Fig 2: DECR2 confers resistance to enzalutamide in prostate cancer cells.a GSEA of peroxisomal Hallmark and KEGG proteins shows positive correlation with acquired resistance to enzalutamide. b DECR2 protein expression is significantly increased in LNCaP acquired apalutamide and enzalutamide resistance organoids and cell lines compared to wildtype LNCaP organoids and cells. Data are represented as violin plots in GraphPad prism. c The correlation of DECR2 expression with overall survival in the SU2C cohort. ARSI = androgen receptor signalling inhibitor. d Cell viability of 22Rv1, V16D and MR49F cells subjected to siRNA-mediated DECR2 knockdown, treated with ENZ (1 or 10 µM). e Cell viability of 22Rv1, V16D and MR49F prostate cancer cell lines treated with thioridazine, TDZ (2.5, 5 and 7.5 µM) and/or ENZ (1 or 10 µM). f 22Rv1, V16D and MR49F cell growth in 3D spheres, treated with TDZ (2.5 and 5 µM) and/or ENZ (1 or 10 µM). Spheroid volumes were determined after four days of culturing the cells in 96-well microplates; spheres were assessed using the ReViSP software. Scale bar, 100 µm. g Growth of hDECR2 and hControl LNCaP cells under charcoal-stripped (androgen-deprived) conditions (DCC). h Growth of hDECR2 and hControl LNCaP cells under full serum conditions and in response to enzalutamide (ENZ, 10 µM) and apalutamide (APA, 10 µM). The same controls were used in both graphs in (h). All data are representative of at least two independent experiments and presented as mean ± s.e.m of triplicate wells. Data in (c) were statistically analysed using a two-sided log-rank test. Statistical analysis was performed using ordinary one-way or two-way ANOVA. ns = non-significant, *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001.

Fig 3: DECR2 knockdown inhibits prostate cancer cell growth in vitro and in vivo.a Cell viability of androgen-dependent LNCaP, castrate-resistant 22Rv1 and V16D, and enzalutamide-resistant MR49F prostate cancer cell lines subjected to siRNA-mediated DECR2 knockdown. b Cell viability of LNCaP cells with stable/inducible shRNA DECR2 knockdown (shDECR2) and (c) stable overexpression of DECR2 (hDECR2). d LNCaP colony formation was evaluated in cells with stable/inducible shRNA DECR2 knockdown (shDECR2) or (e) stable DECR2 overexpression (hDECR2). f LNCaP stable DECR2 overexpression (hDECR2) cell lines were assessed for cell migration using a transwell migration assay. Scale bar, 100 µm. g LNCaP cells with stable/inducible shRNA knockdown of DECR2 (shDECR2+dox, n = 11) or control (shDECR2-dox n = 10) were analysed for orthotopic LNCaP tumour growth in mice, representative bioluminescent tumour images (right). h Ki67 quantification (left) and representative IHC staining (right) of orthotopic LNCaP tumours. Scale bar, 100 µm. This panel includes data from mice with sufficient sized tumours for analysis (shDECR2+dox n = 5, shDECR2-dox n = 8). Scale bar, 50 µm. i Tumour growth and tumour weight of intraprostatically injected LNCaP cells with stable DECR2 overexpression (hDECR2, n = 6) or control (hControl, n = 10), representative bioluminescent tumour images (right). j Lung luminescence readings of stable DECR2 overexpression tumours in mice. All in vitro data are representative of at least two independent experiments and presented as mean ± s.e.m of triplicate wells. Statistical analysis was performed using ordinary one-way ANOVA or two-tailed student’s t-test: *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001.

Fig 4: Global lipidomics of DECR2 knockdown and overexpression in prostate cancer cells reveal strongly altered lipid states.a Global lipidomic study overview. (b) Lipidomic analysis of LNCaP, V16D and MR49F prostate cancer cell lines subjected to siRNA-mediated DECR2 knockdown relative to control. Values are shown as quantitative log2-fold changes. Lipidomics data were from 6 replicates and are represented as means. Each dot represents a lipid species. Dot size is proportionate to statistical significance (see Supplementary Data 2). c Quantitative abundance per saturation group in LNCaP, V16D and MR49F cells. d Lipidomic analysis of DECR2 overexpression (hDECR2) cells relative to control (hControl). e Quantitative abundance. f Relative abundance per saturation group in DECR2 overexpression cells. g Lipid ontology (LION) enrichment analysis of relative lipid abundance in siControl versus siDECR2 LNCaP, V16D and MR49F prostate cancer cells. Statistical analysis was performed using two-tailed student’s t test: *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001.

Fig 5: DECR2 is overexpressed in prostate cancer.a Illustration of fatty acid oxidation in the peroxisome and mitochondria. Thioridazine (TDZ) is an inhibitor of perFAO. b Heatmap of peroxisomal β-oxidation (perFAO) gene expression in Taylor and Grasso cohorts. We manually curated a list of perFAO genes based on Gene Ontology pathway. Cell viability of (c) castrate-resistant C42B, 22Rv1 and V16D, and enzalutamide-resistant MR49F prostate cancer cell lines across a range of TDZ doses. d C42B and 22Rv1 prostate cancer cell lines treated with 2.5 µM TDZ were assessed for cell migration using transwell migration assay. Scale bar: left, 100 µm; right, 200 µm. e Immunostaining for proliferative marker Ki67 in vehicle (VEH) or TDZ-treated (20 μM) patient-derived explants (PDEs). Immunohistochemical staining and quantification of the proliferative marker Ki67 is shown (n = 11). Scale bar, 50 µm. f DECR2 expression with respect to tumour progression in four independent datasets. DECR2 levels were analysed in normal, primary, and metastatic castrate-resistant or neuroendocrine tissue samples. g The association of DECR2 expression and disease-free survival in the MSKCC (Taylor) cohort. h DECR2 protein expression in non-malignant prostate cell lines (PNT1 and PNT2) and prostate cancer cell lines (LNCaP, VCaP, C42B, 22Rv1, V16D, PC3), including enzalutamide-resistant prostate cancer cell line (MR49F). i Left: Representative DECR2 IHC staining of benign prostate tissues and prostate cancer tissues. Scale bar, 50 µm. Middle: DECR2 protein expression in a validation cohort consisting of benign prostate tissues (n = 3) and prostate cancer tissues (n = 10). Right: Intra-tissue IHC analysis of DECR2 expression in prostate cancer tissues (n = 10). All cell line data are representative of at least two independent experiments and presented as mean ± s.e.m of triplicate wells. Statistical analysis was performed using ordinary one-way ANOVA or two-tailed student’s t test. Data in (e) were statistically analysed using paired t-test. Data in (g) were statistically analysed using a two-sided log-rank test. *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001.