**3. Discussion**

This study reports on the connection between CPT1A overexpression (OE) and the metabolic and genetic consequences of its increased activity. Particularly, we found that CPT1A OE cells produced a significant number of acyl-carnitines that promote growth and resistance to stress insults, like excess lipids and androgen withdrawal. These later conditions are characteristic of advanced PCa, where androgen deprivation and excess circulating lipids frequently exists [40].

At the molecular level, this work provides insights into the metabolic changes precipitated by the excess use of lipid oxidation in the mitochondria, particularly the upregulation of the overall lipid catabolic process. As expected, not only CPT1A was increased, but also the ability to hydrolyze

lipid stores via PNPLA2 (triglyceride lipase or ATGL) to increase the supply of fatty acids for the mitochondria (Figures 1B and 2D). It is possible that new inhibitors for ATGL could be of therapeutic value for cancers with increased CPT1A activity [41]. Another molecular aspect of the increased CPT1A activity was the re-wiring of metabolism towards the serine biosynthesis pathway, which has been recently shown to be important in supporting mitochondrial function [30], and a driver in NEPC [42]. These unexpected amino acid metabolism changes are likely promoted by the strong activation of the adaptive ER Stress response, as indicated by the transcriptional upregulation of genes linked to the ER stress response, Figure 2. In fact, ATF4 is a key transcription factor translationally induced upon activation of the unfolded protein response or UPR [43]. This induction triggers an anti-stress response that promotes adaptation, or in the case of a chronic unresolved stress, it can promote cell death. Increased ATF4 regulates serine and glycine metabolism genes to drive de novo serine and glycine production, which can be used for antioxidant defense and glutathione production [44]. In the CPT1A-OE cells, the mitochondrial SHMT2 gene was significantly increased in OE versus KD cells, suggesting a strong induction of serine/glycine synthesis in the mitochondria. This is likely to provide antioxidant defense via the mitochondrial folate cycle that can generate NADPH [45], and carbon units to produce cystathionine, an intermediate in the synthesis of cysteine and ultimately glutathione (Figure 3).

The ER stress observed in the OE cells compared to the KD cells is likely a response to the high ROS production in the mitochondria and low SOD2 expression. This is a known mechanism to promote cancer growth [46]. This was an unexpected result considering the OE cells did not show signs of distressed or fragmented mitochondria. A possible explanation is the upregulation of the glutathione homeostasis pathway. We found that the glutathione-degrading and ER response gene CHAC1 (cation transport regulator homolog1) was increased in OE cells. This gene was discovered in a co-regulated group of genes enriched for components of the ATF4 pathway, including CCAAT/enhancer-binding protein beta (CEBPB), which also binds to the CHAC1 promoter [47,48]. All this evidence would suggest that increased expression of a glutathione-degrading enzyme and activation of the ER stress pathway will lead to cancer cell death. However, we did not observe such changes in the OE cells as the levels of glutathione did not change and cells were able to grow in the presence of exogenously added lipids (Figures 1 and 4). Other studies in breast and ovarian cancer have shown that CHAC1 expression correlates with tumor differentiation and survival [49], suggesting that the observed ER stress in our models is likely stress-resolving and can promote disease progression. In fact, when we challenged the cells with commonly circulating fatty acids in the absence of androgens (a stressful environment), more SOD2 expression was observed in the OE cells and less on the KD cells. This underscores the increased antioxidant response capabilities of the OE cells and potential for survival and growth. This environment may promote adaptation of the OE cells to androgen deprivation, supporting progression to lethal disease. Recent studies have also shown that increased SOD2 activity can protect prostate cells when exposed to radiation [50].

How oxidative changes in the mitochondria connect with ER stress remains unknown. ER and oxidative stress have overlapping and intertwined functions in cancer [21]. Both promote epithelial mesenchymal transition, a key step of metastasis and tissue invasion of many tumor cells. In addition, detachment from the extracellular matrix activates the ATF4-HSPA5 branch of the UPR, which protects from anoikis by stimulating both autophagy and antioxidative stress responses [51]. As the CPT1A-OE cells prefer to grow in suspension [10], they might be using the ER stress response and the mitochondrial oxidative environment to transform to more aggressive tumors [52].

This study and the public databases provide evidence that lipid catabolism driven by CPT1A is associated with more aggressive disease. CPT1A-OE cells showed more SYP and ENO2 neuroendocrine marker expression compared to the KD cells. This suggests that in CRPC tumors, CPT1A activity can rewire metabolism to promote growth and transformation via activation of serine biosynthesis, folate cycle, and glutathione homeostasis, all geared to maintain an adequate redox balance in the

cancer cells. The role of mitochondrial ROS in activating these pro-tumor antioxidant pathways warrants further investigation.
