Characterization of a 2 -Adrenoceptor Binding Sites in Rabbit Ciliary Body Membranes

Transcription

1 Characterization of a -Adrenoceptor Binding Sites in Rabbit Ciliary Body Membranes Yingjin, Annita Verstappen, and Thomas Yorio Purpose. This study sought to identify and characterize subtypes of a -adrenoceptors in rabbit ciliary body. Methods. Radioligand binding assays were performed with the a -agonist ligand [ 15 I]-/?- iodoclonidine ([ 15 I]PIC) and the antagonist ligand [ 3 H]rauwolscine. Results. Both [ 15 I]PIC and [ 3 H]rauwolscine bound to a single population of receptors in membrane preparations of rabbit ciliary body. The densities for the two ligands were the same (about 640 fmol/mg). However, the affinity of [ 15 I]PIC (dissociation constant, K^ = 1.91 ± 0.4 nm) was about threefold higher than that of [ 3 H]rauwolscine (dissociation constant = 6.79 ± 1.5 nm), and binding of [ 15 I]PIC exhibited guanine nucleotide sensitivity. Inhibition of [ I5 I]PIC binding by epinephrine, idazoxan, and amiloride was examined to differentiate between a -adrenoceptors and the imidazoline-preferring receptor. Epinephrine and idazoxan competed for all of the [ 15 I]PIC binding; relative potency was epinephrine > idazoxan >> amiloride. Subtypes of a -adrenoceptors were further studied by competition for [ 15 I]PIC binding by subtype-selective compounds. [ 15 I]PIC binding sites showed the pharmacologic characteristics of an a A -adrenoceptor (oxymetazoline > chlorpromazine» prazosine), and competition by oxymetazoline and chlorpromazine was best fit by a one-site model. Likewise, the relative potency of inhibition of [ 3 H]rauwolscine binding was oxymetazoline > chlorpromazine. However, inhibition of [ 3 H]rauwolscine-binding by oxymetazoline was better fit by a two-site model, which was converted to a one-site model in the presence of nm 5'-guanylimidodiphosphate. Conclusions. A large number of «-adrenoceptors are present in rabbit ciliary body. They are not the imidazoline-preferring receptor but are «-adrenergic-specific receptors of the a A subtype. Invest Ophthalmol Vis Sci. 1994;35: he ability of a -adrenoceptor agonists to reduce the intraocular pressure (IOP) has been well documented, 1 " 4 but not completely understood. These drugs appear to reduce IOP by decreasing aqueous humor formation. 5 Ligand binding studies demonstrated the presence of a large number of a -receptors in rabbit iris ciliary body (ICB). 6 " 8 It was suggested that these receptors may account for the ocular hypotensive effects of a -receptor selective-agonists. Recently, From the Department of Pharmacology, North Texas Eye Research Institute, University of North Texas Health Science Center at Fort Worth, Texas. Supported in part by grants from Alcon Laboratories and from Farrnos, Finland. Submitted for publication April 13, 1993; revised September 16, 1993; accepted November 9, Proprietary interest category: N. Reprint requests: Thomas Yorio, Department of Pharmacology, University of North Texas Health Science Center at Fort Worth, 3500 Camp Bowie Boulevard, Fort Worth, TX a receptor type, "imidazoline-preferring receptor" (IPR), has emerged from the «-adrenergic family of receptors. It is activated by imidazoline but not catecholamine agonists. 9 "" Therefore, IPR is not an adrenergic receptor. Many a -agonists, like lofexidine, detomidine and clonidine, can bind to both a -adrenoceptors and the IPR. The exact receptor types mediating the IOP response to these a -agonists remain unclear. Furthermore, there is convincing evidence to demonstrate that a -adrenoceptors do not represent one homogenous population of receptors. a A -> «ir> a c -, and a D -Adrenoceptor subtypes have been proposed and pharmacologically characterized. 1 " 14 Selective ligands for identification of these subtypes have been developed in the past few years. With the development of imidazoline selective compounds, a more thorough pharmacologic characterization of adreno- 500 Investigative Ophthalmology & Visual Science, April 1994, Vol. 351 No. 5 Copyright Association for Research in Vision and Ophthalmology

2 a -Adrenoceptor Binding Sites in Rabbit Ciliary Body 501 ceptors in ocular tissue is possible. The subtypes of a -adrenoceptors in the ciliary body have not yet been reported. In the current study, we sought to identify and characterize subtypes of a -adrenoceptors in rabbit ciliary body using the high specific activity partial a -agonist ligand, [ 15 I]-/Modoclonidine ([ 15 I]PIC) and the a -antagonist ligand, [ 3 H]-rauwolscine. Our results suggest that the prominent a -receptor subtype in rabbit ciliary body membrane preparation is of the a A -adrenoceptors, a receptor normally associated with inhibition of adenylate cyclase. METHODS Materials [ 15 I]PIC (00 Ci/mmol) and [ 3 H]rauwolscine (81 Ci/mmol) were obtained from New England Nuclear and Amersham, respectively. ( )-Epinephrine bitartrate, yohimbine, prazosin, DL-dithiothreitol, phenylmethylsulfonyl fluoride, ethyleneglycol-bis-(b-aminoethyl ether)n,n'tetra-acetic acid (EGTA), and 5'-guanylimidodiphosphate (Gpp(NH)p) were purchased from Sigma Chemical Co. Idazoxan, amiloride, clonidine and /?-iodoclonidine were from Research Biochemical Inc. (RBI), oxymetazoline from ICN Biochemical, rauwolscine from ROTH and chlorpromazine from Smith Kline & French Labs. All procedures involving animals conformed with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Rabbit Ciliary Body Membrane Preparation For each preparation, freshly enucleated albino rabbit eyes were obtained from a local slaughterhouse. Adhering tissue was cut away from the eyes and the eyes were cut into anterior and posterior poles. The vitreous body and the lens were gently removed from the open anterior segment of the eyes and the ICB was carefully peeled off. The iris was cut away from the ICB and individual ciliary bodies were homogenized by a motor driven glass homogenizer in 3 ml homogenizing buffer (Tris-HCl, 50 mm; DL-dithiothreitol, 1 mm; EGTA, 0.1 mm; and phenylmethylsulfonyl fluoride, 0.1 mm; ph = 7.5). The homogenate was centrifuged at 500g for 10 minute at 4 C and the supernatant was collected. The pellet was homogenized and centrifuged again as above. The supernatant from two centrifugations were combined and centrifuged at 30,000g for 0 minutes at 4 C. This pellet was rinsed twice with 30 to 40 volumes of [ 15 I]PIC binding buffer (Tris-HCl, 50 mm; MgCl, 10 mm and EGTA, 1 mm; ph = 7.6) or [ 3 H]-rauwolscine binding buffer (Tris-HCl, 50 mm; EGTA, 1 mm; ph = 7.6). The final pellet was suspended in the corresponding binding buffer. Protein concentrations were determined by the method of Bradford 15 using bovine serum album (BSA) as a standard. The membrane was frozen under liquid nitrogen and stored at 80 C until use. All procedures were conducted on ice except where otherwise indicated. Radioligand Binding Assay Both [ 15 I]PIC and [ 3 H]-rauwolscine were used in the binding assay. Assays were initiated by the addition of membrane preparations (15 iig for [ 15 I]PIC binding and 30 ixg for [ 3 H]-rauwolscine binding) to the reaction tubes containing radioligand, other drugs and binding buffer (mentioned above). For saturation binding experiments, 6 to 7 concentrations of [ 15 I]- PIC (0.05 to 5 nm) or [ 3 H]rauwolscine (1.5 to 5 nm) were used. For competition experiments, the radioligand concentration was 0.3 to 0.5 nm for [ 15 I]PIC and 1 nm for [ 3 H]-rauwolscine, with 8 to 16 various concentrations of unlabeled competing ligands spaced on either side of the approximate IC 50 concentration. Nonspecific binding was defined by 5 fim clonidine for [ 1) I]PIC and 10 /xm yohimbine for [ 3 H]-rauwolscine binding. In experiments with epinephrine, all samples contained ascorbic acid in a final concentration of 0.001%. Binding reactions were conducted in 00 ix\ for 60 minutes at 5 C and terminated by addition of 5 ml of ice-cold washing buffer (Tris-HCl, 50 mm; MgCl, 10 mm; ph = 7.6 for [ 15 I]PIC and Tris-HCl, 50 mm, ph = 7.6 for [ 3 H]-rauwolscine) and filtration over a glass fiber filter (Whatman GF/C). The filters were washed with additional 15 ml of the washingbuffer. Samples contain [ 15 I]PIC was counted in a gamma counter at an efficiency of 78% (Beckman 700), whereas samples from [ 3 H]-rauwolscine binding was placed in a scintillation vial, covered with 10 ml of scintillation cocktail and counted in a scintillation counter at 40% efficiency (Packard Tricarb 460). The amount of radioligand bound was less than 4% of the total radioligands added. Data Analysis Binding data from saturation experiments were transformed according to the method of Scatchard, 10 and values for dissociation constant (KJ (affinity) and B m:ix (density) were obtained by linear regression analysis of transformed data. 17 Hill coefficients or pseudo-hill coefficients (n H ) were obtained from Hill plot transformations of saturation or competition curves by linear regression. 18 Data from both saturation and competition experiments were analyzed using a nonlinear least-squares curve fitting program (Sigma Plot 5.0; Jandel Scientific; Corte Madera; 199 and Tablecurve, 1.0, 199, Jandel Scientific). Iterative curve fitting of both one-site and two-site models to experimen-

3 50 Investigative Ophthalmology & Visual Science, April 1994, Vol. 35, No. 5 tal data provided IC 50 values, which were converted to inhibition constant (K ; ) values using the Cheng-Prusoff correction 190 for a one-site model. The one-site and two-site models were compared using the following equation 1 : F = [(SS, - SS )/(df, - df )]/(SS /df ) where SS, and SS are the sum of squared residuals for the one-site and two-site model respectively, df] and df are the corresponding degrees of freedom. The two-site model was taken to be a significantly better fit than the one-site model when the calculated F value was greater than or equal to the tabulated value at the prescribed degrees of freedom and the probability of a one-tailed error of Results are expressed as mean ± SE. Both paired and unpaired Student's t test was used to determine the statistical significance of the treatment \ \ \ i * \ A \ A control Kd =.54 nm Bmax = 51 fmol/mg 10 /JM Gpp(NH)p Kd =.94 nil Bmax = 388 fmol/mg i. \. i. i\ Bound (fmol/mg) RESULTS Saturation Analysis A representative saturation curve for [ 15 I]PIC binding in membrane preparations from rabbit ciliary body is shown in Figure 1. Binding was saturable (0.15 to 5 nm) and specific binding was about 85% of total binding at the Kt, concentration of the ligand. Scatchard transformation of specific binding data produced a XT Total Specific A Non-specific [ 15 I]PIC (nm) FIGURE 1. Saturation binding of [ I5 I]PIC in rabbit ciliary body membrane. Specific binding was defined as difference of binding between the absence (total) and presence (nonspecific) of 5 fim clonidine. Results shown are from a single experiment but are similar to those of five other experiments. FIGURE. Effect of Gpp(NH)p on [ IW I]PIC binding. Saturation experiment was performed in the presence and absence (control) of 10 fim Gpp(NH)p. Results shown are from a single experiment but are similar to those of two other experiments. straight line (r = 0.9), with K^ and B niax values of 1.91 ± 0.4 nm and 67 ± 9 fmol/mg protein (n = 6), respectively. The n H was These findings indicate that [ 15 I]PIC binds to a single high-affinity site in rabbit ciliary body. To test whether this binding was sensitive to the functional state of the coupled guanine nucleotide binding protein (G protein), saturation experiments were performed in the presence and absence of 10 /um Gpp(NH)p, a nonhydrolyzable GTP analog. As shown in Figure, the inclusion of 10 nm Gpp(NH)p in the assay decreased the B max by about 5% (P < 0.05 for paired t test) but it did not alter the K^, of the residual binding (P > 0.05). The K^ in the presence of Gpp(NH)p was.78 ± 0.07 nm (n = 3). Binding in the presence and absence of Gpp(NH)p was best fit by a single population of high-affinity binding sites. Therefore, the receptor site identified by [ 15 I]PIC exhibited the guanine nucleotide sensitivity. Binding of [ 3 H]rauwolscine to ciliary body membrane was also saturable (1.5 to 5 nm), and specific binding versus total binding was about 55% at the K^ concentration of the ligand, which is lower than [ 1 "T]- PIC binding. The scatchard plot of [ 3 H]rauwolscine binding was linear (r = 0.9). The B max was similar to that of [ 15 I]PIC binding, but the K^, was about threefold higher than the K,, of [ 15 I]PIC binding (Table 1). The Hill coefficient (n H ) was close to unity (0.93). Therefore, like [ 15 I]PIC, [ 3 H]rauwolscine appeared to bind to the membrane preparations with a single

4 ayadrenoceptor Binding Sites in Rabbit Ciliary Body 503 TABLE l. Comparison of Equilibrium Binding Parameters for a ~R a dioligands in Rabbit Ciliary Body B max No. of K d (nm) (fmol/mg) Experiments [ 15 I]PIC 1.91 ± ±9 [ s H]Rauwolscine 6.79 ± 1.5* 653 ± 5f The statistical significance of difference between parameters for two ligands was determined by a two-tailed unpaired Student's l-test: *P< 0.05, fp > affinity. However, unlike [ 15 I]PIC, the inclusion of 10 nm Gpp(NH)p in the binding assay of [ 3 H]rauwolscine did not alter either B max or Kj. In the presence of Gpp(NH)p the K., and B max were 4.81 ± 0.6 nm and 68 ±15 fmol/mg protein (n = 3), respectively. Specificity of Binding To characterize further the binding sites labeled by [ 15 I]PIC in rabbit ciliary body, we performed competition curves with: (1) IPR-preferring ligands, idazoxan, 91 and amiloride 10 ' ; () a -adrenoceptor selective ligands, rauwolscine and epinephrine; (3) a A - adrenoceptor selective ligand, oxymetazoline 14 ; (4) o: B -adrenoceptor selective ligand, chlorpromazine 14 and prazosin 14 ; and (5) the nonselective partial agonists, clonidine and iodoclonidine. As shown in Figure 3, epinephrine and idazoxan competed for all of the specific [ 15 I]PIC binding with K; values of 1 ± nm and 116 ± 6 nm, respectively. In contrast, amiloride at concentrations up to 10~ 4 M only inhibited about 48% of specific binding with a calculated K; value of 80 ± 19 /um. The K ; values of these three ligands obtained from the current study are similar to the reported K ; values for these agents at a -adrenoceptors in rabbit renal cortex. The relative potency of these ligands to inhibit [ 15 I]PIC bindingwas epinephrine > idazoxan >> amiloride, which is pharmacologic characteristic of a -adrenoceptors in rabbit tissues. When data were analyzed by computerized nonlinear curve-fitting, it was found that a onesite model fit the idazoxan data. However, for epinephrine, a two-site model was preferred over a one-site model (P < 0.05). The estimated high-affinity IC 50 was 14 nm ± nm (83%) and the low-affinity IC 50 was 800 ± 00nM(17%). Oxymetazoline, chlorpromazine and prazosin have been reported to differentiate between o: A - and a ir adrenoceptor subtypes. Inhibition of [ 15 I]PIC binding by these three subtype selective compounds was therefore studied (Fig. 4). Oxymetazoline and chlorpromazine completely inhibited the specific [ 1 T]PIC binding. However, prazosin at 10~ 4 M only A epinephrine idazoxan amiloride 0 oxymetazoline chlorpromazine prazosin log [INHIBITORS] (M) FIGURE 3. Inhibition of [ I5 I]PIC binding by epinephrine, idazoxan, and amiloride. Total specific binding was the specific binding in the absence of inhibitors. Results shown are the mean (± SE) of two to three experiments from different membrane preparations log [INHIBITORS] (M) FIGURE 4. Inhibition of [ ias I]PIC binding by oxymetazoline, chlorpromazine and prazosin. Total specific binding was defined same as in Fig. 3. Results shown are the mean (± SE) of two to three experiments from different membrane preparations.

5 504 Investigative Ophthalmology 8c Visual Science, April 1994, Vol. 35, No. 5 TABLE. Pharmacologic Specificity of ] 15 I]PIC Binding in Rabbit Ciliary Body K ( (nm) n H I max (%) No. of Experiments Oxymetazoline Iodoclonidine Clonidine ( )-Epinephrine Rauwolscine Idazoxan Chlorpromazine Amiloride Prazosine.75 ± ± ± ± 3 ± 116 ± 6,70 ± 50 80,000 ± 19, ,000 ± 19, I max = the maximum inhibition of binding (as percentage of total specific binding) observed in the presence of the highest concentration of the compound tested (up to nm). inhibited about 44% of specific binding. The relative potency of these three compounds to inhibit [ I5 I]PIC binding was oxymetazoline > chlorpromazine > prazosin (Table ). Therefore, the binding sites labeled by [ 1r T]PIC in ciliary body membranes displayed relatively high affinity for oxymetazoline and low affinity for chlorpromazine and prazosin. These observations are consistent with the pharmacologic characteristics of the «A- a drenoceptor subtype. The data were further analyzed by computerized curve fitting and for oxymetazoline and chlorpromazine inhibition of [ 15 I]- PIC binding a one-site model fit the data best. These observations indicate that the a -adrenoceptors detected by [ If T]PIC in the preparation are of a A subtype. To characterize this binding site further, inhibition of [ 15 I]PIC binding by unlabeled/?-iodoclonidine was studied, p-iodoclonidine inhibited all of the specific [ 15 I]PIC binding with a Kj of 7.49 nm, a value which is relatively close to its directly measured K ti. The Hill coefficient (n H ) was 0.99 and the computer fitting analysis showed a one-site model. The validity of using 5 /xm clonidine to define nonspecific binding was also examined by comparing the maximum inhibition of [ 15 I]PIC binding by various compounds. When competing for 0.3 to 0.5 nm [ 15 I]- PIC, the binding remaining in the presence of 10 /um idazoxan, epinephrine, rauwolscine, 1 fxm oxymetazoline and 60 /xm chlorpromazine was not different from that remaining in the presence of 5 ixm clonidine (Table ). Thus, these agents, at their maximal concentrations, appeared to block the same population of sites which specially bound [ I5 I]PIC, indicating that 5 /nm clonidine provided an appropriate definition for nonspecific binding in this assay. Inhibition of [ 3 H]rauwolscine binding by a - adrenoceptor subtype selective compounds, oxymetazoline and chlorpromazine, was examined in the same membrane preparation. As with [ 15 I]PIC binding, both oxymetazoline and chlorpromazine competed for all of the [ 3 H] rauwolscine binding with the Kj values of 6.8 ± 3 nm and 90 ± 30 nm (n = 3), respectively. These values are similar to those obtained in [ 1 T]PIC competition binding studies (Table ). However, different from [ I T]PIC binding, oxymetazoline competition for [ 3 H] rauwolscine binding was biphasic and the data were best fit by a two-site model (P < 0.05, Fig. 5). The IC 50 for the high-affinity site was log [OXYMETAZOLINE] Gpp(NH)p (M) -4-3 FIGURE 5. Inhibition of [ 3 H]rauwolscine binding by oxymetazoline both in the absence (control) and presence of jim Gpp(NH)p. Total specific binding was the specific binding in the absence of oxymetazoline. Results shown are the mean (± SE) of three experiments.

6 a -Adrenoceptor Binding Sites in Rabbit Ciliary Body 4.15 ± 1.67 nm (n = 3) and represented 84% of the binding sites, whereas the IC 50 for the low-affinity site was.4 ± 1 ixm (n = 3), but represented only 16% of the sites. When the competition experiments were repeated in the presence of /xm Gpp(NH)p, the inhibition curve was steep and monophasic (Fig. 5). Computer analysis of the curve indicated a one-site model (P> 0.05). Competition of [ 3 H]rauwolscine binding by iodoclonidine was also studied. Iodoclonidine competed for all of the [ 3 H]rauwolscine binding with a Kj value of 6 ± 5 nm (n = ) and n H of The curve was best fit by a one-site model (P > 0.05). Therefore, iodoclonidine appeared to bind with a single affinity to the [ 3 H]rauwolscine binding sites. The same amount of nonspecific binding was observed in the presence of a 10 /im concentration of oxymetazoline, chlorpromazine, iodoclonidine and yohimbine (data not shown), which further supported the use of 10 /im yohimbine to define nonspecific [ 3 H]- rauwolscine binding. DISCUSSION The current study represents the first report of the binding properties of the radioiodinated a -adrenoceptor agonist ligand, [ 15 I]PIC, in rabbit ciliary body. The ligand binding properties of [ 1r T]PIC were compared with the tritiated antagonist ligand, [ 3 H]- rauwolscine and the results demonstrate that a - adrenoceptors in this tissue are of the a A subtype. [ 1r T]PIC possesses both very high specific radioactivity and high affinity for a -adrenoceptors. This offers clear advantages over other a -radioligands in experiments with small tissue samples like the ciliary body, where binding to low amounts of protein (15 ^g) resulted in tens of thousands of counts specially bound (Fig. 1). In contrast, [ 3 H]rauwolscine would give much fewer counts under the same conditions. Therefore, increasing amounts of protein had to be used, reducing the numbers of experiments that could be performed from each membrane preparation by 50%. Another important characteristic of [ 15 I]PIC is its capacity to bind to the IPR as well as a A and a B subtypes of a -adrenoceptors. 3 This makes it an ideal radioligand for differentiating subtypes within each receptor class. The presence of a large number of a -adrenoceptors in rabbit ciliary body was demonstrated by saturation experiments with both [ 15 I]PIC and [ 3 H]- rauwolscine. The B m;>x obtained from these experiments was close to that (764 fmol/mg) of Jumblatt et al 8 but was about.5-fold more than the B max (about 00 to 50 fmol/mg protein) determined from [ 3 H]- yohimbine or [ 3 H]rauwolscine binding assay reported 505 by Mittag and Tormay, 6 Mittag et al, 4 and Mallorga et al. 7 This may be due to the relatively more purified membrane preparation in the current study, where only the ciliary body was used for the membrane preparation (the iris was removed from the iris-ciliary body) and the homogenate of the ciliary body was centrifuged in two steps (500g first with pellet discarded and then 30,000g with the supernatant discarded). On the other hand, binding of [ 3 H]rauwolscine to ciliary body membrane with a single affinity (K^ = 6.79 nm) is in good agreement with earlier binding data from this tissue. 6 " 8 The current observations suggest that [ 1r T]- PIC and [ 3 H]rauwolscine bind to the same population of receptors. Observations supporting this included: (1) the receptor density detected by these two ligands was the same (Table 1); () unlabeled iodoclonidine competed for all of the [ 3 H]rauwolscine specific binding; and (3) unlabeled rauwolscine inhibited all of the [ 15 I]PIC specific binding. Interestingly, although both [ 15 I]PIC and [ 3 H]rauwolscine bound to the ciliary body membrane with a single affinity, their sensitivity to guanine nucleotides was different. Inclusion of Gpp(NH)p in the assay significantly decreased the B I1Klx of [ 1 T]PIC binding, but not of [ 3 H]rauwolscine. Modulation of binding of agonists, but not antagonists, by guanine nucleotides has been reported in human platelets. 3 ' 5 "' 3 It has been suggested that agonist, but not antagonist, binding to «-adrenoceptors can either induce or stabilize the formation of a ternary complex between agonist, receptor and G protein. 7 This state of the a -adrenoceptor possesses considerably higher affinity for agonists than the free receptor. Addition of a guanine nucleotide like GTP or Gpp(NH)p appears to destabilize the ternary complex so that only the initial interaction of the agonist with the receptor can be detected. In the current study, binding sites detected by [ 1 T]PIC, a partial agonist, should be in the high-affinity state (coupled with G protein). The presence of 10 ^M Gpp(NH)p converted part of the receptors in the high-affinity state into a low-affinity state, which were undetectable under the current experimental conditions, leaving other receptors in the same affinity state as in the absence of Gpp(NH)p. Therefore, it appears that there was a decrease in B m:ix. In contrast, binding sites detected by [ 3 H] rauwolscine, an antagonist, should be in the low-affinity state (free receptor) so that Gpp(NH)p had no effect on its bind- ing- Experimental evidence demonstrates that a - adrenoceptor radioligands with an imidazoline structure bind with high affinity not only to «-adrenoceptors but also to sites that are not competed for by catecholamines. 9 These nonadrenergic sites have been termed IPR and identified in various tissues and species, including rabbit. 8 " 31 However, the pharmaco-

7 506 Investigative Ophthalmology 8c Visual Science, April 1994, Vol. 35, No. 5 logic characteristics of IPR differ considerably, depending on the species investigated and the radioligand used. Terminology of I] and I imidazoline sites has been proposed for the binding sites labeled by [ 3 H] clonidine and [ 3 H] idazoxan, respectively. 11 ' 3 [ 1r T]PIC is an imidazoline, so it could potentially bind to these IPR as well as a -adrenoceptors. However, significant binding of [ 15 I]PIC to rabbit ciliary body IPR in our experiments was unlikely as epinephrine competed for all of the specific [ 15 I]PIC binding. If there was substantial binding of [ 15 I]PIC to imidazoline sites, epinephrine would have not displaced all of the [ 1r T]PIC binding. Although inhibition of [ 15 I]- PIC binding by epinephrine was best fit by a two-site model, the ratio (57) of estimated IC 50 values for the high-affinity and low-affinity sites and the IC 50 value of the low-affinity site (IC 50 = 800 nm) do not support the possibility that the low-affinity site was an imidazoline binding site. It has been reported that epinephrine showed approximately 4,000 to 10,000-fold selectivity for a -adrenoceptors over imidazoline binding sites 933 and furthermore, the Ki value of epinephrine at imidazoline binding sites was reported to be > 1,000,000 nm. 30 Epinephrine inhibition of [ 15 I]PIC binding in a manner consistent with a two-site model could be explained as the result of negative cooperativity or different affinity states of a single receptor. Amiloride, which has been reported to have a very low affinity to a -adrenoceptors but relatively high affinity to IPR in rabbit tissues, 10 ' could not compete for all of [ 15 I]PIC binding (only about 48%) and the calculated Kj value was close to its K, value at a -adrenoceptors. Idazoxan, which is considered to be more selective for imidazoline binding sites than a -adrenoceptors in rabbit tissues, ' 34 inhibited the [ I5 I]PIC binding completely with a single affinity. This is in good agreement with the affinity reported for idazoxan at a -adrenoceptors of rabbit kidney. If there was significant binding of [ 15 I]PIC to imidazoline binding sites, competition study of idazoxan would have detected this and the computer analysis of the binding data would have indicated a two-site model. Instead, a one-site model best fit the data. In addition, the same amount of nonspecific [ 15 I]PIC binding was observed without regards to structure and the B max values obtained from the imidazoline radioligand, [ 1l I]PIC, and the nonimidazoline radioligand, [ 3 H]- rauwolscine, were also equivalent. These observations indicate that there were little, if any, imidazoline binding sites in our rabbit ciliary body membrane preparations. a -adrenoceptors have been subclassified into a A» «B» a C> an d «D subtypes pharmacologically and into a -C10, -C4 and -C by molecular cloning techniques 13,35,36 A recent study shows that pharmacologically defined a A -, a B -, and a C -adrenoceptor subtypes correspond to the cloned human a-c10, -C and -C4 receptor subtypes, respectively. 37 Of these subtypes, the most studied are a A - and a B -adrenoceptors, which can be distinguished on the basis of the relative potencies of oxymetazoline, which is a A -selective, as well as prazosin and chlorpromazine, which are a B -selective. The results from this study suggest that rabbit ciliary body contains the a A -adrenoceptor subtype. There are several observations supporting this conclusion. First, inhibition of [ 15 I]PIC binding by three subtype-selective compounds displayed pharmacologic characteristics of an a A subtype (oxymetazoline > chlorpromazine > prazosin). Second, computer modeling indicated the existence of a single class of binding sites for both oxymetazoline and chlorpromazine. Thirdly, the K; values of these three compounds obtained from the current study are close to the reported Kj values of them in tissues containing only a A subtype. 14 ' 3 Inhibition of [ 15 I]PIC binding by epinephrine in a manner consistent with a two-site model is not a indication of existence of a A and a B subtypes, because epinephrine has been reported not to differentiate between these two subtypes. 14 To further our pharmacologic characterization of a -adrenoceptors in rabbit ciliary body, data from [ 1: T]PIC binding experiments were compared with the binding of the nonselective a -antagonist ligand, [ 3 H]rauwolscine. It was shown that the relative potency for inhibiting [ 3 H]rauwolscine binding was oxymetazoline > chlorpromazine with K ; values consistent with that obtained from the [ 15 I]PIC binding study. However, inhibition of [ 3 H]rauwolscine binding by oxymetazoline was best fit by a two-site model. This was probably due to oxymetazoline binding to the two different states of the single receptor subtype. This is supported by the observation that nm Gpp(NH)p converted the two-site inhibition curve to a one-site fit (Fig. 5). Therefore, the results from both [ 15 I]PIC and [ 3 H]rauwolscine binding experiments demonstrated that most, if not all, a -adrenoceptors in our rabbit ciliary body membrane preparation were of the a A subtype. Furthermore, earlier finding that prazosin exhibited a low affinity for [ 3 H]rauwolscine binding in rabbit ICB membranes 7 ' 4 suggests the presence of a A -adrenoceptor subtype in this tissue. In summary, characterization of the a -adrenoceptor binding with the radioiodinated «-agonist ligand, [ 15 I]PIC, was studied in rabbit ciliary body membrane preparations. The results demonstrate that the rabbit ciliary body contains a large number of a - adrenoceptors and they are of the a A subtype. The identification of an a -adrenoceptor subtype in ciliary body creates the possibility of the development of subtype selective antiglaucoma drugs, which might have

8 a -Adrenoceptor Binding Sites in Rabbit Ciliary Body 507 fewer side effects with more selective therapeutic action, if the side-effects are not related to the a A subtype. Key Words rabbit ciliary body, a -adrenoceptors, radioligand binding assay, [ 1<5 I]-/?-iodoclonidine, [ 3 H]rauwolscine Acknowledgments The authors thank Mike Martin for helpful discussions and for assistance with the curve-fitting programs. References 1. Burke JA, Potter DE. Ocular effect of a selective «-agonist (UK-14,304-18) in cats, rabbits and monkeys. Curr Eye Res. 1986;5: Elko EE, Tran T, Lai H, Yorio T. Ocular hypotensive effects of lofexidine, an a^-adrenoreceptor agonist. Drug Dev Res. 1988;14: Lee DA, Topper JE, Brubaker RF. Effects of clonidine on aqueous humor flow in normal human eyes. Exp Eye Res. 1984;38: Jin Y, Wilson S, Elko EE, Yorio T. Ocular hypotensive effects of medetomidine and its analogs. J Ocular Pharmacol. 1991;7: Chiou GCY. Effects of a, and a. activation of adrenergic receptors on aqueous humor dynamics. Life Set. 1983;3: O4. 6. Mittag TW, Tormay A. Adrenergic receptor subtypes in rabbit iris-ciliary body membranes: Classification by radioligand studies. Exp Eye Res. 1985;40: Mallorga P, Buisson S, Sugrue MF. Alpha,-adrenoceptors in the albino rabbit ciliary process. J Ocular Pharmacol. 1988,4: Jumblatt J, Liu JGH, North GT. Alpha -adrenergic modulation of norepinephrine secretion in the perfused rabbit iris-ciliary body. Curr Eye Res. 1987;6: Lehmann J, Koenig-Berard E, Vitou P. The imidazoline preferring receptor. Life Set. 1989; 14: Michel MC, Insel PA. Are there multiple imidazoline binding sites? TIPS. 1989;10: Michel MC, Ernsberger P. Keeping an eye on the I site: lmidazoline-preferring receptors. TIPS. 199; 13: Bylund DB. Heterogeneity of alpha -adrenergic receptor. Pharmacol Biochem Behav. 1985;: Bylund DB. Subtypes of a -adrenoceptor: Pharmacologic and molecular biological evidence converge. TIPS. 1988;9: Bylund DB, Ray-Prenger C, Murphy TJ. Alpha- A - and alpha B -adrenergic receptor subtypes: Antagonist binding in tissues and cell lines containing only one subtype. J Pharmacol Exp Ther. 1988;45: Bradford M. A rapid and sensitive method for the quantification of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem. 1976;7: Scatchard G. The attractions of proteins for small molecules and ions. Ann NY Acad Sci. 1949;51: Tallarida RJ, Murray RB. Manual of Pharmacologic Calculations with Computer Programs. nd ed. New York: Springer-Verlag; 1987: Hill AV. The possible effects of the aggregation of the molecules of haemoglobin in its dissociation curves. J Physiol (Lond). 1910;40: Cheng YC, Prusoff WH. Relationship between the inhibition constant (Kj) and the concentration of inhibitor which causes 50 percent inhibition (IC 50 ) of an enzyme reaction. Biochem Pharmacol. 1973;: Chou TC. Relationships between inhibition constants and fractional inhibition in enzyme-catalyzed reactions with different numbers of reactants, different reaction mechanisms, and different types of mechanisms of inhibition. Mol Pharmacol. 1974;] 0: Munson PJ. Ligand: A computerized analysis of ligand binding data. Methods Enzymol. 1983;9: Parin A, Coupry I, Graham RM, Uzielli I, Atlas D, Lanier SM. Characterization of an imidazoline/guanidinium receptive site distinct from the a -adrenergic receptor. J Biol Chem. 1989;64: Gerhardt MA, Wade SM, Neubig RR. p-[ lb I]- Iodoclonidine is a partial agonist at the a -adrenergic receptor. Mol Pharmacol. 1990;38: Mittag TW, Tormay A, Severin C, Podos SM. Alphaadrenergic antagonists: Correlation of the effect on intraocular pressure and on a -adrenergic receptor binding specificity in the rabbit eye. Exp Eye Res. 1985;40: Nerbig RR, Gantzos RD, Brasier RS. Agonist and antagonist binding to a -adrenergic receptors in purified membranes from human platelets. Mol Pharmacol. 1985;8: Hoffman BB, Michel T, Brenneman TB, Lefkowitz RJ. Interactions of agonists with platelet a -adrenergic receptors. Endocrinology. 198;110: Limbird L. Alpha -adrenergic systems: Models for exploring hormonal inhibition of adenylate cyclase. Ti PS. 1983;4: Coupry I, Podevin RA, Dausse JP, Parini A. Evidence for imidazoline binding sites in basolateral membrane from rabbit kidney. Biochem Biophys Res Comm. 1987;147: Ernsberger P, Meeley MP, Mann JJ, Reis DJ. Clonidine binds to imidazoline binding sites as well as a - adrenoceptor in the ventrolateral medulla. Eurf Pharmacol. 1987; 134: Ernsberger P, Feinland G, Meeley MP, Reis D. Characterization and visualization of clonidine-sensitive imidazole sites in rat kidney which recognize clonidine-displacing substance. Am] Hypertens. 1990;3: Piletz JE, Andorn AC, Unnerstall JR, Halaris A. Binding of [ :1 H]-p-aminoclonidine to a -adrenoceptor states plus a non-adrenergic site on human platelet

9 508 Investigative Ophthalmology & Visual Science, April 1994, Vol. 35, No. 5 plasma membranes. Biochem Pharmacol. 1991;4: Ernsberger P. Heterogeneity of imidazoline binding sites: Proposed I, and I subtypes. Fundam Clin Pharmacol. 199;6:55s. 33. Ernsberger PR, Westbrook KL, Christen MO, Schafer SG. A second generation of centrally acting antihypertensive agents act on putative I,-imidazoline receptors. J Cardiovasc Pharmacol. 199;4:sl Yablonsky F, Riffaud JP, Lacolle JY, Dausse JP. Evidence for non-adrenergic binding sites for [ 8 H]- idazoxan in the smooth muscle of rabbit urethra. EurJ Pharmacol. 1988;154: Nicholas AJ, Ruffolo RR, Jr. a-adrenoceptor subclassification. In: Ruffolo RR, Jr, ed. a-adrenoceptor: Molecular Biology, Biochemist?y and Pharmacology. Basel: Karger; 1991: Ray-Prenger C, Bylund DB. Alpha-A and alpha-b adrenergic receptor subtypes: Attenuation of cyclic AMP production in cell lines containing only one receptor subtype. J Pharmacol Exp Ther. 1989;51: Bylund DB, Blaxall HS, Iversen LJ, Caron MG, Lefkowitz RJ, Lomasney JW. Pharmacological characteristics of a " a drenergic receptors: Comparison of pharmacologically defined subtypes with subtypes identified by molecular cloning. Molecular Pharmacol. 199;4:l-5.