AAV1.NT-3 gene therapy prevents age-related sarcopenia

rAAV.NT-3 vector production and potency

scAAV1.tMCK.NT-3 design (Supplementary Figure 1A), and production followed previously described methods at Nationwide Children`s Hospital, Columbus [10]. scAAV1.tMCK.NT-3, at 1 × 10¹¹ vg dose, was delivered to the gastrocnemius muscle of 18 months old C57BL/6 mice. Blood samples from terminally anesthetized treated and untreated mice were collected by cardiac puncture at six months post gene injection, and serum was assayed for NT-3 levels using a capture ELISA, as previously reported [10]. No detectable NT-3 levels were found in the untreated C57BL/6 mice at 2 years of age, in contrast to the treated cohort (Supplementary Figure 1B), showing significant improvements in functional and histologic outcome measures as described below. Similar to our previous observations in younger age groups, we found no effect of sex on serum NT-3 levels in the aged C57BL/6 mice.

NT-3 gene therapy improved function, attenuated age-related musculoskeletal and skin changes in the aged C57BL/6 mouse

Run to exhaustion treadmill test was used to assess the efficacy of NT-3 gene therapy on motor function at 2-, 4- and 6-months post gene delivery. AAV1.NT-3 improved treadmill performance, leading to remarkable increases in running distances and time until exhaustion compared to the untreated counterparts at all time points tested. We found a 63% (NT-3, 135.8 ± 17.5, n = 12 vs. UT, 83.5 ± 11.1, n = 11; p = 0.0184), 73% (NT-3, 180.6 ± 16.2, n = 12 vs. UT, 104.4 ± 4.8, n = 11; p = 0.0003) and 63.5% (NT-3, 166.1 ± 13.8 n = 11, vs. UT, 101.6 ± 7.2, n = 8; p = 0.0077) increase in distance run at 2-, 4- and 6-months post gene delivery, respectively (Figure 1A). Treated females performed significantly better at 4 months post-injection (p = 0.0394) compared to treated males (Figure 1B), while there was no sex difference found in the untreated cohort (Figure 1C). We also assessed the efficacy of NT-3 gene therapy on motor coordination of aged C57BL/6 mice with rotarod test. NT-3 treated cohort displayed significantly better motor coordination with 39.2% longer duration compared to untreated at endpoint (NT-3, 56.9 ± 5.9, n = 10 vs. UT, 40.9 ± 6.9, n = 8; p = 0.038, Figure 1D). Treated females performed better than males without reaching statistical significance (Supplementary Figure 2A).

In vivo muscle contractility assay performed in the gastrocnemius muscle of C57BL/6 mice at endpoint showed an increase in the maximum twitch response with NT-3 gene therapy (NT-3, 2.22 ± 0.2, n = 11 vs. UT, 1.57 ± 0.1, n = 8; p = 0.015, Figure 1E) while the maximum tetanic response did not improve significantly (NT-3, 11.1 ± 0.4, n = 10 vs. UT, 10.1 ± 0.6, n = 9; p = 0.142, Figure 1F). When treatment effect on genders was analyzed, both the maximum twitch and maximum tetanic responses were found to be higher in males, without reaching statistical significance (Supplementary Figure 2B, 2C). The treatment did not alter the endpoint weight of the treated cohort (NT-3, n = 11, 37.0 ± 1.95 g; UT, n = 10, 37.6 ± 1.94 g; p = 0.84, Supplementary Figure 2D). Two untreated mice, aged 84 and 100 weeks, and one treated mouse, aged 99 weeks, died before reaching to the end point of the experiment.

We also documented improvements in age related musculoskeletal and skin changes in the NT-3 treated C57BL/6 mice including reduction of kyphosis, dermatitis, and alopecia (Supplementary Figure 3A). AAV1.NT-3 injected, and untreated cohorts were photographed at the end point, and the severity of kyphosis, dermatitis, and alopecia were assessed using a semiquantitative scoring system (0: none, 0.5: mild-moderate, 1: severe). Varying degrees of kyphosis were present in 83% of untreated mice, half of which were severe, while only 50% of the AAV1.NT-3 injected mice had kyphosis and one third were severe. Dermatitis, observed in 33% of untreated mice, was not seen in the treated group. Moderate to severe alopecia around the ears was present in 67% of the untreated mice, whereas this ratio decreased to 33% percent in treated mice (Supplementary Figure 3B–3D).

Reversal of age-related neuromuscular histopathology with NT-3 gene therapy

We examined the efficacy of NT-3 gene therapy upon muscle fiber size and fiber type composition in tibialis anterior, gastrocnemius, quadriceps, and triceps muscles at 6 months post gene injection. Succinic dehydrogenase stain was used for quantification [15], which delineates fibers according to mitochondria content: fatigue-resistant slow-twitch oxidative (STO) fibers with darkest staining, the fast-twitch oxidative (FTO) with intermediate staining and the fast-twitch glycolytic (FTG) with lightest staining intensity (Figure 2A). Fiber size measurements were done on the representative images from deep, intermediate, and superficial zones of untreated and NT-3-treated samples as illustrated in the tibialis anterior and gastrocnemius muscles [15] (Figure 2B, 2C). Both muscles from the untreated cohort showed atrophic angular fibers with the majority belonging to fast twitch type 2 fibers, typical of sarcopenic muscle (Figure 2A). Compared to younger 10-month-old C57BL/6 mice, the mean muscle fiber size of all fiber types from the hindlimb muscles was smaller at 2 years of age, corroborating previous reports [16] and an overall fiber size increase was evident with treatment. Fiber type diameters (mean ± SEM μm; derived from the mean of measurements individually made in each mouse) from untreated and treated 2-year-old C57BL/6 mice of both sexes as combined, and separately for males and females, along with inclusion of data from the same strain at 10 months of age in all 4 muscles are detailed in Supplementary Figures 4–7 and Supplementary Tables 1–12.

With AAV1.NT-3 gene therapy, we observed an overall reversal of the sarcopenic effect on fiber size and alterations in fiber type composition of all muscles examined. Changes in fiber type distribution of STO, FTO, and FTG (as a percent of total) and percent loss as “sarcopenic” (compared to 10-month-old) and percent gain as “NT-3 effect” in the glycolytic (FTG) and oxidative (STO) fiber size are shown in a summary format in Figure 3. Interestingly, alterations in fiber type composition showed different patterns for anterior and posterior compartment muscles of the distal hindlimb. In the tibialis anterior muscle from males, there was a reversal of sarcopenia-related decline in STO fibers, with a notable percentage increase with treatment, along with a decrease in FTG fiber type; suggesting a switch from FTG to STO fiber type. This glycolytic to oxidative fiber type switch resulted in a significant increase in STO fiber size (from 19.76% loss to 15.03% increase). In female tibialis anterior muscle however, an opposite pattern, from oxidative to glycolytic fiber type switch was noted, which was associated with prominent FTG fiber size increase (from 13.1% loss to 18.54% increase, Figure 3A, 3B, Supplementary Table 2). Contrarily, in the gastrocnemius muscle, FTG to STO fiber type switch was notable in females along with a significant increase in STO fiber diameter (from 12. 47% loss to 20.19% increase), while neither the fiber type composition nor the fiber diameter was altered with treatment in males significantly (Figure 3C, 3D, Supplementary Tables 5, 6).

Compared to 10 months old C57BL/6 mice, the greatest size reduction in FTG and STO occurred in quadriceps muscle from 2-year-old in both sexes and was more prominent in males. With treatment, the fiber type switch from glycolytic to oxidative was associated with reversal of sarcopenic effect on oxidative fiber size, in both sexes. The size increase in FTG was meager for both sexes (Figure 3E, 3F, Supplementary Tables 8, 9). Interestingly, triceps muscle appeared more resistant to the age-related size change contrasting with other muscles. The decrease in FTG fiber diameter, compared to those from 10 months old C57BL/6 mice was milder; a size increase occurred in STO and FTO fibers in males, while the STO fiber size decrease was minimal in females (Supplementary Tables 10–12). With treatment, the fiber type switch from glycolytic to oxidative reflected as size increase for both STO and FTO fibers in both sexes, was more prominent for FTO than STO (11.3% in FTO vs. 4% in STO) in males. In addition, males showed 11.38 % size increase in FTG. A trend for STO and FTO fiber type increase was present with aging in males when untreated 2-year-old samples were compared to those from 10-month-old (Figure 3G, 3H).

Aging process in Schwann cells, can easily be identified on the high resolution semi-thin cross sections of peripheral nerves which includes an overall thinning of myelin [17, 18] and age-related myelin alterations, such as increased myelin corrugation, infoldings and outfoldings reflecting axonal atrophy. We observed these changes in sciatic and tibial nerves distally in 2-year-old C57BL/6 mice. To assess the efficacy of NT-3 gene therapy on the aging process in peripheral nerves, we chose to evaluate G ratio (axon diameter/fiber diameter), a measure of myelin thickness in the distal tibial nerve, from aged C57BL/6 mice at 6 months post gene delivery, and compared to untreated counterparts (Figure 4A, 4B). Thinning of myelin in rodent nerves with aging has been reported previously, although mean G ratio in sciatic nerves from aged group was not found to be significantly different from younger groups, explained by the presence of concomitant axon size decrease [17–19]. In distal tibial nerves however, we found significantly increased mean G ratio in aged C57BL/6 mice, reflecting an age related thinning of myelin, when compared to data obtained from 1 year old mice (Figure 4C, 4D; 2-year-old, 0.727 ± 0.0013 vs. 1-year old, 0.665 ± 0.0015; p < 0.0001). The percent of fibers with a G ratio ≥ 0.7, a reflection of thin myelin in aged mice, constituted 67.79%, whereas in 1 year old mice, this was only 27.32% of total fibers. NT-3 gene therapy resulted in an attenuation of age-related myelin change, indicated by treated and untreated groups having significantly different slopes of G ratio scatterplots (Figure 4E). The mean G ratio was significantly improved toward optimal peripheral nerve G ratio of 0.6 reflecting thicker myelin at 6 months post AAV1.NT-3 gene delivery (NT-3, 0.672 ± 0015 vs. untreated, 0.727 ± 0.0013; p < 0.0001). The percent fibers with G ratio ≥ 0.7 representing thin myelin reduced to 32.62%, which constituted 67.79% of total fibers in the untreated cohort (Figure 4F).

We also assessed the status of NMJ in response to NT-3 gene therapy in this model as there is evidence that changes in endplate morphology and NMJ remodeling occur with aging and precede loss of fast motor units [20]. Using immunohistochemistry-based parameters [11], we analyzed a total of 330 NMJs derived from intrinsic foot muscles of NT-3 treated and untreated aged C57BL/6 mice. This analysis showed that AAV1.NT-3 gene therapy, at 6 months of treatment gave rise to a 29.7% increase of innervated NMJs (p = 0.0123). A decrease in the denervated and partially denervated/innervated NMJs was also noted (Figure 4G, 4H). We believe that these quantitative histopathological assessments of muscle, NMJ and nerve collectively provide strong support to the functional improvements observed in aged C57BL/6 mice with NT-3 gene therapy.

NT-3 induced remodeling of muscle metabolism in age-related sarcopenia

Studies have shown that mtDNA and mRNA abundance and mitochondrial ATP production, all decline with advancing age [21–23]. We first started by investigating whether the mitochondria biogenesis marker Pgc1α transcripts and mtDNA copy number per genomic DNA are altered in muscle at 6 months post NT-3 gene therapy in aged C57BL/6 mice. In the untreated tibialis anterior muscle, Pgc1α transcripts showed a significant decline in both sexes at 2 years of age, compared to 10 months old mouse samples (Figure 5A). The response to NT-3 gene therapy, however, showed sex difference; with males having significantly more Pgc1α transcripts than females. The mitochondria copy number in females from the untreated cohort was significantly lower than 10-month-old samples, while the treated cohort showed an increase. Interestingly, mitochondria copy number in the untreated and treated muscle from males did not differ from the 10-month-old males (Figure 5B). We then analyzed the expression levels of Cox1 and Cox3, which are mtDNA encoded subunits of cytochrome c oxidase of respiratory complex IV and Atp5d subunit of complex V, encoded by genomic DNA. We also investigated if intensity-quantification of histochemical staining for cytochrome c oxidase (COX) enzyme activity in these muscles show any correlation with the expression levels of these subunits, and/or directly reflect the functional mitochondria content in the muscle. As expected, Cox1 and Cox3 transcripts significantly declined in the untreated tibialis anterior muscle at 2 years of age compared to the levels from younger C57BL/6 controls (Figure 5C, 5D). AAV1.NT-3 treatment gave rise to a trend of increased Cox1 expression, contrasting with highly significant increases in nuclear encoded Atp5d transcripts in both sexes, which was more prominent in females (Figure 5C–5E). We also found COX enzyme stain-intensity quantification correlating well with increased Atp5d expression levels and mitochondria content, more notable in females compared to the untreated muscle (Figure 5F–5J).

Considering the significant size increase in FTG fibers, along with increased glycolytic fiber type in the female tibialis anterior muscle, we next explored the possibility of mTORC1 activation, as previously shown with NT-3 in the neurogenic muscle from trembler J mouse (Tr^J) [9]. In western blot analysis, we found significant increases in the downstream targets of mTORC1, phosphorylated S6 kinase and 4E-BP1 proteins, indicating the activation of mTORC1 which correlate with increased radial growth of FTG fibers (Figure 5K–5M). In addition, we found a trend for increased expression of glycolytic enzymes Pfkm and Hk-1 in females, suggesting increased carbohydrate metabolism in muscle. In contrast, expression of phosphorylated S6 protein and HK-1 transcripts were lower in males compared to untreated samples, in keeping with the sex-dependent pattern of changes in this muscle (Figure 5N, 5O).

We found similar changes with aging in the untreated gastrocnemius muscle at 2 years of age for both Pgc1α transcripts as well as the mitochondria copy number, which were significantly lower than 10-month-old counterparts (Figure 6A, 6B). At 6 months post NT-3 gene therapy, although not statistically significant, there was further decline in Pgc1α relative expression levels compared to untreated counterparts, without sex difference (Figure 6A). The mitochondria copy number increased in the female muscle without reaching significance, contrasting with a decline in the samples from males (Figure 6B). Cox1 and Cox3 transcripts from untreated muscles at 2 years of age significantly declined, while Atp5d expression increased compared to the levels from 10 months old C57BL/6 muscle (Figure 6C–6E). With NT-3 gene therapy, we found no significant change in the expression levels of these transcripts, although there were sex-dependent differences in staining intensity of COX enzyme in the NT-3 treated muscle, compared to Ringer’s lactate-treated controls (Figure 6F–6I). Females exhibited a stronger staining intensity with treatment, while an opposite pattern was present in males, showing stronger staining intensity in the untreated muscle (Figure 6J). The COX enzyme stain-intensity quantification reflected the changes in Cox1, Cox3, and Atp5d transcripts and mitochondria content between males and females, compared to untreated muscle. Contrasting with tibialis anterior, in western blot analysis, phosphorylated S6 kinase and 4E-BP1 proteins showed significant increases in males, which reflected a small radial growth only in the FTO fibers, while no size increase was noted in FTG fibers in response to treatment (Figure 6K–6M, Figure 3D, Supplementary Table 6). Moreover, the Pfkm and Hk-1 transcripts were lower in males compared to untreated samples suggesting that males did not respond to NT-3 induced anabolic stimulation by increasing carbohydrate metabolism in the gastrocnemius muscle. Females, nonetheless, showed an increase trend for phosphorylated S6 kinase protein levels and the Pfkm and Hk-1 transcripts without reaching significance (Figure 6L–6O).

Animals and treatment groups

Naturally aged C57BL/6 mice were included in the study (JAX stock #000664) and all animal experiments were performed according to the guidelines approved by The Research Institute at Nationwide Children’s Hospital Animal Care and Use Committee that operates in full accordance with the Animal Welfare Act and the Health Research Extension Act (IACUC approval number = AR18-00076). 18 months old C57BL/6 (6 males and 6 females, n = 12) mice received 1 × 10¹¹ vg dose of AAV1.tMCK.NT-3, via intramuscular (IM) injection into the gastrocnemius muscle. Age- and sex-matched C57BL/6 (6 males and 6 females, n = 12) mice were injected with Ringer’s lactate as controls. Mice were tested functionally with treadmill, rotarod, and in vivo muscle contractility assay and they were sacrificed six months post-injection by an over-dosage of xylazine/ketamine anesthesia for harvesting blood, sciatic nerves, as well as upper and lower limb muscles including lumbricals, at six months post gene injection.

AAV1.tMCK.NT-3 vector production and potency

Construct of self-complimentary (sc) AAV serotype 1 vector with muscle specific tMCK promoter was described previously [10]. The vector was produced in our Viral Vector Core at Nationwide Children’s Hospital, Columbus (Andelyn Biosciences). Aliquots of virus were stored at −80°C until used. Serum was separated from blood samples that were collected by cardiac puncture at six months post gene injection and serum NT-3 levels were detected by ELISA as previously reported (NT-3, n = 9; UT, n = 8) [10].

Run to exhaustion test

Run to exhaustion treadmill performance test was performed at two-, four- and six-months post injection. Mice were exercised to exhaustion via treadmill (Columbus Instruments, Exer-6M Treadmill), as described previously [15]. Mice were acclimated to the treadmill prior to data collection. Mice were run to exhaustion with increasing treadmill speed by 1 meter/min each minute, starting at an initial 7 meter per minute velocity. Lanes have a shock plate that pulses at a frequency of ~3 Hz. Mice were considered at “exhaustion” level when they were unable to re-engage the treadmill for 3 seconds after resting on the shock-plate. Run duration was recorded and used to calculate the distance ran (2 months post-injection: NT-3, n = 12; UT, n = 11; 4 months post-injection, NT-3, n = 12; UT, n = 11; 6 months post-injection: NT-3, n = 11; UT, n = 8).

In vivo muscle contractility assay

In vivo muscle contractility assay was performed at the endpoint, as described previously [15]. Hind paw of the anesthetized mouse was placed on footplate, which was attached to a dual-mode lever, and the tibia was aligned perpendicular to the lever. Subcutaneous EMG electrodes were used to stimulate the tibial nerve. Gastrocnemius muscle torque around the ankle joint was measured by muscle physiology apparatus (Aurora Scientific, ON, Canada) using isometric contraction (maximum twitch response) and fatigue (maximum tetanic response) protocols (Max twitch: NT-3, n = 11; UT, n = 8; Max tetanic: NT-3, n = 10; UT, n = 9).

Rotarod

Rotarod test was performed at six months post-injection to assess motor function and balance of the mice. Mice were acclimated to rotarod apparatus (Columbus Instruments, Ohio, USA) prior to data collection. Rotarod protocol included a 5-rpm run with a constant acceleration of 0.2 rpm/s. The averages of the best two out of three runs were calculated (NT-3, n = 10; UT, n = 8).

Musculoskeletal and skin changes

Age-related musculoskeletal and skin changes including kyphosis, dermatitis, and alopecia were documented semi-quantitatively. Six mice, from both treated and untreated cohorts, were photographed at the end point and the severity of kyphosis, dermatitis, and alopecia were scored. Severe changes were scored as 1, mild-moderate changes as 0.5, and no changes as 0. For kyphosis, arbitrary lines were drawn tangential to the scoliosis curvature to delineate the scoliosis angle (Supplementary Figure 3A). Angle range 90°–110° is considered as severe, 11°–150° as mild-moderate and >150° as none.

Histological analysis

Immunohistochemical analysis of neuromuscular junctions (NMJ)

Lumbrical muscles collected from treated and untreated mice (n = 4 for both cohorts with equal sex distribution) were processed as described previously [11]. Muscles were fixed and stained with primary antibodies (Acetylcholine receptor (AChR) antibody, α Bungarotoxin, T1175, 1:500; Anti Neurofilament 200 antibody, N4142, 1:500; SV2 antibody, AB_2315387, 1:50) [49, 50], followed by incubation with secondary antibodies (Alexa Fluor 488 conjugated anti-rabbit and anti-mouse IgG, 1:500). Samples were imaged at 60× magnification using a Zeiss LSM 800 confocal microscope. NMJs were considered to be innervated when nerve completely overlapped the AChRs, as partially innervated when some parts of the AChRs were not overlapping with nerve, and as denervated when there was not any nerve co localizing with AChRs [11, 51]. An average of 41.3 NMJs per mouse were evaluated from NT-3-treated and UT-mice (n = 4 mice per group with equal sex distribution).

Expression analysis

COX staining density analysis

Histochemical enzyme activity of cytochrome c oxidase (COX) was assessed by quantifying the COX stain intensity in tibialis anterior and gastrocnemius muscles from treated and untreated mice. 12 μm thick-fresh frozen sections were stained using COX enzyme histochemistry protocol established in our clinical neuromuscular pathology laboratory. Muscle sections representing deep, intermediate, and superficial zones from gastrocnemius (n = 8 treated, n = 7 untreated) and tibialis anterior (n = 6 treated, n = 6 untreated) muscles were photographed at 10× magnification, using an Olympus BX41 microscope and SPOT camera, with even distribution of sex. Image data were collated and processed utilizing the Python programming language. [55] Images were calibrated using representative selections from each image to establish ranges of intensity for each fiber-type; dark (oxidative/higher mitochondria content/type 1) and light (glycolytic/lower mitochondria content/type 2) fibers, including COX deficient pale fibers. Ratio of dark to light fibers was calculated and presented as mean ± SEM.

Statistics

Adequate sample size was determined according to our previous studies that performed analogous experiments [11, 15]. All statistical analyses were performed in GraphPad Prism 9.0 software. Two tail Student t-test, one-way ANOVA with Tukey’s multiple comparison test, two-way ANOVA with Sidak’s multiple comparison test or linear regression analysis were performed when applicable, and significance level was set at P ≤ 0.05. The tests that meet the best assumptions of the data were chosen. Results were given as mean ± SEM in all experiments and the number of animals was mentioned in figure legends along with the name of the statistical analysis performed. Other than the functional tests and quantitative muscle analysis, no blinding was used.