HIV Integration “Hiding Spots” Vary by Tissue, Highlighting Brain as a Unique Reservoir

A new study published in Communications Medicine reveals that HIV‑1 doesn’t integrate randomly into host DNA—but instead shows distinct preferences depending on the anatomical site. In particular, integration in brain tissue appears markedly different from that in blood and gut, a finding that may help explain the virus’s stubborn persistence in the central nervous system.

The researchers analyzed matched tissue samples (esophagus, stomach, duodenum, colon) and blood from people living with HIV subtype B, collected before modern antiretroviral therapy became widespread. They also examined brain samples from a separate cohort of ART‑naive individuals who died of AIDS. Across these samples, they mapped nearly 930 unique HIV‑1 integration sites, then compared their locations relative to genomic features such as genes, repetitive elements, and DNA structural motifs.

They found that, in most tissues (blood, esophagus, stomach, duodenum, colon), HIV integration heavily favored genes—typically more than 70 % of integration events occurred in genic regions. In contrast, in brain tissue only about 57 % of integrations fell within genes. Brain samples were also distinguished by greater enrichment in repetitive elements such as SINEs and a reduced association with non‑B DNA motifs like G‑quadruplexes, cruciforms, triplex DNA, and Z‑DNA, compared to other tissues.

The brain also stood out in hotspot behavior: nearly half of brain-derived integration events clustered within genomic hotspots (defined as multiple events within 100 bp), more so than any other tissue examined. Those hotspots disproportionately mapped to chromosome 9, with a dense focal cluster within a small 313‑base region. Meanwhile, some hotspots were shared across tissues, and some integration sites appeared in multiple individuals—suggesting convergent targeting even across anatomical compartments. 

To rule out artifacts (for example, PCR contamination), the team performed detailed sequence analyses on one universal hotspot in the TBC1D5 gene (chromosome 3). They observed high sequence diversity and characteristic mutation patterns (including APOBEC3‑mediated G→A hypermutations), consistent with independent integration events.

These differences likely reflect a combination of chromatin environment, nuclear architecture, and cell‑type specific host factors. In the brain, diminished expression of LEDGF/p75—a cofactor that guides HIV toward active genes—along with heightened APOBEC3 activity might push viral integration toward transcriptionally silent or repetitive regions. 

Still, the authors caution that the unmatched nature of the brain cohort (ART‑naive) versus other tissues (mostly from ART‑treated individuals) could influence integration patterns. In untreated brains, ongoing replication might produce a broader integration landscape than would survive selective pressures under ART in other compartments. 

Overall, this work emphasizes that HIV integration is not uniform—and that the CNS may harbor viral reservoirs with distinct genomic footprints. Understanding those tissue‑specific integration biases will be vital for designing strategies aimed at eliminating latent HIV from sanctuary sites like the brain.