Formation of Stable Nanobubbles on Reconstituting Lyophilized ...

Journal of Pharmaceutical Sciences 105 (2016) 2249-2253

Contents lists available at ScienceDirect

Journal of Pharmaceutical Sciences journal homepage: www.jpharmsci.org

Lessons Learned

Formation of Stable Nanobubbles on Reconstituting Lyophilized Formulations Containing Trehalose Chen Zhou 1, Derrick Cleland 1, Jared Snell 2, Wei Qi 3, Theodore W. Randolph 2, John F. Carpenter 1, * 1 2 3

Department of Pharmaceutical Sciences, University of Colorado, Anschutz Medical Campus, Aurora, Colorado 80045 Department of Chemical and Biological Engineering, University of Colorado, Boulder, Colorado 80309 Malvern Instruments, Inc., Columbia, Maryland 21046

a r t i c l e i n f o

a b s t r a c t

Article history: Received 11 April 2016 Accepted 29 April 2016 Available online 8 June 2016

During an investigation of subvisible particles found in lyophilized formulations of intravenous immunoglobulin, we used resonant mass measurement techniques and discovered the presence of nanobubbles (NBs) when a 5% trehalose formulation was reconstituted. This discovery prompted studies to characterize these NBs in placebo formulations as a function of processing conditions and solution compositions. Degassing the reconstituted solutions by applying vacuum removed micron-sized bubbles but did not substantially affect the concentration of NBs. Samples that were annealed in the frozen state before lyophilization had reduced surface areas and, on reconstitution, yielded fewer NBs. Trehalose formulations with added polysorbate 20 (PS20) and formulations with higher ionic strength also had smaller numbers of NBs. Zeta potentials of the bubbles were negative in each of the formulations tested, but the negative zeta potentials were decreased in magnitude with increasing ionic strength and with addition of PS20. When incubated at 4 C, the number of NBs was largely unchanged at the end of 11 days, whereas the number of micron-sized bubbles gradually decreased during the 11-day incubation. Because of their exceptional stability, NBs are expected to contribute to the numbers of submicron particles that can be detected in reconstituted lyophilized protein formulations. © 2016 American Pharmacists Association®. Published by Elsevier Inc. All rights reserved.

Keywords: protein lyophilization particle analysis nanobubble protein particles resonant mass measurement

Introduction Nanobubbles (NBs) are small (100-200 nm), negatively charged air bubbles which may exist in aqueous solutions for prolonged periods.1,2 Recent studies have demonstrated that NBs can be stable for at least a week.1 The persistence of such nanometer-scale air bubbles in liquid solutions is surprising due to the expectation that surface tension forces would generate large pressures within the bubbles,1,3 which in turn would facilitate their rapid dissolution. The unusual persistence of NBs and their potential practical applicationsdwhich include water treatment, medical imaging, surface cleaning, foamed food production, and carbonated drinks1,4,5dhave led to a burgeoning research field focused on characterization and commercial exploitation of NBs. However, to date, there have not been published studies on the topic of the

* Correspondence to: John F. Carpenter (Telephone: (303) 724-6110; Fax: (303) 724-7266). E-mail address: [email protected] (J.F. Carpenter).

current work: NBs formed during reconstitution of lyophilized pharmaceutical formulations. Lyophilization is one of the most important formulation strategies for conferring long-term storage stability to therapeutic proteins. As of 2012, there were over 140 therapeutic protein drugs on the market, among which 64 were marketed as lyophilized products.4 Despite the general stability benefits of freeze-dried formulations, some degree of protein aggregation and particle formation may still be detected in reconstituted, lyophilized formulations.5-7 On account of the potential of particulate matter in therapeutic protein formulations to induce unwanted immune responses,8-11 quantification and characterization of aggregates and particles within these formulations has become a major area of research focus. In the present study, we report the discovery and characterization of NBs in reconstituted lyophilized formulations, examine their stability and investigate the effects of excipients on their formation during reconstitution. The size distribution and concentrations of micron-sized bubbles (MBs) were measured by flow imaging microscopy and those of NBs were measured by resonant mass measurement (RMM) and nanoparticle tracking

http://dx.doi.org/10.1016/j.xphs.2016.04.035 0022-3549/© 2016 American Pharmacists Association®. Published by Elsevier Inc. All rights reserved.

2250

C. Zhou et al. / Journal of Pharmaceutical Sciences 105 (2016) 2249-2253

analysis (NTA). Zeta potentials of air bubbles were measured with electrophoretic light scattering. We also investigated the effects of changing process variables such as the addition of an annealing step during lyophilization and degassing of the reconstituted formulations on MB and NB stability. Materials and Methods

Then primary drying was initiated, and the cycle was completed as described previously for the standard lyophilization cycle. Reconstitution of Freeze-Dried Formulations To obtain the original 1-mL volume of samples, 999-, 995-, 990-, and 950-mL milliQ water (Millipore), respectively, was used to reconstitute lyophilized samples prepared with 0.1, 0.5, 1%, and 5% trehalose, respectively.

Materials Intravenous immunoglobulin (IVIG; Gammagard® Liquid, Baxter HealthCare Corp, Deerfield, IL) was purchased from the University of Colorado at Boulder's Wardenburg Pharmacy. IVIG (Gammagard® Liquid) is a polyclonal human IgG product that includes Ig 1, 2, 3, and 4 and is formulated in 0.25-mM glycine at pH 4.6-5.1.12 High-purity trehalose dihydrate was purchased from Pfanstiehl, Inc. (Waukegan, IL). All other chemicals were purchased from Fisher Scientific (Hampton, NH) and were of reagent grade or higher quality. Methods Preparation of Formulations Unless otherwise noted, all liquid formulations contained 2-mM sodium citrate and had a pH of 4.6. To evaluate the effect of trehalose concentration, formulations containing trehalose at concentrations of 0.1, 0.5, 1, and 5% (w/v) were prepared. In addition, to evaluate potential ionic strength effects on bubble formation and stability in formulations containing 5% trehalose, NaCl was added at 10, 50, 100, and 150 mM. To evaluate the effect of nonionic surfactant and bulking agent, 0.03% (w/v) polysorbate 20 (PS20) and 250-mM glycine, respectively, were added to 5% trehalose solutions. IVIG from the original formulation was dialyzed against 1000-fold volumes of buffer solutions (2-mM sodium citrate at pH 4.6) with 3 changes of the external solution. To study NBs in a protein formulation, an IVIG (1 mg/mL) solution was prepared containing 2-mM sodium citrate and 5% trehalose at pH 4.6. All samples were prepared and studied with 3 or more replicates. Lyophilization of Placebo and Protein-Containing Formulations Before lyophilization, 220-nm polyvinylidene fluoride and polyethersulfone syringe filters (Millipore, Billerica, MA) were used to filter placebo and IVIG-containing formulations, respectively. IVIG solutions were also ultracentrifuged at 100,000 g for 3 hours to remove protein particles. Aliquots of 1-mL liquid solutions were pipetted into a 5-mL glass vials. For standard lyophilization, the shelf temperature was initially set at 10 C, and samples were allowed to equilibrate for 1 h. Then shelf temperature was reduced to 5 C at 1 C/min and held at this temperature for 20 min. The shelf temperature was then reduced to 45 C at 1.3 C/min. After holding samples at 45 C for 400 min, primary drying was per formed by increasing the shelf temperature to 20 C at 2.5 C/min, setting chamber pressure at 70 mTorr and maintaining these conditions for 1400 min. Then the shelf temperature was increased to 33 C at 0.3 C/min to initiate a 4-h secondary drying period, during which time shelf temperatures were maintained at 33 C and the chamber pressure was kept at 70 mTorr. After completion of the freeze-drying cycle, nitrogen gas was used to back-fill the chamber, and vials were then stoppered. To evaluate the effects of an annealing step, the lyophilization cycle was modified.13 After holding the samples at 45 C for 400 min, the shelf temperature was increased to 5 C during 30 min and held at 5 C for 6 h. Then the shelf temperature was reduced to 45 C at 1.3 C/min and held at this temperature for 6 h.

Stability of Bubbles at 4 C After reconstitution, placebo formulations were stored quies cently at 4 C for 11 days and analyzed for bubbles on days 0, 4, and 11. For each sample at each time point, triplicate samples were prepared and analyzed. Degassing Procedure To degas reconstituted samples, vials with stoppers partially raised to allow air movement were placed in a desiccator connected to a vacuum (~15 inches Hg) line for 1 h. Samples were analyzed and compared before and after degassing. Triplicates were prepared and analyzed for each sample. Particle Characterization Flow imaging microscopy (FlowCAM® VS1 Benchtop B3 system; Fluid Imaging Technology, Inc., Scarborough, ME) was used to measure particle sizes and concentrations in the size range of 1 mm. A 0.5-mL sample was loaded into the sample holder, and 0.3 mL was analyzed at flow rate of 0.08 mL/min. Both dark and white pixel thresholds were set at 15. RMM (Archimedes®; Malvern Instruments Ltd., Malvern, UK) was used to measure particle sizes and concentrations in the size range of ~0.1 to 5 mm. Densities of air bubbles and protein particles were set at 0.00123 g/cm3 and 1.4 g/cm3, respectively. Samples (~150 mL) were loaded and analyzed either for a total particle number of 100 or 10 min, whichever came first. In the initial study of a formulation containing 5% trehalose and 1-mg/mL IVIG, NTA NS300 (Malvern Instruments Ltd.) was used to characterize particle numbers and sizes in the range of 50-1000 nm. A 0.8-mL aliquot of sample was loaded into a 1-mL silicone oilfree syringe and was continuously injected, as three 30-second videos were recorded. The NanoSight NTA 3.0 software was used for both data acquisition and analysis. A solution viscosity of 1.3 centipoise was used in the calculation of particle sizes. Zeta Potential Measurement Zeta potentials of NBs were measured by electrophoretic light scattering with a Zetasizer® (Malvern Instruments Ltd.) instrument. The particle refractive index was set at 1.003. Solution refractive index, viscosity, and dielectric constant were calculated based on the solution components using Zetasizer software. Surface Area Measurement The specific surface area of lyophilized formulations was measured by nitrogen BrunauereEmmetteTeller analysis with a Quantachrome Autosorb-1 (Boynton Beach, FL). The sample preparation and experimental procedure were the same as described previously.13 Results NBs in Freeze-Dried IVIG Formulations After Reconstitution In the initial study, 1-mg/mL IVIG samples formulated with either 5% trehalose or 5% trehalose plus 0.03% PS20 were reconstituted with milliQ water. Before lyophilization, submicron particle


levels were below the detection limits of the NTA and RMM instruments. After lyophilization and reconstitution, the submicron particle levels in the 5% trehalose formulation without PS20 were above the upper limit of the NTA instrument (>4 109 particles/ mL). In contrast, NTA detected only 2 108 nanoparticles/mL in the 5% trehalose plus 0.03% PS20 formulation. RMM detected both negatively (7.9 106 particles/mL) and positively buoyant particles (1.2 106 particles/mL) in the reconstituted 5% trehalose formulation. In this relatively simple formulation, the negatively buoyant particles are composed of protein, whereas the positively buoyant particles cannot be attributed to any formulation component. Thus, we conclude that the positively buoyant particles are gas-filled NBs. After reconstitution of the 5% trehalose plus 0.03% PS20 formulation, the concentration of negatively buoyant particles was below the lower detection limit of the RMM instrument. The level of positively buoyant NBs (1.4 106 particles/mL) was slightly higher than that observed in the reconstituted 5% trehalose formulation.

2251

3 106 particles/mL vs. ca. 1 106 particles/mL, respectively). Because NB levels were much higher in these placebo formulations, we used protein-free formulations for subsequent experiments to examine roles of excipients and processing conditions on NB formation and stability. Stability of Air Bubbles To evaluate the stability of the bubbles formed in our study, we incubated the reconstituted 5% trehalose placebo formulations at 4 C for 11 days. The total numbers of MBs initially were relatively low (ca. 10,000 particles/mL) and decreased further during the incubation, becoming essentially undetectable by day 11 (Fig. 1a). In contrast, NB concentrations did not change over the course of 11-day incubation (Fig. 1b). Zeta potentials were measured in the 5% trehalose formulation immediately after reconstitution of lyophilized cakes and after 4 and 11 days of incubation. Zeta potentials, which reflect the surface charge of NBs, did not change over the incubation period (Fig. 1c).

NB in Placebo Versus IVIG Formulations Effect of Degassing Procedure After rehydration, NBs were also found in placebo formulations containing 5% trehalose, but their level was consistently higher than that in the equivalent formulation with 1-mg/mL IVIG (ca.

To further investigate the stability of air bubbles in the reconstituted placebo solution, a vacuum degassing procedure was

Figure 1. Stability of MBs and NBs during incubation of reconstituted 5% trehalose placebo formulation at 4 C, measurements were performed at days 0, 4, and 11. (a) Levels of MBs; (b) Levels of NBs; (c) Zeta potential. Error bar represents standard deviation of 3 independent samples.

2252


applied to the samples. After degassing, the concentration of MBs dropped by an order of magnitude (from approximately 10,000 particles/mL to approximately 1000 particles/mL), but the level of NBs (approximately 3 106 particles/mL) was not affected by degassing. Moreover, their zeta potentials (approximately 30 mV) were also unchanged after degassing. Effect of Excipients and Annealing To further characterize the formation of NBs in reconstituted lyophilized formulations, the effects of formulation excipients were studied. The cake structures after freeze drying of different placebo formulations were visually different. With the 5% w/v trehalose formulation, an elegant cake structure was obtained, whereas those in formulations with higher salt concentrations tended to collapse. The presence of PS20 or the bulking agent glycine did not visibly affect the cake structures. Before lyophilization, no particles were detected with either FlowCAM or RMM (data not shown). After reconstitution of the placebo formulations, both MBs and NBs were detected. The concentrations of MBs and NBs increased with increasing trehalose concentration (Figs. 2a and 2b). Low salt concentrations or the presence of glycine had minimal effect on the concentrations of bubbles. In contrast, the presence of 0.03% PS20 greatly reduced the levels of MBs and NBs. Formulations that contained 150-mM NaCl were collapsed and yielded very few bubbles on reconstitution. This observation led to the hypothesis that the bubbles that appeared after reconstitution might have arisen from microvoids within the lyophilized cakes. To test this hypothesis, we reduced the number of these

microvoids by annealing frozen formulations before lyophilization and characterized the number of NBs formed after reconstitution. After annealing, the specific surface area of the freeze-dried cake, which reflects the surface area of voids left behind after the sublimation of ice, was 0.69 ± 0.04 m2/g, in contrast to 1.34 ± 0.10 m2/g for the cake generated using the standard freeze-drying procedure. As expected, addition of an annealing step before drying significantly reduced the level of NBs formed after reconstitution (6.1 105 particles/mL) compared to levels in samples prepared without annealing (2.7 106 particles/mL). Nanobubble Surface Charges In all placebo formulations, the zeta potentials of bubbles formed after reconstitution were negative (Fig. 2c). For the formulations with trehalose only or with bulk agent glycine, the zeta potentials were approximately 30 mV, indicating a colloidally stable NB population. Notably, the zeta potential value became less negative with increasing salt concentration, likely due to charge shielding effects.2 Discussion Considering the current regulatory emphasis on measuring and controlling subvisible particles in therapeutic protein products, it is important that these particles be characterized as fully as possible. NBs are formed during reconstitution of lyophilized formulations and thus will contribute to counts of submicron particles. It has long been recognized that reconstitution results in visible and

Figure 2. (a) Size distribution of MBs, (b) Number of NBs after reconstitution, and (c) Zeta potential of air bubbles before and after reconstitution of lyophilized placebo formulations. Error bar represents standard deviation of 3 independent samples.


micron-sized bubbles. These relatively large bubbles are not stable and can readily be removed by allowing time for them to float to the surface of the solution, a process that may be accelerated by applying reduced pressure. In contrast, we observed that the concentrations of NBs were unaffected by prolonged incubation or by application of a low-pressure degassing procedure. Because of their stability and the inability of standard techniques to remove them, NBs will remain in solutions during analysis. To understand their contribution to total particle counts, it is critical that they be differentiated from other nanoparticles. The low density of NBs, coupled with RMM techniques, allows facile identification and quantification of this subpopulation of particles. We recommend that such techniques be routinely applied to the analysis of reconstituted formulations of therapeutic proteins. The stability of NBs should be considered in light of the Laplace equation, which states that the pressure difference between the inside and outside of a bubble is proportional to the surface tension divided by the bubble radius

DP ¼

2s R

(1)

where DP is the difference in pressure between the interior and exterior of the bubble, s is the surface tension, and R is the bubble radius. If the surface tension at the interface between the bubble and the solution were typical of those at macroscopic air-water interfaces, the small bubble radii would result in pressures within the NBs on the order of tens of atmospheres. As a result, diffusion of the gas out of the bubbles should result in their rapid collapse. We and others1,2,14-16 have observed that the NBs surfaces are highly negatively charged, which has been explained in terms of adsorption of OH- at the NB interface. As bubble radii decrease (e.g., due to diffusion of gases from their interiors), increased charge repulsion and decreased surface tensions effectively negate anticipated pressure increases, thus stabilizing the NBs. Furthermore, unlike microbubbles that float to the surface, Brownian motion of the small NBs is sufficient to allow them to remain suspended in solution for extended periods (e.g., at least 11 days in the present study). We speculate that microvoids within the dried formulation matrix may provide the mechanism by which NBs are formed during reconstitution. Microvoids are formed during lyophilization when microscopic ice crystals are sublimed away from the glassy matrix of the frozen formulation. Reconstitution of the lyophilized cakes results in bubbles of a wide range of sizes, reflecting the size distribution of voids. NBs may be formed directly as a subpopulation of these bubbles. In addition, they could potentially form when larger MBs shrink as gas diffuses out of them into the bulk solution. These general mechanisms are consistent with our observation that annealing frozen samples caused a decrease in surface area of the cakes, and a concomitant decrease in the number of NBs that were formed on reconstitution. The presence of PS20 was associated with lower concentrations of NBs, and the zeta potentials of the NBs that were present showed that the presence of surfactant reduced their negative charge. This could be the result of competitive adsorption of PS20 that lowered the surface OH concentration and destabilize the NBs. The presence of NaCl caused multiple effects. Massive cake collapse was observed in formulations with 150-mM NaCl, suggesting that microvoids would have been greatly reduced in these

2253

samples, consistent with the lower NB concentrations seen in the presence of NaCl. Zeta potentials of NBs found in the presence of NaCl were less negative than in the absence of salt, presumably due to charge screening and double layer effects. This surface charge reduction could also destabilize NBs. An interesting question is whether protein molecules interact with NBs. In the absence of surfactant, protein molecules are expected to rapidly adsorb to air-water interfaces such as those present by MBs and NBs. Thus, protein-coated bubbles could be a source of nanoparticulate material. Protein-coated bubbles would likely be positively buoyant and indistinguishable from uncoated NBs by RMM analysis. Adsorption of protein to NB-water interfaces might be expected to exert similar effects as the adsorption of surfactants, and we observed that PS20 and IVIG had equivalent effects on the reduction of the number of NBs observed after reconstitution of lyophilized 5% trehalose formulations. Future studies will focus in more detail on the extent and consequences of protein-NB interactions.

Acknowledgments This project was supported by Malvern Instrument Ltd. The authors thank Dr. Hans H. Funke for the specific surface area measurement.

References 1. Ushikubo FY, Furukawa T, Nakagawa R, et al. Evidence of the existence and the stability of nano-bubbles in water. Colloid Surf A. 2010;361(1-3):31-37. 2. Jin F, Li J, Ye X, Wu C. Effects of pH and ionic strength on the stability of nanobubbles in aqueous solutions of alpha-cyclodextrin. J Phys Chem B. 2007;111(40):11745-11749. 3. Seddon JR, Lohse D, Ducker WA, Craig VS. A deliberation on nanobubbles at surfaces and in bulk. Chemphyschem. 2012;13(8):2179-2187. 4. Meyer BK, Shameem M. 1-Commercial therapeutic protein drug products. In: Meyer BK, ed. Therapeutic Protein Drug Products. Cambridge, UK: Woodhead Publishing; 2012:1-11. 5. Dong A, Prestrelski SJ, Allison SD, Carpenter JF. Infrared spectroscopic studies of lyophilization- and temperature-induced protein aggregation. J Pharm Sci. 1995;84(4):415-424. 6. Davis JM, Zhang N, Payne RW, et al. Stability of lyophilized sucrose formulations of an IgG1: subvisible particle formation. Pharm Dev Technol. 2013;18(4): 883-896. 7. Kreilgaard L, Frokjaer S, Flink JM, Randolph TW, Carpenter JF. Effects of additives on the stability of recombinant human factor XIII during freeze-drying and storage in the dried solid. Arch Biochem Biophys. 1998;360(1):121-134. 8. Rosenberg AS. Effects of protein aggregates: an immunologic perspective. AAPS J. 2006;8(3):E501-E507. 9. Kotarek J, Stuart C, De Paoli SH, et al. Subvisible particle content, formulation, and dose of an erythropoietin peptide mimetic product are associated with severe adverse postmarketing events. J Pharm Sci. 2016;105(3):1023-1027. 10. Fradkin AH, Carpenter JF, Randolph TW. Glass particles as an adjuvant: a model for adverse immunogenicity of therapeutic proteins. J Pharm Sci. 2011;100(11): 4953-4964. 11. Van Beers MMC, Gilli F, Schellekens H, Randolph TW, Jiskoot W. Immunogenicity of recombinant human interferon beta interacting with particles of glass, metal, and polystyrene. J Pharm Sci. 2012;101(1):187-199. 12. Gammagard Liquid Immune Globulin Infusion [package insert]. Deerfield, IL: Baxter Healthcare Corporation; 2014. 13. Xu Y, Grobelny P, Von Allmen A, et al. Protein quantity on the air-solid interface determines degradation rates of human growth hormone in lyophilized samples. J Pharm Sci. 2014;103(5):1356-1366. 14. Ebina K, Shi K, Hirao M, et al. Oxygen and air nanobubble water solution promote the growth of plants, fishes, and mice. PLoS One. 2013;8(6):e65339. 15. Cai WB, Yang HL, Zhang J, et al. The optimized fabrication of nanobubbles as ultrasound contrast agents for tumor imaging. Sci Rep. 2015;5:13725. 16. Cho S-H, Kim J-Y, Chun J-H, Kim J-D. Ultrasonic formation of nanobubbles and their zeta-potentials in aqueous electrolyte and surfactant solutions. Colloid Surf A. 2005;269(1-3):28-34.