MRI elucidates differential drug mechanism of action for banoxantrone and atovaquone. Banoxantrone did not cause growth inhibition in (A) Calu6 xenografts or (B) U87 xenografts. Atovaquone did not cause growth inhibition in (C) Calu6 xenografts. D and E, Calu6 and U87 xenografts show reduction in volume of hypoxic tumor with banoxantrone by day 3. Similarly, (F) Calu6 xenografts show reduction in volume of hypoxic tumor with atovaquone by day 7. G and H, Calu6 and U87 xenografts show companion increases in volume of necrotic tumor, whereas (I) no change in volume of necrotic tumor is detected in Calu6 xenografts treated with atovaquone. J and K, No change in volume of normoxic tumor is detected in Calu6 or U87 xenografts treated with banoxantrone, whereas (L) Calu6 xenografts treated with atovaquone show increase in normoxic tumor. Changes described are all relative to baseline.
ARTICLE ABSTRACT
Oxygen-enhanced MRI (OE-MRI) has shown promise for quantifying and spatially mapping tumor hypoxia, either alone or in combination with perfusion imaging. Previous studies have validated the technique in mouse models and in patients with cancer. Here, we report the first evidence that OE-MRI can track change in tumor oxygenation induced by two drugs designed to modify hypoxia. Mechanism of action of banoxantrone and atovaquone were confirmed using in vitro experiments. Next, in vivo OE-MRI studies were performed in Calu6 and U87 xenograft tumor models, alongside fluorine-18–fluoroazomycin arabinoside PET and immunohistochemistry assays of hypoxia. Neither drug altered tumor size. Banoxantrone reduced OE-MRI hypoxic fraction in Calu6 tumors by 52.5% ± 12.0% (P = 0.008) and in U87 tumors by 29.0% ± 15.8% (P = 0.004) after 3 days treatment. Atovaquone reduced OE-MRI hypoxic fraction in Calu6 tumors by 53.4% ± 15.3% (P = 0.002) after 7 days therapy. PET and immunohistochemistry provided independent validation of the MRI findings. Finally, combined OE-MRI and perfusion imaging showed that hypoxic tissue was converted into necrotic tissue when treated by the hypoxia-activated cytotoxic prodrug banoxantrone, whereas hypoxic tissue became normoxic when treated by atovaquone, an inhibitor of mitochondrial complex III of the electron transport chain. OE-MRI detected and quantified hypoxia reduction induced by two hypoxia-modifying therapies and could distinguish between their differential mechanisms of action. These data support clinical translation of OE-MRI biomarkers in clinical trials of hypoxia-modifying agents to identify patients demonstrating biological response and to optimize treatment timing and scheduling.Significance: For the first time, we show that hypoxic fraction measured by oxygen-enhanced MRI (OE-MRI) detected changes in tumor oxygenation induced by two drugs designed specifically to modify hypoxia. Furthermore, when combined with perfusion imaging, OE-MRI hypoxic volume distinguished the two drug mechanisms of action. This imaging technology has potential to facilitate drug development, enrich clinical trial design, and accelerate clinical translation of novel therapeutics into clinical use.
