Imidazolinic radioligands for the identification of hamster adipocyte α2-adrenoceptors

European Journal of Pharmacology, 171 (1989) 145-157

145

Elsevier EJP 51042

Imidazolinic radioligands for the identification of hamster adipocyte ct2-adrenoceptors Jean-Srbastien Saulnier-Blache, Christian Carprnr, Dominique Langin 1 and Max Lafontan L N.S.E.R.M., Unit~ 317, Institut de Physiologie, Universit~ Paul Sabatier, Rue Frangois Magendie, 31400 Toulouse, France and 1 National Institute of Agronomic Research, LN.R.A., Laboratoire de la Lactation, 63122 Ceyrat, France

Received 23 January 1989, revised MS received 15 August 1989, accepted 29 August 1989

Imidazolinic radioligands ([3H]UK 14304, [3H]idazoxan and [3H]RX 821002) were used for the identification of a2-adrenoceptors on hamster fat cell membranes since there are limitations to the use of [3H]yohimbine and [3H]clonidine, which suggest a2-adrenoceptor heterogeneity. Biological assays (lipolysis measurements) were performed on isolated fat cells and binding studies were carried out on fat cell membranes. The imidazolinic derivative, UK 14304, was a full agonist as compared to clonidine. Idazoxan and RX 821002 (2-(2-methoxy-l,4-benzodioxan-2yl)2-imidazoline), a recently developed a2-antagonist, were more potent t~2-antagonists than yohimbine in this fat cell model. [3H]UK 14304 was the most suitable agent for the quantification of the 'high-affinity state' aE-adrenoceptors in binding studies since it did not exhibit the sensitivity to the composition of the buffer shown by [3H]clonidine. Although it is a potent a2-antagonist , [3H]idazoxan had major limitations for use in the identification of a2-adrenoceptors in this cell model since it also bound to 'non-adrenaline displaceable' binding sites which were revealed when imidazolinic derivatives (phentolamine) were used instead of adrenaline to determine the non-specific binding. We demonstrated that [ 3H]RX 821002 was a more suitable radioligand than [ 3H]yohimbine for labelling hamster fat cell aE-adrenoceptors (K o = 1.0 _ 0.1 nM, Bmax = 776 5:60 fmol/mg protein). Moreover, since it exhibited low affinity for 'imidazoline-preferring sites', it represents a valuable ligand even in tissues possessing such binding sites. We suggest that [3H]RX 821002 can be used to identify a2-adrenoceptors in various tissues when these sites cannot be labelled with [ 3H]yohimbine. et2-Adrenoceptors; Adipocytes; Lipolysis; [3H]Yohimbine; [3H]UK 14304; [3H]Idazoxan; [3H]RX 821002; Fat ceils; 'Imidazohne-preferring sites'; (Hamster)

1. Introduction H a m s t e r white fat cells, like human, dog and rabbit (Lafontan et al., 1985) adipocytes, possess a 2 - a d r e n o c e p t o r s , the stimulation of w h i c h mediates an antilipolytic response (Garcia-Sainz and Fain, 1980; Schimmel et al., 1980; C a r p r n 6 et al., 1980). Extensive studies of the transduction

Correspondence to: M. Lafontan, INSERM U.317, Institut de Physiologie, Universit6 Paul Sabatier, Rue Francois Magendie, 31400 Toulouse, France.

mechanisms involved in the mediation of ol2adrenergic responsiveness have shown that a 2adrenoceptor stimulation promotes a reduction of intracellular c A M P levels (Garcia-Sainz and Fain, 1980) by inhibition of adenylyl-cyclase activity (Aktories et al., 1980) mediated by the Gi transducing protein (Martinez-Olmedo et al., 1984). The postsynaptic az-adrenoceptor of the hamster adipocyte has also been used as a pharmacological model to define the az-adrenergic properties of various drugs in biological assays (Carprn6 et al., 1983b; P u s h p e n d r a n and Garcia-Sainz, 1984). In contrast to the clear results obtained b y measuring

0014-2999/89/$03.50 © 1989 Elsevier Science Publishers B.V. (Biomedical Division)

146

a2-adrenergic antilipolytic effects, characterization of hamster fat cell a2-adrenoceptors with radioligand binding assays have provided less convincing results. The binding characteristics of the radioligands used were very different from those obtained with human or dog adipocytes and did not demonstrate the presence of an a2-adrenoceptor population similar to that found in dog or human tissues (Taouis et al., 1987b; Galitzky et al., 1989). In fact, neither the mixed al-a 2antagonist, [ 3H]dihydroergocriptine (Pecquery and Giudicelli, 1980), nor the selective a2-antagonist, [3H]yohimbine (Pecquery et al., 1984; 1988), displayed high-affinity binding to hamster fat cell membrane preparations since their K D values were around 20 nM. The only radioligand, to our knowledge, which exhibits high-affinity binding to hamster adipocyte membranes is [3H]clonidine (K D = 3 nM) (Carp4n6 et al., 1983a). However, the receptor population defined with this a 2agonist is greater than that defined with the previously described antagonists, a result which is in contradiction with the classical views on the identification of adrenoceptor affinity states (Hoffman et al., 1980). More suitable radioligands are therefore required for the characterization of hamster adipocyte a2-adrenoceptors. Moreover, these new radioligands could be used in other tissues when the a2-adrenoceptors cannot be accurately labelled with [3H]yohimbine. The reliable quantification of the adipocyte a2-receptor population actually requires the use of both agonist and antagonist radioligands for at least two reasons. First, the hamster adipocyte is a cell model used to study the hormonal regulation of a2-adrenoceptors, where the coupling of the a2-receptors with their transducing system and the affinity of physiological amines for the sites need to be investigated more deeply (Carprn6 et al., 1983a; Taouis et al., 1987a; Pecquery et al., 1988). Second, ct2-adrenoceptors are not a homogeneous class of receptors and can be divided into several subtypes which are distinguished mainly by their pharmacological properties defined in radioligand binding assays (Bylund, 1985; Lanier et al., 1988) and recently confirmed by genetic approaches (for reviews: Lefkowitz and Caron, 1988; Bylund, 1988). Many

a2-adrenoceptor agonists and antagonists have been synthetised over the last years, but only few of them are as yet available as radioligands. We performed the present study to test whether imidazolinic derivatives such as the aE-agonist, UK 14304 (Cambridge, 1981), and the antagonists, idazoxan and RX 821002 (Doxey et al., 1983; Dabir6 et al., 1983; Stillings et al., 1985), can provide suitable pharmacological tools for the characterization of hamster adipocyte a2-adrenoceptors. In this study we compared their relative potencies in biological assays and binding studies. We found that the imidazolinic derivative, UK 14304, was the best aE-agonist in this model and that it was the most suitable agent to use to study biological effects and to quantify 'high-affinity state' aE-adrenoceptors in binding studies. Idazoxan and RX 821002 were potent a2-antagonists. We demonstrated that only [3H]RX 821002 was a useful radioligand for further investigations of hamster adipocyte a2-adrenoceptors. [3H]Idazoxan had major limitations for aE-adrenoceptor identification in this species since it also bound to non-adrenaline displaceable sites, which could limit the accurate definition of a2-adrenoceptors.

2. Materials and methods

2.1. Animals All the studies were performed with male golden hamsters (Mesocricetus auratus), weighing between 110 and 140 g, that were housed at 20-22°C and received food ad libitum. The animals were killed by decapitation after overnight fasting. White adipose tissues from different fat deposits (perirenal, epididymal and s.c.) were immediately removed and pooled.

2.2. Preparation of isolated adipocytes and lipolysis measurements Adipocytes were isolated according to the method of Rodbell (1964) with minor modifications. The adipose tissue was cut into ~mall pieces and incubated for 30-40 min at 37°C, under vigourous shaking, in Krebs Ringer bicarbonate

147

buffer (pH 7.5) containing 35 mg/ml of bovine serum albumin, 6 mM glucose (KRBA buffer) and 1.6 mg/ml of collagenase. After collagenase digestion of the tissue, the adipocytes were separated from the stromavascular fraction by flotation and were washed 3 times in KRBA buffer. Isolated adipocytes (20-30 mg of total cell lipids) were dispersed in 1 ml of KRBA containing 10/~1 of the test agents. After 90 rain incubation at 37°C, the reaction was stopped on ice and an aliquot (200 #1) of the incubation medium was taken to determine glycerol according to the method of Wieland (1957). The et2-adrenergic antilipolytic effects were explored on adenosine deaminase-stimulated lipolysis; adenosine deaminase (1-2 #g/ml) was added to inactivate endogenous adenosine released during incubation and which could exert an antilipolytic action. The antilipolytic effects of et2-adrenoceptor agonists are expressed as percent inhibition of lipolysis. The lipid content of the incubation vials was determined gravimetrically after extraction according to the method of Dole and Meinertz (1960). Lipolytic activities are expressed in /~mol of glycerol released per 100 mg of total cell lipid per 90 min.

2.3. Preparation of crude adipocyte membranes and radioligand binding studies Crude membranes were obtained after hypotonic lysis of the adipocytes. The lysing medium was composed of 2.5 mM MgC12, 1 mM KHCO3, 2 mM Tris-HC1, 100 ~tM EGTA, pH 7.5 (35 mOsm), and contained the following protease inhibitors: leupeptin (1/zg/ml), benzamidine (0.1 mM), phenyl methyl sulfonyl fluoride (100/~M). Lysis of adipocyte suspensions was performed at 20-22 ° C in order to minimize trapping of membranes in the coagulating fat cake. Fat cell ghosts were separated from the fat cake by centrifugation (40000×g, 10 rain at 15°C). The pellet was resuspended in 1 ml of lysing medium and frozen immediately at - 8 0 °C until the binding studies were performed. et2-Adrenoceptor binding sites were identified with five different radioligands: [3H]clonidine,

[3H]yohimbine, [3H]UK 14304, [3H]idazoxan and [3H]RX 821002. Thawed crude membranes were diluted in a large volume (30 ml) of 50 mM Tris-HC1, 0.5 mM MgC12, pH 7.4 and centrifuged (40000 ×g, 10 min at 4 o C). The pellet was homogenized, filtered through a nylon mesh and adjusted to a final protein concentration of 4-6 mg/ml. The incubation medium consisted of 100/~1 of radioligand and 100 /~1 of membrane suspension made up to a final volume of 400/d with 50 mM Tris-HC1, pH 7.5, containing 0.5 mM MgC1z unless otherwise stated. Incubations were carried out at 25 °C in a water bath for 30 min under constant shaking. At the end of the incubation period, the suspensions were diluted in 4 ml of ice-cold 10 mM Tris-HC1 buffer and filtered through Whatman G F / C glass fiber filters on a Millipore manifold. The filters were washed twice with 10 ml portions of ice-cold buffer, placed in minivials containing 4 ml of scintillation cocktail and counted in a Packard beta counter (efficiency 4550%). For saturation studies, the radioligand concentrations ranged from 0.2 to 15-20 nM and the specific binding was determined as the radioactivity bound to the membranes (defined as total binding) minus the binding determined in the presence of 10 -5 M phentolamine or 2.10 -4 M (-)-adrenaline, depending on the assay. Binding parameters (Bmax, KD) were determined from the equilibrium binding studies by using a computerassisted linear transformation of the saturation binding data (Scatchard analysis, Hill plot (nil)) (Barlow, 1983). For competition studies, a fixed concentration (7-10 nM) of radioligand was used in the presence of increasing concentrations of different competitors (10-9-10 -4 M), the residual binding was expressed as the percent of total radioligand binding. The relative affinity of the different competitors were determined, with one- or two-site models, by computer-assisted calculations (Barlow, 1983). The equilibrium dissociation constant was calculated from the following equation: K i = ECs0/( 1 + 3 H.ligand/KD ligand) where ECs0 is the concentration of pharmacological agent causing 50% inhibition of 3H-ligand binding; 3H-ligand

148

is the concentration of ligand in the assay and K D ligand is the equilibrium dissociation constant obtained in saturation studies. The values presented are the means_ S.E. The significance of differences was tested with Student's t-test. The protein content of membrane preparation was measured with the method of Lowry et al. (1951) with bovine serum albumin as standard.

10

-log agonist (M) 9 7 6

I

0

I

"

L

I

5 I

I

=_o 2.4. Chemicals [O-Methyl-3H]yohimbine (specific activity 88 Ci/mmol) and [3H]idazoxan (60 Ci/mmol) were obtained from Amersham International (Amersham, UK). [3H]Clonidine (specific activity 48 Ci/mmol) and [imidazolyl-4,5-3H]UK 14304 (82.7 Ci/mmol) were obtained from New England Nuclear (Boston, MA). Yohimbine HC1, ( - ) adrenaline bitartrate, bovine serum albumin (Fraction V) were purchased from Sigma (St. Louis, MO). Phentolamine mesylate was obtained from Ciba Geigy (Basel, Switzerland). Idazoxan (RX 781094), RX 821002 and [3H]RX 821002 were kindly given by Reckitt & Colman (Hull, UK). UK 14304 and prazosin were obtained from Pfizer (Sandwich, UK). Clonidine HC1 and BHT 920 were kindly given by Boehringer Ingelheim (Reims, France). Guanylyl-imidodiphosphate (GppNHp), collagenase, adenosine deaminase and enzymes for the glycerol assay came from Boehringer Mannheim (Mannheim, FRG). Oxymetazoline was obtained from Schering (Kenilworth, N J). All other chemicals and organic solvents were of reagent grade.

3. R e s u l t s

3.1. Antilipolytic action of oarious ct2-adrenoceptor agonists on hamster adipocytes The different a2-adrenergic compounds used in the present study included several imidazoline derivatives (UK 14304, clonidine, idazoxan, RX 821002 (2-(2-methoxy-l,4-benzodioxan-2yl)-2-imidazoline) and phentolamine), as well as non-imidazolinic compounds ((-)-adrenaline, yohimbine and BHT 920).

ae

\ \

100 Fig. 1. Antilipolytic effect of different a2-adrenoceptor agonists on hamster fat cells. The adipocytes were incubated in the presence of 2 ~ g / m l adenosine deaminase and increasing concentrations of ( -)-adrenaline + propranolol 10- 5 M (A), clonidine (©), BHT 920 (m) or U K 14304 (e). Glycerol was measured as described in Materials and methods. The results are expressed as the percent inhibition of adenosine deaminase-induced lipolysis in the presence of the corresponding a2-agonist concentrations. Each point represents the mean _+S.E. of five separate experiments.

In order to determine the most powerful ot2agonist on the hamster fat cells, the relative potencies of various a2-adrenergic agents were compared by measuring their antilipolytic effect on adenosine-deaminase stimulated lipolysis. This experimental procedure constitutes a valuable model for the investigation of a2-adrenergic potencies, as described previously for human (Berlan and Lafontan, 1985) and hamster white fat cells (Carp6n6 et al., 1983a). As shown in fig. 1, the most potent aE-agonist tested was UK 14304: it exhibited the greatest maximal response (85__+ 3% inhibition at 10 -5 M, n = 5) and the lowest IC50 value (5.10 -8 M). Clonidine was slightly less active than UK 14304, especially at 10 -6 M ( 5 2 + 9 vs. 79_6%, P<0.05) whereas BHT 920 was the least potent agent (only 49 _ 7% inhibition at the higher dose used 10 -5 M). When 0.3 mM theophylline was used to stimulate lipolysis instead of adenosine deaminase, UK 14304

149

was again the most powerful a2-adrenoceptor agonist (data not shown). In addition, none of the following aE-agonists, B H T 933, xylazine, paraaminoclonidine, guanfacine or tramazoline, were as efficient as U K 14304 in this cell model (not shown). The physiological agonist, ( - ) adrenaline (10-9-10 -6 M), also induced dose-dependent antilipolysis when its fl-lipolytic component was blocked by 10 -5 M propranolol, with a maximal effect comparable to that of 10 -6 M clonidine. However, the catecholamine-induced antilipolysis was reversed at 10 -5 M. This biphasic response is due to the highly effective fladrenergic stimulation of lipolysis, which was only partially prevented by the r-antagonist at 10-5 M (fig. 1). To further characterize the aE-adrenoceptors implicated in this aE-dependent antilipolysis we also tested the blocking properties of various a-antagonists.

!

3.2. Blocking properties of a-antagonists on the a2-adrenergic antilipolytic response In order to determine the relative order of potency of the a-antagonists tested in this model, their capacity to counteract U K 14304-induced antilipolysis was measured. All the antagonists tested were able to completely abolish the response initiated by 10 -6 M U K 14304 (fig. 2). While phentolamine, RX 821002 and idazoxan were equipotent, with mean ICs0 values around 5 - 1 0 -8 M, yohimbine was less efficient, with an ICs0 of 5 . 1 0 -7 M. A similar rank order of effectiveness (phentolamine = RX 821002 = idazoxan > yohimbine) was also obtained when the compounds antagonized the antilipolysis induced by 10 -s M BHT 920, a non-imidazolinic drug (not shown). As the imidazolinic derivatives were at least as potent as the older, established adrenergic compounds, the binding properties of tritiated U K 14304, RX 821002 and idazoxan were studied and compared with the binding properties of the more classical a2-adrenergic radioligands: [3H]clonidine and [ 3H]yohimbine.

3.3. Identification of the hamster adipocyte a2-receptors by [ 3H] clonidine and [ 3H]yohimbine

.o

.E 50 E

I00

i9

8

7

I

I

6

5

-log antagonist (M) Fig. 2. Antagonism of the antilipolytic effect of U K 14304 by different aE-adrenoceptor antagonists on hamster fat cells. The adipocytes were incubated with 2 /~g/ml adenosine deaminase and 10 -6 M U K 14304 in the absence (100%) or presence of increasing concentrations of phentolamine (O), idazoxan (11), RX 821002 ( o ) or yohimbine (A). Glycerol was measured as described in Material's and methods. The results are expressed as percentages of the maximal inhibition obtained with U K 14304. Each point represents the mean of four different experiments, S.E. were deleted for clarity and were less than 10%.

The binding characteristics of the two most commonly used radioligands under the experimental conditions described previously by us for human fat cells (Lafontan et al., 1985), i.e. 25 min of incubation at 25 ° C in 50 m M Tris, 10 mM MgC12 buffer, pH 7.5, and with 10 -5 M phentolamine for definition of non-specific binding are shown in table 1. [3H]Clonidine labelled a homogeneous population of high affinity sites on adipocyte membranes, with a binding capacity of 180 f m o l / m g protein, whereas [3H]yohimbine exhibited a very weak binding affinity (K D higher than 10 nM), with a Bma~ similar to that of [3 H]clonidine. The binding characteristics of both ligands were also determined when the MgC12 concentration in the incubation buffer was reduced. Under these conditions, the maximal binding of [3H]clonidine was increased without any change in K D. Conversely, the maximal [3H]

150 TABLE 1 Hamster fat cell a2-adrenoceptor binding parameters determined with [3H]clonidine and [3H]yohimbine. Fat cell membranes were incubated with various concentrations of radioligands in 50 mM Tris 10 mM MgC12, pH 7.5 or 50 mM Tris 0.5 mM MgC12, pH 7.5, for [3H]clonidine, and 50 mM Tris 10 mM MgC12, pH 7.5 or 25 mM glycyl-glycine for [3H]yohimbine. Specific binding was determined as described in Materials and methods; non-specific binding was determined in the presence of 10 -3 M phentolamine. Bma~: total number of specific binding sites. KD: equilibrium dissociation constant. Non-specific percent (N.S.) at K D is expressed as a % of total binding. Each value is the mean 5: S.E. of n separate experiments. Student's t-test was used to compare the values obtained under the two different conditions for the same radioligand, a p < 0.001; b p < 0.01. Bmax KD (fmol/mg prot) (nM)

N.S.

n

(%)

[ 31t]Clonidine MgCI 210 mM 180+20 MgCI2 0.5 mM 468+76 a

2.35:0.2 2.95:0.5

305:5 10 16+4 b 6

[ 3HIYohimbine MgCl 210 mM Glycyl-glycine

1625:36 246 5:58

13.95:1.3 565:3 8.9 5:1.3 b 50 5:4

6 5

y o h i m b i n e specific b i n d i n g was n o t m o d i f i e d whatever the b u f f e r used b u t the affinity of [3 H ] y o h i m b i n e was significantly i n c r e a s e d b y the use of 25 m M glycylglycine buffer, as r e p o r t e d previously for a n o t h e r cell system ( T u r n e r et al., 1985). W h a t e v e r the i n c u b a t i o n c o n d i t i o n s used for the b i n d i n g assays, the Bm~ value o b t a i n e d with the a n t a g o n i s t ([3H]yohimbine) was never higher t h a n the Bmax value d e f i n e d with the agonist [3H]clonidine. In a d d i t i o n , [ 3 H ] y o h i m b i n e b i n d i n g was c h a r a c t e r i z e d b y a high p r o p o r t i o n ( > 50%) of non-specific binding. In an a t t e m p t to see w h e t h e r the non-specific b i n d i n g of [3H]yohimbine, as r o u t i n e l y assessed b y the residual b i n d i n g in the presence of 10 -5 M p h e n t o l a m i n e , was misleading, we tested the d i s p l a c i n g capacities of various c o m p e t i n g agents t h a t h a d a different c h e m ical nature. N e i t h e r U K 14304, clonidine, a d r e n aline, y o h i m b i n e , i d a z o x a n n o r R X 821002 (at 10 -5 or 10 - 4 M) were a b l e to d i s p l a c e m o r e t h a n 50% of the total [ 3 H ] y o h i m b i n e b i n d i n g ; all the agents gave similar non-specific b i n d i n g values. A parallel s t u d y of [3H]clonidine b i n d i n g showed that all the c o m p e t i n g agents m a x i m a l l y d i s p l a c e d

70-80% o f the t o t a l [3H]clonidine b o u n d shown).

(not

3.4. Identification of a2-adrenoceptors with the imidazolinic agonist, [3H] UK 14304 S a t u r a t i o n b i n d i n g studies with [ 3 H ] U K 14304 were p e r f o r m e d . T h e a g e n t c o m m o n l y used to d e f i n e n o n - s p e c i f i c b i n d i n g , p h e n t o l a m i n e , was o m i t t e d since it is an i m i d a z o l i n i c derivative. T h e p a r t of the t o t a l b i n d i n g which was d i s p l a c e d b y 200 # M a d r e n a l i n e was d e f i n e d as specific b i n d i n g at a 2 - a d r e n o c e p t o r s (fig. 3). T h e non-specific b i n d i n g r e p r e s e n t e d , u n d e r these c o n d i t i o n s , 35% o f the t o t a l b i n d i n g at the K D values a n d was p e r f e c t l y l i n e a r (fig. 3). T h e s a m e value was o b t a i n e d after the a d d i t i o n of p r o t e c t i v e agents (pargyline, catechol, a s c o r b i c acid) or 10 -5 M p r o p r a n o l o l with a d r e n a l i n e . T h e specific b i n d i n g was saturable, a n d the c a l c u l a t e d Bm~x a n d K D values o b t a i n e d f r o m S c a t c h a r d analysis were similar to those o b t a i n e d w i t h [3H]clonidine in the s a m e b u f f e r ( t a b l e 1). T h e S c a t c h a r d p l o t s were

't

~.40 *\.

.; ~.

201_. \ 800 0

~

600

o

400

250 50f] Bound

200 -

0

,I,

i

I

I

I

i

I

2

4

6

8

10

12

3H-UK

14304

I 14

(nM)

Fig. 3. [3H]UK 14304 specific binding to hamster fat cell membranes. Fat cell membranes were incubated with increasing concentrations of [ 3H]UK 14304. Non-specific binding was determined with 200 #M adrenaline. Total (A) and non-specific (m) [3H]UK 14304 binding curves. Each point is the mean + S.E. of five different experiments. The binding parameters obtained by Scatchard plot analysis were: Bmax=438+40 fmol/mg protein, KD= 2.21+0.36 nM, n = 5. Inset: Scatchard plot of one representative experiment (Bm~ = 462 fmol/mg protein, K D = 2.43 nM).

151 non-adrenaline-displaceable binding sites labelled with [3H]UK 14304 on hamster fat cell membranes. Displacement of the non-adrenaline-displaceable binding of [3H]UK 14304 by the a 2adrenergic compounds tested gave the following rank order of potency (based on their relative affinity and maximal inhibition of binding): U K 14304 > idazoxan > clonidine = RX 821002 = phentolamine > yohimbine (fig. 4).

100-

~ 00

\3o, o\ \

O

,e 60

?

~ 40

3.5. Identification of ae-adrenoceptors with the imidazolinic antagonists [3H]idazoxan and [3H]RX 821002

2o

1

I

I

I

l

I

8

8

7

6

5

4

-log

competitor

(M)

Fig. 4. Competition of [3H]UK 14304 binding to hamster fat cell membranes by different az-adrenergic agents. Fat cell membranes were incubated with 5-6 nM [3H]UK 14304 in the absence (100%) or presence of increasing concentrations of (-)-adrenaline (Or), yohimbine (A), phentolamine (O), clonidine (zx),idazoxan (It), UK 14304 (n) or RX 821002 (O). The results are expressed as the percent of total [3H]UK 14304 binding. Each point is the mean of three different experiments; S.E. were deleted for clarity and were less than 5%. linear (fig. 3) and the Hill coefficient was not different from unity (0.98 + 0.01), indicating that [3H]UK 14304 bound to only one class of binding site. When 10 m M MgC12 instead of 0.5 m M MgC12 was used in the 50 mM Tris binding buffer, the [3H]UK 14304 binding parameters were not modified (data not shown). We tested the capacity of different a2-adrenergic compounds to inhibit [3H]UK 14304 binding to hamster fat cell membranes (fig. 4). The relative order of IC50 values estimated with reference to the non-specific binding obtained with 10 -4 M adrenaline (i.e. the a2-adrenoceptor sites) was: U K 14304 = idazoxan > clonidine = RX 821002 > adrenaline > phentolamine >> yohimbine. Yohimbine appeared to be the least effective competitor at a2-adrenoceptors labelled by [3H]UK 14304. In competition assays, all the imidazolinlc compounds ( U K 14304, idazoxan, phentolamine, RX 821002) exhibited a maximal inhibition of [3H]UK 14304 binding that was higher than that produced by adrenaline. These results reveal the existence of

Saturation isotherms obtained with the two recently developed radiolabelled a2-antagonists, [3H]idazoxan and [3H]RX 821002, on hamster fat cell membranes (same buffer conditions and same batches of membranes) are shown in fig. 5. The non-specific binding measured in the presence of 200/~M adrenaline (in order to displace only the ligand bound to a2-adrenoceptor sites) was linear and represented 10% of the total binding at 10 nM [3H]RX 821002 (fig. 5B). The non-specific binding curve for [3H]idazoxan was mainly curvilinear (fig. 5A) and the non-specific binding reached 67% of the total binding at 10 nM. Specific [3H]RX 821002 binding was saturable, reversible and of high affinity (fig. 5B). Scatchard plots of [3H]RX 821002 specific binding were linear (correlation coefficient not different from unity) and allowed us to calculate the binding parameters, Bmax = 776 + 60 f m o l / m g protein, Kt) = 1.0 + 0.1 riM; n = 8. In contrast, saturation was not clearly reached with [3H]idazoxan and it was impossible to calculate accurately the binding parameters from an analysis of the Scatchard plots. The capacities of different a2-adrenergic agents to inhibit [3H]idazoxan (fig. 6A) and [3H]RX 821002 (fig. 6B) binding to hamster fat cell membranes were tested. The compounds inhibited the binding of the two radioligands with very different patterns. Maximal displacement of [3H]RX 821002 total binding was around 90% whatever drug used. Displacement of radioligand binding with adrenaline, corresponding to the specific inhibition of binding on a2-adrenoceptors, showed a maximal inhibition of only 30% with [3H]idazoxan and 90%

152 total

•

[] n o n - s p e c i f i c

1400-

•

1200-

~

I0005

E

iooo-

~ 8oo-

+

800-

600-

600-

400-

~ 400-

,/'

,/T-

-o

~'

~'°°-I ~. ~0~s01 ",.%

P,

~, X zoo-

ZOOc]

O! 0

, 406 Bound

12

14

",, 800

"I"

0 2 3H-Idazoxan

4

6

(nM)

8

10

3H-RX821002

~M)

Fig. 5. Saturation curves for [3H]idazoxan and [3H]RX 821002 binding to hamster fat cell membranes. Fat cell membranes were incubated with increasing concentrations of [3H]idazoxan (A) or [3H]RX 821002 (B). Total (zx,A) and non-specific binding (D,I) were determined as described in Materials and methods. Each point represents the m e a n + S.E. of four (A) and eight (B) different experiments. The binding parameters obtained by Scatchard plot analysis of [3 H]RX 821002 binding were: Bmax = 776 + 61 fmol/mg protein, K D = I . 0 0 + 0 . 0 6 nM, n = 8 . The inset in (B) shows the Scatchard plot of one representative experiment (Bmax ~ 782 fmol/mg protein, K D = 1.36 nM). B

A I00-

100-

..-,.

8o-

"0 ._= r~

%-7-,.

80-

0 0

x~

oJ 00

60-

\

I "1-

40-

40-

2 20-

20-

O--

60-

X rr

I

9

I

I

8 7 6 5 -log competitor (M)

I

I

I

4

0-

I

/

9

8

I

I

T 6 5 -log competitor (M)

I

I

4

Fig. 6. Inhibition of [3H]idazoxan and [3H]RX 821002 binding by different a2-adrenergic competitors. Fat cell membranes were incubatd with 8-10 nM [3H]idazoxan (A) or 9-10 nM [3H]RX 821002 (B) in the absence (100%) or in presence of increasing concentrations of (-)adrenaline (*), phentolarnine (o), clonidine (zx), RX 821002 (o) idazoxan (11), UK 14304 (D) or yohimbine (A). Binding was determined as described in Materials and methods. The results are expressed as the percent of total [3H]idazoxan or total [3H]RX 821002 binding. Each point is the mean of three different experiments. S.E. were deleted for clarity and were less than 5%.

153

with [3H]RX 821002. The relative order of potency of the various pharmacological agents defined from the IC50 values (obtained with 200/~M adrenaline for non-specific binding) was: U K 14304 > RX 821002 > idazoxan > phentolamine > clonidine > adrenaline > yohimbine for [3H]RX 821002. As expected the agonists (UK 14304 and adrenaline) gave shallow competition curves (n H - 0.5) while the antagonists have steeper curves (n H not different from 1). Yohimbine appeared to be the least effective a2-adrenergic competitor. With [3H]idazoxan as radioligand the relative order of potency was: RX 821002 > phentolamine = U K 14304 = idazoxan > adrenaline = yohimbine. For both radioligands, yohimbine appeared to be the least effective competitor at et2-adrenoceptors. With [3H]idazoxan as radioligand all the imidazolinic competitors (UK 14304, idazoxan, phentolamine, clonidine, RX 821002) produced a maximal inhibition of total [3H]idazoxan binding that was always greater than that obtained with adrenaline, as was also observed with [3H]UK 14304 (fig. 4). Such a result suggested the ex-

istence of non-adrenaline-displaceable sites labelled with [3H]idazoxan (fig. 6A), which were not revealed with [3H]RX 821002 (fig. 6B). The binding of [3H]idazoxan to the non-adrenalinedisplaceable binding sites was displaced completely, with high affinity, by U K 14304 and idazoxan and with a lower affinity by clonidine, phentolamine and RX 821002. The competition of [3H]idazoxan binding with RX 821002 and phentolamine appeared to be biphasic, with a plateau corresponding to the level of maximal inhibition obtained with adrenaline. An accurate study of the displacement of [3H]RX 821002 by the full agonist, U K 14304, was undertaken in the absence and presence of 10 -4 M G p p N H p to confirm the existence of GppNHp-sensitive, high-affinity sites for the agonist (fig. 7A). The results were analysed by computer-assisted curve fitting for one- and twosite models. As predicted by the Hill coefficient values (n H = 0.55 + 0.01), the analysis of the data obtained with U K 14304 alone was best fitted to a two-site model and allowed clear distinction be-

B

:00"

80-

~11.1~ : t TT.'~'t~ ::

100-

o

~ r T ~'~;~\:UK+GppNHp

60-

I '-.l_f.....

t~

J'o\ \

I\ '~,

\\Yo. \

_o ,O.

~< rc

Oxymet ? \ \

\

40-

20-

0-1

"6

}

-log competitor(M)

;

200

,

,

,

-logcompetitor(M)

Fig. 7. Comparative inhibition of [3H]RX 821002 binding by oxymetazoline, yohimbine, prazosin and U K 14304. Fat cell membranes were incubated with 5 n M (A) or 10 n M (B) [3H]RX 821002 in the absence (100%) or in presence of increasing concentrations of different competitors. (A) U K 14304 alone (O) or in the presence of G p p N H p 10 - 4 M (A). (B) oxymetazoline (o), yohimbine ( - ) or prazosin (zx). K i values (nM) were 2.6 +0.9 (oxymetazoline), 33.7 + 9.6 (yohimbine), 2261 + 112 (prazosin). The results are expressed as the percent of total [ 3H]RX 821002 binding. Each point is the m e a n + S.E. of three different experiments. A Student's t-test was used to compare the values with and without G p p N H p ( * * * P < 0.001; * * P < 0.01).

154

tween high (RH)- and low (RL)-affinity state receptor. The K i value of U K 14304 at the high-affinity state receptor (Kin) was 1.13 _+ 0.13 nM whereas that at the low-affinity state receptor (KiL) was 120.8 _+ 11.4 nM. The addition of GppNHp caused, as expected, a shift to the right of the competition curve, which became steeper (n H non-significantly different from unity). The [3H]RX 821002 displacement curve for U K 14304 in the presence of GppNHp was best fitted to a one-site model (K i = 63.2 + 4.0 nM). The capacity of oxymetazoline (an Ot2g-Subtype adrenergic agent) to displace [3H]RX 821002 binding on hamster fat cell az-adrenoceptors was compared to that of yohimbine and prazosin (fig. 7B). The affinity of each agent (K i values) was determined with 200/~M adrenaline to define the non-specific binding. Oxymetazoline was the best competitor, with an affinity 13 times greater than that of yohimbine and 870 times greater than that of prazosin. Prazosin had a very low capacity to displace [3H]RX 821002 binding.

4. Discussion

The aim of the study was to use new a2-adrenergic radioligands to correctly identify the a2adrenoceptor population of hamster fat cells, which is apparently different from that defined in humans, and to study various imidazolinic derivatives. The comparative approach with selective agonists and antagonists focused attention on the use of the new selective mE-antagonist, [3H]RX 821002, for the identification of the total number of az-adrenoceptors in a tissue that has a limited capacity to bind [3H]yohimbine. Moreover, the conditions for the use of [3H]clonidine and [3H]UK 14304 were also clarified. Finally, we demonstrated that [3H]UK 14304 and [3H]idazoxan also labelled a large population of non-adrenaline-displaceable binding sites, which could lead to the misinterpretation of binding data if nonspecific binding values are badly defined. For studies on receptor regulation and function, it is important to be able to identify highand low-affinity states of the az-adrenoceptor population by the use of suitable labelled agonists

and antagonists. The agonist labels the high-affinity state of the receptor while a suitable antagonist labels the whole population. Basically, the antagonist labels more sites than the agonist (Hoffman et al., 1980). This has not been clearly established with hamster adipocytes since the a 2antagonist, [3H]yohimbine, labels fewer sites than the agonist, [3H]clonidine. In the present study with hamster adipocytes, U K 14304 behaved as a full agonist as compared to clonidine (fig. 1), as shown in other cell models (Galitzky et al., 1989; Paris et al., 1989). [3H]UK 14304 binding was used to identify the high-affinity az-receptor population. To our knowledge, this is the first time this radioligand has been used to identity Ctz-adrenoceptors on hamster white fat cells. The Bmax values obtained with the full agonist were similar to those defined with [3H]clonidine with the same buffer conditions (table 1, fig. 3). [3H]Clonidine was used previously in our laboratory to identify the az-adrenoceptors on hamster fat cells (Carprn4 et al., 1983a). This radioligand exhibited high affinity binding and the Bmax values were evaluated around 200 f m o l / m g protein. In the present study we showed that this Bma~value could be increased (450 vs. 200 f m o l / m g protein) by using a 50 mM Tris buffer containing a lower concentration of MgC1 z as compared with that used previously (0.5 vs. 10 mM) (table 1). The MgC1 z-dependent variation of [3 H]clonidine binding was marked; it has also been described recently for human fat cells (Galitzky et al., 1989). The modification of the estimated Bm~, values could be ascribed to an alteration of the coupling between the Ctz-adrenoceptor and the Gi transducing protein at high Mg 2+ concentrations. The only az-adrenoceptor antagonist radioligand used until now to identify the whole population (high- and low-affinity states) of hamster fat cell az-adrenoceptors was [3H]yohimbine (Pecquery et al., 1984; 1988). The binding conditions used did not give very accurate results since this radioligand bound with a rather low affinity to membrane preparations (K D values reaching 20 nM). Moreover, the total number of az-adrenoceptors defined with [3H]yohimbine (mean Bmax values around 180-200 f m o l / m g protein) was always lower than that identified with the partial

155 and full agonist radioligands ([3H]clonidine or [3H]UK 14304) used in the present study. Such a result is in contradiction with the receptor occupation theory applied to /3- or a:-adrenoceptors, which predicts the labelling of more receptors by the antagonist than by the agonist. The possible involvement of the buffer composition was explored: when 25 mM glycyl-glycine buffer, proposed by Bylund et al. (1988), was used for [3 H]yohimbine binding assays instead of the commonly used 50 mM Tris, 10 mM MgC12 buffer, the affinity of the radioligand for the hamster fat cell az-adrenoceptor increased (8 vs. 20 nM) but the Bmax values did not significantly change (table 1). The lower affinity of yohimbine in the binding assays was well correlated with its poor efficacy to inhibit the antilipolytic action of UK 14304 as compared with the efficacy of other a2-adrenoce ptor antagonists (idazoxan, phentolamine, RX 821002) (fig. 2). This characteristic of [3H]yohimbine binding (low affinity for az-sites ) has been reported previously for rat tissues by Bylund (1985) who demonstrated that it was a characteristic of rodent type a2-adrenoceptors. Thus it appears that [3H]yohimbine is not the most suitable antagonist radioligand to use for the identification of the total hamster fat cell a2-adrenoceptor population. Two recently developed a2-antagonists were therefore used. Both agents (idazoxan and RX 821002) are potent and selective a2-adrenoceptor antagonists (Doxey et al., 1983; Dabir6 et al., 1983; Stillings et al., 1985). [3H]Idazoxan was used previously to identify a2-adrenoceptors in the brain and other tissues of various species. The new a2-adrenoceptor antagonist, [3H]RX 821002, has not been used until now as a radioligand. On the basis of functional assays, these two compounds were better antagonists than yohimbine in this cell model (fig. 2). Thus, they could be suitable radioligands. Nevertheless, it was impossible to obtain a valid saturation of [3H]idazoxan binding (fig. 5A). [3H]RX 821002 binding (obtained with adrenaline for non-specific binding definition), however, gave reliable binding parameters and a very low non-specific binding. The Bmax values obtained with [3H]RX 821002 and [ 3H]UK 14304 or [3H]clonidine revealed there

were twice as many antagonist binding sites than agonist binding sites, as assessed with the high-affinity component for U K 14304 in competition studies (fig. 7A). These results are consistent with the receptor occupation theory. Competition studies of [3H]RX 821002 binding (fig. 6B) with different a2-adrenergic compounds confirmed the a2-nature of the receptor labelled with the compound and revealed that, as described for [3H]UK 14304 here (fig. 3) and in biological assays, yohimbine was a less effective competitor. Nevertheless, when oxymetazoline and prazosin were used in competition studies of [3H]RX 821002 binding, the relative order of potency of both compounds was consistent with an a2A-Subtype (Bylund et al., 1988); however, the validity of this conclusion is only valid if [3H]RX 821002 does not exhibit subtype selectivity. In the present study we have demonstrated that [3H]RX 821002 is a very good radioligand to identify the whole population of hamster fat cell a2-adrenoceptors. This ligand has already been described as a good tool for the identification of a2A-adrenoceptors in the HT29 cell line (Langin et al., 1989). Thus, the use of [3H]UK 14304 and [3H]RX 821002 will make it possible to perform experiments on the regulation of hamster fat cell a2-adrenoceptors, thus leading to a correct evaluation of the high- and low-affinity states of the a2-adrenoceptor population. In contrast to RX 821002, the other new imidazolinic radioligands that could be used to identify a2-adrenoceptors have certain limitations. [3H]UK 14304 and [3H]idazoxan clearly labelled non-adrenaline-displaceable sites, and were susceptible to competition by phentolamine (figs. 4 and 6A), while the other radioligands [3H]clonidine, [3H]yohimbine and [3H]RX 821002 were not. Competition studies showed that the binding of [3H]UK 14304 and [3H]idazoxan to these sites was only displaceable by imidazolinic compounds. Thus one could propose the existence of 'imidazolinic-preferring sites' on hamster adipocytes. Such 'imidazolinic' sites have been described by others for rabbit kidney (Coupry et al., 1987; Hamilton et al., 1988), rabbit urethra (Yablonsky et al., 1988) rabbit adipocytes (Langin and Lafontan, 1989) and pig kidney (Vigne et al., 1989). Our

156 study shows that these sites exist o n h a m s t e r adipocytes, a n d that these adipocytes could be a suitable model for further investigations. Moreover, a t t e n t i o n should b e focused u p o n the use of p h e n t o l a m i n e to define non-specific b i n d i n g ; its use for this p u r p o s e is q u e s t i o n a b l e i n tissues which possess this k i n d of imidazoline-preferring sites since it leads to a n overestimation of ' t r u e ' a2-adrenoceptors. T o conclude, our experiments show that various imidazolinic derivatives are highly selective a 2 - a d r e n o c e p t o r a g o n i s t s ( U K 14304) a n d antagonists (idazoxan, R X 821002) o n h a m s t e r fat cells. The imidazolinic derivative, [ 3 H ] U K 14304, is a suitable radioligand for the q u a n t i f i c a t i o n of ' h i g h - a f f i n i t y state' a2-adrenoceptors. We have shown that [3H]RX 821002 is a useful raclioligand for a2-adrenoceptor identification in h a m s t e r fat cell m e m b r a n e s whereas [3H]idazoxan has m a j o r limitations for a2-adrenoceptor identification i n this species since it also b i n d s to n o n - a d r e n a l i n e displaceable b i n d i n g sites. W h e n imidazolinic radioligands are used, the use of p h e n t o l a m i n e (or other imidazoline derivatives) for the d e f i n i t i o n of non-specific b i n d i n g is highly q u e s t i o n a b l e w h e n a2-sites are b e i n g investigated i n tissues which possess imidazoline-preferring sites. R X 821002, which exhibits a low affinity for these sites, is a valuable radioligand for the identification of a 2adrenoceptors in tissues which c a n n o t be suitably labelled by [3H]yohimbine.

Acknowledgements The authors gratefully acknowledge the assistance of Drs. Mike R. Stillings and John C. Doxey from Reckitt and Colman (Hull, U.K.) who provided cold and labelled idazoxan and RX 821002 and contributed to discussion of the data. The authors are also grateful to Mrs. Dauzats for her competent technical assistance and preparation of the figures, Mrs. Mathern for her skillful typing and Mrs. Berg6 for her conscientious animal breeding.

References Aktories, K., G. Schultz and K.H. Jakobs, 1980, Regulation of adenylate cyclase activity in hamster adipocytes, NannynSchmiedeb. Arch. Pharmaocl. 312, 167.

Barlow, R.B., 1983, Biodata Handling with Microcomputers (Elsevier Sciences Publishers, Amsterdam, The Netherlands). Berlan, M. and M. Lafontan, 1985, Evidence that epinephrine acts preferentially as an antilipolytic agent in human subcutaneous fat cells. Assessment by analysis of beta- and alpha2-adrenoceptor properties, European J. Clin. Invest. 15, 341. Bylund, D.B., 1985, Heterogeneity of alpha2-adrenergic receptors, Pharmacol. Biochem. Behav. 22, 835. Bylund, D.B., 1988, Subtypes of a2-adrenoceptors: pharmacological and molecular biological evidence converge, Trends Pharmacol. Sci. 9, 356. Bylund, D.B., C. Ray-Prenger and T.J. Murphy, 1988, Alpha2A and alpha2a adrenergic receptor subtypes: antagonist binding in tissues and cell lines containing only one subtype, J. Pharmacol. Exp. Ther. 245, 600. Cambridge, D., 1981, UK-14,304, a potent and selective a 2agonist for the characterization of a-adrenoceptor subtypes, European J. Pharmacol. 72, 413. Carprnr, C., M. Berlan and M. Lafontan, 1983a, Influence of development and reduction of fat stores on the antilipolytic a2-adrenoceptor in hamster adipocytes: comparison with adenosine and fl-adrenergic lipolytic responses, J. Lipid Res. 24, 766. Carprnr, C., M. Lafontan and M. Berlan, 1980, A postsynaptic a2-receptor: the alpha-adrenergic receptor of hamster white fat cells, Experientia 36, 1413. Carprnr, C., M. Lafontan and M. Berlan, 1983b, Mise en 6vidence d'une potentialit6 a2-bloquante de la nicergoline par utilisation du rrcepteur a-adrrnergique du tissu adipeux, J. Pharmacol. (Paris) 14, 57. Coupry, I., R.A. Podevin, J.P. Dause and A. Parini, 1987, Evidence for imidazoline binding sites in basolateral membranes from rabbit kidney, Biochem. Biophys. Res. Commun. 147, 1055. Dabirr, H., J.P. Dausse, P. Mouille, H. Schmitt and P. Meyer, 1983, In vitro studies with(imidazolinyl-2)-2-benzodioxane 1,4[(+)170150], a new potent a2-adrenoceptor blocking agent, European J. Pharmacol. 86, 87. Dole, V.P. and H. Meinertz, 1960, Microdetermination of long chain fatty acids in plasma and tissues, J. Biol. Chem. 235, 2595. Doxey, J.C., A.G. Roach and C.F.C. Smith, 1983, Studies on RX 781094: a selective, potent and specific antagonist of a2-adrenoceptors, Br. J. Pharmacol. 78, 489. Galitzky, J., M. Lafontan, H. Paris and M. Berlan, 1989, Human fat cell alpha2-adrenoceptors - II: comparative study of partial and full agonist binding parameters using [3H]clonidine and [3H]UK 14304, J. Pharmacol. Exp. Ther. 249, 592. Garcia-Salnz, J.A., B.B. Hoffman, S. Li, R.J. Lefkowitz and J.N. Fain, 1980, Role of alphal-adrenoceptors in the turnover of phosphatidyl-inositol and alpha2-adrenoceptors in the regulation of cyclic AMP accumulation in hamster adipocytes, Life Sci. 27, 953. Hamilton, C.A., J. Reid and M.A. Yakubu, 1988, [3H]Yohim-

157 bine and [3H]idazoxan bind to different sites on rabbit forebrain and kidney membranes, European J. Pharmacol. 146, 345. Hoffman, B.B., D. Mullikin-Kilpatrick and R.J. Lefkowitz, 1980, Heterogeneity of radioligand binding to a-adrenergic receptors. Analysis of guanine nucleotide regulation of agonist binding in relation to receptor subtypes, J. Biol. Chem. 255, 4645. Lafontan, M., M. Berlan and C. Carp6n6, 1985, Fat cell adrenoceptors: inter and intraspecific differences and hormone regulation, Int. J. Obesity. 9 (Suppl. 1), 117. Langin, D. and M. Lafontan, 1989, [3H]Idazoxan binding at non-a2-adrenoceptor in rabbit adipocyte membranes, European J. Pharmacol. 159, 199. Langin, D., M. Lafontan, M.R. Stillings and H. Paris, 1989, [3H]RX 821002: a new tool for the identification of a2Aadrenoceptors, European J. Pharmacol. 167, 95. Lanier, S.M., C.J. Homcy, C. Patenande and M. Graham, 1988, Identification of structurally distinct a2-adrenergic receptors, J. Biol. Chem. 263, 14491. Lefkowitz, R.J. and M.G. Caron, 1988, Adrenergic receptors, J. Biol. Chem. 263, 4993. Lowry, O.H., N.J. Rosebrough, A.L. Farr and R.J. Randall, 1951, Protein measurement with the Folin phenol reagent, J. Biol. Chem. 193, 265. Martinez-Olmedo, M.A. and J.A. Garcia-Sainz, 1983, Effect of pertussis toxin on the hormonal regulation of cyclic AMP levels in hamster fat cells, Biochim. Biophys. Acta 760, 215. Pads, H., J. Gahtzky and J.M. Senard, 1989, Interaction of full and partial agonist with HT29 alpha2-adrenoceptor: comparative study of [3H]UK 14304 and [3H]clonidine binding, Mol. Pharmacol. 35, 345. Pecquery, R. and Y. Giudicelli, 1980, Heterogeneity and subcellular localization of hamster adipocyte a-adrenergic receptors - Evidence of a 1- and a2-subtypes , FEBS Lett. 116, 85. Pecquery, R., M.C. Leneuveu and Y. Giudicelli, 1984, In vivo desentization of the fl, but not the a2-adrenoceptor-coupled-adenylate cyclase system in hamster white adipocytes after administration of epinephrine, Endocrinology 114, 1576. Pecquery, R., M.C. Leneveu and Y. Giudicelli, 1988, Influence of androgenic status on the a2/fl-adrenerglc control of

lipolysis in white fat cells: predominant a2-antilipolytic response in testosterone-treated-castrated hamsters, Endocrinology 122, 2590. Pushpendran, C.K. and J.A. Garcia-Sainz, 1984, RX781094 a potent and selective a2-adrenergic antagonist. Effects in adipocytes and hepatocytes, European J. Pharmacol. 99, 337. Rodbell, M., 1964, Metabolism of isolated fat cells, J. Biol. Chem. 239, 375. Schimmel, R.J., R. Serio and K.K. McMahon, 1980, a-Adrenergic inhibition of cyclic AMP accumulation in hamster adipocytes, Biochim. Biophys. Acta 632, 544. Stillings, M.R., C.B. Chapleo, R.C.M. Butler, J.A. Davis, C.D. England, M. Myers, P.L. Myers, N. Tweddle, A.P. Welbourn, J.C. Doxey and C.F.C. Smith, 1985, a-Adrenoceptor reagents. Synthesis of some 2-substituted 1,4-benzodioxans as selective presynaptic a/-adrenoceptor antagonists, J. Med. Chem. 28, 1054. Taouis, M., M. Berlan and M. Lafontan, 1987a, az-Adrenergic turnover in adipose tissue and kidney: irreversible blockade of az-adrenergic receptors by benextramine, Mol. Pharmacol. 31, 89. Taouis, M., M. Berlan, P. Montastruc and M. Lafontan, 1987b, Characterization of dog fat cell adrenoceptors: variations in alpha 2- and beta-adrenergic receptors distribution according to extent of fat deposits and anatomical location, J. Pharmacol. Exp. Ther. 242, 1041. Turner, J.T., C. Ray-Prenger and D.B. Bylund, 1985, Alpha zadrenergic receptors in the human cell line HT29. Characterization with the full agonist radioligand [3H]UK 14304 and inhibition of adenylate cyclase, Mol. Pharrnacol. 28, 422. Vigne, P., M. Lazdunski and C. Frelin, 1989, Guanabenz, guanochlor, guanoxan and idazoxan bind with high affinity to non-adrenergic sites in pig kidney membranes, European J. Pharmacol. 160, 295. Wieland, O., 1957, Eine enzymatische Methode zur Bestimmung von Glycerin, Biochem. Z. 239, 313. Yablonsky, F., J.P. Riffand, J.Y. Lacolle and J.P. Dausse, 1988, Evidence for non-adrenergic binding sites for [3H]idazoxan in the smooth muscle of rabbit urethra, European J. Pharmacol. 154, 209.