N-Hydroxyimides and hydroxypyrimidinones as inhibitors of the DNA repair complex ERCC1-XPF

Article abstract of DOI:10.1016/j.bmcl.2015.08.024

A high throughput screen allowed the identification of N-hydroxyimide inhibitors of ERCC1-XPF endonuclease activity with micromolar potency, but they showed undesirable selectivity profiles against FEN-1. A scaffold hop to a hydroxypyrimidinone template g

Full text of DOI:10.1016/j.bmcl.2015.08.024

Bioorganic & Medicinal Chemistry Letters 25 (2015) 4104–4108
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
N-Hydroxyimides and hydroxypyrimidinones as inhibitors
of the DNA repair complex ERCC1–XPF
a
a
a
a
Timothy M. Chapman a, , Claire Wallace , Kevin J. Gillen , Preeti Bakrania , Puneet Khurana ,
⇑
Peter J. Coombs a, Simon Fox a, Emilie A. Bureau a, Janet Brownlees a, David W. Melton b, Barbara Saxty a
a Centre for Therapeutics Discovery, MRC Technology, 1-3 Burtonhole Lane, Mill Hill, London NW7 1AD, UK
b MRC Institute of Genetics and Molecular Medicine, University of Edinburgh, MRC Human Genetics Unit, Western General Hospital, Crewe Road, Edinburgh EH4 2XU, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
A high throughput screen allowed the identification of N-hydroxyimide inhibitors of ERCC1–XPF endonu-
clease activity with micromolar potency, but they showed undesirable selectivity profiles against FEN-1.
A scaffold hop to a hydroxypyrimidinone template gave compounds with similar potency but allowed
selectivity to be switched in favour of ERCC1–XPF over FEN-1. Further exploration of the structure–activ-
ity relationships around this chemotype gave sub-micromolar inhibitors with >10-fold selectivity for
ERCC1–XPF over FEN-1.
Received 24 June 2015
Revised 6 August 2015
Accepted 7 August 2015
Available online 14 August 2015
Keywords:
DNA repair
Ó 2015 Elsevier Ltd. All rights reserved.
ERCC1–XPF
N-Hydroxyimide
Hydroxypyrimidinone
Platinum-based chemotherapeutics such as cisplatin result in
several forms of DNA damage, however their effectiveness can be
limited by efficient DNA repair processes. The ERCC1–XPF complex
is essential for one such process, nucleotide excision repair (NER),
and there is evidence that it is a good therapeutic target for poten-
tial intervention in a range of cancers,1,2 where inhibition of
ERCC1–XPF could enhance the effectiveness of chemotherapeutic
agents.
A high-throughput screen using a fluorescence-based in vitro
biochemical assay3 allowed the identification of hit compounds 1
and 2 as inhibitors of the endonuclease activity of ERCC1–XPF, with
micromolar potency and good ligand efficiency (Fig. 1). Despite
their flat structures and the N-hydroxyimide motif which was con-
sidered undesirable, they did not show DNA intercalation of our
assay substrate and we noted the similarity of the structure to
known flap endonuclease 1 (FEN-1) inhibitors4 and the natural
product flutimide,5,6 an inhibitor of influenza endonuclease. We
therefore sought to explore whether it was possible to gain addi-
tional potency and selectivity for ERCC1–XPF over other nucleases
including FEN-1 from this starting point. The testing of initial ana-
logues showed that the N-hydroxy group was essential for inhibi-
tory activity since replacement of the N-OH by either N-H or
N-methyl led to complete loss of activity. The mode of inhibition
of FEN-1 by N-hydroxyimides has been proposed to involve chela-
tion of metal ions at the endonuclease active site;4 the enzymatic
activity of ERCC1–XPF is also metal-dependent and the presence
of Mn2+ is required for substrate turnover in the biochemical assay,
so it was our hypothesis that the hits are also able to inhibit
ERCC1–XPF through binding to a metal ion at the endonuclease
active site.
Synthesis and testing of known FEN-1 inhibitors 3–64 (Table 1)
showed that they could also inhibit ERCC1–XPF at micromolar
levels, although due to their high potency against FEN-1 they were
highly selective for this target versus ERCC1–XPF. In order to mon-
itor the selectivity of inhibitors during the optimization process,
we employed counter screen assays against the non-structure
specific nuclease DNase I as well as FEN-1. The phenyl analogue
7 showed lower potency than the thiophene 5, although the addi-
tion of aniline (8) or pyrrolidine (9) substituents allowed some
potency to be regained. The acetamido- substituted positional vari-
ants (e.g., 10–12) all showed similar potency against ERCC1–XPF
but encouragingly suggested that it might be possible to tune the
selectivity of these inhibitors towards ERCC1–XPF inhibition and
away from FEN-1 (11, 12), although it still remained in favour of
FEN-1. The N-substituted analogue 13 was consistent with 6 in
suggesting that substitution at this position is not tolerated, and
14 showed good potency for its size although was still more potent
against FEN-1.
In order to advance the series with the aim of improving
potency and selectivity we investigated a scaffold hop to introduce
additional vectors for exploration and avoid the requirement for
the N–O bond, which was considered a liability from an ADMET
⇑
Corresponding author.
http://dx.doi.org/10.1016/j.bmcl.2015.08.024
0960-894X/Ó 2015 Elsevier Ltd. All rights reserved.

T. M. Chapman et al. / Bioorg. Med. Chem. Lett. 25 (2015) 4104–4108
4105
OH
N
OH
N
OH
N
O
OH
O
O
O
O
O
O
O
R2
O
NH
HN
N
R1
R
1
2
Cl
Figure 2. Scaffold hop to hydroxypyrimidinone motif.11
ERCC1-XPF IC50 = 30 µM
Ligand efficiency = 0.38
FEN-1 IC50 = 0.3 µM
ERCC1-XPF IC50 = 12 µM
Ligand efficiency = 0.39
FEN-1 IC50 = 0.7 µM
metal-binding motif, through either N-methylation (19) or O-
methylation (20), led to losses in potency. These initial examples
represented a promising start with respect to potency against
ERCC1–XPF and achieving selectivity against FEN-1 and DNase I.
The carboxylic acids and esters are highly polar and carry lim-
ited vectors for additional substitution for exploring potency and
selectivity gains, so amides instead of acids were investigated
(Table 3). It was found that a range of amides was well tolerated
and potency gains to the sub-micromolar level against ERCC1–
XPF were achieved, although the SAR appeared relatively flat in
this region. The primary carboxamide 21, methyl amide 22, small
lipophilic alkyl chains (23, 26, 27) and appended polar groups
(24, 28, 32) were all well tolerated, but the tertiary amide 25
was less potent. Aryl and heteroaryl groups also showed similar
potency (29, 31). They showed variable FEN-1 selectivity but in
the best case, with the primary carboxamide 21, greater than 20-
fold selectivity was observed.
Figure 1. Initial hit compounds.
perspective. The scaffold chosen was hydroxypyrimidinone (Fig. 2),
which is known to be able to bind metal ions (typically Mg2+
in a similar way to the N-hydroxyimide and is present in the
blockbuster HIV integrase drug raltegravir,7 inhibitors of HCV
polymerase8,9 and RNase H.10
Compounds containing this scaffold were synthesized and
tested, and gratifyingly initial examples showed comparable inhi-
bition of ERCC1–XPF to the N-hydroxyimides (Table 2). With either
thiophene or phenyl as the R1 substituent, the carboxylic acids 15
)
and 17 showed IC50 values around 3 lM and approximately 4-fold
selectivity for ERCC1–XPF over FEN-1. Their methyl ester ana-
logues 16 and 18 were found to have weaker potency than the
corresponding carboxylic acids and modifications around the
Table 1
Initial activity and selectivity data based around known N-hydroxyimide FEN-1 inhibitors and their analogues
OH
N
OH
N
OH
N
OH
N
OH
N
O
O
O
O
O
O
O
O
O
O
N
N
N
N
N
R1
R3
R3
R3
R4
O
S
R1
R3
S
A
B
C
D
R2
E
R1
R3
R2
R2
R1
R1
R2
R2
Compound
Core
A
R1
R2
R3
H
R4
—
ERCC1–XPF IC50
/
lM
FEN-1 IC50
0.057
/
lM
DNase I IC50
>100
/lM
3
H
1.5
4
5
B
C
H
H
H
H
—
—
1.8
3.9
0.003
0.009
>100
>100
H
H
6
7
8
C
H
H
H
—
H
H
80.4
18.1
3.9
0.68
0.30
0.23
>100
>100
30.2
D
D
H
H
H
HN
N
9
D
D
D
H
H
H
H
H
H
H
9.1
6.1
6.4
1.1
27.9
>100
NH
10
11
H
0.038
2.1
O
O
NH
H
>100
O
12
13
D
D
H
H
H
H
H
6.6
2.0
2.1
87
N
H
H
H
Me
96.2
>100
14
E
H
—
1.5
0.63
>100

4106
T. M. Chapman et al. / Bioorg. Med. Chem. Lett. 25 (2015) 4104–4108
Table 2
Table 3
Initial activity data from scaffold hop
Activity data from variations at amide position
O
O
O
O
OH
OH
R1
OH
R1
OH
O
O
O
R2
R2
R2
O
R1
O
O
O
N
HN
N
N
N
N
N
HN
S
N
R2
A
B
C
R1
Compound Core R1
R2
ERCC1–XPF
FEN-1
IC50
Ma
DNase I
IC50
Ma
IC50
/l
M
/
l
/l
Compound NR1R2
ERCC1–XPF
FEN-1
IC50
DNase I
IC50
Ma
IC50
/
lM
/
l
M
/l
15
16
A
A
H
2.9
10.4
10.1
>100
>100
NH2
S
S
21
1.0
0.6
23.3
12.0
>100
67.6
N
22
H
Me 27.3
N
23
2.7
0.7
4.2
4.0
>100
>100
H
OH
N
H
24
25
17
18
A
A
H
3.9
16.4
53.4
>100
>100
N
4.6
0.6
13.7
1.3
nt
Me 16.4
26
27
>100
N
H
19
20
B
C
H
H
48.8
88.9
nt
>100
nt
0.5
0.8
3.5
5.0
nt
S
S
N
H
H
N
>100
N
H
28
29
62.6
O
F
a
nt = not tested.
0.8
10.5
nt
N
H
Although the potency gains from introducing the amides were
N
H
modest relative to the carboxylic acids, a better balance in com-
pound properties and ADMET profiles was obtained (Table 4).
While the carboxylic acids typically showed logD values of less
than zero and were not permeable in the PAMPA assay, modifying
the amide substituent enabled tuning of the logD of the com-
pounds which led to corresponding improvements in the perme-
ability. They also displayed good stability in both mouse and
human liver microsomes.
30
31
4.2
0.7
2.1
8.6
>100
>100
F
N
N
N
H
O
32
1.0
4.1
58.9
N
H
With the isopentyl group in place as the amide substituent,
changes at the pyrimidinone 2-position were explored (Table 5).
In this region it was found that pendant groups could be added
to a phenyl ring without significant losses in potency, such as basic
groups in 33 and 34, but the introduction of saturated ring systems
led to losses in potency (e.g., 35–38).
In order to address the possibility that the compounds could be
inhibiting enzyme activity in a non-specific manner, such as
through general metal chelation, it was important to test the com-
pounds in biophysical assays to confirm binding to ERCC1–XPF.
Compound 22 was profiled by surface plasmon resonance (SPR),
a
nt = not tested.
(data not shown). In view of the cell activity shown by a compound
with similar biochemical potency,12 it is somewhat surprising that
22 showed no activity in these assays and there are a number of
reasons why this might be the case. Despite acceptable physical
and ADMET properties, it might be that insufficient compound
concentration at the target was achieved, or alternatively that inhi-
bition of ERCC1–XPF activity in these cells is not sufficient to stop
DNA repair. Unlike the catechol series,12 no cell toxicity was
observed with compound alone for the hydroxypyrimidinones,
which could suggest that the cellular effects seen with the catechol
series are due to inhibiting targets and pathways in addition to
ERCC1–XPF. Further studies to measure target engagement in the
cell and probe on- versus off-target effects for both the catechols
and hydroxypyrimidinones would be required to confirm this,
but it might be the case that a combination approach will be
required to achieve efficacy in a therapeutic.
In summary, we have shown that compounds with the N-hy-
droxyimide metal-binding chemotype are able to inhibit the activ-
ity of ERCC1–XPF, although they are also highly potent against
FEN-1. Promisingly, performing a scaffold hop to a hydroxypyrim-
idinone template allowed sub-micromolar inhibitors of ERCC1–
XPF to be obtained with >10-fold selectivity over the other nucle-
ases FEN-1 and DNase I; the inhibitors were also shown to bind to
where it displayed a KD of 4 lM with fast binding kinetics
(Fig. 3) and also in a thermal shift assay, where it showed a
DTm
of 4 °C.
Compound 22 was tested (up to 10 lM) in cell based assays
using A375 melanoma cells to assess effects on DNA repair. Three
assays were used to measure the following parameters: direct
nucleotide excision repair (NER) using a recombinant GFP reporter
assay,3 potentiation of cisplatin induced cell death and a high con-
tent imaging assay measuring levels of
cH2AX foci, a marker of
DNA repair.12 We did not observe any inhibition of repair of dam-
aged GFP DNA by compound 22 in the NER assay and in the cell
viability assay run over 5 days, the compound alone did not cause
toxicity but neither did it enhance the cell death seen with cis-
platin (dose response range of cisplatin of 0.1–3
we did not observe a delay in the repair of DNA damage caused
by cisplatin by addition of compound 22 in the H2AX foci assay
lM). Similarly,
c

T. M. Chapman et al. / Bioorg. Med. Chem. Lett. 25 (2015) 4104–4108
4107
Table 4
Measured in vitro ADMET profiles for selected compounds
Compound
LogD
Cytotox Hep-G2 v-50/
lM
MLM t1/2/min
HLM t1/2/min
PAMPA Papp/nm sÀ1
21
22
26
À0.4
0.5
2.1
>10
>10
>10
150
>400
>400
139
181
>400
7
31
97
Table 5
Activity data from variations at 2-position
O
OH
R1
O
N
H
HN
N
Compound
R1
ERCC1–XPF IC50
/
lM
FEN-1 IC50
38.1
/lM
O
33
0.8
N
34
35
36
1.7
6.3
8.4
71.3
>100
N
NH
N
>100
37
38
25.3
6.9
>100
>100
N
N
Figure 3. SPR sensorgram for compound 22.
ERCC1–XPF in SPR and thermal shift assays. Further work to under-
stand the binding mode of these compounds, for example, if a crys-
tal structure could be obtained, might reveal additional
opportunities to drive improvements in enzyme affinity and cell
activity and advance their utility as probes for ERCC1–XPF biology.
Acknowledgements
The authors thank David Tickle, Sadhia Khan, Nathalie Bouloc
and Zaynab Isseljee for in vitro ADMET and Martin Wear for pro-
duction of ERCC1–XPF protein.

4108
T. M. Chapman et al. / Bioorg. Med. Chem. Lett. 25 (2015) 4104–4108
5. Tomassini, J. E.; Davies, M. E.; Hastings, J. C.; Lingham, R.; Mojena, M.;
Supplementary data
Raghoobar, S. L.; Singh, S. B.; Tkacz, J. S.; Goetz, M. A. Antimicrob. Agents
Chemother. 1996, 40, 1189.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2015.08.
024.
6. Singh, S. B.; Tomassini, J. E. J. Org. Chem. 2001, 66, 5504.
7. Serrao, E.; Odde, S.; Ramkumar, K.; Neamati, N. Retrovirology 2009, 6, 25.
8. Stansfield, I.; Avolio, S.; Colarusso, S.; Gennari, N.; Narjes, F.; Pacini, B.; Ponzi, S.;
Harper, S. Bioorg. Med. Chem. Lett. 2004, 14, 5085.
9. Koch, U.; Attenni, B.; Malancona, S.; Colarusso, S.; Conte, I.; Di Filippo, M.;
Harper, S.; Pacini, B.; Giomini, C.; Thomas, S.; Incitti, I.; Tomei, L.; De Francesco,
R.; Altamura, S.; Matassa, V. G.; Narjes, F. J. Med. Chem. 2006, 49, 1693.
10. Lansdon, E. B.; Liu, Q.; Leavitt, S. A.; Balakrishnan, M.; Perry, J. K.; Lancaster-
Moyer, C.; Kutty, N.; Liu, X.; Squires, N. H.; Watkins, W. J.; Kirschberg, T. A.
Antimicrob. Agents Chemother. 2011, 55, 2905.
11. Note that we have depicted a hydroxypyrimidinone tautomeric form but it
may exist as alternative tautomeric forms depending on the environment.
12. Chapman, T. M.; Gillen, K. J.; Wallace, C.; Lee, M. T.; Bakrania, P.; Khurana, P.;
Coombs, P. J.; Stennett, L.; Fox, S.; Bureau, E. A.; Brownlees, J.; Melton, D. W.;
Saxty, B. Bioorg. Med. Chem. Lett. 2015, 25, 4097.
References and notes
1. McNeil, E. M.; Melton, D. W. Nucleic Acids Res. 2012, 40, 9990.
2. Song, L.; Ritchie, A. M.; McNeil, E. M.; Li, W.; Melton, D. W. Pigment Cell
Melanoma Res. 2011, 24, 966.
3. McNeil, E. M.; Ritchie, A. M.; Astell, K. R.; Shave, S.; Houston, D. R.; Bakrania, P.;
Jones, H. M.; Khurana, P.; Wallace, C.; Chapman, T.; Wear, M. A.; Walkinshaw,
M. D.; Saxty, B.; Melton, D. W. DNA Repair 2015, 31, 19.
4. Tumey, L. N.; Bom, D.; Huck, B.; Gleason, E.; Wang, J.; Silver, D.; Brunden, K.;
Boozer, S.; Rundlett, S.; Sherf, B.; Murphy, S.; Dent, T.; Leventhal, C.; Bailey, A.;
Harrington, J.; Bennani, Y. L. Bioorg. Med. Chem. Lett. 2005, 15, 277.