Because every cell within the body has the same genetic information, a significant problem in biology is to understand how cells within a tissue express genes selectively. A sophisticated network of physical and biochemical signals converge in a highly orchestrated manner to bring about the exquisite regulation that governs gene expression in diverse tissues. Thus, the ultimate decision of a cell to proliferate, express tissue-specific genes, or apoptose must be a coordinated response to its adhesive, growth factor, and hormonal milieu. The unifying hypothesis examined in this overview is that the unit of function in higher organisms is neither the genome nor the cell alone but the complex, three-dimensional tissue. This is because there are bidirectional connections between the components of the cellular microenvironment (growth factors, hormones, and extracellular matrix) and the nucleus. These connections are made via membrane-bound receptors and transmitted to the nucleus, where the signals result in modifications to the nuclear matrix and chromatin structure and lead to selective gene expression. Thus, cells need to be studied "in context", i.e., within a proper tissue structure, if one is to understand the bidirectional pathways that connect the cellular microenvironment and the genome. In the last decades, we have used well-characterized human and mouse mammary cell lines in "designer microenvironments" to create an appropriate context to study tissue-specific gene expression. The use of a three-dimensional culture assay, developed with reconstituted basement membrane, has allowed us to distinguish normal and malignant human breast cells easily and rapidly. Whereas normal cells become growth arrested and form organized "acini," tumor cells continue to grow, pile up, and in general fail to respond to extracellular matrix and microenvironmental cues. By correcting the extracellular matrix-receptor (integrin) signaling and balance, we have been able to revert the malignant phenotype when a human breast tumor cell is cultured in, or on, a basement membrane. Most recently, we have shown that whereas beta1 integrin and epidermal growth factor receptor signal transduction pathways are integrated reciprocally in three-dimensional cultures, on tissue culture plastic (two-dimensional monolayers), these are not coordinated. Finally, we have demonstrated that, rather than passively reflecting changes in gene expression, nuclear organization itself can modulate cellular and tissue phenotype. We conclude that the structure of the tissue is dominant over the genome, and that we may need a new paradigm for how epithelial-specific genes are regulated in vivo. We also argue that unless the structure of the tissue is critically altered, malignancy will not progress, even in the presence of multiple chromosomal mutations.