Functional Expression of MT2 (Mel1b) Melatonin Receptors in Human PAZ6 Adipocytes

Abstract

Several reports have demonstrated that the pineal hormone, melatonin, plays an important role in body mass regulation in mammals. To date, however, the target tissues and relevant biochemical mechanisms involved remain uncharacterized. As adipose tissue is the principal site of energy storage in the body, we investigated whether melatonin could also act on this tissue. Semiquantitative RT-PCR analysis revealed the expression of MT1 and MT2 melatonin receptor mRNAs in the human brown adipose cell line, PAZ6, as well as in human brown and white adipose tissue. Binding analysis with 2-[¹²⁵I]iodomelatonin (¹²⁵I-Mel) revealed the presence of a single, high affinity binding site in PAZ6 adipocytes with a binding capacity of 7.46 ± 1.58 fmol/mg protein and a K_d of 457 ± 5 pm. Both melatonin and the MT2 receptor-selective antagonist, 4-phenyl-2-propionamidotetraline, competed with 2-[¹²⁵I]iodomelatonin binding, with respective K_i values of 3 × 10⁻¹¹ and 1.5 × 10⁻¹¹m. Functional expression of melatonin receptors in PAZ6 adipocytes was indicated by the melatonin-induced, dose-dependent inhibition of forskolin-stimulated cAMP levels and basal cGMP levels with IC₅₀ values of 2 × 10⁻⁹ and 3 × 10⁻¹⁰m, respectively. Modulation of the cGMP pathway by melatonin further supports functional expression of MT2 receptors, as this pathway was shown to be specific for that subtype in humans. In addition, long-term melatonin treatment of PAZ6 adipocytes was found to decrease the expression of the glucose transporter Glut4 and glucose uptake, an important parameter of adipocyte metabolism. These results suggest that melatonin may act directly at MT2 receptors on human brown adipocytes to regulate adipocyte physiology.

MELATONIN IS A hormone produced and secreted at night by the pineal gland. Its secretion is proportional to the duration of darkness, and it thus acts as a neuroendocrine transducer of photoperiodic information. Important physiological functions of melatonin in mammals are regulation of the circadian clock in the hypothalamic suprachiasmatic nuclei (SCN), regulation of seasonal reproduction, and inhibition of dopamine release from the retina (1).

For several years, melatonin has been known to also play a role in energy expenditure and body mass regulation in mammals (2). Hibernating species, for example, undergo dramatic changes in body weight, particularly in fat mass, in response to photoperiodic changes. These responses are mainly mediated by the length of nocturnal melatonin release (3). Direct evidence for melatonin’s effect on body weight comes from infusion experiments with pinealectomized hamsters (4) and injection of melatonin receptor agonists or antagonists into Garden dormice (5).

Variations in melatonin secretion patterns and body mass have also been observed in species that exhibit less dramatic seasonality. In rats and humans, for example, visceral fat levels increase with age, whereas nocturnal plasma melatonin peak concentrations decline (6–8). Daily melatonin supplementation to middle-aged rats restored melatonin levels to those observed in young rats and suppressed the age-related gain in visceral fat (9).

Clinical data also support a link between photoperiodic changes and control of energy balance in man. For example, certain human pathologies associated with desynchronized circadian rhythms and melatonin secretion pattern such as seasonal affective disorder share features common to the prehibernating mammal, such as oversleeping, carbohydrate craving, and overall weight gain (10). Notably, these changes may be triggered by seasonal changes in light, and light treatment has proven useful for correcting disturbances in melatonin rhythm (11, 12).

Mechanisms underlying melatonin’s effect on body weight regulation have yet to be characterized. In Djungarian and Syrian hamsters, the melatonin-induced decrease in fat mass has been associated with thermogenic activation of brown adipose tissue (BAT) (13, 14). In contrast, in other species, such as the dormouse, ground squirrel, or Syrian hamster, melatonin has been shown to induce an increase in fat mass (5) and to induce BAT hypertrophy (2).

Effects of melatonin on adipose tissue may also vary during development. Nocturnal melatonin levels are high in young children and decline with age. Interestingly, human BAT is best developed during early childhood, when melatonin levels are high, and is largely reduced in adults (7, 15). The role of melatonin in BAT may therefore be most significant during the early stages of development.

Melatonin mediates its effects through high affinity G protein-coupled receptors. In mammals, two distinct receptor subtypes have been cloned and named MT1 (Mel1a) and MT2 (Mel1b) (16–18). Both melatonin receptor subtypes inhibit adenylyl cyclase via pertussis toxin-sensitive G_i proteins. In addition, MT1 receptors have been shown to stimulate calcium mobilization through pertussis toxin-insensitive G_q/11 proteins, and MT2 receptors to couple to cGMP inhibition (19, 20). The MT1 receptor is localized in the hypothalamic SCN and hypophyseal pars tuberalis and is thought to mediate circadian and reproductive responses to melatonin. MT2 is localized in the SCN and retina and is thought to mediate melatonin’s effects on circadian rhythms and retinal physiology (21). It is therefore possible that melatonin’s effect on body mass regulation is mediated through activation of central receptors, resulting in changes in metabolic rate via sympathetic nervous activity or altered feeding behavior. However, a direct effect of melatonin on peripheral tissues, such as adipose tissue, is also possible. Indeed, there is now a substantial amount of evidence supporting the expression of 2-[¹²⁵I]iodomelatonin-binding sites in several peripheral tissues, including lymphocytes, arteries, kidneys, and gastrointestinal tract (22 ; see Ref. 23 for review). However, in most cases, the subtype of melatonin receptor and the functional effects of melatonin in these tissues remain unclear.

In the present study we investigated whether there are functional melatonin receptors present in human adipocytes. The presence of these receptors would suggest that melatonin could act directly at adipose tissue and regulate its physiology. We used as a model, PAZ6 adipocytes, an immortalized human brown preadipocyte cell line, which retains the ability to differentiate into mature adipocytes and to express adipocyte-specific marker genes (24).

Materials and Methods

Culture and differentiation of PAZ6 cells

The immortalization of the human preadipocyte cell line PAZ6 has been described previously (24). All cell culture reagents and antibiotics were obtained from Life Technologies, Inc. (Gaithersburg, MD). Preadipocytes were cultivated in DMEM-Ham’s F-12 (1:1, vol/vol) containing Glutamax (Life Technologies, Inc.) supplemented with 15 mm HEPES, penicillin (100 U/ml), streptomycin (0.1 mg/ml), and 5% FCS (medium 1) in an atmosphere of 92.5% air-7.5% CO₂. For differentiation, cells were grown in medium 1 until confluent and then for an additional 14 d in medium 1 supplemented with 33μ m biotin, 17 μm pantothenate, 1 nm T₃, 500 nm insulin, 1μ m pioglitazone, and 100 nm dexamethasone. The medium also contained ascorbate (100μ m) and 3-isobutyl-1-methylxanthine (IBMX; 0.25 mm) for the initial 4 d of differentiation.

RT-PCR

Total RNAs were extracted from PAZ6 preadipocytes, PAZ6 adipocytes, and BAT from three patients with pheochromocytoma and from white adipose tissue (WAT) from two patients undergoing surgery to remove mammary fat. RT-PCR reactions were performed as described previously (24, 25). Polyadenylated[ poly(A)⁺] mRNA from 63 pooled human hypothalami was obtained from CLONTECH Laboratories, Inc. (Palo Alto, CA). Analysis of mRNA levels was performed during the exponential phase of the amplification, which was assessed in preliminary experiments for each pair of primers. Briefly, 100 ng reverse transcribed RNA were used in a PCR reaction where the number of cycles was gradually increased (usually from 20 to 37 cycles). The products of the reaction were visualized on a gel by ethidium bromide staining and analyzed by densitometric scanning. The exponential phase of the reaction was determined by plotting, on a logarithmic scale, the intensity of the signal against the number of amplification cycles. During this phase, when the reaction was performed with variable amounts of reverse transcribed mRNA (from 50–200 ng), the signal obtained was proportional to the amount of cDNA used in the reaction. The same procedure was used for the poly(A)⁺ mRNA from human hypothalamus. Samples containing total mRNA and poly(A)⁺ mRNA were normalized to the same amount of the cyclophilin amplification product. Primers for the following cDNAs were designed with the program Oligo4, using the following GenBank entries: MT1 receptor (accession no. U14108): sense, 5′-TCAACCGCTACTGCTACATC-3′, 5′-annealing position 400; antisense, 5′-TCATCAGTGGAGACGGTTTC-3′, 5′-annealing position 1031; MT2 receptor (accession no. U25341): sense, 5′-TCATCGGCTCTGTCTTCAATA-3′, 5′-annealing position 383; antisense, 5′-ACTGGGTGCTGGCGGTCTGGA-3′, 5′-annealing position 611; Mel1c receptor (accession no. U67882): sense, 5′-CTTCAACATAACAGCCATAGC-3′, 5′-annealing position 360; antisense, 5′-TGCTTGATTGTTGTTGGTTAC-3′, 5′-annealing position 1051; melatonin receptor-related gene (accession no. U52219): sense, 5′-GAAGGAGATGGCAGGCAAGA-3′, 5′-annealing position 861; antisense, 5′-TGGTGGGTAGAGGCAGATTT-3′, 5′-annealing position 1281; PPARγ (accession no. L40904): sense, 5′-AGACAACAGACAAATCACCAT-3′, 5′ annealing position 894; antisense 5′-CTTCACAGCAAACTCAAACTT-3′, 5′-annealing position 1294; cyclophilin (accession no. Y00052): sense, 5′-AGCACTGGAGAGAAAGGATT-3′, 5′ annealing position 132; antisense 5′-GGAGGGAACAAGGAAAACAT-3′, 5′-annealing position 659; Glut 4 (accession no. M20747): sense, 5′-TCCTGCTGCCCTTCTGTC, 5′-annealing position 653; antisense, 5′-GGCCTACCCCTGCTGTCT, 5′-annealing position 961; leptin (accession no. U43653): sense, 5′-GCTGTGCCCATCCAAAAAGT, 5′-annealing position 61; antisense, 5′-ACTGCCAGTGTCTGGTCCAT, 5′-annealing position 242; uncoupling protein 1 (UCP1; accession no. XM011103): sense, 5′-TTAGGAAGCAAGATTTTAGC-3′, 5′-annealing position 337; antisense, 5′-AAGTCGCAAGAAGGAAGGTA-3′, 5′-annealing position 835; UCP2 (accession no. U76367): sense, 5′-TGTGCTGAGCTGGTGACCTATGAC-3′, 5′-annealing position 571; antisense, 5′-AAGGGAGCCTCTCGGGAAGTGCAG-3′, 5′-annealing position 926; lipoprotein lipase (LPL; accession no. M76722): sense, 5′-GAGATTTCTCTGTATGGCACC-3′, 5′-annealing position 267; antisense, 5′-CTGCAAATGAGACACTTTCTC-3′, 5′-annealing position 391; and TNFα (accession no. X01394): sense, 5′-CAGAGGGAAGAGTTCCCCAG, 5′-annealing position 327; antisense, 5′-CCTTGGTCTGGTAGGAGACG, 5′-annealing position 651.

Radioligand binding assays and competition experiments

Binding assays were performed on PAZ6 adipocytes after 14 d of differentiation. The medium was aspirated and replaced by 0.25 ml binding buffer [DMEM-Ham’s F-12 (1:1 vol/vol) supplemented with 15 mm HEPES and 0.5% fatty acid-free BSA], containing 16–1200 pm¹²⁵I-Mel (DuPont-NEN Life Science Products, Boston, MA) as radioligand. Specific binding was defined as binding displaced by 1μ m melatonin (Sigma, St. Louis, MO). Culture plates were incubated at 25 C for 90 min, and the reaction was terminated by transferring plates to ice for 10 min. The medium was aspirated, and each well was rinsed twice with 1 ml ice-cold PBS. Cells were solubilized with 0.5 ml NaOH (1 m) and transferred to Eppendorf tubes, and radioactivity was counted with aγ -counter. Competition experiments were performed as described above with approximately 300 pm¹²⁵I-Mel and varying concentrations of cold ligands: melatonin, 5-methoxycarbonylamino-N-acetyltryptamine (GR 135,531), and 4-phenyl-2-propionamidotetraline (4-P-PDOT) (Tocris Cookson Ltd., Bristol, UK).

Determination of intracellular cAMP levels

PAZ6 cells differentiated in six-well dishes were incubated for 15 min at 37 C in medium 1 with or without forskolin (1μ m Sigma) and increasing concentrations of melatonin. The incubation buffer was discarded, and cells were extracted in 0.5 ml ice-cold 65% ethanol. Cell extracts were centrifuged in a microcentrifuge at 17,000 × g for 10 min. The supernatant was concentrated in a Speed-Vac (Savant, Farmingdale, NY), and the pellet was diluted in 10 mm Tris (pH 7.4) and 1 mm EDTA. cAMP concentrations were determined using a[ ³H]cAMP assay system (Amersham Pharmacia Biotech, Arlington Heights, IL).

Determination of intracellular cGMP levels

PAZ6 cells differentiated in six-well dishes were incubated for 15 min at 37 C in medium 1 with IBMX (1 mm) and increasing concentrations of melatonin. The medium was then discarded, and cells were extracted in 0.5 ml ice-cold 65% ethanol. Cell extracts were centrifuged at 20,000 × g for 15 min at 4 C. Supernatants were dried using a Speed-Vac, and pellets were resuspended in 0.25 ml assay buffer, acetylated, and assayed for cGMP, according to the instructions of the manufacturer of the enzyme immunoassay kit (Amersham Pharmacia Biotech).

Membrane preparation, SDS-PAGE, and immunoblotting

Crude membranes were prepared as described recently (26) and denatured in 25 mm Tris (pH 6.8), 4 m urea, 180 mm dithiothreitol, 2.5% SDS, 5% glycerol, and 0.05% bromophenol blue overnight at room temperature. Proteins were separated by 10.5% SDS-PAGE and transferred to nitrocellulose. Immunoblot analysis was carried out with a rabbit anti-Glut4-specific antibody (1:10,000). Immunoreactivity was revealed using appropriate secondary antibodies coupled to horseradish peroxidase and the ECL chemiluminescent reagent (Amersham Pharmacia Biotech).

Glucose uptake

PAZ6 cells were differentiated in the differentiation medium in the presence or absence of 10 nm melatonin. Twenty-four hours before the experiment cells were incubated in differentiation medium without insulin and with 2% FCS. Cells were then washed twice with KREBS puffer (supplemented with 2% fatty acid-free BSA, 1 mm pyruvate, and 0.1 mm glucose) and incubated in this buffer for 2 h at 37 C. Stimulation was performed in the same buffer supplemented with 0.5 μCi/ml 2-[1,2-³H]deoxy-d-glucose (Amersham Pharmacia Biotech) at 37 C for 5 min in the presence or absence of 200 nm insulin. Cells were washed twice with ice-cold PBS and counted in a β-counter.

Results

Expression of melatonin receptors in PAZ6 adipocytes and human BAT and WAT

We recently developed a human brown preadipocyte cell line, PAZ6, which has been shown to be a valuable model of human adipocytes (24). These cells express upon differentiation several adipocyte specific markers, such as UCP1, LPL, and hormone-sensitive lipase (24, 25). The β-adrenergic response and its regulation upon agonist stimulation are fully maintained in these cells (26).

Semiquantitative RT-PCR experiments were used to study the expression of melatonin receptor mRNA in PAZ6 preadipocytes, PAZ6 adipocytes, and human BAT, WAT, and hypothalamic tissue (Fig. 1). BAT was obtained from patients with pheochromocytoma. In this disease, BAT hyperplasia develops in fat tissue surrounding the tumor (27). WAT was obtained from patients undergoing surgery to remove mammary fat. Poly(A)⁺ mRNA pooled from different human hypothalami was used as a reference for melatonin receptor expression. Amplification products corresponding to both receptor subtypes were detected in PAZ6 adipocytes as well as in WAT, BAT, and hypothalamus (Fig. 1). Expression of the MT1 receptor transcript was most abundant in the hypothalamus and much weaker in BAT and PAZ6 adipocytes, which expressed equivalent amounts. WAT expressed less MT1 transcript than BAT. The MT2 transcript was most abundant in BAT, slightly less abundant in hypothalamus and PAZ6 adipocytes, and barely detectable in WAT. MT1 receptor transcripts were weakly expressed in PAZ6 preadipocytes, whereas MT2 transcripts were undetectable. Expression of both subtypes clearly increased with adipocyte differentiation. Melatonin receptor transcripts were not detected in other human peripheral tissues tested (myometrium, gallbladder, and liver; data not shown), supporting the specificity of melatonin receptor expression in human adipose tissue. Primers specific for two additional members of the melatonin receptor family, the Xenopus Mel1c receptor and the human melatonin-related receptor, failed to amplify the corresponding or closely related amplification products (data not shown). These data show that both melatonin receptor subtypes are expressed in human adipocytes at higher levels in BAT compared with WAT. Expression levels of the MT2 transcript in BAT are comparable to those in the hypothalamus, the principal expression site of central melatonin receptors. Comparable results were obtained in PAZ6 adipocytes and BAT, suggesting that PAZ6 adipocytes are an appropriate model for studying the role of melatonin receptors in human fat tissue.

Fig. 1

Expression of melatonin receptor genes in human adipose tissue. Total mRNA was isolated from BAT, WAT, PAZ6 preadipocytes (PA), and PAZ6 adipocytes (A) and analyzed by semiquantitative RT-PCR. Poly(A)⁺ mRNA from human hypothalamus (Hyp) was used as a reference for melatonin receptor expression. PCR products were visualized on agarose gels by ethidium bromide staining. As an internal control, expression of the cyclophilin (cyclo) gene was tested. No amplification was observed when experiments were performed in the absence of reverse transcriptase (not shown). Cloned melatonin receptor cDNA (1fg) was used as a positive control for specific amplification (Cont.). The results are representative of three independent experiments.

¹²⁵I-Mel-binding sites in PAZ6 adipocytes

Saturation binding studies were performed with increasing concentrations of ¹²⁵I-Mel (15–1200 pm) to verify the expression of melatonin receptor protein in PAZ6 adipocytes. Scatchard analysis revealed a single high affinity binding site with a K_d of 457 ± 5 pm. Binding was found to be saturable with a binding capacity of 7.46 ± 1.58 fmol/mg protein (Fig. 2A). ¹²⁵I-Mel binding to PAZ6 adipocytes was further characterized in competition experiments. Melatonin displaced ¹²⁵I-Mel binding with a K_i of approximately 3 × 10⁻¹¹m (Fig. 2B), confirming the expression of high affinity melatonin receptors. Recently, an atypical ¹²⁵I-Mel-binding site was identified in Siberian hamster brown adipocytes (28). A key feature of this binding site was its sensitivity to GR135531, a compound that has no affinity for MT1 and MT2 receptors (28). In PAZ6 adipocytes ¹²⁵I-Mel binding was not competed by GR135531 at concentrations of up to 1 μm, suggesting that this binding site is different from that described in Siberian hamster brown adipocytes. To discriminate between MT1 and MT2 receptors we used the melatonin receptor antagonist 4P-PDOT, which has a more than 1000 times higher affinity for MT2 receptors than for MT1 receptors (29). The competition curve for 4P-PDOT was monophasic with a K_i of approximately 1.5 × 10⁻¹¹m, which is in good agreement with values reported for human MT2 receptors expressed in CHO cells (29). Combined results from both molecular and pharmacological analyses, support the presence of high affinity melatonin receptors in PAZ6 adipocytes, predominantly corresponding to the MT2 subtype.

Fig. 2

Pharmacological specificity of ¹²⁵I-Mel binding in PAZ6 adipocytes. A, Saturation isotherm of ¹²⁵I-Mel binding; nonspecific binding was measured in the presence of 1 μm melatonin. Inset, Scatchard representation of the data. B, Competition binding of ¹²⁵I-Mel (300 pm) on PAZ6 adipocytes, with GR135531 (○), melatonin (▪), and 4-P-PDOT (▵). Data represent the mean values of a single experiment, performed in triplicate, that is representative of three independent experiments. Data were fitted using the GraphPad Plot program version 4.04 (San Diego, CA).

Melatonin inhibits forskolin-stimulated cAMP accumulation in PAZ6 adipocytes

We next verified whether melatonin receptors were functionally coupled to signal transduction pathways in PAZ6 cells. A pathway common to all cloned, high affinity melatonin receptors is the inhibition of forskolin-stimulated cAMP accumulation. Incubation of PAZ6 adipocytes with 1 μm forskolin increased intracellular cAMP concentrations approximately 10-fold. This increase was inhibited by melatonin in a dose-dependent manner, with an IC₅₀ of 2 × 10⁻⁹m and maximal inhibition levels of approximately 20% (Fig. 3). Melatonin alone had no effect on basal cAMP levels.

Fig. 3

Modulation of forskolin-stimulated cAMP accumulation by melatonin in PAZ6 adipocytes. PAZ6 adipocytes were stimulated for 15 min at 37 C with forskolin (1 μm) in the presence of the indicated concentrations of melatonin; intracellular cAMP levels were determined as described in Materials andMethods. The 100% value corresponds to the mean cAMP value in the presence of 1μ m forskolin. Data are the mean ± se of three independent experiments performed in triplicate (*, P < 0.05). Data were fitted using the GraphPad Plot program version 4.04.

Inhibition of intracellular cGMP levels by melatonin in PAZ6 adipocytes

Recently, we showed that MT2, but not MT1, melatonin receptors modulate intracellular cGMP levels when transfected into HEK 293 cells (20). The inhibitory effect of melatonin on cGMP levels was observable when cGMP degradation was blocked by IBMX, a nonspecific inhibitor of phosphodiesterases. Incubation of PAZ6 adipocytes with IBMX indeed induced a 3-fold increase in basal cGMP levels, which was inhibited by melatonin in a dose-dependent manner with an IC₅₀ value of approximately 3 × 10⁻¹⁰m and a maximal inhibition level of 60% (Fig. 4). Both parameters are in good agreement with those previously observed for the cloned human MT2 receptor (20) and support functional expression of this receptor subtype in PAZ6 adipocytes.

Fig. 4

Modulation of cGMP accumulation by melatonin receptors. PAZ6 adipocytes were incubated for 15 min at 37 C with the indicated concentrations of melatonin in the presence of 1 mm IBMX. Intracellular cGMP levels were determined as described in Materials and Methods. Data are presented as a percentage of the IBMX-stimulated basal level (IBMX stimulation was 3-fold, basal cGMP levels in the absence of IBMX were 20–40 fmol/mg protein). Data are the mean ± se of three independent experiments performed in triplicate. Data were fitted using the GraphPad Plot program version 4.04.

Effect of melatonin on the expression of adipocyte markers in PAZ6 adipocytes

The effect of melatonin treatment on the expression of several adipocyte-specific marker genes was tested by semiquantitative RT-PCR in PAZ6 adipocytes. Cells were treated either throughout the entire differentiation process or for 2 d in the fully differentiated state. Melatonin markedly inhibited Glut4 expression in both experimental protocols (Fig. 5). In contrast, melatonin had no significant effect on the expression of genes encoding for other adipocyte markers: uncoupling proteins UCP1 and UCP2, PPARγ, TNFα, leptin (ob), LPL (Fig. 5), and hormone-sensitive lipase (not shown). Melatonin also had no effect on the amount of leptin protein secreted into the culture medium, confirming the RT-PCR data (not shown). These results demonstrate that melatonin can specifically modify the expression of Glut4, a gene important for adipocyte homeostasis.

Fig. 5

Effect of melatonin on the expression of adipocyte markers in PAZ6 adipocytes. PAZ6 adipocytes were cultured in the presence of melatonin (1 μm) during the differentiation process (14 d) or for 2 d after differentiation. For each experimental condition, duplicate culture wells were extracted independently, and semiquantitative RT-PCR was performed on total RNA from each culture well. PCR products were made visible on agarose gels by ethidium bromide staining. As a control, the signals for cyclophilin (cyclo) expression were determined. Results are representative of three independent experiments.

Melatonin decreases Glut4 protein levels and glucose uptake in PAZ6 adipocytes

To verify whether the effect of melatonin on Glut4 mRNA levels has functional consequences we measured Glut4 protein levels and glucose uptake in PAZ6 adipocytes. Cells were treated with 10 nm melatonin either throughout the entire differentiation process or for 1 d in the fully differentiated state (Fig. 6). One-day melatonin treatment did not change either Glut4 protein levels or glucose uptake significantly, whereas long-term melatonin treatment markedly decreased both parameters.

Fig. 6

Melatonin decreases Glut4 protein levels and glucose uptake in PAZ6 adipocytes. PAZ6 adipocytes were cultured in the presence of melatonin (10 nm) during the differentiation process (14 d) or for 1 d after differentiation. A, Crude membranes were prepared and submitted to SDS-PAGE. Immunoblot analysis was performed with a polyclonal anti-Glut4 specific antibody. B, Cells were incubated for 24 h in differentiation medium with 2% FCS and without insulin. 2-[1,2-³H]Deoxy-d-glucose uptake was measured at 37 C for 5 min in the presence or absence of 200 nm insulin as outlined in Materials and Methods. Results are representative of two additional experiments and are presented as a percentage of the control value in the presence of insulin.

Discussion

Several years ago it was observed that variations in melatonin secretion pattern could induce body weight changes in mammals (2, 4). To date, however, only a limited number of studies have attempted to address the underlying mechanisms. Recent cloning of melatonin receptors revealed the existence of two melatonin receptor subtypes in mammals, MT1 and MT2 (16, 17). These receptors are expressed at several peripheral and central sites, including the hypothalamus. Activation of central receptors may result in changes in metabolic rate via sympathetic nervous activity or altered feeding behavior. Activation of peripheral receptors, for example in adipose tissue, may influence energy storage by modulating adipocyte metabolism or proliferation. Results from the present report suggest that MT2 receptors expressed in human adipose tissue could participate in adipocyte homeostasis.

Several lines of evidence support functional expression of the MT2 receptor subtype in human adipocytes. MT2 receptor mRNA has been detected in PAZ6 adipocytes, BAT, and WAT, but not in several other peripheral tissues. The quantity of MT2 receptor mRNA in BAT is comparable to that in human hypothalamus, a well established expression site of melatonin receptors. Pharmacological analysis indicated the expression of a single high affinity ¹²⁵I-Mel-binding site in PAZ6 adipocytes. The K_d of 450 pm for ¹²⁵I-Mel binding fits best with expression of the MT2 subtype, as K_d values are typically lower for MT1 (20–150 nm) than for MT2 receptors (100–500 nm) (18, 21, 30). The binding capacity of 7.5 fmol/mg protein corresponds to receptor densities reported for high affinity melatonin receptors identified at other peripheral sites, such as lymphocytes and prostate epithelial cells (23, 31). Importantly, ¹²⁵I-Mel binding in PAZ6 adipocytes was competed by the MT2-selective melatonin receptor ligand 4P-PDOT with high affinity. Furthermore, melatonin inhibited forskolin-stimulated cAMP levels with an IC₅₀ within the range of physiological melatonin concentrations (32). Maximal inhibition levels were weak (20%). However, considerable variation has been observed for this parameter (10–80%) in transfected cells as well as in cells expressing endogenous receptors (1, 20, 23, 33), reflecting the variable importance of this pathway depending on the cellular background. Importantly, melatonin also modulated a further second messenger, cGMP, a phenomenon that has been shown to be selective for the MT2 receptor subtype in humans (20). The IC₅₀ value for cGMP modulation was within the range of circulating melatonin levels. Maximal inhibition levels (60%) fitted well with those observed in HEK293 cells transfected with the human MT2 receptor (20).

Expression of ¹²⁵I-Mel-binding sites has previously been reported in BAT of Siberian hamsters, a species known to exhibit dramatic body weight changes in response to changing photoperiod (28). Pharmacological analysis suggested expression of an atypical binding site, different from cloned melatonin receptors and different from pharmacologically characterized ¹²⁵I-Mel-binding sites in hamster brain (30). In agreement with this observation, melatonin did not inhibit cAMP accumulation, and no MT1 receptor transcript was detected in these adipocytes. Expression of the MT2 subtype could be ruled out, because Siberian hamsters have been reported to be natural knockouts for this receptor (34). Taken together, these results suggest that melatonin receptors expressed in BAT are different in humans and Siberian hamsters.

Identification of functional MT2 receptors in human adipocytes raises the question of the physiological role of melatonin in this tissue. To address that question, we investigated whether treatment of PAZ6 adipocytes with melatonin could alter the expression of key genes involved in adipocyte homeostasis. Long-term melatonin treatment of PAZ6 adipocytes markedly decreased Glut4 mRNA levels. This response appears to be specific, as the expression of several other adipocyte marker genes (UCP1, UCP2, PPARγ, TNFα, leptin, hormone-sensitive lipase, and LPL) was not altered by melatonin. The functional relevance of melatonin’s effect on Glut4 mRNA levels was confirmed by measurements showing that long-term melatonin treatment reduces Glut4 protein levels and glucose uptake in PAZ6 adipocytes.

Recently, melatonin has been shown to regulate the expression of another gene in BAT. Treatment of isolated hamster adipocytes with melatonin specifically induced the mitochondrially encoded cytochrome b mRNA (35). The functional importance of this effect is to date unknown.

Glut4 plays a predominant role in insulin-mediated glucose transport. Transgenic mice bearing only one allele of the Glut4 gene were shown to express reduced Glut4 protein levels, resulting in a progressive diabetic phenotype, characterized by impaired glucose homeostasis (36). Adipocyte-selective targeting of the Glut4 gene impaired insulin action in muscle and liver, demonstrating that the expression level of Glut4 in adipose tissue has major effects not only on the adipose tissue but also on the glucose metabolism of the whole body (37). Overexpression of Glut4 in adipose tissue leads to a selective increase in fat cell number and a 2- to 3-fold increase in total body lipids. Changes in Glut4 gene expression have been observed under different physiological and pathological situations. In general, Glut4 mRNA expression is down-regulated in states of relative insulin deficiency, such as streptozotocin-induced diabetes and chronic fasting. In diabetic rodents and humans, Glut4 expression is suppressed in adipocytes in association with insulin resistance. In contrast, exercise has been shown to cause a 1.5-fold increase in Glut4 mRNA expression (38).

Melatonin decreased Glut4 mRNA levels in PAZ6 adipocytes when applied during the whole differentiation process. Two important transcription factors for adipocyte differentiation, PPARγ and C/EBPα, are required for expression of the Glut4 gene during adipocyte differentiation (39). Our results suggest that melatonin has a dominant effect on the increase of Glut4 mRNA levels during differentiation, even in the presence of the differentiation cocktail. A direct effect of melatonin on PPARγ expression can be excluded, because PPARγ mRNA levels were not modified by melatonin treatment.

Incubation of fully differentiated PAZ6 adipocytes with melatonin also decreases Glut4 mRNA levels. Several molecules, such as insulin, cAMP, TNFα, and arachidonic acid, were shown to decrease Glut4 mRNA levels in differentiated adipocytes (40–43). As melatonin receptors are negatively coupled to the adenylyl cyclase system, cAMP is unlikely to be involved in the effect of melatonin on Glut4 mRNA levels. TNFα was shown to decrease Glut4 mRNA levels in PAZ6 adipocytes (44). However, melatonin treatment did not change TNFα mRNA levels, arguing against the involvement of this mechanism. A 48-h treatment with arachidonic acid was shown to suppress Glut4 mRNA accumulation in differentiated 3T3-L1 adipocytes by several different mechanisms (42). Melatonin receptors stimulate the inositol phosphate pathway and potentialize arachidonic acid release in transfected NIH-3T3 cells (45). Thus, a plausible scenario may involve the stimulation of arachidonic acid production by melatonin in PAZ6 adipocytes, followed by a decrease in Glut4 mRNA levels. Further possibilities include new, but still poorly characterized, transcription factors that have been reported to specifically regulate the human Glut4 promoter in transgenic mice (46).

In conclusion, we have shown that melatonin receptors of the MT2 subtype are functionally expressed in human adipocytes. PAZ6 adipocytes constitute the first human cell line expressing endogenous MT2 receptors. Additional studies will be necessary to further define the physiological role of melatonin receptors and their spacio-temporal regulation in human adipocytes.

Acknowledgments

We thank V. Zilberfarb, K. Pernet, and Dr. Issad for expert advice on RT-PCR experiments; L. Camoin for leptin protein determinations; J.-L. Guillaume for help with the Western blotting experiments; and Dr. Couraud for support. We also thank Dr. Duclos (Hôpital St. Joseph, Paris, France) for providing us with brown adipose tissue from patients suffering from pheochromocytoma, and Drs. Cherif-Zahar and Léandri (Clinique Duhesme, Paris, France) in collaboration with Dr. Pietri-Rouxel for providing us with WAT. The anti-Glut4-specific antibody was kindly provided by Dr. Guerre-Millau (Paris, France).

This work was supported by grants from the Association pour la Recherche sur le Cancer (no. 9890 and 5513), Centre National de la Recherche Scientifique, INSERM, Institut de Recherches Internationales Servier, and the Université de Paris VII.

Abbreviations

 
• BAT

  Brown adipose tissue

 
• IBMX

  3-isobutyl-1-methylxanthine

 
• LPL

  lipoprotein lipase

 
• ¹²⁵ I-Mel

  2-[¹²⁵I]iodomelatonin

 
• 4P-PDOT

  4-phenyl-2-propionamidotetraline

 
• poly(A)⁺

  polyadenylated

 
• SCN

  suprachiasmatic nuclei

 
• UCP1

  uncoupling protein 1

 
• WAT

  white adipose tissue