Background
Interleukin-6 (IL-6) is a pleiotropic cytokine involved in several brain diseases as a detrimental factor playing a causal or exacerbating role in neuroinflammation and neurodegeneration [
1‐
7]. Elevated levels of IL-6 are typical for brains from animal models or humans suffering from multiple sclerosis, Alzheimer's disease, Parkinson's disease, lethal sepsis, meningitis and stroke [
8‐
12]. Additionally, long-term exposure of neurons or astrocytes to IL-6 as well as over-activation of IL-6 signaling by IL-6/sIL-6R fusion protein lead to a robust induction of neuroinflammatory response and to neuronal death [
6,
13‐
16]. Therefore, suppression of IL-6 signaling or of IL-6 expression itself is believed to represent a powerful strategy for the treatment or prevention of neuroinflammation and subsequent neurodegeneration. This is supported by diminished neuroinflammation induced by spinal cord injury after infusion of a monoclonal antibody against IL-6 receptor [
17]. Furthermore, the potency of drugs to inhibit IL-6 expression
in vitro and
in vivo correlates with their anti-neuroinflammatory and neuroprotective properties [
18‐
22].
Astrocytes, the main glial cell type of the brain, respond in general to multiple kinds of acute and chronic brain insults with a reaction known as astrogliosis [
23]. This reactive astrogliosis involves morphological, structural and biochemical features including thickened cellular processes, increased expression of glial fibrillary protein and the induction of pro-inflammatory cytokines including IL-6 [
24,
25]. Different types of signaling molecules are able to trigger the astrocytic IL-6 mRNA expression via distinct intracellular signaling pathways [
26]. For example, lipopolysaccharide (LPS) activates the IL-1 receptor-associated kinase (IRAK)-dependent pathway including IκB kinase and nuclear factor κB (NF-κB) [
27]. Another potent group of IL-6 inducers are cytokines such as tumor necrosis factor α, interleukin-1β, oncostatin M (OSM) and leukaemia inhibitory factor (LIF) [
28‐
30]. Interestingly, OSM and LIF belong together with IL-6 to the same cytokine family. These IL-6-type cytokines are characterized by using of glycoprotein gp130 to induce gene expression via JAK/STAT (Janus kinase/signal transducer and activator of transcription) and MAPK (mitogen-activated protein kinase) cascades in a NF-κB-dependent manner [
31,
32]. Thus, blocking of such pathological IL-6-driven gene expression by low molecular weight inhibitors provides a possible strategy for targeting the onset or further propagation of astrogliosis and, subsequently, secondary neuronal cell death.
In the present study, the time- and dose-dependent stimulation of IL-6 expression by OSM was characterized in human U343 glioma cells. Subsequently, our compound libraries were screened for inhibitory effects on OSM-induced IL-6 expression. We identified bioactive compounds belonging to the chemical class of heteroarylketones (HAK). These HAK compounds were able to suppress the LPS-induced IL-6 expression in primary mouse and rat astrocytes as well as in an acute septic shock mouse model in vivo. Finally, the underlying molecular mechanism of HAK compounds interfering with key signaling molecules of OSM-induced signal transduction cascade was analyzed. We demonstrate a selective suppression by HAK compounds of the OSM-mediated phosphorylation of STAT3 at serine 727, which affects STAT3 binding to the NF-κB subunit p65.
Methods
Primary cultures of murine astrocytes
According to Löffler [
33], astrocyte-rich primary cell cultures were started with brains of newborn mice and rats and were maintained in Dulbecco's modified Eagle's medium (DMEM) for 33 days at 37°C in a humified atmosphere with 95% air/5% CO
2.
Cell culture
The human glioma cell line U343 [
34] was maintained in DMEM containing 10% fetal bovine serum and 60 μg/ml gentamycin (Invitrogen, Darmstadt, Germany) at 37°C in a 10% CO
2 atmosphere.
Lipopolysaccharide (LPS)-induced acute septic shock model
A total number of 20 C57/B6 mice at the age of 5 months with an initial body weight of 22 g to 30 g was analyzed in this study. Mice were randomly assigned to 2 groups of control mice (injection of saline or LPS; n = 6 each) and 1 group of LPS-treated mice co-injected with compound HAK-2 (n = 8). Animals were housed at the Experimental Animal Core Facility of the University of Leipzig. The study was approved by the Regierungspräsidium Leipzig, License # TVV 28/07 on November 14, 2007. To induce septic shock and acute inflammation, 14 mice were injected intraperitoneally (i.p.) with 1 mg LPS (Serotyp 055:B5)/kg body weight. Saline injection i.p. was used as control treatment (n = 6). Compound HAK-2 was co-injected i.p. to 8 LPS-treated mice at a concentration of 10 mg/kg body weight. Two hours post treatment mice were sacrificed by CO2 inhalation and tissue samples from hippocampus and cortex as well as plasma were prepared as described later.
Mouse brain samples for qRT-PCR were prepared from 6 - 8 animals per group. After removal of the brain, the tissue samples were flushed shortly with ice cold saline and placed briefly on filter paper. Cortex and hippocampus were prepared, weighted and snap frozen in liquid nitrogen. Samples were stored at -80°C until RNA isolation. Frozen tissue samples were homogenized in RNA lysis buffer (Macherey and Nagel, Düren, Germany) using Precellys ceramic beads (diameter of 1.4 mm; Peqlab, Erlangen, Germany).
At the end of the experiment blood samples were collected in ice-cooled tubes and centrifuged within the next 20 min (10 min, 1500 × g, 4°C). Plasma samples were aliquoted into fractions of 50 μl, shock frozen and stored at -80°C until analysis.
Quantitative real-time PCR
Total RNA of cultured cells and homogenized mouse brain tissue samples was isolated using RNA isolation kit "NucleoSpin RNA II" from Macherey and Nagel. RNA quantity was measured by spectrophotometrical quantification (NanoDrop 2000, Peqlab, Erlangen, Germany). Total RNA (0.5 μg) was transcribed to cDNA using Superscript III and oligo-dT primer (Invitrogen, Darmstadt, Germany). Quantitative real-time PCR was performed with QuantiFast SYBR Green PCR Master mix (Qiagen, Hilden, Germany) using Rotorgene 3000 system (Corbett, Sydney, Australia). Gene expression was normalized to the expression of three reference genes glyceraldehyde 3-phosphate dehydrogenase (GAPDH), glucose-6-phosphate dehydrogenase (G6PDH) and hypoxanthine-guanine phosphoribosyltransferase (HPRT). Primers for PREP (5'-TGAGCAGTGTCCCATCAGAG-3'; 5'- CATCTTCGCTGAACGCATAA-3') and IL-6 (5'-AAAGGACGTCACATTGCA-3'; 5'-GCCTCAGACATCTCCAGTCC-3') were designed using the Primer3 Software (
http://frodo.wi.mit.edu/). Data were analyzed by Rotorgene 3000 Software version 6.0 by comparative quantitation. The appropriate size of PCR products was confirmed by gel-electrophoresis stained with ethidium bromide.
ELISA
Measurements of IL-6 in conditioned medium and murine plasma samples were done by means of human- or murine-specific IL-6 Cytoset kits (Invitrogen, Darmstadt, Germany) according to manufacturer's protocols.
Preparation of whole-cell extracts for western blotting and immunoprecipitation
Human U343 cells were lysed with cell lysis buffer (Invitrogen, Darmstadt, Germany) supplemented with 0.2 mg/ml sodium orthovanadate (Sigma, Taufkirchen, Germany), protease inhibitor mix complete mini (Roche, Mannheim, Germany) and 1 mM AEBSF (Sigma, Taufkirchen, Germany) for 30 min on ice. Lysates were centrifuged for 15 min at 15,000 × g and 4°C. Protein content in supernatants was quantified by Bradford assay (Sigma, Taufkirchen, Germany) according to the manufacturer's protocol and stored at -20°C.
Western blotting
Western blotting was performed with 30 μg protein of whole-cell extracts, mixed with 4 x SDS sample loading buffer (Invitrogen, Darmstadt, Germany) and denatured for 10 min at 85°C. Cell extracts separated by 4 - 12% Novex Bis-Tris Mini Gel system (Invitrogen, Darmstadt, Germany) were transferred to Roti-NC nitrocellulose membranes (Roth, Karlsruhe, Germany). Membranes were probed with primary antibodies against STAT3 (#9132) and P-STAT3S727 (#9134) from Cell Signaling (Frankfurt/M., Germany) as well as with anti-αtubulin (Sigma, Taufkirchen, Germany) to confirm equal loading and blotting of protein samples. Proteins were visualized using HRP-conjugated secondary antibodies (1:4000, cell signaling, Frankfurt/M., Germany) and the SuperSignal West Pico system (Thermo Fisher Scientific, Karlsruhe, Germany).
Small interfering RNA (siRNA)
Human U343 cells (0.25 × 106) were seeded in 24-well plates (Greiner, Frickenhausen, Germany) and transfected immediately with 2.5 μl A. dest. (mock), 100 μM non-target control (NTC) or PREP-specific siRNA ON-TARGETplus SMARTpool (#L-006006-00, Dharmacon, Schwerte, Germany) using DharmaFECT 1 siRNA transfection reagent (Dharmacon, Schwerte, Germany) according to the manufacturer's protocol. After 48 h adherent cells were transfected a second time under identical conditions for further 24 h and subsequently stimulated with OSM (100 ng/ml) for additional 6 h. For IL-6 specific ELISA 5 - 40 μl of conditioned media were utilized. To analyze IL-6 mRNA expression by qRT-PCR, total RNA was isolated and reversely transcribed as described above.
Immunocytochemistry
Human U343 cells (1 × 105 cells/well) were grown on cover slips in 24-well plates (Greiner, Frickenhausen, Germany) for 24 h. After the time of treatment indicated cells were fixed in ice-cold methanol for 10 min on ice, and then incubated with rabbit anti-phospho-STAT3 antibody (#9134, Cell Signaling, Frankfurt/M., Germany) overnight at 8°C. Subsequently, cells were incubated with goat anti-rabbit IgG Cy2-conjugated secondary antibody (Dianova, Hamburg, Germany) at room temperature for 45 min. Finally, cover slips were mounted on microscope slides and approximately 250 cells/sample were evaluated densitometrically by fluorescence microscopy (Zeiss, Jena, Germany) and MetaMorph Image Analysis Software (Universal Imaging Corporation, USA).
Immunoprecipitation
Cell lysates from 4 × 106 U343 cells/sample were obtained as described above. From each sample 200 μg of total protein were immunoprecipitated with 2 μg rabbit anti-p65 antibody (Santa Cruz Biotechnology; Heidelberg, Germany) or A. dest. (control) overnight at 4°C. Immunoprecipitated samples were incubated with 20 μl Dynabeads Protein G (Thermo Fisher Scientific, Karlsruhe, Germany) for 1 h at 4°C. Beads were washed 3 times with 1 ml PBS. Preparation of samples for SDS-PAGE analysis was done by means of dilution with SDS-PAGE sample buffer (Invitrogen, Darmstadt, Germany), followed by denaturation at 90°C for 10 min. SDS-PAGE analysis were performed as described before.
PREP enzymatic activity assay
The activity of human recombinant prolyl endopeptidase (PREP) was determined photometrically using the substrate Z-Gly-Pro-pNA (Bachem, Weil am Rhein, Germany) usually at a concentration of 150 μM. As assay buffer 50 mM HEPES buffer pH 7.6, containing 200 mM NaCl, 1 mM EDTA, and 1 mM dithiothreitol was used. Release of pNA was monitored continuously at 405 nm for 15 min at 30°C in a 96-well plate reader (Spectrafluor, Tecan, Crailsheim, Germany). PREP activity was calculated from the slope of the time product curve with the help of a pNA standard [
35]. For determination of Ki values three different substrate concentrations (62.5 μM; 125 μM and 250 μM) and seven different inhibitor concentrations were analyzed. Substrate concentrations were selected to be at the Km value of Z-Gly-Pro-pNA (125 μM) as well as half and twice the Km. Data were fitted by non-linear regression to the competitive inhibitor equation (Graphit 5.0, Erithacus software).
Statistical analyses
Values are expressed as mean ± SD. Standard unpaired t test was used for analyses of statistically significance. Differences between treatments were considered significant when p < 0.05.
Discussion
Neuropathological situations with extended astroglial activation are associated with increased levels of pro-inflammatory mediators including IL-6.
In vitro and
in vivo studies have demonstrated that IL-6 plays a pivotal role in the initiation of neuroinflammatory cascades and in secondary neuronal cell death [
6,
45]. Thus, prevention of the neuroinflammatory response to primary lesions has a neuroprotective potential.
The present study was performed to identify new small molecular weight inhibitors acting on the pathway that results in IL-6 expression. For screening of our in-house compound libraries the human glioblastoma cell line U343 was used, because glioblastoma cell lines were shown to respond with increased IL-6 expression to different neuroinflammatory stimuli like LPS [
27], Substance P [
46], tumor necrosis factor α, interleukin-1β, leukemia-inhibitory factor and OSM [
30]. Analysis of conditioned media revealed, that in our experimental setup only OSM treatment significantly induced the expression of IL-6 in human U343 glioma cells. This result is consistent with published data, showing that U343 cells express the OSM-receptor components LIFR and OSMRβ as well as the common signal transducer gp130 [
47]. Furthermore, the OSM-mediated activation of signal components of the Jak/STAT- and MAPK-pathways was described for U343 and U373 glioma cells, respectively [
32]. We observed a biphasic induction pattern of OSM-induced IL-6 mRNA expression, which was described earlier also for human U373 astroglioma cells [
32]. The time-course is characterized by a first strong, rapid and transient IL-6 mRNA expression peak at 1 h followed by a second one at 6 h with a less strong, but prolonged induction. The same type of expression pattern was observed for tissue factor mRNA in OSM-treated smooth muscle cells [
48]. Thus, biphasic induction seems to be an OSM-specific feature with general relevance for OSM action.
All potent inhibitors of IL-6 secretion identified in the compound library screen (see Table
1) belong to the chemical class of HAK and are structurally related to inhibitors of PREP [
38]. This observation is in line with the hypothesis, that PREP is involved in regulation of intracellular protein transport and secretion [
49]. However, there was no correlation between PREP siRNA- (knock-down) and pharmacological inhibition of PREP on one hand and the potency of these compounds to suppress the OSM-induced IL-6 expression on the other. Furthermore, our data on the temporal profile of IL-6 suppression suggest that the bioactivity of HAK compounds is most likely based on interference with IL-6 mRNA synthesis but not on disturbed intracellular transport and secretion mechanisms. Therefore, PREP can be excluded as the IL-6 relevant molecular target of HAKs and HAK compounds appear to interact with at least one further molecular target. Interestingly, only the second IL-6 mRNA peak was affected by HAKs indicating that the molecular target of HAK compounds is involved 3 to 6 h post OSM stimulation at earliest. Theoretically, the biological target of HAKs can pre-exist in untreated cells or be induced by OSM treatment and subsequently incorporated in signaling pathways. Notably, in an experiment analyzing the puromycin sensitivity of OSM-induced IL-6 mRNA expression, it was demonstrated that OSM induces the protein synthesis of signaling molecules essential for the second IL-6 mRNA expression peak [
32]. Whereas puromycin completely abrogated the second IL-6 expression peak it showed no effect on the first OSM-induced IL-6 mRNA peak. This demonstrates a requirement for de novo protein synthesis exclusively for the second IL-6 expression peak of this biphasic response signaling. The relationship between HAK-mediated suppression of OSM-induced IL-6 release and the effect of HAK compounds exclusively on the second mRNA peak suggests that more than 75% of secreted IL-6 is based on the second phase of OSM-induced IL-6 mRNA expression. Thus, the mRNA induced in the first phase appears to have regulatory functions rather than acting as a template in protein synthesis. Such a regulatory role of mRNA molecules was recently described by Poliseno et. al. [
50] showing that mRNA molecules from pseudogenes or long non-coding RNAs can act as competitive endogenous RNAs sequestering microRNA molecules.
To elucidate whether the HAK-mediated suppression of OSM-induced IL-6 expression is cell line-specific or valid in general, experiments with primary murine astrocytes were performed. In contrast to human U343 glioblastoma OSM did not induce IL-6 expression in mouse and rat primary astrocytes. However, LPS, known to act as a powerful stimuli of cytokines [
51], significantly increased IL-6 expression in primary murine astrocytes. Co-treatment with HAK compounds markedly suppressed levels of OSM-stimulated IL-6 expression in both rat and mouse astrocytes. These data demonstrate that the anti-inflammatory bioactivity of HAKs is not limited to a single OSM-based cell culture model but also valid for a series of pathophysiological conditions contributing to neuroinflammation and neurodegeneration.
We were also interested to reveal whether HAK compounds are bioactive under inflammatory conditions in vivo. For this study, compound HAK-2 was selected based on its beneficial features concerning toxicity, bioavailability and blood brain barrier passage. In accordance with the data obtained from primary murine astrocytes, compound HAK-2 significantly suppressed LPS-induced IL-6 levels in brain- and plasma-derived samples from septic mice. This result strongly indicates the anti-inflammatory potency of HAK compounds in vivo for possible treatment of central nervous system diseases.
To get more information about the underlying molecular mechanism of HAK bioactivity, the signal-transduction pathways involved in OSM-mediated IL-6 expression were dissected in more detail. Interestingly, LPS- and OSM-induced signal pathways are based on the same molecular mechanism such as STAT3 or NF-κB activation [
31,
52], indicating that HAK compounds may target a common cellular event. Two major signaling cascades the JAK/STAT as well as the MAPK pathways are switched on by binding of OSM to the receptor heterodimers OSMR/gp130 or LIFR/gp130 [
53]. Subsequent activation of signal tyrosine kinases of the JAK family leads to phosphorylation of pivotal signal molecules such as STAT3 and Erk1 and 2 respectively [
31,
32]. The essential role of receptor subunits as well as of downstream signaling molecules as STAT3, Erk1 and p65 for OSM-triggered IL-6 expression in U343 cells was confirmed by siRNA-based knock-down experiments (data not shown). Furthermore, Erk1/2 and STAT3 were phosphorylated 6 h post OSM treatment, which was identified as the critical time point for the HAK bioactivity. Immunoblotting and immunofluorescence experiments revealed that neither OSM-induced pErk1/2
T202/Y204 phosphorylation nor pSTAT3
Y705 phosphorylation were modified by HAK compounds. However, HAK treatment led to a significant reduction of OSM-stimulated pSTAT3
S727 phosphorylation. Importantly, the HAK-based inhibition profiles for IL-6 expression and pSTAT3
S727 phosphorylation are strongly correlating with each other. Thus, suppression of OSM-induced phosphorylation of pSTAT3
S727 is most likely the relevant molecular mechanism of the HAK compound bioactivity to suppress IL-6 expression. In contrast to pSTAT3
Y705, which is essential for dimerization, nuclear translocation and DNA binding [
54,
55], the physiological role of pSTAT3
S727 is discussed controversially [
56‐
59]. Depending on the specific promoter and/or the cellular context pSTAT3
S727 can influence transcriptional activity of target genes [
60,
61].
However, in the case of the IL-6 promoter, where activated NF-κB binds directly to DNA, no cis-regulatory elements for STAT3 binding were identified so far [
62‐
64]. Based on these observations, we hypothesize that pSTAT3
S727 may regulate IL-6 gene expression by an alternative pathway. It is known that STAT3 is complexed with transcription factors such as c-Jun, c-Fos, forkhead and endothelial cell-derived zinc finger protein, respectively [
65‐
68]. Furthermore, it was shown that physical interaction of the STAT3 DNA-binding domain with the NF-κB subunit p65 led to a reduced promoter activity of inducible nitric oxide synthase gene [
69].
Together, these findings strongly suggest that physical interaction between STAT3 and p65 may result in a functional coupling important for the STAT3-dependent regulation of p65 responsive genes. Indeed, we here demonstrated by co-immunoprecipitation that p65 and STAT3 interact with each other in an OSM-dependent manner. Noteworthy, the OSM-stimulated STAT3 and p65 complex formation is quite sensitive against treatment with HAK compounds. This supports our hypothesis and indicates for the first time a regulatory function for pSTAT3S727 in OSM-triggered STAT3/NF-κB interaction.
In summary, HAK compounds selectively target the expression of genes with promoters co-regulated by pSTAT3S727-dependent signaling. Based on this mechanism, kinases phosphorylating STAT3 at serine 727 such as MAPKs, mTOR, NLK and PKCδ may represent direct molecular targets of HAK compounds. Thus, further studies are required to identify the precise molecular mechanisms and the neuroinflammatory-related genes sensitive to HAK-treatment. This will enable the therapeutic development of HAK compounds for treatment of neurological diseases including Alzheimer's disease, multiple sclerosis, Parkinson's disease and traumatic brain injury.
Competing interests
IS, CE, AJN, KM, AK and AM are employees of Probiodrug AG. HUD serves as CSO of Probiodrug AG.
Authors' contributions
IS initiated the project and wrote the paper. IS, AJN, CE, UZ, KM, AM and SR performed the experiments and analyzed the data. AK assisted with the design and interpretation of qRT-PCR experiments. AK, HUD and SR participated in study design and coordination as well as drafted and revised the manuscript. All authors read and approved the final manuscript.