Influence of Adrenomedullin on Tone of Isolated Bovine Retinal Arteries

Abstract

purpose. To assess and characterize the vasorelaxing effect of adrenomedullin (AM) on isolated bovine retinal arteries (BRAs).

methods. Retinal arteries were isolated from bovine eyes and mounted in a wire myograph for isometric tension recording. Concentration–response curves were generated by cumulative addition of AM (1 pM to 0.1 μM) to the organ bath.

results. AM caused a concentration-dependent relaxation of the BRAs. Removal of the endothelium of the BRAs, inhibition of nitric oxide synthase with \(\mathit{N}^{\overline{{\omega}}}\) -nitro-l-arginine (l-NA) or inhibition of soluble guanylyl cyclase with 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ) significantly reduced the AM response. Cyclooxygenase inhibition with indomethacin or sodium diclofenac did not reduce, but rather increased, vasodilation. The AM-receptor antagonist AM 22-52 slightly, but significantly, reduced the AM response, whereas the CGRP-receptor antagonist CGRP 8-37 caused a more pronounced reduction. The adenosine receptor antagonist 8-(p-sulfophenyl) theophylline (8-SPT) did not affect AM-induced vasorelaxation. Inhibition of several intracellular calcium ([Ca²⁺]_i)-reducing mechanisms failed to block the relaxation induced by AM. Only inhibition of the plasma membrane Ca²⁺-adenosine triphosphatase (ATPase) with vanadate significantly attenuated the AM response.

conclusions. AM induces vasodilation in isolated bovine retinal arteries. Endothelium-derived NO and stimulation of CGRP- and AM-receptors appear to be involved in the AM response, whereas prostanoids and activation of adenosine receptors are not involved. Activation of Ca²⁺-extrusion by the plasma membrane Ca²⁺-ATPase may elicit the relaxation of BRAs in response to AM.

Adrenomedullin (AM) is a 52-amino-acid peptide originally identified in human pheochromocytoma. ¹ AM is produced in a broad variety of tissues and cells, including retinal pigment epithelium (RPE), neurons, astrocytes, vascular endothelium, and vascular smooth muscle cells. ^{2 3 4 5 6 7 8} A remarkably broad range of actions has been described for AM (for a review, see Hinson et al. ⁸ ). In several studies, investigators have reported a potential role for AM in the normal physiology of the eye, and in the pathophysiology of some ocular diseases. For example, a relaxing effect on the iris sphincter muscle, ⁹ a role in controlling intraocular pressure, ¹⁰ a protective role in ischemic conditions in retinal pigment epithelium, ³ a role in the pathophysiology of inflammatory eye diseases, ^{2 11} and a role in the pathophysiology of intraocular and orbital tumors ¹² have been described or suggested in the recent literature (for a review of the role of AM in the eye, see Udono-Fujimori et al. ¹³ ). 

AM induces relaxation in different vascular beds. ^{8 14 15 16 17 18} However, the influence of AM on intraocular retinal arteries is as yet unknown. In the present study we therefore investigated the direct effects of AM on isolated bovine retinal arteries (BRAs). A marked variation has been reported in the mechanisms involved in the AM-induced vasodilation in different vascular beds (for a review, see Hinson et al. ⁸ ). Therefore, we sought to characterize pharmacologically the mechanisms involved in the AM response in the BRAs. Finally, we tried to evaluate the hypothesis that AM is the as yet unidentified retinal relaxing factor (RRF) recently described to be released from the retina of different animal species. ^{19 20 21 22}  

Materials and Methods

Tension Measurements

Bovine eyes, obtained from the local abattoir, were enucleated within half an hour after the animals were killed and were transported to the laboratory in ice-cold Krebs-Ringer bicarbonate (KRB) solution. The anterior segment and the vitreous were removed, and the eyecup was placed in cold and oxygenated (5% CO₂ in O₂) KRB solution for further preparation. A segment located between the optic disc and the first branch of the most prominent retinal artery was excised with surrounding retinal tissue. The arterial segments were mounted in an automated dual small vessel myograph (model 500 A; J. P. Trading, Aarhus, Denmark) with a tissue chamber filled with 10 mL of KRB solution. Two stainless steel wires (40 μm diameter) were guided through the lumen of the segments (±2-mm length). One wire was fixed on a holder connected to a force-displacement transducer, and the other was fixed on a holder connected to a micrometer. In most experiments, the adhering retinal tissue was completely removed after the first wire was fixed. In the experiments on arteries with adhering retinal tissue, only part of the retinal tissue was trimmed away, so that a small strip of retinal tissue remained attached to the artery. After mounting, the preparations were allowed to equilibrate for approximately 30 minutes in the KRB solution at 37°C, bubbled with 95% O₂ and 5% CO₂ (pH 7.4). Subsequently, the optimal lumen diameter of the vessels was calculated on the basis of the passive wall tension–internal circumferences relationship. ^{23 24} Immediately after the vessels were stretched to their optimal lumen diameter (219.73 ± 2.37 μm, n = 118), they were activated twice with a KRB solution containing 120 mM K⁺ and once with a KRB solution containing 120 mM K⁺ and 30 μM PGF_2α to assess maximal contractility. 

After the mounting and preparation procedures were concluded, the retinal arteries were contracted by adding 30 μM PGF_2α to the organ bath. When a stable contraction was reached, increasing concentrations of adrenomedullin were added to the organ bath to generate a concentration–response curve. In the experiments on the influence of potassium ions on the AM response, bovine retinal arteries were also contracted either by replacing the standard KRB solution in the organ bath by a KRB solution containing 30 mM K⁺ and 30 μM PGF_2α or by a KRB solution containing 120 mM K⁺ and 30 μM PGF_2α. 

Removal of the Endothelium

The arteries were first unstretched in the myograph. Subsequently, an L-shaped micropipette was positioned at the proximal end of the vessel, and 95% O₂/5% CO₂ was bubbled through the lumen for 1 minute. Subsequently, the artery was stretched by resetting the wires to their original positions, and the vessel was allowed to re-equilibrate for half an hour. 

Drugs

The experiments were performed in a KRB solution of the following composition (mM): NaCl 135, KCl 5, NaHCO₃ 20, glucose 10, CaCl₂ 2.5, MgSO₄ 1.3, KH₂PO₄ 1.2, and EDTA 0.026 in H₂O. KRB solutions containing 30 mM K⁺ (K₃₀) and 120 mM K⁺ (K₁₂₀) were prepared by equimolar replacement of NaCl by KCl. Adrenomedullin (AM), \(\mathit{N}^{\overline{{\omega}}}\) -nitro-l-arginine (l-NA), 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ), indomethacin, sodium diclofenac, acetylcholine chloride, adrenomedullin fragment 22-52 (AM 22- 52), 8-(p-sulfophenyl)theophylline (8-SPT), adenosine, thapsigargin, cyclopiazonic acid (CPA), amiloride hydrochloride, nifedipine, and sodium metavanadate were obtained from Sigma-Aldrich (St. Louis, MO), prostaglandin F_2α (PGF_2α, Dinolytic) from Upjohn (Puurs, Belgium), sodium nitroprusside (SNP) from Merck (Darmstadt, Germany), calcitonin gene-related peptide 8-37 (CGRP 8-37, rat) from Tocris (Bristol, UK), and 1,3-dimethyl-2-thiourea (DMTU) from Sigma-Aldrich (Steinheim, Germany). Stock solutions were made in water, except for ODQ, CPA, amiloride, and nifedipine (dissolved in dimethylsulfoxide); indomethacin and thapsigargin (dissolved in ethanol); and acetylcholine chloride (dissolved in phthalate buffer, pH 4.0). The final concentration of both ethanol and dimethylsulfoxide in the organ bath never surpassed 0.1%. 

Statistical Methods

The data were computed as the mean ± SEM and evaluated statistically with Student’s t-test for paired samples or repeated measures ANOVA with Bonferroni’s post hoc test when appropriate. Two groups of data were considered to be significantly different when P < 0.05. Relaxations are expressed as a percentage decrease in tone (n = number of preparations tested). Estimated EC₅₀ and estimated maximal response are calculated using nonlinear regression curve fit. 

Results

Effect of Adrenomedullin on PGF_2α-Induced Contractions

After PGF_2α induced a stable contraction (mean: 5.80 ± 0.29 mN, n = 21), cumulative addition of AM (1 pM to 0.1 μM) caused a concentration-dependent relaxation of the BRAs (Fig. 1) . The estimated pEC₅₀ is 7.90 ± 0.03, and the estimated maximum relaxation is 59.20% ± 1.00% (n = 21). In a next series of experiments, three consecutive concentration–response curves were constructed to assess the reproducibility of the AM response. No difference in AM response was observed in these repeated trials (results not shown). 

Influence of Removing the Endothelium

The involvement of the endothelium in the AM response was evaluated by comparing the AM-induced relaxations before and after endothelium denudation. The relaxing effect of acetylcholine was used to evaluate the effectiveness of endothelium removal. After endothelium denudation, the acetylcholine-induced relaxation was significantly decreased (0.1 mM: 24.86% ± 4.70% before, and 5.14% ± 1.30% after removal of the endothelium, n = 7, P < 0.05). PGF_2α-induced contraction was slightly reduced (from 5.30 ± 0.73 mN to 4.84 ± 0.56 mN, n = 7, P < 0.05), and the relaxations induced by 10 and 30 nM AM were reduced from 13.86% ± 1.47% and 38.14% ± 4.90% to 6.71% ± 1.38% (n = 7, P < 0.05) and 25.71% ± 4.32% (n = 7, P < 0.05), respectively (Fig. 2) . In this series of experiments, two of nine BRA segments were rejected because acetylcholine induced only a minor relaxation (<10% relaxation at 0.1 mM) in control conditions. It was assumed that the endothelium was not completely intact in these vessels. 

Influence of l-NA and ODQ

The NO-synthase inhibitor l-NA was used to study the involvement of nitric oxide in the AM-induced relaxation of the BRAs (Fig. 3A) . PGF_2α-induced contractions in the BRAs were slightly, but significantly, increased after treatment with 0.1 mM l-NA for 10 minutes (6.89 ± 0.74 mN before vs. 7.23 ± 0.81 mN after treatment, n = 6, P < 0.05). The mean relaxations induced by 10 and 30 nM AM under control conditions were 8.50% ± 1.43% and 25.50% ± 1.61%, respectively (n = 6). Treatment with l-NA significantly reduced the AM response to 0.67% ± 0.49% (n = 6, P < 0.05) and 7.17% ± 0.60% (n = 6, P < 0.05), respectively. The relaxation elicited by acetylcholine was also significantly diminished after treatment with l-NA (0.1 mM: 27.33% ± 5.23% relaxation in the absence vs. 3.67% ± 0.67% in the presence of l-NA, n = 6, P < 0.05). 

To evaluate further the role of NO in the AM response, the influence of the potent and selective soluble guanylyl cyclase inhibitor ODQ ^{25 26} was investigated (Fig. 3B) . Treatment of the BRAs with 1 μM ODQ for 20 minutes induced a small, but statistically significant, increase in PGF_2α-induced tone (4.79 ± 0.44 mN in the absence and 5.21 ± 0.49 mN in the presence of ODQ, n = 8, P < 0.05). AM-induced relaxations at 10 and 30 nM were significantly reduced by ODQ from 17.62% ± 3.96% and 36.70% ± 4.55% to 8.73% ± 2.04% (n = 8, P < 0.05) and 22.53% ± 1.63% (n = 8, P < 0.05), respectively. The relaxations induced by sodium nitroprusside 0.1 mM were also significantly reduced by ODQ (27.83% ± 3.34% in the absence of ODQ, 1.57% ± 0.45% in the presence of ODQ, n = 8, P < 0.05). 

Influence of Indomethacin and Sodium Diclofenac

In these experiments, the influence of the cyclooxygenase (COX) inhibitors indomethacin and sodium diclofenac on the AM response was studied to evaluate the involvement of prostanoids in the relaxations induced by AM. Treatment of the BRAs for 20 minutes with 10 μM indomethacin caused a small, but significant decrease in PGF_2α-induced contraction from 5.26 ± 0.58 mN to 4.61 ± 0.65 mN (n = 8, P < 0.05). The 10-nM AM-induced relaxation showed a substantial, but not significant, increase from 13.73% ± 3.89% to 36.54% ± 11.26% (n = 8, P > 0.05). The increase in 30-nM AM-induced relaxation, in contrast (from 41.89% ± 9.33% to 69.66% ± 9.73%), was statistically significant (n = 8, P < 0.05). Also the water-soluble COX-inhibitor sodium diclofenac caused a small, but statistically significant, decrease in PGF_2α-induced contraction (4.97 ± 0.57 mN in the absence and 4.48 ± 0.55 mN in the presence of sodium diclofenac, n = 6, P < 0.05) and a marked increase in both the 10-nM AM-induced relaxations (35.72% ± 7.13% in the absence and 50.07% ± 11.04% in the presence of sodium diclofenac, n = 6, P < 0.05) and the 30-nM AM-induced relaxations (55.26% ± 10.53% in the absence and 87.92% ± 9.83% in the presence of sodium diclofenac, n = 6, P < 0.05). 

Influence of Increasing Concentrations of Potassium Ions

In these experiments, the influence of increasing concentrations of K⁺-ions on the AM response in BRAs was investigated. The arteries were contracted successively by adding 30 μM PGF_2α to the standard KRB solution in the organ bath (containing 5 mM K⁺), to a KRB solution containing 30 mM K⁺, and to a KRB solution containing 120 mM K⁺. During the steady state of each one of these contractions, a concentration–response curve for AM was constructed. The results of these experiments are presented in Table 1 . The contraction induced by 30 mM K⁺ and PGF_2α was not significantly different from the tone induced by PGF_2α in standard KRB (n = 8, P > 0.05). The contraction induced by 120 mM K⁺ and PGF_2α was significantly larger than the contraction induced by PGF_2α in standard KRB (n = 8, P < 0.05). Both the 10- and 30-nM AM-induced relaxations were slightly, but not significantly reduced in the presence of 30 mM K⁺ (n = 8, P > 0.05). However, in the presence of 120 mM K⁺, both the 10- and the 30-nM AM-induced relaxations were significantly reduced (n = 8, P < 0.05). 

Influence of AM and CGRP Receptor Antagonists

Incubation of the BRAs for 5 minutes with the proposed AM receptor antagonist AM 22-52 (1 μM) caused a statistically significant but small reduction in both the 10-nM AM-induced relaxation (27.67% ± 4.54% in the absence and 22.00% ± 4.31% in the presence of AM 22-52, n = 6, P < 0.05) and the 30-nM AM-induced relaxation (55.33% ± 9.65% in the absence and 47.50% ± 8.26% in the presence of AM 22-52, n = 6, P < 0.05; Fig. 4A ). 

Incubation of the BRAs for 5 minutes with the proposed CGRP₁ receptor antagonist CGRP 8-37 (1 μM) significantly affected both the 10-nM AM-induced relaxation (31.39% ± 5.56% in the absence and 4.79% ± 1.93% in the presence of CGRP 8-37, n = 6, P < 0.05), and the 30-nM AM-induced relaxation (55.43% ± 10.53% relaxation in the absence and 17.77% ± 6.57% in the presence of CGRP 8-37, n = 6, P < 0.05; Fig. 4B ). 

In the next series of experiments, the influence of CGRP 8-37 on the NO-independent part of the AM-induced relaxation was studied. Therefore, three consecutive concentration–response curves for AM were constructed: the first in control conditions, the second after treatment of the BRAs with both the NO-synthase inhibitor l-NA (0.1 mM) and the soluble guanylyl cyclase inhibitor ODQ (1 μM), and the third in the presence of l-NA, ODQ, and CGRP 8-37. The results of these experiments are presented in Figure 5 . As expected, the AM response is significantly reduced in the presence of l-NA and ODQ (P < 0.05, n = 4). In the third concentration–response curve, the additional presence of CGRP 8-37 caused no significant decrease in AM-induced relaxations (P > 0.05, n = 4). 

Influence of CGRP 8-37 on the Relaxing Effect of the RRF

To exclude AM from being the recently described RRF, ^{19 20} the influence of CGRP 8-37 on the relaxations induced by the RRF in BRAs was studied. Because the RRF has not yet been isolated or identified, no standard concentration–response curve can be generated. However, the RRF is released from retinal tissue and its effect on BRAs can therefore be studied with a previously described ¹⁹ technique in which the perivascular retinal tissue is incompletely removed from the BRA segments. Concentration–response curves for PGF_2α (0.1–30 μM) were generated in BRAs with a small strip of adhering retinal tissue before and after incubation with CGRP 8-37 (1 μM, 5 minutes). As no difference was noted in the PGF_2α-induced contraction (n = 4, P > 0.05) in the presence of CGRP 8-37, it was concluded that the effect of the RRF released from the adhering retinal tissue was not inhibited by CGRP 8-37 (Fig. 6) . 

Influence of 8-SPT

Treatment of the BRAs for 10 minutes with the adenosine receptor antagonist 8-SPT (0.1 mM) did not affect the 10-nM AM-induced relaxations (33.85% ± 3.32% in the absence and 31.09% ± 4.10% in the presence of 8-SPT, n = 8, P > 0.05), nor did it affect the 30-nM AM-induced relaxations (48.12% ± 4.75% in the absence and 49.34% ± 4.13% in the presence of 8-SPT, n = 8, P > 0.05). The vasodilation induced by 10 μM of adenosine, however, was significantly reduced by 8-SPT (38.10% ± 2.85% in the absence and 5.10% ± 1.06% relaxation in the presence of 8-SPT, n = 6, P < 0.05). 

Involvement of [Ca]_i-Reducing Mechanisms

The influence of several inhibitors of [Ca²⁺]_i-reducing mechanisms on the relaxation induced by AM was investigated. The results of these experiments are summarized in Table 2 . Treatment of the BRAs with the sarcoplasmic reticulum Ca²⁺-ATPase (SERCA) inhibitors thapsigargin (2 μM; 45 minutes of incubation) or CPA (20 μM; 20 minutes of incubation) significantly affected PGF_2α-induced tone in the BRAs. The mean AM-induced relaxations were not decreased, but rather were increased in the presence of a SERCA inhibitor. Two inhibitors of the Na⁺/Ca²⁺ exchanger were used: 1,3-dimethyl-2-thiourea (DMTU) and amiloride. After treatment with DMTU 25 mM for 30 minutes, PGF_2α was unable to induce a significant contraction in the BRAs, thereby making the generation of a concentration–response curve for AM impossible. Amiloride (0.1 mM; 20 minutes of incubation) caused a smaller, but also significant reduction of the PGF_2α-induced tone. The AM response was not blocked after treatment of the BRAs with amiloride. On the contrary, the relaxation percentages were somewhat increased. To evaluate the potential involvement of closure of l-type Ca²⁺ channels, the AM response was examined in the presence of the L-type Ca²⁺ channel blocker nifedipine (0.1 μM). Nifedipine also caused a decrease in PGF_2α-induced contraction and an apparent increase in AM response (Table 2) . In the experiments with thapsigargin, CPA, amiloride, and nifedipine, the mean AM-induced relaxations before and after treatment were not statistically compared because of the significant difference in PGF_2α-induced contraction. The plasma membrane Ca²⁺-ATPase inhibitor vanadate (1 mM) caused a stable contraction of 4.26 ± 0.42 mN (n = 8) in the BRAs. After a 90-minute incubation, the vanadate-induced tone was supplemented with 30 μM PGF_2α to match the tone induced by 30 μM PGF_2α in the absence of vanadate. Total contraction was not significantly different from contraction in control conditions. The AM-induced relaxations were significantly reduced in the presence of vanadate. 

Discussion

This study clearly demonstrates that AM relaxes isolated BRAs. In experiments in which two or three consecutive concentration–response curves were constructed, the response of BRAs to AM showed to be reproducible. This indicates that tachyphylaxis does not develop in isolated BRAs, although it has been reported to occur in the brain in vivo. ²⁷ Consequently, this protocol was used for further characterization of the mechanisms underlying the AM-induced relaxations in the BRA. 

Many stimuli are known to relax isolated blood vessels in an endothelium-dependent manner. ^{28 29 30} Also the vasodilatory actions of AM have shown to be endothelium-dependent in different blood vessels. ^{14 17} In our experiments, removal of the endothelium of the BRA resulted in a significant decrease in the AM-induced relaxations. Dysfunction of the endothelium was illustrated by the fact that the acetylcholine-induced relaxation, which is known to be endothelium-dependent, ³¹ was significantly reduced. However, the vasorelaxation in response to AM was not completely abolished after removal of the endothelium, and a substantial fraction of the vasorelaxing influence of AM remained unaffected. These experiments suggest that the AM response in the BRAs is only in part endothelium dependent. 

The vascular endothelium can mediate the regulation of vascular tone through the release of several smooth muscle relaxants such as nitric oxide (NO), prostaglandins and the endothelium-derived hyperpolarizing factor (EDHF). ²⁹ We found that the relaxations induced by AM were significantly attenuated by both the NO-synthase inhibitor l-NA and the soluble guanylyl cyclase inhibitor ODQ. The effectiveness of these inhibitors in our experiments was illustrated by the fact that respectively the acetylcholine-induced relaxation and the SNP-induced relaxation were significantly reduced in their presence. A small increase in contractile tone was noticed after treatment of the BRAs with either l-NA or ODQ. An inhibition of the basal release of NO from the endothelium probably accounts for this increase. After treatment of the BRAs with the COX inhibitors indomethacin or sodium diclofenac, the AM response was not decreased, but in contrast even increased. Previous research revealed that also the adenosine and sodium nitroprusside-induced relaxations in the BRA were increased in the presence of indomethacin (Boussery K, et al., unpublished results, 2002), suggesting that this effect is not specific for AM. We can conclude, however, that the AM response is certainly not diminished after inhibition of the production of prostaglandins in the BRAs, thereby excluding vasodilator prostanoids as mediators of the AM response. A major role for the EDHF is excluded by the fact that the AM response was only slightly, but not significantly reduced in the presence of 30 mM of K⁺, known to abolish EDHF-responses. ³² Taking all evidence together, we conclude that the endothelium-dependent part of the AM-induced relaxations in the BRAs is probably mediated by the release of NO from the endothelium. 

Although there is some controversy concerning the receptors responsible for AM-induced vasorelaxation, it has been proposed that activation of CGRP₁ receptors ³³ and specific AM receptors ^{34 35} may be involved. In our experiments, the relaxations induced by AM were markedly antagonized by the proposed CGRP₁ receptor antagonist CGRP 8-37. The proposed AM receptor antagonist AM 22-52 caused a significant, but rather small, reduction in AM-mediated vasodilation. In the interpretation of these results, it should be taken into account that some doubts have been expressed concerning the potency and selectivity of these receptor antagonists. ^{8 36 37} Nevertheless, our results suggest the involvement of CGRP₁ receptors and specific AM receptors in the AM-induced vasorelaxation in the BRAs. From our experiments, one could conclude that the specific AM receptor only seems to play a minor role, although the observations that provoke this assumption could also be accounted for by a combination of a poor antagonist potency of AM 22-52 and a low selectivity of CGRP 8-37. Development of more potent and selective blockers is necessary to clarify the relative importance of the CGRP and AM receptors in the AM induced vasorelaxation. 

As described earlier, the AM-induced relaxation in the BRAs is in part NO-dependent. As a significant part of the AM response remains unaffected by inhibitors of the NO pathway, we also tried to evaluate the influence of CGRP 8-37 on the NO-independent part of the AM response. As expected, the AM response was significantly reduced after inhibition of the NO pathway by a combination of l-NA and ODQ. After inhibition of the NO pathway, additional treatment with CGRP 8-37 caused only a negligible reduction in the AM-induced relaxation. These experiments suggest that CGRP 8-37–sensitive receptors are involved in the NO-dependent part of the AM-induced relaxations, rather than in the NO-independent part. However, clarification of the exact role of CGRP- and AM-receptors in the AM response, and their localization on the endothelium or vascular smooth muscle cells, will only be possible when more potent and selective antagonists for these receptors become available. 

Sabates et al. ³⁸ suggested that the coronary vasodilator effect of AM in dogs is initiated by activation of adenosine receptors. This led us to examine whether adenosine receptors could also be involved in the AM response in BRAs. The adenosine receptor antagonist 8-SPT significantly blocked the adenosine-induced relaxation in the BRAs, but had no effect on the AM response. On the basis of these results, the involvement of adenosine-receptors in the AM-induced vasodilation in the BRAs can be excluded. 

In the next series of experiments, we evaluated the potential involvement of several [Ca²⁺]_i-reducing mechanisms in the AM response. Vasorelaxation may involve a decrease in [Ca²⁺]_i in the smooth muscle cells due to the activation of SERCA, the plasma membrane Ca²⁺-ATPase, and the Na⁺-Ca²⁺ exchanger. In addition, [Ca²⁺]_i reduction may be due to the closure of L-type Ca²⁺ channels. ^{39 40 41 42} Inhibition of SERCA with thapsigargin or CPA resulted in a significant reduction of PGF_2α-induced contraction. The AM response, however, seemed to be increased in the presence of the SERCA inhibitors, perhaps because the AM response is expressed as a relaxation percentage. The major decrease in preexisting tone in the second contraction causes an increase in the relaxation percentages, without increasing the absolute relaxing influence of AM. Therefore, the AM relaxations in these experiments were not statistically analyzed. Treatment of the BRAs with the Na⁺/Ca²⁺ exchanger inhibitors DMTU or amiloride also resulted in a significant decrease in PGF_2α-induced contraction. Only after treatment with amiloride, an acceptable contraction could be induced to generate a concentration–response curve. In analogy with the results for the SERCA inhibitors, the AM response rather seemed to be augmented in the presence of amiloride. Also in the presence of the L-type Ca²⁺ channel blocker nifedipine, the contraction was reduced and the AM response somewhat increased. Again, this apparent increase can be attributed to the significant decrease in preexisting tone. It is clear, however, that the AM response is not blocked in the presence of thapsigargin, CPA, amiloride, or nifedipine, suggesting that it is unlikely that activation of SERCA, activation of the Na⁺/Ca²⁺ exchanger, or closure of L-type Ca²⁺ channels is involved in the vasodilator response to AM in BRAs. 

Inhibition of the plasma membrane Ca²⁺-ATPase with vanadate ⁴³ caused a significant contraction in the BRAs. This can probably be attributed to an inhibition of the basal activity of plasma membrane Ca²⁺-ATPase. After addition of PGF_2α, total contraction was not affected. In contrast, the AM response was significantly reduced in the presence of vanadate. These results suggest that Ca²⁺-extrusion by the plasma membrane Ca²⁺-ATPase may play a role in the AM-induced relaxation in the BRAs. However, vanadate may also influence other mechanisms ^{44 45} and no selective plasma membrane Ca²⁺-ATPase blocker is available. Therefore, these experiments suggest rather than conclusively show the involvement of plasma membrane Ca²⁺-ATPase. 

Okamura et al. ¹⁸ described that AM has a potent, but endothelium- and NO-independent relaxing effect in isolated canine central retinal arteries and that AM may exert selective vasodilator influences on ocular circulation. As yet, no data are available on a possible role for AM in regulating retinal blood flow by influencing the smooth muscle tone in the intraocular retinal arteries. The intraocular retinal arteries are quite different from the central retinal artery. As the central retinal artery passes through the lamina cribrosa, the vessel wall is reduced, the internal elastic lamella is lost, and the medial muscle coat becomes incomplete. The retinal arteries within the eye are therefore arterioles rather than arteries. ⁴⁶ The findings described in the present study demonstrate that intraocular retinal arteries are also prone to the vasorelaxing capacities of AM. Therefore, the present study supports the suggestion made by Okamura et al. ¹⁸ that AM may have a role in the regulation of ocular blood flow. Udono et al. ³ described that hypoxia increases AM expression in human retinal pigment epithelial cell lines. The authors state in the discussion that, because AM is expressed in various types of cells, including neurones and astrocytes, it is plausible that AM production is also increased in the neuroretina under ischemic conditions. These suggestions, taken together with the findings of the present study, raise the question of whether AM could be a paracrine modulator of retinal arterial tone that could also play a role in hypoxic vasodilation in retinal arteries. Further research is necessary to assess whether AM has a role in the regulation of retinal circulation. 

Still, the hypothetical role for AM shows some similarities with the potential role of an as yet unidentified relaxing factor released from retinal tissue. We recently described the existence of such a vasorelaxing factor and gave it the name retinal relaxing factor or RRF. ¹⁹ We also demonstrated that the RRF or a similar retina-derived relaxing factor may be responsible for hypoxic vasodilation of retinal arteries. ²⁰ It is however rather unlikely that AM is the as yet unidentified RRF, because the RRF-response in BRA was shown to be endothelium-independent. ¹⁹ To exclude further the possibility of AM’s being involved in the RRF response, the effect of CGRP 8-37 on the RRF-induced relaxation was tested. Studying the effect of any inhibitor on the RRF-response is hampered, however, by the fact that the RRF has not yet been identified nor isolated. It is therefore impossible to generate a standard concentration–response curve for the RRF. As the RRF is released from retinal tissue, a technique was used in which a small strip of retinal tissue remained attached to the BRAs. In that way, the RRF released from the adhering retinal tissue can continuously exert its relaxing influence on the BRA smooth muscle cells. Addition of increasing concentrations of a vasoconstrictor such as PGF_2α resulted therefore in a weaker contraction than in preparations without retinal tissue. The presence of CGRP 8-37 did not affect this PGF_2α concentration–response curve, suggesting that the continuous relaxing influence of the RRF released from the adhering retinal tissue was not diminished by CGRP 8-37. As CGRP 8-37 did not affect the RRF-induced relaxation, whereas it significantly affected the AM-induced relaxation, AM can be excluded as a possible candidate for the RRF. 

In conclusion, the present study clearly demonstrates that AM induces an endothelium-dependent vasodilation in isolated bovine (intraocular) retinal arteries. Thereby, our findings support the hypothesis that AM may play a role in the regulation of ocular blood flow. Activation of both CGRP and AM receptors seems to be involved, although further research is necessary to clarify their relative importance. 

 Supported by a grant from Ghent University; the Fund for Scientific Research, Flanders; and the Fund for Research in Ophthalmology, Belgium.

 Submitted for publication July 17, 2003; revised September 8, 2003; accepted September 13, 2003.

 Disclosure: K. Boussery, None; C. Delaey, None; J. Van de Voorde, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Concentration–response curve for adrenomedullin (AM) in the BRAs. Rings of the BRAs were contracted with 30 μM PGF_2α (n = 21).

Figure 2.

 

Relaxations of BRAs (expressed as the percentage decrease of the tone induced by 30 μM of PGF_2α) in response to 30 nM AM (n = 7) and 0.1 mM acetylcholine (ACh, n = 7) before (control) and after (−endo) removal of the endothelium of the BRAs (*P < 0.05).

Figure 3.

 

(A) Relaxations of BRAs (expressed as the percentage decrease of the tone induced by 30 μM of PGF_2α) in response to 30 nM AM (n = 6) and 0.1 mM acetylcholine (ACh, n = 6) in the absence (control) and presence of 0.1 mM l-NA (*P < 0.05). (B) Relaxations of BRAs (expressed as the percentage decrease of the tone induced by 30 μM of PGF_2α) in response to 30 nM AM (n = 8) and 0.1 mM sodium nitroprusside (SNP, n = 8) in the absence (control) and presence of 1 μM ODQ (*P < 0.05).

Table 1.

 

Influence of Increasing Concentrations of K⁺ in the Organ Bath on PGF_2α-Induced Contraction and on Both 10- and 30-nM AM-Induced Relaxations in the BRAs

K+ Concentration in the Organ Bath	Contraction (mN)	AM 10 nM (% Relaxation)	AM 30 nM (% Relaxation)
5 mM (standard KRB)	6.35 ± 0.51	19.22 ± 3.55	39.72 ± 9.19
30 mM	5.93 ± 0.57	12.58 ± 2.64	34.52 ± 3.86
120 mM	6.86 ± 0.52*	8.41 ± 2.13*	15.60 ± 2.96*

  n = 8.

*   P < 0.05, vs. 5 mM K⁺.

Figure 4.

 

Concentration–response curve for AM in the BRAs in (A) the absence (control) and presence of the proposed AM-receptor antagonist AM 22-52 (1 μM, n = 6), and (B) the absence (control) and presence of the proposed CGRP₁ receptor antagonist CGRP 8-37 (1 μM, n = 6). Relative responses are expressed as the percentage decrease of the tone induced by 30 μM PGF_2α (*P < 0.05).

Figure 5.

 

Concentration–response curve for AM in the BRAs in control conditions, in the presence of both l-NA 0.1 mM and ODQ 1 μM, and in the presence of l-NA 0.1 mM, ODQ 1 μM, and CGRP 8-37 1 μM (n = 4). Relative responses are expressed as the percentage decrease of the tone induced by 30 μM of PGF_2α (*P < 0.05, control versus l-NA+ODQ).

Figure 6.

 

Contractions (expressed as millinewtons of force) of BRAs with adhering retinal tissue, induced by increasing concentrations of PGF_2α in the presence and absence (control) of 1 μM CGRP 8-37 (n = 4).

Table 2.

 

Influence of Several Inhibitors of [Ca²⁺]_i-Reducing Mechanisms on PGF_2α-Induced Contraction and on Both 10- and 30-nM AM-Induced Relaxations in the BRAs

	PGF2α (mN)			AM 10 nM (% Relaxation)			AM 30 nM (% Relaxation)
		Control	Treatment	Control	Treatment	Control	Treatment
Thapsigargin (n = 4)	4.97 ± 0.65	2.52 ± 0.81*	30.09 ± 6.32	62.57 ± 9.71, †	54.05 ± 7.19	85.12 ± 3.80, †
CPA (n = 4)	5.89 ± 0.50	2.71 ± 0.53*	33.75 ± 10.70	65.00 ± 18.50, †	64.48 ± 18.34	94.00 ± 1.08, †
DMTU (n = 4)	5.85 ± 0.90	0.15 ± 0.04*	21.03 ± 4.85	—, ‡	39.68 ± 6.17	—, ‡
Amiloride (n = 4)	4.76 ± 0.68	2.46 ± 0.45*	53.00 ± 15.88	56.32 ± 15.14, †	69.00 ± 14.82	82.43 ± 9.00, †
Nifedipine (n = 4)	4.98 ± 0.24	1.68 ± 0.19*	27.94 ± 3.95	86.11 ± 2.38, †	40.06 ± 6.54	91.95 ± 3.04, †
Vanadate (n = 8)	5.75 ± 0.39	5.29 ± 0.46	16.90 ± 2.73	7.19 ± 0.78*	36.39 ± 4.35	24.64 ± 2.94*

*   P < 0.05.

†  Mean AM-induced relaxation before and after treatment was not statistically compared because of the significant difference in contractile tone.

‡  After treatment, PGF_2α 30 μM failed to induce a significant contraction in the BRAs. Therefore, no concentration–response curve for AM was constructed.

The authors thank Eric Tack for excellent technical assistance.