Impact of specific productivity and operation mode upon the biophysical properties of HIV-1 Gag-based virus-like particles

Abstract

Virus-like particles (VLPs) are non-infective vaccine candidates that have gained interest given their natural ability to elicit strong immune responses. Particularly, HIV-1 Gag-based VLPs are one of the most described platforms for vaccine development, provided their ability for successful pseudotyping either by genetic engineering or click chemistry. When Gag polyprotein is recombinantly expressed, VLPs are naturally assembled in the vicinity of the cell membrane and then secreted by cell budding, taking part of the host cell membrane. Their properties are dependent upon the cell line and manufacturing method. Although great advancements toward the implementation of analytical methods have been made, VLP quality attributes are quite unclear whenever production is enhanced by metabolic engineering or process intensification strategies. This work offers a comparative study of VLP quality attributes upon transient gene expression (TGE) in HEK293 cell cultures operated in batch and perfusion mode. Moreover, the impact of specific productivity is also studied by ataxia telangiectasia mutated (ATM) gene silencing, which has been reported to enhance fourfold VLP production. A linear negative correlation was found between the ratio of Gag monomers/VLP and specific productivity. 3100 ± 100 monomers/VLP were obtained for the standard batch production, dropping to 1900 ± 100 and 800 ± 60 for the perfusion and batch ATM-knockdown conditions, respectively. Furthermore, functionalization rates were measured in terms of Cy5 per total particles (TP). Both perfusion-derived nanoparticles achieved functionalization rates of 2800 Cy5/TP. On the contrary, those nanoparticles produced in batch yielded functionalization rates below 1000 Cy5/TP. Moreover, a complete lipidome analysis revealed a relative decrease in the quantity of lipid/particle for all studied conditions in comparison to the standard batch production. Finally, all VLP samples were characterized to assess the impact of the differential physicochemical properties upon purification and stability rates. KEY POINTS: • VLP quality inversely correlates with Gag-specific productivity and operation mode. • Functionalization and lipid content drop with metabolic burden or ATM silencing. • Perfusion enables high VLP recovery and lyophilization with preserved morphology.

Keywords: ATM; Bioprocess development; Functionalization; HEK293; Lipidomics; Perfusion; TGE; Virus-like particles (VLPs).

Fig. 1

Experimental design and workflow of the experiments performed in this work. VLPs were harvested for both batch and perfusion productions and then subjected to physicochemical characterization tests. Resistance to ultracentrifugation, monomer/VLP, and membrane protein/TP were quantified. Structural differences were assessed using Cryo-TEM. Furthermore, lipidomic analysis of ultracentrifuged VLP samples was done for each studied group. A final purification comparison study was performed for supernatants coming from perfusion. VLP, virus-like particle; TEM, transmission electron microscopy

Fig. 2

Summarized data upon monomer/VLP determination for each group. A Free unassembled Gag monomer to VLP ratio for each studied condition. VLP concentrations are directly measured by NTA. B Size and particle distribution of VLPs produced in batch measured by NTA. C Size and particle distribution of VLPs produced in perfusion measured by NTA. D Monomer/VLP ratio of each studied group calculated by ELISA. E Correlation between monomer/VLP and average specific production calculated for each study group. Significance was calculated using one-way ANOVA and Tukey’s test. NTA, nanoparticle tracking analysis; VLP, virus-like particle; ELISA, enzyme-linked immunosorbent assay

Fig. 3

Particle functionalization. A Click chemistry assay for each studied group expressed as Cy5 molecules/total particle. In this case, click chemistry is used as an approximation of membrane proteins/TP. B Compilation of pictures taken by SRFM for each Cy5 functionalized group, where red and green fluorescence correspond to Cy5 and VLPs, respectively. VLP, virus-like particle; EV, extracellular vesicle; TP, total particles; SRFM, super resolution fluorescence microscopy

Fig. 4

Lipidomic study performed upon ultracentrifugation for all four groups. A Relative abundance of all studied lipids expressed as a percentage of each quantified lipid in respect to the total sum of lipids for each sample. B Total quantified lipids measured in pmol equivalent normalized by protein and/or number of particles. C Fold change (FC) of each studied lipid in respect to batch Gag condition. Fold changes were normalized as log₂ considering batch Gag as control. Lipids were arbitrarily classified depending on their relative abundance in three main groups depicted above. Significance was calculated using one-way ANOVA and Tukey’s test

Fig. 5

VLP purification. A Schematic representation of the workflow conducted from purification to lyophilization. B DLS profile of the pGag::eGFP-loaded supernatant and eluted fraction containing a major nanoparticle population at 140 nm. C DLS profile of the purified pGag::eGFP eluted fraction and lyophilized sample containing a major nanoparticle population at 140 nm. D HPLC profile of the purified pGag::eGFP VLP fraction and its corresponding lyophilization. E DLS profile of the pGag::eGFP-shATM-loaded supernatant and eluted fraction containing a major nanoparticle population at 140 nm. F DLS profile of the pGag::eGFP-shATM-loaded supernatant and eluted fraction containing a major nanoparticle population at 140 nm. G HPLC profile of the purified pGag::eGFP-shATM VLP fraction and its corresponding lyophilization. VLP, virus-like particle; DLS, dynamic light scattering; HPLC, high performance liquid chromatography; Abs, absorbance

Fig. 6

Schematic representation of the physicochemical parameters for all studied VLPs. This is a simplification of the internal structure of the VLPs. It does not imply the presence of a regular Gag lattice. It serves as a conceptual aid rather than a structural model.TP, total particle; VLP, virus-like particle; eq, equivalent (pmol of equivalent normalized by number of total particles)