Mechanism of Adrenomedullin-Induced Relaxation in Isolated Canine Retinal Arteries

Tomio Okamura* Kazuhide Ayajiki,* Kenji Kangawa,^ and Noboru Toda*

Purpose. To analyze the mechanism of action of adrenomedullin (AM), a peptide recently isolated from human pheochromocytoma, in isolated canine central retinal arteries and to compare the action of calcitonin gene-related peptide (CGRP). Methods. Changes in isometric tension were recorded in helical strips of the arteries with and without the endothelium. 9 Results. Both AM and CGRP produced relaxation: EC50s were 2.62 and 0.71 X 10~ mol/1, respectively, and maximal relaxations were 85.1% and 84.3%, respectively. The AM-induced relaxation was endothelium-independent and unaffected by indomethacin, NG-nitro-L-arginine, methylene blue, and glibenclamide. Treatment with [8-37]CGRP markedly inhibited the relaxations caused by AM and CGRP. Treatment with a high concentration of sodium nitroprusside abolished the relaxation caused by nitroglycerin and atrial natriuretic peptide and reduced the relaxation caused by AM and CGRP. A high concentration of beraprost, a stable analog of prostaglandin I2, suppressed the response to AM and CGRP but not to nitroglycerin. Conclusions. Endothelium-independent relaxations to AM of canine retinal arteries may be mediated primarily by intracellular cyclic adenosine monophosphate by stimulation of CGRPi receptors and partially by cyclic guanosine monophosphate; cyclic guanosine monophosphate is unlikely to be produced by methylene blue-sensitive soluble guanylate cyclase. Prostanoids, nitric oxide, and adenosine triphosphate-dependent K+ channel opening do not appear to be involved in the AM-induced relaxation. Invest Ophthalmol Vis Sci. 1997;38:56-61.

Adrenomedullin (AM) is a recently discovered peptide Adrenomedullin shows homology in chemical in human pheochromocytoma that consists of 52 amino structure with calcitonin gene-related peptide acids.1 It distributes in a variety of organs and tissues, (CGRP), a potent endogenous vasodilator, and both including blood vessels, kidneys, and heart2 and it pro- compounds reportedly share the same receptor in the duces hypotension, diuresis, and natriuresis.3'4 However, perfused rat mesenteric artery6 and rat cultured vascu- actions of this peptide on isolated blood vessels are incon- lar smooth muscle cells7 but not in rat kidneys.4 Ra- sistent and relatively weak, and tachyphylaxis rapidly de- dioligand-binding studies do not always support the velops (unpublished data, 1996). Heterogeneity of vascu- idea that the AM action is mediated by CGRP recep- lar actions also is reported.5 Among isolated canine blood tors.8 Involvement of endogenous nitric oxide (NO) vessels so far tested, we found that central retinal arteries in the vasodilator action of AM have been reported.3'9 responded to the peptide with consistent, marked relax- In the current study, we sought to determine the ations that were reproduced by repeated trials. Such a relaxant action of AM in isolated canine retinal arter- preliminary result led us to speculate that AM may have ies in reference to the endothelium and NO, to com- selective vasodilator actions on ocular circulation and that pare its action with that of CGRP, and to analyze phar- the retinal artery is appropriate for the analysis of charac- macologically the mechanism of AM action. teristic features of AM action on blood vessels. METHODS From the * Department of Pharmacology, Shiga University of Medical Science, Ohtsu, and the \ National Cardiovascular Research institute, Osaka, Japan. Preparations of Isolated Central Retinal Submitted for publication June 3, 1996; revised August 30, 1996; accepted September 5, 1996. Arteries Proprietary interest cateogy: N. All experimental procedures that used animals con- Reprint requests: Noboru Toda, Department of Pharmacology, Shiga University of Medical Science, Seta, Ohtsu 520-21, Japan. formed to the ARVO Statement for the Use of Animals

Downloaded from iovs.arvojournals.org on 09/27/2021 Adrenomedullin in Retinal Artery 57

in Ophthalmic and Vision Research. The institutional review board at our university approved the use of animal blood vessels in this study. Mongrel dogs of either sex, each weighing 8 to 14 kg, were anesthetized with intravenous injections of sodium pentobarbital (30 mg/kg) and were killed by bleeding from the carotid arteries. The eyeballs attached with optic nerves, and extraocular tissues were removed rapidly from the orbital cavities. Retinal central arteries (out- side diameter, 0.3 to 0.5 mm) were isolated and cut into helical strips approximately 20 mm in length. 100 L 9 100 Special care was taken to avoid the endothelial dam- 10" 10-10 1Q-9 age of the strips. The specimens were vertically fixed Cone, of AM (M) Cone, of AM (M) between hooks in a muscle bath of 20 ml capacity, FIGURE l. Concentration-response curves for adrenomedul- containing the modified Ringer-Locke solution main- lin (AM) in canine central retinal arteries at repeated trials tained 37°C ± 0.3°C and aerated with a mixture of {left) and with and without the endothelium (right). Arterial 95% O2 and 5% CO2. The hook anchoring the upper strips were contracted with prostaglandin F2fV. Relaxations end of the strip was connected to the lever of a force- induced by 10~4 mol/1 papaverine were taken as 100%. displacement transducer. Resting tension was adjusted to 0.7 g,10 which is optimal for inducing the maximal contraction. The composition of the bathing medium bing the intimal surface with a cotton ball. Endothelial was as follows (in mol/1): NaCl, 120; KC1, 5.4; CaCl2, function was determined by the relaxant response to 8 10 2.2; MgCl2, 1.0; NaHCO3, 25.0; and dextrose 5.6. The 10~ mol/1 substance P. Unless otherwise men- pH of the solution was 7.35 to 7.42. Before the start tioned, the endothelium was not actively removed of the experiments, all strips were allowed to equili- from the strips. brate for 60 to 90 minutes in the bathing medium, during which time the fluid was replaced every 10 to Statistics and Drugs 15 minutes. Results shown in the text, figures, and table were ex- pressed as mean values ± SE. Statistical analyses were Tension Recording made using the Student's paired and unpaired t-test Isometric mechanical responses were displayed on an for two groups and the Tukey's method after one-way ink-writing oscillograph. The contractile response of analysis of variance for more than three groups. Drugs the arterial strips to 30 mmol/1 K+ was obtained, and used were AM (synthesized by Dr. Kangawa), CGRP, [8-37]CGRP, substance P, atrial natriuretic peptide the strips were washed repeatedly with fresh medium c and equilibrated. Concentration-response curves for (ANP), N -nitro-L-arginine (L-NA) (Peptide Institute, AM (10"10 to 10~8 mol/1) and CGRP (10"11 to 3 X Minoh, Japan), indomethacin, glibenclamide (Sigma, 10~9 mol/1) were obtained by applying the peptides St. Louis, MO), beraprost (Toray, Tokyo, Japan), cumulatively to the bathing medium. When compar- methylene blue trihydrate (Nacalai Tesque, Kyoto, Ja- ing the responses to two or more drugs in single con- pan), sodium nitroprusside (Merck, Darmstat, Ger- centrations in a same preparation, the order of the many), nitroglycerin (Nihonkayaku, Tokyo, Japan), drug application was randomized. Strips were con- PGF2a (Upjohn, Tokyo,Japan), and papaverine hydro- tracted partially with prostaglandin (PG) F2a (3 to 20 chloride (Dainippon, Osaka, Japan). X 10~7 mol/1); the contraction ranged between 25% and 38% of the contraction induced by 30 mmol/1 RESULTS K+. At the end of each series of experiments, papaverine (10~4 mol/1) was added to obtain the maximal Relaxation Induced by Adrenomedullin relaxation. Relaxations induced by test drugs were pre- In retinal arterial strips partially contracted with sented as relative values to the responses induced by 10 8 4 PGF2a, the addition of AM (10~ to 10~ mol/1) pro- 10" mol/1 papaverine. Preparations were treated for duced concentration-related relaxation. Mean values 20 minutes or longer with blocking agents before the of the apparent median effective concentration (EC50) response to agonists was obtained. When the blocking were (2.62 ± 0.33) X 10~9 mol/1, and the maximal agents produced relaxation, the arterial tone was relaxation obtained from 26 strips from separate dogs raised by adding supplemental doses of PGF2a. In was 85.1% ± 2.5%. Tachyphylaxis did not develop by some experiments, responses to AM were compared repeated trials (Fig. 1, left); die dose-response curves in a pair of arterial strips with and without the endo- at first to third trials are almost identical. The relaxant thelium obtained from the same dogs. The endothe- responses to AM did not differ in the strips widi and lium of each arterial strip was removed by gendy rub- without the endothelium (Fig. 1, right). Endothelial

Downloaded from iovs.arvojournals.org on 09/27/2021 58 Investigative Ophthalmology & Visual Science, January 1997, Vol. 38, No. 1

TABLE l. Effects of Antagonists on the Relaxation Caused by 10 9 mol/1 Adrenomedullin in Canine Central Retinal Arterial Strips Relaxation (%)*' by Adrenomedullin

Treatment n Control Experimental

L-NA 10-" mol/1 9 30.4 ± 5.9 30.4 ± 6.9 Indomethacin 10~6 mol/1 11 24.3 ± 5.1 25.6 ± 5.1 Methylene blue 10~s mol/1 6 26.2 ± 7.3 31.0 ± 7.6 Glibenclamide 10""fi mol/1 6 30.0 ± 10.7 34.1 ± 9.8

n = number of strips from separate dogs; L-NA = NG-nitro-L-arginine. * Relaxations relative to those caused by 10~4 mol/1 papaverine.

integrity was determined by 10 8 mol/1 substance P; susceptibility to beraprost treatment of relaxations in- mean values of the response in endothelium-intact duced by AM, CGRP, and nitroglycerin were com- and -denuded strips were 62.7% ± 7.8% and 4.8% ± pared in the same arterial strips. Typical recordings 1.7% (n = 11; P< 0.001, unpaired /-test), respectively. are shown in Figure 5. Treatment with 10~° mol/1 The AM-induced relaxation in the endothelium-intact beraprost abolished the responses to AM (3 X 10~9 strips was not influenced by L-NA (10~4 mol/1), indo- mol/1) and CGRP (10"9 mol/1) but did not alter the methacin (10~G mol/1), methylene blue (10~a mol/1), response to 10~8 mol/1 nitroglycerin. Mean values of and glibenclamide (10~6 mol/1) (Table 1). the response to nitroglycerin before and after treat- The concentration-dependent relaxation by AM ment with beraprost are summarized in Figure 6. was inhibited markedly by treatment with 10~7 mol/1 On the other hand, the nitroglycerin (10~8 mol/ [8-37]CGRP, a CGRP, receptor antagonist11 (Fig. 2). Effects of [8-37]CGRP were compared in the re- 0 r O- sponses to AM (3 X 10~9 mol/1), ANP (10~8 mol/1), and CGRP (10~9 mol/1). Typical tracings are illus- trated in Figure 3. The CGRP receptor antagonist markedly suppressed the responses to AM (n = 4) and CGRP but did not alter the relaxation induced by ANP. Similar results were obtained in three additional strips from separate dogs; mean values of the ANP- induced relaxation before and after [8-37]CGRP were 29.7% ± 7.7% and 29.0% ± 7.1% (n = 4), respec- 8 c tively. Beraprost (10~ M)-induced relaxations were o unaffected by the CGRP antagonist (42.3 ± 3.7 versus 50 35.0 ± 9.5, n= 5). s

Comparison of the Responses to Adrenomedullin, Calcitonin Gene-Related Pep tide, and Nitroglycerin • Control Treatment with 10~6 M beraprost, a stable analog of PGI2,12 abolished the response to AM in low concen- O C8-371CGRP trations (up to 3 X 10~9 mol/1) and markedly attenuated the relaxation at 10~8 mol/1 (Fig. 4, left). Retinal arterial strips contracted with PGF2« responded to CGRP (10"10 to 3 X 10~9 mol/1) with a dose-related 100 L 10-10 10-9 10-8 relaxation. Mean values of the EC50 were (0.71 ± 0.05) 9 9 X 10~ mol/1, and the maximal relaxation at 3 X 10~ Cone, of AM (M)

mol/1 in 10 strips from separate dogs was 84.3% ± 7 4.0%. Although the maximal relaxation was similar FIGURE 2. Modification by 10 mol/1 [8-37] calcitonin to the value widi AM (85.1%), the EC value was gene-related peptide (CGRP) of the response to adre- 50 nomedullin (AM) in canine central retinal arteries precon- significandy lower than that with AM (2.62 ± 0.33) X 4 9 tracted with prostaglandin F2a. Relaxations induced by 10~ 10" mol/1 (P< 0.01). The relaxant response to CGRP mol/1 papaverine were taken as 100%. Significantly different was inhibited markedly by treatment with the high from control, (a) P< 0.001. (b) P< 0.01 (unpaired Rest). concentration of beraprost (Fig. 4, right). Contrasting Five different dogs were used for the experiment.

Downloaded from iovs.arvojournals.org on 09/27/2021 Adrenomedullin in Retinal Artery 59

DOG RETINAL ARTERY

Control • •

o '•S 50

AM ANP CGRP PA [8-37] CGRP 0.2g

100 100 L 10-10 1Q-9 1o-io 10-a Cone, of AM (M) Conc. of CGRP (M) 6 FIGURE 4. Modification by 10 mol/1 beraprost of the re- After wash sponse to adrenomedullin (AM; left) and calcitonin gene- related peptide (CGRP; right) in canine central retinal arteries precontracted with prostaglandin F2o. Relaxations in- 4 10min duced by 10~ mol/1 papaverine were taken as 100%. Sig- 9 nificantly different from control, (a) P < 0.001. (b) P < FIGURE 3. Typical responses to 3 X 10~ mol/1 adrenomedul- 0.01. (c) P< 0.05 (unpaired West). lin (AM), 10~8 mol/1 atrial natriuretic peptide (ANP), 10~9 mol/1 calcitonin gene-related peptide (CGRP) of a canine central retinal artery strip in the presence (middle) and ab- enced by endothelium denudation nor by treatment sence (top and bottom) of 10~7 mol/1 [8-37] calcitonin gene- G 7 with N -nitro-L-arginine, a nitric oxide synthase inhib- related peptide. The strip was contracted with 4 X 10~ itor, or methylene blue, a soluble guanylate cyclase mol/1 prostaglandin (PG) F ; upward arrows indicate the 2a inhibitor. This discrepancy may be the result of differ- supplemental application of PGF2a to maintain the arterial tone. PA = 10~4 mol/1 papaverine. ences in the experimental conditions, animal species, or blood vessels used. Indomethacin or glibenclamide also failed to reduce the AM-induced relaxation, suggesting that neither prostanoids nor an ATP-sensidve l)-induced relaxation was abolished by treatment with + 3 X 10~(l mol/1 sodium nitroprusside (Fig. 6). Relax- K channel opening is involved in the response. ations caused by AM were reduced by sodium nitro- Similarities of the amino acid sequence and the prusside (3 X 10~6 mol/1), as were those by CGRP C-terminal amide structure in AM and CGRP, another 13 (Fig. 7). Atrial natriuretic peptide (10~8 mol/l)-in- potent endogenous vasodilator, have been indi- duced relaxations were attenuated to a greater extent than those induced by AM and CGRP by treatment DOG RETINAL ARTERY — Adrenomedullin with sodium nitroprusside (45.8% ± 7.9% versus 12.7% ± 4.8%; n = 5; P < 0.01; unpaired West).

DISCUSSION Control The current study demonstrated the potent relaxing action of AM in isolated canine retinal arteries. The magni- tude of the relaxation was much greater, and the EC50 value was at least 10 times less than those seen in the AM NTG other canine arteries, such as cerebral, mesenteric, renal, Beraprost coronary, and femoral arteries5 (unpublished data, 0.2g 1996). These results suggest that AM may have selective vasodilator actions on ocular circulation. Further, in con- 10min trast to the other arteries easily acquiring tachyphylaxis, 9 FIGURE 5. Typical responses to 3 X 10~ mol/1 adrenomedul- the response to AM of canine retinal arteries was repro- 8 9 ducible; therefore, retinal artery is suitable for analyzing lin (AM), 10" mol/1 nitroglycerin (CGRP), and 10" mol/ the mechanism of vasorelaxant action of AM. 1 calcitonin gene-related peptide (NTG) of a canine central retinal artery strip in the presence (bottom) and absence (top) It has been reported that AM-induced vasodilata- of 10~6 mol/1 beraprost. The strip was contracted with 5 X 9 7 tion is mediated by endogenous nitric oxide in rat 10~ mol/1 prostaglandin (PG) F2a; upward arrows indicate 3 and canine kidneys. In contrast, in the canine retinal the supplemental application of PGF2a to maintain the arte- artery, the relaxation induced by AM was not influ- rial tone. PA = 10~4 mol/1 papaverine.

Downloaded from iovs.arvojournals.org on 09/27/2021 60 Investigative Ophthalmology & Visual Science, January 1997, Vol. 38, No. 1

Nitroglycerin 10 8 M in the rat kidney4 and human vascular endothelial cells.17 Heterogeneity of the efficacy of the CGRP receptor antagonist may be die result of the difference in the relative ratios of CGRP and AM receptors in the tissue. A recent radioligand-binding study has demonstrated a possible AM-specific binding site in die rat heart, which shows 18 20 5000-fold higher affinity for AM dian CGRP. As for as the canine retinal artery is concerned, the vasodilator action of AM might be mediated primarily by CGRPi receptors, and the affinity of binding for AM might be c o similar to that for CGRP. However, the existence of AM 3 40 receptors in the retinal artery cannot be ruled out, unless cross-reactivity of [8-37]CGRP to AM receptors is ex- a) cluded in the artery. CE In a previous article,19 we demonstrated that treatment with sodium nitroprusside in concentrations suf- 60 ficient to cause an evident increase in the level of cyclic guanosine monophosphate (cGMP) depresses the responses to ANP, even though the cGMP level is C SNP elevated by this peptide. In addition, the same treat- n=9 ment markedly inhibits the responses to 8-bromo- (> cGMP and nitroglycerin. Based on these findings, FIGURE 6. Modification by 3 X 10~ mol/1 nitroprusside (SNP, left) and by 10"6 mol/1 beraprost (BP, right) of the involvement of the nucleotides in the signal transduc- response to 10~8 mol/1 nitroglycerin in canine central reti- tion associated with AM actions was analyzed. Treat- nal arteries precontracted with prostaglandin F2a. Relax- ment of canine retinal arteries with a high concentra- ations induced by 10~'' mol/1 papaverine were taken as tion of sodium nitroprusside significantly attenuated 100%. Significantly different from control, (a) P < 0.001 the response to AM and CGRP and almost abolished (unpaired /!-test). the nitroglycerin-induced relaxation, suggesting that the cGMP-dependent mechanism was involved par- cated.1'1 In the current study dealing with canine reti- tially in the AM- and CGRP-induced relaxations. Eleva- nal arteries, vasodilator potencies of these peptides tion of cGMP by AM was unlikely to be elicited by the were similar. However, the potency of AM is reportedly activation of soluble guanylate cyclase because methyl- less than that of CGRP in canine basilar, mesenteric, ene blue did not reduce the response to AM. Whether renal, coronary, or femoral arteries.5 Adrenomedullin elevation was caused by the activation of participate is more potent in vasodilating the cat pulmonary vas- guanylate cyclase, or by die inhibition of cGMP-depen- cular bed than CGRP,15 and the opposite is true in dent phosphodiesterase was not determined in this the rat mesenteric vascular bed.6 The diversity of the response to AM may be associated with the affinity and the population of receptors responsible for the vascular action. To determine the receptor involved, the susceptibility to [8-37]CGRP, a CGRP, receptor antagonist," of AM and CGRP actions were compared. Retinal arterial relaxations induced by AM were inhibited markedly by 5 so treatment with the antagonist, as were the responses to CGRP, whereas those induced by ANP were not influenced. Adrenomedullin appears to act as an agonist on CGRPi receptors in the canine retinal artery. Similar findings have been reported in the microvasculature of 100 100 L 1o-io 10-9 10-a 1O-1O 10-9 10-e 14 the hamster cheek pouch, cultured rat vascular smooth Cone, of AM (M) Cone, of CGRP (M) muscle cells,7 and hepatic stellate cells.16 However, the FIGURE 7. Modification by 3 X 10 6 mol/1 nitroprusside SNP inhibitory potency of [8-37]CGRP is different, de- of the response to adrenomedullin (AM, left) and calcitonin pending on the organ used—equipotent to both pep- gene-related peptide (CGRP, right) in canine central retinal tides in the hamster cheek pouch, more potent against arteries precontracted with prostaglandin F2a. Relaxations CGRP than AM in the cultured cells, and more potent induced by 10~4 mol/1 papaverine were taken as 100%. Sig- against AM in the hepatic cells. In contrast, the receptor nificantly different from control, (a) P< 0.01; (b) P< 0.05 antagonist does not inhibit the effects of AM effectively (unpaired Hest).

Downloaded from iovs.arvojournals.org on 09/27/2021 Adrenomedullin in Retinal Artery 61

study. On the other hand, treatment with a high con- 6. Nuki C, Kawasaki H, Kitamura K, et al. Vasodilator centration of beraprost, a stable analog of PGI2 known effect of adrenomedullin and calcitonin gene-related to relax the artery by elevating intracellular cyclic peptide receptors in rat mesenteric vascular beds. Bio- adenosine monophosphate,12 markedly reduced or chem Biophys Res Commun. 1993; 196:245-251. abolished the relaxations induced by both AM and 7. Eguchi S, Hirata Y, Kano H, et al. Specific receptors for adrenomedullin in cultured rat vascular smooth CGRP but did not affect the response to nitroglycerin. muscle cells. FEBS Lett. 1994;340:226-30. Therefore, the cyclic adenosine monophosphate-de- 8. Ishizaka Y, Ishizaka Y, Tanaka M, et al. Adrenomedul- pendent mechanism seems to be involved in the re- 20 lin stimulates cyclic AMP formation in rat vascular sponses to AM as well as to CGRP. These results smooth muscle cells. Biochem Biophys Res Commun. suggest that signal transduction mechanisms responsi- 1994; 200:642-646. ble for the relaxation caused by AM are similar to 9. Hirata Y, Hayakawa H, Suzuki Y, et al. Mechanisms of those by CGRP in canine retinal arteries. adrenomedullin-induced vasodilation in the rat kid- In conclusion, AM is likely to elicit endothelium- ney. Hypertension. 1995;25(part 2):790-795. independent relaxation mediated primarily by intra- 10. Kitamura Y, Okamura T, Kani K, Toda N. Nitric oxide- cellular cyclic adenosine monophosphate by the stim- mediated retinal arteriolar and arterial dilatation induced by substance P. Invest Ophthalmol Vis Sci. ulation of CGRP i receptors and partially by cGMP in 1993; 34:2859-2865. canine retinal arteries. Further experiments are re- 11. Dennis T, Fournier A, Cadieux A, et al. hCGRPg_s7, quired to clarify the reason for noticeable heterogene- calcitonin gene-related peptide antagonist revealing ity of the vascular actions of AM. The fact that the calcitonin gene-related peptide receptor heterogene- vasodilator potency of AM in the central retinal artery ity in brain and periphery. / Pharmacol Exp Tlier. was the highest among the canine arteries tested (ref. 1990;254:123-128. 5 and Okamura et al, unpublished data, 1996) sug- 12. Toda N. Beraprost sodium. Cardiovasc Drug Rexj. gests possible roles of circulating AM and the peptide 1988;6:222-238. liberated locally from endothelial cells21 in the regula- 13. Brain SD, Williams TJ, TippinsJR, Morris HR, Maclnt- tion of ocular blood flow. Further, AM may have some yre I. Calcitonin gene-related peptide is a potent vaso- potential as a clinical ocular vasodilator in the future. dilator. Nature. 1985; 313:54-56. Although in the canine and feline brain, circulating 14. Hall JM, Siney L, Lippton H. Interaction of human adrenomedullin 13—52 with calcitonin gene-related angiotensin II increases cerebral vascular resis- 22 23 peptide receptors in the microvasculature of the rat tance, ' whether and how much the blood retinal and hamster. Br J Pharmacol. 1995; 114:592-597. barrier limits the passage of circulating peptides 15. DeWitt BJ, Cheng DY, Caminiti GN, et al. Comparison through tight junctions of endothelial cells is yet to of responses to adrenomedullin and calcitonin gene- be determined. related peptide in the pLilmonary vascular bed of the cat. Eur J Pharmacol. 1994; 257:303-306. Key Words 16. Kawada N, Inoue M. Effect of adrenomedullin on he- adrenomedullin, calcitonin gene-related peptide, central patic pericytes (stellate cells) of the rat. FEBS Lett. retinal artery, cyclic adenosine monophosphate, cyclic gua- 1994;356:109-113. nosine monophosphate 17. KatoJ, Kitamura K, Kangawa K, Eto T. Receptors for adrenomedullin in human vascular endothelial cells. References Eur J Pharmacol. 1995;289:383-385. 18. Owji AA, Smith DM, Coppock HA, et al. An abundant 1. Kitamura K, Kangawa K, Kawamoto M, et al. Adre- and specific binding site for the novel vasodilator ad- nomedullin: A novel hypotensive peptide isolated renomedullin in the rat. Endocrinology. 1995; from human pheochromocytoma. Biochem Biophys Res 136:2127-2134. Commun. 1993; 192:553-560. 19. Matsumoto T, Okamura T, Kinoshita M, Toda N. In- 2. Ichiki Y, Kitamura K, Kangawa K, et al. Distribution teractions of niuovasodilators and atrial natriuretic and characterization of immunoreactive adrenomed- peptide in isolated dog coronary arteries. Eur J Phar- ullin in human tissue and plasma. FEBS Lett. 1994; macol. 1993; 237:31-37. 338:6-10. 20. Poyner DR. Calcitonin gene-related peptide: Multiple 3. Miura K, Ebara T, Okumura M, et al. Attenuation of G actions, multiple receptors. Pharmacol Tlier. 1992; adrenomedullin-induced renal vasodilatation by N - 56:23-51. nitro-L-arginine but not glibenclamide. BrJ Pharmacol. 21. Sugo S, Minamino N, Kangawa K, et al. Endothelial 1995; 115:917-924. cells actively synthesize and secrete adrenomedullin. 4. Elhawary AM, Poon J, Pang CC. Effects of calcitonin Biochem Biophys Res Commun. 1994;201:1160-1166. gene-related peptide receptor antagonists on renal ac- 22. Lowe RF, Gilboe DD. Canine cerebrovascular response tions of adrenomedullin. Br J Pharmacol. 1995; to nitroglycerin, acetylcholine, 5-hydroxytryptamine, 115:1133-1140. and angiotensin. AmJPhysiol. 1973;225:1333-1338. 5. Nakanuira K, Toda H, Terasako K, et al. Vasodilative 23. Faraci FM, Mayhan WG, Schmid PG, Heistad DD. Ef- effect of adrenomedullin in isolated arteries of the fects of arginine vasopressin on cerebral microvascu- dog. Jpn J Pharmacol. 1995;67:259-262. lar pressure. AmJPhysiol. 1988;255:H70-H76.

Downloaded from iovs.arvojournals.org on 09/27/2021