Fig 2: Upregulation of ketogenic/ketolytic enzymes in preclinical and clinical progression of PCa.A Schematic representation of ketogenesis/ketolysis pathway. The three ketone bodies are marked within a pink oval. Enzymes are depicted in black ovals and the other named metabolites in blue ovals. B Representative photomicrograph images of sections of MDA PCa 183 tumors growing in mice and (C) corresponding IHC quantification (Control [n = 5], ERC [n = 6] and Relapse [n = 3]). Magnification 100X. Data are represented as mean ± SD. One-way ANOVA followed by Tukey’s multiple comparisons test was used to assess statistical significance (P < 0.05). D Representative photomicrograph images of sections of samples of the corresponding human donor of MDA PCa 183 before ADT (treatment naive) and after progression (castration resistant) and (E) corresponding IHC quantification. Magnification 200X. Data are represented as mean ± SD. Two-tailed t-test was used to assess statistical significance (P < 0.05). Samples were stained with H&E and immunostained for ACAT1, OXCT1, and BDH1. ERC early response to castration, H&E Hematoxylin and eosin, ACAT1 acetyl-CoA acetyltransferase, OXCT1 3-oxoacid CoA-transferase 1, BDH1 3-Hydroxybutyrate Dehydrogenase 1, HMGCS2 hydroxy-methylglutaryl-CoA synthase 2, HMGCL hydroxy-methylglutaryl-CoA lyase.

Fig 3: Bioinformatics analysis of ketogenic/ketolytic enzymes and MCT transporters as risk predictors of clinical outcome in CRPC.A Kaplan–Meier (KM) curves for progression-free survival (PFS) in months for PCa patients with low (green) or high (red) ACAT1, OXCT1, BDH1, HMGCL, HMGCS2, SLC16A1 and SLC16A7 expression in the TCGA-PRAD dataset (n = 497). Log-rank test and Cox proportional hazard model regression were employed to assess statistical significance. B Multivariable Cox proportional hazard model regression analysis for ACAT1, OXCT1, BDH1 and HMGCL presented as forest plots for PFS. C Heatmap depicting low (blue) or high (red) ACAT1, OXCT1, BDH1 and HMGCL mRNA expression for PCa patients according to the TCGA-PRAD dataset. D KM curves for PFS in months for PCa patients subgroups with different expression levels of ACAT1, OXCT1, BDH1 and HMGCL in TCGA-PRAD: [1] low ACAT1, OXCT1, BDH1 and HMGCL expression (n = 64); [2] low ACAT1, low OXCT1, low BDH1 and high HGMCL expression (n = 221); [3] high BDH1 and HMGCL and low ACAT1 and OXCT1 expression (n = 32), [4] low ACAT1 and high OXCT1, BDH1 and HMGCL expression (n = 7), [5] high ACAT1 and BDH1 and low OXCT1 and HMGCL expression (n = 7); [6] high ACAT1, OXCT1, BDH1 and HMGCL expression (n = 4). The table indicates the number of patients assessed every 12 months. Log-rank test was employed to assess statistical significance. HR: hazard ratio. Statistical significance: P < 0.05.

Fig 4: The hippocampal and cortical levels of ketone bodies and enzymes involved in ketone turnover in the cells. The level of hippocampal (A,B, n = 5) and cortical (C,D, n = 5) β-hydroksybutyrate (BHB) and acetoacetate levels (AcAc). MCT1, BDH1, SCOT, HMGCS2, and HMGCL matched β-actin blots performed in the hippocampus (E) and cortex (G). Quantification of MCT1 levels in the hippocampus (F) and cortex (H). Quantification of BDH1 levels in the hippocampus (J) and cortex (L). Quantification of SCOT levels in the hippocampus (K) and cortex (M). Quantification of HMGCS2 level in the hippocampus (N) and cortex (P). Quantification of HMGCL level in the hippocampus (O) and cortex (Q). The data is presented as a mean ± standard deviation, ∗ indicates statistically significant changes (p < 0.05). SD marks a group of control animals fed with standard chow, while experimental mice were maintained on the ketogenic diet composed of fat of animal (KA) or plant (KP) origins.