The identification and optimization of a N-hydroxy urea series of flap endonuclease 1 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2004.10.086

We describe the identification and SAR of a series of low nanomolar inhibitors of FEN1, an enzyme involved in DNA repair pathways. These inhibitors have been shown to sensitize certain cell lines to DNA damage. Flap endonuclease-1 (FEN1) is a key enzyme i

Full text of DOI:10.1016/j.bmcl.2004.10.086

Bioorganic & Medicinal Chemistry Letters 15 (2005) 277–281
The identification and optimization of a N-hydroxy urea series
of flap endonuclease 1 inhibitors
L. Nathan Tumey,* David Bom, Bayard Huck, Elizabeth Gleason, Jianmin Wang,
Daniel Silver, Kurt Brunden, Sherry Boozer, Stephen Rundlett, Bruce Sherf,
Steven Murphy, Tom Dent, Christina Leventhal, Andrew Bailey, John Harrington and
Youssef L. Bennani
Athersys, Inc., 3201 Carnegie Ave., Cleveland, OH 44115, USA
Received 30 July 2004; revised 28 October 2004; accepted 30 October 2004
Available online 25 November 2004
Abstract—Flap endonuclease-1 (FEN1) is a key enzyme involved in base excision repair (BER), a primary pathway utilized by mam-
malian cells to repair DNA damage. Sensitization to DNA damaging agents is a potential method for the improvement of the ther-
apeutic window of traditional chemotherapeutics. In this paper, we describe the identification and SAR of a series of low nanomolar
FEN1 inhibitors. Over 1000-fold specificity was achieved against a related endonuclease, xeroderma pigmentosum G (XPG). Two
compounds from this series significantly potentiate the action of methyl methanesulfonate (MMS) and temozolamide in a bladder
cancer cell line (T24). To our knowledge, these are the most potent endonuclease inhibitors reported to date.
Ó 2004 Elsevier Ltd. All rights reserved.
Flap endonuclease-1 (FEN1) is a 43kDa divalent metal-
dependent nuclear enzyme that exhibits both DNA
structure-specific endonuclease activity and 50 exonuc-
lease activity. Arguably, the most well-studied aspect
of FEN1 function is its role in the cleavage of Okazaki
fragments during DNA replication.1 However, FEN1
also acts to cleave 50 DNA flaps generated during a vari-
ety of other cellular processes including double-strand
break repair,2 homologous recombination,3 and base
excision repair (BER).4 FEN1 is necessary for normal
embryonic development and homozygous mouse knock-
outs of FEN1 are lethal.5,6 Interestingly, however, yeast
knockouts of FEN1 are viable, albeit at a slower growth
rate.7
DNA damaging agents may improve the therapeutic
window of classical chemotherapeutics by lowering the
minimum effective dose.10 Recent reports describe sev-
eral small molecule inhibitors of DNA repair proteins
including poly(ADP-ribose) polymerase-1 (PARP)11
and O6-alkylguanine-DNA alkyltransferase (ATase or
MGMT).12 These inhibitors are reported to potentiate
the activity of various chemotherapeutic agents includ-
ing temozolamide13 and topotecan.11 In light of this
recent work, we embarked on a strategy to identify
selective small-molecule inhibitors of FEN1 for use as
chemopotentiating agents.
FEN1 is highly homologous to a related endonuclease,
xeroderma pigmentosum G (XPG).14 XPG is part of a
repair pathway that excises DNA containing pyrimidine
dimers, a common form of damage caused by exposure
to UV light. Defects in XPG are known to cause hyper-
sensitivity to UV light, resulting in light-induced skin
lesions and carcinoma.15 Therefore, selective inhibition
of FEN1 over XPG was a key goal of this program.
BER is an important cellular mechanism for the repair
of DNA damage caused by alkylating agents.8 The role
of FEN1 in BER is clearly exemplified in a recent report
that shows nuclease-defective FEN1 results in increased
cellular sensitivity to methyl methanesulfonate (MMS),
a potent DNA alkylating agent.9 Sensitization to
A high-throughput screen was performed with our com-
pound library in order to identify compounds with
inhibitory activity against FEN1. Selected compounds
were subsequently screened against XPG. FEN1 and
Keywords: FEN1; Endonuclease; Cancer; DNA repair.
*
Corresponding author. Tel.: +1 2164319900x498; fax: +1
2163619495; e-mail: ntumey@athersys.com
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.10.086

278
L. N. Tumey et al. / Bioorg. Med. Chem. Lett. 15 (2005) 277–281
moved with HBr), dimethoxy benzyl23 (removed with
TFA), and allyl (removed with Pd(OAc)2/PPh3).
OH
N
OH
N
M
M
M
M
O
O
O
O
O
N
O
N
O
O
Cl
O
O
N
Compound 1 (R = H) proved to be a 0.08lM inhibitor
of FEN1, significantly more potent than our original
lead. Addition of the ortho-chlorine, compound 2, in-
creased potency against FEN1 but not XPG, resulting
in a compound approximately 20-fold more potent than
the lead, ATH-0013974 (Table 1).
ATH-0013974
Flutimide
Two possible binding mechanisms
Figure 1. Structures of ATH-0013974 and Flutimide.
XPG inhibition assays were performed using a fluoro-
genic substrate consisting of a triple-labeled, double-
stranded DNA molecule containing an internal 2-nucleo-
tide gap.16 Several active compounds were identified
from the screen, with the most potent one, ATH-
0013974, having an IC50 of 0.31lM against FEN1 and
1.98lM against XPG (Fig. 1).
Thiophenes are well-known isosteres of benzene rings.
This prompted us to explore heterocyclic N-hydroxy ur-
eas exemplified by compounds 3–11. Compound 3, while
being a very potent inhibitor, was only 4-fold selective
for FEN1 over XPG. Interestingly, the addition of a
methyl group at the 4-position (compound 4) resulted
in slightly increased potency and significantly improved
selectivity (ꢀ10-fold). It is likely that the methyl group
of 4 mimics the chlorine of compound 2. Reversal of
the thiophene (compounds 6–10) resulted in potent com-
pounds provided that a hydrophobic group was present
at the 4 position (compounds 8 and 9). Hydrophobic
substitution at the 5-position generally resulted in de-
creased potency (compounds 6 and 7). Interestingly,
benzothiophene 10 is approximately 7-fold more potent
than the corresponding benzofuran 11, presumably due
to the increased hydrophobicity of sulfur (Table 1).
This compound bears an interesting resemblance to
Flutimide, a recently reported inhibitor of the influenza
endonuclease.17 We hypothesized that these compounds
inhibit the endonucleases by coordination of the two
divalent metal ions needed for catalysis, as illustrated
in Figure 1. These two metals are essential for catalytic
turnover in FEN118 and related endonucleases.19
crystal structure of Pyrococcus furiosus FEN1 shows
A
20
˚
˚
that the metal ions are ꢀ5A apart, only about 0.5A
more than the distance between the two carbonyl oxy-
gens in ATH-0013974. The hypothesized interaction of
the N-hydroxyl with an active site metal is supported
by the observation that O-substitution of ATH-
0013974 completely abolishes its FEN1 inhibitory activ-
ity (data not shown).
Alkylation of the urea nitrogen with small alkyl groups
such as methyl (12) decreased the affinity to both FEN1
Table 1. Activity of compounds 1–11
Compda Structure
IC50 (lM)
FEN1 XPG
0.310 1.98
One of our first strategies was to eliminate one of the
fused aromatic rings in order to avoid any potential
problems associated with flat, polyaromatic systems,
such as nonselective DNA intercalation and toxicity.
As such, we envisioned a related series of N-hydroxy ur-
eas that are readily accessible from ortho amino esters,
as illustrated in Scheme 1.21 Generally, the ortho amino
ester is heated in the presence of carbonyl diimidazole
and then treated with an appropriately protected hydr-
oxylamine. Base-promoted ring closure and deprotec-
tion affords the desired products.22 The ring system is
highly stable, allowing a variety of N-hydroxy protect-
ing group strategies to be employed including benzyl (re-
ATH-13974 (see Fig. 1)
OH
O
O
N
O
O
1
2
R = H
R = Cl
0.079 0.160
0.014 0.185
R
NH
OH
N
3
4
5
R,R0 = H
R = Me, R0 = H
R,R0 = –(CH2)4–
0.025 0.101
0.014 0.145
0.009 0.157
NH
R
O
S
R'
OH
N
O
OPG
N
OR
NH2
O
OR
O
6
7
8
9
R = tBu, R0 = H
R = Ph, R0 = H
R = H, R0 = Ph
0.134 0.258
0.146 0.202
0.012 0.028
O
O
NH
a
H
H
N
b
S
R
N
O-PG
NH
Ar
Ar
O
Ar
R'
R = H, R0 = SO2Ph 0.010 0.023
OH
N
,c
d
c
O
O
X
OH
N
OH
N
O
O
O
O
R
NH
10
11
X = S
X = O
0.043 0.062
0.285 0.257
NH
N
Ar
Ar
Scheme 1. Synthesis of cyclic ureas (a) CDI/toluene/heat then H2NO-
PG; (b) aq NaOH/heat; (c) TFA or HBr or Pd(OAc)2/PPh3; (d) DMF/
K2CO3/alkyl halide/heat.
a All compounds were purified by preparative HPLC and were evalu-
ated for proper identity and purity by analytical HPLC-MS and by
1H NMR.

L. N. Tumey et al. / Bioorg. Med. Chem. Lett. 15 (2005) 277–281
Table 3. Activity of furan derivatives 22–25
279
and XPG (Table 2). Unexpectedly, the potency against
Compda
Structure
IC50 (lM)
FEN1 XPG
FEN1 was regained by the addition of a benzyl group
(13) while the potency against XPG remained un-
changed. This dramatically boosted selectivity for the
desired target, from ꢀ10-fold (compound 4) to nearly
150-fold. In many cases, substituted benzyl groups
(i.e., 15–16) gave up to 240-fold selectivity.
OH
22
R = H
0.143
0.082
0.034
0.076
0.023
0.031
0.276
0.128
0.230
0.327
0.136
0.215
23a
23b
23c
23d
24
R = 4-Me
R = 4-OMe
R = 4-Br
R = 4-Ph
R = 3-OMe
O
O
N
O
NH
Reversal of the thiophene also provided exceptionally
potent and selective compounds (17–21). meta-Substi-
tuted biaryl groups (19–20) provided the best potency
in this series. Interestingly, the addition of a methyl
group at the 4-position (21) lowers the activity by over
300-fold. We believe that the presence of the methyl
group forces the conformation of the benzyl group to
be shifted toward the urea carbonyl. This conformation
is evidently not tolerated in the binding site.
R
OH
N
O
O
O
25
0.038
0.061
NH
a All compounds were purified by preparative HPLC and were evalu-
ated for proper identity and purity by analytical HPLC-MS and by
1H NMR.
A related series of furans was also investigated (Table 3).
These compounds were synthesized by the route illus-
Table 2. Activity of N-alkylated compounds 12–21
Compda
Structure
IC50, (lM)
FEN1
XPG
R = Me, R0 = Me
R = Me, R0 = Bn
R = H, R0 = Bn
0.338
0.028
0.230
0.035
4.62
3.99
2.00
8.45
OH
12
13
14
15
O
N
O
N
R = Me, R0 = Bn-4-OMe
R
R'
S
OH
N
O
O
N
16
0.018
3.04
S
OMe
O
O
OH
O
N
O
N
O
S
17
0.011
0.292
O
OH
N
O
O
O
S
N
18
19
20
R = Br
R = 3-(NHAc)-Ph
R = 4-(CONH2)-Ph
0.072
0.004
0.003
0.550
0.586
0.226
R
OH
N
O
S
N
21
R = 4–(CONH2)–Ph
1.08
30.2
R
1
a All compounds were purified by preparative HPLC and were evaluated for proper identity and purity by analytical HPLC-MS and by H NMR.

280
L. N. Tumey et al. / Bioorg. Med. Chem. Lett. 15 (2005) 277–281
OH
N
100
90
80
70
60
50
40
30
20
10
0
O
O
OEt
NH2
O
O
O
0 µM 17
3 µM 17
6 µM 17
NaO
CN
R
CN
a
c
b
NH
R
R
R
Scheme 2. Synthesis of furan-based cyclic ureas (a) ethyl formate,
NaOMe, MeOH, heat; (b) Diethyl chloromalonate/DMF then DBN/
EtOH/heat; (c) (1) CDI/toluene/heat; (2) H2NO-allyl; (3) aq NaOH/
heat; (4) Pd(OAc)2/PPh3.
trated in Scheme 2.24 Substituted acetonitriles were trea-
ted with ethyl formate then alkylated with diethyl chlo-
romalonate. Subsequent treatment with base promoted
decarboxylation and cyclization (Scheme 2). The amino
ester was converted to the N-hydroxy urea by the afore-
mentioned pathway.25 Consistent with the previous
observation of benzofuran 11 being less potent than
benzolthiophene 10, furan 22 was significantly less ac-
tive than thiophene 8. Interestingly, however, para-sub-
stituents (23a–d) seem to significantly increase the
activity. We believe that the para-biaryl motif in 23d
mimics the meta-biaryl motif in compounds 19–20. Un-
like the thiophene series, substitution on nitrogen gener-
ally had a detrimental effect on activity (data not
shown).
0µM
110µM
150µM
200µM
Concentration of MMS
Chart 1. T24 Colony forming assay of MMS +/ꢁ compound 17.
toxicity of MMS by approximately 4-fold at 3 and
6lM, respectively (Table 4). The data for compound
17 is plotted graphically in Chart 1. As can be seen,
the data is suggestive of a dose-dependent enhancement.
Compounds 16 and 17 also significantly enhanced the
effects of temozolomide, a recently approved chemo-
therapy agent that acts by methylating DNA (data not
shown). Interestingly, none of the FEN1 inhibitors
potentiated the activity of bleomycin. Unlike MMS
and temozolomide, bleomycin does not induce damage
that is repaired by the BER pathway. Therefore, no
enhancement of its activity would be expected with a
FEN1 inhibitor. Taken together, these results suggest
that these FEN1 inhibitors are sensitizing cells to dam-
age that is normally repaired by the BER pathway.
Moreover, these results confirm the FEN1 knockout
phenotype reported by Shibata and Nakamura.9
The high selectivity of compounds 15–20 against XPG,
also a metalo-endonuclease, makes it implausible that
the inhibitors act strictly via metal chelation. Moreover,
it is highly unlikely that metal chelation alone would re-
sult in sub-10nM inhibitors such compounds as 19 and
20. With these potent and selective inhibitors now in
hand, the compounds described in Tables 1–3 were
screened for their ability to sensitize cells to DNA dam-
aging agents. Previous studies have shown that FEN1
null-mutations in a T24 cell line (bladder carcinoma) re-
sulted in increased sensitivity to MMS.9 Accordingly,
the compounds above were initially screened for their
ability to sensitize T24 cells to MMS. The most potent
sensitizers were studied further for their ability to poten-
tiate the effects of other chemotherapeutic agents.
In summary, we have identified a novel and potent series
of FEN1 inhibitors. These compounds were shown to
sensitize bladder carcinoma cells to DNA damage that
is normally repaired by the BER pathway. Further stud-
ies have been undertaken to evaluate the clinical useful-
ness of this approach as a cancer therapy. To the best of
our knowledge, these compounds are approximately
100-fold more potent than any endonuclease inhibitor
reported to date. Moreover, this series of compounds
may illustrate a useful chemotype for the inhibition of
other metallo-nucleases.
The results for compounds 16 and 17 illustrate the
chemopotentiation ability of this class of molecules
(Table 4). These two compounds show little or no effect
on the growth of T24 cells at the tested concentrations,
indicating that the compounds have surprisingly little
effect on DNA replication. However, they enhance the
References and notes
1. Badmbara, R. A.; Murante, R. S.; Henricksen, L. A. J.
Biol. Chem. 1997, 272, 4647–4650.
2. Harrington, J. J.; Lieber, M. R. Embo. J. 1994, 13, 1235–
1246.
Table 4. Survival assay (colony forming) in T24 bladder carcinoma
cellsa
3. Pont, K. G.; Dawson, R. J.; Carroll, D. Embo. J. 1993, 12,
23–24.
4. Harrington, J. J.; Lieber, M. R. Genes Dev. 1994, 8, 1344–
1355.
Compd
+DMSO
100
+110lM
MMS
+150lM
MMS
+200lM
MMS
DMSO
44
22
5
5. Kucherlapati, M.; Yang, K.; Kuraguchi, M.; Zhao, J.; Lia,
M.; Heyer, J.; Kane, M.; Kunhua, F.; Russell, R.; Brown,
A. M. C.; Kneitz, B.; Edelmann, W.; Kolodner, R. D.;
Lipkin, M.; Kucherlapati, R. PNAS 2002, 99, 9924–9929.
6. Larsen, E.; Gran, C.; Saether, B. E.; Seeberg, E.; Klun-
gland, A. Mol. Cell. Biol. 2003, 23, 5346–5353.
16 (1lM)
16 (3lM)
100
86
24
11
12
6
2
0
17 (3lM)
17 (6lM)
96
87
18
11
6
5
4
5
a Represents the average of two independent experiments.

L. N. Tumey et al. / Bioorg. Med. Chem. Lett. 15 (2005) 277–281
281
7. Moreau, S.; Morgan, E. A.; Symington, L. S. Genetics
2001, 159, 1423–1433.
18. Kim, C.; Shen, B.; Park, M. S.; Olah, G. A. J. Biol. Chem.
1999, 274, 1233–1239.
8. Parikh, S. S.; Mol, C. D.; Hosfield, D. J.; Tainer, J. A.
Curr. Opin. Struct. Biol. 1999, 9(1), 37–47.
19. Doan, L.; Handa, B.; Roberts, N. A.; Klumpp, K.
Biochemistry 1999, 38, 5612–5619.
9. Shibata, Y.; Nakamura, T. J. Biol. Chem. 2002, 277, 746–
754.
20. Hosfield, D. J.; Mol, C. D.; Shen, B.; Tainer, J. A. Cell
1998, 95, 135–146.
10. For a review of chemosensitization: Gesner, T. G.;
Harrison, S. D. Annu. Rep. Med. Chem. 2002, 37, 115–
124.
21. The synthesis of a related series of compounds has been
reported Romine, J. L.; Martin, S. W.; Meanwell, N. A.;
Epperson, J. R. Synthesis 1994, 846–850.
11. White, A. W.; Almassy, R.; Calvert, A. H.; Curtin, N. J.;
Griffin, R. J.; Hostomsky, Z.; Maegley, K.; Newell, D. R.;
Srinivasan, S.; Golding, B. T. J. Med. Chem. 2000, 43,
4084–4097.
22. General protocol for the synthesis of compounds 12–21:
The ortho amino ester was refluxed in toluene (0.1M)
along with CDI (1.25equiv) for 2h. The protected
hydroxylamine (1.5equiv) was added and refluxing was
continued for 1h. The solution was concentrated, sus-
pended in EtOH and treated with 2M NaOH. The
solution was heated for 1h and then cooled. The resultant
O-protected imide generally precipitated upon cooling and
acidification with HOAc. The cyclic imide was dissolved in
DMF (0.1M) and treated with K2CO3 followed by an
alkyl halide. After heating to 80°C for 2–12h, the reaction
mixture was cooled and partitioned between water and
EtOAc. The organic layer was evaporated to give the
crude product, which was generally used without purifi-
cation. The O-protecting group was removed by standard
procedures.
12. McElhinney, R. S.; Donnelly, D. J.; McCormick, J. E.;
Kelly, J.; Watson, A. J.; Rafferty, J. A.; Elder, R. H.;
Middleton, M. R.; Willington, M. A.; McMurry, B. H.;
Margison, G. P. J. Med. Chem. 1998, 41, 5265–5271.
13. Middleton, M. R.; Kelly, J.; Thatcher, N.; Donnelly, D. J.;
McElhinney, R. S.; McMurry, B. H.; McCormick, J. E.;
Margison, G. P. Int. J. Cancer 2000, 85, 248–252.
14. Gary, R.; Ludwig, D. L.; Cornelius, H. L.; MacInnes,
M. A.; Park, M. S. J. Biol. Chem. 1997, 272, 24522–24529.
15. Berneburg, M.; Lehmann, A. R. Adv. Genetics 2001, 43,
71–102.
16. This molecule, referred to as BVT substrate, was prepared
by annealing oligonucleotide VT (50-VIC-ccctccgccg-
tcgcgttt-TAMRA; Applied BioSystems) with oligonucleo-
tide B (50-aaacgcgacggcggagggtcttgctcagtgtc gtctccgacact-
gagcaa-Black Hole Quencher; Integrated DNA
Technologies). High-throughput enzyme inhibition assays
were performed using 60lL reaction mixes (50mM Tris–
HCl; pH8.0, 10mM MgCl2, 0.5mM 2-ME, 6lg BSA,
2.5lg circular plasmid, 180 U FEN1 or 50ng XPG, 25lM
test compound, and 5pmol BVT substrate) contained in
black 96-well plates. Reactions were incubated at rt for
90min, stopped through the addition of 40lL stop buffer
(0.025% SDS, 12.5mM EDTA) and fluorescence was
measured using a Fluoroscan plate-reading fluorometer
fitted with 485nm excitation/538nm emission filters.
17. Tomassini, J. E.; Davies, M. E.; Hastings, J. C.; Lingham,
R.; Mojena, M.; Raghoobar, S. L.; Singh, S. B.; Tkacz, J.
S.; Goetz, M. A. Antimicrob. Agents Chemother. 1996, 40,
1189–1193.
23. Pirrung, M. C.; Tumey, L. N.; McClerren, A. L.; Raetz, C.
R. H. J. Am. Chem. Soc. 2003, 125, 1575–1586.
24. Liskowski, V.; Vu, D. N.; Feng, X.; Rault, S. Synthesis
2002, 6, 753–756.
25. General protocol for the synthesis of the furan amino
esters: The substituted acetonitrile was added to a solution
of ethyl formate (1.1equiv) and sodium methoxide
(1.1equiv) in MeOH (0.1M). After refluxing for 2h, the
resulting precipitate was filtered and washed with ether to
give the crude hydroxyacrylonitrile. Diethylchloromalo-
nate (1.0equiv) was added to a stirred solution of the
crude hydroxyacrylonitrile in DMF (0.1M). After stirring
5h at rt, the solvent was removed and the resulting oil was
dissolved in ethanol. DBN (1.0equiv) was added and the
reaction mixture was refluxed overnight. After concentra-
tion, the crude product was dissolved in EtOAc, washed
with water, and purified by silica gel chromoatography
(EtOAc/Hex) to afford the desired furan amino ester.