Serum HIV Neutralization Prediction After Antibody Treatment

Research has identified 3 approaches that may satisfactorily predict serum neutralization of HIV infection and may result in more efficient use of serum samples identified. The results of this study were published in the Journal of Acquired Immune Deficiency Syndromes.

VRC01 is an immunoglobulin G1 (IgG1) broadly neutralizing monoclonal antibody (bnAb) that is currently being evaluated in 2 harmonized phase 2b Antibody Mediated Prevention (AMP) efficacy trials (ClinicalTrials.gov Identifiers: NCT02716675 and NCT02568215). VRC01 targets the CD4 binding site of the HIV-1 envelope glycoprotein. A key function to consider when evaluating the efficacy of bnAbs in preventing HIV infection is virus neutralization activity; however, more resources are required for assays measuring serum neutralizing activity than for measuring serum concentrations of postadministration bnAbs.

Identifying an approach for the prediction of serum neutralization may result in a more efficient use of serum samples and significant resource savings. Therefore, this was the first study that systemically assessed the predictability of postadministration serum neutralization activity based on serum concentrations and in vitro neutralization titers of bnAB.

To predict serum neutralization ID₅₀ titers, the following 3 approaches were used: observed serum concentration divided by inhibitory concentration (IC₅₀), pharmacokinetics model-predicted serum concentration divided by IC₅₀, and joint modeling of the longitudinal serum concentrations and ID₅₀ titers. Serum concentration was measured using an anti-idiotypic enzyme-linked immunosorbent assay, and serum neutralization titers and in vitro neutralization potency IC₅₀ were measured against env-pseudoviruses. In total, 1079 serum samples were collected from 84 participants postadministration of VRC01.

Results showed that VRC01 serum neutralization could be accurately predicted, especially when using pharmacokinetic models. All 3 approaches yielded satisfactory prediction of neutralization titers against viruses of varied sensitivities. The median fold-differences of observed-over-predicted ID₅₀ titers ranged from 0.95 to 1.37. Approaches 1 and 2 showed a slightly lower and more variable prediction performance; however, approach 2 performed marginally better than approach 1. Approach 3 performed the best in terms of prediction, with median fold-differences ranging from 0.95 to 0.99, and a <70% mean squared prediction error when compared with approach 1. Additionally, results were similar for the neutralizing ID₈₀ titers.

In conclusion, approach 1 was shown to be highly resource-efficient, and the incorporation of popPK modeling of longitudinal serum concentrations increased prediction accuracy, which highlights the importance of the collection of pharmacokinetic samples at postadministration times; approach 2 could be used to save resources from assaying longitudinal neutralization titers once the bnAbs are better characterized against more diverse panels of viruses. Approach 3 prediction accuracy improves when paired neutralization longitudinal data and paired serum concentration are used.

Overall, the study authors concluded that, “Our findings have implications for the clinical development of future bnABs, and are also timely in the planning of assays for the AMP trials.”

Reference

Huang Y, Zhang Y, Bailer R, et al. Prediction of serum HIV-neutralization titers after passive administration of VRC01 [published online December 17, 2019]. JAIDS. doi:10.1097/QAI