MICROSCOPY RESEARCH AND TECHNIQUE 43:379–384 (1998) Adhesion-Guided In Vitro Morphogenesis in Pure and Mixed Cell Cultures MARK J. POWERS AND LINDA G. GRIFFITH* Department of Chemical Engineering and Center for Biomedical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139 KEY WORDS organ morphogenesis; adhesion; pure and mixed cell cultures ABSTRACT The ability to understand and control the morphogenesis of mammalian cells is a fundamental objective of cell and developmental biology and tissue engineering research. Numerous processes, both biochemical and biophysical in nature, have been studied in an attempt to elucidate the mechanisms underlying this behavior. We focus here on the contributions of biophysical phenomena to the morphogenetic behavior of pure and mixed cell populations on solid surfaces in vitro. These principles are illustrated using characteristic liver tissue cells as a model system. The studies discussed demonstrate that cell-substratum and cell-cell adhesive forces are critical determinants of the ultimate morphology, cytoarchitecture, and organization achieved by these cells in vitro. Microsc. Res. Tech. 43:379–384, 1998. r 1998 Wiley-Liss, Inc. INTRODUCTION Morphogenesis of three dimensional tissue-like structures in vitro is of increasing interest for basic studies of differentiated cell function as well as applications in the growing field of therapeutic tissue engineering (Langer and Vacanti, 1993). Although the roles of specific signaling processes initiated by ligation of growth factor or adhesion receptors are currently being delineated, cell behavior in vivo ultimately results from integration of many signals within the tissue environment, which are as yet unknown or unquantified. Co-cultures or feeder layers are often used to provide some of these missing signals, but failure to achieve cell functions in vitro that are known to occur in vivo, such as infection of human liver cultures by hepatitis B virus, suggests that the hierarchichal structure of tissue and the accompanying hierarchical delivery of matrix and cytokine signals may also be an essential factor to reproduce in vitro in order to maintain true tissue function. Here, we focus on factors that influence morphogenesis of tissue-like structures in pure and mixed primary cell cultures, and emphasize a systematic approach to identifying culture parameters that can predictably be manipulated to control such morphogenesis. Enormous interest in the intrinsic cellular clues and mechanisms guiding the morphogenetic process was spawned by Townes and Holtfreter, who demonstrated that the dissociated cells of vertebrate embryonic organs would not only re-aggregate but also reorganize in an appropriate manner (Townes and Holtfreter, 1955). It is generally agreed that relatively few mechanisms are required to generate tissue organization, even if it is not known exactly how they lead to the formation of most structures. A variety of mechanisms have been studied in terms of their contribution to various aspects of morphogenesis including directed cell movement, cytoskeletal contraction, cell proliferation, differences in cell adhesion, and contact guidance (reviewed by Bard, 1990). r 1998 WILEY-LISS, INC. Pioneering work by Steinberg suggests that the sorting-out of intermixed embryonic cells, the spreading of one tissue over the surface of another, and the specific inside/outside tissue stratification that arises by either process are guided in large part by the differential intercellular adhesivities of the cell types involved (Steinberg, 1963, 1970, 1996). An important implication of this analysis is that different molecular configurations could lead to identical outcomes provided that they produce comparable adhesion forces. A similar analysis involving cellular forces was perfomed by Oster et al. (1983), who demonstrated that a variety of intricate structural arrangments and patterns could be predicted based strictly on biophysical force balances. Thus, while morphogenesis is a complex process involving numerous signals and guidance clues that can be both chemical and/or physical in nature (Gumbiner, 1996) the fundamental mechanisms responsible for this process are very possibly biophysical in nature. The findings discussed here detail investigations into the role played by biophysical cell-substratum and cell-cell forces in the morphogenesis and pattern formation of homogeneous and heterogeneous cell cultures. Liver is used as the representative tissue system in these studies based on the observed importance of morphology on in vitro cellular behavior. The basic functional unit of liver, the acinus, exhibits a wellcharacterized architecture consisting of continuous plates or cords of a single layer of hepatocytes separated from the sinusoidal blood supply on both sides by a layer of endothelial cells interspersed with stellate cells and Kupffer cells; the plates continuously branch *Correspondence to: Linda G. Griffith, Department of Chemical Engineering and Center for Biomedical Engineering, Massachusetts Institute of Technology, 66–466, Cambridge, MA 02139. E-mail: [email protected] Received 16 August 1998; accepted in revised form 25 August 1998 Contract grant sponsor: National Science Foundation; Contract grant numbers: BCS-9157321, BES-963–2714; Contract grant sponsor: National Institutes of Health; Contract grant number: GM-50047; Contract grant sponsor: Harvard Dental School/MIT. 380 M.J. POWERS AND L.G. GRIFFITH Fig. 1. Comparison of rat hepatocyte cellsubstratum adhesion strengths with cell traction forces. Primary rat hepatocytes were cultured on Matrigel substrata which ultimately lead to the formation of monolayers (䊏) or spheroids (䊊, ⴛ). The shear force required to detach these cells was determined after 24 hours. in culture. Shaded areas represent cell traction force values for fibroblasts (F) (Oliver et al., 1994), keratocytes (K) (Lee et al., 1994), and the estimated range for hepatocytes (H). Ref. Powers et al. (1997). and interconnect between the portal tract and the central vein. Three distinct epithelial domains are present in the hepatic plates: an apical domain forming the bile canaliculi, a lateral domain consisting of epithelial junctions that interconnect the hepatocytes, and a basal (sinusoidal) domain that is adjacent to the sinusoidal endothelial cells and other non-parenchymal cell types. At least some semblance of this acinar cytoarchitecture and organization seems to be essential for the expression of liver-specific function in hepatocellular cultures (Clayton et al., 1985; Dunn et al., 1991; Moghe et al., 1996). Investigations into the morphogenetic behavior of pure cultures of hepatocytes and mixed cultures of hepatocytes and endothelial cells— the two predominant cell types defining the architecture of liver— therefore provide a useful paradigm with which to study the morphogenetic behavior of dispersed tissue cells as they re-form into native tissue structures. MORPHOLOGICAL BEHAVIOR OF PURE HEPATOCYTE CULTURES Aggregation of primary hepatocytes has been observed on numerous surfaces comprised of different architectures and/or molecular compositions (Koide et al., 1989, 1990; Landry et al., 1985; Parsons-Wingerter and Saltzman, 1993; Peshwa et al., 1994; Saltzman et al., 1991). These substrata have been shown to play an important role in the ultimate morphology achieved by these cell aggregates (Koide et al., 1989; Landry et al., 1985). Since a substratum is capable of imparting both biochemical and biophysical signals to cells, it is essential to delineate the contributions of each to the morphogenetic process. We have shown previously (Powers et al., 1997) that hepatocytes are capable of forming both monolayered and spheroidal aggregates on a diverse array of rigid extracellular matrix substrata, including laminin, fibronectin, collagen I, and Matrigel. Hepatocytes attach to each of these matrices through different sets of cell-surface receptors (Clement et al., 1989; Forsberg et al., 1990; Gullberg et al., 1992; Pujades et al., 1992; Stamatoglou et al., 1990a,b), suggesting that differences in biochemical signalling, if present on the different substrata, do not significantly affect the resulting aggregate morphologies on these surfaces. Rather, the morphology of the hepatocyte aggregates correlates strongly with the ligand density present on each surface. For each type of matrix studied, high surface ligand concentrations leads almost exclusively to the formation of monolayered aggregates. Conversely, low matrix concentrations induce the formation of spheroidal aggregates. Quantitation of cell-substratum adhesion strength on these surfaces indicates that it is the biophysical forces present in the system that provide a basis for the observed morphological results: substrata on which monolayers form promote higher levels of cell-substratum adhesion than do those surfaces on which spheroids form. On a general level these analyses provide a basis for simple assessment of the relative affinity of cells for these different substrata (i.e., high or low) based on the morphology of the resulting aggregates. Investigations into the mechanisms responsible for the observed aggregation and morphological behavior provide further clues as to the impact of other cell-based forces in this system. Hepatocytes on rigid substrata aggregate through a mechanism involving active cell membrane protrusion leading to cell-cell attachment and followed by what appears to be a discrete contraction event which often causes a rapid retraction of the peripheral edges of the aggregating cells (Powers and Griffith-Cima, 1996). The membrane extension or spreading of one cell on top of another was not observed to occur (i.e., cells did not form directly into multilayered aggregates). These observations suggest that cell-generated contractile forces, which have been identified as key determinants in the morphogenesis of other cell systems (Bard, 1990; Oster et al., 1983), are critical in the determination of hepatocyte aggregate morphology. Further evidence for this supposition is provided by studies of hepatocyte aggregation on biochemically identical gel substrata of varying deformabilities. On such substrata, spheroids form on malleable gels that are easily deformed by the cytoskeletal forces that the aggregating cells generate; ROLE OF ADHESION IN ORGAN MORPHOGENESIS when the gels are present in a rigid, non-deformable state monolayers result (Coger et al., 1997; Linblad et al., 1991). These results can be interpreted by considering the contributions of cytoskeletal contraction to hepatocyte aggregate morphology. It has been demonstrated that the balance between cell-generated traction forces, which represent that part of the total contractile force transmitted to the substratum (Lauffenburger and Horwitz, 1996), and cell-substratum adhesive forces play an important role in the determination of hepatocyte aggregate morphology (Powers et al., 1997). Figure 1 illustrates the data that support this hypothesis, showing that the estimated hepatocyte traction forces are lower than the expected cell-substratum adhesive forces of a majority of cells attached to the ‘‘monolayerpromoting’’ substrata but greater than most of the cell-substratum adhesive forces present on the ‘‘spheroid-promoting’’ surfaces. In other words, spheroids form when cytoskeletal contractile forces can act to ‘‘pull’’ the cells together, while monolayers result when the cellsubstratum adhesion strength is insurmountable. While this analysis provides convincing evidence that the interplay between cell-substratum adhesion and cellular contraction is critical to hepatocyte aggregate morphogenesis, cell-cell adhesion may also play a role in this process. Martz et al. (1974) outline a model system by which variation in cell-cell and cell-substratum adhesive interactions alone could account for multiple morphological states. In this system, aggregate morphology evolves through a thermodynamic free energy minimization process. As cells aggregate, physical remodeling results from small perturbations in the local molecular (i.e., receptor-based) equilibrium as the cell aggregate attempts to increase high adhesion (either cell-cell or cell-substratum) contact regions. This behavior is analagous to that observed by water droplets interacting with surfaces: the water molecules will bead on hydrophobic surfaces and spread on a hydrophilic surfaces. The model therefore predicts that, when cell-substratum attachment sites are sparse, surfacefree energy is minimized by maximizing the more highly adhesive intercellular adhesions, and spheroids result. Conversely, when cell-substratum attachment sites are plentiful, surface-free energy is minimized by the maximizing the number of cell-substratum adhesion bonds, resulting in monolayers. Active, directional activation of the cytoskeletal machinery does not play a role in this model. In this sense, morphogenesis would occur through a passive mechanism with the relative strengths of cell-cell vs. cell-substratum adhesive interactions ultimately determining the outcome. In our system, previously cited observations (i.e., lack of direct multilayering of coupled cells, the occurrence of a discrete contractile event, cell behavior on gel substrata) suggest that contractile force is the defining event in the initiation of morphogenesis; however, the maturation of aggregate morphology into multi-layer spheroids is likely to be at least partially due to the the free-energy minimization phenomena described above. It is important to stress that both parameters are likely to contribute to the ultimate morphological state of the aggregates, with the contractile behavior of the cells allowing the morphogenesis to occur on a kinetically favorable time scale. 381 Fig. 2. Predicted morphological outcomes for cells sorting on surfaces based on the adhesive properties of the system. Examples are shown for cells that exhibit different levels of cell-cell and cellsubstratum affinity. When cell-cell forces are dominant, spheroidal aggregates would be predicted to form with the high affinity cells sorting internally to the low affinity cells. When cell-surface forces are dominant, monolayers would be expected, with the cells sorting from one another only within this morphological framework. Two intermediate cases are also shown. These analyses indicate that the morphological state of a cellular aggregate provides valuable clues as to the biophysical ‘‘hierarchy’’ present in the system. When monolayers are present, the cell-substratum forces are dominant over cell traction forces and/or cell-cell adhesion forces,while the formation of spheroids indicates that the opposite is true. MORPHOLOGICAL BEHAVIOR OF MIXED CELL CULTURES Due to the observed importance of biophysical properties in the morphogenesis of homogeneous cell cultures, 382 M.J. POWERS AND L.G. GRIFFITH Fig. 3. Phase contrast images of primary rat hepatocytes and bovine aortic endothelial cells coated on high (left), intermediate (middle), and low (right) concentrations of Type I collagen. Images were recorded 24 hours (top) or 4 days (bottom) post-seeding. Examples of cellular morphology are labeled H (hepatocytes) or E (endothelial cells) where discernible. we now focus on the role played by these phenomena in the sorting of mixed cocultures on rigid substrata. Since most tissues are comprised of multiple cell types, these studies provide a more tissue-like system with which to investigate these phenomena. For the purposes of this analysis, we again focus on the liver. Because of the importance of hepatocytes and endothelial cells in acinar architecture, we have chosen to study interactions between these two cell types. The role played by biphysical cell-cell and cell-surface phenomena in the sorting of these two cell types from one another therefore presents an intriguing case study. In considering our approach to these analyses, it should be noted that the nature of interaction between hepatocytes and endothelial cells in vitro is largely unknown. In vivo interactions between hepatocytes and endothelial cells in the postnatal, nonregenerating liver are typically modulated through the extracellular matrix present in the Space of Disse. In vitro, cells are devoid of endogenous matrix at the time of initial plating, but matrix deposited during culture may ultimately control hepatocyte-endothelial cell interactions in vitro. Hepatocytes are indeed capable of synthesizing and secreting extensive arrays of matrix proteins in vitro (Stamatoglou et al., 1987; Sudhakaran et al., 1986; Tamkun and Hynes, 1983) with which endothelial cells are capable of associating. Direct hepatocyte-endothelial adhesion may also be possible through cadherin-based adhesion. Although these cell types express different cadherin receptors —E-cadherin in hepatocytes (Vestweber et al., 1985) and VE- and N- cadherin in endothelial cells (Navarro et al., 1998) — it has been postulated that nonidentical cadherin systems on different cell types could mediate intercellular adhesion (Brackenbury et al., 1981; Mur- phy-Erdosh et al., 1995; Volk et al., 1987). But regardless of the nature of hepatocyte-endothelial interaction, these associations are likely to be more temporal and less able to resist cytoskeletal forces than the stronger, more irreversible epithelial adhesion complexes present in aggregated hepatocytes (Farquhar and Palade, 1963; Miettinen et al., 1978; Stamatoglou and Hughes, 1994), leading us to focus on the role played by heterotypic cell-cell adhesion in this system. The hypothesis that intercellular adhesion phenomena provide a basis for cell sorting behavior was first proposed by Steinberg (1963), who theoretically and experimentally demonstrated that basic sorting behavior could be predicted through phase ordering principles (Steinberg, 1963, 1970, 1996; Steinberg and Takeichi, 1994). Given a culture system comprised of two cell types in which a cell-cell affinity hierarchy is established (i.e., cell type A is more ‘‘cohesive’’ than cell type B), it was shown that the cells will sort from one another so as to minimize the free energy of the system. That is, the high affinity (or high surface ‘‘free energy’’) cells form a spherical core, surrounded by a shell of the low affinity (low surface ‘‘free energy’’) cells. This is analogous to the behavior seen when two immiscible liquids of different surface tensions are mixed. The principles set forth by Steinberg can now be combined with the previously described model of Martz et al. (1974) to allow for cell sorting on a surface. This analysis adds another level of complexity to the system as cell-surface interactions are included in the aforementioned cell affinity hierarchy (Fig. 2). Using the principles of free energy minimization, sorting outcomes for cells on high, intermediate, and low surface free energy cases can be predicted. As seen in these examples of ROLE OF ADHESION IN ORGAN MORPHOGENESIS 383 the formation of spheroidal aggregates with hepatocytes sorting internally to a ‘‘shell’’ of endothelial cells. This architecture has also been observed to occur in other hepatocyte co-culture systems (Landry et al., 1985; Takezawa et al., 1992). According to our model these results suggest that homotypic interactions between hepatocytes are of a higher affinity than homotypic interactions between endothelial cells, not a surprising result given the epithelial nature of hepatocyte intercellular adhesions. Surfaces with intermediate ligand densities lead to the attachment of endothelial cells on top of monolayered hepatocyte aggregates, indicating that the adhesion of endothelial cells to hepatocytes is still thermodynamically preferred over endothelial attachment to a surface. Both hepatocytes and endothelial cells attach as monolayers to high ligand concentration surfaces, a conformation in which interaction with the surface is maximized and interaction with each other is minimized. These results suggest that the morphology and architecture of heterologous cellular structures are guided by biophysical phenomena. Furthermore, this behavior can be predicted based simply on the principles outlined here, potentially providing a simple and effective alternative to more costly and time-consuming approaches (e.g., staged/directed cell seeding; surface patterning) for achieving histotypical architectures in vitro. Extensions of this analytic approach to evolution of structure in more complex three-dimensional systems may ultimately enable in vitro organogenesis of functional tissue (Griffith et al., 1997). REFERENCES Fig. 4. Vertical histological sections (3 µm) of hepatocyte (H)endothelial cell (E) cultures as in Figure 3. High (top), intermediate (middle), and low (bottom) ligand density cultures were fixed on day 4, embedded in glycol methacrylate resin, sectioned, and stained with Hematoxylin & Eosin. Note similarities to predicted outcomes from Figure 2. Figure 2, a wide variety of outcomes could result simply by changing the affinity of cell-substratum interactions. In order to test these predictions of cell sorting on surfaces, mixtures of primary rat hepatocytes and bovine aortic endothelial cells were cultured on substrata of high, intermediate, and low extracellular matrix ligand densities (Fig. 3). Vertical histological sections of these cultures indicate the formation of cellular structures with intercellular organizations closely resembling those predicted from biophysical principles (Fig. 4). Furthermore, inspection of the specimens from intermediate and low ligand density surfaces reveals structures reminiscent of the histotypic arrangement of cells in the acinar architecture. Low surface ligand concentrations appear to promote a situation where cell-cell forces are dominant, leading to Bard, J. (1990) Morphogenesis. In: Developmental and Cell Biology Series. P.W. Barlow, D. Bray, P. B. Green and J. M. W. Slack, eds. Cambridge University Press, New York. Brackenbury, R., Rutishauser, U. and Edelman, G.M. (1981) Distinct calcium-independent and calcium-dependent adhesions systems of chicken embryo cells. J. Biol. Chem., 252:6835–6840. Clayton, D.F., Harrelson, A.L., and Darnell, J.E., Jr. (1985) Dependence of liver-specific transcription on tissue organization. Mol. Cell Biol., 10:2623–2632. Clement, B., Segui-Real, B., Hassell, J.R., Martin, G.R., and Yamada, Y. (1989) Identification of a cell surface-binding protein for the core protein of the basement membrane proteoglycan. J. Biol. Chem., 264:12467–12471. Coger, R., Toner, M., Moghe, P., Ezzell, R.M., and Yarmush, M.L. (1997) Hepatocyte aggregation and reorganization of EHS matrix gel. Tissue Eng., 3:375–390. Dunn, J.C.Y., Tompkins, R.G., and Yarmush, M.L. (1991) Long-term in vitro function of adult hepatocytes in a collagen sandwich configuration. Biotechnol. Prog., 7:237–245. Farquhar, M.G., and Palade, G.E. (1963) Junctional complexes in various epithelia. J. Cell Biol., 17:375–412. Forsberg, E., Paulsson, M., Timpl, R., and Johansson, S. (1990) Characterization of a laminin receptor on rat hepatocytes. J. Biol. Chem., 265:6376–6381. Griffith, L.G., Wu, B., Cima, M.J., Powers, M.J., Chaignaud, B., and Vacanti, J.P. (1997) In vitro organogenesis of liver tissue. Ann. N.Y. Acad. Sci., 831:382–397. Gullberg, D., Gehlsen, K.R., Turner, D.C., Ahlen, K., Zijenah, L.S., Barnes, M.J., and Rubin, K. (1992) Analysis of ␣1␤1, ␣2␤1 and ␣3␤1integrins in cell-collagen interactions: Identification of conformation dependent ␣1␤1 binding sites in collagen type I. EMBO J., 11:3865–3873. Gumbiner, B.M. (1996) Cell adhesion: the molecular basis of tissue architecture and morphogenesis. Cell, 84:345–357. Koide, N., Shinji, T., Tanabe, T., Asano, K., Kawaguchi, M., Sakaguchi, K., Koide, Y., Mori, M., and Tsuji, T. (1989) Continued high albumin production by multicellular spheroids of adult rat hepatocytes formed in the presence of liver-derived proteoglycans. Biochem. Biophys. Res. Commun., 161:385–391. Koide, N., Sakaguchi, K., Koide, Y., Asano, K., Kawaguchi, M., 384 M.J. POWERS AND L.G. GRIFFITH Matsushima, H., Takenami, T., Shinji, T., Mori, M., and Tsuji, T. (1990) Formation of multicellular spheroids composed of adult rat hepatocytes in dishes with positively charged surfaces and under other nonadherent environments. Exp. Cell Res., 186:227–235. Landry, J., Bernier, D., Ouellet, C., Goyette, R., and Marceau, N. (1985) Spheroidal aggregate culture of rat liver cells: histotypic reorganization, biomatrix deposition, and maintenance of functional activities. J. Cell Biol., 101:914–923. Langer, R.S., and Vacanti, J.P. (1993) Tissue engineering. Science, 260:920–926. Lauffenburger, D.A., and Horwitz, A.F. (1996) Cell migration: A physically integrated process. Cell, 84:359–369. Lee, J., Leonard, M., Oliver, T., Ishihara, A., and Jacobson, K. (1994) Traction forces generated by locomoting keratocytes. J. Cell Biol., 27:1957–1964. Linblad, W.J., Scheutz, E.G., Redford, K.S., and Guzelian, P.S. (1991) Hepatocellular phenotype in vitro is influenced by biophysical features of the collagenous substratum. Hepatology, 13:282–288. Martz, E., Phillips, H.M., and Steinberg, M.S. (1974) Contact inhibition of overlapping and differential cell adhesion: a sufficient model for the control of certain cell culture morphologies. J. Cell Sci., 16:401–419. Miettinen, A., Virtanen, I., and Linder, E. (1978) Cellular actin and junction formation during reaggregation of adult rat hepatocytes into epithelial cell sheets. J. Cell Sci., 31:341–353. Moghe, P.V., Berthiaume, F., Ezzell, R.M., Toner, M., Tompkins, R.G., and Yarmush, M.L. (1996) The role of culture matrix configuration and composition in maintenance of hepatocyte polarity and function. Biomaterials, 17:373–385. Murphy-Erdosh, C., Yoshida, C.K., Paradies, N., and Reichardt, L.F. (1995) The cadherin-binding specificities of B-cadherin and LCAM. J. Cell Biol., 129:1379–1390. Navarro, P., Ruco, L., and Dejana, E. (1998) Differential localization of VE- and N-cadherins in human endothelial cells: VE-cadherin competes with N-cadherin for junctional localization. J. Cell Biol., 140:1475–1484. Oliver, T., Lee, J., and Jacobson, K. (1994) Forces exerted by locomoting cells. Semin. Cell Biol., 5:139–147. Oster, G.F., Murray, J.D., and Harris, A.K. (1983) Mechanical aspects of mesenchymal morphogenesis. J. Embryol. Exp. Morphol., 78:83– 125. Parsons-Wingerter, P.A., and Saltzman, W.M. (1993) Growth versus function in the three-dimensional culture of single and aggregated hepatocytes within collagen gels. Biotechnol. Prog., 9:600–607. Peshwa, M.V., Wu, F.J., Follstad, B.D., Cerra, F.B., and Hu, W. (1994) Kinetics of hepatocyte spheroid formation. Biotechnol. Prog., 10:460– 466. Powers, M.J., and Griffith-Cima, L.G. (1996) Motility behavior of hepatocytes on extracellular matrix substrata during aggregation. Biotechnol. Bioeng., 50:392–403. Powers, M.J., Rodriguez, R.E., and Griffith, L.G. (1997) Cell- substratum adhesion strength as a determinant of hepatocyte aggregate morphology. Biotechnol. Bioeng., 53:415–426. Pujades, C., Forsberg, E., Enrich, C., and Johansson, S. (1992) Changes in cell surface expression of fibronectin and fibronectin receptor during liver regeneration. J. Cell Sci., 102:815–820. Saltzman, W.M., Parsons-Wingerter, P., Leong, K.W., and Lin, S. (1991) Fibroblast and hepatocyte behavior on synthetic polymer surfaces. J. Biomed. Mat. Res., 25:741–759. Stamatoglou, S.C., and Hughes, R.C. (1994) Cell adhesion molecules in liver function and pattern formation. FASEB J., 8:420–427. Stamatoglou, S.C., Hughes, R.C., and Lindahl, U. (1987) Rat hepatocytes in serum-free primary culture elaborate an extensive extracellular matrix containing fibrin and fibronectin. J. Cell Biol., 105:2417– 2425. Stamatoglou, S.C., Rui-Chang, G., Mills, G., Butters, T.D., Zaidi, F., and Hughes, R.C. (1990a) Identification of a novel glycoprotein (AGp110) involved in interactions of rat liver parenchymal cells with fibronectin. J. Cell Biol., 111:2117–2127. Stamatoglou, S.C., Sullivan, K.H., Johansson, S., Bayley, P.M., Burdett, I.D.J., and Hughes, R.C. (1990b) Localization of two fibronectinbinding glycoproteins in rat liver and primary hepatocytes. J. Cell Sci., 97:595–606. Steinberg, M.S. (1963) Reconstruction of tissues by dissociated cells. Science, 141:401–408. Steinberg, M.S. (1970) Does differential adhesion govern selfassembly processes in histogenesis? Equilibrium configurations and the emergence of a hierarchy among populations of embryonic cells. J. Exp. Zool., 173:395–434. Steinberg, M.S. (1996) Adhesion in development: an historical overview. Dev. Biol., 180:377–388. Steinberg, M.S., and Takeichi, M. (1994) Experimental specification of cell sorting, tissue spreading, and specific spatial patterning by quantitative differences in cadherin expression. Proc. Natl. Acad. Sci. U.S.A., 91:206–209. Sudhakaran, P.R., Stamatoglou, S.C., and Hughes, R.C. (1986) Modulation of protein synthesis and secretion by substratum in primary cultures of rat hepatocytes. Exp. Cell Res., 167:505–516. Takezawa, T., Yamazaki, M., Mori, Y., Yonaha, T., and Yoshizato, K. (1992) Morphological and immuno-cytochemical characterization of a hetero-spheroid composed of fibroblasts and hepatocytes. J. Cell Sci., 101:495–501. Tamkun, J.W., and Hynes, R.O. (1983) Plasma fibronectin is synthesized and secreted by hepatocytes. J. Biol. Chem., 258:4641–4647. Townes, P., and Holtfreter, J. (1955) Directed movements and selected adhesion of embryonic amphibian cells. J. Exp. Zool., 128:53–120. Vestweber, D., Ocklind, C., Gossler, A., Odin, P., Obrink, B., and Kemler, R. (1985) Comparison of two cell-adhesion molecules, uvomorulin and cell-CAM 105. Exp. Cell Res., 157:451–461. Volk, T., Cohen, O., and Geiger, B. (1987) Formation of heterotypic adherens-type junctions between L-CAM-containing liver cells and A-CAM-containing lens cells. Cell, 50:987–994.