Purchase this article with an account.
Tomoyo Miyabe, Shinnosuke Machida, Kazuya Konoura, Takeshi Nakajima, Mitsuyoshi Azuma; Structure-based in silico approach for selection of kinase inhibitors. Invest. Ophthalmol. Vis. Sci. 2016;57(12):5019. doi: https://doi.org/.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Protein kinases play an important role in several cell processes such as proliferation, transcription, and pathologic changes. Kinase inhibitors have thus been proposed for treatment against diseases including cancer. In ophthalmology, ROCK inhibitor has already been launched as an antiglaucoma drug in Japan. However, discovery of small molecule inhibitors for specific protein kinases is still challenging. This is because approximately 500 protein kinases exist and more than 2,000 other purine-binding proteins share similar ATP binding pockets of kinases. The purpose of this study was to develop an in silico method for identifying potential kinase inhibitors, and the anticancer drug axitinib was used as a representative kinase inhibitor.
Sequences for 9 typical kinases were compared to VEGFR tyrosine kinase, the three amino acid sequences on the hinge region of each kinase were identified, and the surfaces of the ATP binding cavities in the kinases were analyzed in silico (Molecular Operating Environment, MOE). Inhibitory activities of axitinib against the identified kinases were evaluated by a mobility shift assay. To confirm the results of our MOE identification process, axitinib was optimized by functional group conversion, and inhibitory activity of the modified axitinib against the kinases was assessed again.
Axitinib strongly inhibited VEGFR tyrosine kinase. Axitinib did not inhibit other kinases with less than 60% sequence identity to the ATP binding pocket in VEGFR kinase, and moderately inhibited the kinases with more than 70% sequence identity. Axitinib did not inhibit kinases where 2 of the 3 conserved amino acids in the hinge region were replaced. Aurora-A serine/threonine kinase was an exception, and was inhibited by axitinib in spite of the low sequence identity and 2 amino acid replacements. Compared to Aurora-A kinase, VEGFR kinase showed a narrow entrance to deeper, back-pocket ATP binding cavity. Substituting 4,5-dihydro-2-thiazolyl for the methyl residue in axitinib prevented the modified inhibitor from settling into the pocket of Aurora-A. Unfortunately, this substitution approach was not sufficient for the selectivity of VEGFR kinase against kinases with more than 70% sequence identity.
The in silico approach using three-dimensional structure was useful for designing selective kinase inhibitors. Additional approaches are needed for developing higher specificities.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.
This PDF is available to Subscribers Only