Announcements

Authors: Kentaro Yamada¹, Di Cao¹, Ning Li¹, Roland W. Herzog¹, Dongsheng Duan², Sandeep R.P. Kumar¹ 

¹Department of Pediatrics, Herman B Wells Center for Pediatric Research, Indiana University, Indianapolis, IN, USA;
²Department of Molecular Microbiology and Immunology, School of Medicine, University of Missouri, Columbia, MO, USA. 

Correspondence:
Sandeep R. P. Kumar, Ph.D.
Herman B Wells Center for Pediatric Research
Department of Pediatrics
Indiana University
1044 W. Walnut Street
Indianapolis, IN 46202, USA
E-mail: sankuma@iu.edu 

In the context of muscle gene delivery, we recently demonstrated in a murine model that the innate sensing of adeno-associated virus (AAV) vectors and their encoded transgenes in muscle tissue is quite complex and involves multiple innate sensors and associated pathways¹. Innate immune sensors recognize unique molecular structures associated with invading microbes and trigger a rapid cascade of events, which enables communication with the key players involved in adaptive immunity to mount an effective anti-microbial response. Though critical to fighting off invading microbes, innate immune responses are one of the major impediments to the success of viral vector-based gene therapies to treat genetic diseases. Among the various viral vectors, AAV-based vectors are preferred for in vivo gene delivery due to their high degree of tissue tropism and comparatively low immunogenic profile². However, the immunotoxicity observed in multiple clinical trials clearly indicates that AAV vectors are capable of activating the host immune system and rendering gene therapy ineffective. Despite the fact that multiple AAV-based gene therapy products have received regulatory approval, the immunogenicity of AAV and its derivative vectors remains incompletely understood. Over the past 15 years, multiple innate immune sensors such as toll-like receptor (TLR) 2 (AAV capsid sensing), TLR9 (AAV genome sensing), and MDA5-MAVS (dsRNA sensing) have been implicated in AAV immunogenicity³. Among these, TLR9 and the downstream adaptor molecule, myeloid differentiation primary response 88 (MyD88), are most firmly established as the drivers of anti-AAV responses. To prevent TLR9-mediated immune responses, strategies such as the removal of unmethylated CpG motifs (which are potential TLR9 agonists) or the incorporation of TLR9 inhibitory DNA sequences (such as inflammation-inhibiting oligonucleotide 2 (io2)) in the therapeutic gene expression cassette are being employed^4,5. Recently, we have uncovered a TLR9-independent innate sensing pathway that activates cellular immune responses to AAV-encoded transgene products in the liver⁶. This pathway instead utilizes cytokines IL-1a and IL-1b for the induction of IL-1R1–MyD88 signaling but does not rely on the inflammasome machinery (Fig. 1a). 

Figure 1. Innate immune sensors implicated in cellular response to AAV encoded transgene product following hepatic (a) and muscle (b) directed gene delivery. Strategies to prevent TLR9 activation such as depletion of unmethylated CpG motifs from AAV expression cassette are already in clinical use. TLR9 inhibitory sequences such as io2 can be incorporated in the expression cassette to further prevent TLR9 activation due to unmethylated CpG motifs in AAV ITRs. Anakinra, a recombinant version of naturally occurring IL-1 receptor antagonist (IL-1Ra) and Rilonacept, a fusion protein to neutralize IL-1a and IL-1b are already in clinics to treat other disease indications and can be repurposed for AAV gene therapy application. TLR3 antagonists are only available experimentally but not as approved medication. 

Skeletal muscle is one of the target tissues for AAV-mediated gene therapy to treat neuromuscular diseases⁷. Therefore, it is prudent to understand the mechanisms of AAV-driven immune responses following intramuscular gene delivery. Compared to hepatic gene transfer, the intramuscular route of vector administration is typically more prone to immune responses against the transgene product, including CD8⁺ T cell responses through TLR9>MyD88-mediated activation. In our study¹, we asked two questions. First, whether IL-1 signaling is important for CD8⁺ T cell activation against the transgene product expressed in the skeletal muscle, and second, whether additional innate sensing pathways are involved¹. To address these questions, we utilized a variety of strategies to dampen TLR9 activation (either by reducing the CpG content or the incorporation of io2 in the expression cassette) or to inhibit both TLR9 activation and IL-1R1 signaling. 

We utilized muscle-tropic AAV serotype 1 vectors that encoded chicken ovalbumin (OVA) under the control of the ubiquitous cytomegalovirus enhancer/human elongation factor-1α (CMV/EF1α) promoter or the muscle-specific promoter CK8. Additional AAV expression cassettes that either lacked CpG motifs or carried io2 alone or in combination with CpG depletion were also packaged in AAV1. In total, we evaluated five different AAV expression cassettes (Table 1). Two different doses of CMV/EF1a AAV1 vectors (2x10¹⁰ or 2x10¹¹ vector genomes/mouse) were injected into the quadriceps muscle of mice lacking individual innate sensors. In some experimental groups, IL-1 signaling was blocked by a combination of a-IL-1a and a-IL-1b antibodies. Flow cytometry was employed to monitor the kinetics of OVA-specific CD8⁺ T cells in peripheral blood, and an enzyme-linked immunosorbent assay (ELISA) was used to quantitate a-OVA IgGs in the plasma over 6 weeks. 

 Table 1: AAV constructs used in this study.

We first noticed a discrepancy in the need for TLR9 innate sensing between single-stranded (ss) and self-complimentary (sc) AAV vectors. The innate sensing of scAAV seemed to rely more strongly on TLR9 sensing. However, the innate sensing of ssAAV involved a number of different innate signaling pathways including TLR9, IL-1R1, and potentially TLR3 (Fig. 1b). The requirement for a single pathway was vector-dose-dependent. At a low dose, the inhibition of either of these signaling pathways significantly reduced the CD8⁺ T cell response against the transgene product. Interestingly, at a high vector dose, cellular responses were driven by a number of redundant pathways, including TLR9 and IL-1R1 signaling. Blocking one of these innate sensing pathways did not impact the response against the transgene product, whereas simultaneous blockade of both TLR9 and IL-1R1 signaling reduced the CD8⁺ T cell responses (Fig. 1b). Moreover, mice lacking downstream adaptor molecules such as MyD88 and TBK1 had minimal CD8⁺ T cell responses. We also found evidence for the involvement of RNA innate sensing via the TLR3-TRIF pathway. Taken together, these results clearly indicate that multiple innate pathways play a role in driving cellular responses during intramuscular gene transfer. Nevertheless, we did not find evidence of cytoplasmic DNA or RNA sensors being required, with the caveat that these results were entirely based on experiments performed in knockout mice (which may have skewed the results and therefore, they are not entirely conclusive). 

We examined whether genome engineering by CpG depletion or the incorporation of io2 in the AAV expression cassette would affect the immune response at a high dose. Interestingly, individually, these modifications had no impact. However, the presence of io2 in the AAV expression cassette in combination with the inhibition of IL-1R1 signaling was effective in reducing (but not preventing) the CD8⁺ T cell response against the transgene product. Surprisingly, the combined use of CpG depletion and IL-1R1 blockade failed to have a similar effect. Combining CpG depletion with io2 in a single AAV expression cassette reduced the CD8⁺ T cell response (which, however, was not further reduced when combined with IL-1R1 inhibition). These results illustrate the complexities and challenges of eliminating CD8⁺ T cell activation through innate immune blockade at high intramuscular vector doses. Even very low frequencies of CD8⁺ T cell responses in peripheral blood were associated with a loss of expression in the muscle. Nonetheless, we ultimately succeeded in achieving sustained transgene expression in skeletal muscle with minimal circulating or infiltrating CD8⁺ T cells when using a muscle-specific promoter along with CpG depletion and inclusion of the io2 sequence (raising the question of whether io2 inhibits additional pathways besides TLR9). 

In conclusion, we demonstrated the dose-dependent innate sensing of AAV-encoded transgene products following muscle gene transfer. Multiple innate signaling pathways (including TLR9, IL-1R1, and, potentially, TLR3) contribute to CD8⁺ T cell activation in AAV muscle gene transfer. Because of this redundancy, the blockade of multiple innate signaling pathways would be necessary to reduce CD8⁺ T cell responses against the transgene product. Further, we provided evidence that genome engineering (such as CpG motif removal or the incorporation of the TLR9 inhibitory sequence io2), could be beneficial in minimizing these cellular responses. Future work may reveal why the combination of TLR9 and IL-1R1 inhibition was effective in some but not other instances. It should be pointed out that blocking multiple innate signaling pathways, while helping to reduce CD8⁺ T cell responses, had no impact on antibody formation against the transgene product. Similarly, though the use of the muscle-specific promoter mitigated CD8⁺ T cell responses and provided long-lasting local transgene expression, this approach failed to prevent the antibody response. 

References:

1. Li N, Kumar SRP, Cao D, et al. Redundancy in Innate Immune Pathways That Promote CD8(+) T-Cell Responses in AAV1 Muscle Gene Transfer. Viruses 2024;16(10), doi:10.3390/v16101507
2. Colella P, Ronzitti G, Mingozzi F. Emerging Issues in AAV-Mediated In Vivo Gene Therapy. Mol Ther Methods Clin Dev 2018;8(87-104, doi:10.1016/j.omtm.2017.11.007
3. Cao D, Byrne BJ, de Jong YP, et al. Innate Immune Sensing of Adeno-Associated Virus Vectors. Hum Gene Ther 2024;35(13-14):451-463, doi:10.1089/hum.2024.040
4. Wright JF. Quantification of CpG Motifs in rAAV Genomes: Avoiding the Toll. Mol Ther 2020;28(8):1756-1758, doi:10.1016/j.ymthe.2020.07.006
5. Chan YK, Wang SK, Chu CJ, et al. Engineering adeno-associated viral vectors to evade innate immune and inflammatory responses. Sci Transl Med 2021;13(580), doi:10.1126/scitranslmed.abd3438
6. Kumar SRP, Biswas M, Cao D, et al. TLR9-independent CD8(+) T cell responses in hepatic AAV gene transfer through IL-1R1-MyD88 signaling. Mol Ther 2024;32(2):325-339, doi:10.1016/j.ymthe.2023.11.029
7. Wang JH, Gessler DJ, Zhan W, et al. Adeno-associated virus as a delivery vector for gene therapy of human diseases. Signal Transduct Target Ther 2024;9(1):78, doi:10.1038/s41392-024-01780-w.