BIOLOGY
OF REPRODUCtiON
Platelet-Activating
ABRAHAM
FRANK
(1989)
41, 578-586
Factor
in Human
Luteal
Phase
Endometrium1
A. ALECOZAY,4.6
BERTIL G. CASSLEN,3.4
ROBERT
D. DELEON,6
MICHAEL
J. K. HARPER,2’4
MONICA
TERI A. NOUCHI,5
and DONALD
J. HANAHAN5
Departments
of Obstetrics
University
and Gynecology,4
of Texas
Health
San Antonio,
Department
Texas Tech
Texas
M. piEiiL,1
SIL VA,4
and Biochemistry5
Science
Center
78284
and
of Obstetrics
and Gynecology6
University
Health
Science
Center
Lubbock,
Texas
79409
ABSTRACT
Platelet-activating
factor (PAF; 1-0-alkyl-2-acesyl-sn-glycero-3-phosphorykholine)
is one of the most potent
mediators
of vascular
permeability.
PAF levels change in the rabbit endometriwn
just prior to implantation,
which suggests
that PAF may be a key substance
transducing
preimplantation
embryonic
signals.
To study
whether PAF was present in the hwnan endometriwn,
and if so, to determine
the cellular origin and hormonal
regulation
of endometrial
PAF, specimens
were obtained from 14 women (aged 23-42 yr) undergoing
elective
hysterectomy
during the luteal phase of the cycle (plasma progesterone
levels >2 ng/ml). No specimens
were
taken from women with malignant
uterine pathology.
Stromal cells and epithelial
glandular
cells were separated
by coliagenase
and DNAse digestion,
and then cultured
to confluence
in vitro in medium 199. Radioiminunoassays of prostaglandin
F (PGF) and prolactin
in the culture media were used to confirm cell type and viability.
PGF release into the culture mediwn from stromal cells was low (control
1.52 ± 020 ng/ml), and unchanged
by
hormone
treatment.
In contrast,
release
of PGF from unstimulated
glandular
cells was 6.05 ± 052 nglml, and was
significantly
increased
(p<O.OS)by estradiol or progesterone
plus estradiol,
to 12.17±
1.67, and 8.60 ± 0.81,
respectively.
Progesterone
alone was without effect. Prolactin
was secreted
by stromal cell cultures,
increasing
steadily from 24 to 120 h. The levels in the medium were increased
by progesterone.
PAF activity was assessed by
rabbit platelet
aggregation
and serotonin-release
bioassays
after lipid extraction
and separation
by thin-layer
chromatography.
Basal PAF concentration
in unstimulated
stromal
cells during the 24-h period after confluence
was 0.14
± 0.05
PAF in stromal
(mean ± SEM:pmol/mg
cells in a dose-dependent
protein/24
manner,
h). Progesterone
combined
with estradiol
a 1-riM dose giving the optimal
response.
(1 IJM) was as effective
as when combined
with estradiol
(p<O.OS)
in stromal
cells, the values
being 0.99 ± 0.15
stimulatory
effect. PAF concentration
in the medium
was
treatment.
PAF was not found
and its cellular concentration
cal
Platelet-activating
mediator,
has
factor (PAF),
been identified
University
of Lund,
S-22l
to the stromal
cells of the endometrium,
alkyl-2-acetyl-sn-glycero-3-phosphotylcholine
a potent lipid chemichemically
as 1-0-
85 Lund.
(Demop-
oulos et al., 1979). PAF is produced
in response
to
various
stimuli
in a variety of cells including
neutrophils, basophils,
monocytes,
and mast cells (Camussi
et
a!., 1977; Lynch et a!., 1979; Pinckard
et at., 1979;
Clark et at., 1980; Lotner et at., 1980) and is one of the
most potent
inflammatory
mediators
formed
and released by mammalian
tissues.
PAF causes platelets
to
Accepted
Ma
15 1989
Received
March
10, 1989.
tThis work was supported
in part by a technical services agreement
from
the Special Programme
of Research,
Development
and Research
Training in
Human Reproduction.
World Health Organization
(WHO 87007), and NIH
grants (HD 14048 and 10202, RIA core).
Reprint requests:
Michael 3K. Harper. Department
of Obstctncslt3ynecology, The University
of Texas Health Science Center at San Antonio, l’x
7824-7836.
address:
Dr. B. 0. Cassl#{233}n,Department
of Obstetrics
Gycccology.
(10 tiM) to increase
PAF concentrations
significantly
and 1.0 ± 0.13, respectively.
Estradiol
alone had no
undetectable
in most cultures,
and did not change
with
in the glands. PAF appears to be restricted
is increased
by progesterone.
INTRODUCTION
(10 tiM) increased
Progesterone
alone
.
aggregate
0.1-0.001
and to secrete
serotomn
within
I nun at
pM (Pinckard
et at., 1979). PAF has been
.
..
shown to imtiate a number of biological
reactions,
such
as contraction
of smooth muscle (Findlay
et at., 1981;
Tokumura
Sweden.
578
et at.,
1984),
negative
ionotropic
cardiac
PAF IN HUMAN
effects
(Benveniste
et a!., 1983; Levi et a!., 1984),
vasoconstriction
(BjOrk and SmedegArd,
1983; Buxton
et at., 1986), exocrine
gland stimulation
(Soling et at.,
1984),
glycogenolysis
in the fed (perfused)
liver
(Shukia et a!., 1983), and initiation
of labor (Billah and
Johnston,
1983; Billah et at., 1985).
Previous
studies have indicated
production
of PAF
from the mouse embryo
(O’Neill,
1985a,b),
and assigned a role for this substance
during implantation
in
rabbits (Angle et at., 1988). The endometrial
concentration of PAF appears to be increased
significantly
in the
uteri of pregnant rabbits prior to implantation
and then
returns to the control
(estrous)
level at implantation
(Angle
et a!., 1985, 1988).
In pregnancy,
PAF-like
factors have been implicated
in the activation
of platelets and induction
of a mi!d thrombocytopenia
seen
during the first week following
conception
in humans
(O’Neill
et at., 1985). Furthermore,
since a correlation
has been found between in vitro production
of PAF by
human embryos
and subsequent
pregnancy
outcome
after their replacement
(O’Neill
et at., 1985, 1987), it
has been suggested
that the activation
of platelets involves
the release
of biologically
active substances
which are a necessary prerequisite
for the establishment
of pregnancy.
This
suggestion
is further supported
by
studies in estrous mice and in rabbits in which PAF
induced secretion of early pregnancy
factor (Orozco et
at., 1986; Sueoka et at., 1988).
Although
the presence
of PAF has been detected
in
rabbit and rat endometrium
(Angle et a!., 1985, 1988;
Yasuda
et at., 1986, 1988),
there has been no such
previous
study on human endometrium.
However,
PAF
release
by human early-stage
zygotes
in culture
has
been observed
(O’Nei!l
et at., 1985). Implantation
requires an embryo-endometria!
biochemical
interaction;
thus, it is important
to examine
PAF synthesis
and
secretion
in the human
endometrium
to determine
whether PAF may be a mediator in this species.
The
objective
of this study was to assess the cellular origin
and hormonal
regulation
of PAF production
in human
luteal phase endometrial
tissue cultured
in vitro. A
preliminary
account of this work has been made (Alecozay et at., 1988). Radioimmunoassays
of prostaglandin
F (PGF) and pro!actin
were used to confirm cell type
and viability.
Previous
work has shown that glandular
cel!s release more PGF into the medium than do stroma! cells (Schatz et at., 1987), that prolactin is secreted
by stromal cells but not by glandular cells, and that this
secretion
is amplified
by progestational
agents (Maslar
and Ansbacher,
1986; Huang et at., 1987; Cass!#{233}n
et at.,
1990).
ENDOMETRIUM
579
MATERIALS
AND
METHODS
Chemicals
Chemicals
were obtained
from (a) Sigma Chemical
Co. (St. Louis, MO): 17(-estradiol,
progesterone,
hydrocortisone,
Triton X-100,
and insulin-transferrin-sodium selenite media supplement;
(b) Gibco Laboratories (Grand Island,
NY): collagenase,
medium
199
(Ml 99) with L-glutamine,
fetal calf serum, penicillin/
streptomycin
and fungizone;
(C) Collaborative
Research
(Bedford,
MA): epidermal
growth factor,
(d) ParkeDavis
(Grand Prairie,
TX): thrombin;
(e) ICN Immunobiologicats
(Irvine, CA): bovine
serum albumin
(BSA); (f) NEN Research
Products
(Wilmington,
DE):
5-[i ,2-3H(N)J-hydroxytryptamine
creatinine
sulfate
([3H]-serotonin,
sp. act. 15-30 Ci/mmol);
(g) National
Diagnostics
(Somerville,
NJ): Liquiscint
scintillation
cocktail;
and (h) Boehringer-Mannheim
Biochemicals
(Indianapolis,
IN): mouse monoclonat
antibodies,
anticytokeratin
component
18 and anti-vimentin,
and antimouse
IgG-FITC-labeled
polyclonal
antibody
from
sheep.
Tissue
Procurement
Endometriat
specimens
were
obtained
from
14
women (aged 23-42
yr) undergoing
elective
hysterectomy for nonendometrial
pathology,
i.e., cervical dysplasia or uterine prolapse. The day of the ovarian cycle
at the time of hysterectomy
was determined
on the
basis of histologic
criteria (Noyes
et at., 1950). The
luteal phase was further verified by measurement
of the
concentration
of progesterone
in peripheral
plasma
on
the day of surgery (progesterone
concentration
>2.0 ng/
ml).
Tissue
Processing
and
Culture
A sample of uterine tissue not required
for pathological evaluation
was immediately
brought to the laboratory on ice. Isolated cell preparations
were made according
to the method
of Schatz
et at. (1987),
as
modified
by Cassl#{233}net at. (1990).
Briefly, endometrial
tissue was dissected
from the uterine specimen
under
aseptic conditions
in a laminar flow tissue culture hood.
After the removal of blood clots, the tissue was cut into
pieces (- 1 mm3), then transferred
into a dissociation
solution
containing
collagenase
(1 mg/ml),
and incubated at 37#{176}C
in an atmosphere
of 5% CO2 and 95% air
for 2 h. The enzymatic
dissociation
was helped
me-
580
ALECOZAY
chanicatly
by frequent
gentle pipetting.
The enzymetreated
tissue was strained
through
a 350 un nylon
mesh to remove any undigested
tissue, and then passed
through
a second
nylon mesh with a pore size of 35
p.m. The endometrial
epitheiat
cells containing
mostly
glandular
cells were retained,
whereas
single stromal
cells passed through
the 35-p.m nylon mesh. The stromat cells were washed
two times with M199, resuspended
in medium,
and plated
directly.
Specimens
heavily contaminated
with erythrocytes
were not used.
Undigested
stromal cell clumps together with the glandular cells, both of which had been recovered
on the
35-I.Lm mesh were resuspended
in M199 medium
and
removed
by aspiration
of the supernatant
after glands
had been allowed
to sediment
for 5-10 mm. If necessary, this procedure
was repeated
to obtain
a pure
glandular
cell preparation.
Stromal cells and glands were plated in (2 ml) basal
medium,
which consisted
of M199
with 10% heatinactivated
fetal calf serum, 2 mM glutamine,
100,000
IU penicillin/i,
100 mg streptomycinhl,
0.25 mg fungizone/I,
10 mg insulin/I,
10 mg transferrin/l,
10 p,g
selenite/1,
1 p.M hydrocortisone,
10 pg epidermal
growth factor/i,
and 10 nM estradiol.
All cultures
were
done in Corning
6-well
plastic
tissue
culture
plates
(Corning
Glass Works,
Corning,
NY). Stromal
cells
were plated at 106 cells/35-mm
well in basal medium.
The glands were plated at a density of 400/cm2 (± 20)
in 35-mm wells. Cultures
were incubated
in moist air
with 5% CO2 at 37C. The experiments
were started 24
h after initial plating
of the cells, by which time the
cells
had reached
confluence.
Preliminary
studies
showed
that cells in such cultures
excluded
Trypan
Blue dye, which was used as a crude test for viability.
Confluent
cultures
of stromal
and glandular
cells
were exposed
to a constant
concentration
of esiradiol
(10 nM) for 24 h, and then to the same concentration
of
estradiol
and/or progesterone
(1 pM) for a further 24 h
before collection
for PAF assay. Control cultures
without hormonal
addition
were also perfonned.
Progesterone dose curves
were determined
by the addition
of
progesterone
in concentrations
ranging from 10 pM to
10 nM during the second 24-h culture period after the
initial exposure
to estradiol.
For each 6-well plate, one
cell sample was used for protein estimation
(using the
Bio-Rad
procedure:
Bio-Rad
Laboratories,
Richmond,
CA) and one served as control without any hormonal
addition.
When
experiments
involved
PAF measuremerits were terminated
72 h after initial
plating, protein
concentrations
for glandular
and stromat
cell cultures
were not significantly
different.
El
AL
Similarly,
for the POF measurements
by radioimmunoassay,
1-mi samples
of medium
were taken from
other cultures
at the same time as cells were harvested
for PAF assays. The results are expressed
as ng/ml/24
h. For the prolactin
measurements,
the cultures
were
continued
longer.
One milliliter
of medium
was removed for radioimmunoassay
every 24 h and replaced
with 1 ml of conditioned
medium.
The amount
of
prolacun
measured
at each time period thus included a
contribution
of prolactin
remaining
in the culture well
from the previous
period now diluted by a factor of 2.
All prolactin
values have been adjusted for the amount
of prolactin
remaining
at each time period on the assumption
that none had been metabolized.
The values
for prolactin
release into the medium
are expressed
as
ng/m1124 h and reflect the release
rate only over the
previous
24 h.
Assessment
of
Purity
of Stromal
Cell
Cultures
Purity of the stromal cell cultures
was assessed
immunocytochemically.
Stromal
cell preparations
were
aliquoted
in culture medium
into 8-well chamber slides
(Labtek#{174},Nunc Laboratories,
Inc., Naperville,
IL) and
cultured
for 48 h under conditions
of temperature
and
gas phases described
previously.
Subsequently,
the cells
were fixed for 10 mm in methanol,
prechilled
to -20#{176}C,
and washed twice in phosphate-buffered
saline (PBS).
Fixed cells were permeabiized
by a 5-mm incubation
in PBS containing
0.1% Triton X-100, followed
by two
washes in PBS. The fixed and penneabilized
cells were
then exposed
to 100% fetal calf serum for 20 mm at
room temperature,
and the calf serum was removed by
aspiration.
The cells were incubated
for 45 mm in the
presence
of undiluted
preparations
of the mouse monoclonal antibodies
anti-cytokeraun
18 or anti-vimentin.
Control wells were incubated
with PBS alone. Subsequently,
all wells were washed
twice with PBS and
incubated
for 45 mm with a 1:30 dilution
of FITClabeled anti-mouse
IgG at 37#{176}C
in a warm, humidified
chamber.
The cells were washed
twice in PBS. The
plastic well-forming
unit of the slide was manually
removed,
the slid was affixed
with cover slips, and
aquamount
was applied.
Indirect
immunofluorescence
was observed with the aid of an Olympus
BH2 Microscope with epifluorescence
capabilities.
Photographs
(35 mm, Ektachrome
400 ASA) were taken of the
phase-contrast
image
and fluorescent
image of each
microscopic
field of each well incubated
without primary antibody
as a control and either monoclonal
antibody to the intermediate
ifiaments
or cytokeratin.
PAF IN HUMAN
581
ENDOMETRIUM
Radioimmunoassay
The PGF in tissue culture media was measured
by
the radioimmunoassay
procedure
described
by Harper
et at. (1981),
as modified
by Jones et at. (1986). The
antibody
was obtained
from Advanced
Magnetics,
Inc.,
(Cambridge,
MA). The inter- and intraassay
coefficients of variation
were, respectively,
7.0% and 8.7%.
Prolactin
in the tissue culture
media was measured
by a radioimmunoassay
previously
described
by Ehara
et at. (1973),
with modifications.
Highly purified
human
pituitary
prolactin
provided
by
NIAMDD
(NIADDK-hPRL-I-b,
batch No. AFP 2284C2)
was radio-iodinated
and 50 pg/tube was used. A rabbit antiserum to prolactin
was provided
by NIAMDD
(antihPRL-3,
batch No. AFP-C
11580) and was used at a
final concentration
of 1:300,000.
The standard
was the
hPRL-I-b,
batch No. AFP-2284C2.
For the assay, the
samples
or standards
were incubated
for 24 h at 4#{176}C,
followed
by addition of the label, incubation
for 24 h at
4#{176}C,
addition
of the second
antibody,
and a further
incubation
for 24 h at 4#{176}C.
Separation
of bound from
free prolactin
was accomplished
by centrifugation.
The
sensitivity
of the assay was 0.1 rig/tube. The intraassay
and interassay
coefficients
of variation were 5.4% and
18.4%, respectively,
at 25% of maximum
binding.
All
prolactin
measurements
were done in one assay. Progesterone
in the plasma samples
were measured
in a
routine
radioimniunoassay
previously
described
by
Pauerstein
et at. (1978).
in 150 p.1 of chloroform/methanol
by TLC.
(1:1)
for separation
Silica gel G plates (500
pm, 10 x 20 cm) (Anattech,
Newark, DE) were washed
in a neutral solvent
system
of chloroform/methanol/
water (65:35:6),
allowed to air dry, and then activated
by heating
at about
150#{176}C
for 20 mm. Plates were
allowed
to cool and then were divided into 4 lanes. A
standard
consisting
of lyso-phosphatidyicholine
(lysoPC), sphingomyelin,
and PC was applied to the first
lane, an extracted
sample (from above procedure)
to the
second,
another
extracted
sample to the third, and the
standard
to the fourth lane. Plates were then developed
in the same neutral solvent system. After development,
only the standard
lanes were sprayed
with either
2-p-toluidinylinaphthylene-6-sulfonate
(TNS) (Jones et
at., 1982) or phosphorus-detecting
spray to visualize
and mark standards.
Silica gel from the sample lanes
was scraped
from an area corresponding
to just above
lyso-PC to just below PC, with the standards as references. This PAP-containing
fraction also includes
any
sphingomyelin
present.
Experiments
have shown that
the amounts
of sphingomyelin
present
do not affect
PAF-induced
aggregation
of, or serotonin
secretion
from, washed
rabbit platelets.
In control
samples
in
which there was no detectable
PAP activity, the level of
sphingomyelin
was the same as in the stimulated
cells.
Further addition
of sphingomyelin,
at a 10 p.M concentration, to rabbit platelets
caused no shape change,
or
aggregation
and secretion.
Also, it did not inhibit the
behavior of any PAP added to the same assay system.
The collected
silica gel was mixed with 5 ml chloroIsolation
of PAF from
Endometrium
Samples
form/methanol/water
(1:2:0.8 v/v) and allowed
to stand
Extraction
of samples
for thin-layer
chromatography
for 30 mm. Samples
were then centrifuged
at 800 x g
(TLC).
Samples were extracted according
to the method
for 10 mm to pellet the silica gel, and the supernatant
of Bligh and Dyer (1959).
Samples
were mixed with
was collected.
Chloroform
and water were then added
chloroform
and methanol
so that the proportions
of
so that proportions
were chloroform/methanol/water
(2:
chloroform/methanol/water
were (1:2:0.8),
where the
2:1.8) to effect phase separation.
After vortexing,
samoriginal sample constitutes
the water portion.
Samples,
ples were centrifuged
at 800 x g for 10 mm. The lower
if not processed
immediately,
were stored at -20#{176}C. (chloroform-rich)
phase was collected
and evaporated
The usual volume of the extract was about 7 ml. After
under N2 for testing by biologicat
assay.
thawing,
the samples were centrifuged
at 800 x g for 10
mm to pellet any cell debris, and the supernatant
was
Biological
Assay
(Aggregation
and Secretion)
decanted.
The supematant
was then partitioned
into
aqueous
and organic phases by the addition of chloroAggregation.
Samples were assayed for PAP-induced
form and water, so that the proportions
were now
aggregation
of washed rabbit platelets.
Rabbit platelets
are preferred
since they are at least 100-fold
more
chloroform/methanol/water
(2:2:1.8).
Samples
were
sensitive
than human platelets.
Further, they are more
vortexed
well and then centrifuged
at 800 x g for 10
consistent
in response
and more
stable
over the
mm to facilitate
partition
of the phases.
The lower
8-h assay period.
The same is not true of human
(chloroform
rich) layer, containing
the extracted
lipids,
platelets.
Inasmuch
as the structure
of PAP from human
was collected
and evaporated
under N2 and redissolved
TLC
and
final
extraction.
582
ALECOZAY
tissues has been shown to be the same as from animal
tissues,
PAP behaves
the same on rabbit and human
platelets,
and can be validly assayed with rabbit platelets. In a cuvette containing
a magnetic
stir bar, 100 p.1
washed rabbit platelets (1.25 x 106 platelets/p.l)
and 400
p.1 Ca2-containing
Tyrode’s
buffer,
pH 7.2, were
mixed. Cuvettes
were warmed
to 37#{176}C
and placed in a
Payton Aggregometer
(Payton Associates,
Inc., Buffalo,
NY). The unit was calibrated
by setting the platelet
suspension
as 0% transmission
and deionized
water as
100% transmission.
A standard curve was run using
standard
1-O-hexadecyl-2-acetyl-sn-glycero-3-phosphorylcholine
(AGEPC;
PAP C 16:0, Bachem,
Philadelphia,
PA), where aggregation
induced by 0.5 riM PAP (fmat
concentration)
represented
100% aggregation.
Samples
for assay were suspended
in 50 p.1 of 2.5 mg/mI bovine
serum albumin in 0.9% NaC1 (BSA/saline),
and 10 p1
was added to the platelet
suspension.
Samples
were
selected
for secretion
assay on the basis of significant
detectable
inducement
of aggregation.
The lower limit
of detection
of PAP-induced
aggregation
is about 15
pM AGEPC
(fmal concentration).
Secretion
assay. Samples
were also tested for PAP
induction of [3H]serotomn
release from rabbit platelets.
Samples
were serially
diluted
(if needed)
and 4-gl
samples were placed in reaction tubes. The assay also
included
a blank (BSA/saline
alone), Triton X-100 (4
p.1, 2.5%, to represent 100% [3Hlserotonin
release), and
thrombin
(2 p1, 10 units/mI)
(to check platelet
response).
A PAP curve was assayed by serially diluting
standard
PAP and assaying
4-p.l samples.
Platelets
([3H]serotonin-labeled)
were diluted 1:5 in Ca2-contaming Tyrode’s
buffer, pH 7.2, and 200 p.1 was added
to each sample-containing
reaction tube. The reaction
was stopped
at I mm by addition
of 20 p.1 ice-cold
formal in (formaldehyde/saline,
1:9) and the tubes were
centrifuged
at 800 x g for 10 mm at 4#{176}C.
Samples were
then counted by removing
25-p.l aliquots of supematant,
in duplicate,
from each sample
tube to scintillation
counting
vials and adding 5 ml Liquiscint
counting
fluid. Samples
were counted
with a Beckman
LS6800
scintillation
counter for presence
of radioactive
label.
In early experiments,
the identity of the PAP activity
was further established
by treatment
of the samples
with 0.5 N KOH in methanol
for 15 mm at room
temperature,
which gives rise to the inactive lyso-PAF.
This abolished any effects on aggregation
and secretion.
Subsequent
acetylation
of this sample with acetic anhydride with catalytic
amounts of perchloric
acid, which
reconverts
the lyso-PAF
to PAP, completely
restored
biological
activity.
ET AL.
The amount of PAP (in pmol) was calculated
by
plotting
percentage
secretion
of [3H]serotonin
versus
PAP concentration
(known by PAP curve assay). The
PAP standard
curve was plotted
first and then the
samples.
If the samples’
percentage
of secretion
was
above the standard curve, the pmol amount PAP could
not be calculated,
and the samples had to be diluted and
reassayed.
Since the original volume of the tested samples was 50 p.1 and that of the AGEPC was 200 p.1, the
maximum
amount of PAP standard in a tube was 5.0
pmol and for the samples,
1.25 pmol.
Statistical
Analysis
Analysis
of variance (SAS Institute Inc., Cary, NC)
was used to compare
the groups, and then differences
between
means were evaluated
by Student-NewmanKeuls (SNK) multiple range test. Probability
values of
p<0.05
were considered
significant.
RESULTS
The purity of the stromal cell cultures was assessed
by analysis of photographic
images of the same microscopic fields for fluorescence
of each intermediate
filament in the labeled
and control
slides. Stromal cell
populations
were strongly fluorescent
for anti-vimentin
and did not fluoresce
with anti-cytokeratin
18. Approximately
10% of the total cell population
in a given
microscopic
field was positive
for anti-cytokeratin
18
and negative
for anti-vimentin.
These results indicate
that the stromal cell population
was approximately
90%
pure and had 10% contamination
with epithelial
or
glandular-like
cells. Contamination
with other cell types
appeared to be minimal.
Since, in previous
studies in
our laboratory,
glandular cell cultures had not secreted
prolactin,
it was felt that stromal cell contamination
of
these cultures was likely to be minimal.
The viability of the glandular
and stromal cell cultures was assessed
by measurement
of their ability to
release prolactin
and PGF into the medium
and by
exclusion
of Trypan Blue dye. In unstimulated
glandular cell cultures, basal levels of prolactin in the medium
were low (<0.2 ng/ml at 48 h of culture).
In similar
cultures receiving
estradiol
(10 nM) and progesterone
(1 pM) over the same period, the basal level of prolactin release
approximately
doubled
(data not shown).
These low values suggest only a minimal contamination
of stromal cells in the glandular
cell preparations.
Unstimulated
stromal cell cultures released about 0.5
ng/mI/24
h prolactin
over the first 48 h, and signifi-
PAP IN HUMAN
8
#{149}
Control
(Bosoi
o Treated
(Estrodiot and P
a
terone)
ENDOMETRIUM
TABLE 1. PGF concentrations
in medium of lutes! phase glandular and stroinal
cell cultures and the effects of addition of progesterone
(1 pM) and cstradiol (10
n.
Hormonal
E
C.....
#{163}4
C
0
at
I.-
e
0
24
48
72
Time
96
120
(h)
FIG.!. ProLactin production
by lutcal phase endometrial
stroinal cells after
different periods of culture and stimulation
by estradiol (10 nM) and progesterone (1 pM). One milliliter of mednun was removed at each 24-h period and
replaced with 1 ml fresh, conditioned
medium. All values have bean adjusted
for the contribution
of residual prolactin after each sampling, and are expressed
as ngflnhl24 h. reflecting the release over the previous 24-h period. Means have
been calculated
from the values derived from 6 wells (onelatient)
for basal and
5 wells (one4,atient)
for hormone-treated
groups. Except where shown, standard error bars were smaller than the ymbo1 denoting the mean value. Significant differences
were determined
by ANOVA followed by Student-NewmanKeul’s test. Letters indicate significant
differences
atp<0.05 (a is different from
b,etc.)
candy greater amounts
over succeeding
days, reaching
2.07 ± 0.23 (mean ± SEM) mg/mI/24 h at 120 h (Fig. 1).
In contrast,
addition
of estradiol
and progesterone
to
these stromal
cell cultures
doubled
prolactin
release
over the fIrst 48 h (about 1 ng/ml/24
h), and significandy elevated release during the next 72 h over these
and basal values to reach 6.69 ± 0.02 mg/mI/24
h at 120
h (Fig. 1).
Both glands and stroma released POF into the medium, but significantly
higher values were observed
for
the glands (Table 1). Estradiol
atone significantly
increased POF release by glandular cells compared
to all
other groups.
Estradiol
combined
with progesterone
also stimulated
POF release from glandular
cells compared to control, but to a lesser extent than estradiol
alone. Progesterone
alone was without effect. Neither
progesterone
nor estradiol
alone or in combination
stimulated
PGF release
from stromal cells. All PGF
values for the siromal
cell cultures
were significantly
lower than for glandular cell cultures. In these experiments, POE release
was not measured.
From these same cultures,
the cells were extracted
for PAP measurements.
No PAP was detected
in the
glandular cells irrespective
of hormonal treatment.
PAP
was never detected
in glandular
cell culture
medium,
and only occasionally
and in very small amounts
in
stromal
cell culture
medium.
PAP was, however,
de-
583
treatment
Patients
Sirotnal cells
None
Estradiol
Progesterone
Pzogestcaone&cs*radiol
Glandular
cells
None
Eatradiol
Progesterone
Progesterone
& esnadiol
(n)
PGF concentration
in medium
Mean ± SEM (nghnl/24
h)
14
14
14
14
1.52
2.41
1.34
1.67
14
14
14
14
6.05
12.17
5.28
8.60
±
±
±
±
0.20’
0.19’
018d
0.l3
± O.52c
± 1.67’
± O.6O’
±
‘Leuer
superscripts
indicate significant
differences
between
p’cO.OS (a is different from b, etc).
S(fl
from each individual
patient were plated on one 6.wcll
each well receiving
a different
means
plate
at
with
treatment.
tected
in the stromal cells. Thus, PAP appears
to remain
cell-associated,
at least under the present experimental
conditions.
In unstimulated
stromat cell cultures,
basal PAP levels were low. The control values did not differ between
experiments
(see Tables 2 and 3). In the presence
of
estradiol
alone, PAP concentrations
were unchanged,
but when increasing
doses of progesterone
were added
to the cultures in addition to estradiol, a dose-dependent
increase in PAP was observed
(Table 2). The maximal
response
was achieved
with a dose of 1 pM, and for
subsequent
experiments,
this concentration
of progesterone was used. Even low doses of progesterone
combined with estradiol
produced an increase of PAP concentration
over basal or esiradiol-treated
alone levels.
Table 3 shows that this response
to progesterone
is
not dependent
on the presence
of estradiol,
since similady elevated PAP concentrations
were observed
when
the stromal cell cultures were treated with progesterone
atone or with progesterone
plus estradiol.
DISCUSSION
Prolactin
has been shown to be secreted
by human
endometriat
explants
in vitro (Maslar
and Riddick,
1979; Daly et at., 1983), and its secretion
is stimulated
from such explants by progesterone
(Maslar and Ansbacher, 1986). Recently,
it has been determined
that
prolactin
is secreted
by isolated stromal cells in culture,
and that such production
is increased
dose-dependently
by medroxyprogesterone
acetate (Huang et at., 1987) or
progesterone
Cassl#{233}n
et at., 1990). In the present experiments, the stromal cell cultures
responded
similarly,
ALECOZAY
584
TABLE 2. PAF concentrations
concentration
of progesterone.
Hormonal
in luteal phase stromal
treatment
None
Estradiol (10
Estradiol (10
progesterone
Progesterone
Progesterone
Progesterone
dLetter
Palients*
5
superscripts
indicate
each well receiving
5
to
5
5
significant
differences
TABLE 3. PAF concentrations
hormonal stimulmio
Hormone
h)
treatment
between
means
at
on one 6-well
plate
(n)
stromal
cell cultures
PAF concentration
Mean ± SEM
(pinol/mg protein/24
14
14
14
0.14
14
1.00 ± 0.13’
0.99
a different
in relation
to
h)
± 005b
±
± 015’
‘‘Leuer
superscripts
indicate significant
differences
between
pc0.05 (a is different from b, etc.)
*Cfls from each individual
patient were plated on one 6-well
each well receiving
were plated
in lutes! phase
Patients
None
Estradiol (10 nM)
Progesterone
(1 pM)
Esiradiol (10 aM) plus
progesterone
(1 pM)
0.39 ± 0.04’
0.50 ± O.O4
1.00 ± 0.12’
0.80 ±
5
a different
PAF concentration
Mean ± SEM
(prnol/mg proteiW24
in relation
0.05 ± 002d
0.Q7 ± 002d
5
aM)
nM) plus
(10 nM)
(100 nM)
(1 pM)
(10 &M)
p<0.05 (a is different from b, etc.)
*Cells from each individual
patient
with
(a)
cell cultures
ET AL.
means
plate
at
with
treatment.
with
treatment.
prolactin
levels increasing
with time in culture.
This
provides
evidence
that these cultures
contained
mainly
stromal
cells, which were viable. In addition,
stromal cell cultures
were estimated
to be 90% pure on
the basis of positive
staining
for anti-vimentin
and
negative staining
for anti-cytokeratin
18. These conclusions are based on the results of previous studies (Centola et at., 1984; Mungyer et at., 1987; von Koskull and
Virtanen,
1987) on the nature of the intermediate
filament proteins
in human uterine
cells in vitro, which
have shown
that cytokeratin
positivity
is associated
with luminal epithelial
cells and that stromal cells stain
strongly
for vimentin.
Indeed,
it is possible
that the
stromal cell population
is 100% pure, since, the 10%
“contamination”
by presumed
epithelial
cells may represent
the induction
of cytokeratin
expression
in
decidualized
stromal cells in vitro as shown by others
(von Koskull
and Virtanen,
1987).
Schatz
et at. (1987)
have demonstrated
that both
stromal and glandular
epithelial
cells from both proliferative and secretory
phase human endometrium
release
POF
into the culture
medium
to approximately
the
same extent. However,
estradiol
(10 nM) was able to
stimulate
epithelial
cells to release POF
to a much
greater extent (about 10-fold) than stromal cells. In the
present
experiments,
estradiol
significantly
increased
POF release
in the glandular
cell cultures
and was
without effect in the stromal cell cultures. According
to
the results of Schatz et at. (1987), this indicates
that the
glandular
cell cultures
were predominantly
epitheiat
cells with little stromal
cell contamination.
PAP has been shown to be present in the uterus of
rats (Yasuda
et at., 1986, 1988) and rabbits (Angle et
at., 1985, 1988). It is also present in human
amnion
(Bilah
and Johnston,
1983; Billah et at., 1985). Our
previous
studies
in the rabbit
(Angle
et at., 1988)
revealed
that PAP was found mainly in the endometriurn, with only small amounts
in the myometrium,
and
that the uterine concentrations
increased
concomitantly
with the known rise of peripheral
plasma progesterone
levels
(Spilman
and Wilks,
1976;
Harrington
and
Rothermel,
1977).
The present studies with isolated human endometriat
cells in culture indicate,
first, that PAP is present in the
human uterus also. Thus, PAP is likely to be present in
the uteri of most mammals.
Second,
PAP was found
only in stromal
cell cultures,
and not in glandular
epithelial
cell cultures.
Thus, it is probable
that the
stromal cells are the site of PAP synthesis.
Third, PAP
was not found in the medium. This may imply absence
of secretion,
despite the presence
of serum containing
albumin,
which is known
to be involved
in the cellular
secretion
of PAP from polymorphonuclear
leukocytes
(Ludwig
et at., 1985). Thus, PAP appears to remain
cell-associated
as it does in stimulated
endotheliat
cells
(Whatley
et at., 1988). However,
secretion
followed
by
binding
to membrane
receptors
or degradation
in the
medium could also have occurred.
Last, PAP levels in
stromat cells were elevated in a dose-dependent
manner
by addition of progesterone,
irrespective
of the presence
of estradiol.
This is in agreement
with the results obtained for PAP concentration
in rabbit uterus up to Day
6 of pregnancy,
at a time
when peripheral
estradiol
levels are low (Challis
et at., 1973).
PAP is thought to exert its biological
effect through
specific membrane
binding sites (for review, see Godfroid and Braquet,
1987), and the presence
of such
binding sites have been demonstrated
in rabbit endometrial
membrane
preparations
(Kudolo
and Harper,
1989). The exact uterine cell type carrying such binding
PAP IN HUMAN
sites is not yet known.
However,
Smith
and Kelly
(1988) have demonstrated
that PAP stimulated
release
of POE2, but not PGF2a, from human glandular
epithehat cells in culture.
PAP did not change PG release
from stromal
cell cultures.
These findings
imply that
glandular
epithelial
cells contain specific
binding sites
for PAP.
The presence
of PAP in the human
uterus
and its
hormonal
amplification
by progesterone
suggests
that
this pro-inflammatory
mediator
may play a significant
role in normal physiological
functions.
REFERENCES
Alecozay AA. DcLeon FD, Harper MJK, Nouchi T, Hanahan DJ, 1988. Platelet-activating
factor (PAF) in human endometrium.
Biol Rcprod (Suppl.
1)38:78 (Absir. 90)
Angle Ml, Jones MA, McManus
LM, Pinckard RN, Harper MJK, 1988. Platelet-activating
factor in the rabbit uterus during early prcgny.
J Rcprod
Fenil 83:711-fl
Angle MJ. Jones MA. Pinckard RN, McManus
LM. Harper MJK, 1985. Platelet-activating
factor (PAF) in the rabbit uterus during early pregnancy.
Biol Reprod (Suppt. 1)32:143
(Abstr. 212)
Benveniste
I, Boullet C, Brink C, Labat C. 1983. The actions of Paf-acether
(platelet-activating
factor) on guinea-pig
isolated heait preparations.
Br J
Pharmacol
80:81-83
Blllah MM, Di Rcnzo GC, Ban C, Truong CT, Hoffman DR, Anceschi
MM.
BleasdaleJE,
Johnston JM, 1985. Platelet-activating
factor metabolism
in
human amman and the responses
of this tissue to extracdllular
plateletactivating factor. Prostaglandins
30:841-50
Billah MM. Johnston JM, 1983. Identification
of phospholipid
platelet-activating factor (l-O-alkyl-2-acetyl-SN-glycero-3-phosphocholine)
in human
amniotic fluid and urine. Biochein Biophys Res Commun
113:51-58
Bjork J. Smedeg#{224}rdG, 1983. Acute microvascular
effect of PAF-acether,
as
studied by intravital microscopy.
Eur J Pharmacol
96:87-94
Bligh EG, Dyer WJ, 1959. A rapid method of total lipid extraction and purification. Can) Biochem Physiol 37:911-17
Buxton DB, Fisher RA, Hi,hn
Di, Olson MS. 1986. Platelet-activating
factor-mediated
vasoconstriction
and glycogenolysis
in the perfused rat liver. J Biol Cheni 261:644-49
Camussi G, Mencia-HuertaiM,
Benvenisie
i, 1977. Release of platelet-activating factor and histamine.
I. Effect of immune complexes,
complement
and
neutrophils
on human and rabbit mastocytes
and basophils.
Immunology
33:523-34
Cassl#{233}n
BG, Siler-Khodr
TM. Harper MJK, 1990. Progesterone
regulation
of
prolactin release from human endometrial
stromal cells in culture: potential bioassay for progestational
activity. Acts Endcrinol Copenh (In press)
CentolaGM,
CisarM, Knab DR. 1984. Establishment
and morphologic
characterization
of normal human endometrium
in vitro. In Vitro 20.451-62
Challis JRG, Davies U, Ryan KJ, 1973. The concentrations
of progesterone,
estrone and eatradiol-17
in the plasma of pregnant rabbits. Endocrinology 93:971-76
Clark P0, Hanahan DJ, PinckardRN,
1980. Physical and chemical properties of
platelet-activating
factor obtained
from human neutrophils
and monocytes and rabbit neutrophils
and basophils.
Biochim Biophys Acts 628:
69-75
Daly DC, Maslar IA, Riddick DH, 1983. Prolactin production
during in vitro
decidualizaiicm
of proliferative
endometrium.
Am) Obstet Gynecol
145:
672-78
Demopoulos
CA, Pinckard RN, Hanahan DJ, 1979. Platelet-activating
factor.
Evidence for l-O-alkyl-2-acetyl-snglyceryl-3-phosphcsylcholine
as the
active component
(a new class of lipid chemical mediators).)
Biol Qiem
254:9355-58
Ehara Y, Siler T, Van<knRerg
0, Sinha YN, Yen SCC, 1973. Ciroulating
prolactin levels during the menstrual cycle: episodic release and diurnal variation. Am J Obstet Gynecol
117962 -70
ENDOMETRIUM
585
Findlay SR, Lichtenstcin
LM, Hanahan DJ, Pinckard RN, 1981. Contraction
of
guinea pig heal smooth muscle by acetyl glyceryl ether pbosphorylcholine. Am) Physiol 241:C130-33
Godfroid ii. Braquet P, 1986. PAF-acether
specific binding sites: 1. Quantitative SAR study of PAF-acether
isosteres. Trends Pharmacol
Sd
7:368-96
Harper MJK, Norris CJ. Friedncha
WE, Moreno A, 1981. Poly l:C accelerates
ovum transport
in the rabbit by a prostaglandin-mediated
mechanism.
I
Reprod Fertil 63:81-89
Hatrington
FE, Rothermel
JD. 1977. Daily changes in peripheral
plasma progesterone concentrations
in pregnant and pscudopregnantrabbitL
Life Sd
20:1333-40
Huang JR, Tseng L, Bischof P, Janne OA, 1987. Regulation
of prolactin production by progestin, estrogen, and relaxin in human endometrial
stroinal
cells. Endocrinology
121:201 1-ti
Jones MA, Cao Z-d, Anderson
W, Norris C, Harper MJK, 1986. Capillary
permeability
changes inthe uteri of recipient rabbits after transfer of blastocysts from indomethacin-trealed
donors. J Reprod Fertil 78:261-73
Jones MA, Keenan
RW. Horowitz
P, 1982. Use of 6-p-toluidino-2-naphthalenesulfonic
acid to quantitate lipids after thin layer chromatography.)
Chro.natogr 273:522-24
Kudolo GB, Harper MIX. 1989. Characterization
of platelet-activating
factor
binding sites on uterine membranes
from pregnant rabbits
Biol Reprod
41:587-603
Levi R, Burke JA, Gun Z-O, Hatted Y, Happens CM, McManus
LM. Hanahan
DJ, Pinckard
RN, 1984. Acetyl
glycezyl
ether phosphorylcholine
(AGEPC).
A putative mediator of cardiac anaphylaxis
in the guinea pig.
CirRes 54:117-24
Lomer GZ. Lynch JM, Ben Si, Henson PM, 1980. Human neutrophil..derived
platelet-activating
factor.)
lmmunol
124:676-84
Ludwig IC, Happens CL, McManus LM, Molt GE, Pinckard RN, 1985. Modulation of platelet-activating
factor (PAF) synthesis
and release from homan polymorphonuclear
leukocytes
(PMN): role of extracellular
albumin. Arch Biochean Biophys 241:337-47
Lynch JM. Lower GZ, Bets Si, Henson PM. 1979. The release of a plateletactivating
factor by stimulated
rabbit neutrophils.
J Immunol
123:
1219-26
Maslar IA, Ansbachcr
R, 1986. Effects of progesterone
on decidual prolactin
production
by organ cultures of human endometrium.
Endocrinology
118:
2102-08
Maslar IA, Riddick DH, 1979. Prolacun
production
by human endometnum
during the normal menstrual
cycle. Am I Obstet Gynecol
135:751-54
Mungyer 0, Willeniacn
WNP, Rolland R. Vemer HM, Ramackers
FCS, isp
PHK. Pods LO, 1987. Cells of the mucous membrane
of the female genital tract in culture: a comparative
study with regard to the histogenesis
of
endometriosis.
In Vitro Cell Dcv Bio123:111-17
Noyes RW, Hertig AT, Rock J, 1950. Dating the endoinetrial
biopsy. Femi
Steril 1:3-25
O’Neill C, 1985s. Exmnin2iion
of the causes of early pregnancy-associated
thrombocytopenia
in mice. J Reprod Fertil 73 :567-77
O’Neill C, 1985b. Pstuial characterization
of the embryo-derived
platelet-activating factor in mice. J Reprod Peril 75:375-80
O’Neill C, Gidley-Baird
AA, Pike IL. Saunders
DM, 1987. Use of a bioassay
for embryo-derived
platelet-activating
factor as a means of assessing
qualily and pregnancy
potential
of human embryos.
Peril Steril 47:
969-75
O’Neill C. Pike II... Porter RN. Gidley-Baird
AA, Sinosich Mi. Saunders FM,
1985. Maternal recognition
of pregnancy
prior to implantation:
methods
for monitoring
embryonic
viability in vitro and in vivo. Ann NY Acad Sd
442:429-39
Orozco C. Perkins T, Clarke FM, 1986. Platelet-activating
factor induces the
expression
of early pregnancy
factor activity in female mice. I Reprod
Fertil 78349-55
Pauerstein Cl, Eddy CA, Croxano HD, Hess R, Siler-Khodr
TM. Croxatto HB,
1978. Temporal
relationships
of estrogen,
progesterone,
and lutcinizing
hormone levels to ovulation
in women and infrahuman
primates.
Am I
Obstet Gynecol 130:876-84
Pinckard
RN, Fair RS, Hanahan
DJ, 1979. Physicochemical
and functional
identity of rabbit platelet-activating
factor (PAF) released in vivo during
EgE anaphylaxis
with PAF released In vitro from IgE sensitized basophils.
I Immunol
123:1847-57
Schatz F, Markiewicz
L, Gurpide E, 1987. Differential
effects of estradiol.
arachidooic
acid, and A23 187 cci prostaglandin
F
output by epithelial
586
ALECOZAY
and siroinal cells of human endometrium.
Endocrinology
120:1465-71
SD, Buxtoc DB, Olson MS, Hanahan
Di, 1983. Acctylglyceryl
ether
phosphosylcholine.
A potent activator of hepatic phosphoinositide
metabolism and glycogenolysis.
I Biol Qiem 258:1(2l2-14
Smith SK, Kelly RW, 1988. Effect of platelet-activating
factor on the release of
PGF-2a and PGE-2 by separated
cells of human endoinetrium.
I Reprod
Fertil82:271-76
Soling H-D, Elbi H, Feat W, 1984. Acelylcholine-like
effects of 1.O-alkyl2-acetyl-sn-glycero-3-phosphocholine
(“platelet-activating
factor’5 and
its analogues
in exocrine secretory glands. Eur I Biochem 144:65-fl
Spilman
CH, Wilks 1W, 1976. Peripheral
plasma progesterone
during egg
transport in the rabbit. Proc Soc Exp Biol Med 151:726-29
Sueo&a K, Dbarmaijan
AM, Miyazaki T, AtiasSJ, Wallach EE, 1988. Plateletactivating factor-induced
early pregnancy
factor (EPF) activity from the
Shukla
ET AL.
perfused rabbit ovary and oviduct. Am) Obstet Gynecol
159:1580-84
Tokumura
A, Fukuzawa
K, Tsukaiani
H, 1984. Contractile
effect of 1-0hexadecyl-2-O-acetyl-sn-glycero-3phosphocholine
on strips of isolated
rat intestine. 3 Pharm Pliarmacol
36:210-12
von Koahull H, Vinanen I, 1987. Induction of cytokeratin
expression
in human
mesenchymal
cells.) Cell Physiol 133:321-29
Whatley RE, Zimmecman
GA. Mcintyre TM, Prescott SM, 1988. Endothelium
from diverse vascular sources synthesizes
platelet-activating
factor. Arteriosclerosis
8:321-3 1
Yasuda K, Sazouchi K. Nakayama
R, Saito K. 1988. Acyl type platelet-activating factor in normal rat uterus determined
by gas chromatography
mass
spectromeny.
Bloused Environ Mass Spectrom
16:137-41
Yasuda K, Satouchi K, Saito K, 1986. Platelet-activating
factor in normal rat
uterus. Biochcsn Biophys Res Commun
138:1231-36