Introduction

Adeno-associated virus (AAV) based vectors are attracting substantial interest for gene therapy to correct or supplement defective genes and to deliver therapeutics, such as antibodies, that improve disease phenotypes1,2,3. Recombinant AAV (rAAV) vectors have been administered in over 300 clinical trials with three gene therapy drugs approved by the US Food and Drug Administration (FDA) in clinical use (clinicaltrials.gov, 16.02.2023)4. Efficient and scalable manufacture of high-quality rAAV to support preclinical, and clinical trials, remains a significant challenge. The AAV replicative life cycle requires the presence of a helper virus, such as adenovirus, to provide essential factors for regulating AAV gene expression and DNA replication5. rAAV vectors are missing the rep and cap genes from the AAV genome, essential for viral DNA replication and capsid formation, to make space for a transgene expression cassette. Only the viral inverted terminal repeats (ITRs) are retained in the transfer vector, palindromic sequences that confer a DNA hairpin structure required in cis for replication and encapsidation of the single-stranded viral genome6,27. This modification was introduced into TESSA-E1, and the resulting TESSA-E1-tsDNA was infected into HeLaRC32 cells alongside transfection with plasmid pAAV-EGFP (Fig. 6a). In measuring the production of TESSA-E1-tsDNA (adenovirus) genomes, as predicted significantly fewer genomes were produced at 37 °C compared to 32 °C both in the presence of doxycycline and in its absence (Fig. 6b). Intriguingly, for both regular TESSA-E1 and the temperature-sensitive variant, more genomes were made in the absence of doxycycline at both temperatures, a phenomenon seen earlier and presumably reflecting the fewer cellular resources being deployed into transcription in the absence of doxycycline, because the MLP is not activated. Against this background we were testing the hypothesis that AAV rep and cap genes would be amplified in the absence of MLP activation (i.e. without doxycycline) when adenovirus genome replication was impaired, i.e. using the TESSA-E1-tsDNA at 37 °C and not at 32 °C. Our data support this hypothesis, with the temperature-based inhibition of E2B transcription yielding at least tenfold increases in CARE-dependent amplification (Fig. 6c) and replication of rAAV-EGFP, demonstrating successful rAAV genome replication (Fig. 6d). Taken altogether these data suggest that sufficient 22/33K proteins are made to enable CARE amplification even without MLP activation, provided that adenovirus genome replication is limited.

Figure 6
figure 6

Temperature-sensitive DNA polymerase mutant TESSA-E1-tsDNA enables recovery of rep and cap amplification in HeLaRC32 cells. (a) Schematic diagram for production of rAAV2 from HeLaRC32 cells. Cells were co-infected using rAAV2-EGFP (MOI of 100 GC/cell) with TESSA-E1 or TESSA-E1-tsDNA and cultured with DMSO or doxycycline. Cells were incubated at 32 °C and 37 °C. Total DNA was extracted at 72 hpi and quantified by (b) hexon, (c) AAV2 rep, and (d) EGFP-specific qPCR. Data as mean ± SD of triple biological replicates. Analysed by one-way ANOVA followed by Tukey post hoc test. ****p ≤ 0.0001.

Full size image

Discussion

While TESSA was fully capable of providing helper functions to enable rAAV production when the rep and cap genes were provided in trans via plasmid transfection or from a TESSA vector encoding these genes15, we found that in AAV packaging HeLaRC32 cells TESSA-E1 was unable to support rep/cap DNA amplification and rAAV production when its MLP was repressed. Through a series of siRNA and complementation assays, we showed that the adenovirus intermediate-late proteins 22K and 33K were essential for CARE-dependent rep/cap DNA amplification.

As Kruger-Haag et al. showed that while the 100K-deleted adenovirus was unable to support rAAV production, an observation consistent with TESSA-E1 during MLP repression, the pTP-deleted adenovirus successfully promoted rep/cap DNA amplification despite this pTP-mutant being unable to undergo DNA replication with concomitant poor MLP activation and only limited 22/33K20. This apparently contradictory finding led us to hypothesize that, perhaps, sufficient 22/33K proteins are expressed from its internal L4P but can be titrated away from the CARE-amplification process by large amounts of replicating adenovirus genomes within the cell. Indeed, upon siRNA knockdown of the E2B transcription unit, encoding for the virus DNA polymerase and pTP, we were able to recover rep/cap DNA amplification from TESSA-E1 during MLP repression. This result confirmed our initial hypothesis and also provide a mechanistic explanation for the inability of the 100K-deleted adenoviruses to support rAAV production from stable AAV packaging cell lines.

The precise molecular role of the 22K/33K proteins currently remains uncertain, including how their coding sequences are involved in amplification of AAV rep/cap genes from stable producer cells. While we showed that expression of 22K/33K proteins in trans from expression plasmids resulted in a modest five- to tenfold increase in rep and cap DNA amplification from TESSA-E1, despite the absence of doxycycline, it did not fully recover to levels observed from the doxycycline group. Possibly, temporal expression of 22K/33K proteins are required for optimal amplification of AAV rep/cap, as their expression in situ are tightly controlled from the adenovirus18,19. Additionally, expression of the 22K or 33K from the CMV expression plasmids both resulted in AAV rep/cap amplification may suggest functional redundancies between these proteins. The L4-22K and -33K proteins, both sharing the first 105 amino acids but differing towards the C-terminus, are multifunctional and involved in regulating the expression of early and late adenoviral genes, RNA splicing and the genome encapsidation process16. While the L4-22K protein was shown to bind A repeats in the adenovirus packaging signal at a conserved TTTG motif, L4-33K interacts with the IVa2 protein that associates with the packaging domain of the Ad genome and functions as DNA packaging ATPase28,29. It was also shown by co-immunoprecipitation assays and confocal-microscopy that the L4-33K protein co-localizes with adenovirus DBP, which was reported to be essential for rep/cap DNA amplification from AAV packaging cells12,30. Possibly, physical binding of 22/33K proteins to the adenovirus genome limits their availability required for the CARE-amplification process.

To enable the use of TESSA-E1 for production of rAAV from stable rep/cap packaging cells, we introduced the temperature-sensitive DNA polymerase mutation ts36 to generate TESSA-E1-tsDNA. This helper TESSA-E1-tsDNA is effectively produced in HEK293 cells at 32 °C and supplemented with doxycycline, to enable full activity of the ts36 polymerase and whilst inhibiting activity of the TetR for full MLP functionality, respectively. While replication of TESSA-E1-tsDNA genomes was reduced by 15-fold at the non-permissive temperature of 37 °C (with doxycycline), a modest twofold reduction in TESSA-E1-tsDNA genomes was observed in the absence of doxycycline likely due to the increase efficiency of adenoviral genome replication observed from repression of the MLP15. Interestingly, at 37 °C, the temperature used for rAAV production, a modest reduction in the replication of TESSA-E1-tsDNA significantly recovered amplification of the rep/cap gene for rAAV replication in HeLaRC32 cells. We propose TESSA-E1-tsDNA as an efficient helper virus for contaminant-free manufacture of rAAV from stable rAAV packaging cell lines. Future development may explore approaches to further repress TESSA-E1-tsDNA genome replication by employing expression of shRNA, small RNAs, or MLP-TSS-sRNA shown to inhibit adenovirus, or heterologous expression of 22/33K to enhance rAAV production from stable AAV packaging cells31. In addition to recent novel approaches designed to improve rAAV production, and as a DNA delivery vector, this range of options should provide a raft of strategies for efficient scalable manufacture of all AAV serotypes using stable packaging cells, avoiding contamination issues and dramatically decreasing the associated costs32,33,34,35.

Methods

Virus generation and titration

Construction of the TESSA genome plasmid (pSU390) has been described previously (Su et al. 2022). To generate TESSA-E1 and the control Ad5-E1 plasmids, the Ad5 E1 coding sequence was amplified by PCR from plasmid encoding the wildtype Ad5 genome (T38, OXGENE) using forward primer 5’-GACGGCGATCGCTGGGGCGGCCGCCATCGATGGGTTAATTAATGTAGTGTATTTATACCCGG’-3 and reverse primer 5’-GACCCCCACCTTATATATTCTTTCCCACCCTTAATTAAGCCACGCCCACACATTTCAG TACCTC-3’ and inserted into pSU390 and OG268 plasmids linearised with PacI via Gibson DNA assembly (New England Biolabs, MA, USA), to generate TESSA-E1 (pSU708) and Ad5-E1 plasmids (pSU704), respectively. For incorporating the ts36 mutation into the DNA polymerase coding sequence of TESSA-E1, a shuttle plasmid (pSU180), which encodes a portion of the Ad5 E2B polymerase fragment, was used as a template for PCR using forward primer 5’- CTGCACGTATTCGCGCG -3’ with reverse primer 5’- CTCAGCTCCCCTGAAGAGtttACCTACGAGGAACTT -3’, and forward primer 5’- CCCACAATGTAAAaTTCCAAGAAGC-3’ with reverse primer 5’- AAGTTCCTCGTAGGTaaaCTCTTCAGGGGAGCTGAG -3’, to generate two DNA fragments incorporating the ts36 mutation. These fragments were inserted into DraIII and BsrGI (New England Biolabs, MA, USA) linearized pSU180 via Gibson DNA assembly, generating plasmid pSU1278. The ts36-E2B fragment was subsequently excised from pSU1278 using SphI and NheI (New England Biolabs, MA, USA) and inserted into pSU390 (linearized with XbaI and BstZ17I) via Gibson DNA assembly to generate pSU1294. The Ad5 E1 genes were extracted from pSU708 using AsiSI and BstZ17I (New England Biolabs, MA, USA) and inserted into PacI-linearized pSU1294 via Gibson DNA assembly, generating TESSA-E1-tsDNA (pSU1316). Positive clones were isolated and virus recovery, and infectious titration, carried out in HEK293 cells as previously described15.

Determining intact adenovirus particles by PicoGreen assay

Intact adenovirus particles in purified stocks were determined using the PicoGreen dsDNA assay (Invitrogen, UK) which uses a fluorescent nucleic acid stain specific for double-stranded DNA. Purified viral stocks were diluted in TE buffer containing 0.1% SDS, 25mM EDTA and incubated at 37 °C for 15 min to disassociate the virus capsid. Samples and the standard curve generated using bacteriophage lambda DNA were measured using the PicoGreen dsDNA assay kit according to the manufacturer’s instructions. Concentration of intact particles was calculated based upon 1 μg is equivalent to 2.7 × 1010 genome intact particles36. The ratio of intact adenoviral particles, determined by the PicoGreen assay, and the Tissue Culture Infectious Dose (TCID50) is routinely determined to be < 20:1.

Cell culture

HEK293 cells and HeLaRC32 cells were purchased from Cell Biolabs (AD-100, CA, USA) and ATCC (CRL-2972, VA, USA), respectively, and cultured in Dulbecco’s Modified Eagle Medium (DMEM; Sigma-Aldrich, MO, USA) supplemented with 10% (v/v) heat-inactivated foetal bovine serum (FBS; Gibco, MA, USA). Cell lines were maintained at 5% CO2, 37 °C, and 95% humidity.

Plasmids and siRNA transfection

siRNA transfections were performed using Lipofectamine RNAiMAX (ThermoFisher, USA) according to the manufacturer’s protocol. Reverse transfection of HEK293 cells with siRNA was carried out in 48-well plate format. For each well, 3 pmol of siRNA diluted in 30 μL of OptiMEM (Life Technologies, UK) was complexed with 1 μL of RNAiMAX suspended OptiMEM. Reaction was incubated at ambient temperature for 10 min before adding to the well. Subsequently, 9 × 104 cells per well suspended in DMEM containing 10% FBS was added to each well and the plate was transferred to a 37°C incubator. After 24 h, cells were infected with adenovirus vectors as indicated in the experiments. siRNA sequences are provided in Supplementary Table S1. All CMV expression plasmids were provided by OXGENE (Oxford, UK) and the pAAV-EGFP was described previously15. Cells were seeded at 9 × 104 cells per well for 24 h and each well was transfected with 750 ng of plasmid DNA complexed using linear PEI 25kDA (Polysciences) at a 1:3 DNA to PEI mass ratio.

Quantification of viral genomes using qPCR

For quantification of total adenovirus genomes in HeLaRC32 cells, total DNA was extracted from culture media and cellular lysates using DNeasy Blood and Tissue kit (Qiagen, Venlo, Netherlands). Five microlitres of DNA eluent were used in qPCR reactions using TaqMan Fast Advanced Master Mix (Applied Biosystems, CA, USA) in a StepOnePlus Real-Time PCR System (Applied Biosystems, CA, USA). Primer sequences for targeting Ad5 hexon are forward 5′-CACTCATATTTCTTACATGCCCACTATT-3′, reverse 5′- GGCCTGTTGGGCATAGATTG-3′ and TaqMan probe 5′-AGGAAGGTAAC TCACGAGAACTAATGGGCCA-3′. Primer sequences targeting the EGFP are forward 5′-GAACCGCATCGAGCTGAA-3’, reverse 5′-TGCTTGTCGGCCATGATATAG-3′, and TaqMan probe 5′-ATCGACTTCAAGGAGGACGGCAAC-3′. Primer sequences for targeting AAV2 rep forward 5′-GGCCTCATACATCTCCTTCAAT-3′, reverse 5′- AGTCAGGCTCATAATCTTTCCC-3′ and TaqMan probe 5′- TCCAACTCGCGGTCCCAAATCAA-3′. Primer sequences for targeting AAV2 cap are forward 5′-CGACCCAAGAGACTCAACTTC-3′, reverse 5′-GAACCGTGCTGGTAAGGTTAT-3′ and TaqMan probe 5′-AAAGAGGTCACGCAGAATGACGGT-3′. PCR cycles were as follows: 95 °C 10 min; 40 times (95 °C 1 s, 60 °C 20 s). Extraction and quantification of DNAse-resistant rAAV genomes were carried out as previously described15.

Statistical analysis

Data presented as mean ± standard deviation (SD), unless otherwise stated. Significance evaluated using one-way ANOVA followed by Tukey post hoc test, unless otherwise stated, and denoted on the graphs as *P ≤ 0.05 **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.