Gastrointestinal stromal tumor (GIST), the most common sarcoma of the gastro-intestinal tract, is characterized by ligand-independent, constitutively activating mutations in either KIT or platelet-derived growth factor receptor A (PDGFRA) receptor tyrosine kinases that drive tumor cell proliferation. Because of their dependence on these activating mutations, GISTs are a striking example of “oncogene addiction,” which has been exploited therapeutically using imatinib mesylate (Gleevec; Novartis Pharmaceuticals), a small molecule tyrosine kinase inhibitor used as first-line therapy against this tumor. Imatinib has revolutionized GIST treatment, with an approximately 80% response rate. However, GIST patients generally respond to treatment with quiescent/stable disease; complete regression is only seen in approximately 2% of patients. Due to clinically stable disease, patients require lifelong therapy with imatinib. Moreover, approximately 50% of such patients develop secondary resistance within two years of the initiation of imatinib therapy due to the occurrence of additional intra-allelic KIT mutations that abrogate imatinib binding. To more effectively treat GIST patients, it is critical to elucidate how GIST cells survive imatinib therapy, in order to develop therapeutic strategies that augment the action of KIT inhibition by imatinib. Accordingly, we have developed tissue culture models for imatinib-induced quiescence in GIST cells. Upon treatment with imatinib, multiple GIST cell lines demonstrate KIT-dependent cell cycle and translational arrest. However, akin to what is seen in vivo in GIST patients, apoptosis is both modest and variable, resulting in significant clonogenic survival in all cell lines. Furthermore, cell cycle and translational arrest are reversible upon removal of imatinib, further supporting the validity of these in vitro GIST models.
Since autophagy has been increasingly implicated as a resistance mechanism during cancer chemotherapy, we asked whether autophagy was induced as a survival pathway in response to imatinib therapy in GIST. Treatment with imatinib in vitro resulted in robust autophagic flux, as evidenced by LC3 turnover in the lysosome, as well as p62/SQSTM1 degradation. Furthermore, we performed immunohistochemistry for LC3 in formalin-fixed paraffin-embedded GIST samples from patients that had been treated with imatinib for either 3, 5 or 7 days. Although the number of samples evaluated was too small for meaningful statistical analysis, we were struck by the trend of increasing numbers of autophagosomes in biopsies from patients treated for longer times with imatinib. Based on sequencing analysis for KIT or PDGFRA mutational status in these patients, the induction of punctate LC3 positively correlated with predicted sensitivity to imatinib. Considerable evidence indicates that autophagy and apoptosis are intimately related; inhibition of autophagy increases apoptosis in both normal and cancer cells (Fig. 1). We similarly discovered that inhibiting autophagy using RNAi-mediated depletion of ATGs leads to increased apoptosis and decreased clonogenic survival of imatinib-treated GIST cells.
Cancer cells exhibiting autophagy are often sensitive to lysosomal inhibitors that block the terminal stages of autophagic proteolysis; such agents include commonly used anti-malarials such as chloroquine and quinacrine. We discovered that the pharmacological inhibition of autophagy using these lysosomotropic agents synergized with imatinib to induce GIST cell death over a wide range of physiologically relevant drug concentrations. In further support, RNAi depletion of LAMP-2 corroborated our results obtained with anti-malarials. Furthermore, to validate the in vitro results, autophagy inhibitors were combined with imatinib in vivo to treat a mouse GIST xenograft model; this combination regimen results in increased apoptosis within tumors, as measured by cleaved caspase-3 immunohistochemistry. Notably, in the case of imatinib plus quinacrine, partial tumor regression was observed after only 15 days of treatment.
To our knowledge, we are the first to utilize quinacrine to suppress autophagy in a preclinical model for cancer therapy. Our studies corroborated that quinacrine accumulates in lysosomes and robustly inhibits both baseline and starvation-induced LC3-II turnover, demonstrating that quinacrine functions similarly to other lysosomotropic agents with regard to its ability to inhibit autophagy at its later stages. More importantly, both in vitro and in vivo experiments demonstrated that quinacrine, when used in combination with imatinib, results in increased cell death in comparison to chloroquine, a more commonly utilized autophagy inhibitor in diverse therapeutic settings. Because quinacrine also inhibits NFκB and other cellular processes, we cannot exclude that the cytotoxic effects of quinacrine may be augmented by modulation of other cellular pathways in addition to its ability to inhibit autophagy. This possibility is currently under investigation in our laboratories.
Our cell-based models also revealed that imatinib-induced quiescence in GIST cultures is notable for low levels of tumor cell proliferation. Theoretically, this slow proliferation may facilitate the outgrowth of GIST cells with secondary KIT or PDGFRA mutations resistant to imatinib. Upon combining lysosomal inhibitors with imatinib therapy, we found significantly reduced outgrowth of imatinib-sensitive GIST cells in comparison with cultures treated with imatinib therapy alone. This broaches the exciting possibility that combining autophagy inhibition with imatinib may prevent the development of secondary resistance in GIST patients treated with imatinib; further studies are underway to test this hypothesis.
Overall, our results demonstrate that autophagy is induced in GIST cells in response to imatinib, facilitating the establishment of a quiescent state in which a subset of GIST cells can survive indefinitely (Fig. 1). Inhibition of autophagy synergizes with imatinib to increase GIST cell death and retard the outgrowth of residual viable GIST cells in comparison with imatinib therapy alone. Important unresolved issues include the identification of cellular triggers that activate autophagy in response to imatinib therapy; the elucidation of these proteins/pathways may lead to new therapeutic strategies to specifically inhibit the signals that stimulate autophagy downstream of this kinase inhibitor. Moreover, we have initiated studies to determine whether or not autophagy plays a role in the survival of GIST progenitor cells. Finally, we are exploring whether combination regimens of imatinib and anti-malarials can be moved into clinical trials to treat GIST patients, especially in consideration that quinacrine and chloroquine are both well-tolerated, inexpensive drugs with a long history of use in humans. Ultimately, these studies may improve clinical outcome in GIST patients currently undergoing chronic therapy with imatinib.