The studies described in this dissertation demonstrate that basement membrane may be required for maintenance of organized epithelial tissue architecture. The structural organization of normal rat pancreatic acinar epithelium is fully characterized in order to analyze its neoplastic disorganization within a pancreatic acinar cell tumor. Transmission electron microscopy and indirect Or immunofluorescence microscopy of frozen semithin tissue sections (0.5 μm) are utilized to localize different components of the plasmalemma (leucine aminopeptidase, γ-glutamyl transferase, and a basolateral integral membrane protein), cytoskeleton (F-actin), basement membrane (laminin, fibronectin, and types IV and V collagens), and connective tissue (types I and III collagens). Normal acinar cells sit on a continuous basement membrane and display a polarized distribution of intracellular organelles, cytoskeletal elements, and distinct, membrane domains while these organized cell relations are lost within the parenchyma of the pancreatic acinar carcinoma. This tumor-associated disorganization of normal epithelial cell relations correlates directly with absence of integral basement membrane within the parenchyma of this tumor. Interestingly, pancreatic acinar tumor cells also retain the ability to reorient in vivo when in direct contact with intact basement membrane which appears only along the connective tissue boundary.

The rare ability of the pancreatic acinar carcinoma to reorganize in vivo makes this tumor an excellent model system with which to test my thesis that epithelial organization may be triggered by basement membrane and so be directed by forces originating outside the cell. In vitro investigations utilize mechanically dispersed tumor cells and capitalize upon the availability of human amnion as a source of intact, basement membrane. Acinar tumor cells will not spontaneously adhere to standard culture substrata although they can be maintained on amniotic stroma in vitro if washing steps are minimized. Dispersed tumor cells remain round, do not accumulate basement membrane, and display circumferential cortical rings of actin-containing material when cultured on stroma. On the other hand, acinar tumor cells rapidly adhere, spread, and assemble actin into fibrous bundles in the absence of new protein synthesis when cultured on amniotic basement membrane. Studies using dishes coated with purified extracellular matrix components indicate that tumor cell attachment, can be mediated by isolated laminin and type IV collagen although cell spreading is promoted primarily by the former. Importantly, cell shape is shown to be dependent upon the structural integrity of the anchoring substratum as laminin-coated native collagen gels that are malleable mediate tumor cell attachment but can not support alterations in cell form. In addition, ³H-thymidine autoradiography is used to show that the growth rate of tumor cells on basement membrane in vitro is less than that previously described for this tumor in suspension immediately after dispersion.

Amniotic basement membrane and stroma are also utilized for studies designed to determine if basement membrane is sufficient to trigger pancreatic epithelial tumor cell polarization in vitro. Computerized morphometric analysis shows that acinar tumor cells remain randomly oriented on amniotic stroma in vitro as they do within the tumor parenchyma in vivo. However, consistent polarization of intracellular organelles and ultrastructural surface features (e.g., apical microvilli, basal plasmalemmal flattening) are observed within tumor cells adherent to amniotic basement membrane. This repolarization is associated with intracellular cytoskeletal alterations and does not require new protein synthesis although full junctional complexes and regeneration of distinct membrane domains are not observed at the time points studied.

Thus, basement membrane is necessary and sufficient for reorientation of independent epithelial tumor cells in vitro and so the organization of a polarized epithelium may be directed by forces from outside the cell via basement membrane. Stability of epithelial form may require continued presence of intact basement membrane and its dissolution could be responsible for neoplastic disorganization of epithelial architectural relations as observed within the pancreatic acinar cell carcinoma. A model for architectural regulation of tissue structures with basement membrane as a transducer of physical forces is presented.