Optical mapping was performed in 6 mice from each group. After performing a median sternotomy, the hearts were quickly excised. The hearts were stained with RH237 (Invitrogen, California, USA), and were excited with quasi-monochromatic light (520±30 nm) from two green LED lamps. A camera with a 610-nm long pass filter (MiCAM Ultima, BrainVision, Tokyo, Japan) was used to collect the emitted light at 1 ms/frame and 100 × 100 pixels with a spatial resolution of 0.5×0.5 mm²/pixel.⁸ To determine if the activation of Htr3 could shorten the action potential duration (APD), we measured the APD₉₀ after infusion of serotonin (100 μmol/L), the Htr3 agonist, 1-(m-Chlorophenyl)-biguanide (m-CPBG, 200 nmol/L), and the Htr3 antagonist, ondansetron (1 μmol/L).

All physiological tests were performed during non-pregnant (NP) and late pregnant (LP; G16–18) stages. All procedures were performed after an anesthetic (isoflurane/O2) was administered via inhalation, and the heart was quickly removed after deep anesthesia was established. The heart was cannulated using a 24 G needle and then retrogradely perfused via the aorta on a Langendorff apparatus. During coronary perfusion, all perfusates were maintained at 37°C and equilibrated with 100% O2. Initially the heart was perfused with normal Tyrode solution for 2–3 min to clear the blood. This was followed by perfusion with a Ca²⁺ free solution for 3 min, after which, the heart was perfused with an enzyme solution containing 0.14 mg mL⁻¹ collagenase (Yakult) in Ca²⁺ free solution for 12 min. After perfusion with the enzyme solution, the atria were separated from the ventricles, and chopped into small pieces. From these small pieces, single cells were dissociated in a high-K+ and low-Cl⁻ solution using a blunt-tip glass pipette, and stored in the same solution at 4°C until use.

The recording protocols for Kv currents were also performed in ventricular myocytes, which were isolated from adult (8–10 weeks of age) C57BL6 mice as previously described.² Briefly, we used pre-pulse to −40 mV for 25 ms to inactivate I_Na. Kv current was recorded in voltage-clamp mode with 400 ms pulses from a holding potential of −70 mV, with different test potentials increasing from −40 mV to +60 mV in 10 mV steps. The currents were normalized to the cell membrane capacitance (Cm) and averaged. The values were not corrected for the junction potential (pipette offset); this was compensated prior to giga-seal formation.

Membrane isolation was conducted as explained earlier.³ All procedures were performed at 4°C. Ventricular myocytes were homogenized in a homogenization buffer (20 mmol/L Tris–HCl [pH 7.4], 1 mmol/L EDTA, and a protease inhibitor cocktail (2.5 μL/ mL, Sigma). Nuclei and debris were pelleted by centrifugation at 500 g for 10 min. The supernatant was further centrifuged at 25,000 g for 30 min. The pellet thus obtained was resuspended in homogenization buffer and centrifuged again at 25,000 g for 30 min. The final pellet was stored at −80°C. The protein concentration was determined by the Pierce BCA Protein Assay (Sigma-Aldrich).

The primary antibodies used were anti-Kv4.3 (1:200, Santa Cruz Biotechnology), anti-Htr3a (SR-3A, 1:500, Santa Cruz Biotechnology), anti-pan cadherin (1:500, abcam), and anti-EEA1 (1:1000, BD Transduction Laboratories).

Cell lysates were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto nitrocellulose membranes (Bio-Rad, Richmond, CA, USA). After blocking with 5% non-fat dry milk in 1X tris-buffered saline (TBS) containing 1% Tween 20 (for 1h), the membrane was processed for immunoblot analysis with indicated primary antibodies. The primary antibodies were detected with horseradish peroxidase (HRP)-conjugated anti-rabbit or anti-goat immunoglobulin G (IgG) secondary antibodies. Detection was performed using an enhanced chemiluminescence detection system (ECL, Amersham Pharmacia Biotech, Piscataway, NJ, USA).

Confocal microscopy was performed on isolated mouse ventricular myocytes as described previously.⁴ Briefly, ventricular myocytes were plated on laminin (10 μg/mL)-coated coverslips for 3 h at 4°C, fixed with 4% formaldehyde in phosphate-buffered saline (PBS) on ice, permeabilized in 2% donkey serum in PBS with 0.1% Triton X-100 for 30 min, and then incubated with the primary antibody against Kv4.3 (1:200, Santa Cruz Biotechnology) and Htr3a (SR-3A, 1:200, Santa Cruz Biotechnology), overnight at 4°C, followed by incubation with the Alexa Fluor 633-conjugated goat anti-mouse secondary antibody or Alexa Fluor 633-conjugated goat anti-rabbit secondary antibody and Alexa Fluor 488-conjugated donkey anti-goat secondary antibody at a 1:200 dilution for 1 h at room temperature. The immunostained cells were visualized using an LSM 700 confocal laser scanning microscope (Carl Zeiss). Cells were randomly selected and used for imaging and analysis, and immunostaining experiments were repeated at least five times.

To evaluate the effect of ondansetron on ventricular repolarization, we examined the effects of ondansetron on APD between NP and LP mice (Figure 1). When the animals were pretreated with an Htr3 antagonist, ondansetron (1 μmol/L), serotonin failed to shorten APD₉₀. These results indicate that serotonin treatment shortens the APD and consequently the QT interval.

We blocked serotonin receptors by applying ondansetron on freshly isolated LV cardiomyocytes from the NP and LP animals. Even at a high concentration, ondansetron (50 μM) did not decrease Kv current densities in NP mice. In contrast, ondansetron (50 μM) significantly decreased the Kv current densities in LP mice (44.3±1.9 vs. 35.8±2.7 pA/pF, p<0.01) (Figure 2).

To measure the effect of ondansetron on the Kv current densities, LV myocytes were treated with various concentrations of ondansetron (0, 1, 5, 50 μM) for 10 min. As shown in Figure 3A, at higher concentrations (5, 50 μM), ondansetron significantly inhibited the Kv current densities.

Kv channels control the height and duration of action potentials in the myocardium and are the primary determinants of action potential repolarization. To identify the molecular determinants of serotonin-induced APD shortening, we examined Kv channel properties; whole-cell patch clamp recordings were obtained from LV myocytes isolated from adult NP and LP mice. Figure 4A shows the waveforms of the Kv currents from NP (upper panels) and LP (lower panels) mice, respectively. Treatment with serotonin (100 μM) and m-CPBG (30 μM) for 10 min increased Kv current densities in NP, but not in LP mice. Co-application of ondansetron (50 μM) for 10 min blocked the increase in Kv current densities observed in NP mice caused by serotonin treatment.

Figure 4B shows that serotonin increased the amplitude of peak Kv current densities (at +60 mV) in NP, but not in LP mice (55.0±2.5 vs. 72.7±4.9 pA/pF, p<0.001). Likewise, the Htr3a agonist, m-CPBG (30 μM), increased peak Kv current densities (at +60 mV) in NP mice, but not in LP mice (55.0±2.9 vs. 65.0±2.6 pA/pF, p<0.001). This data demonstrated that the Kv current was increased by serotonin and the agonist.

We evaluated the interaction between Kv4.3 and Htr3a in NP mice using confocal microscopy. Figure 5 revealed that Htr3a and Kv4.3 were co-localized on the surface of the membrane, T-tubules, and cytosol. With serotonin pre-treatment (100 μM) for 1 h, Kv4.3 and Htr3a were translocated to the surface membrane in NP mice. Co-application of ondansetron (50 μM) for 1 h erased the increase in membrane trafficking observed in NP mice caused by serotonin treatment.

Because serotonin treatment caused the translocation of Kv4.3 and Htr3a to the membrane, we performed cell fractionation (see method). Serotonin and the agonist increased the abundance of Kv4.3 protein in the membrane fraction. On the other hand, co-application of serotonin with ondansetron (50 μM) for 1h erased this phenomenon in NP (Figure 6A [upper panel]), but not in LP mice (Figure 6A [lower panel]). This data supports the suggestion that the serotonin and agonist treatment caused the translocation of Kv4.3 and Htr3a to the membrane.

Ondansetron prolonged the QT interval with a decrease in Kv current densities and reduced the amplitude of peak current densities in a dose-dependent manner in the hearts obtained from pregnant mice, but not non-pregnant mice. During pregnancy, Htr3a-mediated Kv4.3 membrane trafficking was significantly impaired. This data suggests that ondansetron should be used cautiously during pregnancy, which is physiologically related with QT prolongation.

In this study, high concentrations of ondansetron decreased Kv current densities in pregnant animals. Voltage-dependent Na⁺ and K⁺ channels are important determinants of the human action potentials. Yuri et al. reported that the submicromolar affinity of ondansetron for the hERG K⁺ channel is likely responsible for the prolongation of cardiac repolarization by this drug (YURI A et al., 2000). However, its underlying mechanisms in other Kv channels have not been revealed. Moreover, even though this drug is frequently used as an antiemetic agent in pregnant women, the effect of ondansetron on cardiac repolarization during pregnancy has not been evaluated.

The maternal heart significantly adapts to the circulatory needs of pregnancy, but the effect of pregnancy on ventricular repolarization is poorly understood. Pregnancy is associated with electrocardiogram disturbances such as a longer QT interval. Eghbali et al.^5,9 reported that I_to,f and I_K,slow are down-regulated in pregnancy. In particular, cardiac Kv4.3 channel gene expression was down-regulated by 3- to 5-fold, and was accompanied by a reduction in the transient outward K⁺ currents, a longer action potential, and prolongation of the QT interval. Interestingly, in this study, the Htr3a antagonist, ondansetron, further decreased Kv current densities in pregnant mice. However, the effect of serotonin and the Htr3 agonist on cardiac repolarization was not observed in pregnant mice, whereas they caused shortened QT interval with increased Kv currents in non-pregnant mice. This finding suggests that although the effect of serotonin on cardiac repolarization is reduced, serotonin compensates for the QT prolongation during pregnancy by increasing Kv currents via Htr3.

Unlike the other serotonin receptor genes, which encode G-protein coupled serotonin receptors, Htr3a encodes subunits of the ligand-gated cation channel, and is essential for the formation of functional Htr3 channels.^6,8,106,8,10 The channel is predominantly selective for Na⁺ and K⁺, and its serotonin-activated opening depolarizes the cell membrane by permeating an inward Na⁺ current.^11,12 The precise role of Htr3 as an ion channel expressed in the heart has not been fully evaluated. Our data demonstrates that serotonin increased Kv currents by increasing the membrane trafficking of Kv4.3 in non-pregnant mice. Therefore, it can be speculated that the effect of serotonin on cardiac repolarization was mainly mediated by an increase in Kv current via Htr3. However, this effect was not observed in NP mice pretreated with the Htr3 antagonist, ondansetron.

KCNH2 channel trafficking is well reported. The KCNH2 channel traffics relatively inefficiently by itself¹³ and depends on chaperone proteins for proper folding¹⁴ and trafficking to the plasma membrane.14 However, the Kv4.3 channel trafficking is not well identified.

Our data shows that serotonin increases the Kv current densities via Htr3. Although we have demonstrated the relationship between the Htr3a and Kv channels using co-immunoprecipitation, it is still not clear whether the activated Htr3 increase influenced the Kv current densities directly or indirectly. Moreover, serotonin and Tph1 were increased during pregnancy. However, we could not exactly identify the source of the serotonin that affected the heart.