Interactions of human cytomegalovirus with human fibroblasts

Vonka, Vladimir (Baylor University College of Medicine, Houston, Tex.), and Matilda Benyesh-Melnick. Interactions of human cytomegalovirus with human fibroblasts. J. Bacteriol. 91:213-220. 1966.-Virus attachment of human cytomegalovirus to human embryo lung fibroblasts was found to be temperature-independent, from 4 to 37 C. Prolonged incubation at 4 C, however, resulted in inactivation of a high proportion of attached virus. Virus penetration seemed to be temperature-dependent, occurring at 37 C but not at 4 C. Detailed studies of the growth curve of the virus were made. Cell-associated virus preceded the appearance of virus in the fluid phase by 2 to 5 days. Complement-fixing antigen could be detected, but only when the cytopathic effect was advanced, and it was demonstrable only in the cell-associated fraction. Under methyl cellulose, decreasing the bicarbonate concentration in the overlay from 0.225 to 0.15% resulted in marked increase in plating efficiency with all strains tested. However, varying the concentration of bicarbonate from 0.3 to 0.15% in fluid medium did not influence the growth of virus.     

Cell cycle arrest by human cytomegalovirus 86-kDa IE2 protein resembles premature senescence

Primary human embryo lung fibroblasts and adult diploid fibroblasts infected by the human cytomegalovirus (HCMV) display beta-galactosidase (beta-Gal) activity at neutral pH (senescence-associated beta-Gal [SA-beta-Gal] activity) and overexpression of the plasminogen activator inhibitor type 1 (PAI-1) gene, two widely recognized markers of the process designated premature cell senescence. This activity is higher when cells are serum starved for 48 h before infection, a process that speeds and facilitates HCMV infection but that is insufficient by itself to induce senescence. Fibroblasts infected by HCMV do not incorporate bromodeoxyuridine, a prerequisite for the formal definition of senescence. At the molecular level, cells infected by HCMV, beside the accumulation of large amounts of the cell cycle regulators p53 and pRb, the latter in its hyperphosphorylated form, display a strong induction of the cyclin-dependent kinase inhibitor (cdki) p16(INK4a), a direct effector of the senescence phenotype in fibroblasts, and a decrease of the cdki p21(CIP1/WAF). Finally, a replicative senescence state in the early phases of infection significantly increased the number of cells permissive to virus infection and enhanced HCMV replication. HCMV infection assays carried out in the presence of phosphonoformic acid, which inhibits the virus DNA polymerase and the expression of downstream genes, indicated that immediate-early and/or early (alpha) genes are sufficient for the induction of SA-beta-Gal activity. When baculovirus vectors expressing HCMV IE1-72 or IE2-86 proteins were inoculated into fibroblasts, the increase of p16(INK4a) (observed predominantly with IE2-86) was similar to that observed with the whole virus, as was the induction of SA-beta-Gal activity, suggesting that the viral IE2 gene leads infected cells into senescence. Altogether our results demonstrate for the first time that HCMV, after arresting the cell cycle and inhibiting apoptosis, triggers the cellular senescence program, probably through the p16(INK4a) and p53 pathways.     

Unrestricted replication of human cytomegalovirus in hydrocortisone-treated macrophages.

Monocytes differentiated in the presence of phytohemagglutinin P-stimulated T cells could be infected with human cytomegalovirus AD169 and produced low levels of infectious virus. Additional treatment with therapeutic levels of hydrocortisone resulted in a 10- to 100-fold increase in infectious virus production. Hydrocortisone-treated cells demonstrated immediate-early protein kinetics similar to that observed with human fibroblasts, whereas a delay of up to 24 h was observed with untreated cells. Late protein production was barely detectable by immunostaining without hydrocortisone treatment. In treated cells, however, late protein was detected and the levels correlated with the number of cells producing infectious virus. This system provides a model for human cytomegalovirus infection of macrophages in humans.     

Human cytomegalovirus proteins PP65 and IEP72 are targeted to distinct compartments in nuclei and nuclear matrices of infected human embryo fibroblasts

The cellular distribution of the human cytomegalovirus (HCMV)-specific UL83 phosphoprotein (pp65) and UL123 immediate-early protein (IEp72) in lytically infected human embryo fibroblasts was studied by means of indirect immunofluorescence and confocal microscopy. Both proteins were found to have a nuclear localization, but they were concentrated in different compartments within the nuclei. The pp65 was located predominantly in the nucleoli; this was already evident with the parental viral protein, which was targeted to the above nuclear compartment very soon after infection. The nucleolar localization of pp65 was also observed at later stages of the HCMV infectious cycle. After chromatin extraction (in the so-called in situ nuclear matrices), a significant portion of the pp65 remained associated with nucleoli within the first hour after infection, then gradually redistributed in a perinucleolar area, as well as throughout the nucleus, with a granular pattern. A quite different distribution was observed for IEp72 at very early stages after infection of human embryo fibroblasts with HCMV; indeed, this viral protein was found in bright foci, clearly observable in both non-extracted nuclei and in nuclear matrices. At later stages of infection, IEp72 became almost homogeneously distributed within the whole nucleus, while the foci increased in size and were more evenly spread; in several infected cells some of them lay within nucleoli. This peculiar nuclear distribution of IEp72 was preserved in nuclear matrices as well. The entire set of data is discussed in terms of the necessity of integration for HCMV-specific products into the pre-existing nuclear architecture, with the possibility of subsequent adaptation of nuclear compartments to fit the needs of the HCMV replicative cycle.     

Search for human tumour viruses by transfection: uptake of melanoma and Epstein-Barr virus DNA by human cells.

In a model system, consistent transfection of chick embryo fibroblasts (CEF) by DNA from the XC cell line occurred, with recovery of infectious Rous sarcoma virus. The techniques were then applied in attempts to recover possible human tumour viruses. Even with various modifications of the XC technique, DNA from three human malignant melanoma cell lines failed to infect adult or foetal human fibroblasts, although melanoma DNA was taken up into nuclei of target cells. XC DNA did not transfect human foetal fibroblasts and melanoma DNA was ineffective in CEF. DNA from the Raji (Epstein-Barr virus non-producer) and QIMR-WIL (producer) lymphoblastoid cell lines did not transfect human cord blood lymphocytes or amnion cells. These broadly applicable techniques therefore failed to recover EB virus, the putative melanoma retrovirus, or other potential tumour virus.     

Characteristics of MNNG induced repair synthesis and DNA synthesis induced by human cytomegalovirus

DNA synthesis induced by N-methyl-N'-nitro-N-nitrosoguanidine in mouse embryo and human embryo cells was compared with DNA synthesis induced in these cells by human cytomegalovirus. In virus infected human embryo cells grown in the medium depleted of arginine DNA synthesis showed resistance to hydroxyurea and arabinofuranosylcytosine, similarly as repair synthesis induced by MNNG. DNA synthesis induced by the virus in mouse embryo cells was partially sensitive to both inhibitors.     

Replication of wild-type and mutant human cytomegalovirus in life-extended human diploid fibroblasts

A cDNA encoding the catalytic subunit of human telomerase was used to generate life-extended derivatives of primary human diploid fibroblasts. The life-extended cells supported efficient human cytomegalovirus (HCMV) replication. A subclone of the life-extended cells was generated containing the HCMV UL82 gene and used to isolate and propagate a virus that exhibited a profound growth defect after infection at a low input multiplicity.     

Low tumorigenicity of canine cells transformed by the human cytomegalovirus

Dog embryo kidney cells transformed by the human cytomegalovirus (HCMV) were obtained after non-permissive infection or transfection with viral DNA digested by restriction endonuclease EcoR I. The transformed cells, growing rapidly and showing an unlimited division potential, could use medium with only 2% serum for growth, contained nuclear virus antigens, and formed small colonies (less than 0.2 mm) in agarose. From 40 mice inoculated with transformed canine cells, only one eventually developed a tumor. Results indicate that dog cells are immortalized but not tumorigenically transformed by the human cytomegalovirus.     

Host cellular annexin II is associated with cytomegalovirus particles isolated from cultured human fibroblasts.

A significant amount of host cellular annexin II was found to be associated with human cytomegalovirus isolated from cultured human fibroblasts (approximately 1,160 molecules per virion). This composition was established by four different analytical approaches that included (i) Western blot (immunoblot) analysis of gradient-purified virions with a monoclonal antibody specific for annexin II, (ii) peptide mapping and sequence analysis of virus-associated proteins and proteins dissociated from virus following EDTA treatment, (iii) electron microscopic immunocytochemistry of gradient-purified virions, and (iv) labeling of virus-associated proteins by lactoperoxidase-catalyzed radioiodination. These results indicated that annexin II was primarily localized to the viral surface, where it bound in a divalent cation-dependent manner. In functional experiments, a rabbit antiserum raised against annexin II inhibited cytomegalovirus plaque formation in human foreskin fibroblast monolayers in a concentration-dependent manner. Cumulatively, these studies demonstrate an association of host annexin II with cytomegalovirus particles and provide evidence for the involvement of this cellular protein in virus infectivity.     

Appearance of IgG (Fc) receptor(s) on cultured human fibroblasts infected with human cytomegalovirus.

Cultured human diploid fibroblasts (WI-38) after infection with human cytomegalovirus (CMV) but not when uninfected, could hemadsorb sheep red blood cells (SRBC) coated with rabbit anti-SRBC IgG. The adsorption of IgG-coated SRBC to virus-infected cells was completely abolished if the tests were carried out in the simultaneous presence of rabbit antiserum elicited against CMV. Normal sera of rabbit or human origin as well as purified human IgG but not Fab fragment of human IgG could also abolish the binding of sensitized SRBC to CMV-infected fibroblasts. Active metabolism on the part of CMV-infected fibroblasts proved to be an important requisite for demonstrating binding of sensitized SRBC to their surfaces. By using an indicator Staphylococcus aureus to which rabbit antiserum against normal human IgG, IgM, or IgA was bound via Fc fragments, evidence has been obtained which suggests the existence of receptor(s) on CMV-infected WI-38 cells that react specifically with Fc region of human IgG.     

Inhibition of cytomegalovirus-stimulated human cell ornithine decarboxylase by alpha-difluoromethylornithine.

1. The relationship between synthesis of putrescine, human cytomegalovirus DNA synthesis, cell DNA synthesis, and human cytomegalovirus replication has been studied. 2. Stimulation of ornithine decarboxylase activity by shifting low serum-arrested whole human embryo cells to high serum medium is inhibited more than 99% by 2.5 mM DL-alpha-difluoromethylornithine. The addition of DL-alpha-difluoromethylornithine to human cells arrested in low serum and subsequently stimulated by the addition of fresh high serum-containing medium, causes a greater percent inhibition of ornithine decarboxylase activity than when the drug is added to growing human cells. 3. Increased ornithine decarboxylase activity produced by infection of low serum-arrested human cells was inhibited by 5.0 mM of DL-alpha-difluoromethylornithine. However, at a concentration of 5.0 mM, neither DL-alpha-methylornithine nor DL-alpha-difluoromethylornithine affected human cytomegalovirus growth or was toxic to these cells. These data suggest that the increased putrescine synthesis produced by infection is not required for virus replication. 4. The addition of 5.0 mM DL-alpha-difluoromethylornithine had no effect on human cytomegalovirus DNA synthesis or human cytomegalovirus-induced stimulation of cell DNA synthesis. However, 5.0 mM DL-alpha-difluoromethylornithine significantly reduced the stimulation of cell DNA synthesis caused by treatment with mock infecting fluid.     

Exogenous thymidine is preferentially incorporated into human cytomegalovirus DNA in infected human fibroblasts.

The effect of human cytomegalovirus infection on cellular DNA synthesis in human fibroblasts was measured by fluorometry and by incorporation of radiolabeled thymidine. The results show that although HCMV infection stimulates cellular DNA synthesis in both quiescent and serum-stimulated cells, radiolabeled thymidine is almost exclusively incorporated into viral DNA.     

A human monoclonal antibody to cytomegalovirus (CMV)

We report the development of a human monoclonal antibody to cytomegalovirus (CMV) produced from a human X human hybridoma. This hybrid was developed by fusion of an EBV-transformed cell line making antibody to CMV and a human lymphoblastoid cell line WI-L2. The antibody is directed to a CMV-specific antigen primarily in the nucleus of CMV-infected human fibroblasts. It cross-reacts with at least 10 different strains of CMV and may provide a method for the rapid in vitro diagnosis of CMV infections. The production of CMV-specific human man monoclonal antibodies from human-human hybridomas for future therapeutic use is now technically feasible with this specific method of production.     

Differential effect of arabinofuranosylthymine of the replication of human herpesviruses.

The thymidine analog 1-beta-arabinofuranosylthymine (ara-T) has previously been found to selectively inhibit herpes simplex virus replication. At a relatively nontoxic conentration (50 microgram/ml), ara-T reduced herpes simplex virus yields by 4 to 5 log10. Ara-T was also effective in inhibiting the replication of varicellazoster virus (VZV) in vitro in human embryo fibroblasts, completely preventing VZV-specific cytopathic effects. The inhibition of VZV was reversible upon drug removal at 48 h after addition but was not reversible after 5 days of treatment. ara-T also reduced cell-free virus infectivity and the plaque-forming cell yield of VZV. Compared with the untreated controls, which demonstrated a 1-log10 increase over input plaque-forming cells at 24 h after infection, 50 microgram of ara-T per ml resulted in a 1-log10 decrease. In contrast to herpes simplex virus and VZV, cytomegalovirus replication was relatively resistant to ara-T. Neither cytopathic effects nor the incorporation of [3H]thymidine into acid-insoluble material in cytomegalovirus-infected cells was markedly affected. Analysis of the newly synthesized labeled DNA by CsCl buoyant density determinations indicated that the same relative proportions of cell and virus DNA were synthesized with or without added drug. Interpretation of these results with regard to virus-induced deoxypyrimidine kinase is discussed.     

Transformation of human and murine fibroblasts without viral oncoproteins

Murine embryo fibroblasts are readily transformed by the introduction of specific combinations of oncogenes; however, the expression of those same oncogenes in human cells fails to convert such cells to tumorigenicity. Using normal human and murine embryonic fibroblasts, we show that the transformation of human cells requires several additional alterations beyond those required to transform comparable murine cells. The introduction of the c-Myc and H-RAS oncogenes in the setting of loss of p53 function efficiently transforms murine embryo fibroblasts but fails to transform human cells constitutively expressing hTERT, the catalytic subunit of telomerase. In contrast, transformation of multiple strains of human fibroblasts requires the constitutive expression of c-Myc, H-RAS, and hTERT, together with loss of function of the p53, RB, and PTEN tumor suppressor genes. These manipulations permit the development of transformed human fibroblasts with genetic alterations similar to those found associated with human cancers and define specific differences in the susceptibility of human and murine fibroblasts to experimental transformation.     

Lack of correlation between plasminogen activator production and the continous expression of the transformed phenotype of chicken and mammalian cells as shown with S-adenosyl homocysteine analogues

The effect of S-adenosyl-homocysteine and some of its natural and synthetic analogues was studied on plasminogen activator production in chick embryo fibroblasts transformed by a wild type and by a temperature-conditional mutant of Rous Sarcoma Virus (RSV), in a RSV transformed hamster cell line and in human KB cells. When the analogues were added to the cells infected by RSV before the morphologic transformation took place, those molecules which were previously known as inhibitors of cell transformation, inhibited also plasminogen activator production. However when the same molecules were added to already transformed cells, plasminogen activator was still inhibited very strongly, but the cells conserved their transformed morphology. These results show a lack of coordination between plasminogen activator production and the maintenance of the transformed phenotype, and the possibility to inhibit plasminogen activator production by certain transmethylase inhibitors.     

Nonacylated human transferrin receptors are rapidly internalized and mediate iron uptake

The human transferrin receptor is post-translationally modified by the addition of a fatty acyl moiety. In earlier studies, transient expression in Cos cells of human transferrin receptors in which Cys62 or Cys67 was altered to serine provided evidence that Cys62 is the major acylation site of the receptor (Jing, S., and Trowbridge, I. S. (1987) EMBO J. 6, 327-331). To determine whether acylation of the receptor is required for high efficiency endocytosis and iron uptake, wild type and mutant human transferrin receptors have been stably expressed in chick embryo fibroblasts using a helper-independent retroviral vector. In marked contrast to Cos cells, both Cys62 and Cys67 of the wild type human transferrin receptor were acylated in chick embryo fibroblasts. Moreover, their modification to serine did not abolish palmitate labeling, implying that one or both of these serine residues could serve as alternative lipid attachment sites in these cells. The relative labeling of mutant receptors with palmitate and the susceptibility of their lipid moieties to cleavage by hydroxylamine were consistent with Ser67 but not Ser62 serving as a lipid attachment site. Consequently, to obtain human transferrin receptors lacking covalently bound lipid in the chick embryo fibroblasts, it was necessary to alter Cys62 and Cys67 to alanine. Functional studies indicated that these non-acylated mutant receptors were internalized efficiently and mediated iron uptake from human transferrin at a similar rate to that of wild type receptors. We conclude, therefore, that acylation of the human transferrin receptor is not essential for endocytosis and recycling.     

Biosynthesis and secretion of collagen in normal and SV-40 virus-transformed human embryonal fibroblasts

The biosynthesis and secretion of collagen proteins was studied in cultures of normal human embryo fibroblasts at different passages and growth stages as well as in cultures of human embryo fibroblasts transformed by oncogenic virus SV-40. It was found that normal fibroblasts maintain at a constant level the collagen synthesis throughout 20 passages, which is typical of proliferating and resting cells. Virus-transformed cells produce 3-4 times less collagen proteins on a per cell count. Normal and transformed fibroblasts do not differ in terms of total protein synthesis. Secretion of collagen and non-collagen proteins in transformed cell cultures appeared to be much lower than in normal cell cultures. Study of synthesized proteins by polyacrylamide gel electrophoresis showed that both types of cells secrete collagen proteins predominantly as polymers containing interchain S-S bonds of 3-helix molecules. Study of the protein-synthesizing activity of two polysomal fractions, i.e. membrane bound and free polysomes, isolated from the cells of both types in a cell-free system showed that membrane-bound polysomes from transformed fibroblasts synthesize collagen much less actively in comparison with normal cells. However, in transformed cells free polysomes, in contrast with normal cells, are active participants of a cell-free collagen protein synthesis.     

Virus-specific changes in mouse cells inoculated with a strain of human cytomegalovirus.

Cytopathic changes and virus-specific antigens developed in, then disappeared from, mouse fibroblasts infected by a strain of human cytomegalovirus (CMV), but their disappearance was delayed in cells treated with idoxuridine prior to infection. The replication of vesicular stomatitis virus and herpes simplex virus was restricted in human CMV-infected mouse cells as long as human CMV-specific antigens were present. Virus-specific antigens could be induced by treatment with idoxuridine or arginine deficiency in mouse cells which had previously turned "negative".