INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

PREGNANCY is accompanied by decreased systemic vascular resistance (SVR) and contractile response to vasoconstrictors as well as increased vasodilator responsiveness (1, 9, 23). Since recognition of the importance of the endothelium in modulating vascular tone and reactivity, several studies have examined the contribution of increased endothelial nitric oxide (NO) production or activity to these pregnancy-associated changes. Supporting a role for NO are observations of increased NO biosynthesis and activity (23, 30), enhanced responsiveness to the endothelium-dependent vasodilator ACh (1, 32), and restoration of contractile responsiveness with NO inhibition (19, 31). However, against the likelihood that NO is responsible for all the pregnancy-associated alterations in vascular responsiveness are the observations that NO inhibition failed to reverse pregnancy-associated alterations in vasoconstrictor or vasodilator responses in some whole animal studies (20, 29) and in most nonuterine vascular beds (14, 27, 31).

Endothelium-dependent vasodilation is thought to be due to the release of NO and an as yet unidentified, nonprostanoid, endothelium-derived hyperpolarizing factor (EDHF) (3). The vasodilator action of EDHF as well as other synthetic and endogenous vasodilators operates through membrane hyperpolarization caused by opening of the ATP-sensitive K⁺ (K⁺_ATP) channel (28) and other K⁺ channels such as the calcium-activated K⁺ channel (13). Hyperpolarization of the mesenteric artery vascular smooth muscle cell membrane in pregnant rats has been reported (15), which could be due to changes in potassium conductance. The inability of NO synthase inhibitors to completely block ACh relaxation in vessels isolated from pregnant animals (14, 23) is consistent with an increase in EDHF.

The effect of pregnancy on vascular smooth muscle K⁺_ATP channels has not been previously studied. Previous studies have shown that pregnancy increases the expression of Na⁺ and Ca²⁺ channels in uterine smooth muscle. Furthermore, the expression of K⁺ channels in uterine smooth muscle varies during the estrus cycle and can be stimulated by estrogen treatment (2, 24). Vascular smooth muscle and endothelial cell K⁺_ATP channels have been implicated in vasodilator responses to various stimuli. However, the physiological role of K⁺_ATP channels in vasoregulation is incompletely understood. We hypothesized that pregnancy increased K⁺_ATP channel activity, which, in turn, was responsible for decreasing SVR and vasoconstrictor responsiveness during pregnancy. We therefore sought to determine whether inhibition of K⁺_ATP channel activity during pregnancy raised systemic and/or regional vascular resistance and increased vasoconstrictor response to ANG II compared with that observed in nonpregnant animals. ANG II was chosen because it is the vasoconstrictor whose effects have been the most consistently observed to be blunted during pregnancy (5, 9, 10). We treated equal numbers of nonpregnant and pregnant awake, intact guinea pigs with glibenclamide, a known K⁺_ATP channel inhibitor (7, 25), or vehicle. Systemic and regional (uterine, coronary, cerebral, renal) vascular resistance, blood pressure, and blood flows were compared in nonpregnant or pregnant animals before treatment with glibenclamide or vehicle, after treatment, and after treatment during ANG II infusion.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Animals consisted of nonpregnant and pregnant (day 54 ± 3, term = 63 days) female Hartley guinea pigs (Sasco, Omaha, NE) assigned at random to nonpregnant vehicle (n = 8), nonpregnant glibenclamide (n = 8), pregnant vehicle (n = 8), or pregnant glibenclamide (n = 8) groups. The number of days pregnant was estimated as the number of days from the observation of a vaginal plug. Total body weight was greater in the pregnant than the nonpregnant animals, but maternal body weight (total weight minus uterus and uterine contents) was similar in all groups (Table 1). An additional five guinea pigs (4 nonpregnant, 1 pregnant) were studied to determine the effectiveness of glibenclamide blockade. All guinea pigs ate chow (Prolab guinea pig formula, Agway) and drank water ad libitum before study. The University of Colorado Animal Care and Use Committee approved the care of the guinea pigs and the conduct of the study.

Surgery and instrumentation. Approximately 1 wk before study, catheters were implanted in each guinea pig under ketamine (50 mg/kg im) and xylazine (2.5 mg/kg im) anesthesia as previously described (5). A polyvinyl catheter (0.28 mm ID, 0.64 mm OD) was placed in the right carotid artery, and the tip was advanced into the left ventricle to inject radiolabeled microspheres for the measurement of regional blood flows. A polyethylene catheter (PE-50) was inserted into the right carotid artery to measure blood pressure. An additional polyethylene catheter and two polyvinyl catheters were introduced into the right jugular vein and advanced into the superior vena cava to infuse ANG II and measure cardiac output as previously described (5). Each catheter was routed subcutaneously, exteriorized, and protected inside a cap sutured to interscapular skin. Catheters were filled with chloramphenicol and heparinized saline to keep them patent and free of infection. Each guinea pig was eating and drinking normally before study.

Drug preparation. Glibenclamide (Sigma, St. Louis, MO) was dissolved in dimethyl sulfoxide (DMSO, Sigma) and administered at 3.5 mg/kg iv (4). When it is assumed that 70% of total guinea pig body weight is water, this dose approximates a 10 µM final concentration. An equal volume of vehicle (DMSO) was given. Cromakalim (Sigma) was infused at 5 µg · kg¹ · min¹ iv in accordance with previous studies assessing the efficacy of glibenclamide blockade (26). ANG II (human sequence acetate salt, Sigma) was infused at 540 ng · min¹ · kg total body wt¹ for 10 min, which we have shown previously to produce maximal vasoconstriction (9).

Experimental protocol. Each guinea pig was studied under three conditions: before treatment, after treatment, and after treatment during ANG II infusion. Before-treatment studies were performed after at least 30 min of quiet rest when arterial pressure had stabilized to within 2-3 mmHg and cardiac output to within 10%. Vehicle or glibenclamide was then administered and studies were repeated 20 min later, by which time blood pressure and cardiac output had stabilized. The final set of measurements was performed after 10 min of ANG II infusion, which we have shown previously to be sufficient for the rise in blood pressure to plateau (9).

Arterial pressure was monitored using a Statham 23 Db transducer referred to midthoracic (dorsoventral) level. Diastolic and systolic values were averaged over 1-min intervals and used to calculate mean arterial pressure using custom software in Pascal computer language and an Intel digital computer.

Cardiac output was measured by dye dilution (indocyanine green) as previously described (5). Measurements were repeated until duplicate values were obtained. Radiolabeled 15-µm-diameter microspheres (New England Nuclear, Boston, MA) were injected on completion of the cardiac output measurements to measure regional blood flows (5). One of three microspheres (⁴⁶Sc, ¹¹³Sn, ⁸⁵Sr) was injected into the left ventricle before treatment, after treatment, and after treatment during ANG II infusion. Approximately 2 µCi of each isotope were injected to obtain adequate suprabackground counts and the minimum number of microspheres necessary for precise flow determination. This amount of radioactivity corresponded to a minimum of 10⁵ microspheres/injection. The cardiac output in the vehicle-treated animals before treatment and during ANG II infusion agreed closely with previous reports (9, 10), suggesting that neither the left ventricular catheters nor the microspheres affected the cardiac output obtained. After the final microsphere injection, the guinea pig was killed by pentobarbital sodium overdose. The uterus, placentas, kidneys, heart, lungs, and brain were dissected free from adherent tissue and placed in vials, and radioactivity was counted [in counts per minute (cpm)] with an Auto-Gamma Scintillation Spectrometer (Packard Instrument, Laguna Hills, CA). Organ flows (ml/min) were calculated as [(cardiac output) × (organ cpm)]/(total injected cpm) and were expressed per organ or per 100 g organ weight. Blood flow to paired organs or tissues (kidney, brain) were within 1%, indicating adequacy of microsphere mixing. In the pregnant animals, uteroplacental blood flow was expressed as the sum of blood flow to the uterus, comprising myometrial and endometrial layers, and the placenta (maternal portion). An average of 10% of total uteroplacental flow went to the uterus, consistent with previous reports (19). SVR was calculated as mean arterial pressure divided by cardiac output, and organ vascular resistance was calculated as mean arterial pressure divided by organ blood flow or organ blood flow per 100 g organ weight.

Statistical analyses. Values are expressed as means ± SE in text, Tables 1 and 2, and Figs. 1-5. Effects of pregnancy, glibenclamide treatment, and ANG II infusion were determined using two-way ANOVA with repeated measures (General Linear Model), and comparisons of cell means were made by specific contrasts when the overall comparison was significant (SuperANOVA, Abacus Concepts, Berkeley, CA). Because all comparisons were specified in advance as part of a single analytical procedure, post hoc multiple-comparison corrections were not required. Differences were considered significant when the two-tailed P was <0.05 and were considered as trends when 0.05 < P < 0.10.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Comparison of animals assigned to the vehicle vs. glibenclamide groups before treatment. Groups were well matched for total or maternal body weight, days pregnant, litter size (Table 1), SVR, uterine and uteroplacental resistance and blood flow, coronary resistance and blood flow, renal resistance, renal blood flow in the nonpregnant animals, and cerebral blood flow and cerebral vascular resistance in the pregnant animals (see Figs. 1-5, Table 2). However, blood pressure was higher in the nonpregnant vehicle group and cardiac output was greater in the pregnant vehicle group than in the corresponding glibenclamide group (Fig. 1). Cerebral vascular resistance was also lower in the nonpregnant glibenclamide- than vehicle-treated group (see Fig. 5), and renal blood flow was lower in the pregnant glibenclamide-treated than vehicle-treated animals (Table 2).

View larger version (33K): [in this window] [in a new window]   Fig. 1.   Before treatment, pregnant (Preg) animals assigned to vehicle (Veh) group had lower systemic vascular resistance (SVR) and mean arterial pressure (MAP) than their nonpregnant (Nonpreg) counterparts, and pregnant animals assigned to vehicle group had higher cardiac output (CO). After treatment with 3.5 mg/kg iv glibenclamide (Glib), nonpregnant or pregnant animals had higher SVR and lower CO than Veh-treated animals and, in addition, higher MAP in the pregnant group. After treatment and during 540 ng · min1 · kg total body wt1 ANG II infusion, SVR and MAP rose and CO fell in all groups. SVR was higher in the Glib- than in the Veh-treated nonpregnant or pregnant groups, and MAP was higher and CO lower in pregnant animals only. Symbols: * P < 0.05, Glib vs. Veh; + P < 0.05, after treatment vs. before treatment, or after treatment plus ANG II vs. after treatment;  P < 0.05, nonpregnant Veh vs. pregnant Veh, or nonpregnant Glib vs. pregnant Glib.

Comparison of pregnant vs. nonpregnant groups before treatment. Pregnancy lowered SVR and blood pressure and raised cardiac output in the vehicle group. Values did not differ in the pregnant vs. nonpregnant animals assigned to the glibenclamide group (Fig. 1). Regardless of treatment group, the uteroplacental circulation of the pregnant animals had lower vascular resistance and higher blood flow and received a larger percent cardiac output than did the uterine circulation in the nonpregnant animals (Tables 1 and 2, 12.5 ± 1.2 vs. 1.0 ± 0.1%, respectively, P < 0.05). The rise in blood flow was proportional to the increase in tissue weight (Table 1) such that the nonpregnant animal uterine blood flow per 100 g tissue weight was similar to the pregnant animal uteroplacental flow per 100 g tissue weight (Table 2). Comparing the uterine (nonplacental) portion, resistance per 100 g tissue weight was higher and blood flow lower in the pregnant than in the nonpregnant animals (Fig. 2, Table 2). In the other circulations examined, vascular resistance and tissue weights did not differ consistently between nonpregnant and pregnant animals (Figs. 3-5).

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Glib treatment raised uterine vascular resistance per 100 g tissue wt above before-treatment values in nonpregnant (A) and pregnant guinea pigs (B, top). During ANG II infusion, Glib treatment increased uterine vascular resistance (B, top) or uteroplacental vascular resistance in pregnant animals (B, bottom). Symbols: * P < 0.05, Glib vs. Veh; + P < 0.05, after treatment vs. before treatment, or after treatment plus ANG II vs. after treatment;  P < 0.05, nonpregnant Veh vs. pregnant Veh, or nonpregnant Glib vs. pregnant Glib.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Glib treatment raised coronary vascular resistance above levels present in Veh-treated animals. Values remained higher during ANG II infusion in pregnant but not nonpregnant Glib-treated groups. * P < 0.05, Glib vs. Veh; + P < 0.05, after treatment vs. before treatment, or after treatment plus ANG II vs. after treatment;  P < 0.05, nonpregnant Veh vs. pregnant Veh, or nonpregnant Glib vs. pregnant Glib.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   In pregnant animals, Glib raised renal vascular resistance above pretreatment values to levels that tended (P = 0.09) to be higher than those of Veh-treated pregnant animals. ANG II infusion increased renal vascular resistance in Veh- but not Glib-treated nonpregnant or pregnant animals. + P < 0.05, after treatment vs. before treatment, or after treatment plus ANG II vs. after treatment.

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Glib treatment increased cerebral vascular resistance in pregnant but not nonpregnant animals. Vascular resistance rose during ANG II infusion in Glib- and Veh-treated pregnant animals. * P < 0.05, Glib vs. Veh; + P < 0.05, after treatment vs. before treatment, or after treatment plus ANG II vs. after treatment;  P < 0.05, nonpregnant Veh vs. pregnant Veh, or nonpregnant Glib vs. pregnant Glib.

Effects of glibenclamide treatment in nonpregnant and pregnant animals. In a separate series of five nonpregnant or pregnant guinea pigs, the glibenclamide dosage employed (3.5 mg/kg) reversed a 19.4% fall in SVR produced by the K⁺_ATP channel agonist cromakalim (5 µg · kg¹ · min¹). Higher dosages of glibenclamide (20-26 mg/kg) raised SVR substantially above levels observed at the 3.5 mg/kg dose and terminated in the death in one animal.

Glibenclamide raised SVR, increased blood pressure, and decreased cardiac output in the pregnant animals (Fig. 1). Glibenclamide also raised SVR, lowered cardiac output, and increased blood pressure compared with before-treatment values in the nonpregnant animals, but blood pressure in the nonpregnant glibenclamide-treated animals did not exceed values in the vehicle-treated group. Pregnant glibenclamide-treated animals had higher blood pressures than the nonpregnant glibenclamide-treated group (Fig. 1).

In the uterine circulation, glibenclamide raised vascular resistance and lowered blood flow in the nonpregnant animals (Fig. 2, Table 2). In the pregnant animals, glibenclamide raised vascular resistance and lowered blood flow in the uterine portion of the uteroplacental vascular bed. Uteroplacental resistance or flow did not change, implying that vasoconstriction occurred in the uterine and not the placental portion of the uteroplacental vascular bed (Fig. 2, Table 2).

In the coronary circulation, glibenclamide increased vascular resistance and decreased blood flow in the nonpregnant and pregnant animals (Fig. 3, Table 2). The 62% vascular resistance increase in the pregnant group above vehicle values was greater than the 44% increase in the nonpregnant group (P < 0.05), but the blood flow changes were similar. Heart rate fell to achieve similar values in the pregnant and nonpregnant glibenclamide-treated groups (Table 2).

In the renal circulation, glibenclamide did not change vascular resistance in the nonpregnant animals (Fig. 4). In the pregnant animals, glibenclamide raised renal vascular resistance and lowered renal blood flow when compared with before-treatment values. Flow was lower and resistance also tended (P = 0.09) to be greater in the pregnant glibenclamide vs. vehicle-treated groups (Fig. 4, Table 2).

In the cerebral circulation, glibenclamide raised vascular resistance in pregnant but not nonpregnant animals (Fig. 5). Flow was unchanged after glibenclamide treatment in the pregnant or nonpregnant groups (Table 2).

Effects of ANG II infusion in vehicle- vs. glibenclamide-treated, nonpregnant or pregnant groups. ANG II increased SVR, raised blood pressure, and lowered cardiac output in all groups (Fig. 1). Glibenclamide augmented the ANG II-induced rise in SVR and did so to a similar extent in the nonpregnant and pregnant groups (SVR from after-treatment values = 0.36 ± 0.07 and 0.32 ± 0.09 mmHg · ml¹ · min, respectively, P = NS). Whereas SVR was lower during ANG II infusion in the vehicle-treated pregnant than nonpregnant group, SVR was similar in glibenclamide-treated pregnant and nonpregnant animals (Fig. 1).

In the uterine circulation, ANG II infusion did not change vascular resistance in nonpregnant animals. In pregnant animals, ANG II infusion raised uterine vascular resistance in the vehicle- and glibenclamide-treated groups and increased uteroplacental vascular resistance in the glibenclamide group (Fig. 2). Blood flow fell per 100 g tissue weight during ANG II infusion in the nonpregnant vehicle group but remained unchanged in the other groups (Table 2).

Coronary vascular resistance did not rise during ANG II infusion in any group. Values remained higher in the pregnant glibenclamide-treated group than in the pregnant vehicle-treated animals or the nonpregnant glibenclamide-treated group (Fig. 3). Coronary blood flow rose during ANG II infusion in the nonpregnant glibenclamide-treated group but otherwise was unchanged (Table 2).

In the renal circulation, ANG II increased vascular resistance and reduced blood flow in the nonpregnant animals (Fig. 4, Table 2). In pregnant animals, ANG II raised renal vascular resistance and diminished blood flow in the vehicle-treated group. Renal vascular resistance tended (P = 0.06) to rise in glibenclamide-treated pregnant animals, but blood flow was unchanged (Fig. 4, Table 2).

In the cerebral circulation of the nonpregnant animals, ANG II did not change vascular resistance or blood flow (Fig. 5, Table 2). In the pregnant animals, ANG II infusion increased cerebral vascular resistance in the vehicle- and glibenclamide-treated groups but did not change blood flow.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Decreased vascular contractility during pregnancy is observed in multiple vascular beds by using in vitro as well as in vivo preparations, suggesting that the responsible mechanism involves a change in a factor intrinsic to the vessels or a circulating substance that invokes a stable "switch" or change in vasoreactivity that is retained in vitro. We hypothesized that enhanced K⁺_ATP channel activity was a mechanism by which this switch occurred. We found that glibenclamide, a K⁺ channel blocker with relative specificity for the K⁺_ATP channel, raised systemic and regional vascular resistance in selected vascular beds (uterine and coronary), and SVR response to ANG II infusion in nonpregnant and pregnant guinea pigs. This suggested that K⁺_ATP channel activity modulates systemic, uterine, and coronary vascular resistance and opposes ANG II-induced systemic vasoconstriction and does so similarly in nonpregnant and pregnant guinea pigs. However, glibenclamide increased coronary vascular resistance relative to levels seen in vehicle-treated animals to a greater extent in the pregnant than nonpregnant group and increased renal and cerebral vascular resistance and the uteroplacental resistance response to ANG II in the pregnant animals only. Thus pregnancy appears to augment K⁺_ATP channel activity in the coronary, renal, cerebral, and uterine vascular beds and in the uteroplacental circulation during ANG II infusion. Taken together, these observations suggest that increased K⁺_ATP channel activity during guinea pig pregnancy influences regional control of vascular resistance but cannot account for the characteristic decrease in SVR and ANG II-induced vasoconstrictor response.

Glibenclamide is the most potent of the sulfonylurea drugs with relatively specific K⁺_ATP channel inhibitory effects. It is thought to act at a site in the cell membrane that is different from that at which ATP binds (22). Glibenclamide has been shown to inhibit numerous K⁺_ATP activators (7, 25) and to increase the vasoconstrictor response to BAY K 8644 (a Ca²⁺ entry-promoting agent), suggesting that glibenclamide does not interfere with Ca²⁺ entry through dihydropyridine voltage-dependent channels (16). Although some studies have employed higher dosages of glibenclamide (8, 16), those approaching 20-30 mg/kg proved toxic in our study and previous studies (25). We found that the fall in SVR after infusion of the K⁺_ATP channel agonist cromakalim was fully reversed in nonpregnant as well as pregnant animals by the 3.5 mg/kg glibenclamide dosage employed. Because we did not compare the effectiveness of glibenclamide blockade to a range of cromakalim doses, we were not able to determine if the sensitivity to K⁺_ATP channel blockade differed in the pregnant vs. nonpregnant animals. However, our dosage was within the range required to produce K⁺_ATP channel inhibition in isolated vascular smooth muscle membrane patches (21) and the same as that used by other investigators who showed that it blocked the effects of the K⁺_ATP channel agonists in intact lamb, isolated perfused lung, and isolated vessel preparations (4, 12). Thus we concluded that the data obtained supported the likelihood that effective blockade was achieved in the pregnant and nonpregnant groups.

In agreement with previous reports, SVR was lower and the SVR response to ANG II was diminished in the pregnant compared with the nonpregnant vehicle-treated groups. After glibenclamide treatment, SVR and SVR response to ANG II no longer differed in the pregnant and nonpregnant groups, suggesting that lower SVR and response to ANG II resulted from greater K⁺_ATP channel activity. However, despite random group assignment, the pregnant and nonpregnant animals in the glibenclamide groups had similar SVR before treatment. Thus it was unclear whether the similarity in SVR and SVR response to ANG II in the pregnant and nonpregnant groups after glibenclamide treatment was due to glibenclamide or the retention of similarities present before treatment. The latter possibility seemed unlikely, given the large rise in SVR after treatment and with ANG II infusion. The similarity in SVR after treatment and during ANG II infusion, however, indicated that the effect of glibenclamide was not greater in the pregnant than the nonpregnant animals. We therefore concluded that K⁺_ATP channel activity contributed to the determination of SVR and helped to oppose ANG II-induced vasoconstriction but did so to a similar degree in pregnant and nonpregnant guinea pigs.

The effects of pregnancy on vascular resistance and vasoconstrictor response vary among vascular beds in ways that are likely important for the distribution of blood flow (14, 23). K⁺_ATP channel activity also varies among tissues (22). We were therefore interested in the possibility that pregnancy-associated alterations in regional vascular resistance or vasoconstrictor response to ANG II could be attributed to differences in K⁺_ATP channel activity. Because microsphere-based measurements of blood flow can only be performed a limited number of times, we measured regional blood flow and vascular resistance at the dose of ANG II producing maximal vasoconstriction (9). This meant, however, that we were not able to address the possibility of differences in ANG II sensitivity among the various vascular beds. In keeping with previous reports (5, 8, 23), we interpreted a rise in vascular resistance and blood pressure as demonstrating vasoconstriction. An accompanying fall in organ blood flow was interpreted as resulting from the regional vasoconstriction. We recognized that blood flow could also be unchanged, as the result of autoregulation or other factors serving to raise regional resistance in proportion to the increase in pressure, or increased if greater vasoconstriction occurred in other vascular beds. A problem for this as well as other studies is that only systemic arterial pressure could be measured and this value might not always correspond to the pressure present in the particular vascular bed.

The uterine circulation underwent a reduction in vascular resistance and a rise in blood flow with pregnancy that were proportional to the increase in organ weight. The addition of new vessels as well as the expansion and remodeling of existing vessels and the vasodilatory effects of prostacyclin, NO, and other substances (see Ref. 23 for review) likely contributed to this pregnancy-associated fall in uteroplacental vascular resistance. Our study indicates that K⁺_ATP channel activity also contributes, as demonstrated by the greater rise in uterine vascular resistance after glibenclamide treatment in the pregnant than in the nonpregnant animals. The apparent increase in K⁺_ATP channel activity was confined to the uterine portion of the vascular bed and did not yield greater net uteroplacental K⁺_ATP channel activity. Characteristically low uterine or uteroplacental vascular resistance was maintained in the vehicle-treated animals during ANG II infusion. In pregnancy, greater K⁺_ATP channel activity in the uterine and placental portions of the uteroplacental vascular bed contributed to the lack of ANG II-induced vasoconstrictor response as demonstrated by the resistance rise after K⁺_ATP channel inhibition. Thus we concluded that K⁺_ATP channel activity modulates uterine vascular resistance and opposes ANG II-induced uteroplacental vasoconstriction so as to help maintain blood flow during pregnancy.

In the coronary circulation, K⁺_ATP channels are important in the maintenance of vascular tone, having been shown to influence membrane resting potential in isolated coronary artery smooth muscle cells (17) and to regulate the vasodilator response to acute hypoxia and ischemia in the isolated, perfused guinea pig heart (6). The present, whole animal studies are consistent with these reports, showing that glibenclamide increased coronary vascular resistance and decreased coronary blood flow in nonpregnant as well as pregnant guinea pigs. The fall in cardiac output after glibenclamide treatment in this and other studies (8) likely resulted from increased afterload stemming from vasoconstriction in the systemic vascular bed. Negative inotropic effects could also have been involved although higher glibenclamide dosages are required to inhibit K⁺_ATP channels in the myocardium than in vascular smooth muscle (22). K⁺_ATP channels appeared to exert a greater influence on coronary vascular resistance in the pregnant than the nonpregnant animals as demonstrated by higher coronary vascular resistance after K⁺_ATP channel inhibition. The higher values in pregnant animals were not the result of lower heart rates, acting to decrease myocardial oxygen demand and raise coronary vascular resistance, because heart rate was similar in the two groups. Higher coronary vascular resistance was sustained and blood flow remained lower during ANG II infusion in the glibenclamide-treated pregnant than nonpregnant animals, indicating that K⁺_ATP channel activity helped maintain coronary blood flow. It is possible that glibenclamide had caused maximal constriction such that no further response to ANG II was possible. Blood flow returned to before-treatment values in the glibenclamide-treated nonpregnant and pregnant groups, suggesting autoregulation occurred during ANG II infusion in the coronary circulation. We concluded that K⁺_ATP channel activity contributed to the regulation of coronary vascular tone and blood flow in the nonpregnant as well as particularly in the pregnant guinea pig.

Increased K⁺_ATP channel activity during pregnancy also appeared to modulate renal and cerebral vascular resistance. K⁺_ATP channel inhibition doubled renal vascular resistance and halved blood flow in the pregnant animals. However, renal vascular resistance was not lower in the pregnant than in the nonpregnant animals before treatment, suggesting that the increased K⁺_ATP channel activity may have been opposed by some other vasoconstrictor(s). The marked rise in renal vascular resistance during ANG II infusion was somewhat diminished in the glibenclamide-treated group, suggesting that K⁺_ATP channel activity did not oppose ANG II-induced vasoconstriction. Possibly, glibenclamide had already sufficiently constricted the pregnant renal vessels so that little further change was possible. In the cerebral circulation, the pregnancy-associated increase in K⁺_ATP channel activity was accompanied by lower cerebral vascular resistance, suggesting that K⁺_ATP channel activity helps maintain low cerebral vascular resistance, but glibenclamide did not affect the cerebral vascular resistance response to ANG II differently in the pregnant vs. the nonpregnant animals. Cerebral blood flow remained unchanged by glibenclamide or ANG II, indicating perhaps that autoregulation served to maintain flow. Thus it appeared that increased K⁺_ATP channel activity during pregnancy contributes to the maintenance of normal renal and low cerebral vascular resistance but does not oppose ANG II-induced vasoconstriction in either vascular bed.

Despite considerable study, the mechanisms by which pregnancy decreases SVR and vasoconstrictor response remain elusive. Pregnancy stimulation of endothelial NO production and/or activity has been linked to increased endothelium-dependent vasodilation in several studies (see Ref. 23 for review). However, the considerable heterogeneity among vascular beds and uncertainty as to whether changes in NO activity fully account for the effects of pregnancy make it likely that additional factors are involved. The present studies, to the best of our knowledge, are the first to investigate the contribution of increased K⁺_ATP channel activity to these pregnancy-associated changes. Membrane hyperpolarization in pregnant compared with nonpregnant mesenteric vascular smooth muscle cells has been reported (15). This membrane hyperpolarization could be due to endothelium-derived hyperpolarizing factor (EDHF) acting through K⁺_ATP channels (3, 28). Although other types of K⁺ channels could also be involved, the availability of a relatively specific and well-tolerated K⁺_ATP channel inhibitor (glibenclamide) led us to consider whether pregnancy increased systemic and/or regional K⁺_ATP channel activity. Our data indicate that K⁺_ATP channel activity is important in general in the guinea pig, helping to maintain systemic and regional (uterine and coronary) vascular resistance and for opposing ANG II-induced vasoconstriction. Pregnancy selectively increased K⁺_ATP channel activity in the uterine, renal, coronary, and cerebral vascular beds and in the uteroplacental vasculature during ANG II infusion. Despite this evidence of regional stimulation of K⁺_ATP channel activity during pregnancy, the similarity in the systemic effects of K⁺_ATP channel inhibition in pregnant and nonpregnant animals led us to conclude that increased K⁺_ATP channel activity did not account for the decreased SVR and vasoconstrictor responsiveness of pregnancy. However, pregnancy stimulation of K⁺_ATP channel activity is likely important for the regulation of vascular resistance and maintenance of blood flow in these regional vascular beds. The pregnancy-specific effects of glibenclamide that we observed were likely due primarily to the blockade of vascular smooth muscle K⁺_ATP channels because inhibitory actions of glibenclamide are seen in endothelial cell-free preparations (22). However, because glibenclamide also partially inhibits flow-associated endothelial cell ATP release (11), we cannot exclude the possibility of an endothelial site of action as well. Further studies in isolated vessel or cell preparations from the vascular beds in which pregnancy-specific effects of K⁺_ATP channel inhibition were observed will be useful for determining whether increased K⁺_ATP channel activity during pregnancy is due to direct effects on vascular smooth muscle cells or is modulated indirectly through the endothelium.