The RIO Trial’s Viral Rebound Data Just Complicated Every HIV Functional Cure Strategy on the Market

Picture a regulatory affairs director at an HIV cure-focused biotech opening the secondary and exploratory outcomes paper from the RIO trial in Nature Medicine on a Thursday morning. She already read the primary results published in The Lancet HIV on July 30, 2026. She knows the headline: 3BNC117-LS and 10-1074-LS, a dual broadly neutralizing antibody (bNAb) combination, significantly delayed viral rebound after analytical treatment interruption (ATI) compared to placebo. Her pipeline deck is already built around that signal. Then she opens the secondary outcomes data and finds the resistance tables. The question that follows is the one every HIV functional cure sponsor now has to answer: what exactly are you controlling, and for how long?

The RIO trial, a Phase 2 double-blind, randomized, placebo-controlled study, enrolled adult males living with HIV who had initiated antiretroviral therapy (ART) during recent HIV infection. The primary endpoint measured time to viral rebound in the weeks following ART interruption. Viral rebound was defined according to a pre-specified plasma HIV RNA threshold, with a rigorous multi-measurement rule designed to filter signal from noise., a definition rigorous enough to generate credible primary data in a regulatory context. The bNAb combination met that primary endpoint. But the secondary outcomes published in Nature Medicine are where the operational complexity begins, and where most functional cure development programs will find either their runway or their ceiling.

What the Resistance Data Actually Shows

The secondary analyses revealed something that disrupts a convenient assumption embedded in most bNAb development narratives: that a longer rebound delay necessarily translates into a durable immunological advantage. Participants who received the 3BNC117-LS and 10-1074-LS combination and eventually experienced viral rebound showed evidence of resistance mutations to one or both antibodies at rebound. This is not a peripheral safety footnote. Resistance emergence under bNAb pressure is a protocol design problem, a future-indication problem, and in some cases a combinatorial strategy problem that resets the entire development calculus.

Consider what this means structurally. The RIO bNAbs target two distinct HIV envelope epitopes: 3BNC117-LS binds the CD4 binding site, while 10-1074-LS targets the V3 glycan supersite. The dual-targeting rationale is precisely designed to raise the genetic barrier to resistance, the same logic that drove the shift from mono- to combination ART in the 1990s. Finding resistance at rebound suggests HIV’s escape pathways remain more accessible than the dual-epitope hypothesis predicted, at least in the subset of participants whose virus rebounded on-study. The exploratory analyses in the Nature Medicine paper parsing the breadth and potency of the immune response at rebound are where sponsors need to focus their reading.

The comparison to VRC01 is instructive here. The HVTN 704/HPTN 085 and HVTN 703/HPTN 081 trials, which tested VRC01 for HIV acquisition prevention and were published in the New England Journal of Medicine, demonstrated that single-antibody approaches face an intrinsic ceiling: VRC01-resistant viral strains circulate at meaningful frequency in the population, blunting efficacy in exposed participants before the antibody even starts its job. The RIO trial advances the field by combining two antibodies, but the resistance signals at rebound suggest the ceiling has been raised rather than removed.

The ATI Design Problem Sponsors Keep Avoiding

Analytical treatment interruption is the methodological engine of virtually every HIV functional cure trial being designed right now, and it carries regulatory risks that sponsors persistently underestimate. ATI requires participants to stop suppressive ART under close monitoring, accepting transient viremia in exchange for information about whether an investigational intervention can sustain control. Regulatory guidance specifically addressing ATI design standards remains limited, a gap that creates interpretive latitude but also exposes sponsors to variable reviewer expectations.

The RIO trial addressed this by building in intensive viral load monitoring during the ATI period and defining viral rebound with a six-consecutive-measurement rule, a threshold stringent enough to filter signal from noise. That design choice generates credible primary data. It also means the secondary outcomes, including immune response measures and the resistance profiles at rebound, carry unusual weight, because the primary endpoint integrity allows the exploratory findings to be interpreted without the confounder of a poorly controlled ATI window. Sponsors designing their own ATI protocols should treat the RIO monitoring schedule as a floor, not a ceiling.

Gilead Sciences is running directly adjacent to this territory. At CROI 2025, Gilead presented Phase 2 data showing its investigational twice-yearly regimen combining lenacapavir with bNAbs teropavimab (GS-5423) and zinlirvimab (GS-2872) met its primary endpoint in a Phase 2 study. That program uses a different antibody pair and combines bNAbs with a capsid inhibitor rather than testing bNAbs in isolation during ATI. The RIO secondary data creates an immediate interpretive obligation for Gilead’s team: if dual-bNAb pressure without a backbone antiviral generates resistance mutations at rebound, does adding lenacapavir suppress that escape, or does it merely defer it to a later, harder-to-characterize timepoint?

That question does not yet have an answer in the published record.

The Operational Takeaway for Functional Cure Protocol Teams

The most counterintuitive lesson from the RIO secondary outcomes is about what success looks like in a bNAb ATI trial. The conventional instinct is to optimize for rebound delay: longer is better, and a participant who rebounds at week 18 versus week 6 represents a better outcome. But the resistance data inverts that logic in a critical scenario. A participant who rebounds quickly, before significant antibody-driven selection pressure accumulates, may rebound with a virus that is still sensitive to the study antibodies. A participant who rebounds late, after sustained exposure to 3BNC117-LS and 10-1074-LS, may rebound with a partially resistant virus that complicates re-treatment. The duration of viral suppression and the quality of immune reconstitution are not the same variable, and the RIO secondary endpoints make that distinction visible in ways the primary rebound timing data cannot.

For sponsors building Phase 2b or Phase 3 ATI designs, several operational decisions follow from this. First, pre-specify resistance testing at rebound as a primary secondary endpoint, not an exploratory add-on. The RIO team did this; sponsors who demote resistance testing to exploratory status risk generating data the FDA will treat as hypothesis-generating rather than inferential. Second, build viral sequencing into the monitoring schedule at multiple rebound timepoints, not just the first measurement that meets the rebound definition. Third, if your protocol includes any provision for re-initiating ART after rebound, define the re-initiation criteria prospectively and tie them to both quantitative viral load thresholds and resistance phenotype, not quantitative load alone. Protocol-specified re-initiation criteria, rather than reactive ones, are essential to maintaining GCP compliance and supporting a credible regulatory submission.

The RIO trial also carries a demographic specificity that future sponsors must either replicate or explicitly justify diverging from. The trial enrolled adult males, a design choice that controlled hormonal and pharmacokinetic variability but limits the generalizability of the secondary immune response data to female participants, who represent a substantial share of new HIV diagnoses globally. Any Phase 3 program that cites the RIO secondary outcomes as its scientific rationale will need an explicit diversity plan under the FDA’s established expectations for enrollment that reflects the disease burden, including in women and in populations with non-subtype B virus, where bNAb breadth data are still thin.

Which brings the analysis back to that regulatory affairs director and her pipeline deck. The RIO primary data gave her a rebound delay signal. The secondary outcomes gave her a resistance map. Those two datasets together define not just what 3BNC117-LS and 10-1074-LS can do, but where the next generation of combination strategies has to go: higher genetic barrier, earlier immune intervention, or both. Sponsors who design Phase 3 protocols that build the RIO resistance findings into their primary endpoint structure, rather than treating them as a background footnote, will be better positioned to shape the evidentiary standard for future functional cure filings.

References

1. Nature Medicine, “Broadly neutralizing antibodies in adult males living with HIV undergoing analytical treatment interruption: secondary and exploratory outcomes of the phase II randomized controlled RIO trial”
2. The Lancet HIV / PubMed, “Time to HIV rebound after infusion of long-acting broadly neutralising antibodies 3BNC117-LS and 10-1074-LS and analytical treatment interruption (the RIO trial): a double-blind, randomised, placebo-controlled trial” (July 30, 2026)
3. New England Journal of Medicine, “Two Randomized Trials of Neutralizing Antibodies to Prevent HIV-1 Acquisition” (HVTN 704/HPTN 085 and HVTN 703/HPTN 081, VRC01 trials)
4. Gilead Sciences, “Gilead Presents New HIV Treatment and Cure Research Data at CROI 2025, Including an Investigational Long-Acting, Twice-Yearly Therapy Option” (2025)
5. PubMed, RIO trial protocol: randomised, placebo-controlled, double-blinded Phase II study of dual long-acting bNAbs for post-treatment viral control in recent HIV infection