Date Published: June 12, 2019
Publisher: Public Library of Science
Author(s): Aviva Rabin-Court, Marcos R. Rodrigues, Xian-Man Zhang, Rachel J. Perry, Ming Tan.
Obesity is associated with increased incidence and worse prognosis of more than one dozen tumor types; however, the molecular mechanisms for this association remain under debate. We hypothesized that insulin, which is elevated in obesity-driven insulin resistance, would increase tumor glucose oxidation in obesity-associated tumors. To test this hypothesis, we applied and validated a stable isotope method to measure the ratio of pyruvate dehydrogenase flux to citrate synthase flux (VPDH/VCS, i.e. the percent of total mitochondrial oxidation fueled by glucose) in tumor cells. Using this method, we found that three tumor cell lines associated with obesity (colon cancer [MC38], breast cancer [4T1], and prostate cancer [TRAMP-C3] cells) increase VPDH/VCS in response to physiologic concentrations of insulin. In contrast, three tumor cell lines that are not associated with obesity (melanoma [YUMM1.7], B cell lymphoma [BCL1 clone 5B1b], and small cell lung cancer [NCI-H69] cells) exhibited no oxidative response to insulin. The observed increase in glucose oxidation in response to insulin correlated with a dose-dependent increase in cell division in obesity-associated tumor cell lines when grown in insulin, whereas no alteration in cell division was seen in tumor types not associated with obesity. These data reveal that a shift in substrate preference in the setting of physiologic insulin may comprise a metabolic signature of obesity-associated tumors that differs from that of those not associated with obesity.
Obesity is well-known to increase the prevalence and mortality of more than one dozen tumor types. In spite of the prevalence of obesity and its privileged place in public health discourse, the metabolic and molecular mechanisms underpinning the relationship between obesity and cancer remain contentious. Hyperinsulinemia has emerged as a focal point of research on obesity-related tumors, with increased plasma insulin concentrations independently predicting increased risk and mortality in prostate [1, 2], colon [3–7], breast [8–13], endometrial [12, 14, 15], and pancreatic cancer [16, 17], as well as several other tumor types. The idea that hyperinsulinemia may promote cancer risk is bolstered by the fact that biguanides such as metformin and phenformin, the most commonly prescribed class of diabetes drug worldwide, slow tumor growth associated with reductions in plasma insulin concentrations [18–30], although this class of agents has also shown efficacy in vivo independent of changes in plasma insulin concentrations in a minority of studies [31–33]. We recently showed that both metformin and a novel insulin sensitizer, a controlled-release mitochondrial protonophore, slows tumor growth in two models of colon cancer, and that the tumor-slowing effects of both agents were dependent on reversal of hyperinsulinemia , demonstrating a causative role for hyperinsulinemia in these mouse models.
Hyperinsulinemia has been discussed as a potential mediator of obesity-related cancer growth: activating mutations in the PI3K/Akt pathway are common (10–30% incidence) and confer a poorer prognosis in humans with colon cancer [54, 55], breast cancer [56, 57], and prostate cancer [58, 59]. Activating mutations in the PI3K/Akt pathway have also been observed–albeit at a lower incidence (2–9%)–in melanoma [60, 61] and small cell lung cancer [62, 63], although, to our knowledge, they have not been observed in B cell lymphoma [64, 65]. Mutations in this pathway do not divide as clearly along obesity-associated versus independent lines in the cell lines examined in this study: whereas mutations in this pathway have been described in 4T1  and NCI-H69 cells , they have not been described in the other obesity-associated or -independent cell lines studied here. Evidence for a critical role for insulin in directly promoting tumor cell division is provided by data showing that while insulin promotes tumor cell division in vivo [30, 68–71], pharmacologic agents which reverse hyperinsulinemia slow tumor growth unless exogenous insulin replacement is provided . However, the mechanisms undergirding insulin’s effect upon certain tumors have remained unclear, in part due to a lack of methods to assess tumor substrate preference and the impact of variations in hormones and substrates on tumor fuel preference. To that end, we adapted a stable isotope method which we have previously applied in vivo for tissue-specific measurements of glucose oxidation via pyruvate dehydrogenase relative to total citrate synthase flux [30, 35]. Validation studies demonstrate the sensitivity of this method to alterations in glucose and fatty acid oxidation, and further demonstrate that glucose oxidation is not maximized under these conditions; therefore the lack of an oxidative response to insulin in obesity-independent cell lines does not reflect an inherent limitation in mitochondrial glucose utilization in these cells (Fig 3D).