A serum factor requirement for the passage of cultured Vero cells through G2.

When Vero cells, a line derived from and African Green Monkey kidney, are grown under conditions where the saturation density is limited by serum, they deplete the growth medium of a factor necessary for cell division. The factor is a component of serum. When Vero cells are plated at low density (2 X 10(4)/cm2) in this depleted growth medium (after dialysis against serum-free Dulbecco's Modified Eagle's Medium) they initiate an unbalanced program of growth. Protein synthesis proceeds at the same rate as parallel cells in fresh serum, and and the cells accumulate protein as a function of time. DNA synthesis is also initiated in these cells, and the amount of DNA per cell increases for the next four days plating. However the cells quickly stop dividing. Measurements of DNA per cell using microspectrofluorometry show that the cells are accumulating in the late S and G2 period during this time. Thus we conclude that these cells cannot pass through a transition point in G2. When fresh serum is added to cells after three days in depleted growth medium, they divide before they begin to synthesize DNA. This further confirms that they are in late S and G2. Cell division is promoted in Vero cells in depleted growth medium by bovine fetuin, and to a lesser extent by bovine albumin. Cell division is not promoted by insulin, hydrocortisone, dexamethasone, linolenic acid, calcium, and typsin inhibitor form ovomucoid. From these data we conclude that transit through G2 requires the prescence of an extracellular factor.     

Mitogenic activity of pituitary hormones on cell cultures of normal and carcinogen-induced tumor epithelium from rat mammary glands

Cell suspension containing normal or tumor epithelium were readily obtained by enzymatically digesting rat mammary glands from perphenazine-treated (prolactin-hypersecreting) cycling, female virgin animals or hormone- responsive mammary tumors from animal treated with dimethylbenzanthracene. Cell suspensions were fractioned into predominantly epithelial and predominantly stromal cells by their differential rates of attachment to culture dishes. Both normal mammary and tumor epithelial cells were characterized by the presence of specific cell-junctional complexes, desmosome-like structures, surface microvilli, and their ability to synthesize casein. Serum-dependent protease activity was greater in cultures derived from tumors, and cells from such cultures grew in agarose whereas those from the non-neoplastic gland did not. The addition of prolactin to the culture medium stimulated DNA synthesis in primary or secondary epithelial cultures from tumors, whereas additional insulin and hydrocortisone with prolactin were required for similar levels of DNA synthesis in cultures from non-neoplastic glands. The fraction of cells synthesizing DNA was, however, smaller than that with 10 percent serum measured in the same time period. Both growth hormone and epidermal growth factor stimulated DNA synthesis but to a lesser extent than did prolactin. Prolactin with hydrocortisone and insulin were relatively inactive in promoting DNA synthesis of the nonepithelial cells whereas pituitary fibroblast growth factor was more active. These mitogenic effects were obtained when the hormones were added to the medium at near physiological concentrations, and paralleled the known activities of the hormones in control of mammary gland growth and development in the rat.     

The relationship of serum fibronectin and cell shape to thrombin-induced inhibition of dna synthesis in human fibroblasts

In a previous report it was shown that inhibited DNA synthesis and altered morphology resulted when human fibroblasts (HF) were plated in 3% fetal calf serum (FCS) medium preincubated with thrombin (Hall and Ganguly, 1980a). This was in contrast to the stimulatory effects of this enzyme when added to cells several hours after subculture. Those observations suggested that thrombin may act upon serum components of the growth medium necessary for initial culture establishment following cell plating. In this report, the relationship of serum fibronectin (FN) to this thrombin-mediated inhibitory phenomena was investigated. It was found that the development of altered morphology and inhibited DNA synthesis could be completely prevented by the addition of this glycoprotein to medium preincubated with thrombin. Cell shape and DNA synthesis appeared to be closely related and both parameters showed a dose-dependent sensitivity to added fibronectin. To further investigate this, a technique was developed in which cell shape could be selectively varied and DNA synthesis measured in the absence of serum or thrombin. These studies indicated that cell shape was closely related to DNA synthesis and morphologies identical to that seen in thrombin-treated medium were produced. As observed in the thrombin system, normal cellular appearance and DNA synthesis could be restored by the addition of fibronectin. The results of this work suggest that thrombin acts upon medium components necessary for normal morphological development, possibly fibronectin, in cells following subculture. Inhibited DNA synthesis and growth seem to arise as a direct consequence of this effect.     

Regulation of mitotic activity and the cell cycle in primary chick muscle cells by neurotransferrin

We previously demonstrated that neurotransferrin (NTF), a transferrin extracted from adult chicken peripheral nerves, promotes growth of primary chick muscle cells in the absence of embryo extract. NTF was shown to stimulate DNA synthesis and cell proliferation. In the present study, we demonstrate that NTF is a mitogen using two independent methods; counts of orcein-stained mitotic figures and analysis of cell cycle kinetics with a fluorescence-activated cell sorter. In low-density cultures mitotic activity increases with increasing doses of NTF followed by a plateau at concentrations greater than 6 μg/ml. Residual, embryonic mitotic activity progressively declines with time after plating muscle cells in the absence of NTF. Absence of NTF for 2 days causes cells to lose irreversibly their myogenic potential. In the presence of NTF, mitotic activity increases for 2 days followed by a decline concurrent with myoblast fusion and formation of myotubes. Cell cycle analysis showed that NTF addition causes cell populations to shift from G_t to S and G₂ + M within 18.5 hr. Muscle cells, plated at high densities in the absence of NTF, show mitotic activities similar to those plated at low densities in the presence of NTF. Addition of NTF to high-density cultures is ineffective in stimulating mitosis. These studies show that at typical cell plating densities, NTF is a required mitogen for primary chick muscle cell cultures.     

Beta-type transforming growth factor specifies organizational behavior in vascular smooth muscle cell cultures

In culture, vascular smooth muscle cells (SMC) grow in a "hill-and-valley" (multilayered) pattern of organization. We have studied the growth, behavioral organization, and biosynthetic phenotype of rat aortic SMC exposed to purified platelet-derived growth regulatory molecules. We show that multilayered growth is not a constitutive feature of cultured SMC, and that beta-type transforming growth factor (TGF-beta) is the primary determinant of multilayered growth and the hill-and-valley pattern of organization diagnostic for SMC in culture. TGF-beta inhibited, in a dose-dependent manner, the serum- or platelet-derived growth factor-mediated proliferation of these cells in two-dimensional culture, but only when cells were plated at subconfluent densities. The ability of TGF-beta to inhibit SMC growth was inversely correlated to plating cell density. When SMC were plated at monolayer density (5 X 10(4) cells/cm2) to allow maximal cell-to-cell contact, TGF-beta potentiated cell growth. This differential response of SMC to TGF-beta may contribute to the hill-and-valley pattern of organization. Unlike its effect on other cell types, TGF-beta did not enhance the synthesis of fibronectin or its incorporation into the extracellular matrix. However, the synthesis of a number of other secreted proteins was altered by TGF-beta treatment. SMC treated with TGF-beta for 4 or 8 h secreted markedly enhanced amounts of an Mr 38,000-D protein doublet whose synthesis is known to be increased by heparin (another inhibitor of SMC growth), suggesting metabolic similarities between heparin- and TGF-beta-mediated SMC growth inhibition. The data suggest that TGF-beta may play an important and complex regulatory role in SMC proliferation and organization during development and after vascular injury.     

Hyaluronic acid bonded to cell-culture surfaces stimulates chondrogenesis in stage 24 limb mesenchyme cell cultures

The influence on the differentiation of stage 24 chick limb mesenchymal cells of hyaluronic acid (HA) covalently bonded onto plastic substrates has been examined. Under control conditions, stage 24 cells express phenotypes related to the initial plating density: When plated at high density (5 X 10(6) cells/35-mm culture dish), these cells express a chondrogenic phenotype collectively visualized as a mound or nodule of cartilage. Cartilage nodules are not found in cultures plated at intermediate or low densities, 2 X 10(6) and 1 X 10(6) cells/35-mm dish, respectively. However, when cells are plated onto HA surfaces, expression of the cartilage phenotype occurs at all three plating densities in roughly comparable frequencies. This increase in cartilage nodule formation does not appear to be due to an increased plating efficiency or increased replication rate. The observed effect is dependent on HA concentration; with an increase in bound HA, an increase in the number of cartilage nodules is observed. Digestion of HA substrates with hyaluronidase abolishes the stimulation in chondrogenesis, while no effect is observed if the HA substrates are treated with either trypsin or alkaline borohydride. No other glycosaminoglycan, except for the HA analog, unsulfated chondroitin, exhibits this unique stimulation of chondrogenic expression. While the rate of radiolabeled sulfate incorporation is dramatically increased with cells plated onto HA substrates, the protein biosynthetic rate, as evidenced by radiolabeled proline incorporation, remains unaffected. This dramatic increase in chondrogenic expression is considered in contrast to the previously reported inhibitory effect of HA substrates on myogenesis. These observations suggest that HA may have a regulatory role in the chondrogenic differentiation of chick limb mesenchymal cells.     

Maintenance of the keratocyte phenotype during cell proliferation stimulated by insulin

Keratocytes normally express high levels of aldehyde dehydrogenase and keratocan. They proliferate and lose their keratocyte markers when they become fibroblastic during corneal wound healing. Keratocytes cultured in fetal bovine serum also become fibroblastic, proliferate, and lose these markers. In this report, we studied the effects of three serum growth factors, fibroblast growth factor-2, insulin, and platelet-derived growth factor-BB, on keratocyte proliferation and the maintenance of the keratocyte markers in 7-day cultures in cells plated at low (5,000 cells/cm2) and high (20,000 cells/cm2) density in serum-free medium. Keratocyte proliferation was measured by [3H]thymidine incorporation and by DNA content of the cultures. Cytosolic aldehyde dehydrogenase and keratocan accumulated in the medium were quantified by Western blot. The results showed that all the growth factors stimulated proliferation, but insulin stimulated proliferation more consistently. The keratocyte markers aldehyde dehydrogenase and keratocan were maintained after 7 days in culture in all growth factors, but keratocyte cell morphology was only maintained in medium containing insulin. Most of the proteoglycans were degraded in cultures of keratocytes plated at low density and cultured in the absence of growth factors. This degradation was prevented when keratocytes were cultured in the presence of the growth factors or when keratocytes were plated at high density. The results of this study show that insulin can expand keratocytes in vitro, maintain their phenotype, and prevent proteoglycan degradation.     

Initiation of primary cell cultures from human intracranial tumors on extracellular matrix from bovine corneal endothelial cells

Tissue specimens from 105 human gliomas and 57 human meningiomas were obtained at surgery, dissociated into single cells and small cell aggregates and then plated onto plain plastic tissue culture dishes and dishes which had been precoated with an extracellular matrix (ECM) derived from bovine corneal endothelium. In 80% of the glioma cases we observed a marked improvement in initial plating efficiency, colony formation and speed of attachment when cells were plated on ECM. In 5 cases cells attached only to the ECM-coated dishes but remained afloat in the untreated dishes. In addition it could be noted that over the first 2 days, those cells which had been initiated on ECM showed more signs of morphological differentiation, i.e., extension of cytoplasmic processes or formation of fiber networks between cell groups. If adaptation occurred and proliferation began in vitro, either immediately or after a several days' lag phase, both the ECM-cultured cells as well as those which slowly had adapted to culture on plastic could be passed on to untreated culture ware and perpetuated thereon. In the case of well-differentiated low-grade gliomas where no growth in culture took place, the cultures on ECM could at least be used for initial experiments in the primary cultures (P0). Meningiomas usually attached well to both, plastic or ECM. In 50% of our cases the plating efficiency was higher on ECM but after successful initial culture, the delay until the cells on plastic reached confluence in comparison with those on ECM was 1 or 2 days. Again there were 2 cases in which the cells would not plate on plastic. Here the cells which after 1 day were still afloat plated to more than 80% within the first 2 h after transfer to ECM. In all cases the cells from plastic and ECM cultures were indistinguishable and could be passed onto untreated dishes henceforth. In later culture stages ECM offers several advantages: It is easier to shift cells to serum-free defined culture conditions, the cells will grow at a faster rate on ECM when in higher passages and the maximal number of passages possible is higher on ECM.     

Multiplication of Swiss 3T3 cells in a serum-free medium

Gently trypsinized Swiss 3T3 cells inoculated into medium MCDB 402 attach readily to polylysine-coated surfaces and remain viable for several days in the absence of exogenously added protein. Short-term multiplication under defined conditions can be obtained by supplementing the MCDB 402 with fibroblast growth factor (FGF), insulin (INS), and dexamethasone (DEX). Addition of bovine plasma fibronectin further improves attachment and viability. This system does not require initial plating in serum or the addition of poorly defined extracts for cellular attachment or for multiplication. In the complete system minus FGF, cells plated at a low density attach to the culture surface and become quiescent. The addition of FGF or PDGF 48–72 h after plating stimulates a high level of DNA synthesis during the following 24 h. EGF also stimulates DNA synthesis in these cells, but to a lesser extent. Insulin and dexamethasone are not needed for the initial DNA synthesis response to FGF, but are needed for continuing multiplication over a period of several days. This system provides a means for studying the effects of specific mitogens on Swiss 3T3 cells in the absence of undefined supplements, and without complications due to density-dependent inhibition.     

Effect of insulin at dilution endpoint concentrations on macromolecular synthesis in normal mouse mammary and human breast cancer epithelial cells

Employing defined media conditions, the insulin sensitivities of mouse mammary gland epithelial cells in primary culture and MCF-7 human mammary epithelial cells were determined. Insulin stimulated the rates of (3H] uridine incorporation into RNA and [3H] leucine incorporation into protein in both primary mouse mammary gland epithelial cell cultures and MCF-7 cell cultures at concentrations approximating the dilution endpoint of the hormone (10-21 M). Insulin stimulated the rate of [3H] thymidine incorporation into DNA in primary mouse mammary gland epithelial cells at the dilution endpoint concentrations. However, MCF-7 cells required insulin concentrations 100-1000-times that necessary in mouse mammary epithelial cultures to elicit an increased rate of [3H] thymidine incorporation into DNA. Evidence is presented which suggests that the increased rates of uptake of 3H- uridine, [3H] thymidine and [3H] leucine into their respective precursor pools is not responsible for the apparent stimulation of RNA, DNA and protein synthesis.     

Initiation of DNA synthesis in mammary epithelium and mammary tumors by lithium ions

The growth promoting effects of lithium and insulin on cultures of mammary gland epithelium and CZF mouse mammary tumor cells were investigated. Lithium chloride exerts a 450-fold increase in the rate of DNA synthesis in mammary epithelium from mid-pregnant mice in organ culture or monolayer culture. There is an increase in both the percentage of cells initiating DNA synthesis and the net accumulation of DNA. The most effective lithium concentration is 10 mM, and the maximally effective rate of stimulation is reached 48 hours after addition. The magnitude of response to lithium varies with the physiological state of the mammary epithelial cell donor: epithelium from non-pregnant or lactating mice is less responsive than that from mid-pregnant mice. In combination, insulin and lithium produce either a synergistic or an additive effect on the growth of epithelium dependent upon the physiological state of the donor animal. Lithium also promotes the growth of mammary tumor cells in the absence of serum or other mitogens. The action of lithium on DNA synthesis appears to be a direct effect on the epithelial cells.     

Basal serum-free culture system which supports growth of fetal rat adrenal cells in primary monolayer cell culture

A serum-free culture system has been developed which allows primary cultures of fetal rat adrenal cells to divide in response to insulin 10 micrograms/l, yet retain their ability to respond to adrenocorticotropic hormone (ACTH) by increased corticosterone production. The essential components of this culture system are: (1) a high initial plating number of 125-130 x 10(3) cells/cm2; (2) a low oxygen concentration of 2.5% O2; (3) an initial plating period of 48 h in greater than or equal to 5% (v/v) fetal bovine serum. In contrast to reported requirements for the serum-free growth of adult adrenal cells, the fetal adrenal cells do not require the provision of any special matrix or a competence factor for serum-free growth or steroidogenic activity. This suggests that they are capable of rapidly generating their own matrix and competence factor in primary culture.     

DNA synthesis in primary cultures of adult rat hepatocytes in a defined medium: effects of epidermal growth factor, insulin, glucagon, and cyclic-AMP

Epidermal growth factor (EGF) especially in combination with insulin and glucagon, has been shown to stimulate DNA synthesis in liver cells, both in the whole animal and in cell cultures. As a further development we have found that in primary monolayer cultures of freshly isolated adult rat liver parenchymal cells, in which contamination with nonparenchymal cells was negligible, DNA synthesis was substantially stimulated by these substances. In control cultures, incorporation of [3H]thymidine into DNA and labeling of nuclei in autoradiographs was low. The stimulation by EGF was enhanced by insulin and glucagon, whereas these hormones by themselves exhibited only limited activity. These observations were made in cultures of hepatocytes that were never exposed to serum, even during cell isolation and plating. Hence for stimulation of DNA synthesis under these conditions neither serum factors nor interactions with other types of cells or their products were required. The effects of glucagon were reproduced by substances that elevate intracellular concentration of cyclic-AMP, including cholera toxin, isoproterenol, and methylisobutylxanthine. These various substances, especially EGF, glucagon, or cyclic-AMP, altered the morphological characteristics of the cultures during early stages, promoting cellular spreading and aggregation.     

Bidirectional concentration-dependent effects of glucagon and dibutyryl cyclic AMP on DNA synthesis in cultured adult rat hepatocytes

Glucagon and dibutyryl cyclic AMP exerted both stimulatory and inhibitory effects on hepatocyte DNA synthesis when added to primary monolayer cultures in the presence of serum, dexamethasone, insulin and epidermal growth factor. The stimulation occurred at low concentrations of glucagon (1 pM-1 nM) or dibutyryl cyclic AMP (1 nM-1 microM), while the agents inhibited DNA synthesis at higher concentrations (usually glucagon at over 10 nM or dibutyryl cyclic AMP at over 10 microM). The stimulatory effect was stronger at low cell densities (less than 20 X 10(3) hepatocytes/cm2). When the hepatocytes were cultured at higher densities, stimulatory effects were reduced or absent and the inhibition of (hormone-induced) DNA synthesis by a high concentration of glucagon was much more pronounced than at low cell densities. These results indicate dual, bidirectional, effects of cyclic AMP on hepatocyte DNA synthesis.     

Plating Efficiency for Primary Hamster Embryo Cells as an Index of Efficacy of Fetal Bovine Serum for Cell Culture

Attachment and growth of mammalian cells plated at low cell density require optimum conditions for the cells to form colonies. Reliability, reproducibility, and validity of the plating efficiency test for evaluating cell culture sera were determined by measuring the plating efficiency of 37 lots of fetal bovine serum obtained from 8 suppliers (5 lots from each of 7, 2 lots from 1 supplier), by using hamster embryo fibroblasts plated at low cell density. The test revealed considerable variation between lots of serum and between suppliers. The five lots from some suppliers had consistently high plating efficiencies, whereas one or more lots from other suppliers had quite low efficiencies. The results were reproducible in repeated tests, and control experiments indicated that the test measured the efficiency of the test serum independently of the efficiency of the serum used for the primary outgrowth of the hamster embryo cells.     

Oligonucleotides antisense to catalytic subunit of cyclic AMP-dependent protein kinase inhibit mouse mammary epithelial cell DNA synthesis

Mouse mammary epithelial cells were plated onto 24-well culture plates (50,000 per well), allowed to attach and serum starved for 24 h. Following serum starvation, DNA synthesis was induced by the addition of 10% fetal calf serum and determined by a 1-h pulse with [3H]thymidine from 17 to 18 h after serum addition. Addition of oligonucleotides antisense to the translation start region of cyclic AMP-dependent protein kinase (kinase A) mRNA inhibited thymidine incorporation into DNA (total or percentage of cells incorporating thymidine, as measured by autoradiography). This inhibition was apparent whether compared to controls with no oligonucleotide addition, sense oligonucleotides, or mismatch oligonucleotides. Enzymatic assays indicated that the antisense oligonucleotides lowered kinase A activity in cells. Time course studies indicated that the inhibition in DNA synthesis was not an artifact of the time at which DNA synthesis was estimated. Long-term (4 day) cultures indicated that effects on induction of DNA synthesis were reflected in long-term cell proliferation.     

Participation of plasma membrane proteins in the formation of tight junction by cultured epithelial cells

Measurements of the transepithelial electrical resistance correlated with freeze-fracture observations have been used to study the process of tight junction formation under various experimental conditions in monolayers of the canine kidney epithelial cell line MDCK. Cells derived from previously confluent cultures and plated immediately after trypsin- EDTA dissociation develop a resistance that reaches its maximum value of several hundred ohms-cm(2) after approximately 24 h and falls to a steady-state value of 80-150 ohms- cm(2) by 48 h. The rise in resistance and the development of tight junctions can be completely and reversibly prevented by the addition of 10 μg/ml cycloheximide at the time of plating, but not when this inhibitor is added more than 10 h after planting. Thus tight junction formation consists of separable synthetic and assembly phases. These two phases can also be dissociated and the requirement for protein synthesis after plating eliminated if, following trypsinization, the cells are maintained in spinner culture for 24 h before plating. The requirement for protein synthesis is restored, however, if cells maintained in spinner culture are treated with trypsin before plating. Actinomycin D prevents development of resistance only in monolayers formed from cells derived from sparse rather than confluent cultures, but new mRNA synthesis is not required if cells obtained from sparse cultures are maintained for 24 h in spinner culture before plating. Once a steady-state resistance has been reached, its maintenance does not require either mRNA or protein synthesis; in fact, inhibition of protein synthesis causes a rise in the resistance over a 30-h period. Following treatments that disrupt the junctions in steady- state monolayers recovery of resistance also does not require protein synthesis. These observations suggest that proteins are involved in tight junction formation. Such proteins, which do not turn over rapidly under steady-state conditions, are destroyed by trypsinization and can be resynthesized in the absence of stable cell-cell or cell-substratum contact. Messenger RNA coding for proteins involved in tight junction formation is stable except when cells are sparsely plated, and can also be synthesized without intercellular contacts or cell-substratum attachment.     

Repair of potentially lethal damage in unfed plateau phase cultures of Ehrlich ascites tumour cells. II. Monolayer cultures

Monolayer cultures of EAT cells when plated immediately after irradiation show a desrease in survival as they "age" in the plateau phase of growth. This decrease, which is manifest as a diminution of the shoulder width of the survival curve down to values approaching zero, is reversible if the cells are kept in their growth medium for some hours after irradiation before trypsinization and plating. Survival curves obtained by this holding procedure are similar in shape to those shown by exponentially growing or early plateau phase cells. We interpret this effect in terms of repair of potentially lethal damage which occurs after immediate plating in young cultures but only declared during plating in cultures which have "aged" in the plateau phase. The kinetics of this repair and the effects caused by the addition of serum after irradiation in the cultures have been studied.     

In vitro culture of cryopreserved bovine mammary cells on collagen gels: synthesis and secretion of casein and lactoferrin

The preparation, cryopreservation, and culture on type I collagen gels of lactating bovine mammary cells with prolonged milk protein synthesis and secretion in vitro is described. Cryopreserved cells prepared as acinar fragments from either lactating or developing mammary glands attached to the collagen substratum within 24-48 hr after plating in serum and hormone supplemented medium. During continued culture in hormone-supplemented (insulin, cortisol, and prolactin) serum-free medium outgrowth of cells from the attached acinar fragments was observed beginning on day 2, with continued outgrowth to near confluence by day 6. Two morphologically distinct cell types were evident; initial outgrowth was by large polygonal cells that were subsequently overlain by spindle-shaped cells. Cells from both lactating and developing mammary glands sustained substantial milk protein secretion for at least 14 days in culture. Alpha S1-casein synthesis and secretion in cultures of lactating mammary cells was dependent on a critical minimum cell population density, below which alpha S1-casein was not secreted. In contrast, lactoferrin (LF) secretion into the medium increased linearly with the increase in cell population density. Cells cryopreserved up to 16 months secreted LF at levels comparable to fresh cultures of the same cells.     

Stimulation of growth of primary cultured adult rat hepatocytes without growth factors by coculture with nonparenchymal liver cells

DNA synthesis of adult rat parenchymal hepatocytes alone in primary culture can be stimulated only by the addition of humoral growth factors to the culture medium. However, when parenchymal hepatocytes were cocultured with nonparenchymal liver cells from adult rats, their DNA synthesis was markedly stimulated in the absence of added growth factors or calf serum. DNA synthesis of parenchymal hepatocytes was not stimulated by conditioned medium from nonparenchymal liver cells and was greatest when the parenchymal cells were plated on 24-h cultures of nonparenchymal liver cells. A dead feeder layer of nonparenchymal cells was almost as effective as a feeder layer of viable nonparenchymal cells. These results suggest that the stimulation of DNA synthesis in parenchymal hepatocytes was not due to some soluble factors secreted by nonparenchymal liver cells but to an insoluble material(s) produced by the nonparenchymal liver cells. This insoluble material(s) was collagenase- and acid-sensitive, suggesting that it was a protein containing collagen. The effect of nonparenchymal liver cells was specific: coculture with hepatoma cells, liver epithelial cells, or Swiss 3T3 cells did not stimulate DNA synthesis in parenchymal hepatocytes. Added insulin and epidermal growth factor showed additive effects with nonparenchymal cells in the cocultures. These results suggest that DNA synthesis in parenchymal hepatocytes is stimulated not only by various humoral growth factors but also by cell-cell interaction between parenchymal and nonparenchymal hepatocytes, possibly endothelial cells. This cell-cell interaction may be important in repair of liver damage and liver regeneration.