Ontogeny of Vascular Angiotensin II Receptor Subtype Expression in Ovine Development

Abstract

Angiotensin II (ANG II) increases arterial pressure in fetal sheep and may modulate cardiovascular adaptation before and after birth. The type 1 angiotensin II receptor (AT₁R) predominates in adult vascular smooth muscle (VSM) and mediates vasoconstriction. In contrast, AT₂R predominate in fetal tissues and are not known to mediate contraction. Although sheep are commonly used to study cardiovascular development, the ontogeny and distribution of VSM ATR subtypes is unknown. We examined ATR binding characteristics and subtype expression across the umbilicoplacental vasculature and in aorta, carotid, and mesenteric arteries from fetal (n = 44; 126-145 d gestation) and postnatal (n = 65; 1-120 d from birth) sheep using plasma membranes from tunica media and tissue autoradiography. Binding density (B_max) was similar throughout the umbilicoplacental vasculature (p = 0.5), but only external umbilical arteries and veins and primary placental arteries expressed AT₁R, whereas subsequent placental branches and fetal placentomes expressed only AT₂R. Systemic VSM B_max and binding affinity did not change significantly during development (p > 0.1). Fetal systemic VSM, however, expressed only AT₂R, and binding was insensitive to GTP_γS. Transition to AT₁R in systemic VSM began 2 wk postnatal and was completed by 3 mo. Before birth, umbilical cord vessels are the primary site of AT₁R expression in fetal sheep, and AT₂R seem to predominate in systemic VSM until 2-4 wk postnatal.

Main

The renin-angiotensin system (RAS) is functional in the developing fetus and newborn and is considered to be an important regulator of arterial pressure and cardiovascular adaptation before and after birth^(1–8). Circulating levels of angiotensin II (ANG II) increase in the presence of fetal hemorrhage and hypovolemia as well as after birth^(1,4,5,8). Inhibition of the RAS with nonspecific and specific ANG II receptor (ATR) ligands or angiotensin-converting enzyme inhibitors results in a fall in basal arterial pressure and accentuation of hypovolemic episodes^(2,7,8). Receptor blockade also modifies the baroreflex and, in newborn sheep, the reflex control of renal sympathetic nerve activity^(9,10). Thus, there is substantial evidence suggesting that ANG II plays an important role in regulating circulatory function during development.

The vascular effects of ANG II are mediated by activating the ATR, which belongs to the seven transmembrane super family of receptors⁽¹¹⁾. We have shown that the ATR are expressed in ovine fetal vascular smooth muscle (VSM) from various vascular beds and have binding characteristics similar to that seen in adults⁽¹²⁾. That is, there is a single class of saturable, high affinity binding sites resembling the ATR in adult VSM. Similar observations have been reported for the human placenta⁽¹³⁾. Further, the ATR in ovine fetal aorta and placental arteries does not change its binding density or affinity during the last third of gestation, and values are similar to that in adult sheep⁽¹²⁾. This has not been carefully assessed after birth. More recently, two major subtypes of the ATR have been identified. The AT₁R is the predominant receptor expressed in nearly all adult tissues, including VSM. To date, the majority of biologic functions of ANG II seem to be mediated by activating this receptor through G-protein and calcium-dependent mechanisms, including smooth muscle contraction, cell growth, and fluid and electrolyte regulation⁽¹¹⁾. The AT₂R is the product of a separate gene located on the X chromosome and has about 40% amino acid sequence homology with the AT₁R⁽¹⁴⁾. It is highly expressed in the fetal and newborn rat^(15–18) but also is found in select adult tissues, including myometrium^(19,20), adrenal gland^{(17,18,21–23)}, kidney^(17,21,22), uterine artery^(19,24), and the cerebral vasculature of the adult rat^(25,26). Importantly, it does not modulate smooth muscle contractions^{(19,20,27,28)}. Thus, its functions and mechanism(s) for activation are less clear than those of the AT₁R and may be cell specific⁽²⁹⁾. Although AT₂R expression is high in tissues from the fetal rat^(15–17), it is no longer detected 72 h postnatal⁽¹⁶⁾. In these studies, determination of tissue-specific expression was limited, and it was unclear which subtype was expressed in VSM, if expression was similar in all vascular beds during development, and when in development ATR subtype expression in VSM changed.

Fetal sheep have been used to study cardiovascular development and function⁽³⁰⁾. Because they are easily studied in vivo, much of our understanding of the function and development of the RAS and ANG II has been derived from this model. Compared with the rodent, ovine development occurs over a longer period of time. Therefore, we and others^(31–34) have used this species to more clearly delineate the ontogeny of vascular changes. Studies of the ontogeny of ATR expression and regulation are limited, and reports to date relate primarily to expression by the kidney, adrenal gland, placenta, and brain^{(12,13,22,23,25,26,35)}. The purpose of the present study, therefore, was to determine 1) if ATR binding characteristics in VSM are developmentally regulated, 2) if ATR subtype expression and distribution in VSM are developmentally regulated and tissue specific, and 3) if differences exist in subtype expression between the umbilicoplacental and the systemic vasculature. The latter is of particular interest as the umbilicoplacental vascular bed has been reported to be more responsive to the vasoconstricting effects of ANG II than the fetal systemic vasculature^{(2,3,36–38)} and to exhibit differential responsiveness to ANG II within this vascular bed⁽³⁹⁾.

METHODS

Tissue collection. ATR binding characteristics and subtype expression were determined in the umbilicoplacental vasculature of 17 late third trimester ovine pregnancies (127-140 d gestation; term ∼145 d) and in the aorta, carotid, and mesenteric arteries from fetal (n = 27; 126-145 d gestation) and four groups of postnatal sheep: 1 wk (n = 20; 4-9 d), 2 wk (n = 15; 13-17 d), 1 mo (n = 14; 30-37 d), and ≥3 mo (n = 16; 90-120 d). The aorta and carotid artery were chosen to represent vessels from the systemic vasculature, which were used in prior studies of receptor ontogeny in the rodent^(16,26,40). We also examined a representative muscular artery, i.e. the mesenteric artery, that has not previously been studied to compare ATR binding characteristics and subtype expression between two types of systemic arteries. Pregnant ewes were euthanized with i.v. pentobarbital sodium (50 mg/kg), which also euthanizes the fetus. Each fetus was rapidly delivered, dried, weighed, and measured to confirm gestational age. Immediately thereafter, the intra-abdominal (internal) and extra-fetal (external) portions of the umbilical arteries and the 1st-4th generation placental arteries were obtained as well as samples of umbilical vein and fetal cotyledon, which is the fetal portion of the ovine placenta. The entire abdominal aorta, carotid arteries, and 1st-4th generation mesenteric arteries were collected from all fetal and postnatal animals. Arteries were removed with minimal trauma, residual blood was expressed, and arterial segments were placed in 20 mL of icecold 8 mmol/L PBS (pH 7.4) containing 100 µL of the protease inhibitor phenylmethylsulfonyl fluoride (0.5 mmol/L). Fat, connective tissue, and adventitia were dissected from the vessels, and the endothelium was removed with a cotton-tipped applicator. The remaining medial layer was the source of VSM used in the plasma membrane preparations. These studies were approved by the Institutional Review Board for Animal Research.

Membrane preparation. The VSM from all study arteries (∼1.0 g) was minced in 20 mL of fresh 8 mmol/L PBS, to which 100 µL of each of the protease inhibitors phenylmethylsulfonyl fluoride, leupeptin (5 µg/mL), and aprotinin (5 µg/mL) were added. Tissues that were not assayed on the day of collection were rapidly frozen in liquid nitrogen after mincing and stored at -80°C until the time of assay. We have previously demonstrated the stability of tissues minced in the presence of protease inhibitors and stored at -80°C for up to 6 mo (24). The minced tissues were transferred into 20 mL of ice-cold 0.25 mol/L sucrose and 25 mmol/L Tris buffer, pH 7.4. Plasma membranes were prepared at 4°C using methods previously described^{(12,19,20,24)}. Briefly, tissues were homogenized three times for 10 s using a Polytron (PT-20 probe, Brinkman Instruments, Westbury, NJ) and allowing the probe to cool between homogenizations. An additional 100 µL of each protease inhibitor was then added to the homogenate, which was centrifuged at 10 000 × g for 20 min at 4°C. The supernatant was removed, filtered through two layers of gauze, and ultracentrifuged at 45 000 × g for 30 min at 4°C. The resulting pellets were resuspended in 5 mL of 0.6 mol/L KCl, 30 mmol/L histidine buffer (pH 7.0) to solubilize actin and myosin and were re-ultracentrifuged at 45 000 × g for 30 min at 4°C. The final pellets were resuspended in 25 mmol/L Tris buffer (pH 7.4) containing 10 mmol/L MgCl₂, 10 µg/mL bacitracin, and 0.2% BSA.

Radioligand binding studies and ATR subtype determination. We determined the ATR subtypes in each of the arteries studied using methods previously described^{(12,19,20,24)}. ATR binding and subtype assays were performed with 100 µL of membrane preparation in a total volume of 150 µL of the 25 mmol/L Tris buffer with 0.2% BSA. Specific binding of ¹²⁵I-ANG II to plasma membranes reached equilibrium by 60-90 min, consistent with our previous reports^{(12,18,19,23)}; therefore, 90-min incubations were used in all binding studies. Incubations were performed at 18°C and were terminated by rapid addition of 4 mL of ice-cold 25 mmol/L Tris buffer with 0.2% BSA. Bound and free ligand were separated by filtration through Whatman GF/C filters (Whatman International Ltd, Maidstone, England) under vacuum followed by three rinses of the filters with the buffer. After the filters were dry, the radioactivity was measured with a Packard Scintillation Counter (Packard Instruments, Downers Grove, IL) with an efficiency of 83% for ¹²⁵I.

In competitive radioligand binding studies, tyrosyl ¹²⁵I-[5-L-isoleucine]ANG II (¹²⁵I-ANG II, 2200 Ci/mmol; NEN, Boston) was added in concentrations from 0.3 to 0.4 nmol/L. Unlabeled ANG II was added in increasing concentrations ranging from 10^-11 to 10^-8 mol/L, and the binding characteristics (receptor density and affinity) were determined from analysis of the displacement of labeled ligand. Experiments were performed in duplicate at each concentration used.

ATR subtype determinations were performed with 10^-11 to 10^-6 mol/L of the peptide ligand [Sar¹,Ile⁸]ANG II, which displaces both subtypes equally, and 10^-11 to 10^-5 mol/L of the nonpeptide ligands losartan (kindly provided by DuPont Merck, Wilmington, DE), which is specific for the AT₁R, and PD123319 (kindly provided by Parke-Davis Pharmaceutical, Ann Arbor, MI), which is specific for the AT₂R subtype. Specific binding was calculated as the difference between total ¹²⁵I-ANG II binding and nonspecific binding measured in the presence of 10^-5 mol/L [Sar¹,Ile⁸]ANG II. The percentages of the two receptor subtypes were extrapolated from the percentage of inhibition observed when the specific displacement curves were compared^(19,20,24). The percent concentration of receptor subtypes also was determined by subtracting the 10^-6 mol/L binding value of the respective subtype-selective antagonist from specific binding^(19,20,24). Additional plasma membrane preparations were made from selected arteries and preincubated with either 10^-6 mol/L losartan or PD123319. This concentration is sufficient to saturate the respective subtype without affecting the alternate subtype^(19,20,24). With the predominant ATR subtype blocked, competitive binding studies were performed to confirm the presence of the alternative population of ATR subtype. Displacement curves for each ATR ligand and their respective IC₅₀, as well as the percentages of ATR subtypes, were calculated from the specific binding data using a modification of the computer program LIGAND adapted for microcomputers by McPherson (Elsevier BIOSOFT, Cambridge, UK).

Effects of GTP_γS on ATR binding. To determine whether the ATR identified in VSM of fetal animals are coupled to G-proteins, additional plasma membrane preparations were made from selected fetal and postnatal arteries, i.e. internal and external portions of the umbilical arteries, aorta, and mesenteric artery. Plasma membranes were preincubated with 10^-5 mol/L GTP_γS, an analogue to GTP, and radioligand binding assays were then performed to confirm sensitivity or a lack thereof to GTP_γS.

Emulsion autoradiography. ATR localization within fetal and adult VSM arteries was determined using methods previously described^(16,24). Briefly, frozen sections (6-8 µm thick) of fetal carotid (n = 3; 127-139 d gestation) and femoral (n = 1; 139 d gestation) arteries and adult carotid artery from a pregnant ewe (n = 1; 145 d gestation, AT₁R control vessel) were cut on a microtome cryostat (model OTF 5030; Hacker-Bright Instruments Co., Fairfield, NJ) at -20°C, thaw-mounted onto positively charged slides, dried in vacuo overnight at -5°C over silica gel, and stored with silica gel in sealed Bakelite boxes at -80°C. Immediately before use, the sections were dried for an additional 2-4 h in vacuo at room temperature. The endogenously bound ligand was removed by preincubation with 10 mmol/L phosphate-buffed saline (pH 7.4) containing 150 mmol/L NaCl, 5 mmol/L Na₂EDTA, 0.3 mmol/L bacitracin, and 0.2% BSA for 20 min at room temperature. This was replaced with the same buffer containing 300 pmol/L ¹²⁵I-[Sar¹,Ile⁸]ANG II, and the sections were incubated for 90 min at 18°C in a humidified chamber. Sections were washed four times for 60 s each in 50 mmol/L Tris (pH 7.4) at 0°C and dried in a stream of cool air. The receptor-ANG II complexes were fixed by exposure to paraformaldehyde vapors at 80°C for 2 h in a closed chamber. The vapors were evacuated from the chamber, and the sections were left under a drafted hood overnight. Sections then underwent three washes in phosphate-buffed saline to ensure removal of residual vapors. Serial sections were subsequently coated with NTB2 photographic emulsion (Eastman Kodak Co., Rochester, NY) and air dried lying flat. The sections were exposed for 14 d at 4°C, stained with hematoxylin-eosin, and examined by bright/dark-field microscopy using an Olympus BX50 microscope and Olympus OM/SC35 photomicrographic system (Olympus Optical Co., LTD., Lake Success, NY). Routine controls to evaluate the emulsion for altered chemography included coating a blank slide (to detect background levels of silver grains in the emulsion) and coating sections that has no radioligand in the receptor binding buffer. In addition, specificity of binding was evaluated with 10^-5 mol/L [Sar¹,Ile⁸]ANG II. Receptor subtypes were identified by inhibition of radioligand binding with the AT₁R ligand losartan (10^-5 mol/L) and the AT₂R ligand PD123319 (10^-6 mol/L). Each experiment was repeated a minimum of three times with highly reproducible results.

Statistical methods. Differences in ATR binding density and affinity within groups were examined by Welch's approximation to one-way analysis of variance (ANOVA). Data are presented as the mean ± one SEM. Data were considered significant at p < 0.05 unless otherwise stated.

RESULTS

ATR binding characteristics. Binding characteristics were not determined for placental arteries as we previously reported these observations throughout the last third of ovine gestation⁽¹²⁾. As with the placental arteries, ligand analysis of the displacement of ¹²⁵I-ANG II by unlabeled ANG II demonstrated a one-site model for receptors in umbilical arteries (p < 0.05, r > 0.93) and systemic arteries from fetal (p < 0.05, r > 0.94) and postnatal sheep at 1 wk (p < 0.05, r > 0.94), 2 wk (p < 0.05, r > 0.93), 1 mo (p < 0.05, r > 0.92), and 3 mo (p < 0.05, r > 0.94). Thus, all VSM examined expressed a single class of high affinity saturable binding sites. Specific binding at the dissociation constant (k_d; nmol/L) for ATR in all of the umbilicoplacental arteries exceeded 83%. In systemic arteries, this was 78%, >75%, >74%, >74%, and >76%, respectively, in the five study groups noted above. The vast majority of nonspecific binding was accounted for by binding of the radioligand to the GF/C filters. This was not altered by prewetting the filters with albumin-containing buffer. Hill coefficients for umbilicoplacental arteries ranged from 0.99 to 1.01, whereas values for systemic arteries ranged from 0.94 to 1.28, 0.91 to 1.12, 0.98 to 1.01, 0.98 to 1.07, and 0.97 to 1.02, respectively, indicating that cooperatively was not involved in the binding of ¹²⁵I-ANG II to the VSM receptors in the vessels studied⁽⁴¹⁾. The ability of several ANG II analogues and the unrelated peptide hormone AVP to displace ¹²⁵I-ANG II binding from plasma membrane fractions prepared from placental arteries and fetal aorta has been reported by us⁽¹²⁾ and is characteristic of binding to VSM ATR. This was not different in the arteries included in the present investigation (data not shown).

The average total receptor binding density (B_max) in VSM from segments of the internal and external umbilical artery and umbilical vein within the cord from near-term fetal sheep ranged from 112 to 183 fmol/mg protein and did not differ significantly (Table 1; p = 0.5). Whereas the k_d for the external umbilical artery and vein was similar, values were twice that in the internal umbilical artery (p < 0.05) and 2- to 5-fold greater (p < 0.001) than that in all fetal systemic arteries studied (Table 1).

Table 1 Total angiotensin II receptor binding density (fmol/mg protein) and affinity (nmol/L) in umbilical and systemic vascular smooth muscle during ovine development

The average total ATR binding density in fetal systemic arteries ranged from 110 to 195 fmol/mg protein (Table 1), and values did not differ between vessels, including those determined for the umbilical vasculature (ANOVA, p = 0.48). Similar comparisons were made at each postnatal time period, and except for vessels obtained 1 mo postnatal, there were no significant differences between arteries within a time period (ANOVA, p > 0.40). At 1 mo postnatal, binding density in aorta exceeded that in mesenteric and carotid arteries (p < 0.05), which did not differ. There also were no significant changes in B_max across development, i.e. from fetal to >3 mo postnatal (ANOVA, p > 0.1), for any vessel studied. ATR binding affinity in systemic arteries followed a pattern identical to that for B_max. That is, the k_d did not differ (p > 0.5) within developmental periods except at 1 mo and was unchanged across the period of development studied (ANOVA, p > 0.1). At 1 mo, the k_d for aorta, like B_max, exceeded that for both carotid and mesenteric arteries (p < 0.001; Table 1).

ATR subtype expression. The distribution of ATR subtypes was investigated in all fetal and postnatal arteries by determining the inhibition of ¹²⁵I-ANG II binding by ATR subtype-specific ligands as described under "Methods." We used ligand analysis of the displacement of ¹²⁵I-ANG II by [Sar¹,Ile⁸]ANG II, which has equal affinity for both ATR subtypes, to determine 100% binding of VSM receptor subtypes.

Receptor subtype expression across the umbilicoplacental vascular bed was determined by examining the 1) internal or intra-abdominal portion of the umbilical artery, 2) external portion of the umbilical artery, which lies within the umbilical cord, 3) primary placental artery as it branches directly from the umbilical artery, 4) 2nd-4th generation placental arteries, which extend up to the placentome, 5) fetal placentome, and 6) the portion of the umbilical vein within the cord. AT₁R blockade with losartan failed to inhibit ¹²⁵I-ANG II binding (IC₅₀ > 100 000 nmol/L) in the internal umbilical artery (Table 2; Fig. 1A), the 2nd-4th generation placental arteries, and the fetal component of the placentome (Table 2). In contrast, the AT₁R ligand completely inhibited ¹²⁵I-ANG II binding in segments of the umbilical artery (Fig. 1B) and vein derived from the umbilical cord (Table 2). Notably, the AT₁R and AT₂R ligands inhibited ¹²⁵I-ANG II binding equally in VSM from the primary generation of placental artery (Table 2, Fig. 1C), demonstrating a transitional zone in receptor subtype expression. To ascertain if the internal umbilical artery and 2nd-4th generation placental arteries expressed any AT₁R, additional plasma membrane preparations of internal umbilical (n = 4) and 2nd-4th generation placental (n = 3) arteries were preincubated with 10^-6 mol/L PD123319. By blocking the predominantly expressed AT₂R in these arteries, the absence or presence of a small population of AT₁R can be elucidated^19,20,24,38). Although two of the internal umbilical arteries had no evidence of AT₁R expression, displacement of ¹²⁵I-ANG II binding by losartan was seen in the remaining two arteries, which had IC₅₀ values of 123 and 225 nmol/L. The IC₅₀ values reported for each ATR-specific ligand are similar to those previously reported by us^{(12,18,19,23,37)} and others^(11,27,35). Analysis of these displacement curves suggests that the AT₁R represents ≤5% of total ATR binding in the internal umbilical artery. Examination of the 2nd-4th generation placental arteries demonstrated no displacement by losartan, suggesting suggesting the exclusive presence of AT₂R binding sites.

Table 2 Distribution of angiotensin II receptor subtype expression throughout the umbilicoplacental vascular bed of third trimester fetal sheep

Figure 1

Displacement curves for inhibitors of ¹²⁵I-ANG II binding to plasma membranes prepared from ovine umbilical and placental arteries obtained at 128-139 d gestation (values represent a mean of three experiments with each ligand, each performed in duplicate). (A) Intra-abdominal umbilical artery; (B) external umbilical artery; (C) primary fetal-placental artery. The y axis represents the percent of specifically bound ¹²⁵I-ANG II.

Similar competitive inhibition binding studies were performed using plasma membranes derived from systemic arteries of fetal and postnatal sheep to determine whether differences existed between vascular beds and/or across development (Table 3). In tissues from the late third trimester fetus, the AT₁R ligand failed to displace ¹²⁵I-ANG II binding in VSM from the aorta, carotid, and 1st through 4th generation mesenteric arteries, the IC₅₀ exceeding 100 000 nmol/L for each vessel. In contrast, PD123319 completely displaced ¹²⁵I-ANG II binding in all three arteries. To verify the absence of even a small population of AT₁R, we blocked the predominant AT₂R in plasma membranes prepared from additional aorta (n = 2) and carotid (n = 2) and 1st through 4th generation mesenteric (n = 2) arteries by preincubation with 10^-6 mol/L PD123319. The IC₅₀ for losartan remained > 100 000 nmol/L in all arteries, demonstrating the absence of AT₁R expression in large systemic arteries as well as smaller generation mesenteric arteries. Displacement curves for the mesenteric artery, a representative muscular systemic artery, are illustrated in Figure 2A. To determine whether the absence of AT₁R recognition was because of the method of membrane preparation, we examined the discarded fractions of the membrane preparation, including the fat and adventitia, sucrose supernatant, and the potassium-histidine supernatant in additional samples of aorta (n = 1) and carotid (n = 2) arteries. Only a minuscule amount of binding was detected in the fat and adventitial component, which was the AT₂R subtype.

Table 3 Characterization of vascular smooth muscle angiotensin II receptor subtype expression in systemic arteries during ovine development

Figure 2

Displacement curves for inhibitors of ¹²⁵I-ANG II binding to ovine mesenteric artery plasma membrane fractions at 130-139 d gestation (A), 13-17 d postnatal (B), and ≥3 mo postnatal (C). Values are means of four experiments with each ligand, each performed in duplicate at each age. The y axis represents the percent of specifically bound ¹²⁵I-ANG II.

Because the developing rat seems to switch from AT₂R to AT₁R expression between 24 and 72 h following birth (16), we examined the ATR binding characteristics in systemic VSM obtained from three sheep 12-24 h after birth. There was no difference compared with term fetal animals. PD123319 completely displaced ¹²⁵I-ANG II binding in the aorta (n = 2), carotid (n = 1), and mesenteric (n = 1) arteries, IC₅₀ values of 0.5, 7.51, and 10.3 nmol/L, respectively, whereas losartan had no effect (IC₅₀ > 100 000 nmol/L). The IC₅₀ for [Sar¹,Ile⁸]ANG II was 3.7, 0.98, and 1.09 nmol/L, respectively.

Receptor subtype expression was unchanged 1 wk postnatal and resembled that observed in fetal and postnatal animals <24 h old. That is, ¹²⁵I-ANG II binding was unaffected by losartan and completely inhibited by PD123319 (Table 3). By 2 wk postnatal, both receptor subtype ligands displaced ¹²⁵I-ANG II binding in a variable manner in all three systemic arteries, demonstrating the expression of both ATR subtypes (Table 3, Figure 2B). At 1 mo postnatal, only the carotid artery continued to exhibit a mixed population of receptor subtypes as evidenced by variable displacement of ¹²⁵I-ANG II binding, whereas the aorta and mesenteric artery predominantly expressed AT₁R (Table 3). Preincubation studies performed with additional aortic membrane preparations demonstrated the presence of a small residual population of AT₂R. The transition to the AT₁R was complete at 3 mo postnatal, at which time PD123319 failed to displaced ¹²⁵I-ANG II binding in all three arteries, demonstrating an IC₅₀ >100 000 nmol/L (Table 3; Figure 2C). Preincubation studies of plasma membranes prepared from additional carotid (n = 1) and mesenteric (n = 1) arteries showed evidence of only a small residual population of AT₂R (Fig. 3).

Figure 3

Displacement curves for inhibitors of ¹²⁵I-ANG II binding to ≥3 mo postnatal ovine mesenteric artery plasma membrane fractions. Displacement by [Sar¹,Ile⁸]ANG II is shown as well as by PD123319 (PD) after preincubation with 10^-6 mol/L losartan to uncover undetected AT₂R; values represent one experiment performed in duplicate. The y axis represents the percent of specifically bound ¹²⁵I-ANG II.

To further characterize the ATR expressed in fetal and placental VSM, we determined if fetal receptor binding was coupled to G-proteins as reported in adults animals^{(11,13–15,17,27,41)}. This was accomplished by performing competitive inhibition binding studies on membrane preparations preincubated with 10^-5 mol/L GTP_γS. The presence of GTP_γS did not change the dissociation of ¹²⁵I-ANG II binding by PD123319 in the internal umbilical artery (Fig. 2A), fetal aorta, or fetal mesenteric artery (Fig. 4C). IC₅₀ values averaged 1.4, 2.1, and 1.9 nmol/L, respectively, for [Sar¹, Ile⁸]ANG II and 7.9, 6.6, and 2.7 nmol/L, respectively, for PD123319, demonstrating AT₂R insensitivity to GTP_γ. It also had no effect on binding displacement by losartan in the fetal mesenteric artery (Fig. 4C), which does not express the AT₁R. In contrast, preincubation with GTP_γS completely inhibited the dissociation of ¹²⁵I-ANG II binding by losartan in the external umbilical artery VSM (Fig. 4B), resulting in an IC₅₀ value >100 000 nmol/L. The IC₅₀ for [Sar¹,Ile⁸]ANG II averaged 2.1 nmol/L. Thus, the AT₁R, which is predominantly expressed in the external umbilical artery, is sensitive to GTP_γS, demonstrating coupling to G-proteins by this receptor.

Figure 4

Displacement curves for inhibitors of ¹²⁵I-ANG II binding to plasma membrane fractions in the presence of 10^-5 mol/L GTP_γS (values are means of two experiments with each ligand, each performed in duplicate). (A) Intra-abdominal umbilical artery from 136 d gestation, which is predominantly AT₂R; (B) external umbilical artery from 130-140 d gestation, which is only AT₁R; and (C) mesenteric artery from fetal animals, which expresses only AT₂R. The y axis represents the percent of specifically bound ¹²⁵I-ANG II.

In situ ATR localization. To further identify and localize the expression of ATR subtypes in fetal systemic arteries, we examined carotid and femoral arteries using emulsion autoradiographic visualization of ¹²⁵I-[Sar¹,Ile⁸]ANG II radioligand binding in the presence and absence of ATR ligands. A carotid artery from a pregnant ewe, representing the adult pattern of ATR expression, was studied for comparison. In the presence of ¹²⁵I-[Sar¹,Ile⁸]ANG II alone, both bright field (not shown) and dark field microscopy demonstrated highly abundant silver grains in the endothelium, elastic lamina, medial smooth muscle, and serosal surface in fetal carotid (Fig. 5A) and femoral arteries obtained at 127-137 d gestation and in one adult carotid artery. These silver grains were markedly diminished by 10^-5 mol/L unlabeled [Sar¹,Ile⁸]ANG II (Fig. 5B), demonstrating specificity of radioligand binding. Binding in the fetal carotid (Fig. 5C) and femoral arteries was unaffected by the AT₁R ligand (10^-5 mol/L). However, 10^-6 mol/L PD123319, the AT₂R ligand, virtually eliminated radioligand binding (Fig. 5D). In contrast, in the adult carotid artery, the AT₁R ligand (10^-5 mol/L) inhibited radioligand binding in the endothelium, intima, elastic lamina, and medial smooth muscle, whereas the AT₂R ligand had no effect (data not shown).

Figure 5

Dark-field photomicrographs of ¹²⁵I-[Sar¹,Ile⁸]ANG II radioligand binding visualized using emulsion autoradiography (20× magnification). Radioligand binding to ATR in cross-sections of a fetal carotid artery obtained at 127 d gestation was visualized by silver grains overlying cells expressing ATR. (A) ¹²⁵I-[Sar¹,Ile⁸]ANG II radioligand binding; (B) displacement of radioligand binding by unlabeled [Sar¹,Ile⁸]ANG II; (C) radioligand binding in the presence of 10^-5 mol/L losartan; (D) radioligand binding in the presence of 10^-6 mol/L PD123319. The m designates medial smooth muscle, whereas l designates the vessel lumen.

DISCUSSION

The components of the RAS are present in the placental unit and elsewhere in the developing fetus and newborn^{(1–13,15–18,21,25,35–39)}. However, it remains unclear when in development RAS function resembles that in adult animals, how it contributes to circulatory homeostasis, and what other roles it may play during ontogeny. Although bioactive ANG II is present throughout much of development^(1,37,42), receptors capable of mediating its effects must also be available. At least two receptor subtypes exist, AT₁R and AT₂R^(11,14). In the human, receptor mRNA is present as early as stage 16 in the embryo⁽¹⁸⁾. In the mouse and rat, ATR mRNA is also present early in embryogenesis, but binding is not detected until >12 d postconception, with values increasing thereafter⁽⁴²⁾. These receptors were subsequently identified as being predominantly the AT₂R subtype and were replaced by AT₁R soon after birth⁽¹⁶⁾. The ovine fetus also expresses a single class of high affinity ATR in aortic and placental artery VSM during the last third of gestation, but in contrast to the rat, the binding characteristics were unchanged⁽¹²⁾, and the ATR subtype was not identified.

It is generally agreed that the AT₂R does not mediate its effects through classical signal transduction pathways, and its function and mechanism(s) of action remain controversial. Thus, its role in the developing animal is unclear. Further, its distribution in the developing cardiovascular system is not well characterized. In adult animals AT₁R predominate and account for smooth muscle contraction and other biologic effects of ANG II via G-protein coupling and activation of phospholipase-mediated phosphoinositide hydrolysis and calcium mobilization. Although the AT₂R accounts for 60-85% of binding in human and ovine myometrium, we^(19,20) and others⁽²⁷⁾ could not demonstrate a role for AT₂R-mediated contraction, suggesting that smooth muscle responses are mediated solely by the AT₁R. This raises important questions about the distribution of ATR subtypes and cardiovascular function of the AT₂R in development. In the present study, we characterized the distribution of ATR subtypes in the umbilicoplacental vasculature and specific systemic arteries of developing fetal and postnatal sheep as well as the ontogenic changes that occur during ovine development. Our findings raise provocative questions regarding the role of the RAS in cardiovascular function during development.

Dawes⁽³⁴⁾ suggested that the umbilicoplacental vascular bed might be important in regulating total fetal peripheral resistance because it accounts for ∼40% of fetal cardiac output. We^(36,38) and others^(3,39) reported that this vascular bed is extremely sensitive to infused ANG II. Adamson et al.⁽³⁹⁾ demonstrated that sensitivity to ANG II varied within the ovine umbilicoplacental vascular bed, the major component of vascular resistance occurring primarily between the umbilical artery and its major tributaries with minimal responses occurring distally in either the fetal cotyledons or umbilical vein. We now report that the AT₁R, which mediates VSM contraction in the adult, is expressed only in the umbilical artery and its primary tributaries, whereas subsequent branches and placentomes express only AT₂R. The latter is consistent with that found in the porcine placentome⁽³⁵⁾ but differs from the human, where the AT₁R is found in the villus branches of the placental arteries⁽¹³⁾. If the AT₂R does not mediate VSM contraction, as recently reported by Arens et al.⁽³²⁾ for the internal umbilical artery, our observations explain the differential sensitivity within the umbilicoplacental circulation observed by Adamson et al.⁽³⁹⁾.

Although AT₁R are expressed in umbilical vein VSM, Adamson et al.⁽³⁹⁾ did not observe a rise in umbilical venous resistance in intact animals. This likely reflects the capacity of the ovine placenta to clear 80-90% of plasma ANG II in one passage⁽³⁷⁾. Thus, little or no ANG II would be expected to reach the umbilical vein. Alternatively, the AT₁R in umbilical vein VSM may not be coupled or a potent ANG II antagonist is locally produced. There are no data available to suggest that the umbilical vein AT₁R cannot transduce a contraction response, and umbilicoplacental responses to ANG II infusions are completely inhibited by local infusion of an AT₁R-specific antagonist⁽³⁸⁾. The production of a local antagonist to ANG II also is unlikely as Yoshimura et al.⁽⁴³⁾ reported that the ovine placental vein does not produce substantial amounts of prostacyclin, and ANG II does not induce additional prostanoid synthesis as seen in placental arteries. Thus, the change in ATR subtype distribution across the fetal umbilical vascular bed plus placental clearance of ANG II are sufficient to explain the differential sensitivity to ANG II described by Adamson et al.⁽³⁹⁾. The predominance of AT₁R expression by the umbilical artery also would account for the greater sensitivity to ANG II by the umbilical circulation than the maternal uteroplacental vasculature⁽³⁷⁾, which in women and sheep predominantly expresses AT₂R^(19,24).

Ovine umbilical vascular responses to infused ANG II exceed those elicited by either the fetal peripheral vasculature as a whole^(3,36,37) or by specific vascular beds^(3,38). The mechanism for this differential responsiveness is unknown. In studies comparing the hindlimb and umbilical vasculature, Kaiser et al.⁽³⁸⁾ observed that the hindlimb was less sensitive to local ANG II infusions. Furthermore, there was evidence of autoregulation in the hindlimb, suggesting peripheral responses to ANG II may be the result of myogenic responses. Support for this is obtained from the studies of Iwamoto and Rudolph⁽³⁾ who reported that, relative to the umbilical circulation, peripheral organ and tissue blood flow was maintained during ANG II-mediated increases in arterial pressure. We now provide evidence for the first time that the AT₂R seems to be the predominant subtype in fetal systemic VSM, confirming observations in the ovine fetal femoral artery⁽³⁸⁾ and fetal rat aorta⁽⁴⁰⁾. This could contribute to or explain differences between fetal umbilical and systemic responses to infused ANG II. Because we predominantly examined binding in large systemic arteries, it is possible smaller resistance arteries contain AT₁R. This, however, was not the case in the mesenteric artery, which did not demonstrate AT₁R binding in 1st through 4th generation branches before and after preincubation with the AT₂R antagonist. Alternatively, the decreased peripheral responsiveness to infused ANG II may be the result of local synthesis of an antagonist, such as prostacyclin, as in the ovine umbilical artery^(43,44) and uterine artery of pregnant ewes⁽⁴⁵⁾. Although there is substantial basal prostacyclin and PGE₂ synthesis by fetal mesenteric arteries⁽⁴³⁾, this is unaffected by ANG II or α-agonist, and it is unclear if these prostanoids modulate peripheral ANG II responsiveness. Another possibility is that contractile protein expression in fetal systemic VSM is immature as reflected by an abundance of nonmuscle myosin heavy chain-B and decreased contents of actin, myosin, and smooth muscle myosin heavy chain isoforms compared with the adult^(31–33). When umbilical and fetal systemic artery protein expression and function were compared with adult vessels, only the umbilical artery resembled adult arteries⁽³²⁾. Thus, there are several potential explanations for the differences in fetal umbilical and systemic responses to ANG II that will need to be investigated.

AT₁R and AT₂R expression is developmentally regulated^{(11,15–18,21,22)}. However, studies to date primarily describe changes in ATR subtype mRNA and not receptor binding or protein. Thus, it is difficult to establish firm conclusions about the presence of functional receptors and their role in modulating responses by the RAS in development. For example, the fetal rat aorta contains substantial AT₁R mRNA^(21,22), but >80% of binding is the result of AT₂R⁽⁴⁰⁾. Because much of our understanding of the role of ANG II in cardiovascular development has been obtained in fetal and postnatal sheep^{(1–5,7–10,12,30,32,34,36–39,43,44)}, we characterized ATR expression and subtype distribution in systemic VSM during ovine development. Whereas binding density and affinity were unchanged, subtype expression was predominantly AT₂R until 2 wk postnatal, when there was a transition to AT₁R expression, which was generally completed by 1 mo. Similar changes occur in rat aorta, where the transition begins 2 wk postnatal and the AT₁R accounts for >70% of binding by 8 wk⁽⁴¹⁾. Although AT₁R were not evident in VSM until 2 wk postnatal, ANG II increases arterial pressure in newborn sheep at 2-3 h postnatal and throughout the next 8 wk^{(9,10,46–48)}, and inhibition of angiotensin-converting enzyme decreases mean arterial pressure (MAP) in the first 2 h after birth⁽⁹⁾. Furthermore, carotid arteries, which have no AT₁R binding until 2 wk postnatal, demonstrate dose-dependent responses to ANG II at 1 wk postnatal⁽⁴⁹⁾. However, large doses of ANG II were needed, and responses were ∼16% of adult arteries⁽⁴⁹⁾. Importantly, responses increased ∼3-fold at 4 wk postnatal consistent with the observed change from AT₂R to AT₁R expression.

If the AT₂R is the predominant receptor in peripheral VSM prenatally and postnatally, it is unclear how exogenous ANG II elicits pressor responses. In the fetus, this may occur by constricting the umbilicoplacental vascular bed, which accounts for 40% of cardiac output^(34,38). Another possibility is that the AT₂R is capable of transducing contraction responses in the fetus and newborn. This is unlikely, as Kaiser et at.⁽³⁸⁾ did not see a pressor response to ANG II during AT₁R blockade or a change in pressor responses after infusion of a 10-fold greater dose of the AT₂R antagonist. Further, ANG II mediates VSM contractions via G-protein-coupled mechanisms, and we now show that GTP_γS does not alter AT₂R binding in fetal VSM, providing additional evidence that this receptor is not involved. As noted earlier, resistance arteries may express the AT₁R, but data from the mesenteric circulation do not seem to support this. Alternatively, ANG II may mediate its systemic responses via another mechanism(s). In adults, ANG II increases adrenal catecholamine release⁽⁵⁰⁾, which seems to explain its vasoconstrictive effects on the uteroplacental vasculature, another AT₂R predominant vascular bed⁽⁵¹⁾, and increases sympathetic outflow by a central mechanism⁽⁵²⁾. Tsutsami et al.^(25,53) reported that ATR binding and subtype expression in the fetal and neonatal rat brain are developmentally regulated. Importantly, there was AT₁R binding in areas accessible to blood-born ANG II that are associated with cardiovascular control. More recently, Segar et al.⁽¹⁰⁾ demonstrated that the AT₁R antagonist losartan infused either systemically or into the lateral ventricle of newborn sheep decreased arterial pressure similarly and reset heart rate and renal sympathetic nerve activity baroreflexes. These observations and our data on peripheral VSM ATR subtype expression suggest that pressor responses to infused ANG II may be the result of central effects, demonstrating the importance of the sympathetic nervous system in cardiovascular control before and after birth^(9,10,54). Finally, ANG II also modulates endothelin release⁽⁵⁵⁾, but this has not been examined in the developing animal.

We have described for the first time the binding characteristics of the ATR in fetal and postnatal VSM and the distribution of ATR subtypes in umbilicoplacental and systemic VSM. We have shown that the AT₂R seems to be the predominant subtype in systemic arteries before 2 wk postnatal and that only the umbilical artery and its primary tributaries express AT₁R. These data, therefore, explain the differential sensitivity to ANG II within the umbilical circulation observed by Adamson et al.⁽³⁹⁾ and support not only the thesis of Dawes⁽³⁴⁾ some 30 y ago, regarding its role in modulating peripheral resistance, but also more recent observations by Kaiser et al.⁽³⁸⁾ that ANG II-induced increases in fetal arterial pressure may primarily reflect increases in umbilical resistance. These data, however, raise important questions regarding the role of ANG II in modulating peripheral resistance and the mechanism(s) whereby it increases arterial pressure after birth. The present data support the thesis that ANG II mediates its pressor effects in the fetus and neonate by releasing another mediator or by acting on cardiovascular centers of the brain, thereby supporting recent observations by Segar et al.⁽¹⁰⁾. The role of the VSM AT₂R remains unclear. Finally, the present receptor data parallel changes in VSM protein expression seen by Arens et al.⁽³²⁾, suggesting receptor subtype expression may be associated with or parallel vascular growth and remodeling during development. This will require future studies directed toward understanding smooth muscle growth and differentiation.