Culture of various types of cells on gelled, reconstituted extracellular matrices results in decreased cellular proliferation. In the present study, we evaluated several possible mechanisms for this inhibition, as applied to cultured bovine retinal microvascular endothelial cells (EC) or to retinal pigment epithelial (RPE) cells: whether the inhibition might be related to (a) inactivation of fibroblast growth factor (FGF) by binding of the molecules present in the medium to a matrix component; (b) release of an inhibitor by the matrix in culture; or (c) inhibitory properties of the matrix macromolecules themselves. Our results suggest that mechanism (c) is most likely. The reasons are, first, that culture of EC or RPE cells on several different extracellular matrix substrates in the presence of logarithmically increasing concentrations of acidic or basic fibroblast growth factors (aFGF or bFGF) leads to a vertical shift of the plots of cell number after 4 days in culture vs. log growth factor concentration for both types of cells. The same result obtains when cells are cultured with logarithmically increasing concentrations of all-trans retinoic acid, which inhibits EC but not RPE cell proliferation in a dose-dependent fashion. This is consistent with mechanism (b) or (c), but not (a), for which one would expect a horizontal shift. Second, washing the matrices prior to the plating of cells with 1M NaCl, which elutes aFGF and partially elutes bFGF molecules from basement membranes, does not alter the growth of cells plated after the wash. This suggests also that growth factor binding to the matrix is not a likely mechanism for the observed inhibition. Incubation of matrices with culture medium prior to plating cells does not usually alter the ability of the medium thus "conditioned" to support cell growth, arguing against the possibility that the matrices release a soluble activator or inhibitor of such growth. However, in some experiments performed with lots of Matrigel (a commercially available basement membrane extract from a murine tumor) obtained prior to mid-1991, media "conditioned" by incubation with this matrix did show enhanced ability to facilitate EC and RPE cell proliferation. Finally, if RPE cells or EC are plated on various substrates, allowed to attach for 24 hr., and then the same or other substrates are poured over the cells, the effect on proliferation of the matrices plated on the apical surfaces of the cells is often less than that of matrices plated adjacent to their basal surfaces. Although in most cases these differences are not statistically significant, there is an apparent trend.(ABSTRACT TRUNCATED AT 400 WORDS)