This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Pacák, K. Articles by Palkovits, M. Search for Related Content PubMed PubMed Citation Articles by Pacák, K. Articles by Palkovits, M.

Stressor Specificity of Central Neuroendocrine Responses: Implications for Stress-Related Disorders

Pediatric and Reproductive Endocrinology Branch (K.P.), National Institute of Child Health and Human Development and Clinical Neurocardiology Section (K.P.), National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, Maryland; Laboratory of Genetics (M.P.), National Institute of Mental Health, National Institutes of Health, Bethesda, Maryland 20892-1583; and Laboratory of Neuromorphology, Semmelweis University, 1094 Budapest, Hungary (M.P.)

	Abstract

Top Abstract I. Introduction II. Stress Concept III. Brain Regions Involved... IV. Methods Used for... V. Stressor Specificity of... VI. Stressor-Specific Activation... VII. Clinical Relevance of... References

 
Despite the fact that many research articles have been written about stress and stress-related diseases, no scientifically accepted definition of stress exists. Selye introduced and popularized stress as a medical and scientific idea. He did not deny the existence of stressor-specific response patterns; however, he emphasized that such responses did not constitute stress, only the shared nonspecific component. In this review we focus mainly on the similarities and differences between the neuroendocrine responses (especially the sympathoadrenal and the sympathoneuronal systems and the hypothalamo-pituitary-adrenocortical axis) among various stressors and a strategy for testing Selye’s doctrine of nonspecificity. In our experiments, we used five different stressors: immobilization, hemorrhage, cold exposure, pain, or hypoglycemia. With the exception of immobilization stress, these stressors also differed in their intensities. Our results showed marked heterogeneity of neuroendocrine responses to various stressors and that each stressor has a neurochemical "signature." By examining changes of Fos immunoreactivity in various brain regions upon exposure to different stressors, we also attempted to map central stressor-specific neuroendocrine pathways. We believe the existence of stressor-specific pathways and circuits is a clear step forward in the study of the pathogenesis of stress-related disorders and their proper treatment. Finally, we define stress as a state of threatened homeostasis (physical or perceived treat to homeostasis). During stress, an adaptive compensatory specific response of the organism is activated to sustain homeostasis. The adaptive response reflects the activation of specific central circuits and is genetically and constitutionally programmed and constantly modulated by environmental factors.

I. Introduction

II. Stress Concept

A. Definition of stress

B. Classification of stressful stimuli

C. Selye’s doctrine of nonspecificity revisited

III. Brain Regions Involved in Neuroendocrine Responses to Stress

A. Central autonomic system

B. Central aminergic systems

C. Noncatecholaminergic brainstem neurons

D. Thalamus

E. Neuroendocrine hypothalamus

F. Limbic system

IV. Methods used for Mapping Stressor-Specific Neuronal Circuits

A. Intracerebral microdialysis

B. Protooncogene-"immediate early genes"—immunohistochemistry

V. Stressor Specificity of Central Neuroendocrine Responses

A. Immobilization stress

B. Cold stress

C. Insulin-induced hypoglycemia

D. Hemorrhage

E. Pain stress

VI. Stressor-Specific Activation of Other Neuroendocrine Systems

VII. Clinical Relevance of Stressor Specificity and Future Perspectives

	I. Introduction

Top Abstract I. Introduction II. Stress Concept III. Brain Regions Involved... IV. Methods Used for... V. Stressor Specificity of... VI. Stressor-Specific Activation... VII. Clinical Relevance of... References

 
HANS SELYE DESERVES much of the credit for introducing the term "stress," and for popularizing the concept of stress in the scientific and medical literature of the 20^th century. He wrote in his book, Stress of Life: "This lack of distinction between cause and effect was, I suppose, fostered by the fact that when I introduced the word stress into medicine in its present meaning, my English was not yet good enough for me to distinguish between the words stress and strain. It was not until several years later that the British Medical Journal called my attention to this fact, by the somewhat sarcastic remark that according to Selye stress is its own cause. Actually I should have called my phenomenon the strain reaction and that which causes it ‘stress,’ which would parallel the use of these terms in physics. However, by the time that this came to my attention, biological stress in my sense of the word was so generally accepted in various languages that I could not have redefined it. Hence, I was forced to create a neologism and introduce the word stressor, for the causative agent, into the English language, retaining stress for the resulting condition" (1).

Confusion still arises regarding what one believes defines and constitutes stress. Despite the fact that thousands of research articles have been written about stress and stress-related diseases, until now no scientifically accepted definition of stress exists (2). This results in the view that stress can be practically anything that contributes to virtually any disease in humans. Most scientists view stress as the situation when the hypothalamo-pituitary-adrenocortical (HPA) axis, represented mainly by elevated ACTH levels, is activated (3). Others suggest that activation of other systems with or without an elevation in ACTH may reflect stress-induced disturbed homeostasis (4, 5). Several review articles and book chapters have summarized data from hundreds of stress-related studies and drawn conclusions relating to different aspects of the stress response (2, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17). This review focuses on two major points: 1) evidence that specific stressors may elicit specific responses, and 2) different stressors may activate different brain systems by using specific pathways within the central nervous system. Particular attention has been paid to Selye’s doctrine of nonspecificity of stress responses, which has been tested in our laboratory. Based on our data using five different acute stressors (immobilization stress, hypoglycemia, nontraumatic hemorrhage, pain stress, and cold stress) and several previous reports by others (4, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29), we turned our attention to identifying stressor-specific neuronal circuits in brain and their involvement in stress-related diseases.

	II. Stress Concept

Top Abstract I. Introduction II. Stress Concept III. Brain Regions Involved... IV. Methods Used for... V. Stressor Specificity of... VI. Stressor-Specific Activation... VII. Clinical Relevance of... References

 
A. Definition of stress
Cannon (30, 31, 32) was the first to introduce the term "homeostasis" to describe the "coordinated physiological processes which maintain most of the steady states in the organism." He turned his attention to the sympathetic nervous system as an essential homeostatic system that serves to restore stress-induced disturbed homeostasis and to promote survival of the organism. Cannon was also the first to touch on the issue of specificity of stress responses since he showed, for example, that the specific stabilizing or homeostatic reaction to lack of oxygen is quite different from that with which the body responds to exposure to cold; this, in turn, is virtually the reverse of that required to resist heat (1). However, Cannon never used the term "stress."

Selye introduced and popularized stress as a medical and scientific idea. The starting point for the elaboration of his stress theory was his report, published as a letter to Nature in 1936 (33), describing a pathological triad (adrenal enlargement, gastrointestinal ulceration, and thymicolymphatic involution) elicited by any of a variety of stressors. From this pathological triad he developed a theory of stress that attained wide popularity and aroused intense research interest but also incited controversy, which persists to the present. He defined stress as the nonspecific response (revealed after subtraction of the specific components from the total response) of the body to any demand, emphasizing that the same pathological triad— "stress syndrome"— would result from exposure to any stressor. According to Selye, these demands on the body included bacterial infection, toxins, x-irradiation, and various physical stimuli such as surgery and muscular exercise. Selye’s stress theory did not deny the existence of stressor-specific response patterns; however, he emphasized that such responses did not constitute stress, which was the shared nonspecific component.

Selye mainly focused on the HPA axis as the key effector of the stress response. He considered the adrenal cortex "the organ of integration which participates in the normal and pathological physiology of virtually all tissues in the body," by virtue of its endocrine function (34). Indeed, administration of ACTH can elicit all three components of the pathological triad (34). However, Selye did not assert that HPA activation attending stress reflected the organ pathology in the pathological triad. If anything, Selye asserted the converse.

Selye also introduced the term "general adaptation syndrome" with its three successive phases: the alarm, resistance, and exhaustion stages. He wrote that during the stages of the "general adaptation syndrome" the intensity of the stress response might vary; however, the neural and endocrine patterns characterizing the stage of "alarm" would be essentially the same as those characterizing the other stages. He and others proposed an immense list of diseases of adaptation including hyperfunctional and dysfunctional conditions such as Cushing’s disease, hypertension, adrenal tumors, and others. Hypofunctional states included Addison’s disease and cancer (1, 2, 6, 7, 17, 34). Later, Selye proposed that most of the stressful stimuli induce two types of responses: 1) a general stress response, which is common to all stressors and involves the release of ACTH and adrenal corticosterone, and 2) individual stress responses mediated by "conditioning factors," such as genetically determined predispositions (17).

In contrast to Selye, Cannon recognized the importance of psychological as opposed to physical responses during stress (30, 35). From an evolutionary perspective he questioned whether a stereotyped response pattern could be adaptive, recognizing that a nonspecific stress response would not have provided an advantage in natural selection and thus, would not have evolved. Others, like Mason, properly noted that in response to different stressors, activity of the HPA axis could increase, decrease, or remain unchanged, implying that the presence of a pathological triad may not indicate the occurrence of stress (2, 13, 14, 15). Mason proposed that elicitation of an emotion such as anxiety or fear constituted the basis for similar neuroendocrine responses to different stressors.

Many current views concerning what stress means and how to define and approach it exist, but none has been widely accepted. Many of these theories were discussed in detail and described by Goldstein (Ref. 2 and Table 1). Important contributions to these theories have been made by Weiner (36) and Chrousos (6) and Chrousos and Gold (7). Weiner correctly pointed to specificity of stressor responses by describing stressors as selective pressures from the physical and social environment that threaten or challenge an organism and elicit compensatory response patterns. Chrousos and Gold defined stress as a state of disharmony or of threatened homeostasis, evoking physiologically and behaviorally adaptive responses that can be specific to the stressor or generalized and nonspecific and that usually occur stereotypically, producing a "nonspecific" stress syndrome when the threat to homeostasis exceeds a threshold. They included genetic polymorphisms as well as alterations in the expression of genes and environmental factors as important determinants of individual stress responses.

Recently, McEwen (11) introduced the term "allostasis" into stress research. Allostasis, which may be defined as an ability to maintain stability of the internal milieu through change, was originally proposed by Sterling and Eyer (37). As discussed recently in detail by McEwen (11), allostasis refers to the active process of adaptation by productions of various mediators such as adrenal steroids, catecholamines, cytokines, tissue mediators, and immediate early genes. Upon exposure to a chronic stressful situation, physiological responses are initiated, leading to allostatic (adaptive) responses. These responses involve major systems similar to the stress effector systems that were described previously. If allostatic responses are efficient, adaptation occurs and the organism is protected from damage. In situations where allostatic responses are prolonged, inadequate, overstimulated by repeated "hits" from multiple stressors or if a lack of adaptation occurs, allostatic load results in maladaptation and damage to various organs (11, 16). In contrast to homeostatic mechanisms, allostatic regulations are broader and do not depend on set-point mechanisms, signals are not constant, and anticipation of need is an important element. Another aspect of this theory is that allostatic load also reflects aspects of lifestyle (e.g., eating a high-fat diet, lack of exercise, etc.) and disturbances of diurnal rhythms (e.g., sleep deprivation) that result from overexposure of various tissues to stress mediators. Allostatic theory also continues Selye’s notion of "conditioning factors" to explain individual differences in stress responses.

Based on our previous findings of the existence of stressor-specific neuroendocrine responses and mapping of stressor-specific central circuits that participate in these responses (see below), we attempted to define stress as a state of threatened homeostasis (physical or perceived treat to homeostasis). During stress, an adaptive compensatory specific response of the organism is activated to sustain homeostasis. The adaptive response reflects the activation of specific central circuits and is genetically and constitutionally programmed and constantly modulated by environmental factors.

Another "mainstream" theory of stress has been offered recently by molecular biologists regarding the role of heat shock proteins in cellular survival (38). Ironically, their theory posits essentially the same doctrine of nonspecificity that Selye espoused; regardless of the insult, cells respond in the same way.

Sapolsky and co-workers (39, 40, 41) and McEwen et al. (42) introduced and discussed in great detail new aspects of stress in terms of its adverse effects on various brain regions, especially the hippocampus. Upon exposure to stressors, glucocorticoids are released and act on target cells including brain cells. This central action of glucocorticoids is associated with behavioral, neurochemical, and neurodegenerative changes. Neurodegenerative changes are of great importance since they occur in the hippocampus, one of the brain regions involved in memory processes and other cognitive functions as well as in the regulation of the HPA axis (43, 44). Prolonged exposure to high glucocorticoid levels, as commonly seen upon exposure to chronic stress, causes premature age-related changes in hippocampal electrical activity (45) and dendritic and neuronal atrophy often accelerated upon exposure to neurological insults (e.g., hypoxia) (41). In contrast to these neurodegenerative changes, glucocorticoids also evoke responses that are neuroprotective during exposure to stress (46). For example, various stressors and glucocorticoids increase mRNA expression for oligodendrocyte markers such as glycerol-3-phosphate dehydrogenase and neuronal neurotrophin-3 (46, 47). According to Nichols et al. (46), the activation of these markers serves an important adaptive mechanism in promoting oligodendrocyte survival in response to high glucocorticoids levels. In general, glucocorticoids are viewed as key stress hormones that permit, stimulate, or suppress ongoing stress responses, or are preparative during exposure to a subsequent stressor (48).

Oxidative stress is another type of stressor that participates in neurodegeneration of brain cells (49, 50). Expression of mRNA for glial fibrillary acidic protein, an intermediate filament of astrocytes, is increased by oxidative stress, resulting in astrocyte hyperactivation and subsequent damage (49, 50). Recently, CRH and mifepristone, a potent antagonist of glucocorticoid and progesterone receptors, have been shown to protect against neuronal cell death upon exposure to oxidative stress (51, 52). CRH has a neuroprotective action in CRH receptor type 1-expressing neurons against oxidative cell death (52). This CRH protective function is accompanied by increased release of nonamyloidogenic soluble amyloid ß-precursor protein and by suppression of nuclear factor-B. The neuroprotective activity of these drugs may play an important role in new therapeutic interventions for neurodegenerative conditions such as stroke or Alzheimer’s disease (51).

B. Classification of stressful stimuli
A stressor may be viewed as a stimulus that disrupts homeostasis. In general, stressors can be divided into four main categories: 1) physical stressors that have either a negative or, in some situations, a positive psychological component; 2) psychological stressors that reflect a learned response to previously experienced adverse conditions; 3) social stressors reflecting disturbed interactions among individuals; and 4) stressors that challenge cardiovascular and metabolic homeostasis (4, 10, 53). Physical stressors include cold, heat, intense radiation, noise, vibration, and many others. Chemical stressors include all poisons. Pain stress may be elicited by many different chemical and physical agents. Psychological stressors profoundly affect emotional processes and may result in behavioral changes such as anxiety, fear, or frustration. Social stressors include an animal’s placement into the territory of a dominant animal, and in humans, unemployment and marital separation, among others, are considered social stressors. Stressors that disturb cardiovascular or metabolic homeostasis include exercise, orthostasis, upright tilt, heat exposure, hypoglycemia, and hemorrhage. Many of the stressors described above and used in animal research, however, are mixed and act in concert, such as handling, immobilization stress, anticipation of a painful stimulus, and hypotensive hemorrhage.

In terms of duration, stressors may be divided into two main categories: acute (single, intermittent, and time-limited exposure vs. continuous exposure) vs. chronic (intermittent and prolonged exposure vs. continuous exposure) stressors. It should be noted that many stressors differ in their intensity.

The adaptive responses that are elicited in response to an acute stressor include the physiological and behavioral processes that are essential to reestablish homeostatic balance. During an acute stress response, physiological processes are important to redirect energy utilization among various organs and selectively inhibit or stimulate various organ systems or their components to mobilize energy reserves and to be prepared for exposure to additional, unpredictable challenges. Thus, upon exposure to metabolic stressors, certain tissues tend to reduce their consumption of energy while others, especially those that are important for locomotor activity, receive sufficient nutrients to function properly. The central nervous system also has priority during metabolic stress responses and preferentially receives a sufficient amount of nutrients from the circulation. The increased supply of energy to "crucial" organs is achieved preferentially by release of catecholamines and glucocorticoids that, in general, increase gluconeogenesis and glycogenolysis, inhibit glucose uptake, and enhance proteolysis and lipolysis. The immune system is another essential component of these physiologically adaptive stress responses.

In terms of health consequences during exposure to various stressors, the mechanisms of coping with stress and relevant feedback mechanisms are essential for an organism to develop less severe stress-related health consequences and to survive (54, 55, 56, 57). Coping responses during stress may be defined as cognitive and behavioral responses to manage stress (54, 55, 56, 57, 58). Cohen and Lazarus (59) defined five primary goals for successful coping with stress: 1) reduce harmful environmental conditions and enhance the prospect for recovery; 2) tolerate or adjust to negative events; 3) maintain a positive self-image; 4) maintain emotional equilibrium; and finally 5) preserve social relationships. There are a number of factors that determine whether an individual will cope effectively with a particular stressor. One of these factors, called the "relevant feedback," is the appropriate feedback from coping responses (55, 56, 57). For example, if the relevant feedback to a stressor (unsignaled shock) is low, stress- related symptomatology, e.g., gastric ulceration, increases, while if the relevant feedback is high (signaled shock) less symptomatology is present. Other factors involve appropriate neuroendocrine responses.

The role of neuroendocrine responses in coping with stress is well recognized, since without these responses an organism would be less likely to survive many stressful situations (60). One important feature of successful coping with stress is that physiological systems are not only turned on efficiently by a particular stressor but are also turned off again after a stressor has ceased (60, 61). Thus, when these systems (e.g., neuroendocrine systems) are not rapidly mobilized and then appropriately reduced, elevated hormone levels become dangerous for an organism, resulting in various stress-related diseases (e.g., hypertension, stroke, diabetes, obesity, autoimmune and inflammatory disorders, etc.) (60). The extent to which an individual can cope with stressful situations varies, and these differences are a product of genetics, developmental influences, experience, training, social support, and current mental and physical health (58, 60, 61).

C. Selye’s doctrine of nonspecificity revisited
The parameters of stress observed by Selye were all derived from release of ACTH, which elicits hormonally mediated responses. Stressors, however, also elicit neuronally mediated sympatho-adrenomedullary responses, which although recognized by Selye from Cannon’s work, remained unmeasured and therefore were not considered in the syndrome described by Selye (1, 17, 34). Selye described that stressors do not differ in terms of the "patterns" of stress responses. Only after stressor-defined patterns were removed from consideration could one approach the stereotyped stress syndrome. This syndrome could be graded in intensity, but the pattern of response would not be defined by the stressor.

Testing Selye’s hypothesis in our laboratory was possible only by comparing the relative magnitudes of several independent neural and hormonal responses at different intensities of stressors using a simplifying assumption: that the magnitudes of both the specific and nonspecific components vary directly with the intensity of the stressor over the whole range of stressor intensities, i.e., that there is no ceiling for the specific component, and no threshold for the nonspecific component (4). Thus, if there is a single unitary response to all stressors, then at two different intensities of the same stressor, the ratios of the increments in the responses should be the same for all parameters, regardless of the stressor. By comparing ratios of differences in response to low- and high-intensity stressors, we examined the theory of nonspecific response patterns.

We measured arterial plasma NE, epinephrine (EPI), and ACTH concentrations in conscious Sprague Dawley rats after exposure to one of five different stressors: immobilization (2 h), hemorrhage (10% or 25% of estimated blood volume; the latter producing hemorrhagic hypotension), cold exposure (4 C or -3 C), pain (evoked by subcutaneous administration of 1% or 4% formalin), or hypoglycemia (evoked by intravenous injected insulin at one of three doses: 0.1, 1.0, or 3.0 IU/kg). For each plasma measure for each animal, an area under the curve (concentration x time) was calculated.

For all three plasma measures, net total responses varied by more than 50-fold across stressors (Fig. 1). At their highest intensity, all of the stressors resulted in significant increases in levels of ACTH, NE, and EPI compared with control values obtained after intravenous saline injection. Immobilization stress evoked large increases in plasma levels of ACTH, NE, and EPI, but other stressors induced disproportionately large NE or EPI responses compared with ACTH responses (Fig. 1). Thus, whereas for ACTH, immobilization stress evoked the largest responses and cold was relatively ineffective, for NE, cold evoked the largest responses, and for EPI, insulin evoked the largest responses. The largest increment in plasma EPI levels after administration of insulin is consistent with the homeostatic effect of EPI in antagonizing the actions of insulin and promoting release of glucose from the liver. Similarly, the largest NE responses after cold are consistent with sympathetic activation to conserve heat by piloerection, vasoconstriction, and energy expenditure. No clear specific responses were found for immobilization stress, hemorrhage, and pain stress; therefore, these stressors were suitable for testing of Selye’s doctrine of nonspecificity. Immobilization stress could not be used because the intensity of this stressor could not be varied. Thus, only data for ACTH, NE, and EPI responses to hemorrhage and formalin were appropriate for testing the doctrine of nonspecificity.

View larger version (22K): [in this window] [in a new window]   Figure 1. Responses of plasma ACTH, NE, and EPI during exposure of conscious rats to various stressors. Each bar represents the mean value for the net area under the curve (AUC) for that stressor, where net AUC for each animal was calculated from the baseline AUC subtracted from the total AUC. SAL, Saline; INS, insulin; HEM, hemorrhage; FORM, formalin; IMMO, immobilization.

As shown in Fig. 2, for plasma EPI, the ratio of the response for the 25% hemorrhage to the 10% hemorrhage was smaller than the ratio of the response for 4% formalin to 1% formalin. The doctrine of nonspecificity would predict that the difference between hemorrhage and formalin would also obtain for plasma ACTH; in fact, however, for plasma ACTH, the ratio of the response for the 25% hemorrhage to the 10% hemorrhage was much larger than the ratio of 4% formalin to 1% formalin. The increment in plasma EPI levels between the two intensities of formalin was larger than the increment in plasma ACTH levels; yet the increment in plasma EPI levels between the two intensities of hemorrhage was smaller than the increment in plasma ACTH levels. Clearly, the response patterns to these two stressors are not identical.

View larger version (29K): [in this window] [in a new window]   Figure 2. Ratios of plasma NE, EPI, and ACTH responses to greater or lesser intensities for hemorrhage and formalin. The application of Selye’s doctrine of nonspecificity would predict that arrows should be parallel to each other and of the same length.

The central nervous system plays a crucial role in elicitation and modulation of compensatory stress response patterns. Although a large number of neurotransmitters, neuropeptides, and neuromodulators are activated in various brain regions during exposure to stress, one can predict that specific neuronal circuits exist to optimize effective, rapid, and efficient responses to restore disturbed homeostasis and ensure minimal damage to the organism. This is supported by the elegant work of Gaillet et al. (62), who suggested a differential involvement of PVN noradrenergic pathways in the regulation of the HPA axis according to the nature of the stressor. Thus, identification of such "stressor-specific" anatomical and functional circuits would be extremely important in developing future therapies for stress-related disorders. In this review, based on our studies and the work of others, we have attempted to describe stressor-specific anatomical circuits. The mapping of these stressor-specific neuroanatomical circuits is the first step to move the field of stress research in a new direction.

	III. Brain Regions Involved in Neuroendocrine Responses to Stress

Top Abstract I. Introduction II. Stress Concept III. Brain Regions Involved... IV. Methods Used for... V. Stressor Specificity of... VI. Stressor-Specific Activation... VII. Clinical Relevance of... References

 
Although the entire central nervous system is involved in the maintenance of internal homeostasis and participates in the organization of stress responses, some areas may have specific roles in these regulatory mechanisms. They are summarized briefly in this section.

Stressful stimuli may reach the central nervous system through somato- or viscerosensory pathways through spinal or brainstem sensory neurons (Fig. 3). Somatosensory signals are detected by noxious, mechanical, thermosensitive, etc., or specific (photic, acoustic, taste, equilibral) receptors and carried by spinal and cranial sensory nerves. Viscerosensory signals arise from the body and may reach spinal and supraspinal receptors by neural (from interoceptors) or humoral pathways. (Accordingly, stressful stimuli have been previously classified as neurogenic and systemic stressors.)

View larger version (27K): [in this window] [in a new window]   Figure 3. Neuronal circuits in the organization of stress responses. Horizontal thick and thin lines indicate "short circuit": autonomic (sympathoadrenal and/or parasympathetic) and defense (withdrawal) spinal reflexes in response to stressful stimuli. Thin lines represent "long circuit": ascending (afferent) and descending (efferent) neuronal loops between the spinal cord/medulla and "higher" brain centers. Dashed line indicates neurohumoral hypothalamo-pituitary outflow. CA, Brainstem catecholaminergic neurons.

A. Central autonomic system
Preganglionic neuronal cell groups constitute the output in the effector loop of stress responses. These cholinergic neurons in the medulla and the spinal cord are activated during almost all types of stress responses that influence sympathetic or parasympathetic outflows.

The parasympathetic preganglionic neurons are located in the medulla oblongata and (a minor portion) in the sacral spinal cord. In the medulla, they form distinct cell groups (dorsal motor vagal nucleus, superior and inferior salivatory nuclei). In addition, cells are arranged diffusely in the caudal part of the medulla and form a cellular arc between the dorsal vagal and ambiguus nuclei. The sympathetic preganglionic neurons form a longitudinal cell line in the thoracic (and first lumbar) spinal cord at the lateral portion of the gray matter, and this is referred to as the intermediolateral cell column (IML).

Signals to both types of preganglionic neurons arise through two strictly organized projections: short circuit (reflex) and long circuit (modulatory) neurons (Fig. 3). The short circuit afferents to the parasympathetic preganglionic neurons in the dorsal motor vagal nucleus arise from spinal or cranial sensory (both somato- and viscerosensory) neurons.

Some of the sensory signals (like respiratory) reach the preganglionic cell directly (monosynaptic reflex), but the vast majority of the inputs are relayed by sensory neurons in the nucleus of the solitary tract (NTS). The efferent fibers of the short circuit reach the ganglionic cells (in the vegetative ganglia or intramural ganglionic cells) through the vagal nerve. In addition to the inputs to the dorsal motor vagal nucleus, sensory signals ascend from the NTS to brainstem (parabrachial), hypothalamic, and limbic areas constituting the ascending loop of the long circuit (Fig. 3). In addition, noradrenergic and adrenergic neurons within and around the NTS (A2 noradrenergic and C2 adrenergic cell groups, respectively) receive stressful stimuli through sensory pathways innervating the NTS (63). The long circuit afferents to the dorsal motor vagal neurons (descending loop of the circuit) (Fig. 3) arise from the limbic, hypothalamic, and brainstem nuclei, partly directly, partly relayed by neurons in brain regions such as the lateral hypothalamus, the bed nucleus of the stria terminalis, the parabrachial nuclei, and the periaqueductal central gray (64).

The short circuit afferent fibers to the spinal sympathetic preganglionic neurons in the IML arise in dorsal root ganglion cells with a relay by dorsal horn interneurons. Preganglionic efferents leave the spinal cord through the ventral roots and terminate on sympathetic ganglionic cells located either in the peripheral sympathetic ganglia or in the innervated organs (intramural ganglionic cells). Descending fibers to the IML (long loop efferents) (Fig. 3) arise in limbic, hypothalamic, and brainstem nuclei (64).

From a functional point of view, central biogenic amine-containing neurons can be considered as a part of the central autonomic system. While biogenic amines are present in the peripheral autonomic system, central aminergic neurons represent a very specific "one-way" regulatory system. Having neuronal inputs through both somato- and viscerosensory fibers and feedback signals from hypothalamic, limbic, and cortical areas, aminergic neurons are unique projecting neurons with hundreds of axon collaterals and ten of thousands of axon terminals. All of their nerve endings terminate within the central nervous system and none of them project to the periphery. Therefore, their characteristics are briefly summarized in a separate subsection.

B. Central aminergic systems
Brain adrenergic, noradrenergic, and serotonergic neurons are involved in the central processing of stress responses. The role of dopaminergic neurons is controversial in this respect.

Brainstem catecholaminergic neurons receive direct somatosensory input from spinal cord and trigeminal sensory neurons as well as viscerosensory input from the NTS. Their activation is stressor specific: certain stressors, such as immobilization or pain-related stimuli, activate them rapidly and substantially, while others may have only minor influences.

1. Norepinephrine synthesizing neurons. Neurons in the ventrolateral and the dorsomedial medulla oblongata are the major sources of noradrenergic nerve terminals in the hypothalamus and the limbic system (65, 66, 67). In addition to these, noradrenergic cells in the locus coeruleus also contribute to the central organization of the stress response (67, 68). Lesions of brainstem catecholaminergic cell groups or their ascending fibers block or reduce stress-induced changes in the HPA axis (21).

a. The A1 noradrenergic cell group consists of the most caudal noradrenergic cells in the ventrolateral medulla. They are topographically arranged from the level of the medulla -spinal cord junction up to the level of the area postrema. Axons of these cells, comprising the ventral noradrenergic bundle, ascend to the forebrain and innervate mainly hypothalamic and limbic structures. The highest density of noradrenergic terminals is found in the parvocellular subdivision of the PVN that contains the majority of CRH-synthesizing neurons (67, 69).

b. A2 noradrenergic cells are present in the dorsomedial medulla, partly in the nucleus of the solitary tract (NTS), but a number of NE cells are dispersed into the neighboring nuclei. Ascending noradrenergic fibers from this cell group join the ventral noradrenergic bundle and participate in the noradrenergic innervation of the neuroendocrine hypothalamus (65, 66, 67, 69).

c. Locus coeruleus neurons increase their activity dramatically in response to certain stressful stimuli (70). The cerebral cortex, the cerebellum, and the basal ganglia are the major targets of these neurons, but they also participate in the noradrenergic innervation of the hypothalamus and the spinal cord (67, 71, 72). The locus coeruleus is involved 1) in the conduction of stress signals to forebrain areas, and 2) in the organization of stress responses.

ad 1) Locus coeruleus neurons receive stress signals through somato- and viscerosensory pathways via the spinoreticulothalamic tract. Noradrenergic fibers from the locus coeruleus innervate almost the entire forebrain including cortical, limbic, and hypothalamic structures.

ad 2) Noradrenergic axons from neurons located in the locus coeruleus and in the subcoeruleus area descend in the pontomedullary reticular formation and the lateral spinal funiculus and innervate the spinal cord. The spinal projection of the locus coeruleus has been demonstrated by retrograde tract tracing (73, 74, 75). Large multipolar cells in the ventral part of the locus coeruleus and the subcoeruleus area project to the spinal cord (76). Using a transneuronal viral labeling technique, these neurons have been demonstrated 3 d after injection of pseudorabies virus directly into the lateral- intermediate zones of the thoracic-lumbar spinal cord (Fig. 4).

View larger version (87K): [in this window] [in a new window]   Figure 4. Retrogradely labeled, pseudorabies-infected cells in the locus coeruleus (ventral portion) and the subcoerulear area 4 d after microinjection of the virus into the thoracic spinal cord (into the intermediolateral cell column and the intermediate zones). LC, Locus coeruleus; MT, mesencephalic trigeminal nucleus; RF, pontine reticular formation; SC, subcoeruleus area; IV, fourth ventricle. Bar scale: 100 µm.

2. Epinephrine synthesizing neurons. Adrenergic neurons are present in the middle portion of the ventrolateral medulla (between the A1 and A5 cell groups, rostrocaudally). A separate population of C1 neurons gives rise to a long ascending projection to the endocrine hypothalamus, while others project to the spinal cord to innervate sympathetic preganglionic neurons in the intermediolateral cell column (82, 83, 84, 85). The ascending fibers join the ventral noradrenergic bundle. In addition to C1 neurons, adrenergic neurons are also present in the dorsomedial medulla (C2 cell group) just rostral to the A2 noradrenergic cell group. Axons from these adrenergic neurons also join the ventral noradrenergic bundle and participate in the adrenergic innervation of the hypothalamus and the limbic system (84, 85, 86).

3. Serotonergic neurons. Serotonergic neurons are found in the lower brainstem (raphe nuclei) and in the hypothalamic dorsomedial nucleus. The rostral raphe nuclei (dorsal, midbrain and linear raphe nuclei) project to the hypothalamus and limbic regions (87, 88), while dorsomedial serotonergic neurons participate in the innervation of the pituitary gland (89, 90). Serotonergic neurons in the raphe magnus and raphe pallidus (rostral ventromedial medulla) project to the spinal cord (91, 92). TRH-, substance P-, and serotonin-synthesizing cells in the raphe obscurus (and probably in the raphe pallidus) innervate the dorsomedial medulla, including the dorsal motor vagal nucleus and the nucleus of the solitary tract (90).

Serotonergic neurons react sensitively to certain stressful stimuli (restraint, cold, pain) as has been demonstrated by increased c-fos activation. Especially, neurons in the raphe pallidus are very sensitive to immobilization stress and formalin-induced pain (refer to Section V). Despite a large number of studies, their contribution to the organization of stress responses is still not completely understood.

C. Noncatecholaminergic brainstem neurons
1. Medulla oblongata. The ventrolateral medulla contains stress-sensitive tyrosine hydroxylase-negative neurons. They are present in the lateral reticular and peritrigeminal nuclei. The latter neurons constitute the medullary thermosensitive area and respond to cold stress by rapid c-fos activation. In the dorsomedial medulla, NTS neurons are the principal recipients of first-order vagal and glossopharyngeal afferents, which carry viscerosensory signals (baroreceptor, respiratory, gastrointestinal, taste, etc.) to the central nervous system (93). In addition to catecholaminergic neurons (A2 and C2 cell groups), the NTS contains a variety of peptidergic neurons with hypothalamic and limbic projections (94, 95). In addition to these ascending fibers (they form the ascending loop of the "long circuit"), some of the NTS neurons serve as relay neurons that transfer viscerosensory signals directly to brainstem autonomic preganglionic ("short circuit") and catecholaminergic neurons (Fig. 3).

Cells in the ventromedial medulla (serotonin-, TRH- and substance P-containing neurons in the raphe magnus, paragigantocellular, and magnocellular reticular nuclei) project to the spinal cord, both to the dorsal horn and the intermediolateral cell column (92). They may not receive direct nociceptive signals; painful stimuli are carried from the spinal cord by the spinomesencephalic tract initially to the periaqueductal central gray. From here, enkephalin- and dynorphin-synthesizing neurons project down to the ventromedial medulla and disinhibit -aminobutyric acid (GABA)ergic interneurons. The activated serotonin-, TRH-, substance P-containing neurons innervate dorsal horn-inhibitory (mainly enkephalin-containing) interneurons that can block or reduce acute pain modulation (96).

2. Pons. Neurons in the parabrachial nuclei (medial, lateral, and Kölliker-Fuse nuclei) may serve important roles as intermediate stations that are modulated by both ascending and descending pathways. The lateral parabrachial nucleus is the main site for the relay of viscerosensory information from the NTS to the forebrain (97, 98, 99, 100, 101, 102). The parabrachial nuclei also receive direct neuronal information from the spinal cord and the spinal trigeminal nucleus (98).

3. Midbrain. Cell columns of the periaqueductal gray matter are involved in behavioral, autonomic, and antinociceptive changes. These neurons respond to several stressful stimuli with c-fos activation. Considerable data have been accumulated regarding central gray inhibition of pain through the activation of neurons in the rostral ventromedial medulla. Neurons in the lateral and ventrolateral cell columns of the periaqueductal gray matter project to the medullary parasympathetic preganglionic neurons (ambiguus and dorsal motor vagal nuclei) as well as to viscerosensory NTS neurons (103, 104). Other midbrain structures like the colliculi and the geniculate bodies may participate in the organization of responses to specific (optic, audiogenic) stressful stimuli.

D. Thalamus
The midline and intralaminar thalamic nuclei are strongly involved, especially in mammals, in nociceptive mechanisms (refer to the section on Pain for details). Fibers of the spinoreticulothalamic tract terminate here, and nociceptive signals transfer to limbic cortical areas (cingulate, piriform, entorhinal cortex). These neurons influence behavioral responses to certain stressful stimuli. The other sensory thalamic nucleus (ventral posterior thalamic nucleus), which receives nociceptive signals through spinothalamic, trigeminothalamic, and medial lemniscus fibers, represents the subcortical relay center for the discriminative and topographic recognition of sensory signals that terminate in the somatosensory cortex.

Neurons of the midline thalamic nuclei respond to stress with rapid c-fos activation (63, 105, 106, 107, 108). In various experimental conditions, even the mildest interventions, such as handling and control saline injections, may elicit c-fos activation in the midline thalamic nuclei. Thus, Fos positivity in these nuclei after exposure to various stressors should be assessed with caution.

E. Neuroendocrine hypothalamus
Almost all of the medial hypothalamic nuclei participate in the organization of responses to some stressors. The paraventricular, arcuate, and medial preoptic nuclei project to both the median eminence (neurohumoral output) and brainstem/spinal cord autonomic centers (neuronal output). Descending fibers may terminate on autonomic preganglionic neurons directly (64, 109, 110, 111, 112, 113, 114, 115) or they may exert their effect through brainstem (A5) catecholaminergic neurons (Fig. 5). Other nuclei, like the ventromedial, dorsomedial, perifornical, and supramamillary nuclei, contain stress- responsive neurons with mainly intrahypothalamic projections. The magnocellular neurosecretory neurons in the paraventricular, supraoptic, and accessory magnocellular nuclei are sensitive to stressors influencing body water and electrolyte homeostasis.

View larger version (27K): [in this window] [in a new window]   Figure 5. Stress-related hypothalamo-autonomic neuronal circuits. 1, Stressful stimuli; 2, spinal and medullary viscero- and somatosensory inputs to the hypothalamus; 3, spinal and medullary signals to the hypothalamus through brainstem catecholaminergic pathways; 4, direct hypothalamic projections to spinal and medullary autonomic centers; 5, hypothalamic projections to spinal and medullary autonomic centers through brainstem catecholaminergic neurons; 6, hypothalamic feedback to sensory neurons; 7, stress response to the periphery through the central autonomic system. 1. 8. 7., Short neuronal (periphery-spinal/brain stem-periphery) circuit in response to stressful stimuli.

The lateral hypothalamus may be viewed as comprising a combination of several ascending and descending fibers between the medial hypothalamus, the limbic system, and the autonomic nervous system with thousands of interneurons. Almost all of the stress-conducting fibers enter the hypothalamus in this lateral area. These fibers and a high percentage of medial hypothalamic afferent and efferent fibers are relayed here.

F. Limbic system
Both cortical and subcortical limbic structures are involved in the organization of stress responses. The subcortical areas include the amygdala, septum, habenula, and related structures, while the limbic cortex consists of the hippocampal formation (hippocampus, dentate gyrus, subiculum) and entorhinal, piriform, prelimbic, intralimbic, and cingulate cortices (119, 120).

A great variety of behavioral responses to stress are organized by the limbic system. Accordingly, limbic areas receive neuronal input from brainstem and spinal viscero- and somatosensory neurons (ascending loop of the "long circuit"), and they project to brainstem and spinal autonomic preganglionic neurons (descending loop in the "long circuit"; Fig. 3). With neuronal projections to the hypothalamus, limbic areas may influence the activity of the neuroendocrine hypothalamo-pituitary system (69, 120, 121).

The central nucleus of the amygdala occupies a special position in the organization of stress responses. This nucleus, with four subdivisions, contains various types of peptidergic (CRH, somatostatin, neurotensin, enkephalin, galanin) neurons and receptors (122, 123). They receive brainstem and hypothalamic inputs and project back to these regions directly or through the bed nucleus of the stria terminalis (69, 124). The hypothalamic targets are the PVN and cells within and between the arcuate and ventromedial nuclei (Fig. 6). A significant portion of the brainstem/spinal cord projections of the central amygdala and the bed nucleus of the stria terminalis are relayed by the parabrachial nuclei (Fig. 6).

View larger version (24K): [in this window] [in a new window]   Figure 6. Projections from the central nucleus of the amygdala to the hypothalamus and to the central autonomic system. AN, Arcuate nucleus; HPA, hypothalamo-pituitary-adrenocortical axis; NIST, bed nucleus of the stria terminalis; VM, ventromedial nucleus.

Limbic cortical regions are sensitive to stress, especially if the stressor exceeds a noxious threshold. These regions are neuronally connected (directly or through entorhinal neurons) with the hippocampus and are responsible for stress-related motivational and behavioral responses (57).

	IV. Methods Used for Mapping Stressor-Specific Neuronal Circuits

Top Abstract I. Introduction II. Stress Concept III. Brain Regions Involved... IV. Methods Used for... V. Stressor Specificity of... VI. Stressor-Specific Activation... VII. Clinical Relevance of... References

 
Several approaches have been introduced to localize and characterize brain structures and neuronal pathways that are involved in the organization of stress responses. Among them, intracerebral microdialysis, immunohistochemistry and in situ hybridization for neurotransmitters, neuropeptides and protooncogenes, and tract-tracing techniques are very useful tools, especially when combined with experimental brain surgery (lesioning of brain nuclei, transection of neuronal pathways). Here, we discuss two powerful techniques: intracerebral microdialysis and c-fos immunohistochemistry, in detail.

A. Intracerebral microdialysis
Initially, the participation of various brain regions in stress-induced neuroendocrine responses were studied by measuring tissue concentrations of neurotransmitters or related substances. When lower tissue concentrations of a neurocompound of interest were found in a particular brain area, it was assumed that activation of that area had occurred, and the area was considered to be part of a stressor-specific anatomical and functional circuit. However, a comprehensive understanding of neuronal regulation required the development of approaches for assessing simultaneously the rate of delivery of various neurotransmitters into the synaptic cleft and the magnitude of receptor-mediated postsynaptic responses. It was assumed that a positive correlation would exist, at least in acute stress responses, between neurotransmitter release, its synaptic cleft concentrations, and the activation of an effector system. Much attention was also paid to developing in vivo methods that would be applicable in awake animals. Therefore, microdialysis as a new in vivo method for manipulating and monitoring neurotransmitter release and inactivation, as well as for evaluating receptor-mediated biochemical effects, was introduced (125, 126, 127, 128, 129).

The microdialysis technique makes use of a simple principle. The dialysis membrane is permeable to water and solutes below a certain molecular mass. Perfusion of the microdialysis probe with artificial cerebrospinal fluid or a solution containing a drug of interest creates a concentration gradient across the membrane, causing diffusion of substances across the membrane. In particular, catecholamines, in the extracellular fluid space, diffuse across the dialysis membrane and enter the perfusate, assays of which can then reflect extracellular fluid concentrations of the endogenous compounds. Liquid chromatographic-electrochemical assays of substances in the microdialysate require only small amounts of material, and the samples, under some circumstances, can be assayed directly because the dialysate lacks protein. Microdialysis may also be combined with other techniques such as lesions, local chemical stimulation, pharmacological interventions, and anatomical evaluations, or localized delivery of drugs to specific brain regions, enhancing the value of microdialysis as a tool in experimental neuroendocrinology, neurology, and pharmacology. Our group has applied microdialysis in several studies relating to catecholaminergic innervation of the PVN and stressor-specific NE release in the PVN and its relationship to activation of the HPA axis (24, 25, 130, 131, 132). We introduced and described the usefulness of microdialysis in small brain areas such as the central nucleus of the amygdala and the bed nucleus of the stria terminalis (133). We also introduced simultaneous measurements of extracellular fluid levels of NE and its metabolites in various brain regions to provide a comprehensive assessment of synthesis, turnover, release, metabolism, and uptake of monoaminergic neurotransmitters under basal and pathophysiological conditions (24). In all of our studies, we performed our microdialysis experiments no earlier than 20 h after a probe was inserted to minimize factors that could affect neurotransmitter release into extracellular fluid (e.g., surgical stress, acute tissue damage after probe insertion, etc.).

B. Protooncogene-"immediate early genes"- immunohistochemistry
With an ever-expanding knowledge of brain function, attempts have been made to study the activity of individual neuroendocrine cells. Immediate early genes such as c-fos, c-jun, jun D, or zif268 represent one promising avenue of research. These genes are expressed immediately in response to appropriate extracellular stimuli and then may play important roles in signal transduction and transcriptional regulation in normal cells. Thus, the expression of various immediate early genes in particular neurons is proposed to correlate with their functional activation, and their activation is followed by the production of cell-specific neuroactive substances (28 28A, 107, 134, 135).

c-fos Has been the most frequently used immediate early gene. The pattern of brain activation in response to acute stressful stimuli is examined by using Fos immunohistochemistry (immunostaining the Fos protein product) or in situ hybridization (expression of c-fos) as markers for neuronal activity. After an appropriate stimulus, c-fos expression occurs rapidly, usually within a few minutes, with a peak response within 30 min from the time of the initiation of stress. Fos protein is detectable by immunohistochemistry somewhat later, with maximal levels at 60–90 min after the stressor. The synthesized Fos protein has a half-life of about 2 h.

Initially, it was proposed that responses of immediate early genes to different stressors were rather similar and stereotyped but an increasing number of later studies have clearly demonstrated stressor specificity of c-fos responses in different brain regions (for review, see Refs. 19, 28, 107 , and 135).

As with other methods, detection of Fos immunoreactivity has some limitations. The identification of Fos in individual neurons can be used for functional anatomical mapping with one caveat, that an absence of c-fos induction does not necessarily indicate a lack of neuronal activity (134). The absence of Fos may indicate 1) that a population of neurons does not express c-fos, 2) some other immediate early genes and their products are responsible for neuronal stimulation, 3) the signal at the cell body may be insufficient to induce c-fos expression, 4) the thresholds for c-fos induction may differ in different neurons, or 5) the activating transmitter or the second messenger needed to induce c-fos expression was not present, functioned abnormally, or was bypassed (134). In contrast, the appearance of c-fos mRNA is not necessarily followed by the production of Fos protein. Another situation in which c-fos activation is dissociated from neuronal firing is the presence of a persistent or recurring stimulus. Sustained expression of c-fos in response to long-lasting stimuli has been observed in a variety of stressful conditions. Chronic stress may cause persistent stimulation of c-fos gene expression, or neurons of certain brain nuclei or regions are activated alternatively. Whether this is due to activation of other intracellular mechanisms or to the intensity of the stimulus remains to be determined.

Recently, NGFI-B (nerve growth factor, type 1-B), also known as nur77, N10, and TISI in mice, has also been used to map neuronal activation at multiple levels of stress-related neuroendocrine circuitry (for review, see Ref. 135). Stressor-specific induction of NGFI-B and c-fos expression established NGFI-B as a useful alternative or adjunct marker to c-fos for revealing neuronal activation in the neuroendocrine hypothalamus.

	V. Stressor Specificity of Central Neuroendocrine Responses

Top Abstract I. Introduction II. Stress Concept III. Brain Regions Involved... IV. Methods Used for... V. Stressor Specificity of... VI. Stressor-Specific Activation... VII. Clinical Relevance of... References

 
Various stressors elicit different neural and endocrine responses. We analyzed five different stressors under acute conditions by recording such parameters as plasma ACTH, corticosterone, NE, and EPI levels and extracellular NE levels in the PVN (and in some experiments in limbic nuclei) and investigated c-fos activation by Fos immunostaining, and CRH mRNA expression by in situ hybridization. Thus, we were able to collect enough information to propose the possible routes and targets of these stressors in the central nervous system. We found that stressors differ not only in their evoked responses but also in their neuronal circuits. It should be pointed out, however, that each stressor may activate several brain structures, and the proposed pathways are the most probable but not the exclusive ones.

A. Immobilization stress
Hans Selye was the first researcher who used immobilization stress, which led in rats to the manifestation of his stress syndrome, i.e., adrenal hypertrophy, gastric ulceration, and thymicolymphatic involution (136). Selye’s original restraint procedure involved tying a rat’s legs together and wrapping the rat tightly in a towel.

All types of restraint stress should be viewed as a mixture of physical and psychological stressors, including decreased body temperature and pain stress as important components of some restraint procedures. Thus, immobilization stress-induced patterns of activation of various stress effector systems result from restraint, pain stress, and changes in body temperature.

The maximal responses of the stress effector systems are usually seen within the first 30 min after the beginning of immobilization stress. The magnitude of central stress responses usually diminishes upon exposure to chronic intermittent immobilization stress, most likely reflecting habituation as well as exhaustion of stress effector systems. This is well documented by continuous basal or stress-induced decreases in NE release into extracellular fluid in the PVN, most likely due to maximal NE release that is not matched by its ongoing stimulated synthesis in rats exposed to chronic immobilization stress (132, 137).

In our studies, immobilization stress consisted of taping each rat’s limbs to a metal frame with each rat kept in a prone position. Our data are in substantial agreement with a number of previous reports of c-fos activation in brain nuclei after various types of immobilization or restraint stress.

1. c-fos Expression in the central nervous system after immobilization stress. Strong c-fos activation are observed in several brain regions 30–120 min after immobilization stress, indicating that numerous systems were influenced by this stressor (Table 1).

a. Central catecholaminergic system. Induction of c-fos expression in response to immobilization stress is evident within brainstem catecholaminergic cell groups previously shown to play a role in stress-induced activation of the HPA axis (Table 1). Two hours of immobilization stress induces a strong Fos-like immunostaining in the A1, A2, A5, A6 (locus coeruleus), and A7 noradrenergic cell groups (26, 29, 63). Almost all of the tyrosine hydroxylase-positive cells in the ventrolateral medulla and the locus coeruleus are also Fos positive. In contrast to these, only 40% of tyrosine hydroxylase-positive cells (A2 and C2 cell groups) showed Fos immunopositivity in the dorsomedial medulla (Fig. 7). The double-labeled neurons are mainly in the commissural part of the NTS. Tyrosine hydroxylase-positive dopaminergic neurons in the higher and lower brainstem all remain Fos negative after immobilization stress. These included cells in the substantia nigra and the ventral tegmental area (A8, A9, and A10 cell groups), the A11 cell group in the posterior hypothalamus, the A12 cell group in the arcuate nucleus, the A13 cell group in the zona incerta, and the A14 cell group in the preoptic and hypothalamic periventricular nuclei. Acute immobilization stress-induced Fos expression in several brainstem catecholaminergic areas and various brain nuclei has been reported by others as well (19, 28, 106, 107, 138, 139).

View larger version (74K): [in this window] [in a new window]   Figure 7. Fos immunoreactive cell nuclei in tyrosine hydroxylase-positive A2 and tyrosine-hydroxylase-negative neurons in the nucleus of the solitary tract after 3 h of immobilization stress in rats. Note: a number of tyrosine hydroxylase-positive neurons contain Fos-immunonegative nuclei. ST, Solitary tract. Bar scale, 50 µm.

Immobilization stress also activates noncatecholaminergic neurons in the dorsomedial medulla (26, 28, 63, 107, 139). A fairly high percentage of tyrosine hydroxylase-negative, neuropetidergic neurons are activated in the NTS after immobilization stress (Fig. 7). These neurons project directly to the hypothalamus and the limbic system (94, 95). Whether Fos immunoreactivity in the paratrigeminal and parabrachial nuclei, including the Kölliker-Fuse nucleus, after immobilization stress is dependent on somatosensory or unconditioned aversive stimuli is unknown.

c. Thalamus. A widespread pattern of Fos immunoreactivity and c-fos mRNA expression (26, 138) are detected in the midline thalamic nuclei 60–120 min after immobilization stress. Like after noxious stimuli (26, 28, 105, 107), the central medial, paraventricular, rhomboid, anterodorsal, reuniens, and intermediodorsal thalamic nuclei, as well as the medial subdivision of the lateral habenula, exhibit marked c-fos activation in response to immobilization stress.

d. Hypothalamus. Strong, bilateral Fos immunoreactivity was found in noncatecholaminergic neurons in the hypothalamus after immobilization stress (Table 1). Thirty minutes after immobilization stress, cells in the parvocellular PVN (Fig. 8A) and in the dorsal part of the supraoptic nucleus, most likely oxytocin-containing cells, show c-fos activation. However, 2 to 3 h after immobilization stress, pronounced Fos immunoreactivity can be observed in the entire supraoptic nucleus. Thus, it is possible that alterations in salt and water balance as well as changes in blood volume could contribute to these time-related changes in vasopressin- synthesizing supraoptic neurons. Similar to our results, Miyata et al. (143) also described increases in Fos immunoreactivity in oxytocin-containing neurons in the PVN and supraoptic nucleus. Cells in the magnocellular subdivision of the PVN and vasopressin-containing cells in the supraoptic nucleus remained Fos negative.

View larger version (105K): [in this window] [in a new window]   Figure 8. Fos-immunoreactive cell nuclei of hypothalamic paraventricular neurons. A, After 3 h of immobilization stress; B, 60 min after a 4% formalin injection (subcutaneous) into the hind paw; C, 60 min after withdrawing 25% of the estimated blood volume (hypotensive hemorrhage), and D, 90 min after administration of 3.0 IU/kg insulin (acute hypoglycemia). m, Magnocellular paraventricular subdivision; pm, medial parvocellular paraventricular subdivision; V, third ventricle. Bar scales, 100 µm.

In the preoptic area, dense populations of Fos-positive cells can be located in the ventral and commissural subdivisions of the medial preoptic nucleus but not in its central part or in the lateral and periventricular preoptic areas. In the anterior hypothalamus, Fos-positive cells occupied the ventral subdivision of the anterior hypothalamic nucleus and the lateral hypothalamic area. Fos-immunopositive but not tyrosine hydroxylase-positive cells were seen in the dorsomedial and arcuate nuclei. The ventromedial nucleus and nuclei in the premamillary region were devoid of Fos immunoreactivity (63).

A prominent increase in c-fos mRNA expression was detected in the medial parvocellular subdivision and to a much lesser extent in the magnocellular subdivision of the PVN, as well as in the dorsomedial, arcuate, supramamillary, and posterior hypothalamic nuclei, and in the medial preoptic area (138). c-fos mRNA levels generally peaked at 30 min after stress, and by 120 min were markedly reduced or had returned to basal levels (138).

e. Limbic system. After three hours of immobilization stress, Fos immunoreactivity was found in the lateral subdivision of the central amygdaloid nucleus and in the medial amygdaloid nucleus. CRH-immunopositive cells in the intermediate subdivision of the central nucleus of the amygdala failed to show c-fos activation in response to immobilization stress. Arnold et al. (144) and Chen and Herbert (139) did not find c-fos activation in the central nucleus of the amygdala after 15–60 min of restraint stress. Similarly, Senba and Ueyama (28) and Arnold et al. (144) did not report any changes in c-fos activation in the central nucleus of the amygdala after 15–120 min of immobilization stress. Moderate numbers of immobilization stress-induced Fos-immunopositive neurons were demonstrated in the bed nucleus of the stria terminalis, mainly in its dorsolateral subdivision, which contains CRH cell bodies. Strong c-fos activation was observed in limbic cortical areas (piriform, cingulate, and entorhinal cortices) after immobilization stress (26, 63). Marked c-fos mRNA expression was also detected in the bed nucleus of the stria terminalis, medial and central amygdaloid nuclei, the lateral septal nucleus, medial preoptic area, and the cingulate cortex (138). A marked increase in c-fos mRNA expression was detected in the bed nucleus of the stria terminalis, medial and central amygdaloid nuclei, the lateral septal nucleus, and the cingulate cortex 30 min after immobilization stress (138).

f. Cerebral cortex. Neurons in the parietal somatosensory cortex, mainly in layers 3–5, showed c-fos activation 30–120 min after immobilization stress (26, 138).

g. Activation of other immediate early genes by immobilization stress. In an extensive study by Cullinan et al. (138) c-jun and zif/268 mRNA expression was studied after exposure of rats to restraint stress (30 min in plastic cylinders).

Induction of c-jun mRNA after immobilization stress was detected in the NTS, the medial preoptic area, the dorsal premamillary nucleus, the ventral subdivision of the lateral septal nucleus, superficial layers of the parietal cortex, and the infralimbic cortex (138). Prominent expression of transcripts encoding zif/268 was noted in the parvocellular subdivision of the PVN, the anterior and lateral hypothalamic areas, and the lateral septal nucleus. Moderate increases in zif/268 mRNA expression were found in the dorsomedial nucleus, the premamillary region, the supramamillary nucleus, and the zona incerta. Relatively minor changes were detected in the brainstem. Thus, as is evident from this study, immobilization stress-induced activation of various brain regions included those regions that participate in the activation of the HPA axis as well as in its inhibition to maintain optimal glucocorticoid levels to "repair" disturbed homeostasis. In terms of HPA axis inhibition, GABA-ergic projections from the anterior hypothalamic area, the medial portions of the bed nucleus of the stria terminalis, and the retrochiasmatic area were proposed to participate in the inhibition of CRH neurons in the PVN (138,