High-frequency establishment of human immunoglobulin-producing lymphoblastoid lines from normal and malignant lymphoid tissue and peripheral blood

Human hematopoietic tissue and lymphocytes separated from 10-20 ml samples of peripheral blood have been grown in vitro in a lens-paper and a gelatin foam (Spongostan) grid organ culture. Lymphoblastoid cell lines were established from the lymph nodes, and in one case from the spleen, of 22/23 consecutive, unselected adult individuals without manifest malignancy or infectious mononucleosis. Biopsies from 5/8 patients with malignancy were successful. The blood tines were derived from 5/10 patients with and 4/10 donors without malignancy. The very high frequency of success from normal tissue confirms the assumption made before that the spontaneous establishment of lymphoblastoid cell lines is unrelated to manifest malignancy of the donor. The results indicate that lymphoid cells with a potential for infinite proliferation (“lymphoblastoid transformation”) are present in almost all adult individuals. The Spongostan grid culture is a superior instrument to select and/or adapt these cells in vitro. All lymphoblastoid lines produced immunoglobulins. The majority started with a “polyclonal” pattern of immunoglobulin production but changed towards stable “ monoclonality “ during the course of long-term cultivation. It is suggested that lymphoblastoid lines have a polyclonal origin and that the reason for development of a monoclonal line is a selection of one cell clone either in the organ culture during establishment or in long-term culture.     

Properties of a baboon lymphotropic herpesvirus related to Epstein-Barr virus.

Three lymphoblastoid cell lines were established from splenic lymphocytes of a lymphomatous baboon (Papio hamadryas) by co-cultivation of the lymphocytes with X-irradiated cells of marmoset or baboon lymphoblastoid cell cultures; the baboon splenic lymphocytes failed to grow when cultured alone. A herpesvirus, associated with each cell line, was identified by immunofluorescence, molecular hybridization and electron microscopy. Antigenic comparison with Epstein-Barr virus (EBV) showed that the baboon herpesvirus and EBV shared cross-reacting viral capsid antigens (VCA): 20 of 20 (100%) anti-VCA (EBV)-positive human sera and 55 of 62 (89%) baboon sera reacted with the baboon lymphoblastoid cells and baboon sera stained EBV VCA in P3HR-1 and EB-3 cells. No nuclear antigen, as assayed by anti-complement immunofluorescence tests, was detected in baboon lymphoblastoid cells when human or baboon anti-VCA positive sera were used. Baboon anti-VCA-positive sera also failed to stain EBV nuclear antigens (EBNA) in Raji or P3HR-1 cells. Preliminary molecular hybridization studies showed only approximately 40% homology between viral DNA of baboon cell lines and DNA of EBV derived from P3HR-1 cells.     

Cellular localization of an Epstein-Barr virus (EBV)-associated complement-fixing antigen in producer and non-producer lymphoblastoid cell lines

Anti-complement immunofluorescence (ACIF) was used to study the complementfixing antigens of human lymphoblastoid cell lines. These cell lines carry the Epstein-Barr virus (EBV) genome although only producer cultures synthetize EBV-specific antigens (virus capsid antigen, VCA and early antigen, EA) detectable by direct and indirect immunofluorescence, usually in less than 5% of the cells. The ACIF test revealed an antigen localized in the nucleus of the lymphoblastoid cells. In contrast to EA and VCA, this antigen was present in over 90% of the cells of both producer and non-producer cultures. The antigen was shown to be specific for EBV by comparing the reactions of 52 sera in the ACIF test. Sera giving the nuclear reaction contained antibodies to VCA, EA or antigens detectable by complement fixation tests on cell extracts, but sera without EBV antibodies failed to give the reaction. Weak, equivocal or discordant reactions occurred with six sera with low titres in VCA, EA or complement fixation tests. Cell lines derived by transformation of human and primate lymphocytes by EBV gave the nuclear reaction. Control cells with no known association with EBV were non-reactive. These included foetal lymphocytes transformed by phytohaemagglutinin, cell lines derived from breast cancer, glioma, normal glia, pleuritis maligna and myeloma, and two marmoset lymphoid lines carrying Herpesvirus saimiri (HVS). In preliminary experiments, the ACIF test was used as a tool to trace the EBV genome at the cellular level. Cells from two Burkitt lymphoma biopsies, one tested after biopsy and one after passaging in nude mice, contained an EBV-specific antigen. Three clones of cells derived from hybrids of mouse somatic cells and a human lymphoblastoid cell line also contained such an antigen, but the number of reactive cells varied from clone to clone. A fourth clone was non-reactive.     

Epstein-Barr virus (EBV)-associated antibody patterns in malignant lymphoma and leukemia. II. Chronic lymphocytic leukemia and lymphocytic lymphoma

Sera from Swedish patients with chronic lymphocytic leukemia (CLL) and lymphocytic lymphona (LL) were titrated for antibodies to viral capsid antigens (VCA) of the Epstein-Barr virus (EBV) by the indirect immunofluorescence technique and to EBV-determined cell membrane antigens by blocking of direct cell membrane immunofluorescence (BI). Previously tested sera from healthy Swedish donors served as controls. While 40% of the sera from CLL patients showed relatively high serological reactivities, the mean anti-VCA and BI values did not differ significantly from the control group. The LL group of sera showed significantly higher reactivities than the controls. The most highly reactive sera were found in patients with poorly differentiated LL at frequencies and with mean values approaching those seen in Burkitt's lymphoma, nasopharyngeal carcinoma and Hodgkin's sarcoma.     

Epstein-Barr virus-associated complement-fixing and nuclear antigens in Burkitt lymphoma biopsies

Cells from Burkitt lymphoma (BL) biopsies were examined for Epstein-Barr virus (EBV)-associated antigens by complement fixation (CF) tests and by the anti-complement immunofluorescence (ACIF) test. In CF tests, anticomplementary factors made it difficult to test all the biopsies available. However, biopsies from 19 patients were effectively tested and 12 of these (including two from one patient) contained antigen reacting with a battery of human sera with antibody to EBV but not with sera lacking such antibody. Technical difficulties prevented further characterization of the EBV-related antigens in the biopsies. Application of the ACIF test to BL revealed the presence of EBV-related nuclear antigen in biopsies from 11 of 13 patients. Absorption studies indicated that the nuclear antigens of the biopsies were closely related antigenically to the EBV-determined nuclear antigen (EBNA) previously described in lymphoblastoid cell lines. It is concluded that cells of BL biopsies may contain EBNA in addition to the EBV-related membrane antigen previously described. The results provide further evidence that BL cells from African patients resemble non-producer lymphoblastoid cell lines in containing EBNA and therefore appear to be transformed by EBV.     

Identity of the soluble EBV-associated antigens of human lymphoid cell lines

The reactions of human sera with antigens prepared from five human lymphoid cell lines were studied by immunodiffusion. Two major precipitation lines occurred with QIMR-WIL antigens; one (S line) due to a soluble antigen and the other (F line) due to a sedimentable antigen. A close association in human sera of precipitins producing the S line with antibody to Epstein-Barr virus (EBV) detectable by immunofluorescence (Henle test) and by complement fixation (CF) tests indicated a specific association of the S line with EBV, but the nature of the F line was not clarified. Soluble antigens prepared from five lymphoid cell lines (including Raji) contained at least one common identical heat-resistant component, defined as the antigen giving the S line. The ease of detection of the S antigen in cell lines by immunodiffusion appeared to be related to the titre of the soluble EBV-associated CF antigen, and the evidence suggested that the soluble CF and ID antigens were probably identical.     

Epstein-Barr virus (EBV)-associated antibody patterns in malignant lymphoma and leukemia. I. Hodgkin's disease

Sera from Swedish patients with Hodgkin's disease as well as control sera from donors without known disease were titrated for antibodies to the Epstein-Barr virus (EBV) by the indirect immunofluorescence technique. In the same sera antibodies capable of blocking the direct membrane immunofluorescence reaction obtained between Epstein-Barr virus carrying lymphoblastoid cell lines and a fluorescein-conjugated reference serum from a patient with Burkitt's lymphoma (F-Mutua conjugate) were tested. The serological reactivity of the control sera was very similar to that described in previous reports. In contrast, sera from Hodgkin's disease patients showed an increased reactivity in both tests. When the donors were grouped in relation to clinical and histological data of prognostic importance, an inverse relationship was found between the frequency of lymphoid cells and EBV-associated serological reactivity. Whereas paragranuloma cases did not exceed the reactivity level of the controls, patients with Hodgkin's sarcoma were highly reactive in both tests, reaching levels comparable to those seen in Burkitt's lymphoma and nasopharyngeal carcinoma. The granuloma group was intermediate with regard to serological reactivity.     

Studies on the infectivity and cytopathology of Epstein-Barr virus in human lymphoblastoid cells

The infectivity of EB virus recovered from cultures of P-3J and the HR1K clone of this cell line has been demonstrated in established lymphoblastoid cell lines derived from healthy human donors and from patients with neoplastic disease. Filtered virus concentrates induced cell alterations similar to those produced by other members of the herpes group of viruses. Cell responses observed were: 1) an acute cytopathic effect evident within 16–72 h and characterized by cell swelling, polykaryocyte formation and progressive degeneration, and 2) a less severe CPE with minimal cell swelling and degeneration followed by either the establishment of a carrier culture or more often a culture in which the virus eventually disappeared. Characteristic non-enveloped herpes-virus-type particles were seen in the cell nucleus as early as 16 hours post exposure to virus and, subsequently, enveloped virions were found in the cytoplasm. A non-productive or abortive infection of the lymphoblastoid cells was indicated by the higher incidence of cells showing viral antigen by immunofluorescence, than the presence of virus particles. The infectivity of EBV was neutralized by human sera containing antibodies reactive with the viral envelope.     

High irradiation sensitivity of epstein-barr-virus (ebv) in lymphoblastoid-cells from patients with ataxia-telangiectasia and ebv genome-positive burkitts-lymphoma

Epstein-Barr virus (EBV)-immortalized lymphoblastoid cell lines derived from the peripheral blood of patients with ataxia telangiectasia (AT) and EBV genome-positive Burkitt's lymphoma (BL) were tested for expression of EBV-related lytic antigens by means of irradiation. We used 1 Gy in each experiment, according to the results of the P3HR-1 (derived from African BL) cell line. Significantly higher expression of early antigens (EA) and viral capsid antigen (VCA) was demonstrated in lymphoblastoid cell lines derived from both patients with AT and those with EBV genome-positive BL, as compared to those derived from healthy individuals. These results suggested that defective regulating mechanisms on B lymphocytes, responsible for EBV infection, may underlie for the pathogenesis of development of lymphoproliferative diseases both in patients with AT and EBV genome-positive BL.     

Sensitivity of Epstein-Barr virus (EBV) producer and non-producer human lymphoblastoid cell lines to superinfection with EB-virus

Twenty-nine lymphoblastoid lines and one IgE-producing myeloma line of human origin were exposed to Epstein-Barr virus (EBV) concentrates in vitro. The adsorption of the virus to the outer cell membrane was assessed by counting the number of direct membrane fluorescence-positive cells immediately after infection and by the direct radioimmune membrane labelling method. Reference reagents were derived for both tests from the same serum (“Agnes”), containing antibodies against EB-viral envelope and capsid antigens. The intracellular course of infection was followed by counting the number of cells that responded with the development of early antigen (EA) 48 h after infection. The 11 lymphoblastoid lines that produced no EBV-determined membrane and early antigens adsorbed the virus, although there were quantitative differences between them. EA-positive cells appeared in significant numbers in only seven of them, however. Four lines remained EA negative in spite of a relatively good adsorbing capacity. The IgE-producing myeloma line showed neither virus adsorption nor EA development. Eighteen lymphoblastoid lines were “producers”, or “abortive producers”, i.e.a small proportion of the cells continuously generated two or three of the immunofluorescence-detectable viral antigens, MA, EA and VCA. Nine lines failed to adsorb significant virus quantities and showed no certain increase of EA-positive cells. The resistance of these lines to superinfection is probably determined at the level of viral receptors. Five lines showed a relatively good virus adsorption, but this was not followed by any significant increase in the number of EA-positive cells. Four lines showed good adsorption and also responded with a significant increase in the number of EA-positive cells. The same responses can thus be found to EBV-superinfection in producer and non-producer lines, but the producer lines show a strong preponderance of superinfection-resistant lines with an adsorption block at the receptor level.     

Activation of Epstein-Barr virus in hybrid cells.

Human-primate hybrid cell lines were established by fusion of African green monkey kidney cells (VERO) with lymphoblastoid cells from patients with infectious mononucleosis (IM)(IMK101) and from Burkitt's lymphoma culture (HR1K). Both Epstein-Barr virus (EBV)-specific antigens and EBV particle-containing cells increased in the hybrid lines (IMK1-1/VERO,HR1K/VERO). Treatment of the hybrids with 5-bromodeoxyuridine induced more antigen-positive and more virus-containing cells. EBV could be activated from IM lymphoblastoid cells by fusion of the lymphoblastoid cells with the VERO cells. The increase of viral antigens and virus particles may have been due to the cellular interaction between VERO cells and the lymphoblastoid cells or to a possible derepressor supplied by the VERO component of the hybrid. Virus derived from the HR1K cell line was replicated in the human-primate hybrid, but further investigation may be necessary to determine if it was identical to the EBV derived from the human cell line.     

B cells from leukemic patients are relatively resistant to in-vitro EBV transformation

Samples of peripheral blood mononuclear cells (PBMC) from normal donors or from leukemic patients were used to obtain Epstein-Barr virus (EBV)-transformed lymphoblastoid cell lines (LCL). Whereas the rate of transformation was around 85% with normal donors, it was only around 20% with leukemic patients. However no explanation for this resistance of B cells from leukemic patients to in-vitro EBV transformation could be found. Indeed no correlation exists between this EBV resistance and the age, sex, diagnosis or chemotherapeutic regimen. Furthermore no correlation is apparent between the percentage of B2+ cells, which should have EBV receptors, and the successful EBV transformation.     

Sensitivity of different target cells to the killing action of peripheral lymphocytes stimulated by autologous lymphoblastoid cell lines

Peripheral lymphocytes obtained from two individuals with a previous history of infectious mononucleosis were exposed to mitomycin-treated cells of the autologous lymphoblastoid cell line (LCL) established during the acute phase of the disease. This resulted in a stimulation of DNA synthesis, comparable to or even exceeding a one-way MLC with allogeneic lymphocytes. The cytotoxic effect of the stimulated lymphocytes was tested by colony inhibition or ⁵¹Cr release, against a large LCL panel, including the autologous line and allogeneic lines from patients with Burkitt's lymphoma (BL), nasopharyngeal carcinoma (NPC), infectious mononucleosis (IM), leukemia, myeloma, or normal donors. While the majority of the lines were highly sensitive to the killing action, three were relatively resistant. The same pattern of sensitivity was obtained with effector cells stimulated by autologous LCL derived from IM or BL patients. The majority of the target LCLs had B-cell characteristics and carried the EBV genome, but three cell lines that were devoid of the EBV genome, were also sensitive. These lines included two lymphoid cell lines, one of them a T-cell line, and a myeloma line. Fresh peripheral lymphocytes from normal donors or acute IM patients, PHA-induced blasts and blast cells from a case of acute myeloid leukemia were resistant.     

Lymphoblastoid cells transfected with c-myc: Downregulation of EBV-lytic antigens and impaired response of autologousCD4+ T cells in vitro

Normal EBV-positive lymphoblastoid B-cell lines (LCL) were transfected with vectors containing the c-myc oncogene (pHEBO-Eμ- myc) or control vectors (pHEBO-Eμ) and analyzed for the expression of EBV-lytic and latent antigens. While EBV-latent antigens were normal in the c-myc transfectants there was an almost complete downregulation of EBV-lytic antigens, including BZLFI, EA(D), gp340 and VCA. These observations were consistently repeated on 6 different LCLs transfected with c-myc. Unlike control LCLs, the c-myc transfectants did not release infectious EBV. PCR analysis demonstrated that BZLFI mRNA was virtually absent in c-myc transfectants, possibly suggesting that the deregulated c-myc imposed a block in the EBV-lytic cycle at this particular level. c-myc transfectants failed to sustain the proliferative response of autologous CD4⁺ T-cell clones with specificity for EBV-lytic antigens. However, they regained this capacity after incubation with ultraviolet-inactivated EBV or gp340 antigen in vitro, also indicating that their antigen-presenting capacities were not impaired. c-myc transfectants failed to elicit a secondary proliferative response by autologous CD4⁺ T cells purified from the peripheral blood of EBV-seropositive donors. Exposure of c-myc transfectants to UV-inactivated EBV again resulted in a proliferative CD4⁺ T-cell response comparable to that elicited by the control LCLs. Collectively, our data provide evidence for the remarkable ability of an oncogene to influence the life cycle of a virus and to modify the antigenicity of the infected cells. © 1996 Wiley-Liss, Inc.     

Transformation of lymphocytes by Herpesvirus papio.

Cotton-topped (CT) or white-lipped (WL) marmoset lymphocytes were transformed in vitro with herpesvirus papio (HVP) into permanently growing lymphoblastoid cell lines (LCL). Five of 9 HVP-transformed CT cell lines contained cells with antigens reacting with antibodies to Epstein-Barr virus (EBV) capsid antigen (VCA) and/or to EBV-induced early antigens (EA). None of 12 WL LCL revealed such antigen-producing cells. Cells from both groups of cultures failed to react with antibodies to the EBV-specified nuclear antigen (EBNA). Exposure of baboon circulating lymphocytes to X-irradiated HVP or EBV-carring cells, or to suspensions of EBV resulted in establishment of LCL which all contained VCA and/or EA-positive, but no EBNA-positive cells. Nuclear antigens were undetectable also with anti-VCA-positive sera from baboons, chimpanzees, or other non-human primates. DNA-complementary RNA (cRNA) filter hybridization with EBV cRNA showed that with one exception transformed CT or WL marmoset cells contained at least 1-2 virus genome equivalents per cell, while at least 12-25 virus genome equivalents per cell were detected in transformed baboon cells. These data need confirmation by DNA-DNA reassociation kinetics.     

Heterogeneity among Epstein-Barr virus-seropositive donors in the generation of immunoblastic B-cell lymphomas in SCID mice receiving human peripheral blood leukocyte grafts.

Epstein-Barr virus (EBV) is associated with B-cell malignancy in immunosuppressed humans and SCID mice receiving human peripheral blood leukocyte grafts (hu-PBL-SCID). We have further characterized the process of lymphoma development in hu-PBL-SCID mice. We report that EBV-seropositive donors differ markedly in the capacity of their PBL to give rise to immunoblastic lymphomas in SCID mice; some donors (high incidence) generated tumors rapidly in all hu-PBL-SCID mice, other donors (intermediate-low incidence) gave rise to sporadic tumors after a longer latent period (greater than 10 weeks), and some donors failed to produce tumors. B-cell lymphomas arising from high incidence donors were multiclonal in origin, and EBV replication was detected in all tumors. Tumors derived from intermediate-low incidence donors were monoclonal or oligoclonal and often had no evidence of viral replication. All tumors, regardless of the donor, resembled EBV-transformed lymphoblastoid cell lines in surface phenotype but differed from lymphoblastoid cell lines by having less Epstein-Barr nuclear antigen 2 and CD23 expression. The variable patterns of lymphomagenesis seen among different EBV-sero-positive donors may be explained by lower levels of specific immunity to EBV in high incidence donors, permitting activation of EBV replication and potential transformation of secondary B-cell targets. In addition, there may be differences in the transforming potential of EBV infecting different donors. The use of the hu-PBL-SCID model may help predict patients at high risk for posttransplant or acquired immunodeficiency syndrome-associated lymphomas.     

Epstein-Barr virus: transformation of non-human primate lymphocytes in vitro

Continuous lymphoblastoid cell cultures were established from marmoset (Saguinus sp.), squirrel (Saimiri sciureus), owl (Aotus trivirgatus) and cebus (Cebus apella) monkeys after culturing their peripheral lymphocytes with lethally X-irradiated cells carrying Epstein-Barr virus (EBV). Transformation also was achieved by exposing simian lymphocytes to infectious, cell-free EBV derived from the simian lymphoblastoid cell cultures. Simian lymphocytes were not transformed after exposure to cell-free EBV derived from HR-1 cells. The simian cell cultures were similar to cell cultures derived from Burkitt's lymphoma or infectious mononucleosis patients. EBV-induced early, viral capsid and membrane antigens, intranuclear inclusion bodies and herpesvirus virions were demonstrable in most cultures. Seven cultures were insusceptible to superinfection with EBV and treatment of the cultures with halogenated pyrimidines was relatively ineffective for inducing synthesis of early or viral capsid antigens. All cell cultures had B-cell characteristics: they produced immunoglobulins but did not form spontaneous rosettes with sheep erythrocytes. Four of six marmoset monkeys, inoculated with EBV-transformed marmoset lymphocytes, developed antibodies to EB viral capsid antigens and one marmoset inoculated with autochthonous transformed cells also developed heterophile antibodies. Seven marmosets, inoculated with cell-free EBV derived from HR-1 cell cultures, developed no detectable levels of antibodies to EBV-specified antigens or heterophile antibodies. No overt clinical abnormalities were detected in any of the marmosets inoculated with HR-1 or Kaplan EBV but one of five marmosets inoculated with B95hyphen;8 EBV developed a lymphoma.     

Classification and biological nature of established human hematopoietic cell lines.

Over 200 established human hematopoietic cell lines of normal and malignant origin have been investigated by morphological and functional parameters. Employing morphology as the overriding parameter four types of lines were identified. (1) Lymphoblastoid cell lines, derived from normal and neoplastic hematopoietic tissue, were characterized by the wide morphologic flexibility of individual lymphoblastoid cells, constant association with Epstein-Barr virus (EBV), polyclonal derivation, differentiation for immunoglobulin production (secretion) and their diploids. (2) Lymphoma cell lines. This type of line was established at a high frequency from Burkitt's lymphoma and rarely from other types of lymphoma, but never from patients without malignancy or with non-lymphoma malignancies. Important characteristics were morphologic stereotypia within each line, monoclonal derivation, common but not obligatory association with EBV, variability in the expression of Ig synthesis (no production, or membrane bound Ig, or secretion) and aneuploidy. (3) Myeloma cell lines could only rarely be obtained from patients with myeloma. The basis for classification of these lines is their production of Ig identical to the myeloma protein in vitro. Other important distinguishing features were: plasma cell morphology, absence of EBV and aneuploidy. (4) The leukemia cell line (MOLT 4) was the only line with T-cell characteristics and was easily distinguished from the other types. Important characteristics were a typical surface ultrastructure, absence of EBV and absence of immunoglobulin production, Individual lymphoblastoid lines were in principle identical whereas each line of the other three types had its own characteristic profile. The phenotypic characteristics of the lymphoblastoid lines were very stable during prolonged serial cultivation. Only in a few cases were secondary chromosomal, functional or morphologic alterations noted. We conclude that EBV-carrying lymphoblastoid lines can be obtained from non-neoplastic precursor cells from healthy as well as from diseased individuals. Lymphoma, myeloma and leukemia lines are only obtained from the respective neoplastic tissue but generally at a low frequency. With the exception of Burkitt's lymphoma, malignant hematopoietic tissue and leukemia frequently give rise to established cell lines in vitro of the lymphoblastoid type rather than lines derived from the neoplastic cells;     

Re-exposure of human lymphoblastoid cell lines to Epstein-Barr virus

Human lymphoblastoid lines of various origins which harbour Epstein-Barr virus (EBV)-specific nucleic acid were re-exposed to EBV. Following infection, cells of the non-virus-producing lines, Raji and S 95, predominantly synthesized EBV-specific early antigens (EA), whereas only a small percentage of cells revealed viral capsid antigens (VCA). In Raji cells, the number of VCA-producing cells was paralleled by the percentage of virus-specific DNA-synthesizing cells. In S 95 cells, however, viral DNA-synthesizing cells exceeded the number of VCA-producing cells by a factor of more than 10. Induction of EA in Raji cells was dose-dependent and inversely related to cell growth. Irradiation of the virus by ultraviolet light prior to infection led to reduced infectivity. This reduction seemed to follow single-hit kinetics. Raji cells, previously re-exposed to EBV, showed reduced EA induction after re-infection with EBV, as compared to Raji control cells not previously exposed. Of 10 lines which spontaneously synthesize EBV-specific antigens, seven lines proved to be refractory to re-infection, whereas three were as susceptible as the Raji and S 95 controls. From three of the refractory lines infectious virus could be recovered from the culture medium prior to infection. These results permit the following interpretations: (1) the response of human lymphoblastoid cells after re-infection with EBV results from the infecting virus and not from stimulation of endogenous genomes; (2) cells demonstrating EA synthesis ultimately die; (3) re-exposure to EBV increases the resistance to re-infection of the surviving cells; and (4) cell lines producing infectious EBV are refractory to re-infection. It is suggested that the spontaneous synthesis of infectious virus favours the selection of resistant cells.