Implantation of the embryo in the uterus is a highly regulated process in which the uterus must undergo specific changes to become ‘receptive’ to the implanting embryo. The final positioning of the embryo, that is implantation into the antimesometrial side of the uterine lumen with the inner cell mass facing the mesometrial side is a well-established phenomenon. However, mechanisms controlling this phenomenon are currently unknown. One hypothesis involves a differential oxygen tension within the uterine lumen. However, previous results from our laboratory indicate that an increase in aquaporin 1 (AQP1) in the mesometrial myometrium compared to the antimesometrial muscle may contribute to the antimesometrial positioning of the embryo within the uterine lumen.

At the time of implantation there is a dramatic reduction in uterine luminal fluid volume as the uterine lumen closes down and there is close apposition between the uterine epithelium and trophoblast. This fluid may be removed from the lumen by crossing the uterine epithelial barrier by either paracellular or transcellular pathways. Tight junctions regulate the paracellular barrier and freeze-fracture studies have demonstrated that at the time of implantation there is an increase in the depth of the tight junctions and more branching of the tight junctional strands. Hence, it is thought that the tight junctions play a role in modulating the contents of the luminal fluid. This regulation mechanism of fluid transport via the paracellular route between uterine epithelial cells has been studied, but to date the mechanisms of transcellular fluid transport through these cells remain unclear. Hence, the presence of aquaporins in uterine epithelial cells is of interest because of the role they play in the transcellular route of fluid transport.

Water transport across cells may occur by diffusion alone or through specialised transporters, called aquaporins, which are intrinsic membrane proteins and exhibit a greater capacity of water movement when compared with simple diffusion alone. The direction of water transport by aquaporins is dependent on osmotic gradients and to date every fluid transporting epithelium investigated contains one or more of the 11 aquaporin isoforms. Aquaporins consist of two tandem repeats of three-membrane spanning -helices and are classified according to permeability characteristics. AQP 1, 4 and 5 are all members of the classical aquaporin family and are water selective.

AQP1, a 28 kDa protein initially discovered in red blood cells, has been found to be widely expressed in a variety of tissues. It has been localised in the kidney, capillary endothelium, male reproductive tract, vascular smooth muscle cells and uterine myometrium. AQP4 is also widely distributed in kidney, skeletal muscle, respiratory tract and brain astroglial cells where it is considered to be involved in regulation of the blood–brain barrier. AQP5 has mainly been localised in apical plasma membranes of various secretory glands including salivary and lacrimal glands as well as submucosal glands of the respiratory tract. In addition, AQP5 is also present in the upper gastrointestinal tract and on the apical surface of type I pneumocytes.

Studies on the localisation of aquaporins in the female reproductive tract are just beginning with most of the work being carried out in relation to AQP1. Human and rat uteri were found to contain AQP1 mRNA. In situ hybridisation studies of the pregnant uterus revealed faint expression in endometrial stroma with intense expression in the myometrium, both of which increased at the time of implantation. Immunohistochemical studies have demonstrated the absence of AQP1 in nonpregnant rat uterus whereas very recently AQP1 was found to be present in the inner circular myometrial layer and its expression increased at the time of implantation. Furthermore, it was found that AQP1 was much more concentrated in mesometrial myometrium as compared to antimesometrial myometrium.

Studies using reverse transcriptase PCR demonstrated the presence of both AQP4 and AQP5 in pregnant mouse uterus. In situ hybridisation studies showed AQP4 mRNA to be present in uterine epithelium in early pregnancy with the signal diminishing prior to implantation. AQP5 protein and mRNA were both present in glandular epithelium at the time of implantation in the mouse.

The aim of the current study was the detection of distributional changes in AQP1, 4 and 5 in rat uterine epithelial cells during early stages of pregnancy using light and electron microscopy.

Twenty-five female adult virgin Wistar rats were used in this study. Afternoon vaginal smears indicated which females were in proestrus and these animals were then placed in a cage overnight with a male of proven fertility. A morning vaginal smear indicating the presence of spermatozoa was used as confirmation of mating and this time point was designated as day 1 of pregnancy.

Five animals in each stage of early pregnancy (days 1, 3, 6, 7 and 9) were killed with an intraperitoneal injection of sodium pentobarbitone (0.6 ml; Nembutal; Merial Australia, Parramatta, Australia). Uterine horns were excised and randomly assigned for either light or electron microscope studies.

Immediately after excision, uterine tissue was placed in ice-cold 0.1 M phosphate buffer (pH 7.4) before being cut into 5-mm pieces and covered with OCT (Tissue Tek, Torrance, CA, USA). Blocks were immersed briefly in supercooled isopentane (BDH Laboratory Supplies, Poole, England) and stored in liquid nitrogen until sectioning.

Two blocks per animal were sectioned using a CM3050 cryostat (Leica, Heerbrugg, Switzerland). At least eight randomly selected sections (8-?m thick) per animal were placed onto gelatin-coated slides and allowed to air-dry briefly. Sections were then fixed in acetone at ?20°C for a minimum of 1 h. The remaining steps were all carried out at room temperature and all antibodies were diluted with phosphate buffered saline (PBS) containing 1% bovine serum albumin (Sigma, St Louis, MO, USA). All sections were placed in a humid chamber, incubated in a pre-blocking solution of PBS containing 1% bovine serum albumin prior to incubation in the primary antibody solution for at least 3 h.

Primary antibodies used in this study were directed against unique sequences of AQP1, AQP4 or AQP5 (Alpha Diagnostics International, San Antonio, TX, USA) and were used at concentrations of 2.5, 20 and 3.3 ?g/ml, respectively. After washing in PBS, sections were incubated in FITC-conjugated goat-anti-rabbit secondary antibodies (Zymed, San Francisco, CA, USA) at a concentration of 7.5 ?g/ml for 30 min and kept in the dark to prevent quenching of the fluorescent signal. Slides were then washed in PBS, mounted with Vectashield (Vector, Burlingame, CA, USA), coverslipped and visualised using a Diaplan microscope (Leica). Digital images were taken using a Leica DC 200 camera and micrographs were produced using Photoshop software (Adobe Systems, San Jose, CA, USA). Final magnifications were calculated using a stage micrometer.