The role of cell-cell contact in "contact" inhibition of cell division: a review and new evidence

The term “contact inhibition of cell division” was borrowed from “contact inhibition of cell movement.” We prefer the term “postconfluence inhibition of cell division” as being more operational and less mechanistically biased; it is operationally defined as a pronounced depression of the mitotic rate in a postconfluent culture which displays a stationary density despite periodic nutrient renewal, the inhibition being locally reversibly by removal of the adjacent cells. The mechanism of postconfluence inhibition is of considerable interest because of the inverse correlation between postconfluence inhibition and the tumorigenicity of a number of cell lines. Several hypotheses, involving direct cell-to-cell contacts or locally restricted diffusion gradiens, could explain postconfluence inhibition. With the goal of discriminating among these hypotheses, time-lapse films were taken of carefully regulated, perfused cultures of 3T3 mouse cells, in which the transition from rapid growth to the stationary phase was recorded. Measurements of cell-to-cell contact, local cell density, and generation times were made on an individual cell level and analyzed with the aid of a computer. We observed that all-around cell-cell contact or a high local cell density present throughout G1 often did not produce immediate inhibition of cell division. We conclude that either (i) simple visible cell-cell contacts or a high local cell density are not the direct cause of postconfluence inhibition of cell division, or (ii) their effects often do not inhibit cell division until after a delay of about one cell generation time. Such a delay may be partly responsible for the 50% overshoot past the stationary density that we observed in 3T3 cultures.     

Contact regulation of cell division in an epithelial-like cell line

The rate of cell division in an epithelial-like cell line, 1S1, was examined by time-lapse cinemicrography. When precautions were taken to insure a sufficient nutrient supply, the number of mitoses per unit time in any given area of a confluent monolayer remained constant. This “contact regulation of cell division” resulted in a steadily decreasing frequency of mitosis per cell as the culture became crowded. With the decrease was associated a gradual change in cell shape, from maximally flattened to maximally compact, due to contact inhibition of the movement of cells across one another. When cells were removed along a line scraped on a dense culture, the cells at the edge of the scrape flattened, migrated into the vacant area, and subsequently increased their frequency of mitosis to that characteristic of non-confluent cells. Inhibition of mitosis caused by a limitation on the nutrient supply was also reversed at a line-scrape. These observations suggest that cell flattening promoted mitosis by causing the cell membrane to expand, thereby facilitating the uptake of nutrients. The cell membrane would thus function in the mechanism of contact regulation as a transducer, for converting the pressure of the surrounding cell population into a restraining force upon the metabolism of cell division.     

Contact inhibition of what? An analytical review

Quite a number of phenomena having to do with cells' influences upon one another's movements have come to be regarded as expressions of “contact inhibition.” However, no single, central mechanism has been shown to underlie them all. Consequently, the term “contact inhibition” should not be used without operational modifiers. Inhibitions of individual cell movements imputed to be mediated by cell-cell contacts include inhibition of overlapping (which results in monolayering), of colony expansion, of cell speed (nuclear translocation), of ruffling, of orthogonal movement (proposed to explain spontaneous parallel alignment of cells), and of neighbor exchanges. The six inhibitions listed above are operationally distinct, and only two (overlapping and colony expansion) are known to result from a common mechanism. A seventh phenomenon, so-called “contact inhibition of cell division” (more operationally termed postconfluence inhibition of cell division) is in a separate category and is not considered here. Evidence eliminating action-at-a-distance is available only for the first three, and hence only these should at present be termed contact inhibitions. Inhibition of neighbor exchanges is yet hypothetical; at its extreme, it would immobilize cells in a confluent monolayer, but such immobilization has been found not to occur. Contact inhibition of overlapping, the most studied of the six, is not displayed by invasive cells with respect to normal cells; invasive tumor cells overlap freely upon normal cells, although not necessarily upon one another. Contact inhibition of overlapping, and its loss by invasive cells, can readily be interpreted, by means of the differential adhesion hypothesis, as consequences of cell-type-specific differences in cell-cell and cell-substratum “strengths of adhesion.” These strengths of adhesion are formulated as specific interfacial free energies, which are the only parameters of cellular adhesiveness that have been rigorously shown to determine equilibrium configurations of cell populations.     

Regulation of protein secretion in Chinese hamster ovary cells by cell cycle position and cell density : Plasminogen activator, procollagen and fibronectin

To study the relationship between cell growth control, cell contact, and protein secretion, we examined the production of plasminogen activator, procollagen, and fibronectin by Chinese hamster ovary (CHO) fibroblasts, both as a function of position in the cell cycle and as a function of cell density. CHO fibroblasts that were synchronized at hourly intervals throughout the cell cycle by mitotic selection in an automated roller bottle apparatus secreted plasminogen activator only during the G2 and M phases of the cell cycle (10–14 h after mitotic selection). Cell-associated plasminogen activator activity was variable during G1 and S, but was greatly reduced during G2 and M. In contrast, secretion of the connective tissue matrix proteins, procollagen and fibronectin, was controlled by cell density rather than by cell cycle position. Type III procollagen and fibronectin were secreted throughout the cell cycle with no pronounced variations. Type I procollagen was not secreted by cycling cells and was observed in confluent cultures only after 24–48 h. To correlate these changes in protein secretion patterns with cell shape and contact, we used scanning electron microscopy (SEM) to study the appearance of CHO cells after mitotic selection. Actively dividing cells retained a high proportion of rounded, ruffled, and blebbed cells during all phases of the cell cycle. Only with increased cell density in contact-inhibited confluent cultures did most cells begin to flatten and spread. Thus, secretion of and attachment to extracellular matrix did not occur in rapidly dividing cells, but appeared to require the increased cell-cell contact and spreading that accompanies contact inhibition of growth. On the other hand, increased secretion of plasminogen activator was directly related to cell division and may be part of a sequence of events that allows cells growing in culture to loosen extracellular attachments in preparation for rounding and cytokinesis.     

Flattening,movement and control of division of epithelial-like cells

The kinetics of cell division and movement in four epithelial-like cell lines, grown in continuously perfused culture medium, were studied by time-lapse cinemicrography. One line exhibited “contact regulation of cell division,” so that the rate of mitosis per cell decreased steadily as population density increased. In the other three lines mitosis was not controlled as a function of population density until the cells became very crowded. An explanation for this difference was sought in terms of the hypothesis that the rate of division depends on the area of the cell membrane. Cells of the contact-regulated line flattened uniformly on the substrate. Their motility was restrained by adhesion between their borders. As they crowded together, contact inhibition of cell overlap caused a steady decrease in average surface area per cell. All three of the non-controlled lines also had contact inhibition of overlap. Cells of two of them flattened on the substrate; but these cells had little mutual adhesion and were highly motile, so that they continually changed their shapes. The areas of their cell membranes were therefore not subject to a restraint that could control the rate of division. Cells of the fourth line remained rounded or only slightly flattened during culture growth, so that no change in cell membrane area occurred that could change the rate of division.     

Population density and regulation of cell division in 3T3 cells. I. Inorganic phosphate levels, uptake and release.

Triggering mechanisms for initiating density dependent inhibition of cell division in 3T3 cell monolayers are activated approximately two to three population doublings prior to cessation of cell division at monolayer confluency. This activation occurs at a critical contact cell density of approximately 8 X 10(3) cells/cm2. During this period there are selective controls on transport and storage of required low molecular weight nutrients. A possible correlation between orthophosphate and rates of cell division has been investigated. We have demonstrated a relationship between cellular concentrations of orthophosphate and initiation of density dependent inhibition of cell division. Prior to critical intercellular contact, the [Pi] in 3T3 is 10 mM. During critical contact, this concentration is quickly reduced to approximately 2 mM and remains at this concentration to confluency. Similar alterations do not occur in Py 3T3 cells, which maintain a concentration of approximately 2 mM Pi regardless of cell density. After confluent 3T3 cells are released from inhibition of cell division the [Pi] must increase several-fold before DNA synthesis commences. These are physiological changes in 3T3 cellular [Pi] as a function of cell density, and cannot be attributed to nutrient depletion, altered transport of Pi into the cell, increased [ATP], or increased [PPi] levels. The controlled modulation of [Pi] may regulate glycolysis and coordinate counter-ion changes (Ca++) may regulate mitochondrial activity.     

Cell division, ciliary regeneration and cyclic AMP in a unicellular system

The average speed of nuclear translocation of 3T3 cells, recorded in a time-lapse film of a perfused culture, was negatively correlated with the number of contacting cells, and, to a lesser degree, with the amount of a cell's perimeter in contact with other cells. When a cell was in contact with five or more other cells, its speed was reduced by 50%, on the average, although the variation in individual cell speed was considerable at each level of contact. A partial correlation analysis showed that any extracellular soluble factors governed by the local cell density had little or no effect on speed, relative to the prominent effect of the number of cell-cell contacts, and hence that 3T3 cells display true contact inhibition of speed. This confirms the original demonstration by Abercrombie and Heaysman (1952), who studied chick embryo heart fibrpolasts. In our study, the relationships between average speed and age of the culture was such that a possible independent contribution of a time-associated factor other than contact to the diminution in average speed, although not necessary to account for the data, could not be excluded. The same intercellular contacts found to inhibit speed in this study were previously reported to cause no immediate prolongation of individual cell generation times, despite the fact that the filmed culture was undergoing so-called “contact” inhibition of cell division. In the present study, moreover, no correlation was observed between the average speeds of individual cells and their generation times. Hence, postconfluence inhibition of cell division and contact inhibition of speed of cell movement seem to be independent phenomena.     

Enhanced rates of fluid pinocytosis during exponential growth and monolayer regeneration by cultured arterial endothelial cells

Rates of fluid pinocytosis by bovine aortic endothelial cells were measured during various manipulations of growth status in vitro. Sparsely seeded cultures grew exponentially until a confluent monolayer was formed, at which time growth slowed. This change in growth rate coincided with a decline in the rate of pinocytosis to about one-third that in the growing cultures. During the subsequent attainment of maximal cell density in the confluent monolayer, the pinocytic rate remained constant. There was close correlation between ³H-thymidine labelling indices, as measured by autoradiography, and the rates of pinocytosis. Mechanical “wounding” of the confluent monolayer resulted in cell migration and proliferation. Twenty-four hours after “wounding,” rates of pinocytosis per mg. cell protein were significantly enhanced. When regeneration of the monolayer was blocked by cytochalasin B, pinocytosis remained at the same rate as in the uninjured, confluent monolayer. These experiments support, and extend to endothelium, earlier observations that in growing cells pinocytosis proceeds at a higher rate than in non-growing, quiescent cells. Furthermore, they raise the possibility that the transendothelial transport of macromolecules such as lipoproteins by receptor-in-dependent fluid pinocytosis in vivo may be altered by the growth status of the endothelium.     

Contact inhibition, polyribosomes, and cell surface membranes in cultured mammalian cells

An “overlay” method for rapidly and synchronously inducing contact inhibition in normal cultured cells has been developed. Using this method, disaggregation of cytoplasmic polyribosomes has been observed to occur within a matter of hours after overlay, followed by a decrease in cellular ribosomal RNA. Polysome disaggregation was influenced by the extent of cell-cell interaction and was inhibited by pretreatment of overlay cells with cycloheximide. Treatment of underlay cells with cytosine arabinoside also induced polysome disaggregation, but only after an appreciable lag as compared to that observed in overlaid cultures. Disaggregation could be induced by this method in cultured cells derived from normal tissue but not in cells derived from cancerous tissue. Polysome synthesis in growing “normal” cells (as measured by incorporation of tracer uridine into RNA) was markedly decreased when a cell surface membrane preparation was added to cultures.     

Uniform hypothesis of cell behavior--movement, contact inhibition of movement, adhesion, chemotaxis, phagocytosis, pinocytosis, division, contact inhibition of division, fusion.

The cell has been represented as a charged liquid drop. Contrary to the DLVO-theory, the effect of the surface potential upon the value of the interfacial tension of the cell membrane has also been taken into consideration. The cell membrane has visco-elastic properties and its constituents may move against each other. Cell movement is caused by the appearance of a small number of the electrically charged constituents of the cell membrane on the leading edge of the cell. This produces a local decrease in the surface tension and the cell membrane expansion. At the moment of contact between two cells proton transfers occur between the strongly negatively charged microvilli of one cell and the body of the other, analogous to a condenser breakdown. This, through the effect on the surface tension, causes contact inhibition of movement. The distribution of the proton dissociable groups modifies the interaction between the cells (differentiation) and between the cell and the substratum (adhesion). Adsorption of the charged compounds at the surface of the cell membrane, decreasing the surface potential and increasing the surface tension, causes the phenomena of chemotaxis, phagocytosis and pinocytosis. Cell division, considered in the terms of the surface energy, requires an adequate supply of considerable quantities of energy inversely proportional to the surface potential value. In case of a reduction of the distance between the cells, their surface potential and the energetic barrier of the cell division processes increases, and causes contact inhibition of cell division. Due to their high charge, division of neoplastic cells is inhibited much later than division of normal cells, or is completely ininhibited due to geometric conditions. Fusion of the cell membrane in the intra-cellular and intercellular processes is a reverse process in relation to the cell division.     

Phosphorylation-dephosphorylation of retinoblastoma protein not necessary for passage through the mammalian cell division cycle

Phosphorylation of the retinoblastoma protein (Rb) during the G1-phase of the mammalian cell division cycle is currently believed to be a controlling element regulating the passage of cells into S-phase. We find, however, that the suspension-grown cell lines U937, L1210, and MOLT-4 contain exclusively hyperphosphorylated Rb. Furthermore, when adherent NIH3T3 cells are grown at very low densities to avoid overgrowth and contact inhibition, they also contain only hyperphosphorylated Rb. NIH3T3 cells exhibit hypophosphorylation when the cells are grown at moderate to high cell densities. We propose that cultures of adherent cells such as NIH3T3, when grown to moderate cell densities, are made up of two populations of cells: (a) cells that are relatively isolated and therefore growing exponentially without contact inhibition, and (b) cells that are growth-inhibited by local cell density or contact inhibition. The common observation in adherent cell lines, that Rb is both hyper- and hypophosphorylated in the G1-phase and only hyperphosphorylated in the S- and G2-phases, is explained by the effects of cell density and contact inhibition. Thus, phosphorylation-dephosphorylation of Rb protein during the G1 phase is not a necessary process during the NIH3T3, L1210, MOLT-4, and U937 division cycles. We propose that phosphorylation-dephosphorylation of Rb is independent of the division cycle and is primarily determined by growth conditions throughout the division cycle.     

Cell-cell contact and growth regulation of pinocytosis in 3T3 cells

In sub-confluent cultures of Balb/c-3T3 cells, pinocytosis rates were increased after exposure to specific growth factors (serum; platelet-derived growth factor, PDGF; epidermal growth factor, EGF). Conversely, as cells became growth-inhibited with increasing culture density, there was a corresponding decline in pinocytosis rate per cell. In order to test whether density-inhibition of pinocytosis was influenced either by the growth cycle or by cell contact independently of growth, cells were induced into a quiescent state at a range of subconfluent and confluent densities. Under such conditions, cell density did not significantly inhibit pinocytosis rate. When confluent quiescent cultures in 2.5% serum were exposed to 10% serum, the resulting round of DNA synthesis was accompanied by enhanced pinocytosis per cell, even though the cells were incontact with one another. Furthermore, in a SV40-viral transformed 3T3 cell line, both the growth fraction and the pinocytosis rate per cell remained unchanged over a wide range of culture densities. These studies indicate that density-dependent inhibition of pinocytosis in 3T3 cells appears to be secondary to growth-inhibition rather than to any direct physical effects of cell–cell contact.     

Impact of cell growth morphology on retroviral transduction: effect of contact inhibition

Retrovirus-mediated gene transfer is one of the most commonly used methods to deliver, integrate, and express the gene of interest because the retrovirus can insert the desired gene into the chromosome of the target cells with high stability. However, to deliver the gene successfully, the retrovirus requires active division to integrate reversely transcribed DNA into the chromosome of target cells. In this study, we focused on the effect of cell-cell contact inhibition on the efficiency of retroviral transduction with two anchorage-dependent cell lines: NIH 3T3 and 293 cells. These two cell lines have very different cell morphologies and growth patterns on surfaces. Human embryonic kidney epithelial 293 cells tend to stick together after dividing, while NIH 3T3 cells migrate to occupy available surface and spread. Experimental data indicate that the abatement of the transduction rate of 293 cells was initiated in the early stage of the culture, whereas effect of contact inhibition of NIH 3T3 cells on the transduction rate became dominating at the end of the culture period. Experimental results were also quantitatively illustrated by plotting normalized multiplicity of infection (MOI) versus normalized cell density. According to the outcomes, cell inoculation density plays an important role in optimizing the retroviral transduction rate. The optimal time of retroviral transduction should be confined to the accelerating growth phase for 293 cells and at the exponential growth phase for NIH 3T3 cells. The implication drawn from this study is that contact inhibition effect on retroviral transduction should be taken into account for large-scale gene transfer systems such as the microcarrier bioreactor.     

Optical waveguide lightmode spectroscopy (OWLS) to monitor cell proliferation quantitatively

The use of microscopic observations used for in situ monitoring of cell proliferation in the production of epidermal autografts is not satisfactory. In particular, the identification of the projected cell area from microscopic pictures by image analysis (IA) depends on intensity edges and level of contrasts and is thus limited to subconfluent cultures. Some of these problems can be solved by using optical waveguide lightmode spectroscopy (OWLS), which measures the effective refractive index of a thin layer above an Si(Ti)O(2) waveguide surface. In this study the use of OWLS to monitor cell adhesion, spreading, and growth was studied. The sensitivity of the method was investigated by using three different cell lines, two fibroblasts and one hepatoma cell line. Cell proliferation of two strains of fibroblasts and hepatoma cells was monitored up to 2 days with the OWLS. In parallel, cell density was determined at different time points microscopically using an additional window in the measuring chamber. The cell density of fully spread cells ( approximately 4 h after attachment) was found to be proportional to the OWLS signal. In long-term cultures the influence of the cell density from single cells to confluent cell cultures upon the OWLS signal was investigated. The exponentially growing number of hepatoma resulted in a linear increase of the sensor signal. Due to this and to the fact that the proliferating cells exhibit contact inhibition, it was concluded that the cell contact area must decrease exponentially. The results show the strength of OWLS for monitoring the adhesion and proliferation of anchorage-dependent cells in applications where an on-line indicator of the total biomass is needed. Additionally, OWLS provides metabolic information through detection of the cell mass in close contact with the waveguide.     

Regulation of the cell cycle of 3T3 cells in culture by a surface membrane-enriched cell fraction

Addition of a suspension of a surface membrane enriched fraction prepared from confluent 3T3 cells to sparse 3T3 cells in culture results in a concentration dependent and saturable decrease in the rate of DNA synthesis. The inhibition of cell growth by membranes resembles the inhibition of cell growth observed at confluent cell densities by a number of criteria: (1) In both cases the cells are arrested in the G₁ protion of the cell cycle; (2) the inhibition by membranes or by high local cell density can to a large extent be compensated for by raising the serum concentration or by addition of fibroblast growth factor plus dexamethasone. Membranes prepared from sparse cultures inhibit less well than membranes from confluent cultures in a manner which suggests that binding of membranes to cells is not by itself sufficient to cause inhibition of cell growth. The inhibitory activity has a subcellular distribution similar to phosphodiesterase (a plasma membrane marker) and appears to reside in one or more intrinsic membrane components. Maximally, membranes can arrest about 40% of the cell population in each cell cycle. Plasma membranes obtained from sparse 3T3 cells are less inhibitory than membranes obtained from confluent cells. This suggests either that the inhibitory component(s) in the plasma membrane responsible for growth inhibition may be in part induced by high cell density, or that this component(s) may be lost from these membranes during purification.     

A mammalian somatic "cell cycle" mutant defective in G1

Variants or “mutants” temperature-sensitive (ts) for growth have been isolated by selection from a near-diploid mouse cell line. Thus far. 10 ts mutants which grow normally at 33° C, but not at 39° C, have been isolated. These ts mutants were then studied to determine if any manifested their defect at a unique point or stage in the cell cycle. This type of ts mutant is termed a “cell cycle” mutant. The first screen involves observing individual cells of an asynchronous culture for residual division after a shift from 33° C (permissive temperature) to 39° (nonpermissive temperature). A cell cycle mutant should show some fraction of the cells dividing only once at a normal rate after the shift. The ts variant B54 met this first criterion for a cell cycle mutant (i.e., 50% residual division) and was further analyzed. The second screening technique monitors (1) the rate of entry into S, (2) the length of G₂, and (3) the rate and duration of cells entering mitosis after a shift of an asynchronous culture to 39°. This experiment with B54 revealed that cells in G₁ at the time of the shift to 39° failed to enter S while cells already into S completed the cycle at 39°. These results suggest that B54 is defective in a G₁ function which is required for entry into S, but which is no longer needed once cells have entered S. Other results are presented which also support this hypothesis. In addition the ts function of B54 is apparently required for recovery from a “high density” G₁ arrest.     

Migration and proliferation of diploid human fibroblasts following "wounding" of confluent monolayers

Six diploid human fibroblast strains were grown in confluent monolayers. Holes were scraped in these monolayers and the number of cells proliferating into these “wounds” with time were determined. The migration and mitotic aspects of the proliferation of fibroblasts into these wounds were analyzed separately. Small amounts of undialysed or dialysed serum were essential for cell division but not migration. Saline extracts of skin could not substitute for serum in the medium. Neither zinc nor cupric ion at tolerable concentrations (10^?5M) increased the rate of cell proliferation. Normal human fibroblasts did not immediately start to divide from confluency into the “wound” space. Their generation time was about 32–39 hours. Fibroblasts from patients with cystic fibrosis began to divide almost immediately into the “wounded” area. Their generation time was about 48 to 56 hours.     

Effect of cell growth on hepatitis C virus (HCV) replication and a mechanism of cell confluence-based inhibition of HCV RNA and protein expression

An intimate relationship between hepatitis C virus (HCV) replication and the physiological state of the host liver cells has been reported. In particular, a highly reproducible and reversible inhibitory effect of high cell density on HCV replication was observed: high levels of HCV RNA and protein can be detected in actively growing cells but decline sharply when the replicon cells reach confluence. Arrested cell growth of confluent cells has been proposed to be responsible for the inhibitory effect. Indeed, other means of arresting cell growth have also been shown to inhibit HCV replication. Here, we report a detailed study of the effect of cell growth and confluence on HCV replication using a flow cytometry-based assay that is not biased against cytostasis and reduced cell number. Although we readily reproduced the inhibitory effect of cell confluence on HCV replication, we found no evidence of inhibition by serum starvation, which arrested cell growth as expected. In addition, we observed no inhibitory effect by agents that perturb the cell cycle. Instead, our results suggest that the reduced intracellular pools of nucleosides account for the suppression of HCV expression in confluent cells, possibly through the shutoff of the de novo nucleoside biosynthetic pathway when cells become confluent. Adding exogenous uridine and cytidine to the culture medium restored HCV replication and expression in confluent cells. These results suggest that cell growth arrest is not sufficient for HCV replicon inhibition and reveal a mechanism for HCV RNA inhibition by cell confluence.     

Effect of contact inhibition on the regulation of cholesterol metabolism in cultured vascular endothelial cells.

Cholesterol synthesis in actively growing bovine vascular endothelial cells is regulated by low density lipoprotein (LDL) at a step prior to mevalonate formation, in a manner comparable to that found in aortic smooth muscle cells. LDL uptake by these cells is associated with induction of cholesterol esterification, an increase in total cell cholesterol, and an inhibition of endogenous sterol synthesis. In contrast, cholesterol metabolism in confluent contact-inhibited endothelial cultures was not significantly affected by LDL even though the cells bind the lipoprotein at high affinity receptor sites. Lysosomal degradation and subsequent regulatory effects on cellular cholesterol metabolism, however, were observed in contact-inhibited endothelial cells incubated with cationized rather than native LDL. Cationized LDL enter the cells independently of the high affinity sites. Therefore, the primary regulation of cholesterol metabolism in these cells is neither through the appropriate intracellular enzymes nor through the high affinity surface receptors, but via an inhibition of LDL internalization. It is suggested that this inhibition is due to a strict contact-inhibited morphology which enables the endothelium of the larger arteries to function as a selective barrier to the high circulating levels of plasma LDL.