Cardiovascular Research

Cardiovascular Research 2005 67(2):263-273; doi:10.1016/j.cardiores.2005.03.006

This Article Abstract FREE Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Request Permissions Google Scholar Articles by Azene, E. M. Articles by Li, R. A. Search for Related Content PubMed PubMed Citation Articles by Azene, E. M. Articles by Li, R. A. Social Bookmarking          What's this?

Copyright © 2005, European Society of Cardiology

Non-equilibrium behavior of HCN channels: Insights into the role of HCN channels in native and engineered pacemakers

Ezana M. Azene, Tian Xue, Eduardo Marbán, Gordon F. Tomaselli and Ronald A. Li^*

Department of Medicine, Johns Hopkins University, 720 Rutland Avenue/Ross 1165, Baltimore MD 21205, United States

^* Corresponding author. Tel.: +1 410 614 0035. Email address: ronaldli{at}jhmi.edu

Received 2 September 2004; revised 9 March 2005; accepted 11 March 2005

	Abstract

Top Abstract 1. Introduction 2. Methods 3. Results 4. Discussion References

 
Objective: I_f, encoded by the hyperpolarization-activated, cyclic nucleotide-modulated (HCN) channel gene family, modulates cardiac pacing. During cardiac pacing, changes in membrane potential are rapid, preventing the very slow HCN channels from reaching equilibrium. Here, we examined the properties of HCN channels under non-equilibrium conditions to shed insight into how different HCN isoforms contribute to cardiac pacing.

Methods and results: HCN1, 2 and 4 channels were heterologously expressed in Xenopus laevis oocytes or mammalian Cos7 cells and subjected to voltage clamp. We found that HCN1 channel activation (V_1/2) depended strongly on the holding potential (V_H) for short (100 ms; V_1/2=–118 mV, –78 mV and –19 mV for V_H=+70, –75 and –140 mV, respectively, in Xenopus oocytes) but not long (300-ms) test-pulses, hinting that shifts of V_1/2 under non-equilibrium conditions may alter the impact of I_f in different phases of the cardiac circle. Consistent with this notion, when a train of SA nodal-like action potentials was applied in voltage-clamp experiments, HCN1 exhibited pronounced current–voltage (IV)-hysteresis. Using computational modeling, we demonstrate that the intrinsically sluggish HCN1 activation kinetics underlie their IV-hysteretic behavior and do not hinder the ability to modulate cardiac pacing. By contrast, HCN4 did not exhibit IV-hysteresis. This difference can be attributed to the relatively large activation time constant and markedly delayed onsets of time-dependent HCN4 currents. Indeed, HCN2 channels, which have intermediate activation time constants and delays, displayed and intermediate hysteretic phenotype.

Conclusion: We conclude that non-equilibrium properties of HCN channels contribute to cardiac pacing. These results provide insight for tuning the firing rate of endogenous and induced pacemakers using engineered HCN constructs with distinct gating phenotypes.

KEYWORDS HCN channel; Hysteresis; Ion channel; Non-equilibrium; Gating; Pacemaker; Sino atrial node

	1. Introduction

Top Abstract 1. Introduction 2. Methods 3. Results 4. Discussion References

 
I_f, encoded by the hyperpolarization-activated, cyclic nucleotide-modulated (HCN) channel gene family, is known to contribute significantly to cardiac pacing [1]. To date, four mammalian HCN isoforms (HCN1–4) have been identified, each with a distinct pattern of gene expression, tissue distribution [2] and unique kinetic properties. Of the two predominant isoforms found in the rabbit SA node [3], HCN1 reaches equilibrium in response to a hyperpolarizing test-pulse approximately 10 to 25 times faster than HCN4 [3,4].

The importance of HCN channel function to normal cardiac automaticity in mice [5] was recently corroborated in humans by the description of HCN4 mutations in two patients with idiopathic sinus node dysfunction [6,7]. Up-regulation of HCN2 and HCN4 in the ventricles during heart failure and ventricular hypertrophy has also been suggested to predispose affected individuals to certain arrhythmias [8]. However, despite recognizing the importance of I_f in normal and pathological conditions, how I_f exerts its physiological influence in the heart still remains poorly defined.

During cardiac pacing, changes in membrane potential are rapid and dynamic, preventing the establishment of an equilibrium condition. Therefore, traditional measurements of equilibrium variables including channel availability and midpoint activation may be inadequate to explain certain biological processes during the cardiac cycle. Recently, Clancy et al. reported that non-equilibrium gating of cardiac Na⁺-channels during action potential repolarization triggers the potentially lethal arrhythmias caused by certain long-QT syndrome SCN5A mutations [9]. When measured using standard electrophysiological protocols, the only apparent functional effect of the SCN5A disease mutation I1768V was a 2-fold increase in the rate of recovery from inactivation. By examining non-equilibrium gating, however, these investigators uncovered an abnormally large transient inward Na⁺-current during action-potential repolarization that promotes the formation of pro-arrhythmic early after-depolarizations [9].

Given the functional importance of HCN channels in normal cardiac automaticity, we sought to examine HCN gating under non-equilibrium conditions. We hypothesized that I_f behaves differently during different phases of the cardiac cycle. In other words, it is possible that HCN-encoded currents exhibit current–voltage (IV)-hysteresis during cardiac pacing. If present, IV-hysteresis may have important regulatory and physiological consequences. To test our hypothesis, the properties of HCN1, HCN2 and HCN4 channels were examined under non-equilibrium conditions. We found that HCN1 channels exhibit pronounced IV-hysteresis during simulated SA-nodal pacing that is dependent upon their inherent kinetic properties. In contrast, IV-hysteresis of HCN2 channels is much less pronounced, while HCN4 channels do not exhibit any hysteresis. We attribute the attenuation of IV-hysteresis for HCN2 and HCN4 channels to their markedly delayed onset of activation and relatively slow activation kinetics. These data are discussed in the context of our understanding of the physiological roles of I_f in cardiac pacing and potential implications for genetically fine-tuning cell- and gene-based pacemakers. A preliminary report of our findings has appeared elsewhere [10].

	2. Methods

Top Abstract 1. Introduction 2. Methods 3. Results 4. Discussion References

 
2.1. Molecular biology and heterologous expression in Xenopus laevis oocytes and Cos7 cells
mHCN1 and mHCN2 were subcloned into the pGH expression vector [11]. cRNA was transcribed from linearized DNA using T7 RNA polymerase (Promega, Madison, WI). rbHCN4 (Dr. Harunori Ohmori, Kyoto University) was subcloned into the pCI mammalian expression vector [4]. Channels were heterologously expressed and studied in Xenopus oocytes as described in our previous publications [12,13].

mHCN1 was subcloned into the AdCGI-HCN1 adenoviral shuttle vector [14]. Recombinant mHCN1 adenovirus was made by Cre-recombination of AdCGI-HCN1 DNA with 5 viral DNA in the Cre4 cell line. Mammalian Cos7 cells were cultured in 6-well dishes containing 2 ml DMEM with 10% FBS and 1% penicillin–streptomycin. Cells were transfected with 2 µl/well of AdCGI-HCN1 adenovirus (5 x 10⁶ plaque-forming units) 2 days before recording.

2.2. Electrophysiology
Two-electrode voltage-clamp recordings using micro-injected Xenopus oocytes were performed at 23–25 °C as described previously [13]. The standard recording bath solution contained (in mM): 96 KCl, 2 NaCl, 2 MgCl₂ and 10 HEPES (pH 7.6). A more physiological, high-Na⁺ bath solution containing (in mM) 25 KCl, 120 NaCl, 2 MgCl₂ and 10 HEPES (pH 7.6) was also used as indicated. Whole-cell currents were evoked by one of two protocols using the P-Clamp software package (Axon Instruments, Foster City, CA): Protocol 1: Test-voltages ranging from –140 to 0 mV with various test-pulse durations as indicated were applied to channels from a holding potential of +70, –75 or –140 mV. For HCN1 and HCN2, the holding potential was long enough to achieve equilibrium (1000 ms for HCN1 and 4000 ms for HCN2), as determined by the plateau of whole-cell currents. For HCN4, a near-equilibrium state was achieved by holding for 5000 ms. Longer durations at some voltages resulted in damage to the oocytes after multiple sweeps. After each test-voltage, tail currents were measured after pulsing to +50 mV. Protocol 2: A train of action potentials generated by a computational model (described below) was used as command input in the voltage-clamp protocol. The protocol was shifted negatively by 40 mV for HCN2 and HCN4 to adjust for their more depolarized activation in mammalian cells [15].

Whole-cell patch clamp recordings on Cos7 cells were performed at room temperature using an EPC-9 patch clamp amplifier (HEKA electronic, Heidelberg, Germany) [16]. A modified version of Protocol 1 (above) was implemented using holding voltages of –30, –60 and –90 mV. Pipette electrodes with resistances of 1–3 M were backfilled with the internal pipette solution containing (in mM): 110 K⁺-aspartate, 20 KCl, 1 MgCl₂, 5 Na⁺-phosphocreatine, 10 EGTA, 10 HEPES, 0.1 GTP, 5 Mg-ATP, pH adjusted to 7.2 with KOH. The bath solution contained (in mM): 140 NaCl, 5 KCl, 2 CaCl₂, 1 MgCl₂, 10 HEPES, 10 glucose, pH adjusted to 7.4 with NaOH.

For activation curves, tail current data from Protocol 1 was plotted as a function of the test-voltage (V_t) and fitted to the Boltzman function using the Marquardt-Levenberg algorithm (y=1/[1+exp[(V_t–V_1/2)/k]]) in a non-linear least-squares procedure to obtain the mid-point voltage (V_1/2) and slope factor (k). All data reported are mean ± S.E.M. Statistical significance was determined using one-way ANOVA and Tukey's post-hoc test at the 5% level.

2.3. In silico analysis of the effect of I_f on heart rate
For our computational experiments, a model of HCN1 current (Eqs. (1)–(4)) was inserted into the Luo-Rudy (LRd) guinea pig ventricular cell model [17] with or without 100% suppression of the inward rectifier K⁺-current I_K1. I_K1 suppression has been shown to induce pacemaker activity in silico [18].

Experimentally determined data from our laboratory [12] were fit to estimate the voltage dependence of equilibrium open-probability (y) and the time constants for activation/deactivation (_y, ms). Equations for the Na⁺ and K⁺ components of maximum HCN1 conductance, g_f,Na and g_f,K, were taken from the pacemaker cell model of Kurata et al. [19]. The total maximum HCN1 conductance, g_f, was increased 5-fold to compensate for an almost 17-fold increase in the slow delayed rectifying K⁺ (I_Ks) current conductance (as compared to Kurata model). The estimate of 1.785 mS/µF is well within the range of maximum conductances in the LRd model (0.03 to 16.0 mS/µF). E_Na and E_K (mV) are the Na⁺ and K⁺ Nernst potentials, respectively. I_f is the total HCN current (pA/pF) and v is the transmembrane voltage (mV).

	3. Results

Top Abstract 1. Introduction 2. Methods 3. Results 4. Discussion References

 
3.1. HCN1 activation shifts under non-equilibrium conditions
As our first step, we compared HCN1 activation under non-equilibrium and equilibrium conditions. Fig. 1A demonstrates that activation of HCN1 channels expressed in Xenopus oocytes was strongly dependent on the holding potential (V_H) when short 100-ms test-pulses were used. The activation midpoint (V_1/2) shifted significantly from –118 ± 1.3 mV (n = 6) to –78 ± 3.3 mV (n = 6) to –19 ± 3.8 mV (n = 6) when V_H was changed from +70 to –75 mV to –140 mV (p<0.05). Similar V_H-dependent shifts of V_1/2 were also observed when the same protocols were employed in a more physiologically relevant high-Na⁺ bath solution (Fig. 1B), except that activation midpoints were positively shifted (V_{1/2,+70 mV}=–80 ± 4.8 mV (n = 4), V_{1/2,–140 mV}=+5 ± 2.0 mV (n = 3); p<0.05), consistent with our previous finding that lowering extracellular K⁺ causes depolarizing activation shifts [12]. In contrast, activation curves of HCN1 channels obtained using 3000-ms test-pulses were insensitive to V_H (V_{1/2,+70 mV}=–73 ± 1.1 mV (n = 6), V_{1/2,–75 mV}=–74 ± 1.4 mV (n = 6) and V_{1/2,–140 mV}=–71 ± 0.9 mV (n = 6); Fig. 1C; p>0.05).

View larger version (32K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 1 HCN1 activation in Xenopus oocytes depends on holding potential under non-equilibrium conditions. (A) Representative tail currents (left) from HCN1 channels after short (100-ms) test-pulses generated using the protocol provided in the inset (a=–140 mV, b=–80 mV, c=–40 mV, d=–20 mV, e=0 mV). The activation curve (right) shifts positively at hyperpolarized holding potentials (V1/2,+70 mV=–118 ± 1.3 mV (n = 6), V1/2,–75 mV=–78 ± 3.3 mV (n = 6), V1/2,–140 mV=–19 ± 3.8 mV (n = 6); p<0.05). The dotted line represents the equilibrium activation curve. (B) Substituting extracellular potassium with sodium shifts both activation curves positively but similar dependence on VH is still observed (V1/2,+70 mV=–80 ± 4.8 mV (n = 4), V1/2,–140 mV=+5 ± 2.0 mV (n = 3); p<0.05). (C) Normalized activation curves after long test-pulses (3000-ms) do not display the same holding potential-dependent shifts of V1/2 observed in (A) (V1/2,+70 mV=–73 ± 1.1 mV (n = 6), V1/2,–75 mV=–74 ± 1.4 mV (n = 6) and V1/2,–140 mV=–71 ± 0.9 mV (n = 6); p>0.05).

 
As shown in Fig. 2, qualitatively similar results were also observed with HCN1 channels expressed in mammalian Cos7 cells, indicating that the V_H dependence of activation is an intrinsic property of HCN1 channels. When short 100-ms test-pulses were used (Fig. 2A), V_1/2 shifted significantly from –91 ± 1.7 mV (n = 7) to –59 ± 7.2 mV (n = 5) to +6 ± 5.5 mV (n = 5) when V_H was changed from –30 to –60 mV to –90 mV (p<0.05). Activation curves obtained using 3000-ms test-pulses did not differ from each other significantly (V_{1/2,–30 mV}=–60 ± 1.0 mV (n = 16), V_{1/2,–60 mV}=–58 ± 4.1 mV (n = 3) and V_{1/2,–90 mV}=–60 ± 3.8 mV (n = 6); Fig. 2B; p>0.05).

View larger version (17K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 2 HCN1 activation in mammalian Cos7 cells also depends on holding potential. (A) The activation midpoint (V1/2) shifts positively as the holding potential is hyperpolarized (V1/2,–30 mV=–91 ± 1.7 mV (n = 7), V1/2,–60 mV=–59 ± 7.2 mV (n = 5), V1/2,–90 mV=+6 ± 5.5 mV (n = 5)). The dotted line represents the equilibrium activation curve. (B) Activation curves obtained using 3000-ms test-pulses did not differ from each other significantly (V1/2,–30 mV=–60 ± 1.0 mV (n = 16), V1/2,–60 mV=–58 ± 4.1 mV (n = 3) and V1/2,–90 mV=–60 ± 3.8 mV (n = 6)).

 
To explore the nature of the strong dependence of V_1/2 on V_H for short but not long test-pulses, we next examined the transition of V_1/2 during intermediate test-pulse durations (100 to 800 ms). Fig. 3 shows that, when V_H=+70 mV, V_1/2 increased exponentially with a transition time constant () of 234 ± 18 ms (n = 6) as test-pulse duration was gradually prolonged. The activation midpoint after reaching equilibrium (V_1/2,) was –82 ± 2 mV. For V_H=–140 mV, V_1/2 decreased in a similar exponential manner: and V_1/2, were 182 ± 37 ms and –79 ± 5 mV (n = 6), respectively.

View larger version (15K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 3 HCN1 activation midpoint changes exponentially as a function of test-pulse duration. As the test-pulse duration is increased from 100 ms to 800 ms, the V1/2 for VH=+70 mV increases exponentially ( = 234 ± 18 ms, V1/2,=–82 ± 2 mV) and the V1/2 for VH=–140 mV decreases exponentially ( = 182 ± 37 ms, V1/2,=–79 ± 5 mV).

 
3.2. HCN1 channels exhibit marked IV-hysteresis in response to stimulation by a train of action potentials
Since HCN1 activation depended strongly on V_H for short (non-equilibrium) but not long (equilibrium) test-pulses, we postulated that HCN1 current properties in response to a train of action potentials would vary in different phases of the cardiac cycle depending upon whether the cell was depolarizing or hyperpolarizing. Fig. 4A shows the IV-trajectory of HCN1 channels (n = 6) obtained by applying the SA nodal-like voltage protocol given in the inset. Consistent with our hypothesis, HCN1 channels displayed pronounced IV-hysteresis. Interestingly, the so-called hyperpolarization-activated HCN1 currents increased despite membrane depolarization from the maximum diastolic potential (MDP–67 mV) to the maximum current potential (MCP–61 mV; a to b as marked in Fig. 4). This current peak was followed by a gradual decrease in amplitude during the remainder of diastolic depolarization. Thus, HCN1 current was largest during diastolic depolarization and the action potential upstroke, but was nearly inactive at the same voltages during repolarization.

View larger version (12K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 4 HCN channels exhibit varying degrees of IV-hysteresis in response to a train of action potentials. Hysteresis phase-plot of transmembrane voltage versus (A) HCN1 current, (B) HCN2 current and (C) HCN4 current (n = 6 for all three) generated using the voltage protocol provided in the inset.

 
3.3. IV-hysteresis and pacing rate have a similar dependence on gating kinetics
To explore the potential role of HCN channel IV-hysteresis in pacing and to test whether it can be exploited to modulate pacing rate, we performed an in silico analysis using the LRd model [17] with and without a component of normal or engineered HCN1 current (Table 1 and Fig. 5). Spontaneous pacing at a rate of 171 bpm with a MDP of –59 mV was achieved by completely eliminating the inward rectifier potassium current, I_K1 (Fig. 5B) [18]. Introducing WT HCN1 as a current component (g_f=1.785 mS/µF; Fig. 5C) further increased the pacing rate to 178 bpm without significantly altering MDP. The modified model also qualitatively reproduced the experimental IV-hysteresis relationship (Fig. 6A).

View this table: [in this window] [in a new window]   Table 1 Summary of pacing rate (PR), maximum diastolic potential (MDP) and hysteresis loop-depth for simulations of induced pacemakers with varying fractions of IK1 current (fIK1) and HCN1 activation time constant (fy)

 

View larger version (25K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 5 In silico analysis of the effect on pacing rate of modifying IK1 and If. (A) No spontaneous activity at baseline IK1 levels. (B) Rate=171 bpm with complete IK1 inhibition. (C) Introducing If without IK1 increases pacing rate to 178 bpm. Reducing y by 50% (D) and 90% (E) decreases pacing rate to 165 and 157 bpm, respectively. Pacing could not be induced by introducing If alone unless: (F) If conductance 0.935 mS/µF when V1/2=–70 mV (50% of baseline conductance) or (G) V1/2 –62 mV (when gh=0.357 mS/µF of native SA nodal If). (H) Introducing HCN1 current into the baseline LRd model (i.e. 100% IK1) results in spontaneous pacing at 357 bpm.

 

View larger version (32K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 6 Computer simulation of HCN1 IV-hysteresis. (A) The HCN1 IV-hysteresis obtained from the modified LRd model (circles) is qualitatively similar to our experimental observations. (B) Overall, the relationship between loop-depth and y is "peaked" with a maximum near the baseline y. The relationship between pacing rate and y is also peaked. (C) As a result, loop-depth and pacing rate are positively correlated (r2=0.89). (D) HCN1 current was uncoupled from transmembrane voltage and y was increased 20-fold. The action-potential stimulation protocols provided in the inset were applied to the modified HCN1 current. Loop-depth increases as pacing rate decreases.

 
Both our experimental (Fig. 4)) and computational (Fig. 6) data demonstrate that HCN1 current magnitude depends strongly on the preceding transmembrane potential history. This response is likely due to the very slow activation of HCN1 channels on the time-scale of cardiac pacing. Reducing (i.e. accelerating) the HCN1 gating time constant, _y, to 50% (Fig. 5D) and 10% (Fig. 5E) of the original wild-type level (without changing other channel properties, such as conductance and V_1/2) decreased the pacing rate to 165 and 157 bpm, respectively. These changes were associated with depolarization of MDP by 1 and 4 mV (Table 1). Conversely, increasing _y by 50% increased pacing rate to 180 bpm without changing the MDP (Table 1). However, further increase of _y actually decreased pacing rate, resulting in a "peaked" relationship between _y and pacing rate (open circles, Fig. 6B). All related parameters are summarized in Table 1 and Fig. 5.

To quantify changes in IV-hysteresis as a function of _y, we studied hysteresis loop-depth, defined as the percent increase in current magnitude between the MDP and the MCP. Loop-depth corresponds to the percent increase in current from point a to point b in Fig. 4 and is a representation of the fractional increase of HCN1 current during membrane depolarization from the MDP. The IV phase-plots for _y (baseline), 0.1 x _y and 10 x _y are shown in Fig. 6A. Loop-depth was decreased for both 0.1 x _y and 10 x _y. Similar to the relationship between pacing rate and _y, that between loop-depth and _y also displayed a bell-shaped distribution (solid squares, Fig. 6B). Since both loop-depth and pacing rate have a "peaked" relationship with _y, the relationship between loop-depth and pacing rate for all values of _y is linearly correlated (Fig. 6C; r²=0.89).

3.4. HCN2 and HCN4 have markedly attenuated IV-hysteresis
Based on our model of HCN1 IV-hysteresis and the bell-shaped relationship between loop-depth and _y (Fig. 6B), HCN4, which activates 10–25 times slower than HCN1 (i.e. larger _y), should exhibit significantly less hysteresis (i.e. smaller loop-depth). Indeed, our electrophysiological experiments demonstrate that the response of HCN4 to nodal pacing displayed virtually no IV-hysteresis (Fig. 4C). As anticipated from these results, the IV-hysteresis of HCN2 channels, which activate 4–5 times slower than HCN1 but still faster than HCN4, was also attenuated; although the phenotype was intermediate between HCN1 and HCN4 (Fig. 4B).

Based on the above results, we further postulated that the attenuated loop-depths observed for HCN2 and HCN4 channels resulted from their larger time constants relative to the pacing rate. To test this notion, we performed another in silico experiment by uncoupling I_f from the total ionic current and increasing its _y by 20-fold (Fig. 6D). Uncoupling enabled us to observe the effect of transmembrane voltage on IV-hysteresis without the confounding effect of HCN current feed-back to transmembrane voltage. As the pacing rate was decreased from 125 to 61 to 36 bpm, loop-depth increased from 22% to 65% to 96% (Fig. 6D). These pacing rates correspond to excitation periods of approximately 626, 984 and 1667 ms (18%, 37% and 63% of 20 x _y (2663 ms) at –60 mV). Thus, consistent with our postulation, when the period of excitation is increased (i.e. fewer bpm), slowly activating HCN channel isoforms are able to generate enough current to increase the loop-depth.

3.5. HCN2 and HCN4 but not HCN1 channels exhibit marked activation delays and instantaneous currents
Unlike HCN1, both HCN2 and HCN4 have marked activation delays associated with significant instantaneous current components (I_inst) [20,21]. To explore whether these differences correlate with the different IV-hysteresis phenotypes among HCN isoforms, we examined the first 500 ms of HCN1, HCN2 and HCN4 currents in response to hyperpolarizing voltage steps. During the first 500 ms, HCN1 activation is strongly time- and voltage-dependent (Fig. 7D). In contrast, both HCN2 (Fig. 7E) and HCN4 (Fig. 7F) display an early activation "shoulder" preceding their time-dependent components, giving these isoforms their characteristic sigmoidal activation profiles. The magnitude of I_inst during the HCN2 and HCN4 activation shoulders changes almost proportionally to voltage (from 0 to –60 mV; Fig. 7E and F).

View larger version (25K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 7 HCN1, HCN2 and HCN4 all have Iinst but vary markedly in activation delay. The top three panels are 3000-ms whole-cell currents of HCN1 (A), HCN2 (B) and HCN4(C), after test-pulses (0 to –80 mV) from a holding potential of –30 mV. The lower three panels are expanded views of the first 500 ms of HCN1 (D), HCN2 (E) and HCN4 (F) whole-cell currents, indicated in the top panels by dashed boxes. Iinst is present in all three isoforms; however, the duration of Iinst (e.g. dinst=activation delay at –80 mV) is isoform-dependent. Longer activation delays are associated with slower activation kinetics (HCN4>HCN2>HCN1).

 
For HCN1, the duration of I_inst at –80 mV (i.e. d_{inst,–80 mV}, the delayed onset of activation of the time-dependent component, defined as the time required for a 5% increase in current magnitude from the start of the test-pulse) is insignificant (d_{inst,–80 mV}=4.3 ± 1 ms, n = 3, p<0.05; Fig. 7D). However, HCN4 has a much longer I_inst (d_{inst,–80 mV}=263 ± 11 ms, n = 3, p<0.05; Fig. 7F). The delay in HCN2 activation (d_{inst,–80 mV}=98 ± 9 ms, n = 3, p<0.05; Fig. 7E) is much longer than that of HCN1 but much shorter than that of HCN4, in accordance with the intermediate IV-hysteresis phenotype of HCN2.

Thus, on the time-scale of diastolic depolarization (200–300 ms in humans and <200 ms in rodents), I_inst provides HCN4 and HCN2 currents with the paradoxical ability to respond more rapidly to changes in membrane potential than HCN1 currents (albeit with lower absolute current magnitude). Furthermore, the long activation delays (d_inst) of HCN4 and HCN2 channels enables I_inst to dominate over the time-dependent current component on the time-scale of cardiac pacing. According to our computational model, reducing the frequency of action potential excitation (by decreasing the slope of diastolic depolarization) allows the time-dependent component of slowly activating HCN channels to develop and contribute to IV-hysteresis (Fig. 5D). Therefore, the differences in hysteresis among HCN 1, 2 and 4 may result from a combination of both the slow kinetics of HCN2 and HCN4 channels relative to the pacing rate, in addition to their activation delays and instantaneous currents.

3.6. Effects of HCN channels on pacing depend on parameters other than their own biophysical properties
To further investigate how I_f can functionally influence pacing, we examined the effects of altering other cellular and channel parameters on pacing rate. In the presence of 100% native I_K1, pacing could not be induced by introducing HCN1 alone unless at least one of the following conditions was met: (1) HCN1 conductance 0.935 mS/µF (when V_1/2=–70 mV; Fig. 5F) and (2) V_1/2 –62 mV (when HCN1 conductance=0.357 mS/µF of native SA nodal I_f; Fig. 5G). When HCN1 was just below either of the threshold levels needed to induce pacing, the resting membrane potential (RMP) became significantly depolarized (–64 mV) even in the presence of the RMP stabilizer I_K1 (RMP=–85 mV in the baseline LRd model).

Interestingly, pacing was twice as fast when HCN1 was incorporated in the presence of I_K1 (357 bpm; Table 1, Fig. 5H) than in the absence of I_K1 (178 bpm; Table 1, Fig. 5C), probably as a result of increased HCN channel activation at the more hyperpolarized MDP (–80 mV with I_K1 vs. –59 mV without I_K1). Thus, the strategies of genetic suppression of I_K1 [22] and over-expression of I_f [23] for generating induced pacemaker activities are not necessarily synergistic. A balance between these two "opponents" may be needed to fine-tune pacing.

	4. Discussion

Top Abstract 1. Introduction 2. Methods 3. Results 4. Discussion References

 
4.1. Mechanism and physiological implications of HCN1 IV-hysteresis
Based on their inherently slow kinetics, we made the straight-forward prediction that HCN1 channels exhibit IV-hysteresis. This prediction was experimentally tested to explore the physiological role of I_f in cardiac pacemaking, and how this cardiac membrane current can be effectively engineered to modulate endogenous or induced pacing [22,23]. At 3000 ms, HCN1 currents have reached equilibrium and activation is independent of holding potential (V_H). However, such a long test-pulse is irrelevant during dynamic cardiac pacing (the cycle is only 250 ms in guinea pigs and 1000 ms in humans). In response to shorter test-pulses (100–400 ms), non-equilibrium HCN1 activation is strongly dependent on V_H. Indeed, we observed IV-hysteresis when HCN1 channels were stimulated with a train of action potentials, a fundamentally non-equilibrium situation. These observations could be reproduced using computational modeling and are similar to those described by Bruening-Wright and Larsson [24].

IV-hysteresis of HCN1 channels depends on the magnitude of the activation time constant (_y) relative to the pacing period. Indeed, HCN1 channels exhibit IV-hysteresis because _y is just slow enough to confer HCN1 with a sort of voltage "memory". As demonstrated by our analyses, both very rapid and very slow kinetics would reduce IV-hysteresis. Consider the example of a population of channels subjected to periodic stimulation between a voltage that opens all channels and a voltage that closes all channels. If the stimulation period is T and _y=0.05T, an equilibrium distribution of open and closed channels will be reached in 0.25T (or 5 time constants). Thus, by the time of the next pulse (0.75 T later), all memory of the prior pulse will have been lost, preventing IV-hysteresis. On the other hand, if _y is increased to 20T, more than 80 times the stimulation period is needed to reach equilibrium. In this case, only 5% (1–e^–T/20T) of the closed channels will open. The remaining will be effectively locked in the closed state, again preventing any significant IV-hysteresis. An intermediate time constant of _y=2T would open about 40% (1–e^–T/2T) of closed channels. This would be enough to produce a measurable change in current magnitude and, with 60% of channels still affected by the previous voltage history, exhibit IV-hysteresis.

The rate-dependent non-equilibrium behavior of HCN1 channels is somewhat similar to frequency- or use-dependent channel blockade, which is derived from the drug unbinding time constant being longer than the stimulus interval, thereby imparting a sort of memory [25]. The IV-hysteresis of HCN1 channels is also analogous to that of many enzymes [26]. For instance, in the heart, phosphofructokinase (PFK) enzymatic activity exhibits a pH-dependent hysteresis that has been postulated to play a role in the metabolic regulation of the myocardium during ischemia [27]. The pH-hysteresis of PFK is attributed to slow, pH-dependent conformational changes that enable the enzyme to "memorize" the recent pH history of the environment. Such "memory" is indeed a characteristic of any system under non-equilibrium conditions. On the time-scale of cardiac pacing, HCN1 channels "memorize" the recent transmembrane voltage history because their very slow kinetics forces them to operate under non-equilibrium conditions during the cardiac cycle. Consequently, HCN1 activation is greatly influenced by the preceding transmembrane voltages. This memory explains the increase in HCN1 current amplitude despite depolarization from maximum diastolic potential (MDP) to maximum current potential (MCP). In other words, the increase in current from MDP to MCP (i.e. loop-depth) does not occur in response to membrane depolarization, but, rather, as a delayed response to membrane hyperpolarization during the previous action potential repolarization. Thus, the slow activation of HCN1 channels does not hinder their ability to drive pacing, but, in fact, is central to modulation of firing rate. Our computational model supports this notion since accelerating HCN1 activation (by decreasing _y) decreases the pacing rate.

In addition to its proposed role as a direct contributor to the pacemaker current, I_f, whose density is highest at the periphery of the SA node, may also serve to protect nodal cells from any over-hyperpolarizing influence from the bulk atrial tissue during action potential repolarization. In essence, I_f may stabilize diastolic membrane voltage so that other pacemaker currents can operate within the proper voltage range [27]. In support of this notion, our in silico studies indicate that, when HCN1 is incorporated into the model, MDP and RMP, even in the presence of I_K1, both become more depolarized (whether or not pacing has been induced). Rosen and colleagues [28] were able to modulate pacing in vitro and in vivo by coupling cardiomyocytes to human mesenchymal stem cells (hMSCs) that express HCN2 (heterologously) and connexin-43 (endogenously). HCN2 expression promoted pacemaker activity by generating current sources (via gap junctions) that depolarized the RMP (or MDP) of neighboring quiescent (or oscillating) cardiomyocytes. Ludwig et al. also propose a protective role for HCN current based on their characterization of an HCN2 knock-out mouse that exhibits a highly variable RR-interval on EKG and an unstable MDP in isolated SA nodal cells [5]. Thus, as proposed previously [29] and further supported by our present results, I_f serves a dual role in cardiac pacing by both directly contributing to diastolic depolarization and by stabilizing the MDP at an appropriate voltage so that other currents involved in diastolic depolarization can function properly.

4.2. Physiological roles of the initial activation delay and instantaneous component of I_f
In the current study, we propose that the attenuated IV-hysteresis of HCN2 and HCN4 (compared to HCN1) is due to a combination of slow kinetics and pronounced activation delays with instantaneous currents. In voltage clamp experiments, both endogenous I_f and heterologously expressed HCN channels have been described as consisting of an initial instantaneous current (I_inst) followed by a time-dependent current [20,21]. Macri and Accili have proposed that I_inst may represent current flow through a "leaky" HCN channel closed state, similar to the large instantaneous currents of inwardly rectifying plant K⁺-channels [30]. These investigators also suggest that I_inst may contribute to the Na⁺-sensitive background current (I_b,Na). Other than these studies, however, little is known about the physiological significance of the I_inst and d_inst components [21]. In fact, these parameters are often ignored when HCN gating is studied. For instance, when fitting HCN currents with an exponential function, the initial delay during the activation shoulder is excluded. Since HCN4 is present in the SA node [3], HCN4 I_inst (and not the time-dependent current component) dominates on the time scale of cardiac pacing, and HCN4 (and not HCN1 [31] or HCN2 [5]) knock-out mice die in-utero from severe bradycardia and chronotropic incompetence; further study of HCN4 d_inst and I_inst is warranted.

4.3. "Fine-tuning" cardiac pacing by engineering HCN-encoded I_f
To date, at least three approaches have been taken to induce ectopic pacemaker activity: heterologous β₂-adrenergic receptor expression in pig atria [32], dominant-negative suppression of I_K1 in guinea pig ventricle [22] and over-expression of I_f [23]. Our present and recent results may provide additional pragmatic implications for inducing and fine-tuning pacemaker activity. For instance, the dominant-negative construct HCN1-AAA (i.e. GYG_349–351AAA [13]) could be used to suppress native I_f for tuning the frequency of pacing. Similarly, such recombinant HCN constructs as S3–S4 linker mutants [33], which exhibit a range of gating properties, may be useful for customizing the firing rate of native or engineered pacemakers [34,35,36].

In our computational model, MDP is dependent upon the relative expression levels of I_K1 and HCN1. I_K1 hyperpolarizes the MDP, resulting in a higher HCN1 channel open-probability (y) during early diastolic depolarization. A higher y when I_K1 is present, as compared to when I_K1 is inhibited, facilitates pacing by substantially increasing I_f and loop-depth (loop-depth=670% with I_K1 vs. 228% without I_K1; Table 1). Therefore, the strategies of I_K1 suppression and HCN over-expression to induce pacing may not be synergistic. Moreover, the modulatory effects of HCN1 on pacing are the function of a number of extrinsic cellular parameters (such as MDP and I_K1) as well as its own intrinsic properties (expression level, IV-hysteresis, equilibrium and kinetic gating properties, etc.). Careful modification of these parameters and properties is needed to fine-tune the firing frequency of gene- and cell-based pacemakers. Additionally, given the relatively long activation delays of HCN2 and HCN4, mutations that augment I_inst [30] may also promote ectopic pacing in vivo.

4.4. Conclusions
As a result of their intrinsically slow activation kinetics, HCN1 channels exhibit IV-hysteresis in response to stimulation by a train of action potentials. The degree of IV-hysteresis, when quantified by loop-depth, is strongly correlated with pacing rate. Given the bell-shaped relationship between activation time constant and pacing rate, it appears that the slow kinetics of HCN1 channels do not hinder their ability to modulate pacing. Slow kinetics confers upon HCN1 channels a voltage "memory", enabling their currents to increase in magnitude despite membrane depolarization during early diastole. This mechanism may play an important role in the modulation of cardiac and neuronal pacing.

For HCN2 and HCN4, slow activation kinetics and a long activation delay work in concert to attenuate IV-hysteresis. Thus, non-equilibrium methods which do not obscure the early instantaneous current (I_inst) may sometimes be more appropriate for studying HCN channel gating; especially in the context of inferring how these channels are involved in the regulation of non-equilibrium processes like cardiac pacing. Taken collectively, our present results provide an experimental and theoretical framework for fine-tuning gene- and cell-based pacemakers.

	Acknowledgements

 
This work was supported by a grant from the NIH (R01 HL-52768 and HL72857 to R.A.L.). R.A.L. received salary support from the Cardiac Arrhythmias Research and Education Foundation, Inc. during the tenure of this project.

	Notes

 
Time for primary review 14 days

	References

Top Abstract 1. Introduction 2. Methods 3. Results 4. Discussion References

 

1. Ludwig A., Zong X., Jeglitsch M., Hofmann F., Biel M. A family of hyperpolarization-activated mammalian cation channels. Nature (1998) 393:587–591.[CrossRef][Medline]
2. Shi W., Wymore R., Yu H., Wu J., Wymore R.T., Pan Z., et al. Distribution and prevalence of hyperpolarization-activated cation channel (HCN) mRNA expression in cardiac tissues. Circ Res (1999) 85:e1–e6.[Abstract/Free Full Text]
3. Ludwig A., Zong X., Stieber J., Hullin R., Hofmann F., Biel M. Two pacemaker channels from human heart with profoundly different activation kinetics. EMBO J (1999) 18:2323–2329.[CrossRef][ISI][Medline]
4. Ishii T.M., Takano M., Ohmori H. Determinants of activation kinetics in mammalian hyperpolarization-activated cation channels. J Physiol (Lond) (2001) 537:93–100.[Abstract/Free Full Text]
5. Ludwig A., Budde T., Stieber J., Moosmang S., Wahl C., Holthoff K., et al. Absence epilepsy and sinus dysrhythmia in mice lacking the pacemaker channel HCN2. EMBO J (2003) 22:216–224.[CrossRef][ISI][Medline]
6. Schulze-Bahr E., Neu A., Friederich P., Kaupp U., Breithardt G., Pongs O., et al. Pacemaker channel dysfunction in a patient with sinus node disease. J Clin Invest (2003) 111:1537–1545.[CrossRef][ISI][Medline]
7. Ueda K., Nakamura K., Hayashi T., Inagaki N., Takahashi M., Arimura T., et al. Functional characterization of a trafficking-defective HCN4 mutation, D553N, associated with cardiac arrhythmia. J Biol Chem (2004) 279:27194–27198.[Abstract/Free Full Text]
8. Hiramatsu M., Furukawa T., Sawanobori T., Hiraoka M. Ion channel remodeling in cardiac hypertrophy is prevented by blood pressure reduction without affecting heart weight increase in rats with abdominal aortic banding. J Cardiovasc Pharmacol (2002) 39:866–874.[CrossRef][ISI][Medline]
9. Clancy C.E., Tateyama M., Liu H., Wehrens X.H., Kass R.S. Non-equilibrium gating in cardiac Na+ channels: an original mechanism of arrhythmia. Circulation (2003) 107:2233–2237.[Abstract/Free Full Text]
10. Azene E.M., Li R.A. Marked hysteresis in the current-voltage relationship of HCN-encoded pacemaker channels: insights into the physiological role of I_f in cardiac pacing. Suppl Circ (2003) 108:IV-88.
11. Santoro B., Liu D.T., Yao H., Bartsch D., Kandel E.R., Siegelbaum S.A., et al. Identification of a gene encoding a hyperpolarization-activated pacemaker channel of brain. Cell (1998) 93:717–729.[CrossRef][ISI][Medline]
12. Azene E., Xue T., Li R.A. Molecular basis of the effects of potassium of heterologously-expressed pacemaker (HCN) channels. J Physiol (2003) 547(2):349–356.[Abstract/Free Full Text]
13. Xue T., Marban E., Li R.A. Dominant-negative suppression of HCN1- and HCN2-encoded pacemaker currents by an engineered HCN1 construct: insights into structure-function relationships and multimerization. Circ Res (2002) 90:1267–1273.[Abstract/Free Full Text]
14. Hardy S., Kitamura M., Harris-Stansil T., Dai Y., Phipps M.L. Construction of adenovirus vectors through Cre-lox recombination. J Virol (1997) 71:1842–1849.[Abstract]
15. Robinson R.B., Siegelbaum S.A. Hyperpolarization-activated cation currents: from molecules to physiological functions. Annu Rev Physiol (2003) 65:453–480.[CrossRef][ISI][Medline]
16. Hamill O.P., Marty A., Neher E., Sakmann B., Sigworth F.J. Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches. Pflugers Arch (1981) 391:85–100.[CrossRef][ISI][Medline]
17. Luo C., Rudy Y. A dynamic model of the cardiac ventricular action potential: I. Simulations of ionic currents and concentration changes. Circ Res (1994) 74:1071–1096.[Abstract/Free Full Text]
18. Silva J., Rudy Y. Mechanism of pacemaking in I(K1)-downregulated myocytes. Circ Res (2003) 92:261–263.[Abstract/Free Full Text]
19. Kurata Y., Hisatome I., Imanishi S., Shibamoto T. Dynamical description of sinoatrial node pacemaking: improved mathematical model for primary pacemaker cell. Am J Physiol Heart, Circ Physiol (2003) 283:H2074–H2101.[ISI]
20. Ishii T.M., Takano M., Xie L.H., Noma A., Ohmori H. Molecular characterization of the hyperpolarization-activated cation channel in rabbit heart sinoatrial node. J Biol Chem (1999) 274:12835–12839.[Abstract/Free Full Text]
21. Proenza C., Angoli D., Agranovich E., Macri V., Accili E.A. Pacemaker channels produce an instantaneous current. J Biol Chem (2002) 277:5101–5109.[Abstract/Free Full Text]
22. Miake J., Marban E., Nuss H.B. Gene therapy: biological pacemaker created by gene transfer. Nature (2002) 419:132–133.[Medline]
23. Qu J., Barbuti A., Protas L., Santoro B., Cohen I.S., Robinson R.B. HCN2 overexpression in newborn and adult ventricular myocytes: distinct effects on gating and excitability. Circ Res (2001) 89:E8–E14.[CrossRef][ISI][Medline]
24. Bruening-Wright A., Larsson H. Investigating hysteresis in HCN channel activation. Suppl Biophys J (2004) 86:282a.
25. Johnson E.A., Mc K.M. The differential effect of quinidine and pyrilamine on the myocardial action potential at various rates of stimulation. J Pharmacol Exp Ther (1957) 120:460–468.[Abstract/Free Full Text]
26. Aon M.A., Cortassa S., Hervagault J.F., Thomas D. pH-induced bistable dynamic behaviour in the reaction catalysed by glucose-6-phosphate dehydrogenase and conformational hysteresis of the enzyme. Biochem J (1989) 262:795–800.[ISI][Medline]
27. Boyett M.R., Honjo H., Kodama I. The sinoatrial node, a heterogeneous pacemaker structure. Cardiovasc Res (2000) 47:658–687.[Abstract/Free Full Text]
28. Potapova I., Plotnikov A., Lu Z., Danilo P. Jr, Valiunas V., Qu J., et al. Human mesenchymal stem cells as a gene delivery system to create cardiac pacemakers. Circ Res (2004) 94:952–959.[Abstract/Free Full Text]
29. Stieber J., Hofmann F., Ludwig A. Pacemaker channels and sinus node arrhythmia. Trends Cardiovasc Med (2004) 14:23–28.[CrossRef][ISI][Medline]
30. Macri V., Accili E.A. Structural elements of instantaneous and slow gating in hyperpolarization-activated cyclic nucleotide-gated channels. J Biol Chem (2004) 279:16832–16846.[Abstract/Free Full Text]
31. Nolan M.F., Malleret G., Lee K.H., Gibbs E., Dudman J.T., Santoro B., et al. The hyperpolarization-activated HCN1 channel is important for motor learning and neuronal integration by cerebellar Purkinje cells. Cell (2003) 115:551–564.[CrossRef][ISI][Medline]
32. Edelberg J.M., Huang D.T., Josephson M.E., Rosenberg R.D. Molecular enhancement of porcine cardiac chronotropy. Heart (2001) 86:559–562.[Abstract/Free Full Text]
33. Tsang S.Y., Lesso H., Li R.A. Dissecting the structural and functional roles of the S3–S4 linker of pacemaker (hyperpolarization-activated cyclic nucleotide-modulated) channels by systematic length alterations. J Biol Chem (2004) 279:43752–43759.[Abstract/Free Full Text]
34. Rosen M.R., Brink P.R., Cohen I.S., Robinson R.B. Genes, stem cells and biological pacemakers. Cardiovasc Res (2004 Oct 1) 64(1):12–23.[CrossRef]
35. Kehat I., Khimovich L., Caspi O., Gepstein A., Shofti R., Arbel G., et al. Electrochemical integration of cardiomyocytes derived from human embryonic stem cells. Nat Biotechnol (2004 Oct) 22(10):1282–1289.[CrossRef][ISI][Medline]
36. Xue T., Cho H.C., Akar F.G., Tsang S.Y., Jones S.P., Marban E., et al. Functional integration of electrically active cardiac derivatives from genetically engineered human embryonic stem cells with quiescent recipient ventricular cardiomyocytes: insights into the development of cell-based pacemakers. Circulation (2005 Jan 4) 111(1):11–20. Epub 2004 Dec 20.[CrossRef]

CiteULike    Connotea    Del.icio.us    What's this?

This article has been cited by other articles:

				R. A. Rose, M. G. Kabir, and P. H. Backx Altered Heart Rate and Sinoatrial Node Function in Mice Lacking the cAMP Regulator Phosphoinositide 3-Kinase-{gamma} Circ. Res., December 7, 2007; 101(12): 1274 - 1282. [Abstract] [Full Text] [PDF]

				A. K. Lyashchenko, K. J. Redd, J. Yang, and G. R. Tibbs Propofol inhibits HCN1 pacemaker channels by selective association with the closed states of the membrane embedded channel core J. Physiol., August 15, 2007; 583(1): 37 - 56. [Abstract] [Full Text] [PDF]

				T. Xue, C.-W. Siu, D. K. Lieu, C.-P. Lau, H.-F. Tse, and R. A. Li Mechanistic Role of If Revealed by Induction of Ventricular Automaticity by Somatic Gene Transfer of Gating-Engineered Pacemaker (HCN) Channels Circulation, April 10, 2007; 115(14): 1839 - 1850. [Abstract] [Full Text] [PDF]

				W. Lin, U. Laitko, P. F. Juranka, and C. E. Morris Dual Stretch Responses of mHCN2 Pacemaker Channels: Accelerated Activation, Accelerated Deactivation Biophys. J., March 1, 2007; 92(5): 1559 - 1572. [Abstract] [Full Text] [PDF]

				F. Elinder, R. Mannikko, S. Pandey, and H. P. Larsson Mode shifts in the voltage gating of the mouse and human HCN2 and HCN4 channels J. Physiol., September 1, 2006; 575(2): 417 - 431. [Abstract] [Full Text] [PDF]

				R. B. Robinson, M. R. Rosen, P. R. Brink, I. S. Cohen, T. Xue, and R. A. Li Letter Regarding the Article by Xue et al, "Functional Integration of Electrically Active Cardiac Derivatives From Genetically Engineered Human Embryonic Stem Cells With Quiescent Recipient Ventricular Cardiomyocytes" * Response Circulation, August 9, 2005; 112(6): e82 - e83. [Full Text] [PDF]

This Article Abstract FREE Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Request Permissions Google Scholar A