(W1030-11-73) Investigation of a Ternary Amorphous Solid Dispersion via Confocal Raman Spectroscopy: Where Is the Drug?

Purpose: The formulation principle of amorphous solid dispersions (ASDs) is a well-established technique for overcoming poor water solubility of drugs by embedding the drug amorphously into the polymer matrix. Combining two polymers into a homogenous ternary ASD can lead to synergistic interactions, resulting in higher drug stability or superior dissolution performances compared to the corresponding binary ASDs [1]. However, miscibility of polymers for forming single-phased and homogeneous systems is not always given [2]. For a mechanistic understanding of the synergistic interaction of the two polymers in respect of the dissolution performance of the drug, elucidating the specific drug polymer interaction in phases of various polymer ratios is important. Therefore, an inhomogeneous non-single-phased ternary ASD, comprising the polymers Eudragit L 100-55 (EL 100-55) (45%), Hydroxypropyl cellulose (HPC)-SSL (45%) and the drug celecoxib (CXB) (10%) was prepared via vacuum compression molding (VCM). The distribution of CXB in the polymer matrix was investigated via confocal Raman spectroscopy (CRS). To get an impression about the maximum solid-state miscibility of CXB in each single polymer, highly drug loaded binary ASDs (CXB: EL 100-55 and CXB: HPC-SSL) were prepared via VCM and investigated using differential scanning calorimetry (DSC).

Methods: Preparation of ASDs: Physical mixtures (PMs) were prepared by blending the components in a MM400 ball mill (30 Hz, 3x 5 min milling cycles). Approx. 500 mg of the PMs were loaded into a VCM device with a 20 mm disc geometry. ASD discs were obtained after molding the PMs under vacuum using an annealing temperature of 160 °C for 15 min, followed by cooling. Confocal Raman spectroscopy (CRS): The solid-state distribution of the polymers and CXB in the ternary VCM disc was investigated by CFR using an alpha 500R confocal Raman microscope, equipped with a 532 nm excitation laser, an UHTS 300 spectrometer, and a DV401-BV CCD camera. A 40x 0.6 NA objective was selected and the laser power was set to 25 mW. In order to distinguish the components in the Raman spectra of the ternary ASD, single spectra of the neat substances were determined. Differential Scanning Calorimetry (DSC): DSC analysis was carried out, utilizing a DSC 2 instrument. Prior DSC investigation, the VCM discs were milled using a MM400 ball mill. Approx. 10 mg of the ASDs were weighed into an aluminum pan with pierced lid. The glass transition temperatures and the potential melting points were investigated conducting temperature programs with an underlying heat rate of 2 K/min from 0 to 170 °C (TOPEM mode).

Results: The investigation of the single Raman spectra of the neat substances revealed differences in peak shapes and intensities (Figure 1). CXB and EL 100-55 showed characteristic peaks at 1615 cm^-1 (green arrow) and 1730 cm^-1 (red arrow), respectively, hence these wavenumbers were used for identification of the substances. Additionally, HPC-SSL demonstrated a characteristic double peak between 2850 and 2950 cm^-1. Figure 2 represents an arbitrary selected investigated area of the CXB: EL 100-55: HPC-SSL ASD. Prior investigation, the absence of residual crystallinity in the ASD was confirmed by X-ray powder diffraction and by DSC (data not shown). The green color represents the intensity at 1615 cm^-1, indicating high presence of CXB. The red color represents the intensity at 1730 cm^-1 for detecting EL 100-55 compared to HPC-SSL. CRS demonstrated an inhomogeneous ternary ASD, as different intensities regarding the presence of CXB and EL 100-55 were detected. Interestingly, the green and the red color were complementary to each other. In areas with low content of CXB high contents of EL 100-55 were detected (A). Instead, higher intensity of CXB was associated with lower intensity of the characteristic peak of EL 100-55 (1730 cm^-1) and the double peak between 2850 and 2950 cm^-1 appeared, indicating higher presence of HPC-SSL (C). Obviously, CXB was more distributed and localized in HPC-SSL than in EL 100-55. To find a potential explanation for this observation, the maximum kinetic solid-state solubility of CXB in EL 100-55 and in HPC-SSL alone was investigated via DSC (Figure 3). In case of CXB: EL 100-55, first residual crystallinity was detected using a drug load of 30%, while 50% CXB was embedded amorphously into HPC-SSL. The higher solid-state sink conditions in HPC-SSL than in EL 100-55 was likely the reason for the preferred distribution of CXB in HPC-SSL.

Conclusion: The aim of this study was to provide mechanistic understanding of drug distribution in ternary ASDs. By using confocal Raman spectroscopy this study visualized inhomogeneities of the ternary CXB: EL 100‑55: HPC-SSL ASD. Interestingly, not only the polymers EL 100-55 and HPC-SSL demonstrated phase separation phenomena, but also CXB was inhomogeneously distributed in the polymer matrix. Due to higher sink condition of CXB in HPC-SSL than in EL 100-55 during the molding process of the ternary ASD, the CXB concentration was higher in HPC-SSL than in EL 100-55.

References: [1] Prasad, D. et al. Amorphous Stabilization and Dissolution Enhancement of Amorphous Ternary Solid Dispersions: Combination of Polymers Showing Drug–Polymer Interaction for Synergistic Effects. Journal of Pharmaceutical Sciences 103, 3511–3523 (2014).
[2] Monschke, M. & Wagner, K. G. Impact of HPMCAS on the Dissolution Performance of Polyvinyl Alcohol Celecoxib Amorphous Solid Dispersions. Pharmaceutics 12, 541 (2020).

Acknowledgements:The authors would like to thank Nisso Chemical Europe GmbH for supporting and funding this research project.

Figure 1: Single Raman spectra of the neat substances.

Figure 2: Images of confocal Raman spectroscopy of the ternary CXB: EL 100-55: HPC-SSL ASD. The green color represents the intensity at 1615 1/cm for detecting CXB (upper image) and the red color represents the intensity at 1730 1/cm for detecting EL 100-55 (lower image).

Figure 3: DSC thermograms of binary highly drug loaded CXB: EL 100-55 ASDs (A) and CXB: HPC-SSL ASDs (B).