| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Departments of Obstetrics and Gynecology (T.S., B.C.) and Pathology (A.E.), University Hospital, S-221 85 Lund, Sweden
Address all correspondence and requests for reprints to: Bertil Casslén, M.D., Ph.D., Department of Obstetrics and Gynecology, University Hospital, S-221 85 Lund, Sweden.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Cell migration, i.e. the locomotion of a cell over a substratum, involves extension of the cell at the leading edge and retraction at the trailing edge. Focal adhesion sites function as the transition between cytoplasmic actin fibrils and the extracellular matrix. During migration the urokinase plasminogen activator receptor (uPAR) is located at focal adhesion sites in the leading edge of the cell, and activation of uPAR by ligand binding initiates assembly of focal adhesion proteins involving integrin receptors, vitronectin, caveolin, p130cas, and focal adhesion kinase (7, 8, 9).
We have previously shown that endometrial tissue explants as well as separated stromal and glandular epithelial cells in culture release uPA, and that human microvascular and umbilical vein endothelial cells migrate in response to uPA (10, 11, 12, 13, 14). Our data, furthermore, demonstrated that the catalytic site of uPA was not required, as uPA in complex with plasminogen activator inhibitor-1 was equally efficient to stimulate cell migration. In contrast, binding of uPA to uPAR was critical, because interference with this interaction inhibited the migratory response. The present study of microvascular endothelial cells aimed to characterize the chemotactic signal released from endometrial tissue with respect to cyclic variation, molecular nature, and mechanism of action.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Endometrial tissue was obtained from uteri removed for benign nonendometrial pathology, i.e. cervical dysplasia, dysmenorrhea, uterine prolapsus, adenomyosis, and fibromyomas. All patients were parous, 30–45 yr of age, and had regular menstrual cycles. None used oral contraceptives or hormone replacement therapy. Permission to use part of the endometrium was granted by the university review board for studies on human subjects. The endometrial samples were gently scraped off the upper part of the uterine cavity immediately after removal of the uteri and were transferred to the laboratory in sterile Hanks’ Balanced Salt Solution (HBSS). Endometrial pathology was excluded by the histopathological examination. All endometrial samples were classified with respect to the phase of the menstrual cycle (15, 16). The proliferative phase was subdivided into early (days 5–7), mid (days 8–10), and late (days 11–14) stages, and the secretory phase was similarly divided into early (days 15–19), mid (days 20–22), and late (days 23–28) stages. The late proliferative and early secretory stages were subsequently included in the midcycle phase. Unless specified, endometrial tissue samples used in the experiments were obtained in the proliferative, midcycle, and secretory phases.
Endothelial cell culture
A human microvascular endothelial cell line (HMEC-1) was obtained from CDC (Atlanta, GA). These cells were grown on uncoated plastic in medium MCDB 131 supplemented with EGF (10 ng/mL), hydrocortisone (1 µg/mL), and 15% FBS. Some experiments were also performed with two additional endothelial cell types, HUVEC and ECV-304. HUVEC, obtained from human umbilical cords as we previously described (14), were cultured in plates coated with 0.1% gelatin in medium 199 supplemented with 20% FBS. ECV 304, a spontaneously transformed immortal human endothelial cell line obtained from American Type Culture Collection (Manassas, VA), were grown on uncoated plastic in medium 199 supplemented with 10% FBS. All endothelial cells were grown in six-well tissue culture plates at 37 C in 5% CO2–95% air.
Wound assay
At confluence a wound was made by pressing a razor blade through the cell sheet and gently scraping off the cells, essentially as described by Sato and Rifkin with modifications (14, 17). The cultures were rinsed with HBSS to eliminate debris, and medium without supplements was added. Culture medium and HBSS were phenol red free. Explants (1-mm3 pieces) of endometrial tissue (50 mg/well) were added to the denuded area, i.e. not in physical contact with the endothelial cells, and the wounded cultures were subsequently incubated for 24 h. The cultures were then fixed in absolute methanol for 5 min and stained with 12% Giemsa for 15 min. The number of cells that had migrated across the wound line was counted in four to six high power fields in each well, and the mean was taken as the result. Migration in stimulated wells was subsequently calculated as a percentage of that in unstimulated wells of the same experiment. Wound repair assays, like that described here, have been shown not to be affected by variations in cell proliferation (18, 19). The design of the migratory assay allowed us to study the effect of signal molecules released in serum-free medium from endometrial tissue explants and affecting microvascular endothelial cells, which were not in contact with the explants. By using the same amount of endometrial tissue (50 mg) in all experiments, the response could be quantitatively related to the cyclic phase of the explants. EGF, bFGF, VEGF, phosphatidyl inositol-specific phospholipase C (PiPLC), and monoclonal antibodies to EGF, VEGF, and bFGF were obtained from Sigma (St. Louis, MO), and Aspergillus niger (control antibody) was purchased from Dakopatts (Copenhagen, Denmark). Blocking monoclonal antibody to uPAR, R3, was a gift from Dr. G. Høyer-Hansen (Finsenlab, Copenhagen, Denmark) (20).
Assay for EGF and uPA
The concentrations of EGF in conditioned medium were measured with a commercial enzyme-linked immunosorbent assay (ELISA; Quantikine EGF immunoassay kit, R & D Systems, Inc., Abingdon, UK). This is a solid phase ELISA that uses a monoclonal catching antibody and a horseradish peroxidase-linked polyclonal detecting antibody. We used 3.1 pg/mL as the detection limit. The intraassay variation was 3%, and the interassay variation was 5%. Conditioned medium concentrations of uPA were assayed with a commercial TintElize uPA (Biopool, Umea, Sweden). The assay has two monoclonal catching antibodies and a horseradish peroxidase-polyclonal antibody conjugate detection system. It includes a system for internal subtraction of nonspecific signal. The ELISA measured the single chain form and the high molecular weight form of uPA with high efficiency. The low molecular weight form of uPA was measured with about 40% efficiency. The detection limit was 0.1 ng/mL. At 0.7 ng/mL, the intraassay variation was 10%.
Purification and radiolabeling of uPA
uPA was purified from Ukidan (Serono, Geneva, Switzerland) by affinity chromatography on a benzamidine-Sepharose column as previously described (12). The active enzyme fraction was further separated in high (50 kDa) and low (33 kDa) molecular mass fractions by gel filtration on Sephadex G-100. The peak fraction of high molecular mass uPA, which was 98–99% pure in immunoblotting, was used for 125I labeling with the lactoperoxidase method. Specific radioactivity ranged from 0.4–0.6 megabecquerels/g protein.
Assay of [125I]uPA binding to HMEC-1 cultures
Confluent cultures were given serum-free medium and stimulated with EGF (10 ng/mL) or HBSS as a control for 24 h at 37 C before being incubated on ice for 2 h with radiolabeled uPA in HBSS containing BSA (10 g/L). After removal of the experimental buffer, cells were washed six times with ice-cold HBSS and subsequently lysed with NaOH (1 mol/L). The radioactivity of the lysate was counted in a 1260 Multigamma counter (Pharmacia Biotech, Uppsala, Sweden). Nonspecific binding was assayed with the same procedure, but in the presence of a 100-fold molar excess of unlabeled uPA, and specific binding was calculated as the difference between total and nonspecific binding. To measure the total number of receptors, endogenously bound uPA was removed by briefly (2–3 min) exposing the cells to 75 mmol/L acetate buffer, pH 3.0, containing 2.5 mmol/L CaCl2, 0.5 mmol/L MgCl2, and 0.3 mol/L NaCl, and then rinsed four times with ice-cold HBSS before the incubation with radiolabeled uPA. The number of occupied receptors was calculated as the difference between the numbers of total and free receptors.
Statistical methods
Results were expressed either as the median and percentile or as the mean ± SE. Differences between groups were tested for significance using Wilcoxon signed rank test for paired comparison and Mann-Whitney U test for nonpaired comparison.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). Incubation of the endometrial
explants with estradiol for 24 h before the wound assay increased
the chemotactic signal, whereas preincubation with progesterone alone
resulted in a slight reduction (Fig. 2).
This negative effect of progesterone was, however, reversed by the
presence of estradiol.
|
|
). Also, this antibody blocked about
half the stimulatory effect of endometrial tissue explants, suggesting
that the endometrial migratory signal has at least two components, one
that engages the uPA system and one that does not. In an additional
experiment we performed the cell migration assay with endometrial
explants in the presence of PiPLC, which efficiently cleaves the
phosphatidyl inositol anchor of uPAR and thus depletes the cell surface
of uPAR. Treatment with PiPLC resulted in more than 50%
reduction of the endometrial stimulation (not shown). These results
suggest that the major part of the response to endometrial tissue in
HMEC-1 was mediated by uPAR.
|
Several growth factors that are known to be present in the endometrium
have effects on endothelial cells. Thus, we performed the migratory
assay with endometrial tissue explants in the presence of antibodies to
VEGF, bFGF, and EGF. The results show that antibodies to EGF, but not
antibodies to bFGF, VEGF, or control monoclonal antibodies, reduced
HMEC-1 migration by more than 50% (Fig. 4). This finding suggests that EGF is
involved in the migratory signal released from endometrial tissue. To
make sure that this effect of the explants was not specific to HMEC-1,
we performed the wound assay with HUVEC in the presence of explants.
Migration was increased to 203 ± 28% (n = 5;
P = 0.04) of the control value in the presence of
explants and was reduced to 158 ± 14% (n = 5;
P = 0.04) by the concomitant presence of anti-EGF
antibodies (not shown). To verify that EGF was actually released from
endometrial tissue, we analyzed 24-h conditioned medium from 21
endometrial tissue explant cultures, again using the same amount of
tissue as in the migratory assay, i.e. 50 mg. The release of
EGF tended to be lower at midcycle than earlier in the proliferative
phase and later in the secretory phase (Fig. 5). We also found that the conditioned
medium concentration of EGF was higher in three explant cultures that
had been stimulated with 1 nmol/L estradiol than in control cultures
(not shown). Four cultures had EGF levels below the detection limit and
did not respond to estradiol. Two of the nonresponding cultures, but
none of the responding cultures, were obtained in the midcycle phase.
Conditioned media from HMEC-1 cultures had no detectable EGF.
|
|
Next, we studied the effect of EGF on the endothelial cells and
analyzed the interaction between EGF and the uPA system. EGF stimulated
migration in HMEC-1, and this effect was dose dependent in the range
0.001–10 ng/mL (not shown). Migration was also stimulated by EGF in
HUVEC (168 ± 8% of control; n = 5; P =
0.04) and in ECV 304 (212 ± 4% of control; n = 5;
P = 0.04). These cells do not require EGF for growth.
We subsequently stimulated HMEC-1 with 10 ng/mL EGF in the wound assay
in the presence and the absence of antibodies to uPAR. The stimulatory
effect of EGF on HMEC-1 migration was completely abolished by a
blocking antibody to uPAR (Fig. 6). When
the same amount of endometrial tissue was incubated with or without
HMEC-1, we found in two explant cultures with detectable levels of EGF
that the conditioned medium concentration of EGF was decreased in the
presence of HMEC-1, suggesting that endometrial EGF is, in fact, bound
to the endothelial cells. These results together suggest that
endometrial EGF up-regulates the uPA system in endothelial cells, which
subsequently in an autocrine mode promotes migration. To further
clarify whether this was true for the migratory signal from endometrial
tissue, we compared the effects of antibodies to EGF and uPAR
separately and together. Each of the two antibodies by themselves was
as effective as the combination in inhibiting the endometrial increase
in cell migration (Fig. 7). This suggests
that EGFR and uPAR are mutually dependent actors within a common
sequence of events leading to cell migration, and that both receptors
need to be activated. To analyze plausible up-regulation of the uPA
system by EGF, we assayed conditioned medium from HMEC-1 cultures for
uPA and found the concentration of uPA in EGF-treated cultures to be
decreased by 20 ± 4% (n = 6; P = 0.046; not
shown in the figure). We then quantitated the effect of EGF on the
expression of uPAR in HMEC-1 using a radioligand binding assay. The
amount of free, i.e. endogenously not occupied, receptor
sites was increased by 31 ± 2% after EGF treatment (n
= 8; P = 0.01; not shown in the figure). In contrast,
the amount of occupied receptor sites was not different between
EGF-treated and control cultures. Thus, the reduction of extracellular
uPA in EGF-treated cultures probably results from increased
receptor binding. As, however, the amount of endogenously occupied
receptor sites was not increased in EGF-treated cultures it is likely
that internalization and degradation of uPA were also increased. This
process involves uPAR, and up-regulation of uPAR is a key event in
increasing the turnover of uPA in endometrial cells
(13).
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Migration can be stimulated in HMEC-1 by uPA via binding to and activation of uPAR (14). In this study we found that EGF-dependent migration in these cells was also mediated via uPAR. The effect, however, did not result from increased release of uPA, but from up-regulating the number of uPAR-binding sites. As a result, an increased number of potential interactions between uPAR and integrin receptors may take place at focal adhesion sites. Migration in various cell types is reportedly dependent on up-regulated expression of either uPA or uPAR (14, 21, 22, 23, 24, 25), and antibodies against uPA (21, 26) as well as uPA antisense oligonucleotides inhibit cell migration (27).
Increased expression of uPAR may have several different effects, reflecting various functions of the receptor and presumably various locations on the cell surface. At focal adhesion sites an increased density of uPAR increases the number of signaling events upon activation by ligand binding, and eventually this results in increased cell migration. In addition, increased enzymatic activity of receptor-bound uPA generates more plasmin for extracellular proteolysis, which is required during the invasive process (28). However, uPAR has yet another function. The turnover of uPA involves internalization of the enzyme bound to uPAR, a process mediated by members of the low density lipoprotein receptor family at clathrin-coated pits (28). Up-regulation of uPAR increases internalization and degradation of uPA, and this is, in fact, the mechanism by which endometrial cells decrease uPA activity in response to progesterone (13, 29). Apparently, the increased expression of uPAR in HMEC-1 stimulated with EGF resulted in increased internalization and degradation of uPA, as we found a decreased concentration of uPA in the conditioned medium and no increase in cell-bound uPA.
As the effect of antibodies to EGF and uPAR combined was not superior
to the effect of each of them separately, we suggest that activations
of EGFR and uPAR are successive events along the pathway leading to
cell migration, and that activation of both is necessary for any
response to occur. It has been shown that the full effect of growth
factor receptor signaling requires activation and assembly of proteins
at focal adhesion sites, as extracellular matrix contact is required by
some adherent cells to allow growth factor-stimulated signaling through
the mitogen-activated protein kinase pathway (30, 31). An
additional possibility may be a requirement for complex formation
between activated tyrosine kinase receptors and other receptors, as has
been reported for platelet-derived growth factor receptor and
vß3 integrin and for
EGF receptor (EGFR) and G protein-linked receptors (32, 33). We found that ligand activation of EGFR resulted in
up-regulation of uPAR, which is an integral part of focal adhesion
sites. Furthermore, when uPAR was blocked by antibodies and could not
be activated, there was no migratory response to EGF. Thus, our
observations are fully compatible with the above reports and further
support a key role for focal adhesion site activation and assembly in
the process of endothelial cell migration in response to EGF.
The mechanism by which EGF increases uPAR expression has not been elucidated. However, it is possible that the mitogen-activated protein kinase pathway is involved, and that downstream targets involve transcription factors and activator proteins that bind to the promotor region of the uPAR gene (34). In addition, activation of uPAR by ligand binding may initiate a positive feedback loop on its own expression, as activation of Src kinases has been found to result in increased transcription of the uPAR gene (35).
We found endothelial cell migration to be stimulated mostly by endometrial tissue samples taken in the early to midproliferative phase and the mid to late secretory phase. Samples taken at midcycle, i.e. the late proliferative and early secretory phases, had a much lower stimulatory effect. This pattern of angiogenic activity was also found by Ferenczy et al., who studied uptake of radiolabeled thymidine by endometrial tissue in vitro (36). These researchers found a higher uptake by endothelial cells of the functional endometrium in the early to midproliferative and the mid to late secretory phases and a lower uptake at midcycle. Rogers et al. (37), studying migration of endothelial cells in response to endometrial tissue, had a similar finding, although these researchers found an additional peak in the mid/late proliferative phase. Differences between their findings and ours may relate to their use of bovine aortic endothelial cells and a semiquantitative technique. However, in an immunohistochemical study of proliferating cell nuclear antigen expression in endometrial endothelial cells Goodger et al. found no cyclic variation (38). Using the chorioallantoic membrane assay, Peek et al. found no significant variation over the cycle in the angiogenic signal released from endometrial explants (39). These conflicting results may reflect differences in study design and assay systems.
The early proliferative peak of angiogenic activity noted by Ferenczy et al. and Rogers et al. has also been observed morphologically (40) and is further supported by our data. Furthermore, Peek et al. showed that estradiol stimulated proliferation in human decidual endothelial cells (41). The reduction of explant-induced migration at midcycle may appear difficult to explain, as several reports agree that the endometrial expression of EGF is higher in the late than in the early proliferative phase (5, 6), and studies in mice have shown that endometrial expression of EGF is induced by estradiol (42). Also, we found increased signal from endometrial tissue explants, which had been exposed to estradiol. One possible explanation relates to the impact of EGFR, as endometrial expression of EGFR messenger ribonucleic acid is at a peak in the midproliferative phase, and EGFR protein is increased in the late proliferative phase (6, 43, 44), and autocrine binding of EGF to EGFR in the endometrial cells may reduce the release of EGF to the extracellular fluid. In fact, Taketani and Mizuno found that binding of [125I]EGF to endometrial tissue increased during the proliferative phase to high levels during midcycle and was subsequently very low during the mid to late secretory phase (45). This pattern of EGF binding is perfectly opposite the pattern found by Ferenzy et al. for endothelial cell proliferation (36) and also to that found in our study for endothelial cell migration and for endometrial release of EGF. These data taken together strongly suggest that autocrine binding of EGF by endometrial cells is a mechanism by which paracrine effects of EGF are regulated.
Another possible explanation involves a negative feedback mechanism induced by EGF. As reported, an additional effect of extracellular signal-regulated kinase 2, which is different from its tyrosine-phosphorylating capacity, is to reduce integrin receptor activity by decreasing its ligand binding affinity (46). That negative feedback loop, by which activation of the EGFR down-regulates endothelial cell migration at the integrin level, could alternatively explain the decreased migratory signal at midcycle.
A third possible explanation involves a concomitant endometrial release of a cofactor for EGF, provided that this cofactor acts to amplify the response to EGF and has a different cyclic pattern. Support for this possibility comes from our observation that a 1000-fold higher concentration was required for a similar migratory response when EGF was added exogenously compared with the concentration of EGF in endometrial explant-conditioned medium. This cofactor is probably not uPA, because uPA was, in fact, released by the endothelial cells themselves. Hypothetically, the cofactor may activate either a different signaling receptor that subsequently interacts with EGFR (32, 33) or other cell surface components that participate in the assembly of focal adhesion sites.
The second peak, which occurs in the mid and late parts of the secretory phase, may not be related to progesterone per se, as progestagenic steroids, at least in experimental systems, seem to be angiostatic (47). This is also suggested by our own observation of a slightly decreased stimulatory effect of endometrial tissue explants after preincubation with progesterone. Also, an inhibitor of angiogenesis, thrombospondin-1, is up-regulated by progesterone in the human endometrium and may be responsible for spatial or temporal limitation of angiogenesis (48). However, endometrial expression of EGF in the secretory phase was particularly prominent in stromal cells surrounding small arteries, an observation that may indicate a functional relationship (5). Most studies of endometrial EGF and EGFR have not distinguished among the early, mid, and late parts of the secretory phase (5, 6, 44) and are thus not very helpful in understanding the EGF-dependent migratory signal in this part of the cycle. However, Taketani and Mizuno, who assayed binding of [125I]EGF to endometrial tissue day by day during the menstrual cycle, found that binding was very much decreased in the mid to late secretory phase (45). This potentially allows a greater percentage of EGF to be released in the extracellular fluid. We found that endometrial release of EGF tended to be higher in the mid to late secretory phase than in the midcycle phase. We suggest that this is the most likely mechanism by which endometrial cells modulate the paracrine stimulation of angiogenesis.
A minor part of the migratory signal was not blocked by antibodies to either EGF or uPAR, suggesting that a different cytokine using a different mechanism is involved. We studied VEGF, which is known to occur in the endometrium and have important functions during vasculogenesis (1, 49), and bFGF, which is also present and is known to stimulate endothelial cell migration (2, 22). However, our data do not support that these growth factors were released to the medium by endometrial tissue explants, although HMEC-1 migration was sensitive to these agents. If bFGF or VEGF had been released by the endometrial explants and if either of them had been responsible for the residual part of the migratory response, their effects would be expected to be inhibited by both the specific antibody and the uPAR antibody. Thus, we conclude that the remaining part of the migratory signal is caused by an unidentified cytokine, and its response does not involve activation of uPAR.
Received April 28, 2000.
Revised October 4, 2000.
Revised December 12, 2000.
Accepted December 22, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2-macroglobulin receptor. Mol Hum
Reprod. 4:585–593.vß3 integrin
associates with activated insulin and PDGFß receptors and potentiates
the biological activity of PDGF. EMBO J. 16:5600–5607.[CrossRef][Medline]
-related factor. J Biol Chem. 274:4702–4714.This article has been cited by other articles:
 |
|
 |
|
  C. Print, R. Valtola, A. Evans, K. Lessan, S. Malik, and S. Smith Soluble factors from human endometrium promote angiogenesis and regulate the endothelial cell transcriptome Hum. Reprod., October 1, 2004; 19(10): 2356 - 2366. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Trisciuoglio, A. Iervolino, A. Candiloro, G. Fibbi, M. Fanciulli, U. Zangemeister-Wittke, G. Zupi, and D. Del Bufalo bcl-2 Induction of Urokinase Plasminogen Activator Receptor Expression in Human Cancer Cells through Sp1 Activation: INVOLVEMENT OF ERK1/ERK2 ACTIVITY J. Biol. Chem., February 20, 2004; 279(8): 6737 - 6745. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. D. Albrecht, J. S. Babischkin, Y. Lidor, L. D. Anderson, L. C. Udoff, and G. J. Pepe Effect of estrogen on angiogenesis in co-cultures of human endometrial cells and microvascular endothelial cells Hum. Reprod., October 1, 2003; 18(10): 2039 - 2047. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. S. Gambino, N. G. Wreford, J. F. Bertram, P. Dockery, F. Lederman, and P. A.W. Rogers Angiogenesis occurs by vessel elongation in proliferative phase human endometrium Hum. Reprod., May 1, 2002; 17(5): 1199 - 1206. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
). The
stimulatory effect was completely inhibited by antibodies to uPAR
(P = 0.04; 
). Also, basal migration was reduced
by antibodies to uPAR (P = 0.04).