We studied the importance of endothelium-derived hyperpolarizing factor (EDHF) vs. nitric oxide (NO) and prostacyclin (PGI2) in bradykinin (BK)-induced relaxation in isolated small subcutaneous arteries from normal pregnant women. We also explored the contribution of cytochrome P-450 (CYP450) product of arachidonic acid (AA) metabolism, hydrogen peroxide (H2O2), and gap junctions that have been suggested to be involved in EDHF-mediated responses. Isolated arteries obtained from subcutaneous fat biopsies of normal pregnant women (n = 30) undergoing planned cesarean section were mounted in a wire-myography system. In norepinephrine-constricted vessels, incubation with NG-nitro-l-arginine methyl ester (l-NAME) resulted in a significant reduction in relaxation to BK. Simultaneous incubation with l-NAME and indomethacin failed to modify this response further. BK-mediated dilatation in the presence of K+-modified solution was decreased to similar level as obtained after incubation with l-NAME. Incubation with l-NAME abolished BK-induced responses in K+-modified solution. Sulfaphenazole, a specific inhibitor of CYP450 epoxygenase, and catalase (an enzyme that decomposes H2O2) did not affect the EDHF-mediated relaxation because concentration-response curves to BK were similar in arteries after incubation with l-NAME vs. l-NAME + sulfaphenazole and l-NAME + catalase. The inhibitor of gap junctions, 18α-glycyrrhetinic acid, significantly reduced BK-mediated relaxation both without and with incubation with l-NAME. We found that both NO and EDHF, but not PGI2, are involved in the endothelium-dependent dilatation to BK. BK-induced relaxation is almost equally mediated by NO and EDHF. CYP450 epoxygenase metabolites of AA or H2O2 do not account for EDHF-mediated response; however, gap junctions are involved in the EDHF-mediated responses to BK in subcutaneous small arteries in normal pregnancy.
- endothelium-dependent relaxation
- gap junctions
- endothelium-derived hyperpolarizing factor
- hydrogen peroxide
maternal cardiovascular adaptation to normal pregnancy is associated with decreased peripheral vascular resistance that plays an important role for blood pressure reduction despite an increase in plasma volume and cardiac output. The vascular adaptation to normal pregnancy is believed to be dependent on an enhanced endothelium-dependent dilatation (3, 18, 25). However, the pathways underlying these changes are not fully understood and need further investigation because an impaired cardiovascular adaptation is associated with the development of preeclampsia (18, 25, 42). It is generally accepted that preeclampsia is an endothelial cell disorder, and the increased vascular resistance and blood pressure during this pregnancy-related disease are due to endothelial dysfunction and alterations in endothelium-dependent relaxation (18). The mechanisms underlying endothelium-dependent relaxation in normal human pregnancy have attracted considerable interest due to the rationale to design novel therapeutic models to prevent endothelial dysfunction in preeclampsia.
It has been suggested that the vasodilatory capacity of endothelium is achieved by combined effects of nitric oxide (NO), prostacyclin (PGI2), and endothelium-derived hyperpolarizing factor (EDHF). A contribution of these endothelium-derived substances varies among vessels of different size, vascular bed studied, or agonist used. Thus in vitro studies have shown that EDHF-mediated dilatation tends to be more important in small arteries than in large vessels, where it is believed that NO-mediated dilatation is dominant (32).
The chemical nature of EDHF and the mechanism of EDHF-mediated responses at the level of smooth muscle cells are still not fully understood. It is generally agreed that EDHF activates K+ channels and induces hyperpolarization of smooth muscle cells. This results in a closure of voltage-operated Ca2+ channels and reduction of cytosolic Ca2+ followed by subsequent relaxation (4). The identity of EDHF, however, remains controversial, and such major potential candidates as cytochrome P-450 (CYP450) products of arachidonic acid (AA) (38), potassium ions (13), and hydrogen peroxide (H2O2) (31) have been suggested (Fig. 1). Endothelial cell hyperpolarization may also be transmitted to smooth muscle cells through myoendothelial gap junctions (MEGJ) (6). Busse et al. (4) suggested that endothelium-dependent hyperpolarization may be mediated through a combination of both chemical and electric transmissions; however, the contribution of them varies among species and different vascular beds.
The present study was undertaken to clarify the mechanism involved in the EDHF-mediated vasodilator responses in isolated small arteries from subcutaneous circulation in normal pregnant women. In the present study, we have defined EDHF as the l-NAME- and indomethacin-insensitive component of the dilatation to BK.
MATERIALS AND METHODS
Biopsy Collection and Experimental Procedures
The study was approved by the Ethical Committee of Huddinge University Hospital, and the subcutaneous fat biopsy specimens were obtained from 30 pregnant women (15 nulliparous) with a median age of 27 yr (range 18–42) and a median gestational age of 38 wk (37–40) undergoing planned cesarean section due to breach presentation (n = 14), previous cesarean section (n = 8), and psychological reasons (n = 8).
At cesarean delivery subcutaneous fat biopsies were taken from the incision edge and immediately placed in iced physiological salt solution (PSS) of the following composition (mmol/l): 119 NaCl, 4.7 KCl, 2.5 CaCl2, 1.17 MgSO4, 25 NaHCO3, 1.18 KH2PO4, 0.026 EDTA, and 5.5 glucose. In total, 99 resistance arteries with approximate internal diameter 240 μm were dissected from 30 biopsies and cleaned from surrounding tissue under a dissection microscope. Arteries were stored at physiological saline at 4°C until they were mounted in a wire myography system (multimyograph, model 610, Danish Myo Technology; Aarhus, Denmark).
Depending on the biopsy, two or more segments of subcutaneous arteries were taken and mounted in the organ baths of a four-channel multimyograph. Artery segment was mounted on two stainless wires (40 μm in diameter). One wire was attached to a force transducer and the other to a micrometer to measure the vessel tension.
After all arteries were mounted, they were allowed to equilibrate for 30 min at 37°C, while continuously being oxygenated with 5% carbon dioxide in oxygen. All solutions, including the incubation solutions, were refreshed every 30 min. A standardized normalization procedure was then performed to allow calculation of the artery diameter at which the in vivo transmural pressure of the relaxed artery would have been 100 mmHg. Arteries were then set at 0.9 times this diameter, since it is generally accepted that this diameter enables optimal contractile ability for the arteries with a low resting tension. Myodac Software was used for these calibrations and for data registration (version 2.1, Danish Myo Technology).
After the normalization procedure the arteries were left to equilibrate for 30 min, and five reference constrictions were elicited to elucidate if the arteries are suitable for experiments. The first, second, and fifth contractions were produced using a high (124 mmol/l)-potassium solution (KPSS, made by equimolar substitution of KCl for NaCl in PSS) containing 1 μmol/l norepinephrine (NE). The third and fourth were obtained with NE (1 μmol/l) or KPSS alone. Arteries that did not achieve at least 60% relaxation to 3 μmol/l of bradykinin (BK) were excluded from the study.
All arteries were preconstricted with NE (3 μmol/l), resulting in a contraction level of ≅90% of the initial response to KPSS. In experiments in which K+-modified PSS (35 mmol/l equimolar exchange of NaCl with KCl) was used, the concentration of NE was reduced to 1 μmol/l to achieve a preconstriction level similar to that used in previous experiments. A concentration-response curve to BK (1 nmol/l to 3 μmol/l) was then performed.
Contribution of NO, PGI2, and EDHF to BK-mediated relaxation.
The concentration-response curves to BK were obtained before and after incubation with l-NAME (100 μmol/l, 30 min) alone or in combination with indomethacin (Indo, 10 μmol/l, 30 min). The concentration of l-NAME was sufficient to inhibit the production of NO, since additional application of Nω-nitro-l-arginine (300 μmol/l, 30 min), another inhibitor of NO synthase, had no further inhibitory effect on the BK-induced relaxation (data not shown).
In a separate series of experiments, the concentration-response curves to BK were performed in the K+-modified PSS (35 mmol/l KCl) before and after incubation with l-NAME.
Contribution of AA metabolites, H2O2, and gap junctions to BK-mediated relaxation.
EDHF-mediated contribution to BK-mediated relaxation was investigated by evaluating the BK-mediated relaxation in the presence of the specific inhibitor of CYP450 epoxygenase, sulfaphenazole (10 μmol/l, 30 min), alone or in combination with l-NAME.
To evaluate if H2O2 is mediating the EDHF-typed responses to BK, an initial concentration-response curve to BK was constructed followed by a second concentration-response curve but in the presence of catalase (1,250–6,250 U/ml, an enzyme that dismutates H2O2 to form water and oxygen) alone or in combination with l-NAME.
To elucidate the contribution of gap junctions to the EDHF-typed responses in BK-mediated relaxation, concentration-response curves were obtained after incubation with a reversible inhibitor of gap junctions, 18α-glycyrrhetinic acid (18α-GA, 100 μmol/l, 15 min), alone or in combination with l-NAME.
Chemicals were obtained from Sigma (St. Louis, MO). To prepare stock solution, the substances were dissolved in the corresponding solvent and further diluted in PSS as appropriate. NE and catalase were dissolved in PSS directly before every experiment. Indo was dissolved in pure ethanol, and 18α-GA and sulfaphenazole were dissolved in DMSO at a concentration of 10−2 M. The highest concentration of ethanol and DMSO in the chamber was 1% (vol/vol), and the final bath concentration of combined solvents used simultaneously was not higher than 3% (vol/vol). All concentrations represent the final steady-state concentrations in the chamber.
The force developed by the artery per millimeter of artery segment during application of a certain concentration of a vasoactive substance was calculated using Myodata (Danish Myo Technology). Data were then transferred to STATISTICA (version 5.5, StatSoft, Uppsala, Sweden), in which all statistical analyses were performed. All absolute measurements were corrected for the baseline force developed by the arteries. The relaxation to BK was calculated as a percentage of the contraction induced by NE. Negative log concentration (in mol/l) required to achieve 50% of the maximum response (pEC50) was calculated by nonlinear regression analysis (BioDataFit 1.02).
Multivariate ANOVA for repeated measures was used to compare BK concentration-response curves before and after incubation with particular inhibitors and for differences in BK-mediated relaxation between experimental groups. ANOVA was used to test the differences in diameter, reference, and NE preconstriction in arteries used for different experimental protocols. Paired and unpaired Student's t-test as appropriate was used to compare pEC50 value and NE preconstriction before and after incubation with different substances in arteries used for different experimental protocols. All data are presented as means ± SE; P < 0.05 was considered to be statistically significant.
Mean normalized internal diameter of the arteries was 241 ± 10 μm (range 127–498 μm). There were no differences in luminal diameter, the magnitude of contraction to KPSS, and preconstriction level among the arteries used for different experimental protocols (data not shown).
Contribution of NO, PGI2, and EDHF to BK-Mediated Relaxation
In NE (3 μmol/l)-preconstricted arteries, increased concentrations of BK caused approximately 85–95% relaxation in isolated subcutaneous arteries from healthy pregnant women (Figs. 2–6). The concentration-response curves to BK were similar before and after 30-min incubation with PSS alone, indicating that BK-mediated dilatation is reproducible [percent relaxation to BK at 3 μmol/l (here and in the following text): 89 ± 7% vs. 91 ± 9%; pEC50: 7.4 ± 0.17 vs. 7.7 ± 0.15 (n = 5) in a 1st and a 2nd concentration-response curve, respectively].
Incubation with l-NAME resulted in a significant reduction in relaxation to BK compared with that obtained in PSS (67 ± 8% in l-NAME vs. 85 ± 6% in PSS; pEC50: 6.68 ± 0.18 vs. 7.73 ± 0.13 in the same arteries, n = 8, respectively, P < 0.01, Fig. 2). Simultaneous incubation with l-NAME and Indo failed to modify this response further [56 ± 5% after l-NAME + Indo (n = 16) vs. 67 ± 8% after l-NAME alone (n = 8); pEC50: 6.76 ± 0.16 vs. 6.68 ± 0.18, P > 0.5, Fig. 2].
BK-mediated dilatation in the presence of K+-modified PSS was decreased compared with that obtained in normal PSS (54 ± 4% vs. 88 ± 7% in PSS; pEC50: 6.26 ± 0.11 vs. 7.45 ± 0.11 in the same arteries, n = 7, P < 0.001, Fig. 3). This reduction was similar to that obtained after incubation with l-NAME in PSS (57 ± 8%; pEC50: 6.35 ± 0.17, n = 8). Inhibition of NOS with l-NAME abolished BK-induced responses in K+-modified solution (4 ± 1% in KCl + l-NAME, n = 7, vs. 86 ± 5% in PSS, n = 22, P < 0.001, and 57 ± 8% with l-NAME in PSS, n = 8, P < 0.001, Fig. 3).
Contribution of AA Metabolites, H2O2, and Gap Junctions to BK-Mediated Relaxation
Sulfaphenazole, a specific inhibitor of CYP450 epoxygenase, did not affect the EDHF-mediated relaxation, since concentration-response curves to BK were similar in arteries after incubation with l-NAME vs. l-NAME + sulfaphenazole (64 ± 9%, n = 8 and 59 ± 8%, n = 6, P = 0.7, pEC50: 6.78 ± 0.19 vs. 6.65 ± 0.17, P = 0.7, Fig. 4). Incubation with sulfaphenazole alone did not effect BK-mediated relaxation (85 ± 10%, n = 4, compared with PSS in the same arteries 89 ± 11%, pEC50: 7.38 ± 0.2 vs. 7.55 ± 0.18, Fig. 4).
The contribution of H2O2 to the EDHF-typed responses was explored by using catalase that dismutates H2O2 to form water and oxygen. Incubation with catalase (1,250 U/ml) alone did not alter concentration-response curves to BK (79 ± 5%, n = 6 vs. 87 ± 6 in PSS; pEC50: respectively, 7.1 ± 0.24 vs. 7.56 ± 0.3 for the same arteries, P = 0.4, Fig. 5). Similarly, the inhibitory effect of catalase to EDHF-mediated responses in the presence of l-NAME was negligible [47 ± 6% (n = 17) compared with l-NAME alone 54 ± 5 (n = 12), P = 0.35, pEC50: 6.71 ± 0.18 vs. 6.49 ± 0.13, P = 0.5, Fig. 5]. Thus catalase had no significant effect on vasorelaxation in the absence or presence of l-NAME. In addition, increased concentration of catalase (5-fold, 6,250 U/ml) had no effect on relaxation to BK independently of the presence of l-NAME (51 ± 10% and 81 ± 8%, n = 4, P > 0.1 after incubation with and without l-NAME, respectively).
The inhibitor of gap junctions 18α-GA significantly reduced BK-mediated relaxation both without and with incubation with l-NAME (Fig. 6). Thus incubation with 18α-GA alone resulted in a significant reduction in relaxation to BK compared with that obtained in PSS (59 ± 7% n = 7 compared with PSS in the same arteries, 83 ± 6%, P < 0.01, pEC50: 6.75 ± 0.17 vs. 7.52 ± 0.14, P < 0.005, Fig. 6). Simultaneous incubation with l-NAME and 18α-GA modified this response further [20 ± 3% l-NAME + 18α-GA (n = 12) vs. 65 ± 8% l-NAME (n = 9), P < 0.001, Fig. 5].
Effect of Different Inhibitors of Endothelial Function on the Degree of Preconstriction
NO and PGI2 inhibitors did not change the degree of NE-induced preconstriction of the arteries. Indeed, the magnitude of 3 μmol/l NE contractions was 3.1 ± 0.2 mN/mm2 (n = 30) before vs. 3.3 ± 0.3 mN/mm2 (n = 30, P > 0.1) after incubation with l-NAME.
Sulfaphenazole also did not change the response of arteries to 3 μmol/l NE [3.0 ± 0.6 mN/mm2 (n = 4) alone and 3.1 ± 0.8 mN/mm2 (n = 6) in combination with l-NAME vs. 3.0 ± 0.5 mN/mm2 (n = 4, P > 0.1) in the same arteries in PSS and 3.2 ± 0.7 mN/mm2 (n = 8, P > 0.1) after incubation with l-NAME alone].
Catalase (1,250–6,250 U/ml) did not affect the response of arteries to 3 μmol/l NE [3.5 ± 0.7 mN/mm2 (n = 6) alone and 3.4 ± 0.3 mN/mm2 (n = 17) in combination with l-NAME vs. 3.2 ± 0.6 mN/mm2 (n = 6, P > 0.1) in PSS in the same arteries and 3,6 ± 0,6 mN/mm2 (n = 12, P > 0.1) after incubation with l-NAME alone].
In contrast, incubation with 18α-GA alone significantly reduced NE-induced tone (3.0 ± 0.5 mN/mm2 PSS vs. 1.6 ± 0.4 mN/mm2 18α-GA, n = 7, P < 0.05, Fig. 7). However, NE-induced contraction was similar after incubation of arteries with l-NAME alone (2.9 ± 0.5 mN/mm2 PSS vs. 3.1 ± 0.7 mN/mm2 l-NAME, n = 10, P = 0.8) or in combination with 18α-GA (2.9 ± 0.5 mN/mm2 PSS vs. 2.6 ± 0.5 mN/mm2 l-NAME + 18α-GA, n = 12, P = 0.6, Fig. 7). However, 18α-GA alone did not change KPSS-induced contraction (before incubation 3.8 ± 0.7 mN/mm2 and after 20 min preincubation with 18α-GA 3.0 ± 0.5 mN/mm2, n = 3, P = 0.7).
The solvents used did not affect the vasodilatory responses in the isolated subcutaneous arteries. Ethanol that was used for dilution of Indo did not influence the BK-induced dilatation (e.g., 90 ± 7% and pEC50: 7.45 ± 0.22, n = 4) compared with BK-mediated dilatation in PSS alone (89 ± 6%; pEC50: 7.42 ± 0.19, n = 4). Similarly, DMSO that was used for dilution of 18-αGA and sulfaphenazole had no effect on BK-mediated dilatation (88 ± 8% and pEC50: 7.5 ± 0.26, n = 5). The final bath concentration of combined solvents used in our study (i.e., DMSO + ethanol + distilled water), which was not higher than 3% (vol/vol), did not affect the vasodilatory response to BK (89 ± 9% and pEC50: 7.4 ± 0.25, n = 4) compared with concentration-response curve to BK in PSS alone.
Here we report that endothelium-dependent relaxation is almost equally dependent on NO and EDHF release, whereas PGI2 appears to play no role in BK-induced responses in isolated subcutaneous resistance arteries from normal pregnant women. Neither H2O2 nor the product of CYP450 epoxygenase is the primary mediator of NO- and prostanoid-independent relaxation in these arteries. Gap junctions seem to be responsible for BK-induced EDHF-mediated relaxation in isolated subcutaneous arteries from normal pregnant women.
Relative Contribution of NO, PGI2, and EDHF to BK-Induced Relaxation
It is now generally accepted that NO plays an important role as a systemic vasodilator in pregnancy (39). More recently, a role for EDHF in the increased endothelium-dependent relaxation during pregnancy has been gradually recognized. Thus several studies demonstrated that EDHF-type responses to several endothelium-dependent agonists might be increased in rat (15) or human resistance arteries in normal pregnancy (26, 37). Our results indeed further strengthen the importance of EDHF to endothelium-dependent dilatation. In the present study, we have demonstrated that the contribution of NO and EDHF in BK-mediated dilatation seems to be almost equal, since l-NAME and KCl reduced the dilatation to a similar level, whereas combination of both abolished the response.
In contrast to NO and EDHF, PGI2 is more likely to play a role as a local autacoid, a conclusion upheld by data from pregnant animals (21) and women (40), which have shown that infusion of Indo did not affect blood pressure or peripheral resistance. A minor or no role of endothelium-derived PGI2 in vascular reactivity of peripheral circulation in human pregnancy is supported by our study demonstrating a similar relaxation to BK after incubation with l-NAME + Indo and l-NAME alone. The minor if any influence of PGI2 on BK-mediated dilatation in subcutaneous arteries is consistent with previous studies on vessels from pregnant (26) and nonpregnant (5, 9, 33) humans.
Thus our study indicates that NO and EDHF are the main endothelium-derived vasodilators involved in endothelium-dependent dilatation to BK in human subcutaneous resistance arteries in normal human pregnancy. The experimental approach was based on the evidence that endothelium-dependent hyperpolarization of smooth muscle cells, as a mechanism of EDHF-mediated vasodilator response, depends on the opening of K+ channels in the smooth muscle plasmalemma. An increase in extracellular K+ should then reduce the electrochemical gradient for K+ and inhibit the EDHF responses (32). Therefore, evaluation of BK-induced relaxation in the presence of a raised extracellular K+ solution (>25 mmol/l) should represent the NO-dependent contribution to BK-mediated relaxation. Such a pharmacological approach of raising extracellular K+ or inhibition of NO production using NO synthase inhibitors to account for EDHF-mediated responses is currently used in several laboratories instead of the electrophysiological measurements of alterations in membrane potential, which are associated with meticulous technical difficulties when applied to small arteries (1, 9, 24, 37).
The NO and EDHF component in BK-induced dilatation can be determined as the l-NAME-sensitive or KCl-insensitive value and l-NAME-insensitive or KCl-sensitive value of relaxation, respectively. However, in our study there was a marked difference between these values. For example, the NO component calculated as l-NAME-sensitive value was ≈35% of the response to 3 μmol/l BK while the NO component considered as KCl-insensitive value was ≈65%. On the other hand, the KCl-sensitive and l-NAME-insensitive parts of endothelium-dependent relaxation, which represent the contribution of EDHF, accounted for ≈35% and ≈65%, respectively, whereas the combination of l-NAME and partial depolarization induced by 35 mmol/l K+ in PSS abolished the vasodilator responses to BK. It might be suggested, therefore, that NO and EDHF play almost an equal role in mediating endothelium-dependent relaxation to BK in resistance subcutaneous arteries from normal pregnant women. Moreover, our results indicate that in the absence of NO, an EDHF mediates vasodilatation of resistance arteries to BK or vice versa, thus providing a factor of safety in endothelium-dependent mechanisms in normal pregnancy (24).
EDHF: Identification and Mechanisms of Action
To our knowledge, our study is the first to explore underlying mechanisms of EDHF-mediated responses in subcutaneous resistance arteries from normal pregnant women. We aimed to identify whether epoxygenase products of AA, endogenous H2O2, or gap junctional communications account for EDHF activity in human small subcutaneous arteries isolated from pregnant women. We excluded that K+ per se is responsible for EDHF activity in human subcutaneous arteries, since three recent independent studies failed to prove it (5, 9, 10).
Role for AA Metabolites in EDHF-Mediated Responses
In the present study, we explored the role of CYP450 products in the EDHF-mediated responses in isolated arteries from normal pregnant women. Human endothelial cells contain CYP450 epoxygenase (36), and in our study we used sulfaphenazole, a specific inhibitor of it. This inhibitor had no detectable effect on the BK-induced relaxation in the presence of l-NAME, suggesting that CYP450 metabolites of AA could not account for the EDHF-mediated responses in human subcutaneous arteries from normal pregnant women. Our results are in line with those in human subcutaneous arteries from nonpregnant subjects (5). However, they are in contrast with observations in human isolated coronary (35) and internal mammary arteries (2), forearm microcirculation (19), and even in the gluteal subcutaneous arteries from nonpregnant subjects (9), where derivatives of AA have been suggested to be involved in EDHF-mediated responses. The differences among the studies in human subcutaneous arteries may result from different CYP450 inhibitors used, which have their own limitations. For example, it has been shown that nonspecific inhibitors of CYP450 such as micanazole and other imidazoles may also block K+ channels (14) or impair the agonist-induced increase in intracellular Ca2+ in endothelial cells (20). This could prevent EDHF-mediated effects on smooth muscle rather than affecting the synthesis/release of EDHF itself. An alternative explanation is that the mechanisms of EDHF-mediated response are dependent on the heterogeneity of basic clinical details of the volunteers recruited for studies, since in the above-mentioned studies subcutaneous fat biopsies were obtained from healthy female and male volunteers (9) or in patients with cancer or cardiovascular complications (5).
Role for H2O2 in EDHF-Mediated Responses
Pregnancy is accompanied by high metabolic demands and increased requirements for tissue oxygen, resulting in an increased production of reactive oxygen species (16). Women undergoing normal pregnancy experience increased oxidative stress and lipid peroxidation compared with nonpregnant subjects (29). Thus the level of endogenous production of H2O2 might therefore be increased in pregnancy, although the oxidative stress during normal pregnancy is counteracted by a parallel increase in antioxidant capacity (8). In our study the role of H2O2 in EDHF-mediated responses in small arteries from normal pregnant women was evaluated by using different concentrations of an enzyme catalase, known as a scavenger of extracellular H2O2. We found that EDHF-mediated responses to BK were not due to H2O2. In contrast, several other investigations in isolated arteries from nonpregnant humans (30, 34) and animals (27, 31) have demonstrated that H2O2 is a primary candidate for EDHF-type action. It has been shown that generation of endothelium-derived H2O2 contributes to BK-induced (30) and flow-induced (34) vasodilatation in isolated human arteries; however, these findings were evident only in the diseased condition associated with oxidative stress when generation of superoxide is enhanced. It is therefore possible that contribution of H2O2 as endothelium-derived vasodilatator could be increased during preeclampsia, since the evidence for oxidative stress in this disorder is undoubtedly recognized.
Role of Gap Junctions in EDHF-Mediated Responses
Finally, the assessment of gap junction contribution to EDHF-mediated effects in small arteries was performed, since these vessels have the highest density of MEGJ. In our study we used a derivative of glycyrrhetinic acid (GA), 18α-GA, that disrupts gap junction structure and inhibits intercellular electrical communications. We found that 18α-GA significantly attenuated BK-induced vasodilatation irrespective of whether it was used alone or in combination with l-NAME. This indicates that gap junctions play a critical role in endothelium-dependent relaxation to BK in isolated subcutaneous resistance arteries from normal pregnant women. Our findings are in line with recent experimental evidence of gap junctions mediating EDHF-type responses in isolated arteries from both nonpregnant (6, 17) and pregnant animals (11). In support of our findings, Kenny et al. (23) demonstrated that several distinct agents that interfere with gap junctions inhibited non-NO, non-prostanoid-mediated responses to BK in isolated small myometrial arteries from normal pregnant women. This could suggest that gap junctions contribute to EDHF-type responses in several vascular beds in human pregnancy.
Our experimental data support the role of gap junctions in human subcutaneous circulation during pregnancy, and it is likely that either an electrical current or EDHF as the factor per se or even a combination of both might be transferred directly from endothelial to the adjacent vascular smooth muscle cells. Recently, it has been suggested that EDHF does not exist as a substance, and EDHF-related responses in human subcutaneous circulation are the consequence of the electrical current movement (10). Whether this finding is applicable to pregnancy remains unclear.
It is important to notice that 18α-GA affected the EDHF- more than NO-mediated component of the vasodilator response to BK (see Fig. 6), since the inhibitor effect of 18α-GA was ≈69% vs. ≈34% of maximal relaxation to BK with and without combined incubation with l-NAME, respectively. The involvement of gap junctions in the NO-mediated relaxation remains uncertain. The NO dependence on gap junctions has been demonstrated in isolated arteries from nonpregnant animals (22). In contrast, other study has shown that inhibitors of gap junctions do not affect responses to sodium nitroprusside, an exogenous source of NO (11). Moreover, it is difficult to perceive that gap junctions could significantly increase the movement of NO toward smooth muscle cells, since NO itself is a highly diffusible molecule through cell membrane. Thus the most relevant explanation for inhibitor effects of 18α-GA to BK-induced relaxation in the absence of l-NAME should be that ≈34% of maximal relaxation to BK in resistance subcutaneous arteries from normal pregnant women achieved in our study is accounted for EDHF in conditions of a preserved NO-mediated pathway. This is in accordance with the above-calculated contribution of EDHF-mediated responses, which was calculated as the KCl-sensitive part of endothelium-dependent relaxation and was found to account for ≈35% of maximal relaxation to BK.
It should be noticed that inhibitors of gap junctions might have additional pharmacological actions, including changes in ion channel properties, altered properties of second messengers, and changes in myofilament calcium sensitivity (28). This was the main reason for using the α-form of GA. This form is a more specific and less toxic inhibitor of gap junction than other GA compounds (12, 17). Moreover, 18α-GA at concentrations used in our study had no effect on levcromakalin-induced (K+ channel activator) and verapamil-induced relaxations (17) and did not impair the activation of endothelial adenylate cyclase (41), excluding the possibility of direct inhibition of K+ and Ca2+ channels in vascular smooth cells or cAMP homeostasis in endothelial cells.
In our study, 18α-GA altered contractile response to NE. Similar alterations to other contractile agonists such as vasopressin, endothelin, 5-HT, and PGF2α but not K+-modified PSS have been reported previously (7). It remains unclear why the contractility to NE was decreased significantly after incubation with 18α-GA alone but not in combination with l-NAME. However, it cannot be excluded that disruption of intercellular communication within the wall in some way induces activation of NO production. To our knowledge this observation of gap junctional/NO-dependent reduction in NE-induced contractility in arteries isolated from pregnant women has not yet been reported, and further investigations are warranted to clarify the mechanisms behind it.
The results of our study clearly demonstrate that gap junctions play a critical role for EDHF-mediated responses in small arteries from normal pregnant women; however, the component, either chemical or electrical, that is communicating between endothelial and underlying smooth muscle cells remains to be identified. Moreover, the importance of future clarification of EDHF's role and EDHF-type action in endothelium-dependent dilatation in preeclampsia becomes apparent because this disorder severely impairs the endothelium. It might be anticipated that mechanisms of EDHF-mediated effects could be altered as the specific situation requires to preserve the vasodilative capacity of the endothelium.
In conclusion, this study demonstrates that NO and EDHF almost equally account for endothelium-dependent relaxation to BK in subcutaneous arteries isolated from normal pregnant women. Our investigation also suggests a negligible role for AA metabolites and endogenous H2O2 as a candidate for EDHF action in endothelium-dependent dilatation. Gap junctions are responsible for EDHF-mediated responses to BK in small subcutaneous arteries from normal pregnant women.
This study was supported by grants from Karolinska Institutet, Harald Jeanssons, The Harald and Greta Jeanssons foundation, Swedish Research Council, Swedish Heart and Lung foundation, and Swedish Institute.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society