Oncogene (2015) 34, 3215–3225 © 2015 Macmillan Publishers Limited All rights reserved 0950-9232/15 www.nature.com/onc REVIEW Cancer systems biology: embracing complexity to develop better anticancer therapeutic strategies W Du and O Elemento The transformation of normal cells into cancer cells and maintenance of the malignant state and phenotypes are associated with genetic and epigenetic deregulations, altered cellular signaling responses and aberrant interactions with the microenvironment. These alterations are constantly evolving as tumor cells face changing selective pressures induced by the cells themselves, the microenvironment and drug treatments. Tumors are also complex ecosystems where different, sometime heterogeneous, subclonal tumor populations and a variety of nontumor cells coexist in a constantly evolving manner. The interactions between molecules and between cells that arise as a result of these alterations and ecosystems are even more complex. The cancer research community is increasingly embracing this complexity and adopting a combination of systems biology methods and integrated analyses to understand and predictively model the activity of cancer cells. Systems biology approaches are helping to understand the mechanisms of tumor progression and design more effective cancer therapies. These approaches work in tandem with rapid technological advancements that enable data acquisition on a broader scale, with finer accuracy, higher dimensionality and higher throughput than ever. Using such data, computational and mathematical models help identify key deregulated functions and processes, establish predictive biomarkers and optimize therapeutic strategies. Moving forward, implementing patient-specific computational and mathematical models of cancer will significantly improve the specificity and efficacy of targeted therapy, and will accelerate the adoption of personalized and precision cancer medicine. Oncogene (2015) 34, 3215–3225; doi:10.1038/onc.2014.291; published online 15 September 2014 INTRODUCTION approaches, the clinical failure rate of novel anticancer therapeutic 7 As a disease caused by corruption of normal biological circuits and molecules has been high; moreover, resistance to targeted processes to sustain uncontrolled proliferative growth, cancer is anticancer therapies is a ubiquitous theme, often occurring almost always characterized by a complex spectrum of alterations through a variety of mechanisms, many of which are linked to 8–10 that affect multiple scales ranging from molecular activity within complex signaling pathways and networks. It is our view that cells onto communication between cells and tissues (Figure 1). systems biology approaches that can manage and model this At each of these scales, highly complex patterns of interactions are complexity will be increasingly needed to identify and validate observed and this complexity leads to great challenges in targets, biomarkers and discover more effective and less toxic understanding cancer progression and designing effective cancer therapeutic strategies. Such approaches will complement and therapy. At the genomic level, frequent disruption of the DNA extend more traditional reductionist approaches. maintenance machinery1 and epigenetic modifiers2 result in We define systems biology as the study of complex interactions thousands of sequence alterations3 and global epigenetic in biological systems and the emergent properties that arise from reprogramming,4 greatly complicating the discovery of underlying such interactions. In the field of cancer, systems biology aims at drivers of tumor progression. At the protein level, the frequent developing an increasingly holistic view of cancer development involvement of complex signaling networks makes it difficult to and progression. Systems biology approaches help understand anticipate the influences of oncogenic perturbations and to how complex cancer-associated deregulations coordinately shape predict how to effectively reverse those influences with pharma- malignant states and phenotypes, sometimes across multiple cological agents. At the tissue level, interactions between a tumor biological scales. In parallel, rapid technological advances now and its local environment consisting of a multitude of cell types enable high-throughput and systematic profiling of cancer that dynamically coevolve critically influence tumor growth and genomes, transcriptomes, proteomes, metabolomes and of the invasion.5 The complexity of the networks involved, the large tumor microenvironment. Together with clinical information, number of cells and cell types, the patterns of coordination these data are critical for the establishment of integrated and between molecules, cells and tissues and the constantly changing predictive cancer models. Even though computational and and evolving nature of the disease dictated by a Darwinian mathematical approaches used in systems biology are highly process6 cannot be fully addressed by traditional reductionist versatile, a few categories of general methodologies have approaches. Perhaps as a result of relying solely on reductionist emerged for specific purposes in cancer research. One class is Laboratory of Cancer Systems Biology, Sandra and Edward Meyer Cancer Center, Department of Physiology and Biophysics, Institute for Computational Biomedicine and Institute for Precision Medicine, Weill Cornell Medical College, New York, NY, USA. Correspondence: Dr O Elemento, Laboratory of Cancer Systems Biology, Sandra and Edward Meyer Cancer Center, Department of Physiology and Biophysics, Institute for Computational Biomedicine and Institute for Precision Medicine, Weill Cornell Medical College, 1305 York Avenue, New York, NY 10021, USA. E-mail: [email protected] Received 6 June 2014; revised 11 August 2014; accepted 11 August 2014; published online 15 September 2014 Cancer systems biology W Du and O Elemento 3216 Figure 1. Cancer alterations that affect complex molecular interactions across multiple biological scales. Within a cancer cell, thousands of genetic and epigenetic alterations result in aberrant protein expression or expression of proteins with abnormal functions. Consequently, the homeostasis of cellular signaling networks is frequently disrupted, leading to uncontrolled activation of survival and proliferation factors that drive tumor progression. At the tissue level, cancer cells closely interact with multiple cell types in the local environment. Through these interactions, the tumor microenvironment is educated by cancer cells to promote tumor growth and invasion. In this figure, we use activated B-cell-like diffuse large B-cell lymphoma (ABC-DLBCL) featuring aberrant B-cell receptor (BCR) signaling as an example to illustrate cancer alterations across multiple biological scales. integrative statistical analysis of large-scale cancer multi-omics proved important in profiling the activity and states of cancer cells and clinical data. These unbiased data-driven analyses and cell populations. Finally, we discuss how cancer systems have identified key biological processes underlying cancer biology will facilitate the adoption and implementation of pathogenesis,11 prognostic biomarkers12 and predictive signatures personalized and precision cancer medicine. for drug response.13,14 Another class is mathematical modeling of interaction networks such as intracellular signaling pathways or extracellular crosstalks between tumor and the microenvironment. PROBING CANCER COMPLEXITY AT THE GENOMIC LEVEL These models have proved useful at unraveling mechanisms of Tumors are characterized by a broad array of genetic and drug resistance and in optimizing combinatorial targeted epigenetic disruptions 15–17 therapy. Furthermore, evolutionary models that simulate Cancer is a disease driven by the accumulation of genomic and tumor growth and progression have provided important insights epigenomic alterations. Genomic alterations such as point into the evolution dynamics of tumor and have led to the 18,19 mutations, translocations and copy number variations result in discoveries of more effective dosing schedules. Overall, the aberrant gene expression or expression of mutated genes with application of systems biology approaches have led to substantial abnormal function.3 Genomic abnormalities are frequently accom- improvements in our understanding of cancer initiation and panied by epigenomic alterations,4,20,21 for example, hyper- or progression and to the discovery and implementation of more hypomethylation of specific regions and genes or changes in effective anticancer therapeutic strategies (Figure 2). levels of histone modifications.22,23 Like genomic alterations, As observed in many cancer types, substantial heterogeneity of epigenomic alterations are thought to affect expression levels of molecular alterations in patient population yields highly variable certain genes.4 It is generally thought that epigenetic alterations clinical responses to the same treatment. Consequently, persona- are brought about by mutations or aberrant expression of lized and precision cancer medicine, in which treatment of chromatin-modifying enzymes, a hypothesis supported by the fi patients is selected according to their speci c molecular dereg- ubiquitous presence of mutations in epigenetic enzymes across ulations, is urgently needed to improve therapeutic specificity and nearly all tumor types.2 An alternative model is that epigenetic efficacy. We believe that systems biology approaches that alterations occur by