Summary Although autophagy controls cell death and survival, underlying mechanisms are poorly understood and it is unknown if autophagy just affects whether or not cells die or also controls other aspects of programmed cell death. MAP3K7 is a tumor suppressor gene associated with poor disease-free survival in prostate cancer. Here we report that Map3k7 deletion in mouse prostate cells sensitizes to cell death by TNF-related apoptosis inducing ligand (TRAIL). Surprisingly, this death occurs primarily through necroptosis, not apoptosis, due to assembly of the necrosome in association with the autophagy machinery, mediated by p62/SQSTM1 recruitment of RIPK1. The mechanism of cell death switches to apoptosis if p62-dependent recruitment of the necrosome to the autophagy machinery is blocked. These data show that the autophagy machinery can control the mechanism of programmed cell death by serving as a scaffold rather than by degrading cargo.

Kinase inhibitors are effective cancer therapies, but tumors frequently develop resistance. Current strategies to circumvent resistance target the same or parallel pathways. We report here that targeting a completely different process, autophagy, can overcome multiple BRAF inhibitor resistance mechanisms in brain tumors. BRAFV600Emutations occur in many pediatric brain tumors. We previously reported that these tumors are autophagy-dependent and a patient was successfully treated with the autophagy inhibitor chloroquine after failure of the BRAFV600E inhibitor vemurafenib, suggesting autophagy inhibition overcame the kinase inhibitor resistance. We tested this hypothesis in vemurafenib-resistant brain tumors. Genetic and pharmacological autophagy inhibition overcame molecularly distinct resistance mechanisms, inhibited tumor cell growth, and increased cell death. Patients with resistance had favorable clinical responses when chloroquine was added to vemurafenib. This provides a fundamentally different strategy to circumvent multiple mechanisms of kinase inhibitor resistance that could be rapidly tested in clinical trials in patients with BRAFV600E brain tumors.DOI: http://dx.doi.org/10.7554/eLife.19671.001

Autophagy is a multistage process. Progress within the field has led to the development of agents targeting both early (initiation) and late (fusion) stages of this process. The specific stage of autophagy targeted may influence cancer treatment outcomes. We have previously shown that central nervous system (CNS) tumors with the BRAFV600E mutation are autophagy dependent, and late-stage autophagy inhibition improves the response to targeted BRAF inhibitors (BRAFi) in sensitive and resistant cells. Drugs directed toward initiation of autophagy have been shown to reduce tumor cell death in some cancers, but have not been assessed in CNS tumors. We investigated early-stage inhibition for autophagy-dependent CNS tumors. BRAFi-sensitive and resistant AM38 and MAF794 cell lines were evaluated for the response to pharmacologic and genetic inhibition of ULK1 and VPS34, two crucial subunits of the autophagy initiation complexes. Changes in autophagy were monitored by western blot and flow cytometry. Survival was evaluated in short- and long-term growth assays. Tumor cells exhibited a reduced autophagic flux with pharmacologic and genetic inhibition of ULK1 or VPS34. Pharmacologic inhibition reduced cell survival in a dose-dependent manner for both targets. Genetic inhibition reduced cell survival and confirmed that it was an autophagy-specific effect. Pharmacologic and genetic inhibition were also synergistic with BRAFi, irrespective of RAFi sensitivity. Inhibition of ULK1 and VPS34 are potentially viable clinical targets in autophagy-dependent CNS tumors. Further evaluation is needed to determine if early-stage autophagy inhibition is equal to late-stage inhibition to determine the optimal clinical target for patients.