Am J Physiol Cell Physiol 289: C609 –C616, 2005.
First published May 4, 2005; doi:10.1152/ajpcell.00065.2005.
Do rat cardiac myocytes release ATP on contraction?
Stefanie Gödecke,1 Thomas Stumpe,1 Hilmar Schiller,2 Hans-J. Schnittler,3 and Jürgen Schrader1
1
Institut für Herz- und Kreislaufphysiologie, Heinrich-Heine-Universität, Düsseldorf; 2Cardion, Erkrath;
and 3Institut für Physiologie, Medical Fakultät Carl-Gustav-Carus, Dresden, Germany
Submitted 15 February 2005; accepted in final form 21 April 2005
cardiomyocytes; luciferase
ATP acts as a potent agonist on a variety of
different cell types, including cardiomyocytes (20), inducing a
broad range of physiological responses. The cellular effects
mediated by ATP are determined by the subtypes of P2
purinergic receptors expressed in the particular cell type. In
cardiomyocytes, the expression of ionotropic P2X1–P2X7 receptors and metabotropic P2Y1, P2Y2, P2Y4, P2Y6, and P2Y11
receptors have been described (37). The sensitivity of these
receptors toward extracellular ATP is characterized by their
EC50-ATP values, which are reported to be in the low micromolar range (37). The diversity of ATP receptors expressed in
cardiomyocytes reflects the variety of ATP effects described
for single cells as well as for the whole organ. In the single
cardiomyocyte, micromolar levels of extracellular ATP increase plasma membrane permeabilities for cations (28) and
EXTRACELLULAR
Address for reprint requests and other correspondence: S. Gödecke, Institut
für Herz- und Kreislaufphysiologie, Heinrich-Heine-Universität, Düsseldorf,
Universitätsstrasse 1, D-40225 Düsseldorf, Germany (e-mail: stefanie.
[email protected]).
http://www.ajpcell.org
for Cl⫺ (18), intracellular calcium transients (37, 43), the rate
of phosphoinositide hydrolysis (29, 41), and the contraction
amplitude (24, 27, 43). Moreover, ATP can stimulate phospholipase C (PLC) (27, 29), influence pH and inhibit glucose
transport (11). On the organ level, ATP acts as a positive
inotropic agent (24) and can induce various forms of arrhythmia (37). Furthermore, ATP can induce vasodilation by interacting with P2Y receptors of the vascular endothelium.
Low levels of ATP can be detected in the effluents of
isolated, saline-perfused hearts (4). Because extracellular ATP
is known to be rapidly degraded by ecto-ATPases present on
endothelial cells and cardiomyocytes (44), much higher local
concentrations of ATP can be assumed to exist locally at the
site of release. Kuzmin et al. (21) estimated the baseline
interstitial fluid ATP concentration in hearts in situ to be ⬃40
nM. However, a variety of conditions, such as electrical stimulation (1), hypoxia and ischemia (10, 21, 40), increased blood
flow (39), ␤-adrenergic stimulation (4), or mechanical stretch
(36) have all been reported to be associated with a marked
increase in the cardiac release of ATP.
Extracellular ATP in the cardiovascular system may originate from different cellular sources. ATP is well known to be
a cotransmitter in perivascular sympathetic nerve endings (5).
Activated platelets release substantial amounts of ATP during
hemostasis (14). Furthermore, endothelial cells (3), vascular
smooth muscle cells (26), erythrocytes (32), and inflammatory
cells (8) were described as possible sources for extracellular
ATP. Under ischemic conditions, myocytes release ATP (10,
11, 40), and it has also been speculated that exercising heart
muscle cells are a source of extracellular ATP, similar to what
has been reported for the working skeletal muscle cell (12, 25).
Finally, cells disrupted, e.g., on hypoxia or vascular injury,
release substantial amounts of their cytosolic nucleotides.
Mechanical strain leads to the release of ATP from a variety
of cell types. Endothelial cells release ATP in response to fluid
shear stress, whereas this is not the case for vascular smooth
muscle cells (3). Mechanical perturbation of the cell membrane
during hypoosmotic swelling mediates ATP release from prostate cancer cells (30), epithelial cells (15), and also neonatal
cardiomyocytes (10). Furthermore, it has been reported that in
cultured neonatal cardiomyocytes, ATP release (36) and a
variety of stretch-induced biochemical events such as activation of protein kinases or an increase of protein synthesis (42)
can be stimulated by application of mechanical force. During
the cycle of systolic contraction and diastolic relaxation, the
sarcolemmal membrane is subjected to continuous and pronounced mechanical deformations. In this study, we addressed
the question whether this mechanical stimulus is sufficient to
The costs of publication of this article were defrayed in part by the payment
of page charges. The article must therefore be hereby marked “advertisement”
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
0363-6143/05 $8.00 Copyright © 2005 the American Physiological Society
C609
Downloaded from http://ajpcell.physiology.org/ by 10.220.33.4 on June 18, 2017
Gödecke, Stefanie, Thomas Stumpe, Hilmar Schiller, Hans-J.
Schnittler, and Jürgen Schrader. Do rat cardiac myocytes release ATP
on contraction? Am J Physiol Cell Physiol 289: C609 –C616, 2005. First
published May 4, 2005; doi:10.1152/ajpcell.00065.2005.—ATP is released by numerous cell types in response to mechanical strain. It then
acts as a paracrine or autocrine signaling molecule, inducing a variety
of biological responses. In this work, we addressed the question
whether mechanical force acting on the membranes of contracting
cardiomyocytes during periodic longitudinal shortening can stimulate
the release of ATP. Electrically stimulated isolated adult rat cardiomyocytes as well as spontaneously contracting mouse cardiomyocytes
derived from embryonic stem (ES) cells were assayed for ATP release
with the use of luciferase and a sensitive charge-coupled device
camera. Sensitivity of soluble luciferase in the supernatant of cardiomyocytes was 100 nM ATP, which is ⬃10-fold below the EC50
values for most purinergic receptors expressed in the heart (1.5–20
␮M). Light intensities were not different between resting or contracting adult rat cardiomyocytes. Similar results were obtained with
ES-cell-derived contracting mouse cardiomyocytes. ATP release was
measurable only from obviously damaged or permeabilized cells. To
increase selectivity and sensitivity of ATP detection we have targeted
a recombinant luciferase to the sarcolemmal membrane using a wheat
germ agglutinin-IgG linker. Contraction of labeled adult rat cardiomyocytes was not associated with measurable bioluminescence. However, when human umbilical vein endothelial cells were targeted with
membrane-bound luciferase, shear stress-induced ATP release could
be clearly detected, demonstrating the sensitivity of the detection
method. In the present study, we did not detect ATP release from
contracting cardiomyocytes on the single cell level, despite adequate sensitivity of the detection system. Thus deformation of the
contracting cardiomyocyte is not a key stimulus for the release of
cellular ATP.
C610
ATP RELEASE FROM CONTRACTING CARDIOMYOCYTES
METHODS
Preparation and culture of adult rat cardiomyocytes. Adult Ca2⫹sensitive cardiomyocytes of Wistar rats (250 –300 g) were prepared as
described previously (34). After preparation cells were seeded on
laminin (10 ␮g/ml)-coated coverslips (⬃28 mm) and incubated in
serum-free medium 199 (M199), supplemented with nonessential
amino acids, vitamins, gentamicin (50 ␮g/ml), penicillin (250 IU/ml),
streptomycin (250 ␮g/ml), and insulin (1,000 IU/l). The medium was
changed after 3 h to remove nonadherent cells. After at least 12 h, and
a maximum of 2 days, cardiomyocytes were used for experiments.
ES cell culture and generation of genetically selected cardiomyocytes. ES cell-derived mouse cardiomyocytes were generated as described (19). Undifferentiated J1 ES cells were transfected via electroporation with a construct carrying ␣-myosin heavy chain
(MHC)-aminoglycoside phosphotransferase (MHC-neor) and a phosphoglycerate kinase (pGK)-hygromycin-resistance transgene. Transfected cells were selected by incubation in growth medium containing
hygromycin B (200 ␮g/ml), resulting in the isolation of clone CM7/1.
For differentiation, 4 ⫻ 106 ES cells per 100-mm bacterial dish were
plated and cultured in growth medium lacking leukemia inhibitory
factor. After 4 days of being cultured, the resulting embryoid bodies
were plated onto 35-mm cell culture dishes. When spontaneous
contractile activity was noticed, growth medium supplemented with
G418 (400 ␮g/ml) was added to eliminate noncardiomyocytes.
Preparation and culture of HUVECs. Human umbilical cords were
obtained from a normal placenta. The umbilical vein was cannulated
with blunt needles and perfused to wash the blood out. The vein was
then filled with collagenase IA (Sigma) 0.025%, warmed at 37°C.
Isolated endothelial cells were seeded in 25-cm2 flasks coated with
0.2% gelatin A in endothelial cell basal medium (Promocell) and were
incubated at 37°C and 5% CO2 in a humidified incubator. Trypsinization was performed when endothelial cells reached 70 –90% confluence.
Coupling of firefly luciferase to surface of adult rat cardiomyocytes
or HUVECs. The principle underlying the coupling of firefly luciferase to the membranes of cardiomyocytes performed in this study is
depicted in Fig. 3. N-Acetylglucosamine residues present on cell
membranes in high copy numbers as parts of glycosylated membrane
proteins are specifically recognized by the lectin portion of the
WGA-IgG complex. The IgG component in turn interacts specifically
with the Z-domains of the protein A-luciferase fusion protein.
Generation of wheat germ agglutinin-IgG complexes. Rabbit antimouse IgG (whole molecule) was conjugated to wheat germ agglutinin (WGA) with sulfosuccinnimidyl 4-[N-maleimidomethyl]cyclohexane-1-carboxylate (sulfo-SMCC) as a cross-linking molecule with
the use of the Controlled Protein-Protein Cross-Linking Kit (Pierce).
In brief, 7 mg rabbit IgG were activated with sulfo-SMCC according
to the manufacturer’s instructions for generating maleimide-IgG. Free
AJP-Cell Physiol • VOL
sulfhydryl groups were introduced into 1 mg of WGA by reaction
with N-succinimidyl-S-acetylthioacetate, followed by deprotection of
the masked SH-groups with hydroxylamine. Immediately after deprotection, sulfhydryl-WGA was mixed with maleimide-IgG in approximately equal molar amounts. The cross-linked conjugate was dialyzed against cardiomyocyte stimulation buffer (see below) or PBS
and concentrated using Centriprep YM-3 centrifugal filter devices
(Amicon/Millipore). Final protein concentrations were adjusted to 100
␮g/ml. Aliquots (100 ␮l) were rapidly frozen and stored at ⫺80°C.
Generation of protein A-luciferase fusion protein. For construction
of a fusion gene containing four synthetic Z-domains based on the IgG
binding domain of protein A fused to the NH2 terminus of the firefly
luciferase gene, the following cloning strategy was used: plasmid
pGem-luc (Promega) was linearized with BamHI at the 5⬘ end of the
luciferase gene. A KpnI site was inserted by oligonucleotide ligation.
The resulting plasmid pGem-K was linearized with NotI, treated with
Klenov polymerase to generate blunt ends and cut with KpnI. A
388-bp FspI KpnI fragment isolated from plasmid pEZZ18 (Amersham Pharmacia) containing two synthetic Z-domains was integrated,
thereby reconstituting the NotI site 5⬘ to the Z-domains. To amplify
the number of IgG binding motifs, the resulting plasmid pGem-Z was
linearized with NotI and blunt ends were generated by Klenov treatment. A 373-bp FspI EcoRI fragment from pEZZ18 containing the
sequence identical to the above-mentioned FspI KpnI fragment was
inserted, resulting in plasmid pGem-ZZ. This construct contains an inframe fusion of a short part of the 5⬘ coding region of the lacZ
␣-peptide, followed by four synthetic Z-domains and the coding
region of the luc gene under control of the inducible lac promoter.
Plasmid pGem-ZZ was transformed into competent E. coli HB101
cells. Bacterial lysates of isopropyl-␤-D-thiogalactopyranoside-induced cultures were prepared as described by others for an almost
identical fusion protein (2). However, due to the high degree of
luciferase activity loss during IgG-agarose affinity purification, we
omitted the final affinity purification step. Instead, after the filtration
step, the bacterial lysate was dialyzed against cardiomyocyte stimulation buffer (see below) or PBS and concentrated with the use of
centrifugal filter devices (Centriprep YM-3, Amicon/Millipore). Final
protein concentrations were adjusted to 7.5 ␮g/␮l. Aliquots (100 ␮l)
were snap-frozen and stored at ⫺80°C. Luciferase activity of bacterial
extracts was determined in a total volume of 50 ␮l by mixing 40-␮l
reaction buffer, which contained (in mmol/l) 20 tricine, 0.033 coenzyme A, 1.07 (MgCO3)4Mg(OH)2, 0.1 EDTA, 33.3 DTT, 2.67
MgSO4, 0.53 ATP, and 0.47 luciferin, pH 7.3, with 10 ␮l of various
dilutions of bacterial extract. Luminescence was monitored in 10-s
intervals at 25°C with a Berthold luminometer (model Biolumat LB
9500 T). Luciferase activity was measured as 9 ⫻ 1010 relative light
units 䡠 mg protein⫺1 䡠 s⫺1.
The specific activity of the luciferase was not altered by addition of
the NH2-terminal Z-domains, as could be shown by comparative
activity assays and Western blot experiments using anti-luciferase
antibodies (Promega) on bacterial pGEM-luc and pGem-ZZ extracts
(data not shown).
Attachment of luciferase to cell surfaces. For experiments with cell
surface-bound luciferase, rat cardiomyocytes were cultured as described above on laminin-coated glass coverslips. Culture medium
was carefully removed and the coverslips were rinsed twice with
cardiomyocyte stimulation buffer that contained (in mmol/l) 137
NaCl, 5.4 KCl, 1.0 NaH2PO4, 0.8 Mg2SO4, 2.0 CaCl2, 5.5 glucose,
and 5.0 Tris, pH 7.4. Cells were incubated in stimulation buffer
containing WGA-IgG complex (diluted 1:10) for 30 min at room
temperature, allowing the WGA residues to bind to carbohydrates on
the cell surfaces. Subsequently, cells were carefully washed with
stimulation buffer and overlaid with a 1:10 dilution of bacterial extract
containing the luciferase fusion protein. During a 30-min incubation
period at room temperature, the Z-domains of the fusion protein could
bind to the surface-coupled IgG residues. After being washed, the
coverslips were brought into the perfusion/stimulation chamber con-
289 • SEPTEMBER 2005 •
www.ajpcell.org
Downloaded from http://ajpcell.physiology.org/ by 10.220.33.4 on June 18, 2017
induce the release of ATP from the contracting heart muscle
cell. This cardiomyocyte-derived ATP may then act in an
autocrine way on the different P2 receptors expressed by this
cell type, by modulating, for example, the contractile function
of the cell.
We measured ATP release from contracting rat cardiomyocytes, derived as primary cultures of adult cells or genetically
generated from mouse embryonic stem (ES) cells, respectively.
ATP was assessed using soluble luciferin/luciferase reagent or
a luciferase-protein A fusion protein targeted to the cell membrane in combination with a highly sensitive charge-coupled
device (CCD) camera, allowing the observation of single cells.
The sensitivity of the detection system was confirmed by using
shear stress-stimulated human umbilical vein endothelial
cells (HUVECs) as a positive control for local and transient
release of ATP.
ATP RELEASE FROM CONTRACTING CARDIOMYOCYTES
AJP-Cell Physiol • VOL
cells, and luminescence was then recorded and normalized for the
basal luminescence.
RESULTS
In initial experiments, freshly isolated adult rat cardiomyocytes were attached to a laminin-coated coverslip, allowing the
cells to contract on electrical stimulation. Cells were then
brought into cuvettes containing soluble luciferase/luciferin
reagent dissolved in cardiomyocyte stimulation buffer, which
then was inserted into a luminometer and stimulated electrically with two platinum electrodes. To avoid cell damage,
stimulation parameters were chosen to be just above the stimulation threshold (biphasic pulse, 200 ␮s, 10 ␮A). ATP release
was clearly detectable on stimulation (data not shown). However, microscopical observation of the cells revealed that
several cardiomyocytes were apparently damaged during electrical stimulation, visible by the loss of their typical rod-shaped
phenotype. Hence, it was not possible to distinguish in these
experiments whether the ATP was selectively released from
beating cardiomyocytes or originated from cells damaged during the experimental procedure.
Because overall measurements may have yielded erroneous
results, we decided to measure ATP release from single beating
cardiomyocytes using a sensitive CCD camera system. A
specially designed cardiomyocyte perfusion chamber allowing
both electrical stimulation and microscopical observation was
used (35). As shown in Fig. 1, A–C, for a typical experiment,
we defined regions that matched individual cells (regions 1–3).
Only rod-shaped cells, which contracted on electrical stimulation (biphasic pulse, 200 ␮s, 10 ␮A), were chosen for further
experiments (regions 1 and 2). Cells that evidently had lost
their rod-shaped phenotype at the beginning of the experiment
and did not contract on stimulation (region 3) were excluded
from the analysis. Bioluminescence determined from resting,
not electrically stimulated, cells (Fig. 1B) was compared with
the bioluminescence of identical cells stimulated electrically to
contract (Fig. 1C). Altogether, 21 resting and 16 contracting
cells derived from 5 independent experiments were analyzed.
As depicted in Fig. 1D, light intensities were 1,117.9 ⫾ 83.7
arbitrary units per pixel for the resting cells and 1,065.3 ⫾ 78.3
arbitrary units per pixel for the contracting cells. We conclude
that no detectable ATP release is observed in contracting
cardiomyocytes compared with resting cells. The weak ATP
signal observed in rounded cells (region 3) was independent of
electrical stimulation and was most probably a consequence of
lost membrane integrity and slow leakage of intracellular ATP
due to cell damage.
To arrive at an estimate of sensitivity of our measuring
system, we mixed different concentrations of ATP to the
luciferin/luciferase solution in our perfusion chamber and recorded luminescence. ATP concentrations in the midnanomolar range (100 nM, data not shown) could clearly be detected.
To exclude the possibility that membrane properties of
isolated cardiomyocytes might have been altered due to enzymatic digestion, a further series of experiments were carried
out using genetically selected cardiomyocytes derived from
mouse ES cells. A stable ES cell line containing a vector with
both MHC-neor and a pGK-hygromycin resistance transgene
was selected to generate spontaneously contracting cardiomyocytes, as described in detail by Klug et al. (19). For ATP
measurements, embryoid bodies were plated on 35-mm-thick
289 • SEPTEMBER 2005 •
www.ajpcell.org
Downloaded from http://ajpcell.physiology.org/ by 10.220.33.4 on June 18, 2017
taining stimulation buffer, and cells were allowed to equilibrate for 30
min before being monitored with a CCD camera.
For experiments with HUVECs, cells were cultured as described
above on gelatin-coated glass coverslips. Loading of cells was performed with WGA-IgG complex and luciferase extracts in PBS,
diluted 1:10 in endothelial cell basal medium.
Bioluminescence measurement on adult rat cardiomyocytes. Coverslips with luciferase-labeled adult cardiomyocytes were introduced
into a perfusion chamber, allowing both microscopical observation
and electrical stimulation of the cells. The exact geometry of the
chamber has been described previously (35). After an equilibration
period of 30 min, the chamber was perfused at constant flow (0.5
ml/min) with cardiomyocyte stimulation buffer (see above) containing
150 ␮M D-luciferin (Sigma), 100 ␮g/ml coenzyme A (Sigma), and in
some experiments, various ATP concentrations. In some experiments,
cells were electrically stimulated to contract using the lowest possible
stimulation strength plus 10% to minimize cell damage (stimulator
model G270, Strotmann, Aachen, Germany; 1 Hz, biphasic impulses
between 10 and 20 ␮A) using platinum electrodes. Light emission was
monitored using a Zeiss Axiovert 35 microscope in combination with
a CCD camera (Micromax 512 BFT System, Visitron Systems,
Puchheim, Germany). Luminescence values were calculated using
MetaMorph Software (Universal Imaging).
For experiments with soluble luciferase, coverslips with adult
cardiomyocytes were introduced into the perfusion chamber and
perfused with contraction buffer containing 2 mg/ml luciferin/luciferase reagent (Sigma). It was especially important to use freshly
diluted luciferin/luciferase reagent for each measurement because
luciferase activity is generally known to be very delicate and dependent on buffer composition. In particular, the high-chloride concentration of the cardiomyocyte buffer (140 mM NaCl) necessary for
cell viability constitutes suboptimal reaction conditions for the
enzyme (7).
Stimulation was performed as described above. For measurements
with ES cell-derived mouse cardiomyocytes grown on 35-mm cell
culture dishes, the culture medium was exchanged with medium
containing luciferin/luciferase reagent (2 mg/ml). Measurements were
started after the cells began to contract spontaneously. All measurements were performed with prewarmed buffers at 37°C.
Bioluminescence measurement on HUVECs under fluid shear
stress. Shear stress-induced ATP release from HUVECs was detected
either luminometrically or with a CCD camera. For luminometric
detection of ATP, HUVECs were grown to confluency on a gelatincoated cover slides (0.75 cm2) and loaded with recombinant luciferase. The slides were then inserted into a self-designed, parallel-plate
flow chamber that was integrated into a luminescence device
(Berthold Biolumat LB 9500). A peristaltic pump was used to repeatedly apply 5-s shear-stress pulses of 15 dyn/cm2 under constant
removal of the flow-through and online luminometric measurements
at 37°C. The cross-section of the flow chamber was 0.4 ⫻ 0.05 cm2.
Wall shear stress was calculated using the equation 3yQ/2a2b (22),
where y is the viscosity (poise) of the medium at 37°C, Q is the
volumetric flow rate (in ml/s), a is the half-channel height (in cm), and
b is the channel width (in cm). Luminescence rates were recorded
automatically in 2-s intervals.
For experiments performed using image analysis (CCD camera), a
cone-and-plate rheological device as described by Schnittler et al. (31)
was used to create defined levels of laminar shear stress (6) at 37°C in
a closed system. HUVECs were seeded onto cross-linked gelatincoated perfectly plan-parallel and round glass coverslips (diameter 28
mm, Assistent), grown to confluency, and loaded with recombinant
luciferase. The rheological device is equipped with a microscope
coupled to a Micromax 512 BFT System allowing online observation
of shear stress-exposed cells by time lapse recording. Basal luminescence was measured for 30 s before the beginning of the experiment.
Repeated 5-s shear stress impulses of 15 dyn/cm2 were applied to the
C611
C612
ATP RELEASE FROM CONTRACTING CARDIOMYOCYTES
cell culture dishes and used for CCD camera monitoring 4 –7
days after the appearance of spontaneously contracting cell
regions. At that time, cells had not yet formed a confluent
monolayer. Discrete clusters of cells could be distinguished, of
which ⬃50% showed spontaneous contraction. The representative experiments demonstrated in Fig. 2, A–C, show an ATP
signal over both cell clusters (Fig. 2B), whereas only cluster 1
showed contractile activity. Calculation of average intensities
obtained from three independent experiments revealed that
there was no difference between the marked regions, indicating
that the ATP detected was not released as a consequence of
contraction (Fig. 2D). Inhibition of cell contraction with 10
mM BDM (Fig. 2C) reduced light intensities. This effect,
however, can be fully explained by a direct inhibitory effect of
BDM on the luciferase observed in separate luminometric
measurements (data not shown). Release of intracellular ATP
by disruption of the cell integrity with digitonin was clearly
visible. Digitonin dose-dependently increased light intensities
3- and 25-fold over resting values (Fig. 2D).
Recent studies have shown that detection of ATP released
from astrocytoma cells in the bulk extracellular medium was
possible only after pharmacological inhibition of ecto-ATPase
activity. However, detection of submicromolar ATP concentrations in the cell surface environment by surface-targeted
luciferase was possible without pharmacological intervention,
indicating an efficient and rapid degradation of released ATP
by surface ectonucleotidases (17). Because ectonucleotidase
activity has been reported on the cardiomyocyte cell surface
AJP-Cell Physiol • VOL
(23, 44), we decided to bring luciferase molecules closer to the
region of potential ATP release, namely the sarcolemmal
membrane. For this purpose, we have targeted a recombinant
luciferase fusion protein containing four IgG binding motifs of
Staphylococcus aureus protein A to the cell surface by using a
chemically coupled WGA-IgG linker molecule. WGA is a
lectin that binds mainly to N-acetylglucosamine (GlcNAc)
residues from glycosylated molecules in the glycocalyx, and it
has been shown to bind to the surface of isolated rat cardiomyocytes (33). Figure 3 schematically depicts the principle of
the WGA-IgG-mediated luciferase binding. For measurement
of potential ATP release, isolated rat cardiomyocytes were
loaded with WGA-IgG and ZZ-luciferase, as described in
METHODS.
For CCD camera monitoring, regions were defined matching
the positions of rod-shaped, excitable cells (Fig. 4A). Light
intensities were sampled and corrected for luminescence values
from neighboring cell-free control regions. Bioluminescence
measurements were performed in the absence (Fig. 4B) and
presence of electrical stimulation (Fig. 4C). Altogether, 31
contracting and 26 resting cells taken from 5 independent
experiments were analyzed. The corrected light intensities
were 0.655 ⫾ 6.08 arbitrary units/pixel for resting cells and
⫺0.052 ⫾ 8.922 arbitrary units/pixel for contracting cells. No
increase in light intensity could be detected in the immediate
vicinity of contracting cells.
To confirm membrane binding of ZZ-luciferase, ATP was
added at the end of each experiment (Fig. 4D). To determine
289 • SEPTEMBER 2005 •
www.ajpcell.org
Downloaded from http://ajpcell.physiology.org/ by 10.220.33.4 on June 18, 2017
Fig. 1. Bioluminescence measurement over resting and contracting adult rat cardiomyocytes using soluble luciferase for ATP detection. A: lightmicroscopical image of two intact (regions 1 and
2) and one hypercontracted (region 3) adult rat
cardiomyocyte perfused with contraction buffer
containing luciferin/luciferase reagent. B: chargecoupled device (CCD) camera image of the same
cells without electrical stimulation. C: CCD camera image of the same cells under electrical
stimulation leading to contraction of cells 1 and
2 but not cell 3. D: average light intensities
measured over resting (n ⫽ 21) and contracting
(n ⫽ 16) cells from 5 independent experiments.
AU, arbitrary units. The scale bar in A is 25 ␮m.
ATP RELEASE FROM CONTRACTING CARDIOMYOCYTES
C613
the sensitivity of our system, the cardiomyocytes were loaded
with ZZ-luciferase and light emission was recorded over resting cells after the addition of various concentrations of ATP.
From the data shown in Fig. 5, it is evident that threshold
concentration for ATP producing measurable bioluminescence
was 3.3 ⫻ 10⫺5 M. However, the sensitivity was quite variable
in individual experiments, and in some experiments, even 330
nM could be clearly detected.
To demonstrate that membrane-targeted luciferase can sensibly detect cellular ATP release, we measured shear stressinduced ATP secretion reported for cultured HUVECs (3).
Figure 6A shows a typical luminometrical detection of shear
stress-induced ATP release from HUVECs loaded with membrane-bound luciferase. Sharp luminescence peaks can be observed immediately on stimulation. According to the literature
(3), repeated shear stress impulses caused reduced ATP signals. Perfusion with medium containing 1 ␮M ATP at the end
of the experiment showed that the observed ATP release falls
into the micromolar range. As shown in Fig. 6B, shear stressinduced ATP release from HUVECs can also be detected by
the CCD camera.
DISCUSSION
Fig. 3. Schematic diagram of luciferase targeting to the sarcolemmal membrane. GlucNac, N-acetylglucosamine; WGA, wheat germ agglutinin. For
details, see RESULTS. ?, ATP release is considered hypothetical.
AJP-Cell Physiol • VOL
Release of ATP from the contracting heart has been reported
in several studies (3, 9, 13). Besides nonmyocyte cells (reviewed in Ref. 37), cardiomyocytes have also been speculated
to be a source of extracellular ATP. Recently, Dutta and
coworkers (10) have described an ATP conductive pathway in
neonatal rat cardiomyocytes that responds to hypoxia and
ischemia but also mechanical stretch through osmotic swelling
with pronounced ATP release. The maxi-anion channel involved in this pathway is also present, albeit less frequently in
289 • SEPTEMBER 2005 •
www.ajpcell.org
Downloaded from http://ajpcell.physiology.org/ by 10.220.33.4 on June 18, 2017
Fig. 2. Bioluminescence measurement over
resting and contracting embryonic stem (ES)
cell-derived mouse cardiomyocytes using
soluble luciferase for ATP detection. A: lightmicroscopical image of one spontaneously
contracting (region 1) and one noncontracting cluster (region 2) of mouse cardiomyocytes genetically selected from ES cells, perfused with contraction buffer containing luciferin/luciferase reagent. B: CCD camera
image of the same cell regions. C: CCD
camera image of the same cell regions perfused with contraction buffer containing luciferin/luciferase reagent and 10 mM butanedione monoxime (BDM) to completely block
contraction. D: average light intensities measured over resting (n ⫽ 5) and contracting
(n ⫽ 11) cell regions from 3 independent
experiments. In 2 experiments, digitonin was
added to release intracellular ATP.
C614
ATP RELEASE FROM CONTRACTING CARDIOMYOCYTES
Fig. 5. Bioluminescence of cell surface attached luciferase bound to adult
rat cardiomyocytes. Bioluminescence of 26 quiescent and 31 contracting
cells from 5 independent experiments in luciferin-containing buffer was
determined by CCD camera monitoring. The sensitivity of the system was
determined by monitoring luciferase-loaded cells after addition of a range
of ATP concentrations. Numbers in parentheses refer to the cells monitored
for each ATP concentration, taken from 2–5 independent preparations.
ANOVA, post hoc test Bonferroni (P ⬍ 0.0001 for 10⫺4 and 10⫺3 M ATP);
unpaired Student’s t-test (P ⬍ 0.0001 for 3.3 ⫻ 10⫺5, 10⫺4, and 10⫺3 M
ATP) vs. quiescent cardiomyocytes.
AJP-Cell Physiol • VOL
adult rat cardiomyocytes. This study provides clear evidence
that mechanical stretch of the sarcolemmal membrane during
regular contraction is not associated with a detectable release
of ATP from isolated cardiomyocytes. This was measured at
the single cell level with soluble luciferase as well as luciferase
tagged to the cell surface of cardiomyocytes in combination
with a sensitive CCD camera allowing high-resolution detection.
Measurement of ATP release from a whole organ such as the
heart comprises a variety of different cell types, including,
aside from cardiomyocytes, nerve terminals, coronary endothelial cells, smooth muscle cells, and red blood cells, all of
which have been described to be potential sources of ATP (37).
It is therefore very difficult to clearly assign a measured
extracellular ATP signal to a defined cell type in experiments
using, e.g., isolated perfused hearts. Furthermore, experimental
manipulations like electrical stimulation almost inevitably lead
to the damage of an undefined amount of cells, which as a
consequence release their intracellular nucleotides to the surrounding medium. ATP within cardiomyocytes is normally
present in millimolar concentrations, so that few damaged cells
can give rise to a substantial extracellular ATP signal. Considerations similar to those for the intact organ apply to isolated
cells in culture. Also, in this reduced system, it is technically
very difficult to distinguish between a nonselective release
mechanism and a selective release from intact cells. In fact, in
initial experiments, we have observed extracellular ATP on
electrical stimulation when using luminometric detection of
289 • SEPTEMBER 2005 •
www.ajpcell.org
Downloaded from http://ajpcell.physiology.org/ by 10.220.33.4 on June 18, 2017
Fig. 4. Bioluminescence measurement over resting and contracting adult rat cardiomyocytes using cell surface-attached luciferase for ATP detection. A: light-microscopical image of adult rat
cardiomyocytes loaded with membrane attached
protein A-luciferase and perfused with contraction buffer containing luciferin. Scale bar is 25
␮m. B: CCD camera image of the same cells
without electrical stimulation. C: CCD camera
image of the same cells with electrical stimulation. D: CCD camera image showing the membrane-bound luciferase visualized by perfusion
with buffer containing luciferin and 100 ␮M ATP.
ATP RELEASE FROM CONTRACTING CARDIOMYOCYTES
signals derived from cardiomyocytes attached to coverslips.
However, we also observed that on electrical stimulation, some
cardiomyocytes hypercontracted and lost their rod-shaped
structure, indicative of cell damage and ATP release.
To circumvent these obvious problems, we have decided to
analyze ATP release by microscopical observation of single
cells. Only single cardiomyocytes showing regular contraction
on electrical stimulation were chosen to investigate the release
of ATP, thereby clearly excluding nonmyocytes, as well as
damaged cells as potential contaminating sources. The amount
of ATP released by a single cell experiment is naturally lower
than the quantities released by overall measurements in cell
culture. To compensate for this drop in signal intensity we
utilized a liquid nitrogen-cooled CCD camera, presently the
most sensitive device to detect low light intensities, together
with spatial allocation of the signal to a defined region: one
contracting cardiomyocyte.
For the detection of ATP we used the luciferin-luciferase
reaction known to be a highly sensitive method for the detection of ATP in aqueous solution. It should be noted, however,
that the enzyme activity is impaired by suboptimal reaction
conditions, such as a temperature of 37°C or physiological salt
solutions necessary to provide a normal physiological environAJP-Cell Physiol • VOL
ment for contracting cardiomyocytes. Luciferase activity has
been reported to have an optimal reaction temperature of 25°C
and to be very sensitive to salt conditions, such as high Cl⫺
concentrations (6). In our experiments, we have added high
concentrations of soluble luciferase to the supernatant (2 mg/
ml) to compensate for lowered enzyme activity. The limit of
detection in these experiments was ⬃100 nM ATP. In the
experiments with genetically selected, ES-cell-derived mouse
cardiomyocytes, there was always a detectable but contractionindependent release of ATP over the cells. Whether this basal
ATP release is a general feature of ES-cell derived cardiomyocytes or whether it relates to the genetic selection procedure
used is presently not known. Nevertheless, our detection system is sensitive enough to measure ATP release with high local
resolution.
Several possible mechanisms might have hampered the detection of released ATP by soluble luciferase. The released
ATP is rapidly diluted by diffusion and in addition is rapidly
degraded by ectonucleotidases present on the cardiomyocyte
surface (44). Furthermore, the release of ATP might be spatially confined to specific regions of the cell surface. Therefore,
measurement of ATP in the bulk supernatant might underestimate the actual ATP concentrations at the location of release,
namely the cell surface. We have thus attempted to measure
ATP by targeting the luciferase to the cell surface. To this end,
we have modified a method developed by Beigi et al. (2) using
antibodies against surface molecules of a target cell to tag a
luciferase-protein A fusion protein to the cell surface. This
elegant method, however, has several limitations, most prominently the availability of the appropriate antibodies against
surface properties of target cells. We therefore have used a
more general feature of cell surfaces to couple the luciferase to
the membrane, namely the appearance of glycosylated residues, which can be recognized by lectins. Because glycosylated residues are a general characteristic feature of cell surfaces of every cell type, lectins chemically cross-linked to
unspecific rabbit IgG instead of a specific antibody can be used
to target protein A-fusion proteins to any cell type. In our
particular case, we used WGA, a lectin specifically recognizing
N-acetylglucosamine residues. The WGA-IgG linker molecule
was then employed to target protein A-tagged luciferase to the
cell membrane of cardiomyocytes. With the use of this method,
we routinely were able to measure ATP when added in the
midmicromolar range; in two of five experiments, we clearly
detected 1 ␮M ATP. Beigi et al. (2) report in their work a
detection level in the midnanomolar range. It should be mentioned, however, that the cell numbers in these experiments
were between 1 ⫻ 106 (for promyelocytes) and 7 ⫻ 107 (for
platelets) for each measurement, whereas for reasons discussed
above, we were measuring single cells. The lower sensitivity of
the tagged luciferase for exogenously added ATP is most likely
due to the lower absolute quantity of enzyme molecules bound
to the cell surface. Occasionally, we observed light signals over
rounded and therefore damaged cells (see Fig. 1, region 3),
indicating leakage of intracellular ATP. The tagged luciferase,
however, is sufficiently sensitive to measure cellular ATP
secretion. Nonlytic release of ATP from HUVECs after stimulation with fluid shear stress served as a positive control and
clearly demonstrated that our detection system can measure the
release of “physiological” amounts of ATP at the cellular level.
289 • SEPTEMBER 2005 •
www.ajpcell.org
Downloaded from http://ajpcell.physiology.org/ by 10.220.33.4 on June 18, 2017
Fig. 6. Bioluminescence of shear stress-induced ATP release from human
umbilical vein endothelial cells (HUVECs) measured with cell surface-attached luciferase. A: luminometric detection of ATP released from HUVECs
loaded with membrane-attached protein A-luciferase and stimulated with 5-s
pulses of shear stress (15 dyn/cm2) in a flow-through system. At the end of the
experiment, the medium in the chamber was exchanged from basal medium to
basal medium containing 1 ␮M ATP. Luminescence was measured continuously in 2-s intervals. B: shear stress-induced ATP release from luciferasetagged HUVECs. Luminescence was measured with a CCD camera at selected
time points over 30 s.
C615
C616
ATP RELEASE FROM CONTRACTING CARDIOMYOCYTES
REFERENCES
1. Abood LG, Koketsu K, and Miyamoto S. Outflux of various phosphates
during membrane depolarization of excitable tissues. Am J Physiol 202:
469 – 474, 1962.
2. Beigi R, Kobatake E, Aizawa M, and Dubyak GR. Detection of local
ATP release from activated platelets using cell surface-attached firefly
luciferase. Am J Physiol Cell Physiol 276: C267–C278, 1999.
3. Bodin P, Bailey D, and Burnstock G. Increased flow-induced ATP
release from isolated vascular endothelial cells but not smooth muscle
cells. Br J Pharmacol 103: 1203–1205, 1991.
4. Borst MM and Schrader J. Adenine nucleotide release from isolated
perfused guinea pig hearts and extracellular formation of adenosine. Circ
Res 68: 797– 806, 1991.
5. Burnstock G. Purinergic nerves. Pharmacol Rev 24: 509 –581, 1972.
6. Buschmann MH, Dieterich P, Adams NA, and Schnittler HJ. Analysis
of flow in a cone-and-plate apparatus with respect to spatial and temporal
effects on endothelial cells. Biotechnol Bioeng 89: 493–502, 2005.
7. Denburg JL and McElroy WD. Anion inhibition of firefly luciferase.
Arch Biochem Biophys 141: 668 – 675, 1970.
8. Di Virgilio F, Mutini C, Chiozzi P, Falzoni S, Di Susino M, Sanz JM,
and Ferrari D. A purinergic hypothesis for immunomodulation. Ital
J Biochem 45: 195–203, 1996.
9. Doyle TB and Forrester T. Appearance of adenosine triphosphate in the
perfusate from working frog heart. Pflügers Arch 405: 80 – 82, 1985.
10. Dutta AK, Sabirov ZS, Uramoto H, and Okada Y. Role of ATPconductive anion channel in ATP release from neonatal rat cardiomyocytes in ischaemic or hypoxic conditions. J Physiol 559: 799 – 812, 2004.
11. Fischer Y, Becker C, and Loken C. Purinergic inhibition of glucose
transport in cardiomyocytes. J Biol Chem 274: 755–761, 1999.
12. Forrester T. An estimate of adenosine triphosphate release into the
venous effluent from exercising human forearm muscle. J Physiol 224:
611– 628, 1972.
13. Forrester T and Williams CA. Release of adenosine triphosphate from isolated
adult heart cells in response to hypoxia. J Physiol 268: 371–390, 1977.
14. Holmsen H. Platelet metabolism and activation. Semin Hematol 22:
219 –240, 1985.
15. Jans D, Srinivas SP, Waelkens E, Segal A, Lariviere E, Simaels J, and
Van Driessche W. Hypotonic treatment evokes biphasic ATP release
across the basolateral membrane of cultured renal epithelia (A6). J Physiol
545: 543–555, 2002.
16. Jin J, Daniel JL, and Kunapuli SP. Molecular basis for ADP-induced
platelet activation. II. The P2Y1 receptor mediates ADP-induced intracellular calcium mobilization and shape change in platelets. J Biol Chem 273:
2030 –2034, 1998.
17. Joseph SM, Buchakjian MR, and Dubyak GR. Colocalization of ATP
release sites and ecto-ATPase activity at the extracellular surface of human
astrocytes. J Biol Chem 278: 23331–23342, 2003.
18. Kaneda M, Fukui K, and Doi K. Activation of chloride current by
P2-purinoceptors in rat ventricular myocytes. Br J Pharmacol 111: 1355–
1360, 1994.
AJP-Cell Physiol • VOL
19. Klug MG, Soonpaa MH, Koh GY, and Field LJ. Genetically selected
cardiomyocytes from differentiating embryonic stem cells form stable
intracardiac grafts. J Clin Invest 98: 216 –224, 1996.
20. Kunapuli SP and Daniel JL. P2 receptor subtypes in the cardiovascular
system. Biochem J 336: 513–523, 1998.
21. Kuzmin AI, Lakomkin VL, Kapelko VI, and Vassort G. Interstitial
ATP level and degradation in control and postmyocardial infarcted rats.
Am J Physiol Cell Physiol 275: C766 –C771, 1998.
22. Lawrence MB, Smith CW, Eskin SG, and McIntire LV. Effect of
venous shear stress on CD18-mediated neutrophil adhesion to cultured
endothelium. Blood 75: 227–37, 1990.
23. Meghji P, Pearson JD, and Slakey LL. Regulation of extracellular
adenosine production by ectonucleotidases of adult rat ventricular myocytes. Am J Physiol Heart Circ Physiol 263: H40 –H47, 1992.
24. Mei Q and Liang BT. P2 purinergic receptor activation enhances cardiac
contractility in isolated rat and mouse hearts. Am J Physiol Heart Circ
Physiol 281: H334 –H341, 2001.
25. Parkinson PI. Proceedings: The effect of graduated exercise on the concentration of adenine nucleotides in plasma. J Physiol 234: 72P–74P, 1973.
26. Pearson JD and Gordon JL. Vascular endothelial and smooth muscle cells in
culture selectively release adenine nucleotides. Nature 281: 384–386, 1979.
27. Podrasky E, Xu D, and Liang BT. A novel phospholipase C- and
cAMP-independent positive inotropic mechanism via a P2 purinoreceptor.
Am J Physiol Heart Circ Physiol 273: H2380 –H2387, 1997.
28. Puceat M, Clement O, Scamps F, and Vassort G. Extracellular ATPinduced acidification leads to cytosolic calcium transient rise in single rat
cardiac myocytes. Biochem J 274: 55– 62, 1991.
29. Puceat M and Vassort G. Purinergic stimulation of rat cardiomyocytes
induces tyrosine phosphorylation and membrane association of phospholipase C␥: a major mechanism for InsP3 generation. Biochem J 318:
723–728, 1996.
30. Sauer H, Hescheler J, and Wartenberg M. Mechanical strain-induced
Ca2⫹ waves are propagated via ATP release and purinergic receptor
activation. Am J Physiol Cell Physiol 279: C295–C307, 2000.
31. Schnittler HJ, Franke RP, Akbay U, Mrowietz C, and Drenckhahn D.
Improved in vitro rheological system for studying the effect of fluid shear
stress on cultured cells. Am J Physiol Cell Physiol 265: C289 –C298, 1993.
32. Sprague RS, Ellsworth ML, Stephenson AH, and Lonigro AJ. ATP:
the red blood cell link to NO and local control of the pulmonary
circulation. Am J Physiol Heart Circ Physiol 271: H2717–H2722, 1996.
33. Stegemann M, Meyer R, Haas HG, and Robert-Nicoud M. The cell
surface of isolated cardiac myocytes: a light microscope study with use of
fluorochrome-coupled lectins. J Mol Cell Cardiol 22: 787– 803, 1990.
34. Stumpe T and Schrader J. Phosphorylation potential, adenosine formation, and critical PO2 in stimulated rat cardiomyocytes. Am J Physiol Heart
Circ Physiol 273: H756 –H766, 1997.
35. Stumpe T and Schrader J. Short-term hibernation in adult cardiomyocytes is PO2 dependent and Ca2⫹ mediated. Am J Physiol Heart Circ
Physiol 280: H42–H50, 2001.
36. Uozumi H, Kudoh S, Zou Y, Harada K, Yamazaki T, and Komuro I.
Autocrine release of ATP mediates mechanical stress-induced cardiomyocyte hypertrophy (Abstract). Circulation 98: I-624, 1998.
37. Vassort G. Adenosine 5⬘-triphosphate: a P2-purinergic agonist in the
myocardium. Physiol Rev 81: 767– 806, 2001.
38. Velazquez B, Garrad RC, Weisman GA, and Gonzalez FA. Differential
agonist-induced desensitisation of P2Y2 nucleotide receptors by ATP and
UTP. Mol Cell Biochem 206: 75– 89, 2000.
39. Vials AJ and Burnstock G. ATP release from the isolated perfused
guinea pig heart in response to increased flow. J Vasc Res 33: 1– 4, 1996.
40. Williams CA and Forrester T. Possible source of adenosine triphosphate
released from rat myocytes in response to hypoxia and acidosis. Cardiovasc Res 17: 301–312, 1983.
41. Yamada M, Hamamori Y, Akita H, and Yokoyama M. P2-purinoceptor
activation stimulates phosphoinositide hydrolysis and inhibits accumulation of
cAMP in cultured ventricular myocytes. Circ Res 70: 477– 485, 1992.
42. Yamazaki T, Komuro I, Kudoh S, Zou Y, Nagai R, Aikawa R, Uozumi
H, and Yazaki Y. Role of ion channels and exchangers in mechanical
stretch-induced cardiomyocyte hypertrophy. Circ Res 82: 430 – 437, 1998.
43. Zhang BX, Ma X, McConnell BK, Damron DS, and Bond M. Activation of purinergic receptors triggers oscillatory contractions in adult rat
ventricular myocytes. Circ Res 79: 94 –102, 1996.
44. Zinchuk VS, Okada T, and Kobayashi T. Ecto-ATPase activity in the
rat cardiac muscle: biochemical characteristics and histocytochemical
localisation. Cell Tissue Res 298: 499 –509, 1999.
289 • SEPTEMBER 2005 •
www.ajpcell.org
Downloaded from http://ajpcell.physiology.org/ by 10.220.33.4 on June 18, 2017
Baseline concentration of ATP in the interstitial space of the
rat heart has been determined to be ⬃40 nM (21). Given the
sensitivity of our measurements, it can be concluded that the ATP
concentration at the cell surface does not fall within the range of
EC50-ATP values reported for cardiomyocyte P2 receptors. In
case of P2X receptors EC50-ATP values between 1 ␮M for P2X1
and P2X2 and 20 ␮M for P2X4 were reported (37), whereas for
P2Y2 receptors the EC50-ATP values were found to be 3 ␮M for
P2Y1 (16) and in the range of 1.5–5.8 ␮M for P2Y2 (38).
In summary, the reported release of ATP from intact beating
hearts under normoxic conditions is unlikely to be derived from
cardiomyocytes. ATP is rather released from nonmuscle cells,
such as coronary endothelium or nerve terminals. Cardiomyocytes, however, may contribute to the general release of nucleotides under conditions of ischemia, anoxia, or other conditions of
cell damage, but not as a consequence of mechanical deformation
of the sarcolemmal membrane during contraction.