Fig 2: HIF1a and HIF2a dependence varies across VHL mutant ccRCC cell lines based on genome-wide CRISPR-Cas9 screens. EPAS1 and HIF1A dependency (CERES) scores of VHL mutant ccRCC cell lines from the DepMap database of genome-wide CRISPR-Cas9 dependency screens. CERES scores estimate gene dependency, taking into account the copy number–related fitness effects of CRISPR-Cas9 genome editing.

Fig 3: TP53 mutation is not required for the HIF2a independence of 786-O HI cells. (A) Relative ratios of mixtures of 786-O HI cells expressing 1) GFP and either a nontargeting control sgRNA (sgNT) or the indicated sgRNA targeting EPAS1, or 2) mCherry and sgNT, over time, as determined by flow cytometry. Ratios were normalized such that the ratio at T = 1 wk after lentiviral introduction of the reporters and sgRNAs was 1. Data represent the mean ± SEM of at least two independent replicates. (B) Representative photomicrographs of soft agar colonies formed by 786-O and 786-O HI cells that were grown in the presence of 2 µM PT2399 or DMSO vehicle control. (C) Quantification of soft agar colonies formed by the cell lines in B. Shown are mean colony numbers in the presence of 2 µM PT2399 relative to DMSO. Data are mean ± SEM of at least two independent replicates. (D) Relative ratio of mixtures of 786-O HI cells (filled circles, cells expressing nontargeting sgRNA; open circles, cells expressing sgTP53) that were then engineered to express 1) GFP and either a nontargeting control sgRNA (sgNT) or the indicated sgRNA targeting EPAS1, or 2) mCherry and sgNT, over time, as determined by flow cytometry. Ratios were normalized such that the ratio at T = 1 wk after lentiviral introduction of the reporters and sgRNAs was 1. Data represent the mean ± SEM of at least two independent replicates. (E) Representative photomicrographs of soft agar colonies formed by 786-O HI cells as in D that were grown in the presence of 2 µM PT2399 or DMSO vehicle control. (F) Quantification of soft agar colonies formed by the cell lines as in D. Shown are mean colony numbers in the presence of 2 µM PT2399 relative to DMSO. Data are mean ± SEM of at least two independent replicates.

Fig 4: HIF2a dependence varies across VHL mutant ccRCC cell lines based on cellular competition and anchorage-independent growth assays. (A) Relative ratios of mixtures of cells expressing 1) GFP and either a nontargeting control sgRNA (sgNT) or the indicated sgRNA targeting EPAS1, or 2) mCherry and sgNT, over time, as determined by flow cytometry. Ratios were normalized such that the ratio at T = 1 wk after lentiviral introduction of the reporters and sgRNAs was 1. Data represent the mean ± SEM of at least two independent replicates. (B) Representative photomicrographs of soft agar colonies formed by the indicated VHL mutant ccRCC cell lines that were grown in the presence of 2 µM PT2399 or DMSO vehicle control. (C) Quantification of soft agar colonies formed by the indicated VHL mutant ccRCC cell lines as in B. Data represent mean ± SEM of at least two independent replicates.

Fig 5: Differential sensitivity to mTOR inhibition amongst functional groups of transcripts induced upon VHL loss.(a) The figure is identical to Fig. 5b, but genes involved in the HIF signaling pathway are labelled. (b) Gene set enrichment analysis among genes induced upon VHL loss (FDR < 0.1 and mRNA fold change > 1.5) and either hypersensitive (green) or resistant (pink) to mTOR inhibition in RCC4 cells (see Methods for definition). The top 5 most enriched gene ontology or KEGG orthology terms are shown. n = 124 and 122 mTOR hypersensitive or resistant genes were considered for the analysis. (c) Schematic showing inverse changes in mRNA abundance upon VHL loss to changes in translational efficiency upon mTOR inhibition for genes encoding glycolytic enzymes in RCC4 cells. (d) Boxplots comparing changes in translational efficiency for the indicated classes of HIF-target gene in response to treatment with an mTOR inhibitor in this and previous published studies14,21. Data are expressed as log2 fold change in mean ribosome load (MRL) for the HP5 data and as translational efficiency for the ribosome profiling data. The distributions in each data set were compared using the two-sided Mann-Whitney U test; the number of genes in each class is indicated in parenthesis. The horizontal lines show the median changes in MRL or translational efficiency of all expressed genes. The colours indicate whether the changes are above or below the median value of all genes (blue and red respectively). PC3: human prostate cancer cells; MEF: mouse embryonic fibroblasts. (e) Proportion of genes with the indicated number of uORFs (top panel), CDS length (middle panel), and the length of TOP motif (bottom panel), amongst the specified functional class of HIF-target genes. For the plot of the CDS length, the two most mTOR sensitive groups are indicated by a red box. These analyses were performed for genes for which the 5' UTR sequence could be predicted for the majority of expressed mRNAs (n = 11 and 27 for glycolysis and angiogenesis or vascular process genes respectively). (f) Boxplots showing changes in mRNA abundance of HIF-target genes upon VHL loss in 786-O cells (that is induction of HIF2A) as a function of the encoded protein class. The changes in mRNA abundance of angiogenesis or vascular process genes were compared against those of all other HIF-target genes using the two-sided Mann-Whitney U test (n = 27 and 305 respectively). Boxplots show the median (horizontal lines), first to third quartile range (boxes), and 1.5× interquartile range from the box boundaries (whiskers). Source data