Novel Pharmacological Approaches to the Treatment of Depression

Dennis L. Murphy, Philip B. Mitchell, and William Z. Potter

The search to find safer, more effective and more rapidly acting antidepressants that might also benefit currently treatment-resistant patients continues unabated in the mid 1990s. This review summarizes recent trends in antidepressant drug development. It focuses on agents with relatively novel suggested mechanisms of action, and only briefly surveys other antidepressants under development that are not yet available in the United States whose mechanisms are thought to resemble those of already marketed drugs (36,69 ). Chemical structures of many of these novel agents are depicted in Fig. 1 and Fig. 2. Only an abbreviated reference list, citing reviews or more recent publications is provided, because of space limitations (see Brain Imaging in Mood Disorders).

Major progress in new antidepressant development has been slow, with the notable exception of a group of serotonin selective reuptake inhibitors (SSRIs) introduced in the last 5 years. Although the SSRI's therapeutic effects are accompanied by fewer serious side effects and overdosage hazards than the first- and second-generation antidepressants, their principal mechanism of action, neurotransmitter uptake inhibition, is certainly not a novel one.

Before considering specific drug classes, several criteria need to be considered. Discussion of the pharmacological treatment of depression depends on what patients are subsumed under this broad diagnostic category and what assessment measures and criteria are accepted as indicating a meaningful antidepressant effect. Diagnostic criteria employed in many recent outpatient studies yield populations that sometimes show a high (>50%) 6-week placebo response rate, requiring large numbers of subjects (over 100) to demonstrate significant therapeutic advantage over placebo for a new, active antidepressant. In these populations, even such established antidepressants as imipramine sometimes do not emerge as clearly superior to placebo. European and some U.S. studies of potential antidepressants that are based on comparisons with standard tricyclics alone often may reveal no difference, suggesting equal efficacy for the new drug. However, judgment may need to be reserved as to claims of efficacy when adequate comparisons with a placebo group are unavailable. Possible differences in the clinical characteristics of patients studied in U.S. versus European settings must also be taken into account in assessing response data.

In addition, some psychotherapeutic agents such as alprazolam and trazodone have sometimes uncritically been referred to as antidepressants in the broad sense, that is, they have been reported to be equivalent to firstgeneration tricyclic or other antidepressants in some studies and thus, by implication, considered equally effective in severely depressed, nonpsychotic hospitalized patients. It is doubtful whether most experienced clinicians would consider using alprazolam or trazodone as the mainstay treatment for a patient so dysfunctional with depression as to require hospitalization. On the other hand, recent comparisons provide difficult-to-ignore evidence that certain agents [e.g., the irreversible monoamine oxidase (MAO) inhibitors] may be superior to tricyclic antidepressants in subgroups of depression such as bipolar or atypical depression (66). All of this highlights a level of uncertainty about selecting and evaluating novel pharmacological approaches for treating patients with depression and allied disorders.

One possible organizing principle to survey potential novel antidepressants would be to select compounds on the basis of their activity in animal models sensitive to antidepressant drug effects. However, this approach does not answer the question, because many animal models have multiple limitations for identifying novel antidepressants, ranging from circularity (e.g., the test merely reflects a biochemical property of standard drugs) to failure to detect any activity of certain established agents (e.g., inability of the MAO inhibitors to reverse passiveavoidance deficits) (78, 84). Although some widely used behavioral paradigms (for example, the learned helplessness paradigm, the forced swim test, and the restraint stress-induced reduction of locomotor activity) yield positive results for most currently used antidepressant drugs, false positives and false negatives have been found for some novel compounds. Only clinical testing can definitively establish the utility of a new compound in the treatment of depression, whether narrowly or broadly defined, and a strong case can be made for clinical serendipity being involved in the discovery of most effective antidepressant agents thus far identified. Ultimately, we must depend on the observations of skilled clinicians employing well-targeted populations to find truly novel antidepressants with distinct efficacy in areas of need, rather than on drug–placebo differences established solely in large trials with relatively fewer impaired outpatients.

The single largest class of new antidepressants in the last decade is the SSRIs, but because several of these agents, namely fluoxetine, sertraline, and paroxetine, have recently been approved for use in the United States, only a few features of these agents specifically relevant to the development of new drugs affecting serotonin transport will be noted in this chapter. These agents, however, have drawn special attention to the potential involvement of brain serotonin subsystems in the modulation of mood, anxiety, sleep, and other physiological functions relevant to depressive symptoms (18).

The inhibition of serotonin (5-HT) uptake is by far the most established initial mechanism of action of serotonergic antidepressants; other non-uptake-inhibiting serotonergic drugs are active in animal models sensitive to antidepressant drugs, and have at least a modicum of clinical evidence supportive of antidepressant effects, including the following: (a) several 5-HT1A partial agonists, including buspirone, gepirone, ipsapirone, and flesinoxan (39); (b) a 5-HT2C partial agonist, m-chlorophenylpiperazine, with additional antagonist effects at 5-HT2A and 5-HT3 sites (58); (c) other 5-HT2A/2C antagonists such as ritanserin, etoperidone and nefazodone (3, 26, 37); and (d) an enhancer of 5-HT uptake, tianeptine (44). Inhibitors of serotonin metabolism by MAO type A (MAO-A) inhibition, for which enzyme serotonin is a highly selective substrate, such as moclobemide, brofaromine, and clorgyline, are also effective antidepressants but are not available in the United States (15, 47, 83).

Reconciliation of the diverse serotonin receptor and related neurochemical mechanisms of action for the antidepressant effects of these agents is not immediately apparent. When consideration is given to adaptational events that follow repeated administration of these agents (as chronic administration is required for clinical therapeutic efficacy), even more complex schema are required. In the case of the SSRI's, a case for a net change leading to enhanced serotonergic neurotransmission responsible for antidepressant efficacy has been postulated, based upon drug-induced changes in electrophysiological events at serotonergic synapses and changes in releasable serotonin measured in microdialysis studies (28). However, because the more than 13 molecularly identified serotonin receptors may have opposing rather than potentiating effects measured on even one target cell, to say nothing of the multiple other neurotransmitter systems (dopamine, g-aminobutyric acid, norepinephrine, and peptide systems) that are modulated by the widespread serotonin subsystems (45), adequate interpretations of mechanisms of action of these drugs would seem some distance in the future.

Buspirone is the one representative of this drug class available in the United States; it is approved as an anxiolytic but not yet as an antidepressant. Studies of buspirone and related azapirones including gepirone and ipsapirone, as well as the prototypical 5-HT1A agonist, 8-hydroxy-2-(di-n-propylamino)tetralin(8-OHDPAT), revealed potential antidepressant effects in such animals models as the forced swim test and other stress-induced behavioral deficit tests (71).

Buspirone and gepirone both had significantly greater effects than placebo in patients with major depression in a series of studies that included over a hundred patients, each studied under double-blind conditions (39). Melancholic patients improved to an equal extent as nonmelancholic patients, which was interpreted as evidence that the moderate overall degree of improvement was unlikely to represent only a reduction in anxiety symptoms associated with depression. A smaller study that compared ipsapirone and a placebo in neurotic depressed patients demonstrated significantly greater antidepressant effects for ipsapirone (35). However, the lack of additional, more recent major reports on antidepressant efficacy of the azapirones is of concern.

Studies of the preclinical neuropharmacology of buspirone and related azapirones went through several phases, because the initial focus of the investigations was on antianxiety models. Additionally, buspirone elicited dopamine system effects not found with gepirone or ipsapirone that were puzzling to interpret. It has since become increasingly clear that the activity of these agents in animal models of anxiety and depression fits best with their 5-HT1A partial agonist effects (89).

The predominant initial effect of these drugs in rodents is a slowing of the firing rate of dorsal raphe neurons and, to a considerably lesser extent, medial raphe neurons, a consequence of a direct full agonist effect of these agents on autoreceptor 5-HT1A sites located in the midbrain raphe areas (9). In addition, these drugs act as partial agonists at postsynaptic 5-HT1A sites in the hippocampus and other forebrain areas, although drug concentrations required are up to tenfold higher at these sites, and, among both the azapirones and some newer 5-HT1A agonists with different chemical structures, partial agonist and antagonist properties differ (9, 74). With sustained administration for 10 days or more, some desensitization of the raphe response occurs when measured electrophysiologically (9). However, microdialysis studies indicate sustained reductions in hippocampal 5-HT release during chronic treatment (74). Lesion studies have suggested that the actions of these drugs in animal models of anxiolytic drug effects fit best with a primarily presynaptic mode of action. In contrast, the action of these drugs in animal antidepressant models is blocked by 5-HT1A antagonists but is essentially unaffected by lesions or by drugs that decrease presynaptic function such as the tryptophan hydroxylase inhibitor, p-chlorophenylalanine, which suggests that postsynaptic 5-HT1A partial agonist effects are more important for their antidepressant effects (8, 51, 71). An additional mechanism that might contribute is the metabolism of the azapirones to 1-(2-pyramidal)-piperazine (1-PP), which achieves brain concentrations tenfold higher than the parent component (13); 1-PP is an a2-adrenergic antagonist, a drug class that has been hypothesized to have antidepressant properties (see below). In one clinical study, 1-PP plasma levels were significantly correlated with improvement in depressive symptoms in patients treated with buspirone (79).

The moderate efficacy of the selective 5-HT2A/5-HT2C antagonist ritanserin in several double-blind controlled trials in depressed, dysthymic, and anxiety disorder patients suggested the possibility of a different profile of therapeutic action than that found with earlier, less serotonin receptor subtype-selective 5-HT1/2 antagonists such as methysergide, cyproheptidine, or metergoline. Only one study using ritanserin was placebo controlled; in others, ritanserin's effects equaled those of amitriptyline or flupenthixol, and were superior to trazodone (3). Another related agent, nefazodone, like trazodone, has prominent 5-HT2A and probably 5-HT2C antagonist effects; these effects may contribute more to the clinical effects of nefazodone and trazodone than the relatively weak uptake inhibitory effects of these agents on 5-HT (both) or 5-HT and norepinephrine (nefazodone) (26, 37). Of note, mianserin, an earlier antidepressant in wide use in Europe and other regions, but not available in the United States, also has prominent 5-HT2A,2C antagonist properties (12, 37).

The recently available SSRIs are more selective than the tricyclics and other earlier antidepressants in their lack of effects at cholinergic, adrenergic and other neurotransmitter receptors, which improves their side-effect profile and reduces the risk of overdosage. Other SSRIs, such as citalopram, fluvoxamine, and Ro-15-8081 are also under development, along with some novel compounds that combine serotonin uptake-inhibiting actions with 5-HT1A agonist effects (LY 233708) or 5-HT3 antagonist effects (litoxetine) (69). With advancing knowledge of serotonin receptors, it is of note that (-)-norfluoxetine, an active fluoxetine metabolite with a longer half-life than the parent compound, has only slightly lower affinity for the 5-HT2C receptor than for the 5-HT transporter in rat brain (87). Other examples of more complicated molecular interactions are the newly discovered 5-HT6 and 5-HT7 receptors, which have very high affinities for a number of tricyclic and other psychotropic drugs, including clomipramine, amitriptyline, ritanserin, mianserin, and clozapine, suggesting additional sites of action for these drugs (50, 57).

Another interesting focus has been on drugs with combined 5-HT and norepinephrine uptake-inhibiting effects, but lacking tricycliclike interactions with neurotransmitter receptors, a therapeutic strategy that reflects recent evidence that 5-HT selective and norepinephrine selective uptake inhibitors act through separate serotonergic or noradrenergic mechanisms to yield equivalent antidepressant effects (55); the dual action uptake inhibitors might affect patient subgroups insensitive to 5-HT or norepinephrine changes alone. Venlafaxine and duloxetine (formerly LY 248686) are two examples of such agents; a substantial number of controlled clinical trials have demonstrated definite antidepressant efficacy for venlafaxine, which has a side-effect profile similar to that of the SSRIs (36, 86).

A novel 5-HT uptake modulatory agent is tianeptine, which appears to enhance 5-HT uptake in vivo—in contrast to the 5-HT uptake-inhibiting action of the SSRIs and earlier tricyclics. Only limited clinical data is available for tianeptine. One double-blind study in 265 patients with dysthymic disorder and anxiety found tianeptine to be equipotent with amitriptyline (32). A similar result was found for tianeptine and amitriptyline in depressed alcoholics (49). In general, tianeptine appears to be well tolerated, with no anticholinergic or cardiovascular effects. Adverse effects leading to treatment discontinuation in these clinical trials included insomnia, anxiety, irritability, dizziness, nausea, and vomiting.

In a number of animal models sensitive to antidepressant drugs, tianeptine's profile was indicative of antidepressant activity: Tianeptine decreased aggressive behavior in the isolated mouse, decreased reserpine-induced ptosis and hypothermia, suppressed behavioral despair in the forced swim test, and also blocked pontogeniculooccipital waves induced by Ro 4-1284 in the cat (44). Tianeptine also antagonized deficits in open-field activity after a two-hour restraint, as well as deficits induced by immobilization stress; it also was active in the learned-helplessness paradigm (21, 78).

Tianeptine is an atypical tricyclic molecule with a substituted dibenzothiadiazepine nucleus containing two heteroatoms, and a long amino-heptanoic chain with a terminal acidic group. It is similar in structure to the older antidepressant, amineptine (discussed below) (44). Enhanced 5-HT uptake was first reported in brain synaptosomes from rats treated either acutely or chronically with tianeptine; it did not result from a reduction in 5-HT release, and was not found upon in vitro addition of the drug to synaptosomal preparations. In some studies, tianeptine led to an enhanced depletion of 5-HT, increased brain 5-hydroxyindoleacetic acid brain (5-HIAA) concentrations, and an attenuation of the potassium-induced rise in hippocampal extracellular 5-HT (44).

In an electrophysiological study that examined the effect of both tianeptine and clomipramine on the 5-HT-induced reduction of firing rates of rat hippocampal CA1 pyramidal cells, tianeptine accelerated the recovery of firing rates after 5-HT iontophoresis, in contrast to the delay in recovery found after clomipramine—a result consistent with tianeptine reducing intrasynaptic 5-HT concentrations (25). A similar study demonstrated a reduction in the time required for recovery of neuronal firing activity following microiontophoretic application of 5-HT in rats treated chronically, but not acutely, with tianeptine. Further studies indicated that the enhancement of 5-HT uptake by chronic tianeptine did not appear to be from an adaptive alteration of the 5-HT transporter (65), a finding consistent with earlier studies which found that tianeptine did not affect 3H-imipramine binding. Thus, the mechanism of this 5-HT uptake-enhancing action of tianeptine remains elusive.

Enhancement of 5-HT uptake by tianeptine also occurs in humans, as demonstrated by a study of platelet 5-HT uptake in depressed patients that found an increase in apparent Vmax which persisted after chronic treatment (16). Tianeptine has no effect on either 5-HT receptors (5HT1A/1B/2) or b-adrenergic receptors. It reduced acetylcholine release, an effect apparently mediated by an action on serotonergic neurons (6). Tianeptine also increased concentrations of dopamine and its metabolites in rat striatum, nucleus accumbens, brainstem and cerebral cortex (38). There is some uncertainty as to whether these dopaminergic effects are primary effects of tianeptine or secondary to the serotonergic actions of this agent.

Drugs that directly or indirectly facilitate dopamine function have often been used as either primary agents or adjuncts in the treatment of depression, especially for patients not responding to standard noradrenergic or serotonergic uptake inhibitors. Such compounds have included the psychostimulants amphetamine and methylphenidate, the antiparkinsonian bromocriptine, the MAO inhibitors, and, while it was available, the dopamine uptake inhibitor, nomifensine. Nomifensine gained a reputation for being particularly effective in treatment-resistant patients with chronic depression. Preclinical investigations have demonstrated that most antidepressants, whatever their other actions, affect dopamine function. Facilitation of dopamine function is striking after repeated electroconvulsive therapy. Whether antidepressant effects on dopamine are primary or secondary (e.g., to primary facilitation of 5-HT, see above) is unknown (for reviews see refs. 67 and 73). Nonetheless, it has been hypothesized that some depressed patients require compounds with more potent and direct dopaminergic effects to achieve a therapeutic outcome.

Amineptine was originally marketed as an antidepressant in France and is now available in some other European, African, and South American countries. In 11 controlled studies, amineptine has generally been found to be of equivalent therapeutic efficacy to established antidepressants, although only two of these studies exceeded 50 patients per treatment group, many used relatively low doses of standard drugs, and the only study that included a placebo group was of depressed inpatients with psychomotor retardation, who showed greater improvement with amineptine compared to placebo, clomipramine, or minaprine (68). This result may be pertinent to the fact that amineptine has stimulant effects and associated abuse potential.

Amineptine is an atypical tricyclic, differing from the typical compounds in its 7-aminoheptanoic acid side chain. In animal screening models, amineptine antagonized reserpine-induced ptosis and hypothermia, decreased immobility in a behavioral despair test, and ameliorated impaired social behavior in monkeys; it also antagonized apomorphine-induced hypothermia and behavioral responses and had direct locomotor stimulatory effects in mice (14, 30). Amineptine selectively inhibited the uptake of dopamine into rat striatal synaptosomes, and, at higher concentrations, also caused a release of dopamine, norepinephrine, and serotonin. A microdialysis study found dose-dependent elevations of dopamine in the striatum, nucleus accumbens, and frontal cortex, without changes in dopamine metabolites (38). Chronic administration of amineptine reduced the number of striatal dopamine binding sites (14).

Among other dopamine uptake inhibitors, medifoxamine is a monoamine reuptake inhibitor that preferentially inhibits dopamine uptake (70). Although in clinical use in France, there has been only one report of antidepressant efficacy (23), and there are few studies of its neurochemistry. GBR12909, GBR12783 and GBR13069 are piperazine compounds with potent, highly selective dopamine uptake blocking effects and locomotor stimulatory effects; GBR12909 is active in animal models for antidepressant drug effects, but apparently has not yet been evaluated in humans (2!popup(ch107).

Several direct dopamine agonists were evaluated for antidepressant effects in the 1980s, including bromocriptine and piribedil (40). The lack of recent studies may be because of both a high incidence of adverse effects and a rapid loss of efficacy with continued administration. A recent animal study found no effect of bromocriptine in the forced swim test (60). Lisuride is a centrally acting dopamine and serotonin agonist of the ergot type, which is marketed in Europe as an antidepressant (81). Animal screening studies indicate that lisuride may have clinical antidepressant properties. Unfortunately, there are no published controlled clinical studies to confirm its possible antidepressant activity. Lisuride also has very high affinity for the 5-HT6 and 5-HT7 receptor sites, which also exhibit high affinity for a number of other agents with antidepressant and antipsychotic efficacy (57). Roxindole is a potent dopamine autoreceptor agonist, which also inhibits 5-HT uptake and has 5-HT1A agonistic actions. It was originally developed for the treatment of schizophrenia and has been found to be particularly effective in the treatment of the negative features of schizophrenia. A recent open trial of roxindole in patients with major depression indicated that it may also possess antidepressant properties (31).

Somewhat paradoxically, certain drugs that block D2 and/or D1 receptors and are marketed as atypical antipsychotics, sulpiride (in Europe) and clozapine, are also reported to have antidepressant properties. Sulpiride has been reported to have equivalent antidepressant properties to amitriptyline in a double-blind comparison (76) and clozapine's combination of properties has been suggested to be particularly beneficial in schizoaffective disorder. Antidepressant properties of clozapine could be related either to its 5-HT2 or a2-adrenergic antagonistic properties, the latter suggested by the finding that in a controlled study the addition of the a2-antagonist idazoxan (see below) to a standard neuroleptic produced improvement similar to that seen with clozapine in refractory schizophrenic patients (48). No controlled trials, however, of clozapine in any form of depression, even psychotic depression, have yet been reported.

Classification of putative antidepressants as primarily noradrenergic agents rests on preclinical demonstrations of selective norepinephrine (NE) uptake inhibition, a2-adrenergic antagonism, and/or agonistic effects at a1- or b-adrenergic receptors. It is noteworthy that, to date, among drugs with these actions, only NE uptake inhibition has been consistently associated with antidepressant responses in double-blind, placebo-controlled trials. Perhaps because alterations of noradrenergic function are so often found in depressed patients and often follow chronic treatment with any type of antidepressant, the search for drugs selectively facilitating NE throughput continues.

For many years, it has been hypothesized that inhibitors of a2 receptors would functionally have a sufficient effect on central nervous system (CNS) a2-adrenergic autoreceptors to increase intrasynaptic NE and therefore reproduce the consequences of NE reuptake-inhibiting antidepressants (64). The clinical efficacy of mianserin, which has some a2-blocking properties in the absence of other effects on NE sites, has been invoked as supporting this notion (24), although it is now appreciated that mianserin more potently antagonizes 5-HT2A and 5-HT2C receptors (see above).

Preclinical studies showing that the addition of an a2 antagonist to a tricyclic accelerated the rate of badrenergic receptor down-regulation in rat cortex led to a trial in which yohimbine, the most selective a2 antagonist available for clinical use at the time, was added in patients who had failed to respond to monotherapy with desipramine. In this group, the addition of yohimbine did not produce significant improvement (17).

More recently, both the imidazoline, idazoxan, and the benzodioxinopyrrole, fluparoxan, which show favorable a2/a1 potency ratios have been available for clinical studies (34). Although controlled trials of both are said to have been carried out in several hundred depressed patients, no reports have emerged. Thus, it is uncertain whether monotherapy with selective a2 antagonists is effective in unipolar depression. A preliminary report suggests that idazoxan may be an antidepressant in bipolar depression alone or in combination with lithium (Osman O. T., et al., personal communication). Furthermore, it is possible that the compounds available to date do not show the effects that would be achieved with a truly selective, full a2 antagonist without partial agonist activity and without the binding at imidazoline sites shown by idazoxan. A Belgium compound, R47,243, may be such a potent full antagonist (20) but it does not appear to have been developed for clinical trials.

Three selective NE uptake inhibitors have been recently reported, two of which are nontricyclics: tomoxetine, a benzenepropanamine; viloxazine, a morpholine; and the newer tricyclic, lofepramine. Since these do not involve a novel mechanism, they are not discussed further.

Modafanil is an interesting compound; it is structurally unrelated to any known antidepressant. It has been used in France for a number of indications including depression, although there have as yet been no published studies establishing clinical antidepressant efficacy. Most of the published investigations have examined its psychostimulant properties in humans and animals (10). It has been tentatively classified as an a1-adrenergic agonist based on the ability of a1 antagonists to block its behavioral activity in animals. Another French compound, adrafinil, is also classified as a central a1 agonist and is said to be helpful for depression and other symptomatology in cognitively impaired subjects (22).

There continues to be an interest in indirect noradrenergic agonists, although salbutamol and clenbuterol, the latter almost exclusively a b2 agonist, show at best mixed evidence of antidepressant efficacy (reviewed in ref. 67). Because those established antidepressants shown to down-regulate b-adrenergic receptors consistently reduce the b1 subtype and clenbuterol reduces only b2 receptors, there remains the possibility that a different type of b agonist without limiting peripheral effects would possess antidepressant potency. SR58611A may be such an agent and is being proposed for clinical studies based on its novel identity as a selective atypical b agonist (75).

The first true antidepressants were nonselective, irreversible inhibitors of both MAO-A and MAO-B (e.g., phenelzine and tranylcypromine). The discoveries that antidepressant efficacy required only MAO-A and not MAO-B inhibition (47) and that the hazardous potentiation of pressor responses to dietary tyramine that accompanied MAO-A inhibition could be markedly attenuated by reversible inhibitors of MAO-A led to the current situation where there are at least ten new MAO-A inhibitors under development; unfortunately, none are yet available in the United States (15). A similar number of selective MAO-B inhibitors are also under development, primarily because neuroprotective and moderate treatment effects were reported for the MAO-B inhibitor selegiline (formerly named (-)deprenyl) in Parkinson's disease and dementia (15, 59), although it is noteworthy that when given in higher non-MAO-B selective doses, selegiline is an effective antidepressant in treatment-resistant elderly depressed patients as well as younger patients (77).

Moclobemide is probably the most well-studied reversible, selective MAO-A–inhibiting antidepressant now available in Europe, Canada, and much of the world, except the United States. Efficacy has been demonstrated in multiple comparisons with placebo and conventional antidepressants of all types, including irreversible MAO inhibitors (27). A metaanalysis of over 2000 patients yielded evidence of equivalent efficacy for unipolar and bipolar patients, endogenous and nonendogenous, melancholic and nonmelancholic patients and younger and older patients. The drug is generally well tolerated with no anticholinergic or cardiovascular side effects; only nausea appeared more frequently as a side effect after moclobemide versus placebo treatment (27).

Dietary restrictions for moclobemide have been minimized to include only avoidance of very large servings of possibly tyramine-rich cheeses; the drug is taken after meals to lessen potential interactions with foods containing tyramine. Moclobemide is a weak MAO inhibitor in vitro, but apparently is rapidly and extensively biotransformed to a yet unknown predominantly MAO-A selective metabolite(s). Its effects on rapid-eye-movement sleep suppression and on catecholamine metabolite changes are somewhat less than those found with irreversible MAO-A selective inhibitors; as with earlier MAO-inhibitors, whether moclobemide's clinical effects result from delayed consequences of the enhanced synaptic availability of neurotransmitter amines like serotonin and norepinephrine or from adaptive responses at autoreceptors and other receptors and their signal transducing systems remains conjectural (15).

Another example of a reversible selective MAO-A inhibitor is brofaromine. Its MAO-A selectivity (100-fold) and relative reversibility over 48 hr depend upon in vivo kinetics, as brofaromine in vitro manifests only slight selectively for MAO-A versus MAO-B and only slowly dissociates from the enzyme (15, 83). Brofaromine has a second property at higher doses—serotonin uptake inhibition in vitro and in vivo—which has been suggested to contribute to its clinical efficacy, because MAO-A inhibition becomes maximum at 50 mg/day doses, although clinical studies indicate increasing antidepressant efficacy with increasing doses up to at least 150 mg/day (83).

Minaprine is a nontricyclic compound that appears to enhance dopaminergic, serotonergic, and cholinergic transmission. Of the nine double-blind studies of minaprine published, eight compared minaprine with standard antidepressants and reported equivalent efficacy, although no study had 50 or more subjects per treatment group and doses of the comparison agents were often low. The single placebo-controlled study is worthy of note, because it compared four doses of minaprine (100 to 400 mg/day) with placebo, and found a significant difference from placebo only at 400 mg/day, a finding of particular concern as the usual dose range employed in other studies was {ewc MVIMG, MVIMAGE,!lesseq.bmp}300 mg/day (1). In general, minaprine was well tolerated. Animal studies indicated no potential for abuse despite its enhancement of dopamine function. There have been no clinical reports of stimulant effects or abuse, and the most common side effects are nausea, anxiety, and insomnia.

Minaprine, which is a 3-amino-6-phenylpyridazine derivative, has an atypical molecular structure. In animal models for antidepressant drugs, minaprine antagonized reserpine-induced ptosis in mice, decreased immobility in a behavioral despair test, and antagonized muricidal behavior in rats (7). Minaprine has been suggested to enhance dopaminergic, serotonergic, and cholinergic neurotransmission, although there are relatively few basic pharmacological studies of this compound. Minaprine given to rats reduced striatal homovanillic acid (HVA) and dihydroxyphenylacetic acid (DOPAC) concentrations, while concomitantly increasing levels of 3-methoxytyramine in a dose-dependent fashion; such findings are consistent with increased dopaminergic transmission, although the mechanism of this enhancement is unclear. A weak reversible inhibition of MAO-A might best explain these neurochemical changes (41). Minaprine has no effect on dopamine release or uptake; it also has no affinity for dopamine D1 or D2 receptors. Low doses of minaprine antagonized haloperidol-induced catalepsy, and higher doses induced stereotypic behavior in rats, consistent with enhancement of dopamine activity (7).

Additionally, minaprine has been shown to increase 5-HT and decrease 5-HIAA concentrations in rat cortex, striatum, brainstem, and hypothalamus (7). As with its dopaminergic actions, these effects would be consistent with MAO inhibition. Minaprine affected neither the release nor uptake of 5-HT. Similarly, it had no affinity for either 5-HT1 or 5-HT2 receptors. Minaprine also produced an increase in acetylcholine concentrations in rat striatal, hippocampal, and cortical regions (29). Minaprine displaced the M1-muscarinic antagonist pirenzapine from cortical and hippocampal binding sites and antagonized intrastriatal pirenzapine-induced rotations in rats. These findings suggested the possibility that the cholinomimetic action of minaprine may be mediated by affinity for the M1 receptor. However, other studies have suggested that the cholinomimetic action may be primarily mediated by effects on a serotonergic subsystem by 5-HT1B or 5-HT2 receptors (11). Whatever the specific mechanism, minaprine has aroused interest as a potential agent for patients with cognitive impairment, particularly because there is some evidence that it may improve memory in animal models (88).

Rolipram (a dialkoxyphenyl-2-pyrrolidone) is one of a series of isoenzyme-selective phosphodiesterase inhibitors that have been developed for the treatment of a broad range of medical disorders (33). It increases the availability of cyclic adenosine monophosphate (cAMP) by inhibiting cAMP phosphodiesterase and enhances the availability of NE by stimulating tyrosine hydroxylase and NE release. Despite marked optimism in the 1980s that this agent appeared to be a unique antidepressant, recent controlled trials have thrown considerable doubt upon the clinical efficacy of this compound. Although early controlled trials indicated clinical antidepressant efficacy, more recent studies have found rolipram to be less effective than either imipramine or amitriptyline (72). For example, in the latter study of 50 patients with DSM-III major depression, only 8% recovered with rolipram compared to 44% with amitriptyline. These two negative studies have clouded the future investigation of this compound as a potential antidepressant. Rolipram's action in animal screening models indicated its potential as a clinical antidepressant. It reversed the effects of reserpine on body temperature and locomotor activity and reduced muricidal behavior in olfactory bulbectomized rats (56, 82). There is some uncertainty, however, about the mechanism of these effects. Because they were not prevented by depletion of monoamines or blockade of central b-adrenergic or dopamine receptors, rolipram appeared to be acting by a postreceptor mechanism (82). However, forskolin activation of adenyl cyclase did not replicate the behavioral effects of rolipram, suggesting that rolipram might not act by increasing cAMP (63).

Inositol is essential for the synthesis of phosphatidyl inositol, the precursor of the important second messenger for many neurotransmitter receptors, phosphatidyl inositol triphosphate. Although inositol crosses the blood–brain barrier with difficulty, large inositol doses (12 g) increased inositol concentrations in human cerebrospinal fluid (CSF) by 70% and led to improvement in depression ratings in an open trial (46). A recent double-blind controlled trial demonstrated improvement in depressed patients treated with 12 g/day inositol compared to placebo (46). Although inositol apparently has not been studied in any animal models sensitive to antidepressant drug effects, antidepressants from several classes increase cellular concentrations of inositol triphosphates.

Captopril, one of the angiotensin-converting enzyme (ACE) inhibitors, has been reported to have mood elevating properties in some patients (19). This is consistent with the observation that animal screening models indicate putative antidepressant activity for this compound (54), although there has been one negative study (5). Formal clinical trials in depressed patients have not been reported.

Investigations of patients with bipolar disorder who demonstrated increased CSF calcium levels in the depressed phase have suggested a possible role for this cation in mood disorders, and a potential role of calcium channel modulators in their treatment. Most studies have focused on nifedipine. Animal studies indicate that nifedipine blocks immobility in behavioral despair tests, suggesting a potential for antidepressant action (42). The only clinical trial of nifedipine, however, found no antidepressant efficacy (43). This finding would appear to have been emphatically confirmed by other reports that nifedipine may, in fact, induce depression (52).

As elevated corticotropin-releasing hormone (CRH)/adrenocorticotropic hormone (ACTH)/cortisol activity is frequently associated with depression, consideration has been given to treatments that reduce glucocorticoid function. Positive open clinical trials have been reported for amino glutethamide, ketoconazole, and metapyrone in treatment-resistant depressed patients (58) and ketoconazole in hypercortisolemic depressed patients (85). Animal model studies using either normocortisolemic animals or the hypercortisolemic Fawn-hooded rat strain that manifests depressionlike behavior (4) have not yet been completed.

In two models for antidepressant drug effects, the forced swim test in mice and rats and tail suspension-induced immobility in mice, a noncompetitive N-methyl-D-aspartic acid (NMDA) antagonist, dizolcipine (MK-801) and a partial agonist at the glycine regulatory site of the NMDA receptor, 2-amino-7-phosphoheptanoic acid (AP-7), produced antidepressant-like effects (53, 80). Although apparently no clinical antidepressant trial data with tolerated agents that have comparable effects on the NMDA receptor are available, these results have engendered considerable interest, as imipramine, AP-7, and 17 representatives of almost all major classes of antidepressant drugs given chronically, but not acutely, led to common, possibly adaptive changes in ligand binding at the cortex glycine binding receptor site of the NMDA receptor (61, 62) and to associated changes in NMDA receptor mRNA (K.-P. Lesch, et al., personal communication).

In the present search for novel pharmacological approaches to the treatment of depression and allied disorders, we have taken a broad inclusive perspective of depression. Given that drugs principally classified as antidepressants are also effective in other disorders (e.g., obsessive–compulsive disorder, panic disorder, social phobia) and in some other neuropsychiatric patient populations, this strategy seemed justified. Thus, any compound that has shown some significant therapeutic benefit in any form of depression has been considered to be a possible candidate, as long as it also appeared biochemically distinct from currently available antidepressants. In many instances, considerable additional clinical-trials data in appropriate patient populations and subpopulations employing placebo groups will be needed to clearly establish efficacy. Regrettably, there exists at present an almost universal inverse relationship between novelty of action and convincing clinical evidence of efficacy for the new agents under development considered in this review. Nonetheless, there is some basis for cautious enthusiasm in the small hints from both clinical investigations and animal model studies that at least a few of the novel agents considered in this review may eventually contribute to our therapeutic armamentarium.

published 2000