Background:
BEAT2 (NCT03588715) tested combined peg-IFN-α2b with two broadly neutralizing antibodies (bnAbs) (3BNC117 and 10-1074) at antiretroviral treatment interruption (ATI) in 14 people living with HIV. Entry criteria included bnAb sensitivity by PhenoSense HIV mAb Assay (IC90 < 2.0 µg/mL (3BNC117) and < 1.5 µg/mL (10-1074)). Primary clinical outcomes are presented elsewhere; here, we report the plasma rebound virus phenotype and evolution.
Methods:
HIV-1 env single genome sequencing (SGS) was performed at rebound and longitudinally through ATI. Rebound lineage consensus Envs were tested for neutralization sensitivity by TZM.bl assay and compared with PhenoSense results. Two participants withdrew during the study, 12 were analyzed.
Results:
SGS of first detectable rebound virus (n=210 sequences) revealed a median of 1 (range 1-3) reactivating virus populations. Two participants rebounded with high levels of both bnAbs during the 26-week period of immunotherapy. Early rebound Envs had complete resistance to 10-1074 (IC90 >10µg/ml), and increased resistance to 3BNC117 compared to entry criteria (median IC90 of 5.9 µg/ml). Eight participants rebounded after cessation of immunotherapy, between 26 and 30 weeks post-ATI, with waning 10-1074 and 3BNC117 plasma levels (median of 71.0 and 8.1 µg/ml, respectively). Rebound Envs in 6/8 participants were resistant to 10-1074, while 5/8 retained sensitivity to 3BNC117 (median IC90 of 1.52 µg/ml). Finally, 2 participants with rebound >50 weeks post-ATI as both bnAb levels fell to < 5 µg/ml, retained sensitivity (median IC90s=0.63 to 10-1074 and 0.28 µg/ml to 3BNC117). Rebound Env neutralization sensitivities largely agreed with PhenoSense cutoffs obtained from similar samples (100% concordance for 10-1074, 92% for 3BNC117). Rebound Env resistance to 3BNC117 correlated with time to rebound (rs=0.82, p=0.013) and weakly correlated with concurrent 3BNC117 plasma levels (rs=0.67, p=0.069). Longitudinal sequencing over ATI showed increasing env diversity via within-lineage evolution and addition of new rebound lineages, both of which led to further increases in resistance to 3BNC117 (median 2.5-fold increase in IC50).
Conclusions:
Parallel analyses of rebound Env sensitivity to administered bnAbs using independent research and commercial assays gave similar results. Rebound Env lineages revealed a high frequency of resistance to both bnAbs (10-1074 > 3BNC117), which increased over the duration of ATI and correlated with time to rebound and diminishing bnAb levels.