Genetic Analysis of the Multidrug Resistance Phenotype in Tumor Cells

化疗耐药发生在癌细胞中，是因为特定蛋白质表达的内在或获得性变化。我们已经研究了对天然产物化疗药物的耐药性，如阿霉素、长春花生物碱和紫杉醇，以及最近的组蛋白去乙酰化酶抑制剂和靶向激酶抑制剂。在大多数情况下，由于细胞内药物浓度降低，细胞同时对多种药物产生耐药性。对于天然产物药物，这种交叉耐药通常是由于被称为p -糖蛋白（P-gp）的能量依赖性药物外排系统（ABC转运蛋白）的表达，p -糖蛋白是MDR1或ABCB1基因的产物，或ABC转运蛋白家族的其他成员，包括ABCG2和ABCB5。我们的实验室和其他实验室的研究表明，一些药物对表达p- gp的细胞的毒性比不表达p- gp的细胞更大，这提示了一种治疗耐多药癌症的新方法。已经确定了几种具有这种性质的不同化学类别，包括硫代氨基脲（例如NSC73306）。通过对NSC73306类似物的定量结构活性分析，对NCI-60细胞系的进一步相关性分析，以及对美国药书中杀死表达p- gp细胞的化合物的高通量筛选，我们发现了许多额外的化合物，它们具有更好的选择性杀死表达p- gp细胞的能力，而且还具有更好的溶解度。为了研究这些药物对表达P-gp的癌细胞的作用，我们建立了人类肾上腺皮质癌的小鼠异种移植模型，这是一种本质上表达高水平P-gp的癌症。ABC转运蛋白不仅与癌症的耐药性有关，而且还是血脑屏障（BBB）和血胎盘屏障的主要组成部分。血脑屏障中三种最重要的转运蛋白是ABCB1、ABCC1和ABCG2。我们开发了一种分析血脑屏障和血胎盘屏障中ABCG2表达的系统，基于荧光素是这些屏障中的ABCG2底物，并且在荧光素酶在血脑屏障或血胎盘屏障中表达的转基因小鼠中可以检测到其进入大脑或进入发育中的胎儿。由于对小鼠血脑屏障的研究既耗时又昂贵，我们正在对斑马鱼的血脑屏障进行平行分析，因为斑马鱼血脑屏障的成分似乎与哺乳动物的血脑屏障非常相似。为了了解P-gp的结构如何决定其多特异性，以及特异性如何随着折叠的变化而改变，我们与LCB的其他高级研究人员（包括Di Xia， Suresh Ambudkar和Sriram Subramaniam）合作。Cryo-EM研究表明载脂蛋白P-gp具有两个atp结合位点分离或靠近的动态结构。ATP的结合使P-gp的构象处于后一种状态，ATP的水解导致ATP位点的分离。以小鼠P-gp为模型的晶体学研究表明，ATP位点之间的分离决定了底物结合的跨膜（TM）螺旋的间距，这表明，当ATP位点一起或分开移动时，TM螺旋暴露出不同的残基，使其能够与许多不同的底物结合。对小鼠-人嵌合P-gps的研究揭示了这两种进化相关转运体的相似结构-功能关系。我们已经创建了一种高度敏感的定量检测方法，用于检测ABC转运蛋白mrna和其他与培养癌细胞耐药相关的mrna。我们对人类卵巢癌、肝细胞癌（HCC）和急性髓性白血病（AML）的样本进行了一些详细的研究。在卵巢癌中，有一个11个基因的耐多药特征与化疗反应差有关。在HCC中，信号更为复杂，但能准确区分预后差和预后好的肿瘤。将基因表达模式从不良预后模式转变为良好预后模式的药物也使培养的HCC细胞对抗癌药物敏感。对于AML，研究了同一患者化疗前后的样本。在这种情况下，每个病例的耐药表现出ABC基因和其他耐多药基因的不同表达模式，这表明需要个性化的治疗耐药方法。ABCB5是ABCB1的分子近亲。它在大脑和眼睛的色素细胞以及黑色素瘤中表达。在与Richard Cannon（新西兰奥特加大学）的合作中，我们已经证明，当ABCB5在酵母中表达时，它是一种多药物转运体。ABCB5基因敲除小鼠对主要镇静剂氟哌啶醇敏感，这与该转运体在大脑中的作用一致（与斯坦福医学院的Gary Peltz合作）。

Resistance to chemotherapy occurs in cancer cells because of intrinsic or acquired changes in expression of specific proteins. We have studied resistance to natural product chemotherapeutic agents such as doxorubicin, Vinca alkaloids, and taxol and more recently, histone deacetylase inhibitors and targeted kinase inhibitors. In most cases, cells become simultaneously resistant to multiple drugs because of reductions in intracellular drug concentrations. For the natural product drugs, this cross-resistance is frequently due to expression of an energy-dependent drug efflux system (ABC transporter) known as P-glycoprotein (P-gp), the product of the MDR1 or ABCB1 gene, or to other members of the ABC transporter family, including ABCG2 and ABCB5. Work from our laboratory and others has revealed that some drugs are more toxic to P-gp-expressing cells than to non-expressors, suggesting a novel approach to treatment of MDR cancers. Several different chemical classes with this property, including thiosemicarbazones (e.g., NSC73306), have been identified. A quantitative structure activity analysis of NSC73306 analogs, a further correlation analysis in the NCI-60 cell lines, and a high-throughput screen for compounds in the U.S. Pharmacopeia that kill P-gp-expressing cells have yielded many additional compounds with improved ability to kill selectively P-gp-expressing cells, but also with improved solubility properties. In order to study the effect of these agents that target cancer cells expressing P-gp, we have developed a mouse xenograft model of human adrenocortical carcinoma, a cancer that intrinsically expresses high level of P-gp. Not only are ABC transporters responsible for drug resistance in cancer, but they are a major component of the blood-brain barrier (BBB) and blood-placental barrier. The three most prominent transporters at the blood-brain barrier are ABCB1, ABCC1, and ABCG2. We have developed a system for analysis of ABCG2 expression at the blood-brain and the blood-placental barriers based on the fact that luciferin is an ABCG2 substrate at these barriers and its passage into the brain or into developing fetuses can be detected in transgenic mice in which luciferase is expressed at the blood-brain barrier or blood-placental barrier. Because studies of the BBB in mice are time-consuming and expensive, we are developing a parallel analysis in zebrafish, as components of the zebrafish BBB appear to be very similar to those of the mammalian BBB. To understand how the structure of P-gp determines its polyspecificity and how specificity is altered with changes in folding, we have collaborated with other senior investigators in the LCB, including Di Xia, Suresh Ambudkar, and Sriram Subramaniam. Cryo-EM studies have demonstrated that apo P-gp has a dynamic structure in which the two ATP-binding sites are either separated or close together. Binding of ATP fixes the conformation of P-gp in the latter state and ATP hydrolysis results in separation of the ATP sites. Crystallography studies using mouse P-gp as a model show that the separation between the ATP sites determines the pitch of the transmembrane (TM) helices where substrates bind, suggesting the hypothesis that as the ATP sites move together and apart, the TM helices expose different residues that enable binding to many different substrates. Studies on mouse-human chimeric P-gps have revealed similar structure-function relationships for these two evolutionarily related transporters. We have created a highly sensitive, quantitative assay for ABC transporter mRNAs and other mRNAs associated with drug resistance in cultured cancer cells. We have studied samples from human ovarian cancer, hepatocellular cancer (HCC), and acute myelogenous leukemia (AML) in some detail. In ovarian cancer, there is an 11-gene MDR signature associated with poor response to chemotherapy. In HCC, the signature is more complex, but accurately distinguishes poor prognosis vs. better prognosis cancer. Agents that change the pattern of gene expression from poor prognosis to better prognosis patterns also sensitize cultured HCC cells to anti-cancer drugs. For AML, samples from the same patients before and after chemotherapy were studied. In this case, resistance in each case shows a different pattern of expression of ABC genes and other MDR genes, suggesting that individualized approaches to resistance to therapy will be needed. ABCB5 is a close molecular relative of ABCB1. It is expressed in pigmented cells in the brain and eye, and in melanoma. In collaboration with Richard Cannon (University of Otega, New Zealand) we have shown that when expressed in yeast, ABCB5 is a multidrug transporter. ABCB5 knock-out mice are sensitive to the major tranquilizer haloperidol, consistent with a role of this transporter in the brain (with Gary Peltz, Stanford School of Medicine).