| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 1.6 MB | Adobe PDF |
Authors
Abstract(s)
Os alvos terapêuticos utilizados atualmente para o tratamento de doenças constituem uma percentagem muito reduzida do proteoma humano. Os mesmos alvos são recorrentemente utilizados no processo de desenvolvimento de fármacos, nomeadamente na identificação de pequenas moléculas que possam modelar a sua atividade. Uma vez que apenas uma percentagem muito resumida do proteoma é utilizado no desenvolvimento de fármacos, uma das principais estratégias para a inovação na descoberta é a validação de “novos” alvos terapêuticos. Aqui, propomos uma nova estratégia para a validação de alvos terapêuticos que sabemos estarem envolvidos em doenças, mas para os quais se tenha mostrado difícil o desenvolvimento de moléculas com efeito farmacológico ou para os quais não exista informação da existência de moléculas que modelem a sua atividade. Este objetivo será alcançado com recurso a um conjunto de métodos computacionais baseados na similaridade estrutural de cavidades entre, os dois tipos de alvos previamente mencionados e proteínas bem estudadas para as quais foram já desenvolvidos potenciais modeladores, alguns já aprovados ou em ensaios clínicos. Assim, foram identificados e agrupados todos os alvos difíceis de modelar, resultando num total de 15 potenciais alvos terapêuticos agrupados como DOPE, e todos os alvos bem estudados foram agrupados como Well-Knowns. As cavidades de cada uma das proteínas foram determinadas utilizando o programa CAVIAR, seguido da uma procura de similaridade estrutural (PocketMatch). O centro ativo da proteína Large neutral Amino acid Transporter small subunit 1 (LAT1) apresentou similaridade com a butirilcolinesterase (BChE) e foi considerado o alvo mais promissor para o desenvolvimento da estratégia delineada. O último processo consistiu no screening virtual de um conjunto de moléculas com efeito biológico (inibidores já conhecidos) na butirilcolinesterase obtidas a partir da base de dados ChEMBL e na identificação e seleção dos compostos para os quais foi prevista uma afinidade significativa para com a proteína LAT1. De uma biblioteca de 837 compostos únicos, seis mostraram boa afinidade para com a LAT1, dos quais três foram capazes de mimetizar interacções (proteína-ligando) que os inibidores da LAT1 possuem nas estruturas cristalográficas do PDB. Demonstrou-se assim que métodos computacionais baseados na pesquisa por similaridade estrutural das cavidades das proteínas é uma estratégia promissora para a identificação de “novos” alvos terapêuticos.
Currently, the therapeutic targets used for the treatment of diseases are saturated with the amount of small molecules developed to biologically target them, leading to a significant increase in the cost of discovering new drugs. Since only a small percentage of the proteome is being used for drug development, one of the main alternative solutions is to search for new potential targets in the unexplored part of the proteome and validate them. Thus, we propose a strategy to identify targets in which it is difficult to obtain a pharmacological effect, or which do not yet have any associated information regarding molecules that target them, through a set of computational methods based on the search and similarity of protein cavities between these two types of previously mentioned targets and well studied, and targeted proteins, some of them with approved or in clinical trials drugs. First, we identified and listed all candidate targets classified as undruggable or dark based on work by Oprea and collaborators and grouped them as DOPE, amounting to a total of 15 proteins. Then we grouped all well-studied targets together and named them Well-Knowns. After obtaining the two groups, we proceeded for cavity determination with a program called CAVIAR, followed by a search for structural similar protein cavities using PocketMatch. The cavity corresponding to the active center of LAT1 exhibited a good similarity score with butyrylcholinesterase (BChE) and was considered the most suitable target candidate for further studies, so it proceeded to the final phase. The last step consisted of the virtual screening of small molecules with proven biological activity on butyrylcholinesterase (inhibitors retrieved from ChEMBL) and filtering and identifying the potential best compounds according to their affinities to the new target. From a library of 837 unique compounds, six showed good affinity for Large neutral Amino acid Transporter small subunit 1 (LAT1) and three of them were even able to mimic the interactions established between proteins and inhibitors for which (three-dimensional) experimental structures are available. This study demonstrates that computational methods based on searching for structural similarity of protein cavities is a promising strategy for the identification/validation of new targets and ultimately potential therapeutic agents.
Currently, the therapeutic targets used for the treatment of diseases are saturated with the amount of small molecules developed to biologically target them, leading to a significant increase in the cost of discovering new drugs. Since only a small percentage of the proteome is being used for drug development, one of the main alternative solutions is to search for new potential targets in the unexplored part of the proteome and validate them. Thus, we propose a strategy to identify targets in which it is difficult to obtain a pharmacological effect, or which do not yet have any associated information regarding molecules that target them, through a set of computational methods based on the search and similarity of protein cavities between these two types of previously mentioned targets and well studied, and targeted proteins, some of them with approved or in clinical trials drugs. First, we identified and listed all candidate targets classified as undruggable or dark based on work by Oprea and collaborators and grouped them as DOPE, amounting to a total of 15 proteins. Then we grouped all well-studied targets together and named them Well-Knowns. After obtaining the two groups, we proceeded for cavity determination with a program called CAVIAR, followed by a search for structural similar protein cavities using PocketMatch. The cavity corresponding to the active center of LAT1 exhibited a good similarity score with butyrylcholinesterase (BChE) and was considered the most suitable target candidate for further studies, so it proceeded to the final phase. The last step consisted of the virtual screening of small molecules with proven biological activity on butyrylcholinesterase (inhibitors retrieved from ChEMBL) and filtering and identifying the potential best compounds according to their affinities to the new target. From a library of 837 unique compounds, six showed good affinity for Large neutral Amino acid Transporter small subunit 1 (LAT1) and three of them were even able to mimic the interactions established between proteins and inhibitors for which (three-dimensional) experimental structures are available. This study demonstrates that computational methods based on searching for structural similarity of protein cavities is a promising strategy for the identification/validation of new targets and ultimately potential therapeutic agents.
Description
Tese de mestrado, Química Medicinal e Biofarmacêutica, 2022, Universidade de Lisboa, Faculdade de Farmácia.
Keywords
Target validation Computational methods Cavity similarity LAT1 Teses de mestrado - 2022