Abstract A long standing problem in anti-cancer drug development has been the limited value of preclinical mouse tumor models to reliably predict subsequent clinical activity. All too often highly encouraging preclinical results in mice are followed by complete failure in clinical trials, especially at the randomized phase III level. There are many possible reasons that have been postulated for this preclinical/clinical discrepancy. One which we have been studying for the past decade is the failure to use mouse models which duplicate the challenging circumstance of treating advanced (established) visceral metastatic disease after primary tumors have been surgically resected. Instead, preclinical treatment of established primary tumors or low volume (micro)metastatic disease, often confined to the lungs, have been the historical preclinical model norms. To address this problem we have developed several models of postsurgical advanced metastatic disease involving human tumor xenografts grown in SCID mice, including breast, colorectal or kidney cancer, and melanoma (1). This approach has been extended more recently for postsurgical adjuvant therapy of early stage microscopic metastatic disease (2,3). The established cell lines used for in vivo studies are variants previously selected in vivo for more aggressive spontaneous metastatic capability, which then are sometimes stably tagged with luciferase to permit whole body bioluminescent imaging. Using these models to evaluate the impact of several anti-cancer treatments, consisting mostly of antiangiogenic drugs and/or chemotherapy, either standard maximum tolerated dose or low-dose ‘metronomic’, has highlighted the critical contribution of the extent of metastatic disease to differential therapeutic outcomes, and the prospect of better correlation with clinical outcomes. For example, primary orthotopic tumors, e.g. breast cancer in the mammary fat pad, respond well to treatment with an antiangiogenic drug such as sunitinib, pazopanib or anti-VEGFR-2 antibodies whereas mice with advanced metastases in sites such as the liver or lungs do not, e.g. no prolongation of survival is observed (4). Adding chemotherapy, e.g. paclitaxel to sunitinib did not change the results, whereas adding DC101, an anti-VEGFR-2 antibody, to the same chemotherapy regimen did result in a modest survival improvement (thus mimicking the metastatic breast cancer E2100 phase III results of bevacizumab plus paclitaxel chemotherapy) (4). The observed lack of sunitinib efficacy alone or with chemotherapy when treating mice with advanced metastases mimics three failed phase III clinical trial results of this drug with or without chemotherapy in metastatic breast cancer patients (5). In addition, we have noted that successful treatment of mice with advanced systemic metastatic breast, melanoma or renal cell cancer, e.g. with low-dose metronomic chemotherapy plus an antiangiogenic drug such that overall survival is meaningfully prolonged, sometimes results in the emergence of overt spontaneous brain metastases(1,6,7). Thus, in mice, the brain appears to be a protective sanctuary for the survival and progressive growth of microscopic into macroscopic metastases, as already well known in the clinic. More recent studies have indicated how the brain microenvironment can contribute to the development of melanoma metastases in this organ environment e.g. the interaction of endothelins (ETs) with (elevated) endothelin receptor B expression by the brain melanoma metastatic variants (8). In summary, models of postsurgical advanced metastatic disease to undertake experimental therapeutic studies appear to have a greater degree of clinical relevance compared to most conventional primary tumor therapy models. We are now extending this approach to the development of postsurgical models of early stage microscopic metastatic disease to mimic adjuvant therapy in the clinic (2,9,10); some of our previous (2009) results indicated that adjuvant antiangiogenic therapy may actually worsen eventual survival outcomes of mice with early stage disease (2), a finding for which there is now some preliminary clinical support based on a recent phase III clinical trial assessing treatment of postsurgical early stage colorectal cancer patients with bevacizumab plus chemotherapy (11,12).
Read full abstract
Aug 22, 2022
Hsiang-Lin Tsai + 10
Open Access
