- Email: [email protected]
Placental Superoxide is Increased in Pre-eclampsia
Placenta (2001), 22, 304–308 doi:10.1053/plac.2001.0629, available online at http://www.idealibrary.com on
Placental Superoxide is Increased in Pre-eclampsia J. M. Sikkemaa,e, B. B. van Rijna, A. Franxa, H. W. Bruinsea, R. de Roosb, E. S. G. Stroesc and E. E. van Faassend a Department of Obstetrics, Gynaecology and Neonatology, b Department of Nephrology, c Department of Vascular Medicine, University Medical Centre Utrecht and d The Debye Institute, Section Interface Physics, University of Utrecht, The Netherlands Paper accepted 5 January 2001
One of the current hypotheses on the pathophysiology of pre-eclampsia (PE) states that the placenta secretes one or more cytotoxic factors resulting in maternal endothelial dysfunction. Among the candidate factors are the products of increased oxidative stress. Although there is circumstantial evidence of such an increase, direct evidence is still lacking. Electron paramagnetic spin trap resonance (EPR), the most direct method to detect free radicals in tissues, was used to measure superoxide levels in placentae from normal pregnancies (n=13) and pregnancies complicated by PE (n=10). The superoxide level was significantly increased in the placental tissue of pre-eclamptic women. Moreover, upon inhibition of Cu-Zn superoxide dismutase (SOD) activity the relative increase of the superoxide levels was significantly smaller in the placentae from the PE patients, implying decreased basal Cu-Zn SOD activity. These findings lend direct support to the hypothesis that oxidative stress in placental tissue is increased in PE. Placenta (2001), 22, 304–308 2001 Harcourt Publishers Ltd
INTRODUCTION Pre-eclampsia (PE) is a leading cause of maternal mortality and morbidity, fetal growth retardation and iatrogenic preterm birth (National report, 1990). PE is commonly defined as a syndrome consisting of de novo hypertension and proteinuria in the second half of pregnancy. The pathogenesis of PE is not fully unravelled. Early in normal pregnancy trophoblast invasion of the myometrial spiral arteries changes these arteries from a high to a low resistance vascular bed insensitive to vasoactive agents. In early onset PE the trophoblast invasion is defective. As a consequence the high flow resistance and sensitivity to vasoactive agents in the placental vascular bed are preserved, leading to underperfusion. One of the current hypotheses is that the underperfused and thus ischaemic placenta would release one or several yet unidentified factors causing a generalized maternal endothelial dysfunction, manifesting as vasoconstriction (hypertension), platelet aggregation and vascular leakage (Roberts et al., 1989). Increased oxidative stress in the placenta is thought to induce this maternal endothelial dysfunction in PE (Walsh, 1998). Oxidative stress can be defined as an imbalance between the production of free radicals and antioxidant forces. Oxygen derived free radicals or reactive oxygen species (ROS), like the superoxide anion radical O2 (superoxide), are e
To whom correspondance should be addressed at: University Medical Centre Utrecht, Department of Obstetrics KE4.134.2, P.O. Box 85090, 3508 AB Utrecht, The Netherlands. Fax: 31-30-2505320; E-mail: [email protected] 0143–4004/01/040304+05 $35.00/0
continuously produced in vivo by many cell types. Increased production of ROS disturbs cellular function in many ways. Superoxide is converted into hydrogen peroxide by the activity of superoxide dismutase (SOD) that is present in virtually all cell types. In PE there is circumstantial evidence of increased placental production of ROS. First, both the production rate in the placenta and the maternal plasma concentration of lipid peroxides are elevated in PE (Wang et al., 1991; Walsh, Wang and Kay, 1992). Lipid peroxides are formed by the reaction of ROS with polyunsaturated fatty acids in the cellular membranes. Since several antioxidants, like vitamin E, are present in plasma it is assumed that lipid peroxidation does not occur in the maternal circulation. Therefore circulating lipid peroxidation products are thought to originate from the placenta (Walsh, 1994). In the process of entering the maternal circulation these lipid peroxides could affect maternal endothelial cellular membranes, and thus contribute to the maternal endothelial dysfunction. Second, the presence of nitrotyrosine residues in placental tissue is elevated in PE compared to normal pregnancy (Myatt et al., 1996). Nitrotyrosine residues are indicative of the production of peroxynitrite. This very potent free radical is formed by the very rapid reaction of superoxide and NO. Finally, an impaired scavenging of superoxide could contribute to the high levels of superoxide as well. The activity of the main intracellular superoxide scavenger Cu-Zn-SOD has been reported to be about 30 per cent lower in the placenta in PE (Wang and Walsh, 1996; Poranen et al., 1996). 2001 Harcourt Publishers Ltd
Sikkema et al.: Superoxide in Placental Tissue
So, there is a large body of circumstantial evidence that PE is associated with increased placental production of ROS, like superoxide. Nevertheless, direct evidence of increased production of superoxide is still lacking. The difficulty is the unstable and reactive character of superoxide impeding its measurement. Electron paramagnetic spin trap resonance (EPR) is the most direct method to detect free oxygen radicals (Rosen, 1999). In this study we used EPR to study the levels of superoxide in normal and pre-eclamptic placentae.
MATERIALS AND METHODS Subjects Samples of placental tissue were collected from 10 preeclamptic pregnancies and 13 normal control pregnancies. PE was defined according to the criteria of the International Society for the Study of Hypertension in Pregnancy (ISSHP): a diastolic blood pressure d90 mmHg at two occasions at least 4 h apart and proteinuria d300 mg/24 h in the second half of pregnancy in a previously normotensive woman (Davey and McGillavray, 1988). In patients with chronic hypertension PE was diagnozed when proteinuria occurred in the second half of pregnancy. A concomitant HELLP syndrome was defined as a platelet count <100103/mm3, a serum amino-aspartatetransferase level >60 U/l and a lactate dehydrogenase level >600 U/l (Sibai et al., 1986). The fifth Korotkoff tone was used to measure diastolic blood pressure. All patients and controls were delivered by caesarean section. None of the patients and controls had labour before caesarean section. Controls did not suffer from medical conditions, such as diabetes or obesity and none had a history of small for gestational age (SGA) or hypertensive disorders in any previous pregnancy. Controls were normotensive throughout pregnancy and delivered babies with a birthweight >5th centile for the Dutch population (Kloosterman, 1970). A baby was considered SGA if its birthweight was <5th centile.
Sampling It is conceivable that different regions of the same placenta are exposed to different levels of oxidative stress, leading to variable quantities of PBN adducts. Therefore, the within placental variation in yield of PBN adducts was tested by comparing the yield of PBN adducts in different sections of the same placenta of pre-eclamptic women. No significant differences in yield were found, as long as evidently necrotic tissue was avoided. Samples were collected immediately after caesarean section. A macroscopically non-infarcted full-thickness sample of placental tissue was removed from the centre of the placenta, rinsed in saline to remove excess maternal blood and denuded of the fetal membranes. The samples were divided into two sections of 2.5 g each.
305
Electron paramagnetic spin trap resonance From each placenta, the two collected sections of tissue were carefully chopped into small pieces with a diameter ranging from 2 to 4 mm. The first section was pre-incubated for 15 min at 37C with phosphate-buffered saline (PBS), the second section with PBS buffer containing 10 m diethyldithiocarbamate (DETC; Sigma-Aldrich, St Louis, USA). This potent copper chelator efficiently inactivates Cu-Zn SOD. Next, the lipid soluble spin trap N-t-Butyl-aphenylnitrone (PBN; Molecular Probes Europe, Leiden, The Netherlands) was added to a final concentration of 50 m. After incubation for 15 min at 37C the tissue was denaturated by adding methanol. Next the spin adducts formed were extracted by the addition of 1 ml of chloroform. Placental tissue was then removed. The resulting mixture contained methanol, chloroform and pbs buffer at a ratio of 1 : 1 : 2 respectively. After centrifugation at 3500 rpm (2200 g) for 5 min, the chloroform layer containing the lipophilic PBNspin adducts was collected by syringe and stored at 80C. Prior to EPR assessment, the chloroform was air-evaporated and the residue reconstituted with fresh chloroform to a standard volume of 300 l. Spurious quantities of water will reduce EPR sensitivity, and therefore were removed by addition of some anhydrous sodium sulfate. The sample was transferred to a quartz tube (i.d. 2 mm). Artificial EPR linewidth broadening due to dissolved molecular oxygen was avoided by bubbling the dehydrated sample with nitrogen gas. After this preparation, the sample was kept under nitrogen atmosphere, the quartz tube sealed and placed at the centre of a Bruker ST1 rectangular cavity (operating in TE102 mode with unloaded Q=3800). The spectra were recorded on a modified ESP300E spectrometer (Bruker, Rheinstetten, Germany) operating at X-band frequencies. The magnetic field was modulated at a frequency of 100 kHz and an amplitude of 0.4 Gauss. Microwave power and frequency were 20 mW and 9.402 GHz respectively. The detector operated with gain=5.106, time constant=650 ms and ADC conversion time=164 ms. The magnetic field was swept over an interval ranging from 3330 to 3380 Gauss. All EPR experiments were performed at room temperature. A typical EPR spectrum is given in Figure 1. The superoxide adducts can be identified as a triplet of doublets with hyperfine couplings of AN =14.8 Gauss and AH =2.8 Gauss (Thornalley, 1986). Often, the presence of another PBN adduct with a larger proton hyperfine coupling of about 5 Gauss could be observed. The nature of this secondary adduct was not investigated as we attributed it to PBN adducts from carbon-based radicals resulting from lipid peroxidation. Moreover, in samples containing DETC, a weak single line at g=2.01 was observed and attributed to the paramagnetic Cu-(DETC)2 complex formed upon chelation of copper by DETC. In principle this complex has g=2.025 at a temperature of 77 Kelvin with hyperfine structure due to coupling to the copper nucleus. In our spectra at room temperature the third hyperfine line appears at a position corresponding to
Placenta (2001), Vol. 22
306
A
C
B
D
Figure 1. Typical absorption spectra of a control without (A) and with diethyldithiocarbamate (DETC) pre-incubation (C) and a pre-eclamptic patient without (B) and with DETC pre-incubation (D). The arrows mark the positions of the three pairs of absorption peaks of the PBN-superoxide adducts.
g=2.01 due to hyperfine interaction (Suzuki et al., 1997). In our work we disregarded these contaminating EPR signals, and determined the EPR intensity from the superoxide-PBN adducts only [expressed as signal to noise (s/n) ratio].
Statistics The differences in s/n ratio and relative increase in s/n ratio after preincubation with DETC (defined as the s/n ratio with DETC divided by the s/n ratio without DETC), gestational age at delivery and birthweight between the two groups were compared by Student’s t-test for two independent samples. The difference in parity and gravidity between the two groups were compared by the Mann–Whitney U test for two independent samples. A P value <0.05 was considered to indicate statistical significance.
RESULTS In Table 1 the main characteristics of the patients and controls are given. As opposed to the controls the majority of the patients experienced an iatrogenic preterm delivery. Consequently, mean gestational age was shorter and birthweight lower in the patient group. In the control group, six women were parous and in the patients group there were three parous women. Additionally, five PE patients delivered SGA babies. Two of the patients had HELLP syndrome as well. Two others were known with chronic hypertension, and one patient was diagnozed with underlying pheochromocytoma during her pregnancy.
Table 1. Patient characteristics
Maternal age (years)a Gestational age at delivery (days)a Birthweight (g)a Gravidityb Parityb
Controls (n=13)
Patients (n=10)
32.2 (4) 273 (19) 3463 (523) 2 (1–5) 1 (0–3)
30.0 (5) 216 (20) 1316 (617) 1 (1–6) 0 (0–5)
ns P<0.001 P<0.001 ns ns
a Data are expressed as mean (s.d.), b data are expressed as median (range). ns indicates not significant differences.
Figure 1 shows typical absorption spectra of the PBN-spin adducts in a control (A and C) and a patient (B and D). The specific absorption spectrum of the PBN-superoxide adduct consists of three pairs of peaks (marked by the black arrows). In the absorption spectrum of the patient (B) these peaks are much more pronounced than in the spectrum of the control (A). After pre-incubation with DETC the absorption peaks of the superoxide-PBN adducts increase in both the patient (D) and control (C) to the same level. Figure 2 depicts the mean s/n ratios in controls and patients with and without Cu-Zn SOD inhibition by DETC. The mean (s.d.) s/n ratios in patients and controls were 4.6 (1.2) and 3.3 (0.5), respectively (P<0.01). After Cu-Zn SOD inhibition by DETC the s/n ratio increased in both groups. The mean (s.d.) relative increase were 0.63 (0.27) in the controls and 0.35 (0.11) in the patients (P<0.01). The differences were not significantly affected when patients with concomitant pre-existing hypertension or HELLP syndrome were excluded.
S/N ratio
Sikkema et al.: Superoxide in Placental Tissue
10 9 8 7 6 5 4 3 2 1 0
307
*
Controls
PE
Controls (DETC)
PE (DETC)
Figure 2. Signal to noise ratio (s/n ratio) in pre-eclamptic (PE) patients (n=10) and controls (n=13) of samples with PBN (difference significant P<0.01; Student’s t-test for two independent samples) and with PBN after pre-incubation with diethyldithiocarbamate (DETC). * indicates statistically significant differences from controls.
There was no difference in median s/n ratio between nulliparous and parous women in the control group; 3.2 (n=7) versus 3.3 (n=6) respectively. Neither there was a difference in relative increase after inhibition by DETC.
DISCUSSION To our knowledge this is the first study providing direct evidence of increased superoxide levels in the placenta of pre-eclamptic women. We showed that basal Cu-Zn SOD activity is impaired in the placentae of these patients. In the current hypothesis on the pathogenesis of PE, increased oxidative stress generated in the placenta plays a crucial role. Up to now only circumstantial evidence of such an increase had been published. The increases in superoxide levels we found in the placenta in PE could be due to a higher production rate of superoxide or decreased scavenging by one or several SOD isoforms or a combination of both. The strong inhibitor of Cu-Zn SOD DETC was used to stimulate the PBN-superoxide signal. The s/n ratio after pre-incubation with DETC was similar in both groups. It appeared that the relative increase in superoxide was significantly smaller in the PE patients, suggesting decreased Cu-Zn SOD activity. This corroborates previous reports of reduced total SOD activity (Poranen et al., 1996) and Cu-Zn SOD activity (Wang and Walsh, 1996) in placental homogenates of pre-eclamptic women. Since Cu-Zn SOD is inactivated under conditions of high oxidative stress (Salo et al., 1990), the reduction in its activity in the placenta in PE may be a consequence of prolonged exposure to high levels of superoxide. Besides Cu-Zn SOD there are two other SOD isoforms present in the placenta. First, the mitochondrial manganese SOD (MnSOD), which is mainly located in the endothelium (Myatt et al., 1997). MnSOD is not inhibited by DETC (Halliwell and Gutteridge, 1999). Due to the increases in the amount of mitochondria and the aberrant morphology of these mitochondria in the placenta in PE (Jones and Fox, 1980),
MnSOD may play a role in the excess of superoxide in the placenta in PE. Immunohistochemical studies found no difference in expression of MnSOD in the placenta between normontensive and pre-eclamptic pregnancies (Myatt et al., 1997). However, the activity of MnSOD has not been compared between normotensive and pre-eclamptic pregnancies. The other isoform of SOD present in the placenta is the extracellular form of SOD (ecSOD). In one study, there was no difference in both expression and activity of ecSOD between normal and preeclamptic placentae. Moreover, the contribution of ecSOD was very small compared to CuZN-SOD and MnSOD (Boggess et al., 1998). Whether ecSOD is inhibited by DETC is not known, but since ecSOD has copper in its active centre it is very likely that it is inhibited by this strong copper chelator as well. In the placenta in PE there are a few candidate sources for the increases in placental superoxide production. First, in PE the amount of mitochondria is increased in the placenta and the mitochondria have an abnormal appearance as well (Jones and Fox, 1980). This could result in increased leakage of superoxide into the cytosol, since superoxide is continuously produced in the mitochondrial electrontransportchain. Second, endothelial NO-synthase activity may be increased in the placenta in PE (Morris et al., 1995; Brennecke et al., 1997). We have demonstrated earlier that endothelial NO-synthase (ecNOS) is able to generate superoxide besides nitric oxide (Stroes et al., 1998). Third, xanthine oxidase can act as a source of superoxide. Increased activity of xanthine oxidase is present in the placental bed but not in the placenta itself (Many et al., 2000). The fourth possible source is NAD(P)Hoxidase. To our knowledge there are no data available on the expression and activity of these enzymes in placental tissue. Our results have to be interpreted with some caution. Gestational age differed substantially between patients and controls: controls were all full term at delivery and the patients predominantly in their late second or early third trimester. A gestational age dependent rise in SOD activity or decrease in superoxide production, rather than PE itself, would explain the increased superoxide level in the placentae of the patients in this study. Two longitudinal studies on placental SOD demonstrated that SOD activity remains stable from 32 weeks gestation until term (Sekiba and Yosioka 1979; Takehara, Yosioka and Sasaki, 1990). These studies are, however, hampered by the fact that only a few samples were collected between 32 to 36 weeks of gestation and by lack of differentiation between the activity of the various SOD isoforms. In conclusion it is not known, whether superoxide production in the placenta decreases with advancing gestation. There is only limited longitudinal data on the production of superoxide throughout uncomplicated pregnancies. Lipid peroxides are often used as markers of oxidative stress. Unfortunately data on the course of placental lipid peroxidation in the second half of pregnancy are not consistent. One study suggested a decrease towards term, but only very few samples were collected in the second trimester (Sekiba and Yosioka, 1979). In another study, concentrations of lipid peroxides in human
308
placental tissue did not change between 32 weeks gestation and term (Takehara, Yosioka and Sasaki, 1990). In conclusion, our study lends direct support to the current hypothesis that oxidative stress in the placenta is increased in PE. This increase is likely to be attributed to a reduction in Cu-Zn SOD activity.
REFERENCES Brennecke SP, Gude NM, Di Iulio JL & King RG (1997) Reduction of placental nitric oxide synthase activity in pre-eclampsia. Clin Sci, 93, 51–55. Boggess KA, Oury TD, Kay HH & Crapo JD (1998) Extracellular superoxide dismutase localization and activity within the human placenta. Placenta, 1, 417–422. Davey DA & MacGillivray I (1988) The classification and definition of hypertension disorders in pregnancy. Am J Obstet Gynecol, 158, 892–898. Halliwell B & Gutteridge JCM (1999) Free Radicals in Biology and Medicine. New York: Oxford Science Publications. Jones CJP & Fox H (1980) An ultrastructural and ultrahistochemical study of the human placenta in maternal pre-eclampsia. Placenta, 1, 61–76. Kloosterman GJ (1970) On intrauterine growth. Int J Obstet Gynecol, 8, 895–912. Many A, Hubel CA, Fisher SJ, Roberts JM & Zhou Y (2000) Invasive trophoblast manifest evidence of oxidative stress in preeclampsia. Am J Pathol, 156, 321–333. Morris NH, Sooranna SR, Learmont JG, Poston L, Ramsey B, Pearson JD & Steer PJ (1995) Nitric oxide synthase activities in placental tissue from normotensive, pre-eclamptic and growth retarded pregnancies. Br J Obstet Gynaecol, 102, 711–714. Myatt L, Rosenfield RB, Eis ALW, Brockman DE, Greer IA & Lyall F (1996) Nitrotyrosine residues in placenta: evidence of peroxynitrite formation and action. Hypertension, 28, 488–493. Myatt L, Eis ALW, Brockman DE, Kossenjans W, Greer IA & Lyall F (1997) Differential localization of superoxide dismutase isoforms in placental villous tissue of normotensive, pre-eclamptic, and intrauterine growth restricted pregnancies. J Histochem Cytochem, 45, 1433–1438. National High Blood Pressure Education Program Working Group (1990) Report on High Blood Pressure in Pregnancy. Am J Obstet Gynecol, 163, 1689–1712. Poranen AK, Ekblad U, Uotila P & Ahotupa M (1996) Lipid peroxidation and antioxidants in normal and pre-eclamptic pregnancies. Placenta, 17, 401–405.
Placenta (2001), Vol. 22 Roberts JM, Taylor RN, Musci TJ, Rodgers GM, Hubel CA & McLaughlin MK (1989) Preeclampsia: an endothelial cell disorder. Am J Obstet Gynecol, 161, 1200–1204. Rosen GM (Ed.) (1999) Free Radicals: Biology and Detection by Spin Trapping. Oxford: Oxford University Press. Salo DC, Pacifici RE, Lin SW, Giulivi C & Davies KJA (1990) Superoxide dismutase undergoes proteoysis and fragmentation following oxidative modification and inactivation. J Biol Chem, 265, 11919–11927. Sekiba K & Yoshioka T (1979) Changes of lipid peroxidation and superoxide dismutase activity in the human placenta. Am J Obstet Gynecol, 135, 368–371. Sibai BM, Mabie BC, Taslimi MM, Nazar A, Amon E & Ryan GM (1986) Maternal–perinatal outcome associated with the syndrome of hemolysis, elevated liver enzymes, and low platelets in severe preeclampsiaeclampsia. Am J Obstet Gynecol, 155, 501–509. Stroes ESG, van Faassen EE, Londen van GJ & Rabelink TJ (1998) Oxygen radical stress in vascular disease: the role of endothelial nitric oxide synthase. J Cardiov Pharmacol, 32(Suppl 3), S14–S21. Suzuki Y, Fuji S, Tominga T, Yoshimoto T, Yoshimura T & Kamada H (1997) The origin of an EPR signal observed in dithiocarbamate loaded tissue: copper (II)-dithiocarbamate complexes account for the narrow hyperfine lines. Biochem Biophys Acta, 1335, 242–245. Takehara Y, Yosioka T & Sasaki J (1990) Changes in the levels of lipoperoxide and antioxidant factors in human placenta during gestation. Acta Med Okayama, 44, 103–111. Thornalley PJ (1986) Theory and biological applications of the ESR technique of spin trapping. Life Chem Rep, 4, 57–112. Walsh SW, Wang Y & Kay HH (1992) Placental lipid peroxides and thromboxane are increased and prostacyclin is decreased in women with preeclampsia. Am J Obstet Gynecol, 167, 946–949. Walsh SC (1994) Lipid peroxidation in pregnancy. Hypertens Pregnancy, 13, 1–25. Walsh SW (1998) Maternal-placental interactions of oxidative stress and antioxidants in preeclampsia. Sem Reprod Endocr, 16, 93–104. Wang Y, Walsh SW, Guo J & Zhang Y (1991) The imbalance between thromboxane and prostacyclin pre-eclampsia is associated with an imbalance between lipid peroxides and vitamin E. Am J Obstet Gynecol, 165, 1695– 1700. Wang Y & Walsh SW (1996) Antioxidant activities and mRNA expression of superoxide dismutase, catalase, and gluthatione peroxidase in normal and preeclamptic placentas. J Soc Gynecol Investig, 3, 179–184.
Placental Superoxide is Increased in Pre-eclampsia J. M. Sikkemaa,e, B. B. van Rijna, A. Franxa, H. W. Bruinsea, R. de Roosb, E. S. G. Stroesc and E. E. van Faassend a Department of Obstetrics, Gynaecology and Neonatology, b Department of Nephrology, c Department of Vascular Medicine, University Medical Centre Utrecht and d The Debye Institute, Section Interface Physics, University of Utrecht, The Netherlands Paper accepted 5 January 2001
One of the current hypotheses on the pathophysiology of pre-eclampsia (PE) states that the placenta secretes one or more cytotoxic factors resulting in maternal endothelial dysfunction. Among the candidate factors are the products of increased oxidative stress. Although there is circumstantial evidence of such an increase, direct evidence is still lacking. Electron paramagnetic spin trap resonance (EPR), the most direct method to detect free radicals in tissues, was used to measure superoxide levels in placentae from normal pregnancies (n=13) and pregnancies complicated by PE (n=10). The superoxide level was significantly increased in the placental tissue of pre-eclamptic women. Moreover, upon inhibition of Cu-Zn superoxide dismutase (SOD) activity the relative increase of the superoxide levels was significantly smaller in the placentae from the PE patients, implying decreased basal Cu-Zn SOD activity. These findings lend direct support to the hypothesis that oxidative stress in placental tissue is increased in PE. Placenta (2001), 22, 304–308 2001 Harcourt Publishers Ltd
INTRODUCTION Pre-eclampsia (PE) is a leading cause of maternal mortality and morbidity, fetal growth retardation and iatrogenic preterm birth (National report, 1990). PE is commonly defined as a syndrome consisting of de novo hypertension and proteinuria in the second half of pregnancy. The pathogenesis of PE is not fully unravelled. Early in normal pregnancy trophoblast invasion of the myometrial spiral arteries changes these arteries from a high to a low resistance vascular bed insensitive to vasoactive agents. In early onset PE the trophoblast invasion is defective. As a consequence the high flow resistance and sensitivity to vasoactive agents in the placental vascular bed are preserved, leading to underperfusion. One of the current hypotheses is that the underperfused and thus ischaemic placenta would release one or several yet unidentified factors causing a generalized maternal endothelial dysfunction, manifesting as vasoconstriction (hypertension), platelet aggregation and vascular leakage (Roberts et al., 1989). Increased oxidative stress in the placenta is thought to induce this maternal endothelial dysfunction in PE (Walsh, 1998). Oxidative stress can be defined as an imbalance between the production of free radicals and antioxidant forces. Oxygen derived free radicals or reactive oxygen species (ROS), like the superoxide anion radical O2 (superoxide), are e
To whom correspondance should be addressed at: University Medical Centre Utrecht, Department of Obstetrics KE4.134.2, P.O. Box 85090, 3508 AB Utrecht, The Netherlands. Fax: 31-30-2505320; E-mail: [email protected] 0143–4004/01/040304+05 $35.00/0
continuously produced in vivo by many cell types. Increased production of ROS disturbs cellular function in many ways. Superoxide is converted into hydrogen peroxide by the activity of superoxide dismutase (SOD) that is present in virtually all cell types. In PE there is circumstantial evidence of increased placental production of ROS. First, both the production rate in the placenta and the maternal plasma concentration of lipid peroxides are elevated in PE (Wang et al., 1991; Walsh, Wang and Kay, 1992). Lipid peroxides are formed by the reaction of ROS with polyunsaturated fatty acids in the cellular membranes. Since several antioxidants, like vitamin E, are present in plasma it is assumed that lipid peroxidation does not occur in the maternal circulation. Therefore circulating lipid peroxidation products are thought to originate from the placenta (Walsh, 1994). In the process of entering the maternal circulation these lipid peroxides could affect maternal endothelial cellular membranes, and thus contribute to the maternal endothelial dysfunction. Second, the presence of nitrotyrosine residues in placental tissue is elevated in PE compared to normal pregnancy (Myatt et al., 1996). Nitrotyrosine residues are indicative of the production of peroxynitrite. This very potent free radical is formed by the very rapid reaction of superoxide and NO. Finally, an impaired scavenging of superoxide could contribute to the high levels of superoxide as well. The activity of the main intracellular superoxide scavenger Cu-Zn-SOD has been reported to be about 30 per cent lower in the placenta in PE (Wang and Walsh, 1996; Poranen et al., 1996). 2001 Harcourt Publishers Ltd
Sikkema et al.: Superoxide in Placental Tissue
So, there is a large body of circumstantial evidence that PE is associated with increased placental production of ROS, like superoxide. Nevertheless, direct evidence of increased production of superoxide is still lacking. The difficulty is the unstable and reactive character of superoxide impeding its measurement. Electron paramagnetic spin trap resonance (EPR) is the most direct method to detect free oxygen radicals (Rosen, 1999). In this study we used EPR to study the levels of superoxide in normal and pre-eclamptic placentae.
MATERIALS AND METHODS Subjects Samples of placental tissue were collected from 10 preeclamptic pregnancies and 13 normal control pregnancies. PE was defined according to the criteria of the International Society for the Study of Hypertension in Pregnancy (ISSHP): a diastolic blood pressure d90 mmHg at two occasions at least 4 h apart and proteinuria d300 mg/24 h in the second half of pregnancy in a previously normotensive woman (Davey and McGillavray, 1988). In patients with chronic hypertension PE was diagnozed when proteinuria occurred in the second half of pregnancy. A concomitant HELLP syndrome was defined as a platelet count <100103/mm3, a serum amino-aspartatetransferase level >60 U/l and a lactate dehydrogenase level >600 U/l (Sibai et al., 1986). The fifth Korotkoff tone was used to measure diastolic blood pressure. All patients and controls were delivered by caesarean section. None of the patients and controls had labour before caesarean section. Controls did not suffer from medical conditions, such as diabetes or obesity and none had a history of small for gestational age (SGA) or hypertensive disorders in any previous pregnancy. Controls were normotensive throughout pregnancy and delivered babies with a birthweight >5th centile for the Dutch population (Kloosterman, 1970). A baby was considered SGA if its birthweight was <5th centile.
Sampling It is conceivable that different regions of the same placenta are exposed to different levels of oxidative stress, leading to variable quantities of PBN adducts. Therefore, the within placental variation in yield of PBN adducts was tested by comparing the yield of PBN adducts in different sections of the same placenta of pre-eclamptic women. No significant differences in yield were found, as long as evidently necrotic tissue was avoided. Samples were collected immediately after caesarean section. A macroscopically non-infarcted full-thickness sample of placental tissue was removed from the centre of the placenta, rinsed in saline to remove excess maternal blood and denuded of the fetal membranes. The samples were divided into two sections of 2.5 g each.
305
Electron paramagnetic spin trap resonance From each placenta, the two collected sections of tissue were carefully chopped into small pieces with a diameter ranging from 2 to 4 mm. The first section was pre-incubated for 15 min at 37C with phosphate-buffered saline (PBS), the second section with PBS buffer containing 10 m diethyldithiocarbamate (DETC; Sigma-Aldrich, St Louis, USA). This potent copper chelator efficiently inactivates Cu-Zn SOD. Next, the lipid soluble spin trap N-t-Butyl-aphenylnitrone (PBN; Molecular Probes Europe, Leiden, The Netherlands) was added to a final concentration of 50 m. After incubation for 15 min at 37C the tissue was denaturated by adding methanol. Next the spin adducts formed were extracted by the addition of 1 ml of chloroform. Placental tissue was then removed. The resulting mixture contained methanol, chloroform and pbs buffer at a ratio of 1 : 1 : 2 respectively. After centrifugation at 3500 rpm (2200 g) for 5 min, the chloroform layer containing the lipophilic PBNspin adducts was collected by syringe and stored at 80C. Prior to EPR assessment, the chloroform was air-evaporated and the residue reconstituted with fresh chloroform to a standard volume of 300 l. Spurious quantities of water will reduce EPR sensitivity, and therefore were removed by addition of some anhydrous sodium sulfate. The sample was transferred to a quartz tube (i.d. 2 mm). Artificial EPR linewidth broadening due to dissolved molecular oxygen was avoided by bubbling the dehydrated sample with nitrogen gas. After this preparation, the sample was kept under nitrogen atmosphere, the quartz tube sealed and placed at the centre of a Bruker ST1 rectangular cavity (operating in TE102 mode with unloaded Q=3800). The spectra were recorded on a modified ESP300E spectrometer (Bruker, Rheinstetten, Germany) operating at X-band frequencies. The magnetic field was modulated at a frequency of 100 kHz and an amplitude of 0.4 Gauss. Microwave power and frequency were 20 mW and 9.402 GHz respectively. The detector operated with gain=5.106, time constant=650 ms and ADC conversion time=164 ms. The magnetic field was swept over an interval ranging from 3330 to 3380 Gauss. All EPR experiments were performed at room temperature. A typical EPR spectrum is given in Figure 1. The superoxide adducts can be identified as a triplet of doublets with hyperfine couplings of AN =14.8 Gauss and AH =2.8 Gauss (Thornalley, 1986). Often, the presence of another PBN adduct with a larger proton hyperfine coupling of about 5 Gauss could be observed. The nature of this secondary adduct was not investigated as we attributed it to PBN adducts from carbon-based radicals resulting from lipid peroxidation. Moreover, in samples containing DETC, a weak single line at g=2.01 was observed and attributed to the paramagnetic Cu-(DETC)2 complex formed upon chelation of copper by DETC. In principle this complex has g=2.025 at a temperature of 77 Kelvin with hyperfine structure due to coupling to the copper nucleus. In our spectra at room temperature the third hyperfine line appears at a position corresponding to
Placenta (2001), Vol. 22
306
A
C
B
D
Figure 1. Typical absorption spectra of a control without (A) and with diethyldithiocarbamate (DETC) pre-incubation (C) and a pre-eclamptic patient without (B) and with DETC pre-incubation (D). The arrows mark the positions of the three pairs of absorption peaks of the PBN-superoxide adducts.
g=2.01 due to hyperfine interaction (Suzuki et al., 1997). In our work we disregarded these contaminating EPR signals, and determined the EPR intensity from the superoxide-PBN adducts only [expressed as signal to noise (s/n) ratio].
Statistics The differences in s/n ratio and relative increase in s/n ratio after preincubation with DETC (defined as the s/n ratio with DETC divided by the s/n ratio without DETC), gestational age at delivery and birthweight between the two groups were compared by Student’s t-test for two independent samples. The difference in parity and gravidity between the two groups were compared by the Mann–Whitney U test for two independent samples. A P value <0.05 was considered to indicate statistical significance.
RESULTS In Table 1 the main characteristics of the patients and controls are given. As opposed to the controls the majority of the patients experienced an iatrogenic preterm delivery. Consequently, mean gestational age was shorter and birthweight lower in the patient group. In the control group, six women were parous and in the patients group there were three parous women. Additionally, five PE patients delivered SGA babies. Two of the patients had HELLP syndrome as well. Two others were known with chronic hypertension, and one patient was diagnozed with underlying pheochromocytoma during her pregnancy.
Table 1. Patient characteristics
Maternal age (years)a Gestational age at delivery (days)a Birthweight (g)a Gravidityb Parityb
Controls (n=13)
Patients (n=10)
32.2 (4) 273 (19) 3463 (523) 2 (1–5) 1 (0–3)
30.0 (5) 216 (20) 1316 (617) 1 (1–6) 0 (0–5)
ns P<0.001 P<0.001 ns ns
a Data are expressed as mean (s.d.), b data are expressed as median (range). ns indicates not significant differences.
Figure 1 shows typical absorption spectra of the PBN-spin adducts in a control (A and C) and a patient (B and D). The specific absorption spectrum of the PBN-superoxide adduct consists of three pairs of peaks (marked by the black arrows). In the absorption spectrum of the patient (B) these peaks are much more pronounced than in the spectrum of the control (A). After pre-incubation with DETC the absorption peaks of the superoxide-PBN adducts increase in both the patient (D) and control (C) to the same level. Figure 2 depicts the mean s/n ratios in controls and patients with and without Cu-Zn SOD inhibition by DETC. The mean (s.d.) s/n ratios in patients and controls were 4.6 (1.2) and 3.3 (0.5), respectively (P<0.01). After Cu-Zn SOD inhibition by DETC the s/n ratio increased in both groups. The mean (s.d.) relative increase were 0.63 (0.27) in the controls and 0.35 (0.11) in the patients (P<0.01). The differences were not significantly affected when patients with concomitant pre-existing hypertension or HELLP syndrome were excluded.
S/N ratio
Sikkema et al.: Superoxide in Placental Tissue
10 9 8 7 6 5 4 3 2 1 0
307
*
Controls
PE
Controls (DETC)
PE (DETC)
Figure 2. Signal to noise ratio (s/n ratio) in pre-eclamptic (PE) patients (n=10) and controls (n=13) of samples with PBN (difference significant P<0.01; Student’s t-test for two independent samples) and with PBN after pre-incubation with diethyldithiocarbamate (DETC). * indicates statistically significant differences from controls.
There was no difference in median s/n ratio between nulliparous and parous women in the control group; 3.2 (n=7) versus 3.3 (n=6) respectively. Neither there was a difference in relative increase after inhibition by DETC.
DISCUSSION To our knowledge this is the first study providing direct evidence of increased superoxide levels in the placenta of pre-eclamptic women. We showed that basal Cu-Zn SOD activity is impaired in the placentae of these patients. In the current hypothesis on the pathogenesis of PE, increased oxidative stress generated in the placenta plays a crucial role. Up to now only circumstantial evidence of such an increase had been published. The increases in superoxide levels we found in the placenta in PE could be due to a higher production rate of superoxide or decreased scavenging by one or several SOD isoforms or a combination of both. The strong inhibitor of Cu-Zn SOD DETC was used to stimulate the PBN-superoxide signal. The s/n ratio after pre-incubation with DETC was similar in both groups. It appeared that the relative increase in superoxide was significantly smaller in the PE patients, suggesting decreased Cu-Zn SOD activity. This corroborates previous reports of reduced total SOD activity (Poranen et al., 1996) and Cu-Zn SOD activity (Wang and Walsh, 1996) in placental homogenates of pre-eclamptic women. Since Cu-Zn SOD is inactivated under conditions of high oxidative stress (Salo et al., 1990), the reduction in its activity in the placenta in PE may be a consequence of prolonged exposure to high levels of superoxide. Besides Cu-Zn SOD there are two other SOD isoforms present in the placenta. First, the mitochondrial manganese SOD (MnSOD), which is mainly located in the endothelium (Myatt et al., 1997). MnSOD is not inhibited by DETC (Halliwell and Gutteridge, 1999). Due to the increases in the amount of mitochondria and the aberrant morphology of these mitochondria in the placenta in PE (Jones and Fox, 1980),
MnSOD may play a role in the excess of superoxide in the placenta in PE. Immunohistochemical studies found no difference in expression of MnSOD in the placenta between normontensive and pre-eclamptic pregnancies (Myatt et al., 1997). However, the activity of MnSOD has not been compared between normotensive and pre-eclamptic pregnancies. The other isoform of SOD present in the placenta is the extracellular form of SOD (ecSOD). In one study, there was no difference in both expression and activity of ecSOD between normal and preeclamptic placentae. Moreover, the contribution of ecSOD was very small compared to CuZN-SOD and MnSOD (Boggess et al., 1998). Whether ecSOD is inhibited by DETC is not known, but since ecSOD has copper in its active centre it is very likely that it is inhibited by this strong copper chelator as well. In the placenta in PE there are a few candidate sources for the increases in placental superoxide production. First, in PE the amount of mitochondria is increased in the placenta and the mitochondria have an abnormal appearance as well (Jones and Fox, 1980). This could result in increased leakage of superoxide into the cytosol, since superoxide is continuously produced in the mitochondrial electrontransportchain. Second, endothelial NO-synthase activity may be increased in the placenta in PE (Morris et al., 1995; Brennecke et al., 1997). We have demonstrated earlier that endothelial NO-synthase (ecNOS) is able to generate superoxide besides nitric oxide (Stroes et al., 1998). Third, xanthine oxidase can act as a source of superoxide. Increased activity of xanthine oxidase is present in the placental bed but not in the placenta itself (Many et al., 2000). The fourth possible source is NAD(P)Hoxidase. To our knowledge there are no data available on the expression and activity of these enzymes in placental tissue. Our results have to be interpreted with some caution. Gestational age differed substantially between patients and controls: controls were all full term at delivery and the patients predominantly in their late second or early third trimester. A gestational age dependent rise in SOD activity or decrease in superoxide production, rather than PE itself, would explain the increased superoxide level in the placentae of the patients in this study. Two longitudinal studies on placental SOD demonstrated that SOD activity remains stable from 32 weeks gestation until term (Sekiba and Yosioka 1979; Takehara, Yosioka and Sasaki, 1990). These studies are, however, hampered by the fact that only a few samples were collected between 32 to 36 weeks of gestation and by lack of differentiation between the activity of the various SOD isoforms. In conclusion it is not known, whether superoxide production in the placenta decreases with advancing gestation. There is only limited longitudinal data on the production of superoxide throughout uncomplicated pregnancies. Lipid peroxides are often used as markers of oxidative stress. Unfortunately data on the course of placental lipid peroxidation in the second half of pregnancy are not consistent. One study suggested a decrease towards term, but only very few samples were collected in the second trimester (Sekiba and Yosioka, 1979). In another study, concentrations of lipid peroxides in human
308
placental tissue did not change between 32 weeks gestation and term (Takehara, Yosioka and Sasaki, 1990). In conclusion, our study lends direct support to the current hypothesis that oxidative stress in the placenta is increased in PE. This increase is likely to be attributed to a reduction in Cu-Zn SOD activity.
REFERENCES Brennecke SP, Gude NM, Di Iulio JL & King RG (1997) Reduction of placental nitric oxide synthase activity in pre-eclampsia. Clin Sci, 93, 51–55. Boggess KA, Oury TD, Kay HH & Crapo JD (1998) Extracellular superoxide dismutase localization and activity within the human placenta. Placenta, 1, 417–422. Davey DA & MacGillivray I (1988) The classification and definition of hypertension disorders in pregnancy. Am J Obstet Gynecol, 158, 892–898. Halliwell B & Gutteridge JCM (1999) Free Radicals in Biology and Medicine. New York: Oxford Science Publications. Jones CJP & Fox H (1980) An ultrastructural and ultrahistochemical study of the human placenta in maternal pre-eclampsia. Placenta, 1, 61–76. Kloosterman GJ (1970) On intrauterine growth. Int J Obstet Gynecol, 8, 895–912. Many A, Hubel CA, Fisher SJ, Roberts JM & Zhou Y (2000) Invasive trophoblast manifest evidence of oxidative stress in preeclampsia. Am J Pathol, 156, 321–333. Morris NH, Sooranna SR, Learmont JG, Poston L, Ramsey B, Pearson JD & Steer PJ (1995) Nitric oxide synthase activities in placental tissue from normotensive, pre-eclamptic and growth retarded pregnancies. Br J Obstet Gynaecol, 102, 711–714. Myatt L, Rosenfield RB, Eis ALW, Brockman DE, Greer IA & Lyall F (1996) Nitrotyrosine residues in placenta: evidence of peroxynitrite formation and action. Hypertension, 28, 488–493. Myatt L, Eis ALW, Brockman DE, Kossenjans W, Greer IA & Lyall F (1997) Differential localization of superoxide dismutase isoforms in placental villous tissue of normotensive, pre-eclamptic, and intrauterine growth restricted pregnancies. J Histochem Cytochem, 45, 1433–1438. National High Blood Pressure Education Program Working Group (1990) Report on High Blood Pressure in Pregnancy. Am J Obstet Gynecol, 163, 1689–1712. Poranen AK, Ekblad U, Uotila P & Ahotupa M (1996) Lipid peroxidation and antioxidants in normal and pre-eclamptic pregnancies. Placenta, 17, 401–405.
Placenta (2001), Vol. 22 Roberts JM, Taylor RN, Musci TJ, Rodgers GM, Hubel CA & McLaughlin MK (1989) Preeclampsia: an endothelial cell disorder. Am J Obstet Gynecol, 161, 1200–1204. Rosen GM (Ed.) (1999) Free Radicals: Biology and Detection by Spin Trapping. Oxford: Oxford University Press. Salo DC, Pacifici RE, Lin SW, Giulivi C & Davies KJA (1990) Superoxide dismutase undergoes proteoysis and fragmentation following oxidative modification and inactivation. J Biol Chem, 265, 11919–11927. Sekiba K & Yoshioka T (1979) Changes of lipid peroxidation and superoxide dismutase activity in the human placenta. Am J Obstet Gynecol, 135, 368–371. Sibai BM, Mabie BC, Taslimi MM, Nazar A, Amon E & Ryan GM (1986) Maternal–perinatal outcome associated with the syndrome of hemolysis, elevated liver enzymes, and low platelets in severe preeclampsiaeclampsia. Am J Obstet Gynecol, 155, 501–509. Stroes ESG, van Faassen EE, Londen van GJ & Rabelink TJ (1998) Oxygen radical stress in vascular disease: the role of endothelial nitric oxide synthase. J Cardiov Pharmacol, 32(Suppl 3), S14–S21. Suzuki Y, Fuji S, Tominga T, Yoshimoto T, Yoshimura T & Kamada H (1997) The origin of an EPR signal observed in dithiocarbamate loaded tissue: copper (II)-dithiocarbamate complexes account for the narrow hyperfine lines. Biochem Biophys Acta, 1335, 242–245. Takehara Y, Yosioka T & Sasaki J (1990) Changes in the levels of lipoperoxide and antioxidant factors in human placenta during gestation. Acta Med Okayama, 44, 103–111. Thornalley PJ (1986) Theory and biological applications of the ESR technique of spin trapping. Life Chem Rep, 4, 57–112. Walsh SW, Wang Y & Kay HH (1992) Placental lipid peroxides and thromboxane are increased and prostacyclin is decreased in women with preeclampsia. Am J Obstet Gynecol, 167, 946–949. Walsh SC (1994) Lipid peroxidation in pregnancy. Hypertens Pregnancy, 13, 1–25. Walsh SW (1998) Maternal-placental interactions of oxidative stress and antioxidants in preeclampsia. Sem Reprod Endocr, 16, 93–104. Wang Y, Walsh SW, Guo J & Zhang Y (1991) The imbalance between thromboxane and prostacyclin pre-eclampsia is associated with an imbalance between lipid peroxides and vitamin E. Am J Obstet Gynecol, 165, 1695– 1700. Wang Y & Walsh SW (1996) Antioxidant activities and mRNA expression of superoxide dismutase, catalase, and gluthatione peroxidase in normal and preeclamptic placentas. J Soc Gynecol Investig, 3, 179–184.