BIOLOGY OF REPRODUCTION 52, 609-615 (1995)
Fibroblast Cells from Term Human Decidua Closely Resemble Endometrial Stromal
Cells: Induction of Prolactin and Insulin-Like Growth Factor Binding Protein-1
Expression'
RANDALL G. RICHARDS, 2 ANOOP KICBRAR, GRAEME R. FRANK, SCOTT M. HARTMAN, and HIROAKI JIKIHARA
Division of PediatricEndocrinology and the PerinatalResearch Institute, Children'sHospital and the
University of CincinnatiMedical Centers, Cincinnati, Ohio 45229-3039
ABSTRACT
Term decidual tissue consists of terminally differentiated decidual cells, bone marrow-derived cells, and fibroblast cells. Since
undifferentiated decidual cells are fibroblast-like cells of the endometrial stroma, the possibility exists that the fibroblast cells in
term decidua are undifferentiated decidual cells. To test this hypothesis, a purified population of fibroblast cells was isolated
from term decidua and treated under conditions that are known to induce differentiation of endometrial stromal cells. By flow
cytometry and immunocytochemistry, the fibroblast cells from term decidua were shown to be free of cells expressing bone
marrow-derived cell-surface antigens and the epithelial cell marker cytokeratin. In addition, they tested positive for the cytoskeletal protein vimentin, thus establishing that they were mesenchymal cells. As with endometrial stromal cells, continuous treatment of the decidual fibroblast cells with the progesterone analog medroxyprogesterone acetate and estradiol in combination
with either dibutyryl-cAMP or prostaglandin E, induced cell aggregation and the expression of prolactin (PRL) and insulin-like
growth factor-binding protein-i (IGFBP-1). When cells were plated at an initial cell density of 0.25 x 106 cells/well in a 24well culture dish with medium changes every three days, PRL was first detected on Days 4-6, and the peak of averaged 24 h-PRL
release (30 ng/well) occurred on Days 26-28. The mRNA for decidual fibroblast PRL followed a temporal pattern corresponding
to that of the released hormone. The size of the PRL mRNA was 1.15 kb, corresponding to the alternately spliced PRL mRNA
reported for decidualized endometrial stromal cells and other extrapituitary sources of PRL In contrast to PRL release, IGFBP1 release was first observed during treatment Days 1-3, and high levels of IGFBP-1 were then released into the medium from
Day 4 through at least Day 12. A significant decrease in IGFBP-I release was first observed during treatment Days 17-19, and
the levels of IGFBP-1 then gradually decreased. Release of IGFBP-1 was accompanied by a corresponding increase in IGFBP-1
mRNA. However, whereas the protein levels diminished by Day 19, the mRNA levels appeared to remain elevated until Day 24.
The changes in morphology and the induction of PRL and IGFBP-1 expression in decidual fibroblast cells closely resembled the
response of endometrial stromal cells cultured under similar conditions. The results suggest that the decidualized uterine endometrium maintains a proliferating population of cells that can be recruited for differentiation to the decidual cell phenotype
throughout pregnancy.
INTRODUCTION
Decidual tissue isolated at term is considered to be fully
differentiated. Vince et al. reported that 47% of the cells in
term decidua are of bone marrow origin, comprising 18%
macrophages, 3% large granular lymphocytes, and 8% T cells,
with the remaining portion of the bone marrow-derived
cells being CD-16 positive granulocytes [1]. However, in
addition to bone marrow-derived cells, at least two other
cell-types are present in term decidua, the terminally differentiated endometrial stromal cells (decidual cells) and
fibroblast cells. Although a great deal is known about the
bone marrow-derived cells and the decidual cells in term
decidua, the fibroblast cells have never been systematically
characterized. In light of the fact that undifferentiated decidual cells are fibroblast-like cells of the endometrial stroma
[2], the possibility exists that the fibroblast cells in term decidua are undifferentiated decidual cells.
In vitro studies of decidual cell differentiation have focused on the differentiation of stromal cells isolated from
the proliferative and secretory endometrium. In pure primary culture, these cells have a morphological appearance
similar to that of fibroblasts; they do not express bone marrow-derived cell-surface antigens, and they stain positive for
the cytoskeletal protein vimentin and negative for the epithelial cell marker cytokeratin ([2-6] and unpublished data).
Decidualization of endometrial stromal cells in vitro occurs
under a variety of conditions utilizing steroid hormones or
growth factors alone and in combination. Factors that promote the differentiation of endometrial stromal cells include progesterone [7, 8], estradiol (E2) [8], epidermal growth
factor [9], prostaglandin E2 (PGE 2) [10], cAMP [11], relaxin
[8,12], insulin-like growth factor-1 and -2 [13], androgens
[14], and prolactin (PRL) [15]. PRL and insulin-like growth
factor binding protein-1 (IGFBP-1) are two proteins expressed by decidual cells that are not expressed by endometrial stromal cells. Both proteins have been used as
markers of endometrial stromal cell differentiation
[3,7, 16,17].
In this study, fibroblast cells were isolated from term
decidua and characterized with respect to morphology, cell
Accepted October 27, 1994.
Received May 31, 1994.
'Supported in part by NIH Grant HD-15201.
2Correspondence: Division of Endocrinology, Children's Hospital Medical Center, 3333 Burnet Avenue, Cincinnati, OH 45229-3039. FAX: (513) 559-7486.
609
610
RICHARDS ET AL.
surface antigens, cytoskeletal proteins, and their expression
of PRL and IGFBP-1 in response to decidualizing agents.
The results indicate that the fibroblast cells in term decidua
closely resemble endometrial stromal cells, suggesting that
the decidualized uterine endometrium of pregnancy maintains a proliferating population of cells that can be recruited for differentiation to the decidual cell phenotype.
MATERIALS AND METHODS
Isolation and Culture of Term Decidua Fibroblast Cells
Term human placenta from cesarean section or vaginal
delivery following uncomplicated pregnancies were obtained with Institutional Review Board approval from the
Christ Hospital, Cincinnati, OH, and the University of Cincinnati and Children's Hospital Medical Centers, Cincinnati,
OH. The isolation of cells from decidua parietalis tissue dissected from the fetal membranes was begun within 1-2 h
after delivery. Collagenase-dispersed cells were isolated as
described previously [18] with the addition of a 5-min 37°C
incubation in ammonium chloride solution (0.8 g tissue
culture grade NH 4CI, 0.084 g NaHCO 3, and 0.037 g EDTA
in 100 ml H20) to lyse red blood cells. The isolated cells
were plated on plastic (107 cells/25 cm2 surface area) and
cultured in RPMI 1640 medium (containing 50 mM Hepes,
2% NaHCO 3, 25 U/ml penicillin G, 25 ptg/ml streptomycin,
and 5 pzg/ml amphotericin B, pH 7.2) supplemented with
10% fetal bovine serum (10% FBS). After 24 h, the nonadherent cells (which included most of the terminally differentiated decidual cells and bone marrow-derived cells)
were freed by agitation and removed by a total exchange
of medium. The cells were grown to confluence and subpassaged. By subpassage 3, essentially all the cells were
proliferating fibroblasts. Cells from more than 3 subpassages were used in these studies. For experiments designed
to monitor proteins released into the medium, cells were
initially plated at a density of 0.25 x 106 cells/well in 24well culture dishes and maintained in 1 ml of medium. In
experiments for isolation of mRNA, cells were plated at an
initial cell density of 1 x 106 cells/well in 6-well culture
dishes and maintained in 4 ml of medium. The cell cultures
were maintained at 37°C in a humidified atmosphere of 95%
air:5% CO 2.
Decidual fibroblast cells were also frozen in 10% dimethylsulfoxide/10% FBS-supplemented RPMI-1640 medium and stored for up to one year at -1500C. Similar results have been obtained in 8 experiments using cells isolated
from 4 different decidua, before and after storage at -150°C,
and at subpassages ranging from 4 to 10.
Isolation and Culture of Endometrial Stromal Cells
Proliferative endometrial tissue (within the first 14 days
of the menstrual cycle) from normally menstruating women
was obtained by endometrial suction biopsy at the time of
elective tubal ligation through Dr. Marcel Cedars, Department of Obstetrics and Gynecology, University of Cincinnati Medical Center. Informed consent was obtained, and
the study was approved by the Institutional Review Boards
of the University of Cincinnati and the Children's Hospital
Medical Centers, Cincinnati, OH. The tissue was collected
in sterile Dulbecco's Modified Eagle's Medium (DMEM), the
cells were dispersed by collagenase digestion, and the
stromal cells were separated from contaminating epithelial
cells by selective trypsinization, as described previously [10].
Cells were plated at an initial cell density of 0.3 x 106 cells/
well in 6-well culture dishes and cultured in DMEM (containing penicillin G, streptomycin, and amphotericin B as
above, pH 7.2) supplemented with 2% FBS.
Treatment Regimens
Cells were plated in FBS-supplemented medium and allowed to attach overnight before treatments were begun.
Treatments consisted of the iBS-supplemented medium
alone (control), or FBS-supplemented medium containing
1) the steroids medroxyprogesterone acetate (MPA; 1 ,aM)
and E2 (10 nM); 2) the steroids in combination with dibutyryl-cAMP (db-cAMP; 50 ,uM); 3) the steroids in combination with PGE 2 (1 RiM); 4) db-cAMP alone; or 5) PGE2
alone. MPA, E2, and PGE 2 were dissolved in ethanol and
added to the culture medium. The final concentration of
alcohol never exceeded 0.09% (vol/vol). The medium was
changed every third day and stored at -20°C.
Antibody Labeling and Flow Cytometry
Cells (passage 6) were removed from the culture dish
by treatment with 0.03% EDTA in PBS. Cells (1 x 106) were
then washed with PBS containing 0.2% BSA and 0.01% sodium azide. Nonspecific binding sites were blocked by incubation with mouse IgG (Sigma Chemical Co., St. Louis,
MO) for 10 min at 4C. The cells were labeled with 20 Rl
of fluorescein isothiocyanate (FITC)-conjugated mouse
anti-human CD-45 (Parmingen, San Diego, CA), FITC-conjugated mouse anti-human HLA-DR (Parmingen), FITCconjugated HLA-A,B,C (Parmingen), or Phycoeruthrin (PE)conjugated mouse anti-human CD-14 for 30 min on ice. The
control for antibody labeling was FITC- or PE-conjugated
mouse isotype IgG (Parmingen). A total of 10 000 cells were
sorted for each antigen in a FACScan with use of LYSIS II
software (Becton-Dickinson, San Jose, CA). The human
monocyte cell line U-937 was used as positive control.
Immunocytocbemistry
Vimentin and cytokeratin were detected by immunocytochemical staining essentially as described by Wilfinger et
al. [19]. Cells were removed from the plastic culture dish
by treatment with 0.03% EDTA in PBS, cytospun onto glass
slides, and fixed with acetone. The mouse monoclonal antibody to cytokeratin (AE1/AE3 cocktail; BioGenex Labo-
611
UNDIFFERENTIATED DECIDUAL CELLS IN TERM HUMAN DECIDUA
ratories, San Ramon, CA) was used at 1/40 dilution, and
the mouse monoclonal antibody to vimentin (Cat. #814318;
Boehringer Mannheim, Indianapolis, IN) was used at 1/4
dilution, as suggested by their respective manufacturers. Instead of using horseradish peroxidase conjugated second
antibody, a biotinylated anti-mouse IgG was used, followed
by detection with horseradish peroxidase conjugated avidin
(Vectastain Elite ABC Kit; Vector Laboratories, Burlingame,
CA) and color development with 3,3'-diaminobenzidine/
H2 0 2. The choriocarcinoma cell line JAR was used as positive control for cytokeratin.
RIA
PRL was measured by homologous RIA [20] using material provided by the National Hormone and Pituitary Program, NIDDK. Analysis of variance techniques were used to
determine whether the effect of different treatment regimes was similar for cells treated with and without steroids (tested as the interaction term of treatment and steroid). Treatment group, steroid group, and time were
included in the model as independent variables.
Analysis of mRIVA
Total RNA was extracted from cells using Tri Reagent
(Molecular Research Center, Inc., Cincinnati, OH) and precipitated with isopropanol as suggested by the manufacturer. The RNA samples were denatured in formaldehyde/
formamide, separated in 1.5% agarose/formaldehyde gels
(10 Rig/lane), and transferred to a Nytran membrane
(Schleicher and Schuell, Keene, NH) by standard procedures. Equal loading of transferred RNA was verified by
methylene blue staining of ribosomal bands.
The probe for PRL was a Pst I-digested 566-base fragment
of PRL cDNA provided by Dr. Henry Friesen. The probe for
IGFBP-1 was a 482-base fragment encompassing bases 20512 of IGFBP-1 cDNA with a 10-base deletion (bases 155164) provided by Dr. David Clemmons, University of North
Carolina, Chapel Hill, NC. Both probes were labeled with
[ct- 32p]dCTP (3000 Ci/mmol; Dupont-New England Nuclear,
Boston, MA) via random-primed synthesis (Stratagene Corporation, La Jolla, CA) to a specific activity of 5-10 x 108
cpm/pig. Hybridizations were carried out at 420C in 50%
formamide, 1.25 M sodium chloride, 0.075 M sodium citrate, 100 Rg/ml denatured salmon sperm DNA, 0.1% each
polyvinylpyrrolidone, BSA, Ficoll, and SDS. The maximally
stringent wash employed was 0.15 M sodium chloride, 0.015
M sodium citrate, and 0.1% SDS at 65°C. Specifically hybridizing bands were detected by autoradiography. The sizes
of the mRNA bands were determined by using 0.17-1.77
kb and 0.25-9.5 kb RNA ladders (Gibco-BRL, Gaithersburg,
MD).
Western Blot Analysis
Proteins in 35 Il of 72-h conditioned medium were separated in standard 12% SDS polyacrylamide gels [21] and
FIG. 1. Isolated decidual fibroblast cells stained with hematoxylin and
eosin. x168 (original magnification x200).
transferred to a supported nitrocellulose 0.2-,um pore size
membrane (Schleicher and Schuell). Nonspecific binding
sites were blocked by incubation in 5% instant milk. The
polyclonal antibody used to detect human IGFBP-1 was
purchased from Upstate Biotechnology Incorporated, Waltham, MA (#06-106). The first antibody was detected by
chemiluminescence using horseradish peroxidase conjugated donkey anti-rabbit IgG (Chemicon, Temecula, CA) in
combination with the Renaissance detection system (Dupont-New England Nuclear).
300-
-HLA-DR
200-
100Lu
I
I
I
I
10
102
103
104
10
102
10 3
104
z
-J
-J
L
300
)
200
100
INTENSITY
FIG. 2. Detection of cell surface antigens by flow cytometry. Arrows
indicate negative controls of FITC-conjugated isotype mouse IgG (HLA-DR;
CD45; and HLA-A,B,C) or PE-conjugated isotype mouse IgG (CD14).
612
RICHARDS ET AL.
FIG. 4. Morphological changes in primary cultures of endometrial
stromal cells (ESC) and term decidual fibroblast cells (TDFC) cultured in
control medium (A,C,E) or treated with db-cAMP, MPA, and E2 (B,D,F). Endometrial stromal cells are shown at treatment Day 12 (A,B). Decidual fibroblast cells are shown at treatment Days 7 (C,D) and 25 (E,F). x82 (original magnification x100).
RESULTS
Cellular Morphology and Immunocharacterization
FIG. 3. Detection of cytoskeletal proteins by immunocytochemistry. Cells
were cytospun onto glass slides and probed with mouse monoclonal antivimentin (A), mouse monoclonal anti-cytokeratin (B), or nonimmune mouse
IgG (C). x200.
The decidual fibroblast cells used in these studies appeared homogenous when stained with hematoxylin and
eosin (Fig. 1). Analysis of cellular antigens by flow cytometry (Fig. 2) or immunocytochemistry (Fig. 3) indicated
that the cell population expressed the major histocompatibility antigens HLA-A,B,C, and the cytoskeletal protein vimentin. They did not express cell surface antigens typical
of bone marrow-derived cells (CD45, CD14, and HLA-DR)
or the cytoskeletal protein cytokeratin. These results are
summarized in Table 1.
MorphologicalChanges in Response to Steroids and cAMP
TABLE 1. Immunocharacterization of decidual fibroblast cells.
Antigen
CD45
CD14
HLA-DR
HLA-A,B,C
Vimentin
Cytokeratin
Method*
Cell specificity
Reaction
FC
FC
FC
FC
ICC
ICC
Bone marrow-derived cells
Macrophage
Macrophage
Non-trophoblast
Non-epithelial
Epithelial (uterine epithelium, trophoblast)
-
*FC = Flow Cytometry. ICC = Immunocytochemistry.
+
+
-
Endometrial stromal cells can be induced to decidualize
by incubation with the progesterone analog MPA and E2.
The addition of PGE 2 or cAMP to the culture medium markedly enhances the effects of these steroids [10]. In vitro decidualization of endometrial stromal cells is accompanied
by striking changes in morphology characterized by cell aggregation into clusters and the loss of distinct cell borders.
Photographs of endometrial stromal cell cultures on Day
12 treated with and without steroids and dibutyryl-cAMP (dbcAMP) are shown in Figure 4 (A and B).
UNDIFFERENTIATED DECIDUAL CELLS IN TERM HUMAN DECIDUA
B
A
35
35-
30
30-
25
25-
20
20-
Q 15
I
-J
L
I
10
1510-
5-
5
00
613
10
DAYS
1
2'0
30
DAYS INCULTURE
FIG. 5. Release of PRL by endometrial stromal cells and decidual fibroblast cells. All values represent accumulated release of PRL during a
3-day period (average of triplicate wells + SEM). Results are reported as
average daily PRL release and plotted on the last day of each sampling
period. A) Release of PRL by endometrial stromal cells cultured in medium
containing MPA and E2 (solid circles); db-cAMP, MPA, and E2 (solid triangles); or PGE2, MPA, and E2 (solid inverted triangles). B) Release of PRL by
decidual fibroblast cells cultured in control medium (open squares), or in
medium containing MPA and E2 (solid circles); PGE 2 (inverted open triangles); db-cAMP (open triangles); PGE 2, MPA, and E2 (inverted solid triangles); db-cAMP, MPA, and E2 (solid triangles). Values reported for treatment Day 16 represent the average daily release over 4 days instead of 3
days. The values determined for PRL in medium from cells cultured in control medium and in medium containing MPA and E2 were below the limits
of detection throughout the culture period.
Decidual fibroblasts were treated under similar conditions and monitored for changes in morphology. As with
the endometrial stromal cells, marked changes were observed in the decidual fibroblast cultures treated with MPA,
E2, and db-cAMP (Fig. 4, C-F). Decidual fibroblast cells treated
for 7 days with control medium had a long, spindle-shaped,
fibroblast appearance (Fig. 4C) whereas the decidual fibroblast cells treated during the same period with MPA, E2, and
db-cAMP appeared fatter and aggregated into cell clusters
(Fig. 4D). The differences observed between control and
treated cultures were more evident by treatment Day 25.
As shown in Figure 4E, the morphology of individual de-
FIG. 6. Northern blot analysis of PRL mRNA in decidual fibroblast cells
cultured in control medium, or in medium containing db-cAMP, MPA, and
E2 for the indicated days. PIT = mRNA from human pituitary.
FIG. 7. Western blot analysis showing time course of IGFBP-1 released from decidual fibroblast cells cultured in control medium alone; or
in medium containing db-cAMP, MPA, and E2; or in medium containing
PGE 2, MPA, and E2. Each time-point represents the IGFBP-1 accumulated
during a 3-day culture period. The indicated days are the last day of the 3day period. M = 10% FBS-supplemented medium.
cidual fibroblast cells in control medium did not change,
but the cell culture became grossly overgrown, forming large
aggregated cell masses. In contrast, the cells treated with
MPA, E2, and db-cAMP (Fig. 4F) were not overgrown by Day
25, and the aggregated cell clusters observed on Day 7 were
still a major morphological feature. These cultures also had
a smooth texture associated with the loss of visibly distinct
cell borders. Thus, when cultured under conditions that induce decidualization of endometrial stromal cells, the decidual fibroblast cultures progressed through a series of
morphological changes that appeared very similar to those
observed during endometrial stromal cell decidualization.
Expression of Decidual Cell-Marker Proteins
Endometrial stromal cells do not express PRL or IGFBP1 when cultured in FBS-supplemented medium alone.
However, when endometrial stromal cells are treated with
agents that induce decidualization, both proteins are expressed. The graph in Figure 5A shows the induction of PRL
expression in endometrial stromal cells treated for 12 days
with MPA and E2, or by MPA and E2 in combination with
either db-cAMP or PGE2.
Decidual fibroblast cells treated under similar conditions also expressed PRL (Fig. 5B). As with endometrial
stromal cells, no PRL was expressed in the untreated decidual fibroblast cell cultures (Fig. 5B). However, cultures
FIG. 8. Northern blot analysis of IGFBP-1 mRNA in decidual fibroblast
cells cultured in control medium, or in medium containing db-cAMP, MPA,
and E2 for the indicated days. Analysis of corresponding PRL mRNA is shown
in Figure 4.
614
RICHARDS ET AL.
treated with MPA, E2, and db-cAMP or MPA, E2, and PGE 2
were induced to express PRL. Release of prolactin in response to MPA, E2, and db-cAMP was first observed during
treatment Days 4-6. PRL release then continued to increase
through treatment Days 26-28 followed by a gradual decline. A similar temporal profile of PRL release was observed when the cells were treated with MPA, E2, and PGE 2.
Treatment of the decidual fibroblast cells with MPA and E2
in the absence of db-cAMP or PGE2 did not induce PRL
expression; however, treatment of the cells with db-cAMP
alone induced a statistically significant increase in PRL
expression compared to that of controls (p < 0.0001). The
level of PRL expression with db-cAMP alone was approximately 10% of the PRL expression observed with db-cAMP
and steroids together (steroid by treatment interaction significant at p < 0.001 by analysis of variance). In contrast,
no statistical difference in PRL expression was detected between PGE2 -treated cells and controls.
The induction of PRL expression in the decidual fibroblast cells was confirmed by Northern blot analysis of mRNA.
Steady-state levels of PRL mRNA in decidual fibroblast cells
treated with MPA, E2, and db-cAMP followed a temporal
profile similar to the release of PRL into the medium (compare Figs. 6 and 5B). The mRNA was 1.15 kb, corresponding
to the alternatively spliced PRL mRNA found in other extrapituitary sources of prolactin [22].
IGFBP-1 expression was also induced in decidual fibroblast cells treated with MPA and E2 in combination with
db-cAMP or PGE 2. A time course of IGFBP-1 expression in
control cultures and cultures treated with MPA and E in
combination with db-cAMP or PGE2 is shown in Figure 7.
No IGFBP-1 was detected in the FBS-supplemented medium itself, or in the control culture medium after 3, 19,
and 31 days of treatment. However, significant IGFBP-1 release was observed during the first 3-day culture period
when decidual fibroblast cells were treated with MPA and
E2 in combination with db-cAMP or PGE2. High levels of
IGFBP-1 were then released into the medium from Day 4
through at least Day 12. A significant decrease in the release
of IGFBP-1 was first observed during the treatment period
on Days 17-19, followed by a trend of gradually decreasing
IGFBP-1 release throughout the remaining treatment period.
The steady-state level of IGFBP-1 mRNA in decidual fibroblast cells treated with MPA, E2, and db-cAMP peaked on
or before treatment Day 11 and remained at or near maximal levels through treatment Day 24 (Fig. 8). In addition
to the previously reported 1.55 kb band [23], a larger band
of 3.47 kb also hybridized to the IGFBP-1 probe, and its
relative intensity followed that of the 1.55 kb band. The nature of this second band is currently unknown.
DISCUSSION
The results of this study indicate that a population of
undifferentiated decidual cells is present in term decidual
tissue, possessing properties that are similar, but not identical, to those of endometrial stromal cells. Although a direct link cannot be made between endometrial stromal cells
and decidual fibroblast cells, the evidence strongly suggests
that both cell-types are precursor decidual cells arising from
the same cell lineage.
The population of decidual cells used in these studies
possessed a typical fibroblast morphology and did not express bone marrow-derived cell-surface antigens. Although
it would be unlikely that uterine epithelium would be present in term decidua, chorionic trophoblast cells (which are
epithelial cells) could be present in the cultures. However,
no epithelial cells were detected by immunocytochemical
staining of the intermediate filament, cytokeratin. In contrast, all cells stained positive for the cytoskeletal protein
vimentin, a typical marker for fibroblast (mesenchymal) cells.
The presence of the histocompatibility antigens HLA-A,B,C
also indicated that the cells were not chorionic trophoblasts.
In addition to possessing morphological features and
cellular antigens similar to those of endometrial stromal
cells, the decidual fibroblast cells responded to treatment
with db-cAMP, PGE 2, and steroids in a manner almost identical to that of endometrial stromal cells. When the decidual
fibroblast cells were treated with the steroids MPA and E2
in combination with db-cAMP or PGE2, the cells aggregated
and, under phase-contrast microscopy, the cell borders appeared to become less distinct over time. These changes in
morphology were accompanied by the appearance of the
decidual cell marker proteins PRL and IGFBP-1 in the culture medium, and by the detection of the corresponding
mRNAs. The relative effects of db-cAMP and PGE 2 on PRL
and IGFBP-1 expression were also similar for the two celltypes. Dibutyryl-cAMP had a greater effect on PRL and IGFBP1 expression than PGE2, both with and without the addition
of MPA and E2. In the absence of steroids, the level of dbcAMP-induced PRL expression was approximately 10% of
that expressed with steroids present.
Although the cell surface antigens, cytoskeletal proteins,
changes in morphology, and induction of PRL and IGFBP1 expression were found to be almost identical between
the two cell-types, some characteristics were not identical.
In contrast to their response to steroid treatment in combination with db-cAMP or PGE 2, the decidual fibroblast cells
and the endometrial stromal cells did not respond equally
to steroid treatment in the absence of db-cAMP or PGE 2.
Endometrial stromal cells express detectable levels of PRL
and IGFBP-1 in response to steroids alone [10]; however,
decidual fibroblast cells required the presence of db-cAMP
or PGE2 to induce PRL and IGFBP-1 expression. If PGE 2 and
cAMP act via up-regulation of the progesterone receptor as
has been demonstrated in the rat [24], the lower sensitivity
of the decidual fibroblast cells to steroid treatment might
be explained by a significantly lower basal level of progesterone receptor in the decidual fibroblast cells. This hy-
UNDIFFERENTIATED DECIDUAL CELLS IN TERM HUMAN DECIDUA
pothesis is currently under investigation. In addition to their
steroid insensitivity, decidual fibroblasts cultured in the absence of steroids and db-cAMP or PGE2 exhibited a higher
rate of proliferation and a lack of cell contact inhibition,
while treated decidual fibroblasts were characterized by a
longer induction period for PRL expression compared to
endometrial stromal cells (not shown). Taken together, these
observations seem to suggest that the decidual fibroblast
cells are less "primed" to differentiate than the endometrial
stromal cells. In all other respects, however, the two celltypes appeared to be essentially identical.
This study indicates that the decidualized uterine endometrium maintains a population of stem cells that have
the capacity to proliferate or to terminally differentiate in
vitro. It is tempting to conclude from these results that differentiation of the uterine endometrium is a dynamic process that occurs throughout gestation since PGE2 , progesterone, and estradiol are all synthesized and secreted locally
by the decidua and/or placenta [25, 26]. However, since decidual PRL in amniotic fluid reaches a maximum at week
20 and then decreases through the remaining weeks of gestation [27], it is unclear to what extent these decidual fibroblast cells actively differentiate in vivo and how they might
be regulated. It could be that cellular differentiation is under inhibitory control through interactions with the extracellular matrix or local factors secreted from other cells in
the decidua. Future studies will need to specifically address
this question and consider the role of these cells during
gestation.
Finally, it should be noted that the decidual fibroblast
cells used in this study were isolated from a heterogeneous
population of collagenase-dispersed term decidual cells.
Thus, when these preparations are used to study decidual
cell function, the potential influence of undifferentiated decidual cells on the experimental results must be taken into
consideration.
ACKNOWLEDGMENTS
The authors express appreciation to Carry L. Moeykens, George Whitaker, and
Kristin M. Czerwinski for their technical assistance in carrying out these studies, and
to Drs. Edith Markoff and Stuart Handwerger for their helpful suggestions and critical
review of the manuscript.
REFERENCES
1. Vince GS, Starkey PM, Jackson MC, Sargent IL, Redman CWG. Flow cytometric
characterisation of cell populations in human pregnancy decidua and isolation
of decidual macrophages. J Immunol Methods 1990; 132:181-189.
2. Kirk D, Irwin JC. Normal human endometrium in cell culture. Methods Cell Biol
1980; 218:51-77.
3. Irwin JC, Kirk D, King RJB, Quigley MM, Gwatkin RBL. Hormonal regulation of
human endometrial stromal cells in culture: an in vitro model for decidualization. Fertil Steril 1989; 52:761-768.
4. Osteen KG, Hill GA, HargroveJT, Gorstein F. Development of a method to isolate
and culture highly purified populations of stromal and epithelial cells from human endometrial biopsy specimens. Fertil Steril 1989; 52:965-972.
615
5. Lysiak jJ, Lala PK In situ localization and characterization of bone marrow-derived cells in the decidua of normal murine pregnancy. Biol Reprod 1992; 47603613.
6. Martelli M, Campana A, Bischof P. Secretion of matrix metalloproteinases by human endometrial cells in vitro. J Reprod Fertil 1993; 98:67-76.
7. Daly DC, Maslar IA, Riddick DH. Prolactin production during in vitro decidualization of proliferative endometrium. Am J Obstet Gynecol 1983; 145:672-678.
8. Huang Jk, Tseng L, Bishof P, Janne OA. Regulation of prolactin production by
progestin, estrogen and relaxin in human endometrial stromal cells. Endocrinology 1987; 121:2011-2017.
9. Irwin JC, Utian WH, Eckert RL. Sex steroids and growth factors differentially regulate the growth and differentiation of cultured human endometrial stromal cells.
Endocrinology 1991; 129:2385-2392.
10. Frank GR, Brar AK, Cedars MI, Handwerger S. Prostaglandin E2 enhances human
endometrial stromal cell differentiation. Endocrinology 1994; 134:258-263.
11. Tang B, Guller S, Gurpide E. Cyclic adenosine 3',5'-monophosphate induces prolactin expression in stromal cells isolated from human proliferative endometrium. Endocrinology 1993; 133:2197-2203.
12. Tseng L, GaoJ, Chen R, Zhu HH, Mazella J, Powell DR Effect of progestin, antiprogestin, and relaxin on the accumulation of prolactin and insulin-like growth
factor-binding protein-1 messenger ribonucleic acid in human endometrial stromal
cells. Biol Reprod 1992; 47:441-450.
13. Irwin JC, Fuentes Ldl, Dsupin BA, Giudice LC. Insulin-like growth factor regulation of human endometrial stromal cell function: coordinate effects on insulinlike growth factor binding protein-i, cell proliferation and prolactin secretion.
Regulat Pept 1993; 48:165-177.
14. Narukawa S, Kanzaki H, Inouea, Imai K,Higuchi T, Hatayama H, Kariya M, Mori
T. Androgens induce prolactin production by human endometrial stromal cells
in vitro. J Clin Endocrinol &Metab 1994; 78:165-168.
15. Tabanelli S, Tang B, Gurpide E. In vitro decidualization of human endometrial
stromal cells. J Steroid Biochem Mol Biol 1992; 42:337-344.
16. Bell SC, Jackson JA, Ashmore J, Zhu HH, Tseng L Regulation of insulin-like growth
factor-binding protein-1 synthesis and secretion by progestin and relaxin in long
term cultures of human endometrial stromal cells. J Clin Endocrinol & Metab
1991; 72:1014-1024.
17. Giudice LC, Dsupin BA, Irwin JC. Steroid and peptide regulation of insulin-like
growth factor-binding proteins secreted by human endometrial stromal cells is
dependent on stromal differentiation. J Clin Endocrinol &Metab 1992; 75:12351241.
18. Markoff E, Zeitler P, Peleg S, Handwerger S. Characterization of the synthesis and
release of prolactin by an enriched fraction of human decidual cells. J Clin Endocrinol &Metab 1983; 56:962-968.
19. Wilfinger WW, Larsen WJ, Downs TR, Wilbur DL. An in vitro model for studies
of intercellular communication in cultured rat anterior pituitary cells. Tissue &
Cell 1984; 16:483-497.
20. Sinha YN, Selby FW, Lewis UJ, Vanderlaan WP. A homologous radioimmunoassay
for human prolactin. J Clin Endocrinol & Metab 1973; 36:509-516.
21. Laemmle UK. Cleavage of structural proteins during the assembly of the head of
bacteriophage T4. Nature 1970; 227:680-685.
22. DiMattia GE, Gellersen B, Duckworth ML, Freisen HG. Human prolactin gene
expression: the use of an alternative noncoding exon in decidua and the IM-9
P3 lymphoblast cell line. J Biol Chem 1990; 265:16412-16421.
23. Lee PDK, Conover CA, Powell DR Regulation and function of insulin-like growth
factor-binding protein-I. Proc Soc Exp Biol Med 1993; 204:4-29.
24. Aronica SM, Katzenellenbogen BS. Progesterone receptor regulation in uterine
cells: stimulation by estrogen, cyclic adenosine 3',5'-monophosphate, and insulin-like growth factor I and suppression by antiestrogens and protein kinase inhibitors. Endocrinology 1991; 128:2045-2052.
25. Behrman HR, Romero RJ. Prostaglandins and prostaglandin-like products in reproduction: eicosanoids, peroxides, and oxygen radicals. In: Yen SSC, Jaffe RB
(eds.), Reproductive Endocrinology. Philadelphia: W.B. Saunders Company; 1991:
238-272.
26. Yen SSC. Endocrine-metabolic adaptations in pregnancy. In: Yen SSC, Jaffe RB
(eds.), Reproductive Endocrinology. Philadelphia: W.B. Saunders Company; 1991:
936-981.
27. Clements JA, Reyes FI, Winter JSD, Faiman C. Studies on human sexual development IV. Fetal pituitary and serum, and amniotic fluid concentrations of prolactin. J Clin Endocrinol &Metab 1977; 44:408-413.