77

Molecular and Cellular Endocrinology, 81 (1991) 77-80 0 1991 Elsevier Scientific Publishers Ireland, Ltd. 0303-7207/91/$03.50

MOLCEL

02604

Analyses of ovine corpora lutea for tumor necrosis factor mRNA and bioactivity during prostaglandin-induced luteolysis I. Ji, R.G. Slaughter,

J.A. Ellis, T.H. Ji and W.J. Murdoch

Departments of Animal Science, Molecular Biology and Veterinary Science, Unioersity of Wyoming, Laramie, WY82071, (Received

Key words: Corpora

lutea; Tumor

necrosis

factor;

22 April 1991; accepted

Prostaglandin;

Luteolysis;

lJ.S.A

19 June 1991)

(Sheep)

Summary

It has been suggested that tumor necrosis factor (TNF) participates in the mechanism of regression of the corpus luteum. We measured luteal expression of TNFa mRNA and biological activity during prostaglandin-induced luteolysis in sheep. Initiation of functional luteolysis was marked by a sharp decline in concentrations of progesterone in luteal tissue beginning 4 h after administration of luteolysin. Structural regression of corpora lutea was manifested by a reduction in glandular weight at 16 h. A luteal cytotoxic factor with TNFcr-like bioactivity was isolated after the decrease in tissue progesterone had occurred, but before evidence of luteal resorption. We were unable to detect temporal alterations in TNFa mRNA in luteal samples by classical Northern blot or in situ hybridization analyses. These results imply that luteal TNFa is derived primarily as a preformed entity from an extraovarian source, such as infiltrating leukocytes. These results raise the possibility that this cytokine might not be involved in the early stages of luteal regression in the ewe, yet could play a secondary role, perhaps in the subsequent opsonization and removal of degenerating cells.

Introduction

Tumor necrosis factor (TNF) is a cytotoxic agent secreted by activated macrophages. It appears to play a role in the relatively uncommon occurrence of spontaneous tumor regression (Olds, 1985). In many respects the corpus luteum resembles a solid tumor that undergoes cyclic degeneration. Leukocytes, to include mononuclear cells, infiltrate the involuting corpus luteum

Address for correspondence: W.J. Murdoch, of Animal Science, P.O. Box 3684, University Laramie, WY 82071, U.S.A.

Department of Wyoming,

of several mammals (Paavola, 1979; Murdoch, 1987; Bagavandoss et al., 1988, 1990). Corpora lutea of rabbits produced a TNFa-like substance in response to lipopolysaccharide (Bagavandoss et al., 1988, 1990). A hallmark of tumor regression induced by TNF entails endothelial cell necrosis (Olds, 1985). Vascular damage and tissue ischemia are amongst the earliest morphological signs of luteal regression (Keyes and Wiltbank, 1988). Thus, a cause-and-effect relationship could exist between luteal accumulation of TNF and luteolysis. The objectives of the following investigations were to monitor luteal alterations in TNFa-like bioactivity and expression of mRNA encoding for

78

TNFcv during the time-course of luteolysis induced by prostaglandin (PG) F,, in the sheep. materials and methods Western-range ewes were observed daily for estrous behavior in the presence of vasectomized rams. The first day of estrus was considered day 0 of the estrous cycle. General anesthesia was induced on day 10 by intravenous (jugular) injection of sodium thiopental. Reproductive organs were exteriorized through a midventral abdominal incision using aseptic technique. An ovary containing a corpus luteum was surgically removed at 0, 2, 4, 8 and 16 h (6-10 animals/ group) following intramuscular injection of 10 mg dinoprost tromethamine (Upjohn). Ewes in the O-h group served as controls and were not given PGF,,. Corpora lutea were isolated from ovaries by dissection and weighed. Segments of tissue from each corpus luteum were processed for progesterone radioimmunoassay (Rahmanian and Murdoch, 19871, TNFa! bioassay, and mRNA hybridization analyses (Northern and in situ) as described beiow. Statistical comparisons among weights of corpora lutea, tissue concentrations of progesterone and TNFcu-like bioactivity due to time of collection were assessed by one-way analysis of variance and Dunnett’s test. Tumor necrosis factor bioactivity was measured in supernatants of luteal samples homogenized in RPMI-1640 culture media and filtered through a 0.2 pm membrane. Extracts were assayed calorimetrically for cytotoxicity toward the Wehi 164, clone 13 mouse fibrosarcoma cell line as outlined in detail elsewhere (Hansen et al., 1989). Briefly, 50,000 cells in 0.1 ml were cultured with 0.1 ml volumes of extracts or log dilutions of standard recombinant bovine TNFcv and 10% normal rabbit serum or 10% mono-specific antiTNFcv rabbit serum. The TNF antiserum was prepared in our laboratory using rabbit TNFa as the immunogen. This antiserum, when included in the Wehi assay, blocked the cytoiytic activity of rabbit TNFa and TNFa activity in supernatants from lipopolysaccharide-stimulated ovine macrophages. After a 20 h incubation at 37 o C, 0.025 ml of 5 mg tetrazolium salt per ml of phosphate-

buffered saline was added. Following a 2 h incubation at 37 o C, 0.1 ml of the supernatant was removed and 0.1 ml of extraction buffer containing sodium dodecyl sulfate and ~,~-dimethyl formamide added. Color development was quantified at 595 nm after an overnight incubation. Messenger RNA was isolated using a FASTTrack kit (Invitrogen) according to the manufacturer’s instructions. Tissues were cut into 3-5 mm pieces and ground in liquid nitrogen in a mortar with a pestle. Powdered samples were transferred into sterile centrifuge tubes, mixed with lysis buffer, and homogenized with an Omni 1000 microhomogenizer for 30 s. After incubation in a shaking water bath at 45” C for 1 h, lysate was mixed with NaCI to a concentration of 0.5 M and rocked for 30 min. Oligo dT cellulose was pelleted by centrifugation, thoroughly washed with binding buffer, packed into a column, and eluted with elution buffer. Messenger RNA was prccipitated with sodium acetate and ethanol and quantified at 260/280 nm. Preparations showed numerous fine bands on 1% agarose gel and trace of 28s and 18s ribosomal RNA bands, indicating no visible degradation of RNA. An equal amount of isolated mRNAs (5 pg/lane) were electrophoresed on 1% agarose gel and vacuumblotted to Hybond-N nylon membrane (Amersham). Membranes were hybridized with “Priboprobes which were prepared from human TNFcr and drosophila p-actin cDNAs cloned in pGEM3Z (Promega~. The TNFcv clone was prepared by inserting the Hind111 cut (600 bpj of the TNFol coding sequence from pAW711 clone (American Type Culture Collection) to the Hind111 site in pGEM3Z. The actin coding sequence was cut with Hind111 and the resulting 1.9 kb fragment was introduced into the Hind111 site of the polylinker in pGEM3Z. The Northern hybridization technique was sensitive to < 1 pg mRNA. Pieces of luteal tissue for in situ hybridizations were washed in 0.1 M phosphate-buffered saline, dehydrated in a graded series of ethanol, cieared in xylene, infiltrated with paraffin, and embedded in paraffin blocks using standard techniques. Tissue sections (6 pm) were floated on distilled water (50 o C) containing 0.02% diethylpyrocarbonate (DEP-H20), transferred onto subbed mi-

79

slides, and air-dried. Sections were deparaffinized in xylene, rehydrated to DEP-H,Q, post-fixed for 30 min in 4% paraformaldehyde, and washed in DEP-H,O (4 X 5 min). Slides were placed in prehybridization buffer 60% deionized forma~de; 0.1% denatured salmon sperm DNA; 0.05% yeast RNA; 10 mM Tris (pH 7.4); 1 mM EDTA; 0.3 M NaC1/0.03 M sodium citrate (2 X SSC); 0.02% polyvinylpyrolidone; 0.02% bovine serum albumin; 0.02% Ficoll 400; 10% dextran sulfate; 10 mM dithiothreito1) for 1 h at 25 “C and drained. Hybridization buffers (0.02 ml prehybridization buffer containing either ~00,000 cpm 35S-labelled TNFa, actin (positive control) or pGEM3Z plasmid (negative control) cRNA probes) were aliquoted over sections. The pGEM32 construct containing 600 bp human TNFa cDNA was cut with Ha&f and a 210 bp antisense riboprobe for TNFrv was prepared using SPB polymerase. The actin construct was cut with HaeIII and a 409 bp antisense riboprobe for actin was prepared using T? polymerase. A 72 bp antisense riboprobe containing the polyljnker of pGEM3Z was prepared using SP6 polymerase after cutting the blank vector with EcoRI. Slides were incubated overnight at 37 o C in a humidified chamber, rinsed in 2 X SSC (12 X 15 min) at 37 o C, and air-dried. Slides were dipped (total darkness) in Kodak NTB-2 photographic emulsion diluted 1: 1 with sterile water and ptaced in a vertical position on absorbent paper. The emulsion was allowed to dry for 2 h at 25 o C. Sections were exposed for 5 days at 4 o C in a sealed box containing desiccant. Autoradiograms were developed in Kodak D-19 for 3 min at 16°C Slides were rinsed in distilled water, fixed for 3 min (Kodak rapid fix), washed under running water for 30 min, and air-dried. Sections were stained for 10 min in Mayer hematoxylin, washed (3 X 10 min), air-dried, and mounted under Permount (Fischer Scientific). Quantifications of mRNA were made with the aid of a microcomputer image analysis system (Ladd Research Industries), Silver grains were counted directly from images (eight randomly selected areas per specimen) projected f x 1000; constant bright-field settings) onto a sensor pad coupled to a data analyzer. Average values for each animal were computed. Differences in group means due to time of tissue collec-

croscope

tion were contrasted ance.

by one-way analysis of vari-

Results and discussion Initiation of functional futeolysis was denoted by a precipitous decline in concentrations of progesterone in luteal tissue (per mg wet weight) beginning 4 h after injection of PGF,,,. Structural regression of corpora lutea was indicated on a gross basis by a reduction in luteal weight at 16 h {Fig. 1). These observations are consistent with the concept that morphological evidence of prostaglandin-induced luteal regression in sheep is prefaced by an abrupt decline in circulatory progesterone. This provides a convenient model by which to closely study temporal associations in luteolysis, A cytotoxic factor, with bioactivity analogous to that of TNFcy, was extracted in relatively large amounts from corpora lutea after the primary drop in tissue progesterone had occurred, but before evidence of luteal resorption (Fig. 1). This activity was blocked by TNF antiserum. Alterations in TNFcr bioactivity were localized to the luteal compartment of the ovary, and did not take place, for example, within ovarian interstitium idata not shown). Somewhat surprising&, TNFrw mRNA was not detected in luteal homogenates by Northern blot analysis. This was not due to failure of the cRNA

Fig. 1. Ewes were treated with PGF,, and ovaries containing a corpus luteum were surgically removed at 0, 2,4, 8 and 15 h later. Corpora iutea were excised from the ovaries and weighed. Tissues were analyzed for concentrations of progesterone, TNFrv bioactivity and mRNA expression (in situ hybridization). Means + standard errors are plotted. Asterisks denote differences (P < 0.05) from 0 h.

80

pql

Corpus Luteum rl25

1.1 kb-+

I

0

0

2

4

816 I

Hours

after PGF2a

Treatment

Fig. 2. Messenger RNA was isolated from corpora lutea of ewes collected at various times as described in the legend to Fig. 1. An equal amount of mRNA from each sample was electrophoresed on an agarose gel, blotted onto a nylon membrane, and hybridized with ‘*P-riboprobe for TNFu mRNA. As positive controls, equal amounts of mRNA isolated from liver or corpus luteum of an untreated ewe were hybridized with TNFu or actin probes, respectively.

probe to detect heterologous TNF, because liver tissues hybridized intensely (Fig. 2). Furthermore, this observation was corroborated by results of the in situ hybridization experiment. No significant differences (P > 0.05) in luteal expression of mRNA were detected due to time of sample collection (Fig. 1); mRNA was localized to a small proportion of resident mononuclear cells at 16 h. Luteal and liver sections hybridized intensely with actin and TNFa cRNA probes, respectively ( > 350 grains/field; X 1000; IZ= 6). Hybridization of luteal samples with the plasmid cRNA was negligible ( < 20 grains/field; X 1000; iz = 8). It appears that TNFa mRNA represents only a small quantity of the luteal mRNA pool. These results indicate that the TNFa gene is probably not inherently expressed by the involuting ovine corpus luteum. A TNFa-like molecule has been reported to originate from granulosa cells of rat, cow and human follicles (Roby and Terranova, 1989; Zolti et al., 1990). Conceivably preformed TNF is imported into the corpus luteum by mononuclear leukocytes. Polymorphonu-

clear leukocytes (predominantly eosinophils) invaded the corpus luteum of the sheep within 2 h of injection of PGF,,; mononuclear cells were not observed in luteal tissues until 16 h (Murdoch, 1987). The mode by which TNFa acts to regulate ovarian steroidogenesis remains uncertain. There are conflicting observations as to whether it inhibits or stimulates production of progestins (Roby and Terranova, 1988; Adashi, 1990). Regardless, the abrupt drop in luteal concentrations of progesterone before a detected rise in TNFa bioactivity indicates that it is unlikely that this cytokine is involved in the onset of functional luteolysis in sheep. It is still possible that TNF is of some relevance in morphological demise of the ovine corpus luteum. Acknowledgements

The technical assistance of G. Johnson and E. Van Kirk is appreciated. Wehi cells and bovine TNFa were kindly provided by Drs. M. Lairmore and M. Campos, respectively. This work was supported in part by NIH HD-18702. References Adashi, E.Y. (1990) Endocr. Rev. 11, 454-464. Bagavandoss, P., Kunkel, S.L., Wiggins, R.C. and Keyes, P.L. (1988) Endocrinology 122, 1185-1187. Bagavandoss, P., Wiggins, R.C., Kunkel, S.L., Remick, D.G. and Keyes, P.L. (1990) Biol. Reprod. 42, 367-376. Keyes, P.L. and Wiltbank, M.C. (1988) Annu. Rev. Physiol. 50, 465-482. Murdoch, W.J. (1987) Am. J. Reprod. Immunol. Microbial. 15, 52-56. Olds, L.J. (1985) Science 230, 630-632. Paavola, L.G. (1979) Adv. Exp. Biol. Med. 112, 527-533. Rahmanian, M.S. and Murdoch, W.J. (1987) J. Anim. Sci. 64, 648-655. Roby, K.S. and Terranova, P.F. (1988) Endocrinology 123, 2952-2954. Roby, K.S. and Terranova, P.F. (1989) in Growth Factors and the Ovary (Hirshfield, A.N., ed.), pp. 273-278, Plenum Press, New York. Zolti, M., Meirom, R., Shemesh, M., Wollach, D., Mashiach, S., Shore, L. and Rafael, Z.B. (1990) FEBS Lett. 261, 253-255.