Protection from cytomegalovirus viremia following glycoprotein B vaccination is not dependent on neutralizing antibodies

Significance

Conventionally, vaccines are screened for induction of a neutralizing antibody response in human volunteers before proceeding to late-stage clinical trials. We present results from a human cytomegalovirus (HCMV) challenge study suggesting that this paradigm may not apply universally to all viruses. Instead, viruses like HCMV, which establish lifelong infections and grow both cell-free and cell-associated, may be controlled independently of a potent neutralizing antibody response. Our results suggest that more detailed laboratory studies are required to identify correlates of immune protection for such viruses and that failure of a vaccine to induce a neutralizing antibody response should not necessarily be considered as a key go-no-go decision point in the design of future vaccine studies.

Abstract

Human cytomegalovirus (HCMV) is an important pathogen in transplant patients and in congenital infection. Previously, we demonstrated that vaccination with a recombinant viral glycoprotein B (gB)/MF59 adjuvant formulation before solid organ transplant reduced viral load parameters post transplant. Reduced posttransplant viremia was directly correlated with antibody titers against gB consistent with a humoral response against gB being important. Here we show that sera from the vaccinated seronegative patients displayed little evidence of a neutralizing antibody response against cell-free HCMV in vitro. Additionally, sera from seronegative vaccine recipients had minimal effect on the replication of a strain of HCMV engineered to be cell-associated in a viral spread assay. Furthermore, although natural infection can induce antibody-dependent cellular cytotoxicity (ADCC) responses, serological analysis of seronegative vaccinees again presented no evidence of a substantial ADCC-promoting antibody response being generated de novo. Finally, analyses for responses against major antigenic domains of gB following vaccination were variable, and their pattern was distinct compared with natural infection. Taken together, these data argue that the protective effect elicited by the gB vaccine is via a mechanism of action in seronegative vaccinees that cannot be explained by neutralization or the induction of ADCC. More generally, these data, which are derived from a human challenge model that demonstrated that the gB vaccine is protective, highlight the need for more sophisticated analyses of new HCMV vaccines over and above the quantification of an ability to induce potent neutralizing antibody responses in vitro.