Short Communication: Persistence of HIV After Allogeneic Bone Marrow Transplant in a Dually Infected Individual

Abstract

Allogeneic bone marrow transplant (alloBMT) with continuous antiretroviral therapy alone has not been shown to completely eradicate HIV, possibly due to HIV persistence in rare residual host cells or infection of donor cells. Within a trial of alloBMT in individuals with hematological malignancies and HIV (ClinicalTrials.gov, NCT01836068), we measured HIV reservoirs longitudinally using a quantitative viral outgrowth assay. We sequenced the reverse transcriptase region of pol for replication-competent virus and performed maximum-likelihood phylogenetic reconstruction. Replacement of host cells was measured using short-tandem repeats. In one participant who had ≥99.5% donor cell replacement, HIV reservoirs declined from 2.2 infectious units per million to undetectable levels at post-alloBMT time points except for week 64. Sequence analysis revealed dual infection pre-alloBMT. Replication-competent virus isolated at week 64 post-alloBMT was identical to a pre-alloBMT variant. This report provides proof-of-concept that minor replication-competent HIV variants can persist at low levels despite ≥99.5% donor cell engraftment post-alloBMT.

Keywords: HIV latent reservoir; HIV persistence; allogeneic bone marrow transplant; dual infection.

Conflict of interest statement

C.M.D. reports grants from Abbvie, grants from GlaxoSmithKine, and from Gilead Sciences, out of the submitted study. All other authors have nothing to disclose.

Figures

FIG. 1.

HIV persistence after alloBMT. (A) Measurement of HIV persistence after alloBMT. The top graph shows the clinical parameters of HIV disease for absolute CD4 count (blue circles) and HIV plasma RNA (red circles). Open symbols indicate below the LOD (red dotted line <20 HIV RNA copies/mL). The bottom graph shows donor engraftment of unsorted PBMCs (black squares) and sorted CD3+ T cells (gray squares) as well as the size of the inducible replication-competent HIV reservoir (purple triangles). Open symbols indicate below the LOD (purple dotted line <0.007 IUPM based on cell input). Relevant clinical history is denoted by arrows. The shaded gray area indicates period of febrile syndrome. The dotted arrow shows week 74 when the LN biopsy was taken. (B) Maximum-likelihood phylogenetic reconstruction of participant 7 based on the RT-pol region (lnL = −552.57). Maximum-likelihood inference was under the TN92 model (corrected Akaike information criterion = 1211.04) of evolution with bootstrapping of 1,000 pseudoreplicates and nearest-neighbor interchange branch swapping. The tree was rooted on subtype D reference with four subtype B reference sequences included. The scale is proportional to genetic variation. Samples isolated 6.1 weeks pre-alloBMT (red) and 64 weeks post-alloBMT (blue) are denoted as colored circles. Majority consensus sequences from deep sequences were generated for the four p24-positive wells where next-generation sequencing was used instead of Sanger sequencing. The “major” and “minor” groups are denoted to indicate dual infection. Bootstrap support (%) is indicated on the branches. IUPM, infectious units per million cells; LN, lymph node; LOD, limit of detection; PBMCs, peripheral blood mononuclear cells; RT, reverse transcriptase; alloBMT, allogeneic bone marrow transplant. Color images are available online.