Modular adeno-associated virus (rAAV) vectors used for cellular virus-directed enzyme prodrug therapy

Abstract

The pre-clinical and clinical development of viral vehicles for gene transfer increased in recent years, and a recombinant adeno-associated virus (rAAV) drug took center stage upon approval in the European Union. However, lack of standardization, inefficient purification methods and complicated retargeting limit general usability. We address these obstacles by fusing rAAV-2 capsids with two modular targeting molecules (DARPin or Affibody) specific for a cancer cell-surface marker (EGFR) while simultaneously including an affinity tag (His-tag) in a surface-exposed loop. Equipping these particles with genes coding for prodrug converting enzymes (thymidine kinase or cytosine deaminase) we demonstrate tumor marker specific transduction and prodrug-dependent apoptosis of cancer cells. Coding terminal and loop modifications in one gene enabled specific and scalable purification. Our genetic parts for viral production adhere to a standardized cloning strategy facilitating rapid prototyping of virus directed enzyme prodrug therapy (VDEPT).

Figure 1. Idempotent genetic assembly strategy.

(a) Schematic representation of the restriction enzyme recognition patterns that build up the Prefix and Suffix, respectively. Flanking every BioBrick, these two elements represent the fundamental structure of the idempotent cloning strategy. Note that NgoMIV and AgeI are optional. (b) Genetic elements constructed using the standard BioBrick assembly for the expression of modified VP2 fusion proteins. (c) Plasmids used for production cell line transfection. Names are listed next to the graphical representation of genetic elements. Modified virus-like particles (right) are formed upon combined expression of all elements, see text for description.

Figure 2. Western blots of viral capsids equipped with targeting modules, concentrated by centrifugal ultrafiltration.

Viral particles were separated on a 7.5%–15% SDS gel and blotted onto a PVDF membrane. Particles with deleted HSPG motive served as controls (lanes A_1, B_1) using B1 antibody. (a) Western blot analysis of AAV-2_Affibody. Capsid proteins were detected using B1 antibody (lane A_2). VP2-Affibody was specifically detected via anti-Flag antibody (lane A_3) (b) Western blot analysis of AAV-2_DARPin. Capsid proteins were detected using B1 antibody (lane B_2). VP2-DARPin was specifically detected via anti-Flag antibody (lane B_3).

Figure 3. Viral particle concentration, depletion of unassembled capsid proteins, affinity purification and surface display of targeting modules.

(a) Four days post transfection the supernatant from HEK-293 producer cells was harvested (supernatant) and concentrated via centrifugal ultrafiltration (concentrate, filtrate). The number of packed viral particles was determined via qPCR (n = 3, mean ± SD). (b) Samples from viral particle producing HEK-293 cells (supernatant), ultrafiltration filtrate and concentrate and fresh cell medium (neg. control) were incubated on an ELISA plate. Assembled particles or individual capsid proteins were detected using the A20 or the B1 antibody, followed by incubation with HRP-coupled anti-mouse antibody (n = 3, mean ± SD) (c) Ultrafiltration concentrated viral particles were incubated with Ni-NTA resin washed and capsids were eluted with PBS containing 500 mM imidazole. Subsequently, 4 × 10⁹ purified and non-purified particles were incubated in an ELISA plate, coated with anti-His antibody. Viral particles were detected using the capsid specific A20 antibody, followed by HRP-coupled anti-mouse antibody (n = 4, mean ± SD) (d) 1.8 × 10⁹ His-affinity purified AAV-2_Affibody^Flag and AAV-2_DARPin^Flag particles were incubated on an ELISA plate coated with anti-Flag antibody. Assembled capsids were detected using the A20 antibody, followed by biotin-coupled isotype specific IgG3 antibody and streptavidin-coupled horseradish peroxidase (n = 3, mean ± SD).

Figure 4. Binding and infection specificity of rAAV-2 particles for different cell types.

(a) 1.8 × 10⁹ viral particles were incubated on fixed 2.0 × 10³ A431 or HeLa cells, respectively. Unbound particles were removed by washing and intact capsids were detected via A20 antibody. All values are expressed relative to binding of AAV-2^ΔHSPG to HeLa cells, which was set to 100% (n = 5, mean ± SD). (b) 10⁴ A431, HeLa, or MCF7 cells were incubated with 3 × 10⁸ genomic viral particles for 24 h. Subsequently, the cells were harvested, digested with proteinase K and incubated with S1 nuclease. Double-stranded viral DNA was then quantified using qPCR (n = 3, mean ± SD).

Figure 5. Cell viability assay.

10⁴ A431 or HeLa cells were transduced with 3 × 10⁸ genomic viral particles. After 48 h fresh medium containing 5-FC was added to the cells. On day six an MTT assay was carried out and the amount of reduced formazan quantified. Values were calulated relative to the cells-only signal and then the latter was substracted to reflect the decrease in viability.

Figure 6. Apoptosis induction in A431 relative to HeLa cells.

2 × 10³ A431 or HeLa cells, respectively, were transduced with 3 × 10⁸ genomic viral particles on day one. On day three, the culture medium was exchanged and prodrug solutions were added. On day six, caspase-3 and caspase-7 activity was measured using the Apo-ONE Homogeneous Caspase-3/7 Assay (Promega). Samples were transferred to a 96 well microplate, and fluorescence was measured using a microplate reader. The caspase activity ratio of A431 to HeLa cells for each prodrug concentration is given relative to the respective AAV-2^ΔHSPG ratio (n = 3, mean ± SD). As labeled, panels (a) to (d) show experiments with combinations of Affibody or DARPin and CD (5-FC) or mGMK-TK30 (ganciclovir), respectively. Note that the scale in panel c differs from the other panels.

Figure 7. Effect of rAAV-2 infection on apoptosis induction in mixed cell cultures.

20 × 10³ A431 cells were mixed with 80 × 10³ HeLa cells or 80 × 10³ MCF7 cells, and the mixture was transduced with either His-tag affinity purified AAV-2_Affibody, AAV-2_DARPin or AAV-2^ΔHSPG capsids delivering the cytosine deaminase. Cells only served as a negative control and A431 cells incubated with 20 μM camptothecin for 4 h served as a positive control for apoptotic cells. After addition of the viral particles, cells were incubated in DMEM containing 250 μM 5-FC. 48 h later, cells were detached, washed and incubated with an anti-EGFR-Alexa488 antibody on ice for 1 h. An anti-Annexin-V-PE antibody was added and cells were analyzed by flow cytometry. For every sample 30,000 events were recorded. Dead cells were excluded using initial gating based on forward scatter height versus sideward scatter height. The anti-Annexin-V fluorescence is plotted against the EGFR fluorescence and the percentages of cells in each quadrant are given below each plot. Experiments show that EGFR expression of A431 cells decreases upon apoptosis (right panel). rAAV without targeting does not mediate prodrug activated apoptosis (middle panels). Targeted AAV primarily induces EGFR dependent apoptosis, but a bystander effect might harm nearby cells.