Thrombin Facilitation of Voltage-Gated Sodium Channel Activation in Human Cardiomyocytes
Implications for Ischemic Sodium Loading
Background— Thrombin plays a role in mediating ischemic injury and cardiac arrhythmias, but the mechanisms involved are poorly understood. Because voltage-gated sodium channels (VGSCs) have not previously been considered, putative effects of thrombin on VGSC function were investigated in human isolated cardiomyocytes.
Methods and Results— Sodium current (INa) was recorded by the whole-cell patch-clamp method. Thrombin increased peak INa amplitude in an activity-dependent manner, from 1 to 100 U/mL, with an apparent EC50 of 91±16 U/mL. When tested at 32 U/mL, thrombin-increased INa was abolished by tetrodotoxin (50 μmol/L). Thrombin effects on INa were reversible and repeatable, and 100 U/mL doubled peak INa amplitude. Thrombin (32 U/mL) shifted INa activation to hyperpolarized potentials without affecting steady-state inactivation, producing unusually large increases in window current. Hirudin (320 U/mL) or haloenol lactone suicide substrate (10 μmol/L) failed to significantly affect these effects of thrombin. In current-clamped cardiomyocytes, thrombin (32 U/mL) depolarized resting membrane potential by 10 mV.
Conclusions— Facilitation of VGSC activation causing large increases in window current is a major mechanism by which thrombin may promote ischemic sodium loading and injury.
Received May 29, 2002; revision received July 29, 2002; accepted August 6, 2002.
Besides its well-known role in the circulation during coagulation in cleaving fibrinogen, thrombin also exerts direct actions on cardiomyocytes.1 It has been postulated that the serine protease plays a significant role in mediating ischemic injury, malignant ventricular arrhythmias, and sudden death during acute thrombotic occlusion.2–6⇓⇓⇓⇓ However, the molecular mechanisms underlying these deleterious effects of thrombin are poorly understood.
In cardiomyocytes, thrombin elevates [Na+]i by Na+-H+ exchange activation7 or via production of membrane-derived lysophosphatidylcholine.4,5,8⇓⇓ A possible involvement of voltage-gated sodium channels (VGSCs) in mediating thrombin-induced increases in [Na+]i has not, however, been previously considered. Increases in [Na+]i in ischemic cardiomyocytes generate Ca2+ loading via reverse Na+-Ca2+ exchange, leading to contractile dysfunction and ultimately cell death.9–11⇓⇓ Na+-H+ exchange activity during ischemia is inactive or weak12 because the exchanger is subject to inhibition by extracellular acidosis, 13 which suggests a more significant role for VGSCs as mediators of ischemic Na+ loading.
A large fraction of VGSCs become rapidly nonrecruitable in ischemic tissues after resting membrane potential depolarization, and action potentials (APs) initially shorten and subsequently cease with exhaustion of cellular ATP. Na+ influx continues, however, through noninactivated VGSCs, giving rise to persistent window currents14,15⇓ and further cardiomyocyte depolarization. Moreover, the slowly inactivating component of Na+ current also increases substantially during ischemia, amplifying Na+ influx.15–18⇓⇓⇓
The putative effects of thrombin on VGSCs in human atrial myocytes were studied by the patch-clamp technique. The data demonstrate, for the first time, that thrombin acts as a powerful activator of human cardiac VGSCs, constituting a major mechanism by which the protease increases [Na+]i in ischemic cardiomyocytes.
Heart Tissue Samples
Protocols for obtaining human cardiac tissue were approved by the Hôpital Bicêtre, Université Paris XI Ethics Committee. Specimens of human right atrial appendages were obtained from hearts of patients (54 to 77 years of age) undergoing heart surgery for coronary artery bypass graft or valve replacement. Of the patients involved, 8% had received β-blockers, 38% calcium antagonists, 44% antiulcer drugs, 55% analgesics, 10% diuretics, and 22% antithrombotics. Treatments were usually stopped 24 hours before operation. Patients with atrial dilation were avoided, and none had a history of supraventricular arrhythmias.
Human atrial myocytes were isolated enzymatically as previously described.19 Only quiescent rod-shaped myocytes with clear cross-striations, sharp edges, and well-delineated cell membranes were chosen for experiments. Small myocytes were preferred to optimize spacial voltage-clamp.
Solutions and Drugs
For whole-cell current recordings, the intracellular pipette solution contained (in mmol/L): NaCl 5, CsF 130, MgCl 2, CaCl2 1, EGTA 15, HEPES 10, and MgATP 4, pH 7.2 with CsOH. The external solution contained: NaCl 25, CsCl 108.5, CoCl 2.5, CaCl2 0.5, MgCl 2.5, 4-AP 5, HEPES 10, and glucose 10, pH 7.4 with CsOH. To record APs, myocytes were bathed in Tyrode’s solution (in mmol/L): NaCl 135, KCl 4, MgCl 2, CaCl2 1.8, HEPES 10, NaH2PO4 1, Na-pyruvate 2.5, and glucose 20, pH 7.4 with NaOH. The pipette solution contained (in mmol/L): K-aspartate 115, NaCl 5, KCl 15, MgCl 2, HEPES 10, glucose 10, MgATP 3, tris-phosphocreatine 5, and EGTA 5, pH 7.2 with KOH. Hirudin (leech, ≈2000 U/mg) was from Boehringer Mannheim. Haloenol lactone suicide substrate (HELSS) from TEBU France was dissolved in dimethyl sulfoxide (DMSO) that did not exceeded 0.05% and thrombin (human plasma ≈1000 U/mg). Other chemicals were purchased from Sigma.
Current and AP Recordings
Ionic currents were recorded by the whole-cell patch-clamp technique with a patch-clamp amplifier (Axopatch 200B, Axon Instruments). Patch pipettes (Corning Kovar Sealing code 7052, WPI) had resistances of 0.5 to 1 MΩ. Currents were filtered at 20 kHz (−3 dB, 8-pole low-pass Bessel filter) and digitized at 50 kHz (Digidata 1200, Axon Instruments).
Cell membrane capacitance was determined according to Momtaz et al.20 Cell membrane capacitance was 47.9±2.4 pF (n=113 cells, from 68 donors). Series resistance was compensated at 80% to 95%, resulting in voltage errors of <3 mV, whereas neither cell membrane capacitive current nor leakage current was compensated. In sodium current (INa) recording media, the average holding current densities at a holding potential (HP) of −100 mV were −1.8±0.5 pA/pF in control, −1.9±0.4 pA/pF under 32 U/mL thrombin, −2.0±0.4 pA/pF under tetrodotoxin, and −1.9±0.4 pA/pF after washout (not significant respective to control [n=17]). Peak INa amplitude was monitored according to a steady-state pulse protocol: a 300-ms depolarizing pulse to −30 mV from a HP of −100 mV at 0.2 Hz. An equilibration period was allowed until peak INa reached steady state and remained stable without evidence of a leftward shift of the availability-voltage relationship (hα-Vm). This protocol was designed to detect any such shifts. After each sequence of 5 depolarizing pulses, HP was set to −140 mV. Consequently, during the stabilization period when peak current amplitude was higher after a HP at −140 mV compared with −100 mV, the recording was discarded. Similarly, because many cells showed no effect of thrombin on the hα-Vm relationship, recordings showing an irreversible hyperpolarizing shift of hα-Vm after thrombin application were also discarded.
The steady-state pulse protocol was applied before activation-Vm, availability-Vm, and reactivation protocols in control, thrombin and washout. For the activation-Vm protocol, currents were elicited by 100-ms depolarizing pulses applied at 0.2 Hz from HP −140 mV, in 5- or 10-mV increments between −100 and +50 mV. The steady-state availability-Vm protocol was from a HP of −100 mV; a 2-s conditioning prepulse was applied in 5- or 10-mV increments between −140 and −30 mV followed by a 300-ms test pulse to −20 mV at 0.2 Hz.
The recovery from inactivation protocol was from a HP of −100 mV; a 500-ms pulse followed by a 30-ms pulse was applied to −10 mV. The interval between the two pulses varied from 3 to 1000 ms.
APs were recorded in the whole-cell current-clamp configuration. The recording pipette had a resistance of 1 to 3 MΩ. Under control conditions, cells were current-clamped to a resting potential of −80 mV. APs were then elicited by a 2-ms hyperpolarizing current pulse (to reach a Vm of ≈−100 mV), followed by a 2-ms depolarizing pulse of twice diastolic threshold at 0.2 Hz.
Experiments were carried out at room temperature (22° to 25°C).
Statistics and Data Analysis
Data are expressed as mean±SEM of n determinations or myocytes, and when relevant, the number (N) of preparations was specified. Statistical significance was estimated by paired or unpaired Student’s t test or ANOVA, as appropriate. P<0.05 was considered significant.
Data for activation-Vm and steady-state availability-Vm relationships of INa were fitted to the Boltzmann equation: equation
where Vm is the membrane potential, V0.5 is the half-activation or half-availability potential, and k is the inverse slope factor. For activation-Vm curves, Y represents the relative conductance and k is >0. For availability-Vm curves, Y represents the relative current (INa/INamax) and k is <0.
The same algorithm was applied to the Hill form (see legend to Figure 2) of the thrombin activity-response relationship and to the 2-exponential model for the time course of recovery from inactivation of INa (see legend to Figure 5).
Thrombin Increases Peak INa
Figure 1 shows representative recordings of INa obtained in the absence and presence of thrombin. At 0.1 U/mL, thrombin had no detectable effect on INa (data not shown). Figure 1A shows that thrombin (1 or 10 U/mL) induced a reversible increase of mean peak INa amplitude normalized to cell capacitance, from −141.7±37.2 to −148.1±37.3 pA/pF (9 myocytes; N=6; P<0.05) and from −143.6±12.8 to −187.2±18.8 pA/pF (n=25; N=14; P<0.01), respectively. Similarly, Figure 1B shows that thrombin (32 U/mL) induced a marked increase of peak INa, from −118.9±12.6 to −191.7±20.8 pA/pF (mean of 23 myocytes; N=17; P<0.001), which was abolished by 50 μmol/L tetrodotoxin, a specific voltage-gated sodium channel blocker. Current activation by thrombin and its blockade by tetrodotoxin were fully reversible during washout (Figure 1B). Thrombin (100 U/mL) exerted more marked and reversible increases in peak INa (−117.9±34.6 versus −245.2±61.1 pA/pF, n=7; N=5; P<0.001) (Figure 1C) after repeated applications. These increases occurred despite a reversible hyperpolarizing shift of steady-state availability-Vm. A clear indication of this shift is given by the INa points elicited from a HP of −140 mV (Figure 1C). In addition, thrombin (100 U/mL) induced a marked increase of the slow component of INa measured at the end of the depolarizing test pulse (−1.9±0.3 versus −3.1±0.5 pA/pF, n=7; N=5; P<0.01). Time to peak INa remained unchanged in the presence of thrombin (0.1 to 10 U/mL), whereas with 32 and 100 U/mL, it was significantly decreased from 0.88±0.18 to 0.75±0.17 ms (n=17; N=17; P<0.001) and from 0.93±0.13 to 0.62±0.14 ms (n=9; N=5; P<0.001), respectively. Application of heat-inactivated thrombin, at a nominal activity of 32 U/mL, failed to increase peak INa (−67.9±14.6 in control versus −77.1±17.8 pA/pF under heat-inactivated thrombin; n=7; N=4; NS).
Thrombin increased peak INa in an activity-dependent manner as shown in Figure 2. The maximal effect was reached at 100 U/mL thrombin because higher concentrations elicited myocyte hypercontracture. Thrombin (100 U/mL) doubled peak INa compared with controls, indicating an unusually large augmentation of INa.
Current Density-Voltage Relationships
Figure 3A shows typical examples of INa recordings obtained in the absence and presence of 32 U/mL of thrombin with incremental test pulses of the activation-Vm protocol. The current density-voltage relationships obtained are presented in Figure 3B. The peak INa density obtained with a depolarizing test pulse to −40 mV was significantly increased from −64.7±6.6 pA/pF in controls to −132.4±12.2 pA/pF (P<0.001) by thrombin. Membrane potential values (Vm) associated with maximal amplitude of peak current were −20 mV in controls and ≈−40 mV with thrombin. The activation-Vm relationship was displaced toward more negative membrane potentials. The V0.5 (Vm at which half-activation occurs) was significantly shifted by −9.2 mV (Table and Figure 4B). This effect was reversible on thrombin washout (Table and Figure 4B). Finally, thrombin (32 U/mL) was devoid of significant effect on the INa reversal membrane potential (+34 versus +36 mV in control, NS; Figure 3B) indicating that sodium channel selectivity was unaffected.
Steady-State Availability-Voltage Relationships
Figure 4A shows typical examples of INa recordings obtained in absence and presence of thrombin (32 U/mL) according to the steady-state availability-Vm protocol. Figure 4B and the Table demonstrate that thrombin (32 U/mL) was devoid of significant effect on steady-state availability relationships (Table).
The overlap of the steady-state availability-Vm curve and the activation-Vm curve in the potential range −85 to −40 mV determines the sodium window current amplitude as depicted in Figure 4B and 4C. Because sodium channels are not totally inactivated, a small fraction of these channels remain open and mediate window current. In the presence of thrombin, the INa activation curve was displaced toward hyperpolarizing potentials, which gave rise to large increases in window current amplitude in the −95 to −40 mV range of potentials (Figure 4C).
The time course of INa reactivation was assessed at a HP of −100 mV by the conventional double-pulse method (Figure 5). Time courses of relative current in control and thrombin (32 U/mL) were fitted by a double exponential function (Methods). The data in the Table confirm that there was no difference between the two constants of reactivation in controls versus thrombin.
Effects of Hirudin and HELSS
Figure 6 shows a typical recording of a relatively high concentration of hirudin (320 U/mL), a direct thrombin inhibitor,21 on thrombin-increased INa elicited according to the steady-state pulse protocol (see Methods). Hirudin (320 U/mL) per se failed to significantly affect peak INa (−58.7±13.4 pA/pF in controls versus −60.1±17.3 pA/pF with hirudin; n=6, NS). Even at this high concentration, hirudin failed to affect thrombin-induced increases in peak INa (−122.5±22.4 pA/pF under thrombin, −100.4±21.6 pA/pF after thrombin plus hirudin, n=7, NS). HELSS (10 μmol/L), an inhibitor of calcium-independent phospholipase A2,8 was unable to inhibit the effects of thrombin on peak INa (−79.1±16.3 pA/pF after a 30-minute preincubation with HELSS versus −120.6±20.3 pA/pF under thrombin plus HELSS, n=6; N=6; P<0.01).
As thrombin markedly increased INa window, sarcolemmal depolarization is an expected consequence. Figure 7 shows the effect of thrombin (32 U/mL) on averaged APs recorded from current-clamped single myocytes. Resting membrane potential was significantly shifted to depolarized potentials, and maximum upstroke velocity and AP amplitude were significantly and reversibly decreased by thrombin (Table).
The effects of thrombin on VGSC function were studied in human isolated atrial cardiomyocytes. For the first time, it is demonstrated that thrombin markedly increased peak INa and window current amplitude. These effects were reversible and repeatable, and they are attributable to a shift in the INa activation-voltage relationship to hyperpolarized potentials without affecting steady-state voltage-inactivation characteristics. Moreover, thrombin depolarized cardiomyocytes. Because the direct thrombin inhibitor, hirudin, failed to attenuate VGSC activation by thrombin, enzymatic peptide cleavage is unlikely to be involved. The data suggest that thrombin could promote ischemic sodium loading and arrhythmogenesis by facilitating VGSC activation.
INa Facilitation by Thrombin
Thrombin is considered to play a role in mediating ischemic injury and ventricular arrhythmias during coronary occlusion.2–5⇓⇓⇓ However, the mechanisms underlying these deleterious effects are not well understood. In cardiomyocytes, thrombin increases [Na+]i by Na+-H+ exchange activation7 or via lysophosphatidylcholine production, 4,5⇓ but no consideration has been given to putative activation of VGSCs by the serine protease. The present study therefore attempted to determine whether VGSC activation by thrombin occurred in human cardiomyocytes.
Thrombin induced an increase in peak INa amplitude caused by a marked hyperpolarizing shift of the current activation curve. Consequently, thrombin potentiates the Na+ window current, generated by overlap of activation- and steady-state voltage-inactivation INa curves.18 In the presence of thrombin, the marked activation shift to more negative potentials produced abnormally large increases in window current amplitude. This generates a large, sustained depolarizing current and potentially massive Na+ influx at the resting membrane potential during ischemia. The consequences are a depolarization of resting membrane potential associated with a decrease in AP maximum upstroke velocity. Persistent Na+ current is caused by sustained VGSC openings.15 These late channel reopenings and bursting behavior occur with INa in animal22,23⇓ and human24 ventricular cardiomyocytes. This increase in steady-state Na+ window current resulting from a prolongation of VGSC opening directly increases free [Na+]i.9 The present results suggest that thrombin, by facilitating Na+ channel activation causing abnormally large increases in window current, may amplify Na+ accumulation during ischemia, leading to elevated [Ca2+]i through reverse Na+-Ca2+ exchange.
Because the large, sustained sodium window current increase elicited by thrombin occurred in the voltage range −95 to −40 mV, this inward current will be further augmented by sarcolemmal depolarization during ischemia.15,17,24⇓⇓ Thus, the current- clamp data demonstrate that thrombin depolarized resting membrane potential by 10 mV, which could be a direct consequence of sodium window current amplification. These results corroborate those reported2 in depolarized ventricular fibers, in which thrombin increased automaticity. Depolarized fibers are thought to provide a substrate for arrhythmogenesis resulting from abnormal automaticity in myocardial infarction.18,2⇓ Human atrial cardiomyocytes express Na+ window current25 with operational characteristics similar to their ventricular counterparts,24 although we acknowledge that possible electrophysiological differences between atrial and ventricular cardiomyocytes may exist. The mechanism for thrombin-evoked automaticity could therefore be explained by the markedly increased sodium window current, which depolarizes resting membrane potential and which in addition may promote triggered activity.14,18,24⇓⇓
Thrombin Access to VGSCs
Thrombin concentrations that activated VGSCs in the present study correspond to those that are found during clotting26 and thrombosis.27 Nevertheless, in order to activate VGSCs, the enzyme must access cardiomyocytes across the capillary wall of the coronary microcirculation. Available evidence would indicate this to be highly likely to occur.28,29⇓
Mechanism of Action
Because the direct thrombin inhibitor, hirudin, failed to affect thrombin activation of VGSCs, proteolytic cleavage was not required. Consequently, an involvement of protease-activated receptors1 can be excluded. Thrombin may also increase [Na+]i by calcium-independent phospholipase A2–dependent production of lysophosphatidylcholine.4,5,8⇓⇓ However, HELSS, an inhibitor of cardiomyocyte phospholipase A2, failed to significantly affect thrombin increases in peak INa in the present study, which is suggestive of another mechanism. This issue merits further study, because the phospholipase A2–dependent4,5,8⇓⇓ and presently unidentified mechanism of thrombin-induced cardiomyocyte sodium loading are complementary and may well occur concomitantly. The precise molecular mechanism of thrombin increases in peak INa therefore remains to be elucidated. A direct nonproteolytic effect on VGSC α or β subunits or unidentified G protein–coupled receptors cannot be excluded at present.
In conclusion, thrombin has been shown for the first time to be a powerful activator of cardiac VGSCs by selectively shifting the INa voltage-activation relationship to hyperpolarized potentials. The consequences of such an effect are 2-fold: amplified ischemic sodium loading and arrhythmogenesis. These effects of thrombin are not dependent on proteolytic activity. Facilitation of VGSC activation may therefore be considered as a key molecular mechanism by which thrombin mediates ischemic injury.
Part of this work was supported by a grant from the Bonus Qualité Recherche (Université Paris-Sud XI) and from the Fondation de France. We thank Jean-Michel Talmant for technical support.
Presented in part at the International Society for Heart Research World Congress, Winnipeg, Manitoba, Canada, July 2001, and published in abstract form (J Mol Cell Cardiol. 2001;33:A94).
- ↵Steinberg SF, Robinson RB, Lieberman HB, et al. Thrombin modulates phosphoinositide metabolism, cytosolic calcium, and impulse initiation in the heart. Circ Res. 1991; 68: 1216–1229.
- ↵Goldstein JA, Butterfield MC, Ohnishi Y, et al. Arrhythmogenic influence of intracoronary thrombosis during acute myocardial ischemia. Circulation. 1994; 90: 139–147.
- ↵Park TH, McHowat J, Wolf RA, et al. Increased lysophosphatidylcholine content induced by thrombin receptor stimulation in adult rabbit cardiac ventricular myocytes. Cardiovasc Res. 1994; 28: 1263–1268.
- ↵Yasutake M, Haworth RS, King A, et al. Thrombin activates the sarcolemmal Na+-H+ exchanger: evidence for a receptor-mediated mechanism involving protein kinase C. Circ Res. 1996; 79: 705–715.
- ↵Haigney MC, Lakatta EG, Stern MD, et al. Sodium channel blockade reduces hypoxic sodium loading and sodium-dependent calcium loading. Circulation. 1994; 90: 391–399.
- ↵Rochitte CE, Kim RJ, Hillenbrand HB, et al. Microvascular integrity and the time course of myocardial sodium accumulation after acute infarction. Circ Res. 2000; 87: 648–655.
- ↵Eigel BN, Hadley RW. Antisense inhibition of Na+/Ca2+ exchange during anoxia/reoxygenation in ventricular myocytes. Am J Physiol Heart Circ Physiol. 2001; 281: H2184–H2190.
- ↵Undrovinas AI, Fleidervish IA, Makielski JC. Inward sodium current at resting potentials in single cardiac myocytes induced by the ischemic metabolite lysophosphatidylcholine. Circ Res. 1992; 71: 1231–1241.
- ↵Coulombe A, Coraboeuf E, Malecot C, et al. Role of the “Na window” current and other ionic currents in triggering early after-depolarizations and resulting re-excitations in Purkinje fibers. In: Zipes DP, Jalife J. eds. Cardiac Electrophysiology and Arrhythmias. Orlando, London: Grune and Stratton; 1985.
- ↵Sakmann BF, Spindler AJ, Bryant SM, et al. Distribution of a persistent sodium current across the ventricular wall in guinea pigs. Circ Res. 2000; 87: 910–914.
- ↵Zygmunt AC, Eddlestone GT, Thomas GP, et al. Larger late sodium conductance in M cells contributes to electrical heterogeneity in canine ventricle. Am J Physiol Heart Circ Physiol. 2001; 281: H689–H697.
- ↵Maltsev VA, Sabbah HN, Higgins RS, et al. Novel, ultraslow inactivating sodium current in human ventricular cardiomyocytes. Circulation. 1998; 98: 2545–2552.
- ↵Sakakibara Y, Wasserstrom JA, Furukawa T, et al. Characterization of the sodium current in single human atrial myocytes. Circ Res. 1992; 71: 535–546.
- ↵Vogel SM, Gao X, Mehta D, et al. Abrogation of thrombin-induced increase in pulmonary microvascular permeability in PAR-1 knockout mice. Physiol Genomics. 2000; 4: 137–145.