(523b) Impact of Droplet Generation Principles on Protein Stability during Drying: A Comparative Study Using a Monodispersed Droplet Generator (MDG) and a Bi-Fluid Nozzle

Droplet generation is an important step during the spray drying of any liquid formulations, as it has a great influence on the properties of the dried particles For instance, it is a challenge to produce large particle sizes (> 50 um) by bi-fluid atomizer. Thus, alternative droplet generation technologies could be of interest when trying to produce large flowable particles via solution drying. Within this context, monodispersed droplet generators (MDGs) have gathered some interest¹. The MDG described in this study, DROPPO^® has been used previously to produce polymeric micro-particles. With this nozzle, droplets are generated using a piezo-electric transducer (PZT)². By adjusting the frequency and amplitude of vibration, droplets with different sizes can be formed and particles with different characteristics attained³. Using DROPPO^®, the main goal of this study was to compare MDG droplet generation with a conventional bi-fluid nozzle and understand the impact of these atomization principles on protein stability and particle properties.

Materials and Methods:

Materials:
A Protein X (ProtX) with a size of ~ 50 kDa, was kindly provided by Takeda (Vienna, Austria). The prepared formulations for spray drying contained ProtX at 20%ww. For the shear experiments one of the following saccharides: (same mass of ProtX and one saccharide) the disaccharide Trehalose dihydrate (TD) (Merck KGaA, Germany; Mw = 378.33 g/mol) and the cyclic oligosaccharide Hydroxypropyl β-cyclodextrin, Kleptose® HP ORAL GRADE (HβCD) (Roquette Frères, France; Mw = 1135 g/mol) was added to the 20%ww of protein solution. Before the experiments, all formulations were filtered with syringe filters (CHROMAFIL disposable filter A-45/25 cellulose mixed ester (MV) hydrophilic membrane, 0.45 μm 25 mm; Mercateo, Germany). Two different nozzles have been compared during this study, namely a conventional bi-fluid nozzle (4M8-TriX, ProCepT, Belgium) and DROPPO^® (MAAG GmbH, Germany).

Shear Experiments:
To gain indepth insight into possible induction of shear by the droplet generation method, the prepared ProtX-saccharide formulations with a solid content of 0.2 g/mL, have been fed at a pump speed of 60% (~4.8 g/min) and then collected at the nozzle outlet. The sizes of the nozzles were Ø 1.2 mm for the bi-fluid (at 0.6 bar) and Ø 0.08 mm for DROPPO^® (at 500 Hz frequency and 70% amplitude).

Spray Drying Experiments (SD):
The ProtX-formulations were dried on a lab-scale spray dryer (4M8-TriX, ProCepT, Belgium) operated in an open-air loop and equipped with a long drying chamber (2.1 m). Either a bi-fluid nozzle or the DROPPO^® was used for droplet generation. By applying an air inlet temperature of 150 °C and an airflow rate of 0.11 – 0.15 m³/min, an outlet temperature of 23 – 30°C for the bi-fluid nozzle and of 30 – 39°C for the DROPPO^® was obtained. The generated powders in the collector’s vessel were maintained at this outlet temperature for a maximum of 10 min. The moisture content, aggregation, and particle size of the dried protein powders were analyzed.

Size Exclusion Chromatography (SEC):
The phosphate buffer (pH= 7) was used as the mobile phase at a flow rate of 1 mL/min and the SEC column temperature was set to 20°C to separate the protein sample by size. The injection volume was 25 μL and a BIO-RAD Gel Filtration Standard (Bio-Rad Laboratories Ges.m.b.H., Austria) was used for the relative quantification of monomer, dimer, and aggregate species.

Karl-Fischer-Titration (moisture content):
The water content of the spray-dried protein powders (n = 3) was determined by Karl-Fischer titration (Titroline 7500 KF trace, SI Analytics, Germany) at 21.0 °C and 64.40 – 64.80 % relative humidity. 4 – 5 mg of powder were filled into HPLC vials, to which dry methanol mixed with dry formamide was added in a ratio of 1:1 (Aquastar® CombiMethanol, Merck KGaA, Germany) Extraction was performed for 1 hour at room temperature. Then the samples were transferred into the titrator cell and quantified.

Scanning Electron Microscopy (SEM):
For SEM analysis, particles were coated by sputter deposition with gold palladium and operated on a Zeiss Ultra 55 scanning microscope at 5 kV (Zeiss, Germany).

Results and Discussion:
ProtX-formulations with different compositions of either TD, HPβCD, or without any saccharides were examined in terms of shear effects and/or spray drying performance.

Shear experiments:
The SEC results showed, that both saccharides seem to aid in stabilizing the protein in a liquid state during the droplet generation step. In Figure 1A, it can be seen that, as the formulation is fed through the bi-fluid nozzle, aggregation slightly decreases, and when fed through DROPPO®, the aggregation slightly decreases even more, compared to the liquids before feeding through the nozzles. These results suggest that both alone and in the presence of saccharides the protein is stable during the droplet generation step and that DROPPO® could be a gentler technology to generate protein droplets for drying. Significant differences in aggregates were observed between ProtX alone and with HPβCD on DROPPO® (Figure 1A). Generally, the aggregation using DROPPO® was lower or comparable to the bi-fluid nozzle.

Spray Drying Experiments:
During lab-scale drying of aqueous 20%ww ProtX-formulations, it was observed that the drying capacity of the equipment used was not enough to assure the drying of the large DROPPO® droplets, and likewise, large losses were seen as well as high moisture contents (Table 1). However, as we aimed to understand how different droplet generation technologies affect protein stability and particle characteristics, further analyses of the powders were done via SEC and SEM. By comparing the two nozzles, it was observed that ProtX is stable when dried, irrespective of the droplet generation principle used (Table 1). After performing paired comparison analysis of aggregates and monomer contents, significant differences were observed between the two nozzle types, DROPPO® showed the formation of less aggregates, however, less monomer was found too when compared to the bi-fluid nozzle.

Concerning the particle size, optical microscopy analysis resulted in larger particles produced by the bi-fluid nozzle and smaller particles produced by DROPPO^®, which was verified by SEM (Figure 1 B and C), where it was observed that the particle sizes by DROPPO^® were between 500 nm – 30 µm, whereas the bi-fluid nozzle gave larger, spherical particles between 10 – 60 µm in diameter. This was surprising as larger particles were expected with DROPPO^® generating larger droplets. We hypothesize that smaller particles are a result of satellite droplets generated during the jet breakage, which are smaller in size after drying, and were probably the only particles able to be dried in our lab-scale equipment⁴. The SEM analysis showed, that DROPPO^® led to overall more spherical particles, some of which had a rough and some had a smooth surface. With the bi-fluid nozzle, mostly particle fragments with a very structured surface were observed.

Implications and outlook:

In this work, we showed that it is possible to produce stable protein powders with DROPPO^® with comparable performance to the bi-fluid nozzle in terms of stability. In the future, we aim to use equipment with higher drying capacities to adequately dry the droplets generated by this nozzle.

Literature:

(1) Schaefer, J.; Lee, G. Making Large, Flowable Particles of Protein or Disaccharide in a Mini-Scale Spray Dryer. Pharm. Dev. Technol. 2016, 21 (7), 803–811. https://doi.org/10.3109/10837450.2015.1063649.

(2) Wu, W. D.; Patel, K. C.; Rogers, S.; Chen, X. D. Monodisperse Droplet Generators as Potential Atomizers for Spray Drying Technology. Dry. Technol. 2007, 25 (12), 1907–1916. https://doi.org/10.1080/07373930701727176.

(3) Zhu, P.; Tang, X.; Wang, L. Droplet Generation in Co-Flow Microfluidic Channels with Vibration. Microfluid. Nanofluidics 2016, 20 (3), 1–10. https://doi.org/10.1007/s10404-016-1717-2.

(4) Ivey, J. W.; Bhambri, P.; Church, T. K.; Lewis, D. A. Experimental Investigations of Particle Formation from Propellant and Solvent Droplets Using a Monodisperse Spray Dryer. Aerosol Sci. Technol. 2018, 52 (6), 702–716. https://doi.org/10.1080/02786826.2018.1451818.