Morphogenesis of normal human salivary gland cells in vitro.

Primary cultured human salivary gland cells were transfected with ori-defective mutant DNA of SV40. Following 2-3 weeks of transfection, slowly expanding colonies consisting of small compact cells emerged, while mock-transfected cells did not grow any more and eventually entered crisis, followed by cell death. Using limited dilution technique, we isolated 4 cell clones with distinct morphology from a single colony. Morphological observation of cells cultured on plastic dishes precisely revealed the characteristics of the constituent cells of salivary gland; i.e., three cell clones showing cuboidal- (NS-SV-DC), spindle- (NS-SV-MC), and flattened morphology (NS-SV-SC) were similar to duct-, myoepithelial-, and squamous phenotype, respectively. A remaining cell clone, polygonal in shape and with numerous secretory granules (NS-SV-AC), resembled an acinar cell. Characterization of cell clones by ultrastructural examination and search for specific antigens showed the similarity of NS-SV-DC, NS-SV-MC, NS-SV-AC, and NS-SV-SC to duct, myoepithelial, acinar, and squamous cells, respectively. Anchorage-independent growth in soft agar and tumorigenicity in nude mice were not recognized in all cell clones. These results demonstrate that establishment of cell clones with duct-, myoepithelial-, acinar-, or squamous phenotype was accomplished in the in vitro system, and that based on the evaluation of colony-forming ability in soft agar and tumorigenicity in nude mice these cell clones can be considered to be non-tumorigenic. Using the above in vitro system, we examined the effect of a reconstituted basement membrane extract, Matrigel, on the morphogenesis of cultured normal human salivary gland cells. When NS-SV-DC or NS-SV-MC were seeded on Matrigel in serum-free culture conditions, they formed round or zonal clusters on day 1; failing however, to develop into a salivary gland morphogenesis. Semithin sections of cell clones cultured on Matrigel exhibited multicellular aggregates on day 1, while on days 2 and 3 these cells lost both cell-cell and cell-Matrigel interactions and eventually entered crisis. In an attempt to understand the mechanism involved in this phenomenon, we have investigated proteolytic enzymes and their inhibitors secreted by cell clones. Although cell clones produced almost identical levels of gelatinases, they released increased amounts of plasminogen activators (PAs) as compared with a neoplastic human salivary gland cell line (HSG), which had already been demonstrated to differentiate into acinar cells when cultured on Matrigel. Obvious difference of expression level of tissue inhibitor of metalloproteinases-1 (TIMP-1) was not observed in these cells.(ABSTRACT TRUNCATED AT 400 WORDS)