Functional study of a KCNH2 mutant: Novel insights on the pathogenesis of the LQT2 syndrome

Abstract The K+ voltage‐gated channel subfamily H member 2 (KCNH2) transports the rapid component of the cardiac delayed rectifying K+ current. The aim of this study was to characterize the biophysical properties of a C‐terminus‐truncated KCNH2 channel, G1006fs/49 causing long QT syndrome type II in heterozygous members of an Italian family. Mutant carriers underwent clinical workup, including 12‐lead electrocardiogram, transthoracic echocardiography and 24‐hour ECG recording. Electrophysiological experiments compared the biophysical properties of G1006fs/49 with those of KCNH2 both expressed either as homotetramers or as heterotetramers in HEK293 cells. Major findings of this work are as follows: (a) G1006fs/49 is functional at the plasma membrane even when co‐expressed with KCNH2, (b) G1006fs/49 exerts a dominant‐negative effect on KCNH2 conferring specific biophysical properties to the heterotetrameric channel such as a significant delay in the voltage‐sensitive transition to the open state, faster kinetics of both inactivation and recovery from the inactivation and (c) the activation kinetics of the G1006fs/49 heterotetrameric channels is partially restored by a specific KCNH2 activator. The functional characterization of G1006fs/49 homo/heterotetramers provided crucial findings about the pathogenesis of LQTS type II in the mutant carriers, thus providing a new and potential pharmacological strategy.

the pore domain from each of the four subunits lining the central ion conduction pathway. In addition, the KCNH2 protein contains large cytoplasmic NH2-terminal and COOH-terminal domains with several regulatory sites (for review, see 2 ).
KCNH2 channels can exist in closed, open or inactivated states, and like other voltage-gated K + channels, they contain multiple positive charges in the S4 domain and this acts as the primary voltage sensor for channel opening. 3 The inactivation process of KCNH2 channels, however, exhibits several unusual features: the kinetics of inactivation is much more rapid than the kinetics of activation, 4 and the inactivation process is voltage-dependent. 5 This unique combination of biophysical properties underlies the physiological role that KCNH2 plays in cardiac repolarization.
About 300 mutations in the KCNH2 gene have been identified to date (http://www.fsm.it/cardm oc/). A variety of mechanisms underlie the dysfunction of the KCNH2 channel in these mutants, including the failure of the mutant channels to reach the cell surface, defective activation/inactivation, loss in permeation selectivity and dominant-negative suppression of the function of wild-type channels. 6 In LQTS2, syncope is often the first presenting symptom and the risk of sudden cardiac death is very high. Genetic testing can give invaluable information to support diagnosis in the patient and prognosis in familiars and to address treatment in both. The management of syncope survivors and relatives at high risk often consists in the use of an implantable cardioverter defibrillator (ICD) to correct or interrupt life-threatening arrhythmias.
In LQTSs, general emotional stress and intense exercise play a prominent role in triggering ventricular arrhythmias. Thus, the only pharmacological approach so far consists of the use of β-adrenoceptor antagonists to reduce sympathetic or adrenergic stimulation. 7 However, as β-blocker efficacy is variable, ICD is the only treatment proved effective in treating the resulting arrhythmia and preventing sudden death in patients affected by LQTSs. Indeed, the pharmacotherapy in LQTSs is a challenging issue in the actual clinical management and deciphering the exact mutation-specific pathogenetic mechanism is fundamental to design therapies tailored upon patients' requirements.
Here, we characterized by electrophysiology the biophysical properties of a C-terminus-truncated mutant of KCNH2 (G1006fs/49) that causes QT interval prolongation in members of an Italian family. We identify new insights regarding the role of the C-terminal tail of the KCNH2 channel, the pathogenetic mechanisms underlying the LQTS type II identified in the mutant carriers and a possible pharmacological strategy.

| Patients
The heterozygous c.3017delG mutation in KCNH2 gene (NM_000238) exon 13, identified for the first time in 2003 (www. ncbi.nlm.nih.gov/varia tion/tools/ 10 0 0g enome s/?gts=rs794 728504), 8 has been found in five members of an Italian family screened in our Clinical Unit dedicated to cardiomyopathies. The mutation consists of the G deletion at position 3017 in the KCNH2 coding region, causing both a translational frame shift starting from the glycine 1006 and a premature protein termination 49 amino acids forward. This KCNH2 variant will be indicated throughout the manuscript as G1006fs/49.

| Confocal analysis
Transfected cells were plated on sterile 15-mm-diameter coverslips (Thermo Fisher Scientific) coated before cell plating with Poly-L-Lys for 20 minutes at RT (Sigma-Aldrich). After adhesion, the cells were fixed in ice-cold 100% methanol (Sigma-Aldrich) for 5 minutes. The coverslip was mounted upside down on a microscope slide using a drop of mounting medium composed by 50% PBS 2×, 50% glycerol 100% and 1% m/v n-propyl gallate; pH = 8. Confocal images were obtained with a confocal laser-scanning fluorescence microscope (Leica TSC-SP2).

| Electrophysiological recordings
After gigaseal formation and whole-cell access, pipette capacitance

| Statistical analysis
GraphPad Prism 6 was used for the statistical analysis and graph representation of the electrophysiological data. Data are given as mean ± standard error of the mean. Statistical analysis was performed using two-way ANOVA test for multiple comparisons.

| Clinical data of the Italian family carrying the G1006fs/49 variant
The heterozygous c.3017delG mutation in KCNH2 gene consists of the G deletion at position 3017 in the KCNH2 coding region, causing both a translational frame shift starting from the glycine 1006 and a premature protein termination 49 amino acids forward; indeed, the KCNH2 variant will be indicated throughout the manuscript as G1006fs/49. Figure 1A, the pedigree of the family, spanning  the Table S1. The proband (subject II-3) aged 23 years, after sudden awakening from sleep following a loud noise (her brother's intense crying), suffered from cardiac arrest due to ventricular fibrillation resuscitated by transthoracic DC shock. During the same night, her brother (subject III-5, aged 11 years) experienced syncope (with fall to the ground associated with trauma to the right cheekbone). ECGs from the proband ( Figure 1B) and her brother ( Figure 1C) showed prolonged corrected QT intervals (QTc). Both received an implantable cardioverter defibrillator (ICD). All mutant carriers were treated with β-blocker therapy (nadolol). During a follow-up of 15 years, the proband suffered from episodes of palpitations related to emotions and, 11 years after ICD implant, she had a syncope triggered by intense emotional stress. Rapid rate non-sustained ventricular tachycardia episodes were recorded and reported by ICD ( Figure 1D).

| Confocal microscopy analysis and Western blotting
Confocal images showed that both KCNH2 and G1006fs/49 homotetramers were able to reach the plasma membrane (Figure 2A

| Activation kinetics
KCNH2 current was elicited and recorded using a voltage protocol consisting of a holding potential of −80 mV, 4s depolarizing steps from −60 to +50 mV (with increments of 10 mV), and a third step at −100 mV, used to record the tail current. 4,9,10 Figure   When KCNH2 and G1006fs/49 channels were co-expressed in the same cells, the I-V curve reached the maximum value of current F I G U R E 1 A, Pedigree of the family carrying the G1006fs/49 mutant of the KCNH2 channel. Filled symbols indicate clinically affected individuals; + indicates positive for the mutation; arrow indicates the index patient. B, The index patient and (C) her brother had prolonged QT intervals corrected for heart rate (QTc) on electrocardiogram. D, The index patient was treated with beta-blocker therapy and received an implantable cardioverter defibrillator (ICD) but continued to exhibit tachyarrhythmic episodes that were recorded on ICD interrogation amplitude at +40mV. Although the maximum outward current recorded for KCNH2-G1006fs/49 heterotetramers (14.438 ± 2.740 pA/pF) was reached at more positive potential, the maximal outward current was not statistically different than those recorded for either KCNH2 or G1006fs/49 homotetramers. These results suggest that the G1006fs/49 homotetramers and more strikingly the KCNH2-G1006fs/49 heterotetramers have a delay in the voltage sensitivity of the activation, although they have a functional transporting pore.
After every depolarizing step, a further hyperpolarization at −100 mV allowed the recording of the tail current and used to construct the activation curve of the KCNH2 currents. The activation curves in Figure 3E were obtained fitting to a Boltzmann relationship of the K + tail current normalized to the maximum tail current amplitude.
The activation curve of KCNH2 homotetramers confirmed a threshold voltage close to −40 mV and that it was fully activated around +10 mV with a half-maximum activation voltage (V 1/2 ) of −13.082 ± 1.368 mV and a slope factor (κ) of 6.485 ± 0.182 ( Figure   3E,F, green trace). On the other hand, G1006fs/49 homotetramers exhibited a threshold voltage of around −20 mV and were fully activated at near +20 mV with a V 1/2 of −6.851 ± 1.984 mV and a slope factor (κ) of 8.022 ± 0.406 ( Figure 3F, red trace). Indeed, the normalized tail current from −20 mV to +0 mV and the half-maximum activation voltage (V 1/2 ) were significantly right-shifted for G1006fs/49 homotetramers compared to the KCNH2 homotetramers ( Figure   3E,F, green and red traces). In addition, the current of KCNH2-G1006fs/49 heterotetramers (orange trace) showed a threshold voltage around −40 mV and a fully activation close to +40 mV, with a V 1/2 of −0.976 ± 4.49 mV a slope factor (κ) of 8.4 ± 0.755 ( Figure   3F, orange trace).

| Inactivation kinetics
KCNH2 inactivation is well known to proceed faster than channel activation; thus, its investigation required a three-pulse voltage protocol. 4,9,10 The currents were activated (and inactivated) by a 200ms step to +60 mV. Then, the hyperpolarization of the membrane for 2 ms at −100 mV was used to allow the recovery from inactivation and finally a second set depolarizing steps at voltages between    Figure 4D shows the average of the time constant of inactivation measured at voltages between 0 and +60 mV for each type of channels. G1006fs/49 channels (red trace) were characterized by a faster inactivation compared to KCNH2 homotetramers at +20 mV, which was drastically enhanced in KCNH2-G1006fs/49 heterotetramers at all voltages tested.

| Recovery from inactivation
As previously reported, 4,9,10 once activated by depolarization, The recovery from inactivation was investigated for either KCNH2 ( Figure 5A), G1006fs/49 homotetramers ( Figure 5B) or KCNH2-G1006fs/49 heterotetramers ( Figure 5C) and plotted as mean ± SEM in Figure 5D. We found that for potentials more positive than −60 mV, the recovery from the inactivation was signifi-

| Effect of NS1643 on KCNH2 currents
First, the effect of the specific agonist NS1643 11 was investigated on KCNH2 channels activation, inactivation and recovery from inactivation as described in the previous paragraphs (see paragraphs In Figure 7B,

| D ISCUSS I ON
In this work, we functionally characterized a C-terminal truncating mutant of the KCNH2 channel, G1006fs/49, identified in members of an Italian family affected by LQTS type II.  Figure 3A,B,C).

Interestingly, both the C-terminal truncated KCNH2 variants
R1014PfsX39 and V1038AfsX21 did not show any disturbance in trafficking and maturation, 14 in contrast to the mutant R1014X, which had impaired trafficking towards the plasma membrane. 13 These findings suggest that the novel amino acids introduced by shifts in the reading frame may be still able of masking the ER retention signal in the KCNH2 C-terminus and of conferring stability to the channel. What we instead found of novelty in the C-terminus function mechanisms is that the sequence downstream the position 1006 seems to be important for the voltage sensitivity of the channels.
Interestingly, the mutant R1014PfsX39 did not show any alteration in the activation kinetics compared to the WT channel 14 leading to even narrow the C-terminal sequence involved in the We already found that the study of the pathomechanisms underlying other inherited cardiomyopathies is particularly important for the identification of mutation-specific therapy. [30][31][32] Moreover, the fact that NS1643 can recapitulate the wild-typelike activation kinetics in both G1006fs/49 homo-and heterotetramers channels paves a venue for a pharmacological intervention in the patients affected by mutations in the same region of G1006fs/49.
Interestingly, in vivo treatment with NS1643 in rabbit models of acquired long QT syndrome due to a dofetilide-induced KCHN2 inhibition shortened the QT interval in these animals, indicating that this drug therapy is actually able also in vivo to restore the KCHN2 inhibition. 33 We are aware of the potential limitations of this study. A more meaningful experimental approach would have been to analyse the K + currents described in this paper over cardiomyocytes derived

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.