The heart is the first organ to form and function in the developing embryo (Fishman and Chien, 1997). Formation of the heart involves differentiation of the primary heart tube into the four major cardiac chambers, the right and left ventricles and atria (Pexieder, 1975b). This process involves complex temporal and spatial patterns of gene expression and dynamic changes in cell proliferation and programmed cell death (Olson and Srivastava, 1996; Pexieder, 1975a; Zhao and Rivkees, 2000). Whereas the influence of growth factors and transcription factors on cardiac development has received considerable attention (Olson and Srivastava, 1996), little is known about the influence of hormonal agents, including neurotransmitters, on cardiac structural development.
During early cardiac development, adenosine appears to be the dominant regulator of embryonic cardiac function (Hofman et al., 1997; Rivkees, 1997). Adenosine is a nucleoside that is produced and released from cells into the extracellular space (Phillis, 1991). Levels of adenosine are dynamically regulated and rise during increased cellular activity or tissue ischemia. In the embryo and fetus, adenosine levels also increase markedly with fetal stress and hypoxia (Rivkees, 1997).
After its release, adenosine acts via specific receptors that belong to the G protein-coupled receptor superfamily (Olah and Stiles, 1995). These include A2a and A2b receptors that couple with Gs, and A1 and A3 receptors that couple to Gi (Olah and Stiles, 1995). Adenosine receptors also regulate the activity of ion channels (Cooper and Caldwell, 1990).
During development, A1ARs may be especially important. In the embryonic heart, A1ARs are expressed at very early stages of development and are among the earliest expressed G protein-coupled receptors expressed in the myocardium (Rivkees, 1995). Showing the functional significance of adenosine action in the fetal heart, beginning with the inception of spontaneous cardiac contractility, A1AR activation can slow heart rates and override the effects of adrenergic receptor activation (Hofman et al., 1997).
Whereas the effects of adenosine on embryonic cardiac function are well recognized, the potential influences of adenosine on cardiac structural development are not known. In other systems, it has been shown that adenosine can influence rates of cell division and slow rates of cell growth (Lelievre et al., 1998; Tey et al., 1992). Thus, we speculate that activation of A1ARs may also impact cardiac development. Using several complementary approaches including studies of pregnant dams, isolated embryos, and cardiac explants, we have examined the effects of A1AR activation on heart formation. We now show that A1AR activation inhibits cardiac cell proliferation and can induce cardiac hypoplasia.
Effects of A1AR Activation on Heart Development In Vivo
To investigate if A1AR activation influences embryonic cardiac development in vivo, the A1AR agonist CPA was injected into pregnant mice at various stages of gestation, ranging from E9.5 to E13.5. In studies in which pregnant dams were injected with 5 mg/kg of CPA along with tracer amounts of [3H]CCPA (2-chloro-N6- cyclopentyladenosine; 1 nM; Dupont/New England Nuclear; n = 3 dams), we found that tissue concentrations of [3H]CCPA in the embryo were about 55% of levels in the dam, showing that CPA compounds crossed the placenta to reach the fetus.
Doses of CPA ranging from 1 to 10 mg/kg were given to pregnant dams. Visible effects of treatment on the developing embryos and hearts were observed after injection of 5 or 10 mg/kg CPA doses. Thus, the lowest effective dose (5 mg/kg) was used for subsequent experiments. In the CPA-treated animals, embryos were smaller than in vehicle-treated animals (Fig. 1A). Quantitative analysis of weights and coronal-rump lengths confirmed the observations (Table 1). Hyperdermal edema, dilated thoracic cavity, and hyperemia of the skin, particularly in the interdigital tissue of the footpad, were also observed in the embryos treated with CPA at E11.5 and E13.5 (Fig. 1A).
Table 1. Effects of CPA Treatment on the Development of Embryos and the Hearts In Vivoa
P < 0.05 represents significant difference. LV, left ventricle.
Weight of embryo (g)
0.1554 ± 0.0048 (n = 9)
0.1361 ± 0.0050 (n = 10)
Length of embryo (mm)
10.39 ± 0.1389 (n = 9)
9.600 ± 0.1247 (n = 10)
Weight of heart (mg)
8.900 ± 0.6289 (n = 9)
6.480 ± 0.2913 (n = 10)
Length of LV (mm)
1.311 ± 0.0564 (n = 9)
1.120 ± 0.0327 (n = 10)
Weight of embryo (g)
0.3517 ± 0.0120 (n = 23)
0.2473 ± 0.0123 (n = 29)
Length of embryo (mm)
13.50 ± 0.1633 (n = 23)
11.00 ± 0.2046 (n = 29)
31.6667 ± 5.3831 (n = 6)
10.1667 ± 2.1200 (n = 6)
1.3333 ± 0.3333 (n = 6)
1.6667 ± 0.6667 (n = 6)
Gross anatomical examination of the hearts revealed smaller ventricles and slightly dilated auricula of the atria of CPA-treated embryos (Fig. 1B). The weights of the whole hearts and lengths of the left ventricle also showed significant differences between the vehicle and CPA-treated groups (Table 1). Histological examination of the embryos showed smaller ventricles, dilated atria, and dilated thoracic cavities in CPA-treated embryos (Fig. 1C,D) confirming the gross anatomical observations.
Effects of A1AR Activation on Heart Development In Vitro
We next assessed if the effects seen in vivo reflected direct effects by examining cultures of whole embryos. This approach was used as it provides a means to directly examine the effects of agents on intact embryos without potential confounding influences of the placenta or the dam (Sturm and Tam, 1993).
We first investigated E7–7.5 embryos, in which the heart tube has not formed. As expected, embryos cultured for 48 hr developed into E9.0-appearing embryos (n = 16). The heart tube had differentiated into discrete segments in control embryos, as normally occurs. Looping of the heart tube was nearly complete, with the primitive right and left ventricles on the right and left sides of the body, respectively (Fig. 2A). However, following CPA treatment (1 μM; n = 18), the boundary between the right ventricle and outflow tract was not formed, although the left ventricle was recognizable. Heart looping was also inhibited, with the primitive right and left ventricles located in an anterior-posterior direction, rather than on the right and left sides of the body (Fig. 2B).
Next, to examine the effects of A1AR activation on cardiac chamber development, E10.5 embryos were studied. At this stage of normal development, the major segments of the heart tube have differentiated into the ventricles, atria, and outflow tract. When embryos were cultured for 24 hr in standard conditions, they grew into E11.0-like embryos. However, when embryos were treated with CPA (1 μM) for 24 hr, they had smaller ventricles than controls on gross examination. Histological examination of the hearts revealed thinner ventricular wall and less ventricular trabeculation (Fig. 3).
Effects of A1AR Activation on Cardiac Cell Proliferation
The above observations showed that activation of A1ARs inhibits cardiac development. Next, to evaluate mechanisms of adenosine action, we examined if rates of mitosis and/or apoptosis were altered. When apoptosis was examined, we found no differences in the numbers or distribution of apoptotic cells between CPA- and vehicle-treated embryos in vivo (Table 1). However, when changes of cell proliferation were examined, a marked reduction in BrdU labeling was seen in the hearts of CPA-treated embryos in vivo (Fig. 1C,D; Table 1). The areas with decreased cell proliferation included the walls of the ventricles and atria, interventricular septum, outflow tract, and aorta (Fig. 1C,D).
Similar results were also obtained from in vitro experiments of whole embryos (Figs. 2–4), indicating that CPA treatment directly affected rates of cardiac cell proliferation, rather than rates of apoptosis. Importantly, when BrdU labeling was compared in control and CPA-treated embryos, similar levels of BrdU labeling were seen in the liver (vehicle, 66.3 ± 11.6; CPA, 60.8 ± 12.5) and facial processes (vehicle, 81.8 ± 11.5; CPA, 76.6 ± 10.2; Fig. 3), suggesting that the growth inhibitory effects of CPA were specific to the heart.
To further examine the effects of A1AR activation on cardiac cell proliferation, in vitro experiments were also performed using isolated ventricular tissue from E11.5 and E12.5 embryos cultured using the medium/air interface organ culture system. Explants were treated with CPA (10 nM, 100 nM, or 1 μM) for 4, 8, and 24 hr. Significant reduction in BrdU labeling was observed in the treatments with 1 μM CPA lasting 8 and 24 hr (Figs. 5,6). Effects on BrdU incorporation were also concentration-dependent (Fig. 6).
To examine the specificity of CPA effect, in some studies explants were pre-treated with the highly selective A1AR antagonist DPCPX (8-cyclopentyl-1,3-dipropylxanthine). DPCPX (1 μM) treatment inhibited the effects of CPA (Fig. 6), indicating that CPA was acting through A1ARs. To examine if A1AR antagonists could influence rates of cell division, we also examined cultures treated with DPCPX (1 μM) alone. Suggesting that adenosine does not tonically inhibit cardiac cell division, no differences were observed between control and DPCPX-treated cultures in the proportion of cells manifesting BrdU uptake (P > 0.05).
Cell death was also examined in the explants in the above studies. As observed in vivo, no significant differences were seen between control, CPA, or DPCPX-treated specimens in the number of TUNEL positive cells per unit area (P > 0.05).
Of the known G protein–coupled receptors (GPCRs), A1ARs are among the earliest expressed in the developing heart. When the heart is a primitive cardiac tube that has not begun to loop or beat, A1ARs are expressed in the cardiac cylinder and cardiac A1AR expression continues throughout gestation (Rivkees, 1995). Showing the importance of the adenosinergic system on cardiac structural development during the prenatal period, we find that A1AR activation inhibits cardiac cell proliferation at embryonic stages leading to cardiac hypoplasia.
Following injection of dams with an A1AR agonist during embryonic periods of cardiac differentiation and chamber remodeling, we found cardiac hypoplasia, as indicated by thin ventricular walls and reduced heart weights. We also found that the embryos were smaller after maternal CPA injections, indicating that there was intrauterine growth retardation. Because adenosine can influence maternal vascular function and possibly affect placental blood flow (Mullance and Williams, 1990; Rivkees, 1997), it is possible that the effects observed in utero reflected these factors. Thus, to overcome this potential limitation we studied isolated intact embryos.
Similar to that observed in vivo, studies of isolated embryos revealed that A1AR activation was associated with hypoplasia of the heart, showing that A1AR activation can have direct effects on the embryo. Showing the specificity of A1AR action, reduced rates of cell division were seen in the heart after CPA treatment, but not in sites that do not express A1ARs.
To examine the cellular mechanisms by which A1AR activation inhibits cardiac development, we examined rates of cell death and cell division. Our data suggest that A1AR activation involves effects on cell proliferation, rather than effects on apoptosis, as we did not observe any differences in the proportion of TUNEL-positive cells between control and CPA-treated embryos. However, using BrdU to assess rates of cell division, we observed reduced cell proliferation both in vivo and in vitro.
Another interesting finding after in vivo treatment with A1AR agonist was signs suggestive of cardiac failure in the embryos, including edema, hyperemia, and atrial dilation. It has been previously shown that A1AR activation can slow heart rate in vitro (Hofman et al., 1997). In the embryo and fetus, heart rate influences cardiac output (Heymann, 1985). Thus, A1AR activation may decrease heart rate, leading to reduced cardiac output and heart failure.
Of the known adenosine receptors, the ligand binding properties of A1ARs make this receptor subtype ideally suited to transduce the effects of adenosine in the embryonic heart. In comparison with A2a, A2b, or A3 adenosine receptors, A1ARs have the highest affinity for adenosine and can be activated with modest increases in adenosine levels (Olah and Stiles, 1995). Thus, increases in adenosine levels similar to those reported to occur in utero, are expected to activate A1ARs (Rivkees, 1997).
It is also interesting to note that relatively high concentrations of CPA were needed to inhibit cardiac cell division. These findings suggest that considerable receptor occupancy is needed to inhibit cardiac cell division, a phenomenon that has been observed in other systems (Rice et al., 1996; Srinivas et al., 1997; Van Der Graaf et al., 1997).
The clinical implications of our findings are highly significant, as we have identified a possible mechanism by which embryonic stress can result in fetal cardiac failure and lead to cardiac growth retardation. In contrast to other neuromodulators and neurotransmitters that are discretely released, adenosine is produced and released by all cells (Olah and Stiles, 1995; Rivkees, 1997). In the embryo and fetus, adenosine levels rise markedly in response to stressors that include reduced placental blood flow and hypoxia (Rivkees, 1997). Thus, we postulate that under conditions of fetal stress, increased adenosine levels in the embryo will lead to decreased heart rates, reduce fetal cardiac output, and slow cardiac cell proliferation. The consequences of A1AR activation, as shown here, include fetal heart failure and cardiac hypoplasia.
Overall, our observations show that activation of A1ARs can influence cardiac structural development. Further work is indicated to assess the role of A1AR activation in the pathogenesis of congenital cardiac defects and fetal heart failure.
In Vivo Drug Treatment
This use of animals in our studies was approved by the Yale Animal Care and Use Committee. C57BL/6J mice were paired overnight. The next morning was designated embryonic day (E) 0.5 if a vaginal plug was present. For in vivo treatment studies, pregnant mice were injected with the A1AR agonist N6-cyclopentyladenosine (CPA) or vehicle every 12 hr, for a total of two injections. At least 4 mice were used in each group at each stage. At 22 hr of treatment, BrdU (5′-bromo-2′-deoxyuridine, Sigma; 50 mg/kg) was injected to label dividing cells for 2 hr. Animals were then killed, and embryos dissected and frozen in OCT freezing medium (Sakura).
Embryos (E7.0 and E10.5) were cultured in the roller system as described by Sturm and Tam (1993). Briefly, embryos (E7.0 or E10.5; n = 2–5/bottle) were cultured with 5 ml rat serum that was prepared as described by Hogan et al. (1994) at 38°C for 24–48 hr. Embryos were treated with CPA in the presence of adenosine deaminase (ADA; 0.3 U/ml) during the entire course of culture and labeled with BrdU (30 μg/ml) for 1–4 hr, before being harvested. Control embryos were treated with vehicle alone. Embryos were either fixed with 4% paraformaldehyde (PFA) in phosphate buffered saline (PBS, pH 7.4) or embedded and frozen in OCT medium.
For in vitro explant studies, dissected ventricular tissues (500 × 500 μm) were placed on Millipore filters (Millipore) and cultured in Dulbecco's modified Eagle medium (DMEM; GIBCO) containing 10% fetal bovine serum (GIBCO) in modified Trowel's organ culture system (Vainio et al., 1993; Zhao and Rivkees, 2000). Cultures were incubated at 37°C in 5% CO2 and 95% room air. Explants were pre-incubated with ADA (0.3 U/ml) for 30 min to degrade adenosine in the medium and tissues. ADA was included in the media in all subsequent drug treatments.
For drug treatments, CPA was added to the medium to achieve final concentrations of 10 nM, 100 nM, or 1 μM. Control incubations were treated with vehicle. To block A1AR activation, some explants were pre-treated with the A1AR antagonist 8-cyclopentyl-1,3-dipropylxanthine (DPCPX; 1 μM) for 1 hr, and then treated with CPA in the presence of DPCPX. Explants were collected after 4, 8, or 24 hr of treatment. One hour before explants were harvested, BrdU (30 μg/ml final concentration) was added to the media (Zhao and Rivkees, 2000). Specimens were then embedded and frozen in OCT medium.
Frozen sections (8 μm) were cut on a cryostat. Adjacent sections of each specimen were collected on two separate slides and used for TUNEL or BrdU assays.
To assess apoptosis, the TUNEL assay was performed as described by Zhao and Rivkees (2000). Briefly, sections were fixed in 4% PFA in PBS for 20 min, permeated with 0.1% Triton X-100 (10 min), and incubated with 10 μg/ml proteinase K (10–15 min) at room temperature. TUNEL labeling reaction was carried out following the manufacturer's (Boehringer Mannheim) instructions. Color detection of alkaline phosphatase was carried out with 4-nitro blue tetrazolium chloride and x-phosphate/5-bromo-4-chloro-3-indolyl-phosphate (Boehringer Mannheim) in 100 mM Tris-HCl, pH 9.5, 100 mM NaCl, 50 mM MgCl2, 2 mM levamesole. Sections were then counterstained with 0.1% safranine A or Eosin.
Cell Proliferation Assays
BrdU incorporation was used to assess cell proliferation as described by Zhao and Rivkees (2000). Briefly, sections or whole mount specimens were fixed with 4% PFA, incubated with 4 M HCl for 30 min, and neutralized with 100 mM Tris-HCl (pH 8.5). BrdU labeling was detected with anti-BrdU antibody (Sigma).
Microscopic images were captured using Image Pro program (Media Cybernetics) attached to an Olympus microscope. Photographs of whole mount specimens were taken with a digital camera (Olympus) under a Zeiss dissecting microscope. Pictures were arranged using Adobe® Photoshop®.
Measurements and Statistical Analysis
After in vivo treatment with CPA, embryos and hearts were dissected in PBS. The coronal-rump length of the embryos and length of the left ventricles from the atrioventricular junction to the apex were measured. Individual embryos and hearts were weighted after removing excessive solution.
For counting BrdU positive cells in the tissue sections, the number of cells per unit area on every other section throughout the specimen was counted. Differences between control and treatment were analyzed using the t-test, with P < 0.05 designated as the level of significance.