A glycoprotein in the plasma membrane matrix as a major potential substrate of p60v-src.

A potential substrate of p60v-src in Rous sarcoma virus-transformed cells was found to be a 130-kilodalton (kDa) glycoprotein which binds to lectin-Sepharose and can be immunoprecipitated by an anti-phosphotyrosine antibody. This glycoprotein was shown to be distinct from the fibronectin receptor and a cellular protein phosphorylated in p60v-src immune complexes. The protein was a transmembrane protein localized in the plasma membrane and resistant to extraction with Triton X-100. The 130-kDa protein was also highly phosphorylated in cells transformed by Fujinami sarcoma virus or Y73 but not in cells infected with Rous sarcoma virus mutants that encode p60v-src lacking myristoylated N termini. Phosphorylation of this glycoprotein was temperature dependent in cells infected with temperature-sensitive mutants. The good correlation between its phosphorylation and morphological transformation, together with its relative abundance among phosphorylated proteins and its subcellular localization, suggests that phosphorylation of the 130-kDa glycoprotein is one of the primary events important for cell transformation by p60v-src and related oncogene products.     

Rous sarcoma virus mutant dlPA105 induces different transformed phenotypes in quail embryonic fibroblasts and neuroretina cells.

dlPA105 is a spontaneous variant of Rous sarcoma virus, subgroup E, which carries a deletion in the N-terminal portion of the v-src gene coding sequence. This virus was isolated on the basis of its ability to induce proliferation of quiescent quail neuroretina cells. The altered v-src gene encodes a phosphoprotein of 45,000 daltons which possesses tyrosine kinase activity. DNA sequencing of the mutant v-src gene has shown that deletion extends from amino acid 33 to 126 of wild-type p60v-src. We investigated the tumorigenic and transforming properties of this mutant virus. dlPA105 induced fibrosarcomas in quails with an incidence identical to that induced by wild-type virus. Quail neuroretina cells infected with the mutant virus were morphologically transformed and formed colonies in soft agar. In contrast, dlPA105 induced only limited morphological alterations in quail fibroblasts and was defective in promoting anchorage-independent growth of these cells. Synthesis and tyrosine kinase activity of the mutant p45v-src were similar in both cell types. These data indicate that the portion of the v-src protein deleted in p45v-src is dispensable for the mitogenic and tumorigenic properties of wild-type p60v-src, whereas it is required for in vitro transformation of fibroblasts. The ability of dlPA105 to induce different transformation phenotypes in quail fibroblasts and quail neuroretina cells is a property unique to this Rous sarcoma virus mutant and provides evidence for the existence of cell-type-specific response to v-src proteins.     

v-mil induces autocrine growth and enhanced tumorigenicity in v-myc-transformed avian macrophages

MH2, an avian retrovirus containing the v-myc and v-mil oncogenes, rapidly transforms chick hematopoietic cells in vitro. The transformed cells belong to the macrophage lineage and proliferate in the absence of exogenous growth factors. Here we analyze a series of MH2 deletion mutants and show that these two oncogenes together establish an autocrine growth system in which v-myc stimulates cell proliferation, while v-mil induces the production of chicken myelomonocytic growth factor (cMGF). We also demonstrate that these two oncogenes cooperate in vivo. MH2 efficiently induces monocytic leukemias and liver tumors, while deletion mutants lacking either a functional v-mil or v-myc do not.     

Differentiation of retrovirus-infected avian neuroretina cells

The effects of oncogenic retroviruses on the expression of differentiation markers were studied in monolayer cultures of chick and quail embryo neuroretinas. Transformation by Rous sarcoma virus (RSV) did not affect the appearance of synapses, and the expression of glutamic acid decarboxylase was stimulated by pp60v-src, the product of the src gene. Quail embryo neuroretina cells transformed by Mill Hill 2 (which contains the two oncogenes v-mil and v-myc) were induced to proliferate into permanent cultures that synthesized crystallins and produced lentoid bodies. In contrast, transformation with a temperature-sensitive mutant of RSV reversibly blocked the production of crystallins and lentoid bodies. These data show that given cellular genes can respond differently to distinct oncogenes.     

C127 cells resistant to transformation by tyrosine protein kinase oncogenes

C127 is a nontumorigenic mouse cell line widely used in in vitro transformation assays due to its normal morphological appearance and its very low levels of spontaneous transformation. We now report that C127 cells are resistant to transformation by tyrosine protein kinase oncogenes derived from growth factor receptors such as the retroviral v-fms and the human trk transforming genes. In contrast, these cells could be efficiently transformed by members of the ras oncogene family and by serine/threonine kinase oncogenes such as v-mos and v-raf. C127 cells were also found to be resistant to transformation by v-src, the prototype of a large family of tyrosine protein kinase oncogenes whose products are associated with the inner side of the plasma membrane. However, morphologically normal C127 cells expressing pp60v-src acquired a transformed phenotype upon continuous passage in vitro. Somatic cell hybrids (neoR, hygroR) obtained by fusion of G418-resistant C127 cells expressing p70trk (neoR) and hygromycin-resistant NIH3T3 cells (hygroR) exhibited transformed properties as determined by their ability to grow in semisolid agar. In contrast, no such growth was observed when these neoR p70trk-containing C127 cells were fused to control hygroR C127 cells. These results indicate that C127 cells may either lack or express insufficient levels of certain critical substrate(s) necessary for the onset of transformation by tyrosine protein kinase oncogenes.     

Temperature-sensitive mutants of MH2 avian leukemia virus that map in the v-mil and the v-myc oncogene respectively.

MH2 is an avian retrovirus that contains the v-mil and v-myc oncogenes. In vitro it transforms chick macrophages that are capable of proliferation in the absence of growth factor. Earlier work showed that v-myc induces macrophage transformation and that v-mil induces the production of chicken myelomonocytic growth factor (cMGF), thus generating an autocrine system. We describe the isolation of temperature-sensitive (ts) mutants of MH2 virus. As suggested by marker rescue experiments, one mutant bears a ts lesion in v-mil, whereas the other carries a mutation in v-myc. Ts v-mil MH2-transformed macrophages become factor-dependent at the non-permissive temperature (42 degrees C), while ts-v-myc MH2-transformed macrophages cease growing and acquire a more normal macrophage phenotype at 42 degrees C irrespective of the presence of cMGF. Both phenotypes can be reversed by backshift to the permissive temperature. These results suggest that the gene products of v-mil and v-myc function independently of each other and that v-mil is necessary for the maintenance of autocrine growth, whereas v-myc is required to maintain the transformed phenotype.     

Lack of induction of neuroretinal cell proliferation by Rous sarcoma virus variants that carry the c-src gene.

Expression of p60v-src of Rous sarcoma virus in cultured chicken embryo neuroretinal cells was previously shown to result in the transformation and sustained proliferation of normally quiescent cell populations. We show here that Rous sarcoma virus variants that encode p60c-src, the cellular homolog of p60v-src, lack the ability to induce morphological transformation and cell proliferation of cultured neuroretinal cells. Neuroretinal cells infected with c-src-containing viruses, however, possess no less p60 protein kinase activity assayed in the immune complex than those infected with the transformation-defective Rous sarcoma virus mutants PA101 or PA104, which do stimulate the growth of these cells.     

Coordinate regulation of myelomonocytic phenotype by v-myb and v-myc.

Both avian myeloblastosis virus (by the action of v-myb) and avian myelocytomatosis virus MC29 (by the action of v-myc) transform cells of the myelomonocytic lineage. Whereas avian myeloblastosis virus elicits a relatively immature phenotype, cells transformed by MC29 resemble mature macrophages. When cells previously transformed by v-myb were superinfected with MC29, their phenotype was rapidly altered to that of a more mature cell. These superinfected cells expressed both v-myb (at a level similar to that found before superinfection) and v-myc. It therefore appears that the expression of v-myc can elicit certain properties of a more differentiated phenotype. In addition, unlike cells transformed by v-myb alone, the cells expressing both v-myb and v-myc could not be induced by the tumor promoter 12-O-tetradecanoylphorbol-13-acetate to differentiate to fully mature macrophages. Cells with a morphology similar to that of the superinfected cells were elicited by simultaneously infecting yolk sac macrophages with avian myeloblastosis virus and MC29. Such cells expressed both v-myb and v-myc. These results indicate that expression of v-myb and v-myc in infected cells coordinately regulates myelomonocytic phenotype and that the two viral oncogenes vary in their ability to interfere with tumor promoter-induced differentiation. Our findings also sustain previous suggestions that the oncogenes v-myb and v-myc may not transform target cells by simply blocking differentiation.     

Transformation of chicken embryo fibroblasts and tumor induction by the middle T antigen of polyomavirus carried in an avian retroviral vector.

The middle T antigen of polyomavirus transformed primary chicken embryo fibroblasts when expressed from a replication-competent avian retrovirus. This in vitro-constructed retrovirus, SRMT1, is a variant of Rous sarcoma virus that encodes the middle T antigen in place of v-src. Inoculation of SRMT1 into 1-week-old chickens rapidly induced hemangiomas and hemangiosarcomas. As shown with mammalian cells infected with polyomavirus, polyomavirus middle T antigen appears to be associated with p60c-src in chicken cells infected with SRMT1. When lysates of SRMT1-infected cells immunoprecipitated with either a monoclonal antibody against p60src or anti-T serum were assayed in an in vitro kinase reaction, the middle T antigen was heavily phosphorylated. To see whether an excess of p60c-src could alter the extent of phosphorylation of the middle T protein or the process of cell transformation by middle T, cells were doubly infected with SRMT1 and NY501, a virus which overexpresses p60c-src. Doubly infected chicken embryo fibroblasts transformed with the same kinetics and were morphologically indistinguishable from chicken embryo fibroblasts infected with SRMT1 alone. Phosphorylation of the middle T antigen was elevated two- to fivefold relative to cells infected only with SRMT1.     

Short-term treatment with gamma interferon induces stable reversion of ras-transformed mouse fibroblasts.

Persistent revertants have been generated from NIH 3T3 cells transformed by an activated human Ha-ras gene after short-term gamma interferon treatment in the presence of the cardiac aminoglycoside ouabain. Normal fibroblastlike morphology and anchorage dependence are restored in revertants. Tumorigenicity in nude mice is abolished. The revertants continue to express high steady-state levels of the ras oncogene. Partial retransformation of reverted cells is induced after 5-azacytidine treatment or after infection with retrovirus vectors carrying the v-abl, v-fes, v-myc, or v-src oncogene. The revertants resist the transforming activities of the v-Ha-ras and v-mos oncogenes.     

Distinct and different effects of the oncogenes v-myc and v-src on avian sympathetic neurons: retroviral transfer of v-myc stimulates neuronal proliferation whereas v-src transfer enhances neuronal differentiation

Immature avian sympathetic neurons are able to proliferate in culture for a limited number of divisions albeit expressing several neuron- specific properties. The effect of avian retroviral transfer of oncogenes on proliferation and differentiation of sympathetic neurons was investigated. Primary cultures of 6-d-old quail sympathetic ganglia, consisting of 90% neuronal cells, were infected by Myelocytomatosis virus (MC29), which contains the oncogene v-myc, and by the v-src-containing Rous sarcoma virus (RSV). RSV infection, in contrast to findings in other cellular systems, resulted in a reduction of neuronal proliferation as determined by 3H-thymidine incorporation (50% of control 4 d after infection) and in increased morphological differentiation. This is reflected by increased neurite production, cell size, and expression of neurofilament protein. In addition, RSV- infected neurons, unlike uninfected cells, are able to survive in culture for time periods up to 14 d in the absence of added neurotrophic factors. In contrast, retroviral transfer of v-myc stimulated the proliferation of immature sympathetic neurons preserving many properties of uninfected cells. The neuron-specific cell surface antigen Q211 and the adrenergic marker enzyme tyrosine hydroxylase were maintained in MC29-infected cells and in the presence of chick embryo extract the cells could be propagated over several weeks and five passages. Within 7 d after infection, the number of Q211-positive neurons increased approximately 100-fold. These data demonstrate distinct and different effects of v-src and v-myc-containing retroviruses on proliferation and differentiation of sympathetic neurons: v-src transfer results in increased differentiation, whereas v- myc transfer maintains an immature status reflected by proliferation, immature morphology, and complex growth requirements. The possibility of expanding immature neuronal populations by transfer of v-myc will be of considerable importance for the molecular analysis of neuronal proliferation and differentiation.     

Cellular aging is a critical determinant of primary cell resistance to v-src transformation.

Primary cell cultures are in general resistant to the transforming effect of a single oncogene, a finding considered consistent with the multistage theory of carcinogenesis. In the present studies, we examined whether cellular age, differentiation stage, and/or tissue origin of primary cells plays a role in determining their response to v-src transformation. To study the role of cellular age, rat mammary fibroblasts were isolated from a 50-day-old female rat and infected with a recombinant retrovirus carrying a v-src gene after 2, 7, 14, 21, and 28 days of continuous growth. To determine whether cellular differentiation is important, fibroblasts were isolated from embryos at 12 and 16 days of gestation, from newborns, and from a 30-day-old rat and similarly infected. Finally, the role of primary-cell histogenesis was assessed by infecting primary cultures of fibroblasts isolated from the mammary gland, dermis, and lungs of a mature rat. When compared to 3Y1 cells, all preparations of primary cultures exhibited considerable resistance to v-src transformation. However, whereas primary cells isolated from different tissues responded similarly to the transforming effect of the oncogene, major differences were observed when cells were transduced at different stages of their in vitro life span. v-src was capable of inducing formation of foci and growth in soft agar in early-passage cells but failed to do so in primary cultures infected after 14 days of continuous passaging. Similarly, both the number of foci and the number of colonies in soft agar decreased with tissue donor age. The differential response of young and senescing cells could not be explained by mutations in v-src provirus, by differences in functional v-src expression, or by growth stimulation or suppression via paracrine mechanisms. Furthermore, v-src cooperated with an immortalizing gene, like simian virus 40 large T, polyomavirus large T, E6 and E7 of human papillomavirus, or an activated p53 mutant, to induce anchorage-independent growth of primary cultures but failed to do so with cytoplasmic transforming genes, like v-abl, v-ras, or v-raf, which did not confer indefinite division potential. These studies indicate that cellular aging is a critical determinant of primary-cell resistance to v-src transformation. It is suggested that v-src requires a nuclear auxiliary function for transformation which is present in early-passage cells, particularly when these cells are derived from embryonic tissue, but is lost as cells approach replicative senescence. This auxiliary function is provided by nuclear oncogenes but not cytoplasmic transforming genes.     

Two nuclear oncogenic proteins, P135gag-myb-ets and p61/63myc, cooperate to induce transformation of chicken neuroretina cells

Several studies have shown that full transformation of primary rodent fibroblasts can be achieved in vitro through the cooperation of two oncogenes (usually one nuclear and one cytoplasmic) classified on the basis of different complementation groups. We have shown previously that cooperation between v-mil (cytoplasmic, serine-threonine kinase product), and v-myc (nuclear, DNA-binding product) is required to transform 7-day-old chicken neuroretina cells, which in usual culture medium do not rapidly proliferate. v-mil induces sustained growth of chicken neuroretina cells without transformation; v-myc fails to stimulate the proliferation of chicken neuroretina cells but is required to achieve transformation of the proliferating cells. Here, we present results indicating that the P135gag-myb-ets nuclear protein of avian erythroblastosis virus E26 is able to induce proliferation but not transformation of chicken neuroretina cells. v-myc is required in addition to P135gag-myb-ets to achieve chicken neuroretina cell transformation. In contrast, we found that the P135gag-myb-ets and P100gag-mil proteins are not able to cooperate in this system.     

Insulin-like growth factor I receptor signaling in transformation by src oncogenes.

R- cells, a line of mouse embryo fibroblasts with a targeted disruption of the insulin-like growth factor I (IGF-I) receptor genes, are refractory to transformation by several viral and cellular oncogenes. Using colony formation in soft agar as a measure of full transformation, we report here that R- cells can be transformed by v-src, although they still cannot be transformed by the activated c-src527 (mutation at tyrosine 527 to phenylalanine), which readily transforms mouse embryo cells with a wild-type number of IGF-I receptors (W cells). Although v-src is a more potent inducer of tyrosine phosphorylation than c-src527, the extent of phosphorylation of either insulin receptor substrate 1 or Shc, two of the major substrates of the IGF-I receptor, does not seem sufficiently different to explain the qualitative difference in soft agar growth. v-src, however, is considerably more efficient than c-src527 in its ability to tyrosyl phosphorylate, in R- cells, the focal adhesion kinase, Stat1, and p130cas. These results indicate that v-src, but not c-src527, can bypass the requirement for a functional IGF-I receptor in the full transformation of mouse embryo fibroblasts and suggest that qualitative and quantitative differences between the two oncogenes can be used to identify some of the signals relevant to the mechanism(s) of transformation.     

Effects of cellular transformation on expression of plasminogen activator inhibitors 1 and 2. Evidence for independent regulation

Expression of plasminogen activators (PA) has been reported to be associated with invasive tumor growth and increased metastatic ability. In order to delineate changes in PA and PA inhibitor (PAI) expression that accompany cellular transformation, we studied oncogene-containing variants of the Rat-1 cell line. We report here that transfection of the oncogenes v-src, erbB, c-myc, v-myc, N-myc, and EJras into these cells does not result in detectable PA activity in conditioned media or cell extracts. In addition, Northern blot analysis fails to demonstrate urokinase mRNA in Rat-1 cells or transfectants. Moreover, cells transformed by EJras and v-src but not other oncogenes secrete an active placental-type PAI, PAI-2. Using inducible EJras constructs, we find that increased PAI-2 gene expression is detectable within 6-12 h after treatment with the inducing agent. Peak expression of PAI-2 mRNA is increased 10-15-fold over base line, and high levels are maintained for at least 72 h. In contrast to the results with PAI-2, secretion of endothelial-type PAI-1 into conditioned media is sharply down-regulated by several oncogenes. Thus, we have found that PAI-1 and PAI-2 are independently regulated in transformed variants of Rat-1 cells. The specific induction of PAI-2 in cells transformed by oncogenic ras and src suggests that this protease inhibitor may have a previously unsuspected role in malignancy.     

Synergism of v-myc and v-Ha-ras in the in vitro neoplastic progression of murine lymphoid cells.

Murine bone marrow was either singly or doubly infected with retroviral vectors expressing v-myc (OK10) or v-Ha-ras. The infected bone marrow was cultured in a system that supports the long-term growth of B-lineage lymphoid cells. While the v-myc vector by itself had no apparent effect on lymphoid culture establishment and growth, infection with the v-Ha-ras vector or coinfection with both v-myc and v-Ha-ras vectors led to the appearance of growth-stimulated cell populations. Clonal pre-B-cell lines stably expressing v-Ha-ras alone or both v-myc and v-Ha-ras grew out of these cultures. In comparison with cell lines expressing v-Ha-ras alone, cell lines expressing both v-myc and v-Ha-ras grew to higher densities, had reduced dependence on a feeder layer for growth, and had a marked increase in ability to grow in soft-agar medium. The cell lines expressing both oncogenes were highly tumorigenic in syngeneic animals. These experiments show that the v-myc oncogene in synergy with v-Ha-ras can play a direct role in the in vitro transformation of murine B lymphoid cells.     

Neoplastic transformation of normal and carcinogen-induced preneoplastic Syrian hamster embryo cells by the v-src oncogene.

The ability of cloned Rous sarcoma virus (RSV) DNA encoding the v-src oncogene to neoplastically transform normal, diploid Syrian hamster embryo (SHE) cells was examined. Transfection of RSV DNA into early passage SHE cells resulted in a low but significant number of tumors when treated cells were injected into nude mice. Tumors formed with a low frequency (two tumors out of ten sites injected) and only after a long latency period (14 weeks). In contrast to the normal SHE cells, several different carcinogen-induced preneoplastic immortal SHE cell lines were highly susceptible to transformation by the v-src oncogene to the neoplastic phenotype. Tumors formed with high efficiency and a short latency period (less than 3 weeks). Further studies were performed to determine the basis for the inefficient transformation of the normal SHE cells. NeoR clones isolated after cotransfection of SHE cells with pSV2-neo and RSV DNAs were neither morphologically altered nor immortal and did not contain detectable levels of the v-src gene product. These results suggest that neoplastic transformation by v-src DNA in the normal cells is initially suppressed. However, cells from a v-src-induced tumor expressed v-src RNA, and antibody to v-src protein precipitated from the tumor cells a 60,000-molecular-weight protein which displayed protein kinase activity. Karyotypic analyses confirmed that the tumor was derived from Syrian hamster cells and suggested that it was clonal in nature. These results indicate that the v-src oncogene was primarily responsible for neoplastic transformation of SHE cells. In contrast to the results with the v-src oncogene, our previous studies showed that v-Ha-ras oncogene alone is unable to induce neoplastic transformation of SHE cells. Furthermore, the v-myc oncogene was able to compliment v-Ha-ras to neoplastically transform SHE cells, while cotransfection with v-src plus v-myc did not increase the incidence of tumors.     

Generation of macrophage cell line from fresh bone marrow cells with a myc/raf recombinant retrovirus

We have studied the effects of infection of fresh murine bone marrow (BM) cells by recombinant retroviruses carrying v-raf and v-myc oncogenes, either alone or in combination. Viruses containing v-raf or v-myc alone failed to induce BM proliferation in 24 out of 27 experiments performed so far, only the J2 virus containing both v-raf and v-myc oncogenes induced BM proliferation. Exogenous growth factors (GF) were not required to sustain the mitogenic effect of J2 virus. Infection with retroviruses carrying only v-raf or v-myc did not induce BM cell growth, indicating that co-expression of the two oncogenes was needed to provide the mitogenic signal(s) for BM proliferation. The kinetics of growth of the J2 virus-infected cells (J2 cells) were characteristically biphasic. The initial burst of proliferation was always followed by a quiescent phase culminating in cell death, which could not be reversed by addition of exogenous GF. In contrast, active proliferation of the quiescent monolayers could be restored by addition of dextran-based beads to the cultures, showing that the growth arrest of J2 cells was a reversible process. J2 cells actively growing in the presence of CT-beads could be expanded and cloned and subsequently grew continuously independent of the CT-beads. Eighteen clones obtained from different infections were all macrophages (M phi) by morphological criteria and all of them expressed the same membrane phenotype compatible with M phi, demonstrating that J2 virus infection leads to immortalization of the same BM-derived monocytic subpopulation. When injected in vivo, J2 cells produced histiocytic tumors in nude mice, but did not grow in immunocompetent syngeneic mice. The cells induced to proliferate in vitro in response to J2 virus infection appeared to be limited to the BM compartment, since spleen cells, thymocytes, peritoneal M phi and liver large granular lymphocytes did not grow in vitro in response to J2 virus. The immortalization of BM cells by J2 virus infection represents a novel reproducible experimental system to deliberately generate M phi lines, which proliferate in response to viral oncogenes and do not require exogenous GF to initiate or to sustain their continuous proliferation.     

Genetic analysis of p60v-src domains involved in the induction of different cell transformation parameters.

The expression of p60v-src in chicken cells infected with Rous sarcoma virus causes stimulation of cell proliferation, morphological alteration, and anchorage independence. PA101 and PA104 are temperature-sensitive variants encoding mutant p60v-src proteins that are partially defective in the induction of these transformation parameters. To define the structural basis for the transformation defectiveness of the p60v-src mutants, the v-src genes of PA101 and PA104 were molecularly cloned and analyzed. Amino- and carboxy-terminal coding regions of the cloned mutant genes were exchanged with the corresponding regions of cloned wild-type v-src and chicken c-src genes, reconstructed into viral DNA, and expressed in infected cells maintained at various temperatures. This analysis revealed that lesions within the tyrosine kinase domains of the two mutant proteins confer temperature sensitivity on all three transformation functions of p60v-src. An amino-terminal region of the PA101 mutant protein, which coincides with the proposed modulatory domain and appears to interact with the kinase domain, affects morphological alteration in a temperature-independent manner. Our results suggest that the function of the kinase domain is essential to all three parameters examined, whereas the amino-terminal domain is important in determining cell morphology.     

Coinfection with viruses carrying the v-Ha-ras and v-myc oncogenes leads to growth factor independence by an indirect mechanism.

The concomitant expression of certain oncogenes can transform normal diploid rodent cells into transplantable tumorigenic cells. The mechanism by which these oncogenes collaborate is unclear. Recent findings (M. Oshimura, T. M. Gilmer, and J. C. Barrett, Nature [London] 316:636-639, 1985) raise the possibility that karyotypic changes, including monosomy for chromosome 15, are required to induce tumorigenicity in Syrian hamster embryo cells transfected in vitro with v-Ha-ras and v-myc DNAs. We studied the effect of the oncogenes v-Ha-ras and v-myc, introduced by viral infection, on murine hematopoietic cells. The induction of growth factor independence by the two oncogenes was used as an in vitro correlate of tumorigenicity. After a period of reduced growth rate reminiscent of the growth rate of cells in crisis, the doubly infected cells became growth factor independent. These cells showed a great variability in their karyotypes.