DEVELOPMENTAL BIOLOGY 151,297-e% (19%)

Interleukin-1, Interleukin-6, and Tumor Necrosis Factor a Are Produced

in the Mouse Uterus during the Estrous Cycle and Are Induced
by Estrogen and Progesterone
MAMATADE,THOMAS R. SANFORD,ANDGARY W. WOOD
Department of Pathology and Oncology, University of Kansas Medical Center, 39th and Rainbow Boulevard, Kansas City, Kansas 66103

Accepted February J+, 1992

The ovarian steroids, estrogen and progesterone, regulate cellular and molecular changes which occur in the uterus
during the estrous cycle. Cycles of protein synthesis, cell proliferation and differentiation, and cell death are the direct
results of changes in hormone concentration. To explore the possibility that cytokines, which stimulate proliferation
and differentiation of numerous types of cells, might be associated with those cyclic changes, the production of IL-l,
IL-6, and TNFa was examined in the mouse uterus. Cytokine mRNA expression, bioactivity, and immunoreactivity were
quantitated during the estrous cycle, following ovariectomy and exposure of ovariectomized mice to estrogen and
progesterone. IL-l, IL-6, and TNFa mRNA was detected, and mRNA levels for each of the eytokines varied with the
stage of the cycle. Cytokine bioactivity was expressed throughout the cycle, but levels of each cytokine were highest
during proestrus and/or estrus. Immunoreactivity paralleled bioactivity. Uterus from ovariectomized mice contained
little or no cytokine activity, and systemic administration of estrogen or progesterone resulted in the induction of IL-la
and IL-lb mRNA expression. Significant amounts of IL-6 and TNFol mRNA appeared only following the exposure of
ovariectomized mice to estrogen plus progesterone. Cytokine bioactivity and immunoreactivity also appeared following
the administration of estrogen and/or progesterone. The highest activity levels for each cytokine were observed follow-
ing the injection of estrogen plus progesterone. Cyclic expression of IL-l, IL-6, and TNFa in the uterus and their
apparent regulation by estrogen and progesterone raise the possibility that cytokines and factors which are induced by
cytokines are part of the regulatory process which is induced by ovarian hormones in the uterus of reproductive age
feIK%kS. 0 1992 Academic Press, Inc.

INTRODUCTION of production of additional locally acting factors, regu-

The estrous cycle in mammals involves a series of es-
trogen- and progesterone-regulated events during
which uterine cells undergo cycles of proliferation, dif-
ferentiation, and death (Psychoyos, 1973; Short, 1984).
The cellular and molecular changes which control these
changes are functionally associated with ovulation, suc-
cessful mating, and healthy pregnancy (Psychoyos,
1973; Short, 1984). Although reproductive events are de-
pendent on local actions of estrogen and progesterone,
some or all of the hormonal effects may result from the
action of local mediators, such as prostaglandins, leuko-
trienes, histamine, and growth factors (DiAugustine
ab, 1988; Kennedy, 1977; Malathy et ah, 1986; Murphy et
ah, 1988). Few locally acting factors have been identified
and definitively shown to be involved as mediators of
reproductive changes.
IL-l, IL-6, and TNFa are cytokines which stimulate
the proliferation and differentiation of a wide variety of
cell types (Balkwill and Burke, 1988; Dinarello, 1989;
Durum and Oppenheim, 1989) and under certain condi-
tions also can produce cell death. IL-l, IL-6, and TNFa,
either alone or through autocrineiparacrine induction
late the bodys reparative response to pathologic stimuli
(Balkwill and Burke, 1988; Dinarello, 1989; Durum and
Oppenheim, 1989). Although hormonally induced intra-
uterine events are physiological, numerous investiga-
tors have noted superficial similarities between these
changes and the cellular changes which are associated
with tissue repair. Recent observations suggested that
there may be a fundamental relationship between IL-l,
IL-6, and TNFcr production and reproductive events.
Ovulation is associated with an increase in body temper-
ature (Buxton and Atkinson, 1948). Increases in body
temperature are the result of actions of IL-l and/or IL-6
et on the hypothalamus (Helle et aZ., 1988; LeMay et al.,
1990). IL-l has been shown to be elevated during the
luteal phase of the estrous cycle (Cannon and Dinarello,
1985) and IL-l mRNA was shown by in situ hybridiza-
tion to be higher in the mouse uterus than in other nor-
mal organs (Takacs et al, 1988). Also, very high levels of
TNFa! were expressed in the ovary and may be function-
ally related to ovulation (Adashi, 1990; Adashi et al.,
1989; Roby and Terranova, 1988). Increased cytokine pro-
duction during portions of the estrous cycle when estro-
gen and progesterone levels are elevated may result

297 0012-1606192 $5.00

Copyright 0 1992 by Academic Press, Inc.
All rights of reproduction in any form reserved.
298 DEVELOPMENTAL BIOLOGY VOLUME 151,1992

from the ability of estrogen and progesterone to stimu- poly(A)+ RNA samples was at least 2%, but less than
late mononuclear phagocytes to produce IL-l (Flynn, 3% of the amount of RNA added. Hoffmann-LaRoche
1986; Hu, et al, 1988; Polan, et al., 1988). generously provided us with a cDNA clone for murine
To test the hypothesis that estrogen and progesterone IL-la (Lomedico et al., 1984). E. I. DuPont de Nemours
stimulate and regulate intrauterine IL-l, IL-6, and and Co. generously provided us with a cDNA clone for
TNFa production, those cytokines were quantitated in murine IL-lb (Huang et ah, 1988). DNAX Research Insti-
normal cycling mice. IL-l, IL-6, and TNFa were ex- tute of Molecular and Cellular Biology (Palo Alto, CA)
pressed in the uterus. Cytokine concentrations approxi- generously provided us with a cDNA clone for murine
mately paralleled changes in ovarian hormone concen- IL-6. Genentech, Inc. (San Francisco, CA) generously
trations. Each of the cytokines was produced in the provided us with a cDNA clone for murine TNFa. IL-la
uterus of ovariectomized mice in response to estrogen (500 bp), IL-16 (750 bp), IL-6 (650 bp), and TNFa (300 bp)
and progesterone. were subcloned into pSP6 or pGEM vectors (Promega
Biotech, Inc., Madison, WI). Anti-sense IL-la, IL-lb, IL-
MATERIALS AND METHODS 6, and TNFa riboprobes were synthesized using [32P]-
Animals GTP (>600 Ci/mM) and had specific activities in the
range of 2-4 X lo* dpm/pg (Melton et ah, 1984). RNA in
Swiss/Webster mice (7-8 weeks old) were purchased individual samples was measured by uv absorbance at
from Harlan/Sprague-Dawley (Indianapolis, IN), and 260 n&f. Cytokine-specific mRNA was localized by
maintained in the University of Kansas Medical Center. Northern blotting (Andrews et ah, 1987). The positive
Four- to 6-week-old BALB/c mice also were obtained control for each of the cytokine probes was total RNA
from Harlan/Sprague-Dawley for use in IL-l bioassays. obtained from Escherichia coli lipopolysaccharide
The stages of the estrous cycle were determined by vagi- (LPS)-stimulated 5774 cells. The American Type Cul-
nal smearing, which was subsequently confirmed by ture Collection (Bethesda, MD) provided us with the
gross morphological examination of uterine tissue. Ani- 5774 cell line. Five X 10 in 500 ml were stimulated with
mal use was approved by the University of Kansas LPS (10 rig/ml) for 3 hr and RNA was extracted using
Medical Center Institutional Animal Care and Use the phenol/chloroform method. The density of bands
Committee. was quantitated by scanning densitometry. Peak areas
were measured by laser densitometry using a Biomed
Ovariectomy and Hormone Replacement lD/2D Laser Densitometer (Biomed Instruments, Ful-
lerton, CA). Band area is reported in arbitrary units.
Mice were bilaterally ovariectomized through a single
dorsal incision site. Animals were rested for at least 6
days prior to receiving any treatment. Ovariectomized Cytokine Quantitation
mice were injected subcutaneously with 0.2,2, or 20 mg
Mice were killed by cervical dislocation. Uterus was
of progesterone (endotoxin-free, tissue culture grade
removed and frozen in liquid N,. Uteri from at least 10
D4-pregnene-3,20-dione; Sigma Chemical Co., St. Louis,
mice were pooled, homogenized in RPM1 1640 using a
MO) or 0.1, 10, 100, or 1000 ng of estrogen (endotoxin-
Dounce homogenizer, and sedimented at 27,OOOg for 20
free, tissue culture grade 17/3 estradiol; Sigma) dissolved
min. The supernatant was removed and sterilized by fil-
in sesame oil.
tration through 0.22 p&f pore-size Millipore filters.
Tissue supernatants were analyzed for cytokine activ-
Messenger RNA Detection and Quantitation ity using bioassays. IL-l was quantitated by a modifica-
Mice were killed by cervical dislocation. Uterus was tion of the thymocyte proliferation assay (Wood et ah,
removed and frozen in liquid N,. Frozen uteri from at 1988). Samples were assayed in quadruplicate. Specific-
least 10 mice were pooled and homogenized with a poly- ity was demonstrated by inhibiting the activity in each
tron. Total RNA was isolated by the guanidine isothio- of the supernatants with polyclonal anti-mouse IL-1
cyanate method (Han et al., 1987). Poly(A)+ RNA was (Genzyme, Inc., Cambridge, MA). Recombinant human
isolated from total RNA by adsorption to and elution IL-la (generously provided by Immunex Corp., Seattle,
from oligo(dT)-cellulose (type 3) (Aviv and Leder, 1972). WA) was used as the positive control. IL-6 was assayed
A single column was used to isolate all RNA samples. using the IL-6-dependent lymphoblastoid cell line, B9.
One milligram of total RNA was added to the column Assays were performed in triplicate. Dr. M. Kluger (De-
and the total amount of RNA eluted from the column partment of Physiology, University of Michigan, Ann
was quantitated spectrophotometrically. The yield was Arbor, MI) generously performed IL-6 assays (LeMay et
calculated by dividing the amount of eluted RNA into ah, 1990). Polyclonal anti-IL-6 inhibited the bioactivity,
the amount added to the column. The yield for all thereby demonstrating the specificity of the IL-6 bioac-
 DE, SANFORD, AND WOOD IL-l, X-6, and TNFa Production in Estrous Cycle
tivity. TNFa was assayed by the L929 cytotoxicity
method (Mosmann, 1983). Assays were performed in
triplicate. Polyclonal anti-mouse TNFa! (Genzyme
Corp.) inhibited activity to demonstrate specificity. Re-
combinant human TNFa (obtained from K. Ebner,
KUMC) was used to generate standard curves.
Immunocytochemistry
Mice were killed by cervical dislocation, and uteri
were surgically removed. Uteri from four mice were
pooled for each analysis. Cell suspensions of uterine tis-
sue were prepared by digestion with Type I collagenase
(1 mg/ml, Worthington Biochemical) and Type XV pro-
tease from Bacillus polymyxa (5 mg/ml, Sigma). The
dispersed cells were washed with RPM1 1640 containing
5% fetal bovine serum and directly cytocentrifuged onto
slides. The cells were vacuum dried, fixed in periodate-
lysine-paraformaldehyde buffer for 10 min, and imme-
diately stained with anti-cytokine antibodies by the indi-
rect immunoperoxidase method using Vectastain ABC
kits (Vector Laboratories). Polyclonal anti-IL-l (Gen-
zyme) and TNFcz (Genzyme) were diluted 1:250 in all
assays. The cells were counterstained with methyl
green. Six slides were stained at each data point. Slides
then were coded and scored from negative to four plus in
a double-blind fashion. Grading was performed by two
of the authors. To control for nonspecific staining, a sim-
ilar number of slides were taken through the staining
process with phosphate-buffered saline (PBS) that was
substituted for the primary antibody. To control for the
possibility that the specific rabbit anti-cytokine anti-
sera might nonspecifically stick to cytocentrifuged
uterine cells, normal rabbit serum (NRS) was substi-
tuted for the primary antibody.
Statistics
Statistical analysis of differences between cytokine
concentrations was clone using the Newman-Keuls mul-
tiple comparison test following the one-way analysis of
variance (ANOVA for repeated measures) test using the
IBM SAS computer program.
RESULTS
ILl, IL6, and TNFcY mRNA and Protein Product
Detection in Mouse Uterus during the Estrous Cycle
Local synthesis of cytokine gene products was demon-
strated by Northern blotting of uterine poly(A)+ RNA.
During diestrus, the period during which ovarian hor-
mone levels and uterine cellular activity are lowest
(Psychoyos, 1973; Short, 1984), mRNA for each of the
cytokines was detected (Fig. 1). As the animals moved
into the proestrus stage of the cycle, IL-la, IL-lp, and
299
TNFQ! mRNA expression decreased, while IL-6 expres-
sion increased. Between proestrus and estrus, IL-la, IL-
10, and TNFcv mRNA increased, while IL-6 decreased.
IL-la, IL-lp, and TNFa transcripts from uterine tissue
and 5774 cells exhibited the same electrophoretic mobil-
ity. However, IL-6 mRNA bands from uterus did not
migrate as far in the gel as did IL-6 mRNA from 5774
cells. This probably resulted from the fact that cells
producing IL-6 in the intact tissue were not exclusively
macrophages. Different-sized cytokine transcripts can
result from alternative splicing.
There was no clear temporal relationship between the
quantitative expression of cytokine mRNA and the
stage of the estrous cycle. Maximal gene expression was
not anticipated during cliestrus, the portion of the cycle
exhibiting least cellular activity, yet IL-10 and TNFa
expression was highest at that time. However, message
levels may not directly reflect intrauterine cytokine ac-
tivity. Since transcription and translation may be inde-
pendently related, peaks of bioactivity could differ from
peaks of message expression. Supernatants of homoge-
nized uterine tissue contained significant amounts of
bioactive IL-1 and IL-6 and small amounts of TNFa (Ta-
ble 1). IL-1 bioactivity was high throughout the cycle,
but the concentration was highest during proestrus and
estrus. The increase between diestrus and proestrus was
statistically significant (P < 0.05). IL-6 bioactivity was
undetectable during cliestrus, but high concentrations
were detected during proestrus. IL-6 was detectable
during estrus, but the concentration was substantially
decreased. The amount of TNFa also was low during
diestrus, but increased significantly (P < 0.05) from
cliestrus to proestrus and from proestrus to estrus (P <
0.05). Those data demonstrated that, despite peak mes-
sage expression during cliestrus, the highest concentra-
tions of cytokine bioactivity occurred during proestrus
and estrus. To determine whether immunologically de-
tectable cytokine was expressed in a similar manner in
the uterus, uterine cell suspensions were stained with
polyclonal antibodies to IL-1 and TNFa. The immunocy-
tochemistry and bioassay results were similar (Table 2).
Cells isolated from uteri from mice in cliestrus were posi-
tive for both IL-1 and TNFol. Cells isolated from uteri
from mice in proestrus and estrus also were positive for
IL-1 and TNFa, but the staining intensity was much
higher. The staining intensity and the total number of
cells stained were similar during proestrus and estrus.
Most of the cells in the population were positive. Con-
trols (cells stained with normal rabbit serum or PBS)
always were completely negative.
Induction of ILla, IL@, IL-S, and TNFu by Estrogen
and Progesterone in Uterus from Ovariectomixed Mice
The data derived from cycling female mice demon-
strated the estrous stage-specific expression of IL-l, IL-
300 DEVELOPMENTAL BIOLOGY VOLUME 151.1992

a b 6, and TNFor. The fact that cytokine concentrations were

Cl23 Cl23 highest during proestrus and estrus suggested that a
temporal relationship between hormone production and
intrauterine cytokine expression existed. The following
2.2kb ILl-a 1.8kb experiments were performed to determine whether es-
trogen or progesterone would induce intrauterine pro-
duction of IL-l, IL-6, or TNFol. Uterus from ovariecto-
mized mice contained trace amounts of IL-la and IL-l/3
200 mRNA and no detectable IL-6 or TNFa mRNA (Fig. 2).
80 Similarly, supernatants of uterine tissue from ovariec-
d 5
2 '$ 60
tomized mice contained no detectable IL-l, IL-6, or
100
TNFcr bioactivity (Table 3). When uterine cells from
2 2 40 ovariectomized mice were stained immunocytochemi-
ii a 20 tally with antibodies to IL-l or TNFo, the cells were
completely negative (Table 2).
The absence of detectable cytokine in uteri from ovar-
iectomized mice suggested that estrogen and/or proges-
terone could be required to maintain intrauterine cyto-
kine production. To determine whether estrogen or
28s ,**.a 28s *mm* progesterone induced cytokine gene expression, ovariec-
18s m** 18s m*O tomized mice were exposed to increasing concentrations
of estrogen or progesterone and tested by Northern
Cl23 Cl23 blotting 12 hr later. Progesterone induced both IL-la
and IL-10 mRNA (Fig. 3). Maximal expression occurred
in mice given 2 mg of progesterone. When the kinetics of
gene expression were assessed, IL-la already was in-
C d creased at 30 min and IL-10 induction was evident at 3
Cl23 Cl23
hr (Fig. 4). IL-la and IL-10 reached maximal concentra-
r tions 12 and 24 hr respectively after hormone injection
i (Fig. 4). Only trace amounts of IL-l message was de-
1.6kb IL-6 TNF-a tected in uteri from mice injected 24 hr earlier with se-
same oil alone. Detection of a high level of IL-l bioacti-
vity 24 hr after exposure of mice to progesterone demon-
strated that the IL-l mRNA was translated (Table 3).
100
Translation of message into protein product was docu-
mented further by the observation that uterine cells
E 80 d from progesterone-treated mice were positive when
a2 60 2 100 stained with anti-IL-l (Table 2).
k! 40 2 Estrogen also induced IL-la and IL-lb mRNA (Fig. 3).
20 k! Maximal message expression occurred in mice given 10
ng estrogen. When the kinetics of gene expression were
0 0 examined, IL-la and IL-l@ mRNA were evident at 6 and
12 3 12 3
12 hr, respectively, after hormone injection (Fig. 4).
Maximal expression of IL-la and IL-lp occurred at 12
28s -am* and 24 hr (Fig. 4). Detection of a high level of IL-l bioac-
28s r*mo* 18s m*O tivity 24 hr after exposure of mice to estrogen demon-
18s m-0 strated that the IL-l mRNA was translated (Table 3).
Translation of message was documented further by the
Cl23 Cl23

FIG. 1. Northern blots for (a) IL-la, (b) IL-l& (c) IL-6, and (d)
TNFa. Lanes 1-3 contain 2.2 fig of uterine poly(A)+ RNA from dies-
trus, proestrus, and estrus, respectively. Control (C) was 1.1 pg RNA
from LPS-stimulated macrophages. Upper panel is the Northern blot.
Middle panel is the spectrodensitometric analysis. Lower panel is an
acridine orange-stained gel containing 1.1 pg RNA from LPS-stimu-
lated macrophages and 6.6 Kg of the total uterine RNA from which the
poly(A)+ RNA was prepared, which is included to demonstrate load-
ing uniformity.
 DE, SANFORD, AND WOOD IL-l, IL-6, and TNFol Production in Estrous Cycle 301

TABLE 1
QUANTITATION OF CYTOKINE BIOACTIVITY IN THE MOUSE UTERUS
DURING THE ESTROUS CYCLE

Diestrus Proestrus Estrus

IL-l 16.7 t 2.3 63.4 k 1.9 54.7 f 3.9

IL-6 ND 92 12
TNFa 36 70 130

IL-1 (cpm x lo-), IL-6 (units/pg), and TNFa (units/mg) were

quantitated as described under Materials and Methods. Treatment
with 1.0, 0.5, 0.025, and 0.0125 ng rIL-la produced cpm of 72.1, 49.1,
29.7, and 16.0 x lo-, respectively. ND, below detectable limits.

observation that uterine cells from estrogen-treated

mice were positive when stained with anti-IL-l (Ta-
ble 2). IL-6 TNFcc
When ovariectomized mice were exposed to a combina- FIG. 2. Northern blots of 6.6 pg uterine RNA from ovariectomized
tion of estrogen and progesterone, the total amount of (OVEX) mice. Control (5774) was 1.1 pg RNA from LPS-stimulated
IL-la and IL-lfi mRNA was not appreciably higher than macrophages.
was observed following exposure to a single hormone
(Fig. 3). Despite the lack of obvious synergism between
estrogen and progesterone in the induction of IL-l posed to estrogen plus progesterone than that after ex-
mRNA, IL-1 bioactivity was significantly (P < 0.05) posure to either hormone alone (Fig. 3). IL-6 mRNA was
higher after exposure to both hormones than that after not detectable until 6 hr after injection of the two hor-
exposure to either hormone alone (Table 3). Uterine mones (Fig. 4). The amount of bioactive IL-6 detected in
cells from mice treated with both hormones were uteri from mice exposed to estrogen plus progesterone
strongly positive when stained with anti-IL-l (Table 2). was higher than that in uteri from mice treated with
Very small amounts of IL-6 mRNA were detected fol- either progesterone or estrogen alone (Table 3).
lowing exposure of ovariectomized mice to estrogen or No TNFa mRNA was detectable following exposure of
progesterone (Fig. 3) regardless of when RNA was ob- ovariectomized mice to estrogen or progesterone (Fig. 3)
tained (Fig. 4). No IL-6 mRNA was detected in ovariec- regardless of when RNA was obtained (Fig. 4). No TNFa
tomized mice injected with sesame oil alone. However, mRNA was detected in mice injected with sesame oil
IL-6 bioactivity was detected after ovariectomized mice alone. Also, no TNFL~ bioactivity was detectable after
received either estrogen or progesterone (Table 3). In exposure of ovariectomized mice to either estrogen or
contrast to the findings with IL-l, the IL-6 mRNA con- progesterone (Table 3). As was observed with IL-6, a
centration was substantially higher after mice were ex- significant amount of TNFa mRNA was induced when
mice received both estrogen and progesterone (Fig. 3).
TNFa mRNA was not detectable until 6 hr after injec-
TABLE 2 tion of hormones (Fig. 4). A significant (P < 0.001)
QUANTITATION OF CELL-ASSOCIATED CYTOKINE
BY IMM~NOCYTO~HEMI~TRY~
TABLE 3
PBS/NRS* IL-1 TNFa CYTOKINE BIOACTIVITY IN UTERUS FROM OVARIECTOMIZED MICE
STIMULATED WITH ESTROGEN OR PROGESTERONE
Diestrus - + +
Proestrus - +++ +++ Estrogen and
Estrus - +++ +++ Ovariectomy Estrogen Progesterone progesterone
Ovariectomy - - -
Estrogen - + + IL-1 7.4 * 0.8 58.0 f 1.0 52.7 it 2.4 92.7 -+ 3.1
Progesterone - + + IL-6 ND 56 48 139
Estrogen and TNFa ND 6 8 150
progesterone - +++ +++
IL-1 (cpm X 10e3), IL-6 (units/rg) and TNFa (units/mg) were
Staining intensity was scored from (-) to (++++). quantitated as described under Materials and Methods. Treatment
*PBS, phosphate-buffered saline; NRS, normal rabbit serum. Both with 1.0, 0.5, 0.025, and 0.0125 ng rIL-lol produced cpm of 72.1, 49.1,
were substituted for the primary antibodies in control stained slides. 29.7, and 16.0 x 10-3, respectively. ND, below detectable limits.
302 DEVELOPMENTAL BIOLOGY VOLUME 151.1992

a b
Cl 2345678 C 12345678

2.2kb

1.8kb

200

l3
5
100
2
w
z

Cl2345678
Cl2345678

c d
C 12345678

c 12345678
1.6kb

l.lkb TNF-a

I 2 3 4 5 6 7 8

Cl2345678
Cl2345678

FIG. 3. Northern blot for (a) IL-k, (b) IL-l& (c) IL-6, and (d) TNFa. Lanes 1-8 contain 6.6 wg uterine RNA from ovariectomized mice injected
12 hr earlier with: (1) 100 ng estrogen (E) and 2 mg progesterone (P); (2) 0.2 mg P; (3) 2 mg P; (4) 20 mg P; (5) 1 ng E; (6) 10 ng E; (7) 100 ng E; (8)
1000 ng E. Control (C) was 1.1 bg RNA from LPS-stimulated macrophages. Upper panel is the Northern blot. Middle panel is the spectrodensito-
metric analysis. Lower panel is an acridine orange-stained gel containing the same amount of RNA as was blotted in the upper panel.

amount of bioactive TNFa was detected in uteri from DISCUSSION

mice exposed to estrogen and progesterone (Table 3).
Uterine cells from mice treated with estrogen and pro- Two general conclusions were reached regarding cy-
gesterone were strongly positive for TNFcv (Table 2). tokine production in nonpregnant, reproductive-age
 DE, SANFORD, AND WOOD
rz Progesterone Estrogen Progestrone
$ .5 3 6 12 24 .5 3 6 12 24 .5
1.8Kb
l.lKb
FIG. 4. Northern blots for cytokines following injection of ovariecto-
mized mice with 2 mg progesterone, 100 ng estrogen, or 100 ng estro-
gen plus 2 mg progesterone. Each well received 6.6 pg RNA. Mice were
killed at increasing time in hours following injection of hormone. Con-
trol (5774) was 1.1 pg RNA from LPS-stimulated macrophages.
mice. (1) IL-l, IL-6, and TNFcv were produced in the
uterus during the estrous cycle, and the concentration of
each cytokine increased as the cycle stage advanced
from diestrus to proestrus and estrus. (2) Uterine IL-l,
IL-6, and TNFa production was dependent on the levels
of circulating estrogen and progesterone. The second
conclusion is based on the following observations: (1)
Ovariectomy eliminated intrauterine cytokine produc-
tion. (2) IL-l, IL-6, and TNFa were restored to concen-
trations seen in cycling mice by exposing ovariecto-
mized mice to estrogen or progesterone. (3) Restoration
was hormone concentration dependent, and estrogen
plus progesterone generated higher levels than either
hormone alone.
The data suggested that progesterone directly in-
duced uterine cells to produce IL-la and IL-lb. That
might account for the fact that IL-1 was detected during
diestrus, when little estrogen is produced. Further, the
observation that estrogen and progesterone together
produced higher tissue levels of each of the cytokines
than did either hormone alone may explain the high
tissue levels of IL-l, IL-6, and TNFa observed during
IL-l, K-6, and TNFa Production in Estrous Cycle 303
Estrogen t proestrus and estrus, when circulating estrogen and
progesterone concentrations are elevated.
3 6 12 24
Delayed induction of IL-la and IL-lp message by es-
trogen and delayed induction of IL-6 and TNFa message
by estrogen plus progesterone suggested that those hor-
monal effects were mediated through secondary messen-
gers. The most likely candidate for IL-6 and TNFa in-
duction is IL-l. IL-1 is an autocrine/paracrine inducer
of both IL-6 and TNFcY (Balkwill and Burke, 1988; Din-
arello, 1989; Durum and Oppenheim, 1989). At present,
IL-1 I3
we have no explanation for the delayed induction of IL-
la and IL-l@ by estrogen. The kinetics of IL-l, IL-6, and
TNFL~ appearance, both during the estrous cycle and fol-
lowing exposure of ovariectomized mice to hormones,
suggested that a strictly regulated sequence of induc-
tion exists, with IL-1 appearing first, followed by IL-6,
and then TNFa. Thus, IL-1 was high throughout the
cycle, but increased during proestrus. IL-6 levels in-
creased dramatically from diestrus to proestrus, and
TNFa was highest during estrus. When mice were ex-
posed to hormones, IL-1 mRNA was rapidly induced by
TNFa
progesterone, and high concentrations of bioactivity
were observed following the exposure of mice to estro-
gen or progesterone. IL-6 bioactivity appeared following
the exposure of ovariectomized mice to either estrogen
or progesterone, despite the appearance of little mRNA,
and concentrations were substantially higher following
the exposure to estrogen and progesterone. TNFol mes-
sage and bioactivity appeared only following the expo-
sure of mice to both progesterone and estrogen.
Previous investigations have demonstrated the IL-1
message in the uterus and TNFa message in the ovary
(Takacs et al., 1988; Adashi, 1990; Adashi et ab, 1989;
Roby and Terranova, 1988), but little or no cytokine pro-
duction in nonreproductive organs (Balkwill and Burke,
1988; Dinarello, 1989; Durum and Oppenheim, 1989). If
cytokine production results from stimulation of a single
population of cells, the uterus must contain a unique
population of hormone-responsive cytokine-producing
cells. That is not likely, because nonuterine macro-
phages produce IL-1 in response to ovarian hormones
(Flynn, 1986; Hu et al., 1988; Polan et al., 1988). Another
possible explanation is that intrauterine IL-1 produc-
tion results from a combination of effects. For example,
the effects may result from a direct hormonal stimula-
tion of one population of uterine cells to produce IL-l,
and a secondary stimulation of another population of
cells to produce CSF-1 (Arceci et al., 1989; Bartocci et al.,
1989; Pollard et al., 1987), which also stimulates macro-
phage cytokine production (Hume et al., 1988; Warren
and Ralph, 1986). Thus, ovarian hormones and CSF-1
could be synergistic with respect to IL-6 and TNFa pro-
duction. The resolution of these questions awaits future
studies.
304 DEVELOPMENTAL BIOLOGY
The cell source(s) for each cytokine will need to be
identified before the mechanisms involved in cytokine-
mediated intrauterine cell-cell interaction can be fully
understood. Macrophages are a major source for IL-l
and TNFa (Balkwill and Burke, 1988; Dinarello, 1989;
Durum and Oppenheim, 1989). IL-6 also is produced by
macrophages, but substantial amounts of IL-6 are pro-
duced by fibroblasts stimulated by IL-l (Balkwill and
Burke, 1988; Dinarello, 1989; Durum and Oppenheim,
1989). Although other uterine cell types, e.g., endothelial
cells, mast cells, and perhaps even epithelial cells, may
produce IL-l, it is possible to understand the results of
the current study in relation to known changes in uter-
ine macrophages. In the normal cycling mouse, approxi-
mately 10% of total uterine cells are macrophages (De
and Wood, 1990). Thus, adequate numbers of cells are
present to account for cytokine production in cycling
mice. Uterine macrophage numbers are controlled by
ovarian hormones (De and Wood, 1990). Macrophage
numbers decrease to very low levels following ovariec-
tomy, and exposure of ovariectomized mice to estrogen
or progesterone returns the macrophage numbers to
normal levels (De and Wood, 1990). Those numerical
changes appear to be the result of CSF-1 production
(unpublished observations). Previous studies have dem-
onstrated that macrophage function, assayed by intra-
uterine phagocytic activity, also is controlled by estro-
gen and progesterone (Vernon-Roberts, 1969; Nicol and
Vernon-Roberts, 1965). Our hypothesis is that cytokine
production is another parameter of uterine macrophage
function controlled by estrogen and progesterone.
Data were generated using three techniques for cyto-
kine quantitation, each of which is based on a funda-
mentally different principle. This was done to demon-
strate data reproducibility, to exclude potential difficul-
ties which might arise from attempting to interpret
data obtained using any one of the techniques by itself,
and to exclude the possibility that the data resulted
from artifacts in one of the technical approaches. Detec-
tion of mRNA demonstrated that cytokines were synthe-
sized locally. Detection of mRNA in the uterus from cy-
cling mice excluded artifactual explanations, e.g., endo-
toxin contamination, for appearance of message and
bioactivity following exposure of mice to hormones. The
lack of message in uterine RNA from ovariectomized
mice, the dose dependence of the mRNA response to
hormones, and the concordance of uterine bands with
bands in the positive control demonstrated that cyto-
kine mRNA was not being nonspecifically detected in
uterine RNA. Detecting cytokine activity both by bioas-
say and immunocytochemistry demonstrated that, not
only were genes transcribed, but the transcribed mes-
sage was translated into bioactive product. Finally, the
results obtained with the three techniques were quanti-
VOLUME 151,1992
tatively consistent, suggesting that induction of cyto-
kine production by estrogen and progesterone could be
biologically important.
One final observation which merits discussion is the
fact that cytokine mRNA expression and levels of bioac-
tive cytokine in vivo were not coincident. The antici-
pated result would be that, as estrogen or progesterone
levels increase, mRNA would appear first and would be
followed by production of biologically active cytokine.
With the exception of IL-6, the peak of bioactivity ap-
peared to precede the mRNA peak during the estrous
cycle. The most likely explanation for those data is that,
because of the cyclic nature of the changes, there may be
both positive and negative regulation, and regulation of
cytokine production occurs at both transcriptional and
post-transcriptional levels (Balkwill and Burke, 1988;
Dinarello, 1989; Durum and Openheim, 1989). Thus,
while RNA continued to be transcribed in response to
initial hormonal stimuli, the cytokines themselves may
have induced negative regulatory factors, such as prosta-
glandins or TGF/3, which inhibited cytokine production
but failed to block mRNA synthesis. RNA synthesis de-
creased later as hormone concentrations went back
down. Thus, what we see is a complex interplay between
positive and negative signals. Regardless of the explana-
tion, and others are possible, the data demonstrated
that it is essential to look at levels of translated, biologi-
cally active product in addition to mRNA levels when
attempting to relate cytokines to their function in viva.
The data suggesting that estrogen and progesterone
induce IL-l synthesis have numerous nonreproductive
implications. Estrogen and/or progesterone have
known modulatory effects on a variety of biological pro-
cesses, e.g., immune responses, osteoporosis, and tumor
progression. In each such situation, the possibility must
be considered that the effects are at least partially me-
diated through the induction of cytokines.
We thank G. K. Andrews for extensive help and advice with our
molecular biology. We thank Dr. M. Kluger, University of Michigan,
Ann Arbor, Michigan, for performing the IL-6 bioassays. Supported in
part by NIH Grant HD17678.
REFERENCES
Adashi, E. Y. (1990). The potential relevance of cytokines to ovarian
physiology: The emerging role of resident ovarian cells of the white
blood cell series. Endocrinol. Rev. 1,454-464.
Adashi, E. Y., Resnick, C. E., Packman, J. N., Hurwitz, A., and Payne,
D. W. (1989). Tumor necrosis factor-a inhibits gonadotropin hor-
monal action in non-transformed ovarian granulosa cells. J. Biol.
Chem. 264,l.
Andrews, G. K., Lehman, L. D., Huet, Y. M., and Dey, S. K. (1987).
Metalothionein gene regulation in the preimplantation rabbit blas-
tocyst. Development 100,463-469.
Arceci, R. J., Shanahan, F., Stanley, E. R., and Pollard, J. W. (1989).
Temporal expression and location of colony stimulating factor 1
 DE, SANFORD, AND WOOD X-1,
(CSF-1) and its receptor in the female reproductive tract are consis-
tent with CSF-l-regulated placental development. Proc. Natl. Acad.
Sci. USA 86,8818-8822.
Aviv, H., and Leder, P. (1972). Purification of biologically active globin
messenger RNA by chromatography on oligothymidylic acid cellu-
lose. Proc. Natl. Acad. Sci. USA 69,1409-1412.
Balkwill, F. R., and Burke, F. (1989). The cytokine network. Immunol.
Today 10,299-304.
Bartocci, A., Pollard, J. W., and Stanley, E. R. (1986). Regulation of
colony-stimulating factor 1 during pregnancy. J. Eq. Med. 164,
956-961.
Buxton, C. L., and Atkinson, W. B. (1948). Hormonal factors involved
in the regulation of basal body temperature during the menstrual
cycle and pregnancy. J. Clin. Endocrinol Metab. 8,544-549.
Cannon, J. G., and Dinarello, C. A. (1985). Increased plasma interleu-
kin-l activity in women after ovulation. Science 227,1247-1249.
De, M., and Wood, G. W. (1991). Influence of oestrogen and progester-
one on macrophage distribution in the mouse uterus. J. EndocrinoL
126,417-424.
DiAugustine, R. P., Petrusz, P., Bell, G. I., Brown, C. F., Korach, K. S.,
McLachlan, J. A., and Teng, C. T. (1988). Influence of estrogens on
mouse uterine epidermal growth factor precursor protein and mes-
senger ribonucleic acid. Endocrinology 122,2355-2363.
Dinarello, C. A. (1989). Interleukin 1 and its biologically related cyto-
kines. In Advances in Immunology (F. Dixon, Ed.), Vol. 44, pp.
153-205. Academic Press, San Diego, CA.
Durum, S., and Oppenheim, J. J. (1989). Macrophage-derived media-
tors: Interleukin 1, tumor necrosis factor, interleukin 6, interferon
and related cytokines. In Fundamental Immunology (W. E. Paul,
Ed.), pp. 639-661. Raven Press Ltd., New York.
Flynn, A. (1986). Expression of Ia and the production of interleukin 1
by peritoneal exudate macrophages activated in vivo by steroids.
Lzlfe Sci. 38,2455-2460.
Han, J. H., Stratowa, C., and Rutter, W. J. (1987). Isolation of full-
length putative rat lysophospholipase cDNA using improved meth-
ods for mRNA isolation and cDNA cloning. Biochemistry 26,1617-
1625.
Helle, J. P., Brakenhoff, J. P. J., de Groot, E. R., and Aarden, L. A.
(1988). Interleukin 6 is involved in interleukin-l-induced activities.
Eur. J Immunol. l&553-559.
Hu, S. K., Mitcho, Y. L., and Rath, N. C. (1988). Effect of estradiol on
interleukin 1 synthesis by macrophages. Int. J. Immunopharmacol.
10,247-252.
Huang, J. J., Newton, R. C., Rutledge, S. J., Horuk, R., Matthew, J. B.,
Covington, M., and Lin, Y. (1988). Characterization of murine IL-lfl:
Isolation, expression, and purification. .I ImmunoL 140, 3838-3843.
Huet-Hudson, Y. M., Chakreborty, C., De, S. K., Suzuki, Y., Andrews,
G. K., and Dey, S. K. (1990). Estrogen regulates the synthesis of
epidermal growth factor in mouse uterine epithelial cells. Mol. En-
docrinol 4, 510-523.
Hume, D. A., Pavli, P., Donahue, R. E., and Fidler, I. J. (1988). The
effect of human recombinant macrophage colony stimulating factor
(CSF-1) on the murine mononuclear phagocyte system in viva J.
Immunol. 141,3405-3409.
Kennedy, T. G. (1977). Production of leukotrienes and prostaglandins
X-6, and TNFa Production in Estrous Cycle 305
in the rat uterus during the periimplantation period. Biol. Reprod.
16,286-290.
LeMay, L. G., Vander, A. J., and Kluger, M. J. (1990). Role of interleu-
kin 6 in fever in rats. Am. J. Physiol 258, R798-R803.
Lomedico, P. T., Gubler, U., Hellman, C. P., Dukovich, M., Giri, J. G.,
Pan, Y. C. E., Collier, K., Semionow, R., Chua, A. O., and Mizel, S. B.
(1984). Cloning and expression of murine interleukin-1 cDNA in
Escherichia coli. Nature (London) 312,458-462.
Malathy, P. V., Cheng, H. C., and Dey, S. K. (1986). Production of
leukotrienes and prostaglandins in the rat uterus during the peri-
implantation period. Prostaglandins 32,605-612.
Melton, D. A., Krieg, P. A., Rebagliati, M. R., Maniatis, T., Zinn, K.,
and Green, M. R. (1984). Efficient in vitro synthesis of biologically
active RNA and RNA hybridization probes from plasmids contain-
ing a bacteriophage SP6 promoter. Nucleic Acids Res. 12,7035-7056.
Mosmann, T. (1983). Rapid calorimetric assay for cellular growth and
survival: Application to proliferation and cytotoxicity assays. J. Im-
munol. Methods 65,55-63.
Murphy, L. J., Murphy, L. C., and Friesen, H. G. (1988). Estrogen in-
duces insulin-like growth factor expression in the rat uterus. Mol.
Endocrinol. 1,445-450.
Nicol, T., and Vernon-Roberts, B. (1965). The influence of the estrus
cycle, pregnancy and ovariectomy on RES activity. J. Reticuloen-
dothel. Sot. 2, 15-29.
Polan, M. L., Daniele, A., and Kuo, A. (1988). Gonadal steroids modu-
late human monocyte interleukin-1 (IL-l) activity. Fert. Steril. 49,
964-968.
Pollard, J. W., Bartocci, A., Arceci, R., Orlofsky, A., Ladner, M. B., and
Stanley, E. R. (1987). Apparent role of macrophage growth factor
(CSF-1) in placental development. Nature (London) 330,484-486.
Psychoyos, A. (1973). Endocrine control of egg implantation. In
Handbook of Physiology-Endocrinology II, Section 7 (R. 0.
Greep, E. B. Astwood, and S. R. Geiger, Eds.). Vol. II, Part 2. pp.
187-215. American Physiological Society, Washington, DC.
Roby, K. F., and Terranova, P. F. (1988). Tumor necrosis factor alpha
alters follicular steroidogenesis in vitro. Endocrinology 123, 2952-
2959.
Short, R. V. (1984). Oestrous and menstrual cycles. In Reproduction
in Mammals: Hormonal Control of Reproduction, Book 3 (C. R.
Austin and R. V. Short, Ed.), pp. 115-152. Cambridge Univ. Press,
New York.
Tachi, C., and Tachi, S. (1986). Macrophages and implantation. Ann.
N. Y Acad Sci. 476,158-168.
Takacs, L., Kovacs, E. J., Smith, M. R., Young, H. A., and Durum, S. K.
(1988). Detection of IL-1 alpha and IL-1 beta gene expression by in
situ hybridization: Tissue localization of IL-1 mRNA in the normal
C57BL/6 mouse. J. ImmunoZ. 141,3081-3095.
Vernon-Roberts, B. (1969). The effects of steroid hormones on macro-
phage activity. Int. Rev. Cytol. 25, 131-159.
Warren, M. K., and Ralph, P. (1986). Macrophage growth factor CSF-1
stimulates human monocyte production of interferon, tumor necro-
sis factor and colony stimulating activity. J. Immunol. 137, 2281-
2285.
Wood, G. W., Kamel, S., and Smith, K. (1988). Immunoregulation and
prostaglandin production by mechanically-derived and enzyme-
derived murine decidual cells. J Reprod Immunol. 13, 235-248.