Immobilized Drug Transporters

膜转运蛋白可能在几种药物的药代动力学（在正常组织中表达时）和/或对几种抗癌药物的耐药性（在癌症中表达时）中发挥作用。这些蛋白中有几个属于ABC （ATP结合盒）蛋白超家族，包括p糖蛋白（Pgp, ABCB1）、多药耐药蛋白（MRP1, ABCC1）、米托沙酮耐药蛋白（MXR, ABCG2）、小管多特异性有机阴离子转运蛋白（cMOAT, ABCC2）、胆汁盐输出泵（BSEP, ABCB11）。其他转运蛋白，如有机阳离子转运蛋白（OCT1和OCT2）和有机阴离子转运蛋白（OAT）不属于ABC超家族。这些蛋白质似乎共享几种底物药物，可能包括一些抗癌药物。不同转运体之间抗癌和非抗癌底物的部分共享可以解释用于抑制特定转运体功能的药物的一些副作用。例如，环孢素A已被临床测试为pgp引起的多药耐药抑制剂。它的副作用之一是黄疸，这可能是由于它干扰了cMOAT对共轭胆红素的转运。因此，转运蛋白抑制剂的临床疗效（无论是用于逆转抗癌耐药，还是用于调节药代动力学）都可以通过优化其转运蛋白选择性来提高。Pgp是目前研究最彻底的转运蛋白。Pgp通过将底物从细胞膜内部运送到外小叶或细胞外空间的方式将底物从细胞外排。Pgp生物学的核心是其结合多种底物和抑制剂的能力，这表明可能存在多个结合位点。由于缺乏这些位点的定义和Pgp晶体结构的不可获得性，迄今为止阻碍了针对Pgp功能的药物的合理设计。我们建议亲和层析可以作为表征不同转运体的结合位点和转运周期的手段，并最终定义有效的和蛋白质选择性的药物载体，以药理抑制转运体的功能。我们目前正在表征和优化亲和层析模型的Pgp。这一过程也将为其他ABC和非ABC转运体的建模提供一个原型。将Pgp固定到色谱柱的固定相分为三步：首先，用洗涤剂从Pgp阳性细胞（MDA435/LCC6MDR1人乳腺癌细胞）的膜中提取Pgp；第二，通过透析将Pgp重组为二氧化硅颗粒（IAM珠）上的固定化人工膜；最后，将ppp - iam珠装进不同种类的色谱柱中（因此，我们使用了Amersham的HR5/2玻璃色谱柱或peek管）。阴性对照柱使用亲代pgp阴性LCC6/MDA435细胞的膜获得。正面和分区色谱证实了Pgp底物的保留增加，这在Pgp的阳性柱中尤为明显。底物位移研究在正面色谱允许计算底物亲和力和结合位点的数量。亲和层析和经典过滤结合分析的比较，显示了在不同底物的亲和的定义部分重叠。长春花碱和阿霉素得到了类似的结果，但维拉帕米和环孢素A没有得到类似的结果，这表明IAM模型可能不能完全代表Pgp在其天然膜环境下的结合，应该评估其他固定方法。亲和色谱法可用于评估不同底物的互移，因此也可用于评估两个底物是否与相同或不同的位点结合。到目前为止，我们的评估结果提供了关于几种底物之间相互作用性质的建议，在没有添加ATP的情况下。维拉帕米结合Pgp上的多个位点，其中至少有一个与长春花碱共享。环孢素A在没有ATP和/或其他底物的情况下不与Pgp结合，但它的结合是由长春花碱引起的。然而，环孢素A取代长春花碱，表明相互作用既不是竞争也不是变构，因为它不是互惠的。（+）甲氟喹和（-）甲氟喹与环sprin A和长春花碱相互作用，前者对环sprin A和长春花碱具有对映选择性，后者则没有。通过正面色谱和非线性区域色谱分析对底物位移进行更广泛的评估，将通过确定哪些底物共享相同的位点来完成结合位点的功能定义。必须指出的是，在没有ATP的情况下，这些评价可能不是最佳的，因为当ATP与Pgp结合时，结合位点的构象可能与ATP水解前的构象大不相同。后一种构象可能与有效药效团的定义更相关。我们观察到，ATP的加入改变了其对底物的亲和力：对长春碱和维拉帕米的亲和力降低，而对环孢素A的亲和力增加。在ATP存在的情况下与Pgp结合，必然代表了Pgp对其底物亲和的“平均”，因为它经历了运输周期的不同步骤，包括ATP催化ADP和无机磷酸盐的之前和之后的步骤。可以预期，在催化后阶段，当底物必须释放到膜的细胞外侧时，对有效底物的亲和力将降低。在过滤结合分析的情况下，亲和层析可以用来模拟Pgp的运输周期的不同阶段。例如，使用不可水解的ATP类似物将允许在“预催化”阶段对Pgp进行建模，并且在底物存在的情况下，用钒酸盐和ATP对Pgp柱进行预处理将使其在催化阶段“冻结”。在初步研究中，我们使用了带状亲和色谱法来评估长春花碱与“钒酸盐捕获”pgp柱的结合。结果证实长春花碱对催化后Pgp的结合有抑制作用。这些评价将扩展到不同的基材和运输周期的不同阶段。总之，我们已经获得并正在表征一个亲和色谱模型，用于评估Pgp结合位点和运输周期的功能。我们将尝试进一步优化我们的亲和色谱模型，例如，我们可以评估使用不同方法将Pgp固定在其天然膜中的色谱柱。然后，我们计划通过识别在运输周期的不同阶段共享相同位点的底物，对Pgp的结合位点进行全面的功能表征。位点共享底物的结构亲和关系将用于定义药效团。其他ABC和非ABC转运体（如MRP1和OCT1）将被建模，结果的比较将用于定义转运体选择性药效团。

Membrane transporter proteins may play a role both in the pharmacokinetics of several drugs (when expressed in normal tissues) and/or resistance to several anticancer agents (when expressed in cancers). Several of these proteins belong to the ABC (ATP Binding Cassette) protein superfamily, including, P-glycoprotein (Pgp, ABCB1), the Multidrug Resistance Protein (MRP1, ABCC1), the Mitoxantrone Resistance Protein (MXR, ABCG2), the canalicular Multispecific Organic Anion Transporter (cMOAT, ABCC2), the Bile Salt Export Pump (BSEP, ABCB11). Other transporter proteins, such the Organic Cation transporters (OCT1 and OCT2) and the Organic Anionic Transporter (OAT) do not belong to the ABC superfamily. These proteins seem to share several substrate drugs including, possibly, some anticancer agents. The partial sharing of anticancer and non-anticancer substrates between different transporters may explain some of the side effects of drugs used to inhibit the function of specific transporters. For example, Cyclosporin A has been tested clinically as an inhibitor of Pgp-caused multidrug resistance. One of its side effects is jaundice, which might be caused by its interference of the transport of conjugated bilirubin by cMOAT. So, the clinical efficacy of transporter inhibitors (whether used to reverse anticancer drug resistance, or to modulate pharmacokinetics) may be improved by optimizing their transporter selectivity. Pgp is the transporter that has been most thoroughly studied. Pgp effluxes its substrates from cells by transporting them from the cell membrane's inner to the outer leaflet or to the extracellular space. Central to the biology of Pgp is its ability to bind a wide array of diverse substrates and inhibitors, suggesting the possible existence of multiple binding sites. The lack of definition of these sites and the unavailability of a crystal structure for Pgp have so far hindered a rational design of drugs targeting Pgp function. We propose that affinity chromatography can be used as a means to characterize the binding sites and transport cycle of the different transporters and, ultimately, to define effective and protein-selective pharmacophores for the pharmacological inhibition of transporter function. We are presently characterizing and optimizing affinity chromatography models of Pgp. This process will also provide a prototype for the modeling of other ABC and non-ABC transporters. Pgp immobilization into the stationary phase of chromatographic columns is obtained in three steps: first, Pgp is extracted by detergent from the membranes of Pgp positive cells (MDA435/LCC6MDR1 human breast cancer cells); second, Pgp is reconstituted by dialysis into Immobilized Artificial Membranes on silica particles (IAM beads); finally, the Pgp-IAM beads are packed into different kinds of chromatographic columns (so, we have used either Amersham's HR5/2 glass chromatographic columns or peek tubing). Negative control columns are obtained using membranes form parental Pgp-negative LCC6/MDA435 cells. Frontal and zonal chromatography has confirmed an increased retention of Pgp substrates, which is specifically evident in Pgp' positive columns. Substrate displacement studies in frontal chromatography allow to calculate substrate affinity and number of binding sites. Comparison of affinity chromatography and classical filtration binding assays, shows a partial overlapping in the definition of the affinity of different substrates. Similar results were obtained for vinblastine and doxorubicin, but not for verapamil and cyclosporin A, suggesting that the IAM model may not be completely representative of binding to Pgp in its native membrane environment, and that alternative methods of immobilization should be evaluated. Affinity chromatography can be used to evaluate reciprocal displacement by different substrates, and so also to evaluate whether two substrates bind to the same or to different sites. The results of our evaluations so far provide suggestions about the nature of interactions between several substrates, in the absence of added ATP. Verapamil binds to multiple sites on Pgp, at least one of which is shared with vinblastine. Cyclosporin A does not bind to Pgp, in the absence of ATP and/or other substrates, but its binding is elicited by vinblastine. Cyclosporin A, however, displaces vinblastine, suggesting an interaction which is neither comeotitve nor allosteric as it is not reciprocal. (+)Mefloquine and (-)mefloquine interact with cyclosprin A and vinblastine in a manner which is enantio-selective for the former, but not for the latter. More extensive evaluations of substrate displacement by frontal chromatography and by non-linear zonal chromatography analysis will allow to complete a functional definition of binding sites by identifying which substrates share the same sites. It has to be noted that it is possible that these evaluations are not optimal in the absence of ATP, as the conformation of binding sites may be quite different when ATP is bound to Pgp and before ATP hydrolysis. The latter conformation may be more relevant to the definition of effective pharmacophores. We have observed that addition of ATP changes the affinity for its substrates: affinity for vinblastine and verapamil is decreased, while that for cyclosporin A is increased. Binding to Pgp in the presence of ATP is bound to represent the "average" of Pgp's affinities for its substrates as it goes through the different steps of the transport cycle, including the steps that immediately precede and follow the catalysis of ATP to ADP and inorganic phosphate. It is to be expected that affinity for effective substrates will be decreased in the post-catalytic stage, when the substrate has to be released to the extracellular side of the membrane. As in the case of filtration binding assays, affinity chromatography can be used to model the different stages of Pgp's transport cycle. For example, use of non-hydrolysable ATP analogs will allow to model Pgp in the "pre-catalytic" stage, and pre-treatment of the Pgp column with vanadate and ATP in the presence of a substrate will "freeze" it in the catalytic stage. In preliminary studies, we have used zonal affinity chromatography to evaluate vinblastine binding to a "vanadate-trapped" Pgp-column. The results have confirmed inhibition of vinblastine binding to post-catalytic Pgp. These evaluations will be extended to different substrates and to different stages of the transport cycle. In summary, we have obtained and are characterizing an affinity chromatography model for the functional evaluation of Pgp binding sites and transport cycle. We will try to further optimize our affinity chromatography models and we may, for example, evaluate chromatographic columns where different approaches are used to immobilize Pgp in its native membrane. We plan then to proceed to full functional characterization of Pgp's binding sites by identifying those substrates that share the same site, at different stages of the transport cycle. Structure-affinity relationships of site-sharing substrates will be used to define pharmacophores. Other ABC and non-ABC transporters (such as MRP1 and OCT1) will be modeled and comparison of results will be used to define transporter-selective pharmacophores.