(W1330-02-09) Using Cryo-TEM to Characterize Structural Details of LNPs during Manufacturing

Purpose: Successful lipid nanoparticle (LNP)-based vaccines and therapeutics require optimal stability, encapsulation efficiency, and therapeutic efficacy. This is achieved via careful consideration of the formulation components, but also the choice and optimization of the manufacturing processes as they can directly influence the LNPs’ physicochemical characteristics such as size, uniformity, payload distribution, and overall structure. Here, we show how direct visualization using cryo-TEM and image analysis techniques can be used to reveal differences on the per-particle level between LNP formulations that are prepared using four different mixing platforms, each with a unique design and operating principle. Cryo-TEM is able to show subtle differences in size and mRNA encapsulation, as well as a distinct difference in bleb formation depending on the mixing platform.

Methods: Grids were prepared undiluted using a Vitrobot (TFS). Electron microscopy was performed using a Thermo Fisher Scientific Glacios Cryo Transmission Electron Microscope (Cryo-TEM) operated at 200kV and equipped with a Falcon 3 direct electron detector. After identifying potentially suitable target areas for imaging at lower magnifications, high magnification images were acquired at nominal magnifications of 150,000x (0.096 nm/pixel), 73,000x (0.197 nm/pixel) and 28,000x (0.510 nm/pixel) using Leginon software^1,2. The images were acquired at a nominal underfocus of -5.5μm to -2.0μm and electron doses of ~10-25 e-/Å2. Image analyses were carried out both manually and using a custom AI-based segmentation algorithm using a minimum of 750 particles per dataset, or all particles available. The LNP formulation was a standard four lipid component system (MC3, Cholesterol, DSPC. DMG-PEG-2000) containing hEPO mRNA (N/P ratio = 6) in a final buffer of 20 mM Tris-HCl, pH 7.4. LNPs were prepared by Phosphorex using the following mixing systems: a T-mixer, the Ignite+ (Precision NanoSystems), an impingement jet mixer, and the Sunshine (Unchained Labs). The mixing flow rates on each of the four mixing platforms were optimized to achieve LNPs that had a similar size, uniformity, and mRNA encapsulation efficiency according to DLS and Ribogreen measurements.

Results: In general, the morphology and size of the LNPs prepared using the various mixing platfoms was comparable in the cryo-TEM images (see figure 1). To better quantify any differences between samples, image analyses methods were applied. The size of the LNPs generated using the Ignite +, the Jetmixer, and Sunshine, are very similar and their average size ranges from 60-63nm. However, particles generated using the T-mixer are slightly smaller and show an average size of 52nm with some fairly large outliers (see figure 2). The particle fraction that possibly contains mRNA is slightly higher for samples generated by the Ignite+ and Sunshine as compared to the T-mixer and Jetmixer (see figure 3). And surprisingly, all samples contained 40-46% of particles with bleb formations, except for the LNPs generated by the T-mixer. The T-mixer LNPs showed much fewer particles with blebs, namely only 27% (see figure 3).

Conclusion: In summary, subtle differences in size and payload distribution were observed using cryo-TEM. And most notably, a distinct difference in bleb formation between the LNPs prepared using the T-mixer and the LNPs prepared by the other three mixing systems was seen. Structural differences, including bleb formation, cannot be observed by DLS or other rapid in-line characterization methods. And although bleb formation in LNPs is still a topic of discussion in the field^3,4, they can make all the difference when accelerating screening studies during initial LNP development or when explaining potential variations in activity or potency observed during the scale-up process or later during manufacturing.

References: 1. Cheng et al, Protein Science, 2021, 30, 136-150
2. Suloway et al, J Struct Biol, 2005, 151, 41-60
3. Cheng et al, Adv. Mater. 2023, 35, 2303370. https://doi.org/10.1002/adma.202303370
4. Pratsinis et al, International Journal of Pharmaceutics, 2023, 637, 122874, https://doi.org/10.1016/j.ijpharm.2023.122874

Figure 1. Cryo-TEM images of LNP formulations prepared using a T-mixer, the Ignite+, an impingement jet mixer, and the Sunshine.

Figure 2. Size measurements of LNPs following image analyses of cryo-TEM images. Left: plots showing the probability distribution function (top) and cumulative distribution function (bottom. Right: Mean Area Equivalent Diameter (AED), range, and circularity calculated for each of the four LNP preparations (top); The average diameters calculated from cryo-TEM data compared DLS measurements (bottom).

Figure 3. Top: Cryo-TEM images of LNP formulations prepared using four different mixing systems. Bottom: The left table shows the percentage of particles in each preparation that had a mottled appearance, indicative of the presence of nucleic acid. The table on the right shows the percentage of particles in each preparation with bleb structures with ‘other’ including multivesicular or multilamellar structures.