Extended Data Fig. 4: Cellular pathways for ribose salvage from uridine.

a–c. Relative metabolic activity (RMA) of PDA cell lines depicting the preferential use of uridine at (a,b) low glucose concentration (0-1 mM) but not at c) the high glucose concentration (10 mM), over a 96 h culture. d. Schematic depicting metabolic pathways for uridine utilization. e. Expression of PGM2, UCK1, and UCK2 in non-tumour (NT) and PDA tissue samples from the GSE71729 dataset. Number of samples: NT = 46, PDA = 145. f–h. Expression of PGM2, UCK1 and UCK2 in TCGA (human PDA tumour) and CCLE (human cell line) data separated into UPP1 low (L) and high (H) subsets. i. Western blot for PGM2 in PDA cell lines. Presented in bold are cells that express high UPP1. These samples are the same batch as the data shown in Fig. 1e and the blot was generated during one of the technical replicates of that western blotting. kDa, unit for molecular weight. j. qPCR for PGM2 in ASPC1 cells transfected with siPGM2 compared to non-targeting (siNT) control. On the right: RMA in PGM2 knockdown cells +/− uridine (1 mM) or glucose (1 mM). k. qPCR for UCK1 in ASPC1 cells transfected with siUCK1 compared to non-targeting (siNT) control. On the right: RMA in UCK1 knockdown cells +/− uridine (1 mM) or glucose (1 mM). l. qPCR for UCK2 in ASPC1 cells transfected with siUCK2 compared to non-targeting (siNT) control. On the right: RMA in UCK2 knockdown cells +/− uridine (1 mM) or glucose (1 mM). Statistics and reproducibility: a–c, n = 3 biologically independent samples. Statistical significance for data in a-b was measured using one-way ANOVA with Tukey’s multiple comparisons test. PATU8988S (comparison between no glucose and no glucose + uridine [0.1 mM]: **P = 0.0017; 0.01 mM glucose and 0.01 mM glucose + uridine: ***P = 0.0008; 0.1 mM glucose and 0.1 mM glucose + uridine: ***P = 0.0002; 1 mM glucose and 1 mM glucose + uridine: ****P < 0.0001). CAPAN2 (comparison between no glucose and no glucose + uridine: P = ns (0.5673); 0.01 mM glucose and 0.01 mM glucose + uridine: P = ns (0.0541); 0.1 mM glucose and 0.1 mM glucose + uridine: P = ns (0.092); 1 mM glucose and 1 mM glucose + uridine: ***P = 0.0007. ^#All four group comparisons have significant P: 0.0468, 0.014, 0.0089, 0.0222, respectively, when directly compared using two-tailed unpaired t test). DANG (comparison between no glucose and no glucose + uridine: ****P < 0.0001; 0.01 mM glucose and 0.01 mM glucose + uridine: ****P < 0.0001; 0.1 mM glucose and 0.1 mM glucose + uridine: **P = 0.0051; 1 mM glucose and 1 mM glucose + uridine: **P = 0.0051). ASPC1 (comparison between no glucose and no glucose + uridine: *P = 0.0203; 0.01 mM glucose and 0.01 mM glucose + uridine: **P = 0.0031; 0.1 mM glucose and 0.1 mM glucose + uridine: ***P = 0.0003; 1 mM glucose and 1 mM glucose + uridine: ****P < 0.0001). Statistical significance for data in c was measured using two-tailed unpaired t test and P = ns (0.0852, 0.3509, 0.3021 and 0.3875 for PATU8988S, CAPAN2, DANG and ASPC1, respectively, in the comparison of no uridine vs 0.1 mM uridine groups in the presence of 10 mM glucose). d. Uridine can be channeled into DNA or RNA synthesis by direct phosphorylation from UCK1/2. Uridine can also be catabolized via UPP1 to produce uracil and ribose 1-phosphate. Ribose 1-phosphate is converted to ribose-5-phosphate by PGM2 and fuel pentose phosphate pathway, nucleotide biosynthesis and glycolysis. e. Statistical significance was measured using two-tailed unpaired t test with Welch’s correction. Comparison between NT and PDA: PGM2, ****P < 0.0001; UCK1, ****P < 0.0001; UCK2, *P = 0.018. Box plot statistics – PGM2 (NT: minima = 3.097, maxima = 5.527, median = 4.335, 25th percentile = 3.992, 75th percentile = 4.74; PDA: minima = 2.386, maxima = 6.433, 25th percentile = 4.424, median = 4.961, 75th percentile = 5.457), UCK1 (NT: minima = 3.7, maxima = 5.1, median = 4.5, 25th percentile = 4.2, 75th percentile = 4.7; PDA: minima = 3.6, maxima = 5.1, median = 4.2, 25th percentile = 4, 75th percentile = 4.4), UCK2 (NT: minima = 4.034, maxima = 7.615, median = 5.577, 25th percentile = 5.095, 75th percentile = 5.9; PDA: minima = 4.556, maxima = 6.93, median = 5.727, 25th percentile = 5.458, 75th percentile = 6.059). f-h. Number of samples: TCGA – UPP1 low = 75, high = 75; CCLE – UPP1 low = 22, high = 22. Statistical significance was measured using two-tailed unpaired t test with Welch’s correction. Comparison between L and H groups in TCGA (PGM2: P = ns (0.1226), UCK1: P = ns (0.311); UCK2: *P = 0.0327). In the CCLE L versus H comparison, PGM2, UCK1 and UCK2 have P = ns (0.3486, 0.4645, 0.4381, respectively). TCGA – The Cancer Genome Atlas, CCLE – Cancer Cell Line Encyclopaedia. i. Vinculin is used as a loading control. j. Number of samples: qPCR = 3, RMA = 3 biologically independent samples per group. qPCR – statistical significance was measured using unpaired t test; comparison between siNT and siPGM2: ***P = 0.0007. RMA – statistical significance was measured using multiple unpaired t tests with two-stage two-step method; comparison of siNT and siPGM2 in the presence of 1 mM glucose and no uridine: *P = 0.0452, and **P = 0.0014 in the presence of 1 mM uridine and no glucose. k. Number of samples: qPCR = 3, RMA = 3 biologically independent samples per group. qPCR – statistical significance was measured using two-tailed unpaired t test; comparison between siNT and siUCK1: ***P = 0.0004. RMA – statistical significance was measured using multiple unpaired t tests with two-stage two-step method. Comparison of siNT and siUCK1 knockdown samples in the presence of 1 mM glucose and no uridine: P = ns (0.8652), and P = ns (0.131) in the presence of 1 mM uridine and no glucose. l. Number of samples: qPCR = 3, RMA = 3 biologically independent samples per group. qPCR – statistical significance was measured using unpaired t test; comparison between siNT and siUCK2: ****P < 0.0001. RMA – statistical significance was measured using multiple unpaired t tests with two-stage two-step method; comparison of siNT and siUCK1 knockdown cells in the presence of 1 mM glucose and no uridine: *P = 0.035, and P = ns (0.8653) in the presence of 1 mM uridine and no glucose. Data (a–c, f–h, j–l) are shown as mean ± s.d. The experiments were performed once (a–c, k), and twice (j,l) with similar results.