Measurement of Adenovirus-Based Vector Heterogeneity

Abstract

Adenovirus vectors have become an important class of vaccines with the recent approval of Ebola and COVID-19 products. In-process quality attribute data collected during Adenovirus vector manufacturing has focused on particle concentration and infectivity ratios (based on viral genome: cell-based infectivity), and data suggest only a fraction of viral particles present in the final vaccine product are efficacious. To better understand this product heterogeneity, lab-scale preparations of two Adenovirus viral vectors, (Chimpanzee adenovirus (ChAdOx1) and Human adenovirus Type 5 (Ad5), were studied using transmission electron microscopy (TEM). Different adenovirus morphologies were characterized, and the proportion of empty and full viral particles were quantified. These proportions showed a qualitative correlation with the sample's infectivity values. Liquid chromatography-mass spectrometry (LC-MS) peptide mapping was used to identify key adenovirus proteins involved in viral maturation. Using peptide abundance analysis, a ∼5-fold change in L1 52/55k abundance was observed between low-(empty) and high-density (full) fractions taken from CsCl ultracentrifugation preparations of ChAdOx1 virus. The L1 52/55k viral protein is associated with DNA packaging and is cleaved during viral maturation, so it may be a marker for infective particles. TEM and LC-MS peptide mapping are promising higher-resolution analytical characterization tools to help differentiate between relative proportions of empty, non-infectious, and infectious viral particles as part of Adenovirus vector in-process monitoring, and these results are an encouraging initial step to better differentiate between the different product-related impurities.

Keywords: Adenovirus-based vaccine; Analytical characterization; Critical quality attributes; In-process testing; Mass spectrometry; Transmission electron microscopy.

Figure 1

Schematic of the processes that are involved in the formation and maturation of adenovirus particles. Capsids are first assembled with one vertex acting as a portal. Packing proteins associate with the portal vertex and the viral DNA, transporting the DNA into the assembled capsid. Maturation to an infectious adenovirus particle occurs through cleavage of key adenovirus proteins by adenovirus protease (AVP) and condensation of the viral DNA. Images Created with BioRender.com.

Figure 2

Characterization of appearance and composition of Ad5 samples during CsCl ultracentrifugation or TFF-AEX purification. (A) Flow diagram of purification of Ad5 fractions. (B) Representative TEM micrographs of Ad5 samples taken from high density fraction of 2nd CsCl spin and eluant from AEX with 100 nm scale bars. (C) Representative RP-UHPLC chromatograms of CsCl (black trace) or AEX (blue trace) purified Ad5. The Ad5 protein comprising each peak were identified through LC-MS peptide mapping.

Figure 3

Characterization of appearance and composition of ChAdOx1-GFP samples during CsCl purification. (A) Flow diagram of ChAdOx1-GFP fractions. (B) Representative TEM micrographs of LDF or HDF ChAdOx1-GFP from the 1st or 2nd CsCl purification step with 100 nm scale bars. (C) Representative RP-UHPLC chromatograms of CsCl purified ChAdOx1-GFP. The first or second CsCl low density fractions (LDF) are shown in green or red, respectively, while the first or second CsCl high density fractions (HDF) are shown in black or blue, respectively.

Figure 4

Comparison of relative protein quantification approaches of LC-MS peptide map datasets from ChAdOx1-GFP samples from purification process. (A) Relative RP-UHPLC peak comparison between the different ChAdOx1-GFP CsCl fractions. RP-UHPLC peak abundancies in each CsCl fraction were normalized to the Protein IX peak. Error bars represent triplicate MS measurements. The asterisks (*) denotes a peak not observed in the HDF samples. (B) A radar plot visualization of the same datasets in panel (A). Proteins indicated in grey are below the limit of quantitation. (C) Relative peptide abundance for representative viral proteins in the ChAdOx1-GFP CsCl fractions when compared pairwise to the 2nd high density fraction. Box represents interquartile range, error bars represent range, dots represent outliers and X represents mean. Number of peptides compared is dependent on samples and protein, but n >4.

Figure 5

ChAdOx1 nCoV-19 peptide analysis. Relative peptide abundance for representative proteins in the ChAdOx1 nCoV-19 samples as measured by LC-MS peptide mapping, Samples analysed were taken from one large-scale AEX preparation and two different lab-scale CsCl preparations with different P:I ratios and results were compared pairwise for the ratio of peptide abundance for four different viral proteins. 3 different comparisons are shown: blue boxes - Large-scale AEX (P:I ratio 97) and Lab-scale CsCl 1 (P:I ratio 68); orange boxes - Large-scale AEX (P:I ratio 97) and Lab-scale CsCl 2 (P:I ratio 31); grey boxes - Lab-scale CsCl 1 (P:I ratio 68) and Lab-scale CsCl 2 (P:I ratio 31). Box represents interquartile range, error bars represent range, dots represent outliers and X represents mean. Number of peptides compared is dependent on samples and protein but always >4.