Hormonal regulation of the human QT interval has long been assumed, but underlying mechanisms are poorly understood. Here, we report that estradiol shortens QTc intervals via classical receptor-mediated activation of KCNH2 trafficking. This mechanism could potentially be exploited as a therapeutic approach for patients at risk for arrhythmias associated with prolonged repolarization, although substantial additional cautious and detailed research is needed to ensure safety and efficacy before application.
Keywords: Repolarization, Hormones, Ion channels, Gender, Long-QT syndrome
Abstract
Background
Modulation of cardiac repolarization by sexual hormones is controversial and hormonal effects on ion channels remain largely unknown. In the present translational study, we therefore assessed the relationship between QTc duration and gonadal hormones and studied underlying mechanisms.
Methods and results
We measured hormone levels and QTc intervals in women during clomiphene stimulation for infertility and women before, during, and after pregnancy. Three heterozygous LQT-2 patients (KCNH2-p.Arg752Pro missense mutation) and two unaffected family members additionally were studied during their menstrual cycles. A comprehensive cellular and molecular analysis was done to identify the mechanisms of hormonal QT-interval regulation. High estradiol levels, but neither progesterone nor estradiol/progesterone ratio, inversely correlated with QTc. Consistent with clinical data, in vitro estradiol stimulation (60 pmol/L, 48 h) enhanced IKCNH2. This increase was mediated by estradiol receptor-α-dependent promotion of KCNH2-channel trafficking to the cell membrane. To study the underlying mechanism, we focused on heat-shock proteins. The heat-shock protein-90 (Hsp90) inhibitor geldanamycin abolished estradiol-induced increase in IKCNH2. Geldanamycin had no effect on KCNH2 transcription or translation; nor did it affect expression of estradiol receptors and chaperones. Estradiol enhanced the physical interaction of KCNH2-channel subunits with heat-shock proteins and augmented ion-channel trafficking to the membrane.
Conclusion
Elevated estradiol levels were associated with shorter QTc intervals in healthy women and female LQT-2 patients. Estradiol acts on KCNH2 channels via enhanced estradiol-receptor-α-mediated Hsp90 interaction, augments membrane trafficking and thereby increases repolarizing current. These results provide mechanistic insights into hormonal control of human ventricular repolarization and open novel therapeutic avenues.
See page 651 for the editorial comment on this article (doi:10.1093/eurheartj/ehv475)
Translational perspective.
Hormonal regulation of the human QT interval has long been assumed, but underlying mechanisms are poorly understood. Here, we report that estradiol shortens QTc intervals via classical receptor-mediated activation of KCNH2 trafficking. This mechanism could potentially be exploited as a therapeutic approach for patients at risk for arrhythmias associated with prolonged repolarization, although substantial additional cautious and detailed research is needed to ensure safety and efficacy before application.
Introduction
Hormonal influences on cardiac repolarization have been postulated since gender-related differences in QT intervals were noted.1 Generally, women have longer QTc intervals than men and there is QTc variability during menstrual cycle and pregnancy, suggesting intrinsic hormonal regulation in female hearts.
Recent experimental work in rabbits reported a potentially detrimental effect of oestrogen on the development of cardiac arrhythmias, while progesterone was found protective.2 Rabbits showed enhanced L-type calcium current with oestrogen and reductions with progesterone potentially contributing to this effect. In humans, potential cellular mechanisms regulating human cardiac repolarization by hormones have not yet been identified. Interestingly, data from patients with LQT-2 suggest protective effects of pregnancy-related hormones and increased arrhythmia susceptibility postpartum.3
Accordingly, we hypothesized that female sexual hormones alter QTc and assessed QTc in females exposed to defined hormonal changes. Additionally, we determined the role of female gonadal hormones in QTc modulation of female LQT-2 patients during the menstrual cycle and studied underlying cellular mechanisms.
Methods
Patient recruitment, electrocardiogram, and hormone measurements
Hormone concentrations [estradiol, progesterone, luteinizing hormone (LH), follicle-stimulating hormone (FSH)] were measured from peripheral blood samples and QT intervals (leads II and V5) from surface electrocardiograms (ECGs) in 21 women.
Eleven healthy females (non-LQTS) were studied during a course of clomiphene-citrate therapy for infertility according to standard gynecology procedures. Clomiphene treatment produces an almost exclusive (and supra-physiological) estradiol increase. QTc intervals and hormone levels were obtained at baseline and during clomiphene stimulation. Five additional healthy women (non-LQTS) who became pregnant after stimulation were followed throughout pregnancy. ECGs and hormone levels were determined before pregnancy, during first and third trimester and 1–6 weeks postpartum.
Moreover, three female LQT-2 patients (two symptomatic, one without cardiac events; all were heterozygous carriers of the trafficking-deficient KCNH2 mutation, c.2255G>C, p.Arg752Pro) and two genotype-negative female family members were recruited. All had active menstruation cycles and were without hormonal contraception. ECGs and blood samples were obtained at four visits during a menstrual cycle (day 2/3, 10/11, 12/13, and 21/22).
ECGs were recorded at ∼10.00 AM in all cases. The room was mildly dimmed and temperature comfortable. QT-intervals were determined in triplicate with the tangential method, by one reader (S.B.) blinded to the hormonal/gestational status. QTc values (according to the Bazett, Fridericia, and Framingham formulas) were calculated upon three consecutive intervals. The study protocol was approved by the ethics committee of Frankfurt University (Nr. 70/06). All participants provided written informed consent.
Experimental procedures
Cultured HEK 293B cells were used for in vitro experiments. A plasmid coding for KCNH2-R752P mutation was generated by site-directed mutagenesis. Protein biochemical studies, RNA experiments, electrophysiological recordings and immunoflurescent analyses were performed using standard techniques (see Supplementary material onlinefor details).
Statistical analysis
SPSS (Chicago, USA), Clampfit (Molecular Devices, Ismaning, Germany), and GraphPad Prism (GraphPad Software, San Diego, USA) were used for data analysis (see Supplementary material online for details). P < 0.05 was considered significant. Data are presented as mean ± SD or median values.
Results
QTc intervals during clomiphene stimulation and pregnancy
Clomiphene treatment significantly shorted QTc intervals in healthy women (mean 423 ± 6 vs. 390 ± 5 ms, P = 0.003, n = 11, Figure 1A and B right panel) and induced large increases in estradiol (mean 102 ± 64 vs. 1360 ± 388 pg/mL, P = 0.003) whereas progesterone levels remained unaffected (mean 0.8 ± 0.1 vs. 3.0 ± 1.6 ng/mL, P = 0.14). Accordingly, estradiol was the only hormone significantly associated with changes in QTc (P = 0.000001, Figure 1E), while progesterone was not significantly associated (P = 0.14, Supplementary material online, Figure 1B). Among pregnant women (N = 5, Figure 1C and D) a similar tendency could be observed for estradiol. During pregnancy there was a significant increase in heart rate (HR, mean HR 75 ± 5 b.p.m. at baseline; mean HR 81 ± 6 at third trimester; P = 0.036). QTc values during pregnancy were consistently shortened at higher estradiol levels (Figure 1D, Supplementary material online, Figure 1A). Analyses with three different QT correction formulae (Bazett, Fridericia, Framingham) provided consistent results. The association of estradiol remained significant if data from all study subjects (N = 16) were analysed collectively, even after Bonferroni's correction (P = 0.000). Progesterone, in contrast, had no significant effect on QTc in the overall population (P = 0.10).
Figure 1.
QTc modulation in response to stimulation and pregnancy. (A) Exemplary illustration of in vivo clomiphene effects on hormones and QTc (lead II) in a healthy female. (B) Median values of baseline and clomiphene-stimulated conditions (n = 11). (C) Exemplary illustration of hormone levels and QTc in a healthy female before, during, and after pregnancy. (D) Courses of estradiol and QTc of five females during pregnancies. Of note, QTc shortening coincided with increased estradiol levels. Lines were added for illustration, and symbols represent probands. E, pooled data (pregnant and stimulated; n = 16). Higher estradiol concentrations were associated with QTc shortening (P = 0.000001). E2, estradiol; prog., progesterone; 1st, 1st trimester; 3rd, 3rd trimester; pp, postpartum.
Effect of estradiol stimulation on KCNH2 current
To uncover potential mechanisms underlying the estradiol-induced QTc-shortening, experiments in cells transiently transfected with KCNH2 and oestrogen receptor-α (ERα) were performed. In accordance with the in vivo data, addition of estradiol (60 pmol/L, 48 h) roughly doubled the amplitude of the delayed rectifier current (IKCNH2) with no effect on its biophysical properties (n = 8–10 cells, P = 0.03). The effect of estradiol was concentration dependent and plateaued at concentrations >60 pmol/L (Supplementary material online, Figure 2). Half-activation voltages of IKCNH2 were −17 ± 3 vs. −18 ± 2 mV (P = 0.79) and slope factors: 8.1 ± 0.7 vs. 8.3 ± 1.1 mV (P = 0.87) from cells without or with estradiol stimulation. Ionic current kinetics remained unaffected. After depolarization to +50 mV, deactivation time-constants were (τ1) 1241 ± 254 and (τ2) 167 ± 37 ms without estradiol vs. (τ1) 964 ± 168 and (τ2) 143 ± 21 ms with estradiol stimulation (P = 0.33 and P = 0.52).
The effect of estradiol was dependent on binding of the hormone to ERα. Estradiol-induced increases in IKCNH2 were prevented by the oestrogen-receptor antagonist tamoxifen (200 nmol/L, n = 8, Figure 2A and B) and no direct membrane delimited effect could be observed during acute estradiol exposure (Figure 2C). There was also no effect of estradiol on IKCNH2 when ERα transfection was omitted or when ERα was replaced by ERβ (Supplementary material online, Figure 3A and B). Finally, estradiol had no effect on currents recorded in HEK cells expressing only ERα without KCNH2 (Supplementary material online, Figure 4).
Figure 2.
Long-term estradiol enhances IKCNH2 through oestrogen receptor-α-dependent pathway. (A) Representative currents of HEK cells transfected with KCNH2 and oestrogen receptor-α cDNA (protocol in inset). IKCNH2 increased after estradiol stimulation (60 pmol/L, 48 h). This was abolished by co-incubation with tamoxifen (200 nmol/L). Mean ± SD values for depolarization (‘step’) and repolarization (‘tail’)-induced currents are shown (B, n = 4–6 cells). (C) Acute estradiol exposure did not affect KCNH2 currents in identically transfected cells (mean ± SD, n = 10 each). E2, estradiol; ER, oestrogen receptor; tam, tamoxifen; TP, test potential; pA, pico Ampere; pF, pico Farad.
Estradiol stimulation activates ERα-mediated KCNH2 membrane trafficking
Estradiol did not alter the overall cellular protein amount of KCNH2 (Figure 3A). Membrane analysis revealed that estradiol increased the abundance of KCNH2 protein at the plasma membrane (Figure 3B). This effect was observed only in cells expressing ERα and could not be mimicked by transfection of ERβ (Figure 3C). The response was specific for KCNH2, since under identical conditions KCNQ1 localization was unaffected by estradiol (Figure 3D).
Figure 3.
Estradiol does not affect KCNH2 synthesis but enhances trafficking. Blots are representative of 3–6 experiments with input and pharmacological interventions outlined above each panel, bar graphs at the side represent mean ± SD arbitrary optical densities (OD) for individual lanes. (A) Crude membrane preparations illustrating no change in overall KCNH2 protein expression with ERα co-transfection or estradiol stimulation (60 pmol/L, 48 h). Calnexin expression indicated equal protein loading. (B) Plasma membrane preparations of cells transfected with KCNH2 and ERα showed greater KCNH2 plasma membrane expression with ERα co-transfection and estradiol stimulation (60 pmol/L, 48 h, lane 4, P = 0.047 vs. all others). This effect was abolished by co-incubation with tamoxifen (200 nmol/L). Arrows indicate fully- and core-glycosylated KCNH2 bands. Membrane-marker spectrin was present in the plasma membrane fraction, while calnexin (endoplasmic reticulum marker) was absent. (C) Plasma membrane preparation with KCNH2 and ERβ showed no difference in KCNH2 expression suggesting specificity of the ERα pathway. (D) No effect of estradiol stimulation for plasma membrane preparations from ERα and KCNQ1 co-transfected cells. NT, non-transfected; Q1, KCNQ1; kD, kilo Dalton; AU, arbitrary units. *ANOVA. Other abbreviations are as above.
Confocal microscopy revealed that in the absence of ERα transfection KCNH2 ion-channel protein was predominantly localized intracellularly (Figure 4A) with or without estradiol treatment. After ERα transfection without estradiol application, the receptor resided mainly inside the cytoplasm. Upon estradiol stimulation, ERα translocated to the nucleus and KCNH2 redistributed to the plasma membrane. In contrast, ERβ failed to increase KCNH2 plasma membrane representation in analogous experiments despite successful activation of the receptor as evidenced by nuclear translocation (Figure 4B). Moreover, estradiol failed to induce the translocation of KCNQ1 in HEK cells expressing ERα (Figure 4C) suggesting that the observed effect is specific for KCNH2 channels.
Figure 4.
Enhanced KCNH2 membrane trafficking in response to ERα-mediated estradiol stimulation. (A) (left to right) Cells transfected with KCNH2 (green) alone; with estradiol stimulation (60 pmol/L); KCNH2 and ERα (red); additional estradiol stimulation. Yellow indicates co-localization. Estradiol had no effect on KCNH2 localization without the receptor. Stimulation of co-transfected cells (KCNH2 + ERα) led to nuclear translocation and more prominent KCNH2 expression at the plasma membrane. (B) Analogous experiments with KCNH2 and ERβ. Estradiol led to receptor translocation into the nucleus without effect on KCNH2 membrane localization. (C) Experiments with KCNQ1 and ERα similarly show unchanged KCNQ1 distribution. Images are representative of at least 4–6 experiments. The insets in (A–C) illustrate the optical density along the white bar across the cells. (D) Estradiol stimulation significantly increased the plasma membrane/endoplasmic reticulum ratio in stimulated ERα co-transfected cells. There was no difference for KCNH2 + ERβ or KCNQ1 + ERα (mean ± SD). Bars = 5 µm. End-ret, endoplasmic reticulum. Other abbreviations are as above.
Estradiol improves KCNH2 trafficking through chaperone interaction
KCNH2 mRNA expression remained unchanged in response to estradiol suggesting a post-transcription effect (Figure 5A). Heat-shock cognate protein 70 (Hsc70) and heat-shock protein 90 (Hsp90) are involved in folding during KCNH2 maturation and contribute to its trafficking.4 Estradiol did not affect overall Hsc70 and Hsp90 protein or transcription levels (Figure 5B and C). There was also no change in the expression other KCNH2-associated chaperones in response to estradiol (STIP1, DNAJA1, Grp94, FKBP38, Figure 5B). However, stimulation with the hormone resulted in an enhanced physical interaction of Hsc70 and Hsp90 with KCNH2 as revealed by co-immunoprecipitation experiments (Figure 5D). Collectively, these observations suggest that estradiol enhances membrane trafficking of KCNH2 through a quantitatively greater association of the channel with chaperone complexes.
Figure 5.
Enhanced chaperone/channel interaction with estradiol stimulation. (A) Estradiol stimulation did not alter mRNA transcription of KCNH2 or ERα. (B) Western blots for various KCNH2 chaperones indicated unchanged expression levels under various transfection conditions (indicated above the blots) with or without estradiol stimulation. The open arrow indicates the specific (62 kDa), the black arrow the unspecific band of STIP-1. (C) mRNA levels of Hsc70 and Hsp90 remained unchanged in various transfection constellations with or without estradiol stimulation. (D) Co-immunoprecipitation of Hsc70 and Hsp90 with KCNH2 illustrated enhanced precipitation of both chaperones with KCNH2 during estradiol stimulation, and bar graphs illustrate mean ± SD data (N = 4). Grp94 did not interact with KCNH2 protein. Hsc70, Heat-shock cognate protein 70kD; Hsp90, Heat-shock protein 90kD; STIP1, Hsc70/Hsp90 organizing protein; DNAJA1, Heat-shock protein 40kD; Grp94, 94kD glucose regulated protein; FKBP38, 38kD FK506-binding protein. *ANOVA. Other abbreviations are as above.
Finally, the highly selective Hsp90 inhibitor geldanamycin (1 µg/mL) not only decreased KCNH2 membrane localization and IKCNH2 currents, it also blocked the stimulatory effect of estradiol (Figure 6).
Figure 6.
Inhibition of Hsp90 function abrogates enhanced KCNH2 trafficking and ionic current. (A) Plasma membrane preparations of cells transfected with KCNH2 and ERα and exposed to estradiol stimulation (60 pmol/L, 48 h) in the presence or absence of the Hsp90 inhibitor geldanamycin (1 µg/mL). The level of KCNH2 membrane expression was greatly reduced by geldanamycin exposure. (B) KCNH2 currents were similarly reduced with geldanamycin (mean ± SD, n = 10 and 3 cells, respectively). Gelda, geldanamycin. Other abbreviations are as above.
Modulation of QTc intervals by gonadal hormones in female LQT-2 patients
In order to study whether the mechanism was also operative in LQT-2 patients, females from a large German LQT-2 kindred (KCNH2 mutation c.2255G>C; p.Arg752Pro, Figure 7A) were studied. Exemplary ECGs from one genotype-positive and one genotype-negative family member during the menstrual cycle are shown in Figure 7B and C.
Figure 7.
QTc changes during menstrual cycle of female LQT-2 patients. (A) Pedigree of the KCNH2-R752P family. Arrows indicate symptomatic index patients who survived cardiac arrest. Grey are genotype-positives, and hatched lines are unknown genotypes. Asterisks indicate women participating in our study. (B and C) Examples of QTc measurements and respective hormone levels (including estradiol/progesterone ratio) for one genotype-positive (IV-5) and one genotype-negative family member (IV-10) during the menstrual cycle. (D) QTc values from genotype-positive (n = 3) and genotype-negative (n = 2) individuals stratified according to their median of hormone concentrations. High estradiol was associated with abbreviated QTc, whereas no QTc difference was apparent in relation to progesterone, estradiol/progesterone ratio, and serum potassium. Bottom and top of the box are 25th and 75th percentile, the band represents the median, and whiskers illustrate 95% confidence intervals. Broken, horizontal lines beginning at the median of the higher strata have been added for better visual discrimination. Abbreviations are as above.
On cycle day 2/3 (menstrual phase), estradiol levels were low (20 pg/mL each) and on day 21/22 mean estradiol levels increased (148 and 84 pg/mL, respectively). Simultaneously, the QTc shortened from 488 to 454 ms (genotype positive) and 399 to 332 ms (genotype negative). Shorter QTc values were again associated with higher estradiol levels (Figure 7D). Spearman correlation coefficient (r) was −0.483 for the five patients (P = 0.002). In analogy to our findings in healthy women, changes in progesterone and estradiol/progesterone ratio were unrelated to the QTc for LQTS women (Figure 7D).
Estradiol fails to restore membrane trafficking of mutant KCNH2 channels
KCNH2-R752P is severely trafficking deficient and did not confer IKCNH2 upon heterologous expression (Figure 8). Immunocytochemical analysis of KCNH2-R752P demonstrated the absence of plasma membrane staining and accumulation in the endoplasmic reticulum (Supplementary material online, Figure 5).
Figure 8.
Enhanced trafficking depends on functional KCNH2 channels. (A) Currents recorded from cells transfected with the trafficking-deficient KCNH2-R752P displayed prominently distorted biophysical behaviour with a rapidly activating and inactivating current that did not give rise to repolarization induced ‘tail’ current. (B) Current–voltage relations for cells expressing KCNH2-R752P alone or in combination with wild-type KCNH2 cDNA. An increase in IKCNH2 in response to estradiol occurred if wild-type and mutant were co-transfected. (C) Plasma membrane preparations of KCNH2 and KCNH2-R752P indicate almost absent KCNH2-R752P at the plasma membrane and no rescue of mutant protein with estradiol (n = 3). (D) Quantifications of mRNA illustrate unchanged KCNH2-R752P and oestrogen receptor expression. (E) Confocal images confirm lack of enhanced plasma membrane trafficking of KCNH2-R752P with estradiol. Bars = 5 µm. Images are representative of at least 4–6 experiments. (F) Co-immunoprecipitation of KCNH2-R752P with Hsc70 and Hsp90 indicated an unaltered level of interaction. Mean ± SD of n = 12 and 13 cells, respectively. *ANOVA for time series. Abbreviations are as above.
Co-transfection of KCNH2-R752P with ERα and estradiol stimulation failed to restore this trafficking defect (Figure 8A). In contrast, estradiol stimulation of cells co-transfected with wild-type and mutant KCNH2 (to mimic heterozygosity) increased ion currents (Figure 8B) suggesting an effect on trafficking competent wild-type channels. Overall, there was less KCNH2-R752P than wild-type protein detected in transfected cells suggesting greater protein degradation (Figure 8C). mRNA expression of the mutant allele was similar between groups (Figure 8D). Confocal images of cells transfected with KCNH2-R752P and ERα similarly showed no estradiol-induced increase in KCNH2-R752P membrane localization (Figure 8E). Co-IP of KCNH2-R752P with Hsc70 and Hsp90 demonstrated no enhanced interaction in the context of estradiol stimulation (Figure 8F).
In summary, our data suggest that only trafficking-competent, wild-type KCNH2 channels, but not trafficking-deficient mutant channels such as KCNH2-R752P, can be promoted by estradiol.
Discussion
Main findings
The present translational study provides evidence for an estradiol-dependent acceleration of ventricular repolarization in women through enhanced KCNH2-channel trafficking. This modulation was observed in healthy females and in a small series of LQT-2 patients during the menstrual cycle. We prospectively demonstrated shortening of the QTc interval associated with estradiol elevation after clomiphene stimulation for in vitro fertilization and during pregnancy. Clomiphene stimulation leads to an almost exclusive (and supra-physiological) estradiol elevation and represents a human in vivo validation experiment for our in vitro observations. At the cellular level, the rapid delayed rectifier current was specifically enhanced through increased KCNH2 membrane trafficking caused by greater interaction with Hsp90 chaperone proteins.
Hormonal modulation of the human QT-interval
The QTc interval is longer in women than in men,1,5 and women are at increased risk of drug-induced QT-interval prolongation and torsade-de-pointes tachycardia.6,7 Systematic studies of hormonal effects (e.g. menstrual cycle) are sparse. In one study, healthy women exhibited shorter QTc intervals in the luteal phase (438 ± 16 ms) compared with the menstrual (446 ± 15 ms) and the follicular phases (444 ± 13 ms).8 However, these cycle-dependent differences were only present after double autonomic blockade and absent under control conditions. Studies of hormonal regulation of the QT-interval have often relied on pooled data, potentially flawed by large inter-individual inherent variability in QT-intervals.9,10 To address this large variability, we dichotomized measurements into QTc obtained at ‘high’ vs. ‘low’ estradiol concentrations to overcome this confounder.
While our study does not provide information regarding a potential underlying basis of gender-specific differences, it provides evidence for regulation of cardiac repolarization in female hearts. A similar effect on QTc intervals might be applicable to male hearts in a potentially therapeutic fashion. However, such hypothesis will need evaluation in a prospective trial.
Animal studies of hormonal effects
A large number of animal studies of hormonal regulatory mechanisms of cardiac repolarization have been performed, but in many cases results were inconsistent—potentially owing to inter-species differences.9 Furthermore, most mechanistic studies of estradiol have used acute exposure rather than prolonged, chronic stimulation with sexual hormones.11 While classical activation of steroid receptors by hormones leads to nuclear translocation and transcription activation, acute exposure appears less suitable for studies of expression regulation,12 although an electrophysiological effect of acute application might contribute to an overall cardiac substrate.
A previous study reported an increased risk of sudden cardiac death in postpartum transgenic rabbits expressing an LQT-2 mutation,2 apparently resembling events occurring postpartum in LQT-2 women.13 Animals were treated with estradiol pellets and had serum estradiol levels around 100–200 pg/mL, comparable with human levels during the menstrual cycle. In that rabbit study an increased number of ventricular arrhythmias occurred, potentially mediated by an estradiol-induced increase in L-type calcium current. While we did not study in vitro effects on calcium channels, the result of increased oestrogen in humans as shown here is QTc shortening rather than prolongation as observed in the rabbit study.
Cellular mechanism
The KCNH2 channel traffics relatively inefficiently by itself14 and depends on chaperone proteins for proper folding4 and trafficking to the plasma membrane.15 Our study provided evidence that enhanced plasma membrane localization of KCNH2 was associated with quantitatively greater Hsc70/Hsp90 interaction despite unchanged levels of chaperone or ion-channel expression. A similar type of non-genomic oestrogen action on Hsp90 function has previously been described for neuronal nitric oxide synthase (nNOS) in coronary artery smooth muscle cells.16 Estradiol time-dependently enhanced the interaction between nNOS and Hsp90 as evidenced by co-immunoprecipitation of the two partners. Our work suggests that enhanced formation of chaperone/channel complexes may similarly occur in response to estradiol stimulation, enhancing channel trafficking and/or membrane stability via improved folding and trafficking. The estradiol-mediated effect required trafficking-competent channels, as evidenced by results with the trafficking-deficient mutant KCNH2-R752P. Studies in different types of KCNH2 mutations (trafficking deficient vs. mutations with biophysical effects) should be performed in the future.
Trafficking deficient LQT-2 mutants interact with Hsc70 and Hsp90 when retained in the endoplasmic reticulum.4 The recovery of channel trafficking was coupled to the dissociation of channel–chaperone complexes, suggesting that exit from the endoplasmic reticulum was linked to the dissociation of KCNH2 from this complex.4 Oestrogen may therefore enhance trafficking of KCNH2 from the endoplasmic reticulum to the Golgi apparatus, ultimately resulting in greater movement of functional channels to the membrane.
Implications for human long-QT syndrome
Long-QT syndrome (LQTS) is characterized by prolonged QTc intervals with increased risk of cardiac arrhythmia and sudden death. QTc intervals are considered prolonged if >460 ms in women and >450 ms in men.17 Women with genetically defined LQT-2 subtype (KCNH2 mutations) have a mildly reduced risk of cardiac events during pregnancy but increased risk during postpartum.18–20 Emotional stress and frequent sleep disruption during postpartum have been suggested as explanation for this observation.21 While results of this study cannot explain changes that occur during postpartum, lower rates of cardiac events during pregnancy are well compatible with our data.20
Data regarding cardiac events during pregnancy in LQTS patients are controversial. While Rashba et al. showed evidence for an increased number of cardiac events during pregnancy and postpartum, Seth et al. documented a trend towards reduced risk in LQT-1 and LQT-2 patients during pregnancy.18,20 There were more events postpartum, particularly in the LQT-2 cohort of the study. In another study of 27 female LQT-1 patients who had cardiac symptoms before pregnancy, only one patient had an event during pregnancy.21
Limitations
The use of a heterologous expression model allowed us to dissect pathways regulating IKr in detail. This model has clear limitations with regard to human physiological relevance. However, the use of animal models would have presented other limitations in terms of physiologically relevant oestrogens, types of receptors, and ion channels. The use of cardiomyocytes derived from induced human pluripotential stem cells might be helpful, but there are issues about phenotypic fidelity to human ventricular cardiomyocytes and the detailed characterization and studies that would be necessary are out of the scope of the present manuscript. There is diverse heterogeneous data—mostly regarding acute hormonal regulation—in various animal species regarding the role of a variety of ionic currents and other factors related to arrhythmogenicity. These may additionally be modulated by autonomic nervous system in vivo. Further work is needed to examine the complex broader context.
Based on our data estradiol effects on repolarization are unlikely to explain the characteristically greater QTc intervals in women vs. men. Beyond hormonal regulation of ionic currents, different expression patterns of ionic currents between genders need to be taken into account.22 Potentially, androgens have additional effects on repolarization. This issue should be addressed in future studies.
Novelty and therapeutic implications
This work provides novel insights into the regulation of human repolarization by sex hormones. Our data implicate higher estradiol levels in clinically relevant QTc shortening. We identified estradiol-increased KCNH2 membrane trafficking dependent on enhanced chaperone interaction as a potential underlying mechanism. Direct extrapolation to the clinic should be cautious. If our results are confirmed by further work in complementary studies, treatment with estradiol may warrant consideration as a therapeutic concept for patients at risk of arrhythmia due to prolonged repolarization.
Supplementary material
Supplementary material is available at European Heart Journal online.
Authors’ contributions
L.A., Z.G., S.N., and J.R.E. performed statistical analysis. R.P.B., S.H.H., T.K., E.S.-B., S.N., and J.R.E. handled funding and supervision. L.A., S.B., P.V., P.B., C.F., L.X., Z.G., I.T., S.K., I.W., S.Z., B.S. acquired the data. L.A., S.B., P.V., R.P.B., S.H.H., T.K., E.S.-B., S.N., and J.R.E. conceived and designed the research. L.A., E.S.-B., and J.R.E. drafted the manuscript. R.P.B., S.H.H., T.K., and S.N. made critical revision of the manuscript for key intellectual content.
Funding
J.R.E. received funding from ‘Deutsche Herzstiftung’, E.S.-B. from the Fondation Leducq, Paris, the IZKF of the Medical Faculty of the University of Münster, and S.N. from the Canadian Institutes of Health Research (MOP6892) and the Quebec Heart and Stroke Foundation.
Conflict of interest: none declared.
Supplementary Material
Acknowledgements
We thank Sabine Harenkamp, Christin Lössl, Ellen Schulze-Bahr, Johanna Ising, and Martina Raetz for technical assistance, Maya Marmabachi for support with molecular cloning, and Armin Heisel, MD and Hendrik Bonnemeier, MD for supporting patient examinations. David G. Monroe, PhD, provided oestrogen-receptor-α and -β cDNA.
References
- 1. Bazett HC. An analysis of the time relationship of electrocardiograms. Heart 1920;7:353–370. [Google Scholar]
- 2. Odening KE, Choi BR, Liu GX, Hartmann K, Ziv O, Chaves L, Schofield L, Centracchio J, Zehender M, Peng X, Brunner M, Koren G. Estradiol promotes sudden cardiac death in transgenic long QT type 2 rabbits while progesterone is protective. Heart Rhythm 2012;9:823–832. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Sauer AJ, Moss AJ, McNitt S, Peterson DR, Zareba W, Robinson JL, Qi M, Goldenberg I, Hobbs JB, Ackerman MJ, Benhorin J, Hall WJ, Kaufman ES, Locati EH, Napolitano C, Priori SG, Schwartz PJ, Towbin JA, Vincent GM, Zhang L. Long QT syndrome in adults. J Am Coll Cardiol 2007;49:329–337. [DOI] [PubMed] [Google Scholar]
- 4. Ficker E, Dennis AT, Wang L, Brown AM. Role of the cytosolic chaperones Hsp70 and Hsp90 in maturation of the cardiac potassium channel HERG. Circ Res 2003;92:e87–100. [DOI] [PubMed] [Google Scholar]
- 5. Pham TV, Rosen MR. Sex, hormones, and repolarization. Cardiovasc Res 2002;53:740–751. [DOI] [PubMed] [Google Scholar]
- 6. Lehmann MH, Hardy S, Archibald D, Quart B, MacNeil DJ. Sex difference in risk of torsade de pointes with d,l-sotalol. Circulation 1996;94:2535–2541. [DOI] [PubMed] [Google Scholar]
- 7. Lehmann MH, Hardy S, Archibald D, MacNeil DJ. JTc prolongation with d,l-sotalol in women versus men. Am J Cardiol 1999;83:354–359. [DOI] [PubMed] [Google Scholar]
- 8. Burke JH, Ehlert FA, Kruse JT, Parker MA, Goldberger JJ, Kadish AH. Gender-specific differences in the QT interval and the effect of autonomic tone and menstrual cycle in healthy adults. Am J Cardiol 1997;79:178–181. [DOI] [PubMed] [Google Scholar]
- 9. Yang PC, Clancy CE. Effects of sex hormones on cardiac repolarization. J Cardiovasc Pharmacol 2010;56:123–129. [DOI] [PubMed] [Google Scholar]
- 10. Zhang Y, Ouyang P, Post WS, Dalal D, Vaidya D, Blasco-Colmenares E, Soliman EZ, Tomaselli GF, Guallar E. Sex-steroid hormones and electrocardiographic QT-interval duration: findings from the third National Health and Nutrition Examination Survey and the Multi-Ethnic Study of Atherosclerosis. Am J Epidemiol 2011;174:403–411. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Tanabe S, Hata T, Hiraoka M. Effects of estrogen on action potential and membrane currents in guinea pig ventricular myocytes. Am J Physiol 1999;277(Pt 2):H826–HH33. [DOI] [PubMed] [Google Scholar]
- 12. Dahlman-Wright K, Cavailles V, Fuqua SA, Jordan VC, Katzenellenbogen JA, Korach KS, Maggi A, Muramatsu M, Parker MG, Gustafsson JA. International Union of Pharmacology. LXIV. Estrogen receptors. Pharmacol Rev 2006;58:773–781. [DOI] [PubMed] [Google Scholar]
- 13. Brunner M, Peng X, Liu GX, Ren XQ, Ziv O, Choi BR, Mathur R, Hajjiri M, Odening KE, Steinberg E, Folco EJ, Pringa E, Centracchio J, Macharzina RR, Donahay T, Schofield L, Rana N, Kirk M, Mitchell GF, Poppas A, Zehender M, Koren G. Mechanisms of cardiac arrhythmias and sudden death in transgenic rabbits with long QT syndrome. J Clin Invest 2008;118:2246–2259. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Zhou Z, Gong Q, Epstein ML, January CT. HERG channel dysfunction in human long QT syndrome. Intracellular transport and functional defects. J Biol Chem 1998;273:21061–6. [DOI] [PubMed] [Google Scholar]
- 15. Walker VE, Atanasiu R, Lam H, Shrier A. Co-chaperone FKBP38 promotes HERG trafficking. J Biol Chem 2007;282:23509–23516. [DOI] [PubMed] [Google Scholar]
- 16. Han G, Ma H, Chintala R, Fulton DJ, Barman SA, White RE. Essential role of the 90-kilodalton heat shock protein in mediating nongenomic estrogen signaling in coronary artery smooth muscle. J Pharmacol Exp Ther 2009;329:850–855. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Goldenberg I, Moss AJ, Zareba W. QT interval: how to measure it and what is “normal”. J Cardiovasc Electrophysiol 2006;17:333–336. [DOI] [PubMed] [Google Scholar]
- 18. Rashba EJ, Zareba W, Moss AJ, Hall WJ, Robinson J, Locati EH, Schwartz PJ, Andrews M. Influence of pregnancy on the risk for cardiac events in patients with hereditary long QT syndrome. LQTS Investigators. Circulation 1998;97:451–456. [DOI] [PubMed] [Google Scholar]
- 19. Khositseth A, Tester DJ, Will ML, Bell CM, Ackerman MJ. Identification of a common genetic substrate underlying postpartum cardiac events in congenital long QT syndrome. Heart Rhythm 2004;1:60–64. [DOI] [PubMed] [Google Scholar]
- 20. Seth R, Moss AJ, McNitt S, Zareba W, Andrews ML, Qi M, Robinson JL, Goldenberg I, Ackerman MJ, Benhorin J, Kaufman ES, Locati EH, Napolitano C, Priori SG, Schwartz PJ, Towbin JA, Vincent GM, Zhang L. Long QT syndrome and pregnancy. J Am Coll Cardiol 2007;49:1092–1098. [DOI] [PubMed] [Google Scholar]
- 21. Heradien MJ, Goosen A, Crotti L, Durrheim G, Corfield V, Brink PA, Schwartz PJ. Does pregnancy increase cardiac risk for LQT1 patients with the KCNQ1-A341V mutation? J Am Coll Cardiol 2006;48:1410–1415. [DOI] [PubMed] [Google Scholar]
- 22. Gaborit N, Varro A, Le Bouter S, Szuts V, Escande D, Nattel S, Demolombe S. Gender-related differences in ion-channel and transporter subunit expression in non-diseased human hearts. J Mol Cell Cardiol 2010;49:639–646. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.