Ensemble Machine Learning Approaches Based on Molecular Descriptors and Graph Convolutional Networks for Predicting the Efflux Activities of MDR1 and BCRP Transporters

Overview

Journal AAPS J

Specialty Pharmacology

Date 2023 Sep 12

PMID 37700207

Authors

Asahi Adachi

Tomoki Yamashita

Shigehiko Kanaya

Yohei Kosugi

Affiliations

Soon will be listed here.

Abstract

Multidrug resistance (MDR1) and breast cancer resistance protein (BCRP) play important roles in drug absorption and distribution. Computational prediction of substrates for both transporters can help reduce time in drug discovery. This study aimed to predict the efflux activity of MDR1 and BCRP using multiple machine learning approaches with molecular descriptors and graph convolutional networks (GCNs). In vitro efflux activity was determined using MDR1- and BCRP-expressing cells. Predictive performance was assessed using an in-house dataset with a chronological split and an external dataset. CatBoost and support vector regression showed the best predictive performance for MDR1 and BCRP efflux activities, respectively, of the 25 descriptor-based machine learning methods based on the coefficient of determination (R). The single-task GCN showed a slightly lower performance than descriptor-based prediction in the in-house dataset. In both approaches, the percentage of compounds predicted within twofold of the observed values in the external dataset was lower than that in the in-house dataset. Multi-task GCN did not show any improvements, whereas multimodal GCN increased the predictive performance of BCRP efflux activity compared with single-task GCN. Furthermore, the ensemble approach of descriptor-based machine learning and GCN achieved the highest predictive performance with R values of 0.706 and 0.587 in MDR1 and BCRP, respectively, in time-split test sets. This result suggests that two different approaches to represent molecular structures complement each other in terms of molecular characteristics. Our study demonstrated that predictive models using advanced machine learning approaches are beneficial for identifying potential substrate liability of both MDR1 and BCRP.

References

Yamamoto S, Kosugi Y, Hirabayashi H, Moriwaki T . Impact of P-Glycoprotein on Intestinal Absorption of an Inhibitor of Apoptosis Protein Antagonist in Rats: Mechanisms of Nonlinear Pharmacokinetics and Food Effects. Pharm Res. 2018; 35(10):190. DOI: 10.1007/s11095-018-2470-x. View

Breedveld P, Pluim D, Cipriani G, Wielinga P, van Tellingen O, Schinkel A . The effect of Bcrp1 (Abcg2) on the in vivo pharmacokinetics and brain penetration of imatinib mesylate (Gleevec): implications for the use of breast cancer resistance protein and P-glycoprotein inhibitors to enable the brain penetration of imatinib.... Cancer Res. 2005; 65(7):2577-82. DOI: 10.1158/0008-5472.CAN-04-2416. View

Sakata A, Tamai I, Kawazu K, Deguchi Y, Ohnishi T, Saheki A . In vivo evidence for ATP-dependent and P-glycoprotein-mediated transport of cyclosporin A at the blood-brain barrier. Biochem Pharmacol. 1994; 48(10):1989-92. DOI: 10.1016/0006-2952(94)90601-7. View

Schinkel A, Smit J, van Tellingen O, Beijnen J, Wagenaar E, van Deemter L . Disruption of the mouse mdr1a P-glycoprotein gene leads to a deficiency in the blood-brain barrier and to increased sensitivity to drugs. Cell. 1994; 77(4):491-502. DOI: 10.1016/0092-8674(94)90212-7. View

Kodaira H, Kusuhara H, Ushiki J, Fuse E, Sugiyama Y . Kinetic analysis of the cooperation of P-glycoprotein (P-gp/Abcb1) and breast cancer resistance protein (Bcrp/Abcg2) in limiting the brain and testis penetration of erlotinib, flavopiridol, and mitoxantrone. J Pharmacol Exp Ther. 2010; 333(3):788-96. DOI: 10.1124/jpet.109.162321. View

Enokizono J, Kusuhara H, Ose A, Schinkel A, Sugiyama Y . Quantitative investigation of the role of breast cancer resistance protein (Bcrp/Abcg2) in limiting brain and testis penetration of xenobiotic compounds. Drug Metab Dispos. 2008; 36(6):995-1002. DOI: 10.1124/dmd.107.019257. View

Hancock J, Khoshgoftaar T . CatBoost for big data: an interdisciplinary review. J Big Data. 2020; 7(1):94. PMC: 7610170. DOI: 10.1186/s40537-020-00369-8. View

Kosugi Y, Hosea N . Direct Comparison of Total Clearance Prediction: Computational Machine Learning Model versus Bottom-Up Approach Using In Vitro Assay. Mol Pharm. 2020; 17(7):2299-2309. DOI: 10.1021/acs.molpharmaceut.9b01294. View

Kosugi Y, Hosea N . Prediction of Oral Pharmacokinetics Using a Combination of In Silico Descriptors and In Vitro ADME Properties. Mol Pharm. 2021; 18(3):1071-1079. DOI: 10.1021/acs.molpharmaceut.0c01009. View

10.

Korotcov A, Tkachenko V, Russo D, Ekins S . Comparison of Deep Learning With Multiple Machine Learning Methods and Metrics Using Diverse Drug Discovery Data Sets. Mol Pharm. 2017; 14(12):4462-4475. PMC: 5741413. DOI: 10.1021/acs.molpharmaceut.7b00578. View

11.

Wang Y, Liu H, Fan Y, Chen X, Yang Y, Zhu L . In Silico Prediction of Human Intravenous Pharmacokinetic Parameters with Improved Accuracy. J Chem Inf Model. 2019; 59(9):3968-3980. DOI: 10.1021/acs.jcim.9b00300. View

12.

Kosugi Y, Mizuno K, Santos C, Sato S, Hosea N, Zientek M . Direct Comparison of the Prediction of the Unbound Brain-to-Plasma Partitioning Utilizing Machine Learning Approach and Mechanistic Neuropharmacokinetic Model. AAPS J. 2021; 23(4):72. PMC: 8131289. DOI: 10.1208/s12248-021-00604-x. View

13.

Kamiya Y, Handa K, Miura T, Yanagi M, Shigeta K, Hina S . Prediction of Input Parameters for Simplified Physiologically Based Pharmacokinetic Models for Estimating Plasma, Liver, and Kidney Exposures in Rats after Oral Doses of 246 Disparate Chemicals. Chem Res Toxicol. 2021; 34(2):507-513. DOI: 10.1021/acs.chemrestox.0c00336. View

14.

Ekins S, Waller C, Swaan P, Cruciani G, Wrighton S, Wikel J . Progress in predicting human ADME parameters in silico. J Pharmacol Toxicol Methods. 2001; 44(1):251-72. DOI: 10.1016/s1056-8719(00)00109-x. View

15.

Eric S, Kalinic M, Ilic K, Zloh M . Computational classification models for predicting the interaction of drugs with P-glycoprotein and breast cancer resistance protein. SAR QSAR Environ Res. 2014; 25(12):939-66. DOI: 10.1080/1062936X.2014.976265. View

16.

Hazai E, Hazai I, Ragueneau-Majlessi I, Chung S, Bikadi Z, Mao Q . Predicting substrates of the human breast cancer resistance protein using a support vector machine method. BMC Bioinformatics. 2013; 14:130. PMC: 3641962. DOI: 10.1186/1471-2105-14-130. View

17.

Zhong L, Ma C, Zhang H, Yang L, Wan H, Xie Q . A prediction model of substrates and non-substrates of breast cancer resistance protein (BCRP) developed by GA-CG-SVM method. Comput Biol Med. 2011; 41(11):1006-13. DOI: 10.1016/j.compbiomed.2011.08.009. View

18.

Gantner M, Peroni R, Morales J, Villalba M, Ruiz M, Talevi A . Development and Validation of a Computational Model Ensemble for the Early Detection of BCRP/ABCG2 Substrates during the Drug Design Stage. J Chem Inf Model. 2017; 57(8):1868-1880. DOI: 10.1021/acs.jcim.7b00016. View

19.

Sedykh A, Fourches D, Duan J, Hucke O, Garneau M, Zhu H . Human intestinal transporter database: QSAR modeling and virtual profiling of drug uptake, efflux and interactions. Pharm Res. 2012; 30(4):996-1007. PMC: 3596480. DOI: 10.1007/s11095-012-0935-x. View

20.

Garg P, Dhakne R, Belekar V . Role of breast cancer resistance protein (BCRP) as active efflux transporter on blood-brain barrier (BBB) permeability. Mol Divers. 2014; 19(1):163-72. DOI: 10.1007/s11030-014-9562-2. View