Semicarbazide-Sensitive Amine Oxidase Substrates Stimulate Glucose Transport and Inhibit Lipolysis in Human Adipocytes

Materials and Methods

Chemicals.

[¹⁴C]Benzylamine (57 mCi/mmol) came from Amersham Pharmacia Biotech (Arlington Heights, IL). [α-³²P]dCTP and 2-[1,2-³H]deoxyglucose (2-DG; 26 Ci/mmol) were from PerkinElmer Life Science Products (Boston, MA). Selegiline was purchased from RBI (Natick, MA). Sodium orthovanadate, pargyline, semicarbazide, clorgyline, collagenase, cytochalasin B, bovine serum albumin (fraction V), and other chemicals were purchased from Sigma Aldrich (St. Quentin, France). All electrophoresis reagents were obtained from Bio-Rad (Richmond, CA), except the prestained molecular weight standards that were from Novex (San Diego, CA). Anti-SSAO antibodies were produced from rabbit after immunization with SSAO purified from bovine lung (Lizcano et al., 1998). Enzymes and cofactors used for the determination of glycerol release by isolated fat cells were from Boehringer Mannheim (Mannheim, Germany).

Subjects.

Samples of subcutaneous adipose tissue were obtained from a total of 60 healthy nonobese women: mean body mass index (BMI) was 25.2 ± 0.6 kg · m⁻², and age ranged from 22 to 58 year. Mammary or abdominal adipose tissue was obtained from patients undergoing dermolipectomy (not suction lipectomy) of localized fat depots for cosmetic reasons in the Department of Plastic Surgery of Toulouse Rangueil Hospital. All patients had fasted overnight, and the operations were performed in the morning under general anesthesia. The study was approved by the Ethical Committee of Toulouse University Hospital. Samples of subcutaneous tissues were transported to the laboratory in less than 15 min and were promptly subjected to homogenization or adipocyte preparation through collagenase digestion.

Adipocyte Isolation, Adipose Tissue Homogenization, and Crude Membrane Preparation.

Samples of adipose tissue were dissected and digested by collagenase (1 mg/ml) in Krebs-Ringer solution, pregassed with 95% O₂/5% CO₂, containing bovine serum albumin (35 mg/ml), 15 mM sodium bicarbonate, and 10 mM Hepes (KRBHA buffer, pH 7.4). After digestion for 35 to 45 min at 37°C under agitation, isolated fat cells were filtered and washed three times in the same buffer without collagenase. Freshly isolated adipocytes were adjusted to a suitable dilution (around 50 mg of cellular lipids/ml of KRBHA, equivalent to approximately 300,000 cells/ml) and immediately dispensed in plastic vials for the determination of amine oxidase, glucose transport, or lipolytic activities in intact cells. In several sets of experiments, isolated fat cells were subjected to RNA extraction or disrupted for crude membrane preparation by hypo-osmotic lysis and centrifugation (40,000g for 15 min) as previously described (Pizzinat et al., 1999). RNA extracts and membrane pellets were stored at −80°C until Northern blot, Western blot, or amine oxidase assay. Alternatively, fresh adipose tissue samples were homogenized with an ultraturrax (30 s at 24,000 rpm) in 200 mM potassium phosphate buffer in the presence of antiproteases (protease inhibitor cocktail from Sigma Aldrich at 50 μl/g of tissue/20 ml). Homogenates were then immediately used, without washing or centrifugation, for the determination of [¹⁴C]benzylamine oxidation, protein, and lipid contents. Small samples of adipose tissues (0.1–1 g) were also stored at −80°C up to 2 weeks before homogenization and measurement of amine oxidase activity on thawed material.

Determination of Amine Oxidase Activity.

Amine oxidase activity was determined radiochemically at 37°C as previously described (Enrique-Tarancon et al., 1998) using [¹⁴C]benzylamine as substrate, in both homogenates and intact cells. Fifteen minutes of preincubation with 1 mM semicarbazide or 0.5 mM pargyline was used for complete inhibition of SSAO or MAO, respectively. Isotopic dilutions of [¹⁴C]benzylamine (maximal concentration: 2.5 mM) were incubated at 37°C for 15 min in a final volume of 500 μl for intact cells (in KRBHA buffer) or for 30 min in 200 μl for homogenates (in 200 mM phosphate buffer containing approximately 50 μg of protein). Deaminated labeled products were extracted in toluene/ethyl acetate and counted as previously described (Pizzinat et al., 1999). The large amount of bovine serum albumin in KRBHA buffer, necessary for fat cell viability (protection against intracellular accumulation of fatty acids and cell breakage) was contaminated by bovine plasma amine oxidase that was responsible for less than 10% of the benzylamine oxidation due to cell suspensions and that was subtracted using blanks with KRBHA alone. In addition, SSAO activity was expressed as nanomoles of amine oxidized/100 mg of cell lipids/minute for a better comparison between fat cell preparations and tissue homogenates. Lipid content of homogenates or adipocyte suspensions was gravimetrically determined (Carpéné et al., 1994).

For the assay of amine oxidation by crude membranes, thawed membrane preparations from different individuals were pooled, then protein content was determined, and 100 μg of membrane protein were incubated for 30 min at 37°C in 200 mM potassium phosphate buffer, pH 7.6. Amine oxidation was determined using a continuous spectrophotometric method coupled to peroxidase (Holt et al., 1997). Briefly, 4-aminoantipyrine is oxidized by the hydrogen peroxide formed during amine oxidation and then condenses with vanillic acid to give a red quinone imine dye. The absorbance at 498 nm is proportional to the amount of hydrogen peroxide. The molar absorption coefficient of the dye at pH 7.6 is 4654 M⁻¹ · cm⁻¹. To determine the respective contribution of MAO and SSAO activity, the counterpart oxidase was inhibited by preincubating with 1 mM semicarbazide or 1 mM clorgyline for 30 min, respectively.

For all the determinations of oxidase activity, total amine oxidation was also measured without previous inhibition. Time course assays were used to ensure that initial rates of the reaction were established and proportionality to enzyme concentration was determined. Unless otherwise stated, the kinetic parameters were calculated by using the nonlinear regression analysis of the GraphPad Prism (GraphPad Software, Inc., San Diego, CA).

Immunoblot Analyses.

Protein extracts were obtained from diverse human tissues by ultraturrax homogenization of frozen nonpathological biopsies collected at the hospital. For the two adipose depots studied (mammary or abdominal), fat cell membrane preparations obtained from a pool of individuals were fractionated by electrophoresis on 10% SDS-polyacrylamide gels and transferred to nitrocellulose membranes. Blots were blocked for 1 h at room temperature with 5% nonfat dried milk in 20 mM Tris-HCl, pH 7.5, 150 mM NaCl, and 0.1% Tween 20, before incubation with primary anti-SSAO antibody overnight at 4°C, followed by incubation for 1 h with secondary antibody. Immunoreactivity was detected with the avidin-biotin peroxidase technique. After color development, the membrane was washed with distilled water, air dried, and photographed.

Northern Blot Analyses.

Total RNAs were extracted from isolated fat cell preparations using the guanidium isothiocyanate/phenol/chloroform method (Chomczynski and Sacchi, 1987) and reverse-transcribed using Superscript II and oligo-dT under the manufacturer's conditions (Life Technologies, Inc., Gaithersburg, MD). The polymerase chain reaction was performed on these cDNAs as previously described for the placenta (Zhang and McIntire, 1996), and a 600 bp fragment was cloned into pGEM-T easy (Promega, Madison, WI). Sequencing of this insert was performed with Amplitaqfs dye terminator sequencing kit (PerkinElmer, Norwalk, CT) and a ABI 373 sequencer. The sequence showed 100% homology with the hpao cDNA (Zhang and McIntire, 1996). The 600 bp fragment was excised byEcoRI restriction enzyme digestion and gel purified from the vector before random priming labeling with the Pharmacia random priming kit and [α-³²P]dCTP. Twenty micrograms of total RNA were loaded on a 0.8% agarose gel. After electrophoresis and transfer on nylon N⁺ (Amersham) under standard conditions, hybridization was carried out in Quickhyb solution (CLONTECH) for 2 h with the labeled probe (approximately 2 · 10⁶ cpm/ml). After washing, the membrane was exposed for 3 h and revealed in a PhosphorImager 445 SI (Molecular Dynamics).

Hexose Transport and Lipolysis.

Plastic vials containing 0.4 ml of cell suspension in KRBHA, 2 mM sodium pyruvate, and the tested drugs were incubated for 45 min at 37°C. Then, an isotopic dilution of 2-deoxy-d-[³H]glucose was added at a final concentration of 0.1 mM (approximately 1,300,000 dpm/vial), assays were further incubated for 10 min, and then stopped with 100 μl of 100 μM cytochalasin B. Aliquots (200 μl) of cell suspension were centrifuged as described by Olefsky (1978) in microtubes containing dinonyl phthalate, which allowed the adipocytes to separate from the buffer. After centrifugation, the fat cells (upper part of the tubes) were placed in scintillation vials, and the intracellular radioactivity was counted as previously described (Marti et al., 1998). Extracellular 2-DG present in the cell fraction, which was determined using adipocytes whose transport activity had been previously blocked by cytochalasin B, did not exceed 1% of the maximum 2-DG transport in the presence of insulin. For the determination of lipolytic activity, freshly isolated fat cells were incubated 90 min in a final volume of 0.5 ml of KRBHA without 2-DG or pyruvate, replaced by 6 mM glucose. Lipolytic activity was stopped by cooling the vials in an ice bath and allowing the fat cells to coalesce at the surface. Aliquots (300 μl) of the infranatant medium were then taken for enzymatic determination of glycerol and used as a lipolysis index as previously described (Carpéné et al., 1994).

Results

Immunological and Biochemical Detection of SSAO in Human Subcutaneous Adipose Tissue.

The use of a rabbit antibody directed against bovine lung SSAO (Lizcano et al., 1998) allowed the detection of an immunoreactive band in various human tissues (Fig.1). Under reducing conditions, the apparent molecular weight of the band was approximatively 98 kDa in all the tissues tested [i.e., similar to the typical apparent molecular weight of SSAO protein reported in pigs (Raimondi et al., 1992), rats (Enrique-Tarancon et al., 1998), and cattle (Lizcano et al., 1998)]. Spleen and kidney expressed low amounts of SSAO, whereas clear-cut positive signals were found in the intestine, lung, aorta, pancreas, and adipose tissue. The fact that, in humans, like in rodents, adipose tissue belongs to the group of tissues that highly express SSAO and prompted us to study further the activity of human adipose SSAO. In abdominal adipose tissue homogenates, the oxidation of [¹⁴C]benzylamine, a well known SSAO substrate, was resistant to the irreversible MAO inhibitor pargyline, but was quite completely inhibited by 1 mM semicarbazide (Fig.2A). More than 90% of benzylamine oxidation could be therefore attributed to a SSAO activity in human fat depots. Moreover, semicarbazide inhibited in a dose-dependent manner the oxidation of 0.5 mM benzylamine with an IC₅₀value of 80 ± 12 μM, but it was not the most potent inhibitor of SSAO since hydralazine and aminoguanidine, already described as SSAO inhibitors (Lyles, 1996), exhibited IC₅₀ values of 0.5 ± 0.1 and 38 ± 7 μM, respectively (Fig. 2B). The selective MAO-B inhibitor selegiline was totally inefficient, as was the MAO-A inhibitor clorgyline (not shown).

The SSAO of Human Subcutaneous Adipose Tissue Is Located at the Fat Cell Surface.

To assess whether the [¹⁴C]benzylamine oxidation observed in adipose tissue was due to the adipocytes themselves and not to other cell types present in this tissue (fibroblasts, nerve endings, and vascular cells), we compared the amine oxidase activity in tissue homogenates and isolated fat cell preparations. Both shared the same maximal oxidation velocity when activity was expressed as nmol of benzylamine oxidized/min/100 mg of lipids (Fig. 3):V _max was 14.0 ± 1.3 versus 15.0 ± 2.3 for tissue homogenates and isolated fat cells, respectively (n = 5, N.S.). Thus, adipocytes that contained all the lipids stored in adipose tissue also contained all the SSAO activity. The K _m values were 163 ± 24 and 482 ± 88 μM for homogenates and adipocytes, respectively (n = 5, p < 0.02). TheseK _m values for benzylamine were in the range previously described for human SSAO in adipose tissue (650 μM) (Raimondi et al., 1992) or other organs (110–222 μM) (Lyles, 1996).

To evaluate further the subcellular location of fat cell SSAO, we investigated whether adipocyte membranes were able to oxidize benzylamine. The V _max found in membrane preparations was 99 ± 3 nmol/mg of protein/min. It was higher than those obtained in isolated cells or tissue homogenates, which were 38.2 ± 9.5 and 8.6 ± 1.1 nmol/mg of protein/min, respectively (n = 5). This argued for a subcellular location of SSAO on the plasma membrane, as previously demonstrated for rat fat cells (Morris et al., 1997; Enrique-Tarancon et al., 1998). Although benzylamine is not naturally occuring in the organism, methylamine has been proposed as a putative physiological substrate for SSAO (Yu, 1998). Figure 4 shows that oxidations of unlabeled benzylamine and methylamine by adipocyte membranes were very similar, when SSAO activity was measured according to the spectrophotometric detection of hydrogen peroxide (Holt et al., 1997). In these conditions, the Michaelis constantK _m was 64 ± 9 and 140 ± 27 μM for benzylamine and methylamine, respectively, (n= 3, N.S.). Other biogenic amines were also oxidized by human SSAO, although to a lesser extent (e.g., octopamine and tryptamine), whereas histamine and acetylspermine were not oxidized at all (not shown).

Comparison of the SSAO Activity between Abdominal and Mammary Fat Depots.

It is widely accepted that there are strong differences in various biological parameters between the different human adipose tissues, depending on their anatomical location (e.g., density in adrenoceptors) (Mauriege et al., 1995). Thus, we compared SSAO activity in two anatomical locations: namely the abdominal and the mammary subcutaneous fat depots. Oxidation of benzylamine (from 0.02 to 2.5 mM) was studied in the homogenates of abdominal adipose tissues from a group of 15 subjects and in mammary fat depots from a matched group (n = 15) with no significant difference in age (43 ± 2 versus 40 ± 3) or BMI (24.7 ± 0.7 versus 26.5 ± 1.2). The K _m was 193 ± 16 and 198 ± 18 μM for abdominal and mammary depots, respectively (N.S.). No difference was detected in theV _max values of SSAO: 10.6 ± 1.6 versus 8.5 ± 1.1 nmol of benzylamine oxidized/mg of protein/min. Western blot analysis did not reveal a further difference in the expression or electrophoretic mobility of SSAO protein between the two fat depots (Fig. 5A). In keeping with this, there was no apparent influence of the anatomical location on the abundance of the SSAO mRNA in fat cells (Fig. 5B), and the apparent size of human SSAO mRNA was in agreement with that found for VAP-1 in human vascular tissue: 4.0 to 4.2 kb (Smith et al., 1998).

Effect of Amine Oxidase Substrates on Glucose Transport in Human Isolated Fat Cells.

Since SSAO generates hydrogen peroxide during the oxidative deamination reaction (Lyles, 1996), and since hydrogen peroxide stimulates glucose transport in adipocytes (Ciaraldi and Olefsky, 1982), we tested whether amine substrates could mimic the action of insulin on glucose transport in human adipocytes. Increasing concentrations of benzylamine or methylamine induced a dose-dependent stimulation of 2-[1,2-³H]deoxyglucose transport (Fig. 6). Stimulation of glucose uptake was detected with 10 μM of both amines, and maximal effect was reached at a millimolar dose (i.e., in the range of concentrations undergoing oxidation by intact adipocytes). The maximal stimulation of 2-DG uptake by SSAO substrates was partial: it accounted for one-third of the maximal insulin effect, whatever the anatomical location of the subcutaneous adipocytes, mammary or abdominal. Complementary experiments showed that benzylamine did not hamper insulin effect on glucose uptake (K _act of insulin was 2.7 ± 0.8 and 4.1 ± 1.2 nM, without and with 0.1 mM benzylamine, respectively, n = 6) and that there was hardly a detectable additivity of their respective action (100 nM insulin + 0.1 mM benzylamine accounted for 120 ± 7% of the maximal insulin stimulation). A more powerful insulin-like effect of amines has been previously reported in rat adipocytes, but only in the presence of vanadate (Marti et al., 1998; Enrique-Tarancon et al., 2000). When vanadate was tested on human adipocytes, the human fat cells behaved differently from rat adipocytes: the addition of 0.1 mM vanadate for 45 min did not affect basal or insulin-stimulated glucose transport (vanadate effect was equivalent to 8 ± 5% alone or to 103 ± 10% with insulin, when basal is set at 0% and 100 nM insulin at 100%). Vanadate potentiated the effect of 1 mM hydrogen peroxide, but failed to potentiate benzylamine or methylamine effects (not shown). In this regard, human adipocytes showed similarities with 3T3-L1 adipocytes, which require longer times of exposure to vanadate to observe a synergism with SSAO substrates on glucose uptake (Enrique-Tarancon et al., 2000).

Inhibition of Benzylamine Effect on Glucose Transport by Amine Oxidase Inhibitors and by Antioxidants.

To demonstrate the involvement of SSAO-dependent mechanism in the amine-induced stimulation of glucose transport, we tested several amine oxidase inhibitors and hydrogen peroxide scavengers. Semicarbazide altered the effect of 0.1 mM benzylamine alone, although it was ineffective on 2-DG transport in basal (vanadate 0.1 mM) or in stimulated conditions (100 nM insulin) (Fig. 7). The MAO inhibitor pargyline did not significantly inhibit the response to benzylamine when tested alone at 1 mM, whereas its combination with semicarbazide totally blocked the benzylamine stimulation without affecting the insulin-dependent transport. Noteworthy, wortmannin, an inhibitor of phosphatidylinositol 3-kinase (PI3-kinase), dose dependently inhibited the transport promoted by 100 nM insulin or 0.1 mM benzylamine (not shown) to reach a total blockade when tested at 1 μM (Fig. 7). This suggests that, among the diverse intracellular events activated by insulin or SSAO substrate, PI3-kinase activation represents a common mechanism leading to glucose uptake.

The addition of reduced glutathione or catalase to the incubation medium did not modify basal or insulin-stimulated glucose transport but totally hampered the effect of hydrogen peroxide (either exogenously added or endogenously generated during benzylamine oxidation) (Fig.8). The addition of 1 mM oxidized glutathione did not inhibit the benzylamine action on transport (not shown). The antioxidant N-acetyl-cysteine was also without effect on basal glucose transport, but inhibited the stimulation by hydrogen peroxide or benzylamine (Fig. 8). Similar results were obtained with 1 mM methylamine (not shown), arguing that the hydrogen peroxide generated during SSAO-dependent oxidation of substrates was very likely the cause of the observed increase of glucose uptake.

Correlation between the Capacity of Amines to Be Oxidized and to Stimulate Glucose Transport in Human Adipocytes.

Diverse amines were tested in parallel for oxidation by adipocyte membranes and for stimulation of glucose transport in intact fat cells. Figure9 shows that, among the substrates tested at 1 mM, there was a close correlation between their ability to be oxidized and their capacity to stimulate glucose uptake. Methylamine and benzylamine were the most efficient in both generation of hydrogen peroxide and stimulation of glucose transport, reaching one-third of the maximal effect of insulin. On the contrary, histamine was a poor amine oxidase substrate and did not alter glucose transport in human subcutaneous adipocytes.

Antilipolytic Effect of Benzylamine and Methylamine in Human Fat Cells.

Insulin is not only known for its stimulation of glucose transport, but also for its antilipolytic action. Whether SSAO substrates are also able to mimic this insulin effect was tested on the lipolysis stimulated by 1 mM 3-isobutyl-1-methylxanthine (IBMX). Since peroxovanadium compounds are also able to inhibit lipolysis in human fat cells (Eriksson et al., 1996), the study was conducted with and without 0.1 mM sodium orthovanadate. As shown in Fig.10, vanadate did not modify the antilipolytic effect of benzylamine and methylamine. However, inhibition of lipolysis is not only an insulin-dependent effect, since many other agents (e.g., α₂-adrenergic agonists, purinergic agents) are antilipolytic (Lafontan et al., 1996). We tested such hypothesis, since benzylamine or methylamine could have exerted some direct adrenergic or purinergic effects. Benzylamine and methylamine were unable to stimulate lipolysis, hence devoid of β-adrenergic agonist properties. In addition, blockade by the SSAO inhibitors semicarbazide and hydralazine and the inefficiency of the α₂-adrenergic antagonist RX 821002 brought evidence that the benzylamine-dependent antilipolysis was not related to α₂-adrenergic activation and that the antilipolytic effect of the α₂-agonist UK 14304 was not altered by SSAO inhibition (Fig.11). The amine-dependent inhibition of lipolysis was unaltered by the addition of adenosine deaminase (8 IU/ml), which hampers the antilipolyic effect of adenosine: the lipolysis stimulated by adenosine deaminase + IBMX (0.78 ± 0.03 μmol of glycerol/100 mg of lipid/90 min, n = 7) was partially inhibited by 0.1 mM benzylamine (0.57 ± 0.05) and returned to basal with 1 mM amine (0.30 ± 0.02), making unlikely the involvement of endogenous adenosine or direct activaction of A1-purinergic receptors in the benzylamine antilipolytic action. There was no difference between mammary or abdominal adipocytes regarding the effects of benzylamine on lipolysis (not shown).

Determination of SSAO Activity in Small Biopsies of Human Subcutaneous Fat Depots.

In humans, the most studied SSAO is the circulating soluble form because it can be easily determined in blood samples. Although its exact origin is still unknown, numerous clinical investigations have shown variations of SSAO plasma levels in many diseases (Lyles, 1996) (Boomsma et al., 1999), such as an increase in type I (Ekblom, 1998) and type II diabetes (Boomsma et al., 1999;Garpenstrand et al., 1999; Meszaros et al., 1999). To verify whether the human adipocyte SSAO could also be a candidate as a biological marker, we attempted to determine its kinetic parameters on small pieces of subcutaneous adipose tissue, weighing only around 100 mg (i.e., similar to that obtained by needle microbiopsy) (De Glisezinski et al., 1998). To obtain sufficient data with such small samples, kinetic parameters were calculated from double-reciprocal plot analyses of [¹⁴C]benzylamine oxidation at three different concentrations only. Under these conditions,K _m was 260 ± 38 μM andV _max was 7.5 ± 1.2 nmol/mg of protein/min in homogenates from small biopsies (corresponding to a total of 1.7 ± 0.2 mg of proteins, n = 12). These estimates were not different from that obtained with larger samples of around 1 g of wet tissue tested in parallel (containing 16 ± 1 mg of proteins), in which K _m was 218 ± 38 μM and V _max was 6.5 ± 1.2 nmol/mg of protein/min (n = 12, N.S., paired t test). Thus, data obtained with subcutaneous microbiopsies appear to be accurate enough for the in vitro estimation of adipose tissue SSAO activity.

Discussion

Our study demonstrates that SSAO is highly expressed in human subcutaneous fat depots and that in vitro oxidation of several amines by isolated adipocytes led to the stimulation of glucose transport and inhibition of lipolysis. The membrane-bound and the soluble SSAOs are known, like MAOs, as scavengers for biogenic or dietary amines and have been associated with the generation of more or less toxic products, such as ammonia, aldehydes, and hydrogen peroxide (Callingham et al., 1995; Lyles, 1996; Yu, 1998). However, three novel functions have been recently associated with SSAO. First, VAP-1, present at the endothelium surface and involved in lymphocyte binding (Bono et al., 1998), has been shown to be identical (product from the same gene) to human placenta SSAO (Zhang and McIntire, 1996) and to exhibit amine oxidase activity, although it is not known whether the oxidase activity is necessary for the adhesion properties (Smith et al., 1998). Second, the oxidation of diverse substrates by SSAO present in intracellular vesicles and at the cell surface of rat adipocytes (Enrique-Tarancon et al., 1998) has been involved in the potentiation of vanadate effects on glucose transport (Marti et al., 1998). Third, SSAO has been proposed as late marker of adipocyte differentiation of murine preadipocyte lineages (Moldes et al., 1999). Our present findings further strengthen the view that adipocyte SSAO is relevant in the regulation of glucose and lipid metabolism.

In fat cells and adipose tissue homogenates, most of the oxidation of benzylamine was sensitive to semicarbazide inhibition. This is in agreement with the fact that: a) benzylamine is considered mainly as a SSAO substrate, although it can also be oxidized by MAO-B (Yu, 1986); b) MAO-B is largely less expressed than MAO-A in human adipocytes (Pizzinat et al., 1999). The high level of SSAO activity found in human subcutaneous adipose tissues has been already reported in a comparative study in which a higher V _max for SSAO was observed in human rather than in mouse, rabbit, or pig white adipose tissues (6.6 versus 0.4–1.6 nmol of benzylamine oxidized/mg of protein/min) (Raimondi et al., 1992). However, our study describes for the first time SSAO activity in intact human fat cells and its consequences on metabolism. Benzylamine oxidation was characterized by similar V _max values in both tissue homogenates and intact adipocytes when expressed per milligram of lipids, indicating that, among the diverse cell types present in adipose tissue, adipocytes that contain most of the lipids stored also contained most of the SSAO activity. However, maximal benzylamine oxidation was higher in adipocyte crude membranes than in cell suspensions or in tissue homogenates, when calculated on a per milligram of protein basis. This enrichment is consistent with the location of SSAO in the membrane fraction as already demonstrated in rat adipocytes (Morris et al., 1997; Enrique-Tarancon et al., 1998).

Adipose tissue does not belong to the group of organs involved in the clearance of circulating amines, like liver, kidney, lung, and gut and justifies search for novel functions of adipocyte SSAO. Besides their scavenger action on biogenic or dietary amines, amine oxidases have been involved in oxidative stress and apoptosis in various cell types via their oxidation products (Callingham et al., 1995; Yu, 1998). However, concerning white adipocytes, the literature on apoptosis and oxidative stress is very scarce, whereas hydrogen peroxide is known for its insulin-mimicking effect (Ciaraldi and Olefsky, 1982). In fact, it has been repeatedly reported that millimolar doses of hydrogen peroxide stimulate glucose transport in rat adipocytes (Hayes and Lockwood, 1987; Marti et al., 1998), but metabolic responses to endogenously produced H₂O₂ have never been reported so far in human fat cells. We have already observed an insulin-like effect of SSAO substrates in rat adipocytes, but only in the presence of 0.1 mM vanadate (Enrique-Tarancon et al., 1998; Marti et al., 1998). The present results demonstrate that, in human adipocytes, benzylamine and methylamine stimulate glucose transport when tested alone, and that the addition of vanadate does not further enhance this effect. The reason for this difference in vanadate sensitivity is still unknown, but could probably be attributed to species-specific differences in the oxidative metabolism of vanadate or in the antioxidant defenses of fat cells.

Evidence for an involvement of oxygen reactive species in the effect of SSAO substrates on hexose uptake was supplied by the influence ofN-acetylcysteine, catalase, and/or glutathione that did not affect the basal glucose transport, but abolished the effects of H₂O₂, benzylamine, or methylamine. Comparing the inhibitions of benzylamine oxidation and benzylamine-induced glucose transport by pargyline and semicarbazide allowed further assessment that SSAO activity was involved in both phenomena. Taken together, these data suggest that an SSAO-dependent mechanism was predominantly involved in the benzylamine action on glucose transport. The presence of a small MAO-B activity found in human adipocytes (Pizzinat et al., 1999) and the partial loss in MAO selectivity of pargyline at 1 mM can likely explain the weak SSAO-independent effect of benzylamine on glucose transport. Thus, in the presence of adequate substrates, the continuous hydrogen peroxide production by human adipocyte SSAO was sufficient to reproduce the effect of high doses of exogenous hydrogen peroxide on glucose uptake. An indirect demonstration of the involvement of PI3-kinase activation in response to benzylamine was shown by the sensitivity to wortmannin. This is in accordance with the phosphorylation of insulin receptor substrates and the subsequent PI3-kinase stimulation we recently reported in rat adipocytes treated by tyramine and benzylamine (Enrique-Tarancon et al., 2000). Therefore, it can be assessed that several steps of the insulin-signaling cascade are activated under SSAO activation. However, the cellular targets primarily affected by the products of SSAO activity and leading to the partial insulin-mimicking effects are still unknown. Glucose transport is involved in lipogenesis, one of the major functions of the adipocyte, another one being lipolysis. Our attempt to verify whether benzylamine and methylamine was able to mimic the antilipolytic action of insulin showed that SSAO activity was also involved. Moreover, the lack of effect of the α₂-adrenergic antagonist RX 821002 on the benzylamine or methylamine-induced antilipolysis rendered very unlikely the involvement of an adrenergic mechanism in their antilipolytic effect. It can also be assessed that the release of endogenous adenosine and/or the stimulation of A1-purinergic receptors are not implicated in the antilipolytic effect of SSAO substrates, since benzylamine was able to inhibit the lipolytic effect of IBMX and adenosine deaminase (a purinergic receptor antagonist and a scavenger of adenosine, respectively).

Whether the SSAO present in human adipose tissue is physiologically relevant is an important question that deserves a better knowledge of the pharmacokinetics of trace amines, especially in postprandial states, since the consumption of dietary amines has been estimated to reach 37 mg/day in humans (Pfudstein et al., 1991). The novel interaction between SSAO substrates and adipocyte biology therefore makes it conceivable to quantify adipocyte SSAO in future clinical studies concerned with adipose tissue physiology, especially in the intra-abdominal fat pads that are the most exposed to changes in intestinal blood flow and nutrient level in postprandial states. Further information on this membrane-bound form of SSAO will be probably as useful as those collected on the circulating SSAO in diverse diseases, especially in diabetes (Ekblom, 1998; Bono et al., 1999; Boomsma et al., 1999; Garpenstrand et al., 1999; Meszaros et al., 1999) in which links were evidenced between glucose or lipid metabolism, oxidative stress, and vascular complications.

Our proposal of the determination of adipocyte SSAO in clinical investigation arises from three observations we made on human adipose tissue: a) SSAO is highly expressed in adipose tissue (activity per milligram of protein is greater than in plasma); b) substrates of SSAO stimulate glucose transport and inhibit lipolysis in isolated adipose cells; and c) there is no difference between abdominal and mammary subcutaneous fat depots regarding the SSAO content and the insulin-like effects of SSAO substrates. The technique we propose for SSAO assay on subcutaneous microbiopsies can be further miniaturized, but as it is, it could bring additional information to the diverse assays of soluble SSAO in blood samples already included in many clinical studies. In this context, it is important to note that the use of microbiopsies of adipose tissue has brought valuable information on the hormonal regulation of lipolysis and therapeutic approaches to obesity (De Glisezinski et al., 1998).

Finally, the interplay between SSAO activity and fat cell metabolism suggests that this enzyme may have beneficial and toxicological effects upon cellular function. Future investigations are needed to assess whether the most convenient therapeutic approach concerning SSAO activity in diabetes would be either to inhibit the increased circulating SSAO activity (to limit the putative oxidative stress at the vasculature level) or activate the SSAO activity with selected substrates to improve the glucose uptake in adipose tissue and other peripheral tissues.