π-Icosa Derived AAV Capsid PG007 Enhanced Micro-Dystrophin Delivery in DMD Model

Our prior publication detailed the successful application of the π-Icosa platform in the identification of PG007, a novel adeno-associated virus (AAV) capsid variant. Notably, PG007 exhibited significantly reduced liver tropism while demonstrating enhanced muscle tropism in C57BL/6J mice, directly tackling a major limitation in systemic AAV delivery. This article represents a continuation of our efforts to fully characterize the potential of PG007, a prime example of the π-Icosa platform’s output, as a highly specific muscle-directed gene therapy vector. To this end, we present comprehensive in vivo data evaluating PG007’s transduction efficiency and selectivity across a broader range of muscle tissues and key non-target organs, including the brain, kidney, and lung. Moreover, we extend our evaluation to a nonhuman primate (NHP) model to assess the translational relevance of this π-Icosa-derived capsid, analyzing its biodistribution and transgene expression patterns. We also explore the therapeutic potential of PG007, a testament to the platform’s ability to generate functional vectors, by investigating its capacity to deliver functional micro-dystrophin in Duchenne muscular dystrophy (DMD) mouse models. The findings presented herein aim to definitively validate the muscle-specific targeting and liver de-targeting characteristics conferred upon PG007 by its π-Icosa-mediated design, further underscoring the platform’s transformative potential in the field of gene therapy.

PG007 demonstrates high efficiency and specificity for muscle targeting in Balb/c mice

To assess muscle transduction efficiency following systemic PG007 administration in different mouse strains, we intravenously injected adult Balb/c mice with 2E+11 vg (~8E+12 vg/kg) of AAV9-, AAV2-, PG007-, and MyoAAV 4A-CAG-luciferase-P2A-EGFP. Three weeks post-injection, in vivo luciferase activity and transgene expression were analyzed across tissues. Whole-organ bioluminescence imaging revealed that PG007 exhibited significantly higher luciferase expression in limb muscles than AAV2, AAV9, and even MyoAAV 4A (Fig. 5A, 5B). Consistently, luciferase mRNA levels in the biceps, abdomen, and heart were also significantly elevated in PG007-treated mice compared to MyoAAV 4A (Fig. 5D). Western blot analysis further confirmed higher luciferase protein levels in triceps, quadriceps, and heart tissues with PG007 compared to AAV2, AAV9, and MyoAAV 4A (Fig. 5C). These findings underscore PG007’s enhanced muscle-targeting efficiency, supporting the hypothesis that tissue-specific peptide incorporation can precisely direct PG007 to desired tissues. Additionally, hepatic de-targeting was evident, as luciferase mRNA and protein levels in the liver were markedly reduced in PG007-treated mice compared to other vectors (Fig. 5C, 5E). In non-muscular tissues, including the brain, lung, and kidney, all groups exhibited low expression with no significant differences from vehicle-injected controls (Fig. 5E).

Fig 5. PG007 effectively and selectively targets muscle in Balb/c mice. A. Whole-body bioluminescence images of 8-week-old Balb/c mice injected with 2×10^11 vg/mouse of PG007-CAG-luciferase-P2A-EGFP. B. Firefly luciferase luminescence in hindlimbs. C. Western blots for EGFP and GAPDH in triceps, heart, and liver. D. Fluc mRNA levels in biceps, triceps, quadriceps, gastrocnemius, abdominal muscles, and heart (qPCR). E. Fluc mRNA in de-targeted tissues (liver, brain, lung, kidney).

PG007 exhibits efficient muscle transduction with minimal liver toxicity in primates

To evaluate the translational potential of PG007 in NHPs, given that mouse-derived tissue-specific AAV serotypes often fail to translate universally to NHPs (1,2), we administered a single intravenous infusion of PG007- or MyoAAV 4A-CAG-luciferase-P2A-EGFP to adult Macaca fascicularis at a dose of 3×10^13 vg/kg. Hepatotoxicity was assessed by monitoring serum levels of alanine aminotransferase (ALT) and aspartate aminotransferase (AST), established markers of liver injury. Serum analysis revealed that MyoAAV 4A induced a peak in AST and ALT levels on day 3 post-injection, with increases of 17.9-fold and 13.2-fold, respectively, compared to pre-injection levels (Fig. 6A,6B). In contrast, PG007-treated animals showed no significant elevation in AST or ALT levels throughout the 14-day monitoring period, maintaining levels comparable to baseline (p>0.05), indicating a superior hepatic safety profile following systemic administration. No muscle damage was observed in either group, as assessed by creatine kinase (CK) levels within 14 days (Fig. 6C). To examine transduction efficiency, we quantified vector genome copies. In the liver, PG007 exhibited significantly lower vector genome accumulation and mRNA expression compared to MyoAAV 4A (Fig. 6D), with Western blotting confirming the significantly reduced level of EGFP protein expression (Fig. 6F). Luciferase mRNA levels varied in muscle tissues:  PG007 displayed higher expression in biceps, while MyoAAV 4A showed increased expression in triceps, with no significant differences in quadriceps or gastrocnemius (Fig. 6E). These mRNA patterns were corroborated by EGFP protein levels via Western blotting (Fig. 6F) and fluorescence imaging of muscle tissues (Fig. 6G). Collectively, these results demonstrate that PG007 achieves superior liver de-targeting compared to MyoAAV 4A while maintaining comparable muscle-targeting efficiency in NHPs, highlighting its potential for safer and effective muscle-directed gene therapy.

Figure 6. PG007 Lowers Liver Toxicity in NHPs While Retaining Transduction Efficiency. A–B. ALT and AST levels in Macaca fascicularis serum before and after 3×10^13 vg/kg IV injection of PG007- and MyoAAV 4A -CAG-luciferase-P2A-EGFP. C. Creatine kinase (CK) levels in NHPs serum. D. Vector genome copies per diploid genome and Fluc mRNA fold change in liver. E. Fluc mRNA fold change in biceps, triceps, quadriceps, and gastrocnemius. F. Western blots for EGFP and GAPDH in biceps, triceps, quadriceps, gastrocnemius, and liver. G. Fluorescence images of biceps, triceps, quadriceps, and gastrocnemius cross-sections (EGFP).

PG007 systemically delivered functional micro-dystrophin in DMD mouse models

To evaluate PG007 as a vector for in vivo delivery of therapeutic transgenes, we compared its efficacy to AAV9, currently under investigation in clinical trials for Duchenne muscular dystrophy (DMD), using a dystrophin-deficient B10 DMD-KO mouse model (4 bp deletion in Dmd exon 4). We systemically administered a microdystrophin (μDys) transgene under the muscle-specific MHCK7 promoter via AAV9 or PG007 vectors (2×10^11 vg/mouse). At 14 weeks post-injection, a whole-limb grip strength test revealed that PG007-μDys-treated mice exhibited significantly greater muscle strength recovery compared to AAV9-μDys-treated and vehicle-injected controls (Fig. 7A). By 20 weeks, serum CK levels in PG007-μDys-treated mice were significantly lower compared to AAV9-μDys-treated mice (Fig. 7B), indicating substantial protection against muscle damage. Vector genome copy number analysis showed higher μDys distribution in muscles (quadriceps, biceps, triceps, gastrocnemius) of the PG007-μDys group but lower in non-muscle tissues (liver, kidney, brain, lung) (Fig. 7C). Similarly, qPCR confirmed higher μDys mRNA levels in targeted muscles and lower levels in non-targeted tissues in the PG007-μDys group (Fig. 7D). Immunofluorescence staining further demonstrated restored dystrophin expression in biceps, triceps, quadriceps, and gastrocnemius of PG007-μDys-treated mice, surpassing AAV9-μDys levels (Fig. 7E). Overall, PG007-μDys exhibited superior therapeutic efficacy in DMD-KO mice, enhancing muscle performance, reducing damage, and increasing dystrophin expression at both transcriptional and translational levels, highlighting its potential for DMD gene therapy.

Figure 7. Systemic PG007-MHCK7-microdystrophin improves muscle function and expresses micro-dystrophin in adult DMD mice. (A) Grip strength in wild-type, vehicle-treated DMD-KO, and DMD-KO mice treated with AAV9 or PG007-MHCK7-μDys. (B) Serum CK levels at 14 weeks post-injection. (C) Micro-dystrophin mRNA quantification across tissues. (D) Relative viral genome DNA levels in various tissues. (E) Fluorescent images of biceps, triceps, quadriceps, and gastrocnemius muscles.

Conclusion

This study demonstrates the potential of the novel AAV capsid variants, PG007, as highly effective vectors for muscle-directed gene delivery. Through systematic engineering and screening using the π-Icosa platform, PG007 was identified as a capsid with reduced liver tropism and enhanced muscle specificity, a profile further validated in both rodent and NHP models. In Balb/c mice, PG007 exhibited robust muscle transduction efficiency, with significantly higher luciferase expression in limbs compared to AAV9 and MyoAAV 4A, while maintaining minimal off-target expression in the liver, brain, lung, and kidney. In Macaca fascicularis, PG007 demonstrated a superior safety profile, with no significant elevation in hepatic enzyme levels (ALT and AST) post-administration, unlike MyoAAV 4A, which induced marked hepatotoxicity. Additionally, PG007 achieved comparable muscle transduction to MyoAAV 4A in NHPs, with selective mRNA and protein expression in biceps, triceps, quadriceps, and gastrocnemius, alongside reduced liver transduction. When delivering a therapeutic microdystrophin (μDys) transgene in a DMD-KO mouse model, PG007-μDys outperformed AAV9, restoring muscle strength, reducing serum creatine kinase levels, and enhancing microdystrophin expression across multiple muscle groups, but with lower levels in non-targeted tissues. These findings underscore PG007’s ability to balance efficacy and safety, minimizing hepatic toxicity while maximizing muscle-specific gene delivery. The success of PG007 highlights the power of the π-Icosa platform, a cutting-edge tool for high-throughput AAV capsid screening and optimization, enabling the rapid identification of tissue-specific vectors with improved safety and efficacy profiles for gene therapy applications. We invite researchers and industry partners to explore the π-Icosa platform for their own capsid engineering needs, as its precision and scalability can accelerate the development of next-generation gene therapies.

References

1. Hordeaux J, W.Q., Katz N, Buza EL, Bell P, Wilson JM. , The Neurotropic Properties of AAV-PHP.B Are Limited to C57BL/6J Mice. . Mol Ther, 2018. 26(3): p. 664-668.
2. Bunuales M, G.A., Chillon M, Bosch A, Gonzalez-Aparicio M, Espelosin M, Garcia-Gomara M, Rico AJ, Garcia-Osta A, Cuadrado-Tejedor M, Lanciego JL, Hernandez-Alcoceba R., Characterization of brain transduction capability of a BBB-penetrant AAV vector in mice, rats and macaques reveals differences in expression profiles. Gene Ther, 2024. 31(9-10): p. 455-466