Agomelatine, melatonin and depressive disorder

Abstract

Alteration of nocturnal melatonin production, along with circadian rhythm disturbance, has been demonstrated in several psychiatric disorders. It has been postulated that such disturbances might be causal reflecting a more fundamental abnormality of the function of the suprachiasmatic nucleus (SCN). The SCN contains the body's master ‘clock' while the pineal-SCN nexus is intricate to the nighttime production of melatonin. The more compelling case for causality is made for major depressive disorder (MDD). Lending weight to this proposition is the introduction of agomelatine as an antidepressant agent. Through its actions on melatonin receptors agomelatine can resynchronise circadian rhythms. The circadian hypothesis would posit that normalisation of disturbance would be sufficient of itself to alleviate the symptoms of MDD. Thus, strategies designed to bring about resynchronisation of circadian rhythms should be therapeutically effective in depression. Critical examination of the efficacy of such interventions in MDD suggests that the circadian alteration may be necessary but is not sufficient for an antidepressant effect. Exogenous melatonin administration and bright light therapy have mixed results in limited controlled clinical evaluations. Furthermore, agomelatine has other actions which pre-clinical studies suggest are as important to its therapeutic effects as are its actions on melatonin receptors ipso facto its resynchronising properties. Whether circadian effects are antidepressant remains a moot point and awaits the clinical evaluation of highly selective resynchronising agents.

Keywords::

• agomelatine
• antidepressant
• circadian rhythms
• clock genes
• major depression
• melatonin
• melatonin receptors
• resynchronisation

1. Introduction

A recent paper in this journal canvassed the proposition that normalisation of circadian rhythms are therapeutic in the alleviation of the symptoms of major depressive disorder (MDD) [1]. Abnormalities of circadian rhythms associated with MDD have attracted renewed interest due to the demonstration that the novel naphthalenic derivative, agomelatine is effective as a treatment [1]. Further, the demonstration that agomelatine is capable of entrainment of the 24-h cycle (circadian cycle) provides a point of conjunction [2]. Efficacy of agomelatine as an antidepressant has been established in clinical trials and extensively, critically reviewed in the literature [3-8]. Coupled with its efficacy, the relatively benign side effect profile, at least in the doses used in clinical trials (up to 50 mg/day), would suggest that the compound is ideally suited to the treatment of the majority of cases of MDD [1]. On the other hand, clinical experience with the agent suggests that such therapeutic optimism is misplaced and that like all other extant antidepressants, agomelatine suffers similar well-known shortcomings: delayed onset of action, not effective in all patients and some emergent treatment-related events of concern (e.g., potential liver toxicity) [9]. Nevertheless, that the drug is effective at all is heuristically significant in that its mechanism of action represents a departure from the well-trodden path of monoamine influencing agents (reuptake inhibitors, neurotransmitter catabolic enzyme inhibitors). Of greater significance is whether agomelatine will represent the ‘first in class' of therapeutic agents for the treatment of depression, or if the circadian process can be exploited as a potential resource for therapies. The evidence as set out by Srinivasan et al. is a moot point indeed.

2. Circadian rhythms and depression

It is not the purpose here to review extensively the relationship between depression and circadian rhythm abnormalities which has been reviewed elsewhere [10]. However, an alteration of the endogenous circadian rhythm in depression has been suggested since antiquity and ascribed by Burton as causal in the middle of the 17th century. Confirmatory evidence for such putative abnormalities was not forthcoming until the 20th century when sophisticated methodologies were available. Demonstrable differences in patients compared with control subjects have been established for symptom rhythms, some hormones and for basal body temperature [11]. A common presenting feature of major depression is diurnal variation in mood with a marked improvement towards evening [12]. Additionally, early morning insomnia, another common symptom of depression, points to disturbances of the circadian sleep cycle [13]. Recent studies have demonstrated a relationship between misalignment of circadian phase and severity of depression: the more delayed the more severe the symptoms of illness [14]. Misalignment of the internal clock and sleep timing is ‘depressogenic' in vulnerable patients.

Circadian abnormalities may be specific to different subtypes of depressive disorder. For example, the well-recognised hyper-cortisolaemia of depression occurs in only a subgroup of patients and could reflect a shift in circadian phase of the rhythm (both phase advances and delays have been reported dependent on factors such as age and gender) [10]. Similarly, although the amplitude of the nocturnal melatonin rhythm is frequently reported as blunted in major depression, this too is not universally observed, perhaps suggesting a subtype-specific phenomenon [10]. The blunting of the amplitude might be indicative of more generalised receptor subsensitivity (β₁-adrenoceptors?) including the pineal gland, although this is highly speculative. Furthermore, melatonin secretion is markedly affected by certain medications, most notably benzodiazepines and β-blockers, which suppress melatonin production [15]. Thus, findings of suppressed nocturnal melatonin secretion in depression may be an artefact of hypnotic use.

At a more fundamental biological level, an association of polymorphisms in genes controlling circadian rhythms (so-called ‘clock genes') with the diagnosis of depression (i.e., depression as a distinct entity) has not been consistently demonstrated. On the other hand, two lines of evidence suggest a link: the association of polymorphisms of some circadian genes and features of depression (such as age at onset, diurnal activity pattern, insomnia) and the effects of antidepressant medications on circadian gene expression. A recent study suggested an association between haplotypes in the cry1 and npas2 genes and unipolar major depression [16]. This study provides some support for the involvement of the circadian system with depression, suggesting that variations may, at least, represent a vulnerability factor(s) for the disorder. One major difficulty for genetic studies is that ‘depression' is unlikely to represent a single nosological entity. Until such times that clearly defined depressive subtypes are delineated, genetic associations are likely to be obfuscated by the noise of using an indistinct category of disorder.

Gene expression studies have shown that the repeated administration of the serotonin-specific reuptake inhibitor (SSRI) antidepressant, fluoxetine has effects on the expression of clock genes in the hippocampus and striatum of mice [17]. Thus, after repeated doses of 10 mg/kg there was increased expression of clock, bmal1 and npas2 and suppressed expression of per1 in the hippocampus of treated mice compared with controls. Fluoxetine decreased per1 and per2 expression in the caudate. Given the putatively critical role of the hippocampus in mediating the effects of antidepressants [18], changes observed in this region may be particularly relevant. Since the changes occurred after repeated, not single, doses, this suggests relevance to the therapeutic effects of the drug.

The central pacemaker of the circadian system is located in the suprachiasmatic nucleus (SCN) of the hypothalamus. The effects of lesions of the SCN on measures of anxiety and depression in rodents have been used to examine the role of the circadian pacemaker in these disorders. Bilateral lesions of the SCN in rats were found to have an antidepressant-like effect in the forced swim test: decreased immobility time and greater swimming time [19]. On the other hand, in one study SCN lesions did not offer protection against ‘depression-like behaviour' in the social defeat paradigm [20]. Thus, the SCN may be important for mediating the effects of antidepressants and possibly in depression/depressogenic effects.

3. Is manipulation of the circadian rhythm antidepressant?

Putative circadian abnormalities in depression would suggest that antidepressant effects might be achievable by re-synchronisation of rhythm. It is well recognised that melatonin and light are capable of shifting circadian rhythms dependent on the time of application [21]. As noted previously, agomelatine can also shift rhythms. In the case of agomelatine, shift of rhythm alone is probably not sufficient (but may be necessary) for its antidepressant effects. Thus, in pre-clinical studies both the MT1 and MT2 agonist properties of the compound coupled with its 5-HT2C antagonist actions have been shown to be essential for antidepressant-like activity. In most pre-clinical models, agomelatine has been shown to be superior to melatonin on the behavioural end points. Furthermore, in some studies agomelatine was superior to a selective 5-HT2C antagonist, S-32006. For example, in the olfactory bulbectomy test neither melatonin nor S-32006 was able to fully reverse hyperactivity in the open field. On the other hand, agomelatine showed a dose-dependent reversal of the behaviour [22].

Putative biomarkers of antidepressant activity in rodent models are also dependent on both melatonergic and serotonergic activity. Increases in hippocampal BDNF (brain-derived neurotrophic factor) production, following chronic antidepressant administration, is common to most clinically effective agents. Increased BDNF response has been observed following agomelatine administration but does not occur with either melatonin or the selective 5-HT2C antagonist S-32006 [23]. This further supports the role of both actions as mediating the antidepressant effect.

4. Expert opinion

Agomelatine may yet prove to be the prototype of a new generation of antidepressant agents focussed around actions on circadian rhythm normalisation. However an action on rhythm alone, while necessary for the activity of agomelatine is not sufficient for the compound to alleviate the symptoms of depression. The addition of the pharmacological property of 5-HT2C antagonism, with its consequential action to stimulate noradrenergic and dopaminergic function in pre-frontal cortex, appears to be, at least from pre-clinical studies, necessary. This is reinforced by the therapeutic utility of medications acting as melatonin agonists. Neither of the two licenced melatonin agonists, ramelton or tasimelton, appear to have antidepressant actions [24]. Similarly, melatonin itself has little or no antidepressant effects. This lack of antidepressant effect is despite the fact that all three compounds have demonstrable effects on circadian rhythms. Bright light therapy, which also has potent circadian phase shifting effects, has been demonstrated to have antidepressant effects but principally in the ill-defined condition of seasonal affective disorder. In major depression, light therapy may be useful as an adjunct to conventional antidepressant agents, but has rarely been investigated as therapy alone. The more recent discovery of selective agonists at MT2 receptors has so far, in pre-clinical evaluations at least, demonstrated only anxiolytic actions [24]. Along with selective MT1 agonists, further pharmacological evaluation in relevant pre-clinical models is awaited in order to assess their antidepressant potential. Final proof of activity requires clinical trials.

Thus, while the direct manipulation of circadian rhythms would prima facie suggest a therapeutic antidepressant potential, this does not appear to be the case with the molecules available so far. Additional pharmacological actions would seem to be necessary if antidepressant activity is to be manifest. Just what additional properties are necessary is a moot point, but clearly some action on monoamine mechanisms might be useful, at least based on the experience with agomelatine. There is, however, ample evidence from the pre-clinical literature that other actions may also be profitably exploited. For example, the hypothalamic–pituitary–adrenal (HPA) axis has long been mooted as a potential target of antidepressants, while non-monoamine mechanisms such as glutamate and GABA have recently been a focus of attention. While such molecular combinations must remain speculative for the present, it is difficult to go beyond the conclusion that molecules combining multiple activities are desirable for the alleviation of the multifaceted symptoms of depressive disorders.

Declaration of interest

TR Norman has received honoraria for talks and travel from Servier, Australia Pty Ltd. He has been in receipt of research funds from Institut de Recherches Internationales Servier (IRIS) for work on the pre-clinical activity of agomelatine. He has also received funding for similar activities from Eli Lilly, Bristol-Myers Squibb, Lundbeck, Wyeth, Organon and GlaxoSmithKline.