Background
Therapeutic levels of broadly neutralizing antibodies (bNAbs) can sustain undetectable HIV viraemia. We investigated if post-bNAb control persists after an ART-protected ‘wash-out’ period. In the RIO trial, Arm B participants initially received placebo during a first analytical treatment interruption (ATI-1). Following rebound, participants resumed ART and were offered 3BNC117-LS and 10-1074-LS, and later underwent a second ATI (ATI-2), allowing analysis of viral control when bNAbs had waned.
Methods
Thirty-four participants (18-60 years) were randomized to placebo in Arm B and underwent ATI-1. After rebound, participants resumed ART and received 1-2 bNAb doses (>20wks apart), then underwent ATI-2 a median 6.5 months post-last dose. HIV viral load (VL) was measured weekly during ATI, and within-subject comparisons were made between ATI-1 and ATI-2 for participants who initiated ATI-2, including viral doubling time (DT), peak VL, and HIV-specific IFN-y ELISpot responses. PK analysis of bNAb levels was measured by ELISA. Parameters were compared between early rebound (≤9 weeks; threshold defined by ATI-1 rebound mean+2SD) vs delayed rebound (>9 weeks) in ATI-2.
Results
Of 34 Arm B participants, 29 rebounded in ATI-1, resumed ART, and received bNAbs (3 withdrew; 2 undetectable VL>3 yrs), and 25/29 have initiated ATI-2 to date. At ATI-2 start, median modelled serum levels of 10-1074-LS and 3BNC117-LS were 40.2 and 43.6 ug/mL, respectively – similar to levels observed at rebound in Arm A. No anti-drug antibodies were detected. Compared to ATI-1, ATI-2 showed lower median peak VL (21,400 vs 100,000 cps/mL, p=0.001), and slower median viral DT (0.57 vs 0.31 weeks, p=0.04). Two rebound phenotypes emerged in ATI-2: 12/25 participants showed early rebound with viral kinetics similar to ATI-1, with viral rebound (VL x6>1000 or x2>100,000 cps/mL) within 9 weeks of ATI, and no difference in ATI-2 peak VL or viral DT. In contrast, 13/25 participants demonstrated delayed rebound with sustained control >9 weeks, and significantly lower ATI-2 median peak VL (3,895 vs 19,700 cps/mL, p=0.004) and slower median viral DT (1.16 vs 0.35 weeks, p=0.02) compared to ATI-1. ATI-1 viral kinetics, and bNAb levels and IFN-y ELISpot T cell responses at ATI-2 start, were comparable between rebound phenotypes.
Conclusions
Over half of RIO Arm B participants to date demonstrated delayed rebound and attenuated viral kinetics during ATI-2, suggesting a shift in host-viral dynamics consistent with post-bNAb HIV inhibition.
Table or Figure
