Modulation of interferon-γ-induced glial cell activation by transforming growth factor β1: A role for STAT1 and MAPK pathways


  • Rodrigo Herrera-Molina,

    1. Departamento de Neurología, Laboratorio de Neurosciencias, Facultad de Medicina, Pontificia Universidad Católica de Chile, Santiago, Chile
    2. Department of Neurochemistry and Molecular Biology, Leibniz Institute for Neurobiology, Magdeburg, Germany
    Search for more papers by this author
    • Both authors contributed equally to this work.

  • Betsi Flores,

    1. Departamento de Neurología, Laboratorio de Neurosciencias, Facultad de Medicina, Pontificia Universidad Católica de Chile, Santiago, Chile
    Search for more papers by this author
    • Both authors contributed equally to this work.

  • Juan A. Orellana,

    1. Departamento de Neurología, Laboratorio de Neurosciencias, Facultad de Medicina, Pontificia Universidad Católica de Chile, Santiago, Chile
    Search for more papers by this author
  • Rommy von Bernhardi

    1. Departamento de Neurología, Laboratorio de Neurosciencias, Facultad de Medicina, Pontificia Universidad Católica de Chile, Santiago, Chile
    Search for more papers by this author

Address correspondence and reprint requests to Rommy von Bernhardi, Laboratorio de Neurosciencias, Facultad de Medicina, Pontificia Universidad Católica de Chile, Marcoleta 391, Santiago, CHILE. E-mail:


Overactivated glial cells can produce neurotoxic oxidant molecules such as nitric oxide (NO·) and superoxide anion (O2·). We have previously reported that transforming growth factor β1 (TGFβ1) released by hippocampal cells modulates interferon-γ (IFNγ)-induced production of O2· and NO· by glial cells. However, underlying molecular mechanisms are not completely understood, thereby, the aim of this work was to study the effect of TGFβ1 on IFNγ-induced signaling pathways. We found that costimulation with TGFβ1 decreased IFNγ-induced phosphorylation of signal transducer and activator of transcription-type-1 (STAT1) and extracellular signal-regulated kinase (ERK), which correlated with a reduced O2· and NO· production in mixed and purified glial cultures. Moreover, IFNγ caused a decrease in TGFβ1-mediated phosphorylation of P38, whereas pre-treatment with ERK and P38 inhibitors decreased IFNγ-induced phosphorylation of STAT1 on serine727 and production of radical species. These results suggested that modulation of glial activation by TGFβ1 is mediated by deactivation of MAPKs. Notably, TGFβ1 increased the levels of MAPK phosphatase-1 (MKP-1), whose participation in TGFβ1-mediated modulation was confirmed by MKP-1 siRNA transfection in mixed and purified glial cultures. Our results indicate that the cross-talk between IFNγ and TGFβ1 might regulate the activation of glial cells and that TGFβ1 modulated IFNγ-induced production of neurotoxic oxidant molecules through STAT1, ERK, and P38 pathways.

Abbreviations used

Dulbecco’s modified Eagle’s medium


dimethyl sulfoxide


extracellular signal-regulated kinase


glial fibrillary acidic protein




inducible nitric oxide synthase


Janus-activated kinase


c-Jun N-terminal kinase




mitogen-activated protein kinase


MAPK phosphatase-1


nitro blue tetrazolium


phosphate-buffered saline


phosphorylated ERK


phosphorylated P38


STAT1 phosphorylated on S727


STAT1 phosphorylated on Y701


signal transducer and activator of transcription-type-1


Tris-buffered saline


transforming growth factor β1

Neuroinflammation is a generally beneficial process mediated by the activation of glial cells in response to injury, illness, or infection. Astrocytes and microglia release inflammatory mediators like cytokines and radical species like nitric oxide (NO·) and superoxide anion (O2·) to eliminate the noxious agent and repair damaged tissues. Unfortunately, this process sometimes gets out of balance, as appears to occur in aging (von Bernhardi et al. 2010) or when neuroinflammation persists after removal of the triggering stimulus. Chronic neuroinflammation can become a self-perpetuating response including long-standing glial activation and sustained release of inflammatory cytokines as well as production of oxidative and nitrosative stress. Indeed, chronic neuroinflammation plays a key role in the progression of neurodegenerative diseases (Block and Hong 2005; von Bernhardi 2007; von Bernhardi et al. 2007; Gao and Hong 2008; Ramírez et al. 2008) as the release of NO· and O2· by glial cells induces neuronal injury (Thannickal and Fanburg 2000; Calabrese et al. 2007; von Bernhardi and Eugenín 2012) because of protein carbonylation, lipid peroxidation, and DNA oxidation (Christen 2000; Penkowa et al. 2000; Bazan et al. 2002; Adibhatla et al. 2003).

Glial cells-mediated radical species production involves cross-talk of a complex network of intracellular pathways triggered by inflammatory cytokines, such as interferon-γ (IFNγ). In response to IFNγ, glial cells produce NO· by up-regulation of inducible NO· synthase (iNOS) and also microglial cells release O2· by a nicotinamide adenine dinucleotide phosphate (NADPH)-oxidase-mediated mechanism (Hu et al. 1995; Calabrese et al. 2007). IFNγ potently activates microglia (Ng et al. 1999; Klegeris et al. 2005), and it has been shown to increase in the aged brain, although its endogenous cell source in the brain remains unidentified (Lyons et al. 2011). The main signaling pathway induced by IFNγ is the signal transducer and activator of transcription-type-1 (STAT1), which is activated by a Janus-activated kinase (JAK)-dependent phosphorylation on tyrosine Y701 (pSTAT1tyr) to translocate into the nucleus and induce gene expression (Platanias 2005; Gough et al. 2008). STAT1 full transcriptional activity requires a second phosphorylation on serine S727 (pSTAT1ser) (Wen et al. 1995). Other important pathways activated by IFNγ are mitogen-activated protein kinase (MAPK)s including extracellular signal-regulated protein kinases (ERKs), stress-activated protein kinases c-Jun N-terminal kinase (JNK), and p38 MAP kinase (P38). Activated MAPKs move in the cytoplasm or translocate into the nucleus phosphorylating transcription factors. Noteworthy, ERK and P38 appear to be key actors in the production of free radicals by glia (Bhat et al. 1998; Marcus et al. 2003; Qian et al. 2008), and we have reported that the ERK pathway is modulated by pro- and anti-inflammatory cytokines, regulating the timing of microglia activation (Saud et al. 2005). On the other hand, MAPK signaling is ended by a group of MAPK phosphatases (MKP), being MAPK phosphatase-1 (MKP-1) an archetypal member of this family (Liu et al. 2007; Boutros et al. 2008).

Transforming growth factor β1 (TGFβ1) is a cytokine that regulates multiple cellular processes, such as growth, apoptosis, and inflammation. Its downstream signaling involves Smad family members and MAPKs, although their activation is highly variable and cell type dependent (Schmierer and Hill 2007). In fact, there are reports showing that TGFβ1 modulates glial activity both inhibiting inflammatory cytokines and radical species production (Hu et al. 1995; Ledeboer et al. 2000; Lieb et al. 2003), as well as inducing NO· production when murine astrocytes were pre-treated with TGFβ1 for 24 h (Hamby et al. 2006, 2008). We have reported that TGFβ1 released by hippocampal cells decreases IFNγ-induced O2· and NO· production by glia (Herrera-Molina and von Bernhardi 2005; Saud et al. 2005). However, molecular mechanisms underlying these effects remain to be elucidated.

Here, we examined the effect of TGFβ1 over IFNγ-induced activation of signaling pathways in cultures of mixed and purified glial cells. Our results indicate that TGFβ1 regulates the IFNγ-induced production of radical species through the modulation of STAT1 and ERK1/2 activation. In addition, we report a novel mechanism to explain the regulatory effect of TGFβ1 on neuroinflammation, through the induction of MKP-1 mainly in microglial cells. Moreover, IFNγ decreased TGFβ1-induced activation of P38 suggesting a reciprocal regulation of the signaling pathways triggered by TGFβ1 and IFNγ in glial cells.

Materials and methods

Glial cultures

Primary mixed glial cell cultures were prepared from brain cortices of newborn (2 days) rats, as previously described (Tichauer et al. 2007). Animals were obtained from the institutional animal facility. They were anesthetized with ether before being killed and all procedures were performed in agreement with the animal handling and bioethical requirements established by the Pontificia Universidad Católica de Chile Ethics Committee. Briefly, meninges and blood vessels were removed and the cortices were placed in 0.25% trypsin (Sigma, St. Louis, MO, USA) in buffer Hanks, Ca+2/Mg+2 free, pH 7.2, at 37°C for 10 min, and mechanically dissociated. The cell suspension was plated in 24-well plates for respiratory burst and nitrite production assays, in 35 mm Petri plates for western blots of signaling pathway mediators and MKP-1, or in 100 mm Petri plates for microglia and astrocytes purification, in Dulbecco’s modified Eagle’s medium (DMEM) with F-12 (Gibco, Carlsbad, CA, USA), 10% fetal bovine serum (Hyclone, Logan, UT, USA), and 1% penicillin/streptomycin (Gibco). This protocol generates mixed glial cultures containing approximately 80% astrocytes and 20% microglia. Cells were maintained at 37°C under a humidified 5% CO2 atmosphere. Microglia were collected at 10 days in culture by shaking 100 mm Petri dishes. Microglia were seeded onto 16-mm-diameter 24-well plastic plates (2 × 104 cells/well) in DMEM containing 10% FCS. These cultures contained > 97% of cells that were positive for both isolectin B4 and Iba1. To purify astrocytes, after 10 days in culture, confluent mixed glia was treated with 1 μM cytosine-arabinoside during 3 days to eliminate proliferating cells. Medium was changed twice a week and cultures were used after 3 weeks. These cultures contained > 95% of cells that were positive for glial fibrillary acidic protein (GFAP). No neuronal cells were detected as judged by MAP2 staining.

Determination of superoxide anion

The production of O2· by microglial cells was assessed by the reduction of nitro blue tetrazolium assay (Rook 1985). Briefly, inflammatory activation was induced by addition of 10 ng/mL IFNγ (R & D Systems, Minneapolis, MN, USA), or 10 ng/mL IFNγ plus 1 μg/mL Lipopolysaccharide (LPS; Sigma) with or without 1 ng/mL TGFβ1 (R&D) at 37°C for 24 h, and in the absence or presence of pre-treatment with MAPK inhibitors. For the inhibition of MAPKs, glial cells were pre-treated with 20 μM PD98059 (ERK1/2 inhibitor; Calbiochem, San Diego, CA, USA) or 20 μM SB203580 (P38 inhibitor; Calbiochem) in 0.11% dimethyl sulfoxide (DMSO) in phosphate-buffered saline (PBS) for 30 min previous to the inflammatory stimulation. Control cells were exposed to the same concentration of DMSO. After stimulation, the culture medium was replaced with 1 mg/mL nitro blue tetrazolium in phenol red-free DMEM/F-12 containing 1 mg/mL bovine serum albumin. After the assay, glial cell cultures were fixed with 100% methanol at 23°C. Cells were photographed using bright field optics (Leyca DMIL, Solms, Germany).

Determination of nitrites

Nitrite presence (a stable metabolite of NO·) was determined in the cell culture medium by the Griess assay as previously described (Pfeiffer et al. 1997). In brief, cultures were incubated with 10 ng/mL IFNγ or 10 ng/mL IFNγ plus 1 μg/mL LPS at 37°C for 24 h with or without cotreatment with 1 ng/mL TGFβ1, and in the absence or presence of siRNA transfection or pre-treatment with MAPK inhibitors. Aliquots of 50 μL of medium were mixed with 10 μL EDTA/H2O 1 : 1 (0.5 M, pH 8.0) and 60 μL of fresh Griess reagent [20 mg N-(1-naphtyl)- ethylendiamine, and 0.28 g sulphanilamide (Merck, Whitehouse Station, NJ, USA) dissolved in 20 mL of 5% phosphoric acid, w/v]. Standard curves were established with 1–80 μM NaNO2. Absorbency was measured at 570 nm in a microplate auto reader (ANTHOS 2010, Anthos Labtec Instrument, Salzburg, Austria).

Western blot analysis

After stimulation, cells were rinsed once with PBS and scrapped in 50 μL of sample buffer supplemented with complete protease inhibitor cocktail (CompleteTM; Boehringer Mannheim, Mannheim, Germany) and phosphatase inhibitors (1 mM Na3VO4 and 100 mM NaF, Sigma) at 4°C. Samples were immediately heated at 95°C for 8 min and then maintained at −20°C until use. Aliquots of 20 μL of samples were electrophoretically separated in 10% polyacrylamide gels and transferred to nitrocellulose (0.45 μm pore size; BioScience, San Jose, CA, USA). The nitrocellulose was incubated with blocking solution (5% w/v non-fat dry milk, 0.1% tween-20 in Tris-buffered saline (TBS), pH 7.4) at 23°C for 1 h. The blots were probed with primary antibodies against STAT1 phosphorylated on tyrosine-701 or serine-727, P38 MAPK phosphorylated at threonine-180/tyrosine-182, total STAT1 or P38 MAPK (Cell Signaling Technology, Danvers, MA, USA) according to the manufacturer’s protocol, and then incubated with the secondary antibody, goat anti-rabbit HRP-conjugated IgG (1 : 2000; Calbiochem) at 23°C for 1 h. Primary antibodies against ERK1/2 phosphorylated on tyrosine 204, total ERK1/2, MKP-1 (Santa Cruz Biotechnology, Santa Cruz, CA, USA), and β-Tubulin (Sigma) diluted (1 : 500) in 1% w/v non-fat dry milk, 1% tween-20 in TBS, pH 7.4 at 23°C for 1 h, and then incubated with the secondary antibody, goat anti-rabbit or anti-mouse HRP-conjugated IgG (1 : 5000) at 23°C for 1 h. Bands were visualized by chemiluminescence (PerkinElmer, Boston, MA, USA). The nitrocellulose was stripped between different probes with 100 mM 2-mercaptoethanol, 2% w/v sodium dodecyl sulfate in TBS pH 6.9, at 60°C for 30 min.

Cell labeling

Glial cells (5 × 104) were seeded on glass coverslips in a 24-well plate and maintained in 5% CO2, 95% air at 37°C. Cells were rinsed with cold PBS and fixed with 4% paraformaldehyde at 23°C for 30 min. Samples were incubated with α-glial fibrillary acidic protein antibody (anti-GFAP; 1 : 200; Cell Signaling Technology) and with an anti-MKP-1 antibody (Santa Cruz Biotechnology) in 10% goat serum in PBS at 4°C overnight. Later, cells were labeled with secondary antibodies conjugated with Alexa Fluor 488 or 633 at 23°C for 3 h or with Alexa Flour 568-conjugated Isolectin GS-IB4 (Griffonnia simplicifolia; microglia identity marker) at 4°C overnight. Nuclei were stained with Hoechst or DAPI (Molecular Probes, Eugene, OR, USA). Cells were photographed using fluorescence microscopy (Olympus, Tokyo, Japan) or a confocal microscope (Olympus IX81).

siRNA transfection

Glial cell cultures were transiently transfected with a rat-specific pool of three target-specific 20–25 nt MKP-1 siRNA (Santa Cruz Biotechnology) according to the manufacturer’s instructions. Transfections were performed 24 h prior to stimulation with 10 ng/mL IFNγ with or without 1 ng/mL TGFβ1, using a siRNA transfection reagent with the siRNA at a concentration of 0.08 μM in transfection medium (Santa Cruz Biotechnology). Control cultures were treated with a scrambled sequence that does not lead to specific degradation of any known cellular mRNA (Santa Cruz Biotechnology). Efficiency of siRNA transfection was evaluated by immunofluorescence using an FITC-conjugated siRNA (Santa Cruz Biotechnology).

Signaling network analysis

The signaling network was generated using Ingenuity Pathway Analysis ( The data set containing cell signaling proteins of interest was uploaded to obtain a graphical representation of a network that illustrates the functional relationship with lines and arrows. Primary functional connections were corroborated with the Ingenuity pathway database.

Statistical analysis

Statistical analysis was performed with the Kruskal–Wallis One-Way anova and the Wilcoxon Rank Sum/Mann–Whitney U-test. Evaluation was performed using the GBStat statistical software (Dynamic Microsystems, Inc., Silver Spring, MD, USA). Differences were considered significant for p < 0.05.


TGFβ1 decreased IFNγ-induced O2· and NO· production by glial cells

Microglial cells present in primary mixed glial cultures produced high amounts of O2· following inflammatory stimulation with IFNγ or LPS + IFNγ (LI) for 24 h (Fig. 1a, dark cells). Cotreatment with TGFβ1 resulted in a conspicuous decrease in the number of O2·-producing microglial cells (Fig. 1a). NO· production increased four-fold over control levels when cells were incubated with IFNγ and nine-fold after stimulation with LPS + IFNγ for 24 h. Cotreatment with TGFβ1 abolished induction of NO· by IFNγ and decreased LPS + IFNγ-induced NO· production by 50% (Fig. 1b). Moreover, TGFβ1 showed to be effective to decrease IFNγ-induced NO· production in purified cultures of astrocytes and microglia (see Figs. 1c and 6d), confirming that TGFβ1 can be an effective modulator of the NO· and O2· release by both glial cell types in culture.

Figure 1.

 TGFβ1 reduces interferon-γ (IFNγ)-induced O2· and NO· production by glial cells. Cells were incubated with IFNγ or lipopolysaccharide (LPS)+IFNγ (LI) with or without recombinant TGFβ1 for 24 h. (a) Bright field microphotograph of mixed glial cells. The production of O2· by microglial cells was evaluated by the nitro blue tetrazolium (NBT) assay (dark cells). Representative images of three independent experiments are shown. (b) Nitrites production measured by the Griess reaction. Results are expressed as NO2 concentration. (c) Quantification of O2· and nitrites production by isolated astrocytes and microglial cells, and their reduction by recombinant TGFβ1. Results are expressed as fold-number increase of O2· and NO2 concentration compared with control cultures. Data correspond to the mean ± SEM of three to four independent experiments performed in triplicate. **p < 0.01; ***p < 0.001 versus control cultures; #p < 0.05 or ##p < 0.01 versus cultures stimulated only with inflammatory molecules. Scale bar = 200 μm.

Figure 6.

 TGFβ1 modulation of interferon-γ (IFNγ)-mediated glial activation is mediated by an increment of MAPK phosphatase-1 (MKP-1) levels. (a) MKP-1 protein levels evaluated by western blot in cultures treated with IFNγ and/or TGFβ1. Tubulin was used as loading control. Values correspond to mean ± SEM of four independent experiments. *p < 0.05 versus control. (b) MKP-1 expression (MKP-1, cyan) in astrocytes (GFAP-positive cells, red) and microglia (lectin IB4-positive cells, green) present in mixed glial cell cultures exposed to the indicated experimental conditions, were evaluated by immunolabeling and confocal microscopy. Nuclear staining with DAPI was added on merge image (Merge DAPI) for single cell identification, as shown in digital magnifications. Scale bar = 20 μm. (c) Mixed glial cultures were transfected with control siRNA (CsiRNA) or MKP-1 siRNA (MsiRNA) and then exposed to IFNγ and/or TGFβ1 for 24 h. Graph: Nitrites production determined by the Griess assay. Insert: Blot showing the down-regulation of MKP-1 by MKP-1 siRNA. Values correspond to the mean ± SEM of five independent experiments. **p < 0.01; ***p < 0.001 versus control; ###p < 0.001 versus treatment with IFNγ alone; &p < 0.05 versus cotreatment with IFNγ and TGFβ1. (d) Nitrite production by purified cultures of astrocyte and microglia was evaluated as in (c). Data are mean ± SEM of three independent experiments. **p < 0.01 versus control; *p < 0.05; ##p < 0.01 versus treatment with IFNγ alone; &p < 0.05 versus cotreatment with IFNγ and TGFβ1. (e) GFAP-positive astrocytes (green, arrows) and BH4-positive microglia (arrows head, red) can also be identified by differences in their nuclear morphology. Mock transfection (Mock-t) of mixed glial cultures showed no FITC signal, whereas FITC-conjugated siRNA (siRNA) showed that both microglia (arrow heads) and astrocytes (arrows) were transfected.

Activation of STAT1 and MAPK pathways by IFNγ and TGFβ1 in glial cells

Glial cells showed STAT1 phosphorylation on Y701 only after being exposed to IFNγ. The ratio pSTAT1tyr/total STAT1 observed at 15 min of stimulation increased by two-fold at 30 min (Fig. 2a). In contrast to pSTAT1tyr, STAT1 phosphorylated on S727 was observed in non-stimulated cultures. The ratio pSTAT1ser/total STAT1 progressively increased in a time-dependent manner reaching 4.5-fold compared with control conditions in cultures stimulated with IFNγ for 60 min (Fig. 2a).

Figure 2.

 Activation of signal transducer and activator of transcription-type-1 (STAT1) and mitogen-activated protein kinase pathways by interferon-γ (IFNγ) and TGFβ1 in glial cells. Phosphorylation of STAT1 (a) or ERK1/2 and P38 mitogen-activated protein kinase (b) in mixed glial cell cultures exposed to IFNγ or TGFβ1 for 15, 30, or 60 min. Values correspond to the mean ± SEM of three independent experiments. *p < 0.05 versus control; #p < 0.05 shows significant differences between different treatment durations.

Phosphorylation of ERK1/2 (pERKs) and P38 MAPK (pP38) was low in control cultures and increased when cultures were stimulated with IFNγ. The ratio pERKs/total ERK and pP38/total P38 increased by 40–50% after 15 min and increased up to three to four-fold after 30–60 min of stimulation with IFNγ (Fig. 2b). We also examined the phosphorylation of another MAPK, JNK. Phosphorylated JNK in glial cell cultures stimulated with IFNγ for 15 min or 24 h was similar to that observed in control cultures (data not shown), suggesting that IFNγ did not activate JNK under our experimental conditions.

TGFβ1 also activated ERK1/2 and P38 MAPK in mixed glial cell cultures. Glial cells exposed to TGFβ1 for 15 min showed a two to three-fold increase of pERK1/2 and pP38 compared with control cultures. After the early increment of pERK1/2 and pP38, pERK1/2 gradually decreased to control levels, whereas pP38 maintained a three-fold increase in glial cultures exposed to TGFβ1 for 60 min (Fig. 2b).

Effect of MAPK inhibition on IFNγ-induced production of radical species and activation of STAT1 in glial cells

Involvement of ERK1/2 and P38 in glial activation and NO· and O2· production was tested using inhibitors specific for ERK (PD98059) and P38 (SB203580) pathways. The chosen inhibitor concentrations were those needed to decrease IFNγ-induced phosphorylation of the corresponding MAPK by 90% (data not shown). Production of radical species by glial cells treated with vehicle (0.11% DMSO in PBS) was similar to control cells treated or not with inhibitors (data not shown). A robust O2· production by microglial cells (dark cells, Fig. 3a) was observed in mixed glial cultures exposed to IFNγ for 24 h. Pre-treatment with SB203580 or PD98059 did not modify O2· production by microglial cells under control conditions. However, PD98059 but not SB203580 decreased the number of microglial cells producing O2· during IFNγ-induced respiratory burst (Fig. 3a). The production of NO· by glial cells under IFNγ stimulation increased five-fold compared with cultures maintained in control conditions. SB203580 and PD98059 partially decreased IFNγ-induced NO· production by 35% and 30%, respectively. NO· production decreased by 70% in cultures cotreated with SB203580 and PD98059 (Fig. 3b).

Figure 3.

 Effect of mitogen-activated protein kinase (MAPK) inhibition on interferon-γ (IFNγ)-induced production of radical species in glial cells. (a) Representative images of a nitro blue tetrazolium (NBT) assay showing the production of O2· by microglial cells under control conditions and after exposure to IFNγ for 24 h (dark cells) and the effect of pre-treatment with MAPK inhibitors. (b) Quantification of microglial cells producing O2· (NBT+ cells) and (c) Nitrites production determined by the Griess assay in mixed glial cultures pre-treated with MAPK inhibitors and exposed to IFNγ for 24 h. (-), vehicle-pre-treated cells; (PD), ERK-specific inhibitor PD98059 and (SB), P38-specific inhibitor SB203580. Values correspond to the mean ± SEM of three to five independent experiments in duplicate/triplicate. **p < 0.01 treatment with IFNγ versus control cells; ##p < 0.01 compares treatment with IFNγ, with and without MAPK inhibitors.

The existence of a cross-talk between STAT1 and MAPK pathways was also examined. Pre-treatment with PD98059 or SB203580 did not change the timing or amount of pSTAT1tyr induced by IFNγ in glial cells (Fig. 4ai and ii) as confirmed by densitometric analysis (not shown). In contrast, cells pre-treated with PD98059 or SB203580 and stimulated with IFNγ showed a consistent decrease by 30–40% of pSTAT1ser/total STAT1 ratio compared with cultures exposed to IFNγ for 15, 30, or 60 min after pre-treatment with the inhibitors’ vehicle (Fig. 4ai and ii). This effect was confirmed in other experiments performed at 60 min (Fig. 4b). Consistently, each MAPK inhibitor significantly reduced IFNγ-induced pSTAT1ser. Moreover, pre-treatment with both inhibitors nearly abolished IFNγ-mediated increment of pSTAT1ser (Fig. 4b).

Figure 4.

 Effect of mitogen-activated protein kinase (MAPK) inhibition on interferon-γ (IFNγ)-induced signal transducer and activator of transcription-type-1 (STAT1) phosphorylation in glial cells. Western blot of mixed glial cell cultures pre-treated with vehicle or MAPK inhibitors SB 203580 (SB) or PD 98059 (PD) for 30 min, and stimulated with IFNγ for 15, 30, or 60 min (ai and aii), or only 60 min (b). Densitometric analysis for pSTAT1ser/total STAT1 ratio (a i and ii) and from three to five independent experiments (graph in b) are shown. (-), vehicle-pre-treated cells; (PD), ERK-specific inhibitor PD98059 and (SB), P38-specific inhibitor SB203580. White bar, control conditions. Black bars, cells treated with IFNγ. ***p < 0.001 treatment with IFNγ versus control cells; #p < 0.05 or ###p < 0.001 compares treatment with IFNγ, with and without MAPK inhibitors.

Effect of cotreatment with IFNγ and TGFβ1 on the activation of STAT1, ERK1/2, and P38 MAPK pathways in glial cells

Cotreatment with TGFβ1 and IFNγ for 15 min resulted in a two-fold increase of pERK1/2 compared with the effect of IFNγ alone (calculated as pERK1/2/ERK1/2 ratio, Fig. 5a). TGFβ1, after treatment for up to 60 min did not decrease IFNγ-induced pERK1/2 in glial cultures (Fig. 5a), but inhibition of ERK1/2 phosporylation was observed after 24 h of stimulation with both cytokines (Fig. 5b). pP38 level was low in glial cells stimulated with IFNγ, but increased after exposure to TGFβ1 for 24 h. Cotreatment with both cytokines decreased the induction of pP38 by TGFβ1 alone (Fig. 5b). pSTAT1tyr and pSTAT1ser showed an increase in glial cultures exposed to IFNγ for 24 h compared with control cultures. IFNγ also induced a slight increase of total STAT1. Cotreatment with TGFβ1 decreased IFNγ-induced pSTAT1tyr, pSTAT1ser, and total STAT1 (Fig. 5b).

Figure 5.

 Effect of the cotreatment with interferon-γ (IFNγ) and TGFβ1 on the activation of signal transducer and activator of transcription-type-1 (STAT1), ERK1/2 and P38 mitogen-activated protein kinase pathways in glial cells. Western blots of mixed glial cells stimulated with IFNγ in the presence or absence of TGFβ1 for short (15–60 min, a) and long times (24h, b). (a) Data are the mean of densitometric analysis of pERK1/2/ERK1/2 ratio from three to five independent experiments. *p < 0.05 for IFNγ + TGFβ1 versus control; #p < 0.05 for treatment with IFNγ + TGFβ1 versus IFNγ alone. (b) Tubulin was used as loading control.

TGFβ1 modulation of IFNγ-induced glial cell activation is mediated by an increment of MKP-1 levels

We further explored the mechanism involved in the modulatory effect of TGFβ1. Because it was observed after long times of treatment, induction of gene transcription and de novo protein synthesis could be involved. As recently MKP-1 expression has been found to be involved in glial reactivity (Huo et al. 2011; Lee et al. 2012), we evaluated changes in MKP-1 expression in mixed and purified cultures of astrocytes and microglia. MKP-1 protein levels were increased by 2.5-fold in glial cells, exposed to TGFβ1 for 24 h, over control cells. IFNγ neither induced MKP-1 expression nor modified the induction of MKP-1 expression by TGFβ1 (Fig. 6a). Whether MKP-1 is expressed by astrocytes and/or microglia was evaluated in mixed glial cultures using antibodies against MKP-1, GFAP, and IB4, indirect immunofluorescence, and confocal microscopy. MKP-1 levels were low in both astrocytes (GFAP-positive cells, Fig. 6b) and microglia (IB4- positive, Fig. 6b) in control conditions. As expected, treatment with TGFβ1 largely increased MKP-1 levels in both cell types. Moreover, increased MPK-1 expression was also observed during IFNγ + TGFβ1 cotreatment (Fig. 6b). Then, we down-regulated MKP-1 expression by MKP-1 siRNA transfection, obtaining an efficiency of approximately 30% for both glial cell types as revealed by FITC-conjugated siRNA transfection (Fig. 6e and data not shown). MKP-1 down-regulation reversed the effect of TGFβ1 on IFNγ-induced NO production by 29% in mixed glial cell (Fig. 6c). To further evaluate whether MKP-1 expression is involved in the TGFβ1 anti-inflammatory effect, purified cultures of astrocytes and microglia were transfected with MKP-1 siRNA, and treated with TGFβ1 and/or IFNγ. NO· production was quantified (Fig. 6d). Pure astrocytes produced slightly less NO· than microglia in basal and stimulated conditions. Accordingly, MKP-1 down-regulation prevented TGFβ1 reduced IFNγ-induced NO· production in microglia, whereas the effect was slightly less pronounced in astrocytes (Fig. 6d).


Modulation of glial cell activation exerted by TGFβ1 has been documented. However, little is known about the molecular mechanisms which are involved. Here, we show that inhibition of IFNγ-induced NO· and O2· production by TGFβ1 depended on the cross-talk between MAPKs and STAT1 signaling pathways. Indeed, after a long-lasting stimulation, TGFβ1 regulated IFNγ-induced activation of STAT1 via dephosphorylation of ERK1/2. Notably, we found that the phosphatase MKP-1 may be involved in this regulatory mechanism.

IFNγ induces radical species production via activation of STAT1/MAPKs pathways

Glial cell cultures exposed to IFNγ for short (< 60 min) and long (24 h) times showed increased phosphorylation of STAT1 on Y701 and S727 positions. STAT1 is a key signaling pathway involved in the up-regulation of iNOS and NO· production in several cells types (Dell’Albani et al. 2001; Gough et al. 2008). Inhibition of ERK1/2 and P38 MAPK decreased IFNγ-induced pSTAT1ser, which correlated with a reduction in NO· production. Decrease of pSTAT1ser and NO· production was additive after pre-treatment with both MAPK inhibitors, suggesting that ERK1/2 and P38 are needed for full activation of the STAT1 pathway in glial cells, as described for other cell types (Blanchette et al. 2003; Platanias 2005; Gough et al. 2008). Another finding was that O2· production induced by IFNγ depended on increased levels of pERK1/2, but not pP38, as previously reported (Bhat et al. 1998; Dang et al. 2003). Moreover, phosphorylation of ERK1/2 was elevated after 24 h of treatment with IFNγ, whereas phosphorylation of P38 decreased to control levels. Differential temporal contribution of ERK1/2 and P38 MAPK suggests that whereas both ERK and P38 contribute to STAT1 modulation at short times, only ERK1/2 participates after long-time stimulation with IFNγ. It is also known that sustained activation of ERK signaling pathway in astrocytes plays a pivotal role in stellation and astrogliosis (Chu et al. 2004) and NMDA-induced neuronal injury (Katayama et al. 2010). Such association suggests that ERK signaling may be a potential target for therapeutic applications for neurodegenerative diseases.

TGFβ1 regulates inflammatory response modulating IFNγ-activated signaling pathways

Cotreatment with TGFβ1 resulted in a decrease of IFNγ-induced pERK1/2 and pSTAT1ser levels, mimicking the effects of pre-treatment with MAPKs inhibitors. Thus, suppression of pSTAT1ser was probably mediated by TGFβ1-induced decrease of pERK1/2 i.e. via MKP-1 expression (see below). We also observed, in presence of TGFβ1, an inhibition of IFNγ-induced pSTAT1tyr that could depend on TGFβ1-induced decrease of total STAT1. Little is known about the suppression mechanisms of the JAK-STAT1 pathway via TGFβ1. Nevertheless, and consistent with our results, it has been described that TGFβ1 inhibits iNOS mRNA transcription by suppressing STAT1 activation in IFNγ-stimulated macrophage-like cell line RAW 264.7. It has also been demonstrated that TGFβ receptor I interacts with and phosphorylates IFNγ receptor 1, preventing STAT1 activation in these cells (Takaki et al. 2006). Therefore, certain signaling pathways that were active during single cytokine stimulation became silent during the simultaneous activation of multiple signaling pathways activated by both cytokines. Thus, the final cell response would be mediated by a balance between pro- and anti-inflammatory signals, and maybe the deactivation of the ERK pathway is determinant for the regulatory effect of TGFβ1 over IFNγ-induced glial cell activation.

Importantly, the existence of regulatory interactions between TGFβ1 and IFNγ also has been described in tissue repair in vivo. IFNγ-null mice show an increased amount of TGFβ1 and activation of TGFβ1-induced signaling pathways, suggesting that IFNγ exerts a negative modulation of TGFβ1 activity (Ishida et al. 2004). On the other hand, TGFβ1-null mice show elevated plasma levels of IFNγ and high levels of STAT1, iNOS, and NO· production, indicating a deregulation of IFNγ pathway and its target genes in the absence of TGFβ1 (McCartney-Francis and Wahl 2002). Besides, some levels of interaction between IFNγ- and TGFβ- induced signaling pathways have been described in vitro. For example, IFNγ suppresses TGFβ signaling through up-regulation of the inhibitory Smad7 in U4A cell line (Ulloa et al. 1999) and inhibits TGFβ1 responses via STAT1-mediated sequestration of the nuclear coactivator p300/cAMP response element-binding protein, preventing its association with Smads and blocking Smad transcriptional activity in primary fibroblasts (Ghosh et al. 2001). However, these events do not occur in other cell types evaluated, such as T cells (Yoshimura et al. 2010). Taking all together, abundant evidences not only support our proposition that TGFβ1 modulates the inflammatory response induced by IFNγ, but also suggest the existence of a dynamic signaling cross-talk between both cytokines.

MKP-1 as a key participant in the anti-inflammatory effect of TGFβ1

We observed TGFβ1-induced MKP-1 expression in both glial cells. In addition, MKP-1 expression was not affected by IFNγ, suggesting that TGFβ1 increase MKP-1 expression to regulate activation in glial cells. Indeed, it was confirmed by siRNA targeting of MKP-1. MKP-1 down-regulation prevented the modulation of TGFβ1 on IFNγ-induced NO· production. In similar experimental conditions, we found that IFNγ-induced pERK1/2, but not pP38, was decreased by TGFβ1 indicating that MKP-1 played a role in decreasing pERK1/2. It has been described that MKP-1 dephosphorylates preferentially P38 and JNK, but it also dephosphorylates ERK1/2 in some cell types (Liu et al. 2007; Boutros et al. 2008). There is also a report that pre-treatment with TGFβ1 for 48 h reduced the production of inflammatory mediators induced by Aβ1-42, which was associated with reduction of P38 and NF-κB activation and an increase in MKP-1 levels (Flores and von Bernhardi 2012). Furthermore, it has been also shown that induction of MKP-1 leads to an anti-inflammatory response via ERK dephosphorylation in microglia (Romero-Sandoval et al. 2009), whereas manganese inhibits MKP-1 expression resulting in enhanced MAPK activity and microglial inflammatory phenotype (Crittenden and Filipov 2011). Moreover, an increased MKP-1 level has been reported to be the action mechanism for several anti-inflammatory molecules, including glucocorticoids (Kassel et al. 2001; Jang et al. 2007; King et al. 2009). Noteworthy, the anti-inflammatory effect of 15-Deoxy-Δ12,14-Prostaglandin J2 and 5,8,11,14-eicosatetraynoic acid in astrocytes (Lee et al. 2008, 2012), and dexamethasone in microglia (Huo et al. 2011) have also been attributed to an increase of MKP-1 levels. Besides, it has been demonstrated that this phosphatase participates in STAT1 dephosphorylation (Venema et al. 1998). Thus, TGFβ1-induced MKP-1 expression represents a novel mechanism to explain their regulatory effects on MAPK and STAT1 pathways during inflammation, constituting a mechanism actively regulating both microglia and astrocytes.

TGFβ1–IFNγ cross-talk: a working model

A working model for the interaction of IFNγ and TGFβ1 signaling pathways is proposed in Figure 7. IFNγ induces STAT1 translocation into the nucleus by the JAK-dependent phosphorylation of Y701. ERK1/2 is persistently and P38 is transiently phosphorylated, inducing pSTAT1ser and increasing transcription of target genes and NO production. ERK1/2 also participates in the release of O2·. TGFβ1 decreases IFNγ-induced pSTAT1ser through reduction of ERK1/2 activation, inhibiting the production of radical species. Decreased ERK1/2 activation depends on the TGFβ1-mediated induction of MKP-1. IFNγ-dependent induction of STAT1 protein was abolished in the presence of TGFβ1. In contrast, TGFβ1 induced a persistent increase of pP38, which was inhibited by IFNγ.

Figure 7.

 Modulation of interferon-γ (IFNγ)-induced signaling pathways by TGFβ1. Working network obtained according to Ingenuity database ( as indicated in Methods. IFNγ activates Janus-activated kinase (JAK)- signal transducer and activator of transcription-type-1 (STAT1) pathway, increasing pSTAT1tyr, pERKs and to a lesser extent pP38 levels. Active mitogen-activated protein kinase (MAPK)s potentiates STAT1 activation by phosphorylation in a serine residue. Glial cells respond to IFNγ stimulation by increasing their production of radical species, i.e. by transcriptional up-regulation of inducible nitric oxide synthase (iNOS). TGFβ1 is also able to activate MAPKs. Particularly, TGFβ1 can produce recruitment of TRAF6 (Sorrentino et al. 2008) to activate P38 and regulate MAPK phosphatase-1 (MKP-1) expression (reviewed by Akbarian and Davis 2010). Our results show that, particularly in microglial cell, TGFβ1 inhibits IFNγ-induced STAT1 activation and iNOS expression via a MKP-1-mediated inhibition of ERK1/2. MKP-1 may also modulate the ERK1/2-promoted superoxide production.

Homeostasis of the nervous tissue is maintained by a finely tuned interaction between glial cells and neurons, involving a complex network of signaling pathways induced by simultaneous stimuli. In neurodegenerative diseases, the homeostatic cell balance is altered and cell damage is induced by elevated production of inflammatory cytokines and radical species. In this work, our results show the modulation of inflammatory activation of glial cells through a cross-talk between IFNγ and TGFβ1, which regulates the production of radical species through the activation of STAT1, ERK1/2, and P38 pathways. Induction of MKP-1 appears to be part of the endogenous response mediated by TGFβ1, capable of decreasing potentially cytotoxic activation of microglia and astrocytes. Thus, MKP-1 could be a therapeutic target for neurodegenerative diseases.


Research was supported by grants FONDECYT 1090353 and NIH R03TW008019 to R.v.B. R.H-M thanks financial support from the International Society for Neurochemistry and the Committee for Aid and Education in Neurochemistry (CAEN). We thank Dr. Constanze I. Seidenbecher for critical comments on the manuscript and Dr. Karl-Heinz Smalla for his helpful advice on Ingenuity-based signaling network design.

The authors have no conflict of interest to declare.