RIO Trial: Long-Acting bNAbs 3BNC117-LS and 10-1074-LS Delay HIV Rebound and Accelerate Reservoir Decay After Treatment Interruption
核心洞察
In the randomized RIO trial, 24% of bNAb-treated participants remained off antiretroviral therapy at 96 weeks versus 6% of placebo recipients, with median time to ART restart of 45.4 versus 4.6 weeks.
Early viral rebound before week 20 occurred in 75% of the bNAb arm versus 11% of controls, and rebound viruses were predominantly resistant to 10-1074 rather than 3BNC117.
Pre-existing autologous neutralizing antibodies correlated with prolonged viral control in bNAb recipients, with mean time to ART restart of 108 weeks versus 27.5 weeks in those without such activity.
Two long-acting broadly neutralizing antibodies (bNAbs), 3BNC117-LS and 10-1074-LS, substantially delayed HIV-1 (搜索) viral rebound and prolonged antiretroviral therapy (ART)-free control in adults who initiated treatment during primary or early infection, according to mechanistic analyses from the randomized, placebo-controlled RIO trial published in Nature Medicine.
The RIO trial enrolled 68 participants between 17 May 2021 and 30 July 2024, randomizing them 1:1 to receive the two bNAbs (arm A) or placebo (arm B) two days before analytical treatment interruption (ATI). All participants were men with a median age of 40 years, and most identified as white (58/68, 85.3%). Baseline demographic and clinical characteristics were balanced between arms.
Early rebound before week 20 occurred in 8 arm A and 30 arm B participants — 75% versus 11% (P < 0.001). Yet the median time to ART restart was 45.4 weeks in arm A, where 7 of 29 participants (24%) remained off ART at 96 weeks, compared with 4.6 weeks in arm B, where only 2 of 32 (6%) maintained undetectable HIV-1 (搜索) levels throughout 96 weeks of observation (P < 0.001). Notably, 13 of the 29 individuals in arm A restarted ART before meeting virological restart criteria.
Reservoir Size and Baseline Antibody Activity
Investigators assayed preintervention peripheral blood mononuclear cells using droplet digital PCR (ddPCR) and Q4PCR. The geometric mean number of intact proviruses per million CD4 (搜索)+ T cells was 8.58 and 6.94 by ddPCR (P = 0.5) and 0.27 and 0.28 by Q4PCR (P = 0.5) for arms A and B, respectively — a relatively small reservoir consistent with early therapy. Despite this, intact proviruses showed a median clonality of 57% among individuals with three or more intact proviral sequences.
Baseline reservoir size showed no significant correlation with time to ART restart or peak viremia in either arm. In a non-prespecified exploratory analysis, only 8 of 23 individuals tested showed autologous antibody neutralizing activity, with an 80% inhibitory concentration (IC80) below 250 µg ml−1 of purified IgGs against their reservoir proviruses. No increase in autologous neutralizing activity or expansion of neutralization breadth was observed in samples collected a mean of 68 weeks after ATI.
Escape From 10-1074, Not 3BNC117
Rebound viruses rarely overlapped with dominant latent intact proviral clones. Among 33 intact latent proviral reservoir clones (91 sequences) from 9 control-arm participants, only 2 overlapped with rebound sequences. Rebound viruses in the three control participants with measurable autologous neutralizing antibodies to the reservoir were resistant to their own antibodies.
Selection for bNAb resistance was evident in arm A. The geometric mean neutralizing IC80 for 10-1074 against reservoir and rebound pseudoviruses was 0.09 and 14.7 µg ml−1, respectively, with most rebound viruses completely resistant (IC80 > 50 µg ml−1; P < 0.001). By contrast, the difference for 3BNC117 did not reach statistical significance (P = 0.53), and only 3 of 19 bNAb recipients showed rebound viruses uniformly resistant to 3BNC117. The authors conclude that 10-1074 escape variants were far more likely to emerge than 3BNC117-resistant variants during dual bNAb therapy.
Correlates of Prolonged Control
Among the six bNAb recipients with baseline autologous neutralizing activity, mean time to ART restart was 108 weeks versus 27.5 weeks in the 13 without (P = 0.008). Preinfusion autologous neutralizing antibody titers correlated with time to ART restart in the bNAb arm (n = 14, r = −0.69, P = 0.004). Time to ART restart also correlated with initial reservoir proviral sensitivity to 10-1074 (r = −0.79, P < 0.001), with only a trend for 3BNC117 (r = −0.45, P = 0.06).
Persistent low antibody levels may contribute: mean concentrations of 3BNC117-LS and 10-1074-LS at 96–123 weeks after infusion in seven arm A controllers were 0.70 and 0.96 µg ml−1, respectively. The median estimated 3BNC117-LS concentration at rebound was 52.4 µg ml−1, similar to the median 3BNC117 IC80 of 41.8 µg ml−1 at rebound, supporting the interpretation that the antibody was subtherapeutic at that point.
Fluctuating Viremia and Reservoir Dynamics
In a post hoc subgroup analysis, 11 of 16 remaining arm A participants (68%) maintained fluctuating viremia without meeting ART restart criteria for 16 to more than 58 weeks, compared with only two control-arm participants. Viruses obtained from seven participants during fluctuating viremia were all resistant to preinfusion autologous antibodies when present, and five showed high resistance to 10-1074 while retaining 3BNC117 sensitivity.
Comparing baseline with prerebound samples collected a mean of 52 weeks after infusion (n = 22), the intact reservoir — but not the defective one — was significantly smaller after bNAb therapy, with an estimated half-life of 36 weeks (0.75 years), compared with prior estimates of 4–7 years. No measurable change occurred in arm B (n = 20). The change in reservoir size did not correlate with time to ART restart.
Limitations and Implications
The authors note that sequence information could not be obtained from 33 of 68 participants, limiting sample availability, largely reflecting technical limitations of single-genome amplification in early-treated individuals with small reservoirs. Sampling was restricted to peripheral blood and may not reflect tissue reservoirs, and associations between autologous antibodies and delayed rebound are correlative. Reservoir decay estimates are limited by participants with values below assay detection.
The investigators conclude that a substantial fraction of people living with HIV who receive bNAb therapy can partially control viremia for prolonged periods. Correlates of prolonged control include pre-existing autologous antibodies and pre-existing or enhanced CD8 (搜索)+ T cell responses, suggesting that vaccination to elicit or enhance pre-existing host immunity may increase the number of individuals exhibiting post-bNAb control.