Cholesterol in the plasma membrane of uterine epithelial cells: a freeze-fracture cytochemical study with digitonin

0 Departments of Human Anatomy and Zoology, University of Oxford , England 1 Department of Histology and Embryology, The University of Sydney , Australia Freeze-fracture cytochemistry with digitonin has been used to examine the cholesterol content of the plasma membrane of uterine epithelial cells during the early stages of pregnancy in the rat. Lesions caused by digitonin complexing with cholesterol were seen on both lateral and apical portions of the membrane but tight junctions and desmosomes were lesion-free. Compared with day 1 of pregnancy, lesions on the apical plasma membrane were much more extensive and some were of different morphology on day 6 - the day of blastocyst implantation. We consider mechanisms of lesion formation and interpret the results to indicate a higher content and perhaps a different organization of cholesterol in the apical plasma membrane on day 6 of pregnancy. We also suggest how this increase may occur. - INTRODUCTION The plasma membrane of uterine epithelial cells undergoes a series of dramatic changes during the oestrous cycle and early pregnancy. The apical plasma membrane bears long, thin, regular microvilli on day 1 of pregnancy but by days 5-6, the time of blastocyst implantation, these are replaced by short, irregular, flattened microvilli and various other irregular projections (Enders & Schlafke, 1967; Tachi, Tachi & Lindner, 1970; Ljungkvist, 1972; Murphy & Rogers, 1981). The lateral plasma membrane also undergoes changes during this time, with alterations in the geometry and depth of tight junctions (Murphy, Swift, Mukherjee & Rogers, 1981, 19826; Winterhager & Kuhnel, 1982). These changes are under the control of maternal ovarian hormones (Ljungkvist, 1972; Murphy & Rogers, 1981). Previous work has shown that the changes in microvillar profile of the apical plasma membrane are associated with more fundamental changes in membrane structure, as seen by freeze-fracture studies on density and organization of intramembranous particles (IMPs) (Murphy, Swift, Mukherjee & Rogers, 1979, 1982a). Alterations in carbohydrates of the cell surface coat have also been reported (Hewitt, Beer & Grinnell, 1979; Murphy & Rogers, 1981). This plasma membrane is therefore an interesting system, because changes in profile and organization can be readily observed. Nevertheless, little attention has so far been directed towards the involvement of membrane lipids in the abovementioned phenomena. We recently studied anionic phospholipids in the membrane by freeze-fracture cytochemistry (Murphy & Martin, 1984) and have also examined cholesterol in the large vesicles characteristic of progesterone-dominated epithelial cells (Murphy & Martin, unpublished). The purpose of the present study is, then, to extend our investigations on this plasma membrane into its lipid components, with special attention to cholesterol: this, especially, because the large vesicles that we have already shown to have cholesterol-rich membranes are thought to fuse with the apical plasma membrane (Parr, 1982; Murphy & Martin, unpublished). Fifteen young adult virgin female Wistar rats were used for the study: six on day 1 of pregnancy and five on day 6. Control tissue, treated as described below, was taken from these animals but two additional animals at each time were also used to provide further control tissue. The rats were maintained on tap water and rat cubes, and kept at 20C on a 14h light/10h dark cycle (lights on at 0730 h). Vaginal smears were followed for several days and females showing a pro-estrous smear (Everett, 1948) were caged overnight with males of proven fertility and were separated the following morning. Insemination was determined by the presence of sperm in the vaginal smear at the time of separation, and the day on which sperm were found was designated day 1 of pregnancy. At 1100 h on days 1 and 6 of pregnancy animals were anaesthetized with sodium pentobarbitone, the abdominal cavity was opened and uteri were excised. Tissue was processed from only two animals on any one occasion, using freshly made and identically prepared batches of digitonin each time: on some occasions day 1 and day 6 tissues were processed together and on others, day 1 or day 6 tissues or tissue from control animals were processed together. As recommended by Elias, Goerke & Friend (1978), Robinson & Karnovsky (1980) and Severs & Robenek (1983) tissue was fixed before exposure to the sterol-binding cytochemical: 2 mm transverse slices of uterus were placed in 2-0% glutaraldehyde in 0-1 M-cacodylate buffer (pH 7-4) for 60 min and to ensure that the epithelial cells were freely exposed to the cytochemical agent, the slices of uterus were cut in half at right angles to the transverse cut. Tissue was then transferred to fresh fixative as above containing 2mg/ml digitonin (Sigma; 80%) and incubated for 120 min. Control tissue was treated similarly but without digitonin in the second solution. Following three washes in fresh buffer, the tissue was transferred to 30 % glycerol in buffer. Samples of 1 mm3 were frozen in freshly thawed Freon-22 cooled with liquid N, transferred to liquid nitrogen and then fractured without etching at 115C in a Balzers Apparatus (model BAF 300). After platinum shadowing with an electron gun and carbon deposition, the replicas were separated from tissue in sodium hypochlorite, washed in water, placed on copper grids and examined in Siemens Elmiskop 102 or JEOL 100CX electron microscopes operating at 80 kV. Freeze-fracture 'runs' were repeated until successful replicas were obtained from at least two samples of uterus from each of the 15 rats. The terminology of Branton et al. (1975) is used in referring to membrane fracture faces. Quantification of membrane area occupied by lesions Although not without difficulty, owing to the variability in lesion size and morphology, we attempted to estimate the percentage of apical plasma membrane occupied by lesions as follows. A transparent grid with 1 cm X 1 cm squares drawn on it was placed over pictures enlarged to X75 000. and the numbers of such squares that fell over lesions were expressed as a percentage of the total number of squares that fell over fractured membrane. From two to five pictures were measured for each animal and at least 1000 squares were counted for each group of animals. Incubation of tissues in digitonin produced lesions in the plasma membrane of uterine epithelial cells that were similar in appearance and dimensions to those reported for other cell membranes (Elias et al. 1978; Meyer, 1981; Severs & Robenek, 1983, for a review). In general, digitonin lesions seen in this study were the typical tubular scallops, with dimensions of 50-70 nm wide and 200-400 nm long, but scalelike plates with irregular dimensions were also seen. The form and distribution of digitonin-induced lesions were different, however, depending on the day of pregnancy on which cells were examined. On day 1 of p (...truncated)