Profound Differences in Metabolism and Their Consequences
The metabolism of melatonin is highly complex [4, 7, 10, 32, 65], although this fact is not always perceived by researchers. In this regard, one has to distinguish between circulating and tissue melatonin. The hormone, as released from the pineal gland to the circulation, or entering the blood via oral or enteral administration, is mainly metabolized by hepatic P450 monooxygenases to 6-hydroxymelatonin, which is conjugated and preferentially excreted as 6-sulfatoxymelatonin. When giving melatonin in pharmacological doses, either orally or enterally, it is important to be aware that much of the indoleamine is also loaded to the gastrointestinal tract, which can act at the same time as a source and sink of melatonin [10], and where the indoleamine can undergo enterohepatic cycling without being metabolized [10, 66]. Especially pharmacological concentrations of melatonin seem to enter many other tissues, by virtue of its amphiphilicity, which allows this indoleamine to cross any membrane. In extrahepatic tissues, a substantial fraction of melatonin can be metabolized via other pathways, and since tissue melatonin exceeds the circulating amounts by orders of magnitude [9, 10], the relevance of the alternate pathways should not be underrated. A potentially important route is that of pyrrole ring cleavage, which leads to 5-methoxylated kynuramines (Figure 1) [10, 65] and their secondary products [4, 67, 68]. The most actual estimations assume that about one third of total melatonin may be metabolized via the kynuric pathway [69]. This conclusion is insofar important, as the kynuric metabolites are also bioactive compounds [4]. Additionally, several bioactive indolic metabolites are formed in the central nervous system [32], but, in quantitative terms, the relevance can be hardly judged on the basis of the available data. Moreover, 2- and 3-hydroxylated indolic metabolites are formed under oxidative conditions [70], such as cyclic 3-hydroxymelatonin and 2-hydroxymelatonin, which is in equilibrium with its indolinone tautomer (Figure 2), compounds which have been frequently found in cells or as excretion products after exposure to oxidants or to radiation.
A look at the structural formulas of the melatoninergic agonists discussed here shows that none of them can be metabolized in the same way as the natural hormone, melatonin. In the most closely related compound, P-methyl-6-chloromelatonin (Figure 2), hydroxylation at C-atom 6 is impossible because of the chlorine atom at this position. Therefore, an analog of the major melatonin metabolite, 6-hydroxymelatonin, cannot be formed. This difference is likely to explain the longer half-life of the synthetic analog in the circulation. A full spectrum of metabolites from the chlorinated agonist has not been published. It seems possible that kynuramine analogs are generated in the metabolism of P-methyl-6-chloromelatonin, because the related 6-chloromelatonin is as easily oxidized via pyrrole ring cleavange as melatonin itself, at least in radical generating systems [71]. Whether or not P-methyl-6-chloromelatonin is equally well accepted as a substrate by melatonin-cleaving enzymes, such as indoleamine 2,3-dioxygenase and myeloperoxidase [4], is still inknown.
The naphthalenic analog, agomelatine (Figure 2), exhibits, for fundamental reasons, even more profound deviations in its metabolism. Formation of kynuric compounds and derivatives is principally impossible, since a pyrrole ring is absent in the molecule. Two major metabolites have been described, #-[2-(7-hydroxynaphth-1-yl)ethyl] acetamide (= S21517) and #-[2-(3-hydroxy-7-methoxynaphth-1-yl)ethyl] acetamide (= S21540) (Figure 2) [53]. Dealkylation at the methoxy group, as occurring in S21517 formation, is similarly possible with melatonin [6], by the action of respective cytochrome P450 subforms, such as CYP2C19 and CYP1A2. The resulting 7-hydroxylated compound S21517 is, in structural terms, more serotonin-like and, in fact, exhibits affinity to 5-HT2C receptors. Whether additional serotoninergic effects via different receptor subtypes may exist should be a matter of further investigation. With regard to its molecular structure, S21517 should not substantially contribute to melatonergic effects. While agomelatine is an efficient inhibitor of 5-HT2C receptors, the 3-hydroxylated metabolite S21540 shows considerably weaker binding to this subtype [53]. It should also be noted that the hydroxylation at this position is profoundly different from corresponding reactions in the metabolism of melatonin, although hepatic P450 monooxygenases are responsible for these reactions with melatonin and agomelatine as well. This deviation is an unavoidable consequence of the replacement of the indolic moiety by a naphthalene group.
The metabolism of ramelteon (Figure 3) differs even more from that of melatonin and the other melatoninergic agonists mentioned before. The various changes in the molecule, compared to melatonin, lead to specific consequences for reactions that are allowed or not allowed. The furan ring, which replaces the 5-methoxy group, largely prevents dealkylation and is maintained in metabolites M-II, M-III and M-IV. However, this ring can be oxidized, as is the case in metabolite M-I (Figure 3). Although this creates a structure partially reminiscent of N-acetylserotonin, a serotoninergic activity is rather unlikely, because the newly formed carboxyl group changes the biochemical and physical properties of the molecule considerably. The replacement of the N-acetyl residue by an N-propionyl group makes deacylation less likely, as it is occurring in the case of melatonin by actions of a specific melatonin deacetylase or less specific aryl acylamidases, including the aryl acylamidase activity of choline esterase [4, 32]. Although the unspecific enzymes can also remove other acyl residues, such metabolites from ramelteon have not been described, but one should take the possibility into account that resulting primary amines may be easily oxidized and give rise to further products. The presence of an #-propionyl group is reason for another type of oxidation reaction that is impossible in the N-acetylated amines, namely, a hydroxylation at C2 of the propionamide, which is found in metabolites M-II and M-IV. This hydroxyl group seems to be decisive for a change in live-time and, thus, blood plasma concentrations. The much longer half-life of M-II and the concentrations exceeding that of ramelteon by more than one order of magnitude, as mentioned in the previous section, are obviouly caused by this molecular difference, since the two compounds are otherwise identical. Another significant change in the chemical design of ramelteon concerns the replacement of the pyrrole ring-containing indole by an indene. As in the naphthalenic agomelatine, the 5-atom ring devoid of a nitrogen fully prevents ring opening, so that the kynuric metabolism is excluded. However, this also leads to a change in oxidation of the ring system, taking place at a position comparable to that of the N-atom in the indole, which would not allow such a reaction. Obviously, P450 subforms which are capable of hydroxylating melatonin at C-atom 6 do not accept ramelteon in the same way, and the isoenzymes responsible for the indene oxidation, as found in metabolites M-III and M-IV, may be worth of further investigation. With regard to the formation of M-IV, one should also note that two different sequences of oxidation steps are possible (Figure 3).
Figure 3. The metabolism of ramelteon.
The different metabolism of ramelteon does not only exclude pathways known from melatonin. Apart from the kynuric route, hydroxylations corresponding to C-atoms 6, 2, and 3 of melatonin are not observed. However, the design of the molecule creates a situation that is highly unusual for drug metabolites: aliphatic hydroxylation to M-II generates a product which exceeds by far the blood levels of the parent compound and additionally exhibits melatoninergic properties. Although its affinity to MT1 and MT2 receptors is by one order of magnitude lower than that of ramelteon, this should be compensated by the higher concentrations attained. Therefore, M-II must be concluded to substantially contribute to the melatoninergic effects of ramelteon.
A further aspect of metabolism which may be relevant to all three synthetical agonists discussed here concerns possible long-term toxicity. Even in the absence of acute toxicity, chlorinated aromates, naphthalenes and other polycyclic aromates may be problematic in the long run. These are necessary general considerations, especially for a hydroxylatable naphthalene, which is anyway under suspicion of eventual carcinogenicity. In the case of ramelteon, information provided by Takeda [72] indicates a no-effect level for the induction of hepatic tumors in male mice that was only 3 times of the concentration of the metabolite M-II measured after the therapeutic dose. Since M-II is structurally very similar to the parent compound, its eventual toxicity has to be taken into consideration. Moreover, micronuclei formation was observed with ramelteon, in Chinese hamster lung cells, after metabolic activation [72]. The same information sheet mentions no mutagenicity in the Ames test, but does not refer in this case to metabolic activation, which should be routinely done with this assay, too, including respective tests for M-II. In summary, acute toxicity is presumably not a problem with the compounds discussed, except for some diseases or co-treatments with other drugs for different medicinal purposes [18]. Nevertheless, their suitability for long-term treatment would require further substantiation. With good reasons, the paucity of information available on extended treatment with ramelteon has been criticized [73]. This is even more valid for agomelatine and P-methyl-6-chloromelatonin. Therefore, these drugs should be prescribed for long-term treatment, even in trials, only with caution.
Versatility Versus Selectivity: Advantage or Disadvantage?
The remarkable pleiotropy of melatonin [4, 6, 7, 74, 75] is based on several, collectively exceptional properties: (i) wide distribution of receptors and other binding sites, (ii) multiplicity of binding sites found in different subcellular compartments, (iii) parallel signaling, (iv) cross-talking between signaling pathways, and (v) involvement of bioactive metabolites. This full spectrum of actions, which should be considered biologically meaningful for an agent with orchestrating functions, clearly exceeds in many ways the role of a sleep-promoting substance and, presumably, even that of a chronobiotic. Therefore, investigators from the sleep research field may question whether such a multiplicity of actions is really desired for treating insomnia, or may even regard this as a disadvantage. Without any doubt, a drug like ramelteon is much more specific for the membrane receptors MT1 and MT2. As outlined above, several other actions of melatonin related to QR2, to calmodulin, to direct radical reactions and activities of metabolites are excluded in the case of ramelteon. To a certain extent, especially with regard to metabolism, this seems to be valid for agomelatine as well, although the respective informations are scarce. Binding of P-methyl-6-chloromelatonin to intracellular proteins has been insufficiently studied to date – or, at least, not been published.
While, on the one hand, actions not mediated by MT1 and/or MT2 may appear superfluous at first glance for a soporific drug, one should, on the other hand, also consider sleep induction and support as being related to readjustments of the circadian system, e.g., after a jet lag, for returning to a desired phase in shift workers, or in cases of circadian dysfunctions. Under these conditions, the full spectrum of orchestrating effects as exerted by melatonin, which affects more or less the entire body, may be thought to be advantageous. Therefore, a decision on this point is not that easy. Whatsoever, the efficacy of a compound like ramelteon justifies its prescription to insomniacs, at least, for short-term treatment. In the case of agomelatine, the situation is insofar different as the main reason for its use is the combination of antidepressive and hypnotic effects, whereas the other compounds including melatonin can counteract depression only if the disorder is caused by pathologically dysphased circadian rhythms [19, 20, 56, 57]. For P-methyl-6-chloromelatonin, no real advantage was apparent in comparison to melatonin [29-31].
Selectivity for receptor types should not be mistaken as total absence of pleiotropic effects. Even if a presumably purely MT1/MT2-specific agonist like ramelteon is administered, its actions are by far not restricted to the SCN, but should be found in numerous places of the body where these receptors are also located, including vasculature, immune cells and various vegetative organs [4, 6, 14, 74, 75]. In other words, a host of additional effects has to be expected in parallel to the hypnotic action, including influences on vasomotor and immune functions. In this regard, ramelteon shares properties with melatonin, but, by virtue of its higher receptor affinities and prolonged lifetime, its actions should be expected to be stronger and longer-lasting. Whether this has to be regarded as an advantage or a disadvantage can be hardly judged at the moment. Perhaps, an extension in bioavailability of the natural compound melatonin by slow-release formulations may appear as a safer way for treatment than the use of a synthetic agonist with elevated affinity and a long-lived metabolite of uncertain properties.
Conclusion
The melatoninergic agonists discussed here differ with respect to receptor affinity, selectivity, pharmacokinetics and metabolism. The compound with highest affinity and selectivity, ramelteon, has the additional advantage of extended persistence in the blood, but in conjunction with these properties, attention seems due with regard to the metabolite M-II, which is also melatoninergic, attains much higher plasma levels, and should be more thoroughly investigated.
Agomelatine shows affinities similar to those of melatonin, is less selective, has a somewhat longer lifetime and is converted to two metabolites whose properties require further detailed studies, the dealkylated one with regard to eventual other serotoninergic actions, and both of them under toxicological aspects. An advantage of agomelatine may exist in its additional antidepressive effects, which are combined with the sleep-promoting properties.
P-Methyl-6-chloromelatonin does not seem to be superior to melatonin in terms of hypnotic efficacy.
Therefore, the bottomline is whether any of these compounds should be preferred to melatonin as a sleeping pill. In the case of agomelatine, such an advantage may concern the treatment of depression, especially when this disorder is not related to circadian dysfunctions, but the decision on a future clinical application has to depend on long-term safety. This might be critical with a naphthalenic compound and its structurally similar metabolites. In the case of ramelteon, higher affinity and extended bioavailability may be seen as the main advantages, but, on the other hand, an extended bioavailability has been achieved also in the case of melatonin, with the prolonged-release formulation of CircadinTM. So one could ask what is preferable, a chemically designed drug of longer half-life, for which, however, long-term studies are urgently required, or a natural compound which has proved remarkably well tolerable? Also in the case of melatonin, some investigators have recommended caution, and the concerns were frequently directed to doses of the hormone which were smaller than those recommended for the synthetic drugs. To many investigators, this appears highly inconsequent. However, one should also take notice of the fact that melatonin has already been given in much higher doses. For more than a year, ALS patients received either 30 or 60 mg/day orally as slow-release pills [76], or 300 mg/day enterally as suppositories [77]. The aim was, of course, delay of disease progression, but no negative experiences were made in these studies. On the contrary, melatonin was well tolerated even at the extreme doses, along with hypnotic effects, from which the patients also profited. Although these high amounts would not be recommended for a normal sleeping pill, such findings should largely dispel the concerns of some researchers about the use of melatonin.
