Read time: 6 minutes

Certain next-gen vaccines and gene therapy applications rely on the usage of adeno-associated viruses (AAV) as a delivery vehicle. To ensure the safety and efficacy of viral vector drugs, multiple critical quality attributes (CQAs) need to be well characterized.

In a recent webinar, available on-demand, Jane Luo (PhD), Senior Scientist at SCIEX, demonstrated how capillary electrophoresis (CE) was used to assess multiple CQA analysis of AAV using a single platform. She shared data acquired on nucleic acid loading, viral protein profiles, ratio, and titer, empty and full capsid ratio and residual nucleic acids from the same samples on a single platform within one day.

In this blog, I summarize Jane’s answers to the most pressing questions to help you improve your AAV analysis. For further information, check out the viral vector solution guide or contact us.

Experimental setup

How much sample do you need to perform the comprehensive AAV analysis with CE?

Jane Luo: We used 30 µL to 40 µL of the sample with a titer in the range of ~1 x 10¹¹ to ~1 x 10¹³   gene copies (GC)/mL. You can find further information in this technical note.

Do you need eight points for the calibration curve to determine sample titers?

Jane Luo: The short answer is no; you do not need to have eight points. From my perspective, there are two aspects associated with this question. One is linearity and the other one is detection range. For establishing linearity, the ICH guidelines recommend using a minimum of five concentration endpoints. For the dynamic range, the ICH guidelines recommend the range to cover the test concentration and up to 30% above and below the test concentration. The method I presented covers about three orders of magnitude, which is suitable for most AAV samples.

Can you elaborate on the different time requirements for the capsid protein analysis and the genome analysis?

Jane Luo: For the genome analysis, an extraction of nucleic acids from the AAV samples is necessary, which is not the case for the capsid protein assay. This sample preparation adds a bit of time to the overall assay time for the genome analysis.

When doing the comprehensive AAV analysis (capsid protein analysis and genome analysis within a day), my recommendation is to start the genome analysis and then work on the capsid analysis. This arrangement saves time, and you can use a single CE platform to perform all analyses for these critical quality attributes (CQAs) within a typical workday.

Data processing

Do you need specific software for calculating the full and empty ratio?

Jane Luo: You do not need special software to calculate the full and empty ratio. It is a simple division of the genome titer by the capsid titer. A standard calculator or Microsoft Excel will be fine.

Which signal-to-noise ratio did you use for determining the limit of quantitation (LOQ)?

Jane Luo: We follow the ICH guidelines. For the determination of the LOQ, the signal-to-noise ratio was ten or slightly above ten.

How do you determine the corrected peak area? Is it an automatic determination by software?

Jane Luo: Yes, it is done automatically. The BioPhase 8800 system comes with an intuitive software that lets you preset analysis parameters. The corrected peak area is then automatically determined.

If there is a migration time difference between different serotypes, does this impact the corrected peak area?

Jane Luo: The corrected peak area is not affected by migration time changes. The software is great in doing the automated corrected peak area analysis and will correct for migration time shifts.

What is the linear dynamic range for the genomic titer determination by CE?

Jane Luo: The linear range for the method I presented range from 1×10¹⁰ GC/mL to 2×10¹³ GC/mL. This is a linear dynamic range of 3.3 orders of magnitude and was suitable for the samples we analyzed.

Data results

You mentioned a VP variant peak. Can you explain what this peak is?

Jane Luo: Absolutely. The VP3 variant peak, sometimes also called VP3 prime (VP3’), is a shorter version of the VP3 protein that is derived from an alternative translation initiation site. For further information on the identification of the VP3 prime variant, I suggest the following publication from a research group in Japan (Human gene therapy. 32(21–22):1403-1416). This paper explains that the translation initiation site at M203 can be skipped and an initiation at M211 led to an eight amino acid shorter protein. As you can see from my results, the VP3 variant is very well separated from VP3, demonstrating the great resolution of the CE method.

In addition, the detection of a VP3 fragment with LC-UV-MS was described. This fragment was linked to the hydrolysis of the VP3 at the C terminus, caused by the low pH of the mobile phase and the high temperature used for the column oven during the liquid chromatography analysis, which are standard settings needed for LC-UV-MS analysis.

Scientists using LC for analyzing capsid proteins should be aware of this and might want to consider complementary evaluation with CE.

You mentioned that the additional peaks in your genome analysis with shorter migration time may be partial genomes. Did you confirm their ID?

Jane Luo: We have not done sequencing analysis to confirm IDs ourselves. These peaks being partial genomes is an assumption that is consistent with published literature that used PCR-based methods for confirmation.

CE in comparison to other techniques

Are you aware of full-and empty capsid assessments using ratios based on ELISA and PCR results? Can you comment on the differences between that method and yours?

Jane Luo: Yes, there are publications for which a ratio calculation based on capsid titer from size exclusion chromatography (SEC) or ELISA and the genome titer from qPCR was used for determining full and empty ratios. These ratios rely on two vastly different methodologies, and therefore, data will have compounded variability. That is one downside to consider. In addition, the genome titer from qPCR methods often targets only the regions of inverted terminal repeat (ITR) sequences. This can lead to overestimating the genome titer since capsids with partial sequences or partial genomes, which contain the ITR but not the gene of interest, will be considered. With the presented method, we clearly separate the intact genome from the partial genome and the small size impurities, and therefore avoid overestimation of the genome titer.

Can you elaborate on the time needed for different analytical techniques in comparison to your method?

Jane Luo: The exact time requirements for techniques, such as PCR/ELISA, electron microscopy (EM), analytical ultracentrifugation (AUC), etc., will depend on the specific setup being used. I presented estimates for each technique in my webinar, which are based on published literature.

From the comparison, you can see that it will take 2-3 workdays to perform comprehensive CQA analysis using a combination of techniques and instrumentation, while these parameters can be assessed within a typical workday using a single CE platform instead.

Certain PCR workflows use DNAse or benzonase treatments before samples are analyzed. Do samples need any pre-processing steps prior to CE analysis?

Jane Luo: My recommendation is to do a simple extraction of the nucleic acid using commercially available kits and heat the sample to avoid secondary structure formation prior to CE analysis. Since we do not need to rely on amplification for CE-based genome integrity analysis while achieving high sensitivity with fluorescent dye and laser-induced fluorescence detection, a pre-processing step is not required. You can add a benzonase treatment step and subsequent inactivation and removal of benzonase before nucleic acid extraction. Comparing the results of benzonase-treated to non-treated samples helps to decipher the amount and size range of nucleic acid impurities present outside of the AAV capsid.