Fig 2: LDHB acts in parallel with GPX4 to suppress mitochondria-associated ferroptosis.Immunoblot analysis of LDHB, GPX4, DHODH, LDHA, SLC7A11, and FSP1 in HT1080, MSTO-211H, A549, PANC-1 siRNAs cells, n = 3–6 independent repeats, and the statistical analysis was shown in Supplementary Fig. 3a, b (a). Colocalization analysis by immunofluorescence in HT1080 siRNAs cells after 72 h of transfection (b, scale bar, 20 µm). Images of microscopic analysis (c, 20x objective) and assessment of mitochondrial lipid peroxidation by flow cytometry in HT1080 sgNTC and sgGPX4 cells after 72 h of transfection with siRNAs, n = 3 independent replicates (d, e). Clonogenic assays of parental HT1080 (HT1080-P) and HT1080 cells treated with 4 µM RSL3 after ten cycles (HT1080-R) (f). Immunoblot analysis of LDHB, GPX4, DHODH, and LDHA in HT1080 parental and RSL3 resistant cells. The experiment was repeated three times, yielding consistent results (g). Cell viability of HT1080-P and HT1080-R cells after transfection with siRNAs treated with DMSO or RSL3 alone or in combination with 5 µM FER1 for 48 h, following pretreatment with vehicle, 5 µM FER1 for 24 h, n = 3 independent replicates (h). Immunoblot analysis of LDHB, GPX4, LDHA in HT1080 control and LDHB overexpression cells (i). Cell viability assays of HT1080 cells and HT1080 LDHB ORF cells treated with RSL3 with or without 5 µM FER1 for 48 h, following pretreatment with vehicle, 5 µM FER1 for 24 h. n = 3 independent replicates (j). Clonogenic assay of HT1080 siRNAs cells treated with DMSO or RSL3 alone or in combination with 2 mM glutathione reduced ethyl ester (GSH-mee) for 48 h after pretreatment with vehicle, 2 mM GSH-mee for 24 h (k). Tumor volume and weight of A549 sgNTC shCTRL (n = 12), sgNTC shLDHB (n = 11), sgGPX4 shCTRL (n = 12), sgGPX4 shLDHB (n = 12), sgGPX4 shLDHB treated with liproxstatin-1 (n = 12) xenograft tumors from different mice (l, m), tumor volume and weight of HT1080 sgNTC shCTRL (n = 10), sgNTC shLDHB (n = 10), sgGPX4 shCTRL (n = 10), sgGPX4 shLDHB (n = 12), sgGPX4 shLDHB treated with liproxstatin-1 (n = 12) xenograft tumors from different mice (o, p), 4-HNE expression of A549 and HT1080 sgNTC shCTRL, sgNTC shLDHB, sgGPX4 shCTRL, and sgGPX4 shLDHB xenograft tumors treated with or without liproxstatin-1, n = 10 or n = 12 different regions from different tumors from different mice respectively (n, q). Data were presented as mean ± SEM (l, o) or mean ± SD (e, h, j, m, n, p, q). Two-way ANOVA (h, j, l, o), Unpaired, two-tailed t-test (e, m, n, p, q). ns no significant difference, *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. Source data are provided as a Source Data file.

Fig 3: Pyrimidine synthesis inhibition exacerbates nucleotide insufficiency in net gain aneuploid cells. (A) Diagram depicting the labeled carbon aspartate tracing through the pyrimidine biosynthesis pathway. (B) Bar graphs showing the ratio of labeled UTP to unlabeled UTP from the aspartate tracing experiment in representative diploid, 2N-range aneuploid (clone 2), and 4N-range aneuploid (+1q +2p −4q +7 ++8q −10p +10q +12 +20q −22 −X) HMECs. (C) Ratio of labeled UTP to unlabeled UTP normalized for total chromosomal content relative to the diploid (46 chromosome) baseline. (D) Total UTP (labeled + unlabeled) levels in each cell type, per-cell-normalized. (E) Total UTP (labeled + unlabeled) levels in each cell type, normalized for total chromosomal content. (F) Ratio of reduced (GSH) to oxidized (GSSG) glutathione in each cell type. (G,H) Bar graphs showing the relative abundance of metabolites that increase significantly with low-dose DHODH inhibition (2 nM brequinar), including the DHODH substrate dihydroorotate (G), or decrease with DHODH inhibition (H) in diploid, 2N-range aneuploid, and 4N-range aneuploid HMECs. P-values were calculated using t-tests corrected for multiple testing. Data are summarized from biological replicates across multiple representative clones for each ploidy class, described in more detail for I. (I) Heat map summary of LC/MS-MS-derived metabolomic profiles of diploid (two clones), 2N-range aneuploid (two clones; clones 1 and 2 from Fig. 1C), and 4N-range aneuploid (two clones; +1q +2p −4q +7 ++8q −10p +10q +12 +20q −22 −X and −3q +5 ++8 +11 −15 −18 +20) cells treated with DHODHi with or without uridine rescue, as compared with control conditions for each line. Conditions shown are untreated, treated with 2 nM brequinar, and treated with 2 nM brequinar + 12 μM uridine. FDR values were calculated from edgeR-based analysis corrected for multiple hypothesis testing.

Fig 4: Nucleotide pool stress signaling is activated by low-dose DHODHi and DNA-damaging agents in net gain aneuploid HMECs. (A) Heat map summary of transcriptomic profiles of representative diploid, 2N-range aneuploid (clone 2), and 4N-range aneuploid (−3q +5 ++8 +11 −15 −18 +20) HMECs treated with the DHODH inhibitor brequinar or with a combination of brequinar and UMP, relative to control media conditions. Relative expression levels of ribosomal genes and p53 target genes are outlined with a dotted black box. (B,C) Bar graphs showing relative reads per kilobase per million reads (RPKM) of several ribosomal genes (B) and p53 target genes (C) in diploid, 2N-range aneuploid, and 4N-range aneuploid cells treated with brequinar, brequinar and UMP, or UMP or under control media conditions. (D) Cell viability of untreated diploid, 2N-range aneuploid, and 4N-range aneuploid HMECs treated with 0.5 μM oxaliplatin compared with control media conditions. P-values were calculated using t-tests. Representative HMEC clones are the same as those used in A. (E) Cell viability of diploid (shown in black) and 2N-range aneuploid HMECs (shown in yellow; combined data from clones 1 and 2) that have undergone either no pretreatment (top) or pretreatment with 0.25 μM oxaliplatin (bottom) and were subsequently challenged with 0–40 nM brequinar– with 0–10 μM uridine. (F) Heat map summary of transcriptomic profiles of representative diploid, 2N-range aneuploid (clone 2), and 4N-range aneuploid (+1q +2p −4q +7 ++8q −10p +10q +12 +20q −22 − X) HMECs treated with either oxaliplatin or a combination of oxaliplatin and brequinar, as compared with control conditions.

Fig 5: Net gain aneuploid cells are sensitive to pyrimidine synthesis disruption. (A) Diagram of the pyrimidine biosynthesis pathway. (B) Density plots showing DNA stained with propidium iodide in parental HMECs with AAVS1 (control) knockout, DHODH knockout, and DHODH knockout supplemented with uridine. DHODH knockout causes cells to arrest in S phase, which is rescued by uridine. (C) Cell viability of AAVS1 and DHODH knockout HMECs in diploid and net gain aneuploid (clone 2 from Fig. 1C) backgrounds under uridine supplementation ranging from 0 to 8 μM. The significance of the difference between the AAVS1 knockout and the DHODH knockout cell viability in aneuploid and diploid backgrounds was calculated with t-tests at each uridine dose assayed and is indicated by the asterisks in the table above the graphs. (****) P-value < 0.0001, (***) P-value < 0.001, (**) P-value < 0.01, (*) P-value < 0.05. The aneuploid HMEC line used for this experiment was clone 2 in Figure 1C. (D) The DHODH inhibitor (DHODHi) brequinar phenocopies DHODH knockout. (E) Cell viability of diploid (solid lines/circles) and aneuploid (dotted lines/triangles; clone 1 from Fig. 1C) HMECs after treatment with 0.1–40 nM brequinar doses and 0, 4, and 12 μM uridine doses (depicted using purple, red, and yellow, respectively).