| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Nencki Institute, Warsaw, Poland
ABSTRACT I. ORGANIZATION OF THE AMYGDALA A. An Outline of the Amygdala Anatomy B. The Major External Connections of the Amygdala C. Theories of the Functional Organization of the Amygdala II. GENE ACTIVITY AS A MAPPING TOOL: THEORETICAL AND TECHNICAL CONSIDERATIONS III. AMYGDALA ACTIVATION FOLLOWING BEHAVIORAL TRAINING A. Introduction B. Pattern of c-Fos Expression in the Amygdalar Nuclei in Aversively Motivated Standard Behavioral Tests C. Pattern of Expression of Other Activity Markers in the Amygdalar Nuclei in Aversively Motivated Standard Behavioral Tests D. Pattern of c-Fos Expression in the Amygdalar Nuclei in Aversively Motivated, Ethologically Relevant Behavioral Tests E. Pattern of Expression of Other c-fos Gene Activity Markers in the Amygdalar Nuclei in Aversively Motivated, Ethologically Relevant Behavioral Tests F. Pattern of c-Fos Expression in the Amygdalar Nuclei in Appetitively Motivated Standard Behavioral Tests G. Pattern of Expression of Other Activity Markers in the Amygdalar Nuclei in Appetitively Motivated Standard Behavioral Tests H. Pattern of c-fos Expression in the Amygdalar Nuclei in Appetitively Motivated, Ethologically Relevant Behavioral Tests 1. Sexual interaction/stimulation 2. Maternal/paternal interaction 3. Nonagonistic social interaction I. Pattern of Expression of Other Gene Activity Markers in the Amygdalar Nuclei in Appetitively Motivated, Ethologically Relevant Behavioral Tests J. The IntelliCage System: In-Between of Standard Tests and Ethologically Based Behavioral Paradigms K. Comparison of c-Fos Versus Zif268 Expression Patterns Following Behavioral Training IV. GENE ACTIVITY MARKERS AND DRUGS OF ABUSE A. Model Systems 1. Forced drug administration: acute and chronic 2. Instrumental intravenous drug self-administration 3. Conditioned place preference and avoidance 4. Voluntary alcohol drinking B. Cocaine-Evoked c-Fos Expression C. Cocaine and Other Gene Activity Markers D. Amphetamines and c-Fos Expression E. Amphetamines and Other Gene Activity Markers F. Methylphenidate and c-Fos Expression G. Nicotine and c-Fos Expression H. Nicotine and Other Than c-fos Gene Activity Markers I. Opioids, Morphine Derivatives (Morphine, Heroin), and c-Fos Expression J. Cannabinoids and c-Fos Expression K. Hallucinogens and c-Fos Expression L. Ethyl Alcohol and c-Fos Expression V. DIFFERENT BEHAVIORAL TASKS INVOLVE SPECIFIC SUBDIVISIONS OF THE AMYGDALA A. c-Fos in Neuronal Plasticity B. Basolateral Amygdala C. Central Amygdala D. Medial Amygdala E. Cortical Amygdala VI. CONCLUSIONS: THE FUNCTIONAL HETEROGENEITY OF THE AMYGDALA ACKNOWLEDGMENTS REFERENCES
 |
ABSTRACT |
|---|
 Top NextReferences |
|---|
|
I. ORGANIZATION OF THE AMYGDALA |
|---|
|
TopPreviousNextReferences |
|---|
The amygdaloid complex is a small, spherical gray mass located, in most mammals, in the medial wall of the temporal lobes, close to the terminal part of the inferior horn of the lateral ventricle and adjacent to the hippocampus. This phylogenetically old, almond-shaped structure, like many other parts of the limbic system, has been considered for a long time as controlling various emotional and motivated behaviors. In recent years, the amygdala has emerged as the key forebrain structure mediating inborn and acquired emotional responses, as well as processing, interpreting, and integrating various aspects of biologically and/or emotionally important information. Since the amygdala was described for the first time in the human temporal cortex by Burdach in the early 19th century, many subdivisions of the amygdalar complex have been identified, and the classification of its subdivisions, the extent of its outer border, as well as the nomenclature and number of nuclei have remained controversial (see Ref. 381 for comments).
Different investigators have grouped the amygdaloid nuclei in various ways, using different criteria. One of the first anatomical descriptions was proposed by Johnston (201), who divided the amygdala into a two-part system based on a detailed analysis of comparative vertebrate material. He distinguished 1) a phylogenetically older corticomedial group of nuclei, composed of the central, medial, and cortical nuclei and the nucleus of the lateral olfactory tract and 2) a phylogenetically newer basolateral group, composed of the lateral, basal, and basomedial nuclei. It is worth noting that the nucleus of the lateral olfactory tract (see Ref. 483) and the cortical nuclei (141, 142) were sometimes excluded from the corticomedial part of the amygdaloid complex. Therefore, some authors used the term centromedial and/or dorsomedial (121) to designate Johnston's corticomedial division. On the basis of neuronal morphology (299), the amygdaloid nuclei were grouped into two major subdivisions: 1) the cortex-like nuclei that contain mostly pyramidal or modified pyramidal projection neurons and 2) the noncortex-like nuclei that do not have pyramidal-like neurons. According to McDonald (299), the cortex-like nuclei include the whole Jonston's basolateral group, as well as the cortical nuclei, whereas the noncortex-like nuclei include the central and medial nuclei.
Obviously, the nomenclatures that were set, especially in earlier studies (see Refs. 47, 77, 78, 123, 256), were based mostly on appearance of the tissue, without regard to organization of its connections with other brain structures. On the other hand, more recent anatomical descriptions combine examination of the intrinsic and extrinsic connections of separate amygdaloid nuclei with their cytoarchitectonic and histochemical characteristics. For instance, Price et al. (381) identified three distinct groups of nuclei, differently connected with other brain structures. The first, the basolateral group (called also the group of deep nuclei), is composed of the lateral, basal, and basomedial nuclei (the basomedial nucleus is also termed the accessory basal nucleus). This group is characterized by substantial interconnections with the neocortex. The second, the corticomedial or superficial group, is made up of the periamygdaloid cortex, the anterior and posterior cortical nuclei, the medial nucleus, and the nucleus of the lateral olfactory tract, which are directly connected with the olfactory and accessory olfactory system. The third group of nuclei is composed of the central nucleus and the anterior amygdaloid area, which are strongly interconnected with the autonomic control centers in the lateral hypothalamus and the brain stem structures.
Recently, McDonald (299) as well as Swanson and Petrovich (463) have further developed the parcellation introduced by Price et al. (381). Taking into consideration currently available cytoarchitectonic, chemoarchitectonic, and fiber connections data, they divided the amygdala into three parts: 1) the deep or basolateral group, which is constituted by a ventromedial extension of the deepest layer of the cortex (the lateral, basal, and basomedial nuclei); 2) the centromedial group, which is specialized ventromedial expansion of the striatum (the central and medial nuclei, as well as the amygdaloid part of the bed nucleus of stria terminalis); and 3) the superficial or cortex-like group being a part of the caudal olfactory cortex (the anterior and posterior cortical nuclei, the nucleus of the olfactory tract, and the periamygdaloid cortex). Several further subdivisions have been also distinguished. For instance, in the lateral nucleus, two components (the anterior and posterior) were described by Krettek and Price (245). On the other hand, Alheid et al. (5) divided the lateral nucleus into three regions (dorsal, ventrolateral, and ventromedial). According to Pitkanen et al. (373), the basal nucleus is comprised of three regions: the magnocellular, intermediate, and parvocellular subdivisions, and the basomedial nucleus is divided into the magnocellular and parvocellular subdivisions. Moreover, three regions were distinguished in the central nucleus: the capsular, lateral, and medial subdivisions. The medial nucleus was divided into the rostral, central, and caudal subdivisions, whereas in the periamygdaloid cortex the medial and sulcal regions were distinguished (see Ref. 373). Furthermore, Price et al. (381) and some other authors (see Refs. 10, 373, 374, 427) included to the amygdaloid complex the anterior amygdaloid and amygdalo-hippocampal areas, as well as groups of the intercalated nuclei. The present understanding of the internal organization of the amygdala is shown on the Figure 1.
|
The amygdaloid complex is characterized by widespread and well-organized connections with many forebrain, midbrain, and hindbrain structures. The stria terminalis (ST) and the ventral amygdalofugal pathway (VAFP) are two major bundles of fibers connecting the amygdala with other areas of the brain. The ST, which arises mainly from the corticomedial group of nuclei, conveys fibers to several subcortical structures: the ventromedial hypothalamus as well as the septal and medial preoptic areas (93, 173, 270). The fibers of ST project also to the autonomic centers in the periaqueductal gray matter, parabrachial nucleus, and sympathetic preganglion neurons in the spinal cord. Moreover, the basal nucleus of the amygdala, which projects to the bed nucleus of the stria terminalis, influences a variety of hypothalamic, ventral, and dorsal striatum structures, as well as the brain stem areas. The latter also receive direct projections from the corticomedial amygdala. On the other hand, the VAFP is a rather diffuse fiber tract, which connects mostly the basolateral part of the amygdala with the thalamus, hypothalamus, septum, nucleus accumbens, and other structures of the ventral striatum, parahippocampal gyrus, and certain parts of the cingulum, piriform, and orbitofrontal cortices. The VAFP connects also the lateral and central amygdala with the lateral hypothalamus and with the dopaminergic neurons in the brain stem reticular formation.
From the functional point of view, all connections of the amygdaloid complex can be divided into three major systems (380). The first one provides sensory information to the amygdala, thus supporting amygdaloid modulation of sensory processing. It is composed mostly of the reciprocal projections and connects the amygdala with the olfactory cortex and ascending taste/visceral pathways. Moreover, it transfers sensory information from the posterior thalamus and sensory association cortex. The second system forms outputs of the amygdaloid complex to the hypothalamus and brain stem structures, which modulate visceral functions in relation to the emotional significance of internal and external stimuli. The last forebrain circuit is composed of the amygdala connections with the ventromedial frontal, rostral insular, and rostral temporal cortical areas, as well as with the medial thalamus and ventromedial basal ganglia. Through these connections the amygdala is able to be directly involved in regulation of emotional behavior and to influence several somatic motor responses.
It should be noted that each individual subdivision of the amygdala is characterized by a specific pattern of internal and external connections. An overview of those is presented in Figures 2–4. Their importance for understanding of the functional heterogeneity of the amygdala is discussed below.
|
|
|
More than 100 years ago, Brown and Schafer (49) demonstrated in monkeys that after extensive bilateral damage to the temporal lobes, the animals became unnaturally fearless and tame and that this amazing transformation in the emotional behavior was not associated with apparent sensory changes. Similar emotional transformation was later described by Klüver and Bucy (225) after bilateral destruction of the amygdala and inferior temporal cortex. They observed that rhesus monkeys, earlier fierce and rather apprehensive, after the lesion approached fear-inducing stimuli with no display of anger or fear. The animals became emotionally dulled, and their facial expression and vocalization were less expressive. They were unable to recognize or use previously familiar objects, although their vision was not impaired, and they tended to explore every object. The animals also showed inappropriate sexual behavior, changes in food preferences, and disruption of social behavior evidenced by a loss in rank or even withdrawal from social life.
Later, the amygdala function was studied with the use of a variety of methods, including ablations and/or partial lesions, electrical brain stimulation, neurochemical intra-amygdala injections, and single-unit recordings. Notably, considering anatomical heterogeneity in the amygdala, the most significant conclusions resulting from these studies clearly indicated an advantage in performing subtotal amygdalar lesions over the total amygdalectomy (see Ref. 426).
One of the earliest concepts explaining the functional organization of the amygdala was proposed by Wutz and Olds (518), starting from the point that one can discriminate phylogenetically and morphologically two major subdivisions of the amygdalar complex: the dorsomedial and basolateral groups of nuclei (see above). Wutz and Olds, relying on the results of the self-stimulation studies, suggested the former as a rewarding and the latter as a punishing system. Thus they pointed at the importance of the valence of motivation in the functional descriptions of the amygdala. Although, for the next decades most of the studies were focused on the involvement of the amygdala in negative emotions, recent evidence supports a role of this structure in processing positive emotions, in addition to the negative ones (28, 216).
Nearly all studies on the functional organization within the amygdala have been focused on the role of the basolateral and the central nuclei, while the functions of the cortical and the medial nuclei remain more elusive. Countless studies of conditioned fear showed very clearly that the amygdala is involved in Pavlovian conditioning of aversive emotional responses (see below). Lesions of either the basolateral or central nuclei impaired freezing that indicated the conditioned state of fear (146, 258). According to the well-established model of fear conditioning, information flows to the lateral nucleus and is conveyed directly or via the basal nucleus to the central nucleus, which is primarily seen as the universal output of the amygdala, able to coordinate the autonomic, endocrine, and behavioral responses via brain stem arousal and response systems. The selective lateral nucleus damage demonstrated that it is a "sensory interface" of the amygdala in fear conditioning (259, 401), at least with respect to the auditory and contextual cues (146; but see Ref. 468). Moreover, the basolateral complex of the amygdala is seen as a place of formation of the conditioned stimulus (CS)-unconditioned stimulus (US) associations in fear conditioning (11, 259, 261, 288, 291, 292, 336, 401, 431, 494).
A widely held model, in which the basolateral complex (primarily the lateral nucleus) acts as the associative site for stimulus-outcome representations and the central nucleus provides the output pathway through which these associations gain access to appropriate responses, such as the conditioned freezing response, is known as a serial model of the basolateral/central amygdala function (401) (Fig. 5). This single, serially organized hierarchical lateral-to-central flow of information has been challenged by Killcross, Everitt, and others (see Refs. 22, 218, 219), who showed that the instrumental avoidance responses were impaired in the basolateral amygdala-lesioned animals, whereas Pavlovian conditioned suppression required intact central nucleus. Moreover, an analogous double dissociation using an appetitively motivated task was reported (179, 359). Thus Killcross et al. proposed the parallel model of information processing in the amygdala (22, 61, 218, 219, 335) (see Fig. 5).
|
|
II. GENE ACTIVITY AS A MAPPING TOOL: THEORETICAL AND TECHNICAL CONSIDERATIONS |
|---|
|
TopPreviousNextReferences |
|---|
The fact that among IEGs those encoding transcription factors are the most often investigated in the brain, should be noted here, as it bears two very important consequences. Functionally, it suggests that we are just touching upon a tip of an iceberg of the genomic responses driven by those transcription factors, responses very poorly defined so far. It cannot, however, pass unnoticed that orchestrating such potentially very far-reaching responses is the only functional meaning of the phenomenon of the IEGs' expression. As a matter of fact, coordination of neuronal plasticity has not only been suggested as IEGs' function, but in a few cases has already been proven (consider, e.g., a pathway of extracellular matrix remodeling driven by AP-1; Refs. 197, 209, 338). Another important consequence of studies on IEGs/transcription factors is nuclear localization of the IEGs' protein products that greatly facilitates their immunostaining-based detection, because of excellent spatial resolution and rather simple way of numerical evaluation. The fact that extracellularly regulated gene expression was found to be often very transient has proven to be also very convenient, as it allows for a good temporal resolution of phenomena analyzed.
To help to appreciate the results obtained with IEG-based mapping, it is worthwhile to describe briefly a molecular scenario of their cellular expression pattern. Usually, their mRNA and protein levels are very low at basal conditions, i.e., in a variety of nonstimulated cells, including neurons. Increased neuronal expression results from stimulation of membrane receptors and subsequent rise in second messengers, including Ca2+, followed by kinase activation. The temporal pattern of this activation is quite uniform. An increase in mRNA levels is observed within a few minutes after the signal arrives at the cell membrane, and next the protein is accumulated, which occurs roughly between 30 and 90 min. Both mRNA and protein increases are transient. The transcription is soon shut off, and because of the aforementioned short half-lives of mRNA and protein, those gene products are gone within a couple of hours from the activated cells. Hence, if their expression is still observed in a group of neurons at later times, it suggests that it either results from novel stimuli that arrived later at the cell membrane, or we deal with a subset of late-activated cells. It is also of note that, especially in the case of c-fos, the gene expression in the brain is responsive to the first encounter to a novel stimulus. Repeated animal exposure to the same conditions gradually, within just a handful of sessions, diminishes c-Fos levels to virtual nothing. However, any novel element in otherwise very familiar set-up alerts the animal and results in c-fos activation (13, 343).
Two major experimental approaches to visualize IEGs' products expression are 1) in situ hybridization, detecting mRNA, and 2) immunocytochemistry for protein visualization. The former should provide unequivocal results, whereas the latter may be less specific, as different IEGs' proteins share structural similarities and there are antibodies available that do not differentiate among various Fos proteins. In situ hybridization may either be based on radioactive probe detection, and in this case, typically, provide resolution not allowing to visualize precisely single cells. This drawback has been circumvented recently with efficient nonradioactive probe labeling and detection. On the other hand, immunocytochemistry easily allows for a single-cell resolution at the levels of individual neuronal nuclei. Notably, whereas it has been shown that glia may express IEGs (see, e.g., Ref. 74), this apparently has not been a case in the brain in vivo in a context of neuronal plasticity, learning and memory, drugs of abuse, etc. It is also of note that excitatory neurons comprise the major population of IEG expressing neurons; however, inhibitory interneurons can also do so (65, 115).
In this analysis we have focused on the gene activity markers in the amygdala. These data are presented in a context of wealth of available information gathered on internal heterogeneity of the amygdala with other methods. One has to be, however, aware of the methodical limitations of the approaches employed. In all reported experiments, expression of the gene activity markers within amygdalar nuclei or their subdivisions was compared in the experimental groups with that of the control groups, and the statistically significant differences were determined. For such results to be valid, all the factors affecting the amygdala activation in the control groups should also be carefully taken into account. However, in many different experiments, nonequivalent procedures of handling or habituation to the experimental conditions in the control animals have been used (see also Refs. 98, 227). Furthermore, our recent study shows that in addition to the effects of the designed experimental treatment on the amygdala, one shall also consider information transfer among the animals that also activates this brain structure (228). These could seriously affect the conclusions of such studies leading to apparent discrepancies. Moreover, very often different statistical tests are used; therefore, the level of reliability is different.
The amygdaloid complex consists of several cytoarchitectonically well defined and internally distinguishable nuclei. However, furthermore, in concert with the anatomical data, there are also functional differences between various nuclei (see, e.g., Refs. 80, 111, 113, 162, 217, 229, 235, 282, 332, 385) or even subnuclei (see, e.g., Refs. 159, 385, 391). Therefore, it seems valuable to apply more precise dissection of the nuclei of the amygdala in analyzing the functions of the amygdala in a context of behavior. It is of note that this could be easily achievable with the use of activity markers immunolabeling approach that provides a mapping tool enabling a single-cell resolution. Unfortunately, this important approach has not routinely been employed.
Another obstacle in drawing the conclusions about the functional role of different amygdalar nuclei is also unavailability of the global picture of the amygdala activation for many experimental situations, namely, activation of the medial and cortical nuclei of the amygdala has predominantly been studied in sexual and social behaviors, whereas that of the basolateral and central amygdala has been studied in aversively motivated learning paradigms.
Finally, one additional, very poorly addressed important factor influencing the obtained results might be the amygdala lateralization. Notably, Holahan and White (181) as well as Scicli et al. (435) have recently shown the substantial differences in the c-Fos response between the left and right amygdala during contextual fear learning and memory retrieval.
|
III. AMYGDALA ACTIVATION FOLLOWING BEHAVIORAL TRAINING |
|---|
|
TopPreviousNextReferences |
|---|
In this section we present the results of the studies that have documented the patterns of the amygdalar nuclei activation measured with various gene activity markers under different conditions of behavioral training. To systematize the existing data we have dissected them accordingly to either appetitive or aversive motivation, which the behavioral tests employ, as well as we distinguish a category of the behavioral tests with an ethological relevance.
The evolutionary foundation of the very complex system of emotional responses, that one can observe in mammals, has apparently a two-factor motivational organization, i.e., affects are organized by brain systems that adaptively respond to two basic kinds of stimulation, appetitive and aversive (see Ref. 255). Among the first to propose such a two-factorial system were Schneirla (433) and Konorski (240), who divided behavioral responses according to their adaptational and motivational significance into either approach versus withdrawal behaviors or preservative versus protective reflexes, respectively. Approach and/or preservative reactions included ingestion, copulation, or nurture of progeny, whereas withdrawal and/or protective ones, either escape from or rejection of noxious agents. Herein, we would like to compare the patterns of the amygdalar nuclei activation caused by appetitively versus aversively motivated behaviors.
An involvement of the amygdalar nuclei in emotionally charged behavior and memory is studied mainly in the well-established experimental models of conditioning, such as Pavlovian fear conditioning, active and inhibitory avoidance, or bar-pressing for food reinforcement. All of them employ artificial signaling stimuli and procedures designed for laboratory conditions. However, there is also recent interest in experimental paradigms that exploit more ethologically relevant approaches, using such stimuli as predators, sexual partners, or pups' presence. These paradigms employ evolutionarily and adaptively formed unconditioned emotional states, e.g., fear of predators (see Ref. 408). In this review, we compare the pattern of the amygdalar nuclei activation evoked by these two types of experimental conditions.
Thus far, the most frequently used neuronal activity marker has been c-fos mRNA and its protein (c-Fos). Therefore, we first analyze the data on c-fos activation and then compare them with the results on expression of other, less widely investigated gene activity markers, such as FosB, c-Jun, JunB, JunD, Zif268, NGFI-B, Arc, ICER, phospho-CREB, and brain-derived neurotrophic factor (BDNF).
B. Pattern of c-Fos Expression in the Amygdalar Nuclei in Aversively Motivated Standard Behavioral Tests
The majority of the studies on the expression of gene activity markers and behavioral training involved fear conditioning. Most often, two main training/testing procedures were employed: 1) contextual fear conditioning that is a training paradigm in which the animal is placed for a short time in a chamber, where it receives an inescapable foot shock or foot shocks (unconditioned stimulus, US). Then, it is brought back to the home cage and after some delay it is reexposed to the same apparatus and tested without the US. The memory of the situation is examined by measuring the freezing reaction (freezing or tonic immobility is a temporary state of profound motor inhibition induced by situations that supposedly generate intense fear, with the objective to protect the animal from attacks by predators) during the immediate post-shock period and the retention test. 2) The procedure of cued fear conditioning is similar, except that the original exposure to the US is accompanied by a clear sensory stimulus, e.g., a tone, light, or scent (conditioned stimulus, CS), and the testing is carried out in an experimental cage different from the training chamber, but in the presence of the CS. It is noteworthy that the animals exposed to a foot shock immediately upon entering a novel environment do not efficiently acquire a conditioned freezing reaction. The reduction of contextual conditioning can also be achieved by a preexposure to the training environment (termed as latent inhibition). These two phenomena are often used as control conditions. However, one should bear in mind that these conditions do not exclude learning as such. For instance, the latent inhibition relies on learning that the experimental set-up is initially safe, and this produces a difficulty in learning that it is unsafe later. Notably, recently, Levenson et al. (271) have shown that latent inhibition results in increased acetylation of histone H4 in the hippocampus that in turn may imply rather widespread changes in the gene expression.
Nearly all studies on the expression of gene activity markers within the amygdala during acquisition of conditioned fear have been focused on the basolateral amygdala (and its components, lateral and/or basal nuclei), as well as the central nucleus. Most of the results point to an involvement of the basolateral amygdala in both contextual and cued fear acquisition (for reference, see Table 1). In most cases, no differences in the pattern of c-Fos expression in the amygdalar nuclei between contextual and cued fear conditioning have been reported. However, Radulovic et al. (383) noted the increased c-Fos expression in the basolateral amygdala in cued but not in contextual fear conditioning. Notably, in most of the studies, the lack of activation of c-Fos in the central amygdala was observed (see Table 1).
|
To sum up, it seems that the basolateral part and, less certainly, the medial nucleus of the amygdala are activated in both contextual and cued fear conditioning, whereas the level of c-Fos expression is not augmented in the central and cortical nuclei following such training.
During retrieval trials in the contextual fear conditioning, an animal is placed in the same experimental context as during the training but in an absence of the shock, whereas in cued fear conditioning, it is placed in a different experimental context and the CS cue is presented. In both conditions as a measure of the strength of the memory trace, the level of freezing response is evaluated (the other measures, however less frequently used, are suppression of previously learned licking behavior, ultravocalization, heart rate changes, and defecation frequency). Importantly, this procedure leads to an extinction of the training-evoked behavior, which results in a decrease of the fear evoked by the context, because the animal learns that the context no longer predicts the shock. It has been demonstrated that although the formerly established associations remain strong, the performance in the presence of the extinguished stimulus is diminished (44). Thus the extinction seems to be an active learning process that suppresses, rather than removes, the traces of former learning. Furthermore, the retrieval of the fear memory appears to involve an active, protein synthesis-sensitive component termed reconsolidation (337). Hence, one would expect gene-expression component to the retrieval sessions. Importantly, the gene expression changes observed following the retrieval session (without any foot shock) clearly argue against a notion that painful experience alone is responsible for those changes.
The basolateral part of the amygdala (and its parts, lateral and/or basal nuclei) were the most frequently described as activated, as far as c-Fos expression is concerned, in both contextual and cued fear memory retrieval (for reference, see Table 1). The activation of the basolateral part of the amygdala has also been shown in extinction of fear (176). In contrast to the original training (see above), in most of the studies activation of the central nucleus in both contextual and cued fear memory retrieval was observed (see Table 1). The c-Fos activation within the medial and cortical parts of the amygdala following fear memory retrieval was investigated less frequently, and the existing results are rather contradictory (see Table 1).
To summarize, it appears that the basolateral and central parts of the amygdala are active during contextual and cued fear memory retrieval, whereas the data on the activity of the medial and cortical nuclei of the amygdala are inconsistent.
The two-way (active) avoidance behavior is acquired in a shuttle-box apparatus that consists of two compartments. Both of them are equipped with a source of a CS and a gridded floor through which a US, a foot shock, can be delivered. The animal is originally placed in one of the two compartments, then the CS is presented and followed by the US within a few seconds. The animal is supposed to learn the signaling value of the CS and to avoid the US by moving to the opposite compartment. The training includes several sessions, which are usually composed of a number of trials.
There are reports of the activity of the basolateral, medial, and cortical nuclei, but not the central nucleus, of the amygdala after one session of two-way avoidance training (29, 99, 315, 428). Notably, Savonenko et al. (428) applied factor analysis to separate and group a number of behavioral components observed during the training. Surprisingly, no correlation was found between the level of c-Fos expression observed in any of the 13 amygdalar subdivisions that were analyzed and either avoidance reaction or sum of the shock received. On the other hand, c-Fos expression in the cortical amygdala correlated with grooming behavior (reflecting the lack of fear), whereas the c-Fos levels in the basolateral and medial amygdala correlated with anticipatory anxiety.
In the conditioned emotional response training, the animals that previously acquired an instrumental response, e.g., bar-pressing for a food reward, are subsequently trained to learn that an acoustic stimulus signals a foot shock. The suppression of the instrumental reaction is treated as a measure of acquired emotional response. The training is usually composed of four or five sessions, each including several trials.
The increased c-Fos expression after the first session of such training was observed in the basolateral and medial, but not in the cortical, nuclei of the amygdala (229, 455). The results on c-Fos expression in the central amygdala are inconsistent (see Table 1).
When consumption of a food with a novel taste is followed by exposure to a toxin, animals will avoid that taste in the future. This phenomenon is used in the conditioned taste aversion (CTA) test, in which learning is achieved in a single trial. The water-deprived animals learn to associate a novel taste (e.g., sucrose or saccharin) with the malaise caused by a toxin (e.g., lithium chloride, injected after the onset of drinking of sweetened water). Thus CTA is a form of classical conditioning in which animals avoid a taste (CS) that has previously been paired with a treatment that produces transient illness (US). However, CTA differs from most forms of classical conditioning in the speed and efficacy with which the animals acquire the conditioned response. CTA learning is rapidly gained, tolerates long delays between CS and US (they can reach even a few hours), and is very robust.
The involvement of the amygdala in conditioning of taste aversion measured by c-Fos expression was studied apparently only in the central and basolateral nuclei of the amygdala. The conditioning of taste aversion was shown to induce the elevated level of c-fos and its protein expression in the central amygdala (for reference, see Table 1). However, there are some noticeable discrepancies between the results of those studies. Navarro et al. (340), who compared the level of c-Fos expression in the animals in which saccharin was paired with lithium chloride either once or three times, observed activation of the central amygdala only when three pairings of the stimuli were applied. In this study, the control group consisted of animals in which lithium chloride was replaced by NaCl. Notably, the authors observed the behavioral effects, i.e., rejection of saccharin, following both the first and third trial. In contrast, Lamprecht and Dudai (254), Koh and Bernstein (231), and Wilkins and Bernstein (513) showed activation of the central nucleus after a single trial of taste aversion conditioning. However, Lamprecht and Dudai (254) did not observe the difference in the c-fos expression between CTA-trained animals and lithium chloride-injected rats. On the other hand, Koh and Bernstein (231) as well as Wilkins and Bernstein (513) showed much lower levels of c-Fos expression in the central amygdala following either the lithium chloride injection itself or pairings of the familiar taste with the lithium chloride injection (familiarity of the CS precluded learning of CTA) compared with the animals trained in the CTA.
In the above-mentioned studies, the distribution of the c-Fos-positive neurons within subdivisions of the central amygdala was not investigated. Interestingly, Yamamoto et al. (519) showed that the injection of the lithium chloride induced the elevated c-Fos expression mainly in the lateral part of the central nucleus of the amygdala. However, the influence of the injection itself was not controlled in this study. In contrast, drinking of sucrose elicited the increased c-Fos expression mainly in the medial part of the central nucleus. On the other side, the intraoral infusion of sucrose after CTA induced c-Fos expression evenly in the central amygdala (519).
The data on c-Fos activation in the basolateral amygdala following conditioning of taste aversion are not consistent (for reference, see Table 1). Interestingly, Wilkins and Bernstein (513) found the significant difference in the pattern of c-Fos expression following CTA established by exposing rats to a novel taste CS through a bottle or through intraoral infusion. Conditioning rats with the bottle method led to the increased c-Fos expression in the basolateral and central amygdala, whereas intraoral infusion induced the activation restricted to the central nucleus of the amygdala.
The injection of lithium chloride alone induced the elevated level of c-fos and its protein expression in the central nucleus of the amygdala, whereas the data on activation of the basolateral amygdala following the lithium chloride injection are less consistent (for reference, see Table 1).
It is known that rats rapidly become anorectic when eating an amino acid-imbalanced diet that induces a deficiency of an indispensable amino acid. Recognition of amino acid deficiency that leads to learned taste aversion was studied by Wang et al. (496). They found the increase in c-Fos expression in the central amygdala of rats following the introduction of a diet imbalanced in threonine.
In aggregate, it appears that the central and basolateral nuclei of the amygdala are activated during conditioned taste aversion learning and retrieval, whereas no central amygdala activation was noted during extinction (see Fig. 6). However, it is not clear, especially in the case of the central amygdala, if central amygdala activation is evoked by learning of associations between taste CS and US or coding information about taste CS and US themselves.
|
The expression of other members of Fos/Jun family in the amygdala was studied following CTA acquisition, as well as after lithium chloride injection alone. The increased level of fosB expression was not observed in the central and basolateral amygdala following either CTA training or lithium chloride injection (254, 462). On the other hand, it was shown that c-Fos upregulation in the central nucleus after the injection of lithium chloride is accompanied by the increased expression of JunB (462).
Besides c-Fos, one of the most frequently applied gene activity markers is Zif268 (also known as NGFI-A, Egr-1, Krox-24, ZENK, and TIS-8; see Ref. 227). Almost all studies on the expression of Zif268 within the amygdala were carried out following fear conditioning, and they have been focused on the lateral nucleus of the basolateral amygdala and the central nucleus. Unfortunately, the results of the Zif268 expression following fear conditioning in the lateral amygdala are rather inconsistent. The increase in the zif268 mRNA expression was demonstrated following contextual fear conditioning (284–286, 410), as well as following cued fear acquisition (176). In contrast, Hall et al. (163) reported nonspecific induction of zif268 mRNA, i.e., the expression of zif268 was increased in the lateral nucleus of the amygdala not only in the experimental group but also in all control groups exposed to the training chamber compared with naive controls. Moreover, Weitemier and Ryabinin (504) have shown lack of differences in the Zif268 expression within the amygdala between fear-conditioned and naive mice after the training phase. In this study, both delayed conditioning (the US delivered simultaneously with the last period of CS application) and trace conditioning (the US delivered after a temporal gap after the CS) procedures were applied. In contrast to the lateral nucleus of the amygdala, none of the studies reported activation of the central nucleus following fear acquisition (163, 284, 285).
Similarly inconsistent and incomplete are the results on the expression of Zif268 in the lateral nucleus of the amygdala following fear memory retrieval (161, 176, 286, 410).
Taken together, the data on the Zif268 expression in the lateral nucleus of the amygdala following fear conditioning and memory retrieval are inconclusive. The observed discrepancies might have resulted from different procedures that were used by the authors (for extensive discussion on this issue, see Ref. 227). On the other hand, it appears that no elevated Zif268 expression occurs within the central amygdala following fear conditioning. Interestingly, the expression of another transcription factor protein related to Zif268, NGFI-B, was found to be augmented in the lateral nucleus of the amygdala following contextual fear conditioning, but not after contextual fear memory retrieval (286).
The expression of other immediate early genes, Arc and ICER, was investigated only following very few kinds of behavioral trainings. For instance, the increased arc mRNA expression was found in the basolateral and central nuclei of the amygdala after contextual memory retrieval (529). Furthermore, the expression of ICER was shown to be increased in the central, but not the basolateral, amygdala following both lithium chloride injection and learning of CTA (254, 456; see also Ref. 319).
Phosphorylated form of cAMP-responsive element binding transcription factor (P-CREB) is also believed to serve as a marker of neuronal activity. The increased expression of P-CREB was noted in the basolateral amygdala after both contextual and cued fear conditioning (457, 502). Moreover, the increased level of P-CREB expression was shown in the central amygdala following contextual fear conditioning (457). On the other hand, the elevated expression of P-CREB was observed in the basolateral amygdala, but not in the central nucleus of the amygdala, following cued fear memory retrieval (162). Furthermore, the expression of P-CREB in the basolateral, central, medial, and cortical nuclei of the amygdala was reported after one session of two-way avoidance training (420).
The elevated level of BDNF mRNA expression was seen in the basolateral amygdala after both visually and olfactory cued fear conditioning, but not in the medial amygdala following visually cued fear conditioning (390).
Summing up, the pattern of expression of the other gene activity markers in the amygdalar nuclei following different behavioral tests often, although not always, resembles the pattern seen with the use of c-Fos expression. For instance, fear conditioning induced the elevated level of expression of c-Fos and P-CREB in the basolateral part of the amygdala. Similarly, both injection of lithium chloride and learning of CTA caused the increased level of c-Fos, JunB, and Zif268 expression in the central amygdala. However, there are also clear discrepancies, e.g., on activation of the basolateral part of the amygdala measured with the c-Fos and Zif268 expression following fear conditioning and retrieval.
D. Pattern of c-Fos Expression in the Amygdalar Nuclei in Aversively Motivated, Ethologically Relevant Behavioral Tests
In addition to the standard behavioral tests, there are also paradigms that employ more natural stimuli, in a hope to evoke evolutionarily and adaptively formed unconditioned emotional states. Notably, they are often used and interpreted in different ways: in the standard tests predominantly learning processes are assessed, while in the more ethologically relevant behaviors mainly the evoked emotional states are considered. However, as in the standard tests, in the ethologically based behavioral paradigms both learning and emotions are involved and should be taken into account. For instance, the existence of learning processes during exposure to a predator is proven by the observation that rats develop conditioned fear to both context and cues that have been paired with, e.g., a cat odor (for review, see Ref. 97). Another piece of evidence is provided by the result of Figueiredo et al. (114), who observed a clear decrease of c-fos expression in the medial amygdala after the seventh day of habituation to the cat presence compared with the level of expression after the first session of such exposure. Another example, supporting the involvement of learning processes, is a phenomenon of avoiding a familiar winner by the subordinate animals 1 day after a fight (249). These data suggest that the animals learned to recognize the winner.
To evoke an innate fear, in a more ethologically relevant approach, exposure to a high, open, and/or highly illuminated space, a live predator, a predatory cue (e.g., an odor), an alarm signal (e.g., ultrasonic vocalization), or aggressive conspecifics are used. The elevated plus-maze and the elevated T-maze are considered as ethologically based animal models of anxiety. The plus-maze consists of four arms, which form a regular cross and are elevated over the ground. Two opposing arms are enclosed by side walls, and the remaining two are open. Due to the innate fear of height and open space, the animals remain longer in the enclosed arms. The more pronounced this innate fear is, the longer they stay in the enclosed arms, e.g., it is known that the animals receiving anxiolytic drugs enter more often and remain longer in the open arms. The elevated T-maze is derived from the elevated plus-maze and consists of three elevated arms, one enclosed and two open. An animal is supposed to perform two tasks in the apparatus: the one-way escape and the inhibitory avoidance. For the former task, an animal is placed in an open arm, from which it tends to escape. The time of escape latency is measured. This procedure is supposed to represent innate fear. For the inhibitory avoidance an animal is placed in the enclosed arm and the time to leave the arm is measured. Due to innate fear the animals will learn to avoid the open arms and stay longer in the enclosed arm; therefore, this procedure is supposed to represent conditioned fear. This test usually involves several trials.
The elevated c-Fos expression was observed in the basolateral, medial, and cortical, but not in the central, parts of the amygdala after exposure of rats to the elevated plus-maze (99, 450, but see Ref. 274 as well as Table 2 and Figure 7). The expression of c-Fos after exposure to the one-way escape task of the elevated T-maze was increased in the basolateral part of the amygdala, whereas after exposure to the inhibitory avoidance task, in the medial amygdala (451).
|
|
In the predatory exposure procedure, in which rats were confined to a small box and exposed to a live cat, Figueiredo et al. (114) found the increased c-fos mRNA expression in the medial nucleus of the amygdala. Similarly, the exposure to a cat odor in rats was shown to induce an increased level of c-Fos expression in the medial, but not the basolateral and central nuclei of the amygdala (96, 308). Moreover, Dielenberg et al. (96) showed lack of activation of c-Fos in the cortical nuclei of the amygdala after exposure to the cat odor. In both studies the examined rats were presented to the odor in the arena equipped with the small chamber, which made it possible for the rats to hide.
Using a fox odor, Funk and Amir (131) showed the activation of the basolateral, but not the central, parts of the amygdala. In contrast, Day et al. (89) applying 2,5-dihydro-2,4,5-trimethylthiazoline (TMT), an isolated chemical component from fox feces, observed elevated c-fos mRNA expression in the central and medial, but not basolateral and cortical, nuclei of the amygdala. In this study, the rats were presented to the fox odor in small cages without any place to hide.
Summing up, the predatory presence, or even its odor, seem to evoke the augmented level of c-Fos expression in the medial, but not cortical, nuclei of the amygdala. The data on the c-Fos expression in the basolateral and central amygdala are less consistent. The discrepancies might stem from different odors or odor concentrations that were used in the above-mentioned studies, but also from different experimental conditions, e.g., presence or absence of a place to hide.
Recently, we have described an experimental rat model of between-subject transfer of emotional information and effects of the transferred fear on activation of the amygdala. Briefly, the rats were kept in pairs, and one animal ("demonstrator") was treated either to a foot-shock reinforced context conditioning or exposed to a novel cage without any foot-shocks. Then the demonstrator rats returned to the home cages, to the other animals (called "observers"). We have found that by measuring startle reflex and exploration index we can distinguish between observers paired with demonstrators of both kinds. When we examined the influence of the demonstrators' presence on the activity of the amygdalar nuclei of the observers, we found that the observers had in the all amygdalar subdivisions the same levels of c-Fos expression as demonstrators, except for the central nucleus, where the observers displayed more c-Fos than the demonstrators (see Ref. 228; see also Fig. 8).
|
20 kHz, which were able to induce defense behavior. Similarly, Mongeau et al. (327) observed the increase in c-fos mRNA expression in the basolateral, medial, and cortical nuclei of the amygdala in mice exposed to the ultrasonic vocalization. In contrast, they did not show the increased c-fos expression in the central nucleus of the amygdala. The exposure to aggressive conspecifics is the other behavioral paradigm used to evoke innate fear or aggression. In rodents, aggression against conspecifics is a form of social behavior in which adult animals fight to establish dominance relationships. This is used in the resident-intruder test, in which an unfamiliar animal, an "intruder," is placed into the home cage of the other animal, "resident," to provoke an encounter. Kollack-Walker and Newman (235) investigated the pattern of c-Fos expression in the male Syrian hamster amygdala in the resident-intruder test. They found activation of the medial and cortical, but not the basolateral, part of the amygdala in both dominant and subordinate animals. Furthermore, Kollack-Walker et al. (237) studied the c-fos mRNA expression in the hamster amygdala in similar conditions, except for much intense habituation to handling and to the novelty of another male's cage and odors. They found the increased c-fos expression in the medial, but not basolateral, central and cortical nuclei of the amygdala in the dominant males. In contrast, the animals that experienced social defeat had activated the central and medial, but not basolateral and cortical, nuclei of the amygdala. The activation of the central nucleus of the amygdala after experiencing social defeat was also observed in the study of Kollack-Walker et al. (234). Overall, in the dominant animals, the increased level of c-Fos expression was consistently seen in the medial nucleus of the amygdala (for reference, see Table 2), in both males and females, in hamster, mouse, and prairie vole. Moreover, the activation of the cortical nuclei was often shown in such conditions, whereas the results for the basolateral and central nuclei of the amygdala are less consistent (see Table 2).
Taken together, it appears that the aggressive encounter consistently induced the increased level of c-Fos expression within the medial amygdala (see Fig. 7), whereas the data on activation of the basolateral, central, and cortical nuclei of the amygdala are not consistent. These discrepancies might have been a result of using different experimental conditions, as well as various species of animals employed.
The obtained data also show that the pattern of c-Fos expression in the amygdala may depend on the result of aggressive encounter, being different for its losers and winners. It is known that after an aggressive encounter, a loser selectively avoids his own, familiar winner but does not avoid other males. This suggests that the losers are able to recognize and remember the familiar individual. The neural basis of this phenomenon was studied by Lai et al. (248). They showed activation of the basolateral, but not central and medial, parts of the amygdala in such conditions.
E. Pattern of Expression of Other c-fos Gene Activity Markers in the Amygdalar Nuclei in Aversively Motivated, Ethologically Relevant Behavioral Tests
There are very few pieces of data on the expression of gene activity markers other than c-fos in the amygdala following ethologically relevant behavioral tests. Moreover, they are often not consistent with the c-fos expression results. For instance, in the predatory exposure procedure, in which rats were exposed to a live cat, Rosen et al. (409) did not find the increased zif268 mRNA expression in the lateral nucleus of the amygdala. In this study the examined rats were presented to the cat in the large arena without separate compartments for both animals. The authors compared the expression of zif268 between rats either exposed to a live cat or just handled or confined to a small box. Furthermore, Lai et al. (248) studied Zif268 expression in a loser that selectively avoided his own, familiar winner but did not avoid other males. They investigated the basolateral, central, and medial parts of the amygdala and did not observe the elevated expression of Zif268 in any of the studied nuclei. Activation of the nuclei of the amygdala following aggressive encounter was studied also by Gammie and Nelson (137). Using P-CREB, they did not find the increased activation of the medial and cortical amygdala.
F. Pattern of c-Fos Expression in the Amygdalar Nuclei in Appetitively Motivated Standard Behavioral Tests
There is much less data on the pattern of c-fos expression in the amygdalar nuclei in the appetitively motivated behavioral tests than in the aversively motivated ones. In most of the studies, palatable food or sweet solutions were applied as reinforcements in classical and instrumental conditioning. The increased c-Fos expression in the central nucleus of the amygdala was a result of just drinking such sweet solutions as sucrose or saccharin (for reference, see Table 3). However, Lamprecht and Dudai (254) did not observe the elevated c-fos mRNA expression in the central amygdala following drinking of sweet solution; however, they used rather diluted, 0.1% saccharin. Interestingly, sucrose caused stronger activation of the central nucleus than saccharin. This effect was observed after both drinking and intragastric infusions (519). Furthermore, Koh et al. (232) showed that a novel taste of saccharin induced a greater c-Fos response in the central nucleus of the amygdala than a familiar taste. This effect was dependent on taste intensity, i.e., 0.5% solution evoked stronger response than 0.15% solution of saccharin. A similar effect of novelty was observed by Barot and Bernstein (23) for isotonic concentration of NaCl but not for a highly palatable polysaccharide preparation Polycose (for extensive discussion of possible causes of the observed discrepancies, see Ref. 23). Moreover, it was shown that the increased level of c-Fos expression in the central amygdala induced by drinking of sweet solutions was restricted to the medial part of the nucleus (519).
|
Park and Carr (357) exposed rats to a palatable meal and the meal-paired environment. They did not observe the increased level of c-Fos expression in the central and basolateral parts of the amygdala. Noteworthy, the c-Fos expression was measured after the 11th of daily sessions; thus the animals could have been well habituated both to the cage and the food. Interestingly, Timofeeva et al. (472) observed the augmented level of c-fos expression in the central amygdala following the ingestion of a meal after 24 h of fasting (refeeding procedure). Summing up, the palatable food or sweet solutions induced the increased level of c-Fos expression in the central amygdala, but rather not in the basolateral amygdala.
In Pavlovian appetitive conditioning task, in which a signaling stimulus (CS) is paired with a reinforcement (US), animals acquire two distinct responses: an orienting response to the CS source and approaching the site of reinforcement delivery. Lee et al. (264) examined c-Fos expression in the central amygdala in a Pavlovian appetitive conditioning task, in which a visual CS was paired with food. They found activation of the medial, but not lateral subdivisions of the central nucleus in such training. Moreover, the more intense the training was, the more c-Fos expression in the central nucleus of the amygdala was observed.
In the experiment of Hess et al. (177), rats were trained to nose-poke for water reward. The authors found c-fos activation within the basolateral, medial, and central nuclei of the amygdala following such training, compared with the home-caged controls. However, the expression was studied after several training days; thus it could have been much less pronounced than after the first training session (see above).
Hess et al. (177) examined also the level of c-fos mRNA expression in the amygdalar nuclei in two-odor discrimination test. In this task, the animals learned to nose-poke at a positive odor (peppermint) port for a water reward and to avoid a negative odor (amyl acetate) port, which was punished with a brief strobe flash. They found activation of the basolateral and medial, but not central, nuclei of the amygdala. Noteworthy, the animals were well habituated to the training apparatus before the training session.
Tronel and Sara (476) investigated c-Fos expression in the rat amygdala in appetitively motivated odor-discrimination task. In this test animals poke their noses into a sponge soaked with a target odor to obtain a reward (usually palatable food). They learn to discriminate between different odors (in the experiment of Tronel and Sara: lemon, almond, and mint). The authors found activation of the basolateral but not central amygdala compared with the yoked control. The yoked control animals were exposed to the reward immediately before the trial and then explored the apparatus containing the three odorous sponges without the reinforcement for the same amount of time as the experimental animals. In the retrieval trial 24 h after the training, activation of neither the basolateral nor central amygdala was observed. The retrieval trial consisted of one reinforced trial, similar to that used during the training.
The pattern of activation of the amygdalar nuclei following bar-pressing response reinforced with food was investigated by Knapska et al. (229). Rats that had already acquired a bar pressing response to a partial food reinforcement were further trained to learn that an acoustic stimulus signaled continuous food reinforcement (more frequent than before). Significantly greater levels of c-Fos expression were observed in the central, cortical, and lateral nuclei of the amygdala after the first session of such training, compared with the prolonged training (Fig. 9).
|
G. Pattern of Expression of Other Activity Markers in the Amygdalar Nuclei in Appetitively Motivated Standard Behavioral Tests
David et al. (86) reported that the expression of P-CREB was increased in the central and medial nuclei of the amygdala following the training of bar-pressing for a delayed food reinforcement (the reward was delivered 1 min after a lever press). In parallel to our studies on c-Fos (229), we have also investigated expression of Zif268 following bar-pressing response reinforced with food. As shown on the Figure 9, no significant differences were observed between Zif268 expression after the 1st versus the 10th training session.
H. Pattern of c-fos Expression in the Amygdalar Nuclei in Appetitively Motivated, Ethologically Relevant Behavioral Tests
Results concerning the pattern of c-fos expression in the amygdalar nuclei that was studied in appetitively motivated, ethologically relevant behavioral tests are summarized in Table 4.
|
In all the studies, in which the response of the amygdalar nuclei to mating was examined, the medial nucleus was shown to be active (for reference, see Table 4; see also Fig. 10). Furthermore, very similar results were obtained for males and females; rat, hamster, ferret, musk shrew, and macaque; and both sexually experienced and naive animals. In those experiments, the following control groups were used: the home-caged animals, the animals exposed to the testing cage, or mounting without intromissions. Similarly, artificial vaginocervical stimulation induced the increased level of c-fos mRNA and protein expression in the medial nucleus of the amygdala (81, 368, 469).
|
Interestingly, Erskine and Hanrahan (102) found that paced females (i.e., the animals that were able to pace the timing of the intromissions and ejaculations) had a higher level of c-Fos expression in the medial amygdala than the nonpaced females with the same number of intromissions. On the other hand, Alexander et al. (4), who studied activation of the medial nucleus of the amygdala in high-sexually performing, low-sexually performing, and male-oriented rams, following noncontact sensory stimulation from estrous ewe or other males, did not observe any differences in the level of c-Fos expression between type of ram and sex of stimulus animals.
In all but one study, in which the c-Fos expression in the cortical nuclei after mating was investigated, their activation was seen (for reference, see Table 4). Similarly, the activation of the cortical nuclei was shown after the artificial vaginocervical stimulation (368).
The data on the involvement of the basolateral and central nuclei of the amygdala in the sexual behaviors are very scarce, mostly because the authors did not pay much attention to those regions of the amygdala. Pfaus et al. (368) observed the slightly increased level of c-Fos expression in the basolateral and central amygdala after mating and the artificial vaginocervical stimulation. In addition, Greco et al. (149) showed the elevated c-Fos expression in the central amygdala after mating. However, in the other studies, the change in the level of c-Fos expression in the central and the basolateral amygdala was not observed (for reference, see Table 4).
It was also seen that exposure to an inaccessible estrous female or chemosensory cues from sexually mature partner are sufficient to evoke the increased level of c-Fos expression in the medial nucleus of the amygdala (for reference, see Table 4). Activation of the cortical nuclei was not observed in male rats after exposure to inaccessible estrous female (217). On the other hand, the data on the c-Fos expression in the cortical nuclei following exposure to chemosensory cues from sexual partner are not consistent (see Table 4). The increased level of c-Fos expression was not seen in the basolateral and central parts of the amygdala following exposure to the inaccessible estrous female in rats (278) and in the basolateral amygdala after exposure to the chemosensory cues in male rats (222; but see Ref. 326).
There are some data indicating that subdivisions of the medial nucleus of the amygdala may play different functional roles. For instance, Hosokawa and Chiba (190), who studied the effect of sexual experience on conspecific odor preference in male rats, found that sexually experienced males, which preferred the odor from an estrous female as opposed to the odor from a sexually active male, had the elevated level of c-Fos expression in both the anterodorsal and posterodorsal parts of the medial amygdala. In contrast, naive males, which did not show the preference for an estrous odor, had the increased c-Fos expression only in the posterodorsal part of the medial amygdala. Interestingly, Fernandez-Fewell and Meredith (111) found that the anterior part of the medial amygdala is active in naive male hamsters following mating behavior, whereas the level of c-Fos expression in this part of the amygdala is significantly lower in anosmic males.
Summing up, mating or just the cues from sexually mature partner were shown to induce c-Fos expression in the medial and cortical nuclei of the amygdala (Fig. 10); however, it is not clear which aspects of sexual contact are crucial for the gene induction. It is also conceivable that studying the gene expression in different parts of the medial nucleus might help elucidate this problem. Activation of the basolateral and central amygdala was investigated only in a few studies (see Table 4); thus the data are not conclusive. Though, it seems that both the central and basolateral nuclei of the amygdala can be activated by mating.
Some aspects of explicit learning with the use of sexual reinforcements were also investigated. For instance, Kollack-Walker and Newman (236) observed the elevated c-Fos expression in the medial amygdala of male hamsters after exposure to the experimental context previously associated with mating. Moreover, MonchoBogani et al. (326) studied involvement of the amygdalar nuclei in creating associations between male-derived volatile and nonvolatile pheromones. In mice, nonvolatile pheromones contained in soiled bedding are innately attractive for adult females, whereas male-derived volatiles become attractive only if paired with nonvolatile pheromones in the Pavlovian associative learning. After the preference test, in which chemically inexperienced and experienced females choose between volatiles from clean or male-soiled bedding, the authors found that the cortical nuclei of the amygdala responded to male-derived volatiles in both experienced and inexperienced mice, while the basolateral amygdala was activated selectively in the experienced animals. The basolateral amygdala was not only activated during expression of this memory, but also during learning in the first exposure of females to pheromones and volatiles from male-soiled bedding (326). Furthermore, Kippin et al. (222) observed that male rats exposed to bedding scented with an almond odor paired previously with copulation had the increased level of c-Fos expression in the basolateral but not in the medial nucleus of the amygdala (see Table 4).
Taken together, it seems that the basolateral amygdala is especially engaged in social tasks that require learning. Owing to the limited data on the c-Fos expression within the other nuclei of the amygdala during such learning, it is difficult to draw the conclusion about their involvement.
2. Maternal/paternal interaction
The expression of c-Fos in the amygdalar nuclei after mother-litter interaction was studied in females exposed to the pups or sensory cues from the pups. In the former paradigm, the test animals are either postpartum females from which all pups are removed during parturition or virgin females becoming maternal by exposing them to the foster pups. The maternal behaviors, like licking, sniffing, nest building, retrieving, crouching, or suckling, are scored. In the latter paradigm, the pups hidden in the closed perforated box that is placed in the testing cage are presented to the females. Thus the test animals are exposed to olfactory and auditory cues from the pups. Following these paradigms, the c-Fos expression was measured in the brains of females after some time (30 min to 2 h) of interacting with the pups. The animals exposed to the adult familiar animal, unfamiliar virgin females that did not behave maternally, novel food, or home-caged animals were used as controls.
In almost all studies, in which the response of the amygdalar nuclei to mother-litter interaction, as well as to exposure to the sensory cues from pups was examined, the medial nucleus was shown to be activated (for reference, see Table 4; see also Fig. 10). Similarly, the elevated c-Fos expression in the cortical nuclei of the amygdala was observed in almost all studies in which this issue was investigated (see Table 4). The increased level of c-Fos expression in the basolateral and central amygdala in the maternally behaving animals was also seen (118, 120, 495). However, this increase was much weaker than in the medial and cortical nuclei of the amygdala. Moreover, there are other studies, in which activation of the basolateral amygdala was not observed following maternal interaction (119), as well as, of the basolateral and central amygdala after exposure to sensory cues from pups (118).
Interestingly, Kirkpatrick et al. (223) found the elevated level of c-Fos expression in the medial amygdala of both males and females of the highly social prairie vole following exposure to the pup compared with a nonsocial olfactory stimulus. Thus these results provide the evidence that the medial nucleus is activated also during paternal behavior. To sum up, an exposure to the pups or sensory cues from the pups activated the medial and cortical and, less certainly, the basolateral and central nuclei of the amygdala.
The question of involvement of the parts of the amygdala in forming lamb olfactory memory was directly addressed by Keller et al. (215). In sheep, recognition of the lamb by its mother depends on learning of the olfactory signature of the lamb after parturition, which takes place within the first 2 h postpartum. The authors studied the pattern of c-Fos expression in the amygdala of intact and anosmic ewes 2 h after parturition (the anosmic ewes display maternal behavior but not individual lamb recognition). They found the increased level of c-Fos expression in the cortical amygdala of intact ewes compared with the anosmic animals. Interestingly, neither the basolateral, nor the medial nuclei of the amygdala were more active in the intact ewes.
The c-Fos activation within the amygdalar nuclei was also investigated following a reexposure to pups and pup-associated cues. Unfortunately, the acquired data are highly inconsistent, which may stem from profound differences in the experimental paradigms that were used. For instance, Fleming and Korsmit (118) used a paradigm in which postpartum rats were given a 2- or 4-h interactive experience with 2-day-old pups on day 1 postpartum and then were reexposed to 2- or 3-day-old pups located in a perforated box or to a neutral stimulus (perforated box only) 4 or 10 days later. The authors found the consistent activation of the basolateral, but not central, medial, and cortical, nuclei of the amygdala after reexposure of pups and pup-associated cues compared with inexperienced females. In a similar behavioral paradigm, Fleming and Walsh (120) observed the involvement of the basolateral, central, medial, and cortical nuclei of the amygdala in the first contact of parturient rat dams with the pups on day 1 postpartum. However, reexposure to the pups or to the sensory cues from them induced the elevated c-Fos expression in the basolateral and central amygdala regardless of prior maternal experience. In contrast, Li et al. (272) found selective activation of the medial nucleus of the amygdala in animals exposed to their pups, both interacting with the pups and exposed only to the sensory cues from them. They test the dams on day 11 postpartum after 48 h of separation from their litter, comparing the effects of interaction with the pups, exposure to pups separated with the wire screen, and no exposure to the pups. Moreover, Lonstein et al. (276) showed activation of the medial amygdala of lactating rats that were reexposed to the pups on day 7 postpartum after 48 h of separation from their litter. They compared c-Fos activation to the level of its expression in unstimulated dams. However, in the same experimental paradigm, Lonstein et al. (277), comparing activation of the amygdala of dams 1) interacting with the pups that were either capable or incapable of suckling, 2) exposed to the inaccessible pups in a wire-mesh box, and 3) exposed to a wire-mesh box alone, did not observe differences in the level of c-Fos expression in the medial amygdala. The expression was elevated only compared with the home-caged controls. In contrast, activation of the cortical nuclei of the amygdala was induced only by nonsuckling pups but not by the suckling ones (the nonsuckling pups had their mystacial pads anesthetized to make them incapable of attaching to a nipple).
In summary, reexposure to pups and pup-associated cues induced the elevated c-Fos expression in the basolateral part of the amygdala (see Table 4). The data on involvement of the medial as well as the central and cortical amygdala are inconclusive.
3. Nonagonistic social interaction
The social interaction with unfamiliar conspecifics of the same or other sex that did not evoke sexual or aggressive behaviors was shown to activate the medial amygdala in mouse, hamster, macaque, and prairie vole (for reference, see Table 4). However, this activation was weaker than in either aggressively or sexually behaving animals. The data on c-Fos activation in the cortical and central nuclei in such situation are inconsistent (see Table 4).
Lopez and Ettenberg (278) investigated activation of the amygdalar nuclei after exposure to cues from conspecific. They showed that exposure of male rats to a nonestrous female located behind a perforated, Plexiglas partition did not elevate the level of c-Fos expression in the basolateral and central amygdala. Moreover, Meredith and Westberry (311), who studied the c-Fos expression in the medial amygdala following exposure to various chemosensory stimuli from conspecifics and heterospecifics, showed that the medial nucleus was activated by both, conspecific and heterospecific, stimuli. However, they observed that the anterior part of the medial amygdala responded to both conspecific and heterospecific chemosignals, whereas the posterior part responded distinctively to the conspecific stimuli. This observation indicates again that subdivisions of the medial nucleus of the amygdala may play different functional roles.
Taken together, it seems that the medial nucleus of the amygdala responded during both social interaction and exposure to cues from conspecific, whereas the data on the c-Fos expression in the cortical, central, and basolateral nuclei of the amygdala in such situations are sparse and contradictory.
I. Pattern of Expression of Other Gene Activity Markers in the Amygdalar Nuclei in Appetitively Motivated, Ethologically Relevant Behavioral Tests
Very little is known on expression of gene activity markers other than c-Fos in the amygdala following appetitively motivated ethologically relevant behavioral tests. Moreover, the existing results are not fully consistent with the data gathered on c-Fos expression. For instance, the increased level of c-Jun, but not JunB and JunD, induced by mating was shown in the medial nucleus (117, but see Ref. 508).
The level of zif268 expression was investigated following mating and maternal behavior only in the medial nucleus of the amygdala. It was shown that mating induced the elevated level of Zif268 expression in the medial nucleus (375, 508); however, such effect was not observed after the vaginocervical stimulation and maternal behavior.
J. The IntelliCage System: In-Between of Standard Tests and Ethologically Based Behavioral Paradigms
The IntelliCage system was designed to employ more natural stimuli in well-controlled experimental conditions, as well as to minimize the stress of handling and separation in animals. It is an automated learning apparatus assessing spontaneous and learning behavior of group-caged mice (New-Behavior AG; http://www.newbehavior.com). It employs a large standard rat cage (for a detailed description, see Ref. 136). A coverplate holds four operant learning chambers that fit into the corners of the housing cage. Access into the chamber is provided via a tubular antenna reading the transponder codes that allow recognizing individual animals. This design restricts access to the learning chamber for a single mouse only. The chamber, equipped with a proximity sensor, contains two openings permitting access to the nipples of drinking bottles. These openings are crossed by photobeams recording nose-pokes of the mice. Access to the tubes can be barred by small motorized doors. Aversive stimulation can be delivered in the form of air-puffs directed to the head of the mouse through tubing controlled by electric valves. In addition, each cage contains a sleeping shelter in the center, on which the animals could climb to reach the food. The whole setup is controlled by a microcomputer recognizing visits, nose-pokes, and tube-lickings of individual mice, and delivering reward (by opening the access to water after a nose-poke) or punishment, an air-puff (by entering the test chamber) according to preprogrammed schedules depending on the assignment of the mice to different test groups within the same cage. The system can run continuously for several days, with behavioral activity of the mice being monitored from the experimenter office, e.g., via internet.
Knapska et al. (229) carried out the place preference (reinforced with sweetened water) and place avoidance (reinforced with air-puffs) trainings in the IntelliCage system. They found the elevated c-Fos expression in the lateral nucleus of the amygdala following place preference and place avoidance. However, the central nucleus of the amygdala was activated only by the place preference training (see Tables 1 and 3).
K. Comparison of c-Fos Versus Zif268 Expression Patterns Following Behavioral Training
As mentioned above, c-Fos and Zif268 have sometimes been studied in parallel, revealing significant differences between their expression patterns. For instance, Rosen et al. (410) studied c-fos and zif268 mRNA expression in the same rats following contextual fear conditioning. The pattern of c-fos expression did not parallel those seen for zif268. They observed the increased expression of c-fos only in the medial nucleus of the amygdala following fear conditioning; however, it was not specific, i.e., it was increased in both the control and experimental groups. In contrast, zif268 expression was elevated in the lateral nucleus of the amygdala in the fear-conditioned group compared with the control animals. Similarly, Da Costa et al. (81) showed the increased c-fos mRNA expression in the medial nucleus of the amygdala following artificial vaginocervical stimulation and maternal behavior; however, they did not observe such increase in zif268 mRNA expression. Correspondingly, Lai et al. (248), who studied c-Fos and Zif268 expression in a loser that selectively avoided his own familiar winner, have seen the discrepancy between the patterns of the expression of both markers. They showed activation of the basolateral, but not central and medial, parts of the amygdala as measured by c-Fos expression and activation of none of the above-mentioned nuclei as measured with Zif268 expression. Moreover, Dardou et al. (84) showed increased level of c-Fos and Zif268 expression in the basolateral amygdala following taste-potentiated odor aversion retrieval; however, in the same test in the central amygdala, they observed no changes in the c-Fos expression and the elevated level of Zif268 expression. Also our results presented on the Figure 9 clearly show discrepant patterns of c-Fos and Zif268 expression in the amygdala following appetitive conditioning.
In contrast, Lamprecht and Dudai (254) observed a close resemblance between the pattern of c-fos and zif268 mRNA expression following conditioning of taste aversion and injection of lithium chloride alone, namely, they showed induction of c-fos and zif268 expression in the central but not basolateral nuclei of the amygdala in both paradigms. It has also been shown that both c-Fos and Zif268 expression was elevated in the medial nucleus of the amygdala following mating (375, 508). Furthermore, Herry and Mons (176) observed the elevated level of both c-Fos and Zif268 expression in the basolateral amygdala following cued fear acquisition, as well as following extinction of cued fear.
In aggregate, the observed differences convincingly show that both transcription factor proteins serve different functions in the active neurons, as has been stressed elsewhere (see Refs. 206, 208).
|
IV. GENE ACTIVITY MARKERS AND DRUGS OF ABUSE |
|---|
|
TopPreviousNextReferences |
|---|
In the present review we describe effects of such abused substances as psychostimulants (cocaine, amphetamine, methamphetamine, MDMA, nicotine, and methylphenidate), opioids (morphine and heroin), cannabinoids (THC, H-210, CP 55,940), hallucinogen (LSD), and ethyl alcohol, following single and repeated, forced and voluntary drug administration, withdrawal, reinstatement of drug-seeking behavior, as well as exposure to drug-paired context and/or cues. We focus our considerations on the expression pattern of the most commonly mapped gene activity marker, c-Fos. Additionally, we also mention much less frequent studies on FosB, FRAs, Zif268, and JunB.
Herein, we give only a brief description of experimental paradigms most commonly used in the studies of gene expression induced by drugs of abuse. The animal models of addiction have extensively been described elsewhere (105, 106, 280, 439, 440, 479).
1. Forced drug administration: acute and chronic
Direct drug injection has been the most often employed approach to study the drug effects on gene expression, including the gene activity markers in the amygdala. To get closer to the drug abuse conditions, researchers have used not only a single (termed often as "acute") but also repeated ("chronic") drug application. It has to be, however, noted that these experimental paradigms, whereas capable of showing the pharmacological effects of the drug, only distantly model the drug abuse and addiction phenomena that have important motivational and emotional components, better exemplified in the self-administration studies. Furthermore, the issue of the effects of the injection itself should be carefully considered. It has been repeatedly shown that injection on its own evokes the gene expression, and animals can habituate to this response (see, e.g., Ref. 212). Hence, optimal application protocol, based on several saline/vehicle injections followed by the drug application, should be obeyed, and unfortunately, this has not always been the case.
2. Instrumental intravenous drug self-administration
Drugs of abuse are readily self-administered by animals and, in general, drugs that are self-administered correspond to those that have high abuse potential. Procedure of intravenous drug self-administration demands, however, initiation phase when instrumental reaction, such as pressing the active lever or poking the active hole, is reinforced by natural reward such as food pellets given to starved animals.
During consecutive self-administration acquisition sessions, accomplishment of instrumental reaction results in drug delivery. Drug delivery may be accompanied by presentation of stimuli, such as light illumination or noise, which, in the course of training, acquire secondary rewarding properties and become rewarding CS. Alternatively, the animals are trained to differentiate discriminative stimuli: one signaling drug availability (S+) and another for nonavailability (S–). Drug may be delivered according to different schedules: simple (the number of responses required for an infusion of drug is set at a fixed number), second-order [completion of an individual part of the schedule produces the terminal event (drug infusion) according to another overall schedule], or on progressive ratio (increasing the response requirements for each successive reinforcement). These different schedules of drug delivery allow for measuring drug-motivating properties. Next the maintenance phase follows, when animals self-administer drug for a longer period under stable conditions to develop drug dependence.
In the extinction sessions, the animals are placed into the experimental self-administration cage, but the performance of the instrumental reaction has no programmed consequences (no drug delivery, no cue presentation). Extinction procedures can provide measures of the incentive or motivational properties of drugs by assessing the persistence of drug-seeking behavior in the absence of response-contingent drug delivery.
For the period of withdrawal, the animals are usually kept in their home cages. During this time both drugs and drug-predictive stimuli are not available. Acute withdrawal may be also induced by administration of respective antagonist, such as naloxone or naltrexone following chronic morphine treatment. Abnormal animal behaviors and physiological alterations, observed during this phase, defining withdrawal syndrome, may reflect physical dependence.
Relapse (reinstatement of drug-seeking reaction) may be induced in drug-dependent animals after a period of abstinence by presentation of drug-associated cues or context, stress (e.g., starvation, electric foot shock), or a passive drug delivery. In all but the last case, the instrumental reaction for the drug (drug seeking) may be separated from those induced by the drug. Hence, it may be an objective measure of the drug's reinforcing properties and may reveal the impact of drug-associated cue on such a response (106). Reinstatement session may follow extinction sessions (exposure to drug paired context without drug delivery) immediately (within-session design) or may be carried out on the other day (between-session design). Reinstatement session may also immediately follow the abstinence period without intervening extinction. Extinction and reinstatement sessions may be repeated once or many times, which allows for within-subject Latin square design of the experiment. This phenomenon also shows the endurance of drug craving (439).
3. Conditioned place preference and avoidance
The CPP and conditioned place avoidance (CPA) procedures are used to evaluate rewarding or aversive properties of the drugs. Primary rewarding/aversive properties of the drug are used as an US that is paired with initially neutral environment, which acquire, in the course of conditioning, secondary motivational properties, such that it can act as a CS (479).
In the habituation phase of CPP/CPA training, to avoid the effect of neophobia, the animals are allowed to freely explore experimental apparatus, which consist typically of two or three compartments connected with gates. The time spent in each compartment is measured. Compartments should have different wall paintings (e.g., charts vs. stripes) and floor structure, to be easily distinguished by the animals. During the conditioning phase, the animals are injected with tested drug and immediately after placed in one of the compartments for a short period (e.g., 15–20 min). The entry to the other compartment should be blocked at this stage. On the other day, the animals are injected with saline, or any other control solution, and placed in the opposite compartment. The injections may be repeated several times. On the test day, the animals are placed in the experimental apparatus and are allowed to freely explore all compartments. The time spent in each compartment is measured once again. Preference or avoidance of one of the compartments is the measure of its secondary rewarding/aversive properties acquired due to association with the tested drug.
Alcohol can be drunk by the experimental animals either as a partial substitution of their diet or in alcohol-containing full liquid diet (530). In the commonly used "sucrose fade-away" procedure (242, 243, 423, 501), the animals are encouraged to drink alcohol by mixing low alcohol solutions (2%) with sucrose (10%). Additionally, to increase the motivation for drinking, the animals may have limited for a certain time access to water. In the course of the initiation phase, the alcohol concentration is steadily increased up to 10–16% and the sucrose is excluded. At the end of the initiation phase, part of the animals start to voluntarily drink 10% alcohol with no sucrose added even though the access to water is not limited anymore. During the following maintenance period, animals drink 10–16% alcohol with no sucrose added.
It can be concluded from multiple studies that the animal models of drug addiction are thought to well predict drug abuse potential. Moreover, these models allow for careful extraction of factors involved in acquisition, maintenance, extinction, and reinstatement of drug reinforcement. Furthermore, the animal models of drug addiction may provide an excellent means for studying both the behavioral and biological basis of drug addiction, including gene expression mapping of the brain regions involved in the development and maintenance of the process.
B. Cocaine-Evoked c-Fos Expression
Cocaine is a powerful addictive drug (85, 145). It binds to monoamine transporters and blocks the reuptake of catecholamines into the presynaptic nerve terminals, leading to increased level of dopamine and other monoamines in the synaptic cleft (9, 412). The effects of cocaine include constricted blood vessels, dilated pupils, and increased temperature, heart rate, and blood pressure.
Brown et al. (48) demonstrated that c-Fos expression can be induced in the rat amygdala following single intraperitoneal cocaine injection (10 mg/kg, see Table 5). The following studies have shown that c-Fos may be induced specifically in the lateral (88, 384), basolateral (88, 384, 418), and central amygdala (88, 191, 384, 418). Furthermore, more detailed analysis of the central nucleus indicated that all its subdivisions may be activated, namely, lateral, capsular, and medial (88, 418). The data on the medial amygdala are more controversial: some authors observed elevated c-Fos expression in this nucleus (384, 418), whereas others did not (88). The work of Day et al. (88) points out, however, that the context of drug administration (novel versus familiar) may be, in fact, critical for c-Fos expression in the medial amygdala, but not the drug on its own. Furthermore, the lack of c-Fos activation was reported for the posterior part of basomedial nucleus (384). It appears that there are no studies analyzing c-Fos expression in the cortical amygdaloid nucleus following acute cocaine treatment.
|
Apparently only a single report describes c-Fos expression in the amygdala following repeated cocaine treatment. Radwanska et al. (384) demonstrated that repeated intraperitoneal cocaine administration augments c-Fos expression in the central, lateral, basolateral, and posterior part of basomedial and medial amygdaloid nuclei compared with acutely treated animals.
There is also a group of studies analyzing c-Fos expression following drug self-administration. Kuzmin and Johansson (247) investigated C57Bl/6J mice that were trained to perform nose-poking reaction, which resulted in intravenous infusion of either cocaine or saline. The yoked group received passive infusions. In the yoked group, the infusions were noncontingent to the nose-poking. Self-administration session lasted 30 min, and the animals were decapitated 1 h later. In situ hybridization analysis showed that active cocaine administration during the first session increased c-fos mRNA levels in the lateral and basolateral, but not medial, cortical, or central, amygdala compared with passive cocaine injections. Interestingly, noncontingent cocaine administration (in the yoked group) decreased c-fos level in the lateral and basolateral nuclei compared with the saline controls (247). Contrasting results were obtained by Ciccocioppo et al. (71). In their hands, a single session of intravenous cocaine self-administration induced c-Fos protein expression only in the central, but not basolateral, amygdala. The observed discrepancies may be the result of different species used, mice in Kuzmin's and Johansson's study versus rats by Ciccocioppo et al., as well as due to different instrumental protocols: nose-poke versus bar-pressing. Additionally Ciccocioppo et al. (71) showed that a single challenge session following chronic cocaine self-administration (2 wk) induced c-Fos expression at the level observed in the saline-treated animals.
Brown et al. (48) analyzed c-Fos protein expression in the amygdala following exposure to cocaine-paired environment. In their experiment, the animals were injected with cocaine for 10 days prior to placement into activity cages, or with saline prior to placement into the home cage. The control animals were injected with cocaine in the home cages and with saline prior to placement into activity cages. Two days after the last cocaine injection, the animals were placed once again into the activity cages and killed 90 min later. No specific analysis of the amygdalar nuclei was carried out, but the area marked by the authors apparently covered only the central, basolateral, and lateral nuclei. The majority of the following studies showed, however, that exposure to cocaine-paired environment induced c-Fos protein expression particularly in the basolateral but not the central amygdala (101, 297, 316, 317). Other nuclei of the amygdala were apparently not considered. Contrasting results were obtained by Franklin and Druhan (125), who found no difference in c-Fos expression either in the basolateral or central amygdala in the rats exposed to cocaine-paired environment compared with the animals exposed to saline-paired context. The discrepancy between this and the previously mentioned reports may be due to too early time point following CPP used by Franklin and Druhan (125) for analysis of c-Fos protein expression (1 h vs. 1.5–2 h in the other reports).
It has been also shown that exposure to either cocaine self-administration environment (18, 341) or drug-discriminative cues (71), after long-term abstinence, induces c-Fos protein expression in the basolateral, but not central, medial, lateral, or cortical amygdaloid nuclei. This increase was not observed in the animals that went through extinction training, i.e., multiple exposures to drug self-administration environment without drug administration (341). Furthermore, Ciccocioppo et al. (71) have found that presentation of drug discriminative stimuli even after short-term abstinence period induces c-Fos expression in the basolateral nucleus.
In conclusion (see Table 5), it seems that acute, passive cocaine administration activates most consistently c-Fos within the central and basolateral amygdala. Furthermore, cocaine-induced c-Fos expression in the lateral and basolateral nuclei seems to be regulated by many factors including individual life history or a general arousal status. In contrast, the data on c-Fos expression following a single session of cocaine self-administration are inconsistent. Moreover, the data on c-Fos expression in the amygdala following repeated cocaine administration are scarce. On the other hand, it is relatively consistent that an exposure of an animal to cocaine-paired environment (after CPP) or to cocaine self-administration environment and cocaine-associated cues activates c-Fos expression in the basolateral nucleus. Lack of c-Fos protein expression in the central, medial, and cortical nuclei has been found following both exposure of an animal to cocaine-paired environment (after CPP) as well as to drug self-administration environment and drug-associated cues.
C. Cocaine and Other Gene Activity Markers
Pich et al. (371) found that after 2 wk of cocaine self-administration, the density of the nuclei detected with antibody against FRA is in the central and basolateral amygdala the same as after saline injections. On the other hand, acute cocaine injection was found to induce JunB expression in the lateral, basolateral, and central, but not in the anterior part of medial and posterior part of basomedial amygdaloid nuclei. The pattern of JunB expression was the same following chronic drug treatment (384). Furthermore, Radwanska et al. (384) have reported that acute, cocaine injection activates also Zif268 expression within the lateral, basolateral, central, medial, and posterior part of basomedial amygdala. Following chronic cocaine treatment, Zif268 expression was shown to be decreased in all analyzed nuclei and observed only in the central nuclei (384). In the studies by Mutschler et al. (334), rats were trained to self-administer cocaine until the animals showed stable responding on the lever rewarded with cocaine infusion during 3 consecutive days. Next, the animals were allowed continuous access to cocaine during a 16-h period ("binge"). The animals were killed either immediately or 1 or 14 days following the binge period. In situ hybridization for zif268 mRNA indicated that long-term, but not short-term, withdrawal led to a decreased zif268 level in the basolateral nucleus, compared with saline or passive cocaine-treated animals. No change in zif268 expression was observed either in the medial or central amygdala (334). Also, exposure to the cocaine self-administration environment and drug paired cues was shown to induce zif268 mRNA expression in the basolateral, but not lateral, amygdaloid nucleus (471). Zif268 mRNA expression in the central nucleus, although not reaching statistically significant difference between control and "paired" groups, was increased in the latter. Moreover, a single session of cocaine self-administration did not change ngfi-a/zif268 mRNA in any of the analyzed amygdala nuclei, namely, neither in the lateral nor basolateral, basomedial, medial, or central (247).
D. Amphetamines and c-Fos Expression
Amphetamine compounds [amphetamine, methamphetamine, and 3,4-methylenedioxymethamphetamine (MDMA)] cause an efflux of biogenic amines from neuronal synaptic terminals (indirect sympathomimetics) (437, 461, 516). They also inhibit specific transporters responsible for reuptake of biogenic amines from the synaptic nerve ending and presynaptic vesicles and inhibit monoamine oxidase, which degrades biogenic amine neurotransmitters intracellularly (9, 412). The net effect is an increase of neurotransmitter release into the synapse. Amphetamine increases heart rate, blood pressure, and metabolism.
Acute, passive amphetamine administration has consistently been shown to induce c-fos mRNA and protein expression in the central (63, 88, 90, 100, 309, 323, 350) and basolateral amygdala (88, 100, 309, 323, 350). More controversial is c-Fos in the medial nucleus; some authors indicate that amphetamine activates c-Fos expression in this part of the amygdala (323), while some indicate that it does not (88, 350). The discrepancies observed for the medial nucleus might have, however, arisen due to different doses of amphetamine used (4 vs. 0.5–2 mg/kg). There are virtually no data on c-Fos expression in the cortical amygdaloid nucleus. Interestingly, c-Fos expression was found to be still increased in the central and basolateral amygdala after long-term drug administration (1 injection weekly for 9 wk) (309).
In a series of papers, Day, Ostrander, and co-workers (88, 90, 350) analyzed the impact of different "processive" stressors (demanding the higher brain structures for perception) on amphetamine-induced c-fos expression in the amygdala. They have found that amphetamine-induced c-fos in the central nucleus is decreased by restraint stress, stressful noise (100–110 dB), and novel environment (88, 90, 350). At the same time, the amphetamine-induced c-fos mRNA expression level in the basolateral complex was found to be higher in a novel environment compared with a home cage. Interestingly, only a novel environment, but not amphetamine, induced c-fos expression in the medial nucleus (88, 90, 350). Furthermore, repeated amphetamine self-administration in the home environment did not change c-fos mRNA expression in the amygdala compared with acute treatment, while repeated amphetamine administration in a novel environment decreased c-fos expression in the central nucleus, but not basolateral complex, compared with single injection (88, 350). On the other hand, Carr and Kutchukhidze (63) have found that chronic stress of starvation potentiates amphetamine-induced c-Fos in the central amygdala (63). Furthermore, history of social defeat stress potentiates amphetamine-induced c-Fos expression in the anterior part of the medial nucleus, but not in the basolateral or central nuclei (344).
Mead et al. (309) trained mice to discriminate between amphetamine- and saline-paired environments in the test of conditioned place preference during 8 consecutive days. The next day the animals were reexposed to experimental environment. This procedure induced c-Fos expression in the basolateral and central amygdaloid nuclei.
There are only a few reports on the gene expression induced in the amygdala by amphetamines other than D-amphetamine. Acute methamphetamine injection has been shown to induce c-Fos protein expression in the central and medial (480) or in the central and basolateral nuclei (265, 266). In all these studies, other amygdalar nuclei were not analyzed. Single injection of MDMA was found to activate c-Fos in the central nucleus, compared with a single saline injection (459).
In aggregate (see Table 6), it appears that acute, passive amphetamine administration induces c-Fos expression predominantly in the central and basolateral amygdala. Moreover, the activation of the basolateral complex seems to be potentiated by an arousal, associated, e.g., with a novelty. On the other hand, c-Fos activation in the central amygdala is decreased by several "processive" stressors. The data on c-Fos expression in the medial nucleus although contradictory seem to indicate that high but not low doses of amphetamine activate this nucleus. Chronic, passive amphetamine treatment has been shown to activate c-Fos expression both in the basolateral and central nucleus. There are apparently no studies on c-Fos expression in the amygdala following amphetamine self-administration. Furthermore, exposure to amphetamine paired context was shown to increase c-Fos in the basolateral and central amygdala.
|
Acute amphetamine injection has been shown to elevate zif268 mRNA expression in the central (75, 90, 449), lateral (449), and medial (75), but not in the basolateral amygdala (75). Furthermore, Shilling et al. (449) analyzed zif268 mRNA expression in the amygdala following either escalating doses of amphetamine (from 1 to 8 mg/kg) administered 3 times daily for 4 days or "binge" injections for 9 days (8 mg/kg, 4 times daily). Both schedules of amphetamine administration induced zif268 mRNA expression in the lateral and central amygdaloid nucleus. This expression was however downregulated compared with the effects observed after the acute drug treatment (449). It has been also shown that amphetamine-induced zif268 mRNA expression may be downregulated in the central nucleus by restraint stress and stressful noise (100–110 dB) (90).
FosB expression was analyzed in the amygdala of the rats 4 days after chronic amphetamine treatment. The rats received during 6 days escalating (from 1 to 5 mg/kg, 3 injections daily) or intermitted (1.5 mg/kg, once daily) injections of amphetamine and were killed 4 days after the last injection. This treatment induced elevated FosB expression in the basolateral, but not medial, nucleus of the rats treated with escalating, but not intermitted, doses of amphetamine. FosB expression was also induced in the central nucleus, but this increase did not reach statistical significance (333).
F. Methylphenidate and c-Fos Expression
Methylphenidate is a central nervous system stimulant, commonly used to treat attention deficit-hyperactivity disorder (ADHD) (116). Methylphenidate, similarly as cocaine, blocks dopamine transporters (490). It has effects similar to, but more potent than, caffeine and less potent than amphetamines.
Trinh et al. (475) found that acute methylphenidate injection induced c-Fos protein expression in the basolateral, but not central, nucleus of the mice. Interestingly, dopamine transporter knockout mice show a decrease in drug-induced c-Fos expression in basolateral nucleus and an increase in the central nucleus (475). Papa et al. (356) found that a 24-h- but not 72-h-long withdrawal from chronic methylphenidate treatment induced c-Fos expression in the anterior part of basomedial nucleus and medial nucleus but not lateral, basolateral, or central amygdaloid nuclei.
G. Nicotine and c-Fos Expression
Nicotine acts on the nicotinic acetylcholine receptors. In small concentrations it increases the activity of these receptors that inter alia leads to an increased flow of epinephrine. The release of epinephrine causes an increase in heart rate, blood pressure, and respiration, as well as higher glucose levels in the blood. In high doses, nicotine will cause a depolarizing block of the nicotinic acetylcholine receptor, which is the reason for its toxicity (67, 83, 372).
Single, passive nicotine injection, not depending on a route of administration, has been found to induce c-Fos protein expression in the central (16, 295, 296, 354, 438), but not lateral or basolateral, amygdala (16, 482). Furthermore, more detailed analysis of the central nucleus showed that acute nicotine administration activated c-Fos expression in its capsular and lateral, but not medial, part (16, 482). Interestingly, it has been shown that nicotine injection did not induce c-Fos expression in the central amygdala, when the peripheral action of nicotine was prevented by treatment with hexamethonium (421). It appears that the medial and cortical nuclei of the amygdala have never been analyzed after nicotine treatment.
Interestingly, acute, subcutaneous nicotine challenge, which follows chronic nicotine perfusion, is still able to induce c-Fos protein expression in the central amygdala. The effect of subcutaneous treatment was also observed after short- (24 h) and long-term (72 h) withdrawal (422). Apparently, there are no data on c-Fos expression induced by nicotine self-administration.
c-Fos expression pattern in the amygdala in withdrawal of nicotine-dependent rats was analyzed with the use of the nicotinic receptor antagonist mecamylamine. Mecamylamine treatment of nicotine-dependent rats was found to induce c-Fos expression in the central, but not basolateral, nucleus (355, 478). On the other hand, no c-Fos induction in the central nucleus was observed in nicotine-dependent rats, neither after 24- nor 72-h-long nicotine abstinence, compared with control saline-treated animals (422).
In conclusion (see Table 7), passive administration of nicotine activates c-Fos expression in the central, but not lateral or basolateral, amygdaloid nuclei. It has also been shown that nicotine challenge following long-term nicotine treatment may induce c-Fos in the central amygdala. Moreover, nicotine withdrawal induced by mecamylamine, but not nicotine abstinence, induces c-Fos in the central amygdala.
|
It has been found that acute passive nicotine administration may decrease FRA proteins' level in the basolateral nucleus, without affecting FRA expression in the central amygdaloid nucleus (371). Furthermore, no difference in FRA protein level was observed either in the basolateral or central nuclei after 2 wk of nicotine compared with saline self-administration (371). Acute nicotine administration has been also shown to have no influence on Zif268 protein expression either in the lateral, basolateral, or central amygdala (16).
I. Opioids, Morphine Derivatives (Morphine, Heroin), and c-Fos Expression
Morphine is an opioid agonist which exerts its main effect by binding to the µ-type of the opioid receptors in the central nervous system that are believed to mediate analgesia, euphoria, physical dependence, and respiratory depression. Morphine also binds to 
- and 
-opioid receptors, which are thought to mediate spinal analgesia, meiosis, and sedation. The most commonly used animal model of morphine withdrawal is an acute treatment of morphine-dependent animals with µ-opioid receptor competitive antagonists, naloxone or naltrexone. These treatments have been demonstrated to give clear behavioral signs of withdrawal (143, 283, 434).
Acute morphine injection has been shown to induce c-Fos protein expression in the central, medial, and cortical amygdaloid nuclei of the guinea pig (43) and the central, basolateral, but not medial nuclei of the rat (454). c-Fos expression was observed neither within the central nor medial nucleus after chronic implantation of morphine palettes (138, 460). It has been, however, shown that chronic, intravenous morphine administration may induce c-Fos expression in the lateral and basolateral, but not central, nucleus of the amygdala, when the drug is injected in morphine- but not in saline-paired environment (322). Apparently, there are no data on c-Fos expression following morphine self-administration.
Hayward et al. (172) and Stornetta et al. (460) revealed that naltrexone-precipitated withdrawal in morphine-dependent rats may induce c-fos mRNA and protein expression in the amygdala. Careful analysis of the photographs included in their papers indicates that in fact only the central nucleus of the amygdala was activated. In the subsequent studies, activation of c-Fos expression, following naloxone- and naltrexone-precipitated withdrawal, has been found in the central (138, 148, 198, 200, 204, 339, 489), medial (138, 198, 200), but not basolateral amygdala (198, 200) in the rat and guinea pig. On the other hand, Jin et al. (199) have not found c-Fos in the medial nucleus following naloxone-precipitated withdrawal, and Frenois et al. (126, 127) showed dose-dependent decrease in c-Fos expression in the basolateral nucleus after naloxone-precipitated withdrawal.
Harris and Aston-Jones (15, 168) reported that exposure to morphine-associated place activates c-Fos expression in the central and basolateral amygdala. Furthermore, these effects were dependent on the longitude of "morphine history" and were potentiated by chronic morphine pretreatment, compared with a single drug injection. Moreover, c-Fos expression level in the basolateral, but not central, nucleus correlated with place preference (15, 168).
Heroin is a powerful opiate pain killer that produces euphoria and blissful apathy. It is known for leading to addiction and difficult physical withdrawal symptoms. Heroin has µ-, -, and -opioid receptor activity. Similarly, as in the case of morphine, acute heroin administration activated c-Fos expression in the central nucleus of the rat (452, 453).
In summary (see Table 8), both morphine and heroin activate c-Fos protein expression within the central nucleus of the amygdala. It is difficult to conclude about the activation pattern of c-Fos expression within the basolateral, medial, or cortical nuclei due to scarce and incongruous data. Moreover, c-Fos expression in the basolateral nucleus was found to be downregulated by the higher doses of the drug. The gene expression pattern may be different for different species as revealed by comparison of the rat and guinea pig. It is also not quite clear whether chronic morphine treatment may induce c-Fos expression in the amygdala. The data seem to indicate that there may be differences depending on the mode of morphine treatment: passive injection activates the basolateral nucleus, while chronic implantation of morphine pallets does not lead to c-Fos expression. In all analyzed studies both naltrexone- and naloxone-precipitated withdrawal induced c-Fos expression in the central but not the basolateral nucleus. On the other hand, exposure to morphine-associated place has been shown to activate c-Fos both in the central and basolateral amygdala.
|
9-Tetrahydrocannabinol (THC) is a psychoactive compound in herbal cannabis. It appears that it exerts all of its known central nervous system effects through the CB1 cannabinoid receptor (195). In humans, at low doses it tends to induce a sense of well-being and a dreamy state of relaxation, which may be accompanied by a more vivid sense of sight, smell, taste, and hearing, as well as by subtle alterations in thought formation and expression. High doses may result in image distortion, a loss of personal identity, problems with memory and learning, fantasies, and hallucinations.
Acute THC injection in the low doses (2.5–5 mg/kg) was found to induce c-Fos expression in the central (307, 361, 482) but not lateral, basolateral, or medial amygdala of the mouse (361, 482). On the other hand, THC administrated in a high dose (10 mg/kg) induced c-Fos expression in the basolateral and central, but not anterior, part of the medial nucleus of the rat (7). There are also reports on the influence of psychological factors and other drugs on THC-induced c-Fos expression in the amygdala. Restrain stress following THC administration potentiated c-Fos expression in the central but not basolateral or medial nuclei (362). Furthermore, THC-induced c-Fos expression was found to be potentiated both in the central nucleus and basolateral complex by coadministration of nicotine (482). In agreement with studies on THC, passive administration of cannabinoid receptor agonist HU-210 or CP 55,940 induced c-Fos protein expression in the central nucleus of the rats (14, 397). In addition, also CB1R antagonist SR141716 activated c-Fos expression in the central and basolateral nucleus (361). Moreover, administering SR141716A to rats chronically treated with cannabinoid drug HU-210 also increased c-Fos expression in the central nucleus (397).
In conclusion (see Table 9), THC as well as CB1R agonists and antagonists activate c-Fos expression in the central amygdaloid nucleus. Only the high doses of THC activate c-Fos expression in the basolateral nucleus. Furthermore, THC does not activate c-Fos expression in the medial part of the amygdala.
|
Lysergic acid diethylamide (LSD) produces altered perception, cognition, and mood as well as hallucinations in humans. There is a growing body of evidence in animals that these effects of hallucinogens are mediated by 5-HT2A receptors (1). Acute LSD injection was reported to induce c-Fos expression in the central amygdaloid nucleus of the rat (151).
L. Ethyl Alcohol and c-Fos Expression
Ethyl alcohol (ethanol) is one of the most common and strong psychoactive substances used by humans. It can specifically alter the function of several ligand-activated ion channels including N-methyl-D-aspartate (NMDA), serotonin (5-HT3), glycine, and GABAA receptors (2). Ethanol causes decreased motor functions and decreased consciousness level. At high concentrations, ethanol is an anesthetic and can produce autonomic dysfunction (e.g., hypothermia, hypotension), coma, and death from respiratory depression and cardiovascular collapse.
There are multiple studies which show that acute intraperitoneal or intragastric administration of ethanol increases c-Fos expression specifically in the central nucleus of the amygdala (64, 76, 175, 180, 226, 415, 417, 470). c-Fos protein level increases, in a dose-dependent manner, already 1 h after alcohol injection and may last up to 16 h (64). The study by Ryabinin and co-workers (16, 418) additionally showed that single alcohol injection induced c-Fos expression in all subdivisions of the central nucleus, i.e., capsular, lateral, and medial. Interestingly, acute alcohol injection stimulates c-Fos expression at the same level in the central nucleus of both alcohol-preferring (P and Alko-Alcohol) and nonpreferring rat lines (NP and Alko-Nonalcohol) (470). On the other hand, the data on c-Fos expression in the lateral, basolateral, and medial amygdala are not that clear. In two studies from Ryabinin's groups, the authors showed that intraperitoneal alcohol injection either activated (17) or had no influence (418) on c-Fos expression in the lateral and basolateral amygdala. The observed discrepancies might have arisen, however, due to different times of habituation: 4 days (418) versus 13 days (16). Also, Knapp et al. (226) observed c-Fos expression neither in the lateral nor in the basolateral amygdala. The data on the medial amygdala are again not contradictory: one study showed that c-Fos might be activated there (418); the other did not find any c-Fos induction, either after intraperitoneal or intragastric alcohol infusion (226). Again, the reasons for these discrepancies are unclear. Finally, there is only one study which showed that acute alcohol injection had no influence on c-Fos expression in the cortical amygdala (226).
Moreover, chronic, passive alcohol injection has been shown to reduce c-Fos expression in the central nucleus, compared with acute administration, if measured 2 h following the last of the chronic injections (64).
c-Fos protein expression pattern in the amygdala was also analyzed following acute and chronic alcohol drinking. Drinking for the first time the highly concentrated ethanol (10%), but not low concentrated (2%), induced c-Fos expression in the lateral, but not medial, part of the central nucleus of the amygdala. Drinking 10% alcohol has been also shown to have no influence on c-Fos expression either in the lateral or basolateral nuclei, compared with water-drinking animals. Surprisingly, drinking sucrose induced a pattern of c-Fos activation similar to alcohol. This pattern was different from the one observed in water-drinking controls (414).
Bachtell et al. (17) have reported that repeated drinking of alcohol induced c-Fos expression in the posterioventral portion of the medial part of the central nucleus of the amygdala, but not in the rostral portion of the medial part or lateral and capsular parts of the central nucleus, as well as lateral and basolateral nuclei. On the other hand, the same group (414) described c-Fos activation in the lateral part of the central nucleus, compared with water- and sucrose-drinking animals. Moreover, in another study of Ryabinin et al. (416), in the same model of alcohol administration, no c-Fos activation was observed in any of the analyzed nuclei of the amygdala (basolateral, lateral, and medial part of the central) following repeated alcohol drinking, compared with water- or sucrose-drinking animals. The only described difference in procedures was that the animals received alcohol either during the dark (416) or light phase of a day (414). Additionally, Sharpe et al. (442) showed that chronic alcohol drinking led to the decreased c-Fos protein expression level in the basolateral nucleus, compared with water-drinking animals. They also did not observe any difference in the central nucleus between water and chronic alcohol drinking animals.
c-Fos expression in the amygdala, induced by alcohol withdrawal, was mostly analyzed following discontinuation of alcohol containing diet. The first abstinence period imposed on the animals fed previously an alcohol-containing diet was demonstrated to have no influence on c-Fos protein expression either in the central (42, 330) or basolateral (42) or medial amygdala (330). In contrast, a second abstinence period activated c-Fos expression in the central and basolateral, but not lateral, nucleus (42).
As far as the reinstatement to alcohol drinking is concerned, we have found that exposure of alcohol-dependent animals to alcohol self-administration context after a 1-mo abstinence does not induce c-Fos expression in any of the analyzed nuclei of the amygdala (the ventral and dorsal tips of the lateral nucleus, basolateral, central, or medial nucleus). On the other hand, exposure to alcohol self-administration context with alcohol-paired discrete cues activates c-Fos expression in the ventral tip of the lateral nucleus, basolateral and central nuclei, but not in the dorsal tip of the lateral nucleus or medial nucleus (K. Radwanska, unpublished observations).
As far as other molecular markers are concerned, acute alcohol injection has been demonstrated to induce Zif268 expression in the central but not lateral or basolateral amygdala (16). Furthermore, after chronic drinking of alcohol, there were no differences in Zif268 expression either in the central or lateral or basolateral nuclei, compared with sucrose-drinking animals (17). Zif268 protein expression has been also found to be elevated only in the central, but not basolateral or lateral, nuclei of the animals which experienced the first, but not the second, abstinence period after chronic alcohol consumption (42).
In conclusion (see Table 10), acute, intraperitoneal alcohol administration has been repeatedly shown to activate c-Fos expression in the central nucleus of the amygdala. The data on the basolateral, medial, and cortical nuclei are contradictory and limited; thus drawing any conclusions is difficult. Furthermore, it has been demonstrated that chronic passive alcohol administration does not upregulate c-Fos expression in the central nucleus. In contrast to either intraperitoneal or intragastric routes of ethyl alcohol administration, it seems that acute alcohol drinking does not activate c-Fos expression either in the central or basolateral nucleus. The data on c-Fos expression pattern following chronic alcohol drinking are again contradictory. Finally, the data on alcohol withdrawal are too scarce to draw any ultimate conclusions, but it seems that not the first abstinence period but rather the following ones are able to induce c-Fos expression in the central and basolateral, but not lateral amygdala.
|
|
V. DIFFERENT BEHAVIORAL TASKS INVOLVE SPECIFIC SUBDIVISIONS OF THE AMYGDALA |
|---|
|
TopPreviousNextReferences |
|---|
Before addressing the specific pattern of c-Fos expression within the amygdala in the context of behavior, we would like to consider again the biological meaning of c-Fos. In an undisturbed brain or in response to well-habituated behaviors, there is virtually no c-Fos in neurons (see Ref. 205). This is true even despite intense ongoing neuronal activity. Thus c-Fos expression is in a striking contrast to, e.g., Zif268, whose expression is observed in active brain regions of naive animals, and whose expression levels appear to be tightly linked to ongoing neuronal activity, at least in certain brain regions (including amygdala; see also Fig. 9 and Ref. 227). Hence, c-Fos is a very peculiar "activity" marker that differs from other labels, including metabolic activity indicators. c-Fos expression, however, can be easily evoked by any novel information processed (and then remembered) by the brain (see Refs. 206, 207, 210). This expression occurs throughout the various structures of the brain, however, as clearly seen from the pattern of expression extensively described above; even within the amygdala c-Fos appears to label neurons in a rather specific manner. Thus such widespread, but nonetheless specific, c-Fos expression may imply that the novel information is widely elaborated in various brain areas, however, always in a behavioral context-specific fashion.
One shall also never forget that the only molecular function of this protein, we are aware of, is to control gene expression. Two major hypotheses have been put forward as far as the biological function of this protein is concerned (206–208): 1) involvement in molecular replenishment and 2) control of neuronal plasticity. The latter suggests that c-Fos is a transcriptional regulator of genes encoding proteins subserving the plasticity (e.g., acting at the synapse and involved in synaptic strength). The former relates to c-Fos as being a transcriptional factor for genes encoding proteins that are actively used and thus "worn-out" during periods of intense neuronal activity. One can easily imagine that elaboration of novel information evokes a sudden burst of neuronal activity that results in depletion of a specific pool of proteins, such as neurotransmitter-producing enzymes, receptors, proteins that are released from activated neurons, etc. These proteins should be then replenished, and this process in turn requires the gene expression. c-Fos might be involved in a transcriptional control of those genes (see, e.g., Refs. 197, 209).
Considering the above, we suggest that the neurons displaying increased c-Fos expression have been involved in processing of novel information of specific value, and furthermore, probably, are susceptible to plastic changes. Thus we predict that 1) connections of those neurons should allow for specific information to be delivered to them, 2) the lesions or functional inactivation carried out after the behavioral exposure to a novel condition should affect the performance of the acquired reaction (the lesions as well as activation/inactivation attempts carried out before either the training or drug exposure are more difficult to interpret because of possible confounding effects of compensatory mechanisms), and 3) those neurons should display propensity for plastic changes. With these considerations in mind, in the following sections we synthesize the data on c-Fos expression pattern in each major subdivision of the amygdala with its connectivity as well as information regarding the effects of specific lesions and pharmacological approaches to either activate or inactivate neurons as well as existing knowledge on neuronal/synaptic plasticity. We first summarize the c-Fos expression data with an aim to produce a hypothesis about possible function(s) of the specific amygdalar subdivision, and next we confront this hypothesis with results of studies using other approaches. In the introductory remarks to this review, we have presented current understanding of internal and external organization of the amygdala (see Figs. 1–4). Notably, it is clear from the aforementioned description of the gene activity markers in the amygdala expressed following a variety of behavioral treatments that the data available force us to use in the following considerations a simplified picture of the amygdalar heterogeneity (Fig. 11).
|
Activation of the basolateral part of the amygdala (the lateral and basal nuclei), as indexed by the gene activity markers, was seen as a result of acquisition of a variety of behaviors reinforced with either aversive or appetitive stimuli. The list includes active avoidance learning, fear conditioning with an auditory cue, place preference and place avoidance learning, bar-pressing response reinforced with food, odor discrimination learning, and conditioned emotional response. Furthermore, the basolateral amygdala was also responding with a marked increase in c-Fos expression following exposure to sexually conditioned odor, exposure to associated volatiles, ultrasound-inducing defense behavior, feared conspecific's presence, avoiding the familiar winner, exposure to alarm pheromones, and memory retrieval. This second list is of particular importance, as it shows that increased gene expression in this brain region is not a mere result of sensory perception and nociception, in particular. As a matter of fact, even following the first training session of the active avoidance conditioning, c-Fos expression in the amygdala did not correlate with the sum of shock the animal received (428). We would like to summarize these data as being consistent with the suggested role of the basolateral part of the amygdala in formation of current stimulus-value associations (see also Refs. 61, 104, 419).
In several other experimental paradigms, the increased expression of c-Fos in the basolateral amygdala was not observed. These studies include exposure to natural rewards or treats (drinking of taste solutions, eating palatable food), exposure to estrous female, and predatory, e.g., cat odor, subordinate behavior in an aggressive encounter. Notably, the value of these stimuli is generally the same and does not need updating.
There is also a set of experimental tasks in which activation of c-Fos expression is ambiguous, either due to scarce or contradictory data. Among those procedures we can mention conditioning of taste aversion, contextual fear conditioning, elevated-plus maze, predatory odor (red fox, TMT), dominant aggressive behavior (attacking), mating, exposure to estrous female, and exposure to chemosensory cues from sexually mature partner.
As far as the drugs of abuse are concerned, the basolateral amygdala has been consistently found to display the elevated c-Fos expression following acute administration of cocaine, amphetamine (both drugs involving enhanced dopamine function and improving learning), as well as THC (high dose). On the other hand, neither nicotine nor ethyl alcohol augments c-Fos expression in this nucleus. Data on morphine effects are inconsistent. The basolateral amygdala was also activated in terms of c-Fos expression after a series of cocaine, amphetamine, or morphine injections, which appears very surprising in the context of the well-known phenomenon of diminishing c-Fos activation after repetitions of the behavioral treatments. Most notably, an exposure to cocaine-, amphetamine-, morphine-, or alcohol-paired environment or cocaine- and alcohol-paired discrete cues also induced c-Fos protein expression in the basolateral amygdala, thus showing that the behaviorally relevant stimuli, as in the case of learning (see above), and not the drugs themselves are important in eliciting the gene activity response. Such a gene expression pattern points again to a possible role of the basolateral amygdala in evaluating and updating of the current stimulus-value associations.
It should be stressed that the aforementioned explanation was first inferred from studies using the lesion method, which we briefly review below. The fact that the pattern of the gene expression in the basolateral amygdala correlates well with the results obtained with other approaches should be therefore viewed as an additional argument in favor of this hypothesis.
It was observed that the neurotoxic lesions of the basolateral nuclei neither caused motivational deficits nor changed the locomotor activity (52). Furthermore, it was shown that rats with the basolateral lesions were able to acquire conditioned instrumental responses (21, 39, 50, 52, 170, 219). However, the responding did not have the flexibility of that seen in unoperated animals, being insensitive to postconditioning changes in the value of the reinforcer. For instance, in the reinforcer devaluation procedure reducing the value of the food reward by pairing its ingestion with malaise, subsequent responding to the stimulus that had been previously paired with that food was unchanged in the basolateral amygdalalesioned rats, in contrast to the intact and the central amygdala-lesioned animals that were able to spontaneously adjust their responses to the current value of the reinforcer (170). Moreover, both monkeys and rats with neurotoxic lesions of the basolateral amygdala, in which the selective satiation procedure was applied to devalue the reinforcer, were insensitive to changes in the value of the reward (21, 287). Similarly, in second-order Pavlovian conditioning, in which the CS was paired with a stimulus that provided acquired rather than intrinsic motivational value, Everitt et al. (103) observed a decrease in instrumental responses maintained by a visual conditioned reinforcer previously paired with the sexual reinforcement in the basolateral amygdala-lesioned male rats. Furthermore, the basolateral amygdala-lesioned animals were shown to exhibit a significant, selective reduction in responding on the lever providing a conditioned reinforcer (a stimulus previously paired with water, food, or intravenous infusion of cocaine, Refs. 52, 170, 358, 511). Another phenomenon that seems to rely on current stimulus-value associations and is impaired by the basolateral lesions is the potentiation of feeding by the CS in rats that are already sated (186, 189). On the other hand, the basolateral amygdala-lesioned rats were not impaired in the general form of Pavlovian-to-instrumental transfer (PIT; Refs. 160, 186) in which an appetitive Pavlovian CS enhanced instrumental responding; however, Blundell et al. (39) have recently demonstrated an important role of the basolateral amygdala in the specificity with which conditioned stimuli influence instrumental responding (reinforcer-selective PIT).
Moreover, Balleine et al. (21) have shown that the basolateral amygdala-lesioned rats were able to discriminate the performance of lever pressing from chain pulling during acquisition and when they performed these actions in series to gain access to reward, but they were unable to learn to discriminate rewarded from unrewarded actions with the use of specific properties of the instrumental outcomes (pellet and maltodextrin) in a free operant discrimination situation. Thus the rats produced a clear deficit in sensitivity of instrumental performance to posttraining changes in the incentive value of the instrumental outcome, pointing to a deficit in encoding the motivational significance associated with sensory features of instrumental outcomes.
These results suggest that the basolateral amygdala is involved in particular in the representation of the sensory aspects of motivationally significant events. This is in line with the earlier results showing the involvement of the basolateral amygdala in utilizing physical and ethological properties of stimuli employed in learning (507). The authors developed the experimental design, which allowed delineating the modality effects on learning and performance of the two-way active avoidance response in rats. They examined postlesion acquisition of two-way avoidance and subsequent transfer to other warning signal of different modality. In one group the animals were originally trained with less salient CS (darkness), then transferred to more salient compound (darkness and white noise), and finally to white noise CS. In the second group, the opposite arrangement of CS during the subsequent stages of the training was used. They observed that in control animals transfer to more salient CS enhanced avoidance performance, whereas change to less salient CS decreased it. In contrast, the basolateral amygdala-lesioned animals were almost totally insensitive to any CS modality changes. Therefore, it may be concluded that the basolateral amygdala-lesioned rats were unable to utilize the saliency or the emotional value of the warning signal.
All above-mentioned studies clearly showed that the basolateral complex of the amygdala is involved in making associations between a stimulus and its current motivational value. This hypothesis is also supported by the results of electrophysiological and imaging studies (332, 345, 346, 474, 481). For example, recently, Paton et al. (363) conditioned monkeys to associate abstract images with either positive or negative value and then reversed acquired image value assignments. Recording activity of individual amygdala neurons during such reversal, they showed that changes in the values of images modulated neural activity and that this modulation occurred rapidly enough to account for monkeys' learning.
In addition to the lesion data, the pharmacological studies, in which the basolateral amygdala was blocked or stimulated, showed its involvement in acquisition of various behavioral tasks, memory consolidation/modulation, as well as in extinction of learned associations. For instance, in almost all tested learning tasks, namely, contextual, cued, and olfactory Pavlovian fear conditioning, conditioned fear-potentiated startle, second-order fear conditioning, discriminated approach response to food, inhibitory avoidance and two-way active avoidance training, acquisition was impaired by infusion of NMDA receptor antagonist into the basolateral amygdala (51, 108, 139, 147, 220, 273, 321, 400, 429, 492, 493). However, some results suggested that the basolateral amygdala might not be involved in contextual fear conditioning (493; see also Ref. 484). Moreover, efficiency of active avoidance task learning was impaired following stimulation of the basolateral amygdala with glutamic acid, as well as blocking histamine receptors within this nucleus (8). Similarly, instrumental learning of lever pressing for food reinforcement was shown to require dopamine receptor activation within the basolateral amygdala (12). Moreover, inactivation of the basolateral amygdala with GABAA receptor agonist impaired acquisition of instrumental avoidance learning, as well as taste-potentiated odor aversion (94, 112, 377).
Blocking or stimulation of the basolateral amygdala immediately after the training was also shown to influence strongly memory consolidation and/or modulation. For example, blocking NMDA receptors within the basolateral amygdala disrupted consolidation of inhibitory avoidance response (399, 411). Similarly, infusion of other, non-NMDA receptors antagonists impaired memory consolidation. For instance, infusion of AMPA and metabotropic receptors antagonists after the training of inhibitory avoidance task produced retrograde amnesia (36, 40). Moreover, Hitchcott and Phillips (179) showed that postsession infusion of dopamine receptor agonist into the basolateral amygdala abolished ability of the reward-associated stimulus to support acquisition of a novel instrumental response in the second-order conditioning. Inactivation of the basolateral amygdala with the GABAA receptor agonist also caused retrograde amnesia for instrumental avoidance response (196, 377, 411). Furthermore, it was frequently demonstrated that the basolateral amygdala is involved in modulation of memory consolidation (for review, see Ref. 305). For instance, administration of 
-adrenergic and muscarinic receptor agonists or benzodiazepine receptor antagonists after training of inhibitory avoidance task was shown to enhance retention performance (82, 196, 250). Such enhancement required activation of dopamine receptors within the basolateral amygdala (250–252). It was also observed that histamine receptor agonist infused into the basolateral amygdala after the training enhanced subsequent expression of fear memory (19), whereas histamine receptor antagonist caused impairment of memory consolidation of contextual fear conditioning (360). Involvement of the basolateral amygdala in memory modulation was also shown in the conditioned taste aversion paradigm, in which blockade of noradrenergic receptors during learning affected taste memory formation (320). Furthermore, norepinephrine, as well as the GABAergic antagonist bicuculline, when infused into the basolateral amygdala, caused clear enhancement of extinction of conditioned fear (33).
The basolateral amygdala was also shown to be involved in extinction of learned associations. For instance, extinction of conditioned fear to tone, light, and contextual stimuli, as well as extinction of fear-potentiated startle response engaged NMDA receptors within the basolateral amygdala (107, 262). Similarly, infusion of GABAA receptor agonist into the basolateral amygdala following an extinction session of auditory conditioned fear facilitated extinction (3).
Thus these results collectively provide further support for the hypothesis that the basolateral amygdala is an important part of the neural system subserving stimulus-reinforcer associations. Furthermore, it was shown that infusion of GABAA receptor agonist during selective satiation blocked reward devaluation. However, when the basolateral amygdala was inactivated after satiation, the shift in object preference was not observed (505). This result seems to be consistent with the hypothesis that the basolateral amygdala is necessary for the appropriate registration of the change in the reinforcer value.
Similar conclusions may be derived from the research on the basolateral amygdala involvement in the processes of drug addiction. Multiple lesion studies demonstrate the contribution of this region in the association and updating of drug-rewarding properties with drug administration context and related cues, and thus formation and updating of secondary rewarding properties of conditioned contextual and discrete stimuli. Tetrodotoxin (TTX)-induced inactivation of the basolateral amygdala, before conditioning session, has been shown to impair formation of Pavlovian cocaine-cue associations as measured during cue-induced reinstatement to drug seeking (246). Furthermore, TTX inactivation of the basolateral amygdala before the reinstatement session has been demonstrated to block cue- or context-induced reinstatement (130, 152, 246). Also, lidocaine and quinolinic acid lesions of the basolateral amygdala impair cue-induced relapse (214, 310, 511, 525). All these lesions have, however, no effect on cocaine self-administration and thus cocaine primary rewarding properties (129, 130, 214, 310, 436, 511). Moreover, TTX infusion into the basolateral amygdala has also no effect on priming-injection induced reinstatement (130).
Furthermore, the pharmacological studies in the addiction indicate that manipulations with different neurotransmitter systems within the basolateral amygdala affect the evaluation and updating of the secondary rewarding/aversive properties of conditioned stimuli. Microdialysis studies have shown that almost all drugs of abuse (including amphetamine, cocaine, nicotine, and alcohol) (158, 443, 515, 523, 524) as well as exposure to drug-paired environment (503) induced dopamine (DA) release in the basolateral amygdala. These observations were also confirmed by immunocytochemical detection of DA in the basolateral amygdala following amphetamine administration (369, 370). Importantly, infusion of D1R antagonist (SCH23390) during the conditioning session when animals received passive cocaine injection paired with compound stimulus blocked the formation of cocaine-stimulus association, as measured during CS-induced reinstatement (32). Additionally, raclopride, a D2R antagonist, infused at low doses potentiated, and at high doses inhibited, cue-induced reinstatement (32). Moreover, it was shown that while infusion of dopamine receptor antagonists (SCH-23390, 
-fluopenthixol) into the basolateral amygdala before the reinstatement session significantly reduced responding for cocaine-conditioned reward, it did not affect cocaine self-administration (95, 436). On the other hand, infusion of D-amphetamine into the basolateral amygdala before reinstatement session potentiated cue-induced reinstatement (257). Furthermore, NMDA infusion into the basolateral amygdala has been demonstrated to induce reinstatement to cocaine self-administration in a self-administration environment (171). On the other hand, infusion of AMPA/kainate receptor antagonists (AP-5, CNQX) into the basolateral amygdala had no effect on responding for cocaine-conditioned reward, and did not affect cocaine self-administration (95, 436). It was also demonstrated that GABAergic and cholinergic neurotransmission within the basolateral amygdala are involved in processing of the secondary rewarding properties of drug-associated context. Inactivation of the basolateral amygdala with GABA receptor antagonists was shown to block expression of conditioned place preference (393), and intrabasolateral amygdala administration of the GABAA receptor agonist muscimol decreased acquisition of the conditioned place preference induced by morphine. On the other hand, intrabasolateral amygdala injection of the GABAA receptor antagonist bicuculline in combination with an ineffective dose of morphine elicited a significant conditioned place preference (526). It was also shown that microinjections of cholinergic antagonists in the vicinity of the basolateral amygdala produced a marked reduction of amphetamine-induced hyperactivity (144). The muscarinic receptor antagonist atropine, when injected into the basolateral amygdala, blocked morphine-induced place preference. At the same time, the nicotinic receptor antagonist mecamylamine potentiated morphine-induced place preference (527).
The involvement of the basolateral amygdala in memory storage still remains a matter of debate. It was shown that lesions made after conditioning produced profound deficits in fear potentiated startle response, as well as contextual and tone freezing. Moreover, the temporal gradient of this retrograde amnesia was not observed (133, 221, 267, 289). The memory was tested up to 16 mo, which time spans the adult lifetime of rats (133). However, the data on functional inactivation, as well as receptor blocking within the basolateral amygdala, immediately after conditioning are equivocal. For instance, the engagement of NMDA receptors within the basolateral amygdala was observed in different conditioned fear responses, namely, freezing, expression of fear potentiated startle, 22-kHz ultrasonic vocalization, defecation, and analgesia (109, 263, 290). In contrast, the others did not observe impaired expression of conditioned fear-potentiated startle (57, 321), as well as conditioned contextual and cued fear (396) after infusion of NMDA receptor antagonist into the basolateral amygdala. Similarly, retrieval tests of inhibitory avoidance task and discriminated approach response to food were not disrupted after infusing NMDA receptor blocker into the basolateral amygdala (24, 51). Moreover, infusion of GABAA receptor agonist before testing of the well-established avoidance response failed to disrupt its performance (377). In contrast, infusion of GABAA receptor agonist attenuated expression of conditioned fear (174). It was also observed that AMPA receptor antagonist infused before testing hindered the retrieval of inhibitory avoidance task (24). However, Walker et al. (493) observed that pretest infusion of AMPA receptor antagonist disrupted fear-potentiated startle to odor but not to context CS. On the other hand, retrieval of inhibitory avoidance was enhanced by infusing of norepinephrine into the basolateral amygdala and impaired by -adrenergic antagonist (25). In contrast, it was shown that performance of well-established instrumental response of lever pressing for food reinforcement did not require dopamine receptor activation within the basolateral amygdala (12).
Summing up, it appears that the basolateral amygdala plays a role in encoding and storing some emotional aspects of the conditioned fear. However, the existing data clearly point to contribution of other brain regions to storage of memory of certain other features of the conditioning situation (124). There are some indications that it may be appropriate to consider the basolateral amygdala as a system of functionally different subunits. Addressing this issue requires high-resolution imaging, which cannot be achieved by lesions, but lies within the range of applicability of gene expression mapping and electrophysiological methods. There are examples of both kinds of studies that consistently support the hypothesis of the heterogeneity of the basolateral amygdala.
The basolateral amygdala has also been shown as susceptible to neuronal plasticity both in vivo and in vitro (see Ref. 419 for review). The electrophysiological study of Repa et al. (391) reported very fine functional dissection of the lateral amygdala, showing that its dorsal part (LaD) consists of two neuronal populations with different profiles of electrophysiological responses evoked by fear-conditioning procedure. The cells in the dorsal part of LaD (LaD/d) showed short-latency responses to CS and were only transiently activated. The neurons in the ventral tip (LaD/v), on the other hand, exhibited longer latency responses, and their activation was more persistent. The authors concluded that these two populations of cells might be differently involved in either initiation or storage of long-lasting memories, respectively. This result has been excellently matched by c-Fos labeling study, in which Radwanska et al. (385) showed that aversive learning induced c-Fos expression specifically in the ventral but not dorsal tip of LaD.
Connectivity of the basolateral amygdala also makes it very well suited to be involved in integration of the sensory stimuli with their emotional values and then to transfer this integrated information within amygdala, as well as to other brain structures.
In particular, the lateral nucleus is undeniably a major recipient of sensory information, and it is often considered as the main gateway to the rest of the amygdaloid nuclei (10). Its outputs, which are apparently less widespread as the inputs, go mainly to the frontal cortex, the agranular insular, and the polymodal sensory association areas of the perirhinal and entorhinal cortices (see Refs. 299, 373 for review; Fig. 3). The lateral nucleus receives substantial cortical afferents from the various modality-specific cortical pathways that provide activation by integrated sensory information of each modality (402). It also receives projections from the gustatory and visceral primary areas located in the anterior and posterior insular cortices (299, 304, 373, 445). A prominent subcortical sensory input to the lateral nucleus comes from the posterior thalamic nuclei, including the auditory relay medial geniculate and taste relay ventral posterior nucleus (260, 353, 477, 521). The dorsolateral part of the lateral nucleus is the main target of cortically processed information from all the nonolfactory sensory modalities (299). It receives auditory and visual inputs coming from the unimodal and polymodal association areas of the temporal cortex (299, 403, 445). Moreover, the dorsolateral subdivision of the lateral nucleus is the target of somatosensory information from the visceral dysgranular parietal insular cortex in the parietal lobe (299, 446). Moreover, all of the cortical regions that project to the dorsolateral part of the lateral nucleus have also a light projection to the ventral part of this nucleus (299). The ventral, ventrolateral, and especially ventromedial subdivisions of the lateral nucleus receive also highly integrated visual information via either lemniscal (originating in the lateral geniculate) or nonlemniscal (originating in the posterior thalamic nuclei) thalamo-cortico-amygdalar pathways (444, 447). The ventromedial part of this nucleus receives light to moderate projections from the olfactory/visceral/somatosensory regions of the temporal cortex and substantial innervation from the prefrontal cortex (299). Furthermore, nociceptive information from the thalamic posterior intralaminar nucleus targets all subdivisions of the lateral nucleus (34, 41).
Inputs to the basal nucleus from the sensory-related cortical areas are less widespread than those to the lateral nucleus (see Fig. 3). The basal nucleus receives heavy or moderate projections from many gustatory, visceral, and somatosensory cortical areas, including the dysgranular and agranular insular, as well as the parietal rhinal cortex. Moreover, the prelimbic, infralimbic, medial, precentral, and dorsal anterior cingulated cortices, as well as the primary olfactory, perirhinal, entorhinal, and occipital cortices innervate the basal nucleus (46, 293, 299, 302, 304, 352). Furthermore, the ventrolateral portion of the basal nucleus receives substantial pain information from the parietal rhinal cortex, and the anteroventral part of the basal nucleus seems to be the target of nociceptive inputs coming from the posterior internuclear nucleus of the thalamus (299, 446). Generally, it is supposed that the cortical processing of nociceptive inputs provides information about physical and temporal features of pain (such as quality, intensity, duration, and location of painful stimuli), whereas subcortical regions that are involved in pain transmission are involved more in motivational/affective aspects of nociceptive information (514). Substantial subcortical innervation of the basal nucleus arises from the auditory relay medial geniculate body. Apparently weaker brain stem afferents originate in the ventral tegmental area and locus coeruleus. Several electrophysiological and anatomical studies indicated that also the parabrachial projections to the basal nucleus transmit nociceptive information (486). On the other hand, the basal nucleus is the source of the most numerous cortical projections from the amygdala. It projects heavily to the sensory association regions of the temporal cortex, including the posterior agranular insular, perirhinal, and entorhinal cortices (92, 303, 352, 373, 446, 485).
In aggregate, the connectivity of the basolateral group of amygdalar nuclei appears to be well suited to support the proposed role of this brain region in the evaluation of the motivational value of the stimulus, forming, and modulation of CS-US associations according to several subtle features of the conditioned stimulus, as well as in control of subject's attention to modality, saliency, and temporal attributes of the external cues. In conclusion, the basolateral amygdala appears to play a role in a plasticity of the current stimulus-value associations, by means of their evaluating and updating.
Induction of the expression of c-Fos in the central amygdala was observed following fear memory retrieval, conditioned taste aversion learning and retrieval, lithium chloride injection, consumption of palatable food or sweetened solutions, Pavlovian appetitive conditioning task, training of nose-poking for water reward, and training of bar-pressing response reinforced with food. Thus the increased level of gene expression was seen in the central amygdala following almost all appetitively motivated standard task, in which the incorrect responses were not punished. In some behavioral paradigms, activation of the central amygdala was seen inconsistently, e.g., following conditioned emotional response training, exposure to predatory odor, aggressive encounter, or mating. In contrast, the central nucleus of the amygdala has almost never been activated after fear conditioning, two-way avoidance, and exposure to the elevated mazes. Unfortunately, very little is known on activation of the central amygdala following appetitively motivated ethologically more relevant tests, such as mating, social interaction, or maternal behavior. However, relatively scarce data suggest that this kind of stimulation is also able to induce c-Fos expression in the central amygdala.
As mentioned above, most of the studies investigating the gene expression in conditioned fear learning showed the lack of involvement of the central amygdala. However, there are a few studies in which the increased level of c-Fos expression in the central amygdala was observed after such training. However, it cannot be excluded that in those studies, the animals had the opportunity to communicate (one or more animals per cage and the animals kept in the same or different room). This is of critical importance as, according to our recent results (see Ref. 228), the observed c-Fos expression in the central amygdala could be attributed to the aroused cagemates rather than to the fear conditioning itself. Furthermore, it has been recently shown that the appetitively motivated behaviors induced the augmented c-Fos expression only in the medial part of the central amygdala (229, 264). On the other hand, activation of the lateral part of the central amygdala was observed after introduction of a novel stimulus that had no signaling meaning (229).
As far as drug addiction is concerned, most notably, all analyzed drugs of abuse upon acute, forced treatment were shown to activate c-Fos expression in the central amygdala. Moreover, the detailed analyses indicate that drugs activate in particular the lateral, but not the medial part of this nucleus. Interestingly, these data relate to psychostimulants (cocaine or amphetamine), sedatives (alcohol), and hallucinogens (LSD). Furthermore, also passive, chronic cocaine and amphetamine administrations have been shown to activate c-Fos expression in the central amygdala. Additionally, not only acute drug administration but also acute withdrawal induced by specific antagonists activated c-Fos expression in the central nucleus. This time, however, c-Fos expression was observed both in the lateral and medial parts of the central nucleus. Again, the data seem to be consistent as far as all analyzed drugs are concerned (morphine, cannabinoids, and nicotine). Moreover, exposures to amphetamine-, morphine-, or alcohol paired-context and cues have all been shown to activate the central amygdala, as far as c-Fos expression is concerned.
On the other hand, long-term cocaine self-administration and implantation of morphine pallets did not activate this nucleus in terms of gene expression. Furthermore, contrary to pharmacologically induced withdrawal, it seems that simple cessation of drug administration does not induce gene expression in the central amygdala. Notably, all the studies that analyzed amygdalar c-Fos expression after exposure to cocaine (but not other drugs) paired-context/cues found no expression of this protein in the central amygdala. The data on chronic alcohol drinking are inconsistent.
Thus, in summary, it seems that the medial part of the central amygdala is activated mainly by such conditions as food consumption, mating, or caring for pups, and thus involves activation of the parasympathetic nervous system. Moreover, c-Fos expression in the medial part of the central amygdala may be activated by multiple hormones and modulator neurotransmitters. The same region is also activated by such acute drug withdrawal that is known to regulate potently the activity of the parasympathetic nervous system. Thus the medial part of the central amygdala may be involved in integrating information from external (e.g., food) and internal worlds (e.g., hunger/satiety). This integration allows for promoting adequate physiological and behavioral responses.
On the other hand, the gene activity markers are labeling the lateral part of the central amygdala in such experimental paradigms as conditioned fear, a novel stimulus without defined meaning, anxiolytic and anxiogenic drugs, as well as drugs of abuse or withdrawal of drugs of abuse. Thus it can be hypothesized that the lateral part of the central amygdala is activated by stimuli which modulate attention and vigilance, the conjecture which is in agreement with the hypothesis of Holland and Gallagher, who implicated this nucleus in regulation of attention processes (134, 187).
Notably, c-Fos protein expression in the central nucleus of the amygdala is not activated when an animal is exposed to such aversive conditions as foot shocks, open arena, novel environment, restraint, loud noise, or air puffs. All of those have in common that they immediately activate simple, reflex-driven behaviors of either freezing or flight. Simplistically, one could assume that those conditions do not induce gene expression; however, one may alternatively consider that they may cause an active inhibition of the central nucleus. The latter suggestion is consistent with the recent results of Day et al. (90) showing that cocaine- and amphetamine-induced c-fos and zif268 mRNA and protein expression in the lateral part of the central nucleus of the amygdala can be blocked by so-called processive stressors, such as novel environment, loud noise, or restraint stress, and that the degree of inhibition depends on the intensity of the stressor.
In aggregate, the results of the gene expression studies point to the following (not mutually exclusive) interpretations of the functional role of the central amygdala: 1) the medial and lateral subdivisions of the central amygdala should be considered separately, 2) the medial part appears to be particularly involved in alimentary or even gustatory behaviors, 3) the medial part seems to be intimately linked with the functions of the parasympathetic nervous system, and 4) the lateral amygdala activation appears as a default response that could be actively blocked by the strong aversive stimuli.
Are those interpretations supported by the results obtained with the other approaches to understand functions of the central amygdala? It is known that lesions of the central nucleus, similarly to the lesions of the basolateral complex, impair fear-dependent classical conditioned responses, such as freezing and fear-potentiated startle (56, 70, 87, 178, 184, 187, 258, 291, 405–407, 425, 506, 512, but see Ref. 241). On the other hand, the experiments of Killcross et al. (219) showed that the intact central nucleus is not essential to acquire instrumental defensive responses. There is also a set of Pavlovian conditioning tasks that are impaired only after the central, but not the basolateral, lesions, e.g., conditioned approach response in the Pavlovian appetitive conditioning (autoshaping, Refs. 61, 359), conditioned orienting (134, 170), as well as appetitive and aversive Pavlovian-to-instrumental transfer (160, 186). Since taste aversion learning was shown to be undisturbed by the central amygdala lesions (329, 473), it appears that the CS-US associations that lead an animal to avoid the CS are formed elsewhere. However, the central amygdala was consistently demonstrated to be activated during CTA learning, following either presentation of novel tastes or injection of lithium chloride. Taking into account the involvement of the central nucleus in alimentary conditioning, it seems that this nucleus may be involved in encoding information about specific sensory properties of the stimuli that evoke visceral responses.
As far as electrophysiology is concerned, Nishijo et al. (346), who recorded neuronal activity in the amygdala during discrimination of conditioned sensory stimuli and ingestion of tasty solutions, observed that the taste responsive neurons were located mainly in the central nucleus of the amygdala. Interestingly, these taste neurons also responded to other sensory stimuli, and taste quality was processed based on palatability. Hence, these results suggest that the central amygdala plays a role in evaluation of taste palatability and in association of taste stimuli with other sensory stimuli. Furthermore, it has been shown that the central amygdala is involved in modulating activity of the gustatory neurons in the nucleus of the solitary tract, as well as in the parabrachial nucleus (68, 192). Stimulation of the central nucleus increased the chemical selection of the taste neurons in the parabrachial nucleus, while the central amygdala lesions depressed such effect (192). The modulatory role of the central amygdala seems to depend on the GABAA receptors located within this structure (213). On the other hand, in the nucleus of the solitary tract, modulatory effects of the central amygdala stimulation were mostly excitatory and enhanced responses to taste stimuli (68). The central nucleus of the amygdala was also shown to be involved in the organization of conditioned metabolic endocrine responses, namely, it was observed that lesions of the central amygdala abolished the conditioned insulin response but not the early insulin elevation during the presentation of food (407). Hence, all of these observations link the medial part of the central amygdala with alimentary behaviors.
An involvement of the central nucleus of the amygdala in the mechanisms underlying control of feeding behavior has also been shown by pharmacological approaches. For instance, Mogenson and Wu (325) showed that injecting either dopamine agonist or dopamine antagonist into the central amygdala increased the consumption of NaCl solution. In addition, Minano et al. (318) found that microinjections of GABAA receptor agonist into the central amygdala produced a dose- and time-dependent decrease of food intake in both the satiated and fasted rat. In contrast, microinjection of GABAA receptor antagonist significantly blocked this inhibitory effect. Moreover, Baldo et al. (20) observed that GABAA agonist infusion into the central amygdala reduced feeding elicited either by stimulation of the nucleus accumbens shell or by food deprivation. Similarly, infusion of TTX into the central amygdala blocked hyperphagia induced by the nucleus accumbens shell stimulation. It was also shown that instrumental learning of lever pressing for food reinforcement required dopamine receptor activation within the central amygdala (12). Additionally, Hitchcott and Phillips (179) observed that postsession infusion of dopamine receptor agonist into the central nucleus of the amygdala enhanced acquisition of a conditioned approach response. Conversely, pretest infusion of dopamine receptor agonist attenuated expression of the conditioned approach response. On the other hand, modulation of dopaminergic transmission within the central amygdala did not influence second-order conditioning, in which conditioned stimulus was previously paired with food reward.
In contrast, the central amygdala appears not to be of critical importance for aversive conditioning. For example, Fanselow and Kim (108) showed that administration of NMDA receptor antagonist into the central nucleus of the amygdala did not prevent acquisition of fear. Furthermore, it was observed that stimulation of the central amygdala induced responses opposite to those evoked by fear. For instance, Ciriello and Roder (72) studied contribution of GABAergic inputs in the central nucleus of the amygdala to the cardiovascular responses elicited from this nucleus. They showed that microinjections of glutamate into the central amygdala evoked decreases in mean arterial pressure and heart rate responses. These depressor responses elicited during stimulation of the central amygdala were significantly attenuated after microinjections of GABAA receptor antagonist. Moreover, LeitePanissi and Menescal-de-Oliveira (269) observed that both cholinergic and opioidergic stimulation of the central amygdala decreased the duration of freezing in guinea pigs.
The central amygdala appears to be also involved in regulation of anxiolytic/anxiogenic effects of different drugs. For example, Koch (230) showed that injection of metabotropic glutamate receptor agonist into the central amygdala increased the acoustic startle response in rat. Moreover, infusion of serotonin receptor agonist blocked the fear-potentiated startle response showing anxiolytic properties (153). Conversely, administration of histamine increased an anxiogenic response measured in the elevatedplus maze (528). Interestingly, serotonin receptor agonists, which reduce anxiety, were shown to increase c-Fos expression in the central nucleus of the amygdala (73).
The central nucleus was suggested to regulate the processing of cues, when a predictive relationship between events is first noticed or just altered (134, 135, 184, 187, 188). According to most modern learning theories, the discrepancy between expected and obtained outcomes strongly influence acquisition of learned associations. Examination of the role of the central nucleus in the Pavlovian appetitive conditioning and the blocking/unblocking paradigm with the use of a foot shock as the US revealed that it is involved in orienting or attention responses to signals for significant events rather than in behaviors elicited by the reinforcement itself (134, 184, 187). Moreover, McDannald et al. (298) and Groshek et al. (154), using reversible inactivation of the central nucleus, have shown that it is critical to acquisition of information required for conditioned orienting to visual and auditory CS during its pairing with food, but not to maintain this information for later use. Very recently, Holland and Gallagher (185) have found that surprise-induced enhancement of subsequent learning about a stimulus depended on intact central amygdala at the time of surprise but not during testing of behaviors. Additionally, Gallagher, Holland, and co-workers (66, 166, 167) have shown that dopamine projections to the striatum (the amygdalo-nigrostriatal pathway) as well as cholinergic connections with the substantia innominata and the nucleus basalis are necessary for the control of attention aspects of stimulus processing by the central nucleus. Thus the central amygdala seems to be involved in regulation of attention function, which, in turn, influences associability of stimuli. Interestingly, it was shown that lesions of the central nucleus of the amygdala interfered with surprise-induced enhancement of event processing evoked by omitting an expected signal stimulus or reward (185, 187); however, presenting an unexpected additional reward enhanced the formation of associations in both control and central amygdala-lesioned animals (183).
The lesion and pharmacological research in the animal models of addiction further support our considerations. In particular, dopaminergic and GABAergic systems have been shown to be involved in the perception of drug primary rewarding properties, similarly as it can be involved in perception of and response to the natural rewards. Thus far, microdialysis studies have shown that almost all drug of abuse, including amphetamine, cocaine, nicotine, and alcohol (158, 443, 515, 523, 524), as well as exposure to drug-paired environment (503) induced DA release in the central amygdala. These observations were also confirmed by immunocytochemical detection of DA in the central nucleus of the amygdala following amphetamine administration (369, 370). Furthermore, DA receptor antagonist, SCH 23390, injected into the central amygdala, increases cocaine self-administration (54). Moreover, GABAergic system within the central amygdala is also linked with perception of alcohol rewarding properties. Both acute and chronic alcohol administration to rats was shown to increase GABA release and GABAergic transmission in the central nucleus of the amygdala (394, 395). In agreement with the electrophysiological data, alcohol-induced c-Fos expression in the central amygdala occurs mostly in GABAergic cells (76, 328). Finally, GABAA receptor antagonists, SR 95531 or CCt, injected into the central amygdala were demonstrated to decrease ethanol self-administration (122, 194).
Furthermore, glutamatergic and adrenergic transmission within the central nucleus has been involved in the perception of aversive properties of the withdrawal state and production of the somatic signs of withdrawal. Excitotoxic lesion of the central amygdala with NMDA significantly attenuated withdrawal-induced CPA (500). Similar effects were observed after inhibition of glutamatergic transmission. Microinjection of the AMPA/kainate-glutamate-receptor antagonist CNQX into the central amygdala, as well as the NMDA receptor antagonists MK-801 or D-CPPene significantly attenuated the naloxone-precipitated withdrawal-induced CPA, as well as some somatic signs of withdrawal (only in case of AMPA receptor antagonist) in morphine-dependent rats (467, 498). Furthermore, pretreatment with the NMDA antagonists MK801 and LY274614 or the 2-adrenergic agonist clonidine blocked behavioral signs of morphine withdrawal, as well as withdrawal-induced increase in c-fos mRNA in the central amygdala (388, 389). Morphine withdrawal also transiently elevates extracellular norepinephrine level within the central amygdala (499). In agreement with this observation, injections of -adrenoceptor antagonists, propranolol and timolol, into the central amygdala has been shown to significantly attenuate the morphine withdrawal-induced CPA. Similarly, the 1-antagonist atenolol or the 2-antagonist butoxamine significantly attenuated the CPA. Furthermore, propranolol affected morphine withdrawal-induced somatic signs (499).
Thus the lesion and pharmacological studies in the animal models of addiction indicate that the central amygdala may serve as an interface between perception of the rewarding properties of the drugs or aversive properties of the withdrawal state and production of the appropriate behavioral and physiological reactions driven also by the parasympathetic system.
Neuroanatomical data allow distinguishing at least three major subdivisions of the central amygdala: capsular, lateral, and medial (see Fig. 1 and Ref. 419 for review). This division is based mainly on differential connectivity. For instance, the inputs from the entorhinal (303, 351, 485) and perirhinal cortices (92, 301) provide a large projection to the capsular division of the central nucleus (299, 303), whereas its lateral division is strongly innervated by the amygdalopiriform transition area (202, 303).
Interestingly, most of the taste responsive neurons within the amygdalar complex were found in the central and lateral nuclei (299). It seems that one of the most significant functions of the central nucleus may be regulation of appetitive behavior, by providing a link between environmental stimuli and those brain structures that are involved in the direct control of feeding behavior (for review, see Ref. 28). Substantial projections to the lateral subdivision of the central nucleus come from the anterior agranular and the anterior dysgranular gustatory insular cortex. Light to moderate projections from these cortical areas to the medial subdivision of the central nucleus (299) were also seen. Furthermore, among numerous thalamic sensory and nonspecific projection nuclei (see Fig. 2), the ventral posteromedial nucleus that receives gustatory and visceral inputs from the nucleus solitarius projects to the central nucleus (353, 477, 522). The medial part of the parabrachial nucleus that is significantly involved in processing gustatory information also sends inputs to the medial subdivision of the central nucleus (299). Finally, the central nucleus is strongly and reciprocally connected with the lateral hypothalamus, and the dorsal vagal complex (6, 491).
Plasticity of the central amygdala neurons has not been extensively studied. However, Samson and Paré (424) as well as Fu and Shinnick-Gallagher (128) reported phenomena of the central amygdala LTP, evoked by stimulation of pathways originating from the basolateral/lateral amygdala.
In aggregate, the most plausible explanation for the functional role of the medial part of the central amygdala seems to be that it serves to coordinate acquisition of alimentary/gustatory behaviors. On the other hand, the lateral subdivision appears to be mainly involved in attention and vigilance.
An elevated expression of c-Fos in the medial amygdala was shown following two-way avoidance training, conditioned emotional response training, two-odor discrimination test in which incorrect responses were punished, exposure to predatory presence or odor, aggressive encounter, mating or cues from sexually mature partner, exposure to the pups or sensory cues from the pups, social interaction, and exposure to conspecific pheromones. Frequently, but not always, activation of the medial nucleus of the amygdala was observed after fear conditioning and exposure to the elevated mazes. On the other hand, activation of the medial amygdala was rather not found following fear memory retrieval. c-Fos protein expression seems to be also activated in the medial amygdala by acute, passive administration of cocaine. The data on other drugs and procedures are either very scarce or inconsistent.
Notably, activation of the medial amygdala following training of different defensive responses has consistently been demonstrated. An involvement of the medial nucleus in governing such responses seems to be supported by the results of some of lesion and pharmacological studies. However, one should keep in mind that the role of the medial amygdala (as well as the cortical nuclei, see below) in learning has not often been paid too much attention. For instance, Cahill and McGaugh (53) used the animals with partially spared medial nuclei to show that the lesions of the amygdala impaired the retention of one-trial aversive learning, but not one-trial appetitive task. This is also the case for Lehmann et al. (268), who showed that inactivation of the amygdala spared anterograde memory for shock-probe fear conditioning, but the medial and cortical nuclei remained intact in the applied procedure. Therefore, there is very little known about the effects of the medial amygdala lesions on explicitly studied learning. Oakes and Coover (348) demonstrated that the medial nucleus is not involved in passive avoidance of drinking. In contrast, Holahan and White (182) showed that the medial nucleus lesions impaired conditioned avoidance response, as well as posttraining memory modulation by previously conditioned aversive stimuli. Moreover, it was observed that freezing response in contextual conditioned fear and fear potentiated acoustic startle are impaired by the medial amygdala lesions or blocking (38, 182, 493). The specific role of medial amygdala in controlling defensive behaviors was also shown with the use of pharmacological approach. For instance, Muller and Fendt (331), who injected GABAA receptor agonist into the medial amygdala, showed that such temporary inactivation completely blocked TMT-induced freezing. It was also observed that infusion of AMPA receptor antagonist into the medial amygdala before testing blocked fear-potentiated startle (493).
Primarily, the medial amygdala has been postulated to be involved in processing social olfactory signals (45), in particular in regulation of sexual behaviors (for review, see Refs. 35, 367, 487). Such a view is supported by the effects of lesions of the medial amygdala that were shown to impair sexual reactions (238, 239, 294, 306, 386). The specific role of the medial amygdala in sexual activities is consistent also with the results obtained with pharmacological approach. For instance, Polston et al. (376) showed that pharmacological activation of the medial amygdala with the glutamate agonist NMDA initiated nocturnal prolactin surges that cause a pseudopregnancy state in female rats. Such prolactin surges are normally evoked by vaginocervical stimulation. Furthermore, pregnancy or pseudopregnancy initiation by mating was prevented by intra-amygdalar infusion of the NMDA receptor antagonist administered before mating. The medial amygdala is also a target for many hormonal and chemosensory signals regulating reproductive behavior (for detailed analysis of this issue, see Ref. 342).
However, Baum and Everitt (26), who compared c-Fos activation in noncopulating male rats, males who had free access to females though did not mount, and males copulating to ejaculation, observed that the level of c-Fos expression in the medial amygdala did not depend on the degree of copulatory experience under those conditions (but see also Refs. 375, 413). The medial amygdala has been also hypothesized as engaged in controlling of agonistic behavior. Such behaviors were indeed shown to increase c-Fos expression in the medial nucleus of the amygdala (203, 235). However, it was also shown that animals simply interacting with a conspecific in a nonaggressive fashion had the same c-Fos expression pattern in the medial amygdala as did the aggressive animals (203, 312).
These findings might have pointed out the role of the medial amygdala in response to the social interaction; however, further analysis of the results of studies investigating the pattern of genes expression after various behavioral tests shows that the medial amygdala is also active during very different, not clearly social tasks, e.g., fear conditioning, two-way avoidance training, exposure to the elevated mazes, or nose-poking for water reward. In aggregate, one can come to the conclusion that the expression of gene activity markers in the medial amygdala is evoked by every novel aspect of the experimental situation (see Ref. 88). However, some authors did not observe activation of the medial amygdala evoked by learning in which the novel components of experimental situation were certainly introduced (e.g., Refs. 177, 274, 275, 315, 383, 410, 451).
Another alternative is functional dissection of the medial amygdala. Such functional compartmentalization of this nucleus has already been suggested (69, 488), with the posterodorsal part of the medial amygdala being proposed to regulate reproductive behaviors, while posteroventral and anterior ones are defensive (69). The data on c-Fos expression, which could be used to test this hypothesis, are, however, either scarce or inconsistent, thus not allowing for drawing any firm conclusions.
All above-mentioned discrepancies might be explained by relatively high level of c-Fos expression observed in the medial amygdala in the control groups. Such high level of expression might cause a "ceiling effect" that obscures obtained results. In this context, it should be, however, reminded that in very well habituated animals c-Fos expression is negligent. Hence, the solution to the problem might be found in the longer experimental designs, in which the animals are initially very well habituated to the experimental situation. For instance, Hess et al. (177) observed the lack of c-fos mRNA expression in the medial amygdala of the overtrained rats. Similar habituation (five 1-h daily sessions of habituation to the experimental situation) allowed observing the clear difference not only between the experimental and control groups but also between the level of c-Fos expression induced by appetitively and aversively motivated behaviors (229).
The connectivity of the medial amygdala seems to support its role in controlling defensive as well as sexual responses, especially in processing of olfactory cues. The medial amygdala is innervated by various thalamic nuclei, including the nonspecific projection, somatosensory, and auditory relay nuclei (353, 486, 522; Fig. 2). The medial nucleus also receives substantial projection from the dorsal and lateral hypothalamic areas (6, 352); however, most of its hypothalamic inputs come from the periventricular and medial areas. Thus the medial nucleus is innervated by the paraventricular nucleus, ventromedial and arcuate nuclei, as well as by the premammillary nuclei and medial preoptic area (352, 398). It also receives fibers from the bed nucleus of stria terminalis, substantia innominata, and ventral pallidum and is a considerable target of olfactory inputs coming from the dorsal endopiriform nucleus, accessory olfactory bulb, and the nucleus of the lateral olfactory tract (430). Moreover, it receives inputs from several basal forebrain and brain stem structures, such as the dorsal raphe, parabrachial, and peripendicular nuclei (351). The medial amygdala sends projections to the cortical areas involved in olfactory functions, such as the entorhinal and piriform cortices. Moreover, it innervates the infralimbic and prelimbic cortices, claustrum, and ventral subiculum (59) (Fig. 2). The medial amygdala provides also extensive output to a variety of hypothalamic autonomic regions located mainly in the medial and periventricular areas, but also in the posterior and lateral hypothalamus (59, 382). On the other hand, it sends fibers only to a few nonspecific thalamic nuclei. Furthermore, many basal forebrain structures, including the bed nucleus of stria terminalis, substantia innominata, lateral and medial septal nuclei, are innervated by the medial nucleus. It also substantially projects to various olfactory-related forebrain regions, for instance, to the olfactory tubercle, anterior olfactory nucleus, accessory olfactory bulb, endopiriform nucleus, and the bed nucleus of the accessory olfactory tract (59, 91, 92). Finally, the medial nucleus is a source of afferents that terminate in several brain stem structures that are related to defensive functions, for instance, the raphe nuclei, ventral tegmental area, and pontine central gray (59).
In conclusion, the media amygdala appears to play a particular role in processing novel aspects of external environment, both of social and nonsocial nature.
It seems that the cortical group of nuclei, likewise the medial amygdala, is activated by many different behaviors. Induction of the expression of c-Fos within the cortical nuclei was shown following the aversively and appetitively motivated behaviors, both in standard and especially ethologically more relevant tests, e.g., mating and pup recognition. Interestingly, Savonenko et al. (428), who studied c-Fos expression following two-way avoidance acquisition, found that activation of the cortical amygdala correlated with grooming behavior (reflecting the lack of fear). In the longer experimental designs, in which the animals were well habituated to the experimental situation, the cortical nuclei of the amygdala were activated only following appetitively motivated behaviors (229). Moreover, their involvement in forming olfactory memory of a lamb was also observed (215). Thus it appears that the pattern of activation of the cortical amygdala might depend on some aspects of behavioral tests. However, existing data on activation of the cortical nuclei of the amygdala are still incomplete and often inconsistent; hence, this interesting issue requires further studies. Nevertheless, it might be of interest to see whether data provided by other than gene expression experimental approaches are in line with proposed role of this part of the amygdala in processing of fear (428).
The role of the cortical nuclei (even more than in the case of the medial amygdala, see above) has often been neglected as far as other approaches are concerned. Hence, there are almost no data on the role of the cortical amygdala in learning processes. However, increase of the threshold for rewarding medial forebrain bundle (MFB) stimulation was shown after electrolytic lesions to the cortical and adjacent amygdala subnuclei (37, 314), suggesting that the cortical nuclei may modulate MFB reward signals.
As far as the cortical amygdala connectivity is concerned, it seems that it is well positioned to be involved in controlling a variety of behaviors, especially associated with processing of olfactory cues. Sensory-related inputs to the cortical nuclei originate in several temporal areas, including the posterior agranular and gustatory dysgranular insular cortices, as well as in the piriform, entorhinal, and perirhinal cortices (299, 301, 303, 352, 446, 448; Fig. 4). Substantial inputs from several areas of the frontal cortex, such as the ventral agranular insular and infralimbic cortices, as well as inputs from the cingulate cortex and subiculum have also been noted (60, 193, 299, 304, 352, 465). The anterior cortical nucleus is innervated by numerous thalamic midline nuclei, as well as by the mediodorsal nucleus and medial geniculate body (324, 353, 477, 486, 522). Moreover, several basal forebrain and brain stem structures innervate the cortical nuclei. These include the bed nucleus of stria terminalis, substantia innominata, dorsal raphe, and parabrachial nuclei (34, 155, 156, 352). Substantial olfactory projections come from the olfactory bulb and accessory olfactory bulb, as well as from the endopiriform nucleus (31, 299, 379, 430). The cortical nuclei project to several brain structures that are involved in the olfactory functions, such as the entorhinal and piriform cortices, the accessory olfactory bulb, olfactory tubercle, and endopiriform nucleus (58, 91, 281, 365, 464). Relatively strong projections terminate in the bed nucleus of stria terminalis and in the substantia innominata (58, 156). Furthermore, some projections from the anterior cortical nucleus to the caudatoputamen and nucleus accumbens were noted (300). On the other hand, the posterior cortical nucleus sends substantial outputs to the agranular insular and infralimbic cortices, ventromedial and premammillary hypothalamic nuclei, and the bed nucleus of stria terminalis (58, 244).
In aggregate, the aforementioned data do not allow for drawing unequivocal conclusions about the cortical amygdala. However, its link with processing of safety signals, in particular involving olfactory cues, appears to be well supported.
|
VI. CONCLUSIONS: THE FUNCTIONAL HETEROGENEITY OF THE AMYGDALA |
|---|
|
TopPreviousNextReferences |
|---|
As discussed at the beginning of this review, there are two major contemporary theoretical explanations for amygdala heterogeneity, namely, serial versus parallel processing of information (see Fig. 5 and Ref. 22 for discussion). Those proposals tend to focus on the paths of the information flow through the amygdala. However, at least three levels of neuronal functions should be considered here: 1) connectivity that allows for an information flow, without modification of its content; 2) computational elaboration of the transduced information, producing an output that results from comparing various inputs; and 3) plastic modifications that produce lasting changes in the neurons that carried out the information, thus allowing them to transfer the next portion of the information somewhat differentially than before. The serial versus parallel processing hypotheses apparently do not address those three functions specifically. Furthermore, the c-Fos expression patterns reviewed herein (see, e.g., Fig. 12) are difficult to reconcile with either "serial" or "parallel" approaches. In particular, it appears that the amygdala involvement is neither serial (i.e., sequentially engaging basolateral and central nuclei) nor parallel (i.e., either engaging basolateral or central nuclei). On the other hand, it shall be stressed that taking into account the possible c-Fos function in neuronal plasticity, its activity mapping appears to describe mainly the propensity of neurons within the given amygdalar subdivision to undergo the plastic changes and not just an activation linked purely to the information flow.
|
Alternatively, behavioral sequences may be classified according to their physiological attributes. Some environmental conditions, such as either sudden unexpected danger or painful stimulation, evoke very fast and robust phasic responses that involve increased sympathetic tone overriding parasympathetic activity. This kind of behavioral strategy aims at defending the subject's life. However, such a dramatic phasic response is unnecessary in most situations. It is reasonable to suppose that in most environmental conditions the biologically appropriate responding is equally determined by external information and current physiological state of an organism. For instance, nearly all alimentary, sexual, exploratory, and social behaviors, as well as subject's emotional responses to undefined and indirect environmental dangers, are predominantly composed of tonic responses, in which parasympathetic or vagal tone either masks or overrides sympathetic effects.
Therefore, we hypothesize that plastic changes within the central nucleus of the amygdala mediate tonic behavioral responses, whereas plastic changes in the basolateral amygdala enable phasic behavioral responses (see Fig. 13).
|
|
ACKNOWLEDGMENTS |
|---|
|
TopReferences |
|---|
Address for reprint requests and other correspondence: L. Kaczmarek, Nencki Institute, 02-093 Warsaw, Poland (e-mail: L.Kaczmarek{at}nencki.gov.pl).
|
REFERENCES |
|---|
|
Top |
|---|
