The proinflammatory cytokines TNF-alpha and IL-1 beta impair economy of contraction in human myocardium.

Abstract

Considerable experimental evidence has accumulated over the past years that proinflammatory cytokines, especially TNF-alpha and IL-1beta, impair myocardial function in different animal species. On the other hand, several prospective clinical trials studying TNF-alpha antagonist in patients with chronic heart failure were not able to demonstrate a benefit. As there might be a relevant species-related discrepancy, we intended to prove our previous results demonstrating impaired myocardial economy after exogenous administration of recombinant TNF-alpha in rat myocardium. In the present study, both TNF-alpha and IL-1beta not only revealed an immediate negative inotropic effect but also increased specific oxygen demand in human right-atrial myocardium. Enhanced oxygen consumption was not caused by an elevated basal metabolism but an impaired economy of contraction. Our results suggest that proinflammatory cytokines have a considerable effect on myocardial mechano-energetic parameters in human myocardium as well.

Full text

Page 1

Short Communication
The proinflammatory cytokines TNF-a and IL-1b impair economy
of contraction in human myocardium
U. Hofmanna,*, S. Heuera, K. Medera, J. Boehlerb, V. Langeb,
T. Quaschningc, G. Ertla, A. Bonza
aUniversity of Wurzburg, Medizinische Klinik und Poliklinik I, Abteilung Kardiologie, Germany
bUniversity of Wurzburg, Klinik fu ¨r Herz-und Thoraxchirurgie, Germany
cUniversity of Freiburg, Abteilung Nephrologie, Germany
Received 1 April 2007; received in revised form 8 June 2007; accepted 16 July 2007
Abstract
Considerable experimental evidence has accumulated over the past years that proinflammatory cytokines, especially TNF-a and IL-
1b, impair myocardial function in different animal species. On the other hand, several prospective clinical trials studying TNF-a antag-
onist in patients with chronic heart failure were not able to demonstrate a benefit. As there might be a relevant species-related discrep-
ancy, we intended to prove our previous results demonstrating impaired myocardial economy after exogenous administration of
recombinant TNF-a in rat myocardium. In the present study, both TNF-a and IL-1b not only revealed an immediate negative inotropic
effect but also increased specific oxygen demand in human right-atrial myocardium. Enhanced oxygen consumption was not caused by an
elevated basal metabolism but an impaired economy of contraction. Our results suggest that proinflammatory cytokines have a consid-
erable effect on myocardial mechano-energetic parameters in human myocardium as well.
? 2007 Elsevier Ltd. All rights reserved.
Keywords: Proinflammatory cytokines; Myocardium; Energy metabolism; TNF-a; IL-1b
1. Introduction
Cardiac diseases associated with cytokine activation
include ischemia-reperfusion [1], heart failure [2], myocar-
ditis [3], cardiac allograft rejection [4] and sepsis-associated
cardiac dysfunction [5]. There is evidence from a broad
range of in vivo and ex vivo studies that proinflammatory
cytokines elicit acute as well as delayed negative inotropic
effects, cardiomyocyte apoptosis, interstitial fibrosis and
ventricular dilation [6–8]. We have previously shown that
TNF-a impairs chemo-mechanical energy transduction in
rat myocardium [9]. Several prospective clinical trials failed
to show a clinical benefit of new TNF-a antagonizing drugs
in patients with chronic heart failure [10,11]. Instead, trials
have been stopped prematurely as soon as a potential ben-
efit became unlikely. In the ATTACH trial, treatment with
infliximab in high dose was even associated with worsening
of the clinical status [10].
It has been suggested that available TNF-a antagonists
might even magnify the toxicity of TNF-a. Etanercept, a
TNF-receptor fusion protein, may stabilize circulating
TNF-a leading to cardiac toxicity by retaining complexes
of cytokine and antagonist-drug in the circulation [12]. Inf-
liximab, the other compound used in clinical studies, can
induce cell lysis in cells expressing transmembrane TNF-a
[13]. Another likely hypothesis is a species-related differ-
ence of cytokine effects on myocardium. We followed this
hypothesis and studied the effect of TNF-a on function
and chemo-mechanical energy transduction of human
myocardial tissue. As TNF-a is not the only proinflamma-
tory cytokine impairing myocardial function, we further
1043-4666/$ - see front matter ? 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.cyto.2007.07.185
*Corresponding author. Fax: +49 931 201 36280.
E-mail address: hofmann_u2@klinik.uni-wuerzburg.de (U. Hofmann).
www.elsevier.com/locate/issn/10434666
Cytokine 39 (2007) 157–162

Page 2

tested the hypothesis that IL-1b might show similar effects
on human myocardium.
2. Materials and methods
2.1. Human atrial muscle preparations
Human right-atrial tissue was derived from clinically
stable patients undergoing elective coronary bypass surgery
(n = 17), aortic valve replacement (n = 1) or aortic graft
implantation (n = 1). Mean age was 62 (range from 37 to
77), 14 patients were male and 5 female. All patients were
in clinical heart failure class NYHA I–II, no EF < 30%
was documented by echocardiography or laevocardiogra-
phy. Patients with atrial fibrillation were excluded from
the study. Patients received beta-blockers 62%, ACE-inhib-
itors or aldosteron receptor blockers 46%, statins 31%, oral
anti-diabetic drugs 23%, insulin 8%. To establish extracor-
poreal circulation, a section of the right atrial appendage is
removed to insert the venous cannula. This resected piece
of atrial myocardium was used for the present study. All
patients individually gave their informed written consent
at least one day before surgery. The use of human myocar-
dial tissue was reviewed and approved by the Committee of
Human Ethics of the University of Wurzburg. The study
conforms to the Code of Ethics of the World Medical
Association (Declaration of Helsinki).
2.2. Simultaneous measurement of isometric force and
oxygen consumption
The tissue was immediately stored in a pre-oxygenated
ice-cold Krebs–Henseleit solution containing 30 mmol of
2,3-butanedione. Further preparation of muscle fibers
was performed in this solution to reduce cutting injury.
The experimental setup (Scientific Instruments, Heidelberg,
Germany) and the fiber mounting procedure was described
in detail previously [9]. In brief, the muscle preparation
(diameter 200–400 lm) is fixed by two tweezers, one con-
nected to a force transducer, the other connected to a
micrometer screw, which allows stretch of the preparation.
Within the specially designed perfusion chamber that main-
tains a constant temperature of 37 ?C in the perfusate
(Krebs–Henseleit buffer containing (mmol/l) 152.0 Na+,
3.6 K+, 135.0 Cl?, 25.0 HCO3?, 1.3 HPO4
0.6 Mg2+, 1.25 Ca2+, 11.2 glucose), a Clark-type oxygen
probe is placed perpendicular to the long axis of the mus-
cle. The setup allows the simultaneous determination of
the oxygen partial decrease during a short period of perfu-
sion stop. From these data, the oxygen consumption is cal-
culated. Preparations were stimulated electrically 25%
above threshold. Recording of isometrically developed
force, force–time-integral (FTI) and other mechanical
parameters is done by ‘‘twitch’’ software (Scientific Instru-
ments). FTI (in N · s/min/mm2) is defined as the area
between peak systolic force and diastolic force during the
stimulation interval. It represents an equivalent of work
2?, 0.6 SO4
2?,
in isometrically contracting myocardium and is a major
determinant of MVO2[14].
After equilibration each muscle preparation was step-
wise stimulated from 60 to 140 beats per minute in order
to acquire baseline force data. Thereafter, this protocol
was repeated twice, after 10 min and 30 min continuous
perfusion with the respective cytokine. Simultaneously to
mechanical data acquisition, oxygen consumption was
measured by a Clark-type oxygen probe. FTI was then
plotted against MVO2. The slope of the MVO2–FTI regres-
sion line (Fig. 3c) reflects chemo-mechanical force trans-
duction by cross-bridge cycling and relates mechanical
output to chemical energy input. Its reciprocal indicates
the contractile economy. Y-axis intercept indicates basal
oxygen consumption.
2.3. Materials
All buffer ingredients were from Sigma (Deisenhofen,
Germany), recombinant cytokines were supplied by R&D
systems (Wiesbaden, Germany). Cytokines were solubi-
lized as indicated by the manufacturer and diluted to final
concentrations in Krebs–Henseleit buffer.
2.4. Statistical analysis
Statistical analysis was performed using WinStat soft-
ware (Benecke & Schwippert, Staufen, Germany), implying
Excel spread sheets. Significance was determined by Wilco-
xon signed-rank test for paired data and by Mann–Whit-
ney U-test for unpaired data. Multiple comparisons were
done by ANOVA analysis for repeated measurements. P-
values < .05wereconsidered
expressed as means ± SE.
significant.Dataare
3. Results
3.1. Effects on isometric force
As depicted in Fig. 1, both TNF-a and IL-1b immedi-
ately reduced developed force of human myocardial tissue.
The effect increased during continuous stimulation with
cytokine: At 60 beats per minute stimulation frequency
TNF-a (5 ng/ml) reduced systolic force generation to
64 ± 4% of baseline (after 10 min) and 51 ± 4% (after
30 min), respectively.
IL-1b (1 ng/ml) led to a comparable time-dependent loss
in contractility. Force development was reduced to
61 ± 8% and 53 ± 8% of baseline at equivalent time points.
Both cytokines impaired developed force throughout the
full stimulation range, from 60 to 140 beats per minute.
Cytokines did not affect force–frequency relationship
(Fig. 1). Diastolic force was not changed by TNF-a alone
or with additional IL-1b (Fig. 2).
As depicted in Fig. 2, adding 1 ng/ml IL-1b to atrial
muscle preparations being already exposed to 5 ng/ml
TNF-a for 20 min did not further reduce developed force.
158
U. Hofmann et al. / Cytokine 39 (2007) 157–162

Page 3

0
0.5
1
1.5
2
2.5
3
6080100120 140/ min
mN / mm2
*
*
*
*
*
0
1
2
3
4
5
6080100 120140
/ min
mN / mm2
* *
* * *
Fig. 1. Human atrial preparations (n = 7) were electrically paced with increasing stimulation frequencies. After each change in stimulation frequency they
were allowed to equilibrate, then developed force was measured. The stimulation protocol was repeated after 10 (grey bars) and 30 (white bars) min
continuous superfusion with 5 ng/ml TNF-a (a) or 1 ng/ml IL-1b (b).*p < 0.05 vs. baseline (black bars).
0
20
40
60
80
100
120
baseline 5 min. 10 min.15 min.20 min. 25 min. 30 min. 35 min.40 min.
*
**
*
TNF
TNF + IL-1
%
Fig. 2. Human atrial preparations (n = 5) were electrically stimulated (120/min). Baseline developed force was determined after equilibration. Addition of
5 ng/ml TNF-a produced a significant effect on developed force (white bars,*p < 0.05). Then myocardial preparations were additionally stimulated with
1 ng/ml IL-1b. Stimulation with both TNF-a, and IL-1b evoked no significant further deterioration in mechanical performance. Both cytokines did not
affect diastolic force (black bars).
U. Hofmann et al. / Cytokine 39 (2007) 157–162
159

Page 4

3.2. Effects on myocardial oxygen consumption
There was a highly linear correlation of MVO2and FTI,
both before and after cytokine stimulation in each experi-
ment (Fig. 3b). Both cytokines increased specific oxygen
consumption(MVO2/FTI).
MVO2–FTI regression lines of preparations treated with
5 ng/ml TNF-a or 1 ng/ml IL-1b indicated impaired
chemo-mechanical economy of contraction. The effect on
economy was time-dependent becoming statistically signif-
icant after 30 min of cytokine stimulation. Basal oxygen
Increasedslopes ofthe
consumption (y-axis intercepts) was not altered (data not
shown) Fig. 4.
4. Discussion
The main result of the study is that both TNF-a and
IL-1b reduce contractile function and impair cardiac
chemo-mechanical energy transduction in human myocar-
dium. To our best knowledge, this is the first experimental
study demonstrating the effect of proinflammatory cyto-
kines on human myocardial energy metabolism.
0
2
4
6
8
10
12
14
60 80100 120140 / min
60 80100120140/ min
developed force [mN/mm²]
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
0.045
0 50 100150200 250300
FTI [N x s /mm2x min]
MVO2 [ml/mm3xmin]
Fig. 3. Representative experiment. (a) After equilibration, each muscle preparation was consecutively stimulated from 60 to 140 beats per minute in order
to acquire baseline force data (black bars). Thereafter, this protocol was repeated twice, after 10 min (grey bars) and 30 min (open bars) continuous
superfusion with the respective cytokine. (b) Original registration of the baseline force–frequency relationship. (c) Simultaneously to mechanical data
acquisition, oxygen consumption was measured by a Clark-type oxygen probe to determine MVO2–force–time-integral (FTI) relationship. Data points
reflect repeated measurements of FTI (mN s/mm2· min) and MVO2(ml/min) at stimulation frequencies ranging from 60 to 140 beats per minute under
baseline conditions (circle symbol, r2= 0.94), 10 (triangle symbol, r2= 0.96) and 30 min (square symbol, r2= 0.99) after start of superfusion with IL-1b,
respectively. The slope of the MVO2–FTI regression line reflects chemo-mechanical force transduction by cross-bridge cycling and relates mechanical
output to chemical energy input. Its reciprocal indicates contractile economy.
160
U. Hofmann et al. / Cytokine 39 (2007) 157–162

Page 5

We chose concentrations presumably representing clini-
cally relevant doses which have caused comparable effects
on contractility in rat myocytes [15]. Confirming our previ-
ous findings in animal studies [9], the present results sug-
gest that both cytokines have a relevant acute impact on
human myocardial performance.
Acute myocardial upregulation of TNF-a and IL-1b
during infarction was demonstrated in different animal spe-
cies and man [16]. Not only acute experimental coronary
ligation but also microembolization led to myocardial
TNF-a release and contractile dysfunction in an in vivo
animal model [17]. We showed in patients that even elective
percutaneous coronary interventions induced a significant
release of TNF-a measurable in peripheral blood [18], sup-
portive of a potential relevance of cytokine mediated
impairment of mechanical function in ischemic heart
disease.
While the negative inotropic effect of TNF-a has previ-
ously been established, we demonstrate in the present study
impairment of myocardial economy by TNF-a and IL-1b.
This adds to its detrimental effects in situations with limited
oxygen supply, such as acute coronary events.
Because of the restricted availability of human tissue we
were not able to elucidate the downstream events leading to
the mechano-energetic effects we observed here. However,
there are results from experimental studies in animal cells
indicating that impaired contractile economy might be
explained by alterations in calcium handling shifting the
balance of cytosolic calcium extrusion from sarcoplasmatic
reticulum calciumuptake
removal, which is energetically more wasteful [19–21].
Accordingly, Miyano et al. could demonstrate in an iso-
lated blood perfused canine heart preparation that oxygen
cost of contractility was increased by TNF-a infusion. As
TNF-a did not affect the response to extracellular calcium,
they concluded that not a decrease in calcium sensitivity
but a change in intracellular calcium handling might be
responsible [22].
There are two main limitations of the present study that
have to be addressed: Obviously the protocol does not
allow concluding on chronic effects. Sustained cytokine
stimulation might downregulate or desensitize receptors
thus the detrimental effects of acute cytokine exposure
might be mitigated. For the lack of enough material, we
could further not perform experiments in human left-ven-
tricular tissue. It is not quite clear whether the results
obtained from atrial tissue can be extrapolated to ventric-
ular myocardium. However, we were not aware of any
data concerning differences in the response to cytokines
between atrial and ventricular myocardium. The congru-
ence of the present data with our previous findings from
rat ventricular myocardium [9] further suggests that our
present results should be applicable to human ventricular
tissue as well.
The finding that TNF-a and IL-1b severely impair
chemo-mechanical energy transduction and contractility
in human myocardium implicates clinical significance.
First, the present study confirms that TNF-a effects
observed in several animal studies should be relevant in
human myocardium as well. Second, IL-1b representing
just another member of a still growing family of proin-
flammtory cytokines showed similar effects on human
myocardium. This suggests that targeting just one member
of the cytokine family might fail. In addition, in acute car-
diac injury the relative importance of TNF-a within the
cytokine activation pattern might be different from that
in chronic disease states like heart failure. In chronic rheu-
matoid arthritis, TNF-antagonist represented a new mile-
stone in therapy [23]. Patients with rheumatic arthritis
have substantial increase in risk for cardiovascular disease
including heart failure [24,25] which has not been consid-
ered in the context of cytokine pathophysiology and
potential therapy.
towardtrans-sarcolemmal
Acknowledgements
The study was supported by a grant from the Deutsche
Stiftung fu ¨r Herzforschung to A. Bonz and the Deutsche
Forschungsgemeinschaft, SFB 355 ‘‘Pathophysiologie der
Herzinsuffizienz’’.
0
5
10
15
20
25
30
35
base lineTNF 10'TNF 30'
economy [N*s*mm/ ml]
*
0
5
10
15
20
25
30
baselineIL-1 10' IL-1 30'
economy [N*s*mm/ ml]
*
Fig. 4. The reciprocal of the MVO2–FTI slope indicates economy of
contraction. Electrically stimulated contracting human myocardial muscle
strip preparations were allowed to equilibrate, then baseline data were
gathered. Thereafter, continuous superfusion with 5 ng/ml TNF-a or 1 ng/
ml IL-1b was started. After 30 min both TNF-a (a) and IL-1b (b) showed
a significant effect on myocardial economy (n = 7, p < 0.05 vs. baseline).
U. Hofmann et al. / Cytokine 39 (2007) 157–162
161