688
Increased Messenger RNA Level of the
Inhibitory G Protein a Subunit Gia,2
in Human End-Stage Heart Failure
Thomas Eschenhagen, Ulrike Mende, Monika Nose, Wilhelm Schmitz, Hasso Scholz,
Axel Haverich, Stefan Hirt, Volker Doring, Peter Kalmair,
Wolfgang Hoppner, and Hans-Jorg Seitz
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
In human heart failure the positive inotropic and cAMP-elevating effects of both j,-adrenoceptor agonists
and phosphodiesterase inhibitors are diminished. This has been explained at least in part by an increase
in the inhibitory signal-transducing G protein (G1) and unchanged stimulatory G protein (Gs). In the
present study we determined the mRNA expression pattern of the ce subunits of Gp.l, Gi.2, Gi-3, and Gs in
myocardial tissue samples of patients undergoing heart transplantation. Northern blot analysis of total
RNA extracted from left ventricles with 32P-labeled cDNAs demonstrated expression of Gi..2, ,.3, and
G.. mRNA. In contrast, Gi,,l mRNA was not detectable. To investigate whether the increased ratio of
Gi/G, might be due to altered gene expression, we compared mRNA levels of Q,,.2, Ga.3, and Gsa in left
ventricular myocardium from failing hearts with idiopathic dilated cardiomyopathy (n=8) and ischemic
cardiomyopathy (n=6) and from nonfailing hearts from transplant donors (n=8). Compared with
nonfailing control hearts, the Gia.2 mRNA was increased by 75+±26% (p<O.05) in idiopathic dilated
cardiomyopathy hearts and 90±26% (p<0.05) in ischemic cardiomyopathy hearts. Gi. 3 and Gsa mRNA
levels were similar in the three groups. The results suggest that as in other mammalian species, Gia2 and
Gia3 mRNA are the predominant Gia mRNA subtypes in human ventricular myocardium. An upregulation
of Gi.,2 but not of Giar.3 mRNA probably underlies the increase in Gia protein and might thus be involved
in the pathophysiological process leading to reduced responsiveness to cAMP-increasing agents in
end-stage heart failure. (Circulation Research 1992;70:688-696)
KEY WORDS * heart failure * G proteins * mRNA expression
In studies of ventricular preparations from patients
undergoing heart transplantation, it has been
shown that the positive inotropic and cAMPelevating effects of both 8-adrenoceptor agonists and
phosphodiesterase inhibitors are diminished in the failing heart. In contrast, drugs bypassing the ,3-adrenoceptor/adenylate cyclase signaling pathway, such as dibutyryl-cAMP, dihydroouabain, and Ca>2, exhibit an
unchanged positive inotropic effect.1-3 These results
demonstrate the ability of failing myocardium to exert a
maximal contractile response and suggest a defective
adenylate cyclase stimulation with deficient cAMP production as an important cause of decreased myocardial
contractility.4.5
It is well established that 13-adrenoceptor downregulation occurs in heart failure and might partly explain
From the Abteilung Allgemeine Pharmakologie (T.E., UM.,
M.N., W.S., H.S.) Abteilung ffir Herz-, Thorax- und GefaBchirurgie
(V.D., P.K.), and Abteilung Endokrinologische Biochemie (W.H.,
H.-J.S.), Universitats-Krankenhaus Eppendorf, Universitat Hamburg, and Abteilung fiOr Herz-, Thorax- und GefaBchirurgie, Medizinische Hochschule Hannover (A.H., S.H.), FRG.
Supported by the Deutsche Forschungsgemeinschaft.
Address for reprints: Dr. Thomas Eschenhagen, Abteilung
Allgemeine Pharmakologie, Universitats-Krankenhaus Eppendorf, Universitat Hamburg, MartinistraBe 52, W-2000 Hamburg
20, FRG.
Received July 24, 1990; accepted November 21, 1991.
the diminished inotropic and cAMP response to
,B-adrenoceptor agonists.6,7 However, there is increasing
evidence for an important additional role of G proteins
in regulating adenylate cyclase activity and hence inotropy in human end-stage heart failure. G proteins are
heterotrimeric proteins composed of a, ,3, and y subunits, which occupy a key regulatory role as transducers
of many pathways: Gs couples 81 - and 32-adrenoceptors;
H2 receptors and prostaglandin E2 receptors couple to
the adenylate cyclase in a stimulatory way; Gi couples
a2-adrenoceptors, M2-cholinoceptors, and A1-adenosine
receptors to the adenylate cyclase in an inhibitory way.8
Evidence for an additional direct regulation of cardiac
Ca>2 and Na+ channels or K' channels by a subunits of
G, or Gi, respectively, further underlines the importance
of G proteins in various signal-transducing pathways.9
Recently, a 35-40% increase in the 40-kd a subunit
of Gi in failing human hearts with idiopathic dilated
cardiomyopathy (IDC) has been demonstrated by pertussis toxin-catalyzed ADP-ribosylation.10-12 Studies
with cholera toxin-catalyzed labeling revealed unchanged Gs, levels.0,13 Furthermore, reconstitution experiments in cyc- cells showed that the functional
activity of Qsa is also unchanged in the failing human
heart.10 Alterations in the G protein amount and/or
function could account for the discrepancies between
changes in ,32-adrenoceptor density and adenylate cy-
Eschenhagen et al G Protein mRNA Expression in Human Heart Failure
clase-stimulating and positive inotropic effects of
adrenoceptor agonists in humans14-16 as well as in
animal models.'7 Furthermore, they might explain the
reduced efficacy of receptor-independent phosphodiesterase inhibitors in heart failure.
By now, more than 12 different G protein subunits
have been identified18; at least seven of these (four Gsa
gene splice variants and one each of Gi,1, Gia2, and
Gia3) are thought to be involved in the adenylate cyclase
signaling pathway. There is no clearcut evidence so far
for functional differences between the four Gs subunits or the three Gi subunits; equipotent effects on
the adenylate cyclase and Ca21 channels (G,a) or on
atrial K' channels (Gia) have been shown.9 However,
tissue-specific and developmental differences in the
expression of the three inhibitory G protein subunits
suggest that differences in the coupling to receptors or
effectors might exist.'9,20
Because pertussis toxin-catalyzed ADP-ribosylation
does not discriminate between different Gia subunits, it
is currently not known which of the three Gia subunits is
responsible for the measured increase of the 40-kd
signal in heart failure. With the use of cDNA probes the
present study focused on the following main questions:
1) Is the increase in Gia protein accompanied by an
increase in GiQ mRNA levels in the failing heart? 2) Are
Gsa mRNA levels changed in the failing heart? 3) Are
there differences in Ga mRNA levels between IDC and
ischemic cardiomyopathy (ICM)? Therefore, we characterized the Ga mRNA expression pattern qualitatively
in the human heart with cDNA probes encoding Gai,,
Gia2, Gjia3, and Gsa and quantified Gia2, Gia3, and Gsa
mRNA levels in left ventricular myocardium from patients with IDC and ICM compared with nonfailing
human myocardium.
12-
a
a
a
a
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
Materials and Methods
Cardiac Tissue
Failing human hearts were obtained from patients
undergoing heart transplantation because of end-stage
heart failure caused by IDC (n=8) or ICM (n=6). The
diagnosis of ischemic heart disease was confirmed by
left heart catheterization and coronary angiography.
Patients with other forms of cardiomyopathy, such as
muscular dystrophy or acute myocarditis, were excluded
from the study. All of the patients were classified as
New York Heart Association (NYHA) class IV with
markedly abnormal pretransplant hemodynamics (Table 1). There were no significant differences between
the two groups in mean age, mean cardiac index, mean
pulmonary capillary wedge pressure, or mean left ventricular ejection fraction. Before cardiectomy all patients received conventional medical treatment including cardiac glycosides, organic nitrates, diuretics,
angiotensin converting enzyme inhibitors, and other
vasodilators in varying combinations. No patient reor f3-adrenoceptor antagoceived catecholamines,
nists, or phosphodiesterase inhibitors. Nonfailing control hearts (n=8) were obtained from prospective
multiple organ donors whose hearts could not be transplanted for surgical reasons or because of blood group
incompatibility. Aortic and pulmonary valves were excised and later used for valve replacements. Some of
these patients had a history of hypertension or had
a-
689
TABLE 1. Hemodynamic Data From Patients
Sex
Dilated
M
M
M
M
F
M
M
M
Mean
+SEM
Age
NYHA
(years)
cardiomyopathy
49
46
61
42
38
34
41
54
IV
IV
IV
IV
IV
IV
IV
IV
45.6
3.1
Ischemic cardiomyopathy
IV
M
41
IV
M
51
IV
M
50
IV
M
56
IV
M
54
IV
M
53
PCW
(mm Hg)
CI
EF
(%)
(1/min. m')
28
19
18
30
11
31
24
21
10
16
17
14
19
19
22
13
1.4
2.1
2.4
2.0
1.7
2.0
1.9
1.6
22.8
2.4
16.3
1.4
1.9
0.1
19
32
25
22
27
35
11
14
23
18
23
17
1.5
1.6
2.3
2.0
2.7
1.7
17.7
2.0
26.7
0.2
2.0
2.5
NYHA, functional class of the New York Heart Association;
PCW, pulmonary capillary wedge pressure; EF, radionuclidedetermined left ventricular ejection fraction; CI, cardiac index.
Mean
-+-SEM
50.8
2.1
arteriosclerosis of great vessels on surgical inspection
(Table 2). However, none had a history of heart failure,
and they were thus classified as nonfailing donors. The
mean age of organ donors did not differ significantly
from heart recipients but did tend to be lower. Before
explantation, seven of eight patients received low doses
of dopamine to maintain renal function; one of them
also received dexamethasone. Another patient received
a single dose of an a-adrenoceptor antagonist. Written
informed consent was obtained from all patients or the
family of prospective heart donors before cardiectomy.
Tissue samples from the free walls of the left
ventricles were obtained at the time of explantation
and rapidly frozen in liquid N2. Care was taken not to
take scarred, fibrotic, or adipose tissue, endocardium,
epicardium, or great vessels. Tissue samples were
stored at -80°C until use.
Total RNA Preparation
Preparation of total RNA was performed according
to the method of Auffray and Rougeon (procedure C)21
with modifications that have been previously described
in detail.22 Samples (1-1.5 g) of frozen tissue were
minced to fine powder in liquid N2 with a mortar,
homogenized in 4 M LiCl/8 M urea with a polytron
homogenizer (PT 10-35, Kinematica GmbH, Lucerne,
Switzerland) for 15 seconds at setting 5, and incubated
overnight at 4°C. Total cellular RNA was precipitated
by centrifugation at 17,000g for 30 minutes at 0°C. After
a proteinase K (Boehringer Mannheim, Mannheim,
FRG) digestion in 0.01 M Tris and 0.5% sodium dode-
690
Circulation Research Vol 70, No 4 April 1992
TABLE 2. Clinical and Anamnestic Data From Nonfailing Heart Donors and Steady-State Levels of Gi 2, Gi...3, and
GS, mRNA
Age
(years)
43
M
F
56
46
Diagnosis
CB
SA bleeding
Cerebral venous
thrombosis,
cerebral edema
Polytrauma, CB
CA. CB
M
36
CA, CB
M
41
CB
F
27
CB
Sex
M
M
M
44
53
Others
Fatty liver, obesity*
Fatty liver*
Suspicion of cerebral
cancer*
Arteriosclerosis
Obesity, hypertension,
mild arteriosclerosis
Arteriosclerosis, nicotine
abuse
Wernicke
encephalopathy*
Drug addiction*t
Drugs at HTX
DA 50 mg/hr
DA 6 mg/hr
DA 50 mg/hr
Dexamethasone 20
mg bolus
DA 50 mg/hr
DA 6 mg/hr
Gia 2
GM3
2.37
3.05
2.39
6.53
3.68
8.96
DA 6 mg/hr
3.64
DA 6 mg/hr
1.81
PB 1.5 mg/kg bolus
Mean:
±SEM
2.94
2.79
2.25
3.89
14.1
13.1
6.66
15.6
Gsa
2.74
6.21
3.27
2.58
4.64
4.82
3.41
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
7.43 2.92
9.5
3.8
0.25C
1.5
0.45
Given are the diagnoses leading to clinical death, other diagnoses in patients' history or from pathological
examination, medication at time of heart transplantation, and steady-state levels of Ga2, GMa3, and Gsa mRNA per
amount of total RNA in arbitrary units as described in "Materials and Methods." HTX, heart transplantation; CB,
intracerebral bleeding; SA bleeding, subarachnoidal bleeding; CA, aneurysm of cerebral artery; DA, dopamine; PB,
phenoxybenzamine.
*Patient's history stated that there were no clues for any disease of the heart or circulatory system.
tDrugs used were alcohol, diamorphine, and codeine.
cyl sulfate (SDS) (pH 7.5) at 37°C for 2 hours followed
by three phenol/chloroform extractions, the RNA was
precipitated with 0.2 M sodium acetate and 70% ethanol overnight, followed by centrifugation at 10,000g for
30 minutes at -10C. The RNA pellet was dissolved in
Tris/EDTA buffer. RNA concentration was determined
photometrically at 260 nm in triplicate. The mean
optical density ratio (OD260/OD280) was 2.08±0.04
(n = 22). Total RNA yield was not significantly different
between controls, IDC, and ICM with mean values
(micrograms per gram tissue wet weight) of 238±54
(n =8), 336±45 (n =8), and 377+43 (n =6), respectively.
RNA (15-20 ,ug) was denatured with 42% formamide and 5.8% formaldehyde at 95°C for 2 minutes and
size-fractionated by electrophoresis in 1% agarose gels
containing 0.5 gg/ml ethidium bromide (Fluka Chemie,
Buchs, Switzerland). RNA base length was determined
with an RNA ladder from Boehringer Mannheim. RNA
was transferred to Hybond N nylon membranes (Amersham, Braunschweig, FRG) by Northern blot capillar
transfer overnight using 20x SSC (3 M NaCl and 0.3 M
sodium citrate, pH 7.0) as the transfer medium. Transfer
was controlled on an ultraviolet transilluminator by staining the blot membrane with 0.05% methylene blue according to Herrin and Schmidt.23 For quantification of specific
Ga mRNA levels, total RNA (5 ,ug for determinations of
Gi,2 and Gs5 and 10 jig for Gia 3) was denatured with 50%
formamide and 5.9% formaldehyde for 5 minutes at 56°C
and applied to Hybond N nylon membranes by vacuum
filtration with a Bio Dot SF slot blot apparatus (Bio-Rad
Laboratories, Richmond, Calif.).
cDNA Probes
Plasmids (pGEM-2) with cDNA inserts for rat Gi,1,
Gia2, Gja3, and G,a were gifts from Dr. R.R. Reed.24 A
plasmid (bluescribe) with the cDNA insert for human
Gia3 was supplied by Dr. E.J. Neer.25 The plasmids were
transformed into Escherichia coli, and positive clones
were picked and grown in rich medium. Large-scale
preparation was performed according to the alkaline
lysis method with cesium chloride (Paesel GmbH,
Frankfurt, FRG) density gradient centrifugation. Inserts were cut with EcoRI (Boehringer Mannheim) and
gel-purified. Insert sizes and purity of the cDNA were
determined in 1% agarose gels with a DNA molecular
weight standard from Boehringer Mannheim. Sizes
were as follows: 1,950 base pairs (bp) (Gia 1), 1,750 bp
(Gia 2), 3,070 bp (Gia 3), 620±1,120 bp (EcoRI fragments
of Gs5), and 2,200 bp (human Gia-3). The full-length
cDNA of Ga-2 and the 1,120-bp cDNA fragment of G,,,
were used for probing.
Because of a 94% nucleotide homology in the coding
regions between rat Giail and Gia3,24 we used a 600-bp
Xba IlEcoRI fragment of the 3' noncoding end of Gi,ai
and a 625 -bp EcoRVIEcoRI fragment of the 3' noncoding end of rat Gia3 for hybridization analysis. Comparative probing with the full-length cDNA of both Gjai
and Gia3 revealed identical results. Because the rat Gia
probe did not hybridize with total RNA extracted from
human heart, all hybridizations with human hearts were
performed with an EcoRI/HindIII (550-bp coding region) restriction fragment of the human Gia3 probe.
Analysis with an Xba I/EcoRI fragment (1,000-bp 3'
noncoding end) revealed identical results (not shown).
The cDNA probes were 32p labeled (nick translation
kit, Amersham Buchler, Braunschweig, FRG) with
[32P]dCTP (3,000 Ci/mmol, New England Nuclear-Dupont, Bad Homburg, FRG) to a specific activity of
32-8.0x 108 dpm/,ug. Unbound radioactivity was separated by gel filtration with Sephadex G-50 DNA grade
(Pharmacia Fine Chemicals, Uppsala, Sweden).
Hybridization Procedure
Blot membranes were prehybridized at 42°C overnight in a solution containing 50% formamide, 5 x
Eschenhagen et al G Protein mRNA Expression in Human Heart Failure
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
Denhardt (1 mg/ml Ficoll, polyvinylpyrrolidone, and
bovine serum albumin), 0.9 M NaCl, 0.06 M NaH2PO4,
0.006 M EDTA, 0.1% SDS, and 200 ,ug/ml tRNA from
yeast. Radiolabeled probes were added to fresh hybridization solution (prehybridization solution) in a concentration of 1-2x 106 dpm/ml. Hybridization was performed in 50 ,l/cm2 at 42°C for 48 hours. After
hybridization, blot membranes were washed briefly with
2x SSC and 0.1% SDS, followed by 10 minutes of
incubation in 2x SSC and 0.1% SDS at room temperature and three 20-minute washes in 0.2x SSC and
0.1% SDS at 65°C. Wet blot membranes were sealed in
plastic wrap and exposed to medium-sensitive medical
x-ray film (R2, 3M, Italy) for 1-20 days at -80°C by
using intensifier screens. For subsequent hybridizations
with other radiolabeled cDNAs, the former probe was
removed by pouring boiling 0.1% SDS over the blot
membrane and allowing it to cool down to room temperature. Filters were exposed to highly sensitive x-ray
film (X-OMAT AR, Kodak) for 48-96 hours to verify
complete removal of the probe. It has been shown in
preliminary experiments that this time of exposure was
sufficient to control removal of the former probe.
Quantification of Specific mRNA
Hybridization intensity of autoradiographic signals on
slot blots was measured quantitatively by two-dimensional densitometry (TLC II, CAMAG, Berlin, FRG).
All determinations were done in triplicate. Total RNA (5
,ug each) was applied for determination of Gi,2 and Gsa
mRNA levels; 10 ,ug total RNA was used for Gi,a3 mRNA
since lower signal intensities were expected. Slot blots
were hybridized sequentially two or three times. Control
hybridizations with the same probe had been performed
before and revealed reproducibility of signal ratios on
one blot up to the third or fourth hybridization.
The intensity of autoradiographic signals of identical
RNA samples on different blots varied between films
because of minor differences in amount and specific
activity of the labeled cDNA probe and exposure time
of the films. Therefore, blots were standardized by a
total RNA standard (prepared in milligram quantities
from failing human hearts). Autoradiographic density
of hybridization signals was plotted versus the applied
amount of total RNA standard (Figure 1). A curvefitting by a second-order polynomial was performed
(using SIGMAPLOT [Jandel Scientific, Corte Madera,
Calif.]) to determine the relative amount of specific
mRNA for all other RNA samples on this blot in
arbitrary units (anx axis value of 1 [1 jug RNA standard]
was defined as 1 arbitrary unit).
Nonspecific binding was determined by parallel probing of rRNA (Pharmacia) on each blot in triplicate.
Nonspecific binding was seen only with G,a cDNA.
Hybridization signals amounted to maximally 15% of
sample values and were substracted from all other
signals on the blot.
To correct measurements for minor differences in the
amount of total RNA applied on the blot membrane
and for possible differences in the amount of intact
poly(A+) RNA in the RNA samples, all blots were
reprobed with 32P-labeled oligo[d(T)12-18] (Pharmacia).
Oligo[d(T)12-18] was radiolabeled with [32P]dTTP to a
specific activity of 5x109 dpm/,ug by using terminal
691
hi
._
L-O
00
a
._
r=O.990
5
7
9
11
RNA- standard (pg)
FIGURE 1. Plot of densitometrically determined autoradiographic density versus the amount of total RNA standard
applied (in micrograms) on slot blots. Shown is a representative result of hybridization of total RNA from human left
ventricular myocardium with radiolabeled Gsa cDNA under
high stringency conditions as described in "Materials and
Methods." The relative amount of Gs, mRNA in all other
individual RNA samples on the blot was calculated by
referring the autoradiographic density of each point to the
standard curve. The standard curve is a computed secondorder regression graph.
1
3
deoxyribonucleotidyl transferase (Pharmacia) according to Hoppner et al.26 Hybridization was performed
under the same conditions as with Ga cDNA except for
the lack of formamide in the hybridization solution.
Washing of the blots was performed at a final stringency
of 0.2x SSC and 0.1% SDS at 42°C for 1 hour. Blots
were exposed on medium-sensitive film for 5-16 hours,
and autoradiographic signals were quantified as described for Ga mRNA measurements. There was no
correlation between G, mRNA or poly(A+) RNA levels
and patient age or sex or the duration of storage of the
frozen tissue (not shown).
Statistics
Values presented for steady-state levels of specific
mRNAs are arithmetic mean+SEM. Statistical significance was estimated using Student's t test for
unpaired observations. A value of p<0.05 was considered significant.
Results
Northern Blot Analysis
Northern blots of 20 ,ug total RNA extracted from
human ventricular myocardium and, for comparison,
from rat ventricular myocardium or rat brain are shown
in Figure 2. In human and rat myocardium, a nicktranslated 32P-labeled cDNA probe encoding rat Gia-2
hybridized to a predominant band at 2.4 kilobases (kb).
The rat Gia cDNA hybridized to a 3.5 kb mRNA in total
RNA extracted from rat ventricular myocardium and rat
brain but did not detect any specific signal in total RNA
from human myocardium (not shown). In contrast,
hybridization with two different restriction fragments of
a human Gia3 cDNA resulted in a predominant band at
2.8 kb and two faint bands at about 4.0 and 1.7 kb in
human ventricular myocardium (Figure 2 shows hybridization with the HindIII/EcoRI fragment). There was
692
Circulation Research Vol 70, No 4 April 1992
Ge -mRNA
Giu-2-mRNA I total RNA
G i <-2
Gjc_-3
GsOG
kIb
A 6- -200
U)
1 50
c
-74
-5.3
28S18S-
e_le
-2.8
wI
100
a
2.0
50o
-1.9
- 1.6
n]
U
I
T
- I _1
L ii
l
R-
h.-+
;.
. .
t,.,,,
i.,, ,,.+,,+.,.,++b
F.< ',+* j.*.j*
*-w *
oUs{
*+
,-* +', ++'2,++*,
so -N w+, ++ *
"-+ +4
'.-"'-'*' "'
- ' -"'- ",
R.. 5 +uX
...^A
*E *
Gjo-2- mRNA poly AV RNA
B
HH RH
HH RH RB
HH RH
HH RH
FIGURE 2. Northern analysis of 20 gg total cellular RNA
from failing human left ventricular myocardium (HH) with
idiopathic dilated cardiomyopathy, rat ventricular myocarDownloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
dium (RH), and rat brain (RB). Hybridization was performed with `P-radiolabeled cDNA encoding rat Gi,,l
(600-bp Xba Ifragment), rat Gia-2 (full-length), human Gia13
(550-bp HindIII fragment), and Gsa (1,100-bp EcoRI fragment) under high stringency conditions. Autoradiographic
2 days (Ga,), 5 days
exposure time was 10 days (Gsa
(Gsa 9, and 1 day (GSaY). The positions of the 28S and 18S
rRNA and of the RNA size ladder in kilobases (kb) are
indicated.
only weak hybridization to a faint band at about 3.5 kb
in rat heart total RNA. G.i. mRNA was not detectable
in human or rat ventricular myocardium even after a
long exposure time of 10 days (Figure 2). There was,
however, hybridization of the Gia- cDNA to a single
band at about 3.5 kb in total RNA from rat brain serving
as a positive control (Figure 2). The Gsa cDNA detected
one predominant band at 1.9 kb and a small but clearly
distinguishable band at about 1.8 kb in human myocardium. Interestingly, hybridization of total RNA from rat
myocardium demonstrated the smaller Gsa mRNA band
at 1.8 kb to be the predominant message. In both human
and rat tissues a further signal at about 4.5 kb could be
seen after long exposure time (not distinguishable in
Figure 2), which is in accordance with a detailed analysis
of Gsa mRNA subtype expression in human and dog
myocardium.27 There were no qualitative differences between failing and nonfailing myocardium (not shown).
Moreover, probing of total RNA extracted from human
failing and nonfailing atria revealed an expression pattern
similar to that in ventricular myocardium (not shown).
Slot Blot Quantification of Specific mRNA Levels
Relative concentrations of Gia2, Gia3, and Gs, mRNAs
in myocardium from failing and nonfailing hearts (Figures
3-5) were measured on slot blots. In failing hearts, the
steady-state level of GQ2 mRNA (Figure 3A) was increased by 75±26% in IDC (n=8, 4.95±0.65 arbitrary
units) and increased by 90±26% in ICM (n=5, 5.31 +0.72
arbitrary units) compared with nonfailing hearts (n =8,
2.79±0.25 arbitrary units). In contrast, steady-state levels
of Gi,3 and Gs, mRNA were not significantly different
within the three groups studied (Figures 4A and 5A).
To correct for minor differences in the applied
amount of intact mRNA per sample, individual values
in
_
a
._
NF
IDC
ICM
NF
P<O.OS
FIGURE 3. Graph showing steady-state levels of Gia2
mRNA in failing human hearts with idiopathic dilated cardiomyopathy (IDC) and ischemic cardiomyopathy (ICM) and
in nonfailing hearts (NF). Panel A: Ga-2 mRNA level per
total RNA. Panel B: Gia-2 mRNA level per amount of
VS.
poly(A ') RNA, which was determined by reprobing slot blots
with 2P-labeled oligo[d(T)/. The ordinates give mRNA levels
in arbitrary units and in percent of the Gia2- mRNA level in
nonfailing hearts. Arbitrary units were calculated by referring
individual autoradiographic densities on slot blots to an RNA
standard curve as depicted in Figure 1 and described in
"Materials and Methods." Values are mean-+-SEM. The
number of hearts is indicated in the columns.
obtained from hybridization with Ga-2, Gia3, and Gs,,
mRNA were referred to autoradiographic density values from hybridization with 32P-labeled oligo[d(T)]
(panel B of Figures 3-5). These experiments revealed
almost identical results.
Discussion
Ga mRNA Expression Pattern
The present study demonstrates expression of mRNA
encoding subunits of the inhibitory G proteins Gia-2
and Gia-3 and of the stimulatory G protein GQ, in human
ventricular myocardium. The results suggest that Gia-2
and Gia3 are the predominant Gsa mRNA subtypes in
quantity, whereas Gsa mRNA was not detectable. Gsa
eDNA detected one predominant band at 1.9 kb and a
small but clearly distinguishable band at about 1.8 kb in
the human heart. In contrast, hybridization with total
RNA from rat heart revealed the smaller Gsa mRNA
band at 1.85 kb to be the predominant message. The two
bands in human myocardium probably reflect different
splicing products of the unique Gsa gene, four of which
are known so far: Gsa +2 corresponding to M, 52,000
a
Eschenhagen et al G Protein mRNA Expression in Human Heart Failure
693
Gsu - mRNA / total RNA
Gid-3-mRNAI total RNA
A
A
U,
fl._
.,
a
a0
1-
4
b-
Gijd-3-rmRNA I poly A* RNA
0_100
Gsd - mRNA I poly A' RNA
B
V)
c
b..
a
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
2
50
b.
n
NF
ICM
IDC
FIGURE 4. Graphs showing steady-state levels of Gia3
mRNA in failing human hearts with idiopathic dilated cardiomyopathy (IDC) and ischemic cardiomyopathy (1CM) and
in nonfailing hearts (NE). Panel A: Relative G,a 3 mRNA level
per total RNA. Panel B: Relative Gia3 mRNA level per
amount ofpoly(A+) RNA, which was determined by reprobing
slot blots with 32P-labeled oligo[d(T)]. The ordinates give Gia3
mRNA levels in arbitrary units and in percent of the Gia3
mRNA level in nonfailing hearts. Arbitrary units were calculated by referring individual autoradiographic densities on slot
blots to an RNA standard curve as depicted in Figure 1 and
described in "Materials and Methods." Values are
mean ±SEM. The number of hearts is indicated in the
columns.
protein and Gsa3±4 corresponding to Mr 45,000 protein.27
In both human and rat myocardium, a further signal at
about 4.5 kb could be seen after long exposure time (not
shown), which has previously been described in total
RNA from human ventricles.27 The identity of the
4.5-kb message is currently not known.
The Ga mRNA expression pattern in human and rat
ventricular myocardium (as well as the expression pattern in human atria [not shown]) and the base length of
the different mRNAs reported here are consistent with
reports of rat, hamster, and canine myocardium.20 24 2829
In these animal species, Gja 2 has been reported to be the
predominant Gia, subtype in the heart, and Gja 3 has been
consistently found at low expression levels. The lack of
Gi,,. is consistent with the reported mRNA expression
pattern in adult human30 and rat20 24 myocardium.
There are, however, discrepancies concerning the
expression level and the message size of Giar3 mRNA in
the human heart. Feldman et aP30 reported Gia3 mRNA
to be the predominant message of Gia proteins in human
left ventricles. In this study, a rat Gia3 cDNA was used
for Northern analysis. The same probe did not result in
any hybridization with human total RNA in our exper-
1
NF
IDC
1CM
FIGURE 5. Graphs showing relative levels of Gs, mRNA in
failing human hearts with idiopathic dilated cardiomyopathy
(IDC) and ischemic cardiomyopathy (ICM) and in nonfailing
hearts (NF). Panel A: Relative Gsa mRNA level per total
RNA. Panel B: Relative Gs, mRNA level per amount of
poly(A+) RNA, which was determined by reprobing slot blots
with 32P-labeled oligo[d(T)]. The ordinates give Gsa mRNA
levels in arbitrary units and in percent of GS,X mRNA level in
nonfailing hearts. Arbitrary units were calculated by referring
individual autoradiographic densities on slot blots to an RNA
standard curve as depicted in Figure 1 and described in
"Materials and Methods." Values are mean ±SEM. The
number of hearts is indicated in the columns.
iments, irrespective of the cDNA fragment or the tissue
(ventricles, atria, liver, kidney) studied (not shown).
Without giving exact information on the size of the
cDNA (fragment) used, the authors described an unusual, short Gia3 mRNA at 1.85 kb. In other studies, a
larger Gi,3 mRNA has been reported. The group that
cloned the human Gia3 cDNA and screened different
human tissues described low levels of Gia.3 mRNA in
fetal atrium and determined a message size of 2.8 kb.25
In rat, hamster, and canine heart, Gia3 mRNA was
consistently reported to be 3.5 kb in base length.2428'29
In accordance with these studies, we determined a
predominant Gia3 message size of 2.8 kb in human
myocardium and one of about 3.5 kb in rat myocardium.
The nature of the two additional faint bands in human
myocardium at about 4.0 and 1.7 kb is not known but
might result from cross-hybridization to other members
of the family of highly conserved G protein and subunits. Because of the use of rat cDNA for hybridization
analysis of Gi,2 and human cDNA in the case of Gia3,
the intensity of hybridization signals cannot be compared directly with each other. Thus the present study
does not allow a clear statement of whether Gia2 or Gia3
mRNA is the major Gia subtype in human myocardium.
However, approximately similar signal intensities of
694
Circulation Research Vol 70, No 4 April 1992
Gia-2 and Gita3 hybridization were seen in spite of the
nonhomologous hybridization and a shorter exposure
time for Gia-2 (Figure 2). Therefore, quantitative predominance of Ga2 mRNA may be suggested in human
myocardium. This would be in agreement with antibody
studies of bovine31 and human12 myocardium, which
described at the protein level the majority of G,a protein
to be of the Gia2 subtype.
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
GaQ mRNA Levels: Failing Versus Nonfailing
Human Hearts
The main result of the present study is that steadystate levels of Gia2 mRNA were increased in failing
human hearts compared with nonfailing hearts. This
suggests that the increase in Gi, protein in failing heart
membranes10-12 might be due to increased Gia2 mRNA
levels and thereby due to enhanced de novo synthesis.
Gia 2 mRNA levels were similarly increased in different
causes of heart failure as IDC and ICM. This suggests
that alterations in the expression of Gia, protein-might be
a secondary phenomenon, independent from the pathogenesis of heart failure.
Our finding that GMa2 mRNA levels are increased
both in IDC and ICM was surprising, since results
obtained from pertussis toxin labeling of failing human
heart membranes from our own and other laboratories"1',2 revealed an increased Gi,a protein level in IDC
and an unchanged Gi, protein level in ICM compared
with nonfailing controls. This discrepancy cannot
readily be explained. However, it has been reported in
invertebrates (sea urchin) that mRNA can be stored in
an inactive form temporarily,32 indicating that mRNA
levels do not necessarily directly correlate to translational activity. In addition, there is some evidence for a
physiological role of endogenous ADP-ribosylation
and/or phosphorylation of Ga subunits. Tanuma et a133
purified an endogenous ADP-ribosyltransferase C,
which ADP-ribosylates a cysteine residue in Gi, as does
pertussis toxin. Watanabe et a134 reported that phosphorylation of Gia by the cAMP-dependent protein
kinase A resulted in a decrease in the ADP-ribosylation
by pertussis toxin. These results indicate that the accessibility of Gia for pertussis toxin can be influenced by
posttranslational modifications and hence pertussis toxin-catalyzed ADP-ribosylation does not necessarily reflect the true protein level of GQ. This is probably the
main reason for the discrepancies between mRNA and
pertussis toxin measurements since recent preliminary
results, obtained by using antibodies for quantification
of Gia proteins, showed increased Gi, in both IDC and
ICM.35
What are the possible mechanisms for upregulation of
Gia2 mRNA levels? One of the common features in
end-stage heart failure is an elevated plasma catecholamine level.36 It is noteworthy in this context that a
consensus sequence of a "cAMP response element" has
been described in the promoter region of the Gia2 gene37
and is lacking in the Gsa gene.38 Thus, it seems reasonable
to assume that in severe heart failure an increased adrenergic drive via the cAMP cascade might cause an increased transcription rate of the Gi,2 gene and hence
increased Gi,2 mRNA levels. Such a mechanism would
also explain the lack of an increase in Gs mRNA demonstrated in the present study. Support for this hypothesis
comes from a recent study in which chronic /3-adrenergic
stimulation under in vivo conditions caused a selective
increase in myocardial Gi, and Ga,3 mRNA and unchanged Gsa, mRNA steady-state levels in rats.22
In light of these results, it was surprising that in the
present study Gia3 mRNA levels were not increased in
parallel with Gia2. Furthermore, this finding is in apparent contrast to the study of Feldman et al,3" who
reported an increase of Gia3 mRNA in failing human
hearts. In addition to the methodological problems
discussed above, it is interesting that in the study of
Feldman et al all Ga mRNA levels investigated were
significantly increased by 60-90% (Gia-3 and G5a) or
tended to be higher (Ga,, 36%). In this study normalization of dot blots was performed by reprobing the
blots with a ,3-actin probe, although it is not known if
alterations in the expression of /B-actin occur in human
heart failure. Indeed, regulation of /3-actin mRNA
levels is suggested by isoform shifts in the actin and
myosin filament family in experimentally induced cardiac hypertrophy and/or failure39,40 and by developmental decreases in the expression of cardiac f3-actin.41
Thus, discrepancies between studies may result from
the general problem of unavailability of appropriate
standards for quantification. At present, this problem
is not resolved satisfactorily. We used standardization
of blots with oligo[d(T)] and yielded nearly identical
results as by referring individual mRNA values to the
amount of total RNA. Our finding of an unchanged
Gia3 mRNA suggests that Ga-3 is either not regulated
at all or that mechanisms other than cAMP-mediated
increases in gene transcription are involved in the
regulation of Gi, subunits in human heart. Distinct
regulation of Gia2 and Gja3 mRNA levels like those
shown here suggests different functions of the two Gi,a
subunits in the pathogenesis of heart failure. It would
be important in this context to investigate with subtype-specific antibodies the proportion of Gia-2 and
Gia 3protein in the upregulation of Gi,r in heart failure.
This needs further elucidation.
It might be argued that the results of the present study
were compromised by an inhomogenous control group
containing patients with a history of hypertension and
patients who received drugs at time of transplantation
(Table 2). In addition, the influence of cardioactive drugs
such as cardiac glycosides and angiotensin converting
enzyme inhibitors on Q, mRNA levels in the IDC and
ICM groups cannot be excluded. This is certainly true.
However, as seen in Table 2 there was no correlation
between any anamnestic data and Ga mRNA levels in the
control group. Furthermore, neither now nor in the
future will biochemical studies be possible on hearts from
untreated patients. Moreover, because of an efficient
transplantation program in Europe, the availability of
nonfailing control hearts will always be extremely limited
for research without any objections against its use as a
donor organ.
In conclusion, the present study demonstrates a significant increase of mRNA levels of Gia2 in human
end-stage heart failure caused by IDC and ICM without
changes in Gia-3 and Gsa mRNA levels. The increase in
Gia2 mRNA very likely underlies the increase in Gja
protein in end-stage heart failure, which at least in part
explains deficient cAMP production by the adenylate
cyclase and the diminished positive inotropic effect of
Eschenhagen et al G Protein mRNA Expression in Human Heart Failure
f3-adrenoceptor agonists and phosphodiesterase inhibitors. Because the Gi,2 gene is under cAMP-dependent
control, it is suggested that the increased adrenergic
drive might increase Gia2 gene expression via cAMP.
Upregulation of Gia might thus play an important
pathophysiological role as a basic mechanism of adaptation of the myocardial cell to increased adrenergic
stimulation in severe heart failure.
16.
17.
18.
Acknowledgments
We are grateful to Dr. R.R. Reed (Baltimore, Md.) and Dr.
E.J. Neer (Boston, Mass.) for providing the rat and human Gi,
cDNAs, respectively, and to Dr. H.H. Arnold (Hamburg,
FRG) for his methodical advice. We also thank Ms. A. Diiwel
and Ms. A. Harneit for their generous and excellent help in
various technical aspects of the work.
19.
20.
21.
References
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
1. Bristow MR, Ginsburg R, Minobe W, Cubiccioti RS, Sageman WS,
Lurie K, Billingham ME, Harrison DC, Stinson EB: Decreased
catecholamine sensitivity and beta-adrenergic receptor density in
failing human hearts. N Engl J Med 1982;307:205-211
2. Schmitz W, Scholz H, Erdmann E: Effects of alpha- and betaadrenergic agonists, phosphodiesterase inhibitors and adenosine
on isolated human heart muscle preparations. Trends Pharmacol
Sci 1987;8:447-450
3. Bdhm M, Beuckelmann D, Brown L, Feiler G, Lorenz B, Nabauer
M, Kemkes B, Erdmann E: Reduction of beta-adrenoceptor
density and evaluation of positive inotropic responses in isolated,
diseased human myocardium. Eur Heart J 1988;9:844-852
4. Feldman MD, Copelas L, Gwathmey JK, Phillips P, Warren SE,
Schoen FJ, Grossman W, Morgan JP: Deficient production of
cyclic AMP: Pharmacologic evidence of an important cause of
contractile dysfunction in patients with endstage heart failure.
Circulation 1987;75:331-339
5. Danielsen W, Leyen vd H, Meyer W, Neumann J, Schmitz W,
Scholz H, Starbatty J, Stein B, Ddring V, Kalmar P: Basal and
isoprenaline-stimulated cAMP content in failing versus nonfailing
human cardiac preparations. J Cardiovasc Pharmacol 1989;14:
171-173
6. Fowler MB, Laser JA, Hopkins GL, Minobe W, Bristow MR:
Assessment of the beta-adrenergic receptor pathway in the intact
human heart: Progressive receptor down-regulation and subsensitivity to agonist response. Circulation 1986;74:1290-1302
7. Brodde OE, Schiller S, Kretsch R, Brinkmann M, Borst HG,
Hetzer R, Reidemeister JC, Warnecke H, Zerkowski HR:
Regional distribution of beta-adrenoceptors in the human heart:
Coexistence of functional betal- and beta2-adrenoceptors in both
atria and ventricles in severe congestive cardiomyopathy.
J Cardiovasc Pharmacol 1986;8:1235-1242
8. Neer EJ, Clapham DE: Roles of G protein subunits in transmembrane signalling. Nature 1988;333:129-134
9. Brown AM: Regulation of heartbeat by G protein-coupled ion
channels. Am J Physiol 1990;259:H1621-H1628
10. Feldman AM, Cates AE, Veazey WB, Hershberger RE, Bristow
MR, Baughman KL, Baumgartner WA, Van Dop C: Increase in
the 40,000-mol wt pertussis toxin substrate (G-protein) in the
failing human heart. J Clin Invest 1988;82:189-197
11. Neumann J, Schmitz W, Scholz H, von Meyerinck L, Doring V,
Kalmar P: Increase of myocardial Gi-proteins in human heart
failure. Lancet 1988;2:936-937
12. Bdhm M, Gierschik P, Jakobs KH, Pieske B, Schnabel P, Ungerer
M, Erdmann E: Increase of Gi, in human hearts with dilated but
not ischemic cardiomyopathy. Circulation 1990;82:1249-1265
13. Schnabel P, Bohm M, Gierschik P, Jakobs KH, Erdmann E:
Improvement of cholera toxin-catalysed ADP-ribosylation by
endogenous ADP-ribosylation factor from bovine brain provides
evidence for an unchanged amount of Q,, in failing human
myocardium. J Mol Cell Cardiol 1990;22:73-82
14. Bristow MR, Port JD, Sandoval AB, Rasmussen R, Ginsburg R,
Feldman AM: Beta-adrenergic receptor pathways in the failing
human heart. Heart Failure 1989;5:77-90
15. Kaumann AJ, Hall JA, Murray HJ, Wells FC, Brown MJ: A
comparison of the effects of adrenaline and noradrenaline on
human heart: The role of ,3- and f32-adrenoceptors in the stimu-
22.
23.
24.
25.
26.
27.
695
lation of adenylate cyclase and contractile force. Eur Heart J
1989;10(suppl B):29-37
Hall JA, Kaumann AJ, Brown MJ: Selective /I-adrenoceptor
blockade enhances positive inotropic responses to endogenous
catecholamines mediated through P2-adrenoceptors in human
atrial myocardium. Circ Res 1990;66:1610-1623
Vatner DE, Young MA, Knight DR, Vatner S: ,3-Receptors and
adenylate cyclase: Comparison of nonischemic, ischemic and postmortem tissue. Am J Physiol 1990;258:H140-H144
Gilman A: Regulation of adenylyl cyclase by G proteins, in
Nishizuka Y (ed): The Biology and Medicine of Signal Transduction.
New York, Raven Press, Publishers, 1990, pp 51-57
Foster KA, McDermott PJ, Robishaw JD: Expression of G proteins in rat cardiac myocytes: Effect of KCl depolarization. Am J
Physiol 1990;259:H432-H441
Luetje CW, Tietje KM, Christian JL, Nathanson NM: Differential
tissue expression and developmental regulation of guanine nucleotide binding regulatory proteins and their messenger RNAs in rat
heart. J Biol Chem 1988;263:13357-13365
Auffray C, Rougeon F: Purification of mouse immunoglobulin
heavy-chain messenger RNAs from total myeloma tumor RNA.
Eur J Biochem 1980;107:303-314
Eschenhagen T, Mende U, Nose M, Schmitz W, Scholz H, Wfistel
JM: Isoprenaline-induced increase in mRNA levels of inhibitory
G-protein a-subunits in rat heart. Naunyn Schmiedebergs Arch
Pharmacol 1991;343:609-615
Herrin DL, Schmidt GW: Rapid, reversible staining of northern
blots prior to hybridization. BioTechniques 1988;6:196-198
Jones DT, Reed R: Molecular cloning of five GTP-binding protein
cDNA species from rat olfactory neuroepithelium. J Biol Chem
1987;262:14241-14249
Kim S, Ang SL, Bloch DB, Bloch KD, Kawahara Y, Tolman C, Lee
R, Seidman JG, Neer EJ: Identification of cDNA encoding an
additional alpha-subunit of a human GTP-binding protein: Expression of three alphai subtypes in human tissues and cell lines. Proc
Natl Acad Sci USA 1988;85:4153-4157
Hdppner W, Rasmussen UB, Abuerreish G, Wohlrab H, Seitz HJ:
Thyroid hormone effect on gene expression of the adenine nucleotide transporter in different rat tissues. Mol Endocrinol 1988;2:
1127-1131
Ishikawa Y, Bianchi C, Nadal-Ginard B, Homcy CJ: Alternative
promoter and 5' exon generate a novel Q,, mRNA. J Biol Chem
1990;265:8458-8462
28. Katoh Y, Komuro I, Takaku F, Yamaguchi H, Yazaki Y: Messenger RNA levels of guanine nucleotide-binding proteins are
reduced in the ventricle of cardiomyopathic hamsters. Circ Res
1990;67:235-239
29. Holmer SR, Stevens S, Homcy CJ: Tissue- and species-specific
expression of inhibitory guanine nucleotide-binding proteins. Circ
Res 1989;65:1136-1140
30. Feldman AM, Cates AE, Bristow MR, Van Dop C: Altered
expression of alpha-subunits of G-proteins in failing human hearts.
JMol Cell Cardiol 1989;21:359-365
31. Mumby SM, Pang IH, Gilman AG, Sternweis PC: Chromatographic resolution and immunological identification of the alpha40
and alpha41 subunits of guanine nucleotide-binding regulatory
proteins from bovine brain. J Biol Chem 1988;263:2020-2026
32. Showman RM, Wells DE, Anstrom J, Hursh DA, Raff R: Message
specific sequestration of maternal histone mRNA in the sea urchin
egg. Proc Natl Acad Sci U S A 1982;79:5944-5947
33. Tanuma S, Kawashima K, Endo H: Eukaryotic mono(ADPribosyl)transferase that ADP-ribosylates GTP-binding regulatory
G, protein. J Biol Chem 1988;263:5485-5489
34. Watanabe Y, Imaizumi T, Misaki N, twakura K, Yoshida H: Effects
of phosphorylation of inhibitory GTP-binding protein by cyclic
AMP-dependent protein kinase on its ADP-ribosylation by pertussis
toxin, islet activating protein. FEBS Lett 1988;236:372-374
35. Bohm M, Larisch K, Erdmann E: Quantification of Gi, by a novel
radioimmunoassay in cardiomyopathic human hearts (abstract).
Naunyn Schmiedebergs Arch Pharmacol 1991;343(suppl):R54
36. Cohn JN, Levine TB, Olivari MT, Garberg V, Lura D, Francis GS,
Simon AB, Rector T: Plasma norepinephrine as guide to prognosis
in patients with congestive heart failure. N Engl J Med 1984;311:
819-823
37. Weinstein LS, Spiegel AM, Carter AD: Cloning and characterization of the human gene of the a-subunit of Gi2, a GTP-binding
signal transducing protein. FEBS Lett 1988;232:333-340
696
Circulation Research Vol 70, No 4 April 1992
38. Kozasa T, Itoh H, Tsukamoto T, Kaziro Y: Isolation and characterization of the human G,a gene. Proc Natl Acad Sci U S A
1988;85:2081-2085
39. Bugaisky LB, Anderson PG, Hall RS, Bishop SP: Differences in
myosin isoform expression in the subepicardial and subendocardial
myocardium during cardiac hypertrophy in the rat. Circ Res
1990;66:1127-1132
40. Bakerman PR, Stenmark KR, Fisher JH: a-Skeletal actin messenger RNA increases in acute right ventricular hypertrophy. Am J
Physiol 1990;258:L173-L178
41. Lohse P, Arnold HH: The down-regulation of the chicken cytoplasmic ,B-actin during myogenic differentiation does not require
the gene promotor but involves the 3' end of the gene. Nucleic
Acids Res 1988;16:2787-2803
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
Increased messenger RNA level of the inhibitory G protein alpha subunit Gi alpha-2 in
human end-stage heart failure.
T Eschenhagen, U Mende, M Nose, W Schmitz, H Scholz, A Haverich, S Hirt, V Döring, P
Kalmár and W Höppner
Circ Res. 1992;70:688-696
doi: 10.1161/01.RES.70.4.688
Circulation Research is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1992 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7330. Online ISSN: 1524-4571
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circres.ahajournals.org/content/70/4/688
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Circulation Research can be obtained via RightsLink, a service of the Copyright Clearance Center, not the
Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further information
about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation Research is online at:
http://circres.ahajournals.org//subscriptions/