Tumor Necrosis Factor Receptor-1 is Essential for LPS-Induced Sensitization and Tolerance to Oxygen—Glucose Deprivation in Murine Neonatal Organotypic Hippocampal Slices

Abstract

Inflammation and ischemia have a synergistic damaging effect in the immature brain. The role of tumor necrosis factor (TNF) receptors 1 and 2 in lipopolysaccharide (LPS)-induced sensitization and tolerance to oxygen—glucose deprivation (OGD) was evaluated in neonatal murine hippocampal organotypic slices. Hippocampal slices from balb/c, C57BL/6 TNFR1^−/-, TNFR2^−/-, and wild-type (WT) mice obtained at P6 were grown in vitro for 9 days. Preexposure to LPS immediately before OGD increased propidium iodide-determined cell death in regions CA1, CA3, and dentate gyrus from 4 up to 48 h after OGD (P < 0.001). Extending the time interval between LPS exposure and OGD to 72 h resulted in tolerance, that is reduced neuronal cell death after OGD (P < 0.05). Slices from TNFR1^−/- mice showed neither LPS-induced sensitization nor LPS-induced tolerance to OGD, whereas both effects were present in slices from TNFR2^−/- and WT mice. Cytokine secretion (TNFα and interleukin-6) during LPS exposure was decreased in TNFR1^−/- slices and increased in TNFR2^−/- as compared with WT slices. We conclude that LPS induces sensitization or tolerance to OGD depending on the time interval between exposure to LPS and OGD in murine hippocampal slice cultures. Both paradigms are dependent on signaling through TNFR1.

Keywords

excitotoxicity hippocampus inflammatory mediators neonates < experimental system tissue culture TNFα-induced tolerance

Introduction

Inflammation and ischemia are commonly induced in fetal circulation in pregnancies complicated by infection or placental dysfunction and are implicated in the pathogenesis of perinatal brain damage (Anslow, 1998). These insults target distinct cell populations and may result in damage of the developing brain with subsequent functional impairment.

It is now established that infants exposed to infection in utero are at higher risk of developing cerebral neuropathologies. Common antecedents are perinatal remote infection and the ensuing fetal or neonatal inflammatory response (Viscardi et al, 2004). The fetal inflammatory response is characterised by the accumulation of inflammatory cells in umbilical cord vessels, with a concomitant increase in the production of proinflammatory cytokines in the peripheral circulation (Dammann and Leviton, 2000). The immature fetal blood—brain barrier allows leakage into the brain parenchyma, leading to the exposure of brain parenchyma to inflammatory mediators, triggering the innate immune system of the fetal brain, which is causally linked to the development of perinatal central nervous system pathologies (Ogunyemi et al, 2003).

Perinatal inflammation typically results in white matter damage (Cai et al, 2003), with relatively little or no damage reported to the neuronal population. Animal studies have shown that systemic inflammation in the fetal circulation (Mallard et al, 2003) or locally induced inflammation within the central nervous system in 5–7-day old rats (Cai et al, 2003) results in white matter damage with histopathologic similarities to the white matter disease observed in preterm human infants. Conversely, inflammation induced by the endotoxin lipopolysaccharide (LPS) immediately before ischemia aggravates and extends brain injury to also include neuronal cells, including those in the hippocampus (Wang et al, 2006). Animal studies show that the temporal separation between induced inflammation and subsequent cerebral ischemia will determine whether sensitization or tolerance occurs (Eklind et al, 2005).

Lipopolysaccharide activates the innate immune system, primarily by binding to the Toll-like receptor-4, which leads to the induction of proinflammatory genes, among which is tumor necrosis factor α(TNFα), by the activation of several transcriptional pathways such as nuclear factor kappa B (NF-κB). Tumor necrosis factor α is a key proinflammatory cytokine that is rapidly upregulated during infection and exerts its biologic effects by signaling through two receptors, TNF receptor (TNFR) 1 and TNFR2. Hypothetically, TNFα signaling may be implicated in the mechanisms of both LPS-induced sensitization and tolerance to ischemia. Sensitization in response to TNFα stimulation has been shown to occur through a posttranslational increase in α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor density at synaptic terminals (Stellwagen et al, 2005), through the activation of TNFR1, and a concomitant endocytosis of γ-aminobutyric acid (GABA)_A receptors from inhibitory receptor terminals (Stellwagen et al, 2005). Ischemic tolerance in response to LPS exposure is induced as a result of an attenuation of the cytotoxic effects of TNFα in ischemia (Rosenzweig et al, 2007).

In vitro models are useful for revealing cellular and molecular responses that are not readily observed in vivo. In addition, extracellular conditions are more easily controlled, and the temporal extent of inflammation and ischemia more easily defined. The hippocampal slice culture allows for quantitative evaluation of cell damage over time in a piece of intact central nervous system tissue with various cell types preserved in a three-dimensional tissue architecture. Primarily, it was our aim to develop a model of inflammation and ischemia in the immature brain that would enable the study of interactions between the innate immune system of the brain and its specific mediators with those of induced experimental ischemia, and to study the levels of cell death that result over time. Using this model, we examined the role of the two receptors for TNFα, TNFR1 and TNFR2, respectively, in LPS-induced sensitization and tolerance by (1) blocking the activity of the two receptors for TNFα selectively using specific antagonistic antibodies against the receptors for TNFR1 and TNFR2, and (2) by using slices from transgenic animals lacking either of the two receptors.

Materials and methods

Preparation and Maintenance of Organotypic Hippocampal Tissue Cultures

All animal experiments were approved by the Malmo/Lund ethical committee on animal experiments (approval number M 73-04). Offspring from date-mated balb/c mice (Harlan, Scandinavia, Denmark) and C57BL/6 wild-type (WT), C57BL/6 TNFR1^−/-, and C57BL/6 TNFR2^−/- were used for experiments. In-house breeding generated TNFR1^−/- and TNFR2^−/- mice by crossing TNFR1/R2^−/- mice (Peschon et al, 1998; obtained from Jackson Laboratory, Bar Harbor, ME, USA) with WT C57BL/6 mice (B&K Universal, Stockholm, Sweden) and interbreeding heterozygous offspring followed by selection of appropriate genotypes as determined by polymerase chain reaction. Wild-type C57BL/6 mice were used as controls for TNFR1^−/- and TNFR2^−/- mice, respectively.

In brief, hippocampal organotypic tissue cultures were prepared essentially according to the method of Stoppini, as described earlier (Cronberg et al, 2004; Stoppini et al, 1991), from 6-day-old mice. In short, hippocampi were dissected in ice-cold Hank's balanced salt solution, with 20 mmol/L HEPES (4-(2-hydroxyethyl)-1-piperazine-ethanesulfonic acid), 100 units penicillin—streptomycin per milliliter, and 3 mg/mL D-glucose and cut into 250-μm-thick slices using a McIllwain Tissue Chopper. Sections were plated onto Millicell culture inserts, one per insert (0.4 μm Millicell-CM, 12 mm in diameter, Millipore Corp, Bedford, MA, USA). Slices were cultured in 35°C and 90% to 95% humidity in a CO₂ incubator for 9 days before experiment. The culture medium consisted of 50% modified Eagle's medium with Earle's balanced salt solution (MEM), 25% horse serum, and a 18% Hank's balanced salt solution and was supplemented with 4 mmol/L L-glutamine, 50 units penicillin—streptomycin per milliliter, and 20 mmol/L D-glucose. The pH was adjusted to 7.2 using NaHCO₃. During the first week of culture, 2% of the supplement B27 was included in the medium and thereafter omitted and replaced with the same volume of water. On DIV9, when the experiment was initiated, horse serum was omitted and the volume was replaced with MEM. This medium was used throughout the experimental protocol. Before experimental start, the culture medium was changed the day after the preparation and thereafter three times a week. All substances were obtained from Invitrogen (Carlsbad, CA, USA) with the exception of D-glucose, which was from Sigma-Aldrich (St Louis, MO, USA).

Experimental Protocol

All cultures used in one experiment were prepared from mice pups from one to two females with litters born on the same day. Experiments were started on DIV9, and slices were assorted into four groups with one slice from each individual mouse per experimental group and six slices per group. Slices were incubated with propidium iodide (PI, Sigma-Aldrich) for 1 h to assess the background levels of cell death. Photomicrographs were taken using fluorescent light microscopy and slices were again sorted to include only undamaged slices displaying no PI uptake. Light microscopy was used to include only slices with an organotypic anatomy. Propidium iodide was present in all groups throughout the experimental protocol and all experiments were carried out in serum-free medium. The protocol was performed in four groups: Control = cultures exposed to medium changes at corresponding time points as experimental groups, preincubation period of 24 h followed by transfer to fresh medium for 48 h; LPS = incubation with LPS (1 μg/mL; Sigma-Aldrich) for 24 h followed by transfer to fresh medium for 48 h; Oxygen—glucose deprivation (OGD) = preincubation without LPS for 24 h, followed by OGD and thereafter transfer to fresh medium for 48 h; LPS + OGD = incubation with LPS for 24 h followed by OGD and thereafter transfer to fresh medium for 48 h. To study the influence of time between LPS incubation and OGD, sets of two groups, one control group and one LPS group, were transferred to fresh medium after the initial 24 h of incubation with or without LPS and incubated for 24, 48, or 72 h, with no additional changes of medium, before OGD. An overview of the experimental protocol is given in Figure 1.

Figure 1

Schematic illustration of experimental design. LPS, lipopolysaccharide; PI, propidium iodide; OGD, oxygen—glucose deprivation; IHC, immunohistochemistry. Cytokine release in medium was determined only in experiments with a time interval of 0 h between insults.

Before OGD, cultures were washed once in prewarmed phosphate-buffered saline (PBS) and transferred to the anaerobic incubator (Elektrotek Ltd, Keighley, UK) in an empty prewarmed plate according to previously described protocols for OGD (Cronberg et al, 2004). The anaerobic incubator had an atmosphere comprising 10% H₂, 5% CO₂, and 85% N₂, and the temperature was maintained at 35.0°C ± 0.3°C. Inside the incubator, the slices were transferred to wells containing preequilibrated OGD medium. The OGD medium consisted of, in mmol/L concentrations, 2 CaCl₂, 125 NaCl, 25 NaHCO₃, 2.5 KCl, 1.25 NaH₂PO₄, 2 MgSO₄, and 10 sucrose and had a pH of 7.4. After 15 mins of OGD, the slices were transferred to fresh oxygenated culture medium and placed in the CO₂ incubator. In all groups, slices were cultured for 48 h after the time point of OGD, and photomicrographs were taken in parallel at 4, 12, 24, and 48 h after OGD.

Quantification of Cell Damage

Propidium iodide was used as a marker for cell death and was included in the medium in all groups throughout the experimental protocol (Cronberg et al, 2004). Photomicrographs of PI-stained slices were obtained before experimental start and at time points of 4, 12, 24, and 48 h after OGD in all groups. Densitometric analysis was performed on photomicrographs to obtain quantitative values of cell death. Propidium iodide intensity was measured using standardized prototypes that were applied over corresponding neuronal regions of interest: the CA1, CA3, and the dentate gyrus. Background uptake was measured in an undamaged area outside the CA2 region in each individual slice, and values of cell death in regions of interest were obtained by subtracting the levels of mean fluorescent intensity in the background region in each slice from mean fluorescent intensity values in the three regions of interest. For image processing, the commercial software Image-Pro Plus 4.0 (Media Cybernetics, MD, USA) was used.

Evaluation of Tumor Necrosis Factor Receptors 1 and 2

To evaluate the roles of TNFR1 and TNFR2 in LPS-induced sensitization and in pure OGD, antagonistic antibodies specific against TNFR1 (50 μg/mL; BioLegend, San Diego, CA, USA) and TNFR2 (50 μg/mL; BioLegend, San Diego, CA, USA), respectively, were included in the medium during 24 h of LPS stimulation or in the control medium before OGD. The experiment was performed as described above, using slices from balb/c mice, by adding the respective blocking antibodies to the medium during 24 h of LPS stimulation. To exclude nonspecific interactions of the respective antibodies with the slice culture other than specific antigen binding and blocking of receptor function, we performed the same experiment using heat-inactivated antibodies at the same final concentration.

The roles of TNFR1 and TNFR2 in LPS-induced sensitization and in LPS-induced tolerance were further investigated using TNFR1^−/- and TNFR2^−/- C57BL/6 mice, using the same experimental protocol as was applied to WT slices as indicated and described above.

To evaluate the involvement of AMPA and N-methyl-D-aspartate (NMDA) receptor activation in cell death after LPS exposure followed by OGD, we preincubated slices for 1 h before OGD and coincubated during OGD with the specific AMPA receptor antagonist CNQX (Sigma-Aldrich) and the specific NMDA receptor antagonist MK-801 (Sigma-Aldrich) at a final concentration of 100 μmol/L in both cases. Effects on subsequent cell death were determined as described above.

Immunohistochemistry

Slices were fixed for immunohistochemistry 48 h after OGD. Unsectioned slices were immunostained for the microglial marker F4/80 and for the neuronal nuclear marker NeuN. Slices were fixed with 4% paraformaldehyde in PBS for 10 mins and then stored at 4°C in PBS until used.

Endogenous peroxidase activity was quenched with 3% hydrogen peroxide in 10% methanol for 30 mins . After rinsing, slices were blocked with 5% normal horse serum for a NeuN stain and normal swine serum for an F4/80 stain, and diluted in PBS with 1% Triton X-100 for 1 h before overnight incubation with the primary antibody monoclonal mouse anti-NeuN (Chemicon, Temecula, CA, USA, diluted 1:1,000 in PBS containing 5% normal horse serum and 1% Triton X-100) and monoclonal rat anti-F4/80 (Serotec, Kidlington, Oxford, UK, diluted 1:1,000 in PBS containing 5% normal rabbit serum and 1% Triton X-100). After rinsing, slices were incubated with a biotinylated secondary antibody, horse anti-mouse IgG antibody for a NeuN stain, and rabbit anti-rat IgG antibody for an F4/80 stain. Both antibodies were diluted 1:1,000 in PBS containing 1% Triton X-100 and 2% normal horse serum and swine serum for the respective stains (as indicated above) for 90 mins . After rinsing, slices were incubated with the avidin—biotin—peroxidase complex (Vectastain Elite, Vector Laboratories Inc., Burlingame, CA, USA) for 1 h before reaction with the chromogen 3,3 diaminobenzidine supplemented with 0.06% nickel (DAB-Safe; 0.5 μg/mL; Saveen-Werner AB, Malmo, Sweden). Both sera, both secondary antibodies, and the avidin—biotin—peroxidase complex (Vectastain Elite) were from Vector Laboratories.

Micrographs of immunostained slices were prepared using a microscope (Olympus BX51) equipped for bright-field microscopy with a digital camera (Olympus U-PMTVC). Figures show original, unmodified photomicrographs.

Cytokine Measurements

Levels of secreted cytokines in culture medium were determined at 12 and 24 h after the start of LPS preincubation and at 12, 24, and 48 h after OGD in slices from WT balb/c, WT C57BL/6 mice and in C57BL/6 TNFR1^−/- and TNFR2^−/- strains, respectively. Samples of medium (50 μL) were immediately frozen on dry ice and were thereafter stored at −80°C until analyzed in one batch. Levels of TNFα, interferon-γ, monocyte chemoattractant protein-1 (MCP-1), interleukin (IL)-6, IL-10, and IL-12p70 were determined by cytometric bead assay (Becton Dickinson (BD) Biosciences, San Jose, CA, USA) and flow cytometry according to the manufacturer's recommendations and as described earlier (Hansen-Pupp et al, 2005). Two-color flow cytometric analyses were performed using an FACSCalibur Flow Cytometer (BD Biosciences). Data were acquired and analyzed using BD Biosciences CBA software. The lower limit of detection for the various cytokines evaluated ranged from 2 to 10 pg/mL. For results above the upper limit of detection, serial dilution of the sample was performed to accurately determine cytokine levels. A level of < 0.1 pg/mL was regarded as nondetectable.

Statistics

Statistical analysis was performed using SPSS v14.0 for Windows. Data are expressed as mean^± s.e.m. All compared groups were run in parallel inside the anaerobic incubator. Experiments on the effects of genetic knockouts for TNFR1 and TNFR2 were run and analyzed separately. Differences between groups were assessed using ANOVA for repeated measures with post hoc Bonferroni correction for multiple group comparisons. Variability between experimental dates was adjusted for by including date as a variable. P-values of < 0.05 were considered significant.

Results

Sensitizing Effect of Lipopolysaccharide to Oxygen—Glucose Deprivation

Cell damage induced by pure OGD was present from 4h, reached its maximum by 12 h, and was mainly localized to the CA1 region with some damage in the CA3 region and the dentate gyrus. Pre-incubation with LPS during 24 h before OGD severely aggravated PI-determined cell damage in all studied neuronal sub-regions as compared with sole exposure to OGD (P < 0.001, respectively). Increased cell damage resulting from LPS exposure before OGD was present from 4 h and onward. The time course of cell damage was similar to pure OGD with maximum cell damage by 12 h. Sole exposure to LPS was associated with a small but significant increase in cell damage in the dentate gyrus as compared with control slices (P < 0.05). Control slices exhibited no cell damage as determined by PI. Representative photomicrographs and mean values of PI-determined cell death are given in Figure 2.

Figure 2

Lipopolysaccharide-induced sensitization to OGD in murine hippocampal slice cultures. Temporal development of cell damage after LPS exposure (1 μg/mL) for 24 h, 15 mins oxygen—glucose deprivation (OGD), or 24h LPS + 15 mins OGD and in control slices. (A:1) Fluorescent images of representative cultures showing development of propidium-iodide (PI) determined cell damage. The same culture is shown at repeated time points of recovery. (A:2) Mean PI intensity in the standardized neuronal areas CA1, CA3, and the dentate gyrus (DG) in control slices (shaded bars), only 24 h LPS (black bars), 15 mins OGD (hatched bars), and after 24 h LPS + 15 mins OGD (white bars) at repeated points of recovery. Data are shown as means ± s.e.m. with n = 18. LPS + OGD versus OGD, P < 0.001 in the CA1, CA3 and DG; OGD versus LPS and control, P < 0.01, respectively, in the CA1 region and P < 0.05, respectively, in the CA3 region; LPS versus control, P < 0.05 in DG. Repeated-measures ANOVA with posthoc Bonferroni. (B) NeuN-stained slice (1) and fluorescent image (3) of propidium-iodide determined cell damage at 48 h after 15 mins OGD. Corresponding illustrations at 48 h after LPS (1 μg/mL for 24 h) + 15 mins OGD (2 + 4). (C) Effect of NMDA receptor antagonist (MK-801) and AMPA receptor antagonist (CNQX) on cell damage after LPS exposure followed by OGD and after pure OGD. Mean PI intensity in the standardized neuronal areas CA1, CA3, and the dentate gyrus after 24 h LPS + 15 mins OGD (dark-shaded bars), after 15 mins OGD (light-shaded bars), and in control slices (white bars) at a time corresponding to 24 h after OGD. Data are shown as means ± s.e.m. with n = 12. The NMDA receptor antagonist MK-801 reduced cell damage in the CA1 region (P < 0.01) and in the dentate gyrus (P < 0.01) after LPS + OGD, whereas the AMPA receptor antagonist CNQX had a less pronounced effect in the corresponding subregions (P < 0.05, respectively). Both antagonists reduced cell damage after pure OGD in the CA1 and CA3 regions (all P < 0.05). Repeated-measures ANOVA with post hoc Bonferroni.

Results from staining with NeuN at 48 h correlated well with distribution and degree of PI-determined cell death in neuronal subregions. Control cultures exhibited an intact organotypic distribution of NeuN-positive cells, and sole exposure to LPS did not affect the density and distribution of NeuN-positive neurons as compared with controls. Pure OGD revealed a loss of viable neurons largely restricted to the CA1 region. Cultures exposed to LPS followed by OGD showed a substantial loss of neuronal cell layers in CA1, CA3 and the dentate gyrus and exhibited condensed nuclei indicative of cell death. Representative illustrations of immunostaining with NeuN and PI-determined cell death are given in Figure 2B.

Excitotoxicity as a cause of neuronal death after exposure to LPS followed by OGD was investigated by blocking AMPA and NMDA receptor functions using the receptor antagonists CNQX and MK-801, respectively, as described above. Both receptor antagonists prevented cell death after pure OGD in the CA1 and CA3 regions (all P < 0.05). In cultures exposed to LPS followed by OGD, blocking of NMDA receptor function with MK-801 prevented cell death in the CA1 region (P < 0.01) and in the dentate gyrus (P < 0.01). Incubation with the AMPA receptor antagonist CNQX reduced levels of cell death in the CA1 region and in the dentate gyrus after OGD with LPS preexposure (P < 0.05), although less efficiently than the NMDA receptor antagonist MK-801. Neither CNQX nor MK-801 had any effect on cell death in the CA3 region after OGD with LPS preexposure. Effects of blocking NMDA and AMPA receptor functions on mean values of PI-determined cell death are given in Figure 2C.

Microglial Activation during Lipopolysaccharide Exposure and after Oxygen—Glucose Deprivation

Hippocampal cultures from the respective groups were stained with the microglial marker F4/80 at 12 h after the start of LPS exposure and at 24 and 48 h after subsequent OGD. Changes in staining intensity for F4/80 and morphology have been described as early sensitive indicators of microglial activation (Andersson et al, 1991). Control cultures displayed characteristically quiescent microglia with ramified cell bodies at all time points. LPS-exposed cultures exhibited microglia with a swollen ameboid morphology at 24 h after LPS preexposure and OGD that was more pronounced at 48 h, with microglia presenting a rough, densely stained plasma membrane. Microglial morphology did not differ between slices exposed to LPS only and those exposed to LPS + OGD, although after LPS + OGD the slices exhibited increased numbers of activated microglia. Pure OGD resulted in a characteristic distribution of activated microglia that were localized to the CA1-CA3 cell band. Representative illustrations of immunostaining with F4/80 in slices at a time point corresponding to 48 h after OGD are given in Figure 3A.

Figure 3

Lipopolysaccharide induces immune activation in murine hippocampal slice cultures. (A) F4/80-stained slices illustrating microglial morphology. 1 = control culture only exposed to medium changes; 2 = 48 h after 15 mins OGD; 3 = 48 h after LPS (1 μg/mL for 24 h); 4 = 48 h after LPS (1 μg/mL for 24 h) + 15 mins OGD. Scale bars in 1 through 4 = 400 μm. (B) Left panel = mean concentrations of TNFα (dot), IL-6 (circle), and MCP-1 (triangle) in medium at 12 and 24 h during LPS exposure. Control levels (squares). TNFα, IL-6, and MCP-1 were increased in cultures exposed to LPS as compared with control cultures (P < 0.01, respectively). Right panel = mean concentration of MCP-1 in medium at 12, 24, and 48 h after OGD with LPS preexposure (triangle) and without LPs (square). N = 4 per experimental group at each time point. Error bars denote 1 s.e.m.

Secreted levels of the cytokines TNFα, IL-6, and MCP-1 were increased at 12 h and further increased at 24 h in cultures exposed to LPS as compared with control cultures (P < 0.01, respectively). Secreted levels of IL-10, interferon-γ, and IL-12p70 were not increased in LPS-exposed cultures as compared with controls and were frequently below detection limits.

Secreted levels of MCP-1 were significantly increased at 24 and 48 h after LPS exposure followed by OGD as compared with pure OGD and controls (P < 0.01, respectively). None of the remaining cytokines exhibited increased levels after either combined exposure to LPS and OGD or sole OGD as compared with controls. Mean levels of cytokines during LPS exposure and after OGD are given in Figure 3B.

Role of Tumor Necrosis Factor Receptors in Sensitizing Effect of Lipopolysaccharide on Oxygen—Glucose Deprivation

To investigate the roles of TNFR1 and TNFR2, respectively, blocking antibodies against the two receptors were added during the 24 h of LPS exposure and in control medium before pure OGD. Blockade of TNFR1 or TNFR2 had no modifying effect on cell death after pure OGD (data not shown). The addition of antibodies against TNFR1 resulted in an almost complete reversal of the sensitizing effect of LPS to OGD as determined by PI uptake, whereas the addition of blocking antibodies against TNFR2 resulted in augmentation of cell death after LPS exposure and OGD (P < 0.01). The effects of blocking antibodies to both TNFR1 and TNFR2 were most pronounced in regions CA1 and dentate gyrus and not significant in the CA3 region. The addition of heat-denaturated antibodies had no modifying effect on cell death (data not shown). Representative photomicrographs and mean values of PI-determined cell death are given in Figure 4.

Figure 4

Evaluation of the role of TNFR1 and TNFR2 in LPS-induced sensitization to OGD using antagonistic antibodies. Temporal development of cell damage after LPS exposure (1 μg/mL) for 24 h before 15 mins oxygen—glucose deprivation (OGD) after adding antibodies to either TNFR1 or TNFR2 (50 μg/mL, respectively) during LPS exposure. (A) Fluorescent images of representative cultures showing the development of propidium iodide (PI) determined cell damage. The same culture is shown at repeated time points of recovery. (B) Mean PI intensity in the standardized neuronal areas CA1, CA3, and the dentate gyrus in cultures exposed to 24 h LPS followed by 15 mins OGD (hatched bars), 24 h LPS + abTNFR1 followed by 15 mins OGD (black bars), and 24 h LPS + abTNFR2 followed by OGD (white bars) at repeated points of recovery. Data are shown as means ± s.e.m. with n = 12. Antibody against TNFR1 reduced the sensitizing effect of LPS on OGD in CA1 and DG (P < 0.01, respectively), whereas antibody to TNFR2 increased the sensitizing effect in CA1 and DG (P < 0.05, respectively). Repeated measures ANOVA with post hoc Bonferroni.

To further investigate the role of TNFRs 1 and 2, we examined the effect of LPS exposure before OGD on PI-determined cell death in hippocampal slices from TNFR1^−/-, TNFR2^−/-, and C57BL/6 WT mice, respectively. The sensitizing effect of LPS on subsequent OGD was successfully replicated using WT C57BL/6 mice. The respective knockout strains were tested in separate experiments. In experiments using slices from TNFR1^−/- mice, the sensitizing effect of LPS exposure before OGD was absent in all evaluated subregions, whereas slices from TNFR2^−/- and C57BL/6 WT mice exhibited a similar and pronounced sensitizing effect of LPS exposure on OGD in the CA1, CA3, and the dentate gyrus. These results thus correlated well with the experiments using antagonistic antibodies to TNFR1 and TNFR2, respectively. Representative photomicrographs and mean values of PI-determined cell death are given in Figure 5.

Figure 5

Roles of TNFR1 and TNFR2 in LPS-induced sensitization to OGD evaluated in TNFR1^−/- and TNFR2^−/- hippocampal slice cultures. Cell damage after LPS exposure (1 μg/mL) for 24 h before 15 mins oxygen—glucose deprivation (OGD) in hippocampal slice cultures from C57BL/6 WT, TNFR1^−/-, and TnFR2^−/- mice. (A) Fluorescent images of representative cultures showing development of propidium iodide (PI) determined cell damage. The same culture is shown at repeated time points of recovery. (B) Mean PI intensity in the standardized neuronal areas CA1, CA3, and the dentate gyrus after 24h LPS + 15 mins OGD (white bars), 15 mins OGD (hatched bars), only 24 h LPS (black bars), and in control slices (shaded bars) at a time point corresponding to 24 h after OGD. Data are shown as means ± s.e.m. with n = 12. Preexposure to LPS before OGD aggravated cell death in the CA1 and CA3 area, and in the dentate gyrus as compared with pure OGD in cultures from WT and TNFR2^−/- mice (P < 0.01 for all comparisons). Cultures from TNFR1^−/- mice did not exhibit any sensitizing effect of LPS exposure to OGD. Repeated measures ANOVA with post hoc Bonferroni.

Cytokines secreted from slices from TNFR1^−/-, TNFR2^−/-, and C57BL/6 WT mice, respectively, were determined during LPS exposure as described earlier. The levels of TNFα, IL-6, and MCP-1 were considerably decreased during LPS exposure from TNFR1^−/- slices as compared with the corresponding levels from TNFR2^−/- and C57BL/6 WT mice (P < 0.01, respectively). Slices from TNFR2^−/- mice secreted higher levels of TNFα and IL-6 during LPS exposure as compared with slices from WT mice (P < 0.05, respectively). The mean levels of secreted cytokines from slices from the respective mice strains during LPS exposure are given in Figure 6.

Figure 6

Roles of TNFR1 and TNFR2 in LPS-induced cytokine release. Mean levels of TNFα, MCP-1, and IL-6 at 0, 12, and 24 h during LPS exposure secreted from hippocampal slices from C57BL/6 WT (dots), TNFR1^−/- (squares), and TNFR2^−/- (triangles) mice, respectively. Data are shown as means ± s.e.m. with n = 4. Mean levels of TNFα, MCP-1, and IL-6 from TNFR1^−/- slices were lower than those from WT and TNFR2^−/- slices (P < 0.01 for all comparisons). Mean levels of TNFα and IL-6 from TNFR2^−/- slices were higher than those from WT slices (P < 0.05, respectively). Repeated measures ANOVA with post hoc Bonferroni.

Effect of Varied Time Interval between Insults on the Sensitizing Effect of Lipopolysaccharide on Oxygen—Glucose Deprivation

The effect of increasing the time interval between LPS exposure and OGD from 0 to 24, 48, and 72 h was evaluated on PI-determined cell damage. The sensitizing effect of LPS on OGD was no longer significant with a 24-h interval between insults in any of the studied neuronal subregions. Extending the interval between LPS exposure and OGD to 72 h resulted in a significant decrease in PI-determined cell damage after LPS preexposure to OGD as compared with only OGD (P < 0.05) in the CA1 region and in the dentate gyrus (P < 0.05, respectively) but not in the CA3 region.

The same protocol was applied to hippocampal slice cultures from TNFR1^−/-, TNFR2^−/-, and C57BL/6 WT mice. In experiments using cultures from TNFR1^−/- mice, exposure to LPS at 24, 48, and 72 h before OGD did not result in significant differences in PI-determined cell damage in neuronal subregions as compared with slices subjected only to OGD at any of the studied time intervals. Cultures from both TNFR2^−/- and C57BL/6 WT mice exhibited a sensitizing effect of LPS exposure on OGD with a time interval of 0 h but not at 24 and 48 h, whereas a significant decrease in PI-determined cell damage was observed at a time interval of 72 h between LPS exposure and subsequent OGD in the CA1 region and the dentate gyrus (P < 0.05, respectively). The mean values of PI-determined cell death in slices from the respective mice strains are given in Figure 7.

Figure 7

Roles of TNFR1 and TNFR2 in LPS-induced tolerance to OGD. Cell damage after LPS exposure (1 μg/mL) for 24 h with a 72-h time interval before 15 mins oxygen—glucose deprivation (OGD) in hippocampal slice cultures from C57BL/6 WT, TNFR1^−/-, and TNFR2^−/- mice. Mean PI intensity in the standardized neuronal areas CA1, CA3, and dentate gyrus (DG) in cultures exposed to 24 h LPS + 15 mins OGD (hatched bars) and in cultures only exposed to 15 mins OGD (white bars). Preexposure to LPS 72 h before OGD decreased cell death in the CA1 area and the DG in cultures from WT (P < 0.05, respectively) and from TNFR2^−/- mice (P < 0.05, respectively). Preexposure to LPS in TNFR1^−/- cultures at 72 h before OGD had no modifying effect on cell death. Data are shown as means ± s.e.m. with n = 12. Repeated measures ANOVA.

Discussion

Together, our results provide evidence for the role of both receptors for TNFα, TNFR1 and TNFR2, in LPS-induced sensitization and tolerance to subsequent OGD in immature organotypic hippocampal slice cultures. We show that LPS induces sensitization and tolerance to subsequently induced OGD as a consequence of varied time intervals between the insults. Our results indicate that TNFα plays a central role in LPS-induced sensitization and tolerance to OGD by activating TNFR1, as both effects were absent in slices lacking the gene for TNFR1 or in an experimental set-up where the activation of TNFR1 was blocked by specific antagonistic antibodies. TNFR2 appears to have a modulatory role in LPS-induced sensitization to OGD, as blocking this receptor increased neuronal cell death.

Inflammation and ischemia are important causes of disturbed brain development in the newborn. To study the interaction between these insults, we developed an in vitro model of immature brain tissue using murine hippocampal slices obtained at P6 with experiments performed after 9 days of culture. The murine brain at P15 has a maturity similar to that of the human term brain (Muller et al, 1993). Neurons in hippocampal slices during culture have been shown to evolve through different stages of development at a pace similar to that found in vivo (De Simoni et al, 2003). We propose that the results obtained in the presented model are representative of the interaction of inflammation and ischemia in the human newborn brain.

Lipopolysaccharide-Induced Cytokine Release and Activation of Microglia

Lipopolysaccharide exposure in hippocampal slices activated the innate immune response in slices, as evidenced by increased levels of secreted proinflammatory cytokines, and an activated microglial morphology, as assessed by immunohistochemistry for F4/80. The inflammatory response to LPS in rat hippocampal cultures has been described (Huuskonen et al, 2005), and we here replicate some of these findings in murine hippocampal slices. During LPS exposure, increased levels of secreted cytokines were found for TNFα, IL-6, and MCP-1. The levels of secreted TNFα during LPS exposure reached 5 ng/mL corresponding to concentrations used by others for administration of exogenous TNFα in both in vivo (Rosenzweig et al, 2007) and in vitro studies (Kuno et al, 2005). Previous studies show that exposure to LPS intracerebroventricularly (Pang et al, 2003) does not result in neuronal cell death, which is also the case in rat hippocampal slices in vitro (Huuskonen et al, 2005). In agreement with those results, we observed that LPS exposure and cytokine exposure at high concentrations in hippocampal cultures did not result in neural cell death in the CA cell band. Significant but scarce levels of neuronal cell death in LPS-exposed slices were found in the granule cells of the dentate gyrus, implying differential susceptibility of granule cells and pyramidal cells to proinflammation.

Neither TNFα nor IL-6 was increased after pure OGD, which differs from previous studies, showing a prominent release of these cytokines after ischemia in vivo (Saito et al, 1996) and in hippocampal slices in vitro (Fernandez-Lopez et al, 2005). In contrast, the levels of MCP-1 increased after OGD in LPS preexposed slices, suggesting a role for this chemokine in postischemic processes. Monocyte chemoattractant protein-1 has been shown to be released by neurons during perinatal ischemia in vivo (Ivacko et al, 1997), and is implicated in the pathogenesis of ischemic cell death, as transgenic mice deficient in the gene for MCP-1 display smaller infarcts in a model of middle cerebral artery occlusion (Hughes et al, 2002).

Lipopolysaccharide-Induced Sensitization and Tolerance to Oxygen—Glucose Deprivation

Depending on the applied time interval between the insults in organotypic hippocampal slice cultures, LPS induced both sensitization and tolerance to subsequent OGD. At 0h time interval between insults, LPS sensitized all neuronal regions of the hippocampus to subsequent OGD, whereas a time interval of 72 h resulted in ischemic tolerance. Our results correlate well with in vivo data in neonatal rats showing LPS-enhanced vulnerability to hypoxia/ischemia at a time interval of 2 to 6 h between insults (Coumans et al, 2003; Eklind et al, 2001; Eklind et al, 2005). Conversely, a preconditioning effect of LPS to subsequent cerebral hypoxia/ischemia in vivo was observed at 24 h between insults, whereas an interval of 72 h produced a sensitizing effect (Eklind et al, 2005). The difference in the required time interval between the in vitro slice culture system and that previously described in vivo may have several explanations. The activation of peripheral organ systems present in vivo may serve to both amplify the immune response, which is elicited upon exposure to LPS, and accelerate and amplify counter-regulatory mechanisms, which result in a tolerant state. Amplification of the inflammatory response may result from the activation of the blood-borne immune system and the endothelium of the blood—brain barrier (Vlasov et al, Counter-regulatory or anti-inflammatory mechanisms, which are activated in vivo and not present in tissue culture in vitro, include the activation of the sympathetic nervous system (Prass et al, 2003), the neuroendocrine system (Rybnikova et al, 2007), and glucocorticoid release (Ikeda et al, 2006). These factors have been shown to be involved in induced ischemic tolerance and their activation may accelerate the processes that eventually result in preconditioning. Using a culture system with a preserved architecture of glial and neuronal cells, we show that sensitization and tolerance in response to LPS exposure is operant within the brain and does not require systemic interaction.

The mechanisms involved in LPS-induced tolerance are not well understood. There is emerging evidence indicating a role for cytokines not only in LPS-induced sensitization but also in LPS-induced ischemic tolerance. Sole exposure to proinflammatory cytokines, for example TNFα (Nawashiro et al, 1997), is sufficient to induce a state of ischemic tolerance. Interestingly, proinflammation, including Toll-like receptor and cytokine signaling initially, promotes a period of inflammation and later suppresses innate immunity. This is mediated by the activation of counter-regulatory mechanisms that result in an immune suppressed state, characterized by lower levels of released cytokines upon a second challenge with LPS or during subsequent ischemia (for a review see Kariko et al, 2004). The development of tolerance, unlike sensitization, requires de novo protein synthesis (Bordet et al, 2000), providing a possible explanation for the requirement of a time lag for tolerance to develop. Those data may explain how the manipulation of one experimental parameter, the temporal separation between LPS exposure and subsequent experimental ischemia, may distinguish sensitization from tolerance in our model.

Excitotoxicity in Lipopolysaccharide-Induced Sensitization

Glutamate release and glutamate toxicity is a main component of ischemic neuronal damage and is mediated by binding of glutamate to the ionotropic glutamate receptors, the NMDA, AMPA, and kainate receptors, and to the metabotropic glutamate receptors. The NMDA receptor antagonist MK-801 and the AMPA receptor antagonist CNQX were used to show the level of glutamate involvement in the ensuing neuronal cell death after OGD with or without LPS preexposure. In agreement with previous reports (Newell et al, 1995), we observed that cell death induced by pure OGD was inhibited by NMDA and AMPA receptor blockers. Blocking NMDA receptor function in cultures preexposed to LPS provided almost full protection to subsequent OGD in the CA1 region and in the dentate gyrus, whereas AMPA receptor blockers reduced cell death to levels that were still significantly higher than those induced by pure OGD. Thus, the severe increase in neuronal death in the CA1 region and in the dentate gyrus after LPS exposure before OGD was mediated by glutamate toxicity. In analogy with our results, Zou and Crews (2005) showed that TNFα, in combination with a submaximal dose of glutamate, resulted in accelerated and increased glutamate neurotoxicity in hippocampal slices. The ensuing excitotoxic neuronal cell death was blocked by NMDA receptor antagonists but not by AMPA receptor antagonists. Recent evidence shows that exposure to TNFα increases glutamate toxicity by increasing AMPA receptor density but not NMDA receptor density at synaptic terminals (Stellwagen et al, 2005). This suggests that higher relative concentrations of AMPA than NMDA receptor antagonists are needed to block AMPA receptor function after LPS and OGD. Blocking NMDA or AMPA receptor function had no effect on neuronal death in the CA3 region, suggesting that mechanisms other than excitotoxicity are involved in neuronal death in the CA3 region after LPS and OGD. Differential NMDA receptor density in hippocampal neuronal subregions might explain the finding that the response in neurons of the CA3 region to MK-801 was lower than in neurons of the CA1 region and the dentate gyrus.

The Roles of Tumor Necrosis Factor Receptors 1 and 2 in Lipopolysaccharide-Induced Sensitization and Tolerance

Tumor necrosis factor α is a potent proinflammatory cytokine, capable of inducing severe tissue damage and cell death, but might also induce antiapoptotic and trophic responses. Tumor necrosis factor α exerts its effect by binding to its receptors, TNF receptor 1 and 2, which are present on both neuronal cells and microglial cells. Both receptors activate, among others, the transcription factor NF-κB, which activates transcription of target genes that regulate cell survival and apoptosis. Nuclear factor-κB may induce apoptosis by signaling through Fas-associated death domain and caspase 8 (Muppidi et al, 2004), and may activate mechanisms mediating cell survival by increasing the expression of cell survival factors that provide protection against cerebral ischemia (Wilde et al, 2000). The protective and deleterious effects of TNFα stimulation have been speculated to be mediated by its two receptors (Marchetti et al, 2004).

We investigated the role of the receptors for TNFα, TNFR1, and TNFR2 by using genetically modified mice lacking either receptor subtype, or by the addition of antagonistic antibodies specific for either receptor subtype. We show that signaling through TNFR1 mediates the neurodestructive effects of TNFα in LPS-induced sensitization and the protective effects in LPS-induced ischemic tolerance. Mechanisms activated downstream of TNFR1 are thus essential for both LPS-induced sensitization and tolerance.

Secretion of TNFα, IL-6, and MCP-1 was severely inhibited during LPS exposure in slices devoid of TNFR1. A previous study has shown that signaling through TNFR1 promotes an autocrine loop, thereby amplifying TNFα secretion during LPS exposure or exposure to TNFα. (Kuno et al, 2005) This mechanism appears to be present in our model of hippocampal cultures and, importantly, is associated with subsequent sensitivity to in vitro ischemia. The lack of functional amplification of cytokine production in TNFR1^−/- slices is likely to be central to the absence of both sensitization and tolerance, as sufficient production of TNFα should be essential for the operation of both paradigms. It has been shown that preconditioning using endotoxin upregulates the expression of soluble TNFR1 resulting in decreased circulatory extracellular TNFα (Rosenzweig et al, 2007). The activation of TNFR1 has also been shown to be essential for the upregulation of glutamate transporters, a suggested protective effect in preconditioning resulting in decreased excitotoxic cell death (Persson et al, 2005).

The role of TNFR2 is less clear in LPS-induced sensitization and tolerance. We show that blocking TNFR2 by antagonistic antibodies resulted in increased neuronal damage after LPS and OGD, implying a protective role for TNFR2 in acute excitotoxicity. Indeed, Bernardino et al (2005) reported that TNFα exposure dose-dependently induced either neuroprotection or enhanced neurotoxicity when combined with AMPA in murine hippocampal slices through selective activation of either TNFR1 or TNFR2 (Bernardino et al, 2005). This indicated opposing roles for TNFR1 and TNFR2 in acute excitotoxicity, and a protective role for signaling through TNFR2. We showed that slices from TNFR2^−/- mice secreted significantly higher levels of TNFα and IL-6 during LPS exposure as compared with slices from WT mice. Similar findings have been observed in cultured microglia by blocking TNFR2 during LPS exposure (Kuno et al, 2005). Thus, signaling through TNFR2 appears to reduce LPS-induced proinflammatory cytokine response resulting in reduced neuronal cell death after LPS and OGD. However, the role of TNFR2 in central nervous system inflammation warrants further studies, and it cannot be excluded that protection mediated by the activation of TNFR2 involves the activation of signaling pathways involved in other aspects of neuroprotection.

Conclusions

Signaling through TNFR1 amplifies proinflammatory cytokine response during LPS exposure, whereas signaling through TNFR2 is associated with a decreased proinflammatory cytokine response. Signaling by TNFα through TNFR1 is essential for LPS-induced sensitization and protection to OGD. The organotypic hippocampal slice culture appears well suited for studies addressing the interaction between inflammation and ischemia, with results correlating well to those of existing in vivo models. We propose this in vitro model as a valuable tool for future studies focusing on neuroprotective agents in the context of inflammation and ischemia in the immature brain.

Footnotes

Acknowledgements

We thank Blanka Boberg for excellent technical support.

The authors state no conflict of interest.

References

Andersson

Perry

Gordon

(1991) The kinetics and morphological characteristics of the macrophage—microglial response to kainic acid-induced neuronal degeneration. Neuroscience 42:201–14.

Anslow

(1998) Birth asphyxia. Eur J Radiol 26:148–53.

Bernardino

Xapelli

Silva

Jakobsen

Poulsen

Oliveira

Vezzani

Malva

Zimmer

(2005) Modulator effects of interleukin-1beta and tumor necrosis factor-alpha on AMPA-induced excitotoxicity in mouse organotypic hippocampal slice cultures. J Neurosci 25:6734–44.

Bordet

Deplanque

Maboudou

Puisieux

Robin

Martin

Bastide

Leys

Lhermitte

Dupuis

(2000) Increase in endogenous brain superoxide dismutase as a potential mechanism of lipopolysaccharide-induced brain ischemic tolerance. J Cereb Blood Flow Metab 20:1190–6.

Cai

Pang

Lin

Rhodes

(2003) Differential roles of tumor necrosis factor-alpha and interleukin-1 beta in lipopolysaccharide-induced brain injury in the neonatal rat. Brain Res 975:37–47.

Coumans

Middelanis

Garnier

Vaihinger

Leib

Von Duering

Hasaart

Jensen

Berger

(2003) Intracisternal application of endotoxin enhances the susceptibility to subsequent hypoxic—ischemic brain damage in neonatal rats. Pediatr Res 53:770–5.

Cronberg

Rytter

Asztely

Soder

Wieloch

(2004) Glucose but not lactate in combination with acidosis aggravates ischemic neuronal death in vitro. Stroke 35:753–7.

Dammann

Leviton

(2000) Role of the fetus in perinatal infection and neonatal brain damage. Curr Opin Pediatr 12:99–104.

De Simoni

Griesinger

Edwards

(2003) Development of rat CA1 neurones in acute versus organotypic slices: role of experience in synaptic morphology and activity. J Physiol 550:135–47.

10.

Eklind

Mallard

Arvidsson

Hagberg

(2005) Lipopolysaccharide induces both a primary and a secondary phase of sensitization in the developing rat brain. Pediatr Res 58:112–6.

11.

Eklind

Mallard

Leverin

Gilland

Blomgren

Mattsby-Baltzer

Hagberg

(2001) Bacterial endotoxin sensitizes the immature brain to hypoxic—ischaemic injury. Eur J Neurosci 13:1101–6.

12.

Fernandez-Lopez

Martinez-Orgado

Casanova

Bonet

Leza

Lorenzo

Moro

Lizasoain

(2005) Immature rat brain slices exposed to oxygen—glucose deprivation as an in vitro model of neonatal hypoxic—ischemic encephalopathy. J Neurosci Methods 145:205–12.

13.

Hansen-Pupp

Harling

Berg

Cilio

Hellstrom-Westas

Ley

(2005) Circulating interferon-gamma and white matter brain damage in preterm infants. Pediatr Res 58:946–52.

14.

Hughes

Allegrini

Rudin

Perry

Mir

Wiessner

(2002) Monocyte chemoattractant protein-1 deficiency is protective in a murine stroke model. J Cereb Blood Flow Metab 22:308–17.

15.

Huuskonen

Suuronen

Miettinen

van Groen

Salminen

(2005) A refined in vitro model to study inflammatory responses in organotypic membrane culture of postnatal rat hippocampal slices. J Neuroinflammation 2:25.

16.

Ikeda

Yang

Ikenoue

Mallard

Hagberg

(2006) Endotoxin-induced hypoxic—ischemic tolerance is mediated by up-regulation of corticosterone in neonatal rat. Pediatr Res 59:56–60.

17.

Ivacko

Szaflarski

Malinak

Flory

Warren

Silverstein

(1997) Hypoxicischemic injury induces monocyte chemoattractant protein-1 expression in neonatal rat brain. J Cereb Blood Flow Metab 17:759–70.

18.

Kariko

Weissman

Welsh

(2004) Inhibition of toll-like receptor and cytokine signaling—a unifying theme in ischemic tolerance. J Cereb Blood Flow Metab 24:1288–304.

19.

Kuno

Wang

Kawanokuchi

Takeuchi

Mizuno

Suzumura

(2005) Autocrine activation of microglia by tumor necrosis factor-alpha. J Neuroimmunol 162:89–96.

20.

Mallard

Welin

Peebles

Hagberg

Kjellmer

(2003) White matter injury following systemic endotoxemia or asphyxia in the fetal sheep. Neurochem Res 28:215–23.

21.

Marchetti

Klein

Schlett

Pfizenmaier

Eisel

(2004) Tumor necrosis factor (TNF)-mediated neuroprotection against glutamate-induced excitotoxicity is enhanced by N-methyl-D-aspartate receptor activation. Essential role of a TNF receptor 2-mediated phosphatidylinositol 3-kinase-dependent NF-kappa B pathway. J Biol Chem 279:32869–81.

22.

Muller

Buchs

Stoppini

(1993) Time course of synaptic development in hippocampal organotypic cultures. Brain Res Dev Brain Res 71:93–100.

23.

Muppidi

Tschopp

Siegel

(2004) Life and death decisions: secondary complexes and lipid rafts in TNF receptor family signal transduction. Immunity 21:461–5.

24.

Nawashiro

Tasaki

Ruetzler

Hallenbeck

(1997) TNF-alpha pretreatment induces protective effects against focal cerebral ischemia in mice. J Cereb Blood Flow Metab 17:483–90.

25.

Newell

Barth

Papermaster

Malouf

(1995) Glutamate and non-glutamate receptor mediated toxicity caused by oxygen and glucose deprivation in organotypic hippocampal cultures. J Neurosci 15: 7702–7711.

26.

Ogunyemi

Murillo

Jackson

Hunter

Alperson

(2003) The relationship between placental histopathology findings and perinatal outcome in preterm infants. J Matern Fetal Neonatal Med 13:102–9.

27.

Pang

Cai

Rhodes

(2003) Disturbance of oligodendrocyte development, hypomyelination and white matter injury in the neonatal rat brain after intracerebral injection of lipopolysaccharide. Brain Res Dev Brain Res 140:205–14.

28.

Persson

Brantefjord

Hansson

Ronnback

(2005) Lipopolysaccharide increases microglial GLT-1 expression and glutamate uptake capacity in vitro by a mechanism dependent on TNF-alpha. Glia 51:111–20.

29.

Peschon

Torrance

Stocking

Glaccum

Otten

Willis

Charrier

Morrissey

Ware

Mohler

(1998) TNF receptor-deficient mice reveal divergent roles for p55 and p75 in several models of inflammation. J Immunol 160:943–52.

30.

Prass

Meisel

Hoflich

Braun

Halle

Wolf

Ruscher

Victorov IV Priller

Dirnagl

Volk

Meisel

(2003) Stroke-induced immunodeficiency promotes spontaneous bacterial infections and is mediated by sympathetic activation reversal by post-stroke T helper cell type 1-like immunostimulation. J Exp Med 198:725–36.

31.

Rosenzweig

Minami

Lessov

Coste

Stevens

Henshall

Meller

Simon

Stenzel-Poore

(2007) Endotoxin preconditioning protects against the cytotoxic effects of TNFαlpha after stroke: a novel role for TNFalpha in LPS-ischemic tolerance. J Cereb Blood Flow Metab 27:1663–74.

32.

Rybnikova

Mironova

Pivina

Tulkova

Ordyan

Nalivaeva

Turner

Samoilov

(2007) Involvement of the hypothalamic—pituitary—adrenal axis in the antidepressant-like effects of mild hypoxic preconditioning in rats. Psychoneuroendocrinology 32:813–23.

33.

Saito

Suyama

Nishida

Sei

Basile (1996) Early increases in TNF-alpha, IL-6 and IL-1 beta levels following transient cerebral ischemia in gerbil brain. Neurosci Lett 206:149–52.

34.

Stellwagen

Beattie

Seo

Malenka

(2005) Differential regulation of AMPA receptor and GABA receptor trafficking by tumor necrosis factor-alpha. J Neurosci 25:3219–28.

35.

Stoppini

Buchs

Muller

(1991) A simple method for organotypic cultures of nervous tissue. J Neurosci Methods 37:173–82.

36.

Viscardi

Muhumuza

Rodriguez

Fairchild

Sun

Gross

Campbell

Wilson

Hester

Hasday

(2004) Inflammatory markers in intrauterine and fetal blood and cerebrospinal fluid compartments are associated with adverse pulmonary and neurologic outcomes in preterm infants. Pediatr Res 55:1009–17.

37.

Vlasov

Korzhevskii

Polyakova

(2005) Ischemic preconditioning of the rat brain as a method of endothelial protection from ischemic/repercussion injury. Neurosci Behav Physiol 35:567–72.

38.

Wang

Rousset

Hagberg

Mallard

(2006) Lipopolysaccharide-induced inflammation and perinatal brain injury. Semin Fetal Neonatal Med 11: 343–53.

39.

Wilde

Pringle

Sundstrom

Mann

Iannotti

(2000) Attenuation and augmentation of ischaemia-related neuronal death by tumour necrosis factor-alpha in vitro. Eur J Neurosci 12:3863–70.

40.

Zou

Crews

(2005) TNF alpha potentiates glutamate neurotoxicity by inhibiting glutamate uptake in organotypic brain slice cultures: neuroprotection by NF kappa B inhibition. Brain Res 1034:11–24.