Page 120 - scheppingen

five
118
afterwards. Flow cytometric analysis was performed using a FACSCanto Flow Cytometer together with FACSDiva software (BD Biosciences, San Diego, CA, USA). Data analysis was performed with FlowJo 7.6 software (TreeStar, Ashland, OR, USA). Levels of IL-6 were measured in culture supernatants using the PeliKine CompactTM IL-6 ELISA kit (Sanquin, Amsterdam, the Netherlands) according to the manufacturer’s instructions.
Statistical analysis
Statistical analyses were performed with GraphPad Prism® software (GraphPad soft- ware Inc., La Jolla, CA, USA) using non-parametric statistical testing. Mann-Whitney U test was used for analyzing cell culture data. To assess differences in RNA expression between multiple groups in human samples, non-parametric Kruskal–Wallis testing with correction for multiple comparisons (Dunn’s method) was used. Correlations were assessed using the Spearman’s rank correlation test. P<0.05 was assumed to indicate a significant difference.
Results
miR21, miR146a and miR155 expression by real time qPCR in TSC
Expression of miR21, miR146a and miR155 was studied using qPCR in control human cere- bral cortex and in cortical tubers and perituberal cortex from TSC patients. Expression of all three miRNAs was increased in cortical tubers compared to control cortex (Fig. 1A-C, pairwise comparison following Kruskal-Wallis, miR21: P<0.01, miR146a: P<0.05, miR155: P<0.05); expression in perituberal cortex (n=4) was not different compared to con- trols or tubers. As previously shown, there were no differences in expression between autopsy and surgical control samples 33, 35. A positive correlation was observed between all three miRNAs and GFAP IRS within the tuber (miR21: r=0.569, P=0.001; miR146a: r=0.47, P=0.008; miR155: r=0.47, P=0.007). Expression of IL-1β was increased in cortical tubers compared to control cortex (Supp. Fig. 1, P<0.001) and was positively associated with the expression of miR21 and miR155 (miR21: r=0.668, P<0.001; miR155: r=0.371, P=0.043).
No association was found between miR21, miR146a and miR155 expression and clinical features, including gender, age at epilepsy onset, age at surgery, pre-operative seizure frequency or duration of epilepsy.
miR21, miR146a and miR155 expression by in situ hybridization in TSC
The cellular distribution of miR21, miR146a and miR155 in control, perituberal cortex,
Figure 2 In situ hybridization of miR21 in control, perituberal cortex, TSC cortical tuber and SEGA. A-B: miR21, control grey (A) and white matter (B); miR21 was expressed at low levels in neurons (arrows in A) and was undetectable in glial cells. C: perituberal cortex (peri-ctx), showing mod- erate miR21 expression in neurons (arrows; insert in D) and glial cells (arrowheads). D-I (cortical tuber); miR21 was expressed in neurons (arrows, D-E), glial cells (arrowheads, D-E) and giant cells (asterisk in E). F-I (double labeling); F: expression of miR21 in dysmorphic neurons (NeuN posi- tive, red). G: expression of miR21 in GFAP positive cells (red). H: absence of co-localization with HLA-DR (microglia, red). I: expression of miR21 in pS6 positive cells (red). J-M: miR21 expression in subependymal giant cell astrocytoma (SEGA); K-M (double labeling): expression of miR21 in GFAP positive tumor cells (red; K); absence of colocalization with HLA-DR (microglia, red; L); expression of miR21 in pS6 positive cells (red, M). Scale bar in M: A-D: 60 μm. E-M: 40 μm.