Enzymatically amplified linear dbDNATM as a rapid and scalable solution to industrial lentiviral vector manufacturing

Traditional bacterial fermentation techniques used to manufacture plasmid are time-consuming, expensive, and inherently unstable. The production of sufficient GMP grade material thus imposes a major bottleneck on industrial-scale manufacturing of lentiviral vectors (LVV). Touchlight’s linear doggybone DNA (dbDNATM) is an enzymatically amplified DNA vector produced with exceptional speed through an in vitro dual enzyme process, enabling industrial-scale manufacturing of GMP material in a fraction of the time required for plasmid. We have previously shown that dbDNATM can be used to produce functional LVV; however, obtaining high LVV titres remained a challenge. Here, we aimed to demonstrate that dbDNATM could be optimised for the manufacture of high titre LVV. We found that dbDNATM displayed a unique transfection and expression profile in the context of LVV production, which necessitated the optimisation of DNA input and construct ratios. Furthermore, we demonstrate that efficient 3’ end processing of viral genomic RNA (vgRNA) derived from linear dbDNATM transfer vectors required the addition of a strong 3’ termination signal and downstream spacer sequence to enable efficient vgRNA packaging. Using these improved vector architectures along with optimised transfection conditions, we were able to produce a CAR19h28z LVV with equivalent infectious titres as achieved using plasmid, demonstrating that dbDNATM technology can provide a highly effective solution to the plasmid bottleneck.

. Increasing vgRNA abundance does not translate to enhanced infectivity of dbDNA TM LVV Viral production cells (VPC) were transfected to produce LVV with 1 µg/mL plasmid (mass construct ratio of 2:1:1:1) or 0.5 µg/mL dbDNA at the indicated molar construct ratios, and cells were harvested 72h post transfection of analysis (a) Gene expression measured by RT-qPCR. Probe vgRNA was used to quantify full length viral genomic RNA, and probe eGFP was used for total RNA transcribed from the transfer vector LV-eGFP (b) Mean fluorescence intensity (MFI) measured by flow cytometry (c) Infectious titre of LVV produced using 1 µg/mL plasmid (mass ratio 2:1:1:1) or 0.5 µg/mL dbDNA at the indicated molar construct ratios. Experiments were performed once; error bars represent standard deviation between replicates.

Figure S3. Increasing ratio of transfer vector and VSVg leads to modest improvement in infectivity
(a) VPC were transfected with 0.5 µg/mL dbDNA at the indicated molar construct ratios. Transfer vector was increased from 0.5 to 4, and VSVg was increased from 1 -2. Standard plasmid conditions were used as a control. Supernatants were harvested at 72h post transfection for analysis of infectious titre by flow cytometry of transduced HEK293T cells (b) The ratio of transfer vector and VSVg were further increased to 10 and 4, respectively, as a final test to determine whether significantly increasing the abundance of vgRNA could improve infectivity. To maintain sufficient amounts of accessory proteins under these conditions, the total amount of input dbDNA was slightly increased at higher ratios of transfer vector, as indicated. Supernatants were harvested 72h post transfection for analysis of infectious titre. Experiments were performed once.

Figure S4. Plasmid and dbDNA transfer vectors display different expression profiles
RNA-sequencing was performed as described in the methods to evaluate the transcriptional profile of (a) pDNA and (b) dbDNA LV-eGFP transfer vectors. HEK293F cells were transfected in triplicate with 1 µg/mL pDNA or dbDNA and harvested for RNA extraction 48 h post transfection. Data normalization was done by dividing the sense and antisense counts at each nucleotide position by the total number of reads that aligned to the construct reference sequence. Grey bars represent the indicated regions of interest within the sequence. Anti-sense reads are shown in red and sense reads are shown in blue. Areas of blue shading represent variability between samples prepared in triplicate.

Figure S5. Transfer vector architectures
Shown is a schematic of the dbDNA transfer vector architectures used in this study. RS = random spacer, CMV = Cytomegalovirus promoter, LTR = long terminal repeat, WPRE = woodchuck hepatitis virus posttranscriptional regulatory element, SV40 pA = Simian virus 40 late poly(A) Figure S6. Addition of termination element and downstream spacer improves 3' end processing of vgRNA RNA-sequencing was performed as described in Figure S4 to evaluate the transcriptional profile of (a) dbDNA-LV-eGFP-pA and (b) dbDNA LV-eGFP-pA-RS1 dbDNA transfer vectors. Grey bars represent the indicated regions of interest within the sequence. Anti-sense reads are shown in red and sense reads are shown in blue. Areas of blue shading represent variability between samples prepared in triplicate.

Figure S7. Particle titres and vgRNA abundance are unchanged with addition of poly(A) and RS1kb
LVV productions were carried out using 0.5 µg/mL total dbDNA (molar ratio 4:3:3:4) and the indicated transfer vectors. Supernatants and cells were harvested 72h post transfection for analysis of total particle titre (VP/mL) and vgRNA abundance (a) Total particle titre measured by p24 ELISA (b) Transfer vector RNA abundance measured by RT-qPCR using probes vgRNA for full length viral genomic RNA and eGFP for total RNA. Experiments were performed once; error bars represent standard deviation between replicates. Figure S8. Increasing input dbDNA to 0.7 ug/mL in the context of dbDNA-LV-eGFP-pA-RS1 further boosts infectious titre VPC were transfected to produce LVV using 0.5 -1.0 µg/mL total input dbDNA TM and the dbDNA-LV-eGFP-RS1 transfer vector (a) Infectious titre (TU/mL) measured by flow cytometry of transduced HEK293 T cells 72 h post transduction (b) Cell counts 72 h post transduction. Experiments were performed once; error bars represent standard deviation between replicates.