Vol. 58, No.4, October 1992

FERTILITY AND STERILITY

Printed on aci4-free paper in U.S.A.

Copyright c 1992 The American Fertility Society

Ovarian steroids modulate human monocyte tumor necrosis factor alpha messenger ribonucleic acid levels in cultured human peripheral monocytes*

Randall A. Loy, M.D.t Jill A. Loukides, Ph.D.:!: Mary L. Polan, Ph.D., M.D.§ Department of Obstetrics and Gynecology, Section of Reproductive Endocrinology, Yale University School of Medicine, New Haven, Connecticut

Objective: To determine whether tumor necrosis factor alpha (TNF-a) messenger ribonucleic acid (mRNA) levels in human peripheral monocytes are regulated by ovarian steroids. Design: Human granulosa-luteal cells and cultured, activated human peripheral monocytes were subjected to Northern blot analysis for TNF-a mRNA. Setting: Academic research laboratory. Patients: Two human female volunteers of reproductive age and in the luteal phase of the menstrual cycle. Results: Human granulosa-luteal cells produce TNF -a mRNA. Physiological levels of progesterone (P) and estradiol (E 2 ) modulate TNF -a mRNA from peripheral blood monocytes with an apparent inverse relationship between steroid concentration and TNF -a message. Conclusions: Progesterone and E2 at physiological concentrations regulate TNF-a mRNA production. The P antagonist mifepristone (RU486) and the E2 agonist/antagonist tamoxifen modulate total TNF -a mRNA levels, suggesting involvement of specific receptors. Fertil Steril 1992;58:733-9 Key Words: Tumor necrosis factor, monocytes, ovarian steroids

Monocytes in the peripheral circulation migrate into various tissues and undergo tissue-specific differentiation. The resulting tissue macrophages have morphological and sometimes functional properties that are characteristic for the tissue in which they reside (1). Such resident macrophages and monocytes are present in the human ovarian follicle and Received July 23, 1991; revised and accepted June 3, 1992.

* Supported by Senior Fellowship grant HD07081 from the National Institutes of Health, Bethesda, Maryland to M.L.P. t Reprint requests and present address: Randall A. Loy, M.D., Center for Infertility and Reproductive Medicine, 3435 Pinehurst Avenue, Orlando, Florida 32804. :j: Present address: Department of Obstetrics and Gynecology, University of Maryland, Baltimore, Maryland. § Present address: Department of Gynecology and Obstetrics, Stanford University Medical Center, Stanford, California. Vol. 58, No.4, October 1992

corpus luteum (CL) and represent important potential paracrine mediators of normal tissue homeostasis (2). Macrophage products, especially cytokines, influence ovarian function. Tumor necrosis factoralpha (TNF-a) (cachectin), a 17-kd polypeptide hormone secreted by activated macrophages, has pleiotropic actions and has emerged as an especially important mediator of inflammatory response. This cytokine has recently been immunohistochemically localized in rat, rabbit, bovine, and human ovaries (3-5) and has also been shown to alter ovarian steroidogenesis (6, 7). Macrophage products affect granulosa-luteal cells, and certain steroids are likewise known to influence monocyte secretion of cytokines (8). For example, glucocorticoids in nanomolar concentrations strongly inhibit TNF -a gene transcription and mobilization of any TNF -a mesLoy et al.

E2 and P modulate TNF-a mRNA

733

senger ribonuclic acid (mRNA) synthesized, effectively preventing any cytokine action (9). Tumor necrosis factor-alpha shares many of its diverse properties with another cytokine, interleukin -113 (IL-1j3), despite separate receptors and distinct protein structures. Work from this laboratory has demonstrated that lower concentrations of estradiol (E 2 ) and progesterone (P) (10-8 to 10- 10 M), induce maximal IL-1j3 production and concentrations > 10-7 M significantly reduce IL-1j3 activity (10). The present studies were undertaken to determine whether TNF -0: mRNA levels in human peripheral monocytes are also modulated by ovarian steroids. MATERIALS AND METHODS

Using heparinized syringes, 100 to 140 mL blood was collected from two healthy female volunteers during the luteal phase of the menstrual cycle. Immediately thereafter, to separate the whole blood, equal volumes ofthe anticoagulated whole blood and Sepracell-MN (Septratech Corp., Oklahoma City, OK) at 4°C were mixed and centrifuged at 1,500 X g for 20 minutes in a Beckman J6-B Centrifuge (Beckman Instruments, Palo Alto, CA) with swinging bucket rotor. The mononuclear cell band was aspirated and mixed with four volumes of phosphatebuffered saline (PBS) containing bovine serum albumin (BSA) (Septratech Corp.). Cell removal and washing were performed via centrifugation at 1,350 X g for 10 minutes at 4°C; the supernatant was discarded, and the pellet was resuspended in volume PBS-BSA and centrifuged at 1,300 X g for 2 minutes at 4°C. The monocyte band, just beneath the meniscus was then removed, resuspended in four volumes PBS-BSA, and centrifuged for 10 minutes at 700 X g at 4°C two times. Monocyte fractions were examined microscopically by hematology technicians and found to be more than 85 % pure via Giemsa-Wright stain. (Although Sepracell-MN is a specific centrifugation technique for the isolation of monocytes, there were also platelets present in this cell band and also possibly small numbers ofT lymphocytes and B lymphocytes). The monocytes were cultured at a density of 1 to 3 X 106 cells/mL in 2 mL Roswell Park Memorial Institute (RPMI)-1640 medium (GIBCO Laboratories, Grand Island, NY), pH 7.4, containing 25 mM (N - [2-hydroxyethyl ]piperazine- N'- [2-ethanesulfonic acid]) (Hepes) buffer, 1% penicillin-streptomycin, 2 mM L-glutamine, 0.05 mg/mL gentamicin, 0.25 mM j3-mercaptoethanol, 0.05 ng/mL

t

734

Loy et al.

E2 and P modulate TNF-a mRNA

nystatin (mycostatin), and either 5% fetal calf serum (FCS; GIBCO) or 0.1% BSA (fraction V; Calbiochem, La Jolla, CA). Although the cell culture media and steroid preparations tested negative for endotoxin, low levels of endotoxin « 10 endotoxin U /mL) were present in the prepared FCS. Progesterone or E2 (0 to 10-5 M, 20 ~L in absolute ethanol) was added to cell cultures in dose-response experiments. Steroid dose-response studies were performed in complete RPMI medium supplemented with 0.1 % BSA or 5% FCS as noted in the figure legends. Lipopolysaccharide (LPS; 10 ~g/mL; Sigma Chemical Co., St. Louis, MO) and E2 or P (Sigma Chemical Co.) at 0 to 10-5 M concentrations were added at time 0 in all dose-response studies. The remaining samples were incubated for 3 hours at 37°C in an atmosphere at 95% air-5%C0 2 • At the end of this brief incubation, there was nearly complete viability of cultured cells, via Trypan Blue exclusion sampling, at all concentrations. (The cultures were maintaineQ for 72 to 96 hours with major morphogenic changes and detachment from the substrate noted in most cultures at the time of discard). Cells were then pelleted and washed twice with 500 ~L PBS in preparation for ribonucleic acid (RNA) extraction. In the time-course studies, LPS (10 ~g/mL) was added to all incubations at time 0, and at indicated times cells were centrifuged and washed twice with 500 ~L PBS in preparation for RNA extraction. Cells from a single sample containing LPS alone were centrifuged at time 0, and the pellet was washed twice with 500 ~L PBS served as the negative control. Because each experiment required 100 to 140 mL fresh whole blood, replicate concurrent sampling was not performed. However, time-course studies were repeated a minimum of four times, and dose-response studies with 0.1 % BSA or 5% FCS were repeated a minimum of two or three times, respectively. Ribonucleic Acid Preparation

Total cellular RNA was isolated by acid guanidinium thiocyanate-phenol-chloroform extraction, according to the method of Chomczynski and Sacchi (11). Ribonucleic acid was also isolated from fresh, washed granulosa-luteal cells from pooled follicular fluid (FF) aspirates of patients who had undergone oocyte retrieval in the in vitro fertilization program and from 12-0-tetradecanoylphorbol-13-acetate (TPA, 10 ~M) stimulated HL60 cells, a human promyelocytic leukemia cell line, and A431 cells, a human ep.idermal carcinoma cell line, both known to Fertility and Sterility

produce TNF -a mRNA. The purity of total cellular RNA was assessed by the optical density (OD) 260 to 280 nm (> 1.6 in all cases). The total amount of cellular RNA isolated was denatured in 15 JLL solution containing glyoxal (1 M), dimethylsulfoxide (50%) and NaP0 4 (10 mM), pH 7.0 at 56°C for 1 hour (12). A mean of 3.15 JLg of total RNA was added per each lane of the agarose gels. Denatured samples underwent 1.0% agarose (Bio-Rad, Richmond, CA) gel electrophoresis in 0.01 M NaP0 4 buffer, pH 7.0 at 100 vol for 6 hours. After electrophoresis, RNA was transferred onto nylon membranes (Zetabind; Cuno Inc., Meriden, CT) via the capillary transfer method with 20mx SSC Buffer (3 M NaCI, 0.3 M Na3 Citrate-2H20, pH 7.0, in 1 L H 20). Filters were washed in 2x SSC, air dried, and photographed under shortwave ultraviolet light using type 667 film (Polaroid, Inc. Cambridge, MA). Visible RNA bands were present at 18s and 28s, allowing visual confirmation of total RNA transferred to the nylon membrane. Preparation of TNF-a Complementary Deoxyribonucleic Acid (eDNA) Probe

The TNF-a cDNA probe is a 650-bp Pst fragment in phTNF -5. Plasmid was prepared and purified on CsCI gradients (13) then digested with Pst I to release the 650-bp TNF -a fragment from the vector (14). The digested DNA was electrophoresed through a 1 % agarose gel. The 650-bp fragment was excised and eluted using an IBI Electroluter (New Haven, CT) according to manufacturer's specifications. The purified cDNA fragment was labeled using a random primed DNA labeling kit (BoehringerMannheim Inc., Mannheim, Germany). The nylon filters were prehybridized in a solution of 1 % BSA, 7% sodium dodecyl sulfate (SDS), a 0.5 M NaP0 4 (pH 7.0), and 1 mM ethylenediamine tetraacetic acid (EDTA) at 65°C for 10 to 12 hours in plastic bags. [32P]TNF-a cDNA probe was added to the bags at a minimum concentration of 10-6 cpm/mL and incubated for 12 to 14 hours at 65°C. Thereafter, the filters were washed three times for 10 minutes each in a solution of 0.5% BSA, 5% SDS, 40 mM NaP0 4 (pH 7.0), and 1 mM EDTA at 65°C and then washed three to four times for 10 minutes each in a solution of 1 % SDS, 40 mM NaP0 4 (pH 7.0), and 1 mM EDTA at 65°C. The filters were air dried and then exposed to Kodak XRP -1 film (Eastman Kodak, Rochester, NY) using image intensifying screens at -70°C. The films were developed, and hybridization was quantitated using a JoyceLoebel Chromoscan 3 densitometer (Malden, MA). Vol. 58, No.4, October 1992

Data are expressed as the total integral of hybridization (densitometry measurement) per total micrograms of RNA (via ~OD 260 nm measurement) loaded onto the gel.

RESULTS Production of IL-1~ and TNF-a mRNA activity is induced by the addition of endotoxin to monocytes in culture. The time course of ±LPS (10 JLg/mL) - stimulated TNF -a mRNA production is shown in Figure 1. Maximal mRNA production occurred upon incubation for 3 hours. At 1.5 hours, the TNF-a mRNA from unstimulated cells actually exceeds that of the LPS-stimulated monocytes. There was little mRNA expression after 7, 16, and 24 hours of incubation, especially in the absence of LPS. To confirm the specificity of the prepared TNF-a cDNA probe for TNF-a mRNA and to test human cells from IVF follicular aspirates (predominantly granulosa-luteal cells) for TNF-a mRNA production, positive and negative Northern analysis controls were performed. Total cellular RNA was isolated from several human cell lines known to elaborate IL-1~ and TNF-a mRNA (HL-60, a promyelocytic leukemia cell line; A431, a vulvar squamous cell carcinoma cell line; and monocytes incubated for 6 hours with LPS at 10 JLg/mL), from aspirated follicular cells and a negative control, Jurkat RNA. This nylon filter-bound total cellular RNA was hybridized with 32P-TNF-a probe and is shown in Figure 2. Jurkat cells acted as a negative control, with no TNF-a mRNA production (not shown in Fig. 2). Tumor necrosis factor-alpha is known to

TNFa. RNA! 3000 119 total RNA

2000

1000

o

1.5

3.0

INCUBATION TIME

7.25

+/- LPS

16.25

24

(Hours)

Figure 1 The time course of LPS (10 JLg/mL) - stimulated TNF -a mRNA production by human luteal peripheral monocytes cultured with 5% FeS. Quantitative densitometry of TNF-a mRNA levels in human luteal peripheral monocytes as a function of time in culture. Loy et al.

E2 and P modulate TNF-a mRNA

735

have two distinct bands on Northern analysis migrating near the 28s and 18s ribosomal RNA bands. Tumor necrosis factor-alpha mRNA activity was studied in LPS-stimulated, human luteal phase monocytes cultured in the presence of 5% FeS as a function of P (0 to 10-5 M). Quantitative densitometry graphing of the mRNA levels allows for determination of TNF -a mRNA production over the total J.Ig RNA originally loaded onto the gel (Fig. 3). The P antagonist RU486 at 10-6 M concentration induces maximal amounts of TNF -a RNA. Also, 10-9 P concentration minimally stimulated while higher P doses inhibited TNF-a mRNA production. The possibility exists that low levels of FeS endotoxin or other serum factors could explain the stimulation ofTNF-a. Although both E2 and P in the FeS were assayed and found to be