TGF-β1 is a master cytokine in immune regulation, orchestrating both pro- and anti-inflammatory reactions. Recent studies show that whereas TGF-β1 induces a quiescent microglia phenotype, it plays a pathogenic role in the neurovascular unit and triggers neuronal hyperexcitability and epileptogenesis. In this study, we show that, in primary glial cultures, TGF-β signaling induces rapid upregulation of the cytokine IL-6 in astrocytes, but not in microglia, via enhanced expression, phosphorylation, and nuclear translocation of SMAD2/3. Electrophysiological recordings show that administration of IL-6 increases cortical excitability, culminating in epileptiform discharges in vitro and spontaneous seizures in C57BL/6 mice. Intracellular recordings from layer V pyramidal cells in neocortical slices obtained from IL-6treated mice show that during epileptogenesis, the cells respond to repetitive orthodromic activation with prolonged after-depolarization with no apparent changes in intrinsic membrane properties. Notably, TGF-β1induced IL-6 upregulation occurs in brains of FVB/N but not in brains of C57BL/6 mice. Overall, our data suggest that TGF-β signaling in the brain can cause astrocyte activation whereby IL-6 upregulation results in dysregulation of astrocyteneuronal interactions and neuronal hyperexcitability. Whereas IL-6 is epileptogenic in C57BL/6 mice, its upregulation by TGF-β1 is more profound in FVB/N mice characterized as a relatively more susceptible strain to seizure-induced cell death.

Epilepsy is one of the most common neurologic disorders and is estimated to affect up to 1% of the population worldwide (1). Postinjury epilepsy often develops following brain insults, including ischemic or traumatic injury, as well as following infectious and inflammatory diseases (2). Recent studies suggest that vascular injury, and specifically blood–brain barrier (BBB) dysfunction and the extravasation of serum albumin, plays a key role in postinjury epilepsy (37) (for reviews, see Refs. 8, 9).

Epileptic brains often show glial activation, which was suggested to play an important role in neuronal hyperexcitability (10), synaptogenesis (11), and epileptogenesis (12, 13). Under BBB breakdown, serum albumin binds to TGF-βR2, activates TGF-β signaling cascade (through SMAD2/3 phosphorylation), and induces an astrocytic transcriptional response with proinflammatory characteristics (10, 1416). Furthermore, losartan, previously identified as a blocker of peripheral TGF-β signaling, effectively blocks albumin-induced brain TGF-β signaling and prevents epilepsy (16).

Although accumulating experimental data from animal models strongly support a role for inflammatory responses in epileptogenesis (17, 18), several important questions remain unanswered: What triggers the epileptogenic inflammatory process? Which specific cell populations are involved? What inflammatory mediators are critical to the epileptogenesis process? And what are the mechanisms by which inflammatory cytokines affect neuronal excitability? Because activation of TGF-β signaling in glial subsets has been implicated in both pro- and anti-inflammatory processes (14, 1925), this study aims to uncover the molecular and cellular mechanisms promoting a proinflammatory TGF-β1 signaling in non-neuronal populations and their potential role in the induction of epilepsy.

C57BL/6 mice were purchased from The Jackson Laboratory (Bar Harbor, ME). FVB/N mice were purchased from Harlan Laboratories (Rehovot, Israel). Mice were bred and maintained in a local specific pathogen-free animal facility. All surgical and experimental procedures were approved by the Institutional Animal Care and Use Committee of Ben-Gurion University of the Negev, Israel.

Primary glial cultures were prepared from the cerebral cortex of 1-d-old C57BL/6 mice as previously described (19). Briefly, cortices were excised, cleaned from meningeal tissues, and digested with 2.5% trypsin (solution C, Biological Industries, Beit Haemek, Israel) supplemented with 0.5 mg/ml DNase I (Worthington Biochemical, Lakewood, NJ) for 5 min at 37°C. DNase I (5 mg/ml) was then added prior to passing the digested tissue through a thin pipette several times. The cells were then suspended in DMEM (Life Technologies, Paisley, U.K.) supplemented with 10% FBS (Thermo Fisher Scientific, Logan, UT), 4 mM l-glutamine, 100 U/ml penicillin, 1 μg/ml streptomycin, 2.5 U/ml nystatin, 10 mM HEPES, 1 mM sodium pyruvate, 10 mM nonessential amino acids, and 50 μM 2-ME (all purchased from Biological Industries) and seeded onto poly-d-lysine (Sigma-Aldrich, Jerusalem, Israel)–coated flasks and kept at 37°C, 5% CO2 and 95% humidity. Growth medium was replenished after 24 h and every 2–3 d thereafter. The culture reaches confluence after 12–14 d and is then subjected to quantitative PCR (qPCR), ELISA, and Western blot analyses as whole glia or purified cell subsets as described below.

Cell separation procedure was conducted according to the standard manufacturer’s protocol. In brief, glial cells were harvested between days 14 and 20 of the culture using a 0.25% trypsin solution (solution A, Biological Industries), washed, and incubated with PE-conjugated anti-CD11b magnetic beads (Stemcell Technologies, Vancouver, BC, Canada). Cells were then placed onto a magnet for separation. Subsequently, cells were stained with PE-conjugated anti–GLAST-1 Ab (Milteny Biotec, Bergisch Gladbach, Germany) and analyzed with FACSCalibur (BD Biosciences, Franklin Lakes, NJ). Glial cultures were further analyzed using PE-conjugated anti-CX3CR1, allophycocyanin-Cy7–conjugated anti-F4/80, PE/Dazzle-conjugated anti-CD11c, Alexa Fluor 700–conjugated anti–MHC class II, Brilliant Violet 421–conjugated anti-Ly6C (BioLegend, San Diego, CA), Pacific Orangeconjugated anti-CD45.2 (Invitrogen, Grand Island, NY), and FITC-Alexa Fluor 488–conjugated anti-CCR2 (R&D Systems, Minneapolis, MN). Isotype control staining was used for all Abs (BioLegend). Samples were analyzed with Gallios (Beckman Coulter, Brea, CA).

Mixed glia, purified astrocytes, and microglia were seeded in 6-well (1 × 106 cells/well) or 48-well (1 × 105 cells/well) plates and were treated with 10 ng/ml TGF-β1 (PeproTech, Rocky Hill, NJ) for 8–48 h as indicated in the figure legends. Cells were cultured in the absence or presence of 10 μM SB431542 (an ALK4/5/7 inhibitor; Sigma-Aldrich), 10–20 μM SIS3 (a SMAD3 inhibitor; Santa Cruz Biotechnology, Dallas, TX), or 10 μM SJN2511 (an ALK5 inhibitor; Tocris Bioscience, Bristol, UK). Supernatants were then analyzed by ELISA array (for IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-10, IL-12, IL-17, MCP-1, IFN-γ, TNF-α, MIP-1α, RANTES, and GM-CSF) (Quansys Bioscience, Logan, UT) or with a single cytokine sandwich ELISA (for IL-6 and TNF-α) (BioLegend) according to the manufacturers’ instructions. Total RNA was extracted with TRI Reagent (Sigma-Aldrich) according to the manufacturer’s instructions. RNA was stored at −80°C. Two micrograms RNA was reverse transcribed with a high-capacity cDNA reverse transcription kit (Applied Biosystems/Invitrogen, Carlsbad, CA). IL-6, IL-1β, and TNF-α gene expression was analyzed with TaqMan real-time PCR (Applied Biosystems/Invitrogen). TGF-βR gene expression was quantified with SYBR Green real-time PCR (Roche, Basel, Switzerland) using the following primers: forward, 5′-CCTCGAGACAGGCCATTTGTA-3′, reverse, 5′-GCTGACTGCTTTTCTGTAGTTGG-3′; TGF-βR2, forward, 5′-TTTGCGATGTGAGACTGTCC-3′, reverse, 5′-GAGTGAAGCCGTGGTAGGTG-3′. Samples were run in triplicates. The GAPDH gene was used as an endogenous control to normalize gene expression.

Mixed glia and purified astrocytes were seeded in six-well plates at a density of 6 × 105 cells/well. Cultures were treated with 10 ng/ml TGF-β1 for 1 h in the presence or absence of 10 μM SB431542. Cell lysates were prepared using a lysis buffer containing 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM EGTA, 1% Nonidet P-40, and cocktails of protease (Sigma-Aldrich) and phosphatase (Santa Cruz Biotechnology) inhibitors. Cell lysates were quantified with the Bradford protein assay (Bio-Rad, Hercules, CA) using BSA (Sigma-Aldrich) as standard. Forty-microgram samples were then separated by SDS-PAGE (10% acrylamide) followed by a gel transfer to 0.45 μM nitrocellulose membrane (Cole-Parmer, Vernon Hills, IL). Blots were blocked with 5% nonfat dry milk in TBS-T (10 mM Tris, 135 mM NaCl [pH 7.4], 0.05% Tween 20) and incubated overnight at 4°C with primary anti-SMAD2/3 diluted 1:1000 (Cell Signaling Technology), anti–phospho-(Ser465/467)-SMAD2 diluted 1:1000 (Cell Signaling Technology, Beverly, MA), or anti–TGF-βR2 diluted 1:1000 (Abcam, Cambridge, U.K.). Anti–β-actin, diluted 1:10,000 (MP Biomedicals, Solon, OH), was used to control protein load. Blots were then incubated with peroxidase-conjugated secondary Abs (anti-rabbit [GE Healthcare Life Sciences, Little Chalfont, U.K.] or anti-mouse [Jackson ImmunoResearch Laboratories, West Grove, PA]) for 1.5 h at room temperature. Detection of immunoreactive bands was carried out with ImageQuant LAS 4000 (GE Healthcare Life Sciences) using ECL.

Primary glial cells or purified astrocytes were cultured on eight-well Lab-Tek culture slides (1.5 × 105 cells/well) treated with 10 ng/ml TGF-β1 for 1 h, fixed in 4% buffered paraformaldehyde for 15 min, washed in PBS, and permeabilized in methanol for 10 min at −20°C. After PBS washing, cells were blocked for 2 h at room temperature with Dyna Ab diluent (GBI Labs, Bothell, WA) and were then incubated for 24 h at 4°C with primary anti-CD68 diluted 1:250 (BioLegend) or anti–glial fibrillary acidic protein (GFAP) diluted 1:500 (Invitrogen) together with anti-SMAD2/3 and anti–p-SMAD3 diluted 1:200 (Cell Signaling Technologies). Anti-CNPase was diluted 1:150 (Abcam). Cells were then washed with PBS and incubated with Alexa Fluor 488, 546, or 633 Abs (Invitrogen) diluted 1:500. TO-PRO-3 (Invitrogen) diluted 1:1500 was used for counterstaining. All images were obtained with an Olympus FluoView FV1000 confocal microscope (Olympus, Hamburg, Germany). Z-stack images were taken at 0.5-μm intervals across a 20-μm sample thickness. Confocal images were analyzed with Volocity image analysis software (Impovision, Waltham, MA). Fluorescence intensities of SMAD2/3 or phospho-SMAD3 staining were obtained from the TO-PRO-3–labeled area of GFAP+ or CD68+ cells. In each experiment, image acquisition settings were initially adjusted to avoid saturation and were then used for all experimental groups. At least 10 images were obtained for each experimental group from randomly selected fields.

Hippocampi and cortices were dissected from the brain and lysed in a buffer containing 3 mM NaCl, 250 mM HEPES, 2% Triton X-100, and cocktails of protease (Sigma-Aldrich) and phosphatase (Santa Cruz Biotechnology) inhibitors. Total RNA was extracted with TRI Reagent (Sigma-Aldrich) according to the manufacturer’s instructions and stored at −80°C. RNA quality was analyzed using Bioanalyzer (Agilent Technologies, Santa Clara, CA). Two micrograms RNA was reverse transcribed with a high-capacity cDNA reverse transcription kit (Applied Biosystems/Invitrogen) and IL-6, IL-1β, and TNF-α gene expression was quantified with TaqMan real-time PCR (Roche). Samples were run in triplicates. The GAPDH gene was used as an endogenous control to normalize gene expression.

For extracellular electrophysiological experiments, 12- to 14-wk-old C57BL/6 mice were anesthetized with isoflurane, brains were removed, and transverse cortico-hippocampal slices (400 μm thick) were prepared with a vibratome according to established methods (5). Slices were maintained in a humidified, carbogenated (5% CO2 and 95% O2) gas atmosphere at 36.1°C and perfused with artificial cerebrospinal fluid (ACSF; 124 mM NaCl, 26 mM NaHCO3, 1.25 mM NaH2PO4, 2 mM MgSO4, 2 mM CaCl2, 3 mM KCl, 10 mM glucose [pH 7.4]) in a standard interface chamber (3, 5). Slices were incubated with either ACSF or IL-6 (100 ng/ml) for 2–3 h before transfer to the perfusion chamber. Recordings were performed with a constant ACSF perfusion. Glass microelectrodes (3 MΩ, 154 mM NaCl) were positioned in layers four through five of the sensory-motor neocortex. Slices were stimulated with brief (0.1 ms) pulses using bipolar stimulation electrodes placed at the border between white and gray matter in the same cortical column.

Somatic whole-cell recordings were obtained from layer five pyramidal neurons in somatosensory neocortical coronal slices prepared from C57BL/6 mice implanted either with ACSF (control) or IL-6 (5 ng/h using a stock solution of 5 μg/ml) containing osmotic pumps (Alzet osmotic pumps, Cupertino, CA) for 3–4 d (26). Mice were anesthetized with Nembutal (60 mg kg−1) and decapitated. Coronal slices (300 μm) from primary somatosensory cortex were sectioned with a vibratome (VT1200, Leica, Wetzlar, Germany) and placed in a holding chamber containing oxygenated ACSF at room temperature; they were transferred to a recording chamber after >1 h of incubation.

Cells were viewed with a ×60× water-immersion lens (Olympus) in an Olympus BX51WI microscope mounted on an X–Y translation stage (Luigs and Neumann, Ratingen, Germany). Whole-cell recordings were conducted using patch pipettes pulled from thick-walled borosilicate glass capillaries (1.5 mm outside diameter, Hilgenberg, Malsfeld, Germany). The pipette solution contained 130 mM potassium gluconate, 6 mM KCl, 2 mM MgCl2, 4 mM NaCl, and 10 mM HEPES (pH adjusted to 7.25 with KOH). Pipettes had resistances of 5–7 MΩ when filled with this solution. Recordings were made using an Axoclamp-2A amplifier equipped with an HS-2-x0.01MU headstage (Molecular Devices, Sunnyvale, CA) in bridge mode; data were low pass filtered at 30 kHz (−3 dB, single-pole Bessel filter) and digitized at 100 kHz. Brief (0.1 ms) orthodromic stimuli were delivered via a bipolar tungsten electrode (resistance ∼0.5 MΩ) placed at the border between gray and white matter. Trains of pulses, 10-fold the threshold intensity, were applied at 50 Hz for 1 s. Input resistance was measured by injecting long (150–500 ms) low amplitude (<25 pA) hyperpolarizing current pulses. Three control and six IL-6–injected (5 ng/h) mice were used for these experiments. All recordings were made at room temperature (22 ± 2°C). Electrophysiological data analysis was accomplished using pCLAMP 7.0 (Axon Instruments) and Igor Pro 5.05 (Wavemetrics, Lake Oswego, OR).

Experiments were performed on 12- to 14-wk-old C57BL/6 mice. For electrode implantation, animals were deeply anesthetized (isoflurane, 1.5–2%) and positioned in a stereotactic frame. Two screw electrodes were implanted 3 mm posterior and 2 mm lateral to bregma and were fixed to the skull with bone cement. Alzet osmotic pumps were implanted in position 0.5 mm posterior, 1 mm lateral, and 3 mm deep relative to bregma and contained ACSF only (1 μl/h, n = 6) or ACSF supplemented with IL-6 (1 μg/ml [1 ng/h], n = 3 or 5 μg/ml [5 ng/h], n = 8). Perfusion lasted 1 wk. The pump and transmitter (Data Science International, Saint Paul, MN) were placed s.c. in the dorsal part of the animal’s neck. Following the procedure, the incision was sutured and the animals were s.c. injected with buprenorphine (0.1 mg/kg) and then allowed to recover from the anesthesia.

After recovery, the animals were moved to a behavior room and maintained on a 12-h circadian cycle with ad libitum access to food and water. Single-channel electrographic signals were acquired from freely moving mice for 2 wk using a telemetric electroencephalogram monitoring system (Data Science International), as previously described (27).

At the end of the monitoring period, the animals were sacrificed and their brains were extracted. Two mice were excluded due to wound infection (n = 1) or a gross inflamed lesion on the surface of the cortex (n = 1). For unbiased detection of seizures, recordings were analyzed retrospectively using an in-house written algorithm, based on feature extraction and artificial neural network classification as previously described (16).

For ELISA, Western blot, immunocytochemistry (ICC), and qPCR analyses, statistical significance was tested with unpaired one-way ANOVA. Comparison of amplitude and duration of seizure-like events were made with an unpaired Student t test. Graphs indicate mean ± SD, with n representing the number of mice, slices, or cultures used for analysis, as indicated in the figure legend. A χ2 test was used to compare the frequency of epileptic slices in extracellular recording, and a Fisher exact test was used to compare the frequency of epileptic animals in treated and control groups. Seizure duration and amplitude was compared by two-tailed Student t test.

To explore the role of TGF-β signaling in glial inflammatory response, we treated primary glial cultures with TGF-β1 and quantified cytokine expression at the mRNA and protein levels using qPCR and ELISA, respectively. Of all cytokines and chemokines analyzed (see 2Materials and Methods), only IL-6 and MCP-1 (CCL2) are significantly increased 24 h after treatment (Fig. 1A, 1B). A time kinetics experiment showed that the upregulation of IL-6 mRNA is evident already 8 h after treating the cells with TGF-β1 and peaked 24 h after treatment (12.19- and 22.87-fold increases, respectively; F = 20.62, p < 0.001) (Fig. 1C). The transcriptional response is associated with increased secreted levels of IL-6 in the supernatant at 8, 24, and 48 h after treatment (379.5 ± 59.5, 825.1 ± 25.3, and 1473.6 ± 153.7 pg/ml, respectively) (Fig. 1D). A qPCR analysis shows a milder increase in the mRNA levels of IL-1β and TNF-α at 24 h after treatment (F = 13.57, p < 0.01; Fig. 1C); however, ELISA does not demonstrate significantly elevated protein levels (data not shown).

FIGURE 1.

TGF-β1 induces rapid IL-6 upregulation in primary glial cultures. Cultures prepared from newborn C57BL/6 mice were treated with 10 ng/ml TGF-β1 or left untreated. Cytokine and chemokine expression and secretion were quantified by qPCR and ELISA. (A and B) Multiplex ELISA analysis performed 24 h following TGF-β1 treatment. (C and D) qPCR analysis of IL-1β, TNF-α, and IL-6 (C), and ELISA analysis of secreted IL-6 (D), at 8, 24, and 48 h following TGF-β1 treatment. Bars and data points represent means ± SD obtained from one experiment out of at least three performed. *p < 0.05, **p < 0.01, ***p < 0.001 compared with untreated cells; (A and B) unpaired two-tailed Student t test, (C and D) one-way ANOVA test.

FIGURE 1.

TGF-β1 induces rapid IL-6 upregulation in primary glial cultures. Cultures prepared from newborn C57BL/6 mice were treated with 10 ng/ml TGF-β1 or left untreated. Cytokine and chemokine expression and secretion were quantified by qPCR and ELISA. (A and B) Multiplex ELISA analysis performed 24 h following TGF-β1 treatment. (C and D) qPCR analysis of IL-1β, TNF-α, and IL-6 (C), and ELISA analysis of secreted IL-6 (D), at 8, 24, and 48 h following TGF-β1 treatment. Bars and data points represent means ± SD obtained from one experiment out of at least three performed. *p < 0.05, **p < 0.01, ***p < 0.001 compared with untreated cells; (A and B) unpaired two-tailed Student t test, (C and D) one-way ANOVA test.

Close modal

To determine the glia cell subsets that express IL-6 following exposure to TGF-β1, we enriched astrocyte and microglia cultures with anti-CD11b magnetic beads as described in 2Materials and Methods. Both microglia and astrocytes were used at purities >95% (Fig. 2A). The purified astrocytic culture contained ∼2.07 ± 0.07% CNPase+ oligodendrocyte precursors (data not shown). The microglial subset contained primarily CD11b+F4/80+CX3CR1+ cells and were negative for CD11c, CD45.2, CCR2, Ly6C, and MHC class II staining (data not shown). The cultures were incubated with TGF-β1 for 8, 24, or 48 h. Compared with untreated astrocytes, IL-6 is significantly upregulated in TGF-β1–treated astrocytes at both the mRNA (17.4-, 42.2-, and 34.3-fold increase at 8, 24, and 48 h, respectively, F = 44.6, p < 0.001 for 24 and 48 h, p < 0.05 for 8 h; Fig. 2B) and protein levels (396.2 ± 9.2, 722.9 ± 32.6, and 1459.5 ± 188.1 pg/ml at 8, 24, and 48 h, respectively, F = 64.7, p < 0.001 for each time point; Fig. 2C), with kinetics similar to those of mixed glia (Fig. 1). This is in contrast to purified microglia, which show no significant increase in IL-6 mRNA levels (Fig. 2B) and only a mild accumulation of secreted IL-6 (64 ± 0.5, 165.3 ± 4.7, and 266.2 ± 23.5 pg/ml at 8, 24, and 48 h after treatment, respectively, F = 324.8, p < 0.01; Fig. 2C). LPS stimulation of glia or the astrocyte- and microglia-purified subsets acting via the TLR pathway show, as expected, a more profound secretion of IL-6 by microglia than by astrocytes (Fig. 2D) along with increased secreted levels of IL-1β and TNF-α (data not shown). Two separate blockers of SMAD2/3 phosphorylation, namely ALK 4/5/7 (SB431542) and ALK5 (SJN2511), inhibit TGF-β1–induced IL-6 secretion in cultures of mixed glia (F = 29.56, p < 0.01), enriched astrocytes (F = 51.26, p < 0.01), or enriched microglia (F = 35.25, p < 0.01) (Fig. 2E). A SMAD3-specific inhibitor (SIS3) decreases IL-6 secretion although to a lesser extent, namely by 60 and 57% in cultures of mixed glia and astrocytes, respectively (p < 0.01; Fig. 2E). Taken together, these data indicate that TGF-β1 induces a SMAD2/3-dependent upregulation and secretion of IL-6, as the main proinflammatory cytokine, preferentially in astrocytes.

FIGURE 2.

TGF-β1 induces IL-6 upregulation primarily in astrocytes. Primary glial cultures were first purified to astrocytic and microglial cultures, and cells were then treated with 10 ng/ml TGF-β1. The kinetics of IL-6 expression were evaluated by qPCR and ELISA at 8, 24, and 48 h. (A) FACS analysis of mixed glia, purified microglia, and purified astrocytes. Microglia were labeled with and anti-CD11b, anti-F4/80, and anti-CX3CR1. Astrocytes were labeled with anti–GLAST-1. (B and C) qPCR analysis of IL-6 mRNA (B) and ELISA analysis of secreted IL-6 (C) in astrocyte (As.) and microglia (MG) cultures. Datum points represent the mean ± SD fold change (RQ) or IL-6 levels at each time point compared with untreated controls (UT). (D) ELISA analysis of secreted IL-6 in mixed glia, astrocyte, and microglia cultures treated with LPS. (E) ELISA analysis of secreted IL-6 in mixed glia and astrocyte cultures treated with TGF-β1 in the absence or presence of SB431542 (SB), SJN2511 (SJN), or SIS3, and in microglia cultures treated with TGF-β1 in the absence or presence of SB431542. The results shown represent one experiment out of at least three performed. *p < 0.05, **p < 0.01, ***p < 0.001 (one-way ANOVA test).

FIGURE 2.

TGF-β1 induces IL-6 upregulation primarily in astrocytes. Primary glial cultures were first purified to astrocytic and microglial cultures, and cells were then treated with 10 ng/ml TGF-β1. The kinetics of IL-6 expression were evaluated by qPCR and ELISA at 8, 24, and 48 h. (A) FACS analysis of mixed glia, purified microglia, and purified astrocytes. Microglia were labeled with and anti-CD11b, anti-F4/80, and anti-CX3CR1. Astrocytes were labeled with anti–GLAST-1. (B and C) qPCR analysis of IL-6 mRNA (B) and ELISA analysis of secreted IL-6 (C) in astrocyte (As.) and microglia (MG) cultures. Datum points represent the mean ± SD fold change (RQ) or IL-6 levels at each time point compared with untreated controls (UT). (D) ELISA analysis of secreted IL-6 in mixed glia, astrocyte, and microglia cultures treated with LPS. (E) ELISA analysis of secreted IL-6 in mixed glia and astrocyte cultures treated with TGF-β1 in the absence or presence of SB431542 (SB), SJN2511 (SJN), or SIS3, and in microglia cultures treated with TGF-β1 in the absence or presence of SB431542. The results shown represent one experiment out of at least three performed. *p < 0.05, **p < 0.01, ***p < 0.001 (one-way ANOVA test).

Close modal

To explore the mechanisms underlying the differential signaling of TGF-β1 in astrocytes and microglia, we first evaluated basal expression levels of TGF-βR1 and TGF-βR2 in cultured cells. qPCR analysis shows that the levels of both units of the TGF-βR are higher in microglia compared with astrocytes (5.175 ± 0.6- and 6.154 ± 0.59-fold increase of TGF-βR1 and TGF-βR2, respectively, p < 0.001; Fig. 3A) and similarly higher is the TGR-βR2 protein in microglia (Fig. 3B). We then exposed cultures of mixed glia, purified astrocytes, and purified microglia to TGF-β1 and quantified the expression, phosphorylation, and nuclear translocation of SMAD2/3 (compared with actin as a control) using Western blot and ICC analysis. In the presence of TGF-β1, no change is observed in total SMAD2/3 levels, which are higher in cultures of mixed glia and purified astrocytes compared with cultures of purified microglia (Fig. 3C, 3D). Concomitantly, significantly increased levels of p-SMAD2 are observed in purified astrocyte cultures than in purified microglia cultures (Fig. 3E).

FIGURE 3.

Astrocytes show more pronounced SMAD2/3 phosphorylation and nuclear translocation after TGF-β1 treatment compared with microglia. (A and B) Primary glial cultures were purified to astrocytes and microglia as described in Fig. 2 and analyzed for TGF-βR1 and TGF-βR2 expression levels using qPCR and Western blot. RNA and whole-cell lysates were prepared from purified astrocytes and purified microglia after CD11b separation. (A) qPCR analysis performed for TGF-βR1 and TGF-βR2 in purified astrocytes and microglia. (B) Western blot analysis performed for TGF-βR2 in purified astrocytes and microglia. The black line indicates joined parts of the image. Bars display TGF-βR2 to actin ratio in both astrocytes and microglia. (C and D) Primary glial cultures were treated with TGF-β1 and analyzed for SMAD2 and SMAD3 phosphorylation and nuclear translocation with Western blot and ICC. Whole-cell lysates were prepared following 1 h treatment with TGF-β1 of mixed glia, astrocyte, and microglia cultures. Untreated cultures were used as controls. (C) Western blot analysis performed for total SMAD2/3 and p-SMAD2 in mixed glia, purified astrocyte, and purified microglia cell cultures. (D and E) Quantification analysis of total SMAD2 (D) and p-SMAD2 (E) as their ratio (mean ± SD) to actin in each culture. (FI) Immunocytochemical analysis of mixed glia. (F and G) Cultures were treated with TGF-β1 for 1 h and then coimmunolabeled with anti-CD68 or anti-GFAP together with anti-SMAD2/3 (F) or anti–p-SMAD3 (G). Scale bars in each set represent 20 μm and apply to the entire set. (H and I) Quantification analysis of nuclear SMAD2/3 (H) and p-SMAD3 (I) in CD68+ microglia and GFAP+ astrocytes. Bars represent nuclear fluorescence intensity of SMAD2/3 in microglia and astrocytes (H) and fold change of p-SMAD3 (I) in TGF-β1–treated compared with untreated glial cultures. The results shown represent one experiment out of at least three performed. **p < 0.01, ***p < 0.001 (two-tailed Student t test).

FIGURE 3.

Astrocytes show more pronounced SMAD2/3 phosphorylation and nuclear translocation after TGF-β1 treatment compared with microglia. (A and B) Primary glial cultures were purified to astrocytes and microglia as described in Fig. 2 and analyzed for TGF-βR1 and TGF-βR2 expression levels using qPCR and Western blot. RNA and whole-cell lysates were prepared from purified astrocytes and purified microglia after CD11b separation. (A) qPCR analysis performed for TGF-βR1 and TGF-βR2 in purified astrocytes and microglia. (B) Western blot analysis performed for TGF-βR2 in purified astrocytes and microglia. The black line indicates joined parts of the image. Bars display TGF-βR2 to actin ratio in both astrocytes and microglia. (C and D) Primary glial cultures were treated with TGF-β1 and analyzed for SMAD2 and SMAD3 phosphorylation and nuclear translocation with Western blot and ICC. Whole-cell lysates were prepared following 1 h treatment with TGF-β1 of mixed glia, astrocyte, and microglia cultures. Untreated cultures were used as controls. (C) Western blot analysis performed for total SMAD2/3 and p-SMAD2 in mixed glia, purified astrocyte, and purified microglia cell cultures. (D and E) Quantification analysis of total SMAD2 (D) and p-SMAD2 (E) as their ratio (mean ± SD) to actin in each culture. (FI) Immunocytochemical analysis of mixed glia. (F and G) Cultures were treated with TGF-β1 for 1 h and then coimmunolabeled with anti-CD68 or anti-GFAP together with anti-SMAD2/3 (F) or anti–p-SMAD3 (G). Scale bars in each set represent 20 μm and apply to the entire set. (H and I) Quantification analysis of nuclear SMAD2/3 (H) and p-SMAD3 (I) in CD68+ microglia and GFAP+ astrocytes. Bars represent nuclear fluorescence intensity of SMAD2/3 in microglia and astrocytes (H) and fold change of p-SMAD3 (I) in TGF-β1–treated compared with untreated glial cultures. The results shown represent one experiment out of at least three performed. **p < 0.01, ***p < 0.001 (two-tailed Student t test).

Close modal

ICC analysis of glial cultures shows that the TGF-β1 treatment promotes nuclear translocation of SMAD2/3 that is significantly more colocalized with the astrocytic marker (GFAP) than with the microglial marker CD68 (Fig. 3F, 3H). Additionally, immunoreactivity of p-SMAD3 is evident only upon TGF-β1 treatment (Fig. 3G) and is significantly more intense in astrocytes than in microglia (Fig. 3I, p < 0.001).

Additional support for SMAD2/3-mediated IL-6 upregulation comes from analyzing the promoter region of IL-6 for SMAD-binding elements (see review in Ref. 28). We identified several putative SMAD3/4-binding sequences such as GTCT, AGAC, and CAGAC up to 2000 bp upstream to the IL-6 coding sequence. Whether these sequences indeed serve as SMAD3/4 binding sites that promote TGF-β1–induced IL-6 transcription in astrocytes is yet to be determined.

Overall, our data show that both astrocytes and microglia express TGF-βR1 and TGF-βR2 as well as SMAD2 and SMAD3, which undergo phosphorylation following TGF-β1 treatment. However, although higher levels of TGF-βR are expressed in microglia, the expression, phosphorylation, and nuclear translocation of SMAD2/3 are more abundant in astrocytes than in microglia and may thus be directly associated with the upregulation of IL-6 via the TGF-β1 signaling pathway.

Because both albumin and TGF-β1 induce epileptiform activity when applied onto brain slices (14), and because IL-6 was the primary cytokine upregulated by TGF-β1 in cultured astrocytes, we tested whether IL-6 is sufficient to induce hypersynchronous neuronal activity. We performed extracellular recordings in acute brain slices incubated for 2–4 h with IL-6 (100 ng/ml). Indeed, a brief white matter electrical stimulation induces hypersynchronous propagating epileptiform activity in 85% of all IL-6–treated slices but in none of the control slices (n = 7 and 5, respectively; p = 0.003; Fig. 4A, 4B). We next tested whether an intraventricular application of IL-6 can induce epileptiform activity in vivo. Either low (1 ng/h) or high (5 ng/h) concentrations of IL-6 (IL-6 low and IL-6 high, respectively) were injected to mice (n = 3 and 8, respectively) with an osmotic pump, and cortical activity was continuously monitored for 14 d thereafter. Controls were injected with ACSF (n = 6). Seizure-like activity (>5 s duration) was detected with a custom-built semiautomatic algorithm (see 2Materials and Methods).

FIGURE 4.

IL-6 induces ex vivo and in vivo epileptiform activity. (A) Evoked cortical field potentials were recorded in response to a single brief electrical stimulation. Slices were incubated for 2–4 h in ACSF (n = 6, top trace) in or 100 ng/ml IL-6 (n = 7, bottom trace). (B) The integral of the field potentials was calculated for 0- to 50-ms and for 50- to 200-ms intervals after stimulation. (C) Electroencephalogram showing a seizure-like event detected by the seizure detection algorithm in IL-6–treated mouse (5 ng/hr). Inserts show magnification of a region indicated by an asterisk. (D) Histograms describing the distribution of durations of detected seizures during 14 recording days. Mice were injected ICV with IL-6 at 1 ng/h (left panel) or 5 ng/h (right panel). (E) Mean ± SD duration of seizures detected in mice injected with low (1 ng/h) (n = 20) or high (5 ng/h) (n = 156) doses of IL-6 compared with control (ACSF-injected) mice (n = 3). (F) A temporal analysis showing the number of seizure-like events (SLEs) occurring at different days after surgery (day 0) in the different groups. *p < 0.05, **p < 0.01, ***p < 0.001 (two-tailed Student t test).

FIGURE 4.

IL-6 induces ex vivo and in vivo epileptiform activity. (A) Evoked cortical field potentials were recorded in response to a single brief electrical stimulation. Slices were incubated for 2–4 h in ACSF (n = 6, top trace) in or 100 ng/ml IL-6 (n = 7, bottom trace). (B) The integral of the field potentials was calculated for 0- to 50-ms and for 50- to 200-ms intervals after stimulation. (C) Electroencephalogram showing a seizure-like event detected by the seizure detection algorithm in IL-6–treated mouse (5 ng/hr). Inserts show magnification of a region indicated by an asterisk. (D) Histograms describing the distribution of durations of detected seizures during 14 recording days. Mice were injected ICV with IL-6 at 1 ng/h (left panel) or 5 ng/h (right panel). (E) Mean ± SD duration of seizures detected in mice injected with low (1 ng/h) (n = 20) or high (5 ng/h) (n = 156) doses of IL-6 compared with control (ACSF-injected) mice (n = 3). (F) A temporal analysis showing the number of seizure-like events (SLEs) occurring at different days after surgery (day 0) in the different groups. *p < 0.05, **p < 0.01, ***p < 0.001 (two-tailed Student t test).

Close modal

Significantly more IL-6–treated mice show distinct paroxysmal seizure-like activity compared with control mice (82 and 16.67%, respectively; p < 0.05). Moreover, the number and duration of seizures depend on IL-6 dose. In comparison with control mice (n = 6; three seizures overall, each duration was limited to 5 s), recordings from mice treated with a low dose of IL-6 (1 ng/h) revealed overall 20 seizures with a mean duration of 16.6 s (n = 3; median duration, 5 s; range, 5–60 s; p < 0.05), whereas recordings from high-dose IL-6–treated mice (5 ng/h) revealed overall 156 seizures with a mean duration of 30.4 s (n = 8; median duration, 16 s; range, 5–60 s; p < 0.01) (Fig. 4C–E). Additionally, seizure amplitude is higher in IL-6 high compared with IL-6 low mice (3.72 ± 3.47 and 1.52 ± 1.42 μV2, respectively; p < 0.01). A temporal analysis (Fig. 4F) shows that the earliest seizures are detected 3 d following IL-6 injection, after which their frequency increases during the first week and decreases during the second week after injection.

We next used somatic whole-cell recordings from neocortical pyramidal neurons to test the direct effect of an intracerebroventricular (ICV) injection of IL-6 on neuronal activity. Recordings were made during epileptogenesis, that is, 72–88 h after implanting the ICV pumps and prior to the appearance of the first seizure.

Passive membrane properties are similar between ACSF-treated controls and IL-6–injected mice (n = 8 and 10 neurons, respectively), including resting membrane potential (−60.6 ± 1.8 and −61.1 ± 2.9 mV, respectively; p > 0.05) and input resistance (133 ± 29 and 149 ± 66 MΩ, respectively; p > 0.5). Action potential properties are also similar between controls (n = 8) and IL-6–injected mice (n = 11), including the threshold (−47.4 ± 3.8 and −47.0 ± 5.9 mV, respectively; p > 0.85), the width at −20 mV (1.41 ± 0.09 and 1.42 ± 0.12 ms, respectively), and the amplitude (99 ± 5 and 95 ± 7 mV, respectively). No difference between the groups is observed also in repetitive firing properties: 150-pA, 300-ms-long current pulses elicit, on average, 5.3 ± 1.4 and 4.7 ± 0.9 action potentials in neurons from control (n = 7) and in IL-6–treated mice (n = 9), respectively (p > 0.25). Current clamp recordings likewise reveal no gross difference in the amplitude and frequency of spontaneous synaptic events at the level of resting potential. The minimal stimulus intensity required to elicit an orthodromic action potential is significantly lower in neurons from IL-6–treated mice compared with controls (10 ± 4 and 15 ± 5 V, n = 7 and 10, respectively; p < 0.02). Notably, a robust difference is found in the membrane potential time course following repetitive (50 Hz, 1 s) afferent extracellular stimulation (Fig. 5A–C). While membrane potential in neurons from control mice rapidly returns to resting values, often with a tendency toward a small after-hyperpolarization, a prolonged (up to 4 s) poststimulus depolarization is observed in most neurons from IL-6–treated cortices following extracellular (Fig. 5C) but not intracellular stimulation (Fig. 5D–F).

FIGURE 5.

White matter stimulation induces prolonged after-depolarization in pyramidal neurons from IL-6–treated mice. (A) Responses to white matter stimulation (50 Hz, 1 s) recorded in whole-cell current clamp from representative layer V cortical neurons obtained from one IL-6–treated mouse (5 μg/ml) (top) and one control (ACSF-treated) (bottom) mouse. (B) Magnification of the dashed regions shown in (A) (black, IL-6 treated; gray, ACSF treated). The broken gray line corresponds to the resting membrane potential (Erest). (C) The average poststimulus potential (PSP) between 0.2 and 0.3 s after termination of synaptic stimulation in neurons obtained from control and IL-6–treated mice. The thick horizontal bar indicates the median, the box limits the 25th and 75th percentiles, and the whiskers indicate the range. (D) Responses to intracellular stimulation (50 Hz, 1 nA, 1 s overall, 5-ms pulses) recorded in whole-cell current clamp from representative layer V cortical neurons obtained from an IL-6 and a control mouse (top and bottom traces, respectively). (E) Magnification of the dashed regions shown in D (black, IL-6 treated; gray, ACSF treated). The broken gray line corresponds to the resting membrane potential (Erest). (F) The average poststimulus potential (PSP) between 0.2 and 0.3 s after termination of synaptic stimulation in neurons obtained from control and IL-6–treated mice. Annotations same as in (C). **p < 0.01 (two-tailed Student t test).

FIGURE 5.

White matter stimulation induces prolonged after-depolarization in pyramidal neurons from IL-6–treated mice. (A) Responses to white matter stimulation (50 Hz, 1 s) recorded in whole-cell current clamp from representative layer V cortical neurons obtained from one IL-6–treated mouse (5 μg/ml) (top) and one control (ACSF-treated) (bottom) mouse. (B) Magnification of the dashed regions shown in (A) (black, IL-6 treated; gray, ACSF treated). The broken gray line corresponds to the resting membrane potential (Erest). (C) The average poststimulus potential (PSP) between 0.2 and 0.3 s after termination of synaptic stimulation in neurons obtained from control and IL-6–treated mice. The thick horizontal bar indicates the median, the box limits the 25th and 75th percentiles, and the whiskers indicate the range. (D) Responses to intracellular stimulation (50 Hz, 1 nA, 1 s overall, 5-ms pulses) recorded in whole-cell current clamp from representative layer V cortical neurons obtained from an IL-6 and a control mouse (top and bottom traces, respectively). (E) Magnification of the dashed regions shown in D (black, IL-6 treated; gray, ACSF treated). The broken gray line corresponds to the resting membrane potential (Erest). (F) The average poststimulus potential (PSP) between 0.2 and 0.3 s after termination of synaptic stimulation in neurons obtained from control and IL-6–treated mice. Annotations same as in (C). **p < 0.01 (two-tailed Student t test).

Close modal

To determine whether IL-6 is upregulated in brains of mice during the induction of epileptogeneis, both mouse strains FVB/N and C57BL6 exhibiting a general high and low sensitivity to seizure induction, respectively, were ICV injected with TGF-β1. FVB/N mice (8–10 wk of age) were implanted with TGF-β1–containing (100 ng/ml, 0.1 ng/h) or ACSF-containing osmotic pumps. After 72 h, brains were dissected into hippocampi and cortices and examined for mRNA levels of IL-6, IL-1β, and TNF-α with qPCR as described in 2Materials and Methods. As shown in Fig. 6A–C, compared with untreated mice, all three proinflammatory genes are significantly upregulated in the hippocampi of TGF-β1–injected mice (IL-6, 3.53 ± 2.39, p < 0.01; IL-1β, 19.46 ± 17.12, p < 0.01; and TNF-α, 4 ± 2.11, p < 0.001). The slight upregulation of IL-6, IL-1β, and TNF-α obtained in the hippocampi of the ACSF group at 3 d after injection resembles the normal inflammatory response to injury (Fig. 6A–C). At 3 d after injection, TGF-β1 does not upregulate IL-6, IL-1β, and TNF-α in the cortex over the levels induced by ACSF (data not shown). Notably, in contrast to FVB/N mice, C57BL/6 mice were resistant to TGF-β1–induced IL-6 upregulation at both low (100 ng/ml, 0.1 ng/h for 24–48 and 72 h) and high (1000 ng/ml, 1 ng/h for 72 h) doses (Fig 6D, 6E), suggesting that immunoregulatory mechanisms controlling the proinflammatory role of TGF-β1 are more stringent in C57BL/6 mice than in FVB/N mice and may partially underlie findings demonstrating the relative resistance of C57BL/6, but not FVB/N mice, to seizure-induced cell death (29, 30).

FIGURE 6.

TGF-β1–induced IL-6 secretion is genetic background–dependent. (AC) FVBN mice (8-10 wk old) were implanted with 100 ng/ml TGF-β1–containing (n = 10) or ACSF-containing (n = 10) osmotic pumps. (D and E) C57BL/6 mice were implanted with osmotic pumps containing either 100 ng/ml TGF-β1 (D, n = 4), 1000 ng/ml TGF-β1 (E, n = 4), or ACSF (n = 4). Untreated mice served as control (FVB/N, n = 10; C57BL/6, n = 3). At 72 h after surgery, brains were removed and hippocampi were dissected. RNA was extracted from the tissues and analyzed for IL-6, IL-1β, and TNF-α gene expression using qPCR. *p < 0.05, **p < 0.01, ***p < 0.001 (one-way ANOVA test).

FIGURE 6.

TGF-β1–induced IL-6 secretion is genetic background–dependent. (AC) FVBN mice (8-10 wk old) were implanted with 100 ng/ml TGF-β1–containing (n = 10) or ACSF-containing (n = 10) osmotic pumps. (D and E) C57BL/6 mice were implanted with osmotic pumps containing either 100 ng/ml TGF-β1 (D, n = 4), 1000 ng/ml TGF-β1 (E, n = 4), or ACSF (n = 4). Untreated mice served as control (FVB/N, n = 10; C57BL/6, n = 3). At 72 h after surgery, brains were removed and hippocampi were dissected. RNA was extracted from the tissues and analyzed for IL-6, IL-1β, and TNF-α gene expression using qPCR. *p < 0.05, **p < 0.01, ***p < 0.001 (one-way ANOVA test).

Close modal

The present study sought to elucidate the mechanisms underlying TGF-β signaling–induced epileptogenesis in mice. Using glial cultures, we demonstrate that TGF-β1 induces an astrocyte-specific response that includes a differential SMAD2/3 phosphorylation and nuclear translocation and a rapid upregulation and secretion of IL-6. Interestingly, glial response to LPS is significantly different, with a robust, TLR4–mediated IL-6 secretion from microglia. Finally, electrophysiological experiments support a direct epileptogenic effect of IL-6 both ex vivo and in vivo. IL-6–induced neuronal hyperexcitability was associated with a prolonged depolarizing afterpotential following repetitive orthodromic activation with no apparent changes in intrinsic membrane properties in the recorded neurons. Thus, overall, we suggest that under conditions in which vascular insult and BBB dysfunction occur, TGF-β signaling–induced early upregulation of IL-6 in astrocytes is sufficient to trigger epileptogenesis. Owing to the resistance to TGF-β1–induced IL-6 upregulation in C57BL/6 mice in vivo, this pathway may be limited to certain genetic backgrounds and/or genetic predisposition.

We previously demonstrated TGF-β signaling and a robust inflammatory response in the brain of rats and FVB/N mice that were exposed to serum albumin (10, 11, 14, 16). Furthermore, TGF-β1 pathway blockers efficiently blocked epileptogenesis in BBB breakdown and albumin models of epilepsy in mice and rats (11, 16). Although TGF-β1 is a well-characterized immunoregulatory cytokine (19, 20, 25, 3134), it can also exert autoimmunity, glial activation, and brain inflammation (22, 23, 35, 36). In the present study, we demonstrate that glial exposure to TGF-β1 induces a rapid and robust IL-6 upregulation at both the mRNA and protein levels. Such upregulation of IL-6 by TGF-β1 seems to be a specific inflammatory response: it differs from the effect of LPS, occurs primarily in astrocytes (and not in microglia), and involves only a slight consequent upregulation of IL-1β and TNF-α. Our data demonstrate that both microglia and astrocytes express the TGF-β receptor as well as the trans-signaling molecules SMAD2 and SMAD3. However, whereas both subunits of the TGF-βR are expressed at higher levels in microglia, the expression, phosphorylation, and nuclear translocation of SMAD2/3 are significantly more abundant in astrocytes. Together with our findings that the SMAD3-specific inhibitor SIS3 markedly suppressed the upregulation of IL-6 in astrocytes, we suggest that SMAD3 is required for the enhanced expression of IL-6 in astrocytes presumably via direct binding to SMAD-binding elements in the IL-6 promoter region (28, 37, 38). The signals triggering SMAD3 expression and the molecular pathway of p-SMAD3–induced IL-6 upregulation, alone or with additional cofactors (37, 39), are yet to be characterized.

Of the milieu of proinflammatory cytokines tested in the present study, we found TGF-β1 to induce first, and most prominently, the upregulation of IL-6. Numerous studies demonstrated significant IL-6 levels in the CSF and blood serum of patients suffering from traumatic brain injuries or neurodegenerative processes (40). Interestingly, in most of these disorders, BBB dysfunction has been documented together with increasing likelihood for seizures. In this study, we show that IL-6 is sufficient to promote network hypersynchronization and epileptiform activity both ex vivo and in vivo and thus highlight IL-6 with the potential to play a key role in the development of seizures following brain insults and breakdown of the BBB. Expressed in the brain with receptors on all neural lineages, IL-6 was reported to be involved in neurogenesis and synaptic activity (see reviews in Refs. 40, 41). A significant body of evidence also shows that, similar to TGF-β1, IL-6 is upregulated in neurodegenerative processes with functions ranging from protective to detrimental. As such, IL-6 has been shown to either protect from excitotoxicity in several in vitro and in vivo models, or to enhance N-methyl-d-asparate–induced excitotoxicity in cerebellar granule neurons (40). Notably, transient inhibition of STAT3 phosphorylation (a signaling cascade induced by IL-6) in a rat model of pilocarpine-induced status epilepticus has recently been shown to significantly reduce disease severity along with downregulation of STAT3-regulated genes (42). Our current investigation demonstrates that brain exposure to IL-6 is sufficient to induce seizures in vivo with duration and amplitude depending on the dose of injected IL-6. Interestingly, ICV administration of TGF-β1 induces a significant IL-6 upregulation in FVB/N but not in C57BL/6 mice. These results may at least partially underlie previous studies showing that C57BL/6 mice exhibit a general low sensitivity to seizure induction (4345) and that FVB/N mice have a lower threshold than do C57BL/6 mice for seizure-induced cell death (29, 30, 46), further suggesting that IL-6 upregulation plays an important role in the brain inflammatory response associated with TGF-β–induced epileptogenesis.

The mechanisms underlying neuronal hypersynchronization and epileptogenesis following exposure to IL-6 are not yet fully understood. Increase in intrinsic excitability (47), selective excitatory synaptogenesis (48), and reduction in inhibitory transmission (14) were all reported to occur during epileptogenesis. Our recordings from neocortical pyramidal neurons indicate activity-dependent increase in excitability that does not require a prominent change in intrinsic properties and/or spontaneous synaptic transmission, and is probably astrocyte mediated (10). Indeed, recent findings from human epileptic tissue and animal models of epilepsy suggest a key role for astrocytic dysfunction in epileptogenesis, seizure generation, and seizure propagation (10, 12, 49, 50). In particular, a significant role has been suggested for the downregulation of glial inward rectifying potassium (Kir) channel 4.1, which underlies impaired buffering of extracellular potassium (10, 12, 13). Because our previous molecular and physiological data in mouse and rat models of albumin- and TGF-β–induced seizures showed early activation of astrocytes (3, 10), and specifically a robust excessive extracellular potassium accumulation upon repetitive stimulation at physiologically relevant frequencies (10–50 Hz) (10), we tested a potential role of such stimulation on neuronal excitability in the IL-6–treated cortices. Our recordings demonstrate a long-lasting depolarization (∼8 mV) upon afferent but not intracellular stimulation, predicting a 25% increase in the accumulation of extracellular potassium during neuronal activation, consistent with the notion of astrocytic dysfunction and reduced potassium buffering after insult (51, 52) or during epileptogenesis (3, 10). Overall, although additional experiments in different neuronal populations and experimental conditions (e.g., voltage clamp experiments under different holding membrane potentials) are required to rule out additional changes in synaptic properties, our experiments at an early time point during epileptogenesis suggest that IL-6 is sufficient to facilitate stimulus-dependent neuronal depolarization and hyperexcitability, likely due to failure in buffering of extracellular potassium.

In summary, our study highlights a novel proinflammatory reaction, which does not essentially involve microglial activation but, rather, is executed via TGF-β1 signaling in astrocytes to promote the release of IL-6. This astrocytic response may be dependent on genetic background and is further associated with impaired extracellular homeostasis upon neuronal activation and altered neuronal–astrocytic interactions, leading to spontaneous seizures. Because increased CSF or blood serum levels of IL-6 have been documented in a variety of CNS injuries that are often followed by seizures (53), therapeutic strategies to knock down IL-6 expression and/or signaling in the CNS may prove beneficial.

We thank Dr. Ram Gal for valuable editorial comments.

This work was supported by Israel Science Foundation Grants 713/11 (to A.F.) and 531/11 (to A.M.), German Israeli Foundation Grant 124/2008 (to A.F.), National Institute for Neurological Disorders and Stroke Grant 1rO1N5066005 (to A.F.), European Union’s Seventh Framework Program FP7/2007–2013 Grant 602102 (EPITARGET, to A.F.), and by a local institutional grant (Ben-Gurion University of the Negev, Faculty of Health) (to A.M. and A.F.).

Abbreviations used in this article:

ACSF

artificial cerebrospinal fluid

BBB

blood–brain barrier

GFAP

glial fibrillary acidic protein

ICC

immunocytochemistry

ICV

intracerebroventricular

qPCR

quantitative PCR.

1
Engel
J. J.
, and
Pedley
T. A.
.
2007
.
Introduction: what is epilepsy?
In Epilepsy: A Comprehensive Textbook.
Engel
J.
 Jr.
,
Pedley
T.
, eds.
Lippincot Williams & Wilkins
,
Philadelphia
,
p. 54
66
.
2
Hesdorffer
D. C.
2007
.
Risk factors
. In
Epilepsy: A Comprehensive Textbook.
Engel
J.
 Jr.
,
Pedley
T.
, eds.
Lippincot Williams & Wilkins
,
Philadelphia
,
p. 156
170
.
3
Ivens
S.
,
Kaufer
D.
,
Flores
L. P.
,
Bechmann
I.
,
Zumsteg
D.
,
Tomkins
O.
,
Seiffert
E.
,
Heinemann
U.
,
Friedman
A.
.
2007
.
TGF-β receptor-mediated albumin uptake into astrocytes is involved in neocortical epileptogenesis.
Brain
130
:
535
547
.
4
Janigro
D.
1999
.
Blood-brain barrier, ion homeostatis and epilepsy: possible implications towards the understanding of ketogenic diet mechanisms.
Epilepsy Res.
37
:
223
232
.
5
Seiffert
E.
,
Dreier
J. P.
,
Ivens
S.
,
Bechmann
I.
,
Tomkins
O.
,
Heinemann
U.
,
Friedman
A.
.
2004
.
Lasting blood-brain barrier disruption induces epileptic focus in the rat somatosensory cortex.
J. Neurosci.
24
:
7829
7836
.
6
Tomkins
O.
,
Friedman
O.
,
Ivens
S.
,
Reiffurth
C.
,
Major
S.
,
Dreier
J. P.
,
Heinemann
U.
,
Friedman
A.
.
2007
.
Blood-brain barrier disruption results in delayed functional and structural alterations in the rat neocortex.
Neurobiol. Dis.
25
:
367
377
.
7
Tomkins
O.
,
Shelef
I.
,
Kaizerman
I.
,
Eliushin
A.
,
Afawi
Z.
,
Misk
A.
,
Gidon
M.
,
Cohen
A.
,
Zumsteg
D.
,
Friedman
A.
.
2008
.
Blood-brain barrier disruption in post-traumatic epilepsy.
J. Neurol. Neurosurg. Psychiatry
79
:
774
777
.
8
Abbott
N. J.
,
Friedman
A.
.
2012
.
Overview and introduction: the blood-brain barrier in health and disease.
Epilepsia
53
(
Suppl. 6
):
1
6
.
9
Shlosberg
D.
,
Benifla
M.
,
Kaufer
D.
,
Friedman
A.
.
2010
.
Blood-brain barrier breakdown as a therapeutic target in traumatic brain injury.
Nat Rev Neurol
6
:
393
403
.
10
David
Y.
,
Cacheaux
L. P.
,
Ivens
S.
,
Lapilover
E.
,
Heinemann
U.
,
Kaufer
D.
,
Friedman
A.
.
2009
.
Astrocytic dysfunction in epileptogenesis: consequence of altered potassium and glutamate homeostasis?
J. Neurosci.
29
:
10588
10599
.
11
Weissberg
I.
,
Wood
L.
,
Kamintsky
L.
,
Vazquez
O.
,
Milikovsky
D. Z.
,
Alexander
A.
,
Oppenheim
H.
,
Ardizzone
C.
,
Becker
A.
,
Frigerio
F.
, et al
.
2015
.
Albumin induces excitatory synaptogenesis through astrocytic TGF-β/ALK5 signaling in a model of acquired epilepsy following blood-brain barrier dysfunction.
Neurobiol. Dis.
78
:
115
125
.
12
Heinemann
U.
,
Kaufer
D.
,
Friedman
A.
.
2012
.
Blood-brain barrier dysfunction, TGFβ signaling, and astrocyte dysfunction in epilepsy.
Glia
60
:
1251
1257
.
13
Steinhäuser
C.
,
Seifert
G.
,
Bedner
P.
.
2012
.
Astrocyte dysfunction in temporal lobe epilepsy: K+ channels and gap junction coupling.
Glia
60
:
1192
1202
.
14
Cacheaux
L. P.
,
Ivens
S.
,
David
Y.
,
Lakhter
A. J.
,
Bar-Klein
G.
,
Shapira
M.
,
Heinemann
U.
,
Friedman
A.
,
Kaufer
D.
.
2009
.
Transcriptome profiling reveals TGF-β signaling involvement in epileptogenesis.
J. Neurosci.
29
:
8927
8935
.
15
Frigerio
F.
,
Frasca
A.
,
Weissberg
I.
,
Parrella
S.
,
Friedman
A.
,
Vezzani
A.
,
Noé
F. M.
.
2012
.
Long-lasting pro-ictogenic effects induced in vivo by rat brain exposure to serum albumin in the absence of concomitant pathology.
Epilepsia
53
:
1887
1897
.
16
Bar-Klein
G.
,
Cacheaux
L. P.
,
Kamintsky
L.
,
Prager
O.
,
Weissberg
I.
,
Schoknecht
K.
,
Cheng
P.
,
Kim
S. Y.
,
Wood
L.
,
Heinemann
U.
, et al
.
2014
.
Losartan prevents acquired epilepsy via TGF-β signaling suppression.
Ann. Neurol.
75
:
864
875
.
17
Vezzani
A.
,
French
J.
,
Bartfai
T.
,
Baram
T. Z.
.
2011
.
The role of inflammation in epilepsy.
Nat Rev Neurol
7
:
31
40
.
18
Vezzani
A.
,
Viviani
B.
.
2015
.
Neuromodulatory properties of inflammatory cytokines and their impact on neuronal excitability.
Neuropharmacology
96 (Part A): 70–82
.
19
Abutbul
S.
,
Shapiro
J.
,
Szaingurten-Solodkin
I.
,
Levy
N.
,
Carmy
Y.
,
Baron
R.
,
Jung
S.
,
Monsonego
A.
.
2012
.
TGF-β signaling through SMAD2/3 induces the quiescent microglial phenotype within the CNS environment.
Glia
60
:
1160
1171
.
20
Brionne
T. C.
,
Tesseur
I.
,
Masliah
E.
,
Wyss-Coray
T.
.
2003
.
Loss of TGF-β1 leads to increased neuronal cell death and microgliosis in mouse brain.
Neuron
40
:
1133
1145
.
21
Lanz
T. V.
,
Ding
Z.
,
Ho
P. P.
,
Luo
J.
,
Agrawal
A. N.
,
Srinagesh
H.
,
Axtell
R.
,
Zhang
H.
,
Platten
M.
,
Wyss-Coray
T.
,
Steinman
L.
.
2010
.
Angiotensin II sustains brain inflammation in mice via TGF-β.
J. Clin. Invest.
120
:
2782
2794
.
22
Luo
J.
,
Ho
P. P.
,
Buckwalter
M. S.
,
Hsu
T.
,
Lee
L. Y.
,
Zhang
H.
,
Kim
D. K.
,
Kim
S. J.
,
Gambhir
S. S.
,
Steinman
L.
,
Wyss-Coray
T.
.
2007
.
Glia-dependent TGF-β signaling, acting independently of the TH17 pathway, is critical for initiation of murine autoimmune encephalomyelitis.
J. Clin. Invest.
117
:
3306
3315
.
23
Luo
J.
,
Lin
A. H.
,
Masliah
E.
,
Wyss-Coray
T.
.
2006
.
Bioluminescence imaging of Smad signaling in living mice shows correlation with excitotoxic neurodegeneration.
Proc. Natl. Acad. Sci. USA
103
:
18326
18331
.
24
Tesseur
I.
,
Zou
K.
,
Esposito
L.
,
Bard
F.
,
Berber
E.
,
Can
J. V.
,
Lin
A. H.
,
Crews
L.
,
Tremblay
P.
,
Mathews
P.
, et al
.
2006
.
Deficiency in neuronal TGF-β signaling promotes neurodegeneration and Alzheimer’s pathology.
J. Clin. Invest.
116
:
3060
3069
.
25
Town
T.
,
Laouar
Y.
,
Pittenger
C.
,
Mori
T.
,
Szekely
C. A.
,
Tan
J.
,
Duman
R. S.
,
Flavell
R. A.
.
2008
.
Blocking TGF-β-Smad2/3 innate immune signaling mitigates Alzheimer-like pathology.
Nat. Med.
14
:
681
687
.
26
Fleidervish
I. A.
,
Libman
L.
,
Katz
E.
,
Gutnick
M. J.
.
2008
.
Endogenous polyamines regulate cortical neuronal excitability by blocking voltage-gated Na+ channels.
Proc. Natl. Acad. Sci. USA
105
:
18994
18999
.
27
Weissberg
I.
,
Reichert
A.
,
Heinemann
U.
,
Friedman
A.
.
2011
.
Blood-brain barrier dysfunction in epileptogenesis of the temporal lobe.
Epilepsy Res. Treat.
2011
:
143908
.
28
Morikawa
M.
,
Koinuma
D.
,
Miyazono
K.
,
Heldin
C. H.
.
2013
.
Genome-wide mechanisms of Smad binding.
Oncogene
32
:
1609
1615
.
29
McLin
J. P.
,
Steward
O.
.
2006
.
Comparison of seizure phenotype and neurodegeneration induced by systemic kainic acid in inbred, outbred, and hybrid mouse strains.
Eur. J. Neurosci.
24
:
2191
2202
.
30
Schauwecker
P. E.
2012
.
Strain differences in seizure-induced cell death following pilocarpine-induced status epilepticus.
Neurobiol. Dis.
45
:
297
304
.
31
Zhu
Y.
,
Yang
G. Y.
,
Ahlemeyer
B.
,
Pang
L.
,
Che
X. M.
,
Culmsee
C.
,
Klumpp
S.
,
Krieglstein
J.
.
2002
.
Transforming growth factor-β1 increases bad phosphorylation and protects neurons against damage.
J. Neurosci.
22
:
3898–
-
3909
.
32
Qian
L.
,
Wei
S. J.
,
Zhang
D.
,
Hu
X.
,
Xu
Z.
,
Wilson
B.
,
El-Benna
J.
,
Hong
J. S.
,
Flood
P. M.
.
2008
.
Potent anti-inflammatory and neuroprotective effects of TGF-β1 are mediated through the inhibition of ERK and p47phox-Ser345 phosphorylation and translocation in microglia.
J. Immunol.
181
:
660
668
.
33
Makwana
M.
,
Jones
L. L.
,
Cuthill
D.
,
Heuer
H.
,
Bohatschek
M.
,
Hristova
M.
,
Friedrichsen
S.
,
Ormsby
I.
,
Bueringer
D.
,
Koppius
A.
, et al
.
2007
.
Endogenous transforming growth factor β1 suppresses inflammation and promotes survival in adult CNS.
J. Neurosci.
27
:
11201
11213
.
34
Naert
G.
,
Rivest
S.
.
2011
.
CC chemokine receptor 2 deficiency aggravates cognitive impairments and amyloid pathology in a transgenic mouse model of Alzheimer’s disease.
J. Neurosci.
31
:
6208
6220
.
35
Wyss-Coray
T.
,
Borrow
P.
,
Brooker
M. J.
,
Mucke
L.
.
1997
.
Astroglial overproduction of TGF-β1 enhances inflammatory central nervous system disease in transgenic mice.
J. Neuroimmunol.
77
:
45
50
.
36
Wyss-Coray
T.
,
Loike
J. D.
,
Brionne
T. C.
,
Lu
E.
,
Anankov
R.
,
Yan
F.
,
Silverstein
S. C.
,
Husemann
J.
.
2003
.
Adult mouse astrocytes degrade amyloid-β in vitro and in situ.
Nat. Med.
9
:
453
457
.
37
Feng
X. H.
,
Derynck
R.
.
2005
.
Specificity and versatility in TGF-β signaling through Smads.
Annu. Rev. Cell Dev. Biol.
21
:
659
693
.
38
Mullen
A. C.
,
Orlando
D. A.
,
Newman
J. J.
,
Lovén
J.
,
Kumar
R. M.
,
Bilodeau
S.
,
Reddy
J.
,
Guenther
M. G.
,
DeKoter
R. P.
,
Young
R. A.
.
2011
.
Master transcription factors determine cell-type-specific responses to TGF-β signaling.
Cell
147
:
565
576
.
39
Zhang
Y.
,
Feng
X. H.
,
Derynck
R.
.
1998
.
Smad3 and Smad4 cooperate with c-Jun/c-Fos to mediate TGF-β-induced transcription.
Nature
394
:
909
913
.
40
Erta
M.
,
Quintana
A.
,
Hidalgo
J.
.
2012
.
Interleukin-6, a major cytokine in the central nervous system.
Int. J. Biol. Sci.
8
:
1254
1266
.
41
Spooren
A.
,
Kolmus
K.
,
Laureys
G.
,
Clinckers
R.
,
De Keyser
J.
,
Haegeman
G.
,
Gerlo
S.
.
2011
.
Interleukin-6, a mental cytokine.
Brain Res. Brain Res. Rev.
67
:
157
183
.
42
Grabenstatter
H. L.
,
Del Angel
Y. C.
,
Carlsen
J.
,
Wempe
M. F.
,
White
A. M.
,
Cogswell
M.
,
Russek
S. J.
,
Brooks-Kayal
A. R.
.
2014
.
The effect of STAT3 inhibition on status epilepticus and subsequent spontaneous seizures in the pilocarpine model of acquired epilepsy.
Neurobiol. Dis.
62
:
73
85
.
43
Bankstahl
M.
,
Müller
C. J.
,
Wilk
E.
,
Schughart
K.
,
Löscher
W.
.
2012
.
Generation and characterization of pilocarpine-sensitive C57BL/6 mice as a model of temporal lobe epilepsy.
Behav. Brain Res.
230
:
182
191
.
44
Yang
J.
,
Houk
B.
,
Shah
J.
,
Hauser
K. F.
,
Luo
Y.
,
Smith
G.
,
Schauwecker
E.
,
Barnes
G. N.
.
2005
.
Genetic background regulates semaphorin gene expression and epileptogenesis in mouse brain after kainic acid status epilepticus.
Neuroscience
131
:
853
869
.
45
Borges
K.
,
Gearing
M.
,
McDermott
D. L.
,
Smith
A. B.
,
Almonte
A. G.
,
Wainer
B. H.
,
Dingledine
R.
.
2003
.
Neuronal and glial pathological changes during epileptogenesis in the mouse pilocarpine model.
Exp. Neurol.
182
:
21
34
.
46
Schauwecker
P. E.
,
Steward
O.
.
1997
.
Genetic determinants of susceptibility to excitotoxic cell death: implications for gene targeting approaches.
Proc. Natl. Acad. Sci. USA
94
:
4103
4108
.
47
Becker
A. J.
,
Pitsch
J.
,
Sochivko
D.
,
Opitz
T.
,
Staniek
M.
,
Chen
C. C.
,
Campbell
K. P.
,
Schoch
S.
,
Yaari
Y.
,
Beck
H.
.
2008
.
Transcriptional upregulation of Cav3.2 mediates epileptogenesis in the pilocarpine model of epilepsy.
J. Neurosci.
28
:
13341
13353
.
48
Li
H.
,
Graber
K. D.
,
Jin
S.
,
McDonald
W.
,
Barres
B. A.
,
Prince
D. A.
.
2012
.
Gabapentin decreases epileptiform discharges in a chronic model of neocortical trauma.
Neurobiol. Dis.
48
:
429
438
.
49
Carlen
P. L.
2012
.
Curious and contradictory roles of glial connexins and pannexins in epilepsy.
Brain Res.
1487
:
54
60
.
50
Steinhäuser
C.
,
Boison
D.
.
2012
.
Epilepsy: crucial role for astrocytes.
Glia
60
:
1191
.
51
Anderová
M.
,
Antonova
T.
,
Petrík
D.
,
Neprasová
H.
,
Chvátal
A.
,
Syková
E.
.
2004
.
Voltage-dependent potassium currents in hypertrophied rat astrocytes after a cortical stab wound.
Glia
48
:
311
326
.
52
D'Ambrosio
R.
,
Maris
D. O.
,
Grady
M. S.
,
Winn
H. R.
,
Janigro
D.
.
1999
.
Impaired K+ homeostasis and altered electrophysiological properties of post-traumatic hippocampal glia.
J. Neurosci.
19
:
8152
8162
.
53
Vossel
K. A.
,
Beagle
A. J.
,
Rabinovici
G. D.
,
Shu
H.
,
Lee
S. E.
,
Naasan
G.
,
Hegde
M.
,
Cornes
S. B.
,
Henry
M. L.
,
Nelson
A. B.
, et al
.
2013
.
Seizures and epileptiform activity in the early stages of Alzheimer disease.
JAMA Neurol.
70
:
1158
1166
.

The authors have no financial conflicts of interest.