Site-Directed Mutagenesis Improves the Transduction Efficiency of Capsid Library-Derived Recombinant AAV Vectors

Abstract

Recombinant adeno-associated virus (rAAV) vectors selected from capsid libraries present enormous advantages in high selectivity of tissue tropism and their potential use in human gene therapy applications. For example, rAAV-LK03, was used in a gene therapy trial for hemophilia A (ClinicalTrials.gov: NCT03003533). However, high doses in patients resulted in severe adverse events and subsequent loss of factor VIII (FVIII) expression. Thus, additional strategies are needed to enhance the transduction efficiency of capsid library-derived rAAV vectors such that improved clinical efficacy can be achieved at low vector doses. In this study, we characterized two commonly used library-derived rAAV vectors, rAAV-DJ and rAAV-LK03. It was concluded that rAAV-DJ shared similar transport pathways (e.g., cell surface binding, endocytosis-dependent internalization, and cytoplasmic trafficking) with rAAV serotype 2, while rAAV-LK03 and rAAV serotype 3 shared similar transport pathways. We then performed site-directed mutagenesis of surface-exposed tyrosine (Y), serine (S), aspartic acid (D), and tryptophan (W) residues on rAAV-DJ and rAAV-LK03 capsids. Our results demonstrated that rAAV-DJ-S269T and rAAV-LK03-Y705+731F variants had significantly enhanced transduction efficiency compared to wild-type counterparts. Our studies suggest that the strategy of site-directed mutagenesis should be applicable to other non-natural AAV variants for their optimal use in human gene therapy.

Keywords: cancer targeting; gene therapy; hepatocellular carcinoma; library selection; rAAV vector; site-directed mutagenesis; transduction efficiency; vector distribution.

© 2020 The Authors.

Figures

Figure 1

Transduction Efficiency of rAAV2-CMV-egfp and rAAV-DJ-CMV-egfp Vectors in HEK293 Cells, Treated by Various Specific Pharmacological Inhibitors (A) HEK293 cells were infected with GFP-expressing rAAV2 or rAAV-DJ in the presence of specific inhibitors (100 μg/mL heparin, 2 μM jasplakinolide, 10 μM EIPA, 200 nM bafilomycin A1, 10 μg/mL brefeldin A, and 20 μM MG132). GFP expression was measured at 24 h post-infection by fluorescence microscope. Scale bars represent 300 μm. (B) Quantitation of transduction efficiency of rAAV2 and rAAV-DJ with different inhibitors. Data were normalized by rAAV2- and rAAV-DJ-infected and mock-treated groups, respectively. (C) Binding and internalization assays were performed using vector-transduced HEK293 cells with or without heparin-treatment. All values shown are means ± standard deviations. ∗p 

Figure 2

Transduction Efficiency of Site-Directed rAAV-DJ Capsid Mutants in Human Cell Lines (A) +++, significant increase; ++, no differences; +, significant reduction; −, complete loss of transgene expression; versus WT. (B) HEK293, Huh7, and HepG2 cell lines were transduced with the WT, S269T, Y446F, or S269T+D496E mutant rAAV-DJ-CMV-egfp vectors at an MOI of 5,000 vg/cell. Transgene expression was analyzed by fluorescence microscopy 72 h post-transduction. (C) Quantitation of transduction efficiency of rAAV-DJ mutants in cell lines. All values shown are means ± standard deviations. ∗p < 0.05, ∗∗p < 0.01 versus WT.

Figure 3

Transduction Efficiency of rAAV-DJ-WT and rAAV-DJ-S269T Vectors In Vivo Mice were tail-vein injected with the indicated vectors containing a CMV-egfp gene at 1 × 1011 vg/mouse (n = 3). Major tissues were obtained at 4 weeks post-injection. (A) rAAV-DJ-S269T mediated higher egfp expression, detected by fluorescence microscopy of liver sections. Quantitative data are presented. Scale bars represent 200 μm. (B) Immunofluorescence staining of liver sections at ×400 original magnification. The merged images showed that both vectors preferentially transduce hepatocytes, but not CK19+ cholangiocytes. Green indicates EGFP; red indicates CK19; blue indicates DAPI staining for nucleus. Scale bars represent 50 μm. (C) Tissue DNA was isolated and qPCR was performed to measure the AAV genome copies per 100 ng of DNA. All values shown are means ± standard deviations. ∗∗p < 0.01 versus WT.

Figure 4

Transduction Efficiency of rAAV2-CMV-egfp and rAAV-LK03-CMV-egfp Vectors in Huh7 Cells, Treated by Various Specific Pharmacological Inhibitors (A) Huh7 cells were infected with GFP-expressing rAAV2 or rAAV-LK03 in the presence of specific inhibitors (100 μg/mL heparin, 2 μM jasplakinolide, 10 μM EIPA, 200 nM bafilomycin A1, 10 μg/mL brefeldin A, and 20 μM MG132). GFP expression was measured at 24 h post-infection by fluorescence microscopy. Scale bars represent 300 μm. (B) Quantitation of transduction efficiency of rAAV2 and rAAV-LK03 with different inhibitors. Data were normalized by rAAV2- and rAAV-LK03-infected and mock-treated groups, respectively. (C) Transduction efficiency of rAAV2 and rAAV-LK03 with different concentrations of heparin (0.1–100 μg/ml). All values shown are means ± standard deviations. ∗p 

Figure 5

Transduction Efficiency of rAAV3B, rAAV-LK03, and Their Capsid-Modified Counterparts in Human Hepatic Cell Lines In Vitro (A and B) Transgene expression of (A) rAAV-LK03-CBA-gluc, rAAV3B-CBA-gluc, and (B) capsid-modified rAAV3B-S663V+T492V-CBA-gluc vectors in the indicated human hepatocellular carcinoma cell lines. The GLucs were measured in the supernatant of cell cultures that were transduced with the viral vectors at an MOI of 5 × 103 vg/cell for 48 h. (C and D) Capsid structure of AAV3B (PDB: 3KIC) with the viral asymmetric unit outlined in black (bold triangle). Key residues for the improved transduction within rAAV3B and rAAV-LK03 are highlighted: T492 (yellow), S663 (orange), Y705 (blue), and Y731 (green). (C) Surface representation of the viral capsid viewed down the 2-fold (2f) axis of symmetry. From this orientation S262 is visible (red) tucked underneath the base of the 3-fold (3f) protrusions. S262 is important for species-limited transduction of human hepatocytes. (D) The spherical viral capsid projected onto a two-dimensional roadmap and viewed from the exterior capsid surface. Three out of four residues involved in proteasome trafficking cluster near the icosahedral 2f symmetry axis (T492, Y705, and Y731). However, S663 lies within the canyon that surrounds the 5-fold (5f axis). Residues implicated in receptor binding are located within 3f protrusions and are outlined in purple (R447 and R594). 2-fold, black-filled elliptoid; 3-fold, black-filled triangle; 5-fold, black-filled pentagon. (E) Transgene expression of capsid-modified rAAV-LK03-CBA-fluc vectors in Huh7 cells. The FLucs were measured in Huh7 cells that were transduced with the rAAV-LK03 mutants at an MOI of 5 × 103 vg/cell for 48 h. All values shown are means ± standard deviations. ∗p < 0.05, ∗∗p < 0.01 versus rAAV-LK03.

Figure 6

rAAV-LK03-Y705+731F Vectors Mediated Higher Transduction Efficiency in a Human HCC Xenograft Model following Intravenous Vector Injection Huh7 tumor-bearing NSG mice were tail vein injected with rAAV-LK03-CBA-fluc or of rAAV-LK03-Y705+731-CBA-fluc vectors at 1 × 1011 vg/mouse. (A) Whole-body bioluminescence images of NSG mice were obtained at 3 days post-injection. (B) Quantitative data of bioluminescent signals in mouse tumor xenografts at days 1, 3, 5, and 7 after vector administration. All values shown are means ± standard deviations. ∗p < 0.05, ∗∗p < 0.01 versus WT.