Epstein-Barr virus (EBV) is transmitted by saliva and is a major cause of cancer, particularly in people living with HIV/AIDS. Here, we describe the frequency and quantity of EBV detection in the saliva of Ugandan adults with and without HIV-1 infection and use these data to develop a novel mathematical model of EBV infection in the tonsils. Eligible cohort participants were not taking antiviral medications, and those with HIV-1 infection had a CD4 count >200 cells/mm^3. Over a 4-week period, participants provided daily oral swabs that we analysed for the presence and quantity of EBV. Compared with HIV-1 uninfected participants, HIV-1 coinfected participants had an increased risk of EBV detection in their saliva (IRR=1.27, 95% CI=1.10-1.47) and higher viral loads in positive samples. We used these data to develop a stochastic, mechanistic mathematical model that describes the dynamics of EBV, infected cells, and immune response within the tonsillar epithelium to analyse potential factors that may cause EBV infection to be more severe in HIV-1 coinfected participants. The model, fit using Approximate Bayesian Computation, showed high fidelity to daily oral shedding data and matched key summary statistics. When evaluating how model parameters differed among participants with and without HIV-1 coinfection, results suggest HIV-1 coinfected individuals have higher rates of B cell reactivation, which can seed new infection in the tonsils and lower rates of an EBV-specific immune response. Subsequently, both these traits may explain higher and more frequent EBV detection in the saliva of HIV-1 coinfected individuals.

Data on EBV shedding in adults from Kampala, Uganda and Seattle, Washington, USA, are included here. 

Uganda_ebv_data.csv contains information on EBV shedding in the saliva, genital tract, and in the blood of Uganda cohort participants.

Uganda_baff_cd4_hiv.csv contains information on the B cell activation factor amounts, CD4 T cell counts, and HIV RNA loads found in the blood of Uganda cohort participants.

Seattle_ebv_data.csv contains information on EBV shedding in the saliva of Seattle cohort participants.

Key for Data Labels

• Participant: unique identifying number for each participant of a cohort
• Site: site of sample collection 
• StudyDay: number of days since beginning the observation of a participant
• hiv_status: HIV-1 infection status (NEG = negative, POS = positive)
• CD4: CD4 T cell count/mm^3 in blood
• BAFF: B cell activation factor levels in pg/ml of serum
• HIV: HIV-1 RNA copies/ml of blood
• V: EBV DNA copies/ml as measured by qPCR. Note the threshold of detection was 150 DNA copies/ml for mucosal samples and 50 DNA copies/ml for serum samples