Identification of a myotropic AAV by massively parallel in vivo evaluation of barcoded capsid variants

Parts used

Show 17 moreparts→

Key findings

• AAVMYO, originally designated AAV9P1, is an AAV9 capsid that displays the peptide P1 (RGDLGLS) in an exposed capsid loop. Among capsids injected intravenously into adult C57BL/6J mice as pools of barcoded vectors, AAVMYO ranks above all other 156 variants of the second and third libraries for transcriptional efficiency across the entire musculature, comprising skeletal muscle, diaphragm and heart.
• On the AAV mRNA level in a barcoded library screen in adult C57BL/6J mice, AAV9 is approximately 10.6-, 7.2- and 1.5-fold less efficient than AAVMYO in diaphragm, quadriceps femoris and heart, respectively. More than 70% of the AAVMYO-associated barcodes recovered in that screen are detected in muscle tissues, reflecting the myospecificity of this capsid.
• After individual intravenous injection of 1 x 10^11 vg per C57BL/6J mouse, an AAVMYO vector carrying a CMV promoter-driven eyfp reporter yields 61-, 17- and 11-fold higher viral eyfp transcript levels than the same genome in AAV9 in diaphragm, quadriceps femoris and heart, respectively, measured one week later by RT-qPCR against a POLR2A housekeeper. AAVMYO is concurrently nine-fold detargeted from the liver relative to AAV9.
• With a Firefly luciferase gene under the muscle-specific SPc5-12 promoter delivered intravenously at 4 x 10^10 vg per CB17-SCID mouse, AAVMYO gives an 11.6-fold increase in luciferase mRNA over AAV9 in the diaphragm, a 37.8- up to 50.1-fold increase across various skeletal muscle types and a 5.8-fold increase in the heart, together with 23.2-fold detargeting from the liver, four weeks after injection.
• None of the systemically delivered muscle-targeting benchmarks AAVM41, AAV-B1, AAV9LD and AAV587MTP, nor the AAV9 peptide-display mutants AAV9P3 (RGDAVGV) and AAV9-RGDLRVS that carry sequences similar to P1 (RGDLGLS), match the in vivo efficiency and specificity of AAVMYO in the musculature of intravenously injected C57BL/6J mice.
• Recapitulating the AAV9 double-point mutation P504A/G505A in the AAVMYO capsid yields AAVMYO_LD, which shows enhanced liver detargeting relative to AAVMYO but is concurrently less active in the musculature. The muscle biodistribution of AAVMYO is therefore not merely a consequence of its pronounced liver detargeting.
• AAVMYO delivered intravenously at 5 x 10^11 vg per C57BL/6 mouse transduces slow-twitch type I as well as fast-twitch type IIa and type IIb fibers of the quadriceps femoris. In mdx mice injected intravenously with 2 x 10^11 or 1 x 10^12 vg, AAVMYO drives higher micro-dystrophin expression in quadriceps femoris than AAV9 four weeks after delivery, by immunofluorescence and by western blotting.
• AAV5 ranks among the most efficient vectors in the mouse liver on the vector DNA level yet is outperformed there by numerous other capsids on the mRNA level, falling to position 41 of 91, 45 of 82 and 124 of 157 across three barcoded libraries. Most AAV5-delivered genomes in the mouse liver therefore fail to express mRNA, a phenotype that does not apply to the human liver.
• Among 12 wild-type AAV serotypes compared in the same intravenously injected mice, AAV1 to AAV3, AAV5 to AAV8, AAVrh.10 and AAV12 predominantly target the liver, AAV4 is strongly detargeted from the liver and instead transduces the lung, and AAVpo.1 is largely inactive in the liver but transduces the musculature, especially diaphragm and quadriceps femoris. AAV9 shows the broadest distribution and the highest efficiency in most organs, although roughly 50% of this capsid still ends up in the liver.
• AAV4 peptide-display mutants carrying insertions in variable region VIII retain the high lung efficiency and low liver efficiency of wild-type AAV4 after intravenous injection in mice, indicating that the capsid sequences or domains responsible for AAV4 lung tropism lie outside variable region VIII.
• Inserting a 15-nucleotide barcode into the 3-prime untranslated region of a CMV promoter-driven EYFP reporter in a self-complementary AAV genome, one barcode per capsid variant, permits each capsid to be tracked simultaneously on the vector DNA and on the mRNA level by next-generation sequencing of the same tissue samples. Restricting the pool to barcodes of exactly 15 nucleotides, with no homopolymer longer than 3 nucleotides and a Hamming distance of at least five from every other barcode, yields a usable set of 159 barcodes.
• AAVMYO can be produced to high titers akin to its parent serotype AAV9 and is compatible with all standard AAV purification protocols including affinity columns, so its manufacturing imposes no additional constraint relative to AAV9.

Abstract

Adeno-associated virus (AAV) forms the basis for several commercial gene therapy products and for countless gene transfer vectors derived from natural or synthetic viral isolates that are under intense preclinical evaluation. Here, we report a versatile pipeline that enables the direct side-by-side comparison of pre-selected AAV capsids in high-throughput and in the same animal, by combining DNA/RNA barcoding with multiplexed next-generation sequencing. For validation, we create three independent libraries comprising 183 different AAV variants including widely used benchmarks and screened them in all major tissues in adult mice. Thereby, we discover a peptide-displaying AAV9 mutant called AAVMYO that exhibits superior efficiency and specificity in the musculature including skeletal muscle, heart and diaphragm following peripheral delivery, and that holds great potential for muscle gene therapy. Our comprehensive methodology is compatible with any capsids, targets and species, and will thus facilitate and accelerate the stratification of optimal AAV vectors for human gene therapy.