BIMG-05. TO BE OR NOT TO BE GLYCOLYTIC: DEUTERATED GLUCOSE-BASED ASSESSMENT OF THE WARBURG EFFECT ALLOWS NON-INVASIVE IMAGING OF TUMOR BURDEN AND TREATMENT RESPONSE IN MUTANT IDH GLIOMAS IN VIVO

The Warburg effect, characterized by elevated glucose uptake and flux to lactate, is a metabolic hallmark of cancer. Recent studies have identified deuterium (2)H-magnetic resonance spectroscopy (MRS) using 6,6’-(2)H-glucose as a novel method of imaging the Warburg effect in high-grade primary glioblastomas (GBMs). However, its utility for imaging low-grade gliomas has not been tested. The goal of this study was to determine whether 6,6’-(2)H-glucose can be used for imaging tumor burden and treatment response in mutant isocitrate dehydrogenase (IDHmut) low-grade gliomas in vivo. We examined mice bearing orthotopic tumors of the patient-derived BT257 astrocytoma model. (1)H-MRS, providing a readout of steady-state metabolite levels, confirmed the presence of 2-hydroxyglutarate, the product of IDHmut, in BT257 tumor tissue but not normal brain. Previous studies comparing IDHmut gliomas with GBMs suggest that IDHmut gliomas undergo lactate dehydrogenase silencing, potentially leading to a non-glycolytic phenotype. Nevertheless, our results indicated that, compared to normal brain, glucose uptake and concomitant flux to lactate were significantly higher in BT257 tumor tissue. Importantly, 6,6’-(2)H-glucose metabolism to lactate was observed in BT257 tumor-bearing mice, but not tumor-free mice. Furthermore, imaging studies confirmed spatial localization of lactate production to the tumor vs. contralateral normal brain. We then examined the ability of 6,6’-(2)H-glucose to assess treatment response. Poly-(adenosine 5′-diphosphate-ribose) polymerase inhibitors (PARPi) inhibit IDHmut glioma growth and are in clinical trials for IDHmut glioma patients. Treatment with the PARPi niraparib reduced 6,6’-(2)H-glucose flux to lactate in BT257 tumor-bearing mice. Importantly, this reduction was observed at early time-points when no difference in tumor volume could be detected using anatomical imaging, pointing to the ability of 6,6’-(2)H-glucose to assess pseudoprogression. Collectively, our results suggest that IDHmut gliomas display a glycolytic phenotype amenable to non-invasive (2)H-MRS-based imaging of tumor burden and treatment response.