AAV capsid engineering identified two novel variants with improved in vivo tropism for cardiomyocytes

Figures

• 

  Figure 1. AAV peptide display library screening on hypertrophic cardiomyocytes in vivo (A) Schematic illustration of the library screening. (B) NGS results illustrating the peptide inserts detectable after three rounds of selection in the CMC fraction (CMC—3×), in the non-cardiomyocyte fraction (non—CMC 3×), in liver tissue (liver—3×) and in the cardiomyocyte fraction after the first round of selection (CMC—1×). (C) Variants that accumulated after three rounds of AAV peptide display library screening on hypertrophic CMCs. Shown are sequences of peptide inserts presented with amounts of more than 1.5%. cap, gene encoding capsid proteins (VP1-3) modified by peptide insertion as indicated, assembly activating protein (AAP) and membrane associated accessory protein (MAAP); rep, gene encoding AAV Rep proteins; GFP, gene encoding the enhanced green fluorescence protein (under control of a CMV promoter), CMC, cardiomyocytes; non-CMC, all cardiac cells except cardiomyocytes.

• 

  Figure 2. In vitro characterization of novel AAV capsid variants (A and B) Transduction efficiency of AAV vector variants and the parental AAV2 on non-cardiac cells in vitro. (A) Serial dilutions of respective vector preparations were used for the transduction of HEK293 cells in a 12-well format. Genomic titers of vectors used in this assay were in a comparable range. Transduction efficiency was determined by flow cytometry for GFP-positive cells 48 h after transduction (n = 3). (B) The ratio of infectious particles (on HEK293 cells) per vector genome was calculated based on transduction efficiencies obtained in (A) and the genomic titer measured by qRT-PCR. (C) Heparin competition assay. AAV vector variants and the parental AAV2 were pre-incubated with heparin or not. Transduction efficiency on HEK293 cells was determined by FACS for GFP positive cells 48 h after transduction (n = 3). (D) Thermal stability assay. AAV vector particles were subjected to different temperatures for 15 min followed by dot blotting using the A20 antibody for detection of assembled capsid proteins (intact AAV vector particles). (E) Performance of novel variants on cardiomyocyte cell model. Human iPSC-CMCs were…

• 

  Figure 3. Vector copy number analysis (A) Indicated AAV vectors expressing GFP were injected in sham- and TAC-operated mice 6 weeks after sham and TAC surgery, respectively (note that each experimental group started with four mice each, but final group size differs owing to the loss of individual mice; n as number of mice is depicted for each group). Organs and cells were harvested 2 weeks after AAV vector injection for vector copy number and GFP expression analysis by qRT-PCR. (B) Overview of vector copy number analysis in different organs and cell type. (C and D) Comparison of the vector copy number in target cell types (cardiomyocytes [C]) and the main off-target organ (liver [D]). In (C) and (D), the mean vector copy numbers of pooled sham- and TAC-operated cohorts are shown for AAV2, AAV9, AAV2-THPTPAD, and AAV2-NLPGSGD (vectors that did not show noticeable differences between sham and TAC according to [B]). Data are means ± standard deviation. The p values were determined by one-way ANOVA with Dunnett’s multiple comparison with the current gold standard AAV9.

• 

  Figure 4. Transduction efficiency and specificity of AAV vector variants in vivo based on GFP expression AAV vectors (AAV2, AAV9, AAV2-THGTPAD, AAV2-NLPGSGD, and AAV2-LPSRPSL, expressing GFP) were injected in sham- and TAC-operated mice 6 weeks after sham or TAC surgery (note that each experimental group started with four mice each, but the final group size differs owing to the loss of individual mice; n as the number of mice is depicted for each group). GFP expression in cardiac cell types and different organs was analyzed by qRT-PCR 2 weeks after vector injection. (A) GFP expression analysis (separately for all different groups: sham or TAC with respective AAV vector variants) and (B) analysis based on pooled groups (sham + TAC mice, in case of AAV2, AAV9, AAV-THGTPAD, and AAV2-NLPGSGD—vectors that did not show noticeable differences between sham and TAC according to (B) (n = 3–7). (C) CMC to liver expression ratio normalized to AAV9 (n = 3–7). Data are means ± standard deviation. The p values were determined by one-way ANOVA with Dunnett’s multiple comparison with the current gold standard AAV9. GFP, self-complementary vector genome conformation encoding for enhanced green fluorescence…

• 

  Figure 5. Subcellular vector distribution and episome formation in heart and liver tissue (A–D) Subcellular distribution of indicated AAV vector in different cell compartments (cytosol, nucleus, membrane; measured by qPCR and displayed as GFP transgene copy number concentration or as percentage distribution upon different cell fractions of heart (A and B) and liver tissue (C and D) from sham-operated mice 14 days after administration of respective vector variants. (E and F) Percentage proportion of vector genome formed as episomes in the nucleus of heart and liver tissue (n = 3). Data are means ± standard deviation. The p values were determined by one-way ANOVA with Dunnett’s multiple comparison with the parental serotype AAV2.

• 

  Figure 6. Serum neutralization assay (A) Neutralization assay of indicated AAV vectors by pre-incubation with human IVIG before transduction of HEK293 cells. For each AAV variant, as well as for each reference, the AAV vector amount, which induced 30%–40% of GFP-positive cells in the absence of IVIG (positive control) was chosen and incubated with indicated serum dilutions. The percentage of GFP-positive cells 48 h after transduction was analyzed by flow cytometry (IC50; n = 3 independent experiments). (B) IC50 was calculated as fold change to the parental AAV serotype AAV2. Neutralization assay of AAV variant as well as for references AAV vector variants on HEK293 cells by pre-incubation with pooled serum of mice which were previously injected with (C) AAV9, (D) AAV2, (E) AAV2-NLPGSGD and (F) AAV2-THGTPAD. For each AAV vector variant as well as for reference, the AAV vector amount which induced 30–50% of GFP-positive cells in the absence of any serum (positive control = without any serum) was chosen and incubated with indicated serum dilutions. The percentage of GFP-positive cells 48 h after transduction was analyzed by FACS. (n = 3 independent experiments; Data are means ± standard…

• 

  Figure 7. AAV2-THGTPAD and AAV2-NLPGSGD mediate efficient transgene expression in the heart despite previous AAV9 vector administration (A) AAV vectors encoding for GFP were injected 2 weeks after the administration of indicated AAV vectors delivering the Crimson transgene (4 × 1010 vector genomes or PBS as control). (B and C) Overview of vector copy number analysis (B) and expression analysis (C) of AAV GFP transgene encoding vector genomes (n = 5). Data are means ± standard deviation. The p values were determined by one-way ANOVA with Dunnett’s multiple comparison to the group which previously received PBS in the first injection. Crimson vector copy number and expression analysis are displayed in Figure S1.

• 

  Figure 8. Improved gene therapeutic efficiency by AAV2-THGTPAD and AAV2-NLPGSGD delivering anti-hypertrophic lncRNA H19 (A) Schematic representation of the experimental design. (B) Representative hearts of mice from the respective treatment groups. (C–F) Echocardiographic analysis at the end point of the study. (G) Heart weight to tibia length ratios. (H–J) (H) Expression levels of H19 and (I, J) vector copy number analysis of H19 in heart and liver tissue of sham- and TAC-operated mice treated with respective AAV variants delivering the lncRNA H19 or control (AAV9-empty) (n = 10–12). Data are means ± standard deviation. The p values were determined by one-way ANOVA with Dunnett’s multiple comparison with the TAC AAV9-empty group.