Overexpression of the SK3 channel alters vascular remodeling during pregnancy, leading to fetal demise

Abstract

The maternal cardiovascular system undergoes hemodynamic changes during pregnancy via angiogenesis and vasodilation to ensure adequate perfusion of the placenta. Improper vascularization at the maternal-fetal interface can cause pregnancy complications and poor fetal outcomes. Recent evidence indicates that small conductance Ca²⁺-activated K⁺ channel subtype 3 (SK3) contributes to vascular remodeling during pregnancy, and we hypothesized that abnormal SK3 channel expression would alter the ability of the maternal cardiovascular system to adapt to pregnancy demands and lead to poor fetal outcomes. We investigated this hypothesis using transgenic Kcnn3^tm1Jpad/Kcnn3^tm1Jpad (SK3^T/T) mice that overexpress the channel. Isolated pressurized uterine arteries from nonpregnant transgenic SK3^T/T mice had larger basal diameters and decreased agonist-induced constriction than those from their wild-type counterparts; however, non-receptor-mediated depolarization remained intact. In addition to vascular changes, heart rates and ejection fraction were increased, whereas end systolic volume was reduced in SK3^T/T mice compared with their wild-type littermates. Uterine sonography of the fetuses on pregnancy day 14 showed a significant decrease in fetal size in SK3^T/T compared with wild-type mice; thus, SK3^T/T mice displayed an intrauterine growth-restricted phenotype. The SK3^T/T mice showed decreased placental thicknesses and higher incidence of fetal loss, losing over half of their complement of pups by midgestation. These results establish that the SK3 channel contributes to both maternal and fetal outcomes during pregnancy and point to the importance of SK3 channel regulation in maintaining a healthy pregnancy.

successful pregnancy requires rapid and diverse changes in maternal cardiovascular physiology to ensure appropriate blood flow through the placenta (20). Dynamic changes in maternal heart rate, stroke volume, venous pressure, blood volume, and systemic vascular resistance result from formation of new blood vessels at the maternal-fetal interface and dilation of existing vessels (1). These adaptive changes are essential for providing oxygen and nutrients for, as well as eliminating waste from, the fetus. In humans, hemodynamic changes start at ∼3 wk of gestation, with maximal variations from the nonpregnant state beginning in the second trimester. These two phases of hemodynamic change correlate temporally with common time frames for miscarriage and fetal loss. In mice, changes in heart rate and mean arterial pressure can be detected in the mother as early as pregnancy days 6–8 compared with nonpregnant states (5). In both mice and humans, underlying medical conditions that compromise the ability of the cardiovascular system to adapt, including improper remodeling of the maternal uterine circulation, are associated with pregnancy complications. These maternal complications and associated fetal complications include intrauterine growth restriction, preterm delivery, and fetal demise (23). Given that ∼50% of conceptions end in miscarriages (15, 27), understanding the relationship between cardiovascular maladaptation, remodeling, and pregnancy complications is central to assuring maternal and fetal health.

Potassium channels regulate vascular smooth muscle dilation, proliferation, and angiogenesis (19) and thus can be a contributing factor in vascular remodeling during pregnancy. The small conductance Ca²⁺-activated K⁺ channel subtype 3 (SK3 channel) has emerged as an important ionic regulator of membrane excitability during pregnancy. Transgenic Kcnn3^tm1Jpad/Kcnn3^tm1Jpad (SK3^T/T) knock-in mice, which exhibit approximately threefold increased expression of the SK3 channel, have compromised parturition (2, 21). Another notable phenotype of the SK3^T/T mice is abnormal vessel branching, and increased diameters, of mesenteric arteries (28). Although the SK3 channel is absent in vascular smooth muscle, it is abundant in endothelial cells, where its primary role is to modulate arterial tone and blood pressure (4, 28). In this context, the SK3 channel contributes to sustained hyperpolarization of the endothelial membrane potential, which is conveyed through diffusible factors to the membranes of the vascular smooth muscle cells. Endothelial SK3 channels have a substantial tonic influence on vascular tone in the absence of any exogenous endothelial stimulation (28). SK3 channels mediate endothelium-dependent hyperpolarizing factor (EDHF)-induced vascular relaxation, and blocking SK3 channels nearly abolishes EDHF-induced vasodilation in human systemic omental and myometrial arteries (9) and late-pregnant rat uteroplacental resistance vessels (43). For this reason, the SK3 channel is thought to be a potential target for pregnancy-related vascular diseases, like preeclampsia, in which patients have abnormal EDHF-mediated vasodilation (9). Functional SK3 knockout (SK3^Dox) mice, obtained via doxycycline treatment of the SK3 transgenic mice (2), have a substantial elevation of blood pressure, due in part to disruption of EDHF-mediated vasodilation (3).

Despite the extensive branching found in the mesenteric circulation in SK3^T/T mice, other vascular beds have not been investigated, and little is known of the role of the SK3 channel in mediating pregnancy-induced cardiovascular changes. These studies address the hypothesis that overexpression of SK3 channels alters the ability of the maternal cardiovascular system to adapt to pregnancy demands, resulting in poor fetal outcomes.

MATERIALS AND METHODS

Mouse Breeding

All animal procedures complied with guidelines for the care and use of animals set forth by the National Institutes of Health. All protocols were approved by the Animal Care and Use Committee at the University of Iowa and the Animal Studies Committee at Washington University in St. Louis. Kcnn3^tm1Jpad/Kcnn3^tm1Jpad (SK3^T/T) mice on a C57BL/6 background were used for this study. Adult female mice were mated at 8 wk of age or older for 2-h time periods with an adult WT littermate male to ensure heterozygous offspring. Day 0 of pregnancy (P0) was determined by the presence of a copulatory plug. Litter size was determined by counting the number of pups at delivery or after excising the uterus of the mouse on P7–19.

Measurement of Uterine Artery Basal Diameter and Vasoconstriction

Nonpregnant (NP) SK3^T/T and WT mice were euthanized by CO₂ inhalation, and the uterus was excised and placed into ice-cold Krebs buffer (in mM: 118.3 NaCl, 4.7 KCl, 1.2 MgSO₄, 1.2 KH2PO₄, 25 NaHCO₃, 2.5 CaCl₂, and 5 d-glucose, pH 7.4). The main uterine artery was removed from adjacent adipose tissue and tied with ophthalmic silk sutures (10-O) onto glass micropipettes in an organ bath filled with continuously oxygenated (5% CO₂-20% O₂, balanced N₂) Krebs buffer. Vessels were allowed to equilibrate for 30–45 min. Viability of vessels was tested at the beginning and end of each experiment by measuring constriction in response to 25–50 mM KCl. The diameters of the vessels, in both the presence and absence of Ca²⁺, were measured at 20, 40, 60, and 80 mmHg using a video dimension analyzer system. Responses to the thromboxane A₂ mimetic U-46619 (10–50 nM) were compared between WT and SK3^T/T uterine arteries. After development of a stable constriction to U-46619, acetylcholine (1 μM) was added to test endothelial function. After washing, apamin (500 nM) was added for 15 min to determine SK3 contribution to basal tone of the uterine arteries from WT and SK3^T/T mice.

Measurement of Blood Pressures

Heart rate (HR) (8) and systolic blood pressure (SBP) were measured using a computerized tail cuff procedure (BP-2000 system; Visitech Systems, Apex, NC). Cardiovascular measurements were taken for 5 days for habituation prior to breeding. Measurements of blood pressures were restarted 6–7 days after identification of a copulatory plug and continued every day throughout gestation. The mean pressure over 20 min each day was obtained for each animal.

Ultrasound Imaging and Analysis

Mice were imaged when nonpregnant, and on P14 and P18. After sedation with midazolam (1.2–5 mg delivered subcutaneously), mice remained awake but docile. The anterior thorax, abdomen, and pelvis were shaved, and warmed gel was applied to optimize the acoustic interface. The animal was grasped by the nape of the neck and cradled in the left lateral recumbent position in the imager's hand. A 40 MHz linear array probe was applied to the chest. The imaging probe was coupled to a Vevo 2100^R imaging system (VisualSonics, Toronto, ON, Canada), generating ∼180–200 two-dimensional (2-D) frames per second in both short- and long-axis left ventricular planes. Uterine and fetal sonography was performed immediately thereafter, with application of the probe to the lower abdomen. Fetal cardiac activity and maternal uterine and umbilical artery flow velocity were visualized using 2-D images, and pulse-wave Doppler evaluation was used to measure flow velocity in small branches of the uterine arteries adjacent to the gestational sacs. Fetal size and placental thickness were assessed using 2-D images measured at the point of insertion of the placenta.

Sonographic images were analyzed in a blinded fashion. Endocardial and epicardial borders were traced in the short-axis plane at end diastole and end systole. The lengths from left ventricular outflow tract to endocardial apex and to epicardial apex were measured at end diastole and end systole, respectively. Left ventricular mass was calculated using the biplane area-length method, which was previously validated in our laboratory (7). This method was also used to calculate end-diastolic and end-systolic left ventricular volumes, cardiac output, stroke volumes, and ejection fractions. Fetal viability was ascertained by 2-D visualization of morphological features characteristic of gestational age, and by verification of fetal cardiac motion and Doppler velocities. Sonography was used to measure crown-rump length (fetal size), placental thickness (at the point of umbilical cord insertion), fetal HR, umbilical artery resistive index, and maternal uterine vessel resistive index. Resistive index was calculated as 1 − (1/pulsatility index), with pulsatility index being defined as systolic/diastolic flow velocity by Doppler, as an indicator of vascular resistance downstream of the uterine artery.

Statistical Analysis

All data are presented as means ± SE. Statistical significance was determined by either a Students t-test or a two-way ANOVA where appropriate, followed by a Bonferroni post hoc test. A Fisher exact test was used to determine significance of the fetal demise data; n refers to number of animals in all cases.

RESULTS

Basal Diameter and Vasoconstriction in Mouse Uterine Arteries

The maternal vasculature of the placenta plays an important role in regulation of blood flow to the fetus. Both morphological and functional vascular abnormalities were reported in mesenteric arteries from SK3^T/T mice (28). On the basis of these findings, we sought to determine whether a similar functional adaptation occurs in the uterine vasculature. In NP states, the diameters of the WT mouse uterine arteries were similar to those in previous reports (14); however, the Ca²⁺-free baseline diameters of SK3^T/T uterine arteries were significantly larger than WT arteries, suggesting vascular remodeling (Fig. 1A). Ca²⁺-free buffer had little effect on the basal diameter or the pressure-diameter relationship, suggesting the differences in the diameters between SK3^T/T and WT may be due to structural changes in uterine artery development as opposed to alterations in vasodilation alone. Apamin (500 nM), an inhibitor of the SK3 channel, tended to constrict arteries from SK3^T/T mice to a greater extent than arteries from WT mice, but this did not reach significance (Fig. 1B; P ≤ 0.07). Percent constrictions of uterine arteries from WT and SK3^T/T mice to KCl (25 and 50 mM) were similar (Fig. 1C), suggesting that contractile responses were intact despite the greater basal diameter in the SK3^T/T arteries. However, constriction in response to U-46619, a mimetic of the vasoconstrictor thromboxane A₂, was attenuated in SK3^T/T mice compared with WT (Fig. 1D). These results indicate that SK3^T/T vessels constrict normally to smooth muscle depolarization but that G protein-coupled receptor-mediated vasoconstriction is attenuated.

Fig. 1.

A: pressure-diameter curve of nonpregnant (NP) WT and SK3^T/T uterine artery in the presence and absence of Ca²⁺. Pressure curves between WT and SK3^T/T in Ca²⁺-free conditions (n = 8). B: effect of apamin (500 nM) on uterine artery diameters between WT and SK3^T/T mice (n = 10). C: percent constriction of uterine to KCl (n = 10). D: U-46619 (n = 3–6). Constriction in response to U-46619 was significantly reduced in uterine arteries from SK3^T/T vs. WT. Values are means ± SE. *P < 0.05 vs. WT.

Plethysmography, Echocardiograms and Ultrasonography Measurements in SK3^T/T and WT Mice

Maternal parameters.

To assess whether the increased uterine artery diameter seen with overexpression of the SK3 channel affected maternal hemodynamics during pregnancy, we investigated the cardiovascular adaptations to pregnancy in SK3^T/T vs. WT mice by use of tail cuff plethysmography, echocardiography, and ultrasound imaging.

Although functional knockout of the SK3 channel results in elevated blood pressure (7), systolic blood pressures were similar between WT and SK3^T/T mice in both the NP state and throughout gestation, suggesting that abnormal blood pressure regulation cannot account for the abnormalities during pregnancy in SK3^T/T mice (n = 3–7; Fig. 2A).

Fig. 2.

A: maternal systolic blood pressure for preganancy days (P)7–18 WT and SK3^T/T mice. Maternal cardiac parameters for NP, P14, and P18 WT and SK3^T/T mice. Comparison of maternal heart rate (B), end-systolic volume (C), and ejection fraction (D) in NP, P14, and P18 WT and SK3^T/T mice had differences in these cardiac parameters at most gestations. No differences were noted in NP stages. Values are means ± SE. *P < 0.05, **P < 0.001 vs. WT.

In the NP state, no significant differences were seen between WT and SK3^T/T mice in any of the maternal cardiac parameters measured, but, as expected, cardiovascular parameters changed dynamically during pregnancy in both WT and SK3^T/T mice (Table 1). Pregnancy-associated increases in stroke volume (55%), cardiac output (71%), and HR (7%) of the P18 WT mice were comparable to values in other studies of cardiovascular adaptations in the mouse in late pregnancy (Table 1) (13). The SK3^T/T P18 mice had similar pregnancy-associated increases in stroke volume (48%), cardiac output (57%), and HR (8%) compared with the NP controls (Table 1). The SK3^T/T mice, despite adapting hemodynamically, appeared to have significantly altered cardiac parameters compared with the WT mice. One alteration detected in nonpregnant SK3^T/T mice was HR, and this effect was also seen in the SK3^T/T mothers, which had significantly higher HR than WT dams by P14 and P18 (Fig. 2B). Despite the increase in HR in SK3^T/T mice during pregnancy, there was no significant difference in cardiac output between WT and SK3^T/T mice at any stage of pregnancy (Table 1). Other cardiac changes were as expected, with an increase in the left ventricular mass in both WT and SK3^T/T mice (Table 1). Although no significant difference was measured between the NP genotypes, the hearts from WT mice grew more dramatically between NP and P14 states than those measured in SK3^T/T mice (Table 1).

Table 1.

Cardiac parameters for NP, P14, and P18 WT and SK3^T/T mice

Strain and Pregnancy Stage	HR, beats/min	EDV, μl	ESV, μl	Mass, mg	SV, μl	CO, ml/min	EF	N
WT NP	572.2 ± 35.6	32.98 ± 4.51	10.36 ± 2.14	72.21 ± 4.99	22.62 ± 3.40	12.54 ± 1.80	0.695 ± 0.035	12
SK3T/T NP	645.4 ± 21.15	36.43 ± 3.24	10.96 ± 1.17	83.95 ± 4.87	25.46 ± 2.69	16.64 ± 1.99	0.695 ± 0.023	10
WT P14	546.6 ± 25.98	49.02 ± 3.37	14.05 ± 1.88	87.90 ± 3.73	34.97 ± 2.55	18.66 ± 1.30	0.716 ± 0.026	15
SK3T/T P14	655.9 ± 11.60**	40.34 ± 3.59	7.13 ± 0.842*	87.96 ± 4.17	33.20 ± 3.61	21.91 ± 2.57	0.813 ± 0.024*	14
WT P18	611.5 ± 18.85	47.64 ± 4.08	12.53 ± 1.62	86.99 ± 2.78	35.11 ± 3.52	21.50 ± 2.32	0.731 ± 0.034	10
SK3T/T P18	699.6 ± 18.20*	46.60 ± 3.36	8.88 ± 1.53	89.93 ± 3.51	37.72 ± 2.26	26.18 ± 1.35	0.817 ± 0.020*	11

Values are means ± SE. WT, wild type; SK3, small conductance Ca²⁺-activated K⁺ channel subtype 3; NP, nonpregnant; P, pregancy day; HR, heart rate; EDV, end-diastolic volume; ESV, end-systolic volume; SV, stroke volume; CO, cardiac output; EF, ejection fraction; Mass, left ventricular mass.

^*P < 0.05,

^**P < .001 vs. respective WT.

Surprisingly, no differences in end-diastolic volume (EDV) or stroke volume (SV) were detected between SK3^T/T and WT mothers at any gestational stage. Conversely, end-systolic volume (ESV) significantly differed at the P14 stage; a drop in ESV was detected in the SK3^T/T P14 dams (vs. SK3^T/T NP controls), whereas ESV increased in the WT mothers from NP to P14 (Fig. 2C). This difference was reconciled by P18. Ejection fraction (EF) was similar in both NP genotypes, but at both P14 and P18 stages EF was increased in the SK3^T/T mothers compared with WT dams (Fig. 2D). The SK3^T/T mothers' EFs increased to a greater extent during pregnancy than did those of the WT dams. These altered cardiac parameters in the SK3^T/T pregnant mice indicate that the SK3 channel influences cardiac function during pregnancy.

Fetal parameters.

We sought to determine whether the altered maternal cardiac and vascular changes measured in the pregnant SK3^T/T mice impacted fetal development. Using ultrasonography, we compared fetal HR, crown-rump length (fetal size), placental thickness, umbilical artery resistive index, and maternal uterine vessel resistive index between WT and SK3^T/T mice at P14 and P18. We observed a significant elevation in fetal HR in the P14 pups of SK3^T/T mothers compared with the pups of WT mothers (Fig. 3A). Additionally, fetal size and placental thickness in the P14 SK3^T/T pups were significantly decreased compared with the WT pups at the same gestational stage (Fig. 3, B and C), suggestive of intrauterine growth restriction. Despite the disparities seen at P14, no significant difference was observed at P18 for any of the measured parameters in SK3^T/T compared with WT fetuses.

Fig. 3.

Ultrasound measurement of fetal heart rate (A), fetal size (B), and placental thickness (C) from WT and SK3^T/T P14 mice. All significantly differed in SK3^T/T compared with WT. Values are means ± SE; n = 13–14. *P < 0.05, **P < 0.001 vs. P14 WT.

WT and SK3^T/T Litter Counts and Fetal Demise Rates

To determine whether these abnormalities in maternal cardiac function and vasculature affect fetal outcomes, we examined the fetuses by sonography at P14 and P18. We obtained multiple images of empty, fluid-filled sacs and sacs with debris (Fig. 4) that we correlated with evidence of fetal demise at time of maternal death. It was also interesting to note that the blood perfusion around the gestational sac of the fetal demise led away from the implantation sites, diverting the blood flow to other areas of the uterine horn.

Fig. 4.

Representative images of fetal demise and a viable pup by uterine sonography at P14, respectively.

After the ultrasound observation, we counted the number of pups in WT vs. SK3^T/T mice at various stages of pregnancy. Although litter sizes were similar in WT and SK3^T/T mice at P7–9, by P13–14 and P18–19, the number of viable pups in the SK3^T/T dams was significantly decreased compared with WT (Fig. 5A). Consequently, although fetal demise occurred prior to P14, reabsorption of the products of conception was incomplete by late pregnancy. The fetuses that survived were comparable in size to the WT fetuses at the time of birth, as there were no significant differences in wet weight of the fetuses isolated from P18 WT and SK3^T/T mothers (1.014 ± 0.03 g and 1.043 ± 0.05 g, respectively). Finally, we compared the percentage of mothers losing pups and observed that the percentage of SK3^T/T mice exhibited a higher rate of fetal loss, with 75% of dams exhibiting some fetal loss compared with only 32% of the WT dams (Fig. 5B). Moreover, 51% of the SK3^T/T implantation sites ended in fetal loss compared with only 8% of WT sites (Fig. 5B).

Fig. 5.

A: average pups per litter at P7–9, P13–14, and P18–19 for WT and SK3^T/T mice (n for each group is indicated under each data bar). Values are means ± SE. *P < 0.05,; **P < 0.001 vs. WT at the same gestation stage. B: percentage of SK3^T/T and WT mothers exhibiting at least one fetal demise and percentage of total implantation sites for SK3^T/T and WT mothers lost to fetal demise. *P < 0.05, **P < 0.001 vs. WT.

DISCUSSION

The discovery that SK3^T/T mice exhibit abnormal vascular morphology in the mesenteric circulation raised the question whether these abnormities were restricted to this vascular bed, or whether SK3 channels also influenced morphology and physiology of other organs or vascular beds. We observed that basal diameters of uterine arteries from SK3^T/T mice were increased compared with those of WT mice, and that these arteries have reduced receptor-mediated vasoconstriction to the thromboxane A₂ mimetic U-46619 but intact vasoconstriction to KCl. This implies that uterine arteries from SK3^T/T mice constrict normally in response to direct membrane depolarization but exhibit an attenuated response to G protein-coupled pathways. In the vasculature, SK3 channels are expressed exclusively in the endothelial layer (4, 28), which could account for the normal smooth muscle response to KCl (24). Despite the increases in diameter of uterine arteries in SK3^T/T mice, calcium-free solution had no effect on artery diameters of either WT or SK3^T/T mice. We attribute this difference in diameters between SK3^T/T and WT mice to structural differences between the two genotypes. Although myogenic tone has been shown to be regulated in uterine artery during pregnancy (10), the NP uterine arteries do not appear to have a substantial change. The mechanism by which these vascular changes affect both maternal cardiovascular remodeling and fetal outcomes remains unclear.

An inability to properly remodel the vasculature during pregnancy can cause placental abnormalities and fetal complications. During pregnancy in humans, maternal blood volume increases 45–50%, and the vascular system must adapt to these increases in volume without dramatic changes in blood pressure. Either low or high maternal blood pressure during pregnancy can result in poor fetal outcomes in humans; however, systolic blood pressures in the SK3^T/T mice were similar to WT mice (16, 29). A decreased peripheral vascular resistance in the SK3^T/T mice, as suspected from both uterine and mesenteric vessel changes, may decrease afterload of the heart, resulting in enhanced EF in these mice during pregnancy. However, despite the alterations in maternal cardiovascular parameters (Fig. 2), we detected no differences in umbilical flow velocity or resistive index of the maternal uterine vessels between WT and SK3^T/T mice during pregnancy (data not shown). As expected, left ventricular mass and EDV increased during pregnancy. However, the magnitude of increase was smaller in SK3^T/T than in WT mice. The increased HR in the SK3^T/T mice in late pregnancy could result in reduced ventricular filling yet maintain cardiac output and blood pressure. Alternatively, a 27% increase in blood volume during pregnancy in mice (17) also could underlie the dramatic differences observed in cardiac mechanics in WT vs. SK3^T/T mice during pregnancy that are not observed in the NP state. Finally, differences in diameter and number of vessels at the fetal-maternal interface after angiogenesis could contribute to the decrease in afterload, producing reflex increases in HR and left ventricular contractility.

Although maternal cardiac dysfunction alone does not account directly for the aberrations seen in fetal parameters, it appears to create a cascade of maladaptive events in development, starting with cardiac abnormalities in utero. Fetuses from the SK3^T/T mothers had significantly elevated HR by 14 days gestation. Early during pregnancy, pups of SK3^T/T mothers were growth restricted, perhaps as a result of altered placental perfusion, potentially due to overoxygenation too soon after placentation. During the early stages of pregnancy, the placenta develops in a low-oxygen environment due to endovascular cytotrophoblast cells that block uterine spiral arterioles, limiting blood flow to the fetus (16). Opening these arterioles later in pregnancy rapidly increases oxygen tension and initiates rapid growth and differentiation of the placenta, a step that is essential to maintain the fetus. Premature increases in oxygen concentration can lead to damage of DNA, lipids, and proteins and have detrimental fetal effects due to the low levels of antioxidants in the placenta during the first trimester (24).

Interestingly, there was no significant difference in either placental thickness or fetal weight at P18, indicating that the SK3^T/T fetal growth was able to catch up to WT by the time of birth. Similarly, there was no significant difference in the resistive index of the maternal vessels supplying the fetal sacs at either of the studied days of gestation despite the increase in uterine artery diameter. We posit that SK3^T/T mice adapted by creating a shunt around the site of placentation within the uterine vessels, thus decreasing the resistance and altering blood flow to the viable pups, as the fetal reabsorptions had already occurred by P14. A precedent for this has been observed during connection of the utero-placental vascularization in humans, when an arteriovenous shunt is created to allow for an increase in blood flow (26). In the SK3^T/T mice, reduced vascular tone in uterine vessels may direct ambient blood inflow away from the higher-resistance placental path to the lower-resistance uterine venous circuit (26), creating a relative physiological shunt. Evidence for reduced fetal nutrient blood flow includes elevated fetal HR, a sign of fetal distress, and this was detected in the SK3^T/T mice.

Despite abnormalities in the vasculature of SK3^T/T mice, the role of the SK3 channel in promoting fetal demise has not been examined. The reduction in the number of pups per litter from SK3^T/T mice demonstrates that ∼50% of their pups were being reabsorbed. At P7–9, the litter size of the SK3^T/T mice was similar to that of WT mice, but as gestation progressed, litter sizes were reduced by more than one-half by gestational days 13–14 in the SK3^T/T mice, often exhibiting fetal demise and resorption. This suggests that SK3 overexpression affects fetal development and not implantation. Demise occurs temporally with development of the placenta (P9.5), leading us to believe that survival of the SK3^T/T fetuses is dependent on proper placentation. Not only did the SK3^T/T mice have a higher percentage of pups lost, 75% of the dams had some fetal loss compared with only 32% of the WT mothers. The abnormal branching of mesenteric arteries and the enhanced basal diameter of uterine arteries suggest that defects in blood vessel development may be a generalized phenomenon in SK3^T/T mice and may underlie the reduction in fetal viability.

Potassium channels in endothelial cells contribute to angiogenesis and proliferation of blood vessels. Recently the intermediate conductance calcium-activated channel (IK1), a channel closely related to the SK3 channel, was shown to directly affect angiogenesis of human umbilical veins (11). Blocking IK1 resulted in reduction of vessel outgrowth in vivo. EDHF, another factor that regulates angiogenesis (18), was completely blocked by inhibiting SK3 and IK1 with apamin and TRAM-1 blockers, respectively (6). SK3 channels also regulate cell migration in breast cancer (22), which could contribute to endothelial cell migration, an essential step in angiogenesis (25). Further investigation is needed to determine the role that SK3 channels play in angiogenesis, which could greatly influence many organ systems and pathologies, including abnormal development of the maternal vasculature of the placenta.

Our findings indicate that alterations in SK3 channel expression or function may be an underlying factor in cardiovascular-related disorders during pregnancy. Recently, genome-wide studies identified an association between a single nucleotide polymorphism within an intron in KCNN3 (the SK3 channel gene) and lone atrial fibrillation in humans (8), supporting a role for SK3 channels in cardiovascular regulation, which could potentially affect pregnancy outcomes. The increased cardiac demands of pregnancy may be a contributing factor to exacerbated abnormalities in both the vascular morphology and altered cardiac function in SK3^T/T mice (28). The elevated HR of the SK3^T/T mice could be compensating for other cardiovascular defects. Doppler ultrasound indicates that an alteration in the blood flow in the placenta increases the rate of miscarriage and fetal loss (12). Whether abnormal uterine vascular tone would impact fetal outcomes is unknown. We did not detect a change in uterine artery resistive index; however, overexpression of SK3 channels in the smaller uteroplacental vessels may result in vasodilation and abnormally high flow to the fetus, which was undetectable by ultrasound. Consequently, this increase in blood flow could occur prematurely during pregnancy, causing a defect in placental development and leading to fetal demise. With a better understanding of the contribution of SK3 channels to both cardiac function and vessel formation and growth, we can improve our comprehension of the key mechanisms to target for prevention of fetal loss. Information gained about vascular remodeling and placental abnormalities will add significantly to our understanding of the pathophysiology of pregnancy loss and has the potential to direct therapies to prevent or limit pregnancy losses.

GRANTS

This work was supported by grants from the National Institute of Health (R01 HD-037831 and the American Heart Association (#12GRNT10990002) to S.K.England, the University of Iowa Clinical and Translational Science Program UL1 RR024979-03S3 to S. K. England and N. C. Bowdler, and S10RR026293 to R. M. Weiss) and the March of Dimes (21-FY08-566 to S. K. England). S. L. Pierce was supported by an American Heart Association Predoctoral Fellowship (09PRE2280322). This work was supported, in part, by resources and the use of facilities at the Department of Veterans Affairs Iowa City Health Care System, Iowa City, IA (BX000543-03 to K. G. Lamping).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: C.C.R., S.L.P., D.W.N., K.Z., N.C.B., and R.M.W. performed experiments; C.C.R., S.L.P., K.G.L., N.C.B., and S.K.E. analyzed data; C.C.R., S.L.P., K.G.L., N.C.B., R.M.W., and S.K.E. interpreted results of experiments; C.C.R., S.L.P., and S.K.E. prepared figures; C.C.R., S.L.P., and S.K.E. drafted manuscript; C.C.R. and S.K.E. approved final version of manuscript; S.L.P., K.G.L., N.C.B., R.M.W., and S.K.E. conception and design of research; S.L.P., K.G.L., N.C.B., R.M.W., and S.K.E. edited and revised manuscript.