Molecular Engineering of Adeno-Associated Virus Capsid Improves Its Therapeutic Gene Transfer in Murine Models of Hemophilia and Retinal Degeneration

Abstract

Recombinant adeno-associated virus (AAV)-based gene therapy has been promising, but several host-related transduction or immune challenges remain. For this mode of therapy to be widely applicable, it is crucial to develop high transduction and permeating vectors that infect the target at significantly low doses. Because glycosylation of capsid proteins is known to be rate limiting in the life cycle of many viruses, we reasoned that perturbation of glycosylation sites in AAV2 capsid will enhance gene delivery. In our first set experiments, pharmacological modulation of the glycosylation status in host cells, modestly decreased (1-fold) AAV2 packaging efficacy while it improved their gene expression (∼74%) in vitro. We then generated 24 mutant AAV2 vectors modified to potentially create or disrupt a glycosylation site in its capsid. Three of them demonstrated a 1.3-2.5-fold increase in transgene expression in multiple cell lines (HeLa, Huh7, and ARPE-19). Hepatic gene transfer of these vectors in hemophilia B mice, resulted in a 2-fold increase in human coagulation factor (F)IX levels, while its T/B-cell immunogenic response was unaltered. Subsequently, intravitreal gene transfer of glycosylation site-modified vectors in C57BL6/J mice demonstrated an increase in green fluorescence protein expression (∼2- to 4-fold) and enhanced permeation across retina. Subretinal administration of these modified vectors containing RPE65 gene further rescued the photoreceptor response in a murine model of Leber congenital amarousis. Our studies highlight the translational potential of glycosylation site-modified AAV2 vectors for hepatic and ocular gene therapy applications.

Keywords: AAV2; Leber congenital amaurosis; glycosylation; hemophilia B.

Figure 1

In vitro transduction efficiency of AAV2 vectors in HeLa, Huh7, and ARPE-19 cells in the presence of glycosylation modulators. (A) Cells were pretreated with N-glycosylation inhibitors—tunicamycin and swainsonine or (B) O-glycosylation inhibitors—benzyl-α-GalNAc and alloxane (C) glycosylation enhancer—all-trans retinoic acid and infected with scAAV2-EGFP at 1 × 10³ vgs/cell. The transduction efficiency of these conditions was compared to scAAV2 alone treated cells. N = 6 replicates, *p < 0.05, **p < 0.01, ***p < 0.001 vs scAAV2-WT-treated cells.

Figure 2

Transduction efficiency of AAV2 glycosylation site-modified vectors in human cells in vitro. HeLa cells were either mock-infected or infected with 5 × 10³ vgs/cell of AAV2-WT or AAV2 mutants (A) and cells analyzed for EGFP expression 48 h later by flow cytometry. The percentage EGFP positive cells post-transduction are shown. Similar experiments were carried out in Huh7 (B) or ARPE-19 cells (C). The data depicted are the mean of two independent experiments (n = 6). *p < 0.05, ***p < 0.001 vs AAV2-WT infected cells.

Figure 3

In vitro neutralization assay with AAV2 mutants in the presence or absence of IVIG. scAAV2-WT or mutant vectors were pre-incubated with IVIG at a dilution of 1:256 and assessed for their transduction in HeLa cells, 48 h later. The data is the representation of one independent experiment with three replicates. Inner subset image depicts the percent inhibition of transduction by scAAV2-WT and AAV2 N705Q vectors in the presence of IVIG. Data depicted in this panel is a representation of four independent experiments (n = 12). *p < 0.05 vs AAV2-WT infected cells without pretreatment with IVIG.

Figure 4

Glycosylation site-modified AAV2 vector demonstrate improved hepatic gene transfer of factor (F)IX in a murine model of hemophilia B. Groups of hemophilia B mice were administered with either PBS (mock, n = 5) or scAAV2-LP1-hFIX (n = 5) or scAAV2-T14N-LP1-hFIX (n = 5) vectors at a dose of 5 × 10¹⁰ vgs via tail vein. (A) Plasma levels of human FIX were measured at 4, 10, and 12 weeks after hepatic gene therapy. *p < 0.05 between AAV2-WT vs AAV2-T14N-injected mice at 4 and 12 weeks. (B) Representative images of immunohistochemistry for human FIX expression in scAAV2-LP1 hFIX and scAAV2-T14N-LP1-hFIX administered hemophilia B mice. Arrow marks denote the nonspecific signal seen in the immunostained section. Magnification is 400×.

Figure 5

Evaluation of immune response in hemophilia B mice administered with AAV2-WT or AAV2-T14N mutant encoding human factor (F)IX transgene. Blood samples collected at 12 weeks after hepatic gene transfer was assessed for T-cell and B-cell markers by flow cytometry. Data for CD3+ T lymphocytes (A); cytotoxic T cells (CD3+CD8+) (B); and helper T cells (CD3+CD4+) (C); B-lymphocytes (E) from peripheral blood and regulatory T cells (CD4+CD25+ & FoxP3) in splenocytes (D) from mock-injected or AAV2-injected animals (n = 5) are shown. (F) Number of spots generated by 1 × 10⁶ splenocyte cells stimulated by the AAV2 capsid-specific peptide in the ELISPOT assay. (G) Representative images of the treatment groups, positive controls [concanavalin (Con) A], reagent control (without cells), cells without stimuli as blank controls, in ELISPOT plate are provided below. Data are shown as the mean ± standard deviation of the number of spots obtained from splenocytes seeded in duplicate wells for each of the five mice per group. P values were not significant (p > 0.05) vs WT-AAV2 injected hemophilia B mice.

Figure 6

Ocular gene transfer efficiency of mutant AAV2 vectors. About 3 × 10⁸ vgs of either scAAV2-WT or AAV2 mutants were administered intravitreally into the eyes (n = 6) of C57BL6/J mice. Four and six weeks after ocular gene transfer, the gene expression was measured by fundus imaging in a Micron IV imaging system. Representative data of (A) 4 weeks after ocular gene transfer from scAAV2-EGFP, scAAV2-T14N, scAAV2-Q259N, and scAAV2-N705Q vector injected eyes are shown. The quantitative data of (A) are represented in (B), after image J analysis (n = 4). The fold difference in mean GFP intensity is provided in comparison to scAAV2-WT at 4 and 6 weeks. **p < 0.01 vs WT-AAV2-injected mice. (C) Immunohistochemical analysis of retina from C57BL6/J mice injected with AAV2 vectors. Retinal sections were probed by immunohistochemistry at 16 weeks after intravitreal administration of either PBS or scAAV2-WT or scAAV2-glyco-engineered vectors containing EGFP. GCL, ganglion cell layer; ONL, outer nuclear layer; INL, inner nuclear layer; OS, outer segment; RPE, and retinal-pigmented epithelium. Magnification is 200×.

Figure 7

Visual function rescue in rd12 mice by AAV2 T14N mutant vector. (A) Representative images of the rescue in ERG wave forms in AAV2-T14N-injected eyes when compared to wild type vector injected and mock controls in rd12 mice after 6 weeks and (B) 10 weeks. A rescue in physiological vision as represented by the regain in qualitative wave form was noted. (C) Dot plot for “a wave” and “b wave” plotted against the mean amplitude obtained at 3.1 log cd s/m² shows significant rescue in “a wave” form (left graph) and “b wave” (right graph) in the mutant vector injected group when compared to wild type vector injected and mock controls at 6 weeks and (D) 10 weeks post gene transfer. n = 4−6 eyes. Values represented are mean ± standard deviation. **p < 0.01, ***p < 0.001.