Engineered adeno-associated virus 3 vector with reduced reactivity to serum antibodies

Mika Ito,Shin-ichi Muramatsu,Takamasa Nomura,Naomi Takino,Akihiko Kan

doi:10.1038/s41598-021-88614-9

Mika Ito, Shin-ichi Muramatsu + Show 3 more

Open Access

https://doi.org/10.1038/s41598-021-88614-9

Copy DOI

Abstract

The natural serotypes of adeno-associated virus (AAV) or their variants, such as AAV8 and AAV5, are commonly used as vectors in the clinical programs for liver-targeted gene therapy. While AAV8 vectors are not highly efficient at targeting primary human hepatocytes, AAV3 vectors have recently demonstrated remarkable efficiency at targeting both human and non-human primate hepatocytes. However, the presence of high levels of neutralizing antibodies (NAbs) impedes transduction into hepatocytes, representing a major obstacle to the clinical application of AAV3 vectors. Herein, we engineered the viral capsid to reduce its reactivity with pre-existing NAbs, thereby enhancing the transduction efficiency. By introducing three substitutions (S472A, S587A, and N706A) on the surface loop of AAV3B capsid protein, we generated a triple mutant AAV3 (AAV.GT5) vector with less reactivity to anti-AAV capsid NAbs. While the transduction efficiency of AAV.GT5 into human hepatocellular cell lines was similar to those of parental AAV3B, it was 50-fold higher for hepatocytes derived from humanized mice compared to AAV8 vectors. Moreover, the AAV.GT5 vector yield was similar to those of the AAV2 and AAV3B vectors. Thus, high resistance to pre-existing NAbs makes AAV.GT5 a promising candidate for future liver-targeted gene therapy clinical trials.

Highlights

The natural serotypes of adeno-associated virus (AAV) or their variants, such as AAV8 and AAV5, are commonly used as vectors in the clinical programs for liver-targeted gene therapy
AAV8-based vectors are approximately 10- to 100fold more efficient than AAV2- or AAV5-based vectors in mouse liver transduction, accumulating preclinical and clinical data indicate that AAV8- and AAV5-based vectors, which are used in most current clinical protocols are not efficient at targeting primary human hepatocytes
Based on the previous reports on the capsid structures of AAV2, AAV3A, AAV3B, AAV8, AAVrh[8], and AAV914–16, as well as antigenic epitope mapping of AAV1, 2, 5 and AAV817, we identified three variable regions (IV, VIII, and IX) in the surface loops of VP3 as important for reducing the immunogenicity

Summary

Introduction

The natural serotypes of adeno-associated virus (AAV) or their variants, such as AAV8 and AAV5, are commonly used as vectors in the clinical programs for liver-targeted gene therapy. The data from preclinical models are not predictive of their clinical performance in h umans[7,8] Another hurdle facing liver-directed gene therapy is the high prevalence of neutralizing antibodies (NAbs), which diminishes the efficacy of AAV-based therapies after systemic administration. Even low titers, such as 1:17 for AAV29 and 1:1 for AAV-Spark[100] (engineered capsid derived from AAVrh74)[3], have been associated with reduced, or abrogated, therapeutic efficacy. Triple substitution of amino acids, S472A, S587A, N706A, on the surface loop of the AAV3B capsid protein was studied to assess their transduction efficiency into primary human hepatocytes and enhanced resistance to pre-existing human NAbs

Methods

Results

Conclusion