Here, a suspension HEK293 cell-based AAV production protocol is presented, resulting in reduced time and labor needed for vector production using components that are available for research purposes from commercial vendors.
Adeno-associated viral vectors (AAVs) are a remarkable tool for investigating the central nervous system (CNS). Innovative capsids, such as AAV.PHP.eB, demonstrate extensive transduction of the CNS by intravenous injection in mice. To achieve comparable transduction, a 100-fold higher titer (minimally 1 x 1011 genome copies/mouse) is needed compared to direct injection in the CNS parenchyma. In our group, AAV production, including AAV.PHP.eB relies on adherent HEK293T cells and the triple transfection method. Achieving high yields of AAV with adherent cells entails a labor- and material-intensive process. This constraint prompted the development of a protocol for suspension-based cell culture in conical tubes. AAVs generated in adherent cells were compared to the suspension production method. Culture in suspension using transfection reagents Polyethylenimine or TransIt were compared. AAV vectors were purified by iodixanol gradient ultracentrifugation followed by buffer exchange and concentration using a centrifugal filter. With the adherent method, we achieved an average of 2.6 x 1012 genome copies (GC) total, whereas the suspension method and Polyethylenimine yielded 7.7 x 1012 GC in total, and TransIt yielded 2.4 x 1013 GC in total. There is no difference in in vivo transduction efficiency between vectors produced with adherent compared to the suspension cell system. In summary, a suspension HEK293 cell based AAV production protocol is introduced, resulting in a reduced amount of time and labor needed for vector production while achieving 3 to 9 times higher yields using components available from commercial vendors for research purposes.
Adeno-associated virus (AAV) was discovered in 1965 and has since been used in a myriad of applications1. AAVs have been applied in neuroscience research to study gene and neuronal function, map neurocircuits, or produce animal models for disease2. Traditionally, this is done by injecting directly at the site of interest, as most natural serotypes do not cross the blood-brain barrier or need a high dose to do so1,2,3.
With the discovery of AAV.PHP.B4 and next-generation capsids such as AAV.PHP.eB5 and AAV.CAP-B106, it is possible to target the central nervous system (CNS) using a simple systemic injection. Spatial mapping reveals the cells targeted by AAV.PHP.eB at cellular level6,7. In combination with specific promoters/enhancers, these capsids offer extensive opportunities for neuroscientists to study genes and brain function by non-invasive AAV delivery4,8.
While a lower dose is needed for AAV.PHP.eB (typically 1 to 5 x 1011 Genome Copies (GC)/mouse) compared to AAV9 (4 x 1012 GC/mouse)7, still more vector needs to be produced compared to direct injection strategies (typically 1 x 109 GC/µL injection). Most natural serotypes can be produced using the classical adherent cell culture system in combination with iodixanol purification9,10,11,12. For AAV.PHP.eB this entails a labor-intensive process to culture and transfect cells to obtain sufficient vectors for one experiment8. Therefore, the production of AAV in suspension cell culture in conical tubes was developed. Conical tubes, with a capacity of up to 300 mL, are compact, saving both incubator space and plastics. Suspension cells are much easier to culture and handle in large amounts than adherent cells on 15 cm plates. The transfection components of the protocol remain the same. Therefore, plasmids previously used with the adherent system can easily be used in this protocol based on production in suspension cells. The protocol was successfully transferred to other researchers in the laboratory and successfully used for various capsids and constructs.
All experimental procedures were approved by the institutional animal care and use committee of the Royal Netherlands Academy of Sciences (KNAW) and were in accordance with the Dutch Law on Animal Experimentation under project number AVD8010020199126. In Figure 1, a schematic overview of the complete protocol is provided. From seeding cells to AAV purification, the protocol takes 6 days to complete.
1. Reagent preparation
2. Culture of HEK293 suspension cells
3. Transfection
4. Harvesting the cells
5. Iodixanol purification
Most academic labs use adherent HEK293T cells for AAV production8,9. While this works relatively well when small amounts of AAV are needed for direct injection, a 100-fold higher titer (minimally 1 x 1011 GC/mouse) is needed to achieve similar transduction with systemic capsids such as AAV.PHP.eB.
In this protocol, the production of AAV using suspension HEK293 cells cultured in conical tubes was established. Small-scale cultures can be done in 50 mL tubes and large-scale in 600 mL tubes. For specific shakers (see Table of Materials), a rack can be used in which up to 16 large tubes can be cultured simultaneously. This rack is used in combination with special inserts for 50 mL tubes (Figure 2A). This system offers a significant gain in time for this protocol's culturing and transfection segment. The initial setup of production was done with reporter constructs for luciferase and green fluorescent protein (GFP)10,11. In parallel, the GFP construct was produced in 12, 15 cm2 plates as described in a previously published protocol12. On average, a yield of 7.7 x 1012 GC total was achieved using polyethylenimine (PEI) and 2.4 x 1013 GC using TransIt (Figure 2B). For PEI, this is an almost 3-fold improvement; for TransIt, it is a 9.2-fold improvement over titers obtained with adherent culture (2.6 x 1012). When used to transfect suspension cells, TransIt gives approximately a 3-fold improvement in yield compared to PEI (Figure 2B).
Subsequently, the performance of the AAV vectors was tested in vivo. GFP vectors were produced with PEI and TransIt and compared to GFP vectors produced with the classical adherent system. At 4 weeks after injection, mice (6 weeks C57BL/6 female mice weighing 20-25 g) were sacrificed, and GFP expression in mouse's brains was evaluated by histology (Figure 3A). Comparative analysis was done on sagittal brain sections. No difference was observed in the transduction pattern of the virus produced using these production methods (Figure 3B). The luciferase activity of the vectors produced using either method was also assessed. Transduction was measured in vivo for 3 weeks; the expression pattern over time is similar no matter which transfection reagent is used (Figure 4).
Next, the protocol was tested and implemented by other members of the team (Table 2). The production of other capsids with PEI was successful, yielding an average yield of 3.5 x 1012 GC vectors. For capsid B10 production, a 3-fold improvement can be achieved by using TransIt versus PEI. For another project, several constructs packaged in AAv.PHP.eB gave an average yield of 3.7 x 1012, which is enough to inject 7 mice intravenously at a dose of 5 x 1011 GC per mouse. These results illustrate that the protocol can be successfully used by several researchers for various capsids and construct combinations.
Figure 1: Schematic overview of production. On day 1 (D1), cells are cultured in 300 mL tubes. On day 2 (D2), cells are transfected with the relevant plasmids. On day 5 (D5), cells are harvested and lysed using a 10x tween lysis buffer. After lysis, cell debris is removed by centrifugation. Subsequently, the cell lysate is filtered to remove large proteins, and the AAV virus is precipitated using Polyethylene glycol (PEG) 8000. On day 6 (D6), the iodixanol purification and concentration are performed. The PEG concentrate from the previous day is treated with DNase and afterward placed through an iodixanol gradient. The purified vectors are subsequently desalted and concentrated using a centrifugal filter. Please click here to view a larger version of this figure.
Figure 2: Incubator set-up and AAV yields. (A) Viral production cells are cultured in a shaking incubator with a 50 cm shaking diameter. This incubator is equipped with an adapter plate for 600 mL tubes and special inserts for 50 mL tubes produced by the mechatronics department. (B) Following the initial setup, reporter constructs using either polyethyleneimine (PEI; 7.7 x 1012 GC total, n=5) or TransIt (2.4 x 1013 GC total, n=5) as transfection method and compared to the yields achieved with the classical adherent method (2.6 x 1012 GC total, n=5) as described in Verhaagen et al.12. Data are presented as mean ± Standard deviation. Please click here to view a larger version of this figure.
Figure 3: GFP expression pattern after in vivo evaluation. (A) Schematic depiction of a typical AAV vector construct containing the ubiquitous promoter CMV immediate-early enhancer, chicken β-actin promoter, and rabbit β-Globin splice acceptor site (CAG) followed by Green fluorescent protein (GFP), the Woodchuck Hepatitis Virus (WHV) Posttranscriptional Regulatory Element (WPRE) and a polyadenylation tail (pA)13. Here, 6-week-old female mice (n=3) received a tail vein injection containing 5 x 1011 total genomic copies (gc)/mouse. At 4 weeks after injection, mice were sacrificed, and brains were analyzed by histology. (B) Representative depiction of the transduction pattern in sagittal sections, using adherent HEK293T cells and suspension cells with polyethyleneimine (PEI) or TransIt transfection reagent, a similar transduction pattern is observed irrespective of production method used. Scale bar = 1000 µm. Please click here to view a larger version of this figure.
Figure 4: Similar luciferase activity irrespective of transfection method. (A) Schematic depiction of the construct used containing the ubiquitous promoter CMV immediate-early enhancer, chicken β-actin promoter, and rabbit β-Globin splice acceptor site (CAG) luciferase (LUC) with a V5 tag and a bovine growth hormone polyadenylation tail (BGHpA), 5 x 1011 GC total/mouse (n=3) was injected via the tail vein, luciferase activity in the cranial area (circled in red, in example image) as bioluminescent image was measured weekly. (B) Bioluminescent activity was measured as relative luminescence units (RLU) depicted as a blue bar for PEI production and a brown bar for TransIt. No significant difference in luciferase activity in the selected brain area was observed between groups. Data are presented as mean ± Standard deviation. Please click here to view a larger version of this figure.
Table 1: Transfection parameters. Please click here to download this Table.
Table 2: Yields of the implemented protocol. Suspension protocol used by other researchers to produce AAV. Please click here to download this Table.
Systemic administration of AAV is a powerful tool for gene transfer to the CNS; however, the production of AAV is an expensive and laborious process. By using suspension cells, labor and plastics are reduced compared to the adherent culture of HEK293T on 15 cm2 plates. Furthermore, the conical tubes implemented here are easy to handle and maximize the use of laboratory space. The protocol was set up by two researchers and subsequently applied by others in the lab. A series of productions by three independent researchers yielded, on average, enough vectors per batch to inject 6-7 mice.
The use of HEK293 suspension cells has been described earlier13,14. However, the suspension cell line described is not freely available for research purposes. This is a bottleneck for academic researchers when applying the described methods. The suspension cell line described in this protocol is freely available for research purposes. The culture of suspension cells was done in conical tubes instead of Erlenmeyer's to save space and time. With Erlenmeyer's, up to 4 productions can be cultured simultaneously, versus 12 productions using conical tubes. Conical tubes can be directly transferred to the centrifuge for harvest.
One alternative possibility for a suspension cell culture system is the baculoviral system. An advantage of the baculo system is that these cells and media are publicly available and can easily be applied. For academics, repositories such as Addgene are available, containing a large library of ready-to-use AAV transfer plasmids12. While these can also have caveats, such as faulty ITRs, they serve as a good source for plug-and-play experiments. The choice was made to keep a HEK293-based production system as available plasmids can directly be used for production without the transfer of the construct to baculovirus.
For transfection, polyethyleneimine is relatively cheap and is, therefore, the transfection reagent of choice. As an alternative, a novel transfection method (TransIt) was evaluated to increase the yield of vectors. A three-fold improvement was observed using half the amount of plasmids. This makes TransIt an interesting alternative transfection reagent when larger stocks are needed. A limitation of this is the cost, making it less interesting for small batches.
In summary, here, a protocol is presented that can be used to produce high-yield AAV in suspension cells in an academic laboratory setting. Tools are described that can be used in an academic laboratory setting to produce AAV at a high yield.
The authors have nothing to disclose.
This work was supported by a grant from the Royal Netherlands Academy of Arts and Sciences (KNAW) research fund and a grant from Start2Cure (0-TI-01). We thank Leisha Kopp for her input and advice in the setup of the protocol. Figures were created using Biorender.
39 mL, Quick-Seal Round-Top Polypropylene Tube, 25 x 89 mm – 50Pk | Beckman Coulter | 342414 | |
Adapter 600 mL conical tubes, for rotor S-4×1000, | eppendorf | 5920701002 | |
Adapter Plate fits 16 bioreactors of 600 ml | Infors HT/ TPP | 587633 | |
Aerosol-tight caps, for 750 mL round buckets | eppendorf | 5820747005 | |
Centrifuge 5920 R G, 230 V, 50-60 Hz, incl. rotor S-4×1000, round buckets and adapter 15 mL/50 mL conical tubes | eppendorf | 5948000315 | |
Distilled Water | Gibco | 15230147 | |
DNase I recombinant, RNase-free | Roche | 4716728001 | |
DNase I recombinant, RNase-free | Roche | 4716728001 | |
DPBS, calcium, magnesium | Gibco | 14040091 | |
DPBS, no calcium, no magnesium | Gibco | 14190144 | |
Fisherbrand Disposable PES Filter Units 0,2 | Fisher | FB12566504 | |
Fisherbrand Disposable PES Filter Units 0,45 | Fisher | FB12566505 | |
Holder for 50 ml culture tubes also fits falcon tube | Infors HT/ TPP | 31362 | |
Holder for 600 ml cell culture tube | Infors HT/ TPP | 66129 | |
Incubator Minitron 50 mm | Infors HT | 500043 | |
LV-MAX Production Medium | Gibco | A3583401 | |
N-Tray Universal | Infors HT/ TPP | 31321 | |
OptiPrep – Iodixanol | Serumwerk bernburg | 1893 | |
PEI MAX – Transfection Grade Linear Polyethylenimine Hydrochloride (MW 40,000) | Poly-sciences | 24765-100 | |
Phenol red solution | Sigma-Aldrich | 72420100 | |
Poly(ethylene glycol) 8000 | Sigma-Aldrich | 89510 | |
TransIT-VirusGEN | Mirus | Mir 6706 | |
Trypan Blue Solution, 0.4% | Gibco | 5250061 | |
TubeSpin Bioreactors-50ml | TTP | 87050 | |
TubeSpin Bioreactors-600ml | TTP | 87600 | |
Viral Production Cells | Gibco | A35347 | |
Vivaspin 20 MWCO 100 000 | Cytvia | 28932363 |