Efficacious Androgen Hormone Administration in Combination with Adeno-Associated Virus Vector-Mediated Gene Therapy in Female Mice with Pompe Disease

Abstract

Pompe disease is an autosomal recessive lysosomal storage disorder caused by deficiency of acid α-glucosidase (GAA), resulting in skeletal muscle weakness and cardiomyopathy that progresses despite currently available therapy in some patients. The development of gene therapy with adeno-associated virus (AAV) vectors revealed a sex-dependent decrease in efficacy in female mice with Pompe disease. This study evaluated the effect of testosterone on gene therapy with an AAV2/8 vector containing a liver-specific promoter to drive expression of GAA (AAV2/8-LSPhGAA) in female GAA-knockout (KO) mice that were implanted with pellets containing testosterone propionate before vector administration. Six weeks after treatment, neuromuscular function and muscle strength were improved as demonstrated by increased Rotarod and wirehang latency for female mice treated with testosterone and vector, in comparison with vector alone. Biochemical correction improved after the addition of testosterone as demonstrated by increased GAA activity and decreased glycogen content in the skeletal muscles of female mice treated with testosterone and vector, in comparison with vector alone. An alternative androgen, oxandrolone, was evaluated similarly to reveal increased GAA in the diaphragm and extensor digitorum longus of female GAA-KO mice after oxandrolone administration; however, glycogen content was unchanged by oxandrolone treatment. The efficacy of androgen hormone treatment in females correlated with increased mannose-6-phosphate receptor in skeletal muscle. These data confirmed the benefits of brief treatment with an androgen hormone in mice with Pompe disease during gene therapy.

BACKGROUND

Pompe disease, or glycogen storage disease type II, is an autosomal recessive metabolic myopathy resulting from pathogenic variants in the acid alpha-glucosidase (GAA) gene. GAA is a lysosomal hydrolase that breaks down glycogen. GAA deficiency results in accumulation of lysosomal glycogen with an associated myopathy. Pompe disease has a spectrum of severity from infantile-onset forms, where the presentation is more acute and to milder, late-onset forms, which correlates with the relative deficiency of GAA.¹ In the most severe cases, progressive weakness of respiratory muscles may culminate in cardiorespiratory failure.² Although late-onset forms typically do not develop cardiomyopathy, symptomatic patients typically present with progressive skeletal myopathy or protracted respiratory failure at any age.¹

In 2006, the Food and Drug Administration (FDA) approved enzyme replacement therapy (ERT) with glucosidase alfa for the treatment of Pompe disease. Although this therapy has reduced glycogen accumulation and improved patient symptoms and survival, ERT requires frequent intravenous injections and is very costly.³ Furthermore, some patients with a good ERT response continue to have motor weakness as noted by residual weakness in the neck flexor and dorsiflexor muscles, and myopathic facies, ptosis, and strabismus.^3–5

Owing to the expense and the limited efficacy of ERT, alternatives in the form of gene therapy with adeno-associated virus (AAV) vectors are being explored as a treatment for Pompe disease.^6,7 Studies in in vivo mice models have shown that administration of AAV has resulted in the long-term correction of cardio and skeletal muscle myopathies.¹ However, they have also demonstrated a sex-dependent decrease in the efficacy of AAV vectors in female mice,^8–10 which could impair the response of female patients if the same phenomenon occurs in humans. A similar sex-dependent decrease in the efficacy of AAV vector also occurs in female mice with Pompe disease.^11,12

We evaluated the possibility that androgens such as testosterone enhance the efficacy of an AAV vector in female mice with Pompe disease, which could work through multiple mechanisms: (1) increased transduction from the AAV vector⁸; (2) increased Igf-1 expression¹³ that will increase cation-independent mannose-6-phosphate receptor (CI-MPR) expression and promote receptor-mediated uptake of GAA in muscle^6,14; (3) muscle hypertrophy and increased lean body mass^13,15; and (4) suppression of abnormally increased macroautophagy (“autophagy”) through activating Akt and suppressing mTOR pathway.^16,17

The previously observed efficacy of increasing CI-MPR expression with a small molecule during gene therapy support the strategy of administering drug therapy during gene therapy in Pompe disease.^18–20

The current studies evaluated the effect of testosterone or oxandrolone upon AAV vector-mediated gene therapy in the GAA knockout (KO) mouse model for Pompe disease. Initially, mice were treated with testosterone before vector administration. Immunoblotting was conducted to measure CI-MPR uptake in skeletal muscle, showing improvement in mice treated with both vector and testosterone as compared with gene therapy alone. Neuromuscular function, muscle strength, and biochemical correction were evaluated for mice treated with both testosterone and vector, in comparison with gene therapy alone.

Results demonstrated a beneficial effect from androgen hormone administration in female mice with Pompe disease. In a second experiment, mice were treated with vector and oxandrolone in combination with various timing of oxandrolone exposure. Biochemical correction and muscle strength improved for oxandrolone-treated female mice with Pompe disease, demonstrating partial efficacy of oxandrolone in combination with gene therapy.

METHODS

Study 1: testosterone in combination with AAV vector-mediated gene therapy

The AAV vector was prepared as described and administered intravenously to female GAA-KO mice with a C57BL/6 background.^21,22 Rotarod and wirehang testing was performed before treatment and 6 weeks after vector administration as described.⁷ Mice were treated with tail vein injection of AAV2/8 vector containing a liver-specific promoter to drive expression of GAA (AAV2/8-LSPhGAA) (5 × 10¹⁰ vector genomes [vg]) on day 1. Female GAA-KO mice (8 per group) were implanted subcutaneously with pellets containing testosterone propionate or with placebo pellets on day 5. Female GAA-KO mice (n = 8) were injected with PBS as a negative control for vector administration, and implanted with either testosterone or placebo pellets.

Age-matched female GAA-KO mice were housed in groups of 3–5, and mice from different groups were co-housed when possible. GAA activity and glycogen content were analyzed as described using biological assays and normalized to protein content.²² Western blot analysis was also performed. Groups of GAA-KO mice were killed for tissue analysis at 6 weeks after testosterone (or sham) treatment and vector administration. All animal procedures were carried out in accordance with Duke University Institutional Animal Care and Use Committee-approved guidelines.

Study 2: oxandrolone in combination with AAV vector-mediated gene therapy

The AAV2/8-LSPhGAA vector AAV2/8-LSPhGAA (5 × 10¹⁰ vg) was administered at 3 months of age on day 1. To achieve a prolonged effect from oxandrolone in mice, we dosed it continuously in drinking water. We chose the duration of 3 weeks to imitate the duration of slow-release testosterone. We also dosed early to attempt to enhance transduction with the AAV vector, and dosed late to evaluate for an effect on CI-MPR–mediated uptake of GAA. Equal numbers of female and male GAA-KO mice were dosed with oxandrolone in drinking water in two different treatment groups of 4–5 mice each after vector administration at two time points: weeks 1–4 and weeks 14–18.

Group sizes were as follows: Females—Group 1 (n = 5), Group 2 (n = 4), Group 3 (n = 5), Group 4 (n = 5), Group 5 (n = 7). Males—Group 1 (n = 7), Group 2 (n = 4), Group 3 (n = 4), Group 4 (n = 5), Group 5 (n = 4). Control vector-treated GAA-KO mice (n = 5 of each sex) received regular drinking water, and drug control mice received oxandrolone without any vector injection for weeks 14–18. Mock-treated GAA-KO mice (n = 5 of each sex) were injected with PBS as a negative control group for this experiment. Mice were housed following the protocol described in Study 1.

Rotarod and wirehang testing occurred as described before treatment and 6, 12, and 18 weeks after treatment as described.⁷ GAA and glycogen content were analyzed through GAA activity assays, glycogen content assays, and total protein assays as described, with GAA and glycogen data normalized to protein content.²³ In addition, immunoblot analysis was performed as described.²⁰ All mouse groups were killed for tissue analysis at 18 weeks after vector administration. All animal procedures were carried out in accordance with Duke University Institutional Animal Care and Use Committee-approved guidelines, as in Study 1.

Statistics

Multiple comparisons were assessed with a one-way analysis of variance ANOVA and Sidak's multiple comparisons test using Prism software (GraphPad, La Jolla, CA), using an alpha of 0.05. Statistically significant differences are indicated by a horizontal line extending between two groups, with asterisks denoting the p-values as described in the figure legends.

RESULTS

Study 1: testosterone in combination with AAV vector-mediated gene therapy

The AAV vector evaluated here, AAV2/8-LSPhGAA, has been efficacious with regard to biochemical correction and immune tolerance induction to human GAA in male and female adult GAA-KO mice, although with greater efficacy in male mice.^12,23–25 The testosterone or placebo was administered 4 days after vector administration to groups of 3-month-old female GAA-KO mice (Fig. 1A).⁸ In contrast to the earlier study by Davidoff and colleagues, no oophorectomy was performed, thus evaluating the effects of testosterone in females with normal sexual function. Male mice were not examined in this experiment, which was intended to evaluate enhanced efficacy from testosterone in female mice. However, both sexes were evaluated in the study of a second androgen hormone, oxandrolone (Fig. 1B).

Figure 1.

Schematic of experimental design for testosterone (A) and oxandrolone (B). In an 8-week study, GAA-KO mice were treated with placebo, testosterone, a single injection of vector (AAV2/8-LSPhGAA), or both vector and testosterone (A). In an 18-week study, GAA-KO mice were treated with placebo, oxandrolone delivered through drinking water for 4 weeks, a single injection of vector (AAV2/8-LSPhGAA), both vector and early doses of oxandrolone (weeks 1–4), or both vector and late doses of oxandrolone (weeks 14–18). AAV, adeno-associated virus; AAV2/8-LSPhGAA, AAV2/8 vector containing a liver-specific promoter to drive expression of GAA; GAA, acid α-glucosidase; KO, knockout.

Transduction and transgene expression were evaluated in female mice to better understand the effects of testosterone during gene therapy in Pompe disease. Transduction was evaluated using quantitative polymerase chain reaction (qPCR) of liver DNA for both the testosterone and oxandrolone experiments, producing analogous results. Equivalent transduction was detected in the liver for the vector-treated groups (Fig. 2A). Similarly, expression was unchanged after administration of vector and testosterone in comparison with vector-alone group, as evaluated using reverse transcriptase qPCR (Fig. 2A; p < 0.05).

Figure 2.

Transduction and expression of vector and biochemical correction in the liver with testosterone. Liver qPCR and RT-qPCR were conducted to determine the transduction and expression of the vector genome in mice treated with testosterone (A). GAA-KO mice were evaluated for GAA activity and glycogen content in the liver in each treatment group (n = 5) (B). Mean ± SD is given. p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), p < 0.0001 (****). qPCR, quantitative polymerase chain reaction; RT, reverse transcriptase; SD, standard deviation.

In contrast, liver GAA activity was significantly increased by the administration of vector and testosterone, in comparison with placebo or testosterone alone (Fig. 2B; p < 0.01). Consistent with correction of GAA deficiency, liver glycogen content was uniquely decreased by the administration of both vector and testosterone, in comparison with placebo or testosterone alone (Fig. 1B; p < 0.05). These data demonstrated that enhanced efficacy from testosterone treatment during gene therapy was attributable to improved biochemical correction, rather than increased transduction or transgene expression.

Efficacy from testosterone was further evaluated through Rotarod and wirehang testing, which have been validated as endpoints in GAA-KO mice (Fig. 3).^6,7 Rotarod latency was significantly increased in female mice treated with vector and testosterone, in comparison with placebo (Fig. 3A; p < 0.05).

Figure 3.

Testosterone muscle and weight endpoint measurements. Female GAA-KO mice in each group (n = 5) were evaluated for change in latency on rotarod tests (A) and wirehang tests (B). Multiple body weight measures and ratios were calculated, including change in body weight (C) and ratio of left ventricle to body weight (D). Mean ± SD is given. p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), and p < 0.0001 (****).

Similarly, wirehang latency was significantly elevated for mice treated with vector and testosterone, in comparison with placebo (Fig. 3B; p < 0.01). Body weight decreased significantly in mice treated with vector and testosterone, in comparison with placebo (p < 0.05; Fig. 3C). Left ventricle weight increased slightly in groups treated with testosterone, although vector administration ameliorated this effect (Fig. 3D; p < 0.05). These changes in body weight and left ventricular size could reflect detrimental effects of testosterone; however, muscle function improved as demonstrated by Rotarod and wirehang testing after testosterone treatment during gene therapy in female GAA-KO mice.

Subsequently, the biochemical correction of skeletal muscle was evaluated to determine the effect of testosterone on the secretion and uptake of liver-expressed GAA. The GAA activity in diaphragm was increased after vector administration, when compared with placebo and testosterone control groups (Fig. 4A; p < 0.05). The diaphragm's glycogen content significantly decreased with the administration of both vector and testosterone, in comparison with each of the other conditions, which demonstrated a benefit in the diaphragm from testosterone treatment during gene therapy (Fig. 4A).

Figure 4.

Testosterone GAA activity and glycogen content. Female GAA-KO mice in each group (n = 5) were evaluated for GAA activity levels and glycogen content in the heart, diaphragm, and quadriceps (A), and the soleus, EDL, and gastrocnemius (B). Mean ± SD is shown. p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), p < 0.0001 (****). EDL, extensor digitorum longus.

Similarly, testosterone treatment improved biochemical correction of the soleus muscle, as demonstrated by increased GAA activity and decreased glycogen content after the administration of both vector and testosterone, in comparison with each of the other conditions (Fig. 4B).

GAA activity in the quadriceps increased significantly after the administration of both vector and testosterone in comparison with each of the other conditions (Fig. 4A). However, quadriceps glycogen content did not decrease after vector administration (Fig. 4A), consistent with the resistance of skeletal muscle to biochemical correction in Pompe disease.²⁵ The pattern of biochemical correction for the gastrocnemius was similar to the quadriceps (Fig. 4B).

The relevance of this lack of glycogen reduction was explained by the similar pattern for the extensor digitorum longus (EDL), a muscle comprising primarily type II myofibers (Fig. 4B). The glycogen content in these muscles was not decreased in the face of increased GAA activity, which is consistent with the resistance to biochemical correction for muscles comprising mainly type II myofibers in Pompe disease.²⁶ On the contrary, the soleus primarily comprises type I myofibers and glycogen content was decreased by the higher GAA activity present after testosterone treatment (Fig. 4B).

Heart GAA was increased by vector administration with and without testosterone (Fig. 4A), in comparison with either placebo or testosterone administration (p < 0.0001). The heart showed significantly decreased glycogen content in vector treatment groups, both with and without testosterone (Fig. 4A). Overall, the enhanced biochemical correction of the heart, diaphragm, and soleus demonstrated a unique response to vector and testosterone administration that was consistent with the good response for these muscles to GAA replacement.^25,26

Study 2: oxandrolone in combination with AAV vector-mediated gene therapy

Another androgenic hormone, oxandrolone, was evaluated subsequently to determine its effects on gene therapy in Pompe disease. Oxandrolone is a synthetic anabolic steroid medication that promotes the growth of muscle tissue, and an agonist of testosterone.²⁷ AAV2/8-LSPhGAA was administered at 13 weeks of age to groups of GAA-KO mice treated with oxandrolone in drinking water, either early or late, or with vehicle (regular water) during an 18-week study (Fig. 1B). Both sexes were evaluated to measure effects of oxandrolone on vector-mediated efficacy. The transduction in the female GAA-KO mouse liver was unchanged after oxandrolone administration (Fig. 5A), similar to the result after testosterone administration (Fig. 2A).

Figure 5.

Transduction of vector and biochemical correction in the liver with oxandrolone. Liver qPCR was conducted to determine the transduction of the vector genome in female GAA-KO mice treated with oxandrolone (A). GAA-KO mice were evaluated for GAA activity and glycogen content in the liver (B). Group 1 (n = 5), Group 2 (n = 4), Group 3 (n = 5), Group 4 (n = 5), Group 5 (n = 7). Mean ± SD is given. p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), p < 0.0001 (****).

Liver GAA activity was significantly increased in all vector treatment groups, when compared with untreated and oxandrolone-only groups (p < 0.0001; Fig. 5B). Liver glycogen content was significantly decreased after treatment with vector and late oxandrolone administration, in comparison with oxandrolone and vector control groups (p < 0.01; Fig. 5C). Furthermore, administration with vector and early oxandrolone administration, in comparison with oxandrolone alone (Fig. 2B; p < 0.05), demonstrated efficacy with regard to biochemical correction in the liver.

Rotarod and wirehang testing were conducted to evaluate muscle function, and benefits were demonstrated in female GAA-KO mice (Fig. 6). Rotarod latency was significantly increased in female GAA-KO mice for the vector and late oxandrolone group, in comparison with the untreated group (p < 0.05; Fig. 6A). Wirehang latency was unchanged by any treatment (Fig. 6B). Body weight uniquely increased after vector and early dose oxandrolone administration, in comparison with the untreated group (Fig. 6C; p < 0.01). Left ventricle size decreased significantly in all vector-treated groups of female GAA-KO mice, in comparison with the untreated group (Fig. 6D), demonstrating efficacy from gene therapy.

Figure 6.

Oxandrolone muscle and weight endpoint measurements. Female GAA-KO mice were evaluated for change in latency on Rotarod tests (A) and wirehang tests (B). Multiple body weight measures and ratios were calculated, including change in body weight (C) and ratio of left ventricle to body weight (D). Group 1 (n = 5), Group 2 (n = 4), Group 3 (n = 5), Group 4 (n = 5), Group 5 (n = 7). Mean ± SD is given. p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), p < 0.0001 (****).

For male mice, Rotarod latency was significantly increased in the vector and early dose oxandrolone group, in comparison with the vector-alone group (Supplementary Fig. S1A; p < 0.05), whereas no effect was observed in wirehang latency (Supplementary Fig. S1B). Body weight increased in male groups receiving vector alone or vector and late dose oxandrolone, in comparison with the untreated group (Supplementary Fig. S1C; p < 0.05) and cardiac hypertrophy decreased in all male groups receiving vector, with and without oxandrolone, in comparison with the untreated group (Supplementary Fig. S1D; p < 0.0001). Thus, efficacy was demonstrated with regard to muscle function and weight changes for both sexes from gene therapy and from the addition of oxandrolone during gene therapy.

After vector administration, biochemical correction was evaluated in the heart and skeletal muscles. The GAA activity was increased in diaphragm after administration of vector and late oxandrolone, in comparison with the untreated and oxandrolone alone groups (Fig. 7A). The diaphragm's glycogen content significantly decreased in all vector treated groups, in comparison with untreated and oxandrolone-alone groups (Fig. 7A). Quadriceps GAA activity was significantly increased after the administration of vector and late oxandrolone, in comparison with oxandrolone alone (Fig. 7A). Quadriceps glycogen content was uniquely decreased after vector and early dose oxandrolone administration, in comparison with the untreated group (p < 0.05; Fig. 7A).

Figure 7.

Oxandrolone GAA activity and glycogen content. GAA-KO mice were evaluated for GAA activity levels and glycogen content in the heart, diaphragm, and quadriceps (A), and the soleus, EDL, and gastrocnemius (B). Group 1 (n = 5), Group 2 (n = 4), Group 3 (n = 5), Group 4 (n = 5), Group 5 (n = 7). Mean ± SD is shown. p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), p < 0.0001 (****).

GAA activity was increased after administration of vector and early oxandrolone, in comparison with the untreated and oxandrolone alone groups (Fig. 7B). Similar to the effect of testosterone, oxandrolone and vector treatment failed to reduce glycogen content in either the EDL or gastrocnemius, despite the presence of elevated GAA activity (Fig. 7B).

Heart GAA was increased by vector administration with and without oxandrolone at all time points (Fig. 7A), in comparison with either untreated mice or oxandrolone administration (p < 0.0001). The heart's glycogen content was decreased in all vector-treated groups, in comparison with untreated or oxandrolone groups (p < 0.0001; Fig. 7A). The soleus had significantly increased GAA activity after treatment with vector and late oxandrolone, in comparison with placebo and oxandrolone control groups (Fig. 7B).

The soleus exhibited significantly decreased glycogen content after administration of vector and both early and late oxandrolone, in comparison with placebo and oxandrolone control groups (Fig. 7B). However, the soleus and EDL both contained increased glycogen content in the late oxandrolone and vector group, compared with the vector-alone group (Fig. 7B). These data demonstrated a lack of correlation between increased GAA activity and decreased glycogen content after oxandrolone treatment in both the soleus (type I myofibers) and EDL (type II myofibers) in female GAA-KO mice.

The effect of oxandrolone in decreasing muscular glycogen content was slightly diminished in male GAA-KO mice during gene therapy in comparison with female mice. Significant increases in GAA activity in males were observed in all groups receiving vector, with and without oxandrolone, in the liver (Supplementary Fig. S1A; p < 0.0001).

GAA activity was increased in the quadriceps for the vector-only and vector and late oxandrolone groups, in comparison with oxandrolone alone (p < 0.05); in the EDL in the vector and late oxandrolone group, in comparison with the untreated group (p < 0.05); in the soleus in the vector and late oxandrolone group, in comparison with oxandrolone alone (p < 0.05); and in the heart in the vector-only and vector and late oxandrolone group, in comparison with untreated and oxandrolone-alone groups (p < 0.001; Supplementary Fig. S2A, B).

There were no significant increases in GAA activity observed in the gastrocnemius. Glycogen content was significantly decreased in the quadriceps in the vector-only group, in comparison with placebo and oxandrolone control groups; in the EDL in the vector-only group, in comparison with placebo; in the soleus in the vector-only group, in comparison with placebo and oxandrolone control groups; and in the heart in all groups receiving vector, when compared with placebo and oxandrolone control groups (Supplementary Fig. S2A, B). Concerns were raised by the observation of significantly increased glycogen content for the quadriceps and EDL in late and early oxandrolone vector-treated groups, in comparison with the vector-only group (Supplementary Fig. S2A, B).

Mechanism for increased correction of GAA deficiency

The effect of androgens upon gene therapy in female GAA-KO mice was further evaluated by immunoblotting of skeletal muscles. A marker for abnormal autophagy in Pompe disease,²⁸ microtubule-associated protein 1A/1B-light chain 3 (LC3), was quantified in the gastrocnemius and quadriceps (Fig. 8). Testosterone and vector administration decreased LC3 in the quadriceps, in comparison with untreated controls, consistent with correction of autophagy after testosterone administration (Fig. 8A).

Figure 8.

Immunoblotting of skeletal muscles. Microtubule-associated protein 1A/1B-LC3, p62, and CI-MPR were quantified in each treatment group in the gastrocnemius and quadriceps of GAA-KO mice. (A) Testosterone study (n = 3 per group). (B) Female mice in oxandrolone study, including untreated (n = 4), vector alone (n = 4), and vector with late oxandrolone treatment (n = 5). (C) Male mice in oxandrolone study, including untreated (n = 4), vector alone (n = 4), and vector with late oxandrolone treatment (n = 5). Mean ± SD is given. p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), p < 0.0001 (****). CI-MPR, cation-independent mannose-6-phosphate receptor; LC3, light chain 3.

However, the CI-MPR was increased in gastrocnemius after the administration of vector-only, in comparison with untreated GAA-KO muscle (Fig. 8A). Therefore the mechanism for increased GAA activity in skeletal muscle after testosterone treatment was potentially increased GAA expression, because higher CI-MPR expression to increase receptor-mediated uptake was not observed after testosterone administration. Immunoblotting was also performed to evaluate whether oxandrolone produced similar effects (Fig. 8B, C).

In female mice, treatment with either vector alone or both vector and oxandrolone resulted in increased CI-MPR in the quadriceps (Fig. 8B). However, only combined vector and oxandrolone resulted in consistently decreased LC3 in the quadriceps and gastrocnemius (Fig. 8B). For male mice, the vector-only treatment resulted in decreased LC3 in the quadriceps (Fig. 8C). However, the oxandrolone and vector treatment group demonstrated increased SQSTM1/p62 in the gastrocnemius as compared with the vector-only group (Fig. 8C). The latter effect indicated exacerbation of elevated autophagy in the muscle of male mice after oxandrolone administration, which provides insight into the elevated glycogen content in the gastrocnemius for male mice receiving oxandrolone (Supplementary Fig. S2).

DISCUSSION

In this study, multiple beneficial effects of testosterone were demonstrated during AAV vector-mediated gene therapy in female GAA-KO mice, including increased CI-MPR–mediated uptake of GAA and suppressed autophagosome formation. Surprisingly, these effects were independent of liver transduction, which remained equivalent with or without testosterone treatment despite previous studies reporting that testosterone increased transduction in the liver of female mice (Fig. 2A). Vector genomes were not further increased in the liver by the addition of testosterone or oxandrolone treatment, indicating a lack of an effect from androgen hormones upon transduction (Figs. 2 and 5).

However, it is possible that we achieved high-level transduction of the liver, and therefore transduction was not limiting the female response to gene therapy in this study. The benefits from androgen treatment during gene therapy in GAA-KO mice included improved biochemical correction from increased GAA uptake in muscle correlated with increased muscle strength and neuromuscular function that were detected with the wirehang and Rotarod tests, respectively. Benefits were greater in female mice than in male mice for oxandrolone. Other effects of androgens, including muscle hypertrophy from testosterone administration may potentially be beneficial even in the absence of increased CI-MPR expression owing to the enhanced uptake of circulating GAA from the blood into muscle.

Previous studies of androgen hormone effects in mice during gene therapy with AAV vectors focused upon differences in liver transduction between male and female mice.^8–10 Davidoff and colleagues investigated the effect of testosterone administration following oophorectomy, and demonstrated increased single- to double-strand DNA conversion after testosterone treatment.⁸ This study administered testosterone to intact female mice, and the greatest benefit was observed in conjunction with increased CI-MPR–mediated uptake of GAA. Increased CI-MPR correlated with increased GAA activity in muscle as previously described.^6,29 The increased CI-MPR expression is attributable to testosterone increasing Igf-1 expression¹³ that will increase CI-MPR expression and promote receptor-mediated uptake of GAA in muscle.^6,14

Similarly, the second study administering oxandrolone to evaluate broader androgen hormone efficacy in enhancing vector-mediated gene therapy found that oxandrolone increased the GAA activity in muscle, suggesting increased CI-MPR–mediated uptake.^3,11 GAA activity increased significantly in each tissue and muscle group that received vector or a combination of vector and a late dose of oxandrolone.

Glycogen content decreased significantly in the diaphragm and soleus in female mice treated with vector and testosterone. Glycogen content was unchanged despite increased GAA activity after testosterone treatment during gene therapy in the quadriceps, gastrocnemius, and EDL, indicative of resistance of skeletal muscle and muscle comprised mainly of type II myofibers to biochemical correction, as often observed with Pompe disease.^16,17

This difference in treatment efficacy depending on muscle type may be attributed to a lower concentration of endocytotic proteins and lysosomal enzymes, including lower expression of CI-MPR and AP-2, and increased autophagy in type II myofibers.¹⁷ All liver and muscle groups in female mice receiving vector or a combination of vector and oxandrolone decreased significantly in glycogen content, except for the gastrocnemius. Conversely, in male mice receiving vector and oxandrolone, the androgen drug in combination with gene therapy produced significantly increased glycogen content in most muscle groups, not inversely correlating with observed increases in GAA activity as expected.

It can be concluded that oxandrolone treatment does not effectively lower glycogen content, despite increasing GAA activity in muscle during gene therapy in Pompe disease. However, the observed beneficial effects of androgen hormone treatment in concert with vector-mediated gene therapy warrant further examination of androgens other than oxandrolone.

The potential efficacy from androgen hormone treatment might justify the administration of a drug with androgenic effects and acceptable toxicity profile to females in the context of gene therapy. Limitations of this study include the relatively small group size, which might have prevented detection of treatment effects in some muscles owing to inadequate statistical power. Larger studies of androgen hormones to replicate potential benefits during gene therapy are warranted. Similarly, the enhanced receptor-mediated uptake of GAA could be beneficial during ERT, if a transient androgen treatment could be safely administered at the outset or after stabilization of benefits from ERT that usually occur after 1 year of treatment.³⁰

Women have been treated with androgen hormones during assisted reproduction³¹ and for decreased libido³² without significant adverse events. Thus, testosterone or a similar androgen could be administered to enhance the effect of GAA replacement to improve efficacy from gene therapy or ERT.

AUTHOR DISCLOSURE

No competing financial interests exist.

FUNDING INFORMATION

This study was supported by the NIH Grant No. R01AR065873 from the National Institute of Arthritis and Musculoskeletal and Skin Disorders, and by Genethon.

SUPPLEMENTARY MATERIAL

Supplementary Figure S1

Supplementary Figure S2