Improving the metabolic fidelity of cancer models with a physiological cell culture medium

Abstract

Currently available cell culture media may not reproduce the in vivo metabolic environment of tumors. To demonstrate this, we compared the effects of a new physiological medium, Plasmax, with commercial media. We prove that the disproportionate nutrient composition of commercial media imposes metabolic artifacts on cancer cells. Their supraphysiological concentrations of pyruvate stabilize hypoxia-inducible factor 1α in normoxia, thereby inducing a pseudohypoxic transcriptional program. In addition, their arginine concentrations reverse the urea cycle reaction catalyzed by argininosuccinate lyase, an effect not observed in vivo, and prevented by Plasmax in vitro. The capacity of cancer cells to form colonies in commercial media was impaired by lipid peroxidation and ferroptosis and was rescued by selenium present in Plasmax. Last, an untargeted metabolic comparison revealed that breast cancer spheroids grown in Plasmax approximate the metabolic profile of mammary tumors better. In conclusion, a physiological medium improves the metabolic fidelity and biological relevance of in vitro cancer models.

Fig. 1. Plasmax sustains cancer cell growth and increases colony formation.

(A and B) Comparison of the formulation of Plasmax and DMEM-F12. na (not applicable) refers to components not present in DMEM-F12, with the exception of linoleate, putrescine, lipoic acid, and thymidine, which are not present in Plasmax. (C) Cell number at the end of a proliferation assay with BT549, CAL-120, and MDA-MB-468 cells, performed in Plasmax or DMEM-F12 under normoxic (N) and hypoxic (H) conditions, 21 and 0.1% O₂, respectively. Means ± SEM; n = 3 independent experiments. (D) Doubling time of MDA-MB-468 cells as determined after each of the 10 consecutive passages in either DMEM-F12 or Plasmax. Means ± SEM. (E) Micrographs showing the morphology of MDA-MB-468 cells cultured in DMEM-F12 or Plasmax for eight passages. (F) Representative images and (G and H) quantification of a colony formation assay performed with BT549, CAL-120, and MDA-MB-468 cells preincubated (2 days), seeded at 500 cells per well, and incubated (12 days) with DMEM-F12 (D) or Plasmax (P) as indicated. Means ± SEM (n = 3 independent experiments). (C, D, G, and H) Each dot represents an independent experiment, and P values refer to a two-tailed t test for paired homoscedastic samples.

Fig. 2. The Plasmax component sodium selenite prevents ferroptosis and thereby enhances the colony formation capacity of TNBC cells.

(A) Images of colony formation assays performed with MDA-MB-468 cells seeded 500 cells per well and incubated for 14 days in 1:1 DMEM-F12/Plasmax mix or in DMEM-F12 supplemented with the different mixed stock solutions or individual components of Plasmax. Images are representative of three independent experiments. (B) Representative images and (C) quantification of colony formation assays performed with MDA-MB-468 cells seeded at 500 cells per well in DMEM-F12 with different concentrations of Na₂SeO₃ and incubated for 14 days. Means ± SEM; n = 3 independent experiments. (D) Representative images and (E) quantification of the colony formation assays performed with MDA-MB-468 cells seeded at the indicated cell density and incubated for 7 days. Means ± SEM; n = 3 independent experiments. (F) Quantification of colony formation assays performed with MDA-MB-468 cells seeded at 5000 cells per well and incubated for 7 days in DMEM-F12 supplemented with FBS and 50 nM Na₂SeO₃, as indicated. Means ± SEM; n = 3 independent experiments. (G) Representative images and (H and I) quantification of colony formation assays performed with MDA-MB-468 cells seeded at 5000 cells per well and incubated for 7 days in DMEM-F12 with 50 nM Na₂SeO₃, 1 mM N-acetylcysteine (NAC), or 100 μM Trolox, as indicated. Means ± SEM; n = 3 independent experiments. (J) Western blot showing GPX4 levels in MDA-MB-468 cells seeded at low (2000 cells/cm²) or high (10,000 cells/cm²) density and incubated in DMEM-F12 for 3 days with 50 nM Na₂SeO₃, as indicated. Images are representative of three independent experiments. (K) GPX activity measured in MDA-MB-468 cells seeded and incubated as for (J). Means ± SEM; n = 3 independent experiments. (L) Representative images and (M) quantification of colony formation assays performed with MDA-MB-468 cells seeded at 5000 cells per well (L) or at 50,000 cells per well (H) and incubated for 7 days in DMEM-F12 supplemented with 50 nM Na₂SeO₃ and 250 nM (1S,3R)-RSL3. Means ± SEM; n = 2 independent experiments. (N) Lipid peroxidation levels of MDA-MB-468 cells seeded as in (J) and incubated in DMEM-F12 for 48 hours with 50 nM Na₂SeO₃ and for the last hour with 50 μM t-butOOH, as indicated. At the end of the incubation, 1 μM BODIPY 581/591 C11 was added to the cells for 15 min. Means ± SEM; n = 3 or 4 independent experiments. (O) Representative images of MDA-MB-468 cells seeded at 10,000 cells/cm² and incubated in DMEM-F12 with Na₂SeO₃ for 72 hours and for the last 24 hours with 50 μM t-butOOH, as indicated. Images are representative of three independent experiments. (P and R) Representative images and (Q and S) quantification of colony formation assays performed with MDA-MB-468 cells seeded at 5000 cells per well and incubated for 7 days in DMEM-F12 supplemented with 50 nM Na₂SeO₃, 2.5 μM deferoxamine, and 50 nM liproxstatin-1 (Lip-1), as indicated. Means ± SEM; n = 2 (P and Q) or n = 3 (R and S) independent experiments. (T) Schematic representation of the factors affecting the colony-forming capacity of TNBC cells. “X” refers to a factor present in the medium of confluent MDA-MB-468 cells and not yet identified. (F, H, I, K, M, N, Q, and S) Each dot represents an independent experiment, and P values refer to a two-tailed t test for paired homoscedastic samples.

Fig. 3. Plasmax induces cell line–specific transcriptomic alterations and prevents the pseudohypoxic gene expression signature of cells cultured in commercial media.

(A) PCA of gene expression obtained from RNA-seq data of BT549, CAL-120, and MDA-MB-468 cells cultured in Plasmax or DMEM-F12, in normoxia. n = 3. Each symbol represents an independent experiment. (B) Venn diagram showing the number of genes that were differentially regulated in these cell lines when cultured in Plasmax and DMEM-F12. (C) Heat map of genes that were significantly [false discovery rate (FDR) of 10%] and coherently regulated (absolute log₂ fold change ≥ 0.585) by culturing cells in Plasmax, in normoxia, in at least two of three cell lines. n = 3, log₂ (mean fold change). (D) Correlation analysis of genes regulated by hypoxia in Plasmax (y axis) and genes regulated by Plasmax in normoxia (x axis). Each dot represents the mean of three independent experiments. (E) Expression levels of HIF1α target genes CA9, TXNIP, PDK1, and BNIP3 in BT549 cells relative to control condition (normoxia, DMEM-F12). Means ± SEM; n = 3, each dot represents an independent experiment, and P values refer to a two-tailed t test for unpaired homoscedastic samples. (F) Western blot showing HIF1α levels in BT549 cells in Plasmax (P) or DMEM-F12 (D), in normoxia (N) and hypoxia (H). (G) Western blot showing HIF1α levels in BT549, CAL-120, and MDA-MB-468 cells cultured in DMEM-F12 or Plasmax in normoxia. (H) Western blot showing HIF1α levels in BT549 cells cultured in normoxia in DMEM-F12, Plasmax, DMEM, and RPMI 1640. (I) Western blot showing pyruvate-dependent HIF1α levels in BT549 cells, cultured in DMEM-F12, Plasmax, and DMEM in normoxia. (F to I) Images are representative of three independent experiments. P values refer to a two-tailed t test for unpaired homoscedastic samples.

Fig. 4. Culture medium defines cell metabolic landscape by altering nutrient exchange rates and metabolic pathways.

(A) Consumption and secretion rates of amino acids and intermediates of glycolysis and the urea cycle in cells cultured in DMEM-F12 and Plasmax. Means ± SEM; n = 3 independent experiments. (B) Correlation analysis between the differential availability of neutral proteinogenic amino acids in Plasmax compared to DMEM-F12 and their respective consumption rates. Means ± SEM; n = 3 independent experiments. P values refer to a two-tailed Pearson test. (C) Consumption of the amino acid bound nitrogen. Means ± SEM; n = 3 independent experiments. (D to I and K) Intracellular abundance of metabolites and metabolite ratios in cells cultured in Plasmax and DMEM-F12. Means ± SEM; n = 3 independent experiments. (J) Western blot showing ASS1 levels in cells cultured in DMEM-F12. Images are representative of three independent experiments. (L) Schematic representation of the urea cycle and expected labeling of argininosuccinate from ¹³C₆ arginine upon reversed ASL activity. (M to O) Intracellular levels of ¹³C argininosuccinate in CAL-120 (M), BT549 (N), and MDA-MB-468 (O) cells cultured for 48 hours in Plasmax and DMEM supplemented with ¹³C₆ and ¹³C₀ arginine, as indicated. Means ± SEM; n = 3 (M) or n = 2 (N and O) independent experiments. P values refer to a two-tailed t test for paired homoscedastic samples.

Fig. 5. Cells cultured in Plasmax approximate the in vivo tumor metabolome better.

(A) Schematic representation of the experimental setup applied to compare the untargeted metabolic profiling of CAL-120 cells cultured in vitro (as 2D monolayers or 3D spheroids) in Plasmax and DMEM-F12, with CAL-120–derived orthotopic tumors. (B) PCA of three independent experiments or n = 3 mice (nine tumor fragments) performed as detailed in (A). (C) Weighted distance between each culture condition and the mean values of tumor samples, as calculated from the PCA reported in (B). P values refer to a two-tailed t test for unpaired homoscedastic samples. a.u., arbitrary units. (D to F) Relative abundance and ratio of urea cycle intermediates in CAL-120 cells grown in Plasmax and DMEM-F12 as 2D and 3D cultures and orthotopic mammary tumors. Means ± SEM; n = 3 independent experiments or n = 3 mice. P values refer to a two-tailed t test for unpaired homoscedastic samples comparing the in vitro conditions to the tumor samples. (G to I) Micrographs of CAL-120–derived mammary tumor tissue showing (G) hematoxylin and eosin (H&E) staining and IHC for (H) Ki67 and (I) ASS1. Arrows indicate stromal cells, whereas the arrowhead denotes vessel endothelium. T, tumor; FP, fat pad.