Melatonin (N-acetyl-5-methoxytryptamine) is involved in the regulation of several cellular functions. It is secreted by the pineal gland but also produced in many other sites such as retina, skin, gut, and bone marrow cells, although the physiological implication of these extrapineal sites, with the exception of retina, is mostly obscure ((1, 2) and ref. therein cited).
In humans, the melatonin-generating system is photosensitive displaying a circadian rhythm, with the highest concentration levels produced at night in the darkness (3, 4).
Popular for preventing jetlag or as adjuvant in elderly people with sleep-problems, melatonin functional versatility (5, 6) has been so far explained by an established antioxidant activity (7). Melatonin exerts its action either directly via bulk reactions or indirectly via cell receptor-mediated signaling (8), both favoring maintenance of the cell redox balance. Relevant to cell bioenergetics melatonin is highly liposoluble, exerting the antioxidant action also in the mitochondrion (9, 10). The molecular mechanisms by which melatonin acts as a hormone at this level and correlation with its circadian synthesis still remain largely obscure.
The receptor-mediated melatonin signaling involves different types of cell surface and nuclear receptors, expressed to a different extent by cell-lines and tissues, including skin. Among keratinocytes, the HaCat cell lines used in this study express the cell surface MT2-type receptor (isoform MT2b) and the nuclear receptors, that is, the retinoid orphan receptor (RORα) and the NQO2 flavoprotein, also known as the MT3 melatonin binding site (11). The receptor mediated chemistry triggered by melatonin appears complex and is still matter of debate (8, 12); it is worth considering that an uptake of melatonin by HaCaT cells has been already described, although at melatonin concentration values in the μM–mM range (13, 14).
Nitric oxide (NO) is produced endogenously by the nitric oxide synthase isoforms, NOSs (15); the constitutive NOSs, that is, the endothelial NOS (eNOS) and the neuronal NOS (nNOS), produce NO in the nM range, whereas the inducible NOS (iNOS) releases up to μM NO (16, 17). NO exerts a controlled inhibitory role on the mitochondrial respiratory chain through the reaction with complex I (18) and complex IV (19–21). The inhibition of complex IV is rapid and reversible (20, 22, 23); a pulse of NO induces a depression of oxidative phosphorylation (OXPHOS) and activation of glycolysis, in cells able to sustain it (24, 25). Stimulated by the suggested implication of NO in the circadian cycle (26, 27) and by the finding that the major NO oxidation products, nitrite and nitrate, also follow a circadian rhythm with a night peak (28), we have investigated the interplay between melatonin, NO, and cell bioenergetics by monitoring the mitochondrial response of cultured keratinocytes. We have followed the parameters of interest over several hours incubation with nM melatonin to grossly mimic a night physiological cell exposure to melatonin. On this time scale, we have observed the rise of the nNOS mRNA, paralleled by the production of NOx and leading to a shift of the cell metabolism from OXPHOS to glycolysis.
MATERIALS AND METHODS
Dulbecco's modified Eagle's medium (DMEM) and fetal bovine serum (FBS) were from Invitrogen Life Technologies (GIBCO) (Paisley, UK) and from PAA (Linz, Au). Melatonin, JC-1, and all other reagents were purchased from Sigma (St. Louis, MO, USA), unless otherwise specified.
Stabilised human keratinocytes (HaCaT, ATCC, cell lines USA) were grown at 37 °C, 5% CO2, 95% air in DMEM containing 4.5 g/L glucose, 10% FBS, 0.05 mg/mL gentamycin, and 2 mM L-glutamine in 25-cm2 flasks or multiwell plates. Before melatonin treatment, cells were incubated ∼ 24 h in 1 g/L glucose DMEM (w/o FBS and phenol red). When necessary, HaCaT cells were harvested by trypsinization and centrifugation (1,000 × g) and carefully suspended in the working medium, at suitable density (see text). Cell lysis was performed by CelLytic™M Cell Lysis reagent in the presence of the Protease Inhibitor Cocktail or by TRIzol (Invitrogen, Paisley, UK); protein content was determined according to Bradford (29).
NOS mRNA Determination
NOS mRNA was determined by Quantitative Real Time Polymerase Chain Reaction (QRT-PCR); HaCaT cells were harvested (∼ 3 × 106 cells) and total RNA was isolated (30). The RNA (1 μg) reverse transcription was performed using SideStep™ II QRT-PCR cDNA Synthesis Kit (Stratagene). QRT-PCR was performed using primers designed by BioRad Laboratories (Hercules, CA, USA) (Software Beacon Designer) and purchased by PRIMM (Milano, Italy). SYBR green-based (Brilliant_ SYBR_ Green QPCR Master Mix, Stratagene) QRT-PCR was performed using a MJ Mini Opticon Detection System (BioRad Laboratories). The following protocol was used: denaturation step (95 °C for 5 min), amplification step (95 °C for 10 sec, 55 °C for 30 sec) repeated 45 times. All reactions were performed in triplicate. Melting curve analysis was performed at the end of every run to ensure a single amplified product for each reaction. β-actin gene (PRIMM) was used for normalization. QRT-PCR reagents, from Stratagene (Santa Clara, CA, USA). Quantification was performed using the Gene Expression analysis for iCycler iQ Real-Time PCR detection system (Version 1.10, BioRad Laboratories).
NOx (Nitrate/Nitrite) Determination
The accumulation of NOx was determined fluorimetrically (Fluorimetric Assay Kit, Cayman Chemical Co., Ann Arbor, MI, USA) in the culture medium of cells grown under standard conditions, and after 6 h incubation with melatonin (see text).
The cell adenosine-5′-triphosphate (ATP) concentration was quantified by chemiluminescence, under stationary conditions or kinetically by following the rate of ATP production. Cells were incubated overnight in an antibiotic/FBS-free DMEM medium, and the following day, melatonin (final concentration 1 nM) was added to the cells (∼ 5 × 104 cells/well) for 6 h.
The stationary ATP measurements were performed using both melatonin treated and control cells suspended in phosphate-buffered-saline (PBS) containing L-glutamine (2 mM), in the presence or absence of glucose, 11 mM; when necessary, oligomycin (2.5 μg/mL) was added over the last 1.5 h incubation. The rate of ATP production was evaluated as described in (31) and after cell membrane permeabilization with digitonin (60 μg/mL, 20 min at 25 °C). The assay was performed in the presence of iodoacetamide (2 mM) and the adenylate kinase inhibitor, P1, P5-di(adenosine-5′) pentaphosphate pentasodium salt (25 μM). The ATP synthesis was induced by adding to the permeabilized cells, succinate (20 mM) and ADP (0.5 mM), in the presence of rotenone (4 μM) (31). A reference sample was assayed in the absence of succinate and in the presence of antimycine A (18 μM) and oligomycin (2 μM).
ATP measurements (Perkin Elmer, ATPlite) were performed using a VICTOR™ Multilabel Counter (Perkin Elmer, USA) equipped with 96-well plates (ViewPlate-96, white, Perkin Elmer, Waltham, USA).
Cells (∼ 3 × 106), precultured ∼ 24 h in antibiotic/FBS-free DMEM, were incubated with melatonin, 1 nM, for 6 h and according to (32): 2 h in PBS containing glucose (1 mM), followed by 1 h in PBS without glucose. Thereafter, over the further 3 h incubation, glucose was readded to the cells, in the presence or absence of myxothiazol and antimycin A (10 μM each); the reaction was stopped using HClO4 (33). Lactate was determined spectrophotometrically on the cell supernatant.
Mitochondrial Membrane Potential Measurements
The mitochondrial proton electrochemical potential gradient (δμH+) of HaCaT cells after ∼ 6 h incubation with melatonin (1–100 nM), was measured following the accumulation into the mitochondrial matrix of the fluorescent, cationic probe JC-1 (Sigma Chem. Co). Briefly, JC-1 (0.4 μM) was added to cells (∼ 1 × 106) premixed with nigericin. The kinetics of JC-1 accumulation was measured at 595 nm (VICTOR™ Multilabel Counter, Perkin Elmer) in the presence of ouabain, 0.5 μM, to avoid aspecific cell fluorescence (34).
Cells (1 × 106) were lysed and centrifuged at 20,000 × g and supernatants assayed for citrate synthase (35) and for total protein content (29). HaCaT cells used as such or after incubation with melatonin display a similar protein content and citrate synthase activity (not shown).
The presence of melatonin was evaluated by immunoassay using the “Direct saliva Melatonin ELISA kit” Bühlmann (Schönenbuch, CH), adapted to cell lysates. Cells (3 × 106) were incubated with melatonin (1 nM) for 1 h and 5 h. After incubation, both melatonin-treated and control cells were harvested, resuspended (at 106/mL density), and lysed. Aliquots of lysates were loaded on enzyme-linked immunosorbent assay (ELISA) plate containing antimelatonin polyclonal antibody.
The number of independent measurements is indicated in figure legend. Significance was determined using the Student t-test, run by Excel (Microsoft Windows platform). The error bars correspond to the standard error of the mean (SEM); all P values correspond to two-sided sample t-test assuming unequal variances. A P value ≤ 0.05 was considered significant.
NOSs Gene Expression
The intracellular mRNA expression levels of the eNOS, nNOS, and iNOS were measured by quantitative RT-PCR in HaCaT cells incubated for different times with increasing amounts of melatonin. As shown in Fig. 1, compared with controls, melatonin used at 0.5 and 1 nM concentrations produced a 2.15- and 4-fold increase, respectively, in nNOS mRNA level, whereas under similar conditions the eNOS mRNA did not vary significantly. The iNOS mRNA assayed in parallel remained undetectable under all conditions (not shown). The time dependent profiles of the nNOS expression stimulated up to 7.5 h by 1 nM melatonin, are shown in Fig. 2. The nNOS mRNA expression level increases with the incubation time, reaching a 1.95-fold increase after 3 h and a maximum 3.75-fold increase after 6 h incubation.
The accumulation of NOx in the medium after 6 h incubation with increasing amounts of melatonin is shown in Fig. 3. The production of NOx at ∼ 1 nM melatonin is ∼ 2 times higher than in controls. Melatonin used at 10 and 100 nM concentrations did not affect the NOx production.
Melatonin Determination in HaCaT Cells
We investigated the presence and stability of melatonin in HaCaT cells by immunoassay (ELISA). Cells were incubated with melatonin, 1 nM, for 1 h and 5 h, and the amount of melatonin up-taken by the cells was measured after cell washing and lysis. The results are reported in Table 1 and show that under the conditions chosen a significant amount of melatonin enters the cells, without significant changes over the incubation time.
Table 1. Cell melatonin concentration
pmoles/1 × 106 cells (± SEM)
P < 0.05.
HaCaT cells (3 × 106) were incubated with melatonin (1 nM) for 1 h and 5 h. After incubation, the melatonin-treated and control cells were harvested, resuspended (at 106/mL density), and lysed for ELISA immunoassay (see Materials and Methods section).
The ATP concentration levels of cells treated for 6 h with melatonin, 1 nM, was measured either at steady state in the presence or absence of glucose or kinetically by following the rate of ATP production (Fig. 4).
When measuring the stationary levels of ATP, glucose starvation was performed to promote OXPHOS and minimize the glycolytic contribution to the ATP production (Warburg effect) (36). In the absence of glucose, the addition of oligomycin induces a dramatic decrease of the ATP level, due to OXPHOS inhibition (Fig. 4, panel a); under these conditions, the difference between the ATP measured in the absence and presence of oligomycin, defined in Fig. 4 as ΔATP is, within a gross approximation, indicative of the ATPOXPHOS (arrows in Fig. 4, panel a). The results suggest that the ATPOXPHOS is lower in melatonin treated cells with respect to control cells. In the presence of glucose, as glycolysis can take place and compensate for loss of ATPOXPHOS the effect of oligomycin is no longer evident, and the ATP concentration levels of melatonin treated cells and controls are similar (Fig. 4 panel b and c).
The existence of a melatonin-induced depression of the ATPOXPHOS production was verified directly by measuring the rate of succinate-driven ATP synthesis of the cells (31). After 6 h incubation with melatonin, 1 nM, the rate of ATP production is 0.51 (± 0.044) nmoles/min × 106 cells to be compared with 0.69 (± 0.028) nmoles/min × 106 cells (Fig. 4 panel d). Taken together, the results suggest that melatonin induces a significant ∼ 25% depression of ATPOXPHOS production.
The effect of nM melatonin on the relative contribution of OXPHOS and glycolysis to the overall ATP production was independently evaluated by comparing the concentration of lactate produced in the presence of glucose and in the absence or presence of the respiratory chain inhibitors, antimycin A and myxothiazole (Fig. 5). The basal lactate concentration in controls is on average 68 μg/ml × 106 cells (Fig. 5a); this value rises to 80 μg/ml × 106 cells after melatonin treatment. The increased lactate production induced by melatonin appears consistent with the conclusion that melatonin partially inhibits OXPHOS, and glycolysis is compensating: similarly, in the presence of antimycin A and myxothiazole, the lactate increases in controls as in melatonin treated cells; in these latter, to a minor extent as expected if the effect was additive. The basal lactate, that is, the lactate produced in the absence of myxothiazol and antimycin A, is proportional to glycolytic ATP, whereas the difference between the lactate detected in the presence of myxothiazol and antimycin A, and the basal lactate (defined as Δlactate) is proportional to the OXPHOS ATP. The basal lactate value divided by Δlactate was taken as indicative of the ratio between ATPglycolytic and ATPOXPHOS (37, 38), as confirmed by (32). As shown in Fig. 5b, the ATPglycolytic/ATPOXPHOS ratio is ∼ 1.1 in controls and ∼ 2.1 in the presence of melatonin.
Mitochondrial Membrane Potential
The effect on mitochondrial ΔμH+ of melatonin added to the cell culture medium was evaluated by following the ΔΨ-dependent accumulation in the mitochondrial matrix of the fluorescent probe JC-1.
Typical fluorescence kinetics of JC-1 aggregates formation is shown in Fig. 6. After 6 h incubation with 1 nM melatonin, the ΔF ∝ΔΨ is decreased, by ∼ 20%. This decrement reaches 36% when cells experience starvation from arginine. The mitochondrial ΔΨ-depression induced by melatonin was reverted by washing cells and allowing further 3 h incubation in a melatonin-free medium.
The key finding of this work is that in the presence of nanomolar amounts of melatonin and after a few hours incubation, the basal level of the cellular nNOS expression rises by a factor of 4 thereafter returning to basal level. Almost synchronously also the production of NOx increases while the mitochondrial membrane potential decreases. A partial decrease of ATP OXPHOS production is also observed, as well as a compensatory increase of glycolytic ATP, expected on the basis of the Warburg effect typically occurring in the presence of glucose sustaining glycolysis. Taken together, these findings suggest that, triggered by melatonin and over the time window explored, a fraction of mitochondrial complex IV reacts with NO and is reversibly inhibited (20, 23, 39).
The inhibition of ATP OXPHOS production is better observed in the absence of glucose, when the contribution of glycolysis is minimized. Under these conditions, the overall ATP level is almost halved, compared with cells maintained in glucose (40) and the addition of oligomycin causes a dramatic decrease of ATP, whose amplitude is indicative of the OXPHOS contribution to the overall ATP detected (ΔATP in Fig. 4). Interestingly, the ΔATP induced by oligomycin is reproducibly smaller in melatonin-treated cells, likely due to the (just mentioned) partial inhibition of complex IV because under the conditions explored, that is, low NO concentration, the involvement of complex I, though possible, is less likely (41). It is worth noting that when the cells were starved from the NOS substrate arginine, the electrophoretic import of JC-1 was the highest, confirming that cultured cells may express a basal NOS activity, releasing NO and depressing the mitochondrial potential to a measurable extent (42). The decrease of OXPHOS ATP induced by melatonin, though statistically significant is small, likely due to glycolytic compensation; therefore OXPHOS impairment was further substantiated measuring lactate (32). After exposure to melatonin, the ATPglycolytic/ATPOXPHOS ratio, worked out through lactate measurements, is ∼ two-fold higher, confirming that nM melatonin drives the equilibrium between ATPOXPHOS and ATPglycolytic, toward the latter (Fig. 5b).
Within the limits of the heterogeneous distribution of melatonin among organ and tissues (43), it is intriguing that the concentrations of melatonin, in the nM range, inducing the observed changes of the nNOS mRNA are compatible with the concentration of melatonin circulating in the human blood at night peak time, that is, ∼ 100–200 pg/mL (1, 6), corresponding to ∼ 1 nM melatonin. Even more interestingly, under these conditions, the upregulation of glycolytic enzymes has been observed in the rat pineal gland (44), suggesting a correlation between melatonin and mitochondrial metabolism.
The up-regulation of nNOS herein reported seems specific because, over the same time period, the eNOS expression remained constant and the iNOS was not detectable. Consistently with a receptor mediated nuclear DNA-activated process, no effects of melatonin were observed at incubation times earlier than 4–5 h. At melatonin concentrations higher than nM, and/or by further extending the melatonin incubation time the nNOS expression decreases, suggesting the existence of a time- and concentration-dependent negative feed-back control of the nNOS mRNA expression, at least in HaCaT cells. The mRNA expression level observed at 1 nM melatonin is about the double of that observed at 0.5 nM, whereas the corresponding NOx levels are similar. This finding is of complex interpretation because possibly related to post-transcriptional regulation of nNOS, as well as to different stationary stability of the nNOS mRNA and the NOx: a unique answer demands further investigation.
Indeed, it is possible that the mitochondrial effects attributed to melatonin (this article) might be also related to still active melatonin degradation products (2, 13), whose effectiveness remains to be elucidated. Regardless of their involvement, however, it is worth to consider that depending on the fraction of cytochrome-c-oxidase inhibited by NO, the O2 consumption and thereby the OXPHOS ATP synthesis, may be differently affected, particularly depending on the functional state (ATP synthesis) prevailing in mitochondria (19, 45). In state IV respiration, the electron flux through the respiratory chain is controlled at the level of complex I rather than at cytochrome-c-oxidase, so that inhibition of a relatively small fraction of cytochrome-c-oxidase may not result in a significant decrease of O2 consumption, while increasing the reduction level of the uphill respiratory chain components, hence the production of ROS (46, 47). On the contrary, when the rate-control of the respiratory chain is transferred to the cytochrome-c-oxidase, as it occurs typically with state III respiring mitochondria (21, 48), the rate of respiration together with synthesis of ATP OXPHOS decreases; in this case, a higher fraction of O2 might become available for close by cells (O2-diversion). HaCaT cells, whose relative population of state III and state IV mitochondria under the conditions explored is hard to determine and presently unknown, do not display a clear melatonin-induced depression of respiration (not shown). Still, in the presence of nM melatonin, we observe a depression of mitochondrial Δψ as well as a clear metabolic shift toward glycolysis.
In summary, based on the data provided by this study, nM melatonin in the environment of melatonin-sensitive cells produces a transient but substantial rise of the constitutive nNOS. The induced increase of NO production leads to transient, partial inhibition of the respiratory chain with a decrease of mitochondrial Δψ, depression of OXPHOS, and rise of glycolysis. All together these findings suggest that we are facing a new role of melatonin, involving mitochondria probably in a circadian context.
This work was partially supported by Ministero dell'Istruzione, dell'Università e della Ricerca of Italy (PRIN 2008FJJHKM_ 002 to P.S.).