虛擬老鼠:系統性電腦預測臨床前 ADMET 參數

在現代藥物開發的過程中，如何增加藥物開發成功的機率，一直是個十分重要的問題。數以百計的候選藥物，往往只有數個最後可以進入市場。為了增加藥物開發成功的機率，在進入臨床試驗前，進行藥物的 ADMET(吸收、分佈、代謝、分泌、毒性)性質研究，已經是普遍性的作法。然而，若以傳統方式，全以生物實驗來進行臨床前測試，有著損耗大量金錢、時間的缺點。因而近年來，許多電腦模型被提出來做這些 ADMET 性質測試的替代方案。以電腦模型來做測試有著快速、省錢的優勢，同時也能有不錯的預測效果。儘管如此，現有被提出的工具有著整合性低、預測缺乏結構資訊的缺點。在此研究中，我們完成了一個以虛擬老鼠為名的軟體原型，結合多種重要的臨床前測試所會用到的指標並提供預測基於的結構資訊。包括 hERG 抑制劑預測、細胞色素 P450 抑制劑預測、致突變性檢測(Ames 測試)、血腦障壁穿透預測、細胞毒性預測、Caco-2 細胞吸收模型預測。輔以方便使用的圖形介面與可輸出報告的功能。讓藥物的臨床前測試能以快速、有效的方式進行。同時我們也對使用到的模型進一步以上市後下架、開發過程中中止開發的藥物做驗證。證實我們所提出的軟體、模型是確實有其功能與預測的效果。

關鍵字

新藥開發；性質預測

並列摘要

In drug discovery, a critical challenge is that how to improve the drug development method to increase the probability of success of the projects. On average, the success rate of a drug from phase I to approval (to market) is only 9.6%. It’s not an easy task. Particularly, the attrition rate is around 75% in preclinical stage, which is the most challenging phase of drug development projects. To make drug candidate to enter phase I, it has been become standard process recently to preform ADMET (Absorption, Distribution, Metabolism, Excretion, Toxicity) profiling before entering clinical trial. However, experimental determination of these properties is time-consuming and costly. For example, to test Caco-2 permeability, it takes one and half of month at the price of 1,200 US dollars from non-GLP (Good Laboratory Practice) laboratory. Therefore, computational methods have been proposed in many preclinical studies to predict those as an alternative approach in preclinical stage. It’s obvious the computational methods have the advantage of being fast and cheap compared to traditional experiments. However, only few web-based tools have integrated predictions of different ADMET properties. The web-based tools are concerned with information security during internet transmission. Those tools also have drawbacks that they don’t provide structural information and only simply provide prediction results. In this study, we finished a prototype, named as Virtual Rat, which not only provide predictions of important ADMET properties, but also the structural information that the prediction results based on. The properties we included in Virtual Rat are predictions of hERG inhibition, cytochrome P450 (CYPs) inhibition, mutagenicity (Ames test), blood-brain barrier penetration, cytotoxicity, and Caco-2 permeability. All of these properties are the reasons of high attrition rate and withdrawn rate in drug discovery. As a stand-alone application with easy-use user interface and functionality to export report in Microsoft word format for further usage, Virtual Rat makes preclinical screening of properties mentioned above fast and effective. We also validate our models using 578 withdrawn or discontinued drugs since 1960s to confirm the availability of Virtual Rat.

並列關鍵字

QSAR ； ADMET ； property prediction.

參考文獻

1. Thomas, D. W.; Burns, J.; Audette, J.; Carrol, A.; Dow-Hygelund, C.; Hay, M., Clinical development success rates 2006-2015. San Diego: Biomedtracker/Washington, DC: BIO/Bend: Amplion 2016.

Google Scholar

2. Pammolli, F.; Magazzini, L.; Riccaboni, M., The productivity crisis in pharmaceutical R&D. Nature reviews Drug discovery 2011, 10 (6), 428-438.

Google Scholar

3. Hodgson, J., ADMET--turning chemicals into drugs. Nature Biotechnology 2001, 19 (8), 722.

Google Scholar

4. Health, U. D. o.; Services, H., Innovation or stagnation? Challenge and opportunity on the critical path to new medical products. http://www. fda. gov/ScienceResearch/SpecialTopics/CriticalPathInitiative/CriticalPathOpportu nitiesReports/ucm077262. htm Accessed July 19, 2010 2004.

Google Scholar

5. Lynch, T. a. P., A, The effect of cytochrome P450 metabolism on drug response, interactions, and adverse effects. American family physician 2007, 76, 391.

Google Scholar

國際替代計量

虛擬老鼠:系統性電腦預測臨床前 ADMET 參數

全文下載

主題瀏覽