WHY INVESTIGATE NEW TARGETS BEYOND MONOAMINES?

Despite an explosion in psychiatric neuroscience research in the past decade, there is consensus regarding a significant ‘pipeline’ problem in neuropsychopharmacology (Agid et al, 2007; Pangalos et al, 2007). Given the high lifetime prevalence of DSM-IV mood disorders (20.8%), with first onset often in childhood or adolescence (Kessler et al, 2005), the public health urgency of identifying safe and more effective treatments for these common and disabling conditions is clear. For unipolar and bipolar disorder, and the commonly comorbid anxiety disorders, there have been very few significant innovations and no genuine breakthrough drugs in the past 2 decades (Spedding et al, 2005). Novel therapeutic strategies that accelerate speed of onset, enhance efficacy, and tolerability, while addressing core pathophysiological features of these illnesses are needed.

The limitations of current approaches to drug discovery in neuropsychiatric disorders have been extensively scrutinized (Roth et al, 2004; Wong and Licinio, 2004; Spedding et al, 2005; Berton and Nestler, 2006; Agid et al, 2007). Much of the previous (and current) focus of investigation for mood disorders regards the biology and neural circuitry most relevant to the monoaminergic systems (serotonin (5-HT), norepinephrine (NE), and dopamine (DA)), given the therapeutic benefit of the selective 5-HT reuptake inhibitors (SSRIs), tricyclic antidepressants (TCAs), monoamine oxidase inhibitors (MAOIs), and dopaminergic medications (Mann, 2005; Dunlop and Nemeroff, 2007; Nestler and Carlezon, 2006). A recent modernization of the classical monoamine theory of depression has posited that elevated density of monoamine oxidase A (MAO-A) (the enzyme that metabolizes 5-HT, NE, and DA) is the primary monoamine-lowering process during a major depressive episode, while monoamine transporter density variability plays a secondary role in influencing loss of specific monoamines, explaining symptomatic heterogeneity during major depressive episodes (Meyer et al, 2006). Areas of active investigation in monoamine biology include (1) catecholamine depletion studies using acute tryptophan depletion, phenylalanine/tyrosine, or alpha-methyl-para-tyrosine (see Ruhe et al, 2007 for review); (2) in vivo brain imaging studies of specific receptors and transporters for these neurotransmitters (eg, Oquendo et al, 2007; Parsey et al, 2006a, 2006b; Bhagwagar et al, 2006; Moses-Kolko et al, 2007); (3) cerebrospinal fluid (CSF) and neuroendocrine challenge studies; (4) postmortem studies; and more recently, (5) large-scale genetic association studies (Hu et al, 2007). Inasmuch as the phenotype of depression is heterogeneous, biological investigations of monoamine systems in diverse patient groups has hindered the ability to reliably uncover abnormalities and has stymied identification of biomarkers. Contradictions in the literature, rather than reflecting minimal involvement of monoaminergic systems in disorder pathophysiology, likely reflect heterogeneity of biological substrates. Phenotypic contributors to mood disorder heterogeneity, which likely impact expression of monoamines (and response to SSRIs and similar drugs), include sex, age of onset (early onset vs mid-life vs late-onset), race/ethnicity, subtype (melancholic vs atypical vs dysthymic), current mood state (euthymic vs depressed), chronicity of illness, previous number of episodes, family history of psychopathology, and axis I–III disorder comorbidity (Trivedi et al, 2006).

Thus, while SSRIs and similar monoaminergic drugs are the mainstays of pharmacological treatment for mood disorders, their limitations in sustaining disorder remission are increasingly apparent. The first phase of the STAR*D study, the largest effectiveness study of its kind in ‘real world’ patients, measured the efficacy of the SSRI citalopram in outpatients with unipolar depression (n=2876). Remission rates were 28%, a remission rate similar to that achieved in standard randomized placebo-controlled acute efficacy trials (Trivedi et al, 2006). As the presence of residual symptoms is a strong predictor of relapse or recurrence (Keller, 2004), optimal therapeutic strategies require a focus on achieving and sustaining remission, by presumably addressing core pathophysiological processes. Non-monoaminergic biological systems have received increasing attention from translational researchers for their potential therapeutic application for mood disorders. The most actively pursued drug targets for unipolar depression include amino-acid neurotransmitter receptors and transporters, neuropeptides (corticotropin-releasing factor (CRF), substance P), neurotrophic factors (brain-derived neurotrophic factor (BDNF)), and steroid receptors/hormones (glucocorticoid receptor, DHEA, testosterone, melatonin; see Hasler et al, 2004; Berton and Nestler, 2006; Martinowich et al, 2007 for reviews). It is noteworthy, however, that despite numerous potentially novel targets, relatively few non-monominergic-based drugs will gain FDA approval by the end of this decade, while additional monoaminergic medications, with minimal advantages over existing drugs, will likely gain regulatory approval over the next few years. Table 1 lists medications in phase II and III clinical trials for unipolar major depression (accessed on www.clinicaltrials.gov on 24 July 2007).

Table 1 Medications in US FDA Clinical Trials for Unipolar Major Depression

In this paper, we highlight several promising areas for therapeutics for mood disorders (unipolar depression and bipolar disorder) based upon two complementary approaches:

  1. 1)

    Understanding the presumed therapeutically relevant biochemical targets of medications currently in use, and using that knowledge to design new drugs directed at these targets. This includes not only direct biochemical targets, but also downstream targets that are regulated by chronic drug administration (ie, consistent with the clinical temporal profile; Gould et al, 2004a, 2004b).

  2. 2)

    Understanding the pathophysiology of the illness, and using that knowledge to design therapeutics to attenuate or prevent those pathological processes (Quiroz et al, 2004). As space limitations do not permit an exhaustive discussion of every promising candidate using both strategies, we emphasize compounds with particularly high potential for clinical translation.

AGENTS INTERACTING WITH NON-CATECHOLAMINE NEUROTRANSMITTER SYSTEMS

Amino-Acid Neurotransmitter Modulators

Glutamate, the major excitatory neurotransmitter in the brain, is ubiquitous in the mammalian CNS, where it is used in up to 60% of synapses. The emerging interest in glutamate-modulating agents derives from both a pathophysiological perspective regarding amino-acid neurotransmitters’ role in mood disorders as well as evidence that conventional antidepressants have effects on specific glutamate receptor subtypes (Manji et al, 2003; Charney and Manji, 2004). Recent pathophysiological hypotheses proffer a view of dynamic regional amino-acid transmitter dysfunction in severe, recurrent mood disorders, with implications for experimental therapeutics that target-specific ionotropic or metabotropic receptors, glial glutamate transporters, or vesicular glutamate release (Kugaya and Sanacora, 2005; Zarate et al, 2003; Sattler and Rothstein, 2007; see Figure 1, pathways c–f). One prominent model for unipolar depression proposes that disrupted glial cell function in structures relevant to emotional processing such as amygdala, observed in postmortem mood disorder samples (Hamidi et al, 2004), may result in decreased uptake of glutamate, with a resultant elevation of extracellular glutamate levels (Kugaya and Sanacora, 2005). Disruptions in glutamate (Glu)–glutamine (Gln) neuronal–glial cycling are believed to cause decreased glutamate release and decreased cortical GABA, examined in vivo in human subjects using techniques such as 13C-MRS (Shen, 2006; Mason et al, 2007). A reduction in glial cell function would result in decreased flux through the Glu/Gln and GABA/Gln cycles, findings that are consistent with reduced rates of Glu/Gln cycling and lowered cortical GABA concentrations in individuals with MDD (Kugaya and Sanacora, 2005; Sanacora and Saricicek, 2007). Microarray analysis has demonstrated significant dysregulation of a specific subset of genes encoding the astroglial high-affinity glutamate transporters SLC1A2 (EAAT2; GLT1) and SLC1A3 (EAAT1; GLAST); glutamate-ammonia ligase (glutamine synthetase); and various subunits of glutamate receptors and GABAA receptors in anterior cingulate (area 24) and left dorsolateral prefrontal cortex (areas 9 and 46) (Choudary et al, 2005), lending support for regionally specific abnormalities in glutamate function in mood disorders.

Figure 1
figure 1

Plasticity regulators as targets for the development of novel agents for the treatment of mood disorders. This figure depicts the multiple targets by which neuroplasticity and cellular resilience can be increased in mood disorders. (a) Phosphodiesterase inhibitors increase the levels of pCREB; (b) MAP kinase modulators increase the expression of the major neurotrophic protein Bcl-2; (c) mGluR II/III agonists modulate the release of excessive levels of glutamate; (d) drugs such as lamotrigine and riluzole act on Na+ channels to attenuate glutamate release; (e) AMPA potentiators upregulate the expression of BDNF; (f) NMDA antagonists like ketamine and memantine enhance plasticity and cell survival; (g) novel drugs to enhance glial release of trophic factors and clear excessive glutamate may have utility for the treatment of depressive disorders; (h) CRF antagonists and (i) glucocorticoid antagonists attenuate the deleterious effects of hypercortisolemia, and CRF antagonists may exert other beneficial effects in the treatment of depression via non-HPA mechanisms; (j) agents which upregulate Bcl-2 (eg, pramipexole, shown to be effective in bipolar depression; Zarate et al, 2004b). These distinct pathways have convergent effects on cellular processes such as bioenergetics (energy metabolism), neuroplasticity, neurogenesis, resilience, and survival. Nature Reviews Drug Discovery. Modified and reproduced with permission from Charney and Manji (2004).

The impairment in glial uptake of Glu, and conversion to Gln, would lead to an elevation of extracellular Glu during intense neuronal activity, and serve to activate extracellular N-methyl-D-aspartate (NMDA) receptors associated with various forms of excitotoxicity (Pittenger et al, 2007). Changes in extrasynaptic Glu concentrations can have wide-ranging effects on diverse physiological functions, including activation of signal transduction pathways involved in the regulation of neurotrophic factors, neuroplasticity, and cellular resilience (Pittenger et al, 2007; see Figure 1). GABAergic involvement in the pathophysiology and treatment of mood disorders is equally compelling, and supported by preclinical studies showing stress-related changes in GABAergic function, GABAergic effects of existing antidepressant medications, antidepressant/mood-stabilizing efficacy associated with GABAergic drugs, and robust GABAergic abnormalities and genetic associations in depressed patients (see Sanacora and Saricicek, 2007 for recent review).

Antidepressants/mood stabilizers that impact primarily glutamate receptors such as NMDA receptor antagonists, metabotropic glutamate receptor (mGluR) agonists and antagonists, and positive modulators of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors have demonstrated antidepressant-like properties, with potential common downstream mechanisms of action (Paul and Skolnick, 2003; Zarate et al, 2003; Sanacora et al, 2003; Mathew et al, 2005b; see Figure 1). Alterations in neural plasticity in critical limbic and reward circuits, mediated by increasing postsynaptic AMPA to NMDA throughput, may represent a convergent mechanism for therapeutics of severe mood disorders (Schloesser et al, 2008; Figure 1, pathways e, f). In the sections below we discuss several glutamate-modulating approaches that may hold particular promise for unipolar and bipolar depression.

Riluzole (Major Depressive Disorder And Bipolar Depression)

Riluzole (2-amino-6-trifluoromethoxy benzothiazole) is a neuroprotective agent that is the only FDA-approved medication for amyotrophic lateral sclerosis (ALS), although median survival is only prolonged by about 2–3 months (Miller et al, 2007). Two open-label studies support the potential utility of riluzole in treatment-resistant unipolar depression (Zarate et al, 2004a; Sanacora et al, 2007), and one open-label add-on study found improvements in bipolar depression (Zarate et al, 2005). Interestingly, the patients who responded to riluzole tended to be in remission (ie, almost symptom-free), suggesting that there may be a subgroup of mood disorder patients for whom glutamatergic strategies have considerable utility. Riluzole has complex mechanisms of action; among its actions, it inhibits release of glutamate through inhibition both of voltage-dependent sodium channels and of P/Q-type calcium channels, enhances glutamate reuptake, and upregulates AMPA receptors (reviewed in Zarate et al, 2006b; Mathew et al, 2005b; see Figure 1, pathway d). Riluzole modulates neuroplasticity via stimulation of growth factor synthesis such as BDNF (Mizuta et al, 2001; Katoh-Semba et al, 2002), promotion of neuritogenesis, neurite branching, and neurite outgrowth (Shortland et al, 2006), increasing concentrations of the neuronal marker N-acetyl-aspartate in hippocampus (Mathew et al, in press), and enhancement of surface expression of the hippocampal AMPA receptor subunits (GluR1 and GluR2) (Du et al, 2007). The latter process is implicated in the regulation of activity-dependent synaptic strength and postsynaptic receptor responsiveness (Du et al, 2007).

Riluzole's efficacy has also been reported for patients with generalized anxiety disorder (GAD) (Mathew et al, 2005a; Mathew et al, in press). The relationship between GAD and mood disorders is particularly strong (Moffitt et al, 2007; Kessler et al, 2005), and patients with ‘anxious depression’ likely represent a common and valid subtype of mood disorder (Fava et al, 2004, 2006) and are less likely to achieve remission with standard SSRIs (Trivedi et al, 2006). One intriguing mechanism that could explain riluzole's anxiolytic efficacy is evidence that at higher concentrations, riluzole strongly potentiated postsynaptic GABAA receptor function in cultured hippocampal neurons (He et al, 2002). Commonly used anxiolytics such as benzodiazepines bind to and allosterically interact with GABAA receptors, which mediate most fast inhibitory neurotransmission. Thus, riluzole may confer antidepressant and anxiolytic activity through numerous mechanisms, and randomized placebo-controlled studies are warranted.

NMDA Receptor Antagonists (Major Depressive Disorder And Bipolar Depression)

There is substantial evidence that NMDA receptor antagonists have antidepressant properties (Krystal, 2007). Chronic treatment with NMDA antagonists such as MK-801 and AP-7 resulted in antidepressant-like behavioral effects in models such as chronic mild stress, learned helplessness, footshock-induced aggression, and olfactory bulbectomy (reviewed in Paul and Skolnick, 2003). Chronic, but not acute, treatment with NMDA antagonists also resulted in reductions in density of forebrain β-adrenoreceptors (Paul et al, 1992; Maj et al, 1993) and 5-HT2 receptors (Papp et al, 1994). Similar to monoamine-based antidepressants, NMDA antagonists given to animals processed in the forced swim test resulted in rapid downregulation of forebrain β-adrenoceptors (Wedzony et al, 1995). Repeated administration of conventional antidepressants has been found to alter the expression of mRNA that encodes NMDA receptor subunits (Boyer et al, 1998), through expression of BDNF (Brandoli et al, 1998).

The above preclinical reports provided a sound rationale for recent exploratory work using NIMDA receptor antagonists in mood disorder patients. Ketamine, a high-affinity non-competitive NMDA antagonist, has been used as a standard anesthetic agent for many years both in pediatrics and in adults, with doses as high as 2 mg/kg intravenous (i.v.). Two placebo-controlled studies of a single subanesthetic dose of i.v. ketamine (0.5 mg/kg infusion over 40 min) in unipolar depression (Berman et al, 2000; Zarate et al, 2006a, 2006b), reported rapid although transient antidepressant efficacy. Euphoric and psychotomimetic side effects were observed acutely (2–4 h following i.v. administration), but were temporally distinct from the amelioration of core depressive symptoms, which in some cases persisted for up to 1 week. It was shown more recently in rodent models that AMPA receptor throughput is required for the antidepressant-like effect of ketamine, and that ketamine may result in rapid antidepressant effect by enhancing AMPA relative to NMDA throughput in critical neuronal circuits (Du et al, 2006; Maeng et al, 2007; Figure 1, pathway f).

In contrast to the dramatic effects observed in these two i.v. ketamine studies, a placebo-controlled study of the low-to-moderate-affinity non-competitive NMDA antagonist memantine (oral dosing) did not show antidepressant effects (Zarate et al, 2006c). While higher-affinity NMDA antagonists may be required for antidepressant efficacy, i.v. administration may also be an important factor. Important next steps for i.v. ketamine studies in mood disorders include (1) dose-finding studies of i.v. ketamine, as all studies thus far have used fixed 0.5 mg/kg dosage, in an attempt to balance antidepressant efficacy with side effects; (2) repeated dose administration studies; (3) investigations of whether particular medications may be used in combination with i.v. ketamine to potentiate (and extend) the antidepressant effects; and (4) studies in bipolar depression. Given ketamine's potential for acute psychotomimetic side effects, however, more selective subtype NMDA antagonists, such as NR2B receptor subtype, with lower liability for these side effects, will continue to be explored (Zarate et al, 2007a). Indeed, the NR2B antagonist Ro 25-6981 had antidepressant-like properties in the forced swim test (Maeng et al, 2007). From a safety perspective, this compound also has not been associated with vacuolization in rodents (Gil et al, 2002), in contrast to the reversible vacuolization at high doses observed with noncompetitive NMDA antagonists such as ketamine and dizocilpine (Olney et al, 1989).

AMPA Receptor Potentiators/Ampakines (Major Depressive Disorder)

AMPA receptors are a subfamily of ionotropic glutamate receptors that mediate the fast component of excitatory neurotransmission, and which, like NMDA receptors, are involved in learning and memory. Several classes of compounds can allosterically modulate AMPA receptors. These compounds (so-called AMPA receptor-positive modulators or AMPA receptor potentiators, ARPs) do not activate AMPA receptors themselves, but slow the rate of receptor desensitization and/or deactivation in the presence of an agonist. Several AMPA receptor potentiators in development have demonstrated antidepressant effects in animal models of depression, including exposure to inescapable stressors, the forced swim test, the tail-suspension-induced immobility test, and learned helplessness models (reviewed in Zarate et al, 2006b). In contrast to traditional antidepressants, this group of compounds does not appear to affect the extracellular concentration of monoamines; however, these drugs can enhance the neurotrophic actions of BDNF mRNA and protein in primary neuronal cultures and in rat brain (Bleakman et al, 2007; Figure 1, pathway e). No placebo-controlled clinical trials in mood disorder patients have been published to date; a phase II trial of an AMPA receptor potentiator for major depression was recently halted due to preclinical toxicity (www.clinicaltrials.gov identifier NCTG00113022).

NEUROPEPTIDES AS THERAPEUTIC TARGETS FOR MAJOR DEPRESSION

Neuropeptides are short-chain amino acids that act as neurotransmitters in brain circuits implicated in mood and anxiety regulation. Recognition of the large number of neuropeptide targets and the role of stress-related neuropeptides in preclinical models has prompted investigation into the clinical utility of small-molecule neuropeptide modulators. Below we review two classes of neuropeptide receptor antagonists that are well represented in phase II and III RCTs of unipolar depression (Table 1): (1) corticotropin-releasing factor-1 (CRF1) receptor, also known as corticotropin-releasing hormone (CRH), antagonists, and (2) neurokinin (NK) receptor antagonists.

Corticotropin-Releasing Factor Antagonists

Over 20 years of preclinical research have suggested that the 41-amino-acid peptide CRF is an important mediator of the stress response, coordinating the neuroendocrine, autonomic, immune, and behavioral responses to stress, while hyperactivity of the hypothalamic–pituitary–adrenal (HPA) axis has been associated with subgroups of patients with major depression (Holsboer, 2000). Hypothalamic and extrahypothalamic neurons containing CRF1 and CRF2 receptors are located throughout key brain stem, limbic, and cortical regions implicated in affective and anxiety, with CRF1 the predominate subtype in limbic regions and in pituitary regulation of HPA axis activity (Valdez, 2006). Animal models of early life stress have been associated with hyperactivity of CRF neurons, and direct CRF administration into CNS is associated with depression and anxiety-like symptoms (Gorman et al, 2002). Clinical evidence for the putative role of CRF in patients with major depression includes (1) increased concentrations of CRF in CSF (Nemeroff et al, 1984); (2) blunted adrenocorticotropin (ACTH) response to systemic injection of exogeneous CRF, suggesting pituitary CRF receptor downregulation due to chronic CRF hypersecretion (Holsboer et al, 1984; Gold et al, 1984); (3) postmortem studies showing increased CRF and/or messenger RNA (mRNA) expression in hypothalamic paraventricular nucleus, brain stem, and frontal cortex (reviewed in Heim et al, in press); and (4) characteristic escape from suppression with elevated ACTH and cortisol response to the combined dexamethasone/CRF test (Ising et al, 2005).

To date, one CRF1 receptor antagonist (R121919) has been tested in an open-label clinical trial for major depression (Zobel et al, 2000), with improvements noted in depressive and anxiety symptoms. Hepatotoxicity ended the clinical development of this particular compound, although several CRF1 receptor antagonists are currently in worldwide development for major depression (see Table 1; Figure 1, pathway h) and for anxiety disorders, including post-traumatic stress disorder and social phobia.

Neurokinin Receptor Antagonists

Substance P is the most abundant brain tachykinin (a class, which includes NK A and B among others), and is the endogeneous ligand for the neurokinin 1 (NK-1) receptor (Ebner and Singewald, 2006). The initial interest in this neuropeptide and receptor family for mood and anxiety conditions reflected its extensive colocalization with monoamine neurotransmitters (Ebner et al, 2004) and abundant expression in stress neural circuitry. Genetic or pharmacologic blockade of NK-1 receptors were observed to induce many of the same long-term neural effects as standard antidepressants on cell signaling molecules such as BDNF and hippocampal neurogenesis (Blier et al, 2004; van der Hart et al, 2005), providing the rationale for experimental therapeutic trials in mood and anxiety disorders (Herpfer and Lieb, 2005). Despite the promising preclinical data, a recent pooled analysis of five 8-week randomized, double-blind, placebo-controlled, multicenter studies of the NK-1 receptor antagonist aprepitant in over 2500 patients with major depression failed to demonstrate efficacy (Keller et al, 2006). Notably, paroxetine 20 mg was used as an active comparator for three of the trials and demonstrated benefit over placebo on primary outcomes, suggesting that the trials had appropriate assay sensitivity. Despite the setback for this particular compound, several NK-1 antagonists remain in active development for unipolar depression (Table 1).

Saredutant (SR 48968) is a selective non-peptide NK-2 receptor antagonist that blocks the effects of NK A, the endogenous ligand for NK-2. This compound has shown antidepressant and anxiolytic-like effects in preclinical models (Salomé et al, 2006) and has progressed the furthest in clinical development. Of four phase III studies conducted in major depression, two showed statistically significant results and two studies were not statistically significant vs placebo; the drug manufacturer anticipates an application for regulatory approval in 2008 (source: Sanofi-Aventis press release, 17 September 2007).

COMMON INTRACELLULAR PATHWAYS OF MONOAMINERGIC DRUGS AND MOOD STABILIZERS/LITHIUM RELEVANT TO NOVEL THERAPEUTICS

Evidence that antidepressants and mood stabilizers exert major effects on signaling pathways regulating cellular plasticity has reshaped views about the neurobiological underpinnings of these disorders (Manji et al, 2001; D’Sa and Duman, 2002; Nestler et al, 2002; Schloesser et al, 2008). In this section we review several therapeutic targets and common pathways that monoaminergic antidepressants and mood stabilizers/lithium impact either directly or indirectly, with relevance for novel unipolar and bipolar disorder therapeutics.

BDNF/Growth Factors (Major Depressive Disorder)

Genetic/neurodevelopmental factors, repeated affective episodes (and likely elevations of glucocorticoids), and illness progression may all contribute to the impairments of cellular resilience, volumetric reductions, and cell death/atrophy observed in mood disorders (Charney and Manji, 2004). According to the neurotrophic hypothesis of depression (Duman and Monteggia, 2006), decreased expression of BDNF and other growth factors likely contributes to mood disorders, while upregulation of BDNF and its receptor TrkB may be a critical component of all antidepressants’ actions (Figure 1, pathway j). Structural imaging studies in unipolar depression have found a decrease in the gray matter volume of multiple areas of the orbital, medial, and dorsolateral prefrontal cortex, with the most prominent reduction reported in the left (but not right) subgenual prefrontal cortex (see Drevets, 2001, 2007). An increase in ventricular size also has been consistently reported in patients with bipolar disorder, and progressive gray matter loss in hippocampal, fusiform, and cerebellar regions (Moorhead et al, 2007; see Beyer et al, 2004; Manji et al, 2003 for reviews). The trophic effects are postulated to reverse illness-related atrophic changes, thereby reinstating the neurochemical throughput in critical circuitry regulating affective, cognitive, motoric and neurovegetative functions (Charney and Manji, 2004). Human phase I/II trials of recombinant methionyl human BDNF have already been undertaken, wherein the BDNF was administered by intrathecal infusion to patients with ALS (Ochs et al, 2001). Unfortunately, treatment-limiting side effects were encountered at higher doses, precluding further study. The recent report of dysregulation of several fibroblast growth factor (FGF) system transcripts in frontal cortical regions of brains from human subjects with major depression suggests that this growth factor family may also represent an important target for the development of novel therapeutics (Evans et al, 2004).

Erk MAP Kinase Pathway and Bcl-2 (Bipolar Disorder)

Lithium and mood stabilizers such as valproate activate the extracellular receptor-coupled kinase (ERK) mitogen-activated protein (MAP) kinase pathway, which is an intracellular signaling cascade implicated in neuroplasticty (Schloesser et al, 2008). An increasing number of strategies are being investigated to develop small-molecule switches for protein–protein interactions, which have the potential to regulate MAP kinase cascades, and interactions between homo- and heterodimers of the Bcl-2 family of proteins (Figure 2, pathway b; see Schloesser et al, 2008 for recent review). Both TrkA and TrkB utilize the phosphatidylinositol-3 kinase (PI3K)/Akt and ERK MAP kinase pathways to bring about their neurotrophic effects (see Schloesser et al, 2008). The ERK MAP kinase cascade also increases the expression of Bcl-2 via its effects on cAMP response element-binding protein (CREB).

Figure 2
figure 2

GSK-3 and intracellular signaling. GSK-3 regulates diverse signaling pathways in the cell. These include insulin/insulin-like growth factor (IGF-1) signaling, neurotrophic factor signaling, and Wnt signaling. Insulin signaling through its tyrosine receptor kinase (Trk) receptor activates PI3K-mediated signaling, resulting in inhibition of GSK-3. Inhibition of GSK-3 activates glycogen synthase and eukaryotic initiation factor 2B (eIF2B), while inhibiting insulin receptor substrate-1 (IRS-1, an inhibitor of the insulin receptor). Insulin is generally thought to have minimal effects on CNS neurons; however, IGF-1 interacting with its cognate receptor appears to have similar functions. Neurotrophins (NT) act through Trk receptors A, B, and C to activate PI3K, AKT, and inhibit GSK-3. Many effectors have been implicated in GSK-3's neurotrophic effects, including transcription factors (eg, heat-shock factor-1 (HSF-1), C-Jun, and CREB) and, recently, the proapoptotic Bcl-2 family member BAX. In the Wnt-signaling pathway, secreted Wnt glycoproteins interact with the frizzled family of receptors and, through disheveled-mediated signaling, inhibit GSK-3. Stability of this process requires the scaffolding proteins AXIN and adenomatous polyposis coli (APC). Normally active GSK-3 phosphorylates β-catenin, leading to its ubiquitin-dependent degradation. However, when GSK-3 is inhibited in the Wnt pathway, β-catenin is not degraded, allowing for its interaction with T cell-specific transcription factor (TCF) to act as a transcription factor. β-Catenin activity is modulated by the intracellular estrogen receptor (ER), which also affects transcription of an independent set of genes. As shown in the figure and described in the text, medications useful for the treatment of mood disorders have both direct and indirect effects on GSK-3, and GSK-3-regulated cell signaling pathways. This includes direct effects of lithium and indirect effects of antipsychotics, amphetamine, and selective 5-HT SSRIs. These distinct pathways have convergent effects on cellular processes such as bioenergetics (energy metabolism), neuroplasticity, neurogenesis, resilience, and survival. Thus, our hypothesis is that lithium (and other medications) may act by enhancing these processes through inhibition of GSK-3. Gi refers to Gi/Go; Gq refers to Gq/G11. Modified and reproduced with permission from Gould and Manji (2005).

It has been hypothesized that lithium's antidepressant effects might be associated with its upregulation of Bcl-2 (Charney and Manji, 2004). Bcl-2 has traditionally been viewed as a ‘long-term neuroprotective protein’; however, Bcl-2 is a key regulator of mitochondrial function, and there is a growing appreciation of the diverse functions that mitochondria play in regulating integrated CNS function. Thus, increasing evidence suggests that mitochondrial Ca2+ sequestration has a key role in modulating the tone of synaptic plasticity in a variety of neuronal circuits, and that regulation of mitochondrial function is likely to play an important role in regulating synaptic strength of neuronal circuitry mediating complex behaviors. Bcl-2 attenuates apoptosis by sequestering proforms of death-driving cysteine proteases (called caspases) by preventing the release of mitochondrial apoptogenic factors such as calcium, cytochrome c, and apoptosis-inducing factor (AIF) into the cytoplasm, and by enhancing mitochondrial calcium uptake (Charney and Manji, 2004; Schloesser et al, 2008). It is also noteworthy that pramipexole, a drug used in Parkinson's disease, also upregulates Bcl-2 in several brain areas, and has been shown to exert antidepressant effects in a double-blind, placebo-controlled trial in patients with bipolar II depression (Zarate et al, 2004b). While the DA agonistic effects of pramipexole may clearly also contribute to its purported antidepressant effects, its robust neurotrophic effects suggest that it may have broader utility as an antidepressant potentiator.

Phosphodiesterase Inhibitors (Major Depressive Disorder)

As seen in Figure 1 (pathway a), one approach to enhance the activity of CREB is to use an inhibitor of phosphodiesterase (PDE), the enzyme responsible for the breakdown of cAMP. Indeed, PDE4A and PDE4B may be relevant targets for development of agents that possess antidepressant effects either as monotherapy or in combination with agents that increase intrasynaptic monoamine levels, due to the possible synergism of effects on the cAMP cascade. The idea that PDE inhibitors may have potential antidepressant activity is not a new one, and was initially proposed by Wachtel in the early 1980s (Wachtel and Schneider, 1986). In the 1980s and early 1990s, a number of open-label and controlled clinical trials demonstrated that rolipram, a specific inhibitor of the high-affinity cAMP PDE4, may have antidepressant efficacy in depressed patients (reviewed in Manji et al, 2003). In addition, there is some evidence that rolipram may have a faster onset of response compared with standard antidepressants. Despite these data, the potential use of rolipram for depression was limited because of side effects such as nausea and emesis. Second-generation compounds with markedly improved tolerability are being developed (Dyke and Montana, 2002; Huang et al, 2001), and it is anticipated that the availability of CNS-penetrant PDE4 inhibitors may lead to the development of a novel class of antidepressants.

Glycogen Synthase Kinase-3 (Bipolar Disorder and Major Depression)

Glycogen synthase kinase-3 (GSK-3) is a ubiquitous, constitutively active, multi-substrate serine/threonine kinase that regulates, and is regulated by, numerous diverse signaling pathways (eg, the Wnt pathway, PI3K pathway, protein kinase A, protein kinase C, among many others; Doble and Woodgett, 2003; see Figure 2). Initial interest in GSK-3 as a therapeutic target for mood disorders arose from the seminal observations that lithium directly inhibited the enzyme (Klein and Melton, 1996; Figure 1, pathway j, while more recent data have highlighted the important role of GSK-3 regulation in the pathophysiology of mood disorders (see O’Brien and Klein, 2007; Gould and Manji, 2005; Gould et al, 2006 and references therein). Major findings relevant to both pathophysiology and potential therapeutic mechanisms of action include:

  1. i)

    GSK-3 phosphorylation is markedly regulated by agents that increase 5-HT transmission such as d-fenfluramine, fluoxetine, and imipramine such that inhibitory control of GSK-3 might occur in mood disorders marked by 5-HT dysregulation (Li et al, 2004).

  2. ii)

    GSK-3 is a major regulator of apoptosis and cellular plasticity/resilience (Crowder and Freeman, 2000; Liu et al, 2004; Franco et al, 2004; Zhao et al, 2007). Generally, increased activity of GSK-3 is proapoptotic, whereas inhibiting GSK-3 attenuates or prevents apoptosis (see Gould et al, 2006 for review).

  3. iii)

    GSK-3 has a major effect on regulating the circadian period in diverse species, an effect it shares with lithium (Gould et al, 2004b). Notably, treatment strategies are being developed derived from a chronobiological model of bipolar disorder (McClung, 2007) based on evidence that the central transcriptional activator of molecular rhythms, CLOCK, may play an important role in the DA system for mood regulation (Roybal et al, 2007).

  4. iv)

    Recent animal behavioral data (from pharmacologic and genetic models) have shown that manipulation of the GSK-3-signaling cascade produces both antidepressant and antimanic effects in models of depression or mania. Two separate GSK-3 inhibitors (AR-A014418; Gould et al, 2004a and L803-mts; Kaidanovich-Beilin et al, 2004) were found to reduce immobility in the forced swim test and attenuate amphetamine-induced hyperactivity. Animals lacking one copy of the GSK-3β gene demonstrate specific behavioral responses to lithium and other GSK-3 inhibitors (Beaulieu et al, 2004; O’Brien et al, 2004), while transgenic overexpression of GSK was associated with increased activity (Prickaerts et al, 2006). This is one of the few manipulations that have been shown to exert both antidepressant and antimanic effects, with two caveats. First, it is not unanimously accepted that amphetamine-induced hyperactivity is a valid model of manic behavior. Second, there is evidence that rolipram (discussed above) also decreases methamphetamine-induced hyperlocomotion (Siuciak et al, 2007).

  5. v)

    An increase in GSK-3β activity was observed recently in patients with major depression in a postmortem study (Karege et al, 2007), and response to lithium augmentation in major depression was associated with the GSK-3β −50T/C single-nucleotide polymorphism (Adli et al, 2007).

In view of their potential therapeutic effects not only in bipolar disorder, but also in Alzheimer's disease (Huang and Klein, 2006) and diabetes (Henriksen and Dokken, 2006), it is not surprising that specific, brain-penetrant GSK-3 inhibitors are in development by numerous pharmaceutical companies. GSK-3 phosphorylates and inhibits glycogen synthase, a downstream effector of insulin action, and overactivity of GSK-3 is associated with insulin resistance of skeletal muscle glucose transport (Schloesser et al, 2008); accordingly, GSK-3 inhibitors have been tested in prediabetic and type 2 diabetes rodent models (Henriksen and Dokken, 2006). Unless side effects prove to be prohibitive, GSK-3 inhibitors may thus represent a completely novel class of treatment for severe mood disorders, germane to both disease pathophysiology as well as to understanding the mechanism of action of existing treatments (Gould and Manji, 2005; Gould et al, 2006).

Protein Kinase C-Signaling Cascade (Bipolar Disorder)

PKC is a shared biochemical target for the actions of chronic lithium and valproate. PKC is highly enriched in the brain, where it plays a major role in regulating both pre- and postsynaptic aspects of neurotransmission. PKC is now known to exist as a family of closely related subspecies, has a heterogeneous distribution in brain (with particularly high levels in presynaptic nerve terminals), and plays a major role in the regulation of neuronal excitability, neurotransmitter release, and long-term alterations in gene expression and plasticity (Manji and Lenox, 1999). A considerable amount of biochemical data support the potential involvement of PKC in the pathophysiology and treatment of bipolar disorder. Evidence supporting this are changes in PKC and its substrates in bipolar patients, and changes in PKC-signaling pathways after treatment with lithium or valproate (Manji and Lenox, 1999). It is noteworthy that psychostimulants, which are capable of triggering manic episodes in susceptible individuals, and induce manic-like behaviors in rodents, are known to activate PKC. Thus, the evidence suggests that two structurally dissimilar antimanic agents—lithium and valproate—attenuate PKC function in a therapeutically relevant time frame, while pro-manic psychostimulants activate PKC. These data suggest that PKC modulation plays a critical role in the treatment of mania. Birnbaum et al (2004) demonstrated that excessive activation of PKC dramatically impaired the cognitive functions of the prefrontal cortex, exposure to stress-activated PKC, and resulted in prefrontal dysfunction, and inhibition of PKC (including indirectly with mood stabilizers) protected cognitive function. Pharmacological inhibition of PKC results in many behavioral changes similar to the ones induced by mood stabilizers. These include attenuation of hyperactivity, risk-taking behavior, and hedonic drive; notably, PKC inhibitors attenuate these important aspects of the manic-like syndrome in rodents (Einat and Manji, 2006; Einat et al, 2007). Importantly, recent preclinical studies have specifically investigated the antimanic effects of tamoxifen per se (since this is the only CNS-penetrant PKC inhibitor available for humans). These studies showed that tamoxifen significantly reduced amphetamine-induced hyperactivity and risk-taking behavior (Einat et al, 2007). Notably, a recent whole-genome association study of BPD has further implicated this pathway. Of the risk genes identified, the one demonstrating by far the strongest association with BPD was diacylglycerol kinase, an immediate regulator of PKC (Baum et al, 2007). These findings led to a single-blind clinical trial investigating possible antimanic properties of the PKC inhibitor tamoxifen (Bebchuk et al, 2000). While best known for its anti-estrogen properties, tamoxifen is also a potent PKC inhibitor at high concentrations. Initial results are encouraging, finding that tamoxifen treatment resulted in a significant decrease in manic symptoms (Bebchuk et al, 2000). Larger double-blind, placebo-controlled studies of tamoxifen conducted at NIMH have shown that tamoxifen indeed possesses robust antimanic effects. Subjects on tamoxifen showed significant improvement in mania compared with placebo as early as 5 days, and the effect size for the drug difference was very large after 3 weeks (Zarate et al, 2007b). PKC has multiple isoforms and PKC-mediated cellular processes are tissue- and isoform specific. This has allowed the modulation of function of individual isozymes; thus, compounds with isozyme selectivity may find themselves in the therapeutic armamentarium for bipolar disorder, although to date no such compounds have been tested in bipolar disorder (Zarate et al, 2006b).

CONCLUSION

Significant paradigm shifts in the psychopharmacology of mood disorders have not occurred in the past several decades, due to poor understanding of disease pathogenesis, imprecise delineation of phenotypic boundaries, and limitations of animal models (Berton and Nestler, 2006). While serendipity will continue to play a role in drug discovery in psychiatry, advances in animal and human genetics, molecular biology, and brain imaging will likely advance the discovery of biomarkers and identify plausible endophenotypes for subgroups of patients with these illnesses. Drugs used commonly in other medical specialties will continue to be explored in mood disorders, based on molecular understanding of their actions. As a prominent recent example, the commonly used β-lactam class of antibiotics were found to increase glutamate uptake via increased expression of glial glutamate transporter (Glt1), thus providing a novel class of compounds that may act to buffer increased glutamatergic release in stress-related conditions (Rothstein et al, 2005).

A major thrust of future drug discovery in mood disorders will continue efforts to identify agents with rapid and sustained onsets of action (such as i.v. ketamine and other NMDA receptor-selective subtype antagonists), thereby minimizing disorder morbidity and mortality in the critical weeks between initial symptom expression and drug efficacy. The development of novel PET radioligands for neurotransmitter receptors not extensively investigated to date (eg, nicotinic and muscarinic cholinergic systems) will shed light on an older literature suggesting that cholinomimetic drugs (ie, muscarinic agonists, acetylcholinesterase inhibitors) exacerbate depressive signs and symptoms such as dysphoria, psychomotor retardation, impairment of attention and memory, HPA axis hyperactivity, and sleep EEG abnormalities (Bertrand, 2005). In a proof-of-concept study, an older antimuscarinic drug, scopolamine (4 μg/kg), was observed to produce potent reductions in depressive severity compared with placebo, with rapid, robust antidepressant responses achieved in 18 patients with unipolar or bipolar depression (Furey and Drevets, 2006).

Finally, while drug discovery efforts have generally focused on specific neurotransmitter receptors or transporters as the target of intervention, there are many additional areas (eg, trafficking proteins, immune modulators) that have not been considered due to space limitations. Next-generation drugs, in addition to treating core symptoms of mood disorders, might be able to target additional important components of these conditions such as enhancing cognition (eg AMPA receptor potentiators), preventing or reversing epigenetic factors that may have long-term negative impact on the course of the illness (eg, histone deacetylase inhibitors), or reducing medical comorbidities such as diabetes (eg, GSK-inhibitors).