Cytokine secretion by human fetal membranes, decidua and placenta at term

Cytokines such as monocyte chemotactic peptide-1 (MCP-1), interleukin-8 (IL-8), RANTES (Regulated on Activation and Normally T-cells Expressed and presumably Secreted) and interleukin-10 (IL-10) are thought to play pivotal roles in immune recognition, acceptance of the fetal allograft, maintenance of pregnancy and parturition. Their secretion and regulation within the third trimester uterus is, however, less well defined. We therefore investigated the release of these cytokines by third trimester amnion, chorion, placenta and decidua, and studied the influence of prostaglandin E2 (PGE2) infusion on their release in a dynamic placental cotyledon perfusion system. MCP-1 was released predominately by the chorion (78.2 +/- 7.3 pg/mg wet tissue weight; mean +/- SEM), decidua (112.4 +/- 5.2 pg/mg) and placenta (101.8 +/- 5.0 pg/mg) with low amounts from the amnion (1.3 +/- 0.4 pg/mg). High concentrations of IL-8 were released by the amnion (39.9 +/- 5.3 pg/mg), chorion (52.8 +/- 1.9 pg/mg), decidua (42.2 +/- 1.5 pg/mg) and placenta (45 +/- 1.3 pg/mg). Release of RANTES was not detectable from the amnion but was detected in moderate amounts from the chorion (6.0 +/- 1.2 pg/mg), decidua (15.2 +/- 1.4 pg/mg) and placenta (26.9 +/- 1.6 pg/mg). Low concentrations of IL-10 were secreted by the chorion (6.8 +/- 0.8 pg/mg), decidua (9.0 +/- 0.9 pg/mg) and placenta (3.3 +/- 0.3 pg/mg) with none detectable from the amnion. MCP-1, IL-8, RANTES and IL- 10 were all released by perfused placental cotyledons. PGE2 stimulated release of MCP-1, IL-8 and IL-10 into the maternal and of MCP-1 and IL- 8 into the fetal circulation of the placenta but had no effect on RANTES release. It is suggested that MCP-1 and IL-8 may be involved in the inflammatory process of parturition and IL-10 in the protection of the fetal allograft. In addition, PGE2 may have an important immunomodulatory role within the uterus at term.      

Expression of insulin-like growth factor in the placenta of intrauterine growth-retarded human fetuses

Many cases of intrauterine growth retardation (IUGR) are the result of placental and fetal tissue insufficiency. Insulin-like growth factor-I (IGF-I) is known to play a role in placental and fetal growth. An immunocytochemical study was performed to localize IGF-I peptides in human placenta and umbilical cords of normal (n = 3) and IUGR (n = 3) fetuses. The peripartum fetal conditions were evaluated as well. Immunoreactive IGF-I was detected in the cytotrophoblast, syncytiotrophoblast, amnion, endothelial cells of fetal capillaries and in the decidua in both normal and IUGR placental tissue. A more robust immunostaining and increased numbers of positively stained cells were found in the decidua of IUGR placenta (p < 0.001). Intense immunostaining was also found in endothelial cells, smooth muscle cells and fibroblasts of the umbilical vein. IGF-I immunoreactivity was also present in stroma (Hofbauer cells and/or fibroblasts) of IUGR villi. Our results indicate that expression of IGF-I is high in specific sites in placenta and umbilical cords, which indicates a paracrine and/or endocrine function. The increased expression of IGF-I in placenta of IUGR fetuses indicates its involvement in restoring normal growth by means of a positive feed-back mechanism.     

Differential activity of the gelatinases (matrix metalloproteinases 2 and 9) in the fetal membranes and decidua,associated with labour

Degradation of the extracellular matrix in fetal membranes has been implicated in the rupture of fetal membranes, the process of parturition and placental detachment from the decidua after parturition. In this study we assessed labour-associated changes in gelatinase activity in cultured human amnion, chorion and decidua, as well as in amniotic fluid. We found that in media conditioned by decidua, following the establishment of uterine contractions, matrix metalloproteinase-2 (MMP-2) activity is increased while the protein tissue inhibitors of matrix metalloproteinase-1 (TIMP-1) level is decreased. The formation of a 130 kDa gelatinase band was also significantly increased after contractions began. In media conditioned by chorion, the initiation of uterine contractions did not change MMP activity or TIMP-1 levels. However, an increase in MMP-9 activity and a decrease in TIMP-1 protein levels were observed following the establishment of uterine contractions in media conditioned by amnion. We suggest that this differential spatial regulation provides a form for modulatory hieratical activity of the MMPs in the onset of labour allowing rupture of the membranes while avoiding premature placental separation.     

Matrix metalloproteinases-2, -3 and -9 in human term placenta

Matrix metalloproteinases (MMPs) are zinc-dependent enzymes that degrade the components of the extracellular matrix (ECM) and are known to be the main mediators of human placentation and parturition. Although there are many studies on the roles and distribution of MMPs in human term placenta, so far none of the studies has investigated the distribution of MMP-2, -3 and -9 in different cells of various placental sites. In this study, we aimed to determine the distribution and enzymatic activities of MMP-2, -3 and -9 with regard to different regions of term human placenta, such as amnion, basal plate, chorionic plate, decidua, chorion laeve, Nitabuch's stria, umbilical cord and placental villi. Eighteen normal human term placentas were obtained after vaginal deliveries. Immunohistochemistry and zymography were performed for MMP-2, -3 and -9 on placental tissue sections and protein extracts, respectively. Nearly all tissues showed immunoreactivity for MMPs. The strongest enzymatic activity for MMP-2 was seen in areas where invasive trophoblast cells invaded maternal tissues. MMP-9 had the highest enzymatic activity at the contact region of fetal and maternal parts, suggesting the importance of MMP-9 in separation of the placenta from the uterine wall during labor. MMP-3 had a similar localization to MMP-9, suggesting that besides gelatinases like MMP-2 and -9, MMP-3 (stromelysin-1) may also have important roles during labor. This study describes the site-specific distribution and activities of MMPs and therefore might help in elucidating the molecular mechanisms in pathologies such as premature rupture of membranes.     

Prostaglandin endoperoxide H synthase mRNA expression in the human amnion and decidua during pregnancy and in the amnion at preterm labour

We have examined the expression of prostaglandin endoperoxide H synthase (PGHS) isoenzymes in the amnion and the decidua during gestation, and the abundance of PGHS mRNA in the amnion at idiopathic preterm labour. PGHS-1 and -2 mRNA abundance in the amnion, determined with ribonuclease protection assays, was significantly (P< 0.05) higher at term than earlier during pregnancy. In contrast, neither PGHS-1 and -2 mRNA values, nor PGHS-specific activity, measured with a cell-free assay, was different in the decidua at term as compared to earlier gestational ages. In individual term patients, PGHS-2 mRNA values in the amnion were positively correlated with PGHS-2 mRNA values in the chorion laeve. PGHS-1 and -2 mRNA abundance was higher (P < 0.05) in the amnion after idiopathic preterm labour than in the absence of labour at the same gestational age (28-35 weeks). Thus, PGHS-1 and -2 are induced in the amnion at term. Furthermore, amniotic PGHS-2 changes in co-ordination with PGHS-2 concentrations in the chorion laeve. PGHS is not induced in the decidua at term. Increased amniotic PGHS expression may contribute to the enhanced intrauterine prostaglandin synthesis before term labour. Both PGHS isoenzymes may participate in the increase of PGHS activity in the amnion at preterm birth.     

Isolation of mesenchymal stem cells of fetal or maternal origin from human placenta

Recently we reported that second-trimester amniotic fluid (AF) is an abundant source of fetal mesenchymal stem cells (MSCs). In this study, we analyze the origin of these MSCs and the presence of MSCs in human-term AF. In addition, different parts of the human placenta were studied for the presence of either fetal or maternal MSCs. We compared the phenotype and growth characteristics of MSCs derived from AF and placenta. Cells from human second-trimester (mean gestational age, 19(+2) [standard deviation, +/- 1(+3)] weeks, n = 10) and term third-trimester (mean gestational age, 38(+4) [standard deviation, +/- 1] weeks, n = 10) AF, amnion, decidua basalis, and decidua parietalis were cultured in M199 medium supplemented with 10% fetal calf serum and endothelial cell growth factor. Cultured cells were immunophenotypically characterized, the adipogenic and osteogenic differentiation capacity was tested, and the growth kinetics were analyzed. The origin of fetal and maternal cells was determined by molecular human leukocyte antigen typing.We successfully isolated MSCs from second-trimester AF, amnion, and decidua basalis as well as term amnion, decidua parietalis, and decidua basalis. In contrast, MSCs were cultured from only 2 out of 10 term AF samples. The phenotype of MSCs cultured from different fetal and maternal parts of the placenta was comparable. Maternal MSCs from second-trimester and term decidua basalis and parietalis showed a significantly higher expansion capacity than that of MSCs from adult bone marrow (p < .05). Our results indicate that both fetal and maternal MSCs can be isolated from the human placenta. Amnion is a novel source of fetal MSCs, likely contributing to the presence of MSCs in AF. Decidua basalis and decidua parietalis are sources for maternal MSCs. The expansion potency from both fetal and maternal placenta-derived MSCs was higher compared with adult bone marrow-derived MSCs.     

Matrix metalloproteinase (MMP)-2 and MMP-9 and their inhibitor, TIMP-1, in human term decidua and fetal membranes: the effect of prostaglandin F(2alpha) and indomethacin.

Degradation and breakdown of gestational membranes and the adjacent decidua are essential processes for the advancement of labour. We have assessed the effect of prostaglandin (PG) synthesis on the expression and activity of matrix metalloproteinase (MMP)-2 and MMP-9 and tissue inhibitor of metalloproteinases (TIMP-1) in fetal membranes at the edge of the placenta and decidua, by using ex-vivo organ culture of the tissues in the absence or presence of PGF(2alpha) (0.1, 1.0 and 10 microg/ml) or a PG synthesis inhibitor, indomethacin (10(-4)-10(-6) mol/l). Conditioned media were assessed for MMP by zymography on gelatin containing sodium dodecyl sulphate-polyacrylamide gels and for TIMP-1 by Western blot analysis. Compared to the membranes, decidua produced significantly more MMP-2 and MMP-9 as well as TIMP-1. PGF(2alpha) caused a 2.4- and 1.9-fold increase in the production of MMP-2 and MMP-9 in the decidua, respectively (P < 0.05), and an 11.3-fold increase of the active form of MMP-2 (62 kDa) which could hardly be detected in basal culture conditions (P < 0.01). PGF(2alpha) decreased TIMP-1 production by 70% in the decidua. The production of MMP-2 and MMP-9 and TIMP-1 by the amniotic and chorionic membranes was not affected by PGF(2alpha). Indomethacin decreased the production of MMP-2 and MMP-9 by 78 and 35% in chorion, and by 70 and 58% in amnion, respectively (P < 0.05), but did not affect production in decidual tissue. Indomethacin increased the production of TIMP-1 in chorion and amnion [by 4.1- and 4.5-fold respectively (P < 0.01)], but had no effect on decidua. Cumulatively, PGF(2alpha) increases decidual gelatinolytic activity. Meanwhile the inhibition of PG production by indomethacin reduces total gelatinolytic activity in fetal membranes, possibly accounting for some of its labour-arresting property.     

Progesterone receptor expression in human decidua and fetal membranes before and after contractions: possible mechanism for functional progesterone withdrawal

In humans, progesterone levels are sustained before the onset of labour. Therefore, the mechanism for parturition that has been proposed for humans is 'functional' progesterone withdrawal. Immunohistochemical staining for the progesterone receptor (PR) was positive in the decidua with a decline after contractions began. Western blot analysis revealed a number of PR isoforms expressed in the decidua, with the PR-B form being dominant. After contractions began, all PR isoforms decreased sharply. PR-B and PR-A decreased by 85.8% +/- 6.7 and 78.2% +/- 7.1, respectively (P < 0.001). Incubation of decidua with Prostaglandin F2alpha 1.0 microg/ml decreased the expression of all forms of PR isoforms. PR-B was reduced by 64% +/- 6.09 (P < 0.01); PR-A was reduced by 77% +/- 5.9 (P < 0.05), while PR-C was reduced by 80% +/- 7.24 (P < 0.05). Progesterone (80 microg/ml) increased the PR-B, PR-C the 45 and 36 kDa isoforms to 150% +/- 7.89, 210% +/- 12.4, 270% +/- 9.7 and 216% +/- 13.5, respectively (P < 0.05). In immunohistochemical studies, the PR was not identified in the amnion or in the chorion, regardless of the presence or absence of contractions. Western blot analysis demonstrated that PR-C (60 kDa) and the 36 kDa isoforms were dominant in the amnion. After contractions began, PR-A decreased significantly by 61.9% +/- 7.1 (P < 0.001).     

HLA-DR positive cells in the human placenta.

A population of heterogeneous HLA-DR positive cells has been identified in the human placenta and decidua using immunochemical and histochemical methods. These cells are found in three areas: the subepithelial layer of the amnion, the decidua, and more sparsely within the chorionic villous stroma. In addition to HLA-A, -B and -C antigens, they also express the leucocyte-common antigen, indicating their origin from bone marrow precursors. The majority have a characteristic stellate shape with many cytoplasmic processes. In the villous stroma these stellate cells can be distinguished from the Hofbauer cells (placental macrophages) by their morphology, stronger expression of HLA-DR and lack of lysosomal enzyme activity. In the amnion and decidua they cannot be clearly distinguished from tissue macrophages. By using monoclonal antibodies specific for foetal or maternal HLA-A or -B allotypes, the HLA-DR positive cells in the chorionic villi and the amnion have been shown to be foetal in origin. In contrast, most of the HLA-DR positive cells in the decidua are maternal; a few adjacent to the basal plate are foetal. The preponderance of these cells in those areas of the placenta where foetal and maternal tissues are in close proximity is striking. The possibility that some of the cells are equivalent to the dendritic cells that have been described in other tissues is discussed.     

Production of granulocyte colony-stimulating factor at the materno-foetal interface in human pregnancy.

A bioassay specific for human granulocyte colony-stimulating factor (G-CSF) was developed and used to measure G-CSF production in human pregnancy tissues. G-CSF was secreted by both foetal chorionic villous and maternal decidual tissues taken in the first trimester and at term. The level of G-CSF production by placental tissue was 6750 (1250-10,000) units of bioactivity per g of tissue in 48 hr in the first trimester and 104 (83-190) U/g at term. Bioactive G-CSF was also secreted by decidual tissue, more in the first trimester than at term. ELISA immunoassays measured 75 (10-820) ng/g/48 hr of G-CSF antigen from first trimester placenta, 15 (10-50) ng/g from first trimester decidua and less than 2 ng/g from term placenta. RNA isolated from decidual and chorionic villous tissue or from cells purified by flow cytometry, contained G-CSF mRNA in both tissues. In decidua, mRNA for G-CSF was confined to the macrophages, and cytotrophoblast from term amniochorion contained no detectable G-CSF mRNA. No G-CSF, measured as bioactivity or as mRNA, was detectable in choriocarcinoma cell lines.     

Preferential expression of endogenous type C viral antigen in Rhesus placenta during ontogenesis

A radioimmunoassay for the major internal structural protein (p26) from Macaca arctoides type C virus (MAC-1) was used to score antigen expression in rhesus placenta and other fetal organs obtained at various times of gestation. Antigen expression was detected in 16 out of 16 placental specimens but not in fetal heart, lung, spleen, liver, kidney, intestines, skeletal muscle, testes, and endometrial scrapings of decidua from eight selected animals, nor in thymus from two animals. The levels of antigen detected in placenta ranged between 2 and 218 ng/mg protein with a correlation between lower antigen expression and term gestation or parity greater than 10. A 10-fold higher level of antigen expression was detected at the external surface of placenta near decidua than in the remainder of the placenta toward the amniotic surface. Thus, even within the placenta, there was a preferred site for endogenous retroviral antigen expression. Last, substantial interindividual variation exists in regard to the amount of placental p26 expression in preterm animals of low or intermediate parity.     

Potentiation of the hCRF-induced release of ACTH in man by an opioid antagonist

Summary Administration of synthetic human corticotropin-releasing factor (hCRF; 2 µg/kg body weight) to six normal male subjects produced a significant rise in plasma ACTH, followed by an increase in circulating cortisol. Simultaneous treatment with the opioid antagonist naloxone (1.6 mg i.v. bolus, followed by an infusion at a rate of 1.2 mg/h) significantly potentiated the hCRF-induced rise in ACTH and enhanced the cortisol response to hCRF. It is suggested that naloxone acts by antagonizing an inhibitory ultra-short-loop feedback effect of coreleased -endorphin on pituitary corticotrophs, thereby amplifying the net effect of hCRF, i.e., the release of ACTH.ACTH adrenocorticotropic hormone - CRF corticotropin-releasing factor - hCRF human CRF - oCRF ovine CRF - EKG electrocardiogram - POMC proopiomelanocortin - RIA radioimmunoassay - S.E.M. standard error of the mean This study was supported by the Deutsche Forschungsgemeinschaft (Go 299/3-2; Mu 585/2-2)     

Increased adrenomedullin protein content and mRNA expression in human fetal membranes but not placental tissue in pre-eclampsia

The relationship between Pre-eclampsia (PE) and placental production of Adrenomedullin (AdM) is not completely understood. This study measured placental and fetal membrane AdM protein concentrations by specific radioimmunoassay and mRNA expression by Northern blot analysis in samples obtained at either term or preterm gestation from women either in labour or not in labour. Samples were obtained from women with normotensive and pre-eclamptic pregnancies. There were significant increases in immunoreactive AdM protein concentration (pg/mg DNA) in choriodecidua and amnion of women with PE compared to normal pregnancy for the preterm not-in-labour group (choriodecidua: control 124 +/- 16, n = 10, PE 361 +/- 35, n = 10; amnion: control 94 +/- 12, n = 10, PE 153 +/- 19, n = 10) and for the term not-in-labour (choriodecidua: control 128 +/- 17, n = 14, PE 459 +/- 51, n = 8; amnion: control 112 +/- 15, n = 14, PE 253 +/- 57, n = 8) and in-labour (choriodecidua: control 531 +/- 74, n = 14, PE 881 +/- 188, n = 8; amnion: control 545 +/- 84, n = 14, PE 1008 +/- 230, n = 8) groups. AdM mRNA relative abundance was greater in preterm, not-in-labour choriodecidual samples in PE, but not in amnion. In addition, this study observed labour-associated increases in choriodecidual and amniotic irAdM in term pre-eclamptic and control patients. However, there were no significant changes in AdM protein or mRNA expressions between any of the groups for placental tissue. These results suggest that fetal membranes, but not placental, production of AdM is increased at the post-translational level during PE in preterm and term tissues and at the pre-translational level during PE in preterm tissues. Fetal membranes, AdM may play an important role in the regulation of feto-placental hemodynamics and fetal physiology during pre-eclampsia.     

The pregnant guinea-pig as a model for studying influenza virus infection in utero: infection of foetal tissues in organ culture and in vivo

Organ cultures of guinea-pig foetal tissues showed a similar pattern of susceptibility to influenza virus to that already observed for human (Rosztoczy et al., 1975) and ferret (Sweet, Toms and Smith, 1977) foetal tissues. Respiratory, alimentary and urogenital tract tissues were susceptible whereas neural and lymphopoietic tissues were insusceptible. However, of the foetal membranes (amnion, chorion, umbilical cord and placenta) only the chorion was susceptible, in contrast to the corresponding ferret tissues, all of which were susceptible. The insusceptibility of the placenta paralleled that of human placenta which is similarly haemomonochorial in structure. Following intracardial inoculation of high titre virus (ca 10(9-4) EBID50) into pregnant guinea-pigs virus was isolated from all foetal membranes (amnion, chorion, umbilical cord and placenta), but in low titre. Although sporadic isolations were made from foetal tissues (intestine, kidney, heart, liver and spleen) there was no evidence for viral replication in these tissues. These results are discussed in relation to possible infection of the human foetus in utero with influenza virus.     

The pregnant guinea-pig as a model for studying influenza virus infection in utero: infection of foetal tissues in organ culture and in vivo.

Organ cultures of guinea-pig foetal tissues showed a similar pattern of susceptibility to influenza virus to that already observed for human (Rosztoczy et al., 1975) and ferret (Sweet, Toms and Smith, 1977) foetal tissues. Respiratory, alimentary and urogenital tract tissues were susceptible whereas neural and lymphopoietic tissues were insusceptible. However, of the foetal membranes (amnion, chorion, umbilical cord and placenta) only the chorion was susceptible, in contrast to the corresponding ferret tissues, all of which were susceptible. The insusceptibility of the placenta paralleled that of human placenta which is similarly haemomonochorial in structure. Following intracardial inoculation of high titre virus (ca 10(9-4) EBID50) into pregnant guinea-pigs virus was isolated from all foetal membranes (amnion, chorion, umbilical cord and placenta), but in low titre. Although sporadic isolations were made from foetal tissues (intestine, kidney, heart, liver and spleen) there was no evidence for viral replication in these tissues. These results are discussed in relation to possible infection of the human foetus in utero with influenza virus.     

Macrophage inhibitory cytokine 1 in fetal membranes and amniotic fluid from pregnancies with and without preterm labour and premature rupture of membranes

The placenta and fetal membranes are the site of expression of macrophage inhibitory cytokine (MIC-1), a member of the transforming growth factor (TGF)-beta superfamily. We hypothesized that MIC-1 may act as an immune regulator in pregnancy complications associated with intrauterine inflammation. Decidual cells, chorionic trophoblasts and amnion epithelial cells were identified by immunohistochemistry as the predominant MIC-1-containing cell type in term membranes. Amnion and choriodecidual explants all produced MIC-1 in culture, the latter having the greatest production rate (206 +/- 74.5 pg/mg tissue/24 h, n=6; mean +/- SEM). Production was not responsive to stimulation by pro-inflammatory cytokines. MIC-1 was detectable in 217 transabdominal amniotic fluid (AF) samples taken from 15 to 41 weeks gestation, concentrations ranging from 0.9-51.1 ng/ml. AF MIC-1 concentrations in pregnancies with premature rupture of membranes (PROM) or preterm labour, either with or without microbial invasion of the amniotic cavity, were not significantly different from those delivered at term either with or without labour. Treatment with MIC-1 (0.25-25 ng/ml) did not alter production of interleukin-6 or -8 by amnion or choriodecidual cells in vitro. We conclude that AF MIC-1 is derived from the fetal membranes and decidua, but that MIC-1 is unlikely to be involved in the pathophysiology of preterm birth or PROM.     

Production of inhibin forms by the fetal membranes, decidua, placenta and fetus at parturition

Inhibins are regulators of paracrine and endocrine function during pregnancy, but their intrauterine sites of secretion are not well established. In amniotic fluid, inhibin A-, inhibin B- and inhibin pro-alphaC-containing isoforms were present in high concentrations, whereas in maternal serum, inhibin A and pro-alphaC forms were present in high amounts, with low concentrations of inhibin B. In fetal cord serum, inhibin pro-alphaC was present in all samples, inhibin B was detectable in male but not female fetuses, with no detectable inhibin A in either sex. From cultured explants, both inhibin A and B were secreted by chorion laeve, whereas only inhibin A was secreted by placenta, with both tissues secreting inhibin pro-alphaC. Only low concentrations of both dimeric inhibins and pro-alphaC forms were secreted by decidua parietalis and amnion. The dual perfused placental cotyledon secreted both inhibin A and pro-alphaC into maternal perfusate, but only inhibin pro-alphaC into the fetal circulation and less than to the maternal side. We conclude that trophoblast is the predominant source of dimeric inhibins, but with markedly different secretion depending on its intrauterine location. There was a significant decrease in inhibin A and pro-alphaC in amniotic fluid collected at term active labour compared to elective Caesarean section (P < 0.001). This may reflect a local change in inhibin/activin processing at labour, likely in chorion laeve trophoblast cells, which may be important in the paracrine control of the feto-maternal communication required to maintain pregnancy and initiate labour.