The chemiluminescent signal detected with a cooled CCD camera (Pierce, USA) was analyzed with ArrayVision 8.0 software (Imaging Research, USA). The sensitivity limit for each molecule was: CCL1 (0.8 pg/mL), CCL2 (0.8 pg/mL), CCL3 (3.1 pg/mL), CCL4 (0.8 pg/mL), CCL5 (0.4 pg/mL), CCL11 (0.5 pg/mL), CCL17 (0.4 pg/mL), CCL22 (0.2 pg/mL) and CXCL8 (0.2 pg/mL)
as provided by the manufacturer. For LMD-samples, all values below the limit of detection were assigned with the corresponding limit value. We strictly followed the manufacturer’s instructions and conducted selleck products the assay in a blinded manner. LMD and plasma samples (with exception of temporal profiles) were assayed twice and the mean value of both measurements was given. For LMD-cell Selleck AZD6244 samples the resulting
chemokine protein concentration was finally corrected by the total protein content and values are given as pg/mg. Plasma results were expressed as pg/mL. Whole analysis was performed with SPSS 15.0 software (SPSS Inc., USA). Shapiro–Wilk test was used to define normally distributed variables (p > 0.05), due to small sample sizes. Normal distribution was analyzed by Students’ t test or ANOVA and mean and SD values were given. Different time points of temporal profiles were compared by ANOVA of repeated measures and paired-t test, while correlations with other continuous variables Quinapyramine were assessed by Pearson test. Non-normal distribution was assessed by Mann–Whitney U or Kruskal–Wallis
tests and median and interquartile range (IQR) were reported. We compared temporal profiles by Friedman and Wilcoxon tests, and analyzed correlations by Spearman test. Pearson chi-squared test was used to compare categorical variables. In all cases, a p-value <0.05 was considered statistically significant at a 95% confidence level. For sample size and statistical power calculation we compared medians by using Ene 3.0 free software (GlaxoSmithKline S.A., Spain; http://sct.uab.cat/estadistica/es). Of the nine chemokines assayed, CCL3, CCL4 and CCL17 were not detected in LMD-cell samples. Among the remaining six chemokines, CCL1 and CCL2 were found at higher levels in neurons than in blood vessels (p = 0.021 in both cases) only in healthy contralateral area. Interestingly, CCL5 and CCL22 were decreased within the vessels and neurons, respectively, when the contralateral region of the brain was compared to the infarcted tissue (both cases with a p = 0.043) ( Fig. 1). All the nine chemokines were detected in plasma samples of ischemic stroke patients and, as shown in Supplementary Table 2, no differences regarding demographic and clinical data were found between both studied cohorts.