BACKGROUND: Naive cells comprise 90% of the CD4 ⁺ T cell population in neonates and exhibit distinct age-specific capacities for proliferation and activation. We hypothesized that HIV-infected naive CD4 ⁺ T cell populations in children on long-term antiretroviral therapy (ART) would thus be distinct from infected memory cells. METHODS: Peripheral blood naive and memory CD4 ⁺ T cells from 8 children with perinatal HIV on ART initiated at age 1.7–17 months were isolated by FACS. DNA was extracted from sorted cells, and HIV proviruses were counted, evaluated for intactness, and subjected to integration site analysis (ISA). RESULTS: Naive CD4 ⁺ T cells containing HIV proviruses were detected in children with 95% statistical confidence. A median 4.7% of long terminal repeat–containing naive CD4 ⁺ T cells also contained HIV genetic elements consistent with intactness. Full-length proviral sequencing confirmed intactness of 1 provirus. In the participant with the greatest degree of naive cell infection, ISA revealed infected expanded cell clones in both naive and memory T cells, with no common HIV integration sites detected between subsets. Divergent integration site profiles reflected differential gene expression patterns of naive and memory T cells. CONCLUSION: These results demonstrate that HIV persisted in both naive and memory CD4 ⁺ T cells that underwent clonal expansion and harbored intact proviruses, and suggest that infected memory T cell clones do not frequently arise from naive cell differentiation in children with perinatal HIV on long-term ART. FUNDING: Center for Cancer Research, NCI; Office of AIDS Research; NCI FLEX; Children’s and Emory Junior Faculty Focused Award.

HIV-1 proviral landscapes were investigated using near-full-length HIV single-genome sequencing on blood samples from five children with vertically acquired infection and on ART for ~7–9 years. Proviral structures were compared to published datasets in children prior to ART, children on short-term ART, and adults on ART. We found a strong selection for large internal proviral deletions in children, especially deletions of the env gene. Only 2.5% of the proviruses were sequence-intact, lower than in the comparative datasets from adults. Of the proviruses that retained the env gene, >80% contained two or more defects, most commonly stop codons and/or gag start mutations. Significantly fewer defects in the major splice donor site (MSD) and packaging signal were found in the children on short or long-term ART compared to the adults, and tat was more frequently defective in children. These results suggest that different selection pressures may shape the proviral landscape in children compared to adults and reveal potentially different genetic regions to target for measuring the intact HIV reservoir and for achieving HIV remission in children.