sign in sign up
<<
Home , Hippocampus

Molecular Psychiatry (1997) 2, 255–262  1997 Stockton Press All rights reserved 1359–4184/97 $12.00

SEMINARS IN MOLECULAR PSYCHIATRY AT THE CLINICAL NEUROENDOCRINOLOGY BRANCH, NIMH, NIH

Possible mechanisms for atrophy of the hippocampus BS McEwen

Harold and Margaret Milliken Hatch Laboratory of Neuroendocrinology, The Rockefeller University, New York, NY 10021, USA

Recently published work using MRI to image the human has revealed that the hippocam- pal formation undergoes a selective atrophy in diverse conditions such as Cushing’s syn- drome, post-traumatic stress disorder, recurrent depressive illness, normal aging preceding and in Alzheimer’s disease. Hippocampal shrinkage is usually accompanied by deficits in declarative, episodic, spatial and contextual performance and the hippo- campal changes provide a neural substrate for changes in cognitive function that have been recognized to accompany these various conditions. The hippocampus has long been known as a target of stress hormones, and it is an especially plastic and vulnerable region of the brain. However, the prominence of the hippocampus as a target has obscured the fact that other factors besides glucocorticoid hormones are involved in the process of hippocampal atrophy. Excitatory amino acids and NMDA receptors are prominent in their involvement in an animal model of hippocampal atrophy as well as in neuronal death. Further- more, the finding of hippocampal atrophy does not necessarily imply a permanent loss of cells, and this aspect deserves careful investigation, both to analyze the underlying anatom- ical changes and to investigate the possibility of pharmacological treatment to reverse the process. In cases where atrophy is due to cell loss, the time course of the disease process will provide much useful information about mechanism and offer the possibility of early inter- vention to arrest or slow the pathological process. Keywords: hippocampus; atrophy; cell loss; stress; ; NMDA receptors

It has been known for some time that the areas, in particular in Cushing’s syndrome, recurrent shows signs of atrophy as a result of elevated glucocort- depressive illness, post-traumatic stress disorder, icoids and severe, traumatic stress (eg holocaust and aging prior to overt dementia survivors).1 However, it has only been very recently (Table 1). The diversity of conditions in which atrophy that brain imaging techniques have allowed for a regional analysis of the atrophy or shrinking of various brain structures to see which ones are most affected. Table 1 Conditions involving atrophy of the human hippocampus Recent evidence indicates that the human hippocam- pal formation is particularly sensitive in this respect Condition References and tends to show greater changes than other brain Recurrent depressive illness 64 PTSD 65, 66 The section Seminars in Molecular Psychiatry at the Clinical Aging, preceding dementia 56, 57 Neuroendocrinology Branch, NIMH is jointly edited by Philip W Dementia 67 Gold, Julio Licinio, Esther Sternberg and Ma-Li Wong. Presented Cushing’s syndrome 68 on 29 April 1997. Schizophrenia 69, 70 Correspondence: BS McEwen, PhD, The Rockefeller University, Box 165, 1230 York Avenue, New York, NY 10021, USA. E-mail: mcewenȰrockvax.rockefeller.edu For another recent summary of hippocampal Received and accepted 7 January 1997 atrophy in the human brain, see Ref. 71. Atrophy of the human hippocampus BS McEwen 256 occurs raises the question whether they reflect a com- impairment in the learning of spatial and short-term mon mechanism and secondly whether the atrophy is memory tasks.4 permanent or reversible. Besides the effect of stress on hippocampal structure, Regarding mechanism, it is tempting to attribute the there are other forms of plasticity in the hippocampus, occurrence of hippocampal atrophy to glucocorticoids. including reversible that is regulated This is because the hippocampus is a primary target by ovarian steroids and excitatory amino acids via area for adrenal steroids in the brain, and adrenal ster- NMDA receptors in female rats and occurs in the CA1 oids have been shown to have effects on hippocampal region21,22 and a reversible atrophy of dendrites of CA3 neuronal plasticity and on the loss of hippocampal during hibernation in ground squirrels and neurons in conditions like and aging.1–4 How- hamsters.23,24 The estrogen-regulated CA1 synaptic ever, other factors play a role, including the endogen- plasticity is a rapid event, occurring during the female ous excitatory amino acid , and one rats′ 5-day estrous cycle, with the taking sev- of the purposes of this short overview will be to sum- eral days to be induced under the influence of estro- marize what we know about the mechanism of neu- gens and endogenous , and then disap- ronal atrophy in several animal models. Another goal pearing within 12 h under the influence of the of this article is to discuss the relationship between proestrus surge of progesterone.21 In contrast, the CA3 reversible atrophy and loss of permanent cells, includ- atrophy found in rats and noted in the preceding para- ing the possibility of therapeutic strategies for revers- graph is a relatively slow process taking normally at ing atrophy and preventing permanent damage. This least 3 weeks to develop under daily stress and a week issue has recently been discussed in a News & Views or so to disappear. However, dendritic atrophy in article in this journal.5 hibernating ground squirrels and hamsters develops as fast as the hibernating state and can be reversed rapidly Hippocampal neuronal plasticity in several animal within several hours (Magarinos, McEwen and Pevet, models unpublished).23,24 Although anatomically similar to the The is an important brain stress-induced atrophy in rats and tree shews, it is not structure in episodic and spatial learning as well as a yet clear if this process involves the same mechanisms; component of the control of a variety of vegetative however, if this is the case, the question becomes what functions such as ACTH secretion.6,7 It is also a plastic factors make the atrophy rapid in hibernation and slow and vulnerable brain structure that is damaged by in relation to repeated stress. stroke and head trauma and susceptible to damage dur- ing aging and repeated stress.1 In 1968, we showed that Animal models of reversible dendritic atrophy hippocampal neurons express receptors for circulating The detailed study of the process of dendritic atrophy adrenal steroids,8 and subsequent work in many lab- in rats and tree shrews is a useful model for under- oratories has shown that the hippocampus has two standing cellular mechanisms and pharmacological types of adrenal steroid receptors, Type I means of intervening and either blocking or reversing (mineralocorticoid) and Type II (glucocorticoid) which hippocampal atrophy. In the hibernation and stress mediate a variety of effects on neuronal excitability, studies, dendritic length and branching is assessed by neurochemistry and structural plasticity.9 morphometry after silver staining neurons with the sin- Recent work in our laboratory has shown that adre- gle-section Golgi technique. More recently, electron nal steroids are involved in three types of plasticity in microscopy has revealed that stress and glucocorti- the hippocampal formation. First, they reversibly and coids alter morphology of presynaptic mossy fiber ter- biphasically modulate excitability of hippocampal minals in the stratum lucidum region of CA3.25 neurons and influence the magnitude of long-term Initially, atrophy of apical dendrites of CA3 pyramidal potentiation, as well as producing long-term neurons, and not dentate granule neurons or CA1 pyr- depression.10–13 These effects may be involved in amidal neurons, was found after 21 days of daily biphasic effects of adrenal secretion on excitability and corticosterone (CORT) exposure and also after 21 days cognitive function and memory during the diurnal of 6 h per day repeated restraint stress.4 Psychosocial rhythm and after stress.14–17 Second, adrenal steroids stress also causes apical dendrites of CA3 pyramidal participate along with excitatory amino acids in reg- neurons to atrophy in rats26 and in an insectivore, the ulating of granule neurons, tree shrew.20 Stress- and CORT-induced atrophy were in which acute stressful experiences can suppress the prevented by the anti-epileptic drug, phenytoin ongoing neurogenesis.18 We believe that these effects (Dilantin), thus implicating the release and actions of may be involved in -related learning and memory, excitatory amino acids, since phenytoin blocks gluta- because of the anatomical connections between the mate release and antagonizes sodium channels and dentate gyrus and the , a brain area important possibly also T-type calcium channels that are acti- in memory of aversive and fear-producing experi- vated during glutamate-induced excitation; this result ences.19 Third, adrenal steroids participate along with was consistent with evidence that stress induces excitatory amino acids in a reversible stress-induced release of glutamate in hippocampus and other brain atrophy of dendrites in the CA3 region of hippocampus regions.4,20,27 Recent work has shown that NMDA of male rats4 and tree shrews,20 a process that affects receptor blockade is also effective in preventing stress- only the apical dendrites and results in cognitive induced dendritic atrophy.4 Atrophy of the human hippocampus BS McEwen 257 A model of the cellular and neurochemical interac- corticosterone secretion,30 but may instead be related tions involved in dendritic atrophy is presented in to its reported effects in enhancing the reuptake of Figure 1, and it emphasizes the interactions between within the hippocampus.31 Further evidence neurons and neurotransmitters. The role of adrenal for serotonin involvement in dendritic atrophy comes steroids will be discussed below. Besides glutamate, from studies of psychosocial stress in rats, in that both other participating neurotransmitters include GABA dominant and subordinate rats show dendritic atrophy and serotonin. Inhibitory have a signifi- as well as down-regulation of 5HT transporter cant role in controlling hippocampal neuronal excit- expression in the CA3 region, indicating either a ability,28 and involvement of the GABA-benzodiazep- reduced density of serotonin terminals or a reduced ine receptor system is strongly suggested by the ability expression of the transporter.26 However, dominant of a benzodiazepine, adinazolam, to block dendritic rats show a greater reduction in 5HT transporter site atrophy (Magarinos, unpublished). As for serotonin, than subordinates, and dominants also show a greater repeated restraint stress in rats and psychosocial stress dendritic atrophy.26 cause changes in the hippocampal formation that Because corticosterone- and stress-induced atrophy include not only atrophy of dendrites of CA3 pyrami- of CA3 pyramidal neurons are both blocked by pheny- dal neurons but also suppression of 5-HT1A receptor toin as well as by tianeptine,4 serotonin released by binding.4,26 stress or corticosterone may interact pre- or post-synap- Serotonin is released by stressors and tianeptine, an tically with glutamate released by stress or corticos- atypical tricyclic antidepressant that enhances sero- terone, and the final common path may involve inter- tonin reuptake and thus reduces extracellular 5HT lev- active effects between serotonin and glutamate els, prevented both stress- and corticosterone-induced receptors on the dendrites of CA3 neurons innervated dendritic atrophy of CA3 pyramidal neurons,29 by mossy fibers from the dentate gyrus. There is evi- whereas several inhibitors of serotonin reuptake, dence of interactions between serotonin and NMDA fluoxetine and fluvoxamine, and desipramine, an receptors, indicating that serotonin potentiates NMDA inhibitor of noradrenaline uptake, failed to block atro- receptor binding as well as activity of NMDA receptors 32,33 phy (Magarinos, Watanabe, unpublished). Thus the and may do so via 5-HT2 receptors. effect of tianeptine on CA3 pyramidal mor- Glucocorticoid treatment causes dendritic atrophy, phology is not due to its reported effects in reducing and stress-induced atrophy is blocked by treatment

Figure 1 Diagram of the hippocampal CA3 region showing schematically the innervation of apical dendrites of CA3 pyramidal neurons by mossy fiber projections from dentate granule neurons. Granule neurons also send mossy fibers to interneurons in the hilus, which, in turn, send inhibitory projections to the CA3 pyramidal neurons. The balance between the excitatory input and the inhibitory tone from the interneurons is presumed to be very important to the excitability of CA3 neurons. Evidence summarized in the text indicates that excitatory amino acid release during repeated stress, aided by circulating glucocorticoids, leads to a reversible atrophy of apical dendrites over 3–4 weeks in rats and tree shrews. Serotonin also participates, possibly by aiding the excitatory amino acid activity at the NMDA receptor, and reduced GABA-benzodiazepine-mediated inhibitory activity at from the interneurons on CA3 pyramidal neurons may also exacerbate the atrophy. Prolonged psychosocial stress in is known to cause a permanent loss of CA3 pyramidal neurons, which may reflect another and more dangerous stage of the actions of the mediators that cause the atrophy shown in the diagram. We envision that the process that is so intense in the CA3 region occurs to some extent throughout the hippocampus, which might help explain why the entire human hippocampus undergoes atrophy of 10–15% in various conditions described in the text and Table 1. Atrophy of the human hippocampus BS McEwen 258 with an adrenal steroid synthesis blocker, cyano- Following upon the widespread activation of NMDA ketone.4 What is the role of adrenal steroids in relation receptors, the increased levels of intracellular calcium to the involvement cited above? A may make the dendritic cytoskeleton become depoly- primary site of action is the release of glutamate, since merized or undergo proteolysis.4 Stress is also reported adrenalectomy markedly reduces the magnitude of the to alter the expression of the , BDNF and EAA release evoked by restraint stress.34 In this con- NT-3, in the hippocampus.45 Very little is known about nection, mossy fiber terminals from dentate granule the adrenal steroid receptor types involved in these neurons in the stratum lucidum of CA3 apical den- effects, or the localization of stress and adrenal steroid drites show morphological alterations as a result of effects on expression within the hippo- .25 These changes involve reorganization campus and their relationship to the conditions of of synaptic vesicles in the synaptic terminal, with a repeated stress which bring about morphological higher density occurring in regions adjacent to active changes. However, conditions which cause dendritic synaptic zones. atrophy, such as repeated restraint stress or psychoso- The stratum lucidum zone of CA3 contains high lev- cial stress, do not appear to change neurotrophin els of kainate receptors, as demonstrated by quantitat- expression in hippocampus (Kuroda, unpublished), ive autoradiography, and, as noted above, these recep- indicating that neurotrophins are probably not directly tors are decreased in density by ADX and restored to involved in the mechanism of dendritic atrophy. This normal by corticosterone replacement.35 Because kain- does not exclude the possibility that neurotrophin ate receptors are feed-forward autoreceptors for excit- depletion or suppression might be involved in perma- atory amino acids on presynaptic mossy fiber nerve nent neuronal loss resulting from more severe and pro- endings (Figure 1), the effects of adrenal steroids are longed stress, as described, for example, by Uno et al.46 consistent with the dependence of stress-induced glut- amate release on the presence of the adrenal glands.34 What is the functional significance of impaired Another possibility is that corticosterone or stress hippocampal function on stress and aging? alter CA3 neuronal atrophy through regulation of Stress and glucocorticoids are known to have specific GABAergic synaptic inhibition (Figure 1). In support of effects on cognitive function in and in animal this notion, low levels of corticosterone alter mRNA models. Adrenal steroids and stressful experiences levels for specific subunits of GABAa receptors in CA3 produce short-term and reversible deficits upon epi- and the dentate gyrus of adrenalectomized rats,36 sodic and in animal models and in whereas stress levels of corticosterone have produced humans,47 whereas repeated stress also impairs cogni- different effects on GABAa receptor subunit mRNA tive function in animal models and glucocorticoid levels and receptor binding in hippocampal subregions elevation or treatment in humans is accompanied by including CA3 (Orchinik, Weiland and McEwen, cognitive dysfunction.48 There are also declines in cog- unpublished). Therefore, it appears that corticosterone nitive function in aging humans that are correlated may alter the excitability of hippocampal neurons with progressive elevations in HPA activity over 3–4 through regulation of GABAa receptor expression, but years.49,50 it remains to be seen if the corticosteroid effects on Acute effects of stress or glucocorticoid adminis- neuronal morphology involve changes to the number tration are evident within a time span ranging from a or pharmacological properties of GABAa receptors. few hours to a day and are generally reversible and Both stress and glucocorticoid treatment cause quite selective to the task or particular situation.47,51 enhanced expression of NMDA receptors in hippocam- Adrenal steroid effects are implicated in both selective pus,37,38 and this is another potential mechanism by attention as well as in ,47 and which adrenal steroids are involved in dendritic atro- such actions are consistent with the effects of adrenal phy (Figure 1). It is puzzling, however, that NMDA steroids on the modulation of long-term potentiation receptors are not expressed at very high levels in the and primed-burst potentiation (see above). However, stratum lucidum where mossy fibers terminate39, given some acute actions of stress may involve other mech- the evidence cited above for the importance of this anisms than glucocorticoids, including endogenous innervation for dendritic atrophy. The presence of opioid in the case of painful stressors NMDA receptors on the more distal aspects of CA3 like shock (summarized in Ref. 4). With regard to non- dendrites suggest that the mossy fiber activation of glu- painful stressors, exposure of rats to a novel environ- tamate release triggers a much more widespread ment resulted in a rapid and reversible impairment of activity of excitatory amino acids affecting the entire plasticity in vivo in the CA1 region and this effect may dendritic tree of the CA3 pyramidal neurons. Indeed, involve the actions of glucocorticoids.17 the and physiology of CA3 neurons is Longer-term stress that produces dendritic atrophy consistent with a recurrent excitation. Neuroanatom- in the CA3 region of the hippocampus has also been ically, CA3 neurons send recurrent axonal projections shown to impair hippocampal-dependent learning. A to other CA3 neurons as well as back to the hilus;40,41 recent study has demonstrated an impairment of per- and physiologically, CA3 neurons are characterized by formance on an 8-arm radial maze in rats that received burst firing42,43 and have a multiplicity of calcium 21 days of restraint stress prior to maze training.52 The channel types that contribute to the activation of cal- stress effect is reversible, in that inhibition of the initial cium currents by low voltage changes.44 learning was evident when rats were trained and evalu- Atrophy of the human hippocampus BS McEwen 259 ated immediately after the end of stress, but not 18 transmitters. Nevertheless the role of glucocorticoids days after the termination of stress. This impairment should not be ignored. Glucocorticoids are elevated in was in the same direction, but not as great as, impair- Cushing’s syndrome and may also be somewhat elev- ment found in aging rats; moreover, stress effects were ated in aging individuals, but this is probably not the prevented by prior treatment of rats with phenytoin or case for any of the other disorders listed in Table 1, at with tianeptine under the same conditions in which least at the time they are studied, except that there are phenytoin was able to prevent the stress-induced atro- elevations in glucocorticoids associated with the diur- phy of CA3 pyramidal neurons.29,52,53 A subsequent nal rhythm and stressful experiences that take place on study showed that the same repeated restraint stress a daily basis. paradigm that causes dendritic atrophy impaired the Moreover, pathological elevations of adrenal steroids short-term (4 h) retention of a spatial recognition mem- are not required for atrophy of hippocampal neurons. ory in a Y-maze, an action that was also prevented by First, in the animal models of stress-induced atrophy, tianeptine treatment during the stress regimen.54 ordinary adrenocortical stress responses are all that is Declines of hippocampally related cognitive func- needed for the process to occur with repetition of the tions, such as spatial and , occur in stressful situation. Second, with regard to human hip- human subjects and are correlated with increases in pocampal atrophy, individual differences in stress HPA activity over 3–4 years.49,50 Recent evidence has responsiveness may play a role in making some people revealed that the most severely impaired individuals more vulnerable to their own stress hormones. One have a significantly smaller hippocampal volume com- example that is instructive in understanding how indi- pared to the least impaired individuals.55 This result is vidual differences might play a role in brain changes consistent with other findings of individual differences is the finding that some individuals who are exposed in cognitive function correlated with hippocampal vol- to repeated psychosocial stress (eg public speaking) fail ume reductions in elderly humans.56,57 to habituate their elevation; these individuals Long-term stress also accelerates a number of bio- lack self-esteem and self-confidence.63 It is therefore logical markers of aging in rats, including increasing not difficult to imagine that individuals with a more the excitability of CA1 pyramidal neurons via a cal- reactive stress hormone profile will expose themselves cium-dependent mechanism and causing loss of hippo- to more cortisol and experience more stress-elevated campal pyramidal neurons.58 An important factor may neural activity, than other people who can more easily be the enhancement by glucocorticoids of calcium cur- habituate to psychosocial challenges. rents in the hippocampus,59 in view of the key role of In this regard, events related to trauma leading to calcium ions in destructive as well as plastic processes PTSD and the course of illness in recurrent depressive in hippocampal neurons. Another aspect making the illness may involve very distinct pathways of selective aging hippocampus more vulnerable may be the per- and repeated elevations of glucocorticoid hormones in sistence of excitatory amino acid release after the ter- relation to the individual experiences and reactivities. mination of a stressful experience.60 It will be In the case of disorders such as post-traumatic stress important to learn how long-term changes in neural disorder, we must learn more about stress responses activity as well as glucocorticoid secretion in such con- and neurochemical changes accompanying the initial ditions as traumatic stress and recurrent depressive ill- trauma, which usually took place months or years ness may alter the structure and function of neurons, before, as well as the ongoing stress responsiveness and particularly in the hippocampus. neurochemical activity (eg brain glucose metabolism) Another aspect of stressful experiences is the devel- in traumatized individuals. For recurrent depressive opmental influence of early stress and of neonatal illness, we are largely ignorant of the same type of his- handling on the life-course of aging and age-related tory of the depressed individual as far as endocrine cognitive impairment. As discussed elsewhere,61,62 function and neurochemical activity, as well as such early experiences can either increase or decrease responses to stressful life experiences. In recurrent the rate of brain aging through a mechanism in which depression, a long-term pattern of increased neuro- the activity of the HPA axis appears to be involved. chemical, autonomic and HPA reactivity to experi- The early experiences are believed to set the level of ences may underlie a progression of neuronal struc- responsiveness of the HPA axis and autonomic nervous tural changes, involving atrophy that might lead to system in such a way that these systems either over- permanent damage, including neuronal loss.5 react in animals subject to early unpredictable stress or Regarding the role of adrenal steroids, we have noted under-react in animals exposed to the neonatal hand- above, however, that glucocorticoids merely have to be ling procedure. present in moderate stress-levels for stress-induced atrophy of CA3 neurons to occur and that serotonin is Understanding the role of glucocorticoids in human involved along with excitatory amino acids acting via hippocampal atrophy NMDA receptors as the final common path for such Although it is tempting to attribute a dominant role in atrophy. Regarding reversibility, treatment with drugs hippocampal atrophy to glucocorticoids, this is like phenytoin or tianeptine, both of which block undoubtedly an over-simplification, and the preceding stress-induced atrophy, is a potential means of testing discussion shows how complex the interactions are both the mechanism and at the same time demonstrat- between adrenal steroids and endogenous neuro- ing the reversibility of human hippocampal atrophy. Atrophy of the human hippocampus BS McEwen 260 There is already some indication that hippocampal 10 Pavlides C, Kimura A, Magarinos AM, McEwen BS. Type I atrophy in Cushing’s syndrome is reversible adrenal steroid receptors prolong hippocampal long-term (Starkman, personal communication). If hippocampal potentiation. NeuroReport 1994; 5: 2673–2677. atrophy in conditions such as recurrent depression and 11 Pavlides C, Kimura A, Magarinos AM, McEwen BS. Hip- post-traumatic stress disorder is reversible through the pocampal homosynaptic long-term depression/ depotenti- ation induced by adrenal steroids. 1995; 68: actions of agents like tianeptine or phenytoin, then 379–385. there must be a sustained hyperactivity of the neuro- 12 Pavlides C, Watanabe Y, Magarinos AM, McEwen BS. transmitter systems involved or a repetitive activity of Opposing role of adrenal steroid Type I and Type II recep- the HPA axis. On the other hand, there may be irrevers- tors in hippocampal long-term potentiation. Neuroscience ible loss of hippocampal neurons, and some of the evi- 1995; 68: 387–394. dence in the MRI of recurrent depressive illness is con- 13 Pavlides C, Ogawa S, Kimura A, McEwen BS. Role of adre- sistent with this possibility.64 nal steroid mineralcorticoid and glucocorticoid receptors In conclusion, in so far as atrophy of the hippocam- in long-term potentiation in the CA1 field of hippocampal pus and accompanying cognitive impairment are signs slices. Brain Res 1996; 738: 229–235. of reversible neuronal atrophy, they may be treatable 14 Barnes C, McNaughton B, Goddard G, Douglas R, Adamec with agents that block the actions of excitatory amino R. Circadian rhythm of synaptic excitability in rat and monkey . Science 1977; 197: 91–92. acids, serotonin and adrenal steroids that are impli- 15 Dana RC, Martinez JL. Effect of adrenalectomy on the cir- cated in neuronal atrophy in animal models. In as cadian rhythm of LTP. Brain Res 1984; 308: 392–395. much as the decreased hippocampal volume is due to 16 Diamond DM, Bennett MC, Fleshner M, Rose GM. neuronal loss, treatment strategies should focus on the Inverted-U relationship between the level of peripheral earlier traumatic or recurrent events, and it may be corticosterone and the magnitude of hippocampal primed possible to devise strategies to reduce or prevent neu- burst potentiation. Hippocampus 1992; 2: 421–430. ronal damage. 17 Diamond DM, Fleshner M, Rose GM. impairs spatial . Behav Neurosci 1996; 110: 661–672. Acknowledgements 18 Cameron HA, Gould E. The control of neuronal birth and survival. In: Shaw CA (ed). Receptor Dynamics in Neural Research in the author’s laboratory on the topic of this Development. CRC Press: New York, 1996, pp 141–157. article is supported by NIH Grant MH 41256 and by 19 LeDoux JE. In search of an emotional system in the brain: funding from the Health Foundation (New York) and leaping from fear to and consciousness. In: Gaz- Servier (France). zaniga M (ed). The Cognitive . MIT Press: Cambridge, 1995, pp 1049–1061. 20 Magarinos AM, McEwen BS, Flugge G, Fuchs E. Chronic References psychosocial stress causes apical dendritic atrophy of hip- pocampal CA3 pyramidal neurons in subordinate tree 1 Sapolsky R. Stress, the and the Mechanisms shrews. J Neurosci 1996; 16: 3534–3540. of Neuron Death. MIT Press: Cambridge, 1992, pp 1–423. 21 McEwen BS, Gould E, Orchinik M, Weiland NG, Woolley 2 Sapolsky R, Krey L, McEwen BS. The neuroendocrinology CS. Oestrogens and the structural and functional plasticity of stress and aging: the glucocorticoid cascade hypothesis. of neurons: implications for memory, ageing and neuro- Endocrin Rev 1986; 7: 284–301. degenerative processes. In: Goode J (ed). Ciba Foundation 3 Landfield PW, Eldridge JC. Evolving aspects of the gluco- Symposium #191. The Non-reproductive Actions of Sex corticoid hypothesis of brain aging: hormonal modulation Steroids. CIBA Foundation: London, 1995, pp 52–73. of neuronal calcium homeostasis. Neurobiol Aging 1994; 22 Gazzaley AH, Weiland NG, McEwen BS, Morrison JH. Dif- 15: 579–588. 4 McEwen BS, Albeck D, Cameron H, Chao HM, Gould E, ferential regulation of NMDAR1 mRNA and protein by Hastings N, Kuroda Y, Luine V, Magarinos AM, McKit- estradiol in the rat hippocampus. J Neurosci 1996; 16: trick CR, Orchinik M, Pavlides C, Vaher P, Watanabe Y, 6830–6838. Weiland N. Stress and the brain: a paradoxical role for 23 Popov VI, Bocharova LS. Hibernation-induced structural adrenal steroids. In: Litwack GD (ed). Vitamins and Hor- changes in synaptic contacts between mossy fibres and mones. Academic Press: San Diego, 1995, pp 371–402. hippocampal pyramidal neurons. Neuroscience 1992; 48: 5 Sheline YI. Hippocampal atrophy in major depression: a 53–62. result of depression-induced neurotoxicity? Mol Psy- 24 Popov VI, Bocharova LS, Bragin AG. Repeated changes of chiatry 1996; 1: 298–299. dendritic morphology in the hippocampus of ground 6 Jacobson L, Sapolsky R. The role of the hippocampus in squirrels in the course of hibernation. Neuroscience 1992; feedback regulation of the hypothalamic-pituitary-adreno- 48: 45–51. cortical axis. Endocrin Rev 1991; 12: 118–134. 25 Magarinos AM, Verdugo Garcia JM, McEwen BS. Chronic 7 Eichenbaum H, Otto T. The hippocampus — what does it restraint stress causes ultrastructural changes in rat mossy do? Behav Neural Biol 1992; 57: 2–36. fiber terminals. Abs Soc Neurosci 1996; 22: 1196. 8 McEwen BS, Weiss J, Schwartz L. Selective retention of 26 McKittrick CR, Magarinos AM, Blanchard DC, Blanchard corticosterone by limbic structures in rat brain. Nature RJ, McEwen BS, Sakai RR. Chronic social stress decreases 1968; 220: 911–912. binding to 5HT transporter sites and reduces dendritic 9 De Kloet ER, Azmitia EC, Landfield PW. Brain cortico- arbors in CA3 of hippocampus. Abs Soc Neurosci 1996; steroid receptors: studies on the mechanism, function, and 22: 2060. neurotoxicity of corticosteroid action. Ann New York 27 Taylor CP, Meldrum BS. Na+ channels as targets for neuro- Acad Sci 1996; 1–746. protective drugs. TIPS 1995; 16: 309–316. Atrophy of the human hippocampus BS McEwen 261 28 Freund TF, Buzaski G. Interneurons of the hippocampus. 47 Lupien SJ, McEwen BS. The acute effects of cortico- Hippocampus 1996; 6: 347–470. steroids on : integration of animal and human 29 Watanabe Y, Gould E, Cameron H, Daniels D, McEwen BS. model studies. Brain Res Rev 1996; (in press). Stress and antidepressant effects on hippocampus. Eur J 48 McEwen BS, Sapolsky RM. Stress and cognitive function. Pharmacol 1992; 222: 157–162. Curr Opin Neurobiol 1995; 5: 205–216. 30 Delbende C, Contesse V, Mocaer E, Kamoun A, Vaudry 49 Lupien S, Lecours AR, Lussier I, Schwartz G, Nair NPV, H. The novel antidepressant, tianeptine, reduces stress- Meaney MJ. Basal cortisol levels and cognitive deficits in evoked stimulation of the hypothalamo-pituitary-adrenal human aging. J Neurosci 1994; 14: 2893–2903. axis. Eur J Pharmacol 1991; 202: 391–396. 50 Seeman TE, McEwen BS, Singer BH, Albert MS, Rowe JW. 31 Whitton P, Sarna G, O’Connell M, Curzon G. The effect Increase in urinary cortisol excretion and memory of the novel antidepressant tianeptine on the concen- declines: MacArthur studies of successful aging. J Clin tration of 5-hydroxytryptamine in rat hippocampal dialys- Exp Endocrinol 1996; (in press). ates in vivo. 1991; 104: 81–85. 51 Kirschbaum C, Wolf OT, May M, Wippich W, Hellhammer 32 Mennini T, Miari A. Modulation of 3H glutamate binding DH. Stress- and treatment-induced elevations of cortisol by serotinin in rat hippocampus: an autoradiographic levels associated with impaired verbal and spatial declara- study. Life Sci 1991; 49: 283–292. tive memory in healthy adults. Life Sci 1996; 58: 1475– 33 Rahmann S, Neumann RS. Activation of 5-HT2 receptors 1483. facilitates depolarization of neocortical neurons by N- 52 Luine V, Villegas M, Martinez C, McEwen BS. Related methyl-d-aspartate. Eur J Pharmacol 1993; 231: 347–354. stress causes reversible impairments of spatial memory 34 Lowy MT, Gault L, Yamamoto BK. Adrenalectomy attenu- performance. Brain Res 1994; 639: 167–170. ates stress-induced elevations in extracellular glutamate 53 Watanabe Y, Gould E, Cameron HA, Daniels DC, McEwen concentrations in the hippocampus. J Neurochem 1993; BS. Phenytoin prevents stress- and corticosterone-induced 61: 1957–1960. atrophy of CA3 pyramidal neurons. Hippocampus 1992; 35 Watanabe Y, Weiland NG, McEwen BS. Effects of adrenal 2: 431–436. steroid manipulations and repeated restraint stress on 54 Conrad CD, Galea LAM, Kuroda Y, McEwen BS. Chronic mRNA leels and excitatory amino acid recep- stress impairs rat spatial memory on the Y-maze and this tor binding in hippocampus. Brain Res 1995; 680: 217– effect is blocked by tianeptine pre-treatment. Behav Neu- 225. rosci 1996; 110: 1321–1334. 36 Orchinik M, Weiland NG, McEwen BS. Adrenalectomy 55 Lupien S, DeLeon M, DeSanti S, Convit A, Tannenbaum selectively regulates GABAa receptor subunit expression BM, Nair NPV, McEwen BS, Hauger RL, Meaney MJ. in the hippocampus. Mol Cell Neurosci 1994; 5: 451–458. Longitudinal increase in cortisol during human aging pre- 37 Bartanusz V, Aubry JM, Pagliusi S, Jezova D, Baffi J, Kiss dicts hippocampal atrophy and memory deficits. Abs Soc JZ. Stress-induced changes in messenger RNA levels of N- Neurosci 1996; 22: 2061. methyl-d-aspartate and subunits in selected 56 Golomb J, Kluger A, de Leon MJ, Ferris SH, Convit A, Mit- regions of the rat hippocampus and . Neuro- telman MS, Cohen J, Rusinek H, De Santi S, George AE. science 1995; 66: 247–252. Hippocampal formation size in normal human aging: a 38 Weiland NG, Orchinik M, McEwen BS. Corticosterone correlate of delayed secondary memory performance. regulates mRNA levels of specific subunits of the NMDA Learning and Memory 1994; 1: 45–54. receptor in the hippocampus but not in cortex of rats. Abs 57 Convit A, de Leon MJ, Tarshish C, De Santi S, Kluger A, Soc Neurosci 1995; 21: 502. Rusinek H, George AJ. Hippocampal volume losses in 39 Monaghan DT, Holets VR, Toy DW, Cotman CW. Anatom- minimally impaired elderly. Lancet 1995; 345: 266. ical distributions of four pharmacologically distinct 3H-l- 58 Kerr S, Campbell L, Applegate M, Brodish A, Landfield P. glutamate binding sites. Nature 1983; 306: 176–179. Chronic stress-induced acceleration of electrophysiologic 40 Ishizuka N, Weber J, Amaral DG. Organization of intra- and morphometric biomarkers of hippocampal aging. J hippocampal projections originating from CA3 pyramidal Neurosci 1991; 11: 1316–1324. cells in the rat. J Comp Neurol 1990; 295: 580–623. 59 Kerr DS, Campbell LW, Thibault O, Landfield PW. Hippo- 41 Li XG, Somogyi P, Ylinen A, Buzsaki G. The hippocampal campal activation enhances volt- CA3 network: an in vivo intracellular labeling study. J age-dependent Ca2+ conductances: relevance to brain Comp Neurol 1994; 339: 181–208. aging. Proc Natl Acad Sci USA 1992; 89: 8527–8531. 42 Hablitz JJ, Johnston D. Endogenous nature of spontaneous 60 Lowy MT, Wittenberg L, Yamamoto BK. Effect of acute bursts in hippocampal neurons. Cell Mol Neurobiol 1981; stress on hippocampal glutamate levels and spectrin pro- 1: 325–334. teolysis in young and aged rats. J Neurochem 1995; 65: 43 Buzsaki G, Penttonen M, Nadasdy Z, Bragin A. Pattern 268–274. and inhibition-dependent invasion of den- 61 Meaney M, Aitken D, Berkel H, Bhatnager S, Sapolsky R. drites by fast spikes in the hippocampus in vivo. Proc Natl Effect of neonatal handling of age-related impairments Acad Sci USA 1996; 93: 9921–9925. associated with the hippocampus. Science 1988; 239: 44 Avery RB, Johnston D. Multiple channel types contribute 766–768. to the low-voltage-activated calcium current in hippocam- 62 Meaney MJ, Tannenbaum B, Francis D, Bhatnagar S, pal CA3 pyramidal neurons. J Neurosci 1996; 16: 5567– Shanks N, Viau V, O’Donnell D, Plotsky PM. Early 5582. environmental programming hypothalamic-pituitary- 45 Smith MA, Makino S, Kvetnansky P, Post RM. Stress and adrenal responses to stress. Semin Neurosci 1994; 6: glucocorticoids affect the expression of brain-derived neu- 247–259. rotrophic factor and neurotrophin-3 mRNAs in the hippo- 63 Kirschbaum C, Prussner JC, Stone AA, Federenko I, Gaab campus. J Neurosci 1995; 15: 1768–1777. J, Lintz D, Schommer N, Hellhammer DH. Persistent high 46 Uno H, Ross T, Else J, Suleman M, Sapolsky R. Hippocam- cortisol responses to repeated psychological stress in a pal damage associated with prolonged and fatal stress in subpopulation of healthy men. Psychosomat Med 1995; primates. J Neurosci 1989; 9: 1705–1711. 57: 468–474. Atrophy of the human hippocampus BS McEwen 262 64 Sheline YI, Wang PW, Gado MH, Csernansky JC, Vannier 68 Starkman M, Gebarski S, Berent S, Schteingart D. Hippo- MW. Hippocampal atrophy in recurrent major depression. campal formation volume, memory dysfunction, and cor- Proc Natl Acad Sci USA 1966; 93: 3908–3913. tisol levels in patients with Cushing’s syndrome. Biol Psy- 65 Bremner DJ, Randall P, Scott TM, Bronen RA, Seibyl JP, chiatry 1992; 32: 756–765. Southwick SM, Delaney RC, McCarthy G, Charney DS, 69 Bogerts B, Lieberman JA, Ashtair M, Bilder RM, De Greef Innis RB. MRI-based measurement of hippocampal vol- G, Lerner G, Johns C, Masiar S. Hippocampus-amygdala ume in patients with combat-related posttraumatic stress volumes and psychopathology in chronic schizophrenia. disorder. Am J Psychiatry 1995; 152: 973–981. Biol Psychiatry 1993; 33: 236–246. 66 Gurvits TV, Shenton ME, Hokama H, Ohta H, Lasko NB, 70 Fukuzako H, Fukuzako T, Hashiguchi T, Hokazono Y, Orr SP, Kikinis R, Jolesz FA, McCarley RW, Pitman RK. Takeuchi K, Hirakawa K, Ueyama K, Takigawa M, Kajiya Reduced hippocampal volume on magnetic resonance Y, Nakajo M, Fujimoto T. Reduction in hippocampal for- imaging in chronic post-traumatic stress disorder. Biol mation volume is caused mainly by its shortening in Psychiatry 1996; 40: 1091–1099. chronic schizophrenia: assessment by MRI. Biol Psy- 67 de Leon MJ, Golomb J, George AE, Convit A, Tarshish CY, chiatry 1996; 39: 938–945. McRae T, De Santi S, Smith G, Ferris SH, Noz M, Rusinek 71 Sapolsky RM. Why stress is bad for your brain. Science H. The radiologic prediction of Alzheimer disease: the 1996; 273: 749–750. atrophic hippocampal formation. Am J Neuroradiol 1993; 14: 897–906.