSC neurons. MK801-toxicity in the RSC was abrogated by i.c.v. and intra-RSC infusions of the AMPA/kainate antagonist 6,7-dinitroquinoxaline-2,3-dione (DNQX). Interestingly, i.c.v. application of neither muscimol or DNQX inhibited MK801-toxicity in the PLCo, suggesting that the mechanism of neuronal death in the RSC and the PLCo might be different. 1-naphthylacetyl spermine trihydrochloride (NASPM), which blocks Ca2+ permeable AMPA/kainate receptors, also reduced MK801-induced toxicity in the RSC. Intra-RSC infusion of AMPA or kainic acid alone promoted death of RSC neurons and was reminiscent of the degeneration induced by the i.p. treatment with MK801. Collectively, these experiments provide evidence for an AMPA/kainate-dependent mechanism of excitotoxicity in the death of RSC neurons after i.p. treatment with MK801.

Neurodegeneration and prolonged immediate early gene expression throughout cortical areas of the rat brain following acute administration of dizocilpine.

N-methyl-d-aspartate receptor antagonist drugs (NMDA-A), such as dizocilpine (MK801), induce long-lasting behavioral disturbances reminiscent to psychotic disorders in humans. To identify cortical structures affected by NMDA-A, we used a single dose of MK801 (10 mg/kg) that caused low and high neurodegeneration in intact and orchiectomized male rats, respectively. Degenerating somas (neuronal death) and axonal/synaptic endings (terminal degeneration) were depicted by a silver technique, and functionally affected cortical neuronal subpopulations by Egr-1, c-Fos, and FosB/DeltaFosB-immunolabeling. In intact males, MK801 triggered a c-Fos induction that remained high for more than 24 h in selected layers of the retrosplenial, somatosensory and entorhinal cortices. MK801-induced neurodegeneration reached its peak at 72 h. Degenerating somas were restricted to layer IV of the granular subdivision of the retrosplenial cortex, and were accompanied by suppression of Egr-1 immunolabeling. Terminal degeneration extended to selected layers of the retrosplenial, somatosensory and parahippocampal cortices, which are target areas of retrosplenial cortex. Induction of FosB/DeltaFosB by MK801 also extended to the same cortical layers affected by terminal degeneration, likely reflecting the damage of synaptic connectivity. In orchiectomized males, the neurodegenerative and functional effects of MK801 were exacerbated. Degenerative somas in layer IV of the retrosplenial cortex significantly increased, with a parallel enhancement of terminal degeneration and FosB/DeltaFosB-expression in the mentioned cortical structures, but no additional areas were affected. These observations reveal that synaptic dysfunction/degeneration in the retrosplenial, somatosensory and parahippocampal cortices might underlie the long-lasting impairments induced by NMDA-A.

Increased rewarding properties of morphine in perinatally protein-malnourished rats.

In the current research, we assessed the influence of a protein malnutrition schedule from the 14th day of gestation up to 40 days of age (D-rats) on the rewarding properties of morphine in adult rats by means of the conditioned place preference paradigm. Well-nourished animals (C-rats) administered with different doses of morphine (0.75, 1.5, 3, 6, 12 or 24 mg/kg i.p.) exhibited a conditioning place preference with doses of 3 and 6 mg/kg, whereas in D-rats such a conditioning effect was observed with doses of 1.5 and 3 mg/kg. No adverse effects were observed in either C- or D-rats for the higher doses of morphine. In addition, when animals of both groups were pretreated twice a day for 3 days with increasing doses of morphine (5, 10 and 20 mg/kg s.c.), only D-rats elicited sensitization to the conditioning effect with the lowest dose of morphine (0.75 mg/kg i.p.). Furthermore, sensitized D-rats showed a selective and significant increase in FosB expression in the nucleus accumbens (core and shell), basolateral amygdala and medial prefrontal cortex, brain areas that are functionally related to the rewarding neural circuit. These results demonstrate that a deficient nutritional status during the perinatal period results in adult subjects having neural alterations, leading to an increased responsiveness to morphine and/or enhanced reinforcement effects, which correlates with an overexpression of FosB in selective brain areas related to the rewarding network.

Effect of retinal ablation on the expression of calbindin D28k and GABA in the chick optic tectum.

The effect of retinal ablation on qualitative and quantitative changes of calbindin D28k and GABA expression in the contralateral optic tectum was studied in young chicks. Fifteen days old chicks had unilateral retinal ablation and after 7 or 15 days, calbindin expression was analyzed by Western blot and immunocytochemistry. Neuronal degeneration was followed by the amino-cupric silver technique. After 15 days, retinal lesions produced a significant decrease in calbindin immunostaining in the neuropil of layers 5-6 and in the somata of neurons from the layers 8 and 10 of the contralateral tectum, being this effect less marked at 7 days post-lesion. Double staining revealed that 50-60% of cells in the layers 8 and 10 were calbindin and GABA positive, 30-45% were only calbindin positive and 5-10% were only GABAergic neurons. Retinal ablation also produced a decrease in the GABA expression at either 7 or 15 days after surgery. At 7 days, dense silver staining was observed in the layers 5-6 from the optic tectum contralateral to the retinal ablation, which mainly represented neuropil that would come from processes of retinal ganglion cells. Tectal neuronal bodies were not stained with silver, although some neurons were surrounded by coarse granular silver deposits. In conclusion, most of calbindin molecules are present in neurons of the tectal GABAergic inhibitory circuitry, whose functioning apparently depends on the integrity of the visual input. A possible role of calbindin in the control of intracellular Ca2+ in neurons of this circuit when the visual transmission arrives to the optic tectum remains to be studied.

Supracapsular bed nucleus of the stria terminalis contains central and medial extended amygdala elements: evidence from anterograde and retrograde tracing experiments in the rat.

Neurons that accompany the stria terminalis as it loops over the internal capsule have been termed collectively the supracapsular bed nucleus of the stria terminalis (BSTS). They form two cell columns, a lateral column and a considerably smaller medial column. The lateral column merges rostrally with the lateral bed nucleus of the stria terminalis and caudally with the central amygdaloid nucleus (central extended amygdala components). The medial column is continuous with the medial bed nucleus of the stria terminalis and the medial amygdaloid nucleus (medial extended amygdala districts). The connections of the BSTS were investigated in the rat by placing injections of Phaseolus vulgaris-leucoagglutinin (PHA-L) or retrograde tracers in different parts of the extended amygdala or in structures related to the extended amygdala. BSTS inputs and outputs were identified, respectively, by the presence of varicose fibers and retrogradely labeled neurons within the stria terminalis. The results suggest that the medial-to-lateral compartmentalization of BSTS neurons reflects their close alliance with the medial and central divisions of the extended amygdala. The medial BSTS contains primarily elements that correspond to the posterodorsal part of the medial amygdaloid nucleus and the medial column of the posterior division of the medial bed nucleus of the stria terminalis, and the lateral BSTS contains elements that correspond to the medial and lateral parts of the central amygdaloid nucleus and lateral bed nucleus of the stria terminalis. These results add strong support to the concept of the extended amygdala as a ring-like macrostructure around the internal capsule, and they are of theoretical interest for the understanding of the organization of the basal forebrain.

The concepts of the ventral striatopallidal system and extended amygdala.

The concepts of the ventral striatopallidal system and extended amygdala have significantly improved our understanding of basal forebrain organization. As a result of these and other advances during the last twenty years, many of the most prominent basal forebrain structures, including the nucleus accumbens, olfactory tubercle, and amygdaloid body, have all but lost their relevance as independent functional anatomical units. In order to appreciate the distinct differences that exist between the ventral striatopallidal system and the extended amygdala, and as a way of explaining the choice of the terms ventral striatopallidal system and extended amygdala, we will review the discovery and subsequent elaboration of these two systems. On the background of these discussions, we will then proceed to dispel some recently published misgivings regarding the usefulness of the extended amygdaloid concept.

The extended amygdala and salt appetite.

Both chemo- and mechanosensitive receptors are involved in detecting changes in the signals that reflect the status of body fluids and of blood pressure. These receptors are located in the systemic circulatory system and in the sensory circumventricular organs of the brain. Under conditions of body fluid deficit or of marked changes in fluid distribution, multiple inputs derived from these humoral and neural receptors converge on key areas of the brain where the information is integrated. The result of this central processing is the mobilization of homeostatic behaviors (thirst and salt appetite), hormone release, autonomic changes, and cardiovascular adjustments. This review discusses the current understanding of the nature and role of the central and systemic receptors involved in the facilitation and inhibition of thirst and salt appetite and on particular components of the central neural network that receive and process input derived from fluid- and cardiovascular-related sensory systems. Special attention is paid to the structures of the lamina terminalis, the area postrema, the lateral parabrachial nucleus, and their association with the central nucleus of the amygdala and the bed nucleus of the stria terminalis in controlling the behaviors that participate in maintaining body fluid and cardiovascular homeostasis.

The neuronal organization of the supracapsular part of the stria terminalis in the rat: the dorsal component of the extended amygdala.

In the present normal anatomical light and electron microscopic study in the rat, histochemical (Nissl, Timm, Golgi) or immunocytochemical (microtubule-associated protein type 2, glutamate decarboxylase, glutamate receptor subunit 1, synaptophysin) stains were used to analyse neurons embedded within the stria terminalis and their associated neuropil. These cells are closely related to the bed nucleus of the stria terminalis and the centromedial amygdala, and have been termed the "supracapsular part of the bed nucleus of the stria terminalis". The largest part of this neuronal complex is located in the ventrolateral part of the stria, where it appears as a round or oval "lateral pocket" in virtually any type of light microscopic preparation because of its collection of neuronal cell bodies and dense neuropil, in addition to a lacework of unmyelinated axons. A much smaller but still distinct "medial pocket" is located in the medial corner of the stria. The large lateral subdivision of the supracapsular stria terminalis is directly continuous with the lateral bed nucleus of the stria terminalis and extends to the central amygdaloid nucleus, containing a column of neurons that is only broken up into cell clusters at the most caudal levels of the stria as it drops vertically toward the amygdala. The considerably smaller medial subdivision appears, in turn, to be directly continuous with the medial part of the bed nucleus of the stria terminalis. The medial column tapers off more rapidly than the lateral part, so that as the middle levels are approached, only small interrupted clusters of cells are seen. Solitary neurons can also be found in practically every part of the stria terminalis except among the ventrally located axons of the commissural component. Most of the neurons are small to medium in size, as viewed in transverse sections of the stria, but larger neurons are also encountered. In sections parallel to the stria, many neurons are fusiform in appearance. The dendrites are often aligned in a longitudinal fashion; many of the dendrites related to the cells in the lateral pocket are moderately to densely spined, whereas those in the medial pocket are more sparsely spined. The neuropil in both the lateral and medial pockets is characterized by boutons, bundles of unmyelinated axons, and dendrites. Based on their vesicle content, the boutons are divided into three major types: (A) round or slightly oval, agranular vesicles of uniform size; (B) pleomorphic, agranular vesicles, many of which are flattened; and (C) pleomorphic agranular vesicles, some of which are considerably larger than the ones in type B boutons. Type A boutons establish contacts with both dendritic spines and shafts, whereas types B and C usually contact dendritic shafts and sometimes somata. These synaptic components are similar to those described earlier for the central and medial amygdaloid nuclei. Overall, our results support the contention advanced in 1923 by Johnston [J. comp. Neurol. 35, 337481] that the cells accompanying the stria terminalis are interconnecting columns of a macrostructure encompassing the bed nucleus of the stria terminalis and centromedial amygdala. More recently, it has been appreciated that columns of neurons below the globus pallidus also belong to this macrostructure [Alheid G. F. et al. (1995) In The Rat Nervous System, 2nd edn, pp. 495 578, Academic, San Diego; de Olmos J. S. et al. (1985) In The Rat Nervous System, pp. 223-334, Academic, Sydney], which has been named the "extended amygdala".

The accumbens: beyond the core-shell dichotomy.

This article highlights recent discoveries related to the accumbens and closely associated structures, with special reference to their importance in neuropsychiatry. The development of "striatal patches" in the accumbens is reviewed in a series of pictures. Neuronal ensembles are discussed as potentially important functional-anatomical units. Attention is also drawn to recent discoveries related to the neuronal circuits that the primate accumbens establishes with the mesencephalic dopamine system. On the basis of histological and neurochemical differences, the accumbens has been divided into core and shell compartments. In the context of this article, the shell, which is an especially diversified part of the accumbens, is the subject of special attention because of its close relation to the extended amygdala and distinctive response to antipsychotic and psychoactive drugs.

Substantia innominata: a notion which impedes clinical-anatomical correlations in neuropsychiatric disorders.

Comparative neuroanatomical investigations in primates and non-primates have helped disentangle the anatomy of the basal forebrain region known as the substantia innominata. The most striking aspect of this region is its subdivision into two major parts. This reflects the fundamental organizational scheme for this portion of the forebrain. According to this scheme, two major subcortical telencephalic structures, i.e. the striatopallidal complex and extended amygdala, form large diagonally oriented bands. The rostroventral extension of the pallidum accounts for a large part of the rostral subcommissural substantia innominata, while the sublenticular substantia innominata is primarily occupied by elements of the extended amygdala. Also dispersed across this region is the basal nucleus of Meynert, which is part of a more or less continuous collection of cholinergic and non-cholinergic corticopetal and thalamopetal cells, which stretches from the septum diagonal band rostrally to the caudal globus pallidus. The basal nucleus of Meynert is especially prominent in the primate, where it is sometimes inappropriately applied as a synonym for the substantia innominata, thereby tacitly ignoring the remaining components. In most mammals, the extended amygdala presents itself as a ring of neurons encircling the internal capsule and basal ganglia. The extended amygdala may be further subdivided, i.e. into the central extended amygdala (related to the central amygdaloid nucleus) and the medial extended amygdala (related to the medial amygdaloid nucleus), which generally form separate corridors both in the sublenticular region and along the supracapsular course of the stria terminalis. The extended amygdala is directly continuous with the caudomedial shell of the accumbens, and to some extent appears to merge with it. Together the accumbens shell and extended amygdala form an extensive forebrain continuum, which establishes specific neuronal circuits with the medial prefrontal-orbitofrontal cortex and medial temporal lobe. This continuum is particularly characterized by a prominent system of long intrinsic association fibers, and a variety of highly differentiated downstream projections to the hypothalamus and brainstem. The various components of the extended amygdala, together with the shell of the accumbens, are ideally structured to generate endocrine, autonomic and somatomotor aspects of emotional and motivational states. Behavioral observations support this proposition and demonstrate the relevance of these structures to a variety of functions, ranging from the various elements of the reproductive cycle to drug-seeking behavior. The neurochemical and connectional features common to the accumbens shell and the extended amygdala are especially relevant to understanding the etiology and treatment of neuropsychiatric disorders. This is discussed in general terms, and also in specific relation to the neurodevelopmental theory of schizophrenia and to the neurosurgical treatment of neuropsychiatric disorders.

Use of an amino-cupric-silver technique for the detection of early and semiacute neuronal degeneration caused by neurotoxicants, hypoxia, and physical trauma.

A new amino-cupric silver protocol is described for detection of neuronal degeneration. We describe its selectivity in visualizing both early and semiacute degeneration after intracerebral or systemic administration of a variety of neurotoxicants in rats, and after transient ischemic episodes in gerbils. As early as 5 min after physical trauma, or 15 min following either intrastriatal injections of glutamate analogs or exposure to ischemic episodes, neuronal silver staining was evident at primary sites of trauma (i.g. injection sites) and at hodologically related secondary sites. With intoxication by peripheral injections of trimethyltin (IP) or intracerebral injections of Doxorubicin, reproducible patterns of degeneration are demonstrable after 24 h or after 9-13 days, respectively. The amino-cupric silver method permits simultaneous detection of all neuronal compartments against a clear background. Degeneration in the neuronal cell bodies, dendrites, axons and terminals, as well as the recruitment of new structures in a progressive pathologic process, could be accurately followed. The inclusion of new reagents increased the sensitivity vis-à-vis previous versions of the cupric-silver method. The advantages and disadvantages of the current method in comparison with other means of neurotoxic assessment are discussed in detail, with special emphasis on its unique ability to discriminate irreversible degenerative phenomena and degeneration of axonal components in cases where the cell body remains apparently intact. The amino-cupric silver method is an especially useful tool for surveying neuronal damage in basic neuroscience investigations and in neuropathologic and neurotoxic assessment.

Silver staining as a tool for neurotoxic assessment.

There is no denying that the silver methods lost their dominant role as tract-tracing methods in the past 10 to 15 years. But it seems equally clear that the silver technique is headed for a dramatic revival in many fields of neuroscience, where the scope and localization of neuronal degeneration are a central issue. Together with the immunostaining of proteins formed or altered in traumatized neurons, the modern silver techniques provide neurotoxicologists and neuropathologists with unparalleled opportunities to detect and study injured and dying neurons. Characterized by great sensitivity and distinct rendition of the morphology of degenerating neurons and their processes, the reduced silver methods constitute the ideal tool for screening irreversible neuronal damage caused by neurotoxic substances including drugs of abuse. Those interested in the rapidly expanding fields of "excitotoxicity" and neurodegenerative disorders (Taylor 1991) are also likely to find increasing use for the silver methods. The pattern of degeneration in so-called "system degenerations" may be predetermined by the neuronal connections (Saper et al. 1987), and as the disease progresses from the destruction of the originally affected neuron population, closely related systems and pathways may be recruited into the pathophysiologic cascade. Any type of trauma to the CNS has the potential to produce this type of "domino effect" of degeneration, through which additional systems are progressively recruited into a degenerative chain reaction of transneuronal degeneration. In other words, even longstanding disorders may exhibit signs of more recent degeneration, and the proper use of silver methods at autopsy may give some important clues regarding the etiology of disease; it may also provide new insights about the anatomy of the human brain. Little can be said at present about the chemical basis of argyrophilia in degenerating and "reactive" neurons, but there is every reason to pay more attention to this subject. One can expect that a continuing and concerted effort will result in a rational understanding of the molecular biological and physicochemical events that fortuitously provide the basis for the selective impregnation of degenerating neuronal elements. This knowledge can be the basis for the development of even more reliable and simple, yet sensitive, silver methods suited for neurotoxic risk assessment on a large scale.

Projections from the amygdala to basoventral and mediodorsal prefrontal regions in the rhesus monkey.

The sources of ipsilateral projections from the amygdala to basoventral and mediodorsal prefrontal cortices were studied with retrograde tracers (horseradish peroxidase or fluorescent dyes) in 13 rhesus monkeys. The basoventral regions injected with tracers included the orbital periallocortex and proisocortex, orbital areas 13, 11, and 12, lateral area 12, and ventral area 46. The mediodorsal regions included portions of medial areas 25, 32, 14, and dorsal area 8. The above sites represent areas within two architectonic series of cortices referred to as basoventral or mediodorsal on the basis of their anatomic location. Each series consists of areas that show a gradual increase in the number of layers and their delineation in a direction from the caudal orbital and medial limbic cortices, which have an incipient laminar organization, towards the eulaminated periarcuate cortices (Barbas and Pandya, J. Comp. Neurol. 286: 353-375, '89). Labeled neurons projecting to the prefrontal cortex were found in the basolateral, basomedial (also known as accessory basal), lateral, and ventral cortical nuclei, and in the anterior amygdaloid and amygdalopiriform areas. The distribution of labeled neurons differed both quantitatively and qualitatively depending on whether the injection sites were in basoventral or mediodorsal prefrontal cortices. Cases with caudal orbital injections had the most labeled neurons in the amygdala, followed by cases with injections in cortices situated medioventrally. The latter received a high proportion of their amygdaloid projections from the basomedial nucleus. The lateral amygdaloid nucleus sent a robust projection to the least architectonically differentiated orbital periallocortex, and a weaker projection to the adjoining orbital proisocortical regions, but did not appear to project to either medial proisocortical sites or to the more differentiated ventrolateral or dorsolateral prefrontal cortices. In addition, there were topographical differences in the origin of projections from one amygdaloid nucleus directed to various prefrontal cortices. These differences were correlated either with the destination of the axons of afferent amygdaloid neurons to basoventral or to mediodorsal prefrontal cortices and/or with their projection to areas with varying degrees of laminar organization within the basoventral or mediodorsal sector. The clearest topography was observed for projections originating in the basolateral nucleus.(ABSTRACT TRUNCATED AT 400 WORDS)

Quantitative determination of collateral anterior olfactory nucleus projections using a fluorescent tracer with an algebraic solution to the problem of double retrograde labeling.

The bilateral projections of the rat anterior olfactory nucleus (AON) were evaluated using retrograde fluorescent tracers. Competitive effects of these tracers led to severe underestimation of bilaterally projecting neurons, when double-labeled cells were counted. The underestimate was corrected using a numerical approach, which is of general utility for problems in double labeling and requires only a single tracer. With this method we estimated that approximately 63% of AON neurons project bilaterally to the olfactory bulbs, except for the external part which projects exclusively to the contralateral olfactory bulb. No other AON neurons project only to the contralateral bulb.

Tracing of two-neuron pathways in the olfactory system by the aid of transneuronal degeneration: projections to the amygdaloid body and hippocampal formation.

Following an olfactory bulb lesion in guinea pig (2 to 3 days), neuronal degeneration occurs in several olfactory-bulb-related areas, primarily in the piriform cortex. The degenerating neurons, which are argyrophilic, are also found in the posterolateral cortical amygdaloid nucleus and the ventrolateral entorhinal cortex. It is suggested that the neurons degenerate because of a transneuronal effect due to a sudden loss of afferent input from the olfactory bulb, although a retrograde effect acting in concert with transneuronal factors cannot be excluded. Terminal degeneration can be identified in several areas outside the olfactory bulb projection area, and is interpreted as degeneration in the axons of the degenerating cortical neurons. Such terminal degeneration, which is best seen 3 to 4 days postoperatively, has been identified in part of the basolateral amygdaloid complex, in the basomedial amygdaloid nucleus, and in the temporal parts of the fascia dentata of the hippocampal formation. Terminal degeneration has also been observed in the deep layers of the anterior olfactory nucleus, the olfactory tubercle, the nucleus of the lateral olfactory tract, and the anterior amygdaloid area. All these projections, apparently, represent the second link in two-neuron pathways, where mitral or tufted cells in the olfactory bulb make up the first neuron. This interpretation was confirmed in control experiments in which areas of argyrophilic neurons coincided with the location of retrogradely labeled neurons following injection of fluorescent substances into several of the above-mentioned areas of terminal degeneration.

A modified cupric-silver technique for the impregnation of degenerating neurons and their processes.

A rapid version of the de Olmos-Ingram cupric-silver technique with higher sensitivity and affinity for degenerative changes is introduced for the staining of mechanically and chemically induced degeneration. Emphasis is laid on induction of degeneration with experimental approaches sparing 'fibers of passage'.

Double and triple labeling of neurons with fluorescent substances; the study of collateral pathways in the ascending raphe system.

A retrograde labeling procedure utilizing fluorescent substances (Granular Blue, Nuclear Yellow and propidium iodide) was used to establish the presence of branching axons in the ascending raphe system of young rats. After injections in septum, medial thalamus and olfactory cortex, the number of double-labeled cells in various combinations was found to be relatively large in the dorsal raphe nucleus, whereas triple-labeled cells occurred more rarely. Each class of neurons, i.e. single-, double- and triple-labeled, were shown to have a predominant distribution within specific parts of the nucleus.

The afferent connections of the main and the accessory olfactory bulb formations in the rat: an experimental HRP-study.

The afferent connections of the main and accessory olfactory bulbs in the rat were examined by injecting horseradish peroxidase (HRP) into one or the other of these structures either by microelectrophoresis or by hydraulic pressure. Alternate sections were stained with newly developed HRP-procedures using either benzidine dihydrochloride (de Olmos and Heimer, '77) or tetramethyl-benzidine. Eighteen to twenty-four hours after unilateral HRP injections confined to the main olfactory bulb, a large number of HRP-labeled perikaria appeared in the following telencephalic structures on the ipsilateral side: All portions of the anterior olfactory nucleus (AON) except its external part, the lateral transitional field (LT) between AON and the paleocortex, the whole extent of the primary olfactory cortex (POC); the medial forebrain bundle area deep to the olfactory tubercle, the nucleus of the horizontal limb of the diagonal band (NHDB) and the nucleus of the lateral olfactory tract (NLOT). A moderate to small number of labeled cells, furthermore, were seen in the dorsal (DT) and medial (MT) transition fields, the ventral praecommissural hippocampus (tt2), the ventral superficial part of the nucleus of the vertical limb of the diagonal band (NVDB), the sublenticular part of the substantia innominata (SI), the anterior amygdaloid area, the posterolateral cortical amygdaloid nucleus (C2) and the transition region (28 L') between the olfactory cortex and the lateral entorhinal area proper. On the contralateral side a large number of labeled cells were found in all parts of the AON, with especially heavy labeling in its external part. A moderate number of labeled cells could also be detected in the lateral transition field (LT) and the NLOT. In the diencephalon and the brain stem a moderate number of HRP-labeled perikaria were observed in the dorsal, perifornical, and lateral hypothalamus, as well as in locus coeruleus and the dorsal and medial raphae nuclei. Following large HRP injections in the main olfactory bulb a moderate to small number of labeled cells were seen also in the posterior and premammillary hypothalamus and in field CA1 of the retrocommissural hippocampus on the ipsilateral side, as well as in POC on the contralateral side. It is possible, however, that the uptake of label took place in an undetected pool of HRP in the very rostal part of AON rather than in the olfactory bulb. HRP injections in the accessory olfactory bulb resulted in labeled neurons in the posterior ventro-lateral part of the bed nucleus of the stria terminalis, the nucleus of the accessory olfactory tract, the rostrodorsal portions of the medial amygdaloid nucleus, and the whole extent of the posteromedial cortical amygdaloid nucleus (C3) on the ipsilateral side. A few lightly labeled cells were seen also in the contralateral C3.