Identifying off-target effects of etomoxir reveals that carnitine palmitoyltransferase I is essential for cancer cell proliferation independent of β-oxidation

Abstract

It has been suggested that some cancer cells rely upon fatty acid oxidation (FAO) for energy. Here we show that when FAO was reduced approximately 90% by pharmacological inhibition of carnitine palmitoyltransferase I (CPT1) with low concentrations of etomoxir, the proliferation rate of various cancer cells was unaffected. Efforts to pharmacologically inhibit FAO more than 90% revealed that high concentrations of etomoxir (200 μM) have an off-target effect of inhibiting complex I of the electron transport chain. Surprisingly, however, when FAO was reduced further by genetic knockdown of CPT1, the proliferation rate of these same cells decreased nearly 2-fold and could not be restored by acetate or octanoic acid supplementation. Moreover, CPT1 knockdowns had altered mitochondrial morphology and impaired mitochondrial coupling, whereas cells in which CPT1 had been approximately 90% inhibited by etomoxir did not. Lipidomic profiling of mitochondria isolated from CPT1 knockdowns showed depleted concentrations of complex structural and signaling lipids. Additionally, expression of a catalytically dead CPT1 in CPT1 knockdowns did not restore mitochondrial coupling. Taken together, these results suggest that transport of at least some long-chain fatty acids into the mitochondria by CPT1 may be required for anabolic processes that support healthy mitochondrial function and cancer cell proliferation independent of FAO.

Fig 1. Etomoxir (EX) inhibits most of fatty acid oxidation (FAO) at a 10 μM concentration in BT549 cells but does not affect cellular proliferation until much higher concentrations are used.

(A) The pool sizes of acylcarnitines (ACs) decrease by over 80% at 10 μM etomoxir. Additional small decreases are observed at 200 μM etomoxir (n = 3). (B) Isotopologue distribution pattern of citrate after BT549 cells were labeled with 100 μM U-¹³C palmitate for 24 hours. The M+2 isotopologue reflects FAO activity (n = 3). (C) Growth curve of BT549 cells when treated with vehicle control, 10 μM etomoxir, or 200 μM etomoxir (n = 4) (doubling time [DT]). All data are presented as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001.

Fig 2. Mitochondrial respiration and nutrient utilization do not show a dose response to etomoxir because 200 μM etomoxir (EX) has an off-target effect on respiratory complex I.

(A) Nutrient utilization after BT549 cells were treated with vehicle control, 10 μM etomoxir, or 200 μM etomoxir for 48 hours (n = 3). (B) Mitochondrial stress test of whole cells (BT549) after treatment with vehicle control, 10 μM etomoxir, or 200 μM etomoxir for 1 hour (n = 4). (C) Measured and calculated parameters of mitochondrial respiration (generated from data in Fig 2B). (D) 200 μM etomoxir leads to changes in state I respiration but does not affect state II respiration, indicating that 200 μM directly inhibits complex I of the electron transport chain (n = 3). (E) Isotopologue distribution pattern of citrate after BT549 cells were labeled with U-¹³C glucose for 12 hours (n = 3). (F) Isotopologue distribution pattern of citrate after BT549 cells were labeled with U-¹³C glutamine for 6 hours (n = 3). All data are presented as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. The oxygen consumption rate (OCR) was corrected for nonmitochondrial respiration. AA, antimycin A; FCCP, carbonyl cyanide p-trifluoromethoxyphenylhdrazone; oligo, oligomycin; rot, rotenone; suc, succinate.

Fig 3. Knockdown of CPT1A inactivates most of fatty acid oxidation (FAO) and decreases cellular proliferation.

(A) Isotopologue distribution pattern of citrate in BT549 cells with scrambled small interfering RNA (siRNA) (scrambled, black) or after CPT1A knockdown (KD, red). Cells were labeled with 100 μM U-¹³C palmitate for 24 hours, starting at 48 hours after siRNA knockdown. The M+2 peak reflects FAO activity (n = 3). (B) Growth curve of control and CPT1A^KD BT549 cells (n = 4) (DT, doubling time). (C) The decrease in cellular proliferation cannot be rescued by various concentrations of acetate (n = 5). Data are presented as mean ± SEM. **p < 0.01, ***p < 0.001.

Fig 4. Knockdown of CPT1A causes mitochondrial uncoupling.

(A) CPT1A^KD cells (KD, red) uptake more glucose, glutamine, glutamate, and fatty acids relative to scrambled small interfering RNA (siRNA) controls (scrambled, black). CPT1A^KD cells also excrete more lactate (n = 4). (B) Mitochondrial stress test for scrambled siRNA controls and CPT1A^KD cells (n = 3). (C) Measured and calculated mitochondrial respiration parameters (generated from data in Fig 4B). Data are presented as mean ± SEM and normalized to the final number of cells after respiration measurements to account for differences in proliferation. We note that coupling efficiencies are calculated as the ratio of the oxygen consumption rate (OCR) required for ATP production to basal OCR in the same sample and therefore are independent of the sample normalization method. *p < 0.05, **p < 0.01, ***p < 0.001. The OCR was corrected for nonmitochondrial respiration. AA, antimycin A; FCCP, carbonyl cyanide p-trifluoromethoxyphenylhdrazone; oligo, oligomycin; rot, rotenone;.

Fig 5. Imaging mitochondrial dysfunction in CPT1A^KD cells.

(A, B) Mitochondria were stained by Mitotracker red, and nuclei were stained by DAPI. Images from scrambled small interfering RNA (siRNA) controls (A) show less fluorescence intensity of Mitotracker red compared to CPT1A^KD cells (B). (C, D) Representative electron microscopy (EM) images of normal mitochondria in wild-type BT549 cells (C) and the abnormal vesicular morphology of mitochondria in CPT1A^KD cells (D).

Fig 6. The levels of complex lipids are altered in the mitochondria of CPT1A^KD cells.

(A) Scatter plot comparing the integrated intensities of 77 lipid species altered between scrambled small interfering RNA (siRNA) controls and CPT1A^KD cells. All lipids profiled that showed a fold difference ≥ 1.5, a p-value ≤ 0.01, and a signal intensity ≥ 10,000 are displayed. The diagonal line represents the equation y = x, so that points below the line represent the 66 lipids that decrease in abundance in CPT1A^KD cells. (B) The identities and absolute concentrations of dysregulated lipids were determined and the relative differences plotted. Signaling lipids are displayed on top, and structural lipids on bottom. CL, cardiolipin; Cer, ceramide; DG, diacylglycerol; Gal/GlcCer, galactosyl/glucosylceramide; KD, knockdown; LacCer, lactosylceramide; LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PG, phosphatidylglycerol; PS, phosphatidylserine; SM, sphingomyelin. Data are presented as mean ± SEM (n = 3). *p < 0.05, **p < 0.01.

Fig 7. Model for the anabolic role of carnitine palmitoyltransferase I (CPT1) in mitochondrial metabolism.

Acyl-CoA species have anabolic fates in the cytosol (1), in addition to catabolic (2) and anabolic (3) fates in the mitochondrial matrix (e.g., phospholipid sidechain remodeling and protein acylation). ACSL, acyl-CoA synthetase; CAT, carnitine-acylcarnitine translocase; CPT1, carnitine palmitoyltransferase I; CPT2, carnitine palmitoyltransferase II; FA, fatty acid.