2023 年 5 巻 2 号 p. 28-35
Several recombinant adeno-associated virus-based gene therapy products have been developed recently. The vector genome (VG) titer is a critical quality attribute associated with the clinical dosing of these products and, thus, requires accurate quality control measures. Typically, VG titers are measured by quantitative polymerase chain reaction (qPCR). Droplet digital PCR is more reliable than qPCR and a powerful analytical tool for quantifying genome copies with high accuracy and precision; however, VG titers cannot be correctly quantified without the appropriate preparation of analytical test samples. In this study, we systematically assessed the role of each component and treatment comprising DNase treatment for free and residual DNA, DNase inactivation, and single-stranded DNA extraction from capsids in VG titration results during pre-analytical sample preparations. Incubation near the single-stranded DNA-release temperature decreased the number of recombinant adeno-associated virus genome copies. Moreover, we developed a simplified three-step pre-analytical procedure with concurrent DNase inactivation and single-stranded DNA extraction at a much higher temperature than the release temperature. Developing an analytical procedure for recombinant adeno-associated virus genome titration by droplet digital PCR based on release temperature is a science- and risk-based approach that would improve quality control testing of recombinant adeno-associated virus-based gene therapy products.
The vector genome (VG) titer is a crucial quality attribute for rAAV-based gene therapy products. The current method for measuring the VG titer concentration is qPCR, which is less accurate than ddPCR. However, appropriate preparation of analytical test samples is required to accurately quantify VG titers using ddPCR. This study investigated the effect of pre-analytical sample preparation, including DNase treatment, DNase inactivation, and single-stranded DNA extraction from capsids on VG titration results, and proposed a simplified three-step pre-analytical procedure for VG titration using ddPCR. This approach can improve quality control testing of rAAV-based gene therapy products.
Adeno-associated virus (AAV) is a parvovirus encapsidating up to 4.7 kb single-stranded deoxyribonucleic acid (ssDNA) flanked by two inverted terminal repeats. The recombinant AAV (rAAV) vector is an essential tool for in vivo gene therapy owing to its low immunogenicity, non-pathogenicity, and long-term transgene expression [1, 2]. Many clinical trials of rAAV-based gene therapy are ongoing worldwide, and six products, Glybera, Luxturna, Zolgensma, Upstaza, Hemgenix, and Roctavian, have been launched to date. To ensure the efficacy and safety of rAAV-based gene therapy, it is important to develop appropriate analytical methods to assay vector titers and quantify impurities using suitable analytical equipment [3]. Although several technologies to engineer rAAV have been developed, the quality control of rAAV vectors remains a key challenge in gene therapy product development.
The vector genome (VG) titer is a key and widely applied quality attribute of rAAV-based gene therapy products associated with efficacy and clinical doses. Real-time quantitative polymerase chain reaction (qPCR) is a standard analytical method for quantifying VG titers [4, 5]. Recently, droplet digital PCR (ddPCR) has been developed to quantify VG titers because of its powerful analytical technique that enables absolute genome quantification [6], resulting in VG titration with higher accuracy and precision than that of qPCR [7,8,9,10]. Although ddPCR is an excellent analytical tool for quantifying VG titers with low variability to obtain appropriate results for the quality control of rAAV vectors, developing a pre-analytical procedure is essential to prepare test samples for rAAV genome titration by ddPCR. Although previous studies have applied ddPCR and qPCR to rAAV genome titration, no standardized pre-analytical procedure has been developed. Various pre-analytical procedures were utilized for each study; however, the process parameters applied in preparing the analytical test samples influenced the results of VG titration [11,12,13,14].
To establish an appropriate rAAV genome titer assay via ddPCR, understanding the pretreatment parameters for analytical test sample preparation, which affect the titration result throughout the analytical procedure development, is necessary. In this study, we investigated the importance of pretreatment parameters for analytical test sample preparation that affect the titration results throughout the analytical procedure to establish an appropriate rAAV genome titer assay via ddPCR. We also demonstrated that the ssDNA-release temperature (Tr) is a critical thermal parameter for ddPCR test sample pretreatment. The Tr value is regarded as a strong analytical parameter for genome titration related to an analytical target profile of the analytical quality by design approach, which will be proposed in the International Council for Harmonization (ICH) Q14 guidelines [15].
The vectors of AAV1- cytomegalovirus (CMV)-ZsGreen 1 (rAAV1), AAV2-CMV-ZsGreen 1 (rAAV2), AAV5-CMV-ZsGreen 1 (rAAV5), and AAV6-CMV-ZsGreen 1 (rAAV6) were produced at the Manufacturing Technology Association of Biologics (Tokyo, Japan) by triple plasmid transfection of HEK293T cells, followed by purification using AVB Sepharose™ High Performance as affinity chromatography (Cytiva, Marlborough, MA, USA). Based on the genome copies (GC) quantified via qPCR for release testing, rAAV vectors were diluted to approximately 108 GC/ml using 0.005% Pluronic F-68 (Thermo Fisher Scientific, Waltham, MA, USA) in 10 mM Tris-HCl buffer containing 1 mM ethylenediaminetetraacetic acid (EDTA) and Na2 as the TE buffer (Promega, Madison, WI, USA) before pre-analytical test sample preparation for ddPCR.
Primers and probesPrimers and probes for the CMV promoter were obtained from Eurofins Genomics (Tokyo, Japan). The sequence of the forward primer was 5′-CATCAATGGGCGTGGATAGC-3′, that of the reverse primer was 5′-GGAGTTGTTACGACATTTTGGAAA-3′, and that of the probe was FAM-ATTTCCAAGTCTCCACCC-BHQ1.
AAV genome titer assay by droplet digital PCRThe ddPCR sample (1 µl) containing the extracted rAAV genomes was mixed with 19 µl of the reaction mixture comprising ddPCR Supermix for Probe (No dUTP) (Bio-Rad, Hercules, CA, USA), primers (final concentrations of 900 nM), probes (final concentrations of 250 nM), and nuclease-free water (Promega). The mixture was encapsulated into droplets using Droplet Generation Oil for Probes (Bio-Rad) on a QX200 Droplet Generator (Bio-Rad) instrument. The droplets were set in a C1000 Touch thermal cycler (Bio-Rad) for PCR with the following conditions: one cycle at 95°C for 10 min for enzyme activation, 40 cycles each at 94°C for 30 sec, 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 software (Bio-Rad). For further information, please refer to https://www.bio-rad.com/webroot/web/pdf/lsr/literature/10031906.pdf. The copies per 20 µL readout were equivalent to GC/µL of the ddPCR sample. The AAV genome titer (GC/ml) was calculated using the formula X=1000 ×(a/b)Y, where X is the AAV genome titer, a is the final volume of the ddPCR sample in pretreatment, b is the volume of rAAV vector used in pretreatment for ddPCR measurement, and Y represents the GC/µL of ddPCR sample.
Evaluation of DNase treatment stepFirst, the rAAV samples were mixed with 500 U/ml recombinant DNase I (TaKaRa Bio, Kusatsu, Japan) and incubated at 30, 35, 37, 39, 41, and 46°C for 30 min each on a T100 thermal cycler (Bio-Rad) in the presence of 20 mM Tris-HCl (pH 7.5), 4 mM MgCl2, and 2.5 mM dithiothreitol (TaKaRa Bio). After DNase I treatment, the samples were mixed with 10 mM EDTA (TaKaRa Bio) in 10 mM Tris buffer (pH 8.0) (Nacalai Tesque, Kyoto, Japan) and incubated at 95°C for 10 min (approximately 107 GC/ml of rAAV genome). The samples were diluted to the appropriate GC concentration for ddPCR using 0.05% Pluronic F-68 in TE buffer.
Evaluation of DNase inactivation stepThe rAAV samples were mixed with 500 U/ml recombinant DNase I in the same buffer as described in the previous subsection and incubated at 37°C for 30 min in a thermal cycler. Subsequently, the samples were mixed with 10 mM Tris buffer with or without 10 mM EDTA and incubated at 50, 55, 60, and 65°C for 30 min and at 95°C for a further 10 min (approximately 107 GC/ml of rAAV genome). After that, the samples were diluted to the appropriate GC concentration for ddPCR using 0.05% Pluronic F-68 in TE buffer.
Determination of ssDNA-release temperature from rAAV capsid by ddPCRThe rAAV samples (4 µl) were incubated at 4°C and 40, 45, 50, 55, 60, 65, 70, and 75°C for 30 min in a thermal cycler. Recombinant DNase I (250 U/ml) was added, followed by incubation at 37°C for 30 min to digest capsid DNA. EDTA (40 mM) in TE buffer was added, and the mixture was incubated at 55°C for 30 min and at 95°C for 15 min to inactivate DNase I. The ddPCR samples were prepared by diluting 0.05% Pluronic F-68 in TE buffer (final volume of 100 µl). Encapsidated genome percentage was calculated using the formula X=(GCT /GC4°C) ×100, where X is the encapsidated genome percentage, GCT is GC/ml of the sample incubated at temperature T, and GC4°C is GC/ml of the sample incubated at 4°C in the first step.
Thermal stability characterization of rAAV capsid via differential scanning fluorimetry (DSF)Thermal stability of the rAAV vectors was characterized by DSF using the Uncle platform (Unchained Laboratories, Pleasanton, CA, USA). The fluorescence intensity of the rAAV vectors (approximately 1012 GC/ml) was monitored at temperatures ranging from 15–95°C and ramping at 1°C/min using the following two methods. (1) Determination of the melting temperature (Tm) and measurement of the fluorescence spectra of capsids excited by a 266 nm laser at each temperature point. The Tm was determined by monitoring the barycentric mean of the fluorescence spectrum from 300 nm–430 nm. (2) Determination of the temperature for ssDNA release (Tr) from rAAV vectors containing SYBR Gold (Thermo Fisher Scientific) diluted to 500-fold and measurement of the fluorescence spectrum of SYBR Gold excited by a 473 nm laser at each temperature point. The Tr was determined by monitoring the peak area of the fluorescence spectrum from 500–650 nm.
Evaluation of concurrent DNase inactivation and ssDNA extraction stepThe rAAV samples were mixed with 500 U/ml recombinant DNase I in the same buffer as previously described and incubated at 37°C for 30 min in a thermal cycler. Subsequently, the samples were mixed with 10 mM Tris buffer and incubated at 75, 85, and 95°C for 5, 10, and 20 min, respectively (approximately 107 GC/ml of rAAV genome) [7, 8, 11, 12]. The samples were further incubated at 95°C for 30, 45, and 60 mins and diluted to the appropriate GC concentration for ddPCR using 0.05% Pluronic F-68 in TE buffer.
The pre-analytical phase of sample preparation for ddPCR generally involves four steps: 1) DNase treatment to remove external DNA; 2) DNase inactivation to protect encapsidated DNA at the subsequent step; 3) ssDNA extraction to encapsulate ssDNA in droplets; and 4) PCR following droplet generation to detect the droplets containing ssDNA by fluorescence (Fig. 1A). First, we investigated whether the thermal parameters in the DNase treatment of rAAV vectors affected the ddPCR results. Typically, residual DNA comprising unencapsulated AAV-derived DNA, contaminating plasmid DNA, host cell DNA, and others in rAAV samples were eliminated by several DNase commercial products at 37°C according to the manufacturer’s instructions. The effect of heating on the outcome was evaluated comprehensively, as the equipment conditions and human error may alter the temperature during the actual test sample preparation. Figure 2 shows no remarkable differences in the GC/ml of the rAAV1, rAAV2, rAAV5, and rAAV6 vectors, representing the AAV serotypes. We also investigated the impact of incubation duration and DNase I concentration on ddPCR results for different rAAV1 vectors and found no remarkable differences in the GC of the vectors (Supplementary Fig.1). Thus, we concluded that the parameters applied in the DNase treatment step slightly impacted the number of GC of the rAAV vectors quantified using ddPCR. DNase inactivation was performed under various conditions (temperature, time, additive reagents, and other parameters) to prevent the degradation of the AAV vector genome in the next step. After DNase treatment, incubation conditions were evaluated based on the thermal effects of a mixture of DNase I and rAAV1 vectors. These temperatures include those used to inactivate DNase I and degrade capsid proteins using proteinase K [12, 13, 16]. The mixture excluded EDTA, a reagent known to inhibit the DNase I reaction, and increasing the incubation temperature gradually decreased the number of GC of the vector (Fig. 3A). However, the addition of EDTA did not reduce the GC number of the vector. Decreasing trends in the number of GC of rAAV2, rAAV5, and rAAV6 across different AAV serotypes were also observed (Fig. 3B). Furthermore, the addition of proteinase K followed by incubation at 55°C and 65°C substantially decreased the GC of the vectors (Supplementary Fig. 2). Moreover, PCR inhibition due to a higher proteinase K concentration was observed, as described in previous studies (data not shown) [11, 13]. This indicated that heating during DNase inactivation released ssDNA from the capsid and digested ssDNA with residual DNase activity prior to inactivation.
Workflows of pre-analytical test sample preparation for ddPCR-based rAAV genome titration. (A) This procedure comprises a four-step process: DNase treatment, DNase inactivation, ssDNA extraction, and polymerase chain reaction (PCR), followed by droplet generation. (B) This procedure comprises three steps: DNase treatment, concurrent DNase inactivation, ssDNA extraction, and PCR, followed by droplet generation.
Impact of incubation temperature on rAAV genome titration results in DNase treatment. The genome copies of rAAV1, rAAV2, rAAV5, and rAAV6 vectors incubated for 30 min at 30, 35, 37, 39, 41, and 46°C with 500 U/ml DNase I were quantified using ddPCR after DNase inactivation and ssDNA extraction. Test sample preparation was performed individually. Error bars indicate the standard deviation of ddPCR results (N=3).
Impact of incubation temperature on rAAV genome titration results in DNase inactivation step. (A) The genome copies of rAAV1 vectors incubated for 30 min at 4, 50, 55, 60, and 65°C in Tris buffer with and without ethylenediaminetetraacetic acid (EDTA) following DNase treatment were quantified using ddPCR after ssDNA extraction. (B) The genome copies of rAAV2, rAAV5, and rAAV6 vectors incubated in Tris buffer without EDTA were quantified using ddPCR after ssDNA extraction. Test sample preparation was individually performed. Error bars indicate the standard deviation of the ddPCR results (N=3).
To further understand the decrease in rAAV GC during DNA inactivation in the 50–65°C range, we determined the temperature required to release ssDNA from the rAAV capsids. First, GC of the rAAV1, rAAV2, rAAV5, and rAAV6 vectors were incubated at various temperatures, followed by DNase I treatment, and then quantified via ddPCR (Supplementary Fig. 3). The retention percentage of encapsidated ssDNA was calculated using the ddPCR results (Fig. 4A). The slight decrease in the GC of the vectors began at 45°C, and a remarkable decrease was observed at 50–60°C; there was no difference in trend among different capsid serotypes. Additionally, the amount of ssDNA released from rAAV1 did not increase remarkably upon the extension of the incubation period (Fig. 4B). Therefore, ssDNA was rapidly released from the capsids within 10 min as the critical temperature for genome release was reached. To associate the temperature required to release ssDNA with the physicochemical properties, we thermodynamically characterized rAAV vectors using the DSF method (Table 1). First, the melting temperature (Tm) of the capsids was determined by monitoring the fluorescence spectrum from 300–430 nm (derived from the autofluorescence of tryptophan excited by a 266 nm laser). Although Tm is a representative parameter associated with protein stability [17] that presents different values depending on the serotype of the capsid, these values were very different from the temperature required to release ssDNA in the ddPCR results. Second, the fluorescence spectrum of the rAAV vectors mixed with SYBR Gold, a fluorescent reagent used to detect DNA excited by a 473 nm laser, was monitored from 500–650 nm. This method identified the temperature at which ssDNA leakage occurred (ssDNA-releasing temperature; Tr) by increasing the fluorescent peak area derived from SYBR Gold-stained ssDNA released from rAAV capsids (Table 1). The Tr of the rAAV vectors was approximately the same as that of the ddPCR results, with no difference among the capsid serotypes.
Impact of incubation temperature on ssDNA leakage from rAAV capsids. (A) The genome copies of rAAV1, rAAV2, rAAV5, and rAAV6 vectors were incubated for 30 min at 4, 40, 45, 50, 55, 60, 65, 70, and 75°C in Tris buffer without EDTA followed by DNase treatment were quantified using ddPCR. The plots indicate the percentage of encapsidated ssDNA after incubation at each temperature. The values were calculated by dividing the GC/ml of the sample incubated at each temperature by the GC/ml of the sample incubated at 4°C (N=3). (B) The genome copies of rAAV1 vectors incubated for up to 40 min at 45, 55, and 65°C in Tris buffer without EDTA followed by DNase treatment were quantified using ddPCR.
| Serotype | Tma | Trb | ||
|---|---|---|---|---|
| Average (°C) | RSD % | Average (°C) | RSD % | |
| 1 | 86.6 | 0.07 | 55.0 | 0.07 |
| 2 | 73.4 | 0.12 | 56.0 | 1.73 |
| 5 | 91.7 | 0.16 | 55.8 | 0.45 |
| 6 | 80.9 | 0.61 | 55.1 | 0.40 |
aTm indicates the melting temperature determined by monitoring the barycentric mean of the fluorescence spectrum from 300 to 430 nm, which was derived from tryptophan in capsids excited by a 266 nm laser. The Tm was measured thrice, and the average and relative standard deviation (RSD) % were calculated accordingly. bTr indicates the ssDNA-releasing temperature from capsids, determined by monitoring the peak area of the fluorescence spectrum from 500 to 650 nm, which was derived from SYBR Gold conjugated with ssDNA, using a 473 nm laser. The Tr was measured thrice, and the average and RSD % were calculated accordingly.
During DNase inactivation, caution should be exercised not to heat the vectors to a temperature that would ensure ssDNA leakage. The temperature should not be the melting temperature, which is commonly used as an indicator of protein stability, but rather a lower temperature that can be identified using the DSF method.
Three steps of pretreatment and impact of thermal parameters on genome titer determination by ddPCRBased on the results described in the previous sections, we believe that a three-step pre-analytical procedure comprising (5) DNase treatment, (6) concurrent DNase inactivation and ssDNA extraction, and (7) PCR following droplet generation (Fig. 1B) is suitable for analytical test sample preparation for ddPCR because it is simple and can reduce the risk of ssDNA leakage. To investigate whether the temperature and incubation time in concurrent DNase inactivation and ssDNA extraction affected the ddPCR results, GC of the rAAV vectors incubated for 5, 10, and 20 min at 75, 85, and 95°C, respectively, were quantified using the ddPCR method (Fig. 5A–5D). The number of GC increased when the incubation temperature reached the Tm (Table 1); however, no further increase was observed when the temperature exceeded the Tm. Further incubation was performed for up to 60 min at 95°C, and the number of GC gradually decreased with increasing temperature (Fig. 6 because excessive heat stress may have led to ssDNA degradation, resulting in a decrease in the number of GC. However, there was no impact of the heating ramp rate up to 95°C in the range of 0.1–1°C/s (Supplemantary Fig. 4). Thus, the thermal parameters in the concurrent DNase inactivation and ssDNA extraction steps had little impact on the number of GC of the rAAV vectors quantified by ddPCR.
Impact of incubation temperature on rAAV genome titration results during concurrent DNase inactivation and ssDNA extraction. The genome copies of (A) rAAV1, (B) rAAV2, (C) rAAV5, and (D) rAAV6 vectors incubated for 5, 10, and 20 min at 75, 85, and 95°C, respectively, in Tris buffer without ethylenediaminetetraacetic acid (EDTA) following DNase treatment were quantified using ddPCR. Test sample preparation was performed individually. Error bars indicate the standard deviation of the ddPCR results (N=3).
Impact of incubation time at 95°C on rAAV genome titration results during concurrent DNase inactivation and ssDNA extraction. The genome copies of rAAV1, rAAV2, rAAV5, and rAAV6 vectors incubated for 10, 30, 45, and 60 min at 95°C in Tris buffer without ethylenediaminetetraacetic acid (EDTA) following DNase treatment were quantified using ddPCR. Test sample preparation was performed individually. Error bars indicate the standard deviation of ddPCR results (N=3).
Several studies on rAAV-based gene therapy have been conducted, and ddPCR has been widely used for rAAV genome titration. However, the procedures for pre-analytical test sample preparation for ddPCR vary between studies, as most methods are based on the experience of researchers and the conventions of individual laboratories and companies. The quality control of rAAV vectors is a major challenge in developing gene therapy products, and pre-analytical procedures should be based on science- and risk-based approaches.
In this study, we investigated the impact of pre-analytical test sample preparation parameters on the results of rAAV genome titration using ddPCR to develop appropriate, scientific, and risk-based analytical procedures. Variations in the titration result for different serotypes, rAAV1, rAAV2, rAAV5, and rAAV6, were evaluated by varying the incubation temperature and time for each step of the DNase treatment, DNase inactivation, and ssDNA extraction. Optimization of proteinase K treatment may be considered in qPCR; however, this was not performed in this study because we decided to reduce the number of pre-analytical steps and heat treatments to develop a simpler and more robust analytical procedure. Therefore, considering its convenience and robustness, it is preferable to sufficiently inactivate the DNase and extract ssDNA from the capsid by heat treatment during the pre-analytical step.
Furthermore, we identified the critical temperature Tr at which ssDNA leakage starts from the rAAV capsids as an important characteristic for developing a pre-analytical procedure for rAAV genome titration via ddPCR. Using DSF, the Tr of all rAAVs used in this study was determined to be approximately 55°C, and a reduction in the number of GC as a result of ddPCR quantification was observed in the DNase inactivation step of the pre-analytical procedure at 50–65°C, regardless of the serotype. As Tr varies depending on the formulation [18], manufacturing process [19], and length of encapsidated ssDNA [20], Tr can be determined according to the characteristics of the rAAV-based product under development. The formulation of the drug product and pre-analytical treatment of the test sample for rAAV genome titration can be performed based on product analysis. As described in this study, various pre-analytical preparation procedures for VG titration by ddPCR, such as the direct embedding of AAV vectors into droplets [7, 14] and the addition of proteinase K to extract ssDNA [12, 13], have been developed. However, the importance of identifying Tr should be considered in testing method development for quality control of rAAV-based gene therapy products because of the heating of AAV vector samples during all pre-analytical methodologies, such as DNase treatment and PCR.
The incubation temperature during the DNase treatment to degrade free DNA had a small impact on the titration results because it was lower than Tr. During DNase inactivation, the incubation temperature, depending on the procedure, carries the risk of reducing the genome titer owing to ssDNA leakage close to Tr. Therefore, the Tr value is a strong target analytical parameter for genome titration. To obtain an accurate genome titer using ddPCR, the test samples were promptly heated to a markedly higher temperature than Tr for concurrent DNase inactivation and ssDNA extraction. Determining the temperature that causes ssDNA leakage from rAAV capsids is important for appropriate pre-analytical procedure development of genome titration via ddPCR, and the three-step pretreatment method is a simpler procedure with a lower risk of reducing rAAV genome titration during the pre-analytical preparation of test samples for quality control testing.
Testing methods that directly embed AAV vectors into droplets without prior ssDNA extraction have been previously demonstrated [7, 14]; however, aggregation of AAV particles, including empty capsids, affects VG titration results based on ddPCR [12]. Depending on the AAV serotype and formulation, aggregation can occur during manufacturing, storage, and pre-analytical treatment. Therefore, it may be preferable to extract ssDNA from AAV capsids before droplet generation to eliminate the influence of the sample conditions.
It is crucial to develop a testing method for vector genome titration based on science- and risk-based approaches for the quality control of rAAV-based gene therapy products. ddPCR is a powerful analytical tool; however, there is a risk of ssDNA leakage from capsids during analytical sample pretreatment. Thus, we propose that the ssDNA-releasing temperature Tr should be identified for each product to be developed, and operations for extended periods around this critical temperature should be avoided during pre-analytical sample preparation, particularly during DNase inactivation. The presented three-step pre-analytical procedure (condition examples: DNase treatment at 37°C for 30 min with 1 U of DNase and concurrent DNase inactivation and ssDNA extraction at 95°C for 10 min in 10 mM Tris buffer (pH 8.0) containing 10 mM EDTA, followed by PCR) aids in quantifying the vector genome titer more reliably via ddPCR.
Upon satisfactory compliance, data supporting the findings of this study can be provided by the corresponding author.
S.S. designed and performed the experiments. A.K. and K.U. supervised the study.
The authors declare no competing interests.
This study was supported by a grant-in-aid from the Project “Research and Development of core technologies for gene and cell therapy” supported by the Japan Agency for Medical Research and Development (AMED) (JP20ae0201002).
