Abstract
Breakthrough bleeding (BTB), a major side effect of long-acting progestogen (p)-only contraceptives in women, is the main reason for discontinuation of their use. To understand the mechanisms of BTB, a mouse model of endometrial breakdown and repair was adapted to evaluate the effects of long-term progestogens on the endometrium. Appropriately prepared mice received either an etonogestrel (ENG)- or levonorgestrel (LNG)-releasing subdermal implant. Forty eight hours after decidualization was induced in one uterine horn the majority of tissues were highly decidualized, designated 0 day (0d). Uteri were collected subsequently at 5-day intervals (to 45d) and both decidualized and non-decidualized horns were analysed for morphological changes, leukocyte infiltration and matrix metalloproteinase expression (MMP). In decidualized horns, large blood vessels (BV) developed and disturbance of tissue integrity was observed at 5d with substantial stromal breakdown by 10d, progressing until 25d when re-epithelialization was initiated. By 45d, the tissue was restored to its pre-decidualized state but with considerable tortuosity of the luminal epithelium. Tissue remodelling was not apparent in the non-decidualized horns before 35d, when hyperproliferation of the luminal epithelium resulted in tortuosity. Changes in morphology were similar with the two progestogens, but occurred more rapidly with LNG. Apart from macrophages, few leukocytes were present in non-decidualized horns but large infiltrates of neutrophils and uterine natural killer cells (uNK) were associated with tissue breakdown in decidualized tissue, many of these cells were MMP9-positive. MMP7 was primarily associated with tissue repair. Therefore, this model mimics some of the changes observed in endometria of women using p-only contraceptives and provides an opportunity for functional studies.
Introduction
Progestogen (p)-only contraceptives, such as the sub-dermal implants Implanon and Norplant or the levonorgestrel-releasing intrauterine system (LNG-IUS), provide long-term, safe and highly effective methods of birth control. However, unpredictable and often prolonged bleeding is associated with these contraceptives in approximately 30% of users and is the main reason given for their discontinuation. The mechanisms underlying this problem remain largely unknown. The different methods of delivery and local uterine concentrations of progestogen can have differing effects on the morphology of the endometrium. The LNG-IUS produces a high local concentration of progestogen and the endometrium becomes highly decidualized (Silverberg et al. 1986), while the slow release subdermal implants result in lower uterine progestogen concentrations and more variable endometrial morphologies. Tissue samples collected from women using Norplant have previously been characterized into three different subtypes: atrophic, shedding and decidualized (p-modified; Marsh et al. 1995, Vincent et al. 1999), although the bleeding episodes in these women did not directly correlate with the morphology of the tissue (Marsh et al. 1995, Vincent et al. 1999, Rhoton-Vlasak et al. 2005).
The normal menstrual cycle is guided by the steroid hormones oestrogen and progesterone which act primarily through and subsequently regulate expression of their specific receptors (oestrogen receptor (ER) and progesterone receptor (PR) respectively). The different methods of delivering contraceptive progestogens to the uterus result in various expression of steroid receptors in the endometrium, which may contribute to the bleeding disturbances. PR is suppressed in women using LNG-IUS, which exposes the endometrium to high levels of LNG (Critchley et al. 1998a). However, when exposed to lower uterine concentrations of progestogens, such as Implanon or Norplant, PR is high (Critchley et al. 1993, Macpherson et al. 1999). ER does not appear to be altered in Implanon users nor women using depot medroxyprogesterone acetate (MPA; Sereepapong et al. 2004), however, it is lower in endometrium from Norplant users when compared with normal proliferative phase endometrium (Critchley et al. 1993). Progestogens induce changes in ER and PR expression, which are likely to be involved in the observed histological differences and bleeding problems associated with specific p-only contraceptives.
Previous studies suggest a number of factors likely to contribute to the bleeding problems. These include an increase in number and size of small endometrial blood vessels (BV) as well as an increase in microvascular fragility (Rogers et al. 1993, Hickey et al. 1996, Pritts et al. 2005). Increased BV density was observed in the endometrium of women recently exposed to Norplant; however, the number of BV was not indicative of increased bleeding or spotting (Hickey et al. 1999). BV density was significantly increased in biopsies with an atrophic or regressed endometrial histology, however, this correlated to fewer bleeding days when compared with other histological appearances, suggesting that vessel integrity and function, not density, is important in bleeding (Hickey et al. 1999). Levels of vascular endothelial growth factor (VEGF) and tissue factor (TF) are also altered in the endometrium of women using Norplant (Runic et al. 1997, 2000, Lockwood et al. 2000, Schatz et al. 2003), and in particular TF was increased in the bleeding sites when comparedwith non-bleeding sites and may contribute to aberrant angiogenesis (Lockwood et al. 2000). Therefore, compromises in the vasculature and overall tissue integrity could play an important role in breakthrough bleeding (BTB).
Under the guidance of oestrogen and progesterone, the endometrium undergoes extensive tissue remodelling during each menstrual cycle. Synthetic steroids used in contraceptives influence the endometrium resulting in distinct morphological changes (Silverberg et al. 1986, Marsh et al. 1995, Hickey et al. 1999, Vincent et al. 1999, Rhoton-Vlasak et al. 2005). It is possible that molecules important for tissue remodelling will be involved in the endometrial changes observed in women using p-only contraceptives and contribute to overall tissue integrity. Many molecules thought to play important roles in tissue remodelling and angiogenesis are produced by leukocytes. During the menstrual cycle, specific subpopulations of leukocytes are recruited into the endometrium and immediately prior to menstruation can contribute up to 40% of the total cellular content of the tissue (Kamat & Isaacson 1987, Jeziorska et al. 1995, King 2000, Salamonsen & Lathbury 2000). Leukocytes produce a number of enzymes including matrix metalloproteinases (MMPs) which are capable of breaking down components of the extracellular matrix. MMPs are selectively upregulated just prior to menstruation (Hampton & Salamonsen 1994, Jeziorska et al. 1996, Zhang et al. 1998, Salamonsen & Woolley 1999, Zhang & Salamonsen 2002, Curry & Osteen 2003, Goffin et al. 2003). Furthermore, inhibition of MMP activity in explant endometrial cultures demonstrated a critical role for these enzymes in endometrial breakdown (Marbaix et al. 1996). Leukocytes and MMPs are also abundant in endometrium exposed to p-only contraceptives, in similar numbers to those observed in normal menstrual phase tissue and are particularly high in endometria displaying a shedding morphology (Galant et al. 2000, Marbaix et al. 2000, Vincent & Salamonsen 2000, Vincent et al. 2000, 2002).
The mechanisms underlying BTB and indeed menstruation and menstrual disorders are difficult to study, given that they occur only in women and a few old world primates (Shaw et al. 1972, Goncharov et al. 1976, Kaiserman-Abramof & Padykula 1989, Rasweiler 1991). The invasive nature of collecting human endometrial samples and the lack of suitable non-primate models limits investigations where these processes can be manipulated and subsequently elucidated. A major difference between women and other mammals is that very few species undergo spontaneous decidualization during a non-fertile cycle as occurs in women, therefore making direct comparisons with women difficult. We previously developed a mouse model for endometrial breakdown and repair (Brasted et al. 2003), in which we observed an increase in MMP expression and an influx of leukocytes (Kaitu’u et al. 2005) after P withdrawal, consistent with what is observed during menstruation in women. Importantly, endometrial breakdown in this model was only observed in the uterine horns in which decidualization had been induced and no changes were seen in the non-decidualized horn (Brasted et al. 2003, Kaitu’u et al. 2005). Therefore, it appears that decidualization, or at least some of the changes that accompany decidualization may be important contributors to endometrial instability.
The aims of the present study were to develop a mouse model, in which the effects of long-term administration of the progestogens used in Implanon and Norplant could be examined and manipulated. Morphology of the tissue and BV, MMP expression and leukocyte influx were examined in both the decidualised and the non-decidualized uterine horns which represent the different endometrial morphologies seen in women using p-only contraceptives.
Materials and Methods
Animals
Female C57BL/6 mice age 8–12 weeks, were obtained from Monash University Animal Services and housed under standard conditions with food and water and allowed to feed ad libitum with a constant light cycle of 12 h (lights on from 0800 to 2000 h). Ethics approval was granted by the Monash University/Monash Medical Centre Animal Ethics Committee B.
Mouse model for long-term progesterone exposure
A previously described mouse model for endometrial breakdown and repair (Brasted et al. 2003, Kaitu’u et al. 2005) was adapted to examine the effects of long-term exposure (up to 45 days) to etonogestrel (ENG) or LNG on decidualized and non-decidualized endometrium. Briefly, ovariectomized mice received three daily (at approximately 0900 h) s.c. injections of 100 ng 17β-estradiol (Sigma Chemical Co.) in arachis oil. The mice were then rested for 3 days before insertion of one-fourth of a standard Implanon rod (containing ENG; gift from Organon, Oss, The Netherlands) or a silastic implant (0.062 inches internal diameter, Dow Corning, Midland, MI: functional size 1 cm length see (Kaitu’u et al. 2005)) containing LNG (Sigma, as contained in Norplant) subcutaneously into the back of the animals. Simultaneously, the mice received a single s.c. injection of 500 ng progesterone (Sigma) in arachis oil and a series of three 5 ng 17β-estradiol injections on this and the following 2 days. At approximately 1000 h on the day of the final 17β-estradiol injection, 20 μl sesame oil was injected into the lumen of uterine horn to induce artificialdecidualization (stimulated horn). The left horn was left untreated (non-stimulated horn). At 48 h following the oil injection, the injected horn had undergone extensive decidualization; this was designated day 0 (0d). Mice were sacrificed at 5-day intervals from 0d until 45d (ENG) or 15d (LNG) and the uteri were collected (n ≥ 3 at each time point). Fig. 1 details the treatment regime. All surgeries were performed under xylazine/ketamine-induced anaesthesia.
Tissue collection
Uteri were cleaned of fat, weighed and each horn was cut into three pieces, which were: (1) fixed overnight at 4 °C in phosphate buffered formalin, (2) fixed for 2 h at room temperature (RT) in Carnoy’s fixative or (3) snap-frozen in liquid nitrogen and stored at −80 °C. Formalin-and Carnoy’s-fixed tissues were subsequently processed to wax.
Histology and Immunohistochemistry
Uterine cross sections (5 μm) of formalin- or Carnoy’s-fixed tissues were processed through Histosol (Sigma) and a graded series of ethanol to distilled water (dH2O) prior to immunostaining or histological staining. For histological analysis, hydrated sections were stained with haematoxylin and eosin using standard staining procedures and mounted using dibutyl polystyrene xylene (DPX, BDH Laboratory Supplies, Poole, England).
Immunohistochemistry was conducted to detect MMP3, MMP7 and MMP9 in cross sections of uterus at different times following decidualization as previously described (Kaitu’u et al. 2005). Rabbit anti-mouse MMP3 was a gift from Dr L Moons and rabbit anti-rat MMP7 was a gift from Prof. J F Woessner Jr (Yu & Woessner 2000). Goat anti-mouse MMP9 was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Antigen retrieval was performed on Carnoy’s-fixed tissue by heating sections in 0.1 M citrate buffer for 5 min in a 700 W microwave set to medium. The slides were allowed to cool to RT and were then rinsed in dH2O. Endogenous peroxidase activity was quenched by immersion of sections in 3% H2O2 in methanol for 30 min at RT. Sections were then incubated with blocking solution containing 10% normal goat serum (MMP3 and MMP7) or 20% normal horse serum (MMP9) in Tris-buffered saline (TBS) pH 7.6 for 20 min at RT. Sections were incubated for 1 h with primary antibodies diluted in 10% foetal bovine serum (FBS)/TBS to 1 μg/ml (MMP3) in the ratio of 1:300 (MMP7) or 1 μg/ml (MMP9), then washed sequentially in TBS, 0.6% (v/v) Tween 20 in TBS and thrice in TBS. Biotinylated goat anti-rabbit IgG for MMP3 and MMP7, and biotinylated horse anti-goat IgG (Dako, Glostrup, Denmark) for MMP9, diluted at a ratio of 1:200 in 10% FCS/TBS, were applied for 45 min at RT and the slides were washed as described above. For all protocols, the StreptABC horse radish peroxidase (HRP) kit and diaminobenzidine (DAB) solution (Dako) were used in accordance with the manufacturer’s specifications to reveal the MMP3, −7 and −9 staining.
Immunolocalization of neutrophils was performed on formalin-fixed tissue using the rat anti-mouse mono-clonal MCA771GA (Serotec, Oxford, UK) diluted to 20 μg/ml in 10% FBS/TBS. Antigen retrieval (0.1 M citrate buffer; 5 min at high level and 3 min at medium-low level in a 700 W microwave) preceded endogenous peroxidase activity quenching (3% H2O2 in methanol for 5 min). Sections were blocked with 20% normal rabbit serum for 5 min at RT then incubated with primary antibody for 30 min at RT. Sections were washed as specified for MMPs above, followed by incubation with biotinylated rabbit anti-rat IgG (Dako; 1:200 v/v in 10% FBS/TBS for 30 min). StreptABC-HRP reagent preceded colour development with DAB (Dako).
Immunolocalization of macrophages was performed on Carnoy’s-fixed tissue using the rat anti-mouse F4/80 pan macrophage antibody (BMA Biomedicals, Rheinstrasse, Switzerland). Microwave antigen retrieval (10 min at high level in a 700 W microwave) was followed by an incubation with a peroxidase blocking agent (3% H2O2 in methanol for 10 min) and the primary antibody (0.81 μg/ml diluted in 10% FBS/TBS) at 37 °C for 30 min. Biotinylated rabbit anti-rat IgG (Dako; 1:200 diluted in 10% FBS/TBS) was incubated with sections for 30 min at RT, followed by StreptABC HRP reaction for 30 min and colour development using DAB (Dako) for 5 min.
To detect uterine NK cells, Carnoy’s-fixed sections were incubated with peroxidase blocking agent (3% H2O2 in methanol for 5 min) prior to blocking in 20% FBS/TBS for 10 min at RT. NK cells were detected using biotinylated lectin-DBA (Sigma) diluted to 50 μg/ml in 10% FBS/TBS for 40 min at 37 °C, followed by StreptABC-HRP (Dako) for 30 min at RT and the fuchsin substrate-chromogen system (Dako) for 5 min at RT. A microcomputer imaging device (AIS: Imaging Research, Brock University, Canada) was used to assess the relative number of uNK cells (expressed as a percentage of total tissue area strongly stained) in each tissue.
Immunohistochemistry was conducted using the BV marker CD34, an affinity purified rat anti-mouse IgG (Serotec). Formalin-fixed tissues were dewaxed and rehydrated prior to quenching of endogenous peroxidase activity with 3% H2O2 in methanol for 5 min at RT. Tissues were blocked in 20% normal rabbit serum in TBS for 5 min at RT before incubating with anti-mouse CD34 (0.5 μg/ml in 10% FBS/TBS) for 30 min at 37 °C. Biotinylated rabbit anti-rat IgG (Dako; 1:200 diluted in 10% FBS/TBS) was incubated with sections for 30 min at RT, followed by StreptABC HRP reaction for 30 min and colour development using DAB (Dako) for 2 min.
Immunolocalization of ER-α was performed using an affinity purified rabbit anti-human polyclonal antibody (ER-α; H-20:sc-543, Santa Cruz Biotechnology). Carnoy’s-fixed tissues were subjected to antigen retrieval (3 min at high, 5 min at medium then 1 min at high level in a 700 W microwave in 0. 1 M citrate buffer) followed by quenching of endogenous peroxidase activity (3% H2O2 in methanol for 10 min at RT) and blocking (20% normal goat serum/TBS 10 min at RT). Primary antibody was incubated with the tissues (2 μg/ml) for 60 min at 37 °C. Goat anti-rabbit/HRP conjugated polymer (Dako EnVision+System Kit) was then incubated with the tissue for 30 min at RT and visualized using DAB (Dako) for 2 min.
Immunolocalization of progesterone receptor (PR) was performed using the rabbit anti-mouse, rat and human polyclonal antibody (PR-A and PR-B; C-20:sc-593, Santa Cruz Biotechnology). Endogenous peroxidase activity was quenched in formalin-fixed sections (3% H2O2 in methanol for 10 min) followed by antigen retrieval (0.1% trypsin in 0.1% calcium chloride for 30 min at RT) and blocking in 20% normal goat serum/TBS for 10 min at RT. Sections were then incubated with primary antibody (2 μg/ml) for 1 h at 37 °C and washed with 0.6% Tween 20/TBS. Tissues were then incubated with goat anti-rabbit HRP (Polymer, Dako Envision+kit) for 30 min at RT and visualized using DAB (Dako) for 3 min.
All sections were lightly counterstained with Harris hematoxylin (Accustain; Sigma Diagnostics), dehydrated and mounted using DPX mounting medium. Negative controls were included for each tissue section by substitution of the primary antibody with a matching concentration of normal rabbit IgG for MMP3 and-7, neutrophils, macrophages, ER and PR, and normal goat IgG (Sigma) for MMP9.
In situ zymography
Tissues were examined for the presence of active forms of gelatinases, by in situ zymography (ISZ), as previously described (Zhang & Salamonsen 2002). In brief, 7 μm frozen sections were cut on to poly-l-lysine coated slides, fixed in 10% buffered formalin for 5 min at 4 °C and washed thrice with cold TBS. The substrate (DQ™ gelatin from pig skin, fluorescein conjugate, Molecular Probes, Inc., Eugene, OR, USA) was dissolved to a final concentration of 25 μg/ml in a mixture of 2% gelatin and 2% sucrose in TBS with 0.02% sodium azide and 100 μl layered over the tissue section, covered with a coverslip and incubated in a darkened humid chamber at 37 °C for 16 h. Each section was viewed using an Olympus Corp. (Birkeroed, Denmark) fluorescent microscope with fluorescein isothiocyanate (FITC) filter.
Statistics
A one way ANOVA with Tukey’s post hoc test was used to determine differences in weights between stimulated or non-stimulated uterine horns at different time points. A P value of < 0.05 was deemed statistically significant. Results are presented as mean weight +s.e.m.
Results
Changes in uterine weight
Mice treated with ENG were culled at 5-day intervals between 0d (48 h after induction of decidualization) and 45d. The average weights (±s.e.m.) of stimulated and non-stimulated horns at each time point are shown in Fig. 2A. Weights of stimulated horns increased to a peak at 10d when they were >sixfold heavier than on 0d (P<0.001). The weights then decreased at 15d, but were still significantly greater than at 0d (P<0.001). By 20d they were no longer significantly heavier than on 0d. The non-stimulated horns showed a small but non-significant increase in weight from 0d to 25d but at 35d and 45d were significantly heavier than at 0d (P<0.05 and < 0.01 respectively).
The stimulated horns collected from mice treated with LNG displayed a similar significant increase in weight to those collected from ENG-treated mice (~fivefold increase from 0d to 5d, P<0.001, Fig. 2B). However, the peak weight was observed at 5d (when compared with 10d for ENG mice) and at 10d the weight of the stimulated horn had decreased and was not significantly different from 0d. Again there was very little variation in the weights of the non-stimulated horn; only at 10d there was a significant increase from 0d (P<0.05).
Morphological changes
The morphology of the uterine tissue was assessed at each time point by haematoxylin and eosin staining. Representative photomicrographs are shown in Fig. 3. At 0d in the stimulated horn, stromal expansion and differentiation of the decidual cells was apparent. Closure of the lumen was observed and small glands were seen in the basal zone (Fig. 3, 0d morphology). At 5d, large BV were obvious in the central decidual zone and confirmed by CD34 staining of endothelial cells (Fig. 3). These were not detected at 0d demonstrating a paucity of such vessels in early decidua. At 5d, there was also some loss of structural integrity and red blood cells (RBC), and leukocytes were evident in the decidual zone close to BV (Fig. 3, 5d). At 10d, there was clear evidence of stromal breakdown, with large areas of destruction within the decidual zone and an abundance of RBC in the lumen (Fig. 3, 10d). Structural integrity of the BV was compromised at 15d (Fig. 3). This breakdown progressed from 15d–25d (Fig. 3) and in some areas the decidual zone appeared to part almost completely from the myometrium (Fig. 3, 20d). By 25d, there was some evidence of re-epithelialization and large numbers of RBC along with cellular debris were isolated within the lumen (Fig. 3). At 35d, the debris was almost entirely cleared from the lumen and re-epithelialization was often observed to be complete (Fig. 3, 35d). At 45d, the stimulated horn had completely restored to a pre-decidualized state (Fig. 3, 45d), however, the luminal epithelium appeared highly tortuous. At these later time points after tissue restoration (25–25d), BV staining was observed throughout the stroma, similar to that of the non-stimulated horns at these times (data not shown).
In comparison, the non-stimulated horn (Fig. 4) showed no signs of stromal differentiation at 0d with glands scattered throughout the stroma (Fig. 4, 0d). Supporting the lack of change observed in the weights of the non-stimulated horns, these tissues remained largely unchanged in morphology from 0d to 25d (Fig. 4). There appeared to be some stromal oedema at 25d which then increased overtime. Furthermore, at 35d and 45d (Fig. 4) the luminal epithelium of the non-stimulated horn became more tortuous, resembling that of the stimulated horn at these time points and possibly contributing to the slight change in weights observed at the later time points.
The morphology of the uteri from LNG-treated animals did not differ from that in the ENG-treated animals. In the stimulated horn (Fig. 5A), there was closure of the lumen at 0d with signs of stromal decidualization and glands were located in the basal area. At 5 and 10d, large BV, similar to those stained for endothelial cells in Fig. 3, were apparent and there was loss of structural integrity and evidence of tissue breakdown. At 15d there were some signs of re-epithelialization and tissue regeneration, reflecting the more rapid change in uterine weight in these animals when compared with those treated with ENG. In contrast, the non-stimulated horn (Fig. 5B) showed no signs of decidualization, but importantly at 15d there was increased tortuosity of the luminal epithelium and stromal oedema similar to that seen at the later time points in ENG-treated mice.
Oestrogen and progesterone receptors
In the stimulated horns collected from ENG-treated mice, nuclear ER staining was present in cells throughout the stroma in tissues that were not yet breaking down (0–10d, representative photomicrograph Fig. 6A, 5d). By 15d, when tissue integrity was clearly compromised, immunoreactive ER was lost from stromal cells and localized to the nuclei of large BV and glands (not shown). At later time points when the tissue was undergoing repair, ER was detected in the endometrial glands located in the basal area of the tissue (Fig. 6A, 35d).
Very little PR staining was detected in any cells at any of the time points (not shown). However, positive control uterine tissue collected from normal cycling mice showed a high level of staining throughout the endometrium (data not shown), confirming that PR are downregulated in decidualized endometrium exposed to long-term-ENG.
Matrix metalloproteinases
MMP9
At 0d, isolated MMP9-positive cells were scattered throughout the decidual zone of the stimulated horns collected from ENG treated mice (Fig. 6B). Numbers were greatly increased at 5d, particularly in areas of breakdown and maintained until 20d (representative photomicrograph Fig. 6B). At 35d, the number of MMP9-positive cells dramatically decreased and only a few isolated positive cells remained (not shown). Gelatinase activity (which includes MMP9 and −2 activities; green fluorescence) was confirmed using ISZ in these tissues at 5d (Fig. 6B 5d, ISZ). No MMP9-positive cells were detected in the non-stimulated control horns at any time point (not shown).
MMP7
Isolated MMP7-positive cells were observed in the stroma and the breakdown areas in the stimulated horns from 0–20d (representative photomicrograph Fig. 6C). At the later time points (25–35d), MMP7 was mainly associated with the process of repair and re-epithelialization and was often observed near areas of newly formed epithelium (Fig. 6C, 35d). At 45d, when the tissue had completely regenerated, there was no MMP7-positive staining (not shown). MMP7 staining was not consistently detected in the non-stimulated horn at any time point, although there were occasional isolated positive cells in some sections (not shown).
MMP3
Immunopositive staining for MMP3 was present throughout the time course in the stimulated uterine horns. Diffuse interstitial staining was observed at isolated foci within the decidua and did not appear to be altered during tissue breakdown or repair (representative photomicrograph Fig. 6D). There was no significant MMP3 staining in the non-stimulted horn at any time point (not shown).
Leukocytes
In the tissue collected from ENG-treated mice, very few uNK cells were observed in the non-stimulated horn or in decidualized endometrium at 0d (not shown). Immunostaining revealed a dramatic increase in uNK cells in the stimulated horn at 5d, particularly in the mesometrial region containing large vessels (Fig. 7A). The uNK cells were located throughout the stroma in this area, with no obvious relationship to BV (Fig. 7A). By 10d, the number of uNK cells had not greatly increased further but they were now mainly located within the breakdown area (Fig. 7A, 10d). At 15d, uNK cell numbers decreased, coincident with the onset of tissue repair (not shown). Semi-quantitative analysis of the proportion of the tissue stained for uNK cells showed that approximately 20% of the tissue stained for uNK cells at 5 and 10d (Fig. 8). This decreased to approximately 5% by 15d. The level of staining at other time points was below the level of detection.
Very few neutrophils were detected in the non-stimulated horns at any time points. Neutrophils were present in low numbers in the stimulated horn at 0d, and increased only slightly throughout the tissue at 5d (Fig. 7B). These cells were mostly located within BV with a few isolated cells located within the stroma. At 10d, when tissue breakdown was first apparent, large numbers of neutrophils were located within these breakdown areas and isolated neutrophils were detected in the surrounding decidua (Fig. 7B). High numbers of neutrophils persisted at 15d, but these were located almost entirely within the cellular debris (not shown). After 15d, coincident with the onset of tissue repair, the numbers of neutrophils decreased until very few remained at 45d when the debris had been entirely cleared from the lumen and the tissue had completely repaired (not shown).
Macrophages were present in the stimulated horn at 0d in the basal area and near the lumen, and were consistent in number from 0–10d (representative photo-microgragh Fig. 7C). At later time points, macrophage numbers increased at sites of tissue repair (Fig. 7C, 20 and 35d).
In the non-stimulated horn, macrophages were abundant throughout the time course. These were located within the stroma and often underlying the luminal epithelium (Fig. 7D).
Discussion
This study has adapted a previously described mouse model of endometrial breakdown and repair (Brasted et al. 2003) to investigate the long-term effects of progestogens contained in certain p-only contraceptives on the endometrium. A range of endometrial morphologies are found in women using long-term progestogens, including atrophic, highly decidualized and tissue undergoing breakdown (Silverberg et al. 1986, Critchley et al. 1998a, Vincent et al. 1999, Phillips et al. 2003). For this reason, we subjected one uterine horn of mice treated with either ENG or LNG to a decidualizing stimulus, while the other remained non-decidualized. The study demonstrated similarities between this mouse model and what is observed in tissue collected from women using p-only contraceptives. The horn in which decidualization was induced, underwent breakdown and repair, while the non-decidualized horn slowly increased in weight but did not undergo major remodelling. In both horns, the luminal epithelium became highly tortuous with prolonged progestogen exposure. Prior to the breakdown of the decidualized endometrium, there was an increase in the number and size of BV and a transient influx of neutrophils and uNK cells that persisted throughout the breakdown period. Immunopositive MMP9 cells were observed in high numbers in areas of breakdown, while MMP7 and -3 were consistently observed in decidualized, breaking down and repairing tissue. MMP3 was present in foci throughout the tissue at all time points. Although MMP7 was initially localized to isolated cells within the stroma and breakdown area, in repairing tissue it was mainly localized to regenerating epithelium.
In women using p-only contraceptives, endometrial morphology varies with the type of contraceptive used and the method of progestogen delivery. When high local concentrations of progestogen are delivered to the endometrium using an intrauterine device, such as the LNG-IUS, extensive decidualization and epithelial atrophy were observed and the tissue remains decidualized until the device is removed up to several years later (Silverberg et al. 1986, Critchley et al. 1998a, 1998b, Phillips et al. 2003). However, the decidua induced by artificial stimulation (such as oil injection) in mice, has a finite life span of approximately 8 days before spontaneous breakdown occurs. This time can be increased to 14 days following the administration of progesterone (Hirabayashi et al. 1999), as seen in mice treated with ENG in the present study. However, in mice treated with LNG, the weights of the stimulated horns were reduced at 10d, highlighting the differences between different progestogens. Furthermore, the finite life-span of the artificial decidua in mice is a clear and fundamental difference between the endometrium of mice and women.
Subdermal implants, such as Norplant or Implanon, result in lower intrauterine progestogen concentrations and produce a more variable effect on endometrial morphology (Marsh et al. 1995, Vincent et al. 1999). Tissue from Norplant users has been morphologically subclassified into atrophic (narrow proliferative-type epithelial glands), shedding and progestogen-modified (highly decidualized) endometrium (Marsh et al. 1995, Vincent et al. 1999). Thus, the non-decidualized horn of the mouse model would be expected to more closely represent the non-decidualized morphologies. However, neither atrophy nor shedding was observed and the most striking feature of non-decidualized horns was the hyperproliferation of the luminal epithelium, resulting in tortuosity, after 25d. This was also observed in the restored but originally-stimulated horn.
Two different progestogens were compared in this study. Although the overall morphology of the tissue was very similar in both LNG- and ENG-treated animals, the processes appeared to be retarded in ENG-treated mice highlighting the different effects of specific progestogens. While endometrial morphology has been well described in Norplant exposed tissue, such information for Implanon-exposed tissue is very limited and the different effects of the two progestogens on endometrial morphology require further investigation.
BTB is thought to arise from superficial vessels within the endometrium which increase in number and fragility in Norplant users (Hickey et al. 1996, Hickey & Fraser 2002). We also observed substantial changes in BV in the mouse model. Large BV with thin walls appeared after 5 days and continued to increase in number and luminal volume over the subsequent 15 days. However, at 15 days these vessels appeared to lose structural integrity and become leaky. Areas of condensed RBC within the stroma, indicative of haemorrhage, were evident. Studies using a different mouse model have also reported changes in endometrial vasculature in response to either MPA or LNG (Girling et al. 2004). However, in this model, normal cycling mice were exposed to progestogens in the absence of any decidualizing stimulus. Despite the differences between the models, the changes observed in the endometrial vasculature highlight the important role that progestogens exert on vessel size and fragility, even in the absence of decidualization.
The normal menstrual cycle is driven by the steroid hormones, oestrogen and progesterone, which function via specific receptors. PR is expressed in the stroma throughout the menstrual cycle with a slight decrease during the menstrual phase (Critchley et al. 1993). In the glands, PR protein is maximal in the mid-late proliferative and early secretory phases, then downregulated during the late secretory phase, remaining low during the menstrual and early proliferative phase (Critchley et al. 1993, Ingamells et al. 1996). In Norplant users, immunoreactive PR is higher in the stroma when compared with normal endometrial samples at any stage of the cycle and is also high in glands (Critchley et al. 1993). ER however, is expressed at lower levels in the stroma and glands in Norplant users when compared with normal proliferative phase endometrium (Critchley et al. 1993). Neither PR nor ER is altered in women using depot MPA experiencing frequent bleeding episodes when compared with women who experienced amenorrhea (Sereepapong et al. 2004). However, in women using LNG-IUS, PR is suppressed (Critchley et al. 1998a). This is likely a result of the high endometrial concentration of progestogen and corresponds to our observations in this mouse model in which the dose of progestogen was high. Therefore, our studies support those of others, showing that ER and PR expression can be altered by progestogens. It is also important to note that the mice in the present study were ovariectomized although the diet contained phytoestrogens which have been shown to affect the uterus (Britt et al. 2005).
Uterine NK cells appear in human endometrium in association with the induction and continuation of decidualization (Marshall & Jones 1988, Croy & Kassouf 1989, King et al. 1989) and have also been observed in high numbers in highly decidualized progestogen-modified human endometrium (Critchley et al. 1998b). Furthermore, high numbers of uNK cells have been associated with the bleeding episodes in menopausal women receiving hormone replacement therapy (Hickey et al. 2005). This is consistent with the high numbers of uNK cells in the highly decidualized tissue prior to breakdown in this mouse model. In mice, uNK cells are a major source of VEGF (Wang et al. 2003) and studies in uNK cell-deficient mice have defined a role for these cells in the modification of spiral arterioles during pregnancy (Guimond et al. 1997, Croy et al. 2003). In the present study, we showed that uNK cells were present in high numbers in tissue that contained large thin-walled BV. Although a functional relationship remains to be established, it is likely that these cells are an important contributor to the BV remodelling in this model.
Consistent with observations in normal pre-menstrual tissue, we observed high numbers of neutrophils infiltrating the breakdown areas of the mouse tissue. These cells also produce many cytokines and proteases including elastase and MMPs, such as MMP9 and MMP7, that are thought to contribute to tissue breakdown (Kamat & Isaacson 1987, Jeziorska et al. 1995, King 2000, Salamonsen & Lathbury 2000, Zhang & Salamonsen 2002). The presence of these cells in the breakdown areas in this model supports this role and is in agreement with observations from progestogen-exposed human tissue where elevated numbers of neutrophils are also seen in areas of breakdown (Vincent et al. 1999).
MMPs degrade components of the extracellular matrix as well as regulating their own activity and that of other bioactive molecules. Inhibition of MMP activity in human explant endometrial cultures demonstrated a critical role for these enzymes in endometrial breakdown (Marbaix et al. 1996). MMPs are also postulated to have an important role in irregular bleeding with p-only contraceptive use (Skinner et al. 1999, Vincent et al. 1999, 2000, Vincent & Salamonsen 2000). In the present model, immunopositive MMP9 was observed in low abundance in decidualized tissue with an increase observed in association with breakdown. This expression pattern supports previous data that showed an upregulation of MMP9 in endometrial tissue, with a shedding morphology collected from Norplant users and in women who experience dysfunctional uterine bleeding, consistent with a likely role in tissue breakdown (Vincent et al. 1999, Galant et al. 2004). MMP3 was consistently expressed at foci during the process of tissue breakdown and repair in the decidualized horn, supporting human studies showing that MMP3 is elevated in endometrium from women using Norplant, DepoProvera and LNG-IUS (Vincent et al. 2000, Galant et al. 2004, Oliveira-Ribeiro et al. 2004). Interestingly, MMP7 was mainly associated with the process of repair and re-epithelialization and was predominantly localized to the regenerating epithelium. This staining pattern supports previous studies showing MMP7 expression is highly important for wound repair and epithelial cell migration in the epithelial tissue of damaged airways (Parks et al. 2001). MMP7 has also been shown to be upregulated in intestinal re-epithelialization (Salmela et al. 2004). The results from the present study would support such a role for epithelial-associated MMP7 as it was localized to newly formed epithelium in the repairing endometrium and is consistent with a similar location in our mouse model for endometrial breakdown and repair, a model that mimics some key features of menstruation in women (Kaitu’u et al. 2005). Although it was shown in the breakdown and repair mouse model that MMPs were not the key contributors to these processes, investigations into the functional role of MMPs in the long-term p-only model have been unsuccessful as we were unable to completely block MMP activity using specific inhibitors. Therefore, we can only speculate on their role in this model based on localization, activity and previous studies. The localization pattern of MMP expression is consistent with the present hypothesis that they play a role in endometrial breakdown and repair.
While this model mimics some characteristics of endometrium of women using p-only contraceptives, we acknowledge some limitations. Very high doses of progestogen were administered in the present model and to reduce this may prove difficult, particularly for ENG which was provided as Implanon rods, prepared for women, not mice and which when cut, present complex pharmacodynamics. In addition, there appear to be fundamental differences between decidua in mice and in women, both with respect to their different finite lives and to their mechanisms of induction.
In summary, the present study describes a mouse model for the study of the effects of long-term progestogen exposure on the endometrium and demonstrates some similar characteristics to those observed in the endometrium of women using p-only contraceptives. Studies in women are limited for obvious reasons and this model hence offers an opportunity for elucidating the potential functions of key modulators of endometrial fragility in BTB.
A time line describing the mouse model of long-term progestogen exposure. Ovariectomized mice were allowed to recover for 7 days before receiving a series of oestrogen injections. P-implants were inserted along with a second series of oestrogen injections to prime the uterus prior for the administration of a decidualizing oil stimulus into one uterine horn. Substantial decidualization had occurred 48 h after the stimulus was given (0d) and mice were culled from this time point at 5-day intervals up to 45 days (0–45d).
Citation: Reproduction 133, 1; 10.1530/rep.1.01171
Average uterine weights of stimulated and non-stimulated horns. Uteri were collected from mice treated with etonogestrel (A) or levonorgestrel (B) at 0–45d and dissected into stimulated (dark bars) and non-stimulated (light bars) horns. These horns were weighed separately and average wet weight is shown in milligrams (+s.e.m.). (A) Etonogestrel; the stimulated horn was significantly increased in weight at 10d and 15d when compared with 0d (a=P<0.001) before decreasing. The non-stimulated horn remained a consistent weight until 35 and 45d when significant increase from 0d was observed (*P<0.05 and †P<0.01 respectively). (B) Levonorgestrel; the stimulated horn significantly increased in weight from 0 to 5d (a=P<0.001) before decreasing. The weight of the non-stimulated horn at 10d was significantly different to the non-stimulated horn at 0d (*=P<0.05).
Citation: Reproduction 133, 1; 10.1530/rep.1.01171
Morphology and the development of blood vessels in the stimulated horn from etonogestrel-treated mice. Cross sections of stimulated uterine horns collected at 0–45 days (d) from mice treated with etonogestrel were stained with haematoxylin and eosin. Blood vessels (BV) were identified in some sections by immunohistochemistry for CD34. The stimulated horn was highly decidualized at d0 with glands located in the basal area (arrow). By 5d, large BV had appeared (as shown by CD34 staining of endothelial cells) along with areas containing red blood cells (RBC). The process of breakdown began around 10d. BV began to lose integrity at 15d (CD34, 15d) and breakdown continued until 25d. At 20d the decidual zone had begun to break away from the myometrium (arrow) and repair of the epithelium was observed around 25d. By 35d the repair was almost complete (arrow) and by 45d the repaired epithelium was very tortuous. Scale bars; CD34 stained sections=50 μm, morphology=100 μm.
Citation: Reproduction 133, 1; 10.1530/rep.1.01171
Morphology of the non-stimulated horn from etonogestrel-treated mice. Cross sections of non-stimulated uterine horns collected from mice treated with etonogestrel were stained with haematoxylin and eosin. The endometrium showed no signs of decidualization at 0 days (d) with glands observed throughout the stroma (0d, arrow). The morphology of the non-stimulated horn remained largely unchanged from 0–25d although some evidence of stromal oedema began at 20d. At 35 and 45d, the luminal epithelium was very tortuous (arrows) similar to the epithelium after repair in the stimulated horns (Fig. 3, 45d). Scale bars=200 μm.
Citation: Reproduction 133, 1; 10.1530/rep.1.01171
Morphology of both stimulated and non-stimulated horns from levonorgestrel-treated mice. Cross sections of stimulated (A) and non-stimulated (B) uterine horns collected from mice treated with levonorgestrel were stained with haematoxylin and eosin. The stimulated horn was highly decidualized at 0 days (d) with glands observed in the basal area (arrow). Large thin-walled blood vessels (BV), similar in morphology to those seen in ENG-treated animals, were observed at 5 and 10d with evidence of epithelial repair at 15d (arrow). The non-stimulated horn (B) altered very little throughout the time course with no evidence of stromal differentiation or tissue breakdown. Similar to ENG-treated animals, at 15d the luminal epithelium appeared very tortuous (arrow) and stromal oedmea was apparent at 10 and 15d. Scale bars=50 μm.
Citation: Reproduction 133, 1; 10.1530/rep.1.01171
Oestrogen receptor staining in the stimulated horn of etonogestrel-treated mice. Cross sections of uterine horns collected from mice treated with etonogestrel were immunostained for oestrogen receptor α (ER). ER-positive cells were observed throughout the stroma in decidual tissue (A 5d, arrows indicating nuclear staining) and during the early stages of tissue breakdown. There was a reduction in staining as breakdown progressed (not shown); however, during tissue repair, staining was detected in glands (A 35d, arrow) in the basal area. Inset=negative control. Scale bars=50 μm. MMP localization and gelatinase activity in the stimulated horn of etonogestrel-treated mice. Cross sections of the stimulated horns from mice treated with etonogestrel were analysed by immunohistochemistry for the presence of MMP9 (B), MMP7 (C) and MMP3 (D) as well as in situ zymography to detect gelatinase activity (B, ISZ). Immunopositive MMP9 (B) was detected in a small number of cells in decidualized tissue (0d arrow). Numbers of positive cells increased during tissue breakdown (5d) and were maintained up to 20d (not shown). Immunopositive MMP7 (C) was detected in isolated cells in decidual and breaking down tissue (10d, arrows). Numbers were increased at sites of re-epithelialization during tissue repair (35d, arrow). Immunopositive MMP3 (D) was detected in isolated instances throughout the time course (representative 10d). Gelatinase activity was readily detectable at 5d, coincident with high numbers of immunopositive MMP9 (B, ISZ). Scale bars=100 μm.
Citation: Reproduction 133, 1; 10.1530/rep.1.01171
Leukocyte localization in stimulated and non-stimulated horns from etonogestrel-treated mice. Stimulated (A–C) and non-stimulated (D) uterine horns were collected from mice treated with etonogestrel at various time points. Cross sections were analysed by immunohistochemistry for uterine natural killer (uNK) cells (A), neutrophils (B) and macrophages (C and D). uNK cells (A) were observed in high numbers in areas of tissue breakdown after 5 and 10d (inset; negative control). The higher power image shows their relationship to BV. Isolated neutrophils (B) were seen throughout the tissue at 5d (arrows), and increased in number in the tissue undergoing breakdown at 10 and 15d (inset; negative control). In the stimulated horn (C), macrophages were observed throughout the decidualized (0d, arrow) and subsequently breaking down tissue and were associated with re-epithelialization during tissue repair (20 and 35d arrows). In the non-stimulated horn, isolated macrophages were observed throughout the time course (D, arrows). Scale bars=50 μm.
Citation: Reproduction 133, 1; 10.1530/rep.1.01171
Proportional area of tissue stained with uterine natural killer cells: stimulated horn of etonogestrel-treated mice. Cross sections of the stimulated horns from mice treated with etonogestrel were analysed using the analytical imaging system (AIS) to determine the proportional area of the stimulated horn stained by uNK cells. Proportional area (percentage+s.e.m.) was determined at each time point (n>3 sections for each time point).
Citation: Reproduction 133, 1; 10.1530/rep.1.01171
We thank Dr Devora Lieberman of Organon for the provision of the Implanon rods. This project received financial support from the National Institute of Health (Grant no. HD43192) and the UND/UNFPA/WHO/World Band Special Programme of Research, Development and Research Training in Human Reproduction, World Health Organization. LAS was funded by the NH&MRC of Australia (#388901). TJK-L was a recipient of an Australian Postgraduate Association scholarship. The authors declare that there is no conflict of interest that would prejudice the impartiality of this scientific work.
References
Brasted M, White CA, Kennedy TG & Salamonsen LA2003 Mimicking the events of menstruation in the murine uterus. Biology of Reproduction 69 1273–1280.
Britt KL, Simpson ER & Findlay JK2005 Effects of phytoestrogens on the ovarian and pituitary phenotypes of estrogen-deficient female aromatase knockout mice. Menopause 12 174–185.
Critchley HO, Bailey DA, Au CL, Affandi B & Rogers PA1993 Immunohistochemical sex steroid receptor distribution in endometrium from long-term subdermal levonorgestrel users and during the normal menstrual cycle. Human Reproduction 8 1632–1639.
Critchley HO, Wang H, Kelly RW, Gebbie AE & Glasier AF1998a Progestin receptor isoforms and prostaglandin dehydrogenase in the endometrium of women using a levonorgestrel-releasing intrauterine system. Human Reproduction 13 1210–1217.
Critchley HO, Wang H, Jones RL, Kelly RW, Drudy TA, Gebbie AE, Buckley CH, McNeilly AS & Glasier AF1998b Morphological and functional features of endometrial decidualization following long-term intrauterine levonorgestrel delivery. Human Reproduction 13 1218–1224.
Croy BA & Kassouf SA1989 Evaluation of the murine metrial gland for immunological function. Journal of Reproductive Immunology 15 51–69.
Croy BA, He H, Esadeg S, Wei Q, McCartney D, Zhang J, Borzychowski A, Ashkar AA, Black GP, Evans SS,et al.2003 Uterine natural killer cells: insights into their cellular and molecular biology from mouse modelling. Reproduction 126 149–160.
Curry TE Jr & Osteen KG2003 The matrix metalloproteinase system: changes, regulation, and impact throughout the ovarian and uterine reproductive cycle. Endocrine Reviews 24 428–465.
Galant C, Vekemans M, Lemoine P, Kokorine I, Twagirayezu P, Henriet P, Picquet C, Rigot V, Eeckhout Y, Courtoy PJ,et al.2000 Temporal and spatial association of matrix metalloproteinases with focal endometrial breakdown and bleeding upon progestin-only contraception. Journal of Clinical Endocrinology and Metabolism 85 4827–4834.
Galant C, Berliere M, Dubois D, Verougstraete JC, Charles A, Lemoine P, Kokorine I, Eeckhout Y, Courtoy PJ & Marbaix E2004 Focal expression and final activity of matrix metalloproteinases may explain irregular dysfunctional endometrial bleeding. American Journal of Pathology 165 83–94.
Girling JE, Heryanto B, Patel N & Rogers PA2004 Effect of long-term progestin treatment on endometrial vasculature in normal cycling mice. Contraception 70 343–350.
Goffin F, Munaut C, Frankenne F, Perrier D’Hauterive S, Beliard A, Fridman V, Nervo P, Colige A & Foidart JM2003 Expression pattern of metalloproteinases and tissue inhibitors of matrix-metalloproteinases in cycling human endometrium. Biology of Reproduction 69 976–984.
Goncharov N, Aso T, Cekan Z, Pachalia N & Diczfalusy E1976 Hormonal changes during the menstrual cycle of the baboon (Papio hamadryas). Acta Endocrinologica (Copenh)82 396–412.
Guimond MJ, Luross JA, Wang B, Terhorst C, Danial S & Croy BA1997 Absence of natural killer cells during murine pregnancy is associated with reproductive compromise in TgE26 mice. Biology of Reproduction 56 169–179.
Hampton AL & Salamonsen LA1994 Expression of messenger ribonucleic acid encoding matrix metalloproteinases and their tissue inhibitors is related to menstruation. Journal of Endocrinology 141 R1–R3.
Hickey M & Fraser IS2002 Surface vascularization and endometrial appearance in women with menorrhagia or using levonorgestrel contraceptive implants. Implications for the mechanisms of breakthrough bleeding. Human Reproduction 17 2428–2434.
Hickey M, Fraser I, Dwarte D & Graham S1996 Endometrial vasculature in Norplant users: preliminary results from a hysteroscopic study. Human Reproduction 11 35–44.
Hickey M, Simbar M, Young L, Markham R, Russell P & Fraser IS1999 A longitudinal study of changes in endometrial microvascular density in Norplant implant users. Contraception 59 123–129.
Hickey M, Crewe J, Goodridge JP, Witt CS, Fraser IS, Doherty D, Christiansen FT & Salamonsen LA2005 Menopausal hormone therapy and irregular endometrial bleeding: a potential role for uterine natural killer cells?. Journal of Clinical Endocrinology and Metabolism 90 5528–5535.
Hirabayashi H, Sato T, Kohno S, Tanaka M, Kobayashi S, Ohta Y & Iguchi T1999 Apoptotic cell death in artificially induced deciduoma of pseudopregnant mice. Anatomical Record 254 205–213.
Ingamells S, Campbell IG, Anthony FW & Thomas EJ1996 Endometrial progesterone receptor expression during the human menstrual cycle. Journal of Reproduction and Fertility 106 33–38.
Jeziorska M, Salamonsen LA & Woolley DE1995 Mast cell and eosinophil distribution and activation in human endometrium throughout the menstrual cycle. Biology of Reproduction 53 312–320.
Jeziorska M, Nagase H, Salamonsen LA & Woolley DE1996 Immunolocalization of the matrix metalloproteinases gelatinase B and stromelysin 1 in human endometrium throughout the menstrual cycle. Journal of Reproduction and Fertility 107 43–51.
Kaiserman-Abramof IR & Padykula HA1989 Ultrastructural epithelial zonation of the primate endometrium (rhesus monkey). American Journal of Anatomy 184 13–30.
Kaitu’u TJ, Shen J, Zhang J, Morison NB & Salamonsen LA2005 Matrix metalloproteinases in endometrial breakdown and repair: functional significance in a mouse model. Biology of Reproduction 73 672–680.
Kamat BR & Isaacson PG1987 The immunocytochemical distribution of leukocytic subpopulations in human endometrium. American Journal of Pathology 127 66–73.
King A2000 Uterine leukocytes and decidualization. Human Reproduction Update 6 28–36.
King A, Wellings V, Gardner L & Loke YW1989 Immunocytochemical characterization of the unusual large granular lymphocytes in human endometrium throughout the menstrual cycle. Human Immunology 24 195–205.
Lockwood CJ, Runic R, Wan L, Krikun G, Demopolous R & Schatz F2000 The role of tissue factor in regulating endometrial haemostasis: implications for progestin-only contraception. Human Reproduction 15 144–151.
Macpherson AM, Archer DF, Leslie S, Charnock-Jones DS, Makkink WK & Smith SK1999 The effect of etonogestrel on VEGF, oestrogen and progesterone receptor immunoreactivity and endothelial cell number in human endometrium. Human Reproduction 14 3080–3087.
Marbaix E, Kokorine I, Moulin P, Donnez J, Eeckhout Y & Courtoy PJ1996 Menstrual breakdown of human endometrium can be mimicked in vitro and is selectively and reversibly blocked by inhibitors of matrix metalloproteinases. Proceedings of the National Academy of Sciences of the United States of America 93 9120–9125.
Marbaix E, Vekemans M, Galant C, Rigot V, Lemoine P, Dubois D, Picquet C, Henriet P, Twagirayezu P, Sufi S,et al.2000 Circulating sex hormones and endometrial stromelysin-1 (matrix metalloproteinase-3) at the start of bleeding episodes in levonorgestrel-implant users. Human Reproduction 15 120–134.
Marsh MM, Butt AR, Riley SC, Rogers PA, Susil B, Affandi B, Findlay JK & Salamonsen LA1995 Immunolocalization of endothelin and neutral endopeptidase in the endometrium of users of subdermally implanted levonorgestrel (Norplant). Human Reproduction 10 2584–2589.
Marshall RJ & Jones DB1988 An immunohistochemical study of lymphoid tissue in human endometrium. International Journal of Gynecological Pathology 7 225–235.
Oliveira-Ribeiro M, Petta CA, De Angelo Andrade LA, Bahamondes L & Hidalgo MM2004 Correlation between endometrial histology, microvascular density and calibre, matrix metalloproteinase-3 and bleeding pattern in women using a levonorgestrel-releasing intrauterine system. Human Reproduction 19 1778–1784.
Parks WC, Lopez-Boado YS & Wilson CL2001 Matrilysin in epithelial repair and defense. Chest 120 36S–41S.
Phillips V, Graham CT, Manek S & McCluggage WG2003 The effects of the levonorgestrel intrauterine system (Mirena coil) on endometrial morphology. Journal of Clinical Pathology 56 305–307.
Pritts EA, Ryan IP, Mueller MD, Lebovic DI, Shifren JL, Zaloudek CJ, Korn AP, Darney PD & Taylor RN2005 Angiogenic effects of norplant contraception on endometrial histology and uterine bleeding. Journal of Clinical Endocrinology and Metabolism 90 2142–2147.
Rasweiler JJt1991 Spontaneous decidual reactions and menstruation in the black mastiff bat, Molossus ater. American Journal of Anatomy 191 1–22.
Rhoton-Vlasak A, Chegini N, Hardt N & Williams RS2005 Histological characteristics and altered expression of interleukins (IL) IL-13 and IL-15 in endometria of levonorgestrel users with different uterine bleeding patterns. Fertility and Sterility 83 659–665.
Rogers PA, Au CL & Affandi B1993 Endometrial microvascular density during the normal menstrual cycle and following exposure to long-term levonorgestrel. Human Reproduction 8 1396–1404.
Runic R, Schatz F, Krey L, Demopoulos R, Thung S, Wan L & Lockwood CJ1997 Alterations in endometrial stromal cell tissue factor protein and messenger ribonucleic acid expression in patients experiencing abnormal uterine bleeding while using Norplant-2 contraception. Journal of Clinical Endocrinology and Metabolism 82 1983–1988.
Runic R, Schatz F, Wan L, Demopoulos R, Krikun G & Lockwood CJ2000 Effects of norplant on endometrial tissue factor expression and blood vessel structure. Journal of Clinical Endocrinology and Metabolism 85 3853–3859.
Salamonsen LA & Lathbury LJ2000 Endometrial leukocytes and menstruation. Human Reproduction Update 6 16–27.
Salamonsen LA & Woolley DE1999 Menstruation: induction by matrix metalloproteinases and inflammatory cells. Journal of Reproductive Immunology 44 1–27.
Salmela MT, Pender SL, Karjalainen-Lindsberg ML, Puolakkainen P, Macdonald TT & Saarialho-Kere U2004 Collagenase-1 (MMP-1), matrilysin-1 (MMP-7), and stromelysin-2 (MMP-10) are expressed by migrating enterocytes during intestinal wound healing. Scandinavian Journal of Gastroenterology 39 1095–1104.
Schatz F, Krikun G, Caze R, Rahman M & Lockwood CJ2003 Progestin-regulated expression of tissue factor in decidual cells: implications in endometrial hemostasis, menstruation and angiogenesis. Steroids 68 849–860.
Sereepapong W, Chotnopparatpattara P, Taneepanichskul S, Markham R, Russell P & Fraser IS2004 Endometrial progesterone and estrogen receptors and bleeding disturbances in depot medroxyprogesterone acetate users. Human Reproduction 19 547–552.
Shaw ST Jr, Elsahwi SY & Moyer DL1972 Menstrual blood quantitation in the rhesus monkey: an experimental tool for improving intrauterine contraceptive devices (IUDS). Fertility and Sterility 23 257–263.
Silverberg SG, Haukkamaa M, Arko H, Nilsson CG & Luukkainen T1986 Endometrial morphology during long-term use of levonorgestrel-releasing intrauterine devices. International Journal of Gynecological Pathology 5 235–241.
Skinner JL, Riley SC, Gebbie AE, Glasier AF & Critchley HO1999 Regulation of matrix metalloproteinase-9 in endometrium during the menstrual cycle and following administration of intrauterine levonorgestrel. Human Reproduction 14 793–799.
Vincent AJ & Salamonsen LA2000 The role of matrix metalloproteinases and leukocytes in abnormal uterine bleeding associated with progestin-only contraceptives. Human Reproduction 15 135–143.
Vincent AJ, Malakooti N, Zhang J, Rogers PA, Affandi B & Salamonsen LA1999 Endometrial breakdown in women using Norplant is associated with migratory cells expressing matrix metalloproteinase-9 (gelatinase B). Human Reproduction 14 807–815.
Vincent AJ, Zhang J, Ostor A, Rogers PA, Affandi B, Kovacs G & Salamonsen LA2000 Matrix metalloproteinase-1 and -3 and mast cells are present in the endometrium of women using progestin-only contraceptives. Human Reproduction 15 123–130.
Vincent AJ, Zhang J, Ostor A, Rogers PA, Affandi B, Kovacs G & Salamonsen LA2002 Decreased tissue inhibitor of metalloproteinase in the endometrium of women using depot medroxyprogesterone acetate: a role for altered endometrial matrix metalloproteinase/tissue inhibitor of metalloproteinase balance in the pathogenesis of abnormal uterine bleeding?. Human Reproduction 17 1189–1198.
Wang C, Tanaka T, Nakamura H, Umesaki N, Hirai K, Ishiko O, Ogita S & Kaneda K2003 Granulated metrial gland cells in the murine uterus: localization, kinetics, and the functional role in angiogenesis during pregnancy. Microscopy Research and Technique 60 420–429.
Yu WH & Woessner JF Jr2000 Heparan sulfate proteoglycans as extracellular docking molecules for matrilysin (matrix metalloproteinase 7). Journal of Biological Chemistry 275 4183–4191.
Zhang J & Salamonsen LA2002 In vivo evidence for active matrix metalloproteinases in human endometrium supports their role in tissue breakdown at menstruation. Journal of Clinical Endocrinology and Metabolism 87 2346–2351.
Zhang J, Nie G, Jian W, Woolley DE & Salamonsen LA1998 Mast cell regulation of human endometrial matrix metalloproteinases: a mechanism underlying menstruation. Biology of Reproduction 59 693–703.
