Vascular delivery of rAAVrh74.MCK.GALGT2 to the gastrocnemius muscle of the rhesus macaque stimulates the expression of dystrophin and laminin α2 surrogates

Abstract

Overexpression of GALGT2 in skeletal muscle can stimulate the glycosylation of α dystroglycan and the upregulation of normally synaptic dystroglycan-binding proteins, some of which are dystrophin and laminin α2 surrogates known to be therapeutic for several forms of muscular dystrophy. This article describes the vascular delivery of GALGT2 gene therapy in a large animal model, the rhesus macaque. Recombinant adeno-associated virus, rhesus serotype 74 (rAAVrh74), was used to deliver GALGT2 via the femoral artery to the gastrocnemius muscle using an isolated focal limb perfusion method. GALGT2 expression averaged 44 ± 4% of myofibers after treatment in macaques with low preexisting anti-rAAVrh74 serum antibodies, and expression was reduced to 9 ± 4% of myofibers in macaques with high preexisting rAAVrh74 immunity (P < 0.001; n = 12 per group). This was the case regardless of the addition of immunosuppressants, including prednisolone, tacrolimus, and mycophenolate mofetil. GALGT2-treated macaque muscles showed increased glycosylation of α dystroglycan and increased expression of dystrophin and laminin α2 surrogate proteins, including utrophin, plectin1, agrin, and laminin α5. These experiments demonstrate successful transduction of rhesus macaque muscle with rAAVrh74.MCK.GALGT2 after vascular delivery and induction of molecular changes thought to be therapeutic in several forms of muscular dystrophy.

Figures

Figure 1

Expression of GALGT2 expression by cytotoxic T-cell (CT) carbohydrate immunostaining in rhesus macaque skeletal muscle. One tibialis anterior (TA) was injected with 5 × 1012 vg rAAVrh74.MCK.GALGT2, whereas the TA in the contralateral limb (control) was mock-injected with vehicle alone. After 8 weeks, a section of each muscle was stained with CT2 to determine transgene expression. Bar = 200 µm (lower panels) and 500 µm (upper panels). rAAVrh74, recombinant adeno-associated virus, rhesus serotype 74.

Figure 2

Heat map of cytotoxic T-cell (CT) glycan expression after isolated focal limb perfusion delivery of rAAVrh74.MCK.GALGT2 to the gastrocnemius muscle. The average percent muscle expression is shown for each block taken from lateral head of a gastrocnemius muscle treated with 2 × 1012 vg/kg rAAVrh74.MCK.GALGT2 12 weeks previously. Staining examples are from one of multiple sections imaged and used for quantification of each block. rAAVrh74, recombinant adeno-associated virus, rhesus serotype 74.

Figure 3

Effect of preexisting anti-rAAVrh74 immunity on muscle GALGT2 transgene expression. All macaques were treated with 2 × 1012 vg/kg rAAVrh74.MCK.GALGT2 via the isolated focal limb perfusion procedure. (a) The average of 12- and 24-week data from serum-naive macaques is compared with macaques with preexisting serum antibodies to rAAVrh74 (preimmune, defined as rAAV antibody enzyme-linked immunosorbent assay (ELISA) at 1:800>1.0). Open bars represent all animals not treated with immune suppression (non-IS). Errors for non-IS cohorts are SEM for n = 6 macaques per condition, *P < 0.05. Dark bars represent all cohorts, including cohorts treated with immune suppression. Errors for all cohorts are SEM for n = 12 macaques per condition, ***P < 0.001. (b) Representative image of positive CT2 immunostaining in rAAVrh74.MCK.GALGT2-treated gastrocnemius muscles and of hematoxylin and eosin (H&E) staining. Inset in left panel shows staining with secondary antibody only. Bar is 200 µm (left panel) or 100 µm (right panel). (c) Average myofiber diameter (mini-Feret) between treated and contralateral untreated muscles was unchanged at 12 and at 24 weeks posttreatment. Errors are SD for n = 3 per condition. (d) rAAVrh74 antibody titers (1:800 dilution ELISA) in naive and preimmune macaques after treatment at time 0 with rAAVrh74.MCK.GALGT2. Errors are SD for n = 3 per condition. rAAVrh74, recombinant adeno-associated virus, rhesus serotype 74.

Figure 4

Recombinant adeno-associated virus (rAAV) biodistribution and GALGT2 gene expression after vascular delivery of rAAVrh74.MCK.GALGT2. Serum-naive rhesus macaques were treated for 12 or 24 weeks with rAAVrh74.MCK.GALGT2. (a) Biodistribution of rAAV vector genomes (vg) was quantified in treated and control tissues per µg of genomic DNA. (b) GALGT2 transcripts arising from rAAVhr74.MCK.GALGT2 treatment were assayed by quantitative real-time-polymerase chain reaction. Errors are SEM for n = 9 measurements per condition in a and SD for n = 3 measurements per condition in b. LN, lymph node; rAAVrh74, rAAV, rhesus serotype 74.

Figure 5

Effect of immune suppression on muscle GALGT2 transgene expression. Rhesus macaques were treated with 2 × 1012 vg/kg rAAVrh74.MCK.GALGT2 via isolated focal limb perfusion. (a) Summary graph of percentage of muscle fibers overexpressing GALGT2 in cohorts (of three macaques each) with low (Time 0 antibody (Ab) not >1 for serum enzyme-linked immunosorbent assay at 1:800 dilution) or high (Time 0 Ab>1 using the same assay) levels of preexisting rAAVrh74 serum antibodies in the presence and absence of immune suppression with prednisolone (Pred) or Pred with tacrolimus (Tac) and mycophenolate mofetil (MMF). Errors are SD for n = 3 animals per condition. (b) Summary box graph of serum rAAVrh74 antibody levels for each cohort represented in (a) measured on the day of vector administration, before treatment. The gray line represents the cohort mean and the box ends represent the 5 and 95% confidence intervals for the cohort. rAAVrh74, recombinant adeno-associated virus, rhesus serotype 74.

Figure 6

Enzyme-linked immunosorbent spot (ELISpot) assay for detection of rAAVrh74 capsid- or GALGT2-specific T cells in peripheral blood mononuclear cells (PBMCs). (a) Peptides comprising the entire AAVrh74 capsid protein were separated into three pools. There were mild transient responses to peptide pools in a few macaques that exceeded 50 spot-forming colonies (SFCs)/106 PBMCs. (b) Peptides comprising the entire GALGT2 protein were separated into two pools. There were significant responses in two macaques to GALGT2 peptide pools, particularly at 20 weeks. (c) Comparison of two ELISpot-positive GALGT2 pools identified a single responsive peptide with two human–rhesus amino acid differences, one of which, threonine (rhesus) to methionine (human), was necessary for positive response to the human P5 GALGT2 peptide. (d) ELISpot analysis of GALGT2 peptides maps the optimal N- and C-terminal ends for presentation of the ELISpot-responsive P5 peptide as glycine and phenylalanine, respectively. IFN-γ, interferon-γ; rAAVrh74, recombinant adeno-associated virus, rhesus serotype 74.

Figure 7

Glycosylation of α dystroglycan after rAAVrh74.MCK.GALGT2 treatment. Different amounts of NP-40 whole-cell lysates from GALGT2-treated and untreated muscles were subjected to precipitation by Wisteria floribunda agglutinin (WFA), a βGalNAc binding lectin known to precipitate α dystroglycan after glycosylation by GALGT2, or wheat germ agglutinin (WGA), a control lectin that precipitates α dystroglycan in the absence of GALGT2 activity. Proteins were separated on 4–12% gradient gels and subjected to Western blot using antibodies to α dystroglycan (αDG), β dystroglycan (βDG), dystrophin (Dys), or utrophin (Utrn). Molecular weights shown are 160 kDa for α dystroglycan, 43 kDa for β dystroglycan, and 400–430 kDa for utrophin and dystrophin. rAAVrh74, recombinant adeno-associated virus, rhesus serotype 74.

Figure 8

Upregulation of dystrophin and laminin α2 surrogate proteins in rhesus macaque muscle after GALGT2 overexpression. Serum-naive rhesus macaques were treated with 2 × 1012 vg/kg rAAVrh74.MCK.GALGT2 via the isolated focal limb perfusion procedure for 12 or 24 weeks and compared with untreated age-matched controls. (a) Western blots of protein expression on 20 µg of protein lysate were done using antibodies to GALGT2 (70 kDa), utrophin (400 kDa), plectin1 (300–600 kDa), laminin α5 (350 kDa), agrin (200–400 kDa), laminin α2 (80–90 kDa, principal band with C-terminal antibody), dystrophin(430 kDa), α dystroglycan(160 kDa), β dystroglycan (43 kDa, 30 kDa), and GAPDH (38 kDa). (b) Semi-quantitative real-time PCR was performed on messenger RNA purified from treated and untreated muscles to measure relative expression of rhesus utrophin, dystrophin, dystroglycan, plectin1, laminin α2, agrin, or laminin α5. Rhesus GAPDH was used as an internal control. Errors are SEM for n = 6–18 measurements per condition *P < 0.05, **P < 0.01, ***P < 0.001. (c) Utrophin immunostaining shows upregulation of utrophin protein along the sarcolemmal membrane in GALGT2-treated muscles. Staining for laminin α5 similarly showed upregulation in extrasynaptic regions of the basal lamina. Bar = 100 µm. rAAVrh74, recombinant adeno-associated virus, rhesus serotype 74.