Lipid degradation promotes prostate cancer cell survival

Abstract

Prostate cancer is the most common male cancer and androgen receptor (AR) is the major driver of the disease. Here we show that Enoyl-CoA delta isomerase 2 (ECI2) is a novel AR-target that promotes prostate cancer cell survival. Increased ECI2 expression predicts mortality in prostate cancer patients (p = 0.0086). ECI2 encodes for an enzyme involved in lipid metabolism, and we use multiple metabolite profiling platforms and RNA-seq to show that inhibition of ECI2 expression leads to decreased glucose utilization, accumulation of fatty acids and down-regulation of cell cycle related genes. In normal cells, decrease in fatty acid degradation is compensated by increased consumption of glucose, and here we demonstrate that prostate cancer cells are not able to respond to decreased fatty acid degradation. Instead, prostate cancer cells activate incomplete autophagy, which is followed by activation of the cell death response. Finally, we identified a clinically approved compound, perhexiline, which inhibits fatty acid degradation, and replicates the major findings for ECI2 knockdown. This work shows that prostate cancer cells require lipid degradation for survival and identifies a small molecule inhibitor with therapeutic potential.

Keywords: ECI2; androgen receptor; cell cycle; lipid degradation; metabolism.

Figure 1. Enoyl-CoA delta isomerase 2 (ECI2) is over-expressed in prostate cancer

(A) ECI2 expression was evaluated in prostate cancer tissue samples. The data shown represents matched normal epithelium and adenocarcinoma from 20 radical prostatectomy specimens. Relative expression of the different transcripts were calculated using the comparative CT method, where the matched benign tissue of the same patient were set to 1 and normalized to the geometric mean CT value of GAPDH, TBP and 18s. Wilcoxon matched-pairs signed rank test was used to test for significance in the differential expression of ECI2 between the matched benign and cancer tissue. (B) Kaplan Meier curves for the low/medium group versus the high ECI2 expressing group. We evaluated whether ECI2 expression levels are associated with survival in prostate cancer patients. The difference in overall survival between the low/medium expressing group and high expressing group was 77 months vs 115 months, p = 0.0086. Here stating that an overview of the clinical cohorts use in Figures 1A and 1B and the statistical analysis are to be found in Supplementary Tables 2, 4 and 5.

Figure 2. Androgen receptor (AR) regulates Enoyl-CoA delta isomerase 2 (ECI2) expression

(A) Chromatin immunoprecipitation (ChIP) of androgen receptor (AR) in VCaP cells. Cells were deprived of androgens for 3 days and treated either with 1nM R1881 or vehicle, as indicated. The putative AR binding site for ECI2 was identified from a published AR ChIP-seq data set [3]. The data shown is representative of two biological replicates. (B) LNCaP and VCaP cells were treated as in A. Total mRNA was isolated at 12 hours and the expression of ECI2 and actin was evaluated using RT-qPCR. The data shown are an average of three independent experiments with SEM, and significance was evaluated using paired samples Student's t-test, *< 0.05. (C) LNCaP and VCaP cells were deprived of androgens for 3 days, treated either with 1nM R1881 or vehicle (−) and protein lysates were collected. Densitometry was used to evaluate ECI2 levels. (D) Western blot was used to confirm ECI2 knockdown in LNCaP and RWPE-1 cells after 96 hours. Densitometry was used to evaluate ECI2 levels. The growth rate of cells was evaluated with life-cell imaging. The data shown are an average of three independent biological replicates with SEM. The significance was evaluated using paired samples Student's t-test, **< 0.01, ***< 0.001. (E) Cell death activation after ECI2 knockdown. LNCaP cells were reverse-transfected and allowed to attach for 24 hours. At this point, a dye detecting caspase 3/7 activation was added and cumulative activation of caspase 3/7 was followed in real-time using life-cell imaging. Data shown is an average of three biological replicates with SEM.

Figure 3. Metabolomic profiling after ECI2 knockdown in LNCaP cells

(A) Reaction catalyzed by Enoyl-CoA Delta Isomerase 2 (ECI2). (B) The level of extra-cellular lactate, glucose and choline were determined using nuclear magnetic resonance after 72 hours of ECI2 knockdown. The data are presented as % of complete media without cells and are an average (with SEM) of five biological replicates. The significance was evaluated using paired samples Student's t-test, *< 0.05, ***< 0.001. (C) The levels of water-soluble intra-cellular metabolites were determined using mass-spectrometry after 72 hours of ECI2 knockdown. Only metabolites that were affected significantly by at least one siRNA are shown. Data shown are an average (with SEM) of five biological replicates. The significance was evaluated using paired samples Student's t-test, *< 0.05, **< 0.01. (D) ECI2 knockdown in LNCaP (for 72 hours) and VCaP cells (for 96 hours). The data shown are representative of at least three biological replicates for LNCaP cells and two replicates for VCaP cells. Densitometry was used to evaluate signal intensity. (E) The levels of intra-cellular phosphatidylserine (PS), phosphatidylcholine (PC), phosphatidylethanolamine (PE) and phosphatidylinositol (PI) were determined using mass-spectrometry after 72 hours of ECI2 knockdown. Identification of lipid composition is provided in Supplementary Figure 2. Data shown are an average (with SEM) of three biological replicates. The significance was evaluated using paired samples Student's t-test, *< 0.05, **< 0.01. (F) The levels of saturated and un-saturated lipids were determined using mass-spectrometry after 72 hours of ECI2 knockdown (same data-set as presented in Figure 3E). Scrambled samples were set to 1 for all four lipid classes, and siECI2 samples were normalized to this. Data shown are an average (with SEM) of three biological replicates. The significance was evaluated using paired samples Student's t-test, *< 0.05, **< 0.01.

Figure 4. RNA-seq after ECI2 knockdown in LNCaP and RWPE-1 cells

The expression of ECI2 was reduced by treating LNCaP and RWPE-1 cells for 48 hours with siRNA and RNA was collected and used for RNA-seq. (A) Venn diagram shows the number of genes that were differentially regulated by both siRNAs in either LNCaP or RWPE-1 cells, and regulated differentially between the two cell lines. (B) Validation of the RNA-seq data using RT-qPCR. The data shown are an average of at least two biological replicates for both RNA-seq and validation, and the significance was evaluated using paired samples Student's t-test, *< 0.05, **< 0.01, ***< 0.001.

Figure 5. Lipid degradation inhibitor perhexiline activates cell death in prostate cancer cells

(A) Oil red O staining of LNCaP cells after 24 hours of perhexiline treatment. Data shown are an average (with SEM) of five biological replicates. The significance was evaluated using paired samples Student's t-test, *< 0.05. (B) LNCaP cells were treated with increasing doses of perhexiline and growth rate was followed using life-cell imaging. Data shown is an average of four biological replicates with SEM. Significance was evaluated using paired samples Student's t-test, *< 0.05, **< 0.01, ***< 0.001. (C) LNCaP cells were treated with perhexiline for 4 hours or 24 hours, and protein lysates were collected for western blotting. The data shown are representative of three biological replicates. Densitometry was used to evaluate signal intensity. (D) LNCaP cells were treated with perhexiline as indicated and cell death activation was evaluated using life-cell imaging detecting activation of caspases 3 and 7. Data shown is an average of four biological replicates with SEM. Significance was evaluated with paired samples Student's t-test, *< 0.05, **< 0.01. (E) LNCaP cells were treated with perhexiline alone or in combination with androgen deprivation therapy (either Abiraterone or MDV-3100) and growth rate was followed using life-cell imaging. Data shown are an average of four biological replicates with SEM. Significance was evaluated using paired samples Student's t-test, *< 0.05. Red stars indicate comparison between perhexiline and combinatorial treatment, while green stars indicate comparison between androgen deprivation and combinatorial treatment.