Jane E. Grundy

Cytomegalovirus-Infected Endothelial Cells Recruit Neutrophils by the Secretion of C-X-C Chemokines and Transmit Virus by Direct Neutrophil-Endothelial Cell Contact and during Neutrophil Transendothelial Migration

Infection of endothelial cells with an endothelial cell-tropic clinical isolate of cytomegalovirus (CMV), C1FE, induced enhanced production of the neutrophil chemoattractant C-X-C chemokines interleukin-8 and GROalpha. Infected endothelial cell supernatants induced neutrophil chemotaxis in a transendothelial migration assay. Neutrophils acquired the CMV structural protein pp65 following either coculture with infected endothelial cells or transmigration through infected endothelium. The lack of CMV p72 expression in the neutrophils indicated that viral replication had not occurred in these cells. Of importance, neutrophils acquired infectious CMV during transmigration across infected endothelium and were subsequently able to transmit infectious virus to fibroblasts. Thus, CMV-infected endothelial cells can recruit neutrophils by the secretion of C-X-C chemokines and can transmit the virus to them by direct cell-to-cell contact and during neutrophil transendothelial migration, suggesting that the neutrophil-endothelial cell interaction plays an important role in virus dissemination in vivo.

Beta 2 microglobulin enhances the infectivity of cytomegalovirus and when bound to the virus enables class I HLA molecules to be used as a virus receptor.

We have previously demonstrated that human cytomegalovirus (CMV) binds the host protein beta 2 microglobulin (beta 2m) from body fluids or from cell culture media. In this report we have examined the effect of the beta 2m on viral infectivity. We have shown that the addition of human purified beta 2m, or a fraction of foetal calf serum corresponding to bovine beta 2m, to culture medium increased the amount of infectious extracellular CMV, compared to that from cells grown in serum-free medium. Metabolic labelling experiments demonstrated that this effect was not due to an increase in the amount of extracellular virus but to an increase in the infectivity of the virus present in extracellular fluids. We concluded that the binding of beta 2m by CMV increased its infectivity. We have shown that CMV and beta 2m compete for binding sites on fibroblasts. As the main binding site on cells for beta 2m is the class I HLA heavy chain we compared the binding of CMV to the Raji and Daudi cell lines which express or lack expression of class I HLA molecules. The binding of radiolabelled beta 2m-coated CMV was significantly higher to Raji cells than to Daudi cells. Furthermore, CMV could compete with beta 2m for binding to Raji cells, although the reverse was not true. These results demonstrate that CMV can use class I HLA molecules as a virus receptor. We propose that when coated with beta 2m, CMV has the capacity to displace beta 2m from the class I HLA heavy chain-beta 2m dimer on the cell surface and bind to cells. The fact that beta 2m enhances infectivity suggests that such binding leads to productive infection of cells.

Augmentation of graft-versus-host reaction by cytomegalovirus infection resulting in interstitial pneumonitis

The severity of the graft-versus-host (GVH) reaction, judged by splenomegaly and immunosuppression, was augmented by murine cytomegalovirus (MCMV) infection. Profound GVH-induced immunosuppression was seen in adult unirradiated MCMV-infected F1, mice even after challenge with extremely low doses of parental spleen cells. Mice receiving MCMV+GVH challenge died from days 16–21, with interstitial pneumonia being the most prominent pathological lesion. Pulmonary disease was unrelated to levels of viral replication in the lung. These results suggest that in human marrow recipients, cytomegalovirus infection may play a primary role both in provoking or accentuating GVH disease, as well as in the development of interstitial pneumonia.

Cytomegalovirus induced up-regulation of LFA-3 (CD58) and ICAM-1 (CD54) is a direct viral effect that is not prevented by ganciclovir or foscarnet treatment

Cytomegalovirus (CMV) is a major pathogen in transplant recipients and AIDS patients, and the virus may also play a role in allograft rejection. Previous work from this laboratory demonstrated increased cell surface expression of the adhesion molecules ICAM-1 (CD54) and LFA-3 (CD58) following CMV infection in vitro. We investigated whether the induction of adhesion molecules by CMV was a direct viral effect or secondary to cytokine induction. Cytokines known to up-regulate ICAM-1, such as TNFalpha or IL-1beta, were not detected in the supernatants of infected fibroblasts, and neutralizing antibodies against these cytokines did not abrogate the induction of either ICAM-1 or LFA-3 by CMV. Infected cell supernatants had increased levels of IL-6, IL-8 and IFNbeta however, the addition of recombinant forms of these cytokines did not affect adhesion molecule expression. Neither virus-free infected cell supernatants nor UV-inactivated virus up-regulated adhesion molecules, demonstrating that the induction of ICAM-1 and LFA-3 by CMV was a direct effect requiring infectious virus. Effective antiviral treatment with ganciclovir or foscarnet accentuated rather than abrogated the up-regulation of adhesion molecules, suggesting that CMV immediate early/early gene expression, which is not blocked by such treatment, was responsible for the adhesion molecule induction. Thus, despite effective antiviral therapy in the transplant recipient, CMV infected cells may continue to provide a focus of proinflammatory activity, which could contribute to immunopathology and/or accentuate graft rejection or graft-versus-host disease in vivo.

Down-regulation of the class I HLA heterodimer and β2-microglobulin on the surface of cells infected with cytomegalovirus

Cytotoxic T cell recognition of virus-infected cells requires the presentation of viral peptides by class I HLA molecules on the cell surface. We report here that cytomegalovirus (CMV) infection of human fibroblasts results in a progressive decrease in the cell surface expression of class I HLA and beta 2-microglobulin (beta 2m) such that in the late stages of infection the majority of infected cells have no detectable surface class I HLA. Coincident with decreased surface expression of class I HLA was an increase in his cytoplasmic expression. Confocal scanning laser microscopic analysis demonstrated that class I HLA and beta 2m accumulate in a perinuclear compartment inside the CMV-infected cell. Our data thus support the concept that CMV infection induces altered transport of class I HLA to the cell surface. We suggest that the virus has evolved this mechanism as a strategy to avoid T cell recognition of infected cells.

CYTOMEGALOVIRUS IN URINE SPECIMENS HAS HOST BETA-2 MICROGLOBULIN BOUND TO THE VIRAL ENVELOPE - A MECHANISM OF EVADING THE HOST IMMUNE-RESPONSE

We have previously reported that human cytomegalovirus (CMV) from urine specimens cannot be captured onto a solid phase by CMV-specific monoclonal antibodies that can capture CMV grown in vitro. We report here that CMV exists in vivo in body fluids such as urine as beta 2 microglobulin (beta 2 m)-coated particles. We have demonstrated the presence of beta 2m on CMV purified directly from urine by Western blotting and have shown that the beta 2m was associated with the viral envelope. Urinary CMV could be specifically bound by an affinity column comprising a monoclonal antibody specific for beta 2m bound to Sepharose. The beta 2m-coated urinary CMV could not be neutralized by hyperimmune globulin, human immune sera or murine monoclonal antibodies that could neutralize CMV grown in cell culture. We conclude that the binding of beta 2m by CMV masks the important antigenic sites necessary for neutralization which are recognized by mans immune response. We propose that CMV has evolved this mechanism of coating itself in a host protein as a mechanism of evading the host immune response and facilitating transmission between individuals.

Use of the polymerase chain reaction to analyse sequence variation within a major neutralizing epitope of glycoprotein B (gp58) in clinical isolates of human cytomegalovirus.

The heterogeneity of low passage human cytomegalovirus (HCMV) strains was determined by HindIII typing of 28 clinical isolates from transplant patients. These data have shown that, in general, each patients strain has a unique restriction profile, usually comprising combinations of HindIII sites present in one or more of the tissue culture-adapted strains AD169, Towne and Davis. To map sequence changes in a more refined manner we performed detailed analyses of 33 low passage clinical isolates, including those aforementioned, analysing a sequence within glycoprotein B containing a major neutralizing epitope. A 149 bp sequence containing the epitope (amino acids 608 to 625) was amplified using the polymerase chain reaction, the products were cloned and their DNA sequence was determined. Comparison of the DNA and deduced amino acid sequences with those of HCMV strain AD169 revealed that there was a high degree of conservation of the epitope between the 33 clinical isolates. However 10 of the isolates possessed silent mutations and three isolates contained mutations producing amino acid changes within the neutralizing epitope. The possible functional significance of these changes is discussed.

Cytomegalovirus strain AD169 binds β2 microglobulin in vitro after release from cells

We previously reported that a host protein, beta 2 microglobulin (beta 2m) inhibited the detection of human cytomegalovirus (CMV) in urine specimens by enzyme immunoassay and postulated that beta 2m bound to the virus particle and masked the viral antigenic determinants. We report here that CMV strain AD169 grown in cell culture bound human beta 2m when this protein was added to cell culture fluids or when the virus was added to urine. Such binding was not seen with herpes simplex virus. CMV could also bind bovine beta 2m from foetal calf serum in cell culture fluids. The use of radiolabelled beta 2m in other experiments showed that CMV bound beta 2m after release from cells and that the bound beta 2m did not represent acquisition of class I HLA molecules during budding from host cell membranes. Immunoprecipitation studies showed that beta 2m was bound by two viral envelope proteins beta 2m BP1 (beta 2m-binding protein 1) and beta 2m BP2 of molecular masses 36,000 and 65,000 daltons respectively. beta 2m could not bind to separated viral proteins under reducing or non-reducing conditions. We propose that interaction of these two proteins on the viral surface is required to enable CMV to bind beta 2m.

Two clinical isolates and the Toledo strain of cytomegalovirus contain endothelial cell tropic variants that are not present in the AD169, Towne, or Davis strains.

The highly fibroblast‐passaged AD169, Towne, and Davis strains of cytomegalovirus (CMV) were found to have a restricted capacity to infect endothelial cells in vitro. Although such replication could be increased by a combination of low speed centrifugation and sodium butyrate treatment, the extracellular virus produced was infectious for fibroblasts but not for endothelial cells. In contrast, the low passage Toledo strain, and a low passage fibroblast‐grown clinical isolate of CMV, C1F, could be continually passaged in endothelial cells, giving the strains C1FE and Toledo.E. Whilst, using the conditions described above, initial infection of endothelial cells with AD169 or C1F resulted in similar titres of extracellular virus as assayed on fibroblasts, only the virus from the C1F strain was infectious for endothelial cells. Passage of C1F in fibroblasts decreased its ability to infect endothelial cells, whilst retaining equal ability to infect fibroblasts. Although endothelial‐cell‐passaged cell‐free C1FE virus was endothelial cell‐tropic, it was still much more infectious for fibroblasts than for endothelial cells. It is concluded that the C1F and Toledo strains, but not the AD169, Towne, or Davis strains, contained endothelial cell tropic variants, which could be lost on passage through fibroblasts, but retained on passage through endothelial cells. Furthermore, virus in an ex vivo source of CMV, a blood specimen, was found to be more tropic for fibroblasts than for endothelial cells, suggesting that in vivo CMV exists as quasi strains with different cell tropism, some of which might be lost in vitro by passage in an inappropriate cell type. J. Med. Virol. 57:298–307, 1999.

Antibody Responses to Murine Cytomegalovirus in Genetically Resistant and Susceptible Strains of Mice

The role of antibodies as mediators of genetically determined resistance to murine cytomegalovirus (MCMV) in mice has not been elucidated. The ability of mice with different MCMV resistance phenotypes to produce an antibody response to MCMV was investigated in order to assess whether the host genotypes that control resistance also influence antibody production. Antibodies to MCMV in the sera of resistant (BALB.K, CBA/CaH, B10.BR) and susceptible [BALB/c, BALB.B, C57BL/10ScSn (B10), B10.D2, B10.A, A/J] mice were determined by ELISA and/or a complement-requiring neutralization assay. IgM antibodies were produced by all strains of mice as early as 3 to 5 days post-infection (p.i.) with maximum titres observed after 10 days p.i. for some strains, whilst IgG antibodies were produced by 5 to 7 days p.i. with maximum titres at 20 days p.i. IgA antibodies were not detected in the sera of MCMV-infected mice. Virulent MCMV induced higher antibody titres than either attenuated or u.v.-inactivated forms of the virus. Although high doses of virulent virus delayed the early production of IgM antibody they did not adversely affect the kinetics of IgG antibody production. High titres of neutralizing antibodies were detected as early as day 3 post-inoculation of virulent virus; when attenuated virus was used in the neutralization assay, this was found to be more easily neutralized than salivary gland-derived virus. Interestingly, although guinea-pig complement greatly enhanced antibody-mediated neutralization of MCMV, mouse complement was also effective at enhancing neutralization. Although genetically determined resistance to MCMV is an early event with the resistant phenotype being demonstrable during the first few days of the infection, there was no evidence that antibodies were responsible for this resistance since neither antibody titres nor the time of first appearance of antibody correlated with resistance status. However, these results do not exclude a more general role for antibody in limiting MCMV infection, especially in immunity to re-infection since passively transferred antibodies from resistant or susceptible mouse strains lowered virus titres in MCMV-infected animals.