Clinical and Experimental Pharmacology and Physiology (2015) 42, 874–880
doi: 10.1111/1440-1681.12415
ORIGINAL ARTICLE
Direct inhibition, but indirect sensitization of pacemaker activity to sympathetic tone by the interaction of endotoxin with HCN-channels Henning Ebelt,* Isabel Geißler,* Sara Ruccius,* Volker Otto,* Sophie Hoffmann,* Heinrich Korth,* Udo Kl€ ockner,† Ying Zhang,* Yi Li,* Claudia Grossmann,† Uwe Rueckschloss,† Michael Gekle,† Juliane Stieber,‡ Stefan Frantz,* Karl Werdan,* Ursula M€ uller-Werdan*# and Harald Loppnow* *Department of Internal Medicine III, †Julius-Bernstein-Institute of Physiology, Martin-Luther-Universit€at HalleWittenberg, Halle (Saale), ‡Institute of Experimental and Clinical Pharmacology and Toxicology, University of Erlangen, Erlangen, and #Chair of Geriatrics, Charite - Universit€atsmedizin Berlin, Berlin, Germany
SUMMARY In critically ill patients regulation of heart-rate is often severely disturbed. Interaction of bacterial endotoxin (lipopolysaccharide, LPS) with hyperpolarization-activated cyclic nucleotide-gated cation-(HCN)-channels may interfere with heart-rate regulation. This study analyzes the effect of LPS, the HCN-channel blocker ivabradine or Ca2+-channel blockers (nifedipine, verapamil) on pacemaking in spontaneously beating neonatal rat cardiomyocytes (CM) in vitro. In vivo, the effect of LPS on the heart-rate of adult CD1-mice with and without autonomic blockade is analyzed telemetrically. LPS (100 ng/mL) and ivabradine (5 lg/mL) reduced the beating-rate of CM by 20.1% and 24.6%, respectively. Coincubation of CM with both, LPS and ivabradine, did not further reduce the beating-rate, indicating interaction of both compounds with HCN-channels, while coincubation with Ca2+-channel blockers and LPS caused additive beating-rate reduction. In CD1-mice (containing an active autonomic-nervous-system), injection of LPS (0.4 mg/kg) expectedly resulted in increased heart-rate. However, if the autonomic nervous system was blocked by propranolol and atropine, in line with the in vitro data, LPS induced a significant reduction of heart-rate, which was not additive to ivabradine. The in vivo and in vitro results indicate that LPS interacts with HCN-channels of cardiomyocytes. Thus, LPS indirectly sensitizes HCN-channels for sympathetic activation (tachycardiceffect), and in parallel directly inhibits channel activity (bradycardic-effect). Both effects may contribute to the detrimental effects of septic cardiomyopathy and septic autonomic dysfunction.
Correspondence: Prof. Dr. Harald Loppnow, Department of Internal Medicine III, Martin-Luther-Universit€at Halle-Wittenberg, 06120 Halle (Saale), Germany. Email: [email protected]
UMW and HL contributed equally as senior authors. Received 10 December 2014; revision 8 April 2015; accepted 19 April 2015. © 2015 Wiley Publishing Asia Pty Ltd
Key words: autonomic blockade, beating-rate, endotoxin, gene deficient mice, HCN-channels, heart cells, heart-rate, ivabradine, neonatal rat cardiomyocytes, sepsis.
INTRODUCTION Heart-rate (HR) is an independent predictor of morbidity and mortality in cardiovascular diseases (CVD).1,2 Lowering HR has been shown to be associated with favourable clinical outcomes in different patient populations.3,4 Regulation of HR is severely disturbed in the vast majority of critically ill patients suffering from severe sepsis and septic shock. These patients show an inadequate tachycardia and in parallel also a significant reduction of heart-rate variability (HRV) which has been shown to indicate an unfavourable prognosis.5 Within the last years, hyperpolarization-activated cyclic nucleotide-gated cation-(HCN)-channels have been recognized as the molecular basis of the ‘funny current’ If and thereby as important regulators in the process of cardiac pacemaking (review6). These channels are transmembrane proteins that serve as ion-channels, generating and regulating rhythmic activity of cardiac pacemaker cells and spontaneously firing neurons. Elevation of intracellular cAMP facilitates activation of the channels by shifting the voltage dependence of HCN-channels to less hyperpolarized potentials and by accelerating the opening kinetics.7 The HCN-channel family, comprising four members, belongs to the superfamily of ‘poreloop’ cation-channels.8 The expression of HCN-channels has been identified throughout the heart and the nervous system. HCN4 is the major isoform expressed in the cardiac conduction system, including sinuatrial node (SAN), atrioventricular node and the Purkinje fibres.9 HCN-channels are also present in atrial and ventricular myocytes and HCN2 is the dominant isoform in these cells. Ivabradine (C27H36N2O5; Iva) is an If-current blocker available for clinical use. The drug molecules are ‘open HCN-channel’ blockers and enter the channel pore from the intracellular side to reach their binding site, thereby closing the channels. Ivabradine slows HR by selectively inhibiting the pacemaker If-current in a dose-dependent manner. It slows the diastolic depolarization slope of SAN-cells, with minimal effect on myocardial contractility, blood pressure, and intracardiac conduction. Based on HR reduction, ivabradine improves ischaemic myocardial blood flow,
HCN-channels and endotoxin as well as contractile function of heart and infarct size.10 Ivabradine has been reported recently to increase the indices of HRV in patients with CVDs. However, the precise mechanism governing this effect is unknown. Since a non-autonomic modulation of HRV has been described in denervated human heart11 an intrinsic mechanism regulating HRV may also exist. Although cardiac automaticity is intrinsic to various pacemaker tissues, HR is, to a great extent, under the control of the autonomic nervous system.12 There is a significant relationship between the autonomic nervous system activity and cardiovascular mortality including sudden death.13 HRV represents autonomic nervous system activities, including sympathetic and vagal activities, making it a promising marker of mortality in CVD and non-cardiac diseases. Autonomic dysfunction has been shown in cardiovascular diseases, as well as in sepsis, systemic inflammatory response syndrome (SIRS), and multiorgan dysfunction syndrome (MODS),5,14 whereby, especially in critical illness, participation of an intracardiac source of HRV in addition to autonomic dysfunction is likely.15 Endotoxins (lipopolysaccharides; LPS) are large molecules in the outer membrane of Gram-negative bacteria,16 consisting of a lipid and a polysaccharide component. They are able to interact with various cell types of the host and can activate cells of the immune and the cardiovascular system, such as endothelial, smooth muscle, and heart cells, to release inflammatory mediators like IL-6, IL-1 and chemokines.16–20 These mediators may contribute to the regulation of inflammatory processes in CVD. In addition to these well-known inflammatory reactions, we and other laboratories have shown previously that LPS can directly interfere with the electrophysiological properties of HCN-channels.21–24 This may explain (at least in part) the disturbance of HR regulation found in severe sepsis mentioned above. However, so far all data regarding the interaction of LPS with HCN-channels have been derived either from heterologous expression systems (HEK293-cells)21,23 or from adult human atrial cardiomyocytes with no spontaneous depolarization.22 Therefore, the experiments presented here aimed to analyze, on the one hand, the effects of LPS at the cellular level in spontaneously beating cardiomyocytes, without interference of the autonomic system. On the other hand, we aimed to analyze the LPS effect in vivo in mice, both, with a functionally active, but also with a pharmacologically blocked autonomic nervous system. These investigations may provide further evidence for the impact of LPS on cardiac pacemaking. The data suggest that LPS interferes with the beating-rate at the level of the HCN-channels.
RESULTS LPS reduces the beating-rate of cultured neonatal rat cardiomyocytes Previous data indicated that LPS may interfere with HCN-channels. In order to analyze at the cellular level whether LPS has an effect on HR, possibly by HCN-channels, experiments were performed with spontaneously beating cardiomyocytes isolated from neonatal rats (CM). In unstimulated CM cultures, the spontaneous beating-rate of the cardiomyocytes was 71.3 2.9 beats per minute (b.p.m.; Fig. 1). In the presence of 100 ng/mL LPS, the beating-rate was reduced significantly to 57.0 5.4 b.p.m.,
875
Fig. 1 LPS reduces the beating-rate of cultured neonatal rat cardiomyocytes. CM (1.2 9 105 cells/cm2) were cultured in 6-well plates for 24 h in CMRL-growth-medium, and subsequently in serumfree CMRL-synthetic-medium. After replacing the medium with fresh CMRL-syntheticmedium, LPS (100 ng/mL) was applied to the respective cultures. Each condition was incubated in duplicate cultures. After 24 h, the beating-rate of the cells was measured by analyzing three regions in each well for 72 s. The data of 54 measurements for each condition were included in this analysis. Significances were calculated by Mann–Whitney test.
representing a 20.1% reduction of the beating frequency in response to LPS. The effect of LPS on the beating-rate of CM is dependent on HCN- but not on Ca2+-channels As shown above, LPS reduced the beating-rate of CM. Previous work from heterologous expression systems indicated a direct interaction of LPS with HCN-channels, without requirement for classical LPS-mediated signaling pathways, such as NF-kB or MAPkinases.21–23 Therefore, we tested the hypothesis that LPS affects the beating frequency via its interaction with HCN-channels, using the HCN-channel blocker ivabradine. First, we analyzed the effect of ivabradine in the absence or presence of LPS. As shown in Fig. 2, increasing concentrations of ivabradine reduced the beating-rate of CM in a dose-dependent manner. LPS (100 ng/mL) did not further reduce the beating-rate. This indicates a common target of ivabradine and LPS – in this case the HCN-channels – resulting in a non-additive action. By contrast, inhibition of voltage-gated calcium-channels with either nifedipine (Fig. 3) or verapamil (data not shown) resulted in an additive effect of the blockers and the LPS, supporting the suggestion that these channels do not contribute to the observed LPS action on the beating-rate. LPS causes tachycardia in conscious mice, which is reduced by blockade of the autonomic system In order to show the influence of LPS on the HR in vivo the NNinterval (i.e., the sinus node cycle length), which is inversely related to the beating-rate (i.e., enhanced NN reflects reduced beating-rate), was measured telemetrically. In vivo, the cycle length of the murine sinus node (NN in milliseconds (ms)) under resting conditions (hatched columns) in the experiments presented in Fig. 4 ranged from 123.3 4.7 ms (≙ 488 b.p.m.) to 134.7 4.4 ms (≙ 444 b.p.m.) with no significant differences between the individual experiments. As expected, a reduction of the mean NNinterval was observed after the injection (black columns) of atropine (132.0 7.5 to 102.4 3.3 ms), as presented in Fig. 4a, whereas the administration of the beta-blocker propranolol led to
© 2015 Wiley Publishing Asia Pty Ltd
H Ebelt et al.
876
(a)
Fig. 2 Ivabradine does not further reduce the effect of LPS on the beating-rate of cultured neonatal rat cardiomyocytes. CM (1.2 9 105 cells/cm2) were cultured in 6-well plates, as described in the legend to Fig. 1. Different concentrations of ivabradine were added to the cultures. Immediately thereafter, 100 ng/mL LPS (black columns) or medium without LPS (white columns) was applied to the respective cultures and the beating-rate of the cells was measured after 24 h. The number of measurements included in the calculation: (0 Iva, n = 54; 5 Iva, n = 30; 10 Iva, n = 6; 25 Iva, n = 18). Significances were calculated by the Mann–Whitney test. ns, not significant.
Fig. 3 Nifedipine further reduces the effect of LPS on the beating-rate of neonatal rat cardiomyocytes. CM (1.2 9 105 cells/cm2) were cultured in 6-well plates. Different concentrations of the calcium-channel blocker nifedipine were added to the cultures. Subsequently, 100 ng/mL LPS (black columns) or medium without LPS (white columns) was applied to the respective cultures and the beating-rate of the cells was measured after 24 h. The data of 30 measurements for each condition are summarized. Significances were calculated by the Mann–Whitney Test.
an elevation of the NN-interval from 124.1 5.8 to 183.8 15.3 ms). As expected, the combined inhibition of the autonomic nervous system by simultaneous injection of both, atropine and propranolol, did not significantly increase the NN interval (134.7 4.4 to 146.1 7.9 ms). On the other hand, sole injection of LPS induced tachycardia in mice, as shown by the reduction of the NN-interval from 129.4 7.9 to 105.2 5.6 ms (Fig. 4b). However, when LPS was given simultaneously with autonomic blockade (atropine + propranolol), a significant increase of the mean NN-interval from 123.3 4.7 to 176.9 5.8 ms was observed (Fig. 4b), in line with the above cell culture experiments (compare Fig. 1). Interference of LPS and ivabradine suggests the HCNchannels as a target of LPS In order to analyze whether the effect of LPS on the HR in vivo depends on the HCN-channels, injection experiments were per-
(b)
Fig. 4 (a, b) LPS causes tachycardia in conscious mice, which is reduced by blockade of the autonomic system. Heart-rate of CD1-mice was determined by telemetrical ECG-recording 15 min before (hatched columns) and 20 min after (black columns) the injection of the indicated compounds. Four mice were used in each group (n ≥ 15 injections per group). A, atropine (1 mg/kg); P, propranolol (20 mg/kg); LPS, lipopolysaccharide (0.4 mg/kg). Statistical analysis was performed using t-test. ns, not significant.
formed in the absence and the presence of the HCN-channel blocker ivabradine (Iva). In accordance with previous studies, a dose-dependent HR-lowering effect of ivabradine was observed up to 10 mg/kg (not shown), whereas a further increase of the dosage did not show a significant effect. A representative tracing is shown in Fig. 5a. Thus, the concentration of 10 mg/kg was used for all further experiments. In the present experiment (Fig. 5b), ivabradine expectedly enhanced the NN-interval. Also shown in Fig. 5b, injection of the HCN-channel blocker ivabradine with subsequent injection of LPS did not significantly alter the HR of the mice, as compared to ivabradine injected solely. Furthermore, in the situation of autonomic blockade induced by atropine and propranolol, as described above, LPS failed to induce a further increase of the NN-interval, if the mice were simultaneously treated with ivabradine (Fig. 5c). These in vivo findings also argue for HCN-channels as the common target of ivabradine and LPS, as suggested by the in vitro data shown above. Besides HCN4, other HCN-channels may be involved in the response to LPS Previous experiments suggested that LPS is not only able to interfere with HCN4, but also with other HCN-isoforms.21,22 Therefore, we analyzed the effect of LPS on the sinus node cycle length in HCN4-knock-out animals. For this purpose, animals were used, where one HCN4-allele was replaced by a tamoxifen-sensitive Crerecombinase (HCN4+/ ; hatched columns), while the second HCN4-allele was marked by loxP-sites, thereby enabling the induction of a complete HCN4-knock-out (HCN4 / ; black columns).25 The HCN4+/ -mice first underwent a series of pharmacological injections before the remaining HCN4-allele was knocked out by tamoxifen (HCN4 / ). The basal sinus node cycle length of the HCN4+/ -mice was 128.0 5.6 ms and did not differ significantly from that of wild type animals (123.3 4.7, compare Fig. 4). As shown in Fig. 6, the complete loss of HCN4-channels (HCN4 / ) did not significantly alter the response of the animals to the injection of atropine, but reduced the effect of the betablockade by propranolol, as compared to HCN4+/ -animals. However, the effect of LPS on HR was found to be unchanged in HCN4 / -mice, both under basal conditions as well as under
© 2015 Wiley Publishing Asia Pty Ltd
HCN-channels and endotoxin
877
(a1)
(a2)
(c)
(b)
Fig. 5 (a, b, c) Interference of LPS and ivabradine suggests the HCNchannel as a target of LPS. (a) Representative examples of ECG tracings in awake mice using telemetric transmitters. (a1) Control; (a2) Ivabradine, 20 min after injection of Iva. (b, c) Heart-rate of CD1-mice was determined by telemetrical ECG-recording 15 min before (hatched columns) and 20 min after the injection (black columns) of the indicated compounds (n ≥ 15 injections per group). A, P, LPS, compare Fig. 4; Iva, ivabradine (10 mg/kg). Statistical analysis was performed using the t-test. ns, not significant.
autonomic blockade, suggesting that further HCN-channels, such as HCN2, may contribute to the LPS effect. Taken together, the above findings are in line with the hypothesis that LPS, in the presence of autonomic innervation, leads to a sensitization of the HCN-channels to sympathetic stimuli, with the consequence of an elevated HR. However, when LPS is given in the absence of autonomic stimulation it directly reduces the HR. The effect of LPS on the sinus node seems to be dependent on HCN-channels because it can be prevented by the HCNchannel blocker ivabradine. The interaction of LPS is not limited to the HCN4-channel, but other HCN-channel subtypes may also interact.
DISCUSSION Functional disturbances of otherwise healthy organs are the hallmarks of severe sepsis and septic shock, leading to septic organ dysfunction. Especially the impairment of renal function or septic encephalopathy are well known in this regard. However, although not recognized so easily, the heart can also suffer from sepsisinduced injury,26,27 predominantly of diastolic nature,26 which
Fig. 6 Besides HCN4, other HCN-channels may be involved in the response to LPS. CD1-mice (wild-type) were used where one HCN4allele was replaced by a tamoxifen-sensitive Cre-recombinase while the second HCN4-allele was marked by loxP-sites (HCN4CreERT2/HCN4loxP). These mice, containing one functionally intact HCN4-allele (HCN4+/ ; hatched columns) first underwent the injections indicated in the Methods, before the remaining HCN4-allele was knocked out by tamoxifen and the injections got repeated 20 days later (HCN4 / ; black columns) (n ≥ 15 injections per group). A, P, LPS, Iva, compare Fig. 5. Statistical analysis was performed using the t-test. ns, not significant.
can be quantified as impairment in cardiac function.27 Important components of septic cardiomyopathy are abnormalities in the regulation of HR and HRV.14 These particular abnormalities have been repeatedly shown to be of high prognostic relevance in critically ill patients at the ICU so that it seems to be of special interest to understand the underlying mechanisms.5 We have shown previously that LPS, which is a component of the cell membrane of Gram-negative bacteria, can interact with HCN-channels, which are crucially involved in cardiac pacemaking.21–23 Experiments with isolated human cardiomyocytes had demonstrated that LPS can directly influence the ‘funny current’ If in two ways: (i) LPS can lead to a sensitization of the HCNchannels to adrenergic stimuli; but (ii) in parallel also can directly inhibit HCN-channels.22 Therefore, we analyzed whether or not the previous findings also apply to functional in vitro cell culture experiments and to the even more complex in vivo situation in mice. A number of experiments have addressed the effects of LPS on HR regulation before.28,29 Many of these studies utilized experimental systems where changes occurred with a delay of several hours after LPS exposure, or last over 1–24 h. This may involve the host’s immune system and alterations of the central nervous system.30 In order to avoid this interference, we performed in vivo analyses measuring immediately after LPS-administration, i.e. within 30 min. The present in vivo studies indeed confirmed the results of the authors’ previous and present cell culture experiments: while in the presence of an intact autonomic nervous system in the mice i.p. injection of LPS led to tachycardia, LPS induced bradycardia when the autonomic modulation was pharmacologically blocked (Figs 4 and 5). In the cell culture experiments with CM – which may be considered as a model of isolated pacemaker cells – only the direct bradycardic effect can be detected, due to the lack of the influence of the autonomic nervous system. The present study also supports the view that the immediately occurring effect of LPS on the HR is based on an interaction with cardiac HCN-channels, which are responsible for the ‘funny
© 2015 Wiley Publishing Asia Pty Ltd
878
H Ebelt et al.
current’ as the molecular basis of pacemaking. This interpretation is based on the finding that both, in mice as well as in experiments with CMs, the effect of LPS is not altered by the blockade of calcium-channels, but can be prevented by the administration of the HCN-channel blocker ivabradine. In mice and humans, the HCN-family is composed of four different members named HCN1 to HCN4.6 In the past, the majority of studies investigating the interplay between LPS and cardiac pacemaking focused on HCN2 for several reasons. First, the initial studies were performed in cardiomyocytes isolated from human right atria that mainly express the HCN2-isoform.22 Additionally, in 2007 we have reported that LPS reduces the beatingrate variability of CM.28 In these cells HCN2 is the main isoform, with a five times higher amount of mRNA detectable than of HCN4.9,31 However, in mice and humans, HR is under normal conditions determined by the activity of the sinus node where HCN4 is expressed in much higher amounts.9 Accordingly, we have shown previously that in HEK293-cells, which stably expressed hHCN4 or hHCN2, respectively, LPS exerts comparable effects on both channel subtypes.21 In order to address the question whether the effects of LPS on sinus node activity in vivo are restricted to HCN4, transgenic animals were included into the study where HCN4 can be knockedout by the injection of tamoxifen. It has been shown previously in detail that HR regulation occurs quite normally in these animals after HCN4-knock-out and that disturbances of cardiac pacemaking are seen especially in situations of increased repolarizing currents, as induced by muscarinic stimulation or during the transition from activated to basal cardiac state while no limitations were found under b-adrenergic stimulation.25 In accordance with this suggestion, we found comparable effects induced by atropine before and after the knock-out of HCN4, whereas the effect of the beta-blocking agent propranolol was reduced in the absence of HCN4. However, injection of LPS led to the same effects on HR both under basal conditions, as well as under autonomic blockade, irrespective whether or not HCN4 was expressed. This indicates that the HCN-channels remaining after HCN4-knockout, such as HCN1 or HCN2, are also affected by LPS. If autonomic dysfunction in MODS, indicated by narrowed HRV and inadequately high HR, is linked to an unfavourable outcome, improvement of autonomic function might improve prognosis of these patients. First therapeutic attempts in this direction have tried to reduce the inadequately high HR in patients with septic shock by the short acting betablocker esmolol,32 as well as by the If-blocker ivabradine.33 In a randomized study with 154 patients with septic shock and a HR ≥ 95 b.p.m., esmolol achieved a mean HR reduction of 18 b.p.m.. The 28-day mortality in the esmolol group was 49.4% and in the control group 80.5% (adjusted hazard ratio 0.39; 95% confidence interval 0.26–0.59; P < 0.001).32 In the randomized MODIFY trial33 70 patients with MODS (APACHE II score ≥ 20), a HR ≥ 90/min and existing contraindications to betablockers, were randomly treated with standard therapy or standard therapy plus a 4-day treatment with ivabradine. The primary endpoint of this trial was the percentage of patients with a reduction of the mean HR of at least 10 b.p.m., 96 h after the start of trial treatment. Recruitment of the study is already closed and data analysis is still ongoing.33 While in the esmolol and the MODIFY trial the inadequately high HR of the patients was reduced by blocking sympathetic
drive and If pharmacologically, the consequence of a direct blockade of LPS by antibodies or immunoglobulins on HR and on HRV have not yet been studied in critically ill patients. Taken together, the data presented here add further evidence to the suggestion that LPS can directly interact with cardiac pacemaking by the interference with sinuatrial HCN-channels. This interaction can be considered to be a mechanism leading to altered heart-rate regulation observed in severe sepsis that is linked to an unfavorable outcome of critically ill patients – but which might also bear therapeutic potential in the future.
MATERIAL AND METHODS Isolation and culture of neonatal rat cardiomyocytes Spontaneously beating neonatal rat cardiomyocytes (CM) were isolated from the heart of 1 to 3 days old Wistar rats by the method described by Werdan and Erdmann34. Briefly, beating hearts were taken from neonates sacrificed by cervical dislocation, rinsed in cold PBS-AG1 solution and bruised using a scalpel. The bruised hearts were incubated in dissociation solution (0.03 % collagenase; 0.12 % trypsin; in PBS-AG1-solution; 37°C) for 10 min and the supernatant was collected. The dissociation procedure was repeated 10 times, the first supernatant was discarded. The tubes containing the CM in CMRL-growth-medium (compare34) were centrifuged (1,400 rpm, 10 min, room temperature) and the CM resuspended in CMRL-growth-medium. This cell suspension was filtered (cell strainers; 70 lm) and transferred into 175 cm2 culture flasks for 1.5 h (37°C) in order to separate CM from non-muscle cells adhering to the culture dish (differential attachment technique). The cell suspension containing the non-adhesive cardiomyocytes was transferred into culture flasks or plates in CMRL-growth-medium (1.2 9 105 cells/cm2). After 24 hours incubation at 37°C, the culture medium was replaced by CMRL-synthetic-medium (compare34) and the CM were used for the experiments after further 24 hours of incubation. Determination of the beating-rate in CM The beating-rate of isolated CM was recorded using a photo-electrical system.28 Cells cultured in 6-well plates were analyzed under the microscope in a chamber keeping the temperature constantly at 37°C. Three regions of each well were measured for 72 s. For each condition, six measurements were performed. The frequency of the CM contraction, as well as parameters of beating-rate variability were analyzed in a program specifically designed using the software LABVIEW.35 ECG recordings in mice All animal experiments shown here comply with the Directive 2010/63/EU of the European Parliament and were approved by the local review board. Adult male CD1-mice were used for the in vivo studies. For transmitter implantation, mice underwent general anaesthesia with 2.5% isofluran, were placed on a prewarmed operation table (38°C) and received a subcutaneous injection of the analgetic carprofen (5 mg/kg). A median laparotomy was performed and an ECG-transmitter (PhysioTelTM
© 2015 Wiley Publishing Asia Pty Ltd
HCN-channels and endotoxin EA-F20; Data Science International, St Paul, MN, USA) was placed into the peritoneal cavity, recording both ECG-data and locomotory activity of the animals. The electrodes were guided through the peritoneum using a needle and placed subcutaneously on the right upper chest and in front of the lower left costal arch. The peritoneal cavity, as well as all skin lesions, were closed with surgical sutures. In order to avoid interference with the healing process, all injection experiments were started not earlier than 15 days after the transmitter implantation. In vivo injection protocol In order to analyze the influence of the autonomic nervous system on the effects of LPS on heart-rate, atropine sulfate (1 mg/ kg; B. Braun AG, Melsungen, Germany), propranolol hydrochloride (20 mg/kg; MIBE GmbH Arzneimittel, Sandersdorf-Brehna, Germany), ivabradine (10 mg/kg; Servier-Deutschland, M€ unchen, Germany), LPS (0.4 mg/kg; Sigma Aldrich, Taufkirchen, Germany) or combinations of these compounds were applied by intraperitoneal injection. All injections were performed not earlier than 15 days after the implantation of the ECG-transmitter and always between 4 p.m. and 6 p.m. in order to minimize circadian fluctuations. During the injections the animals received a shortlasting (approx. 3–5 min) anaesthesia induced by inhalation of 2.5% isoflurane. In the first part of the injection protocol, propranolol followed by atropine, ivabradine and finally the combination of atropine and propranolol were applied with a minimum of 24 h break between each injection. This cycle of injections was performed in triplicate with a 48 h break in between. In the second cycle of experiments, LPS followed by the combination of LPS + atropine + propranolol, ivabradine + atropine + propranolol, and LPS + ivabradine + atropine + propranolol were injected with a minimum of 48 h between the injections. At the end of the experiments, the animals were killed by cervical dislocation under general anaesthesia with 5% isoflurane. Each group consisted of four animals.
879
(ECG-tracings and locomotory activity) was performed using a standard telemetry system (Data Science International, PhysioTel R and MultiplusTM Implant, St. Paul, MN, USA). To analyze the effect of the different injections, a standardized approach was used. First, from both the time span before the injection ( 25 min until 10 min), as well as from the time after the injection (+20 min until +40 min) three intervals each lasting 3 min were manually selected, where physical activity of the mice was especially low or absent, as seen from the locomotory recordings. From these 3 min segments, both the mean NN-interval (meanNN3 min), as well as the locomotory activity score were calculated automatically. If the locomotory activity score was < 1, the meanNN3 min-interval of this segment was used to contribute to the mean NN-interval (meanNNbefore and meanNNafter) that was used finally for statistical analysis. Statistical analysis Statistical analysis was performed using SPSS (SPSS Inc., Chicago, IL, USA). All metric data are shown as mean standard error of the mean (SEM). Comparisons of two groups were performed using student’s t-test (paired or unpaired, respectively), or Mann–Whitney-test, as mentioned in the respective legends. Comparisons of multiple groups were done by ANOVA with post-hoc testing (Bonferroni). *P-values of < 0.05 were considered statistically significant.
ACKNOWLEDGEMENTS The expert technical assistance of S. Koch and C. Pilowski is gratefully acknowledged. This work was supported by grants of the DFG (German Research Foundation) to Ursula M€ uller-Werdan, Henning Ebelt and Michael Gekle (DFG MU 1010/4-1 and GE 905/18-1, respectively), by the Roux-Programm of the Medical Faculty of the Martin-Luther-Universit€at Halle-Wittenberg, and by a grant of Servier-Deutschland to Harald Loppnow, Henning Ebelt, and Karl Werdan.
DISCLOSURE
Conditional HCN4 knock-out mice The generation and phenotype of the tamoxifen-inducible HCN4knock-out mice has been described in detail before.25 In brief, a tamoxifen-inducible Cre-recombinase (CreERT2 construct) was ‘knocked in’ into the HCN4-locus. After crossing these animals with floxed HCN4-mice, complete deletion of the HCN4-gene can be achieved by injection of tamoxifen. Animals with a successful knock-down of HCN4-channels can be identified by their ECG-recordings, since they show regularly occurring sinuatrial blockades.25 In the present experiments, all mice with one functionally intact HCN4 allele (HCN4CreERT2/HCN4loxP; HCN4+/ ) first underwent the injection protocol described above. Afterwards, the animals were subjected to tamoxifen treatment (240 mg/kg for 3 days), leading to the complete HCN4-knockout (HCN4 / ). After an interval of at least 20 days, the initial pharmacological injection protocol was repeated again. ECG data collection During the ECG-recordings, all mice remained in their standard cages and could move normally. Data acquisition and analysis
Harald Loppnow, Henning Ebelt, and Karl Werdan received an unrestricted grant of Servier Deutschland. K. Werdan got honoraria from Servier company for participation in national and international advisory board activities, for his national leadership in the SHIFT study and for lectures.
REFERENCES 1. Kannel WB, Kannel C, Paffenbarger RS Jr, Cupples LA. Heart rate and cardiovascular mortality: The Framingham Study. Am. Heart J. 1987; 113: 1489–94. 2. Seccareccia F, Pannozzo F, Dima F et al. Heart rate as a predictor of mortality: The MATISS project. Am. J. Public Health 2001; 91: 1258–63. 3. Swedberg K, Komajda M, B€ohm M et al. Ivabradine and outcomes in chronic heart failure (SHIFT): A randomised placebo-controlled study. Lancet 2010; 376: 875–85. 4. Hoke RS, M€uller-Werdan U, Lautenschl€ager C, Werdan K, Ebelt H. Heart rate as an independent risk factor in patients with multiple organ dysfunction: A prospective, observational study. Clin. Res. Cardiol. 2012; 101: 139–47.
© 2015 Wiley Publishing Asia Pty Ltd
880
H Ebelt et al.
5. Schmidt H, M€ uller-Werdan U, Hoffmann T et al. Autonomic dysfunction predicts mortality in patients with multiple organ dysfunction syndrome of different age groups. Crit. Care Med. 2005; 33: 1994–2002. 6. Benarroch EE. HCN channels: Function and clinical implications. Neurology 2013; 80: 304–10. 7. Tokimasa T, Akasu T. Cyclic AMP regulates an inward rectifying sodium-potassium current in dissociated bull-frog sympathetic neurones. J. Physiol. 1990; 420: 409–29. 8. Yu FH, Yarov-Yarovoy V, Gutman GA, Catterall WA. Overview of molecular relationships in the voltage-gated ion channel superfamily. Pharmacol. Rev. 2005; 57: 387–95. 9. Shi W, Wymore R, Yu H et al. Distribution and prevalence of hyperpolarization-activated cation channel (HCN) mRNA expression in cardiac tissues. Circ. Res. 1999; 85: e1–6. 10. Heusch G. Heart rate in the pathophysiology of coronary blood flow and myocardial ischaemia: Benefit from selective bradycardic agents. Br. J. Pharmacol. 2008; 153: 1589–601. 11. Bernardi L, Salvucci F, Suardi R et al. Evidence for an intrinsic mechanism regulating heart rate variability in the transplanted and the intact heart during submaximal dynamic exercise? Cardiovasc. Res. 1990; 24: 969–81. 12. Jalife J, Slenter VA, Salata JJ, Michaels DC. Dynamic vagal control of pacemaker activity in the mammalian sinoatrial node. Circ. Res. 1983; 52: 642–56. 13. Schwartz PJ. Cardiac sympathetic innervation and the prevention of sudden death. Cardiologia 1990; 35: 51–4. 14. Werdan K, Schmidt H, Ebelt H et al. Impaired regulation of cardiac function in sepsis, SIRS, and MODS. Can. J. Physiol. Pharmacol. 2009; 87: 266–74. 15. Papaioannou VE, Verkerk AO, Amin AS, de Bakker JM. Intracardiac origin of heart rate variability, pacemaker funny current and their possible association with critical illness. Curr. Cardiol. Rev. 2013; 9: 82–96. 16. Loppnow H, Brade H, Rietschel ET, Flad HD. Induction of cytokines in mononuclear and vascular cells by endotoxin and other bacterial products. Methods Enzymol. 1994; 236: 3–10. 17. Loppnow H, Libby P. Adult human vascular endothelial cells express the IL6 gene differentially in response to LPS or IL1. Cell. Immunol. 1989; 122: 493–503. 18. Schonbeck U, Herzberg M, Petersen A et al. Human vascular smooth muscle cells express interleukin-1beta-converting enzyme (ICE), but inhibit processing of the interleukin-1beta precursor by ICE. J. Exp. Med. 1997; 185: 1287–94. 19. M€ uller-Werdan U, Schumann H, Loppnow H et al. Endotoxin and tumor necrosis factor alpha exert a similar proinflammatory effect in neonatal rat cardiomyocytes, but have different cardiodepressant profiles. J. Mol. Cell. Cardiol. 1998; 30: 1027–36. 20. Westphal E, Li C, Pilowski C et al. Endotoxin-activated cultured neonatal rat cardiomyocytes express functional surface-associated interleukin-1alpha. J. Endotoxin Res. 2007; 13: 25–34.
21. Kl€ockner U, Rueckschloss U, Grossmann C et al. Differential reduction of HCN channel activity by various types of lipopolysaccharide. J. Mol. Cell. Cardiol. 2011; 51: 226–35. 22. Zorn-Pauly K, Pelzmann B, Lang P et al. Endotoxin impairs the human pacemaker current If. Shock 2007; 28: 655–61. 23. Kl€ockner U, Rueckschloss U, Grossmann C et al. Inhibition of cardiac pacemaker channel hHCN2 depends on intercalation of lipopolysaccharide into channel-containing membrane microdomains. J. Physiol. 2014; 592: 1199–211. 24. Scheruebel S, Koyani CN, Hallstr€om S et al. I(f) blocking potency of ivabradine is preserved under elevated endotoxin levels in human atrial myocytes. J. Mol. Cell. Cardiol. 2014; 72: 64–73. 25. Herrmann S, Stieber J, St€ockl G, Hofmann F, Ludwig A. HCN4 provides a ‘depolarization reserve’ and is not required for heart rate acceleration in mice. EMBO J. 2007; 26: 4423–32. 26. Landesberg G, Gilon D, Meroz Y et al. Diastolic dysfunction and mortality in severe sepsis and septic shock. Eur. Heart J. 2012; 33: 895–903. 27. Werdan K, Oelke A, Hettwer S et al. Septic cardiomyopathy: Hemodynamic quantification, occurrence, and prognostic implications. Clin. Res. Cardiol. 2011; 100: 661–8. 28. Schmidt H, Saworski J, Werdan K, M€uller-Werdan U. Decreased beating rate variability of spontaneously contracting cardiomyocytes after co-incubation with endotoxin. J. Endotoxin Res. 2007; 13: 339–42. 29. Gholami M, Mazaheri P, Mohamadi A et al. Endotoxemia is associated with partial uncoupling of cardiac pacemaker from cholinergic neural control in rats. Shock 2012; 37: 219–27. 30. Fairchild KD, Saucerman JJ, Raynor LL et al. Endotoxin depresses heart rate variability in mice: Cytokine and steroid effects. Am. J. Physiol. Regul. Integr. Comp. Physiol. 2009; 297: R1019–27. 31. Herrmann S, Stieber J, Ludwig A. Pathophysiology of HCN channels. Pflugers Arch. 2007; 454: 517–22. 32. Morelli A, Ertmer C, Westphal M et al. Effect of heart rate control with esmolol on hemodynamic and clinical outcomes in patients with septic shock: A randomized clinical trial. JAMA 2013; 310: 1683–91. 33. Nuding S, Ebelt H, Hoke RS et al. Reducing elevated heart rate in patients with multiple organ dysfunction syndrome by the I (f) (funny channel current) inhibitor ivabradine: MODI (f)Y trial. Clin. Res. Cardiol. 2011; 100: 915–23. 34. Werdan K, Erdmann E. Preparation and culture of embryonic and neonatal heart muscle cells: Modification of transport activity. Methods Enzymol. 1989; 173: 634–62. 35. Lehnich H, Kaltenborn G, M€uller-Werdan U, Pauer H-D. Digitalisierung, Verarbeitung und Auswertung von Videoaufzeichnungen in Echtzeit zur Kontraktilit€atsmessung an isolierten Herzzellen. Virtuelle Instrumente in der Praxis 2001; 2001: 312–7.
© 2015 Wiley Publishing Asia Pty Ltd
doi: 10.1111/1440-1681.12415
ORIGINAL ARTICLE
Direct inhibition, but indirect sensitization of pacemaker activity to sympathetic tone by the interaction of endotoxin with HCN-channels Henning Ebelt,* Isabel Geißler,* Sara Ruccius,* Volker Otto,* Sophie Hoffmann,* Heinrich Korth,* Udo Kl€ ockner,† Ying Zhang,* Yi Li,* Claudia Grossmann,† Uwe Rueckschloss,† Michael Gekle,† Juliane Stieber,‡ Stefan Frantz,* Karl Werdan,* Ursula M€ uller-Werdan*# and Harald Loppnow* *Department of Internal Medicine III, †Julius-Bernstein-Institute of Physiology, Martin-Luther-Universit€at HalleWittenberg, Halle (Saale), ‡Institute of Experimental and Clinical Pharmacology and Toxicology, University of Erlangen, Erlangen, and #Chair of Geriatrics, Charite - Universit€atsmedizin Berlin, Berlin, Germany
SUMMARY In critically ill patients regulation of heart-rate is often severely disturbed. Interaction of bacterial endotoxin (lipopolysaccharide, LPS) with hyperpolarization-activated cyclic nucleotide-gated cation-(HCN)-channels may interfere with heart-rate regulation. This study analyzes the effect of LPS, the HCN-channel blocker ivabradine or Ca2+-channel blockers (nifedipine, verapamil) on pacemaking in spontaneously beating neonatal rat cardiomyocytes (CM) in vitro. In vivo, the effect of LPS on the heart-rate of adult CD1-mice with and without autonomic blockade is analyzed telemetrically. LPS (100 ng/mL) and ivabradine (5 lg/mL) reduced the beating-rate of CM by 20.1% and 24.6%, respectively. Coincubation of CM with both, LPS and ivabradine, did not further reduce the beating-rate, indicating interaction of both compounds with HCN-channels, while coincubation with Ca2+-channel blockers and LPS caused additive beating-rate reduction. In CD1-mice (containing an active autonomic-nervous-system), injection of LPS (0.4 mg/kg) expectedly resulted in increased heart-rate. However, if the autonomic nervous system was blocked by propranolol and atropine, in line with the in vitro data, LPS induced a significant reduction of heart-rate, which was not additive to ivabradine. The in vivo and in vitro results indicate that LPS interacts with HCN-channels of cardiomyocytes. Thus, LPS indirectly sensitizes HCN-channels for sympathetic activation (tachycardiceffect), and in parallel directly inhibits channel activity (bradycardic-effect). Both effects may contribute to the detrimental effects of septic cardiomyopathy and septic autonomic dysfunction.
Correspondence: Prof. Dr. Harald Loppnow, Department of Internal Medicine III, Martin-Luther-Universit€at Halle-Wittenberg, 06120 Halle (Saale), Germany. Email: [email protected]
UMW and HL contributed equally as senior authors. Received 10 December 2014; revision 8 April 2015; accepted 19 April 2015. © 2015 Wiley Publishing Asia Pty Ltd
Key words: autonomic blockade, beating-rate, endotoxin, gene deficient mice, HCN-channels, heart cells, heart-rate, ivabradine, neonatal rat cardiomyocytes, sepsis.
INTRODUCTION Heart-rate (HR) is an independent predictor of morbidity and mortality in cardiovascular diseases (CVD).1,2 Lowering HR has been shown to be associated with favourable clinical outcomes in different patient populations.3,4 Regulation of HR is severely disturbed in the vast majority of critically ill patients suffering from severe sepsis and septic shock. These patients show an inadequate tachycardia and in parallel also a significant reduction of heart-rate variability (HRV) which has been shown to indicate an unfavourable prognosis.5 Within the last years, hyperpolarization-activated cyclic nucleotide-gated cation-(HCN)-channels have been recognized as the molecular basis of the ‘funny current’ If and thereby as important regulators in the process of cardiac pacemaking (review6). These channels are transmembrane proteins that serve as ion-channels, generating and regulating rhythmic activity of cardiac pacemaker cells and spontaneously firing neurons. Elevation of intracellular cAMP facilitates activation of the channels by shifting the voltage dependence of HCN-channels to less hyperpolarized potentials and by accelerating the opening kinetics.7 The HCN-channel family, comprising four members, belongs to the superfamily of ‘poreloop’ cation-channels.8 The expression of HCN-channels has been identified throughout the heart and the nervous system. HCN4 is the major isoform expressed in the cardiac conduction system, including sinuatrial node (SAN), atrioventricular node and the Purkinje fibres.9 HCN-channels are also present in atrial and ventricular myocytes and HCN2 is the dominant isoform in these cells. Ivabradine (C27H36N2O5; Iva) is an If-current blocker available for clinical use. The drug molecules are ‘open HCN-channel’ blockers and enter the channel pore from the intracellular side to reach their binding site, thereby closing the channels. Ivabradine slows HR by selectively inhibiting the pacemaker If-current in a dose-dependent manner. It slows the diastolic depolarization slope of SAN-cells, with minimal effect on myocardial contractility, blood pressure, and intracardiac conduction. Based on HR reduction, ivabradine improves ischaemic myocardial blood flow,
HCN-channels and endotoxin as well as contractile function of heart and infarct size.10 Ivabradine has been reported recently to increase the indices of HRV in patients with CVDs. However, the precise mechanism governing this effect is unknown. Since a non-autonomic modulation of HRV has been described in denervated human heart11 an intrinsic mechanism regulating HRV may also exist. Although cardiac automaticity is intrinsic to various pacemaker tissues, HR is, to a great extent, under the control of the autonomic nervous system.12 There is a significant relationship between the autonomic nervous system activity and cardiovascular mortality including sudden death.13 HRV represents autonomic nervous system activities, including sympathetic and vagal activities, making it a promising marker of mortality in CVD and non-cardiac diseases. Autonomic dysfunction has been shown in cardiovascular diseases, as well as in sepsis, systemic inflammatory response syndrome (SIRS), and multiorgan dysfunction syndrome (MODS),5,14 whereby, especially in critical illness, participation of an intracardiac source of HRV in addition to autonomic dysfunction is likely.15 Endotoxins (lipopolysaccharides; LPS) are large molecules in the outer membrane of Gram-negative bacteria,16 consisting of a lipid and a polysaccharide component. They are able to interact with various cell types of the host and can activate cells of the immune and the cardiovascular system, such as endothelial, smooth muscle, and heart cells, to release inflammatory mediators like IL-6, IL-1 and chemokines.16–20 These mediators may contribute to the regulation of inflammatory processes in CVD. In addition to these well-known inflammatory reactions, we and other laboratories have shown previously that LPS can directly interfere with the electrophysiological properties of HCN-channels.21–24 This may explain (at least in part) the disturbance of HR regulation found in severe sepsis mentioned above. However, so far all data regarding the interaction of LPS with HCN-channels have been derived either from heterologous expression systems (HEK293-cells)21,23 or from adult human atrial cardiomyocytes with no spontaneous depolarization.22 Therefore, the experiments presented here aimed to analyze, on the one hand, the effects of LPS at the cellular level in spontaneously beating cardiomyocytes, without interference of the autonomic system. On the other hand, we aimed to analyze the LPS effect in vivo in mice, both, with a functionally active, but also with a pharmacologically blocked autonomic nervous system. These investigations may provide further evidence for the impact of LPS on cardiac pacemaking. The data suggest that LPS interferes with the beating-rate at the level of the HCN-channels.
RESULTS LPS reduces the beating-rate of cultured neonatal rat cardiomyocytes Previous data indicated that LPS may interfere with HCN-channels. In order to analyze at the cellular level whether LPS has an effect on HR, possibly by HCN-channels, experiments were performed with spontaneously beating cardiomyocytes isolated from neonatal rats (CM). In unstimulated CM cultures, the spontaneous beating-rate of the cardiomyocytes was 71.3 2.9 beats per minute (b.p.m.; Fig. 1). In the presence of 100 ng/mL LPS, the beating-rate was reduced significantly to 57.0 5.4 b.p.m.,
875
Fig. 1 LPS reduces the beating-rate of cultured neonatal rat cardiomyocytes. CM (1.2 9 105 cells/cm2) were cultured in 6-well plates for 24 h in CMRL-growth-medium, and subsequently in serumfree CMRL-synthetic-medium. After replacing the medium with fresh CMRL-syntheticmedium, LPS (100 ng/mL) was applied to the respective cultures. Each condition was incubated in duplicate cultures. After 24 h, the beating-rate of the cells was measured by analyzing three regions in each well for 72 s. The data of 54 measurements for each condition were included in this analysis. Significances were calculated by Mann–Whitney test.
representing a 20.1% reduction of the beating frequency in response to LPS. The effect of LPS on the beating-rate of CM is dependent on HCN- but not on Ca2+-channels As shown above, LPS reduced the beating-rate of CM. Previous work from heterologous expression systems indicated a direct interaction of LPS with HCN-channels, without requirement for classical LPS-mediated signaling pathways, such as NF-kB or MAPkinases.21–23 Therefore, we tested the hypothesis that LPS affects the beating frequency via its interaction with HCN-channels, using the HCN-channel blocker ivabradine. First, we analyzed the effect of ivabradine in the absence or presence of LPS. As shown in Fig. 2, increasing concentrations of ivabradine reduced the beating-rate of CM in a dose-dependent manner. LPS (100 ng/mL) did not further reduce the beating-rate. This indicates a common target of ivabradine and LPS – in this case the HCN-channels – resulting in a non-additive action. By contrast, inhibition of voltage-gated calcium-channels with either nifedipine (Fig. 3) or verapamil (data not shown) resulted in an additive effect of the blockers and the LPS, supporting the suggestion that these channels do not contribute to the observed LPS action on the beating-rate. LPS causes tachycardia in conscious mice, which is reduced by blockade of the autonomic system In order to show the influence of LPS on the HR in vivo the NNinterval (i.e., the sinus node cycle length), which is inversely related to the beating-rate (i.e., enhanced NN reflects reduced beating-rate), was measured telemetrically. In vivo, the cycle length of the murine sinus node (NN in milliseconds (ms)) under resting conditions (hatched columns) in the experiments presented in Fig. 4 ranged from 123.3 4.7 ms (≙ 488 b.p.m.) to 134.7 4.4 ms (≙ 444 b.p.m.) with no significant differences between the individual experiments. As expected, a reduction of the mean NNinterval was observed after the injection (black columns) of atropine (132.0 7.5 to 102.4 3.3 ms), as presented in Fig. 4a, whereas the administration of the beta-blocker propranolol led to
© 2015 Wiley Publishing Asia Pty Ltd
H Ebelt et al.
876
(a)
Fig. 2 Ivabradine does not further reduce the effect of LPS on the beating-rate of cultured neonatal rat cardiomyocytes. CM (1.2 9 105 cells/cm2) were cultured in 6-well plates, as described in the legend to Fig. 1. Different concentrations of ivabradine were added to the cultures. Immediately thereafter, 100 ng/mL LPS (black columns) or medium without LPS (white columns) was applied to the respective cultures and the beating-rate of the cells was measured after 24 h. The number of measurements included in the calculation: (0 Iva, n = 54; 5 Iva, n = 30; 10 Iva, n = 6; 25 Iva, n = 18). Significances were calculated by the Mann–Whitney test. ns, not significant.
Fig. 3 Nifedipine further reduces the effect of LPS on the beating-rate of neonatal rat cardiomyocytes. CM (1.2 9 105 cells/cm2) were cultured in 6-well plates. Different concentrations of the calcium-channel blocker nifedipine were added to the cultures. Subsequently, 100 ng/mL LPS (black columns) or medium without LPS (white columns) was applied to the respective cultures and the beating-rate of the cells was measured after 24 h. The data of 30 measurements for each condition are summarized. Significances were calculated by the Mann–Whitney Test.
an elevation of the NN-interval from 124.1 5.8 to 183.8 15.3 ms). As expected, the combined inhibition of the autonomic nervous system by simultaneous injection of both, atropine and propranolol, did not significantly increase the NN interval (134.7 4.4 to 146.1 7.9 ms). On the other hand, sole injection of LPS induced tachycardia in mice, as shown by the reduction of the NN-interval from 129.4 7.9 to 105.2 5.6 ms (Fig. 4b). However, when LPS was given simultaneously with autonomic blockade (atropine + propranolol), a significant increase of the mean NN-interval from 123.3 4.7 to 176.9 5.8 ms was observed (Fig. 4b), in line with the above cell culture experiments (compare Fig. 1). Interference of LPS and ivabradine suggests the HCNchannels as a target of LPS In order to analyze whether the effect of LPS on the HR in vivo depends on the HCN-channels, injection experiments were per-
(b)
Fig. 4 (a, b) LPS causes tachycardia in conscious mice, which is reduced by blockade of the autonomic system. Heart-rate of CD1-mice was determined by telemetrical ECG-recording 15 min before (hatched columns) and 20 min after (black columns) the injection of the indicated compounds. Four mice were used in each group (n ≥ 15 injections per group). A, atropine (1 mg/kg); P, propranolol (20 mg/kg); LPS, lipopolysaccharide (0.4 mg/kg). Statistical analysis was performed using t-test. ns, not significant.
formed in the absence and the presence of the HCN-channel blocker ivabradine (Iva). In accordance with previous studies, a dose-dependent HR-lowering effect of ivabradine was observed up to 10 mg/kg (not shown), whereas a further increase of the dosage did not show a significant effect. A representative tracing is shown in Fig. 5a. Thus, the concentration of 10 mg/kg was used for all further experiments. In the present experiment (Fig. 5b), ivabradine expectedly enhanced the NN-interval. Also shown in Fig. 5b, injection of the HCN-channel blocker ivabradine with subsequent injection of LPS did not significantly alter the HR of the mice, as compared to ivabradine injected solely. Furthermore, in the situation of autonomic blockade induced by atropine and propranolol, as described above, LPS failed to induce a further increase of the NN-interval, if the mice were simultaneously treated with ivabradine (Fig. 5c). These in vivo findings also argue for HCN-channels as the common target of ivabradine and LPS, as suggested by the in vitro data shown above. Besides HCN4, other HCN-channels may be involved in the response to LPS Previous experiments suggested that LPS is not only able to interfere with HCN4, but also with other HCN-isoforms.21,22 Therefore, we analyzed the effect of LPS on the sinus node cycle length in HCN4-knock-out animals. For this purpose, animals were used, where one HCN4-allele was replaced by a tamoxifen-sensitive Crerecombinase (HCN4+/ ; hatched columns), while the second HCN4-allele was marked by loxP-sites, thereby enabling the induction of a complete HCN4-knock-out (HCN4 / ; black columns).25 The HCN4+/ -mice first underwent a series of pharmacological injections before the remaining HCN4-allele was knocked out by tamoxifen (HCN4 / ). The basal sinus node cycle length of the HCN4+/ -mice was 128.0 5.6 ms and did not differ significantly from that of wild type animals (123.3 4.7, compare Fig. 4). As shown in Fig. 6, the complete loss of HCN4-channels (HCN4 / ) did not significantly alter the response of the animals to the injection of atropine, but reduced the effect of the betablockade by propranolol, as compared to HCN4+/ -animals. However, the effect of LPS on HR was found to be unchanged in HCN4 / -mice, both under basal conditions as well as under
© 2015 Wiley Publishing Asia Pty Ltd
HCN-channels and endotoxin
877
(a1)
(a2)
(c)
(b)
Fig. 5 (a, b, c) Interference of LPS and ivabradine suggests the HCNchannel as a target of LPS. (a) Representative examples of ECG tracings in awake mice using telemetric transmitters. (a1) Control; (a2) Ivabradine, 20 min after injection of Iva. (b, c) Heart-rate of CD1-mice was determined by telemetrical ECG-recording 15 min before (hatched columns) and 20 min after the injection (black columns) of the indicated compounds (n ≥ 15 injections per group). A, P, LPS, compare Fig. 4; Iva, ivabradine (10 mg/kg). Statistical analysis was performed using the t-test. ns, not significant.
autonomic blockade, suggesting that further HCN-channels, such as HCN2, may contribute to the LPS effect. Taken together, the above findings are in line with the hypothesis that LPS, in the presence of autonomic innervation, leads to a sensitization of the HCN-channels to sympathetic stimuli, with the consequence of an elevated HR. However, when LPS is given in the absence of autonomic stimulation it directly reduces the HR. The effect of LPS on the sinus node seems to be dependent on HCN-channels because it can be prevented by the HCNchannel blocker ivabradine. The interaction of LPS is not limited to the HCN4-channel, but other HCN-channel subtypes may also interact.
DISCUSSION Functional disturbances of otherwise healthy organs are the hallmarks of severe sepsis and septic shock, leading to septic organ dysfunction. Especially the impairment of renal function or septic encephalopathy are well known in this regard. However, although not recognized so easily, the heart can also suffer from sepsisinduced injury,26,27 predominantly of diastolic nature,26 which
Fig. 6 Besides HCN4, other HCN-channels may be involved in the response to LPS. CD1-mice (wild-type) were used where one HCN4allele was replaced by a tamoxifen-sensitive Cre-recombinase while the second HCN4-allele was marked by loxP-sites (HCN4CreERT2/HCN4loxP). These mice, containing one functionally intact HCN4-allele (HCN4+/ ; hatched columns) first underwent the injections indicated in the Methods, before the remaining HCN4-allele was knocked out by tamoxifen and the injections got repeated 20 days later (HCN4 / ; black columns) (n ≥ 15 injections per group). A, P, LPS, Iva, compare Fig. 5. Statistical analysis was performed using the t-test. ns, not significant.
can be quantified as impairment in cardiac function.27 Important components of septic cardiomyopathy are abnormalities in the regulation of HR and HRV.14 These particular abnormalities have been repeatedly shown to be of high prognostic relevance in critically ill patients at the ICU so that it seems to be of special interest to understand the underlying mechanisms.5 We have shown previously that LPS, which is a component of the cell membrane of Gram-negative bacteria, can interact with HCN-channels, which are crucially involved in cardiac pacemaking.21–23 Experiments with isolated human cardiomyocytes had demonstrated that LPS can directly influence the ‘funny current’ If in two ways: (i) LPS can lead to a sensitization of the HCNchannels to adrenergic stimuli; but (ii) in parallel also can directly inhibit HCN-channels.22 Therefore, we analyzed whether or not the previous findings also apply to functional in vitro cell culture experiments and to the even more complex in vivo situation in mice. A number of experiments have addressed the effects of LPS on HR regulation before.28,29 Many of these studies utilized experimental systems where changes occurred with a delay of several hours after LPS exposure, or last over 1–24 h. This may involve the host’s immune system and alterations of the central nervous system.30 In order to avoid this interference, we performed in vivo analyses measuring immediately after LPS-administration, i.e. within 30 min. The present in vivo studies indeed confirmed the results of the authors’ previous and present cell culture experiments: while in the presence of an intact autonomic nervous system in the mice i.p. injection of LPS led to tachycardia, LPS induced bradycardia when the autonomic modulation was pharmacologically blocked (Figs 4 and 5). In the cell culture experiments with CM – which may be considered as a model of isolated pacemaker cells – only the direct bradycardic effect can be detected, due to the lack of the influence of the autonomic nervous system. The present study also supports the view that the immediately occurring effect of LPS on the HR is based on an interaction with cardiac HCN-channels, which are responsible for the ‘funny
© 2015 Wiley Publishing Asia Pty Ltd
878
H Ebelt et al.
current’ as the molecular basis of pacemaking. This interpretation is based on the finding that both, in mice as well as in experiments with CMs, the effect of LPS is not altered by the blockade of calcium-channels, but can be prevented by the administration of the HCN-channel blocker ivabradine. In mice and humans, the HCN-family is composed of four different members named HCN1 to HCN4.6 In the past, the majority of studies investigating the interplay between LPS and cardiac pacemaking focused on HCN2 for several reasons. First, the initial studies were performed in cardiomyocytes isolated from human right atria that mainly express the HCN2-isoform.22 Additionally, in 2007 we have reported that LPS reduces the beatingrate variability of CM.28 In these cells HCN2 is the main isoform, with a five times higher amount of mRNA detectable than of HCN4.9,31 However, in mice and humans, HR is under normal conditions determined by the activity of the sinus node where HCN4 is expressed in much higher amounts.9 Accordingly, we have shown previously that in HEK293-cells, which stably expressed hHCN4 or hHCN2, respectively, LPS exerts comparable effects on both channel subtypes.21 In order to address the question whether the effects of LPS on sinus node activity in vivo are restricted to HCN4, transgenic animals were included into the study where HCN4 can be knockedout by the injection of tamoxifen. It has been shown previously in detail that HR regulation occurs quite normally in these animals after HCN4-knock-out and that disturbances of cardiac pacemaking are seen especially in situations of increased repolarizing currents, as induced by muscarinic stimulation or during the transition from activated to basal cardiac state while no limitations were found under b-adrenergic stimulation.25 In accordance with this suggestion, we found comparable effects induced by atropine before and after the knock-out of HCN4, whereas the effect of the beta-blocking agent propranolol was reduced in the absence of HCN4. However, injection of LPS led to the same effects on HR both under basal conditions, as well as under autonomic blockade, irrespective whether or not HCN4 was expressed. This indicates that the HCN-channels remaining after HCN4-knockout, such as HCN1 or HCN2, are also affected by LPS. If autonomic dysfunction in MODS, indicated by narrowed HRV and inadequately high HR, is linked to an unfavourable outcome, improvement of autonomic function might improve prognosis of these patients. First therapeutic attempts in this direction have tried to reduce the inadequately high HR in patients with septic shock by the short acting betablocker esmolol,32 as well as by the If-blocker ivabradine.33 In a randomized study with 154 patients with septic shock and a HR ≥ 95 b.p.m., esmolol achieved a mean HR reduction of 18 b.p.m.. The 28-day mortality in the esmolol group was 49.4% and in the control group 80.5% (adjusted hazard ratio 0.39; 95% confidence interval 0.26–0.59; P < 0.001).32 In the randomized MODIFY trial33 70 patients with MODS (APACHE II score ≥ 20), a HR ≥ 90/min and existing contraindications to betablockers, were randomly treated with standard therapy or standard therapy plus a 4-day treatment with ivabradine. The primary endpoint of this trial was the percentage of patients with a reduction of the mean HR of at least 10 b.p.m., 96 h after the start of trial treatment. Recruitment of the study is already closed and data analysis is still ongoing.33 While in the esmolol and the MODIFY trial the inadequately high HR of the patients was reduced by blocking sympathetic
drive and If pharmacologically, the consequence of a direct blockade of LPS by antibodies or immunoglobulins on HR and on HRV have not yet been studied in critically ill patients. Taken together, the data presented here add further evidence to the suggestion that LPS can directly interact with cardiac pacemaking by the interference with sinuatrial HCN-channels. This interaction can be considered to be a mechanism leading to altered heart-rate regulation observed in severe sepsis that is linked to an unfavorable outcome of critically ill patients – but which might also bear therapeutic potential in the future.
MATERIAL AND METHODS Isolation and culture of neonatal rat cardiomyocytes Spontaneously beating neonatal rat cardiomyocytes (CM) were isolated from the heart of 1 to 3 days old Wistar rats by the method described by Werdan and Erdmann34. Briefly, beating hearts were taken from neonates sacrificed by cervical dislocation, rinsed in cold PBS-AG1 solution and bruised using a scalpel. The bruised hearts were incubated in dissociation solution (0.03 % collagenase; 0.12 % trypsin; in PBS-AG1-solution; 37°C) for 10 min and the supernatant was collected. The dissociation procedure was repeated 10 times, the first supernatant was discarded. The tubes containing the CM in CMRL-growth-medium (compare34) were centrifuged (1,400 rpm, 10 min, room temperature) and the CM resuspended in CMRL-growth-medium. This cell suspension was filtered (cell strainers; 70 lm) and transferred into 175 cm2 culture flasks for 1.5 h (37°C) in order to separate CM from non-muscle cells adhering to the culture dish (differential attachment technique). The cell suspension containing the non-adhesive cardiomyocytes was transferred into culture flasks or plates in CMRL-growth-medium (1.2 9 105 cells/cm2). After 24 hours incubation at 37°C, the culture medium was replaced by CMRL-synthetic-medium (compare34) and the CM were used for the experiments after further 24 hours of incubation. Determination of the beating-rate in CM The beating-rate of isolated CM was recorded using a photo-electrical system.28 Cells cultured in 6-well plates were analyzed under the microscope in a chamber keeping the temperature constantly at 37°C. Three regions of each well were measured for 72 s. For each condition, six measurements were performed. The frequency of the CM contraction, as well as parameters of beating-rate variability were analyzed in a program specifically designed using the software LABVIEW.35 ECG recordings in mice All animal experiments shown here comply with the Directive 2010/63/EU of the European Parliament and were approved by the local review board. Adult male CD1-mice were used for the in vivo studies. For transmitter implantation, mice underwent general anaesthesia with 2.5% isofluran, were placed on a prewarmed operation table (38°C) and received a subcutaneous injection of the analgetic carprofen (5 mg/kg). A median laparotomy was performed and an ECG-transmitter (PhysioTelTM
© 2015 Wiley Publishing Asia Pty Ltd
HCN-channels and endotoxin EA-F20; Data Science International, St Paul, MN, USA) was placed into the peritoneal cavity, recording both ECG-data and locomotory activity of the animals. The electrodes were guided through the peritoneum using a needle and placed subcutaneously on the right upper chest and in front of the lower left costal arch. The peritoneal cavity, as well as all skin lesions, were closed with surgical sutures. In order to avoid interference with the healing process, all injection experiments were started not earlier than 15 days after the transmitter implantation. In vivo injection protocol In order to analyze the influence of the autonomic nervous system on the effects of LPS on heart-rate, atropine sulfate (1 mg/ kg; B. Braun AG, Melsungen, Germany), propranolol hydrochloride (20 mg/kg; MIBE GmbH Arzneimittel, Sandersdorf-Brehna, Germany), ivabradine (10 mg/kg; Servier-Deutschland, M€ unchen, Germany), LPS (0.4 mg/kg; Sigma Aldrich, Taufkirchen, Germany) or combinations of these compounds were applied by intraperitoneal injection. All injections were performed not earlier than 15 days after the implantation of the ECG-transmitter and always between 4 p.m. and 6 p.m. in order to minimize circadian fluctuations. During the injections the animals received a shortlasting (approx. 3–5 min) anaesthesia induced by inhalation of 2.5% isoflurane. In the first part of the injection protocol, propranolol followed by atropine, ivabradine and finally the combination of atropine and propranolol were applied with a minimum of 24 h break between each injection. This cycle of injections was performed in triplicate with a 48 h break in between. In the second cycle of experiments, LPS followed by the combination of LPS + atropine + propranolol, ivabradine + atropine + propranolol, and LPS + ivabradine + atropine + propranolol were injected with a minimum of 48 h between the injections. At the end of the experiments, the animals were killed by cervical dislocation under general anaesthesia with 5% isoflurane. Each group consisted of four animals.
879
(ECG-tracings and locomotory activity) was performed using a standard telemetry system (Data Science International, PhysioTel R and MultiplusTM Implant, St. Paul, MN, USA). To analyze the effect of the different injections, a standardized approach was used. First, from both the time span before the injection ( 25 min until 10 min), as well as from the time after the injection (+20 min until +40 min) three intervals each lasting 3 min were manually selected, where physical activity of the mice was especially low or absent, as seen from the locomotory recordings. From these 3 min segments, both the mean NN-interval (meanNN3 min), as well as the locomotory activity score were calculated automatically. If the locomotory activity score was < 1, the meanNN3 min-interval of this segment was used to contribute to the mean NN-interval (meanNNbefore and meanNNafter) that was used finally for statistical analysis. Statistical analysis Statistical analysis was performed using SPSS (SPSS Inc., Chicago, IL, USA). All metric data are shown as mean standard error of the mean (SEM). Comparisons of two groups were performed using student’s t-test (paired or unpaired, respectively), or Mann–Whitney-test, as mentioned in the respective legends. Comparisons of multiple groups were done by ANOVA with post-hoc testing (Bonferroni). *P-values of < 0.05 were considered statistically significant.
ACKNOWLEDGEMENTS The expert technical assistance of S. Koch and C. Pilowski is gratefully acknowledged. This work was supported by grants of the DFG (German Research Foundation) to Ursula M€ uller-Werdan, Henning Ebelt and Michael Gekle (DFG MU 1010/4-1 and GE 905/18-1, respectively), by the Roux-Programm of the Medical Faculty of the Martin-Luther-Universit€at Halle-Wittenberg, and by a grant of Servier-Deutschland to Harald Loppnow, Henning Ebelt, and Karl Werdan.
DISCLOSURE
Conditional HCN4 knock-out mice The generation and phenotype of the tamoxifen-inducible HCN4knock-out mice has been described in detail before.25 In brief, a tamoxifen-inducible Cre-recombinase (CreERT2 construct) was ‘knocked in’ into the HCN4-locus. After crossing these animals with floxed HCN4-mice, complete deletion of the HCN4-gene can be achieved by injection of tamoxifen. Animals with a successful knock-down of HCN4-channels can be identified by their ECG-recordings, since they show regularly occurring sinuatrial blockades.25 In the present experiments, all mice with one functionally intact HCN4 allele (HCN4CreERT2/HCN4loxP; HCN4+/ ) first underwent the injection protocol described above. Afterwards, the animals were subjected to tamoxifen treatment (240 mg/kg for 3 days), leading to the complete HCN4-knockout (HCN4 / ). After an interval of at least 20 days, the initial pharmacological injection protocol was repeated again. ECG data collection During the ECG-recordings, all mice remained in their standard cages and could move normally. Data acquisition and analysis
Harald Loppnow, Henning Ebelt, and Karl Werdan received an unrestricted grant of Servier Deutschland. K. Werdan got honoraria from Servier company for participation in national and international advisory board activities, for his national leadership in the SHIFT study and for lectures.
REFERENCES 1. Kannel WB, Kannel C, Paffenbarger RS Jr, Cupples LA. Heart rate and cardiovascular mortality: The Framingham Study. Am. Heart J. 1987; 113: 1489–94. 2. Seccareccia F, Pannozzo F, Dima F et al. Heart rate as a predictor of mortality: The MATISS project. Am. J. Public Health 2001; 91: 1258–63. 3. Swedberg K, Komajda M, B€ohm M et al. Ivabradine and outcomes in chronic heart failure (SHIFT): A randomised placebo-controlled study. Lancet 2010; 376: 875–85. 4. Hoke RS, M€uller-Werdan U, Lautenschl€ager C, Werdan K, Ebelt H. Heart rate as an independent risk factor in patients with multiple organ dysfunction: A prospective, observational study. Clin. Res. Cardiol. 2012; 101: 139–47.
© 2015 Wiley Publishing Asia Pty Ltd
880
H Ebelt et al.
5. Schmidt H, M€ uller-Werdan U, Hoffmann T et al. Autonomic dysfunction predicts mortality in patients with multiple organ dysfunction syndrome of different age groups. Crit. Care Med. 2005; 33: 1994–2002. 6. Benarroch EE. HCN channels: Function and clinical implications. Neurology 2013; 80: 304–10. 7. Tokimasa T, Akasu T. Cyclic AMP regulates an inward rectifying sodium-potassium current in dissociated bull-frog sympathetic neurones. J. Physiol. 1990; 420: 409–29. 8. Yu FH, Yarov-Yarovoy V, Gutman GA, Catterall WA. Overview of molecular relationships in the voltage-gated ion channel superfamily. Pharmacol. Rev. 2005; 57: 387–95. 9. Shi W, Wymore R, Yu H et al. Distribution and prevalence of hyperpolarization-activated cation channel (HCN) mRNA expression in cardiac tissues. Circ. Res. 1999; 85: e1–6. 10. Heusch G. Heart rate in the pathophysiology of coronary blood flow and myocardial ischaemia: Benefit from selective bradycardic agents. Br. J. Pharmacol. 2008; 153: 1589–601. 11. Bernardi L, Salvucci F, Suardi R et al. Evidence for an intrinsic mechanism regulating heart rate variability in the transplanted and the intact heart during submaximal dynamic exercise? Cardiovasc. Res. 1990; 24: 969–81. 12. Jalife J, Slenter VA, Salata JJ, Michaels DC. Dynamic vagal control of pacemaker activity in the mammalian sinoatrial node. Circ. Res. 1983; 52: 642–56. 13. Schwartz PJ. Cardiac sympathetic innervation and the prevention of sudden death. Cardiologia 1990; 35: 51–4. 14. Werdan K, Schmidt H, Ebelt H et al. Impaired regulation of cardiac function in sepsis, SIRS, and MODS. Can. J. Physiol. Pharmacol. 2009; 87: 266–74. 15. Papaioannou VE, Verkerk AO, Amin AS, de Bakker JM. Intracardiac origin of heart rate variability, pacemaker funny current and their possible association with critical illness. Curr. Cardiol. Rev. 2013; 9: 82–96. 16. Loppnow H, Brade H, Rietschel ET, Flad HD. Induction of cytokines in mononuclear and vascular cells by endotoxin and other bacterial products. Methods Enzymol. 1994; 236: 3–10. 17. Loppnow H, Libby P. Adult human vascular endothelial cells express the IL6 gene differentially in response to LPS or IL1. Cell. Immunol. 1989; 122: 493–503. 18. Schonbeck U, Herzberg M, Petersen A et al. Human vascular smooth muscle cells express interleukin-1beta-converting enzyme (ICE), but inhibit processing of the interleukin-1beta precursor by ICE. J. Exp. Med. 1997; 185: 1287–94. 19. M€ uller-Werdan U, Schumann H, Loppnow H et al. Endotoxin and tumor necrosis factor alpha exert a similar proinflammatory effect in neonatal rat cardiomyocytes, but have different cardiodepressant profiles. J. Mol. Cell. Cardiol. 1998; 30: 1027–36. 20. Westphal E, Li C, Pilowski C et al. Endotoxin-activated cultured neonatal rat cardiomyocytes express functional surface-associated interleukin-1alpha. J. Endotoxin Res. 2007; 13: 25–34.
21. Kl€ockner U, Rueckschloss U, Grossmann C et al. Differential reduction of HCN channel activity by various types of lipopolysaccharide. J. Mol. Cell. Cardiol. 2011; 51: 226–35. 22. Zorn-Pauly K, Pelzmann B, Lang P et al. Endotoxin impairs the human pacemaker current If. Shock 2007; 28: 655–61. 23. Kl€ockner U, Rueckschloss U, Grossmann C et al. Inhibition of cardiac pacemaker channel hHCN2 depends on intercalation of lipopolysaccharide into channel-containing membrane microdomains. J. Physiol. 2014; 592: 1199–211. 24. Scheruebel S, Koyani CN, Hallstr€om S et al. I(f) blocking potency of ivabradine is preserved under elevated endotoxin levels in human atrial myocytes. J. Mol. Cell. Cardiol. 2014; 72: 64–73. 25. Herrmann S, Stieber J, St€ockl G, Hofmann F, Ludwig A. HCN4 provides a ‘depolarization reserve’ and is not required for heart rate acceleration in mice. EMBO J. 2007; 26: 4423–32. 26. Landesberg G, Gilon D, Meroz Y et al. Diastolic dysfunction and mortality in severe sepsis and septic shock. Eur. Heart J. 2012; 33: 895–903. 27. Werdan K, Oelke A, Hettwer S et al. Septic cardiomyopathy: Hemodynamic quantification, occurrence, and prognostic implications. Clin. Res. Cardiol. 2011; 100: 661–8. 28. Schmidt H, Saworski J, Werdan K, M€uller-Werdan U. Decreased beating rate variability of spontaneously contracting cardiomyocytes after co-incubation with endotoxin. J. Endotoxin Res. 2007; 13: 339–42. 29. Gholami M, Mazaheri P, Mohamadi A et al. Endotoxemia is associated with partial uncoupling of cardiac pacemaker from cholinergic neural control in rats. Shock 2012; 37: 219–27. 30. Fairchild KD, Saucerman JJ, Raynor LL et al. Endotoxin depresses heart rate variability in mice: Cytokine and steroid effects. Am. J. Physiol. Regul. Integr. Comp. Physiol. 2009; 297: R1019–27. 31. Herrmann S, Stieber J, Ludwig A. Pathophysiology of HCN channels. Pflugers Arch. 2007; 454: 517–22. 32. Morelli A, Ertmer C, Westphal M et al. Effect of heart rate control with esmolol on hemodynamic and clinical outcomes in patients with septic shock: A randomized clinical trial. JAMA 2013; 310: 1683–91. 33. Nuding S, Ebelt H, Hoke RS et al. Reducing elevated heart rate in patients with multiple organ dysfunction syndrome by the I (f) (funny channel current) inhibitor ivabradine: MODI (f)Y trial. Clin. Res. Cardiol. 2011; 100: 915–23. 34. Werdan K, Erdmann E. Preparation and culture of embryonic and neonatal heart muscle cells: Modification of transport activity. Methods Enzymol. 1989; 173: 634–62. 35. Lehnich H, Kaltenborn G, M€uller-Werdan U, Pauer H-D. Digitalisierung, Verarbeitung und Auswertung von Videoaufzeichnungen in Echtzeit zur Kontraktilit€atsmessung an isolierten Herzzellen. Virtuelle Instrumente in der Praxis 2001; 2001: 312–7.
© 2015 Wiley Publishing Asia Pty Ltd
