Monoclonal Antibody against Phosphatidylserine Inhibits In Vitro

8 downloads 0 Views 1MB Size Report
D11A4 (CL+/PS-) reacted only minimally and at the level of ... BA3B5C4 (CL+/PS+) and another monoclonal aPL anti- body ..... Thromb Res 1988; 51:267-281.
BIOLOGY OF REPRODUCTION 56, 50-58 (1997)

Monoclonal Antibody against Phosphatidylserine Inhibits In Vitro Human Trophoblastic Hormone Production and Invasion' Hiroshi Katsuragawa, 2 ' 3 Hideharu Kanzaki,4 Takuya Inoue, 3 Takeshi Hirano,3 Takahide Mori,3 and Neal S. Rote 5

Department of Gynecology and Obstetrics, Faculty of Medicine, Kyoto University, Sakyo-ku, Kyoto, 606-01 Japan Department of Obstetrics and Gynecology,4 School of Medicine, Kansai Medical University, Osaka, 570 Japan Departments of Microbiology and Immunology, and Obstetrics and Gynecology, Wright State University, School of Medicine, Dayton, Ohio 45435 of successful pregnancy also appear to have placental damage and are at risk for intrauterine growth retardation (IUGR), prematurity, and early and severe pregnancy-induced hypertension (PIH) [1, 2, 9-11]. Therefore, the vast majority of pregnancies in the presence of aPL antibodies result in miscarriage or are severely complicated. Using monoclonal aPL antibodies that differentiate between CL and PS, we have determined that PS-dependent antigens are expressed by the placenta and are reactive with aPL antibodies. The placental surface is covered with branching chorionic villi that are in turn covered by a layer of specialized cells, the trophoblastic layer, which includes an inner layer of mononuclear cytotrophoblast and an outer layer of a multinucleated syncytium (syncytiotrophoblast). As placental development proceeds, the syncytium expands and is renewed by an obligatory intercellular fusion process from the underlying cytotrophoblastic cells [12]. Most of the villous cytotrophoblastic layer becomes depleted by the third trimester. Monoclonal aPL antibody BA3B5C4 (CL+/ PS+) reacts strongly with the villous cytotrophoblast, whereas 3SB9b (CL-/PS+) reacts with the apical surface of the syncytiotrophoblastic layer [13-15]. Trophoblastic cells are also found in extravillous sites. Proliferating villous cytotrophoblasts penetrate the overlying syncytium at the tip of some chorionic villi to migrate into the maternal tissue [16]. Interstitial cytotrophoblasts invade further into the maternal tissue, some to the level of the maternal myometrium. Endovascular cytotrophoblasts, arising from the invading interstitial trophoblasts, extend into the maternal spiral arteries, displacing the endothelial cell lining of the maternal vessels. Both PS-reactive aPL antibodies react with extravillous cytotrophoblast and endovascular trophoblast [13-15]. The PS-dependent antigens are differentiation related. In a choriocarcinoma model of trophoblast differentiation, the BA3B5C4 (CL+/PS+) epitope is expressed in predifferentiation cells, and the 3SB9b (CL-/PS+) epitope is located on the surface of the postdifferentiation cells [17, 18]. Our observations provide direct evidence that trophoblasts externalize PS-dependent antigens during differentiation and that anti-phosphatidylserine (aPS) will react with normal syncytiotrophoblasts, potentially mediating direct pathological effects at the maternal/fetal interface. This hypothesis is supported by the observation that 3SB9b (CL-/ PS+) prevented intertrophoblastic fusion in an in vitro choriocarcinoma model and induced fetal and placental IUGR in a pregnant mouse model [19, 20]. In the presence of BA3B5C4 (CL+/PS+) and another monoclonal aPL antibody, Dl 1A4 (CL+/PS-), fusion proceeded normally and murine pregnancies were not affected. These data suggest that externalization of PS-dependent antigens occurs during

ABSTRACT Naturally occurring antiphospholipid (aPL) antibodies against cardiolipin (CL)- and phosphatidylserine (PS)-dependent antigens are associated with placental dysfunction and unsuccessful pregnancy. Murine monoclonal aPL antibodies react with placental trophoblast and may interfere with normal trophoblastic function. In this study, we evaluated the expression of phospholipid-dependent antigens during trophoblast differentiation and measured the effects of monoclonal aPL antibodies on two in vitro aspects of trophoblast differentiation: hormone production and invasion into filters coated with extracellular matrix. Murine monoclonal IgM aPL antibodies that differentiated between PS and CL were used: 3SB9b reacted only with PS (CL-/PS+), D11A4 reacted only with CL (CL+/PS-), and BA3B5C4 reacted with both CL and PS (CL+/PS+). Isolated trophoblasts were cultured for 4 days, and reactivity with monoclonal aPL antibodies was evaluated daily. BA3B5C4 (CL+/PS+) reacted strongly with most trophoblasts that were freshly isolated (Day 0) and through 2 days of culture, after which time the percentage of cells reactive with BA3B5C4 decreased steadily. 3SB9b (CL-/PS+) reactivity increased during incubation; about 8% of cells reacted initially, but after 1 day of incubation 100% reacted, and this percentage remained stable throughout the 4-day incubation. D11A4 (CL+/PS-) reacted only minimally and at the level of the negative control monoclonal antibody (mAb) with 1- and 2-day cultures. Both mAbs that reacted with PS-dependent antigens completely prevented invasion of matrigel-coated filters by isolated trophoblasts. These mAbs also inhibited trophoblastic hCG and human PL production by more than 45%. Thus, as trophoblasts undergo differentiation, they are reactive with mAbs against PS. These antibodies are inhibitory in vitro to trophoblastic hormone production and invasion. INTRODUCTION The antiphospholipid (aPL) antibody syndrome has been defined as the production of autoantibodies against negatively charged phospholipid-dependent antigens, especially cardiolipin (CL) and phosphatidylserine (PS), and has been clinically associated with thrombocytopenia, thrombosis, pregnancy loss, or a combination of these events [1-8]. Although the aPL antibody syndrome has been considered a disease of systemic thrombosis, the most common complication is disturbance of pregnancy. Patients with aPL antibodies have a 50-90% pregnancy loss rate; half of these pregnancies are lost in the first trimester [2]. The few cases Accepted August 13, 1996. Received May 3, 1996. 'This work was supported in part by Grants-in-Aid Scientific Research (No. 0530439, No. 05454448) from the Ministry of Education, Science and Culture of Japan, and National Institute of Health of grant HD 23697 (N.S.R.). 2Correspondence. FAX: 81-775-21-5414.

50

aPL ANTIBODY INHIBITS TROPHOBLAST FUNCTION trophoblast differentiation and intercellular fusion and that some of the aPL antibody-induced obstetrical complications, such as IUGR, may be mediated by aPL antibodymediated inhibition of the trophoblastic fusion process resulting in defective placentation. Trophoblast differentiation-related activities include hormone production and decidual invasion. In the current study we investigated the effects of monoclonal aPL antibodies on differentiation-related hCG and human (h)PL production and on an in vitro model of trophoblast invasiveness. MATERIALS AND METHODS Specimens First-trimester products of conception were obtained from 10 healthy women undergoing legal abortions at 612 wk of gestation. Normal term (36-41 wk gestation) placentas (n = 5) were obtained immediately from spontaneous vaginal delivery or from uncomplicated cesarean section. The gestational age was confirmed by ultrasound. Informed consent for the use of human tissues in this study was obtained from all of the patients. Isolation of Human Trophoblasts Trophoblasts were prepared using Percoll gradient separation [12]. Villous tissues were subjected to three enzymatic digestions with 0.125% trypsin (Sigma Chemical Company, St. Louis, MO) and 0.2 mg/ml DNase I (Sigma). The collected cells were applied to a 5-70% Percoll gradient (Pharmacia, Uppsala, Sweden). After centrifugation at 1200 x g, the middle band (density: 1.040-1.060 g/ml) was collected, and the cells were rinsed and resuspended in Dulbecco's Modified Eagle medium (DMEM) (Gibco, Grand Island, NY) at room temperature. We generally obtained 3-5 x 107 cells/30 g villous tissue. Any remaining leukocytes were removed by plating the cells (6-well plate; Coming Glass Works, Coming, NY) for 20 min and then aspirating the trophoblast-enriched supernatant. The cell viability was greater than 95% by trypan blue dye exclusion. To assess the purity of these trophoblasts, we used immunoperoxidase labeling with anti-vimentin antibody (specific for endothelial cells and fibroblasts, 1:120 dilution; Dako, Carpintera, CA), anti-CD 9 antibody (reactive with fibroblasts, 1:20 dilution; Dako), and anti-CD45 antibody (reactive with macrophages, 1:20 dilution; Dako). Anti-SPI antibody (directly used; Zymed, South San Francisco, CA) and anti-cytokeratin 4.62 antibody (1:40 dilution; Sigma) were used as trophoblast-specific controls. In some experiments, trophoblasts were evaluated with anti-[-hCG antibody (1:20 dilution; Dako) and anti-hPL antibody (1:200 dilution; UCB Bioproducts SA, L'Alleud, Belgium). All antibodies were diluted in PBS. The purified preparation contained approximately 90-95% cytotrophoblasts [21]. Monoclonal Antibodies (mAbs) against Phospholipids Production and serologic characterization of monoclonal aPL antibodies have been described elsewhere [14, 22]. Each mAb reacted in ELISAs against the negatively charged phospholipids, phosphatidic acid, phosphatidylinositol, and phosphatidylglycerol and did not react with phosphatidylethanolamine or phosphatidylcholine. Each mAb, however, reacted differently with CL and PS; 3SB9b reacted only with PS, DllA4 reacted only with CL (CL+/ PS-), and BA3B5C4 reacted with both CL and PS (CL+/

51

PS+). The concentrations of IgM in undiluted tissue culture supernatant were 3SB9b, 80 pxg/ml; BA3B5C4, 15 ,ag/ml; and D11A4, 30 pxg/ml. Anti-ds-DNA antibody (1:20 dilution, Ch26-1352; ATCC, Rockville, MD) was used as a monoclonal IgM negative control. Immunohistochemistry Trophoblastic cells from term placentas were cultured in DMEM with 20% fetal calf serum (FCS; Dainippon Pharmaceutical Co., Osaka, Japan), 100 IU/ml penicillin, and 100 g/ml streptomycin at 37°C under a 5% CO2, 100% humidified atmosphere. Human cytotrophoblast cells were plated into 10 chambered glass microscope slides (LABTEK, Naperville, IL) at a density of 1 x 104 cells per well. The slides had been coated with fibronectin (Sigma) (20 pxg/ml dilution with DMEM). Every 24 h, the standard medium was changed. Cells were cultured up to 4 days. Cultured cells were either observed directly with phase-contrast microscopy or first rinsed with PBS and then fixed for 10 min at room temperature in 5% formalin or for 20 min at -20°C in acetone. The slides were stored at 4C until they were stained with immunoperoxidase. The supernatant was aspirated, and cytospin smears were prepared using a cytocentrifuge (Shandon Scientific Ltd., Cheshire, England). Immunoperoxidase Staining and Morphologic Assessment Immunoperoxidase staining was performed by standard procedures using a commercial kit (Histostain-SP kit; Zymed). Briefly, endogenous peroxide activity was blocked by immersion in a solution containing 9 parts methanol:l part hydrogen peroxidase. Nonspecific binding was blocked by preincubation in 10% nonimmune rabbit serum. Fifty microliters of test antibody was added to the wells after removal of the blocking solution; the wells were incubated for 1 h at room temperature in a moist chamber and rinsed for 2 min in flowing PBS. Then 50 1l of biotinylated polyclonal anti-mouse IgM was placed into each well. The wells were rinsed with PBS and then incubated for 5 min with the avidin-peroxidase conjugate. The wells were finally incubated for 5-15 min with the kit's substrate reagent. The wells were counterstained with hematoxylin and evaluated. The percentage of positive cells is defined as the number of cells that were stained divided by the total number of cells counted within 15 fields of view (total cells per case, 37-68; average, 55 + 9). Any discernible staining was considered positive. The intensity of the positive cells was defined based on the extent of color development in cells that were stained: total; strong; moderate and partial; weak. The results represent the mean ± SEM of quadruplicate plates from five different experiments. Hormone Production First-trimester human cytotrophoblast cells were cultured in triplicate in 1 ml of DMEM supplemented with 2% BSA (Dainippon Pharmaceutical), 100 IU/ml penicillin, and 100 pxg/ml streptomycin at 37°C under a 5% CO 2, 100% humidified atmosphere. Trophoblast cells were inoculated into each well of a 24-well plate (Corning Glass Works) at a density of 5 x 105 cells/mi. After 12 h, aPL antibodies and control antibody (ds-DNA) were added to the culture medium, and the cells were cultured for up to 72 h. The culture medium was changed every 24 h, with continuous supplementation with aPL antibodies or control

52

KATSURAGAWA ET AL.

antibody. Upon completion of the culture, the culture media were collected, centrifuged, and frozen at -20°C until the hormone assays were performed. The hCG and hPL content of the culture media was determined in duplicate by enzyme immunoassay (EIA) using commercial kits (Mochida Pharmaceutical Co., Tokyo, Japan). The assay method was a tube-supported, sandwichtype EIA using mAbs, and is characterized by a one-step assay without extraction [21, 23]. The detection thresholds for the hCG and hPL assays were 0.2 mIU (Second WHO International Standard)/ml and 2.5 ng/ml, respectively.

aPLs reactivity against human cultured trophoblast

Invasion Assay To quantify invasion, first-trimester cytotrophoblasts were plated on matrigel invasion chamber inserts (9 mm; Becton Dickinson Labware, Bedford, MA) containing polyethylene terephthalate membranes with 8-lm-diameter pores and 24-well tissue culture plates. The upper surface of the filters was coated with 100 txg/cm2 matrigel. Before use, the inserts were rehydrated for 1.5-2.0 h with 0.5 ml of warm DMEM containing 0.1% BSA. Cytotrophoblasts (2 X 105) were plated on each filter in 250 pl1of DMEM with 2% BSA, 100 IU/ml penicillin, and 100 pug/ml streptomycin at 37°C under a 5% CO2, 100% humidified atmosphere; after 6 h, aPL antibodies and control antibody (ds-DNA) were added. The assay was performed in at least triplicate chambers. Medium added to both the top and the bottom of the culture contained the same concentration of the aPL antibodies or control antibody. The cultures were incubated for 72 h. The culture medium was changed every 24 h, with continuous supplementation with aPL antibodies or control antibody. After 72-h culture, the upper surface of the chamber insert membrane was "scrubbed" three times with a cotton swab and DMEM; the cells on the lower surface of the scrubbed membranes were fixed with 5% formalin for 10 min at room temperature, and the cells that traversed the membrane were counted. 3-(4, 5-Dimetythiazol-2-yl)2,5-Diphenyltetrazolium Bromide (MTT) Reduction Assay To examine the viability of trophoblasts after treatment with aPL antibodies, the MTT reduction assay was performed as described previously [24] with minor modification. A tetrazolium salt, MTT, is metabolized to a colored formazan salt by mitochondrial enzyme activity in viable cells. Fifty thousand viable trophoblasts in DMEM supplemented with 20% FCS in the presence or absence of aPL antibodies were inoculated in 96-well flat-bottomed tissue culture plates (Corning Glass Works). Cells were cultured for 3 days, and the culture medium was changed every 24 h. After 3 days of culture, the culture medium was removed, and 50 .I1 DMEM containing 500 pIg/ml MTT (Sigma) was added to each well. The cells were incubated for 3 h. The culture medium was then aspirated, and the cells were lysed with 100 p.1 100% ethanol. MTT reduction was determined by reading the plates at an optical density of 570/630 nm on an automated ELISA plate reader (Molecular Devices, Menlo Park, CA). Statistical Analysis The data were expressed as means + SEM and examined using a one-way analysis of variance, followed by Scheffe's F test for multiple comparisons. Pair-wise comparisons were made using the unpaired t-test.

FIG. 1. Reactivity of aPL antibodies against human cultured term trophoblast. Percentages of mononuclear trophoblasts staining positive with monoclonal aPL antibodies BA3B5C4 (CL+/PS+) (open circles), 3SB9b (CL-/PS+) (solid circles), D11A4 (CL+/PS-) (open squares), with positive control antibody against 3-hCG (triangles), and with negative control antibody against ds-DNA (solid squares).

RESULTS Immunohistochemical Analysis of aPL Antibodies Binding to Isolated Trophoblast Our previous results with use of placental sections and choriocarcinoma models of trophoblast differentiation indicated that monoclonal aPL antibody BA3B5C4 (CL+/ PS+) reacted strongly with the villous cytotrophoblast whereas 3SB9b (CL-/PS+) reacted with the apical surface of the syncytiotrophoblast; they also indicated that predifferentiation choriocarcinoma cells reacted with BA3B5C4 (CL+/PS+), but not with 3SB9b (CL-/PS+), whereas the pattern of reactivity was reversed (BA3B5C4-/3SB9b+) on postdifferentiation cells. In order to determine whether a similar change in reactivity occurred in differentiating trophoblasts, we reacted aPL antibodies with daily cytospin preparations from term trophoblast cultures that had been incubated for 0-4 days (Figs. 1 and 2). A negative monoclonal IgM control, anti-ds-DNA (1:20 dilution), was used, and it showed little reactivity. The only reactivity occurred after 1, 2, 3, or 4 days of culture, at which time ds-DNA 5%, and 19 4% 5%, 20 reacted with 23 + 9%, 24 of the cells, respectively. A positive monoclonal IgG control, anti-3-hCG (1:20 dilution), was used as a trophoblast differentiation marker. Initially (0 time) a small percentage (7 4%) of the cells were positive for hCG production

FIG. 2. Immunoperoxidase staining of cytospin smear preparations of trophoblasts with monoclonal aPL antibodies. Immunoperoxidase staining of term cytotrophoblast (A) with monoclonal aPL antibody BA3B5C4 (CL+/PS+) at 0 time cytospin smears. Many cells stained (arrows). Immunoperoxidase staining of term trophoblast with monoclonal aPL antibody BA3B5C4 (CL+/PS+) at 1 day (B) and 4 days (C) of culture. Some cytotrophoblast-like cells stained (arrows), but syncytiotrophoblast-like cells did not stain (arrowheads). Immunoperoxidase staining of term cytotrophoblast (D) with monoclonal aPL antibody 3SB9b (CL-/PS+) at 0 time cytospin smears. Some cells stained (arrows). Immunoperoxidase staining of term trophoblast with monoclonal aPL antibody 3SB9b (CL-/ PS+) at 1 day (E) and 4 days (F) of culture. Immunoperoxidase staining of term cytotrophoblast with monoclonal aPL antibody D11 A4 (CL+/PS-) (G) at 0 time cytospin smear. Immunoperoxidase staining of term trophoblast with monoclonal aPL antibody D11A4 (CL+/PS-) at Day 1 (H) and Day 4 (1)of culture. Some cells weakly stained (arrows). Bars = 25 Rm.

aPL ANTIBODY INHIBITS TROPHOBLAST FUNCTION

53

54

KATSURAGAWA ET AL. Day 4. 3SB9b (CL-/PS+) stained relatively few cells (8 ± 3%) at the 0 time point. With incubation, however, the percentage rose to 100% after 1 day and remained at more than 90% throughout the incubation period. D 11A4 (CL+/ PS-) reacted minimally or, more commonly, not at all with mononuclear trophoblasts. The only reactivity occurred after 1 or 2 days of culture, at which time DllA4 reacted 9% of the cells, respectively. By with 29 ± 10% and 38 3 days of culture, the trophoblasts were completely unreactive. Although immunoperoxidase techniques are, at best, only semi-quantitative, the amount of color development suggested that the rank of labeling intensity was 3SB9b > BA3B5C4 > DllA4 (Fig. 2). BA3B5C4 (CL+/PS+) reacted strongly with cells cultured for 1 day, and the intensity of the positive cells lessened with continued culturing (Fig. 2, A-C). The staining with 3SB9b (CL-/PS+) remained intense on all samples from 1 through 4 days of culture (Fig. 2, D-F). Compared to staining with the other aPL antibodies, staining with DllA4 (CL+/PS-) was always considerably weaker (Fig. 2, G-I). All cultures contained a certain percentage of apparently multinucleated cells (Fig. 3). Throughout the 4-day culturing procedure, these syncytia did not react with BA3B5C4 (CL+/PS+) (Fig. 3A) or DI lA4 (CL+/PS-) (not shown), but reacted intensely with 3SB9b (CL-/PS+) (Fig. 3B) or with antibody against B-hCG (Fig. 3C). These observations confirm that the expression of BA3B5C4- and 3SB9b-reactive antigens is related to differentiation in normal term trophoblast, as well as in choriocarcinoma models. Effects of aPL Antibodies on hCG and hPL Secretion

FIG. 3. Monoclonal aPL antibody immunoperoxidase staining of multinucleated term trophoblasts within cytospin preparations from cultures that were incubated for 4 days. Preparations were stained with the monoclonal aPL antibodies (A) BA3B5C4 (CL+/PS+) and (B) 3SB9b (CL-/PS+) and with (C) control anti-B-hCG. Syncytiotrophoblast-like cells strongly stained (arrowheads, C). Bars = 25 plm.

(Fig. 1). After 1 day of culture, 69 + 15% of trophoblasts showed positive reactions, and the percentage increased during 2-4 days of culture. Most of the mononuclear cells reacted with BA3B5C4 (CL+/PS+) at the initial 0 time point (Fig. 1). After 1 or 2 days of culture, all the mononuclear trophoblasts were reactive. After 2 days, however, the percentage of cells 11% at Day 3 and 20 9% at staining dropped to 37

The adequate production of trophoblastic hormones is necessary for a successful pregnancy. Isolated first-trimester trophoblasts were cultured in the presence of mAbs, and the levels of hCG and hPL production were measured. The levels of hormone secreted into the culture medium by trophoblasts in the presence of a negative control mAb, antids-DNA, were defined as the basal secretions levels of each hormone. The mean + SEM of the basal secretion levels for Days 1, 2, and 3 of trophoblast culture were 431.4 + 31.2, and 577.5 23.7 mIU/ml of hCG/ml 18.0, 531.3 5.4 ng of hPL/ and 95.8 + 2.4, 100.7 + 6.4, and 114.9 ml, respectively. By Day 3 of culture, BA3B5C4 (CL+/PS+) and 3SB9b (CL-/PS+) had significantly inhibited basal hCG (Fig. 4A) and hPL (Fig. 4B) secretion (p < 0.0005 and p < 0.0003, respectively). BA3B5C4 had reduced hCG and hPL secretion to 55% and 42% of the control levels, and 3SB9b had reduced hCG and hPL secretion to 40% and 34% of the control levels, respectively. In contrast, DllA4 did not inhibit basal hCG or hPL secretion. These experiments were performed using different first-trimester chorionic villi obtained from 8 women, with identical results. Effects of aPL Antibodies on In Vitro Trophoblast Invasion Trophoblast differentiation is characterized by the development of invasive extravillous trophoblasts, which migrate to the level of the maternal myometrium during successful, uncomplicated pregnancy. We investigated whether monoclonal aPL antibodies would affect the cytotrophoblast invasion process in an in vitro model system using matrigel-coated filters (Fig. 5). First-trimester trophoblasts exposed to the PS-reactive aPL antibodies, BA3B5C4 (CL+/PS+) (not shown) and 3SB9b (CL-/PS+) (Fig. 5A),

aPL ANTIBODY INHIBITS TROPHOBLAST FUNCTION

55

FIG. 4. Inhibitory effects of monoclonal aPL antibodies 3SB9b (CL-/ PS+) (squares), BA3B5C4 (CL+/PS+) (solid circles), and D11A4 (CL+/ PS-) (open circles) on (A) hCG and (B)hPL production by first-trimester trophoblasts. Data are expressed as a percentage of the control hormone production, without the addition of aPL antibodies. Values are the mean + SEM of 8 separate placental experiments. * p < 0.0005 (A) or p < 0.0003 (B)compared to the control.

were completely blocked from traversing the filters. No cells were detectable on the filter after 72 h of culture with either aPL antibody. In cultures exposed to Dl 1A4 (CL+/ PS-) (Fig. 5B) or negative control (ds-DNA) (Fig. 5C), 33.5 4.0 and 35.0 + 8.5, respectively, cells per membrane were observed to have traversed the membrane. These experiments were performed using different first-trimester chorionic villi obtained from 5 women with identical results.

FIG. 5. Effects of monoclonal aPL antibodies (A)3SB9b (CL-/PS+), (B) D11A4 (CL+/PS-), and (C) ds-DNA on in vitro first-trimester cytotrophoblast invasion of matrigel filters. Immunoperoxidase staining of extravillous trophoblast with polyclonal anti-hPL antibody at 72 h of culture. The extravillous trophoblastic cells stained. Bars = 25 Im.

Assessment of Cellular Viability In order to determine whether monoclonal aPL antibodies caused decreased hormone release and decreased invasion by direct damage to the first-trimester trophoblast, we measured the ability of these cells to reduce MTT after culture with aPL antibodies for 3 days as compared to cells in cultures supplemented with control antibody. The means of percentages positive ( SEM) were 3SB9b, 94.2 1.6%; BA3B5C4, 95.9 2.5%; DllA4, 97.7 1.4%; and anti-ds-DNA, 98.1 + 2.4%. There was no significant difference within the antibodies tested.

DISCUSSION The presence of naturally occurring aPL antibodies is associated strongly with several obstetrical complications: recurrent miscarriage, IUGR, and PIH. The most commonly proposed mechanism by which these events occur is aPL antibody-mediated thrombosis in maternal spiral arteries, resulting from damage to endothelial cells or platelets, or both [25-32]. Although decidual thrombosis is observed commonly in women with the aPL antibody syndrome, it is not specific for aPL antibodies and may be a manifes-

56

KATSURAGAWA ET AL.

tation of placental damage, regardless of the cause [33-38]. Decidual thrombosis may contribute to aPL antibody-induced placental damage, but other mechanisms must also be considered to explain fully the effects of aPL antibodies. During pregnancy the predominant cell type in contact with maternal circulating aPL antibodies is the trophoblast. Because of the predominance of obstetrical problems in patients with aPL antibodies, it is reasonable that the trophoblast may be a prime target. In order to explore the role of phospholipids in placental trophoblast differentiation and the effects of aPL antibodies on that process, we generated three monoclonal aPL antibodies that differentiate between PS- and CL-dependent antigens in routine diagnostic ELISAs; BA3B5C4 reacted with both CL and PS (CL+/PS+), Dl 1A4 reacted with CL (CL+/PS-), and 3SB9b reacted with PS (CL-/PS+). Our results have suggested that throughout pregnancy the trophoblastic cells express multiple antigenic forms of PS-dependent antigens [13-15, 39]. Monoclonal aPL antibody BA3B5C4 (CL+/PS+) reacted strongly with the villous cytotrophoblast, whereas 3SB9b (CL-/PS+) reacted with the apical surface of the syncytiotrophoblastic layer [13-15]. Both aPL antibodies reacted with extravillous cytotrophoblast. Only monoclonal aPL antibodies with specificity against PS-dependent antigens react with trophoblast, and at least two different PS-dependent antigens are expressed in the placenta. Human aPL antibodies have similar anti-trophoblastic reactivity [15, 39]. We tested four aPL antibody-positive sera, and each contained antibody reactive with the trophoblastic layer. Some of the sera reacted with a pattern similar to that of 3SB9b and others reacted similarly to BA3B5C4. PS vesicles absorbed much of the trophoblast reactivity, and the sera and monoclonal aPS antibodies reciprocally blocked the binding of each other to placenta. The expression of the BA3BS5C4-reactive antigen by prefusion villous cytotrophoblast and of the 3SB9b-reactive antigen by the post-fusion syncytiotrophoblast suggests that these antigens are differentiation related in the villous trophoblast. We investigated that possibility using choriocarcinoma cell lines, malignant trophoblasts that provide models for trophoblast differentiation [17]. BeWo choriocarcinoma cells grow in relatively undifferentiated form unless treated with forskolin or certain cytokines ([40]; unpublished results). Using this model, we have determined that the BA3BS5C4-reactive antigen is expressed in the cytoplasm of predifferentiation cells and that the 3SB9B-reactive antigen is located on the surface of the postdifferentiation cells [17, 18]. The differential reactivity of 3SB9b and BA3B5C4 suggests that the antigenic conformation involving PS on the cytotrophoblast is modulated into a different conformation concurrent with fusion into the syncytium. The relationship of PS expression with intertrophoblastic fusion has been studied using an in vitro model, forskolininduced fusion of the choriocarcinoma line JAR [41]. After treatment with forskolin, 80% of the nuclei were found in multinucleated cells. If, however, the assay was performed in the presence of 3SB9b, intertrophoblastic fusion was completely prevented. Fusion proceeded normally in the presence of BA3B5C4 and Dl1A4. Our observations provide direct evidence that aPS will react with normal syncytiotrophoblast and may directly mediate pathological effects by preventing intertrophoblast fusion. The success of pregnancy, however, is also dependent on factors other than syncytialization. Trophoblast differentiation can be measured by the production of hormones

such as hCG and hPL. Recent observations have suggested that syncytium formation and hormonal differentiation are separate, but parallel, events [42]. Our data demonstrate that aPL antibodies against PS-dependent antigens will prevent the secretion of hCG and hPL. In terms of DllA4 (CL-dependent antigen) and ds-DNA (control antibody), there are discrepancies between the immunohistochemical reactivity and the bioreactivity of suppression of hormone production on Days 1 and 2, although the exact mechanism of these phenomena remains to be discussed. The process by which this occurs in yet unknown. Hormone production appears responsive to signal transduction through membrane-bound phospholipase C and protein kinase C [43, 44]. Sera containing aPL antibodies can effectively block the induction of hCG production by exogenous phospholipase C or by GnRH [45, 46]. The activity of protein kinase C is dependent on membrane phospholipid, particularly PS [47]. It is possible, therefore, that aPL antibodies against PS can interfere with signal transduction in trophoblast and prevent the induction of hormone production by the syncytiotrophoblast. Successful pregnancy is also dependent on extravillous trophoblastic invasion of the decidua. Proliferating villous cytotrophoblasts penetrate the overlying syncytium at the tip of some chorionic villi to form anchoring cell columns connected to the decidua [16]. Trophoblasts migrate from these regions into the maternal tissue, some to the level of the maternal myometrium. Endovascular cytotrophoblasts, arising from the invading trophoblasts, extend into the maternal spiral arteries, displacing the endothelial cell lining of the maternal vessels. In pregnancies complicated by PIH, invasion appears to be incomplete [48]. Our data demonstrate, for the first time, that aPS will prevent trophoblastic invasion in an in vitro model. Again, the mechanism has yet to be clarified. In vitro invasion is dependent upon several factors, including adherence to the extracellular matrix, response to external cytokine signals, expression and alteration of adhesion proteins, and production and secretion of proteases [49]. The effects of aPL antibodies on the individual steps in this process have not yet been studied. How does aPS gain access to membrane PS? PS is a predominant phospholipid on the inner side of plasma membranes and is therefore not accessible to circulating aPL antibodies. Because of the normal inaccessibility of PS, the key to understanding the pathophysiology associated with aPL antibodies lies in understanding the normal physiological conditions under which antigenic phospholipid conformations are externalized. During myoblast fusion to form myotubules, PS is externalized [50, 51]. PS is also externalized on the surface of platelets during activation and fusion between platelet granules and the plasma membrane [21, 52-54]. Throughout normal placental development, the villous trophoblasts appear to externalize PS during a constant and obligatory intercellular fusion process, so that the presence of aPS will block intertrophoblastic fusion. Extravillous trophoblast also seems to externalize PS. Although these cells occasionally undergo intercellular fusion to form decidual giant cells, they generally do not undergo this process [55]. There appears to be a major difference, however, between the expression of PS-dependent antigens in villous and extravillous trophoblast. Villous trophoblast will undergo a transition from being "BA3B5C4 reactive/3SB9b unreactive" to being "BA3B5C4 unreactive/3SB9b reactive." An intermediate form may express both PS-dependent antigens simultaneously. The extravil-

aPL ANTIBODY INHIBITS TROPHOBLAST FUNCTION

lous trophoblast, however, appears to express both antigens throughout its differentiation; trophoblast throughout the decidua, as well as endovascular trophoblast, expresss both markers simultaneously. Thus, the trophoblast externalizes PS during intercellular fusion and differentiation. PS-dependent antigens are expressed in at least two forms that can be differentiated by mAbs. The inhibitory effects of monoclonal aPL antibodies on trophoblast intercellular fusion, hormone production, and invasion suggest that many of the obstetrical complications observed in the aPL antibody syndrome may be due to aPL antibody-induced trophoblast dysfunction.

19.

20.

21.

22.

ACKNOWLEDGMENT We thank Ms. Reiko Sakata for her excellent technical assistance.

REFERENCES 1.

2.

3.

4. 5.

6. 7. 8. 9.

10. 11. 12.

13.

14. 15. 16. 17. 18.

Branch DW, Scott JR, Kochenour NK, Hershgold E. Obstetric complications associated with the lupus anticoagulant. N Engl J Med 1985; 313:1322-1326. Branch DW, Scott JR. Clinical implications of anti-phospholipid antibodies: the Utah experience. In: Harris EN, Exner T, Hughes GRV, Asherson RA (eds.), Phospholipid-Binding Antibodies. Boca Raton, FL: CRC Press; 1991: 335-346. Rote NS. Transplacental alteration of fetal coagulation: antiphospholipid and antiplatelet antibodies. In: Bellisario R, Mizejewski GJ (eds.), Transplacental Disorders: Perinatal Detection, Treatment, and Management (including Pediatric AIDS). New York: Alan R. Liss, Inc.; 1990: 95-120. Rote NS. Antiphospholipid antibodies and disorders of pregnancy. J Clin Immunoassay 1990; 13:34-42. Rote NS, Ng AK, Ault KA, Carmody M, Dostal-Johnson DA, Mitchell J, Ng S, Roberts R, Siekman R. Antiphospholipid antibodies and recurrent pregnancy loss: analysis of monoclonal and polyclonal antiphosphatidylserine antibodies. In: Mettler L, Billington WD (eds.), Reproductive Immunology, 1989. Amsterdam: Elsevier Science Publishers; 1990: 217-226. Rote NS, Walter A, Lyden TW. Antiphospholipid antibodies: lobsters or red herrings? Am J Reprod Immunol 1992; 28:31-37. Scott JR, Rote NS, Branch DW. Immunologic aspects of recurrent abortion and fetal death. Obstet Gynecol 1987; 70:645-656. Harris EN. Antiphospholipid antibodies. Br J Haematol 1990; 74:19. Branch DW, Andres R, Digre KB, Rote NS, Scott JR. The association of antiphospholipid antibodies with severe preeclampsia. Obstet Gynecol 1989; 73:541-545. Francois A, Freund M, Daffos F, Remy P, Aiach M, Jacquot C. Repeated fetal losses and the lupus anticoagulant. Ann Intern Med 1988; 109:993-994. Kochenour NK, Branch DW, Rote NS, Scott JR. A new postpartum syndrome associated with antiphospholipid antibodies. Obstet Gynecol 1987; 69:460-468. Kliman HJ, Feinman MA, Strauss III JE Differentiation of human cytotrophoblasts into syncytiotrophoblasts in culture. Troph Res 1987; 2:407-421. Katsuragawa H, Rote NS, Inoue T, Narukawa S, Kanzaki H, Mori T Monoclonal antiphosphatidylserine antibody reactivity against human first-trimester placental trophoblasts. Am J Obstet Gynecol 1995; 172: 1592-1597. Lyden TW, Vogt E, Ng AK, Johnson PM, Rote NS. Monoclonal antiphospholipid antibody reactivity against human placental trophoblast. J Reprod Immunol 1992; 22:1-14. Rote NS, Lyden TW, Vogt E, Ng AK. Antiphospholipid antibodies and placental development. In: Hunt JS (ed.), Immunobiology of Reproduction. New York: Springer-Verlag; 1994: 285-302. Boyd JD, Hamilton WJ. The Human Placenta. Cambridge: Heffer and Sons, Ltd.; 1970. Lyden TW, Ng AK, Rote NS. Modulation of phosphatidylserine epitope expression on BeWo cells during forskolin treatment. Placenta 1993; 14:177-186. Rote NS, Chang J, Katsuragawa H, Ng AK, Lyden TW, Mori T. Expression of phosphatidylserine-dependent antigens on the surface of

23.

24.

25. 26.

27. 28.

29.

30.

31.

32.

57

differentiating BeWo human choriocarcinoma cells. Am J Reprod Immunol 1995; 33:114-121. Rote NS, Ng AK. Anti-phospholipid antibodies and the placenta. In: Kurpisz M, Fernandez Z (eds.), Immunology of Human Reproduction. Oxford: ios Scientific Publishers; 1995: 377-400. Vogt E, Ng AK, Rote NS. A model for the antiphospholipid antibody syndrome: monoclonal antiphosphatidylserine antibody induces intrauterine growth retardation in mice. Am J Obstet Gynecol 1996; 174: 700-707. Katsuragawa H, Kanzaki H, Inoue T, Hirano T, Narukawa S, Watanabe H, Mori T. Endometrial stromal cell decidualization inhibits human chorionic gonadotropin and human placental lactogen secretion by co-cultured trophoblasts. Hum Reprod 1995; 10:3028-3034. Rote NS, Ng AK, Dostal-Johnson DA, Nicholson SL, Siekman R. Immunologic detection of phosphatidylserine externalization during thrombin-induced platelet activation. Clin Immunol Immunopathol 1993; 66:193-200. Inaji H, Yaoi E, Maeura Y, Matsuura N, Tominaga S, Koyama H, Takatsuka Y, Mori T. Carcinoembryonic antigen estimation in nipple discharge as an adjunctive tool in the diagnosis of early breast cancer. Cancer 1987; 60:3008-3013. Mosmann T. Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays. J Immunol Methods 1983; 65:55-63. Carreras LO, Vermylen JG. "Lupus" anticoagulant and thrombosispossible role of inhibition of prostacyclin formation. Thromb Haemostasis 1982; 48:38-40. Marchesi D, Parbtani A, Frampton G, Livio M, Remuzzi G, Cameron JS. Thrombotic tendency in systemic lupus erythematosus. Lancet 1981; 1:719. Ros JO, Tarres MV, Baucells MV, Maired JJ, Solano JT. Prednisone and maternal lupus anticoagulant. Lancet 1983; 2:576. Walker TS, Triplett DA, Javed N, Musgrave K. Evaluation of lupus anticoagulants: antiphospholipid antibodies, endothelium associated immunoglobulin, endothelial prostacyclin secretion, and antigenic protein S levels. Thromb Res 1988; 51:267-281. Cariou R, Tobelem G, Soria C, Caen J. Inhibition of protein C activation by endothelial cells in the presence of lupus anticoagulant. N Engl J Med 1986; 314:1193-1194. Freyssinet JM, Wiesel ML, Gauchy J, Boneu B, Cazenave JP An IgM lupus anticoagulant that neutralizes the enhancing effect of phospholipid on purified endothelial thrombomodulin activity-a mechanism for thrombosis. Thromb Haemostasis 1986; 55:309-313. Cosgriff TM, Martin BA. Low functional and high antigenic antithrombin III level in a patient with the lupus anticoagulant and recurrent thrombosis. Arthritis Rheum 1981; 24:94-96. Tannenbaum SH, Finko R, Cines DB. Antibody and immune complexes induce tissue factor production by endothelial cells. J Immunol

1986; 137:1532-1537. 33. DeWolf F, Carreras LO, Moerman P Vermylen J, Van Assche A, Renaer M. Decidual vasculopathy and extensive placental infraction in a patient with repeated thromboembolic accidents, recurrent fetal loss, and a lupus anticoagulant. Am J Obstet Gynecol 1982; 142:829-834. 34. Out HJ, Kooijman CD, Bruinse HW, Derksen RHWM. Histopathological findings in placentae from patients with intra-uterine fetal death and anti-phospholipid antibodies. Eur J Obstet Gynecol Reprod Biol 1991; 41:179-186. 35. Lochshin MD, Druzin ML, Goei S, Qamar T, Magid MS, Jovanovic L, Ferenc M. Antibody to cardiolipin as a predictor of fetal distress or death in pregnant patients with systemic lupus erythematosus. N Engl J Med 1985; 313:152-156. 36. AlthabeO, Labarrere C, Telenta M. Maternal vascular lesions in placentae of small-for-gestational-age infants. Placenta 1985; 6:265. 37. Sheppard BL, Bonnar J. An ultrastructural study of utero-placental spiral arteries in hypertensive and normotensive pregnancy and fetal growth retardation. Br J Obstet Gynecol 1981; 88:695. 38. Abramowsky CR, Vegas ME, Swinehart G, Gyves MT. Decidual vasculopathy of the placenta in lupus erythematosus. N Engl J Med 1980; 303:668-672. 39. Rote NS, Lyden TW, Vogt E, Adler RR, Chang J, Katsuragawa H, Lin L, Ng AK, Nicholson S, Patel PN, Shroyer L, Ponder D. Expression of phosphatidylserine epitopes during trophoblast membrane fusion: an alternative hypothesis for antiphospholipid antibody associated pregnancy disorders. In: Dondero E Johnson PM (eds.), Reproductive Immunology. New York: Serono Symposia Publications, Raven Press; 1993: 97:281-284. 40. Taylor RN, Newman ED, Chen S. Forskolin and methotrexate induce

58

41.

42.

43.

44.

45.

46.

47.

KATSURAGAWA ET AL. an intermediate trophoblast phenotype in cultured human choriocarcinoma cells. Am J Obstet Gynecol 1991; 164:204-210. Adler RR, Ng AK, Rote NS. Monoclonal antiphosphatidylserine antibody inhibits intercellular fusion of the choriocarcinoma line, JAR. Biol Reprod 1995; 53:905-910. Rachmilewitz J, Gonik B, Goshen R, Ariel I, Schneider T, Eldar-Geva T, de Groot N, Hochberg A. Intermediate cells during cytotrophoblast differentiation in vitro. Cell Growth Differ 1993; 4:395-402. Iwashita M, Watanabe M, Setoyama T, Mimuro T, Nakayama S, Adachi T, Takeda Y, Sakamoto S. Effects of diacylglycerol and gonadotropin-releasing hormone on human chorionic gonadotropin release by cultured trophoblast cells. Placenta 1992; 13:213-221. Shi CZ, Zhuang LZ. Norepinephrine regulates human chorionic gonadotropin production by first trimester trophoblast tissue in vitro. Placenta 1993; 14:683-693. Gleicher N, Harlow L, Zilberstein M. Regulatory effect of antiphospholipid antibodies on signal transduction: a possible model for autoantibody-induced reproductive failure. Am J Obstet Gynecol 1992; 167:637-642. DiSimone N, DeCarolis S, Lanzone A, Rosisvalle E, Giannice R, Caruso A. In vitro effect of antiphospholipid antibody-containing sera on basal and gonadotropin releasing hormone-dependent human chorionic gonadotropin release by cultured trophoblast cells. Placenta 1995; 16:75-83. Reza E Igarashi K, Tokita S, Asai K, Aoki J, Asaoka Y, Umeda M, Inoue K. Anti-idiotypic monoclonal antibody recognizes a consensus

48.

49. 50. 51. 52.

53.

54.

55.

recognition site for phosphatidylserine in phosphatidylserine-specific monoclonal antibody and protein kinase C. FEBS Lett 1994; 339:229233. Zhou Y, Damsky CH, Chiu K, Roberts JM, Fisher SJ. Preeclampsia is associated with abnormal expression of adhesion molecules by invasive cytotrophoblasts. J Clin Invest 1993; 91:950-960. Graham CH, Lala PK. Mechanisms of placental invasion of the uterus and their control. Biochem Cell Biol 1992; 70:867-874. Sessions A, Horowitz AE Myoblast aminophospholipid asymmetry differs from that of fibroblasts. FEBS Lett 1981; 134:75-78. Sessions A, Horowitz AE Differentiation related differences in the plasma membrane phospholipid asymmetry of myogenic and fibrogenic cells. Biochim Biophys Acta 1983; 728:103-111. Bevers EM, Comfurius P, Zwaal RFA. Changes in membrane phospholipid distribution during platelet activation. Biochim Biophys Acta 1983; 736:57-66. Thiagarajan P, Tait JE Binding of annexin V/placental anticoagulant protein I to platelets. Evidence for phosphatidylserine exposure in the procoagulant response of activated platelets. J Biol Chem 1990; 265: 17420-17423. Zwaal RFA, Bevers EM, Comfurius P Rosing J, Tilly RH, Verhallen PE Loss of membrane phospholipid asymmetry during activation of blood platelets and sickled red cells; mechanisms and physiological significance. Mol Cell Biochem 1989; 91:23-31. Loke YW. Experimenting with human extravillous trophoblast: a personal view. Am J Reprod Immunol 1990; 24:21-28.