Humoral immune responses and cytomegalovirus excretion in children with asymptomatic infection

Cytomegalovirus immunoglobulin G titers do not predict reactivation risk in immunocompetent hosts

Cytomegalovirus (CMV) reactivation occurs in roughly one-third of immunocompetent patients during critical illness, and is associated with worse outcomes. These outcomes have prompted consideration of early antiviral prophylaxis, but two-third of patients would receive unnecessary treatment. Tissue viral load has been associated with risk of reactivation in murine models, and recent work has suggested a relationship between immune responses to CMV and underlying viral load. We, therefore, sought to confirm the hypothesis that serum CMV-specific immunoglobulin G (IgG) correlates with tissue viral load, and might be used to predict the risk of reactivation during critical illness. We confirm that there is a good correlation between tissue viral load and serum CMV-specific IgG after laboratory infection of inbred mice. Further, we show that naturally infected outbred hosts have variable tissue viral DNA loads that do not correlate well with serum IgG. Perhaps as a consequence, CMV-specific IgG was not predictive of reactivation events in immunocompetent humans. When reactivation did occur, those with the lowest IgG levels had longer durations of reactivation, but IgG quartiles were not associated with differing peak DNAemia. Together our data suggest that CMV-specific IgG titers diverge from tissue viral loads in outbred immunocompetent hosts, and their importance for the control of reactivation events remains unclear.

Human antibody reactivity against the lower matrix protein (pp65) produced by cytomegalovirus

The lower matrix protein (pp65) is a major product of many laboratory strains of cytomegalovirus (CMV). It is thus an integral part of many CMV serological assays based on native antigen. Recombinant fragments of pp65 have previously been investigated for their usefulness in more-defined assays. The latter antigens have, however, failed to develop a positive response with serum samples derived from a substantial number of infected individuals. Here we show that the human humoral immune response to CMV pp65 is highly diverse and recognizes at least seven distinct but in some cases partly overlapping epitopes. Most of these epitopes could not be mimicked by any of the investigated recombinant or synthetic antigens. Furthermore, when we investigated the ability of human CMV-seropositive serum samples to block the reactivity of pp65-specific antibodies recognizing five different epitopes within pp65, it was evident that several sera did not contain significant levels of antibodies against any of these or overlapping structures. It was thus concluded that the antibody response against CMV pp65 is weak in some CMV-infected individuals, making this antigen unsuitable for use alone in serological screening systems for CMV infection.