The effect of fetal thyroparathyroidectomy on the transport of calcium across the ovine placenta to the fetus

The ovine fetal placenta has been perfused with autologous fetal blood under controlled conditions in eleven experiments in which the fetus was first removed. Eight of these experiments involved four pairs of twins, one lamb of which had been thyroparathyroidectomized (TXPTX) three to seven days earlier. By this time the normal placental calcium gradient from mother to fetus had either decreased or been reversed. The mean rate of transport of calcium from the mother was unchanged by previous fetal TXPTX, but the final calcium gradient achieved from the mother to the perfusing blood was significantly less than with placentae from intact fetuses. No significant alteration in fetal plasma I,25-dihydroxyvitamin D (I,25(OH)2D) concentration was observed as a result of the fetal TXPTX, indicating that hypocalcaemia can compensate for the lack of PTH in fetal production of I,25(OH)2D. Fetal thyroidectomy with replacement of thyroxine did not lead to reversal of the placental calcium gradient, indicating that calcitonin was not involved. It is suggested that in the ovine fetus, parathyroid hormone promotes the active transport of calcium from mother to fetus, so that in its absence the fetus must obtain its calcium for growth by reducing its calcaemia and thereby allow net diffusion of calcium to replace the action of the placental calcium pump. The price paid for this compensation is marked hypocalcaemia and defective calcification of osteoid.     

Epidermal growth factor binding and receptor distribution in term human placenta

Human placenta has a large number of epidermal growth factor (EGF) receptors when measured either by [125I]iodoEGF binding or by protein yield after purification. To localize EGF receptors in situ in normal human term placenta, two different light microscopic methods were used. To detect unoccupied, accessible EGF binding sites on the extracellular surface of placental cells in intact blocks of tissue, samples were incubated with [125I]iodoEGF, sectioned and autoradiography performed. To detect the total pool of intracellular and extracellular EGF receptors, placental tissue was sectioned, treated with detergent, and then anti-EGF receptor antibody was localized by immunohistoperoxidase techniques. Both [125I]iodoEGF and anti-EGF receptor antibody methods showed that EGF receptors were primarily present on syncytiotrophoblast cells of placental villi. Smooth muscle cells of placental blood vessels also contained EGF receptors. Neither connective tissue cells within the core of terminal chorionic villi nor endothelium of fetal blood vessels had detectable [125I]iodoEGF binding or immunoreactive EGF receptors. Since the quantity of placental smooth muscle cells is only a small fraction compared to trophoblast cells, we conclude that syncytiotrophoblast cells are primarily responsible for the high levels of EGF receptors found in extracts prepared from human term placenta.     

Developmental expression of vascular endothelial growth factor (VEGF) receptors and VEGF binding in ovine placenta and fetal membranes

The receptor tyrosine kinases, kinase-insert domain-containing receptor (KDR) and fms-like tyrosine kinase (Flt-1), and their ligand vascular endothelial growth factor (VEGF) are essential for the development and maintenance of placental vascular function during pregnancy. To further understand the role of VEGF in mediating angiogenesis and vascular permeability during development, the cellular localization of KDR and Flt-1 mRNA and protein, and the distribution of(125)I-VEGF binding sites in placenta, chorion and amnion of ovine fetuses were examined at three different gestational ages. In placentae at 62, 103 and 142 days, the predominant site of KDR mRNA and protein, and VEGF binding was the maternal vascular endothelium. In addition, a specific, although weak, signal for KDR mRNA was found in the maternal epithelium. At 103 and 142 days but not 62 days gestation, KDR mRNA and protein as well as VEGF binding sites were abundantly present in the endothelium of villous blood vessels. In the fetal membranes at 62, 103 and 142 days gestation, KDR mRNA and protein were expressed in the amniotic epithelium and intramembranous blood vessel endothelium, where binding of(125)I-VEGF was strong. There was no KDR mRNA or VEGF binding in the chorionic cytotrophoblast. Flt-1 expression was not detectable in placentae or fetal membranes at the three ages studied.In summary, the results demonstrated that VEGF receptors are present in the maternal and fetal vasculatures of the ovine placenta. This expression is consistent with a capillary growth-promoting function of KDR and its ligand VEGF. Further, the presence of KDR and VEGF binding sites in ovine fetal membranes suggests a role for VEGF in promoting intramembranous vascularity and permeability throughout gestation.     

Serum- and cell-mediated immune protection of mouse placenta and fetus against a Brucella abortus challenge: expression of barrier effect of placenta

When mice are intravenously inoculated with a virulent Brucella abortus strain at day 12 to 14 of pregnancy and killed three to five days later, colonization of placentae, fetuses and spleens can be estimated by the frequency and level (bacterial count) of infection and by linkage between individual placental and paired fetal infections. This linkage indicates the placental barrier effect, defined as the number of non-infected fetuses linked to 100 colonized placentae. Immune mice serum raised against two Brucella fractions injected one day before challenge (1) restricted the placental colonization (the dose required to infect 50 per cent placentae was increased by 50 to 70 times compared to controls), (2) decreased the level of splenic and placental infection, and (3) increased the barrier effect so that most fetuses were protected even when linked to a heavily infected placenta. Immune (B + T) spleen cells from mice vaccinated with a Brucella cell-wall fraction transferred to recipients seven to eight days before mating, that is, 22 days before challenge (1) restricted the frequency of placental and fetal colonization, (2) decreased the level of splenic, placental and fetal infections, and (3) increased the barrier effect. However, separated B- and T-cells were less active, in particular on the level of fetal infection. In contrast with serum, the cells did not decrease infection of the fetuses linked to heavily infected placentae.     

Regional distribution and pharmacological characterization of [125I]endothelin-1 binding sites in human fetal placental vessels

High-affinity binding sites for [125I]endothelin(ET)-1 have been detected in purified membrane preparations of the fetal arteries and veins of the chorionic plate and the stem villi vessels of the human term placenta. Regardless of the vessel type, the apparent dissociation constant was found to be in the picomolar range (26----45 pM), and the Bmax value close to 600 fmol/mg protein. In stem villi vessels, ET-1, ET-2, sarafotoxin S6b and vasocontractor intestinal peptide (VIC) were approximately equipotent in their competitive displacement of [125I]ET-1 binding. The endothelin precursors, human and porcine big-endothelin, recognized ET-1 sites with low affinity (nM range), a finding which reflects their low potency as recognized vasocontractant agents. Interestingly, [125I]ET-1 binding parameters and pharmacological profiles were identical in fetal veins and arteries of the chorionic plate. Similarly, a study carried out in rat aortic membranes, revealed the presence of high affinity [125I]ET-1 binding sites with pharmacological characteristics close to those of the human stem villi vessels. In all vessels investigated, the binding pattern of ET-3 against [125I]ET-1 was of a non-competitive nature. Thus, these results demonstrate the presence of specific [125I]ET-1 binding sites along the vascular tree of the fetal side of the placenta and would support evidence currently available, favouring the existence of distinct ET-1 and ET-3 receptors. Finally, ET-1 in the human placenta may play an important physiological role as regulator of vascular resistance and/or be implicated as a pathological factor in certain pregnancy-related diseases.     

Relation between placental morphometry and fetal growth]

Forty-eight placentae of full term infants, 21 placentae from appropriate for gestational age infants (AGA) and 27 placentae from small for gestational age infants (SGA) were measured by morphometric technic using the automatic image analyzer, in order to find out the extent of fetomaternal exchange which determines the transfer of oxygen and nutrition from mother to fetus and fetal growth. The results of measurement correlated well both with infant birth weight and placental weight. They demonstrated striking quantitative differences when the placentae of SGA were compared with those of AGA. The placenta weights in the group of SGA were notably less than those in the group of AGA. It seems that low birth weight relates to low functional tissue mass of placenta. This reduction of functional tissue is accompanied by diminution of the area for exchange between mother and fetus, both at the villous surface area and at fetal capillary surface area. Thus, the ability of transferring oxygen and nutrition from mother to fetus is curtailed. The results show that the rate of fetal growth is limited by placental function as well as its weight.     

The effect of a reversible period of adverse intrauterine conditions during late gestation on fetal and placental weight and placentome distribution in sheep

Adverse intrauterine conditions occurring during early to mid-gestation or throughout the whole of gestation influence placental weight and the distribution of placentome types in sheep. However, no study to date has investigated the effect of a reversible period of adverse intrauterine conditions during late gestation upon fetal and placental weight and placentome distribution in sheep. Twenty-two sheep fetuses were chronically instrumented with an inflatable cord occluder, amniotic and vascular catheters and with a Transonic flow probe around an umbilical artery. At 125 days (term isca.145 days) the occluder was inflated to reduce umbilical blood flow by ca.30 per cent for 3d in 12 fetuses (umbilical cord compressed, UCC). The occluder was then deflated and umbilical blood flow allowed to return to baseline. The remaining 10 fetuses acted as sham-operated controls in which the occluder remained deflated at all times. At 135-137dGA ewes were humanely killed and tissues collected, weighed and placentomes classified. A reduction in umbilical blood flow by approximately 30 per cent from baseline for 3 days in UCC fetuses led to mild fetal asphyxia throughout the period of cord-compression. After deflation of the occluder cuff, umbilical blood flow returned to a level that was significantly greater than that measured during baseline. Umbilical cord compression had no effect on fetal body weight but significantly increased fetal adrenal weight relative to body weight. While the total number of placentomes was not altered by cord-compression, total placentome weight and the total weight of C/D-type placentomes were both reduced in UCC relative to control placentae. In addition, the mean weight of placentomes, and of C/D-type placentomes specifically, was significantly lower in UCC relative to control placentae. When expressed as a percentage of the total number of placentomes in the placenta, there was a significantly lower percentage of C/D-type placentomes in UCC relative to control placentae. In addition, there was a significant relationship between the total number of placentomes and the percentage C/D-type placentomes in control, but not UCC, placentae. The data suggest that a temporary, reversible period of adverse intrauterine conditions occurring late in gestation in sheep has persisting effects upon the placenta, mean placentome weight and placentome distribution.     

Picornavirus infection in early murine gestation: significance of maternal illness

To evaluate whether maternal illness following picornavirus infection during pregnancy adversely affects placental and fetal health, mice were inoculated with the GDVII strain of Theiler's murine encephalomyelitis virus or control cell lysate during days 4-7 of gestation. Gross appearance, histopathology and viral culture, and in situ hybridization positivity of placentae and fetuses from ill GDVII-infected, healthy GDVII-infected and control mice were compared. Twenty of 34 (59 per cent) GDVII-infected dams became clinically ill. More placenta-fetus pairs from ill mice were grossly abnormal (68 per cent) than from well GDVII-infected (51 per cent;P< 0.01) or control mice (9 per cent;P< 0.001). Virus was detected by in situ hybridization in 73 per cent of placentae and 29 per cent of fetuses from sick GDVII-infected dams, and in 85 per cent of placentae and 19 per cent of fetuses from healthy GDVII-infected mice (differences not significant). Histological abnormalities consisting of necrosis or an increase in hyaline tissue in the vascular labyrinth layer were similarly frequent in placentae from ill and well GDVII-infected mice (58 per cent versus 67 per cent, P=0.5). Viral RNA, inflammation and necrosis were evident in the heart, great vessels, brain and spinal cord of GDVII-infected fetuses. Infection with GDVII in early pregnancy produces a high rate of gross placental and fetal abnormalities. The rate of gross abnormalities exceeds the incidence of fetal infection and more closely parallels the rates of infection and histopathology in the placenta, suggesting that much of the damage to placenta-fetus pairs is a consequence of placental infection. In addition, the occurrence of viral-induced maternal illness is associated with additive risk to placental and fetal health not explained by an increased rate of placental or fetal infection.     

Ovine intramembranous pathway permeability: use of solute clearance to determine membrane porosity.

OBJECTIVE: The contribution of the fetal chorioamniotic membranes (i.e. the intramembranous pathway) to the regulation and maintenance of amniotic fluid (AF) volume and composition has yet to be completely understood. Knowledge of membrane permeability properties is vital to understanding how the intramembranous pathway contributes to the overall maintenance of AF homeostasis. Although there are significant data regarding the regulation of intramembranous water flow, there is little understanding of the regulation of intramembranous solute flow. In the present study, we sought to determine the effect of molecular weight or size of non-polar compounds on intramembranous solute movement in the ovine model. METHODS: Five singleton ovine fetuses (117 +/- 3 days) were chronically prepared with bladder, tracheal, amniotic cavity and femoral arterial and venous catheters and an esophageal occluder. The allantoic membranes were excised. After 5 days' recovery, AF volume was calculated by intraamniotic injection of 99Tc-labelled red blood cells (time -6 to 0 h). At time 0, AF exchange routes were limited to the intramembranous pathway by inflation of the esophageal occluder and external drainage of fetal urine and lung fluid. Following intra-amniotic injection of creatinine (Cr, 1 g, MW 11 000 Da, 4 A) and [125I]albumin (RISA, 250 microCi, MW 69 000 Da, 36 A), maternal and fetal plasma and AF samples were collected at timed intervals during the subsequent 5 h. AF solute clearance (Cl(x)) was determined by the changes in AF total solute content. RESULTS: Cr and RISA disappeared from the AF with a corresponding increase in fetal, though not maternal, plasma levels. The mean Cl(Cr) was significantly greater than Cl(RISA) (2.0 +/- 0.3 ml/min vs. 1.0 +/- 0.2 ml/min; p < 0.04). CONCLUSION: Solute clearance from the amniotic cavity is inversely proportional to solute molecular weight/size. Although the membrane comprising the ovine intramembranous pathway is size restrictive, membrane pores allow passage of non-polar solutes up to 36 A. Knowledge of membrane permeability characteristics is essential for the utilization of the intramembranous pathway for fetal therapeutics.     

Ovine intramembranous pathway permeability: use of solute clearance to determine membrane porosity

Objective: The contribution of the fetal chorioamniotic membranes (i.e. the intramembranous pathway) to the regulation and maintenance of amniotic fluid (AF) volume and composition has yet to be completely understood. Knowledge of membrane permeability properties is vital to understanding how the intramembranous pathway contributes to the overall maintenance of AF homeostasis. Although there are significant data regarding the regulation of intramembranous water flow, there is little understanding of the regulation of intramembranous solute flow. In the present study, we sought to determine the effect of molecular weight or size of non-polar compounds on intramembranous solute movement in the ovine model. Methods: Five singleton ovine fetuses (117 &#45 3 days) were chronically prepared with bladder, tracheal, amniotic cavity and femoral arterial and venous catheters and an esophageal occluder. The allantoic membranes were excised. After 5 days' recovery, AF volume was calculated by intraamniotic injection of 99 Tc-labelled red blood cells (time &#109 6 to 0 h). At time 0, AF exchange routes were limited to the intramembranous pathway by inflation of the esophageal occluder and external drainage of fetal urine and lung fluid. Following intra-amniotic injection of creatinine (Cr, 1 g, MW 11 000 Da, 4 Å) and [ 125 I]albumin (RISA, 250 &#119 Ci, MW 69 000 Da, 36 Å), maternal and fetal plasma and AF samples were collected at timed intervals during the subsequent 5 h. AF solute clearance (Cl x) was determined by the changes in AF total solute content. Results: Cr and RISA disappeared from the AF with a corresponding increase in fetal, though not maternal, plasma levels. The mean Cl Cr was significantly greater than Cl RISA (2.0 &#45 0.3 ml/min vs. 1.0 &#45 0.2 ml/min; p < 0.04). Conclusion: Solute clearance from the amniotic cavity is inversely proportional to solute molecular weight/size. Although the membrane comprising the ovine intramembranous pathway is size restrictive, membrane pores allow passage of non-polar solutes up to 36 Å. Knowledge of membrane permeability characteristics is essential for the utilization of the intramembranous pathway for fetal therapeutics.     

The placenta: Immunologic or hemodynamic protector of the fetus?

Complete separation of the maternal and fetal circulation during normal pregnancy has been regarded as an extremely important protective factor from the immunologic standpoint. This hypothesis was tested in outbred rabbits in which a direct maternal-fetal parabiosis was established during the last week of pregnancy by implanting an intact maternal omental pedicle subcutaneously into the fetus. A functional cross-circulation, which developed after 48 hours as evidenced by maternal⁵¹Cr-labeled erythrocytes, led to pathologic changes culminating in fetal death within 70 to 80 hours postoperatively. Hyperacute graft rejection, previously postulated as the etiology of these fetal changes, was ruled out since the same characteristic syndrome occurred in similarly treated inbred fetuses syngeneic with their mothers. This study indicates that the absence of direct vascular intercommunication between mother and fetus at the level of the placenta is clearly necessary for physiologic as well as immunologic reasons.     

Effects of protein-calorie malnutrition on transplacental kinetics of aminoisobutyric acid in rats.

In order to find out if inefficient transport of amino acids contributes to a decrease in fetal weight during maternal malnutrition, we injected [14C]- and [3H]-labelled aminoisobutyric acid (AIB), respectively, in the mother and its fetuses and determined its transplacental kinetics on day 20 of gestation in rats fed a 21 per cent (control) or a low (5 per cent) protein diet. Rats fed a low protein diet consumed significantly less food than did the rats fed a control diet and thus suffered from protein-calorie malnutrition. A low protein diet led to a significant (P less than 0.05) decrease in maternal and fetal volume of distribution of AIB, a decrease in the clearance of AIB from the mother to the fetus and an increase in the time required for the fetal plasma AIB concentration following maternal injection to exceed the maternal plasma AIB concentration. The clearance of AIB from the fetus into the mother or to outside (e.g. amniotic fluid) was not altered by protein deficiency. It is concluded that a decrease in the efficiency of the placenta to deliver amino acids to the fetus may be a contributing factor in fetal growth retardation during maternal protein malnutrition.     

Transport of polar non-electrolytes across the intact and perfused guinea-pig placenta

The permeability of the mature placenta of the guinea-pig has been measured to several polar non-electrolytes of differing molecular weights and to two cations. Two methods were used: on the one hand, direct analysis of the fetus after it has taken up labelled solutes from maternal blood, and, on the other, perfusion of the umbilical circulation in situ. In the intact condition, fluorescein dextran of 3000 molecular weight and 51Cr-EDTA permeated from mother to fetus at rates proportional to their free diffusion coefficients in water, indicating the presence of large pores. Perfusion of the umbilical circulation with a physiological salt solution containing dextran increased permeability to these two solutes and hence total pore area three to four times. The movement of 14C-erythritol, 24Na and 42K was such as to suggest that they also cross the placenta by an additional route (or routes), which is not enchanced by perfusion. This route presumably traverses the plasma membranes of the syncytiotrophoblast.     

Uptake of vitamin A and retinol-binding protein by human placenta in vitro

The mechanisms involved in the transplacental passage of vitamin A bound to serum retinol-binding protein (RBP) were investigated. Samples of full-term human placenta were incubated for up to 3 h with 125I-RBP or [3H]retinol-RBP. The tissue 125I activity was monitored in washed, whole samples, while [3H]retinol was extracted from homogenized samples and separated by high-pressure liquid chromatography before counting. An accumulation of [3H]retinol and, to a much lesser degree, of 125I-RBP was noted. The placental uptake of both compounds was dependent on temperature and could be inhibited by adding unlabelled retinol-RBP. About 40 per cent of the tritium activity consisted of retinyl esters formed after the uptake of [3H]retinol. When washed placental specimens were postincubated for a few hours in a tracer-free medium, about 25 per cent of the 125I activity and 80 per cent of the tritium activity were released into the medium, provided that it was supplemented with fetal serum. The results illustrate two possible mechanisms by which vitamin A can pass the placental membranes: a direct transfer of retinol-RBP involving cellular uptake and release of the protein-ligand complex; cellular uptake of retinol by a cell-surface receptor specific for RBP, transient esterification of retinol, and release of the vitamin into fetal serum.     

Clearance of inert molecules, Na, and Cl ions across the primate placenta

Transplacental clearances of antipyrine, tritiated water, urea, Na, and Cl ions were studied in the rhesus monkey. In order of decreasing placental permeability determined in vivo, these experiments showed the following order: tritiated water and antipyrine, urea, chloride, and sodium ions. Clearances of tritiated water and antipyrine were equal. The placental clearances of tritiated water and antipyrine were significantly different when determined across the chorion laeve in vitro. These observations suggest that the transplacental diffusion of these molecules is flow limited. The diffusion rates of urea, Na, and Cl ions were limited primarily by the permeability of the placental membrane. The placenta of the rhesus monkey is more permeable to Na and Cl ions than the sheep placenta. Permeability to urea per kilogram of fetal body weight is of the same order of magnitude in both species.     

Effect of a failure of energy supply on adenine nucleotide breakdown in placentae and other fetal tissues from rat and guinea pig

The effects of ischaemia on adenylate energy charge of tissues from fetal rats and fetal guinea pigs were measured. Adult rat and guinea-pig tissues, as well as human placentae, were also studied. The largest differences observed were between the fetuses from different pregnant animals (P = 4.74 X 10(-15). Reductions in energy charge in placentae were slower than in other defined fetal tissues, especially brain. In the rat, an immature species at birth, greater 'stability' was observed in placentae of 14 days of gestation than near term at 20 days of gestation. As contrast, in the guinea pig, a mature species at birth, there was no difference in 'stability' in placenta or other fetal tissues between about 40 days of gestation and near term, about 60 days of gestation. In addition to these tissue and maturity effects in the fetus, it has been confirmed that fetal tissues are more 'resistant' than adult tissues to failures of energy supply. Concentrations of adenosine, uridine, guanosine and cytidine nucleotides in placenta show similar patterns in rats and guinea pigs. Fetal liver contains more uridine nucleotides and brain more cytidine nucleotides. It is suggested that the placenta retains an early fetal ability to maintain itself during ischaemia; this might be advantageous during parturition. Possible endocrine and other mechanisms 'damping' fetoplacental metabolism are linked with a discussion of the large maternal effect.     

Expression of apoptosis in placentae from mice lacking the prostaglandin F receptor

This study aimed to investigate the changes in apoptosis in the placenta and decidua of pregnant mice lacking the prostaglandin F receptor. Mouse placentae were removed from fetuses on days 10-23 of pregnancy. Apoptotic cells were examined by a DNA fragmentation assay and the terminal deoxynucleotidyl transferase-mediated dUDP nick end-labelling (TUNEL) technique. The placenta and decidual weight increased before day 18 and 14 of pregnancy, and then decreased with gestational day. After day 19, the fetuses gradually died in the uterus. All fetuses died in the uterus on day 23 of pregnancy. The number of apoptosis was not significantly different between wild type and FP-deficient mice before day 18 of pregnancy by DNA fragmentation and TUNEL staining. The DNA fragmentation was always more pronounced in decidual tissue on each day of pregnancy. DNA laddering on placentae was more extensive on day 22 than day 18. In placenta, most TUNEL-positive cells were detected in trophoblast and stromal cells. A higher intensity of apoptotic cells was in the decidual basalis. The main area was the centre of the decidual basalis, and was in decrease toward to margin of placenta. The index of TUNEL positive cells increased as gestation progressed toward termination. Especially, it was prominent in the placentae on day 22 compared with that day 18 of pregnancy. The increased TUNEL-positive staining in syncytiotrophoblast surface was found in placenta at post-term, compared with those at term. Apoptosis may provide insights into both normal placental development and placental dysfunction during an abnormal pregnancy from post-term pregnancy.     

Morphological aspects of the placenta in HIV pregnancies

Forty-nine placentae from HIV-seropositive mothers were collected in various hospitals in France and Belgium. Twenty [corrected] placentae with seven fetuses from interrupted pregnancies and 29 [corrected] placentae from spontaneous deliveries, including two stillborns and a set of twins, were studied morphologically. No significant abnormalities were observed in the aborted material. The placentae corresponding to deliveries presented no significant gross abnormalities but the ratio of fetal to placental weight was significantly decreased in the study group compared with the control group (6.13 versus 7.41; P less than 0.001), associated with a congestive and mature aspect of the parenchyma. Histologically a high incidence of chorioamnionitis (43 per cent) was found, contrasting with the absence of villitis. A relative villous hypercellularity was observed in the study group compared with the control group. Ultrastructural studies of 13 placentae corresponding to gestations of 10 to 40 weeks are presented. In six cases, retrovirus-like particles were found at various sites, such as villous fibroblasts, syncytiotrophoblast and endothelial cells, and in the free membranes.     

Electrical potential difference across the mid-term human placenta

Electrical potential difference across the mid-term human placenta was recorded during hysterotomy. The average value registered was 2.7 mV (S.E. of mean=0.4 mV, n=7), fetus negative. From this value and the concentrations of Na, K, Ca, Cl and inorganic phosphate in maternal and fetal plasma the possible mechanisms of net transport from mother to fetus are inferred. It is concluded that of the above ions only the transport of Na is compatible with simple diffusion.     

Placental barrier breakage in preeclampsia: ultrastructural evidence

Objective: It is known that the placenta acts as an immunological barrier between the mother and fetal "graft" allowing two antigenically different organisms to tolerate one another. Preeclampsia may be considered as a fetal rejection consequent to severe damage at placental endothelial and syncytiotrophoblast level. In order to verify this placental barrier damage we undertook the present study by electron microscopy. Study design: 14 placentae from preeclaptic women, and the same number of placentae from healthy controls were examined. Results: The results showed that endothelial cells from preeclamptic placentae express various and severe alterations, consisting of swollen and bulbous cytoplasm, degenerated inter-endothelial junctions with consequent crossing of fetal blood cells outside the vessels. Conclusions: These lesions could be the ultrastructural evidence of the placental barrier breakage leading to rejective reaction we presumed to be basis of preeclampsia.