*4.9. Public Database Analysis*

RMA normalized prostate cancer microarray data was downloaded from GEO (GSE32967, [37]). Only the first replicate for each sample was used for the analysis to deal with uneven sample replicates. The most variable probe was selected to represent each gene. The limma R package was used to compare small cell carcinoma (*n* = 4) to adenocarcinoma (*n* = 3) prostate cancer. TPM normalized gene expression and clinical data from The Cancer Genome Atlas PRAD dataset (Firehose Legacy) [61], was downloaded from cBioPortal [62]. Gene Set Variation Analysis (GSVA) was performed using the GSVA R package [63], to score several gene sets with a Poison distribution. Patients were classified into high and low pathway groups by median score splitting. The survival (https://cran.r-project.org/web/packages/survival/citation. html) and survminer R packages (https://rpkgs.datanovia.com/survminer/index.html) were used to generate Kaplan-Meier curves and perform log rank tests. For the CPT1A and AR gene expression correlations, Log2 normalized whole transcript mRNA expression values from Taylor et al. [38] prostate cancer samples were downloaded from cBioPortal [62]. Expression of AR and CPT1A between primary and metastatic samples were compared using a Wilcoxon rank sum test. Pearson correlation was calculated for the correlation between AR and CPT1A expression in either primary or metastatic tumors.

#### **5. Conclusions**

The overall goal of this study is to understand the lipid metabolic underpinnings of advanced prostate cancer so that metabolic therapies can be designed effectively. Overall, we provide evidence that CPT1A activity may have a relevant role in advanced PCa, including transformation to NEPC. Etomoxir is a potent inhibitor of CPT1 that has been used in clinical trials in Europe [64], but it is not currently being used in the USA. Considering all the toxic chemotherapeutic agents used in cancer treatments, the potential for drugs like etomoxir to impact cancer growth and drug response warrants investigation.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2072-6694/12/11/3431/s1, Figure S1: Colony growth after treatment with different chain fatty acids, Figure S2: Enrichment score plots of the pathways and Gene ontology (GO) pathways shown in Figure 2, Figure S3: Metabolomic enrichment analysis, Figure S4: Uncropped western blots of SOD2 expression and densitometry graphs, Figure S5: cBioportal database correlations of CPT1A and AR expression, Figure S6: Additional progression-free survival plots, Table S1: RNA seq analysis results from comparing the OE cells to the KD cells.

**Author Contributions:** Conceptualization, I.R.S.; Formal analysis, J.K., K.B., E.N.-G., A.G., J.C.C. and I.R.S.; Funding acquisition, I.R.S. and J.C.C.; Investigation, H.E. and A.G.; Methodology, A.D., E.M., K.B., and H.E.; Software, A.G. and J.C.C.; Supervision, I.R.S.; Writing–original draft, I.R.S.; Writing–review & editing, M.J., A.D., A.G., E.N.-G., J.C.C., and I.R.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was supported by the American Cancer Society (129846-RSG-16-256), NIH Bioinformatics and Biostatistics Shared Resource Core, Genomics Shared Resource, Functional Genomics Shared Resource and Cancer Center Support Grant (P30CA046934), NIH U01, (CA231978), Cancer league of Colorado (AWD-193286), Nutrition and Obesity Research Center (NORC).

**Acknowledgments:** We are grateful from the technical support from Maren Salzmann-Sullivan, Gergana Stoykova, and Juliana Oviedo. We are also thankful to Kristofer Fritz for the advice on SOD2 antibodies.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.
