AAV immunity: more than meets the eye

Comment on Scallan et al, page 1810

Gene transfer with viral vectors is compromised by preexisting host immunity. A novel passive immunity mouse model now provides evidence that anti-AAV humoral responses may be more limiting than previously recognized.

In the pursuit of successful strategies for gene transfer approaches to treat inherited diseases, challenges from the host immune response continue to represent a major obstacle. Until there has been a significant improvement in the ability of nonviral protocols to deliver transgenes, most gene transfer studies will continue to use various forms of viral vectors to effect gene delivery. With this in mind, unless the gene transfer protocol is performed ex vivo, there is a high likelihood that components of the host immune system will be activated by the viral vector, the transgene product, or both.

Of the currently available viral vectors, adeno-associated virus (AAV) has proved the least immunogenic option. Studies of the cellular response to AAV transduction have indicated that this process is far less disruptive than adenoviral transduction,¹  and indeed, there appears to be little, if any, activation of innate immunity following in vivo AAV delivery. Nevertheless, there is an adaptive immune response to AAV vectors that is likely to have an adverse influence on both the transduction efficiency and the long-term persistence of transduced cells following AAV gene transfer. This consequence has been demonstrated recently in a clinical trial of AAV2-mediated gene transfer to the liver in hemophilia B. In this study, one patient achieved factor IX levels of 12% within 2 weeks of gene transfer only to experience a transient episode of hepatotoxicity and loss of transgene expression that has been presumed to be related to preexisting cellular immunity to AAV capsid proteins.²  Unfortunately, this outcome was not predicted by preclinical studies in large-animal models of hemophilia³  in which long-term expression of the therapeutic transgene has now been documented in the absence of any preexisting AAV immunity.⁴ 

Scallan and colleagues have now explored the influence of AAV immunity on gene transfer efficacy in a novel mouse model described in this issue of Blood (see figure). In this model of passive immunity, using pooled human immunoglobulins infused into SCID or NOD-SCID mice, the authors have been able to demonstrate a greater than expected inhibition of AAV-mediated gene transfer than previously suggested from in vitro neutralization assays. This inhibitory effect was most apparent with AAV2 and less evident with pseudotypes AAV6 and AAV8. The use of NOD-SCID mice in this study has also shown that there is an adverse influence of innate immunity on the efficacy of gene transfer with AAV vectors.

In addition to its description of the mouse model of AAV immunity, this article also confirms that a significant number of adults have preexisting neutralizing antibodies (and presumably cellular immunity) to AAV2 (62%). This situation will pose a challenge for effective long-term gene transfer in humans, and an analysis of alternative routes and rates of vector administration performed by Scallan and colleagues suggests that these approaches will not be helpful. In future strategies using this promising vector system, novel serotypes may prove more effective at avoiding neutralization, and the use of empty capsid decoys or transient immunosuppression might also prove beneficial. ▪