467. Assay Conditions Significantly Alter AAV Neutralizing Antibody Determination

Abstract

A major limitation of vectors based on current AAVs is the wide prevalence of memory B and T cells reactive to the AAV capsid in human populations arisen following natural AAV infection. The Neutralizing Antibody (NAB) assay guides experimentation and determines inclusion or exclusion of subjects in clinical studies in order to predict in vivo neutralization. Furthermore, evidence of a memory B-cell response to AAV may correlate with anti-capsid memory T-cell populations implicated in the safety and efficacy outcome of clinical AAV liver studies. Based on previously reported findings, NAB assay protocols in current use are unable to robustly predict in vivo neutralization in monkeys (Zinn et al. Cell Reports, 2015 & Wang et al. Mol Ther, 2010) leading to false negatives. In a clinical setting, the enrollment of a seropositive individual exposes the subject to undue risk, and generates outcome variability of the often small clinical studies. Here, we sought to permutate several of the NAB assay conditions in order to evaluate their impact on the NAB titer readout. Specifically, we varied the following variables: cell line, transgene detection, use of adenovirus, and AAV amount. Prior to evaluating these assay permutation, we sought to increase the sensitivity of the readout for transgene expression in the assay in order to evaluate at all conditions both high and low transducing AAV serotypes. This optimization increased our sensitivity by ~200-fold. With this increased sensitivity of readout, we first sought to evaluate the compliance of standard AAVNAB assays with the percentage law, formulated by Andrewes and Elford in 1933, a standard for virological NAB assays, which seeks stoichiometric conditions of NAB and viral load that enables measurement of NAB titers independent of viral input in the assay. Our data shows that at the conditions in common use in the field the percentage law is not upheld, and NAB titers are highly dependent on input viral load of the assay, directly impacting sensitivity and specificity, and leading to the occurrence of false negatives. Indeed, quantitative assessment of seroprevalence in primate populations with assays that reduce the amount of AAV to within percentage law range affects assay outcome qualitatively and quantitatively. We also independently assessed other assay parameters to improve sensitivity, robustness and reproducibility, and to study whether they may also have qualitative effects on titer outcomes. Our studies highlight the importance of pre-existing immunity in AAV gene therapy and limitations of the current methodologies to measure it in a robust and predictive manner. Data indicates the potential for false positive and negative readout for commonly used assay protocols. An optimized protocol was developed for further validation and evaluation. These studies may impact translational and clinical AAV gene therapy studies.