Characterization of Phase Separation Propensity for Amorphous

Jan 9, 2017 - Use of the HPMCAS feasibility map has been proposed as a ..... at 37 °C 100% RH, and representative thermal traces are shown in Figure ...
0 downloads 0 Views 2MB Size
Subscriber access provided by UB + Fachbibliothek Chemie | (FU-Bibliothekssystem)

Article

Characterization of Phase Separation Propensity for Amorphous Spray Dried Dispersions Daniel McNamara, Shawn Yin, Duohai Pan, George Crull, Peter Timmins, and Balvinder Vig Mol. Pharmaceutics, Just Accepted Manuscript • DOI: 10.1021/acs.molpharmaceut.6b00722 • Publication Date (Web): 09 Jan 2017 Downloaded from http://pubs.acs.org on January 10, 2017

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 free 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 accessible to all readers and 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.

Molecular Pharmaceutics 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 24

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

Title: Characterization of Phase Separation Propensity for Amorphous Spray Dried Dispersions

Daniel McNamara†*, Shawn Yin†, Duohai Pan†, George Crull†, Peter Timmins‡§, Balvinder Vig† †

Drug Product Science and Technology, Bristol-Myers Squibb, One Squibb Drive, New Brunswick NJ 08903 Drug Product Science and Technology, Bristol-Myers Squibb, Reeds Lane, Moreton, Merseyside CH46 1QW, UK § current: Department of Pharmacy, University of Huddersfield, Queensgate, Huddersfield HD1 3DH, UK ‡

*

corresponding author

ACS Paragon Plus Environment

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 24

Abstract: A generalized screening approach, applying isothermal calorimetry at 37°C 100%RH, to formulations of spray dried dispersions (SDDs) for two active pharmaceutical ingredients (APIs) (BMS-903452 and BMS-986034) is demonstrated. APIs 452 and 034, with similar chemotypes, were synthesized and promoted during development for oral dosing. Both APIs were formulated as SDDs for animal exposure studies using the polymer hydroxypropylmethlycellulose acetyl succinate M grade (HPMCAS-M). 452 formulated at 30% (wt/wt%) was an extremely robust SDD that was able to withstand 40°C 75%RH open storage conditions for 6 months with no physical evidence of crystallization or loss of dissolution performance. Though 034 was a chemical analog with similar physical chemical properties to 452, a physically stable SDD of 034 could not be formulated in HPMCAS-M at any of the drug loads attempted. This study was used to develop experience with specific physical characterization laboratory techniques to evaluate the physical stability of SDDs and to characterize the propensity of SDDs to phase separate and possibly crystalize. The screening strategy adopted was to stress the formulated SDDs with a temperature humidity screen, within the calorimeter, and to apply orthogonal analytical techniques to gain a more informed understanding of why these SDDs formulated with HPMCAS-M demonstrated such different physical stability. Isothermal calorimetry (Thermal Activity Monitor, TAM) was employed as a primary stress screen wherein the SDD formulations were monitored for 3 days at 37°C 100% RH for signs of phase separation and possible crystallization of API. Powder x-ray diffraction (pXRD), modulated differential scanning calorimetry (mDSC), Fourier transform infrared spectroscopy (FTIR), and solid state nuclear magnetic resonance (ssNMR) were all used to examine formulated SDDs and neat amorphous drug. 452 SDDs formulated at 30% (wt/wt%) or less did not show phase separation behavior upon exposure to 37°C 100% RH for 3 days. 034 SDD formulations from 10 through 50% (wt/wt%) all demonstrated thermal traces consistent with exothermic phase separation events over 3 days at 37°C 100% RH in the TAM. However, only the 15, 30, and 50% containing 034 samples showed pXRD patterns consistent with crystalline material in post-TAM samples. Isothermal calorimetry is a useful screening tool to probe robust SDD physical performance and help investigate the level of drug polymer miscibility under a humid stress. Orthogonal analytical techniques such as pXRD, ssNMR and FTIR were key in this SDD formulation screening to gain physical understanding and confirm or refute whether physical changes occur during the observed thermal events characterized by the calorimetric screening experiments. Keywords: amorphous, amorphous amorphous phase separation, spray dried dispersions, crystallization, physical stability, isothermal calorimetry, SDD formulation screen

ACS Paragon Plus Environment

Page 3 of 24

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

Introduction: As part of a research effort to screen for oral activity, two compounds 452 and 034 were selected and scaled-up for development activities. In order to maximize oral exposure in animal studies amorphous spray dried dispersions (SDDs) of each compound were formulated and evaluated for physical stability and performance. The SDD’s consisted solely of amorphous API and a polymer (HPMCAS M grade). No other excipients were employed to formulate the SDDs. The chemical structures of 452 and 034 are listed in Table 1 along with key physical chemical characteristics for each API. Both APIs can be categorized as BCS Class II demonstrating low aqueous solubility and high membrane permeability. 034 has higher aqueous solubility than 452. As is typical with BCS Class II compounds, the dissolution performance of the oral dosage form is key to maximizing oral delivery and exposure. Based on a Tm/Tg (K/K) ratio versus Log P formulation feasibility map for HPMCAS, a 30% (wt/wt%) drug load was selected for 452 in HPMCAS-M grade as an initial formulation. The formulation feasibility map derives from hands on experience with over 130 compounds formulated in HPMCAS wherein the Tm/Tg ratio for the API is used as a combined thermodynamic (Tm) and kinetic (Tg) parameterization of the propensity of the API to crystallize and log P of the API positions the API relative lipophilicity as a second parameter. Use of the HPMCAS feasibility map has been proposed as a rationale for how the SDD formulator might select drug loads to maximize likelihood of selecting a robust HPMCAS SDD formulation.1,2 The 30% HPMCAS-M SDD of 452 proved to be a robust formulation with exemplary physical stability and consistent dissolution performance as it was able to withstand 6 months storage at 40°C 75% RH open storage conditions. That is, none of the physical characteristics of the SDD had changed over the 6 months (still pXRD amorphous) and the dissolution performance of SDD samples stored for 6 months at 40°C 75% RH were comparable to initial time zero SDD samples. As the physical chemical properties of 034 where grossly comparable to the lead 452, though the Tm/Tg ratio was slightly different, a 30% drug load in HPMCAS-M was initially selected to be the initial trial SDD formulation. 034 at 30% drug load in HPMCAS-M proved to be a physically unstable formulation with poor physical stability and poor dissolution profile characteristics. This study was undertaken to gain a more nuanced understanding of why SDDs formulated with HPMCAS-M for 452 and 034 were so unique and different. In order to develop a broader understanding of the performance and physical stability characteristics of HPMCAS-M SDDs made with 452 and 034, a matrix of SDDs at various API loads were formulated on lab-scale spray drying equipment. 452 SDDs were formulated at: 10, 30, and 50% (wt/wt%), whilst 034 SDDs were made at 10, 15, 30, and 50% (wt/wt%). In addition to making SDDs with varying API compositions, neat amorphous API material for both 452 and 034 were also spray dried and characterized using the ensemble of techniques. mDSC was used to determine glass transition temperatures for both neat amorphous APIs and all SDDs. Water sorption isotherms for both neat APIs at 25°C were also determined. Finally, HPMCAS-M (polymer alone) was also spray dried and characterized.

ACS Paragon Plus Environment

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 24

Experimental Section: Materials: APIs: Single lots of 452 and 034 with purity contents (>99% by internal HPLC assay methods) were used in this study. Polymer: A single lot of HPMCAS-M grade granular from Shin-Etsu was used to make all SDDs. SDDs: Fresh gram-scale batches formulated at: 10, 30, 50% (wt%) 452 and 10, 15, 30, and 50% 034 SDDs were manufactured using a custom built small-scale spray dryer. Spray dry solutions of drug and HPMCAS-M granular grade usually containing 2.5% solids content (API + polymer) were sprayed at 65-70°C from dichloromethane solvent using heated nitrogen gas through a two-fluid spraying nozzle (2050 LC/64AC, Spraying Systems Co.). Within the spray dryer, solid material was collected on a 4” filter paper. After scraping the SDD sample from the filter paper, the solid was vacuum dried overnight in a glass vial at RT and stored over desiccant. Neat 452, 034, and HPMCAS-M materials were also spray dried in the same equipment and vacuum dried overnight at RT. Powder X-ray diffraction (pXRD) was always run after the overnight vacuum to confirm amorphous character of the respective samples. Physical Mixes: Physical mixes of both 452 and 034 were made by weighing and mixing spray dried amorphous API with spray dried HPMCAS-M at 50 wt %. Mixing was accomplished using a glass scintillation vial and a vertical rotating turntable. Physical mixes were also stored over desiccant. TAM procedure: A TAM III (TA Instruments, DE, USA) thermostat equipped with 4mL nanocalorimeters at 37°C was used. Approximately 200 mg of sample was weighed into a 3 mL glass ampoule and a 250 µL glass insert, filled with 200 µL distilled water, was also added to the ampoule. The ampoule was crimped and sealed with an aluminum cap fitted with a silicone gasket. A similar reference glass ampoule was crimped containing a glass insert and distilled water. Sample and reference ampoules were equilibrated for 15 minutes prior to being lowered into the calorimeter measurement position. Normalized enthalpy (J/g) was calculated for thermal trace data by integrating the power-time curve to baseline. pXRD procedure: All Powder X-ray Diffraction (PXRD) patterns of spray dried materials were recorded on a Bruker D4 Endeavor/Lynxeye (Bruker, Germany) X-ray powder diffractometer with Cu Kα radiation (λ = 1.541 Å). The diffractometer was equipped with a ceramic tube which was set at the power level of 40kV and 40mA (1.6 KW), and a position sensitive lynxeye detector. Incident optics consisted of a 0.6º divergence slit. Diffracted optics consisted of Ni-K-Beta filter and a 3º opening on detector window. Data were collected in reflectance geometry, whilst spinning, over a 2θ range of 4–32°, with a step size of 0.03°, and counting time of 1 sec/step in continuous mode. mDSC procedure: A DSC Q1000 (TA Instruments, DE, USA) system was used. 2-5mg of sample was crimped in an aluminum DSC pan and a dry nitrogen purge was applied. A heating ramp of 2.5°C /min. was applied from 25 to 200°C with a modulation amplitude of 1.5°C applied every 60 seconds. Water Isotherm 25°C: Moisture sorption isotherms were collected in a VTI SGA-100 Symmetric Vapor Analyzer using approximately 10 mg of sample. The sample was dried at 25°C until the loss rate of 0.0050 wt %/min was obtained for 10 ACS Paragon Plus Environment

Page 5 of 24

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

minutes. The sample was tested at 25°C and 4, 5, 15, 25, 35, 45, 50, 65, 75, 85, 95% RH. Equilibration at each RH was reached when the rate of 0.010 wt%/min for 35 minutes was achieved or a maximum of 600 minutes. FTIR procedure: The IR spectra were collected using a FTIR spectrometer with a total internal reflection (ATR) (IS50 FTIR, ThermoFisher Scientific, Madison, WI, USA). Each spectrum was measured by accumulating 64 scans at 4 cm-1 spectral resolution over the range from 500 cm-1 to 4000 cm-1 at ambient temperature. ssNMR procedure: All measurements were conducted on a Bruker AV III instrument (Bruker-Biosystems, Billerica, MA) operating at a proton frequency of 400.13 MHz and using a 3 frequency 4 mm MAS probe . Approximately 70 mg of sample was used for each experiment. The samples were confined in a 4 mm ZrO2 rotors. Samples were typically spun at 13KHz. The sample temperature was controlled by cooling the drive gas using a Bruker BCU extreme. Proton MAS and Proton T1 MAS measurements were collected on the same spectrometer. The spectral sweep width was 500 ppm centered at 100 ppm. 4096 data points were acquired and zero filled to 8192 prior to apodization with 20 Hz line broadening. Typically 2096 free induction decays were coadded. The spectra were referenced indirectly to TMS (0 ppm) using water (5 ppm).

ACS Paragon Plus Environment

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 24

Results and Discussion: Neat Amorphous API Characterization: Water isotherms for both neat amorphous APIs are shown in Figure 1. 034 sorbs significantly more water than 452 at all humidity conditions evaluated in the 25°C isotherm. For dry neat amorphous APIs, the Tg values for both APIs are similar (96 and 106°C). Hence, for dry neat meta-stable glasses at room temperature a simple expectation is that both APIs would have similar crystallization propensities. However, when equilibrated at even moderate humidity, it is expected that 034 would be effectively plasticized by sorbed water to a much greater extent than 452 and would be more likely to crystallize. pXRD run on post-isotherm samples for both APIs confirmed that the samples were still amorphous. SDD Characterization: Figure 2 shows the experimentally determined Tgs with API drug load for all SDDs. Also shown in Figure 2 are theoretical fits of Tg versus API load applying the Gordon Taylor (G-T) equation. SDDs made with both APIs show negative deviations from the G-T theoretical predictions, though 034 containing SDDs at 50 and 30% drug load have lower Tgs than pure 034 alone. Generally, negative G-T deviations are assumed to result from weak drug-polymer interaction in the SDD as compared to the drug-drug or polymer-polymer interactions of the pure components at similar compositions.3 It has been suggested that dispersions which exhibit negative G-T deviations have a tendency to de-mix and phase separate.4 Another BCS II API formulated in HPMCAS-M SDDs which showed significant negative G-T deviations, below the Tg of the neat API, was shown to phase separate and crystalize upon exposure to a temperature and humidity challenge when the drug load was greater than 50%.5 Since 034 SDDs demonstrate the most negative G-T deviations, one could infer that 034-HPMCAS-M (drugpolymer) interactions are weaker than 452-HPMCAS-M interactions. The primary goal of this study was to more fundamentally understand the behavior and robustness of SDDs made with 452 or 034. To accomplish this we sought to develop practical screening techniques to assess the phase separation propensity of the formulated SDDs. Our practical experience was that 452 at 30% API load in HPMCAS-M was a robust formulation that withstood temperature and humidity challenges like 40°C 75% RH for six months in open conditions and always demonstrated good in-vitro dissolution performance with no physical changes as detected by pXRD. In a microcentrifuge dissolution evaluation good in-vitro dissolution performance for 452 SDD samples meant that the 452 SDD samples maintained an API concentration of 200µg/mL for over 90 minutes in an aqueous media.1,2 In contrast, 034 at 30% API load in HPMCAS-M performed poorly in initial dissolution evaluations and could not be stored at 50°C 75% RH overnight without demonstrating phase separation and crystallization of API. Since water vapor is a ubiquitous environmental stressor for all pharmaceutical dosage forms, including SDD systems, and aqueous dissolution is an important performance evaluation for SDDs we determined to incorporate moisture exposure into our SDD formulation screening regime. Other researchers interested in amorphous-amorphous phase separation (AAPS) behavior of SDDs have similarly adopted high moisture challenges to stress SDD formulations.6-10 Further, dissolution mechanisms that have been proposed for SDD systems posit: phase separation with API nanostructure formation are likely, or alternatively that erosion of hydrated polymer containing amorphous API without phase separation occurs.2 We hypothesized that if one could characterize how a trial SDD reacted to a moisture challenge, it would lead to a more fundamental understanding of which of the dissolution mechanisms, including phase separation, dominated SDD dissolution performance. Further we were interested to examine how drug loading of the SDD might impact the SDD dissolution mechanism and phase separation propensity. All of these factors led us to carefully evaluate isothermal calorimetry as a possible primary experimental SDD screening technique. Moisture can be introduced to calorimetric samples and modern commercially available instruments are quite sensitive. Also important was that the screen be relatively fast (days) give reproducible results and require small amounts of formulation available during lab-scale preliminary formulation designs. We also recognized that a thermal technique like ACS Paragon Plus Environment

Page 7 of 24

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

calorimetry would require orthogonal analytical techniques like pXRD, ssNMR, and FTIR to confirm possible phase separation or crystallization events that might accompany calorimetrically detected thermal events. As an experimental control, spray dried HPMCAS-M grade (SD HPMCAS-M) was produced and characterized using isothermal calorimetry (TAM III ampoule experiment) using 37°C 100% RH as a stress screening condition. Figure 3 shows the thermal traces for four samples of SD HPMCAS-M measured with the TAM. Initially, after a 15 minute temperature partial equilibration, the ampoule containing the solid sample and vial containing water are lowered into the TAM III thermal detection zone. Within 30 minutes of being lowered and additional equilibrating at 37°C 100% RH the SD HPMCAS-M solid begins an inexorable thermal enthalpic relaxation. The extent and kinetics of this exothermic thermal relaxation are fundamental characteristic of the amorphous glass produced by spray drying. Many pharmaceutical researchers have documented that isothermal calorimetry (specifically TAM) is an applicable technique to characterize exothermic relaxation of pharmaceutical glassy materials.11-14 A previous paper detailed efforts to investigate SDDs phase separation with temperature and humidity challenges via isothermal microcalorimetry applying a TAM 2277.15 However, it was noted by Calahan et al that baseline drift inherent to the TAM 2277 limited the sensitivity and detection of low heat flux ( < 2µW) events especially over long experimental run times.15 General performance test specifications for a well maintained and performing TAM 2277 equipped with a 4 mL ampoule calorimeter and standard amplifier stipulate a baseline drift specification of not more than 300 nW over 24 hours. Comparatively, the TAM III nanocalorimeters used herein have baseline drift specifications of less than 40 nW over 24 hours. From a sensitivity and capacity to detect or characterize small thermal signals over days, the TAM III nanocalorimeter has significant advantages compared to the TAM 2277 4 mL calorimeter. Over the course of 72 hours, -16.7 (+0.5) J/g of heat is given off by the SD HPMCAS-M glass as it structurally relaxes and anneals. pXRD patterns of SD HPMCAS-M samples post-TAM do not show any appreciable differences indicating that no phase separation or crystallization occurred during the 37°C 100% RH screen over 72 hours given the scale of domain resolution consistent with pXRD. However, ssNMR 1H T1 times determined for SD HPMCAS-M of 1.8 s, increase to 2.3 s post-TAM, indicating that the polymer has become more ordered as a result of being exposed to 37°C 100% RH for 72 hours and that this ordering process is consistent with structural annealing of a glass (Table 2). Other researchers have reported similar 1H T1 times for SD HPMCAS-M.5 Since it is generally accepted that ssNMR can be used to probe smaller domain sizes (~50 nm) than wide angle scattering pXRD, it is not surprising that ssNMR did detect a subtle difference in the annealed HPMCAS—M samples post-TAM while the pXRD patterns pre and post were comparable.16 Thermal traces for 10, 30, and 50% drug loaded 452 SDD samples are shown in Figure 4 for exposure over 72 hours at 37°C 100% RH. Both the 10 and 30% SDD samples show the expected exponential decay consistent with enthalpic relaxation of a homogeneous amorphous glass. The 50% drug loaded 452 SDD thermal trace is unusual: an increasingly exothermic heat signal is recorded after some two days at 37°C 100% RH. The exothermic heat signal increases in magnitude from 48 to 72 hours and when the 50% SDD samples were checked by pXRD post-TAM there were peaks in the diffraction pattern indicating phase separation and crystallization of solid 452 had occurred (Figure 5). The exothermic signal detected for the 50% 452 SDD while stressed at 37°C 100% RH is indicative of phase separation (a necessary a priori event to API crystallization) and subsequent API cystallization. ssNMR 1H T1 results for the 50% SDD before and after thermal stress also show a significant difference for the relaxation time after thermal stress (Table 2). Figure 4 shows the thermal trace for a 50:50 physical mix of amorphous 452 and SD HPMCAS-M. When stressed, the physical mixture of amorphous API and SD HPMCAS-M (already phase-separated) demonstrates a fast large exothermic signal which corresponds to crystalline phase being detected in the post-TAM pXRD (Figure 5). Note that both the 10 and 30% 452 postTAM pXRD results confirm that these samples remained amorphous. ssNMR 1H T1 characterization of the postTAM 10 and 30% SDD samples also show no change in relaxation time. ACS Paragon Plus Environment

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 24

Similar TAM screening was performed for 034 SDDs formulated with drug loads from 10 through 50% at 37°C 100% RH and representative thermal traces are shown in Figure 6. All of the 034 SDDs samples, for all drug loads, showed exothermic events in the TAM thermal traces. The thermal signal was so large for the 50% drug load 034 SDD samples that it never returned to baseline even after 72 hours. A duplicate sample of the 50% 034 SDD also did not return to below 130 µW/g for 72 hours. For the 10, 15, and 30% 034 SDDs there is a linear relationship between the API load and the enthalpy associated with the exothermic event signal. Higher drug loading is associated with faster exothermic event peak times and larger enthalpies for the events. Figure 7 shows the linear correlation between drug loading and enthalpy of the exothermic event detected in the TAM screen. pXRD of 034 SDD samples post-TAM show that all of the formulations crystallized except for the 10% API load (Figure 8). Figure 9(a) shows the FTIR second derivative spectra of 452 SDDs along with amorphous API and spray dried polymer in the range from 1600 cm-1 to 1800 cm-1. The C=O peak at 1735 cm-1 for HPMCAS remains unchanged in the SDDs as compared with that of HPMCAS alone. In contrast, a new peak at 1648 cm-1 was found in SDDs formulated with drug loads from 10% to 50%. The appearance of the peak at 1648 cm-1 in SDDs is likely due to the formation of hydrogen bonding between API N=C-N and the hydroxyl group in polymer. The 1648 cm-1 peak intensity increases in the SDDs from 10% to 30%. However, for SDD with a drug load of 50%, this new peak decreases in intensity compared to that in 30%. This result indicates that the strong hydrogen bonding between API and polymer was formed at 10% and 30% drug load, but this interaction is diminished at 50% API loading. FTIR second derivative spectra of 034 SDDs are shown in Figure 9(b). In contrast with the 452 SDDs, no significant changes of IR spectra in the ranges from 1600 cm-1 to 1800 cm-1 in any of the 034 SDDs at 10%, 30% and 50% API loading were detected. In addition, no shift of OH group at 3350 cm-1 of the API (not shown in the figure) was observed in the 034 SDDs. The peak at 1705 cm-1 in SDDs and amorphous API is likely attributable to an API-API interaction. Overall, 034 SDD FTIR comparisons showed that the API-polymer interaction is weak or even nonexistent when compared to that demonstrated by 452 SDDs formulated through 30% drug load. Neat Amorphous API Characterization: To further characterize crystallization propensity of the neat amorphous APIs, multiple samples were studied by TAM at 37°C 100% RH. Figures 10-11 show the heat flux versus time thermal traces for the neat amorphous APIs when exposed to 37°C 100% RH. Both 452 and 034 show large exothermic events (> -45 J/g) in the TAM thermal traces consistent with an amorphous solid becoming phase separated and then crystallizing. Figure 10 for 452 shows a single broad exotherm with a peak at approximately 1.6 days while the 034 thermal trace in Figure 11 shows two exothermic events which peak between 16-20 hours and 40-44 hours, respectively. pXRD for 452 samples after 3 days at 37°C 100% RH confirm presence of crystalline API (Figure 5). Figure 12 shows pXRD performed on a 034 samples after 24 hours and also after 72 hours. pXRD confirms that after 24 hours exposure at 37°C 100% RH, 034 is still pXRD amorphous albeit with a different amorphous halo then seen prior to the thermal temperature and humidity stress imposed in the TAM screen. After 72 hours at 37°C 100% RH, pXRD spectra of 034 shows crystalline peaks consistent with a known solid phase of 034. One interpretation of the TAM thermal tracing is that the initial exothermic event near 18 hours in the TAM for 034 is indicative of the amorphous phase becoming more ordered which is subsequently followed by a solid phase crystallization exotherm at 42 hours. Figure 5 shows pXRD for 452 samples after temperature and humidity stress in the TAM which confirm that solid crystallization did occur; however, the single exotherm indicates that there was not a distinct separate phase ordering event followed by a crystallization: phase ordering and crystallization appear more as a single concomitant event. An isothermal calorimetric screening experiment wherein samples are evaluated at 37°C 100% RH for thermal events appears to be a useful screening technique to evaluate trial HPMCAS-M SDDs for physical robustness and ACS Paragon Plus Environment

Page 9 of 24

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

phase separation tendency. Spray dried amorphous HPMCAS-M polymer shows the expected enthalpic exothermic relaxation at 37°C 100% RH which can be used as a control to compare to thermal traces for trial SDD formulations. ssNMR 1H T1 times confirm that SD HPMCAS-M is more ordered after annealing in the TAM at 37°C 100% RH for 72 hours. This behavior of SD HPMCAS-M is consistent with enthalpic relaxation of an amorphous polymeric glass. During structural relaxation or annealing of an SDD in the calorimeter that is prone to AAPS and possibly crystallization it is expected that additional exothermic events, in addition to the enthalpic relaxation of the polymer, would be detected. As AAPS is a demixing process that is entropically unfavorable, it could be expected that an exothermic enthalpic signal should also accompany this process. Similarly, crystallization of the API within an amorphous-amorphous phase separated SDD should also register as an exothermic enthalpic event. Regarding the differences in physical robustness of SDDs formulated with 452 and 034 in HPMCAS-M, the isothermal calorimetric screening for 452 SDDs at 30% API load and below produced thermal traces without additional exothermic events. These thermal traces are consistent with enthalpic relaxation of a single homogeneous amorphous phase. However, 452 SDD at 50% drug load showed an exothermic thermal trace consistent with phase separation and eventual API crystallization. ssNMR 1H T1 analyses also confirm a significant difference in relaxation time for the 50% SDD after thermal annealing consistent with pXRD evidence of crystalline phase. SDDs formulated at 50% API load for 452 in HPMCAS-M are above the miscibility limit for the drug and polymer when stressed at 37°C 100% RH for days. All of the 034 SDDs formulated from 10 through 50% showed exothermic events in the isothermal calorimetric thermal trace screens. All of these 034 SDDs demonstrated crystalline API post-TAM screening. The only exception to this is the 10% 034 SDDs which did show an exothermic event but did not show crystalline solid phase by pXRD after 72 hours at 37°C 100% RH. It is likely that at 10% 034 in HPMCAS-M, the drug load is low enough in the SDD that API molecules must diffuse so far that a local critical nuclei concentration is never exceeded, hence no crystallization event or crystalline API results even though the SDD demonstrates AAPS exothermic behavior in the TAM stress. FTIR characterization of the API-polymer interaction was instructive and helpful in that the interaction between 452 and HPMCAS-M for SDDs formulated at 30% or below helped explain their inherent physical robustness as compared to the physical liabilities of the 034 SDDs across all drug loads tested. Gordon-Taylor analyses of the 452 and 034 SDD Tgs with HPMCAS-M content further bolsters the hypothesis that 034 interacted less with HPMCAS-M than 452. pXRD results on post-TAM samples were key to interpretation of thermal events observed during the isothermal calorimetric screening. An exothermic event during the 37°C 100% RH screening did not guarantee crystalline API in the SDD. In the case of the 034 10% SDD, we suspect that the TAM-detected exothermic event was associated with general AAPS but, even with phase heterogeneities, the local concentration of the API never exceeded a crystallization nuclei threshold. Examination of the plasticization behavior of neat amorphous API with water is also important to understand how likely the amorphous API is to crystallize when formulated in an SDD. It is well known that HPMCAS-M sorbs significant amounts of water at humidities greater than 75% RH.1 From the water isotherm data alone for neat amorphous APIs, one would expect 034 to be much more susceptible to plasticization by water and hence prone to crystallize when exposed to ambient or higher moisture stresses than 452. In fact, at high humidities using the rule of thumb of 10°C diminution of Tg for every 1% water sorbed, the effective Tg would drop enough that 034 could be expected to crystallize at 30°C. Applying the same water induced Tg plasticization argument, 452 would not be expected to crystallize at high humidities and room temperature. However, in this specific instance, both neat amorphous APIs do crystallize at 37°C 100% RH. Neat 034 does phase order more quickly than 452 from the TAM thermal traces suggesting that kinetically amorphous 034 is more physically labile. ACS Paragon Plus Environment

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 24

Conclusions: This example of two chemically related API’s, with similar Tm/Tg ratios, demonstrates that using the Tm/Tg ratio and log P as the primary metrics to formulate SDDs from an empirically constructed historical feasibility map derived for a polymer may be too simplistic. Application of the Tm/Tg metric may be a reasonable initial starting point to formulate an SDD, but the feasibility map approach does not consider the importance of the drug-polymer interaction and it is particularly uninstructive especially when it fails. The isothermal calorimetric screening approach detailed in this study can be applied as a tripwire to allow the formulator to screen for polymers with favorable API-polymer interactions and define the drug load which would be less likely to result in physically unstable SDDs that phase separate upon a humidity stress. Phase separation behavior does not always guarantee that crystallization or physical failure of the SDD will follow: this axiom was verified in the example of 034 at 10% drug load in a HPMCAS-M SDD which showed evidence of phase separation by TAM, yet there were no peaks in the pXRD results after 72 hours in the TAM screen. However, the 15% SDD containing 034, showing larger exothermic phase separation TAM behavior than the 10% SDD, did physically fail as pXRD peaks consistent with crystallization were detected in the post-TAM stressed sample. Alternatively, 452 formulated at 30% drug load and 10% showed no phase separation tendency. The 30% 452 formulated in HPMCAS-M was a robust SDD proved by real time stability results. A 50% SDD formulated with 452 showed exothermic phase separation behavior by TAM, physical failure by pXRD (crystallization post-TAM), and FTIR results which suggested that the drug-polymer interaction had saturated between 30 and 50% drug load. The 50% 452 SDD demonstrated poor initial dissolution also (data not shown in this paper). At the very least by applying the isothermal calorimetric screen, the formulator may also gain insight into how the SDD behaves mechanistically during dissolution testing. Especially important though is the use of orthogonal analytical techniques like FTIR, ssNMR, and pXRD to round out a genuine understanding of the physical complexities associated with these systems. Formulators would also do well to characterize the water sorptive characteristics of neat amorphous API as they rapidly attempt to create physically robust SDD formulations which can survive commercial expiry periods under humidity temperature stress.

ACS Paragon Plus Environment

Page 11 of 24

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

Acknowledgement: Authors thank C. Nunes for historical development project background and briefings and K. Stefanski for instruction on use of the Bend Mini. DM thanks G. Zografi, S. Gaisford, J. Zhao, S. Nicholson, R. Haskell, and M. Hussain for helpful suggestions and comments on various drafts of the manuscript.

ACS Paragon Plus Environment

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 24

References: (1) Friesen, D. T., Shanker, R., Crew, M., Smithey, D. T., Curatolo, W. J., Nightingale, J. A. S. Hydroxypropyl Methylcellulose Acetate Succinate-Based Spray-Dried Dispersions: An Overview. Molecular Pharm., 2008, 5, pp 1003-1019. (2) Vodak, D. T.; Morgen, M. Chapter 9, Design and Development of HPMCAS-Based Spray-Dried Dispersions. In Amorphous Solid Dispersions Theory and Practice, Shah, N., Sandhu, H., Choi, D. S., Chokshi, H., Malick, A. W., Springer.: New York, 2014; pp 303-322. (3) Crowley, K. J.; Zografi, G. Water Vapor Absorption into Amorphous Hydrophobic Drug/Poly(vinylpyrrolidone) Dispersions. J. Pharm. Sci., 2002, 91, 2150-2165. (4) Shamblin, S. L., Huang, E. Y., Zografi, G. The Effects of Co-Lyophilized Polymeric Additives on the Glass Transition Temperature and Crystallization of Amorphous Sucrose. J. Thermal Analysis, 1996, 47, 1567-1579. (5) Calahan, J. L., Azali, S. C., Munson, E. J., Nagapudi, K. Investigation of Phase Mixing in Amorphous Solid Dispersions of AMG 517 in HPMC-AS Using DSC, Solid-State NMR, and Solution Calorimetry. Molecular Pharm., 2015, 12, pp 4115-4123. (6) Marsac, P. J., Konno, H., Rumondor, A. C., Taylor, L. S. Recrystallization of Nifedipine and Felopdipine from Amorphous Molecular Level Solid Dispersions Containing Poly(vinylpyrrolidone) and Sorbed Water. Pharm. Res., 2008, 25, pp 647-656. (7) Rumondor, A., Taylor, L. S. Effect of Polymer Hygroscopicity on the Phase Behavior of Amorphous Solid Dispersions in the Presence of Moisture. Molecular Pharm., 2010, 7, pp 477-490. (8) Rumondor, A., Wikstrom, H., van Eerdenbrugh, B., Taylor, L. S. Understanding the Tendency of Amorphous Solid Dispersions to Undergo Amorphous-Amorphous Phase Separation in the Presence of Absorbed Moisture. AAPS PharmSciTech, 2011, 12, pp 1209-1219. (9) Qi, S., Moffat, J. G., Yang, Z. Early Stage Phase Separation in Pharmaceutical Solid Dispersion Thin Films under High Humidity: Improved Spatial Understanding Using Probe-Based Thermal and Spectroscopic Nanocharacterization Methods. Molecular Pharm., 2013, 10, 918-930. (10) Purohit, H. S., Taylor, L. S. Phase Separation Kinetics in Amorphous Solid Dispersions Upon Exposure to Water. Molecular Pharm., 2015, 12, pp 1623-1635. (11) Liu, J., Rigsbee, D. R., Stotz, C., Pikal, M. J. Dynamics of Pharmaceutical Amorphous Solids: The Study of Enthalpy Relaxation by Isothermal Microcalorimetry. J. Pharm. Sci., 2002, 91, pp 1853-1862. (12) Kawakami, K., Ida, Y., Direct Observation of the Enthalpy Relaxation and the Recovery Processes of Maltose-Based Amorphous Formulation by Isothermal Microcalorimetry. Pharm. Res., 2005, 20, pp 1430-1436. (13) Kawakami, K., Pikal, M. J. Calorimetric Investigation of the Structural Relaxation of Amorphous Materials: Evaluating Validity of the Methodologies. J. Pharm. Sci., 2005, 94, pp 948-964. (14) Shamblin, S. L., Hancock, B. C., Pikal, M. J. Coupling Between Chemical Reactivity and Structural Relaxation in Pharmaceutical Glasses. Pharm. Res., 2006, 10, pp 2254-2268. (15) Calahan, J. L., Zanon, R. L., Alvarez-Nunez, F., Munson, E. J. Isothermal Microcalorimetry to Investigate Phase Separation for Amorphous Solid Dispersions of AMG 517 with HPMC-AS. Molecular Pharm., 2013, 10, pp 1949-1957. (16) Yuan, X., Sperger, D., Munson, E. R. Investigating Miscibility and Molecular Mobility of Nifedipine-PVP Amorphous Solid Dispersions Using Solid-State NMR Spectroscopy. Molecular Pharm., 2014, 11, 329-337.

ACS Paragon Plus Environment

Page 13 of 24

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

List of Tables TABLE 1: Summary Comparison of Physical Chemical Properties of APIs 452

034

Chemical Structure

Molecular Weight 512 530 cLog P† 2.2 2.8 Log P 4.2 N/A Tm (°C) 248 300 Tg (°C) 96 106 Tm/Tg (K) 1.3 1.5 Aqueous Solubility PBS (µg/mL)‡ 48 hr)) ‡

10% SDD post-TAM 1.9† -30% SDD post-TAM 1.9† -† 50% SDD post-TAM 3.3 -value in brackets () is standard deviation † experimental error for each measurement is typically less than 10% ‡ see Figure 4

ACS Paragon Plus Environment

Molecular Pharmaceutics

List of Figures: 12 10

Weight (%)

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 24

8 6 4 2 0 0

20

40

60

80

100

RH (%)

Figure 1. Water Isotherms for neat amorphous 452 and 034 at 25°C. Key: 452 adsorption open circle, 452 desorption half-filled circle, 034 adsorption open square, 034 desorption half-filled square.

ACS Paragon Plus Environment

Page 15 of 24

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

Figure 2. Tg with API composition for HPMCAS-M SDDs for (A) 452 and (B) 034 at room temperature. Gordon-Taylor theoretical plots in red calculated using true densities for both APIs and SD HPMCAS-M. (Helium pycnometry was used to determine true density, of SD HPMCAS [=] 1.21, 452[=] 1.37, and 034[=] 1.33g/mL, respectively).

ACS Paragon Plus Environment

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

Figure 3. Thermal traces of SD HPMCAS-M at 37°C 100% RH for four samples. Inset plot shows full-scale normalized heat flow thermal trace for one sample. Exothermic heat flow is up.

ACS Paragon Plus Environment

Page 16 of 24

Page 17 of 24

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

Figure 4. Thermal traces for 452 10, 30, and 50% (wt/wt%) SDDs in HPMCAS-M and 50:50 physical mix at 37°C 100% RH. Exothermic heat flow is up.

Figure 5. pXRD of 452 SDD samples and physical mix post-TAM exposure to 37°C 100% RH. (a) simulated 452 N-2 form, all other pXRD results for samples post-TAM: (b) spray dried neat 452, (c) 50:50 physical mix, (d) 50% SDD, (e) 30% SDD, (f) 10% SDD. ACS Paragon Plus Environment

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 18 of 24

Figure 6. Thermal traces for 034 10, 15, 30, and 50% (wt/wt%) SDDs in HPMCAS-M and 50:50 physical mix at 37°C 100% RH. Exothermic heat flow is up.

ACS Paragon Plus Environment

Page 19 of 24

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

Figure 7. Enthalpy as a function of 034 SDD API drug load. Error bars represent (+ one standard deviation). R2=0.97, intercept is not significantly different than zero. Enthalpy normalized for sample weight.

Figure 8. pXRD of 034 SDD samples and neat SD 034 post-TAM exposure to 37°C 100% RH. (a) simulated 034 H2-3 form, (b) simulated 034 N-1 form, all other pXRD results for samples post-TAM: (c) spray dried neat 034, (d) 50:50 physical mix, (e) 50% SDD, (f) 30% SDD, (g) 15% SDD, (h) 10% SDD. All samples show crystallization except for 10% API load.

ACS Paragon Plus Environment

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

Figure 9. FTIR Summary of (a) 452 SDDs and (b) 034 SDDs with drug loading.

ACS Paragon Plus Environment

Page 20 of 24

Page 21 of 24

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

Figure 10. Thermal traces for three neat amorphous 452 samples at 37°C 100% RH. Exothermic heat flow is up.

ACS Paragon Plus Environment

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

Figure 11. Thermal traces for three neat amorphous 034 samples at 37°C 100% RH. One sample was pulled from TAM after 24 hours. Exothermic heat flow is up.

ACS Paragon Plus Environment

Page 22 of 24

Page 23 of 24

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

Figure 12. pXRD for neat amorphous spray dried 034 samples after 37°C 100% RH TAM exposure. (a) simulated 034 N-1 form, (b) 034 after 3 days in TAM, (c) 034 after 1 day in TAM, (d) initial spray dried 034 (no TAM exposure).

ACS Paragon Plus Environment

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

338x190mm (96 x 96 DPI)

ACS Paragon Plus Environment

Page 24 of 24