Machine learning with active pharmaceutical ingredient/polymer interaction mechanism: Prediction for complex phase behaviors of pharmaceuticals and formulations

doi:10.1016/j.cjche.2023.09.006

Abstract

Abstract: The high throughput prediction of the thermodynamic phase behavior of active pharmaceutical ingredients (APIs) with pharmaceutically relevant excipients remains a major scientific challenge in the screening of pharmaceutical formulations. In this work, a developed machine-learning model efficiently predicts the solubility of APIs in polymers by learning the phase equilibrium principle and using a few molecular descriptors. Under the few-shot learning framework, thermodynamic theory (perturbed-chain statistical associating fluid theory) was used for data augmentation, and computational chemistry was applied for molecular descriptors’ screening. The results showed that the developed machine-learning model can predict the API-polymer phase diagram accurately, broaden the solubility data of APIs in polymers, and reproduce the relationship between API solubility and the interaction mechanisms between API and polymer successfully, which provided efficient guidance for the development of pharmaceutical formulations.

Key words: Multi-task machine learning, Density functional theory, Hydrogen bond interaction, Miscibility, Solubility

摘要： The high throughput prediction of the thermodynamic phase behavior of active pharmaceutical ingredients (APIs) with pharmaceutically relevant excipients remains a major scientific challenge in the screening of pharmaceutical formulations. In this work, a developed machine-learning model efficiently predicts the solubility of APIs in polymers by learning the phase equilibrium principle and using a few molecular descriptors. Under the few-shot learning framework, thermodynamic theory (perturbed-chain statistical associating fluid theory) was used for data augmentation, and computational chemistry was applied for molecular descriptors’ screening. The results showed that the developed machine-learning model can predict the API-polymer phase diagram accurately, broaden the solubility data of APIs in polymers, and reproduce the relationship between API solubility and the interaction mechanisms between API and polymer successfully, which provided efficient guidance for the development of pharmaceutical formulations.

关键词: Multi-task machine learning, Density functional theory, Hydrogen bond interaction, Miscibility, Solubility

Kai Ge, Yiping Huang, Yuanhui Ji. Machine learning with active pharmaceutical ingredient/polymer interaction mechanism: Prediction for complex phase behaviors of pharmaceuticals and formulations[J]. Chinese Journal of Chemical Engineering, 2024, 66(2): 263-272.

References

[1] J. Brinkmann, L. Exner, C. Luebbert, G. Sadowski, In-silico screening of lipid-based drug delivery systems, Pharm. Res. 37(12)(2020)249.
[2] A. Prudic, A.K. Lesniak, Y.H. Ji, G. Sadowski, Thermodynamic phase behaviour of indomethacin/PLGA formulations, Eur. J. Pharm. Biopharm. 93(2015)88-94.
[3] C. Luebbert, E. Stoyanov, G. Sadowski, Phase behavior of ASDs based on hydroxypropyl cellulose, Int. J. Pharm. X 3(2021)100070.
[4] V.F. Leimann, O.H. Goncalves, G.D. Sorita, S. Rezende, E. Bona, I.P.M. Fernandes, I.C.F.R. Ferreira, M.F. Barreiro, Heat and pH stable curcumin-based hydrophilic colorants obtained by the solid dispersion technology assisted by spray-drying, Chem. Eng. Sci. 205(2019)248-258.
[5] M.M. Knopp, L. Tajber, Y.W. Tian, N.E. Olesen, D.S. Jones, A. Kozyra, K. Lobmann, K. Paluch, C.M. Brennan, R. Holm, A.M. Healy, G.P. Andrews, T. Rades, Comparative study of different methods for the prediction of drug-polymer solubility, Mol. Pharm. 12(9)(2015)3408-3419.
[6] A. Prudic, Y.H. Ji, C. Luebbert, G. Sadowski, Influence of humidity on the phase behavior of API/polymer formulations, Eur. J. Pharm. Biopharm. 94(2015)352-362.
[7] J.J.F. Chen, D.P. Visco Jr, Developing an in silico pipeline for faster drug candidate discovery:virtual high throughput screening with the Signature molecular descriptor using support vector machine models, Chem. Eng. Sci. 159(2017)31-42.
[8] S. Ekins, A.C. Puhl, K.M. Zorn, T.R. Lane, D.P. Russo, J.J. Klein, A.J. Hickey, A.M. Clark, Exploiting machine learning for end-to-end drug discovery and development, Nat. Mater. 18(5)(2019)435-441.
[9] X.H. Che, S.Y. Chai, Z.Z. Zhang, L. Zhang, Prediction of ligand binding sites using improved blind docking method with a Machine Learning-Based scoring function, Chem. Eng. Sci. 261(2022)117962.
[10] J. Vamathevan, D. Clark, P. Czodrowski, I. Dunham, E. Ferran, G. Lee, B. Li, A. Madabhushi, P. Shah, M. Spitzer, S.R. Zhao, Applications of machine learning in drug discovery and development, Nat. Rev. Drug Discov. 18(6)(2019)463-477.
[11] S.W. Yue, M.C. Muniz, M.F. Calegari Andrade, L.F. Zhang, R. Car, A.Z. Panagiotopoulos, When do short-range atomistic machine-learning models fall short?J. Chem. Phys. 154(3)(2021)034111.
[12] E.D. Miller, M.L. Jones, M.M. Henry, B. Stanfill, E. Jankowski, Machine learning predictions of electronic couplings for charge transport calculations of P3HT, AlChE. J. 65(12)(2019) e16760.
[13] Y.Q. Wang, Q.M. Yao, J.T. Kwok, L.M. Ni, Generalizing from a few examples, ACM Comput. Surv. 53(3)(2021)1-34.
[14] J. Peng, W. Sun, G.M. Zhou, L. Xie, H.S. Han, Y. Xiao, The accurate prediction and analysis of bed expansion characteristics in liquid-solid fluidized bed based on machine learning methods, Chem. Eng. Sci. 260(2022)117841.
[15] G.E. Karniadakis, I.G. Kevrekidis, L. Lu, P. Perdikaris, S.F. Wang, L. Yang, Physics-informed machine learning, Nat. Rev. Phys. 3(6)(2021)422-440.
[16] A. Prudic, Y.H. Ji, G. Sadowski, Thermodynamic phase behavior of API/polymer solid dispersions, Mol. Pharm. 11(7)(2014)2294-2304.
[17] J. M. Prausnitz, R. N. Lichtenthaler, E. G. De Azevedo, Molecular thermodynamics of fluid-phase equilibria. Pearson Education:1998.
[18] J. Gross, G. Sadowski, Perturbed-chain SAFT:an equation of state based on a perturbation theory for chain molecules, Ind. Eng. Chem. Res. 40(4)(2001)1244-1260.
[19] J. Gross, G. Sadowski, Application of the perturbed-chain SAFT equation of state to associating systems, Ind. Eng. Chem. Res. 41(22)(2002)5510-5515.
[20] J.P. Wolbach, S.I. Sandler, Using molecular orbital calculations to describe the phase behavior of cross-associating mixtures, Ind. Eng. Chem. Res. 37(8)(1998)2917-2928.
[21] B. Wang, D.D. Wang, S. Zhao, X.B. Huang, J.B. Zhang, Y. Lv, X.C. Liu, G.J. Lv, X.J. Ma, Evaluate the ability of PVP to inhibit crystallization of amorphous solid dispersions by density functional theory and experimental verify, Eur. J. Pharm. Sci. 96(2017)45-52.
[22] C. Gertig, L. Fleitmann, C. Hemprich, J. Hense, A. Bardow, K. Leonhard, CAT-COSMO-CAMPD:integrated in silico design of catalysts and processes based on quantum chemistry, Comput. Chem. Eng. 153(2021)107438.
[23] S.F. Boys, F. Bernardi, The calculation of small molecular interactions by the differences of separate total energies. Some procedures with reduced errors, Mol. Phys. 19(4)(1970)553-566.
[24] T. Lu, F.W. Chen, Multiwfn:a multifunctional wavefunction analyzer, J. Comput. Chem. 33(5)(2012)580-592.
[25] W. Humphrey, A. Dalke, K. Schulten, VMD:visual molecular dynamics, J. Mol. Graph. 14(1)(1996)33-38.
[26] K. Ge, Y.H. Ji, A thermodynamic approach for predicting thermodynamic phase behaviors of pharmaceuticals in biorelevant media, Chem. Eng. Sci. 261(2022)117973.
[27] K. Ge, R. Paus, V. Penner, G. Sadowski, Y.H. Ji, Theoretical modeling and prediction of biorelevant solubility of poorly soluble pharmaceuticals, Chem. Eng. J. 444(2022)136678.
[28] K. Lehmkemper, S.O. Kyeremateng, O. Heinzerling, M. Degenhardt, G. Sadowski, Impact of polymer type and relative humidity on the long-term physical stability of amorphous solid dispersions, Mol. Pharm. 14(12)(2017)4374-4386.
[29] Y.W. Tian, J. Booth, E. Meehan, D.S. Jones, S. Li, G.P. Andrews, Construction of drug-polymer thermodynamic phase diagrams using Flory-Huggins interaction theory:identifying the relevance of temperature and drug weight fraction to phase separation within solid dispersions, Mol. Pharm. 10(1)(2013)236-248.
[30] Y. Sun, J. Tao, G.G.Z. Zhang, L. Yu, Solubilities of crystalline drugs in polymers:an improved analytical method and comparison of solubilities of indomethacin and nifedipine in PVP, PVP/VA, and PVAc, J. Pharm. Sci. 99(9)(2010)4023-4031.
[31] M. Maniruzzaman, J.Y. Pang, D.J. Morgan, D. Douroumis, Molecular modeling as a predictive tool for the development of solid dispersions, Mol. Pharm. 12(4)(2015)1040-1049.
[32] T.X. Xiang, B.D. Anderson, Molecular dynamics simulation of amorphous indomethacin-poly (vinylpyrrolidone) glasses:solubility and hydrogen bonding interactions, J. Pharm. Sci. 102(3)(2013)876-891.
[33] T.X. Xiang, B.D. Anderson, Effects of molecular interactions on miscibility and mobility of ibuprofen in amorphous solid dispersions with various polymers, J. Pharm. Sci. 108(1)(2019)178-186.
[34] Y. Li, X.M. Liu, Y.Q. Zhang, K. Jiang, J.J. Wang, S.J. Zhang, Why only ionic liquids with unsaturated heterocyclic cations can dissolve cellulose:a simulation study, ACS Sustainable Chem. Eng. 5(4)(2017)3417-3428.
[35] R. Han, H. Xiong, Z. Ye, Y.L. Yang, T.H. Huang, Q.F. Jing, J.H. Lu, H. Pan, F.Z. Ren, D.F. Ouyang, Predicting physical stability of solid dispersions by machine learning techniques, J. Control. Release 311-312(2019)16-25.
[36] S. Goyal, A. Chattopadhyay, K. Kasavajhala, U.D. Priyakumar, Role of urea-aromatic stacking interactions in stabilizing the aromatic residues of the protein in urea-induced denatured state, J. Am. Chem. Soc. 139(42)(2017)14931-14946.
[37] E. Espinosa, E. Molins, C. Lecomte, Hydrogen bond strengths revealed by topological analyses of experimentally observed electron densities, Chem. Phys. Lett. 285(3-4)(1998)170-173.
[38] K. Lehmkemper, S.O. Kyeremateng, O. Heinzerling, M. Degenhardt, G. Sadowski, Long-term physical stability of PVP-and PVPVA-amorphous solid dispersions, Mol. Pharm. 14(1)(2017)157-171.
[39] W.L. Dai, S. Cahyawijaya, Y.J. Bang, P. Fung, Weakly-supervised multi-task learning for multimodal affect recognition, arXiv (2021)2104.11560.
[40] S.D. Thakore, J. Akhtar, R. Jain, A. Paudel, A.K. Bansal, Analytical and computational methods for the determination of drug-polymer solubility and miscibility, Mol. Pharm. 18(8)(2021)2835-2866.
[41] W.Z. Zheng, Z.H. Ma, W.Z. Sun, L. Zhao, Target high-efficiency ionic liquids to promote H₂SO₄-catalyzed C4 alkylation by machine learning, AlChE. J. 68(7)(2022) e17698.
[42] H.Q. Wen, Y. Su, Z.H. Wang, S.M. Jin, J.Z. Ren, W.F. Shen, M. Eden, A systematic modeling methodology of deep neural network-based structure-property relationship for rapid and reliable prediction on flashpoints, AlChE. J. 68(1)(2022) e17402.
[43] K. Nurzynska, J. Booth, C.J. Roberts, J. McCabe, I. Dryden, P.M. Fischer, Long-term amorphous drug stability predictions using easily calculated, predicted, and measured parameters, Mol. Pharm. 12(9)(2015)3389-3398.
[44] M. Przybylek, T. Jelinski, P. Cysewski, Application of multivariate adaptive regression splines (MARSplines) for predicting Hansen solubility parameters based on 1D and 2D molecular descriptors computed from SMILES string, J. Chem. 2019(2019)1-15.
[45] M.D. Moore, P.L.D. Wildfong, Informatics calibration of a molecular descriptors database to predict solid dispersion potential of small molecule organic solids, Int. J. Pharm. 418(2)(2011)217-226.