Because adenosine is a vascular tone modulator, we examined the effect of adenosine and congeners in the vascular reactivity of isolated human placental vessels and in perfused cotyledons. We characterized its vasomotor action and tentatively identified the receptor subtypes and their intracellular signaling mechanisms. We recorded isometric tension from the circular layer of chorionic vessel rings maintained under 1.5 g of basal tension or precontracted with KCl. The relative order of potency of adenosine and structural analogs is consistent with the expression of A2B receptors, 5′-(N-ethylcarboxamido)adenosine (NECA) being the most potent. The maximal contraction ranged from 45% to 60% of the KCl standard response, except for an A2A receptor agonist that did not exceed 15%. Consistently, NECA was 100-fold more potent than adenosine to raise the perfusion pressure of ex vivo perfused cotyledons. In contrast, a selective A3 receptor agonist relaxed precontracted rings of chorionic vessels. Whereas a selective A3 receptor antagonist was ineffective to antagonize adenosine-induced contraction, A2 or A1 receptor antagonists reduced adenosine-induced vasoconstriction concentration-dependently. Denudation of the endothelial layer reduced adenosine- and NECA-induced contractions by 50–70%. Furthermore, indomethacin reduced adenosine- or NECA-induced contractions concentration-dependently in intact and endothelium-denuded rings. A thromboxane receptor antagonist blocked adenosine- and NECA-induced contractions in intact and endothelium-denuded rings, suggesting the involvement of an arachidonic acid metabolite as the mediator of the vasoconstriction. We propose that adenosine A2B receptors mediate the adenosine-induced contraction vasomotor effect in human chorionic vessels and that this involves synthesis of a thromboxane receptor activator or a related prostanoid.
- adenosine-induced vasoconstriction
- human chorionic vessels
- arachidonate metabolite
- thromboxane receptor antagonism
- placental vasculature
the adenosine receptor family comprises four gene products, identified by pharmacological, biochemical, and molecular biology studies (45). Whereas A1 and A3 receptors are coupled to adenylate cyclase through the Gi/Go proteins and therefore decrease intracellular cAMP, the A2A and A2B receptors are coupled through Gs; the human A3 receptor has also been observed to be coupled to Gq and regulates phospholipase C (8, 9, 11, 51, 52). The diverse physiological effects mediated by the different adenosine receptors present in the central nervous, cardiovascular, and immune systems have been confirmed by studies with transgenic mice lacking these receptors (30). Specific adenosine receptor ligands have emerged (23, 26) and proven critical in tentatively identifying functional receptor subtypes, suggesting their involvement in pathophysiological processes (31). The selectivity of these agonists/antagonists is only relative, however, because the selectivity is lost at large concentrations.
In blood vessels, adenosine derives in large part from nucleotides released from platelets, endothelial cells, and perivascular nerve endings or tissue damage. On tissue ischemia, adenosine may protect the myocardium and blood vessels from reperfusion injury, playing a relevant pathophysiological role (39). Adenosine vasodilatates vessels by acting predominantly on adenosine A2A receptors on vascular smooth muscle cells (19). In mammals, including humans, coronary vessels are dilated by adenosine, which increases heart perfusion, an effect due to the activation of adenosine A2B and A3 receptors (18, 39). In contrast, A1 receptors located in the kidneys mediate a vasocontractile response that ensures a transient and local rise in blood pressure (36). In addition to controlling vascular tone, A2B receptors have been implicated in the regulation of mast cell secretion (14, 34), gene expression (15), intestinal function, and neurosecretion (18, 37). In the ovine and human placental vasculature, arachidonic acid metabolites play a critical role in the regulation of fetal-placental circulation in health and disease (41, 55); although debatable, prostaglandin E2 (46) and even prostacyclin were once considered potent vasoconstrictors (40, 55). The role of adenosine in the placental vessels remains poorly investigated. Reid et al. (47) reported that adenosine has a biphasic response in the ovine fetal placental vasculature, an observation that may indicate the expression of a mixed adenosine receptor population along the placental vasculature or signaling mechanisms at variance. On the other hand, our group recently reported (21, 54) that ATP contracts superficial chorionic vessels from the human placenta. Considering that ATP is rapidly inactivated by releasable ectonucleotidases (56), the possibility exists that even though the placenta lacks sympathetic perivascular nerves, the ATP released to the fetal-placental circulation will be rapidly degraded to adenosine and modulate the ATP vasomotor response.
Because adenosine has been proposed to play a modulator role in the regulation of human vascular tone, we aimed at examining the vascular reactivity of adenosine along the human placental vasculature by ascertaining its vasomotor action and typifying the putative adenosine receptor subtypes that might mediate these effects in isolated human chorionic vessels and intact perfused cotyledons. We further aimed at characterizing the intracellular signaling pathways that may govern the adenosine response in a tissue-specific manner. In view of the paramount role of prostanoids in placental vasculature, we ascertained whether eicosanoids might act as the mediators of the adenosine-induced vasomotor action in human chorionic vessels.
The present vascular reactivity assays and ex vivo cotyledon perfusion protocols show that both endothelial cells and vascular smooth muscles from human chorionic vessels and cotyledons have predominantly A2B receptors, which are coupled to the arachidonic acid cascade and cause a vasoconstriction, mediated apparently by the release of a thromboxane or related prostanoid. RT-PCR studies confirmed the expression of A2B receptors in both endothelial and smooth muscles of these vessels. The finding that placental vessels contract to applications of adenosine, in contrast to most other arteries or veins in which adenosine is known to cause vasorelaxation, illustrates the concept that the vascular reactivity of adenosine varies in humans in a tissue-specific manner.
MATERIALS AND METHODS
Adenosine Receptor Ligands and Source of These Compounds
Adenosine and related structural analogs such as 5′-(N-ethylcarboxamido)adenosine (NECA), 2-p-(2-carboxyethyl)phenethylamino-5′-N-ethylcarboxamidoadenosine HCl (CGS-21680), N6-cyclohexyladenosine (CHA), 9-chloro-2-(2-furanyl)-5-([phenylacetyl]amino)-[1,2,4]triazolo[1,5-c]quinazoline (MRS-1220), serotonin hydrochloride (5-HT), 1-(4-chlorobenzoyl)-5-methoxy-2-methyl-3-indoleacetic acid (indomethacin), prostaglandin E2, and prostaglandin F2α were purchased from Sigma-Aldrich (St. Louis, MO). 1-[2-Chloro-6-[[(3-iodophenyl)methyl]amino]-9H-purin-9-yl]-1-deoxy-N-methyl-β-d-ribofuranuronamide (2-Cl-IB-MECA), 8-cyclopentyl-1,3-dipropylxanthine (DPCPX), and 4-(2-[7-amino-2-(2-furyl)[1,2,4]triazolo[2,3-a][1,3,5]triazin-5-ylamino]ethyl)phenol (ZM-241385) were purchased from Tocris Cookson (Ellisville, MO). ([1R-[1α(Z),2β,3β,5α]]-(+)-7-[5-[[1,1′-biphenyl)-4-yl]methoxy]-3-hydroxy-2-(1-piperidinyl)-cyclopentyl]-4-heptenoic acid (GR32191) was provided by Glaxo Wellcome. Analytical grade reagents for buffer preparation were purchased from Merck (Darmstadt, Germany).
Source of Human Placentas
Over 160 full-term placentas from normal pregnancies delivered by vaginal or cesarean section were transported from the delivery room to the laboratory within <20 min after childbirth. The Committee on Ethics of the Medical Research Department approved these protocols and consent forms, and ethical regulations were strictly followed.
Dissection of Placentas to Obtain Rings from Superficial Chorionic Vessels
Immediately after placentas arrived in the laboratory, a 2- to 4-cm segment of second- and third-order superficial chorionic artery and vein vessels were routinely dissected to obtain 0.5-cm rings and mounted on a bath chamber used for vascular reactivity studies as detailed by Valdecantos et al. (54). Most of the protocols were repeated at least four times, using vessels dissected from different placentas each time. Studies on artery or vein rings were performed with intact (E+) or manually endothelium-denuded (E−) vessels (2, 54). Vessel rings were placed in Krebs-Ringer buffer maintained at 37°C within a double-jacketed organ bath chamber bubbled with a 95% O2-5% CO2 mixture. Isometric muscular tension from the circular layer was recorded with a Grass force displacement transducer connected to a Grass 7 multichannel polygraph. Vessel rings were manually adjusted to an artificial tension of 1.5 g, which, as described by Valdecantos et al. (54), is an optimal value obtained from length-tension curves; this tension was maintained throughout the experiment. After an equilibration period of 60 min, with buffer washouts every 15 min, the rings were challenged with three successive applications of 70 mM KCl to evoke the maximal smooth muscle contraction, used to quantify and standardize responses within vessels. This concentration of KCl proved optimal for this caliber of vessels [see Valdecantos et al. (54) and Racchi et al. (44) for human saphenous vein biopsies]. This long-lasting protocol was routinely followed with each vessel ring preparation; it allowed a thorough washout of hormones or drugs that could have been present in the tissues due to parturition or cesarean section surgery.
Vascular Reactivity Bioassays in Intact and Endothelium-Denuded Vessels
Determination of agonist potency.
The contractile potency of adenosine, the endogenous ligand for all four adenosine receptor subtypes, and agonists with preferential selectivity for each of the alleged receptor subtypes, CHA (a preferential A1 receptor ligand), CGS-21680 (a rather selective A2A receptor agonist), NECA (a nonselective A2 receptor agonist), and 2-Cl-IB-MECA (recognized as a relatively selective A3 ligand), was assessed by means of noncumulative concentration-response protocols. Varying adenosine concentrations were added directly to the tissue bath chambers in random order within two or three orders of magnitude concentration range. The adenosine applications were spaced every 45 min, a time lag required to avoid desensitization of the adenosine-induced contraction. Parallel control experiments were conducted to examine whether the adenosine vasomotor response decayed over time during the 4- to 5-h course of the experiment. In the particular case of NECA, CHA, or CGS-21680, because the second application caused a much diminished response, a single concentration was assessed per bioassay, i.e., in the concentration-response curve protocols, each data point is derived from separate vessel rings, each obtained from an individual placenta. In every case, at least three to five rings were used to quantify the contractions elicited by NECA, CHA, or CGS-21680. In addition, we also examined whether, in precontracted vessels, adenosine and analogs also elicited contractions. For this purpose, a single concentration of 19 μM adenosine, 57 μM CHA, 37 μM CGS-21680, or 1 μM NECA was also applied in 20 mM KCl-precontracted rings.
In every single bioassay protocol, rings were challenged with a 70 mM KCl standard at the beginning and the end of the experiment, a concentration previously demonstrated to cause the maximal contracture of these vessels (54). This procedure allowed the standardization of the concentration-response protocols described. Furthermore, adenosine concentration-response studies were performed in E+ and E− vessels. Likewise, in the case of NECA, a single concentration was assayed in tissues with and without the endothelial cell layer.
Studies with adenosine receptor antagonists.
Intact vessel rings were challenged repeatedly with adenosine at 19 μM, a value close to its EC50, which did not evidence desensitization after successive applications. Each antagonist was examined in a wide range of concentrations (6–1,500 nM); the antagonists were preapplied 10 min before the standard 19 μM adenosine challenge. The results are plots of the percentage of 70 mM KCl values of the vasomotor action of the adenosine challenge vs. the antagonist concentration. Independent protocols, performed in separate vessels, examined the potency of DPCPX, ZM-241385, or MRS-1220 to block the standard adenosine-evoked contractions. These antagonists have relative selectivity for the A1, A2A, and A3 receptor subtypes, respectively. The concentration of these antagonists that halved the adenosine-induced contraction was interpolated from the respective concentration-response curves and expressed as the mean ± SE for each receptor blocker. To assess whether more than one adenosine receptor mediates the adenosine response, additional studies were performed applying 200 nM ZM-241385 and 200 nM DPCPX together.
Blockade of tissue cyclooxygenase and thromboxane receptor.
Because eicosanoids often act as mediators of the vasomotor actions elicited by serotonin and other vasoconstrictors in umbilical vessels (3, 7, 13), we assessed the involvement of the arachidonic acid cascade in adenosine-induced vasoconstrictions. For this purpose, tissue cyclooxygenases were blocked with indomethacin (42), a nonselective COX-1/COX-2 inhibitor. The indomethacin concentration necessary to reduce the 19 μM adenosine-induced vasoconstriction was determined in a first set of experiments. Next, chorionic rings were incubated with 100 nM indomethacin for 40 min before the performance of a complete adenosine concentration-response protocol. As control, tissues were maintained in the bath chamber during the same 40 min but without drug treatments, except for the challenge with 19 μM adenosine. Tissue preincubation with 100 nM indomethacin was also used to assess the blockade of the vasoconstriction elicited by 1 μM NECA. Reversibility was examined by challenging the tissues several times until the original contractile response was recovered. Analogous protocols were performed in endothelium-denuded vessels.
In the next series of protocols we assessed whether the thromboxane receptor blocker GR32191 (33) antagonizes the adenosine-induced contraction. Adenosine concentration-response protocols were performed in the absence of GR32191 and then after a 5-min exposure to 10, 70, or 210 nM GR32191. Additional experiments assessed whether 140 nM GR32191 also reduced the vasomotor response elicited by 1 μM NECA in E+ and E− vessels. Separate experiments addressed the specificity of the GR32191-induced blockade of the adenosine-evoked contractions by challenging vessels with 70 mM KCl before and after a 5-min incubation with 140 nM GR32191.
RT-PCR Amplification Studies
Four- to five-centimeter segments of chorionic arteries or veins with and without the endothelial cell layer were placed in RNA stabilizing solution for RNA extraction and PCR studies. Total RNA from each chorionic vessel was extracted by the standard Chomczynski and Sacchi procedure (5). The oligonucleotide amplification primers for each human (h) adenosine receptor subtype and the length of the expected PCR products (in parentheses) were A1: sense 5′-CTCTAGAGATGCCGCCCTCC-3′, antisense 5′-CGGAATTCCCAGGGCCAGGA-3′ (311 bp) (35); A2A: sense 5′-AGATGGAGAGCCAGCCTCTG-3′, antisense 5′-GCTAAGGAGCTCCACGTCTG-3′ (427 bp) (20); A2B: sense 5′-GAGCTGATGGAGCACTCGAG-3′, antisense 5′-ACACCGAGAGCAGGCTGTAC-3′ (342 bp) (22); and A3: sense 5′-ACCCCCATGTTTGGCTG-3′, antisense 5′-GCACAAGCTGTGGTACCTCA-3′ (361 bp) (29).
The PCR reactions were performed in 25-μl final volumes, containing each primer at 0.5 μM, 2 μl of cDNA, 200 μM dNTPs, 1 U of Taq DNA polymerase, and 1× Taq DNA polymerase reaction buffer. The PCR thermal profile for hA1, hA2A, and hA2B were 3 min at 94°C and 30 cycles of 1 min at 94°C, 1 min at 60°C, and 3 min at 72°C. A final extension at 72°C for 7 min was completed. PCR conditions for hA3 were 2 min at 95°C, followed by 35 cycles of 30 s at 94°C, 30 s at 55°C, and 30 s at 72°C and a last step of 10 min at 72°C. A sample without cDNA was subjected to this protocol as a negative control. The PCR products were separated by electrophoresis in agarose gel and visualized with ethidium bromide staining. The specificity of product bands was confirmed by analysis of the nucleotide sequence with an ABI 3100 Sequencing Automatic Analyzer.
Additional protocols were followed to evaluate the presence of human CD31 and human myosin alkaline light chain (MALC) isoform 6, selective endothelium and vascular smooth muscle markers, respectively. The corresponding primers were obtained from the GenBank database as detailed by Buvinic et al. (4). For this purpose, RNA extracts from chorionic vessels with or without the endothelial layer were utilized. All molecular biology reagents and buffers were obtained from GIBCO-BRL Life Technologies, Promega, and Ambion RNA.
Perfusion of Placental Cotyledons and Assessment of Vascular Resistance
Individual cotyledons were perfused with Krebs-Ringer buffer gassed with 95% O2-5% CO2 through plastic tubing cannulas inserted into the corresponding placental artery. After a 40-min equilibration period, 300 μM adenosine or 10 μM NECA dissolved in buffer was perfused through the arterial cannula for 1 min. Perfusion pressure was monitored continually on a Grass polygraph; the changes in perfusion pressure were recorded through a strain gauge connected to the artery. Generally, for adenosine we tested seven concentrations in four different placentas, whereas for NECA four concentrations were tested; a single agonist concentration was examined per cotyledon, and each concentration was examined in at least four separate placentas. This procedure was determined to be optimal for concentration-response determinations, because we noted in a preliminary series of experiments that a second perfusion of NECA, but not adenosine, performed 25 min later, evoked a markedly reduced contraction. In a further set of three protocols, the vasomotor action elicited by 300 μM adenosine was assessed in cotyledons precontracted with 0.3 μM 5-HT, a procedure that allowed us to discard the idea that adenosine might act as a vasorelaxant in precontracted tissues. Results are plotted as the mean changes in perfusion pressure elicited by both adenosine and NECA.
Data Quantification and Statistical Analysis
The median effective contractile concentration (EC50) was interpolated from each agonist concentration-response curve (at least 4 points defined a curve, and each point was repeated in individual vessel rings from at least 3–8 separate placentas) and served to establish the relative order of potency of adenosine and structurally related analogs; EC50 values are expressed as micromolar. The contraction elicited by each agonist concentration was recorded in grams of tension, normalized according to the KCl standard, allowing comparisons within multiple vessel rings, and plotted in a standard concentration-response curve format. With regard to the relaxant effect induced by 2-Cl-IB-MECA, this effect was quantified by normalizing the percent relaxation of the 20 mM KCl-induced tension; the median agonist concentration required to elicit relaxation was interpolated from each concentration-response curve protocol. In the case of the receptor subtype antagonist studies, the concentration required to halve the contractile response to the standard adenosine challenge (IC50) was interpolated for each antagonist experiment (n = 4).
GraphPad software (GraphPad, San Diego, CA) was used to fit concentration-response curves. Analysis of variance established the statistical significance of several treatments on the concentration-response curves. When necessary, the Student’s t-test tables (paired or nonpaired) were used; Dunnett's tables for multiple comparisons with a single control were likewise used when appropriate. Significance was established by a P value <0.05.
Vasomotor Action of Adenosine and Structurally Related Analogs in Intact Vessels
Adenosine contracted isolated chorionic arteries or veins concentration-dependently, reaching slightly more than 40% of the KCl standard contracture. The adenosine EC50 was almost identical in artery and vein and averaged 37 μM (Table 1 and Fig. 1). NECA, a highly potent but nonselective adenosine receptor agonist, was almost 100-fold more potent than adenosine and reached 60% of the KCl contracture, and CHA, a classic A1 receptor ligand, was 2-fold more potent than adenosine and maintained the maximal response; that is, they both reached 40% of the KCl standard contracture, whereas the potency of CGS-21680, a prototype A2A receptor agonist, was indistinguishable from that of CHA, but its maximal contracture did not exceed 15% (Table 1). Furthermore, adenosine and congeners also raised the tension of precontracted vessels; the magnitude of the contraction was similar to that obtained in nonprecontracted tissues (see tracings in Fig. 1), although only 2-Cl-IB-MECA, an alleged A3 receptor agonist, which per se was inactive as a vasoconstrictor, concentration-dependently relaxed KCl-precontracted chorionic vessels (Fig. 2). The potency of 2-Cl-IB-MECA to induce relaxations was 27.5 ± 7 μM (n = 4) in vein rings; a similar value was found in arteries (18.5 ± 8.5 μM n = 3). The maximal relaxation achieved 58.7 ± 7% of the KCl-induced tension (n = 4).
Successive CHA, CGS-21680, or NECA applications in the range of their EC50 values resulted in a gradual and substantial loss of vasomotor activity (Table 2). In contrast, three or more sequential adenosine applications elicited contractions of similar magnitude (Table 2).
The KCl standard challenge elicited a significantly larger contraction in chorionic vein than in artery rings [1.54 ± 013 g (n = 29) vs. 0.88 ± 0.08 g (n = 29), P < 0.001]; however, we did not observe significant differences between the KCl contracture in vessels derived from cesarean section placentas vs. vaginal birth placentas, ruling out the influence of drugs used during the surgical and/or anesthetic procedures.
Adenosine-Induced Vasoconstriction Raises Perfusion Pressure of Isolated Cotyledons
Consistent with the finding that NECA is much more potent than adenosine to contract isolated chorionic ring vessels, NECA was at least 100-fold more potent than adenosine to raise the perfusion pressure of human placenta cotyledons (Fig. 3). Consonant with the vascular reactivity observations in the isolated rings from the chorionic vessels, a second NECA application did not cause a rise in the perfusion pressure, in contrast to adenosine, which showed essentially no desensitization (see recordings in Fig. 3). Moreover, paralleling the observations in isolated vessel rings, adenosine caused a further rise in perfusion pressure in precontracted cotyledons (Fig. 3). Interestingly, the magnitude of the rise in perfusion elicited by 300 μM adenosine in noncontracted and precontracted tissues was the same (28.3 ± 4 mmHg; n = 4).
Additional experiments demonstrated that prostaglandin F2α elicited a concentration-dependent rise in the perfusion pressure (its estimated EC50 was 3 ± 0.43 μM), whereas the application of 5 μM prostaglandin F2α evoked a rise in the perfusion pressure of 137.5 ± 12.5 mmHg (n = 3) and that of 0.5 μM 5-HT caused a 91 ± 15 mmHg rise in perfusion pressure (n = 3). Therefore, prostaglandin F2α is at least 100-fold more potent than adenosine as a human cotyledon vasoconstrictor.
Adenosine Receptor Classification Based on Use of Receptor Antagonists
DPCPX and to a much lesser extent, ZM-241385, concentration-dependently blocked the challenge adenosine-induced contractions. The estimated potency of DPCPX and ZM-241385 to halve the adenosine contraction standard in chorionic arteries was 15 ± 3.7 (n = 6) and 207 ± 42 (n = 3) nM, respectively (Fig. 4). Likewise, in veins, these values were 22 ± 3.8 (n = 6) and 633 ± 84 (n = 3) nM, respectively. MRS-1220 did not significantly block the adenosine-induced contractions in either chorionic arteries or veins [Fig. 4; linear regression r = 0.29 and 0.014 (not significant) in chorionic artery and vein, respectively]. Successive adenosine applications caused a similar vasoconstriction, as exemplified in Table 2, ruling out the idea that the antagonist-evoked reduction of the adenosine vasoconstriction is due to muscular damage or nonspecific metabolic alterations.
In addition, 200 nM ZM-241385 reduced the magnitude of the contractions induced by the 19 μM adenosine standard in artery rings by 47% [from 18.7 ± 2.2% of the KCl contracture (n = 21) to 8.7 ± 1.6% of the maximal KCl response (n = 7); P < 0.05], 200 nM DPCPX caused a 69% reduction of the response (from the control value to 5.8 ± 1.4% of the maximal KCl response; n = 12; P < 0.01). However, the coapplication of both antagonists only reduced the adenosine response by 77% (4.3 ± 0.7%; n = 4; P < 0.05), a value that is not significantly different from that attained by the separate application of each antagonist. Similar findings were also observed in chorionic vein rings (data not shown).
RT-PCR Products Identify Adenosine Receptor mRNAs
RT-PCR further identified the mRNAs coding for multiple adenosine receptors. In E+ and E− vessels, RT-PCR products identified the mRNAs coding for the human adenosine A2A, A2B, and A3 receptors in chorionic arteries or veins (Fig. 5). In addition, RT-PCR products for the A1 receptor were detected only in E+ veins (Fig. 5). The removal of the endothelial cell layer was confirmed by the loss of the CD31 marker, whereas in the presence of the MALC, a smooth muscle tracer was preserved in the absence of endothelium (Fig. 5). This experiment was repeated with mRNA extracted from three separate placentas.
Role of Endothelium
Removal of the endothelial cell layer reduced by about two-thirds the adenosine- and NECA-induced contractions. Endothelial cell removal caused a flattening and downward shift of the adenosine concentration-response curve in both artery and vein rings (P < 0.01, n = 7; Fig. 6). The NECA-induced contraction was likewise more than halved after endothelium removal (P < 0.05; Fig. 6). Acetylcholine, a widely accepted endothelium-dependent vasodilator, is used to functionally evaluate the role of the endothelium in vascular reactivity assays. However, it did not relax the intact isolated chorionic vessels; in fact, 1–10 μM acetylcholine contracted one-third of chorionic rings. Likewise, 10–100 μM acetylcholine failed to relax precontracted perfused human placental cotyledons. Therefore, because the acetylcholine challenge was not useful as a functional test of endothelium denudation, the absence of CD31, revealed in the RT-PCR studies, confirmed in this case the efficacy of the endothelium removal procedure.
Additionally, the 2-Cl-IB-MECA-induced relaxation was independent of the endothelium (Fig. 2). Removal of the endothelial cell layer did not significantly modify the potency or the maximal relaxation evoked by this agonist in isolated chorionic veins [15.4 ± 2.4 (n = 4) vs. 27.5 ± 7 μM (n = 4)].
Involvement of Arachidonic Cascade in Adenosine- and NECA-Evoked Contractions
In view of the fact that arachidonate metabolites have a preponderant role in vascular tone regulation, we assessed the putative role of eicosanoids as mediators of adenosine-evoked contractions.
Indomethacin concentration-dependently inhibited adenosine-evoked contraction in both chorionic artery and vein rings (Fig. 7). Indomethacin blockade was reduced by 30–60% after a 120-min washout period (Fig. 7). Moreover, 100 nM indomethacin obliterated and flattened the adenosine concentration-response curve in both artery and vein rings (Fig. 7). Likewise, 100 nM indomethacin reduced by 92 ± 4% (P < 0.01) the 1 μM NECA-induced contraction in either vessel (Fig. 7, inset). Moreover, 100 nM indomethacin also reduced by 80% the 1 μM NECA-evoked vasoconstriction in vessel rings denuded of the endothelial layer (P < 0.01; n = 3). Furthermore, 2-Cl-IB-MECA-evoked relaxation was also blocked in a concentration-dependent manner by 0.1–1 μM indomethacin (see recordings in Fig. 2).
Thromboxane receptor blockade.
The thromboxane inhibitor GR32191, which per se did not alter the basal tone of the rings, concentration-dependently reduced the vasoconstriction evoked by adenosine (P < 0.05 in arteries and veins), whereas 70 nM GR32191 halved the vasoconstrictions and increasing the concentration to 210 nM obliterated the adenosine concentration-response curve (Fig. 8).In separate experiments, 140 nM GR32191 reduced by 70–90% the 1 μM NECA-induced vasoconstriction (P < 0.05) in artery or vein chorionic rings. Likewise, GR32191 abolished the 19 μM adenosine- and 1 μM NECA-induced vasoconstriction in endothelium-denuded vessels (data not shown). In separate control protocols, 140 nM GR32191 reduced only by 8.7% the 70 mM KCl-evoked contractions (n = 4), evidencing antagonist selectivity. Previous studies from our laboratory (54) showed that the EC50 for prostaglandin F2α is 0.09 ± 0.004 μM, a value 400-fold more potent than adenosine, with a maximal effect that doubled that elicited by the KCl standard challenge. The present results show that prostaglandin E2 is also a constrictor of isolated chorionic vein segments; although its EC50 is 170-fold less than prostaglandin F2α (15.5 ± 2.1 μM; n = 4), it is still 2.4-fold more potent than adenosine.
The present results support the hypothesis that the adenosine-induced vasoconstriction of human chorionic vessels is mediated by the activation of A2B receptors, an effect that appears to be related to the synthesis of an arachidonate metabolite that ultimately activates a thromboxane receptor in the vascular smooth muscles. Figure 9 summarizes and schematizes this proposal, indicating that adenosine contracts chorionic vessels via the activation of a dominating population of adenosine A2B receptors located in both endothelial and vascular smooth muscle cells. These receptors are coupled to the synthesis of an arachidonate metabolite, likely thromboxane A2, which might activate a thromboxane TPτ receptor, as the final effector of the adenosine contractile response. At present we cannot exclude the possible participation of related prostanoids. The mechanism that links A2B receptor signaling to phospholipase activity, and thereafter activation of tissue cyclooxygenase, remains to be further described. The present results highlight for the first time a role for adenosine as an extracellular signal involved in blood flow regulation in the human placental vasculature and links its transduction mechanism to an arachidonic acid cascade metabolite.
Adenosine contracted isolated chorionic arteries and veins and, moreover, increased the perfusion pressure of cotyledons, a finding that does not imply, but suggests, commonalities in receptor distribution along the placental vasculature. We are aware that vasoconstrictors such as angiotensin II or norepinephrine do not have the same responses along the whole fetal-placental circulation, at least in the sheep, a common animal model used to study cardiovascular development. Almost 20 years ago, Adamson et al. (1) showed that angiotensin II increased umbilical artery resistance in the sheep, whereas no change was detected in cotyledon or venous resistances. Likewise, norepinephrine had no effect on cotyledon vascular resistance but constricted downstream vessels (1). Radiolabeling binding studies further confirmed the data of Adamson et al. (1) and established that only umbilical vessels expressed type 1 angiotensin II receptor (6, 50). In addition, the density of angiotensin II receptors in placental primary/secondary stem vascular tissue depends on whether the delivery is vaginal or by cesarean section (25), a finding that has pathophysiological consequences because it depends on the circulating levels of angiotensin II, which change with type of delivery. Furthermore, 2-Cl-IB-MECA, an alleged adenosine A3 receptor ligand, relaxed human chorionic vessels, an effect that is only detectable in precontracted rings and that also appears to involve the activation of the arachidonate cascade. Reviriego et al. (49) and Huidobro-Toro and Valdecantos (21) previously reported that 500 μM adenosine is required to vasodilatate precontracted human chorionic vessels, an indication of a dual mechanism of adenosine on human chorionic vessels. We infer that in the human superficial chorionic vessels, although adenosine might vasodilatate, this effect is hindered by the more potent adenosine A2B receptor-mediated vasoconstriction, revealing that multiple adenosine receptors might regulate the placental vascular tone, a conclusion consonant with the findings of Reid et al. (47), who first observed the biphasic action of adenosine while studying the physiology of ovine placental vessels.
For a comparison, we have not overlooked the fact that adenosine contracts the human bronchia, an effect that is also indirect because adenosine activates mast cell adenosine A2B receptors (53), triggering the release of a variety of chemical mediators, including prostanoids of the leukotriene family. In addition, Nicholls et al. (38) reported that adenosine and congeners contract the rat duodenum muscularis mucosae, in contrast to the rat duodenum longitudinal muscle that is relaxed by adenosine. In common with the present findings, Nicholls et al. (38) inferred that this adenosine-induced contraction is mediated through the preponderant activation of A2B receptors.
Several arguments substantiate our hypothesis that the adenosine-induced vasoconstriction in human placental vessels is mediated predominantly by activation of the A2B receptor. The finding that CGS-21680, a prototype A2A receptor agonist, was inactive at 1 μM suggests that the A2A receptor may be ruled out (28). Furthermore, the potency of NECA is within the range of the potency described for this agonist as an A2B receptor agonist (17). Likewise, and by the same argument, we may also discard the influence of A3 receptors, because 2-Cl-IB-MECA, a high-affinity A3 receptor agonist (27), was inactive as a vasoconstrictor and instead induced relaxations. With regard to the role of A1 receptors, we conclude that the large CHA concentrations required to contract chorionic vessels may be mediated by a nonspecific interaction with the A2B receptor (15). Furthermore, the use of receptor antagonists is consonant with our proposal. In the absence of a selective A2B receptor antagonist, we examined the blockade of the adenosine-induced vasoconstriction with DPCPX, ZM-241385, and MRS-1220. The potency of DPCPX is four- to sixfold lower than what is described for the human A1 receptor, but closer to its potency as an A2B receptor antagonist (28). ZM-241385, an alleged selective and potent A2A receptor antagonist (43), was found to be even less potent than DPCPX, indicating that the nature of the ZM-241385-induced antagonism could be related to a nonspecific blockade of the alleged A2B receptor (32). Confirming the existence of a single adenosine receptor population, and that both antagonists act at a common site, the joint application of ZM-241385 and DPCPX did not result in an increased antagonism. Furthermore, MRS-1220, a specific A3 receptor antagonist (24), did not block the adenosine-induced contraction.
Consistent with this interpretation, RT-PCR studies identified the A2B receptor subtype mRNA in these vessels. A1 receptor mRNA was found only in intact veins, suggesting that the mRNA might be further restricted to endothelial cells. We are aware that the mRNAs for the A1 and A2A receptors were detected by this procedure; however, at present we have no indication for a functional role of these receptors in the placental vasculature, except for the mRNA coding for the A3 receptor, which might be linked to vasodilatation.
Regarding the intracellular signaling mechanisms of adenosine in human chorionic vessels, the present results might be interpreted to indicate that arachidonate metabolites are in all likelihood the mediators of the adenosine effects in the placental vasculature. Furthermore, we infer that the thromboxane TPτ receptor (3, 10, 33) is the final effector involved in the adenosine-induced contraction. The present results highlight the preponderant role of arachidonate metabolites, such as prostaglandin E2, prostaglandin F2α, and likely thromboxane A2, in the control of the placental vascular tone (40, 41). Furthermore, the present data demonstrate the potent and efficacious contractions mediated by prostanoids in isolated vessels as in perfused cotyledons, a finding that further supports their role as mediators of the adenosine contractions. Why the adenosine-induced contractions do not desensitize, in contrast to the rapid loss of response observed with NECA, CHA, or CGS-21680, remains an open question. However, we recognize that the lack of desensitization of the adenosine responses allowed us to study its physiology in human chorionic vessels.
An issue that has gained increasing interest within the past few years, and that may have profound implications for this research and clinically oriented pathophysiological studies, refers to the discovery that tissue hypoxia upregulates the expression of the adenosine A2B receptor while decreasing the mRNA for the A2A receptor (12, 16) in both human umbilical vein endothelial cells and bronchial smooth muscle cells. Because the placentas used in this study were not oxygenated in the 20-min period that elapsed between child delivery and laboratory manipulation, we are aware that some degree of modulation might have occurred and may have a definitive influence on our vascular reactivity studies. The group of Biaggioni (16) demonstrated that 3 h of hypoxia sufficed to triple the mRNA for the adenosine A2B receptor in endothelial cells from the umbilical vein, while it reduced by almost 90% the mRNA coding for the A2A receptor. However, it is difficult at present to evaluate the sole influence of this regulatory mechanism in vascular reactivity.
Although the present research was not aimed at determining the origin of the local placental adenosine production, it is important to keep in mind that adenosine in blood vessels derives from platelet release, endothelial cells, nerve endings, tissue damage, and so forth. These processes may be exacerbated by a variety of stimuli, including, among others, ischemia, local hypoxia, and platelet aggregation. As to the physiological role that adenosine may play in fetal adaptation, the present results indicate the importance of the adenosine receptors in the regulation of fetal-placental circulation in health and disease states, including eclampsia and related diseases.
It remains a challenge to examine whether there is a common intracellular signaling mechanism that mediates the A2B receptor activation in different tissues, confirming the generality of our proposal. Adenosine is known to cause a transient rise in blood pressure, a finding that appears unique to the kidney microcirculation and involves A1 receptor activation (36). Although at present we cannot distinguish whether the indirect adenosine-evoked vasocontractile mechanism operates exclusively in human placental vessels, the vasomotor response developed by serotonin in placental vessels is also indirect in nature and dependent on the synthesis of an arachidonate metabolite (7, 13), suggesting that in the human placenta, which lacks sympathetic innervation (48), the arachidonic acid cascade has a major role in mediating the vasomotor regulation discussed for adenosine. The pathophysiological implications of the present findings for preeclampsia and related vascular diseases during pregnancy are open for future research.
This work was funded by Fondo Nacional de Desarollo Áreas Prioritarias-Biomedicine Grant 13980001; the Instituto Mileniode Biología Fundamental y Aplicada also contributed to the funding of this investigation. R. López did an extended summer rotation at the Department of Physiology as a medical student and was funded by local intramural grants.
We are indebted to Prof. Drs. A. Germain and J. Carvajal and the personnel of the Clinical Hospital Maternity Ward for help with the expeditious handling of the placentas. The collaboration of the personnel of the Obstetrics Department and the fellows of the maternity ward at the Medical School of the Pontifical Catholic University is much appreciated. The thromboxane receptor antagonist GR32191 was a kind donation from Dr. P. Humphrey, Glaxo-Wellcome UK. The editorial assistance of M. Brenda Watt and the review of the manuscript by Prof. E. Contreras are greatly appreciated.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 by the American Physiological Society