UCSF

Unlabelled

Reduction and oxidation (redox) chemistry is increasingly implicated in cancer pathogenesis. To interrogate the redox status of prostate tumors noninvasively, we developed hyperpolarized [1-(13)C]dehydroascorbate ((13)C-DHA), the oxidized form of vitamin C, as an MR probe. In a model of transgenic adenocarcinoma of the mouse prostate (TRAMP), increased reduction of hyperpolarized (13)C-DHA to vitamin C was observed in tumor, as compared with normal prostate and surrounding benign tissue. We hypothesized that this difference was due to higher concentrations of glutathione and increased transport of hyperpolarized (13)C-DHA via the glucose transporters (GLUT1, GLUT3, and GLUT4) in TRAMP tumor. To test these hypotheses, hyperpolarized (13)C-DHA MR spectroscopy (MRS) and (18)F-FDG PET were applied as complementary technologies in the TRAMP model.

Methods

Late-stage TRAMP tumors (>4 cm(3)) were studied at similar time points (MR studies conducted < 24 h after PET) in fasting mice by (18)F-FDG PET and hyperpolarized (13)C-DHA MR imaging on a small-animal PET/CT scanner and a (1)H/(3)C 3-T MR scanner. PET data were processed using open-source AMIDE software to compare the standardized uptake values of tumor with those of surrounding muscle, and (13)C-DHA MRS data were processed using custom software to compare the metabolite ratios (vitamin C/[vitamin C + (13)C-DHA]). After in vivo studies, the tumor glutathione concentrations were determined using a spectrophotometric assay, and thiol staining was performed using mercury orange. Real-time polymerase chain reaction was used to evaluate the relevant transporters GLUT1, GLUT3, and GLUT4 and vitamin C transporters SVCT1 and SVCT2. GLUT1 was also evaluated by immunohistochemistry.

Results

The average metabolite ratio was 0.28 ± 0.02 in TRAMP tumor, versus 0.11 ± 0.02 in surrounding benign tissue (n = 4), representing a 2.5-fold difference. The corresponding tumor-to-nontumor (18)F-FDG uptake ratio was 3.0. The total glutathione was 5.1 ± 0.4 mM in tumor and 1.0 ± 0.2 mM in normal prostate, whereas reduced glutathione was 2.0 ± 0.3 mM and 0.8 ± 0.3 mM, respectively, corresponding to a 2.5-fold difference. In TRAMP tumor, mercury orange staining demonstrated increased thiols. Real-time polymerase chain reaction showed no significant difference in GLUT1 messenger RNA between TRAMP tumor and normal prostate, with immunohistochemistry (anti-GLUT1) also showing comparable staining.

Conclusion

Both hyperpolarized (13)C-DHA and (18)F-FDG provide similar tumor contrast in the TRAMP model. Our findings suggest that the mechanism of in vivo hyperpolarized (13)C-DHA reduction and the resulting tumor contrast correlates most strongly with glutathione concentration. In the TRAMP model, GLUT1 is not significantly upregulated and is unlikely to account for the contrast obtained using hyperpolarized (13)C-DHA or (18)F-FDG.