Fig. 2: Mitchondiral effects of PPARG lead to identification of AKT3 as a novel PPARG effector gene.

A, B Seahorse Bioscience mito stress test assay KO clones & Scrm control, graphs represent three independent experiments and four technical replicates per experiment, error bars show SEM. Statistical significance denoted by * where p ≤ 0.05 determined by 2-way Anova and Dunnetts multiple comparison test. C, D Seahorse Bioscience mito stress test assay OE clones & EV7 control, graphs represent three independent experiments and four technical replicates per experiment, error bars show SEM. Statistical significance denoted by * where p ≤ 0.05 by 2-way Anova and Dunnetts multiple comparison test. E IHC representative images for VDAC1 staining in OE & EV7 derived tumours, scale bar represents 100 μm. F Analysis of IHC images for VDAC1. Graph displays average percentage of positively stained cells over at least three samples with error bars giving SEM. Statistical significance, where found, denoted by * with p ≤ 0.05 determined by Mann–Whitney. G Carbon 13 labelled glucose derived ATP levels from EV7 and OE19 clones grown in 3D, unlabelled not shown. Represents three independent experiments each with three technical replicates. Bars represent average of these experiments with error bars showing SEM. Statistical significance is denoted by * where p ≤ 0.05 determined by 2 way Anova and Sidak’s multiple comparisons test where +5 ATP was highlighted as the significantly different isotopologue. H Subset of top 10 significant hits from RNA-seq analysis performed on tumour samples derived from PC3-M Scrm control, KO2, and DU-145 EV7 and OE19 clones. Three tumour samples from each cell type were used. Scrm vs. KO2 compared to EV7 vs. OE19 to identify hits that were oppositely affected in each comparison. All hits are statistically significant with p ≤ 0.05 and are ordered in terms of statistical significance, further details of analysis are given in the material and methods section. I RNA-scope in situ hybridisation representative images for AKT3 probing in OE & EV7 derived tumours, scale bar represents 50 μm. J Halo software analysis of RNA-scope for AKT3 on tumours derived from OE cell lines. Analysis performed on at least three prostate samples for each cell type used, graph displays average with error bars giving SEM. Statistical significance denoted with * where p ≤ 0.05 determined by Anova and Dunnett’s multiple comparisons test. K Immunoblot analysis from EV7 and OE19 derived tumour samples, three tumours per cell type. Values above each band indicate the densitometry value of the band as normalised to the loading control HSC70 and compared to the first EV7 band. Probed for PGC1a, CRM1, AKT3 and VDAC1, image representative of three independent experiments. L Immunoblot for the ETC complexes, complex I is labelled in blue and complex IV in red to allow the reader to identify the correct densitometry value for the corresponding bands, image representative of three independent experiments. Values above each band indicate the densitometry value of the band as normalised to the loading control HSC70 and compared to the first EV7 band.