Physicochemical properties of antibacterial compounds: Implications for drug discovery
Physicochemical properties of antibacterial compounds: Implications for drug discovery
复制标题
DOI：
10.1021/jm700967e
复制
发表时间：
2008-05-22
影响因子：
7.3
通讯作者：
Moser, Heinz E.
中科院分区：
文献类型：
--
作者：
O'Shea, Rosemarie;Moser, Heinz E.
抗菌药物的发现在上个世纪中期后不久达到顶峰，发现了仍在临床使用的大多数化合物类别。在1962年引入链阳性菌素和喹诺酮类药物后，直到2000年利奈唑胺上市，才发现并批准了新的抗生素类别用于临床。这一事实是相当令人惊讶的，因为在这段时间里，不仅研究得到了大力支持，而且基因组学和高通量筛选等新技术也被引入并应用于提高生产率。随着抗菌药物耐药性的上升，我们将不得不补充抗菌药物的武库，为医生提供未来成功治疗感染的工具。[2]与发现新型抗菌化合物类别相关的部分困难是由对安全的广谱抗生素的严格要求所定义的：靶标必须是必需的，在各种细菌物种中高度保守，并且在人类中不存在，不同或非必需。抑制剂必须是有效的，并且理想地应该显示靶相关的全细胞活性，并且出现耐药性的倾向低。此外，最初的“命中”支架应该服从结构变化，以允许后期“先导”化合物的效力、功效和安全性的优化。许多作者不仅讨论了新型抗菌药物的必要性，以确保未来的治疗方案，而且还讨论了以前在命中识别和铅优化步骤中遇到的困难。3-8虽然抗菌化合物的下降管道可以说反映了这些要求的技术复杂性，但相对较少的出版物提供了有关过去几年各公司遇到的困难的详细信息。葛兰素史克集团最近的一项审查增加了与该主题相关的多个问题的宝贵信息，并深入了解了基于靶向的基因组方法的成功和失败。[9]关于这些化合物本身的性质，作为缺乏新药剂的可能来源，发表的文章相对较少。Payne及其同事的分析说明了抗菌药物发现的挑战，并表明多个参数导致了高损耗率。随着多重耐药病原体的出现和对新型抗生素的需求，尽可能多地了解先前的努力并将所吸取的经验教训应用于未来抗生素的发现至关重要。特别是一个重要的参数之前已经提到，但在我们看来，没有充分分析：抗菌药物的理化性质空间。Lipinski的里程碑式研究11代表了第一次系统地尝试将药物的理化性质与初始命中和随后的后期先导的预测成功匹配相关联。正是这项工作首次将药物的物理化学性质与其口服生物利用度以及临床前和临床开发期间的后续困难和磨损率联系起来。该分析的主要发现是认识到口服候选药物的理想属性空间（MW、亲脂性、氢键供体和受体），以及企业化合物档案已慢慢远离最佳区域的事实。物理化学空间，最有可能是由合成便利性而不是设计驱动的。这种意识对药物发现产生了重大影响，今天在药物发现之前分析这些性质（“五规则”或“Lipinski规则”）是很常见的。
Antibacterial drug discovery peaked shortly after the middle of the past century with the discovery of most compound classes that are still in clinical use. After the introduction of streptogramins and quinolones in 1962, no novel class of antibiotics was identified and approved for clinical use until linezolid was launched in 2000. This fact is rather surprising because not only was research heavily supported during this time but also novel technologies such as genomics and high-throughput screening were introduced and applied to improve productivity. 1 With antibacterial resistance on the rise, we will have to replenish the arsenal of antibacterial drugs to provide physicians with the tools to successfully treat infections in the future. 2 Part of the difficulty associated with the discovery of novel antibacterial compound classes has been defined by stringent requirements for a safe, broad spectrum antibiotic: The target must be essential, highly conserved among various bacterial species, and absent, different, or nonessential in humans. The inhibitor must be potent and should ideally display target-related whole cell activity with a low propensity for the emergence of resistance. Furthermore, the initial “hit” scaffold should be amenable to structural changes to allow for optimization of the potency, efficacy, and safety of later-stage “lead” compounds. A number of authors have discussed not only the necessity of novel antibacterial drugs to ensure future treatment options but also difficulties previously encountered during the hit identification and lead optimization steps. 3–8 While the declining pipeline of antibacterial compounds arguably reflects the technical complexities of these requirements, relatively little has been published providing detailed information about the difficulties encountered at various companies over the past years. A recent review from the group at GlaxoSmithKline added valuable information on multiple issues related to this topic and provided insight into the successes and failures of a target-based, genomic approach. 9 Relatively little though has been published on the nature of the compounds themselves as a possible source for the paucity of new agents. The analysis of Payne and co-workers illustrates the challenge of antibacterial drug discovery and suggests that multiple parameters contribute to the high attrition rate. With the rise of multi-drug-resistant pathogens and the need for novel antibiotics, it is critical to understand as much as possible from prior efforts and to apply learned lessons to the discovery of future antibiotics. One important parameter in particular has previously been mentioned but, in our view, not sufficiently analyzed: the physicochemical property space of antibacterial drugs. 10Lipinski’s landmark study11 represented the first systematic attempt to correlate the physicochemical properties of drugs with the predicted successful matriculation of initial hits and subsequent late-stage leads. It was this work that connected for the first time physicochemical properties of drugs with both their oral bioavailability and their subsequent difficulties and attrition rates during preclinical and clinical development. Major findings from this analysis were the recognition of an ideal property space for orally available drug candidates (MW, a lipophilicity, hydrogen bond donors and acceptors), as well as the fact that corporate compound archives had been slowly moving away from an optimal area of this physicochemical space, most likely driven by synthetic convenience rather than by design. This awareness had a major impact on drug discovery, and today it is common to analyze these properties (the “rule of five” or “Lipinski’s rules”) prior to …