Abstract Body:

Introduction: Glioblastoma is the most common primary brain tumor. Despite evolving understanding of tumor biology, treatment resistance has limited any meaningful survival benefit in the last decade, supporting the need for early biomarkers of treatment response. The noninvasive interrogation of brain tumor metabolism represents an ideal such opportunity, potentially annotating metabolic pathways of response. Given the combined role of glucose metabolism and oxidative stress in response to radiation treatment, in this work we explore the utility of co-hyperpolarized [1-13C]pyruvate and [1-13C] dehydroascorbate (DHA)1-4 to simultaneously evaluate glycolytic metabolism and oxidative stress in orthotopically implanted brain tumor bearing mice. Methods: To assess sensitivity to radiation, U251 and U87 glioblastoma cell lines were plated in triplicate and treated with 0 Gy, 5 Gy, or 10 Gy radiation in one dose. For in vivo tumor experiments, female athymic Foxn1nu mice (n=14) underwent stereotactic orthotopic injection with 5×105 U87 glioblastoma cells (n=8) or 3×105 U251 cells (n=6). Once tumors were detectable on MRI, half of in each group underwent whole brain radiation with 8Gy in one dose. Each mouse was injected via tail vein over 10s with 150uL (U251) or 300μL (U87) of co-hyperpolarized 40 mM [1-13C]DHA and 100 mM [1-13C]pyruvate. MR experiments were performed using a 3T MRI system (Bruker) equipped with a quadrature double-tuned 1H/13C volume coil. The MRI protocol included a T2-weighted sequence for anatomical reference and 13C Chemical Shift Imaging (CSI) sequence (32x32mm2 field-of-view, 5mm thickness) run 25s after substrate injection. When post-contrast imaging was pursued, mice underwent intraperitoneal injection with 1mmol/kg gadolinium and T1-weighted imaging within 60 minutes. Spectroscopic data were processed using a custom Matlab script and analyzed in SIVIC5. Normalization for polarization and concentration were performed and means were compared using a two-tailed t-test. Results: U251 demonstrate increased radiosensitivity that U87 tumor cells after treatment with 10 Gy, resulting in 27%(± 9%, p=0.02) decrement in surviving fraction at 48 hours compared to only 6% in U87s (Fig.A).  Both U87 and U251 orthotopic tumors are hyperintense on T2-weighted MRI, though U251 are better visualized after contrast administration (Fig.B). CSI reveals differential lactate and vitamin C generation over the mouse brain (Fig.C). When comparing means of fractional concentrations of metabolites generated from their respective substrates, U251s produce more lactate (0.48±0.07, p=0.001) and vitamin C (0.15±0.07, p=0.01) than U87s (Fig.D). After treatment with radiation, the radioresistant U87s demonstrate no significant change in either product. However, radiation decreases both lactate and vitamin C generation in the more radiosensitive U251s (0.28±0.07, p=0.02 and 0.087±0.03, p=0.04, respectively). Given that a saturating concentration of both probes is infused, we explored the ability to directly compare the metabolic products of both probes by taking the ratio of lactate generated to vitamin C. As shown in Fig.E, the ratio of lactate / vitamin C recapitulates these findings: treatment with radiation produces no significant difference in U87, but a significant decrement in U251s (-0.96 ± 0.29, p=0.02). Further, the baseline lactate/vitamin C ratio in U251s is significantly higher than that of U87 (1.55±0.44, p=0.02). Conclusion: These findings demonstrate the sensitivity of this novel technique to simultaneously examine metabolism and oxidative stress of mouse models of brain tumors. Further, the addition of hyperpolarized [1-13C] DHA to interrogate redox metabolism reveals unique signatures between tumors of varying radiosensitvities. Decrement in antioxidants such as vitamin c and glutathione may explain a tumors’ sensitivity to an oxidative load such as radiation. This data suggests measurement of lactate to vitamin C ratio has potential as a biomarker for radiosensitivity and response to treatment, and work is ongoing to optimize image acquisition and spatial resolution. Survival data for treated mice is also forthcoming.

Author