Investigation of the Experimental and Theoretical Release of Caffeine from Cosmetic Hydrogels and Patches
356
Presented by: Yimeng Jiao
Introduction
There is an increasing demand for cosmetic products that provide visible skin improvements. During their development, the release of cosmetic actives (CAs) is monitored by applying an In Vitro Release Testing (IVRT) method. Despite IVRT being a wildly used experimental tool, it is time consuming and has poor repeatability. Molecular docking (MD) is a tool that performs computational simulation. In the systems such as solutions, gels, and creams, it can be used to predict types of chemical interactions by calculating binding energies of temporary created complexes between the molecules. This study explores experimental (IVRT) and theoretical (MD) release of caffeine from six different cosmetic hydrogels and patches. The aim of the study was to compare the experimental and theoretical data and determine whether MD can be used as a tool to predict the release of caffeine.
Methods
MD studies of caffeine and gelling agents (GAs) were performed by the Autodock Vina software. GAs were used as target molecules in the MD studies and their 3D structures were optimised by the Molecular Mechanics Universal Force Field or downloaded from the Protein Data Bank. The caffeine structure was used as a ligand in MD studies and optimised using Density Functional Theory (DFT) calculations. A Fused Deposition Modelling 3D printer (Ultimaker, UK) was used to fabricate patches made of polyvinyl alcohol (PVA). The patches were loaded with different hydrogels, containing caffeine as an active substance. The six hydrogels used were made from the following GAs: sodium polyacrylate (SPA), sodium carboxymethyl cellulose (SCMC), xanthan gum (XG), gellan gum (GG), carrageenan (CAR), and hydroxymethyl cellulose (HMC). IVRT of caffeine was performed using a system of 10 vertical diffusion cells (Copley, UK), the cellulose acetate membrane, and phosphate buffer of pH 7.40. The release of caffeine was assessed from the six hydrogels, as well as the PVA patches loaded with the hydrogels, at 30 min, 1 h, 2 h, 4 h, and 6 h.
Results
Binding energies (BEs) between GAs, caffeine, and membrane within their mutual orientations were obtained using the MD method. BEs of a single neutral caffeine molecule with GAs (caffeine-GAs) were found to be from −2.4 kcal/mol, for the caffeine-SPA complex, to −4.1 kcal/mol, for the caffeine-GG complex, while the BEs for GAs-membrane were from −4.7 kcal/mol, for the SPA-membrane complex, to −6.8 kcal/mol, for the GG-membrane complex. The BE for the caffeine-membrane system was −4.1 kcal/mol. The IVRT release profiles of caffeine from the six hydrogels have shown high similarity. After 30 minutes, all hydrogels were in the range of 60-70% release, except the CAR hydrogel which was about 10% lower. After 60 minutes, all six hydrogels have released 90% or more of caffeine. IVRT form PVA patches loaded with the hydrogels showed that after 30 minutes the release from all patches was 30-40%, followed by 50-60% after 60 minutes, and reaching 90-100% release after 6 hours.
Discussion
The BE values of a single neutral caffeine molecule with each of the GAs do not differ considerably, indicating similar effect on their release profiles. The MD calculations on the membrane structure have shown that all GAs possess higher affinity towards membrane than towards caffeine. Furthermore, the caffeine affinity towards membrane is higher compared to its affinity towards any GA. Hence, the MD results indicated that there would be no distinctive difference in caffeine kinetics among the hydrogels studied. The IVRT results showed that the release profiles from all hydrogels were very similar. The release profile from all PVA patches loaded with hydrogels followed the same trend, with PVA impairing the release. Both computational and experimental studies confirmed that the release profiles of caffeine from all hydrogels were similar.
Conclusion
The combined computational and experimental approach was applied, aiming to understand the profile of caffeine release from cosmetic hydrogels and patches. It was found that IVRT and MD results were in accordance. The kinetics of caffeine release does not only depend on its binding strength to the membrane, but it is also defined by the competition of caffeine and GA for the binding sites on the membrane. With further method development, MD can potentially be applied as a tool to predict the release of cosmetic actives from different formulations.
There is an increasing demand for cosmetic products that provide visible skin improvements. During their development, the release of cosmetic actives (CAs) is monitored by applying an In Vitro Release Testing (IVRT) method. Despite IVRT being a wildly used experimental tool, it is time consuming and has poor repeatability. Molecular docking (MD) is a tool that performs computational simulation. In the systems such as solutions, gels, and creams, it can be used to predict types of chemical interactions by calculating binding energies of temporary created complexes between the molecules. This study explores experimental (IVRT) and theoretical (MD) release of caffeine from six different cosmetic hydrogels and patches. The aim of the study was to compare the experimental and theoretical data and determine whether MD can be used as a tool to predict the release of caffeine.
Methods
MD studies of caffeine and gelling agents (GAs) were performed by the Autodock Vina software. GAs were used as target molecules in the MD studies and their 3D structures were optimised by the Molecular Mechanics Universal Force Field or downloaded from the Protein Data Bank. The caffeine structure was used as a ligand in MD studies and optimised using Density Functional Theory (DFT) calculations. A Fused Deposition Modelling 3D printer (Ultimaker, UK) was used to fabricate patches made of polyvinyl alcohol (PVA). The patches were loaded with different hydrogels, containing caffeine as an active substance. The six hydrogels used were made from the following GAs: sodium polyacrylate (SPA), sodium carboxymethyl cellulose (SCMC), xanthan gum (XG), gellan gum (GG), carrageenan (CAR), and hydroxymethyl cellulose (HMC). IVRT of caffeine was performed using a system of 10 vertical diffusion cells (Copley, UK), the cellulose acetate membrane, and phosphate buffer of pH 7.40. The release of caffeine was assessed from the six hydrogels, as well as the PVA patches loaded with the hydrogels, at 30 min, 1 h, 2 h, 4 h, and 6 h.
Results
Binding energies (BEs) between GAs, caffeine, and membrane within their mutual orientations were obtained using the MD method. BEs of a single neutral caffeine molecule with GAs (caffeine-GAs) were found to be from −2.4 kcal/mol, for the caffeine-SPA complex, to −4.1 kcal/mol, for the caffeine-GG complex, while the BEs for GAs-membrane were from −4.7 kcal/mol, for the SPA-membrane complex, to −6.8 kcal/mol, for the GG-membrane complex. The BE for the caffeine-membrane system was −4.1 kcal/mol. The IVRT release profiles of caffeine from the six hydrogels have shown high similarity. After 30 minutes, all hydrogels were in the range of 60-70% release, except the CAR hydrogel which was about 10% lower. After 60 minutes, all six hydrogels have released 90% or more of caffeine. IVRT form PVA patches loaded with the hydrogels showed that after 30 minutes the release from all patches was 30-40%, followed by 50-60% after 60 minutes, and reaching 90-100% release after 6 hours.
Discussion
The BE values of a single neutral caffeine molecule with each of the GAs do not differ considerably, indicating similar effect on their release profiles. The MD calculations on the membrane structure have shown that all GAs possess higher affinity towards membrane than towards caffeine. Furthermore, the caffeine affinity towards membrane is higher compared to its affinity towards any GA. Hence, the MD results indicated that there would be no distinctive difference in caffeine kinetics among the hydrogels studied. The IVRT results showed that the release profiles from all hydrogels were very similar. The release profile from all PVA patches loaded with hydrogels followed the same trend, with PVA impairing the release. Both computational and experimental studies confirmed that the release profiles of caffeine from all hydrogels were similar.
Conclusion
The combined computational and experimental approach was applied, aiming to understand the profile of caffeine release from cosmetic hydrogels and patches. It was found that IVRT and MD results were in accordance. The kinetics of caffeine release does not only depend on its binding strength to the membrane, but it is also defined by the competition of caffeine and GA for the binding sites on the membrane. With further method development, MD can potentially be applied as a tool to predict the release of cosmetic actives from different formulations.