Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Octopamine shares structural and functional homologies with norepinephrine. First, octopamine is the monohydroxylated analog of norepinephrine. Second, both are considered as stress neurohormones, with a noradrenergic system operative in vertebrates only, and a functional octopaminergic system limited to invertebrates, which is especially well documented in insects. Third, the receptors to these transmitters are G protein-coupled receptors sharing some biochemical and pharmacological similarities. Finally, both biogenic amines prepare the animal to "fight or flight," as reviewed by Roeder (1999). Although vertebrates are devoid of the four types of octopamine receptors identified so far, they contain trace amounts of octopamine in plasma (Andrew et al., 1993; Martina et al., 1998), peripheral tissues (Murphy et al., 1975), or central nervous system (Saavedra, 1988). A better understanding of the effects of octopamine in mammals appears relevant regarding the putative toxicity of octopaminergic drugs used as insecticides (Evans and Gee, 1980; Nathanson et al., 1993; Roeder, 1999) and undesirable side effects of medicinal plant extracts rich in octopamine and synephrine that are claimed to be useful as weight-lowering products (Calapai et al., 1999).

When administered to rats, octopamine induces diverse responses, such as changes in locomotor activity and decrease in blood pressure (Delbarre et al., 1982), which have been attributed to activation of adrenergic or dopaminergic systems, depending on the dose and mode of administration. Direct interaction of octopamine with mammalian adrenergic receptors (ARs) has been assessed by using cellular models. In Chinese hamster ovary cells transfected with human _2A-, _2B-, or _2C-ARs, a selective inhibition of adenylyl cyclase was observed with meta-octopamine via an agonism at _2A-ARs (Airriess et al., 1997), whereas the amine coupled the _2B- or _2C-ARs to both activation and inhibition of cAMP accumulation (Rudling et al., 2000). In contrast, no clear ₂-adrenergic agonism was observed for para-octopamine in fat cells from diverse mammals that endogenously express different amounts of adipocyte ₂-AR subtypes (Fontana et al., 2000). Moreover, we observed that octopamine was able to stimulate adipocyte ₃-ARs (Carpéné et al., 1999; Fontana et al., 2000). In fact, octopamine fully stimulates lipolysis in rodent fat cells, with the same maximal effect as norepinephrine but with a potency that is two orders of magnitude lower (Carpéné et al., 1999), whereas in human adipocytes, poorly responsive to ₃-AR agonists (Van Liefde et al., 1994), octopamine does not induce full lipolytic stimulation (Carpéné et al., 1999). In mammals, ₃-ARs are known to be positively coupled to Gs protein and to activate adenylyl cyclase, like ₁- and ₂-ARs (thereby stimulating lipolysis), but ₃-ARs can also be coupled to other effectors, leading to negative inotropic effect in the human heart (Gauthier et al., 1998) or to partial inhibition of insulin effects in rat adipocytes (Carpéné et al., 1993; Klein et al., 1999). Accordingly, it has been proposed that octopamine acts as a selective agonist at ₃-ARs because it counteracted the insulin action on glucose transport into rat adipocytes (Yen et al., 1998; Fontana et al., 2000). To further analyze the ₃-AR agonist properties of octopamine, the present study was aimed at comparing its effects on lipolysis and glucose transport with those of the ₃-AR agonists BRL 37344 and CL 316243 in mice genetically lacking functional ₃-ARs (Susulic et al., 1995). A putative stimulation of _2A-ARs by octopamine was also tested in transgenic mice designed to express human _2A-AR in adipose tissues (Valet et al., 2000).

The present pharmacological characterization of octopamine action also includes a study of its oxidation by monoamine oxidase (MAO) and semicarbazide-sensitive amine oxidase (SSAO) because it has been reported that amine oxidation elicits stimulation of glucose transport into rat (Marti et al., 1998) or human (Morin et al., 2001) adipocytes. Indeed, our recent findings have demonstrated that the deaminative oxidation catalyzed by MAO and SSAO present in adipocytes (Pizzinat et al., 1999; Morin et al., 2001) generates hydrogen peroxide, which exerts insulin-like actions such as stimulation of glucose uptake (Enrique-Tarancon et al., 2000). Octopamine, already described as a substrate of MAO (Youdim and Finberg, 1991) or SSAO (Castillo et al., 1999), was therefore tested for its oxidation-dependent effects, in comparison with two amines of reference: tyramine, an octopamine precursor that activates octopamine receptors in insects and that is a MAO substrate in mammals, and benzylamine, a SSAO substrate (Lyles, 1996).

Our comparative pharmacological approach demonstrates that, in the millimolar range, octopamine is either a ₃-AR agonist, an MAO, or an SSAO substrate. In all, our results show that octopamine exerts, with a low potency, dual action in mammalian fat cells: on one hand, it stimulates lipolysis and inhibits insulin-promoted glucose uptake via agonism at ₃-ARs, whereas, on the other hand, it stimulates glucose uptake via oxidation by MAO or SSAO. This latter action is predominant only when ₃-ARs are genetically inactivated or poorly efficient, as is the case for ₃-AR-deficient mice and human subcutaneous adipocytes, respectively.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. [¹⁴C]Tyramine (45.2 mCi/mmol) and 2-[1,2-³H]deoxyglucose (2-DG, 26 Ci/mmol) were purchased from PerkinElmer Life Science Products (Boston, MA). [¹⁴C]Benzylamine (57 mCi/mmol) was from Amersham Pharmacia Biotech (Les Ullis, France). Enzymes and cofactors for glycerol determination were from Roche Molecular Biochemicals (Mannheim, Germany). Collagenase, adenosine deaminase (ADA), cytochalasin B, 3-isobutyl-1-methylxanthine, bovine serum albumin, (±)-para-octopamine, and other chemicals were from Sigma-Aldrich (St. Louis, MO) unless otherwise stated. CL 316243, BRL 37344, and SR 59230 were kindly provided by Dr. T. H. Claus (American Cyanamid-Lederlé Laboratories, Pearl River, NY), Dr. M. A. Cawthorne (GlaxoSmithKline, Welwyn Garden City, Hertfordshire, UK), and Dr. L. Manara (Sanofi Research Center, Milano, Italy), respectively.

Sampling of Adipose Tissues in Subjects and Animals. Samples of human subcutaneous adipose tissue were obtained from 22 healthy nonobese women (body mass index was 26.5 ± 1.4 kg · m², mean age 42 ± 2 years) undergoing abdominal surgical lipectomy at Rangueil Hospital (Toulouse, France). The study was approved by the Ethical Committee of Toulouse University Hospital. Samples of adipose tissue were transferred in less than 15 min from the Department of Plastic Surgery to the laboratory.

Adipocyte Isolation. Adipose tissues were digested in Krebs-Ringer buffer containing 15 mM sodium bicarbonate, 10 mM HEPES, bovine serum albumin (3.5% w/v), and collagenase (1.5 mg/ml for rodents, 1 mg/ml for humans). After digestion for 35 to 45 min at 37°C, isolated fat cells were filtered and washed three times in the same buffer without collagenase. The cells were adjusted to a suitable dilution in the same buffer supplemented with either 2 mM pyruvate for hexose uptake assays or 5 mM glucose for lipolysis assays. All fat cell suspensions were incubated in plastic vials with the tested drugs in a final volume of 400 µl. The amount of fat cells present in the incubations was assessed by the weight of cell lipids, which was 22 ± 1, 16 ± 2, and 13 ± 1 mg of lipid/assay for human, rat, and murine adipocytes, respectively.

Hexose Transport and Lipolysis. Adipocyte suspensions were incubated at 37°C with the tested drugs during 45 min. Then an isotopic dilution of 2-DG (0.4 µCi) was added at a final concentration of 0.1 mM for an additional 10-min period. Assays were stopped with 100 µl of 100 µM cytochalasin B. Then 200-µl aliquots of the cell suspension were centrifuged in microtubes containing dinonyl phthalate, which allowed for separation of the adipocytes from the buffer and counting of the radioactive intracellular 2-DG as previously described (Morin et al., 2001). Extracellular 2-DG present in the cell fraction was subtracted from the experimental values as previously reported (Carpéné et al., 1993). Lipolysis was determined as already described (Carpéné et al., 1999) by measuring the glycerol released from cell suspensions after an incubation of 60 or 90 min for rodent and human adipocytes, respectively.

Amine Oxidase Activity. Oxidase activity was measured using [¹⁴C]tyramine or [¹⁴C]benzylamine according to the radiochemical method described by Fowler and Tipton (1981). MAO activity was defined as the part of oxidation inhibited by 15-min preincubation with 0.5 mM pargyline, whereas the activity inhibited by 1 mM semicarbazide was ascribed to SSAO. Simultaneous addition of pargyline plus semicarbazide abolished [¹⁴C]tyramine or [¹⁴C]benzylamine oxidation in all the species studied as already reported (Marti et al., 1998; Pizzinat et al., 1999). Briefly, thawed samples of SCWAT were homogenized in 200 mM phosphate buffer, pH 7.4, containing an antiprotease cocktail from Sigma-Aldrich. Homogenates (300 µg of protein/100 µl) were incubated for 15 (human) or 30 min (rodent) at 37°C in 200 µl of phosphate buffer in the presence of 0.5 mM [¹⁴C]tyramine (0.05 µCi) or 0.1 mM [¹⁴C]benzylamine (0.1 µCi), unless otherwise stated. Assays were stopped by adding 50 µl of 4 M HCl. Reaction products were extracted by addition of 1 ml of solvent (toluene/ethyl acetate, v/v). Then 0.7-ml aliquots of the organic phase were transferred into scintillation vials and counted for radioactivity.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Differential Effects of Benzylamine, Tyramine, and Octopamine on Glucose Transport and Amine Oxidase Activity in Rat Adipocytes. As previously reported (Enrique-Tarancon et al., 1998, 2000; Marti et al., 1998), vanadate (0.1 mM) was necessary to observe a clear-cut stimulatory action of a millimolar dose of benzylamine or tyramine on basal glucose transport in rat adipocytes (5.7 ± 0.4 and 4.3 ± 0.8, respectively, versus 1.1 ± 0.3 nmol of 2-DG taken up/100 mg of lipid/10 min, n = 8, p < 0.001). Under these conditions, vanadate did not allow octopamine to stimulate 2-DG uptake (1.1 ± 0.2 nmol/100 mg of lipid/10 min). However, octopamine can be considered as an adequate substrate for the amine oxidases present in rat white adipocytes because, in competition studies of [¹⁴C]tyramine oxidation, its apparent affinity for SSAO activity was comprised between those of benzylamine and tyramine, whereas its affinity toward MAO was less than 7-fold lower than the reference amines (data not shown). The apparent discrepancy between the capacity of octopamine to be oxidized and its lack of vanadate-dependent stimulation of glucose uptake could have been explained by a balance between its putative insulin-like effect and its ₃-adrenergic inhibitory action on glucose uptake, already described in rat adipocytes (Yen et al., 1998; Fontana et al., 2000). Because selective and complete blockade of the ₃-ARs remains difficult to obtain on rat adipocytes with the available ₃-AR-antagonists, we further studied the effects of octopamine in fat cells from mice with genetically disabled ₃-ARs.

Adiposity and Lipolytic Activities of White Adipocytes in Control, ₃ /, and ₂-trans Mice. The ₃-AR knockout mice (₃ /) (Susulic et al., 1995) together with the transgenic mice created on the ₃-AR / background and expressing the human _2A-AR in their adipose tissues (₂-trans) (Valet et al., 2000) were used as valuable tools to delineate the dual action of octopamine as amine oxidase substrate or as ₃- and/or ₂-adrenergic agent. Even though the wild-type FVB/n control and the genetically modified female mice shared nearly identical body mass (24-26 g when 11 weeks old), the adiposity was increased in ₃-AR-deficient animals, as previously reported (Susulic et al., 1995): the INWAT mass was moderately but significantly higher in both ₃ / and ₂-trans mice than in control (0.72 ± 0.06 and 1.00 ± 0.10 versus 0.53 ± 0.06 g, respectively; n = 15, p < 0.05).

View this table: [in this window] [in a new window]   TABLE 1 Lipolytic responses of adipocytes from control, transgenic mice lacking functional 3-AR and/or expressing human 2A-AR Assays were carried out without any addition (basal lipolysis) or with the lipolytic agents at the indicated final concentration. Means ± S.E.M. of six determinations.

Regulation of Glucose Transport by Insulin and Amines in Control, ₃ /, and ₂-trans Mice. Millimolar concentrations of tyramine or benzylamine activated glucose uptake in mouse fat cells in the presence of 0.1 mM vanadate (Fig. 1). Vanadate, which was without noticeable effect on insulin-stimulated uptake, moderately activated basal glucose uptake, reaching only 10 to 20% of maximal insulin regardless of the group studied. In control mice, the synergism between vanadate and tyramine or benzylamine, which allowed both amines to stimulate 2-DG transport up to one-half the maximal effect of insulin, did not apply for octopamine, which remained as inefficient as in rat adipocytes (Fig. 1). However, octopamine plus vanadate clearly promoted glucose transport in adipocytes from both ₃ / and ₂-trans mice. Octopamine alone was inefficient in control but showed a tendency to stimulate glucose uptake in ₃-AR-deficient mice (data not shown). Because the appearance of an insulin-like effect of octopamine was dependent on both the presence of vanadate and the invalidation of ₃-ARs, whereas the presence of ₂-ARs did not clearly influence this phenomenon, we focused our study on the balance between ₃-AR agonism and the amine-oxidase-substrate properties of octopamine.

View larger version (31K): [in this window] [in a new window]   Fig. 1.   Comparison of the effects of insulin, benzylamine, tyramine, and octopamine on glucose transport in the presence of vanadate in mouse adipocytes: influence of genetic alterations of the 3-AR and 2-AR expression. Isolated adipocytes were incubated during 45 min with 0.1 mM vanadate alone (no addition) or in combination with 0.1 µM insulin, 0.1 mM benzylamine, 1 mM tyramine, or 1 mM octopamine before exposure to 0.1 mM [3H]2-DG during 10 min. Results are expressed as percentage of insulin effect on hexose transport with 100% corresponding to the 2-DG uptake stimulated by 0.1 µM insulin alone (6.2 ± 0.7, 4.8 ± 0.8, and 3.4 ± 0.4 nmol/100 mg of lipid/10 min) in control (), 3 / (), and 2-trans mice (), respectively. Mean ± S.E.M. of six experiments. ***p < 0.001 versus control mice.

Comparison of Actions of Octopamine, ₃-AR Agonists, and Tyramine or Benzylamine in Adipocytes from Control and  ₃ / Mice. The lipolytic actions of octopamine and ₃-AR agonists (CL 316243 and BRL 37344) were expressed relative to the maximal lipolysis obtained with 10 µM isoproterenol because this expression of the results allows comparison of intrinsic activities of -adrenergic agents. In control mice, isoproterenol, octopamine, and CL 316243 or BRL 37344 reached the same maximal lipolytic activity. In ₃ / mice, the responses to octopamine, CL 316243, or BRL 37344 were blunted and did not exceed one-third that of isoproterenol (Fig. 2). These data confirmed that, in murine adipocytes, the octopamine lipolysis was mainly dependent on ₃-AR-activation, whereas the effect of isoproterenol was less hampered in ₃ / mice because it resulted from the activation of the three types of -ARs. Note that tyramine and benzylamine were without any lipolytic effect in both control and ₃ / mice.

View larger version (31K): [in this window] [in a new window]   Fig. 2.   Lipolytic effect of 3-AR agonists, octopamine, benzylamine, and tyramine in adipocytes from control and 3 / mice. Lipolytic activity is expressed as percentage of the maximal activity obtained with 10 µM isoproterenol, which was equivalent to 10.7 ± 0.8- and 4.3 ± 0.6-fold the basal glycerol release in control () and in 3 / mice (), respectively. Mean ± S.E.M. of six experiments. *p < 0.05, **p < 0.01, ***p < 0.001 versus control.

View larger version (29K): [in this window] [in a new window]   Fig. 3.   Comparison of the effects of 3-AR agonists and amines on insulin-stimulated glucose transport in adipocytes from control and 3 / mice. Adipocyte preparations, obtained from a pool of two littermates, were incubated during 45 min in the presence of 0.1 µM insulin plus 2 IU/ml ADA and with the indicated concentrations of 3-AR agonists (CL 316243 and BRL 37344) or biogenic amines. The results of hexose uptake assays are expressed as percentage of insulin-stimulated uptake; 100% corresponds to the transport obtained in the presence 0.1 µM insulin plus 2 IU/ml ADA, which was 4.6 ± 0.6 and 3.9 ± 0.7 nmol 2-DG/100 mg of lipid/10 min, in control () and 3-AR / mice (), respectively. The value 0% corresponds to the transport in the presence of ADA alone (1.0 ± 0.1 and 1.0 ± 0.2 nmol of 2-DG/100 mg of lipid/10 min in control and 3 / mice). Mean ± S.E.M. of six experiments. Difference between control and 3 / mice at *p < 0.05, **p < 0.02.

View larger version (38K): [in this window] [in a new window]   Fig. 4.   Prevention by semicarbazide and pargyline of the effects of insulin, benzylamine, and octopamine on glucose transport in 3 / mouse adipocytes. 2-DG uptake into adipocytes was measured for a 10-min period after an incubation of 45 min without (control, ) or with 1 mM semicarbazide () and pargyline () used alone or in combination () and the indicated concentrations of insulin and amines. Each column is the mean ± S.E.M. of four to eight determinations on INWAT adipocytes prepared from 3 / male mice. Different from corresponding control at *p < 0.05, **p < 0.02, ***p < 0.01.

Amine Oxidase Activity in Adipose Tissue from Control and ₃ / Mice. The fact that octopamine, benzylamine, and tyramine stimulate glucose uptake in ₃-AR-deficient mice prompted us to verify putative changes in the amine oxidase activities in adipocytes from ₃ / mice. Figure 5 shows that there was no difference between control and ₃ / mice regarding the capacity of WAT homogenates to oxidize tyramine or benzylamine. As previously reported for rat WAT (Enrique-Tarancon et al., 1998; Marti et al., 1998), tyramine oxidation reached a plateau at 0.5 mM, whereas oxidation of benzylamine was maximal at 0.1 mM in both groups (data not shown). Total inhibition of amine oxidase activities was obtained with the combination of 1 mM semicarbazide (SSAO inhibitor) and 0.5 mM pargyline (MAO inhibitor). Tyramine oxidation was poorly inhibited by 1 mM semicarbazide alone and thus could be mainly attributed to MAO activity (Fig. 5A), whereas more than 80% of benzylamine oxidation was sensitive to semicarbazide and thus SSAO-dependent (Fig. 5B). Taken together, these data demonstrate that 1) murine adipocytes express both MAO and SSAO activities, 2) tyramine is mainly an MAO substrate and benzylamine an SSAO substrate, and 3) there was no change in the MAO and SSAO activities between control and ₃ / adipocytes.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Tyramine and benzylamine oxidation by subcutaneous adipose tissue of control and 3 / mice. Homogenates of SCWAT were preincubated 15 min without (total) or with 1 mM semicarbazide (S) alone or in combination with 0.5 mM pargyline (P + S). Then they were incubated 30 min in a final volume of 200 µl in the presence of 0.5 mM [14C]tyramine (A) or 0.1 mM [14C]benzylamine (B). Means ± S.E.M. of four or seven experiments. (C) Inhibition of benzylamine oxidation by adipose tissue of control and 3 / mice. Homogenates of SCWAT were incubated 30 min with 0.1 mM [14C]benzylamine alone (100%) or in competition with increasing concentrations of tyramine (triangles), octopamine (squares), or benzylamine (circles). Mean ± S.E.M. of three experiments. For A to C, there was no significant difference between control (open symbols or columns) and 3 / mice (closed symbols).

Octopamine Effects on Lipolysis and Glucose Transport in Human Subcutaneous Adipocytes. Human adipocytes possess several similarities with fat cells from the ₃ / mice, because 1) they are poorly responsive to the lipolytic action of both ₃-AR agonists and octopamine (Van Liefde et al., 1994; Carpéné et al., 1998, 1999), and 2) they exhibit high MAO (Pizzinat et al., 1999) and SSAO (Morin et al., 2001) activities. This incited us to verify in human adipocytes whether octopamine was able to activate glucose uptake in an oxidation-dependent manner. Figure 6 shows that human fat cells were less sensitive to the lipolytic action of octopamine than control mice adipocytes: although a significant lipolytic effect could be detected with millimolar concentrations of octopamine, it never reached the maximal lipolysis obtained with isoproterenol. In human adipocytes, the lipolytic effect of 1 mM octopamine was partially inhibited by -AR antagonists. The selective ₂-antagonist ICI 118551 was more efficient than the ₃-antagonist SR 59230A and the ₁-antagonist CGP 20712A, according to the blockade obtained at 0.1 mM: 0.18 ± 0.02, 0.27 ± 0.04, and 0.36 ± 0.05 µmol of glycerol released/100 mg of lipid/90 min, respectively (1 mM octopamine alone being equivalent to 0.38 ± 0. 07 µmol of glycerol/100 mg of lipid/90 min, n = 3). The partial lipolytic action of octopamine was not potentiated by the ₂-AR antagonist RX 821002 (data not shown), whereas it was for norepinephrine (Carpéné et al., 1998)

View larger version (30K): [in this window] [in a new window]   Fig. 6.   Comparative study of octopamine and isoproterenol effects on adipocyte lipolysis in human and control or 3 / mice. Adipocytes from nonobese subjects (), control mice (), or 3 / mice () were incubated during 90 min (human, right y-axis) or 60 min (mouse, left y-axis) without (bas) or with increasing concentrations of octopamine. The glycerol release in response to 10 µM isoproterenol (iso) served as reference for maximal lipolysis. Means ± S.E.M. of six experiments. Different from the corresponding basal lipolysis at *p < 0.05, **p < 0.01, ***p < 0.001.

View larger version (18K): [in this window] [in a new window]   Fig. 7.   Stimulation of glucose transport by octopamine, benzylamine, and insulin in human subcutaneous adipocytes. Adipocytes were preincubated 45 min without (bas) or with the indicated concentrations of octopamine (), 0.1 mM benzylamine (benz), or 100 nM insulin (ins) before a 10-min hexose uptake assay. Mean ± S.E.M. of eight experiments. Different from basal uptake at *p < 0.05, ***p < 0.001.

View larger version (17K): [in this window] [in a new window]   Fig. 8.   Comparison of the capacities of SSAO and MAO to oxidize tyramine (), benzylamine (), and octopamine () in human adipose tissue. [14C]Benzylamine or [14C]tyramine was incubated at 0.5 mM with human SCWAT homogenates during 15 min in the presence of the indicated concentrations of competitors. Oxidation products were extracted and counted as detailed under Materials and Methods. Semicarbazide () and pargyline () are SSAO and MAO inhibitors, respectively. Results are expressed as percentage of the oxidation without any competitor, which was equivalent to 2.8 ± 0.4 and 2.0 ± 0.2 nmol/mg of protein/min for benzylamine and tyramine, respectively. A, mean ± S.E.M. of three experiments; B, mean ± S.E.M. of four to five experiments.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The present comparative approach demonstrates that octopamine, considered as a false neurotransmitter in mammals, exerts in the millimolar range, dual action in fat cells. On one hand, it stimulates lipolysis and counteracts insulin lipogenic action via ₃-AR stimulation, but, on the other hand, it stimulates glucose uptake by a mechanism that is dependent on its oxidation by MAO or SSAO.

The lack of stimulatory action of octopamine on glucose uptake into rat adipocytes was rather unexpected regarding 1) its capacity to be oxidized by rat fat cells, and 2) the diversity of amine oxidase substrates able to stimulate glucose uptake in these cells (Enrique-Tarancon et al., 2000). Previous demonstrations indicating that octopamine counter-regulates insulin stimulation on glucose utilization via ₃-AR activation (Yen et al., 1998; Fontana et al., 2000) prompted us to test the existence of a dual action of octopamine on hexose uptake. The ₃ / mouse model allowed us to demonstrate that the counter-regulation of insulin action by high octopamine concentrations was abolished when ₃-ARs were invalidated. In this model, octopamine behaved as benzylamine and tyramine (which were devoid of lipolytic or antilipogenic action in rats or control mice) because all of them were able to mimic insulin stimulation on hexose uptake in a vanadate-dependent manner. Thus, when its ₃-AR agonism was impaired, octopamine activated glucose uptake like any substrate of the adipocyte MAOs or SSAOs, which, in ₃ / mice, did not change in activity or selectivity.

Moreover, the involvement of ₃-AR activation in the lipolytic effect of octopamine has been evidenced by the parallelism between the net decrease in its action and the loss of responsiveness for ₃-AR-agonists in the WAT of ₃-AR-deficient mice. This observation agrees in full with previous comparative approaches indicating that octopamine was weakly lipolytic in the species in which adipocytes are poorly responsive to ₃-AR agonists (e.g., guinea pig and human) (Carpéné et al., 1998, 1999). In fact, octopamine shared the same maximal effects as ₃-AR agonists when considering stimulation of lipolysis or inhibition of insulin-dependent glucose transport in rat or control mice adipocytes, whereas in human adipocytes, the present data, together with our already reported observations (Carpéné et al., 1999), showed that octopamine, like BRL 37344 or CL 316243, did not reach the maximal lipolytic action of isoproterenol, which is completely blocked by ₁- plus ₂-AR antagonists. Octopamine can therefore be definitively considered as a full ₃-AR agonist, but with a much lower potency than norepinephrine. Nevertheless, we cannot assess that octopamine is totally devoid of ₁-AR-, and especially of ₂-AR agonism, because its feeble lipolytic effect in human adipocytes was more sensitive to inhibition by the ₂-AR antagonist ICI 118551 than by SR 59230A, a ₃-AR antagonist (Manara et al., 1995). No clear agonism at ₂-ARs was evidenced for octopamine in lipolysis experiments as already reported (Fontana et al., 2000), and in glucose transport assays.

In the three species studied, octopamine behaved as a substrate for the amine oxidases present in fat cells, namely, MAO-A and MAO-B (Pizzinat et al., 1999) and SSAO (Lyles, 1996). As already reported for diverse biogenic amines, there were interspecific differences according to the substrate selectivity of amine oxidases (Youdim and Finberg, 1991; Lyles, 1996) because octopamine was a preferential SSAO substrate in rat and murine adipocytes, whereas it exhibited higher affinity for MAO than for SSAO in human adipocytes. Of note is that the effective concentrations of amines toward deaminative oxidation were in the 0.1 to 1 mM range despite the species. These concentrations could appear huge regarding adrenergic pharmacology, but they are in the range of the Michaelis constants of MAOs and SSAOs toward their substrates (Youdim and Finberg, 1991; Lyles, 1996). We have already reported that millimolar doses of tyramine or benzylamine stimulate glucose uptake into rat or human adipocytes, and that this insulin-like effect of amine oxidase substrates was dependent on their oxidation and on the generation of hydrogen peroxide, known to mimic insulin action on glucose transport although its intracellular targets are still undefined (Marti et al., 1998; Morin et al., 2001). The present work extends these observations to murine adipocytes and confirms that vanadate, at a dose ineffective per se, tremendously potentiates the insulin-like effect of amines in rodent adipocytes, whereas this metal, well known for inhibiting phosphatases and mimicking insulin action, is not necessary in human adipocytes. To date, the reason for this difference is not clear but cannot be attributed to interspecific differences because it has already been reported that biogenic amines such as serotonin stimulate glucose uptake into rat cardiomyocytes in an oxidation-dependent manner without need for exogenous vanadate (Fischer et al., 1995). The fact that octopamine stimulation of hexose transport into adipocytes is potentiated by vanadate in ₃ / mice and is abolished by amine oxidase inhibitors in human and ₃ / mice makes it very likely that this effect is dependent on the insulin-like actions of the hydrogen peroxide generated during deaminative oxidation, as previously demonstrated for other amines (Marti et al., 1998; Enrique-Tarancon et al., 2000; Morin et al., 2001).

A reduced lipid mobilization from WAT, which is highly sensitive to ₃-AR activation in mouse, could explain the increase of intra-abdominal and subcutaneous fat depots observed in ₃-AR-deficient mice, without change in body weight gain, as already reported (Susulic et al., 1995). However, the reduction of the ₃-adrenergic inhibitory action on glucose uptake and metabolism could also be involved in the larger development of the fat stores of ₃-AR-deficient mice. Of note, findings establishing that ₃-AR activation is opposed to insulin effects (Carpéné et al., 1993; Klein et al., 1999) are not entirely in disagreement with the antidiabetic properties of ₃-AR-agonists and may explain why, in rodents, such drugs reduce adiposity without body weight loss (Danforth and Himms-Hagen, 1997). First, ₃-AR-agonists inhibit the insulin lipogenic action in fat stores; and second, they promote insulin secretion and action in other peripheral tissues leading to an overall better glucose disposal and a different fuel repartition. Whether endogenous octopamine may participate to this ₃-adrenergic control in any physiological or pathophysiological state seems unlikely, due to its low circulating levels.

Human subcutaneous adipocytes constituted another model in which octopamine hardly activated lipolysis and did not inhibit insulin action. However, octopamine was able to activate glucose uptake into these cells in an oxidation-dependent manner because 1) it constituted an adequate substrate for MAO and SSAO, as reported in other human tissues (Castillo et al., 1999); 2) its effect was blocked by amine oxidase inhibitors, as already reported for benzylamine (Morin et al., 2001); and 3) both MAO and SSAO were substantially expressed in human adipocytes (Pizzinat et al., 1999; Morin et al., 2001). Recent indications obtained on the cellular localization of SSAO in human peripheral tissues also sustained the hypothesis that, in addition to the oxidative deamination of monoamines, amine oxidases might participate in the regulation of physiological processes via hydrogen peroxide generation (Andrés et al., 2001). In humans, plasma levels of octopamine are roughly between 1 and 5 nM (Rossi-Fanelli et al., 1976; Andrew et al., 1993) and can be increased by 3- to 5-fold in individuals older than 70 years, in renal disease (Kinniburgh and Boyd, 1979), and in hepatic encephalopathy (Rossi-Fanelli et al., 1976; Cangiano et al., 1982). Whatever the physiopathological situation, circulating levels of endogenous octopamine are probably not elevated enough to reproduce in vivo the pharmacological effects observed in our in vitro approach. However, because octopamine transport and accumulation occur in human platelets (Murphy et al., 1975), such phenomena remain to be verified in adipocytes, which have an efficient uptake of catecholamines (Pizzinat et al., 1999) and may participate in a cumulative manner to the catabolism of biogenic amines.

To conclude, our data provide evidence that adipocytes from different species, including human, exhibit distinct in vitro responses to octopamine, several being related to the activation of their ₃-ARs and others requiring functional MAO or SSAO. Of note, these responses resulting in changes in glucose utilization are observed with millimolar doses of octopamine and make their physiological relevance unlikely. However, such phenomena could occur during accidental poisoning with octopaminergic pesticides (Evans and Gee, 1980; Costa et al., 1988) or with medicinal plant extracts containing high levels of octopamine-related drugs such as synephrine (Calapai et al., 1999).