Focal ischemia was induced in outbred male Crl:CD1/ICR mice (25 to 35 g) using the filament method as previously described [11, 12]. All procedures were in accordance with the Swiss Federal Law on Animal Welfare and were approved by the Swiss Cantonal Veterinary Office. Briefly, the left middle cerebral artery was occluded for 30 min and the filament was withdrawn to allow reperfusion. Sham animals underwent the procedure without arterial occlusion. Cerebral blood flow and temperature were recorded. Randomly, a single dose of vehicle solution (0.85% sodium chloride) or D-JNKI1 (0.1 mg/kg, NeoMPS, obtained from Xigen SA, Epalinges, Switzerland) was injected intravenously 3 h after ischemia onset. Mice were killed at different time points after MCAO.
Plasma was obtained from cardiac blood samples (1/16 v/v of 4% sodium citrate) by centrifugation (15 min at 2500 rpm, 4°C). Mice were transcardially perfused with PBS. To evaluate the cytokines released by the tissue rather than those present intracellularly, we assayed the culture medium of brain and spleen samples after an overnight incubation [17, 18]. Coronal brain slices 2 mm thick and a segment of the spleen 5 mm thick were incubated in wells containing 1.5 mL of DMEM medium (Dulbecco’s Modified Eagle Medium 1X, 4.5 g/L glucose, L-glutamine, GIBCO, UK), 10% horse serum (Oxoid Ltd., Basingstoke, UK) supplemented with 11.0 mg/mL sodium pyruvate (100 mM, Sigma, USA) and 10 mL/L penicillin:streptomycin (Sigma, USA) for 20 h at 37°C in humidified air with 5% CO2. Control mice were used to distinguish between the release due to MCAO and that resulting from tissue preparation and culture. The medium was centrifuged (5 min at 12,000 rpm, room temperature). The supernatants were frozen at −80°C. Protein concentrations were determined by Bradford assay.
To assess cytokine levels in brain homogenates, the slice adjacent to the one tested for brain cytokine release was directly frozen in liquid nitrogen and conserved at −80°C for further analysis. Frozen sections were cut again into 20 μm slices using a cryostat and the ipsilateral and contralateral hemispheres were separated and collected. Slices were homogenized in buffer (20 mM tris(hydroxymethyl)aminomethane-acetate, pH 7.0; 0.27 M sucrose; 1 mM ethylenediaminetetraacetic acid; 1 mM ethyleneglycoltetraacetic acid; 50 mM sodium fluoride; 10 mM beta-glycerophosphate; 5 mM sodium pyrophosphate; 1 mM sodium vanadate; 1% Triton X) containing protease inhibitors. After centrifugation (15 min at 10,000 rpm, 4°C), supernatants were collected and protein concentrations calculated according to the Bradford method.
An ELISA was performed using half the quantities recommended by the manufacturer for the detection of mouse IL-6 (OptEIA TM Set, BD Biosciences, San Diego, CA) and KC/CXCL1 (DuoSet ELISA Development System, Abingdon, UK) in plasma, supernatants from incubated brains and spleens, and supernatants from brain homogenates. The optical density was measured at 450 nm.
Immunohistochemical labeling and double immunofluorescence of brain sections were performed according to previously published data  using the following antibodies: mouse anti-Neuronal Nuclei (NeuN, 1/500, Millipore), mouse anti-Glial Fibrillary Acidic Protein (GFAP, 1/500, Millipore), rat anti-CD11b (Mac-1, 1/100, AbD Serotec, UK), rabbit polyclonal anti-IL-6 antibody (1/200, ab6672, Abcam, Cambridge, UK), anti-growth-related protein alpha antibody against chemokine KC/CXCL1 (GRO-alpha, 1/100, ab86436, Abcam, Cambridge, UK). Images were acquired with the Aviovision v3.1 software using a Zeiss Axiovision microscope.
Results were expressed as mean ± standard error of the mean (SEM). Cytokine concentration data were square root-transformed to bring distributions closer to normal and reduce variance inequalities before statistical analysis. Comparisons between three or more groups were performed using one- or two-way analysis of variance. For post-hoc comparisons, a Brown-Forsythe test for equality of variances between groups was performed, followed by Tukey’s comparison of means when variances were equal or a Games-Howell test when variances were unequal. Paired statistical tests were used to compare cytokine concentration in the ipsilateral versus the contralateral hemisphere at each time point for both the MCAO+vehicle and the MCAO+D-JNKI1 mice (JMP 9.0.0 and IBM SPSS statistics 19.0.0). P <0.05 was considered significant.