Adeno-Associated Viruses (AAV) and Host Immunity - A Race Between the Hare and the Hedgehog

Abstract

Adeno-associated viruses (AAV) have emerged as the lead vector in clinical trials and form the basis for several approved gene therapies for human diseases, mainly owing to their ability to sustain robust and long-term in vivo transgene expression, their amenability to genetic engineering of cargo and capsid, as well as their moderate toxicity and immunogenicity. Still, recent reports of fatalities in a clinical trial for a neuromuscular disease, although linked to an exceptionally high vector dose, have raised new caution about the safety of recombinant AAVs. Moreover, concerns linger about the presence of pre-existing anti-AAV antibodies in the human population, which precludes a significant percentage of patients from receiving, and benefitting from, AAV gene therapies. These concerns are exacerbated by observations of cellular immune responses and other adverse events, including detrimental off-target transgene expression in dorsal root ganglia. Here, we provide an update on our knowledge of the immunological and molecular race between AAV (the "hedgehog") and its human host (the "hare"), together with a compendium of state-of-the-art technologies which provide an advantage to AAV and which, thus, promise safer and more broadly applicable AAV gene therapies in the future.

Keywords: AAV; antibody response; capsid; cellular response; engineering; immune evasion; neutralizing antibodies; pre-existing immunity.

Figure 1

Immune responses against AAVs. AAVs delivered systemically can be neutralized by pre-existing antibodies prior to entering the cells. If they evade nAb binding, they enter the cells through endocytosis and can then be degraded in the endosomes. In certain cell types, such as DCs, their genome or capsid can be sensed by TLR9 and TLR2, respectively, which induces the innate immune response. Alternatively, AAVs can successfully escape the endosome and traffic in the cytoplasm, where they can be ubiquitinated, resulting in capsid degradation. The ensuing peptides can next be loaded to MHC class I molecules and presented on the surface of the cells, which are then targeted and possibly eliminated by CD8+ T-cells (CTL response). After endosomal escape, AAVs can also successfully transduce the cell and deliver their viral genome to the nucleus, where the transgene is expressed. Any misfolded protein encoded by the transgene can be degraded by the proteasome and the ensuing peptides can be loaded onto MHCI and also provoke a CTL response. Sensing of AAV vector components in plasmacytoid DCs (pDCs) by the innate immune system leads to the activation of conventional DCs (cDCs). cDCs employ antigen presentation to cross-prime CD8+ T-cells towards an effector type (Teff) and to activate CD4+ T-cells. The latter can activate B cells, which in turn produce the nAbs against the capsid and transgene product. After prolonged or inadequate stimulation, the CD8+ T-cells can be eliminated by different mechanisms including T-cell exhaustion, anergy or apoptosis. The tolerogenic environment in the liver can also stimulate the production of regulatory T-cells (Tregs), which can suppress the aforementioned immune responses at different stages.

Figure 2

Strategies to evade immune responses. Several approaches to evade immune responses have been applied in the clinic or explored on a basic research level. They can be roughly divided into two main classes, i.e., 1) modulation or suppression of the subject’s immune system and 2) AAV vector engineering. 1) Modulation/suppression of the immune system. The route of administration is decisive in evading pre-existing immunity. Targeting immune-privileged organs, when possible, provides protection not only by evading immune responses, but also by inducing regulatory responses. The latter can also be induced through multiple exogenous interventions. Additionally, pharmacological immune suppression has been used extensively in the clinic. Pre-existing immunity (nAbs) is particularly difficult to evade or suppress. Promising approaches include plasmapheresis or the use of immunoadsorption columns ex vivo after separation from the cellular parts of the blood. Furthermore, recent studies have used IgG-degrading enzymes in vivo. 2) AAV vector engineering. The use of native serotypes from species other than humans or NHPs, or engineering the capsid of existing serotypes (predominantly from primate species) are advantageous strategies to evade immunity. The current serotypes can be modified in defined positions (rational design) based on acquired knowledge about sequence-structure-function relationships. To increase variability, the rational design strategy can be used to generate libraries, which then can be evolved/selected ex or in vivo. Libraries can also be fully randomized and interrogated for multiple properties, not only immune evasion. Moreover, the viral genome can be optimized to minimize antigen presentation (e.g., via cell-type specific promoters, miRNAs to prevent expression in APCs or peptides to inhibit antigen presentation) or immune system induction (CpG depletion). Finally, the capsid can be protected from the immune system by coating with molecules, such as PEG, engulfment in exosomes, or display of immune evasion peptides on its surface.