Regulation of murine macrophage proinflammatory and anti ...

Regulation of murine macrophage proinflammatory and antiinflammatory cytokines by ligands for peroxisome proliferator-activated receptor-␥: counter-regulatory activity by IFN-␥ David G. Alleva, Eric B. Johnson, Francisco M. Lio, Stefen A. Boehme, Paul J. Conlon, and Paul D. Crowe Neurocrine Biosciences, Inc., San Diego, California

Abstract: The prostaglandin, 15-deoxy ⌬12,14prostaglandin J2 (15d-PGJ2)1, and thiazolidinediones are ligands for the nuclear receptor, peroxisome proliferator-activated receptor (PPAR)-␥, which mediates anti-inflammatory activity by suppressing murine macrophage (M␾) production of the inflammatory mediator, nitric oxide (NO). Here, we elucidated this anti-inflammatory activity further by investigating whether PPAR-␥ ligands regulated a panel of proinflammatory and anti-inflammatory cytokines produced by primary inflammatory murine M␾ (thioglycollate-elicited peritoneal exudate M␾; PEM). Thiazolidinediones and 15d-PGJ2 suppressed lipopolysaccharide (LPS)-induced PEM production of NO and IL-12(p40) to a greater extent than IL-6 and TNF-␣ production. Whereas 15d-PGJ2 showed the greatest extent of suppression of proinflammatory mediator production, the thiazolidinedione, BRL49653, was the most potent compound studied. Surprisingly, treatment with the M␾-activation cytokine, IFN-␥, prevented PPAR-␥ ligands from suppressing the proinflammatory cytokines completely and reduced their suppression of NO production substantially, demonstrating that activation conditions affect PPAR-␥-mediated, anti-inflammatory activity. Western analysis demonstrated that the antagonistic activity of IFN-␥ did not involve modulation of PPAR-␥ expression but showed that IFN-␥ interfered with PPAR-␥ ligand regulation of p42/p44 MAP kinase activation and the cytosolic disappearance of NF-␬B upon LPS stimulation. Finally, we showed that PPAR-␥ ligands did not substantially modulate production of the anti-inflammatory cytokine, IL-10, and that antibody-mediated neutralization of IL-10 did not prevent the ligands from suppressing proinflammatory mediator production. In contrast to studies with noninflammatory human monocytes and M␾, our results demonstrate that primary murine inflammatory M␾ are extremely sensitive to the anti-inflammatory activity of PPAR-␥ ligands. These results suggest that drugs such as thiazolidinediones may be most effective in suppressing M␾ activity early (i.e., in the absence of lymphocyte-

derived IFN-␥) in the inflammatory process. J. Leukoc. Biol. 71: 677– 685; 2002. Key Words: prostaglandin 䡠 thiazolidinedione 䡠 nuclear receptor 䡠 nitric oxide

INTRODUCTION The peroxisome proliferator-activated receptor (PPAR)-␥ is a member of the PPAR family of nuclear receptors, which when activated upon ligand binding, regulates genes involved in lipid metabolism in various tissues [1, 2]. This receptor also displays anti-inflammatory activity, because PPAR-␥ ligands, such as derivatives of the prostaglandin J2 (PGJ2) and the synthetic thiazolidinediones, troglitazone and BRL49653, suppress production of the inflammatory mediator nitric oxide (NO) by down-regulating expression of inducible NO synthase (iNOS) in interferon-␥ (IFN-␥)- and lipopolysaccharide (LPS)activated murine primary macrophage (M␾) and the murine M␾ RAW 264.7 cell line [3–5]. Although several thiazolidinediones and PGJ2 derivatives demonstrate this anti-inflammatory activity, 15-deoxy ⌬12,14-PGJ2 (15d-PGJ2) appears to display the strongest activity [6]. It is interesting that the anti-inflammatory mechanism of PPAR-␥ may be indirect, because the activation of PPAR-␥ induced by ligand binding causes this complex to bind several critical transcription factors including nuclear factor-␬B (NF-␬B), which is involved in iNOS gene transcription, thus inhibiting their transcriptional activity [6]. The claim that PPAR-␥ ligands are potent anti-inflammatory agents has been challenged in at least one study [7], which showed that thiazolidinediones could not regulate murine RAW cell production of the proinflammatory cytokine tumor necrosis factor ␣ (TNF-␣) and interleukin (IL)-6. This study [7] also showed that 15d-PGJ2 suppressed murine RAW cell pro-

Correspondence: Drs. David Alleva and Paul Crowe, Neurocrine Biosciences, Inc., 10555 Science Center Dr., San Diego, CA 92121-1102. E-mail: [email protected] and [email protected] Received November 10, 2000; revised August 18, 2001; accepted November 30, 2001.

Journal of Leukocyte Biology Volume 71, April 2002 677

duction of these cytokines only modestly. Differences in activity between 15d-PGJ2 and thiazolidinediones were also observed in the regulation of TNF-␣, IL-1, and IL-6 produced by activated human monocytes and M␾ [7, 8]. In addition, the M␾ activation stimulus may affect the ability of PPAR-␥ ligands to suppress cytokine production, because 15d-PGJ2 (at nontoxic doses) suppressed human monocyte TNF-␣ and IL-6 production only when monocytes were activated with okadaic acid [8] or phorbol ester [7, 8] but not with LPS [7, 8]. These studies suggest that 15d-PGJ2 and thiazolidinediones display their strongest anti-inflammatory activity in the regulation of murine M␾ NO production but display only modest, if any, regulation of proinflammatory cytokine production, which may depend on the activation stimulus. However, these differences in the regulation of cytokines versus NO production and in the effects of activation stimuli have not been addressed yet in a single study, primarily because noninflammatory human monocytes/M␾ have been used to study the regulation of cytokines, which are cells incapable of producing NO [9]. Furthermore, it is unknown whether synthetic PPAR-␥-specific ligands, such as the thiazolidinediones, regulate other proinflammatory cytokines such as IL-12(p40) or anti-inflammatory cytokines such as IL-10. Therefore, by studying murine peritoneal exudate M␾ (PEM) as well as the RAW 264.7 cell line, we defined further the activity of PPAR-␥ ligands on M␾ activation by determining the regulatory action of 15d-PGJ2 and thiazolidinediones on the production of a panel of proinflammatory and anti-inflammatory cytokines relative to that of NO. We found that production of the proinflammatory mediators NO and IL-12(p40) was far more sensitive to the suppressive effects of both types of PPAR-␥ ligands than the production of IL-6 and TNF-␣, that these ligands did not suppress but modestly up-regulated production of IL-10, and that different M␾ activation conditions, such as the presence of IFN-␥, strongly affected PPAR-␥ ligand activity. Our results suggest that inflammatory conditioning, in the absence of lymphocyte-derived IFN-␥, sensitizes M␾ to the anti-inflammatory activity of PPAR-␥ ligands. Because M␾ activation occurs in the absence of lymphocyte activation (i.e., IFN-␥ production) during early stages of inflammation or diseases such as atherosclerosis, PPAR-␥ ligand drugs such as thiazolidinediones may be most effective during these inflammatory situations.

MATERIALS AND METHODS Animals Four- to six-week-old C57BL/6 male mice were purchased from the Jackson Laboratories (Bar Harbor, ME) and were maintained for 1–2 weeks after arrival under germ-free conditions in a vivarium. The Institutional Animal Care and Use Committee approved all animal experimentation prior to implementation.

Reagents LPS (Escherichia coli: 0111:B4; Sigma Chemical Co., St. Louis, MO) and recombinant murine IFN-␥ (PharMingen, San Diego, CA) were stored at ⫺80°C and diluted in complete medium immediately before use. The PPAR-␥ ligands, 15d-PGJ2 (Biomol, Plymouth Meeting, PA), ciglitazone (Biomol), and BRL49653 (American Radiolabeled Chemicals, St. Louis, MO) were dissolved in 100% dimethyl sulfoxide (DMSO) at 100 mM and stored in aliquots at

678

Journal of Leukocyte Biology Volume 71, April 2002

⫺80°C under nitrogen. Complete medium consisted of RPMI-1640 medium supplemented with 2 mM L-glutamine, 0.5% HEPES (Cellgro, Herndon, VA), 5 ␮g/ml penicillin, and 100 U/ml streptomycin (Gibco-BRL, Grand Island, NY), and 10% fetal bovine serum (BioWhittaker, Walkersville, MD) was used for culturing PEM and RAW cells. No endotoxin (⬍10 pg/ml) was detected using the Limulus amoebocyte assay (BioWhittaker) in the reagents and media described above. Antimurine IL-10 monoclonal antibody (mAb; JES5-2A5) and its immunoglobulin G (IgG)1 isotype-matched, control mAb (R3–34) were purchased from PharMingen.

M␾ isolation, culturing, and assessment of cytokines and NO PEM were elicited by an intraperitoneal injection of 2 ml sterile thioglycollate broth (Remel, Lenexa, KS) 4 days prior to peritoneal lavage as described previously [10]. PEM were isolated by adherence to plastic in which 2 ⫻ 105 peritoneal exudate cells were seeded per well of 96-well, flat-bottom, tissue culture-treated plates and were allowed to adhere for 2 h, and nonadherent cells were washed away. The BALB/c peritoneal M␾ cell line, RAW 264.7 (American Type Culture Collection, Manassas, VA; TIB-71), was seeded (105/well) in 96-well, flat-bottom, tissue culture-treated plates and incubated for 2 h before activation. PEM and RAW cells were activated with LPS (100 ng/ml) with or without IFN-␥ (100 U/ml) in the presence or absence of PPAR-␥ ligands, and conditioned medium was collected at an optimal culture period of 16 h [10] and stored at ⫺20°C for future assessment of cytokines and nitrite levels. TNF-␣, IL-6, IL-10, and IL-12(p40) levels in conditioned medium were assessed by enzyme-linked immunosorbent assay (ELISA; PharMingen), and nitrite (a stable nitrogen oxide formed from the spontaneous oxidation of NO) was assessed using the Griess reagent as described previously [11, 12].

Western blot analysis RAW 264.7 cells and PEM were harvested from cultures (107/well of a six-well, tissue-culture plate) by treatment with trypsin-ethylenediaminetetraacetate (EDTA), washed once with phosphate-buffered saline (PBS), and then lysed in ice-cold buffer containing 10 mM Tris, pH 8, 1 mM EDTA, 150 mM NaCl, 1% Nonidet P-40, 10 ␮g each leupeptin and aprotinin, and 2 mM phenylmethylsulfonyl fluoride. After incubation on ice for 30 min, cell lysates were centrifuged to precipitate-insoluble material, including intact nuclei. This method of protein-lysate preparation maintains nuclear membrane integrity, thereby only cytosolic proteins were examined. The soluble protein extracts were boiled in 2⫻ sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer for 5 min. Proteins were resolved on 10% reducing SDS-PAGE gels (Novex, San Diego, CA) and transferred to nitrocellulose. Nitrocellulose membranes were blocked for 1 h with blocking buffer (10 mM Tris, pH 7.5, 150 mM NaCl, 2% bovine serum albumin, 0.5% ovalbumin, 2.5% blotting grade nonfat dried milk). Then, the nitrocellulose membranes were incubated overnight with the primary antibody at 4°C diluted in blocking buffer. The antibodies used include anti-PPAR-␥ Ab (0.8 ␮g/ml; Santa Cruz Biotechnology, Santa Cruz, CA), anti-iNOS Ab (Santa Cruz Biotechnology), anti-HSP70 Ab (0.5 ␮g/ml; Transduction Laboratories, San Diego, CA), antiphosphorylated p42/p44 mitogen-activated protein (MAP) kinase Ab (clone 12D4; Upstate Biotechnology, Lake Placid, NY), anti-p42/p44 MAP kinase Ab (clone 1B3B9; Upstate Biotechnology), or anti-NF-␬B (cat. #06-886; Upstate Biotechnology). Protein bands were visualized by incubating blots with a secondary peroxidase-labeled antibody (Pierce Chemical Co., Rockford, IL) followed by chemiluminescence detection using SuperSignal West Dura reagent (Pierce Chemical Co.) and Kodak BioMax film (Rochester, NY). Densitometric quantitation of the Western blots was done using the NIH Image Analysis Program.

RESULTS PPAR-␥ ligand regulation of LPS-stimulated, proinflammatory cytokine production by primary murine M␾ We determined whether the PPAR-␥ ligands, 15d-PGJ2, and the thiazolidinediones BRL49653 and ciglitazone regulated http://www.jleukbio.org

Fig. 1. Regulation of murine PEM proinflammatory mediator production by PPAR-␥ ligands. Thioglycollate-elicited, peritoneal M␾ (2⫻105) from C57BL/6 mice were activated with LPS (100 ng/ml) in the presence or absence of the PPAR-␥ ligands 15d-PGJ2, BRL49653, and ciglitazone for 16 h, and culture-conditioned medium was analyzed for cytokines by ELISA and nitrite using the Griess reagent. Values are the mean and SE of triplicate cultures from one experiment representative of at least three experiments. (}) Cytokine or nitrite levels in control cultures activated with LPS alone.

inflammatory murine PEM production of proinflammatory cytokines in a manner similar to that of NO. Thioglycollateelicited PEM were activated with LPS in the presence or absence of different concentrations of the PPAR-␥ ligands for 16 h, and levels of the proinflammatory cytokines TNF-␣, IL-6, and IL-12(p40), in addition to nitrite (a stable form of NO) levels, were measured in conditioned medium. There were differences in the efficacy and potency among these ligands to suppress proinflammatory cytokine and NO production (Fig. 1), and although all three ligands regulated production of NO and cytokines, 15d-PGJ2 suppressed NO, IL-12(p40), and IL-6 to a greater extent (i.e., more efficacious) than the thiazolidinediones, whereas all ligands suppressed TNF-␣ production to the same extent. 15d-PGJ2 also showed the greatest efficacy in suppressing NO production (almost a complete inhibition), followed by the degree of IL-12(p40) suppression (eight- to tenfold), IL-6 (sixfold), and TNF-␣ (twofold; Fig. 1). However, the potency (i.e., lowest effective dose) among the PPAR-␥ ligands differed, and although 15d-PGJ2 showed similar potency to BRL49653 in suppressing NO production (Fig. 1), BRL49653 was the most potent at suppressing the cytokines (Fig. 1). Ciglitazone was the least potent in suppressing production of any proinflammatory mediator (i.e., 10 –100

times less potent than BRL49653), despite the fact that both thiazolidinediones achieved a similar extent of suppression of each cytokine and NO. The suppressive effects of PPAR-␥ ligands on PEM cytokine and NO production were not associated with a reduction in PEM viability or metabolic activity, as measured by conversion of Alamar Blue (Fig. 2A). It is important to note that only the activity of nontoxic concentrations of each ligand was shown in Figure 1; 15d-PGJ2 was toxic above 8 ␮M, BRL49653 was toxic above 16 ␮M, and ciglitazone showed toxicity above 60 ␮M (Fig. 2A). In addition, control cultures with the DMSO vehicle alone at concentrations similar to those present in cultures with the PPAR-␥ ligands (i.e., highest DMSO concentration ⬍0.1%) showed no toxicity and did not modulate cytokine or NO production (unpublished results). Although others have shown that PPAR-␥ ligands suppress NO production via suppression of iNOS mRNA transcription in murine M␾ [3, 5, 13], modulation of cytokine gene transcription by these ligands has not been investigated. Using an effective, nontoxic dose of 15d-PGJ2 and the thiazolidinedione BRL49653, we demonstrated that LPS-induced mRNA levels of several proinflammatory cytokine genes expressed in PEM were suppressed substantially by these PPAR-␥ ligands (Fig.

Alleva et al. PPAR-␥ ligand regulation of murine M␾ cytokines

679

between PEM and this cell line. In contrast to the regulation of PEM, 15d-PGJ2 and BRL49653 showed similar potencies in the regulation of NO and cytokine production by RAW cells. Ciglitazone was the least potent and effective in regulating RAW cell proinflammatory mediator production, which was also observed with PEM (Fig. 4). The effects on RAW cell viability by these ligands were similar to that of primary M␾ (unpublished results), and only nontoxic doses of the ligands were used to modulate RAW cell function. In summary, by using PEM and RAW 264.7 cells, we demonstrated that murine M␾ production of NO and IL-12(p40) was more sensitive than IL-6 and TNF-␣ production to the suppressor activity of PPAR-␥ ligands. Furthermore, 15d-PGJ2 consistently displayed the greatest efficacy in suppressor activity, whereas BRL49653 showed, on average, the greatest potency, which in some cases (i.e., RAW cells) was matched by 15d-PGJ2.

IFN-␥ modulation of PPAR-␥ ligand activity on proinflammatory mediator production Because activation conditions may affect PPAR-␥ ligand regulation of proinflammatory cytokines, we determined whether

Fig. 2. Effects of PPAR-␥ ligands on viability of murine PEM. Thioglycollateelicited, peritoneal M␾ (2⫻105) from C57BL/6 mice were activated with LPS (100 ng/ml) in the presence or absence of the PPAR-␥ ligands 15d-PGJ2, BRL49653, and ciglitazone for 16 h. Culture-conditioned medium was removed, and fresh medium was added with Alamar Blue vital dye. Cell viability was assessed by redox conversion of the dye, which absorbs at 570 nm. Data are the mean and SE of optical density (O.D.) values from triplicate cultures of one experiment representative of at least three experiments. (}) Viability of control cultures activated with LPS alone. *, Significantly different from control culture values (P⬎0.05) using the Student’s t-test to compare means.

3). Further, this experiment illustrates that in addition to mRNA levels of IL-12(p40), p(35), and IL-6, mRNA levels of the proinflammatory cytokines IL-1␣ and IL-18 were also suppressed by these ligands, whereas IL-1␤ mRNA levels were not (Fig. 3). These results are consistent with the observation that activation via ligand binding induces PPAR-␥ to bind and functionally inactivate critical transcription factors involved in cytokine gene activation (reviewed in ref. [6]); however, the ligand/PPAR-␥ complex may also be acting upstream of nuclear events.

PPAR-␥ ligand regulation of LPS-stimulated proinflammatory cytokine production by the murine M␾ RAW 264.7 cell line Next, we determined whether PPAR-␥ ligands showed similar differences in suppressor activities on cytokine and NO production by the murine M␾ RAW 264.7 cell line. Similar to the regulatory patterns observed in PEM, 15d-PGJ2 showed greater suppressor activity than the thiazolidinediones on RAW cell NO and cytokine production, in which the extent of 15d-PGJ2induced suppression of NO production (roughly eightfold) was greater than that of IL-6 and TNF-␣ (roughly fourfold; Fig. 4). IL-12(p40) was not detected in conditioned medium by LPSactivated RAW cells, demonstrating an apparent difference 680


Fig. 3. PPAR-␥ ligand modulation of M␾ proinflammatory cytokine mRNA expression. Thioglycollate-elicited, peritoneal M␾ (107) were cultured in the presence or absence of LPS (100 ng/ml) with or without the PPAR-␥ ligands 15d-PGJ2 and BRL49653 for 8 h, at which time total RNA was extracted and analyzed for mRNA levels of IL-12 p35 and p40 subunits, IL-6, IL-18, IL-1␣, and IL-1␤, in addition to the control genes L32 and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and using the RNase protection assay. Samples were separated on a sequencing gel and visualized by autoradiography (upper panel). RNA was quantified by densitometry and normalized to control L32 mRNA (lower panel).

http://www.jleukbio.org

Fig. 4. Regulation of murine RAW 264.7 cell proinflammatory mediator production by PPAR-␥ ligands. Murine RAW 264.7 cells (2⫻105) were activated with LPS (100 ng/ml) in the presence or absence of the PPAR-␥ ligands 15d-PGJ2, BRL49653, and ciglitazone for 16 h, and culture-conditioned medium was analyzed for cytokines by ELISA and nitrite using the Griess reagent. Values are the mean and SE of triplicate cultures from one experiment representative of at least three experiments. (}) Cytokine or nitrite levels in control cultures activated with LPS alone.

the T-cell-derived cytokine IFN-␥ altered PPAR-␥ ligand activity, which is a potent up-regulator of proinflammatory mediator production by LPS-activated M␾ [14]. IFN-␥ up-regulated LPS-induced production of all proinflammatory mediators (Fig. 5; compare with control values in Fig. 1). However, IFN-␥ treatment completely prevented all PPAR-␥ ligands from suppressing production of the proinflammatory cytokines and allowed only the highest doses of the 15d-PGJ2 and thiazolidinediones to suppress NO production, in which 15dPGJ2 was the most potent and effective (Fig. 5). IFN-␥ did not modulate PPAR-␥ ligand effects on cell viability (see Fig. 2B). These effects of IFN-␥ were also apparent in the RAW 264.7 cell line (Fig. 6); IFN-␥ treatment of LPS-activated RAW cell NO production shifted the 15d-PGJ2 dose-response curve, resulting in reduced potency of this ligand by roughly threefold. These effects of IFN-␥ on NO production were not a consequence of an inhibition of PPAR-␥ protein expression, because IFN-␥ did not modulate LPS-induced PPAR-␥ expression in RAW cells during induction of iNOS expression (Fig. 7). To investigate whether IFN-␥ could modulate LPS-inducible intracellular signaling molecules involved in PPAR-␥-mediated suppressor activity, PEM were activated with LPS or LPS plus IFN-␥ in the presence or absence of the PPAR-␥ ligands 15d-PGJ2 and ciglitazone for 30 min. The phosphorylation status of p42/p44 MAP kinases and the levels of cytosolic NF-␬B were then measured by Western blot analysis (Fig. 8). Our results, consistent with others [15], show clearly that 15d-PGJ2 modulated LPS-induced MAP kinase phosphorylation substantially more than PPAR-␥-specific ligands (i.e., ciglitazone shown in Fig. 8A and BRL49653 shown in ref. [15]). Surprisingly, IFN-␥ caused a dramatic down-regulation of LPS-induced MAP kinase activation and therefore blocked the ability of the PPAR-␥ ligands from modulating

p42/p44 MAP kinase phosphorylation (Fig. 8A). Additionally, although LPS alone induced the disappearance of roughly 85% of cytosolic NF-␬B (presumably as a result of nuclear translocation), 15d-PGJ2 and ciglitazone blocked this event partially, as is evident by the elevation in cytosolic NF-␬B levels (Fig. 8B), which is consistent with their suppression of proinflammatory mediator production. Guyton et al. [15] have demonstrated similar findings (published while this manuscript was prepared) by measuring the degradation of I␬B␣ in the cytosol, a factor that binds and tethers NF-␬B in the cytoplasm, thereby preventing its translocation to the nucleus. Although IFN-␥ did not modulate LPS-induced NF-␬B cytoplasmic levels substantially (Fig. 8B), IFN-␥ prevented PPAR-␥ ligands from interfering with the decrease of cytoplasmic NF-␬B amounts, which is consistent with their lack of regulation of proinflammatory mediator production in the presence of IFN-␥.

Effects of PPAR-␥ ligands on production of the anti-inflammatory cytokine IL-10 It is possible that IFN-␥ prevented PPAR-␥ ligand-induced activity indirectly by modulating production of the anti-inflammatory cytokine IL-10. IL-10 production is induced by LPS and works in an autocrine way to suppress proinflammatory mediator production [16]. It is interesting that IFN-␥ is a potent inhibitor of IL-10 production (unpublished results; and reviewed in ref. [16]), which would be a plausible mechanism underlying the antisuppressor activity of IFN-␥. We addressed the role of IL-10 in PPAR-␥ ligand-induced suppression by determining first whether PPAR-␥ ligands could modulate production of IL-10. Contrary to the regulation of proinflammatory cytokines, LPS-induced PEM IL-10 production was up-regulated modestly by the 15d-PGJ2 and thiazolidinediones


681

Fig. 5. Effect of IFN-␥ on PPAR-␥ ligand-induced, murine PEM proinflammatory mediator production. Thioglycollate-elicited, peritoneal M␾ (2⫻105) from C57BL/6 mice were activated with LPS (100 ng/ml) plus IFN-␥ (100 U/ml) in the presence or absence of the PPAR-␥ ligands 15d-PGJ2, BRL49653, and ciglitazone for 16 h, and culture-conditioned medium was analyzed for cytokines by ELISA and nitrite using the Griess reagent. Values are the mean and SE of triplicate cultures from one experiment representative of at least three experiments. (}) Cytokine or nitrite levels in control cultures activated with LPS alone. *, Significantly different from control culture values (P⬎0.05) using the Student’s t-test to compare means.

(roughly a 10 –30% increase; Fig. 9). Although these results suggest that IL-10 may be involved in PPAR-␥-mediated suppression, the modest elevation in IL-10 production by these ligands does not support such an idea sufficiently. To further address this issue, Ab-mediated neutralization of IL-10 activity in LPS-activated PEM cultures did not alter the suppressor activity of PPAR-␥ ligands substantially (Table 1). As expected, neutralization of IL-10 caused a four- to sixfold upregulation of proinflammatory cytokine production (but not of NO production). Although there was some reduction in the percent (i.e., relative) suppression of cytokines under some of the conditions, the absolute amount of cytokines suppressed by these ligands was actually greater in the presence of anti-IL-10 mAb. In addition, anti-IL-10 Ab treatment modestly reduced the amount of suppression of NO production induced by ciglitazone and BRL49653, however neutralization of IL-10 caused a modest decrease rather than an increase in NO production, which would give the appearance of a diminished, relative suppression induced by these ligands (Table 1). These results demonstrate that PPAR-␥ ligands have a modest stimulatory effect on IL-10 production but that this anti-inflammatory cytokine does not play a significant role in PPAR-␥-induced suppression of proinflammatory mediator production. 682


Fig. 6. Effect of IFN-␥ on PPAR-␥ ligand-induced, murine RAW 264.7 cell NO production. Murine RAW 264.7 cells (2⫻105) were activated with LPS (100 ng/ml) with or without IFN-␥ (100 U/ml) in the presence or absence of the PPAR-␥ ligand 15d-PGJ2 for 16 h, and culture-conditioned medium was analyzed for nitrite using the Griess reagent. Data are normalized values of the mean of triplicate cultures from one experiment (representative of five experiments), in which mean values of NO levels in cultures treated with PPAR-␥ ligands were divided by mean values of their respective, untreated control cultures. SE ⬎ 10% in all calculations. For convenience, the dotted line was included to display differences in EC50% values.

http://www.jleukbio.org

DISCUSSION The nuclear receptor PPAR-␥, when activated by ligands such as derivatives of PGJ2 and thiazolidinediones, is involved in controlling murine M␾ production of the inflammatory mediator NO [3, 13]. Further, we elucidated this anti-inflammatory activity by demonstrating that natural (i.e., 15d-PGJ2) and synthetic (i.e., thiazolidinediones) PPAR-␥ ligands suppressed murine M␾ production of the proinflammatory cytokines IL-12, IL-6, and TNF-␣ (in addition to NO production) but not production of the anti-inflammatory cytokine IL-10. These ligands suppressed gene transcription of the proinflammatory cytokines mentioned above in addition to IL-1␣ and IL-18. Although there was a large difference in potency among the PPAR-␥ ligands, the two thiazolidinediones achieved a similar extent of suppression, suggesting that both compounds use a similar PPAR-␥-mediated suppressor pathway but have different potencies in binding to or activating PPAR-␥. In contrast, 15d-PGJ2 appears to use, perhaps, an additional suppressor mechanism, because it showed a greater consistent degree of suppression than the thiazolidinediones, even in the presence

Fig. 7. Regulation of PPAR-␥ expression in murine M␾. Lysates were prepared from resting RAW264.7 cells (lane 1) or RAW264.7 cells that had been cultured for 16 h with 100 ng/ml LPS (lane 2) or LPS plus 100 U/ml IFN-␥ (lane 3). PPAR-␥ and iNOS proteins were detected by immunoblotting using a mAb specific for each protein. Equivalent protein loading was confirmed by subsequent immunoblotting for HSP-70.

of IFN-␥ (i.e., suppression of NO production). Indeed, additional suppressor mechanisms by 15d-PGJ2 have been shown in which 15d-PGJ2 inhibited I␬B kinase activity directly in activated M␾ [5], which may explain why 15d-PGJ2 but not some thiazolidinediones could suppress human M␾ cytokine (i.e., TNF-␣ and IL-6) production [7]. The differences in suppressor efficacy and potency among PPAR-␥ ligands may be attributed to differences in their PPAR-␥ binding sites and affinities, which would lead to differences in their ability to activate PPAR-␥ for sequestering transcription factors. Ligand-activated PPAR-␥ complexes appear to impart anti-inflammatory activity by binding critical transcription factors such as NF-␬B, AP-1, STAT-1, and Ets

Fig. 8. Phosphorylation status of p42/p44 MAP kinases and NF-␬B levels in the cytoplasm of murine PEM. C57BL/6-adherent PEM (10 million) were cultured per well of a six-well, tissue-culture plate with medium alone, 5 ␮M 15d-PGJ2, or 20 ␮M ciglitazone for 1 h followed by a 30 min incubation in the presence or absence of LPS (100 ng/ml) or LPS plus IFN-␥ (100 U/ml), and cytosolic lysates were analyzed for levels of phosphorylated MAP kinase (A) and NF-␬B (B) by Western analysis. Protein loading was equilibrated by immunoblotting for total p42/p44 MAP kinase, followed by densitometric quantitation as shown.

Fig. 9. Regulation of murine PEM IL-10 production by PPAR-␥ ligands. Thioglycollate-elicited, peritoneal M␾ (2⫻105) from C57BL/6 mice were activated with LPS (100 ng/ml) in the presence or absence of the PPAR-␥ ligands 15d-PGJ2, BRL49653, and ciglitazone for 16 h, and culture-conditioned medium was analyzed for the anti-inflammatory cytokine IL-10 by ELISA. Values are the mean and SE of triplicate cultures from one experiment representative of at least three experiments. (}) IL-10 levels in control cultures activated with LPS alone. *, Significantly different from control culture values (P⬎0.05) using the Student’s t-test to compare means.


683

TABLE 1.

Role of IL-10 Production on PPAR-␥ Ligand-Induced Suppression of M␾ Proinflammatory Cytokine Production LPS-activated M␾ proinflammatory mediator productiona

M␾ treatment Medium 15d-PGJ2 Ciglitazone BRL49653

IL-12 (ng/ml)

IL-6 (ng/ml)

TNF-␣ (ng/ml)

Nitrite (␮M)

None

␣-IL-10

None

␣-IL-10

None

␣-IL-10

None

␣-IL-10

0.58 0.14 (32) 0.05 (11) 0.05 (11)

2.4 0.62 (26) 0.34 (14) 0.39 (16)

6.1 3.7 (61) 5.0 (82)* 2.3 (38)

34.3 25.7 (76) 28.8 (85) 22.6 (67)

2.8 1.9 (69) 1.2 (42) 1.1 (38)

19.6 12.7 (62) 7.5 (37) 7.3 (36)

14.0 3.6 (26) 6.1 (44) 8.4 (60)

11.0 2.9 (29) 7.9 (77) 8.6 (84)

a Thioglycollate-elicited peritoneal M␾ (2⫻105) from C57BL/6 mice were cultured with LPS (100 ng/ml) in the presence or absence of anti-IL-10 mAb (2 ␮g/ml) with or without PPAR-␥ ligands (15d-PGJ2 and BRL49653⫽1 ␮M; ciglitazone⫽10 ␮M) for 16 h, and conditioned medium was analyzed for cytokine levels by ELISA and nitrite levels using the Griess reagent. The mean of triplicate culture values is shown, and the percentage of the maximum (without PPAR-␥ ligands) cytokine production is shown in parentheses. SE for all means was ⬍15%. All values were significantly different from control culture (medium) values (P⬎0.05) using the Student’s t-test to compare means, except where indicated (*).

required for full expression of iNOS and proinflammatory cytokine genes in activated M␾ (reviewed in ref. [6]). Perhaps the dynamics of thiazolidinedione and prostaglandin binding to PPAR-␥ lead to differences in the quantity and quality of transcription-factor recruitment, which would dictate the degree of proinflammatory gene suppression. This model may also offer insight into the differential regulation of proinflammatory and anti-inflammatory genes (i.e., IL-10) by PPAR-␥ ligands. The lack of substantial regulation of IL-10 production by PPAR-␥ is similar to the activity of other nuclear receptors such as the glucocorticoid receptor [17], retinoid X receptor [18], and the 1,25 hydroxy-vitamin D3 receptor [19]. Ligandmediated activation of these receptors in M␾ leads to suppression of one or more proinflammatory cytokines but not of IL-10, in which the proposed suppressor mechanisms also involve sequestering transcription factors. Further investigations should determine whether a similar set of transcription factors binds each of these activated nuclear receptors, which would explain their preferential suppression of proinflammatory but not antiinflammatory cytokines. This is the first study comparing cytokine and NO modulation by PPAR-␥ ligands. Our results demonstrate that the rank order of sensitivity to PPAR-␥ ligand-induced suppression of murine M␾ proinflammatory mediator production was NO ⬎ IL-12(p40) ⬎ IL-6 ⬎ TNF-␣. Results of others who studied the regulation TNF-␣ and IL-6 in RAW cells or in human M␾ [7] appear to support our results, although those studies lacked toxicity data. Thieringer et al. [7] have shown that TNF-␣ and IL-6 produced by RAW cells activated with LPS alone were suppressed by relatively high doses of 15d-PGJ2, yet the toxicity profile of these compounds was not shown. Because our results demonstrate that 15d-PGJ2 was toxic above 8 ␮M (see Fig. 2), it is possible that the only doses shown to suppress TNF-␣ production were toxic (i.e., 25 and 50 ␮M). Accordingly, only one dose of 15d-PGJ2 (12.5 ␮M) that may have been nontoxic in that study suppressed IL-6 production by roughly fourfold [7]. This study [7] is consistent with our results here that production of TNF-␣ is not as sensitive as that of IL-6 to the suppressive activity of nontoxic doses of PPAR-␥ ligands. Our assessment of the toxic range of 15d-PGJ2 has been corroborated by others [3, 20]. Also consistent with our murine M␾ results, others [7, 8] have demonstrated that LPS-induced 684


TNF-␣ production by in vitro, matured human M␾ is only suppressed modestly by the highest nontoxic doses of 15dPGJ2 (i.e., 10 –12.5 ␮M), in which greater effects were seen on IL-6 production. Unlike murine PEM, human monocytes/M␾ do not produce iNOS or NO upon activation with any known agents [9] and require IFN-␥ for induction of IL-12(p40) synthesis (unpublished observation; and reviewed in ref. [21]), an activation requisite that would block PPAR-␥ ligand activity. Therefore, the profound, suppressive effects of PPAR-␥ ligands on murine PEM NO and IL-12(p40) production that we show here demonstrate that the use of murine “inflammatory,” primary M␾ is an ideal system for elucidating the anti-inflammatory activity of PPAR-␥ ligands. The regulatory activity of PPAR-␥ ligands was dependent on the activation stimulus, because suppression was only achieved when M␾ were activated with LPS alone, whereas the presence of IFN-␥ prevented suppression of the cytokines completely and allowed only the highest nontoxic doses of the 15d-PGJ2 (i.e., 8 ␮M) and the thiazolidinediones BRL49653 and ciglitazone (i.e., 16 ␮M and 60 ␮M, respectively) to modulate NO production. This IFN-␥ effect on PPAR-␥ ligand activity is not only novel but is supported by results of others who used a combination of LPS plus IFN-␥ or IFN-␥ alone [3–5, 20] and showed that PGJ2 derivatives modulate murine M␾ NO production only modestly. Two studies [3, 20] show that 12d-PGJ2 or 15d-PGJ2 at the highest nontoxic doses tested (i.e., 10 ␮M and 5 ␮M, respectively) inhibited only NO production by 50% in M␾ activated with LPS plus IFN-␥, which is consistent with the 50% suppression observed by the highest nontoxic 15d-PGJ2 dose of 8 ␮M in our study (see Fig. 5). In addition, Ricote et al. [4] showed that 15d-PGJ2 could suppress NO levels produced by PEM activated with IFN-␥ alone; 15d-PGJ2 appeared to be more potent in this system (lowest effective dose⫽1 ␮M) relative to those that used the combination of LPS plus IFN-␥, probably because PPAR-␥ ligand activity would be greater in cells that are less activated (i.e., by IFN-␥ alone). Nevertheless, consistent with our results, other studies [4, 5] show that nontoxic doses of the thiazolidinediones (i.e., ⬍20 ␮M) showed a modest, if any, modulation of NO production by M␾ activated in the presence of IFN-␥. In addition, we demonstrate here for the first time that IFN-␥ prevented 15d-PGJ2 (and the thiazolidinediones) from suphttp://www.jleukbio.org

pressing cytokine production completely, whereas other studies that showed suppresser activity by PPAR-␥ ligands assessed cytokine production in the absence of IFN-␥ [7, 8, 22]. Because we showed that IFN-␥ did not modulate LPSinduced M␾ PPAR-␥ expression, it is unlikely that IFN-␥ blocked PPAR-␥ ligand-induced suppression via down-regulation of PPAR-␥. Although IFN-␥ is known to inhibit LPSstimulated PEM IL-10 production (our unpublished observation; and reviewed in ref. [16]), we showed that IL-10 does not play a major role in PPAR-␥ ligand-induced suppression of proinflammatory mediator production, suggesting that IFN-␥ prevents suppression by a mechanism other than inhibiting IL-10 production. It has been suggested that ligand-induced activation of PPAR-␥ may suppress M␾ activation (i.e., NO production) by acting as a sequestering complex that binds and inactivates transcription factors necessary for proinflammatory cytokine and iNOS gene transcription (reviewed in ref. [6]). Perhaps IFN-␥ stimulates high levels of these transcription factors that would out-compete the sequestering activity of an activated PPAR-␥ complex. Alternatively, unique, IFN-␥-induced, stimulatory factors that do not bind activated PPAR-␥ may play a role. Here, our results demonstrate that PPAR-␥ ligands affect events upstream of the nuclear sequestering action, because 15d-PGJ2 and ciglitazone partially blocked the LPS-induced disappearance of cytosolic NF-␬B, and that IFN-␥ affected this activity strongly. Our results are consistent with those of Guyton et al. [15], who demonstrated that BRL49653 up-regulated cytosolic levels of I␬B␣ (which correlates with cytosolic NF-␬B levels) by roughly twofold. Our results suggest that the influence of PPAR-␥ ligands on cytokine and NO production interferes with signaling molecules upstream from transcription-factor binding DNA promoter elements. These observations raise the question of whether the PPAR-␥/ligand complexes are acting further upstream than previously believed or whether these ligands act independently of PPAR-␥. In conclusion, we have clarified discrepancies in the literature showing that PPAR-␥ potently regulates NO and proinflammatory cytokines produced by murine inflammatory M␾ activated with LPS alone and that IFN-␥ interferes strongly with this activity. The impact of PPAR-␥ ligands on inflammatory M␾ may be strongest during the early stages of inflammation before substantial inhibitory levels of IFN-␥ are produced by activated lymphocytes. Moreover, PPAR-␥ ligand drugs may be more effective in reducing M␾ proinflammatory activity during atherosclerosis [13], a condition that does not appear to be mediated by T-cell-derived IFN-␥ production.

REFERENCES 1. Lemberger, T., Desvergne, B., Wahli, W. (1996) Peroxisome proliferatoractivated receptors: a nuclear receptor signaling pathway in lipid physiology. Annu. Rev. Cell Dev. Biol. 12, 335–363. 2. Schoonjans, K., Staels, B., Auwerx, J. (1996) The peroxisome proliferator activated receptors (PPARS) and their effects on lipid metabolism and adipocyte differentiation. Biochim. Biophys. Acta 1302, 93–109.

3. Colville-Nash, P. R., Qureshi, S. S., Willis, D., Willoughby, D. A. (1998) Inhibition of inducible nitric oxide synthase by peroxisome proliferatoractivated receptor agonists: correlation with induction of heme oxygenase 1. J. Immunol. 161, 978 –984. 4. Ricote, M., Li, A. C., Willson, T. M., Kelly, C. J., Glass, C. K. (1998) The peroxisome proliferator-activated receptor-gamma is a negative regulator of macrophage activation. Nature 391, 79 – 82. 5. Castrillo, A., Diaz-Guerra, M. J., Hortelano, S., Martin-Sanz, P., Bosca, L. (2000) Inhibition of IkappaB kinase and IkappaB phosphorylation by 15-deoxy-delta(12,14)-prostaglandin J(2) in activated murine macrophages. Mol. Cell. Biol. 20, 1692–1698. 6. Ricote, M., Huang, J. T., Welch, J. S., Glass, C. K. (1999) The peroxisome proliferator-activated receptor(PPARgamma) as a regulator of monocyte/ macrophage function. J. Leukoc. Biol. 66, 733–739. 7. Thieringer, R., Fenyk-Melody, J. E., Le Grand, C. B., Shelton, B. A., Detmers, P. A., Somers, E. P., Carbin, L., Moller, D. E., Wright, S. D., Berger, J. (2000) Activation of peroxisome proliferator-activated receptor gamma does not inhibit IL-6 or TNF-alpha responses of macrophages to lipopolysaccharide in vitro or in vivo. J. Immunol. 164, 1046 –1054. 8. Jiang, C., Ting, A. T., Seed, B. (1998) PPAR-gamma agonists inhibit production of monocyte inflammatory cytokines. Nature 391, 82– 86. 9. Denis, M. (1994) Human monocytes/macrophages: NO or no NO? J. Leukoc. Biol. 55, 682– 684. 10. Alleva, D. G., Kaser, S. B., Beller, D. I. (1998) Intrinsic defects in macrophage IL-12 production associated with immune dysfunction in the MRL/⫹⫹ and New Zealand Black/White F1 lupus-prone mice and the Leishmania major-susceptible BALB/c strain. J. Immunol. 161, 6878 – 6884. 11. Green, L. C., Wagner, D. A., Glogowski, J., Skipper, P. L., Wishnok, J. S., Tannenbaum, S. R. (1982) Analysis of nitrate, nitrite, and [15N]nitrate in biological fluids. Anal. Biochem. 126, 131–138. 12. Alleva, D. G., Burger, C. J., Elgert, K. D. (1994) Tumor-induced regulation of suppressor macrophage nitric oxide and TNF-alpha production. Role of tumor-derived IL-10, TGF-beta, and prostaglandin E2. J. Immunol. 153, 1674 –1686. 13. Ricote, M., Huang, J., Fajas, L., Li, A., Welch, J., Najib, J., Witztum, J. L., Auwerx, J., Palinski, W., Glass, C. K. (1998) Expression of the peroxisome proliferator-activated receptor gamma (PPARgamma) in human atherosclerosis and regulation in macrophages by colony stimulating factors and oxidized low density lipoprotein. Proc. Natl. Acad. Sci. USA 95, 7614 – 7619. 14. Boehm, U., Klamp, T., Groot, M., Howard, J. C. (1997) Cellular responses to interferon-gamma. Annu. Rev. Immunol. 15, 749 –795. 15. Guyton, K., Bond, R., Reilly, C., Gilkeson, G., Halushka, P., Cook, J. (2001) Differential effects of 15-deoxy-delta(12,14)-prostaglandin J2 and a peroxisome proliferator-activated receptor gamma agonist on macrophage activation. J. Leukoc. Biol. 69, 631– 638. 16. Moore, K. W., O’Garra, A., de Waal Malefyt, R., Vieira, P., Mosmann, T. R. (1993) Interleukin-10. Annu. Rev. Immunol. 11, 165–190. 17. Barnes, P. J. (1998) Anti-inflammatory actions of glucocorticoids: molecular mechanisms. Clin. Sci. (Lond.) 94, 557–572. 18. Na, S. Y., Kang, B. Y., Chung, S. W., Han, S. J., Ma, X., Trinchieri, G., Im, S. Y., Lee, J. W., Kim, T. S. (1999) Retinoids inhibit interleukin-12 production in macrophages through physical associations of retinoid X receptor and NFkappaB. J. Biol. Chem. 274, 7674 –7680. 19. D’Ambrosio, D., Cippitelli, M., Cocciolo, M. G., Mazzeo, D., Di Lucia, P., Lang, R., Sinigaglia, F., Panina-Bordignon, P. (1998) Inhibition of IL-12 production by 1,25-dihydroxyvitamin D3. Involvement of NF-kappaB downregulation in transcriptional repression of the p40 gene. J. Clin. Investig. 101, 252–262. 20. Bernardo, A., Levi, G., Minghetti, L. (2000) Role of the peroxisome proliferator-activated receptor-gamma (PPAR-gamma) and its natural ligand 15-deoxy-delta12, 14-prostaglandin J2 in the regulation of microglial functions. Eur. J. Neurosci. 12, 2215–2223. 21. Gately, M. K., Renzetti, L. M., Magram, J., Stern, A. S., Adorini, L., Gubler, U., Presky, D. H. (1998) The interleukin-12/interleukin-12-receptor system: role in normal and pathologic immune responses. Annu. Rev. Immunol. 16, 495–521. 22. Chung, S. W., Kang, B. Y., Kim, S. H., Pak, Y. K., Cho, D., Trinchieri, G., Kim, T. S. (2000) Oxidized low density lipoprotein inhibits interleukin-12 production in lipopolysaccharide-activated mouse macrophages via direct interactions between peroxisome proliferator-activated receptor-gamma and nuclear factor-kappa B. J. Biol. Chem. 275, 32681–32687.


685