Peptidases and smooth muscle cell angiotensin II receptor pharmacology

Peptides. Vol. 14, pp. 345-352, 1993 0196-9781/93 $6.00 + .00 Printed in the USA. Copyright © 1993 Pergamon Press Ltd.

Peptidases and Smooth Muscle Cell Angiotensin II Receptor Pharmacology

R O B E R T B. C O H E N , M A R I A L. W E B B A N D K E N N E T H E. J. D I C K I N S O N l

Department of Cardiovascular Biochemistry, Bristol-Myers Squibb Pharmaceutical Research Institute, P.O. Box 4000, Princeton, NJ 08643-4000

Rece ived 25 S e p t e m b e r 1992

COHEN, R. B., M. L. WEBB AND K. E. J. DICKINSON. Peptidases and smooth muscle cell angiotensin H receptor pharmacology. PEPTIDES 14(2) 345-352, 1993.--Angiotensin (A) il receptors on rat aortic smooth muscle (RASM) cell membranes were characterized using the radioligand [~2sI][SarqleS]All ([~2Sl]SI-AII). Angiotensin I, All, and AIII inhibited specific [12Sl]SI-AIl binding, and their rank order of potencies, and Ki values (nM) were: All (3.7) > AI (32.5) > AIII (54.0), which differed from that observed for rat adrenal cortex: All (0.85) > AIII (3.3) >> AI (100). Similar results were observed for RASM membranes in the presence of guanine nucleotides, and for intact cells in the absence or presence of an internalization inhibitor. Lowering the incubation temperature from 37°C to 4°C, or inclusion of PMSF (I mM), and preparing membranes in the presence of EGTA ( 1 mM) altered the rank order of potencies and K~ values (nM) of the angiotensin peptides to: All ( I. 1) > AIII (7.0) ,> AI (144). [~2Sl]Angiotensin I was metabolized completely over the course of 90 min to small (<tetrapeptide) fragments as measured by HPLC. There was no evidence for formation of All or AIII from AI, which would have explained the unusually high potency of AI. [~25I]Angiotensin l metabolism could be attenuated by inhibitors of serine proteases PMSF, aprotinin, and chymostatin. The beneficial effects of PMSF and EGTA suggested that serine protease(s) and metalloproteases contribute to the observed anomalous pharmacological characteristics of AI and AIII, respectively. The RASM cell membranes contained a homogeneous population of binding sites for losartan, and its K~ value differed in the absence (50 nM) or presence (I 6 nM) of protease inhibitors, which suggests that the receptor may also be a target for these peptidases. These data indicate that the significant angiotensinase activity of RASM cell preparations can markedly influence the potencies of angiotensin peptides and nonpeptidic antagonists.

Angiotensin II receptor Smooth muscle cells Protease inhibitors Losartan

THE angiotensin II (All) receptor contributes significantly to hypertension in man and therefore represents a key target for the design of antihypertensive agents. Angiotensin II receptors in the kidney have been distinguished by the relative potencies of angiotensin peptides (12). Thus, the rank order of potency of angiotensin peptides for tubular epithelial receptors ( A I I > AIII > AI) differed from that in glomeruli ( A I I = AIII > AI). The relative potencies of AII and AIII as pressor agents (AI I> AIII) (25) also differed from that for steroidogenesis (AII= AIII) (3,22). AII receptor heterogeneity has also been suggested from studies showing different rank order of potencies of angiotensin peptides for stimulating responses in adrenal medulla, cortex, and vascular smooth muscle (22). The recent introduction of nonpeptidic AII receptor antagonists (37) has provided further evidence that AII receptors do not represent a homogeneous receptor population (4,7,36). Thus, AII receptors that exhibit high affinity for DuP 753 (losartan) are designated AT~ receptors and are the predominant subtype in vascular smooth muscle (36), zona glomerulosa cells of the adrenal cortex (4), hepatocytes (10), and kidney mesangial cells (23). These receptors represent the classical AII receptor that modulates vascular tone, adrenal aldosterone se-

cretion, liver glycogenolysis and gluconeogenesis, and kidney hemodynamics and glomerular filtration rate.

Vascular smooth muscle cells in culture represent a conve- nient model system for studies of drug effects in the vasculature. Thus, rat aortic smooth muscle cells contain predominantly AT, receptors (36), which are coupled to phosphoinositide metabolism, and Ca 2+ mobilization (16,30). Membrane preparations were prepared from cultured rat aortic smooth muscle cells and their pharmacological characteristics determined using radioligand binding techniques. During the course of these studies, it became clear that these characteristics were significantly different from other AT~ receptors. The present study describes our attempts to understand these atypical pharmacological characteristics.

METHOD

Materials

[125I][Sarqleg]Angiotensin II ([125I]SI-AII) (2200 Ci/mmol) and [ ~2sI] [Tyr4]angiotensin I ([~ 25 I]AI ) (2200 Ci/mmol) were obtained from NEN Research Products (Boston, MA). Angiotensin

Requests for reprints should be addressed to Dr. K. Dickinson.

345

346 COHEN, WEBB AND DICKINSON

peptides were from Peninsula Labs (Belmont. CA), and cell culture reagents were from Gibco (Rockville, MD). Guanyly- limidodiphosphate, chymostatin, n-tosyl-L-phenylalanine chloromethyl ketone, bestatin, soybean trypsin inhibitor, and phosphoramidon were from Boeringer Mannheim (Indianapolis, IN). Phenylmethylsulfonyl fluoride, aprotinin, and phenylarsine oxide were from Sigma Chemical Co. (St. Louis, MO). All other chemicals were obtained from Fisher, Sigma, or Mallinckrodt and were reagent grade. DuP 753 (losartan) was synthesized in the Chemistry Dept., Bristol-Myers Squibb Pharmaceutical Re- search Institute.

Rat Aortic Smooth Muscle (IL4SM) Cell Culture and Membrane Preparation

The RASM cells were generously provided by Dr. Marshall Runge (Emory University). These cells were derived from rat aorta as primary smooth muscle cells and maintained in culture up to the 25th passage. The RASM cells were cultured at 37°C in T-75 flasks under humidified 95% air/5% CO2 in HEPES- buffered Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal calf serum, 50 U/ml penicillin, and 50 ug/ml streptomycin (Gibco Labs). Following the attainment of conflu- ence (5-7 days), cells were trypsinized with 2 ml 0.25% trypsin/ 1 trL~,I EDTA, and collected into buffer A [0.1 mM phenylmethylsulfonyl flouride (PMSF), 10 ttg/ml soybean trypsin inhibitor (STI), 20 mM HEPES, pH 7.4, dissolved in DMEM] at a concentration of 3-4 × 105 cells/ml.

Membrane preparation 1. The RASM cells were washed in phosphate-buffered saline, and freeze-thawed in an acetone:dry ice bath and a 37°C water bath. Cells were homogenized using a Brinkmann Polytron (setting #8 for 10 s and setting #10 for 10 s), and centrifuged at 92,000 × g for 45 min at 4°C. The supernatant was discarded and the pellet resuspended in buffer A, transferred to a 50-ml culture tube, and homogenized as described above. The homogenate was stored at - 8 0 ° C until use.

Membrane preparation 2. The cell suspension was washed with buffer A and homogenized in 50 mM Tris-HCl, I mM EGTA, 10 rm~vI MgClz, 0.24 trypsin inhibitor unit (TIU) units/ ml aprotinin, 0.1 mg/ml 1,10-phenanthroline with a Brinkmann Polytron homogenizer (setting #7, 3 × 6 s). The homogenate was passed through two layers of cheesecloth, and centrifuged at 40,000 × g for 20 rain at 4°C. The supernatant was discarded, and the membranes resuspended in buffer and washed three times. The pellet was resuspended in this buffer at a concentration of 0.2-0.8 mg protein]ml. The cell homogenate was stored in l-ml aliquots at - 8 0 ° C until use.

Rat adrenal cortical membrane preparation. Membranes were prepared as described by Chiu et al. (6). In some cases the 13,000 × g centrifugation step was eliminated.

[1251][SaflIleS]Angiotensin II Binding

Assays were conducted in a total volume of 250 #i in tubes arranged in microtiter plate format (Marsh Biomed Coo) . The incubation mixture contained 50/~l [I25I]SI-AII (80-200 pM, 70,000-180,000 cpm), 25 #l displacing drug or AII to define nonspecific binding ( l uM), and incubation buffer. Binding to RASM cell membranes was conducted in the following assay buffer: 150 m M NaC1, 50 m M Tris-HCl, pH 7.4, 5 m M MgCI2, 0.1% bovine serum albumin, 0.24 TIU units/ml aprotinin, 0.1 mg/ml 1.10-phenanthroline. Binding to rat adrenal cotex (RAC) membranes was routinely conducted in the following assay buffer: 50 m~t Tris-HCI, pH 7.4, 5 m M MgCI2, 0.22% BSA. The binding reaction was initiated by the addition of 100 ul membranes (7-12 ug protein) diluted in incubation buffer. The

tubes were incubated at 37°C for 2 h with continuous shaking (Easyshaker. SLT- Labinstruments, A-5082 Grodig, Austria). Bound and free radioligand were separated by simultaneous filtration on a Tomtec TM cell harvester in combination with a ill- termat B that had been presoaked for 1 h in 0.1% polyethylene- imine (PEI) in order to reduce nonspecific binding. The filtermat was rinsed of excess PEI during a prewash cycle, and the membranes were filtered and washed with 150 mM NaCI, 5 mM Tris- HCI, pH 7.4 at 4°C. The filtermat was removed, and micro- waved, membrane side up, at full power for 3 × 2 rain in a microwave oven. The dried filtermat was placed in a sample bag, a sheet of Meltilex TM solid scintillant wax placed on the membrane side of the filtermat, and the Meltilex sheet melted into the mat using a T-Tray Heat-Sealer (Wallac Pharmacia). The impregnated sheet was counted in a Betaplate TM liquid scintillation counter (L.K.B. Pharmacia. T-tray compatible model # 1205) at 60% efficiency. Protein assays were performed on the membrane preparation using B.C.A. reagent, with bovine serum albumin as standard.

Binding data were analyzed by iterative fitting to a one-site model, and inhibition constants (/~]) were calculated from IC50 values. Competition experiments were routinely conducted with 100-200 pM [L~SI]SI-AII.

Analysis of Angiotensin Metabolic Products

[~25I]Angiotensin 1 (0.1 nM) was incubated with RASM membranes under conditions similar to those used for the radioligand binding assay. Total incubation volume was 250/A. The reaction was terminated by adding 25 ~1 of 250 mM TEAP (triethylammonium phosphate; prepared by adding triethyl- amine to 25 mM phosphoric acid until pH = 3.0) followed im- mediately by centrifugation in an Eppendorf microfuge at 15,000 × g for 5 min. The supernatant ( 10-200 td) was applied to a C4 reverse-phase HPLC column (Waters Delta-Pak, 300 ~,. 5 u) that was preequilibrated with 15% acetonitrile in 25 mM TEAP buffer, pH 3.0. The column was developed with an 8-min linear acetonitrile gradient (15-35%) at a flow rate of 2 ml/min. The effluent was collected in a fraction collector at 15-s intervals and the fractions were subjected to gamma spectroscopy. Nonra- dioactive angiotensin peptides were detected by an on-line UV detector set at 215 nm. Separated products were identified by their retention times, which were 6.75, 5.50, and 3.5 rain for [~-~SI]AI, [l:SI]AII, and [~"sI]SI-AII, respectively. The noniodi- hated peptides had the following retention times (rain): AI, 5.68: A(2-10), 5.18: All, 3.42: SI-AII, 1.69: AIII, 2.76: A(3-8), 3.5: A(4-8), 2.06: A(5-8). 1.36: A(1-4), 0.9.

RESULTS

Pharmacological Characterization of All Receptor on RASM Cell Membranes

[t25I][SarqleS]Angiotensin II bound to RASM cell and rat adrenal cortex (RAC) membranes with KD values of 0.3 _+ 0.04 and 0.46 _+ 0.05 nM (n = 3-6) and with Bma~ values of 2.9 _+ 0.7 and 4.1 pmol/mg protein, respectively. Equilibrium was at- tained after 2 and 1 h at 37°C for RASM and RAC membranes, respectively, and was maintained for a further 2 h with RASM membranes (data not shown). Figure IA and B shows the inhibition curves ofAI, AII, and AIII in competition with [~25I]SI- AII binding to RAC and RASM cell membranes. The rank order of potency for inhibition of binding to RAC membranes (AI I> AIII ~> AI) differed from that determined for RASM cell membranes (AII> AI = AIII). Thus, AI was significantly more potent (threefold) and AIII was significantly weaker (sixteenfold) for

A N G I O T E N S I N II R E C E P T O R P H A R M A C O L O G Y 347

I

10

'~ 80' o o

-o 6O t -

O .o = 40 < :I 00

20

0

100'

7 6 5 4

[3

I I

9 8

-Log (Peptide), M

A 100"

80"

60'

40

20

B

10 9 8 7 6 5 4

-Log (Peptide), M

FIG. 1. Inhibition of[J251]SI-AlI binding to rat adrenal cortex (A) and RASM cell (B) membranes by angiotensin peptides. Membranes (25-50 ~g/ml) were incubated with 0.1 nM [~25I]SI-AII and increasing concentrations of AI (I-1), All (©), or AIII (O), and specific binding assessed as described in the Method section. Results show representative inhibition curves of experiments performed five to I 1 times.

the vascular membrane preparations. The calculated Ki values of the angiotensin peptides for RAC and RASM cell receptors were compared with their ICs0 values reported for a number of AII receptors (Table l). In all cases reported in the literature, AII exhibited greater or equal potency to AIII, and AI was substantially (>fiftyfoid) weaker than AII. These pharmacological characteristics were similar to those obtained for rat adrenal cortical receptors. By contrast, RASM cell receptors exhibited high affinity for AII but similar potency for AIII and AI.

One explanation for this unusual pharmacological profile would be different amounts of agonist-induced, high-affinity receptor states in vascular cell membranes, which resulted in different overall IC~o values for the angiotensin peptides. Guanine nucleotides stabilize angiotensin receptors in a low-affinity, G- protein uncoupled state ( l l). Therefore, competition curves were performed in the presence of a nonhydrolyzable guanine nu-

cleotide, Gpp(NH)p. In the presence of this nucleotide, competition curves were right-shifted, and slope factors were close to unity, indicating a homogeneous population of agonist binding sites (Table 2). The K~ values for all three peptides were increased two to threefold, but the unusually high potency for AI and the weak activity of All l was maintained• These data suggest that the atypical pharmacology of RASM angiotensin receptors did not relate to differential G-protein-dependent states of the vascular receptor. These findings also indicate that occupancy of vascular AII receptors by these angiotensin peptides resulted in receptor-G-protein coupling.

Another possible explanation for the unusual pharmacology was the presence of a distinct population of cytosolic receptors that were exposed following membrane preparation. We tested this hypothesis by performing the same binding experiment in intact RASM cells in the presence of 100/,tM phenylarsine oxide

TABLE l

PHARMACOLOGICAL CHARACTERISTICS OF All RECEPTORS FROM VARIOUS TISSUES

IC5o Values (nM)

Tissue AI All AIII Reference

MASMC membranes 500 3 20 (16) N I E- 115 cell membranes 220 2.1 4.0 (14) Neonatal rat myocytes 1 I0 0.8 7 (29) Rat liver 300 1 5.6 (13) Rabbit adrenal 579 1.5 10.8 (13) Rabbit uterus 116 2.5 1.6 (13) Rat adrenal cortex 100 _+ 22 0.85 _+ 0.33 3.3 _+ 0.4 RASM 32.5 _+ 6.2 3.7 _+ 1.0 54.0_+ 7.0

IC50 values (riM) were from the stated literature sources, or were estimated from data therein. Ki values of peptides for rat adrenal cortical and RASM membranes were calculated from ICso values using the equation of Cheng and Prusoff (5). Results are the mean _+ SEM of 3-11 determinations. MASMC = mesenteric artery smooth muscle cell.


TABLE 2 EFFECT OF G U A N I N E NUCLEOTIDE ON THE INHIBITION

CONSTANTS OF PEPTIDES FOR RASM RECEPTORS

Peptide (-)Gpp(NH)p nn (+)Gpp(NH)p n H

At 28.7 +_ 4.2 0.68 + 0.07 78.2 _+ 10.8 0.95 -+ 0.12 AII 2.5+0.4 0.75_+0.11 6.6+ 1.7 0.88_+0.04 AIII 16.8 _+ 8.3 0.77 +_ 0.04 44.6 _+ 7.9 0.92 _+ 0.17

Competition curves for the angiotensin peptides were determined in the absence and presence of 100 uM guanylylimidodiphosphate, and the slopes of the curves (nM) and K~ values (nM) were determined by com- puter-assisted iterative curve fitting. Results show mean _+ SEM of three determinations.

(PAO). This agent inhibits internalization of AII receptors in vascular smooth muscle cells (19), and after treatment, only cell surface receptors were available for binding. We have shown previously that [~25I]AII is internalized by RASM cells following incubation for 2 h at 37°C (35). Inhibition of [t25I]SI-AII binding by AI, AII, and AIII was examined in the absence and presence of PAO. In control cells the order of potencies, and K, values (nM) were AII (0.4) > AI (2.0) >_ AIII (4.6), which was similar to data obtained in the presence of PAO, AII (1.0) > AI (3.8) > AIII (14.3). The unusually high potency of AI for the cell surface binding sites, either in the absence or presence of receptor internalization, correlated with data obtained with membrane preparations, and suggest that binding was not to a cytosolic receptor.

Metabolism oral in RASM Membranes

Another possible explanation for the unusual binding pharmacology was the presence, on cells and in membrane preparations, of metabolizing enzymes that degraded the angiotensin peptides to more potent (in the case of AI) or less potent (in the case of AIII) inhibitors of ['2sI]SI-AII binding than their corre- sponding parent peptides. In order to investigate angiotensin peptide metabolism by RASM membranes, AI was exposed to RASM membranes (from different passages) under identical conditions to those used in the binding assay (2 h at 37°C). Angiotensin I was measured by radioimmunoassay. Although a degree of variability was observed in the ability of membranes to metabolize AI, it was clear that some membrane preparations had sufficient peptidase activity to completely metabolize A[ (data not shown). In view of these findings, HPLC methods were developed in order to identify metabolites of AI, and the elution protocol was optimized to detect the production of AII or AIII.

In general, the retention times (RT) of angiotensin peptides were directly proportional to the length of the peptides (see the Method section). The RTs ranged from 5.68 rain for the deca- peptide AI to 0.90 min for the tetrapeptide, angiotensin(l-4). The only exception was the hexapeptide AII(3-8), which eluted after the octapeptide AII. When AII was substituted in the l position with sarcosine and in the 8 position with isoleucine, the RT was reduced by approximately one-half. The iodinated peptides had RTs 1-2 min longer than their noniodinated coun- terparts. Figure 2A shows a time course of [125I]AI metabolism by RASM membranes at 37°C. [~25I]Angiotensin I (0.1 n_M) was incubated with l0 pg RASM membranes for various time periods and the supernatant subjected to reverse-phase HPLC. At 0 min the majority of the [z25I] appeared at 6.75 min, which was consistent with the retention time for authentic [125I]AI. This peak decreased over the course of 90 min and reappeared at a RT of

1 rain, which was consistent with a peptide fragment containing four or less amino acids. There was no evidence of radioactive products that corresponded to [~25I]AII or [~2-~I]AIII, and at the end of 90 rain all the [125I]AI was metabolized. Similar results were obtained when noniodinated AI (I nmol) was incubated with RASM membranes for 90 rain at 37°C. Under these conditions, AI (RT = 5.73 min) was completely metabolized to products that eluted with a RT = 1.66 rain or that eluted with the solvent front (data not shown). These findings indicate that RASM membranes have significant peptidase activity that me- tabolizes AI directly to small (<tetra?) peptide fragments.

In order to investigate the possibility that AII or AIII was formed at early time intervals, which subsequently led to formation of the small peptide fragment, [~25I]AI (0.1 riM) was incubated with 10 #g RASM membranes for 1-l0 rain and the products subjected to HPLC. Figure 2B shows [L25I]AI decreased slightly from 0-10 min and a small peak appeared at a RT of 1.25 min. This peak was similar to the peak that was formed during the extended time course (Fig. 2A). No other peaks emerged, indicating that no intermediates, such as AII or AIII. were formed from AI at short time intervals.

When [~25I]AI was incubated with RASM membranes in the presence of the serine protease inhibitor chymostatin, metabolism was significantly attenuated. Figure 3 shows [~zsI]AI (6.75 rain) decreased significantly (74%) over the course of 30-min incubation with RASM membranes at 37°C. The product (RT = 1.0 rain) was similar to that shown in Fig. 2. In the presence

20000 1

15000"]

10000

0 2 _ 0 2

[ 1 2 5 1 ] - A I I [ 1 2 5 1 ] - A I

~' ~U ( A )

. . . . . = i

4 6

T i m e ( m i n )

• 0 rain - - - o , - - 15 min . . . . . ='"" 30 min . . . . * - - - 60 min . . . . "*"'-" 90 rain

L _

- - ' - ~ " " " 1'0

30000"

~" 25000" ¢g.

o,~ 20000"

,.,.g - - 15000 ' 04

- ' - 10000.

5000"

(n)

• lmin --.-...-o-- 2 min . . . . . t - - - 5 rain . . . . o-.-- 10 rain

. i b _ _ ,, .J L

1 2 3 4 5 6 7 8

T i m e ( m i n )

FIG. 2. Time course of [~'Sl]Al metabolism by RASM cell membranes. [~25I]Angiotensin I (0.1 riM) was incubated with RASM membranes for 0-90 rain (A) or 0-10 min (B) under the same conditions used for radioligand binding studies, and the supernatant was subjected to reverse- phase HPLC as described in the Method section. Fractions were collected and radioactivity measured by gamma counting. The arrows show the retention times of radiolabeled AI and All.

ANGIOTENSIN II RECEPTOR PHARMACOLOGY 349

15000

E 1oo00 o,9,

¢~ 5000

o 2 4 6 T ime (rnin)

8 10

FIG. 3. Effect of chymostatin on [t25I]Al metabolism by RASM cell membranes. [t251]Angiotensin I (0.1 nM) was incubated with RASM membranes for 30 min at 37°C in the absence (11) or presence (©) of 100 ~tM chymostatin, and the supematant was subjected to reverse-phase HPLC. Fractions were collected and radioactivity measured by gamma counting.

of chymostatin (100 #M) only 26% of the [~251]AI was metabolized. Similar results were obtained with PMSF, aprotinin, and n-tosyI-L-phenylalanine chloromethyl ketone as the serine protease inhibitors (data not shown). Thus, AI metabolism could be attenuated by the inclusion of inhibitors of serine proteases.

The radioligand [~25I]SI-AII was much less affected by RASM membranes, and after 2 h at 37°C, 92% of the ligand was intact (data not shown). These data suggest that the radioligand was not the target for RASM proteases, and its metabolism was not responsible for the unusual pharmacological characteristics of RASM membranes.

Other potential products of AI metabolism included AI(2- 10) and AI(1-9), which could be derived by aminopeptidase and carboxypeptidase activity, respectively. The ability of these peptides to inhibit [t251]SI-AII binding to RASM cell membranes was determined together with the AII fragments (3-8) and (5- 8). The Ki values of A(2-10), A(l-9), A(3-8), and A(5-8) were 0.11, 0.42, 3.8, and >100 uM. respectively. The low affinity of these fragments for the receptor suggested that the formation of these peptides would not account for the increased potency of AI.

Table 3 shows the effect of a variety of protease inhibitors on the K~ values of AI for RASM cell membranes, lnhibitors of ACE, carboxypeptidase A and B, aminopeptidase, neutral endopeptidase, or trypsin had little influence on the Ki value for AI, which suggested that these enzymes were not responsible for metabolism of AI. At high concentrations, PMSF (1 mM) significantly (p < 0.05) increased the Ki value of AI from 32.5 + 6 to 144 +_ 15 nM, which suggested that a serine protease was responsible for the unusually high potency of AI. Lowering the incubation temperature to 4°C also increased the K, value for AI (K~ = 426 ___ 232 nM). These findings suggested that AI was metabolized by a non-ACE enzyme(s), which was sensitive to serine protease inhibitors. A more rigorous membrane preparation, which included homogenization in EGTA-containing buffers and multiple washes (preparation 2, see the Method section), was particularly effective in decreasing the Ki value of AIll (Ki = 7 +_ 2.5 nM), whereas the Ki value for AI was less influenced by this protocol.

Figure 4 shows inhibition curves for the angiotensin peptides in competition with [125I]SI-AII (0.1 nM) binding to RASM membranes (prepared using protocol 2, see the Method section) using standard assay conditions including 1 mM PMSF. Under these conditions, the potency of AI was significantly decreased and the potency of AIII was markedly increased. The rank order of potency was changed from All > AI = AIII to All > AIII ,> AI (Table 4). Thus, the unusual pharmacology observed with these peptides was likely due to the presence of peptidase activity in the RASM cell preparation.

Table 4 shows that the inclusion of PMSF in the incubation mixture, and use of a more rigorous membrane preparation, also decreased the K~ value oflosartan for RASM cell membranes threefold, from 50.3 nM to 16.2 nit/. Since this nonpeptidic drug was unlikely to be the target for RASM cell proteases, it is possible that proteolysis of the receptor may also have contrib- uted to these affinity changes. The slope factor of losartan (nil = 0.91 _+ 0.03) was not significantly different from unity, and was unaffected by peptidase inhibitors, which suggested that RASM membranes used for these studies contained a homogeneous population of AT~ receptors.

DISCUSSION

This study describes our attempts to validate rat aortic smooth muscle cells as a model system for investigations of the action of All in the vasculature. These cells have been successfully used

TABLE 3 EFFECT OF PROTEASE INHIBITORS ON THE K, VALUE OF AI

FOR RASM CELL MEMBRANES

Inhibitor K~ (nM) Protease Target

Control 33 + 6 Plummer's/benzylsuccinate, 10 #M 45 Carboxypeptidase A or B Captopril, 10/aM 13 ACE Fosinoprilic acid, 10 #M 14 ACE Bestatin, 100 uM 11 Aminopeptidase EDTA, 1 mM 9.9 + 5.9 MetaUopeptidase Soybean trypsin inhibitor, 100/~M 41 Trypsin Phosphoramidon, 10 uM 6 Neutral endopeptidase PMSF, I mM 144 _+ 15 Serine peptidase

Experiments were conducted under standard assay conditions, and the protease inhibitors were included at the stated concentrations. Mean K~ values were derived from IC50 values using the Cheng and Prusoff equation (5).


100

6 o

,4o ~ 20

0 11

[]

i i i i ~ i 10 9 8 7 6 S 4

- Log [Peptide], M

FIG. 4. Effect of PMSF on angiotensin peptide binding to RASM cell membranes. The RASM membranes (prepared by protocol 2, see the Method section) were incubated with 0.1 m~l ['2Sl]SI-AII and increasing concentrations ofAl (rq), All (©), or AIII (I) in the presence of 1 mM PMSF. Specific binding is plotted against peptide concentration. Results are representative of experiments performed three to five times.

to study the biochemical sequelae that follow receptor occupancy by All (16,30), although a detailed characterization of the angiotensin receptor on these cells has not been reported. We were therefore intrigued when the rank order of potencies of the angiotensin peptides for our RASM cell membranes (All > AI = AIII) differed substantially from the rank order obtained in other All-containing tissues (All > AIII ,> AI). Douglas et al. (12) described different rank order of potencies for kidney epithelial and glomerular All receptors, but this difference related to the potency of AllI. There were no reports of an angiotensin receptor that exhibited equal potency for AI and AIII. The reasons for this unusual potency order could have related to: different agonist-induced, G-protein interactions of the peptides; the presence in our membrane preparation of a pool of cytosolic binding sites with different characteristics to those on the cell surface; the presence of metabolizing enzymes that generated peptides with different potencies to the parent peptides; or enzymes that proteolyzed the radioligand or receptor and changed its pharmacological characteristics.

There was no evidence that the agonist-induced, G-protein interaction was substantially different for the angiotensin peptides. Thus, AI, All, and AIII competition curves were all mod- ulated by guanine nucleotides, which implied that the RASM All receptor was coupled via a G-protein to its effector system. In the presence of guanine nucleotides, the competition curves had slopes closer to unity but the relative potencies of AI and AIII were maintained.

The possibility that the membrane preparation contained a pool of cytosolic receptors that differed from those on the cell surface was tested by conducting binding studies with intact cells. Softer et al. (31 ) have described a cytosolic angiotensin receptor from rabbit liver with different pharmacological properties to classical All receptors. However, demonstration of this binding site has been reported to require organomercurials in the incubation buffer. Since AII receptors are known to internalize upon exposure to All at 37°C (32), these experiments were conducted in the absence and presence of the internalization inhibitor PAO (16,19). Thus, control binding represented binding to cell surface

and internalized receptors, and binding to PAO-treated cells was to cell-surface receptors. Results indicated similar pharmacological characteristics under both conditions, which suggested that there was little difference in the potency of All peptides in the absence or presence of receptor internalization. Moreover, AI and AIII had equal affinities for cell surface angiotensin receptors that were similar to results obtained with membranes. These data do not support the concept that the membrane binding sites differed substantially from those on intact cells.

Degradation of angiotensin peptides to smaller fragments has been shown in a variety of tissues (22,26) and by a number of diverse peptidases including: angiotensin converting enzyme (9). aminopeptidases (21), post-proline cleaving enzyme (34). and chymases (33). Thus, we considered it likely that AI and AIII were degraded during the course of the incubation, in spite of our use of aminopeptidase (l,10-phenanthroline) and serine protease (aprotinin) inhibitors in the incubation buffer. Recent reports suggested that AI can be rapidly degraded to All in RASM cells by ACE, and the observed AI-stimulated Ca t+ transient was due to the formation of All in a captopril-sensitive reaction (2). Moreover, the membranes used for our early studies represented a relatively crude preparation, which may have contained intrinsic and extrinsic membrane-bound proteases. Thus, formation of All or AIII from AI seemed a possible explanation for the unusually high potency of AI. Using HPLC techniques to separate angiotensin peptides, or direct radioimmunoassay of AI, indicated that AI was indeed subject to metabolism by RASM membranes. In some cases, all the AI was degraded to immunologically distinct peptide products during 2-h incubation at 37°C. conditions that reproduced the binding assay. Surpris- ingly, degradation of AI by RASM peptidases was not due to ACE, since captopril or fosinoprilic acid had no effect on the potency of AI. The enzyme responsible for the metabolism of AI was unlikely to be a carboxypeptidase, neutral endopeptidase, or aminopeptidase, since specific inhibitors failed to alter the K, value of AI. Lowering the incubation temperature, or inclusion of PMSF, shifted the K, value of AI threefold, which suggested that a temperature-dependent serine protease was responsible for the unusually high potency of AI. Interestingly, the serine protease inhibitor STI did not alter the K, value of Ai. which suggested that the enzyme was not t~psin.

Recently a chymotrypsin-like enzyme has been described in human heart (33) and human skin mast cells (27), which was able to cleave the PheS-His 9 bond of AI to generate All. This

TABLE 4

EFFECT OF PMSF ON THE PHARMACOLOGICAL POTENCIES OF ANGIOTENSIN PEPTIDES

K~ (nM3

Agent PMSF + PMSF

AI 32.5_+ 6.2 144 _+ 15" All 3.7 _+ 1.0 1.1 _+ 0.2* AIII 54.0 + 7.0 7.0 + 2.5* Losartan 50.3 _+ 13.7 16.2 + 4.0*

Competition curves for the stated peptides and drugs were conducted under standard assay conditions using membranes prepared according to protocol 1 (see the Method section) or to membranes prepared according to protocol 2, and supplemented with PMSF. Results show the mean Ki values _+ SEM of 3-13 experiments.

* Significantly different from control value p < 0.05.

ANGIOTENSIN II RECEPTOR PHARMACOLOGY 351

cardiac endopeptidase was inhibited by serine protease inhibitors such as PMSF, STI, and chymostatin, but not aprotinin. These properties differ from the RASM enzyme responsible for AI metabolism, which was sensitive to aprotinin (data not shown) and relatively insensitive to STI (Table 3). However, rat chymase (mast cell protease I) has a different substate specificity to human heart chymase and cleaves the Tyra-lle 5 bond of AI, yielding A(1-4) and A(5-10), but it was unable to generate All (20). Our HPLC studies of AI metabolism by RASM membranes did not indicate any significant formation of either AII or AIII, even after short times of incubation, and the predominant products were small peptides of four or less amino acids. These data are consistent with cleavage of AI via an endopeptidase that is sensitive to the serine protease inhibitors PMSF, aprotinin, and chymostatin.

The presence of the metalloprotease inhibitor EGTA during membrane preparation significantly decreased the Ki value for AIII and slightly increased the value for AI. Abhold et al. (1) showed that metabolism of AIII by membrane-bound peptidases was increased when aminopeptidase inhibitors were present. They suggested that other angiotensinases, such as carboxypep- tidases and endopeptidases, which would cleave the Pro7-Phe s and Tyr4-lle 5 bonds ofangiotensins, would be activated following inhibition of aminopeptidases. Regoli et al. (26) have described the major routes of metabolism of AlI in rat serum and kidney homogenates to be due to aminopeptidases and a chymotrypsin- like endopeptidase that hydrolyzed AII at the Tyrn-lle 5 bond. Our results with RASM membranes would be consistent with the presence of a chymase endopeptidase that hydrolyzed AI, and EGTA-sensitive aminopeptidases that hydrolyzed AIII, although it is likely that the peptides could be substrates for both enzymes. However, since the biological activity of these products would be expected to be less than that of AI, their formation could not explain the increased potency orAl exposed to RASM membranes. The location of these metabolizing enzymes was likely to be on the cell surface, since the anomalous potency of AI was also seen in intact cells.

Other potential substrates for these proteolytic enzymes in- clude radioligand and the receptor. HPLC of [~25I]SI-AII incubated with membranes for 2 h at 37°C indicated little metabolism of radioligand. Sarcosyl-substituted angiotensins are less effective substrates for aminopeptidases, although metabolism may still occur (24). Inhibition of radioligand degradation by inhibitors is normally associated with increased specific binding of radioligand ( 1 ), and we did not consistently observe this phenomenon when PMSF was included in the incubation buffer. Moreover, changes in the potencies of drugs would be expected to be uni- directional if the radioligand was the target site for the RASM peptidases, and our results indicated both increased (AIII) and decreased (AI) affinities in the presence of these protease inhibitors. The potency of the nonpeptidic antagonist losartan was increased threefold by inclusion of PMSF during incubation, and use of a rigorous membrane preparation. Since this com- petitor was not likely to be a substrate for a metabolizing enzyme, the beneficial effect of these inhibitors on the Ki value for losartan suggests that the inhibitors may be protecting the receptor from proteolysis. Robertson et al. (28) have recently shown that the potency of losartan for inhibition of All- or Alll-stimulated contraction of rabbit aorta was little affected by the aminopep-

tidase inhibitor amastatin. These differences may relate to the longer incubation times of binding assays relative to those used for contractile measurements.

Fujimoto et al. (15) have recently examined the effects of peptidase inhibitors on the IC5o values of angiotensin peptides for porcine aortic smooth muscle cell membranes. They showed that whereas the affinity of AII or SI-AII were little affected by amastatin and PMSF, the affinity of AIII was significantly increased by inclusion of these protease inhibitors. Thus, these authors suggested the existence ofamastatin- and PMSF-sensitive peptidases in pig aortic membranes that selectively degrade AIII. Robertson et al. (28) have also shown that the potency of AIII (but not All) to contract rabbit aorta was increased by inclusion of aminopeptidase inhibitor amastatin. It is likely that the peptidases responsible for degradation of AIII and AI are particularly rich in vascular tissues, since AIII and AI exhibit the appropriate affinities for rat adrenal cortical All receptors even in the absence of protease inhibitors (Table !). PC I2W cell membranes, which express AT2 receptors (35), also exhibit anomalous Ki values for AI and AIII unless endogenous peptidase activity is controlled (unpublished observations).

Under conditions where RASM protease activity was inhibited, the potency order for the angiotensin peptides All > AIII ,> AI was identical to that for other All receptors (Table 1). Recently, angiotensin receptor subtypes have been differentiated by the use of nonpeptidic antagonists, which show selective affinity for AT~ or AT2 receptors (4,7,36). Losartan was a potent inhibitor of [t25I]SI-AII binding to RASM AII receptors, and its calculated K~ value (16.2 nM) was similar to its KB value (3.3 nM) for inhibition of AIl-stimulated contraction of rabbit aorta (8). Competition curves had slopes close to unity, which con- firmed reports that the RASM cells contained a single homogeneous population of ATI receptors (37).

In summary, RASM cells contain a homogeneous population of AT~ receptors, as defined by the selective nonpeptidic antagonist losartan. The potency of AI to inhibit [~25I]SI-AII binding to RASM membranes was greater than that expected from literature values. The appropriate potency of AI could be restored by inclusion of the serine protease inhibitor PMSF, or by lowering the incubation temperature. Angiotensin III exhibited decreased potency for the All receptor, which could be increased by preparing membranes in the presence of EGTA. These observations could be explained by metabolism of these angiotensin peptides to products that had increased (for AI) and decreased potency (for AIII) for the receptor. Although AI was actively metabolized by RASM membranes, and this metabolism was attenuated by serine protease inhibitors, we observed no evidence for formation of All or AIII. The products were small (tetra ?) peptides that were unlikely to have increased potency for the receptor. The radioligand was not a target for the RASM proteases. The affinity of the nonpeptidic drug losartan was increased by serine protease inhibitors. Since this drug was unlikely to be a substrate for proteases, it is likely that these anomolous pharmacological characteristics derived, in part, from proteolysis of the All receptor.

ACKNOWLEDGEMENT

The authors thank Drs. Bud Weller and Denis Ryono for synthesis of DuP 753.

REFERENCES

1. Abhold, R. H.; Sullivan, M. J.; Wright, J. W., Harding, J. W. Binding, nopeptidase inhibition with amastatin and bcstatin. J. Pharmacol. degradation and pressor activity ofangiotensins II and III after ami- Exp. Ther. 242:957-962" 1987.

352 COHEN, WEBB A N D DICKINSON

2. Andre, P.; Schott, C.', Nehlig, H.; Stoclet, J-C. Aortic smooth muscle cells are able to convert angiotensin I to angiotensin I!. Biochem. Biophys. Res. Commun. 173:1137-1142; 1990.

3. Campbell, W. B.; Brooks, S. N.; Pettinger, W. A. Angiotensin ll- and angiotensi n Ill-induced aldosterone release in vivo in the rat. Science 184:994-996; 1974.

4. Chang, R. S. L.; Lotti, V. J. Two distinct angiotensin II receptor binding sites in rat adrenal revealed by new selective nonpeptide ligands. Mol. Pharmacol. 37:347-351; 1990.

5. Cheng, Y. C.; Prusoff, W. H. Relationship between the inhibition constant (Ki) and the concentration of the inhibitor which causes 50 percent inhibition (I50) of an enzymatic reaction. Biochem. Pharmacol. 22:3099-3108; 1973.

6. Chiu, A. T.; Carini, D. J.; Johnson, A. L.; McCall, D. E.; Price, W. A.; Thoolen, M. J. M. C.; Wong, P. C.; Taber, R. I.: Timmermans P. B. M. W. M. Non-peptide angiotensin II receptor antagonists, II. Pharmacology of S-8308. Eur J. Pharmacol. 157:13-21 ; 1988.

7. Chiu, A. T.; Herblin, W. F.; McCall, D. E.; Ardecky, R. J.; Carini, D. J.; Duncia, J. V.; Pease, L. J.; Wong, P. C.: Wexler, R. R.; Johnson, A. L.; Timmermans, P. B. M. W. Identification of angiotensin II receptor subtypes. Biochem. Biophys. Res. Com mun. 165:196-203: 1989.

8. Chiu, A. T.; McCall, D. E.: Price, W. A.; Wong, P. C.; Carini, D. J.; Duncia, J. V.; Wexler, R. R.: Yoo, S. E.; Johnson, A. L.; Timmermans, P. B. M. W. M. Non-peptide angiotensin II receptor antagonists, VII. Cellular and biochemical pharmacology of DuP 753. an orally active antihypertensive agent. J. Pharmacol. Exp. Ther. 252:711-718; 1990.

9. Cushman, D. W.; Wang, F. L." Fung, W. C.; Harvey, C. M.; De- Forrest, J. M. Differentiation of angiotensin-converting enzyme (ACE) inhibitors by their selective inhibition of ACE in physiolog- ically important target organs. Am. J. Hypertens. 2:294-306: 1989.

10. De Gaspero, M.; Whitebread, S.; Mele, M.; Motani, A. S.; White- combe, P. J.; Ramjoue, H-P.; Kamber, B. Biochemical characterization of two angiotensin II receptor subtypes in the rat. J. Car- diovasc. Pharmacol. 16:$31-$35; 1990.

11. Delean, A.; Ong, H.; Gutkowska, J.; Schiller, P. W.; McNieoll. N. Evidence for agonist-induced interaction ofangiotensin receptor with a guanine nucleotide-binding protein in bovine adrenal zona glomerulosa. Mol. Pharmacol. 26:498-508: 1984.

12. Douglas, J. G. Angiotensin receptor subtypes of the kidney cortex. Am. J. Physiol. 253:F1-F7; 1987.

13. Dudley, D. T.; Panek, R. L.; Major, T. C.; Lu, G. H.; Bruns, R. B.: Klinkefus, B. A.; Hodges, J. C.; Weishaar, R. E. Subclasses ofangiotensin II binding sites and their functional significance. Mol. Pharmacol. 38:370-377; 1990.

14. Fluharty, S. J.: Reagan, L. P. Characterization of binding sites for the angiotensin II antagonist ~:5l-[Sarc~,lleS]-angiotensin II on murine neuroblastoma N 1 E- 115 cells. J. Neurochem. 52:1393-1400; 1989.

15. Fujimoto, M.; Mihara, S.; Shigeri, Y.; ltazaki, K. Possible implication of peptidase activity in different potency of angiotensins 1I and III for displacing [~25t]anglotensin Ill binding in pig aorta. Eur. J. Phar- macol. 215:259-264; 1992.

16. Griendling, K. K.; Berk, B. C.; Ganz, P.; Gimbrone, M. A.; Alex- ander, R. W. Angiotensin II stimulation of vascular smooth muscle phosphoinositide metabolism. Hypertension 9(Suppl. III):181 - 185; 1987.

17. Gunther, S.; Gimbrone, M. A.; Alexander, R. W. Identification and characterization of the high alSnity vascular angiotensin II receptor in rat mesenteric artery. Circ. Res. 47:278-286; 1980.

18. Jenne, D. E.; Tschopp, J. Angiotensin II-forming heart chymase is a mast-cell-specific enzyme. Biochem. J. 276:567-568:199 I.

19. kessegue, B.; Alexander, R. W.; Clark, M.: Griendling, K. K. An- giotensin ll-induced phosphatidylcholine hydrolysis in cultured ~as- cular smooth muscle cells. Biochem. J. 276:19-25:1991.

20. Le Trong, H.; Neurath, H.: Woodbury, R. G. Substrate specificity of the chymotrypsin-like protease in secretory granules from rat mast cells. Proc. Natl. Acad. Sci. USA 84:364-367; 1987.

21. Palmieri. F. E.: Bausback, H. H.; Ward, P. E. Metabolism of vaso- active peptides by vascular endothelium and smooth muscle aminopeptidase M. Biochem. Pharmacol. 38:173-180; 1989.

22. Peach, M. J. Renin-angiotensin system: Biochemistry and mecha- nism of action. Physiol. Rev. 57:313-370: 1977.

23. Pfeilschifter. J. Angiotensin II B-type receptor mediates phosphoinositide hydrolysis in mesangial cells. Eur. J. Pharmacol. 184:201- 202; 1990.

24. Regoli. D.: Rioux. F.; Park, W. K.; Choi, C. Role of the N-terminal amino acid for the biological activities ofangiotensin and inhibitory analogs. Can. J. Physiol. Pharmacol. 52:39-49: 1974.

25. Regoli. D.; Park. W. K.; Rioux, F. Pharmacology ofangiotensin I1. Pharmacol. Rev. 26:69-123: 1974.

26. Regoli, D.: Riniker, B.: Brunner, H. The enzymatic degradation of various angiotensin II derivatives by serum, .,lasma or kidney homogenates. Biochem. Pharmacol. 12:637-646; 1963.

27. Reilly, C. F.: Tewksbury, D. A.: Schechter, N. M.: Travis, J. Rapid coversion ofangiotensin I to angiotensin II by neutrophil and mast cell proteinases. J. Biol. Chem. 257:8619-8622: 1982.

28. Robertson. M. J.: Mootoosamy, P.; Clark, K. k. Effects ofpeptidase inhibition on angiotensin receptor agonist and antagonist potency in rabbit isolated thoracic aorta. Br. J. Pharmacol. 106:166-172; 1992.

29. Rogers, T. B.: Gaa, S. T.: Allen, 1. S. Identification and characterization of functional angiotensin II receptors on cultured heart myo- c,vtes. J. Pharmacol. Exp. Ther. 236:438-444; 1986.

30. Smith. B. J.: Smith, k. S.: Brown, E. R.: Barnes, D.: Sabir, M. A.: Davis, J. S.: Farese, R. V. Angiotensin II rapidly increases phos- phatidate-phosphoinositide hydrolysis and mobilizes intracellular calcium in cultured arterial muscle cells. Proc. Natl. Acad. Sci. USA 81:7812-7816: 1984.

31. Softer, R. k.: Kiron, M. A. R.: Rosenberg. E.; Bandyopadhyay, S. Angiotensin receptor: Properties and relevance to the development of antagonists. Kidney Int. 34(Suppl. 26):S-69-S-72: 1988.

32. Ullian, M. E.; Linas, S. L. Angiotensin II surface receptor coupling to inositol trisphosphate formation in vascular smooth muscle cells. J. Biol. Chem. 265:195-200: 1989.

33. Urata, H.: Kinoshita, A.: Misono, K. S.; Bumpus, F. M.: Husain, A. Identification of a highly specific chymase as a major angiotensin II-forming enzyme in the human heart. J. Biol. Chem. 265:22,348- 22,357: 1990.

34. Ward, P. E.: Bausback, H. H.: Odya, C. E. Kinin and angiotensin metabolism by purified renal post-proline cleaving enzyme. Biochem. Pharmacol. 36:3.187-3,193: 1987.

35. Webb, M. L.; Liu, E. C-K.; Cohen, R. B.; Hedberg, A.: Bogosian, E. A.; Monshizadegan, H.; Molloy, C.; Serafino, R.; Moreland, S.; Murphy, T. J.: Dickinson, K. E. J. Molecular characterization of angiotensin III type II receptors in rat pheochromocytoma cells. Peptides 13:499-508: 1992.

36. Whitebread, S.; Mele, M.; Kamber, B.; de Gasparo, M. Preliminary.' biochemical characterization of two angiotensin II receptor subtypes. Biochem. Biophys. Res. Commun. 163:284-291: 1989.

37. Wong, P. C.: Price. W. A.: Chiu, A. T.; Duncia, J. V.: Carini, D. J.; Wexler, R. R.; Johnson, A. L.: Timmermans, P. B. M. W. M. Non- peptide angiotensin II receptor antagonists. VIII. Characterization of functional antagonism displayed by DuP 753, and orally active antihypertensive agent. J. Pharmacol. Exp. Ther. 252:719-725: 1990.

Documents

Peptidases and smooth muscle cell angiotensin II receptor pharmacology