(Received for publication, December 16, 1996, and in revised form, March 19, 1997)
From the Falk Cardiovascular Research Center,
Division of Cardiovascular Medicine, Stanford University School of
Medicine, Stanford, California 94305 and the ¶ Molecular
Cardiology Laboratory, Stanford Research Institute International, Menlo
Park, California 94025
Early heart failure is characterized by elevated
plasma atrial natriuretic peptide (ANP) levels, but little is known
about the direct effects of ANP on cardiac myocytes. In neonatal rat cardiac myocytes, ANP induced apoptosis in a dose-dependent
and cell type-specific manner. Maximum effects occurred at 1 µM ANP, with a 4-5-fold increase in apoptotic
cells, reaching a maximum apoptotic index of 19%. In contrast, the
maximum apoptotic index of ANP-treated non-myocytes was 1.1 ± 0.2%, equivalent to control cultures. ANP treatment also sharply
reduced levels of Mcl-1 mRNA, a Bcl-2 homologue, coincident with
the increase in the incidence of apoptosis. ANP induction of apoptosis
was receptor-dependent and mediated by cyclic GMP: the
effect was mimicked by 8-bromo-cGMP, a membrane-permeable analog, and
by sodium nitroprusside, an activator of soluble guanylyl cyclase, and
was potentiated by a cGMP-specific phosphodiesterase inhibitor,
zaprinast. Interestingly, norepinephrine, a myocyte growth factor,
inhibited ANP-induced apoptosis via activation of the -adrenergic
receptor and elevation of cyclic AMP. These results show that ANP is a
specific effector of cardiac myocyte apoptosis in culture via
receptor-mediated elevation of cGMP. Furthermore, at least in this
model, ANP and norepinephrine may have opposing roles in the modulation
of cardiac myocyte growth and survival.
Apoptosis, or programmed cell death, plays a central role in development and homeostasis in all species. Apoptosis differs from necrosis in that it occurs by activation of an energy-requiring molecular suicide program, while necrosis is generally regarded as an accidental or unregulated phenomenon. Apoptosis is seen in cells undergoing embryonic deletion, after serious damage, or following an inappropriate or incomplete growth stimuli. Apoptosis is an important component of organ remodeling in development and during adaptation to stress (1). Both proliferating and quiescent cells are capable of undergoing apoptosis in response to noxious or sublethal events. The pathological activation of apoptosis is now thought to contribute to a variety of disease processes, including heart failure, myocardial infarction, and arrhythmogenic right ventricular dysplasia (2-14). However, the molecular events leading to myocardial apoptosis are poorly understood.
In heart failure, contractile abnormalities are accompanied by multiple compensatory disturbances in myocardial bioenergetics, neural and hormonal circulatory control, and peripheral vascular tone (15-18). Early in the development of chronic congestive heart failure, plasma concentrations of atrial natriuretic peptide (ANP)1 and brain natriuretic peptide increase markedly (16). These peptides have natriuretic, diuretic, and vasorelaxant effects and may decrease cardiac filling pressures and arterial pressure, increase sodium excretion, and inhibit activation of the renin-angiotensin-aldosterone system. Although these observations suggest a compensatory role for ANP in heart failure, the precise contribution of these peptides to the pathophysiology of congestive heart failure is not completely understood. We and others have shown that ANP and CNP inhibit vascular smooth muscle growth (19-21) and that ANP induces apoptosis in cultured vascular smooth muscle cells (22). We accordingly examined the effects of natriuretic peptides on cardiac myocyte apoptosis.
All natriuretic peptides and analogs were
purchased from Peninsula Laboratories Inc. (Belmont, CA) and were
>99% pure as determined by high performance liquid chromatography.
RNAzol B RNA isolation kit was purchased from Tel-Test Inc.
(Friendswood, TX). RT-PCR kits, and cell culture reagents were
purchased from Life Technologies, Inc. Zaprinast was obtained from
Biomol Research Laboratories, Inc. (Plymouth Meeting, PA). Hoechst
33342 (H33342) and propidium iodide were purchased from Molecular
Probes (Eugene, OR). DIG oligonucleotide 3-end labeling kit and DIG
luminescent detection kit were purchased from Boehringer Mannheim. All
other reagents were of the highest purity available and were obtained
from Sigma unless noted.
Enriched cultures of myocyte and non-myocyte cells were obtained from the neonatal rat by stepwise trypsin dissociation as described previously (30). After overnight plating in minimum Eagle's medium + 5% fetal calf serum, cells were maintained in serum-free minimum Eagle's medium supplemented with transferrin and insulin as described previously (30) The final myocyte cultures contained >90% ventricular myocytes at a density of 4 × 106 cells/60-mm dish. Experiments were performed 3 days after plating. To decrease the potential effect of endogenously produced natriuretic peptides, the culture medium was replaced daily.
Non-myocytes from the same cultures were obtained by selective preplating (30). These cells (approximately 95% fibroblasts, with small percentages of smooth muscle and endothelial cells) were used at passages 1-3. For experiments, non-myocytes were plated onto 35-mm dishes or two-welled chamber slides at a density of 1 × 105 cells/cm2 in 5% serum. Following overnight attachment, the non-myocyte cultures were placed in serum-free medium as above, and experiments were performed 3 days later.
Quantitative Analysis of Apoptotic NucleiCells were analyzed for apoptosis following visualization of nuclear morphology with the fluorescent DNA-binding dyes H33342 and propidium iodide. On day 3 of serum-free culture, cells were treated for various times with the indicated test compounds. The monolayers were rinsed with phosphate-buffered saline and then incubated with 5 µg/ml H33342 and 5 µg/ml propidium iodide for 30 min. Individual nuclei were visualized at × 400 using fluorescent microscopy and analyzed for apoptotic characteristics; propidium iodide was used to identify nonviable cells. Propidium iodide-positive cells with near-normal chromatin were counted as necrotic. An average of 800-1000 nuclei from random fields were analyzed for each data point. The apoptotic index (percentage of apoptotic nuclei) was calculated as (apoptotic nuclei/total nuclei) x 100%. Sample identities were concealed during scoring, and at least three samples were scored per group.
Analysis of DNA FragmentationCells were treated with test
compounds for 24 h on days 3-4 of culture. Nuclear DNA was
isolated from treated and untreated cells following cell lysis in 300 mM NaCl, 50 mM Tris, pH 8, 25 mM
EDTA containing 0.15 mg/ml proteinase K and 100 µg/ml RNase A. The
lysate was incubated for 3 h at 37 °C and was then
deproteinated by two phenol/chloroform/isoamyl alcohol extractions. 1 µg of DNA was used for end labeling with DIG-ddUTP and terminal
transferase for 30 min at 37 °C. DNA labeling efficiency was checked
by dot blot followed by detection with an alkaline phosphatase-coupled anti-digoxigenin antibody and the chemiluminescent substrate disodium 3-(4-methoxyspiro{1,2-dioxetane-3,2-(5
-chloro)tricyclo[3.3.1.13,7]decan}-4-yl)phenyl
phosphate. For the analysis of DNA fragmentation, digoxigenin-labeled DNA was size-fractionated in a 2% agarose gel
and cleaved with brief depurination in 0.2 N HCl for 10 min and partial hydrolysis in 0.5 N NaOH, 1.5 M
NaCl for 30 min followed by neutralization with 1 M
Tris-HCl, 1.5 M NaCl. Following transfer to nylon membrane
overnight in 20 × SSC, the DNA was visualized immunochemically as above.
Desmin was used to distinguish cardiac myocytes from non-myocytes as described previously (24). Myocyte and non-myocyte cultures were plated on 2-well chamber slides and cultured for 3 days prior to 24-h treatment with ANP as described above. Cells were fixed and permeabilized with ice-cold methanol, blocked with 5% heat-inactivated horse serum/phosphate-buffered saline and sequentially incubated with anti-desmin monoclonal antibody (1:200 dilution in Tris-buffered saline, Sigma), horse anti-mouse IgG antibody conjugated to biotin (3 µg/ml, Vector Laboratories), and FITC-avidin (5 µg/ml) and H33342. The cells were visualized using fluorescence microscopy and documented by dual exposure photography.
RNA Extraction and PCR AnalysisTotal cellular RNA (1 µg)
was extracted using RNAzol B and analyzed by RT-PCR. Oligonucleotide
primers specific for Mcl-1 were derived from the human sequence (25):
forward, 5 CTG TAC AAG GGA AGC TTT 3
; reverse, 5
GAA AAG CTG TAA GAA
ACA GG 3
(product: 334 base pairs). As an internal control for the
semiquantitative analysis of Mcl-1,
-actin was also amplified.
-Actin primers were derived from the rat sequence, as described
previously (23): forward, 5
TGG AGA AGA GCT ATG AGC TGC CTG 3
;
reverse, 5
GTG CCA CCA GAC AGC ACT GTG TTG 3
(product: 210 base
pairs). For the analysis of Mcl-1 mRNA, the RNA was
reverse-transcribed and then aliquoted into two PCR tubes, one for
amplification of
-actin transcripts and the second for amplification
of Mcl-1 transcripts. PCR was performed for 35 (Mcl-1) or 15 (
-actin) cycles each at 94 °C × 1 min, 55 °C × 2 min, 72 °C × 3 min, and 1 cycle of 72 °C × 15 min.
After reaction, the PCR products for Mcl-1 and
-actin were pooled,
size-separated through 1.5% agarose gel and visualized under UV
light.
Data are presented as mean ± S.E. Analysis was performed by analysis of variance followed by Fischer PLSD, p < 0.05% was considered significant.
To test the
ability of ANP to modulate apoptosis in cardiac myocytes, enriched
cultures of neonatal myocytes were treated with 1 µM ANP
for 24 h on day 4 of culture and examined in situ following staining with H33342 and propidium iodide (Fig.
1A). ANP treatment caused a marked increase
in the percentage of apoptotic cells (19.1% ± 1.45 versus
4.78 ± 0.65 in control cultures, n = 10, p < 0.05). The percentage of necrotic cells was
identical in vehicle and ANP-treated cultures (3.25 ± 0.24% and
3.38 ± 0.52%, respectively, n = 10).
Induction of apoptosis by ANP was dose-dependent and receptor-mediated. A maximum 4.1-fold induction was obtained at an ANP concentration of 1 µM and an approximate ED50 of 5-10 nM (Fig. 1B). This effect was not seen with two different synthetic analogs of ANP, des-(Cys-Cys) rANP and C-ANF, that bind only to the ANP clearance receptor but not the guanylyl cyclase-linked A- and B-type receptors (Fig. 1B and data not shown). Similarly, ANP isolated from the Japanese eel (Anguilla japonica), which has very low affinity for the rat ANP receptor, and exerts 20-fold less vasodepressor and natriuretic activity in the rat (26), had a correspondingly reduced effect on the induction of apoptosis in these rat myocytes (Fig. 1B). The pro-apoptotic activity of ANP is thus likely to be mediated via the A- and B-type natriuretic peptide receptors.
DNA cleavage into nucleosomal sized fragments is a hallmark of apoptotic cell death (1). DNA isolated from control myocytes exhibited a low level of DNA fragmentation, consistent with a low, basal rate of apoptosis under serum-free conditions. On the other hand, DNA from the ANP-treated myocytes was extensively fragmented, yielding the characteristic nucleosomal "ladder" (Fig. 1C).
ANP-induced Apoptosis Is Myocyte-specificThe non-myocytes
from the neonatal heart do not undergo ANP-induced apoptosis. Enriched
cultures of primary myocyte contain a small percentage (5%) of
non-myocytes, of which >90% are fibroblasts and ~5% are smooth
muscle cells by morphological and immunocytochemical criteria
(24).2 In confirmation of an earlier report
(27), RT-PCR analysis showed that both myocyte and non-myocyte
populations express all three known isoforms of the natriuretic
receptors at similar levels (data not shown). Thus, both cell
populations may be competent to respond to ANP. However, ANP did not
appear to induce apoptosis in the non-myocyte cultures. Under basal
conditions, the apoptotic index of the non-myocyte cultures was very
low (<1%) and was unchanged by ANP as assessed by nuclear morphology
(Fig. 1B) or DNA laddering assay (Fig. 1C). Thus,
although non-myocytes express receptors for natriuretic peptides, they
appear to be resistant to ANP-induced apoptosis.
To confirm this observation, we directly examined the apoptotic
cells in non-myocyte and myocyte cultures with a monoclonal mouse
anti-desmin antibody and the nuclear stain H33342 (Fig. 2). Desmin expression is generally a muscle-restricted
protein, and only the cardiac myocytes in these mixed primary cultures are desmin-positive (24). Greater than 95% of the cells in these cultures were desmin-positive and possessed other morphological features of cardiac myocytes (Fig. 2, B and C).
In contrast, fewer than 1% of the cells in the non-myocyte cultures
were desmin-positive, and these typically represented contaminating
myocytes. Virtually all apoptotic cells under basal conditions and
after ANP treatment were also desmin-positive. We conclude that, in
these mixed cultures, ANP-induced apoptosis is limited to cardiac
myocytes.
Non-myocytes Do Not Influence ANP-induced Myocyte Apoptosis
It is possible that ANP-induced apoptosis is an
indirect effect mediated by non-myocytes (or myocytes) through
paracrine/autocrine mechanisms. To rule out this possibility, we
conducted a series of experiments in which media were conditioned by
incubation with myocytes, non-myocytes, or in empty tissue culture
dishes, for varying intervals after the addition of ANP. The results
are summarized in Fig. 3, A and B.
Neither myocyte nor non-myocyte-derived media altered basal rates of
myocyte apoptosis after 12 h of conditioning, regardless of ANP
treatment (Fig. 3A). Furthermore, the ability of ANP-treated
non-myocyte conditioned media to induce apoptosis at earlier time
points exactly paralleled that of ANP-containing medium maintained in
an empty tissue culture dish. Induction of apoptosis by either
medium was maximal immediately after ANP addition and declined in a
linear fashion over a further 8 h of conditioning (Fig.
3B). These data suggest that all of the apoptotic
potential of the conditioned media can be attributed to the added ANP
peptide. Thus, it is unlikely that non-myocytes contribute to the
effects of ANP on cardiac myocyte apoptosis.
cGMP Can Induce Myocyte Apoptosis
As shown above, the
ANPC receptor is not likely to be a mediator of ANP-induced
apoptosis. Of the two other ANP receptor subtypes, ANPA and
ANPB, only the former binds ANP with high affinity (28). The dose-response relationship presented in Fig. 1 is thus consistent with an action mediated by the ANPA receptor. The
ANPA receptor is coupled to generation of cGMP, suggesting
that guanylyl cyclase and cGMP may participate in ANP-induced
apoptosis. To confirm this, we tested a series of cGMP modulating
agents. Sodium nitroprusside (106 to 10
3
M), a guanylate cyclase-activating nitric oxide donor, also
induced cardiac myocyte apoptosis (Fig. 4A).
Zaprinast, a cGMP-specific phosphodiesterase inhibitor that has been
shown to potentiate other effects of ANP on cardiac myocytes (29),
strongly potentiated the induction of apoptosis by a submaximal dose of
ANP (1 nM, Fig. 4B; 3.3-fold versus
2.1-fold in the absence of zaprinast.) Finally, the cell-permeable cGMP
analog 8-bromo-cGMP caused a dose-dependent induction of
apoptosis at 24 h, with a maximum of 3.4-fold over control at 100 µM and an approximate EC50 of 0.08 µM (Fig. 4A). Taken together, these results
strongly suggest that ANP-induced apoptosis is mediated by a guanylyl
cyclase linked receptor via elevation of intracellular cGMP.
ANP-induced Apoptosis Is Blocked by Norepinephrine
Norepinephrine is an adrenergic agonist known to
stimulate hypertrophic growth of cultured myocytes (30). Since growth
factors frequently act in opposition to apoptosis in other systems, we asked whether norepinephrine would have an inhibitory effect on basal
and ANP-induced cardiac myocyte apoptosis. Treatment of cultures with
norepinephrine alone for 24 h did not influence the basal rate of
apoptosis (Fig. 5A). However, norepinephrine reduced the apoptotic index to control levels in ANP-stimulated myocytes (Fig. 5A).
Norepinephrine influences myocyte growth in culture via both
1 and
-adrenoreceptors (30-34), depending on the
extent of cell contact. To examine the receptor class mediating the
anti-apoptotic effect of norepinephrine, myocytes were treated with ANP
and norepinephrine alone or in combination with subtype-selective
adrenergic receptor antagonists, either propranolol (
-selective) or
terazosin (
1-selective). Only the
-adrenoreceptor
antagonist propranolol blocked the anti-apoptotic effect of
norepinephrine (Fig. 5B), suggesting that this effect was
mediated by a
-adrenoreceptor. Neither antagonist had independent effects on basal or ANP-induced apoptosis (Fig.
5B).
-Adrenoreceptor-mediated signaling in cardiac myocytes involves both
an activation of the adenylate cyclase/cAMP generating pathway and a G
protein mediated increase in extracellular calcium influx (31). To
further analyze the protective effects of
-adrenergic stimulation,
we measured ANP-induced apoptosis in the presence of various
combinations of isoproterenol, a
-adrenergic agonist; nifedipine, an
L-type calcium entry antagonist; and KT5720, a highly selective
inhibitor of cAMP-dependent protein kinase (protein kinase
A, Ref. 31). As predicted, isoproterenol (4 µM) was
equipotent to norepinephrine in blocking ANP-induced apoptosis.
Interestingly, inhibition of protein kinase A, but not of calcium
entry, blunted the anti-apoptotic effects of isoproterenol, suggesting
a cAMP-dependent effect (Fig. 5C). Consistent
with these results, the membrane-permeable cAMP analog dibutyryl cAMP
(Bt2cAMP) also inhibited ANP-induced apoptosis (Fig.
5D). The effects of Bt2cAMP were
dose-dependent and were reversed at concentrations higher
than 1 mM (not shown). These results suggest that the
anti-apoptotic effect of
-adrenergic stimulation is mediated through
activation of protein kinase A and does not require calcium influx.
Apoptosis is modulated by the balance of pro- and
anti-apoptotic proteins, including members of the Bcl-2 family (35,
36). Mcl-1, a Bcl-2 homologue (25), has been shown to inhibit apoptosis in Chinese hamster ovary cells (37) and appears to be expressed at
relatively high levels in the heart in comparison with other Bcl-2-related proteins (38). PCR analysis revealed that ANP induced a
time-dependent decrease in Mcl-1 mRNA levels that
reached a nadir at 16 h and returned to base line at 48 h
(Fig. 6). Control PCR-amplified -actin transcript
levels did not change over the same interval. Mcl-1 transcripts were
not detectable in non-myocytes using the same protocol (not shown).
In this paper, we report a novel direct effect of ANP on cardiac myocytes, the induction of apoptosis. ANP-induced apoptosis is dose-dependent, receptor-specific, and involves natriuretic peptide receptor-mediated increases in cGMP. Furthermore, this apoptosis is antagonized by catecholamine-induced increases in cAMP. Given the previously documented ability of the neonatal rat cardiac myocyte cell culture system to identify direct cardiotrophic properties of both soluble and mechanical factors (39-44), our results provide a basis for further investigation of this property of ANP in animal models of cardiac growth and failure.
A number of neurohumoral and autocrine substances, including
endothelin, angiotensin II, and norepinephrine, are known to stimulate
cardiac myocyte and smooth muscle growth (20-22, 41-45). That there
might exist a complementary ANP-mediated mechanism to inhibit growth in
these cells is perhaps not surprising. Growth inhibition is often
accompanied by an increased probability of apoptosis, while
growth-promoting agents tend also to promote survival. For example,
angiotensin II inhibits smooth muscle cell apoptosis (46), while nitric
oxide, transforming growth factor 1 and ANP, all potent inhibitors
of smooth muscle cell growth, induce apoptosis in these cells (46, 47,
22). Significantly, nitric oxide is thought to induce apoptosis in
cardiac myocytes via cGMP generation, similar to ANP (6). Since the
heart is the major source of this natriuretic peptide, we postulate
that ANP may be an autocrine growth inhibitor, opposing the effect of
autocrine and paracrine growth factors.
Enhanced production and secretion of ANP accompanies cardiac myocyte hypertrophy. Although ANP has been shown to inhibit growth of cardiac fibroblasts in culture (48), as yet there is only indirect evidence that natriuretic peptides influence myocyte growth. Transgenic mice overexpressing ANP exhibit a 27% reduction in total heart weight (49) and appear to have a blunted right ventricular hypertrophic response to hypoxia-induced pulmonary hypertension (50), while mice homozygous for disruption of the pro-ANP gene exhibit a significant increase in the ratio of heart weight to body weight (51). Stabilization of ANP with alatriopril, a mixed angiotensin converting enzyme/atriopeptidase inhibitor, resulted in a greater reduction of compensatory hypertrophy compared with an angiotensin converting enzyme inhibitor alone in a rat model of myocardial infarction (52). Although some of these observations may be due in part to ANP effects on blood pressure, it is clear that the net effect of ANP on myocardium in vivo is one that limits myocyte growth.
Another interesting feature of our results is the inverse relationship
between cAMP and cGMP with respect to apoptosis. cGMP itself may either
inhibit (53-56), promote (6, 22, 46, 57-61), or have no effect (see
Figs. 1 and 2) on apoptosis, depending on the cell type. In this case,
ANP induction of apoptosis via cGMP clearly involves cardiac
myocyte-specific signal transduction mechanisms. The -adrenergic
receptor-mediated trophic effects of norepinephrine are likewise
restricted to cardiac myocytes in these cultures (31).
-Adrenergic
stimulation antagonized the pro-apoptotic effects of ANP through a
cAMP/protein kinase A-dependent mechanism associated with
morphological evidence of cell growth (not shown). cAMP and cGMP
exhibit a reciprocal relationship in studies of
-adrenoreceptor-stimulated cardiac contractility (62), and the
molecular mediators of these effects on contractility have been well
characterized. However, the molecular targets for cyclic
nucleotide-modulated cell fate decisions are unknown.
Our data suggests that the anti-apoptotic Bcl-2 homologue, Mcl-1 (25, 35-38), may be one such target in the cardiac myocyte. The Bcl-2 homologue Mcl-1 was initially identified as a protein up-regulated during the differentiation of a monocytoid cell line, ML-1 cells (25), and has been shown to inhibit apoptosis in Chinese hamster ovary cells (37). Interestingly, Mcl-1 is expressed at high levels in the heart (38), an observation confirmed in this study. Although these studies do not establish a causal relationship between the ANP-induced fall in Mcl-1 transcripts and increased apoptosis, it is consistent with the hypothesis that Mcl-1 expression is one determinant of the probability of apoptosis in the cardiac myocyte, along with many other factors (25, 35, 36). Further studies on the role of Mcl-1 and related proteins in ANP-induced apoptosis are in progress.
A number of neuroendocrine effectors, including ANP, brain natriuretic peptide, renin, angiotensin II, norepinephrine, and arginine vasopressin, are present at high levels early in the development of heart failure and may play an important role in its progression (15-18). Our data suggest that the sustained elevation of ANP may have direct effects on the cardiac myocyte independent of its hemodynamic actions. Whether the cumulative effects of ANP and other agents are harmful or beneficial in vivo is likely to depend on a balance of external and intracellular factors. Further studies of this effect of ANP will help to illuminate the precise molecular effectors and targets of growth restriction and apoptosis in the cardiac myocyte.
We thank Barbara Sato for technical assistance and Keith Webster, Pedro Trindade, and Howard Hutchinson for discussion and advice.