Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Amine oxidation of many exogenous or biogenic amines is a widely spread phenomenon occurring in living organisms, from microorganisms to vertebrates. This reaction is catalyzed by two families of amine oxidases: the copper-containing amine oxidase family, which encompasses diamine oxidase, lysyl oxidase, and semicarbazide-sensitive amine oxidases (SSAO) (for review, see Houen, 1999), and the FAD-dependent amine oxidase family, essentially represented by monoamine oxidases A and B (MAO) (Yu, 1986). Tissue-bound SSAOs are dimeric membrane proteins largely oriented to the extracellular space (Salmi et al., 2001). Their inhibition by semicarbazide or carbonyl reagents is due to the presence of a topaquinone in their catalytic site (Lyles, 1996). MAOs are mitochondrial enzymes inhibited by acetylenic compounds like clorgyline, selegiline, and pargyline, which are MAO-A-selective, MAO-B-selective, and mixed inhibitors, respectively (Fowler et al., 1982). It has been thought for a long period that the major function of amine oxidases in mammals was to scavenge circulating amines, including sympathomimetic amines, metabolites of proteolysis, or ingested products. However, the search for other biological functions was prompted by the fact that the products of oxidative deamination, namely, aldehydes, ammonia, and hydrogen peroxide are as reactive as the amines from which they originate (Lyles, 1996).

We have recently shown that several amines such as tyramine or benzylamine markedly stimulate glucose transport in rodent or human adipocytes through their oxidation by SSAO or MAO (Marti et al., 1998; Enrique-Tarancon et al., 2000; Morin et al., 2001). Amine-dependent stimulation of glucose transport was found to be mediated by the hydrogen peroxide generated during amine oxidation since it was prevented by catalase, glutathione, or N-acetylcysteine. In fact, hydrogen peroxide is a well established insulin-mimicking agent that enhances glucose transport (Taylor and Halperin, 1979), stimulates lipogenesis (May and De Haën, 1979), and inhibits lipolysis (Little and De Haën, 1980) in fat cells. These insulin-like effects seem to be mediated, at least in part, by tyrosine phosphorylation of intracellular proteins (Heffetz et al., 1990). Other insulin-like effects, also dependent on hydrogen peroxide generation, have been reported for amine oxidase substrates: stimulation of adipose differentiation in 3T3 preadipocyte lineages (Fontana et al., 2001; Mercier et al., 2001) and inhibition of lipolysis in human fat cells (Morin et al., 2001).

The present work aimed at investigating the tissue distribution of amine oxidase activities and the in vivo relevance of insulin-like effects of their substrates. For this purpose, we have studied the capacity of tyramine to be oxidized by different rat tissues with a special attention to the so-called insulin-sensitive tissues. These tissues, in which a glucose transporter translocation to the cell surface and a subsequent increase in glucose uptake occur upon insulin stimulation, are adipose tissues, and heart and skeletal muscles. The effects of tyramine, which is an indirectly acting sympathomimetic amine (Saavedra, 1988) and also a good substrate for both MAO and SSAO in rodents (Lyles, 1996), were compared with those of benzylamine, an SSAO substrate of reference (Lyles, 1996), since both compounds were shown to elicit phosphorylation of insulin receptor substrates and to stimulate glucose transport in rat fat cells (Enrique-Tarancon et al., 2000). Since a synergism between vanadate and the hydrogen peroxide formed during amine oxidation was observed on isolated rat adipocytes and attributed to the generation of pervanadate, a still more powerful insulin-mimicking agent than either hydrogen peroxide or vanadate (Fantus et al., 1989), our investigations on tyramine effects were conducted in the absence and in the presence of vanadate. Finally, we attempted to reproduce in vivo the tyramine effects on glucose utilization to determine whether the interplay between amine oxidase activation and glucose transporter recruitment could constitute a novel role of peripheral MAOs or SSAOs.

Our results show that tyramine stimulates in vitro hexose uptake not only in adipocytes via its oxidation by MAO and SSAO, but also in cardiomyocytes and skeletal muscles by way of MAO-dependent oxidation. Moreover, in vivo tyramine administration not only exerted transient pressor effects known to be mediated by its norepinephrine-releasing action, but also exhibited antihyperglycemic properties, which were 1) sensitive to SSAO or MAO inhibitors, 2) likely due to a direct increase of hexose uptake in insulin-sensitive tissues, and 3) mostly independent on insulin secretion since it is preserved in insulin-depleted rats.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. [¹⁴C]Tyramine (53.3 mCi/mmol) and 2-[1,2-³H]deoxyglucose (2-DG, 26 Ci/mmol) were purchased from PerkinElmer Life Sciences (Boston, MA). Collagenase, cytochalasin B, fatty acid-free bovine serum albumin, hydrogen peroxide, sodium orthovanadate, tyramine, pargyline, semicarbazide, clorgyline, and other chemicals were from Sigma-Aldrich (St. Louis, MO) unless otherwise stated. Selegiline was kindly provided by Orion-Farmos (Turku, Finland).

Animals and Tissue Sampling. Male Wistar rats were individually housed at 22°C with free access to food and water. The tissues removed from euthanized animals (200-250 g unless otherwise stated) were frozen in liquid nitrogen and stored at 80°C for further homogenate preparation or immediately used for adipocyte or cardiomyocyte isolation and subsequent determinations of glucose transport. The epididymal, retroperitoneal, and perirenal fat pads were pooled as intra-abdominal white adipose tissue (INWAT), while subcutaneous white adipose tissue was composed of inguinal fat pads.

Amine Oxidase Activity. Oxidase activity was measured using [¹⁴C]tyramine according to the radiochemical method described by Tipton and coworkers (Fowler et al., 1982). Briefly, thawed samples of different tissues or organs were homogenized in 200 mM phosphate buffer, pH 7.4, containing an antiprotease cocktail from Sigma- Aldrich. Total homogenates were then diluted to approximately 300 µg of protein/100 µl and were incubated for 30 min at 37°C in 200 µl of phosphate buffer in the presence of 0.5 mM [¹⁴C]tyramine (0.05 µCi) after a 30-min preincubation without (total oxidation) or with inhibitors as detailed below. 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. MAO activity was defined as the part of oxidation inhibited by preincubation with 0.5 mM pargyline, whereas activity inhibited by 1 mM semicarbazide was ascribed to SSAO. Simultaneous addition of pargyline plus semicarbazide was also tested to verify the additivity of MAO and SSAO activities in total tyramine oxidation as already reported (Marti et al., 1998; Pizzinat et al., 1999). For competition studies of tyramine oxidation, the homogenates were preincubated 30 min at 37°C with the indicated concentrations of inhibitors before an additional 30-min incubation with 0.5 mM labeled tyramine (approximately 100,000 dpm in a final volume of 200 µl).

Deoxyglucose Transport in Isolated Adipocytes. Adipose tissues were minced in Krebs-Ringer medium containing 12.5 mM Hepes (KRBH buffer), plus 2 mM pyruvate and bovine serum albumin (BSA; 3.5% w/v). After digestion with collagenase (1.5 mg/ml) for 35 to 45 min at 37°C, isolated fat cells were filtered and washed three times in the same buffer. Fat cell suspensions were incubated with the tested drugs during 45 min at 37°C in plastic vials in a final volume of 400 µl. Then, 2-deoxy-D-[³H]glucose (2-DG, 0.4 µCi) was added at a final concentration of 0.1 mM for 10 min. Assays were stopped with 100 µl of 100 µM cytochalasin B, and aliquots of the cell suspension were centrifuged in microtubes containing phthalic acid dinonyl ester (density 0.98 g/ml), which allowed adipocytes to separate from the buffer and allowed counting of the radioactive intracellular 2-DG. Extracellular 2-DG present in the cell fraction was subtracted from the experimental values and did not exceed 1% of the maximum 2-DG transport in the presence of insulin as previously reported (Morin et al., 2001).

Hexose Transport in Isolated Cardiomyocytes. Cardiomyocytes were obtained by collagenase digestion, and 2-DG uptake was assayed as previously described (Fischer et al., 1995). Briefly, 1.5 ml of cell suspension (approximately 1.5 mg of cell protein in 6 mM KCl, 1 mM Na₂HPO₄, 0.2 mM NaH₂PO₄, 1.4 mM MgSO₄, 128 mM NaCl, 10 mM Hepes, 1 mM CaCl₂, 2% BSA) was preincubated at 37°C and pH 7.4 for 45 min with the tested agents. Samples were then incubated in the presence of 2-[³H]DG (1 µM final) for an additional 30 min. Each assay was done in triplicate and hexose uptake was stopped by 400 µM phloretin. Cells were separated from medium by centrifugation in centrifuge vials containing 1 ml of silicon oil (density 1.03 g/ml), and the radioactivity of 2-DG taken up was counted as previously reported (Fischer et al., 1995).

Hexose Transport in Soleus Muscle. Male Wistar rats of 100 to 140 g were fasted overnight and soleus was dissected under anesthesia (pentobarbital 60 mg/kg i.p.). The isolated muscles were fixed to stainless steel clips to maintain a resting muscular tension during the experiments according to the method described by Stock and coworkers (Liu et al., 1996). Muscles were first incubated for 30 min at 37°C in 4 ml of KRBH buffer, pH 7.4, with 20 mM Hepes, 2 mM pyruvate, and 0.2% BSA, continuously gassed with O₂/CO₂ (95/5). Thereafter, each muscle was incubated for 30 min in 4 ml of fresh KRBH buffer containing the tested agents. A third incubation was conducted for 30 min with the same drugs plus 0.1 mM 2-[³H]DG. Then, the muscles were washed four times in KRBH, blotted, frozen in liquid nitrogen, weighed, and dissolved in 1 ml of Solvable (PerkinElmer Life Sciences, Boston, MA) at 50°C. The radioactivity present in muscle digests was counted by liquid scintillation spectroscopy. An isotopic dilution of [¹⁴C]mannitol (1 mM) was used for the determination of the extracellular space in which 2-DG was also present in soleus without being internalized in the cells. Specific uptake of 2-DG was then calculated by subtracting the values of hexose present in this extracellular space (which increased linearly with muscle sample weight) from total uptake.

Intraperitoneal Glucose Tolerance Tests. Male rats of 200 to 250 g were fasted during 6 h (from 8:00 AM to 2:00 PM) before the tolerance test. Blood samples were drawn from the tail vein of conscious animals at the indicated times before and after glucose load (i.p. bolus of 2 g/kg injected in 6 ml/kg of saline at time 0), and glycemia was immediately determined with the Glucotrend II glucometer (Roche Diagnostics, Mannheim, Germany). A bolus of tyramine at 4 mg/kg (29 µmol/kg) or vehicle (0.9% NaCl, saline) was administered intraperitoneally 15 min before the glucose load in controls or in rats previously treated with semicarbazide (3.2 mg/kg/day) and pargyline (2.4 mg/kg/day) during 1 week. Type 1 diabetic rats were also used for glucose tolerance tests. They were hyperglycemic, hypoinsulinemic and tested at least 2 weeks after streptozotocin treatment (single injection at 65 mg/kg).

In Vivo 2-Deoxyglucose Uptake. Experiments were performed between 9:00 and 11:00 AM on overnight fasted rats under pentobarbital anesthesia (60 mg/kg) with catheters placed in the jugular vein and the carotid artery for drug injections, and under invasive cardiovascular monitoring on a Honeywell recorder, respectively. 2-[³H]DG (25 µCi in 200 µl) was injected into the jugular vein at time 0; then, arterial blood samples were collected after 1, 3, 5, 10, 20, 40, and 60 min for the determination of blood glucose (One-Touch glucometer; Lifescan, Issy les Moulineaux, France) and disappearance of circulating 2-[³H]DG, according to the technique of Sokoloff et al. (1977). Blood samples were immediately deproteinized in Ba(OH)₂/ZnSO₄ and centrifuged (2 min, 13,000g) for 2-DG determination in supernatants. Aliquots of plasma were frozen at -20°C for subsequent insulin radioimmunoassay (ERIA Pasteur, Marnes la Coquette, France) or enzymatic determination of glycerol levels (Roche Diagnostics). One hour after 2-DG injection, rats were killed by cervical dislocation, and tissue samples were transferred in 0.5 ml of 1 M NaOH and heated at 60°C until total digestion. Then 0.5 ml of 1 M HCl was added to each sample. One aliquot (200 µl) of the neutralized solution was added to 1 ml of 6% HClO₄ and another aliquot to 1 ml of Ba(OH)₂/ZnSO₄. After centrifugation, 800 µl of each supernatant served for radioactivity counting in 10 ml of Emulsifier Safe (PerkinElmer Life Sciences). The content of 2-deoxyglucose 6-phosphate internalized was calculated by the difference between the radioactivity found in the HClO₄ (total 2-DG) and Ba(OH)₂/ZnSO₄ (free 2-DG) supernatants.

Estimation of Amine Oxidase Activity and Glucose Index Utilization per Whole Organ or Tissue. The different components of tyramine oxidation, expressed as nanomoles per milligram of protein per minute on tissue samples, served for a lumped estimation of amine oxidase activities per whole organ or tissue. This was based, for each organ or tissue, on 1) direct determination of protein content and, depending on the tissue, 2) wet weight (for liver, kidney, brain, heart, and small intestine) or estimated weight (for total skeletal muscles, and white and brown fat) according to the anatomical tables for laboratory rodents and applied to male rats of 250 g (Suckow, 1998). In vivo glucose utilization rate was estimated as previously described (Ferré et al., 1985), taking into account the rate of glucose analog disappearance in blood and the content of phosphorylated 2-DG in the different anatomical locations studied. Then, extrapolations were made to express the index of glucose utilization per whole tissue or organ (Hom et al., 1984).

Statistical Analyses. Results are given as mean ± S.E.M. Statistical significance was assessed by use of Student's t test.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Tissue Distribution of MAO and SSAO Activities. The capacity of different tissues to oxidize [¹⁴C]tyramine was tested without any inhibitor to determine total oxidation, and in the presence of the following drugs: 1 mM semicarbazide, a selective SSAO-inhibitor that does not affect MAO and thus allows quantification of MAO-dependent activity; 0.5 mM pargyline, a MAO inhibitor (the remaining oxidation being mainly attributed to SSAO activity); and the combination of semicarbazide plus pargyline for the detection of non-MAO-non-SSAO activities. Amine oxidase activities were expressed as the amount of tyramine oxidized per mg of proteins present in homogenates without any subcellular fractioning. Under these conditions, total tyramine oxidation was virtually equivalent to the sum of MAO and SSAO activities in every anatomical location (Fig. 1). Liver was the most oxidative tissue toward tyramine and contained almost exclusively MAO activity. Adipose depots (especially INWAT) also displayed a high oxidative capacity that was accounted for by both MAO and SSAO. Comparably high rates of tyramine oxidation were also found in brain or heart (both containing only MAO), whereas activities were lower in other peripheral tissues. All the adipose depots expressed both MAO and SSAO activities, regardless of their anatomical location, intra-abdominal or subcutaneous, or their nature, white or brown. Adipose tissue, especially INWAT, constituted the major site of SSAO activity, together with aorta and, to a lesser extent, lung and intestine. All the skeletal muscles studied exhibited relatively low levels of tyramine oxidation, mostly attributed to MAO activity. The lowest level of tyramine oxidation was found in blood. Regarding their capacity to metabolize tyramine by both MAO- and SSAO-dependent oxidations, the insulin-sensitive tissues (skeletal and cardiac muscles, adipose tissues) constituted a substantial subset of the MAO/SSAO-containing tissues.

View larger version (35K): [in this window] [in a new window]   Fig. 1.   Comparison of the capacities of SSAO and MAO to oxidize tyramine in diverse tissues of the rat. Tissue homogenates were preincubated during 30 min without (total oxidation, white bars) or with 1 mM semicarbazide, 0.5 mM pargyline, or the combination of both (black columns). Then 0.5 mM [14C]tyramine was added for an additional 30 min. Semicarbazide and pargyline are SSAO and MAO inhibitors, respectively. SCWAT, subcutaneous white adipose tissue; BAT, brown adipose tissue. Mean ± S.E.M. of 5 to 10 experiments.

Tyramine Is Mainly Oxidized by MAO-A in Cardiac and Skeletal Muscles. Figure 2 shows that, in adipose cells, none of the tested amine oxidase inhibitors was able to totally prevent [¹⁴C]tyramine oxidation since only 50% of the oxidation was maximally inhibited by semicarbazide (SSAO inhibitor), clorgyline (MAO-A inhibitor), or selegiline (MAO-B inhibitor). Total inhibition of tyramine oxidation was only achieved when 1 mM semicarbazide was combined with a 0.1 mM concentration of each selective MAO inhibitor (not shown) or with pargyline, as already reported for white adipose tissue homogenates in Fig. 1 and for adipocyte crude membranes (Marti et al., 1998). Although it led to an incomplete blockade, clorgyline exhibited a higher potency than did selegiline in inhibiting amine oxidase activity. Thus, around 50% of tyramine oxidation was essentially resulting from MAO-A rather than MAO-B activity, whereas the 50% remaining was insensitive to MAO inhibitors and, therefore, SSAO-dependent. A different pattern was observed in cardiomyocytes or soleus muscle homogenates: tyramine oxidation was totally blocked by MAO inhibitors, with clorgyline being markedly more potent than selegiline. Semicarbazide produced only a discrete inhibition in soleus muscle and had no influence in cardiac cells (Fig. 2). The same pattern of tyramine oxidation was observed in diaphragm homogenates (not shown). These data suggested that MAO-A is the predominant amine oxidase activity in muscle cells.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Involvement of MAO-A, MAO-B, and SSAO in tyramine oxidation in adipocytes, cardiomyocytes, and skeletal muscle. Homogenates of thawed isolated cells or tissues were preincubated for 30 min at 37°C with the indicated concentrations of inhibitors and then incubated for 30 min with 0.5 mM tyramine. Results are expressed as percentage of the total oxidation of tyramine obtained without any inhibitor, which was equivalent to 4.5, 1.6, and 0.3 nmol/mg protein/min in adipocytes, cardiomyocytes, and skeletal muscle preparations, respectively. Each point represents the mean ± S.E.M. of six to eight determinations in adipocytes, three in cardiomyocytes, and four to six in muscle.

Stimulation of Glucose Transport by Tyramine in Adipocytes, Cardiomyocytes, and Soleus Muscle. Since tyramine (Marti et al., 1998) and benzylamine (Enrique-Tarancon et al., 2000) stimulate glucose uptake in rat adipocytes in the presence of vanadate, it was of interest to test whether two other insulin-sensitive tissues, namely cardiac and skeletal muscles, were also responsive to amine oxidase substrates. Therefore, 2-DG uptake assays were conducted on freshly isolated adipocytes, cardiomyocytes, and soleus muscles. In all models, hydrogen peroxide was able to partially mimic the insulin stimulation of glucose uptake (Fig. 3). The presence of 0.1 mM vanadate did not modify basal or insulin-stimulated transport, whereas it significantly improved the response to hydrogen peroxide via a synergism likely due to pervanadate formation (Fantus et al., 1989). When tested alone on adipocytes, tyramine and benzylamine were hardly able to activate glucose uptake, whereas in the presence of vanadate, they stimulated up to 80% of the maximal insulin effect (Fig. 3). In rat cardiomyocytes, 0.3 mM tyramine activated 2-DG uptake to around 40% of the maximal insulin effect, and benzylamine had no effect. The synergism between vanadate and amines did not occur in isolated cardiomyocytes. Hexose transport in soleus muscle was slightly but significantly stimulated by tyramine, whereas the benzylamine effect remained under the limits of significance. Although vanadate exhibited a tendency to increase the tyramine effect, it did not provoke a clear-cut potentiation, in contrast with the synergism observed with hydrogen peroxide.

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Stimulation of glucose transport by insulin, hydrogen peroxide, vanadate, and tyramine in rat in rat white adipocytes, cardiomyocytes, and soleus muscle. 2-DG uptake was measured as described under Materials and Methods on isolated fat cells (upper panel), cardiomyocytes (middle panel), or soleus muscle (lower panel) after 1 h preincubation without (white columns) or with 0.1 mM vanadate (dark columns) under basal conditions or under stimulation by 100 nM insulin, 1 mM hydrogen peroxide, 1 mM tyramine or 0.1 mM benzylamine (for cardiomyocytes, insulin was tested at 20 nM, tyramine at 0.3 mM). Results are expressed as -fold increase over basal uptake. The respective values for basal and insulin-stimulated 2-DG uptake were 0.9 ± 0.1 and 7.9 ± 0.8 nmol/100 mg lipid/10 min for adipocytes, 22.1 ± 3.1 and 227.6 ± 27.5 pmol/mg protein/60 min for cardiomyocytes, and 57.2 ± 7.9 and 203.6 ± 8.6 nmol/g wet tissue/30 min for soleus muscle. Mean ± S.E.M. of n experiments. Significant increase over basal at , p < 0.05; , p < 0.01; or , p < 0.001. Different from corresponding condition without vanadate at , p < 0.05; , p < 0.01; or , p < 0.001.

View this table: [in this window] [in a new window]   TABLE 1 Effects of amine oxidase inhibitors on the tyramine-dependent stimulation of hexose uptake in cardiomyocytes or adipocytes Rat cardiomyocytes were incubated with 100 µM tyramine for 1 h without and with MAO inhibitor (1 µM clorgyline or selegiline, 100 µM pargyline) or SSAO inhibitor (1 mM semicarbazide), and assayed for 2-DG uptake at 30 min. Adipocytes were incubated 45 min with 0.1 mM vanadate alone or with the indicated concentration of tyramine or insulin in the absence (no inhibitor) or in the presence of 100 µM MAO inhibitors or 1 mM semicarbazide, and assayed for 2-DG uptake at 10 min. Results are expressed as -fold increase over basal 2-DG uptake. Values are mean ± S.E.M. of 4 experiments on cardiomyocytes, and 4 (vanadate alone or with insulin) or 10 (vanadate plus tyramine) experiments on adipocytes. Different from control without inhibitor at: * p < 0.05; ** p < 0.01; *** p  0.001.

Effect of Acute Administration of Tyramine and Vanadate on Glucose Tolerance Test. To investigate whether the insulin-like effect observed in vitro could occur in vivo, we tested the influence of an i.p. tyramine bolus injection in a glucose tolerance test (IPGTT). Tyramine was injected at 4 mg/kg (29 µmol/kg), a dose previously shown to induce pressor responses (Finberg and Youdim, 1988), 15 min before the glucose load. The hyperglycemic response was markedly reduced after the tyramine challenge (Fig. 4). When tested at 4.6 mg/kg (25 µmol/kg), sodium orthovanadate also diminished the hyperglycemic response but was less efficient than tyramine since it allowed the peak increase of glycemia 15 min after glucose load. When combined, tyramine and vanadate tended to have an even stronger antihyperglycemic action, although no statistically significant additivity could be demonstrated.

View larger version (42K): [in this window] [in a new window]   Fig. 4.   Influence of an acute administration of tyramine and vanadate on glucose tolerance test in conscious male rats. Male Wistar rats (6-h fasted) received an i.p. injection (drug, arrowhead) of vehicle (saline, 1 ml/kg), or tyramine (4 mg/kg) and sodium orthovanadate (4.6 mg/kg) alone or in combination, 15 min before a glucose load (2 g/kg i.p., time 0). Blood glucose levels are given in mg/100 ml; each point is the mean of six different rats per group. Inset, areas under the curve (AUC) of the increase in blood glucose, given in arbitrary units. Different from respective control at , p < 0.05; , p < 0.01; or , p < 0.001.

Tyramine-Dependent Improvement of Glucose Tolerance in Rats Treated with MAO and SSAO Inhibitors or with Streptozotocin. Rats were treated daily with semicarbazide (3.2 mg i.p./kg), pargyline (2.4 mg i.p./kg) or vehicle (NaCl 0.9%) one week before glucose tolerance tests. These treatments did not modify body weight gain, plasma glucose or insulin levels (not shown). Semicarbazide treatment abolished SSAO-dependent tyramine oxidation in INWAT (from 0.26 ± 0.04 to 0.06 ± 0.01 nmol/mg protein/min, p < 0.01), whereas pargyline treatment resulted in a 75% decrease of MAO activity when compared with controls treated with saline (from 0.17 ± 0.03 to 0.05 ± 0.01 nmol/mg protein/min, p < 0.02). Figure 5 shows that the antihyperglycemic effect of tyramine observed during IPGTT in controls was not found in rats previously treated with amine oxidase inhibitors.

View larger version (44K): [in this window] [in a new window]   Fig. 5.   Influence of a chronic treatment with semicarbazide or pargyline on the tyramine-dependent improvement of glucose tolerance in the conscious rat. Rats were injected daily i.p. with semicarbazide (3.2 mg/kg/day), pargyline (2.4 mg/kg/day), or saline (control) during 1 week. Rats were fasted during 6 h; then, tyramine (4 mg/kg) or vehicle (saline) was injected i.p. 15 min before glucose load (2 g/kg given i.p. at time 0). Results are given as increase over glucose blood levels at time 0, which were: 91 ± 5 mg/100 ml in control rats (n = 10) receiving either saline (closed circles) or tyramine (open circles), and 109 ± 6 mg/100 ml in semicarbazide (n = 5, NS)- and 118 ± 3 mg/100 ml in pargyline-treated rats (n = 4, p < 0.01). Inset, areas under the curve (AUC) of the increase in blood glucose. Different from respective control at , p < 0.05; , p < 0.01.

View larger version (27K): [in this window] [in a new window]   Fig. 6.   Effect of tyramine plus vanadate on glucose tolerance in the diabetic streptozotocin-treated rat. Type 1 diabetic rats were fasted during 6 h, and then tyramine (4 mg/kg) plus vanadate (0.0046 mg/kg) or vehicle (saline) was injected i.p. 15 min before glucose load (2 g/kg). Results are given as increase over glucose blood levels, which were, at time 0: 361 ± 24 and 341 ± 29 mg/100 ml for saline (closed circles) and tyramine group (open triangles), respectively (n = 5, NS). Different from untreated diabetic at , p < 0.05.

Effect of Tyramine on in Vivo Glucose Utilization in Various Tissues of the Anesthetized Rat. A given drug that improves glucose tolerance may act by increasing insulin secretion or by intensifying glucose disposal (insulin-sensitizer or insulin-mimicker). We therefore verified whether tyramine could promote in vivo glucose uptake or insulin release. A tracer dose of 2-[³H]DG was injected into anesthetized rats 15 min after tyramine or saline challenge to measure putative alterations of glucose consumption in various tissues or organs. The rats were monitored before and after tyramine i.v. injection at 4 mg/kg to determine any change in heart rate or blood pressure. In addition, blood samples were periodically taken for the determination of circulating glucose, insulin, and glycerol (Fig. 7). Tyramine immediately provoked a transient increase in blood pressure with a parallel (although not significant) increase in heart rate. This short-lasting pressor action probably resulted from the previously reported interaction of tyramine with catecholaminergic endings and norepinephrine release (Saavedra, 1988). Similarly, an increase in blood glucose and a peak in plasma insulin were immediately detected after tyramine injection. All these parameters returned to basal levels 15 to 20 min later, i.e., when 2-DG injection was made. No further changes in cardiovascular or metabolic parameters were observed during the 60 min following the hexose tracer administration (Fig. 7). At the end of this period, various tissue samples were taken to determine their in vivo hexose uptake by using internalized radioactive 2-DG-phosphate as an index (Ferré et al., 1985). As illustrated in Fig. 8, the accumulation of 2-DG-phosphate was markedly increased after tyramine treatment in soleus and diaphragm muscles, heart, intra-abdominal white fat pads, and interscapular brown adipose tissue. In contrast, uptake was unchanged in EDL muscle, subcutaneous adipose tissue, jejunum, or brain of tyramine-treated rats when compared with controls. The 2-DG-phosphate levels were not determined in liver and kidney since this phosphorylated glucose analog, poorly metabolized in many tissues, is a substrate for the glucose 6-phosphatase activity highly expressed in these organs (Ferré et al., 1985). Thus, these data showed that, without any glucose load, tyramine administration was able to increase in vivo glucose transport in almost all the insulin-sensitive tissues at a dose which caused only minor changes in cardiovascular parameters or insulin secretion.

View larger version (22K): [in this window] [in a new window]   Fig. 7.   Evolution of circulating levels of glucose, insulin, and glycerol, and changes in cardiovascular parameters in anesthetized rats after i.v. administration of tyramine and 2-deoxyglucose. Rats of 220 to 230 g were anesthetized by pentobarbital. Tyramine (4 mg/kg) or its vehicle (saline) was injected into the jugular vein 15 min before the i.v. administration of a trace amount 2-DG (25 µCi); then, blood samples were taken at the indicated times for the determination of blood glucose, plasma insulin, or glycerol levels. Blood pressure and heart rate were monitored through carotid cannulation. Mean ± S.E.M. of five experiments. Different from respective control (saline) at , p < 0.05 or , p < 0.001.

View larger version (31K): [in this window] [in a new window]   Fig. 8.   In vivo hexose uptake in different tissues after i.v. tyramine administration to anesthetized rats. 2-Deoxyglucose phosphate content was determined as described under Materials and Methods in different tissues 60 min after i.v. injection of 2-DG (25 µCi/rat) in rats previously treated with tyramine (4 mg/kg i.v., closed columns) or with saline (open columns). Mean ± S.E.M. of the five experiments reported in Fig. 7. BAT, brown adipose tissue; WAT, white adipose tissue. Different from hexose uptake in control condition (saline) at , p < 0.05 or , p < 0.01.

Glucose Uptake Stimulation and Amine Oxidase Content per Organ or Tissue. Estimates of glucose consumption by each tissue or organ were obtained by balancing data of 2-DG-phosphate internalization presented above with radioactive 2-DG clearance in blood and with tissue/organ mass. The resulting indexes of glucose uptake were expressed as micrograms per whole organ or tissue per minute and are shown in Fig. 9. Similarly, data of tyramine oxidation obtained on homogenates were transformed to evaluate the whole amount of MAO and SSAO present per tissue/organ (Fig. 9A). Under these conditions, liver was confirmed as the major anatomical site of MAO activity. Although all tissues or organs were not analyzed, our approach showed that, for a male rat weighing 250 g, the whole skeletal muscular mass contained a much larger total MAO activity than did other organs well known for their richness in this mitochondrial enzyme such as brain, kidney, or heart. White adipose tissues represented one of the most important sites of SSAO activity. Regarding glucose uptake capacity, skeletal muscles constituted, as expected, the major component of basal glucose consumption, whereas white fat depots represented a lower amount. Tyramine challenge increased glucose uptake in fat depots, heart, and muscles, whereas it did not modify the glucose utilization of intestine or brain (Fig. 9B). This differential effect was apparently not due to a lack of amine oxidase activity in the latter anatomical locations, since they contained as much MAO activity as heart or adipose tissue (Fig. 9A). In fact, it appeared that tyramine stimulates glucose uptake only in insulin-sensitive tissues in a manner that is apparently related to their amine oxidase content.

View larger version (24K): [in this window] [in a new window]   Fig. 9.   Content of amine oxidase activity and amount of tyramine-dependent glucose uptake in insulin-sensitive tissues and in other organs of the rat. A, tissue distribution of amine oxidase activities in the adult rat. Data of tyramine oxidation of tissue homogenates per milligram of protein, presented in Fig. 1, were used to express amine oxidase activity in nanomoles oxidized per whole organ or tissue per minute, after evaluation of wet weight and total protein content for each tissue or group of tissues, as described under Materials and Methods. B, estimates of in vivo glucose utilization in rat tissues after saline or tyramine injection (4 mg/kg i.v.). As detailed under Materials and Methods, the results of in vivo 2-DG uptake reported in Fig. 8 were used together with data of the radioactive 2-DG clearance in blood for the calculation of glucose utilization index expressed as micrograms per whole organ or tissue per minute. For skeletal muscles and white fat depots, an average of 2-DG uptake values found in the distinct anatomical locations was used for the lumped estimation of glucose utilization. ND, not determined, due to the presence of glucose 6-phosphatase in the corresponding organ.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

This work shows that tyramine exhibits an in vitro insulin-like action in adipocytes and cardiac or skeletal muscles; it also demonstrates that this action occurs in vivo and influences glucose disposal. Both in vitro and in vivo metabolic effects of tyramine are dependent on oxidation by MAO or SSAO, which readily occurs not only in organs well recognized for their catabolism of amines (brain, liver, kidney, intestine, lung) but also in tissues able to increase their glucose uptake under insulin stimulation (adipose tissue, heart, muscles).

Analysis of MAO and SSAO tissue distribution was conducted with one concentration of tyramine (0.5 mM) largely higher than the K_m values we found on cardiomyocyte (83 µM) or on adipocyte preparations (138 µM) (Marti et al., 1998). The observed anatomical distribution, in good agreement with that found in humans (Lewinsohn et al., 1980), showed that tyramine oxidation was 1) mainly MAO-dependent in liver, heart, brain, kidney, and skeletal muscles and 2) due to both MAO and SSAO in adipose tissues and aorta. Our observations also indicated that all the studied muscles expressed low MAO activity. However, skeletal muscles represent around 40% of the body mass; hence, a large proportion of total MAO activity is located to muscles. Similarly, the small proportion of SSAO involved in tyramine oxidation by skeletal muscles, although barely detectable on samples, becomes non-negligible with regard to the whole muscular mass (compare Figs. 1, 2, and 9). A hardly detectable SSAO activity was also found in liver and was likely due to the presence of the recently characterized vascular adhesion protein SSAO/VAP-1 (Salmi et al., 2001) expressed in endothelial cells of hepatic sinusoids (Grant et al., 2002). Aorta was also a location of vascular SSAO activity, in agreement with the high SSAO content found in vascular smooth muscle cells (El Hadri et al., 2002). More importantly, the novel insight of our tissue distribution study is that white adipose depots constitute a major origin of SSAO activity.

The effective concentrations of tyramine in hexose uptake assays were in the 0.01 to 1 mM range. Although in the range of the K_m values of MAOs and SSAOs (Lyles, 1996; Houen, 1999), these concentrations are more elevated than those usually used for in vitro receptor pharmacology and obviously raise the question of the selectivity of action. However, several points argue for an amine oxidase-dependent mechanism in the observed effects of tyramine. First, the influence of tyramine on glucose metabolism was almost abolished when amine oxidases were blocked both under in vitro and in vivo conditions. Second, tyramine was not found to be lipolytic in adipocytes and, therefore, devoid of direct adrenergic agonism (Visentin et al., 2001). Third, stimulation of glucose uptake in fat cells was obtained with other diverse amine oxidase substrates (benzylamine, methylamine) and was always mediated by the oxidation product hydrogen peroxide, as demonstrated by blockade with catalase, or diverse antioxidants (Marti et al., 1998). This was true as well for the improvement of adipose differentiation, another insulin-like effect of these amines (Fontana et al., 2001; Mercier et al., 2001) also mediated by hydrogen peroxide (Taylor and Halperin, 1979). All these points make it very likely that the tyramine-induced improvement of glucose utilization observed in adipose tissue, heart, and muscle is a direct effect mediated by the hydrogen peroxide generated during deaminative oxidation.

High concentrations of vanadate are well known to increase protein phosphorylation (Heffetz et al., 1990). Our data confirm that this compound tremendously potentiates the insulin-like effect of hydrogen peroxide in all of the three insulin-sensitive tissues, as already observed in studies on pervanadate effects (Fantus et al., 1989; Huyer et al., 1997). At a dose that is ineffective per se, vanadate also potentiated tyramine or benzylamine effect in rat adipocytes (Marti et al., 1998; Enrique-Tarancon et al., 2000), whereas it was less effective in potentiating tyramine stimulation of glucose uptake into skeletal muscle or cardiomyocytes. Since SSAO is highly expressed in fat cells, it could be suspected that only the extracellular oxidation products generated by the ecto-enzyme SSAO can be potentiated by vanadate. In good agreement with this is the finding that vanadate potentiates the effect of benzylamine, a preferred SSAO substrate, in fat cells only. Accordingly, rat cardiomyocytes that only express MAO were unresponsive to benzylamine but responsive to tyramine even without vanadate, suggesting that the mitochondrial MAO produced enough hydrogen peroxide to promote the translocation of glucose transporters, but not enoughor not in the functional compartmentto be complexed with exogenous vanadate. Although the exact mechanisms leading to a potentiation between amines and vanadate remain unclear, the observed stimulations of in vitro glucose uptake prompted us to test the effectiveness of tyramine to stimulate glucose utilization in conscious and anesthetized rats. Under these conditions, an antihyperglycemic effect was detected with tyramine alone and reinforced in the presence of exogenous vanadate, an observation in agreement with in vitro data.

The fact that MAO or SSAO inhibitors block the metabolic effects of tyramine clearly contrasts with the well known potentiation that MAO inhibitors exert on tyramine pressor effects (Finberg and Youdim, 1988; Faukhauser et al., 1994). To explain this discrepancy, it is necessary to remember that tyramine is an indirectly acting sympathomimetic amine that releases norepinephrine from catecholaminergic terminals and consequently increases blood pressure. Therefore, the cardiovascular responses to tyramine are increased when metabolism of endogenous norepinephrine is reduced by MAO inhibitors. In contrast, tyramine directly activates glucose uptake in target cells as a consequence of the insulin-mimicking properties of hydrogen peroxide generated by oxidative deamination, sensitive to blockade of MAO and SSAO. In the present work, we compared the cardiovascular and metabolic responses to tyramine. The former were immediate and transient: blood pressure and heart rate returned to basal values in less than 10 min. Increase in plasma insulin was also modest and transient (2-fold increase over less than 15 min, when compared with the 5-fold increase lasting more than 30 min after a classical glucose load), whereas significant improvements in glucose utilization were observed in a period of 1 to 2 h. The tyramine-stimulated glucose utilization probably did not result from this transient insulin secretion but from another long-lasting event. Tyramine metabolism by peripheral amine oxidases may be likely involved in this long-lasting component since antihyperglycemic response was impaired when MAO or SSAO activities were inhibited. On the contrary, the tyramine antihyperglycemic effect was maintained in insulin-depleted diabetic rats. In all, our observations suggest that, in vivo, tyramine readily increases glucose disposal, not via insulin secretion, but by increasing glucose uptake in insulin-sensitive tissues, similarly to that observed in vitro.

Whether endogenous tyramine participates in the physiological control of glucose metabolism appears speculative, since endogenous plasma tyramine concentration is around 10 nM in control fasted subjects (Faraj et al., 1979), i.e., much lower than the doses needed for the observation of in vitro responses. However, tyramine is present in many foods and beverages, together with other alimentary amines (e.g., histamine, tryptamine, octopamine, putrescine) (Kirschbaum et al., 2000). The mere fasting levels of tyramine are therefore not indicative enough against a putative relevance of the in vivo effects of alimentary amines on glucose metabolism, since 1) in vitro insulin-like effects of amines are not limited to tyramine but can be extended to other amine oxidase substrates, including polyamines (Enrique-Tarancon et al., 2000; Mercier et al., 2001); 2) the effects of amines are additive until the maximal capacity of amine oxidases is attained; and 3) it cannot be excluded that, in postprandial states, the total amount of alimentary amines crossing the intestinal barrier (Hasan et al., 1988) could reach the threshold necessary to influence glucose metabolism. In keeping with this, one has to remember that the daily standard food intake contains around 30 mg of alimentary amines in industrialized countries (Pfudstein et al., 1991), and that small amounts of normally harmless pressor amines in foods can lead to a severe hypertensive crisis termed "cheese-effect" in patients treated by irreversible MAO inhibitors (McCabe, 1986). Whether the tyramine-induced improvement of glucose disposal may have a physiological relevance or may constitute a pharmacological approach of diabetes treatment deserves further investigation.

To conclude, our data demonstrate that the functional MAO or SSAO present in all the insulin-sensitive tissues can no longer be neglected in comparison with the more recognized amine oxidase activities present in brain, liver, or kidney. From a quantitative point of view, the overall mass of skeletal muscles holds a MAO activity that is markedly higher than in other organs or tissues, with the exception of the liver. From a qualitative point of view, this muscular MAO activity may, as well as adipose MAO and SSAO, stimulate glucose utilization upon activation by amine substrates as demonstrated here in vivo with tyramine. However, the acute stimulating effects of tyramine on glucose uptake reported in the present work are still insufficient to assess firmly that amine oxidation can constitute a therapeutic approach to diseases linked to glucose intolerance. In fact, tyramine, which is considered as a false neurotransmitter (Saavedra, 1988), has indirect sympathomimetic action and should not be selected for such pharmacological investigations. Structure-activity studies of MAO and SSAO ligands, which have been for a long period dedicated to the design of selective inhibitors, may generate more powerful amines, devoid of undesirable effects. In this regard, it is important to note that prolonged treatment with benzylamine (which does not occur naturally and is devoid of sympathomimetic effect) reduces the hyperglycemia of the streptozotocin-induced diabetic rat (Marti et al., 2001).