Brief Report: Prediction of Serum HIV-1 Neutralization Titers After Passive Administration of VRC01

BACKGROUND: VRC01 is a human IgG1 broadly neutralizing antibody (bnAb) that binds to the HIV-1 envelope glycoprotein. It is being evaluated in two ongoing Phase 2b trials, the first efficacy assessment of a passively-administered bnAb for HIV-1 prevention. HVTN 104 was a phase 1 trial of VRC01. SETTING: We measured serum concentrations and serum neutralization of VRC01 in 1079 longitudinal samples collected after passive administration of VRC01 in 84 HVTN 104 participants. As assays for measuring VRC01 serum neutralization titers are resource-intensive, we investigated approaches to predicting such titers. METHODS: Serum concentration was measured using an anti-idiotypic ELISA assay. Serum neutralization ID50 titers and in vitro neutralization potency IC50 of the VRC01 clinical lot were measured against Env-pseudoviruses. Three approaches were used to predict serum neutralization ID50 titers based on (1) observed serum concentration divided by IC50, (2) pharmacokinetics model-predicted serum concentration divided by IC50, and (3) joint modeling of the longitudinal serum concentrations and ID50 titers. RESULTS: All 3 approaches yielded satisfactory prediction of neutralization titers against viruses of varied sensitivities; the median fold differences (FDs) of observed-over-predicted ID50 titers were between 0.95 and 1.37. Approach 3 generally performed the best with fold differences between 0.95 and 0.99 and <82% mean squared prediction error relative to approach 1. Similar results were obtained for ID80 titers. CONCLUSION: VRC01 serum neutralization could be accurately predicted, especially when using pharmacokinetics models. The proposed prediction approaches could potentially save significant resources for the characterization of serum neutralization of VRC01, including for other bnAbs and bnAb combinations.