The failure of the antidepressant drug discovery process is systemic

Abstract

Current antidepressants are crude compared with the ideal and patents on most have expired. There are therefore strong clinical and commercial pressures for new drugs to replace them. The prospects for this are, however, now markedly reduced as several major pharmaceutical companies have abandoned work in this area whilst many others have sharply decreased their research investment. These changes and the lack of progress over such a long period are indicative of a catastrophic systems failure which, it is argued, has been caused in large part by a logical flaw at the animal modelling stage. This tautology has served to lock the current antidepressant drug discovery process into an iterative loop capable only of producing further variations of that which has gone before. Drugs produced by this approach have proved to be only poorly effective in the context of the clinically depressed population as a whole. Hence, the inevitable failure of the current antidepressant drug discovery process has left little behind that can be salvaged. Therefore, it is suggested that this be urgently reformulated on more rational grounds using more appropriate species in new animal models based upon a thorough understanding of the behavioural expressions of depression in the clinic.

Keywords

Animal models antidepressant depression drug discovery MDD

Introduction

Depression is a widespread disorder that accounts for approximately 12% of the total burden of non-fatal global disease (Üstün et al., 2004). In the developed world around 25% of people can expect to experience this at some point in their lives (Blazer, 2000; WHO, 2002). Females are up to three times more vulnerable than males (Burt and Stein, 2002). First line treatments for depression are mostly drugs based (Hollon et al., 2002), with antidepressants being the third most commonly prescribed medication in the United States (Mojtabai and Olfson, 2011).

Whilst their use is on an upward trajectory throughout the developed world (Lin et al., 2011) the utility of these drugs is nonetheless restricted because not all patients respond to them (e.g. Bielski and Friedel, 1976; Brown et al., 1994; Fava and Davidson, 1996; Moncrieff, 2007). Indeed, these response rates may be as low as 40–50% (e.g. Trivedi et al., 2006), even after several years of treatment (e.g. Colman et al., 2011). This is not greatly higher than the response rate to placebo (Walsh et al., 2002). A further cause for concern are meta-analyses including negative findings submitted to the FDA but not published in the scientific literature that show that even these low figures are exaggerated by publication bias (Piggot et al., 2010; Turner et al., 2008).

There are reviews indicating that small numbers of severely (but not mild or moderately) depressed patients respond to these drugs (e.g. Fournier et al., 2010) but these must be viewed in the light of this publication bias. There is also a marked tendency for the strength of reported drug effects to be related to the age of the publication (Papakostas and Fava, 2009), which has been suggested to reflect the larger amount spent on advertising whilst these drugs were still in patent (Lacasse and Leo, 2005; Leo and Lacasse, 2008).

In addition to concerns about the large number of patients who cannot be treated using these drugs, currently available antidepressants are crude compared with the ideal and have several very serious shortcomings. Relapse rates are high (e.g. Geddes et al., 2003), there is a dangerous ‘therapeutic lag’ of several weeks (e.g. Skolnick, 1999) that leaves even those who do eventually respond feeling temporarily even more depressed (Jick et al., 2004; Möller and Volz, 1996) and side effects remain an issue even though these are reduced in second generation antidepressants compared with first (e.g. Dording et al., 2002).

In consequence official policies no longer recognise drugs-based antidepressant therapies as the panaceas they once did (e.g. Hughes and Cohen, 2009; Middleton et al., 2005; NICE, 2004) and significant numbers of depressives reject these treatments in favour of ‘self-medication’, with alcohol and/or other non-prescription drugs (e.g. Garland et al., 2012; Hendrie et al., 1998).

There are therefore strong clinical and ethical pressures towards the development of a new generation of antidepressants and the commercial rewards for doing so are potentially enormous. It is deeply paradoxical then that so many of the major players in the pharmaceutical industry have declared that they are no longer going to vigorously pursue this goal (Miller, 2010). The aim of the present paper is to explore the reasons why this hiatus has occurred and ways in which this knowledge can be used to help inform efforts to get the antidepressant drug discovery process moving forward once again.

Dissecting the process

Animal models

The drug discovery process seeks to produce drugs that are both efficacious and safe (Paul et al., 2010). Whilst clinical trials are often given a higher profile (e.g. Lipsky and Sharp, 2001) this fails to take into account the pivotal role animal models play in the process leading up to this phase. That is, preclinical models provide information that helps determine whether drugs should move into Phase 1 clinical trials (Herrling, 2005; Verkman, 2004) and, where appropriate, fulfil the regulatory obligation for efficacy to be shown before they are allowed to do so (e.g. Woodcock and Woosley, 2008). That said, the utility of animal models of depression is often not immediately obvious to non-specialists because they tend not to be based on a direct mapping of homologous responses between species (as would be the case with, say, hypertension), their end points often bear no resemblance to the disorder they are seeking to model (see below) and the level of translation between these models and clinical efficacy is low (Harmer et al., 2011).

In this context, animal models of depression should ideally be based on a common aetiology in both animals and man. They are not and so ‘pharmacological validity’ has been used as the defining characteristic instead. That is, all potential new drugs are measured against existing compounds in models that have been developed solely on the basis of their sensitivity to these standards (e.g. Kudryavtseva et al., 1991; Porsolt et al., 1977; Song and Leonard, 2005; Steru et al., 1985; Willner, 1997). These models have in consequence remained mechanistic and are now frequently referred to as ‘assays’ in recognition of their lack of any obvious relationship with clinical depression (e.g. Krishnan and Nestler, 2008).

There are a plethora of such models (for reviews see McArthur and Borsini, 2006; Willner, 1984, 1997) and one of the main reasons for this is that it is impossible to reject particular tests once they have satisfied the ‘sensitivity to known antidepressants’ criterion. This point also extends to those more recent genetically based models (e.g. Cryan and Holmes, 2005; Cryan and Mombereau, 2004). Hence, use of these assays has in recent times become a matter of preference and taste rather than utility, since they have all been developed on essentially the same definitional lines.

Commonly used models of depression include tail suspension (Cryan et al., 2005; Steru et al., 1985), forced swim (Porsolt et al., 1977), mouse killing (e.g. Song and Leonard, 2005), chronic social defeat (e.g. Keeney and Hogg, 1999; Kudryavtseva et al., 1991) and chronic mild stress (e.g. Monleon et al., 1995; Willner et al.,1987).

Logical flaw

Attempting to identify new compounds as potential antidepressants by testing them in animal models of depression that have been defined as such solely on the basis of their sensitivity to existing antidepressants is logically flawed, a tautology (of the form A=B therefore B=A). Thus the argument ‘it is an antidepressant because it has effects in animal models of depression that are animal models of depression because they are sensitive to the actions of antidepressants’ is in context no different in form from the statement ‘God exists because the Bible says he does and the Bible is the infallible word of God’.

This basic error was built into the current antidepressant drug discovery process from its inception (e.g. Chessin et al., 1956; Loomer et al., 1957) and the problems inherent with this approach have become compounded at each stage. Hence the current situation, where first generation antidepressants put on the market before FDA regulations required efficacy to be shown (Goodrich, 1963) were used as templates for a second generation of drugs that we now know to be only poorly effective in the context of the clinically depressed population as a whole (Colman, 2011; Piggot et al., 2010; Trivedi et al., 2006).

This situation has become further confounded in more recent times by the propensity to blur the conceptual differences between ‘stress’ and ‘depression’ and to erroneously use these terms more or less interchangeably (Dias et al., 2009). Further, many of the models currently in use are open to alternative explanations for effects seen in them. For example the end-points used to indicate antidepressant action in the forced swim test could also be the product of effects on learning and memory (e.g. West, 1990). Effects on sucrose consumption in the chronic mild stress model can be accounted for by the food deprivation component of this procedure alone (Forbes et al., 1996) and chronic social stress models where animals are exposed to attack from an aggressive conspecific (e.g. Keeney and Hogg, 1999) could equally be measuring factors that influence the activation of the endorphin system (e.g. Rodgers and Hendrie, 1983; Siegfried et al., 1987) and/or the well documented analgesic actions of the antidepressants themselves (e.g. Micó et al., 2006; Wattiez et al., 2011)

In consequence, the vast majority of researchers in this area operate on the de facto working definition of depression as ‘that which responds to existing antidepressants’ (e.g. Cryan et al., 2002; Cryan and Holmes, 2005; Deussing, 2006; Kelly et al., 1997; Krishnan and Nestler, 2008; Samuels et al., 2011; Song and Leonard, 2005; Willner, 1984; Wong and Licinio, 2004) in any species, including yeast (e.g. Chen et al., 2012) and regardless of the mechanisms underlying these effects or the end-points employed.

False assumptions

If the first flawed premise of the current antidepressant drug discovery process is that existing drugs are suitable templates on which to base the development of new treatments, the second is that rats and mice are suitable species on which to (almost exclusively) base the animal models these drugs are tested in (e.g. Cryan and Holmes, 2005; Krishnan and Nestler, 2008; Porsolt et al., 2001).

Were the antidepressant drug discovery process to have been developed along rational lines, rats and mice would have been selected because of special characteristics that make them particularly suitable for use in these models. They do not, however, have such characteristics (Dias et al., 2009; Hendrie and Pickles, 2009) and there are indeed strong arguments against their continued use in this context (Hendrie et al., 2011; Hendrie and Pickles, 2009). Further, little consideration has been given to species differences between rats and mice except for perhaps, occasionally, their size. Hence there are rat (Detke et al., 1995) and mouse (Porsolt et al., 1977) forced swim tests; rat (Rygula et al., 2005) and mouse (Keeney and Hogg, 1999) social defeat models; rat (Willner et al., 1987) and mouse (Yalcin et al., 2005) chronic unpredictable mild stress models and rat (Chermat et al., 1986), mouse (Steru et al., 1985) and gerbil (Varty et al., 2003) tail suspension tests, etc.

Table 1.

False assumptions underlying the current antidepressant drug discovery process. The resulting practices have led to inappropriate species being used in animal models that have been developed solely on the basis of their sensitivity to drugs that are now known to be only poorly effective in the context of the clinically depressed population as a whole. This tautological reasoning locked the drug discovery process into an iterative loop capable only of producing further variations of that which had gone before and made systems failure inevitable.

Assumption	Practice	Issue	Consequence
Existing drugs are effective	Animal models developed on basis of sensitivity to existing drugs	Existing drugs not effective in 50–60% of depressed population	Candidate drugs struggle to outperform placebo
Animal models must be sensitive to existing drugs	Animal models defined by sensitivity to existing drugs	Logically flawed Tautology	Process locked into iterative loop Capable only of producing more ‘me-too’s
Depression can be studied in any species	Heavy reliance on rats and mice	Depression most likely species-specific and not present in rats or mice	Rats and mice not suitable species for use in this context

There are nonetheless important differences between rats and mice and their significance cannot be ignored. Laboratory rats are derived from Rattus norvegicus. These are colonial animals that live together relatively peaceably under laboratory conditions (e.g. Barnett, 1963; Rodgers and Hendrie, 1982), within which significant levels of full-blooded aggression are seen under only the most unusual of conditions (e.g. Rodgers et al., 1983). Laboratory mice (Mus musculus) are by contrast strongly territorial and show high levels of intra-male aggression whenever they are group housed. Cages of male mice (the sex that is most frequently used) are in consequence a mixture of dominant and subordinate animals that vary greatly in terms of behaviour, physiology and immune functioning (Berry, 1970; Brain, 1971; Hendrie et al., 1996) and this is, at least, a major source of variance in all studies using this species.

Rats and mice are, it seems, used extensively in the field of depression research because there has been no serious consideration that things should be otherwise (e.g. Borsini, 2012; Hendrie et al., 2011; Nestler and Hyman, 2010) and the lumping together of findings made using rats, mice and/or other species as ‘rodent models’ (e.g. Castagné et al., 2011; Duman, 2010; El Yacoubi and Vaugeois, 2007; Krishnan and Nestler, 2011; McArthur and Borsini, 2006; Nestler and Hyman, 2010) ignores the potential importance of these species differences. This is also a very high risk strategy that assumes depression to be a general mammalian feature that can be modelled in any animal of this class.

Systems failure

The ease with which ‘me-same’ compounds could be produced in the 1960s, ’70s and to a certain extent the ’80s (e.g. Domino, 1999; Geddes et al., 2000) once gave this flawed antidepressant drug discovery process an aura of rationality. However, serendipity remained at its base, the illusion could not be maintained indefinitely and the paucity of this approach has now been fully exposed. The glacial rate of progress over the past 60 years has provided improvements in tolerability but not efficacy (Hindmarch, 2001). The most frequently prescribed antidepressants are now more than 20 years old and out of patent (Ciprani et al., 2009). There is little in the pipeline to replace them and only poor prospects of there being progress in the near future now that many of the world’s leading pharmaceutical companies have to a greater or lesser degree reduced their investment in this area (Blier, 2010). There has been what can only be described as a catastrophic systems failure and this has left little behind that can be salvaged.

Taking stock

With the benefit of hindsight it is apparent that even a cursory consideration of the species used in these models would have raised serious questions about the models themselves. Asking ‘what species?’ inevitably raises the question ‘why?’ and any rational answer to that demands a level of understanding about depression itself that we do not currently possess. The emphasis would thus have been on developing that understanding of depression rather than continuing down the line of making high levels of investment in drug-led approaches that have not proved successful. The result of this omission has been an antidepressant drug discovery process that is tautologically locked into an iterative loop capable only of producing further variants of that which has gone before. This error has become critical now the proportion of the clinical population that cannot be treated with these drugs has become clear. There can be no more potent demonstration of the moribund state of the current antidepressant drug discovery process than the growing clinical interest in the Class C street drug ketamine (e.g. Berman et al., 2000; Morgan and Curran, 2012).

Moving forward

The need to develop a greater understanding of clinical depression and to apply that knowledge to the development of new animal models for use in a reformulated drug discovery process highlights one further difficulty that needs to be addressed before that process can begin: that of resolving the different approaches taken by psychiatrists and pre-clinicians so that findings from each can be integrated. The prevalent approach in the clinic has been nosological, with an emphasis on interview and questionnaires that themselves concentrate heavily on mood state (such as those originally developed by Hamilton (1960) and Beck and colleagues (1961)). These methods are clearly not available to those that work with animals. Hence, although there is no widespread tradition of behavioural assessment in psychiatry (e.g. van Praag et al., 1987; van Praag, 2010) a full characterisation of the behavioural expression of depression in the clinic is nonetheless required.

Importance of the ethological approach

Ethological analysis was first applied to laboratory animals in the 1960s (e.g. Grant and MacKintosh, 1963) and has gained use particularly in the fields of anxiety (e.g. Cole and Rodgers, 1993; Holmes et al., 2000, Smolinsky et al., 2009) and depression (e.g. Hendrie and Starkey, 1998; Pickles et al., 2012).

The most important aspects of the ethological approach are Tinbergen’s four principal questions (Tinbergen, 1963). These questions relate to form (what does it look like?), function (what does it do?), ontogeny (how does it change over a lifetime?) and phylogeny (what does it look like and do in other species?) and can be applied to all levels from molecular to societal to give insights that are not obtainable by any other means.

In the context of depression this approach has recently led to an analysis that suggests that this is an adaptation mediated via events in the third ventricle (Hendrie and Pickles, 2010). This has been concluded because the behavioural cluster associated with depression (e.g. hunched posture, avoidance of eye contact, reduced competition for food/sex and sleep disruption) is defensive in nature. This cluster in turn serves to reduce an individual’s attack provoking stimuli and so allows them to remain within social groups that have become hostile to their presence. Depression is most commonly seen in adults and the nature of the adaptation indicates that it would only have developed in those species with a similar social need to our own. The third ventricle is implicated because the hypothalamus lies at one end, the pineal at the other and major pathways from the amygdala and hippocampus pass through it (Hendrie and Pickles, 2009, 2010).

The opportunity to test this hypothesis experimentally has not presented itself as yet and so it may not be supported. Nonetheless, its proposal serves to illustrate the potential significance of the guiding principles of ethology in this context and the value of the pioneering work that has already been done on the behavioural characterisation of human psychiatric illness (e.g. Dixon, 1986; Dixon et al., 1989; Troisi et al., 1990; Troisi, 1999).

Reformulating the process

The failure of the current antidepressant drug discovery process has been contingent upon a number of avoidable factors and it is imperative that these mistakes are not repeated. It is beyond the scope of this paper and the position of those authoring it to suggest in detail how the antidepressant drug discovery process should be reformulated but there are a few general principles that clearly need to be adhered to:

Ethology must be at the heart of this new process, to provide the required ethological description of clinical depression in all its various stages and to enable homologies between animals and man to be identified as they emerge.

New animal models need to be developed on the basis of this knowledge and designed to allow full integration of data generated in animals and man.

These new models must also be constructed on the basis of a full understanding of the natural histories of the species to be used in them. Since it is postulated that depression is a species-specific adaptation developed in response to particular social circumstance this adaptation will only be seen in those species with similar social needs to our own (Hendrie and Pickles, 2009, 2010). The social organisations of rats and mice do not predict that they are amongst the group of species to have developed this adaptation (Hendrie and Pickles, 2009). Mongolian gerbils (Meriones unguiculatus) are, however, a strong candidate in this context (e.g. Pickles et al., 2012) and other species will no doubt be found once they are looked for.

Lastly, the purpose of the drug discovery process is to produce drugs that are both efficacious and safe and the process needs to be shaped to meet those goals, not the other way round. Hence, if rats and mice are not suitable for use in this context then concerns about using different species in new animal models must be set to one side. The continued use of current animal models of depression cannot be justified only on the grounds that they fit in with other aspects of the drug discovery process and the process must evolve to accommodate that.

Conclusions

The ‘annus horribilis’ suffered by European neuroscience in 2010 as the result of GSK, AstraZeneca, Pfizer, Merck and Sanofi all announcing significant reductions in their research efforts into traditional drug discovery for the treatment of neuropsychiatric disorders has been well documented (Nutt and Goodwin, 2011). The catastrophic failure of the antidepressant drug discovery process has undoubtedly been a contributing factor and there are hence pressing clinical, ethical and commercial reasons why this process must be quickly built anew.

Footnotes

Conflict of interest

None declared.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Barnett

(1963) The Rat: A study in Behavior. Chicago: Aldine.

Beck

Ward

Mendelson

. (1961) An inventory for measuring depression. Arch Gen Psychiatry 4: 561–571.

Berman

Cappiello

Anand

. (2000). Antidepressant effects of ketamine in depressed patients. Biol Psych 47: 351–354.

Berry

(1970) The natural history of the house mouse. Field Studies 3: 219–226.

Bielski

Friedel

(1976) Prediction of tricyclic antidepressant response: A critical review. Arch Gen Psychiatry 33: 1479–1489.

Blazer

(2000) Mood disorders: Epidemiology. In: Sadock

Sadock

(eds) Comprehensive Textbook of Psychiatry. New York: Lippincott, Williams and Wilkins, pp.1298–1308.

Blier

(2010) The well of novel antidepressants: Running dry. J Psychiatry Neurosci 35: 219–220.

Borsini

(2012) Models for depression in drug screening and preclinical studies: Future directions World J Pharmacol 1: 21–29.

Brain

(1971) The physiology of population limitation in rodents – a review. Commun Behav Biol 6: 115–123.

10.

Brown

Steinberg

van Praag

(1994) The pathogenesis of depression: Reconsideration of neurotransmitter data. In: den Boer

Sitsen

JMA

(eds) Handbook of Depression and Anxiety: A Biological Approach. New York: Marcel Dekker, pp.317–347.

11.

Burt

Stein

(2002) Epidemiology of depression throughout the female life cycle. J Clin Psychiatry 63: 9–15.

12.

Castagné

Moser

Roux

. (2011) Rodent models of depression: Forced swim and tail suspension behavioural despair tests in rats and mice. Curr Protoc Pharmacol Ch. 8: Unit 8.10A.

13.

Chen

Korostyshevsky

Lee

. (2012) Accumulation of an antidepressant in vesiculogenic membranes of yeast cells triggers autophagy. PLoS ONE 7: e34024. DOI: 10.1371/journal.pone.0034024.

14.

Chermat

Thierry

Mico

. (1986) Adaptation of the tail suspension test to the rat. J Pharmacol 17: 348–350.

15.

Chessin

Dubnick

Kramer

. (1956) Modifications of pharmacology of reserpine and serotonin by iproniazid. Federation Proc 15: 409.

16.

Cipriani

Furukawa

Salanti

. (2009) Comparative efficacy and acceptability of 12 new-generation antidepressants: A multiple-treatments meta-analysis. Lancet 373: 746–758.

17.

Cole

Rodgers

(1993) An ethological analysis of the effects of chlordiazepoxide and bretazenil (Ro16–6028) in the murine elevated plus-maze. Behav Pharmacol 4: 573–580.

18.

Colman

Naicker

Zeng

. (2011) Predictors of long-term prognosis of depression. CMAJ 183: 1969–1976.

19.

Cryan

Holmes

(2005) The ascent of mouse: Advances in modelling human depression and anxiety. Nat Rev Drug Discov 4: 785–790.

20.

Cryan

Mombereau

(2004) In search of a depressed mouse: Utility of models for studying depression-related behavior in genetically modified mice. Mol Psychiatry 9: 326–3577.

21.

Cryan

Markou

Lucki

(2002) Assessing antidepressant activity in rodents: Recent developments and future needs. Trends Pharmacol Sci 23: 238–245.

22.

Cryan

Mombereau

Vassout

(2005) The tail suspension test as a model for assessing antidepressant activity: Review of pharmacological and genetic studies in mice. Neurosci Biobehav Rev 29: 571–625.

23.

Detke

Rickels

Lucki

(1995) Active behaviors in the rat forced swimming test differentially produced by serotonergic and noradrenergic antidepressants. Psychopharmacology 121: 66–72.

24.

Deussing

(2006) Animal models of depression. Drug Discov Today Dis Models 3: 375–383.

25.

Dias

Santos

Takeuchi

. (2009) Depression across the species. Open Psychiatry J 3, 33–38.

26.

Dixon

(1986) Ethological aspects of psychiatry. Schweiz Arch Neurol Psychiatr 137, 151–163.

27.

Dixon

Fisch

Huber

. (1989) Ethological studies in animals and man: Their use in psychiatry. Pharmacopsychiatry 22: 44–50.

28.

Domino

(1999) History of modern psychopharmacology – a personal view with an emphasis on antidepressants. Psychosom Med 61: 591–598.

29.

Dording

Mischoulon

Petersen

. (2002) The pharmacologic management of SSRI-induced side effects: A survey of psychiatrists. Ann Clin Psychiatry 14: 143–147.

30.

Duman

(2010) Models of depression. Vitam Horm 82: 1–21.

31.

Yacoubi

M EI

Vaugeois

(2007) Genetic rodent models of depression. Curr Opin Pharmacol 7: 3–7.

32.

Fava

Davidson

(1996) Definition and epidemiology of treatment-resistant depression. Psychiatr Clin North Am 19: 179–200.

33.

Forbes

Stewart

Matthews

. (1996) Chronic mild stress and sucrose consumption: Validity as a model of depression. Physiol Behav 60: 1481–1484.

34.

Fournier

DeRubeis

Hollon

. (2010) Antidepressant drug effects and depression severity: A patient-level meta-analysis. JAMA 303: 47–53.

35.

Garland

Pettus-Davis

Howard

(2012) Self-medication among traumatized youth: Structural equation modeling of pathways between trauma history, substance misuse, and psychological distress. J Behav Med. DOI: 10.1007/s10865–012–9413–5, (published online: 28 March 2012).

36.

Geddes

Carney

Davies

. (2003) Relapse prevention with antidepressant drug treatment in depressive disorders: A systematic review. Lancet 361: 653–661.

37.

Geddes

Freemantle

Mason

. (2000) SSRIs versus other antidepressants for depressive disorder. Cochrane Database Syst Rev 2: CD001851.

38.

Goodrich

(1963) FDA’s regulations under the Kefauver-Harris Drug Amendments of 1962. Food Drug Cosmet Law J 18: 561–569.

39.

Grant

MacKintosh

(1963) A comparison of the social postures of some common laboratory rodents. Behaviour 21: 246–259.

40.

Harmer

Cowen

Goodwin

(2011) Efficacy markers in depression. J Psychopharm 25: 1148–1158.

41.

Hamilton

(1960) A rating scale for depression. J Neurol Neurosurg Psychiatry 23: 56–62.

42.

Hendrie

Pickles

(2009) Depression as an evolutionary adaptation: Implications for the development of preclinical models. Med Hypotheses 72: 342–347.

43.

Hendrie

Pickles

(2010) Depression as an evolutionary adaptation: Anatomical organisation around the third ventricle. Med Hypotheses 74: 735–740.

44.

Hendrie

Starkey

. (1998) Pairbond disruption in Mongolian gerbils: Effects on subsequent social behaviour. Physiol Behav 63: 895–901.

45.

Hendrie

Phillips

Pickles

(2011) Behavioural scientists appear not to have specialist knowledge of laboratory animal behaviour. J Psychopharm 25: A35–A35.

46.

Hendrie

Sarailly

Starkey

(1998) Evidence to suggest that self-medication with alcohol is not an effective treatment for the control of depression. J Psychopharm 12: 112.

47.

Hendrie

Weiss

Eilam

(1996) Exploration and predation models of anxiety: Evidence from laboratory and wild species. Pharmacol Biochem Behav 54, 13–20.

48.

Herrling

(2005) The drug discovery process. Prog Drug Res 62: 1–14.

49.

Hindmarch

(2001) Expanding the horizons of depression: Expanding the monoamine hypothesis. Hum Psychopharm Clin Exp 16: 203–218.

50.

Hollon

Thase

Markowitz

(2002) Treatment and prevention of depression. Psychol Sci Public Interest 3: 39–77.

51.

Holmes

Parmigiani

Ferrari

. (2000) Behavioral profile of wild mice in the elevated plus-maze test for anxiety. Physiol Behav 71: 509–516.

52.

Hughes

Cohen

(2009) A systematic review of long-term studies of drug treated and non-drug treated depression. J Affect Disord 118: 9–18.

53.

Jick

Kaye

Jick

(2004) Antidepressants and the risk of suicidal behaviours. JAMA 292: 338–343.

54.

Keeney

Hogg

(1999) Behavioural consequences of repeated social defeat in the mouse: Preliminary evaluation of a potential animal model of depression. Behav Pharmacol 10: 753–764.

55.

Kelly

Wrynn

Leonard

(1997) The olfactory bulbectomized rat as a model of depression: An update. Pharmacol Ther 74: 299–316.

56.

Krishnan

Nestler

(2008) The molecular neurobiology of depression. Nature 455: 894–902.

57.

Krishnan

Nestler

(2011) Animal models of depression: molecular perspectives. Curr Topics Behav Neurosci 7: 121–147.

58.

Kudryavtseva

Bakshtanovskaya

Koryakina

(1991) Social model of depression in mice of C57BL/6J strain. Pharmacol Biochem Behav 38: 315–320.

59.

Lacasse

Leo

(2005) Serotonin and depression: A disconnect between the advertisements and the scientific literature. PLoS Med 2(12): e392.

60.

Leo

Lacasse

(2008) The media and the chemical imbalance theory of depression. Society 45: 35–45.

61.

Lin

H-C

Erickson

Balkrishnan

(2011) Antidepressant utilization, adherence, and health care spending in the United States: The case of MDD patients 2000–2007. Health Outcomes Res Med 2: e79–e89.

62.

Lipsky

Sharp

(2001) From idea to market: The drug approval process. J Am Board Fam Pract 14: 362–367.

63.

Loomer

Saunders

Kline

(1957) A clinical and pharmacodynamic evaluation of iproniazid as a psychic energizer. Psychiatr Res Rep Am Psychiatr Assoc 8: 129–141.

64.

McArthur

Borsini

(2006) Animal models of depression in drug discovery: A historical perspective. Pharm Biochem Behav 84: 436–452.

65.

Micó

Ardid

Berrocoso

. (2006) Antidepressants and pain. Trends Pharmacol Sci 27: 348–354.

66.

Middleton

Shaw

Feder

(2005) NICE guidelines for the management of depression are clear for severe depression, but uncertain for mild or moderate depression. BMJ 330: 267–268.

67.

Miller

(2010) Is pharma running out of brainy ideas? Science 329: 502–504.

68.

Mojtabai

Olfson

(2011) Proportion of antidepressants prescribed without a psychiatric diagnosis is growing. Health Aff (Millwood) 30: 1434–1442.

69.

Möller

Volz

(1996) Drug treatment of depression in the 1990s. An overview of achievements and future possibilities. Drugs 52: 625–638.

70.

Moncrieff

(2007) Are antidepressants as effective as claimed? No, they are not effective at all. Can J Psychiatry 52: 96–97.

71.

Monleon

D’Aquila

Parra

. (1995) Attenuation of sucrose consumption in mice by chronic mild stress and its restoration by imipramine. Psychopharmacology 117: 453–457.

72.

Morgan

CJA

Curran

(2012) Ketamine use: A review. Addiction 107: 27–38.

73.

Nestler

Hyman

(2010) Animal models of neuropsychiatric disorders. Nat Neurosci 13: 1161–1169.

74.

NICE (2004) Depression: Management of Depression in Primary and Secondary Care. Clinical Guideline 23. London: National Institute for Clinical Excellence (updated 2009).

75.

Nutt

Goodwin

(2011) ECNP Summit on the future of CNS drug research in Europe 2011: Report prepared for ECNP by David Nutt and Guy Goodwin. Eur Neuropsychopharmacol 21: 495–499.

76.

Papakostas

Fava

(2009) Does the probability of receiving placebo influence clinical trial outcome? A meta-regression of double-blind, randomized clinical trials in MDD. Eur Neuropsychopharm 19: 34–40.

77.

Paul

Mytelka

Dunwiddie

. (2010) How to improve R&D productivity: The pharmaceutical industry’s grand challenge. Nature Rev Drug Discov 9: 203–214.

78.

Pickles

Hagan

Jones

DNC

. (2012) Short-term individual housing induced social deficits in female Mongolian gerbils: Attenuation by chronic but not acute imipramine. Neuropharmacology 62: 1993–1998.

79.

Pigott

Leventhal

Alter

. (2010) Efficacy and effectiveness of antidepressants: Current status of research. Psychother Psychosom 79: 267–279.

80.

Porsolt

Brossard

Hautbois

. (2001) Rodent models of depression: Forced swimming and tail suspension behavioral despair tests in rats and mice. In: Current Protocols in Neuroscience. John Wiley & Sons, Inc., p.8.10A.1–10.

81.

Porsolt

Le Pichon

Jalfre

(1977) Depression: A new animal model sensitive to antidepressant treatments. Nature 266: 730–732.

82.

Rodgers

Hendrie

(1982) Agonistic behaviour in rats: Evidence for non-involvement of opioid mechanisms. Physiol Behav 29: 85–90.

83.

Rodgers

Hendrie

(1983) Social conflict activates status-dependent endogenous analgesic or hyperalgesic mechanisms in male mice: Effects of naloxone on nociception and behaviour. Physiol Behav 30: 775–780.

84.

Rodgers

Hendrie

Waters

(1983) Naloxone partially antagonizes post-encounter analgesia and enhances defensive responding in male rats exposed to attack from lactating conspecifics. Physiol Behav 30: 781–786.

85.

Rygula

Abumaria

Flugge

. (2005) Anhedonia and motivational deficits in rats: Impact of chronic social stress. Behav Brain Res 162: 127–134.

86.

Samuels

Leonardo

Gadient

. (2011) Modeling treatment-resistant depression. Neuropharmacology 61: 408–413.

87.

Siegfried

Frischknecht

Riggio

. (1987) Long-term analgesic reaction in attacked mice. Behav Neurosci 101: 797–805.

88.

Skolnick

(1999) Antidepressants for the new millennium. Eur J Pharmacol 375: 31–40.

89.

Smolinksy

Bergner

LaPorte

. (2009) Analysis of grooming behavior and its utility in studying animal stress, anxiety, and depression. In: Gould

(ed.) Mood and Anxiety Related Phenotypes in Mice. Neuromethods 42. New York: Humana Press, pp.21–36.

90.

Song

Leonard

(2005) The olfactory bulbectomised rat as a model of depression. Neurosci Biobehav Rev 29: 627–647.

91.

Steru

Chermat

Thierry

. (1985) The tail suspension test: A new method for screening antidepressants in mice. Psychopharmacology (Berl) 85: 367–370.

92.

Tinbergen

(1963) On aims and methods of ethology. Zeitschrift fur Tierpsychologie 20: 410–433.

93.

Trivedi

Rush

Wisniewski

. (2006) Evaluation of outcomes with citalopram for depression using measurement-based care in STAR*D: Implications for clinical practice. Am J Psychiatry 163: 28–40.

94.

Troisi

(1999) Ethological research in clinical psychiatry: The study of nonverbal behavior during interviews. Neurosci Biobehav Rev 23: 905–913.

95.

Troisi

Pasini

Bersani

. (1990) Ethological assessment of the DSM-III subtyping of unipolar depression. Acta Psychiatr Scand 81: 560–564.

96.

Turner

Matthews

Linardatos

. (2008) Selective publication of antidepressant trials and its inﬂuence on apparent effcacy. N Engl J Med 358: 252–256.

97.

Üstün

Ayuso-Mateos

Chatterji

. (2004) Global burden of depressive disorders in the year 2000. Br J Psychiatry 184: 386–392.

98.

Van Praag

(2010) Biological psychiatry: Still marching forward in a dead end. World Psychiatry 9: 164–165.

99.

Van Praag

Kahn

Asnis

(1987) Denosologization of biological psychiatry or the specificity of 5-HT disturbances in psychiatric disorders. J Affect Disord 13: 1–8.

100.

Varty

Cohen-Williams

Hunter

(2003) The antidepressant-like effects of neurokinin NK1 receptor antagonists in a gerbil tail suspension test. Behav Pharmacol 14: 87–95.

101.

Verkman

(2004) Drug discovery in academia. Am J Physiol Cell Physiol 286: C465–C474.

102.

Walsh

Seidman

Sysko

. (2002) Placebo response in studies of major depression: Variable, substantial, and growing. JAMA 287: 1840–1847.

103.

Wattiez

A-S

Libert

Privat

A-M

. (2011) Evidence for a differential opioidergic involvement in the analgesic effect of antidepressants: Prediction for efficacy in animal models of neuropathic pain? Br J Pharmacol 163: 792–803.

104.

West

(1990) Neurobehavioral studies of forced swimming: the role of learning and memory in the forced swim test. Prog Neuropsychopharmacol Biol Psychiatry 14: 863–877.

105.

Willner

(1984) The validity of animal models of depression. Psychopharmacology 83: 1–16.

106.

Willner

(1997) Validity, reliability and utility of the chronic mild stress model of depression: A 10-year review and evaluation. Psychopharmacology 134: 319–329.

107.

Willner

Towell

Sampson

. (1987) Reduction of sucrose preference by chronic unpredictable mild stress, and its restoration by a tricyclic antidepressant. Psychopharmacology 93: 358–364.

108.

Wong

Licinio

(2004) From monoamines to genomic targets: A paradigm shift for drug discovery in depression. Nat Rev Drug Discov 3: 136–151.

109.

Woodcock

Woosley

(2008) The FDA Critical Path Initiative and its influence on new drug development. Ann Rev Med 59: 1–12.

110.

World Health Organization (2002) The World Health Report 2002: Reducing Risks, Promoting Healthy Life. Geneva, Switzerland: World Health Organization.

111.

Yalcin

Aksu

Belzung

(2005) Effects of desipramine and tramadol in a chronic mild stress model in mice are altered by yohimbine but not by pindolol. Eur J Pharmacol 514: 165–174.