Transforming growth factor-β1 produced by hippocampal cells modulates microglial reactivity in culture

www.elsevier.com/locate/ynbdi Neurobiology of Disease 19 (2005) 229 – 236

Transforming growth factor-B1 produced by hippocampal cells modulates microglial reactivity in culture Rodrigo Herrera-Molina and Rommy von BernhardiT Department of Neurology, Faculty of Medicine, Pontificia Universidad Cato´lica de Chile, Marcoleta 391, Santiago, Chile Received 1 September 2004; revised 23 November 2004; accepted 4 January 2005 Available online 11 February 2005

S

Activated microglia produce superoxide anion (O2 ) and nitric oxide (NO), both of which can be neurotoxic. To identify regulatory mechanisms that might modulate over-activation of microglia, we evaluated the inhibition of microglial activation by factors secreted by hippocampal cells. Supernatants from hippocampal cell cultures (Hippocampal-Cm) prevented microglial O2 and NO production. LAP-TGFB1 was present in the Hippocampal-Cm as shown by immunoblot and a TGFB1-dependent proliferation-inhibition bioassay. LAP-TGFB1 and TGFB activity increased in hippocampal cultures exposed to proinflammatory conditions (LPS and Interferon-gamma). The inhibition of O2 and NO production by Hippocampal-Cm was mimicked by the addition of recombinant TGFB1. Treating Hippocampal-Cm with an antibody against TGFB1 to neutralize its activity eliminated its ability to inhibit O2 and NO production. Our findings suggest that the TGFB1 secreted by hippocampal cells modulated microglial activity. We propose that in pathological conditions, impairment of this modulatory mechanism could enhance microglia-mediated neurotoxicity. D 2005 Elsevier Inc. All rights reserved.

S

S

S

Keywords: Cytokines; Hippocampal cell; Inflammation; Microglial cells; Neurodegeneration; Nitric oxide; Superoxide anion; TGFh1

Introduction Activated microglial cells and astrocytes are present in many pathological states of the central nervous system including trauma, ischemia, multiple sclerosis, and Parkinson’s and Alzheimer’s diseases (McGeer and McGeer, 1995; Meda et al., 2001; von Bernhardi et al., 2001). Neuronal death is associated with the activation of microglial cells (Bal-Price and Brown, 2001; Penkowa et al., 2000) and it has been hypothesized that microglial cells increase neuronal injury through the synthesis and release of shortlived cytotoxic factors including superoxide radicals (O2S ) and nitric oxide (NO) (McGeer and McGeer, 1995; Meda et al., 2001; Miranda et al., 2000). Acting either alone (Cai and Jones, 1998; T Corresponding author. Fax: +56 2 6321924. E-mail address: [email protected] (R. von Bernhardi). Available online on ScienceDirect (www.sciencedirect.com). 0969-9961/$ - see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.nbd.2005.01.003

Henrich-Noack et al., 1996; van Muiswinkel et al., 1996) or in concert by the generation of peroxinitrite (ONOO ) (Meda et al., 2001; Thannickal and Fanfurg, 2000), these cytotoxic compounds can induce neuronal death. When microglial cells are activated by proinflammatory molecules, such as IFN-g and lipopolysaccharide (LPS), large amounts of O2S are produced by the brespiratory burstQ (BruceKeller et al., 2000; Chao et al., 1995; van Muiswinkel et al., 1996) of such plasma membrane oxidases as NADPH oxidase (Thannickal and Fanfurg, 2000). Proinflammatory priming of microglial cells is necessary for the respiratory burst (Chao et al., 1995; Hu et al., 1995). LPS or IFN-g also elicit the production of high amounts of NO by microglial cells. This is mediated through the activation of nitric oxide synthase (NOS) and especially by the transcriptional up-regulation of its inducible isoform (iNOS) (Ding et al., 1997; Lieb et al., 2003; Meda et al., 2001; Vegeto et al., 2001; von Bernhardi et al., 2001). Microglial activity is modulated in various ways. Enzymes with antioxidant properties like catalase or superoxide dismutase and anti-inflammatory cytokines such as IL-1Ra, IL-10 or IL-4 can limit microglial activation (Wood, 1998). Transforming growth factor-h1 (TGFh1) is another cytokine that regulates microglial activity. TGFh1 inhibits microglial cell production of the proinflammatory molecules IL-1 and TNF-a and the expression of Class II MHC molecules (Benveniste, 1998) and Fas glycoprotein (Lee et al., 2000). TGFh1 also inhibits the induction of NOS and decreases the release of NO (Lieb et al., 2003; McCartney-Francis and Wahl, 2002; Vincent et al., 1997). Recombinant TGFh1 modulates superoxide production by microglial cells (Chao et al., 1995; Hu et al., 1995). In contrast, while these activities are expected to reduce cell damage, TGFh1 is also associated with proinflammatory events and the potentiation of neurotoxicity (Brown, 1999; Henrich-Noack et al., 1996). Hippocampal neurons and glial cells secrete TGFh both in vivo and in vitro (Flanders et al., 1998; Zhu et al., 2000). In vivo it has been also defined the differential secretion of TGFh and their receptors in different neuronal populations (Zhu et al., 2000). There are very low concentrations of TGFh1 in normal brain tissue, whereas its expression is increased in activated glial cells from injured or diseased brain (O’Brien et al., 1994; Zhu et al., 2000).

230

R. Herrera-Molina, R. von Bernhardi / Neurobiology of Disease 19 (2005) 229–236

TGFh1 is increased in brain tissue of patients after ischemic stroke (Krupinski et al., 1996) and appears to be involved in various neurodegenerative pathologies (Flanders et al., 1998). TGFh1 is constitutively produced by hippocampal neurons (Zhu et al., 2000) and it has been associated with neuroprotection in hippocampal cell cultures (Henrich-Noack et al., 1996) through the activation of PI3 kinase/Akt and MAPK/Erk1,2 pathways (Zhu et al., 2002, 2004). TGFh1 produced by neurons also modulates several glial functions including the expression of structural proteins and cell reactivity (Chen et al., 2002; De Sampaio, 2002; Eyu¨poglu et al., 2003; McMillian et al., 1994; Mittaud et al., 2002). Because microglial activation can decrease neuronal survival (Eyu¨poglu et al., 2003), we investigated the role of hippocampal cells in the modulation of O2S and NO production by microglial cells in culture. Cell culture experiments have certain limitations. Some regulatory factors having a role under in vivo conditions are not evaluated in our experimental model. Thus, glial activation and their response to TGFh could differ respect to their activation in vivo. However, it is a reasonable first approach considering the complex interactions involved in inflammatory regulation. We found that conditioned media from hippocampal cells decreased O2S and NO production and that TGFh1 mediated this effect. If the same interactions occurs in vivo, it give us a hint on one of the mechanisms that could be involved in neuroprotection and has the potentiality to become a therapeutic target.

Materials and methods Hippocampal cell cultures Hippocampal cultures were derived from E17–18 rat embryos as previously described (Ca´ceres et al., 1984). Briefly, isolated hippocampi were incubated in 0.25% trypsin (Sigma) in Ca2+/Mg2+ free–Hanks’ buffer, pH 7.2, at 378C for 10 min, and the tissue mechanically dissociated and centrifuged at 100  g for 1 min. The cell pellet was resuspended in minimal essential medium (MEM-10) with 10% horse serum (Gibco), and the cells were seeded at a density of 2–2.5  104 cells/cm2 into polyl-lysine-coated 24-well culture plates and incubated in a humidified atmosphere (5% CO2, 95% air at 378C) for 1 h. The medium was replaced with Neurobasal medium with B27 supplement (NbB27) (Gibco) with 1 mM glutamine (Sigma), and 1% penicillin–streptomycin (Gibco); this rich culture medium supports the growth and development of neurons (Brewer et al., 1993). Experiments were performed with hippocampal cells after 5–6 days in culture; at this time, 65 F 1.4% of the cells were neurons (h-tubulin isotype III positive cells) and the rest were astrocytes (GFAP positive cells).

were maintained in a controlled and humidified atmosphere (5% CO2, 95% air at 378C) for 14 days. Microglial cells were dislodged by agitation at 100 rpm on an orbital shaker at 378C for 20 min. Detached microglial cells were collected by centrifugation (200  g for 10 min), resuspended in DMEM/F-12 with 10% fetal bovine serum and seeded in 96-well plates at a density of 3  104 cells per well. After 2 h at 378C, to remove non-attached cells, cells were rinsed once with fresh culture medium and the medium was changed. Conditioned medium from hippocampal cell cultures Hippocampal conditioned medium was obtained from hippocampal cell cultures incubated with medium alone (HippocampalCm) or from cultures stimulated with both IFN-g, 10 ng/ml, (R&D) and LPS, 1 Ag/ml, (Sigma) (Stimulated-Cm). Media were filtered through inert cellulose acetate membranes (pore size 0.22 Am, Orange Scientific) and maintained at 208C until use. Respiratory burst assay The production of O2S was assessed by the reduction of nitro blue tetrazolium (NBT assay) into a blue precipitate (Bruce-Keller et al., 2000; Rook, 1985). Cells were cultured in fresh medium or Hippocampal-Cm. Inflammatory glial cell activation was elicited by addition of IFN-g, 10 ng/ml, LPS, 1 Ag/ml, or both of these proinflammatory compounds, for 24 h at 378C. After this time, the culture medium was replaced with NBT, 1 mg/ml, in phenolred-free DMEM/F-12 containing 1 mg/ml BSA. Respiratory burst was triggered with phorbol 12-myristate 13-acetate (PMA), 150 ng/ml (Sigma) for 1.5 h. Next, glial cell cultures were fixed with 100% methanol at room temperature. Cells were photographed under bright field microscopy in an inverted microscope (Leyca DMIL). The crystals were dissolved with 50 Al 1:1.15 2 M KOH/ DMSO (Sigma) and absorbency was read at 645 nm in a microplate auto reader (ANTHOS 2010, Anthos Labtec Instrument). MTT assay The ability to reduce 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT) was used as an index of reducing activity of glial cells (Schwartz and Osborne, 1995). 3  104 microglial cells cultured in fresh medium or Hippocampal-Cm were incubated with IFN-g, 10 ng/ml, or both LPS, 1 Ag/ml, and IFN-g, 10 ng/ml, at 378C for 24 h. 10 Al of MTT, 5 mg/ml, was added to each culture and incubated at 378C for 3 h. Any crystals formed by the reduced MTT were dissolved by the addition of 100 Al lysis solution (50% w/v dimethilformamide, 20% w/w SDS) and read at 570 nm with reference to 645 nm in a microplate auto reader (ANTHOS 2010, Anthos Labtec Instrument).

Preparation of microglial cells and mixed glial cell cultures Determination of nitrites (NO2 ) Primary mixed glial cell cultures were prepared from the brain cortices of newborn (2-day) rats. The cortices were separated from meninges and blood vessels, placed in 0.25% trypsin (Sigma) in Hanks’ buffer, Ca2+/Mg2+ free, pH 7.2, at 378C for 10 min, mechanically dissociated and the resultant cell suspension plated in 24-well plates or in 75-cm2 flasks. The culture medium consisted of Dulbecco’s minimal essential medium (DMEM) (Gibco) supplemented with F-12 (Gibco), 10% fetal bovine serum (Hyclone) and 1% penicillin and streptomycin (Gibco). Cultures

3  104 microglial cells cultured in fresh medium or Hippocampal-Cm were incubated with IFN-g, 10 ng/ml, or both LPS, 1 Ag/ml, and IFN-g, 10 ng/ml, at 378C for 24 h. NO2 secreted to the cell culture medium was determined by the Griess assay (Pfeiffer et al., 1997). In brief, 50 Al of medium was mixed with 10 Al EDTA/H2O 1:1 (0.5 M, pH 8.0) and 60 Al of freshly prepared Griess reagent (20 mg N-[1-naphtyl]-ethylendiamine and 0.2 g sulphanilamide dissolved in 20 ml of 5% phosphoric acid, w/v).

R. Herrera-Molina, R. von Bernhardi / Neurobiology of Disease 19 (2005) 229–236

231

Calibration curves were established with 1–80 AM NaNO2. Absorbency was measured at 570 nm in a microplate auto reader (ANTHOS 2010, Anthos Labtec Instrument).

cell activation was elicited by addition of IFN-g, 10 ng/ml or IFNg, 10 ng/ml plus LPS, 1 Ag/ml, for 24 h at 378C. NO2 and O2S were measured as described in the corresponding method sections.

Western blot analysis of conditioned media

Statistical analysis

For the TGFh1 immunoblot, 1.5 ml of filtered conditioned medium from hippocampal cell cultures under control or stimulated conditions was concentrated by centrifugation at 2500  g in a Centricom-10 (Amicom). Equivalent amounts (100 Ag) of protein as determined by the Bicinchoninic acid (BCA) method were electrophoretically separated on 12% poly-acrylamide gels and transferred to nitrocellulose (pore size 0.45 Am, BioScience). The nitrocellulose was incubated with a blocking solution (5% nonfat milk, 0.05% Tween-20 in PBS) at 48C overnight. The blot was probed with polyclonal rabbit antibody specific for rat LAPTGFh1 (1:400, Santa Cruz) and then incubated with a second antibody, goat anti-rabbit horseradish peroxidase-conjugated IgG (1:10,000, Calbiochem). The LAP-TGFh1 protein was visualized by chemiluminescence (Life Science).

Statistical analysis was performed with the Wilcoxon Rank Sum/Mann–Whitney U test. Evaluation was performed using GBstat statistical software (Dynamic Microsystems, Inc). Differences were considered significant for P b 0.05.

Antibody-neutralization of TGFb1 bioactivity in conditioned media from hippocampal cell cultures TGFh1 is secreted bound to a latency-associated protein (LAP). In this form, TGFh1 is inactive. The TGFh1 contained in the conditioned media from control or stimulated hippocampal cell cultures was activated by heating the sample at 808C for 10 min (Gleizzes et al., 1997; Munger et al., 1997; Vincent et al., 1997). Following heat treatment, to neutralize the activity of TGFh1 (Vincent et al., 1997), conditioned media were incubated with antibody specific for active TGFh1 (3 Ag/ml, R&D) at room temperature for 4 h, before being added to the TGFh assay cell cultures. As a control, we incubated the conditioned media (inactivated and heat-activated) with an antibody specific for LAP-TGFh1 (3 and 5 Ag/ml, Santa Cruz).

Results Hippocampal-Cm modulated O2S production by microglial cells Microglial cells in mixed glial cultures produced more O2S following stimulation with IFN-g (dark cells, Fig. 1A). When glial cells incubated with IFN-g were treated with PMA, there were significant increases in O2S production (Fig. 1A). Hippocampal-Cm markedly inhibited O2S production (Fig. 1A). To quantify the magnitude of this inhibitory effect, microglial cultures were exposed to Hippocampal-Cm (Fig. 1B). There was a 4.5-fold increase (compared with cells not incubated with IFN-g) of O2S levels in microglial cultures stimulated with IFN-g in NbB27 ( P b 0.001). When O2S production was induced with IFN-g and PMA in NbB27, O2S production increased 8-fold (compared with control cultures, P b 0.001). Hippocampal-Cm decreased the generation of O2S by microglial cultures exposed to IFN-g by 15% ( P b 0.05) and the PMA-triggered O2S production by 50% (compared with cultures in NbB27, P b 0.001). The levels of O2S in microglial cultures incubated in Hippocampal-Cm and exposed to IFN-g or IFN-g plus PMA were not statistically different. Hippocampal-Cm increased MTT reduction in microglial cell cultures

TGFb bioassay The biological activity of TGFh was assessed using the mink lung epithelial cell (ATCC CCL-64) proliferation inhibition assay (Mouri et al., 2002; Vincent et al., 1997). 2  103 CCL-64 cells were seeded in 96-well plates in supplemented MEM media (2 mM lglutamate, 1.5 g/l sodium bicarbonate, 1 mM sodium pyruvate, Sigma; 10% fetal bovine serum, Hyclone) and incubated for 2–3 h at 378C. Next, an equal volume of hippocampal conditioned medium (a 5% fetal bovine serum was added because the hippocampal medium was serum free) was tested for the presence of activated TGFh1 by adding it to a lung cell culture. After 48 h, the inhibition of cell proliferation was determined by two independent methods: the MTT (Kim et al., 2002) and neutral red assays (Vincent et al., 1997). The amount of inhibition of proliferation as determined by both methods was similar (data not show). The concentration of TGFh was determined by interpolating the degree of inhibition of experimental samples in a standard inhibition curve established with recombinant TGFh1 (0.1–10 ng/ml, R&D). Modulation of NO2 and O2S production by recombinant TGFb1 Cells were cultured in fresh medium, Hippocampal-Cm or fresh medium with 1 ng/ml of recombinant TGFh1. Inflammatory glial

Microglial cultures incubated in Hippocampal-Cm had an increase of 50–70% of MTT reduction compared with cells incubated in NbB27 ( P b 0.02, Fig. 2). The same effect was observed in microglial cultures exposed to IFN-g or LPS + IFN-g (Fig. 2). This result suggested that Hippocampal-Cm contained soluble factors enhancing the reduction metabolism of microglial cells. Hippocampal-Cm and recombinant TGFb1 modulated O2S production in microglial cells Like Hippocampal-Cm (Fig. 1), recombinant TGFh1 inhibited O2S production of mixed glial cultures (Fig. 3I). HippocampalCm and recombinant TGFh1 decreased O2S production in cultures exposed to IFN-g or LPS + IFN-g and decreased the number of microglial cells producing O2S when triggered with PMA (Fig. 3I). Changes in O2S production were also accompanied by changes in microglial cell morphology. Microglial cells producing O2S had a rounded morphology characteristic of activated macrophages (von Bernhardi et al., 2001) but microglial cells exposed to Hippocampal-Cm or recombinant TGFh1 maintained the elongated morphology characteristic of resting cells (Fig. 3II).

232

R. Herrera-Molina, R. von Bernhardi / Neurobiology of Disease 19 (2005) 229–236

contrast, there was no increase in NO2 production when cultures were incubated with IFN-g in the presence of Hippocampal-Cm or when recombinant TGFh1 was added to cultures in NbB27 (Fig. 4). LPS + IFN-g elicited a 32-fold increase of NO2 production in mixed glial cultures incubated in NbB27 (Fig. 4; P b 0.001). Compared with cells incubated with LPS + IFN-g in NbB27, the addition of Hippocampal-Cm or recombinant TGFh1 decreased NO2 production by 48% and 53%, respectively, of cultures incubated with LPS + IFN-g (Fig. 4; P b 0.01). Hippocampal cells produced higher levels of TGFb1 under proinflammatory conditions TGFh1 is secreted bound to LAP with apparent mol. wt. of 53–55 kDa (Gleizzes et al., 1997; Munger et al., 1997). Western blot analysis showed that LAP-TGFh1 was present in Hippocampal-Cm and it increased by 70% when hippocampal cells were incubated with LPS + IFN-g (Figs. 5A and B; P b 0.001). Thus, TGFh bioactivity detected in Hippocampal-Cm (0.4 ng/ml) increased 6.5-fold (2.6 ng/ml) when hippocampal cell cultures were stimulated with LPS + IFN-g (Fig. 5C; P b 0.001). Anti-active TGFb1 antibody reversed the ability of Hippocampal-Cm to inhibit O2S production of mixed glial or microglial cell cultures To determine if endogenous TGFh1 mediated the modulatory effect of Hippocampal-Cm on O2S production, an antibody specific for active TGFh1 was added to Hippocampal-Cm. The antibody-treated Hippocampal-Cm was added to mixed glial (Fig. 6A) or microglial (Fig. 6B) cell cultures. Such antibody treatment reversed the inhibitory activity of Hippocampal-Cm. Thus, mixed glial cells incubated with LPS + IFN-g + PMA in the presence of antibody-treated Hippocampal-Cm had higher levels of O2S production than cells incubated with nonantibody-treated Hippocampal-Cm (Fig. 6A). Likewise, antibody treatment of Hippocampal-Cm reversed its ability to inhibit the O2S production by microglial cells incubated with IFN-g +

S

Fig. 1. Hippocampal-Cm reduces O2 production by microglial cells. Glial cultures were cultured with IFN-g (IFNg) or without (control) in NbB27 or Hippocampal-Cm for 24 h. For some samples, O2 production was triggered with PMA for the last 1.5 h of culture [IFNg (PMA)]. (A) In mixed glial cultures, microglial cells produced superoxide radicals (dark cells) when exposed to proinflammatory molecules. O2 production was inhibited when glial cells were cultured with Hippocampal-Cm. Scale bar = 200 Am. (B) Quantification of O2 production by isolated microglial cells. Results are expressed as fold-number increase of O2 production compared with control cultures. Values correspond to the mean F SE of three independent experiments performed in triplicate evaluated by the NBT colorimetric assay. **P b 0.001 compared with control conditions, #P b 0.05 and ##P b 0.001 compared with stimulated cultures maintained in fresh Neurobasal/B27 media.

S

S

S

S

Hippocampal-Cm and recombinant TGFb1 modulated NO production in mixed glial cultures NO2 production increased 12-fold when mixed glial cultures in NbB27 were incubated with IFN-g (Fig. 4; P b 0.001). In

Fig. 2. Hippocampal-Cm increases reduction metabolism. Microglial cells were treated with IFN-g or LPS + IFN-g (LI) in NbB27 or HippocampalCm for 24 h. The reduction of MTT by microglial cells was higher in the cells incubated with Hippocampal-Cm than in those cultured in NbB27. Values are means F SE from three independent experiments in triplicate. *P b 0.02 compared with cultures maintained in NbB27.

R. Herrera-Molina, R. von Bernhardi / Neurobiology of Disease 19 (2005) 229–236

233

Discussion A main finding of this study was that conditioned medium from hippocampal cell cultures (Hippocampal-Cm) inhibited the respiratory burst (O2S production) of microglial cells. O2S production was inhibited in cultures containing isolated microglial or mixed glial cells. Microglial cells produce less O2S in organotypic slice cultures from hippocampus maintained for 10 days compared with 4 days in vitro (Czapiga and Colton, 1999). The authors suggest that intercellular connections between neurons and microglial cells mediate the functional down-regulation of O2S production (Czapiga and Colton, 1999). Based on our present findings, however, an alternative hypothesis is that hippocampal cells could modulate O2S production by microglial cells through the secretion of soluble proteins like TGFh1 (see below). Another principal finding of our study was that the modulatory effect or ability of Hippocampal-Cm to inhibit O2S production by microglial cells was mimicked by the addition of recombinant TGFh1. Moreover, when Hippocampal-Cm was treated with an antibody specific for active TGFh1, a procedure shown to neutralize TGFh1 activity in a TGFh1 bioassay, the inhibitory activity of Hippocampal-Cm was eliminated. These findings suggest that TGFh1 secreted by hippocampal cells modulated microglial reactivity or O2S production and were supported by our morphological results, since microglial cells exposed to Hippocampal-Cm or recombinant TGFh1 maintained the elongated morphology characteristic of resting cells. On the other hand, microglial cells producing O2S had a rounded morphology characteristic of activated macrophages. There are contradictory reports regarding the modulation of the microglial O2S production by neurons. Whereas we and Czapiga and Colton reported that O2S production is inhibited by neurons (Czapiga and Colton, 1999), Sudo et al. (1998) showed that medium conditioned by cortical neurons up-regulated O2S production by microglial cells. However, in Sudo et al.’s study, cortical neurons were cultured in a sub-optimal neuronal culture medium (serum-free

S

Fig. 3. Hippocampal-Cm and recombinant TGFh1 similarly reduce O2 production by microglial cells. NBT assay of mixed glial cells treated with proinflammatory molecules with and without being triggered with PMA for 1.5 h. Proinflammatory molecules were added to NbB27, HippocampalCm, or NbB27 with TGFh1 (TGF-h1) for 24 h. Dark cells correspond to microglial cells secreting O2 . (I) Microglial cells incubated with recombinant TGFh1 or Hippocampal-Cm produced less O2 . The effect of recombinant TGFh1 and Hippocampal-Cm on microglial O2 production was more marked when cultures were treated with PMA. Representative results of three independent experiments are shown. Scale bar = 250 Am. (II) Higher magnification of the microglial cells present in the cultures incubated with LPS + IFN-g + PMA in: (A) NbB27, (B) Hippocampal-Cm, and (C) NbB27 with recombinant TGFh1. Scale bar = 75 Am.

S

S

S

PMA (Fig. 6B). Hippocampal-Cm treated with a non-neutralizing antibody specific for LAP-TGFh1 maintained its ability to inhibit O2S production (data not shown), indicating that it was the TGFh1 activity in Hippocampal Cm that inhibited O2S production. Relative to control cultures, the production of O2S by microglial cells incubated with IFN-g plus PMA in NbB27 or antibodytreated Hippocampal-Cm increased 6.5-fold ( P b 0.001); nonantibody-treated Hippocampal-Cm decreased the production of O2S by 50% (Fig. 6B; P b 0.001).

Fig. 4. Hippocampal-Cm and recombinant TGFh1 inhibit the production of NO. Mixed glial cultures were incubated with IFN-g or LPS + IFN-g (LI) in NbB27, Hippocampal-Cm, or NbB27 + 1 ng/ml TGFh1 for 24 h. Hippocampal-Cm and recombinant TGFh1 prevented the increase of NO2 production induced by IFN-g or LPS. NO2 production in control cultures was 2–4 AM. Results are expressed as fold-number increase of NO2 production compared with control cultures. Values are means F SE from three independent experiments (n = 6–9). **P b 0.001 compared with control cultures, ##P b 0.001 compared with cultures treated with IFN-g in NbB27 and #P b 0.01 compared with cultures treated with LI in NbB27.

234

R. Herrera-Molina, R. von Bernhardi / Neurobiology of Disease 19 (2005) 229–236

Further work will be needed to determine the participation of TGFh as a protective mechanism in vivo. We are performing intrathecal administration of pro- and anti-inflammatory cytokines to evaluate

Fig. 5. TGFh1 is present in Hippocampal-Cm. The presence of TGFh1 was evaluated by two methods. (A) LAP-TGFh1 was detected by Western blot analysis in Hippocampal-Cm under control condition (basal) and in hippocampal cultures incubated for 48 h with LPS + IFN-g (stimulated). (B) Densitometry analysis of the Western blot. Data correspond to means F SE and are expressed as the percentage of the TGFh1 present in the Hippocampal-Cm under stimulated conditions with reference to the basal secretion. Three independent samples were evaluated. (C) TGFh bioactivity in Hippocampal-Cm was determined by a cellular proliferation inhibition assay. Values are means F SE of 8–10 determinations of three independent cultures. TGFh1 production by hippocampal cells increased in the presence of LPS + IFN-g (**P b 0.001).

MEM) lacking neuronal growth factors and culture lasted for only 3 days. Under these conditions, cortical neurons probably were poorly developed, activated or even damaged by the deprivation conditions. On the other hand, cell culture experiments have certain limitations. It has been shown that microglial cell activation after injury of the optic nerve is reduced in vivo compared to that observed under in vitro conditions (Reichert and Rotshenker, 1996), suggesting that additional mechanisms may have a role in vivo.

Fig. 6. Anti-active TGFh1 antibody reverses the ability of HippocampalCm to inhibit O2 production of mixed glial or microglial cell cultures. Heat-activated Hippocampal-Cm was incubated with an antibody against active TGFh1 for 4 h. The glial cells were incubated with LPS + IFN-g + PMA in NbB27, Hippocampal-Cm or antibody-treated Hippocampal-Cm (+ anti-TGFh1) for 24 h. (A) Microglial cells showed O2 production (dark cells) when they were incubated in NbB27 or in antibody-treated Hippocampal-Cm. Scale bar = 100 Am. (B) Quantification of the production of O2 by microglial cells. Results are expressed as fold-number increase of O2 production compared with control cultures. Values are means F SE (n = 6) from two independent experiments. TTP b 0.001 compared with control cultures, ##P b 0.001 compared with cultures treated with IFN-g (PMA) in NbB27.

S

S

S S

R. Herrera-Molina, R. von Bernhardi / Neurobiology of Disease 19 (2005) 229–236

the effect of TGFh on microglial activation, and using hippocampal slices in culture. In tissue slices, the interaction of the various cell types is better preserved than in culture experiments. Besides its inhibitory effect on respiratory burst, HippocampalCm also decreased the production of NO in mixed glial cell cultures. This effect might have been due to the presence of TGFh1 in Hippocampal-Cm, since it was also mimicked by the addition of recombinant TGFh1. Previous studies have shown that TGFh1 decreases NO production induced by LPS or IFN-g in glial cultures (Ledeboer et al., 2000; Lieb et al., 2003; McCartney-Francis and Wahl, 2002; Vincent et al., 1997). TGFh1 abolished NO2 production induced by IFN-g whereas it only decreased by 50% when cells were treated with LPS + IFN-g. These results suggest that TGFh1 modulated only some of the activation pathways involved in the induction of NO production by proinflammatory molecules. A third main finding in this study was the demonstration of TGFh1 in Hippocampal-Cm. Hippocampal-Cm contained LAPTGFh1 and exhibited TGFh biological activity. Both the amount of LAP-TGFh1 secreted by the cells and the active TGFh increased when hippocampal cell cultures were exposed to proinflammatory molecules, suggesting that stimulated hippocampal cells increased the secretion of TGFh1. The expression of TGFh1 is up-regulated in hippocampal neurons after transient forebrain ischemia (Zhu et al., 2000) and activated astrocytes secrete TGFh1 (Brown, 1999; Meda et al., 2001; Zhu et al., 2000). TGFh1 may play a role in pathology and activated glial cells are an important source of TGFh1 in several pathologic states (Chen et al., 2002; Zhu et al., 2000). Recently, the existence of regulatory interactions between TGFh1 and IFN-g has been proposed (Ishida et al., 2004) as well as the deregulation of IFN-g signaling pathways in TGFh1 null mice (McCartney-Francis and Wahl, 2002). Moreover, TGFh1 appears to inhibit the production of IFN-g by mononuclear cells (Mouri et al., 2002). We propose that the increase in TGFh1 production observed in hippocampal cells in response to LPS + IFN-g could reflect a regulatory mechanism secondary to cell activation as part of a feedback loop. The induction of TGFh1 by proinflammatory molecules would limit the temporal and spatial extent of the inflammatory response. An effect that could involve the activation of NF-nB and ERK1, 2 pathways that appear to be neuroprotective. We are presently examining this possibility. TGFh1 has various functions (Hayashida et al., 2004; Ishida et al., 2004; Mouri et al., 2002). In the central nervous system, TGFh1 has a direct neuroprotective effect by increasing neuronal viability. TGFh1 is constitutively secreted by hippocampal neurons (Zhu et al., 2000) and it is associated with neuronal protection against degeneration caused by transient global ischemia and staurosporine-induced apoptosis (Henrich-Noack et al., 1996; Zhu et al., 2000). TGFh, as well as other endogenous neurotrophic factors, increase with age (Bye et al., 2001; Harry et al., 2000), apparently associated to changes in glial activation during aging (Nichols, 1999); whereas TGFh appears to decrease in several other tissues (Matsunaga et al., 2003). The up-regulation of neurotrophic factors suggests that the aged brain may have compensatory changes in response to aging. TGFh1 is involved in the activation of NF-nB and ERK1, 2 neuroprotective pathways (Zhu et al., 2002, 2004). Activated microglial cells are associated with the lesion-induced dendritic retraction of hippocampal neurons. TGFh1 deactivates microglial cells and abolishes dendritic retraction (Eyu¨poglu et al., 2003). The regional and cellular differences in the cytotoxicity of microglial cells probably depend on multiple factors related to cell identity

235

and environmental conditions. One of the factors could be the differential secretion of TGFh by different cell populations (e.g., the CA1 pyramidal layers of the hippocampus) and its upregulation after injury (Zhu et al., 2000). In summary, our results suggested that TGFh1 produced by hippocampal cells modulates O2S production (respiratory burst), NO2 production, and morphological changes of glial cells. TGFh1, then, could have an indirect neuroprotective effect by preventing over-activation of microglial cells with their attendant neurotoxic activity.

Acknowledgments This work was supported by the grant 1040831 from FONDECYT to RvB. We thank Dr. Howard Etlinger for his helpful suggestions and critical reading of the manuscript. R.H-M. is a Biochemistry student from the Faculty of Chemistry and Pharmaceutical Science, Universidad de Chile.

References Bal-Price, A., Brown, C., 2001. Inflammatory neurodegeneration mediated by nitric oxide from activated glia-inhibiting neuronal respiration, causing glutamate release and excitotoxicity. J. Neurosci. 17, 6480 – 6491. Benveniste, E.N., 1998. Cytokine actions in the central nervous system. Cytokine Growth Factor Rev. 9, 259 – 275. Brewer, G.J., Torricelli, J.R., Evege, E.R., Price, P.J., 1993. Optimised survival of hippocampal neurons in B27-supplemented neurobasal, a new serum-free medium combination. J. Neurosci. Res. 25, 567 – 576. Brown, D.R., 1999. Dependence of neurones on astrocytes in a co-culture system renders neurones sensitive to transforming growth factor h1induced glutamate toxicity. J. Neurochem. 72, 943 – 953. Bruce-Keller, A.J., Keeling, J.L., Keller, J.N., Huang, F.F., Camondola, S., Mattson, M.P., 2000. Antiinflammatory effects of estrogen on microglial activation. Endocrinology 141, 3646 – 3656. Bye, N., Zieba, M., Wreford, N.G., Nichols, N.R., 2001. Resistance of the dentate gyrus to induced apoptosis during ageing is associated with increases in transforming growth factor-beta 1 messenger RNA. Neuroscience 105, 853 – 862. Ca´ceres, A., Banker, G., Steward, O., Binder, L., Payne, M., 1984. MAP2 is localized to the dendrites of hippocampal neurons which develop in culture. Brain Res. 315, 314 – 318. Cai, J., Jones, D.P., 1998. Superoxide in apoptosis. J. Biol. Chem. 273, 11401 – 11404. Chao, C.C., Hu, S., Peterson, P.K., 1995. Modulation of human microglial cell superoxide production by cytokines. J. Leukocyte Biol. 58, 35 – 70. Chen, S., Luo, D., Strit, W., Harrison, J.K., 2002. TGF-h1 up regulates CX3CR1 expression and inhibits fractalkine-stimulated signaling in rat microglia. J. Neuroimmunol. 133, 46 – 55. Czapiga, M., Colton, C.A., 1999. Function of microglia in organotypic slice cultures. J. Neurosci. Res. 56, 644 – 651. De Sampaio, 2002. Neuro–Glia interaction effects on GFAP gene: a novel role for TGFh1. Eur. J. Neurosci. 16, 2059 – 2069. Ding, M., St. Pierre, B.A., Parkinson, J.F., Medberry, P., Wong, J.L., Rogers, N.E., Ignaro, L.J., Merril, J.E., 1997. Inducible nitric-oxide synthase and nitric oxide production in human foetal astrocytes and microglia. J. Biol. Chem. 272, 11327 – 11335. Eyqpoglu, L.Y., Bechmann, L., Nitsch, R., 2003. Modification of microglial function protects from lesion-induced neuronal alteration and promotes sprouting in the hippocampus. FASEB J. 17, 1110 – 1111. Flanders, K.C., Ren, R.F., Lippa, C.F., 1998. Transforming growth factorbetas in neurodegenerative disease. Prog. Neurobiol. 54, 71 – 85.

236

R. Herrera-Molina, R. von Bernhardi / Neurobiology of Disease 19 (2005) 229–236

Gleizzes, P.E., Munger, J.S., Nunes, I., Harpel, J.G., Mazzieri, R., Noguera, I., Rifkin, D.B., 1997. TGF-h latency: biological significance and mechanisms of activation. Stem Cells 15, 190 – 197. Harry, G.J., Lefebvre d’Hellencourt, C., Bruccoleri, A., Schmechel, D., 2000. Brain Behav. Immun. 14, 288 – 304. Hayashida, T., Decaestecker, M., Schnaper, W., 2004. Cross-talk between ERK MAP kinase and Smad signaling pathways TGF-h dependent response in human mesangial cells. FASEB J. 17, 1576 – 1578. Henrich-Noack, P., Prehn, J.H.M., Krieglstein, J., 1996. TGF-h1 protects hippocampal neurons against degeneration caused by transient global ischemia. Stroke 27, 1609 – 1614. Hu, S., Sheng, W.S., Peterson, P.K., Chao, C.C., 1995. Cytokines modulation of murine microglial cell superoxide production. Glia 13, 45 – 50. Ishida, Y., Kondo, T., Takayasu, T., Iwakura, Y., Mukaida, N., 2004. The essential involvement of cross-talk between IFN-g and TGF-h1 in the skin wound-healing process. J. Immunol. 172, 1848 – 1855. Kim, M.S., Ahn, S.M., Moon, A., 2002. In vitro bioassay for transforming growth factor-beta using XTT method. Arch. Pharm. Res. 25, 903 – 909. Krupinski, J., Kumar, P., Kumar, S., Kaluza, J., 1996. Increased expression of TGF-h1 in brain tissue after ischemic stroke in humans. Stroke 27, 852 – 857. Ledeboer, A., Breve´, J.J.P., Poole, S., Tilders, F.J.H., Van Dam, A.M., 2000. Interleukin-10, interleukin-4, and transforming growth factor-h differentially regulate lipopolysaccharide-induced production of proinflammatory cytokines and nitric oxide in co-cultures of rat astroglial and microglial cells. Glia 30, 134 – 142. Lee, S.J., Zhou, T., Choi, Ch., Zheng, W., Benveniste, E.N., 2000. Differential regulation and function of Fas expression on glial cells. J. Immunol. 164, 1277 – 1285. Lieb, K., Engels, S., Fiebich, B.L., 2003. Inhibition of LPS-induced iNOS and NO synthesis in primary rat microglial cells. Neurochem. Int. 42, 131 – 137. Matsunaga, S., Nagano, S., Onishi, T., Morimoto, N., Suzuki, S., Komiya, S., 2003. Age-related changes in expression of transforming growth factor-h and receptors in cells of intervertebral discs. Mittaud, P., Labourdette, G., Zingg, H., Guenot-Di Scala, D., 2002. Neurons modulate oxytocin receptor expression in rat cultured astrocytes: involvement of TGF-h and membrane components. Glia 37, 169 – 177. McCartney-Francis, N.L., Wahl, S.M., 2002. Dysregulation of IFN-g signaling pathways in the absence of TGF-h1. J. Immunol. 169, 5941 – 5947. McGeer, P.L., McGeer, E.D., 1995. The inflammatory response system of brain: implications for therapy of Alzheimer and other neurodegenerative diseases. Brain Res. Rev. 21, 195 – 218. McMillian, M.K., Thai, L., Hong, J.S., O’Callaghan, J.P., Pennypacker, K.R., 1994. Brain injury in a dish: a model for reactive gliosis. TINS 17, 138 – 142. Meda, L., Baron, P., Scarlato, G., 2001. Glial activation in Alzheimer’s disease: the role of Ah and its associated proteins. Neurobiol. Aging 22, 885 – 893. Miranda, S., Opazo, C., Larrondo, L.F., Mun˜oz, F.J., Ruiz, F., Leighton, F., Inestrosa, N.C., 2000. The role of oxidative stress in the toxicity induced by amyloid h-peptide in Alzheimer’s disease. Prog. Neurobiol. 62, 633 – 648. Mouri, H., Sakaguchi, K., Sawayama, T., Senoh, T., Ohta, T., Nishimura, M., Fujiwara, A., Terao, M., Shiratori, Y., Tsuji, T., 2002. Suppressive

effects of transforming growth factor-h1 produced by hepatocellular carcinoma cell lines on interferon-g production by peripheral blood mononuclear cells. Acta Med. Okayama 56, 309 – 315. Munger, J.S., Harpel, J.G., Gleizes, P.E., Mazzieri, R., Nunes, I., Rifkin, D.B., 1997. Latent transforming growth factor-h: structural features and mechanisms of activation. Kidney Int. 51, 1376 – 1382. Nichols, N.R., 1999. Glial responses to steroids as markers of brain aging. J. Neurobiol. 40, 585 – 601. O’Brien, M.F., Lenke, L.G., Lou, J., Bridwell, K.H., Joyce, M.E., 1994. Astrocyte response and transforming growth factor-beta localization in acute spinal cord injury. Spine 19, 2321 – 2329. Penkowa, M., Girallt, M., Carrasco, J., Hanberd, H., Hidalgo, J., 2000. Impaired inflammatory stress and neurodegeneration after brain injury in interleukin-6-deficient mice. Glia 32, 271 – 285. Pfeiffer, S., Gorren, A.C.F., Schmidt, K., Werner, E.R., Hanser, B., Bohle, D.S., Mayer, B., 1997. Metabolic fate of peroxynitrite in aqueous solution. J. Biol. Chem. 272, 3465 – 3470. Reichert, F., Rotshenker, S., 1996. Deficient activation of microglia during optic nerve degeneration. J. Neuroimmunol. 70, 153 – 161. Rook, G.A., 1985. A simple method for the solubilisation of reduced NBT, and its use as a colorimetric assay. J. Immunol. Methods 82, 161 – 167. Schwartz, L., Osborne, B., 1995. Cell Death. Academic Press, London, pp. 260. Sudo, S., Tanaka, J., Toku, K., Desaki, J., Matsuda, S., Arai, T., Sakanaka, M., Maeda, N., 1998. Neurons induce the activation of microglial cell in vitro. Exp. Neurol. 154, 499 – 510. Thannickal, V., Fanfurg, B.L., 2000. Reactive oxygen species in cell signaling. Am. J. Physiol.: Lung Cell. Mol. Physiol. 279, 1005 – 1028. Vincent, V.A., Tielders, F.J., Van Dam, A.M., 1997. Inhibition of endotoxin-induced nitric oxide synthase production in microglial cells by the presence of astroglial cells: a role for transforming growth factor h. Glia 19, 190 – 198. van Muiswinkel, F.L., Veerhuis, R., Eikelenboom, P., 1996. Amyloid h protein primes cultured rat microglial cells for an enhanced phorbol 12-myristate 13-acetate-induced respiratory burst activity. J. Neurochem. 66, 2468 – 2476. von Bernhardi, R., Ramirez, G., Matile, H., Dfbeli, H., 2001. Immobilized amyloid precursor protein constructs: a tool for the in vitro screening of glial cell reactivity. Eur. J. Neurosci. 14, 1 – 12. Vegeto, E., Bonincontro, C., Polli, G., Sala, A., Viappiani, S., Nardi, F., Brusadelli, A., Viviani, B., Ciana, P., Maggi, A., 2001. Estrogen prevents the lipopolysaccharide-induced inflammatory response in microglia. J. Neurosci. 21, 1809 – 1818. Wood, P., 1998. Neuroinflammation, Mechanism and Management. Humana Press Inc., Totowa, USA. Zhu, Y., Roth-Eichhorn, S., Braun, N., Culmsee, C., Rami, A., Krieglstein, J., 2000. The expression of transforming growth factor-beta 1 (TGF-h1) in hippocampal neurons: a temporary upregulated protein level after transient forebrain ischemia in the rat. Brain Res. 866, 286 – 298. Zhu, Y., Yang, G.Y., Ahlemeyer, B., Pang, L., Che, X.M., Culmsee, C., 2002. Transforming growth factor-h1 increases bad phosphorylation and protects neurons against damage. J. Neurosci. 22, 3898 – 3909. Zhu, Y., Culmsee, C., Klumpp, S., Krieglstein, J., 2004. Neuroprotection by transforming growth factor-h1 involves activation of nuclear factor-nB through phosphatidylinositol-3-OH kinase/Akt and mitogen-activated protein kinase-extracellular-signal regulated kinase1,2 signaling pathways. Neuroscience 123, 897 – 906.