Melatonin and its analog 5-MCA-NAT (5-methylcarboxyamino-N-acetyl tryptamine) are active compounds reducing intraocular pressure (IOP). This action is mediated through MT2 and the putative MT3 melatonin receptor, producing a transient reduction of IOP that lasts for a few hours and has not yet been characterized. The use of melatonin and its analog are causing a decrease in chloride efflux from rabbit nonpigmented epithelial cells (NPE), possibly explaining the decrease in IOP. Melatonin and 5-MCA-NAT inhibited rabbit NPE chloride release in a concentration-dependent manner, whereas the pD2 values were between 4.5 ± 1.2 and 4.4 ± 1.0, respectively. Melatonin hypotensive action was enhanced by the presence of MT2 antagonists, such as DH97 (N-pentanoyl-2-benzyltryptamine) and 4-P-P-DOT (4-phenyl-2-propionamidotetralin) and by the nonselective melatonin receptor antagonist luzindole. Prazosin (1.5 µM) partially reverses the melatonin action by acting as a selective MT3 antagonist. However, at 15 nM it acts as an α-adrenergic receptor antagonist, enhancing the melatonin effect. Regarding the intracellular pathways triggered by melatonin receptors, neither phospholipase C/protein kinase C pathway nor the canonical reduction of intracellular cAMP was responsible for melatonin or 5-MCA-NAT actions. On the contrary, the application of these substances produced a concentration-dependent increase of cAMP, with pD2 values of 4.6 ± 0.2 and 4.9 ± 0.7 for melatonin and 5-MCA-NAT, respectively. In summary, melatonin reduces the release of chloride concomitantly to cAMP generation. The reduction of Cl− secretion accounts for a decrease in the water outflow and therefore a decrease in aqueous humor production. This could be one of the main mechanisms responsible for the reduction of IOP after application of melatonin and 5-MCA-NAT.
Melatonin is a relevant hormone controlling various physiologic actions, many of which are related to the photoperiod (Pandi-Perumal et al., 2006). It is generally accepted that blood melatonin levels increase overnight as a consequence of its production and release from the pineal gland (Caprioli and Sears, 1984; Alarma-Estrany and Pintor, 2007; Stehle et al., 2011). This gland is not exclusively responsible for melatonin production. In the orbital cavity and in the eye, for instance, some areas, such as the retina (Cardinali and Rosner, 1971a,b; Alarma-Estrany and Pintor, 2007), ciliary body (Martin et al., 1992; Alarma-Estrany and Pintor, 2007), and harderian glands (Djeridane et al., 1998; Alarma-Estrany and Pintor, 2007), have the ability to synthesize and release melatonin. As occurs in other tissues, the eye melatonin exerts many of its actions by means of membrane receptors termed melatonin receptors, divided into MT1, MT2, and the putative MT3 melatonin receptors (Alarma-Estrany and Pintor, 2007; Dubocovich et al., 2010).
One significant physiologic process in the eye undergoing circadian control is the regulation of the intraocular pressure (IOP). IOP is the result of the balance between the production of the aqueous humor by the ciliary body (Civan and Macknight, 2004; Do and Civan, 2004) and its drainage by the trabecular meshwork and uvesoscleral pathway (An and Ji, 2011; Pattabiraman et al., 2012). In this sense, the circadian fluctuation of IOP has been widely studied (Rowland et al., 1981; Liu et al., 2011) as well as its melatonin levels relationship (Samples et al., 1988). Interestingly, this circadian pattern can be modified by applying exogenously melatonin or any of its analogs (Pintor et al., 2001; Serle et al., 2004). In this sense, the topical application of melatonin on the ocular surface produces a transient reduction in IOP that is enhanced by melatonin analogs, such as 5-MCA-NAT (5-methylcarboxyamino-N-acetyl tryptamine) or IIK7 [N-butanoyl 2-(9-methoxy-6H-isoindolo[2,1-a]indol-11-yl)ethanamine]. Melatonin and these two analogs, acting through MT3 and MT2 melatonin receptors, can modify IOP by acting on the ciliary body (Pintor et al., 2001, 2003; Alarma-Estrany et al., 2007).
From a therapeutic point of view, the implications of melatonin and analogs on IOP control are relevant. IOP is elevated in primary open angle glaucoma, affecting more than 65 million patients all over the world (Quigley and Broman, 2006). In most of the cases, the treatment of the reduction of the abnormally elevated IOP is by means of adrenergic compounds, carbonic anhydrase inhibitors, prostaglandins, or parasympathomimetics (Webers et al., 2008; Lee and Goldberg, 2011; Carta et al., 2012). Because melatonin reduces IOP in experimental models, it would be of interest to see whether this effect is also feasible in humans. In this sense, some ophthalmologists have started to use melatonin to reduce IOP in patients undergoing cataract surgery, indicating the relevance of this molecule as regulator of IOP (Ismail and Mowafi, 2009), suggesting its possible use as a treatment of ocular hypertension.
The actions of melatonin in the reduction of IOP are taking place mainly on the ciliary body as noted above. Interestingly, the actions melatonin and analogs can exert on this part of the eye are not only the short-term IOP reduction (Pintor et al., 2001, 2003) but they also produce a long-term effect. This second aspect of melatonin action is due to the modification in the expression of key genes encoding for proteins relevant in the control of aqueous humor production, such as adrenergic receptors (Crooke et al., 2011) and carbonic anhydrases (Crooke et al., 2012). These proteins indirectly regulate water efflux, therefore controlling aqueous humor production and subsequently IOP.
On the contrary, little is known about the mechanisms of IOP rapid reduction by melatonin or 5-MCA-NAT. It is supposed that melatonin and analogs may also modify the aqueous humor water production, but to date there is no evidence of the possible mechanism involved.
One of the ions driving the water movement from the ciliary body to the posterior chamber to form aqueous humor is chloride (Civan and Macknight, 2004; Do and Civan, 2004). Because there may be a connection between melatonin receptor activation, chloride movement, and aqueous humor production, the present experimental work studies the ability of melatonin and 5-MCA-NAT to modify the chloride efflux in ciliary body nonpigmented epithelial cells.
Materials and Methods
Nonpigmented epithelial cells (NPE), an immortalized cell line of rabbit ciliary nonpigmented epithelium, were kindly supplied by Dr. Coca-Prados (Yale School of Medicine, New Haven, CT). Cells were grown in high glucose Dulbecco’s modified Eagle’s medium (Gibco/Invitrogen, Carlsbad, CA) containing 10% fetal bovine serum (Sigma-Aldrich, St. Louis, MO) and 0.05 mg/ml Gentamicin (Gibco/Invitrogen) at 37°C in humidified atmosphere 5% CO2–95% air.
Chloride Efflux Studies.
Chloride efflux was measured using MQAE [N-(6-methoxyquinolyl) acetoethyl ester; Invitrogen, Carlsbad, CA] as chloride indicator (Lee et al., 1984). Briefly, cells were seeded in 48-wells plates (Iwaki, Tokyo, Japan) at a density of 104 cells/well and grown to confluence. Twenty hours before the experiment, the cells were incubated in Dulbecco’s modified Eagle’s medium containing 1 mM MQAE (West and Molloy, 1996). After incubation, the cells were washed three times in chloride-containing buffer and incubated in this buffer, with or without different antagonists, for 10 minutes (at 37°C) to induce chloride channel activation. This buffer consisted of 2.4 mM K2HPO4, 0.6 mM KH2PO4, 1 mM CuSO4, 1 mM MgSO4, 10 mM Hepes, 10 mM d-glucose, and 130 mM NaCl (Panreac, Barcelona, Spain). After this new incubation, the buffer was replaced by a chloride-free buffer, with or without the corresponding agonist or antagonist. In this buffer, NaCl was replaced by an equimolar concentration of NaNO3 (Panreac). Plates were then read on Fluoroskan FL fluorescence plate reader (Thermo Labsystems Inc., Waltham, MA) following the methodology described by Huete et al. (2011).
Melatonin receptor antagonists luzindole (nonspecific antagonist melatonin receptor), 4-P-P-DOT (4-phenyl-2-propionamidotetralin), DH97 (N-pentanoyl-2-benzyltryptamine; MT2 receptor antagonists; Tocris, Bristol, UK) (100 µM), prazosin (α-1 and MT3 receptor antagonist; Santa Cruz Biotechnology, Dallas, TX) (15 nM, 150 nM, 1.5 µM), specific α-1 antagonist corynanthine (Santa Cruz Biotechnology) (100 µM), protein kinase C (PKC) inhibitors staurosporine (100 nM) and bisindolylmaleimide I (1 µM), phospholipase C (PLC) inhibitor U73122 (1-[6-[[(17β)-3-methoxyestra-1,3,5(10)-trien-17-yl]amino]hexyl]-1H-pyrrole-2,5-dione; 3 µM), and protein phosphatase 1/2A okadaic acid (100 nM; Tocris) were added in chloride-containing buffer and maintained in a chloride-free buffer. Forskolin (40 µM) and IBMX (3-isobutyl-1-methylxanthine; 50 µM) were added only in a chloride-free buffer.
Data are expressed as mean ± S.E.M. of relative fluorescence units normalized to the initial time (Ft − F0), where Ft is the fluorescence at time t and F0 is the initial fluorescence.
Calculation of the Stern-Volmer Constant.
To calculate Stern-Volmer constant (Ksv) in NPE cells, we used the double ionophore technique. As described by West and Molloy (1996), isosmotic buffers with different concentrations of chloride were used to create a range of chloride concentrations. The whole range of concentrations was assayed simultaneously in the same plate to avoid MQAE leakage problems during monitoring of chloride concentrations in the given sample.
To equilibrate intracellular and extracellular buffer chloride concentration, 10 μM tributyltin (Sigma-Aldrich) and 5 μM nigericin (Sigma-Aldrich) were added to the buffer. Applying changes in fluorescence in cells under these different chloride buffers, a Stern-Volmer plot was obtained, responding to the equation:where F0 is the MQAE fluorescence in the absence of chloride, F is the MQAE fluorescence in presence of chloride, and Ksv (Stern-Volmer constant) is the slope of linear plot representing the efficiency of collisional quenching.
To determine concentration-response curves, different concentrations of melatonin (Sigma-Aldrich) and 5-MCA-NAT (Tocris) were tested according to the previous methodology. Concentrations tested varied in ranges from 1 nM to 150 μM. The Fmax and the slope of the straight segment of each dose curve were converted into % of fluorescence (Ft − F0) versus control (taken as 100%) and plotted. Data are plotted as percent of mean fluorescence (versus control) in relative fluorescence units ± S.E.M. versus logarithm of agonist concentration.
Cyclic AMP Studies.
Cyclic AMP accumulation was measured using cAMP Enzyme Immune Assay EIA KIT (Cayman Chemical Company, Ann Harbor, MI). Cells were grown to confluence in 6-well plates (Iwaki). Then medium was replaced with fresh medium containing different concentration of agents. Antagonists were preincubated over 15 minutes, and after 10 minutes of incubation with agonists, the medium was removed and immediately each well was incubated for 20 minutes in medium containing 275 μl HCl 0.1 M. The cells were then scraped and centrifuged at 1000g for 10 minutes, and the supernatant was assayed as indicated in the protocol of cAMP Cayman EIA Kit. The results were expressed as picomoles per milliliter to avoid errors due to the resuspension and protein quantification in small volumes. All wells were examined before the assay, and cells were counted to ensure homogeneity of the study. Results were shown as a mean ± S.E.M.
GraphPad Prism (GraphPad Software Inc., San Diego, CA) was used to obtain the linear regression (for the straight lines), nonlinear regression curves, and calculation of slope (for the straight lines and straight segments of the curves), pD2, and IC50 values. Statistical significance was calculated by analysis of variance (Bonferroni post tests) and Student’s t test, when needed. Value of P < 0.05 was taken as significant.
Effect of Melatonin and 5-MCA-NAT on Ciliary Body Epithelial Cells.
MQAE is a highly sensitive chloride fluorescence probe. This fluorescence is quenched in the presence of chloride so that changes in fluorescence are inversely proportional to changes in chloride concentration. By using the protocol specified in Materials and Methods, it was possible to verify the linear relationship between dye fluorescence and intracellular chloride concentrations in this cell type (Fig. 1A). By using this data it was also possible to calculate the Stern-Volmer constant (Ksv), whose value was 12.18 ± 0.89 M−1 (n = 6), in rabbit NPE cells. This constant permits us to calculate the intracellular Cl− concentration, which was 74.37 ± 3.8 mM (n = 6) in rabbit nonpigmented ciliary epithelial cells when external Cl− concentration was 130 mM.
By using this fluorescence probe, we were able to measure changes in the intracellular chloride concentrations after challenging the cells with melatonin and analogs. In particular, melatonin and 5-MCA-NAT were able to modify intracellular chloride. Normalized plots of fluorescence versus time always presented sigmoid patterns. From these plots, three different parameters were calculated: the maximal fluorescent signal, Fmax, which corresponded to the minimal intracellular chloride concentration ([Cl−]i); t50, which corresponded to the time necessary to produce 50% of Fmax, an indication of how fast was the release of chloride; and finally, the slope of the curve straight segment representing the velocity (V) of the chloride efflux. We used these parameters to evaluate the treatments versus the untreated cells that were taken as controls. These control cells showed the normal release of chloride of this cell type in chloride-free buffer. As previously indicated, nontreated cells depicted a sigmoid behavior that was consistent to a chloride efflux from inside the cells to the extracellular milieu (Fig. 1B). From this curve it was possible to calculate Fmax, t50, and V values presented in Table 1.
Melatonin and 5-MCA-NAT (100 μM) clearly and significantly changed the Cl− efflux as can be seen in Fig. 1B. There were differences in the Fmax, t0.5, and V when comparing melatonin and 5-MCA-NAT with control (Fig. 2; Table 1). Interestingly, both melatonin and 5-MCA-NAT showed a strong inhibition compared with control. Indeed, melatonin completely inhibited chloride release for about 602 seconds (10.0 minutes) and 5-MCA-NAT for roughly 860 seconds (14.3 minutes). After that, and in the presence of these two compounds, the slope of their respective curves was not as steep as the control and, moreover, they did not reach the Fmax the control did (Fig. 1B).
Concentration-Response Curves for Melatonin and 5-MCA-NAT.
To fully study the effect of melatonin and 5-MCA-NAT on chloride fluxes, cells were challenged with graded concentrations of both compounds following the protocol described in Materials and Methods. We focused on how these two compounds were able to diminish cell fluorescence (Fmax) and its concentration dependency. In this sense, and as can be seen in Fig. 2A, both compounds depicted concentration-response curves that were almost identical. From both curves it was possible to obtain pD2 value of 4.5 ± 1.2 for melatonin and 4.4 ± 1.0 for 5-MCA-NAT (n = 5). These values corresponded to EC50 values of 31.6 and 39.8 μM for melatonin and 5-MCA-NAT, respectively.
When, instead of studying the relationship between concentration and cell fluorescence, we analyzed concentration versus changes in the slope (velocity) for melatonin and 5-MCA-NAT, sigmoidal curves were obtained, providing interesting data. As observed in Fig. 2A, melatonin presented a pD2 value of 4.7 ± 0.2, whereas 5-MCA-NAT provided a pD2 value of 5.0 ± 0.1, corresponding to EC50 values of 19.9 and 10 μM for melatonin and 5-MCA-NAT, respectively (n = 5). Interestingly, the Hill slopes for both compounds were different, their values being 0.6 ± 0.3 for melatonin and 1.8 ± 0.2 for 5-MCA-NAT (Fig. 2B) (n = 5).
Studies with Antagonists.
Although the presence of melatonin receptors, mainly MT2 and MT3, has already been described in the ciliary body nonpigmented epithelial cells, we tried to investigate which receptor is involved in the changes in the intracellular chloride concentrations. In this sense, the MT2 antagonist 4-P-P-DOT was unable to modify the effect of melatonin. Interestingly another MT2 antagonist, DH97, and the nonselective melatonin receptor antagonist luzindole, enhanced the effect triggered by melatonin and 5-MCA-NAT. Melatonin effect in the presence of DH97 reduced Cl− efflux from 47.2 ± 5.3 to 21.1 ± 1.4% and luzindole to 19.8 ± 1.1% (n = 6).
5-MCA-NAT effect changed from 43.4 ± 3.4% (alone) to 33.3 ± 1.3% when 4-P-P-DOT was present, to 17.9 ± 2.3% in the presence of DH97, and to 17.5 ± 2.1% when luzindole was present (n = 6) (Fig. 3). In the same sense, prazosin (MT3 antagonist receptor) 15 nM surprisingly enhanced the effect of 5-MCA-NAT to 29.5 ± 4.3%. Nevertheless, the values were very close to those of 5-MCA-NAT alone (44.8 ±3 .5%) when we used 150 nM prazosin and partially reverted 5-MCA-NAT effect (78.3 ±1 6.5%) when it was used at 1.5 µM concentration (n = 6) (Fig. 4A). Similar results were obtained using melatonin (n = 6) (Fig. 4B); however, no statistically significant differences were reached between melatonin and melatonin + 15 nM prazosin. This lack of statistical significance difference probably is due to a relatively high S.E.M regarding “n” used and not to a difference in the behavior of both substances.
Adrenoreceptor α-1 Implications in Chloride Regulation.
The results obtained with prazosin at a low concentration produced an enhancement of 5-MCA-NAT hypotensive effect, whereas at high concentration, it produced the partial inhibition of 5-MCA-NAT effect, suggesting the involvement of an adrenoreceptor α-1 in the regulation of chloride secretion.
As is shown in Fig. 5, A and B (n = 6), corynanthine (α-1 antagonist) was able to enhance the effect of 5-MCA-NAT on chloride efflux from 43.36 ± 5.53 to 28.18 ± 3.46%, confirming the role of α-1 receptors in the regulation of chloride flux. Corynanthine alone has no effect on the chloride secretion (data not shown).
Second Messengers Triggered by Melatonin and 5-MCA-NAT.
It has been claimed that MT3 melatonin receptors are coupled to the PLC/PKC pathway (Huang et al., 2001). To see whether the effect of melatonin and 5-MCA-NAT was triggering this intracellular pathway, different blocking agents of this route were tested in their ability to modify the Cl− effluxes triggered by melatonin and 5-MCA-NAT.
As is shown in Fig. 6, none of the compounds tested to inhibit the PLC/PKC pathway was able to produce a change in the fluorescence signal, either alone or in the presence of melatonin or 5-MCA-NAT.
Involvement of cAMP Pathway.
We decided to investigate the canonical cAMP pathway that has been described to be negatively coupled to both MT1 and MT2 receptors. When adenylate cyclase activity was increased by means of a forskolin and IBMX mixture (see Materials and Methods), we could notice a reduction in Cl− efflux as observed in Fig. 7A. Because the chloride efflux triggered by adenylate cyclase activation resembled the behavior of melatonin and 5-MCA-NAT, we studied the ability of these two substances to increase cAMP concentrations. The results showed that 100 μM melatonin was able to increase intracellular cAMP levels of 58.05 ± 4.22 to 84.59 ± 6.78 pmol/ml (n = 6). In the case of 5-MCA-NAT, 100 μM results were similar, obtaining an intracellular cAMP concentration of 90.55 ± 5.53 pmol/ml (n = 6) in the presence of the melatonin analog.
Different concentrations of melatonin and 5-MCA-NAT were tested to characterize the dose-dependent behavior of these substances. As shown in Fig. 7B, graded concentrations of melatonin and 5-MCA-NAT evoked the accumulation of concomitant amounts of cAMP in rabbit NPE cells. The concentration-response curve provided a pD2 value of 4.6 ± 0.2 for melatonin and 4.9 ± 0.7 for 5-MCA-NAT, which were equivalent to EC50 values of 22.0 and 19.4 μM (n = 6) for melatonin and 5-MCA-NAT, respectively.
To investigate the involvement of the MT3 receptor in this signaling pathway we blocked this receptor using 1.5 µM prazosin (n = 6), measuring the accumulation of intracellular cAMP. The results were presented in Fig. 8. Data showed a partial and significant reversion of intracellular cAMP from 92.34 ± 6.36 (5-MCA-NAT) to 74.36 ± 4.32 pmol/ml (5-MCA-NAT + prazosin).
The present manuscript describes the effect of melatonin and its analog 5-MCA-NAT acting on melatonin receptors of ciliary body nonpigmented epithelial cells. The main action of melatonin and its analog is the modulation of intracellular chloride concentrations, which is important because chloride rules water movement and therefore is the key ion driving the production of the aqueous humor (Civan and Macknight, 2004). The aqueous humor is responsible for the correct eye shape and acts as a nutritional fluid for avascular structures such as the lens or the cornea (Civan, 1998). Under certain circumstances, a lack of drainage of the aqueous humor produces an elevation of IOP possibly responsible for the pathology termed glaucoma.
There are different reports indicating that melatonin application reduces or increases IOP (Caprioli and Sears, 1984). We claim that melatonin and analogs reduce IOP in New Zealand white rabbits, glaucomatous monkeys (Serle et al., 2004), and even humans (Ismail and Mowafi, 2009).
Aqueous humor formation relies on the ability of the ciliary body cells [pigment epithelium (PE) and NPE] to mobilize chloride ions from the stromal part of the ciliary body to the posterior chamber of the eye (Do and Civan, 2004). The results presented in this manuscript suggest that melatonin and 5-MCA-NAT may reduce IOP because they decrease the efflux of chloride from the cytoplasm of NPE toward the extracellular space. The effect of both melatonin and 5-MCA-NAT inhibiting this ion movement was strong during 10 and 14 minutes for melatonin and 5-MCA-NAT (100 μM), respectively. After this interval of inhibition, the rate of efflux increases, although slope and Fmax were always below the control values. This fact indicates that the inhibitory effect does not only affect initial chloride efflux but inhibition was also present when the equilibrium was reached (see Fig. 1). The substantial inhibition of the chloride release and the presumable inhibition of the aqueous humor formation seem to be higher than the IOP reduction observed in vivo (Pintor et al., 2003). In this sense, it is important to notice that the in vitro model we are using, although mimicking the conditions present in the ciliary body, has certain limitations. One of the main in vitro restrictions is that we are measuring the secretory layer NPE cells, producers of the aqueous humor, and not the drainage system, possibly the reason for such differences. Another limitation to take into consideration is that the measurements we performed involve only Cl− efflux processes. This implies that there is a Cl− efflux after a concentration gradient (as occurs in the in vivo model) and, because of the lack of this ion in the extracellular buffer, mechanisms transporting this ion from the extracellular space to NPE cells cytoplasm are not fully activated. Consequently, the data obtained are dependent solely on the state of chloride channels and transporters. Also, we should emphasize the absence of the PE, which takes chloride from stromal and transfers it to NPE layer cells. The importance of PE and its function has been extensively described in the literature (McLaughlin et al., 1998; Do and To, 2000; Do et al., 2004a; Ni et al., 2006). However, the simultaneous study of both layers complicates precise conclusions concerning the contribution of each structure involved.
Although there are some limitations in the model we are using, it is important to emphasize that an analysis of the concentration-curves demonstrates that the effects depicted by melatonin and 5-MCA-NAT were highly similar but not identical (especially in Fmax), as we can see in Fig. 1. This matches the results obtained for both substances on IOP in New Zealand white rabbits, where the effect of melatonin as an hypotensive agent is less robust than that of 5-MCA-NAT (Pintor et al., 2003).
An interesting and unexpected observation was the effect of the classic melatonin antagonists. The MT2 antagonist DH97 and the MT1/MT2 nonselective antagonist luzindole potentiated the inhibitory effect of both melatonin and 5-MCA-NAT on chloride movement. This may suggest that melatonin and 5-MCA-NAT are acting through a different MT1/MT2 or luzindole-sensitive receptor. Moreover, when MT1/MT2 or luzindole-sensitive receptors were blocked, the effect of this putative receptor was enhanced, suggesting a different signaling pathway from the canonical described for melatonin receptors. This indirectly implies that the action of MT1/MT2 may have an opposite effect, increasing the efflux of Cl−. Because MT1 and MT2 melatonin receptors are negatively coupled to adenylate cyclase and there is a concomitant reduction in the concentrations of cAMP, the blockade of these two receptors might involve PLC/PKC pathway as happens in some models (Bowler et al., 1996; Godson and Reppert, 1997; Dortch-Carnes and Tosini, 2013). Interestingly, new studies are appearing indicating that new second messenger systems can be coupled to melatonin receptors. For instance, melatonin receptors produce a reduction in sodium nitroprusside–released nitric oxide and cGMP levels in human nonpigmented ciliary epithelial cells (Dortch-Carnes and Tosini, 2013). These authors indicate that, at least in part, melatonin and analog effects can use this pathway in human NPE cells and that this second messenger system might be responsible for the melatoninergic compound hypotensive effect. These results are perfectly compatible with those presented here and with previous ones in which we suggested that MT2 melatonin receptor agonists reduce IOP (Alarma-Estrany et al., 2008).
Coming back to the second messengers being activated by melatonin and 5-MCA-NAT in our model, we were unable to detect any involvement of the PLC/PKC pathway. Some authors claimed that 5-MCA-NAT is not a selective MT3 receptor agonist but it could activate MT1 or MT2 melatonin receptors (Vincent et al., 2010). If this were so, we would expect reductions in the cytosolic concentration of cAMP, because, as already mentioned above, MT1 or MT2 melatonin receptors are negatively coupled to adenylate cyclase (Vanecek, 1998). The study on the ability of melatonin and 5-MCA-NAT to inhibit cAMP formation pointed in the opposite direction, because it was impossible to see a reduction in cAMP production. On the contrary, we could observe that melatonin and its analog increased cAMP concentrations in a concentration-dependent manner. This is not a common mechanism of signal amplification, but it has been described in some models (Raviola, 1974; Beraldo and Garcia, 2005; Schuster et al., 2005). Indeed it modified chloride efflux, and this effect was mimicked when we applied forskolin and IBMX. The relationship between the increase of cAMP and the decrease of IOP is widely studied in the eye. In rabbits and monkeys, it is known that this decrease in IOP is associated with a significant decrease in AH secretion. Interestingly, it was reported that 5-MCA-NAT is able to produce important increases in cAMP in chick retinas by a mechanism that may involve an MT3 (Mel1c) binding site (Sampaio, 2009).
When the involvement of the putative MT3 melatonin receptor was studied, the use of the only available antagonist prazosin was tested. Low concentrations of this antagonist produced an unexpected potentiation of the action triggered by 5-MCA-NAT. Nevertheless, at higher concentrations (1.5 μM), prazosin had its antagonistic effect blocking melatonin and 5-MCA-NAT actions, indicating that at low micromolar concentrations it acts as a MT3 receptor antagonist as previously described elsewhere (Dubocovich et al., 2003; Pintor et al., 2003; Alarma-Estrany et al., 2011).
Concerning the results obtained with prazosin at nanomolar concentrations, the observed contradictory effect potentiating 5-MCA-NAT action can be explained as an effect performed on α-1 adrenoceptors. This point was confirmed when the same effect was obtained with the selective α-1 adrenoceptor antagonist corynanthine, as was also described regarding the modulation of IOP in rabbits by other authors (Chidlow et al., 2001).
In the same way, we tested the action of prazosin, in micromolar range, on the cAMP accumulation induced by 5-MCA-NAT. Prazosin was able to partially revert the increased intracellular cAMP. This fact relates the cAMP increase with the MT3 melatonin receptor.
It was demonstrated that the chloride efflux from NPE cells is one of the most, if not the most important, ion controlling aqueous humor secretion as previously commented (Jacob and Civan, 1996; Forrester, 2002). However, the nature of the proteins ruling the movement of Cl− (Ritch et al., 1989; Paulmichl et al., 1992; Chen et al., 1999; Do et al., 2006) and the transmitters regulating chloride secretion and secondarily IOP is not clear. For example, traditionally it has been accepted that the blockade of β-adrenergic receptors is a pharmacologic approach to decrease IOP (Freddo, 1987). β-Adrenergic receptors are positively coupled to adenylate cyclase but, surprisingly, cAMP activates some chloride channels. The explanation seems to be that the effects of cAMP itself and β-adrenergic receptors may reflect a complex regulation evolving different mechanisms (McLaughlin et al., 2001) and presumably several channels and transporters as suggested by Do and Civan (2004). Recently, it was demonstrated that the increase in cAMP levels does not only activate protein kinase A but also may have a direct effect on ion channels. Fleishauer and coworkers (2001) demonstrated that cAMP action is conducted via a direct action on chloride channels and not by means of protein kinase A in the ciliary body. Also experiments performed using patch clamp techniques have demonstrated that cAMP can activate chloride channels such as the maxi-Cl−, although its physiologic role in situ has not been fully elucidated (Mitchell and Civan, 1997; Do et al., 2004b). Small changes in cAMP cytosolic concentrations may produce vast changes in the chloride efflux. In this sense, Huang and coworkers (2001) demonstrated that small changes in cAMP levels can cause profound variations in the lung chloride efflux. In addition, transepithelial chloride secretion in the bovine and porcine cilliary body is severely inhibited by cAMP (Do et al., 2004a; Ni et al., 2006). These results may explain and reinforce the results obtained with melatonin and 5-MCA-NAT described in the present manuscript, also matching the results obtained by Sampaio (2009).
The results obtained in this study suggest that a prazosin-sensitive melatonin receptor, different from the MT1 or MT2, is responsible for the decrease in Cl− outflow. Both melatonin and 5-MCA-NAT would bind to the three melatonin receptor subtypes, probably with different affinity. MT1/MT2 receptors produce an opposite effect on intracellular cAMP levels but not so intense as the noncloned melatonin receptor (putative MT3 receptor). This could explain the observed effect when the inhibition of MT1 and MT2 melatonin receptors is performed. Differences in the affinity of the agonists for each receptor subtype may explain the observed changes in cAMP at nonsaturating concentrations of these two agonists. When melatonin acts, it does so mainly through MT1/MT2 receptors, whereas 5-MCA-NAT is through MT2 and preferentially by MT3. Also, there is more and more evidence pointing to melatonin receptor heterodimerization to explain these differences. In this sense, Ayoub and coworkers (2004) described that MT2 melatonin receptors can form heterodimers with MT1 receptors and that the formation of these heterodimers modifies the affinity for the agonists and antagonists. Moreover, there are cases where heterodimerization produces significant changes in ligand binding, signaling, or trafficking, and this may also explain the activation of AC by melatonin and 5-MCA-NAT (Gazi et al., 2002). More work in this area is necessary to fully understand the real mechanism underlying the melatonin receptors involved in the production of the aqueous humor in our model.
In summary, we can conclude that melatonin and 5-MCA-NAT acting through MT3 melatonin receptors can participate in the modulation of certain chloride channels by a process that involves cAMP increase. This might be the potential mechanism to reduce and modulate net aqueous humor secretion, explaining why melatonin and some of its analogs can reduce IOP in experimental animals (Pintor et al., 2001, 2003).
The authors thank Penny Rollinson for help with the English edition of the manuscript.
Participated in research design: Pintor.
Conducted experiments: Huete-Toral.
Performed data analysis: Martínez-Águila.
Wrote or contributed to the writing of the manuscript: Crooke.
- Received July 11, 2014.
- Accepted October 23, 2014.
This work was supported by grants from the Spanish Ministry of Economy and Competition [Grants SAF2010-16024 and SAF-2013-44416-R]; and the Ministry of Health Social Services and Equality RETICS [Grant RD12/0034/0001]. Additionally, this work was supported by Spanish Ministry of Economy and Competition studentship to F.H.-T. and Universidad Complutense de Madrid studentships to A.M.-Á.
- N-butanoyl 2-(9-methoxy-6H-isoindolo[2,1-a]indol-11-yl)ethanamine
- intraocular pressure
- 5-methylcarboxyamino-N-acetyl tryptamine
- N-(6-methoxyquinolyl) acetoethyl ester
- nonpigmented epithelial cells
- pigment epithelium
- protein kinase C
- phospholipase C
- Copyright © 2014 by The American Society for Pharmacology and Experimental Therapeutics