Microstructure Formation for Improved Dissolution Performance of

Publication Date (Web): February 27, 2019. Copyright © 2019 American Chemical ... Wu, Wang, Löbmann, Grohganz, and Rades. 2019 16 (3), pp 1294–130...
0 downloads 0 Views 987KB Size
Subscriber access provided by Washington University | Libraries

Article

Microstructure Formation for Improved Dissolution Performance of Lopinavir Amorphous Solid Dispersions Na Li, and Lynne S. Taylor Mol. Pharmaceutics, Just Accepted Manuscript • DOI: 10.1021/acs.molpharmaceut.9b00117 • Publication Date (Web): 27 Feb 2019 Downloaded from http://pubs.acs.org on March 3, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 49 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

Microstructure Formation for Improved Dissolution Performance of Lopinavir Amorphous Solid Dispersions

Na Li† and Lynne S. Taylor*† †Department

of Industrial and Physical Pharmacy, Purdue University, 575 Stadium Mall Drive, West Lafayette, Indiana 47907, United States

ACS Paragon Plus Environment

1

Molecular Pharmaceutics 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 49

Abstract Amorphous solid dispersions (ASDs), where the drug is dispersed in a polymer, have become increasingly prevalent as a formulation strategy for the oral delivery of poorly-soluble drugs due to their potential for substantial solubility enhancement. However, ASDs are susceptible to amorphous phase separation, which may promote crystallization and/or alter the release performance. Nevertheless, the mechanisms by which phase separation and subtle microstructural changes affect ASD release remain poorly understood. Therefore, understanding the microstructure of ASDs and the subsequent implication for ASD performance are critical to design an optimally performing formulation. In this study, comprehensive investigations of microstructure evolution in lopinavir ASDs, prepared using a solvent-based process, were undertaken. Atomic force microscopy (AFM)-based nanoscale thermal analysis (nanoTA) enabled characterization of local composition at the submicron scale. The formation of heterogeneous domains was found to improve the in vitro release of lopinavir from lopinavirhydroxypropylmethylcellulose (HPMC) ASDs for drug loadings above 33% w/w. The composition and amount of each phase formed, as well as the size and location of drug-rich phases, were found to be critical factors contributing to the altered release kinetics observed. This study highlights the complexity and importance of ASD microstructure, and should contribute to a broader understanding of ASD release mechanisms.

Key words: amorphous solid dispersions, miscibility, phase-separation, drug-loading, dissolution, nanoTA

ACS Paragon Plus Environment

2

Page 3 of 49 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

Abbreviations AFM: atomic force microscopy ASD: amorphous solid dispersions DL: drug loading HPMC: hydroxypropylmethylcellulose HPLC: high performance liquid chromatography LLPS: liquid-liquid phase separation LPV: lopinavir nanoTA: nanoscale thermal analysis NTA: nanoparticle tracking analysis RH: relative humidity

ACS Paragon Plus Environment

3

Molecular Pharmaceutics 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 49

Introduction With the application of structure-based drug design and high throughput screening methods, more and more potential therapeutic agents have poor aqueous solubility. This has been the most problematic physicochemical property compromising drug oral bioavailability

1-3.

The use of

supersaturating formulations, such as amorphous solid dispersions (ASDs), where the drug is dispersed in a polymer, can improve not only the release rate, but also the driving force for passive diffusion across the intestinal wall 4. Consequently, ASDs are widely used to improve the solubility and bioavailability of poorly-soluble drugs5-7. A successful ASD formulation should remain stable throughout its shelf-life, and possess optimal dissolution performance in the human gastrointestinal tract to achieve high bioavailability. In the past, extensive efforts have been directed towards maintaining the solid-state stability of ASDs, including inhibiting drug crystallization8, 9, minimizing amorphous-amorphous phase separation10-12, and increasing drugpolymer interactions.13, 14 However, there is less understanding of how to optimize the release performance. While the impact of solid-state crystallinity on the dissolution performance of ASDs is well documented,15-17 the role of amorphous-amorphous phase separation on drug release is not well understood with conflicting reports on impact18, 19. Drug release from an ASD is dependent on multiple factors, including polymer type20-22, inclusion of surfactants20, drug loading23-25, extent of mixing26, and particle size27. Upon phase separation in the ASD matrix, heterogeneous amorphous domains with different local compositions are created. Therefore, the dissolution performance can be altered through domain size reduction and local compositional changes (altered local drug loading). Reduction in domain size and creation of high surface area can promote ASD dissolution27. Matteucci et al reported that high surface areas were responsible the improved performance of phase-separated amorphous

ACS Paragon Plus Environment

4

Page 5 of 49 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

itraconazole formulations compared to homogeneous ASDs28. However, other studies have suggested that amorphous-amorphous phase separation in ASDs does not always improve dissolution performance18, 29. Chen and colleagues produced ASDs particles at 40% drug loading with compound BMS-817399 and polyvinylpyrrolidone (PVP) by spray drying and subjected them to storage at 95% relative humidity (RH) to induce amorphous-amorphous phase separation. They found surface enrichment of the hydrophobic drug on these particles. The surface normalized dissolution rate of the drug from these systems was not affected, but the polymer release rate was reduced in phase-separated ASDs18. Purohit et al studied ritonavir ASDs, at drug loadings of 10% to 50%, formulated with PVP, copovidone (PVPVA), or hydroxypropyl methylcellulose acetate succinate (HPMCAS) and stored at 97% RH to induce phase separation. Reduced drug release was observed from some ASDs whereby the extent of change in drug release was dependent on both polymer type and drug loading29. The authors suggested that reduced dissolution performance could be attributed to phase separation. Nonetheless, the dissolution mechanisms and factors dominating the drug release rate of phase-separated ASDs remain obscure. Upon phase separation, heterogeneous structures with average diameters ranging from a few nanometers to a few microns can form within the ASD matrix 30. Thus, to characterize the local composition of ASDs, analytical techniques enabling chemical characterization with high spatial resolution are needed. Advances in high-resolution imaging techniques coupled with orthogonal chemical analyses have empowered compositional analysis at the nanoscale. Atomic force microscopy (AFM) based techniques, such as AFM-infrared spectroscopy (AFM-IR)

31-35,

AFM-Raman spectroscopy (AFM-Raman)36, pulsed force mode (PFM)-AFM37, microscale and nanoscale thermal analyses (microTA and nanoTA) resonance (LCR)20,

34, 43,

12, 31-33, 38-42,

and AFM-Lorentz contact

have been reported as approaches useful to characterize the

ACS Paragon Plus Environment

5

Molecular Pharmaceutics 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 49

microstructure and local compositions in various pharmaceutical systems at the submicron scale. In addition, electron microscopy-based elemental analysis, including transmission electron microscopy (TEM)-energy-dispersive X-ray spectroscopy (EDX)43, 44, and TEM- electron energy loss spectroscopy (EELS)45, are powerful tools to characterize nanoscale compositions in ASDs. Nevertheless, IR and Raman spectroscopy have penetration depths of about 1 µm and therefore spectral mixing is often observed if the phase-separated domain size is smaller than this 32, 35. In TEM analysis, the electron beam passes through the specimen and therefore the vertical average composition though the thickness of the sample instead of an individual domain is measured. The current AFM-based LCR analysis is more suitable for qualitative analysis and semi-quantitative analysis, and may introduce artifacts32, 34, 46. AFM-based thermal analysis (nanoTA), however, has a high spatial resolution and sensitivity to surface composition, and is therefore of interest for compositional characterization of nanoscale domains formed in an ASD matrix. A typical spatial resolution of 100 nm and a depth sensitivity of about 50 nm can be achieved with this technique 32,

enabling characterization of small heterogeneous structures. If the glass transition temperature

difference between the two components is significant, the local sample composition can be then estimated. The purpose of this study is to evaluate various ASD formulations in terms of drug release, probing interrelationships between solid-state microstructure and solution-state performance. Specifically, this study explores 1) how variations in drug loading, as well as the route and extent of phase separation cause differences in drug release; 2) how the creation of multiple phases impacts the release mechanism; and 3) how different types of nanoparticles may be formed during ASD disintegration/dissolution. Lopinavir (LPV) was used as the model poorly soluble compound with a low crystallization propensity47. Hydroxypropylmethylcellulose (HPMC) has one of the

ACS Paragon Plus Environment

6

Page 7 of 49 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

highest glass transition temperatures among polymers commonly used in ASD formulation

32,

enabling estimation of local compositions based on local thermal properties, and therefore was used as the model polymer. The chemical structures of lopinavir and HPMC are shown in Figure 1.

Materials and Methods Materials Lopinavir was purchased from Chemshuttle (Wuxi, China). HPMC (Methocel E5) was obtained from the Dow Chemical Company (Midland, MI). HPLC grade methanol, acetonitrile, and dichloromethane were purchased from Mallinckrodt Baker (Phillipsburg, NJ). All other chemicals (citric acid and sodium citrate dihydrate) were purchased from SigmaAldrich (St. Louis, MO). Phosphate buffer solution of pH 6.5 and 50 mM was prepared by dissolving 2.236 g of sodium phosphate dibasic and 4.726 g of sodium phosphate monobasic monohydrate in water to a final volume of 1000 mL.

Methods Physicochemical property characterization The melting (Tm) and glass transition (Tg) temperatures of crystalline and amorphous lopinavir respectively were measured using differential scanning calorimetry (DSC). A TA Q2000 DSC equipped with an RCS90 refrigeration unit (TA Instruments, New Castle, DE) was used. Crystalline lopinavir powder was sieved to obtain powder solids with particle size from 106 to 250 µm. Approximately 6 mg of powder was weighed into Tzero aluminum DSC pans (TA Instruments, New Castle, DE). Samples were equilibrated at 10 °C, ramped up to 110 °C with a heating rate of

ACS Paragon Plus Environment

7

Molecular Pharmaceutics 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 49

1 °C/min, then equilibrated at -40 °C, and ramped up to 210 °C with a heating rate of 5 °C/min and a modulation of  1 °C every 60 seconds. Nitrogen was used as the purge gas at a flow rate of 50 mL/min. The first heating cycle was used to determine Tm, and the second heating cycle was used to determine Tg.

Preparation of amorphous solid dispersion (ASD) films Stock solutions of 4 mg/mL solid content were prepared at different drug loadings using a binary solvent mixture of 1:1 methanol : dichloromethane (v:v). Initially miscible ASD films were prepared by rotary evaporation directly from scintillation vials. To prepare phase-separated ASD systems, different amounts of water were introduced into the stock solution. The mixture was then rotavapped yielding phase-separated ASD films. For storage samples, initially miscible ASDs were stored at 97% RH (generated by a saturated solution of potassium sulfate48) and 25 C for 7 days. These samples were dried under vacuum for at least 48 hours prior to analysis. ASDs were prepared at pre-determined solid loadings, yielding LPV concentrations of 17 µg/mL (the amorphous solubility of LPV49) and 30 µg/mL in 40 mL of dissolution media in subsequent dissolution experiments. Triplicate experiments were performed. For AFM topography and thermal property evaluations, round glass cover slides (i.d. 10 mm) were placed in glass vials for film preparation. ASDs were prepared at LPV concentrations equivalent to 30 µg/mL in 10 mL of dissolution media to obtain films with microstructures similar to samples used for dissolution. Glass slides were chosen to avoid substrate effects on film morphology, such that the film obtained on the slide is the same as the film deposited on the bottom of the glass vial (i.e. similar to the film used for dissolution testing). ASD films were prepared using the protocol described above. After vacuum drying, the glass slide was carefully taken out

ACS Paragon Plus Environment

8

Page 9 of 49 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

from the vial, then taped onto a steel stud for AFM analysis. At least three samples were prepared for each treatment.

Release tests Forty mL of pH 6.5 phosphate buffer was added to vials containing the ASD film to evaluate the dissolution kinetics. At predetermined time points, 0.5 mL of the solution was sampled and filtered through a 0.45 µm polytetrafluoroethylene syringe filter (4 mm diameter) for LPV quantitation. The first 0.25 mL was discarded to saturate the filter. After each sampling point, the dissolution media was replenished with 0.5 mL of blank buffer solution. All dissolution tests were carried out at 251 C. For experiments carried out at 30 µg/mL LPV, in order to characterize fast drug release over the first 5 minutes, aliquots of dissolution media were sampled at 10s, 30s, 50s, 70s, 90s, 120s, 180s, 240s, 300s, 360s, 420s, 480s, 540s, and 600s with equal volumes of fresh media replacement. For experiments carried out at 17 µg/mL LPV (its amorphous “solubility”), sampling time was 5min, 10min, 20min, 30min, 40min, 50min, and 60min with media replacement. For experiments at 200 µg/mL (LPV + polymer), sampling time was 10min, 20min, 30min, and 40min with media replacement. The first derivative of the average lopinavir concentration curve as a function of time was calculated as the drug release rate. An Agilent 1260 Infinity series HPLC (Agilent Technologies, Santa Clara, CA) equipped with a Waters XTerra RP C-18 column (150 mm × 4.6 mm, i.d. 3.5 μm) (Waters Corp., Milford, MA) was used. An isocratic elution method of 60% acetonitrile and 40% water was used. LPV was detected at 210 nm with an injection volume of 10 µL and a flow rate of 1 mL/min. The total run length was 7 min. Calibration curves were constructed to cover the appropriate sample concentration range (5-25 µg/mL, 1-5 µg/mL, and 0.1-1 µg/mL) with R2 values above 0.999. The

ACS Paragon Plus Environment

9

Molecular Pharmaceutics 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 49

average recovery was between 98% to 102%, with standard deviations lower than 2%. The limit of detection of this HPLC method is below 20 ng/mL. The degree of supersaturation was calculated by dividing the lopinavir concentration with the crystalline solubility of lopinavir measured under the same conditions.

Film immersion studies ASD films were prepared by spin-coating at 18% RH or below for buffer immersion studies 43.

Stock solutions of 50 mg/mL were prepared by dissolving lopinavir and HPMC in 1:1 methanol

: dichloromethane solvent mixture. Four percent (v/v) water was added in the stock solution and adequately mixed prior to spin-coating to prepare phase-separated films. A spin coater (Chemat Technology Inc., Northridge, CA) was used. Twenty microliters of stock solution were deposited onto the substrate. It was then spun at 50 rpm for 6 seconds, and 3100 rpm for 30 seconds. All films were vacuum dried for at least 48 hours to remove residual solvent. To determine the evolution of film morphology upon hydration, each dry film was immersed in 2 mL of pH 6.5 phosphate buffer solutions for predetermined time intervals and then air-dried. These samples were further dried under vacuum for at least 48 hours prior to AFM imaging.

Characterization of ASD microstructure A nanoIR2 AFM instrument (Anasys Instruments, Santa Barbara, CA) was used to collect topographical images. Contact mode NIR2 cantilevers (Anasys Instruments, Santa Barbara, CA) were used. A scan rate of 0.5 Hz was used, with an x and y resolution of 256 points. Images were analyzed using Analysis Studio (Anasys Instruments, Santa Barbara, CA). Surface roughness (Ra)

ACS Paragon Plus Environment

10

Page 11 of 49 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

was calculated using Gwyddion (http://gwyddion.net/). The size of the spherical domains found in the AFM height images were measured using the ruler tool in the software. For samples with less surface domains (such as samples prepared with 100% water in solvent), at least 50 domains were counted. For all other samples, at least 2 images were analyzed. Histograms were generated from 0 to 2000 nm range with a bin size of 40 nm using OriginPro 2015 64Bit b9.2.214 (OriginLab, Northampton, MA). D10, D50, and D90 values were calculated based on histograms generated. For nanoscale thermal analysis (nanoTA), Thermalever® cantilevers (Anasys Instruments, Santa Barbara, CA) were used with the NIR2 AFM in the nanoTA mode using the protocol described in a previous study32. Briefly, a calibration curve was obtained by measuring the softening points of polycaprolactone (55 °C), polyethylene (116 °C), and poly (ethylene terephthalate) (235 °C) prior to sample analysis. The local glass transition temperature of ASD films was determined using the peak point in the deflection vs temperature curve. A calibration curve of LPV-HPMC ASD was obtained by determining the glass transition temperature of miscible films obtained by rotary evaporation. The Gordon-Taylor equation was used for fitting and composition calculations 50: 𝑇𝑔 =

𝑤1𝑇𝑔1 + 𝐾(1 ― 𝑤1)𝑇𝑔2 𝑤1 + 𝐾(1 ― 𝑤1)

(equation 1)

Where 𝑇𝑔 is the glass transition temperature of the miscible ASD, 𝑤1 and 𝑇𝑔1 (in Kelvin) are the weight fraction and glass transition temperature of the drug, and 𝑇𝑔2 (in Kelvin) is the glass transition temperature of the polymer. 𝑇𝑔1 and 𝑇𝑔2 values were determined to be 367 K and 482 K by nanoTA. For each phase-separated sample, equal numbers (5-30) of nanoTA scans were performed on the discrete phase and continuous phase respectively to generate distribution statistics.

ACS Paragon Plus Environment

11

Molecular Pharmaceutics 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 49

Particle characterization 33% drug-loading ASDs whereby different amounts of water were added during preparation were dissolved to a final theoretical LPV concentration of 42.5 µg/mL. Miscible ASD films with a 10% drug-loading were used as control samples to generate colloidal species via liquid-liquid phase separation (LLPS). Control samples were dissolved at LPV concentrations of 34 µg/mL. These concentrations were chosen to ensure the number of particles generated were within the optimal concentration for nanoparticle tracking analysis. After 5 minutes, aliquots of the dissolution media were removed for further analysis. Each experiment was repeated three times. The viscosity of the dispersant was measured using an SV-10 AND vibro viscometer (A&D Company Ltd, Tokyo, Japan). The dissolution media were filtered through 0.45 µm cellulose acetate syringe filters to remove particles suspended in solution prior to viscosity analysis. Calibration was performed prior to analysis using reverse osmosis water. Particle size distribution (number based) was measured using a nanoparticle tracking analysis (NTA) system NanoSight LM 100 (Malvern Instruments, Westborough, MA). Most phase-separated samples generated bimodal or multimodal distributions of particles during release testing, and therefore the use of dynamic light scattering (DLS) analysis was not feasible. Samples prepared following the protocol described above were used without filtration. Data collection and analysis were performed using the nanoparticle tracking analysis software version 3.1.46. A flowthrough sample stage and a green laser of 532 nm were used. The measured viscosity values were used for particle size calculations. The camera settings were kept at a screen gain of 3.0 and camera level of 8.0. For analysis, the detection threshold and camera gain were set at 5.0 and 10.0, respectively. D10, D50 (also referred to as median diameter), and D90 were calculated as the

ACS Paragon Plus Environment

12

Page 13 of 49 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

cumulative 10%, 50%, and 90% percentile of the diameters. Three separate samples were prepared, and the average of triplicate experiments was plotted as the particle size distribution. To characterize particles formed in solvent systems, methanolic stock solution containing 10 mg/mL lopinavir was titrated into different amount of water and mixed. Dichloromethane was not used due to its immiscibility with water. To exclude the interference from polymer aggregates, HPMC was not added as it is insoluble in methanol. Because of particle counting difficulties in systems with a large number of particles, NTA measurement was not performed. A Malvern nanoZS Zetasizer (Malvern Instruments, Westborough, MA) was used to perform dynamic light scattering measurements. The mixture was immediately mixed and analyzed to minimize the impact of particle agglomeration. Triplicate samples were prepared and measured. The viscosity and refractive indices of the methanol-water solvent mixture was estimated from the results summarized by Thompson et al 51 and Herraez et al 52.

Results Physicochemical properties Lopinavir has an amorphous “solubility” of 16–18 μg/mL at 25 °C 49. This is also referred as the onset LLPS concentration, i.e. the concentration where a new phase is observed as the concentration of drug in water is gradually increased. In this work, 17 μg/mL was taken as the amorphous “solubility” of lopinavir. The glass transition temperature (Tg) and melting temperature (Tm) of lopinavir were measured to be 69  2 °C 49 and 94  1 °C by DSC. After 48 hours of equilibrium at 97% RH, the Tg of lopinavir was lowered to 39  9 °C 49. Therefore, lopinavir exists in the glassy state in ASDs at 25 °C. When precipitated from the solution in the form of amorphous colloidal species, it is

ACS Paragon Plus Environment

13

Molecular Pharmaceutics 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 49

likely in the supercooled liquid state as the drug is expected to absorb more water compared to when stored at 97% RH, and thus possess a lower Tg. LPV is not expected to crystallize in the presence of HPMC within the time frame tested in this study 49. The calorimetric Tg of HPMC was 140  5 °C as measured by DSC32. Softening temperatures of pure amorphous LPV and HPMC films determined by nanoTA were 97 ± 4 and 209 ± 11 °C32, respectively. Because the nanoTA method of evaluating Tg relies on penetration of the AFM cantilever into the sample, the value measured could be different as compared to the calorimetric Tg32, 53, 54. Herein, the wide window between softening Tgs of HPMC and LPV enables compositional characterization based on local Tg values.

Release performance Impact of phase separation The release profiles of LPV from ASDs with different drug loadings, is summarized in Figure 2. Given that an excess of ASD was added (equivalent to 30 μg/mL drug), and that the solution was filtered to remove undissolved ASD fragments and most colloidal species, the maximum solution concentration is expected to be around the amorphous solubility of LPV (17 μg/mL; this represents the maximum molecularly dissolved concentration of LPV that can exist, at concentrations higher than this, colloidal species form). It can be seen that the maximum solution concentration achieved is indeed close to this expected value, although only the 15% drug loading ASD reached this value. The drug release rate increased with decreasing drug loading in all three systems; these comprised miscible ASDs as well as phase-separated ASDs prepared either by exposure to high RH43, or by adding water to the solvent during preparation43. However, compared to the miscible ASDs, the dissolution performance of the two phase-separated ASD systems at

ACS Paragon Plus Environment

14

Page 15 of 49 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

drug loadings from 33% to 85% were improved significantly, both in terms of the rate and degree of supersaturation generated. The route of phase separation also appears to impact the dissolution performance, with ASDs prepared with 4% water in the solvent possessing higher dissolution rates at drug loadings of 33% to 85% compared to comparable ASDs stored at 97% RH. No significant difference in dissolution rates was observed at 15% and 100% drug loading. The dissolution performance of these ASDs was also evaluated at a higher lopinavir amount (200 μg/mL of drug if all material released) and for added amounts equivalent to the amorphous solubility (17 µg/mL, lopinavir). The results are shown in Figure S1 and S2 in the supporting information. Similar enhanced dissolution performances in phase-separated ASDs were observed. The maximum solution concentration was around 17 µg/mL as excess released lopinavir present as colloidal nanodroplets was removed by filtration. For miscible lopinavir-HPMC systems, the dissolution performance of ASDs were also determined at low drug loadings (from 5% to 33% LPV) and lopinavir concentrations of 17 μg/mL. The results are shown in Figure 2D and H. Compared to Figure 2E, lower dissolution rates were observed at the same drug loadings. This can be attributed to the reduced lopinavir concentrations used. Lopinavir release increased with decreasing drug loading within 5 minutes. This is consistent with results shown in Figure 2A-C. For ASDs with drug loadings of 5% to 20%, the drug release reached a plateau within 2 minutes with similar dissolution rates observed. As the drug loading increased to 25% and 33%, a more gradual drug release pattern was observed.

Impact of water addition To further investigate the extent of phase separation on ASD dissolution performance, different amounts of water were introduced to the solvent system used to prepare the ASDs with

ACS Paragon Plus Environment

15

Molecular Pharmaceutics 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 49

the expectation that greater amounts of water would increase the extent of phase separation32. LPV release after a fixed time interval was determined, and the results are shown in Figure 3A. For 15% drug loading ASDs, the impact of water on LPV release was not significant except when a very large amount of water was added in the solvent (70% and 100% water-to-solvent ratios, corresponding to 0.7:1 and 1:1 water-solvent mixtures). For both 33% and 50% drug loading ASDs, drug release was highly sensitive to the amount of added water over the range of 0% to 6% H2O addition (Region I), whereby increasing amounts of water promoted LPV release. In contrast, when water addition exceeded 40%, lopinavir release was reduced (Region III). Water-to-solvent ratios from 6% to 40% were optimal for ASD dissolution (Region II).

ASD Microstructure ASD film morphology The AFM height images of the ASD films prepared under different conditions are shown in Figure 4. The miscible ASDs showed a smooth surface with very few surface structures. For ASDs prepared with water in the solvent, spherical shaped discrete domains were formed at all drug loadings. As drug loading increased, these features became larger. At 15% drug loading, the domain height is about 50 nm, whereas at 70% drug loading, the domain height increased to around 500 nm. For ASDs stored at 97% RH, the height images showed some variations in topography. Irregularly shaped discrete and semi-continuous domains were formed at all drug loadings. The size of these height features was also non-uniform, ranging from 100 nm to a few microns. The surface roughness of 33% LPV ASD films as a function of added water amount was characterized using AFM measurements (Figure 3B). When the amount of water added increased from 0% to 2%, a steep increase in surface roughness was observed, followed by a plateau with

ACS Paragon Plus Environment

16

Page 17 of 49 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

additional amounts of water. Representative AFM height images can be found in Figure S3, supporting information. The increase in surface area due to increased surface roughness may contribute to some extent, but cannot completely explain the improved dissolution performance noted above. To investigate the evolution of film morphology during dissolution, spin-coated ASD films were immersed in water for different time intervals, dried, and analyzed with the AFM. The results are shown in Figure 5. For the initially miscible ASDs, a change in surface topography, likely indicative of phase separation, occurred within 3 seconds upon contact with water at drug loadings from 15% to 70% (Figure 5A). Finely structured 40% (Figure 3A). For 33% lopinavir ASDs prepared with 2% to 40% water, additional analysis of the data suggested that the Tgs could be clustered, based on composition, into 3-4 groups: drug-rich (low Tg), polymer-rich (high Tg), and intermediate phases. This analysis ignores the observed differences in surface morphology (i.e. discrete and continuous regions), focusing instead on compositional variations across the film. The grouped Tgs and their distributions are shown in Figure 7. Based on this more extensive analysis of variation in the local Tg values, multiple populations with different local compositions can be identified within the film. Hence, considering surface topography alone appears to be insufficient to understand the compositional variations; in other words, heterogeneity occurs within the continuous and discrete regions. Heterogeneity can occur along the xy plane, or along the z-axis (film depth). For systems prepared with 40% or less water, local Tg measurements revealed heterogeneity on the xy plane within each region (continuous or discrete). In contrast, for ASDs prepared with 50% or 100% water added to the solvent, no heterogeneity was seen within each phase in the x-y direction, but mass balance cannot be achieved, suggesting heterogeneity in the z-direction with drug enrichment on the surface. In other words, for the latter systems, both phases yield low Tg values and corresponding high drug loadings (Figure 6B and D). For systems prepared with 2% to 8% water, four compositions could be delineated based on the Tg measurements, while for 10-40% water addition, 3 compositions were extracted. The proportions of each phase (determined from the occurrence of a given analysis region with a given

ACS Paragon Plus Environment

20

Page 21 of 49 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

Tg value) are shown in Figure 7B. For the 10-40% samples, the proportion of the fast-dissolving polymer-rich phase (shown in green, Figure 7B) was about 40% to 60% with slight variations. These observations agree with the results shown in Figure 3A, where drug release was found to be largely similar over this water-to-solvent ratio range. Compared to ASDs prepared with 10% water, samples prepared with 2% water consisted of less polymer-rich phases (green and purple, Tg>150 oC,

LPV%