Current Obstetrics & Gynaecology (1998) 8, 2-7 © 1998 Harcourt Brace & Co. Ltd Mini-symposium: The placenta Aspects of structure and function in human placenta T. M. Mayhew and L. Leach This review summarizes essential features of the functional morphology of human term placenta, concentrating on the processes of proliferation, growth, diffusive transport and microvascular permeability. It introduces the main structures that make up the 'villous membrane' interposed between the maternal and fetal bloods. It then presents an updated view of the proliferation of the principal functional compartment of the membrane, the trophoblast. This is a continuously renewing epithelium: cytotrophoblast cells divide mitotically throughout gestation and are recruited into the overlying syncytium. Contrary to previous dogma, cytotrophoblast is not depleted during gestation. The syncytium loses nuclei in large aggregates (syncytial knots), which detach into the maternal circulation. At least some nuclei are apoptotic and may be phagocytosed by macrophages at extraplacental sites. The villous stroma and fetal endothelium also grow by proliferation. These processes help to expand exchange surface areas and minimize diffusive distances, structural quantities that can be used to estimate placental-oxygen diffusive conductance. The fetal vascular compartment contributes substantially to overall transplacental resistance to solute transport. Fetal vessels are lined by a continuous endothelium with well-differentiated junctional complexes in the paracellular clefts. These complexes contain adhesion molecules that are vulnerable to exogenous agents, and whose expression and localization have been linked with junctional disruption and altered permeability, and altered placental efficiency and permeability. Changes in placental proliferation, growth, diffusive transport and vascular permeability may all play a role in pregnancy-related disorders such as pre-eclampsia and diabetes. OVERVIEW OF PLACENTAL STRUCTURE In the human haemochorial placenta, villous trees bathed by maternal blood circulating through the intervillous space are crucial to placental growth, morphogenesis and function and, hence, to fetal well-being.I,: Because of their number and physical attributes, terminal villi (TV) are the most influential in determining functional activity, and exchanges between maternal and fetal blood occur via the villous T. M. Mayhew, L. Leach, School of Biomedical Sciences University of Nottingham, Queen's Medical Centre, Nottingham NG7 2UH, UK Correspondence to: T.M.M. membrane (VM), which comprises trophoblast, stroma and the endothelium of fetal mierovessels. Trophoblast is a two-compartment epithelium. An inner proliferative zone, the cytotrophoblast (CT), transforms during gestation from a complete layer to a set of dispersed cells from which post-mitotic cells are recruited into an outer terminally-differentiated syncytiotrophoblast (ST). The syncytium is all but unique in human tissues and may have evolved to allow invasion of maternal tissues without breaching the intervascular barrier. Moreover, syncytium also permits the economical regional redistribution of ST mass and, hence, economical adaptations to improve diffusive conductance? Its effectiveness in this regard is related to its mean thickness, variability of thickness Placental structure and function 3 and surface area. During pregnancy the VM becomes thinner, whilst surface area and volume increase enormously, further enhancing placental functional capacity. Most expansion occurs after 20 weeks by the formation of new TV. 2'44 The second major discriminatory barrier of the villous membrane is the fetal endothelium. This is a continuous endothelium with numerous tight and adhaerens junctions present in the intercellular clefts. Fetal microvessels are fairly 'tight' and their permeability values approach those seen for skeletal muscle microvessels rather than the leakier microvessels found in the heart, liver or kidney.7,8 Dilation and peripheralization of fetal microvessels, together with dispersal of CT cells, lead to localized attenuation of the VM (at vasculosyncytial membranes, thickness less than 2 gm), as well as thickened regions where ST nuclei aggregate (syncytial knots, thickness 10-15 gm). Surface area expands faster than volume so the ST becomes progressively thinner overall. This design ensures t h a t maternal and fetal vascular beds are brought into close proximity over an extensive area. The beds are separated by the VM and this paper focuses attention on some recent changes in our perceptions of the VM and its structural ingredients. A NEW VIEW OF TROPHOBLAST PROLIFERATION AND GROWTH The prevailing notion of trophoblast growth has been one of conditional expansion limited by a shrinking pool of CT cells? ,9 Recent stereological studies on the actual numbers of CT and ST nuclei have altered this view to one of a continuously renewing epithelium (like small intestine), in which recruitment and extrusion are regulated. ~°'H From the first trimester onwards, trophoblast expands by the continuous proliferation of CT cells and increases in the absolute numbers of CT and ST nuclei. Because CT cells become more widely dispersed as ST area expands, 6,J°the use of 2D histological sections gives the erroneous impression that CT cells decline in number during gestation. In fact, not only are CT cells not depleted by recruitment into ST, but the numerical ratio of ST:CT nuclei (about 9:1)and volume of protoplasm per nucleus (about 1100 ~tm3) are maintained. Ratios are preserved despite the shedding of nucleated ST fragments as syncytial knots. Maybe as many as 150 000 fragments per day enter the maternal systemic circulation. The proportion of trophoblast volume occupied by CT cells at term (about 15%) also remains remarkably constant over the villous tree,4 suggesting that trophoblast turnover occurs at all levels. The notion of continuous recruitment is consistent with CT cells still being mitotically active near term and with discontinuity of the CT layer, but it is incompatible with nett loss of CT cells. 1,9,12In full-term placentae, CT cells can be seen forming desmosomes and Fig. 1 Electronmicrographof a term placentalchorionicvillus showinga syncytialknot (sk) arisingfrom the underlying syncytiotrophoblast(syn).The nuclearaggregatesin the syncytial knots appear pyknoticand, sometimes,apoptotic.Paler staining cytotrophoblast(cy) and fetal vessels(fv)lined with continuous endothelium(e) can be seen. The villousstroma (s) contains macrophagesand pericytes.Bar = 10 gm adhaerens junctions with the overlying ST. Immunochemistry reveals brightly fluorescent punctate staining of the epithelial adhaerens adhesion molecules E-cadherin and ~-catenin on the apical surface of CT cells (unpublished observation). The gradual transformation from a two-layered to an essentially single-layered epithelium has implications for maternofetal transfer of immunoglobulins. The intact CT layer may act as a barrier to IgG transfer ~3at least until its cells separate and disperse. IgG is known to be transcytosed through endosomal compartments of the ST and fetal endothelium? 4 The qualitative morphology of trophoblast has been examined microscopically.1,2 Such studies raise the question of how a constant volume of trophoblast per nucleus might be conserved. One possibility is via the loss of nucleus-rich ST fragments (Fig. 1). CT nuclei vary in size but tend to be rounded and euchromatic with, occasionally, prominent nucleoli. ST nuclei are smaller, indented and more heterochromatic. In syncytial knot regions, they are more pleiomorphic, densely packed, heterochromatic and convoluted. Finally, certain regions of ST harbour closely packed nuclei with apoptotic features. 15 In other tissues, apoptotic nuclei are removed by macrophages. This appears not to happen in ST, although it is possible that syncytial-knot 4 Current Obstetrics & Gynaecology fragments in the maternal circulation are phagocytosed by macrophages at extraplacental sites. It is probable that extrusion of syncytial knots normally occurs without disrupting the integrity of the VM. Occasionally, the VM is injured leading to local denudation of ST and release of syncytial knot fragments. Apoptosis may be another process that initiates denudation? 5 PLACENTAL GROWTH IS MONOPHASIC AND NOT BIPHASIC Until recently, ~°'1~there was uncertainty about whether placental growth is biphasic (a proliferation phase followed after 36 gestational weeks by a hypertrophic phase) or monophasic (solely proliferative). Wider aspects of villous growth from 12 weeks to term have been reassessed by stereological studies I°,Hwhose findings contradict the claim that placental growth is biphasic. Numbers of CT, ST, stromal and endothelial nuclei were found to increase exponentially and roughly in parallel, outstripping changes in placental volume and suggesting that growth in different compartments is tightly regulated. Growth varied between compartments, but nuclear proliferation was always dominant. In stroma, the volume of tissue per nucleus declined. In capillaries, ,the mean area of an endothelial squame increased and squame density per unit length declined. Immunochemical studies suggest that there are centres of proliferation for epithelium, stroma and vascular endothelium? 2 The growth constant for endothelial nuclei may be greater than that for CT and stromal cells, and this fits the idea that genesis and growth of TV are affected by the linear growth of fetal vessels) Fetal endothelium expresses CD44, thyl and A10-33/1, markers associated with proliferating endothelial cells such as those present in malignant tumours and in culture (unpublished, results). CORRELATING STRUCTURE WITH DIFFUSIVE CONDUCTANCE Variation in thickness of the VM is beneficial. It allows more effective diffusion than would be possible with a uniform membrane of identical arithmetic mean thickness 3and is achieved, partly, by redistributing ST mass following obtrusion by underlying fetal vessels or by active protoplasmic flow. 16'17The outcome is a Combination of vasculosyncytial membranes and syncytial knots. The critical structural determinants of placental diffusive conductance (diffusing capacity, Dp) are exchange surface areas and effective (harmonic mean) diffusion distances. Dp (ml.min-X.kPa -1) measures the ease with which a gas or nutrient diffuses across the placenta, and depends on the physical properties of tissues and physicochemical properties of the diffusing substance. The intervascular pathway for oxygen diffusion can be analysed as a set of up to six tissue layers each of which offers a partial resistance to flow. Being arranged serially, the partial resistances can be summed to obtain overall resistance, which is the reciprocal of total Dr. The six layers allow: (1) dissociation of oxygen from haemoglobin in maternal erythrocytes, me; diffusion across (2) maternal plasma, mp; (3) trophoblast, tr; (4) villous stroma, st; (5) fetal plasma, fp; and (6) association with haemoglobin in fetal erythrocytes, fe. TM Hence, total resistance to oxygen flow can be expressed as: f llDp = 1/Dme + l/Drop + l/Dtr + l/Dst + 1/Dfp + l/Dr. The partial conductances Dm~ and Df~ depend on vascular-space volumes and oxygen-haemoglobin reaction rates. The conductances Drop, D,~r Dst and Dfp are governed by Fick's law of diffusion and, consequently, by exchange surface areas (S), tissue-layer thicknesses (T) and tissue permeability to oxygen (K). Each conductance can be estimated via a modified Fick equation: D = K.S/T h where T h is harmonic mean thickness and S is the average of the upstream (maternal) and downstream (fetal) surfaces of each layer. A modified Fick equation is required because the arrangement of tissue ingredients in human placenta is complex. Trophoblast varies in thickness (from vasculosyncytial membranes to syncytial knots) and so oxygen conductances vary locally. For this reason, it is better to estimate harmonic rather than arithmetic thicknesses 3 thereby giving greater weight to thinner regions. In addition, the downstream side of the stroma is represented, not by a sheet, but by the endothelium of individual capillaries. Consequently, capillary and villous surface areas may be unequal. The same may apply to trophoblast and other layers. Since oxygen must cross both surfaces of each layer, it is preferable to take S as the mean of two surfaces. Provided that certain sampling requirements are met, stereological estimation of the key structural quantities provides a theoretically maximal Dp that might be found under optimal conditions. In reality, malperfusion, mismatching of maternal and fetal blood flows, and arteriovenous shunting reduce the efficiency of oxygen diffusion. Physiological estimation o f Dp is complicated further by the problems of estimating oxygen tensions at the sites of exchange. An added advantage of this morphometric approach is that changes in all compartments of the pathway are assessed and given appropriate weighting. The approach has been used to explore the relative resistance to oxygen transfer contributed by each layer, and has indicated that VM accounts for about 90% of total resistance. 1819 , The surface available for oxygen diffusion increases during gestation whilst the harmonic mean thickness Placental structure and function of the VM falls. Therefore, D P should increase and, indeed, evidence for this has been adduced ~ suggesting that earlier uncertainties about how fetal growth continues despite declining relative volumes and surfaces of villi might be attributed to failure to monitor a sufficient set of structural variables. Influential changes occur in the trophoblast and stroma and, from l0 to 41 weeks, the rise in total D P is commensurate with the gain in fetal weight, indicating that functional maturation of the placenta is matched to fetal growth. Estimates o f Dp for oxygen have also been made in abnormal pregnancies associated with hypoxia. These include high altitude (hypobaric hypoxia), maternal anaemia (normobaric hypoxia), preeclampsia (ischaemic hypoxia) and maternal diabetes mellitus (in which chronic fetal stress is indicated by elevated levels of fetal haemoglobin and erythropoietin). In all cases, there is thinning of the VM (with or without impoverished growth of villi and expansion of the intervillous space) and partial, total or specific diffusion conductances are increased) °43 On the Fick model, exchange surface area and harmonic thickness strongly influence D P. Of the two, the latter has the more impact, and this makes good sense because it is an economical and effective strategy for improving Dp. 19As an adaptive strategy, producing more TV has the possible disadvantage of increasing blood volume and placing an extra burden on the fetal cardiovascular system beyond that afforded by, for example, elevated haematocrits. BARRIER FUNCTION OF THE FETAL MICROVASCULAR ENDOTHELIUM The general consensus of endothelial biologists is that permeability of continuous microvessels is conferred by resistances in series: the luminal gl~ycocalyx, fibre matrix of the inter-endothelial wide zones, tight junctions and basement ,membrane. The structural complexity and molecular organization of these ingredients depend on the tissue in which the vessels are located and on whether they lie in the arteriolar, capillary or venular parts of the local circulation. The ultrastructure of placental fetal endothelium suggests that it is a fairly restrictive barrier to transport of solutes. Unlike other non-brain continuous capillaries, the dilated capillaries in TV contain few caveolae. Coated vesicles, endosomes and free vesicles are present, so the endothelium is capable of vesicular transport? ~ The paracellular clefts between adjoining endothelial cells offer the major transport pathway for hydrophilic solutes, and possess from one to four tight junctional regions? Where adjoining membrane leaflets are closely apposed, but not fused, there is a roughly 4 nm separation. This gap may allow transport of water and solutes less than 4 nm in diameter. Serial sectioning of tight junctions has shown that 5 Fig. 2 Confocal micrographs of term placental fetal microvessels which have been tilted in the Y axis. VE-cadherin, an adhesion molecule exclusive to endothelial adhaerens junctions, can be seen as bright fluorescent blebs (arrowheads) on the luminal membrane of the endothelium lining fetal vessels (fv). Progressive tilting (5 ° tilts, a~:l) reveals that staining is not continuous but is punctate along the paracellular clefts of vessels (arrows). Trophoblast (t) is also indicated. Bar = 10 gm. they are not continuous throughout the length of the capillaries but disappear within four to seven serial sections when section 60-70 nm thick are used. The separation of leaflets in these discontinuities (wide zones) is about 17 nm. Hydrophilic molecules, therefore, have a tortuous route to negotiate as they cross from the abluminal to luminal side of the blood vessel. The discontinuous tight junctions are a common feature of non-brain capillaries. The wide zones of paracellular clefts contain junctional complexes called adhaerens junctions (Fig. 2). These contain transmembrane adhesion molecules that belong to the cadherin group of cell-cell adhesion molecules. The latter are linked to the internal actin cytoskeleton viX peripheral linking molecules ~catenin, [3-catenin, plakoglobin, vinculin and c~actinin. 24 All cadherins have an extracellular portion with calcium binding, adhesive and glycosylation sites, as: well as a transmembrane domain and cytoplasmic tail that possess phosphorylation and ~cytoskeletal binding sites. Thus, their structure suggests susceptibility to both external and internal cues that may affect homophilic binding a n d s o regulate junctional integrity and permeability. The extracellular portions of cadherins may, along with their role in endothelial cell-cell adhesion, be part of the fibre-matrix molecular sieve that influences solute transport. Placental adhaerens junctions are rich in VE-cadherin. 25In vitro studies using human umbilical vein endothelial cell (HUVEC) monolayers have shown tha( VE-cadherin is expressed in cell-cell contact only when cells reach 6 Current Obstetrics & Gynaecology confluence. Furthermore, blocking VE-cadherin with antibodies results in gap formation and leakage of haem proteins? 4 VE-cadherin is linked to cytoplasmic actin via cx-catenin, J3-catenin and plakoglobin. These peripheral linking molecules are also vulnerable to phosphorylation and are thought, therefore, to be ligands for signal transduction. Changes in these molecules, or in actin, may affect localization and binding of VE-cadherin with resultant junctional separation and increased permeability. Perfusion of human placental microvessels for 30rain with 100 gM histamine 26 results in altered localization of VE-cadherin, a twofold increase in the separation observed in tight junctional regions, and an 80% increase in leakage of cyanocobalamin (RMM 1200). Recent observations in our laboratory have revealed that histamine also affects immunolocalization of catenins. In vitro studies using HUVEC cells have demonstrated that adhaerens junctional molecules are Vulnerable to a range of inflammatory mediators and cytokines such as thrombin, bradykinin, tumour necrosis factor-~ and interleukins 1 and 6. Incubations with these vasoactive agents result in changes in both the expression and immunolocalization of these molecules, as well as altering permeability. Endothelial cells can be isolated from microvessels of the term placenta 27 and also express junctional adhesion molecules that appear vulnerable to exogenous agents. The effects of long duration diseases such as diabetes and pre-eclampsia, where elevated glucose levels and abnormal blood flow are major factors, is an area of recent interest. Endothelial cells in culture are vulnerable to high glucose, there are reports of altered expression of PECAM-1 ?8 Ongoing research in our laboratory suggests that diabetes and high glucose may be linked with differential expression of VE-cadherin and PECAM-1 in placental microvessels. PECAM-1, a molecule thought to be involved in adhesion and transmigration of leucocytes, is found on the luminal membrane and the intercellular cleft of placental microvessels. It was localized to separate but neighbouring membrane microdomains to that seen for VE-cadherin. 25 This molecule appears to be vulnerable to histamine ~6 and its downregulation has been linked to increased endothelial permeability. Ultrastructural studies show that" placental endothelium possesses luminal and abluminal caveolae that are connected to one or the other plasma membrane with no interconnecting vesicles or transendothelial channels. The endothelial cells possess coated vesicles, early and late endosomes, lysosomes and free vesicles - organelles for endocytosis and transcytosis of macromolecules such as I g G ? 4'29 The endothelial plasma membranes contain heterogeneous-receptor populations necessary for uptake and transcytosis of macromolecules, such as insulin-like growth factor and IgG? 9 The luminal surface of microvessels is lined by an extensive glycocalyx that extends into the mouth of the paracellular cleft and luminal caveolaeY It is therefore, a constituent of both the transcellular and paracellular diffusion pathway. As stated before, the overall surface area of the microvessels plays an important role in transplacental diffusion of solutes. TV cont]ain very long, looped capillaries that form sinusoids. These may be adaptations to reduce the thickness of the intervascular barrier (thus increasing the diffusion capacity) and slow fetal blood flow (thus increasing the time available for the exchange of substances between maternal and fetal circulations). Both hypobaric hypoxia and gestational diseases, such as pre-eclampsia and diabetes, appear to influence the extent of vascularization and capillary diameter. 19,22 MICROVASCULAR PERMEABILITY Placental physiologists have been concerned mainly with the permeability of the intervascular barrier, the individual contribution of endothelium being largely ignored. However, it has been pointed out that the placenta has the characteristics of a filter with infrequent large pores and numerous small pores in series, and that the latter could well be endothelial. 3°Human placental microvessels appear to be fairly restrictive. Although horseradish peroxidase (HRP, R M M 40 000 Da) crosses the paracellular clefts when placental vessels are perfused with high concentrations of cationic HRP, larger molecules such as IgG appear to take a transcellular route?" By using the single-passage multiple=tracer dilution technique, it has been shown that there is a substantial restriction to the diffusion of radiolabelled cyanocobalamin (RMM 1353, molecular radius 0.84 nm) in term human placental microvessels perfused extracorporeally. 7 Comparing the permeability values for cyanocobalamin and EDTA in placental microvessels with those published for some other microvascular beds shows that term human microvessels are marginally less permeable than skeletal muscle capillaries, and far more restrictive than cardiac capillaries. ACKNOWLEDGEMENTS We wish to thank all our colleagues who have worked with us on the research cited in this paper, including Wan Ismail, an Honours student, for supplying the electron micrograph of Figure 1. Our placental researches have been supported by The Anatomical Society of Gt Britain & Ireland, The Cunningham Trust, Leverhulme Trust and Wellcome Trust. REFERENCES 1. Dearden L, Ockleford CD. Structure of human trophoblast: correlation with function. In: Loke C, White H (eds). Biology of Trophoblast. London: Elsevier, 1983:69-110 2. Kaufmann P, Burton GJ. Anatomy and genesis of the placenta. In: Knobil E, Neil JD (eds). The Physiology of Reproduction. New York: Raven Press, 1994:441-484 3. Jackson MR, Joy CF, Mayhew TM, Haas JD. Stereological studies on the true thickness of the villous membrane in Placental structure and function 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. human term placentae: a study of placentae from highaltitude pregnancies. Placenta 1985; 6:249~58 Sen DK, Kaufmann P, Schweikhart G. Classification of human placental villi. II. Morphometry. Cell Tissue Res 1979; 200:425-434 Boyd PA. Quantitative structure of the normal human placenta from 10 weeks of gestation to term. Early Human Develop 1984; 9:297-307 Jackson MR, Mayhew TM, Boyd PA. Quantitative description of the elaboration and maturation of villi from 10 weeks of gestation to term. Placenta 1992; 13: 357-370. Eaton BM, Leach L, Firth JA. Permeability of perfused term human placental microvessels. J Physiol 1993; 463: 141-155 Leach L, Firth JA. Fine structure of the paracellular junctions of terminal villous capillaries in the perfused human placenta. Cell Tiss Res 1992; 268:447-452 Arnholdt H, Meisel F, Fandrey K, Lohrs U. Proliferation of villous trophoblast of the human placenta in normal and abnormal pregnancies. Virchows Archiv B Cell Pathol 1991; 60:365-372 Mayhew TM, Simpson RA. Quantitative evidence for the spatial dispersal of trophoblast nuclei in human placental villi during gestation. Placenta 1994; 15:837-844 Mayhew TM, Wadrop E, Simpson RA. Proliferative versus hypertrophic growth in tissue subcompartments of human placental villi during gestation. J Anat 1994b; 184:535-543 Blankenship TN, King BE Developmental expression of Ki67 antigen and proliferating cell nuclear antigen in Macaque placentas. Develop Dynamics 1994; 201:324-333 Bright NA, Ockleford CD. Cytotrophoblast cells: a barrier to maternofetal transmission of passive immunity? J Histochem Cytochem 1995; 43:933 944 Leach L, Eaton BM, Firth JA, Contractor SE Immunocytochemical and labelled tracer approaches to uptake and intracellular routing of immunoglobulin-G (IgG) in the human placenta. Histochem J 1991; 23:444449 Nelson DM. Apoptotic changes occur in syncytiotrophoblast of human placental villi where fibrin type fibrinoid is deposited at discontinuities in the villous trophoblast. Placenta 1996; 17:382391 Jackson MR, Mayhew TM, Haas JD. On the factors which contribute to thinning of the villous membrane in human placentae at high altitude. I. Thinning and regional variation in thickness of trophoblast. Placenta 1988a; 9:1-8 Jackson MR, Mayhew TM, Haas JD. On the factors which contribute to thinning of the villous membrane in human 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 7 placentae at high altitude. II. An increase in the degree of peripheralization of fetal capillaries. Placenta 1988b; 9:9-18 Mayhew TM, Jackson MR, Boyd PA. Changes in oxygen diffusive conductances of human placentae during gestation (10-41 weeks) are commensurate with the gain in fetal weight. Placenta 1993a; 14:51-61 Mayhew TM, Jackson MR, Haas JD. Microscopical morphology of the human placenta and its effects on oxygen diffusion: a morphometric model. Placenta 1986; 7:121-131 Jackson MR, Mayhew TM, Haas JD. Morphometric studies on villi in human term placentae and the effects of altitude, ethnic grouping and sex of newborn. Placenta 1987b, 8: 487-495 Mayhew TM, Sorensen FB, Klebe JG, Jackson MR. Oxygen diffusive conductances in placentae from control and diabetic women. Diabetologia 1993b; 36:955-960 Mayhew TM, Sorensen FB, Klebe JG, Jackson MR. Growth and maturation of villi in placentae from well-controlled diabetic women. Placenta 1994a; 15:57-65 Reshetnikova OS, Burton GJ, Teleshova OV. Placental histomorphometry and morphometric diffusing capacity of the villous membrane in pregnancies complicated by maternal iron-deficiency anemia. Am J Obstet Gynecot 1995; 173: 724-727 Dejana E, Corada M, Lampugnani MG. Endothelial cell-cell junctions. The FASEB Journal 1995; 916:910-917 Leach L, Clark P, Lampugnani M-G, Arroyo AG, Dejana E, Firth JA. Immuno-electron characterisation of the interendothelial junctions of human term placenta. J Cell Sci 1993; 104:1073-1081 Leach L, Eaton BM, Westcott EDA, Firth JA. Effect of histamine on endothelial permeability and structure and adhesion molecules of the paracellular junctions of perfused human placental microvessels. Microvasc Res 1995; 50: 323-337 Leach L, Bhasin Y, Clark P, Firth JA. Isolation of endothelial cells from human term placental villi using immunomagnetic beads. Placenta 1994; 15:355-364 Baumgartner-Parzer S, Wagner L, Pettermann M, Gessl A, Waldhausl W. Modulation by high glucose of adhesion molecule expression in cultured endothelial cells. Diabetologia 1995; 38:1367-1370 Bright NA, Ockleford CD, Anwar M. Ontogeny and distribution of Fc gamma receptors in the human placenta. Transport or immune surveillance? J Anat 1994; 184:297-308 Stulc J. Extracellular transport patlaways in the haemochorial placenta. Placenta 1989; 10:113-119