Cancers are complex and heterogeneous pathological "organs" in a dynamic interplay with their host. Models of human cancer in vitro, used in cancer biology and drug discovery, are generally highly reductionist. These cancer models do not incorporate complexity or heterogeneity. This raises the question as to whether the cancer models' biochemical circuitry (not their genome) represents, with sufficient fidelity, a tumor in situ. Around 95% of new anticancer drugs eventually fail in clinical trial, despite robust indications of activity in existing in vitro pre-clinical models. Innovative models are required that better capture tumor biology. An important feature of all tissues, and tumors, is that cells grow in three dimensions. Advances in generating and characterizing simple and complex (with added stromal components) three-dimensional in vitro models (3D models) are reviewed in this article. The application of stirred bioreactors to permit both scale-up/scale-down of these cancer models and, importantly, methods to permit controlled changes in environment (pH, nutrients, and oxygen) are also described. The challenges of generating thin tumor slices, their utility, and potential advantages and disadvantages are discussed. These in vitro/ex vivo models represent a distinct move to capture the realities of tumor biology in situ, but significant characterization work still remains to be done in order to show that their biochemical circuitry accurately reflects that of a tumor.

Two-dimensional (2D) cell cultures growing on plastic do not recapitulate the three dimensional (3D) architecture and complexity of human tumors. More representative models are required for drug discovery and validation. Here, 2D culture and 3D mono- and stromal co-culture models of increasing complexity have been established and cross-comparisons made using three standard cell carcinoma lines: MCF7, LNCaP, NCI-H1437. Fluorescence-based growth curves, 3D image analysis, immunohistochemistry and treatment responses showed that end points differed according to cell type, stromal co-culture and culture format. The adaptable methodologies described here should guide the choice of appropriate simple and complex in vitro models.

One of the applications of electroporation/electropulsation in biomedicine is gene electrotransfer, the wider use of which is hindered by low transfection efficiency in vivo compared with viral vectors. The aim of our study was to determine whether modulation of the extracellular matrix in solid tumors, using collagenase and hyaluronidase, could increase the transfection efficiency of gene electrotransfer in histologically different solid subcutaneous tumors in mice. Tumors were treated with enzymes before electrotransfer of plasmid DNA encoding either green fluorescent protein or luciferase. Transfection efficiency was determined 3, 9, and 15 days posttransfection. We demonstrated that pretreatment of tumors with a combination of enzymes significantly increased the transfection efficiency of electrotransfer in tumors with a high extracellular matrix area (LPB fibrosarcoma). In tumors with a smaller extracellular matrix area and less organized collagen lattice, the increase was not so pronounced (SA-1 fibrosarcoma and EAT carcinoma), whereas in B16 melanoma, in which only traces of collagen are present, pretreatment of tumors with hyaluronidase alone was more efficient than pretreatment with both enzymes. In conclusion, our results suggest that modification of the extracellular matrix could improve distribution of plasmid DNA in solid subcutaneous tumors, demonstrated by an increase in transfection efficiency, and thus have important clinical implications for electrogene therapy.