Fats and Mets, KRAS-Driven Lipid Dysregulation Affects Metastatic Potential in Pancreatic Cancer

Pancreatic cancer is the third leading cause of death in the United States and is projected to become the second most deadly malignancy by 2030. Development of metastases is the most lethal aspect of the cancer, with 52% of patients presenting with de novo metastatic disease, having dismal 5-year survival of less than 3% in this setting (1). Obesity is an identified risk factor, with a 10% excess risk of developing pancreatic cancer for every 5 kg/m² unit increase in body mass index (BMI; ref. 2) and a 1.49 RR of dying from pancreatic cancer in patients with BMI >35 compared with normal weight (3).

Acknowledging that obesity is a contributing risk factor, Rozeveld and colleagues hypothesized that dysregulation of lipolysis may be a novel mechanism by which pancreatic tumors fuel metastasis (4). Intriguingly, in other cancers such as breast cancer, metastatic phenotypes have previously been associated with increased lipid droplet accumulation (5). Lipolysis is a process leading to the breakdown of triacylglycerols stored in lipid droplets and release of fatty acids and glycerol primarily to provide ATP/energy. The key sequential steps in intracellular lipolysis include action of adipose triglyceride lipase (ATGL), which catalyzes the first and rate-limiting step, followed by the actions of hormone-sensitive lipase (HSL) on diacylglycerol and monoacylglycerol lipase (MGL) on monoacylglycerol. Rozeveld and colleagues demonstrated that excess fatty acids stored as lipid droplets are required for cancer cell migration and invasion independently of growth and proliferation, likely by providing a source of ATP. Subsequently, the authors discovered that KRAS downregulates HSL in tumor cells, hence increasing intracellular lipid droplet accumulation. KRAS–HSL axis disruption resulted in lower numbers of lipid droplets, reprogramming of the invasive cell metabolic phenotype, and ultimately inhibited in vitro tumor cell migration/invasion as well as in vivo metastases formation in a syngeneic KPC mouse model. While KRAS downregulated HSL and induced steady-state cells to rely on glycolysis for proliferation, there was a transient switch to oxidative metabolism as measured by optical redox ratio analyses with use of stored lipids to enable invasion and metastasis (4).

These novel results raise a number of broad-reaching mechanistic questions relating to tumor cell energetics. Importantly, what precipitates the conversion of a tumor cell from a noninvasive to an invasive phenotype and how does the availability of a lipid fuel source precipitate this change? Tumor cell extrinsic or intrinsic mechanisms may be considered as hypotheses to both these critical questions. Microenvironmental cues such as hypoxia, cellular interactions, as well as intratumoral nutritional limitations, and epigenomic alterations could all contribute to these conditions.

To elaborate, hypoxia is well documented in pancreatic cancer, as shown by pimonidazole-labeling studies, and contributes to the malignant phenotype of the disease. Thereafter, activation of hypoxia-inducible factor-1α is a prominent driver of the Warburg effect and has been shown to suppress αKG dehydrogenase, a key mitochondrial enzyme complex, which shifts metabolism from the tricarboxylic acid cycles toward isocitrate dehydrogenase–mediated fatty acid synthesis (6). In addition, hypoxia supports epithelial-to-mesenchymal transition (EMT), which is associated with invasion and metastasis. Although Rozeveld and colleagues did not find significant alterations in EMT markers during in vitro lipolysis manipulation in this study, intravital imaging studies have demonstrated that pancreatic tumors have dynamic hypoxia gradients (7), suggesting that the context and depth of hypoxia in relation to the functional measures of invasion may be more relevant in this setting.

Another extrinsic mechanism may relate to the cellular microenvironment. In pancreatic cancer, this is characterized by a dense desmoplastic stroma and dynamic cellular elements whose role in tumor progression and resistance to therapies are well established. Pancreatic stellate cells have been shown to undergo a lipid metabolic shift when differentiating into an activated cancer-associated fibroblast (CAF). CAFs are characterized by a loss of intracellular lipid droplets and secretion of lysophosphatidylcholine, a lipid component subsequently metabolized to lysophosphatidic acid, which has been shown to promote cancer cell migration and pancreatic cancer progression (8).

Intrinsically, pancreatic cancer cells rely on a number of metabolic processes to provide their cellular energy including glycolysis, glutaminolysis, and lipolysis. Carbon source availability (glucose vs. fatty acid vs. glutamine) is dependent on the microenvironment (9). In the presence of metabolic stressors, such as avascularity (causing glucose deprivation), cancer cells may augment their dependency on fatty acid β oxidation stochastically, to maintain homeostasis and, thereafter, precipitate invasion and metastasis.

Finally, increasing fatty acid levels in cells also exerts profound effects on the pancreatic tumor epigenome. Epigenetic modification induced by acetyl-CoA (which is produced in the mitochondria from fatty acid oxidation) has been shown to create a histone profile characterized by transcriptional upregulation of stem-like prosurvival and metastatic genes including FOXA1, GATA5, and Sox2 in pancreatic cancer (10).

The clinical implications of this study are intriguing. For example, it would be ineffective to inhibit fatty acid utilization alone when the tumor cell is in the steady state. Conversely, it would also likely be futile to inhibit fatty acid storage when the tumor cell is already in the invasive phenotype. Future clinical application of this work will likely need to be combinatorial as simply preventing the appearance of new metastases while existing metastases continue to grow will not be beneficial. In addition, the timing, depth, and method of lipolysis manipulation will need to be optimized, for example, through direct (e.g., HSL) or indirect (e.g., MEK and macroautophagy) inhibition routes. While attractive experimentally and with preclinical proof-of-principal data in this article, etomoxir was withdrawn from clinical development in the early 1990s due to hepatotoxicity.

In conclusion, Rozeveld and colleagues have established the importance of tumor cell energetics by highlighting that the ability to store and mobilize lipids are critical for cancer cell invasion and metastases. Importantly, they have linked this to novel aspects of canonical oncogenic changes in pancreatic cancer pathophysiology, the KRAS–HSL axis in modulating lipolysis in pancreatic cancer. This study provides an important benchmark for further research to investigate the larger question of how tumor energy demands relate to phenotype and their implications for clinical care.

No potential conflicts of interest were disclosed.