Identification of miRNAs that affect p16^INK4A expression during oncogene-induced senescence.

<p>(a) Graphical representation of the workflow. The screen was performed in 384 well-format in TIG3 ΔBraf:ER cells with a library containing 471 human pre-miRs. 48 hours after transfection the cells were treated with 4-hydroxy-tamoxifen (4-OHT) for 48 hours and p16^INK4A levels were determined by immunofluorescense (IF) and analyzed automatically. (b) Screen results. The percentage of p16^INK4A positive cells was determined for each well using the Hoechst signal to determine the total cell number. Following automated image analysis, the Z-score was calculated for each miRNA based on the average percentage of p16^INK4A positive cells per plate. All listed miRNAs are considered potential hits; the ones indicated in bold were chosen for further validation. BMI1i, p16i and JMJD3i represent controls in which BMI1, p16^INK4A or JMJD3 were downregulated by siRNAs. SCR: scrambled control. (-) indicates scrambled control cells in which senescence was not induced. (c) Immunofluorescense images taken as part of the screening process. Controls and selected hits are shown. (d) Fold change relative to the plate average of all hits increasing the percentage of p16^INK4A positive cells in the screen (red line indicates the overall average). (e) Fold change relative to the plate average of all hits decreasing the percentage of p16^INK4A positive cells in the screen (red line indicates the overall average). (f) The average of p16^INK4A induction in 5 independent experiments. IF for p16^INK4A followed by automated image analysis was used as a readout and scrambled (SCR) control samples were used for normalization. Averages are shown with SEM and t-tests were used to evaluate differences. Asterisks indicate a significant difference (p<0.001) in p16^INK4A induction.