Successful development of targeted anti-cancer drugs is often coupled with predictive assays that enable selective treatment of patients more likely to benefit from therapy. Immunohistochemistry is often used to assess expression of drug targets, but it suffers from subjectivity and lack of quantitative measures. We developed a method for automated, quantitative analysis (AQUA) for assessing protein levels in situ. In response to PA-08-134, we propose to expand AQUA to simultaneously assess tumors and endothelial cells, and develop models to predict clinical benefit from adjuvant sorafenib and sunitinib for renal cell carcinoma (RCC), as well as models to predict prognosis in untreated patients. RCC has traditionally been a disease that is highly resistant to systemic therapy. However, multiple targeted therapies, including sorafenib and sunitinib, have recently revolutionized the approach to metastatic RCC. Both drugs are effective for subsets of RCC patients, and both are associated with some toxicity. Given their success in metastatic RCC, these drugs are being studied as adjuvant therapies in a large, randomized, double blinded, multi-center trial called E2805. Specimens are being collected on all patients, and they offer a unique opportunity to develop models to predict clinical benefit from these drugs and models to predict prognosis in untreated patients in a multi-center clinical trial setting on a very large cohort. Sorafenib and sunitinb have multiple targets, and our purpose is to identify the most important predictive marker/s. We will study angiogenic markers, members of the MAPK pathway and other known targets. In preliminary studies using AQUA, we showed that RCCs with high levels of vascular endothelial growth factor (VEGF) receptors in tumor cells tend to have lower microvessel density and poor survival. We hypothesize that patients with high VEGF receptor expression in TUMOR cells are more likely to benefit from therapy than those with high microvessel density. We will expand AQUA to enable concurrent assessment of targets in tumor and vessels, by masking the tumor and vessels with different fluorophores. We will establish staining conditions for all known targets of sorafenib and sunitinib and select mediators of angiogenesis in tumor, endothelium and adjacent normal tissue using historical cohorts of untreated RCC patients. We will then assess VHL mutations and expression of sorafenib and sunitinib targets in a training set (67%) of E2805 patients and generate predictive models for each of the drugs, to be validated in a testing set. Standard clinical co-variates and VHL mutational status will be incorporated into the model. We will also use these molecular markers and clinical co-variates to improve current prognostic models in the placebo-treated patients. These models can be used to select patients for the optimal adjuvant therapy for RCC (sorafenib, sunitinib or neither), and this approach can be studied in other clinical settings as well.
In some cancers, targeted therapies (drugs that specifically inhibit certain key proteins in the cancer cell) have dramatically impacted management of the disease, and their success has usually been coupled with identification of the most important drug target/s and selective treatment of those patients whose tumors express the target. We propose to develop predictive models for sorafenib and sunitinib (new drugs that are being studied to decrease development of metastases for kidney cancer), which will enable us to selectively treat those patients that are more likely to derive benefit from these drugs, and spare the rest of the patients the toxicity and cost associated with this therapy. We will use a newly developed method of automated analysis of target levels from biopsy specimens, and in the future we will be able to apply this technology to other diseases and other targeted therapies.
Showing the most recent 10 out of 20 publications
