INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

DESPITE THE FACT THAT IN RODENTS catecholamines and their adrenergic receptors (AR) are present as early as embryonic day 8 and 9 (E8-9) (4, 33), pharmacological experiments suggested that -AR activation by norepinephrine and epinephrine was not essential in the fetus. Pregnant rat dams treated with a nonselective -AR antagonist had pups with modest developmental deficiencies that resolved soon after birth. In addition, -AR blockade did not affect litter size, birth weight, or the number of stillbirths or resorptions (20). Furthermore, direct application of a -AR antagonist to E10-12.5 rat fetuses in culture failed to decrease cardiovascular function (27). These findings, along with the fact that fetal catecholamine neurons lack efferent and afferent connections and mature synaptic machinery (31), supported the view that catecholamines were not necessary during development.

It was therefore a surprise to learn that tyrosine hydroxylase knockout mice (Th^/) die in utero. Depending on the particular null allele and genetic background, between 75 and 100% of catecholamine-deficient animals die before birth, with the majority of lethality occurring during midgestation (E11.5-E14.5), long before synaptic transmission is established, and equivalent to the first trimester in the human fetus. Deficient mice that make it to birth exhibited bradycardia (19, 40) and some had cardiovascular abnormalities (40). Th^/ mice (19, 26, 40) that lack dopamine, norepinephrine, and epinephrine, and dopamine--hydroxylase knockout (Dbh^/) mice (33) that lack norepinephrine and epinephrine, had the same pattern of lethality, suggesting that norepinephrine and/or epinephrine mediate survival. Mutant mice lacking the homeodomain transcription factor Phox2b, which fail to develop sympathetic ganglia and therefore lack peripheral sources of norepinephrine, also exhibit the same fetal lethality as Th^/ and Dbh^/ mice (24, 25). In contrast, mice that lack dopamine, but not norepinephrine and epinephrine, are viable (39), ruling out a critical role for dopamine during fetal development. Taken together, the survival pattern of these four types of mutant mice provides support for an essential role for norepinephrine in the fetus. Although low levels of epinephrine have been detected at midgestation and are synthesized in the embryonic heart (10, 11), biosynthesis of epinephrine does not become substantial until after the peak window of death, suggesting that the essential catecholamine is norepinephrine (33). Because both norepinephrine and epinephrine act at the same receptors, these data show that activation of adrenergic receptors is essential for fetal viability.

The mechanism by which catecholamines maintain fetal survival is unknown. Although norepinephrine and epinephrine could serve a trophic role during development (40) (analogous to the role played by another biogenic amine, serotonin, in cardiac development) (38), we investigated whether norepinephrine and epinephrine might serve an acute physiological function in the fetus. In both Th^/ and Dbh^/ mice, the severity of morphological abnormalities varied from animal to animal, suggesting that the primary defect might be physiological rather than anatomic in nature. In this work, we hypothesize that fetuses release catecholamines to sustain cardiovascular function in response to stress in utero, suggesting that the physiological role for catecholamines in the fetus is analogous to their role in postnatal animals.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Drugs. Bupranolol was a gift of Schwarz-Mann (Liestal, Switzerland). Other drugs and chemicals were purchased from Sigma-Aldrich (St. Louis, MO).

Animals. For experiments examining the in vivo rescue of Th^/ pups to birth, pups were genotyped on the day of birth from matings between albino ICR Th^+/ mice, in which the Th null allele had been back-crossed for six generations onto the ICR strain (26). Albino mice were used in these experiments to avoid small amounts of catecholamines, synthesized via tyrosinase, that are found in pigmented Th^/ fetuses (26). Fetal culture experiments used pigmented wild-type C57bl/6J fetuses (Harlan; Indianapolis, ID), albino ICR fetuses (Taconic, Germantown, NY), and fetuses from matings between Th^+/ mice in which the Th null allele had been back-crossed onto C57bl/6J for six generations. Similar results were obtained from albino ICR Th^+/ matings (data not shown). The presence of a vaginal plug was taken as E0.5 and all fetuses were used at E12.5. Animals were genotyped with the use of a PCR strategy that simultaneously identified the Th wild-type and null alleles. This strategy combined a common forward primer in Th exon 6 (5'-GGGTGAGCCAATTCCCCAC-3') with a reverse primer in Th exon 8 (5'-TGAGTGCATAGGTGAGGAGG-3') and a reverse primer in the neomycin gene from the null allele (5'-CATAGCCGAATAGCCTCTCC-3'). Animals were used in accordance with approved protocols under Institutional Animal Care and Use Committee review.

Rescue of Th^/ pups via maternal drinking water. Multiparous Th^+/ ICR females were mated with Th^+/ ICR males and treated with ascorbic acid at 2.5 mg/ml, isoproterenol HCl (1-100 µg/ml), and/or l-phenylephrine HCl (10 µg/ml) in ascorbic acid from E8.5 until parturition. Drugs were administered in the drinking water, which was changed daily and protected from light.

Tissue preparation for histological examination. E13.5 fetuses were immersion fixed in 10% buffered formalin (Baxter Diagnostic; Deerfield, IL) and embedded in paraffin. The sections (15 µm) were stained with hematoxylin and eosin.

Fetal culture. Fetal culture was adapted from previously published protocols. Pregnant female mice were anesthetized with isoflurane (IsoFlo, Abbott Laboratories, North Chicago, IL) and euthanized by cervical dislocation. Both uterine horns were removed into a petri dish containing room temperature W3 buffer composed of (in mM) 120 NaCl, 5 KCl, 1 NaH₂PO₄, 20 HEPES, and 20 glucose (pH 7.3) that was supplemented with 10 mM MgSO₄, and the fetal masses were gently removed. The fetuses, with their visceral yolk sacs intact, were placed in fresh W3 medium supplemented with 1 mM MgSO₄ and 2 mM CaCl₂, which was the same medium used for culturing. The fetuses were freed of their surrounding membranes but left attached to the yolk sac via the vitelline stalk. Each litter was divided between control and drug treatment groups. Three to four fetuses per group were placed in a 25-ml Falcon T-flask and partially submerged in 4 ml of prewarmed W3 buffer preequilibrated with 95% O₂-5% CO₂. Flasks were then placed on a heated rocker platform that maintained the buffer at 37°C. Flasks were rocked such that the buffer washed over the fetuses 12 times/min without submerging them completely. A constant infusion at 20 ml/min of either 95% O₂ to create a normoxic condition or 55% O₂ to create a hypoxic condition was bubbled into the buffer via a gas line embedded in the lid of the flask.

Heart rate determination in vitro. Heart rate for each fetus was counted in the culture flask under a dissecting microscope for 30 s at 10-min intervals. Culture flasks were kept on a heated stage at all times to maintain buffer temperature at 37°C, because heart rate was very dependent on incubation temperature. Two control counts at 95% O₂-5% CO₂ were taken before the commencement of hypoxia or addition of drugs into the media. Hypoxia was induced by infusing reduced 55% O₂-5% CO₂-balance N₂. Drugs were added directly into the buffer surrounding the fetuses. Except where indicated, antagonists were used at 10 µM: phentolamine (9, 14, 15) should block all subtypes of -ARs, bupranolol (1, 12, 18, 36) should block all subtypes of -ARs, and atropine should block all five classes of cholinergic muscarinic receptors (2).

Statistical analysis. Statistical analysis for fetus culture experiments was performed with Statview software (SAS Institute, Cary, NC) using one-way ANOVA, followed by Fisher's protected least-significant difference post hoc test where appropriate. Heart rate in each treatment group was averaged and the means ± SE calculated. Statistical significance for Figs. 2C and 5 was determined by comparing the averaged heart rate at each time point after the addition of drugs, with the average heart rate of sibling fetuses without drugs. Data from C57bl/6J and ICR fetuses were statistically identical and were pooled for Figs. 2C and 5. Statistical significance for Fig. 4 was determined by comparing Th^/ fetuses with Th^+/ sibling fetuses in the same flasks. For Fig. 3B, fetal heart rate was measured for 60 min during exposure to various concentrations of O₂ and the average of the 40-, 50-, and 60-min heart rate was calculated. This value was then expressed as a percentage of the prehypoxic heart rate. Statistical analysis for Fig. 6 was performed with the use of ² analysis.

Doppler echocardiography. Doppler echocardiography (Sonos 500, Hewlett-Packard Instruments; Palo Alto, CA) was performed essentially as described by Gui et al. (13) with the use of a 5- to 12-MHz probe and a commercial standoff. The pregnant dam was anesthetized with pentobarbital sodium (50 mg/kg Nembutal; Abbott Laboratories). Abdominal fur was removed with depilatory cream.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Catecholamine-deficient fetuses show cardiovascular abnormalities. Th^/ fetuses are deficient in all catecholamines and die in utero starting as early as E9.5 (28). At E12.5-E13.5 about one-half of the Th^/ fetuses show markedly dilated blood vessels and pooling of the blood in the liver, heart (compare Fig. 1, A and C, with Fig. 1, B and D), and vasculature (arrows, Fig. 1B). In the heart, the ventricular septum and walls are thinner and show greater cellular disorganization in Th^/ fetuses (compare Fig. 1, E and F). Atria in the mutant animals are often enlarged with thinner walls compared with Th^+/+ or Th^+/ siblings (compare Fig. 1, G and H).

View larger version (96K): [in this window] [in a new window]   Fig. 1.   Tyrosine hydroxylase knockout (Th/) fetuses show cardiovascular abnormalities. Microscopic structure of Th+/ (A, C, E, and G) and Th/ (B, D, F, and H) fetuses and fetal mouse hearts at embryonic day 13.5 (E13.5). A and B: sagittal sections through E13.5 Th+/ and Th/ mouse fetus showing accumulation of blood in liver (Li), heart (H), lungs (L), and large vessels (arrows) of the Th/ fetus. Th+/ fetuses had the same phenotype as Th+/+ fetuses. Scale bar is 1 mm. C-H: higher power images of fetal hearts. Th/ fetal hearts show accumulation of blood in the atria (a) and ventricles (v) compared with Th+/ siblings (compare D with C). In addition, Th/ fetal hearts show marked dilatation of the atria and ventricles (H and F). Under higher magnification (H), Th/ fetal hearts show a reduction in the number of cell layers in the atria (arrowheads). Scale bar: C and D = 250 µM; E and F = 200 µM; G and H = 100 µM. Tissue sections were stained with hematoxylin and eosin.

AR blockade has no effect on wild-type fetal heart rate under control conditions in culture. To examine the role of norepinephrine in fetal cardiovascular function, we cultured wild-type E12.5 fetuses in the presence of various adrenoreceptor antagonists to block the action of endogenous catecholamines and pharmacologically mimic a catecholamine-deficient phenotype. For these experiments, we developed an acute fetal culture assay by modifying existing techniques in two ways. First, we cultured fetuses without serum, which contains variable amounts of catecholamines. Second, we infused a constant supply of oxygen into the culture flask (Fig. 2A). Fetuses cultured under optimal oxygenation (95% O₂) maintained a steady heart rate (200-220 beats/min) that dropped <15% during 50 min of incubation (Fig. 2B). The high O₂ concentration is necessary (and considered standard) in this type of culture because the fetuses are oxygenated by passive diffusion rather than by placental circulation. Whereas younger fetuses can be cultured at ambient O₂, E11-12 rodent fetuses require 95% O₂ for optimal development in long-term (1-2 days) cultures (7, 22, 23). Rodent fetuses older than E12 require hyperbaric O₂ to survive and develop on schedule (22).

View larger version (38K): [in this window] [in a new window]   Fig. 2.   Adrenergic blockade under normoxic conditions does not affect heart rate. A: E12.5 wild-type mouse fetus in culture demonstrating intact circulation between the fetus and placenta via the umbilical vessels. Individual fetuses are gently removed from the uterus and freed from their surrounding membranes (visceral and parietal yolk sacs). The fetus is transparent at this stage of development and the four chambers of the heart can be seen (atria and ventricle) and heart rate easily monitored. B: fetuses cultured under optimal oxygenation (95% O2), maintained a steady heart rate (200-220 beats/min) that dropped <15% during 50 min of incubation. Data represent 12 experiments including 35 fetuses. C: percent change in fetal heart rate compared with sibling fetuses cultured without drug at each time point. Neither phentolamine [an -adrenergic receptor (-AR) antagonist, 10 µM], nor bupranolol (a -AR antagonist, 10 µM), or the two drugs in combination had any effect on fetal heart rate. bpm, Beats per minute. Drugs added at time 0 (arrow). Data represent 14 experiments using 103 fetuses. Data presented as mean ± SE.

Hypoxia reduces fetal heart rate in vitro and in vivo. Because there was no effect of adrenergic blockade under control conditions, we asked whether the requirement for catecholamines would be unmasked by stress, analogous to the requirement for adrenergic activation in the mature animal. The most likely stressor in utero is hypoxia because periodic contractions of the uterus occur during gestation and transiently reduce fetal oxygenation (PO₂) (37). We first assessed whether midgestation fetuses are sensitive to hypoxia by culturing wild-type fetuses under reduced O₂ (55%). Hypoxia decreased heart rate to 65-70% of control prehypoxic levels (Fig. 3A). This level was maintained unless fetuses were reoxygenated, in which case heart rate was restored to 90% of prehypoxic levels (Fig. 3A). Fetal heart rate decreased in a stepwise manner with decreasing O₂ content; 70% O₂ caused a 25% drop in heart rate, whereas extreme hypoxia (20% O₂) caused fetal death within a few minutes (Fig. 3B).

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Fetal heart rate drops in response to hypoxia. A: solid line is the heart rate of wild-type mouse fetuses cultured for 10 min at 95% O2, 20 min at 55% O2, then returned to 95% O2. The dotted line represents heart rate of fetuses cultured for 10 min at 95% O2, then switched and maintained at 55% O2. Data were collected from 2 experiments using 15 fetuses. B: percent change in heart rate from initial prehypoxic values after culture of fetuses at various O2 concentrations. Data were collected from 82 experiments using 300 fetuses. Data are the average of the last three time points at the indicated level of O2. **P < 0.005 compared with prehypoxic values. C: time course of fetal heart rate in utero during maternal hypoxia as measured by Doppler echocardiography. Data collected from 3 experiments using 6 fetuses. Data are mean ± SE.

Catecholamine-deficient fetuses cannot maintain heart rate during hypoxia. If catecholamines regulate heart rate in response to hypoxic stress, catecholamine-deficient fetuses should show a greater decrease in heart rate than Th^+/+ fetuses. To test this, Th^+/ mice were mated and the fetuses cultured under hypoxic conditions without drugs. Fetal heart rate was monitored for 50 min, and, at the termination of the experiment, fetuses were collected and genotyped. Th^+/ fetuses slowed their heart rate similarly to wild-type sibling fetuses in response to hypoxia (Fig. 4A). In the same flasks, however, the heart rate of Th^/ fetuses was one-half that of Th^+/+ or Th^+/ siblings, dropping to 25-35% of prehypoxic levels, which occurred primarily at later time points (30, 40, and 50 min) (Fig. 4B). The decrease in the heart rate of Th^/ fetuses was reversed by isoproterenol, a -AR agonist, which restored heart rate to ~80% of prehypoxic levels (Fig. 4B). Hence, the extreme bradycardia induced in catecholamine-deficient fetuses is due to the lack of norepinephrine rather than to an anatomic or physiological inability of Th^/ fetuses to respond to -AR activation. Th^/ fetuses in the same flasks restored their heart rate to the same level as Th^+/ siblings (Fig. 4B), demonstrating that the difference between Th^/ and Th^+/ fetuses is eliminated in the presence of a -AR agonist. Even while under 95% O₂, Th^/ fetuses developed a slower heart rate than their Th^+/+ and Th^+/ siblings (Fig. 4C). However, the drop in normoxic Th^/ fetuses (33% reduction compared with Th^+/ fetuses) was less than in hypoxic Th^/ fetuses (55% reduction compared with Th^+/ fetuses), supporting the idea that catecholamines are more important in regulating chronotropy in response to hypoxia than under normoxic conditions. These data support the contention that, as in the mature animal, catecholamines limit the extent of fetal bradycardia in response to hypoxia and play a modest role in maintaining heart rate under normoxic conditions.

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Compared with Th+/+ or Th+/ siblings, Th/ fetuses exhibit greater hypoxia-induced bradycardia, which is reversed by isoproterenol. Time course of heart rate of Th+/+, Th+/, or Th/ sibling fetuses incubated with 55% O2 without drugs (A), 55% O2 with 10 µM isoproterenol (B), and 95% O2 without drugs (C). Data in A represent 12 experiments using 53 fetuses; data in B represent 5 experiments using 23 fetuses; data in C represent 12 experiments using 49 fetuses. Data presented as mean ± SE. *P  0.05, Th+/+ heart rate compared with Th/ heart rate in A and C, Th+/ heart rate compared with null heart rate in B.

-AR blockade causes significant decrease in wild-type fetal heart rate during hypoxia. To determine whether AR activation was responsible for mediating the effects of catecholamines in wild-type animals, we cultured wild-type C57bl/6J and ICR fetuses under hypoxic conditions and examined the effect of various AR antagonists on heart rate. Phentolamine (at 10 µM), an antagonist of all subtypes of -ARs, had no effect on heart rate (Fig. 5A), suggesting that -ARs are not involved in regulating fetal heart rate. In contrast, bupranolol decreased heart rate by 35% compared with sibling fetuses incubated under hypoxia alone in both strains of mice (Fig. 5A). Figure 5B shows that the IC₅₀ for bupranolol is ~10 nM, a concentration at which bupranolol is specific for -ARs. Because bradycardia is often reversed by the blockade of muscarinic acetylcholine receptors in mature animals, we blocked muscarinic receptors with 10 µM atropine (Fig. 5A). Atropine had no effect on heart rate, suggesting that cholinergic activation does not contribute to hypoxia-induced bradycardia. These data suggest that endogenous norepinephrine and epinephrine are likely to act at -ARs, rather than -ARs or muscarinic receptors, to maintain fetal heart rate during hypoxia.

View larger version (26K): [in this window] [in a new window]   Fig. 5.   -AR, but not -AR or muscarinic receptor blockade, causes a significant decrease in E12.5 wild-type C57bl/6J or ICR heart rate when combined with hypoxia. A: percent change in fetal heart rate compared with sibling fetuses cultured without drug at each time point. At time 0 (arrow), drugs were added and fetuses were shifted from 95% O2 to 55% O2. Data represent 23 experiments using 155 fetuses. Data are presented as mean ± SE. **P  0.05. B: bupranolol acts in a dose-dependent manner to reduce heart rate in wild-type E12.5 fetuses. Data represent 12 experiments using 85 fetuses. Data presented as mean ± SE.

/

Isoproterenol, a -AR agonist, rescues Th^/ mice to birth. Our data suggest that -AR activation restores heart rate in response to hypoxia in cultured catecholamine-deficient fetuses and that -AR blockade exacerbates hypoxia-induced bradycardia in wild-type animals. Are the cardiovascular effects observed in cultured fetuses related to survival in vivo? If -AR activation also sustains the survival of fetuses in utero, Th^/ mice might be rescued to birth by -AR activation. To test this, Th^+/ females were mated with Th^+/ males and treated from E8.5 to parturition with various concentrations of isoproterenol, a non-subtype-selective -AR agonist, via the drinking water. Without treatment, only 5% of the Th^/ pups survived to birth (Fig. 6). The vehicle ascorbic acid, included as an antioxidant in the isoproterenol and phenylephrine solutions, by itself increased survival from 5% to 18%. Although it is unclear how ascorbic acid mediates increased survival, it may reduce antiadrenergic effects of oxyradicals. Isoproterenol increased survival in a dose-dependent manner to 100%. The inactive (+) enantiomer did not increase survival over vehicle alone. Isoproterenol is not normally administered orally because it is rapidly degraded in the liver (8). Nevertheless, with the use of HPLC we detected isoproterenol in E12.5 fetuses of all genotypes at levels 3-40 times that of norepinephrine (data not shown), demonstrating that isoproterenol can pass the placenta and accumulate to physiologically active concentrations in the fetus.

View larger version (32K): [in this window] [in a new window]   Fig. 6.   Percentage of Th/ pups surviving to birth after maternal treatment from E8.5 to parturition with adrenergic receptor agonists in the drinking water at the indicated concentrations. n, Number of pups collected in each dose group; Iso, isoproterenol. The expected number of Th/ mice was taken to be one-third of the sum of Th+/+ plus Th+/ pups. **P < 0.005 compared with ascorbic acid vehicle by 2 analysis.

/

/

/

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Physiological role of catecholamines in fetal survival. Catecholamines are essential for fetal survival, as illustrated by the death in utero of various lines of catecholamine-deficient mice. However, the exact role that catecholamines play, as well as their mechanism of action, has not been investigated. Our finding that a specific -AR agonist, isoproterenol, increases survival of catecholamine-deficient fetuses to 100% in a dose-dependent and stereo-specific manner (Fig. 6) suggests that survival requires -AR activation. Rescued pups resemble the small number of Th^/ pups that are born without drug treatment and die by postnatal day 21 because they fail to eat or drink (39).

View larger version (27K): [in this window] [in a new window]   Fig. 7.   Hypothesis for the role of catecholamines in the developing fetus. All fetuses are subjected to transient bouts of hypoxia throughout gestation as a result of uterine contractures. In wild-type fetuses (left), hypoxia induces bradycardia and the release of norepinephrine (NE) into the circulation. NE activates -ARs on the cardiovascular system to increase beat frequency and limit the extent of hypoxia-induced bradycardia. In NE-deficient fetuses (right), no NE can be released, resulting in extreme bradycardia and death of NE-deficient mice. If NE-deficient fetuses are treated with a -AR agonist (isoproterenol), direct activation of -ARs rescues mutant fetuses to birth. In culture (dotted trace), -AR antagonists prevent -AR activation in wild-type fetuses, such that heart rate drops to a level that is insufficient to maintain adequate blood flow, leading to death. -AR blockade makes the wild-type fetuses pharmacologically null, resembling NE-deficient mice.

/

/

Clinical use of -AR antagonists during human pregnancy. -AR antagonists are commonly used as second-line antihypertensives during pregnancy. Despite the fact that most antagonists pass the placenta, their effect on the fetus in the first trimester is largely unknown. Our data in the mouse suggest that -AR blockade could be deleterious to the fetus by exacerbating ischemia during transient hypoxic periods in utero. In clinical studies (35), mothers treated with antihypertensive drugs during pregnancy had newborns that were small for gestational age. Two randomized, controlled studies (3) showed that atenolol, a ₁-AR selective antagonist, significantly reduced birth weights by 26% when administered at the end of the first trimester. If our data can be extrapolated to humans, they might explain these clinical findings by suggesting that -AR blockade is detrimental because it exacerbates transient hypoxic episodes, leading to ischemic damage.