Heterosubtypic, cross-reactive immunity to human Cytomegalovirus glycoprotein B

Cytomegalovirus (CMV) genome is highly variable and heterosubtypic immunity should be considered in vaccine development since it can enhance protection in a cross-reactive manner. Here, we developed a protein array to evaluate heterosubtypic immunity to CMV glycoprotein B (gB) in natural infection and vaccination. DNA sequences of four antigenic domains (AD1, AD2, AD4/5, and AD5) of gB were amplified from six reference and 12 clinical CMV strains, and the most divergent genotypes were determined by phylogenetic analysis. Assigned genotypes were in vitro translated and immobilized on protein array. Then, we tested immune response of variable serum groups (primarily infected patients, reactivated CMV infections and healthy individuals with latent CMV infection, as well gB-vaccinated rabbits) with protein in situ array (PISA). Serum antibodies of all patient cohorts and gB-vaccinated rabbits recognized many genetic variants of ADs on protein array, including but not limited to the subtype of infecting strain. High-grade cross-reactivity was observed. In several patients, we observed none or neglectable immune response to AD1 and AD2, while the same patients showed high antibody response to AD4/5 and AD5. Among the primary infected patients, AD5 was the predominant AD, in antibody response. The most successful CMV vaccine to date contains gB and demonstrates only 50% efficacy. In this study, we showed that heterosubtypic and cross-reactive immunity to CMV gB is extensive. Therefore, the failure of CMV gB vaccines cannot be explained by a highly, strain-specific immunity. Our observations suggest that other CMV antigens should be addressed in vaccine design. We established a cytomegalovirus (CMV) protein array that allows the evaluation of heterosubtypic immunity to the four major antigenic domains (AD1, AD2, AD4/5, and AD5) of glycoprotein B (gB) in natural infection and vaccination. Our results showed that heterosubtypic and cross-reactive immunity to CMV gB is extensive. Therefore, the failure of CMV gB vaccines so far cannot be explained by a highly, strain-specific immunity.