The modulatory role of androgens and progestins in the induction of vasorelaxation in human umbilical artery

Life Sciences 81 (2007) 993 – 1002 www.elsevier.com/locate/lifescie

The modulatory role of androgens and progestins in the induction of vasorelaxation in human umbilical artery Mercedes Perusquía a,⁎, Erika Navarrete a , Lorena González a , Carlos M. Villalón b a

Departamento de Biología Celular y Fisiología, Instituto de Investigaciones Biomédicas, Universidad Nacional Autónoma de México, Apartado postal 70228, 04510 México D.F., México b Farmacobiología, Cinvestav-Coapa, Tenorios 235, Col. Granjas-Coapa, C.P. 14330, México D.F., México Received 15 March 2007; accepted 24 July 2007

Abstract Sex steroids have been described as protectors of the cardiovascular system and one of their relevant actions is inhibition of vascular tone. However, this information has been derived from animal models. The aim of this study was to investigate the vasorelaxant properties of several progestins and androgens on the vascular tone of human umbilical artery (HUA) to elucidate their potential regulatory role on fetoplacental blood flow. HUA rings, obtained from umbilical cords at vaginal deliveries and cesarean section from term uncomplicated pregnancies, were isometrically recorded and precontracted with either KCl or serotonin. Subsequently, dehydroepiandrosterone, testosterone, progesterone and some of their 5-reduced metabolites were added at different noncumulative concentrations on KCl-induced precontraction. There were significant differences in the vasorelaxing responses to these steroids; excluding 5α-pregnandione, the remaining steroids induced concentration-dependent vasorelaxations. In general, androgens were more potent than progestins, with 5β-dihydrotestosterone being the most potent one. These vasorelaxations remained unaffected by inhibitors of transcription and translation, selective steroid receptor antagonists, a nitric oxide synthase inhibitor or specific blockers of K+ channels. Interestingly, the serotonin contraction was significantly less sensitive to steroid-induced vasorelaxation. Moreover, the contraction evoked by Ca2+ in depolarized tissues (by KCl–Ca2+ free solution) was prevented by steroids. These data, taken together, suggest that sex steroids (particularly androgens) induce an acute (nongenomically-mediated) vasorelaxing effect on the HUA which may be mediated by: (i) a nitric oxide-independent pathway; and/or (ii) a decrease in external Ca2+ influx by inactivating Ca2+ channels, but not by activating K+ channels. © 2007 Elsevier Inc. All rights reserved. Keywords: Androgens; Progestins; Vasorelaxation; Umbilical artery; Fetoplacental circulation; Nongenomic steroid effects

Introduction The cardiovascular system is an important target for female steroid hormones in which estrogens (Mendelsohn and Karas, 1999), progesterone (Barbagallo et al., 2001) as well as some natural (Ramírez et al., 1998) and synthetic (Perusquía et al., 2003) progestins induce vasorelaxation. Likewise, androgens such as testosterone (Jones et al., 2003) and some of its 5-reduced metabolites (Perusquía et al., 1996) can also induce vasorelaxation. However, few studies have reported the direct effect of sex steroids on human blood vessels. For example, 17β-estradiol and

⁎ Corresponding author. Tel.: +5255 5622 8963; fax: +5255 5622 9198. E-mail address: [email protected] (M. Perusquía). 0024-3205/$ - see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.lfs.2007.07.024

progesterone produce vasorelaxation in human arteries and veins (Mugge et al., 1993; Omar et al., 1995). Nevertheless, the acute exposure of isolated human blood vessels to different progestins and androgens has not been extensively examined in human umbilical blood vessels (e.g. Silva de Sa and Meirelles, 1977; Ramírez et al., 1998; Fausett et al., 1999). The vasorelaxation to sex steroids may involve both endothelium-dependent (Miller and Vanhoutte, 1991) and -independent (Perusquía et al., 1996) pathways. These divergent results may be due to species differences, the vascular bed under study and/or the experimental conditions. Moreover, two main modes of action might help to explain the vasorelaxing response to male and female sex steroids, namely: (i) the control of [Ca2+]i homeostasis, including inactivation of Ca2+ entry via voltagegated (Zhang et al., 1994, 2002; Scragg et al., 2004) and

994

M. Perusquía et al. / Life Sciences 81 (2007) 993–1002

nonvoltage-gated pathways (Jones et al., 2002); and/or (ii) the activation of K+ channels (Deenadayalu et al., 2001; Ding and Stallone, 2001), principally on the large-conductance Ca2+activated K+ channels (BKCa), voltage-dependent K+ channels (Kv) and adenosine triphosphate (ATP)-sensitive K+ channels (KATP). Hence, we set out to investigate the direct effect of a wide series of sex steroids, including dehydroepiandrosterone (DHEA), testosterone, progesterone and their 5-reduced metabolites, on the vascular tone of the human umbilical artery (HUA). For this purpose, we analyzed: (i) the sensitivity of the vasocontractile responses elicited by high KCl (40 mM) or 5hydroxytryptamine (5-HT; 10 μM) to steroid-induced vasorelaxation; and (ii) the vasorelaxing efficacy of each steroid and their molecular structure–vasorelaxing response relationship. In order to shed further light on the mechanism(s) involved in steroid-induced vasorelaxation, we also explored the possible participation of: (i) intracellular steroid receptors; (ii) nuclear transcription/translation mechanisms; (iii) endothelium-derived relaxing factor (nitric oxide; NO); and (iv) K+ (BKCa, Kv and KATP) and voltage-dependent Ca2+ channels.

order to verify the reproducibility of the responses; their amplitude was constant and the tone was stable for more than 70 min in three consecutive stimuli. After each KCl-induced contraction the tissues were washed with Krebs–Henseleit solution to re-equilibrate for 60 min. The same procedure was followed to analyze the contractile responses to 5-HT (0.1, 0.5, 1.0 or 10 μM); these were added separately to different tissues and the response to each concentration was recorded for 60 min. Preliminary results showed that 10 μM 5-HT was the optimal concentration to maintain a stable contraction tone for at least 40 min, which was replicated three times. Likewise, after each 5-HT-induced contraction the tissues were washed with Krebs–

Materials and methods Samples and general methods This study, approved by the Ethics Committee of Woman's Hospital, Ministry of Health of Mexico (Department of Medical Teaching), was performed in accordance with The Declaration of Helsinki. Umbilical cords were obtained, with written consent, from term healthy pregnancies (38–40 weeks) after vaginal delivery and cesarean section. Segments of umbilical cord (∼ 10 cm in length) were cut and placed in ice cold lowCa2+ Ringer of the following composition (mM): NaCl (110), KCl (5), CaCl2 (0.16), MgCl2 (2), NaHCO3 (10), NaH2PO4 (0.5), glucose (10) and EDTA (0.49), resulting in a pH of 6.9; the clean HUA was stored at 4 °C. Experiments were performed 24 and 48 h after HUA had been refrigerated. The HUA was transferred to Krebs–Henseleit solution with the following composition (mM): NaHCO3 (24.9), NaCl (119.5), KCl (4.7), KH2PO4 (1.2), MgSO4 (1.2), CaCl2 (2.5) and glucose (12.0); this solution was gassed continuously with 95% O2 in 5% CO2 to maintain pH 7.4 and constant temperature (37 °C). The HUA segments were cut into rings of 1 cm length, which were: (i) suspended horizontally between two stainless steel wires; and (ii) bathed in individual 10-ml tissue chambers filled with Krebs–Henseleit solution at 37 °C and constantly oxygenated (O2/CO2 95:5). One wire was attached to a fixed support at the bottom of the chamber and the other to an isometric force transducer (FTO3C; Grass Instrument, Quincy, MA). A passive resting tension of 20 mN (2.0 g) was adjusted throughout the experiment and the isometric tension was recorded by a polygraph (79; Grass Instruments). Tissues were allowed to equilibrate for a 2-h period before conducting the experiments. Then, the vasocontractile responses to 40 mM KCl were induced after replacing Krebs–Henseleit solution with an equimolar substitution of 40 mM KCl and 84 mM NaCl. KCl caused a tonic contraction which was repeated three times in

Fig. 1. Inhibitory effect of steroids on KCl-induced contraction in human umbilical artery. A) Typical recording of the contraction induced by high K+ solution (KCl) and the immediate vasorelaxing effect to 5β-DHT at 120 μM; note the contraction recovery after washout (W), showing that the steroid effect was reversible. B) The vehicle for steroids (absolute ethanol 0.1%) had no significant effect (p N 0.05) on KCl-induced contraction. The solid black line indicates the incubation time with steroid or vehicle. C) Concentration–response curves to androgens and progestins on the contraction evoked by KCl. Each concentration tested was significantly different (p b 0.05) as compared with that of vehicle. 5β-DHT and 5α-pregnandione curves were significantly different (⁎p b 0.05) from all steroid curves. I, II, and III groups were significantly different between them (p b 0.05), but the curves included in each group were not different. Each symbol in the graph represents the mean ± SEM of six independent experiments.

M. Perusquía et al. / Life Sciences 81 (2007) 993–1002

995

Table 1 Steroid-induced vasorelaxation on KCl-induced contraction of human umbilical artery Steriod

IC50 (range)

r

Rmax at 120 μM (% of relaxation)

Potency

276.7 μM⁎⁎ (215–346)

1.0

38.9 ± 1.6⁎⁎

0.4

N1 mM

− 0.1

3.5 ± 0.4⁎⁎⁎

b0.1

933.7 μM⁎⁎⁎ (811–1086)

1.0

30.1 ± 1.1⁎⁎⁎

0.1

N1 mM

0.8

11.3 ± 1.4⁎⁎⁎

b0.1

N1 mM

0.9

14.0 ± 0.8⁎⁎⁎

b0.1

119.9 μM (100–133)

1.0

53.3 ± 1.3

1.0

161.8 μM⁎⁎ (138–187)

1.0

46.3 ± 1.4⁎

0.7

N1 mM

0.9

24.5 ± 1.1⁎⁎⁎

b 0.1

56.6 μM⁎⁎ (55–59)

1.0

65.3 ± 1.3⁎⁎

2.1

104.8 μM (95–121)

1.0

54.6 ± 1.3

1.1

N1 mM

0.9

14.5 ± 1.1⁎⁎⁎

b0.1

The values are mean (n = 6) ± SEM. The half maximal relaxation (IC50) was calculated by straight-line regression from every noncumulative concentration–response curve. IC50 was expressed as median (range). Values above 1 mM were marked as N1 mM. Pearson's correlation coefficient (r) represents the fitness of the straight line. The potency was calculated from IC50 by the formula: IC50 DHEA / IC50 metabolite assuming a value of 1.0 to DHEA. ⁎p b 0.005, ⁎⁎p b 0.0005, ⁎⁎⁎p b 0.00005 compared with DHEA values.

996

M. Perusquía et al. / Life Sciences 81 (2007) 993–1002

Henseleit solution to re-equilibrate for 60 min. The preparations that developed a tension of less than 1 g were discarded. Effect of steroids on vascular tone This protocol was addressed to examine the effect of androgens and progestins on KCl-induced contraction. After the two KCl pre-stimuli, a tonic contraction to 40 mM KCl was recorded for 40 min (control response); then, the tissues were washed with normal Krebs–Henseleit solution (re-polarized). When the tone reached the baseline, tissues were allowed to equilibrate for 60 min and then 40 mM KCl was added. After a stable contractile tension was attained (∼10 min), each of the different androgens (DHEA, testosterone, 5α-DHT, 5β-DHT, androsterone and androstanediol) and progestins (progesterone, 5α-pregnandione, 5β-pregnandione, 5α-pregnanolone and 5βpregnanolone) were added independently at different concentrations (6, 12, 60 or 120 μM) in a noncumulative manner; the effect was determined during 30 min. Finally, the rings were washed to re-equilibrate for 60 min before inducing the last contraction to KCl, which was recorded during 40 min. Limitation of the solubility of the above steroids prevented further exploration at concentrations higher than 120 μM. Only one treatment was made in each experiment, which means that the resulting concentration–response curve to each steroid was determined independently in samples from different donors. The concentration– response curves were plotted, and the inhibitory concentration 50 (IC50) for each steroid was calculated. In a separate group of experiments, KCl-induced contraction was determined after exposure to the vehicle of steroids (absolute ethanol); each concentration tested never exceeded 0.1% v/v. In other HUA rings, the relaxant efficacy of 5β-DHT, androsterone, DHEA, testosterone, progesterone, 5β-pregnandione and 5α-DHT (i.e. the most potent vasorelaxing steroids on KCl-induced contraction), was analyzed on the contractile response to 5-HT. For this purpose, after the tissues were prestimulated twice with KCl, a control response to 10 μM 5-HT was recorded for a 40-min period. The tissues were washed with Krebs–Henseleit solution, re-equilibrated for 60 min and subsequently pre-contacted again with 10 μM 5-HT. After a stable tension was attained (∼ 10 min), 120 μM of each of the above steroids was added; this was the highest concentration to induce the maximal relaxation (Rmax) on KCl-induced contraction. The effect of each steroid was recorded separately for 30 min. Finally, the tissues were washed with Krebs–Henseleit solution and, 60 min later, the last contraction to 5-HT was observed for 40 min. In parallel independent experiments, the effect produced by the final volume of steroid vehicle (0.1% absolute ethanol) was analyzed on 5-HT-induced contraction. The effect induced by 120 μM of each steroid on 5-HT- and KCl-induced contractions was compared. Antihormones and inhibitors of protein synthesis and transcription The involvement of steroid receptors in the vasorelaxation to DHEA, testosterone, 5β-DHT and androsterone was examined

with 10 μM flutamide (an androgen receptor antagonist) and 100 μM RU 486 (a progesterone receptor antagonist) to observe the progesterone effect. In each case, the corresponding antihormones were added to establish the desired concentrations and allowed to equilibrate for 30 min on KCl-induced contraction. Then, the corresponding steroids were added independently to observe their Rmax at 120 μM during a 30min period. Since DHEA and testosterone can be bioconverted into estrogens, these two androgens were also tested in the presence of 1 μM ICI 182,780 (a pure estrogen receptor antagonist). In addition, the effect produced by the final volume of vehicle (0.1% absolute ethanol) for each substance was analyzed in independent experiments. The comparison was

Fig. 2. Typical recordings of the contraction to 10 μM 5-hydroxytriptamine (5HT) on isolated human umbilical artery. A) Inhibitory effect of 5β-DHT at 120 μM; note the contraction recovery after washout (W). B) The vehicle for steroid (absolute ethanol 0.1%) did not modify 5-HT-induced contraction. The solid black line indicates the incubation time with steroid or vehicle. C) Comparison of inhibitory effect induced by progesterone (P), 5βpregnandione (5β-PDO), dehydroepiandrosterone (DHEA), testosterone (T), 5α-dihydrotestosterone (5α-DHT), 5β-dihydrotestosterone (5β-DHT) and androsterone (A) on KCl- and 5-HT-induced contraction, when they were tested equimolarly at 120 μM. Statistical significance: ( ⁎ p b 0.05, ⁎⁎ p b 0.0005, ⁎⁎⁎p b 0.00005) as compared with the steroid effect on 5-HT contraction. Steroid-induced vasorelaxation on 5-HT contraction shows that a (DHEA and 5β-DHT) were different from b (P, 5β-PDO, T and A) (p b 0.01). Moreover, c (5α-DHT) (p b 0.001) as well as b resulted different from c (p b 0.05). Each bar represents the mean ± SEM; n = 6.

M. Perusquía et al. / Life Sciences 81 (2007) 993–1002

made with the Rmax of each steroid without any pretreatment. It is to be noted that the concentration tested for each antihormone is high enough to abolish the androgenic (Nazareth and Weigel, 1996), progestational (Attardi et al., 2004) and estrogenic (Wakeling et al., 1991) actions at the receptor level. In other experiments, HUA rings precontracted with KCl were incubated with the protein synthesis inhibitor cycloheximide (40 μM) or the transcription inhibitor actinomycin D (10 μM) for 30 min. Subsequently, the aforementioned steroids (120 μM each) were separately added and their effect (evaluated during a 30-min period) was compared in the absence and in the presence of the above inhibitors. The concentrations used of these inhibitors were high enough to completely inhibit protein synthesis (Waring, 1990) and transcription (Perry and Kelley, 1970). The potential role of NO in the vasorelaxing effect of steroids The potential role of NO was separately explored for the most potent vasorelaxing steroids resulting from the above experiments. The tissues were incubated with the NO synthase inhibitor, Nω-nitro-L-arginine methyl ester (L-NAME, 10 μM), for 30 min on KCl-induced contraction, before eliciting the relaxation to each steroid (120 μM). The vasorelaxing effect of each steroid was observed during 30 min. Their Rmax in HUA without any pretreatment were compared with the Rmax with L-NAME pretreatment. 10 μM L-NAME abolishes the endothelium-dependent relaxation to acetylcholine in rat aortic rings (Rees et al., 1990). Potential interaction of steroids with K+ channels To analyze the possible interaction of steroids with K+ channels, 1 mM tetraethylammonium (TEA; a selective inhibitor of BKCa channels at this concentration), 1 mM 4aminopyridine (4-AP; an inhibitor of Kv channels) or 10 μM glybenclamide (a selective inhibitor of KATP channels) was applied 10 min after the KCl stimulus. 30 min later, the Rmax for DHEA, testosterone, 5β-DHT, androsterone or progesterone (120 μM each) was evaluated for 30 min. The above steroids produced the most potent vasorelaxations on KCl- and 5-HTinduced contraction. The concentration used of each inhibitor of K+ channels is sufficient to antagonize, selectively, those channels in arterial smooth muscle (Nelson and Quayle, 1995).

997

The results were compared with the percentage of inhibition by each steroid without any pretreatment. Potential Ca2+ channel antagonism The most effective vasorelaxant steroids (DHEA, testosterone, 5β-DHT or androsterone) were analyzed in HUA rings depolarized with a KCl–Ca2+ free solution (KCl–Ca2+∅; a depolarizing 40 mM KCl solution modified by addition of 2 mM EGTA and without CaCl2). A transient contraction was obtained to KCl–Ca2+∅ solution and, when the baseline was reached, 2.5 mM CaCl2 was added to evoke a reproducible tonic contraction, which was recorded for 30 min (control); then, the tissues were preincubated with each steroid at different noncumulative concentrations (6, 12, 60 or 120 μM; added separately) 10 min before the addition of CaCl2 at 2.5 mM. Under these conditions, the contraction induced by CaCl2 was recorded for 30 min in the presence of each steroid at every concentration, which was compared with the corresponding control. Subsequently, the tissues were washed out and a CaCl2-induced contraction was elicited again. The washout was done after all CaCl2 contractions, three times, with KCl–Ca2+∅ solution. The concentration–response curves were plotted and the IC50 for each steroid was calculated. Data presentation and statistical analysis All data in the text and figures are expressed as mean ± SEM (n = 6, where n = 1 represents one patient). Changes in tension are shown as percentage of inhibition of the contraction. The concentration for each compound is expressed as the final concentration in the organ bath. The effect of steroids on the KCl- and 5-HTinduced contraction was evaluated by comparing the amplitude of the contraction before (control 100%) and 30 min after steroid addition. The preventive effect of steroids on CaCl2-induced contraction in depolarized tissues was compared with the experimental responses (the previous control contraction set at 100%). The potency of each steroid is expressed as the IC50 (the steroid concentration required to inhibit 50% the contractile response from the control; expressed as median range) obtained by straight-line regression from every noncumulative concentration–response curve. The curves were compared by means of a two-way ANOVA; for multiple comparisons a one-way ANOVA followed by the Tukey's test was used. For comparison of two

Table 2 Maximal relaxation (% of relaxation at 120 μM) induced by steroids in the absence (control) and in the presence of different blockers on KCl-induced contraction of HUA Steroid

Control

Pretreatment Flut

DHEA T 5β-DHT A P

53.3 ± 1.3 46.3 ± 1.4 65.3 ± 1.3 54.6 ± 1.6 38.9 ± 1.3

50.1 ± 1.1 49.5 ± 1.5 65.2 ± 2.0 58.6 ± 1.6

RU 486

53.8⁎ ± 1.8

ICI 182,780

Cycl

Acti

L-NAME

TEA

4-AP

Gly

51.0 ± 2.6 45.1 ± 1.8

53.3 ± 0.9 44.6 ± 0.8 62.0 ± 2.2 55.2 ± 1.8 38.3 ± 0.8

50.1 ± 2.5 44.6 ± 2.2 63.8 ± 2.3 53.3 ± 1.6 37.4 ± 3.1

54.9 ± 2.4 42.9 ± 2.6 63.3 ± 2.1 53.3 ± 1.8 38.7 ± 1.6

50.1 ± 1.4 49.7 ± 1.0 65.0 ± 1.5 55.9 ± 2.0 36.2 ± 0.9

51.3 ± 1.0 46.9 ± 1.4 63.2 ± 1.4 56.9 ± 2.5 37.2 ± 0.5

51.3 ± 0.9 43.8 ± 1.3 63.8 ± 1.0 53.8 ± 2.3 38.0 ± 1.2

DHEA (dehydroepiandrosterone), T (testosterone), 5β-DHT (5β-dihydrotestosterone), A (androsterone), P (progesterone), Flut (flutamide), Cycl (cycloheximide), Acti (actinomycin D), TEA (tetraethylammonium), 4-AP (4-aminopyridine), Gly (glybenclamide). The effect of each steroid was not blocked by the different pretreatments. ⁎Effect significantly augmented (p b 0.0005). Each value represents the mean ± SEM; n = 6.

998

M. Perusquía et al. / Life Sciences 81 (2007) 993–1002

responses, the Student's t-test for unpaired experiments was used. Statistical significance was accepted at p b 0.05. Drugs and chemicals With the exception of ICI 182,780 (Tocris Cookson, Ellisville, MO), the remaining compounds were all purchased from Sigma (St. Louis, MO) and included: DHEA, testosterone, 5α-DHT (17β-hydroxy-5α-androstan-3-one), 5β-DHT (17βhydroxy-5β-androstan-3-one), androsterone (3α,5α-androsterone; 3α-hydroxy-5α-androstan-17-one), androstanediol (3α,5α-androstanediol; 5α-androstane-3α,17β-diol), progesterone, 5α-pregnandione (5α-pregnane-3,20-dione), 5β-pregnandione (5β-pregnane-3,20-dione), 5α-pregnanolone (3βhydroxy-5α-pregnan-20-one), 5β-pregnanolone (3β-hydroxy5β-pregnan-20-one), flutamide, RU 486, cycloheximide, actinomycin D, L-NAME, 5-HT creatinine sulfate, TEA, glybenclamide and 4-AP. All steroids, antihormones, glybenclamide and the inhibitors of protein synthesis and transcription were dissolved in absolute ethanol as a stock solution, and then diluted to working concentrations in absolute ethanol; 10 μl aliquots were added to the 10 ml bath chambers, which never exceeded 0.1% v/v of vehicle. The remaining drugs were dissolved in distilled water. Actinomycin D was kept in the dark to avoid light-induced degradation. The ice cold low-Ca2+ Ringer was used to abate HUA contractility during transport and storage; the functional integrity is preserved, and under physiological conditions (in Krebs–Henseleit solution) the activity is reestablished.

was: 5β-DHT ≫ androsterone =DHEA N testosterone Nprogesterone ≫ 5β-pregnandione N 5α-DHT = androstanediol = 5βpregnanolone≥ 5α-pregnanolone N5α-pregnandione. Regarding the sensitivity of the contraction induced by 5-HT to steroids, the most potent vasorelaxing steroids were tested equimolarly at 120 μM. As illustrated in Fig. 2A, all selected steroids induced also a rapid (2 min) relaxant effect on 5-HT-induced contraction; after washout, this inhibitory (vasorelaxing) effect was reversible. This vasorelaxation was significantly different (p b 0.00005) from that by the corresponding vehicle (0.1% ethanol, which relaxed 1.7 ± 0.3%; see Fig. 2B). Again, 5β-DHTand DHEA displayed a high efficacy to relax the contraction induced by 5-HT, while 5α-DHT showed a moderate relaxing efficacy (Fig. 2C). It is important to highlight that the KCl-induced contraction was significantly more sensitive to steroid-induced relaxation than the contraction induced by 5-HT, as shown in Fig. 2C.

Effect of antihormones and inhibitors of protein synthesis and transcription The most efficacious steroids to induce inhibition on KCl- and 5-HT-induced contraction (i.e. 5β-DHT, androsterone, DHEA,

Results Vasorelaxation induced by steroids in the HUA Fig. 1A shows that the steroids under study relaxed the KClinduced contraction, an effect which was induced within 1 min. A total recovery of the amplitude and tone of KCl-induced contraction was observed after washout. Moreover, this effect (at all concentration tested), was significant (p b 0.05) when compared with that by vehicle (0.1% ethanol, which relaxed no more than 1.2 ± 0.4%, n = 6; see Fig. 1B). Remarkably, the relaxant efficacy of each steroid was different; with the exception of 5α-pregnandione, the remaining steroids induced concentration-dependent vasorelaxations (Fig. 1C). Clearly, the analysis of these curves indicates that the curves to 5β-DHT and 5αpregnandione were significantly different (p b 0.05) when compared with those of the remaining steroids (for further details, see Fig. 1C). Notably, 5β-DHT elicited the most potent vasorelaxation. In addition, the sensitivity (IC50) of the above steroids on KCl-induced contraction and their maximal relaxation (Rmax) are shown in Table 1; with the exception of androsterone, the differences in IC50 and Rmax values obtained with the steroids were significant (p b 0.005) when compared with those of their precursor DHEA. Furthermore, the IC50 and Rmax for 5β-DHT were significantly different (p b 0.00005 and p b 0.0005, respectively) when compared with those for its precursors, DHEA and testosterone. With respect to IC50 values, the order of potency

Fig. 3. Contraction induced by 2.5 mM CaCl2 (Ca2+) in umbilical artery previously depolarized by KCl–Ca2+ free solution (KCl–Ca2+∅). A) Ca2+-induced contraction was notably prevented by 5β-DHT at 120 μM; note the total contraction recovery after washout (W), showing that the Ca2+ antagonic effect of steroid was reversible. The solid black line indicates the incubation time with steroid. B) Concentration-response curves of the Ca2+ antagonic effect induced by the steroids tested. Each point represents the mean (n =6)± SEM. Inhibitory concentration 50 (IC50) was expressed as median (range).

M. Perusquía et al. / Life Sciences 81 (2007) 993–1002

testosterone and progesterone) were selected. Table 2 shows that the vasorelaxation to DHEA, testosterone and its 5-reduced metabolites was not inhibited (p N 0.05) by pretreatment with the androgen receptor antagonist, flutamide, or with the pure antiestrogen, ICI 182,780 for DHEA and testosterone. Likewise, the antiprogestin RU 486 failed to block progesterone-induced vasorelaxation, which was even significantly augmented; this potentiation may have been due to the vasorelaxing properties of RU 486 per se (11.0 ± 1.0%, n =6, see Table 2). Certainly, further studies, which fall beyond the scope of the present investigation, will be required to analyze in detail the slight vasorelaxation produced by RU 486. Furthermore, cycloheximide, actinomycin D or their vehicle (0.1% absolute ethanol each) failed to affect (pN 0.05) the Rmax induced by each steroid on KCl-induced contraction (Table 2). Effect of L-NAME or K+ channels inhibitors on steroid-induced relaxation The Rmax produced by 5β-DHT, androsterone, DHEA, testosterone or progesterone on the control KCl-induced contraction did not significantly (p N 0.05) differ from that obtained after pretreatment with: (i) L-NAME; or (ii) specific inhibitors of BKCa, Kv and KATP channels (TEA, 4-AP and glybenclamide, respectively). The vehicle for glybenclamide (absolute ethanol 0.1%) had no significant effect on the contraction evoked by KCl (Table 2). Effect of steroids on Ca2+-induced contraction After HUA depolarization with KCl–Ca2+∅ solution, a tonic contraction was induced rapidly by CaCl2, which was: (i) antagonized when tissues were preincubated with each steroid; and (ii) reversible after washout (Fig. 3A). The different steroids induced a concentration-dependent blockade on the CaCl2 contractions. The sensitivity (IC50 values) of CaCl2 contraction in HUA previously depolarized to steroids showed that all steroids tested were powerful Ca2+ antagonists (Fig. 3B); the latter finding reveals that 5β-DHT induced a nearly complete prevention of CaCl2 contractions (95.0 ± 1.0%; n = 6). For the sake of comparison, the prevention of CaCl2-induced contraction to the steroids at 120 μM turned out to be significantly higher than their inhibitory effect at 120 μM on KCl- and 5-HT-induced contraction (p b 0.00005 and p b 0.000005, respectively). The final volume of steroid vehicle (0.1% absolute ethanol) did not significantly prevent CaCl2-induced contraction (1.0 ± 0.5%, n = 6; p N 0.05). Discussion General Our study in the isolated HUA: (i) shows that a series of androgens and progestins produce vasorelaxation; and (ii) sheds further light on the possible mechanisms involved in the regulation by steroid hormones. This is, to the best of our knowledge, the first functional line of evidence describing the relaxant effects of 5-reduced androgens and progestins on

999

vascular tone of HUA which may represent a potential regulatory role on fetoplacental blood flow. Accordingly, other studies have also reported the vasorelaxing properties of female and male sex steroids in experimental animals (Mendelsohn and Karas, 1999; Zhang et al., 2002; Perusquía, 2003). Likewise, μM concentrations of 17β-estradiol and progesterone induce a vasorelaxing effect in the isolated human coronary, mammary, placental and omental arteries (Omar et al., 1995; Belfort et al., 1996; Mugge et al., 1997; Ramírez et al., 1998). Nevertheless, very few studies had analyzed the effects of sex steroids in human umbilical blood vessels; indeed, estrogens (Silva de Sa and Meirelles, 1977; Fausett et al., 1999), progesterone and 5β-pregnandione (Ramírez et al., 1998) from 10 to 100 μM, produce vasorelaxation in umbilical arteries and veins, but there was no evidence for androgens. Our study extends these findings on the HUA by showing that several progestins and androgens produced an acute inhibitory (vasorelaxing) response on the contraction to KCl or 5-HT. With the exception of 5α-progesterone (5αpregnandione), KCl-induced contraction was inhibited by the different androgens and progestins (see Table 1). Admittedly, supraphysiological concentrations of steroids were tested in the present study; however, other in vitro findings have also shown a vasorelaxation at the same μM concentrations of steroids, possibly due to their hydrophobic nature. Accordingly, the in vitro assay is limited when using nonpolar steroids, the effect of which can only be detected at high concentrations. Hence, our findings suggest that: (i) sex steroids may act synergistically in vivo at lower, physiological, concentrations to induce vasorelaxation; and (ii) the HUA vasorelaxation by androgens and progestins lies, amongst other possibilities, in the regulation of fetoplacental blood flow, which is one of the most important rate-limiting factors for normal fetal growth. Chemical structure–vasorelaxing activity relationship of androgens and progestins on HUA The fact that androgens and progestins produced vasorelaxation with different potencies (Fig. 1C and Table 1) suggests a chemical structure–vasorelaxing activity relationship. Hence, the A-ring of the steroid nucleus is planar in the structure of DHEA, testosterone, progesterone and the α/trans configuration at C5 of 5α-reduced metabolites. In contrast, the A-ring bends 90° relative to the steroid nucleus when the C5 hydrogen is β/cisoriented in the case of 5β-reduced steroids (see structural conformations in Table 1). Clearly, the structural change of the 5β configuration is critical for an enhanced vasorelaxing effect. Moreover, a slight vasorelaxation was observed on the subsequent 3α- or 3β-hydroxylation, but the efficacy of the 3α,5α configuration was dependent on the inclusion of a 17-keto group (such as androsterone). Thus, the 17-keto group (also present in DHEA; one of the most potent vasorelaxants), may be relevant to increase the vasorelaxing efficacy. Notably, androgens were better vasorelaxants than progesterone. The same structure–activity relationship of steroids has also been observed on the contractility of pregnant human

1000

M. Perusquía et al. / Life Sciences 81 (2007) 993–1002

myometrium at term. DHEA, testosterone and androsterone are equipotent in this preparation, with the 5β-reduced forms of progesterone and testosterone being the most potent relaxants, while the 5α-reduced metabolites were weaker (Perusquía and Jasso-Kamel, 2001; Perusquía et al., 2005); these studies also show that the androgens displayed a greater inhibitory effect than progesterone. Moreover, previous observations in placental and umbilical blood vessels describe that 5β-pregnandione was equipotent to progesterone, whereas 5α-pregnandione and 5α-pregnanolone were inactive (Ramírez et al., 1998). In keeping with the above findings, 5β-DHT is a very potent: (i) vasodilator (Perusquía et al., 1996; Perusquía and Villalón, 1999) and vasodepressor (Perusquía and Villalón, 2002) agent in rats; and (ii) intestinal (Perusquía, 2003) and uterine (Perusquía et al., 2005) relaxant. Moreover, the acute vasorelaxation to 5β-DHT correlates with 5β-reductase activity, which was significantly lower in essential hypertension (Iki et al., 1994). Hence, reduced levels of 5β-reduced steroids may increase the vasoconstrictor tone which may translate into hypertension. Thus, androgens in general (and 5β-DHT in particular) are good vasorelaxing steroids. Potential mechanisms involved in the mode of action of steroids-induced vasorelaxation Our findings basically suggest a nongenomic action rather than the classical steroid nuclear transcription, based on the short latency (∼ 2 min) and relaxation reversal after washouts. This is reinforced by the fact that, in the presence of antihormones or inhibitors of protein synthesis and transcription, steroids still induced a maximal relaxation (Rmax). These findings exclude that progesterone-induced vasorelaxation in human placental blood vessels is mediated by progesterone receptors (Omar et al., 1995) and, accordingly, the nongenomic action of steroids is unrelated to: (i) intracellular receptor occupancy; and (ii) transcription/translation pathways. Moreover, aromatization can be excluded since: (i) the pure antiestrogen did not block the effects of DHEA and testosterone; and (ii) non-aromatizable steroids (progestins and 5reduced androgens) induced relaxation in HUA (present study) and rat vasculature (Perusquía et al., 1996; Deenadayalu et al., 2001; Ding and Stallone, 2001). Interestingly, changes in steroids structure affect their interaction with intracellular receptors to elicit genomic actions. Indeed, 5β-DHT, which is devoid of androgenic activity (Fang et al., 2003), produces a potent relaxation (present results). In contrast, its 5α-epimer (5α-DHT), which has very high affinity and activity at androgen receptors (Fang et al., 2003), produces a weak relaxation (present study). Other studies suggest that steroids induce an endotheliumdependent vasorelaxion (e.g. Miller and Vanhoutte, 1991). However, the endothelium has an atypical function in the HUA since: (i) acetylcholine-induced endothelium-dependent relaxation is weak as compared with other blood vessels (Chaudhuri et al., 1991); and (ii) the biosynthesis of endothelial nitric oxide (NO) is absent (Xie and Triggle, 1994). On this basis, we did not

remove HUA endothelium, but decided to analyze the role of NO on steroid-induced vasorelaxation by using an inhibitor of NO synthase. The fact that L-NAME did not modify this response suggests a NO-independent mechanism, as shown in other blood vessels (Perusquía and Villalón, 1999; Deenadayalu et al., 2001). Consequently, our findings provide evidence that androgens and progestins induce relaxation by interacting with vascular smooth muscle. Thus, steroids-induced vasorelaxation may be operative in the presence of endothelial dysfunction, as described in the maternal (McCarthy et al., 1993) and fetoplacental (Barber et al., 2001) circulations. Considering that the HUA lacks autonomic innervation (Reilly and Russell, 1977; Fox and Khong, 1990), its vascular tone is regulated by local mediators including prostaglandins, 5-HT and/or some ions such as K+ and Ca2+ (Haugen et al., 1991). Thus, the vasorelaxing effect to sex steroids could be due, at least in part, to inhibition of L-type Ca 2+ channels (Zhang et al., 1994, 2002; Scragg et al., 2004) and/or activation of K + channels (Deenadayalu et al., 2001). However, our results showing that steroids-induced vasorelaxation was not affected by inhibitors of BKCa, Kv and KATP channels rules out the involvement of K+ channels. In line with this view, the blocker of KATP channels, glybenclamide: (i) did not affect testosterone-induced relaxation of rabbit coronary arteries (Yue et al., 1995); and (ii) reduced the effect of testosterone on the canine coronary vasculature (Chou et al., 1996). Likewise, 4-AP, but not glybenclamide or TEA, blocked testosterone-induced vasorelaxation of the rat aorta (Ding and Stallone, 2001). In addition, TEA and iberiotoxin attenuated the relaxation to testosterone on porcine coronary arteries via stimulation of BKCa channels (Deenadayalu et al., 2001). Therefore, the nature of K+ channels stimulated by steroids may be heterogeneous depending on the blood vessel under study, the species, and the particular chemical structure of each steroid. Since the involvement of K+ channels seems unlikely in the HUA, the control of extracellular Ca2+ influx by steroids has been proposed in animal blood vessels (Perusquía and Villalón, 1999; Jones et al., 2002). Since the KCl-induced contraction was significantly more sensitive to steroid-induced vasorelaxation than that by 5-HT, the vasocontractile response to 5-HT and KCl involves, at least, two different mechanisms. With respect to 5HT-induced HUA contraction, 5-HT2A and 5-HT1B receptor subtypes are involved (Lovren et al., 1999); indeed, stimulation of these receptors results in an enhanced influx of extracellular Ca2+ via voltage- or receptor-operated Ca2+ channels (Medeiros and Calixto, 1991). In contrast, the HUA contractile response to KCl is mainly due to the influx of extracellular Ca2+ via voltagedependent Ca2+ channels (Dogan et al., 1991; Wylam et al., 1993). Thus, the potent HUA vasorelaxation to steroids on KCl contraction suggests a preferential blockade of voltage-operated (over receptor-operated) Ca2+ channels. Similarly, the fact that 5β-DHT prevented Ca2+-induced contraction in HUA depolarized with KCl–Ca2+∅ is directly associated with a blockade of voltage-operated Ca2+ channels. Accordingly, electrophysiological data have shown an antagonism by 17βestradiol (Zhang et al., 1994), progesterone (Barbagallo et al.,

M. Perusquía et al. / Life Sciences 81 (2007) 993–1002

1001

References

foetuses with androstenedione and testosterone. Acta Endocrinologica 57 (2), 187–207. Chaudhuri, G., Buga, G.M., Gold, M.E., Wood, K.S., Ignarro, L.J., 1991. Characterization and actions of human umbilical endothelium derived relaxing factor. British Journal of Pharmacology 102 (2), 331–336. Chou, T.M., Sudhir, K., Hutchison, S.J., Ko, E., Amidon, T.M., Collins, P., Chatterjee, K., 1996. Testosterone induces dilation of canine coronary conductance and resistance arteries in vivo. Circulation 94 (10), 2614–2619. Deenadayalu, V.P., White, R.E., Stallone, J.N., Gao, X., Garcia, A.J., 2001. Testosterone relaxes coronary arteries by opening the large-conductance, calcium-activated potassium channel. American Journal of Physiology. Heart and Circulatory Physiology 281 (4), H1720–H1727. Ding, A.Q., Stallone, J.N., 2001. Testosterone-induced relaxation of rat aorta is androgen structure specific and involves K+ channel activation. Journal of Applied Physiology 91 (6), 2742–2750. Dogan, N., Cicek, E., Cenik, A.G., Singirik, E., Kilic, M., Ozcan, A.S., 1991. 5-Hydroxytryptamine-induced contraction of human isolated umbilical artery and its dependence on cellular and extracellular Ca++. Archives Internationales de Pharmacodynamie et de Thérapie 312, 79–85. Fang, H., Tong, W., Branham, W.S., Moland, C.L., Dial, S.L., Hong, H., Xie, Q., Perkins, R., Owens, W., Sheehan, D.M., 2003. Study of 202 natural, synthetic, and environmental chemicals for binding to the androgen receptor. Chemical Research in Toxicology 16 (10), 1338–1358. Fausett, M.B., Belfort, M.A., Nanda, R., Saade, G.R., Vedernikov, Y., 1999. The effects of sex steroids on human umbilical artery and vein. Journal of the Society for Gynecologic Investigation 6 (1), 27–31. Fox, S.B., Khong, T.Y., 1990. Lack of innervation of human umbilical cord. An immunohistological and histochemical study. Placenta 11 (1), 59–62. Haugen, G., Bjoro, K., Stray-Pedersen, S., 1991. Vasoactive effects of intra- and extravascular serotonin, PGE2 and PGF2α in human umbilical arteries. Gynecologic and Obstetric Investigation 31 (4), 208–212. Iki, K., Miyamori, I., Hatakeyama, H., Yoneda, T., Takeda, Y., Takeda, R., Dai, Q.L., 1994. The activities of 5β-reductase and 11β-hydroxysteroid dehydrogenase in essential hypertension. Steroids 59 (11), 656–660. Jones, R.D., English, K.M., Pugh, P.J., Morice, A.H., Jones, T.H., Channer, K.S., 2002. Pulmonary vasodilatory action of testosterone: evidence of a calcium antagonistic action. Journal of Cardiovascular Pharmacology 39 (6), 814–823. Jones, R.D., Pugh, P.J., Jones, T.H., Channer, K.S., 2003. The vasodilatory action of testosterone: a potassium-channel opening or a calcium antagonistic action? British Journal of Pharmacology 138 (5), 733–744. Lovren, F., Li, X.F., Lytton, J., Triggle, C., 1999. Functional characterization and m-RNA expression of 5-HT receptors mediating contraction in human umbilical artery. British Journal of Pharmacology 127 (5), 1247–1255. McCarthy, A.L., Woolfson, R.G., Raju, S.K., Poston, L., 1993. Abnormal endothelial cell function of resistance arteries from women with preeclampsia. American Journal of Obstetrics and Gynecology 168 (4), 1323–1330.

Attardi, B.J., Burgenson, J., Hild, S.A., Reel, J.R., 2004. In vitro antiprogestational/antiglucocorticoid activity and progestin and glucocorticoid receptor binding of the putative metabolites and synthetic derivatives of CDB-2914, CDB-4124, and mifepristone. Journal of Steroid Biochemistry and Molecular Biology 88 (3), 277–288. Barbagallo, M., Dominguez, L.J., Licata, G., Shan, J., Bing, L., Karpinski, E., Pang, P.K., Resnick, L.M., 2001. Vascular effects of progesterone: role of cellular calcium regulation. Hypertension 37 (1), 142–147. Barber, A., Robson, S.C., Myatt, L., Bulmer, J.N., Lyall, F., 2001. Heme oxygenase expression in human placenta and placental bed: reduced expression of placenta endothelial HO-2 in preeclampsia and fetal growth restriction. FASEB Journal 15 (7), 1158–1168. Belfort, M.A., Saade, G.R., Suresh, M., Vedernikov, Y.P., 1996. Effects of estradiol-17β and progesterone on isolated human omental artery from premenopausal nonpregnant women and from normotensive and preeclamptic pregnant women. American Journal of Obstetrics and Gynecology 174 (1 Pt 1), 246–253. Benagiano, G., Mancuso, S., Mancuso, F.P., Wiqvist, N., Diczfalusy, E., 1968. Studies on the metabolism of C-19 steroids in the human foeto-placental unit. 3. Dehydrogenation and reduction products formed by previable

McClamrock, H.D., Adashi, E.Y., 1992. Gestational hyperandrogenism. Fertility and Sterility 57 (2), 257–274. Medeiros, Y.S., Calixto, J.B., 1991. Influence of calcium entry blockers and calmodulin inhibitors on 5-hydroxytryptamine-, potassium- and calciuminduced contractions in human umbilical artery in-vitro. Journal of Pharmacy and Pharmacology 43 (6), 411–416. Mendelsohn, M.E., Karas, R.H., 1999. The protective effects of estrogen on the cardiovascular system. The New England Journal of Medicine 340 (23), 1801–1811. Miller, V.M., Vanhoutte, P.M., 1991. Progesterone and modulation of endothelium-dependent responses in canine coronary arteries. American Journal of Physiology 261 (4 Pt 2), R1022–R1027. Mugge, A., Riedel, M., Barton, M., Kuhn, M., Lichtlen, P.R., 1993. Endothelium independent relaxation of human coronary arteries by 17βoestradiol in vitro. Cardiovascular Research 27 (11), 1939–1942. Mugge, A., Barton, M., Fieguth, H.G., Riedel, M., 1997. Contractile responses to histamine, serotonin, and angiotensin II are impaired by 17β-estradiol in human internal mammary arteries in vitro. Pharmacology 54 (3), 162–168. Nazareth, L.V., Weigel, N.L., 1996. Activation of the human androgen receptor through a protein kinase A signaling pathway. Journal of Biological Chemistry 271 (33), 19900–19907.

2001; Zhang et al., 2002) and testosterone (Scragg et al., 2004) on L-type Ca2+ currents of rat myocytes. Admittedly, further patch-clamp experiments will be required to define whether this mechanism is operative in HUA cells. Physiological relevance Vasoconstriction of the umbilical cord results in growth retardation and intrauterine death of the fetus. Therefore, the capability of androgens and progestins to produce a nongenomic vasorelaxation may be physiologically relevant. Indeed, these steroids are produced in the materno-fetoplacental unit (Benagiano et al., 1968) and have an important uterine relaxing effect in pregnant women (Perusquía and Jasso-Kamel, 2001; Perusquía et al., 2005); hence they may contribute to two major adaptative responses during pregnancy, namely, relaxation of both uterine and vascular smooth muscle. This propregnancy role is consistent with the markedly increased plasma levels, throughout pregnancy, of progesterone, DHEA, androstenedione and testosterone (McClamrock and Adashi, 1992). Thus, it is reasonable to propose that a low level of steroids during pregnancy, particularly the androgens reported here, could partly be the cause of preeclamsia/eclamsia and, consequently, exogenously applied vasorelaxing steroids may be therapeutically relevant for the endothelial dysfunction described in the maternal and fetoplacental circulation as well as in the treatment of gestational hypertension. Acknowledgments We very much appreciate J. García and R. Toledo for their technical assistance, Dr. E. Calixto and the midwives of General Hospital-IMSS No. 4 and No. 98 for collecting umbilical cords as well as the assistance of Professor Marco A. José in the statistical analysis is much appreciated. This study was supported by the Universidad Nacional Autónoma de México (DGAPA/PAPIIT project No. IN202507-3, and No. IN221102-2).

1002

M. Perusquía et al. / Life Sciences 81 (2007) 993–1002

Nelson, M.T., Quayle, J.M., 1995. Physiological roles and properties of potassium channels in arterial smooth muscle. American Journal of Physiology 268 (4 Pt 1), C799–C822. Omar, H.A., Ramírez, R., Gibson, M., 1995. Properties of a progesteroneinduced relaxation in human placental arteries and veins. Journal of Clinical Endocrinology and Metabolism 80 (2), 370–373. Perry, R.P., Kelley, D.E., 1970. Inhibition of RNA synthesis by actinomycin D: characteristic dose-response of different RNA species. Journal of Cellular Physiology 76 (2), 127–139. Perusquía, M., 2003. Androgen-induced vasorelaxation: a potential vascular protective effect. Experimental and Clinical Endocrinology & Diabetes 111 (2), 55–59. Perusquía, M., Villalón, C.M., 1999. Possible role of Ca2+ channels in the vasodilating effect of 5β-dihydrotestosterone in rat aorta. European Journal of Pharmacology 371 (2-3), 169–178. Perusquía, M., Jasso-Kamel, J., 2001. Influence of 5α- and 5β-reduced progestins on the contractility of isolated human myometrium at term. Life Sciences 68 (26), 2933–2944. Perusquía, M., Villalón, C.M., 2002. The vasodepressor effect of androgens in pithed rats: potential role of calcium channels. Steroids 67 (13-14), 1021–1028. Perusquía, M., Hernández, R., Morales, M.A., Campos, M.G., Villalón, C.M., 1996. Role of endothelium in the vasodilating effect of progestins and androgens on the rat thoracic aorta. General Pharmacology 27 (1), 181–185. Perusquía, M., Villalón, C.M., Navarrete, E., García, G.A., Pérez-Palacios, G., Lemus, A.E., 2003. Vasodilating effect of norethisterone and its 5α metabolites: a novel nongenomic action. European Journal of Pharmacology 475 (1-3), 161–169. Perusquía, M., Navarrete, E., Jasso-Kamel, J., Montaño, L.M., 2005. Androgens induce relaxation of contractile activity in pregnant human myometrium at term: a nongenomic action on L-type calcium channels. Biology of Reproduction 73 (2), 214–221. Ramírez, R.J., Gibson, M., Kalenic, J., Einzig, S., Omar, H.A., 1998. In vitro vascular relaxation to progesterone and its metabolites in human umbilical

and placental blood vessels. Journal of Maternal-Fetal Investigation 8 (2), 61–65. Rees, D.D., Palmer, R.M., Schulz, R., Hodson, H.F., Moncada, S., 1990. Characterization of three inhibitors of endothelial nitric oxide synthase in vitro and in vivo. British Journal of Pharmacology 101 (3), 746–752. Reilly, R.D., Russell, P.T., 1977. Neurohistochemical evidence supporting an absence of adrenergic and cholinergic innervation in the human placenta and umbilical cord. Anatomical Record 188 (3), 277–286. Scragg, J.L., Jones, R.D., Channer, K.S., Jones, T.H., Peers, C., 2004. Testosterone is a potent inhibitor of L-type Ca2+ channels. Biochemical and Biophysical Research Communications 318 (2), 503–506. Silva de Sa, M.F., Meirelles, R.S., 1977. Vasodilating effect of estrogen on the human umbilical artery. Gynecologic Investigation 8 (5-6), 307–313. Wakeling, A.E., Dukes, M., Bowler, J., 1991. A potent specific pure antiestrogen with clinical potential. Cancer Research 51 (15), 3867–3873. Waring, P., 1990. DNA fragmentation induced in macrophages by gliotoxin does not require protein synthesis and is preceded by raised inositol triphosphate levels. Journal of Biological Chemistry 265 (24), 14476–14480. Wylam, M.E., Samsel, R.W., Schumacker, P.T., Umans, J.G., 1993. Extracellular calcium and intrinsic tone in the human umbilical artery. Journal of Pharmacology and Experimental Therapeutics 266 (3), 1475–1481. Xie, H., Triggle, C.R., 1994. Endothelium-independent relaxations to acetylcholine and A23187 in the human umbilical artery. Journal of Vascular Research 31 (2), 92–105. Yue, P., Chatterjee, K., Beale, C., Poole-Wilson, P.A., Collins, P., 1995. Testosterone relaxes rabbit coronary arteries and aorta. Circulation 91 (4), 1154–1160. Zhang, F., Ram, J.L., Standley, P.R., Sowers, J.R., 1994. 17β-Estradiol attenuates voltage-dependent Ca2+ currents in A7r5 vascular smooth muscle cell line. American Journal of Physiology 266 (4 Pt 1), C975–C980. Zhang, M., Benishin, C.G., Pang, P.K., 2002. Rapid inhibition of the contraction of rat tail artery by progesterone is mediated by inhibition of calcium currents. Journal of Pharmacy and Pharmacology 54 (12), 1667–1674.