Ring-truncated deguelin derivatives as potent Hypoxia InducibleFactor-1a (HIF-1a) inhibitors
Ho Shin Kim a, Mannkyu Hong a, Su-Chan Lee a, Ho-Young Lee a, Young-Ger Suh a,Dong-Chan Oh b, Ji Hae Seo c, Hoon Choi c, Jun Yong Kim c, Kyu-Won Kim c, d,Jeong Hun Kim e, Joohwan Kim f, Young-Myeong Kim f, So-Jung Park g, Hyun-Ju Park g,Jeewoo Lee a, *
a Research Institute of Pharmaceutical Sciences, College of Pharmacy, Seoul National University, Seoul, 151-742, South Koreab Natural Products Research Institute, College of Pharmacy, Seoul National University, Seoul, 151-742, South Koreac SNU-Harvard NeuroVascular Protection Research Center, College of Pharmacy, Seoul National University, Seoul, 151-742, South Koread Department of Molecular Medicine and Biopharmaceutical Sciences, Graduate School of Convergence Science and Technology, Seoul National University,Seoul, 151-742, South Koreae College of Medicine, Seoul National University, Seoul, 151-742, South Koreaf School of Medicine, Kangwon National University, Kangwon-do, 200-701, South Koreag School of Pharmacy, Sungkyunkwan University, Suwon, 440-746, South Korea
a r t i c l e i n f o
Article history:Received 31 July 2015Received in revised form24 September 2015Accepted 25 September 2015Available online 30 September 2015
Keywords:Hypoxia Inducible Factor-1HIF-1Heat shock protein 90HSP90AntitumorDeguelin
A series of fluorophenyl and pyridine analogues of 1 and 2 were synthesized as ring-truncated deguelinsurrogates and evaluated for their HIF-1a inhibition. Their structureeactivity relationship was system-atically investigated based on the variation of the linker B-region moiety. Among the inhibitors, com-pound 25 exhibited potent HIF-1a inhibition in a dose-dependent manner and significant antitumoractivity in H1299 with less toxicity than deguelin. It also inhibited in vitro hypoxia-mediated angiogenicprocesses in HRMECs. The docking study indicates that 25 occupied the C-terminal ATP-binding pocket ofHSP90 in a similar mode as 1, which implies that the anticancer and antiangiogenic activities of 25 arederived from HIF-1a destabilization by binding to the C-terminal ATP-binding site of hHSP90.
Hypoxia Inducible Factor-1 (HIF-1) is the transcription factorthat regulates the cellular response for the survival of cells inhypoxia. HIF-1 protein is a heterodimer that consists of a consti-tutively expressed b-subunit and an oxygen-regulated a-subunit.Under normoxia, HIF-1a is degradable by the pVHL-mediatedubiquitin protease pathway, which includes hydroxylation byprolyl hydroxylases (PHD), binding to the product of the vonHippel-Lindau (pVHL), being tagged with polyubiquitin and
served.
proteasomal degradation. However, under hypoxia, HIF-1a be-comes stable from proline hydroxylation, accumulates and trans-locates to the nucleus. There, HIF-1a dimerizes with HIF-1b toactivate the HIF-1 complex, which binds to hypoxia-response ele-ments (HRE) in the HIF target genes to control transcription. Thisregulation can induce angiogenesis, proliferation, metastasis andinvasion of cancer cells [1,2]. Therefore, the HIF-1a inhibition caninhibit this angiogenesis, decrease the proliferation of cancer cellsand reduce the chemotherapy resistance [3e9]. In addition, therehas been growing interest in the biology of the HIF-1 pathway andits role in human diseases that are associated with the hypoxicmicro-environment, such as cancer, stroke and heart disease[10e15].
Fig. 2. Pharmcophoric regions of the ring-truncated deguelin analogues.
H.S. Kim et al. / European Journal of Medicinal Chemistry 104 (2015) 157e164158
Heat shock protein 90 (HSP90) is a molecular chaperone thatregulates the post-translational folding, stability and function of itsclient proteins such as ErbB2, Src, c-MET, AKT, Raf-1, MMP-2 andHIF-1a. Because the chaperone inhibition can induce the decom-position of HIF-1a, HSP90 inhibition is considered a new andeffective therapy against angiogenesis-associated diseases such ascancer. Structurally, the HSP90 protein contains three functionaldomains: the ATP-binding, protein-binding, and dimerizingdomain. Most HSP90 inhibitors that were developed as anticanceragents have been identified as the so-called N-term inhibitors,which bind to an ATP-binding domain in the N-terminal. However,these inhibitors cause some problems, and one of them induce aheat shock response (HSR), which ultimately leads to an increase inHSP90 and anti-apototic proteins such as HSP70 and HSP27.Therefore, the inhibition of another ATP-binding site in the C-ter-minal may be an alternative strategy to discover clinically appli-cable HSP90 inhibitors [16e20].
Deguelin, which is a naturally occurring rotenoid, has been re-ported to prevent tobacco carcinogen-induced lung carcinogenesisby blocking the Akt activation; it also exhibits potent apoptotic andantiangiogenic activities against diverse transformed cells andcancer cells in vitro (Fig. 1) [21]. It interferes with the chaperonefunction of HSP90 by inhibiting ATP binding, which induces thedestabilization of HIF-1a and consequent tumor growth reductionin xenograft models of various human cancers [22].
Fig. 1. Deguelin and its ring-truncated surrogates.
Previously, Chang et al. reported that two ring-truncateddeguelin analogues, compounds 1 and 2, exhibited excellent HIF-1a suppression and potent cell growth inhibition in the humannon-small-cell lung carcinoma cell line, H1299 (Fig. 1) [23]. Inaddition, their in vivo antiangiogenic activities were observed in thezebrafish model in a dose-dependent manner. Jo et al. also reportedthat the destabilization of HIF-1a by both compounds suppressedhypoxia-mediated retinal neovascularization and vascular leakagein diabetic retinawithout inducing a definite toxicity in the oxygen-induced retinopathy mouse model [24]. The results indicated thatthe new HSP90 inhibitors 1 and 2 were considered promising leadcompounds for anti-proliferation and anti-angiogenesis.
Structurally, the ring-truncated deguelin scaffold was dividedinto three pharmacophoric parts: A-region (3,4-dimethoxyphenyl),B-region (linker), and C-region (2,2-dimethyl chromene ring)(Fig. 2). To further optimize the leads 1 and 2 as anticancer agents,we decided to investigate their fluorophenyl (X ¼ CeF) and pyri-dine (X ¼ N) derivatives in the A-region to improve the targetbinding for the HIF-1a inhibition and aqueous solubility. We pre-sumed that polarizing the 3,4-dimethoxyphenyl group by incor-porating a polar nitrogen or fluoro atom might provide betterbinding interaction to HSP90 and pharmacokinetic profile. Inaddition, a-methyl carbonyl in the B-region was modified with itsbioisosteres, which include olefin, diol, alcohol, and acyl groups(Fig. 2).
In this paper, we investigated the structure activity relationshipsof the fluorophenyl and pyridine analogues of 1 and 2 for HIF-1ainhibition using western blot assay. With selected potent inhibitorsin the series, we further characterized their cytotoxicities in a tu-mor cell line and anti-angiogenesis in hypoxia-mediated angio-genic processes in human retinal microvascular endothelial cells. Inaddition, we performed the docking study with HSP90 to deter-mine its mode of action.
2. Result and discussion
2.1. Chemistry
The final compounds were generally synthesized through acarbonecarbon bond formation between the phenyl sulfonyl in-termediates of the A-region and 2,2-dimethyl chromene aldehydeof the C-region.
To synthesize the phenyl sulfonyl intermediates (7, 8), fluo-rophenyl aldehyde 3 and pyridine aldehyde 4 as the starting ma-terials were prepared from commercially available o-fluorocatechol and 2-bromo-3-pyridinol, respectively, according to theprocedures in the literature [25,26]. The prepared aldehydes (3, 4)were reduced to the corresponding alcohols (5, 6), which werehalogenated and subsequently converted into the correspondingsulfonyl intermediates (7, 8), respectively, using benzene sulfinicacid sodium salt (Scheme 1). Then, 2,2-Dimethyl chromene C-re-gion aldehyde 9 [27] was coupled with the sulfonyl intermediates(7, 8) using n-BuLi to provide two diasteromeric mixtures, whichwere directly oxidized without separation and subsequentlydesulfonylated to obtain ketones (12, 13). (Scheme 2).
The final alcohol and acyl analogues were synthesized using theconventional routes from the corresponding ketones (Scheme 3).Ketones (12,13) were alkylated with the corresponding alkyl halideto produce a-alkyl ketones (14e17), which were reduced by NaBH4to provide alcohol compounds (18e23). Under this condition, thesyn diastereomer was found as a predominant isomer(syn:anti ¼ ca. 20:1) when R1 was a monoalkyl group. For thestructural analysis, only 25 (syn) and 26 (anti) were separated, andtheir stereochemistries were assigned based on the 1H NMR anal-ysis according to a previous report [28]. The alcohols were furtheracylated to obtain the acylated products (24e36, 41e43) or car-bamoylated in two steps to obtain the carbamoylated products(37e40, 44e45).
To synthesize the olefinic and dihydroxyl analogues of the B-region, alcohol compounds (18, 21) were dehydrated under acidiccondition to provide only trans isomers (46, 47), which weredihydroxylated using OsO4 to obtain the syn-diol compounds (48,49). The oxime (50) and methyl oxime (51) analogues of the pyri-dine A-region were obtained from ketone 13 (Scheme 4).
Scheme 1. Synthesis of the A-region. Reagents and conditions: (a) NaBH4, MeOH, 0 �C to rt; (b) SOCl2, MC, 0 �C to rt; (c) PhSO2Na, DMF, rt.
Scheme 2. Synthesis of the ketone analogs. Reagents and conditions: (a) 7 or 8, n-BuLi, THF, �78 �C; (b) DMP, MC, rt; (c) (n-Bu)3SnH, AIBN, benzene, reflux.
Scheme 3. Syntheses of the alcohol and acyl analogues. Reagents and conditions: (a) R-X, NaH, DMF 0 �C to rt; (b) NaBH4, MeOH, rt; (c) R-COCl, Pyridine, MC, 0�C; (d) i) CDI, MC,0�C; ii) NH4OH or NH2Me or NHMe2, water 40�C.
Scheme 4. Syntheses of the olefin, dihydroxy and oxime analogs. Reagents and conditions: (a) p-TSA, toluene, reflux; (b) OsO4, NMO, acetone, water, rt; (c) NH2OR-HCl, NaOAc,EtOH, water, reflux.
H.S. Kim et al. / European Journal of Medicinal Chemistry 104 (2015) 157e164 159
To synthesize the chiral S-isomer of 16, the pyridine surrogate ofcompound 2, chiral (S)-methyl aldehyde of A-region was preparedfrom alcohol 6 first (Scheme 5). Alcohol 6 was converted to the
nitrile and subsequently hydrolyzed to the corresponding acid 52.Acid 52 was converted to the corresponding (S)-oxazolidinone,which was methylated to provide (S)-methyl product 53. Further
Scheme 6. Synthesis of the chiral 16S. Reagents and conditions: (a) NaNO2, KI, c-HCl, H2O, 0 �C to rt; (b) 54, n-BuLi, THF, �78 �C; (c) DMP, MC, rt.
Table 3Acyl analogues as ring-truncated deguelin.
H.S. Kim et al. / European Journal of Medicinal Chemistry 104 (2015) 157e164160
steps produced the key chiral aldehyde 54 to couple with the C-region. The C-region iodide 56 was synthesized from the corre-sponding amine 55 by Sandmeyer reaction, which was preparedfrom resorcinol as shown in our previous report [23]. The lithiationof iodo compound 56 and subsequent coupling with aldehyde 54provided the corresponding alcohol, which was oxidized to pro-duce the final chiral (S)-methyl ketone 16S (Scheme 6).
2.2. Structure activity relationship for HIF-1a inhibition
The synthesized compounds were evaluated for their HIF-1ainhibition using western blot assay with H1299 cell lines [23]. TheH1299 cells were untreated or treated with deguelin and the ana-logues at 100 nM for 72 h and subsequently incubated underhypoxic conditions for 12 h. The resulting cell lysates were used for
Table 1Ketone analogues as ring-truncated deguelin.
Code X R1, R2 WB (100 nM)
Deguelin 641 C H, H 85.32 C H, (S)eMe 82.112 CeF H, H 67.8 (±37.03)14 CeF H, Me 78.6 (±16.44)15 CeF Me, Me 74.9 (±6.17)13 N H, H 77.6 (±10.86)16 N H, Me 89.3 (±11.54)16S N H, (S)eMe 44.2 (±33.24)17 N Me, Me 99.9 (±10.10)
Table 2Alcohol analogues as ring-truncated deguelin.
Code X R1, R2 WB (100 nM)
18 CeF H, H 74.3 (±33.76)19 CeF H, Me 75.0 (±24.20)20 CeF Me, Me 68.0 (±33.99)21 N H, H 75.2 (±36.33)22 N H, Me 71.2 (±29.4)23 N Me, Me 69.2 (±19.9)
the western blot analysis with a monoclonal antibody against HIF-1a. Because all of the compounds have fluorophenyl or pyridine A-region and 2,2-dimethyl 2H-chromene C-region, their inhibitoryactivities are represented by the structure of the B-region, as shownin Tables 1e4.
Code X R1, R2 R3 WB (100 nM)
24 CeF H, H Me 69.3 (±49.85)25 (Syn) CeF H, Me Me 52.2 (±12.56)26 (Anti) CeF H, Me Me 79.4 (±5.01)27 CeF Me, Me Me 83.4 (±8.36)28 CeF eCH2eCH2e Me 154.6 (±17.09)29 CeF H, H Et 77.3 (±39.35)30 CeF H, Me Et 63.7 (±43.72)31 CeF Me, Me Et 88.2 (±16.37)32 CeF H, H Cyclopropyl 70.3 (±19.92)33 CeF H, Me Cyclopropyl 35.0 (±6.05)34 CeF H, Et Cyclopropyl 87.9 (±20.53)35 CeF Me, Me Cyclopropyl 110.6 (±38.89)36 CeF H, Me Cyclobutyl 69.5 (±18.25)37 CeF H, H NH2 88.9 (±21.10)38 CeF H, Me NH2 48.8 (±2.99)39 CeF H, Me NHMe 76.5 (±28.37)40 CeF H, Me NMe2 64.1 (±20.46)41 N H, H Me 114.0 (±27.98)42 N H, Me Me 96.7 (±19.97)43 N Me, Me Me 113.2 (±6.43)44 N H, H NH2 75.0 (±18.08)45 N Me, Me NH2 107.7 (±25.14)
Table 4Olefin, diol and oxime analogues as ring-truncated deguelin.
Code X Y, Z WB (100 nM)
46 CeF eCH]CHe 133.8 (±40.03)47 N eCH]CHe 133.2 (±24.03)48 CeF eCH(OH)eCH(OH)e 68.1 (±4.23)49 N eCH(OH)eCH(OH)e 68.9 (±1.46)50a N eCH2eC(]NOH)e 78.5 (±25.36)50b N eCH2eC(]NOH)e 79.2 (±34.62)51 N eCH2eC(]NOCH3)e 64.8 (±19.78)
Fig. 3. Antitumor activity of 25 and 33 in H1299.
Fig. 4. Effect of deguelin and 25 on the viability of mouse hippcampus (HT-22).
H.S. Kim et al. / European Journal of Medicinal Chemistry 104 (2015) 157e164 161
First, we explored the ketone B-region analogues (Table 1). Thefluorophenyl and pyridine surrogates (12, 13) of parent 1 showedbetter HIF-1a inhibition than did parent 1. Interestingly, the pyri-dine surrogate (16S) of parent 2 exhibited excellent inhibition,which was ca. 2-fold more potent than 2 and its racemate 16. 16Salso had better activity than deguelin did. This result indicates thatthe corresponding fluorophenyl and pyridine surrogates are betterHIF-1a inhibitors than did the previous ring-truncated deguelinanalogues. The analysis indicates that the inhibitory activity in-creases in the order of no-substituent > a-methyl � a,aʹ-dimethylgroups next to ketone.
Second, we investigated the alcohol B-region analogues(Table 2). Compounds 18 and 19 without substitution at the a-po-sition showed similar inhibitions with the corresponding ketoneanalogues (12, 13). However, unlike ketone, the inhibitory activityincreased in the order of a,aʹ-dimethyl > a-methyl > no-substituentnext to alcohol.
Third, we examined the O-acyl and O-carbamoyl analogues(Table 3). In this series, we incorporated amethyl, ethyl, dimethyl orcyclopropyl group as an a-substituent and acetoxy, propionyloxy,cyclopropanecarbonyloxy, carbamoyloxy, and dimethylcarbamoy-loxy as the acyl and carbamoyl groups. All of the a-monoalkyl an-alogues (25, 30, 33, 34, 36, 38e40, 42) were syn isomers that wereisolated as a predominant form in the B-region. For the SAR anal-ysis, only anti-isomer 26 was separated and found to be much lesspotent than syn 25. In this series, the four compounds (25, 33, 38,40) exhibited better activities than did deguelin. Interestingly, all ofthem had a-methyl moiety next to the O-acyl/carbamate groups,which suggests that a methyl group provided a favorable confor-mation for the HIF-1a inhibition. A more steric environment at thea-position, such as the ethyl (34), dimethyl (27, 31, 35, 43, 45), andcyclopropyl (28) groups, diminished the inhibitory activity.
Finally, we examined olefin, diol, and oxime B-region analogues(Table 4). Although the olefinic derivatives (46, 47) did not showany HIF-1a inhibition, the diol derivatives (48, 49) exhibited goodHIF-1a inhibition, which was comparable to that of deguelin. 48and 49 had better activity than did their corresponding mono-hydroxyl surrogates (18, 21) and similar activity to the ketonesurrogates (12, 13). In addition, the methyl oxime derivative ofpyridine A-region (51) displayed excellent inhibition, which wasbetter than deguelin, but its oxime isomers (50) showed moderateinhibition.
Among the tested compounds, we selected 6 compounds withbetter HIF-1a inhibition than deguelin and evaluated their inhibi-tion at a lowconcentration (10 nM) to examine the dose-dependentinhibition (Table 5). As expected, all of the selected compoundsshowed better inhibition than deguelin at 10 nM. In particular,compound 25 showed the best inhibition.
2.3. Antitumor activity
Deguelin and its analogs exhibited profound antiproliferativeeffects on human malignant bronchial epithelial cells and in vivo
Table 5HIF-1a inhibition of the selected potent compounds.
efficacy in suppressing the lung tumor formation in the micemodel, which indicates that they are promising antitumor candi-dates for lung cancer [21,22]. In this regard, the two compounds 25and 33, which were selected based on the primary HIF-1a inhibi-tion screening result, were tested for antitumor activity in thehuman non-small-cell lung carcinoma cell (H1299) line using theMTT assay (Fig. 3). As expected, both compounds showed prom-ising cytotoxicity in a concentration-dependent manner, where theviability of the cell was reduced to 50% at 20 mM. The result in-dicates that the cytotoxicities are derived from their HIF-1ainhibition.
Because the use of deguelin as anticancer agents is limitedbecause of its potential toxicities, we examined the toxicity of 25 bytesting the effect of 25 on the viability of hippocampal cell (HT-22).Compared to deguelin, compound 25 showed significantly reducedcytotoxicity, which indicates that 25 is a more potent and less toxicantitumor agent than deguelin (Fig. 4).
2.4. Anti-angiogenesis activity
Angiogenesis is the formation of new blood vessels, which are asignificant component of a solid tumor for its proliferation andmetastasis. The VEGF (vascular endothelial growth factor) plays apivotal role in tumor angiogenesis and is up-regulated in tumorcells under hypoxic conditions.
The anti-angiogenesis activity of compound 25 was examinedusing proliferation and migration assays in human retinal micro-vascular endothelial cells (HRMECs), which were previouslydescribed [29], and compared to deguelin at each 100 nM (Fig. 5).Although the VEGF activated angiogenesis in the proliferation andmigration assays, compound 25 effectively inhibited the in vitrohypoxia-mediated angiogenic processes in both assays at a com-parable level to deguelin, which indicates that it can be used as apotential anti-angiogenic agent to suppress the retinal neo-vascularization for the treatment of diabetic retinopathy.
Fig. 5. Anti-angiogenesis activity of 25 in HRMECs (A: proliferation assay; B: migration assay).
H.S. Kim et al. / European Journal of Medicinal Chemistry 104 (2015) 157e164162
2.5. Molecular modeling
We recently demonstrated that L80 (Fig. 6D), which is a cyclizedanalogue of compound 25, strongly diminished the physical inter-action between HIF-1a and HSP90 in H1299 cells by directlymodulating the function of HSP90 using co-immunoprecipitationand pull-down assays [30]. In addition, we showed that L80 couldbind to the C-terminal ATP-binding pocket of HSP90 to hamper theHSP90 function when we studied recombinant HSP90 proteins,which contain the full-length protein, N-terminal/middle domain,middle domain, and C-terminal domain. Furthermore, our recentstudy confirmed that lead 1 also bound to the C-terminal ATP-binding pocket of HSP90, disrupted the HSP90 function andcaused a degradation of client proteins without affecting the HSP70expression (unpublished data). These data suggest that compound25 inhibits the HSP90 function by binding to the C-terminal ATP-binding pocket of HSP90.
Fig. 6. Molecular modeling of 25. (A) Overlay of the docking pose of 25 (purple carbon) anresidues of the binding site are represented by the gray carbon capped stick. The active site ihydrophilic) (B) Docked pose of 25. The dashed ellipsoids are hydrogen-bonding interactioninterface of the open state hHsp90.25 is represented by the purple CPK model. Chain A is
To ensure the binding of 25 to the C-terminal ATP-bindingpocket of HSP90 for the HIF-1a inhibition, we performed a dock-ing study of 25 in the C-terminal ATP-binding pocket of hHSP90homodimer (Fig. 6). The active site was determined based on thereported hHSP90 C-terminal ATP-binding site [31]. To examine thedocking poses, we conducted a docking study of 25 in the C-ter-minal ATP-binding pocket of the hHSP90 homodimer. The C-ter-minal ATP-binding site was assigned as an active site at thedimerization interface (Fig. 6C). After docking, we compared thedocking pose and score of ATP with those of 25. Fig. 6A shows thatATP (white carbon) fits well into the ATP-binding pocket of chain A.The adenine ring is located deep into the hydrophobic pocket. Thephosphate moiety interacts with multiple hydrogen bonds withGlu611, and 30-hydroxy of the sugar moiety interacts with Ser677 inchain A via a hydrogen bond. Compound 25 is partially super-imposed with the sugar-phosphate moiety of ATP. However, 25binds to the active site with a higher docking score (-logKd ¼ 7.80)
d ATP (white carbon) in the active site of the hHsp90 C-terminal. The key amino acids shown as a lipophilicity property surface map (brown color: hydrophobic; blue color:s (<2.8 Å). A and B indicate the chain names. (C) Binding site for 25 in the dimerizationshown as an orange ribbon, and chain B is a cyan ribbon. (D) Structure of L-80.
H.S. Kim et al. / European Journal of Medicinal Chemistry 104 (2015) 157e164 163
than ATP (-logKd¼ 6.29). 25 occupies the centre of the active site byforming multiple hydrogen bonds with both chains A and B (Fig. 6Band C). The oxygen atom of the methoxy group that is attached tothe 2,2-dimethyl chromene ring forms a hydrogen bond with theside chain NH of Lys615 in chain A, which is part of the ATP-bindingpocket. Two oxygen atoms of the acetoxy group form hydrogenbonds with the side chain of Lys615 in chain B. Furthermore, thesubstituents of the 3,4-dimethoxy-2-fluorophenyl group areinvolved in forming hydrogen bonding networks. The fluoro groupand oxygen atom of 3-methoxy form hydrogen bonds with Lys615in chain B. The oxygen atom of 4-methoxy interacts with side chainNH of Asn 622 (chain B). Hugel and coworkers reported that therole of the C-terminal inhibitors is to prohibit the global confor-mational changes, which include N-terminal dimerization and theformation of the ATP binding pocket [32e35]. Our docking resultsdemonstrate that 25 can compete with ATP in binding to the C-terminal ATP binding site and stabilize the open state of C-terminalhHSP90 with the bridging hydrogen bond networks at the interfaceof the homodimer.
3. Conclusion
A series of fluorophenyl and pyridine analogues of 1 and 2,which are lead HIF-1a inhibitors through HSP90 inhibition as ring-truncated deguelin surrogates, were investigated for their HIF-1ainhibition. Their structureeactivity relationship was systematicallyexamined by varying the B-region moiety. Among the studied in-hibitors, 6 compounds showed better HIF-1a inhibition thandeguelin. The two selected inhibitors, 25 and 33, exhibited prom-ising antitumor activity in human non-small cell lung carcinoma(H1299). In addition, compound 25 inhibited in vitro hypoxia-mediated angiogenic processes in human retinal microvascularendothelial cells (HRMECs). The docking study of 25 with thehHSP90 C-terminal ATP-binding site indicated that 25 snuglybound to the C-terminal ATP-binding pocket of the hHSP90homodimer with a high docking score, as found in its cyclizedanalog L80. Overall, compound 25 is a potential anticancer andantiangiogenic agent with an HIF-1a inhibition mechanismthrough binding to the C-terminal ATP-binding site of hHSP90.
Acknowledgments
This work was supported by grants from the National ResearchFoundation of Korea (NRF), the Ministry of Science, Republic ofKorea (NRF-20110019400).
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.ejmech.2015.09.033.
[2] M. Yee Koh, T.R. Spivak-Kroizman, G. Powis, HIF-1 regulation: not so easycome, easy go, Trends Biochem. Sci. 33 (2008) 526e534.
[3] G.L. Semenza, Targeting HIF-1 for cancer therapy, nature reviews, Cancer 3(2003) 721e732.
[4] Hua Zhong, Angelo M. De Marzo, Erik Laughner, Michael Lim, David A. Hilton,David Zagzag, Peter Buechler, William B. Isaacs, Gregg L. Semenza, JonathanW. Simons, Overexpression of hypoxia-inducible factor 1a in common humancancers and their metastases, Cancer Res. 59 (1999) 5830e5835.
[5] K.L. Talks, H. Turley, K.C. Gatter, P.H. Maxwell, C.W. Pugh, P.J. Ratcliffe,A.L. Harris, The expression and distribution of the hypoxia-inducible factorsHIF-1a and HIF-2a in normal human tissues, cancers, and tumor-associatedmacrophages, Am. J. Pathol. 157 (2000) 411e421.
[6] N.J.P. Beasley, R. Leek, M. Alam, H. Turley, G.J. Cox, K. Gatter, P. Millard,
S. Fuggle, A.L. Harris, Hypoxia-inducible factors HIF-1a and HIF-2a in head andneck cancer: relationship to tumor biology and treatment outcome in surgi-cally resected patients, Cancer Res. 62 (2002) 2493e2497.
[7] M. Volm, R. Koomagi, Hypoxia-inducible factor (HIF-1) and its relationship toapoptosis and proliferation in lung cancer, Anticancer Res. 20 (2000)1527e1533.
[8] P. Birner, M. Schindl, A. Obermair, G. Breitenecker, G. Oberhuber, Expression ofhypoxia-inducible factor 1a in epithelial ovarian tumors: its impact onprognosis and on response to chemotherapy, Clin. Cancer Res. 7 (2001)1661e1668.
[9] M.I. Koukourakis, A. Giatromanolaki, J. Skarlatos, L. Corti, S. Blandamura,M. Piazza, K.C. Gatter, A.L. Harris, Hypoxia inducible factor (HIF-1a and HIF-2a) expression in early esophageal cancer and response to photodynamictherapy and radiotherapy, Cancer Res. 61 (2001) 1830e1832.
[10] A. Unruh, A. Ressel, H.G. Mohamed, R.S. Johnson, R. Nadrowitz, E. Richter,D.M. Katschinski, R.H. Wenger, The hypoxia-inducible factor-1 alpha is anegative factor for tumor therapy, Oncogene 22 (2003) 3213e3220.
[11] D.M. Aebersold, P. Burri, K.T. Beer, J. Laissue, V. Djonov, R.H. Greiner,G.L. Semenza, Expression of hypoxia-inducible factor-1a: a novel predictiveand prognostic parameter in the radiotherapy of oropharyngeal cancer,Cancer Res. 61 (2001) 2911e2916.
[12] A. Rapisarda, B. Uranchimeg, D.A. Scudiero, M. Selby, E.A. Sausville,R.H. Shoemaker, G. Melillo, Identification of small molecule inhibitors ofhypoxia-inducible factor 1 transcriptional activation pathway, Cancer Res. 62(2002) 4316e4324.
[13] E.-J. Yeo, Y.-S. Chun, Y.-S. Cho, J. Kim, J.-C. Lee, M.-S. Kim, J.-W. Park, YC-1: apotential anticancer drug targeting hypoxia-inducible factor 1, J. Natl. CancerInst. 95 (2003) 516e525.
[14] S.J. Welsh, R.R. Williams, A. Birmingham, D.J. Newman, D.L. Kirkpatrick,G. Powis, The thioredoxin redox inhibitors 1-methylpropyl 2-imidazolyl di-sulfide and pleurotin inhibit hypoxia-induced factor 1a and vascular endo-thelial growth factor formation, Mol. Cancer Ther. 2 (2003) 235e243.
[15] J.-W. Lee, S.-H. Bae, J.-W. Jeong, S.-H. Kim, K.-W. Kim, Hypoxia-inducible factor(HIF-1)a: its protein stability and biological functions, Exp. Mol. Med. 1 (2004)1e12.
[16] Y. Xia, H.K. Choi, K. Lee, Recent advances in hypoxia-inducible factor (HIF)-1inhibitors, Eur. J. Med. Chem. 49 (2012) 24e40.
[17] J.C. Young, V.R. Agashe, K. Siegers, F.U. Hartl, Pathways of chaperone-mediatedprotein folding in the cytosol, nature reviews, Mol. Cell Biol. 5 (2004)781e791.
[18] L. Neckers, P. Workman, Hsp90 molecular chaperone inhibitors: are we thereyet?, Clinical cancer research, Off. J. Am. Assoc. Cancer Res. 18 (2012) 64e76.
[19] L. Whitesell, S.L. Lindquist, HSP90 and the chaperoning of cancer, nature re-views, Cancer 5 (2005) 761e772.
[20] R. Bagatell, L. Whitesell, Altered Hsp90 function in cancer: a unique thera-peutic opportunity, Mol. Cancer Ther. 3 (2004) 1021e1030.
[21] (a) J. Beliakoff, R. Bagatell, G. Paine-Murrieta, C.W. Taylor, A.E. Lykkesfeldt,L. Whitesell, Hormone-refractory breast cancer remains sensitive to theantitumor activity of heat shock protein 90 inhibitors, Clin. Cancer Res. 9(2003) 4961e4971;(b) B.K. Eustace, T. Sakurai, J.K. Stewart, D. Yimlamai, C. Unger, C. Zehetmeier,B. Lain, C. Torella, S.W. Henning, G. Beste, B.T. Scroggins, L. Neckers, L.L. Ilag,D.G. Jay, Functional proteomic screens reveal an essential extracellular role forhsp90 alpha in cancer cell invasiveness, Nat. Cell Biol. 6 (2004) 507e514;(c) L. Stepanova, X. Leng, S.B. Parker, J.W. Harper, Mammalian p50Cdc37 is aprotein kinase targeting subunit of Hsp90 that binds and stabilizes Cdk4 Gene,Dev. 10 (1996) 1491e1502;(d) S.E. Holt, D.L. Aisner, Joseph Baur, V.M. Tesmer, M. Dy, M. Ouellette,J.B. Trager, G.B. Morin, D.O. Toft, J.W. Shay, W.E. Wright, M.A. White, Func-tional requirement of p23 and Hsp90 in telomerase complexes, Genes Dev. 13(1999) 817e826;(e) P.N. Münster, D.C. Marchion, A.D. Basso, N. Rosen, Degradation of HER2 byAnsamycins induces growth arrest and apoptosis in cells with HER2 over-expression via a HER3, phosphatidylinositol 3ʹ-kinase-AKT-dependentpathway, Cancer Res. 62 (2002) 3132e3137;(f) A.D. Basso, D.B. Solit, G. Chiosis, B. Giri, P. Tsichlis, N. Rosen, Akt forms anintracellular complex with heat shock protein 90 (Hsp90) and Cdc37 and isdestabilized by inhibitors of Hsp90 function, J. Biol. Chem. 277 (2002)39858e39866;(g) J.S. Isaacs, Y.J. Jung, E.G. Mimnaugh, A. Martinez, F. Cuttitta, L.M. Neckers,Hsp90 regulates a von hippel lindau-independent hypoxia-inducible factor-1alpha-degradative pathway, J. Biol. Chem. 277 (2002) 29936e29944.
[22] S.H. Oh, J.K. Woo, Y.D. Yazici, J.N. Myers, W.Y. Kim, Q. Jin, S.S. Hong, H.J. Park,Y.G. Suh, K.W. Kim, Structural basis for depletion of heat shock protein 90client proteins by deguelin, J. Natl. Cancer Inst. 99 (2007) 949e961.
[23] D.J. Chang, H. An, K.S. Kim, H.H. Kim, J. Jung, J.M. Lee, N.J. Kim, Y.T. Han, H. Yun,S. Lee, G. Lee, S. Lee, J.S. Lee, J.H. Cha, J.H. Park, J.W. Park, S.C. Lee, S.G. Kim,J.H. Kim, H.Y. Lee, K.W. Kim, Y.G. Suh, Design, synthesis, and biological eval-uation of novel deguelin-based heat shock protein 90 (HSP90) inhibitorstargeting proliferation and angiogenesis, J. Med. Chem. 55 (2012)10863e10884.
[24] D.H. Jo, H. An, D.J. Chang, Y.Y. Baek, C.S. Cho, H.O. Jun, S.J. Park, J.H. Kim,H.Y. Lee, K.W. Kim, J. Lee, H.J. Park, Y.M. Kim, Y.G. Suh, J.H. Kim, Hypoxia-mediated retinal neovascularization and vascular leakage in diabetic retina issuppressed by HIF-1alpha destabilization by SH-1242 and SH-1280, novel
H.S. Kim et al. / European Journal of Medicinal Chemistry 104 (2015) 157e164164
hsp90 inhibitors, J. Mol. Med. 92 (2014) 1083e1092.[25] A. Adejare, F. Gusovsky, W. Padgett, C.R. Creveling, J.W. Daly, K.L. Kirk, Syn-
theses and adrenergic activities of ring-fluorinated epinephrines, J. Med.Chem. 31 (1988) 1972e1977.
[26] C. Nemecek, W.A. Metz, S. Wentzler, F.X. Ding, C. Venot, C. Souaille,A. Dagallier, S. Maignan, J.P. Guilloteau, F. Bernard, A. Henry, S. Grapinet,D. Lesuisse, Design of potent IGF1-R inhibitors related to bis-azaindoles,Chem. Biol. Drug Des. 76 (2010) 100e106.
[27] G.E. Henry, H. Jacobs, A short synthesis of 5-methoxy-2,2,-dimethyl-2H-1-benzopyran-6-propanoic acid methyl ester, Tetrahedron 57 (2001)5335e5338.
[28] N. Matsumori, D. Kaneno, M. Murata, H. Nakamura, K. Tachibana, Stereo-chemical determination of acyclic structures based on carbon-proton spin-coupling constants. A method of configuration analysis for natural product,J. Org. Chem. 64 (1999) 866e876.
[29] C.K. Kim, Y.K. Choi, H. Lee, K.S. Ha, M.H. Won, Y.G. Kwon, Y.M. Kim, The far-nesyltransferase inhibitor LB42708 suppresses vascular endothelial growthfactor-induced angiogenesis by inhibiting ras-dependent mitogen-activatedprotein kinase and phosphatidylinositol 3-kinase/Akt signal pathways, Mol.Pharmacol. 78 (2010) 142e150.
[30] S.-C. Lee, H.-Y. Min, H. Choi, H.S. Kim, K.-C. Kim, S.-J. Park, M.A. Seong, J.H. Seo,H.-J. Park, Y.-G. Suh, K.-W. Kim, J. Lee, H.-Y. Lee, Synthesis and evaluation of a
novel deguelin derivative, L80, which disrupts ATP binding to the C-terminaldomain of heat shock protein 90, Mol. Pharm. 88 (2015) 245e255.
[31] (a) M. Sgobba, G. Degliesposti, A.M. Ferrari, G. Rastelli, Structural models andbinding site prediction of the C-terminal domain of human Hsp90: a newtarget for anticancer drugs, Chem. Biol. Drug Des. 71 (2008) 420e433;(b) M. Sgobba, R. Forestiero, G. Degliesposti, G. Rastelli, Exploring the bindingsite of C-terminal Hsp90 inhibitors, J. Chem. Inf. Model. 50 (2010) 1522e1528.
[32] C. Soti, A. Racz, P. Csermely, A nucleotide-dependent molecular switch con-trols ATP binding at the C-terminal domain of Hsp90. N-terminal nucleotidebinding unmasks a C-terminal binding pocket, J. Biol. Chem. 277 (2002)7066e7075.
[33] C. Garnier, D. Lafitte, P.O. Tsvetkov, P. Barbier, J. Leclerc-Devin, J.M. Millot,C. Briand, A.A. Makarov, M.G. Catelli, V. Peyrot, Binding of ATP to heat shockprotein 90: evidence for an ATP-binding site in the C-terminal domain, J. Biol.Chem. 277 (2002) 12208e12214.
[34] A.K. Shiau, S.F. Harris, D.R. Southworth, D.A. Agard, Structural analysis of E. colihsp90 reveals dramatic nucleotide-dependent conformational rearrange-ments, Cell 127 (2006) 329e340.
[35] C. Ratzke, M. Mickler, B. Hellenkamp, J. Buchner, T. Hugel, Dynamics of heatshock protein 90 C-terminal dimerization is an important part of its confor-mational cycle, PNAS 107 (2010) 16101e16106.
