Comparison of two- and three-dimensional cancer models for assessing potential cancer therapeutics

Abstract A plethora of molecules with the capacity to kill cancer cells have been identified through intensive research efforts. These agents were typically identified using screens that involved established cancer cell lines grown on plastic substrates in vitro in a two-dimensional (2D) format, allowing for the testing of even poorly soluble molecules. Promising agents, formulated for in vivo administration, can then be examined using some form of xenograft model in immune-compromised mice to obtain a safety to efficacy profile sufficient to warrant clinical testing. While some of the agents initially screened in this way have achieved clinical utility for a variety of cancers, their use commonly comes with dose-limiting side-effects that damage highly proliferating nontumor cells as well as immune cells, ultimately limiting the capacity of a patient to derive durable benefit through selective recognition and clearance of cancer cells. As 2D cell cultures do not recapitulate the three-dimensional (3D) architecture and complexity of human solid tumors, more representative preclinical models are needed to accelerate drug discovery. A greater understanding of how cancer cells can manipulate a 3D tumor microenvironment to thwart or modify the actions of potential therapeutic agents identified in 2D screens has recently been used to explain the poor performance for some of these agents in the clinic. Thus efforts to more accurately assess potential anticancer agents in settings that resemble a tumor microenvironment in vivo have recently been a major focus of academic and industrial investigators where extracellular matrix (ECM) components and multiple noncancerous cells (including stromal fibroblasts, infiltrating immune cells as well as blood and lymphatic vascular networks) have been introduced. Therapeutic approaches where tumor cells and coopted nontumor cells in the microenvironment are simultaneously inhibited may offer a more efficient way to treat cancer. This chapter focuses on technical advances made in 3D in vitro models that could be used to assess potential anticancer strategies in a manner that may be more predictive of clinical outcomes. Such models are particularly important in the testing of drug candidates whose actions are designed to alter the function of ECM components and/or noncancerous cells within a tumor microenvironment.