Relating GPCRs by biased ligands for enhanced therapeutic efficacy

描述(由申请人提供)： 目标：目标是扩大Arrestin和G蛋白偏向配体之间的已知药理关系。从GI偶联的多巴胺D2受体(D2R)开始，我们将预测有偏见的配体，回顾和通过前瞻性实验测试对我们的方法进行评估。然后，我们将这项工作扩展到其他受体，并最终将结果组织到与受体多药理学相关的图中，根据其配体的功能选择性。 意义重大。尽管优先激活arrestin或典范信号的偏向配体可能包含了在结构良好的G蛋白偶联受体中尚未开发的新的治疗机会，但他们的发现往往是偶然的，几乎没有预测的计算模型。该领域缺乏实际的有偏见的配体，其不完全探索的行列甚至包括市场上的药物，如阿立哌唑，这可能归功于这种机制的疗效。新的筛查能力来衡量？-在大规模的1/2招募中，为这里提出的计算方法、实验证实的偏向配体和“偏向多元药理学”图提供了独特和及时的支持。 理论/背景：并不是所有的配体都是平等的。对于GPCRs，在选择性、多药理学和药代动力学方面相似的配体表现出不同的第二信使信号和治疗效果的模式。这种“功能选择性”将激活信号转导的配体区分为典型的(G偶联)和阻滞蛋白1/2(又名阻滞素3/4，以下简称“arrestin”)途径。途径参与是配体特异性的，并有助于药物的结果。例如，产生较少副作用的有效的第三代抗精神病药物最初被认为是D2R的部分激动剂，但可能只通过arrestin2发出信号。 目的1：预测和测试新的芳香素配体。有偏向的配体被认为可以稳定跨膜受体的构象集合，从而优先激活信号通路；然而，在实践中，我们必须更多地从配体的效力推断受体的构象，而不是相反。我们将追求两个目标：a.利用其对配体结构的唯一依赖，按有效性分组，我们将采用相似系综方法(SEA)来区分有偏见的配体基团。B.此外，我们将在D2R预测和测试用于抗精神病药物的新的偏向配体。 里程碑。该系统的基本特征是存在的，并在使用SEA对20种药物的35多个目标进行预测时证明了原理的正确性。在这里，我们提炼了SEA的“靶标”概念来编码配基偏向。有两个务实的里程碑。I.评估五种计算方法，这些方法利用配体中心的观点来预测新的有偏见的配体，从单独的SEA到内部的“标记集”受体签名。二、前瞻性地在实验中测试20种Arrestin偏向的D2R配体预测，并根据需要推进(A)一种药物用于抗精神病动物模型或(B)多达三种临床前化合物用于PK图谱。 目的2：通过有偏向的配体研究GPCRs的药理作用。我们将扩展我们的所有GPCR的地图，以反映有偏见的配体类别。这将使我们能够询问该方法的影响范围和影响：a.确定偏向于arrestin的配体在图谱上与同一受体的正则配体更紧密地联系在一起，还是与不相关受体的偏重arrestin配体联系更紧密。B.采用目标1中的方法来预测和测试进一步治疗受体的偏向配体。 里程碑。我们将通过以下方式将功能选择性与已知GPCR药理学的更广泛背景联系起来：i.将本研究和文献中的有偏见的配体结合到受体关系的全球图谱中。二、通过预测和测试从治疗受体列表中选择的两个受体(总共40个)中的每一个上的另外20个有偏见的配体，展示了广泛的适用性，这些配体具有已知的表现出有偏见的下游信号的配体。尽管这些目标诚然雄心勃勃，但初步结果表明它们是可行的。

DESCRIPTION (provided by applicant): Goals: The goal is to expand the known pharmacological relationships among arrestin and G protein biased ligands. Starting with the Gi-coupled dopamine D2 receptor (D2R), we will predict biased ligands, evaluating our methods both retrospectively and via prospective experimental assay. We then extend this work to additional receptors, and finally organize the results into a map relating receptor polypharmacology by the functional selectivity of their ligands. Significance. Whereas biased ligands that preferentially activate arrestin or canonical signaling may comprise novel therapeutic opportunities unexploited among the well trammeled G protein coupled receptors, their discovery is more often serendipitous and few if any predictive computational models exist for it. The field suffers from a paucity of actual biased ligands, whose incompletely explored ranks include even marketed drugs such as aripiprazole, which may owe its efficacy to this mechanism. New screening capabilities to measure ?-arrestin1/2 recruitment at large scale provide unique and timely support for the computational methods, experimentally confirmed biased ligands, and "biased polypharmacology" maps proposed here. Theory/Background: Not all ligands are created equal. For GPCRs, ligands similar in selectivity, polypharmacology, and pharmacokinetics nonetheless exhibit differing patterns of second messenger signaling and therapeutic efficacy. This "functional selectivity" differentiates ligands that activate signal transduction via the canonical (G coupled) and the ¿arrestin1/2 (aka arrestin3/4, hereafter "arrestin") pathways. Pathway engagement is ligand specific and contributes to drug outcomes. As one example, effective third generation antipsychotics that induce fewer side effects were originally thought to be partial agonists of D2R, but may instead signal exclusively via ¿arrestin2. Aim 1: To predict and test new arrestin-biased ligands. Biased ligands are thought to stabilize transmembrane receptor conformational ensembles to preferentially activate a signaling pathway; in practice however we must more often deduce receptor conformations from ligand efficacy than the other way around. We will pursue two goals: a. Leveraging its sole reliance on ligand structures, grouped by efficacy, we will adapt the Similarity Ensemble Approach (SEA) to differentiate groups of biased ligands. b. Further, we will predict and test new biased ligands at D2R for antipsychotic use. Milestones. The essential features of this system exist, and proof of principle has been demonstrated in predicting new targets for over 35 targets for 20 drugs using SEA. Here we refine SEA's notion of "target" to encode ligand bias. There are two pragmatic milestones. i. Evaluating five computational methods that leverage a ligand centric viewpoint to predict new biased ligands, from SEA alone to in house "marker set" receptor signatures. ii. Prospectively testing 20 arrestin biased D2R ligand predictions experimentally and advancing (a) one drug to antipsychotic animal models or (b) up to three preclinical compounds to PK profiling, as warranted. Aim 2: To relate GPCR pharmacology by biased ligands. We will extend our maps of all GPCRs to reflect biased ligand classes. This will allow us to interrogate the method's reach and impact: a. Determining whether arrestin biased ligands associate more closely in the map with canonical ligands of the same receptor-or instead with the arrestin biased ligands of an unrelated receptor. b. Adapting methods from Aim 1 to predict and test biased ligands for further therapeutic receptors. Milestones. We will relate functional selectivity to the wider context of known GPCR pharmacology by: i. Combining biased ligands from this study and from the literature into a global map of receptor relationships. ii. Demonstrating broad applicability by predicting and testing a further 20 biased ligands at each of two receptors (for 40 total) selected from the list f therapeutic receptors with ligands known to exhibit biased downstream signaling. Whereas these goals are admittedly ambitious, preliminary results suggest that they are feasible.