CPT1A Over-Expression Increases Reactive Oxygen Species in the Mitochondria and Promotes Antioxidant Defenses in Prostate Cancer

Abstract

Cancers reprogram their metabolism to adapt to environmental changes. In this study, we examined the consequences of altered expression of the mitochondrial enzyme carnitine palmitoyl transferase I (CPT1A) in prostate cancer (PCa) cell models. Using transcriptomic and metabolomic analyses, we compared LNCaP-C4-2 cell lines with depleted (knockdown (KD)) or increased (overexpression (OE)) CPT1A expression. Mitochondrial reactive oxygen species (ROS) were also measured. Transcriptomic analysis identified ER stress, serine biosynthesis and lipid catabolism as significantly upregulated pathways in the OE versus KD cells. On the other hand, androgen response was significantly downregulated in OE cells. These changes associated with increased acyl-carnitines, serine synthesis and glutathione precursors in OE cells. Unexpectedly, OE cells showed increased mitochondrial ROS but when challenged with fatty acids and no androgens, the Superoxide dismutase 2 (SOD2) enzyme increased in the OE cells, suggesting better antioxidant defenses with excess CPT1A expression. Public databases also showed decreased androgen response correlation with increased serine-related metabolism in advanced PCa. Lastly, worse progression free survival was observed with increased lipid catabolism and decreased androgen response. Excess CPT1A is associated with a ROS-mediated stress phenotype that can support PCa disease progression. This study provides a rationale for targeting lipid catabolic pathways for therapy in hormonal cancers.

Keywords: CPT1A; ROS; androgen response; fatty acids; oxidative stress; prostate cancer; serine.

Figure 1

Cells with overexpression of carnitine palmitoyl transferase I (CPT1A) show a lipid catabolic phenotype and increased growth when supplemented with fatty acids. (A) mRNA expression of C4-2 cells with decreased (knockdown (KD)) and overexpression (OE) of CPT1A cells and their respective controls (non-targeting (NT) and empty virus (EV)). (B) Intracellular triglyceride lipase assay in OE cells. (C,D) Short and medium (C) and long chain (D) acyl carnitine content in OE (red) versus KD (blue), normalized to their own controls. (E) Acyl carnitine changes in C4-2 parental cells treated with etomoxir (100 μM) for 48 h. (F,G) Colony growth after treatment with C12:0 fatty acid conjugated to BSA for 14 days and normalized to BSA-only treatment. (H,I) Colony growth after treatment with C16:0 fatty acid conjugated to BSA for 14 days and normalized to BSA-only treatment. * p < 0.05, ** p < 0.01, *** p <0.001.

Figure 2

RNAseq analysis of CPT1A OE cells shows increased endoplasmic reticulum (ER) stress response, serine metabolism, and less androgen receptor (AR) signaling. (A) Schematic of the RNAseq analysis paradigm with the CPT1A-KD and OE cells. (B) Gene set enrichment analysis was performed on the fold change of the comparison. Normalized Enrichment Scores (NES) and False discovery rate (FDR) adjusted p-values are shown for select pathways, as well as the rank of those genes in fold change ranking, are plotted. (C) Heatmap of the leading-edge genes for select pathways. (D–F) Gene expression graphs of significant genes associated with lipid catabolism (D), serine biosynthesis (E) and ER stress (F) in OE versus KD comparison after adjustment to their respective control cell lines. Adj p value < 0.001.

Figure 3

Excess CPT1A is associated with serine and glycine metabolism pathway and glutathione homeostasis metabolites. (A) Metabolite set enrichment analysis of OE metabolites versus KD cell metabolites after normalization to their own controls. The p-value is defined by the color scale, and the enrichment by the size of the circle. (B) Relative abundance of metabolites associated with the serine and glycine biosynthesis via the SHMT2 gene in the mitochondria. A scheme of the pathway is shown under the graphs, highlighting the sequence of events from glucose to synthesis of serine, glycine and the one carbon metabolism compound dimethylglycine. The genes significantly altered in OE cells from the RNAseq analysis involved in the pathway are shown in red. (C) Relative abundance of metabolites associated glutathione (GSH) homeostasis. The S-glutathionyl-L-Cys metabolite is associated with ineffective handling of oxidative stress. A scheme to the pathway is shown below. * p < 0.05, ** p < 0.01. (D) Gene expression analysis of OE versus KD cells for CTH (p = 6.29 × 10⁻¹³), CHAC1 (p = 3.6 × 10⁻¹²) and GSTO2 (p = 0.003) genes (adj p-value).

Figure 4

Mitochondrial reactive oxygen species (ROS) are increased in CPT1A-OE cells. (A) Electron paramagnetic resonance spectroscopy (EPR) traces of the four cell lines tested, including the mitochondrial probe by itself. Amplitude and linewidth provide information of the amount of ROS. The concentration was acquired by Spin Fit followed by SpinCount module (Bruker). (B) Quantification of the EPR traces normalized to protein content. *** p < 0.001. (C) Gene expression analysis of OE versus KD cells for SOD1 ns), SOD2 (p = 6.9 × 10⁻⁴), and SOD3 (p = 0.015). Only SOD2 is a mitochondrial enzyme. (D,E) Western blots of SOD2 expression after incubation with fatty acids (Oleic and palmitate mixture; 25 µM each) in FBS (D) or charcoal stripped serum (CSS) (E) for 48 h. (F) Schematic of the role of SOD2 in metabolizing superoxide and the role of glutathione in eliminating H₂O₂. GPX = Glutathione peroxide. GSSG = oxidized glutathione.

Figure 5

Lipid and serine metabolism genes are associated with less androgen signaling and a neuroendocrine phenotype. (A) Gene set enrichment analysis (GSEA) plot of the OE versus KD analysis showing the decrease in the androgen response hallmark in OE cells. (B) Gene expression analysis of OE versus KD cells for neuronal-like markers like ENO2 (p = 6.1 × 10⁻¹⁰), SYP (p = ns), NCAM2 (p = 4.14 × 10⁻⁵), NXPH4 (p = 1.6 × 10⁻¹⁷). (C) GSEA plots of the public GSE32967 dataset showing the comparison between small cell carcinoma (n = 3) versus adenocarcinoma (n = 3) patient xenografts for the Hallmark Androgen Response; GO_Serine and GO_Tetrahydrofolate pathways. Normalized enrichment scores (NES) and statistical significance are indicated in the plots. (D) Heatmap of the leading-edge genes from the serine and tetrahydrofolate functional GSEA plots shown in panel C, which are also genes increased in the OE cells compared to KD cells. CPT1A and AR are also included and show an inverse correlation in one of the samples of each subtype of PCa studied.

Figure 6

Increased lipid catabolism and decreased androgen response is associated with less progression-free survival. The TCGA prostate adenocarcinoma (PRAD) Firehose legacy dataset (n = 492) was divided by median split into high and low pathway scores. Pathways from our OE versus KD RNAseq analyses (Figure 2C) were then studied for progression free survival (PFS) in the TCGA PRAD dataset. Kaplan–Meier (KM) plots for Lipid Catabolism Process (A) and Androgen Response Hallmark (B), with the number at risk and p-value (logrank test) are shown.