The formyl peptide receptor like-1 and scavenger receptor MARCO are involved in glial cell activation in bacterial meningitis

For the production of bacterial culture supernatants, bacteria were grown in stationary broth cultures for 20 h at 37°C until they had reached an optical density of 1.0. One ml of this culture was added to 9 ml of fresh medium and this was incubated for further 24 h in 75 cm² flasks without shaking. Brain heart infusion broth was used for the cultivation of Streptococcus pneumoniae, and thioglycolate broth supplemented with K1 and hemin was used for Neisseria meningitidis. Broth cultures of the test bacteria were centrifuged at 5000×g for 15 min, and the resulting supernatants were filter-sterilized. The supernatants were diluted 1:3 in EpiLife medium (Sigma, Germany) and used for stimulation experiments. Growth medium (diluted 1:100 or 1:200) served as the negative control to evaluate unspecific effects. Cells were challenged with supernatants from Neisseria meningitidis (ATCC 13077, type strain isolated from a fatal case of meningitis; 1:100) or Streptococcus pneumoniae (ATCC 6303; capsula type 3; 1:100) in serum and antibiotic-free DMEM.

FPRL1 antagonist WRW4 was purchased from Dr. P. Henklein (Charité, Berlin, Germany). The peptide was dissolved at 10 mM concentration in DMSO. Pertussin toxin (PTX) and formyl-methionyl-leucyl-proline (fMLF) were obtained from Sigma Chemical Company, Germany. pERK and ERK2 antibodies were obtained from Santa Cruz Biotechnology (CA, USA).

Isolated cerebral cortices and rostral mesencephali from 2-day old rats were stripped of the meninges, minced and dissociated enzymatically with trypsin from bovine pancreas (Sigma) in phosphate-buffered saline and 50 μg/mL DNase I (Roche Molecular Biochemicals, Indianapolis, IN, USA) for 30 min at 37°C and crushed mechanically with Pasteur pipettes. Astrocytes were prepared following the McCarthy and DeVellis method [17], which allows for the preparation of nearly pure cultures of astrocytes (> 97%) and cultivated in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (FCS). Suspended microglial cells were plated in 75 cm² cell culture flasks in microglial cell growth medium and harvested as described [18]. Prior to replating microglial cells for different assays, cell number and viability was estimated by trypan blue exclusion. This procedure increased the purity of the microglial preparation to > 98% with few astrocytes remaining. To test cell purity, cultures were stained with specific cell markers for astrocytes (glial fibrillary acidic protein (GFAP); astrocytes marker; Sigma) and microglia (OX42; microglia/macrophages marker; Sera-Lab, Leicestershire, UK). The production of primary rat meningeal cells was described previously [7].

HEK293 cells (American Type Culture Collection, Rockville, MD, USA) were subcultivated in DMEM supplemented with 10% FCS and 1% penicillin/streptomycin (Carl Roth, Karlsruhe, Germany). The transfection and selection of HEK293 cells expressing hMARCO, hFPRL1 or coexpressing hFPRL1-hMARCO was described previously [16].

The astrocytes transfection was described previously [16]. Briefly, one day before transfection, 3 × 10⁵/well astrocytes were seeded in DMEM containing 10% FCS in 6-well-plates and transfected with Primefect^® (Lonza, Riverside, USA) transfecting agent containing control siRNA (25 nM) and siRNA for target proteins (25 nM), respectively according to the manufacturer's recommendation. Small interfering RNA (siRNA) duplexes corresponding to rat FPRL1 and MARCO cDNA sequences (GenBank accession number XM_001057934, XM_218012 and XM_001054109) or control siRNA were purchased from Qiagen, Valencia, CA, USA. Cells were cultured for an additional 96 h and analyzed for mRNA expression via SYBR green real time PCR.

Total RNA was isolated using the peqGold Trifast reagent (peqlab, Erlangen, Germany) according to the manufacturer's protocol. RNA samples were reverse-transcribed by moloney murine leukemia virus (MMLV) reverse transcriptase (Fermentas, Burlington, Canada) and random hexamer primers (Invitrogen, USA). The cDNA products were used immediately for SYBR green real-time RT-PCR for Camp (rat), FPRL1 as well as MARCO and TaqMan real-time RT-PCR was used for CAMP (gene name for human cathelicidin) and IL-1β. Gene expression was monitored using the StepOne Plus apparatus (Applied Biosystems, USA) according to standard protocol [19]. Relative quantification was performed using the ΔCt method which results in ratios between target genes and a housekeeping reference genes (GAPDH or 18 s). The primer for Camp, FPRL1 and MARCO were manufactured by Qiagen (USA; QuantiTect Primer Assay). The specificity of the amplification reaction was determined with a melting curve analysis. We performed relative quantification of the signals normalizing the various gene signals with glycerinaldehyde-3-phosphate dehydrogenase (GAPDH) signal (forward primer: TCTACCCACGGCAAGTTCAAC; reverse primer: TCTCGCTCCTGGAAGATGGT; Eurofins MWG Operon, Germany) for SYBR Green real time RT-PCR, and with 18 s for TaqMan real time PCR.

For western blot analysis of MAP kinases phosphorylation, glial or HEK293 cells were seeded in DMEM containing 10% FCS. Cells were harvested in a lysis buffer (50 mM Tris pH 7.5, 100 mM NaCl, 5 mM EDTA, 1% Triton, 2 mM sodium orthovanadate, 2.5 mM sodium pyrophosphate, 1 mM glycerol 2-phosphate, 1 mM phenylmethylsulfonylfluoride). Proteins (5 μg for pERK and ERK2) were resolved in SDS sample buffer, and a western blotting procedure was performed as previously described in detail [20]. Membranes were incubated with polyclonal primary antibodies against pERK1/2 overnight at 4°C and subsequent detection was performed with peroxidase-labeled secondary antibodies. Antibody binding was detected via enhanced chemiluminescence (Amersham Pharmacia Biotech, Essex, UK). The membranes were then stripped and re-probed with anti-ERK2 antibody as a loading control. The western blot bands were densitometrically evaluated with the program PC-BAS, the pERK-bands were adjusted with their respective ERK-bands and subsequently, the values were referred to control (= 100%).

1.5 · 10⁵ astrocytes/well, or transfected HEK293 cells or 5 · 10⁵ microglia/well were seeded in 22-mm 12-well dishes with DMEM containing 10% FCS and incubated overnight. The medium was removed and replaced by 0.5 ml of serum-free DMEM medium containing 10 μM forskolin (for astrocytes; Sigma) or 25 μM forskolin (for microglia or HEK293 cells) plus agonists. Different forskolin concentrations were used because of different cell sensitivities to forskolin-stimulated adenylate cyclase activity. The cells were incubated at 37°C for 15 min, and the reaction was terminated by removal of the culture medium and addition of 1 ml of ice-cold HCl/ethanol (1 N; 1 : 100, v/v). After centrifugation the supernatant was evaporated, residues were dissolved in Tris-EDTA (TE) buffer (50 mM Tris-EDTA, pH 7.5), and cAMP content was determined using a commercially available radioimmunoassay kit (Amersham Pharmacia Biotech).

An established model of pneumococcal meningitis in infant rats was used [21]. The animal studies were approved by the Animal Care and Experimentation Committee of the Canton of Bern, Switzerland. Wistar rats were infected on postnatal day 11 by direct intracisternal injection of 10 μl saline solution containing a defined inoculum of Streptococcus pneumoniae (serogroup 3) or Neisseria meningitidis with a 32-gauge needle in the cisterna magna. The animals were sacrificed at 12, 22 and 39 h for SP respectively 24 h for NM after infection (n = 3/time intervall). Uninfected control animals were injected with 10 μl of a sterile saline solution. To document existence of bacterial meningitis, CSF was obtained by puncturing the cisterna magna and cultured quantitatively. Animals were sacrificed, and perfused with 4% paraformaldehyde in PBS via the left cardiac ventricle for immunohistological studies.

Coronary brain sections (10 μm) from rats were cut on a cryostat, collected on Superfrost Slides (Merck, Braunschweig, Germany), air dried for 24 h and stored at −80°C. For immunofluorescence staining, sections were fixed with acetone (5 min), permeabilized with 0.1% Trition × in PBS for 10 min at room temperature, incubated with rabbit anti-MARCO (1:100; sc-68913, Santa Cruz, USA) and mouse anti-GFAP (astrocytes marker; ab10062, Abcam, Cambridge, UK) overnight at 4°C. Finally, the slices were incubated with donkey anti-rabbit AlexaFluor 488 (Molecular Probes, USA) or goat anti mouse Cy3 (Sigma, Germany) for 1 h at room temperature. Primary rat glial or different HEK293 cells were grown on glass coverslips. Cells were then exposed to NM or SP for 12 or 24 h at 37°C. After fixation with ice cold methanol and acetone for 5 min and 20 seconds, cells were blocked in 0.1 M Tris-HCl pH 7.5 containing 1,5% bovine serum albumin (BSA; Sigma Chemical Company, Germany) for 10 min. Coverlips were incubated at 4°C overnight with primary antibodies diluted in TRIS containing 1,5% BSA. The following primary antibodies were used for this study: rabbit anti-MARCO antibody (1:100; Santa Cruz, USA), rabbit anti-FPRL1 antibody (1:100; sc-18191, Santa Cruz, USA), goat anti-rCRAMP antibody (1:100; sc-34170, Santa Cruz, USA), and rabbit anti-LL-37 antibody (1:100; ab64892, Abcam, Cambridge, UK). Finally, the coverlips were incubated with donkey anti-rabbit AlexaFluor 488 or donkey anti-goat AlexaFluor 488 (all 1:250; Molecular Probes, USA) for 1 h at room temperature. Nuclear staining was performed with bisbenzimide (Sigma, Germany). Cells were digitally photographed using a Zeiss Axio Z1 Imager microscope (Zeiss, Göttingen, Germany).

All experiments were performed at least in triplicate and the values represent mean ± SEM. The significance of the difference between test and control groups was analyzed using the ANOVA test, followed by the Bonferroni test. Data from the real-time RT-PCR, from densitometric quantification of western Blots, and from cAMP assay were analyzed using GraphPad Prism 5.0 software.

In a first set of experiments we investigated the expression of both the FPRL1 and MARCO receptor by primary astrocytes, microglia and meningeal cells after stimulation with either NM or SP bacterial supernatants for 0, 6, 12, 24 h by real-time RT-PCR technique. As shown in Figure 1A and 1C, NM or SP does not induce a significant increase of FPRL1 expression in astrocytes as well as microglia. Only in meningeal cells, SP induced a significant increase of FPRL1 expression after 12 h of treatment (4 ± 0.5-fold increase in expression; Figure 1E). In astrocytes, a significant increase of MARCO expression by NM after 12 h of treatment was detected (5 ± 0.8-fold increase in expression; Figure 1B). For SP treatment, we could not detect an increase of MARCO expression. In microglia no significant changes of receptor expressions were observed after NM as well as SP treatment (Figure 1D). In meningeal cells, maximum MARCO expression was reached by NM after 24 h of treatment and by SP after 6 h of treatment (9 ± 0.6- and 6.1 ± 1.3-fold increase in expression, respectively; Figure 1F).

We then examined the FPRL1 and MARCO expression in primary rat astrocytes and microglia cells after treatment with NM and SP bacterial supernatants for 24 h using fluorescence microscopy. For FPRL1, both primary cells showed a clearly protein expression, but NM as well as SP could not induce an increased receptor expression (Figure 2A). For MARCO, we could detect an increase of protein expression in astrocytes after NM treatment, whereas SP induced no change of receptor expression in astrocytes as well as microglia (Figure 2B).

In order to validate our previous in vitro findings in vivo, we used double fluorescence microscopy with receptor specific antibodies to examine the localization of MARCO and GFAP in astrocytes from infant rats with bacterial meningitis with NM or SP. Control rat brains showed a very low positive GFAP staining, since GFAP is a marker for activated astrocytes [22]. In correspondence with real-time RT-PCR and immunofluorescence results, MARCO expression shows a strong increase after 24 h of infection with NM as seen in Figure 3A, with visible co-localization of MARCO and GFAP in the sub-cortical area; Figure 3B. In addition the meninges showed an increase of MARCO expression after NM infection. Positive MARCO staining and MARCO/GFAP co-localization were also shown after 12, 22, 39 h of infection with SP, albeit weaker than in meningitis by NM (Figure 3A). After SP infection an increase of MARCO expression in meninges was also detected. Interestingly we could show an increase of MARCO expression after treatment with NM and SP in meningeal cells. For FPRL1 however, only SP induced increased expression (see Figure 1E and 1F).

Our previous results show a physical and functional interaction between FPRL1 and MARCO [16]. Furthermore we show an NM/SP-induced Camp (rat) expression in glial cells with a maximum at 24 h for astrocytes respectively 6 h for microglia [6]. Therefore, we investigated the possible involvement of FPRL1 in NM/SP-induced Camp expression in glial cells using real-time RT-PCR. As shown in Figure 4A, astrocytes were incubated with NM or SP with or without the FPRL1 antagonist WRW4 or the G-protein inhibitor pertussis toxin (PTX) for 24 h and Camp expression was determined. For microglia (Figure 4B), the cells were stimulated for 6 h. The results show that WRW4 and PTX strongly decreased NM-induced Camp expression (30.0 ± 8.8-fold increase) in astrocytes. For SP treatment, we could not detect a change of Camp expression after WRW4 or PTX co-stimulation, but also SP alone did not significantly induce Camp expression. In microglia, NM-induced Camp expression (11.7 ± 2.9-fold increase) was inhibited by WRW4 or PTX (Figure 4B). For SP treatment, WRW4 as well as PTX also decreased Camp expression.

To confirm the real-time RT-PCR results, we examined rCRAMP protein expression in primary rat astrocytes and microglia cells after treatment with NM and SP with different WRW4 concentrations using fluorescence microscopy. For NM, bacterial supernatants induced rCRAMP protein expression in glial cells after 24 h of treatment in astrocytes and 12 h of treatment in microglia (Figure 4C). NM-induced rCRAMP expression was strongly decreased by 5 and 10 μM WRW4 in astrocytes as well as microglia. Also, SP induced rCRAMP expression in astrocytes and microglia after 24 and 12 h of treatment, respectively (Figure 4D). In astrocytes, only 10 μM WRW4 showed an inhibition of SP-induced rCRAMP expression, whereas in microglia, 5 and 10 μM WRW4 decreased SP-induced rCRAMP expression. WRW4 alone could not induce rCRAMP expression in astrocytes or in microglia.

To investigate the effect of FPRL1 on NM- as well as SP-induced glial cell activation we incubated primary rat astrocytes and microglia with NM or SP to determine ERK1/2 phosphorylation and changes of cAMP accumulation. As shown in Figure 5A to 5D, treatment with NM as well as SP resulted in increased phosphorylation of ERK1/2 in both astrocytes (NM: 1.9 ± 0.2; SP: 1.6 ± 0.1) and microglia (NM: 2.0 ± 0.1; SP: 2.9 ± 0.1). Therefore, as a next step, we tested the influence of NM as well as SP on changes in cAMP levels in glial cells. To determine whether NM and SP affected cAMP formation via G-protein receptor activity, cAMP production in glial cells was induced by forskolin treatment (8.7 ± 1.6 pmol cAMP). Interestingly, NM (2.9 ± 0.3 pmol cAMP) as well as SP (3.0 ± 0.6 pmol cAMP) induced a decrease in forskolin-induced cAMP accumulation in glial cells (Figure 5E and 5F). Apparently their activity is stimulated by an inhibiting G-protein. The FPRL1 agonist fMLF (2.5 ± 0.9 pmol cAMP) was used as a positive control. Next, we tested whether SP- or NM-induced ERK1/2 phosphorylation and changes in cAMP levels are mediated by inhibiting G-protein coupled receptor FPRL1. We used the FPRL1 antagonist WRW4 and pertussis toxin (PTX) as an inhibitor of inhibitory G-proteins. We first investigated the inhibition of ERK1/2 phosphorylation by the antagonists. As shown in Figure 5A and 5C, preincubation with WRW4 and PTX inhibited NM- as well as SP-induced ERK1/2 phosphorylation in astrocytes. In microglia, NM-induced ERK1/2 phosphorylation was decreased using WRW4 or PTX (Figure 5B and 5D). For SP, the induced ERK1/2 phosphorylation was inhibited by PTX, whereas WRW4 co-stimulation did not show a significant decrease. No changes could be detected in ERK1/2 phosphorylation with WRW4 and PTX alone. In addition to the assay mentioned above, the effects of antagonists on NM- or SP-induced cAMP formation was investigated. As shown in Figure 5E and 5F, fMLF and NM-mediated decreases in cAMP levels were inhibited by WRW4 in both astrocytes and microglia. For SP, in astrocytes as well as microglia, WRW4 had no effect on an SP-induced decrease of cAMP level. WRW4 alone did not alter forskolin-stimulated adenylate cyclase activity in glial cells.

To determine whether FPRL1 as well as MARCO are essential for NM- as well as SP-induced Camp as well as cytokine expression and ERK1/2 phosphorylation, we prepared siRNA targeting rat FPRL1 and MARCO in astrocytes. Transfection of astrocytes with FPRL1 and MARCO siRNA, but not with control siRNA, resulted in a significant reduction in FPRL1 and MARCO mRNA levels 96 h after transfection as determined by SYBR green real time RT-PCR (Figure 6A and 6B). The transfection of primary rat microglia did not reach sufficient efficiency (data not shown). To investigate the effect of FPRL1 and MARCO for NM- as well as SP-induced Camp or IL-1β expression, astrocytes were transfected with FPRL1 or MARCO siRNA as described above. Ninety-six hours after transfection, Camp and IL-1β expression was determined using real-time RT-PCR. As shown in Figure 6C, treatment with NM as well as SP for 24 h resulted in a strong increase of Camp expression in astrocytes that had been transiently transfected with control siRNA and also in untransfected control cells. Transfection with FPRL1 but also MARCO siRNA significantly inhibited NM- as well as SP-induced Camp expression in astrocytes. Also the NM- and SP-induced proinflammatory cytokine IL-1β expression was significantly reduced by FPRL1 as well as MARCO siRNA inhibition (Figure 6E). The receptor activity was further determined by ERK1/2 phosphorylation. As shown in Figure 6D and 6F, treatment with NM and SP resulted in an intense phosphorylation of ERK1/2 in astrocytes that had been transiently transfected with control siRNA and also in untransfected control cells (about two- to threefold). Interestingly, transfection with FPRL1 as well as MARCO siRNA completely inhibited ERK1/2 phosphorylation induced by NM and SP.

In an effort to further evaluate the effect of FPRL1 and MARCO in NM/SP- and fMLF-induced signal transduction we used transfected HEK293 cells. Either hMARCO- or hFPRL1-expressing, or hFPRL1 and hMARCO co-expressing cells were generated and ERK 1/2 phosphorylation was analyzed after treatment with NM/SP and fMLF for 5 and 15 minutes. The results of the western blot were confirmed by densitometric quantification. In untransfected cells no treatment led to an increase in ERK 1/2 phosphorylation (Figure 7A and 7E). In hMARCO-transfected cells, a 5-minute treatment with NM (6.5 ± 1.5 fold increase) and SP (11.3 ± 0.3 fold increase) induced a significant increase in ERK1/2 phosphorylation (Figure 7B and 7F). As shown in Figure 7C and 7G, hFRPL1-transfected cells also showed this increase in phosphorylation, but here the treatment with the FPRL1 agonist fMLF (6.6 ± 0.1 fold increase) also led to increased phosphorylation after 5 minutes. Finally, in hFPRL1 and hMARCO co-expressing cells a 5-minute stimulation with all three substances significantly raised the level of ERK phosphorylation (fMLF: 17.1 ± 1.1; NM: 19.7 ± 2; SP: 12 ± 1.8; Figure 7D and 7H). It should also be mentioned that the increased phosphorylation after fMLF and NM treatment in the co-expressing cells was also significantly higher than that of cells transfected with only one of the receptors. The above results are reflected in a change in forskolin-induced adenylate cyclase activity. In untransfected, or just hMARCO-transfected, cells no change in cAMP concentration was detected after treatment with fMLF and NM (Figure 7I and 7J). In hFPRL1-transfected cells only the stimulation with fMLF was able to reduce the cAMP level (2.6 ± 0.6 pmol; Figure 7K). Solely the FPRL1 and MARCO co-expressing cells showed a decrease in adenylate cyclase activity after treatment with either fMLF (4.3 ± 0.2 pmol) or NM (2.6 ± 0.9 pmol; Figure 7L).

In an additional set of experiments we investigated the effect of fMLF, NM and SP treatment for the human cathelicidin CAMP (LL-37) in transfected HEK293 cells mentioned above. After 0, 6, 12 or 24 h treatment with fMLF, NM and SP CAMP expression as analyzed by using real-time RT-PCR. As shown in Figure 8A to 8C, in untransfected and cells transfected only, with either hFPRL1 or hMARCO, stimulation with either substance did not change the expression of CAMP mRNA. Only in hFPRL1 and hMARCO co-transfected cells (Figure 8D) did the treatment with fMLF show a maximum increase of CAMP mRNA expression after 6 h (4-fold increase). NM induced a maximum induction of CAMP expression at 24 h of treatment (3.9-fold increase). SP did not alter CAMP expression in hFPRL1-hMARCO-expressing HEK293 cells.

To confirm the real-time RT-PCR results, we examined LL-37 protein expression in transfected HEK293 cells after treatment with NM, SP and fMLF for 24 h using fluorescence microscopy. As shown in Figure 9 in untransfected and in HEK293 cells transfected only with hMARCO, stimulation with either substance did not change the expression of LL-37 protein. In hFPRL1-transfected cells, fMLF induced a strong, and NM a weak, LL-37 protein expression, whereas in hFPRL1 and hMARCO co-transfected cells treatment with NM and fMLF showed a clear increase of LL-37 expression.