THE EFFECT OF PROGESTERONE AND HUMAN INTERFERON u-2 ON THE RELEASE OF PGF2 AND PGE FROM EPITHELIAL CELLS OF HUMAN BROLIFERATIVE ENDOMFA'RIUM S.N. Mitchell and S.K. Smith Department of Obstetrics and Gynaecology, University of Cambridge Clinical School, The Roeie Maternity Hospital, Robinson Way, Cambridge, CB2 2SW ABSTRACT Progesterone and interferon-like trophoblastic proteins modulate prostaglandin (PG) synthesis from endometrium in early ovine and bovine pregnancy. Enriched epithelial cells were prepared from human endometrium removed in the proliferative phase of the menstrual cycle (n=8). Progesterone at a concentration of 1 PM suppressed PGE release from the cells during the first 24 hours in culture. After 48 hours in culture progesterone at a dose of 100 nM and 1 PM suppressed both the release of PGF2, and PGE from the cells and this suppression was Addition of exogenous 30 PM maintained for a further two days. arachidonic acid (AA) abolished this effect of progesterone on both PGF2, and PGE release. Interferon a-2 did not suppress the basal release of PGFza nor PGE. In the presence of progesterone, interferon a-2 attenuated the progesterone mediated suppression of PGF2, but not PGE release from endometrial cells. These findings suggest that progesterone suppresses the basal release of PGs from human endometrium, but unlike the sheep, interferon a-2 does not exert this action on human endometrium. INTRODUCTION The maintenance of early human pregnancy depends on the continued secretion of progesterone from the corpus luteum (CL) (1). In sheep and cattle, regression of the CL is prevented in early pregnancy, by conceptus secretory proteins (CSP) (2,3,4,5) which prevent the uterine luteolysin, PGFz,, from reaching the ovary. Uterine PGs are not luteolytic in women but suppression of PG synthesis is still needed to prevent menstruation and abortion (6,7). The proteins in CSPs, ovine and bovine trophoblastic proteins (oTP-1 and bTP-I), which exert this action are a group of N-linked glycosilated, interferon like peptides (8,9) which show immunological cross-reactivity and 70% cDNA sequence homology (9).


0 1992 Butterworth-Heinemann

Prostaglandins Ovine TP-1 reduces plasma levels of PGF2, and its metabolite 13,14dihydro-15-keto-PGF2o in the plasma of cyclic ewes (10) and both oTP-1 and bTP-1 attenuate PGF2, release from ovine and bovine endometrial explants in vitro (11, 12). Ovine TP-1 shows 70% sequence homology with human interferon-a (IFN-a) (13,14) and oTP-1 and human IFN-a bind to the same receptor in membrane preparations of ovine endometrium (15). Both oTP-1 and IFN-a suppress PGF2, and PGE release from separated cells of ovine endometrium maintained in vitro (11). Interferon-a is present in human chorionic villous syncytiotrophoblast as early as eight weeks gestation (16) and may thus be involved in the mechanism of suppression of PG synthesis that occurs in early human pregnancy 07,18,19). Progesterone suppresses PG synthesis from human endometrial explants (20), its secretion from the CL being maintained in early pregnancy by the release of chorionic gonadotrophin from the trophoblast. The aim of this study was to investigate the effects of progesterone and IFN-a2 on the synthesis of PGF2, and PGE from separated cells of human endometrium. MATERIALS


Materials.: Hanks balanced salt solution (HBSS), Dulbeccos modified Eagles medium (DMEM) and Hams F12 nutrient medium were obtained from Gibco Ltd., England. Penicillin, streptomycin, collagenase, DNase, insulin, oestradiol-17f3, arachidonic acid (AA), progesterone, collagen type IV, standard prostaglandin Fzu and E2, sodium acetate, methoxyamine hydrochloride, phosphate buffer tablets and glycerol gelatin mountant were obtained from Sigma Chemicals Ltd., Poole, Dorset, England. Fungizone, Hepes buffer, L-glutamine and gentamicin were bought from ILN & Flow Laboratories, High Wycombe, Bucks., England. Ultraser was obtained from Life Science, England and Percoll bought from Pharmacia Ltd., Milton Keynes, England. Human interferon a-2 was bought from Hoffman La Roche, Welwyn Garden City, England. Radioactive prostaglandin Fza and E2 were bought from Amersham International, Bucks., England, and the prostaglandin antiserum was a present from Dr. R. W. Kelly, Edinburgh. Ultima Gold scintillant was obtained from Canberra Packard Ltd., Pangbourne, England. Biotin anti-mouse antibody and avidin conjugated peroxidase were obtained from Dakopatts Ltd., High Wycombe, England. Anti-cytokeratin antibody, PKK-1 was obtained from Labsystems Ltd., England. Falcon flasks (75 ml) and 96 well microtitre plates were bought from Phillip Harris Scientific Ltd., London, England. Endometrial preparation: Human endometrium was obtained from 8 women undergoing either hysterectomy or dilatation and curettage for benign gynaecological conditions. Ethical approval was obtained from the


Ethical Committee of the Cambridge Health Authority. Endometrial strips removed at operation were placed immediately in 5 ml ice cold HBSS containing fungizone (5 pg/ml), penicillin (50 U/ml) and streptomycin (50 ug/ml) and taken to the laboratory. Endometrial fragments were cut into approximately 1 mm3 pieces, placed in 5 ml HBSS containing Hepes buffer (20 nM), collagenase (1 mg/ml), DNase (0.02 mg/ml), penicillin (50 U/ml), streptomycin (50 pg/ml) and fungizone (5 pg/ml) and shaken in a water bath at 37oC for about 20 minutes. An enriched fraction of endometrial glands was prepared as described by Satyaswaroop et al., (21). Briefly, partially digested endometrial fragments were filtered through a 45 pm nylon filter. The glandular elements were retained on the filter and collected by back flushing. This latter fraction was further digested for about 40 to 60 minutes until a single cell preparation was achieved. Red blood cells were removed from this preparation by placing the digested cell suspension over a (60%) Percoll gradient followed by centrifugation at 1500 r.p.m. for 15 minutes. The epithelial cells appeared as a white buffy coat above the Percoll and were aspirated from the interface and washed with 20 ml of HBSS. 10 ~1 of the cell suspension was removed and the cells counted in a haemocytometer. Culture corzditiolzs : The culture medium contained equal volumes of DMEM and Hams F12 nutrient medium, plus serum substitute, Ultraser (2%), L-glutamine (10 mM), Hepes buffer (20 mM), MEM non essential amino acids (100 mM), insulin (10 pg/ml), penicillin(50 U/ml), streptomycin (50 pg/ml), gentamicin (50 ug/ml) and oestradiol-17G (10 nM). Cells were plated out in 200 ~1 in 96 well collagen coated microtitre plates at a final concentration of 106 cells/ml. Cultures were performed in triplicate in 5% C@/95% air at 37OC for 96 hours. Supernatant was removed from each well every 24 hours and non-attached cells removed by centrifugation at 2500 r.p.m. for 10 minutes and replaced in fresh medium in their original wells. The supernatant was mixed with an equal volume of methoxyamine solution (40mg sodium acetate and 5 g methoxyamine hydrochloride in 50 ul of 10% ethanol made up to a 500 ml solution with distilled water, pH 5.6-5.8). Cells obtained from each patient were incubated with three different doses of progesterone (10 and 100 nM and 1 PM) and three doses of IFN CI2 (5, 50, 500 U/ml) both with and without AA (30 nM). In addition, cells were incubated with progesterone (1 PM) and IFN a-2 (500 U/ml) in the presence or absence of AA (30 nM). Control wells consisted of cells incubated without the addition of progesterone or IFN a-2 in the presence or absence of AA (30 nM).



Cell characterisation: In order to determine the relative amounts of epithelial cells in these enriched glandular preparations, cells were stained for their cytokeratin antigenicity. Glandular cells were prepared as described above and after culture for 24 hour, were dispersed with trypsin (0.25%), cytospun onto glass slides and fixed in acetone (60%). Cells were rehydrated with phosphate buffered saline (PBS) for 5 minutes and incubated for 30 minutes at room temperature with the anti-cytokeratin antibody, PKK-1 used at a dilution of l/200. The cells were washed twice with PBS and incubated for 30 min. with the second antibody, a biotin anti-mouse antibody used at a dilution of l/100 in PBS. The cells were washed and incubated for a further 30 minutes with an avidin conjugated peroxidase. Finally, the cells were stained with diaminobenzidine (0.05 g in 100 ml PBS + 100 ~1 of 50% H202) for five minutes. Cells were washed with PBS twice and counterstained with Cerazzi’s haematoxylin. Inspection of 400 cells (magnification x 400) revealed brown staining in 94 % of cells of the presumed enriched glandular preparation. Cell viability: At the end of the four days of culture the cells were assessed by staining with Nigrosin blue dye and approximately 80% of the cells were capable of extruding the dye. Assessmellt of IFN a-2 activity: IFNa-2 (dose 3 MU) was reconstituted in 3 ml of sterile water, diluted to 2x105 U/ml, aliquoted into 500 ~1 and stored at -800C until used. The predictive loss of activity at -200C and -7OOC has been estimated to be minimal (22). The viability of IFNa-2 after 9 months in storage was assessed by determining its anti-viral activity against an International Reference Standard of interferon (Professor A. P. Flint, Royal Zoological Society of London). This assay assessed the effect of incubation of Semliki Forest Virus with MVBK cells which had been incubated for the preceeding 24 hours with IFN a-2. These studies confirmed that the activity of the IFN a-2 used in these experiments, was between 50 to 75 % of the activity of the International Reference Standard. Measurement of PGs: The concentrations of PGF2o and PGE in culture medium were measured by radioimmunoassay WA). The assay and significant cross-reactivities of the antibodies raised against PGF2, and methyloximated PGE were described previously (23). Briefly, 30 ~1 of test sample were mixed with 70 ~1 PBS, 100 ~1 of appropriate antiserum and 100 ~1 of radiolabelled PG, and incubated at 4OC overnight. Bound and unbound antibody was separated with 0.8 ml of polyethylene glycol, the bound antibody being precipitated and removed by centrifugation at 3000 r.p.m. for 10 minutes at 4OC. The precipitate was mixed with scintillant and counted for 1 minute in a Canberra Packard 1500 Liquid Scintillation Analyser. The sensitivity of the assay, being the lowest amount of PG



detectable from the blank was 2 pg/tube. The intra-assay coefficient of variation for PGF2, was 12.6% and for PGE was 6.4 %. The inter-assay coefficients of variation for PGF2, and PGE were 13.8 % and 13.6 % respectively. The cross reactions between PGEl and PGE2 is 53% and between PGEl and PGE3 is 31% (23), in this text PGE therefore refers to the collective PGEs. Statisfical analysis: There was a wide variability with respect to the concentration of PGs released by endometrium obtained from different individuals. Values given in the text and tables are the cumulative median concentrations + the standard error of the means (SEM) and the range in rig/l@@ cells/number of days in culture. The effect of treatments on the release of PGs is expressed as the mean @EM) percentage change of the cumulative amount of PG in the medium, where the percentage change equals the cumulative amount of PG released by the endometrium after treatment xl00 divided by the cumulative concentration of PG found in the culture media of cells maintained without the addition of progesterone or IFN a-2.

Table 1. Cumulative median release + range of PGF2a and PGE release from human proliferative endometrium.














0.1 - 5.7ng/ml

In Ihc prcscnce

of arachidonic





- 2.9ng/ml

1.9 0.7 - 9.lng/ml

0.3 - 3.7nglmI


0.3 - 27.2ngIml

0.4 - 4.0nglml


0.5 - 33.0nglml

0.G - 34Snglml

: 2

1 0.4

PGF 2a

0.2 - 3.4nglml

- 2.nglml

1.0 0.2 - 4.0nglml

3.3 1.9 - 32.0nglml

3 1.2 0.3 -

4.3 2.7 - 44.nglml

4 1.G 0.5 - 6.3nglml

5.4 3.3 - 47.1 nglml

As the data did not demonstrate a normal distribution, the non parametric, Wilcoxon matched pairs rank sum test was used on the raw


462 data to compare differences into culture medium under

in the median concentrations different treatments.

of PGs released

RESULTS The basal amounts of PGF2, and PGE released into the culture medium are shown in Table 1. More PGE was released than PGF2, over all four days in culture and AA (30 PM) increased the synthesis of both PGs by the cells. In the absence of AA, (Figure la,) progesterone, at 1 uM significantly Figure 1. The effect of progesterone on the cumulative from human proliferative endometrium. Cells were absence (a) and presence (b) of AA (30nM).



RAY2 250,



release of PGF2, cultured in the

Lxl3 ice .




w 250






The values given are the mean percentage change of the cumulative amounts of PG (k SEM) released into the culture media. (* p < 0.03, ** p < 0.02.) suppressed the release of PGF2, over the first 48 hours in culture (t =2.2, p=< 0.03). The reduction of PGF2, synthesis was maintained over the third and fourth day in culture with doses of 100 nM and 1 PM, such that



at the end of the fourth day PGFza release was reduced by 38% (t=2.24, p=

The effect of progesterone and human interferon alpha-2 on the release of PGF2 alpha and PGE from epithelial cells of human proliferative endometrium.

Progesterone and interferon-like trophoblastic proteins modulate prostaglandin (PG) synthesis from endometrium in early ovine and bovine pregnancy. En...
844KB Sizes 0 Downloads 0 Views