INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ESTROGEN THERAPY BENEFITS women in the prevention of cardiovascular disease; however, it is unclear how this benefit is derived (22, 26). One possibility is through estrogen-mediated increases in blood flow and vasorelaxation. This is supported by observations that estrogen increases coronary blood flow in vivo (5, 28, 30, 35) and coronary artery relaxation in vitro (11, 25, 43). This effect of estrogen may be mediated by activation of endothelial nitric oxide synthase (eNOS), which increases nitric oxide (NO) synthesis, and is associated with enhanced smooth muscle synthesis of guanylyl cyclase (1, 16, 22-24, 44). However, this relaxation in coronary arteries may in part be endothelium independent (11, 25, 43), suggesting involvement of other mechanisms. Recently, White et al. (43) identified a large-conductance, calcium-dependent potassium channel (BK_Ca) in porcine coronary artery myocytes that was rapidly activated by estradiol-17 (E₂) through a cGMP-dependent mechanism involving type II inducible nitric oxide synthase (iNOS; see Ref. 4). Wellman et al. (42) also observed BK_Ca in rat coronary myocytes, but activation by E₂ was reported to be endothelium dependent. It is unclear if this reflects species differences, if estrogen activates both nitric oxide synthase (NOS) and BK_Ca independently, and, if so, whether these responses are interrelated.

Estrogen also may be responsible for the uterine and systemic vasodilation characteristic of normal pregnancy (33). Its effects on the uterine vascular bed were first studied by Markee (21) in 1932, who observed that increases in circulating estrogen were associated with hyperemia of ocular endometrial explants. Subsequent investigators reported in ovariectomized nonpregnant ewes that uterine blood flow (UBF) rose 50-150% within 2 h of a systemic dose of E₂ (6). However, the potency of E₂ as a vasodilator was demonstrated by Killam et al. (13). In these studies, acute E₂ exposure increased UBF >10-fold within 90-120 min in a reproducible and characteristic pattern in unstressed, ovariectomized nonpregnant ewes. Rosenfeld et al. (20, 35) confirmed these results and demonstrated that E₂ also increased blood flow to several nonreproductive tissues, including the myocardium, and this paralleled increases in cardiac output and decreases in systemic vascular resistance. Although several agents increase UBF transiently (32), none result in the prolonged rise that follows E₂ exposure, and the mechanisms involved remain unclear.

The uterine vascular bed can be isolated and studied in intact animals remote from surgery and other stresses, providing an excellent model in which to investigate the mechanisms responsible for estrogen-mediated vasodilation (33, 34). Cycloheximide inhibits acute E₂-mediated uterine vasodilation (13), whereas actinomysin D has no effect (29, 31). Thus posttranscriptional, nongenomic mechanisms appear to mediate acute uterine vascular responses to E₂. In this model, locally infused nitro-L-arginine methyl ester (L-NAME) dose dependently inhibits the acute E₂-induced uterine vasodilation and the parallel rise in uterine cGMP secretion, suggesting that E₂ enhances local NOS activity (34). At the time of maximum UBF, L-NAME-induced inhibition is rapidly reversed by L-arginine, further demonstrating a role for NO in uterine vascular responses to E₂ (34). However, L-NAME-mediated inhibition does not exceed 60-65% (34, 40), suggesting that NO "contributes" to E₂-induced increases in UBF and that additional factors are involved. Thus we designed studies using uterine artery myocytes and ovariectomized nonpregnant sheep to determine whether 1) the BK_Ca is present in uterine artery myocytes, 2) these channels are involved in mediating E₂-induced uterine vasodilation, and 3) E₂-mediated increases in UBF are due to an interaction between NO and BK_Ca. We also examined the precise time at which E₂-induced increases in UBF began to delineate possible genomic and nongenomic mechanisms.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Tissue collection and cell isolation procedures. Oophorectomized nonpregnant ewes were killed with 120 mg/kg iv of pentobarbital sodium. The abdomen was opened, the uterus was removed in block, and first- through fourth-generation uterine arteries were rapidly dissected from each horn. Arteries were placed in chilled sterile physiological saline solution, cleaned of excess adventitia and intraluminal blood, and transported for cell isolation. Myocytes were isolated from first- and second-generation arteries as previously described (43). Briefly, the adventitia was carefully dissected away, and the endothelium was removed. Each artery was cut into 1-mm strips that were placed in test tubes containing dissociation media [in mM: 137 NaCl, 5.6 KCl, 1 MgCl₂, 0.1 CaCl₂, 0.42 NaHPO₄, 0.44 NaH₂PO₄, 4.2 NaHCO₃, and 10 HEPES, pH 7.4 (NaOH)]. Vascular strips were incubated at 37°C in 2 ml of the dissociation solution with papain (26 U/ml) and dithiothreitol (1 mg/ml) for 35 min at 37°C. The papain solution was then discarded, and the tissues were incubated with dissociation medium containing collagenase (2 U/ml), elastase (75 U/ml), and soybean trypsin inhibitor (1 mg/ml) for 20 min at 37°C. Tissue was then triturated gently in enzyme-free dissociation medium, and the cells were pelleted by centrifugation at 500 g for 6 min at 4°C. The pellet was suspended in fresh medium and kept at 4°C. Experiments were performed within 6-8 h after cell dissociation.

Patch-clamp studies. For cell-attached patches, several drops of cell suspension were placed in a recording chamber (Warner Instruments) containing a solution of the following composition (in mM): 140 KCl, 10 MgCl₂, 0.1 CaCl₂, 10 HEPES, and 30 glucose (pH 7.4, 22-25°C). Single potassium channels were measured in cell-attached patches by filling the patch pipette (2-5 M) with Ringer solution of the following composition (in mM): 110 NaCl, 5 KCl, 1 MgCl₂, 2 CaCl₂, and 10 HEPES; a gigaohm seal was made on a single myocyte. Voltage across the patch was controlled by clamping the cell at 0 mV with the high-concentration extracellular potassium solution described above. Currents were filtered at 2 kHz and digitized at 10 kHz. Average channel activity (NP_o) in patches with multiple BK_Ca channels was determined as described (43). Experiments were performed to determine the effects of E₂ on channel activity and to assess the role of guanylyl cyclase using the inhibitor LY-83583 and 8-bromo-cGMP (Sigma Chemical, St. Louis, MO).

Animal model. The animal model used in the in vivo studies has been described in detail (13, 20, 34). In brief, nonpregnant ewes of mixed Western breed were fasted overnight but were allowed access to water. In the morning, animals were given atropine sulfate intramuscularly, and a percutaneous venous jugular catheter was placed for administration of preanesthetic pentobarbital sodium and ketamine hydrochloride. Animals were intubated, surgically prepared, and administered isoflurane (Mallinckrodt Veterinary, Mundelein, IL) and oxygen via a rebreathing anesthesia machine. Animals were ovariectomized through a midline abdominal incision, and 3.0- to 3.5-mm (ID) electromagnetic flow probes (Carolina Medical, King, NC) were implanted on the middle uterine artery of each uterine horn proximal to the first bifurcation. Polyvinyl catheters containing heparinized saline (100 U/ml) were implanted retrograde 2 cm into a distal branch of the uterine artery of each uterine horn for local intra-arterial infusion of drugs. The abdomen was closed, and, via a groin incision, polyvinyl catheters were implanted in the femoral artery and vein to the level of the abdominal aorta and lower vena cava, respectively. Animals received antibiotics on the day of surgery and the next two days as well as banamine (Schering-Plough Animal Health, Union, NJ) for pain. All animals were allowed 5 days for postoperative recovery. These studies were approved by the Institutional Review Board for Animal Research at the University of Texas Southwestern Medical Center at Dallas.

Experimental protocols. Two protocols were used in these studies. In the first, we determined if tetraethylammonium chloride (TEA; Sigma Chemical), a selective inhibitor of BK_Ca at submillimolar concentrations (27), infused directly into the uterine circulation before systemic E₂ altered baseline UBF and E₂-induced uterine vasodilation. Six nonpregnant ewes were included in these studies, and experiments were performed in each uterine horn if the catheters were patent and the blood flow probes were functional. To establish the presence of maximum UBF responses to systemic E₂, animals were administered 1 µg/kg E₂ (Steraloids, Wilton, NH) via the femoral venous catheter over 1 min beginning on the 5th postoperative day while continuously monitoring UBF, mean arterial pressure (MAP), and heart rate for 120 min. This dose of E₂ increases UBF 5- to 10-fold at 90-120 min and heart rate 15-20% without changing MAP. These responses are reproducible every 24 h (13, 33, 34) and result in plasma concentrations associated with the onset of parturition (18). Studies were begun after maximal UBF responses to E₂ were observed on 2 to 3 consecutive days. On the day of study, a continuous TEA infusion was initiated via a uterine artery catheter after a 30-min control period and was maintained for 120 min. Doses of TEA were randomly chosen and calculated to result in uterine arterial concentrations ranging from 0.05 to 0.6 mM. Arterial concentrations were estimated from the rate of TEA infused in micrograms per minute divided by baseline measurements of UBF in milliliters per minute (34). After 30 min of local TEA infusion, E₂ (1 µg/kg) was systemically infused via the femoral venous catheter over 1 min. Continuous recordings of UBF, MAP, and heart rate were initiated 30 min before the infusion of TEA and were maintained until 90 min after stopping the TEA infusion. After each TEA study, UBF responses to E₂ in the absence of TEA were performed daily until responses resembled those seen before TEA treatment. The TEA study was repeated using another randomly selected dose until responses to five to six doses had been examined in each animal.

Hemodynamic measurements. MAP in the lower abdominal aorta was monitored continuously via a femoral arterial catheter connected to a pressure transducer (type 4-327-0109; Bell and Howell, Pasadena, CA). Heart rate was determined from the phasic signal derived from the arterial pressure monitor. UBF was monitored continuously with square-wave electromagnetic flowmeters (model FM501; Carolina Medical). All measurements were continuously recorded on a six-channel pen recorder (model 3000; Gould, Cleveland, OH).

Statistical analysis. Repeated-measures ANOVA was used to examine changes over time. When significance was observed, Student-Newman-Keuls test was used to determine differences between groups at P < 0.05. Regression analysis by the least squares method was used to determine changes across doses. Where indicated, Student's t-test was used to compare groups. Values are presented as means ± SE.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Potassium channel identification and effects of E₂. Because there are no previous studies measuring single potassium channel activity in myocytes from ovine uterine arteries, our initial patch-clamp studies characterized single-channel activity in cell-attached patches on isolated myocytes. These studies revealed that electrical activity in these membrane patches was dominated by large-amplitude, single-channel openings carrying outward current (Fig. 1A). The activity of this channel was stimulated by increasing the concentration of calcium at the cytoplasmic surface of the membrane as shown in Fig. 1B. Recordings from cell-attached patches revealed minimal channel activity; however, channel gating was increased dramatically when these same patches were excised into an inside-out configuration where "intracellular" calcium concentration was now 100 µM (n = 4). Further biophysical analysis of channel activity demonstrated a microscopic current-voltage relationship with a unitary conductance of 107 ± 7 pS (n = 3-6 cells/point) in physiological gradients of potassium (Fig. 1C). Therefore, these studies identified this channel as the large-conductance, calcium- and voltage-activated potassium (BK_Ca) channel. No other potassium channel exhibits these specific biophysical and pharmacological characteristics.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Expression of calcium-dependent potassium (BKCa) channels in myocytes from ovine uterine arteries. A: traces from the same cell-attached patch recorded at various membrane potentials, as indicated. Channel openings are upward deflections from baseline. Dotted lines indicate channel closed state. B: recordings from the same membrane patch in the cell-attached (on-cell) configuration and then immediately after excision in the inside-out configuration. The concentration of calcium in the bath solution was 100 µM. C: mean single-channel current-voltage relationship obtained from cell-attached membrane patches in physiological gradients of potassium. Each point represents the mean of 3-6 experiments ± SE. Channel conductance was measured as the slope of the line calculated by linear regression (107 ± 7 pS).

View larger version (8K): [in this window] [in a new window]   Fig. 2.   Estradiol-17 (E2) stimulates BKCa channel activity in single uterine artery myocytes. Representative recordings from a cell-attached patch before (A) and 45 min after (B) addition of 100 nM E2. Each tracing is 300 ms of continuous channel activity at +40 mV. Channel openings are upward deflections from the baseline (closed) state, as indicated by the dotted line.

View larger version (13K): [in this window] [in a new window]   Fig. 3.   Estrogen opens BKCa channels via cGMP. Recordings from the same cell-attached patch (+40 mV) before and 30 min after exposure to E2 (1 µM) and then after cumulative addition of 20 µM LY-83583 (15 min). Further addition of 1 mM 8-bromo (Br)-cGMP (25 min) restored channel activity. Channel openings are upward deflections from the baseline (closed) state (indicated by dotted line). Periods of drug exposure are indicated by the lines above the traces.

Early effects of E₂. Although Killam et al. (13) characterized the pattern of the uterine vascular responses to systemic E₂ in 1973, demonstrating a maximum rise in UBF by 90-120 min that persisted for 8-12 h, no one has carefully characterized the early part of this response to determine when the rise in UBF actually begins. This is important as it will define the rapidity of the UBF response in vivo and potentially differentiate between genomic and nongenomic vascular responses to E₂. To address this, the sensitivity of the recorder was increased to more accurately measure UBF in the first 60 min after E₂ infusion (19). As anticipated (13, 33), UBF rose from 24.5 ± 1.9 to 134 ± 7.0 ml/min (P < 0.0001, n = 34) 90 min after E₂ (1 µg/kg) and was associated with a rise in heart rate from 69.5 ± 1.9 to 82.7 ± 1.9 beats/min (P < 0.0001) but no change in MAP. When we examined UBF at 5, 10, 30, and 60 min after E₂, UBF was unchanged until 30 min, after which it gradually rose 3.5-fold from 26.8 ± 1.9 to 86.7 ± 5.1 over the next 30 min (Fig. 4).

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Responses of ovine uterine blood flow (UBF) to systemic E2 in the first 60 min after infusion. Blood flow was unaffected until 30 min after E2, at which time UBF began to gradually rise. Data were analyzed using repeated-measures ANOVA.

Effects of TEA on E₂ responses. Because the patch-clamp studies identified the BK_Ca in uterine artery myocytes and because this channel responded to E₂ exposure, we examined its role in E₂-mediated increases in UBF in ovariectomized nonpregnant ewes. Estimated concentrations of TEA ranging from 0.05 to 0.6 mM, concentrations known to be specific for BK_Ca inhibition (27), were randomly infused through a uterine artery catheter for 120 min as previously described (34). Thirty minutes after an intra-arterial TEA infusion was initiated, E₂ (1 µg/kg iv) was systemically infused. Representative experiments are shown in Fig. 5, where two different doses were infused in alternate uterine horns 48 h apart. Whereas TEA attenuated the UBF response to E₂ in the infused uterine horn, the response in the contralateral uterine horn was unaffected. Furthermore, the inhibitory effects of TEA were not evident 24 h later, demonstrating the reversibility of channel blockade and the absence of a toxic tissue response to local TEA infusion. Thirty-seven infusions were performed in six ewes, each animal receiving five to seven doses using one or both uterine horns. Ipsilateral basal UBF was unaffected by TEA at any dose studied; values were 25.8 ± 2.0 and 23.0 ± 1.9 ml/min (P > 0.1) before and after 30 min of TEA infusion, respectively. However, local TEA dose dependently inhibited the E₂-induced increases in UBF (Fig. 6; P < 0.0001) compared with responses obtained in the absence of TEA on the prior day. Of note, at an estimated arterial concentration of 0.6 mM TEA, inhibition averaged 67 ± 11%. Local TEA had no effect on contralateral UBF responses, heart rate, or MAP, with the first two increasing significantly 90 min after E₂ (Table 1).

View larger version (24K): [in this window] [in a new window]   Fig. 5.   Representative tracings of the uterine vascular response to a systemic infusion of E2 (1 µg/kg) and the ipsilateral inhibitory effects of local continuous intra-arterial infusions of tetraethylammonium ion (TEA). A: acute E2-mediated rise in UBF in both uterine horns in the absence of TEA. B: unilateral inhibitory effects of a continuous right-sided TEA infusion (0.12 mM) on E2-induced uterine vasodilation. C: return of the E2 response on the right at 48 h and the unilateral left-sided inhibition by TEA (0.16 mM) at that time.

View larger version (14K): [in this window] [in a new window]   Fig. 6.   Dose-inhibition curve demonstrating the percentage inhibition of E2-induced increases in UBF by local intra-arterial infusions of TEA. The percentage inhibition was determined from responses to systemic E2 observed 24 h before the local infusion of TEA. Data were analyzed by linear regression analysis.

View this table: [in this window] [in a new window]   Table 1.   Comparison of effects of E2 on contralateral UBF, MAP, and heart rate in the absence and presence of continuous intra-arterial infusions of TEA at varying doses

Effects of L-NAME plus TEA on E₂ responses. After constructing the dose-inhibition curve for local intra-arterial infusions of TEA, it was determined that the arterial concentration of TEA that inhibited E₂-induced uterine vascular responses ~40% was 0.3 mM. Therefore, eight studies were performed in five nonpregnant ewes to examine the effects of simultaneous submaximal inhibition of NOS and BK_Ca. L-NAME was infused via one uterine artery catheter for 10 min at a rate to achieve a local arterial concentration of 5 mg/ml and 40% inhibition (34). Twenty minutes later, TEA was infused locally in the same uterine horn to achieve and maintain a concentration of 0.3 mM for 120 min. Systemic E₂ (1 µg/kg iv) was infused 30 min after starting the TEA infusion. A representative experiment is shown in Fig. 7, demonstrating the unilateral inhibitory effects of the two antagonists on E₂-induced increases in left UBF. On the day before the study of the two antagonists, E₂ alone increased UBF more than sevenfold by 90 min in both uterine horns and increased heart rate 21%; MAP was unaffected (Table 2). Although L-NAME and L-NAME plus TEA did not alter basal ipsilateral or contralateral UBF (Fig. 8), the rise in ipsilateral UBF 90 min after systemic E₂ (1 µg/kg) was completely inhibited by the combination of antagonists (Fig. 8A). However, UBF in the treated horn gradually rose 3.9-fold (P = 0.001) 90 min after stopping the TEA infusion (Figs. 7 and 8A), a value that was 48% of that seen in the presence of E₂ alone. In contrast, UBF in the untreated uterine horn increased 4.4-fold 90 min after E₂ (P < 0.0001, Fig. 8B) and an additional 38% 90 min after stopping TEA. Compared with E₂-induced increases in UBF on the previous day (Table 2), this response was also decreased ~50%, demonstrating cross circulation between the two uterine horns. Local L-NAME infusion alone increased MAP and decreased heart rate. These values, however, were unaffected by TEA, and, although E₂ increased heart rate 15% 90 min after E₂ in the presence of L-NAME plus TEA, MAP was unaltered (Table 2).

View larger version (14K): [in this window] [in a new window]   Fig. 7.   Representative tracing demonstrating the experimental protocol and the effect of a left intra-arterial infusion of nitro-L-arginine methyl ester (L- NAME; 5 mg/ml) for 10 min followed by TEA (0.3 mM) for 120 min on E2-induced increases in UBF. The contralateral uterine horn (right) was not infused and served as an internal control.

View this table: [in this window] [in a new window]   Table 2.   Effects of E2 on MAP and heart rate in the absence and presence of L-NAME and TEA and on UBF in the absence of these antagonists

View larger version (27K): [in this window] [in a new window]   Fig. 8.   Effects of local intra-arterial infusions of L-NAME (5 mg/ml) for 10 min and TEA (0.3 mM) for 120 min on E2-induced increases in UBF in the treated (A) and control (B) uterine horns of nonpregnant ewes (n = 8). Different letters designate significant differences between time periods at P  0.001 when analyzed by repeated-measures ANOVA.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The vasodilating properties of estrogen are believed to be beneficial in preventing cardiovascular disease in women (22) and to be responsible, in part, for the systemic vasodilation characteristic of normal pregnancy (33). However, the mechanisms mediating this vasodilation are not fully understood. Existing evidence suggests that both endothelium-dependent and -independent events may be involved (1, 11, 16, 22, 23, 25, 43, 44). In the present study, we have demonstrated for the first time that uterine artery myocytes express BK_Ca, which are rapidly activated in the presence of low concentrations of E₂, at least in part through a cGMP-dependent mechanism. Furthermore, we have shown for the first time in intact, unanesthetized animals that TEA, a selective BK_Ca antagonist (27), dose dependently attenuates E₂-mediated vasodilation but does not completely inhibit this response. In contrast, local infusions of submaximal inhibitory doses of L-NAME, a NOS antagonist, and TEA result in complete inhibition of uterine vascular responses to E₂. Therefore, these studies demonstrate that both NO and BK_Ca contribute to E₂-mediated vasodilation in intact unstressed animals, that an interaction exists between E₂-mediated activation of NOS and BK_Ca through activation of guanylyl cyclase, and that both endothelium-dependent and -independent mechanisms are probably involved.

Although the mechanisms responsible for estrogen-induced vasodilation have eluded investigators for 60 years, recent studies point to an important role for activation of NOS in the arterial wall. For example, eNOS is expressed in uterine artery endothelium (36, 38), and E₂ upregulates eNOS expression and activity in endothelium from several vascular beds, including the uterine circulation (8-10, 17, 36, 38, 41). However, NOS inhibition in both the uterine and coronary vascular beds with L-NAME only partially attenuates E₂-mediated vasodilation (28, 34, 40), suggesting involvement of additional mechanisms. In addition, endothelium-independent mechanisms are known to contribute to the relaxant effects of E₂ in the coronary circulation (11, 25, 43). Because the uterine vascular bed is known to be responsive to infused E₂ (6, 13, 20, 32, 33, 34) and can be isolated in unanesthetized ewes (20, 34), we used it to further investigate the mechanisms responsible for E₂-mediated vasodilation. Employing in vitro patch-clamp techniques, we identified the presence of BK_Ca in uterine artery myocytes and observed a 70-fold increase in channel activity within 20-40 min after exposure to nanomolar concentrations of E₂. Furthermore, when we added a guanylyl cyclase inhibitor, channel activity decreased ~75%, and this was restored by adding back 8-bromo-cGMP. This is strikingly similar to that observed in coronary myocytes (4, 43). Therefore, E₂ appears to mediate its vasodilatory effects in both vascular beds through endothelium-dependent and -independent mechanisms. Moreover, these data suggest in the latter case that this is mediated by rapid activation of smooth muscle guanylyl cyclase, suggesting the expression of a NOS isoform within the media (36).

To establish the contribution of BK_Ca to E₂-mediated uterine vasodilation in intact ewes, we determined the inhibitory effects of locally infused TEA using the paradigm previously employed with NOS inhibition by L-NAME (34). TEA was used rather than charybdotoxin or iberiotoxin, also specific antagonists of the BK_Ca (27), since it would have been impossible to construct a dose-response curve with these agents due to their cost. At concentrations <1 mM, TEA is a selective BK_Ca antagonist (27). Thus we used estimated arterial TEA concentrations ranging from 0.05 to 0.6 mM, well below that which inhibits other potassium channels (27), and we established, for the first time in vivo, a dose-inhibition curve for E₂-mediated vasodilation in the absence of systemic effects. From the dose-inhibition curve, 50% inhibition would have occurred at an arterial concentration of 0.3-0.4 mM, which is probably even less at the level of the vascular myocyte, and is consistent with the 0.2-0.3 mM observed in vitro (27). Of note, local TEA infusions did not alter basal UBF, suggesting minimal channel activity in uterine arteries in the absence of E₂, which is consistent with our patch-clamp studies. Furthermore, local TEA did not modify E₂ responses in either the contralateral uterine horn or systemic vasculature, e.g., heart rate rose ~20%, demonstrating minimal systemic spillover or that BK_Ca are not involved in systemic responses to E₂. This inhibitory effect on the uterine vasculature was absent 24 h later; thus channel blockade was reversible, and TEA did not manifest any toxic effects at the doses studied. However, as with L-NAME (34), maximum inhibition of uterine vascular responses to E₂ averaged 60-65%. Therefore, BK_Ca contribute to E₂-mediated vasodilation but are not solely responsible.

Because local inhibition of either NOS or BK_Ca alone does not entirely inhibit acute E₂-mediated vasodilation, we determined if the vascular responses to E₂ reflect an interaction between the two pathways. Evidence for this is obtained not only from the present study but also from earlier findings that the E₂-mediated rise in UBF is associated with NOS activation and parallel increases in uterine cGMP production (34) and more recent reports that acute E₂ increases NOS activity in uterine artery endothelium (36, 38). Furthermore, in porcine coronary myocytes, L-NAME blocks BK_Ca activation by E₂, which appears to be mediated through a cGMP-dependent mechanism involving smooth muscle iNOS (4). Submaximal inhibitory doses of L-NAME plus TEA completely inhibited E₂-induced vasodilation. Therefore, both pathways are involved, and they may be interactive. Node et al. (28) reported similar observations in studies of the canine coronary circulation. They observed that, although L-NAME alone prevented increases in coronary vascular cGMP synthesis after acute E₂ exposure, the rise in coronary blood flow was inhibited only 50%. In contrast, local L-NAME plus BK_Ca blockade with iberiotoxin completely inhibited acute E₂-mediated coronary vasodilation. It is unclear, however, if the doses were submaximal, since dose-inhibition curves were not generated. Wellman et al. (42) also observed reversal of E₂-mediated vasorelaxation in intact rat coronary arteries by NOS inhibition and iberiotoxin. They suggested, in contrast to White et al. (4, 43) and the present study, that endothelium-derived NO is essential for BK_Ca activation through a cGMP-dependent mechanism. Darkow et al. (4) suggested that smooth muscle iNOS was involved in porcine coronary arteries, whereas Salhab et al. (36) found neuronal NOS (nNOS) expression in uterine artery myocytes. Although differences may exist in the cellular pathways of different species, available evidence now supports the thesis that E₂ mediates its vasodilatory effect in both the uterine and coronary vasculature by activating both NOS and BK_Ca. The present data, however, demonstrate that no more than 65% of the E₂-mediated vasodilation is due to either NOS or BK_Ca activation alone but do not rule out involvement of other mechanisms.

After the TEA infusion was stopped, UBF rose fourfold, and the response resembled that after acute E₂ exposure (13, 33), i.e., a 30-min delay followed by a rise in UBF that was maximum by 90 min. This was not observed in the coronary circulation, reflecting the acute nature of these experiments (28). However, a similar rise in UBF follows inhibition of E₂-mediated uterine vasodilation with cycloheximide (13). These authors concluded that this was evidence for transient inhibition of new protein synthesis essential to intracellular signal transduction after binding of E₂ to its receptor. Thus, when cycloheximide was removed, protein synthesis was initiated, and the intracellular responses were completed. Local L-NAME infusions result in prolonged inhibition of vascular NOS, as evidenced by the absence of increases in local cGMP synthesis (28, 34). Thus E₂ may activate BK_Ca after receptor binding via a series of enzymatic steps independent of NOS activation or, as recently reported, by directly binding to the -subunit of the BK_Ca (39). Either would explain the rise in UBF after TEA removal and reversal of BK_Ca blockade. This further supports the thesis that the two pathways are interactive but also may function independently of each other. Alternatively, another NO-independent mechanism may be involved in E₂-mediated vasodilation. Inhibition of calcium influx and alterations in intracellular calcium are associated with endothelium-independent relaxation after E₂ exposure (7, 11, 12, 14), and calcium channels have been implicated in uterine vasodilation during the porcine estrous cycle and in early porcine pregnancy (37). Because NOS plus BK_Ca inhibition results in complete inhibition of E₂-mediated vasodilation, the alternative pathway(s) may play a minor role compared with activation of NOS and BK_Ca. This will have to be examined in future studies. It is unlikely, however, that prostaglandins are involved, since indomethacin has no apparent effect on either the uterine or coronary vascular responses to acute E₂ exposure (28, 34).

Vascular responses to estrogen depend on the method and duration of exposure. In nonpregnant ewes continuously infused with E₂, rapid responses begin within 30 min, whereas late effects are evident at 5-7 days, all of which have been characterized by monitoring systemic and uterine hemodynamic responses (18). The late hemodynamic effects likely reflect genomic mechanisms, since they are associated with increases in eNOS protein in uterine artery endothelium, total NOS abundance in whole uterine arteries, and basal UBF and arterial contents of cGMP (36, 38, 40). The rapidity of the immediate vascular responses to acute E₂, however, suggests involvement of nongenomic mechanisms. This is supported by the lack of an effect of actinomycin D on acute E₂-induced increases in UBF (29, 31) and in eNOS activity by cultured endothelium (3). Furthermore, it is now apparent that E₂ can increase eNOS activity within minutes through a receptor-mediated event (2, 3, 15). We clearly demonstrate that uterine vascular responses to bolus doses of E₂ are not evident for 30 min but progress quickly thereafter. This delay in intact sheep may reflect the time needed to maximize newly synthesized NO (36), to increase smooth muscle cGMP to an effective threshold level (34), or to recruit and activate BK_Ca throughout the uterine vascular bed. It is notable that, in the patch-clamp studies, BK_Ca activation may not be maximum until 20-40 min. Once the system is activated, however, vasodilation progresses for 60 min, is stable for 1-2 h, and gradually returns to baseline conditions over 6-12 h (13, 33). The mechanisms responsible for this sequence are unknown.

In the present study, we provide further evidence that the uterine vascular bed in oophorectomized nonpregnant ewes is an excellent model in which to study both the in vivo (34) and in vitro (36) mechanisms responsible for estrogen-mediated vasodilation and that these mechanisms resemble those observed in the coronary vascular bed under less optimal circumstances (28). Our data also provide strong in vivo and in vitro evidence that both NOS and BK_Ca are involved in the vasodilatory responses after acute E₂ exposure, that these pathways are likely initiated by nongenomic mechanisms, and that another mechanism may be involved. From these results, we suggest that acute E₂ exposure initiates a receptor-mediated event that activates eNOS and probably smooth muscle nNOS (36) to produce NO. This NO increases adjacent smooth muscle cGMP, which activates a cGMP-dependent kinase that enhances BK_Ca activity and decreases calcium inflow via voltage-gated calcium channels, resulting in vasodilation. It is unclear, however, if E₂ directly enhances BK_Ca activity and inhibits calcium influx or intracellular release.