| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Carcinogenesis |
Departments of Laboratory Pathology and Medicine [J. T. A., D. G. K.] and Obstetrics and Gynecology [B. A. L.], Division of Reproductive Endocrinology and Infertility, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599, and Helsinki University Central Hospital, Helsinki, Finland [M. S.]
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Epithelial cancers exhibit disruption of normal cell-cell interactions at many levels. Epithelial cancers typically manifest abnormal oncogene or tumor suppressor gene expression and may have lost the ability to be regulated by the local tissue environment (4) . This abnormality affects paracrine signals from the mesenchyme/stroma that are present in both embryonic and normal adult tissues leading to the associated changes in the epithelium-ECM-stroma relationships. It has been postulated that such changes lead to dysregulation of growth and eventually to neoplasia (26) . The loss of regulatory or suppressive influences over epithelial proliferation may also be the result of aging of the stroma (27) , or reversion to a more fetal phenotype, which contributes to the expansion of associated epithelial cells (28) . Such observations raise the possibility that some cancers arise by the escape from normal stromal control in addition to epithelial genetic mutations.
This study presents data that support the hypothesis that endometrial stromal cells regulate growth, differentiation, and hormonal responsiveness of endometrial epithelium. Additionally, stromal cells cultured in the appropriate ECM can modify the malignant phenotype of a well-differentiated endometrial cancer cell line (Ishikawa). Using a novel cell-culture system (29) , we have attempted to reconstitute the normal paracrine-endocrine-ECM relationships in vitro. The cell culture method described below provides a valuable tool for understanding these complex interactions and may lead to better methods of treatment for this and other epithelial cancers.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Endometrial tissue was rinsed in Hanks BSS to remove blood and debris. After gentle centrifugation (600 x g) the supernatant was removed, and the tissue was placed on a 100-mm plastic tissue culture dish (Corning-Costar, Cambridge, MA). The entire procedure was performed under a sterile laminar flow hood. The tissue was minced with sterile scalpels into 1-mm2 fragments and digested with collagenase (2 mg/ml; CLS-1; Worthington Biomedical, Freehold, NJ) in DMEM-H medium (as above) for 2.5 h at 37°C on a shaking rotor. The tissue digest was vigorously pipetted and added to a stacked sterile wire sieve assembly with #100 wire cloth sieve (140 µm size; Newark Wire Co., Newark, NJ), followed by a #400 wire cloth sieve (37 µm). After the endometrial digest was added to the top of the sieve assembly, the epithelial glands were retained in the #100 and #400 sieves while the stromal cells passed through to the receptacle below. Stromal cells were collected from the lower receptacle, pelleted by centrifugation, and resuspended in 3 ml of DMEM-H medium. RBCs were removed by carefully layering cell suspension over 3 ml of Ficoll-Paque (Pharmacia, Piscataway, NJ) in a sterile 15-ml polycarbonate tube. The solution was centrifuged for 8–10 min at 400 x g. The medium/Ficoll interface layer containing the stromal cells was plated onto 100-mm plastic tissue culture dishes. Stromal cells were cultured for 3 days before being plated for CC experiments, as described below.
Ishikawa cells were derived from an endometrial adenocarcinoma and have been characterized previously (30 , 31) . Stromal cell stocks, Ishikawa cells, and the normal human foreskin fibroblasts, NHF-1 cells (32) , were maintained in a 1:1 mixture of M199:Ham’s F12 medium (Life Technologies, Inc.) supplemented with 4% heat-inactivated FBS, ITS+ [containing insulin (0.62 µg/ml), transferrin (0.62 µg/ml), and selenium (0.62 ng/ml), BSA (125 µg/ml) and linoleic acid (52.6 µg/ml; BD Biosciences, Bedford, MA)] plus 100 units of penicillin, 0.1 mg of streptomycin and 0.25 of µg amphotericin B/ml; (Antibiotic/antimycotic solution; Sigma Chemical Co.). FBS was heat inactivated by incubating in a 56°C water bath for 1 h before filter sterilization.
When stromal and Ishikawa cells were grown in CC, the medium was changed to phenol red-free M199 and F12 medium (1:1) with ITS+ [insulin (3.1 µg/ml), transferrin (3.1 µg/ml), selenium (3.1 ng/ml), BSA (625 µg/ml), and linoleic acid (263 µg/ml)] plus 0.1 mM phosphoethanolamine (Sigma Chemical Co.). Charcoal-stripped serum was reduced to 1–2%. In experiments involving steroid hormone treatments, the hormones were diluted in 100% ethanol. Hormones were added to culture medium at the specified dosages, 17ß-estradiol, 10-8 M (E; Sigma Chemical Co.); medroxyprogesterone acetate, 10-6 M (P; Sigma Chemical Co.), and RU486 10-6 M (R; gift from Dr. Richard Little, University of Pennsylvania, Philadelphia, PA). Ethanol concentrations did not exceed 0.5% in culture medium.
Immunohistochemistry for Cytoskeletal Proteins.
Stromal and epithelial cells in CC were visualized by cell-specific staining of cytoskeletal proteins. Primary antibodies to cytokeratin (rabbit antihuman, Novocastra, Newcastle upon Tyne, United Kingdom) or vimentin (mouse antihuman, Dako Corp., Carpinteria, CA) were used to examine epithelial (Ishikawa; Texas Red secondary fluorescence) and stromal (FITC-green)-specific markers, respectively. Nuclei were labeled with a blue immunofluorescence by DAPI (Sigma). Cultures were rinsed with PBS and fixed with 3.7% formaldehyde in PBS for 10 min. They were then treated with ice-cold 100% methanol, then 100% acetone, then 0.4% Triton X-100 in PBS, each for 10 min with PBS rinsing between treatments. Samples were treated with 2% horse serum and 2% goat serum in PBS containing 4% BSA for 30 min at 37°C to block nonspecific binding of primary antibodies. Antibodies were diluted in PBS containing 4% BSA with 2% horse and 2% goat sera at 1:50 for cytokeratin and 1:100 for vimentin. Controls included samples treated with no primary antibody but treated with secondary fluorescent antibodies. The primary antibodies were incubated on cells for 45 min at 37°C then overnight at 4°C. Samples were rinsed three times with PBS. Fluorescent secondary antibodies were added concurrently: fluorescein-antimouse IgG, 1:75; and Texas Red-antirabbit IgG, 1:75 (Vector Labs, Burlingame, CA) in 2% horse and 2% goat sera for 30 min at 37°C. Samples were rinsed three times with PBS, and fresh DAPI was added at 0.2 µg/ml PBS for 2 min at RT. Slides were rinsed with PBS and coverslips added using antifade mounting medium (courtesy of Department of Pathology and Laboratory Medicine, Microscopy Services Lab, University of North Carolina). Fluorescence was visualized using a Nikon Microphot FXA fluorescent microscope with triple-band-pass filter cube containing UV, rhodamine, and fluorescein filters (XF-63; Omega Optical Inc., Brattleboro, VT).
CM Studies.
To examine the effect of soluble stromal factors, medium was conditioned by stromal cells cultured on plastic or embedded in Matrigel, also called basement membrane extract (BME), (Matrigel, growth factor reduced, phenol red-free; Collaborative Biomedical Products, Bedford MA). Stromal cells were plated at 80% confluence on plastic using medium with 4% FBS for cell attachment, and equal numbers of cells/well were embedded in Matrigel. Stromal cells were found previously to be nonproliferative in Matrigel (29)
; thus, the high densities of stromal cells plated on plastic and in Matrigel account for the lack of proliferation of stromal cells in Matrigel. In some studies, a parallel cell line of normal human foreskin fibroblasts (NHF-1) was used as a control and embedded in Matrigel. Cells (normal stromal cells or NHF-1 cells) cultured in Matrigel were prepared as follows. Matrigel was kept at 4°C and was placed undiluted on precooled MilliCell CM filter inserts (0.4 µm, 30 mm diameter; Millipore Corp., Bedford, MA). Stromal cells were trypsinized, pelleted, and then resuspended in undiluted Matrigel at 4°C. Aliquots of 500-µl Matrigel plus stromal cells at 3 x 105 cells/well were evenly spread over the cold inserts and allowed to gel at 37°C for 1–2 h before culture medium was added to each well.
For studies involving steroid hormones, the culture medium was supplemented with 2% charcoal-stripped serum and E, EP, or EPR (as above). Solvent controls (O) included the medium placed on stromal cells on plastic or in Matrigel containing 0.01% ethanol. Cultures were treated with solvent control or hormones for up to 10 days, replacing the medium every 3–4 days. CM was obtained by removing the medium containing hormones and replacing with serum-free, hormone-free culture medium for 24–48 h. The stromal CM was collected, filter-sterilized to remove cellular debris, and stored at -20°C.
Cocultures of stromal and Ishikawa cells were prepared using early passage stromal cells (passage 1–4). Typically, stromal cells were cultured on plastic or embedded in BME (as described above). Stromal cells were plated at 105 cells/well of a 24-well plate (2 cm2) or in BME on MilliCell CM filter inserts (0.4 µm, 12 mm diameter) in triplicate for each data point. Parallel filter inserts were prepared in advance for Ishikawa MCs using the undiluted BME without added stromal cells. Serum-supplemented medium was gently added at 1 ml/well after 1 h. Stromal cells were allowed to incubate overnight. The next day, the Ishikawa cells were seeded at 5 x 104 cells/well onto the filter inserts containing BME alone or BME plus stromal cells. The medium was replaced with the serum-reduced medium, used for CC studies, as described above.
[3H]Thymidine Incorporation.
Tritiated thymidine ([3H]thymidine) incorporation was used to assess cell proliferation. Ishikawa cells were plated in triplicate in MC or CC, as described previously. After 4 or 7 days of growth, [3H]thymidine was added in culture medium at a final concentration of 2.5 µCi/ml. After 24 h, the medium was removed. Cultures were rinsed twice with HBSS, and trichloroacetic acid (5%) was added to each culture well at 4°C for 20 min to fix cells to plate. The plates were rinsed three times with 5% trichloroacetic acid to remove unincorporated radioactive material. The cells were solublized with 0.2 M NaOH for 30 min at 37°C before neutralization with 75 mM HCl. The cell solution was added to scintillation vials and counted for 10 s in a Packard Scintillation Counter (Meriden, CT). Thymidine incorporation was standardized to total cell counts (cpm/cell) or to DNA concentration (fmol thymidine/mg DNA). DNA concentrations were determined by isolating DNA from parallel wells using DNAzol (Molecular Research Center, Cincinnati, OH). Isolated DNA was quantified by spectrophotometric absorbency at 260 nm.
Colony Formation.
To determine the effects of medium conditioned by stromal cells on Ishikawa colony growth, Ishikawa cells in culture medium were plated on plastic substrate at clonal density of 500 cells/well of 12-well dish (4 cm2) or 2000 cells/6-well dish (9.5 cm2). After 24 h, culture medium was removed and medium conditioned by stromal cells (CM; see "Materials and Methods," below) was added to each well at a ratio of 2 parts CM to 1 part CC medium plus 1% charcoal-stripped serum. Cultures were grown for 6–7 days, then dishes were fixed and stained with methylene blue (Sigma). Colonies of >75 cells/colony were counted as positive.
Glycodelin IFMA.
To examine the effect of stromal cells on Ishikawa cell differentiation, Ishikawa cells were plated alone or in CC with normal endometrial stromal cells either on plastic or in Matrigel as described previously (see illustrations of experimental configurations in Fig. 6
). Hormone treatments included either no added hormones (ethanol control) or EP (10-8 M, 10-6 M, respectively). Medium was replaced, and supernatant medium was collected on days 5 and 9 and stored at -80°C. Medium was assayed for glycodelin using an IFMA. The IFMA used a monoclonal antibody to glycodelin for coating the microtiter wells and secondary labeling with europium III chelate, as described previously (33)
. The sensitivity of IFMA is more than 25 times greater than that of RIA and allows detection and accurate quantitation of glycodelin in samples undetectable by RIA. Experimental samples of 25 µl of cell culture supernatants were added to microtiter plates precoated with antibodies to glycodelin. Plates were incubated overnight at RT. Samples were washed, and the secondary antibody, europium III chelate, was added to each well and incubated at RT for 2 h. The wells were washed, and a fluorescence enhancement solution containing 2-naphtoyltrifluoroacetone and tri-n-octylphosphine oxide was added. Fluorescence was measured using a fluorometer. Results are presented as ng/ml means of triplicate determinations.
|
Reverse Transcription Protocol.
For reverse transcription of RNA, 1 µg of RNA sample was added to individual 200-µl microfuge tubes. Amounts of added DEPC water were adjusted to keep concentrations of RNA consistent between samples. Master mix was prepared for a final volume of 20 µl/tube containing Rnasin RNase inhibitor (20 units; Promega), reverse transcriptase (M-MLV, 200 units), random primers (3 µg), DTT (10 mM) 5 x buffer, and dNTP (1 mM; all from Life Technologies, Inc.). Samples were incubated at 37°C for 45 min then 90°C for 5 min. cDNA samples were stored at -20°C.
PCR Protocol.
PCR amplification of cDNA was performed by adding 2 µl of cDNA sample (0.5 µg/ml) to individual 200-µl microfuge tubes. Master mix was prepared for a final volume of 50 µl/sample including Taq polymerase (1.25 units), 10 x buffer (100 mM Tris HCl (pH 8.3), 500 mM KCl, 15 mM MgCl2, 0.1% gelatin), and MgCl2 (1.5 µM; all from Boehringer Mannheim, Indianapolis, IN), DEPC water (32.6 µl/50 µl), Tergitol NP40 (0.15 µl/50 µl; Sigma Chemical Co.), 3'- 5' primers (0.5 µl each/50 µl, synthesized by Lineberger Cancer Center Nucleic Acid Core Facility, Chapel Hill, NC), and dNTP (0.2 mM; Life Technologies, Inc.). Master mix was aliquoted at 44 µl/tube, and mineral oil was layered on top of reaction volume at 25 µl/tube. PCR cycler (Perkin-Elmer, Foster City, CA) was prewarmed to 85°C. PCR samples were placed in cycler, dNTP (4 µl, 2.5 mM) added as a "hot start," tubes were closed, and primer program was initiated. On completion of cycles, samples were stored at 4°C before agarose gel analysis. The PCR primer sets along with optimal cycle programs used for stromal cell characterization are detailed in Table 1
.
|
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. On BME, the cells migrated together within 24 h, despite the fact that they were seeded as individual dispersed cells (Fig. 1b)
. With time, the colonies formed glandular-like tubular and spherical structures reminiscent of the glandular structure of the tumor from which the Ishikawa cells were derived.
|
. When cultured on Matrigel, stromal cells had a distinct appearance of isolated stellate cells with irregular cell borders (Fig. 1d)
.
In addition to the distinctions in morphology between the two cell types in CC, we studied the interactions between stromal and Ishikawa cells by immunofluorescent staining using antibodies to cytoskeletal proteins that distinguish between stromal cells and epithelial cancer cells. Stromal cells were identified using an antivimentin antibody and secondary FITC-labeled mouse IgG antibody (green). Ishikawa cells stained positively using an epithelial-specific anticytokeratin antibody with Texas Red-labeled secondary rabbit IgG antibody. Using this immunohistochemical technique to localize the respective cell types, the stromal cells (green) in BME were found to aggregate and intersperse between the three-dimensional structures of the Ishikawa cells (red; Fig. 2, a–c
). These photomicrographs document how the Ishikawa cells, having been plated as single cells on top of the gel of Matrigel, appear to migrate into the gel, coalescing to form dense structures where bridges of cells create honeycombed patterns (Fig. 2b)
. The stromal cells also having been plated in homogeneous lawn of cells in the Matrigel, appear to congregate or migrate toward the epithelial structures (Fig. 2a)
, and surround them making contact via cellular processes (Fig. 2, b and c)
. This cell-specific migration and association of stromal cells with Ishikawa cells did not occur when plastic was used as a substrate for CC (Fig. 2d)
. On the plastic culture plate, the cells randomly intermixed on the culture plate suggesting that BME mediates a cellular behavior that favors specific cell-cell interaction.
|
. The CM from stromal cells cultured on plastic had no apparent inhibitory effect. These findings suggest that normal endometrial stromal cells exert a paracrine inhibitory influence on the Ishikawa cells that is dependent on exposure to BME components.
|
, Ishikawa colony counts were reduced by 29% when CM was added from stromal cells cultured on plastic (P < 0.01) compared with unconditioned medium controls. When CM from stromal cells embedded in BME was added to Ishikawa cultures, the colony counts were inhibited by 82% (P < 0.001).
|
This CM was then used to determine colony formation in Ishikawa cells. When Ishikawa cells were treated in the absence of stromal cell CM, there were no significant changes noted in cell proliferation in response to E, EP, or EPR (Fig. 5)
compared with solvent control (O). When Ishikawa cells were treated with CM from hormone-treated stromal cells grown in BME plus fresh aliquots of the same hormones, a 2.7-fold increase in colony numbers was seen in response to E (plus stromal factors; P < 0.001). The addition of EP plus EP-treated stromal medium diminished the cell colony count to 64% compared with E alone (P < 0.05), whereas EPR reversed the EP inhibitory effect, bringing the colony counts back to the level of E alone (P < 0.05). The CM from steroid-treated NHF-1 cells showed no growth inhibition or hormonal regulation of Ishikawa cell colony formation. Instead, the CM from the NHF-1 cells was shown to increase Ishikawa colony counts compared with unconditioned medium controls, the colonies being too numerous to count (one representative count shown per group). These data indicate that factors from hormone-treated endometrial stromal cells restore hormone growth responsiveness to the Ishikawa cell line. This hormone-induced growth pattern was repeatedly found in multiple experiments, using not only stromal CM but also direct CC of stroma and Ishikawa cells (data not shown).
|
, Ishikawa cells growing alone in MCs on either plastic or BME showed little detectable glycodelin expression. Glycodelin was not expressed in Ishikawa cells growing on BME and cocultured with stromal cells on plastic. However, when Ishikawa cells were cocultured in contact with stromal cells in BME, glycodelin expression was induced to 3.4 ng/ml (P = 0.02) and to 4.7 ng/ml (P = 0.0002) with EP treatment. These results were replicated at day 9 and in triplicate experiments. These data suggest for the first time that CC with stromal cells under the proper conditions induces expression of a differentiated endometrial gene product in an endometrial cancer cell line. Such a result implies that some normal function can be restored in these neoplastic cells when they are cultured in the presence of stromal factors and ECM. The growth, morphology, and function of stromal cells on Matrigel were strikingly distinctive compared with stromal cells on plastic. These differences indicate potential involvement of the integrins, the receptors for ECM. To determine gene expression of integrins, we collected RNA from stromal cells growing on plastic or within BME, in the presence or absence of E or EP, and analyzed gene expression by RT-PCR. These samples were the same ones used in CM experiments described above and were shown to be effective in regulation of Ishikawa growth and hormonal response.
The expression of integrin subunits 
2, 3, 4, 6, and ß3 by RT-PCR is shown in Fig. 7
and is standardized against actin. One notable finding was that the laminin-collagen receptor 2 was decreased with E and EP when stromal cells were cultured on tissue culture plastic. Less difference was seen in stromal cells in BME with only a slight decrease in the EP-treated cells. The 3ß1 integrin, a receptor for laminin, collagen, and fibronectin, remained consistent in all of the hormone treatments and substrates as assessed by RT-PCR of the 3 subunit. The fibronectin receptor -4 subunit was suppressed by EP in stromal cells when cultured on in both plastic and BME. The laminin receptor 6ß1 was assessed through the 6 subunit. Expression was decreased in control and E treatment when stromal cells were cultured on plastic and was absent with EP treatment. Stromal cells cultured in BME maintained this integrin except for diminished expression with EP treatment. The fibronectin and vitronectin receptor subunit ß-3 was present in all of the conditions but showed increased expression in E-treated stromal cells when cultured in BME.
|
. Increased expression of PR was noted in BME culture compared with plastic independent of hormone treatment. ER expression increased with E treatment and decreased with EP treatment in both matrix conditions.
|
. TGF-ß1 was present in stromal cells on BME treated with EP but was diminished in stromal cells on plastic. PDGF appeared to be slightly increased in stromal cells in BME. |
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The present study examined the effect of stromal cells and Matrigel on the regulation of endometrial cancer cell (Ishikawa) growth and differentiation using a CC method with normal endometrial stromal cells or medium conditioned by stromal cells. The Ishikawa cell line used in this study was first developed by Nishida et al. (35) and contains functional ERs and PRs, as well as most of the endometrial epithelial integrins found in normal endometrial epithelium (30 , 31 , 36) .
Morphological observations indicated a dynamic interaction between the stromal cells and the Ishikawa cells as shown in the immunofluorescent photomicrographs. In the combined CC both cell types within the structures appear to be embedded within the Matrigel with proliferation suppressed in both stromal and Ishikawa cells. When cocultured in the Matrigel, it is likely that stromal (and Ishikawa) cells remodel the ECM around them. It appears that stromal cells migrate out of the Matrigel and assume positions in close contact with the Ishikawa cells. This chemotactic tendency is an interesting phenomenon and one that we plan to additionally investigate.
Ishikawa colony counts and [3H]thymidine incorporation were repeatedly and significantly inhibited in the presence of CM from stromal cells in contact with BME compared with proliferation in the absence of stromal factors. CM from stromal cells cultured on plastic was partially inhibitory to Ishikawa colony growth but showed no inhibition of [3H]thymidine incorporation in Ishikawa cells. These data demonstrate that not only do stromal cells modulate the growth of neoplastic cells in vitro but also that stromal cells in contact with basement membrane proteins provide a greater inhibitory effect on Ishikawa cells growth than the stromal cells not connected to basement membrane. These data suggest that in the endometrium, subepithelial stromal cells may produce unique paracrine factors or different quantities of paracrine factors involved in growth suppression of the epithelial cells compared with factors produced by interstitial stromal cells.
Medium conditioned by stromal cells cultured in the presence of BME not only inhibited the growth of Ishikawa cells but also restored hormonal responsiveness. Steroid hormones E and P, and the antiprogestin R regulated growth of Ishikawa cells in the presence of CM containing diffusible factors released by steroid-treated stromal cells embedded in BME. Steroid hormones had little growth effect on Ishikawa cells in the absence of stromal-induced factors contained in CM. The pattern of growth of the Ishikawa cells with stromal CM reflects the variations in growth of epithelial cells occurring with hormonal changes during the natural endometrial cycle. In the hormonal cycle, the glandular epithelial cells proliferate in the E-dominant phase and exhibit decreased proliferation in response to P in the secretory phase. This pattern of hormonal response was repeated in multiple experiments, which included either stromal CM or direct CC of cells (data not shown). It is important to note that R has multiple complex mechanisms of action. In addition to its functioning as an antiprogestin, it also activates the glucocorticoid receptor. The restoration in colony formation, after addition of EPR, is likely attributable to the negative effect on P and PR activity but could be because of R binding to the glucocorticoid receptor. The R interactions with the glucocorticoid receptor are important, and we are currently working toward a better understanding of glucocorticoid receptors in normal endometrium and Ishikawa cells.
An important point illustrated by these studies was that the hormonal regulation by stromal cells was specific to endometrial stroma. As expected, neither the growth inhibition nor the hormonal response in colony growth was observed when Ishikawa cells were treated with CM from the foreskin fibroblast cell line NHF-1 cultured in BME. This indicated that endometrial stromal cells were unique in producing hormone-induced growth and inhibitory factors. The NHF-1 cells also provided a control for potential competition of nutrients in CM. The Ishikawa colony counts were increased with NHF-1 CM suggesting that this CC protocol supplies sufficient nutrients for both cell types. Thus, Ishikawa cell growth inhibition by stromal cell CM is not attributable to lack of sufficient nutrients. In our model, endometrial stromal cells cultured in BME were unique in that they provided regulation of Ishikawa proliferation via paracrine factors yet to be identified.
The stromal cells cultured in BME in the presence of E and P induced the expression of glycodelin in the Ishikawa cells. Neither neoplastic endometrium nor Ishikawa cells in standard cell culture conditions express glycodelin (37) , suggesting the loss of this differentiated function is an early event in the transformation process. Hackenberg et al. (38) reported glycodelin expression by another endometrial cell line derived from a poorly differentiated endometrial cancer and from several primary tumors. It is interesting to speculate that regulated expression of glycodelin in the well-differentiated Ishikawa cells represents a higher level of cellular organization requiring stromal paracrine signals. It is perhaps fortuitous that these cells, which maintained regulated expression of other hormone-responsive genes (30 , 36) , demonstrated the role of stromal cells in regulating specific gene products.
Glycodelin expression appears to be complex. Whereas considered to be a P-induced protein, its expression is delayed relative to the onset of P secretion, and the timing of glycodelin does not correlate well with serum P levels (39) . Investigators have reported no difference in expression in the endometrium of women with "out-of-phase" endometrium compared with "in-phase" cycles (40) . Recently the peptide hormone relaxin, which is derived from the corpus luteum as well as the endometrium, has been shown to induce glycodelin (39) . Because stromal cells are a target cell for relaxin (41 , 42) , perhaps the effect we observed is mediated indirectly by relaxin through the stromal component of our CC model.
The induction of Ishikawa glycodelin secretion by stromal cells in BME suggests for the first time that CC with stromal cells under the proper conditions may induce differentiation an endometrial cancer cell line. These results suggest that some normal functioning of neoplastic cells remains possible and may be restored when stromal factors and ECM are present. The lack of glycodelin expression in neoplastic endometrium and Ishikawa cells in standard culture conditions may reflect altered gland-stromal interactions that occur in endometrial tumors. Furthermore, these data support the concept that loss of normal epithelial-stromal cell contact and/or loss of basement membrane may contribute to the abnormal phenotype of cancer cells in vivo. This change in phenotype toward a more normal phenotype may have broad implications for the treatment of endometrial cancers in the future.
The initial characterization of molecular differences in stromal cells in plastic and BME included RT-PCR for integrins, steroid receptors, the inhibitory factor TGF-ß1, as well as an important stimulatory factor, PDGF. Stromal cultures were treated with estradiol (E) or EP to represent proliferative and secretory cycle variations, respectively. Whereas the stromal cell morphology varied greatly between plastic and BME, these gene expression data suggest several subtle differences but reveal no striking differences between stromal cells on these substrata. This result is surprising given the great difference in morphology and effects of the stromal culture conditions on epithelial growth. Cycle-specific changes in integrins found in vivo by Lessey et al., (43)
were reflected in the EP suppression of 2, 4, and 6 in stromal cells on both substrates. The increased ß3 expression in E-treated stromal cells in BME is in contrast with the regulation of vß3 in endometrial epithelial cells (44)
.
The steroid receptors ER and PR were studied by RT-PCR. In the normal endometrium, E induces an increase in stromal ER during the proliferative phase, whereas the P in the secretory phase down-regulates ER (45) . This regulation of ER was reflected in in vitro cultures of stromal cells grown on both growth substrates. Estrogen also induces PR in vivo in the proliferative stage, but contrary to ER, the stromal cell PR levels are maintained during the secretory stage (45) . The stromal cells in culture with BME maintained PR expression compared with stromal cells cultured on plastic independent of steroid treatment.
Stromal cell expression of TGF-ß1, TGF-ß2, and PDGF was analyzed by RT-PCR. These growth factors were chosen as potential mediators of stromal paracrine effect based on previous studies (46, 47, 48) . Also, TGF-ß was shown to be involved in P-induced stromal regulation of epithelial metalloproteases (21) . Similar to the RT-PCR results with integrins, the expression of TGF-ß1 and PDGF showed more sensitivity to steroid presence than to substrate. TGF-ß1 showed increased expression with E treatment in both substrates. The lack of expression of TGF-ß in EP-treated stromal cells on plastic may contribute to the loss of regulation of epithelial growth by this culture condition. It is not clear from these studies whether TGF-ß produced by the stromal cells is present in the latent or active form (49) . It is possible that stromal cells could produce an additional protease involved with the activation of TGF-ß (50) resulting an inhibitory signal to the epithelial cells.
This screening of gene expression of known mediators was a preliminary step in determining factors involved in stromal paracrine regulation. There was not any single integrin, growth factor, or hormone receptor that had markedly different expression in stromal cells when grown in BME. These results reflect the complex "vocabulary" of signaling that takes place between integrin receptors, hormone receptors, and growth factor receptors. (3
, 51)
. It may be that the combination of increased expression of the -6 integrin, PR, and TGF-ß1 found in stromal cells in Matrigel treated with EP are necessary for the growth inhibition and induction of differentiation.
The mechanisms of stromal cell inhibition of epithelial cell growth and induction of differentiation may also include the proteolytic release by the stromal cells of growth (or inhibitory) factors present in the Matrigel. This is a very important potential mechanism, especially as the ECM is a storehouse of multiple peptide growth factors (52) . The hormonal induction of stromal metalloproteases plays a critical role in the cyclical growth and sloughing of the endometrium (53) . Additional studies will determine not only the involvement of this mechanism of stromal regulation but also what potential growth or inhibitory factors are involved.
Disruption of the delicate balance of signaling results in dramatic changes in the way cells interact with each other and with the ECM. Bissel and Hall (54) hypothesized that the unit of function in higher organisms is not the individual cell but the tissue itself. Tissue architecture is intimately involved in controlling the signaling processes within cells. Future studies will provide a better understanding of tissue architecture and may reveal new concepts of tissue function.
This work has revealed an additional layer of complexity in cell-cell communications by demonstrating the role of ECM in regulating the paracrine function of stromal cells. This provides a hint that stromal cells immediately next to the basement membrane, the "subepithelial stroma," may function differently than the interstitial stromal cells more remote from the basement membrane, which are in contact with other collagens and fibronectin matrix proteins. Indeed, a similar cellular subpopulation was reported for prostate periepithelial stromal cells that, in the proximity of the basement membrane, had the lost phenotypic expression of an androgen receptor in the progression of prostate cancer (55) .
The concept of stromal cell association with basement membrane proteins provides insights into potential mechanisms involved in neoplastic progression. During tumor metastasis, interruptions are commonly observed in the basal lamina. Abnormal morphology of the basal lamina in some tumors is one indication that stromal-epithelial interactions are perturbed (12) . If stromal cells lose their basement membrane-associated regulating functions as the basement membrane is altered in neoplastic lesions, they may also lose the ability to synthesize appropriate inhibitory factors, thus promoting progression of preneoplastic epithelial lesions.
The results presented here are consistent with data we published previously using normal endometrial epithelial cells (29) . Very similar results were found including stromal-induced inhibition of growth and induction of glycodelin in normal epithelial cells using the same CC configuration. The predominant difference between the CC of normal endometrial cells and the endometrial cancer cells, as described in the current report, was the inability to document a stromal-induced hormonal growth response in normal endometrial cells. Data from primary cells had shown a modest effect by stromal cells on hormonal-induced changes in growth patterns that were repeatable but never strongly significant. This could be attributable to variations in lots of primary cells from their previous hormonal milieu. Growth of the Ishikawa cells was reproducibly and significantly modulated by the stromal factors and hormones.
The development of CC models such as the one presented here is timely and urgently needed if we are to understand the course of cancer induction and progression. Many investigators are recognizing the importance of stroma in the epithelial neoplastic process. It is necessary to rebuild the parts of a tissue to understand cancer as a sum of the interactions of each component. This dilemma in cancer research was stated by Smithers: "Reduction gains precision about parts but at each step loses information about the larger organization it leaves behind" (56) . The CC model presented here represents a tool to explore the next layer of complexity of cellular and tissue functioning. It provides the opportunity to examine communications within the epithelial-stromal-ECM axis between normal and neoplastic tissues and to elucidate how abnormalities in these communications may contribute to disease.
|
ACKNOWLEDGMENTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
FOOTNOTES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1 This work was supported in part by NIH Grant CA31733 (to D. G. K.) and by National Institute of Child Health and Human Development/NIH through cooperative agreement U54 HD-35041 as part of the Specialized Cooperative Centers Program in Reproduction Research, the National Cooperative Program on Markers of Uterine Receptivity for Blastocyst Implantation HD 34824 (to B. A. L.). M. S. was supported by grants from the Finnish Cancer Foundation and the Federation of Finnish Life and Pension Insurance Companies. 
2 To whom requests for reprints should be addressed, at National Center for Complementary and Alternative Medicine/NIH, 8 West Drive MSC 2669, Qtrs. 15B1, Bethesda, MD 20892-2669. E-mail: jarnold{at}mail.nih.gov
3 The abbreviations used are: E, estrogen; ECM, extracellular matrix; BME, basement membrane extract; IFMA, immunofluorometric assay; CC, coculture; CM, conditioned medium; DAPI, 4'6-diamindino-2-phenylindole; ER, estrogen receptor; PR, progesterone receptor; PDGF, platelet-derived growth factor-A; FBS, fetal bovine serum; MC, monoculture; NHF-1 normal human fibroblast cells; TGF, transforming growth factor; dNTP, deoxynucleotide triphosphate; DEPC, diethyl pyrocarbonate; RT-PCR, reverse transcription-PCR; RT, room temperature; R, RU486; EP, estrogen plus progesterone; P, progesterone; EPR, estrogen, progesterone, and RU486; DMEM-H, Dulbecco’s Modified Eagles Medium-High Glucose; BSS, Balanced Salt Solution.
Received 4/ 7/01. Accepted 11/ 1/01.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1 integrin. J. Steroid Biochem. Mol. Biol., 59: 31-39, 1996.[Medline]
This article has been cited by other articles:
|
|
 |
|
  M Seppala, H Koistinen, R Koistinen, L Hautala, P C Chiu, and W S Yeung Glycodelin in reproductive endocrinology and hormone-related cancer Eur. J. Endocrinol., February 1, 2009; 160(2): 121 - 133. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Barbier, H. J. Kloosterboer, and D. G. Kaufman Effects of Tibolone Metabolites on Human Endometrial Cell Lines in Co-culture Reproductive Sciences, January 1, 2008; 15(1): 75 - 82. [Abstract] [PDF]
|
|
|
|
 |
|
 K. Ohta, T. Maruyama, H. Uchida, M. Ono, T. Nagashima, T. Arase, T. Kajitani, H. Oda, M. Morita, and Y. Yoshimura Glycodelin blocks progression to S phase and inhibits cell growth: a possible progesterone-induced regulator for endometrial epithelial cell growth Mol. Hum. Reprod., January 1, 2008; 14(1): 17 - 22. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Seppala, H. Koistinen, R. Koistinen, P.C.N. Chiu, and W.S.B. Yeung Glycosylation related actions of glycodelin: gamete, cumulus cell, immune cell and clinical associations Hum. Reprod. Update, May 1, 2007; 13(3): 275 - 287. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Uchida, T. Maruyama, M. Ono, K. Ohta, T. Kajitani, H. Masuda, T. Nagashima, T. Arase, H. Asada, and Y. Yoshimura Histone Deacetylase Inhibitors Stimulate Cell Migration in Human Endometrial Adenocarcinoma Cells through Up-Regulation of Glycodelin Endocrinology, February 1, 2007; 148(2): 896 - 902. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. N. Jabbour, R. W. Kelly, H. M. Fraser, and H. O. D. Critchley Endocrine Regulation of Menstruation Endocr. Rev., February 1, 2006; 27(1): 17 - 46. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. S. Barbier, K. A. Becker, M. A. Troester, and D. G. Kaufman Expression of Exogenous Human Telomerase in Cultures of Endometrial Stromal Cells Does Not Alter Their Hormone Responsiveness Biol Reprod, July 1, 2005; 73(1): 106 - 114. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Blauer, P.K. Heinonen, P.M. Martikainen, E. Tomas, and T. Ylikomi A novel organotypic culture model for normal human endometrium: regulation of epithelial cell proliferation by estradiol and medroxyprogesterone acetate Hum. Reprod., April 1, 2005; 20(4): 864 - 871. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. S. Grant-Tschudy and C. R. Wira Hepatocyte Growth Factor Regulation of Uterine Epithelial Cell Transepithelial Resistance and Tumor Necrosis Factor {alpha} Release in Culture Biol Reprod, April 1, 2005; 72(4): 814 - 821. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. T. Arnold, H. Le, K. K. McFann, and M. R. Blackman Comparative effects of DHEA vs. testosterone, dihydrotestosterone, and estradiol on proliferation and gene expression in human LNCaP prostate cancer cells Am J Physiol Endocrinol Metab, March 1, 2005; 288(3): E573 - E584. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. J. Jakubowicz, P. A. Essah, M. Seppala, S. Jakubowicz, J.-P. Baillargeon, R. Koistinen, and J. E. Nestler Reduced Serum Glycodelin and Insulin-Like Growth Factor-Binding Protein-1 in Women with Polycystic Ovary Syndrome during First Trimester of Pregnancy J. Clin. Endocrinol. Metab., February 1, 2004; 89(2): 833 - 839. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Mandelin, H. Lassus, M. Seppala, A. Leminen, J.-A. Gustafsson, G. Cheng, R. Butzow, and R. Koistinen Glycodelin in Ovarian Serous Carcinoma: Association with Differentiation and Survival Cancer Res., October 1, 2003; 63(19): 6258 - 6264. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. I. Risinger, G. L. Maxwell, G. V. R. Chandramouli, A. Jazaeri, O. Aprelikova, T. Patterson, A. Berchuck, and J. C. Barrett Microarray Analysis Reveals Distinct Gene Expression Profiles among Different Histologic Types of Endometrial Cancer Cancer Res., January 1, 2003; 63(1): 6 - 11. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Seppala, R. N. Taylor, H. Koistinen, R. Koistinen, and E. Milgrom Glycodelin: A Major Lipocalin Protein of the Reproductive Axis with Diverse Actions in Cell Recognition and Differentiation Endocr. Rev., August 1, 2002; 23(4): 401 - 430. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
