Abbreviations used : DMEM, Dulbecco's modified Eagle medium ; IL-6, interleukin-6 ; MRE, multiple response element ; NF-κB, nuclear factor-κB ; PKC, protein kinase C ; TPA, phorbol 12-myristate 13-acetate.
Address correspondence and reprint requests to Dr. M. Schwaninger at Department of Neurology, University of Heidelberg, Im Neuenheimer Feld 400, 69120 Heidelberg, Germany.
Abstract : Bradykinin, a mediator of inflammation, is produced in the brain during trauma and stroke. It is thought to open the blood-brain barrier, although the mechanism is unclear. We have investigated, therefore, the effect of bradykinin on the expression of interleukin-6 (IL-6), a putative modulator of the blood-brain barrier, in astrocytes. IL-6 gene transcription was evaluated by transient transfection of the human IL-6 promoter linked to the luciferase gene. In murine astrocytes, bradykinin stimulated IL-6 secretion and gene transcription. The effect of bradykinin was blocked by KN-93, an inhibitor of Ca2+/calmodulin-dependent protein kinases, and by bisindolyl-maleimide I, an inhibitor of protein kinase C, suggesting the involvement of these protein kinases. Mutations in the multiple response element and the binding site for nuclear factor-κB (NF-κB), but not in other known elements of the IL-6 promoter, interfered with induction of IL-6 transcription. The involvement of NF-κB was supported further by the binding that overexpression of nmIκBα, a stable inhibitor of NF-κB, inhibited the induction of IL-6 by bradykinin. Bradykinin activated NF-κB in primary astrocytes as shown by increased DNA binding of NF-κB. These data demonstrate that bradykinin stimulates IL-6 expression through activation of NF-κB, which may explain several inflammatory effects of bradykinin.
Bradykinin is a peptidergic mediator of inflammation. It is released from its precursor kininogen by the protease kallikrein in response to tissue damage or injury. In the periphery, bradykinin can elicit all major symptoms of inflammation, such as pain, hyperperfusion, and extravasation of plasma. The components of the kinin-kallikrein system are also expressed in brain (Walker et al., 1995). In the brain, elevated extracellular levels of bradykinin are found in cerebral trauma and ischemia (Ellis et al., 1988 ; Kamiya et al., 1993). As in peripheral tissue, bradykinin can increase the permeability of brain capillaries (Wahl et al., 1988). This effect of bradykinin has been exploited to open the blood-brain barrier for chemotherapy in brain tumors (Matsukado et al., 1996), but may also underlie the vasogenic brain edema found in cerebral ischemia (Kamiya et al., 1993 ; Relton et al., 1997) and trauma (Ellis et al., 1988).
Blood-brain barrier functions of endothelial cells are induced by astrocytes. The precise mechanism, either paracrine or mediated by cell-cell contact, is unknown up to now. However, the cytokine interleukin-6 (IL-6), which is expressed at a low level in normal brain, has been implicated as a regulator of blood-brain barrier properties in endothelial cells. IL-6 induces the expression of marker enzymes of brain capillary endothelial cells (Takemoto et al., 1994 ; Sun et al., 1997). On the other hand, IL-6 can increase the permeability of the blood-brain barrier (Brett et al., 1995 ; de Vries et al., 1996). This may contribute to the disturbed blood-brain barrier in diseases such as cerebral ischemia and trauma, in which IL-6 expression in astrocytes is highly increased (Hariri et al., 1994 ; Maeda et al., 1994).
Here we report that bradykinin can induce the transcription of the IL-6 gene in astrocytes and characterize the intracellular signaling cascade involved.
MATERIALS AND METHODS
Bradykinin, thapsigargin, poly(dI-dC), and phorbol 12-myristate 13-acetate (TPA) were purchased from Sigma (Munich, Germany) ; bisindolylmaleimide I was obtained from Calbiochem, and pRLSV40 from Promega (Mannheim, Germany). DEAE dextran was obtained from Pharmacia (Freiburg, Germany). KN-93 was a kind gift of H. Hidaka. A stock solution of TPA (1 mM) was prepared in dimethyl sulfoxide and further diluted in Dulbecco's modified Eagle medium (DMEM). Thapsigargin and bisindolylmaleimide I were dissolved in dimethyl sulfoxide. KN-93 and bradykinin were dissolved in water. Controls received the solvent only.
Astrocytes were prepared from the brains of neonatal mice (postnatal day 2) by enzymatic digestion as has been described (Marriott et al., 1995). Fourteen days after preparation, oligodendrocytes and microglia were removed from the underlying astrocytic monolayer by shaking the culture flasks on a rotary shaked for 4-6 h at 800 rpm. Astrocytes and the human astrocytoma cell line 1321N1 (Brismar, 1995) were cultured in DMEM containing 10% fetal calf serum, penicillin (50 IU/ml), and streptomycin (50 μg/ml). If IL-6 release was measured, the medium was changed immediately before stimulation. IL-6 was measured in medium 6 h after stimulation by using a commercially available ELISA for murine IL-6 according to the manufacturer's instruction (Pharmaingen, Hamburg, Germany). The detection limit was 10 pg/ml.
Two days before transfection, cells were plated on 6-cm plates. Different from a previous report (Schwaninger et al., 1997), primary astrocytes were transfected with DEAE dextran, as we obtained lower variation of transfection efficiency with this method. Plates were washed with phosphate-buffered saline, and then 2 μg of the first reporter was added in full medium containing 75 μg/ml DEAE dextran for 1 h. Plates were washed again with phosphate-buffered saline. When indicated, 2 μg of pMSC-nmIκBα was cotransfected. pMSV-nmIκBα contains the coding sequence of IκBα with mutations S32A/S36A in the vector pMexNeo I and was a kind gift of R.-P. Ryseck and R. Bravo. To maintain a constnat DNA concentration, Bluescript was added to controls. 1321N1 cells were transfected by using the calcium phosphate method ; after 5 h, cells were exposed for 3 min to 10% glycerol in DMEM. After incubation for 42 h in full medium, cells were stimulated for 6 h and harvested. Firefly luciferase activity was measured as described (Schwaninger et al., 1993). To control for transfection efficiency, 0.1 μg of pRLSV40 per plate was cotransfected and measured with the Dual Luciferase Reporter Assay (Promega). The mutations within the human IL-6 promoter have been described before (Dendorfer et al., 1994) apart from m2κB-IL6Luc, which contains the following mutation within the nuclear factor-κB (NF-κB) binding site : GT AAT ATT TTC CCA T. For the present study, all IL-6 promoter mutations were subcloned into the BamHI/XhoI site of the luciferase vector pXP2 (Nordeen, 1988) except the mutation of the multiple response element (MRE), which was subcloned into the HindIII/XhoI site of the same vector. All constructs were confirmed by sequencing. For statistical analysis, the t test was used.
Gel shift assays
Primary astrocytes on 6-cm plates were stimulated for 1 h. Nuclear extracts were prepared as described (Schreiber et al., 1989). Double-stranded oligonucleotides with 5′-GATC overhangs and the following sequence were annealed and labeled by a fill-in reaction with [α-32P]cCTP and Klenow enzyme : GAT CCA GAG GGG ACT TTC CGA GA. Labeled oligonucleotide and nuclear extract were incubated on ice for 15 min in a buffer containing 10 mM Tris-HCl (pH 7.5), 1 mM EDTA, 50 mM NaCl, 0.15 μg/μl poly(dI-dC), 5% glycerol, 0.1 mg/ml bovine serum albumin, and 1 mM dithiothreitol (total volume, 20 μl). When indicated, antisera (1 μl) that have been described before were added (Weigh et al., 1994). Then the complexes were separated on a 6% polyacrylamide gel [180 V, 0.5× TBE (0.045 M Tris-borate, 1 mM EDTA)]. Gels were analyzed by autoradiography. For competition, the following mutated NF-κB binding site was used : GAT CCA GAC CAT GGT ATC CGA GA.
Stimulation of IL-6 secretion and gene transcription by bradykinin
Addition of increasing concentrations of bradykinin to the incubation medium led to a concentration-dependent release of IL-6 from astrocytes (Fig. 1). The maximal stimulation was 2.1-fold with an EC50 of 3.4 μM for the release of IL-6. As for other cytokines, the release of IL-6 is regulated mainly at the level of gene transcription. In separate experiments, IL-6 gene transcription was investigated by transfecting primary astrocytes with a reporter gene construct containing the coding sequence of luciferase under transcriptional control of the human IL-6 promoter (-1,179/+9). Bradykinin stimulated IL-6 gene transcription in a concentration-dependent manner (Fig. 1). The maximal stimulation obtained by 10 μM bradykinin varied between experiments from 1.5- to 2.5-fold (Figs. 1 and 2B). The EC50 for IL-6 gene transcription was 3.2 μM.
Pharmacological characterization of the intracellular signaling cascade mediating the bradykinin effect on IL-6 gene transcription
Bradykinin receptors are linked via G proteins to several signal-transduction pathways that encompass activation of adenylyl cyclase, phospholipase A, and phospholipase C (Bhoola et al., 1992). In astrocytes, bradykinin activates phospholipase C, induces formation of inositol 1,4,5-trisphosphate, and elevates the intracellular Ca2+ concentration (Gimpl et al., 1992 ; Bender et al., 1993 ; Stephens et al., 1993). To investigate whether Ca2+ -dependent protein kinases are involved in the effect of bradykinin on IL-6 gene transcription, KN-93, a selective inhibitor of Ca2+/calmodulin-dependent protein kinases, was used (Sumi et al., 1991). KN-93 (10 μM) inhibited the stimulation of IL-6 transcription by bradykinin (Fig. 2A). However, the effect of the adenosine analogue 2-chloroadenosine, which is mediated by protein kinase A (data not shown), was unaffected by KN-93 (Fig. 2A) in accordance with the selective inhibition of Ca2+/calmodulin-dependent protein kinases. Bisindolylmaleimide I, a potent inhibitor of protein kinase C (PKC) (Toullec et al., 1991), inhibited the stimulatory effect of bradykinin on IL-6 gene transcription as well (Fig. 2B). Chelerythrine, another inhibitor of PKC, also blocked the stimulatory effect of bradykinin on IL-6 gene transcription but also elevated the basal level of IL-6 transcription more than bisindolylmaleimide I (Fig. 2B, data not shown). In accordance with the notion that the effect of bradykinin on IL-6 transcription is mediated by a Ca2+/calmodulin-dependent protein kinase and PKC, activation of PKC by TPA (300 nM) stimulated IL-6 genetranscription 2.2 ± 0.1-fold (n = 4), and mobilization of intracellular Ca2+ by thapsigargin induced IL-6 transcription 2.1 ± 0.1-fold (n = 4) in primary astrocytes.
Mapping of the IL-6 promoter element responsive to PKC and Ca2+ mobilization in astrocytes
Various response elements have been described in the IL-6 promoter (Sehgal, 1992). To define the elements that mediate the response to activation of PKC and mobilization of intracellular Ca2+ in astrocytes, reporter fusion constructs harboring internal mutations in several elements of the human IL-6 promoter were transfected into the human astrocytoma cell line 1321N1. In contrast to primary astrocytes, thapsigargin alone did not stimulate, but rather inhibited, luciferase activity. The inhibition by thapsigargin alone was not sensitive to mutations in the IL-6 promoter and therefore could be due to a posttranscriptional effect (Fig. 3). However, in the presence of TPA, thapsigargin did enhance IL-6 gene transcription (Fig. 3). Both stimulation by TPA and the synergism between TPA and thapsigargin were unaffected by mutations in the TRE (TPA response element), GRE1 (a putative glucocorticoid response element), and the binding site for C/EBPβ (NF-IL6) (Fig. 3). Also, a mutation in the GRE2 (another putative glucocorticoid response element) did not interfere with the stimulation by TPA or TPA/thapsigargin (data not shown). However, two different mutations in the binding site for NF-κB, as well as a mutation in the MRE, clearly reduced the stimulation by TPA or TPA/thapsigargin. This points to the involvement of NF-κB and one of the factors binding to the MRE in the stimulation of IL-6 transcription by activation of PKC and Ca2+ mobilization.
NF-κB mediates the stimulation of IL-6 transcription by bradykinin in primary astrocytes
To investigate the role of NF-κB in bradykinin stimulation of IL-6 transcription in primary astrocytes, the effect of a mutation in the NF-κB binding site on bradykinin stimulation was tested. Mutation of the NF-κB binding site completely inhibited the stimulatory effect of bradykinin on IL-6 gene transcription (Fig. 4).
In unstimulated cells, NF-κB is retained in the cytosol through association with inhibitory IκB proteins. Following stimulation, IκB is degraded, resulting in NF-κB translocation to the nucleus. nmIκBα, which is stable to degradation in the cell due to mutations at Ser32 and Ser36, can act as a dominant negative inhibitor of NF-κB (Brockman et al., 1995 ; DiDonato et al., 1996). In our experiments, overexpression of nmIκBα reduced the stimulatory effect of bradykinin on IL-6 transcription (Fig. 4).
To test whether bradykinin can activate NF-κB, we investigated DNA binding of NF-κB in gel shift assays. After treatment with bradykinin, two bands (I and II) were enhanced that were specifically competed by the wild type, but not a mutated NF-κB binding site (Fig. 5A). The enhancement was already evident after 20 min of stimulation with bradykinin, but was more pronounced after 1 and 3 h (Fig. 5B). To characterize the two bands further, antisera against three subunits of NF-κB were added to the incubation. A preimmune serum (PI) and an antiserum against p52 had no effect. However, the antiserum against p50 abolished both complexes I and II and the antisera against RelA (p65) reduced complex I (Fig. 5B). This demonstrates that band I consists of RelA/p50 heterodimers, whereas complex II contains p50 homodimers as has been observed in a brain extract (Schneider et al., 1999). In conclusion, bradykinin activates NF-κB in primary astrocytes and thereby induces IL-6 transcription.
This study shows that bradykinin can stimulate the release of IL-6 from astrocytes. Bradykinin has been shown previously to stimulate the release of IL-6 in some other, but not all investigated cell types (Paegelow et al., 1995 ; Rehbock et al., 1997 ; Modéer et al., 1998 ; Wiernas et al., 1998). The parallel stimulation of IL-6 secretion and gene transcription argues that the increased secretion is due at least in part to increased transcription of the IL-6 gene. Little is known about the mechanisms whereby bradykinin stimulates gene transcription.
Astrocytes are known to carry B2 receptors. Like other membrane receptors (McCarthy and Salm, 1991), B2 receptors are found only on a subset of astrocytes. Among astrocytes of type 1, ~60% of the cells carry bradykinin receptors (Stephens et al., 1993). This implies that bradykinin stimulates IL-6 expression in cells with B2 receptors nearly twice as effectively as indicated in our model, which measures effects over all astrocytes. B2 receptors have a high affinity for bradykinin. In astrocytes, a KD of 16 nmol/L has been reported (Cholewinski et al., 1991). However, in accordance with the concentration-response curve obtained for electrophysiological responses (Gimpl et al., 1992 ; Stephens et al., 1993), micromolar concentrations of bradykinin are needed to stimulate IL-6 expression (Fig. 1). An explanation for this apparent discrepancy may be that larger concentrations of bradykinin are required for each step in the signal transduction process from receptor binding to physiological response.
Bradykinin receptors are thought to be coupled to phospholipase C β via interaction with Gq protein (Bhoola et al., 1992). Several authors have reported that in astrocytes bradykinin induces breakdown of phosphoinositide leading to activation of PKC and mobilization of intracellular Ca2+ (Stephens et al., 1993 ; Chen et al., 1995). Activation of both PKC and Ca2+/calmodulin-dependent protein kinases is involved in the induction of IL-6 gene transcription by bradykinin (Fig. 2). Other neurotransmitters, such as substance P, histamine, and serotonin, which also stimulate IL-6 expression in astrocytes, use similar second messenger pathways (Pousset et al., 1996 ; Lieb et al., 1998). Our data confirm previous findings that PKC activates IL-6 gene transcription through the MRE (Ray et al., 1989).
In addition, the transcription factor NF-κB is involved in induction of IL-6 gene transcription by bradykinin. This conclusion is supported by three lines of evidence : (a) mutations within the binding site for NF-κB inhibited the stimulation of IL-6 transcription by bradykinin or PKC and intracellular Ca2+ mobilization (Figs. 3 and 4) ; (b) overexpression of the stable NF-κB inhibitor nmIκBα inhibited the induction of IL-6 gene transcription by bradykinin (Fig. 4) ; and (c) bradykinin activated NF-κB in primary astrocytes as shown by the increased DNA binding of p50 homodimers and p50/RelA heterodimers (Fig. 5). It is known that the IL-6 gene can be transactivated by NF-κB in response to the cytokines interleukin-4, tumor necrosis factor-α, and interleukin-1β (Libermann and Baltimore, 1990 ; Zhang et al., 1990 ; Sparacio et al., 1992). In contrast to bradykinin, cytokines activate NF-κB through a signaling cascade that does not involve PKC or Ca2+ (Meichle et al., 1990). Nevertheless, activation of PKC is a classic stimulus of NF-κB, and also mobilization of Ca2+ from intracellular stores has been reported before to activate NF-κB, although the involved signaling cascades are not yet well defined (Frantz et al., 1994 ; May and Ghosh, 1998). The above data suggest that bradykinin induces IL-6 gene transcription through a PKC- and Ca2+ -dependent activation of NF-κB.
NF-κB plays a critical role in immune and inflammatory responses (May and Ghosh, 1998). Some of the effects of bradykinin as a mediator of inflammation therefore may be due to activation of NF-κB. This mechanism could explain, for example, the stimulation of interleukin-1β and tumor necrosis factor-α expression by bradykinin (Tiffany and Burch, 1989), as the expression of both cytokines is controlled by NF-κB at the transcriptional level (May and Ghosh, 1998). In the brain, induction of IL-6 or other cytokines in astrocytes may contribute to the effect of bradykinin on the blood-brain barrier. Exogenous bradykinin exerts a very fast effect on the blood-brain barrier (Butt, 1995) that is hardly compatible with induction of gene transcription. However, the more long-lived opening of the blood-brain barrier in inflammatory, traumatic, or ischemic brain disease may relate to induction of gene transcription by bradykinin.
We greatly appreciate the generous gifts of KN-93 from H. Hidaka (Nagoya, Japan), of pMSV-nmIκBα from R. P. Ryseck and R. Bravo (Princeton, NJ, U.S.A.), of the antisera from F. Weih (Karlsruhe, Germany), and of 1321N1 cells from J. van der Kaay (Dundee, U.K.). This study was supported by a grant of the DFG (Schw 416/3-1 and 3-2).