Human endometrium possesses a unique immunological environment enabling implantation of the semiallogeneic embryo. Large populations of macrophages and uterine-specific natural killer cells infiltrate the implantation site, believed to be important modulators of trophoblast invasion and decidualization. In the absence of pregnancy, there is a dramatic influx of neutrophils, eosinophils, and macrophages, likely to be critical for focal inflammatory endometrial destruction. However, little is known regarding selective recruitment of leukocyte subtypes. We employed a gene array approach to analyze the expression of 21 chemokines in endometrium. Real-time RT-PCR and immunohistochemistry was conducted to verify expression patterns and determine cellular source. Nine chemokines were highly abundant in human endometrium: monocyte chemotactic protein-3, eotaxin, fractalkine, macrophage inflammatory protein-1beta, 6Ckine, IL-8, hemofiltrate CC chemokine-1 and -4, and macrophage-derived chemokine. Chemokine mRNA was generally up-regulated during endometrial receptivity and early pregnancy, particularly of macrophage and natural killer chemoattractants. Chemokine protein was predominantly localized to epithelial glands, whereas differentiated stromal cells were a major source of chemokines after decidualization. This is the first study to use an unbiased approach to screen for endometrial chemokines, and we report the selective regulation of chemokines, corresponding to the recruitment of distinct leukocyte subpopulations required for pregnancy and menstruation.

Considerable correlative evidence suggests an important role for matrix metalloproteinases (MMPs) in menstruation, a process which occurs naturally in very few species. In this study, MMP expression was examined in a mouse model of endometrial breakdown and repair and the functional importance of MMPs determined. In the model, progesterone support was withdrawn from mice in which endometrial decidualization had been induced; 24 h later, endometrial breakdown was complete, and the entire decidual zone had been shed. Re-epithelialization had occurred by 36 h, and the endometrium had undergone extensive restoration toward a predecidualized state by 48 h. Immunoreactive MMP9 and MMP7 colocalized with leukocyte subsets, particularly neutrophils, whereas MMP13 staining was always extracellular. MMP3 and MMP7 were abundant during re-epithelialization in close proximity to newly reforming epithelium. The functional importance of MMPs in these processes was examined using two MMP inhibitors, doxycycline and batimistat. Both inhibitors effectively reduced MMP activity, as assessed by in situ zymography, but did not have significant effects on endometrial breakdown or repair. This study demonstrates that although MMPs are present in abundance during endometrial breakdown and repair in this mouse model, they are not the key mediators of these processes.

Activin and inhibin subunits are co-expressed by human endometrial epithelial and decidualised stromal cells. Activin A is a potent stimulator of decidualisation in vitro, but the mechanisms are unknown. Matrix metalloproteases (MMPs) are known to be important during decidualisation, as administration of a broad spectrum MMP inhibitor in the rat results in reduced decidualisation. Transforming Growth Factor(TGF)-βs are closely related to activins and inhibit MMP production in endometrial epithelial cells. We hypothesised that activins regulate MMP production during decidualisation and/or trophoblast invasion. Epithelial and stromal cells were isolated from human endometrium and treated for 24 h with activin, inhibin, activin/inhibin, and follistatin. Media were collected and subjected to gelatin and caesin zymography. In epithelial cells, activin A stimulated the expression of latent forms of MMPs-1, -2, -7 and -9, and increased formation of active forms of MMPs-2 and -7. Cotreatment with inhibin prevented this stimulation, whilst inhibin alone completely inhibited MMP production. Treatment with follistatin treatment reduced MMP levels. Similar regulation was seen in stromal cells for MMPs-1, -2 and -9. These data show that activin stimulates the production and activation of MMPs in both endometrial cells, and that inhibin is a potent inhibitor. It is interesting that activin is acting in an opposing manner to TGF-β, indicating that these two closely related proteins have divergent signalling pathways in endometrial cells. Further, this is the first demonstration of a role for inhibin in regulating MMPs and indeed for inhibin action in the endometrium. These findings are of potential importance in understanding regulation of MMPs in the peri-implantation endometrium. Activin is the predominant dimer produced by decidual and epithelial cells, where it may be promoting decidualisation though enhancing MMP production and activation. Furthermore, activin secretion by invasive cytotrophoblasts may stimulate focal decidual MMP production promoting their invasion during embryo implantation.

Strong correlative evidence supports a role for matrix metalloproteinases (MMPs) in the tissue breakdown at menstruation. As menstruation occurs in very few species besides women, there is a lack of suitable and easily accessible animal models available to examine the functional significance of potential key mediators of this process. A mouse model of endometrial breakdown and repair has been developed,1 which morphologically resembles that of human endometrium at menstruation. Previous studies in our laboratory showed that the expression patterns of various MMPs in the mouse model closely resembled those seen in the human.2 Administration of doxycycline, a broad spectrum MMP inhibitor, decreased gelatinase activity, but had no effect on tissue breakdown in this model. The aim of the present study was to further examine the importance of MMPs in endometrial breakdown and repair via administration of batimistat, a highly potent and specific MMP inhibitor. Batimistat was administered I.P to mice 24 h prior to the expected time of endometrial breakdown. The efficacy of batimistat within the uterus was proven using in situ zymography, which identifies MMP activity (rather than latent forms). This demonstrated that batimistat was reaching its target organ and effectively inhibiting MMP activities (both gelatinase and collagenase). Examination of gross uterine morphology revealed no apparent difference between groups, with batimistat treated uteri displaying a similar extent of tissue breakdown and repair to their control counterparts. Measurement of the breaking down area compared to total endometrial area revealed no difference between control and batimistat treatment, with the breaking down areas being 69 ±13% and 72 ± 9.8% of total endometrial cross-sectional area respectively. There was likewise no effect on endometrial repair. The results of this study together with our previous study using doxycycline, indicate that MMPs are not the key mediators of endometrial breakdown in this model. (1)Shen et al. (2004) Reprod. Fert. Dev. 16(Suppl), A265, p. 97.(2)Brasted et al. (2003) Biol. Reprod. 69, 1273.