Human placental syncytiotrophoblast microvillous membranes impair maternal vascular endothelial function

Objective To investigate the hypothesis that, should there be an increase in deported syncytiotrophoblast microvillous membrane fragments in pre-eclampsia, it may cause maternal vascular endothelial dysfunction.
Design Syncytiotrophoblast microvillous membrane (STEM) vesicles, prepared from normal term placentae, were perfused through small subcutaneous arteries isolated from fat biopsies obtained at caesarean section. Endothelial function of these arteries was studied by determining acetylcholine-induced relaxation after preconstriction with noradrenaline. As controls, physiological buffer or red blood cell membranes in physiological buffer were used and endothelial function similarly estimated. Transmission electron microscopy was performed on arteries after perfusion.
Sample STBM vesicles, isolated from the placentae of three healthy women undergoing elective caesarean section for reasons unrelated to pre-eclampsia, were suspended in physiological buffer. Subcutaneous fat arteries were obtained from a separate group of 13 normotensive pregnant women, also undergoing elective caesarean section at term.
Results Perfusion with red blood cell membranes or physiological buffer had no significant effect on the concentration dependent relaxation in arteries preconstricted with noradrenaline. However, after 2 h perfusion with STBM vesicles, arteries showed a significant reduction in relaxation to acetylcholine, indicative of altered endothelial function. Transmission electron microscopy of arteries perfused with STBM vesicles confirmed endothelial disruption.
Conclusions STBM vesicle perfusion specifically altered the relaxation response of preconstricted maternal subcutaneous fat arteries to acetylcholine, suggesting an alteration in endothelial dependent relaxation. Deported microvilli may therefore be capable of producing endothelial cell damage and endothelial dysfunction observed in the maternal syndrome of pre-eclampsia.     

Endothelial function in myometrial resistance arteries of normal pregnant women perfused with syncytiotrophoblast microvillous membranes

Objective To investigate the effects of syncytiotrophoblast microvillous membranes (STBM) in concentrations, found in vivo in women with pre-eclampsia, on endothelial function in isolated resistance arteries.
Setting Department of Obstetrics and Gynaecology, Huddinge University Hospital, Stockholm.
Sample Twenty-nine myometrial resistance arteries isolated from biopsies of healthy term pregnant women, obtained during caesarean section.
Methods The myometrial arteries were mounted in a pressure arteriograph and perfused intraluminally for three hours with STBM (20 to 2000ng/mL) or with erythrocyte membranes or physiological salt solution as controls, all substituted with 0.5% bovine serum albumin. Bradykinin concentration-response curves were performed before and after perfusion.
Main outcome measures The bradykinin concentrationresponse curves were fitted to the Hill equation and maximal dilation and the  pEC₅₀  values were determined from these fits. Differences within groups were analysed with a paired  Student's t test  . Electron microscopic evaluation of the endothelium was performed.
Results Neither STBM nor erythrocyte membrane perfusion affected maximal dilation or the  pEC₅₀  values of the bradykinin concentration-response curves at any concentration. Examination by electron microscopy showed no obvious damage to the endothelium after perfusion with STBM or erythrocyte membranes.
Conclusion Perfusion with STBM in concentrations up to 100 times those reported in pre-eclampsia has no significant effect on bradykinin-mediated dilation in isolated myometrial arteries.     

Adhesion molecules of syncytiotrophoblast microvillous membranes inhibit proliferation of human umbilical vein endothelial cells

It has been shown previously that syncytiotrophoblast microvillous membranes (STBM), isolated from normal or pre-eclampsia placentae, specifically inhibit the proliferation of cultured human umbilical vein endothelial cells (HUVEC) and disrupt the cell monolayer without causing cell death. We have previously shown that this anti-proliferative activity resides in a self-aggregating complex in which eight proteins, namely integrins alpha(5)(CD49e) and alpha(V)(CD51), dipeptidyl peptidase IV (DPP IV, CD26), alpha-actinin, transferrin receptor (TfR, CD71), transferrin, placental alkaline phosphatase (PLAP) and monoamine oxidase A (MAO-A) were identified. In the present study, we investigated which of these components causes the anti-proliferative activity of STBM. Antibodies against integrin alpha(5)and alpha(V)and DPP IV all reduced the STBM-induced inhibition of proliferation of HUVEC, which was also reversed by added fibronectin. A preparation of PLAP inhibited endothelial proliferation, but this was not due to enzymatic activity. The preparation was shown to be impure with more than 12 bands present on Coomassie blue stained SDS-PAGE gels. These included integrins alpha(5)and alpha(V), which could account, at least in part, for the inhibitory activity. We could not exclude, however, the possibility of other unidentified factors being involved. We conclude that adhesion molecules account for a major part of the anti-proliferative activity of STBM; these appear to compete for ligands in the extracellular matrix or serum with the appropriate receptors on HUVEC.     

Proteins associated with activity of Fc receptors on isolated human placental syncytiotrophoblast microvillous plasma membranes

To identify Fc receptors from human placental microvilli, proteins that were liberated by detergents from human placental synctiotrophoblast microvillous membranes (StMPM) were characterized by their abilities to bind human IgG in immune complexes with sheep or goat anti-human IgG and to monomeric rabbit anti-dinitrophenol (DNP) IgG bound to DNP-lysine Sepharose. Three placental IgG-binding proteins coprecipitated with immune complexes (Mr = 68,000, 52,000-56,000, 40,000) and were designated pIBP68, pIBP56 and pIBP40, respectively. Of the three proteins only pIBP56 bound to immobilized monomeric rabbit IgG. It was isolated from detergent lysates of StMPM and LDS/phenol glycoprotein extracts of placental plasma membranes suggesting that pIBP56 was a glycoprotein FcR previously reported (Mikulska et al, 1982). The binding specificities of pIBP56 and pIBP40 appeared to be detergent dependent. Photoaffinity crosslinking of StMPM surface proteins in situ to monomeric rabbit derivatized with N-succinimidyl(4-azidophenyl)-I, 3-dithiopropionate identified IgG-binding proteins identical in size to pIBP56 and pIBP40. Crosslinking further suggested that monomeric IgG covalently bound to a complex of StMPM proteins with a total size of 110,000-120,000 Mr. The findings suggest that pIBP68, pIBP56 and pIBP68 are responsible for IgG binding activity of placental StMPM.     

Effect of syncytiotrophoblast microvillous membrane treatment on gene expression in human umbilical vein endothelial cells

OBJECTIVE: Syncytiotrophoblast membrane fragments (STBM) exist in the peripheral circulation in pregnant women and it has been shown that the level of circulating STBM is significantly increased with pre-eclampsia compared with uncomplicated pregnancies. STBM could be one of the factors which directly causes the endothelial cell dysfunction of pre-eclampsia. This study investigates the effect of STBM on endothelial cell gene expression. DESIGN: Human umbilical vein endothelial cells were cultured in the presence and absence of STBM. At specified time points, total RNA was purified from the cultures and analysed on microarrays. SETTING: A laboratory investigation using placentas obtained from a hospital delivery ward. SAMPLE: Placentas from nine healthy women were obtained. STBM vesicles were isolated from the placentas and umbilical vein endothelial cell cultures were established from the umbilical cords. METHODS: Gene expression was screened by Affymetrix GeneChips and confirmed with real-time polymerase chain reaction or enzyme-linked immunosorbent assay. MAIN OUTCOME MEASURES: Fold changes in gene expression levels between treated and control cultures were calculated from the microarray results. RESULTS: Overall, the results do not show any great changes in gene expression in endothelial cells after STBM treatment (28 genes changed two-fold or more out of approximately 10,000 genes examined by microarray). In general, the changes observed are consistent with inhibition of proliferation of endothelial cells by exposure to STBM. The unfolded protein response in particular may be involved. CONCLUSIONS: STBM may influence endothelial cell function during pregnancy but STBM alone cannot account for the entire range of endothelial dysfunctions observed in pre-eclampsia.     

Neutrophils are stimulated by syncytiotrophoblast microvillous membranes to generate superoxide radicals in women with preeclampsia

OBJECTIVE: This study was undertaken to determine whether syncytiotrophoblast microvillous membranes (STBMs) stimulate maternal neutrophils to produce superoxide radicals in women with preeclampsia. STUDY DESIGN: Serum levels of tissue polypeptide antigen (TPA), which is a marker for STBM, were measured in 25 nulliparous women (10 with mild preeclampsia, 6 with severe preeclampsia, and 9 controls). Superoxide production by maternal neutrophils from cases and controls and by donor neutrophils cocultured with the STBMs from cases and controls was measured spectrophotometrically by reduction of ferricytochrome C. RESULTS: Maternal TPA levels were significantly greater among cases than controls (P=.005). Superoxide production by maternal neutrophils and donor neutrophils cultured with STBM from cases of preeclampsia was greater than controls (P values.006 and.019, respectively), and dose-response relationships were observed. Superoxide production by maternal leukocytes was correlated with superoxide induction by STBMs in culture (P=.007). CONCLUSION: STBMs in maternal blood induce neutrophils to generate superoxide radicals that may cause endothelial dysfunction in women with preeclampsia.     

The effect of alpha-fetoprotein on the growth of placental cells in vitro

Cell cultures of early mouse placentae were studied in the presence of mouse amniotic fluid or isolated alpha-fetoprotein (AFP). Four cell types were observed: polygonal cells, giant cells, small round cells and fibroblasts. The number of polygonal cells increased in placental cultures in the presence of amniotic fluid or AFP, suggesting de novo formation and proliferation in vitro. These preparations were found to stimulate the DNA synthesis in polygonal cells, as demonstrated by 3H-thymidine labelling experiments. Polygonal cells also show a significant increase in 3H-leucine incorporation, indicating more active protein synthesis under the effect of amniotic fluid or AFP. These data suggest that AFP may be one of the fetal factors promoting trophoblastic differentiation.     

Glucose transport and system A activity in syncytiotrophoblast microvillous and basal plasma membranes in intrauterine growth restriction

The mechanisms underlying the reduced fetal plasma concentrations of amino acids and glucose associated with intrauterine growth restriction (IUGR) remain to be fully established. The activity of the amino acid transporter system A has been shown to be reduced in the syncytiotrophoblast microvillous membrane (MVM) in IUGR, however the impact of these changes on transplacental transport is difficult to assess without information on system A activity in the basal plasma membrane (BM). In this study we measured system A activity and mediated D-glucose uptake using radiolabelled substrates and rapid filtration techniques, and glucose transporter isoform 1 (GLUT 1) protein expression using Western blots in MVM and BM isolated from human placentas. In term IUGR (n=11) MVM system A activity was unaltered compared to controls (n=9). In contrast, system A activity in MVM was reduced by 50 per cent (P< 0.05) in preterm IUGR (n=8, gestational age 28-36 weeks) as compared to controls (n=8, gestational age 28-35 weeks). BM system A activity was unaltered in both IUGR groups. Similarly, MVM and BM GLUT 1 expression and mediated D-glucose uptake was not affected by IUGR. In all preterm IUGR pregnancies signs of severe fetal compromise were present whereas term IUGR fetuses were less affected. These data support the view that MVM system A activity is related to the severity of compromise in IUGR. The markedly reduced system A activity in MVM in preterm IUGR together with the unaltered activity in BM is consistent with a decreased transplacental transport of neutral amino acids in this pregnancy complication. The hypoglycemia present in utero in some IUGR fetuses is not caused by a decreased glucose transport capacity across the syncytiotrophoblast plasma membranes.     

Isolation and purification of human placental plasma membranes from normal and pre-eclamptic pregnancies. a comparative study

Human placental syncytiotrophoblast is the main barrier for materno-fetal exchange. Analysis of transplacental transport involves the study of ion channels in both the maternal-facing microvillous membrane (MVM) and the fetal-facing basal membrane (BM). Difficulties in having access to intact placenta with conventional electrophysiological methods favour alternative methodologies, such as isolation and reconstitution of membranes in artificial lipid systems. Pre-eclampsia is a major health problem of human pregnancy. The search for altered physiological processes in pre-eclamptic placentae requires the investigation of events at both the microvillous and basal surfaces. The aim of this study was to obtain reliable syncytiotrophoblast plasma membranes from human normal (N) and pre-eclamptic (PE) pregnancies. We describe a protocol which allows for the simultaneous isolation of MVM and BM. The purity of the membranes isolated was evaluated using enzymatic assays, binding studies, Western blotting and immunohistochemistry. Enrichment of alkaline phosphatase activity for MVM was 17 to 21-fold, with 13-16 per cent protein recovery, for both N and PE. Enrichment of adenylate cyclase activity for BM was 9-fold for N, and enrichment of dihydroalprenolol binding to beta-adrenergic receptors was 12-fold for N and 6-fold for PE, with 14 per cent protein recovery for both N and PE. Cross contamination was low and mitochondrial membrane contamination was negligible. We conclude that MVM and BM isolated from placentae of pre-eclamptic women are similar in enrichment and purity to those of healthy women, thus allowing their use in comparative electrophysiological studies.     

Review: The effects of oxygen on normal and pre-eclamptic placental tissue--insights from metabolomics

Placental dysfunction is central to many complications of human pregnancy including pre-eclampsia (PE), intra-uterine growth restriction (IUGR) and stillbirth. The precise molecular pathophysiology of placental dysfunction in these conditions is not known, although oxidative and nitrative stresses have been implicated. Metabolites are low molecular weight chemicals which play an important role in biological function, primarily through metabolism and regulation of biological processes. The holistic study of metabolites, defined as metabolomics or metabolic profiling, has the objective to detect and identify all, or a large complement of all metabolites. Metabolomics is applied to discover new knowledge regarding biological processes and systems. We hypothesised that a metabolomic strategy could (1) provide a reproducible technique to investigate the intracellular metabolism of placental tissue and also metabolites consumed from or secreted in to the extracellular 'metabolic footprint' of in vitro culture systems (2) identify metabolic related differences in placental tissue culture systems subjected to perturbations in oxygen tension and from pregnancies complicated by PE. We review our early studies which demonstrate that a reproducible experimental protocol is required, including the preparation of culture medium and the site of the placenta applied for sampling tissue. We have detected changes in the intracellular metabolome and metabolic footprint of placental tissue in response to altered oxygen tension and PE. We have demonstrated that placental tissue from uncomplicated pregnancies cultured in 1% oxygen (hypoxia) had metabolic similarities to explants from PE pregnancies cultured at 6% oxygen (normoxia). Metabolites requiring further study include lipids, glutamate and glutamine and metabolites related to tryptophan, leukotriene and prostaglandin metabolism. Metabolomics has the potential to identify changes in clinical conditions, such as PE, that are associated with placental molecular pathophysiology.     

妊娠高血压综合征患者血浆对体外培养的内皮细胞合成血管活 …

目的 探讨妊娠高血压综合征（妊高征）患者的血浆对体外培养的内皮细胞合成血管活性物质的影响,和妊高征患者血浆中是否存在某些引起内皮细胞损伤的物质。方法 应用放射免疫法测定２０例先兆子痫患者（先兆子痫组）及１５例正常晚期妊娠妇女（正常晚孕组）产前血浆前列环素（ＰＧＩ２）代谢产物６－ｋｅｔｏ－ＰＧＦ１α、血栓素（ＴＸＡ２）代谢产物ＴＸＢ２的水平。在体外培养的爬刮动脉内皮细胞中分别加入先兆子痫患者和正常晚     

Uteroplacental blood flow and placental vascular endothelial growth factor in normotensive and pre-eclamptic pregnancy

Objective To determine whether placental vascular endothelial growth factor (VEGF) is increased in pre-eclampsia.
Design Prospective cohort study.
Setting Royal Prince Alfred Hospital, Sydney, Australia.
Sample Eleven normotensive women and eight women with pre-eclampsia matched for age and gestation.
Methods Uterine artery Doppler ultrasound flow velocity profiles were recorded in the third trimester and resistance index calculated as (V_s-V_d)/V_s (V_s= peak systolic flow velocity, V_d= end diastolic flow velocity). Placental tissue at delivery was examined for VEGF distribution with avidin-biotin-peroxidase immunohistochemistry.
Results Uterine resistance index [median (range)] was significantly increased in pre-eclamptic women (normotensive: 0.42 (0.36–0.51); pre-eclampsia: 0.59 (0.40–0.75);   P = 0.005  ). Notching of the uterine artery waveform, consistent with a high resistance circulation, was evident in early diastole in five women with pre-eclampsia but only one normotensive woman (   P = 0.013  ). Placental VEGF was increased in women with pre-eclampsia in the decidual trophoblast (normotensive: 34% (4–59) cells stained for VEGF; pre-eclampsia: 58% (15–95);   P = 0.033  ) and in the villous syncytiotrophoblast (normotensive: VEGF count 1.4 arbitrary units (1.1–2.1); pre-eclampsia: 1.8 arbitrary units (1.4–2.2);   P = 0.041  ). Analysis indicated that uterine artery resistance index was directly correlated with placental VEGF staining, mean arterial pressure and birthweight.
Conclusions Abnormal uterine artery Doppler ultrasound flow velocity profiles in pre-eclampsia indicate increased uteroplacental resistance. The associated increase in placental VEGF may represent a compensatory mechanism attempting to restore blood flow towards normal.     

Neutrophil activation induced by placental factors in normal and pre-eclamptic pregnancies in vitro.

Increased neutrophil activation has been demonstrated in women with pre-eclampsia. Activated neutrophils may play a significant role in the vascular endothelial pathophysiology in this disorder of pregnancy. How neutrophils become activated in pre-eclampsia is unknown. It has been proposed that activating factors could be produced and released by the placenta. To test if placental factors could stimulate neutrophil activation and what mechanism might be involved, neutrophils isolated from healthy female volunteers were exposed to the conditioned medium (CM) derived from either normal (Nor) or pre-eclamptic (PE) placental villous culture. Neutrophil-endothelial adhesion, neutrophil superoxide generation, elastase activity and integrin expression were measured. The data were analysed by ANOVA. A P value less than 0.05 was considered statistically significant. All values are expressed as a mean+/-s.e. We found: (1) neutrophil-endothelial adhesion was significantly increased in neutrophils exposed PE-CM than those exposed to Nor-CM and non-CM, P< 0.01; (2) both Nor-CM and PE-CM could stimulate neutrophils to generate more superoxide radicals; (3) there was no difference in elastase activity after neutrophil exposure to Nor-CM compared to PE-CM, P> 0.1; (4) significant changes in CD62L and CD11b expression were found in neutrophils exposed to PE-CM. We conclude that factors produced by the placenta can activate neutrophils by an increase in superoxide generation and modulation of adhesion molecule expression. Upregulation of surface adhesion molecule CD11 expression may be responsible for the increased neutrophil-endothelial adhesion induced by factors derived from pre-eclamptic placentae.     

Inhibin expression in normal and pre-eclamptic placental tissue.

Serum inhibin levels increase during normal pregnancy, but are significantly higher in patients with pre-eclampsia. The aim of this study was to demonstrate possible increased expression of inhibin within the placentas of women with pre-eclampsia compared with non-pre-eclamptic controls. Cellular expression of inhibin alpha and beta A subunits was studied using immunohistochemistry on formalin-fixed, paraffin-embedded placental sections from cases of pre-eclampsia (n = 23) and gestational age-matched non-pre-eclamptic controls (n = 16). Immunohistochemistry was performed using monoclonal antibodies against inhibin alpha and beta A subunits by the indirect immunoperoxidase technique. Intensity of staining was graded by a semiquantitative scoring method. Differences in distribution and intensity of staining between control and pre-eclamptic placentas were analyzed using a nonparametric Mann-Whitney U test. Staining for both inhibin alpha and beta A was predominantly confined to the cytoplasm of syncytiotrophoblast, with weak expression within intermediate trophoblast. The intensity of staining for inhibin alpha was significantly greater in the syncytiotrophoblast of pre-eclamptic patients (mean staining intensity controls = 0.97, disease = 1.87; p < 0.001). Inhibin beta A staining was generally stronger than for the alpha subunit, and was also significantly increased in pre-eclamptic patients compared with controls (mean controls = 1.72, disease 2.19; p < 0.05). This is the first evidence for increased placental inhibin presence in pre-eclampsia, suggesting increased inhibin production within the placenta, a finding that could account for increased serum inhibin levels in pre-eclampsia.     

Non-gastric H+/K+ ATPase is present in the microvillous membrane of the human placental syncytiotrophoblast

In humans, the non-gastric H(+)/K(+)ATPase (ATP1AL1) has previously been shown to be expressed in the epithelia of skin, kidney and colon. In this study we tested the hypothesis that the non-gastric H(+)/K(+)ATPase is localized to the syncytiotrophoblast, the transporting epithelium of the human placenta. Microvillous (MVM) and basal plasma membranes (BM) of the syncytiotrophoblast were isolated from term placenta and membrane proteins were separated using SDS-PAGE. The ATP1AL1 protein was identified as a 114 kD band in both MVM and BM by Western blot, however, the protein was more abundant in the MVM. Using immunocytochemistry H(+)/K(+)ATPase protein was localized in MVM but not BM. We constructed primers specific for ATP1AL1 and performed RT-PCR on RNA isolated from human placenta and human kidney. A product of the expected size could be detected in both tissues after 30 cycles of amplification. The sequence identity of this 517 nucleotide product was confirmed by sequencing and found to be identical to the human non-gastric H(+)/K(+)ATPase. The activity of this proton pump appears to be low in normal healthy placental at term, however, it is speculated that MVM non-gastric H(+)/K(+)ATPase may be important in pathological states. In conclusion, non-gastric H(+)/K(+)ATPase is present in the microvillous plasma membrane of the transporting epithelia of the human placenta.     

Plasma adiponectin concentrations and placental adiponectin expression in pre-eclamptic women

This study was conducted to compare maternal plasma adiponectin concentrations and adiponectin expression in term placentas between normotensive pregnant women and pre-eclamptic women. Plasma adiponectin concentrations were assessed by a sandwich enzyme-linked immunosorbent assay in 81 normotensive pregnant women, 27 pre-eclamptic women and 15 non-pregnant healthy women. The expression of adiponectin in the placentas was assessed by immunohistochemistry. Plasma adiponectin concentrations in normotensive pregnant women did not show a significant change during pregnancy and postpartum compared with non-pregnant women. However, plasma adiponectin concentrations in pre-eclamptic women were significantly (p < 0.05) lower than in non-pregnant and normotensive pregnant women. No immunoreactive adiponectin was detected in the term placentas of normotensive pregnant women, whereas a positive immunostaining for adiponectin was observed in endothelial cells of chorionic vessels in pre-eclamptic women. Our data suggest that decreased plasma adiponectin concentrations may contribute to the pathophysiology of pre-eclampsia and that adiponectin localized in chorionic vessels may play a role in the restoring of endothelial damage in the feto-maternal units of pre-eclampsia.     

In vitro effect of bioactive natriuretic peptides on perfusion pressure in placentas from normal and pre-eclamptic pregnancies

The number of placental vascular guanylate-coupled receptors, corresponding to bioactive natriuretic peptide receptors is greater in preeclampsia, but there are no clear data about atrial natriuretic peptide (ANP) concentration in preeclampsia. The influence of various doses of ANP and urodilatin (URO) on placental perfusion pressure in preeclampsia was investigated by perfusing 16 human placentas in vitro. The placental vessels were submaximally preconstricted by continuous infusion of N-ω-nitro-l-arginine (NOLA). Perfusion pressure was measured continuously. Over 180 min various doses of αANP or URO were administrated (25, 50, 100, 200 nmol/l, 2 min increments). The effects of pretreatment with the guanylate cyclase inhibitor, LY 83583 was also examined. We found that ANP and URO attenuated NOLA-induced vasoconstriction, that URO given in higher doses produced stronger vasocilation than ANP, and that the mean decrease of perfusion pressure was higher in preeclampsia. The possibility of a non cGMP-mediated pathway of ANP and URO action should be considered. Received: 11 August 1998 / Accepted: 15 January 1999     

Matrix metalloprotease-9, placental syncytiotrophoblast and the endothelial dysfunction of pre-eclampsia

The maternal syndrome of pre-eclampsia is caused by generalized maternal endothelial cell dysfunction, arising directly or indirectly from factors of placental origin. Syncytiotrophoblast membrane microvesicular particles are shed from the placental surface into maternal blood in increased amounts in pre-eclampsia and, in vitro, both inhibit endothelial cell proliferation and cause marked changes in the morphology of the cultured cell monolayers. Because there is evidence that proteolytic activation and degradation of the underlying matrix can cause the same morphological changes, we tested the hypothesis that proteases intrinsic to syncytiotrophoblast microvillous membranes (STBM) are the cause of the in vitro endothelial changes.Purified STBM were analysed by zymography and western blotting. Although we could confirm the presence of urokinase plasminogen activator (uPA) in STBM we could demonstrate no intrinsic activity presumably because of its association with the plasminogen activator inhibitor-2 (PAI-2) which is also a component of STBM. We detected gelatinase activity and showed that it was due to the matrix metalloproteinase-9 (MMP-9). Its presence was confirmed in this location by immunohistocytochemistry.Protease inhibitors caused a small reversal of the effects of STBM on morphology and no effect on inhibition of proliferation. We conclude that the effect of STBM on endothelial cells is unlikely to be caused by intrinsic proteases.