Prostaglandin-E2 regulation of tumor necrosis factor receptor release in human monocytic THP-1 cells.

effector cells in response to inﬂammation, immunologic reactions, and physiological stress in acute tissue in- Recent in vitro studies indicate that tumor necrosis jury or sepsis. TNF is cytotoxic for tumor cells and factor (TNF) production in human monocytic THP-1 appears to be involved in the pathogenesis of endotoxic cells is suppressed by action of arachidonic acid me-shock as well (1). Infusion of human TNF into rats tabolite prostaglandin–E 2 (PGE 2 ). PGE 2 stimulation of human monocytic cell line THP-1 demonstrates that elicits a condition known as cachexia, involving hypo-PGE 2 not only regulates TNF activity at production tension and hemorrhagic necrosis of various vital or-levels, but does so through the release of two soluble gans (1). TNF is also associated with activation and TNF receptors (BP-55, BP-75) as well. PGE 2 can thus growth of immune cells including macrophages, B- and exert a regulatory effect on TNF biologic activity by T-lymphocytes, endothelial cells, neutrophils, and ﬁ- interfering with its ability to reach cell membrane re-broblasts (2–5). ceptors. THP-1 cells were activated with PGE 2 for ei-Like many cytokines, the speciﬁc binding of TNF to ther 2- or 6-hr time periods, and the supernatants sub-its receptor molecules is the essential ﬁrst step in prop- sequently tested by ELISA to quantitate the levels of agating the biological activities of TNF. The TNF binding activity of soluble cytokine receptors has sug- release of soluble TNF receptor shedding. TNF man monocytic cell line THP-1. THP-1 cells found with NEM alone (85% but additionally, it provides evidence of a new pathway of regulating the wide range of immune functions in 88%

Like many cytokines, the specific binding of TNF to ther 2-or 6-hr time periods, and the supernatants subits receptor molecules is the essential first step in propsequently tested by ELISA to quantitate the levels of agating the biological activities of TNF. The TNF lisoluble receptor released. In addition, we examined gand-receptor interaction activates signal transducmechanisms of receptor shedding by investigating the tion pathways, leading to the expression of TNF's wide rate of membrane internalization and the role of serrange of functions. The existence of two distinct TNF ine proteases. PGE 2 -stimulated THP-1 cells showed receptors (TNF-R) has been determined with molecular soluble 55-and 75-kDa TNF receptor release levels weights of 55 and 75 kDa (6)(7)(8). activity (13), yet the precise kinetics of the regulation remains unclear. The inhibition expressed through TNF ligand-receptor binding in vivo leads to its classi-INTRODUCTION fication as TNF blocking protein (14).
The host antitumor response, in which TNF plays an Tumor necrosis factor (TNF) 2 is a polypeptide secreted by activated macrophages and other immune integral role, has been found to elicit a series of reac- tions involving the endocrine system. The theory of 1 10 6 cells per 1 ml, per well. Cells were activated with various concentrations of prostaglandin E 2 (Sigma neuroendocrine influence over immune functions has developed from evidence of nerve endings in organs Chemical Co., St. Louis, MO) and incubated in a CO 2 incubator at 37ЊC for specified time periods. Upon com-and tissues of the immune system such as the thymus gland, bone marrow, spleen, and lymph nodes (15). Re-pletion of incubation periods, culture supernatants were recovered and centrifuged, resulting in a cell-free cent studies have additionally documented the ability of TNF and other cytokines to stimulate the synthesis supernatant to be tested by ELISA. and release of certain hormones. Further investigation ELISA has revealed that not only do cells of the immune system influence those of the neuroendocrine system, but The generation of anti-TNF receptor rabbit sera, puthat the endocrine system, particularly the hypotharification of anti-TNF receptor rabbit IgG, labeling of lamic-pituitary axis, exerts an effect on immune cells anti-TNF receptor rabbit IgG with horseradish peroxias well. A particular hormone of interest which appears dase, and specificity testing were performed utilizing to be associated with TNF is prostaglandin-E 2 (PGE 2 ), methods described (18-21). Recombinant 55-and 75an arachidonic acid metabolite and a potent immuno-kDa TNF receptors (22) used to generate antibodies, suppressive agent. Lipopolysaccharide (LPS) is a as well as to generate ELISA, were kindly supplied by strong inducer of TNF on monocytes, which, in turn, Dr. Tadahiko Kohno of (Amgen, Inc., Thousand Oaks, stimulates the production of PGE 2 . Also, PGE 2 has CA). Briefly, anti-TNF receptor IgG (5 mg/ml in 0.05 M been found to have a regulatory effect on cytokines sodium bicarbonate buffer, pH 9.5) was added to each (16). Exogenous PGE 2 has been found to suppress LPSwell of a 96-well, flat-bottomed, polystyrene ELISA induced TNF mRNA accumulation in a dose-dependent plate (Corning Glass Works). Following an overnight manner (17).
incubation at 4ЊC, the wells were individually washed The focus of this study is to investigate the effects of with 0.2% Tween-20 (Bio-Rad Laboratories, Hercules, PGE 2 on TNF receptors, based on its participation in CA) in phosphate-buffered saline (PBS). Recombinant modulation of TNF biologic activity. Regulation of TNF TNF receptor standards and samples were added in receptor levels may further elucidate the mechanistic duplicate to the wells. The plates were incubated at details of receptor release, as well as possible parame-37ЊC for 3 hr, after which another washing occurred. ters of neuroendocrine involvement. As macrophages Horseradish peroxidase-labeled anti-TNF receptor IgG play a critical role in TNF and TNF receptor production was added to each well, and plates were incubated for and release, in vitro studies were performed using hu-1 hr in 37ЊC. The substrate, consisting of Immunopure man monocytic cell line THP-1. THP-1 cells produce ABTS tablets (2,2-azinobis [3-ethylbenzthiazoline-6high levels of TNF, and it closely mirrors characterissulfonic acid] diammonium salt) (Pierce, Rockford IL), tics found in normal activated macrophages.
hydrogen peroxide (Fisher Scientific, Fair Lawn, NJ), We examined the effect of stimulating THP-1 cells and 0.1 M sodium acetate buffer, pH 4.2, was added, with PGE 2 on TNF receptor release, as well as mRNA followed by room temperature incubation for 20 min. and surface receptor expression. PGE 2 activation of Results were obtained with the use of an EAR 400 AT THP-1 cells resulted in the decrease of TNF release, plate reader measuring OD 405 (SLT-Lab Instruments, as well as the increase of both forms of soluble TNF Salzburg, Austria). Soluble receptor concentrations receptors, allowing us to postulate the existence of two were calculated using the regression line derived from distinct modes of TNF activity regulation. the known standards for each receptor. This assay is not able to distinguish between TNF/TNF-R complex

MATERIALS AND METHODS
and soluble TNF-R. TNF ELISA was also utilized, the method for which was identical to TNF-R ELISA, and Cell Culture also cannot differentiate between TNF/TNF-R complex and TNF. THP-1 is a nonadherent, human, monocytic cell line which was obtained from the American Type Culture Iodination of rTNF-a and Binding Assay Collection (Rockville, MD). These cells were passed twice a week in RPMI 1640 (GIBCO BRL, Life Technol- The iodination of rTNF-a was performed as described ogies, Grand Island, NY), supplemented with 10% by Bolton et al. (23). One hundred microgram of rTNFheat-inactivated fetal bovine serum (GIBCO BRL). The a (Genentech, South San Francisco, CA) in 20 ml of 0.1 cells were cultured in plastic flasks (Corning Glass M sodium borate buffer, pH 8.5, was added to the dried Works, Corning, NY) and incubated in a 37ЊC, CO 2 iodinated ester and the reaction mixture (Boltonincubator. Hunter Reagent, Monoiodo, 125 I, DuPont, Wilmington, DE). Free iodine was removed by gel filtration over a Activation of THP-1 Cells by PGE 2 NAP-5 column (Pharmacia, Sweden) with PBS containing 5 mg gelatin. One hundred fifty-microliter frac-THP-1 cells were seeded in 24-well tissue culture plates (Corning Glass Works) at a concentration of 1 tions were collected and 1-ml aliquots were assayed by AID CI 8152 / 6c10$$$242 05-20-96 10:28:48 cial AP: Cell Immuno Clinnigamma Counter (Pharmacia, LKB, Uppsala, Sweden). Peak fractions were pooled and stored at 4ЊC to be used within 1 month. The procedure for TNF binding assay analysis was performed as described by Gatanaga et al. (24). One million cells were incubated at 37ЊC for 6 hr in 1 ml of binding buffer (RPMI 1640 / 10% FBS), with or without 10 07 M PGE 2 in 24-well tissue culture plates. The cells were then incubated with 1 to 30 ng of 125 I-labeled TNF-a, in the presence or absence of a 100-fold excess of unlabeled TNF-a. Following a 3-to 4-hr incubation at 4ЊC, the buffer was decanted and cells were washed two to four times with ice-cold PBS (25). The cells were then solubilized in 1 ml of 1 N NaOH and radioactivity was determined by Clinnigamma Counter. The number of membrane TNF-a receptors and measure of their affinity were calculated by Scatchard plot analysis.

RNA Purification/Isolation and Analysis by Northern Blot
Total cellular RNA was isolated from THP-1 cells via the guanidinium isothiocyanate/cesium chloride method of purification. The RNA was quantitated spectrophotometrically at wavelength 260. Ten microliters of RNA was electrophoresed in 2.2 M formaldehyde in 1.2% agarose gels. The RNA is then transferred to nitrocellulose filter (Schleicher & Schuell, Keene, NH) for 15 min twice and then with 0.11 SSC, 0.1% SDS for 15 min. The filters were exposed to X-Omat XAR-5 (Kodak, Rochester, NY) X-ray films at 070ЊC. Band intensities were confirmed by analysis via gel digitizer RESULTS (UVP Image Storage-5000) and quantitatively com-Regulation of TNF and TNF Receptor Release from pared to those of b-actin, as well as to bands receiving PGE 2 -Activated THP-1 Cells in Vitro no treatment.
One million THP-1 cells were cultured per well with Inhibition Assays 10 07 M PGE 2 for 6 hr at 37ЊC in a CO 2 incubator. Cellfree supernatants were assayed for TNF, as well as for Several assays were performed to determine the action of inhibitors of certain pathways. One millimolar 55-and 75-kDa TNF receptors by ELISA. Over four trials, the average TNF concentration for the cells incu-concentrations of Pefabloc (Boehringer Mannheim, Indianapolis, IN), a known serine protease inhibitor, bated in media was 229 pg/ml, while the mean for cells in PGE 2 was 161 pg/ml. Thus results show an average were added to THP-1 cells with and without PGE 2 . The cells were incubated in 37ЊC for a period of 6 hr, after 21 { 8.8% down regulation of the release of TNF (Fig.  1). PGE 2 suppression of TNF release in THP-1 cells is which the levels of TNF receptors were measured by ELISA.
consistent with prior findings (17). In contrast, detection of the soluble forms of the TNF receptors revealed Similarly, the THP-1 cells were subjected to incubation with N-ethylmaleimide (NEM) (Sigma Chemical a general increase compared to spontaneous release control levels. Fifty-five-kilodalton receptor release in-Co., St. Louis, MO), a known inhibitor of cell membrane internalization, again with and without PGE 2 . Incuba-creased an average 12.25 { 5.1%, from 104 to 122 pg/ ml, whereas 75-kDa receptor release increased 55.5 { tion of THP-1 cells in low temperature (16ЊC) is also known to exert the same inhibition of internalization 18%, from 199 to 305 pg/ml at the same PGE 2 concentration (Fig. 1). All values were significant by Student's and was utilized as well. All culture supernatants were centrifuged and subsequently analyzed by ELISA. and 10 08 M concentrations was also investigated, as manner as described above and incubated in a CO 2 incubator. The cells were subsequently cultured with well as the effect of a shorter 2-hr stimulation period. 125 I-labeled TNF, with or without an excess amount of Both variations resulted in the same general trend as cold TNF, and analyzed via binding assay. Activation was found with 6-hr, 10 07 M PGE 2 treatment (data not by PGE 2 for 6 hr resulted in a greater than twofold shown).
increase in the number of membrane-bound receptors (Fig. 2). It was determined through Scatchard analysis PGE 2 Regulation of Membrane TNF Receptors that the control group had 1190 surface receptors per To determine the specific regions of PGE 2 regulation cell, whereas a 6-hr incubation with PGE 2 yielded 2533 in THP-1 cells, the expression of cell surface receptors receptors per cell. Affinity constants remained compawas examined. THP-1 cells were cultured in the same rable within groups as the control affinity was 3.07 1 AID CI 8152 / 6c10$$$243 05-20-96 10:28:48 cial AP: Cell Immuno 10 010 M and for PGE 2 was 2.29 1 10 010 M. Again, a shorter activation period of 2 hr was also investigated, and the results indicate a similar increase of cell surface receptors (data not shown).

Northern Blot Analysis of PGE 2 Regulatory Effects of TNF and TNF Receptors
THP-1 cells were stimulated with PGE 2 for a period of 6 hr in 37ЊC, CO 2 incubation and then subjected to RNA purification methods and Northern hybridization. Resulting gels were then analyzed through use of a gel digitizer (UVP, Image Storage-5000, Ultra-Violet Products, San Gabriel, CA). RNA band densities of PGE 2 -stimulated cells were digitally compared to bactin mRNA levels, as well as additionally to mRNA of THP-1 cells receiving no treatment. As RNA band intensities in PGE 2 -treated cells indicate, the level of 55-kDa receptor was nearly identical to the b-actin levels (7% decrease compared to control). Seventy-five-kilodalton mRNA levels in PGE 2 -treated cells were shown to increase by 0.5% of the b-actin control density (Fig. 3). Results indicate no significant regulation of 55or 75-kDa receptors by PGE 2 at mRNA transcription levels. and followed by a 6-hr incubation in 37ЊC, after which soluble receptor levels were tested by ELISA. Cells treated with Pefabloc alone demonstrated depressed receptor numbers. For instance, the 55-kDa receptor by 169%. Pefabloc applied concurrently with PGE 2 rewas detected at a 59% decrease compared to the spontasulted in a depression of receptor numbers comparable neous release amount (Fig. 4A), and the 75-kDa recepto that exhibited with Pefabloc alone (244 { 2.4 versus tor decreased 78% (Fig. 4B). ization is an integral process in PGE 2 -mediated recep-Low-temperature incubation of THP-1 cells produced results similar to those of NEM stimulation. Incubation tor release, THP-1 cells were incubated with 1 mM NEM, an internalization inhibitor, for a period of 6 hr (16ЊC) yielded soluble 55-kDa TNF receptor levels of 1165 pg/ml, a 298% increase compared to its low-tem-in 37ЊC. The cell supernatants were then analyzed for soluble receptor levels by ELISA. Receptor level fluc-perature counterpart which produced only 293 pg/ml.

Inhibition of Serine Protease Mechanism by Pefabloc
The amount of soluble 75-kDa TNF receptors increased tuations were similar to that seen with Pefabloc treatment. Cells treated solely with NEM exhibited consid-by 192% in the cells incubated in normal temperature (37ЊC), as opposed to those treated with low tempera-erable downregulation of soluble TNF receptor levels when tested by ELISA, compared to spontaneous re-ture. lease levels. Seventy-five-kilodalton receptor levels in cells treated with NEM alone were a mere 10% of spon-DISCUSSION taneous release amounts (Fig. 5B), whereas 55-kDa receptor levels with NEM comprised only 16% of control PGE 2 has been found to control TNF activity via two modes: (i) Direct suppression of the TNF molecule (16), levels (Fig. 5A). Contrastingly, stimulation with PGE 2 elicited a 148% increase in 55-kDa TNF receptor levels and (ii) upregulation of TNF receptors which may in turn bind to and subsequently regulate TNF activity and a 144% increase in 75-kDa levels. THP-1 cells stimulated simultaneously with PGE 2 and NEM produced (14). This study not only supports the theory of the integration of neuroendocrine and immune systems, a downregulation in receptor release, one which is comparable to that found with NEM alone (85% for 55 kDa but additionally, it provides evidence of a new pathway of regulating the wide range of immune functions in and 88% for 75 kDa).
which TNF is involved. PGE 2 is a key inhibitor of tumor necrosis factor at both the transcriptional and the posttranscriptional levels (17). This suppression occurs in a dose-dependent manner and is not limited solely to the restriction of TNF, as interleukin-1 represents another cytokine regulated by PGE 2 (26). PGE 2 additionally exerts a mechanism of feedback inhibition upon TNF production, observed as TNF stimulation of fibroblasts and monocytes results in a decreased production of PGE 2 (26). The regulation of TNF production and release makes PGE 2 a significant immunomodulatory factor, an important link, so to speak, in the network of molecules and hormones bridging the immune and neuroendocrine systems together. PGE 2 regulation of TNF in THP-1 cells has been established in several studies (16,17). However, further evaluation revealed regulation of soluble TNF receptor levels with PGE 2 treatment as well. We report here that both 55-and 75-kDa soluble receptor levels increased dramatically with PGE 2 activation. This treatment, furthermore, increased the number of membrane-bound receptors by more than two-fold, with no significant change in affinity. This increase in cell surface and soluble receptor expression prompts us to expect a comparable upregulation at the protein transcription level. Interestingly, TNF receptor mRNA levels, as analyzed by Northern hybridization, remain unchanged within control groups and those which were activated by PGE 2 . These findings suggest the existence of an intracellular pool of TNF receptors situated in the region between transcriptional activities and cell surface expression. The existence of such would account for the variation in surface receptors while the mRNA levels remain unchanged. Yet another possibility for the discrepancy may be a greater translation of receptors than that detectible at the mRNA transcrip-