*4.5. Intracellular Lipase Analysis*

The enzymatic activity of intracellular lipase was measured using a substrate containing 3H Triolein (Perkin Elmer, Waltham, MA, USA) and human serum as a source of ApoC2 as described previously [56]. In brief, LNCaP-C4-2 cells overexpressing CPT1A (OE) and controls (EV) were grown to 80% confluence. Intracellular lipase was made accessible by lysing the cells in heparin containing M-PER cell lysis buffer (Pierce, Rockford, IL, USA). Intracellular lipase activity was determined by incubation with 3H Triolein substrate for 45 min at 37 ◦C. The protein concentration of the lysate was determined to calculate the Lipase-dependent hydrolysis (FFA release) of 3H Triolein per mg, per min.

#### *4.6. Electron Paramagnetic Resonance Spectroscopy*

Mitochondrial ROS production was measured by EPR using the mitochondrial-targeted spin probe 1-hydroxy-4-[2-triphenylphosphonio)-acetamido]-2,2,6,6-tetramethyl-piperidine,1-hydroxy-2,2,6,6 tetramethyl-4-[2-(triphenylphosphonio)acetamido] piperidinium dichloride (mito-TEMPO-H) as previously reported [57]. Cells were grown to 80% confluency prior to the EPR measurements. The mito-TEMPO-H probe was prepared in deoxygenated 50 mM phosphate buffer. Cells were washed and treated with mito-TEMPO-H 0.25 mM in Krebs-HEPES buffer (KHB) containing 100 μM of a metal chelator DTPA to avoid direct oxidation with metal ion or hydroxyl radical generation by Fenton reaction. Cells were incubated for 50 min at 37 ◦C, placed on ice, then gently scraped. 50 μL of cell suspension was loaded in an EPR capillary tube and EPR measurements were performed at room temperature using Bruker EMXnano X-band spectrometer [57]. EPR acquisition parameters were: microwave frequency = 9.6 GHz; center field = 3432 G; modulation amplitude = 2.0 G; sweep width = 80 G; microwave power = 19.9 mW; total number of scans = 10; sweep time = 12.11 s; and time constant = 20.48 ms. mito-TEMPO. Nitroxide radicals concentration was obtained by simulating the spectra using the SpinFit module incorporated in the Xenon software of the bench-top EMXnano EPR spectrometer followed by the SpinCount module (Bruker, Billerica, MA, USA).Total protein was extracted from analyzed samples and quantified with a Bio-Rad DC protein assay kit (Bio-Rad, Hercules, CA, USA), and nitroxide concentrations were normalized to total protein.

#### *4.7. Statistics*

Student *t*-tests or ANOVA tests were used to compare between groups, followed by post hoc tests when appropriate, alpha = 0.05. Analysis was carried out with GraphPad Prism software v8 (GraphPad Software, San Diego, CA, USA). All data represent mean ± SD, unless otherwise indicated.

## *4.8. RNAseq and Pathway Analysis*

All cells (KD, OE, and their respective controls) were grown to 80% confluency before RNA isolation. RNA was extracted using a RNeasy Plus Mini Kit (Qiagen, Valencia, CA, USA). RNA quality was verified using a High Sensitivity ScreenTape Assay on the Tape Station 4200 (Agilent Technologies, Santa Clara, CA, USA) and measured with a Tecan Plate Reader (Thermo Fisher Scientific, Waltham, MA, USA). Library construction was performed using the Universal Plus mRNA Library Kit (NuGen Technologies, Redwood City, CA, USA), and sequencing was performed on the NovaSeq 6000 instrument (Illumina, San Diego, CA, USA) using paired-end sequencing (150 bp) by the University of Colorado Cancer Center Genomics and Microarray Core. Illumina adapters were removed using BBDuk (sourceforge.net/projects/bbmap) and reads <50 bp after trimming were discarded. Reads were aligned and quantified using STAR (2.6.0a) [58]) to the Ensembl human transcriptome (hg38.p12, release 96). Normalization and differential expression were calculated using the limma R package [29]. An interaction model within limma used to directly compare the OE and the KD. Gene set enrichment analysis was performed using the fGSEA R package (v1.10.0) with 10,000 permutations and the Hallmarks and GO Biological Processes gene set collections from the Molecular Signatures Database [59]. Heatmaps were generated with the ComplexHeatmap R package [60] following *z*-score

transformation. RNA-sequencing data have been deposited into the NCBI Gene Expression Omnibus database (accession number GSE161243).
