Research Article: Prediction of drug–ABC-transporter interaction — Recent advances and future challenges☆

Date Published: September 17, 2018

Publisher:

Author(s): Floriane Montanari, Gerhard F. Ecker.

http://doi.org/10.1016/j.addr.2015.03.001

Abstract

With the discovery of P-glycoprotein (P-gp), it became evident that ABC-transporters play a vital role in bioavailability and toxicity of drugs. They prevent intracellular accumulation of toxic compounds, which renders them a major defense mechanism against xenotoxic compounds. Their expression in cells of all major barriers (intestine, blood–brain barrier, blood–placenta barrier) as well as in metabolic organs (liver, kidney) also explains their influence on the ADMET properties of drugs and drug candidates. Thus, in silico models for the prediction of the probability of a compound to interact with P-gp or analogous transporters are of high value in the early phase of the drug discovery process. Within this review, we highlight recent developments in the area, with a special focus on the molecular basis of drug–transporter interaction. In addition, with the recent availability of X-ray structures of several ABC-transporters, also structure-based design methods have been applied and will be addressed.

Partial Text

ATP-binding cassette transporters (ABC-transporters) form a large superfamily of membrane proteins. Members of the ABC-transporters can be found in all living organisms from prokaryotes to mammals. Generally speaking, these transporters participate in active transport, i.e. they hydrolyze ATP and use its energy to transport their substrates. In humans, 49 ABC-transporters are recognized to date and belong to 7 distinct subfamilies [1], ABCA to ABCG. The usual “transport unit” consists of two intracellular nucleotide binding domains and two transmembrane domains. The nucleotide binding domains (NBDs), usually well conserved across subfamilies, bind and hydrolyze ATP. The transmembrane domains create the translocation chamber across which the substrates diffuse. These regions are usually little conserved and are responsible for the substrate specificity of the different transporters. Members of the ABCBA subfamily transport cholesterol and lipids [2]. Members of the B, C and G subfamilies are multi-drug resistance-associated transporters or associated with diseases.

Working with large datasets seems to be the way to build high quality models and derive general trends for compounds interacting with ABC-transporters. Data, however, is scarce, at least for some less studied transporters. Typical medicinal chemistry studies report bioactivities for a small set of chemically related compounds. Scientists wanting to build large datasets must collect and merge together such data, using databases like ChEMBL [56] and Pubchem [57], but also manual search through MEDLINE. Now, what if the groups measuring ABC-transporters substrate or inhibition activities each use their own assay design? Then merging together data becomes a challenging task. Zdrazil et al. [58], studied all bioassays from ChEMBL for P-gp inhibition and transport, when these assays reported IC50, EC50 or Ki values. Subsequently, they annotated assays according to their potential for being combined together in a large QSAR dataset. The results show the importance of overlapping binding sites for the different substrates used in the bioassays, as well as the cell line in which the transporter is expressed.

Due to the tremendous progress in the field of structural biology, structures of transmembrane transporters, including several ABC-transporters, became available. Most of them were from prokaryotes, and only very recently structures from eukaryotic organisms were also resolved in a resolution which allows starting structure-based approaches. However, the only human ABC-transporter crystallized so far is ABCB10 [62]. Nevertheless, the whole field of ABC-transporter research was inspired by the first structures being deposited in the Protein Data Bank [63], and protein homology models of P-gp immediately became available.

Challenges in the field of ABC-transporter are manifold. With respect to the prediction of drug transporter interaction, there are on our point of view several immediate issues which should be mentioned. The most obvious one is the availability of an atomic resolution structure of human P-glycoprotein in complex with a prototype ligand such as verapamil. This would allow benchmarking all docking studies on this structure, which definitely would increase the validity of the binding hypotheses retrieved. Nevertheless, in early drug discovery, in silico models based on machine learning will still be the main tools for prioritization of large compound libraries. In silico classification models will only show high predictivity if the underlying data are of high quality and of a considerable size. In case of P-glycoprotein, the size of the datasets available in the public domain is sufficient, but the use of almost 50 different assays currently does not allow combining all the data and to compile a large, high quality dataset for training the models. For ABC-transporters other than P-glycoprotein, the situation is even worse, as already the available datasets are small. For solving the issue of different assays, a transporter assay ontology combined with the definition of a set of standard reference compounds would be highly recommended.

ABC-transporters represent an integral and important part of the human transportome. Although there are only 49 genes described in humans, they fulfill important roles and are strongly linked to drug absorption, distribution, and elimination. Furthermore, besides cytochromes, they are also involved in drug–drug interactions and thus also toxicity of drugs. With the increasing accessibility of biological data and the tremendous progress of structural biology, our understanding of the molecular basis of ligand–transporter interaction is progressing. In this review, we have outlined recent ligand-based models built on large datasets rather than on congeneric series. While these models allow screening rapidly large databases of molecules to predict their substrate or inhibition properties, their interpretation remains at the level of substructures or general physico-chemical properties. On the structure-based side, the presence of crystal structures of the B subfamily allowed building high quality homology models for P-gp and a mapping of different binding sites has started. Such advances have not yet been noted for other ABC-transporters, but we believe that new structures will appear in the PDB that will allow similar studies to be performed.

 

Source:

http://doi.org/10.1016/j.addr.2015.03.001

 

0 0 vote
Article Rating
Subscribe
Notify of
guest
0 Comments
Inline Feedbacks
View all comments