Supplemental Figure 1. Immune-histochemical staining of PHF8 protein in normal, benign prostate and malignant PrCa. A) and B) are representative samples for benign prostate epithelial cells with absent or weak intensity of PHF8 staining. C) and D) show moderately intense staining patterns, typical for the majority of low grade Gleason score lesions. D) and E) are examples for moderate to strong staining patterns, as frequently found in high grade Gleason tumours.
Supplemental Figure 2. Heterogeneity of Immune-histological staining for PHF8 in malignant PrCa samples. A) Hyperplastic, benign lesion with negative staining for PHF8. B) Low grade Gleason 3 tumour with very weak PHF8 expression. C) Same tumour as in C, showing heterogeneous staining of malignant areas with different (advanced) histological morphology. D) PrCa metastatic lesion with strong staining of PHF8; this image illustrates the characteristic strong nuclear staining pattern of such lesions.
Supplemental Figure 3. Phase contrast images to illustrate the impact of PHF8 silencing on the formation of invasive features in PC-3 PrCa cells grown in organotypic 3-D culture. Top panel: siPHF8 compared to scrambled control transfection after 4 days of 3-D culture. Lower panel: the same area after 6 days of 3-D culture. More invasive structures now dominate the image. PHF8 silencing affects both growth of spheroids as well as shape.
Supplemental Figure 4: Analysis of cell motility in wound-healing assays. The IncuCyte real-time imaging device and the ‘wound maker” tool were used to scape off a defined area to be closed by the wound margins. PHF8 silenced versus scrambled-control siRNA transfected LNCaP cells only show a marginal difference under these cell culture conditions.
Supplemental Table 1: Full content of the Epigenetic Players (Chromatinome) siRNA library, indicating protein domains with epigenetic functions and/or literature-based association reasons for their inclusion.
Supplemental Table 2: GO (gene ontology) categories enriched in clusters identified by primary functional screens, performed on cell spot microarrays.
Supplemental Table 3: Content of the JmjD/Histone Demethylase siRNA library used for secondary screening, including siRNA sequences.
Supplemental Table 4: Cell-line specific hits with z-scores from primary and secondary siRNA screens.
Supplemental Table 5: Genes that correlate with PHF8 mRNA expression in silico transcriptome analyses based on 496 prostate cancer microarrays contained in the GeneSapiens gene expression database.
Supplemental Table 6: Gene Ontology (GO) categories significantly enriched for genes co-expressed with PHF8 in prostate cancer.
Supplemental Table 7: Differential expression of mRNAs in response to specific PHF8 knock-down by siRNA in the LNCaP cell line.
2