Estrogen Contributes to Gender Differences in Mouse Ventricular Repolarization

Abstract

Rationale

Fast-transient outward K⁺ (I_to,f) and ultra-rapid delayed rectifier K⁺ currents (I_Kur or I_K,slow) contribute to mouse cardiac repolarization. Gender studies on these currents have reported conflicting results.

Objective

One key missing piece information in these studies is the animals’ estral stage. We decided to revisit gender-related differences in K⁺ currents, taking into consideration the females’ estral stage.

Methods and Results

We hypothesized that changes in estrogen levels during the estral cycle could play a role in determining the densities of K⁺ currents underlying ventricular repolarization. Peak total K⁺ current (I_K,total) densities (pA/pF, at +40 mV) were much higher in males (48.6±3.0) than in females at estrus (27.2±2.3) but not at diestrus-2 (39.1±3.4). Underlying this change, I_to,f and I_K,slow were lower in females at estrus vs males and diestrus-2 (I_K,slow: male 21.9±1.8, estrus 14.6±0.6, diestrus-2 20.3±1.4; I_to,f: male 26.8±1.9, estrus 14.9±1.6, diestrus-2 22.1±2.1). The lower I_K,slow in estrus was only due to I_K,slow1 reduction without changes of I_K,slow2. Estrogen treatment of ovariectomized mice decreased I_K,total (46.4±3.0 to 28.4±1.6), I_to,f (26.6±1.6 to 12.8±1.0) and I_K,slow (22.2±1.6 to 17.2±1.4). Transcript levels of Kv4.3 and Kv1.5 (underlying I_to,f and I_K,slow, respectively) were lower in estrus vs. diestrus-2 and male. In ovariectomized mice, estrogen treatment resulted in downregulation of Kv4.3 and Kv1.5, but not Kv4.2, KChIP2 and Kv2.1 transcripts. K⁺ current reduction in high estrogenic conditions were associated with prolongation of the action potential duration and corrected QT interval.

Conclusion

Downregulation of Kv4.3 and Kv1.5 transcripts by estrogen are one mechanism defining gender-related differences in mouse ventricular repolarization.

INTRODUCTION

Gender-based differences in cardiac repolarization have been previously shown in humans, with females having longer action potentials and higher susceptibility to arrhythmias than same-aged males¹. In mouse however, studies on gender-related differences in right and left ventricular myocytes have yielded different results when examining fast transient outward K⁺ current (I_to,f) and ultra rapid delayed rectifier K⁺ current (I_Kur or I_K,slow), which play an important role in cardiac repolarization^{2, 3}. Reduction of I_K,slow but not I_to,f⁴, reduction of I_to,f⁵, or no change at all⁶ in females vs. males were reported. One common feature in all these studies was the lack of explicit report of the stage of females’ estral cycle. During the estral cycle there are changes in plasma hormone levels, especially of estrogen (E2)⁷, that have been shown to downregulate cardiac Kv4.3 transcript ⁸ which is one of the molecular correlates of I_to,f. In rats, the estral cycle lasts ~4 days with an estrogen peak at proestrus, hours before estrus; hence, the estrus stage reflects the effects of the high E2 surge at proestrus. These effects decay towards diestrus-2, when the animal has been exposed to low E2 levels⁷ for the longest time (~3 days). Here, we revisited in detail gender differences and E2-mediated changes in K⁺ repolarizing currents by initially measuring E2 levels during the mouse estral cycle, and correlated E2 levels with outward K⁺ currents, action potential duration (APD) and corrected QT (QTc) interval. We measured total outward K⁺ current (I_K,total) densities as well as their main components I_K,slow and I_to,f in males and in females at estrus (under the influence of high estrogen of the preceding proestrus), diestrus-2 (under a prolonged exposure to low estrogen levels), and in E2-treated ovariectomized (OVX) mice. Our data revealed that I_K,total, I_to,f and I_K,slow densities were lower in animals under the influence of high E2 (estrus and E2-treated OVX) when compared with low E2 levels (diestrus-2, males and OVX sham). Lower K⁺ current densities in high estrogenic conditions were associated with downregulation of Kv4.3 and Kv1.5 transcript levels, providing a molecular correlate for the prolongation of both APD and QTc interval in female mice under the influence of high estrogen.

METHODS

2.1. Animals and hormonal treatment

Male, female and ovariectomized (OVX) mice were used (C57BL/6). Protocols received institutional review and committee approval. The female estral cycle and the treatment of OVX mice with E2 are described in the online data supplement.

2.2. Chemicals

Tetrodotoxin, tetraethylammonium (TEA) and 4-aminopyridine (4-AP) were from Sigma and estrogen pellets were from Innovative Research of America.

2.3. Isolation of Ventricular Myocytes, electrophysiology and Electrocardiograms

Whole-cell, voltage-, and current-clamp recordings were obtained at room temperature from right ventricular myocytes. Cell isolation, recording solutions, the pulse protocols and the details of the ECG recordings and analysis are provided in the online data supplement.

2.4. Real-Time PCR

Total RNA from mouse heart right ventricle was isolated using Trizol (Invitrogen) and was reverse transcribed with gene specific primers using Omniscript RT kit (Qiagen). See online data supplement for details.

2.5. Statistics

ANOVA followed by Tukey-Kramer Multiple-Comparison test was used for statistical analysis. Repeated Measures ANOVA was used for I–V curves. Probability values <0.05 were considered statistically significant. Values are mean±SE.

RESULTS

Plasma E2 concentration is highest in females at proestrus which is mimicked in E2-treated OVX mice

Similar to rats⁷, female mice have an estrous cycle of ~four days; proestrus, estrus, metaestrus/diestrus-1 and diestrus-2. E2 plasma concentrations [E2] were measured in male and female mice (proestrus, estrus and diestrus-2 identified from vaginal smears) and in OVX mice control (sham) and E2 treated (Fig. 1). Similarly to rats⁷, [E2] is maximal in the proestrus stage, reaching ~60 pg/ml, and is reduced to ~20 pg/ml at estrus and diestrus-2. [E2] in male mice is ~12 pg/ml, which is significantly lower than [E2] in dietrus-2 (P<0.001). As expected, ovariectomy decreased [E2] to ~7 pg/ml which increased to ~120 pg/ml after E2 treatment. Note that although [E2] is low in estrus and diestrus-2, the estrus stage is primed by a preceding (0.5–1 day) surge of estrogen in proestrus, while diestrus-2 is marked by a prolonged exposure (~3 days) to low estrogen levels. Therefore, in this work we used mice under the influence of high estrogen (estrus and E2-treated OVX) and low estrogen (diestrus-2, OVX sham, and males).

Figure 1. Plasma [E2] in females and males.

[E2] measurements were performed in females at different stages of the estral cycle, in OVX placebo treated (sham) or E2-treated, and in male mice. Values are (pg/ml): males 12.4±1.3 (n=10); proestrus 60±9 (n=5), estrus 17.1±2.5 (n=17), diestrus-2 19.6±1.5 (n=14), OVX sham 7.30±0.9 (n=6) and E2-treated 120.7±12.9 (n=3). Significantly different (+) proestrus vs. male (P<0.05), (#) proestrus vs. diestrus-2 (P<0.05), (&) proestrus vs. estrus (P<0.05), (*) male vs. diestrus-2 (P<0.001), (@) diestrus-2 vs. OVX-sham (P<0.001), ($) estrus vs. OVX-sham (p,0.001) and (^) E2-treated vs. OVX-sham (P<0.05).

Gender and estral cycle determine outward K⁺ current densities

To test our hypothesis that estrogen levels determine outward K⁺ current densities, I_K,total was recorded in right ventricular myocytes isolated from males, estrus and diestrus-2 (Fig. 2A). While inward rectifier K⁺ current (I_Kir) remained constant, I_K,total at potentials higher than −10 mV in male mice and in diestrus-2 was significantly higher than in estrus (Fig. 2B). The steady-state current (I_ss) at the end of the pulse is somewhat smaller in estrus than in males, but similar to diestrus-2 (Fig. 2C). For comparative purposes, the underlying components of I_K,total were isolated by two methods: i) based on exponential fitting⁶ (Fig. 2D) and ii) based on short depolarizing prepulses as to inactivate I_to,f⁴ (Fig. 3). Both methods yielded virtually identical results (Table-1). The amplitudes and decay of I_K,total are well described by the sum of fast (~100 ms) and slow (~1 s) exponential components (corresponding to I_to,f and I_K,slow) and a steady-state current, I_ss (Fig. 2D). Using this method, Fig. 2E demonstrates gender differences in I_to,f, I_K,slow and I_ss.

Figure 2. Gender differences in I_K,total, I_to,f and I_K,slow: current separation by exponential fitting.

A , I_K,_total recorded from isolated right ventricular myocytes. Inset: Pulse protocol with 10 mV increments. Traces at right are at higher time gain. B, Peak I_K,_total-voltage curve in male (○), estrus (●) and diestrus-2 (▲). C, I_K,_total, I_ss and I_Kir for male (M), estrus (E), and diestrus-2 (D2). D, Separation of I_to,f and I_K,slow with exponential fitting. I_ss, baseline current ⁶, ⁸. E, I_to,f, I_K,slow, and I_ss for male (M), estrus (E), and diestrus-2 (D2). I_total, I_to,f and I_K,slow were measured at the peak at +40 mV; I_ss was at steady-state at +40 mV; and I_Kir was at the end of a 100 ms pulse at −100 mV. Values are mean±SE. The number of mice and cells used in Figs 2–4,6 in each condition are given in Table-1. Significantly different (*) estrus vs. male, (#) estrus vs. diestrus-2, (&) diestrus-2 vs. male (Figs 2–4).

Figure 3. Higher I_to,f, and I_K,slow in males than in females: prepulse separation.

A , I_K,slow with I_to,f inactivating prepulse (inset). Capacity transients and higher display are as in Fig. 2A. B, Peak I_K,slow-voltage curves. C, Bar graphs at +40 mV of I_K,slow (male 37 cells from 7 mice, estrus 33 cells from 10 mice, diestrus 34 cells from 10 mice), I_K,slow1 (male:12.5±2.2 16 cells from 4 mice; estrus:7.1±0.7 13 cells from 3 mice and diestrus-2:14.1±1.1, 14 cells from 3 mice) and I_K,slow2 (male:10.1±1 16 cells from 4 mice; estrus:9.5±0.6 13 cells from 3 mice and diestrus-2:9.4±0.8, 14 cells from 3 mice), τ_decay and time to peak. D, I_to,f from I_total-I_K,slow. E, Peak I_to,f-voltage curves. F, Bar graphs at +40 mV of I_to,f density, τ_decay and time to peak. *, #, & are as in Fig. 2. G, I_K,slow at +40 mV (I_to,f inactivated by prepulse) before (I_K,slow) and after application of 50 μM 4-AP (I_K,slow2). The traces are I_ss baseline subtracted. I_K,slow1 = I_K,slow−I_K,slow2.

Table 1.

Gender differences and estrogen regulation of K⁺ currents, APD and QTc interval

	Male	Estrus	Diestrus-2	Sham	E2-treated
	21–37 cells 7–11 mice	20–33 cells 7–10 mice	20–34 cells 7–10 mice	8 cells 3 mice	6 cells 3 mice
Itotal (pA/pF)a	48.6±3.0	27.2±2.3** #	39.1±3.4&	46.4±3.0	28.4+1.6††
Iss (pA/pF)b	4.4±0.2	3.6±0.3*	3.8±0.2	5.2±1.1	4.1±0.4
IKir (pA/pF)c	−11.5±0.7	−10.6±0.5	−11.8±0.6	−10.5±1.0	−10.8±1.1
Ibase (pA/pF)c	4.5±0.2	3.6±0.3*	3.9±0.2	5.4±1.1	4.1±0.3
Ifast (pA/pF)	26.2±1.9	14.3±1.6** #	20.3±1.6&	25.3±2.0	11.8±0.9††
Ifast-τ (ms)	80±7	99±11* #	72±4	64±5	115±15††
IK,slow (pA/pF)c	14.9±1.3	7.6±0.7** #	12.4±1.7	12.8±1.3	11.1±1.0
IK,slow-τ (ms)d	1128±34	1351±75** ##	1089±23	1088±73	1344±82
Ito,f (pA/pF)	26.8±1.9	14.9±1.6** #	22.1±2.1&	26.6±1.6	12.8±1.0††
Ito,f-τ (ms)d	108±14	103±26	95±15	80±11	103±25
Ito,f-time to Peak(ms)d	4.9±0.2	5.2±0.2	5.3±0.4	4.3±0.2	4.6±0.1
IK,slow (pA/pF)	21.9±1.8	14.6±0.6** #	20.3±1.4	22.2±1.6	17.2±1.4†
IK,slow-τ (ms)d	838±20	852±30	840±26	767±37	971±48†
IK,Slow-time to Peak(ms)	14.02±1.40	17.10±1.09	16.84±1.39	16.83±1.69	21.92±1.78
	8 cells, 3 mice	14 cells, 5 mice	8 cells, 4 mice	8 cells, 3mice	15 cells, 4 mice
APD50 (ms)	3.11±0.21	5.21±0.83* #	2.88±0.3	2.42±0.26	5.11±0.86††
APD75 (ms)	5.75±0.59	10.63±2.0* #	5.62±0.58	3.65±0.41	11.44±1.96††
APD90 (ms)	12.52±1.24	29.53±6.21*	11.96±1.19	9.28±1.74	34.01±4.2††
	10 mice	14 mice	11 mice	12 mice	9 mice
RR (ms)	124.4±3.0	123.9±4.0#	119±4.8	122.7±2.46	125.22±3.9
QRS (ms)	9.60±0.22	9.38±0.25	9.36±0.20	9.58±0.28	9.55±0.17
QT (ms)	52.0±1.2	57.0±3.7##	44.6±1.2&	44.1±0.8	52.22±1.5††
QTc (ms)	46.7±1.0	50.9±2.6##	41.05±0.96&	39.8±0.8	46.7±1.4††
HR	492±10	498±13	510±18	491±10	483±15

Outward current values are at +40 mV.

^aCurrent density at the end of 4.5 s pulse.

^bCurrent density at the end of 100 ms pulse at −100 mV.

^cMeasurements after current separation by exponential fitting², ⁶.

^dMeasurements after current separation by 100 ms prepulse to 0 mV⁴.

^*p<0.05 Estrus vs. Male;

^**p<0.01 Estrus vs. Male;

^#p<0.05 Estrus vs. Diestrus-2;

^##p<0.01 Estrus vs. Diestrus-2;

^&p<0.05 Diestrus-2 vs. Male;

^†p<0.05 E2 treated vs. Sham;

^††p<0.01 E2 treated vs. Sham.

Figure 3A shows I_K,slow (in the same cells as in Fig. 2A), elicited after inactivation of I_to,f with a 100 ms prepulse to 0 mV. I_to,f (Fig. 3D) was thus obtained from I_K,total (Fig. 2A) minus I_K,slow (Fig. 3A). While there were no gender differences in the decay time constant (τ_decay) of I_K,slow (3C) and I_to,f (3F), their densities showed exactly the same tendency as in gender differences where the currents were separated by exponential fitting⁶ (Figs. 3C, 3F vs. Fig. 2E). The kinetics of activation was similar for both I_K,slow and I_to,f in males, estrus and diestrus-2 (Fig. 3C, F). These results indicate that the gender differences found in current densities are independent of the method used.

To identify which component(s) of I_K,slow underlies lower I_K,slow in estrus, we used 4-AP (50 μm) to isolate I_K,slow1 (Kv1.5) from I_K,slow2 (Kv2.1)^9,10. We first recorded I_K,slow using a pre-pulse to inactivate I_to,f, then I_Kslow2 in the presence of 50 μM 4-AP which blocks I_K,slow1 and I_Kslow1=I_Kslow−I_Kslow2 (Fig. 3G). We validated this approach by directly recording I_Kslow1 after blocking I_Kslow2 with TEA (25 mM) and calculated I_Kslow2=I_Kslow−I_Kslow1 (not shown). In agreement with quantitative Real Time PCR data (See Fig. 5D,E), I_Kslow1, but not I_Kslow2 were significantly lower in estrus (under the influence of the high estrogen peak at proestrus) when compared to diestrus-2 and male (Fig. 3C,G). This data demonstrates that gender differences in I_K,slow is mediated by the differences only in I_K,slow1, but not I_K,slow2.

Figure 5. Cardiac Kv4.3 and Kv1.5 transcripts levels are lower under the influence of high estrogen.

Mean±SE relative transcripts in males (M), estrus (E) and diestrus-2 (D) normalized to male for Kv4.3 (A), Kv4.2 (B), KChIP2 (C), Kv1.5 (D), Kv2.1 (E) and GAPDH (F) (N=4–6 mice, *p<0.01 vs. male). Mean±SE relative transcripts in OVX sham (S), and E2-treated normalized to sham for Kv4.2 (G), KChIP2 (H), Kv1.5 (I), Kv2.1 (J) and GAPDH (K) (N=4–6 mice*p<0.01 vs. sham).

Inhibition of outward K⁺ currents by E2 treatment in OVX animals

To further test the hypothesis of E2-induced inhibition of outward K⁺ currents as being one mechanism underlying gender differences, we directly tested the effect of E2 on OVX mice. E2 treatment significantly reduced I_K,total at potentials more depolarized than −10 mV, without any change in I_Kir (Fig. 4A,B). I_K,total densities in OVX sham were similar to males and diestrus-2, while I_K,total densities in E2-treated OVX mice were similar to estrus (Fig. 4A–C). Similar results were obtained for I_K,slow and I_to,f, and the inhibitory action of E2 was even more pronounced for I_to,f than I_K,slow (Fig. 4D–Fvs. Fig. 4G–I).

Figure 4. Estrogen-induced inhibition of I_K,_total, I_K,slow and I_to,fA, D, G.

I_K,_total, I_Kir, I_K,slow and I_to,f were obtained as in Fig. 2A and Fig. 3 for sham- and E2-treated OVX mice. A, Capacity transients and higher display are as in Fig. 2A. B, E, H, I_K,_total- I_K,slow- I_to,f –voltage curves for sham (○) and E2-treated (●). C, F, I, Bar graphs for I_K,_total, I_K,slow and I_to,f comparing male (M), estrus (E) and diestrus-2 (D2) (from Figs. 2–3) with OVX-sham (S) and E2-treated (E2). Significantly different (†) OVX-sham vs. OVX-E2-treated.

Kv4.3 and Kv1.5 transcript levels are lower in estrus and are downregulated by E2 treatment in OVX mice

To identify the molecular correlates underlying the lower I_to,f and I_K,slow in estrus compared to male and diestrus-2, we used real-time PCR to quantify relative transcript levels of Kv4.3, Kv4.2, and KChIP2 (molecular correlates of I_to,f) and of Kv1.5 and Kv2.1 (molecular correlates of I_K,slow). Kv4.3 transcript levels were ~2 fold lower in estrus compared to male and diestrus-2, while Kv4.2 and KChIP2 transcript levels were similar in the three conditions (Fig. 5A–C), thus making Kv4.3 as a strong candidate to contribute to the gender differences of I_to,f. Regarding the molecular correlates of I_K,slow, Kv1.5 (Fig 5D) but not Kv2.1 (Fig. 5E) transcript levels were ~2 fold lower in estrus compared to male and diestrus-2, which is in agreement with our electrophysiological data showing that gender differences in I_K,slow are mediated by the differences in I_K,slow1 (encoded by Kv1.5), but not I_K,slow2 (encoded by Kv2.1).

We have previously shown that cardiac Kv4.3 transcript levels were downregulated by E2⁸. We show now that cardiac Kv1.5 transcript levels are also downregulated by E2 treatment (Fig. 5I). Kv4.2, KChIP2 and Kv2.1 (Fig. 5G–J) transcripts were not affected by E2 treatment. These results support the view that only Kv4.3 and Kv1.5 are regulated by E2.

Action potential duration and QTc interval are longer in estrus compared to diestrus-2 and in E2 treated OVX mice

We further investigated whether changes in K⁺ currents in estrus vs. male and diestrus-2 are associated with altered action potential duration and QTc interval. Fig. 6 shows that action potential durations (APD) at 50% (APD50) and 90% (APD90) of the depolarization were similar in males and in diestrus-2 (~3.1 ms and 12 ms respectively), while they were significantly longer in estrus (~5 ms and 30 ms, respectively). E2 treatment of OVX mice increased both APD50 and APD90 (from 2.4 ms to 5.1 ms and from 9.28 ms to 34 ms, respectively) to values similar to estrus (Table-1). This data supports the role of E2 in gender related differences of cardiac repolarization.

Figure 6. Action potential duration and QT interval are longer in estrus compared to male and diestrus-2; and E2 treatment increased both the action potential duration and QT interval.

Action potentials were elicited by 1 ms pulses in male (A), diestrus-2 (B), estrus (C), OVX sham (D) and E2-treated (E). Individual action potential durations (○) and their averages (●) at 50% (APD50, F) and 90% (APD90, G) of depolarization. Typical ECG recordings in dietrus-2 (H), estrus (I), OVX sham (J) and E2 treated (K). ECG parameters are given in Table-1.

Analyses of ECG recordings revealed no differences in the shape of P waves or the QRS complexes in estrus vs. diestrus-2 and in sham vs. E2-treated OVX mice (Fig 6H–K). The heart rate, RR and QRS durations were also similar between estrus, diestrus-2, sham and E2-treated mice (Table-1). Interestingly QTc interval was significantly longer in estrus (51 ms) vs. diestrus-2 (41 ms) and in E2-treated OVX (47 ms) vs. sham (40 ms), indicating that changes in action potential durations in females are associated with altered ventricular repolarization. In male mice however, although the APD was similar to diestrus-2, the QTc interval was significantly longer than diestrus-2, possibly due to the higher body weight and larger cardiac size compared to same-aged female mice (Table-1).

In summary, in female mice, changes in outward K⁺ current densities were associated with changes in action potential duration and ventricular repolarization.

DISCUSSION

Previous studies have shown conflicting data regarding gender related differences in K⁺ current densities and ventricular repolarization in mice^4–6. This conflicting data may arise from the use of female mice at random without establishing their estral stage. Taking into consideration that estrogen levels vary during the female estral cycle (Fig. 1), we revisited gender-related differences in K⁺ current densities and cardiac repolarization in male and female mice at estrus and diestrus-2, and in E2-treated OVX mice. We show for the first time that plasma E2 concentration is one of the key players in determining outward K⁺ current densities as well as affecting ventricular repolarization in mouse heart.

Higher estrogen levels correlate with reduced outward K⁺ current densities

In agreement with the role of E2 in downregulating K⁺ current densities, I_K,total densities were lower in animals under the influence of high plasma E2 (estrus and E2-treated OVX) compared with low plasma E2 levels (diestrus-2, males and OVX; Figs. 2, 4A). The lower I_K,total densities in animals under the influence of high plasma E2 were due to lower densities of both I_K,slow and I_to,f (Figs. 3, 4). Supporting a role of E2 in downregulating I_to,f, we found that males, which have lower levels of [E2] compared to females (Fig. 1), have significantly higher levels of I_to,f densities than females. We also show that I_to,f density in females is regulated during the estral cycle and by E2 treatment. These findings complement our previous work which shows that I_to,f is downregulated at the end of pregnancy⁸, where levels of plasma E2 are high.

Table-2 summarizes previously reported I_K,total, I_K,slow and I_to,f current densities, as well as our present work, in male and female mice. Lower I_K,total and I_K,slow but not I_to,f densities in female vs. male were reported by Trepanier-Boulay et al. using the CD-1 strain ⁴. In contrast to our finding and to those of Trepanier-Boulay et al. for I_K,total and I_K,slow ⁴, Brunet et al.⁶ using C57BL/6 mice reported no gender differences in I_K,total, I_K,slow and I_to,f densities, thus differences in strain is not a suitable explanation for the inconsistency in the results. Similar to our results, and opposite to Trepanier-Boulay et al. ⁴ and Brunet et al.⁶, higher I_to,f densities in males vs. females was previously reported by Wu & Anderson⁵. How can we explain the differences in these findings? One explanation is that if one uses animals at random, there is a lower probability to sample animals in estrus, where the changes are more pronounced. In addition one can argue that the lack of significant differences in the I_to,f densities reported by Trepanier-Boulay et al.⁴ in males and females could be due to recording currents from cells isolated from the whole ventricle and that the regional heterogeneity in the expression levels of I_to,f could have masked gender differences. In this work, we maximized homogeneity of our myocyte population by comparing only currents from cells isolated from the right ventricle, which also have the highest density of I_to,f⁶. However, the regional differences can not explain the discrepancy between our data and Brunet et al⁶, who demonstrated no significant differences in the I_K,total, I_to,f and I_K,slow densities in right ventricular myocytes between male and female mice². The contradiction between our present findings and Brunet et al.⁶ in comparing I_to,f from the same ventricular regions could lie in the experimental method. Brunet et al.⁶ used cardiomyocytes maintained in serum-free medium for up to 24 hrs, in which time they can undergo remodeling. In our study, we only used freshly dissociated cardiomyocytes within 4–6 hr after isolation.

Table 2.

I_K,total, I_K,slow and I_to,f current densities in male and female mice

Currents	Gender dependency	Strain	Region	Fresh vs. Culture	Estral cycle evaluation	Ref
IK,total
	M>F	CD-1	WV	Fresh	Unknown	4
	No-diff.	C57Bl/6	LV	Up-24 h	Unknown	6
	No-diff.	C57Bl/6	RV	Up-24 h	Unknown	6
	M>E D2>E	C57Bl/6	RV	Fresh	Yes	Present work
	No-diff.	db/+	RV	Fresh	Unknown	26
	M>F	C57Bl/6	Septum	Up-24 h	Unknown	6
IK,Slow
	M>F	CD-1	WV	Fresh	Unknown	4
	No-diff.	C57Bl/6	LV	Up-24 h	Unknown	6
	No-diff.	C57Bl/6	RV	Up-24 h	Unknown	6
	M>E D2>E	C57Bl/6	RV	Fresh	Yes	Present work
	M>F	C57Bl/6	Septum	Up-24 h	Unknown	6
Ito,f
	No-diff.	CD-1	WV	Fresh	Unknown	4
	M>F	C57Bl/6 129SVE	LV	Fresh	Unknown	5
	No-diff.	C57Bl/6	LV	Up-24 h	Unknown	6
	M>E D2>E	C57Bl/6	RV	Fresh	Yes	Present work
	No-diff.	C57Bl/6	RV	Up-24 h	Unknown	6
	M>F	C57Bl/6	Septum	Up-24 h	Unknown	6

M indicates male, F female, WV whole ventricle, RV right ventricle, LV left ventricle, Fresh (within 4–6 hr after cell isolation), E female at estrus, D2 female at diestrus-2 and No-diff. indicate no difference.

Higher estrogen levels correlate with prolongation of action potential duration and QTc interval

Here we show that estrogen plays a key role in regulating the ventricular repolarization in mouse heart. Lower I_to,f and I_K,slow densities in female mice in estrus compared to male and dietrus-2 were associated with longer action potential duration in estrus (Fig. 6). E2 treatment of OVX mice resulted also in the prolongation of action potential duration and QTc interval to a comparable level as in estrus, underscoring the role of E2 in the regulation of action potential duration and ventricular repolarization. Supporting a role of E2, in larger animals E2 treatment can downregulate cardiac Kv1.5 (HK2) and KCNE1 transcript levels of rabbit hearts, which is correlated with a prolonged QTc interval¹¹. In addition, E2 can also influence the expression as well as conductance of the L-type calcium channel (I_Ca,L) which plays an important role in the cardiac repolarization phase and in arrhythmogenesis. I_Ca,L channels of cardiomyocytes differ significantly between day 0 (peak of plasma estrogen, proestrus) and day 4 (peak of progesterone-to-estrogen ratio, diestrus-2) of the estral cycle of female guinea pigs, suggesting that I_Ca,L density varies according to estral hormonal changes (See review¹). Moreover, E2 treatment reduced I_Ca,L density of OVX rat¹², and I_Ca,L density was increased in estrogen receptor α (ERα)-deficient mice (See review¹), thus demonstrating that the inhibitory action of E2 was mediated mainly via the ERα. Here we speculate that the large variability in the duration of APD50 and APD90 in estrus and in E2-treated OVX mice (Fig. 6) is due to different expression levels of ERα in different cardiomyocytes, and the variability is more evident under the influence of high estrogen (Fig. 6). Supporting the role of sex hormones in regulating APD and cardiac repolarization, testosterone has been previously proposed to play a role in gender-based cardiac repolarization differences, as androgen deficiency in castrated mice produced a longer APD and prolongation of QTc interval (See review¹). Similar to mice, ventricular repolarization was significantly altered by testosterone treatment in dogs¹³ and rabbits (See review¹).

Mechanism of estrogen action: Regulation of Kv4.3 and Kv1.5 transcript levels

Several reports suggest that the expression of Kv4.3 and Kv1.5 are regulated by sex hormones^{8, 14}. Our data demonstrate that Kv4.3 was the only molecular correlate of I_to,f which its transcript levels were lower under the influence of high estrogen (estrus and E2-primed OVX) compared to low estrogen (diestrus-2, male and OVX sham). E2-induced decreased expression of Kv4.3 was not accompanied with significant alteration in Kv4.2 and KChIP2 transcripts (Fig. 5). The lower Kv4.3 transcripts in estrus vs male and diestrus-2 and in E2-treated OVX mice vs sham coincided with reduced I_to,f densities. Regarding the molecular correlates of I_K,slow, Kv1.5 but not Kv2.1 transcript levels were lower in estrus compared to male and diestrus-2 and were downregulated by E2 treatment. The fact that I_K,slow2 peak current densities were similar (encoded by Kv2.1) in male, estrus and dietrus-2 (Fig. 3C,D), whereas I_K,slow1 (encoded by Kv1.5) peak current densities were significantly lower in estrus compared to male and diestrus-2, further support that the transcriptional changes in Kv1.5 underlie changes in I_K,slow ruling out the possibility of estrogen-induced posttranslational modification of Kv1.5 protein. Higher Kv1.5 transcripts in males than in females has been reported, and testosterone has been suggested to be the main player in this regulation¹⁴. Our data suggest that transcriptional regulation of Kv4.3 and Kv1.5 by E2 is at least one of the mechanisms responsible for the action of estrogen in remodeling cardiac repolarization.

How can estrogen regulate the transcriptional activity of Kv4.3 and Kv1.5 genes? Estrogen can regulate the transcriptional activity of a gene via a direct genomic classical pathway, which requires the presence of Estrogen Response Elements (ERE) in the promoter of the gene. Alternatively, estrogen can stimulate the transcription through the DNA-associated transcription factors known as indirect genomic pathway via estrogen receptors¹⁵. Analysis of 10 kb upstream of Kv4.3 and Kv1.5 transcription start sites failed to identify an ERE, but showed several half-EREs as well as Sp1 and Ap1 elements, that could be involved in reducing expression of Kv4.3 and Kv1.5 by E2. There is also an increasing body of evidence that direct application of estrogens can acutely regulate different types of K⁺ and I_Ca,L currents, affecting action potential durations through non-genomic pathways^{1, 16}. For example, direct application of raloxifene (a selective estrogen receptor modulator) to human atrial myocytes inhibited I_to,f and I_K,slow within minutes in a concentration-dependent manner, as well as prolonged APD¹⁶. Although our Real Time PCR data supports a genomic action of E2, a non-genomic contribution cannot be excluded.

Influence of estrogen on QTc interval and arrhythmia susceptibility during the menstrual cycle and estrogen replacement therapy in women

Here we showed that in female mice, the changes in plasma E2 levels during the estral cycle, as well as by E2 treatment, affect the QTc interval. The QTc is significantly longer in high estrogenic conditions (estrus and E2-treated OVX) in comparison to low estrogenic conditions (diestrus-2 and OVX). In premenopausal women, circulating levels of estrogen and progesterone change dramatically during the menstrual cycle. The early follicular phase (menses) is characterized by a low level of estrogen (~100pmol/l). Estrogen levels steadily increase after onset of menses and reach a peak (~750 pmol/l) at about day 14 of the ovulatory phase¹⁷. Estrogen levels fall during the luteal phase, when progesterone levels begin to increase. Changes in QTc interval during the menstrual cycle has been cited with conflicting results. Several investigators found no change in the QTc interval at different times of the menstrual cycle^17–19, whereas Nakagawa et al. reported a shorter QTc interval during the luteal phase compared to the follicular phase²⁰. The shorter QTc interval in the luteal phase has been attributed to higher levels of plasma progesterone concentration (~5 fold) compared to the follicular phase, as plasma estrogen and testosterone concentrations were similar in both phases²⁰. It is well accepted that women are at a higher risk of developing drug-induced torsades de pointes than men. How the changes in QTc interval during the menstrual cycle might influence the arrythmogenesis in female patients with ventricular arrhythmia or paroxysmal supraventricular tachycardia (SVT) has been of great interest. The prolongation of QTc duration induced by ibutilide, an antiarrhythmic agent, was significantly greater at menses and the ovulatory phase than in the luteal phase¹⁹. The drug-induced prolongation of QTc was inversely correlated with higher levels of progesterone and progesterone-to-estrogen ratio, rather than estrogen alone¹⁹. The occurrence and the duration of SVT episodes in women has been suggested to be affected by the menstrual cycle, being higher in day 28 when progesterone levels are higher, as opposed to day 7 of the menstrual cycle, when estrogen predominates²¹. The incidence of arrhythmia in SVT patients was found to be slowed down by estrogen and facilitated by progesterone²¹.

The effect of estrogen and progesterone replacement therapy on ventricular repolarization has also been investigated in healthy postmenopausal women. Hormone replacement therapy with estrogen alone significantly prolonged QTc interval^22,23, whereas combined estrogen and progestin therapy had no effect²⁴ or reduced the QTc interval^{23, 25}. These findings suggest that sex hormones may directly modulate ventricular repolarization; estrogen is reported to account for the prolongation of QTc interval similar to our results in mice whereas progesterone tends to reduce QTc interval. A detailed study to correlate plasma estrogen and progesterone to the QTc interval during the menstrual cycle and hormone replacement therapy should provide key information to set the appropriate timing of the treatment or electrophysiological procedure in premenopausal women more prone to arrhythmias.

Non-standard Abbreviations and Acronyms

I_to,f

fast-transient outward K⁺ current

I_Kur or I_K,slow

ultra-rapid delayed rectifier K⁺ current

I_ss

steady-state current

I_Kir

inward rectifier K⁺ current

I_Ca,L

L-type calcium channel

E2

estrogen

ERE

Estrogen Response Elements

ERα

estrogen receptorα

OVX

ovariectomized

APD

action potential duration

QTc

corrected QT

TEA

tetraethylammonium

4-AP

4-aminopyridine

SVT

paroxysmal supraventricular tachycardia