探索激酶抑制劑之選擇性與結合機制以開發抗癌藥物

Abstract

激酶在生物體的訊息傳遞中扮演樞紐角色並藉此調控細胞進程，也因此成為包含癌症在內之許多疾病的治療標靶。然而由於所有的激酶都具有演化上保留的三磷酸腺苷結合位，使得設計具選擇性之激酶抑制劑仍是項亟需面對而又艱困的挑戰。從目前已有超過三萬種激酶抑制劑被開發出來，美國食品藥物管理局卻只通過二十八種小分子藥物可見一斑。此低臨床藥化率可歸咎於抑制劑缺乏功效與選擇性、針對特定疾病難以找出適當的治療標靶與現今未考慮標靶在疾病中扮演之角色的對激酶抑制劑選擇性之不佳評估方式所致。 為了克服上述問題，我們提出奠基於以描述蛋白結合位點與其偏好之化合物官能基的錨點輿圖來探討人類磷脂肌醇與蛋白激酶之抑制劑選擇性與結合機制，同時結合『蛋白激酶-抑制劑-疾病譜系輿圖』與考慮標靶癌化角色之『藥化計分方程』指導蛋白激酶抑制劑之最佳化並度量其於癌症治療上之臨床藥化率。針對蛋白激酶，我們提出一個嶄新的概念『蛋白激酶-抑制劑家族』，配合之前利用數個錨點描述蛋白結合位之物化性質與交互作用傾向以解釋蛋白-抑制劑間的結合機制並成功發現許多新型抑制劑的『蛋白結合位點-化合物官能基輿圖』，將能以『蛋白結合位點-化合物官能基輿圖』之共有錨點描述一群具有相似交互作用介面之蛋白激酶-抑制劑交互作用對間的結合機制，並透過比較不同家族間的共有錨點來探討蛋白激酶抑制劑之選擇性。一群具有相似交互作用介面之蛋白激酶-抑制劑交互作用對，我們稱為『蛋白激酶-抑制劑家族』，其定義如下：(1)『蛋白激酶-抑制劑家族』中的蛋白激酶間需具有顯著的序列相似性； (2)『蛋白激酶-抑制劑家族』中的抑制劑間需具有顯著的拓樸相似性；(3)『蛋白激酶-抑制劑家族』中的蛋白激酶-抑制劑交互作用介面間需具有顯著的交互作用相似性。而由保留性作用殘基組成的蛋白結合位點、蛋白結合位點之化合物官能基偏好與包含靜電力、氫鍵、凡得瓦力等蛋白結合位點-化合物官能基間之交互作用型態等三個必要元素構成之『蛋白結合位點-化合物官能基輿圖』錨點恰好適合描述蛋白激酶-抑制劑交互作用介面間的交互作用。 我們更據此建立了全球第一個『蛋白激酶-抑制劑-疾病譜系輿圖』資料庫來探討蛋白激酶-抑制劑家族與蛋白激酶-抑制劑-疾病的關聯性。此資料庫包含了1208個蛋白激酶-抑制劑家族、962個蛋白激酶-抑制劑-疾病關聯性、399種人類蛋白激酶、35788個蛋白激酶抑制劑、186985組蛋白激酶-抑制劑活性試驗，其中具55603組蛋白激酶-抑制劑交互作用 (≦10 μM)、339種疾病與638個疾病等位基因變異體。大規模蛋白激酶抑制活性試驗結果顯示蛋白激酶-抑制劑家族中的蛋白激酶成員常可被抑制劑成員抑制，而此特性結合蛋白激酶-抑制劑-疾病關聯性，有助於舊藥新用與副作用預測。此外，這些錨點還可反映蛋白激酶的構型、功能與抑制劑選擇性。我們亦將此概念運用於探索磷脂肌醇激酶，並發現磷脂肌醇激酶與蛋白激酶有相似的錨點形態，且可被類黃酮所抑制，尤其是槲皮素與萃取自台灣特有大金星蕨之原芹菜素WYC02，後續細胞實驗更證明WYC02可經由影響磷脂肌醇激酶訊息傳遞途徑而抑制子宮頸癌的生長與癌細胞增生。 最後，我們比較了癌症病人與其對應之正常組織的蛋白表現量，藉此了解標靶激酶在特定癌症中扮演的角色，並將其引入蛋白激酶抑制劑選擇性之評估方法中，而提出了『藥化計分方程』以度量蛋白激酶抑制劑於癌症治療上之臨床藥化率。蛋白激酶抑制劑抑制譜結果則顯示具高臨床藥化率之蛋白激酶抑制劑可根據結合蛋白激酶-抑制劑-疾病譜系輿圖的最佳化指引被設計出來。 我們相信『蛋白激酶-抑制劑家族』與『蛋白激酶-抑制劑-疾病譜系輿圖』資料庫可對激酶抑制劑之選擇性與結合機制提供生物性的見解，而『藥化計分方程』得分則可反映設計癌症藥物時的臨床成功率，並可根據病人的蛋白表現情形作為個人化醫療的依據。根據『蛋白激酶-抑制劑-疾病譜系輿圖』與『藥化計分方程』，我們則可針對特定癌症設計出具高臨床藥化率之蛋白激酶抑制劑。

Kinases play central roles in signaling pathways to control cellular processes and are promising therapeutic targets for many diseases, especially in cancers. Designing selective kinase inhibitors is an emergent and challenging task, because kinases share an evolutionary conserved ATP-binding site. To date, over thirty thousand kinase inhibitors have been identified; however, only 28 small molecule drugs have been approved by US FDA. The low clinical development success rates for investigational inhibitors may result from lack of inhibitor efficacy and selectivity, difficulty in drug target validation for particular diseases, as well as incorrect evaluation without considering the roles of target kinases in particular diseases for inhibitor selectivity. To address these issues, we proposed anchor (pocket–moiety interaction preference)-based maps to understand kinase inhibitor selectivity and binding mechanisms for 399 of 518 human protein kinases and human PIK3 family lipid kinases, as well as “Approvance scores” to quantify the clinical development success rates of inhibitors for particular cancers. For protein kinases, we have proposed a new concept of kinase-inhibitor family (KIF) that represents a set of kinase-inhibitor complexes with similar interfaces and consensus binding mechanisms. A KIF can be defined as follows: (1) the kinases in the KIF with significant sequence similarity; (2) the inhibitors in the KIF with significant topology similarity; (3) the kinase-inhibitor interactions (KIIs) in the KIF with significant interaction similarity. Interaction similarities were evaluated by site-moiety maps (SiMMaps). A SiMMap represents physicochemical properties and interaction preferences of a protein binding site by several statistical anchors, consisted of a binding pocket with conserved interacting residues, moiety preferences of the pocket, and pocket-moiety interaction type (electrostatic, hydrogen-bonding, or van der Waals). By integrating SiMMaps, the KIIs within a KIF are often conserved on some consensus KIDFamMap anchors, which represent conserved interactions between the kinase subsites and consensus moieties of their inhibitors. Furthermore, to explore kinase-inhibitor families and kinase-inhibitor-disease (KID) relationships for kinase inhibitor selectivity and mechanisms, we built the first KIDFamMap database, includes 1,208 KIFs, 962 KIDs, 55,603 KIIs, 35,788 kinase inhibitors, 399 human protein kinases, 339 diseases, and 638 disease allelic variants. Our experimental results reveal that the members of a KIF often possess similar inhibition profiles. The KIDFamMap anchors can reflect kinase conformations types, kinase functions, and kinase inhibitor selectivity. The same anchor-based concept was also applied to human PIK3 family lipid kinases. We found that PIK3 family lipid kinases have similar anchor pattern to protein kinases and can be targeted by flavonoids, especially quercetin and WYC02, isolated from the whole native Taiwan fern, Thelypteris torresiana. Further cell-based functional experiments showed that WYC02 inhibited cervical cancer cell proliferation and tumor growth through PIK3 Signaling pathway. Finally, we explored expression data of cancer patients and corresponding normal tissues to understand the roles of target kinases in particular cancers, and introduced them to refine current kinase inhibitor selectivity evaluation methods and proposed “Approvance scores” to quantify the clinical development success rates of inhibitors for particular cancers. Moreover, combining the KIDFamMap optimizing guidance, kinase inhibitors with high approvance scores can be designed. Our results show that the kinase candidates identified from expression data for computing approvance score were highly correlated with cancer-related genes and biological processes of gene ontology, as well as approvance scores are consistent with the efficiencies of inhibitors. Furthermore, kinase profiling results also show that the optimizing guidance of KIDFamMap is able to design kinase inhibitors with high approvance scores. We believe that the concept of KIF and KIDFamMap database can provides biological insights into kinase inhibitor selectivity and binding mechanisms, as well as the approvance scores can reflect an index to design the drugs of a particular disease and provide personalized medicine according to the patient’s gene expressions. According to KIDFamMap and approvance scores, we can design kinase inhibitors for particular diseases with high clinical development success rates.