2,4-dienoyl-CoA reductase regulates lipid homeostasis in treatment-resistant prostate cancer

Abstract

Despite the clinical success of Androgen Receptor (AR)-targeted therapies, reactivation of AR signalling remains the main driver of castration-resistant prostate cancer (CRPC) progression. In this study, we perform a comprehensive unbiased characterisation of LNCaP cells chronically exposed to multiple AR inhibitors (ARI). Combined proteomics and metabolomics analyses implicate an acquired metabolic phenotype common in ARI-resistant cells and associated with perturbed glucose and lipid metabolism. To exploit this phenotype, we delineate a subset of proteins consistently associated with ARI resistance and highlight mitochondrial 2,4-dienoyl-CoA reductase (DECR1), an auxiliary enzyme of beta-oxidation, as a clinically relevant biomarker for CRPC. Mechanistically, DECR1 participates in redox homeostasis by controlling the balance between saturated and unsaturated phospholipids. DECR1 knockout induces ER stress and sensitises CRPC cells to ferroptosis. In vivo, DECR1 deletion impairs lipid metabolism and reduces CRPC tumour growth, emphasizing the importance of DECR1 in the development of treatment resistance.

Fig. 1. AR signalling is conserved in ARI-resistant cells.

a Representative pictures of WT and ARI-resistant LNCaP cells cultured in 2D conditions. Scale bar represents 100 µm. b Representative pictures of WT and ARI-resistant LNCaP organoids embedded in Matrigel. Scale bar represents 50 µm. c Quantification of cell (top panel) and organoid (bottom panel) diameter. d Cell proliferation of WT and ARI-resistant LNCaP cells after 48 and 72 h. Cell count is normalised to initial number of cells at T0. e Cell proliferation of WT and ARI-resistant LNCaP cells treated for 48 h with different AR inhibitors (10 µM). Cell count is normalised to non-treated condition. f Western blot analysis of AR, KLK3 and FKBP5 expression in WT and ARI-resistant LNCaP cells. HSC70 was used as a sample loading control. g RT-qPCR analysis of AR (full length, fl), KLK3 and FKBP5 expression in WT and ARI-resistant LNCaP cells. CASC3 was used as a normalising control. h Immunofluorescence showing nuclear AR expression in WT and ARI-resistant LNCaP cells. Scale bar represents 20 µm. c–e, g Data are presented as mean values +/− SD. c, g *p-value < 0.05 using a 1-way ANOVA with a Dunnett’s multiple comparisons test. d, e *p-value < 0.05 using a 2-way ANOVA with a Tukey’s multiple comparisons test. Source data are provided as a Source Data File.

Fig. 2. Omics analysis of ARI-resistant cells reveal altered glucose metabolism.

a Enriched pathways upregulated in the proteomic analysis of ARI-resistant cells when compared to WT LNCaP. Selected proteins were significantly modulated in at least 2 out of 3 conditions. Pathway enrichment analysis was performed using the STRING database (http://string-db.org). b Steady-state levels of significantly regulated metabolites in ARI-resistant cells when compared to WT LNCaP (FC > 1.5, p < 0.05 using a 1-way ANOVA with a Dunnett’s multiple comparisons test). Selected metabolites were significantly modulated in at least 2 out of 3 conditions. c Isotopologue distribution of selected metabolites in ARI-resistant and WT LNCaP cells following ¹³C-glucose incorporation for 1 h. d Isotopologue distribution of pyruvate and phosphoenolpyruvate in ARI-resistant and WT LNCaP cells following ¹³C-glucose incorporation for 24 h. e Isotopologue distribution of reduced glutathione (GSH) in ARI-resistant and WT LNCaP cells following ¹³C-glucose incorporation for 24 h. b–e Data are presented as mean values +/− SD. c–e *p-value of labelled fraction < 0.05 using a 1-way ANOVA with a Dunnett’s multiple comparisons test. DHAP dihydroxyacetone-phosphate, PEP phosphoenolpyruvate, GSH reduced glutathione, GSSG oxidised glutathione. Source data are provided as a Source data File.

Fig. 3. Lipid metabolism is strongly dysregulated in ARI-resistant cells.

a Labelled palmitate fraction derived from ¹³C-glucose (left panel) and relative isotopologue distribution of palmitic acid in ARI-resistant and WT LNCaP cells following ¹³C-glucose incubation for 72 h (right panel). b Western blot analysis of ACC, ACLY, and phospho-ACLY expression in WT and ARI-resistant LNCaP cells. HSC70 is used as a sample loading control. c Heatmap illustrating the steady-state levels of significantly regulated lipids in ARI-resistant cells when compared to WT LNCaP (FC > 1.5, p < 0.05 using a 1-way ANOVA with a Dunnett’s multiple comparisons test). Values are expressed as log(FC). a Data are presented as mean values +/− SD. a *p-value < 0.05 using a 1-way ANOVA with a Dunnett’s multiple comparisons test. Cer ceramide, CL cardiolipin, DG diacylglycerol, LysoPC lysophosphatidylcholine, PC phosphatidylcholine, PE phosphatidylethanolamine, PS phosphatidylserine, SM sphingomyelin, TG triglyceride. Source data are provided as a Source data File.

Fig. 4. Increased glucose metabolism in ARI-resistant cells depends on AR signalling.

a Expression of labelled isotopologue of selected metabolites in ARI-resistant and WT LNCaP cells silenced for AR and following 1 h ¹³C-glucose incorporation. b Western blot analysis of AR, ACC and ACLY expression in WT and ARI-resistant LNCaP cells following AR siRNA silencing. HSC70 is used as a sample loading control. a Data are presented as mean values +/− SD, *p-value of labelled fraction < 0.05 using a 2-way ANOVA with a Sidak’s multiple comparisons test. DHAP dihydroxyacetone-phosphate, PEP phosphoenolpyruvate. Source data are provided as a Source data File.

Fig. 5. DECR1 is a potential target for CRPC.

a Venn diagrams highlighting proteins commonly modulated (p-value < 0.05, FC > 1.5) in ARI-resistant cells cultured in 2D and 3D conditions. Upregulated proteins are on top; downregulated proteins are into brackets. b Western blot analysis of DECR1 expression in WT and ARI-resistant LNCaP cells. c Western blot analysis of DECR1 expression in AR⁺ prostate cancer cells following acute AR inhibition for 48 h. d Cell proliferation of C4-2 cells silenced for DECR1 expression and treated with enzalutamide (20 µM—48 h). Cell count is normalised to untreated control (siCTL). e Cell proliferation of LNCaP cells overexpressing DECR1 and treated with enzalutamide (20 µM—48 h). Cell count is normalised to untreated empty vector (EV OE). f Western blot analysis of DECR1 expression in hormone naïve (LNCaP, VCaP) and castration-resistant (LNCaP AI, VCaP CR) tumour orthografts. g Immunohistochemical staining (left) and quantification (right) of DECR1 expression in CRPC tissue samples. Data are represented as difference in histoscore between post-treatment and pre-treatment biopsies. Scale bar represents 100 µm. h Percentage of prostate cancer patients showing genomic (copy number gain or amplification) or mRNA alteration (z-score = 1.5) for DECR1 using the TCGA dataset. i Gene expression analysis of DECR1 in normal and tumoural prostate tissues according to the TCGA dataset (n = 498). Centre line corresponds to median of data, top and bottom of box correspond to 75th and 25th percentile, respectively. Whiskers extend to adjacent values (minimum and maximum data points not considered outliers). j Kaplan–Meier survival analysis of prostate cancer patients stratified according to DECR1 expression using the TCGA dataset. b, c, f HSC70 is used as a sample loading control. d, e Data are presented as mean values +/− SD. d, e *p-value < 0.05 using a 1-way ANOVA with a Dunnett’s multiple comparisons test. g statistical analysis was performed using a Wilcoxon matched-pair signed-rank test (n = 14). i statistical analysis was performed using a pairwise ANOVA. j statistical analysis was performed using a logrank test. Source data are provided as a Source data File.

Fig. 6. DECR1 loss alters lipid homeostasis and sensitises CRPC cells to ferroptosis.

a Western blot analysis of DECR1 expression in LNCaP and LNCaP AI cells. b Cell proliferation of DECR1 KO (knockout) and untargeted control (CTL) cells after 72 h. Cell count is normalised to initial number of cells at the start of the experiment. c Hierarchical clustering of significantly altered lipids in DECR1 KO cells when compared to CTL cells (FC > 1.2). Selected lipids were significantly altered in at least one of the two KO cells (p < 0.05 using two-sided Student’s t-test). Values are expressed as log(FC). d Steady-state levels of significantly altered ceramides (Cer) in DECR1 KO cells when compared to CTL cells (FC > 1.2). Selected lipids were significantly altered in at least one of the two KO cells (p < 0.05 using two-sided Student’s t-test). e Absolute concentration of total MUFAs (left panel) and total PUFAs (right panel) in CTL and DECR1 KO cells, quantified by GC-MS. f Absolute concentration of free MUFAs (left panel) and free PUFAs (right panel) in CTL and DECR1 KO cells, quantified by GC-MS. g Western blot analysis of BIP, DNAJC3, XBP1s and CHOP expression in DECR1 KO and CTL cells. h Western blot analysis of GPX4 expression in DECR1 KO and CTL cells. i Cell proliferation of DECR1 KO and CTL cells treated for 48 hours with RSL3 (10 µM). Cell count is normalised to initial number of cells. j Cell proliferation of DECR1 KO and CTL cells treated for 48 h with RSL3 (10 µM) and Trolox (20 µM) or Liproxstatin (50 nM). Cell count is normalised to initial number of cells. a, g, h HSC70 is used as a sample loading control. b, d, e, f, i, j Data are presented as mean values +/− SD. b, e, f, i, j *p-value < 0.05 using a 1-way ANOVA with a Dunnett’s multiple comparisons test. DG: diacylglycerol, PC phosphatidylcholine, PE phosphatidylethanolamine, PG phosphatidylglycerol, PI phosphatidylinositol, PS phosphatidylserine, SM sphingomyelin, TG triglyceride. Source data are provided as a Source data File.

Fig. 7. DECR1 KO affects lipid metabolism and decreases CRPC tumour growth.

a Representative pictures of LNCaP AI CTL (top image) or DECR1 KO (bottom image) tumour orthografts monitored by ultrasound imaging (left). Quantification of tumour volume using ultrasonography (right). b Representative pictures of hematoxylin/eosin staining on orthografts from LNCaP AI CTL (top) or DECR1 KO (bottom) cells. C = cancer cells, N = necrotic region. Quantification of the proportion of cancer cells (right). Scale bar represents 1000 µm. c Representative heatmaps on data from Raman spectroscopy using 532 nm excitation (intensity at frequency 2845 cm^-1/intensity at frequency 2935 cm^-1) on LNCaP AI CTL (top) and DECR1 KO (bottom) derived orthografts (left). Average Raman spectra of LNCaP AI CTL (grey) or DECR1 KO (magenta) orthografts (middle) and quantification of tumour lipid content using the 2845 cm⁻¹-peak (right panel). AU = arbitrary unit. d Heatmap illustrating the steady-state levels of significantly regulated lipids in DECR1-deficient tumours when compared to CTL tumours (FC > 1.5, p < 0.05 using a two-tailed Student’s t-test). Values are expressed as z-score. a–c Data are presented as mean values +/− SD. a–c *p-value < 0.05 using a two-tailed Mann–Whitney U-test. BMP Bis(monoacylglycero)phosphate, Cer ceramide, PC phosphatidylcholine, PE phosphatidylethanolamine, PG phosphatidylglycerol, PI phosphatidylinositol, PS phosphatidylserine, TG triglyceride. Source data are provided as a Source data File.