BIOLOGY OF REPRODUCTION 69, 1273–1280 (2003)
Published online before print 11 June 2003.
Mimicking the Events of Menstruation in the Murine Uterus1
M. Brasted,3C.A. White,3T.G. Kennedy,4and L.A. Salamonsen2,3
Uterine Biology Laboratory,3Prince Henry’s Institute of Medical Research, Clayton, Victoria 3168, Australia
Department of Physiology & Pharmacology,4The University of Western Ontario, London, Ontario, Canada N6A 5C1
Menstruation and endometrial regeneration occur during ev-
ery normal reproductive cycle in women and some Old World
primates. Many of the cellular and molecular events of men-
struation have been identified by correlative or in vitro studies,
but the lack of a convenient model for menstruation in a labo-
ratory animal has restricted functional studies. In this study, a
mouse model for menstruation first described by Finn in the
1980s has been modified for use in a commonly used inbred
strain of mouse. A decidual stimulus was applied into the uterine
lumen of appropriately primed mice and leukocyte numbers and
apoptosis were examined over time following progesterone
withdrawal. Endometrial tissue breakdown was initiated after
12–16 h, and by 24 h, the entire decidual zone had been shed.
Re-epithelialization was nearly complete by 36 h and the en-
dometrium was fully restored by 48 h. Leukocyte numbers in-
creased significantly in the basal zone by 12 h after progester-
one withdrawal, preceding stromal destruction. Stromal apopto-
sis was detected by TUNEL staining at 0 and 12 h but decreased
by 16 h after progesterone withdrawal. This mouse model thus
mimics many of the events of human menstruation and has the
potential to assist in elucidation of the functional roles of a va-
riety of factors thought to be important in both menstruation
and endometrial repair.
apoptosis, female reproductive tract, menstrual cycle, uterus
With every reproductive cycle, the human endometrium
undergoes extensive remodeling that is unparalleled in any
other adult organ. Following menstruation, when most of
the functional layer of the endometrium is shed, tissue res-
toration occurs, initially by very rapid re-epithelialization
of the exposed surface. Increased estrogen levels then stim-
ulate proliferation and reestablishment of the stromal and
vascular components of the tissue. Following ovulation, as
progesterone levels rise, there is considerable cellular dif-
ferentiation, which prepares the endometrium for blastocyst
implantation; this includes decidualization of the stroma,
elongation and increased tortuosity of the glands, and an-
giogenesis producing specialized spiral arterioles. In the ab-
sence of an implanting blastocyst, the corpus luteum, the
1This work was supported by the NH&MRC of Australia (grants 169003,
143798). T.G.K. contributed to this work while on sabbatical leave from
the University of Western Ontario. He was supported in part by a Hudson
2Correspondence: L.A. Salamonsen, Prince Henry’s Institute of Medical
Research, Level 4, Block E, Monash Medical Centre, 246 Clayton Road,
Clayton, Victoria 3168, Australia. FAX: 61 3 9594 6125;
Received: 19 February 2003.
First decision: 20 March 2003.
Accepted: 16 May 2003.
? 2003 by the Society for the Study of Reproduction, Inc.
ISSN: 0006-3363. http://www.biolreprod.org
primary source of circulating progesterone, degenerates and
serum progesterone levels fall, providing a critical trigger
for menstruation .
Many of the molecular and cellular events of menstru-
ation have now been identified [2, 3]. Leukocyte numbers
increase dramatically immediately premenstrually and can
contribute as much as 40% of the total cellular composition
of the tissue at this time . Other events include actions
of matrix metalloproteinases (MMPs) to degrade the tissue
matrix, of vasoactive substances such as prostaglandins, en-
dothelin, and nitric oxide, and of products of a variety of
leukocytes. However, as menstruation occurs naturally only
in women, some Old World primates, the elephant shrew
(Elephantus myuras jamesoni) and the bat (Glossophaga
soricina) , most of the data supporting roles for such
factors, is either correlative or based on in vitro experiments
using tissue explants or cell cultures. Old World primates
are not widely available for experimentation. There is thus
a clear need for a nonprimate animal model of menstrua-
tion, in which the sequence of events leading to menstru-
ation can be established and which can also be used for
functional studies to examine the relevance of individual
molecules to the processes of menstruation and endometrial
During the 1980s, Finn and Pope  developed a model
of endometrial breakdown in the mouse, in which proges-
terone support was withdrawn from artificially decidualized
endometrium, to mimic the fall in serum progesterone that
occurs in women following luteal regression. Features that
occur in the endometrium of women at the time of men-
struation, including influx of leukocytes and tissue degen-
eration, were observed in this model. However, consider-
able temporal variation in events was observed and the
model was not further used.
The purpose of the present study was to optimize this
mouse model for menstruation in an inbred strain of mouse
commonly used for genetic manipulation, to establish the
temporal sequence of events of tissue breakdown and res-
toration, and to assess the numbers of leukocytes and extent
of apoptosis during the phase of tissue breakdown. Proges-
terone was delivered via silastic implants, providing an op-
portunity to induce a rapid decline in serum progesterone
levels and thus reduce variability in responses.
MATERIALS AND METHODS
Female C57BL/6 mice of 8–12 wk of age were obtained from Monash
University Animal Services. Mice were housed in standard conditions with
food and water provided ad libitum and a constant light cycle of 12 h
(lights-on from 0800 h to 2000 h). Ethics approval for this project was
granted by the Monash University/Monash Medical Centre Animal Ethics
BRASTED ET AL.
struation. Ovariectomized (OVX) mice are subjected to a series of injec-
tions of 17?-estradiol (E) and an implant of progesterone (P) is inserted,
prior to a decidualizing injection of oil directly into the uterine lumen.
The P implant is removed 49 h later and the animals are killed between
0 and 48 h after this.
Sequence of experimental steps for the mouse model of men-
Induction of the Mouse Model of Menstruation
All surgeries were performed under xylazine/ketamine-induced anes-
thesia. Mice were ovariectomized 7 days prior to the first of three daily
subcutaneous injections of 100 ng 17?-estradiol (Sigma Chemical Co., St.
Louis, MO) in arachis oil at approximately 0900 h. After resting the mice
for 3 days, progesterone (P) implants were inserted subcutaneously into
the back of each mouse and 5 ng of 17?-estradiol in arachis oil was
injected subcutaneously at approximately 0900 h on that and the subse-
quent 2 days. The progesterone implants were prepared essentially as de-
scribed by Milligan and Cohen  by filling silastic tubes (0.062 inches
inner diameter; Dow Corning, Midland, MI) with progesterone and sealing
the ends with polyethylene plugs, such that the functional length of each
implant was 1 cm. However, crystalline P (Sigma) was used rather than a
slurry of P in oil. Prior to use, the implants were incubated overnight at
37?C in phosphate buffered saline, pH 7.35 (PBS)/1% fetal calf serum
(Trace Biosciences, Sydney, Australia). At approximately 1100 h on the
day of the final 17?-estradiol injection, 20 ?l of sesame oil was injected
into the lumen of the right uterine horn of each mouse to induce deci-
dualization. The left horn remained untreated as a control. The P implants
were removed 49 h later. Mice were killed at the time of implant removal
(0 h) and 12, 16, 20, 24, 36, and 48 h thereafter (6–8 mice per time point)
and the uteri were harvested for further analysis. This sequence of events
is shown diagrammatically in Figure 1. Uteri were cleaned of fat and
weighed. Any mouse in which the oil-treated horn had not decidualized
(as evidenced by weight ?400% of the untreated contralateral horn; ap-
proximately 30% of animals) was excluded from the study. Some uterine
tissue was fixed in phosphate buffered formalin overnight and processed
to wax. Other tissue was snap frozen in liquid nitrogen. Blood was taken
from the mice at the 0- and 12-h time points for estimation of serum
The concentration of progesterone in serum was determined by the
Biochemistry Unit, Southern Cross Pathology (Clayton, Australia) using
a microparticle enzyme immunoassay (Abbott AxSYM System, Abbott
Australasia Pty. Ltd., Doncaster, Victoria, Australia). Concentrations were
expressed as nmol/L (mean ? SEM).
Uterine cross-sections of formalin-fixed tissue were deparaffinized and
hydrated by processing sections through Histosol (Sigma) and a graded
series of ethanol to distilled (d)H2O. The hydrated sections were stained
with hematoxylin and eosin using standard staining procedures. Stained
sections were dehydrated and mounted under coverslips using DPX
mounting medium (BDH Laboratory Supplies, Poole, England).
Morphological Analysis of Murine Endometrium
Hematoxylin and eosin-stained sections of murine uterus were scored
blind using an arbitrary scoring system, ranging from 0 to 4, such that 0
? no, 1 ? minimal, 2 ? moderate, 3 ? extensive, 4 ? profound signs of
destruction of the target cellular structure (including loss of adherence
between cells, rounding of nuclei, large interstitial spaces). Sections were
scored blind, with at least five sections from different mice within each
group analyzed. Semiquantitative analysis of the destruction of the deci-
dualized stromal tissue and the luminal epithelium was performed.
Immunohistochemical Analysis of Leukocytes
Transverse sections of formalin-fixed, paraffin-embedded mouse uteri
were deparaffinized in Histosol and hydrated via immersion in baths of
absolute ethanol, 70% ethanol, and dH2O. They were then immersed in
0.1 M citrate buffer and heated for 5 min in a 700-W microwave set to
medium. Once the slides had returned to room temperature, they were
rinsed in dH2O and immersed in 0.6% H2O2in dH2O for 30 min at room
temperature. After further rinsing in dH2O, the sections were bathed in a
blocking solution of Tris buffered saline (TBA; pH 7.5)/10% normal
mouse serum for 30 min at room temperature, after which time the block-
ing solution was replaced with 100 ?l of either rat anti-mouse CD45 (Cat.
No. 553076; Pharmingen, BD Biosciences, Lexington, KY), 5 ?g/ml in
TBS/10% normal mouse serum, or an equivalent amount of nonspecific
rat IgG for an incubation overnight at 4?C. Sections were washed thrice
in TBS with an additional wash step of TBS/0.5% Tween 20 between the
first and second washes in TBS. A secondary antibody solution consisting
of biotinylated rabbit anti-rat Ig (DAKO [Australia] Pty. Ltd., Botany,
NSW, Australia) in TBS/10% normal mouse serum was applied to each
section for 1 h at room temperature, followed by four washes as above.
The StrepABC kit and DAB solution (DAKO) were used in accordance
with the manufacturer’s specifications to reveal the CD45 staining. Sec-
tions were lightly counterstained with Harris hematoxylin (Accustain; Sig-
ma Diagnostics, Castle Hill, NSW, Australia), dehydrated, and mounted
using DPX mounting medium.
Stereology was used to determine the average number of leukocytes
per unit area (2729 ?m2) . Leukocytes were identified by staining for
CD45, and the DH Castgrid, version 1.6, software package (Olympus,
Denmark) was used to randomly move the counting frame (area of 2729
?m2) throughout the designated region of interest. At least 100 cells in
each category (stained and unstained) were counted. Leukocyte numbers
were analyzed separately in basal and decidualized areas of the tissues.
Cell counting was performed by the same observer and sections were
scored blind. Leukocyte numbers were also assessed in control tissues (no
oil injection, no decidualization).
TUNEL Staining to Detect Apoptosis
Hydrated cross-sections of formalin-fixed mouse uteri were deparaffin-
ized in Histosol and processed through a gradient of ethanol into dH2O.
Sections were immersed in methanol/3% H2O2for 30 min at room tem-
perature. Following two washes of 5 min in 0.01 M PBS (pH 7.4), slides
were placed on an ice-cold tray. DNA labeling mixture (50 ?l) was added
to each section; this solution contained 10 ?l 5? terminal deoxynucleo-
tidyl transferase (TdT) buffer, 1 ?l DIG DNA labeling mix, 2 ?l 2.5 mM
CoCl2, 0.2 ?l TdT (25U/?l) (all from Roche Diagnostics Australia Pty.
Ltd., Castle Hill, NSW, Australia) and 37 ?l MilliQ H2O. A glass coverslip
was used to contain the DNA labeling mix solution over the section during
the 30-min incubation at 37?C in a humid chamber. Control assays re-
placed TdT with an equal volume of MilliQ H2O. Sections were washed
twice in PBS for 5-min duration and then bathed in a blocking solution
of 20% normal rabbit serum in PBS for 10 min at room temperature in a
humid chamber. After this blocking step, approximately 100 ?l of sheep
anti DIG solution (Roche, 1.5 mU/?l in 10% normal rabbit serum/PBS)
was applied to each section and incubated for 30 min at room temperature
in a humid chamber. Sections were washed twice in 0.05 M PBS with
gentle shaking, bathed in a solution of biotinylated rabbit anti-sheep Ig 1:
500 (DAKO) in TBS/5% normal rabbit serum for 30 min at room tem-
perature and then washed twice in TBS. The StrepABC kit and DAB
solution (both from DAKO) were used in accordance with the manufac-
turer’s specifications to reveal the TUNEL staining. Sections were lightly
counterstained with hematoxylin, dehydrated, and mounted using DPX
mounting medium. The relative abundance of stromal cells that were
stained during TUNEL processing were compared by scoring each tissue
with an arbitrary scoring system such that 0 indicated no stained cells in
the tissue, through 4, where more than half of the cells were stained. All
sections were scored blind by the same observer.
All statistical analysis was performed using Prism 3.0. Comparison
between progesterone levels at 0 and 12 h and between weights of unstim-
ulated and stimulated uterine horns was made by paired t-test. Where a
number of groups were compared, analysis of variance was performed
A MOUSE MODEL FOR MENSTRUATION
lowing progesterone withdrawal. Open circles represent individual ani-
mals and bars show the mean value at each time point. * P ? 0.05, 20
h vs. 48 h.
Weight (mg) of stimulated uterine horns at times (in hours) fol-
following Bartlett test for equal variance. Individual differences were an-
alyzed by Dunnett multiple comparison test. In all cases, significance was
taken as P ? 0.05.
Morphology of the Uterus Following Oil Injection
and Progesterone Withdrawal
Following the injection of oil into the uterine lumen of
a sensitized mouse, there was an increase in the weight of
the stimulated horn harvested 49 h later. The weights of
the stimulated horns greatly exceeded those of their non-
stimulated counterparts (66.6 ? 7.5 vs. 16.0 ? 1.5 mg,
respectively, P ? 0.0001, n ? 8 animals). Following the
withdrawal of progesterone, the mean weight of the stim-
ulated uterine horns continued to increase until the 20-h
time point, at which time it was significantly greater than
the weight at 0 h (P ? 0.05), decreasing thereafter (Fig. 2).
The murine endometrium undergoes extensive remodeling
during the process of decidualization, and at 49 h following
the injection of the deciduogenic stimulus into the uterine
lumen (0 h after progesterone withdrawal: Fig. 3A), expan-
sion of the stromal cell population and their differentiation
into decidual cells was apparent. Glands were absent from
the decidual zone, but small glands could be found within
the basal zone, proximal to the myometrium. It has been
previously reported  that extensive neovascularization of
the tissue supports the increase in tissue size, although
blood vessels are largely absent from the primary decidual
zone. By this time, there is closure of the lumen, the surface
of which is generally mostly intact through re-epithelializa-
tion of the lumen damaged by the application of oil. These
features were also apparent in the present study (Fig. 3A)
and contrast with the control horn, to which a decidual
stimulus was not applied (not shown).
The withdrawal of progesterone support by the removal
of the subcutaneous progesterone implants led to a rapid
fall in serum levels of circulating progesterone from 180 ?
33 nmol/L at 0 h to 6 ? 2 nmol/L at 12 h (P ? 0.0001, n
? 3 mice per time point). However, the influence of this
change was not evident by 12 h, as indicated by the mor-
phology of the decidualized endometrium (Fig. 3B). At this
time, the morphology of the uterus was similar to that ob-
served at the 0 time point (49 h following the deciduogenic
stimulus, Fig. 3A), including the infiltration by red blood
cells throughout the tissue, presumably from the highly per-
meable blood vessels that are a feature of the decidual re-
sponse. However, by 16 h following the withdrawal of pro-
gesterone (Fig. 3C), changes in the structural integrity of
the decidualized endometrium became apparent. Large
spaces between decidual cells were evident, and fewer
blood vessels in the basal zone remained intact. This re-
duction in cell-cell contact was even more marked by 20 h
following progesterone withdrawal (Fig. 3D), when large
regions of necrotic decidual tissue were frequently ob-
served. The loosening of connective tissue in the basal zone
presumably contributed to the disassociation of the endo-
metrium from the myometrium, which was further exag-
gerated by 24 h following withdrawal of progesterone (Fig.
3E). Morphology during the subsequent 24 h showed ex-
tensive remodeling of the tissue with isolation of the dead
tissue (evidenced by loss of adherence between cells,
rounding of nuclei, and large interstitial spaces) within the
lumen apparent by 36 h after the withdrawal of progester-
one (Fig. 3F). This cellular debris was isolated within the
lumen following re-epithelialization, which is initiated an-
timesometrially from glands in the basal zone. By 48 h,
there was a relatively small quantity of debris remaining
within the lumen and the endometrium had undergone ex-
tensive restoration toward a predecidualized state (Fig. 3G).
Semiquantitative analysis of the morphological features
of the decidualized endometrium following the withdrawal
of progesterone was performed and representative sections
with a score of 0 (representing no evidence of tissue de-
struction), and 4 (indicating complete tissue destruction) are
seen in Figure 3, A and E, respectively. Such analysis fur-
ther emphasized the progressive destruction of the deci-
dualized zone of stromal cells (Fig. 4A) and luminal epi-
thelium (Fig. 4B). The extent of destruction of the stromal
tissue did not differ at the early time points (0, 12, and 16
h following the withdrawal of progesterone) but was sig-
nificantly greater at the 20- and 24-h time points (P ?
0.0001). The extent of destruction of the luminal epithelium
varied considerably within groups and occurred even in tis-
sues harvested at the early time points. Thus, no statistically
significant differences were recorded. Nonetheless, intact
epithelium was observed infrequently in the uteri harvested
at 24 h after the withdrawal of progesterone compared with
the earlier times.
Transverse sections taken through control uterine horns
(not decidualized) across the range of time points following
progesterone withdrawal did not show signs of tissue break-
Immunohistochemical Analysis of Leukocytes
The importance of leukocytes in the remodeling of the
human endometrium throughout the menstrual cycle 
and in particular, the dramatic increase in their numbers just
prior to menstruation, heralded these cells as targets for
analysis in this model. Leukocytes, identified by immuno-
staining for CD45, a marker of hematopoietic cells, were
present throughout the endometrium in both decidualized
and nondecidualized tissue, often in close association with
the luminal epithelium and throughout the basal zone (Fig.
3, H–J). They were also closely associated with the newly
generated luminal epithelium at the later time points (Fig.
3I). Leukocytes were also localized around glands, which
are located in the basal zone of the decidualized endome-
trium. At the time when breakdown of the decidual zone
BRASTED ET AL.
terone. Given the variability between animals, two representative uteri are shown at each time point following P withdrawal. A through H) Hematoxylin
and eosin-stained transverse sections of representative uterine horns. A through H) stimulated horns at (A) 0, (B) 12, (C) 16, (D) 20, (E) 24, (F) 36, and (G)
48 h after withdrawal of progesterone (images captured at ?20 magnification). H through J) Cells of hematopoietic origin, identifiedbyimmunohistochemical
staining for CD45 (leukocyte common antigen), shown by the brown coloration. At 16 h, an increase in CD45?cells could be seen in the basal zone (H
and insert). By 36 h, an increased abundance of CD45?cells was apparent within the area of breakdown (J) but were also seen in the regenerating tissue,
particularly in subepithelial areas (I). Bars ? 100 ?m (A through G and insert to H), 200 ?m (H), 50 ?m (I), and 300 ?m (J).
Morphologic changes and CD45 immunostaining in the decidualized endometrium at different time points following the withdrawal of proges-
A MOUSE MODEL FOR MENSTRUATION
stromal tissue (A) and the luminal epithelium (B) at different times follow-
ing progesterone withdrawal. Destruction is given in arbitrary units, where
0 ? no destruction and 4 ? complete destruction of the cellular com-
partment and destruction is evidenced by loss of adherence between cells
and increased interstitial spaces. Circles represent data for individual an-
imals. Bars represent the mean for each group. * P ? 0.001, 20 and 24
h compared with 0, 12, and 16 h.
Semiquantitative analysis of the destruction of the decidualized
al zone of stimulated (A) and nonstimulated (B) uteri following the with-
drawal of progesterone. Each circle represents the average number of
leukocytes per field of view (2730 ?m2) in a section from a single mouse.
Only sections of uteri that were (A) or were not (B) decidualized were
included for analysis. Bars represent the mean value within a group.
Stereological analysis of the abundance of leukocytes in the bas-
was well progressed (24 and 36 h), very large numbers of
leukocytes, most of which were identified morphologically
as macrophages, were present within the breakdown area
(Fig. 3J): given the variability between tissues at these time
points, leukocyte numbers were not assessed.
Stereological analysis of leukocyte abundance in the bas-
al zone in decidualized horns (Fig. 5A) revealed an increase
in cells per unit area within 12 h of the removal of the
implant (P ? 0.016 compared with 0-h control). This in-
crease in leukocyte abundance was maintained until 24 h
after progesterone withdrawal (Fig. 5A), at which time
analysis was impaired by the lack of structural integrity of
the tissue (as seen in Fig. 3E). Analysis of leukocyte abun-
dance throughout the entire tissue and decidual zone did
not indicate changes in the leukocyte populations (other
than those in the basal zone), although the changes in tissue
weight and size could potentially mask any such changes
(data not shown). Leukocytes were also counted in the bas-
al zone in control horns (no induction of decidualization).
Although the mean numbers of leukocytes per field was
higher than in the decidualized tissue, no changes in num-
bers were detected following progesterone withdrawal (Fig.
Assessment of Apoptosis
TUNEL staining was used to assess the extent of cell
death in the decidualized endometrium of mice and follow-
ing the withdrawal of progesterone. Quantitative scoring of
these tissues revealed that stromal cells containing nicked
DNA (indicative of cells undergoing the initial stages of
cell death) were present immediately before progesterone
withdrawal (49 h after the decidualizing stimulus) and de-
creased significantly in abundance by 16 h (P ? 0.05 at
each of 16, 20, and 24 h compared with 0 and 12 h; Fig.
6). The lack of a significant difference between the extent
of TUNEL staining at the time of withdrawal of the pro-
gesterone implant compared with 12 h later indicates that
the increase in leukocyte abundance observed at this 12-h
time point does not occur in response to increased cell
death following the withdrawal of progesterone.
BRASTED ET AL.
following progesterone withdrawal. The relative abundance of stromal
cells that were stained during TUNEL processing were compared by as-
signing a score to each section based on the extent of TUNEL staining of
the stromal cells of the decidual zone, whereby 0 ? no staining, 1 ?
scarce staining (1%–5% of cells), 2 ? some staining (6%–20% of cells),
3 ? frequent staining (20% –50% of cells), 4 ? extensive staining (?50%
of cells). Open circles represent the data for individual animals and bars
represent the mean values within each group. * P ? 0.05 for 0 and 12 h
compared with 16, 20, and 24 h.
Apoptosis in stromal cells in stimulated uteri at times (in hours)
This study has refined and evaluated a mouse model of
menstruation, originally described by Finn and Pope ,
whereby progesterone is withdrawn from mice in which the
endometrium has been artificially decidualized. In our
study, endometrial breakdown is observed as early as 16 h
after progesterone withdrawal. By 24 h, the decidual zone
is separated from the rest of the endometrium; by 36 h, re-
epithelialization is well progressed; and by 48 h, the tissue
debris is fully cleared and the endometrium restored to its
predecidual state. This mimics the remodeling events of
menstruation in which both tissue breakdown and re-epi-
thelialization of degraded endometrium proceed very rap-
idly (reviewed in ).
In humans, leukocyte numbers increase dramatically late
in the secretory phase of the cycle . In the present model,
the number of leukocytes per unit area increased signifi-
cantly in the basal zone by 12 h after progesterone with-
drawal, and this increase in numbers preceded the destruc-
tion of the stroma that was first evident at 16 h but signif-
icantly increased by 20 h. Given that the decidualized stro-
ma appears to separate from the basal zone, the increased
numbers of leukocytes at this site supports the contention
that leukocytes play an important role in initiation of tissue
breakdown via their production of MMPs, serine proteases,
and other bioactive factors and is in agreement with pre-
vious findings in the human . Leukocyte numbers per
unit area did not increase in the decidual zone following
progesterone withdrawal. The higher number of leukocytes
per unit area in the control horns was not surprising given
the relatively small stromal area compared with the area in
stimulated horns. However, the constant numbers found in
these control horns following progesterone withdrawal in-
dicates that the signals recruiting leukocytes must arise
from the decidual cells or from cells such as the displaced
glands, requiring cross-talk with the decidual cells.
In the human uterus, there is clear demarcation between
the basal layer of the endometrium and the inner functional
layer and it is largely the latter that takes part in menstru-
ation. Although there is no such clear demarcation in the
mouse endometrium, it does appear that there is a basal
layer of stroma that remains undifferentiated during the de-
cidual reaction in both pregnant and artificially stimulated
uteri . Indeed, it is likely that, in the mouse as in the
human, once the stroma has embarked on differentiation, it
cannot revert to its former state if progesterone support is
removed and must be shed.
Apoptosis has been shown to increase in the human en-
dometrium prior to and during menstruation, predominantly
in the epithelium. Little apoptosis is seen in the stroma until
menstruation has commenced [11, 12]. TUNEL staining
was performed to establish whether the tissue loss seen in
the present study was primarily due to increased apoptosis.
Although apoptosis was detected in the decidual zone im-
mediately before progesterone withdrawal, it was clearly
not induced by this withdrawal. Indeed, by 16 h, there were
significantly less apoptotic cells in the decidual zone than
at the time of progesterone withdrawal. Apoptosis in the
epithelium was not assessed due to the variation in epithe-
lial integrity in sections within each group. This pattern of
apoptosis resembles that seen in artificial menstrual cycles
in spayed macaques . In these animals, menstruation
began 2–3 days after removal of progesterone implants.
However, 12 h after progesterone withdrawal, there was a
consistent, dramatic increase in cell death by apoptosis, es-
pecially in the basalis. Leukocyte invasion into the stroma
occurred 1–2 days after progesterone withdrawal. This se-
quence of events is similar to that seen in our mouse model
and leads to the conclusion that the increase in leukocytes
in the basal zone relies on quite different signals from those
that regulate apoptosis.
No differences were seen between time points when de-
struction of the luminal epithelium was assessed, although
most was lost by 24 h. Although there was considerable
variability between animals, it was unusual to observe tis-
sues in which there was a complete absence of luminal
epithelium: this reflects the situation at menstruation in
women in whom re-epithelialization is thought to occur at
least in part from residual luminal epithelium [14, 15]. The
very rapid restoration of the luminal epithelium (within 48
h) in this model is also in agreement with postmenstrual
repair in the human and the rhesus monkey [16, 17]. Thus,
our mouse model provides also a useful model for studying
the events of endometrial repair.
Silastic implants containing progesterone have been used
to provide long-acting treatment in a variety of species,
including mice: serum hormone levels in treated mice re-
main steady over 21 days compared with the marked fluc-
tuations seen with daily subcutaneous injections . Fur-
thermore, withdrawal of such implants results in a very
rapid fall in progesterone concentrations (95% in 24 h ),
more resembling the rate of fall in women. This was seen
also in the present study in which progesterone levels were
minimal 12 h after implant withdrawal. As substantial var-
iation in morphology was seen at each time point following
cessation of progesterone injections, both in published work
 and in preliminary studies in our laboratory (data not
shown), we chose to use implants to better enable exami-
nation of the kinetics of events following progesterone with-
drawal. The very rapid sequence of events described here is
likely to be a consequence of the rapid fall in serum P.
Variability in the extent of decidualization between mice
A MOUSE MODEL FOR MENSTRUATION
was a problem in these and similar studies and is also seen
in women during the late secretory phase of the menstrual
cycle. In the present study, this was minimized as far as
possible by our choice of 49 h of decidualization as the
starting point. When establishing the model, we started with
24 and 36 h of decidualization as Time 0, but found con-
siderable variability in the extent of decidualization be-
tween mice. Indeed, it was not clear in some mice whether
there was complete failure to decidualize or whether deci-
dualization was merely retarded (data not shown). When 49
h was chosen as Time 0, there was still variability between
the extent of decidualization as reflected in the weights of
the stimulated horns, but any mice that failed to decidualize
could be clearly identified at the time of sacrifice and could
thus be excluded from the study.
One important difference between the mouse model and
menstruation in the human is the extent of decidualization.
During the normal human cycle, decidualization is initiated
close to the spiral arterioles during the mid to late secretory
phase and then spreads throughout the upper two thirds of
the endometrium . This predecidual cell expansion is
not only by cellular hypertrophy but also by mitosis , as
for decidualization in the mouse. Once pregnancy is estab-
lished, decidualization continues throughout the first tri-
mester . Highly decidualized endometrium is not com-
monly found in the absence of pregnancy but is often seen
in women using progestin-only contraceptives, particularly
the levonorgestrel-releasing intrauterine system . By
contrast with this slow progression in women, once deci-
dualization is induced in mice (only in the presence of a
blastocyst or by artificial stimulation), it progresses rapidly
from the site of initiation immediately below the uterine
epithelium and means have not yet been found to restrain
the process following induction. This is one limitation of
the mouse model of menstruation. There may be funda-
mental differences between stromal cells from rodents and
primates, as mouse and rat endometrial stromal cells deci-
dualize spontaneously when put in culture unless they are
derived from ovariectomized animals that have not been
treated with progestin . In contrast, human endometrial
stromal cells derived from cycling women in either the pro-
liferative or secretory phase of the cycle, decidualize in
culture only when treated with progesterone for ?6 days.
This can be accelerated by addition of decidualizing stimuli
such as interleukin-11 , activin A , and prostaglan-
din E2 or most commonly by cAMP.
It appears that decidualization or at least initiation of a
predecidual state in the endometrial stroma is a prerequisite
for menstruation. Although the stroma of rhesus and cy-
nomolgus macaques do not decidualize to the extent seen
in women during the nonfertile menstrual cycle, some en-
largement of the stromal cells is observed, especially
around the spiral arteries. This is likely to be similar in
kind but not in degree to the response of stromal cells to
progesterone in the human . Indeed, it is in the area
around the spiral arterioles that decidualization is initiated
in the human. In the menstruating bat (Glossophaga sori-
cina) during the late luteal phase, a swollen polyp contain-
ing large decidual-like cells is seen in the endometrium,
and this breaks down and bleeds at the end of the cycle
. It is important to note that, in the model described
here, tissue breakdown was not observed in the nondeci-
dualized horn, supporting the contention that at least some
progress toward stromal differentiation is required for ini-
tiation of tissue breakdown. Given the major phenotypic
changes that accompany decidualization , it is likely
that new secretory products from these cells initiate or pro-
mote the tissue breakdown.
This mouse model for menstruation thus demonstrates
many features in common with spontaneous menstruation
in higher primates. Given the wide availability of geneti-
cally modified mice and of inhibitors of many of the factors
thought to be involved in both menstruation and in tissue
repair, the model will allow functional studies to determine
which factors are critical to these processes and closer ex-
amination of the sequence of events. Such studies are es-
sential if we are to progress toward rational treatment of
disorders of menstruation and endometrial repair.
We thank Stuart Milligen and Colin Finn for helpful discussions before
this project commenced, Sue Panckridge and Samantha Park for assistance
in preparation of the manuscript, and Dr. Rebecca Jones for critically read-
1. Giudice LC, Ferenczy A. The endometrial cycle. In: Adashi EY, Rock
J, Rosenwaks Z (eds.), Reproductive Endocrinology, Surgery and
Technology. Philadelphia: Lippincott-Raven; 1996:271–300.
2. Salamonsen LA, Kovacs G, Findlay JK. Current concepts of the
mechanisms of menstruation. In: Smith SK (ed.), Baillie `re’s Clinical
Obstetrics and Gynaecology, Dysfunctional Uterine Bleeding, 13, 2nd
ed. London: Baillie `re Tyndall; 1999:161–179.
3. Dunn CL, Critchley HO, Kelly RW. IL-15 regulation in human en-
dometrial stromal cells. J Clin Endocrinol Metab 2002; 87:1898.
4. Salamonsen LA, Zhang J, Brasted M. Leukocyte networks and human
endometrial remodelling. J Reprod Immunol 2002; 57:95–108.
5. Finn CA, Pope M. Vascular and cellular changes in the decidualized
endometrium of the ovariectomized mouse following cessation of hor-
mone treatment: a possible model for menstruation. J Endocrinol
6. Milligan SR, Cohen PE. Silastic implants for delivering physiological
concentrations of progesterone to mice. Reprod Fertil Dev 1994; 6:
7. Vincent AJ, Malakooti N, Rogers PAW, Affandi B, Salamonsen LA.
Endometrial breakdown in women using Norplant is associated with
migratory cell expression of matrix metalloproteinase-9 (gelatinaseB).
Hum Reprod 1999; 14:807–815.
8. Finn CA, Pope MD, Milligan SR. A study of the early morphological
changes initiated in the uterine luminal epithelium by substances (oil
and carrageenan) which induce the decidual cell reaction in mice. J
Reprod Fertil 1989; 86:619–626.
9. Salamonsen LA. Tissue injury and repair in the human female repro-
ductive tract. Reproduction 2003; 125:301–331.
10. Salamonsen LA, Lathbury LJ. Endometrial leukocytes and menstru-
ation. Hum Reprod Update 2000; 6:16–27.
11. Dahmoun M, Boman K, Cajander S, Westin P, Backstrom T. Apopto-
sis, proliferation, and sex hormone receptors in superficial parts of
human endometrium at the end of the secretory phase. J Clin Endo-
crinol Metab 1999; 84:1737–1743.
12. Vaskivuo TE, Stenback F, Karhumaa P, Risteli J, Dunkel L, Tapanai-
nen JS. Apoptosis and apoptosis-related proteins in human endome-
trium. Mol Cell Endocrinol 2000; 165:75–83.
13. McClellan M, West NB, Brenner RM. Immunocytochemical locali-
zation of estrogen receptors in the macaque endometrium during the
luteal-follicular transition. Endocrinology 1986; 119:2467–2475.
14. Ferenczy A. Studies on the cytodynamics of human endometrial re-
generation II transmission electron microscopy and histochemistry.
Am J Obstet Gynecol 1976; 124:582–595.
15. Ludwig H, Metzger H, Frauli M. Endometrium: tissue remodelling
and regeneration. In: d’Arcangues C, Fraser IS, Newton JR, Odlind
V (eds.), Contraception and Mechanisms of Endometrial Bleeding.
Cambridge: Cambridge University Press; 1990:441–466.
16. Ludwig H, Spornitz UM. Microarchitecture of the endometrium by
scanning electron microscopy: menstrual desquamation and remodel-
ling. Ann N Y Acad Sci 1991; 622:28–46.
17. Okulicz WC, Ace CI, Scarrell R. Zonal changes in proliferation in the
rhesus endometrium during the late secretory phase and menses. Proc
Soc Exp Biol Med 1997; 214:132–138.
BRASTED ET AL.
18. Brenner RM, Maslar IA. The primate oviduct and endometrium. In:
Knobil E, Neill J (eds.), The Physiology of Reproduction. New York:
Raven Press; 1988:303–329.
19. Loke YW, King A. Human Implantation. Cell Biology and Immunol-
ogy. Cambridge: Cambridge University Press; 1995.
20. Critchley HOD, Wang H, Jones RL, Kelly RW, Drudy TA, Gebbie
AE, Buckley CH, McNeilly AS, Glasier AF. Morphological and func-
tional features of endometrial decidualization following long-term in-
trauterine levonorgestrel delivery. Hum Reprod 1998; 13:1218–1224.
21. Sananes N, Weiller S, Baulieu EE, Le Goascogne C. In vitro deci-
dualization of rat endometrial cells. Endocrinology 1978; 103:86–95.
22. Dimitriadis E, Robb L, Salamonsen LA. Interleukin 11 advances pro-
gesterone-induced decidualization of human endometrial stromal cells.
Mol Hum Reprod 2002; 8:636–643.
23. Jones RL, Salamonsen LA, Findlay JK. Activin A promotes human
endometrial stromal cell decidualization in vitro. J Clin Endocrinol
Metab 2002; 87:4001–4004.
24. Brar AK, Frank GR, Kessler CA, Cedars MI, Handwerger S. Proges-
terone-dependent decidualization of the human endometrium is me-
diated by cAMP. Endocrine 1997; 6:301–307.
25. Rasweiler JJ. Early embryonic development and implantation in bats.
J Reprod Fertil 1979; 56:403–416.
26. Popovici RM, Kao LC, Giudice LC. Discovery of new inducible
genes in in vitro decidualized human endometrial stromal cells using
microarray technology. Endocrinology 2000; 141:3510–3513.