Forced degradation analysis of recombinant adeno-associated virus serotype 8 based on analytical anion exchange chromatography coupled to orthogonal characterization

In the present study, we have successfully applied a rapid and robust analytical AEX method based on a novel strong anion-exchanger (quaternary ammonium) column for the assessment of rAAV8 quality during forced degradation. To elucidate the role of deamidation in forced degradation, peptide-mapping technology based on liquid chromatography–tandem mass spectrometry (LC–MS/MS) was employed to monitor the deamidation rate under accelerated storage. In addition, charge detection mass spectrometry (CDMS) was used to facilitate the interpretation of the chromatographic variations identified using AEX. We identified an association between an increase in the deamidation rate and a prolongation of the retention time using the analytical AEX method. Moreover, the potential of this method was further explored by applying it to the monitoring of degradation under a wide range of pHs (pH 2.5–9.5) and at a relatively high temperature (40 °C). The chromatograms showed that the degradation of rAAV8 differed at differing pH, implying that it may have influenced the capsid proteins, the encapsidated genome, or both. Furthermore, several other techniques were used to characterize the effects of accelerated storage on the transgene and capsid proteins of rAAV8. For instance, the titer was determined by digital polymerase chain reaction (dPCR), and genome integrity was evaluated by capillary-gel electrophoresis using laser-induced fluorescence (CE-LIF). In addition, VP composition was determined using sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE), and particle mass and particle binding/unbinding behavior were characterized using mass photometry (MP). Subsequently, an in vitro transduction efficiency assessment of rAAV8 was conducted. These physicochemical approaches revealed that, during accelerated storage, the loss of unique VP1 and VP1/VP2 common regions, and genome fragmentation and release, predominated at lower pH, whereas deamidation and structural variation were the major forms of the degradation at higher pH. Thus, in the present study, we have used AEX coupled to orthogonal characterization to study the pH-dependency of rAAV8 degradation pathways and advanced understanding of rAAV stability.

Materials and methods

Analytical AEX

The AEX method was performed using a Thermo Scientific Vanquish UHPLC system coupled to a fluorescence (FLR) detector (excitation wavelength, 280 nm; emission wavelength, 350 nm). A strong anion exchange (quaternary ammonium) test column (50 × 4.6 mm internal diameter; YMC, Kyoto, Japan) was used for the separation at room temperature. Sample vials were placed at 4 °C in the autosampler, and each injection contained 2.5–5.0 × 10¹⁰ viral particles (vp) for intact sample without stress treatment. Thermo Scientific Chromeleon 7 software was used for data processing and analysis.

Peptide mapping by LC–MS/MS

An Ultimate 3000 LC pump (Thermo Fisher Scientific) coupled to a Q Exactive HF-X mass spectrometer (Thermo Fisher Scientific) was used for LC–MS/MS) analysis of the sample. The peptides were separated using an Acquity UPLC Peptide CSH C18 Column (130 Å, 1.7 μm; 1.0 × 150 mm; 1.7-μm particle size) at 45 °C. Mobile phase A consisted of water containing 0.1 v/v % formic acid (FA), and mobile phase B consisted of acetonitrile containing 0.1 v/v % FA. Separation was performed using a gradient of 5–35 % mobile phase B over 45 min at 50 μL/min. A full MS scan was performed using an ion transfer tube temperature of 250 °C, a resolution of 120,000, a mass range (m/z) of 300–2000, and a funnel radio frequency (RF) level of 40. A subsequent data-dependent MS/MS scan was performed using a higher energy collisional dissociation of 27 % and a resolution of 30,000. The peptides were identified and the deamidation rates were quantified using a Byos v.5.6.68 (Protein Metrics, Cupertino, CA).

Charge detection mass spectrometry (CDMS)

Twenty microliters of each sample was directly buffer exchanged into 200 mM aqueous ammonium acetate solution (Invitrogen, AM9070G) with 0.01 % pluronic F-68 (ThermoFisher Scientific, 24040032), adjusted to pH9.0, using Micro Bio-Spin P-6 gel columns (Bio-Rad, 7326221). Mass analysis was performed using a Waters Xevo™ CDMS instrument with an electrostatic linear ion trap (ELIT). Ions were generated by positive mode nanoelectrospray ionization and were subsequently trapped for 100 ms, with a charge RMSD of ∼0.9 e. Signal processing and visualization were performed using waters_connect™ CDMS Toolkit software. Time-domain signals were Fourier transformed, where the measured frequency and magnitude corresponded to an individual ion’s m/z and z, respectively, enabling direct mass calculation. Individual ion data were binned into histograms of m/z, charge, and mass spectra, along with 2-dimensional heat maps.

Results and discussion

Investigation of the chromatographic retention time prolongation

Samples of rAAV8 EPs and FPs from a new batch were stored at 40 °C for 0, 3, or 7 days and then analyzed using AEX and LC–MS/MS-based peptide-mapping (Fig. 2). The trends of a prolongation of retention time and a decrease in peak area on the AEX elution profile were reproduced for the new batch (Fig. 2A and B). The levels of several modifications (deamidation, phosphorylation, and oxidation) were evaluated, and no trends were observed regarding oxidation or phosphorylation (data not shown). Consistent with the widely reported deamidation of asparagine (N) residues, generating acidic species,^23,34 there were four asparagine residues that were deamidated at rates of >1 % (N57, N66, N94, and N263) in the FPs (Fig. 2C) and EPs (Fig. 2D), and there were significant increases in deamidation at all four sites as the storage time was extended. In particular, N57 exhibited a deamidation level of as high as 28.3 % for FPs and 72.4 % for EPs after 7 days of storage at 40 °C.

In addition, CDMS analysis was performed after storage for 14 days at 40 °C, yielding the charge distributions shown in Fig. 2E and F and the mass plots in Supplementary Fig. S2. Consistent with the greater deamidation, the representative charge number of the EPs (Fig. 2F) decreased from 157.6 to 149.6 after this period of storage. However, this decrease in charge during storage did not occur for the FPs (Fig. 2E), suggesting that deamidation might be the cause of other physicochemical changes. This is discussed below.

Journal of Pharmaceutical Sciences, 2026, 104255, Volume 115, Issue 5: Fig. 2. Investigation on chromatographic retention time prolongation. AEX elution profiles (baseline calibrated) of rAAV8 full particle and empty particle after accelerated storage were shown in A) and B). Deamidation levels after accelerated storage identified by peptide mapping analysis of rAAV8 full particle and empty particle were shown in C) and D). Charge distributions of rAAV8 empty particle and full particle after accelerated storage were shown in E) and F).

Larger increases in deamidation at all four of the asparagine residues were identified for the EPs than for the FPs, which corresponded to greater prolongation of the chromatographic retention for the EPs (Fig. 2B) than for the FPs (Fig. 2A). Of these four residues, N57 and N263 were followed by a glycine (G) residue, whereas N94 and N66 were followed by a histidine (H) residue and an alanine (A) residue, respectively. It is known that the N+1 residue has an essential role in the local flexibility of the backbone and the deamidation rate; for instance, the NG motif has the greatest impact, followed by histidine and serine residues.^35,36 In addition, higher-order structure, solvent accessibility, and the external conditions, including temperature and buffer pH, have effects.^23,37

The deamidation of N263, which is located within the flexible loop regions of the rAAV8 capsid,²⁴ has previously been suggested to be the principal cause of the prolongation.²⁶ In contrast, N57 and N94 are located within VP1-unique regions that are buried inside the capsid and have incompletely resolved structures.³⁸ However, we could not exclude the effects of the deamidation of N57 and N94 of rAAV8 on the retention time, and higher temperatures could cause the VP1u and VP1/2 common regions to be exposed.³⁹ Given the close correlation between capsid deamidation and retention time, the monitoring of rAAV8 using AEX during forced degradation storage can be considered to be a straightforward and efficient preliminary method of characterization.

Application of AEX analysis to forced degradation

To explore the potential utility of the analytical AEX methodology, an in-depth investigation of the effect of pH on rAAV stability was conducted by storing rAAV8 FP samples under multiple buffer conditions at 4 °C or 40 °C for 0, 1, 3, 7 or 14 days. The study of the effect of pHs between 2.5 and 9.5 was designed to simulate the proton transfer equilibrium during rAAV manufacturing and storage. For example, the elution conditions for AC usually involve a pH of 2.5, and pH 9.0 is common during cell lysis and AEX polishing. In addition, pH values of 3–8 are usually evaluated to select the optimal storage buffer for rAAV.^40,41

Conclusion

In the present study, an analytical AEX method was applied and demonstrated to be a sensitive tool for the monitoring of surface property-related changes in rAAV8 under accelerated storage conditions. We identified a close association between a prolongation of the retention time and progressive deamidation of rAAV8 capsid proteins at the N57, N94, and N263 residues. Extension of the AEX method to the characterization of pH-dependent forced degradation indicated the existence of distinct degradation pathways, which were elucidated by comprehensive physicochemical characterization of the integrities of the capsid proteins and encapsidated DNA cargo.

Under neutral and basic conditions, the prolongation of retention time and decrease in peak area were primarily associated with deamidation-related changes in surface charge, aggregation and nonspecific adsorption, which are likely to be the result of the externalization of the VP1/VP2 common region. In contrast, acidic conditions (low pH) were associated with a distinct type of degradation, which involved fragmentation, the release of encapsidated DNA, and the cleavage of viral protein. Notably, rAAV8 demonstrated remarkable stability at relatively high temperature in a pH 5.5 citrate buffer, with preservation of capsid integrity and the vector titer, while a certain amount of DNA fragmentation, underlining the importance of optimizing the buffer environment with wide considerations. Furthermore, functional assessments confirmed that the loss of infectivity is multifactorial, being affected by VP1/VP2 degradation and DNA fragmentation, rather than by deamidation alone, highlighting the need to use a comprehensive set of methodologies for rAAV quality assessment. Overall, the present findings regarding forced degradation analysis provide valuable insight into the optimization of the rAAV manufacturing process, formulation, and stability assessment.