Abstract
Immune responses to human non-self transgenes can present challenges in preclinical studies of adeno-associated virus (AAV) gene therapy candidates in nonhuman primates. Although anti-transgene immune responses are usually mild and non-adverse, they can confound pharmacological readouts and complicate translation of results between species. We developed a gene therapy candidate for Pompe disease consisting of AAVhu68, a clade F AAV closely related to AAV9, that expresses an engineered human acid-alpha glucosidase (hGAA) tagged with an insulin-like growth factor 2 variant (vIGF2) peptide for enhanced cell uptake. Rhesus macaques were administered an intravenous dose of 1x1013 genome copies (GC)/kg, 5x1013 GC/kg, or 1 x 1014 GC/kg of AAVhu68.vIGF2.hGAA. Some unusually severe adaptive immune responses to hGAA presented, albeit with a high degree of variability between animals. Anti-hGAA responses ranged from absent to severe cytotoxic T-cell-mediated myocarditis with elevated troponin I levels. Cardiac toxicity was not dose dependent and affected five out of eleven animals. Upon further investigation, we identified an association between toxicity and a major histocompatibility complex class I haplotype (Mamu-A002.01) in three of these animals. An immunodominant peptide located in the C-terminal region of hGAA was subsequently identified via enzyme-linked immunospot epitope mapping. Another notable observation in this preclinical safety study cohort pertained to the achievement of robust and safe gene transfer upon intravenous administration of 5x1013 GC/kg in one animal with a low pre-existing neutralizing anti-capsid antibodies titer (1:20). Collectively, these findings may have significant implications for gene therapy inclusion criteria.