Despite their apparent success in pre-clinical trials, metalloproteinase (MMP) inhibitors proved to be inefficacious in clinical settings. In an effort to understand the underlying causes of this unanticipated outcome, we modeled the consequences of long-term MMP inhibition by removing one of the major players in tumorigenesis, MMP9, in two complimentary mouse models of pancreatic neuroendocrine carcinogenesis: Myc;BclXl and RIP1-Tag2. By employing gel zymography and a fluoregenic solution assay, we first established that MMP9 is expressed and activated in Myc;BclXl tumors in an interleukin-1b-dependent manner. The genetic deletion of MMP9 in Myc;BclXl mice impairs tumor angiogenesis and growth analogous to its absence in the RIP1-Tag2 model. Notably, tumors that developed in the context of MMP9-deficient backgrounds in both models were markedly more invasive than their typical wild-type counterparts, and expressed elevated levels of pro-invasive cysteine cathepsin B. The increased invasion of MMP9-deficient tumors was associated with a switch in the spectrum of inflammatory cells at the tumor margins, involving homing of previously undetected, cathepsin-B expressing CD11b;Gr1-positive cells to the invasive fronts. Thus, plasticity in the tumor inflammatory compartment is partially responsible for changes in the expression pattern of tumorassociated proteases, and may contribute to the compensatory effects observed on MMP inhibition, hence accounting for the heightened tumor progression described in late stage clinical trials.Oncogene advance online publication, 5 March 2012; doi:10.1038/onc.2012.60Keywords: tumor; metalloproteinase; invasion; cathepsin; Myc; interleukin-1b
INTRODUCTIONMatrix metalloproteinases (MMPs) are a family of zinc-dependent proteinases that contribute to development and tissue homeostasis, and modulate cancer progression by digesting a variety of substrates. Owing to the role of MMP in extracellular matrix (ECM) re-arrangement (also referred to as remodeling or degradation), and in the promotion of neoangiogenesis, it was anticipated that MMP inhibition in vivo would have an anti-tumorigenic effect. However, phase III clinical trials with an MMP inhibitor (MMPI), combined with standard-of-care chemotherapy, failed to demonstrate clinical benefit for late stage cancer patients and, remarkably, was in some cases worse than the chemotherapyonly control arm. For instance, the use of Bay 12-9566 in treating patients with pancreatic cancer reduced both median progressionfree survival and quality-of-life when compared with standard care gemcitabine.1 In another case, treatment of glioblastoma patients rendered no statistically significant survival benefits.2 Additionally, the clinical trials in small-cell lung carcinomas had to be terminated prematurely.3 This failure led to the early cessation of ongoing trials, and to reconsideration of the merit of a variety of MMPIs in clinical development by a number of pharmaceutical companies. 4 --6 The molecular mechanism behind this disappointing outc...