Comparison of 18F-fluoroazomycin-arabinofuranoside and 64Cu-diacetyl-bis(N4-methylthiosemicarbazone) in preclinical models of cancer

Hypoxic regions are present in different types of cancer and are a negative prognostic factor for disease progression and response to therapy. ¹⁸F-fluoroazomycin-arabinofuranoside (¹⁸F-FAZA) and ⁶⁴Cu-diacetyl-bis(N4-methylthiosemicarbazone) (⁶⁴Cu-ATSM) have been widely used to visualize hypoxic regions in preclinical and clinical studies. Although both these radioligands have high signalto- noise ratios, ⁶⁴Cu-ATSM may be suitable for use in in vivo imaging and as a radiotherapeutic agent. Despite encouraging results suggesting that it may have a role as a prognostic tracer, ⁶⁴Cu-ATSM was recently shown to display cell line-dependent kinetics of oxygen-dependent uptake. We set out to evaluate the kinetics of ⁶⁴Cu-ATSM distribution in different cancer models, using ¹⁸F-FAZA as the gold standard. Methods: ¹⁸F-FAZA and ⁶⁴Cu-ATSM uptake were compared ex vivo using dual-tracer autoradiography and in vivo using PET in different xenograft mouse models (FaDu, EMT-6, and PC-3). ¹⁸F-FAZA uptake was compared with ⁶⁴Cu-ATSM uptake in PET studies acquired at early (2 h after injection) and delayed time points (24 h after injection). To evaluate the presence of hypoxia and copper pumps, the tumors from animals submitted to PET were harvested and analyzed by an immunohistochemical technique, using antibodies against carbonic anhydrase IX (CAIX) and copper pumps (Ctr1 and ATP7B). Results: ⁶⁴Cu-ATSM showed a higher tumor-to-muscle ratio than did ¹⁸F-FAZA. In the FaDu mouse model, radioactivity distribution profiles were overlapping irrespective of the hypoxic agent injected or the time of ⁶⁴Cu acquisition. Conversely, in the EMT-6 and PC-3 models there was little similarity between the early and delayed ⁶⁴Cu-ATSM images, and both the radiotracers showed a heterogeneous distribution. The microscopic analysis revealed that ¹⁸F-FAZA-positive areas were also positive for CAIX immunostaining whereas immunolocalization for copper pumps in the 3 models was not related to radioactivity distribution. Conclusion: The results of this study confirm the celldependent distribution and retention kinetics of ⁶⁴Cu-ATSM and underline the need for proper validation of animal models and PET acquisition protocols before exploration of any new clinical applications.