(W1130-01-03) Study of Physiological Buffering Effects on Ezetimibe Supersaturation Using Hydroxypropyl Cellulose as Precipitation Inhibitor

Purpose: Phosphate buffers are predominantly used for in vitro testing even though bicarbonate (or carbonate) buffers are known to be more physiological [1]. Recently, the effect of bicarbonate buffering was evaluated in a gastrointestinal transfer model to study drug supersaturation in presence of polymer whereby a pronounced buffer effect was evidenced for different drug bases [2]. A follow-up study did not employ such a polymer and interestingly, the buffer effect was comparatively subtle and compound-specific [3]. It seems that for sustaining drug supersaturation, the interface between polymer and water with ions plays a critical role. This study aimed at investigating the precipitation inhibiting effects of hydroxypropyl cellulose (HPC) using ezetimibe as model compound in presence of phosphate vs. bicarbonate buffer. The present research further compares the experimental results of the two buffer systems in an atomistic computational study.

Methods: A solvent shift assay with in-line UV analytics (model 440 UV−vis spectrophotometer from SI Photonics Inc.,Tucson, USA) was used in combination with a KM10-2 FLEX CO₂/N₂ gas mixer (WITT-GASETECHNIK GmbH & Co KG, Witten, Germany). The aqueous buffer solution contained 0.01% (w/w) HPC at pH 6.5 in line with other work from the literature [4]. For the in silico study, HPC was first subjected to buffer concentrations according to the literature [2] to then conduct the molecular docking trials. A GLYCAM and general atom force field (GAFF2) of the AMBER12 force field family were employed for HPC (n=20) and drug, respectively, while water was present as TIP3P model. Molecular charges were based on quantum-chemical estimations (AM1BCC). The cut-off for long-range molecular interactions was 8 Å; periodic boundary conditions (NTP ensemble) were selected at 1 bar and 310 K (2x 1fs steps for 5 ns) using YASARA v. 20.12.24 (YASARA Biosciences, GmbH, Vienna, Austria) and this software platform was also used for subsequent molecular docking with the integrated VINA algorithm [5].

Results: A comparison of the kinetic concentration profiles (Fig. 1) showed in both buffer systems effects of drug precipitation as compared to pure drug. Ezetimibe alone started to crush out practically from the start of the solvent-shift experiments, whereas a delay of around an hour (63.7 ± 7.0 min induction time) was achieved in phosphate buffer. This precipitation inhibition of HPC (0.01%, w/w) was less pronounced in the bicarbonate buffer (25.3 ± 3.1 min induction time) but there was still a clear difference evidenced as compared to the kinetics of pure drug. Molecular docking provided mean interaction values of HPC and ezetimibe with -17.1 kJ/mol for the case of bicarbonate buffer, whereas -18.9 kJ/mol resulted for the corresponding drug-excipient interactions in phosphate buffer. Although there was some variation of individual docking configurations given and the buffer effect on molecular docking was rather subtle, there was still a statistically significant effect given (p < 0.0001). This computational result is remarkable as it suggests Hofmeister ion effects to influence the way of how the polymer interacts with the drug. Even though this is a computational result for the bulk phase, there may exist also ion-induced polymer differences at the interface of nucleating or growing drug crystals, The slightly lower molecular interaction energy in physiological buffer likely translated into a less pronounced drug precipitation inhibition effects in vitro. Previous work explained buffer differences on a mechanistic level primarily based on differences in buffer capacity or surface-pH differences, but the latter mechanisms were at least for the present model system not expected to contribute greatly to the observed findings.

Conclusion: Effects of physiological buffering is a topic of much contemporary interest. The present study is pioneering in using computational pharmaceutics to achieve a better mechanistic understanding of buffer effects in the field of drug supersaturation testing. In silico results from molecular docking with polymer were consistent with the experimentally observed differences in precipitation kinetics. Thus, the widely used phosphate buffer may lead to more pronounced drug interactions with polymer and longer induction times than in physiological buffer. This work further supports the view that HPC is indeed promising to use for drug precipitation inhibition. However, more data would be needed to draw any final conclusions.

References: 1. Sheng, J.J., et al., Towards and In Vivo Dissolution Methodology: A Comparison of Phosphate and Bicar-bonate Buffers. Molecular Pharmaceutics, 2008, 6(1), 29-39.
2. Jede, Ch. et al., Improved Prediction of in Vivo Supersaturation and Precipitation of Poorly Soluble Weakly Basic Drugs Using Biorelevant Bicarbonate Buffer in a Gastrointestinal Transfer Model. Molecular Phar-maceutics, 2019, 16, 3938-3947.
3. Krollik et al., The effect of buffer species on biorelevant dissolution and precipitation assays – Compari-son of phosphate and bicarbonate buffer. European Journal of Pharmaceutics and Biopharmaceutics, 2022, 171, 90-101.
4. Niederquell, A., et al., Hydroxypropyl Cellulose for Drug Precipitation Inhibition: From the Potential of Mo-lecular Interactions to Performance Considering Microrheology. Molecular Pharmaceutics, 2022, 19, 690-703.
5. Krieger, E. and Vriend, G. YASARA View-molecular graphics for all devices-from smartphones to work-stations. Bioinformatics, 2014, 30, 2981−2982.

Figure 1: Solvent-shift precipitation experiment (in-line UV analysis) in phosphate (A) vs. physiological buffer (B) (both at pH 6.5). Circles represent pure drug while addition of 0.01% (w/w) added HPC (L-grade) is given as trian-gles.



Figure 2: Molecular docking pose of ezetimibe on an oligomer of HPC (n=20) that was previously energy mini-mized by a molecular dynamics simulation assuming physiological buffering. Details are given in the text.