- Email: [email protected]
Virtual screening and experimental validation identify novel modulators of nuclear receptor RXRα from Drugbank database
Accepted Manuscript Virtual Screening and Experimental Validation Identify Novel Modulators of Nuclear Receptor RXRα from Drugbank database Dan Xu, Lijun Cai, Shangjie Guo, Lei Xie, Meimei Yin, Ziwen Chen, Hu Zhou, Ying Su, Zhiping Zeng, Xiaokun Zhang PII: DOI: Reference:
S0960-894X(16)31337-3 http://dx.doi.org/10.1016/j.bmcl.2016.12.058 BMCL 24547
To appear in:
Bioorganic & Medicinal Chemistry Letters
Received Date: Revised Date: Accepted Date:
1 September 2016 20 December 2016 21 December 2016
Please cite this article as: Xu, D., Cai, L., Guo, S., Xie, L., Yin, M., Chen, Z., Zhou, H., Su, Y., Zeng, Z., Zhang, X., Virtual Screening and Experimental Validation Identify Novel Modulators of Nuclear Receptor RXRα from Drugbank database, Bioorganic & Medicinal Chemistry Letters (2016), doi: http://dx.doi.org/10.1016/j.bmcl. 2016.12.058
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Graphical Abstract To create your abstract, type over the instructions in the template box below. Fonts or abstract dimensions should not be changed or altered.
Virtual Screening and Experimental Validation Identify Novel Modulators of Nuclear Receptor RXRα from Drugbank database
Leave this area blank for abstract info.
Dan Xu, Lijun Cai, Shangjie Guo, Lei Xie, Meimei Yin, Ziwen Chen, Hu Zhou, Ying Su, Zhiping Zeng and Xiaokun Zhang.
Bioorganic & Medicinal Chemistry Letters
Virtual Screening and Experimental Validation Identify Novel Modulators of Nuclear Receptor RXRα from Drugbank database §a Dan Xu , Lijun Cai§a, Shangjie Guoa, Lei Xiea, Meimei Yina, Ziwen Chena, Hu Zhoua, Ying Sua,b, Zhiping a a,b Zeng* and Xiaokun Zhang* a
School of Pharmaceutical Sciences, Fujian Provincial Key Laboratory of Innovative Drug Target Research, Xiamen University, Xiamen 361005, China; Cancer center, Sanford Burnham Prebys Medical Discovery Institute, La Jolla, CA, USA; § These authors contributed equally to this work. b
ARTICLE INFO
ABSTRACT
Article history: Received Revised Accepted Available online
Retinoid X receptor alpha (RXRα), an important ligand-dependent transcription factor, plays a critical role in the development of various cancers and metabolic and neurodegenerative diseases. Therefore, RXRα represents one of the most important targets in modern drug discovery. In this study, Drugbank 2.0 with 1280 old drugs were virtually screened by Glide according to the crystal structure of ligand-binding domain (LBP) of RXRα. 15 compounds selected were tested for their binding and transcriptional activity toward RXRα by Biacore and reporter gene assay, respectively. The identified new scafford ligand of RXRα, Pitavastatin (1), was chemically optimized. Our results demonstrated that statin compounds Pitavastatin (1) and Fluvastatin (4) could bind to the LBP of RXRα (KD = 13.30 μM and 11.04 μM, respectively) and serve as transcriptional antagonists of RXRα. On the contrary, compound (12) (domperidone) and (13) (rosiglitazone maleate) could bind to the LBP of RXRα (K D = 8.80 μM and 15.01 μM, respectively) but serve as transcriptional agonists of RXRα.
Keywords: RXRα Old drug agonist antagonist virtual screening
2016 Elsevier Ltd. All rights reserved.
Retinoid X receptor (RXR, NR2B1-3), a ligand-dependent transcriptional factor, is a unique member of the nuclear receptor superfamily1. So far, three subtypes of RXR have been found: RXRα, RXRβ and RXRγ2. RXRα is ubiquitously expressed in different human tissues with high levels of expression in liver, lung, muscle, kidney, intestinal tract and epithelial tissues. Because of this, RXRα plays an important role in many important physiological processes such as vertebrate development, various aspects of metabolism, cell differentiation, cell death and homeostatis3. Abnormal expression or activity of RXRα in human is implicated in the development of dyslipidemia4, atherosclerosis5, type 2 diabetes6, autoimmune7, cancer8, Alzheimer’s disease9 and allergy. Therefore, RXRα has become a crucial drug target in modern drug discovery, highlighted by the approval of Targretin/Bexarotene, a RXR synthetic ligand, for treating acute T-cell lymphoma10. RXRα is made of several functional domains. The sequence from 1 to 134 is called as A/B domain (modulating N-terminal domain or AF-1 domain)11. Amino acids encoding from 135 to 200 is named as DNA-binding domain (DBD). This domain contains two zinc finger motifs (135-155, 171-195), which could bind to DNA response elements12. Amino acids encoding from 201 to 224 is hinge region, which is primarily responsible for *corresponding authors. Email: [email protected] (Z. P. Zeng) and [email protected] (X. K. Zhang)
linking the DNA-binding domain (DBD) to the ligand-binding domain (LBD), whose encoding amino acids cover from 225 to 46213. Until now, the crystal structure of full-length RXRα is not reported, which could be attributed to the flexibility of the Nterminus and the hinge domain (201-224). But the threedimensional structures of LBD and DBD have been well defined13. RXRα exerts its transcriptional activities in the nucleus by forming homodimer (RXRα:RXRα)14 with itself or heterodimerizing with many other nuclear receptors such as RAR, LXR, PPARγ, VDR15, which is one of the reasons that RXRα participates in the development of various cancers and diseases. RXRα could also form homotetramer16 whose biological significance remains to be established. In addition to its transactivation activity in the nucleus, RXRα can act biologically in the cytoplasm. We recently reported that a truncated form of RXRα (tRXRα) resides in the cytoplasm where it interacts with the p85α subunit of phosphatidylinositol-3-OH kinase (PI3K) leading to activation of PI3K/AKT survival signaling in cancer cells17-19. Significantly, the tumor promoting effect of tRXRα could be suppressed by certain RXRα ligands such as Sulindac and its analogues, resulting in apoptosis of cancer cells17-19. Though many ligands of RXRα have been reported, only three of them were approved by the FDA for clinical applications: Alitretinoin (9-cis-RA)14, 20-22, Acitretin23, 24 and Targretin25, 26
Figure 1. Structure and report gene assay of 15 commercially available compounds. (A)The chemical structures of 15 commercially available compounds. (B)Identification of RXRα-selective agonist and antagonist. Antagonist effect of compound 1 and compound 4, and agonist effect of compound 12 and compound 13. MCF-7 cells cotransfected with the reporter plasmids pCMV-Myc-RXRα, pGL6-RXRE-Luc and pC DNA 3.1-Rellina were treated with 9-cis-RA (10−7 M) alone or together with drug for 24 h. Reporter activities were measured as described above. Data shown are mean ± SD.
(LGD1069, Bexarotene). Unfortunately, all of them have severe side effects when high dose of these drugs is administered or these drugs are taken for a long time against chronic diseases. Thus identifying new chemical entities possessing high efficacy but low toxicity for RXRα is a challenge for modern nuclear receptor drug discovery27. Owning to the favorable profiles of pharmacokinetic and safety properties of approved drugs, repositioning existing drugs for new indications has been an effective drug discovery and development strategy28. One of the approaches under this strategy is to exploit new targets by structure-based virtual screening of existing drugs29. In our previous work, we found that one old drug, Sulindac, could bind to RXRα17. Further optimization led to K-8000325, K-8008 and K-801218 , which show improved binding selectivity and biological activity. However, whether some other old drugs could target RXRα is still unknown. In this paper, FDA-approved drugs downloaded from Drugbank30 were used as the old drug database and virtual screening of this database toward the crystal structure of LBP of RXRα was carried out. Drugs of high-score were selected for further binding and biological evaluation. Some hits were identified and their interactions with RXRα were studied theoretically and experimentally. The big difference between agonistic and antagonistic conformation of RXRα is the location of H12 in its ligand-
Figure 2. Reporter gene assay of Pitavastatin calcium (1), Fluvastatin (4), Domperidone (12) and Rosiglitazone maleate (13). MCF-7 cells cotransfected with the reporter plasmids pCMV-Myc-RXRα, pGL6-RXRELuc and pC DNA 3.1-Rellina were treated with 9-cis-RA (10−7 M) alone or together with drug for 24 h (The concentration of 9-cis-RA is 10−6 M in A and 10−7 M in B-D). (A)Concentration dependence effect of Pitavastatin calcium (1). (B)Concentration dependence effect of Fluvastatin (4). (C)Concentration dependence effect of Domperidone (12). (D)Concentration dependence effect of Rosiglitazone maleate (13). (E)Reporter of Pitavastatin calcium (1), Fluvastatin (4), Domperidone (12) and Rosiglitazone maleate (13) on Nur77 transactivation. MCF-7 cells cotransfected with pG5-Luc and pBind-Nur77-LBD. (E) were treated with compound (10 μM) for 24 h. Reporter activities were measured as described above. Data shown are mean ± SD. (F)Effect of Pitavastatin calcium (1), Fluvastatin (4), Domperidone (12) and Rosiglitazone maleate (13) on RARγ transactivation. MCF-7 cells cotransfected with pG5-Luc and pBind-RARγLBD (F) were treated with compound (10 μM) in the presence or absence of the RARγ agonist AHPN (1 μM) for 24 h. Reporter activities were measured as described above. Data shown are mean ± SD.
binding domain. The transition of agonist to antagonist lead to mobility of H12 from the surface to the other end, indicating that the binding site of antagonistic conformation may be bigger than that of agonist. It is also important to note that Leu436 of H12, which is responsible for the stabilization of interaction with holoH12, will normally produce a steric hindrance effect with antagonist, providing the ration basis for the modulator design of RXRα. Here, in order to cover both the binding pocket of modulators, 3343 old drug conformations produced from 2D structure of 1280 compounds were docked to the agonist-bound and antagonist-bound crystal structures of RXRα-LBD, respectively. The top 24 compounds possessing the highest docking scores were summarized in Table S1 (see supporting information). Interestingly, statin drugs, the type of cholesterol synthesis inhibitors31, could bind to the LBD of RXRα with high score (1, 4, 5, 7). The docking energy of Pitavastatin calcium (1) binding to agonist conformation of LBD of RXRα（PDBID： 1FBY ） is -12.171 kcal mol-1, ranking it the highest among compounds screened. And also its binding energy towards antagonist conformation of LBD of RXRα（PDBID：3A9E）is up to -10.810 kcal mol-1, ranking it sixth in the antagonism
Figure 4. Binding mode of Pitavastatin calcium (1), Fluvastatin (4), Domperidone (12) and Rosiglitazone maleate (13) to RXRα-LBD. Hydrogen bond was labeled as a dotted line. (A)Comparison of Pitavastatin calcium (1, yellow) and LG100754 (green) binding to RXRα. (B)Comparison of Fluvastatin (4, yellow) and LG100754 (green) binding to RXRα. (C)Comparison of Domperidone (12, yellow) and 9-cis-RA (green) binding to RXRα. (D)Comparison of Rosiglitazone maleate (13, yellow) and 9-cis-RA (green) binding to RXRα. Figure 3. Surface plasmon resonance of Pitavastatin calcium (1), Fluvastatin (4), Domperidone (12) and Rosiglitazone maleate (13). Gradient concentrations of each compound were injected through flow cells immobilized with purified RXRα-LBD protein. The chip is being exposed to ligand solution during 0-120 second, and the ligand is dissociated from the chip by running buffer from 120 to 200 second.The kinetic profiles are shown. (A) Sensorgram of Pitavastatin calcium (1) binding to RXRα-LBD. (B) Sensorgram of Fluvastatin (4) binding to RXRα-LBD. (C) Sensorgram of Domperidone (12) binding to RXRαLBD. (D) Sensorgram of Rosiglitazone maleate (13) binding to RXRαLBD. (E) Sensorgram of Compound (25) binding to RXRα-LBD. (F) Sensorgram of Compound (27) binding to RXRα-LBD.
screening. Similarly, other statin drugs such as Fluvastatin (4, 10.751 kcal mol-1 for RXRα agonism and -9.75 kcal mol-1 for RXRα antagonism), Rosuvastatin Calcium (5, -10.722 kcal mol-1 for agonism and -11.189 kcal mol-1 for antagonism), and vastatin Sodium (7, -10.497 kcal mol-1 for agonism and -10.3 kcal mol-1 for antagonism) have good scoring and ranking toward both of agonist and antagonist conformations of RXRα. Like classical ligands of RXRα, these four statin drugs have carboxyl acid group, which could interact with Arg316 in RXRα and the topology of statins is also L-shape, which could fit the ligandbinding pocket of RXRα. Taken together, the docking results indicated that statin drugs might bind to RXRα and could be explored as new RXRα modulators. Another worth-noting result is that Rosiglitazone maleate (13), which was shown to act as the ligand of nuclear receptor PPARγ to regulate blood sugar level32, has a moderate docking energy toward RXRα (-9.737 kcal mol-1 for agonism and -9.024 kcal mol-1 for antagonism). This maybe due to the similarity of the ligand-binding pockets in RXRα and PPARγ. Chemical analysis indicated that Rosiglitazone maleate (13) lacks carboxyl acid group, which is quite different from the classic RXRα ligands. Therefore, whether and how Rosiglitazone maleate (13) binds to RXRα remains to be determined. Among the interested 24 drugs, only 15 of them (1, 3, 4, 5, 6, 7, 8, 9, 11, 12, 13, 15, 16, 19, 24) were commercially available for further testing (see Fig. 1A). All of these compounds may be hydrophobic according to
their chemical structure, and this is coincided with the hydrophobic pocket of LBD of RXRα. In order to confirm our docking results, reporter gene assay of transcriptional activity of RXRα in MCF-7 cells in the absence or presence of 10-7 M 9-cis-RA was used to evaluate the agonist and antagonist effect of the 15 commercially-available compounds (Fig. 1A). When tested at 10 μM concentration, Pitavastatin calcium (1) and Fluvastatin (4) showed good antagonist effect on 9-cis-RA-induced RXRα transactivation (Fig. 1B). Rosuvastatin Calcium (5) also slightly antagonized the effect of 9-cis-RA, while Pravastatin Sodium (7) had no effect. In addition to the statins, Ezetimibe (3), Seratrodast (8), Cilostazol (16), and Butoconazole (19) exhibited antagonist effect on 9-cisRA-induced reporter activity. The order of their antagonism is Cilostazol (16) > Ezetimibe (3) > Butoconazole (19) > Seratrodast (8). Our data also revealed that two old drugs, Domperidone (12) and Rosiglitazone maleate (13), could enhance the effect of 9-cis-RA on inducing RXRα transactivation. Such an agonist effect occurred only in the presence of 9-cis-RA but not its absence. Four drugs with strong effect on RXRα transactivation, Pitavastatin calcium (1), Fluvastatin (4), Domperidone (12) and Rosiglitazone maleate (13), were further evaluated dose-dependently (1 μM, 5 μM, 10 μM) for their effect on RXRα transactivation by reporter assay in MCF-7 cells (Fig. 2). Our results showed that Pitavastatin calcium (1) at 1 μM concentration had better antagonistic effect than Fluvastatin (4). However, Fluvastatin (4) exhibited very nice dose dependent antagonist effect (Fig. 2B). At 10 μM, it almost completed suppressed the effect of 9-cis-RA on inducing RXRα reporter activity, while the antagonist effect of Pitavastatin calcium (1) appeared to reach plateaus at 1 μM concentration (Fig. 2A). When Domperidone (12) and Rosiglitazone maleate (13) were analyzed (Figs. 2C and D), they did not show any effect at concentration of 1 μM. At higher concentrations (5 and 10 μM), both compounds enhanced the effect of 9-cis-RA. Pitavastatin calcium (1), Fluvastatin (4), Domperidone (12) and
Rosiglitazone maleate(13) were also evaluated for their effect on tranactivation
Figure 5. Pitavastatin calcium (1) derivatives and their activity study. (A)Synthesis of Pitavastatin calcium (1) derivatives. (B)Antagonist effect of Pitavastatin calcium (1) and its analogs (25, 26, 27). MCF-7 cells cotransfected with the reporter plasmids pCMV-Myc-RXRα, pGL6RXRE-Luc and pCDNA 3.1-Rellina were treated with 9-cis-RA (10−7 M) alone or together with drug for 24 h. Reporter activities were measured as described above. Data shown are mean ± SD.
of Nur77 and RARγ by reporter assays. Our results showed that none of these compounds showed any apparent agonist or antagonist effect on constitutive Nur77 activity or RARγ activity induced by its agonist AHPN33 (see Fig. 2E and 2F). Thus, the agonist or antagonist effect we observed for these compounds is selective to RXRα and is unlikely due to their nonspecific effect on general transcriptional machinery. To determine whether the effect of Pitavastatin calcium (1), Fluvastatin (4), Domperidone (12) and Rosiglitazone maleate (13) on RXRα transactivation was due to their binding to the receptor, Surface Plasmon Resonance experiment was conducted (Fig. 3), and our results demonstrated that all of the drugs (1, 4, 12, 13) could bind to the purified RXRα-LBD protein. The binding kinetics of most of our compounds are a fast association and a fast dissociation process, making it difficult to determine its Ka and Kd. Thus we could only derivate each affinity constant by direct analysis of steady-state binding data. Until now, it is reported that protein ligand binding site of RXRα is one. Moreover, in the Biacore T200 software, analysis of steady-state binding data by fitting the curve of binding level against concentrations can only be applied to 1:1 Langmuir model, so 1:1 complex model could be reasonable for the fitting of experimental data. But it does not rule out other binding possibilities. The KD for the Pitavastatin calcium (1) binding to the LBD of RXRα is 13.30 μM. The binding kinetics of Fluvastatin (4) is quite similar to Pitavastatin calcium (1), but its binding affinity (KD = 11.04 μM) is substantially stronger than Pitavastatin calcium (1), which is opposite to the reporter assay.
Domperidone (12), as the agonist of RXRα, has the best binding affinity (KD = 8.80 μM) as comparing the other three compounds. The binding kinetics of the compound is similar to Pitavastatin calcium (1) and Fluvastatin (4), but is slightly different from Rosiglitazone maleate (13). It is interesting that the binding model of Rosiglitazone maleate (13) showed a little slow association and fast dissociation process, and its binding affinity is moderate (KD = 15.01 μM). In order to further understand how Pitavastatin calcium (1), Fluvastatin (4), Domperidone (12) and Rosiglitazone maleate (13) bind to RXRα-LBD, their docking results of each complex were performed by 10 ns molecular dynamics and their binding results were analyzed in details (Fig. 4). First, two antagonists, Pitavastatin calcium (1) and Fluvastatin (4), were docked into the antagonist conformation of RXRα-LBD (PDB ID: 3a9e). LG100754-bound structure was set as the reference for comparison (see Fig. 4A and 4B). The MD results showed that Pitavastatin calcium (1) and Fluvastatin (4) bind to a region that overlap with LG100754 binding region which is located in the Lshape of the ligand-binding pocket of RXRα. The fluorobenzene moiety of Pitavastatin calcium (1) and Fluvastatin (4) was located in the same position as the propoxyl moiety of LG100754, which could interact with the Leu436 from H12 by Van Der Waal force, pushing the conformation of RXRα from agonism to antagonism. This explains why Pitavastatin calcium (1) and Fluvastatin (4) displayed antagonistic effect in the transcriptional activity assay. The carboxyl group of Pitavastatin calcium (1) and Fluvastatin (4) were found to form ionic interaction with Arg316 and hydrogen bond with amide group of Ala327 and thus acting as anchor as the carboxyl group of classical RXRα ligands. It is noteworthy that two hydroxyl groups in the chain of Pitavastatin calcium (1) and Fluvastatin (4) do not make positive contribution to their binding in the environment of the hydrophobic pocket, implying that elimination of these polar groups would enhance their binding affinity. In a similar manner, Domperidone (12) and Rosiglitazone maleate (13) were docked into the agonist conformation of RXRα-LBD (PDB ID: 1fby). 9-cis-RA was set as the reference compound for comparison (Fig. 4C and 4D). The results showed that Rosiglitazone maleate (13) could overlap with 9-cis-RA better than Domperidone (12). Unlike 9-cis-RA, both of them lack carboxyl acid group. Nevertheless, the 5-(4-(2(methyl(pyridin-2-yl)amino)ethoxy)benzyl)thiazolidine-2,4dione moiety of Rosiglitazone maleate (13) could place in the position of carboxyl acid group of 9-cis-RA, which could also form hydrogen bond with Arg316 and Ala327, but Domperidone (12) lost this interaction completely. Maybe the hydrophobic interaction compensates and predominates the binding affinity of Domperidone (12) because of its double heterocyclic ring distributes in the hydrophobic pocket of ligand binding site. Taken together, all of the four drugs could bind to the LBP of RXRα in the same fashion as the classical RXRα ligands. According to the theoretical binding study of Pitavastatin calcium (1) to the LBP of RXRα, three derivatives of Pitavastatin calcium (1) were designed and synthesized (see Fig. 5A). Elimination of two hydroxyl groups was conducted by H2SO4/CH3COOH at room temperature. Unexpectedly, a six-ring lactone was produced (25) and only one hydroxyl group was removed in the first step. When this compound was treated with TBAF in THF solution at room temperature, the six-ring lactone was hydrolysis and another hydroxyl group was eliminated. Our target molecule (27) was obtained in the end. Secondly, in order to evaluate the importance of the carboxyl acid group in Pitavastatin calcium (1), this function group was esterificated (26) using H2SO4/C2H5OH as the reagent (see Fig. 5A).
Reporter gene assay and SPR study were used to study the structure-activity relationship of Pitavastatin calcium (1) and its derivatives (25, 26, 27) (see Fig. 5B, 3E, 3F). The results illustrated that block of carboxyl acid group would decrease the antagonistic effect on 9-cis-RA-induced RXRα reporter activity in MCF-7 cells and also lost its binding affinity to 138.3 μM (Fig. 3E). Elimination of two hydroxyl group from Pitavastatin calcium (1) would slightly enhance its antagonistic effect, and it is worth noting that compound (27) at 100 nM could effectively inhibit 9-cis-RA activity in a degree that was similar to the effect of 10 μM and also increase its binding affinity to 5.12 μM (Fig. 3F). Taken together, this SAR study demonstrated that carboxylic acid group is critical for the drugs’ biological activity and the two hydroxyl groups were not, which is consistent with the binding model of Pitavastatin calcium (1) to RXRα (see Fig. 4A). In conclusion, RXRα-based virtual screening of a drug database from Drugbank was conducted and 15 commercially available compounds were obtained for biological testing and binding studies. The results show that two statin drugs, Pitavastatin calcium (1) and Fluvastatin (4), could bind to RXRα and have potential antagonistic effect on RXRα transactivation in MCF-7 cells. Domperidone (12) and Rosiglitazone maleate (13) could also bind to RXRα but enhanced the effect on 9-cis-RA on activating RXRα. Modeling and SAR study concluded that it is the carboxylic acid group and pendant hydrophobic interactions predominate the binding of Pitavastatin calcium (1) to RXRα. Acknowledgements This work was supported by the grants from the National Nature Science Fund of China (NSFC-81301888, NSFC31271453, NSFC-31471318, NSFC-91429306, U1405229, and NSFC-91129302), the Fundamental Research Funds for the Central Universities (2013121038). References and notes 1. 2. 3. 4. 5.
6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.
18.
Evans, R. M.; Mangelsdorf, D. J. Cell 2014, 157, 255. Mangelsdorf, D. J.; Borgmeyer, U.; Heyman, R. A.; Zhou, J. Y.; Ong, E. S.; Oro, A. E.; Kakizuka, A.; Evans, R. M. Genes & Dev. 1992, 6, 329. Szanto, A.; Narkar, V.; Shen, Q.; Uray, I. P.; Davies, P. J. A.; Nagy, L. Cell Death Differ. 2004, 11, S126. Lagu, B.; Lebedev, R.; Pio, B.; Yang, M.; Pelton, P. D. Bioorg. Med. Chem. Lett. 2007, 17, 3491. Claudel, T.; Leibowitz, M. D.; Fievet, C.; Tailleux, A.; Wagner, B.; Repa, J. J.; Torpier, G.; Lobaccaro, J. M.; Paterniti, J. R.; Mangelsdorf, D. J.; Heyman, R. A.; Auwerx, J. Proc. Natl. Acad. Sci. U. S. A. 2001, 98, 2610. Kakuta, H. Yakugaku Z. J. Pharm. Soc. Jpn. 2011, 131, 917. Diab, A.; Hussain, R. Z.; Lovett-Racke, A. E.; Chavis, J. A.; Drew, P. D.; Racke, M. K. J. Neuroimmunol. 2004, 148, 116-126. Zhang, X. K.; Zhou, H.; Su, Y. Acta Biochim. Biophys. Sin. 2016, 48, 49. Crunkhorn, S. Nat. Rev. Drug Discov. 2012, 11, 271. Ottow, E.; Weinmann, H.: Nuclear receptors as drug targets; WileyVCH, 2008. Mascrez, B.; Mark, M.; Krezel, W.; Dupe, V.; LeMeur, M.; Ghyselinck, N. B.; Chambon, P. Development 2001, 128, 2049. Holmbeck, S. M. A.; Foster, M. P.; Casimiro, D. R.; Sem, D. S.; Dyson, H. J.; Wright, P. E. J. Mol. Biol. 1998, 281, 271. Bourguet, W.; Ruff, M.; Chambon, P.; Gronemeyer, H.; Moras, D. Nature 1995, 375, 377. Zhang, X. K.; Lehmann, J.; Hoffmann, B.; Dawson, M. I.; Cameron, J.; Graupner, G.; Hermann, T.; Tran, P.; Pfahl, M. Nature 1992, 358, 587. Mangelsdorf, D. J.; Evans, R. M. Cell 1995, 83, 841. Chen, Z. P.; Iyer, J.; Bourguet, W.; Held, P.; Mioskowski, C.; Lebeau, L.; Noy, N.; Chambon, P. J. Mol. Biol. 1998, 275, 55. Zhou, H.; Liu, W.; Su, Y.; Wei, Z.; Liu, J.; Kolluri, S. K.; Wu, H.; Cao, Y.; Chen, J.; Wu, Y.; Yan, T.; Cao, X.; Gao, W.; Molotkov, A.; Jiang, F.; Li, W.-G.; Lin, B.; Zhang, H.-P.; Yu, J.; Luo, S.-P.; Zeng, J.-Z.; Duester, G.; Huang, P.-Q.; Zhang, X.-K. Cancer Cell 2010, 17, 560. Chen, L.; Wang, Z. G.; Aleshin, A. E.; Chen, F.; Chen, J.; Jiang, F.; Alitongbieke, G.; Zeng, Z.; Ma, Y.; Huang, M.; Zhou, H.; Cadwell, G.;
19. 20. 21.
22. 23. 24.
25. 26. 27.
28. 29. 30. 31. 32. 33. 34.
Zheng, J.-F.; Huang, P.-Q.; Liddington, R. C.; Zhang, X. K. Su, Y. Chem. & Biol. 2014, 21, 596. Zhang, X. K.; Zhou, H.; Su, Y. Acta Biochim. Biophys. Sin. 2016, 48, 49. Heyman, R. A.; Mangelsdorf, D. J.; Dyck, J. A.; Stein, R. B.; Eichele, G.; Evans, R. M.; Thaller, C. Cell 1992, 68, 397. Levin, A. A.; Sturzenbecker, L. J.; Kazmer, S.; Bosakowski, T.; Huselton, C.; Allenby, G.; Speck, J.; Kratzeisen, C.; Rosenberger, M.; Lovey, A.; Grippo, J. F. Nature 1992, 355, 359. Egea, P. F.; Mitschler, A.; Rochel, N.; Ruff, M.; Chambon, P.; Moras, D. EMBO Journal 2000, 19, 2592. Lowe, N. J.; Prystowsky, J. H.; Bourget, T.; Edelstein, J.; Nychay, S.; Armstrong, R. J. Am. Acad. Dermatol. 1991, 24, 591. Bavinck, J. N. B.; Tieben, L. M.; Vanderwoude, F. J.; Tegzess, A. M.; Hermans, J.; Terschegget, J.; Vermeer, B. J. J. Clin. Oncol. 1995, 13, 1933. Gottardis, M. M.; Bischoff, E. D.; Shirley, M. A.; Wagoner, M. A.; Lamph, W. W.; Heyman, R. A. Cancer Res. 1996, 56, 5566. Tooker, P.; Yen, W.-C.; Ng, S.-C.; Negro-Vilar, A.; Hermann, T. W. Cancer Res. 2007, 67, 4425. Assaf, C.; Bagot, M.; Dummer, R.; Duvic, M.; Gniadecki, R.; Knobler, R.; Ranki, A.; Schwandt, P.; Whittaker, S. Br. J. Dermatol. 2006, 155, 261. Ashburn, T. T.; Thor, K. B. Nat. Rev. Drug Discov. 2004, 3, 673. Ma, D. L.; Chan, D. S. H.; Leung, C. H. Chem. Soc. Rev. 2013, 42, 2130. Wishart, D. S.; Knox, C.; Guo, A. C.; Cheng, D.; Shrivastava, S.; Tzur, D.; Gautam, B.; Hassanali, M. Nucleic Acids Res. 2008, 36, D901. Hebert, P. R.; Gaziano, J. M.; Chan, K. S.; Hennekens, C. H. Jama-J. Am. Med. Assoc. 1997, 278, 313. Gerstein, H. C.; Yusuf, S.; Holman, R. R.; Bosch, J.; Anand, S.; Avezum, A.; et. al. Lancet 2006, 368, 1096. Holmes, W. F.; Dawson, M. I.; Soprano, D. R.; Soprano, K. J. J. Cell Physiol. 2000, 185, 61. Sato, Y.; Ramalanjaona, N.; Huet, T.; Potier, N.; Osz, J.; Antony, P.; Peluso-Iltis, C.; Poussin-Courmontagne, P.; Ennifar, E.; Mely, Y.; Dejaegere, A.; Moras, D.; Rochel, N. Plos One 2010, 5, e15119.
S0960-894X(16)31337-3 http://dx.doi.org/10.1016/j.bmcl.2016.12.058 BMCL 24547
To appear in:
Bioorganic & Medicinal Chemistry Letters
Received Date: Revised Date: Accepted Date:
1 September 2016 20 December 2016 21 December 2016
Please cite this article as: Xu, D., Cai, L., Guo, S., Xie, L., Yin, M., Chen, Z., Zhou, H., Su, Y., Zeng, Z., Zhang, X., Virtual Screening and Experimental Validation Identify Novel Modulators of Nuclear Receptor RXRα from Drugbank database, Bioorganic & Medicinal Chemistry Letters (2016), doi: http://dx.doi.org/10.1016/j.bmcl. 2016.12.058
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Graphical Abstract To create your abstract, type over the instructions in the template box below. Fonts or abstract dimensions should not be changed or altered.
Virtual Screening and Experimental Validation Identify Novel Modulators of Nuclear Receptor RXRα from Drugbank database
Leave this area blank for abstract info.
Dan Xu, Lijun Cai, Shangjie Guo, Lei Xie, Meimei Yin, Ziwen Chen, Hu Zhou, Ying Su, Zhiping Zeng and Xiaokun Zhang.
Bioorganic & Medicinal Chemistry Letters
Virtual Screening and Experimental Validation Identify Novel Modulators of Nuclear Receptor RXRα from Drugbank database §a Dan Xu , Lijun Cai§a, Shangjie Guoa, Lei Xiea, Meimei Yina, Ziwen Chena, Hu Zhoua, Ying Sua,b, Zhiping a a,b Zeng* and Xiaokun Zhang* a
School of Pharmaceutical Sciences, Fujian Provincial Key Laboratory of Innovative Drug Target Research, Xiamen University, Xiamen 361005, China; Cancer center, Sanford Burnham Prebys Medical Discovery Institute, La Jolla, CA, USA; § These authors contributed equally to this work. b
ARTICLE INFO
ABSTRACT
Article history: Received Revised Accepted Available online
Retinoid X receptor alpha (RXRα), an important ligand-dependent transcription factor, plays a critical role in the development of various cancers and metabolic and neurodegenerative diseases. Therefore, RXRα represents one of the most important targets in modern drug discovery. In this study, Drugbank 2.0 with 1280 old drugs were virtually screened by Glide according to the crystal structure of ligand-binding domain (LBP) of RXRα. 15 compounds selected were tested for their binding and transcriptional activity toward RXRα by Biacore and reporter gene assay, respectively. The identified new scafford ligand of RXRα, Pitavastatin (1), was chemically optimized. Our results demonstrated that statin compounds Pitavastatin (1) and Fluvastatin (4) could bind to the LBP of RXRα (KD = 13.30 μM and 11.04 μM, respectively) and serve as transcriptional antagonists of RXRα. On the contrary, compound (12) (domperidone) and (13) (rosiglitazone maleate) could bind to the LBP of RXRα (K D = 8.80 μM and 15.01 μM, respectively) but serve as transcriptional agonists of RXRα.
Keywords: RXRα Old drug agonist antagonist virtual screening
2016 Elsevier Ltd. All rights reserved.
Retinoid X receptor (RXR, NR2B1-3), a ligand-dependent transcriptional factor, is a unique member of the nuclear receptor superfamily1. So far, three subtypes of RXR have been found: RXRα, RXRβ and RXRγ2. RXRα is ubiquitously expressed in different human tissues with high levels of expression in liver, lung, muscle, kidney, intestinal tract and epithelial tissues. Because of this, RXRα plays an important role in many important physiological processes such as vertebrate development, various aspects of metabolism, cell differentiation, cell death and homeostatis3. Abnormal expression or activity of RXRα in human is implicated in the development of dyslipidemia4, atherosclerosis5, type 2 diabetes6, autoimmune7, cancer8, Alzheimer’s disease9 and allergy. Therefore, RXRα has become a crucial drug target in modern drug discovery, highlighted by the approval of Targretin/Bexarotene, a RXR synthetic ligand, for treating acute T-cell lymphoma10. RXRα is made of several functional domains. The sequence from 1 to 134 is called as A/B domain (modulating N-terminal domain or AF-1 domain)11. Amino acids encoding from 135 to 200 is named as DNA-binding domain (DBD). This domain contains two zinc finger motifs (135-155, 171-195), which could bind to DNA response elements12. Amino acids encoding from 201 to 224 is hinge region, which is primarily responsible for *corresponding authors. Email: [email protected] (Z. P. Zeng) and [email protected] (X. K. Zhang)
linking the DNA-binding domain (DBD) to the ligand-binding domain (LBD), whose encoding amino acids cover from 225 to 46213. Until now, the crystal structure of full-length RXRα is not reported, which could be attributed to the flexibility of the Nterminus and the hinge domain (201-224). But the threedimensional structures of LBD and DBD have been well defined13. RXRα exerts its transcriptional activities in the nucleus by forming homodimer (RXRα:RXRα)14 with itself or heterodimerizing with many other nuclear receptors such as RAR, LXR, PPARγ, VDR15, which is one of the reasons that RXRα participates in the development of various cancers and diseases. RXRα could also form homotetramer16 whose biological significance remains to be established. In addition to its transactivation activity in the nucleus, RXRα can act biologically in the cytoplasm. We recently reported that a truncated form of RXRα (tRXRα) resides in the cytoplasm where it interacts with the p85α subunit of phosphatidylinositol-3-OH kinase (PI3K) leading to activation of PI3K/AKT survival signaling in cancer cells17-19. Significantly, the tumor promoting effect of tRXRα could be suppressed by certain RXRα ligands such as Sulindac and its analogues, resulting in apoptosis of cancer cells17-19. Though many ligands of RXRα have been reported, only three of them were approved by the FDA for clinical applications: Alitretinoin (9-cis-RA)14, 20-22, Acitretin23, 24 and Targretin25, 26
Figure 1. Structure and report gene assay of 15 commercially available compounds. (A)The chemical structures of 15 commercially available compounds. (B)Identification of RXRα-selective agonist and antagonist. Antagonist effect of compound 1 and compound 4, and agonist effect of compound 12 and compound 13. MCF-7 cells cotransfected with the reporter plasmids pCMV-Myc-RXRα, pGL6-RXRE-Luc and pC DNA 3.1-Rellina were treated with 9-cis-RA (10−7 M) alone or together with drug for 24 h. Reporter activities were measured as described above. Data shown are mean ± SD.
(LGD1069, Bexarotene). Unfortunately, all of them have severe side effects when high dose of these drugs is administered or these drugs are taken for a long time against chronic diseases. Thus identifying new chemical entities possessing high efficacy but low toxicity for RXRα is a challenge for modern nuclear receptor drug discovery27. Owning to the favorable profiles of pharmacokinetic and safety properties of approved drugs, repositioning existing drugs for new indications has been an effective drug discovery and development strategy28. One of the approaches under this strategy is to exploit new targets by structure-based virtual screening of existing drugs29. In our previous work, we found that one old drug, Sulindac, could bind to RXRα17. Further optimization led to K-8000325, K-8008 and K-801218 , which show improved binding selectivity and biological activity. However, whether some other old drugs could target RXRα is still unknown. In this paper, FDA-approved drugs downloaded from Drugbank30 were used as the old drug database and virtual screening of this database toward the crystal structure of LBP of RXRα was carried out. Drugs of high-score were selected for further binding and biological evaluation. Some hits were identified and their interactions with RXRα were studied theoretically and experimentally. The big difference between agonistic and antagonistic conformation of RXRα is the location of H12 in its ligand-
Figure 2. Reporter gene assay of Pitavastatin calcium (1), Fluvastatin (4), Domperidone (12) and Rosiglitazone maleate (13). MCF-7 cells cotransfected with the reporter plasmids pCMV-Myc-RXRα, pGL6-RXRELuc and pC DNA 3.1-Rellina were treated with 9-cis-RA (10−7 M) alone or together with drug for 24 h (The concentration of 9-cis-RA is 10−6 M in A and 10−7 M in B-D). (A)Concentration dependence effect of Pitavastatin calcium (1). (B)Concentration dependence effect of Fluvastatin (4). (C)Concentration dependence effect of Domperidone (12). (D)Concentration dependence effect of Rosiglitazone maleate (13). (E)Reporter of Pitavastatin calcium (1), Fluvastatin (4), Domperidone (12) and Rosiglitazone maleate (13) on Nur77 transactivation. MCF-7 cells cotransfected with pG5-Luc and pBind-Nur77-LBD. (E) were treated with compound (10 μM) for 24 h. Reporter activities were measured as described above. Data shown are mean ± SD. (F)Effect of Pitavastatin calcium (1), Fluvastatin (4), Domperidone (12) and Rosiglitazone maleate (13) on RARγ transactivation. MCF-7 cells cotransfected with pG5-Luc and pBind-RARγLBD (F) were treated with compound (10 μM) in the presence or absence of the RARγ agonist AHPN (1 μM) for 24 h. Reporter activities were measured as described above. Data shown are mean ± SD.
binding domain. The transition of agonist to antagonist lead to mobility of H12 from the surface to the other end, indicating that the binding site of antagonistic conformation may be bigger than that of agonist. It is also important to note that Leu436 of H12, which is responsible for the stabilization of interaction with holoH12, will normally produce a steric hindrance effect with antagonist, providing the ration basis for the modulator design of RXRα. Here, in order to cover both the binding pocket of modulators, 3343 old drug conformations produced from 2D structure of 1280 compounds were docked to the agonist-bound and antagonist-bound crystal structures of RXRα-LBD, respectively. The top 24 compounds possessing the highest docking scores were summarized in Table S1 (see supporting information). Interestingly, statin drugs, the type of cholesterol synthesis inhibitors31, could bind to the LBD of RXRα with high score (1, 4, 5, 7). The docking energy of Pitavastatin calcium (1) binding to agonist conformation of LBD of RXRα（PDBID： 1FBY ） is -12.171 kcal mol-1, ranking it the highest among compounds screened. And also its binding energy towards antagonist conformation of LBD of RXRα（PDBID：3A9E）is up to -10.810 kcal mol-1, ranking it sixth in the antagonism
Figure 4. Binding mode of Pitavastatin calcium (1), Fluvastatin (4), Domperidone (12) and Rosiglitazone maleate (13) to RXRα-LBD. Hydrogen bond was labeled as a dotted line. (A)Comparison of Pitavastatin calcium (1, yellow) and LG100754 (green) binding to RXRα. (B)Comparison of Fluvastatin (4, yellow) and LG100754 (green) binding to RXRα. (C)Comparison of Domperidone (12, yellow) and 9-cis-RA (green) binding to RXRα. (D)Comparison of Rosiglitazone maleate (13, yellow) and 9-cis-RA (green) binding to RXRα. Figure 3. Surface plasmon resonance of Pitavastatin calcium (1), Fluvastatin (4), Domperidone (12) and Rosiglitazone maleate (13). Gradient concentrations of each compound were injected through flow cells immobilized with purified RXRα-LBD protein. The chip is being exposed to ligand solution during 0-120 second, and the ligand is dissociated from the chip by running buffer from 120 to 200 second.The kinetic profiles are shown. (A) Sensorgram of Pitavastatin calcium (1) binding to RXRα-LBD. (B) Sensorgram of Fluvastatin (4) binding to RXRα-LBD. (C) Sensorgram of Domperidone (12) binding to RXRαLBD. (D) Sensorgram of Rosiglitazone maleate (13) binding to RXRαLBD. (E) Sensorgram of Compound (25) binding to RXRα-LBD. (F) Sensorgram of Compound (27) binding to RXRα-LBD.
screening. Similarly, other statin drugs such as Fluvastatin (4, 10.751 kcal mol-1 for RXRα agonism and -9.75 kcal mol-1 for RXRα antagonism), Rosuvastatin Calcium (5, -10.722 kcal mol-1 for agonism and -11.189 kcal mol-1 for antagonism), and vastatin Sodium (7, -10.497 kcal mol-1 for agonism and -10.3 kcal mol-1 for antagonism) have good scoring and ranking toward both of agonist and antagonist conformations of RXRα. Like classical ligands of RXRα, these four statin drugs have carboxyl acid group, which could interact with Arg316 in RXRα and the topology of statins is also L-shape, which could fit the ligandbinding pocket of RXRα. Taken together, the docking results indicated that statin drugs might bind to RXRα and could be explored as new RXRα modulators. Another worth-noting result is that Rosiglitazone maleate (13), which was shown to act as the ligand of nuclear receptor PPARγ to regulate blood sugar level32, has a moderate docking energy toward RXRα (-9.737 kcal mol-1 for agonism and -9.024 kcal mol-1 for antagonism). This maybe due to the similarity of the ligand-binding pockets in RXRα and PPARγ. Chemical analysis indicated that Rosiglitazone maleate (13) lacks carboxyl acid group, which is quite different from the classic RXRα ligands. Therefore, whether and how Rosiglitazone maleate (13) binds to RXRα remains to be determined. Among the interested 24 drugs, only 15 of them (1, 3, 4, 5, 6, 7, 8, 9, 11, 12, 13, 15, 16, 19, 24) were commercially available for further testing (see Fig. 1A). All of these compounds may be hydrophobic according to
their chemical structure, and this is coincided with the hydrophobic pocket of LBD of RXRα. In order to confirm our docking results, reporter gene assay of transcriptional activity of RXRα in MCF-7 cells in the absence or presence of 10-7 M 9-cis-RA was used to evaluate the agonist and antagonist effect of the 15 commercially-available compounds (Fig. 1A). When tested at 10 μM concentration, Pitavastatin calcium (1) and Fluvastatin (4) showed good antagonist effect on 9-cis-RA-induced RXRα transactivation (Fig. 1B). Rosuvastatin Calcium (5) also slightly antagonized the effect of 9-cis-RA, while Pravastatin Sodium (7) had no effect. In addition to the statins, Ezetimibe (3), Seratrodast (8), Cilostazol (16), and Butoconazole (19) exhibited antagonist effect on 9-cisRA-induced reporter activity. The order of their antagonism is Cilostazol (16) > Ezetimibe (3) > Butoconazole (19) > Seratrodast (8). Our data also revealed that two old drugs, Domperidone (12) and Rosiglitazone maleate (13), could enhance the effect of 9-cis-RA on inducing RXRα transactivation. Such an agonist effect occurred only in the presence of 9-cis-RA but not its absence. Four drugs with strong effect on RXRα transactivation, Pitavastatin calcium (1), Fluvastatin (4), Domperidone (12) and Rosiglitazone maleate (13), were further evaluated dose-dependently (1 μM, 5 μM, 10 μM) for their effect on RXRα transactivation by reporter assay in MCF-7 cells (Fig. 2). Our results showed that Pitavastatin calcium (1) at 1 μM concentration had better antagonistic effect than Fluvastatin (4). However, Fluvastatin (4) exhibited very nice dose dependent antagonist effect (Fig. 2B). At 10 μM, it almost completed suppressed the effect of 9-cis-RA on inducing RXRα reporter activity, while the antagonist effect of Pitavastatin calcium (1) appeared to reach plateaus at 1 μM concentration (Fig. 2A). When Domperidone (12) and Rosiglitazone maleate (13) were analyzed (Figs. 2C and D), they did not show any effect at concentration of 1 μM. At higher concentrations (5 and 10 μM), both compounds enhanced the effect of 9-cis-RA. Pitavastatin calcium (1), Fluvastatin (4), Domperidone (12) and
Rosiglitazone maleate(13) were also evaluated for their effect on tranactivation
Figure 5. Pitavastatin calcium (1) derivatives and their activity study. (A)Synthesis of Pitavastatin calcium (1) derivatives. (B)Antagonist effect of Pitavastatin calcium (1) and its analogs (25, 26, 27). MCF-7 cells cotransfected with the reporter plasmids pCMV-Myc-RXRα, pGL6RXRE-Luc and pCDNA 3.1-Rellina were treated with 9-cis-RA (10−7 M) alone or together with drug for 24 h. Reporter activities were measured as described above. Data shown are mean ± SD.
of Nur77 and RARγ by reporter assays. Our results showed that none of these compounds showed any apparent agonist or antagonist effect on constitutive Nur77 activity or RARγ activity induced by its agonist AHPN33 (see Fig. 2E and 2F). Thus, the agonist or antagonist effect we observed for these compounds is selective to RXRα and is unlikely due to their nonspecific effect on general transcriptional machinery. To determine whether the effect of Pitavastatin calcium (1), Fluvastatin (4), Domperidone (12) and Rosiglitazone maleate (13) on RXRα transactivation was due to their binding to the receptor, Surface Plasmon Resonance experiment was conducted (Fig. 3), and our results demonstrated that all of the drugs (1, 4, 12, 13) could bind to the purified RXRα-LBD protein. The binding kinetics of most of our compounds are a fast association and a fast dissociation process, making it difficult to determine its Ka and Kd. Thus we could only derivate each affinity constant by direct analysis of steady-state binding data. Until now, it is reported that protein ligand binding site of RXRα is one. Moreover, in the Biacore T200 software, analysis of steady-state binding data by fitting the curve of binding level against concentrations can only be applied to 1:1 Langmuir model, so 1:1 complex model could be reasonable for the fitting of experimental data. But it does not rule out other binding possibilities. The KD for the Pitavastatin calcium (1) binding to the LBD of RXRα is 13.30 μM. The binding kinetics of Fluvastatin (4) is quite similar to Pitavastatin calcium (1), but its binding affinity (KD = 11.04 μM) is substantially stronger than Pitavastatin calcium (1), which is opposite to the reporter assay.
Domperidone (12), as the agonist of RXRα, has the best binding affinity (KD = 8.80 μM) as comparing the other three compounds. The binding kinetics of the compound is similar to Pitavastatin calcium (1) and Fluvastatin (4), but is slightly different from Rosiglitazone maleate (13). It is interesting that the binding model of Rosiglitazone maleate (13) showed a little slow association and fast dissociation process, and its binding affinity is moderate (KD = 15.01 μM). In order to further understand how Pitavastatin calcium (1), Fluvastatin (4), Domperidone (12) and Rosiglitazone maleate (13) bind to RXRα-LBD, their docking results of each complex were performed by 10 ns molecular dynamics and their binding results were analyzed in details (Fig. 4). First, two antagonists, Pitavastatin calcium (1) and Fluvastatin (4), were docked into the antagonist conformation of RXRα-LBD (PDB ID: 3a9e). LG100754-bound structure was set as the reference for comparison (see Fig. 4A and 4B). The MD results showed that Pitavastatin calcium (1) and Fluvastatin (4) bind to a region that overlap with LG100754 binding region which is located in the Lshape of the ligand-binding pocket of RXRα. The fluorobenzene moiety of Pitavastatin calcium (1) and Fluvastatin (4) was located in the same position as the propoxyl moiety of LG100754, which could interact with the Leu436 from H12 by Van Der Waal force, pushing the conformation of RXRα from agonism to antagonism. This explains why Pitavastatin calcium (1) and Fluvastatin (4) displayed antagonistic effect in the transcriptional activity assay. The carboxyl group of Pitavastatin calcium (1) and Fluvastatin (4) were found to form ionic interaction with Arg316 and hydrogen bond with amide group of Ala327 and thus acting as anchor as the carboxyl group of classical RXRα ligands. It is noteworthy that two hydroxyl groups in the chain of Pitavastatin calcium (1) and Fluvastatin (4) do not make positive contribution to their binding in the environment of the hydrophobic pocket, implying that elimination of these polar groups would enhance their binding affinity. In a similar manner, Domperidone (12) and Rosiglitazone maleate (13) were docked into the agonist conformation of RXRα-LBD (PDB ID: 1fby). 9-cis-RA was set as the reference compound for comparison (Fig. 4C and 4D). The results showed that Rosiglitazone maleate (13) could overlap with 9-cis-RA better than Domperidone (12). Unlike 9-cis-RA, both of them lack carboxyl acid group. Nevertheless, the 5-(4-(2(methyl(pyridin-2-yl)amino)ethoxy)benzyl)thiazolidine-2,4dione moiety of Rosiglitazone maleate (13) could place in the position of carboxyl acid group of 9-cis-RA, which could also form hydrogen bond with Arg316 and Ala327, but Domperidone (12) lost this interaction completely. Maybe the hydrophobic interaction compensates and predominates the binding affinity of Domperidone (12) because of its double heterocyclic ring distributes in the hydrophobic pocket of ligand binding site. Taken together, all of the four drugs could bind to the LBP of RXRα in the same fashion as the classical RXRα ligands. According to the theoretical binding study of Pitavastatin calcium (1) to the LBP of RXRα, three derivatives of Pitavastatin calcium (1) were designed and synthesized (see Fig. 5A). Elimination of two hydroxyl groups was conducted by H2SO4/CH3COOH at room temperature. Unexpectedly, a six-ring lactone was produced (25) and only one hydroxyl group was removed in the first step. When this compound was treated with TBAF in THF solution at room temperature, the six-ring lactone was hydrolysis and another hydroxyl group was eliminated. Our target molecule (27) was obtained in the end. Secondly, in order to evaluate the importance of the carboxyl acid group in Pitavastatin calcium (1), this function group was esterificated (26) using H2SO4/C2H5OH as the reagent (see Fig. 5A).
Reporter gene assay and SPR study were used to study the structure-activity relationship of Pitavastatin calcium (1) and its derivatives (25, 26, 27) (see Fig. 5B, 3E, 3F). The results illustrated that block of carboxyl acid group would decrease the antagonistic effect on 9-cis-RA-induced RXRα reporter activity in MCF-7 cells and also lost its binding affinity to 138.3 μM (Fig. 3E). Elimination of two hydroxyl group from Pitavastatin calcium (1) would slightly enhance its antagonistic effect, and it is worth noting that compound (27) at 100 nM could effectively inhibit 9-cis-RA activity in a degree that was similar to the effect of 10 μM and also increase its binding affinity to 5.12 μM (Fig. 3F). Taken together, this SAR study demonstrated that carboxylic acid group is critical for the drugs’ biological activity and the two hydroxyl groups were not, which is consistent with the binding model of Pitavastatin calcium (1) to RXRα (see Fig. 4A). In conclusion, RXRα-based virtual screening of a drug database from Drugbank was conducted and 15 commercially available compounds were obtained for biological testing and binding studies. The results show that two statin drugs, Pitavastatin calcium (1) and Fluvastatin (4), could bind to RXRα and have potential antagonistic effect on RXRα transactivation in MCF-7 cells. Domperidone (12) and Rosiglitazone maleate (13) could also bind to RXRα but enhanced the effect on 9-cis-RA on activating RXRα. Modeling and SAR study concluded that it is the carboxylic acid group and pendant hydrophobic interactions predominate the binding of Pitavastatin calcium (1) to RXRα. Acknowledgements This work was supported by the grants from the National Nature Science Fund of China (NSFC-81301888, NSFC31271453, NSFC-31471318, NSFC-91429306, U1405229, and NSFC-91129302), the Fundamental Research Funds for the Central Universities (2013121038). References and notes 1. 2. 3. 4. 5.
6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.
18.
Evans, R. M.; Mangelsdorf, D. J. Cell 2014, 157, 255. Mangelsdorf, D. J.; Borgmeyer, U.; Heyman, R. A.; Zhou, J. Y.; Ong, E. S.; Oro, A. E.; Kakizuka, A.; Evans, R. M. Genes & Dev. 1992, 6, 329. Szanto, A.; Narkar, V.; Shen, Q.; Uray, I. P.; Davies, P. J. A.; Nagy, L. Cell Death Differ. 2004, 11, S126. Lagu, B.; Lebedev, R.; Pio, B.; Yang, M.; Pelton, P. D. Bioorg. Med. Chem. Lett. 2007, 17, 3491. Claudel, T.; Leibowitz, M. D.; Fievet, C.; Tailleux, A.; Wagner, B.; Repa, J. J.; Torpier, G.; Lobaccaro, J. M.; Paterniti, J. R.; Mangelsdorf, D. J.; Heyman, R. A.; Auwerx, J. Proc. Natl. Acad. Sci. U. S. A. 2001, 98, 2610. Kakuta, H. Yakugaku Z. J. Pharm. Soc. Jpn. 2011, 131, 917. Diab, A.; Hussain, R. Z.; Lovett-Racke, A. E.; Chavis, J. A.; Drew, P. D.; Racke, M. K. J. Neuroimmunol. 2004, 148, 116-126. Zhang, X. K.; Zhou, H.; Su, Y. Acta Biochim. Biophys. Sin. 2016, 48, 49. Crunkhorn, S. Nat. Rev. Drug Discov. 2012, 11, 271. Ottow, E.; Weinmann, H.: Nuclear receptors as drug targets; WileyVCH, 2008. Mascrez, B.; Mark, M.; Krezel, W.; Dupe, V.; LeMeur, M.; Ghyselinck, N. B.; Chambon, P. Development 2001, 128, 2049. Holmbeck, S. M. A.; Foster, M. P.; Casimiro, D. R.; Sem, D. S.; Dyson, H. J.; Wright, P. E. J. Mol. Biol. 1998, 281, 271. Bourguet, W.; Ruff, M.; Chambon, P.; Gronemeyer, H.; Moras, D. Nature 1995, 375, 377. Zhang, X. K.; Lehmann, J.; Hoffmann, B.; Dawson, M. I.; Cameron, J.; Graupner, G.; Hermann, T.; Tran, P.; Pfahl, M. Nature 1992, 358, 587. Mangelsdorf, D. J.; Evans, R. M. Cell 1995, 83, 841. Chen, Z. P.; Iyer, J.; Bourguet, W.; Held, P.; Mioskowski, C.; Lebeau, L.; Noy, N.; Chambon, P. J. Mol. Biol. 1998, 275, 55. Zhou, H.; Liu, W.; Su, Y.; Wei, Z.; Liu, J.; Kolluri, S. K.; Wu, H.; Cao, Y.; Chen, J.; Wu, Y.; Yan, T.; Cao, X.; Gao, W.; Molotkov, A.; Jiang, F.; Li, W.-G.; Lin, B.; Zhang, H.-P.; Yu, J.; Luo, S.-P.; Zeng, J.-Z.; Duester, G.; Huang, P.-Q.; Zhang, X.-K. Cancer Cell 2010, 17, 560. Chen, L.; Wang, Z. G.; Aleshin, A. E.; Chen, F.; Chen, J.; Jiang, F.; Alitongbieke, G.; Zeng, Z.; Ma, Y.; Huang, M.; Zhou, H.; Cadwell, G.;
19. 20. 21.
22. 23. 24.
25. 26. 27.
28. 29. 30. 31. 32. 33. 34.
Zheng, J.-F.; Huang, P.-Q.; Liddington, R. C.; Zhang, X. K. Su, Y. Chem. & Biol. 2014, 21, 596. Zhang, X. K.; Zhou, H.; Su, Y. Acta Biochim. Biophys. Sin. 2016, 48, 49. Heyman, R. A.; Mangelsdorf, D. J.; Dyck, J. A.; Stein, R. B.; Eichele, G.; Evans, R. M.; Thaller, C. Cell 1992, 68, 397. Levin, A. A.; Sturzenbecker, L. J.; Kazmer, S.; Bosakowski, T.; Huselton, C.; Allenby, G.; Speck, J.; Kratzeisen, C.; Rosenberger, M.; Lovey, A.; Grippo, J. F. Nature 1992, 355, 359. Egea, P. F.; Mitschler, A.; Rochel, N.; Ruff, M.; Chambon, P.; Moras, D. EMBO Journal 2000, 19, 2592. Lowe, N. J.; Prystowsky, J. H.; Bourget, T.; Edelstein, J.; Nychay, S.; Armstrong, R. J. Am. Acad. Dermatol. 1991, 24, 591. Bavinck, J. N. B.; Tieben, L. M.; Vanderwoude, F. J.; Tegzess, A. M.; Hermans, J.; Terschegget, J.; Vermeer, B. J. J. Clin. Oncol. 1995, 13, 1933. Gottardis, M. M.; Bischoff, E. D.; Shirley, M. A.; Wagoner, M. A.; Lamph, W. W.; Heyman, R. A. Cancer Res. 1996, 56, 5566. Tooker, P.; Yen, W.-C.; Ng, S.-C.; Negro-Vilar, A.; Hermann, T. W. Cancer Res. 2007, 67, 4425. Assaf, C.; Bagot, M.; Dummer, R.; Duvic, M.; Gniadecki, R.; Knobler, R.; Ranki, A.; Schwandt, P.; Whittaker, S. Br. J. Dermatol. 2006, 155, 261. Ashburn, T. T.; Thor, K. B. Nat. Rev. Drug Discov. 2004, 3, 673. Ma, D. L.; Chan, D. S. H.; Leung, C. H. Chem. Soc. Rev. 2013, 42, 2130. Wishart, D. S.; Knox, C.; Guo, A. C.; Cheng, D.; Shrivastava, S.; Tzur, D.; Gautam, B.; Hassanali, M. Nucleic Acids Res. 2008, 36, D901. Hebert, P. R.; Gaziano, J. M.; Chan, K. S.; Hennekens, C. H. Jama-J. Am. Med. Assoc. 1997, 278, 313. Gerstein, H. C.; Yusuf, S.; Holman, R. R.; Bosch, J.; Anand, S.; Avezum, A.; et. al. Lancet 2006, 368, 1096. Holmes, W. F.; Dawson, M. I.; Soprano, D. R.; Soprano, K. J. J. Cell Physiol. 2000, 185, 61. Sato, Y.; Ramalanjaona, N.; Huet, T.; Potier, N.; Osz, J.; Antony, P.; Peluso-Iltis, C.; Poussin-Courmontagne, P.; Ennifar, E.; Mely, Y.; Dejaegere, A.; Moras, D.; Rochel, N. Plos One 2010, 5, e15119.