Lauria Antonino, Bonsignore Riccardo, Bartolotta Roberta, Perricone Ugo, Martorana Annamaria, Gentile Carla
Dipartimento di Scienze e Tecnologie Biologiche Chimiche e Farmaceutiche "STEBICEF", Universita di Palermo, Via Archirafi 32 I-90123 Palermo, Italy.
Curr Pharm Des. 2016;22(21):3073-81. doi: 10.2174/1381612822666160224142323.
Polypharmacology, defined as the modulation of multiple proteins rather than a single target to achieve a desired therapeutic effect, has been gaining increasing attention since 1990s, when industries had to withdraw several drugs due to their adverse effects, leading to permanent injuries or death, with multi-billiondollar legal damages. Therefore, if up to then the "one drug one target" paradigm had seen many researchers interest focused on the identification of selective drugs, with the strong expectation to avoid adverse drug reactions (ADRs), very recently new research strategies resulted more appealing even as attempts to overcome the decline in productivity of the drug discovery industry.
Polypharmacology consists of two different approaches: the former, concerning a single drug interacting with multiple targets related to only one disease pathway; the latter, foresees a single drug's action on multiple targets involved in multiple disease pathways. Both new approaches are strictly connected to the discovery of new feasible off targets for approved drugs.
In this review, we describe how the in silico facilities can be a crucial support in the design of polypharmacological drug. The traditional computational protocols (ligand based and structure based) can be used in the search and optimization of drugs, by using specific filters to address them against the polypharmacology (fingerprints, similarity, etc.). Moreover, we dedicated a paragraph to biological and chemical databases, due to their crucial role in polypharmacology.
Multitarget activities provide the basis for drug repurposing, a slightly different issue of high interest as well, which is mostly applied on a single target involved in more than one diseases. In this contest, computational methods have raised high interest due to the reached power of hardware and software in the manipulation of data.
多药理学被定义为通过调节多种蛋白质而非单一靶点来实现预期治疗效果。自20世纪90年代以来，它越来越受到关注。当时，制药行业因药物不良反应导致永久性损伤或死亡，并面临数十亿美元的法律赔偿，不得不撤回几种药物。因此，在此之前，“一药一靶”模式使许多研究人员将兴趣集中在选择性药物的研发上，强烈期望避免药物不良反应（ADR）。然而，最近新的研究策略更具吸引力，尽管其目的是克服药物研发行业生产率的下降。
多药理学包括两种不同的方法：前者涉及一种药物与仅一条疾病途径相关的多个靶点相互作用；后者则预见一种药物对多条疾病途径中涉及的多个靶点起作用。这两种新方法都与发现已批准药物新的可行脱靶效应密切相关。
在本综述中，我们描述了计算机辅助工具如何在多药理学药物设计中起到关键支持作用。传统的计算协议（基于配体和基于结构）可用于药物的搜索和优化，通过使用特定筛选器针对多药理学特性（指纹、相似性等）进行筛选。此外，由于生物和化学数据库在多药理学中的关键作用，我们专门用一段内容进行了阐述。
多靶点活性为药物重新利用提供了基础，这也是一个备受关注的稍有不同的问题，主要应用于涉及多种疾病的单一靶点。在这种情况下，由于硬件和软件在数据处理方面的强大能力，计算方法引起了高度关注。
