Purpose: Thiostrepton (TST) is a natural compound with potent activity against gram-positive pathogens, including Methicillin-resistant Staphylococcus aureus (MRSA). Despite being known for over a century, TST has a limited clinical application due to its low water solubility and lack of intestinal absorption. To overcome these barriers, we encapsulated TST in DSPE-PEG2000 sterically stabilized micelles (SSM) for safe and effective drug delivery via intravenous administration. Using the conventional thin-film hydration method, we encapsulated up to 5 TST molecules per micelle but found this approach unsuitable for large-scale production. As an alternative, we employed the co-solvent freeze-drying method but obtained only one drug molecule per micelle. Therefore, the purpose of this study was to identify optimum conditions to allow maximum drug encapsulation of TST in SSM by the co-solvent freeze-drying method.
Methods: We prepared TST-SSM by the co-solvent freeze-drying method varying the proportion of tert-butanol in the co-solvent system (50% or 75%), reconstitution method (spontaneous, swirling for 5 minutes, or sonication for 5 minutes), presence of phosphate-buffered saline (PBS) salts, and phospholipid concentration (5 to 15mM). More specifically, we dissolved TST in tert-butanol and added an aqueous solution of DSPE-PEG2000 under constant stirring. We aliquoted 1.5mL of the mixture into 4mL vials, placed samples in the -80ºC freezer overnight, and freeze-dried formulations for 24h. For large-scale experiments, we aliquoted 6mL of samples into 20 mL vials for freeze-drying. We recorded the visual appearance of lyophilized cakes and reconstituted samples in aqueous media using gentle swirling. Formulations were characterized regarding particle size by dynamic light scattering, drug concentration by high-performance liquid chromatography (HPLC), and antimicrobial activity against MRSA USA 300 by microdilution assay.
Results: We increased tert-butanol from 50% to 75% (v/v) to evaluate if undissolved drug particles were present in the co-solvent system and affecting the drug loading in micelles. Since no improvement in drug loading was observed, we concluded that TST molecules were completely dissolved before freeze-drying, but aggregates were being formed later during freeze-drying or reconstitution processes. To investigate the impact of the cake reconstitution on drug encapsulation, we prepared the same formulation with 75% tert-butanol to water (v/v) and reconstituted in water using three different methods: spontaneous reconstitution, swirling, or sonication. Although only one drug molecule could be encapsulated per micelle in all reconstitution methods, samples reconstituted with sonication were considerably less turbid than other methods, demonstrating the impact of cake reconstitution on drug encapsulation. We added PBS salts in the formulation before freeze-drying to improve the hydrophilicity of the cake and its affinity for water. However, the presence of salts combined with high tert-butanol concentration (75% v/v) affected the phospholipid solubility and resulted in phase separation. To improve the phospholipid solubility, we reduced the tert-butanol concentration to 50% (v/v) while still keeping the PBS salts, and we were able to obtain a clear system. After lyophilization, the formulation containing PBS salts reconstituted faster in water than cakes prepared without salts and reconstituted with saline (30 seconds versus 3 minutes). In addition, the presence of PBS salts increased the drug loading from 1 to 5 drug molecules per micelle. To strengthen the formulation, we increased the phospholipid concentration to 10mM or 15mM keeping the same ratio of the drug. However, cakes took longer to reconstitute, resulting in drug precipitation. As an alternative, we prepared the formulation with 5mM of DSPE-PEG2000 with a larger fill volume (6mL) and reconstituted the sample in 1/3 of the volume. The reconstitution time was about 1 minute and no other particles besides micelles were observed in the formulation, indicating complete drug encapsulation. The TST solubility was increased 245-fold compared to free drug. Finally, TST-SSM presented better activity than the free drug dissolved in 4% DMSO against MRSA USA 300 (MIC 0.05 μM versus 0.2μM), suggesting additional beneficial roles of the delivery system.
Conclusion: We successfully optimized the scalable co-solvent freeze-drying method to produce TST-SSM with a high loading of 5 drug molecules per micelle. We observed that lyophilized cakes prepared with low phospholipid concentration (5mM) and BPS salts exhibited fast reconstitution in water and efficient drug encapsulation. However, the co-solvent system required an adequate proportion of tert-butanol, water, and PBS salts to maintain both drug and phospholipid in solution before freeze-drying, which in our case was 50% tert-butanol: 2X PBS (v/v). To strengthen the nanomedicine, we concentrated the formulation by reconstituting the cake in 1/3 of the fill volume, increasing the total aqueous solubility of TST by 245-fold. TST-SSM was effective against MRSA USA 300 and offers great potential for clinical development. In addition, findings from this study will contribute for the development of other nanomedicines using the co-solvent freeze-drying method.