Quantification of Empty, Partially Filled and Full Adeno-Associated Virus Vectors Using Mass Photometry

Abstract

Adeno-associated viruses (AAV) are one of the most commonly used vehicles in gene therapies for the treatment of rare diseases. During the AAV manufacturing process, particles with little or no genetic material are co-produced alongside the desired AAV capsid containing the transgene of interest. Because of the potential adverse health effects of these byproducts, they are considered impurities and need to be monitored carefully. To date, analytical ultracentrifugation (AUC), transmission electron microscopy (TEM) and charge-detection mass spectrometry (CDMS) are used to quantify these subspecies. However, they are associated with long turnaround times, low sample throughput and complex data analysis. Mass photometry (MP) is a fast and label-free orthogonal technique which is applicable to multiple serotypes without the adaption of method parameters. Furthermore, it can be operated with capsid titers as low as 8 × 10¹⁰ cp mL^-1 with a CV < 5% using just 10 µL total sample volume. Here we demonstrate that mass photometry can be used as an orthogonal method to AUC to accurately quantify the proportions of empty, partially filled, full and overfull particles in AAV samples, especially in cases where ion-exchange chromatography yields no separation of the populations. In addition, it can be used to confirm the molar mass of the packaged genomic material in filled AAV particles.

Keywords: adeno-associated virus vectors; analytical ultracentrifugation; genomic cargo; partially filled particles; single-molecule mass photometry.

Figure 1

Schematic illustration of a mass photometry measurement. The attachment and detachment of AAVs onto a glass slide result in an interferometric light scattering at the solid–liquid interface. Full and empty particles are visualized in the microscopic image, where white and black circles represent filled and empty AAVs, respectively.

Figure 2

Native microscope images of the solid–liquid interface (a) before and (b) after cleaning with isopropanol, (c) overcrowded with and (d) containing an adequate number of AAV particles.

Figure 2

Native microscope images of the solid–liquid interface (a) before and (b) after cleaning with isopropanol, (c) overcrowded with and (d) containing an adequate number of AAV particles.

Figure 3

Linear correlation of the measured percentage-filled AAV capsids using mass photometry and expected percentage-filled AAV capsids by AUC.

Figure 4

(a) Linear correlation between binding count rate and capsid titer (determined with ELISA) for sample concentrations between 4 × 10¹⁰ cp mL⁻¹ and 8 × 10¹¹ cp mL⁻¹. Sample concentrations ≥ 2.5 × 10¹² cp mL ^–1 result in a significant decrease in the binding count rates. (b) 3D plot of measured dilutions.

Figure 5

The mass distributions of three AAV serotypes ((a) AAV5, (b) AAV8, and (c) AAV9) comprising different proportions of AAV subpopulations. E, PF, F and OF stand for empty, partially filled, full and overfull, respectively. Samples comprising meC (orange) and mfC (green), respectively, were selected for each serotype to set the AAV limits for the analysis of an AAV sample (blue).

Figure 6

Evaluation of the MP performance using three different in-house produced serotypes (a) AAV5, (b) AAV8 and (c) AAV9. Obtained MP results (purple) were compared to the AUC data (blue).