| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Regular Article TA |
a Departments of Pharmacology,
b Obstetrics and Gynaecology,
c Microbiology and Immunology, and
d Perinatal Research Centre, University of Alberta, Edmonton, Alberta T6G 2H7, Canada
ABSTRACT
The large increase in placental surface area and fetal villous vascular development in the third trimester of pregnancy requires degradation and reformation of the placental basal lamina. Degradation is carried out by matrix metalloproteinases (MMPs) secreted by adjacent cells. Although the gelatinases, MMP-2 and MMP-9, which are released by extravillous cytotrophoblasts (CTs) are believed to play crucial roles in early placental expansion, neither has been reported in third trimester villous trophoblasts nor has appropriate (basolateral) release of any MMP by the highly polarized syncytiotrophoblast (ST) been demonstrated. We demonstrated villous trophoblast expression of both MMP-2 and MMP-9 by in situ immunohistochemistry and by Western blot analysis and zymography of lysates and culture supernatants of highly purified villous CTs. We also found that epidermal growth factor (EGF)-stimulated CT differentiation into ST and stimulation by the phorbol diester, PMA, both increase MMP-9 secretion. The direction of MMP release was determined with confluent cultures of ST on porous membranes. We found that >90% of MMP-2 and MMP-9 were released from the basolateral surface. We conclude that villous STs express and release gelatinases from their basolateral surfaces in a regulated manner and suggest that such polarized release may be important to villous tissue remodeling.
kinases, pregnancy, syncytiotrophoblast
INTRODUCTION
The villous placenta is the metabolic interface between a mother and her fetus and, as such, it bidirectionally responds to maternal and fetal signals (reviewed in [1]). During the third trimester of gestation, placental villi respond to accelerated fetal growth with increases in size and umbilical vasculature. This requires a large increase in the surface area of the syncytiotrophoblast (ST), the fetal epithelium that faces maternal blood. Villous growth also requires extensive remodeling of the underlying villous tissue, which consists of the trophoblast basement membrane, villous macrophages and fibroblasts, and closely apposed vascular endothelial cells. It is unknown how this remodeling takes place or what may regulate it.
Tissue remodeling is largely mediated by matrix metalloproteinases (MMPs), which are zinc-dependent endopeptidases that degrade extracellular matrix (ECM) (reviewed in [2]). There are more than 20 known MMPs grouped into subfamilies on the basis of their substrate specificity and homologies. Under physiological conditions, MMPs are involved in ontogenesis as well as angiogenesis and wound healing. Increased expression of MMPs plays a role in the pathogenesis of inflammation and cancer. Aberrant expression of MMP-9 by extravillous cytotrophoblasts (CTs) has also been linked to the pregnancy disease, pre-eclampsia [3, 4]. Metalloproteinases are also involved in processes other than matrix remodeling (i.e., in platelet aggregation [5, 6], vasoconstriction [7], and regulation of ion metabolism [8]).
Metalloproteinases 2 and 9 are enzymes with relative substrate specificity to collagen-containing ECM and are therefore termed gelatinases (reviewed in [2]). They are synthesized, respectively, as 72 and 94 kDa proenzymes that are activated by membrane associated MMPs and extracellular proteases such as plasmin. Although MMP-2 is constitutively expressed by many cells, MMP-9 expression is regulated in osteoclasts, neutrophils, and macrophages by growth factors, cytokines, ECM proteins, and by cell-cell and cell-ECM interactions that alter cell shape. Homozygous mice with a null mutation in the MMP-9 gene show an abnormal pattern of skeletal growth plate vascularization and ossification, which suggests that this gelatinase is crucial for tissue remodeling during angiogenesis [9].
Expression of MMPs by extravillous CT plays an important role in placental invasion into the uterine endometrium during early pregnancy. In particular, MMP-9 is strongly expressed by extravillous CT during trophoblast invasion in the first trimester of gestation [10] and has been shown to be crucial for in vitro trophoblast invasion into collagen gels [11]. However, relatively little is known of MMP expression in the villous trophoblast, the crucial cell layer that faces the maternal blood and performs gas, nutrient, and waste transport functions. Although MMP-1, MMP-3, and MMP-7 were found in human villous trophoblasts, MMP-9 was only occasionally observed in neutrophils [12], which express it at very high levels (reviewed in [2]). These observations suggested that the tissue remodeling that accompanies angiogenesis in the villous stroma may not be mediated by MMP-9, but possibly by other collagen-specific proteases such as MMP-1 or MMP-2.
As a first step in testing this hypothesis, we re-examined MMP-9 expression in the villous placenta and explored the expression of MMP-2. The expression of both MMPs was probed with polyclonal rabbit anti-MMP-2 [13] and MMP-9 immunoglobulin G (IgG). We were surprised to find strong expression of both MMP-2 and MMP-9 in the stromal vascular endothelium and in the trophoblast. Expression was confirmed in highly purified (>99.99%) preparations of villous trophoblasts by Western blot and zymographic analysis, and it was further determined that the gelatinases were released in the fetal (basolateral) direction.
MATERIALS AND METHODS
Cell Isolation and Culture
Villous CTs (>99.99% pure) were isolated from normal term placentas by trypsin/DNase digestion of minced chorionic tissue and elimination of contaminating cells by immunoabsorption onto IgG-coated glass bead columns (Biotex, Edmonton, AB, Canada) as previously described [14–16]. The placentas were acquired by University of Alberta Perinatal Research Centre staff at the Royal Alexandria Hospital, Edmonton, following the ethics guidelines of the Capital Health Authority. The purified cells were routinely cryopreserved and are >90% viable (via trypan-blue-exclusion) after thawing.
Cryopreserved CTs were thawed and plated at 5 x 106 cells in individual wells of a 6-well culture dish (Costar, Cambridge, MA) and cultured in Iscoves modified Dulbeccos medium (IMDM, Gibco, Grand Island, NY) supplemented with 10% fetal bovine serum (FBS, Gibco) and 50 µg/ml gentamicin as previously described [16]. The cells remained predominantly mononuclear after 48 h of incubation in this medium. Differentiated trophoblasts (STs) were prepared by culturing CTs with medium containing rHu EGF (10 ng/ml; Prepro-Tech, Rocky Hill, NJ) to promote syncytialization [17–19]. All cultures were washed 24 h before collection of supernatants with warm IMDM, then cultured for the final day of culture with serum-free IMDM. In some experiments, the cells were cultured for the final day with 100 nM phorbol 12-myristate 12-acetate (PMA; Sigma Canada, Oakville, ON, Canada). Syncytiotrophoblasts cultured in 10% FBS in IMDM with EGF show a low spontaneous rate of apoptosis (<5%, [18]); those cultured with and without serum for 22 h do not have statistically different (P > 0.05) apoptosis frequencies (unpublished observations).
Preparation of Trophoblast Layers on Permeable Insert Membranes
Insert membranes (3-µm pores, 6.4 mm) precoated with fibronectin (Biocoat; Becton Dickinson, Bedford, MA) were soaked in 15% FBS in IMDM for 1 h prior to cell plating. Cytotrophoblasts were thawed, washed once, and 2 x 105 cells in 10% FBS in IMDM were added to each insert. The inserts were incubated at 37°C in a 5% CO2 humidified atmosphere in specialized Falcon companion 24-well tissue culture plates (Becton Dickinson). Nonadherent cells were removed by gentle washing after 4 h, and the cultures were replenished with medium containing EGF. On Days 3 and 7 of culture, freshly thawed and washed CTs from the same placental preparation were added as described earlier. The EGF medium was changed in both insert and lower chambers every 2 days, then replaced with serum-free IMDM 24 h before supernatants were collected (at Day 16 of culture). During the 24-h collection culture, culture medium levels in upper (apical) and lower (basal) culture chambers were maintained at hydrostatic equilibrium (600 µl and 400 µl, respectively) or, where described, at a higher (5 mm) level in the apical chamber to ensure net fluid flow downward (apical to basal).
Transepithelial resistance was measured with an Endohm tissue resistance measurement chamber (World Precision Instruments, Inc., Sarasota, FL) and expressed as ohms. Average net transepithelial resistance in these cultures at Day 16 was 426 ± 32 ohms (n = 5), compared with 212 ± 78 ohms for confluent tight-junctioned MDCK-2 epithelial cells. These values correspond closely to previously observed values (unpublished observations).
Immunoglobulin G Preparations
Polyclonal antipeptide antibodies were generated using a synthetic peptide corresponding to a fragment of gelatin-binding domain of MMP-9. The predicted peptide structure of the human MMP-9 sequence (GenBank database, National Center for Biotechnology Information [NCBI], Bethesda, MD) was analyzed with Wisconsin package software (Genetic Computer Group, Madison, WI). Hydrophilicity, surface probability, chain flexibility, secondary structure, and antigenicity index (Jameson-Wolf) were used as criteria for the choice of peptide. In order to exclude the sequence that was identical or similar to rabbit MMP-9, the selected sequence was compared with protein data banks (NCBI and GenBank) using the pBLAST program. The nine-amino acid fragment (ANYDRDKLF) was chosen, a cysteine was added for the following attachment chemistry, and solid-phase peptide synthesis was carried out at the Biotechnology Support Laboratory, Texas A&M University, College Station, Texas. The resulting peptide was then coupled with m-maleimidobenzoyl-N-hydroxysuccinimide and the activated peptide conjugated to thyreoglobulin. Rabbits were immunized with the resulting conjugate and the serum antipeptide antibody titer was monitored by ELISA. The rabbit IgG was purified by affinity chromatography using Affi-Prep Protein A Support (Bio-Rad, Hercules, CA) and stored at -20°C until used. The preparation of polyclonal rabbit anti-MMP-2 IgG was carried out in a similar fashion and has been previously described [13]. Rabbit IgG (Sigma) was used as a negative control for these studies.
Zymography
The activity of MMP-2 and MMP-9 was measured by zymography under nonreducing conditions as described before [13]. Briefly, 20 µg of protein from conditioned media or cell homogenates was electrophoretically separated on 8% SDS-PAGE with copolymerized gelatin (2 mg/ml) as a gelatinase substrate. The gelatinolytic activities of MMPs were quantified by measuring the density of the cleared bands using a ScanJet 3c scanner (Hewlett-Packard, Boise, ID) and SigmaGel measurement software (Jandel Scientific, San Rafael, CA). Medium conditioned by the human fibrosarcoma cell line, HT1080, was used for proMMP-2 and proMMP-9 size standards as previously described [20].
Immunoblot Analysis
Western blot analyses of MMP-2 and MMP-9 were performed as previously described [5]. Briefly, samples of cell homogenates (20 µg protein) were subjected to 7% SDS-PAGE under reducing conditions, the separated proteins were electroblotted onto polyvinylidene fluoride membranes (Schleicher and Schuell, Keene, NH), and probed with polyclonal antibodies (see earlier discussion) against MMP-9 and MMP-2 (1 µg/ml). Bands corresponding to MMP-9 were detected with an ehanced chemiluminescence kit (Amersham, Piscataway, NJ).
Immunohistochemistry of Tissue Sections
Chunks (1 cm3) from fresh villous placenta were immediately fixed in 5% paraformaldehyde overnight, then embedded in paraffin, sectioned, and the sections were layered onto glass slides by standard procedures. After deparaffination and rehydration, sections were immunohistochemically stained with antibody for MMP-2, MMP-9, or control rabbit IgG (at 1 µg/ml) using the Streptavidin Biotin system (Histostain-SP Kit, Zymed/Intermedico, Markham, CA) with 3-amino-9-ethyl carbazole as the developing chromagen (which yields an intense red color) as previously described [16]. Some membrane inserts cultured with trophoblasts were also paraformaldehyde-fixed, paraffin-embedded, and sectioned at 90° to the membrane, then stained as just described for MMP-2 and MMP-9. Both the tissue and insert sections were counterstained with hematoxylin (Sigma) to visualize nuclei and Stat Stain (VWR, Mississauga, ON, Canada) to visualize cytoplasm.
Statistics
Results are expressed as the mean ± SEM, and comparisons were made using one-way ANOVA followed by the Tukey-Kremer multiple comparison test. When appropriate, Student's t-test for unpaired data was used for confirmation.
RESULTS
MMP-2 and MMP-9 Expression in Term Villous Placental Tissue and Cells
In situ expression of MMP-2 and MMP-9 proteins was evaluated by immunohistochemistry of sections from paraformaldehyde-fixed and paraffin-embedded villous placental tissue using polyclonal rabbit antipeptide IgG preparations as described in Materials and Methods. MMP-9 was found to be expressed strongly in the stromal vascular endothelium, consistently in STs, and weakly in CTs (Fig. 1B). MMP-2 was also expressed in the vascular endothelium and villous trophoblast (Fig. 1D). Trophoblast expression of MMP-2 was not as marked as that of MMP-9 and appeared to be more evenly distributed across the ST. Control immunohistochemistry with rabbit IgG showed no staining (Fig. 1, A and C).
|
The specificity of the antibody preparation to MMP-9 was confirmed by Western blot analysis of trophoblast cell homogenates, which showed detection of a strong 92 kDa band in homogenates of cultured STs (Fig. 2B). Although MMP-2 could be detected in control preparations (homogenates of MMP-2 expressing HT1080 cells), levels in the ST homogenate were below detection (data not shown).
|
The expression of MMP-2 and MMP-9 in villous trophoblasts was further confirmed by zymographic analysis of freshly isolated and cultured villous trophoblasts from term placentas (Fig. 2A). MMP expression was detected in freshly isolated (uncultured) CTs at 94 and 74 kDa, which correspond to proMMP-9 and proMMP-2, respectively, observed in control HT1080 cell culture supernatants (standards). Cytotrophoblasts cultured without EGF for 24 h showed strong expression of MMP-9 in cell homogenates (H lane) but not in the culture medium (M lane). MMP-2 was not detected in either culture compartment. Trophoblasts cultured for 7 days with EGF (which stimulates differentiation into syncytialized cells [15, 16, 18]) maintained cellular expression of MMP-9 but released large amounts of enzyme into culture supernatants (cultured STs, H and M lanes). MMP-2 could be reproducibly detected in ST culture supernatants (M lane) but not in homogenates (H lane).
MMP-9 Expression as a Function of Trophoblast Culture with EGF and Phorbol Diesters
The data suggest that the pattern of MMP-9 expression changed during the differentiation of CT into ST. In order to assess this change in more detail, trophoblasts were cultured for up to 9 days with EGF and the cells and supernatants were analyzed for MMP-9 expression every 2 days. The results (Fig. 3, A and B) show that the relative amount of MMP-9 accumulating in the supernatant during a 24-h period peaked at Day 7 of culture when the cells became completely syncytialized [15], and reproducibly decreased on Day 9. In contrast, cell-associated MMP-9 remained relatively constant for the 9-day culture period (Fig. 3C).
|
In order to determine if MMP-9 expression could be regulated, cultured CTs and STs were stimulated for 24 h with the phorbol diester, PMA, a potent stimulator of protein kinase C (PKC) [21] and a potent stimulator of MMP expression [5, 22–24]. PMA did not stimulate MMP-9 expression from CTs but strongly stimulated its cellular accumulation (Fig. 4A) and release (Fig. 4B) from STs.
|
Polarized Release of MMP-2 and MMP-9 from Cultured Syncytiotrophoblasts
Villous STs are positioned at the interface between maternal blood and the villous stroma; the latter is composed of a basal lamina beneath which are fetal capillaries, villous fibroblasts, and macrophages [25]. Syncytiotrophoblasts are polarized epithelial cells; thus, MMP-2 and MMP-9 could be released either from apical surfaces (facing maternal circulation) or basal surfaces (facing the villous basal lamina). The extent and direction of polarized MMP secretion was determined with STs cultured on permeable membranes (unpublished results). These cultures are confluent and tight-junctioned as demonstrated by high electrical resistance and restricted diffusion of macromolecules from the upper to lower culture chambers. Syncytiotrophoblasts cultured on such membranes express MMP-2 and MMP-9, which can be detected with immunohistochemistry (Fig. 5). After 24 h of culture in serum-free medium, >90% of MMP-2 and MMP-9 were found in the lower culture chamber (Fig. 6). In order to determine whether even the low levels of MMPs found in the upper chamber reflected passive diffusion from the lower to the upper chamber, the experiment was repeated under conditions that minimized upward leakage. The culture medium level in the upper chamber was made 5 mm higher than the level in the lower chamber to maintain a hydrostatic gradient that would discourage upward leakage, and the experiment was carried out for only 2 h. Under these conditions, approximately 20% of total MMP-9 was nonetheless found in the upper chamber (data not shown), suggesting that gelatinases accumulating in the upper chamber were not leaking upward from the lower chamber.
|
|
DISCUSSION
Although MMP-1, MMP-3, and MMP-7 were previously found in villous trophoblasts, MMP-9 was observed only in occasional neutrophils [12], which strongly express it (reviewed in [2]). There are no reports of MMP-2 expression in villous placenta; thus, our observations of MMP-2 and MMP-9 expression in villous trophoblasts are novel. The differences in MMP-9 expression from the earlier study may be attributed to differences in antibody preparations (ours were novel), to the manner of tissue fixation (we used 5% paraformaldehyde overnight) or length of time between delivery and fixation (we collected tissue within 15 min of delivery).
Our observations that PMA can increase MMP-9 production and secretion indicate that MMP-9 production from STs can be regulated via activation of PKC, as has been shown in other cells [23, 24]. It is interesting that undifferentiated villous CTs, although they can produce and release MMP-9, are not regulated by PMA. Cytotrophoblasts also release much less MMP-9 and almost undetectable levels of MMP-2, in contrast to STs. Our observations suggest that production of MMP-2 and MMP-9 is primarily a function of differentiated STs, where they can be regulated by either maternal or fetal stimuli.
MMP-9 from polarized cells can be preferentially secreted either from the apical surfaces (as with a variety of epithelial cells [26, 27]) or basal surfaces (as from bovine aortic endothelial cells [28]). MMP-2 secretion can also be preferentially apical (as with rabbit kidney collecting duct cells [27]) or basal (as with both endothelial cells [28] or human epithelial cells [26]). However, MMP-2 and MMP-9 can be transported in different directions by the same cell [26]. We saw preferential (>90%) MMP-9 and MMP-2 accumulation in the basal supernatant of STs cultured on porous membranes. Given that basal supernatant accumulation requires passage across a nucleopore membrane composed of uniform 3-µm holes occupying <10% of the surface with >90% being impermeable (Fig. 5) and that apical accumulation has no comparable barrier (diffusion is directly into the supernatant), we can conclude that strongly preferential basal release of the MMPs took place under the conditions of our experiments.
We cannot exclude the possibility that STs release a small amount of MMPs from their apical surfaces. A comparison of the present studies with studies of basal release from confluent endothelial cells [28] shows somewhat more apical accumulation from endothelial cells (<20% of the total) than from trophoblasts (<5%). Both culture types allowed similar macromolecule diffusion across confluent cell layers (approximately 1.3% per hour for endothelial cells and <0.6% per hour for trophoblasts (unpublished results). However, our observations of some apical accumulations under conditions that minimize basal to apical chamber release argue that accumulation of MMP-9 in the apical supernatant of ST cultures is not via diffusion through the membrane and cell layer from the basal chamber. The results suggest limited MMP-9 release from the apical (maternal) surface of STs. This suggestion is in accord with observations of an increase in maternal plasma levels of MMP-9 during pregnancy [29].
Whereas apical release of MMP-2 and MMP-9 has implications for maternal vascular functions (platelet aggregation [5] or vasoconstriction [7]), basal release from STs ejects the gelatinases onto the trophoblast basal lamina in the third trimester of gestation, when there are few underlying CTs [1]. The villous placenta also undergoes a rapid expansion in size in the third trimester, which requires a simultaneous increase in villous umbilical angiogenesis [30]. Increased growth of villous STs and vascular endothelia both require remodeling of the intervening trophoblast basal lamina. Thus, the basal release of MMP-9 that we observe from third trimester STs is appropriate for such tissue remodeling.
FOOTNOTES
1 This study was carried out with funds received from MRC grant MT-15479 to L.J.G. and 14074 to M.W.R. M.W.R is a Medical Research Council of Canada Scientist. 
2 Correspondence: Larry Guilbert, Department of Medical Microbiology and Immunology, 6-25 HMRC, University of Alberta, Edmonton, AB, Canada T6G 2S2. FAX: 780 492 9828; larry.guilbert{at}ualberta.ca
Accepted: June 8, 2000.
Received: April 27, 2000.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  H. A. Coppock, A. White, J. D. Aplin, and M. Westwood Matrix Metalloprotease-3 and -9 Proteolyze Insulin-Like Growth Factor-Binding Protein-1 Biol Reprod, August 1, 2004; 71(2): 438 - 443. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Wang, D. F. Lewis, Y. Gu, Y. Zhang, J. S. Alexander, and D. N. Granger Placental Trophoblast-Derived Factors Diminish Endothelial Barrier Function J. Clin. Endocrinol. Metab., May 1, 2004; 89(5): 2421 - 2428. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. E. Curry Jr. and K. G. Osteen The Matrix Metalloproteinase System: Changes, Regulation, and Impact throughout the Ovarian and Uterine Reproductive Cycle Endocr. Rev., August 1, 2003; 24(4): 428 - 465. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Hube, P. Reverdiau, S. Iochmann, S. Trassard, G. Thibault, and Y. Gruel Demonstration of a Tissue Factor Pathway Inhibitor 2 Messenger RNA Synthesis by Pure Villous Cytotrophoblast Cells Isolated from Term Human Placentas Biol Reprod, May 1, 2003; 68(5): 1888 - 1894. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. G. Hemmings and L. J. Guilbert Polarized Release of Human Cytomegalovirus from Placental Trophoblasts J. Virol., June 5, 2002; 76(13): 6710 - 6717. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. Mayers, T. Hurst, L. Puttagunta, A. Radomski, T. Mycyk, G. Sawicki, D. Johnson, and M. W. Radomski Cardiac surgery increases the activity of matrix metalloproteinases and nitric oxide synthase in human hearts J. Thorac. Cardiovasc. Surg., October 1, 2001; 122(4): 746 - 752. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
