A synthetic platform for developing recombinant adeno-associated virus type 8 producer cell lines

Abstract

Recombinant adeno-associated virus (rAAV) is one of the most widely used viral vectors for gene therapy. It is used in very high doses for the treatment of many diseases, making large-scale production for clinical applications challenging. We have established a synthetic biology-based platform to construct stable production cell lines, which can be induced to produce rAAV2. In this study, we extended our cell line construction pipelines for rAAV2 to rAAV8, a serotype whose tropism makes it attractive for gene delivery in multiple tissues. The Genome Module, encoding the rAAV2 genome, and Replication Modules, containing Rep68, DBP and E4orf6 coding sequences, originally used for rAAV2 were retained, but the Packaging Module was modified to replace the AAV2 intron-less cap gene (VP123) with that of AAV8. These three genetic modules were integrated into HEK293 genome to generate four rAAV8 producer cell lines VH1-4, which all produced rAAV8 upon induction. Their productivity was similar to the initial rAAV2 producer cell lines GX2/6 constructed using the same pipeline, but was much lower than conventional triple plasmid transfection. We identified Cap protein production and capsid formation as a potential limiting factor, just as we observed in GX2/6. By integrating more copies of AAV8 VP123 into VH3 clone, the encapsidated rAAV8 titer increased 20-fold to a level comparable to triple transfection. By tuning induction conditions to modulate capsid production, the full particle content could be elevated. This study demonstrated that our rAAV producer cell line development platform is robust and applicable to different AAV serotypes.

Keywords: HEK293; adeno‐associated virus; biomanufacturing; gene therapy; synthetic biology.

FIGURE 1

Schematic diagram of genetic modules used for (a) rAAV8 producer cell line construction and (b) enhancing VP protein and capsid productivity, and (c) rAAV8 production cell line development. (a) Genome module (GM) contains a cargo gene GFP coding sequence (CDS) under the control of inducible LacSwitch promoter flanked by AAV2 ITRs with a lacI repressor gene linked to a puromycin resistance gene driven by a phosphoglycerate kinase promoter (P_PGK). Replication module (RM) contains an inducible TetOn promoter‐driven adenoviral helper E4orf6 and a DNA binding protein (DBP) CDS, and a destabilization domain (DD) and mCherry‐tagged AAV2 Rep68 CDS with a reverse Tet transactivator‐encoding gene rtTA3 linked to a hygromycin resistance gene driven by a P_PGK. rAAV8 Packaging Modules contain an inducible CumateSwitch promoter‐driven native AAV8 cap gene (PM8‐A) or intron‐less AAV8 cap gene (VP123) with an inefficient ACG start codon for VP1 (PM8‐B) linked to a Rep52 CDS via an IRES. In addition, the Rep52 CDS is linked through P2A to a small ultra‐red fluorescent protein (smURFP) CDS. Both PM8‐A and PM8‐B also contain a Cym repressor‐encoding gene CymR linked to a blasticidin resistance gene driven by a P_PGK. Each genetic module was cloned into a transposon vector and flanked by the transposable element ITR 1 (ITR_TE1), which was recognized by Transposase 1. (b) rAAV8 Capsid Modules CM8‐A and CM8‐B for sequential transfection of VP123 and Rep52 linked by IRES (CM8‐A) or VP123 alone (CM8‐B). Both contain a TagBFP CDS encoding a blue fluorescent protein for FACS. Each of these two modules was cloned into a transposon vector and flanked by the transposable element ITR 2 (ITR_TE2), which was recognized by Transposase 2. (c) To construct rAAV8 producer cell lines, GM, RM, PM8‐B, and Transposase mRNA 1 were transfected into HEK293 cells and the three transposon vectors were integrated into the host cell genome. Following antibiotic selection, the VH cell pool was obtained. Single‐cell cloning and rAAV8 productivity screening were performed and four clonal cell lines, VH1 to 4, were isolated. VH1 and VH3 cells were chosen for sequential transfection with CM8‐A or CM8‐B. After rAAV8 productivity assessment on four resulting cell pools, the VH3B pool was chosen for further BFP positive single cell sorting. The rAAV8 productivity of the single‐cell clones was evaluated with RM4 assay cells. VH3B1 and VH3B2 cell lines were established. These figures were created with BioRender.com.

FIGURE 2

Selection of packaging modules for rAAV8 and titer of the resulting rAAV8 producer cell lines. (a) Total VG and capsid of produced by HEK293 cells transiently transfected with GM, RM and PM8‐A or PM8‐B and induced with 10 μg/mL doxycycline and 90 μg/mL cumate (10D90C) for 72 h. The total VG and capsid titer included rAAV8 from both cell lysate and culture supernatant. (b) The total VG titer (left y‐axis) and percentage of total VG secreted to culture supernatant (right y‐axis) of VH cell pool, VH1‐4 clonal cell lines induced with 10D90C and TriX. All the samples were harvested at 72 h post induction or transfection. The data are presented as the means ± SDs (n = 3).

FIGURE 3

Time dynamics of viral proteins, capsid, and total intracellular AAV genome in the production of rAAV8. VH1‐4 clonal cell lines were induced with 10D90C. The producer cell lines and TriX are as labeled on the graphs. Cells were harvested at various time points for absolute protein quantification of (a) Rep78/68, (b) DBP, (c) AAV8_VP1, (d) AAV8_VP1/2/3 (sum of VP proteins) using targeted quantitative proteomics. The total capsid and total intracellular AAV genome are in (e). The data are presented as the mean ± SDs (n = 3). ns: no significant different (p > 0.05, Student's t‐test).

FIGURE 4

Comparable rAAV8 productivity upon boosting VP production. (a) The total VG titer of VH1, VH3 and their derived CM8‐A and CM8‐B‐integrated cell pools VH1A, VH1B, VH3A, and VH3B. (b) The integrated copy number of GM, RM and sum of PM and CM in parent VH3 and two derived clones, VH3B1 and VH3B2. The integrated copy number of each gene module was measured using qPCR by targeting GFP CDS, Rep68 CDS, and AAV8 VP123, respectively, and in reference to host cell GRP15 gene. (c) The total VG and capsid titer of TriX, VH3, VH3B1, and VH3B2. (d) The total VG titer of VH3B1 and VH3B2 before and after 19 cell doublings. *p < 0.05, **p < 0.01, ns: no significant different (p > 0.05, Student's t‐test).

FIGURE 5

Altered induction dynamics of capsid production affected full particle content. For both VH3B1 and VH3B2 cell lines, the induction concentration of cumate was varied and its induction time was delayed by different periods, while both doxycycline (denoted as d) and cumate (denoted as c) induction were terminated at 72 h. The total VG titers of both cell lines with different induction conditions were shown in (a) and (b), respectively. The total VG titer, total capsid titer and full particle content (%full) of both cell lines induced with 10D90C, 10D30C, and 10D15C were shown in (c) and (d), respectively. ns: no significant different (p > 0.05, Student's t‐test).

FIGURE 6

Expression of AAV proteins in TriX, VH3, VH3B1, and VH3B2 cell lines. (a, e): rAAV8 triple plasmid transient transfection of HEK293 cells (TriX); (b, f): VH3 parent cell line; (c, g): VH3B1 cell line; and (d, h): VH3B2 cell line. The data are presented as the means ± SDs (n = 3).