3-Deazaneplanocin A and Neplanocin A Analogues and Their Effects on Apoptotic Cell Death
Eric K. W. Tam,[a] Tuan Minh Nguyen,[a] Cheryl Z. H. Lim,[b] Puay Leng Lee,[b] Zhimei Li,[b]
Xia Jiang,[b] Sridhar Santhanakrishnan,[a] Tiong Wei Tan,[a] Yi Ling Goh,[a] Sze Yue Wong,[a]
Haiyan Yang,[a] Esther H. Q. Ong,[c] Jeffrey Hill,[c] Qiang Yu,*[b] and Christina L. L. Chai*[a, d]

3-Deazaneplanocin A (DzNep) is a potential epigenetic drug for the treatment of various cancers. DzNep has been reported to deplete histone methylations, including oncogenic EZH2 com- plex, giving rise to epigenetic modifications that reactivate many silenced tumor suppressors in cancer cells. Despite its promise as an anticancer drug, little is known about the struc- ture–activity relationships of DzNep in the context of epigenet- ic modifications and apoptosis induction. In this study, a number of analogues of DzNep were examined for DzNep- like ability to induce synergistic apoptosis in cancer cells in
combination with trichostatin A, a known histone deacetylase (HDAC) inhibitor. The structure–activity relationship data thus obtained provide valuable information on the structural re- quirements for biological activity. The studies identified three compounds that show similar activities to DzNep. Two of these compounds show good pharmacokinetics and safety profiles. Attempts to correlate the observed synergistic apoptotic activi- ties with measured S-adenosylhomocysteine hydrolase (SAHH) inhibitory activities suggest that the apoptotic activity of DzNep might not be directly due to its inhibition of SAHH.

1. Introduction

Epigenetic alterations play important roles in cancer develop- ment. These alterations, including DNA hypermethylation and chromatin modifications such as histone methylation and de- acetylation,[1] have provided new targets for therapeutic inter- ventions in cancer cells. While advancements have been made in the last decade in the development of inhibitors targeting histone deacetylation or DNA methylation for cancer therapy,[2]
there is little progress in the development of histone methyla- tion inhibitors. Aberrant histone methylation, such as histone H3 lysine trimethylation (H3K27me3) induced by oncogenic polycomb protein histone methyltransferase EZH2, has been frequently linked to tumorigenesis,[3] so there has been tre-
mendous effort directed towards the development of small molecules that target EZH2.
We previously reported 3-deazaneplanocin A (DzNep) (Figure 1), an S-adenosylhomocysteine (AdoHcy) hydrolase (SAHH) inhibitor that can effectively deplete cellular levels of

[a] Dr. E. K. W. Tam, Dr. T. M. Nguyen, Dr. S. Santhanakrishnan, T. W. Tan, Y. L. Goh, S. Y. Wong, H. Yang, Prof. C. L. L. Chai
Institute of Chemical & Engineering Sciences
Agency for Science, Technology & Research (A*STAR)
8 Biomedical Grove, Neuros #07-01, Singapore 138665 (Singapore) E-mail: [email protected]
[b] C. Z. H. Lim, P. L. Lee, Z. Li, X. Jiang, Dr. Q. Yu
Cancer Therapeutics & Stratified Oncology, Genome Institute of Singapore Agency for Science, Technology and Research (A*STAR)
60 Biopolis Street, #02-01, Singapore 138672 (Singapore) E-mail: [email protected]
[c] E. H. Q. Ong, Dr. J. Hill Experimental Therapeutics Centre
Agency for Science, Technology and Research (A*STAR)
31 Biopolis Way, Nanos Level 3, Singapore 138669 (Singapore) [d] Prof. C. L. L. Chai
Department of Pharmacy, National University of Singapore 18 Science Drive 4, Singapore 117543 (Singapore)
E-mail: [email protected]
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/cmdc.201402315.

Figure 1. Structures of 3-deazaneplanocin A (DzNep) and neplanocin (NPA).

the EZH2 complex, inhibit H3K27me3 and other histone meth- ylations, and induce apoptosis in cancer cells, but not in normal cells, through the reactivation of many silenced tumor suppressors in cancer cells.[4] Although this compound does not appear to be specific for EZH2–H3K27me3, its anticancer activity associated with the depletion of EZH2 has been dem- onstrated in vitro and in vivo in many cancer types, including breast, lung, liver, prostate, leukemia.[5] Recent efforts have led to the discovery of a number of selective EZH2 inhibitors, for example, EPZ005687,[6] GSK126,[7] and EI1.[8] However, these in- hibitors seem to have anticancer activity only in B cell lympho- mas that carry activating mutations of EZH2, but not in epithe- lial tumors in which enhanced EZH2 activity is mainly caused

ti 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemMedChem 0000, 00, 1 – 11 &1&

by increased gene expression. Intriguingly, a recent study showed that although inhibitors of EZH2, such as GSK126, have robust activity in the depletion
of H3K27me3 in epithelial tumors cells, they have little effect on cancer cell growth.[9]
Moreover, a peptide that induces the depletion of EZH2 protein complex is sufficient to inhibit cell growth.[10] This raises a possi- bility that targeting the EZH2 complex, rather than inhibiting H3K27me3, is required for block- ing the EZH2-mediated onco- genic effect. Given the remark- able activities of DzNep in the depletion of EZH2 and in the in- hibition of tumor growth, we were intrigued with the possibili- ties of examining DzNep deriva- tives so as to obtain a better un- derstanding of the structure–ac- tivity relationships.
In this study, we report a com- prehensive cell-based structure– activity relationship study where we screened for DzNep-like ac- tivities for the induction of apoptosis in cancer cells. We also report the IC50 values of se- lected DzNep analogues against human SAHH enzymes with the objective of determining if a cor- relation exists between SAHH in- hibition and the ability of the compounds to induce apoptosis.

2.Results
2.1Synthesis of the target compounds
Our early studies have shown that the natural product, nepla- nocin A (NPA) (Figure 1) is able to induce the same level of apoptosis as DzNep in cellular models (unpublished observa- tions). We thus designed and synthesized a total of 40 com-

Figure 2. Compounds used for structure–activity determination. Series A includes variations in substituents at- tached to the carbocyclic ring system of NPA. Series B and D include variations in the carbocyclic ring system. Series C includes variations in the heterocyclic ring attached. For D7, D8, D14, D18, and D19, the relative stereo- chemistry is shown.

pounds based on variations in the structure of DzNep and NPA (Figure 2). They include 1) variations in the substituents at- tached to the carbocyclic ring system of NPA (series A, com- pounds A1–A10), 2) variations in the carbocyclic ring system (series B and D, compounds B1–B4, D1–D19), 3) variations in
the nature of the heterocyclic ring attached to the carbocyclic ring (series C, compounds C1–C9). Only the synthesis of select- ed compounds is discussed here. The full experimental details on the synthesis of the target compounds can be found in the Supporting Information.

ti 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemMedChem 0000, 00, 1 – 11 &2&

Scheme 1. Key coupling step to target compounds.

In general, the target compounds were synthesized by react- ing the carbocyclic scaffold (“core”) with the appropriate heter- ocyclic bases under Mitsunobu or SN2 conditions (Scheme 1). For compounds in series A and C, the core is comprised of functionalized cyclopentene 1 and 2, while the compounds in series B and D were either commercially available or were syn- thesized following literature procedures (see Supporting Infor- mation).
Cyclopentenol 1 (Z = H) was synthesized in eight steps start- ing from d-ribose,[11] and subsequent transformations gave protected NPA 3 in good yields. With the protected NPA in hand, protected precursors of A2–A10 were accessed from this advanced intermediate as summarized in Scheme 2. Com- pound 4 was directly accessed from 3 through the reaction of the alcohol with 1,1’-thiocarbonyl diimidazole and tributyltin hydride, while the synthesis of 5 was achieved using dimethy- laminosulfur trifluoride (DAST). Methylation, ethylation, and benzylation of 3 gave the O-analogues 6, 7, and 9, respectively. Phenoxy derivative 8 was obtained under Mitsunobu condi- tions, while methyl ester 10 was accessed via a three-step pro-

Scheme 2. Reagents and conditions : a) 1. 1,1’-thiocarbonyl diimidazole, RT, 18 h, CH2Cl2, 56%, 2. tributyltinhydride, 110 8C, 8h, 56%; b) DAST, RT, 3 h, 72%;
c) MeI, NaH, DMF, RT, 16 h, 54%; d) EtI, NaH, DMF, RT, 16 h, 25%; e) PhOH, PPh3, DIAD, THF, RT, 18 h, 78%; f) BnBr, NaH, DMF, RT, 16 h, 39%; g) 1. Dess–Martin periodinane, CH2Cl2, RT, 17 h, 95%, 2. oxone, DMF, RT, 48 h, 74%, 3. (trimethylsilyl)diazomethane, MeOH/toluene (2:3), RT, 48 h, 47%; h) 11. Dess–Martin peri- odinane, CH2Cl2, RT, 17 h, 95%; 2) MeMgBr, THF, 0 8C !RT, 16 h, 86%.

ti 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemMedChem 0000, 00, 1 – 11 &3&

cedure (oxidation to the aldehyde with Dess–Martin periodi- nane, followed by oxidation to the acid and esterification) from alcohol 3. Compounds 11 and 12 were derived from the same aldehyde by treatment with methyl magnesium iodide. In all cases, global deprotection of the tert-butyloxycarbonyl (Boc) and acetonide protecting groups under acidic conditions afforded the target compounds.
For the synthesis of cyclopentenol 15, d-ribose was convert- ed, through a sequence of reactions, to a diastereomeric mix- ture of dienes 13 following modified procedures (Scheme 3).[12]

Scheme 3. Reagents and conditions : a) 1. 2nd-gen. Hoveyda–Grubbs catalyst, CH2Cl2, RT, 4 h, 2. pyridinium chlorochromate, Celite, CH2Cl2, RT, 4 h (Overall yield of 55% from the aldehyde; see Supporting Information for details);
b) NaBH4, CeCl3, 0 8C !RT 85–95%.
Ring-closing metathesis followed by pyridinium chlorochro- mate (PCC) oxidation gave enone 14 in reasonable yields. The ketone functionality of the enone was then reduced under Luche’s conditions to afford a single stereoisomer (15) in high isolated yields. With this alcohol 15 in hand, coupling with N,N-diprotected adenine followed by deprotection gave com- pound A1 in good yields (Scheme 4). Subsequent hydrogena- tion of A1 gave compound D9 in quantitative yields.

Scheme 5. Reagents and conditions : a) I2, pyridine, CH2Cl2/THF (10:1), RT, 18CHh, 75%; b) NaBH4, CeCl3, MeOH, 0 8C !RT, 1 h, 98%; c) TBDPSCl, imidazole,
2Cl2/THF (10:1), RT, 18 h, 93%; d) nBuLi, NFSI, THF, ti 60 8C, 4 h; e) 1. TBAF, THF, RT, 18 h, 2. TsCl, Et3N, DMAP, CH2Cl2, RT, 18 h; f) adenine, NaH, DMF,
80 8C, 18 h, 28% over four steps for 23 ; g) 2 n HCl, RT, 48 h, 79%; h) 10% Pd/C, H2 (50 psi), MeOH, RT, 16 h, 73%; i) 2 n HCl, RT, 48 h, 95%.

Scheme 4. Reagents and conditions : a) 1. PPh3, DIAD, diBoc adenine, THF, 80%, 2. 10% HCl in H2O/MeOH (1:1), 90%; b) H2, 10% Pd/C, H2O, RT, 24 h, 95%.

The synthesis of D4 and D16 are summarized in Scheme 5. Using enone 14 as described in Scheme 4, iodination followed by reduction of ketone 16 provided alcohol 17 in good yields. The free hydroxy group of 17 was protected as its tert-butyldi- phenylsilyl (TBDPS) ether 18. Fluorination of iodo-alkene 18 with excess N-fluorobenzenesulfonimide (NFSI) and BuLi fur- nished an inseparable mixture of compounds 19 and 20 in a ratio of 3:1, which was used as is in the subsequent steps. Deprotection of the TBDPS groups of the mixture of 19 and 20
using tetrabutylammonium fluoride (TBAF) provided an insepa- rable mixture of alcohols. Tosylation of the mixture to give 21 and 22, followed by alkylation with adenine in N,N-dimethyl- formamide (DMF), yielded a mixture of 23 and 24, which was separable by column chromatography. The reduction of the double bond of 23 in the presence of Pd/C under high pres- sure (50 psi) provided 25 in 73% yield. The stereochemistry of the fluorine substituent and the basic skeleton of 25 was un- ambiguously characterized by X-ray crystallography (see Sup- porting Information). Hydrolysis of 25 in acidic medium gave D16 in good yields. Similarly, hydrolysis of 23 furnished D4 in good yields.

ti 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemMedChem 0000, 00, 1 – 11 &4&

Compounds D5, D6, D15, and D17 were synthesized from versatile iodocyclopentenes 17 and 18 through palladium-cata- lyzed reactions followed by a sequence of transformations sim- ilar to that described previously (Scheme 6; see Supporting In- formation for details). The stereochemistry of the vicinal hy- droxy groups of the cyclopentane ring in compounds D15 and D17 is presumed to be syn, based on observation made with D16.

placement of the carbocyclic ring of NPA with other carbocy- clic core structures (series B) indicated that the free hydroxy substituents are critical for DzNep-like apoptotic activity as is the stereochemistry of the hydroxy groups, as B1 and B2 do not show any synergistic activity. Surprisingly, saturation of the double bond (cyclopentane vs cyclopentene; B3) retained the synergistic apoptotic activity.
In view of the simplicity of the structure of A1, further struc- tural modifications were centered on variations of this deriva- tive. Compounds in series C examined the effect of varying the heterocyclic bases, and none of the compounds tested in this series show any significant DzNep-like apoptotic activity. Series D further examined the effect of varying the carbocyclic core. Consistent with the observations made previously, saturation of the double bond of A1 (D9) retained the apoptotic activity. The vinyl fluoride derivative D4 and the saturated compound D16 showed synergistic apoptotic cell death, while all other variations of the carbocyclic core resulted in loss of apoptotic activity. This indicates the sensitivity of the carbocyclic core in inducing apoptosis. As observed with compounds in series B, the location, stereochemistry, and presence of both the secon- dary hydroxy groups are critical for apoptosis. In addition, the ring size is also important. From these studies, three com- pounds D4, D9, and D16 have been identified as showing syn- ergistic apoptotic activity comparable with DzNep.
A secondary screen with selected positive hits was carried

Scheme 6. Synthetic routes to D5, D6, D15, and D17. Reagents and condi- tions : a) PdCl2(PPh3)2, AlMe3, THF, 50 8C, 18 h; b) Phenylboronic acid, PdCl2- (PPh3)2, Na2CO3, toluene/MeOH, 50 8C, 18 h. (See Supporting Information for details).

2.2Cell-based structure–activity relationship studies
We have previously shown that DzNep is able to synergize with histone deacetylase (HDAC) inhibitor, trichostatin A (TSA), to potently induce apoptosis in colon cancer DLD1 cells.[13] We also showed that DzNep potently induces apoptosis upon acti- vation of transcription factor E2F1 in colon cancer cell HCT116.[14] We thus used these two independent in vitro cellu- lar systems to determine the cellular activities of the DzNep- like compounds described here. Initially, DLD1 cells were used to screen for DzNep analogues that can increase apoptosis in combination with TSA. DLD1 cells were treated with DzNep an- alogues (5 mm) for 48 h followed by the HDAC inhibitor TSA (150 nm) for an additional 24 h before harvesting for apoptosis assessment using fluorescent-activated cell sorting (FACS) anal- ysis of DNA content. The positive hits were compounds that show increased apoptotic cell death on drug combination, that is to say, where the apoptotic cell death is greater than the sum of that caused by treatment with the DzNep analogue or TSA alone. The results are summarized in the Supporting In- formation.
From the screening studies, the effect of substitution (R) at the C-4 position of the carbocyclic ring of NPA was investigat- ed. Positive hits are compounds A1–A4, A9, and A10 (i.e., compounds possessing H, Me, CH2F, CH2OMe, CH(OH)CH3 sub- stituents instead of the CH2OH substituent present in NPA). Re-
out with HCT116 cells in which E2F1 is fused to the 4-hydroxy- tamoxifen (4-OHT)-responsive ligand-binding domain of the es- trogen receptor (ER) to form an ER–E2F1 fusion protein, which becomes activated after addition of 4-OHT. We have previously shown that DzNep, upon addition of 4-OHT (to activate E2F1) in cell culture, potently induced apoptosis in this system, while only a modest level of apoptosis was induced in the absence of 4-OHT.[14]
As shown in Figure 3, in HCT116 ER-E2F1 cells, 4-OHT and DzNep alone induced approximately 13% and 20% of cells in apoptosis, respectively, as determined by counting cells in sub- G1 phase using FACS analysis, while the combination of 4-OHT and DZNep resulted in 54% of cells in apoptosis. Compounds A4, D4, D9, and D16 exhibited comparable activities to DzNep, resulting in synergistic apoptosis of 45%, 47%, 49%, and 47%,

Figure 3. In vitro assessment of DzNep analogues. p53-null HCT116-ER– E2F1-expressing cells were treated with DzNep (5 mm) or DzNep analogues A4, D4, D9 or D16 in the presence or absence of 4-OHT (300 nm). After
72 h, cells were harvested, and cell death was assessed by propidium iodide staining using fluorescent-activated cell sorting (FACS).

ti 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemMedChem 0000, 00, 1 – 11 &5&

respectively, when combined with 4-OHT. We also showed that the DzNep analogues have the capacity to deplete histone methylations including H3K27me3, H3K4me3, and H3K20me3, though D9 seems to be more selective on H3K27me3 as shown in Figure 4.

can be concluded that neither D9 and D16 show any signifi- cant inhibition of the hERG channel.

2.5. Antitumor effects of D9 in a xenograft tumor mouse model
To evaluate the anticancer effect of D9 in vivo, an HCT116 sub- cutaneous xenograft tumor mouse model was used. In this model, 5ti106 HCT116 cells were injected into Balb/c nude mice subcutaneously (s.c.), and when the tumors reached an average of 200 mm3 in size, D9 was administered through in- traperitoneal (i.p.) injection. As shown in Figure 5a D9 adminis- tration at 30 or 80 mgkgti 1 five days a week for two weeks re- sulted in marked decrease in tumor growth in a dose-depen- dent manner. During the course of drug treatments, the body weight of mice was monitored daily. D9 did not give rise to obvious signs of toxicity, and only a minimum loss of body weight was observed (Figure 5b). These data suggest that D9

Figure 4. Western blot showing the effect of DzNep analogues on histone methylation. HCT116 cells were treated with increasing concentrations of in- dicated compounds (0.5, 1, 2.5 and 5 mm) for 72 h before being harvested
for Western blot analysis.
has significant antitumor effects while at the effective doses it

2.3SAHH inhibition by selected compounds
A total of 16 compounds (DzNep, NPA, A1, A3–A6, A9–10, D4– 6, D9, D15–17) were tested for inhibition of human SAHH using an in vitro SAHH activity assay. Although the IC50 or Ki values of some of these compounds have already been report- ed in the literature, comparison of the literature values proved to be difficult due to the different assay platforms and the spe- cies and form of SAHH used (recombinant or purified native). A summary of the results is shown in Table 1. From these stud- ies, it is noted that NPA and DzNep are potent inhibitors of SAHH, while D9 is approximately sixfold less potent than DzNep. Interestingly D16 was 18 fold less potent than D9, while the activity of D4 and D16 are similar. The ranking of the compounds in terms of the IC50 values with SAHH inhibition is NPA (most potent) > A3 > DzNep > D9 > A9–A4 > A10 > A6 > A5 > D17 > D16 ti D4 ti D15 > A1 > D5 ti D6 (least potent).

2.4In vitro ADME and hERG studies
Compounds D9 and D16 were selected for in vitro ADME eval- uation and hERG studies. Measurement of the log D showed that both compounds have comparable values (ti 1.6 at pH 3.0 and ti0.5 at pH 7.4). In addition, the solubility of the com- pounds at pH 2 and 6.5 did not change and is approximately 300 mm for both D9 and D16. The half-lives of D9 and D16 with human liver microsomes were measured to be 2636 min and 1029 min, respectively, while that with male rat liver micro- somes and male mouse liver microsomes were 183 min and 307 min for D9 and 836 and 982 min for D16, respectively.
hERG safety studies were performed, and the IC50 values of both D9 and D16 were found to be greater than 100 mm using a FluxOR hERG fluorescence assay, and greater than 10 mm using the electrophysiological hERG test. From these data, it

Figure 5. Effects of D9 in vivo on a) HCT116 xenograft tumor growth and b) body weight change of mice. Group I: 10% DMSO (^); Group II: D9
30 mgkgti 1 (&); Group III: D9 80 mgkgti 1 (~).

ti 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemMedChem 0000, 00, 1 – 11 &6&

Table 1. Inhibition (IC50) of human S-adenosylhomocysteine hydrolase (SAHH) by DzNep derivatives.[a]

Structure IC50 [mm] Structure IC50 [mm]

DzNep
0.23 ti 0.07 [Lit: Ki=23 nm,
human SAHH][25]
NPA
0.04 ti 0.01
[Lit: IC50 = 0.82 mm,
human SAHH][26]

A4

2.22 ti 0.24
[Lit: IC50 > 10 mgmLti 1, rabbit erythrocyte][19]

A5

13.1[b]

A6

7.47 ti 4.30

A3

0.13 ti 0.04
[Lit: IC50 > 10 mgmLti1, rabbit erythrocyte][19]

A9

1.71 ti 0.24
[Lit: IC50 = 0.2 mgmLti 1, rabbit erythrocyte][19]

A9/A10 (15:85)[c]

3.46 ti 0.87

A1

44.59 ti 13.34 [Lit: Ki=35 nm, bovine liver][27]

D4

24.39 ti 0.40

D9

1.35 ti 0.35 [Lit: IC50 = 9 mm, human SAHH][28]

D16

25.18 ti 2.17

D5 > 200 D15 25.7[b]

D6 > 200 D17 20.9[b]

[a] Inhibition was determination using an enzyme assay with recombinant expressed human SAHH; data represent the mean ti SD of n = 2 independent ex- periments, unless specified otherwise. [b] Data from a single independent assay. [c] Tested as a mixture of compounds.

is well tolerated in the recipient mice (Figure 5). The pharma- cokinetic profile of D9 in Sprague–Dawley rats through oral (p.o.) administration (5 mgkgti 1) also suggested that D9 is readily absorbed and has a half-life of 5 h.

3.Discussion
Our studies above have delineated some of the critical struc- tural features required for DzNep-like apoptotic activity (Figure 6). In brief, the structure–activity relationship study identified that the 2,3-dihydroxycarbocyclic core of DzNep, whether saturated or unsaturated, is needed for synergistic ac-
tivity (Figure 6). The heterocyclic base at the C-1 position of Figure 6. Critical structural features for DzNep-like apoptotic activity.

ti 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemMedChem 0000, 00, 1 – 11 &7&

the carbocycle can either be adenine or deazaadenine, but all the other heterocyclic combinations led to loss in activity. Thus compound A1 and D9 contain the minimum pharmacophore required for synergistic apoptotic activity. When considering the structure–activity relationships of DzNep and NPA, it is ap- parent that variation in the substituents at the C-4 position cannot be easily rationalized. The presence of a substituent is not necessary for synergistic apoptotic activity, as seen with compound A1. There is also some tolerance for small substitu- ents at the C-4 position of the carbocyclic ring. Interestingly, a secondary alcohol at the C-4 position does not lead to large losses in biological activity. The presence of substituents at the C-5 position of the carbocyclic ring causes large losses in syn- ergistic apoptotic activity—the exception is when a fluorine atom replaces the hydrogen atom at C-5 of A1 and D9.
SAHH is an important enzyme that catalyzes the hydrolysis of S-adenosylhomocysteine (AdoHcy) to adenosine and l-ho- mocysteine.[15] Accumulation of AdoHcy through inhibition of SAHH can, in turn, lead to inhibition of methyltransferases and thus indirectly affect the epigenetic mechanisms of a cell. DzNep and NPA have long been recognized as potent inhibi- tors of SAHH. Indeed, early studies have focused on the devel- opment of these and related compounds as antivirals through the inhibition of SAHH.[16] The mechanisms of SAHH inhibition are suggested to occur either via a reversible Type I mecha- nism through the oxidation of the cofactor NAD + to NADH, or via an irreversible Type II mechanism of covalent binding to the active site by an inhibitor with a nucleophilic residue.[17] Of the selection of the target compounds measured for SAHH ac- tivity, the most potent SAHH inhibitor was NPA (IC50 = 40 nm), followed by the fluorinated compound A3 (IC50 = 130 nm), then DzNep (IC50 = 230 nm). We note that these values for NPA and DzNep differ from those reported in the literature for
[16b,18] The compounds that were tested from ser- ies A showed that SAHH inhibition is sensitive to modifications of the substituents at the C-4 position. The stereoisomers of the secondary alcohol A9 showed a 1.5-fold difference in SAHH activities as compared with the enriched mixture of A10 (containing 15% A9). From the literature, a large difference in SAHH activities between A9 and A10 was reported, in favor of A9, as measured with rabbit SAHH.[19] For the selected com- pounds studied in series D, the most potent compound is D9 which showed very different activity in comparison with the unsaturated compound A1. The vinyl fluoride D4 was the best compound amongst the unsaturated compounds in series D, and D4 has been reported to inhibit SAHH via the Type II mechanism.[20] The saturated compounds D15–D17 have com- parable SAHH activity.
Our SAHH studies identified NPA, DzNep, methyl ether A4, fluorinated derivative A3, alcohol A9, and D9 as the best inhib- itors of SAHH. Broadly speaking, these compounds are also potent inducers of apoptosis. However there is no discernible direct relationship between SAHH inhibition and synergistic apoptotic activities. Specifically compounds that have reasona- ble SAHH inhibitory activities (e.g., D15) do not display syner- gistic apoptotic activity.

4. Conclusions

The toxicities and the short half-lives of DzNep and NPA have long been known, and this is in part due to the rapid phos- phorylation by adenosine kinase.[20] Through these studies, we have identified two compounds (D9 and D16) that have longer half-lives than NPA and DzNep. In addition, animal stud- ies with D9 show that it has lower toxicity than DzNep and is effective at shrinking tumors. Thus, we have successfully identi- fied at least two compounds, which are potentially superior to DzNep and NPA, for further development as chemotherapeu- tics for epigenetic therapy. Future mechanistic studies will un- doubtedly shed more information on the mode of action of these compounds.

Experimental Section3-deazaneplanocin A
Chemistry
Unless otherwise specified, materials were purchased from com- mercial suppliers and used without further purification. Solvents for reactions were taken from a Glass Contour solvent purification system under nitrogen. 1H NMR and 13C NMR spectra were ac- quired on a Bruker400. UltraShield spectrometer operating at 400 MHz and 100 MHz, respectively, unless specified otherwise. Chemical shifts (d) are given in ppm relative to respective residual solvent peaks (CD3OD: d = 3.31 ppm; [D6]DMSO: d = 2.50 ppm; D2O: d = 4.79 ppm) except for those recorded in CDCl3, which were referenced to tetramethylsilane (TMS) or residual solvent peak at d = 7.26 ppm. For 13C NMR spectra, chemical shifts are given rela- tive to the respective residual solvent peak (CD3OD: d = 49.0 ppm; [D6]DMSO: d = 39.5 ppm; CDCl3 : d = 77.2 ppm). Where necessary, the signals of novel compounds were assigned by 1DNOE differen- ces and/or 2DNMR techniques: 1H–1H COSY, 13C–1H HMQC, and 13C–1H HMBC. These experiments were performed using standard Bruker microprograms. Low-resolution and high-resolution electron impact mass spectra (EIMS) were measured using a Finnigan MAT95XP double-focusing mass spectrometer. Low-resolution elec- trospray ionization mass spectra (ESIMS) were recorded using Waters Quattro MicroTM API, and high-resolution mass spectra (HRMS) were obtained using an Agilent6210 Time-of-Flight LC–MS system. Flash column chromatography was conducted manually using Merck silica gel 60 (35–70 mm) with an overpressure of 200 mbars. Amination reactions using 33% aqueous ammonia were carried out using a high-pressure reactor. Elemental analysis was performed using a EuroEA3000 series CHNS analyzer.
Compounds B1,[22] B2,[23] and D12[24] were synthesized following lit- erature procedures, while B4 is commercially available. Detailed procedures and characterization data for all other compounds are provided in the Supporting Information.

Biology
In vitro evaluation of DzNep analogues
Induction of apoptosis in HCT116 cells : p53 Knock out (KO) HCT116 cells that express inducible E2F1 (ER–E2F1) were used to assess the ability of DzNep analogues to induce E2F1-dependent apoptosis.[14]
To activate E2F1, 4-hydroxytamoxifen (4-OHT) (300 nm) was added to the tissue culture medium. Cell apoptosis induced by DzNep an- alogues was measured in the presence or absence of 4-OHT for

ti 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemMedChem 0000, 00, 1 – 11 &8&

72 h by analysis of DNA content via FACS. To measure the ability of DzNep analogues to synergize with a histone deacetylase inhibitor (HDACi) for apoptosis induction, DLD1 cells were treated with DzNep analogues (5 mm) for 48 h followed by HDACi trichostatin A (TSA) (150 nm) for an additional 24 h before being harvested for FACS analysis. To assess the apoptosis induction, cells were har- vested and fixed in 70% ethanol. Fixed cells were stained with pro- pidium iodide (50 mgmLti 1) after treatment with RNase (100 mgmLti 1). The FACS analysis of stained cells was done in FACS- calibur flow cytometer (Becton Dickinson Instrument, San Jose, CA, USA). Cell cycle fractions were quantified using the CellQuest soft- ware (Becton Dickinson). Apoptosis was determined by measuring the DNA content of cells in sub-G1.
Production and purification of recombinant SAHH : The glutatione S- transferase (GST)-tagged SAHH in the pDEST 565 vector was ex- pressed in Escherichia coli BL21 (DE3) pLysS and inoculated over- night in LB medium with chloramphenicol (34 mgmLti 1 ) and ampi- cillin (100 mgmLti 1) at 37 8C with orbital shaking at 200 rpm. The culture was then diluted 1:250 in LB medium with chloramphenicol and ampicillin, and maintained at 30 8C until the OD600 value reached 0.5–0.6 arbitrary units. After the addition of isopropyl-b-d- 1-thiogalactopyranoside (400 mm), the culture was incubated for an additional 16 h and centrifuged at 6000 rpm for 40 min. at 4 8C. The bacterial pellet was then collected by centrifugation and lysed by sonication in buffer containing Na3(PO4) (10 mm), NaCl (150 mm), DNAseI (5 mgmLti1), lysozyme (0.625 mgmLti1), 1X pro- tease inhibitor (pH 7.4), and 2-mercaptoethanol (5 mm). The over- expressed protein was purified using the Bio-Scale Mini Profinity GST 5 mL cartridges (BioRad) followed by concentration and desalt- ing in Tris buffer (2 mm, pH 8.0) with NaCl (100 mm). The concen- trated SAHH–GST was cleaved with Tobacco Etch Virus (TEV) pro- tease at 1 mg TEV per 50 mg GST-tagged protein in Tris buffer (100 mm, pH 8.0) with NaCl (100 mm), and dithiothreitol (DTT) buffer (2 mm). The cleaved GST was removed using glutathione Se- pharose 4B beads (GE Healthcare) after 1 h incubation in Tris buffer (10 mm, pH 7.5) with NaCl (50 mm).
SAHH activity assay : The thiol-containing products of the SAHH-cat- alyzed reaction were detected in a 96-well format through the flor- escent thiol detection reagent, ThioGlo1 (Calbiochem). Freshly pre- pared SAHH (5 mL; 100 ng mLti1 in 100 mm Tris pH 7.5) was added to inhibitor/DMSO control (5 mL) and freshly prepared assay buffer (15 mL; 100 mm Tris pH 7.5 with 3 mm DTT, 150 mm NAD, and 3 mm EDTA) and incubated at 37 8C for 30 min. After incubation, SAHH (Sigma, 5 mL of 750 mm, freshly prepared in 100 mm Tris, pH 7.5) was added and further incubated at 37 8C for 10 min. 500 mm. Thio- Glo1 freshly diluted in Tris buffer (100 mm, pH 7.5) was added and incubated at 37 8C for another 15 min. The fluorescence signal was detected by Safire2 (Tecan) with a 380 nm excitation and 510 nm emission filter. IC50 measurements were performed using GraphPad Prism version 5.00 for Windows (GraphPad software, San Diego, CA, USA).

In vivo antitumor assessment of DzNep analogues
All animal studies were conducted in compliance with animal pro- tocols approved by the A*STAR–Biopolis Institutional Animal Care and Use Committee (IACUC) of Singapore.
The female athymice BALB/c nude mice (5–8 week-old) were housed in the Biological Resource Centre. Mice were implanted subcutaneously in flank with 5ti10 6 cells of HCT116 parental
human colon carcinoma. When tumors reached ti 200 mm3, the mice were divided into four groups (10 mice per group) and were

treated with vehicle or DzNep analogue D9 at 30 and 80 mgkgti1 by intraperitoneal (IP) injection for 14 days. Tumor growth and the whole body weight changes of mice were monitored every other day.
Student’s t-test is used to determine the statistical significance of tumor volumes and body weight changes between groups. Statis- tical analyses are conducted at a p level of 0.05. SPSS was used for all statistical analyses and graphic presentations.

Acknowledgements

The authors thank the Agency for Science, Technology and Re- search (A*STAR), Singapore (CCOG01-001-2008 and ETPL/10- FS0002) for financial support of this project.

Keywords: deazaneplanocin A · epigenetics · neplanocin A · S- adenosylhomocysteine hydrolase (SAHH) · structure–activity relationships

[1]K. Helin, D. Dhanak, Nature 2013, 502, 480–488.
[2]a) J. S. Garcia, N. Jain, L. A. Godley, Onco. Targets Ther. 2010, 3, 1–13; b) O. A. O’Connor, M. L. Heaney, L. Schwartz, S. Richardson, R. Willim, B. MacGregor-Cortelli, T. Curly, C. Moskowitz, C. Portlock, S. Horwitz, A. D. Zelenetz, S. Frankel, V. Richon, P. Marks, W. K. Kelly, J. Clin. Oncol. 2006, 24, 166–173; c) R. L. Piekarz, R. Frye, M. Turner, J. J. Wright, S. L. Allen, M. H. Kirschbaum, J. Zain, H. M. Prince, J. P. Leonard, L. J. Geskin, C. Reeder, D. Joske, W. D. Figg, E. R. Gardner, S. M. Steinberg, E. S. Jaffe, M. Stetler-Stevenson, S. Lade, A. T. Fojo, S. E. Bates, J. Clin. Oncol. 2009, 27, 5410–5417.
[3]a) F. Crea, L. Fornaro, G. Bocci, L. Sun, W. L. Farrar, A. Falcone, R. Danesi, Cancer Metastasis Rev. 2012, 31, 753–761; b) F. Crea, E. M. Hurt, L. A. Mathews, S. M. Cabarcas, L. Sun, V. E. Marquez, R. Danesi, W. L. Farrar, Mol. Cancer 2011, 10, 40; c) J. A. Simon, C. A. Lange, Mutat. Res. 2008, 647, 21–29.
[4]J. Tan, X. Yang, L. Zhuang, X. Jiang, W. Chen, P. L. Lee, R. K. Karuturi, P. B. Tan, E. T. Liu, Q. Yu, Genes Dev. 2007, 21, 1050–1063.
[5]a) T. Chiba, E. Suzuki, M. Negishi, A. Saraya, S. Miyagi, T. Konuma, S. Tanaka, M. Tada, F. Kanai, F. Imazeki, A. Iwama, O. Yokosuka, Int. J. Cancer 2012, 130, 2557–2567; b) A. Hayden, P. W. Johnson, G. Packham, S. J. Crabb, Breast Cancer Res. Treat. 2011, 127, 109–119; c) J. K. Lee, K. C. Kim, Biochem. Biophys. Res. Commun. 2013, 438, 647–652; d) J. Zhou, C. Bi, L. L. Cheong, S. Mahara, S. C. Liu, K. G. Tay, T. L. Koh, Q. Yu, W. J. Chng, Blood 2011, 118, 2830–2839.
[6]S. K. Knutson, T. J. Wigle, N. M. Warholic, C. J. Sneeringer, C. J. Allain, C. R. Klaus, J. D. Sacks, A. Raimondi, C. R. Majer, J. Song, M. P. Scott, L. Jin, J. J. Smith, E. J. Olhava, R. Chesworth, M. P. Moyer, V. M. Richon, R. A. Copeland, H. Keilhack, R. M. Pollock, K. W. Kuntz, Nat. Chem. Biol. 2012, 8, 890–896.
[7]M. T. McCabe, H. M. Ott, G. Ganji, S. Korenchuk, C. Thompson, G. S. VanAller, Y. Liu, A. P. Graves, A. Della Pietra III, E. Diaz, L. V. LaFrance, M. Mellinger, C. Duquenne, X. Tian, R. G. Kruger, C. F. McHugh, M. Brandt, W. H. Miller, D. Dhanak, S. K. Verma, P. J. Tummino, C. L. Creasy, Nature 2012, 492, 108–112.
[8]W. Qi, H. Chan, L. Teng, L. Li, S. Chuai, R. Zhang, J. Zeng, M. Li, H. Fan, Y. Lin, J. Gu, O. Ardayfio, J. H. Zhang, X. Yan, J. Fang, Y. Mi, M. Zhang, T. Zhou, G. Feng, Z. Chen, G. Li, T. Yang, K. Zhao, X. Liu, Z. Yu, C. X. Lu, P. Atadja, E. Li, Proc. Natl. Acad. Sci. USA 2012, 109, 21360–21365.
[9]Z. N. Wee, Z. Li, P. L. Lee, S. T. Lee, Y. P. Lim, Q. Yu, Cell Rep. 2014, 8, 204–216.
[10]W. Kim, G. H. Bird, T. Neff, G. Guo, M. A. Kerenyi, L. D. Walensky, S. H. Orkin, Nat. Chem. Biol. 2013, 9, 643–650.
[11]a) J. H. Cho, D. L. Bernard, R. W. Sidwell, E. R. Kern, C. K. Chu, J. Med. Chem. 2006, 49, 1140–1148; b) B. Y. Michel, P. Strazewski, Tetrahedron 2007, 63, 9836–9841.
[12]H. Osato, I. L. Jones, A. Chen, C. L. L. Chai, Org. Lett. 2010, 12, 60–63.

ti 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemMedChem 0000, 00, 1 – 11 &9&

[13]X. Jiang, J. Tan, J. Li, S. Kivimae, X. Yang, L. Zhuang, P. L. Lee, M. T. Chan, L. W. Stanton, E. T. Liu, B. N. Cheyette, Q. Yu, Cancer Cell 2008, 13, 529– 541.
[14]Y. Zhao, J. Tan, L. Zhuang, X. Jiang, E. T. Liu, Q. Yu, Proc. Natl. Acad. Sci. USA 2005, 102, 16090–16095.
[15]M. A. Turner, X. D. Yang, D. Yin, K. Kuczera, R. T. Borchardt, P. L. Howell, Cell Biochem. Biophys. 2000, 33, 101–125.
[16]a) P. K. Chiang, Pharmacol. Ther. 1998, 77, 115–134; b) E. De Clercq, Nu- cleosides Nucleotides Nucleic Acids 2005, 24, 1395–1415; c) C. K. Tseng, V. E. Marquez, R. W. Fuller, B. M. Goldstein, D. R. Haines, H. McPherson, J. L. Parsons, W. M. Shannon, G. Arnett, M. Hollingshead, J. S. Driscoll, J. Med. Chem. 1989, 32, 1442–1446.
[17]M. S. Wolfe, R. T. Borchardt, J. Med. Chem. 1991, 34, 1521–1530.
[18]M. Bray, J. Driscoll, J. W. Huggins, Antiviral Res. 2000, 45, 135–147.
[19]S. Shuto, T. Obara, M. Toriya, M. Hosoya, R. Snoeck, G. Andrei, J. Balzari- ni, E. De Clercq, J. Med. Chem. 1992, 35, 324–331.
[20]H. O. Kim, S. J. Yoo, H. S. Ahn, W. J. Choi, H. R. Moon, K. M. Lee, M. W. Chun, L. S. Jeong, Bioorg. Med. Chem. Lett. 2004, 14, 2091–2093.
[21]a) L. L. Bennett, Jr., B. J. Bowdon, P. W. Allan, L. M. Rose, Biochem. Phar- macol. 1986, 35, 4106–4109; b) M. Hasobe, J. G. McKee, D. R. Borcherd- ing, B. T. Keller, R. T. Borchardt, Mol. Pharmacol. 1988, 33, 713–720.

[22]K. Fukukawa, T. Ueda, T. Hirano, Chem. Pharm. Bull. 1983, 31, 1842– 1847.
[23]Y. X. Tan, S. Santhanakrishnan, H. Y. Yang, C. L. L. Chai, E. K. W. Tam, J. Org. Chem. 2014, 79, 8059–8066.
[24]S. M. Siddiqi, X. Chen, S. W. Schneller, S. Ikeda, R. Snoeck, G. Andrei, J. Balzarini, E. De Clercq, J. Med. Chem. 1994, 37, 1382–1384.
[25]R. K. Gordon, K. Ginalski, W. R. Rudnicki, L. Rychlewski, M. C. Pankaskie, J. M. Bujnicki, P. K. Chiang, Eur. J. Biochem. 2003, 270, 3507–3517.
[26]L. S. Jeong, S. J. Yoo, K. M. Lee, M. J. Koo, W. J. Choi, H. O. Kim, H. R. Moon, M. Y. Lee, J. G. Park, S. K. Lee„ M. W. Chun, J. Med. Chem. 2003, 46, 201–203.
[27]R. Snoeck, G. Andrei, J. Neyts, D. Schols, M. Cools, J. Balzarini, E. De Clercq, Antiviral Res. 1993, 21, 197–216.
[28]Y. Kitade, T. Ando, T. Yamaguchi, A. Hori, M. Nakanishia, Y. Ueno, Bioorg. Med. Chem. 2006, 14, 5578–5583.

Received: July 26, 2014
Published online on && &&, 0000

ti 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemMedChem 0000, 00, 1 – 11 &10&

FULL PAPERS

Epigenetic agents against cancer: 3- Deazaneplanocin A and neplanocin A are known to reactivate tumor suppres-
sors leading to apoptosis in cancer cells; however, their exact mode of action is not fully understood. Here, 42 ana- logues were evaluated in cell- and enzyme-based assays. No direct correla- tion between cytotoxicity and SAHH in- hibitory activity was found. The SAR studies identified two compounds with good PK and safety profiles that warrant closer investigation.
E. K. W. Tam, T. M. Nguyen, C. Z. H. Lim,
P.L. Lee, Z. Li, X. Jiang,
S. Santhanakrishnan, T. W. Tan, Y. L. Goh, S. Y. Wong, H. Yang, E. H. Q. Ong, J. Hill,
Q.Yu,* C. L. L. Chai*
&& – &&

3-Deazaneplanocin A and Neplanocin A Analogues and Their Effects on Apoptotic Cell Death

ti 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemMedChem 0000, 00, 1 – 11 &11&