AAVMYO, MyoAAV2A, and MyoAAV4A show higher whole-body transduction efficiency compared to AAV9

To assess the transduction efficiency and organ distribution of myotropic AAVs (AAVMYO, MyoAAV2A, MyoAAV4A) in comparison to AAV9, a cytomegalovirus (CMV)-driven transgene containing genes coding for luciferase and eGFP was inserted in these AAV serotypes for in vivo and ex vivo analyses (Fig. 1A). Bioluminescence was first performed in vivo to measure luminescence intensity based on luciferase activity in 6-week-old CD1 mice injected via tail vein with a low dose of 4.5E + 12 vg/kg (equivalent of 1E + 11 vg per mouse) and measured 5 weeks post injection. The doses used were calculated from qPCR titration, and the results obtained by ddPCR are shown in Figure S1.

Most injected animals showed bioluminescence in the upper and lower limb regions in the prone position and hindlimb region in the back position, suggesting several muscle groups were transduced (Fig. 1B). In addition, luminescence was visible in the plexus region in MyoAAVs-injected mice, potentially the heart. Also, luminescence was visible in the upper abdominal region in only AAV9-injected mice after 1 week, suggesting a potential expression in liver, while the signal disappeared after 5 weeks (Figure S2). The intensity of luminescence, measured by determining the average whole-body radiance, was higher in AAVMYO, MyoAAV2A, and MyoAAV4A compared to AAV9 (Fig. 1C). Luminescence with MyoAAV2A- and MyoAAV4A was respectively 5 and 18 times higher than with AAVMYO. No differences of intensity of bioluminescence were observed between male and female mice, as well as between retro-orbital and tail-vein injections (Figure S3-S4).

Later, to compare muscle transduction efficiency between all the AAVs, we next injected a higher dose (4.5E + 13 vg/kg) of AAV9 into a separate cohort of mice. This high dose of AAV9 showed a similar pattern of bioluminescence localization as in AAVMYO, MyoAAV2A, and MyoAAV4A-injected mice at the 10 times lower dose (Fig. 1B). Although no significant difference in average whole-body radiance was observed between AAV9 HD, MyoAAV2A, and MyoAAV4A, the results showed 7-times lower luminescence with AAVMYO (1.7E + 7 ± 4.2E + 6 p/sec/cm²/sr) compared to AAV9 HD (1.2E + 8 ± 9.6E + 6 p/sec/cm²/sr).

In general, a stronger luminescence related to transgene expression was observed in AAVMYO, MyoAAV2A, and MyoAAV4A compared to AAV9 at the same dose, mainly located in several muscle regions.

Fig. 1

Study design and comparison of in vivo luminescence levels between AAV9 and myotropic AAVs in adult mice 5w after systemic injection. (A) Study design. Genes coding for luciferase and eGFP are contained in the cytomegalovirus (CMV)-driven transgene separated by an IRES (Internal Ribosome Entry Site) sequence that allows translation of both proteins. Wild-type (WT) CD1 and C57BL/6 mice have been injected at 6w (weeks) of age and analysis performed 5w after injection, or at post-natal day 1 and analyzed 4 weeks later. (B) Expression of whole-body in vivo luminescence in WT CD1 mice systemically injected with 4.5E + 12 vg/kg (viral genome / kilogram) (low dose) of AAV9-, AAVMYO-, MyoAAV2A-, MyoAAV4A-CMV-luc-IRES-eGFP or 4.5E + 13 vg/kg (high dose, HD) of AAV9-CMV-luc-IRES-eGFP. Color scale: 5E + 6–3E + 7 photons per second per centimeter square per steradian (p/sec/cm²/sr). (C) Quantification of in vivo luminescence in CD1 mice injected with AAV9-, AAVMYO-, MyoAAV2A-f, MyoAAV4A- at 4.5E + 12 vg/kg or AAV9-CMV-luc-IRES-eGFP at 4.5E + 13 vg/kg (HD) taken at 11 weeks. Luminescence is quantified by measuring the average radiance (p/sec/cm²/sr). Data are presented as mean values +/- SEM (n = 4–13). One-way ANOVA with Tukey correction. ***p < 0.001 and ****p < 0.0001 versus AAV9; ##p < 0.01, ###p < 0.001, ####p < 0.0001 versus AAVMYO

MyoAAV2A and MyoAAV4A present higher transduction in leg muscle and heart while AAVMYO is most efficient in diaphragm

A pilot study was performed with different AAV serotypes in CD1 mice and VCN/dg was measured. MyoAAV2A and MyoAAV4A showed a higher vector copy number (VCN) in TA and heart compared to AAVMYO in both strains, while no increase was observed for AAVMYO compared to AAV9 in these tissues (Fig. 2A). In diaphragm, the three myotropic AAVs exhibited a strong increase in VCN compared to AAV9.

Next, we used other mouse cohorts to compare two different routes of systemic injection and also two mouse strains with different genetic backgrounds. Several tissues (tibialis anterior-TA, diaphragm, heart, and liver) were collected 5 weeks after systemic AAV injection with a low dose (4.5E + 12 vg/kg) in 6-week-old CD1 (tail vein) and C57BL/6 (retro-orbital) mice and eGFP mRNA level and luciferase activity were measured.

The level of eGFP mRNA expression in TA muscle was significantly increased for all myotropic AAVs compared to AAV9 in CD1 mice while it was the case only for MyoAAV2A and MyoAAV4A in C57BL/6 mice with 32 and 28 times respectively (Fig. 2B). In diaphragm, AAVMYO and MyoAAV2A correlated with increased mRNA expression in CD1 mice, while MyoAAV4A showed a decrease mRNA level in C57BL/6 mice compared to AAV9. In heart, only MyoAAV2A showed a mRNA increase in C57BL/6 mice (6.5 times) (Fig. 2B). In C57BL/6 mice injected at high dose, AAVMYO demonstrated a significant increase in transgene mRNA level in TA but not in heart compared to AAV9 (Figure S5).

In correlation with the eGFP mRNA expression, quantification of luciferase activity representing the transgene expression showed a higher transduction efficiency for MyoAAV2A- and MyoAAV4A- compared to AAV9- and AAVMYO-injected CD1 and C57BL/6 mice in TA muscle (Fig. 2C). MyoAAV2A presented the highest luciferase activity in TA in CD1 (3.2E + 4 ± 1.1E + 4 RLU/µg total proteins) mice, whereas MyoAAV4A appeared more efficient in C57BL/6 mice (1.3E + 5 ± 1.2E + 4 RLU/µg total proteins). For diaphragm, AAVMYO was the most efficient to transduce CD1 mice. In heart, MyoAAV2A achieved the highest luciferase activity in both mouse strains. Variations were observed between strains, with a global loss of transduction efficiency in all tissues in CD1 compared to C57BL/6 mice. Overall, the highest expression was noted for AAVMYO in diaphragm (79-fold compared to AAV9).

In conclusion, MyoAAV2A and MyoAAV4A showed the most efficient transduction in TA muscle and heart, whereas AAVMYO was the most efficient in diaphragm among all tested AAVs.

Fig. 2

Comparison of transduction efficiency between AAV9 and myotropic AAVs in several organs from wild-type adult mice. Mice were systemically injected at 6-week-old and organs collected 5 weeks later. (A) Quantification of vector copy number per diploid genome of AAVMYO-, MyoAAV2A-, MyoAAV4A- and AAV9-CMV-luc-IRES-eGFP in WT CD1 and C57BL/6 mice in tibialis anterior leg muscle, diaphragm, heart and liver. Data are presented as mean values +/- SEM (n = 5–6). One-way ANOVA with Tukey correction. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 versus AAV9 ; #p < 0.05, ##p < 0.01, ###p < 0.001, ####p < 0.0001 versus AAVMYO ; $p < 0.05, $$$$<0.0001 versus MyoAAV2A ; §p < 0.05, §§p < 0.01, §§§§p < 0.05 versus MyoAAV4A. (B) Quantification of eGFP mRNA fold change expression of AAVMYO-, MyoAAV2A-, MyoAAV4A- and AAV9-CMV-luc-IRES-eGFP in WT CD1 and C57BL/6 mice in different organs. Data are presented as mean values +/- SEM (n = 5 for C57BL/6 mice, n = 9–12 for CD1 mice). One-way ANOVA with Tukey correction. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 versus AAV9; #p < 0.05, ##p < 0.01, ####p < 0.0001 versus AAVMYO; §p < 0.05, §§p < 0.01 §§§§p < 0.0001 versus MyoAAV4A. (C) Quantification of luciferase activity of AAVMYO-, MyoAAV2A-, MyoAAV4A- and AAV9-CMV-luc-IRES-eGFP in WT CD1 and C57BL/6 mice in different organs. Data are presented as mean values +/- SEM (n = 5 for C57BL/6 mice, n = 9–13 for CD1 mice). One-way ANOVA with Tukey correction. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 versus AAV9; ##p < 0.01, ####p < 0.0001 versus AAVMYO ; $p < 0.05, $$p < 0.01 versus MyoAAV2A ; §p < 0.05 versus MyoAAV4A

AAVMYO, MyoAAV2A and MyoAAV4A transduce more efficiently type IIb myofibers compared to AAV9

To investigate the basis for the better muscle transduction with myotropic AAVs, we investigated if myotropic AAVs show different transduction patterns depending on myofiber types. In mouse muscles, there are barely slow (oxidative) type I fibers, while type IIa are more oxidative than type IIb [38]. Indeed, it has been demonstrated that AAV9 shows a preference for type IIa and IIX myofibers, while it poorly transduces type IIb myofibers [39]. Immunofluorescence was performed on cross-sections of TA muscles systemically transduced with the different AAVs at 4.5E + 13 vg/kg to identify eGFP signal and type IIa and IIb myofibers.

As expected, AAV9-transduced TAs showed a higher rate of transduction in type IIa myofibers than IIb myofibers (Fig. 3A-B). All the myotropic AAVs transduced type IIa myofibers at high level (more than 80% of eGFP + fibers). Furthermore, AAVMYO, MyoAAV2A and MyoAAV4A demonstrated enhanced transduction efficiency in type IIb myofibers compared to AAV9, with the highest transduction rate achieved by MyoAAV4A under these conditions.

Thus, the better muscle transduction with MyoAAV4A in C57BL/6 mice is potentially related to their better efficiency to transduce different myofiber types.

Fig. 3

Correlation between of eGFP and myofiber IIa and IIb in AAV9, AAVMYO, MyoAAV2A, MyoAAV4A-transduced muscles. Tibialis anterior muscle sections collected 5 weeks after injection in 6-week-old C57BL/6 mice with high dose (4.5E + 13 vg/kg) of AAVMYO-, MyoAAV2A-, MyoAAV4A- and AAV9-CMV-luc-IRES-eGFP. (A) Fluorescent co-labeling of eGFP, myofiber IIa and IIb in TA of AAVMYO-, MyoAAV2A-, MyoAAV4A- and AAV9-injected mice. Red arrowhead and yellow arrowhead indicate respectively GFP negative IIb fibers and GFP positive IIa fibers. Scale bar is 200 μm. (B) Quantification of positive GFP (GFP+) fibers depending of the fiber type. Percentage of GFP+ fibers was determined by the number of fibers presenting a colocalization between GFP and IIa/IIb signal divided by the total number of myofiber IIa/IIb. Data are presented as mean values +/- SEM (n = 3). Two-way ANOVA with Bonferroni correction

AAVMYO exhibits the highest transduction efficiency in 1-day post-natal injected mice

Depending on the age of onset and disease severity, patients with muscle diseases might be injected with AAV at different ages, including early stages for disease prevention. To assess if the myotropic AAVs diffuse as effectively when injected into neonates or into adult mice, both CD1 and C57BL/6 mouse strains were injected retro-orbitally at 1 day post-natal (d.p.n) with 4.1E + 12 vg/kg, and the transduction efficiency was evaluated 4 weeks later by measuring eGFP mRNA level and ex vivo luficerase activity from the TA, diaphragm muscles, heart, and liver.

In CD1 mice, AAVMYO presented the highest eGFP mRNA level in TA among all AAVs, also confirmed by luciferase activity in this tissue (Fig. 4A-B). Additionally, in both mouse strains, AAVMYO presented the highest eGFP mRNA expression and luciferase activity in diaphragm (1E + 2 ± 1.7E + 1 RLU/µg protein for CD1; 5.8E + 2 ± 7.2E + 1 RLU/µg protein for C57BL/6). In C57BL/6 mice, MyoAAV2A and AAVMYO showed an increased luciferase activity in TA and diaphragm compared to AAV9 (Fig. 4B). Noteworthy, all myotropic AAVs showed a decrease of eGFP mRNA level in heart compared to AAV9. Surprisingly, MyoAAV4A presented low transduction in TA, diaphragm and heart tissues compared to all the other AAVs after injection at 1 d.p.n, unlike following adult injection (Fig. 4A-B).

In conclusion, myotropic AAVs show variations in transduction efficiency depending on the age at injection, with the highest transduction rate in 1-d.p.n-injected mice with AAVMYO in both mouse strains.

Fig. 4

Comparison of transduction levels in different organs between AAV9 and myotropic AAVs in mice 4 weeks after retro-orbital injection at 1 day post-natal. WT CD1 or C57BL/6 mice were systemically injected at 1 d.p.n. with 4.1E + 12 vg/kg of AAV9-, AAVMYO-, MyoAAV2A-, or MyoAAV4A-CMV-luc-IRES-eGFP and analyzed 4 weeks later. (A) Quantification of eGFP mRNA fold change expression of AAVMYO-, MyoAAV2A-, MyoAAV4A- and AAV9-CMV-luc-IRES-eGFP in WT CD1 and C57BL/6 mice in different organs. Data are presented as mean values +/- SEM (n = 5–6). One-way ANOVA with Tukey correction. *p < 0.05, **p < 0.01, ****p < 0.0001 versus AAV9 ; ##p < 0.01, ####p < 0.0001 versus AAVMYO ; §p < 0.05, §§p < 0.01, §§§p < 0.001, §§§§p < 0.0001 versus MyoAAV4A. (B) Quantification of luciferase activity of AAVMYO-, MyoAAV2A-, MyoAAV4A- and AAV9-CMV-luc-IRES-eGFP in WT CD1 and C57BL/6 mice in different organs. Data are presented as mean values +/- SEM (n = 5–6). One-way ANOVA with Tukey correction. *p < 0.05, **p < 0.01, ****p < 0.0001 versus AAV9 ; $$p < 0.01, $$$p < 0.001 versus MyoAAV2A ; §§p < 0.01, §§§p < 0.001, §§§§p < 0.0001 versus MyoAAV4A

Differential liver de-targeting for the compared AAV serotypes

The different myotropic AAVs showed increased transduction efficacy in muscles and heart. As several clinical trials with AAV8, AAV9 and AAVrh74 were halted due to liver toxicity, sometimes leading to patient death, we compared the different myotropic AAVs for liver transduction at both ages and in both strains.

In CD1 mice injected at 1 day postnatal, AAVMYO and MyoAAV2A exhibited liver de-targeting according to luciferase assay results while such de-targeting was not observed in C57BL/6 mice (Fig. 4B). However, based on eGFP mRNA levels in this mouse strain, AAVMYO and MyoAAV2A did demonstrate a de-targeting effect in the liver in both genetic backgrounds (Fig. 4A). In adult-injected mice, all serotypes tested including AAV9 displayed a lower liver transduction compared to muscles and heart (Fig. 2A-C). AAVMYO significantly de-targeted the liver in both mouse strains compared to AAV9 while this effect was observed for MyoAAV2A in C57BL/6 mice and for MyoAAV4A in CD1 mice (Fig. 2C).

Overall, the liver de-targeting effect of different myotropic serotypes was not consistently present in different mouse strains and varied depending on the age of injection, albeit AAVMYO and MyoAAV4A appeared the most promising.