- Email: [email protected]
Immunohistochemical localization of throm☐ane synthase in human intrauterine tissues
Placenta (1994), 15,389-398
Immunohistochemical Localization of Thromboxane Synthase in Human Intrauterine Tissues B. WETZKA”, W. SCHiiFER’, F. KOMMOSS”, H. BETTENDORF”, R. NihINGb, M. BRECKWOLDT” & H. P. ZAHRADNIKaTC a Universit&fiauenkh’nik II, Ettdokrinologieund Repmduktionsmedizin, HugstetterStr. 55, D-79106 Freiburg, Gevnany ’ UniversitiitKonstanz, Fakultiitfur Biologic, P.O. Box 5560, D78464 Konstanz, Germany ‘ To whom cowespondenceshould be addressed Paper accepted3.12.1993
SUMMARY Uterine tissues are known to be able to synthesize thromboxaneAs mz), but there is little information about the nature of cells actually responsiblefor its production. In this study human placenta, fetal membranes, umbilical cord and pregnant myometrium were investigatedimmunohistochemically.The avidin-biotin methodfor a monoclonalantibodyagainst human thromboxanesynthase (Tii 300) was applied on frozen tissue sections.In placenta, fetal membranes and umbilical cord, staining waspositivefor Hojbauer cellsandjbroblasts. Further, in sectionsof placenta, capillaryendotheliumshowed antigenityfor IXsynthase. Leiomyocytesin the umbilical cord vessels contained the enzyme as well. Preparations of pregnant myometrium were shown to express 7X synthase in leiomyocytes,endothelial cells and connectivetissue cells.Amnion, trophoblastand decidua did not possessantigenityfor this enzyme. Since 7’222 plays an important role for the regulation of vascular tone and aggregationof plateletsand may stimulate myometrialcontractionsduring parturition, the abundance of ‘Ix synthase in pregnancy-specifictissuesconfirmsprevious in vivo and in vitro observations.Further, 7X42 synthesizedby Hofbauer cells may be involved in immunologicalreactionsduring pregnancy, and the number and level of activation of Hofbauer cells may be closely related to the initiation of labour. Thmmboxane production by the endothelium lining the f&al vesselspoints to its reguhztoryrolefor the bloodflow in thefetoplacentalunit.
0143-4004/94/040389
+ 10 $08.00/O
0 1994 W. B. Saunders Company Ltd
390
Placenta(1994), Vol. 15
INTRODUCTION The arachidonic acid (AA) metabolite thromboxane A2 (TXAz) has been known for several years as a potent vasoconstrictor and an activator of platelet aggregation (Hamberg, Svensson and Samuelsson, 1975). It is the main eicosanoid produced in thrombocytes, but was shown to be synthesized in many other cell types in vitro. Recently a role for this prostanoid in tumour proliferation and metastasation (Nigam and Zakrzewicz, 1990) and immunological reactions (Rola-Pleszczynski et al, 1985) was proposed. The cellular effects of TXAz are receptor-mediated with an activation of phospholipase C catalysing the release of inositol-triphosphate and diacylglycerol as second messengers. During pregnancy, a significant increase in plasma levels (Mitchell et al, 1978) and urinary excretion (Noort, de Zwart and Keirse, 1987) of the more stable thromboxane metabolite TXBz could be observed. The elevated levels in urine were even more impressive after labour. The synthesizing capacity for TXAz of human tissues derived from pregnant uteri could be shown in vitro with different culture systems (Harper, Khodr and Valenzuela, 1983; M&ill, Viinikka and Ylikorkala, 1984; Walsh, 1985). In short-term tissue culture experiments conducted in our laboratory, TXAz was the main cyclooxygenase metabolite of AA produced by placenta and fetal membranes (Wetzka et al, 1993a, b). Further, special interest was focused on the TXAz production in tissues derived from women suffering from pregnancy-induced hypertension (PIH). The ratio of local prostacyclin/thromboxane level is supposed to be an important factor for the regulation of uterine blood flow (M&ilH et al, 1986). Several studies support the involvement of a PGIz/TXAz imbalance in the pathogenesis of clinical symptoms of PIH (reviewed in Zahradnik et al, 1991). However, there is little data about the identity and localization of the cells within the different tissue compartments for the pregnant uterus that are actually responsible for the production of TXAz (Niising, Lesch and Ullrich, 1990; Swanson et al, 1992). This study was carried out in order to elucidate the cellular distribution of the TX synthase in human placenta, fetal membranes, umbilical cord and myometrium via immunohistochemistry using a monoclonal antibody (mAb) against human TX synthase (Niising, Wernet and Ulhich, 1990).
MATERIALS AND METHODS Immunohistochemistry was carried out using the avidin-biotin method for mAbs. Vectastain Elite kit and peroxidase substrate 3-amino-9-ethylcarbazole (AEC) were purchased from Vector Laboratories, Burlingham, USA. As specific antibody Tii 300, a mAb against human platelet thromboxane synthase-characterized in Niising, Wemet and Ulhich (1990)-was applied. Tissue specimens were collected immediately after delivery, either vaginal or by caesarean section. ‘Labour’ was defined as the presence of regular myometrial contractions and the beginning of cervical dilatation. Four uncomplicated pregnancies went to term. Two of them delivered vaginally. Two underwent caesarean section, one before the onset of labour because of breech presentation, the other after the onset of labour because of fetal distress. In addition, two women suffering from PIH-defined as blood pressure over 140/ 90mmHg in the second half of pregnancy in women who were normotensive before pregnancy (for details see Zahradnik et al, 1991)- were delivered by caesarean section
391
Wet& et al: Thmmboxane Synthase in Intrauterine Tissues
after the onset of labour in weeks 36 and 38, respectively. All six babies had an uncomplicated perinatal period. Placenta, fetal membranes and myometrium were directly snap-frozen in liquid nitrogen. Pieces of umbilical cord were rinsed prior to freezing in Tyrode’s solution for removing clotted blood in the fetal vessels. The frozen sections (3-5 km) were mounted onto microscopic slides and allowed to dry at room temperature overnight. The structural integrity of the tissue specimens was established by control staining with haematoxylin and eosin. The slides for immunohistochemical procedure were fixed in ice-cold acetone for 1Omin. After drying, the sections were rinsed with 0.01 M phosphate-buffered saline (PBS; pH 7.2) and incubated with 1% (v/v) hydrogen peroxide for 10 min in order to inhibit endogenous peroxidases. After intensive washing with 0.5% (v/v) triton X-100 (Sigma, Deisenhofen, Germany) and PBS, preincubation was performed with horse serum (30 min). Then the slides were incubated with the specific antibody Tti 300 in a dilution of 1:200 in PBS containing 0.1% (v/v) bovine serum albumin for 30min. After rinsing three times in PBS, incubation with the second antibody, a biotinylated horse-anti-mouse antibody, was carried out for 30min. Following removal of surplus of antibody in PBS, the slides were incubated with avidin-peroxidase complex for 45 min at 37 “C. Then the bound peroxidase cleaved AEC at 37 “C for 30-45 min in the dark leading to a red-brown colour. Excess substrate was removed by intensive rinsing with distilled water and counterstaining performed with Mayer’s hemalum solution (Merck, Darmstadt, Germany). Finally the tissue sections were mounted under glass coverslips with Kaiser’s glycerol gelatin (Merck, Darmstadt, Germany). Negative controls for each tissue were performed in the same way as described above. While incubating the positive controls with the specific antibody Tti 300 following preabsorption with horse serum, the negative controls were incubated with plain PBS. Evaluation was performed with conventional light microscopy by Dr Komrnoss and Dr Wetzka. Scoring of staining intensity was done in a relative manner (Table 1). ‘+++’
Table I. Immunohistochemical localization of TX synthase in human uterine tissues at term of pregnancy with monoclonal antibody Tii 300. Staining intensity (+/++/ + + +) indicates positive staining for TX synthase, negative staining is shown as ‘-’ Tissue compartment
Cell types
Staining
Placental villi
Hotbauer cells Fibroblasts of villous core Endothelial cells Syncytiotrophoblast Trophoblast Hofbauer cells Fibroblasts Amnion epithelium Trophoblast Decidua Hotbauer cells Fibroblasts Leiomyoqtes of vessels Endothelial cells Fibroblasts/histiocytes Endothelial cells Leiomyocytes
+++ +++ ++/+++ -
Fetal membranes
Umbilical cord
Myometrium
Intensity:
+ weak, + + medium,
+ + + strong, - negative.
+++ +++ +++ +++ +/++ +++ ++/+-I-+ +/++
intensity
392
Placenta (I 994), Vol. 1.5
(strong) was given for maximum staining observed in our preparations, ‘+’ (weak) for minimum staining, respectively. ‘++’ (medium) scored cells were stained in between maximum and minimum. Identification of cell types was done with the help of Prof. Dr.med. N. BGhm, Chairman of the Department of Pediatric Pathology, University of Freiburg, Germany. Amniotic, trophoblastic, decidual and endothelial cells, leiomyocytes of vessels and myometrial preparations were identified according to their specific localization and presentation within the tissues. Cells were called ‘Hofbauer cells’ when they showed the following microscopic criteria: (1) typical localization in the core of placental villi, in the fibroblast and spongy layer of amnion and the reticular layer of chorion and in the Wharton’s jelly of the umbilical cord; (2) appearance with round, fusiform or stellate shape, eccentric nucleus, various granules or vacuoles (Boume, 1962). In some cases, Hofbauer cells could not be differentiated from fibroblasts because of their morphological resemblance.
RESULTS In all uterine tissues staining was positive for cells of the mononuclear phagocytic system and connective tissue cells. There were no major differences between tissue specimens obtained after vaginal deliveries or caesarean sections and from uncomplicated pregnancies or from patients with PIH, respectively (results not shown). The results for the individual tissue compartments together with the staining intensities of antigen-positive cells are shown in Table 1. In placenta, Hofoauer cells and fibroblasts lying in the villous core were stained strongly. There was no staining of the cytotrophoblast and the syncytiotrophoblast layer [Figure l(A)]. Decidua adjacent to placental villi was negative for the TX synthase [Figure l(B)]. Endothelial cells of some capillaries in the villous core [Figure l(A)] and of a few bigger fetal vessels in stem villi [Figure 1(C)] showed antigen@ for TX synthase. Fetal membranes were positive for Hofbauer cells and fibroblasts situated in the spongy layer of amnion and the reticular layer of chorion, whereas the amnion epithelium remained unstained [Figure 2(A)]. D eci‘dua adhering to the chorion layer exhibited only unspecific staining. Sections of the umbilical cord showed stained Hofbauer cells and fibroblasts in the Wharton’s jelly [Figure 3(A)]. Further, leiomyocytes of the vessel walls were positive for TX synthase [Figure 3(B)]. In the umbilical cord specimens endothelium did not exhibit positive staining. In myometrium preparations, leiomyocytes, connective tissue cells and endothelial cells showed positive antigen@ for TX synthase [Figure 4(A) and (B)]. Negative controls for either tissue are shown in Figures l(D), 2(B), 3(C) and 4(C).
DISCUSSION In accordance with Niising et al (1992), cells belonging to the mononuclear phagocytic system were strongly positive for TX synthase. Hotbauer cells lying in the core of placental villi, in the connective tissue layer between amnion and trophoblast in the fetal membranes, and in the Wharton’s jelly of the umbilical cord showed antigenity for the TX42 synthesizing enzyme. This finding agrees with the observations of TXBa being the major cyclooxy-
Wetzka et al: Thromboxane Synthase in Intrautm’ne Tissues
Fipre
1. Immunohistochemical localization of TX synthase in placenta with monoclonal antibody Tii 300. A, placental villi; B, decidua adjacent to placental villi; C, stem villus; D, negative control. The magnification bar represents 50 pm in A, 80 pm in C and 125 urn in B and D.
Figure 2. Fetal membranes. A, cross section of fetal membrane layers; B, negative control. The magnification bar
represents 125 pm. A, amnion; C, chorion; D, decidua; E, endothelium; F, fibroblast; H, Hotbauer cells; S, syncytiotrophoblast; V, villi of placenta.
394
Placenta(1994), Vol. 15
genase product formed from AA by human mononuclear cells (Moley et al, 1979) and of monocytes having the second highest content of TX synthase after platelets @&sing and Ulhich, 1990). The physiological significance of macrophages within the fetomatemal unit represents a new aspect for the immunology in pregnancy (Bulmer and Johnson, 1984). They seem to play a pivotal role concerning immune interactions between the mother and the fetoplacental graft (Hunt, 1989). Holbauer cells, from fetal origin, appear as early as the 4th week of pregnancy during the process of vasculogenesis in the early placental villi (Demir et al, 1989). They possess antigen-presenting abilities, produce a variety of cytokines and exhibit phagocygotic activity (Castelluci and Kaufmann, 1990). The expression of TX’ synthase in this cell population might play a role in regulation of cell proliferation (Nigam and Zakrzewicz, 1990) and activity of immunocompetent cells (Rola-Pleszczynski et al, 1985). An activation of macrophages by an inflammation during pregnancy might stimulate their TXAa synthesis and start myometrial contractions resulting in pre-term labour. The action of TX synthase on prostaglandin endoperoxides leads to the production of TXA2 and 12-hydroxyheptadecatrienoic acid (HHT) and malondialdehyde in a 1:l ratio (Haurand and Ulhich, 1985). 12-HHT, as well as other hydroxy AA products, can act as a chemotactic stimulant for polymorphonuclear granulocytes (Goetzl and Gorman, 1978). Pregnancy complications like PIH or HELLP syndrome (Haemolysis-elevated liver enzymes-low platelet count) are most probably caused by pathological immune reactions. So a changed TXAa synthesis by uterine macrophages may be expected to play an important role for the development of clinical symptoms. However, in this study there was no major difference in TX synthase expression between tissues from uncomplicated pregnancies compared with PIH pregnancies. Since the pathological processes of PIH start during placentation and effect placental circulation throughout pregnancy with the consequence of placental infarctions, it should be considered that at term of pregnancy most tissue changes are ‘burned-out’. Therefore an immunohistochemical examination of intrauterine tissues from the first and second trimesters of pregnancy could be expected to show interesting results concerning the question of pathogenesis of PIH. In addition, we found positive staining of fibroblasts neighbouring Hofbauer cells in all tissues studied. The predominance of TX synthase in cells derived from the mesoderm rather than in ectodermal cells corresponds to the findings of Bryant-Greenwood, Rees and Tumbull (1990) who described antigen@ for prostaglandin synthase more frequently in the connective tissue part of fetal membranes than in the trophoblast and decidua. These observations point to an important role of connective tissue cells (including tissue macrophages) for intrauterine prostaglandin synthesis. Further, leiomyocytes of vessel walls and pregnant myometrium and endothelial cells were stained specifically for TX synthase. A previous immunohistochemical study of pregnant myometrium conducted with a polyclonal antibody against porcine lung TX synthase showed the expression of the enzyme in myometrial smooth muscle and walls of myometrial blood vessels (Swanson et al, 1992). The authors did not distinguish between the muscle layer and the endothelium of vessel wall, therefore it is not clear whether they found endothelial cells showing antigenity for TX synthase. The results are comparable with our study except for two topics. Swanson et al (1992) described a difference in staining intensity between myometrium obtained before and after labour which could not be found in the present study. This can be explained in part by the use of a different antibody and probably by the smaller number of tissue samples investigated in our study. The negative staining of macrophages (and endothelial cells?) in the Swanson study may result from the use of a
Wet.2 :kn et al: Thmnboxane Synthase in Intrauterine Tis!WCS
Figure 3. Immunohistochemical localization of TX synthase in umbilical cord with mAb Tii 300. A, Wharton’s jelly; B, umbilical artery; C, negative control. The magnification bar represents 50 km in A and 80 pm in B and C. Figure 4. Pregnant myometrium. A, overall view; B, myometrial represents 200 km in A and 50 pm in B and C. E, endothelium; leiomyocytes.
vessel; C, negative control. The magnification bar F, fibroblast; H, Hofbauer cells; Hi, histiocyte; L,
396
Placenta (1994), Vol. 15
different antibody and technique of tissue preparation (paraffin-embedded tissues). The polyclonal antibody against porcine lung TX synthase used by Swanson et al (1992) is known to cross-react with TX synthases of most mammalian species. Since the size of these enzymes deriving from different species can vary, as Swanson et al (1992) showed by immunoblot-TX synthase of bovine lung had 65 kD and TX synthase of human uterus only 55 kD-it can be expected that the staining in immunohistochemistry might be slightly different and might not be as specific as if an antibody against human TX synthase is applied. Furthermore, Niising et al (1992) observed that mAb Tii 300 reacted strongly with TX synthase of human macrophages. The positive staining of endothelial cells for TX synthase in our study is not easy to justi@ with the hypothesis that activated thrombocytes stimulate the endothelial PGIz production by transfer of prostaglandin endoperoxides (Moncada and Vane, 1979). In this model of antagonistic PGHZ metabolism by platelets and endothelium, an expression of TX synthase in endothelial cells seems contradictory. However, human umbilical endothelial cells could be shown to synthesize TXA2 (Griesmaher et al, 1989). Therefore it must be assumed that endothelium expresses TX synthase under certain circumstances, i.e. during pregnancy. In all studies investigating the TX synthase in human tissues (Niising, Lesch and Ullrich, 1990; Niising et al, 1992; Swanson et al, 1992) endothelial cells stained positively for this enzyme in the placenta, the umbilical cord and the myometrium. The observation of elevated levels of TXBZ in plasma and urine during pregnancy (Mitchell et al, 1978; Noort, de Zwart and Keirse, 1987) and of TXAZ being the major cyclooxygenase metabolite of AA in tissue cultures of placenta and fetal membranes (Wetzka et al, 1993a, b) together with the findings described above point to the importance of uterine TXA2 for local regulation of blood flow and haemostasis. Additionally its contracting potency for myometrium (Wilhelmsson, W&land and Wiqvist, 1981) may control myometrial activity during labour. Our finding of negative staining for TX synthase in trophoblast cells seems to contradict the observation of Nelson and Walsh (1989a, 1989b). They separated placental villi into the trophoblast layer and the villous core and found nearly the same TXA2 production in both compartments, whereas intact villi produced only 10% TXA2 compared with the separated villous parts. However, they described vimentin-positive cells in the trophoblast preparation which could be fibroblasts and endothelial cells adhering to the trophoblast layer and could be responsible for the TXA2 production. Further, aspirin affects TXA;! production only in whole villus and villous core, but not in the trophoblast layer (Nelson and Walsh, 1989b). This observation points to a modulatory role of trophoblast cells on the TX synthesis of villous core cells rather than on their own TXAZ production. Therefore the trophoblast compartment might be responsible for maintaining a physiological balance between PGIz and TXA2 production in the placenta. In conclusion this study gives new aspects concerning the localization of TX42 production within the different tissue compartments of the pregnant human uterus. It points to an important role of macrophages in the fetoplacental unit for prostaglandin production. TXA2 synthesis in endothelium and leiomyocytes may be responsible for the control of intrauterine blood flow. Further immunohistochemical studies with the mAb against human TX synthase will be conducted to investigate the number of macrophages. These studies will use a double staining technique with an additional antibody against a monocyte and will compare tissues derived from women in labour/not in labour/pre-term labour/ PIH in order to elucidate the role of TXAZ in pregnancy and labour.
397
Wetzka et al: Thromboxane Synthase in Intrauterine Tissues
ACKNOWLEDGEMENTS We gratefully acknowledge Prof. Dr.med. in interpretation of the tissue preparations.
N. Biihm, Pathologisches
Institut der Universitiit
Freiburg
for his help
REFERENCES Boume, G. (1962) Holbauer cells. In The Human Amnion and Chorion (Ed.) Boume, G. pp. 95-112. London: Lloyd-Luke. Bryant-Greenwood, G. D., Rees, M. C. P. & Tumbull, A. C. (1990) Immunohistochemical localization of relaxin, prolactin and prostaglandin synthase in human amnion, chorion and decidua. Journal ofEn&rinology, 114,491-496. Bulmer, J. N. & Johnson, P. M. (1984) Macrophage populations in the human placenta and amniochorion. Clinical and Experimental Immunology, 57, 393-403. Castelluci, M. & Kaufmann, P. (1990) Hofbauer cells. In Pathology of the Human Placenta 2nd edn. (Ed.) Benirschke, K. & Kaufmann, P. pp. 71-80. New York: Springer. Demir, R., Kaufmann, P., Castelluci, M., Erbengi, T. & Kotowski, A. (1989) Fetal vasculogenesis and angiogenesis in human placental villi. Acta Anatomica (Base& 136, 190-203. Goetzl, E. J. & Gornmn, R. R. (1978) Chemotactic and chemokinetic stimulation of human eosinophil and neutrophil polymorphonuclear leukocytes by 12-L-hydroxy-5,8,10-heptadecatrienoic acid (HHT). j’ournal of Immunology, 120,526-531. Griesmacher, A., We&l, G., Schreiner, W. & Miiller, M. M. (1989) Thromboxane A2 generation by human umbilical endothelial cells. Thrombosis Research, 56, 6117623. Hamberg, M., Svensson, J. & Samuelsson, B. (1975) Thromboxanes. A new group of biologically active compounds derived from prostaglandin endoperoxides. Proceedings of the National Academy of Sciences, USA, 72, 2994-2998. Harper, M. J. K., Khodr, G. S. & Valenzuela, G. (1983) Prostaglandin production by human term placentas in vitro. Prostaglandins, Leukotrienes and Medicine, 11, 121-I 29. Haurand, M. & UIlrich, V. (1985) Isolation and characterization of thromboxane synthase from human platelets as a cytochrome P-450 enzyme. 3ournal of Biological Chemistry, 260,15059-l 5067. Hunt, J. S. (1989) Cytokine networks in the uteroplacental unit: Macrophages as pivotal regulatory cells.3oumal ofR@mductive Immunology, 16, 1-17. Miikil& U. M., Viinikka, L. & YIikorkaIa, 0. (1984) Increased thromboxane A2 production but normal prostacyclin by the placenta in hypertensive pregnancies. Prostaglandins, 27, 87-95. M%kilii, U. M., Jouppila, P., Kirkiien, P. L. & Ylikorkala, 0. (1986) Placental thromboxane and prostaqclin in the regulation of placental blood flow. ObstetM and Gynecology, 68,537-540. Mitchell, M. D., Bibby, J. G., Hicks, B. R., Redman, C. W. G., Anderson, A. B. M. & Tumbull, A. C. (1978) Thromboxane Bz and human parturition: Concentrations in the plasma and production in vitro.3oumal of Endom’nology, 78, 435-441. Moley, J., Bray, M. A., Jones, R. W., Nugteren, D. H. & van Dorp, D. A. (1979) Prostaglandin and thromboxane production by human and guinea-pig macrophages and leucocytes. Pmstaglandins, 17,730-736. Moncada, S. & Vane, J. R. (1979) Pharmacology and endogenous roles of prostaglandin endoperoxides, thromboxane A2 and prostacyclin. Pharmacological Reviews, 30,293-331. Nelson, D. M. SLWalsh, S. W. (1989a). Thromboxane and prostacyclin production by different compartments of the human placental viIlus.3oumal of Clinical Emiocrinology and Metabolism, 68, 676-683. Nelson, D. M. & Walsh, S. W. (1989b). Aspirin differentially affects thromboxane and prostacyclin production by trophoblast and villous core compartments of human placental villi. Ameri~un 3oumal of Obstetric and Gynecoloa, 161, 1593-1598. Nigam, S. & Zakrzewicz, A. (1990) Tumour cell proliferation by thromboxane A?: A receptor mediated event. In Advances in Prostaglandin, Thromboxane and Leukotriene Research (Ed.) Samuelsson, B. pp. 925-928. New York: Raven Press. Noort, W. A., de Zwart, F. A. & Keirse, M. J. N. C. (1987) Increase in urinary thromboxane excretion during pregnancy and labour. Prostaglandins, 34, 413-421. Niising, R. & UIIrich, V. (1990) Immunoquantitation of thromboxane synthase in human tissues. Eicosanoidr, 3, 175-180. Niising, R., Lesch, R. & Ulhich, V. (1990) Immunohistochemical localization of thromboxane synthase in human tissues. Eicosanoids, 3,53-58. Ntising, R., Sauter, G., Fehr, P., Diirtniiller, U., Kasper, M., Gudat, F. & UIlrich, V. (1992) Localization of thromboxane synthase in human tissues by monoclonal antibody Tii 300. Virchow S Archiv A, 421,249-254. Rota-Pleszczynski, M., Gagnon, L., Bolduc, D. & LeBreton, G. (1985) Evidence for the involvement of the thromboxane synthase pathway in human natural cytotoxic cell activity.3oumal oflmmunology, 135,4114-4119.
398 Swanson, M. L., Lei, Z. M., Swanson,
Placenta (1994), Vol. 15
P. H., Rae, Ch. V., Narumiya, S. & Hiram, M. (1992). The expression of thromboxane A2 synthase and thromboxane A2 receptor gene in human uterus. Biology of Reproduction, 47, 105-l 17. Walsh, S. W. (1985) Preeclampsia: An imbalance in placental prostacyclin and thromboxane production. American Journal of Obstetricsand Gynecology,152, 335-340. Wetzka, B., Schiifer, W., Breckwoldt, M. & Zahradnik, H. P. (1993a) Eicosanoid production by frozen tissue in vitro is markedly changed. Prostaglandins,Leukottienes and EssentialFatty Acids (in press). Wetzka, B., Sch%fer, W., Scheibel, M., Niising, R. & Zahradnik, H. P. (1993b) Eicosanoid production by intrauterine tissues before and after labour in short-term tissue culture. Prostaglandins,45, 571-581. Wiibelmsson, L., Wildand, M. & Wiqvist, N. (1981) PGH2, TXAZ and PG12 have potent and differentiated actions on human uterine contractility. Prostaglandins, 21, 277-286. Zahradnik, H. P., Schier, W., Wetzka, B. & Breckwoldt, M. (1991) Hypertensive disorders in pregnancyThe role of eicosanoids. Eicosanoids,4, 123-136.
Immunohistochemical Localization of Thromboxane Synthase in Human Intrauterine Tissues B. WETZKA”, W. SCHiiFER’, F. KOMMOSS”, H. BETTENDORF”, R. NihINGb, M. BRECKWOLDT” & H. P. ZAHRADNIKaTC a Universit&fiauenkh’nik II, Ettdokrinologieund Repmduktionsmedizin, HugstetterStr. 55, D-79106 Freiburg, Gevnany ’ UniversitiitKonstanz, Fakultiitfur Biologic, P.O. Box 5560, D78464 Konstanz, Germany ‘ To whom cowespondenceshould be addressed Paper accepted3.12.1993
SUMMARY Uterine tissues are known to be able to synthesize thromboxaneAs mz), but there is little information about the nature of cells actually responsiblefor its production. In this study human placenta, fetal membranes, umbilical cord and pregnant myometrium were investigatedimmunohistochemically.The avidin-biotin methodfor a monoclonalantibodyagainst human thromboxanesynthase (Tii 300) was applied on frozen tissue sections.In placenta, fetal membranes and umbilical cord, staining waspositivefor Hojbauer cellsandjbroblasts. Further, in sectionsof placenta, capillaryendotheliumshowed antigenityfor IXsynthase. Leiomyocytesin the umbilical cord vessels contained the enzyme as well. Preparations of pregnant myometrium were shown to express 7X synthase in leiomyocytes,endothelial cells and connectivetissue cells.Amnion, trophoblastand decidua did not possessantigenityfor this enzyme. Since 7’222 plays an important role for the regulation of vascular tone and aggregationof plateletsand may stimulate myometrialcontractionsduring parturition, the abundance of ‘Ix synthase in pregnancy-specifictissuesconfirmsprevious in vivo and in vitro observations.Further, 7X42 synthesizedby Hofbauer cells may be involved in immunologicalreactionsduring pregnancy, and the number and level of activation of Hofbauer cells may be closely related to the initiation of labour. Thmmboxane production by the endothelium lining the f&al vesselspoints to its reguhztoryrolefor the bloodflow in thefetoplacentalunit.
0143-4004/94/040389
+ 10 $08.00/O
0 1994 W. B. Saunders Company Ltd
390
Placenta(1994), Vol. 15
INTRODUCTION The arachidonic acid (AA) metabolite thromboxane A2 (TXAz) has been known for several years as a potent vasoconstrictor and an activator of platelet aggregation (Hamberg, Svensson and Samuelsson, 1975). It is the main eicosanoid produced in thrombocytes, but was shown to be synthesized in many other cell types in vitro. Recently a role for this prostanoid in tumour proliferation and metastasation (Nigam and Zakrzewicz, 1990) and immunological reactions (Rola-Pleszczynski et al, 1985) was proposed. The cellular effects of TXAz are receptor-mediated with an activation of phospholipase C catalysing the release of inositol-triphosphate and diacylglycerol as second messengers. During pregnancy, a significant increase in plasma levels (Mitchell et al, 1978) and urinary excretion (Noort, de Zwart and Keirse, 1987) of the more stable thromboxane metabolite TXBz could be observed. The elevated levels in urine were even more impressive after labour. The synthesizing capacity for TXAz of human tissues derived from pregnant uteri could be shown in vitro with different culture systems (Harper, Khodr and Valenzuela, 1983; M&ill, Viinikka and Ylikorkala, 1984; Walsh, 1985). In short-term tissue culture experiments conducted in our laboratory, TXAz was the main cyclooxygenase metabolite of AA produced by placenta and fetal membranes (Wetzka et al, 1993a, b). Further, special interest was focused on the TXAz production in tissues derived from women suffering from pregnancy-induced hypertension (PIH). The ratio of local prostacyclin/thromboxane level is supposed to be an important factor for the regulation of uterine blood flow (M&ilH et al, 1986). Several studies support the involvement of a PGIz/TXAz imbalance in the pathogenesis of clinical symptoms of PIH (reviewed in Zahradnik et al, 1991). However, there is little data about the identity and localization of the cells within the different tissue compartments for the pregnant uterus that are actually responsible for the production of TXAz (Niising, Lesch and Ullrich, 1990; Swanson et al, 1992). This study was carried out in order to elucidate the cellular distribution of the TX synthase in human placenta, fetal membranes, umbilical cord and myometrium via immunohistochemistry using a monoclonal antibody (mAb) against human TX synthase (Niising, Wernet and Ulhich, 1990).
MATERIALS AND METHODS Immunohistochemistry was carried out using the avidin-biotin method for mAbs. Vectastain Elite kit and peroxidase substrate 3-amino-9-ethylcarbazole (AEC) were purchased from Vector Laboratories, Burlingham, USA. As specific antibody Tii 300, a mAb against human platelet thromboxane synthase-characterized in Niising, Wemet and Ulhich (1990)-was applied. Tissue specimens were collected immediately after delivery, either vaginal or by caesarean section. ‘Labour’ was defined as the presence of regular myometrial contractions and the beginning of cervical dilatation. Four uncomplicated pregnancies went to term. Two of them delivered vaginally. Two underwent caesarean section, one before the onset of labour because of breech presentation, the other after the onset of labour because of fetal distress. In addition, two women suffering from PIH-defined as blood pressure over 140/ 90mmHg in the second half of pregnancy in women who were normotensive before pregnancy (for details see Zahradnik et al, 1991)- were delivered by caesarean section
391
Wet& et al: Thmmboxane Synthase in Intrauterine Tissues
after the onset of labour in weeks 36 and 38, respectively. All six babies had an uncomplicated perinatal period. Placenta, fetal membranes and myometrium were directly snap-frozen in liquid nitrogen. Pieces of umbilical cord were rinsed prior to freezing in Tyrode’s solution for removing clotted blood in the fetal vessels. The frozen sections (3-5 km) were mounted onto microscopic slides and allowed to dry at room temperature overnight. The structural integrity of the tissue specimens was established by control staining with haematoxylin and eosin. The slides for immunohistochemical procedure were fixed in ice-cold acetone for 1Omin. After drying, the sections were rinsed with 0.01 M phosphate-buffered saline (PBS; pH 7.2) and incubated with 1% (v/v) hydrogen peroxide for 10 min in order to inhibit endogenous peroxidases. After intensive washing with 0.5% (v/v) triton X-100 (Sigma, Deisenhofen, Germany) and PBS, preincubation was performed with horse serum (30 min). Then the slides were incubated with the specific antibody Tti 300 in a dilution of 1:200 in PBS containing 0.1% (v/v) bovine serum albumin for 30min. After rinsing three times in PBS, incubation with the second antibody, a biotinylated horse-anti-mouse antibody, was carried out for 30min. Following removal of surplus of antibody in PBS, the slides were incubated with avidin-peroxidase complex for 45 min at 37 “C. Then the bound peroxidase cleaved AEC at 37 “C for 30-45 min in the dark leading to a red-brown colour. Excess substrate was removed by intensive rinsing with distilled water and counterstaining performed with Mayer’s hemalum solution (Merck, Darmstadt, Germany). Finally the tissue sections were mounted under glass coverslips with Kaiser’s glycerol gelatin (Merck, Darmstadt, Germany). Negative controls for each tissue were performed in the same way as described above. While incubating the positive controls with the specific antibody Tti 300 following preabsorption with horse serum, the negative controls were incubated with plain PBS. Evaluation was performed with conventional light microscopy by Dr Komrnoss and Dr Wetzka. Scoring of staining intensity was done in a relative manner (Table 1). ‘+++’
Table I. Immunohistochemical localization of TX synthase in human uterine tissues at term of pregnancy with monoclonal antibody Tii 300. Staining intensity (+/++/ + + +) indicates positive staining for TX synthase, negative staining is shown as ‘-’ Tissue compartment
Cell types
Staining
Placental villi
Hotbauer cells Fibroblasts of villous core Endothelial cells Syncytiotrophoblast Trophoblast Hofbauer cells Fibroblasts Amnion epithelium Trophoblast Decidua Hotbauer cells Fibroblasts Leiomyoqtes of vessels Endothelial cells Fibroblasts/histiocytes Endothelial cells Leiomyocytes
+++ +++ ++/+++ -
Fetal membranes
Umbilical cord
Myometrium
Intensity:
+ weak, + + medium,
+ + + strong, - negative.
+++ +++ +++ +++ +/++ +++ ++/+-I-+ +/++
intensity
392
Placenta (I 994), Vol. 1.5
(strong) was given for maximum staining observed in our preparations, ‘+’ (weak) for minimum staining, respectively. ‘++’ (medium) scored cells were stained in between maximum and minimum. Identification of cell types was done with the help of Prof. Dr.med. N. BGhm, Chairman of the Department of Pediatric Pathology, University of Freiburg, Germany. Amniotic, trophoblastic, decidual and endothelial cells, leiomyocytes of vessels and myometrial preparations were identified according to their specific localization and presentation within the tissues. Cells were called ‘Hofbauer cells’ when they showed the following microscopic criteria: (1) typical localization in the core of placental villi, in the fibroblast and spongy layer of amnion and the reticular layer of chorion and in the Wharton’s jelly of the umbilical cord; (2) appearance with round, fusiform or stellate shape, eccentric nucleus, various granules or vacuoles (Boume, 1962). In some cases, Hofbauer cells could not be differentiated from fibroblasts because of their morphological resemblance.
RESULTS In all uterine tissues staining was positive for cells of the mononuclear phagocytic system and connective tissue cells. There were no major differences between tissue specimens obtained after vaginal deliveries or caesarean sections and from uncomplicated pregnancies or from patients with PIH, respectively (results not shown). The results for the individual tissue compartments together with the staining intensities of antigen-positive cells are shown in Table 1. In placenta, Hofoauer cells and fibroblasts lying in the villous core were stained strongly. There was no staining of the cytotrophoblast and the syncytiotrophoblast layer [Figure l(A)]. Decidua adjacent to placental villi was negative for the TX synthase [Figure l(B)]. Endothelial cells of some capillaries in the villous core [Figure l(A)] and of a few bigger fetal vessels in stem villi [Figure 1(C)] showed antigen@ for TX synthase. Fetal membranes were positive for Hofbauer cells and fibroblasts situated in the spongy layer of amnion and the reticular layer of chorion, whereas the amnion epithelium remained unstained [Figure 2(A)]. D eci‘dua adhering to the chorion layer exhibited only unspecific staining. Sections of the umbilical cord showed stained Hofbauer cells and fibroblasts in the Wharton’s jelly [Figure 3(A)]. Further, leiomyocytes of the vessel walls were positive for TX synthase [Figure 3(B)]. In the umbilical cord specimens endothelium did not exhibit positive staining. In myometrium preparations, leiomyocytes, connective tissue cells and endothelial cells showed positive antigen@ for TX synthase [Figure 4(A) and (B)]. Negative controls for either tissue are shown in Figures l(D), 2(B), 3(C) and 4(C).
DISCUSSION In accordance with Niising et al (1992), cells belonging to the mononuclear phagocytic system were strongly positive for TX synthase. Hotbauer cells lying in the core of placental villi, in the connective tissue layer between amnion and trophoblast in the fetal membranes, and in the Wharton’s jelly of the umbilical cord showed antigenity for the TX42 synthesizing enzyme. This finding agrees with the observations of TXBa being the major cyclooxy-
Wetzka et al: Thromboxane Synthase in Intrautm’ne Tissues
Fipre
1. Immunohistochemical localization of TX synthase in placenta with monoclonal antibody Tii 300. A, placental villi; B, decidua adjacent to placental villi; C, stem villus; D, negative control. The magnification bar represents 50 pm in A, 80 pm in C and 125 urn in B and D.
Figure 2. Fetal membranes. A, cross section of fetal membrane layers; B, negative control. The magnification bar
represents 125 pm. A, amnion; C, chorion; D, decidua; E, endothelium; F, fibroblast; H, Hotbauer cells; S, syncytiotrophoblast; V, villi of placenta.
394
Placenta(1994), Vol. 15
genase product formed from AA by human mononuclear cells (Moley et al, 1979) and of monocytes having the second highest content of TX synthase after platelets @&sing and Ulhich, 1990). The physiological significance of macrophages within the fetomatemal unit represents a new aspect for the immunology in pregnancy (Bulmer and Johnson, 1984). They seem to play a pivotal role concerning immune interactions between the mother and the fetoplacental graft (Hunt, 1989). Holbauer cells, from fetal origin, appear as early as the 4th week of pregnancy during the process of vasculogenesis in the early placental villi (Demir et al, 1989). They possess antigen-presenting abilities, produce a variety of cytokines and exhibit phagocygotic activity (Castelluci and Kaufmann, 1990). The expression of TX’ synthase in this cell population might play a role in regulation of cell proliferation (Nigam and Zakrzewicz, 1990) and activity of immunocompetent cells (Rola-Pleszczynski et al, 1985). An activation of macrophages by an inflammation during pregnancy might stimulate their TXAa synthesis and start myometrial contractions resulting in pre-term labour. The action of TX synthase on prostaglandin endoperoxides leads to the production of TXA2 and 12-hydroxyheptadecatrienoic acid (HHT) and malondialdehyde in a 1:l ratio (Haurand and Ulhich, 1985). 12-HHT, as well as other hydroxy AA products, can act as a chemotactic stimulant for polymorphonuclear granulocytes (Goetzl and Gorman, 1978). Pregnancy complications like PIH or HELLP syndrome (Haemolysis-elevated liver enzymes-low platelet count) are most probably caused by pathological immune reactions. So a changed TXAa synthesis by uterine macrophages may be expected to play an important role for the development of clinical symptoms. However, in this study there was no major difference in TX synthase expression between tissues from uncomplicated pregnancies compared with PIH pregnancies. Since the pathological processes of PIH start during placentation and effect placental circulation throughout pregnancy with the consequence of placental infarctions, it should be considered that at term of pregnancy most tissue changes are ‘burned-out’. Therefore an immunohistochemical examination of intrauterine tissues from the first and second trimesters of pregnancy could be expected to show interesting results concerning the question of pathogenesis of PIH. In addition, we found positive staining of fibroblasts neighbouring Hofbauer cells in all tissues studied. The predominance of TX synthase in cells derived from the mesoderm rather than in ectodermal cells corresponds to the findings of Bryant-Greenwood, Rees and Tumbull (1990) who described antigen@ for prostaglandin synthase more frequently in the connective tissue part of fetal membranes than in the trophoblast and decidua. These observations point to an important role of connective tissue cells (including tissue macrophages) for intrauterine prostaglandin synthesis. Further, leiomyocytes of vessel walls and pregnant myometrium and endothelial cells were stained specifically for TX synthase. A previous immunohistochemical study of pregnant myometrium conducted with a polyclonal antibody against porcine lung TX synthase showed the expression of the enzyme in myometrial smooth muscle and walls of myometrial blood vessels (Swanson et al, 1992). The authors did not distinguish between the muscle layer and the endothelium of vessel wall, therefore it is not clear whether they found endothelial cells showing antigenity for TX synthase. The results are comparable with our study except for two topics. Swanson et al (1992) described a difference in staining intensity between myometrium obtained before and after labour which could not be found in the present study. This can be explained in part by the use of a different antibody and probably by the smaller number of tissue samples investigated in our study. The negative staining of macrophages (and endothelial cells?) in the Swanson study may result from the use of a
Wet.2 :kn et al: Thmnboxane Synthase in Intrauterine Tis!WCS
Figure 3. Immunohistochemical localization of TX synthase in umbilical cord with mAb Tii 300. A, Wharton’s jelly; B, umbilical artery; C, negative control. The magnification bar represents 50 km in A and 80 pm in B and C. Figure 4. Pregnant myometrium. A, overall view; B, myometrial represents 200 km in A and 50 pm in B and C. E, endothelium; leiomyocytes.
vessel; C, negative control. The magnification bar F, fibroblast; H, Hofbauer cells; Hi, histiocyte; L,
396
Placenta (1994), Vol. 15
different antibody and technique of tissue preparation (paraffin-embedded tissues). The polyclonal antibody against porcine lung TX synthase used by Swanson et al (1992) is known to cross-react with TX synthases of most mammalian species. Since the size of these enzymes deriving from different species can vary, as Swanson et al (1992) showed by immunoblot-TX synthase of bovine lung had 65 kD and TX synthase of human uterus only 55 kD-it can be expected that the staining in immunohistochemistry might be slightly different and might not be as specific as if an antibody against human TX synthase is applied. Furthermore, Niising et al (1992) observed that mAb Tii 300 reacted strongly with TX synthase of human macrophages. The positive staining of endothelial cells for TX synthase in our study is not easy to justi@ with the hypothesis that activated thrombocytes stimulate the endothelial PGIz production by transfer of prostaglandin endoperoxides (Moncada and Vane, 1979). In this model of antagonistic PGHZ metabolism by platelets and endothelium, an expression of TX synthase in endothelial cells seems contradictory. However, human umbilical endothelial cells could be shown to synthesize TXA2 (Griesmaher et al, 1989). Therefore it must be assumed that endothelium expresses TX synthase under certain circumstances, i.e. during pregnancy. In all studies investigating the TX synthase in human tissues (Niising, Lesch and Ullrich, 1990; Niising et al, 1992; Swanson et al, 1992) endothelial cells stained positively for this enzyme in the placenta, the umbilical cord and the myometrium. The observation of elevated levels of TXBZ in plasma and urine during pregnancy (Mitchell et al, 1978; Noort, de Zwart and Keirse, 1987) and of TXAZ being the major cyclooxygenase metabolite of AA in tissue cultures of placenta and fetal membranes (Wetzka et al, 1993a, b) together with the findings described above point to the importance of uterine TXA2 for local regulation of blood flow and haemostasis. Additionally its contracting potency for myometrium (Wilhelmsson, W&land and Wiqvist, 1981) may control myometrial activity during labour. Our finding of negative staining for TX synthase in trophoblast cells seems to contradict the observation of Nelson and Walsh (1989a, 1989b). They separated placental villi into the trophoblast layer and the villous core and found nearly the same TXA2 production in both compartments, whereas intact villi produced only 10% TXA2 compared with the separated villous parts. However, they described vimentin-positive cells in the trophoblast preparation which could be fibroblasts and endothelial cells adhering to the trophoblast layer and could be responsible for the TXA2 production. Further, aspirin affects TXA;! production only in whole villus and villous core, but not in the trophoblast layer (Nelson and Walsh, 1989b). This observation points to a modulatory role of trophoblast cells on the TX synthesis of villous core cells rather than on their own TXAZ production. Therefore the trophoblast compartment might be responsible for maintaining a physiological balance between PGIz and TXA2 production in the placenta. In conclusion this study gives new aspects concerning the localization of TX42 production within the different tissue compartments of the pregnant human uterus. It points to an important role of macrophages in the fetoplacental unit for prostaglandin production. TXA2 synthesis in endothelium and leiomyocytes may be responsible for the control of intrauterine blood flow. Further immunohistochemical studies with the mAb against human TX synthase will be conducted to investigate the number of macrophages. These studies will use a double staining technique with an additional antibody against a monocyte and will compare tissues derived from women in labour/not in labour/pre-term labour/ PIH in order to elucidate the role of TXAZ in pregnancy and labour.
397
Wetzka et al: Thromboxane Synthase in Intrauterine Tissues
ACKNOWLEDGEMENTS We gratefully acknowledge Prof. Dr.med. in interpretation of the tissue preparations.
N. Biihm, Pathologisches
Institut der Universitiit
Freiburg
for his help
REFERENCES Boume, G. (1962) Holbauer cells. In The Human Amnion and Chorion (Ed.) Boume, G. pp. 95-112. London: Lloyd-Luke. Bryant-Greenwood, G. D., Rees, M. C. P. & Tumbull, A. C. (1990) Immunohistochemical localization of relaxin, prolactin and prostaglandin synthase in human amnion, chorion and decidua. Journal ofEn&rinology, 114,491-496. Bulmer, J. N. & Johnson, P. M. (1984) Macrophage populations in the human placenta and amniochorion. Clinical and Experimental Immunology, 57, 393-403. Castelluci, M. & Kaufmann, P. (1990) Hofbauer cells. In Pathology of the Human Placenta 2nd edn. (Ed.) Benirschke, K. & Kaufmann, P. pp. 71-80. New York: Springer. Demir, R., Kaufmann, P., Castelluci, M., Erbengi, T. & Kotowski, A. (1989) Fetal vasculogenesis and angiogenesis in human placental villi. Acta Anatomica (Base& 136, 190-203. Goetzl, E. J. & Gornmn, R. R. (1978) Chemotactic and chemokinetic stimulation of human eosinophil and neutrophil polymorphonuclear leukocytes by 12-L-hydroxy-5,8,10-heptadecatrienoic acid (HHT). j’ournal of Immunology, 120,526-531. Griesmacher, A., We&l, G., Schreiner, W. & Miiller, M. M. (1989) Thromboxane A2 generation by human umbilical endothelial cells. Thrombosis Research, 56, 6117623. Hamberg, M., Svensson, J. & Samuelsson, B. (1975) Thromboxanes. A new group of biologically active compounds derived from prostaglandin endoperoxides. Proceedings of the National Academy of Sciences, USA, 72, 2994-2998. Harper, M. J. K., Khodr, G. S. & Valenzuela, G. (1983) Prostaglandin production by human term placentas in vitro. Prostaglandins, Leukotrienes and Medicine, 11, 121-I 29. Haurand, M. & UIlrich, V. (1985) Isolation and characterization of thromboxane synthase from human platelets as a cytochrome P-450 enzyme. 3ournal of Biological Chemistry, 260,15059-l 5067. Hunt, J. S. (1989) Cytokine networks in the uteroplacental unit: Macrophages as pivotal regulatory cells.3oumal ofR@mductive Immunology, 16, 1-17. Miikil& U. M., Viinikka, L. & YIikorkaIa, 0. (1984) Increased thromboxane A2 production but normal prostacyclin by the placenta in hypertensive pregnancies. Prostaglandins, 27, 87-95. M%kilii, U. M., Jouppila, P., Kirkiien, P. L. & Ylikorkala, 0. (1986) Placental thromboxane and prostaqclin in the regulation of placental blood flow. ObstetM and Gynecology, 68,537-540. Mitchell, M. D., Bibby, J. G., Hicks, B. R., Redman, C. W. G., Anderson, A. B. M. & Tumbull, A. C. (1978) Thromboxane Bz and human parturition: Concentrations in the plasma and production in vitro.3oumal of Endom’nology, 78, 435-441. Moley, J., Bray, M. A., Jones, R. W., Nugteren, D. H. & van Dorp, D. A. (1979) Prostaglandin and thromboxane production by human and guinea-pig macrophages and leucocytes. Pmstaglandins, 17,730-736. Moncada, S. & Vane, J. R. (1979) Pharmacology and endogenous roles of prostaglandin endoperoxides, thromboxane A2 and prostacyclin. Pharmacological Reviews, 30,293-331. Nelson, D. M. SLWalsh, S. W. (1989a). Thromboxane and prostacyclin production by different compartments of the human placental viIlus.3oumal of Clinical Emiocrinology and Metabolism, 68, 676-683. Nelson, D. M. & Walsh, S. W. (1989b). Aspirin differentially affects thromboxane and prostacyclin production by trophoblast and villous core compartments of human placental villi. Ameri~un 3oumal of Obstetric and Gynecoloa, 161, 1593-1598. Nigam, S. & Zakrzewicz, A. (1990) Tumour cell proliferation by thromboxane A?: A receptor mediated event. In Advances in Prostaglandin, Thromboxane and Leukotriene Research (Ed.) Samuelsson, B. pp. 925-928. New York: Raven Press. Noort, W. A., de Zwart, F. A. & Keirse, M. J. N. C. (1987) Increase in urinary thromboxane excretion during pregnancy and labour. Prostaglandins, 34, 413-421. Niising, R. & UIIrich, V. (1990) Immunoquantitation of thromboxane synthase in human tissues. Eicosanoidr, 3, 175-180. Niising, R., Lesch, R. & Ulhich, V. (1990) Immunohistochemical localization of thromboxane synthase in human tissues. Eicosanoids, 3,53-58. Ntising, R., Sauter, G., Fehr, P., Diirtniiller, U., Kasper, M., Gudat, F. & UIlrich, V. (1992) Localization of thromboxane synthase in human tissues by monoclonal antibody Tii 300. Virchow S Archiv A, 421,249-254. Rota-Pleszczynski, M., Gagnon, L., Bolduc, D. & LeBreton, G. (1985) Evidence for the involvement of the thromboxane synthase pathway in human natural cytotoxic cell activity.3oumal oflmmunology, 135,4114-4119.
398 Swanson, M. L., Lei, Z. M., Swanson,
Placenta (1994), Vol. 15
P. H., Rae, Ch. V., Narumiya, S. & Hiram, M. (1992). The expression of thromboxane A2 synthase and thromboxane A2 receptor gene in human uterus. Biology of Reproduction, 47, 105-l 17. Walsh, S. W. (1985) Preeclampsia: An imbalance in placental prostacyclin and thromboxane production. American Journal of Obstetricsand Gynecology,152, 335-340. Wetzka, B., Schiifer, W., Breckwoldt, M. & Zahradnik, H. P. (1993a) Eicosanoid production by frozen tissue in vitro is markedly changed. Prostaglandins,Leukottienes and EssentialFatty Acids (in press). Wetzka, B., Sch%fer, W., Scheibel, M., Niising, R. & Zahradnik, H. P. (1993b) Eicosanoid production by intrauterine tissues before and after labour in short-term tissue culture. Prostaglandins,45, 571-581. Wiibelmsson, L., Wildand, M. & Wiqvist, N. (1981) PGH2, TXAZ and PG12 have potent and differentiated actions on human uterine contractility. Prostaglandins, 21, 277-286. Zahradnik, H. P., Schier, W., Wetzka, B. & Breckwoldt, M. (1991) Hypertensive disorders in pregnancyThe role of eicosanoids. Eicosanoids,4, 123-136.