Association between adeno‐associated virus genomic titers and intracellular plasmid levels

The recombinant adeno‐associated viral (rAAV) vector is one of the most effective viral vectors in gene therapy because of its low immunogenicity, high transduction efficiency, broad tissue specificity, and long‐term transgene expression ability. HEK293T cells are widely used to produce rAAV vectors, and acquisition of high rAAV‐producing cells is one of the essential processes for higher rAAV vector production. To acquire such cells, a method to identify cells with high expression of transgene protein was developed; however, the relationship between high expression of transgene protein and increased production of rAAV has not been sufficiently studied to date. We used the fluorescent protein ZsGreen1 as a transgene, subdivided the adherent cells according to the intensity of transgene expression protein, and reconfirmed the relationship between rAAV vector production and the amount of intracellular plasmid. We found that the amount of rAAV produced was not correlated with the intensity of expression of the transgene protein but may be correlated with the amount of intracellular plasmid. We also found that cells with high expression of transgene protein might not necessarily produce large amounts of rAAV vector. Based on these results, it was suggested that intracellular plasmid copy number could be used as a new marker for cells producing high levels of rAAV vector.


INTRODUCTION
In recent years, adeno-associated viral (AAV) vector has become a key tool in gene therapy.5][6][7][8] In addition, a large amount of the viral vector is required; 9,10 for example, the current globally approved onasemnogene abeparvovec (Zolgensma) requires 1.1 × 10 14 vg/kg for children weighing 2.6 kg or more, 11,12 thus necessitating mass production of rAAV vectors.Therefore, to achieve efficient cost reduction in rAAV vector production and obtain as high an yield of rAAV vectors as possible, it is necessary to design high-yielding stocks of rAAV vector-producing cells.7][18][19] Wild-type AAV is a dependoparvovirus and depends on helper virus proteins to provide essential genes for a productive infection. 20,21Since the E1A/E1B genes are natively integrated into HEK293 or HEK293T cells, 22,23 other genes from adenovirus, namely, E2A, E4, and virus-associated (VA) genes must be carried by pHelper plasmids.The pRC plasmid carries the Rep and Cap of the wild-type AAV genes, with the Rep proteins playing a central role in AAV DNA replication and a leading role in vector genome packaging and the VP proteins compose the capsid of AAV. 17,19he pAAV-transgene plasmid is usually introduced with the gene of interest, that is transgene, used for gene therapy, excised, and then replicated using Rep proteins, and finally packaged into the rAAV capsid.Gene of interest protein expression is not essential for rAAV vector production; however, similar to other plasmids, this plasmid also expresses the protein.
Gu et al. hypothesized that rAAV vector productivity by triple transfection shows a positive correlation with cell transfection efficiency; thus, the transfected cells were sorted to identify green fluorescent protein (GFP)-positive transfected cells. 24However, the relationship between high expression of transgene proteins and increased production of rAAV vector has not been adequately studied.In this study, we compared transgene protein expression with rAAV titration using a cell sorter.We also employed Zoanthus sp. Green fluorescent protein (ZsGreen1) instead of the transgene of the rAAV2 vector to observe transgene protein expression in detail.

Cells and culture medium
The AAVpro 293 T cell line (Takara Bio, Shiga, Japan) was used as the host cell line in the present study.The cells were maintained in T-flasks with high glucose Dulbecco's modified Eagle Medium (DMEM) (Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% HyClone™ Fetal Bovine Serum (GE Healthcare, UT, USA) and 1% penicillin-streptomycin (Thermo Fisher Scientific, Waltham, MA, USA) at 37 • C in a 5% CO 2 incubator.

Triple plasmid transfection of AAVpro 293 T cells
The cells were plated at the cell density of 4.0 × 10 4 cells/cm 2 or 8.0 × 10 4 cells/cm 2 in 20 mL of complete growth medium in a T-75 flask for approximately 18-24 h and then transfected using the TransIT®-293 transfection reagent (Mirus Bio, Madison, WI, USA).The cells were transfected with 21 μg of total plasmid vectors (pRC2-mi342, pHelper, and pAAV-ZsGreen1 plasmid (Takara Bio) at a ratio of 1:1:1 (w/w)) and 63 μL of TransIT®-293 transfection reagent in 2 mL serum-free DMEM.Approximately 6 h after transfection, the culture medium was completely replaced with fresh DMEM containing 2% fetal bovine serum (FBS).

Sample preparation
Three days after transfection, 250 μL of 0.5 M ethylenediaminetetraacetic acid (EDTA) (pH 8.0) was added to 20 mL of the culture medium containing rAAV2-ZsGreen-producing cells and incubated for approximately 2 min.Twenty microliters of the detached cell suspension medium was dispensed and counted using a TC20 automated cell counter (Bio-Rad, Hercules, CA, USA).Another 100 μL of the detached cell suspension medium for use as a whole titer was then dispensed and centrifuged at 1800 × g at 4 • C for 10 min.The supernatant was discarded and the cells were resuspended in 100 μL of 1× phosphate-buffered saline (PBS) (−).The sample was stored at −80 • C until use.The remaining detached cells were collected in a sterile 50 mL centrifuge tube and centrifuged at 100 × g at 4 • C for 5 min for flow cytometory.The supernatant was removed, and the cell pellet was washed in PBS (−) to remove any residual plasmids and transfection reagents.Finally, the cell pellet was resuspended in PBS (−) and sorted using the SH800S cell sorter (Sony Biotechnology Inc., San Jose, CA, USA).

Flow cytometry and cell sorting using a flow cytometer
Samples were acquired on the SH800S cell sorter (Sony Biotechnology, San Jose, CA, USA).Before fractionation of ZsGreen1-positive cells, single cells were gated based on the forward scatter area (FSC-A) and forward scatter height (FSC-H) properties, followed by selection of intact cells based on back scatter area (BSC-A) and FSC-A scatter properties.Intact cells were also confirmed by LysoTracker® deep red fluorescent dye (Thermo Fisher Scientific, Waltham, MA, USA) and screened by a 638 nm-excitation laser in the SH800S cell sorter (data not shown).Four hundred eighty-eight nm excitation laser for the fluorescein isothiocyanate (FITC) channel was used for the detection of ZsGreen1-positive cells, since ZsGreen1 excitation and emission maxima are 493 and 505 nm, respectively. 25sGreen1-positive cells were identified and extracted based on the FITC fluorescence intensity and were subdivided into four fractions (L, M1, M2, and H fractions) expediently.Each fraction was sorted into 50,000 cells using one well of a 96-well plate in "ultra-purity" mode.The cells were collected into 1.5 mL centrifuge tubes and then centrifuged at 1800 × g at 4 • C for 10 min.The supernatant was discarded and the cell pellet was washed in 1× PBS to remove impurities contained in the sheath solution (Beckman Coulter, Inc., Brea, CA, USA) used in the cell sorter.Finally, the cell pellet was suspended in 125 μL of 1× PBS, and the samples were stored at −80 • C until use.

Titrating rAAV using a droplet digital PCR
To titrate the rAAV genomes, 2 μL of each rAAV solution was added to 18 μL of DNase I solution (5 U, 40 mM Tris-HCl (pH 7.5), 8 mM MgCl 2 , 5 mM dithiothreitol (DTT), 0.05% Pluronic F68) and incubated at 37 • C for 30 min.DNase I was inactivated by adding 1 μL of 0.5 M EDTA (pH 8.0), followed by incubation at 25 • C for 5 min.The sample was heated at 95 • C for 10 min to denature the AAV capsid.This sample was used as a pretreated sample.After proper dilution of the pretreated sample with 0.05% Pluronic F68 in tris-EDTA (TE) solution, 1 μL of the rAAV sample was mixed with 19 μL of reaction mixture with droplet digital PCR (ddPCR) Supermix for Probe (No 2 ′ -deoxyuridine, 5 ′ -triphosphate (dUTP)) following the instructions from Bio-Rad.After droplet generation, ddPCR was performed on a C1000 Touch Thermal Cycler (Bio-Rad).The ddPCR program was performed at 95 • C for 10 min for enzyme activation, 40 cycles at 94 • C for 30 s followed by 60 • C for 1 min, and 98 • C for 10 min for enzyme deactivation.The droplets were analyzed using a QX200 droplet reader and QuantaSoft™ (Bio-Rad).The ddPCR primers and probe for determining the rAAV titer are described in Table 1. a The 5 prime ends of the all probes were labeled with FAM.b The 3 prime ends of all probes were labeled with Black Hole Quencher-1.

2.6
Titrating rAAV using a median tissue culture infectious dose (TCID 50 ) assay Titration of infectious rAAV2 particles based on ZsGreen1 fluorescence intensity was performed following the procedure for the rAAV infectious titer assay of the AAV2 reference standard stock available on the International Society of BioProcess Technology (ISBioTech) website (https://isbiotech.org/ReferenceMaterials/pdfs/AAV2_RSS_Infectious_titer_assays_V2.pdf).The TCID 50 transducing units (tu) were calculated from three independent assays.

Monitoring intracellular plasmids in ZsGreen1-positive cells using ddPCR
Three independent ddPCR reactions were performed on QX200 Droplet Reader (Bio-Rad), C1000 Touch Thermal Cycler (Bio-Rad), and QuantaSoft™ (Bio-Rad).The plasmid copy number detection primers and the custom TaqMan probe used in this study were synthesized by Thermo Fisher Scientific (Table 1).

Isolation and culturing of the M2 or H fraction cells
AAVpro293T cells in T-75 flasks were transfected with 21 μg of pAAV-ZsGreen1 plasmid and 63 μL of TransIT®-293 transfection reagent in 2 mL serum-free DMEM.Approximately 6 h after transfection, the culture medium was completely replaced with fresh DMEM containing 2% FBS.Three days post-transfection, 300,000 cells were collected using the SH800S cell sorter from M2 or H fractions and cultured in a 6 well plate (Zs-M2 or Zs-H fraction cells).Following expansion for 10 days in a T-75 flask, the cells from each fraction were again transfected with 21 μg of each plasmid vector at a ratio of 1:1:1 (w/w) and 63 μL of TransIT®-293 transfection reagent in 2 mL serum-free medium.Approximately 6 h after transfection, the culture medium was completely replaced with fresh DMEM containing 2% FBS.Three days post-transfection, 50,000 cells were collected using the SH800S cell sorter from the Zs-M2 or Zs-H fractions.

Statistical analysis
Statistically significant differences among data sets were determined using Student's t-test.Statistical significance was set at p < 0.05.Correlation coefficient (r) between 0.3 and 0.7 indicated moderately positive linear relationships, and the correlation coefficient between 0.7 and 1.0 indicated strong positive linear relationships. 26

Fractionation of ZsGreen1-positive cells and evaluation of rAAV2 titer
When the plasmid is transfected into the cells, the proteins required for rAAV production are translated, the rAAV capsid is produced, the transgene is replicated between ITRs, and the rAAV vector genome is inserted into the capsid.Then, the transgene protein, which is originally unnecessary, is also translated at the same time (Figure 1).First, we compared transgene protein expression with rAAV titration using a cell sorter.To investigate the mechanism of intracellular rAAV2 vector production, we fractioned the ZsGreen1-positive cells into low, medium1, medium2, high ZsGreen1 protein expression fractions (i.e., L, M1, M2, and H fractions, respectively) according to the amount of ZsGreen1 protein expression and used a cell sorter to collect 50,000 cells from each fraction.In all four fractions, the difference in protein expression level was compared to the difference in the rAAV2 titer or intracellular plasmid copy number.An outline is shown in Figure 2. We used Zoanthus sp.green fluorescent protein (ZsGreen1) instead of the GFP in the transgene of rAAV2 vector to observe transgene protein expression in detail.][29] Triple plasmid transfection was conducted to produce rAAV2-ZsGreen1 in AAVpro293T cells, and 3 days post-transfection, the ZsGreen1-positive cells were extracted using a cell sorter.We focused on altering the initial cell density for preculture approximately 18-24 h before transfection and found no differences in the ratio of ZsGreen1-positive cells among the initial cell density of 4.0 × 10 4 cells/cm 2 or 8.0 × 10 4 cells/cm 2 (Figure 3A).The ZsGreen1 protein-positive cells were subdivided into four fractions, that is, L, M1, M2, and H fractions, based on ZsGreen1 fluorescence intensity (Figure 3A).We then examined the genome titer of 800 cells in each fraction using ddPCR.The rAAV2 genome titer at an initial cell density of 4.0 × 10 4 cells/cm 2 proportionally increased from the L fraction to the H fraction with the ZsGreen1 protein level (Figure 3B), whereas in case of higher initial cell density, the rAAV2 genome titer of each fraction reached a plateau; that is, it did not correlate with the ZsGreen1 protein level (Figure 3B).An infectious titer assay (TCID 50 ) was also performed using this system, and it was found that the rAAV titer using TCID 50 could be detected in 4000 cells theoretically.Investigation of the transduction unit of each fraction showed that the rAAV2 infectious titer per cell was almost the same in all fractions and did show any correlation with ZsGreen1 protein expression (Figure S1).When successive rAAV2 genome titers for each fraction were monitored, the genome titer initially increased from the L to H fraction until about 2 days after transfection, while excessive cell density caused an imbalance in rAAV vector productivity of the cell (Figure S2).These results suggest that high-titer cells with rAAV vector may change depending on the initial cell density, culture days, and cell condition; that is, the transgene protein expression was not essential for rAAV vector production and not significant as a marker for the selection of cells producing high amounts of rAAV vector.

Monitoring intracellular plasmid copy number in rAAV-ZsGreen1-positive cells
Some methods for measuring the amount of plasmid in rAAV vector production have so far been mainly used only to measure the amount of residual plasmid in rAAV vector. 30,31In this study, we attempted to monitor the intracellular plasmid copy numbers in rAAV2-ZsGreen1-positive cells.Four hundred cells in each fraction were used to monitor the intracellular plasmid copy number using ddPCR (Figure 2).When the initial cell density was 4.0 × 10 4 cells/cm 2 , the number of copies of the intracellular plasmid increased from the L fraction to the H fraction, similar to that in case of the rAAV2-ZsGreen1 genome titer (Figure 4).However, when the initial cell density was doubled at 8.0 × 10 4 cells/cm 2 , the plasmid copy number of each fraction reached a plateau similar to that in case of the rAAV2-ZsGreen1 genome titer, which was not correlated with the ZsGreen1 protein expression level (Figure 4).It is speculated that the reason for the constant intracellular plasmid levels in each fraction when the initial cell density was increased could be competition for the plasmid, because the amount of plasmid available was insufficient relative to the number of cells.Interestingly, the copy number ratio between the three plasmids in the cell was equal to the molar ratio of the three plasmids, even though they were transfected in a weight ratio at the time of transfection.Grieger et al. reported that higher rAAV genome titers were obtained using molar ratios of the three plasmids than using weight ratios; 16 It might stand the reason to use the molar ratio.
The rAAV genomic titer per cell, summarized in Figure 5 showed a strong correlation with intracellular plasmid copies (the correlation coefficient for the initial cell density of 4.0 × 10 4 cells/cm 2 ; 0.92, the correlation coefficient for initial cell density of 8.0 × 10 4 cells/cm 2 ; 0.97) rather than with the intensity of transgene protein expression (the correlation

F I G U R E 5
The results of comparison between rAAV genome titer, ZsGreen1 protein expression level, and intracellular plasmid copies.
The median value of the fluorescence intensity in each fraction for the ZsGreen1 protein expression level was determined.rAAV: recombinant adeno-associated virus; FCM: flow cytometer; FITC: fluorescein isothiocyanate; ddPCR: droplet digital PCR; NS: not significant coefficient for 4.0 × 10 4 cells/cm 2 ; 0.85, the correlation coefficient for 8.0 × 10 4 cells/cm 2 ; 0.54) (Figure 5).Similar to genomic titers, intracellular plasmid copies also did not correlate with transgene protein expression levels at varying initial cell densities (the correlation coefficient for 4.0 × 10 4 cells/cm 2 ; 0.97, the correlation coefficient for 8.0 × 10 4 cells/cm 2 ; 0.72) (data not shown).Also, cells with high expression of transgene protein might not necessarily be cells producing high amounts of rAAV vector.

Isolation and culturing of cells with high productivity
To further confirm whether the H fraction was a cell group with greater productivity, we cultivated the H fraction cell group with the M2 fraction cell group as a control, and studied the characteristics.The M2 and H fractions of pAAV-ZsGreen1-transfected cells were sorted and cultured in a 6 well plate (Figure 6A).Following expansion for 10 days in a T-75 flask, the cells from each fraction (Zs-M2 and Zs-H fraction cells) were transfected with three plasmid vectors for the production of rAAV2-ZsGreen1.Three days post-transfection, the Zs-M2 and the Zs-H fraction cells were analyzed by flow cytometer (FCM) and it was found that there was almost no difference in the ratio of ZsGreen1-positive cells between the Zs-M2 and Zs-H fraction cells (Figure 6B).The 50,000 cells of M2 and H fraction were again collected by cell sorter from Zs-M2 or Zs-H fraction cells, respectively, and the rAAV2 genome titer of the M2 or H fraction cells in each fraction was measured by ddPCR.The genome titer also did not differ between the two fractions.(Figure 6C).This suggests the possibility that even if the H-fractionated cell group is extracted, only high ZsGreen1-producing cells cannot be extracted, and a mixture of low and high ZsGreen1-producing cells are likely obtained.However, the AAVpro293T cells used in this study are adherent cells and differ considerably in gene expression and metabolic pathways from the serum-free conditioned suspension HEK293T and HEK293 cells currently used in industry. 32Therefore, different results may be obtained when suspension of HEK293 cells are used.Gu et al. said that due to extensive cell death in the GFP high population, which is most likely attributed to GFP overexpression, they selected the viable GFP-medium cells for clonal cell isolation. 24Since we used ZsGreen1 instead of GFP as the fluorescent protein in this study, cell death was hardly observed during the 10 days culture period.Interestingly, fluorescence microscopy results showed that ZsGreen1 protein expression did not decrease in the transfected cells even after 10 days (data not shown).Therefore, when ZsGreen1 protein is used as an index to isolate high-expressing cells as in Gu et al., 24 ZsGreen1 protein contamination may pose an industrial problem.

CONCLUSION
This study found that using adherent HEK293T cells, the rAAV genomic titer per cell showed a strong correlation with intracellular plasmid copy number rather than with the intensity of transgene protein expression.It was also found that cells with high expression of transgene protein might not necessarily produce large amounts of rAAV vector.Based on these results, it was suggested that intracellular plasmid copy number could be used as a new marker for cells producing high levels of rAAV vector.
Schematic representation of rAAV vector production by triple plasmid transfection.The red circles represent the monitoring factors in this assay, that is, transgene protein expression, rAAV genomic titer, and intracellular plasmid copy number.rAAV: recombinant adeno-associated virus F I G U R E 2 A basic schematic representation of the analyzes of rAAV-producing cells in this assay.rAAV: recombinant adeno-associated virus F I G U R E 3 Comparison of ZsGreen1 protein expression and rAAV genomic titers based on initial cell density.(A) FCM analysis of untreated negative control and three plasmid-transfected cells.The area in which the ZsGreen1-positive cells existed is indicated by an open red square.The ZsGreen1-positive cells were subdivided into 4 classes: L, M1, M2, and H based on the intensity of FITC (488 nm).(B) The result of rAAV genome titer (vg) for each fraction by ddPCR assay (*p < 0.05; **p < 0.01; NS, p > 0.05)

F I G U R E 4
Comparison of intracellular plasmid copy numbers in rAAV2-ZsGreen1-positive cells.Measurement of the residual intracellular plasmid copies for each fraction by the ddPCR assay (*p < 0.05; **p < 0.01; NS, p > 0.05).

F I G U R E 6
Isolation and culturing of cells with high productivity.(A) A basic schematic representation of the analyzes of rAAV-producing cells in this assay.The M2 or the H fraction cells from pAAV-ZsGreen1 plasmid-transfected cells were sorted and cultured for 10 days (Zs-M2 or Zs-H fraction cells, respectively), and then three plasmid transfections were performed in each of the fraction cells.(B) FCM analysis of three plasmid-transfected Zs-M2 and Zs-H fraction cells 3 days after the transfection.The area where the cells were sorted and ddPCR analyzed is shown in the open red square.(C) The result of rAAV genome titer (vg) for the M2 and H fraction from Zs-M2 or Zs-H fraction cells by ddPCR assay.rAAV: recombinant adeno-associated virus; FCM: flow cytometer; ddPCR: droplet digital PCR