In a previous study we showed that pneumococcal adherence to epithelial cells was enhanced by a preceding respiratory syncytial virus (RSV) infection. RSV-glycoproteins, expressed on the infected cell surface, may play a role in this enhanced pneumococcal binding, by acting as bacterial receptors. In the current study, it was attempted to analyze the capacity of pneumococci to interact directly with RSV virions. By flowcytometry, a direct interaction between RSV and pneumococci could be detected. Heparin, an inhibitor of RSV infectivity that interacts with RSV protein-G, blocked RSV-pneumococcal binding, indicating that the latter interaction is indeed mediated by protein-G. RSV-pneumococcal complexes showed enhanced adherence to uninfected human epithelial cells, compared with pneumococcal adherence without bound RSV, and this enhancement was also blocked by heparin. In addition, the significance of these findings in vitro was explored in vivo in a murine model. Both mice that were pretreated with RSV at day 4 before pneumococcal challenge and mice infected with both agents simultaneously showed significantly higher levels of bacteraemia than controls. Simultaneous infection with both agents enhanced the development of pneumococcal bacteraemia most strongly. It was hypothesized that direct viral binding is another mechanism by which RSV can induce enhanced pneumococcal binding to epithelial cells, a phenomenon that is translated in vivo by a higher invasiveness of pneumococci when administered simultaneously with RSV to mice. Apparently, RSV acts in this process as a direct coupling particle between bacteria and uninfected epithelial cells, thereby increasing colonization by and enhancing invasiveness of pneumococci. Clinical and epidemiologic data suggest that respiratory syncytial virus (RSV) infections in humans can be complicated by bacterial superinfection e.g., with Streptococcus pneumoniae, leading to increased morbidity (1-5). Mechanisms underlying bacterial superinfection include virus-induced local destruction of the epithelium, compromising the host's physiologic barrier, and virus-induced modulation of the immune response (6,7).In addition, enhanced bacterial adherence to virus-infected cells is considered an important factor increasing the risk of bacterial superinfection (8 -11). In a previous study in vitro, we obtained evidence for such a mechanism. A preceding RSV infection of human respiratory tract epithelium led to significantly enhanced adherence of S. pneumoniae (12).The basis of RSV-enhanced pneumococcal adherence is not known. RSV infection both leads to expression of viral glycoproteins and up-regulation of cellular molecules on the hostcell membrane. Both could possibly serve as bacterial receptors, as has been described for Neisseria meningitidis: viral glycoprotein G as well as cellular molecules, like CD 14 and CD18, are involved in enhanced binding of N. meningitidis to RSV-infected cells (13,14

In the present study, we analyzed the effect of a preceding respiratory syncytial virus (RSV) infection of human respiratory epithelial cells on the adherence of Streptococcus pneumoniae tested by means of a cytometric fluorescence assay. Adherence of clinically relevant pneumococcal serotypes 3, 9, 14, 18, 19, and 23 was studied using uninfected and RSV-infected monolayers. To this end, monolayers of both human nasopharyngeal cells (HEp-2) and pneumocyte type II cells (A549) were infected with RSV serotype A. Adherence to uninfected epithelial cells varied between pneumococcal serotypes. After RSV infection of the monolayers, all serotypes showed a strongly (2-to 10-fold) and significantly increased adherence when compared with adherence to uninfected monolayers. Enhanced adherence was observed with both cell lines. By fluorescence and scanning electron microscopy, we observed redistribution of pneumococcal adherence over the epithelial surface due to RSV infection, with dense bacterial accumulations near to epithelial syncytia. Streptococcus pneumoniae and RSV belong to the most important pathogens of upper and lower respiratory tract infections in young children (1, 2). There is accumulating evidence for a positive relationship between infections with S. pneumoniae and RSV, especially in the pathogenesis of otitis media, pneumonia, and meningitis (3-8). Epidemiologically, the peak incidences of RSV infections and invasive infections due to S. pneumoniae coincide (9). In a large study on the etiology of community-acquired bacterial pneumonia, a preceding viral infection could be detected by serology in 39% of the children (10). The combination of RSV and S. pneumoniae was seen most frequently in children below age 5 y.The most prevalent pneumococcal serotypes causing invasive pneumococcal disease in children include serotypes 9, 14, 18, 19, and 23. Pneumococcal sepsis in adults is often caused by serotype 3 (11-14).Enhanced pneumococcal adherence, secondary to a preceding RSV infection of the respiratory tract epithelium, is considered one of the mechanisms facilitating bacterial infection (15,16).The aim of the present study was to examine the influence of prior RSV infection on pneumococcal adherence to confluently grown epithelial cells using an assay based on adherence of fluorescent pneumococci. The influence of RSV-preinfection on pneumococcal superinfection with serotypes 3, 9, 14, 18, 19, and 23 was evaluated. MATERIAL AND METHODS Bacteria.Clinical pneumococcal isolates, serotypes 3, 9, 14, 18, 19, and 23, were kindly provided by Dr. C. Neeleman, Intensive Care Department of the University Hospital of Nijmegen, The Netherlands, and stored at Ϫ70°C in micro-banks (Pro-Lab Diagnostics, Austin, TX, U.S.AϾ). Before testing, an aliquot of stored bacteria was transferred from a micro-bank bead to blood-agar plates and was incubated overnight at 37°C in a CO 2 incubator. The next day, the bacteria were inoculated in Todd-Hewitt broth (Difco, Detroit, MI, U.S.A.), supple-

Fifteen monoclonal antibodies (MAs) directed against either the El or E2 glycoprotein of Semliki Forest virus (SFV) were characterized by immunoglobulin subclass, pl traject, hemagglutination inhibition, neutralization of infectious virus, and protection against virulent infection in mice. All MAs except UM8.4 (immunoglobulin M [IgM]) belonged to various subclasses of IgG and predominantly to IgG2a, but all were unique as indicated by their banding patterns in isoelectric focusing. Competitive binding assays with these MAs revealed the presence of at least six distinct antigenic determinants (epitopes) on the El glycoprotein and five epitopes on the E2 glycoprotein. Two of the epitopes on El, as defined by the properties of the MAs, were associated with hemagglutination inhibition (Elc and Eld), three were associated with neutralization (Ela, Elb, and Elf), and five were associated in various degrees with protection (Ela, Elb, E1c, Ei', and Elf) of mice against virulent SFV infection. With the MAs against E2, the epitopes on E2 were similarly defined. Epitopes E2b and E2e were associated with hemagglutination inhibition, E2c and E2d were associated with neutralization, and three epitopes were associated with in vivo protection (E2', E2C, and E2d). Furthermore, for each MA the relative avidity to purified SFV was determined with an enzyme-linked immunosorbent assay. The binding of some MAs to purified SFV was enhanced by a second MA. The relative avidities of individual MAs did not correlate with their neutralizing capacities. From the results, we suggest that the amino acid sequence which makes up determinant E2d and is recognized by the highly protective MA UM5.1 is an excellent candidate for the production of a synthetic vaccine.

Both neutralizing and nonneutralizing immunoglobulin G2a monoclonal antibodies (MAs) directed against the E2 glycoprotein of Semliki Forest virus (SFV) protected mice prophylactically and therapeutically against virulent SFV infection. The neutralizing MAs, however, conferred protection to mice at lower doses than did nonneutralizing MAs. The antibody-dependent, complement-mediated cytolysis of SFV-infected L cells was effectuated by both kinds of antibodies, but again neutralizing MAs were more effective. Removal of the Fc part of the neutralizing MA UM 5.1 by pepsin digestion resulted in a 100-fold reduction of the neutralization titer (104 versus 106) and a complete loss of its capacity to mediate antibody-dependent, complement-mediated cytolysis. Passive protection of infected mice occurred only after administration of relatively high doses of F(ab')2 of MA UM 5.1 (30.0 ,ug versus 0.1 ,ug). F(ab')2 fragments prepared from the nonneutralizing MA UM 4.2 had lost their protective capacity completely. Surprisingly, the nonneutralizing MA UM 4.2 retarded virus growth in mouse fibroblasts (L cells), although inhibition was at much higher doses than with the neutralizing MA UM 5.1. Furthermore, both MAs promoted the uptake of virulent SFV in the Fc receptor-bearing WEHI-3 cells. The results suggest that nonneutralizing MAs protect mice not only by antibody-dependent, complement-mediated cytolysis but also by growth inhibition and enhanced uptake of SFV in the nonpermissive macrophages of BALB/c mice. This hypothesis is supported by the absence of viremia in recipients of nonneutralizing MA UM 4.2 at 24 h after infection.

No abstract