NeuroscienceandBiobehavioralReviews,Vol. 16, pp. 525-534, 1992

0149-7634/92$5.00 + .00 Copyright©1992PergamonPress Ltd.

Printedin the U.S.A. All rightsreserved.

Chronic Mild Stress-Induced Anhedonia: A Realistic Animal Model of Depression PAUL WILLNER, I RICHARD MUSCAT* AND MARIUSZ PAPPt

Department o f Psychology, City o f L o n d o n Polytechnic, Old Castle St., London E1 7NT, UK Received 5 F e b r u a r y 1992 WILLNER, P., R. MUSCAT AND M. PAPP. Chronic mild stress-induced anhedonia: A realistic animal model of depression. NEUROSCI BIOBEHAV REV 16(4) 525-534, 1992.-Chronic sequential administration of a variety of mild stressors causes a decrease in responsiveness to rewards in rats, which is reversed by chronic administration of antidepressant drugs. This paper reviews the validity of chronic mild stress-induced anhedonia as an animal model of depression, and the evidence that changes in hedonic responsiveness in this model are mediated by changes in the sensitivity of dopamine D2 receptors in the nucleus accumbens. The review opens with an analysis of the design features of animal models of depression, and ends with a brief account of other animal models of anhedonia. Chronic mild stress Construct validity

Anhedonia Dopamine

Animal model of depression Predictive validity Nucleus accumbens D2 receptors Sensitization

Face validity

disadvantages of using complex, chronic models are small, relative to the costs of testing an ineffective drug in the clinic. The question of suitability for purpose arises also, not only in relation to drug screening, but also in relation to the different investigative uses of simulations of depression. In general, animal models do not attempt to simulate the total complexity of a psychiatric syndrome, but rather select particular features, which may be more or less appropriate for the problems that the model is then used to address. For example, simulations of depression based on the use of different inbred strains (61,114) are particularly appropriate for investigating issues relating to individual differences in susceptibility to depression, but this advantage may not extend to their suitability for investigating the pathogenesis of depression or the mechanisms of action of antidepressant drug. The essential requirement if a model is to be used to study the pathogenesis of depression is that it must employ realistic inducing conditions. Similarly, the essential requirement in a model designed to elucidate the mechanisms underlying the therapeutic action of antidepressants is that the behavioural changes expressed in the model persist sufficiently for antidepressants to be administered on a clinically relevant time scale: at least three and preferably greater than 6 weeks.

ANIMAL models of depression are used for two distinct purposes: as screening tests in the context of antidepressant discovery and development programs, and as simulations within which to investigate aspects of depression. These two uses have different and to some extent conflicting requirements, and, therefore, a model developed for one purpose may be less suitable for the other. Traditional screening tests are subject to a number of logistical imperatives, such as ease and speed of use, and high capacity. For all of these reasons, a screening test should be as simple as possible and should respond acutely to antidepressant treatment. However, a simulation of depression alms to mimic aspects of the clinical situation, and from this perspective should demonstrate a slow onset of antidepressant action over several weeks of chronic treatment. A simulation should also embody a degree of complexity, to permit investigation of the behavioural features of the model. These tensions have led to a situation in which the tests used routinely within the pharmaceutical industry are of limited or minimal validity as simulations of depression (96). The fact that some features of a valid simulation of depression are incompatible with traditional screening requirements does not, however, mean that drug development programs must dispense with realistic models. Most pharmaceutical companies have by now abandoned the high volume, random screening approach in favour of the development of a small number of compounds specifically designed to meet predetermined pharmacological criteria. In such a program, the place of behavioural screening methods shifts from the discovery phase to the development phase. In this context, the logistical

Validity and Realism in Animal Models of Depression The development of valid animal models of psychiatric disorders can only proceed if procedures are in place by which the validity of animal models can be assessed. A general meth-

* Present address: Department of Biomedical Sciences, University of Malta, Msida, Malta. t Present address: Institute of Pharmacology, Polish Academy of Sciences, Krakow, Poland. ; To whom requests for reprints should be addressed. 525

526 odology exists for assessing the validity of animal models, under which models are assessed with respect to their predictive validity, face validity, and construct validity: Predictive validity refers to the accuracy of predictions made from the model, particularly in respect of drug actions; face validity refers to the phenomenological similarity between the model and the disorder modelled; and construct validity refers to the theoretical rationale for the model. These methods were first used to assess the validity of animal models of depression (96) and have since been applied more widely (19,25,26,80). The central issue in assessing the face validity of a simulation of depression is the degree to which the simulation is realistic. In a classic paper, McKinney and Bunney (49) proposed that an animal model should resemble the condition modelled in terms of its etiology, symptomatology, treatment, and biological basis. However, a second important theoretical contribution (1) pointed out that not all symptoms carry equal weight: A valid animal model should demonstrate a resemblance to the clinically defined core symptoms of the disorder, rather than the subsidiary symptoms. Furthermore, not all of the symptoms of depression can be modelled in animals; symptoms that can only be known by subjective verbal report are, in principle, excluded (96,100). Depression is a complex psychological disorder, which in the DSM-III-R diagnostic system (2) is defined by the presence of at least one core symptom combined with a number of additional subsidiary symptoms. It is clear that certain of the symptoms of depression, which can be defined operationally (eg., weight change, psychomotor retardation), could, in principle, be modelled in animals, though equally clearly, other symptoms, which rely on verbal report (eg., feelings of worthlessness, thoughts of death or suicide), could not. Of particular importance is that DSM-III-R defines two core symptoms for the diagnosis of major depression: loss of interest or pleasure (anhedonia) and depressed mood. Anhedonia can be modelled in animals, but depressed mood cannot. It follows that anhedonia is the essential symptom to reproduce in any attempt to simulate depression realistically. Anhedonia is important for another reason also: In DSMIII, anhedonia was the defining symptom of the principal subtype of major depression, melancholia. DSM-III-R adopted a more relaxed approach to the diagnosis of melancholia, in which anhedonia is one of several symptoms that are given equal weight. However, it is clear from an account of the DSM-III-R decision making process that anhedonia was demoted from its DSM-III preeminence because there were insufficient data at the time (1985) to support its retention, not because evidence had emerged to justify its removal (118). In the light of evidence that pervasive anhedonia may be a central characteristic in depressions with reduced rapid eye movement (REM) sleep latency (24), which is a reliable biological marker of melancholia (40), the decision to downplay the diagnostic significance of anhedonia may have been mistaken. An examination of the behavioural features of animal models of depression reveals that most of the current models provide, at best, a superficial reflection of clinical symptomatology (101). The behaviour most frequently studied in animal models is a decrease in locomotor activity. However, while it is true that psychomotor changes are a central feature of severe depression (60), it cannot be simply assumed that retardation and agitation are equivalent to decreases or increases in locomotor activity. Slowness of thinking is at least as important a feature of psychomotor retardation as a decrease in physical activity, and psychomotor agitation involves such features as hand-wringing and outbursts of complaining,

WILLNER, MUSCAT AND P A P P in addition to a restless increase in locomotion. With the exception of the primate separation model (27), the degree of resemblance between the changes in locomotor activity observed in animal models of depression and the psychomotor features of depression is questionable. Two other behaviours that form the centerpiece of a number of animal models of depression, impulsivity (7,82), and decreased social contact (23,27,108), do not feature prominently among the diagnostic criteria for depression, perhaps because they have low specificity for depression as compared to a range of other psychiatric disorders. There are a variety of social factors that powerfully influence vulnerability to, and recovery from, depression, two of which, lack of social support (12) and low social skill (42), could be said to involve a decrease in social contact. However, decreased social contact appears among the DSM-III-R criteria for depression as an aspect of "loss of interest," rather than as a symptom in its own right. Similarly with impulsivity: This concept is used increasingly by basic neuroscientists to organize the behavioural literature on the functions of the forebrain serotonin (5-HT) projections (88), and may lead to a reconceptualization of the relationship between depression and other disorders in which 5-HT is implicated (99); at present, however, viewed as a symptom of depression, impulsivity is at best a minor aspect of psychomotor agitation. Despite its crucial importance as the essential element in a realistic simulation of depression, anhedonia is rarely modelled. In fact, this symptom is seen in only four models: learned helplessness (114), withdrawal from chronic amphetamine or cocaine (35,37), neonatal treatment with the tricyclic antidepressant clomipramine (92), and chronic stress (31, 109). The present review is concerned primarily with reviewing recent evidence relating to the anhedonic effect of chronic stress, and in particular, with the more recent version of this model in which the level of applied stress is minimized (107,109). We discuss the validity of the chronic mild stress (CMS) procedure as an animal model of depression, and review studies of the neural basis of CMS-induced anhedonia. THE CHRONIC MILD STRESS MODEL The CMS model involves the chronic sequential application, to rats, of a variety of extremely mild stressors. The procedure was originally designed to be a milder version of a chronic varied stress protocol described by Katz and colleagues (31-33), which includes a number of severely stressful elements. In the majority of studies using the original procedure, locomotor activity in an open field was used to assess reactivity to acute stress, which was decreased in chronically stressed animals, and partially restored by concurrent treatment with antidepressant drugs leg. (32,33,70)]. However, in one study, it was reported that chronically stressed rats also failed to increase their fluid intake when sucrose or saccharin were added to their drinking water, and this effect was prevented by concurrent administration of imipramine (31). The possibility that this abnormality of sweet fluid intake might reflect a generalized anhedonia provided the motivation for the development of a less controversial means of achieving this end. In the CMS procedure, all of the individual elements are, at worst, mildly stressful. During 1-3 weeks exposure to CMS, rats display a reduction in sensitivity to rewards, which is usually monitored by a decrease in their consumption of palatable weak sucrose solution. CMS-induced behavioural deficits may be maintained for several months; however, normal

A REALISTIC ANIMAL MODEL OF DEPRESSION behaviour is restored, during continued application of CMS, by chronic treatment with antidepressant drugs (55,57,58, 79,109). The reduction of sucrose intake by CMS is an extremely robust effect, which is not always apparent within the first, or even the second, week of exposure of CMS, but has eventually been observed in every one of more than fifty experiments.

Methodological Issues In the standard version of the CMS protocol, as described by Willner and colleagues (109), male Lister hooded rats are first trained to consume a weak sucrose solution, which was available for 60 min in the home cage, following 20-h food and water deprivation. The concentration of sucrose used was 0.7070 (w : v) in the early experiments, and was slightly higher (1070) in more recent studies. The choice of sucrose concentration is dictated by the fact that the sucrose concentrationintake function is bell-shaped. At low concentrations, on the ascending limb of the concentration-intake function, sucrose intake rises monotonically with concentration, and intake is monotonically related to reward value, as assessed by preference measures in choice tests. However, at high concentrations, on the descending limb of the concentration-intake function, intake is no longer related in a simple way to reward value (54,68). For CMS studies, sucrose concentration was set midway up the ascending limb of the concentration-intake function (1%), so that changes in responsiveness to reward, in either direction, would be reflected in corresponding changes in sucrose intake. The training phase of the procedure typically lasts for 2-3 weeks. Subsequently, half of the animals are subjected to CMS, and sucrose intake tests are conducted once weekly. The standard CMS protocol consists of the sequential application of a variety of mild stressors, each for a period of between 2 and 20 h, in a schedule that lasts for a full week, and is repeated thereafter. The schedule typically consists of: two 20-h periods of food and water deprivation, one immediately prior to the sucrose intake test, the other followed by 2 h of restricted access to food (scattering of a few 45 mg precision pellets in the cage); one additional 16-h period of water deprivation; two periods of continuous overnight illumination; two periods (7 and 17 h) of 45 degree cage tilt; one 17-h period of paired housing; one 17-h period in a soiled cage (100 ml water in sawdust bedding); two periods (3 and 5 h) of intermittent white noise (85 dB); three periods (7, 9, and 17 h) of low intensity stroboscopic illumination (60 flashes/minute). In the paired housing condition, animals are always housed in the same pairs, but the location alternates between the home cages of each member of the pair. CMS reliably causes a decreased intake of sucrose, relative to non-stressed control animals, which, once established, can be maintained by continued application of CMS for 3 months or more, and persists for 2-3 weeks following the termination of CMS (55,57,58,79,109). Although the majority of studies in this model have used Lister hooded rats, sucrose intake is also suppressed by CMS in Long Evans rats (Broekkamp, personal communication). Whether these effects are of similar magnitude has not been evaluated. Strain differences have been reported in other animal models of depression (72,114), and it was recently reported that sucrose intake was more suppressed by CMS in an inbred Sprague-Dawley-derived hypercholinergic strain (FSL) than in the corresponding hypocholinergic strain (FRL) (75). Although control animals are described as "non-stressed," they are, in fact, subjected to two stressors, which could po-

527 tentially confound the results: all animals are housed singly and sucrose consumption tests are routinely carried out following 20-h food and water deprivation, applied equally to "stressed" animals and "controls." However, neither of these factors is responsible for the difference in sucrose intake between CMS-exposed and control animals: Sucrose intake was decreased by CMS to a similar extent in singly-housed and paired-housed animals, and also similar proportional decreases were seen in deprived and non-deprived as in animals (53). However, relative to testing with deprivation, intakes in non-deprived animals were both smaller and more variable, greatly reducing the statistical power of the experiment. Testing is therefore carried out following deprivation in order to reduce the number of animals needed to obtain statistically significant effects (53). Studies have been carried out to examine whether any of the elements of the CMS protocol are either necessary or sufficient to cause anhedonia. In these experiments the CMS timetable was first simplified by simply presenting stressors overnight, rather than at all times of day and night. Using this simplified procedure, the effects of subsets of stress elements were examined. This revealed particularly potent effects of a subset of three elements, each presented twice weekly: paired housing, exposure to wet bedding, and 45 ° cage tilt. The effects of each of these elements individually were therefore examined on single and repeated presentation. Of the various elements used, the only one that by itself reduced sucrose intake was paired housing. However, the effects of a single weekly pairing showed rapid habituation, and the effects of 6 weekly pairings (ie., almost every night) also showed habituation, though more slowly (53). Although paired housing (in animals normally housed singly) appears to be the single most potent element in the CMS protocol, it is not a necessary element. Experiments have also been conducted in which pairing was simply removed from the standard protocol. In three successive replications, the remaining elements, in combination, were found to decrease sucrose intake, despite the fact that none of them did so individually (53). Furthermore, while the effect of 6 weekly pairings habituated after 4-5 weeks, the effects of 2 weekly pairings, in combination with four other elements that in themselves were ineffective, were large and persistent. Thus, no one element of the CMS protocol is either necessary or sufficient to maintain a persistent decrease in sucrose intake, but variety does appear to be essential (53).

Construct Validity of the CMS Model The central question in assessing the validity of the CMS paradigm as a realistic animal model of depression is the assumption that CMS causes anhedonia: that is, that the intake of a weak solution of sucrose represents a valid measure of sensitivity to reward. This assumption has, therefore, been subjected to intensive scrutiny. An initial concern was to establish that CMS does not cause nonspecific changes in consummatory activity. In contrast to the decreased intake of weak sucrose solutions, CMS did not decrease intake of plain water. However, when 1070 sucrose was added to the water, control animals increased their intake, but CMS animals did not (53). Similarly, although CMS usually causes weight loss relative to non-stressed controls, neither food intake (53,55,66) nor the intake of a concentrated (34070) sucrose solution (103,107) were decreased by CMS. However, CMS did decrease the intake of a weak solution of saccharin (0.0207o w : v) (109). These data show that the

528 reduction of sucrose intake by CMS is unrelated to the calorie content of the sucrose. In addition, the fact that there is no generalized decrease in consummatory behaviour also rules out changes in circadian rhythmicity as an explanation of the decreased intake of sucrose or saccharin (53). At first sight, the failure of CMS to decrease food intake (53,55,66) appears to be incompatible with an anhedonic effect. However, this would only be the case if the quantity of food consumed was monotonically relationed to its palatability. In order to examine this question, consumption of a palatable wet mash was examined in rats subjected chronically (410 weeks) to CMS. In control animals, the addition of 30°70 or 40070 sucrose caused a decrease in the quantity of mash consumed, but increased the rate of eating. Both the increase in eating rate and the decrease in intake, at high sucrose concentration, were markedly attenuated in stressed animals, which therefore had higher intakes of very sweet mash but lower rates of eating, relative to control animals (78). Eating rate is usually considered to be closely related to palatability, as assessed, for example in preference studies (5,16,81,94). The failure of stressed animals to adapt their intake to increases in sweetness supports the hypothesis that animals exposed to CMS are anhedonic. In order to provide a measure of sensitivity to reward, independent of consummatory behaviour, and also to investigate responsiveness to a variety of different rewards, an extensive series of experiments was conducted using the place conditioning paradigm. In this paradigm, rewarding properties are inferred if animals display a preference for a distinctive environment in which the reward was previously administered (13). In place conditioning experiments, animals were first subjected to CMS for 2-3 weeks, using the standard protocol, then switched to a modified procedure in which stressors were administered at night only; this was done so that training and testing could be carried out (during the day) uncontaminated by concurrent administration of stressors. The animals were first adapted to an apparatus containing two distinctive compartments (black and white), and preference for the two arms was assessed in a 10 min preconditioning test. They then received a series of pairs of training trials: In one trial of each pair the animal was administered a reward in either the black or the white compartment; the following day it was confined in the alternate compartment and no reward was administered. Finally, preference for the two compartments was reassessed in a 10 rain postconditioning test. Control animals showed marked increases in preference for the reward-associated compartment, irrespective of whether this was the initially preferred (black) or the initially nonpreferred (white) side. However, place preferences were attenuated or abolished by CMS in every experiment. A variety of natural and drug rewards were used, in different experiments, including sucrose solutions (0.7 or 34070 w : v ) ; food pellets; amphetamine (0.5-1.0 mg/kg); quinpirole (100-400 #g/kg); and morphine (0.75 mg/kg) (62,63,66). By contrast, druginduced place aversions (established using naloxone: 0.75 mg/ kg and picrotoxin: 2 mg/kg) were unaffected by CMS (62). Thus, the effect of CMS is properly described as anhedonia (decreased sensitivity to rewards), rather than being a nonspecific motivational impairment or a failure of associative learning. Finally, it has recently been reported that CMS caused a 50070 increase in the threshold for ICSS reward. The threshold current necessary to maintain ICSS, at electrodes implanted in the ventral tegmental area (VTA) increased gradually during three weeks of CMS; this effect was prevented by concur-

WILLNER, MUSCAT AND P A P P rent administration of desipramine (50). Together with the place preference data, these results support the position that decreases in sucrose intake provide a simple means of detecting CMS-induced anhedonia.

Face Validity of the CMS Model By focussing on anhedonia as the central feature, considerable face validity is built in to the design of the CMS model. However, in addition to modelling anhedonia, the model also provides a number of other realistic parallels to depression, particularly in relation to four aspects of antidepressant drug treatment: Firstly, treatment with antidepressant drugs does not commence until a hedonic deficit has been established; thus the model demonstrates reversal of hedonic deficits by antidepressants, rather than prophylactic effects. Secondly, stress continues during the treatment period; similarly in the clinical situation, there would usually be no abrupt change in the patient's circumstances associated with the onset of treatment. Thirdly, the time course of recovery (3-5 weeks) corresponds closely to that observed in the clinic. Fourthly, no changes are seen in the control animals, corresponding to the failure of antidepressants to elevate mood in nondepressed volunteers (55,57,58,79,109). The CMS model also offers a realistic simulation of the etiology of depression. Many studies have implicated chronic low grade stress ("strain") as an important factor in the onset of depression (3,6,12,20,28,30), and the CMS model offers a reasonable approximation to these stresses of daily life. The more conventional stress models, which use a single session of moderate to severe stress are usually likened to "major life events," which also cause a substantial increase in the risk of depression (12,43). However, it is important to recognize that the adverse consequences of life events are present for a prolonged period of 6-12 months. A number of psychological processes have been described that mediate these long-lasting effects; in particular, the impact of minor stressors is exacerbated in the aftermath of a life event (11,12). Life events should not, therefore, be viewed as acute stressors: Indeed, in the case of bereavement a diagnosis of depression is explicitly excluded during the period of acute loss. From this perspective, the CMS model may also be considered to have face validity in relation to the etiological role of life events: perhaps more so than models using acute stress. The data from the CMS model go beyond a general relationship between chronic stress and depression, in suggesting a relationship between chronic stress and anhedonia specifically. To our knowledge, this issue has been addressed in only a single study, which was derived explicitly from the CMS model. In this study, a questionnaire designed to measure the frequency and perceived intensity of mild stress (30) was administered to two groups of depressed inpatients, matched for severity of depression, and an age- and sex-matched control group. One group of patients were anhedonic (DSM-III melancholic); the other were not (DSM-III non-melancholic). The three groups did not differ in their reporting of the frequency of stressful events, but the anhedonic group reported a substantial increase in perceived severity of stress, relative to both of the other groups (110). The confirmation of a clinical relationship between chronic stress and anhedonia, as predicted from the CMS model, both supports the validity of the model, and also demonstrates its heuristic value. Predictive Validity of the CMS Model Studies to date have examined the effects in the CMS model of chronic treatment with three tricyclics (imipramine,

A REALISTIC A N I M A L MODEL OF DEPRESSION desipramine, and amitriptyline) and two atypical antidepressants: the specific 5HT uptake inhibitor fluoxetine, and the specific NE uptake inhibitor maprotiline. In these studies, CMS was applied for several weeks, in order to induce anhedonia, before the onset of drug treatment, and CMS continued throughout the treatment period. Drugs were administered once dally at 5 mg/kg. All five antidepressants caused a return to control levels of sucrose intake, after 3-5 weeks of treatment. The tricyclic effects have been replicated in several studies (55,57,58,79,109). In a further experiment, food-induced place preference conditioning was assessed, following 12 weeks of CMS, in animals treated chronically (8 weeks) with the atypical antidepressants fluoxetine or maprotiline (5 mg/kg/day). There was no evidence of place conditioning in vehicle-treated CMSexposed animals; by contrast, place conditioning was normal in the antidepressant-treated CMS-exposed animals (55). Finally, another recent study showed that concurrent treatment with DMI (10 mg/kg/day) prevented the increase in ICSS threshold in animals exposed for 3 weeks to a modified version of the CMS procedure (50). Thus, chronic antidepressant treatment caused a return to normal responsiveness to rewards, as assessed by a variety of different methods: sucrose intake, place conditioning, and brain stimulation reward. As antidepressant drugs are effective not only in depression, but also in most forms of anxiety (76), it is particularly important to demonstrate that anxiolytics are ineffective in a putative animal model of depression. The non-antidepressant benzodiazepine chlordiazepoxide (5 m g / k g / d a y for 8 weeks) was ineffective in reversing CMS-induced anhedonia: unlike antidepressant-treated animals, chlordiazepoxide-treated CMSexposed animals showed no recovery of sucrose drinking, and no evidence of food-induced place preference conditioning (55). NEURAL MECHANISMSOF CMS-INDUCED ANHEDONIA In recent years, a strong consensus has emerged that the mesolimbic dopamine (DA) projection from the ventral tegmental area (VTA) to the nucleus accumbens plays a crucial role in mediating the behavioural effects of rewards, one of the most important lines of evidence being the suppression of rewarded behaviour by DA receptor antagonists (38,67, 102,106, I I I, 112). There are striking parallels between the effects of DA receptor antagonists and those of CMS. Effects of DA antagonists comparable to those described above for CMS include: I. selective suppression of the intake of, and preference for, dilute sucrose solutions, while sparing consumption of more concentrated solutions (54,69); this effect of neuroleptics has been shown to be localized within the nucleus accumbens (68); 2. attenuation of the intake-reducing and rate-enhancing effects of high sucrose content on the consumption of wet mash (78); 3. attenuation of food-induced place preference conditioning (89); 4. attenuation of amphetamine- and morphine-induced place preference conditioning; again, an effect localized to the accumbens (13); 5. an increase in electrical threshold for ICSS through electrodes implanted in the VTA (67). In view of the extent of behavioural similarities between DA antagonist- and CMS-treated animals, the mesolimbic DA

529 system has formed a natural focus for studies of the neural mechanisms underlying CMS-induced anhedonia and its reversal by antidepressant drugs.

Presynaptic Mechanisms The initial series of experiments examined neurochemical changes in the CMS model. Three or seven weeks exposure to CMS increased the concentration of DA and 5HT, and their metabolites, in limbic areas, but not in the caudate nucleus; concentrations of NA were unaltered by CMS (103). In subsequent experiments, DA release was measured in vivo, in anaesthetized animals, using fast cyclic voltammetry. In these studies, CMS increased the electrically-stimulated release of DA, and again, these effects were observed in the nucleus accumhens only, not in the caudate nucleus (90). A related observation was that CMS also decreased the sensitivity of inhibitory DA autoreceptors, again, in the nucleus accumbens only (90), consistent with earlier behavioural observations that CMS decreased sensitivity to a low (autoreceptor-selective) dose of apomorphine (58). These studies point to the nucleus accumbens as a region significantly involved in mediating the effects of CMS. However, these presynaptic changes cannot in themselves explain the alterations in sensitivity to reward. One immediate problem is that CMS apparently increases DA release in the accumbens, as observed also with acute stressors (8,18,29,91,114), which is difficult to reconcile with a neuroleptic-like behavioural profile. A second problem is that the effects o f chronic imipramine were very similar to those of CMS. Thus, chronic treatment (5 weeks) with imipramine (5 mg/kg/day) normalized sucrose intake, but also increased electrically-stimulated DA release to an extent similar to that seen following CMS. Furthermore, imipramine did not reverse the increased DA release in CMS-exposed animals (90).

Postsynaptic Mechanisms The similarity in the presynaptic effects of imipramine and CMS suggests that changes in sensitivity to reward in this model are more likely to be mediated postsynaptically. Consistent with this hypothesis, a significant decrease in the number of D2 receptors in the nucleus accumbens was observed following prolonged (7 weeks) exposure to CMS (103). Functional evidence of postsynaptic receptor subsensitivity was provided by a series of experiments assessing rewarding and locomotor stimulant effects of the D2/D3 agonist, quinpirole (21,87), administered at postsynaptically active doses (100-400 #g/kg). Locomotor activity was assessed by the distance traversed in a 45-min session in a runway; rewarding effects were assessed using the place conditioning paradigm, with administration o f quinpirole on either the initially nonpreferred (white) or the initially preferred (black) side. In a further place preference experiment, quinpirole was administered directly within the nucleus accumbens (0.75 #g/side). In all experiments, responses to quinpirole were attenuated or abolished following CMS (63). It therefore appears that anhedonia in CMS-exposed animals results from subsensitivity of reward-related D2 receptors in the nucleus accumbens. Our working hypothesis is that this effect is secondary to a prolonged and persistent overexposure to DA, resulting from an increase in DA release. Interestingly, subsensitivity to DA agonists has also been reported in the hypocholinergic FSL strain, which show increased susceptibility to the anhedonic effect of CMS. A small decrease in the number of D2 receptors has also been observed in

530 FSL animals, though this effect was not replicated in a later study (15). Antidepressant drugs have traditionally been assumed to exert their clinical effects through an interaction with NA or 5HT systems. However, after chronic administration, antidepressants have also been found to potentiate the psychomotor stimulant effects of DA agonists, administered systemically or by direct injection into the nucleus accumbens (34,45,98). Strong evidence that antidepressants reverse CMS-induced anhedonia by potentiating DA transmission was provided by a series of studies in which the therapeutic response to tricyclic antidepressants was reversed by DA receptor antagonists. In these experiments, which were carried out following successful chronic treatment of CMS-exposed animals with antidepressants, DA receptor antagonists were administered acutely immediately prior to sucrose intake tests, at low doses that are without effect in untreated animals or in non-stressed animals. The effects of D1 (SCH-23390) and D2 (pimozide, sulpiride, raclopride) antagonists were similar: all decreased sucrose intake in antidepressant-treated stressed animals, but not, at low doses, in any other group (55,57,79). By contrast, the 5HT receptor antagonist metergoline, which reverses certain other actions of antidepressants (10,108), was without effect in stressed animals successfully treated with imipramine (57). The most striking finding from this series of experiments was that the therapeutic effects of fluoxetine and maprotiline in the CMS model, which act primarily as 5HT and NA reuptake inhibitors, respectively, were also reversed by acute administration of raclopride (100 #g/kg). This suggests that sensitization of D2 receptors within the nucleus accumbens may represent a final common pathway for the anti-anhedonic actions of antidepressant drugs (55), and could well explain why it is that so little progress has been made in identifying groups of depressed patients who respond preferentially to 5HT or NE uptake inhibitors.

Behavioural Sensitization: A New Treatment Strategy for Depression? The evidence that antidepressants reverse CMS-induced anhedonia by increasing DA receptor sensitivity in the nucleus accumbens led to a consideration of whether the same end might be achieved by other means. Although the mechanism is currently unknown, intermittent administration of low doses of DA agonists also causes an increase in sensitivity to their locomotor stimulant effects similar to that observed following chronic antidepressant treatment (48,74,77,105). The sensitizing effects of intermittent DA agonist administration and chronic antidepressant treatment are, to some extent, additive (65). Studies were therefore initiated to examine whether intermittent treatment with DA agonists could restore normal behaviour in CMS-treated animals. It was found that 3-4 doses of the D2/D3 agonists quinpirole or bromocriptine, administered at 3-4 day intervals, had effects that were in almost every respect comparable to those observed with tricyclic antidepressants. Sucrose consumption and place preference conditioning returned to normal in treated animals; the time course of these effects was similar to that seen in tricyclictreated animals; no effects were seen in non-stressed control animals; and recovery in treated animals was reversed by acute administration of raclopride (56,64). These data suggest that intermittent administration of DA agonists could represent a novel strategy for the treatment or the adjunctive treatment of depression, and provide a further example of the heuristic value of the CMS model.

WILLNER, MUSCAT AND P A P P OTHER ANIMAL MODELS OF ANHEDONIA In the final section of this review, three other animal models of anhedonia are outlined, and their relationships to the CMS model are examined.

Uncontrollable Shock Uncontrollable shock has been shown to cause a wide variety of subsequent behavioural abnormalities, one of which, an impairment of shock escape learning ("learned helplessness"), has been very widely discussed as an animal model of depression (22,44,83,97). In addition to the escape learning impairment, uncontrollable shock also causes many other resemblances to major depression: so much so that an animal exposed to it could reasonably be said to meet the DSM-III-R criteria (95), were it not for the fact that most of the abnormalities dissipate in considerably less time than the minimum 2 weeks required for this diagnosis. Among its many other effects, a single session of inescapable electric footshock, in mice, causes an impairment of responding for brain stimulation reward, which is not seen following exposure to comparable escapable shock. These studies have mainly used the rate of ICSS as the dependent variable, which is susceptible to a variety of nonspecific influences. However, a specific impairment of sensitivity to reward is indicated by the observation that the effects of inescapable shock are anatomically specific. ICSS elicited from the VTA (the origin of the mesolimbic dopamine (DA) projection), or from the nucleus accumbens or frontal cortex (two of its terminal fields) was suppressed by inescapable shock; however, ICSS elicited from the substantia nigra (the origin of the nigrostriatal DA projection) was unaffected. Furthermore, within the VTA, ICSS from certain regions was affected, while comparable rates of ICSS elicited from other regions were not (114,116,117). The duration of these effects is uncertain: They persist for at least 7 days, but studies using longer durations have not been reported. A long-term (7 weeks) suppression of locomotor activity has been reported following a single shock session, but this effect is unrelated to shock controllability, and so is clearly distinct from other learned helplessness phenomena (113). An important observation in this model is that uncontrollable electric shock has variable behavioural effects (most of which appear to be antidepressant-reversible) in different inbred mouse strains. To take an extreme example, in the C57BL/67 strain, exposure to inescapable shock severely impaired subsequent learning to escape shock, but had no effect on responding for brain stimulation reward, while the D B A / 2J strain showed exactly the opposite pattern of deficits (84,117). Although the pharmacology of shock-induced escape learning deficits has been extensively investigated (73,85), there has been relatively little investigation of the pharmacology of shock-induced anhedonia. Normal sensitivity to brain stimulation reward was restored by chronic, but not acute, treatment with the tricyclic antidepressant DMI, administered prophylactically or following exposure to inescapable shock; however, these effects varied considerably between strains and between stimulation sites (114,115). Related to the uncontrollable shock-ICSS model is the observation that a single session of restraint stress reduced saccharin preference relative to water. This effect was attenuated by an acute administration of a tricyclic antidepressant and a 5-HT uptake inhibitor; however, diazepam was also effective (71).

A REALISTIC ANIMAL MODEL OF DEPRESSION

Withdrawal From Psychomotor Stimulants A number of studies have reported that responding for ICSS was also reduced in the days following withdrawal from chronic amphetamine treatment. In these studies, amphetamine was administered to rats for between 4 and 10 days, typically using several administrations each day, at increasing doses (4,14,35,41,86). The threshold for ICSS was elevated following amphetamine withdrawal, confirming that the rate reduction reflects a subsensitivity to brain stimulation reward rather than a depression of motor activity (14,41). In a single pharmacological study, this effect was attenuated by subacute treatment with imipramine or amitriptyline, and completely reversed by continued treatment (36). In a variant of this procedure, animals self-administer cocaine, rather than having amphetamine administered to them (37). In these experiments, the threshold for brain stimulation reward, administered through electrodes in the posterior lateral hypothalamus, was obtained using a discrete trial procedure (39), which is sensitive to changes in reward value, but not to changes in motor performance (47). Following 24 h of cocaine self-administration, ICSS thresholds were elevated for several hours (37). Acute administration of the DA agonist bromocriptine restored ICSS thresholds to normal (46). Conventional antidepressants have not been tested in this paradigm. In addition to the very limited pharmacological evaluation of these procedures, their ecological validity is questionable: The effects are probably of considerable relevance to the stimulant-withdrawal "crash," but it is far less obvious that these procedures realistically model the etiology of major depression. In terms of physiological mechanism, it is plausible that stimulant-withdrawal anhedonia effects could result from a decrease in D2 receptor responsiveness in the nucleus accumbens secondary to a prolonged overstimulation of these receptors by elevated intrasynaptic concentrations of DA, as apparently occurs in the CMS model.

Neonatal Clomipramine The third method of inducing anhedonia is neonatal treatment with the tricyclic antidepressant clomipramine, which has been reported to cause a variety of abnormalities when the animals reach maturity, including subsensitivity to ICSS. This effect was present in 7-month-old animals, but was not seen when the animals were tested at 4 months (93). Clomipramine treated animals did not differ from controls in their 24-h consumption of sucrose or saccharin, monitored at monthly intervals, except for a decrease in saccharin intake at 7 months only. However, these data were not reported in detail, and it is uncertain whether the methods used were sufficiently sensitive to detect changes in sensitivity to reward (92). The fact that the saccharin result was seen at only a single time-point does raise a question of whether the impairment of ICSS responding seen in 7-month-old animals is similarly transient. It is unclear how, if at all, this model relates to the etiology of depression in humans. Information concerning the pharmacological responsiveness of the model is minimal (92), and the effects of antidepressant drugs on the anhedonic component have apparently not yet been studied. COMPARATIVE E V A L U A T I O N

In most respects, the CMS model compares favourably with other animal models of anhedonia.

531 Firstly, the pharmacological evaluation of CMS-induced anhedonia, albeit limited, has been more extensive than in the other three anhedonia models. Thus, CMS-induced anhedonia has been shown to respond to chronic therapeutic treatment with five antidepressants from three different classes, and not to respond to one non-antidepressant; the time course of these effects (3-5 weeks) closely resembles that seen in the clinic. Although studies of antidepressant effects on shock-induced escape learning deficits are now routine (73), antidepressant treatment of shock-induced anhedonia has only been demonstrated for DMI, administered subacutely (114,115). Similarly, there is only a single publication on the treatment of amphetamine-withdrawal anhedonia by tricyclic antidepressants (36), and, as yet, no information on the antidepressant treatment of anhedonia in either cocaine-withdrawn or neonatally clomipramine-treated animals. Secondly, the anhedonic effects of CMS are known to persist for a period of months, which both permits the demonstration of slow effects of antidepressants, and allows a variety of investigations to be carried out that cannot so readily be performed in models of briefer duration: for example, place conditioning following recovery (55). Reported durations are less than 3 weeks in the uncontrollable shock and stimulant withdrawal models: These effects may well last longer, but this information is currently lacking. In the neonatal clomipramine model, the impairment of ICSS was absent in 4month old rats but present at 7 months, and presumably, thereafter (93); on the other hand, an impairment of saccharin drinking was observed at 7 months, but not in either younger or older animals (92). Thirdly, of the four models reviewed, the CMS model arguably embodies the most realistic etiology. CMS is clearly a more realistic procedure in relation to major depression than stimulant withdrawal, and is probably more realistic than acute administration of uncontrollable shock; the direct clinical relevance of neonatal clomipramine treatment to depression is currently obscure. However, it would be wrong to see these different means of inducing anhedonia as being in any sense inconsistent or in competition. Rather, the concept of anhedonia provides a framework within which to investigate the interactions between different etiological factors, which already include genetic variables (75,114), epigenetic factors (92,93), acute major stressors (37,114), and chronic minor stressors (this review). An understanding of how these variables interact in the genesis of anhedonia would probably be of considerable relevance to the etiology of depression. Fourthly, CMS-induced anhedonia has been demonstrated in relation to a wide variety of rewards, rather than relying exclusively on ICSS. The relevance of ICSS impairments to the clinical phenomenon of anhedonia rests largely on the assumption that ICSS taps into a unified system that mediates and integrate the effects of diverse sources of reward. This view of ICSS forms a productive working hypothesis (9), but it is clearly an oversimplification (66). Unlike the other models, the claim that CMS treatment causes a generalized insensitivity to rewards is based on empirical evidence, rather than on assumptions about the nature of ICSS. The one area in which the CMS model is currently lacking is in its narrow focus on anhedonia. Behavioural studies of the CMS model have concentrated on this symptom because of its crucial significance in establishing the validity of an animal model of depression. However, in the uncontrollable shock and neonatal clomipramine models a wide range of other abnormalities reminiscent of depression have also been

532

WILLNER, MUSCAT AND PAPP

demonstrated (92,95); these models, therefore, have a syndroreal character that the CMS model presently does not. Finally, the implications of these studies for the neurophysiological basis of depressive disorders should not be overlooked. The CMS, uncontrollable shock, and stimulant withdrawal models all point to the mesolimbic D A system as a crucial element in anhedonia. In the CMS model, there is good evidence that antidepressant drugs reverse anhedonia by sensitizing D2 receptors; indeed, sensitization of D2 receptors appears to be sufficient for antidepressant action in this model (56,64). Some recent clinical evidence is consistent with a D A hypothesis of depression (17,51,52), but this hypothesis faces

a number of obstacles, not least o f which is the use o f D A receptor antagonists in the treatment of depression (59). Discussion of these paradoxes falls outside the scope of the present review [see (59,104,107)]. Even so, it is clear that studies that pursue these clinical implications will undoubtedly be o f value both for understanding the physiological basis of m o o d disorders, and in the further development of animal models of anhedonia. ACKNOWLEDGEMENTS The studies described in this review were partly supported by the Medical Research Council of Great Britain.

REFERENCES 1. Abramson, L. Y.; Sehgman, M. E. P. Modeling psychopathology in the laboratory: History and rationale. In: Maser, J. D.; Seligman, M. E. P., eds. Psychopathology: Experimental models. San Francisco: Freeman; 1978:1-26. 2. American Psychiatric Association. DSM III R-Diagnostic and Statistical Manual of Psychiatric Disorders. 3d. ed. APA, Washington, D.C.: American Psychiatric Association; 1987. 3. Aneshensel, C.; Stone, J. Stress and depression: A test of the buffering model of social support. Arch. Gen. Psychiatr. 39: 1392-1396; 1982. 4. Barrett, R. J.; White, D. K. Reward system depression following chronic amphetamine: Antagonism by haloperidol. Pharmacol. Biochem. Behav. 13:555-559; 1980. 5. Bellisle, F. Quantifying palatability in humans. Ann. N.Y. Acad. Sci. 575:363-374; 1989. 6. Billings, A.; Cronkite, R.; Moos, R. Social-environmental factors in unipolar depression: Comparisons of depressed patients and non-depressed controls. J. Abnorm. Psychol. 92:119-133; 1983. 7. Bizot, J. C.; Thiebot, M. H.; LeBihan, C.; Soubrie, P.; Simon, P. Effects of imipramine-like drugs and serotonin uptake blockers on delay of reward in rats: Possible implication in the behavioural mechanism of action of antidepressants. J. Pharmacol. Exp. Ther. 246:1144-1151; 1988. 8. Blanc, G.; Herve, D.; Simon, H.; Lisoprawski, A.; Glowinski, J.; Tassin, J. P. Response to stress of mesocortical frontal dopaminergic neurons in rats after long-term isolation. Nature 284: 265-276; 1980. 9. Bozarth, M. A. The mesolimbic dopamine system as a model reward system. In: Willner, P.; Scheel-Kruger, J., eds. The mesolimbic dopamine system: From motivation to action. Chichester: Wiley; 1991:301-330. 10. Broekkamp, C. L. E.; Garrigou, D.; Lloyd, K. G. Serotoninmimetic and antidepressant drugs on passive avoidance learning by olfactory bulbectomized rats. Pharmacol. Biochem. Behav. 13:643-646; 1980. 11. Brown, G. W. A psychosocial view of depression. In: Bennett, D. H.; Freeman, H., eds. Community psychiatry. London: Churchill-Livingstone; 1989:7 I-114. 12. Brown, G. W.; Harris, T. Social origins of depression. London: Tavistock; 1978. 13. Carr, G. D.; Fibiger, H. C.; Phillips, A. G. Conditioned place preference as a measure of drug reward. In: Liebman, J. M.; Cooper, S. J., eds. The neuropharmacological basis of reward. Oxford: Oxford University Press; 1989:264-319. 14. Cassens, G. P.; Actor, C.; Kling, M.; Schildkraut, J. J. Amphetamine withdrawal effects threshold of intracranial selfstimulation. Psychopharmacology 73:318-322; 1981. 15. Crocker, A. D.; Overstreet, D. H. Changes in dopamine sensitivity in rats selectively bred for difference in cholinergic function. Pharmacol. Biochem. Behav. 38:105-108; 1991. 16. Davis, J. D. The microstructure of ingestive behavior. Ann. N.Y. Acad. Sci. 575:106-119; 1989. 17. Del Zompo, M.; Boccheta, A.; Bernardi, F.; Burrai, C.; Corsini, G. U. Clinical evidence for a role of dopaminergic sys-

18. 19. 20. 21.

22.

23. 24. 25.

26.

27. 28. 29. 30. 31. 32.

33. 34.

tems in depressive syndromes. In: Gessa, G. L.; Serra, G., eds. Dopamine and mental depression. New York: Pergamon Press; 1990:177-184. Dunn, A. J. Stress-related activation of cerebral dopaminergic systems. Ann. N.Y. Acad. Sci. 537:188-205; 1988. Ellenbroek, B. A.; Cools, A. R. Animal models with construct validity for schizophrenia. Behav. Pharmacol. 1:469-490; 1990. Feather, N.; Barber, J. Depressive reactions and unemployment. J. Abnorm. Fsychol. 2:185-195; 1983. Fuller, R. W.; Hemrick-Luecke, S. K. Decrease in hypothalamic epinephrine concentration and other neurochemical changes produced by quinpirole, a dopamine agonist, in rats. J. Neural Transm. 61:161-173; 1985. Gather, J.; Miller, W. R.; Seaman, S. F. Learned helplessness, stress and the depressive disorders. In: Depue, R. A., ed. The psychobiology of the depressive disorders: Implications for the effects of stress. New York: Academic Press; 1979:335-363. Garzon, J.; Fuentes, J. A.; Del Rio, J. Antidepressants selectively antagonize the hyperactivity induced in rats by long-term social isolation. Eur. J. Pharmacol. 59:294-296; 1979. Giles, D. E.; Roffwarg, H. P.; Schlesser, M. A.; Rush, A. J. Which endogenous depressive symptoms relate to REM latency reduction? Biol. Psychiatr. 21:473-482; 1986. Goudie, A. Animal models of drug abuse and dependence. In: Willner, P., ed. Behavioural models in psychopharmacology: Theoretical, industrial and clinical perspectives. Cambridge: Cambridge University Press; 1991:453-484. Green, S.; Hodges, H. Animal models of anxiety. In: Willner, P., ed. Behavioural models in psychopharmacology: Theoretical, industrial and clinical perspectives. Cambridge: Cambridge University Press; 1991:21-49. Henn, F. A.; McKinney, W. T. Animal models in psychiatry. In: Meltzer, H. Y., ed. Psychopharmacology: The third generation of progress. New York: Raven; 1987:697-704. Jahoda, M. The impact of unemployment in the 1930s and the 1970s. Bull. Br. Psychol. Soc. 32:309-314; 1979. Kalivas, P. W.; Abhold, R. Enkephalin release into the ventral tegmental area in response to stress: modulation of mesocorticolimbic dopamine. Brain Res. 414:339-348; 1987. Kanner, A. D.; Coyne, J. C.; Schaefer, C.; Lazarus, R. S. Comparison of two modes of stress measurement: Daily hassles and uplifts versus major life events. J. Behav. Med. 4:1-39; 1981. Katz, R. J. Animal model of depression: Pharmacological sensitivity of a hedonic deficit. Pharmacol. Biochem. Behav. 16:965968; 1982. Katz, R. J.; Sihel, M. Animal model of depression: Tests of three structurally and pharmacologically novel antidepressant compounds. Pharmacol. Biochem. Behav. 16:979-982; 1982. Katz, R. J.; Roth, K. A.; Carroll, B. J. Acute and chronic stress effects on open field activity in the rat: Implications for a model of depression. Neurosci. Biobehav. Rev. 5:259-264; 1981. Klimek, V.; Maj, J. Repeated administration of antidepressant drugs enhanced agonist affinity for mesolimbic D-2 receptors. J. Pharm. Pharmacol. 41:555-558; 1989.

A REALISTIC ANIMAL MODEL OF DEPRESSION 35. Kokkinidis, L.; Zacharko, R. M. Response sensitization and depression following long-term amphetamine treatment in a selfstimulation paradigm. Psychopharmacology 68:73-76; 1980. 36. Kokkinidis, L.; Zacharko, R. M.; Predy, P. A. Postamphetamine depression of self-stimulation responding from the substantia nigra: Reversal by tricyclic antidepressants. Pharmacol. Biochem. Behav. 13:379-383; 1980. 37. Koob, G. F. Anhedonia as an animal model of depression. In: Koob, G. F.; Ehlers, C. L.; Kupfer, D. J., eds. Animal models of depression. Boston: Birkhauser; 1989:162-183. 38. Koob, G. F.; Goeders, N. E. Neuroanatomical substrates of drug self-administration. In: Liebmann, J. M.; Cooper, S. J., eds. The neuropharmacological basis of reward. Oxford: Oxford University Press; 1989:214-263. 39. Kornetsky, C.; Esposito, R. U. Reward and detection thresholds for brain stimulation: Dissociative effects of cocaine. Brain Res. 209:496-500; 1981. 40. Kupfer, D. J.; Thase, M. E. The use of the sleep laboratory in the diagnosis of affective disorders. Psychiatr. Clin. N. Am. 6: 3-25; 1983. 41. Leith, N. J.; Barrett, R. J. Amphetamine and the reward system: Evidence for tolerance and post-drug depression. Psychopharmacology 46:19-25; 1976. 42. Lewinsohn, P. M. A behavioral approach to depression. In: Friedman, R. J.; Katz, M. M., eds. The psychology of depression: Contemporary theory and research. New York: Wiley; 1974: 157-185. 43. Lloyd, C. Life events and depressive disorders reviewed. Arch. Gen. Psychiatr. 37:529-548; 1980. 44. Maier, S. F. Learned helplessness and animal models of depression. Prog. Neuropsychopharmacol. Biol. Psychiatr. 8:435--446; 1984. 45. Maj, J. Behavioral effects of antidepressant drugs given repeatedly on the dopaminergic system. In: Gessa, G. L.; Serra, G., eds. Dopamine and mental depression. New York: Pergamon Press; 1989:39-146. 46. Markou, A.; Koob, G. F. Bromocriptine reverses post-cocaine anhedonia in a rat model of cocaine withdrawal. Am. Coll. Neuropsychopharmacology Abstr. 157; 1989. 47. Markou, A.; Hanley, S. J.; Chehade, A. K.; Koob, G. F. Effects of performance and reward manipulations on current-intensity thresholds and other measures derived from a discrete-trial selfstimulation procedure in rats. Soc. Neurosci. Abstr. 15:35; 1989. 48. Mattingley, B. A.; Gotsick, J. E. Conditioning and experiential factors affecting the development of sensitization to apomorphine. Behav. Neurosci. 103:1311-1317; 1989. 49. McKinney, W. T.; Bunney, W. E. Animal model of depression: Review of evidence and implications for research. Arch. Gen. Psychiatr. 21:240-248; 1969. 50. Moreau, J.-L.; Jenck, F.; Martin, J. R.; Mortas, P.; Haefely, W. E. Antidepressant treatment prevents chronic unpredictable mild stress-induced anhedonia as assessed by ventral tegmental self-stimulation behavior in rats. Eur. Neuropsychopharmacol. 2:43--49; 1992. 51. Mouret, J.; LeMoine, P.; Minuit, M. Marqueurs polygraphiques, cliniques et th~rapeutiques des depressions dopaminod~pendantes (DDD). Comptes Rendus, Serie III, 301-306; 1987. 52. Mouret, J.; LeMoine, P.; Minuit, M.; Robelin, N. La L-tyrosine gu&it, immediatement et a long terme, les depressions dopamino-d~pendantes (DDD). l~tude clinique et polygraphique. Comptes Rendus, Serie III, 93-98; 1988. 53. Muscat, R.; Willner, P. Suppression of sucrose drinking by chronic mild unpredictable stress: A methodological analysis. Neurosei. Biobehav. Rev. 16:507-517; 1992. 54. Muscat, R.; Kyprianou, T.; Osman, M.; Phillips, G.; Willner, P. Sweetness-dependent facilitation of sucrose drinking by raclopride is unrelated to calorie content. Pharmacol. Biochem. Behay. 40:209-213; 1992. 55. Muscat, R.; Papp, M.; Willner, P. Reversal of stress-induced anhedonia by the atypical antidepressants, fluoxetine and maprotiline. Psychopharmacology; in press.

533 56. Muscat, R.; Papp, M.; Willner, P. Antidepressant-like actions of dopamine agonists in an animal model of depression. Biol. Psychiatr. 31:937-946; 1992. 57. Muscat, R.; Sampson, D.; Willner, P. Dopaminergic mechanism of imipramine action in an animal model of depression. Biol. Psychiatr., 28:223-230; 1990. 58. Muscat, R.; Towell, A.; Willner, P. Changes in dopamine autoreceptor sensitivity in an animal model of depression. Psychopharmacology 94:545-550; 1988. 59. Nelson, J. C. The use of antipsychotic drugs in the treatment of depression. In: Zohar, J.; Belmaker, R. H., eds. Treating resistant depression. New York: PMA Corp.; 1987:131-146; 1987. 60. Nelson, J.; Charney, D. S. The symptoms of major depression. Am. J. Psychiatr. 138:1-13; 1981. 61. Overstreet, D. H.; Janowsky, D. S. A cholinergic supersensitivity model of depression. In: Boulton, A.; Baker, G.; MartinIverson, M. Neuromethods. vol. 20. Animal Models in Psychiatry. Basel: Birkhauser; 1992:81-114. 62. Papp, M.; Lappas, S.; Muscat, R.; Willner, P. Attenuation of place preference conditioning but not place aversion conditioning by chronic mild stress. J. Psychopharmacol. 6:352-356; 1992. 63. Papp, M.; Muscat, R.; Willner, P. Subsensitivity to rewarding and locomotor stimulant effects of a dopamine agonist following chronic mild stress. Psychopharmacology (Bed.) (in press). 64. Papp, M.; Muscat, R.; Willner, P. Behavioural sensitization to a dopamine agonist is associated with recovery from stressinduced anhedonia. Psychopharmacology; in press. 65. Papp, M.; Muscat, R.; Willner, P. Additive effects of chronic treatment with antidepressant drugs and intermittent treatment with a dopamine agonist. Eur. Neuropsyehopharmacol. 2:121125; 1992. 66. Papp, M.; Willner, P.; Muscat, R. An animal model of anhedonia: Attenuation of sucrose consumption and place preference conditioning by chronic unpredictable mild stress. Psychopharmacology 104:255-259; 1991. 67. Phillips, A. G.; Fibiger, H. C. Neuroanatomical bases of intracranial self-stimulation: Untangling the Gordian knot. In: Liebmann, J. M.; Cooper, S. J., eds. The neuropharmacological basis of reward. Oxford: Oxford University Press; 1989:66-105. 68. Phillips, G.; Muscat, R.; Willner, P. Different anatomical substrates for neuroleptie-induced response decrement and neuroleptic-induced anhedonia. Behav. Pharmacol. 2:129-141; 1991. 69. Phillips, G.; Willner, P.; Muscat, R. Reward-dependent suppression or facilitation of consummatory behaviour by raclopride. Psychopharmacology 105:355-360; 1991. 70. Pignatiello, M. F.; Olson, G. A.; Kastin, A. J.; Ehrensing, R. H.; McLean, J. H.; Olson, R. D. MIF-I is active in a chronic stress animal model of depression. Pharmacol. Biochem. Behav. 32:737-742; 1989. 71. Plaznik, A.; Stefanski, R.; Kostowski, W. Restraint stressinduced changes in saccharin preference: the effect of antidepressant treatments and diazepam. Pharmacol. Biochem. Behav. 33:755-759; 1989. 72. Porsolt, R. D.; Bertin, A.; Jalfre, M. Behavioral despair in rats and mice: Strain differences and the effects of imipramine. Eur. J. Pharmacol. 51:291-294; 1978. 73. Porsolt, R. D.; Len6gre, A.; MeArthur, R. A. Pharmacological models of depression. In: Olivier, B.; Mos, J.; Slangen, J. L., eds. Animal models in psychopharmacology. Basel: Birkhauser; 1991:137-159. 74. Post, R. M.; Weiss, S. R. B.; Pert, A. Animal models of mania. In: Willner, P.; Scheel-Kruger, J., eds. The mesolimbic dopamine system: From motivation to action. Chichester: Wiley; 1991:273-299. 75. Pucilowski, O.; Overstreet, D. S.; Rezvani, A.; Janowsky, D. S. Effects of acute and chronic stressors on saccharin preference in hypercholinergic rats. Behav. Pharmacol. 3(Suppl. 1):50; 1992. 76. Rickels, K.; Schweizer, E. E. Current pharmacotherapy of anxiety and panic. In: Meltzer, H. Y., ed. Psychopharmacology:

534

77.

78.

79. 80. 81. 82.

83. 84. 85. 86. 87.

88. 89. 90.

91. 92. 93. 94. 95.

96. 97. 98.

WILLNER, MUSCAT AND PAPP The third generation of progress. New York: Raven; 1987:11931204. Robinson, T. E.; Becker, J. B. Enduring changes in brain and behavior produced by chronic amphetamine administration: A review and evaluation of animal models of amphetamine psychosis. Brain Res. Rev. 11 : 157-198; 1986. Sampson, D.; Muscat, R.; Phillips, G.; Willner, P. Decreased reactivity to sweetness following chronic exposure to mild unpredictable stress or acute administration of pimozide. Neurosci. Biobehav. Rev. 16:519-524; 1992. Sampson, D.; Muscat, R.; WiNner, P. Reversal of antidepressant action by dopamine antagonists in an animal model of depression. Psychopharmacology 104:491-495; 1991. Sarter, M.; Hagan, J.; Dudchenko, P. Behavioral screening for cognition enhancers: From indiscriminate to valid testing. Psychopharmacology; in press. Sclafani, A.; Clyne, A. E. Hedonic response of rats to polysaccharide and sugar solutions. Neurosci. Biobehav. Rev. 11:173180; 1987. Selden, L. S.; Dahms, J. L.; Shaughnessy, R. A. Behavioral screen for antidepressants: The effects of drugs and electroconvulsive shock on performance under a differential-reinforcement-of-low-rate schedule. Psychopharmacology 86:55-60. Seligman, M. E. P. Helplessness: On depression, development and death. San Francisco: Freeman; 1975. Shanks, N.; Anisman, H. Stressor-provoked behavioral changes in six strains of mice. Behav. Neurosci. 102:894-905; 1988. Sherman, A. D.; Saquitne, J. L.; Petty, F. Specificity of the learned helplessness model of depression. Pharmacol. Biochem. Behav. 16:449-454; 1982. Simpson, D. M.; Annau, Z. Behavioural withdrawal following several psychoactive drugs. Pharmacol. Biochem. Behav. 7: 5964; 1977. Sokoloff, P.; Giros, B.; Martres, M.-P.; Bouthenet, M.-L.; Schwartz, J.-C. Molecular cloning and characterization of a novel dopamine receptor (D3) as a target for neuroleptics. Nature 347:146-151; 1990. Soubrie, P. Reconciling the role of central serotonin neurons in human and animal behavior. Behav. Brain Sci. 9:319-364; 1986. Spyraki, C.; Fibiger, H. C.; Phillips, A. G. Attenuation by haioperidol of place preference conditioning using food reinforcement. Psychopharmacology 77:379-382; 1982. Stamford, J. A.; Muscat, R.; O'Connor, J. J.; Patel, J. J.; Wieczorek, W. J.; Kruk, Z. L.; WiUner, P. Voltammetric evidence that subsensitivity to reward following chronic mild stress is associated with increased release of mesolimbic dopamine. Psychopharmacology 105:275-282; 1991. Thierry, A. M.; Tassin, J. P.; Blanc, G.; Glowinski, J. Selective activation of the mesocortical dopamine system by stress. Nature 263:242-244; 1976. Vogel, G.; Neill, D.; Hagler, M.; Kors, D. A new animal model of endogenous depression: A summary of present findings. Neurosci. Biobehav. Rev. 14:85-91; 1991. Vogel, G. W.; Neill, D.; Hagler, M.; Kors, D. Decreased intracranial self-stimulation in a new animal model of endogenous depression. Neurosci. Biobehav. Rev. 14:65-68; 1991. Weingarten, H. P.; Watson, S. D. Sham feeding as a procedure for assessing the influence of diet palatability on food intake. Physiol. Behav. 28:401-407; 1982. Weiss, J. M.; Bailey, W. H.; Goodman, P. A.; Hoffman, L. J.; Ambrose, M. J.; Salman, S.; Charry, J. M. A model for the neurochemical study of depression. In: Spiegelstein, M. Y.; Levy, A., eds. Behavioral models and the analysis of drug action. Amsterdam: Elsevier; 1982:195-223. Willner, P. The validity of animal models of depression. Psychopharmacology 83:1-16; 1984. Willner, P. Validating criteria for animal models of human mental disorders: Learned helplessness as a paradigm case. Progress in Neuropsychopharmacol. Biol. Psychiatr. 10:677-690; 1986. Willner, P. Sensitization to the actions of antidepressant drugs.

99.

100. 101. 102. 103. 104.

105. 106.

107.

108.

109.

110. I 11. 112. 113.

114.

115. 116.

117.

118.

In: Emmett-Oglesby, M. V.; Goudie, A. J., eds. Psychoactive drugs: Tolerance and sensitization. Clifton, N J: Humana; 1989: 407--459. Willner, P. Towards a theory of serotonergic dysfunction in depression. In: Bevan, P.; Cools, A.; Archer, T., eds. Behavioural pharmacology of 5-HT. Hillsdale, N J: Erlbaum; 1989: 157-178. Willner, P. Animal models of depression: An overview. Pharmacol. Ther. 45:425-455; 1990. Willner, P. Animal models as simulations of depression. Trends Pharmacol. Sci. 12:131-136; 1991. Wiliner, P.; Scheel-Kruger, J., eds. The mesolimbic dopamine system: From motivation to action. Chichester: Wiley; 1991. Willner, P.; Klimek, V.; Golembiowska, K.; Muscat, R. Changes in mesolimbic dopamine may explain stress-induced anhedonia. Psychobiology 19:79-84; 1991. Willner, P.; Muscat, R.; Papp, M.; Sampson, D. Dopamine, depression and antidepressant drugs. In: Willner, P.; ScheelKruger, J., eds. The mesolimbic dopamine system: From motivation to action. Chichester: Wiley; 1991:385-408. Willner, P.; Papp, M.; Cheeta, S.; Muscat, R. Environmental influences on behavioural sensitization to the dopamine agonist quinpirole. Behav. Pharmacol. 3:43-50; 1992. Wiliner, P.; Phillips, G.; Muscat, R. Suppression of rewarded behaviour by neuroleptic drugs: Can't or won't, and why? In: Willner, P.; Scheel-Kruger, J., eds. The mesolimbic dopamine system: From motivation to action. Chichester: Wiley; 1991: 251-271. Willner, P.; Sampson, D.; Papp, M.; Phillips, G.; Muscat, R. Animal models of anhedonia. In: Soubrie, P., ed. Animal models of psychiatric disorders, vol. 3. Anxiety, Depression and Mania. Basel: Karger; 1991:71-99. Willner, P.; Sampson, D.; Phillips, G.; Fichera, R.; Foxlow, P.; Muscat, R. Effects of isolated housing and chronic antidepressant treatment on cooperative social behaviour in rats. Behay. Pharmacol. 1:85-90; 1989. Wiliner, P.; Towell, A.; Sampson, D.; Sophokleous, S.; Muscat, R. Reduction of sucrose preference by chronic mild stress and its restoration by a tricyclic antidepressant. Psychopharmacology 93:358-364; 1987. Willner, P.; Wilkes, M.; Orwin, O. Attributional style and perceived stress in endogenous and reactive depression. J. Affect. Dis. 18:281-287; 1990. Wise, R. A. Neuroleptics and operant behaviour: The anhedonia hypothesis. Behav. Brain Sci. 5:39-87; 1982. Wise, R. A. The brain and reward. In: Liebmann, J. M.; Cooper, S. J., eds. The neuropharmacological basis of reward. Oxford: Oxford University Press; 1989:377-424. Woodmansee, W. W.; Silbert, L. H.; Maier, S. F. Stressinduced changes in daily activity in the rat are modulated by different factors than are stress-induced escape-learning deficits. Soc. Neurosci. Abstr. 17:146; 1991. Zacharko, R. M.; Anisman, H. Stressor-provoked alterations of intracranial self-stimulation in the mesocortiolimbic dopamine system: An animal model of depression. In: Willner, P.; ScheelKruger, J., eds. The mesolimbic dopamine system, from motivation to action. Chichester: Wiley; 411-442; 1991. Zacharko, R. M.; Bowers, W. J.; Anisman, H. Responding for brain stimulation: Stress and desmethylimipramine. Progress in Neuropsychopharmacol. Biol. Psychiatr. 8:601-606; 1984. Zacharko, R. M.; Bowers, W. J.; Kokkinidis, L.; Anisman, H. Region specific reductions of intracranial self-stimulation after uncontrollable stress: Possible effects on reward processes. Behav. Brain Res. 9:129-141; 1983. Zacharko, R. M.; Lalonde, G. T.; Kasian, M.; Anisman, H. Strain specific effects of inescapable shock on intracraniai selfstimulation from the nucleus accumbens. Brain Res. 426:164168; 1987. Zimmerman, M.; Spitzer, R. L. Melancholia: From DSM-III to DSM-III-R. Am. J. Psychiatr. 146:20-28; 1989.