Applying a GAN-based classifier to improve transcriptome-based prognostication in breast cancer.

Established prognostic tests based on limited numbers of transcripts can identify high-risk breast cancer patients, yet are approved only for individuals presenting with specific clinical features or disease characteristics. Deep learning algorithms could hold potential for stratifying patient cohorts based on full transcriptome data, yet the development of robust classifiers is hampered by the number of variables in omics datasets typically far exceeding the number of patients. To overcome this hurdle, we propose a classifier based on a data augmentation pipeline consisting of a Wasserstein generative adversarial network (GAN) with gradient penalty and an embedded auxiliary classifier to obtain a trained GAN discriminator (T-GAN-D). Applied to 1244 patients of the METABRIC breast cancer cohort, this classifier outperformed established breast cancer biomarkers in separating low- from high-risk patients (disease specific death, progression or relapse within 10 years from initial diagnosis). Importantly, the T-GAN-D also performed across independent, merged transcriptome datasets (METABRIC and TCGA-BRCA cohorts), and merging data improved overall patient stratification. In conclusion, the reiterative GAN-based training process allowed generating a robust classifier capable of stratifying low- vs high-risk patients based on full transcriptome data and across independent and heterogeneous breast cancer cohorts.

Melanoma 2.0. Skin cancer as a paradigm for emerging diagnostic technologies, computational modelling and artificial intelligence.

We live in an unprecedented time in oncology. We have accumulated samples and cases in cohorts larger and more complex than ever before. New technologies are available for quantifying solid or liquid samples at the molecular level. At the same time, we are now equipped with the computational power necessary to handle this enormous amount of quantitative data. Computational models are widely used helping us to substantiate and interpret data. Under the label of systems and precision medicine, we are putting all these developments together to improve and personalize the therapy of cancer. In this review, we use melanoma as a paradigm to present the successful application of these technologies but also to discuss possible future developments in patient care linked to them. Melanoma is a paradigmatic case for disruptive improvements in therapies, with a considerable number of metastatic melanoma patients benefiting from novel therapies. Nevertheless, a large proportion of patients does not respond to therapy or suffers from adverse events. Melanoma is an ideal case study to deploy advanced technologies not only due to the medical need but also to some intrinsic features of melanoma as a disease and the skin as an organ. From the perspective of data acquisition, the skin is the ideal organ due to its accessibility and suitability for many kinds of advanced imaging techniques. We put special emphasis on the necessity of computational strategies to integrate multiple sources of quantitative data describing the tumour at different scales and levels.

The apoptosome molecular timer synergises with XIAP to suppress apoptosis execution and contributes to prognosticating survival in colorectal cancer.

The execution phase of apoptosis is a critical process in programmed cell death in response to a multitude of cellular stresses. A crucial component of this pathway is the apoptosome, a platform for the activation of pro-caspase 9 (PC9). Recent findings have shown that autocleavage of PC9 to Caspase 9 (C9) p35/p12 not only permits XIAP-mediated C9 inhibition but also temporally shuts down apoptosome activity, forming a molecular timer. In order to delineate the combined contributions of XIAP and the apoptosome molecular timer to apoptosis execution we utilised a systems modelling approach. We demonstrate that cooperative recruitment of PC9 to the apoptosome, based on existing PC9-apoptosome interaction data, is important for efficient formation of PC9 homodimers, autocatalytic cleavage and dual regulation by XIAP and the molecular timer across biologically relevant PC9 and APAF1 concentrations. Screening physiologically relevant concentration ranges of apoptotic proteins, we discovered that the molecular timer can prevent apoptosis execution in specific scenarios after complete or partial mitochondrial outer membrane permeabilisation (MOMP). Furthermore, its ability to prevent apoptosis is intricately tied to a synergistic combination with XIAP. Finally, we demonstrate that simulations of these processes are prognostic of survival in stage III colorectal cancer and that the molecular timer may promote apoptosis resistance in a subset of patients. Based on our findings, we postulate that the physiological function of the molecular timer is to aid XIAP in the shutdown of caspase-mediated apoptosis execution. This shutdown potentially facilitates switching to pro-inflammatory caspase-independent responses subsequent to Bax/Bak pore formation.

The FLAME-accelerated signalling tool (FaST) for facile parallelisation of flexible agent-based models of cell signalling.

Agent-based modelling is particularly adept at modelling complex features of cell signalling pathways, where heterogeneity, stochastic and spatial effects are important, thus increasing our understanding of decision processes in biology in such scenarios. However, agent-based modelling often is computationally prohibitive to implement. Parallel computing, either on central processing units (CPUs) or graphical processing units (GPUs), can provide a means to improve computational feasibility of agent-based applications but generally requires specialist coding knowledge and extensive optimisation. In this paper, we address these challenges through the development and implementation of the FLAME-accelerated signalling tool (FaST), a software that permits easy creation and parallelisation of agent-based models of cell signalling, on CPUs or GPUs. FaST incorporates validated new agent-based methods, for accurate modelling of reaction kinetics and, as proof of concept, successfully converted an ordinary differential equation (ODE) model of apoptosis execution into an agent-based model. We finally parallelised this model through FaST on CPUs and GPUs resulting in an increase in performance of 5.8× (16 CPUs) and 53.9×, respectively. The FaST takes advantage of the communicating X-machine approach used by FLAME and FLAME GPU to allow easy alteration or addition of functionality to parallel applications, but still includes inherent parallelisation optimisation. The FaST, therefore, represents a new and innovative tool to easily create and parallelise bespoke, robust, agent-based models of cell signalling.

Convergence of pathway analysis and pattern recognition predicts sensitization to latest generation TRAIL therapeutics by IAP antagonism.

Second generation TRAIL-based therapeutics, combined with sensitising co-treatments, have recently entered clinical trials. However, reliable response predictors for optimal patient selection are not yet available. Here, we demonstrate that a novel and translationally relevant hexavalent TRAIL receptor agonist, IZI1551, in combination with Birinapant, a clinically tested IAP antagonist, efficiently induces cell death in various melanoma models, and that responsiveness can be predicted by combining pathway analysis, data-driven modelling and pattern recognition. Across a panel of 16 melanoma cell lines, responsiveness to IZI1551/Birinapant was heterogeneous, with complete resistance and pronounced synergies observed. Expression patterns of TRAIL pathway regulators allowed us to develop a combinatorial marker that predicts potent cell killing with high accuracy. IZI1551/Birinapant responsiveness could be predicted not only for cell lines, but also for 3D tumour cell spheroids and for cells directly isolated from patient melanoma metastases (80-100% prediction accuracies). Mathematical parameter reduction identified 11 proteins crucial to ensure prediction accuracy, with x-linked inhibitor of apoptosis protein (XIAP) and procaspase-3 scoring highest, and Bcl-2 family members strongly represented. Applied to expression data of a cohort of n = 365 metastatic melanoma patients in a proof of concept in silico trial, the predictor suggested that IZI1551/Birinapant responsiveness could be expected for up to 30% of patient tumours. Overall, response frequencies in melanoma models were very encouraging, and the capability to predict melanoma sensitivity to combinations of latest generation TRAIL-based therapeutics and IAP antagonists can address the need for patient selection strategies in clinical trials based on these novel drugs.

Low expression of pro-apoptotic proteins Bax, Bak and Smac indicates prolonged progression-free survival in chemotherapy-treated metastatic melanoma.

Despite the introduction of novel targeted therapies, chemotherapy still remains the primary treatment for metastatic melanoma in poorly funded healthcare environments or in case of disease relapse, with no reliable molecular markers for progression-free survival (PFS) available. As chemotherapy primarily eliminates cancer cells by apoptosis, we here evaluated if the expression of key apoptosis regulators (Bax, Bak, Bcl-2, Bcl-xL, Smac, Procaspase-9, Apaf-1, Procaspase-3 and XIAP) allows prognosticating PFS in stage III/IV melanoma patients. Following antibody validation, marker expression was determined by automated and manual scoring of immunohistochemically stained tissue microarrays (TMAs) constructed from treatment-naive metastatic melanoma biopsies. Interestingly and counter-intuitively, low expression of the pro-apoptotic proteins Bax, Bak and Smac indicated better prognosis (log-rank p < 0.0001, p = 0.0301 and p = 0.0227 for automated and p = 0.0422, p = 0.0410 and p = 0.0073 for manual scoring). These findings were independently validated in the cancer genome atlas (TCGA) metastatic melanoma cohort (TCGA-SKCM) at transcript level (log-rank p = 0.0004, p = 0.0104 and p = 0.0377). Taking expression heterogeneity between the markers in individual tumour samples into account allowed defining combinatorial Bax, Bak, Smac signatures that were associated with significantly increased PFS (p = 0.0002 and p = 0.0028 at protein and transcript level, respectively). Furthermore, combined low expression of Bax, Bak and Smac allowed predicting prolonged PFS (> 12 months) on a case-by-case basis (area under the receiver operating characteristic curve (ROC AUC) = 0.79). Taken together, our results therefore suggest that Bax, Bak and Smac jointly define a signature with potential clinical utility in chemotherapy-treated metastatic melanoma.

TAK1 suppresses RIPK1-dependent cell death and is associated with disease progression in melanoma.

Melanoma cells are highly resistant to conventional genotoxic agents, and BRAFV600/MEK-targeted therapies as well as immunotherapies frequently remain inefficient. Alternative means to treat melanoma, in particular through the induction of programmed cell death modalities such as apoptosis or necroptosis, therefore still need to be explored. Here, we report that melanoma cell lines expressing notable amounts of RIPK1, RIPK3 and MLKL, the key players of necroptosis signal transduction, fail to execute necroptotic cell death. Interestingly, the activity of transforming growth factor ß-activated kinase 1 (TAK1) appears to prevent RIPK1 from contributing to cell death induction, since TAK1 inhibition by (5Z)-7-Oxozeaenol, deletion of MAP3K7 or the expression of inactive TAK1 were sufficient to sensitize melanoma cells to RIPK1-dependent cell death in response to TNFα or TRAIL based combination treatments. However, cell death was executed exclusively by apoptosis, even when RIPK3 expression was high. In addition, TAK1 inhibitor (5Z)-7-Oxozeaenol suppressed intrinsic or treatment-induced pro-survival signaling as well as the secretion of cytokines and soluble factors associated with melanoma disease progression. Correspondingly, elevated expression of TAK1 correlates with reduced disease free survival in patients diagnosed with primary melanoma. Overall, our results therefore demonstrate that TAK1 suppresses the susceptibility to RIPK1-dependent cell death and that high expression of TAK1 indicates an increased risk for disease progression in melanoma.

Phosphoprotein patterns predict trametinib responsiveness and optimal trametinib sensitisation strategies in melanoma.

Malignant melanoma is a highly aggressive form of skin cancer responsible for the majority of skin cancer-related deaths. Recent insight into the heterogeneous nature of melanoma suggests more personalised treatments may be necessary to overcome drug resistance and improve patient care. To this end, reliable molecular signatures that can accurately predict treatment responsiveness need to be identified. In this study, we applied multiplex phosphoproteomic profiling across a panel of 24 melanoma cell lines with different disease-relevant mutations, to predict responsiveness to MEK inhibitor trametinib. Supported by multivariate statistical analysis and multidimensional pattern recognition algorithms, the responsiveness of individual cell lines to trametinib could be predicted with high accuracy (83% correct predictions), independent of mutation status. We also successfully employed this approach to case specifically predict whether individual melanoma cell lines could be sensitised to trametinib. Our predictions identified that combining MEK inhibition with selective targeting of c-JUN and/or FAK, using siRNA-based depletion or pharmacological inhibitors, sensitised resistant cell lines and significantly enhanced treatment efficacy. Our study indicates that multiplex proteomic analyses coupled with pattern recognition approaches could assist in personalising trametinib-based treatment decisions in the future.

Novel strategies to mimic transmembrane tumor necrosis factor-dependent activation of tumor necrosis factor receptor 2.

Tumor necrosis factor receptor 2 (TNFR2) is known to mediate immune suppression and tissue regeneration. Interestingly, the transmembrane form of tumor necrosis factor (tmTNF) is necessary to robustly activate TNFR2. To characterize the stoichiometry and composition of tmTNF during TNFR2 activation, we constructed differently oligomerized single chain TNF ligands (scTNF) comprised of three TNF homology domain (THD) protomers that mimic tmTNF. Using a variety of cellular and in vivo assays, we can show that higher oligomerization of the scTNF trimers results in more efficient TNF/TNFR2 clustering and subsequent signal transduction. Importantly, the three-dimensional orientation of the scTNF trimers impacts the bioactivity of the oligomerized scTNF ligands. Our data unravel the organization of tmTNF-mimetic scTNF ligands capable of robustly activating TNFR2 and introduce novel TNFR2 agonists that hold promise as therapeutics to treat a variety of diseases.

MToolBox: a highly automated pipeline for heteroplasmy annotation and prioritization analysis of human mitochondrial variants in high-throughput sequencing.

MOTIVATION: The increasing availability of mitochondria-targeted and off-target sequencing data in whole-exome and whole-genome sequencing studies (WXS and WGS) has risen the demand of effective pipelines to accurately measure heteroplasmy and to easily recognize the most functionally important mitochondrial variants among a huge number of candidates. To this purpose, we developed MToolBox, a highly automated pipeline to reconstruct and analyze human mitochondrial DNA from high-throughput sequencing data. RESULTS: MToolBox implements an effective computational strategy for mitochondrial genomes assembling and haplogroup assignment also including a prioritization analysis of detected variants. MToolBox provides a Variant Call Format file featuring, for the first time, allele-specific heteroplasmy and annotation files with prioritized variants. MToolBox was tested on simulated samples and applied on 1000 Genomes WXS datasets. AVAILABILITY AND IMPLEMENTATION: MToolBox package is available at https://sourceforge.net/projects/mtoolbox/.