OUP user menu

Pacemaker activity of the human sinoatrial node: effects of HCN4 mutations on the hyperpolarization-activated current

Arie O. Verkerk, Ronald Wilders
DOI: http://dx.doi.org/10.1093/europace/eut348 384-395 First published online: 25 February 2014

Abstract

The hyperpolarization-activated ‘funny’ current, If, plays an important modulating role in the pacemaker activity of the human sinoatrial node (SAN). If is carried by hyperpolarization-activated cyclic nucleotide-gated (HCN) channels, which are tetramers built of four HCN subunits. In human SAN, HCN4 is the most abundant of the four isoforms of the HCN family. Since 2003, several loss-of-function mutations in the HCN4 gene, which encodes the HCN4 protein, or in the KCNE2 gene, which encodes the MiRP1 accessory β-subunit, have been associated with sinus node dysfunction. Voltage-clamp experiments on HCN4 channels expressed in COS-7 cells, Xenopus oocytes, or HEK-293 cells have revealed changes in the expression and kinetics of mutant channels, but the extent to which these changes would affect If flowing during a human SAN action potential is unresolved. Here, we review the changes in expression and kinetics of HCN4 mutant channels and provide an overview of their effects on If during the time course of a human SAN action potential, both under resting conditions and upon adrenergic stimulation. These effects are assessed in simulated action potential clamp experiments, with action potentials recorded from isolated human SAN pacemaker cells as command potential and kinetics of If based on voltage-clamp data from these cells. Results from in vitro and in silico experiments show several inconsistencies with clinical observations, pointing to challenges for future research.

  • Sinoatrial node
  • Pacemaker activity
  • Hyperpolarization-activated current
  • HCN4
  • MiRP1
  • Computer simulation

Congenital sick sinus syndrome

The ‘sick sinus syndrome’ has been defined as the ‘intrinsic inadequacy of the sinoatrial node (SAN) to perform its pacemaking function due to a disorder of automaticity and/or inability to transmit its impulse to the rest of the atrium’.1 The SAN dysfunction is characterized by an atrial rate that is inappropriate for physiological requirements2 and presents itself with a combination of symptoms such as dizziness, fatigue, and syncope. Sick sinus syndrome is a common cardiac disorder accounting for ≈50% of pacemaker implantations.3 The SAN dysfunction may have several electrocardiographic features including sinus bradycardia, sinus arrest, SAN exit block, or alternating periods of bradycardia and tachyarrhythmias.4

In the literature of the 1960s and 1970s, several reports are available on familial forms of sick sinus syndrome.59 By the end of the 1990s, as reviewed by Lei et al.,10 it became clear that sinus bradycardia could be caused by autosomal dominant mutations in SCN5A.1113 This gene encodes the NaV1.5 protein that constitutes the pore-forming α-subunit of the cardiac fast sodium channel. Often, sinus bradycardia due to mutations in the SCN5A gene is observed as part of the SCN5A overlap syndrome,14 but there are also mutations that present with sinus bradycardia as the sole phenotype (see10 and primary references cited therein).

In 2003, Benson et al.15 revealed autosomal recessive mutations in SCN5A as a molecular cause of sick sinus syndrome. Also, in 2003, studies by Mohler et al.16 and Schulze-Bahr et al.17 linked mutations in the ANK2 and HCN4 genes, respectively, to sick sinus syndrome. In case of the ANK2 gene, which encodes the ankyrin-B protein, sinus bradycardia occurs in a spectrum of symptoms associated with the ‘ankyrin-B syndrome’.18 The HCN4 gene encodes the pore-forming α-subunit of the HCN4 ion channels underlying the hyperpolarization-activated ‘funny current’, If, which is also known as the ‘pacemaker current’. Recently, the KCNE2 gene, which encodes the MinK-related peptide 1 (MiRP1), was reported as another ion channel related gene associated with sick sinus syndrome.19 The MiRP1 protein is an important accessory β-subunit of the If channel20 and other cardiac ion channels, in particular, of the KCNH2 encoded HERG channel that underlies the delayed rectifier potassium current, IKr.21 Furthermore, Holm et al.22 identified MYH6, which encodes the alpha heavy chain subunit of cardiac myosin, as a non-ion-channel-related sick sinus syndrome susceptibility gene. Thus, the number of genes involved in congenital sick sinus syndrome is rapidly evolving.

Starting with the aforementioned report by Schulze-Bahr et al.,17 several mutations have been identified that link sick sinus syndrome to mutations in the hyperpolarization-activated cyclic nucleotide-gated (HCN) gene family that mediates the ‘pacemaker current’, If, in the heart (for reviews, see2327). The HCN channel family comprises of four members, HCN1–HCN4, which can form HCN channels in homomeric as well as heteromeric tetramers. The dominant HCN transcript in human SAN is HCN428 and the HCN4 locus has been identified as a modulator of heart rate in a genome-wide association study.29 Thus far, reports of mutations affecting human If have been restricted to HCN417,3035 or KCNE2, encoding the If modulating MiRP1 subunit.19 Voltage-clamp experiments on wild-type and mutant human HCN4 channels expressed in COS-7 cells, Xenopus oocytes, or HEK-293 cells have revealed changes in the expression and/or kinetics of mutant HCN4 channels, but the extent to which these changes would affect If flowing during a human SAN action potential is unresolved.

Here, we first give an overview of the HCN4 and MiRP1 mutations associated with sick sinus syndrome and their effects on the biophysical properties of the HCN4 current. Next, we briefly review the current-clamp and voltage-clamp data that we obtained from the isolated human SAN pacemaker cells and the mathematical model of human If that we based on these data. Finally, we show how this model can be used in simulated action potential clamp experiments, with action potentials recorded from isolated human SAN pacemaker cells as command potential, to assess the effects of the aforementioned HCN4 and KCNE2 mutations on If flowing during a human SAN action potential.

Mutations in HCN4 and KCNE2 associated with sick sinus syndrome

To date, seven mutations in HCN4 and one in KCNE2 have been associated with the sick sinus syndrome.17,19,3035 Figure 1A indicates the location of each of these mutations on the HCN4 and MiRP1 proteins. The clinical features are summarized in Figure 1B and C and Table 1, whereas changes in expression or biophysical properties associated with the mutations are described below, in chronological order of publication, and summarized in Table 2.

View this table:
Table 1

Clinical observations in carriers of mutations in HCN4 or KCNE2

StudyYear of publicationMutationMutation carriersClinical presentation
Mutations in HCN4
 Schulze-Bahr et al.172003573XSingle index patient (66-year-old female)Idiopathic sinus bradycardia of 41 b.p.m.; chronotropic incompetence; intermittent episodes of atrial fibrillation
 Ueda et al.302004D553NSingle index patient (43-year-old female) and two family membersWide spectrum of cardiac arrhythmias, including severe bradycardia (24 h average of 39 b.p.m.), QT prolongation, and Torsade de Pointes in index patient; QT prolongation in family members
 Milanesi et al.312006S672R15 Members of a single Italian familyAsymptomatic sinus bradycardia; average resting heart rate, corrected for age and gender, of 52.2 ± 1.4 b.p.m. (range 43–60 b.p.m.), in the 15 mutation carriers vs. 73.2 ± 1.6 b.p.m. (range 64–81 b.p.m.) in the 12 non-affected family members
 Nof et al.322007G480R8 Members of a single familyAsymptomatic sinus bradycardia from a young age, with normal chronotropic and exercise capacity; minimum, average, and maximum heart rate of 31 ± 8, 48 ± 12, and 101 ± 21 b.p.m., respectively, in the 8 mutation carriers vs. 55 ± 9, 73 ± 11, and 126 ± 16 b.p.m., respectively, in the 8 non-carriers
 Schweizer et al.332010695X8 Members of a single German familyMarked sinus bradycardia with no signs of chronotropic incompetence; basal heart rate of 45.9 ± 4.6 b.p.m. (range 38–51 b.p.m.) in the 8 mutation carriers vs. 66.5 ± 9.1 b.p.m. in the 6 non-carriers; minimum heart rate of 35.9 ± 5.6 vs. 47.2 ± 5.9 b.p.m.; maximum heart rates of 160.3 ± 12.6 vs. 171.8 ± 18.7 b.p.m.
 Laish-Farkash et al.342010A485V14 Members of three Moroccan Jewish decent familiesSymptomatic familial sinus bradycardia with normal chronotropic and exercise capacity; minimum, average, and maximum heart rate of 37 ± 3, 58 ± 6, and 117 ± 27 b.p.m. in the 14 mutation carriers, respectively, vs. 49 ± 11, 77 ± 12, and 140 ± 32 b.p.m., respectively, in the 6 non-carriers
 Duhme et al.352013K530N6 Members of a single familyFamilial age-dependent tachycardia–bradycardia syndrome and persistent atrial fibrillation; no atrial fibrillation or any other relevant cardiac arrhythmia in non-carriers; mild, asymptomatic sinus bradycardia (50–60 b.p.m.) in index patient
Mutations in KCNE2
 Nawathe et al.192013M54TSingle index patient (55-year-old Caucasian male)History of marked sinus bradycardia; average heart rate of 43 b.p.m. (range 30–125 b.p.m.), along with pauses; daughter died suddenly at the age of 13, post-mortem genetic testing revealed the M54T mutation
  • Mutations are heterozygous with autosomal dominant inheritance.

View this table:
Table 2

Effect of mutations in HCN4 or KCNE2 on biophysical properties of HCN4 current

MutationType of expressionExpression systemShift in V1/2 or activation threshold (mV)Slope factor (mV)Time constant of activationTime constant of deactivationReversal potentialCurrent densityChannel expressionSensitivity to cAMPReference
Mutations in HCN4
G480RHeteromericOocyte, HEK≈−15??=≈50%?=Nof et al.32
HomomericOocyte, HEK≈−30??=≈12%=Nof et al.32
Oocyte?????≈20%??Laish-Farkash et al.34
A485VHeteromericOocyte, HEK≈−30?=≈33%??Laish-Farkash et al.34
HomomericOocyte, HEK≈−60?=≈5%?Laish-Farkash et al.34
K530NHeteromericHEK≈−14=237%===?Duhme et al.35
HomomericHEK======?=Duhme et al.35
D553NHeteromericCOS==≈90%≈110%?≈37%?Ueda et al.30
Oocyte, COS?????≈54%=?Netter et al.36
HomomericCOS==≈90%≈110%?≈8%?Ueda et al.30
Oocyte, COS====?≈12%=Netter et al.36
573XHeteromericCOS=*−1.9*=???=Schulze-Bahr et al.17
HomomericCOS−4.6*==???=Schulze-Bahr et al.17
S672RHeteromericHEK−4.9==≈74%????Milanesi et al.31
HomomericHEK−8.4==≈63%???=Milanesi et al.31
Oocyte−6.1?≈180%≈90%???Xu et al.37
695XHeteromericHEK===??=?Schweizer et al.33
HomomericHEK=−3.572%==??Schweizer et al.33
Mutations in KCNE2
M54THomomericNRVM==192%=?18%??Nawathe et al.19
  • Data are changes relative to wild-type current.

  • ?, not reported; ≈, estimated from presented figures; ↓, decreased; ↑, increased; =, unchanged. *15 s hyperpolarizing pulses; changes reported, but no quantitative data provided; performed in oocytes which lack cAMP modulation due to high basal activity32; oocyte, HEK, COS, and NRVM: Xenopus oocytes, HEK-293 cells, COS-7 cells, and neonatal rat ventricular cardiomyocytes, respectively.

Figure 1

Mutations in HCN4 and MiRP1 associated with sinus bradycardia. (A) Schematic topology of the HCN4 and MiRP1 proteins. The MiRP1 β-subunit has a single transmembrane segment with an extracellular N-terminus and intracellular C-terminus, whereas the HCN4 α-subunit has six transmembrane segments (S1–S6), a pore-forming loop (P), and intracellular N- and C-termini. The voltage sensor of the channel is formed by the positively charged S4 helix. The C-terminus comprises of the C-linker (dotted line) and the cNBD, which is known to mediate cAMP dependent changes in HCN channel gating. Red dots indicate the location of the eight known HCN4 and MiRP1 mutation sites associated with sinus bradycardia. The split dots indicate the truncations resulting from the 573X and 695X non-sense (truncating) mutations. (B, C) Resting heart rate (B) and heart rate during exercise (C) in carriers of each of the eight heterozygous mutations (filled symbols) and unaffected family members (open symbols). Squares indicate data from a single index patient (filled squares) or an estimate for a non-carrier (open square). Bars indicate SEM. #Not corrected for age and gender. Note the difference in ordinate scale and axis breaks.

HCN4-573X

The HCN4-573X mutation results in a truncated HCN4 protein that lacks the cyclic nucleotide-binding domain (cNBD) (Figure 1A). Transiently transfected COS-7 cells indicated normal intracellular trafficking and membrane integration of HCN4-573X subunits.17 Patch-clamp experiments demonstrated a shift in channel activation to more hyperpolarized potentials and a steeper steady-state activation curve (Table 2), but this required 15 s long hyperpolarizing voltage steps (at 20–22°C) to become apparent. Both homomeric and heteromeric channels appeared insensitive to cyclic adenosine monophosphate (cAMP), demonstrating a dominant-negative effect of HCN4-573X on wild-type subunits.

HCN4-D553N

A functional study in COS-7 cells30 showed a reduced membrane expression, and decreased current, because of a dominant-negative trafficking defect of the mutant HCN4-D553N protein. The voltage dependence of activation of the mutant HCN4 channel was comparable with wild-type, but activation was faster while deactivation was slower (Table 2). However, Netter et al.36 reported that D553N mutant channels have normal trafficking, with similar surface expression of D553N and wild-type channels in COS-7, HeLa, and HL-1 cells. In Xenopus oocytes and COS-7 cells, D553N channels generated currents with reduced amplitude, but unaltered kinetics. Furthermore, D553N channels did not respond to adrenergic stimulation.

HCN4-S672R

A functional study in HEK-293 cells revealed a shift in channel activation to more hyperpolarized potentials and faster deactivation of HCN4-S672R mutant channels (Table 2). The cAMP-dependent shifts in voltage dependence of activation, as assessed in inside-out macropatches, were similar in wild-type and S672R mutant channels,31 suggesting that this mutation did not affect the sensitivity to cAMP. However, Xu et al.37 found a reduced sensitivity to cAMP by using inside-out patch-clamp recordings in Xenopus oocytes (but see DiFrancesco27).

HCN4-G480R

Western blot analysis demonstrated significantly reduced membrane expression of homomeric HCN4-G480R channels in HEK-293 cells.32 Functional analysis in Xenopus oocytes and HEK-293 cells revealed a decrease in current density, accompanied by a hyperpolarizing shift in voltage dependence of activation and slowing of activation kinetics (Table 2). In Xenopus oocytes, neither wild-type nor G480R currents were modulated by β-adrenergic stimulation, probably due to the high levels of endogenous cAMP in Xenopus oocytes.38 Thus, whether the G480R mutation affects the sensitivity to cAMP is unresolved. Recently, Laish-Farkash et al.34 confirmed the decrease of current density in HEK-293 cells.

HCN4-695X

The HCN4-695X mutation results in a truncated cNBD (Figure 1A). Patch-clamp experiments in HEK-293 cells33 demonstrated a steeper slope of the activation curve and faster activation of homomeric 695X mutant current as well as insensitivity to cAMP (Table 2). Heteromeric channels failed to respond to cAMP, just as homomeric mutant channels, indicating a dominant-negative suppression of cAMP responsiveness by the mutant subunits.

HCN4-A485V

Western blot analysis revealed significantly reduced membrane expression of homomeric HCN4-A485V channels in HEK-293 cells.34 Functional analysis in Xenopus oocytes and HEK-293 cells demonstrated large hyperpolarizing shifts of the voltage dependence of activation. Moreover, activation and deactivation were slowed and current density was largely reduced (Table 2).

HCN4-K530N

Patch-clamp experiments in HEK-293 cells35 demonstrated that the biophysical properties of homomeric HCN4-K530N mutant channels are similar to wild-type channels (Table 2). Experiments on heteromeric channels, however, indicated that interaction of neighbouring mutant and wild-type subunits is impaired, causing a significant hyperpolarizing shift of the half-maximal activation voltage and slowed activation of the HCN4 current. Heteromeric channels showed larger sensitivity to cAMP than wild-type or homomeric K530N channels as demonstrated by the ≈7 mV larger cAMP-induced shift of the activation curve and larger change in activation time constant.

KCNE2-M54T

Patch-clamp experiments in neonatal rat ventricular myocytes demonstrated that co-expression with M54T mutant MiRP1 decreased HCN4 current density by 80% compared with HCN4 alone or HCN4 co-expressed with wild-type MiRP1. In addition, co-expression with M54T MiRP1 slowed HCN4 activation at physiologically relevant voltages, while HCN4 deactivation and voltage dependence of activation were not affected.19

Additional mutations in HCN4

In 2009, Ueda et al.39 described a novel HCN4 mutation in a Brugada syndrome patient. This mutation was an insertion of four bases at a splicing junction of exon 2 and intron 2, probably resulting in a defective HCN4 protein without the ability to form functional tetramers. The mutation is neither included in Figure 1 nor in our simulations, because sinus node dysfunction was not observed, which was explained by ‘the presence of unaffected splicing’.39

Evans et al.40 recently reported HCN4 mutations in two cases of sudden infant death syndrome (SIDS). Although mutations in HCN4 have not been associated with SIDS thus far,41 the HCN4-A195V and HCN4-V759I mutations were considered ‘potentially pathogenic’.40 Since an association of these mutations with cardiac arrhythmias and/or an in vitro study of the mutations are lacking, these mutations are not included in Table 2 and our simulations.

If in human sinoatrial node cells

Experimental data on the electrophysiology of human SAN cells are scarce. Some years ago we had the opportunity to make patch-clamp recordings from single pacemaker cells that were isolated from a patient without structural heart disease who underwent SAN excision.42 From a total of three cells we recorded action potentials and If. From the acquired data, we constructed a mathematical model of If in human SAN cells.43,44 Here, we present a brief overview of our experimental data and our mathematical model of If. Also, we discuss how modulation of If by cAMP can be implemented into the model.

Experimental data on human If

In voltage-clamp experiments at 36 ± 0.2°C, each of the human SAN pacemaker cells showed an inward current that was activated upon 2 s hyperpolarizing steps from −40 mV. This current became larger and was activated more rapidly at increasingly negative potentials and was strongly reduced by 2 mM Cs+, which characterized it as If. Figure 2A shows the If steady-state activation data that were obtained by plotting the normalized Cs+-sensitive tail current amplitude against potential (filled red circles). The dashed red line is the Boltzmann curve that was fitted to the data. The membrane potentials for half-maximal activation (V1/2) and the slope factor were −96.9 ± 2.7 and 8.8 ± 0.5 mV (mean ± SEM, n = 3), respectively. The open blue circles are recent experimental data on If in rabbit SAN cells.45 With a value of 9.0 mV, the slope factor of the associated Boltzmann curve (dashed blue line) is similar, but the V1/2 of −73.0 mV45 is considerably less negative than that of the human cells. It should, however, be realized that the actual human V1/2, as averaged from the three cells, may be less negative than −96.9 mV due to lack of full activation of If during the 2 s voltage-clamp steps at less negative potentials.42

Figure 2

Pacemaker activity of single pacemaker cells isolated from the SAN. (AC) Characteristics of the hyperpolarization-activated ‘pacemaker current’ (If) in rabbit and human SAN cells. (A) Normalized steady-state activation of If. The dashed curves are the Boltzmann fits to the experimental data. (B) Time constant of If activation (circles) and deactivation (squares). The dashed curves are bell-shaped fits to the experimental data. (C) Current–voltage relationship of the If ‘step current’, normalized to membrane capacitance, at the end of a 2 s voltage-clamp step. (D) Typical action potentials of single rabbit (blue) and human (red) SAN cells. Data are means ± SEM (n = 9 and n = 3 for rabbit and human, respectively). All the membrane potential values are corrected for the estimated liquid junction potential.

The time courses of activation and deactivation of human If could be well fitted by mono-exponential functions, yielding the time constants shown in Figure 2B (filled red symbols). Again, the open blue symbols are recent experimental data on If in rabbit SAN cells.45 Like the activation curve of Figure 2A, the time constant curve of Figure 2B is shifted to more hyperpolarized potentials for the human cells. Figure 2C shows that the If current density in the human SAN cells is considerably smaller than in rabbit SAN cells.24 Figure 2D shows typical action potentials recorded from human and rabbit SAN cells. The major difference is in the longer spontaneous cycle length and slower diastolic depolarization of the human cells, which amounted to 828 ± 15 ms and 49 ± 18 mV/s, respectively.42

Mathematical model of human If

The experimental data of Figure 2A–C were used to construct a mathematical model of If in human SAN cells.43,44 In line with the aforementioned mono-exponential fits, we used a first-order Hodgkin and Huxley-type kinetic scheme, with Embedded Image (1) Embedded Image (2) Embedded Image (3) Embedded Image (4) Embedded Image (5)

where If is expressed in pA/pF, y is the Hodgkin and Huxley-type gating variable, gf is the fully activated If conductance (in nS/pF), Vm is the membrane potential (in mV), Ef is the If reversal potential (in mV), y is the steady-state value of y, and τy is the time constant of (de)activation (in ms).

Equation (3) represents the original Boltzmann fit to the steady-state activation data of Figure 2A. Since this Boltzmann fit levels off at 0.01329 instead of 0, which would result in an overestimation of If if used in our computer simulations, we introduced a separate equation for y at Vm≥ −80 mV [Eq. (4)] as an optimal fit to the data in the physiologically relevant membrane potential range. Equation (5) is our fit to the If (de)activation data of Figure 2B and appears as the dashed red line in Figure 2B. The fully activated If conductance gf of Eq. (1) was set to 0.075 nS/pF and the If reversal potential Ef to −22 mV, in accordance with the fully activated If conductance of 75.2 ± 3.8 pS/pF and If reversal potential of −22.1 ± 2.4 mV (mean ± SEM, n = 3).42

Modulation of If by cyclic adenosine monophosphate

In rabbit SAN cells, the spontaneous beating rate increases upon adrenergic stimulation. This is illustrated in Figure 3A, which shows typical action potentials recorded under basal conditions and in the presence of noradrenalin (NA).46 In eight cells, NA increased the diastolic depolarization rate by 19 ± 5 mV/s, decreased the cycle length by 56 ± 11 ms, and increased the maximum upstroke velocity by 0.7 ± 0.2 V/s, without significant changes in other action potential parameters.46 An important contributor to the increase in rate is If. Direct binding of cAMP results in a shift of the steady-state activation curve by ≈10 mV,47 accompanied by a similar shift of the time constant curve.48 The resulting increase in If leads to an increase in the diastolic depolarization rate, which in turn leads to an increase in the beating rate (Figure 3A).

Figure 3

Effects of adrenergic stimulation on pacemaker activity of single SAN pacemaker cells. (A) Typical action potentials of an isolated rabbit SAN pacemaker cell under control conditions (‘control’, dashed blue line) and in the presence of 1 μM NA (solid red line). (B, C) Shift in If activation curve (B) and time constant curve (C) upon adrenergic stimulation. Horizontal arrow indicates the 15 mV shift.

Data on the modulation of If by cAMP in human SAN cells are lacking. However, experiments on human HCN4 channels in mammalian expression systems have demonstrated similar effects as observed in rabbit SAN cells,4952 with a somewhat larger shift in V1/2 of ≈15 mV.24 Accordingly, we incorporated the modulation of If by cAMP through a 15 mV shift in the y and τy curves, as illustrated in Figure 3B and C.

Functional effects of HCN4 and KCNE2 mutations on human If

The mathematical model of human If can be used in combination with the acquired action potentials to reconstruct If during the time course of a human SAN action potential.24,44 Here, we show how this can be applied to assess the functional effects of the aforementioned HCN4 and KCNE2 mutations on If.

Reconstruction of If

For the numerical reconstruction of If, we used a train of action potentials recorded from one of our human SAN cells. This train was used as the command potential in a simulated action potential clamp experiment, in which If was described by Eqs. (1)–(5), with the parameter settings adapted for the mutation of interest and the modulation by cAMP (Table 3).

View this table:
Table 3

Parameter settings in simulated action potential clamp experiments

MutationScaling factor for gfShift (mV)Shift with cAMP (mV)
HCN4-G480R0.50−150
HCN4-A485V0.33−30−15
HCN4-K530N1−14+7.8
HCN4-D553N0.370+15
HCN4-573X100
HCN4-S672R1−4.9+10.1
HCN4-695X100
KCNE2-M54T0.180+15
  • Scaling factor and shift relative to wild-type.

  • Shift applied to both steady-state activation curve and time constant curve.

In the wild-type case, the numerical reconstruction results in the If trace of Figure 4B (solid blue line). The net membrane current (Inet, noisy green trace) and If are of similar amplitude, but this does not imply that If is the major contributor to Inet, since Inet is the net result of multiple inward and outward currents.24

Figure 4

Effect of mutations on If in human SAN pacemaker cell assessed by simulated action potential clamp. (A) Action potentials recorded from human SAN pacemaker cell used for action potential clamp. (B) Computed ‘wild-type’ (WT) If of human SAN pacemaker cell during action potentials of panel A (blue line). Net membrane current (Inet) computed from Inet = −Cm× dVm/dt, where Cm and Vm denote membrane capacitance and membrane potential, respectively, shown in green. (CJ) Computed If of human SAN pacemaker cell carrying heterozygous mutation in (CI) HCN4 or (J) MiRP1, as indicated, during the action potentials of panel A (solid red line). Wild-type If of panel B shown for reference (dashed blue line). Note the difference in the current scale between panel B and panels CJ.

In the absence of action potentials recorded upon adrenergic stimulation, the same train of action potentials was used to reconstruct If in case of elevated cAMP levels. This was achieved by repeating the simulations with the activation and time constant curves shifted by +15 mV. The resulting If trace is shown as a solid blue line in Figure 5B, together with the control trace of Figure 4B, which is now shown as a dashed red line. It is immediately apparent from Figure 5B that the amplitude of If is more than doubled upon adrenergic stimulation.

Figure 5

Effect of mutations on If in human SAN pacemaker cell in case of adrenergic stimulation assessed by simulated action potential clamp. (A) Action potentials recorded from human SAN pacemaker cell used for action potential clamp. (B) Computed ‘wild-type’ (WT) If of human SAN pacemaker cell upon adrenergic stimulation during action potentials of panel A (solid blue line). Control If of Figure 4B shown for reference (dashed red line). (CJ) Computed If of human SAN pacemaker cell carrying heterozygous mutation in (CI) HCN4 or (J) MiRP1, as indicated, upon adrenergic stimulation during action potentials of panel A (solid blue line). Wild-type If of panel B shown for reference (dashed red line).

To quantify the contribution of If to diastolic depolarization, we computed the charge carried by If (Qf) during the 25 mV, 550 ms depolarization that starts at the maximum diastolic potential of −63 mV (Figure 4A, double-headed arrow). Under control conditions, Qf amounts to 0.018 pC/pF (Figure 6, leftmost blue bar), which is somewhat smaller than the net charge flow of 0.025 pC/pF or, equivalently, 25 mV (Figure 6, dashed grey line). Upon adrenergic stimulation, Qf increases 2.4-fold to 0.042 pC/pF (Figure 6, leftmost red bar). Thus, one might say that wild-type If has a ‘depolarization reserve’ of 24 mV, i.e. the difference between the ‘depolarization power’ of 18 mV under basal conditions and 42 mV upon adrenergic stimulation.

Figure 6

Contribution of If to diastolic depolarization for each of the heterozygous mutations in HCN4 or MiRP1. Charge carried by If (Qf) during the 25 mV, 550 ms spontaneous depolarization from the maximum diastolic potential of −63 mV of the human SAN action potential as indicated in Figure 4A. Left blue bars are computed from the If traces of Figure 4, i.e. under control conditions (‘control’). Right red bars are computed from the If traces of Figure 5, i.e. upon adrenergic stimulation (‘cAMP’). The dashed grey line indicates the charge of 0.025 pC/pF carried by the net membrane current (Qnet) during the 25 mV depolarization.

HCN4-G480R mutation

Experiments on heteromeric HCN4 channels learn that the heterozygous G480R mutation results in a ≈50% reduction in current density as well as a −15 mV shift in the activation curve (Table 2). These effects were implemented in our simulations by scaling the If conductance by 0.5 and shifting the voltage dependence by −15 mV (Table 3), which means that we shifted the steady-state activation curve as well as the time constant curve by −15 mV. As a result, the time constant of activation at a given hyperpolarized potential is increased (Figure 3C), which is in accordance with the experimental data (Table 2).

If is drastically reduced in amplitude (Figure 4C) and Qf is reduced by >80% to only 0.0033 pC/pF (Figure 6), equivalent to a ‘depolarization power’ of only 3.3 mV. This might explain the marked sinus bradycardia in the mutation carriers (Table 1). The HCN4 channels do not loose their sensitivity to cAMP (Table 2). Therefore, we simulated modulation by cAMP by a +15 mV shift in the activation and time constant curves. The ‘depolarization power’ thereby increases to 8.8 mV, which is still only 21% of wild-type (Figure 6), which seems at odds with the normal exercise testing and the absence of any evidence of chronotropic incompetence in the mutation carriers.32

HCN4-A485V mutation

Like the G480R mutation, the A485V mutation results in a reduction in current density, to 33% instead of 50%, as well as a negative shift in the activation curve, by −30 mV instead of −15 mV (Table 2). Accordingly, the effects on the simulated If are more severe, both under control conditions (Figure 4D) and in the presence of cAMP (Figure 5D). The ‘depolarization power’ in terms of Qf is reduced to 5 and 3% of wild-type, respectively (Figure 6). The reduction to 5% under basal conditions is consistent with the marked sinus bradycardia in the mutation carriers, but the reduction to 3% in the presence of cAMP is inconsistent with their normal chronotropic and exercise capacity.34 It should, however, be noted that experimental data on cAMP sensitivity are lacking, so that cAMP effects were simulated by the ‘normal’ +15 mV shift.

HCN4-K530N mutation

The effects of the K530N mutation are limited to a −14 mV shift in voltage dependence (Table 2). Accordingly, If is less affected than in case of the above G480R and A485V mutations (Figure 4E), with a reduction in Qf by 60% (Figure 6). Since the sensitivity to cAMP is increased—heteromeric K530N channels demonstrate a +21.8 mV shift in voltage dependence instead of the +14.3 mV shift of wild-type channels35—this reduction in Qf is limited to 32% upon adrenergic stimulation (Figure 6). The simulation data are consistent with the mild bradycardia observed in the mutation carriers.35

HCN4-D553N mutation

In case of the D553N mutation, there is no shift in voltage dependence. Also, sensitivity to cAMP is not affected. However, the current density of heteromeric D553N channels is reduced to ≈37% in COS-7 cells (Table 2). Accordingly, we reduced gf to 37% of control, which of course leads to an equivalent reduction in If amplitude (Figures 4F and 5F) and Qf (Figure 6). Given the aforementioned differences between the experimental data of Ueda et al.30 and Netter et al.,36 it is difficult to draw any firm conclusion regarding the compatibility of the simulation results with the clinical data of the single index patient.30

HCN4-573X mutation

The main effect of the 573X mutation is that the HCN4 channels loose their sensitivity to cAMP (Table 2). As a result, the ‘depolarization reserve’, in terms of an increase in ‘depolarization power’ upon adrenergic stimulation, is reduced to zero (Figure 6). This is in line with the chronotropic incompetence in the single index patient.17 However, the absence of a mutation effect under basal conditions (Figure 4G) cannot explain the marked sinus bradycardia.

HCN4-S672R mutation

The effects of the S672R mutation seem limited to a slight negative shift in voltage dependence (Table 2). Accordingly, our simulations predict a relatively small decrease in both If (Figures 4H and 5H) and Qf (Figure 6), consistent with the mild asymptomatic sinus bradycardia in the mutation carriers31 and the apparent absence of chronotropic incompetence.27

HCN4-695X mutation

Patch-clamp data on the 695X truncating mutation are essentially identical to those on the 573X truncating mutation (Table 2). Consequently, there are no differences in the simulation results between the two mutations (Figures 4G and I, Figures 5G and I). Like the 573X mutation, the absence of a mutation effect under basal conditions (Figure 4I) cannot explain the marked sinus bradycardia observed in the mutation carriers. Furthermore, the reduction in the ‘depolarization reserve’ (Figure 6) is at odds with the absence of signs of chronotropic incompetence in the mutation carriers.33 Of note, the single index patient with the 573X mutation showed clear chronotropic incompetence.17

KCNE2-M54T mutation

The effects of the M54T mutation in KCNE2 seem limited to a significant decrease in HCN4 current density (Table 2). Accordingly, the mutation effects were simulated by reducing gf by 82%, resulting in an equivalent reduction in both If amplitude (Figures 4J and 5J) and Qf (Figure 6). Since clinical data are limited to a single index patient19 and the effects of a mutation in KCNE2 are certainly not limited to If, it is impossible to draw any reasonable conclusion regarding the compatibility of the simulation results with the clinical data.

Some concluding remarks

The functional effects of the loss-of-function mutations in HCN4 and KCNE2, as assessed in voltage-clamp experiments on homomeric and/or heteromeric mutant channels in expression systems, seem not fully compatible with the clinically observed features. For example, carriers of the 695X mutant in HCN4 have no signs of chronotropic incompetence, while this mutation induces a complete insensitivity to cAMP.33 This might suggest that If is less important in human pacemaker activity during exercise than during basal rhythm, which can be explained by an inhibited activation of If due to the considerably shorter action potentials during exercise, and that the high rate during exercise is largely determined by other inward currents, in particular, the L-type calcium current53 and the sodium–calcium exchange current as part of the ‘calcium clock’.54

In line with the in vitro data, our in silico data of Figures 46 also raise questions regarding some apparent inconsistencies between these data and the clinical observations. However, it should be noted that the in vitro data are often incomplete (cf. Table 2) and that the clinical data are in most cases limited to a small number of patients or even a single index patient. Furthermore, heart rate data are not always corrected for age and gender, which may either exaggerate or obscure differences between mutation carriers and non-carriers.

With regard to the in vitro data, there are several other issues that should be kept in mind. One is the general tendency to draw firm conclusions from observed changes in channel kinetics and ignore the fact that these are not always significant at physiologically relevant membrane potentials. Another issue are the apparently conflicting in vitro data on particular mutations. This holds for the D553N mutation, as assessed by Ueda et al.30 and Netter et al.,36 as well as for the S672R mutation, as assessed by Milanesi et al.31 and Xu et al.37

Another point that requires attention is the common identification of If channels with HCN4 channels. One should realize that HCN4, although abundant, is not the single HCN subunit in human SAN. It is therefore conceivable that a considerable amount of HCN tetramers is not fully built of HCN4 subunits, which may not only be important for the behaviour of the wild-type current51 but also for the mutant current.

An important general issue is that, in addition to MiRP1, several other regulatory elements are known to interact with HCN4 or If channels (‘context dependence’55). These include PIP2, caveolin-3, and SAP97.5659 In case of a mutation in HCN4, this interaction may be affected, thereby altering expression levels and/or kinetics of If, without being directly noticed in the expression systems that are commonly used to assess mutation effects.

Another general issue is that the alterations in cardiac rhythm of the mutation carriers (sinus bradycardia, tachycardia–bradycardia syndrome, atrial fibrillation) may induce a significant remodelling of If. For example, atrial tachyarrhythmias cause a downregulation of If in dogs,60 whereas chronic atrial fibrillation causes an upregulation of If in humans.61

We preferred to study If in simulated action potential clamp experiments, thus ensuring that the action potential followed the course of that of a human SAN cell. One could also incorporate the human If equations into one of the available comprehensive mathematical models of a SAN cell, which are mostly rabbit SAN cells, but one should realize that the results regarding the effects of If on beating rate are largely dependent on the ‘model environment’.62

In conclusion, the identification of HCN4 and KCNE2 mutations in relation to sinus node dysfunction has not only provided us with valuable information, but also with intriguing new questions regarding the role of If in human SAN.

Conflict of interest: none declared.

References

View Abstract