Acidosis Forces Fatty Acid Uptake and Metabolism in Cancer Cells Regardless of Genotype

Abstract

While proteins facilitate fatty acid (FA) partitioning into plasma membranes, movement between membrane leaflets occurs through a "flip-flop" mechanism. This study provides evidence that biological acidosis, as encountered in tumors and ischemic diseases, promotes FA protonation, thereby enhancing neutral, non-ionized FA uptake. This positions the altered lipid metabolism in acid-exposed cells as a consequence, rather than a cause, of preferential FA uptake. Cancer cell vulnerability, independent of their genetic background, directly stems from this paradigm shift, as detoxifying the overload of very long-chain FA (VLCFA) becomes highly dependent on peroxisomal activity. Inhibition of peroxisomal function in acid-exposed cancer cells leads to the rerouting of these fatty acids into triglycerides within lipid droplets, but also into phospholipids, contributing to membrane alterations, triggering ER stress, and ultimately supporting cytotoxicity. Using patient-derived tumor organoids and sera from human volunteers supplemented with polyunsaturated FA (PUFA), it is shown that inhibiting peroxisomal ACOX1 selectively kills acid-exposed cancer cells, an effect exacerbated by pharmacological stimulation of glycolysis. Similar acid-driven FA uptake is observed in endothelial cells and cardiac myocytes, opening new therapeutic avenues not only cancer but also cardiovascular diseases.

Keywords: acidosis; cancer; fatty acid; lipid metabolism; peroxisome.

Figure 1

Extracellular pHe determines the extent of fatty acid (FA) uptake. A) Normalised quantification of ¹⁴C FA uptake during stepwise medium acidification in SiHa cells (vs maintaining stable pHe) for palmitic acid (PA) (left graph) and docosahexaenoic acid (DHA) (right graph) (N = 3, n = 1). B,C) Representative pictures B) and quantification C) of ORO‐stained lipid droplets (LD) in SiHa cells maintained at pHe 7.4 (7.4 > 7.4) or shifted to pHe 6.5 (7.4 > 6.5) upon supplementation for 24 h with the indicated concentration ranges of PA or DHA (N = 3, n = 5 fields). D) Schematic protocol of pHe swap in cancer cells maintained at physiological pHe (7.4) or chronically adapted to acidic pHe (6.5). E) Time‐course quantification of LD accumulation (expressed as a ratio to LD content at time 0) in SiHa performed using Nanolive imaging after supplementation with 25 µm DHA or FA‐free BSA as vehicle (N = 1, n = 2). α = early phase slope values (i.e., before the plateau). F,G) Representative end‐point pictures F) and quantification G) of LD in SiHa cells maintained at the indicated pHe (7.4 > 7.4 and 6.5 > 6.5) or following pHe swapping (7.4 > 6.5 and 6.5 > 7.4) and supplemented with 50 µm DHA or FA‐free BSA as vehicle (N = 2, n = 10 fields). H) Quantification of BODIPY‐stained LD in SiHa cells maintained in 7.4 (7.4 > 7.4) or shifted in 6.5 (7.4 > 6.5) upon supplementation for 6 h with DHA and coupled with FA transporter inhibitors (2 µg mL⁻¹ JC63.1, 10 µm SSO, 2 µm FATP1i, 2 µm FATP2i) or DMSO as vehicle (N = 3, n = 5). I,J) Representative pictures I) and quantification J) of ORO‐stained LD in SiHa cells exposed to the indicated pHe and supplemented with 50 µm DHA or FA‐free BSA as vehicle (N = 2, n = 5 fields). K,L) Representative pictures K) and quantification L) of pHi in HCT116 cells transduced with a ratiometric pH‐sensitive protein (pHluorin2) and exposed for 160 min to 50 µM DHA or FA‐free BSA as vehicle (N = 3, n = 1). M. HPLC‐MS quantification of DHA esterified in phospholipids from SiHa cells maintained at 7.4 (7.4 > 7.4) or shifted to 6.5 (7.4 > 6.5) and supplemented with 50 µm DHA or Methyl‐DHA for 24 h (N = 3, n = 1). All scale bars: 10 µm. Vehicles consist of an equivalent concentration of FA‐free BSA for DHA or DMSO for the drugs. Graphs are presented as mean ± SEM (A, C, J, L), min to max whisker plots, range = Q1 to Q3, and line = median G,H) or mean ± SD (M). N = biological replicates, and n = technical replicates. ns = non‐significant; ***p < 0.001. Significance was determined by two‐way ANOVA with Tukey's multiple comparison test. The p‐values reported for A and C refer to the column factor, whereas those in L indicate the differences between the final points of the kinetics.

Figure 2

Extracellular acidification by hypoxia or stimulated glycolysis enhances DHA uptake across cancer cell types. A) Schematic protocol for inducing hypoxia (1% O₂) in cancer cells cultured in either non‐buffered or buffered bicarbonate buffer, resulting in spontaneous medium acidification or preventing pHe drop, respectively. B,C) Representative pictures B) and quantification C) of ORO‐stained LD in SiHa and HCT116 cells supplemented with 25 µm DHA or FA‐free BSA as vehicle in the indicated conditions (N = 2, n = 10 fields). D) Schematic protocol for reducing or increasing pHe using MPC inhibitor 7ACC2, and PDK inhibitor DCA, respectively. E,F) Representative pictures E) and quantification F) of ORO‐stained LD in SiHa and HCT116 cells supplemented with 25 µm DHA or FA‐free BSA as vehicle after the indicated pharmacological treatments (DMSO as vehicle) (N = 2, n = 10 fields). All scale bars: 10 µm. The vehicle consists of an equivalent concentration of FA‐free BSA for DHA and DMSO for drugs. All graphs are presented as min‐max whisker plots, with the range indicated by Q1 to Q3, and the line representing the median. N = biological replicates, and n = technical replicates. ns = non‐significant; ***p < 0.001. Significance was determined by two‐way ANOVA with Tukey's multiple comparison test.

Figure 3

Enhanced DHA uptake under acidosis sensitizes cancer cells to peroxisomal inhibition. SiHa cells maintained at the indicated pHe (7.4 > 7.4 and 6.5 > 6.5) or following pHe swapping (7.4 > 6.5 and 6.5 > 7.4) were supplemented with 50 µm DHA or FA‐free BSA as vehicle, and treated as indicated. A) Effects of DGAT1 inhibition (15 µm A922500 vs DMSO, 72 h) in the different pHe conditions (N = 3, n = 6). B) Quantification of membrane fluidity in SiHa cells exposed for 48 h to 15 µm DGAT1i (or DMSO as vehicle); Generalized polarization (GP) was calculated for both plasma and endo‐membranes (n = 30 independent cells). C,D) Representative pictures of the Hyper3 oxidized and reduced forms in transduced SiHa cells maintained at acid pH 6.5 C) and quantification of Ex₄₈₈/Ex₄₀₅ Ratio (reflecting H₂O₂ production) normalised to control in SiHa cells exposed for 48 h to 15 µm DGAT1i (or DMSO as vehicle) in each pHe condition (N = 2, n = 10 fields) D). E) Effects of ACOX1 inhibition (60 µm tricosadiynoic acid vs DMSO, 24 h) in the different pHe conditions during supplementation with 50 µm DHA, C22:1, or FA‐free BSA as vehicle (N = 3, n = 6). F,G) Representative pictures F) and quantification G) of ORO‐stained LD in SiHa cells undergoing the indicated pHe changes and exposed to 30 µm ACOX1i for 48 h (N = 2, n = 10 fields). H–J) HPLC‐MS quantification of DHA or C22:1 esterified in TG H) and in Phospholipids I), or their β‐oxidation intermediates J) from SiHa cells maintained at 7.4 (7.4 > 7.4) or shifted to 6.5 (7.4 > 6.5) and supplemented with 50 µm DHA or C22:1 and treated with 7.5 µm DGAT1i or 15 µm ACOX1i (DMSO as vehicle) during 24 h (N = 3, n = 1). All scale bars: 10 µm. The vehicle consists of an equivalent concentration of FA‐free BSA for DHA or DMSO for drugs. Graphs are presented as mean ± SD A,E,H,I,J) or min to max whisker plots, range = Q1 to Q3, and line = median B,D,G). bdl = below detection limit. N = biological replicates, and n = technical replicates. ns = non‐significant; ***p < 0.001. Significance was determined by two‐way ANOVA with Tukey's multiple comparison test.

Figure 4

DHA and ACOX1i combination induce apoptosis without triggering ferroptosis. SiHa and HCT116 cells maintained at acidic pHe 6.5 were supplemented with 50 µm DHA or FA‐free BSA as vehicle, and treated as indicated. A) Representative flow cytometry plots depicting 7‐AAD versus FITC‐Annexin V staining of pH6.5‐adapted SiHa and HCT116 cells treated for 8 h with 60 µm ACOX1i (DMSO as vehicle). B) Quantification of the cell number (%) in the 7‐AAD and Annexin V double‐positive quadrants (N = 3, n = 1). C) Effect of cell death inhibitors (30 µm Z‐VAD‐FMK for apoptosis or 30 µm Ferrostatin‐1 and 20 µm α‐tocopherol for ferroptosis) on the viability of SiHa and HCT116 cells exposed to 40 µM ACOX1i, or 5 µm RSL3 as positive control of ferroptosis (DMSO as vehicle) (N = 3, n = 6). D,E) Representative flow cytometry histograms D) and quantification E) of ROS in SiHa and HCT116 cells treated for 5 h with 60 µm ACOX1i (DMSO as vehicle), as determined using DCFDA (N = 4, n = 1). F) Live cell confocal microscopy assessment of ER morphology in SiHa upon exposure for 5 h to 60 µm ACOX1i (DMSO as vehicle), as determined using 1 µm ER‐tracker green reporter. G) Quantification of membrane fluidity (2 µm Laurdan) in SiHa cells exposed for 24 h to 15 µm ACOX1i (or DMSO as vehicle); Generalized polarization (GP) was calculated for ER membrane thanks to 1 µm ER‐tracker red staining (N = 3, n = 4). H,I) Representative pictures H) and quantification I) of nuclear CHOP immunofluorescence in fixed SiHa cells treated for 4 h with 60 µm ACOX1i (DMSO as vehicle) (n = 60 independent cells). J,K) RT‐qPCR of genes associated with UPR J) and immunoblotting of ubiquitinylated proteins K) in SiHa cells treated for 5 h with 60 µm ACOX1i (DMSO as vehicle) and 50 µm DHA treatments (FA‐free BSA as vehicle) (N = 6 for qPCR and N = 1 for Immunoblotting). All scale bars: 10 µm. Graphs are presented as mean ± SD B,C,E), mean ± SEM J), or min to max whisker plots, range = Q1 to Q3, and line = median G,I). N = biological replicates, and n = technical replicates. ns = non‐significant; ***p < 0.001. Significance was determined by one‐way ANOVA B,E,G,I,J) and two‐way ANOVA C) with Tukey's multiple comparison test.

Figure 5

Acidosis‐inducing MPCi enhances the cytotoxic effects of DHA supplementation. Cancer cells, spheroids, and organoids were pre‐challenged with 10 µM 7ACC2 (DMSO as vehicle) to stimulate glycolysis and induce medium acidification and treated as indicated with ACOX1 inhibitor (tricosadiynoic acid) (DMSO as vehicle) and DHA (FA‐free BSA as vehicle). A) Effects of combining 75 µm DHA and ACOX1 inhibition (60 µm, 72 h) on the viability of SiHa and HCT116 cells under induced acidification (N = 3, n = 6). B,C) Representative pictures B) and quantification C) of ORO‐stained LD in SiHa and HCT116 cells under induced acidification and exposed to 50 µm DHA and 30 µm ACOX1i (N = 2, n = 10 fields). Scale bar: 10 µm. D) Effects of combining 75 µm DHA and DGAT1 inhibition (30 µm, 72 h) (DMSO as vehicle) on the viability of SiHa and HCT116 cells under induced acidification (N = 3, n = 6). E,F) Representative pictures E) and quantification F) of cytotoxic effects measured by Cytotox Green Reagent of combining 100 µm DHA and 60 µm ACOX1i in HCT116 spheroids under induced acidification (N = 3). Scale bar: 300 µm. G,H) Representative pictures G) and quantification H) of cytotoxic effects of combining 50 µm DHA and 60 µm ACOX1i on colorectal cancer organoids under induced acidification (N = 18). Scale bar: 100 µm. Graphs are presented as mean ± SD A,D, H), mean ± SEM F), or min to max whisker plots, range = Q1 to Q3, and line = median C). N = biological replicates, and n = technical replicates. ns = non‐significant; ***p < 0.001. Significance was determined by two‐way ANOVA with Tukey's multiple comparison test.

Figure 6

Acidosis enhances the detrimental effects of combining dietary ω3‐PUFA and ACOX1 and promotes FA uptake in non‐cancer cells. A) Tumor growth in mice injected with SiHa cancer cells transduced with ACOX1 shRNA (vs scramble shRNA). Mice were fed with control (left) or PUFA‐rich (right) diets and treated or not with MPCi 7ACC2 (3 mg kg⁻¹) (N = 5). B) Representative pictures of FaDu tumors collected from mice fed either diet and treated with ACOX1i (10 mg kg⁻¹), MPCI 7ACC2 (3 mg kg⁻¹), or both. Scale bar: 1 mm. C,D) Quantification of cancer cells detected by pan‐keratin staining C) and fibroblasts detected by vimentin staining D) in FaDu tumors collected from mice fed either diet, and treated as in B (N = 5,6). E,F) Representative pictures of liver sections from mice bearing FaDu tumors and treated as in B (E) and quantification of the extent of liver toxicity revealed by vacuolization (F) (N = 6). Scale bar: 30 µm. G,H) Cytotoxic effects of 60 µm ACOX1i (DMSO as vehicle) on human colorectal cancer organoids cultured at pHe 7.4 or 6.5 G) or undergoing medium acidification using 7ACC2 treatment (DMSO as vehicle) H), and supplemented with human serum from 5 healthy donors, before and after 1‐week ω3 PUFA dietary supplementation (N = 5, n = 4). I,J) Representative pictures I) and quantification J) of ORO‐stained LD in endothelial cells cultured at the indicated pHe, and supplemented with 500 or 1000 µm OA (FA‐free BSA as vehicle) (N = 3, n = 4 fields). Scale bar: 10 µm. K) Quantification of ORO‐stained LD in cardiomyocytes cultured at pHe 7.4 or 6.5, either under normoxia or hypoxia (1% O₂), and supplemented with 50 µm PA, OA, or DHA (FA‐free BSA as vehicle) (N = 3, n = 4 fields). Graphs are presented as mean ± SD C,D,F,G, and H), mean ± SEM A), or min to max whisker plots, range = Q1 to Q3, and line = median J and K). N = biological replicates and n = technical replicates. ns = non‐significant; ***p < 0.001. Significance was determined by one‐way ANOVA, Tukey's multiple comparison test (A with endpoints, C, D), and two‐way ANOVA, Tukey's multiple comparison test F,G,H,J,K).