CCNP Award Paper: Unveiling the role of melatonin MT₂ receptors in sleep, anxiety and other neuropsychiatric diseases: a novel target in psychopharmacology

Who’s who?

Two subtypes of mammalian MLT receptors have been cloned and characterized, the MT₁ (Mel1a) and the MT₂ (Mel1b) receptor subtypes. Both subtypes are members of the 7-transmembrane GPCR family (for detailed information on the molecular properties of MLT receptors please refer to the review by Dubocovich and colleagues⁵). By using recombinant MLT receptors it has been shown that the MT₁ receptor is coupled to different G proteins that mediate adenylyl cyclase inhibition by a pertussis toxin-sensitive G protein and phospholipase C β activation.²² The MT₂ receptor is also coupled to inhibition of adenylyl cyclase, and it also inhibits the soluble guanylyl cyclase pathway.²³ Interestingly, the human MT₂ receptor has a lower affinity for 125 I-Mel than the human MT₁ receptor.²⁴ Another interesting difference between MT₁ and MT₂ receptors is that MT₂, unlike MT₁, desensitizes after exposure to the full agonist MLT,²⁵ likely by an internalization mechanism, but it is not clear whether this occurs after short- (minutes) or long-term (5–8 hours) exposure to MLT in vivo.²⁶ Based on these considerations, it appears reasonable that the use of an MT₂ receptor partial agonist might be the best approach to activate a receptor that desensitizes after exposure to its full agonist. Now that selective MT₂ receptor partial agonists are available, this hypothesis warrants further investigation.

Where are they?

The distribution of MLT receptors within the central nervous system has not been completely elucidated, primarily owing to the lack of selective antibodies for MLT receptors. The study of MLT receptor localization has mostly been performed using receptor autoradiography with the nonselective 2-[¹²⁵I]iodomelatonin, or through real-time quantitative reverse transcription–polymerase chain reaction (RT-PCR), marking the MLT receptor mRNA. Messenger RNA expression of MT₁ and MT₂ receptors has been observed in the retina, SCN, thalamus, hippocampus vestibular nuclei and cerebral and cerebellar cortex.^27–30 At the level of the hippocampus, MT₂ receptors were detectable in particular on CA₃ and CA₄ pyramidal neurons, which receive glutamatergic excitatory inputs from the entorhinal cortex, whereas MT₁ receptors were predominantly expressed in CA₁.³¹ Recently, using polyclonal antibodies³² we have pointed out the presence of MT₂ receptors in the reticular thalamus, substantia nigra (pars reticulata), supraoptic nucleus, red nucleus and CA₂ and CA₃ areas of the hippocampus.³³

Using in situ hybridization histochemistry with selective and specific digoxigenin-labelled antisense and sense oligonucleotide probes, a robust hybridization of MT₁ and MT₂ receptor mRNA has been observed in the SCN.³⁴

In humans, MT₂ receptors have been recently localized at the level of the SCN, the supraoptic nucleus and the para-ventricular nucleus using polyclonal specific antibodies. Importantly, the receptors were confined to neurons and nerve fibres, but not to glial cells.³⁵

What are their roles?

The functional characterization of MLT receptors is still an active matter of research. It is known that MT₁ and MT₂ receptors could have either opposing or complementary functions. For example, in hippocampal slices, MT₁ and MT₂ receptor activation appears to differentially modulate γ-aminobutyric acid (GABA)_A receptor function, suggesting that MLT, through activation of different receptor subtypes, may exert opposite effects in the same brain area.³⁶ Similarly, MT₁ and MT₂ receptors have also been shown to act in an opposite manner on the vascular system, producing vasoconstriction or vasodilatation, respectively.¹⁸

Controversial results using in vitro and in vivo techniques have been published concerning the role of MT₁ and MT₂ receptors in circadian regulation and sleep. As mentioned previously, the mammalian SCN is known to be rich in both MT₁ and MT₂ receptors.³⁷ The SCN is the “master clock” that controls behavioural, metabolic and physiologic rhythms,³⁸ including the light-dependent synthesis and release of MLT from the pineal gland.³⁹

In vitro experiments using ligands toward MLT receptors and SCN slices from wild-type and MLT receptor knockout mice showed that activation of MT₁ receptors inhibited the neuronal firing rate of the SCN, whereas activation of MT₂ receptors phase-shifted circadian rhythms of neuronal firing rates.^{26,34,37,40,41} But, in Siberian hamsters in which the MT₁ receptor was functional, but the MT₂ receptor was not, MLT elicited a clear phase shift of the circadian rhythm of SCN electrical activity.⁴² In vivo experiments in MLT receptor knockout mice further complicated the scenario. Dubocovich and Markowska,²⁶ injected MLT 2 hours before the onset of activity in mice kept under constant darkness and found a phase shift in the activity onset in wild-type but not in MT₁ receptor knockout (MT₁^−/−) mice. Moreover, MLT accelerated the entrainment to a new light–dark cycle of wild-type, but not of MT₁^−/− mice. For these reasons, current knowledge does not allow us to rule out a possible role of MT₁ receptors in the phase shift of SCN neuronal activity due to MLT.^37,43

Melatonin and sleep

In rats, contradictory effects of MLT in sleep have been reported. Holmes and Sugden⁶⁵ reported that MLT reduces time to sleep onset and increases both SWS and PS. Others suggest that MLT is involved only in the control of PS regulation since the lesion of the pineal gland or inhibition of MLT synthesis using a β-adrenergic antagonist (propranolol) decreased PS during light and dark periods.^66,67 Melatonin (2.5 mg/kg and 5 mg/kg) increased the number of sleep cycles and the total duration of PS,⁶⁸ and at the dose of 10 mg/kg, MLT decreased the onset to the first episode of SWS and increased the duration of SWS and PS. These latter effects were blocked by the GABA_A antagonists flumazenil and picrotoxin.⁶⁹ In monkeys, MLT (0.3, 1 and 3 mg/kg) does not affect SWS, REMS and light sleep, but does shorten the latency to the first episode of sleep. However, this reduction in the latency to sleep occurred only at the lowest dose.⁷⁰ In a study using rats,⁷¹ MLT affected neither sleep nor circadian rest activity rhythms. In cats, exogenous MLT (0.01–1 mg/kg) significantly increased SWS, but the effect was weak and lasted for only 2 hours.⁷²

There is a large body of research demonstrating hypnotic effects of MLT in clinical studies,^73–77 although some studies failed to find significant effects,^78,79 and therefore its clinical efficacy is still unclear.⁸⁰ These discrepant results may originate from the pharmacological constraints of generic MLT, limiting its clinical use: short half-life (< 1 hr), high first-pass metabolism, binding of multiple receptors and effects dependent on time of the day or phase of circadian rhythm.⁸¹ For example, the hypnotic effect of MLT in humans varies depending on the time of the administration, consistent with the circadian phase resetting properties of MLT.⁸² Although several studies suggest that MLT treatment may be appropriate for sleep disorders, it is not established whether MLT acts directly on sleep regulation or on circadian rhythms associated with sleep. However, the sleep promoting effect of the MT₁/MT₂ receptor agonist ramelteon (TAK-375) strongly suggests that MLT compounds facilitate the induction of sleep rather than modify all sleep architecture or circadian rhythms. Ramelteon has been approved by the U.S. Food and Drug Administration (July 2005) for the treatment of insomnia characterized by difficulty with sleep onset.⁸³ Given that ramelteon is an MLT receptor agonist with high affinity for both MT₁ and MT₂ receptors, it is difficult to dissect the receptor subtype(s) involved in its sleep-promoting properties.

MT₂ receptor ligands (UCM765 and IIK7) and sleep

The pharmacological studies using selective MT₂ receptor ligands have allowed researchers to better identify the role of the MT₂ receptor in sleep function. Since MLT synthesis, and thus its circulating levels as well as the expression of its receptors, follow circadian daily variations, sleep experiments with UCM765 were performed across the entire 24-hour sleep–wake cycle to examine its effects across the light–dark period.³³ Given the short half-life of the compound (T_1/2 = 44 min) due to an extensive first-pass metabolism,^33,84 UCM765 was injected every 4 hours to keep the concentration of the drug within the steady state range. A subcutaneous injection of UCM765 at a dose of 40 mg/kg significantly reduced the latency to NREMS by 59% while increasing the amount of NREMS during the 24-hour period by 38%. The increase in NREMS was mainly evident during the light/inactive phase of the 24-hour light–dark cycle. The amount and latency of REMS were not altered by UCM765. The amount of wakefulness was instead significantly decreased (−22%) due to the increase in NREMS paralleled by no changes in REMS (Fig. 2).

Similar effects were replicated after intraperitoneal injection of IIK7 (10 mg/kg):⁴⁹ the injection led to a 38% decrease in the latency to NREMS and a 250% increase in the amount of NREMS during the first hour after injection. Like UCM765, no effects on REMS amount and latency were reported with IIK7.

Remarkably, the effects of UCM765 on NREMS as well as on NREMS latency were prevented by the MT₂ antagonist 4P-PDOT (10mg/kg), suggesting an MT₂ receptor–dependent mechanism. Administration of 4P-PDOT (10 mg/kg) alone did not significantly modify sleep parameters;³³ however, we cannot rule out that higher or lower doses of the MT₂ antagonist may influence sleep parameters.

Until now, it was hypothesized that MT₁ receptors control sleep, whereas MT₂ receptors are involved in the time shift mainly by controlling the neural activity of the SCN.⁴³ Contrary to this common belief, pharmacological studies have pointed out that MT₂ receptors modulate sleep, particularly NREMS.

Possible changes in sleep spindles and power spectra were also evaluated after the injection of UCM765 and IIK7. According to Steriade,⁸⁵ sleep spindles are considered the epitome of EEG synchronization at sleep onset (NREMS stage 2) and are defined by the association of 2 distinct rhythms: the waxing and waning spindle waves at 7–14 Hz with sequences lasting for 1–2 seconds, and the periodic recurrence of spindle sequences with a slow rhythm of 0.1–0.2 Hz.

Administration of UCM765 significantly increased the number of sleep spindles per minute of NREMS (+16%),³³ whereas no information was reported for IIK7. When we analyzed the effects of UCM765 on the power spectra across the 24-hour period, we found that, unlike the benzodiazepine diazepam, the EEG power of the Δ band of NREMS was significantly enhanced (+14%), whereas REMS θ power was not altered.³³ Neither Δ nor θ powers were affected after IIK7 injection.

Of note, the same 24-hour EEG/EMG experiment carried out with the nonselective MT₁/MT₂ receptor ligand UCM793 (40 mg/kg), which belongs to the same class of N-(substituted-anilinoethyl)amides, and MLT (40 mg/kg) failed to reproduce similar results. Administration of UCM793 slightly decreased the onset to the first NREMS episode (−10.7%) and did not increase the total amount of NREMS.³³ On the other hand, MLT produced a slight but nonsignificant increase of NREMS duration during the inactive phase (unpublished data, 2012). These data confirmed that the MT₂ receptor ligands possess more potent hypnotic properties than the nonselective MT₁-MT₂ ligands.

In our labaratory, we also examined whether MT₂ receptors were present in brain areas or nuclei, other than the SCN, that are known to be involved in the process of sleep. Immunocytochemical, pharmacological and in vivo electrophysiology studies answered these questions. Using specific MT₂ receptor antibodies, we found that basically all neurons in the reticular thalamus bear MT₂ receptors³³ and, as previously reported, most of these neurons (if not all) are GABAergic.⁸⁶ The reticular thalamus is an important nuclei in the neural network controlling sleep; it is implicated in the generation of Δ waves during deep SWS⁸⁷ and also of spindle waves during light SWS, thus being called the spindle “pacemaker.”⁸⁸ During NREMS, the neurons of the reticular thalamus change their activity, from single spikes firing during waking and REMS to rhythmic spike bursts and increased firing mode, which is transmitted to thalamic relay nuclei and modulated by corticothalamic inputs, resulting in a widespread synchronization across neuronal assemblies.^89–91 For detailed information on how the reticular thalamus controls the cortical oscillations and on its role in the thalamocortical networks, see the reviews by Steriade⁹² and by Linás and Steriade.⁹³

Since MT₂ receptors are highly expressed in the reticular thalamic nucleus, and given that UCM765 acts to promote NREMS, we decided to test whether UCM765 was able to enhance the activity of reticular thalamic neurons by using in vivo electrophysiology (Fig. 3, top). Intravenous administration of UCM765 (20 mg/kg) significantly increased the firing (+91%) and burst activities of reticular thalamic neurons. In addition, to demonstrate that the increased activity of the reticular thalamus relies on MT₂ receptors present within the nucleus, we microinfused the selective MT₂ receptor antagonist 4P-PDOT⁹⁴ directly into the nucleus before intravenous injection of UCM765. When 4P-PDOT was injected before UCM765, there was no change in the firing pattern of reticular thalamic neurons. These findings allowed us to conclude that MT₂ receptor agonists promote sleep through the activation of reticular thalamic neurons.

We then examined whether the selective activation of MT₂ receptors into the reticular thalamus was sufficient to promote sleep, particularly NREMS. We bilaterally implanted 2 cannula into the reticular thalamus. Microinfusion of UCM765 into the reticular thalamus produced results that mirrored the subcutaneous administration of the drug: decreased latency to NREMS, increased amount of NREMS and no effects on REMS. Remarkably, the same dose of UCM65 microinfused into the substantia nigra, a brain area not involved in NREMS, but with MT₂ receptors, did not alter the vigilance states. A summary of the effects of UCM765 on vigilance states and reticular thalamic activity is reported in Figure 3.

MT₂ knockout mice and sleep

To support the proof of the concept that MT₂ receptors may represent a novel target for treating sleep abnormalities, the sleep–wake cycle in MT₂^−/− mice has been examined. Knockout mice represent an excellent tool to study the involvement of different neurotransmitters, receptors and enzymes in sleep regulation, as demonstrated for orexins;⁹⁵ histamine deficient L-histidine decarboxylase;⁹⁶ corticotropin-releasing hormone receptor type 1;⁹⁷ dopamine (DA) D₂ receptor;⁹⁸ and 5-HT_1A, 5-HT_1B, 5-HT_2A and 5-HT_2C receptors.⁹⁹ Compared with wild-type controls, MT₂^−/− mice⁴⁰ displayed a significant decrease in the amount of NREMS during the 24-hour period (−17%), with the decrease mainly due to changes occurring during the light/inactive phase of the 12-hour light–dark cycle.³³ No variation in the amount of REMS was found. The decline in the amount of NREMS is likely attributed to the decrease of NREMS Δ and ∑ powers found after analyzing the 24-hour power spectra. Even though some issues remain to be considered, NREMS ∑ power is considered a valid measure of sleep spindle activity.¹⁰⁰ To date, no research has reported on visually scored sleep spindles in MT₂^−/− mice, but the decrease in NREMS ∑ power observed in MT₂^−/− mice suggests that the number of sleep spindles is likely reduced.

In our labaratory, we have also tested UCM765 in MT₂^−/− mice, and we found that it was ineffective at modifying vigilance states. The lack of effect in MT₂^−/− mice concurred with the evidence indicating that the NREMS promoting effect of UCM765 is selectively mediated by MT₂ receptors. In keeping with the findings in rats, in wild-type mice UCM765 significantly increased NREMS duration (wild-type plus vehicle = 64.4 ± 4.0 min, n = 6; wild-type plus UCM765 = 89.3 ± 5.5 min, n = 6, p = 0.004) and decreased wakefulness (wild-type plus vehicle = 110.3 ± 5.6 min, n = 6; wild-type plus UCM765 = 83.82 ± 5.6 min, n = 6, p = 0.005), without affecting REMS duration.

To verify whether the reduction in NREMS observed in MT₂^−/− mice was due to the lack of MT₂ receptors rather than to impairments in other sleep-related neurotransmitter systems, MT₂^−/− mice were treated with the GABA_A agonist diazepam. As observed in wild-type mice, diazepam decreased the latency to the first episode of NREMS (−64.4%), enhanced NREMS duration (+49.1%), and tended to decrease REMS duration (−20.6%) in MT₂^−/− mice, suggesting that the GABAergic system regulating sleep was intact in these mice.

We have also performed EEG/EMG studies in MT₁^−/− mice and found that these mice have a selective perturbation of REMS. In particular, during the light phase, REMS duration was decreased compared with wild-type mice (−46%) and NREMS duration was increased compared with MT₂^−/− mice (+33%).¹⁰¹ This particular phenotype induced by the lack of the MT₁ receptor could explain the reason why a selective MT₂ receptor ligand is a more effective hypnotic agent than a nonselective MLT receptor ligand. We have indeed hypothesized that stimulation of the MT₂ receptor only increases NREMS, whereas MT₁ receptor activation increases REMS, while slightly decreasing NREMS.

General consideration

Pharmacological and genetic studies targeting MT₂ receptors have revealed a new important function of this MLT receptor subtype: the modulation of NREMS. Pharmacological activation of MT₂ receptors using selective partial and full agonists promotes NREMS without affecting REMS. On the other hand, genetic inactivation of MT₂ receptors leads to a decrease in the amount of NREMS without affecting REMS (Table 1). Our results³³ demonstrated that the local pharmacological activation of MT₂ receptors within the reticular thalamus is sufficient to induce and promote NREMS, whereas MT₂ receptor activation in brain nuclei not intimately involved in the neural circuit of sleep, such as the substantia nigra, produces no effects on NREMS. These experiments thus support the concept that MT₂ receptors may be considered a novel target for hypnotic agents.

Since we found MT₂ receptors at the level of the hippocampus³³ and since neocortical and hippocampal activity is coupled on both long and short time scales during SWS,¹⁰² we cannot rule out the possibility that the involvement of these areas may also participate in the promotion of SWS.

Further research needs to be undertaken to determine whether these MT₂ receptor agonists affect the phase shift. Our experiments seem to suggest that at the hypnotic doses, MT₂ receptor agonists do not affect time shifting. However, we cannot discount the possibility that the activation of MT₂ receptors in the SCN, which is involved in time shifting of tSCN firing activity,⁴⁰ may also contribute indirectly to the NREMS process. The development of selective MT₁ or MT₂ receptor agonists will offer the opportunity to better understand and distinguish the role of each receptor in the phase shifting from the sleep promoting effects of MLT. This will ultimately allow for the development of selective pharmacotherapeutic treatment strategies in patients with sleep and time shifting disorders. To date, structure–activity relationships conferring selectivity for the MT₁ receptor subtype are still a work in progress;⁸ therefore, no selective MT₁ receptor agonists have been tested in sleep and time shifting studies.

Insomnia is a common public health problem, with a prevalence ranging from 11% to 16%.^103,104 Primary insomnia is characterized by a difficulty to initiate or maintain sleep for at least 1 month. Currently available drugs to treat insomnia include benzodiazepine receptor agonists (benzodiazepines and nonbenzodiazepines), antidepressants, antipsychotics, antihistamines and ramelteon, a nonselective MT₁/MT₂ receptor agonist. Unfortunately, these hypnotics cannot reproduce a physiologic sleep,¹⁰⁵ and their use, especially chronic use, is associated with adverse effects, such as tolerance to the drug, rebound insomnia, physical withdrawal symptoms when discontinued, sedation, anorexia, anxiety, agitation, tremors, convulsion, and physical and psychological dependence. Therefore, the sleep disorders market has substantial unmet needs,¹⁰⁶ and based on the preclinical evidence, MT₂ receptors may be a promising novel target deserving consideration. Our experiments indicate that MT₂ receptor agonists may have a more favourable pharmacological profile than classical benzodiazepines, such as diazepam. Indeed, in comparison with diazepam, UCM765 showed similar NREMS promoting properties, but unlike the benzodiazepine, it did not produce sedation in the rotorod test (unpublished data, 2012) and open field test (OFT) at the hypnotic dose.¹⁰⁷ Moreover, the typical reduction of NREMS Δ and θ powers and REMS θ power¹⁰⁸ observed with benzodiazepines was not observed with UCM765, which instead increases the Δ power during NREMS and REMS. In addition, UCM765, unlike benzodiazepines, does not increase the peak of β frequency.¹⁰⁹ These pharmacological studies have proven that targeting MT₂ receptors may become a novel pharmacological strategy in sleep research and therapy, representing a valid alternative to benzodiazepines and benzodiazepine derivatives.

Melatonin and anxiety

Anxiety disorder (or generalized anxiety, as per DSM-IV¹¹⁰) is characterized by excessive, uncontrollable and often irrational worry about everyday things that is disproportionate to the actual source of worry lasting for more than 6 months. The psychological symptoms are often associated with physical symptoms, including fatigue, fidgeting, headaches, nausea, numbness and sweating.

Melatonin and the nonselective MT₁/MT₂ receptor agonist agomelatine have displayed anxiolytic-like properties in classical animal paradigms of anxiety.^111–113 For instance, MLT increased the activity within the central area of the OFT¹¹² and the time spent in the open arms of the elevated plus maze test (EPMT).^111,113 Agomelatine increased the number of punished responses in the Vogel test, increased open arm exploration in the EPMT and decreased vocalization time in the conditioned footshock-induced ultrasonic vocalization test.¹¹³

In humans, MLT has been used to reduce or eliminate benzodiazepine administration in elderly patients.¹¹⁴ Still the receptor mediating the anxiety-like effect of MLT was unknown. The recent development of selective MT₂ receptor agonists and the use of selective antagonists have shed light on the distinct role of each MLT receptor subtype in the regulation of anxiety.

The MT₂ receptor ligand UCM765 and anxiety

Using UCM765, we recently made the first attempt, to our knowledge, to investigate the possible role of MT₂ receptors in anxiety.¹⁰⁷ We used 3 well-established animal paradigms to study anxiety: the EPMT, OFT and the novelty-suppressed feeding test (NSFT). The first experiment in the EPMT and NSFT was designed to obtain a dose-response curve using lower doses of UCM765 than those active in promoting NREMS. Administration of UCM765 at a dose of 10 mg/kg, displayed anxiolytic properties by increasing the time spent in the open arms (+208%) of the EPMT and by decreasing the latency to eat in a new environment (−46%) without affecting the latency to eat and the amount of food intake per animal body weight in the home cage. Melatonin was also tested, and it produced results similar to UCM765. By using a nonselective MT₁/MT₂ receptor antagonist (luzindole) and a selective MT₂ receptor antagonist (4P-PDOT), we investigated whether both MT₁ and MT₂ receptors were implicated in the anxiolytic properties of MLT. Administration of UCM765 or MLT did not show anxiolytic effects in the EPMT and in the NSFT in animals pretreated with luzindole (10 mg/kg) or 4P-PDOT (10 mg/kg). These results collectively demonstrate that the anxiolytic effects of UCM765 and MLT are both pharmacologically mediated by MT₂ receptors. Notably, 4P-PDOT or luzindole alone did not affect anxiety levels at the doses used to block the effects of UCM765 and MLT. Accordingly, Nava and Carta¹¹⁵ found that luzindole administered in higher doses than those in our experiments (30 and 60 mg/kg v. 10 mg/kg) did not modify anxiety levels. Administration of UCM765 and MLT was also tested in the OFT, but they did not alter measures of anxiety behaviour in this paradigm. However, unlike diazepam, UCM765 and MLT did not reduce locomotor activity¹⁰⁷ or increase the number of falls in the rotoroad tests (unpublished data, 2012), 2 measures related to the sedative properties of a putative psychotropic compound.

These findings are the first to examine the role of MLT receptors in anxiety, and they indicate that MT₂ receptors mediate the anxiolytic effects of the neurohormone.

MT₂ knockout mice and anxiety

To support the pharmacological studies showing that MT₂ receptors are involved in anxiety, we investigated whether MT₂^−/− mice display altered levels of anxiety in the EPMT, NSFT and OFT (unpublished data, 2012). In the EPMT, the number of entries and the time spent in the open arm was not modified, but interestingly MT₂^−/− mice spent more time in the centre than wild-type controls, suggesting a cognitive impairment at the level of goal-oriented behaviour. In the EPMT, the time spent by animals in the open arm is the most important variable to consider when establishing whether or not a treatment affects anxiety. In MT₂^−/− mice, open arm entries and time spent in the open arm were not affected, but they spent more time than controls in the centre of the maze. On the other hand, the time spent in the central platform of the EPMT has been linked to the decision-making process rather than to anxiety itself.¹¹⁶ In agreement with the notion of possible decision-making impairments in MT₂^−/− mice, Larson and colleagues¹¹⁷ found that these animals, when tested in the EPMT on 2 consecutive days, failed to shorten the transfer latencies to enter a closed arm on the second day, which is an index of learning and/or memory impairments.

In the NSFT, the latency to eat in the new environment was significantly longer in MT₂^−/− mice than wild-type mice (+80%), whereas the latency to eat and the amount of food intake per animal body weight in the home cage were unchanged. In the OFT, the number of entries and the time spent in the centre of the arena were significantly higher in MT₂^−/− mice than wild-type mice, whereas locomotion was unchanged. The results gleaned from these experiments suggest that the genetic manipulation of MT₂ receptors interferes in a complex manner with some paradigms of anxiety behaviour. Anxiety disorders, by definition, are complex and heterogeneous;¹¹⁸ nonetheless, different animal models could be tapping different aspects of anxiety,¹¹⁹ and studies of different mouse strains and different anxiolytic treatments have previously yielded contrasting results, depending on the behavioural test (Table 2).^116,119

General consideration

Pharmacological and genetic studies targeting MT₂ receptors have shed new light into the neurobiological mechanism through which the MLT system modulates anxiety. Although further studies are necessary, the research undertaken until now shows that MT₂ receptors modulate anxiety levels and consequently this receptor may become a novel target for the treatment of anxiety disorders. Common anxiolytic drugs, such as benzodiazepines, targeting the GABAergic system are also used as hypnotic agents at higher dosages. Targeting MT₂ receptors seems to provide similar pharmacological outcomes. Administration of UCM765 at a dose of 40 mg/kg promotes sleep,³³ whereas a lower dose (10 mg/kg) reduces anxiety.¹⁰⁷ These similar pharmacological effects occurring after the activation of the GABAergic and the MLT systems are likely due to the mutual interaction between these 2 systems. We found that almost all of the GABAergic reticular thalamic neurons express MT₂ receptors.³³ Moreover, MLT administration increased GABA levels in several brain regions, such as the hypothalamus, cerebellum and cerebral cortex,¹²⁰ and the administration of flumazenil, a benzodiazepine receptor antagonist, blocked the anxiolytic-like effect of MLT in the EPMT.¹¹¹ However, given that the MT₂ receptor–mediated anxiolytic properties are also associated with low sedation and a lack of abuse potential, as demonstrated by the use of nonselective melatoninergic compounds,¹²¹ targeting this MLT receptor subtype may offer better anxiolytic pharmacological strategies over the GABAergic compounds, especially when treating mild levels of anxiety.¹⁰⁷

Further studies will need to investigate the neurobiological mechanism through which MT₂ receptors modulate anxiety, the effects of MT₂ receptor full agonists and the pharmacological and genetic blockade of MT₂ receptors. Further study is also needed to determine whether MT₂ receptors are localized only on GABAergic neurons or also on other systems, such as 5-HT and glutamate, known to be implicated in anxiety.