Tyrosine kinase inhibitor suppresses coronary arteriosclerotic changes and vasospastic responses induced by chronic treatment with interleukin-1 beta in pigs in vivo.

We recently demonstrated that chronic treatment with IL-1 beta induces coronary arteriosclerotic changes and vasospastic responses to autacoids in pigs in vivo and that those responses are importantly mediated by PDGF. The receptors for PDGF and other major growth factors are known to have tyrosine kinase activity. We therefore investigated the effects of a selective tyrosine kinase inhibitor, ST 638, on those responses induced by IL-1 beta in our swine model. Intimal thickening and coronary vasospastic responses to serotonin and histamine were induced at the site of the coronary artery where IL-1 beta was chronically and locally applied. These responses were significantly suppressed in a dose-dependent manner by cotreatment with ST 638. In addition, ST 494, which is an inactive form of ST 638, did not inhibit those responses. The treatment with ST 638 alone did not affect the coronary vasoconstricting responses to the autacoids. Immunoblotting using an antibody to phosphotyrosines confirmed the inhibitory effects of ST 638 on the tyrosine phosphorylations induced by IL-1 beta. These results thus suggest that tyrosine kinase activation may play an important role in mediating the effects of IL-1 beta, while also suggesting that ST 638 has an inhibitory effect on the arteriosclerotic changes and vasospastic responses to autacoids in our swine model in vivo.

Direct identification of phosphotyrosine-containing proteins in some retrovirus-transformed cells by use of anti-phosphotyrosine antibody.

For direct identification of phosphotyrosine-containing proteins in lysates of various cells, phosphotyrosine (P-Tyr) was coupled to carrier proteins and anti-P-Tyr antibodies were raised in rabbits and mice. The antibodies were highly specific for P-Tyr and did not cross-react with phosphoserine or phosphothreonine. The mean association constant of rabbit anti-P-Tyr antibody to N-acetyl-P-Tyr was about four times that of rabbit anti-azobenzene phosphonate antibody. In addition, anti-P-Tyr antibody scarcely cross-reacted with the 5'-monophosphate of ribosyladenine or the 5'-monophosphate of ribosylinosine, whereas anti-azobenzene phosphonate antibody cross-reacted appreciably with these compounds. Anti-P-Tyr antibody immunoprecipitated three oncogenic gene products from cells transformed with Rous sarcoma virus, Fujinami sarcoma virus, and Abelson murine leukemia virus, respectively. The immunoprecipitates with anti-P-Tyr antibody from cells transformed with these three retroviruses all manifested tyrosine kinase activity including activity for phosphorylations of oncogene products. In addition to the proteins reported previously, the following new phosphotyrosine-containing proteins were immunoprecipitated from the respective retrovirus-transformed cells by anti-P-Tyr antibody: Mr 230,000, 74,000, and 24,000 proteins (Rous sarcoma virus); Mr 230,000, 69,000, and 24,000 proteins (Fujinami sarcoma virus); and Mr 230,000, 62,000, and 54,000 proteins (Abelson murine leukemia virus).

Expression of platelet-derived growth factor-independent phenotypes in BALB/c 3T3 cell variant with high susceptibility to chemically or physically induced neoplastic cell transformation: dissociation from activation of protein kinase C.

A31-I-13, a clonal cell variant of nontransformed BALB/c 3T3 that is highly susceptible to chemically or physically induced malignant cell transformation but is not sensitive to cell killing or susceptible to induced somatic cell mutation compared with another less transformation-susceptible A31-I-1 cell variant, was previously found to be constitutively competent [platelet-derived growth factor (PDGF)-independent] to synthesize DNA (M. Tatsuka et al., J. Cell. Physiol., 139: 18-23, 1989). The present study has demonstrated that density-arrested, quiescent A31-I-13 cells autonomously exhibit disruption of actin filamentous bundles and perturbations of dynamic morphology. PDGF induced these cytoskeletal modulations in quiescent A31-I-1 cells, which require PDGF for the induction of DNA synthesis. Furthermore, the cytoskeletal modulations of quiescent A31-I-13 cells were not accompanied by an increased production of plasminogen activators, activation of protein kinase C, or phosphorylation of a Triton X-100-soluble protein (molecular weight, 90,000) known as 80K, a major substrate for protein kinase C. However, these modulations were accompanied by the tyrosine phosphorylation of Triton X-100-insoluble (cytoskeletal) proteins with molecular weights of 24,000, 32,000-33,000, and 36,000. These Triton X-100-insoluble proteins, as well as the 80K protein, were phosphorylated by the exposure of quiescent A31-I-1 cells to PDGF. Thus the pathway for producing the transformation-susceptible phenotype in A31-I-13 appears to coincide with the PDGF signaling pathway but does not involve the protein kinase C pathway.

Specific inhibitors of tyrosine-specific protein kinases: properties of 4-hydroxycinnamamide derivatives in vitro.

Inhibition by seven synthetic 4-hydroxycinnamamide derivatives, ST 271, ST 280, ST 458, ST 494, ST 633, ST 638, and ST 642, of tyrosine-specific protein kinases (tyrosine kinase) of oncogene or proto-oncogene products (p130gag-v-fps, p70gag-actin-v-fgr, pp60v-src, pp60c-src) and epidermal growth factor (EGF) receptor kinase were investigated. ST 638 (alpha-cyano-3-ethoxy-4-hydroxy-5-phenylthiomethylcinnamamide) strongly inhibited more of the tyrosine kinases than any of the other compounds. The susceptibilities of these tyrosine kinases to ST 638 increased in the following order: EGF receptor greater than p70gag-actin-v-fgr greater than pp60c-src greater than p130gag-v-fps, pp60v-src, with 50% inhibitory concentration values of 1.1, 4.2, 18, 70, and 87 microM, respectively. The phosphorylation of the tyrosine residues in particulate fractions from RR1022 cells expressing pp60v-src was inhibited by ST 638 in a dose-dependent way, while it had a negligible effect on the phosphorylations of threonine and serine residues. Kinetic analysis showed that ST 638 competitively inhibited the phosphorylation of an exogenous substrate by the EGF receptor kinase with a Ki of 2.1 microM. ST 638 noncompetitively inhibited autophosphorylation by EGF receptor kinase. These results indicate that ST 638 is a potent and specific inhibitor of tyrosine kinases in vitro, and that its inhibitory activity is caused by competing with the substrate protein for the tyrosine kinase binding site.

Association of protein tyrosine phosphorylation with B cell differentiation induced by 12-O-tetradecanoylphorbol-13-acetate (TPA).

We have assessed whether tyrosine protein kinase (TPK) is involved in B cell differentiation. In vitro phosphorylation of an endogenous substrate in B cell leukemias showed that leukemic B cells at different stages of differentiation had specific endogenous substrates in tyrosine phosphorylation as well as distinct TPK activity. To clarify the relationship between TPK and the process of B cell differentiation, we studied protein tyrosine phosphorylation in two kinds of leukemic B cells, which showed distinct responses to TPA (12-O-tetradecanoylphorbol-13-acetate) in B cell differentiation. TPA-treated leukemic B cells from patients with B cell chronic lymphocytic leukemia (B-CLL) differentiated into cytoplasmic immunoglobulin (clg)+ plasmacytoid cells, while TPA-treated leukemic B cells from patients with hairy cell leukemia (HCL) did not differentiate into clg+ cells, but showed a peculiar morphological change, spreading. Untreated B-CLL cells and HCL cells showed similar TPK activities and tyrosine protein phosphorylation. When treated with TPA, enhanced phosphorylation was seen in B-CLL cells, while a clear reduction in phosphorylation was found in HCL cells. However, using 4-hydroxycinnamide derivatives which reduce TPK activity, we found that only the reduction of TPK activity did not lead HCL cells to spreading. These data suggest that protein tyrosine phosphorylation and/or dephosphorylation might be involved in B cell differentiation, but only the change of TPK activity in HCL cells is not sufficient to induce effects.

Tyrosine kinase inhibitor suppresses the (re) stenotic changes of the coronary artery after balloon injury in pigs.

OBJECTIVE: Restenosis after percutaneous transluminal coronary angioplasty (PTCA) still remains a serious late complication. Many growth factors induced in restenotic lesions may be responsible for restenosis after PTCA. Most of the receptors for such growth factors possess tyrosine kinase activity. This study was designed to determine whether or not a specific tyrosine kinase inhibitor, ST 638, can prevent (re)stenotic changes of the coronary artery after balloon injury. METHODS: A segment of the porcine coronary artery was aseptically wrapped with cotton mesh absorbing either ST 638 or vehicle, followed by balloon injury. Two weeks after the procedure, coronary stenosis and vasoconstricting responses were examined by coronary arteriography and (re)stenotic changes of the coronary artery were histologically examined. Antiphosphotyrosine immunoblotting was also performed to examine the inhibitory effects of ST 638. RESULTS: Coronary arteriography showed the development of mild stenotic lesions at the balloon-injured sites, where hyperconstrictive responses were repeatedly induced by intracoronary serotonin and histamine. Histologically, neointimal formation was noted at the balloon-injured site, where the total vessel area also tended to decrease (geometric remodeling). The treatment with ST 638 suppressed all the hyperconstrictive responses, the neointimal formation, and the geometric remodeling induced by balloon injury. Immunoblotting for phosphotyrosine proteins demonstrated the elevation of proteins at the balloon-injured site, which was suppressed by ST 638. CONCLUSIONS: These results indicate that tyrosine kinases are activated at the balloon-injured site and the inhibition of such kinase activities is effective in reducing both the (re)stenotic changes (neointimal formation and geometric remodeling) and the hyperconstrictive responses of the coronary artery after balloon injury.

Lipocortin I (annexin I) is preferentially localized on the plasma membrane in keratinocytes of psoriatic lesional epidermis as shown by immunofluorescence microscopy.

Lipocortin I (LPC-I, also called annexin I) is a 35-kD protein that binds phospholipids and actin in a Ca(++)-dependent manner. It is also a major substrate for EGF receptor/kinase and protein kinase C, and a putative inhibitor of phospholipase A2, which produces chemical mediators to cause inflammation. Psoriasis (PS) is an inflammatory skin disease characterized by a rapid turnover of keratinocytes and a defect in keratinization with increased activities of phospholipase C and A2, and EGF receptor. To understand the mechanism of the PS lesion formation and the function of LPC-I, its distribution was studied in the epidermis of PS, subacute eczema and normal skin, and in tumor cells of seborrheic keratosis and Bowen's disease. This study involved immunofluorescence and immunoblotting using affinity-purified polyclonal and monoclonal antibodies specific to LPC-I and to its Ca(++)-bound form. In normal, nonlesional PS and subacute eczema epidermis, LPC-I was detected mainly in the cytoplasm of the suprabasal cells, although it was on the inner aspects of the plasma membrane in some parts of the granular layer. In lesional epidermis of PS, it was localized mainly on the inner aspects of the plasma membrane, but not in the cytoplasm of the whole suprabasal cells as the Ca(++)-bound form, indicating a preferential localization on the plasma membrane. This membrane-binding of LPC-I was also observed in seborrheic keratosis, but not in Bowen's disease. These results suggest that the binding of LPC-I to the plasma membrane occurs actually in living cells, plays a role, not necessarily disease specific, in the PS lesion formation, and has some relevance to normal or abnormal differentiation of keratinocytes.

A 32-kDa protein associated with phospholipase A2-inhibitory activity from human placenta.

Two monomeric 32-kDa proteins, termed 32K-I (pI 5.8) and 32K-II (pI 5.1), were isolated from human placenta, which was solubilized by a Ca2+-chelator. Only 32K-I was associated with PLA2-inhibitory activity. CNBr peptide mapping indicated that 32K-I was distinct from 32K-II and two 36-kDa proteins, called calpactin I and II or lipocortin II and I, which have been shown to possess PLA2-inhibitory activity. 32K-I bound to PS in a Ca2+-dependent manner. 32K-I was detected in many tissues except brain, cardiac and skeletal muscle.

Establishment of cell lines derived from ataxia telangiectasia and xeroderma pigmentosum patients with high radiation sensitivity.

Four human fibroblast cell lines, three of which were derived from a patient with ataxia telangiectasia and the other from a patient with xeroderma pigmentosum, were established after transfection with cloned SV40 DNA. These 4 cell lines showed some phenotypes characteristic of neoplastically transformed cells, and had a human karyotype with heteromorphisms identical to those of the parental fibroblasts. Their sensitivity to the cytotoxic effects of gamma-rays or ultraviolet irradiation was as high as those of their parental fibroblasts.

[Cellular factors participating in cell transformation].

To clarify the cellular target(s) of onc gene products of Moloney murine sarcoma virus (Mo-MSV), we isolated eight mutant cells that exhibit temperature-sensitivity for transformation by wild type Mo-MSV from F2408 (Fischer rat cell line). These mutant cells showed normal growth and normal Mo-MSV production at the nonpermissive temperature, suggesting that cellular and viral replication are not affected by these cellular mutations. These mutant cells are divided into 3 classes on the bases of temperature-sensitivity for transformation by other tumor viruses such as Kirsten sarcoma virus, Rous sarcoma virus (SRD strain), Fujinami sarcoma virus and Abelson leukemia virus. These results have suggested that several kinds of cellular factors are concerned in transformation by Mo-MSV.

High-level expression of human lipocortin I in the fission yeast Schizosaccharomyces pombe using a novel expression vector.

We have developed a novel expression system that allows the fission yeast, Schizosaccharomyces pombe, to be used for the efficient overproduction of heterologous proteins. As an example of the utility of this system, human lipocortin I was expressed to 50 percent of soluble protein, and 150 mg of highly purified material was obtained from 10 grams of wet cell paste. Expression of lipocortin I was driven by the human cytomegalovirus (hCMV) promoter in a vector that also contains a neomycin resistance gene (neo) under the control of the SV40 early promoter, permitting selection for increasing copy-number with increasing concentrations of the antibiotic G418. The purified protein was equivalent to its native counterpart with respect to antigenicity and biochemical properties such as phospholipase A2 inhibition, actin binding and N-terminal acetylation. We have also used this system to produce comparable amounts of other proteins including rat arginase, rat NDP-kinase and human interleukin-6.

A p60v-src-related tyrosine kinase in the acetylcholine receptor-rich membranes of Narke japonica: association and dissociation of phosphatidylinositol kinase activity.

We have isolated a tyrosine-specific protein kinase from the acetylcholine receptor (AChR)-rich membranes of the electric ray Narke japonica. The enzyme is immunologically related to p60v-src, the product of the transforming gene of Rous sarcoma virus. A substantial phosphatidylinositol (PI) kinase activity was associated with this enzyme when it was purified through tyrosine-agarose affinity chromatography used previously for the purification of p60v-src. However, by subsequent chromatography on casein-agarose, most of the associated PI kinase activity was separated from the tyrosine kinase activity. The results suggest that the tyrosine-specific protein kinase in the AChR-rich membranes of N. japonica has no intrinsic PI kinase activity.

Calcium-induced intracellular cross-linking of lipocortin I by tissue transglutaminase in A431 cells. Augmentation by membrane phospholipids.

Covalently cross-linked multimers of lipocortin I are shown to be present in human epidermoid carcinoma A431 cells treated with epidermal growth factor or the calcium ionophore A23187. This intracellular cross-linking of lipocortin I is suggested to be mediated by the action of tissue transglutaminase, a Ca2(+)-dependent protein cross-linking enzyme. Cross-linking of lipocortin I competes with proteolytic digestion of the protein, and pretreatment of the cells with inhibitors for calpain (Ca2(+)-dependent intracellular protease) markedly enhanced the cross-linking of lipocortin I. Cross-linked lipocortin I is shown to be present in the soluble fraction of A431 cells as well as in the particulate fraction; a 34-kDa fragment of lipocortin I was solubilized successfully by plasmin digestion of the latter fraction. Immunofluorescence microscopy using specific antilipocortin-I antibody showed that cross-linked lipocortin I forms an envelope-like structure, which is not extracted with [ethylenebis(oxyethylenenitrilo)]tetraacetic acid (EGTA) or Triton X-100. In vitro incubation of purified lipocortin I with tissue transglutaminase resulted in the formation of covalently cross-linked lipocortin I dimer, tetramer, and so on. Amine incorporation and cross-linking studies using lipocortin I and its N-terminal truncated derivatives indicated that the cross-linking site is localized within the plasmin-susceptible N-terminal 29 amino acids of lipocortin I. The cross-linking of lipocortin I is shown to be accelerated more than 10 times by the addition of phosphatidylserine vesicles, on which lipocortin I molecules are most likely aligned in a conformation suitable for cross-linking. Collectively, these findings suggest that an increase of intracellular calcium concentration results in the attachment of lipocortin I onto the plasma membrane phospholipids through the C-terminal domain of the molecule where the membrane-bound lipocortin I is cross-linked by the action of tissue transglutaminase through the N-terminal domain.

A pathogenic autoantibody, pemphigus vulgaris-IgG, induces phosphorylation of desmoglein 3, and its dissociation from plakoglobin in cultured keratinocytes.

Pemphigus vulgaris (PV) is an autoimmune blistering skin disease, which is characterized by autoantibodies to a specific desmosomal constituent, i.e. desmoglein 3 (Dsg3). In this study, we analyzed phosphorylation of desmosomal proteins and their molecular interactions after PV-IgG binding to Dsg3 using DJM-1 cells, a squamous cell carcinoma cell line, and normal human keratinocytes. Cells were metabolically labeled with 32P inorganic phosphate, followed by stimulation with the IgG fractions from five PV patients or normal individuals for 20 min. Phosphorylation of specific desmosomal components and their molecular interactions were studied in immunoprecipitates using PV-IgG and anti-plakoglobin (PG) antibodies. PV-IgG binding alone induced the phosphorylation of Dsg3 at serine residues. Although Dsg3 and PG were coprecipitated by PV-IgG-immunoprecipitation when treated with normal IgG, PG was not coprecipitated with Dsg3 when stimulated with PV-IgG, suggesting that PV-IgG binding to Dsg3 caused the dissociation of Dsg3 from PG. These results demonstrate that the binding of pathogenic PV autoantibodies to the cell surface antigen Dsg3, which is an adhesion molecule categorized into desmosomal cadherins, caused particular phosphorylation of Dsg3 and its dissociation from PG.

Linear relationship between diphosphorylation of 20 kDa light chain of gizzard myosin and the actin-activated myosin ATPase activity.

The addition of large amounts of myosin light chain kinase to the reconstituted gizzard actomyosin shows diphosphorylation of 20 kDa myosin light chain. Accompanying diphosphorylation, the actin-activated myosin ATPase activity was also enhanced. The extent of diphosphorylation and the myosin ATPase activity were clearly demonstrated to be in a linear relationship. From the time course experiment, the conversion of monophosphorylated light chain into one which was diphosphorylated seemed to be a sequential process. Moreover, analyzing phospho-amino acid by using a two-dimensional electrophoresis technique revealed that monophosphorylated light chain contained phosphoserine and diphosphorylated one contained phosphothreonine in addition to phosphoserine.

Isolation and characterization of three forms of 36-kDa Ca2+-dependent actin- and phospholipid-binding proteins from human placenta membrane.

We purified three forms of 36-kDa proteins, two monomeric 36-kDa proteins, which had pIs of 7.5 (36K-I) and 6.4 (36K-II), and one 36-kDa complex (36K-C) consisting of two subunits, 36-kDa (pI 7.5) and 12-kDa (pI 5.8), from human placenta membrane. The 36-kDa subunit of 36K-C was identical to 36K-I as judged by pI, cyanogen bromide peptide mapping and immunological cross-reactivity. The three proteins showed F-actin- and phosphatidylserine-binding abilities dependent on Ca2+ concentrations at millimolar and micromolar levels, respectively. They all had phospholipase A2 inhibitory activity. Only 36K-II was phosphorylated extensively at tyrosine residue in Ca2+- and EGF- dependent manners in the membrane fraction of A431 cells. 36K-I was the best substrate for src kinase, whereas 36K-II was the best for fps kinase. However, 36K-C was not phosphorylated by any kinases used here.

Effects of vanadate on tyrosine phosphorylation and the pattern of glycosaminoglycan synthesis in rabbit chondrocytes in culture.

The present study examined the effects of high doses of vanadate on glycosaminoglycan (GAG) synthesis and tyrosine phosphorylation in rabbit chondrocytes in confluent cultures. Although 6 microM vanadate increased the incorporation of [3H]glucosamine into chondroitin sulfate proteoglycans twofold, 40-60 microM vanadate suppressed this incorporation fourfold. Although 6 microM vanadate had little effect on [3H]glucosamine incorporation into hyaluronate, 40-60 microM vanadate increased this incorporation threefold. Chemical analyses confirmed that the increase in [3H]glucosamine incorporation into hyaluronate and the decrease in the incorporation into chondroitin sulfate proteoglycan correlated with increased hyaluronate content and decreased chondroitin sulfate content in the cell layers of vanadate-transformed cells. Chondrocytes exposed to 40-60 microM vanadate became typically transformed spindlelike cells. Furthermore, vanadate, at 6 and 60 microM, increased the overall level of phosphotyrosine by 8- and 31-fold, respectively, and 60 microM vanadate enhanced phosphorylation of many phosphotyrosine-containing proteins. These observations suggest that vanadate induces transformation-associated changes in the pattern of GAG synthesis when it induces excess phosphorylation on tyrosine in chondrocyte proteins.

Characterizations of two distinct Ca2+-dependent phospholipid-binding proteins of 68-kDa isolated from human placenta.

Two distinct 68-kDa proteins, named 68K-I (pI 6.4) and 68K-II (pI 5.6), were solubilized from human placenta by treatment with 5 mM EGTA. On DE52 cellulose column chromatography at pH 7.4, 68K-I in the EGTA eluate was recovered in the unadsorbed fractions, whereas 68K-II was retained on the column and eluted with 0.2 M NaCl. The 68K-I protein was obtained in more than 95% purity by further hydroxylapatite and cation exchange chromatographies, while the 68K-II protein was purified further by gel filtration and hydroxylapatite chromatographies. Partial amino acid sequence data showed that 68K-I protein was a novel protein which shared the same sequences as lipocortin I and that 68K-II was the same as human p68/67-kDa calelectrin (Crompton, M. R., Owens, R. J., Totty, N. F., Moss, S. E., Waterfield, M.D., and Crumpton, M. J. (1988) EMBO J. 7, 21-27; Südhof, T. C., Slaughter, C. A., Leznicki, I., Barjon, P., and Reynolds, G. A. (1988) Proc. Natl. Acad. Sci. U.S.A. 85, 664-668). The two proteins bound to acidic phospholipids, phosphatidylserine, and/or phosphatidylinositol, but not to phosphatidylcholine, in the presence of micromolar levels of Ca2+. 68K-I bound to phosphatidylinositol preferentially to phosphatidylserine, whereas 68K-II bound only to phosphatidylserine. Both 68K-I and 68K-II inhibited phospholipase A2 activity, and the inhibition by 68K-II was detectable only in the presence of 100 mM KCl. 68K-I, but not 68K-II, was found to bind to F-actin in a Ca2+-dependent (1 mM) manner. Moreover 68K-I, but not 68K-II, was phosphorylated in vitro at tyrosine residues by fps kinase and by epidermal growth factor receptor/kinase, the latter reaction being dependent on Ca2+ and epidermal growth factor. Western blot analysis with affinity purified anti-68K-I and anti-68K-II antibodies showed that 68K-I was located in only certain tissues, especially human placenta, whereas 68K-II was present in many human and rat tissues.

A rat mutant cell clone showing temperature-dependent transformed phenotypes with functional expression of the src gene product.

The cellular mutant B814 isolated from a Fischer rat cell line shows temperature-sensitivity of focus formation on infection with Moloney murine sarcoma virus (Mo-MSV) and Rous sarcoma virus (RSV). An RSV-transformed clone (S814-2) isolated from B814 cells shows temperature-sensitive transformed phenotypes for morphology, growth in soft agar, and glucose uptake. The expression, phosphorylation, and tyrosine kinase activity of pp60v-src in S814-2 were not affected at the nonpermissive temperature, and virus rescued from this clone had wild-type transforming ability, suggesting that a cellular factor altered in S814-2 is responsible for the cellular steps of transformation after the function of pp60v-src. In addition, the cellular 36K protein, a possible candidate as a target of pp60v-src, was phosphorylated at the nonpermissive temperature in S814-2, indicating that phosphorylation of the 36K protein is not correlated with transformed phenotypes.

A digital image processing system for quantitating dynamic morphology in cultured mammalian cells.

An automated, video-driven system has been developed which can quantitate dynamic cell morphology in cultured mammalian cells. This system is based upon the Personal Image Analysis System and is assisted by a video-enhanced contrast microscopy with a computer-aided digital image processing unit and a time-lapse video technique. Various parameters for cell motility including locomotion (vectorial translation) and accompanying shape changes can be simultaneously analyzed. Here, we describe this system and demonstrate its application in Balb/c 3T3 cell culture. This system represents a new tool for exploring subtleties of mammalian cell behavior.