(M1430-01-05) Running at the Verge of Collapse in Lyophilization: What is the Impact of the Edge Vial Effect on Protein Stability and Particle Morphology?

Purpose: Lyophilization is a key process in the (bio)pharmaceutical industry for preserving sensitive biological products, such as proteins and vaccines.^{1, 2} The lyophilization process is inherently affected by a high degree of inhomogeneity, and different vials within the same batch experience different conditions during freeze-drying.³ One of the main sources of such variability is the so-called edge vial effect.It refers to the difference in product temperature (T_p) observed during a lyophilization process in vials located at the edges or corners of the shelf, i.e., vials with an incomplete hexagonal surrounding of neighbor vials, compared to those placed centrally with a full hexagonal surrounding. The edge vial effect is particularly strong during the primary drying phase of lyophilization, i.e., the phase where ice sublimation takes place, and has been related in previous research to either the number of neighbor vials or to radiation from the chamber walls and door.^4–7 Within this study, we want to assess if the risk of (sub)visible particle formation of a protein is different in edge vs. center vials, and if a different vial position within the batch results in a different morphology of the protein particles.

Methods: A model protein at 25 mg/mL was used in TrisHCl with 150 mM sucrose and 0.4 mg/mL polysorbate 20 (formulation F1). Formulation F2 additionally contained L-arginine hydrochloride. Both formulations were lyophilized within an Epsilon 2-10D freeze-dryer (Martin Christ, Osterode, Germany), applying either a very conservative, or a more aggressive protocol). The aggressive lyophilization protocol was chosen at the verge of collapse, as predicted by a mechanistic model combining heat and mass transfer equations (Figure 1). The lyophilization batches comprised vials having between three and six neighbor vials. Lyophilized cakes from both formulations, lyophilization processes, and all four neighbor vial conditions were evaluated regarding their cake volume fraction by an image-based approach. Besides, the residual moisture and the glass transition temperature of the dried material were determined by Karl Fischer titration and differential scanning calorimetry, respectively. After reconstitution, the vials were visually inspected for visible particles.⁸ Furthermore, changes in particle formation were investigated by dynamic light scattering (DLS), size-exclusion chromatography, and automated micro-flow imaging (MFI) measurements to screen from monomer to micron-sized particles. Obtained MFI particle images ≥5 µm were analyzed by ParticleSentryAI⁹ for the evaluation of the particle morphology. For the training of ParticleSentryAI models, an additional data set obtained after stir stress of both formulations F1 and F2 was added as a control.

Results: A clear correlation between vial position and primary drying time was observed, i.e., a vial with four neighbors dried about 30% faster than a vial with six neighbors. Interestingly, vials with four neighbors dried faster, and experienced higher product temperatures, than vials with 3 neighbors, due to the proximity of the plexiglass door of the lyophilizer. Both the residual moisture content and the glass transition temperature were not impacted by the lyophilization process and number of neighbors. The cake volume fraction was slightly affected by the lyophilization process in F1, with more pronounced shrinkage observed in the aggressive run, compared to the conservative cycle. Also, a slightly higher cake volume fraction was observed in the conservative run for F1 in vials with 4 neighbors than in vials with 5 or 6 neighbors (Figure 2A). In contrast, no significant trends were observed in F2 cake volume fractions (data not shown). In the reconstituted material, no differences were observed by visual inspection, size exclusion chromatography, and DLS when comparing different numbers of neighbor vials or different lyophilization processes. The micron-sized particle burden obtained by MFI was independent of the number of neighbor vials in F1 but was increased due to the more aggressive lyophilization process (Figure 2B). For F2, no clear differences were observed in the MFI data of different lyophilization processes/ numbers of neighbors (data not shown). Lastly, a possible impact of the edge vial effect on the particle morphology was evaluated by training several models with ParticleSentryAI. No differences depending on the lyophilization process and number of neighbor vials could be resolved (Figure 2C), indicating that the edge vial effect did not impact the particle morphology.

Conclusion: In this yet unpublished study, we investigated the impact of the edge vial effect during lyophilization on particle burden and particle morphology applying a mechanistic model to select an aggressive lyophilization process at the verge of collapse. Minor differences in cake volume and particle burden were observed depending on the lyophilization process. No impact of the edge vial effect on particle count and particle morphology was noted, suggesting that the different number of neighbor vials and the exposure to radiation does not significantly affect the extent or mechanism of subvisible protein particle formation.

References: 1. Randolph TW, Schiltz E, Sederstrom D, et al. Do not drop: mechanical shock in vials causes cavitation, protein aggregation, and particle formation. 2015. J Pharm Sci 104: 602–611.
2. Pardeshi SR, Deshmukh NS, Telange DR, et al. Process development and quality attributes for the freeze-drying process in pharmaceuticals, biopharmaceuticals and nanomedicine delivery: a state-of-the-art review. 2023. Futur J Pharm Sci 9: 35.
3. Rambhatla S, Pikal MJ. Heat and mass transfer scale-up issues during freeze-drying, I: atypical radiation and the edge vial effect. 2003. AAPS PharmSciTech 4: E14.
4. Ehlers S, Schroeder R, Friess W. Trouble With the Neighbor During Freeze-Drying: Rivalry About Energy. 2021. J Pharm Sci 110: 1219–1226.
5. Ehlers S, Friess W, Schroeder R. Impact of chamber wall temperature on energy transfer during freeze-drying. 2021. Int J Pharm 592: 120025.
6. Pikal MJ, Bogner R, Mudhivarthi V, et al. Freeze-Drying Process Development and Scale-Up: Scale-Up of Edge Vial Versus Center Vial Heat Transfer Coefficients, Kv. 2016. J Pharm Sci 105: 3333–3343.
7. Matejčíková A, Tichý E, Rajniak P. Experimental investigation of inhomogeneities of primary drying during lyophilization: Impact of the vials packing density. 2022. Journal of Drug Delivery Science and Technology 74: 103550.
8. Ph. Eur. 10.0. 2020. 2.9.20 Particulate contamination: Visible particles. In European Directorate for the Quality of Medicines (EDQM), editor. 10th European Pharmacopoeia (Ph.Eur.10), 10th ed., Stuttgart: Deutscher Apotheker Verlag.
9. Thite NG, Ghazvini S, Wallace N, et al. Machine Learning Analysis Provides Insight into Mechanisms of Protein Particle Formation Inside Containers During Mechanical Agitation. 2022. J Pharm Sci 111: 2730–2744.

Figure 1. Effect of chamber pressure and shelf temperature on the risk of process failure during primary drying of F1. Process conditions in blue result in a predicted risk of failure below 5%, which was computed assuming a maximum primary drying time of 50 hours. The selected, more aggressive primary drying protocol is highlighted by a black rectangle.

Figure 2. (A) Cake volume fraction obtained for F1 by conservative (plane bar) and aggressive (dotted bar) primary drying processes, for all neighbor conditions (3-6 neighbors, 3N-6N). (B) Particle concentrations (for particles ≥ 5 µm) detected by MFI in F1. (C) ParticleSentryAI similarities indices (0= similar, 1= dissimilar) in 6N vials after the F1 conservative lyophilization process (red), 4N vials after the F1 aggressive lyophilization process (green), and F1 vial after stir stress (blue).