Recent data has provided compelling evidence that hypoxic tumors (cervix, head and neck, soft tissue sarcoma) are resistant to the most widely used forms of therapy. In addition, hypoxic tumors are prone to metastasis. Thus, the early detection of tumor hypoxia is likely to play an important role in choosing the most appropriate therapies. A non-invasive assay for tissue hypoxia would be applicable to all tumors, many of which cannot be accessed by invasive methods, to aid in the selection the most appropriate therapy on an individual patient basis. The proposed work involves the detailed testing of two 2-nitroimidazoles, EF1 and EF5, in rodent tumor models. Such compounds are metabolically reduced forming covalent intracellular adducts at a rate which is maximal in severe hypoxia and decreases, with first order inhibition kinetics, as the oxygen concentration increases. The bound adducts can then be detected in various ways as a surrogate for the measurement of tissue hypoxia. Binding of EF5 has been thoroughly characterized using immunohistochemical techniques and has been shown to predict individual tumor response to radiation in rodents. EF5 is FDA approved for use in humans and is currently in phase I clinical trials using such methods. Results from this trial have shown no EF5-related toxicity, and anectdotal information regarding the ability of EF5 to detect prognosis-related hypoxia are encouraging. EF5, and its much more hydrophilic analog, EF1, have recently been labelled with 18F, allowing drug and adduct detection by positron emission tomography (PET). Thus, we will be able to compare invasive measurement of hypoxia, at very high resolution using immunohistochemistry, with non-invasive measurements (PET) at lower resolution. Our planned studies will allow several key aspects of non-invasive hypoxia imaging to be tested in detail: these include the effects of drug concentration, drug polarity and optimization of imaging algorithms and resolution. The last of these will greatly benefit from a new high-resolution PET camera under current development at the University of Pennsylvania. Our overall hypothesis is that assessment of hypoxia using appropriate PET imaging agents will predict radiation response in individual tumors. It should also be noted that the detection of hypoxia has wide-ranging applications in other pathologic conditions including stroke, heart attack, and wound healing.
