Tissue Structure, Nuclear Organization, and Gene Expression in Normal and Malignant Breast I

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 membranebound 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 gehome. 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 threedimensional 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 /]1 integrin and epidermal growth factor receptor signal transduction pathways are integrated reciprocally in three-dimensional cultures, on tissue culture plastic (twodimensional 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 epithelialspecific 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. Received 11/11/98; accepted 2/4/99. 1 Presented at the "General Motors Cancer Research Foundation Twentieth Annual Scientific Conference: Developmental Biology and Cancer," June 9-10, 1998, Bethesda, MD. This work was supported by Contract DE-AC03-76SF00098 from the United States Department of Energy, Office of Biological and Environmental Research (to M. J. B.), and by NIH Grants CA64786 and CA57621 (to M.J.B.). Additional funding is as follows: WHO/IARC and Department of Defense/Breast Cancer Research Program fellowship (to S. A. L.); University of California/Breast Cancer Research Program fellowship (to V. M. W.) and a grant from the Danish Medical Research Council (to O. W. P.); and United States Department of Energy, Office of Biological and Environmental Research (an Alexander Hollacnder Distinguished Postdoctoral Fellowship administered by the Oak Ridge Institute for Science and Education; to K. L. S.). 2 To whom requests for reprints should be addressed, at Lawrence Berkeley National Laboratory, One Cyclotron Road, Berkeley, CA 94720. "Science is built up with facts, as a house is with stones. But a collection of facts is no more science than a heap of stones is a house."-Jules Henri Poincarb (1854-1912) In adult organisms, cells must maintain the program of regulated gene expression that is instituted during development . What are the genomic rules that allow this p rogram of gene expression to be selective? Even after the cells have arrived at their "destinations" within the tissues, they radically alter their patterns of gene expression, both quantitatively and qualitatively, as a result of systemic (hormones, chemokines , and environmenta l insults) and microenvironmental (cell-ECM, cell-cell, local growth factors, local injury, and others) signals. This is observed most dramatically when cells are isolated and cultured outside of the organism, most notably on tissue culture plastic (as 2D 3 monolayers; 1). Along with loss of tissuespecific gene expression, the most striking change is in cellular and nuclear organization and architecture (2). One of us proposed almost two decades ago that in addition to growth factors and hormones, E C M that surrounds tissues in vivo contains signaling molecules that are responsible for maintenance of tissue form and function (3), and furthermore, that there may be both mechanical and biochemical connect ions between the ECM and the nuclear skeleton, leading to changes in chromatin structure and gene expression. The combined work of many investigators, including our own, has conf i rmed the significance of the ECM in every tissue examined (reviewed in Refs. 4 6 ) . In addition, the discovery of the ECM receptors, the most important family of which is the integrins (reviewed in Refs. 7 and 8), has elucidated a mechan i sm by which ECM signaling could be achieved across the cellular membrane. In parallel, the work of structural and molecular biologists has unraveled important new informat ion about the structure of the nucleus and the chromatin. The higher order structure of eukaryotic ch romosomes consists of independent loop domains that are separated f rom each other by the at tachment of specialized genomic sequences (matrix at tachment regions) onto the N M (reviewed in Refs. 9 -11) . This organization is important not only to compact DNA but also for various functions involving DNA. Matrix at tachment regions have been shown to be essential for demethyla t ion of chromatin domains (12) and chromat in accessibility (13), processes directly implicated in regulating gene expression. The compact ion of eukaryotic D N A into chromatin is thought to establish a specific pattern of gene expression. Heterochromatin, defined cytogenetically as regions o f the genome that remain condensed throughout the cell cycle, is known to remain transcriptionally silent. Translocation of a euchromat ic region of the genome to a site adjacent to heterochromatin often yields variable silencing of the translocated genes, as exemplif ied by the process of posit ion effect variegation of the eye color gene brown, in Drosophila (14). Other examples of chromatin packaging associated with long-term transcriptional repression is the transcriptional silencing observed at the mating 3 The abbreviations used are: 2D, two-dimensional; 3D, three-dimensional; ECM, extracellular matrix; NM, nuclear matrix; MEC, mammary epithelial cell; EGF, epidermal growth factor; EGFR, EGF receptor.