Caffeic acid phenethyl ester (CAPE) prevents inflammatory stress in organotypic hippocampal slice cultures

Caffeic acid phenethyl ester (CAPE) prevents inflammatory stress in organotypic hippocampal slice cultures

Molecular Brain Research 115 (2003) 111–120 / locate / molbrainres Research report Caffeic acid phenethyl ester (CAPE) prevents inf...

564KB Sizes 0 Downloads 16 Views

Molecular Brain Research 115 (2003) 111–120 / locate / molbrainres

Research report

Caffeic acid phenethyl ester (CAPE) prevents inflammatory stress in organotypic hippocampal slice cultures ´ ´ de Bock b , Gerard ´ Pascale Montpied a , *, Frederic Rondouin b , Gilles Niel c , Laurence Briant d , ¨ Bockaert b Anne-Sophie Courseau b , Mireille Lerner-Natoli b , Joel a

Faculte´ de Pharmacie, CNRS-UMR 5094, 15 Avenue Charles Flahault, 34060 Montpellier Cedex 2, France b CNRS, UPR 2580, CCIPE, 34094 Montpellier Cedex 5, France c ENSCM, 34296 Montpellier Cedex 5, France d CNRS, UMR 5121, Institut de Biologie, 34060 Montpellier Cedex, France Accepted 30 April 2003

Abstract Caffeic acid phenethyl ester (CAPE) is an antioxidant component of propolis, a natural product secreted by honeybee. Recent literature shows that CAPE inhibits nuclear factor k B (NFkB) activation in cell lines. Since NFkB was shown to be a crucial factor in neuroinflammation and to be associated with some neuropathologies, CAPE might reduce these disorders in brain too and have therapeutic applications. To test this hypothesis we used a model of endotoxic insult (interferon-g, followed by lipopolysaccharide) on rat organotypic hippocampal cultures. Cerebral inflammatory responses were strongly inhibited by CAPE (100 mM): reductions of NFkB nuclear activity, tumor necrosis factor a and nitric oxide productions were observed. At the dose of maximal effects (100 mM), an increase of cAMP-responsive element binding protein (CREB) activity, which anti-inflammatory role is well known, was seen. We compared CAPE effects with those of other drugs: anti-inflammatory as acetyl-salicylate and dexamethasone (glucocorticoid), antioxidant as pyrrolidine dithiocarbamate, or selective permeant inhibitor of NFkB as SN 50 peptide. These studies lead us to conclude that CAPE presents an interesting and original neuropharmacological profile compared to these drugs and might be helpful in the prevention of neurotoxic events due to excessive inflammatory reaction in brain. CAPE interferes with several effectors of neuroinflammation that might have complementary and synergic effects and allows a rather durable control since an acute treatment at the time of endotoxin exposure allows to control inflammatory factors for over 48 h.  2003 Elsevier B.V. All rights reserved. Theme: Disorders of the nervous system Topic: Neurotoxicity Keywords: LPS; Gel shift; NFkB; CREB; NO; TNFa

1. Introduction The caffeic acid phenethyl ester (CAPE), a phenolic compound extracted from beehive product propolis, was found to have many properties: antioxidant, immunomodulatory, anti-inflammatory and anticancerous [21,46,40,48,49]. CAPE molecular mechanisms are not fully elucidated and some have been described at the *Corresponding author. Tel.: 133-4-6714-9388; fax: 133-4-67149387. E-mail address: [email protected] (P. Montpied). 0169-328X / 03 / $ – see front matter  2003 Elsevier B.V. All rights reserved. doi:10.1016 / S0169-328X(03)00178-5

peripheral level or on cell lines. CAPE effects on nuclear factor k B (NFkB) transduction signal [37] suggest that it could also be active on brain inflammatory events observed in several neuropathologies or in response to infectious diseases. Increasing interest has been focused on antiinflammatory or immunosuppressive pharmacology that could prevent or alleviate the consequences of excessive inflammation in the central nervous system [20,15,32,45,4]. Lipopolysaccharide (LPS), a bacterial endotoxin, had been used in several models of neuroinflammation and was shown to reproduce the cascade of reactions observed following an immune alert, and the phenotypic alterations


P. Montpied et al. / Molecular Brain Research 115 (2003) 111–120

described during various neuropathologies [25,24,43,36,29]. Interacting with the CD14 receptors (membrane bound or soluble forms) expressed in microglia and on the parenchymal structures of the circumventricular regions, it produces down stream to this initial event activations of NFkB, secretions of tumor necrosis factor a (TNF-a) and releases of nitric oxide (NO) [36,29]. It is believed that, released in the microdomain where CD14 is anchored, LPS molecules containing a lipid moiety move by lateral diffusion and activate by steric stress several receptor complexes which expression depends on the cell and the state of inflammation [52]. The later has been shown for example to be altered by the exposure to interferon-g (IFN-g) that increases CD14 and Toll like receptor4 and is responsible for numerous transcriptional regulation involved in the neuroinflammatory process. Following an endotoxic shock, in particular, the central inflammation is increased by infiltration of leukocytes that release IFN-g apparently involved in the mediation of coinflammatory signals between microglia and leucocytes [23,30]. The primary steps of LPS-mediated inflammatory signals have been shown to involve various molecules (CD14, HSP 70, CXCR4, GDF5, TLR4); the receptor complex built at this time lead to the activation of NFkB, then the inductions of TNF-a and of the inductible NO synthase (NOS-2) genes. This is believed to be pivotal in the amplification and the persistence of the cascade of cellular and molecular brain immune reactions, since each of the three factors can activate the synthesis of the others. These loops of autoinduction promote phenomenona such as neuronal death and consecutively irreversible alterations in brain function [36,11,27,9]. Extensively studied in the progression of these inflammatory events in the central nervous system, NFkB activation emerged as an appropriate site for the design of a new neuropharmacological approach since, through its blockage, it should be possible to shunt or decrease the induction of all proinflammatory genes containing the regulatory consensus for this nuclear factor [7,17,18]. We have studied, in organotypic hippocampal cultures (OHC) that preserve the cytoarchitecture and cellular dynamics of brain tissue, the potential effect of CAPE against the LPS-mediated neuroinflammatory events. To test CAPE in conditions of maximal inflammation, cultures were pretreated with IFN-g that was shown to enhance the inflammatory responses in our model as it does in vivo (see above Ref. [23]). The inflammatory dynamics were evaluated first at the nuclear level: NFkB activation was measured and in parallel CREB activation was evaluated, as a control and as a second pathway of interest, involved in the down regulation of neuroinflammation [32] and in neuroprotective signals [34]. Secondary to the nuclear transcription factor alterations, the production and the release of TNF-a and NO, two major effectors of the inflammation, were measured. Finally, we performed comparisons of CAPE effects with those of selected com-

pounds reported as modulators of different inflammation steps.

2. Material and methods Experiments were carried out in accordance with the animal welfare guidelines of the Institut National de la ´ Sante´ et de la Recherche Medicale and approved by the ˆ (authorization no. ` de l’Agriculture et de la Foret Ministere 34-103).

2.1. Drugs LPS from Salmonella typhimurium, acetyl salicylate (AS), pyrrolidine dithiocarbamate (PDTC), dexamethasone (Dexa) were obtained from Sigma (St. Quentin Fallavier, France). Recombinant rat IFN-g was purchased from R&D Systems (Abingdon, UK). CAPE and SN 50 peptide [33] were synthesized in our laboratory. Dilutions were made, in water for AS and PDTC, in horse serum used in culture medium for SN 50 peptide, and in 50% ethanol for CAPE (stock solution at 4 mM) and for Dexa (stock solution at 1 mM). In the comparative study between CAPE and the above drugs, the concentrations used for the treatments were chosen in order to obtain maximal effects and were deduced from previous studies [19,53,51,44].

2.2. Culture technique Organotypic hippocampal cultures were prepared according to the membrane technique [50], with a protocol previously published [10]. Briefly, the hippocampi of 5–7day-old Wistar rats (Janvier, Le Genest St Isle, France) were dissected and sectioned transversely at 320 mm. Ten slices were transferred on a 30-mm diameter porous membrane insert (Millicell-CM, Millipore, St Quentin en Yvelines, France) kept in a 100-mm diameter Petri dish filled with 5 ml of culture medium (50% MEM with 25 mM Hepes, 25% heat-inactivated horse serum, 25% HBSS, 6.5 mg / ml glucose, 2 mM glutamine, 25 U / ml penicillin and 25 mg / ml streptomycin). Cultures were maintained in a humidified 5% CO 2 incubator at 35 8C for a period of 2 weeks before the experiments.

2.3. Treatments After a culture period of 2 weeks, each insert carrying 10 organotypic slices was transferred to an independent well in a six-well multidish filled with 1 ml of culture medium. Any given compound used thereafter was added in this volume and below the membrane in order to obtain the desired final concentration and to avoid direct contact of the initial solution with slices. All experiments were effected in presence of horse serum.

P. Montpied et al. / Molecular Brain Research 115 (2003) 111–120

2.4. Protocol for inflammatory stress ( IS) OHC were pretreated during 24 h with IFN-g (500 U / ml or 50 ng / ml) and then exposed to LPS (200 ng / ml) in presence of IFN-g. All drugs, CAPE (4 to 100 mM), SN 50 (25 mg / ml), Dexa (5 mM), PDTC (100 mM), AS (100 mM), were applied 1 / 2 h before and during LPS addition. Four hours after the beginning of LPS application, slices were collected from each well in order to prepare nuclear protein for the measurements of NFkB and CREB activities and the medium was sampled for TNF-a bioactivity evaluation. This time point, corresponding at the maximal activation of NFkB nuclear activities, was chosen in preliminary kinetic investigations. Some wells were conserved during 48 h for NOS-2 immunohistochemistry and nitrite concentrations measurement in the supernatant.

2.5. Evaluation of NFk B and CREB nuclear activity by electrophoretic mobility-shift assay ( EMSA) Nuclear extracts were prepared as previously described [5] with some modifications. Briefly, the 10 hippocampal slices present in a well were collected in ice cold PBS 13. After a quick spin and complete removal of PBS, 500 ml of cold buffer A (containing 10 mM KCl, 2 mM MgCl 2 , 0.1 mM EDTA, 1 mM DTT, 0.1 mM phenylmethylsufonyl fluoride, 4 mg / ml aprotinin and 10 mM Hepes, pH 7.8) was added to the tissue. Tissue dissociation was completed by gentle homogenization using a manual Eppendorf homogenizer. Following incubation for 15 min on ice, 65 ml of 10% Nonidet P-40 were added and homogenized by vortexing with the cells suspensions. Microcentrifugation at 12 000 rpm, 4 8C, for 30 s allows to pellet the nuclear fraction, resuspended then in 200 ml of buffer B (containing 50 mM KCl, 300 mM NaCl, 0.1 mM EDTA, 1 mM DTT, 0.1 mM phenylmethylsulfonyl fluoride, 4 mg / ml aprotinin, 10% glycerol, and 50 mM Hepes, pH 7.8). Incubation in a 4 8C cold chamber on a shaker for 30 min allows the extraction of nuclear proteins. Nuclear extracts were cleaned by centrifugation at 12 000 rpm, 4 8C, for 5 min, and the supernatants and stored in aliquots at 280 8C until used. The NFkB and CREB mobility shift assay was performed using 5 to 7 mg of nuclear proteins, 1310 5 cpm of 32 P-labeled probe consisting in a double stranded oligonucleotide containing either the consensus NFkB sequence and some flanking nucleotides (59GCTGGGGACTTTCCAGGAGGCGT-39) as previously described [5] or the consensus of the cAMP responsive element (CREB) (59-ACGCTGCTGCGTCAGCAAAT-39) previously used [28]. Following the incubation of the proteins with the labeled consensus sequence at room temperature for 20 min in buffer C (containing 100 mM KCl, 1 mM DTT, 1 mM ZnSO 4 , 20% glycerol, 0.01% Nonidet P-40, and 50 mM Hepes, pH 7.8) supplemented with BSA, tRNA, and poly(dI-dC) in a final volume of 20 ml, the reaction was stopped with dying solution, loaded on


a 10% polyacrylamide gel and run at 120 V (distance of migration510–12 cm). Competition experiments were performed by adding 250- or 500-fold excess of unlabeled probe for 30 min prior to the addition of the radiolabeled probe in the reaction mixture. The absence of labeled shifted band was complete with a 250-fold excess of cold probe. Supershift assays were also performed in order to control for the specificity of the nuclear factor binding. Polyclonal antibody against NFkB p65 (Santa Cruz, Biotech, CA) was used and the absence of labeled shifted band was obtained with 1 / 500 dilutions of the antibody (not shown). The optic density (OD) of the shifted bands obtained were quantified in parallel for the different compounds and the different conditions compared. Densitometric measurements were obtained following calibration employing Kodak calibration step tablet 809ST601 as a standard, using Image 1.38 software (W/ Rasband, NIMH). Statistical comparisons of the measurements were made using analysis of variance followed by Fisher posthoc test.

2.6. TNF-a bioassay Biological activity of TNF-a secreted in the medium was evaluated by a cytotoxic assay performed on the L929 murine fibroblast cell lineage using a method previously described [10]. Briefly, 6310 4 L929 cells in 100 ml of medium (RPMI 1640, 10% foetal calf serum, 2 mM glutamine, 20 mg / ml gentamycin) were seeded in 96 microwells for 24 h. Activity of serial dilutions of supernatants was assayed on culture wells, in quadruplicate, and compared with the activity of standard doses of recombinant human TNF-a (from National Institute for Biological Standards and Controls, Potters Bar, UK, ref. 88 / 532). Viability of the remaining cells was estimated by optical density with a colorimetric cell proliferation assay using ` MTS (Promega, Charbonniere, France). Statistical comparisons of the measurements were made using Student’ t-test.

2.7. NOS-2 immunohistochemistry OHC were fixed at 4 8C for 1 h in 4% paraformaldehyde in 0.1 M phosphate buffer. Cryoprotected by 10% saccharose, slices were cut into 12-mm sections using a cryostat (Microm, Francheville, France) and mounted on gelatincoated slides. Briefly, mounted sections were preincubated, for 1 h at room temperature, in PBS-0.25% Triton containing 20% horse serum, then incubated O / N at 4 8C in the same buffer containing only 2% horse serum and a rabbit polyclonal antibody raised against NOS-2 diluted 1 / 200 (M-19 Santa Cruz). After several quick washes, the Cy姠3conjugated antibody raised against rabbit IgGs (1 / 400), was added for the secondary incubation at 4 8C for 3 h.


P. Montpied et al. / Molecular Brain Research 115 (2003) 111–120

The slides were then rinsed and coverslipped with Mowiol (Calbiochem, Meudon, France). Sections were observed using a Leitz DMRB microscope (Leica, Westlar, Germany) equipped for light transmission and epifluorescence. Images were captured with a cooled color camera CCD, with a 139231040 resolution (Cool Snap, Roper scientific, Evry, France) and transferred to Adobe Photoshop  (version 5.02) for image processing.

built using sodium nitrite dilutions in culture medium. Statistical comparisons of the measurements were made using Student’ t-test.

3. Results

3.1. Dose–response effects of CAPE on NFk B and CREB nuclear activities following IS ( Fig. 1)

2.8. Nitrite measurement NO release was determined by measuring nitrites, a stable oxidation breakdown product of NO [13]. Nitrites concentration in the medium was estimated by a microplate colorimetric assay method by mixing 100 ml of sample with 100 ml of Griess reagent (Molecular Probes, Eugene, OR). After 1 h of incubation at room temperature, the absorbance at 570 nm was determined on a microplate reader. Nitrite level was quantified on a standard curve

Nuclear protein preparations obtained from OHC following IS or IS plus treatment with CAPE, were incubated with NFkB and CREB 32 P-labeled sequences. Comparison of the OD measurements of the labeled shifted bands obtained in absence or in presence of various concentrations of CAPE added with LPS showed that CAPE was ineffective on NFkB activation at 4 mM. A concentration of 40 mM blocked 80% (P,0.05) of the nuclear factor translocation. No significant increase of this effect was

Fig. 1. CAPE effects on pro- and anti-inflammatory nuclear factors activation in OHC 4 h following inflammatory stress induction: Left: NFkB nuclear concentration. Right: CREB nuclear concentration. The comparisons between groups are performed using analysis of variance followed by Fisher’s post-hoc test, **P,0.05. Optical density (O.D.) data are mean6S.E.M. from a minimum of three independent determinations. Left panel, **indicate significance between a given group of treatment and the IFN-g plus LPS stimulated OHC. Right panel, letters are added to indicate significance between several groups of treatment, (a) CAPE 4 mM versus stimulated OHC, (b) CAPE 100 mM versus stimulated OHC, (c) CAPE 40 mM versus CAPE 4 mM, (d) CAPE 100 mM versus CAPE 40 mM.

P. Montpied et al. / Molecular Brain Research 115 (2003) 111–120

observed at higher concentrations. Measured in parallel in each condition, CREB nuclear activity was found to progressively increase as CAPE concentrations reach antiinflammatory efficacy. This significant dose-dependent effect of CAPE consists in a variation of CREB binding activity from a low level compared to basal level when only 4 mM of CAPE were added to a level representing a 209% (P,0.05) increase when a dose of 100 mM in CAPE was applied.

3.2. Comparison between CAPE and other antiinflammatory molecules on NFk B and CREB nuclear activations following IS ( Fig. 2) Nuclear protein preparations were obtained in the same conditions than in Section 1. The concentrations, used for the treatment with the various inhibitory drugs, were chosen according to Section 2. In OHC exposed to IS and co-treated with 25 mg / ml of SN50, which sequence comprised a selective domain blocking the nuclear localization sequence of NFkB, a 52% (P,0.01) inhibition of NFkB translocation to the nucleus was obtained whereas no significant effect on CREB translocation (229%; P5


n.s.) was seen in parallel. Dexamethasone, a ligand of the glucocorticoid receptors, used at a dose of 5 mM inhibited NFkB nuclear translocation by 40% (P,0.05) in the OHC and did not modify CREB translocation. Pyrrolidine dithiocarbamate, known for its potent antioxidant effect, did not change NFkB nuclear activity in OHC but increased CREB nuclear activity by 270% (P,0.01), an effect clearly superior to CAPE effect. Acetyl salicylate at 100 mM, already described as an inhibitor of NFkB activation in brain, suppressed 81% (P,0.01) of NFkB activation in OHC undergoing IS, as did 100 mM of CAPE.

3.3. Comparison of CAPE effects on TNF-a and nitrite productions with those of several anti-inflammatory drugs ( Fig. 3) All the drugs, except for PDTC, inhibited NFkB activation in OHC (see above) and all of them, but not PDTC, significantly decreased TNF-a production with various efficacy: the reduction in TNF-a production were about 80% for SN50 and AS, 85% for Dexa, and 50% for CAPE. In contrast, NO production was not responsive to all the

Fig. 2. Comparison between CAPE and different anti-inflammatory drugs on NFkB and CREB activations in OHC 4 h following inflammatory stress induction: Drugs were applied at the following concentrations: SN 50 (25 mg / ml), Dexa (5 mM), PDTC (100 mM), AS (100 mM), CAPE (100 mM). Significance between a given group of treatment and the IFN-g plus LPS stimulated OHC are represented: **P,0.05, Student’ t-test. Optical density (O.D.) data are mean6S.E.M. from three to six independent determinations.


P. Montpied et al. / Molecular Brain Research 115 (2003) 111–120

inhibition of the NFkB inflammatory pathway, decreased NO production by, respectively: 52, 75, 72%.

3.4. Dose–response effect of CAPE on NOS-2 expression and NO release following IS ( Fig. 4) To control that CAPE-mediated decrease in nitrite was effectively due to an effect on the NOS-2 gene expression we performed immunohistochemistry. The IS-mediated induction of NOS-2 expression (which synthesizes NO and consequently nitrites), provided a strong immunolabeling that remained high for at least 48 h. This labeling of NOS-2 was almost abolished by a 100-mM CAPE treatment, for at least 48 h. In parallel, the dose response showed that the significant effect of CAPE on nitrites production is reached at 40 mM, with a maximal decrease of about 75% in levels measured in the medium.

4. Discussion

Fig. 3. Comparison between the suppressive effects obtained with CAPE and different inhibitory drugs on nitrites and TNF-a productions in OHC supernatant during inflammatory stress: Drugs were applied at the concentrations indicated in Fig. 2. Data are mean6S.E.M. from a minimum of five independent determinations. The significance was estimated using Student’ t-test: **P,0.05.

drugs inhibiting NFkB activation. The highly selective inhibitors of NFkB nuclear activity in our model, SN50 and AS, had no effect on NO production whereas Dexa, PDTC and CAPE, that are not only acting through the

Our results demonstrate that CAPE, a compound previously characterized as having anti-inflammatory and immunomodulatory properties on various cell lines and peripheral tissues [21,46,37,40,48,49], is a potent inhibitor of neuroinflammatory responses obtained here in organotypic hippocampal cultures pretreated with IFN-g and exposed to LPS. Aimed to respond to a cellular stress, neuroinflammation can become excessive. Described in various neuropathologies and following endotoxic shock, various molecules are responsible of the initial cellular stress. However it was shown that when infiltration of leucocytes through

Fig. 4. CAPE effects on NOS-2 expression and nitrite oxide production 48 h following an inflammatory stress induction: Left panel: NOS-2 immunoreactivity. Right panel: Nitrite release in culture medium. Data are mean6S.E.M. from minimum six independent determinations (**P,0.05, Student’ t-test, the significance was estimated by comparison to stimulated OHC).

P. Montpied et al. / Molecular Brain Research 115 (2003) 111–120

the BBB occurs, following endotoxic shock for example, a cytokine: IFN-g is particularly important in the orchestration of co-inflammatory mechanisms through intercellular messengers exchanged between leucocytes and microglia [23]. We have developed an in vitro model using a pretreatment with IFN-g followed by LPS exposure in order to test the anti-inflammatory efficacy of CAPE. Compared to in vivo models, organotypic cultures allowed an easy monitoring of the neuroinflammatory factors and a good reproducibility of high levels of inflammation. In these conditions of neuroinflammation, we found that CAPE reduced the NFkB nuclear translocation, the TNF-a secretion, and the nitrite production in brain tissue. The efficiency of CAPE on NFkB activation in brain is quite comparable to what was reported in histiocytic cell line stimulated with TNF-a [37]. Similarly, CAPE effect on TNF-a production is consistent with its effect described in colonic macrophage and epithelial cell lines as well as in vivo following LPS-induced colitis [16,22]. These last authors showed that the suppression of LPS-mediated NFkB activation was not followed by a similar level of TNF-a gene suppression; similarly in OHC, we found that CAPE-mediated effect on LPS-induced TNF-a secretion was not as profound as on NFkB nuclear levels. In comparison to that, the CAPE-mediated decrease in nitrite production seems proportional to CAPE-mediated reduction of NFkB nuclear translocation. It has been shown that CAPE effect on NO production was due to the decrease of NFkB binding to NOS-2 gene promoter [49]. Consistent with this effect of CAPE at the transcriptional level, we found a direct effect of CAPE on LPS-induced NOS-2 protein level of expression. The NOS-2 concentration visualized by immunohistochemistry was decreased in OHC, and remained at the lowest level 48 h after LPS and CAPE incubation (Fig. 4). This discrepancy between the sensitivity of TNF-a and NO productions to CAPE treatment may lay in the differential influence of IFN-g pretreatment on TNF-a and NOS-2 genes. In preliminary experiments aimed to set maximal conditions of neuroinflammation, we observed that IFN-g pretreatment, was not only indispensable to trigger the expression of the major histocompatibility complexes I and II in microglia (verified by immunohistochemistry, not shown), but was in addition increasing TNF-a secretion up to levels hundred times greater compared to a simple exposure to LPS. In contrast, NO production was only increased by 4-fold (not shown). This is consistent with what is known about IFN-g induction of STAT-1 and IRF-1, two nuclear factors acting on several genes carrying the GAS and / or IRF regulatory sequences [23,14,31]. In addition, several works have shown that IFN-g potentialized the LPS-mediated cellular inflammatory stress by increasing the expression of the toll-like receptor 4 essential in the complex cascade relaying LPS binding to the CD14 receptor. The complexities of the NOS-2 promoter expression might be at the origin of a lesser sensitivity to IFN-g [42,38,6] and our


results suggest that, in our model of neuroinflammatory stress, IFN-g pretreatment operates already an alteration of the regulatory pathways involved in the control of TNF-a and NOS encoding genes. NFkB consensus regulatory sequences are found in several genes involved in inflammation (i.e.: NFkB itself, TNF-a, NOS-2). The absence of parallel between the expression levels of these genes is consistent with the fact that the affinity and the influence of nuclear factors related to the Rel / NFkB family is not similar for all the kB consensus which sequence and / or position varied in the different gene promoters. We have observed in our set up experiments that prior incubations with a p65 polyclonal antibody abolished the detection of the double band corresponding to nuclear factor binding to the labeled k consensus, therefore we know that our evaluation concerned only the abundant dimers of NFkB massively induced in our model of neuroinflammation and those are composed of p65 (named also RelA) and most probably of p50 or p52 subunits [9,35]. It is possible that other kB binding factors are expressed in OHC but not detected in our conditions because of there low levels of expression or low affinity for the HIV-k consensus. Our data on CAPE-mediated inhibition of NFkB activation, is limited to the classical pathways and results probably from a cytoplasm action of CAPE on IkB and possibly to a direct interference of CAPE with the binding of the activated dimers to the consensus nuclear sequences [37] although it is yet unknown if in situ CAPE is diffusing up to the nucleus. It is, however, possible that the new k binding factor, described by [35], probably expressed in OHC, might modulate differentially TNF-a and NOS-2 genes, acting as a competitor for their two k sites. Finally, consistent with its wide spectrum of anti-inflammatory actions, we found that CAPE increased CREB nuclear activity at high concentration (100 mM). This effect of CAPE on CREB nuclear factor has not been documented. It is well known that this nuclear factor acts on several genes involved in anti-inflammatory processes and in particular that the CRE consensus sequence is present in the complex promoter of TNF-a and NOS-2 genes [32]. As a consequence, CAPE effect on TNF-a and NOS-2 gene expression results at least from a decrease in NFkB interaction with the k sequences and from an increase in CREB interaction with CRE sequences. This might be at the origin of the differential effect of CAPE on the inductions of NFkB, TNF-a and NOS-2 in brain tissue exposed to inflammatory stress. Similarly, the data on the comparative study between CAPE and other anti-inflammatory drugs (Fig. 3), are suggesting that each antiinflammatory drug interfered in a unique way with the regulation of NFkB, TNF-a, NOS-2, and CREB levels. Although NFkB activation was strongly reduced with AS, SN 50, Dexa, and CAPE, only Dexa and CAPE decreased NO synthesis whereas these four drugs did decrease TNFa secretion. On the other hand, PDTC inactive on glial NFkB pathway [51], as confirmed here, was deprived of


P. Montpied et al. / Molecular Brain Research 115 (2003) 111–120

effect on TNF-a, whereas it was, with CAPE, the strongest inhibitor of NO synthesis. Interestingly, as CAPE, PDTC is well known for its antioxidant properties and in this study it was found to produce a quite strong increase in CREB nuclear activity. Therefore, it is possible to hypothesize that the analogy between their effects on oxidant species, on NO synthesis and on CREB nuclear activity reveals an important similarity of their mechanism of action. It has been shown that ROS decrease CREB nuclear concentration and that, this was involved in the cellular death induced by ROS [47,41]. Therefore, PDTC and CAPE by decreasing ROS might increase the CREB nuclear concentrations that will then suppress NOS-2 expression leading to a lowering of NO synthesis. This would explain that PDTC, a quite strong antioxidant, produced slightly higher increase of CREB but a less effective suppression of NOS gene, since compared to CAPE it was deprived of effect on NFkB. This absence of effect on NFkB is also a potential cause of its absence of control of TNF-a secretion. Concerning Dexa, a drug commonly used in neurology, its effect on both TNF-a and NO levels might be due to a dual effect on NFkB and on the glucocorticoid receptor, its typical target known to be involved in the regulation of numerous inflammatory events. Dexa effect on NFkB is consistent with other reports on leucocytes and mononuclear showing that glucocorticoids increase IkB and decrease expression [1]. Dexa effect on NOS-2 gene expression was characterized in macrophages and thus its effect on microglial nitrite production is certainly mediated through similar mechanisms and probably mostly due to GRE (glucocorticoid responsive element) sequences present on NOS-2 gene [3]. The lack of effects of AS and SN 50, the two selective NFkB inhibitors, on the control of NOS-2 gene activation showed that this selectivity narrows their effect to genes having a promoter with a dominant regulation by NFkB. Acting at least on two complementary pathways, CAPE has therefore a better potential than these selective NFkB inhibitors. Such a dual action had been described for endogenous peptides having anti-inflammatory action in the CNS, these peptides were shown to interact with a nuclear factor that interacts with both the NFkB and the CREB binding. We here demonstrate that CAPE acts consistently as an anti-inflammatory drug in brain, altering several major effectors of inflammation shown to be involved in neurotoxic events in this organ [11,27,7,12,2,39]. In addition, its long lasting control of neuroinflammation (48 h) with a single administration at the time of the inflammatory stress appears as a quite interesting aspect in view of the hypothesized importance in the delayed death events of the inflammatory response [11,7,44,12,2]. At this stage, the pharmacokinetic of CAPE in vivo and the level of passive diffusion across BBB should be investigated. CAPE was shown to have a very good permeability coefficient across membrane due to its

physicochemical properties [8], and on the other hand was reported to have a beneficial effect in an in vivo model of ischemia / reperfusion spinal cord injury [26]. In conclusion, our data give important aspects of CAPE effects on brain tissue undergoing inflammatory stress. Its synergic effect on molecular loops responsible for the amplification and the spreading of inflammation in brain is clearly encouraging for further research on its use as a beneficial and effective drug in neurological diseases, where excessive activation of glial cells may have dangerous consequences for neuronal survival.

References [1] A. Aljada, H. Ghanim, E. Assian, P. Mohanty, W. Hamouda, R. Garg, P. Dandona, Increased IkappaB expression and diminished nuclear NF-kappaB in human mononuclear cells following hydrocortisone injection, J. Clin. Endocrinol. Metab. 84 (1999) 3386– 3389. [2] K.R. Bales, Y. Du, D. Holtzman, B. Cordell, S.M. Paul, Neuroinflammation and Alzheimer’s disease: critical roles for cytokine / Ab-induced glial activation, NF-kB and apolipoprotein E, Neurobiol. Aging 21 (2000) 427–432. [3] P.J. Barnes, I. Acock, Anti-inflammatory action of steroids: molecular mechanisms, Trends Pharmacol. Sci. 14 (1993) 436–441. [4] J. Bauer, H. Rauschka, H. Lassmann, Inflammation in the nervous system: the human perspective, Glia 36 (2001) 235–243. [5] M. Benkirane, K.T. Jeang, C. Devaux, The cytoplasmic domain of CD4 plays a critical role during the early stages of HIV infection in T-cells, EMBO J. 13 (1994) 5559–5569. [6] N. Blondeau, C. Widmann, M. Lazdunski, C. Heurteaux, Activation of the nuclear factor-kB is key event in brain tolerance, J. Neurosci. 21 (2001) 4668–4677. [7] F. Boissiere, S. Hunot, B. Faucheux, C. Duyckaerts, J.J. Hauw, Y. Agid, E.C. Hirsch, Nuclear translocation of NF-kappa B in cholinergic neurons of patients with Alzheimer’s disease, Neuroreport 8 (1997) 2849–2852. [8] G.C. Ceschel, P. Maffei, A. Sforzini, S. Lombardi Borgia, A. Yasin, C. Ronchi, In vitro permeation through porcine buccal mucosa of caffeic acid phenethyl ester (CAPE) from topical mocuadhesive gel containing propolis, Fitoterapia 73 (2002) S44–S52. [9] C.B. Combs, S.C. Karlo, S.C. Kao, G.E. Landreth, Beta-amyloid stimulation of microglia and monocytes results in TNF a-dependent expression of inducible nitric oxide synthase and neuronal apoptosis, J. Neurosci. 15 (2001) 1179–1188. [10] F. de Bock, J. Dornand, G. Rondouin, Release of TNF-a in the rat hippocampus following epileptic seizures and excitotoxic neuronal damage, Neuroreport 7 (1996) 1125–1129. ¨ G. Rondouin, The [11] F. de Bock, B. Derijard, J. Dornand, J. Bockaert, neuronal death induced by endotoxic schock but not that induced by excitatory amino acids requires TNF-a, Eur. J. Neurosci. 10 (1998) 3107–3114. [12] M. Delgado, Vasoactive intestinal peptide and pituitary adenylate cyclase-activating polypetide inhibit CBP-NF-kB interaction in activated microglia, Biochem. Biophys. Res. Com. 297 (2002) 1181–1185. [13] H.A. Ding, C.F. Nathan, D.J. Stuehr, Release of reactive nitrogen intermediates and reactive oxygen intermediates from mouse peritoneal macrophages, J. Immunol. 141 (1988) 2407–2412. [14] Y. Dong, W.M. Rohn, E.N. Benveniste, IFN-g regulation of the type IV class II transactivator promoter in astrocytes, J. Immunol. 162 (1999) 4731–4739.

P. Montpied et al. / Molecular Brain Research 115 (2003) 111–120 [15] U. Fiszer, Does Parkinson’s disease have an immunological basis? The evidence and its therapeutic implications, BioDrugs 15 (2001) 351–355. [16] L.R. Fitzpatrick, J. Wang, J. Le, Caffeic acid phenethyl ester, an inhibitor of nuclear factor-kappaB, attenuates bacterial peptidoglycan polysaccharide-induced colitis in rats, J. Pharmacol. Exp. Ther. 229 (2001) 915–920. [17] J. Glasgow, R. Perez-Polo, One path to cell death in the nervous system, Neurochem. Res. 25 (2000) 1373–1383. [18] M. Grilli, M. Memo, Nuclear factor–kappa B / Rel proteins: a point convergence of signalling pathways relevant in neuronal function and dysfunction, Biochem. Pharmacol. 57 (1999) 1–7. [19] M. Grilli, M. Pizzi, M. Memo, P. Spano, Neuroprotection by aspirin and sodium salicylate through the blockade of NFkB activation, Science 274 (1996) 1383–1385. [20] E. Grunblatt, S. Mandel, M.B. Youdim, Neuroprotective strategies in Parkinson’s disease using the models of 6-hydroxydopamine and MPTP, Ann. NY Acad. Sci. 899 (2000) 262–273. [21] I. Guarini, Z.Z. Su, S. Zucker, J. Lan, D. Grunberger, P.B. Fisher, Growth inhibition and modulation of antigenic phenotype in human melanoma and glioblastoma multiforme cells by caffeic phenethyl ester (CAPE), Cell Mol. Biol. 38 (1992) 513–527. [22] J.J. Haddad, S.C. Land, Nuclear factor Kappa B blockade attenuates but does not abrogate liposaccharide-dependent tumor necrosis factor-alpha biosynthesis in alveolar epithelial cells, Biochem. Biophys. Res. Commun. 285 (2001) 267–272. [23] K.G. Hauesler, M. Prinz, C. Nolte, J.R. Weber, R.R. Schumann, H. Kettenmann, U.K. Hanisch, Interferon-g differentially modulates the release of cytokines and chemokines in lipopolysaccharide and pneumococcal cell wall-stimulated mouse microglia and macrophages, Eur. J. Neurosci. 16 (2002) 2113–2122. [24] B. Hauss-Wegrzyniak, P.D. Vraniak, G.L. Wenk, LPS-induced neuroinflammatory effects do not recover with time, Neuroreport 11 (2000) 1759–1763. [25] A.J. Herrera, A. Castano, J.L. Venero, J. Cano, A. Machado, The single intranigral injection of LPS as a new model for studying the selective effects of inflammatory reactions on dopaminergic system, Neurobiol. Dis. 7 (2000) 429–447. [26] A. Ilhan, U. Koltuksuz, S. Ozen, E. Zu, H. Ciralik, O. Akyol, The effects of caffeic acid phenthyl ester (CAPE) on spinal cord ischemia / reperfusion injury in rabbits, Eur. J. Cardiothorac. Surg. 16 (1999) 458–463. [27] G.H. Jeohn, W.G. Kim, J.S. Hong, Time dependency of the action of nitrite oxide in lipopolysaccharide interferon-gamma-induced neuronal cell death in murine primary neuron-glia co-cultures, Brain Res. 880 (2000) 173–177. [28] K. Kako, H. Wakamatsu, T. Hamada, M. Banasik, K. Ohata, T. Niki-Kuroiwa, S. Suzuki, J. Takeuchi, N. Ishida, Examination of DNA-binding activity of neuronal transcription factors by electrophoretical mobility shift assay, Brain Res. Protoc. 2 (1998) 242– 249. [29] S. Lacroix, D. Feinstein, S. Rivest, The bacterial endotoxin lipopolysaccharide has the ability to target the brain in upregulating its membrane CD14 receptor within specific cellular populations, Brain Pathol. 8 (1998) 625–640. [30] N. Laflamme, S. Rivest, Toll-like 4: the missing link of the cerebral innate immune response triggered by circulating gram-negative bacterial cell wall components, FASEB J. 15 (2001) 155–163. [31] J.F. Lau, C.M. Horvath, Mechanisms of type I interferon cell signaling and STAT-mediated transcriptional responses, Mt Sinai J. Med. 69 (2001) 156–168. [32] G. Levi, Regulation of cyclic AMP synthesis in microglial cells and possible role of cyclic AMP in neuroprotection, in: Patterson, Kordon, Christen (Eds.), Neuroimmune Interactions in Neurological and Psychiatric Disorder, Springer, Berlin, Heidelberg, 2000, pp. 124–138.


[33] Y.Z. Lin, S.Y. Yao, R.A. Veach, T.R. Torgerson, J. Hawiger, Inhibition of nuclear translocation of transcription factor NF-kB by synthetic peptide containing a cell membrane-permeable motif and nuclear localization sequence, J. Biol. Chem. 270 (1995) 14255– 14258. [34] T. Mabuchi, K. Kitagawa, K. Kuwabara, K. Takasawa, T. Ohtsuki, Z. Xia, D. Storm, T. Yanagihara, M. Hori, M. Matsumoto, Phosphorylation of cAMP response element-binding protein hippocampal neurons as a protective response after exposure to glutamate in vitro and ischemia in vivo, J. Neurosci. 21 (2001) 9204–9213. [35] A.M. Moerman, M. Xianrong, M.M. Lucas, S.W. Barger, Characterization of neuronal kB-binding factor distinct from NF-kB, Mol. Brain Res. 67 (1999) 303–315. [36] S. Nadeau, S. Rivest, Role of microglia-derived tumor necrosis factor in mediating CD14 transcription and nuclear factor kappa B activity in the brain during endotoxemia, J. Neurosci. 20 (2000) 3456–3468. [37] K. Natarajan, S. Singh, T.R. Burke, D. Grunbeger, B.B. Aggarwal, Caffeic acid phenethyl ester is a potent and specific inhibitor of activation of nuclear transcription factor NF-kB, Proc. Natl. Acad. Sci. USA 93 (1996) 9090–9095. [38] A.H. Neufeld, B. Liu, Comparison of the signal transduction pathways for the induction of gene expression of nitric oxide synthase-2 in response to two different stimuli, Nitric Oxide 8 (2003) 95–102. [39] H. Neumann, H. Wekerle, Neuronal control of the immune response in the central nervous system: linking brain immunity to neurodegeneration, J. Neuropathol. Exp. Neurol. 57 (1998) 1–9. [40] Z. Orban, N. Mitsiades, T.R. Burkes, M. Tsokos, G.P. Chrousos, Caffeic acid phenethyl ester induces leukocyte apoptosis modulates nuclear factor-k B and suppress acute inflammation, Neuroimmunomodulation 7 (2000) 99–105. [41] S. Pagazhenthi, A. Nesterova, J. Purevsuren, G. Audesirk, M. Kern, L. Cabell, E. Eves, M.R. Rosner, L.M. Boxer, J.E. Reusch, Oxidative stress-mediated down–regulation of bcl-2 promoter in hippocampal neurons, J. Neurochem. 84 (2003) 982–996. [42] S.K. Park, H.L. Lin, S. Murphy, Nitric oxide regulates nitric oxide synthase-2 expression by inhibiting NF-kB binding to DNA, Biochem. J. 322 (1997) 609–613. [43] X. Qiao, D.J. Cummins, S.M. Paul, Neuroinflammation-induced acceleration of amyloid deposition in the APPV717F transgenic mouse, Eur. J. Neurosci. 14 (2001) 474–482. [44] Z.H. Qin, Y. Wang, N. Masami, T.N. Chase, Nuclear Factor kB contributes to excitotoxin-induced apoptosis in rat striatum, Mol. Pharmacol. 53 (1998) 33–42. [45] N.J. Rothwell, G.H. Luheshi, Interleukin 1 in the brain: biology, pathology and therapeutic target, Trends Neurosci. 23 (2000) 618– 625. [46] A. Russo, R. Longo, A. Vanella, Antioxidant activity of propolis: role of caffeic acid phenethyl ester and galangin, Fitoterapia 73 (2002) S21–S29. [47] V. See, J.P. Loeffer, Oxidative stress Induces neuronal death by recruiting a protease and phosphatase-gated mechanism, J. Biol. Chem. 276 (2001) 35049–35059. [48] S. Son, B.A. Lewis, Free radical scavenging and antioxidative activity of caffeic acid amide and ester analogues: structure–activity relationship, J. Agric. Food Chem. 50 (2002) 468–472. [49] Y.S. Song, E.H. Park, G.M. Hur, Y.S. Ryu, Y.S. Lee, J.Y. Lee, Y.M. Kim, C. Jin, Caffeic acid phenethyl ester inhibits nitric oxide synthase gene expression and enzyme activity, Cancer Lett. 175 (2002) 53–61. [50] L. Stoppini, P.A. Buchs, D. Muller, A simple method for organotypic cultures of nervous tissue, J. Neurosci. Methods 37 (1991) 173–182. [51] H. Togashi, M. Sasaki, E. Frohman, E. Taira, R.R. Ratan, T.M. Dawson, V.L. Dawson, Neuronal (type I) nitric oxide synthase


P. Montpied et al. / Molecular Brain Research 115 (2003) 111–120

regulates nuclear factor kappaB activity and immunologic (type II) nitric oxide synthase expression, Proc. Natl. Acad. Sci. USA 94 (1997) 2676–2680. [52] M. Triantafilou, K. Triantafilou, Lipopolysaccharide recognition: CD14, TLRs and the LPS-activation cluster, Trends Immunol. 23 (2002) 301–304.

[53] J.T. Yang, C.N. Chang, T.H. Lee, J.C. Hsu, T.N. Lin, J.H. Wu, Dexamethasone inhibits ischemia-induced transient reduction of neurotrophin-3 mRNA in rat hippocampal neurons, Neuroreport 26 (1998) 3477–3480.