Fig 2: PHD3 knockdown affects cell-cell adhesion and actin cytoskeletal morphology.(A) BxPC3 cells stably transduced with retrovirus containing PHD3Wt (BxPC3-Wt), Vector (BxPC3-Vec) or anti-PHD3 shRNA (BxPC3-KD) were grown on tissue culture dishes for 48 hours and live cells were photographed at 20× using phase contrast microscopy. (B) Stable BxPC3-Wt and BxPC3-KD cells grown on glass coverslips were labeled with PHD3 antibody (green) and phalloidin (red) and DAPI (blue) and photographed at 20× using a Zeiss 510 confocal microscope.

Fig 3: PHD3 depletion leads to impaired ccRCC cell growth and decreases the phosphorylation of p70 S6K and S6 ribosomal proteins. a Western blot analysis of 786-O cells with two siRNA sequences targeting PHD3 under normoxia and hypoxia shows decrease in phosphorylation of p70 S6K and S6 ribosomal protein. Right-hand graphs show phosphorylated (S235/236 and S240/244) pS6 ribosomal protein normalized to total S6 protein. Quantification of two biological replicates, mean ± SEM, fold change to Scr (*p < 0.05, n.s. not significant). b Western blot analysis of RCC4 cells with PHD3 depletion under normoxia and hypoxia decreases the phosphorylation of p70 S6K and S6 ribosomal protein. Quantification of two biological replicates, mean ± SEM, fold change to Scr (*p < 0.05, n.s. not significant). c 786-O and RCC4 3D colony forming in Matrigel®. Quantification of three (786-O) or two (RCC4) biological replicates with triplicate wells for each siRNA treatment, mean ± SEM, fold change to Scr (**p < 0.01, ***p < 0.001). Representative images in the panel acquired by using ×4 objective, scale bar 100 µm

Fig 4: The methylated PHD3 gene in melanoma, prostate, breast and renal carcinoma cell lines is refractory to induction by hypoxia. A) Melanoma, prostate and breast carcinoma cell lines were treated with hypoxia (1% O2) or normoxia (21% O2) for 24 hours. Total RNA was extracted and converted to cDNA by reverse transcription. Quantitative real-time reverse transcription-PCR analysis of PHD3 was performed with normalization to GAPDH gene expression. Relative quantitation was determined by the DDCt method. ND = not detectable. Error bars = SEM. n = 3. B) Renal clear cell carcinoma cell lines were either untreated or treated with hypoxia as in A). The right panel depicts relative PHD3 mRNA levels compared to mRNA extracted from normal renal tissue. The left panel depicts the methylation status at each of 58 CpG dinucleotides present in the PHD3 CpG island of the representative renal carcinoma cell lines. Error bars = SEM. n = 3.

Fig 5: Elevated glucose destabilizes HIF-1a through increased PHD activity.(A) 293T WT cells in LG were treated with the indicated doses of PHD inhibitors CoCl2, daprodustat (Dap), and DMOG for 6 h to stabilize HIF-1a protein. (B) 293T WT cells in LG or HG were treated with 100 µM CoCl2 or 50 µM daprodustat and HIF-1a was assessed by Western blot. Relative changes in HIF-1a expression, normalized to ß-actin by densitometry, are shown below the blot. (C) HRE-luciferase plasmid was transfected into 293T WT cells exposed to acute or chronic glucose and/or treated with 100 µM CoCl2 for 24 h. Changes in HIF-1 transcriptional activity were measured via relative luminescence (normalized to transfection control) and analyzed by two-way ANOVA with Sidak’s multiple comparisons. (D) Metal-assisted protein quantification of HIF-1a and downstream target proteins from WT cells grown in LG or HG. Antibody metal labels: HIF-1a 161Dy, PDGFA 165Ho, VEGFA 164Dy, PHD3 151Eu, p-AKT 159Tb, and actin 171Yb. (E) qRT-PCR analysis of IDH1 mRNA in 293T WT cells grown in LG, 24 h HG, or chronic HG (n = 3, one-way ANOVA). (F) Corresponding IDH1 (47 kD) protein analysis. (G) ELISA quantification of 2-KG in 293T WT LG or HG cells. (H) Schematic showing the up-regulation of IDH1 and 2-KG production in HG, leading to PHD-mediated hydroxylation and degradation of HIF-1a.