Research Paper Mediators of Inflammation, 7, 13–18 (1998)

We have previously described inhibition of the synthesis of three acute-phase inflammatory cytokines in human and rat macrophages by acetate esters of rooperol, a dicatechol of plant origin. Analysing the mechanism of anticytokine activity of rooperol, we compared levels of TNFalpha, IL-1beta and IL-6 mRNAs in the human promonocytic U937 cell line pretreated with phorbol myristate acetate (PMA) and incubated with rooperol tetraacetate (RTA) alone or in combination with LPS (500 ng/ml). It was found that 10 microM RTA decreased the levels of cytokine mRNAs both in the presence and absence of LPS, suggesting pretranslational inhibition of cytokine synthesis. Electrophoretic mobility shift analysis (EMSA) showed that RTA may influence cytokine mRNA expression by decreasing the binding activity of transcription factors NF-kappaB and AP-1.


Introduction
Macrophages, ubiquitous in the body, have the potential to produce several cytokines involved in inflammation, such as TNFa , IL-1b and IL-6. [1][2][3][4][5] Induced synthesis of these cytokines involves multiple cell signalling pathways and can be regulated at the level of gene transcription, mRNA stability and translation, or cytokine precursor processing and secretion. 4,5 Overproduction of cytokines is believed to play a role in the pathogenesis of a wide range of disorders, such as septic shock, rheumatoid arthritis and asthma. [6][7][8][9] Compounds which suppress cytokine production may have beneficial effects in the treatment of such diseases. The list of anti-inflammatory drugs affecting cytokine production is growing steadily and involves glucocorticoids, some antioxidants, 6 bicyclic imidazoles, 10 piperazine derivatives, 11 tenidap 7,8 and many others. We have recently described inhibition of the TNFa , IL-1b and IL-6 (at the protein level) in human and rat alveolar macrophages, human blood monocytes and human promonocytic cell line U937 by a plant dicatechol, rooperol, and its acetate esters. 12 Rooperol derived from the plant Hy poxis ro o peri is a potent inhibitor of 5-lipoxygenase, but in concentrations up to 10 m M has no activity against either isoform of cyclooxygenase. 13 This suggests that the drug might be therapeutically useful in disorders when it is desirable to inhibit the production of leukotrienes but not prostaglandins, such as asthma and inflammatory bowel disease. Evidence is accumulating that pro-inflammatory cytokines also contribute to the pathogenesis of these as well as other inflammatory disorders, so that it would be desirable to inhibit their production as well. Although rooperol exhibit antioxidative properties, 13 and suppresses NO production by rat macrophages 12 and murine endothelial cells, 14 (overproduction of NO by inducible NO synthase may contribute to the pathogenesis of asthma 9 ), the mechanism of its anticytokine activity requires elucidation. The study now reported is an analysis of the mechanism by which rooperol may suppress the production of pro-inflammatory cytokines. We describe the effect of rooperol tetraacetate on the TNFa , IL-1b and IL-6 mRNAs, as well as DNAbinding activity of two transcription factors in U937 cells exposed to rooperol in the presence and absence of bacterial endotoxin. the concentration of 1 mg/ml. Rooperol ((E)-1,5-bis (39 ,49 -dihydroxyphenyl) pent-4-en-1-yne) tetraacetate (RTA) was dissolved in DMSO as described previously 12 to obtain a stock solution (100 mM), and was further diluted in 10% bovine serum albumin in PBS.
Cell culture U937 cells were grown at 37°C in 75-cm 2 flasks containing RPMI supplemented with 8% FBS and antibiotics under a humidified atmosphere of 95% air and 5% CO 2 . Every third day the medium was doubled, and every sixth day cells were centrifuged and suspended (1:4) in the fresh medium. The experiments were carried out on U937 cells differentiated into a macrophage-like cells in the presence of PMA at a concentration of 34 ng/ml for 48 h. During that period viable and differentiated cells adhered to the culture plates. Unattached cells were discarded and tested factors (in RPMI with 2% FBS) added after additional 24 h culturing of the adherent cells w ithout PMA. Cells were cultured in 12-well plates (Costar) (10 6 cells per well) for cytokine estimation by ELISA, in 90 mm Petri dishes (7 3 10 6 cells per dish) for mRNA isolation and in 60 mm Petri dishes (3 3 10 6 cells per dish) for isolation of nuclear proteins.

Cytokine assay
In order to avoid discrepancies related to cytokine secretion, total (medium + cells) TNFa , IL-1b and IL-6 proteins were determined using commercially available ELISA kits (Genzyme). After 24 h of culture w ith the tested factors, U937 cells were scraped from the wells directly into the medium, subjected to three cycles of freezing and thawing, and centrifuged (10 000 g for 3 min) before cytokine determination.

Isolation of RNA and Northern blot analysis
Total U937 RNA was extracted using the phenol extraction method and LiCl precipitation as described by Rose-John et a l. 15 from the cells treated w ith 10 m M RTA for 1 h followed by 2 or 18 h incubation with 500 ng/ml LPS up to the moment of RNA extraction. RNA samples (5 m g) were separated by electrophoresis in 1% agarose gel under denaturing conditions. 16 RNA was then transferred to Hybond-N membranes (Amersham, Arlington, IL) according to the manufacturer's instructions. Hybridization w ith 32 P-labelled probes was performed overnight at 65°C in a mixture containing 1.0 M NaCl, 1% SDS and 10% dextran sulphate. Specific mRNAs were detected using the following probes : 1.1-kb PstI-PstI restriction fragment of human TNFa cDNA (ATCC, Rockville, MA), 1.2-kb EcoRI-HindIII fragment of human IL-6 cDNA (a generous gift of Dr T. Kishimoto), pGEMI containing a 570 bp Ss tI-PvuII cDNA fragment of human IL-1b (kindly provided by Immunex TM , Seattle, WA), a cDNA probe specific for the human glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (ATCC, GenBank/EMBL: M17851), and a 1.5-kb Ps tI-PstI fragment of the pSP64 coding for rat b -actin (kindly provided by Dr J. Sipe, Boston, USA). The probes were labelled with Megaprime Labeling Kit (Amersham). As the reference, 18 S and 28 S rRNA were visualized in UV light after ethidium bromide staining, or 18 S rRNA determined by hybridization with the EcoRI-Sa lI fragment of the plasmid containing human 18 S rRNA cDNA (kindly provided by N. Bhownick, Athens, GA). Non-specifically bound radioactivity was removed by washing and the blots were subjected to autoradiography at -70 o C using intensifying screens. The relative intensity of the bands was evaluated by densitometry using a computer imaging (MCID) system (Imaging Research Inc., Canada).

Nuclear protein extraction
Nuclear proteins were obtained by the mini extraction procedure described by Suzuki et al. 17 PMAprimed U937 cells were cultured for 1 h with 10 m M RTA followed by exposure to 500 ng/ml of LPS up to the moment of cell collection. Cells were collected with a rubber policeman, washed tw ice with cold PBS and centrifuged for 5 min at 400 3 g . The cells, suspended in buffer containing 10 mM HEPES (pH 7.8), 10 mM KCl, 2 mM MgCl 2 , 1 mM DTT, 0.1 mM EDTA, and 0.1 mM PMSF, were incubated on ice for 15 min. Nonidet P-40 was added and samples were centrifuged for 30 s at 14 000 r.p.m. Pelleted nuclei were resuspended in buffer containing 50 mM HEPES, 50 mM KCl, 300 mM NaCl, 0.1 mM EDTA, 1 mM DTT and 0.1 mM PMSF, centrifuged for 5 min at 14 000 r.p.m. at 4°C and the supernatant was frozen in 10% glycerol. The protein concentration was determined by Lowry's method.

EMSA
The following double-stranded oligonucleotides were used for DNA electrophoretic mobility shift assay: the oligonucleotide that contained two binding sites for NF-k B from c-myc oncogene corresponding to bp -1101-1081 (5'-AAGTCCGGGTTTTCCCCAACC-3') 18 and AP-1 oligonucleotide: 5'-TCGACTAGTATGAGT-CAGCCG-3'. 19 These oligonucleotide probes were either end-labelled with [g -32 P]ATP and T4 polynucleotide kinase (NF-k B) or labelled w ith [a -32 P]dCTP and Klenow polymerase (AP-1) (Megaprime Labeling Kit). For gel shift analysis 5 m g of nuclear proteins were incubated with 0.5 ng (c.10 5 cpm) of labelled oligonucleotides. As competitors, a 100-fold excess of the unlabeled oligonucleotide was added to the binding reaction.

Rooperol tetraacetate inhibition of cytokine mRNAs in PMA-primed U937 cells cultured in the presence and absence of LPS
We found previously that LPS-induced synthesis of TNFa , IL-1b and IL-6 proteins was inhibited in human and rat alveolar macrophages, human blood monocytes/macrophages and U937 cells ex posed to rooperol and its acetate esters. 12 This has been confirmed and extended by the data shown in Table 1. The inhibitory effect of rooperol is less pronounced than in blood monocytes 12 because 'control' U937 cells are in fact significantly stimulated by PMA.
Since cytokine gene ex pression is often regulated at the pretranslational stages 1,4,5 we determined the levels of specific mRNAs in U937 cells primed w ith PMA and stimulated with LPS. To compare early and late responses, measurements of the cytokine mRNAs were carried out at 2 and 18 h after exposure to LPS. Figure 1A shows that control cells express TNFa mRNA, due to stimulation with PMA, whereas unprimed and unattached cells do not produce cytokines even after stimulation with LPS ( Fig. 1B and Ref. 20). As expected, stimulation w ith LPS increased the abundance of TNF mRNA, especially after 2 h. We also found that pretreatment of cells with RTA decreased levels of TNFa mRNA both in control and LPS-stimulated cells ( Fig. 1A and B). Similar pictures were obtained in three consecutive experiments. Exposure of U937 cells to RTA also decreased the amounts of IL-1b and IL-6 mRNAs, but the effects were more pronounced after 18 h culture with LPS or/and RTA (Figs 2 and 3). Six independent experiments were carried out and Figs 2 and 3 show typical results. The inhibitory activity of RTA appears to be rather specific for cytokine mRNAs since the amount of 18 S RNA was unaffected both in the presence and absence of LPS. Moreover, RTA at 10 m M concentration had no significant effect on the expression of Inhibitio n of cy to kine s ynthesis in U937 cells  b -actin (a typical structural protein of the cell) and GAPDH (one of the glycolytic pathway enzymes) (Fig.  4). These results suggest that RTA inhibits the synthesis of inflammatory cytokines at the transcriptional level. As shown recently by several authors, translation of cytokine mRNA requires an additional signal provided by stress-responsive kinases and thus in certain cases a dissociation of transcription from translation may occur. 4,21 Our results presented in Table 1 indicate that changes in the total cytokine contents (cell + media) elicited by LPS and RTA in U937 cells roughly correspond to changes in the amount of the specific mRNAs (Figs 1-3), which is consistent with a pretranslational site of action of RTA.

Inhibition of binding activity of transcription factors NF-k B and AP-1 by RTA in PMA-primed U937 cells cultured in the presence and absence of LPS
Cytokine gene promoters contain multiple regulatory elements recognized by several transcription factors, the most important being NF-k B, AP-1 and C/EBP. 4,5,22 It is know n that in U937 cells both PMA and LPS activate NF-k B. 23 -25 Figure 5A shows that RTA decreased NF-k B binding activity in PMA-primed (control) and LPS-stimulated U937 cells 30 min. follow ing exposure to LPS. At shorter (15 min) (Fig.  5B) or longer than 4 h incubation times the effect of RTA was less pronounced (data not show n). The inhibitory effect of RTA was comparable in control (PMA-primed) and LPS-stimulated cells (Fig. 5A). When nuclear extracts from U937 cells were analysed by EMSA (three independent experiments) with the AP-1 probe inhibition of binding by RTA was observed after 30 min and after 1 h exposure to LPS (Fig. 6).

Discussion
The U937 cell line, primed toward macrophage differentiation w ith PMA, is a useful model for studying the regulation of cytokine synthesis. 17,24,25 Although PMA affects protein kinase C and induces synthesis of cytokines (high 'control' values), thereby reducing the relative response to LPS, the U937 cells represent a uniform population that can be easily cultured in amounts sufficient for analysis of RNA and nuclear proteins. Using this model we demonstrated that RTA in 10 m M concentration, which does not show detectable cytotoxic effects such as inhibition of total protein synthesis or leakage of lactate dehydrogenase, 12 decreases abundance of mRNAs coding for TNFa , IL-1b and IL-6 both in control (PMAprimed) and LPS-stimulated U937 cells (Figs 1-3). From comparison of scanning values it is evident that the 'inhibitory index' (control vs. RTA, and LPS vs. LPS+RTA) is very similar suggesting that RTA acts on both PMA-induced and LPS-induced expression of cytokine genes. This may indicate a common target of RTA on the intracellular signalling pathway but the problem requires further studies. On the other hand, mRNAs coding for other cellular proteins (GAPDH, b -actin) were not detectably affected by RTA. Thus, it can be concluded that RTA inhibits synthesis of the inflammatory cytokines at the transcriptional level, in contrast to other inhibitors such as bicyclic imidazoles 22 or piperazine derivatives 11 which affect translation stages, or tenidap, which acts indirectly by inhibiting the cyclooxygenase pathway. 26 The observed decreases in the amounts of cytokine mRNAs in the presence of RTA could result from corresponding changes in transcription rates or mRNA stability. Cytokine mRNAs contain adenylate/uridylate-rich elements which are held responsible for the short half life of their messages. 27 To exclude the possibility that RTA enhances the rate of cytokine mRNA degradation additional experiments with actinomycin D are required.
On the other hand, the results of our experiments point to the interaction of rooperol with transcription factors, known to be required for cytokine gene transcription: NF-k B and AP-1. The activation of NF-k B  6. Inhibition of AP-1 binding to DNA sequences following 60 min RTA treatment or/and 60 min exposure to LPS. Electrophoretic mobility shift assay of nuclear extracts isolated from control, RTA-(10m M), LPS-(500 ng/ml) or LPS+RTA-treated U937 cells primed with PMA. Cells were incubated for 60min with RTA, followed by an additional 60min with LPS and then nuclear extracts were isolated and processed as described in Materials and Methods. Densitometric scans are expressed in arbitrary units.
involves an oxidation step w ith possible involvement of reactive oxygen species 17,28 and several antioxidants showing anti-inflammatory properties act at this stage. 6,28 As show n by van der Merwe et a l. 13 rooperol belongs to the group of biologically active antioxidants, and thus by interfering with NF-k B activation may inhibit synthesis of inflammatory cytokines. Redox regulation of the AP-1 transcription factor complex is also well documented in both U937 and HeLa cells 24,29 and may explain the RTA-elicited reduction of AP-1 binding observed in our experiments (Fig. 6). Since EMSA showed only a transient decrease in the binding capacity of NF-k B other possible sites of anticytokine action of RTA may be identified in future. We can conclude, however, that RTA influences abundance of mRNAs coding for TNFa , IL-1b and IL-6 in human macrophage-like cells and affects binding of transcription factors implicated in cytokine gene expression. This provides a better understanding of the activity of RTA as potential anti-inflammatory and anticytokine drug.