Preeclampsia is associated with an increase of circulating levels of microparticles (MPs), but their role in vascular dysfunction during the course of preeclampsia is not understood. Inasmuch as preeclampsia is a gestational disease, we tested the effect of MPs from preeclamptic women (PrMPs) and MPs from normal pregnant women (CMPs) on vessels from pregnant mice. We exposed aortic rings from pregnant mice to circulating levels of PrMPs or CMPs for 24 h and evaluated their response to serotonin (5-HT). PrMPs, but not CMPs, were able to induce hyporeactivity in response to 5-HT in aortas from pregnant mice. The nitric oxide (NO) synthase inhibitor NG-nitro-l-arginine strongly enhanced the response to 5-HT in PrMP-treated vessels but had no significant effect on CMP-treated vessels. The 5-HT-induced contraction in PrMP-treated vessels was completely abolished by the selective cyclooxygenase-2 (COX-2) inhibitor NS-398 but was only reduced in CMP-treated vessels, suggesting an increased participation of COX-2 vasoconstrictor products in the effect of PrMPs. Consistent with this hypothesis, PrMPs enhanced levels of 8-isoprostane and PGE2 in vessels, despite reduction of thromboxane B2. These results strengthen the main concept that MPs in preeclampsia could act as vectors to stimulate intracellular cascades in vascular cells, leading to an enhanced NO production to counteract increased COX-2 vasoconstrictor metabolites by taking into account pregnancy.
- nitric oxide
one of the most common hemodynamic disorders during pregnancy is hypertension; if it is associated with proteinuria, the gestational hypertension becomes preeclamptic syndrome (23). Preeclampsia is a multiorgan disorder associated with a generalized endothelial dysfunction, resulting in fetal growth retardation. Despite extensive research, the mechanisms involved in this vascular dysfunction are not well understood. The preeclamptic endothelium overexpresses procoagulant factors, and endothelium-dependent regulation of vascular tone is impaired and even abolished (24, 25). Nevertheless, it is unclear whether the altered vascular tone observed in preeclampsia results from changes in circulating vasoactive substances (5, 16, 34) or alterations of the vascular smooth muscle signaling and contraction systems.
Recently, several groups reported elevated circulating levels of shed membrane microparticles (MPs) in preeclampsia and, subsequently, suggested their involvement in the unrelenting hypertension associated with this disease. MPs are fragments released from the plasma membrane of stimulated or apoptotic cells (2, 35). Although the total number of circulating MPs was unaltered in preeclampsia, the proportions of T lymphocyte and granulocyte MPs are increased (30). Circulating MPs from these patients abolish endothelium-dependent relaxation, in contrast to the effect of MPs from healthy pregnant women, in isolated myometrial arteries (32).
We have been interested in comparing the vascular reactivity to vasoconstrictor agonists in MPs from preeclamptic women (PrMPs) with that in MPs from normal pregnant women (CMPs). We recently found that PrMPs, but not CMPs, induce vascular hyporeactivity in human omental small arteries and aortas from nonpregnant mice, which is associated with release of proinflammatory factors and a complex interaction between nitric oxide (NO) synthase (NOS) and cyclooxygenase (COX) systems (18). However, these studies were performed in vessels from nonpregnant mice. Preeclampsia is a hypertensive disorder of pregnancy. Moreover, inasmuch as pregnancy is characterized by specific hemodynamic modifications (21), experiments were designed to compare effects of CMPs with effects of PrMPs on vessels from pregnant animals to study the mechanisms by which pregnancy might modify the effects of PrMPs. The results would strengthen the main concept of a role of MPs as vectors that are able to activate intracellular cascades in vascular cells, which could promote vascular activation in preeclamptic women by taking into account pregnancy.
The Ethics Committee of the University Hospital of Strasbourg approved this study. MPs were harvested from women with (n = 21) and without (n = 17) preeclampsia.
According to standard criteria (4), preeclampsia is defined by the de novo appearance of hypertension (systolic blood pressure ≥140 mmHg or diastolic blood pressure ≥90 mmHg) associated with new-onset proteinuria, defined as a dipstick measurement of ≥2 per 24 h detected for the first time after 20 wk of pregnancy. Symptoms of the patients enrolled in the present study were persistent headache (75%), abdominal pain (62%), and systolic blood pressure of 164 ± 20 mmHg and diastolic blood pressure of 111 ± 10 mmHg. They exhibited a new-onset proteinuria ⩾3 associated with clinical edema and subsequent preterm birth and intrauterine growth restriction. Although the patients had the severe form of preeclampsia, they did not display characteristics of central nervous system, renal, or respiratory failure. All patients were admitted to the intensive care unit for monitoring and control of their blood pressure. Patients with preexistent hypertension, gestational diabetes mellitus, coagulation abnormalities, and previous renal or hepatic diseases were not included in the study. Blood samples from normal pregnant and preeclamptic women were collected in 13 × 75 mm 4.5-ml glass plasma tubes (BD Vacutainer) containing 0.138 mol/l trisodium citrate solution at a final volume ratio of 9:1. Circulating MPs were harvested after the plasma samples were centrifuged for 15 min at 1,500 g and then for 2 min at 14,000 g, the MPs were pelleted and recovered in 1 ml of RPMI medium, and the samples were centrifuged again for 45 min at 13,000 g. Supernatants corresponding to the last MP wash medium were used as control.
MPs were quantified as nanomolar phosphatidylserine equivalents (nM PS eq) and phenotyped with specific monoclonal antibodies (anti-GPIβα, anti-CD11a, and anti-CD31), as described elsewhere (12). As we previously reported (11), overall circulating MP levels were significantly greater in preeclamptic patients (n = 21) than in normal pregnant women (n = 17): 11.68 ± 1.09 vs. 7.5 ± 1.35 nM PS eq (P < 0.05). Furthermore, levels of lymphocyte- and platelet-derived MPs were elevated in preeclamptic patients compared with normal pregnant women: 5.85 ± 0.61 vs. 2.9 ± 0.65 nM PS eq for lymphocyte-derived MPs and 5.83 ± 0.65 vs. 4.2 ± 0.94 nM PS eq for platelet-derived MPs (18). Thus we found twice as many platelet- as lymphocyte-derived MPs in the normal pregnant women, whereas the levels of the two types of MPs were similar in the preeclamptic patients. Therefore, the total levels of MPs were increased and the proportions of lymphocyte- and platelet-derived MPs were modified in blood of preeclamptic patients.
Endotoxin levels in all MP preparations were assessed using the Limulus amoebocyte lysate kit (catalog no. QCL-1000, Cambrex) and were found to be below the lower detection limit of the kit (<0.1 endotoxin unit/ml).
Female C57BL/6 mice that were 12 wk old at the 18th day of pregnancy were anesthetized intraperitoneally with a mixture of ketamine (100 mg/kg), medetomidine (50 μg/kg), and heparin (500 U/kg). They were killed by decapitation, and aortic rings were harvested. The experiments were conducted in compliance with statutory requirements by an accredited research scientist (authorization no. 02812 from the French Department of Agriculture).
After 24 h of randomized incubation in 50% RPMI-50% M199 supplemented with 10% bovine fetal serum with or without overall circulating levels of PrMPs or CMPs for each patient, mouse aortas were mounted on a wire myograph (EMKA Technologies, Paris, France). The cellular origins of the circulating PrMPs or CMPs differed (i.e., composition similar to that observed in the patients) to better mimic their effect in vivo in the patients’ blood. After they were centrifuged, MPs were pelleted and recovered in 0.9 ml of RPMI without serum and stored at −80°C until subsequent use. For each patient, 0.1 ml of serum was added to the MP suspension, and this 10% serum was used to incubate vessels for 24 h.
Mechanical activity was recorded isometrically by a force transducer. The experiments were performed at 37°C with physiological salt solution continuously bubbled with 95% O2-5% CO2.
After the vessels were set to their working length, they were challenged twice with KCl (100 mM)-depolarized physiological salt solution containing 10 μM phenylephrine (Sigma-Aldrich, Saint Quentin Fallavier, France) to elicit reproducible contractile responses to test the maximal contractile capacity of the vessels. The vessels were washed and allowed to rest for an additional 20 min, and the function of the endothelium was tested by the presence of acetylcholine-induced relaxation subsequent to 80% precontraction with 1–3 μM phenylephrine. The vessels were washed, and concentration-response curves were elicited by cumulative application of 0.1 nM–10 μM 5-HT (Sigma-Aldrich) to endothelium-intact vessels in the absence or presence of inhibitors: the NO synthase inhibitor NG-nitro-l-arginine (l-NNA, 100 μM; Sigma-Aldrich), the selective COX-2 inhibitor N-(2-cyclohexyloxy-4-nitrophenyl) methanesulfonamide (NS-398, 10 μM; Sigma-Aldrich), and l-NNA + NS-398. The inhibitors were added in the bath 30 min before addition of 5-HT.
The treatment of vessels with MPs for 24 h corresponded to the time required for maximal effect. Thus all experiments were performed under these conditions. Moreover, regardless of the circulating level of MPs from preeclamptic women, we obtained a reduction of vascular contraction similar to that in vessels from nonpregnant mice (18).
NO spin trapping and electronic paramagnetic resonance studies.
NO production was detected using the electronic paramagnetic resonance (EPR) technique with Fe2+-diethyldithiocarbamate (DETC; Sigma-Aldrich) as spin trap according to Mulsch et al. (19). After administration of 5-HT, vessels preincubated for 24 h with circulating levels of PrMPs or CMPs or in a medium without MPs were treated with 250 μl of colloid Fe(DETC)2 and incubated for 1 h at 37°C. These studies were performed using the instrument settings previously described (29).
Determination of prostanoid production.
After incubation in a medium with or without PrMPs or CMPs, vessels were treated with 10 μM 5-HT for 20 min at 37°C. The medium was collected, and thromboxane B2, prostaglandin E2 metabolites, and 8-isoprostane were measured (pg/mg dry tissue wt) by enzyme immunoassays kits (Cayman Chemical, SPI-BIO, Montigny le Bretonneux, France).
One-way ANOVA, Kruskal-Wallis, and Mann-Whitney U tests or one-way ANOVA for repeated measures followed by Tukey's post hoc test were performed using Statview version 5.0 software (SAS Institute, Cary, NC). P < 0.05 was considered to be statistically significant. pD2 represents the negative logarithm of the half maximally effective molar concentration. Values are means ± SE for n experiments; n represents the number of animals.
PrMPs decrease 5-HT-stimulated contraction in pregnant mouse aortic rings.
The contractile response to any of these vasoconstrictor agents did not differ between vessels incubated with medium without MPs and vessels incubated with CMPs (Table 1).
5-HT produced a concentration-dependent increase in tension in vessels with functional endothelium. Incubation of mouse aortic rings with CMPs for 24 h did not significantly affect the contractile response to 5-HT compared with vessels not incubated with MPs [Table 1, Fig. 1; pD2 = 6.71 ± 0.21 vs. 6.80 ± 0.14, n = 5, P = not significant (NS)]. In contrast, PrMPs decreased 5-HT-induced vascular contraction, but not sensitivity to 5-HT (Table 1, Fig. 1; pD2 = 6.89 ± 0.21, n = 5).
Together, these data demonstrate that 1) the mechanism by which PrMPs reduce contraction is independent of receptor coupling and 2) the reduced response is related to PrMPs, inasmuch the response in the presence of CMPs is identical to that in vessels that are not incubated with MPs. These data also show that if there is basal release of substances from the particles or the tissue, they are not different in the absence of CMPs or after incubation of vessels with at least CMPs.
l-NNA did significantly enhance 5-HT-stimulated contraction in arteries from pregnant mice incubated with CMPs [maximum contractile response (Emax) = 1.31 ± 0.21 vs. 1.01 ± 0.13 mN/mm, pD2 = 7.08 ± 0.26 vs. 6.74 ± 0.19, P = NS; Fig. 2A] or without MPs (not shown). However, l-NNA markedly increased the response to the same agonist in PrMP-treated vessels [Emax = 1.01 ± 0.14 vs. 0.53 ± 0.08 mN/mm (P < 0.001), pD2 = 7.01 ± 0.24 vs. 6.88 ± 0.21 (P = NS), n = 5; Fig. 2B]. Thus, in presence of l-NNA, 5-HT-induced contraction in PrMP-incubated vessels was restored to that in control and CMP-treated arteries.
Moreover, endothelium-intact aortic rings from pregnant mice treated with PrMPs and preincubated with Fe(DETC)2 exhibited an EPR feature of signals derived from NO-Fe(DETC)2. Incubation of aortas for 24 h with PrMPs elicited a nearly twofold increase in the NO-Fe(DETC)2 EPR signal compared with that obtained from CMP-incubated vessels (Fig. 2C). NO production in control vessels was not significantly different from that in CMP-treated aortas (Fig. 2C). Together, these data strongly suggest that enhanced NO production by PrMPs accounts for the hyporeactivity of aortas in response to 5-HT, which is prevented in the presence of the NO synthase inhibitor l-NNA.
The selective inhibitor of COX-2, NS-398, significantly reduced the contractile response to 5-HT in vessels treated with CMPs [Emax = 0.33 ± 0.06 vs. 0.99 ± 0.11 mN/mm (P < 0.01), pD2 = 6.3 ± 0.35 vs. 6.74 ± 0.19 (P = NS); Fig. 3A], and, surprisingly, NS-398 almost abolished the contraction to the same agonist in arteries from mice treated with PrMPs [Emax = 0.08 ± 0.01 vs. 0.53 ± 0.08 mN/mm (P < 0.01), pD2 = 6.62 ± 0.36 vs. 6.88 ± 0.21 (P = NS); Fig. 3B]. These results suggest that CMPs and PrMPs stimulate the release of COX-2 vasoconstrictor products. NS-398 did not affect the contractile response to 5-HT in vessels incubated in medium without MPs (not shown). Finally, the association of NO and COX-2 inhibitors did not modify 5-HT-dependent contraction in CMP-treated aortas (Emax = 1.08 ± 0.03 vs. 0.99 ± 0.11 mN/mm, pD2 = 6.7 ± 0.13 vs. 6.74 ± 0.19, P = NS; Fig. 3A); in contrast, the combination of the two inhibitors strongly enhanced the contraction in response to 5-HT administration in aortas from pregnant mice treated with PrMPs [Emax = 0.94 ± 0.03 vs. 0.53 ± 0.08 mN/mm (P < 0.01), pD2 = 6.77 ± 0.11 vs. 6.88 ± 0.21 (P = NS); Fig. 3B]. Thus the concentration-response curves to 5-HT were not significantly different from those obtained in control or CMP-treated vessels.
MPs stimulate release of vasoconstrictor prostanoids.
Assay for thromboxane B2, the stable metabolite of thromboxane A2, showed less thromboxane A2 production in aortas from pregnant mice treated with PrMPs than in those incubated with CMPs (Fig. 4). The release of PGE2 was enhanced in PrMP-treated arteries compared with CMP-treated vessels (Fig. 4). Moreover, 8-isoprostane synthesis was increased in aortas treated with PrMPs compared with those treated with CMPs (Fig. 4). Prostanoid levels in vessels incubated in a medium without MPs were not significantly different from those in CMP-treated aortas (Fig. 4).
We provide evidence that PrMPs induced a decrease of 5-HT-stimulated vascular contraction in aortic rings from pregnant mice. These effects were observed at circulating levels of PrMPs and were associated with increased participation of NO and COX-2 metabolites, PGE2 and 8-isoprostane, despite reduced release of thromboxane A2. In this context, the balance of the release of these products is shifted toward enhanced participation of NO to modulate vascular reactivity. PrMP-induced reduction of 5-HT-induced vascular contraction in human omental arteries and aortas from nonpregnant mice (18), as well as in vessels from pregnant mice, suggests that PrMPs, per se, can elicit vascular hyporeactivity to vasoconstrictor agonists independently of pregnancy and arteries from different species. These data suggest that PrMPs could play a main role in regulation of the increased response to vasoconstrictors in preeclamptic patients.
It is known that preeclampsia is associated with systemic vascular inflammation by neutrophil activation (15) and oxidative stress (7, 11), leading to alteration of vascular reactivity at the level of the endothelium and vascular smooth muscle cells by decreased production of endothelium-relaxing factors (16). Moreover, this syndrome is characterized by an increased pressure response to vasoconstrictors such as endothelin-1 and angiotensin II (5), leading to enhanced vascular reactivity (14). Despite extensive research, little is known about the mechanisms that drive changes in vascular reactivity during preeclampsia, especially the possible role of circulating MPs in these patients. In a previous study (30), although the same total levels of circulating MPs were detected in normal pregnancy and preeclampsia, increased T cell and granulocyte MP rates were found in the blood of preeclamptic women. Moreover, a close relationship between endothelial dysfunction and circulating levels of endothelial MPs has been reported in preeclampsia (9). Platelet-derived MPs were reduced (30) or unchanged (9) in the two groups of patients in these studies.
We previously found (18) elevated levels of leukocyte- and platelet-derived MPs in the bloodstream of women affected by preeclampsia compared with normal pregnant women.
As expected, PrMPs could have a deleterious effect on hypertension, which characterizes this disease. On the contrary and in accordance with recent reviews (17, 31), we found a reduction in vascular contraction in PrMP-treated vessels. Nevertheless, we could provide a hypothesis in light of our previous study (18), in which we examined individually the effect of non-platelet- and platelet-derived PrMPs. Our results suggest that platelet-derived PrMPs might have a protective role in the vascular wall because of their main contribution to the overproduction of NO with subsequent reduction of vascular contraction, whereas non-platelet-derived PrMPs are mainly involved in COX-2-dependent vasoconstrictor production, with a possible role in the induction of preeclamptic hypertension. Thus platelet-derived PrMPs are mainly protective and non-platelet-derived PrMPs are mainly deleterious in the vascular changes observed in preeclampsia.
Furthermore, the circulating level of MPs from preeclampsic patients and normal pregnant women had a similar effect in terms of reduced vascular contraction. One can advance the hypothesis that the alteration in vascular reactivity was linked to the different composition and the constituents of PrMPs.
In our previous study (29), we found that MPs act on smooth muscle cells independently of the presence of the endothelium and, alternatively, change the balance between relaxant and constrictor factors in smooth muscle. We report a significant participation of NO in the effects of PrMPs in aortas from pregnant mice. The origin of this increase in NO release is most likely the consequence of the ability of PrMPs to induce inducible NOS (iNOS) expression in the vascular wall, as previously reported (18).
The discrepancy between the release of elevated amounts of NO and hypertension is striking in light of a simultaneous increase in participation of COX-2 (26, 27), which has been reported in pregnant animals (26) and in preeclamptic women (13). Indeed, in pathophysiological conditions where COX-2 is upregulated in the vascular wall and NO is released from iNOS, peripheral resistance is usually low in the presence of hypotension (28). Two studies have established that MPs are able to upregulate COX-2 (1, 2); moreover, in preeclamptic women, plasma concentrations of substances such as vasoconstrictor eicosanoids are increased (33) and placental isoprostane level in biological fluids is enhanced (34). In the present work, PrMPs enhanced the release of vasoconstrictor metabolites such as PGE2 and 8-isoprostane (8) by a mechanism sensitive to the COX-2 inhibitor NS-398 in aortas from pregnant mice. This suggests that PrMPs, but not CMPs, directly or indirectly modify the activity of 5-HT with a shift toward substantial vasoconstriction mediated by PGE2 and isoprostanes (6, 22).
Pregnancy is characterized by a decrease in vascular resistance attributed to a modification of hormonal production and an increased role of endogenous vasodilators (10, 21), especially the enhanced level of circulating NO by upregulation of endothelial NOS (20). Pregnancy might also alter the effects of PrMPs, with enhanced production of vasoconstrictor metabolites, leading to hypertension, which is the main characteristic of preeclampsia. Surprisingly, PrMP-induced vascular inflammation, by inducing iNOS and COX-2 expression, and subsequent vascular hyporeactivity to 5-HT occur independently of pregnancy (18). The only differential effect of PrMPs in arteries from pregnant and nonpregnant mice is due to the changes in the COX-2 metabolite released. Thus PrMPs enhanced PGE2 release in vessels from pregnant mice, whereas they did not modify the release of the same metabolite in arteries from nonpregnant mice (18). Therefore, pregnancy might induce biological diversion of the nature of the COX-2 metabolite release, although the functional consequence of PrMPs on vascular reactivity remains unchanged in pregnant and nonpregnant mice. The reason for such changes requires further investigation to understand whether it is linked to COX-2 or to an environmental condition of the vessels.
In conclusion, our previous work (18) and the present study provide evidence that PrMPs have a role in vascular function during preeclampsia that is associated with their proinflammatory properties by their ability to induce iNOS and COX-2 expression in the vascular wall via NF-κB. Interestingly, although PrMPs induce NO overproduction with subsequent reduction of vascular contraction, at the same time they increase the production of vasoconstrictor metabolites from COX-2, which has a possible role in the increase in blood pressure during preeclampsia (18). This is strengthened by the elevated levels of 8-isoprostane found in this study in PrMP-treated vessels and previously detected by Walsh et al. (34) in the placenta of preeclamptic women. Finally and most importantly, pregnancy does not modify the functional role of PrMPs as proinflammatory bioeffectors in promoting changes in vascular contraction during preeclampsia.
Partial financial support for this study was provided by School of Medicine of the Université Louis Pasteur (Strasbourg) Grant 2002-5196 and Fondation de la Recherche Médicale Grant INE20050303433 (to R. Andriantsitohaina).
We thank Drs. L. Grunebaum and D. Desprez for help with MP preparation. We gratefully acknowledge the valuable technical assistance of J. F. Poirier.
Present address of F. Meziani: Département de réanimation médicale et médecine hyperbare, CHU Angers, France.
↵* A. Tesse and F. Meziani contributed equally to this work.
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 © 2007 by the American Physiological Society