Primary microglial cultures were prepared as previously described [3] (Kim et al., 2006). In brief, cells dissociated from the cerebral hemispheres of 1 day-old postnatal rat brains (Sprague-Dawley strain) were seeded at a density of 1.2×10⁶ cells/ml in Dulbecco's modified Eagle's medium (DMEM; Gibco, Carlsbad, CA, USA) containing 10% FBS (Hyclone, Logan, UT, USA) and 1% penicillin-streptomycin (Gibco, Carlsbad, CA, USA). Two weeks later, microglia were detached from flasks by mild shaking and filtered through a cell strainer (BD Falcon, Bedford, MA, USA) to remove astrocytes. After centrifugation (1,000×g) for 5 min, cells were resuspended in fresh DMEM containing 10% FBS and 1% penicillin-streptomycin and plated at a final density of 1.5×10⁵ cells/well on a 24 multiwell culture plate. After 2 hrs, the medium was changed for DMEM containing 5% FBS and 500 µM B27 supplement (Gibco, Carlsbad, CA, USA).

Primary microglial cells were treated with serum-free DMEM containing 100 ng/ml of LPS (Sigma, St. Louis, MO, USA) for 18 hrs. After washing with twice with DMEM, medium was replaced with fresh DMEM. LPS-conditioned media (LCM) was collected 24 hrs later and concentrated from 600 µl to 20 µl using a Centricon 10 (Millipore, Billerica, MA, USA). For control, primary microglial cells were treated with serum-free DMEM for 24 hrs.

Primary microglia cells (1.5×10⁵) were seeded in 24-well plates and 1 day later treated with LCM or LPS (100 ng/ml). To measure the amount of NO produced, 50 µl of conditioned medium was mixed with an equal volume of Griess reagent (0.5% sulfanilamide, 0.05% N-naphthylene-diamine-H-chloride, and 2.5% H₃PO₄) and incubated for 5 min at room temperature. Absorbances of mixtures were measured at 550 nm using a microplate reader. NaNO₂ standards were used to calculate NO₂^- concentrations.

Cells were seeded at a density of 1.5×10⁵ cells per well in 24-well plates. All transfections were performed using oligofectamine (Invitrogen, Carlsbad, CA, USA) as a carrier. SMART pool siRNA specifically targeting HMGB1 (ON-TARGET plus SMART pool siRNA L-114889, accession no.NM_ 001109373; Dharmacon, Lafayette, CO, USA) and non-specific siRNA (on-TARGET plus Non-targeting pool D-001810, Dharmacon, Lafayette, CO, USA) were used. Transient transfections were carried out according to the manufacturer's instruction. siRNA and lipid complexes were added to the wells of 96-well culture plates to a final concentration of 40 pM siRNA and 1.8 µl oligofectamine.

Microglial cells were treated with LCM or LPS (100 ng/ml) and total RNA was purified with TRI reagent (Sigma-Aldrich, St Louis, MO) according to the manufacturer's instructions. First-strand cDNA was synthesized using the Takara RNA PCR kit (Takara Bio, Otsu, Japan) in a total volume of 20 µl containing 1 µg of total RNA. Real-time PCR was performed in a final volume of 20 µl containing 10 µl of 2x SYBR Green supermix (Takara Bio, Otsu, Japan), 1 µl each of 5 pmol/µl forward and reverse primers, and 5 µl of cDNA (50 ng; 1/100 dilution) using the Mini-Opticon Real-Time PCR System Detector (Bio-Rad, Richmond, CA). PCR was performed using: 5 minutes at 95℃, followed by 40 cycles of 30 seconds at 95℃, 30 seconds at 57℃, and 30 seconds at 72℃. The specificity of amplification was determined by DNA melting curve analysis using the built-in software. Differences in amplification fold were calculated based on the real-time PCR amplification of the target gene versus GAPDH using the built-in Gene Expression Analysis software in an iCycler iQ Real-Time RCR Detection System (Bio-Rad, Richmond, CA). The following primers sets were used; 5'-TCATTGAC CTCAACTACATGGT-3' and 5'-CTAAGCAGTTGGTGGTGCAG-3' for GAPDH. Real-Time PCR was performed in quadruplicate.

To analyze the binding of HBHP to HMGB1, a pull-down assay was performed using a 20 µl of streptavidin agarose beads (50% slurry) (Pierce, Rockford, IL, USA). Pull-down complexes were separated by 12% SDS-PAGE and analyzed by Western blotting. Biotin-labeled HBHP (HMSKPVQ; 1 or 5 µg/ml) or scrambled HBHP (PMQSKHV; 5 µg/ml) was incubated with LCM (containing HMGB1) at 4℃ for 4 hrs with rotation. A competition assay was performed using 5 µg/ml of HBHP, which was preincubated with the A box domain (1, 5, or 10 µg/ml) or the B box domain (1, 5, or 10 µg/ml) of HMGB1 for 1 hr at 4℃ with rotation. The biotin complexes obtained were incubated with streptavidin beads for 1 hr at 4℃ with rotation, centrifuged at 8,000 rpm for 1 min, and analyzed by Western blotting.

Cells were washed twice with cold PBS and lysed with RIPA buffer (50 mM Tris-HCl (pH 7.4), 1% NP40, 0.25% sodium-deoxycholate, 150 mM NaCl) containing 1 complete Mini protease inhibitor cocktail tablet (Roche diagnostics, Basel, Switzerland). For protein preparations from the supernatant, media were collected and concentrated as previously described and cells were treated with RIPA buffer. The lysates obtained were incubated in ice for 15 min and centrifuged at 12,000 rpm for 15 min at 4℃ and supernatants were loaded onto 12% SDS-PAGE gels. The primary antibodies (diluted at 1:1,000) used were as follows: anti-HMGB1 (Abcam, Cambridge, UK) and anti-α-tubulin (Santa Cruz Biotechnology, Santa Cruz, CA, USA). Primary antibodies were detected using a chemiluminescence kit (Roche, Mannheim, Germany) using goat anti-rabbit or anti-mouse horseradish peroxidase-conjugated secondary antibody (1:2,000, Millipore, Billerica, MA, USA).

Statistical analysis was performed by analysis of variance (ANOVA) followed by the Newman-Keuls test. All data are presented as means±SEMs and statistical significance was accepted at the 5% level.

HMGB1 levels in primary microglia culture media were examined at after 1, 3, 6, 12, 18, or 24 hrs of LPS treatment (100 ng/ml). HMGB1 accumulation was detected after 3 hrs of LPS treatment and its levels gradually increased to peak at 18 hrs (Fig. 1A). On the other hand, HMGB1 levels in whole cell lysates decreased gradually from 3 hrs (Fig. 1B), whereas α-tubulin levels were similar at all time points (Fig. 1B).

To obtain LPS-free LCM, culture media were replaced with fresh medium after 18 hrs of LPS treatment and LCM was collected at 12 or 24 hrs later. Immunoblot analysis revealed HMGB1 accumulation in LPS-free LCM, and the level of HMGB1 at 24 hrs after medium replacement was slightly higher than that at 12 hrs (Fig. 1C). When fresh primary microglia cultures were incubated with LCM for 24 hrs, NO was induced (Fig. 1D) and more NO was induced when cells were treated with concentrated LCM (Fig. 1D). In following experiments, we used LCM collected 24 hrs after medium replacement that was concentrated two-fold.

To determine whether HMGB1 in LCM was responsible for microglial activation, the proinflammatory potency of LCM was examined after HMGB1 knockdown. HMGB1 levels in primary microglia decreased to 28.6±3.9% of the control level at 12 hrs after HMGB1-siRNA transfection (Fig. 2A). Furthermore, the LCM of HMGB1-siRNA transfected cells induced significantly less NO (59.9±0.9%) than LCM-treated control cells (Fig. 2B). However, NO levels were unchanged, when LCM from non-specific siRNA (NS)-transfected cells was used (Fig. 2B). These results suggest that HMGB1 plays an important role in LCM-induced microglial activation. LCM-induced NO production was also significantly suppressed by co-treating cells with HMGB1 A box (50 ng/ml) to 46.2±2.3% (n=4, p<0.01), a well-known antagonist of HMGB1 [20], or by co-treating anti-HMGB1 antibody (500 ng/ml) to 50.5±3.3% (n=4, p<0.01). These findings further support the critical role played by HMGB1 in LCM-induced microglia activation (Fig. 2C).

To determine whether HBHP affects LCM-induced microglia activation, NO levels were measured after co-treating cells with LCM and HBHP. NO levels were suppressed to 56.7±5.1%, 30.8±1.6%, and 32.5±10.8% of the control when HBHP was treated at 10, 50, or 100 ng/ml, respectively (Fig. 3A). Whereas scrambled HBHP (PMQSKHV; sc-HBHP, 50 or 100 ng/ml) did not produce such effects. In addition, proinflammatory cytokine productions in LCM-treated primary microglia cultures were also suppressed by HBHP treatment (10, 50, or 100 ng/ml), but not by scrambled HBHP (sc-HBHP, 50 or 100 ng/ml) (Fig. 3B), further supporting the notion that HBHP has anti-inflammatory effects in LCM-treated primary microglia cultures.

Next, we examined whether HBHP binds to the HMGB1 released by activated microglia to LCM. A biotin pull-down assay using biotin-labeled HBHP revealed the presence of binding between HBHP and HMGB1 in LCM in a HBHP-dose dependent manner, whereas, this binding was not detected when scrambled HBHP was used (Fig. 4A). Moreover, the formation of HBHP-HMGB1 was significantly suppressed when the LCM of HMGB1-ablated cells was used (Fig. 4B). When LCM was incubated with biotinylated HBHP (5 µg/ml) in the presence of HMGB1 A box (1, 5, or 10 µg/ml) or B box (5 or 10 µg/m), HMGB1/HBHP binding was clearly inhibited dose dependently by HMGB1 A box but not by HMGB1 B box (Fig. 4C). Accordingly, these results suggest that HBHP binds directly and specifically with HMGB1 A box in LCM and that the suppression of LCM-induced microglial activation by HBHP is due to this binding.