Validating bladder cancer xenograft bioluminescence with magnetic resonance imaging requires correction for hypoperfusion, hypoxia and necrosis. Free

493

Purpose: Bioluminescent imaging allows longitudinal assessment of tumor burden in orthotopic tumor models. One limitation we have encountered in bladder xenografts is poor correlation between tumor size (weight or volume) and luminescence in several cell lines. We wished to confirm this discrepancy with magnetic resonance imaging (MRI). We hypothesized that the lack of correlation was due to hypoxia and necrosis in the tumors.

Materials and Methods: Orthotopic bladder tumors were established in 5 week old nude mice with 100,000 KU7 or 500,000 253J-BV cells (10 mice each) that had been previously transduced with a lentiviral luciferase construct. Bioluminescence (In Vivo Imaging System, Xenogen Corp., Almeda, CA) and MRI imaging (Bruker Biospin 4.7T scanner) were performed weekly within 24 hours of each other. Tumor volume was calculated from T1 and T2 weighted images at each time point and dynamic contrast studies were completed for the final two time points. The mice were sacrificed at four weeks 45 minutes after intraperitoneal injection of 60 mg/kg piminidazole. The tumors were excised, weighed and frozen whole. Step sections were obtained at 1 mm intervals. Hematoxylin and eosin staining and immunohistochemical analysis of piminidazole adduct formation were performed to assess hypoxia and necrosis. CD31 staining was used to quantitate vascularity. Relative areas of each section containing total tumor, hypoxic tumor and necrotic tumor were quantitated with ImagePro software (Media Cybernetics, Silver Spring, MD)

Results: KU7 xenograft bioluminescence showed minimal correlation to tumor volume on MRI during the first two weeks of imaging but none thereafter (R²<0.1) and also none to tumor weight. KU7 xenografts showed extensive hypoxic and necrotic areas lacking blood vessels, so that bioluminescence was found to correlate to the volume of perfused, non-hypoxic, non-necrotic tumor. The correlation between MRI volume and luminescence was stronger (R²=0.93 at 4 weeks) in 253J-BV xenografts, where hypoxia and necrosis were less prominent. MRI tumor volume tripled in size between the first and second week in KU7 but remained constant in 253J-BV.

Conclusion: The correlation between tumor bioluminescence and tumor size depends on the growth kinetics of the individual cell line used in xenograft studies. These parameters do not correlate in xenografts that rapidly outgrow their vascular supply and develop broad areas of hypoxia and necrosis. In these cases, however, the bioluminescence does yield information about the amount of viable tumor, and should therefore still be considered a useful parameter.