Implementing Patient-Derived Xenografts to Assess the Effectiveness of Cyclin-Dependent Kinase Inhibitors in Glioblastoma.

Glioblastoma (GBM) is the most common primary brain tumor with no available cure. As previously described, seliciclib, a first-generation cyclin-dependent kinase (CDK) inhibitor, down-regulates the anti-apoptotic protein, Mcl-1, in GBM, thereby sensitizing GBM cells to the apoptosis-inducing effects of the death receptor ligand, tumor necrosis factor-related apoptosis-inducing ligand (TRAIL). Here, we have assessed the efficacy of seliciclib when delivered in combination with the antibody against human death receptor 5, drozitumab, in clinically relevant patient-derived xenograft (PDX) models of GBM. A reduction in viability and significant levels of apoptosis were observed in vitro in human GBM neurospheres following treatment with seliciclib plus drozitumab. While the co-treatment strategy induced a similar effect in PDX models, the dosing regimen required to observe seliciclib-targeted responses in the brain, resulted in lethal toxicity in 45% of animals. Additional studies showed that the second-generation CDK inhibitor, CYC065, with improved potency in comparison to seliciclib, induced a significant decrease in the size of human GBM neurospheres in vitro and was well tolerated in vivo, upon administration at clinically relevant doses. This study highlights the continued need for robust pre-clinical assessment of promising treatment approaches using clinically relevant models.

Sensitization of glioblastoma cells to TRAIL-induced apoptosis by IAP- and Bcl-2 antagonism.

Due to the lack of effective treatments for glioblastoma (GBM), we here studied the responsiveness of GBM cell lines to the combination of death ligand, TRAIL and the IAP antagonist, TL32711 (Birinapant). Responses were highly heterogeneous, with synergistic apoptosis as well as treatment resistance observed. Caspase-8 and Bid, together with caspase-3, form a nonlinear signalling hub that efficiently induced apoptosis in responder cell lines. Cells resistant to TRAIL/TL32711 expressed low amounts of procaspase-8 and Bid and poorly activated caspase-3. We therefore hypothesised that improving caspase-8 activation or sensitising mitochondria to truncated Bid (tBid) could convert non-responder GBM cell lines to responders. Mathematical simulations of both strategies predicted mitochondrial sensitization to tBid would outperform enhancing caspase-8 activation. Indeed, antagonising Bcl-2 by ABT-199 allowed TRAIL/TL32711 response synergies to manifest in otherwise TRAIL resistant cell lines. These findings were further corroborated in experiments with a translationally relevant hexavalent TRAIL variant. Our study therefore demonstrates that a high caspase-8/Bid signature is associated with synergistic TRAIL/TL32711-induced apoptosis in GBM cells and outlines Bcl-2 antagonism as a highly potent intervention to sensitize highly TRAIL-resistant GBM cells to TRAIL/TL32711 combination treatment.

Simulating and predicting cellular and in vivo responses of colon cancer to combined treatment with chemotherapy and IAP antagonist Birinapant/TL32711.

Apoptosis resistance contributes to treatment failure in colorectal cancer (CRC). New treatments that reinstate apoptosis competency have potential to improve patient outcome but require predictive biomarkers to target them to responsive patient populations. Inhibitor of apoptosis proteins (IAPs) suppress apoptosis, contributing to drug resistance; IAP antagonists such as TL32711 have therefore been developed. We developed a systems biology approach for predicting response of CRC cells to chemotherapy and TL32711 combinations in vitro and in vivo. CRC cells responded poorly to TL32711 monotherapy in vitro; however, co-treatment with 5-fluorouracil (5-FU) and oxaliplatin enhanced TL32711-induced apoptosis. Notably, cells from genetically identical populations responded highly heterogeneously, with caspases being activated both upstream and downstream of mitochondrial outer membrane permeabilisation (MOMP). These data, combined with quantities of key apoptosis regulators were sufficient to replicate in vitro cell death profiles by mathematical modelling. In vivo, apoptosis protein expression was significantly altered, and mathematical modelling for these conditions predicted higher apoptosis resistance that could nevertheless be overcome by combination of chemotherapy and TL32711. Subsequent experimental observations agreed with these predictions, and the observed effects on tumour growth inhibition correlated robustly with apoptosis competency. We therefore obtained insights into intracellular signal transduction kinetics and their population-based heterogeneities for chemotherapy/TL32711 combinations and provide proof-of-concept that mathematical modelling of apoptosis competency can simulate and predict responsiveness in vivo. Being able to predict response to IAP antagonist-based treatments on the background of cell-to-cell heterogeneities in the future might assist in improving treatment stratification approaches for these emerging apoptosis-targeting agents.

Predicting the cell death responsiveness and sensitization of glioma cells to TRAIL and temozolomide.

Genotoxic chemotherapy with temozolomide (TMZ) is a mainstay of treatment for glioblastoma (GBM); however, at best, TMZ provides only modest survival benefit to a subset of patients. Recent insight into the heterogeneous nature of GBM suggests a more personalized approach to treatment may be necessary to overcome cancer drug resistance and improve patient care. These include novel therapies that can be used both alone and with TMZ to selectively reactivate apoptosis within malignant cells. For this approach to work, reliable molecular signatures that can accurately predict treatment responsiveness need to be identified first. Here, we describe the first proof-of-principle study that merges quantitative protein-based analysis of apoptosis signaling networks with data- and knowledge-driven mathematical systems modeling to predict treatment responsiveness of GBM cell lines to various apoptosis-inducing stimuli. These include monotherapies with TMZ and TRAIL, which activate the intrinsic and extrinsic apoptosis pathways, respectively, as well as combination therapies of TMZ+TRAIL. We also successfully employed this approach to predict whether individual GBM cell lines could be sensitized to TMZ or TRAIL via the selective targeting of Bcl-2/Bcl-xL proteins with ABT-737. Our findings suggest that systems biology-based approaches could assist in personalizing treatment decisions in GBM to optimize cell death induction.

Determining the contributions of caspase-2, caspase-8 and effector caspases to intracellular VDVADase activities during apoptosis initiation and execution.

Apoptosis signaling crucially depends on caspase activities. Caspase-2 shares features of both initiator and effector caspases. Opinions are divided on whether caspase-2 activity is established during apoptosis initiation or execution in response to DNA damage, death receptor stimulation, or heat shock. So far, approaches towards measuring caspase-2 activity were restricted to analyses in cell homogenates and extracts, yielded inconsistent results, and were often limited in sensitivity, thereby contributing to controversies surrounding the role of caspase-2 during apoptosis. Furthermore, caspases overlap in substrate specificities, and caspase-8 as well as effector caspases may cleave the optimal VDVAD recognition motif as well. We therefore generated a highly sensitive Förster resonance energy transfer (FRET) substrate to determine the relative contribution of these caspases to VDVADase activity non-invasively inside living cells. We observed limited proteolysis of the substrate during apoptosis initiation in response to death receptor stimulation by FasL, TNFα and TRAIL. However, this activity was attributable to caspase-8 rather than caspase-2. Likewise, no caspase-2-specific activity was detected during apoptosis initiation in response to genotoxic stress (cisplatin, 5-FU), microtubule destabilization (vincristine), or heat shock. The contribution of caspase-2 to proteolytic activities during apoptosis execution was insignificant. Since even residual, ectopically introduced caspase-2 activity could readily be detected inside living cells in our measurements, we conclude, in contrast to several previous studies, that caspase-2 activity does not contribute to apoptosis in the scenarios investigated, and that instead caspase-8 and effector caspases are the most significant VDVADases during canonical apoptosis signaling.