Rational Vaccine Design Against Cytomegalovirus

Abstract

Human cytomegalovirus (CMV) is the most common cause of congenital infection worldwide, affecting approximately 1 in 150 infants, and is a leading cause of morbidity and mortality among transplant recipients. Congenital CMV (cCMV) infection can lead to permanent hearing loss, brain damage, and neurodevelopmental delay, and cCMV alone is responsible for nearly 25% of all infant hearing loss. Chronic CMV infection has also been associated with a heightened inflammatory state and increased risk of aging-related diseases, including as cardiovascular disease and type 2 diabetes. Over the last fifty years, there have been many efforts to develop vaccines that can prevent CMV disease. However, vaccine development for CMV faces many challenges, including a limited understanding of the immune responses protective against infection.Like other beta herpesviruses, CMV can establish a lifelong, persistent infection in hosts, marked by periods of latency and reactivation, and preexisting immunity does not protect against reinfection. Moreover, CMV is known to predominantly spread by direct cell-to-cell transmission, complicating efforts to design vaccines that not only prevent viral entry via antibody neutralization but also inhibit the spread of cell-associated virus. To identify targets for vaccine development, we investigated the antibody immune responses associated with protection from CMV in historical vaccine trials and the lineage maturation of neutralizing antibodies elicited in natural infection.The most efficacious CMV vaccine to-date is the glycoprotein B (gB) subunit vaccine combined with the MF59 adjuvant (gB/MF59), which achieved 50% protection against primary CMV acquisition in multiple Phase 2 clinical trials. CMV gB is a viral envelope protein that mediates fusion with host cell membranes and is required for viral entry into all known cell types and for cell-cell spread. Previous studies had found that in cohorts of CMV-seronegative postpartum women and renal transplant recipients, gB/MF59 vaccination did not elicit broadly neutralizing antibodies but instead generated robust nonneutralizing antibody responses, namely antibody-dependent cellular phagocytosis (ADCP). These studies lacked the statistical power to determine whether ADCP responses contributed to the partial efficacy of the gB/MF59 vaccine. We aimed to define the immune responses correlated with protection from primary CMV acquisition in Phase 2 gB/MF59 clinical trials in cohorts of adolescent girls and postpartum women. We first evaluated the vaccine-elicited sera IgG binding, neutralizing, and nonneutralizing responses against CMV, and we observed distinct immunogenicity profiles in the adolescent and postpartum cohorts, wherein adolescent vaccinees but not postpartum vaccinees developed broadly neutralizing CMV antibodies. We then compared the sera antibody responses between vaccinees who acquired infection and those who remained uninfected during the course of the trial. By multiple logistic regression analysis, we found that protection against primary CMV acquisition was associated with the presence of sera IgG binding to cell-associated gB, but not IgG binding to soluble gB as used in the gB/MF59 vaccine. These results suggested that there may be conformational differences between cell-associated and soluble gB. Supporting this, we identified gB-specific monoclonal antibodies (mAbs) that differentially recognized these gB structures. Our findings indicated the importance of the native, cell-associated gB conformation in future CMV vaccine design.In our immunogenicity studies of gB/MF59, we observed that the vaccine failed to elicit antibodies against the gB antigenic domain 2 site 1 region (AD-2S1), which is a highly conserved, linear epitope at the far N terminus of gB that is known to be a target for potently neutralizing antibodies in natural infection. The presence of sera antibodies against gB AD-2S1 in naturally infected individuals have been associated with decreased risk for cCMV transmission and less severe CMV disease in transplant recipients. Yet, only about half of naturally infected individuals develop anti-gB AD-2S1 antibodies, and gB-based vaccines to-date have failed to elicit these responses. Thus, it remained unclear how to generate neutralizing antibodies against this poorly immunogenic epitope by vaccination. With the goal to identify a gB AD-2S1 structure that could elicit these potently neutralizing antibodies from the germline, we employed a B cell lineage-targeted vaccine strategy. We mapped the phylogeny of a well-characterized, potently neutralizing anti-gB AD-2S1 mAb from its germline precursor, then by empiric testing, we identified the antibody mutations that conferred neutralizing function. We found that a single heavy chain mutation in the CDR1 region was both necessary and sufficient to confer neutralizing function to the otherwise nonneutralizing IGHV3-30/IGKV3-11 germline ancestor mAb. This study identified a critical, early B cell receptor mutation that can serve a target for mutation-guided CMV vaccine design.Only a limited number of gB AD-2S1 mAb sequences have been published to-date, with few clonally related members. To identify additional antibodies for lineage analysis and for evaluation as potential passive therapeutics, we developed a pipeline to sequence anti-CMV gB AD-2S1 mAbs and estimate their antigen binding in high-throughput. This study leveraged modern advances in B cell sequencing technologies to expedite the discovery of potently neutralizing CMV antibodies. In sum, this work contributes to our understanding of protective immunity to CMV. We found that the immune correlate of protection for the gB/MF59 vaccine was sera IgG binding to cell-associated gB, suggesting that the next generation of gB-based vaccines should be designed to adopt a cell-associated or native conformation. Then, we investigated the lineage maturation of potently neutralizing antibodies against CMV gB AD-2S1 and identified a critical mAb mutation from the germline associated with the development of neutralization function, which will may serve as target for immunogen design. Additionally, we developed a pipeline for the high-throughput isolation of anti-gB AD-2S1 mAbs. These findings provide insight into antibody-mediated protection from CMV and illuminate paths forward for rational vaccine design.