517 - Prevalence of CMV DNA and of infectious virus detected in saliva of young children in childcare centers

Publication Number: 3498.517

John Holik, University of Massachusetts Chan Medical School, Worcester, MA, United States; Anne Mirza, University of Massachusetts Medical School, worcester, MA, United States; Olesea Cojohari, University of Massachusetts Medical School, Worcester, MA, United States; Timothy F. Kowalik, UMass Chan Medical School, Worcester, MA, United States; Laura Gibson, UMass Chan Medical School, Worcester, MA, United States

Background: Young children with primary cytomegalovirus (CMV) shed virus in saliva and urine for prolonged periods. Contact with children who attend large group programs is a risk factor for congenital CMV infection. The CMV Transmission and Immune Tracking (TransmIT) Study seeks to understand CMV spread within childcare centers and to inform future efforts to reduce transmission from children to pregnant caretakers.

Objective: The objective of this study was to measure prevalence of CMV shedding and to compare CMV detection by molecular and culture assays in saliva of children attending childcare. We hypothesized that higher PCR viral load would correlate with presence and higher level of infectious virus.

Design/Methods: Children < 36 months of age (n=64) were enrolled at 13 childcare centers. One saliva sample was collected from each child. CMV DNA and cellular infection were quantified by digital PCR (dPCR) and by modified shell vial (infectivity) assay using fibroblasts or epithelial cells, respectively.

Results: CMV shedding was detected by dPCR in 58 of 64 (90.6%) saliva samples. Viral loads in this assay ranged from 0 to 2.14x10^5 copies/mL of saliva. CMV shedding was detected by infectivity assay in 47 of 64 (73.4%) total samples and in 47 of 58 (81.0%) PCR+ samples. Viral loads in this assay ranged from 0 to 1.04x10^5 infectious virus units/mL of saliva. Overall, 8 of 64 (12.5%) samples were reproducibly discordant with six PCR+/infectivity- and two PCR-/infectivity+. Higher PCR viral load predicted detection and higher amount of infectious virus in saliva, especially for >10^4 genomes/mL. Unexpectedly, CMV in only 24 of 47 (51.1%) infectivity+ samples could infect both cell types used in the assay. CMV in the remaining 23 of 47 (48.9%) samples infected a single cell type: 18 fibroblasts only (38.3% of 47 infectivity+ and 78.3% of 23 single cell type samples) and 5 epithelial cells only (10.6% of 47 infectivity+ and 21.7% of single cell type samples).

Conclusion(s): Most young children in childcare centers shed CMV in saliva detected by dPCR or by infectivity assay. CMV PCR viral load >10^4 genomes/mL may be a biomarker of infectious saliva. Nearly half (48.9%) of infectivity+ viruses infected only one cell type, but over one-third (38.3%) did not infect epithelial cells. Our findings highlight the potential impact of mitigating CMV transmission in childcare settings. They also raise the possibility that distinct CMV populations in salivary glands differ by cell tropism and/or escape immune control and that CMV vaccines targeting epithelial entry proteins may not be optimally effective.