(M1330-08-49) CryoTEM Characterization of Vaccine Materials in Development

Purpose: The increasing complexity of vaccine materials necessitates characterization using a multitude of analytical methodologies to obtain sufficient product understanding. We have applied cryo transmission electron microscopy (cryoTEM) to assess distribution of surface proteins on a live virus vaccine candidate that affects viral particle biodistribution, replication in vivo, and overall immune responses, as well as structural changes in the viral particles after thawing from the frozen storage state.

Methods: In one study, a virus harvest fluid, containing recombinant vesicular stomatitis virus (rVSV) displaying the SARS-CoV-2 spike (s) protein on its surface, was treated with an endonuclease to digest host cell nucleic acids, column purified (CP), ultrafiltered, and concentrated with a default pump (UFP-1) or a low shear pump (UFP-2). CryoTEM images of the three process intermediates were acquired, and fraction counting analysis categorized the viral particles as densely covered (multiple stretches of membrane 50nm in length that contains 3 or more spike proteins) or less densely covered with surface spikes. In a second study, rVSV vaccine materials stored for the same duration at 2-8^oC (refrigerated) and -70^oC and then thawed (F/T) were imaged neat or after concentration via ultracentrifugation.

Results: CryoTEM showed more particle damage in UFP-1 than CP, in the form of nucleocapsids without a surrounding membrane or undergone uncoiling, as well as less densely decorated with surface spikes. UFP-2 contained more particles displaying a denser layer of surface spikes and less particle damage compared to UFP-1. Fraction counting analysis confirmed this trend, yielding 73%, 58% and 77% spike protein density for CP, UFP-1 and UFP-2. Thus usage of a low shear pump for viral particle concentration correlates with reduced particle damage and surface spike protein loss.

CryoTEM showed that the refrigerated, pre-concentration rVSV material contained bullet-shaped particles typical of VSV. The F/T, pre-concentration material consisted of a mixture of bullet-shaped and round particles. After concentration, the refrigerated material had mostly bullet-shaped particles but also a very small number of round particles. The equivalently treated F/T material contained a mixture of bullet-shaped and round particles. This finding demonstrates that although sample concentration via ultracentrifugation and/or resuspension of the resultant pellets slightly reduced the viral particle integrity (round morphology), particle damage was already present in the pre-concentration F/T material.

Conclusion: These case studies illustrate that cryoTEM elucidation of the ultrastructure of viral particles can be used to determine process-induced impact on vaccine materials and guide process development and improvement during vaccine development.

Acknowledgements:We thank our partner IAVI, Brooklyn, NY, for their collaboration on this work; Annette Schneeman and partners at NanoImaging Services for the fraction counting analysis and imaging of some rVSV samples; and Josef Vlasak, Andrea Gomez, Christopher Tubbs, Matthew Troutman, and Douglas Richardson for their help with cryoTEM imaging method development and helpful discussions.

Figure 1. CryoTEM of rVSV process intermediates CP (a to c), UFP-1 (d to f), and UFP-2 (g to i). See the text for sample detail. All three samples consisted of mostly bullet-shaped particles, but CP and UFP-2 contained fewer nucleocapsids without a surrounding membrane or had undergone uncoiling, as well as a denser layer of surface spike proteins, than UFP-1. Scale bars: 500 nm in a, d, and g; 200 nm in rest of the figures.

Table 1. Fraction counting analysis of bullet-shaped particles densely covered with spike proteins in rVSV process intermediates. See the text for sample detail. Particles categorized as densely covered (multiple stretches of membrane 50 nm long containing ≥3 S proteins) or less densely covered with the protein. UFP-2, generated by concentration of CP material using a low shear pump, contained a higher percentage of particles densely covered with spikes compared to UFP-1, generated using a default pump. UFP-2 and CP had comparable percentages of particles densely covered with spikes.

Figure 2. CryoTEM images of a rVSV vaccine material subjected to refrigeration (refrigerated) or a freeze-thaw (F/T) cycle. The material was imaged neat (pre-concentration, a to d) or after concentration by ultracentrifugation (post-concentration, e to h). See the text for sample detail. After concentration, the refrigerated material had mostly bullet-shaped particles (red arrow) typical of VSV, but also a few round particles (blue arrow). The F/T material consisted of a mixed population of bullet-shaped and round particles even before sample concentration. Empty vesicles (green arrow) and un-enveloped and unwound nucleocapsid (orange arrow) were also present in some of the samples. Scale bars: 500 nm.