Novel indole inhibitors of IMPDH from fragments: Synthesis and initial structure-activity relationships

Article abstract of DOI:10.1016/j.bmcl.2006.01.090

The elaboration of previously reported indole fragments as inhibitors of inosine monophosphate dehydrogenase (IMPDH) is described. The synthesis, in vitro inhibitory values for IMPDH II, PBMC proliferation and physicochemical properties are discussed.

Full text of DOI:10.1016/j.bmcl.2006.01.090

Bioorganic & Medicinal Chemistry Letters 16 (2006) 2539–2542
Novel indole inhibitors of IMPDH from fragments:
Synthesis and initial structure–activity relationships
Rebekah E. Beevers, George M. Buckley, Natasha Davies, Joanne L. Fraser,
Francis C. Galvin, Duncan R. Hannah,* Alan F. Haughan, Kerry Jenkins,
Stephen R. Mack, William R. Pitt, Andrew J. Ratcliffe, Marianna D. Richard,
Verity Sabin, Andrew Sharpe and Sophie C. Williams
UCB Celltech, Granta Park, Great Abington, Cambridge, CB1 6GS, UK
Received 21 November 2005; revised 19 January 2006; accepted 19 January 2006
Available online 17 February 2006
Abstract—The elaboration of previously reported indole fragments as inhibitors of inosine monophosphate dehydrogenase
(IMPDH) is described. The synthesis, in vitro inhibitory values for IMPDH II, PBMC proliferation and physicochemical properties
are discussed.
Ó 2006 Elsevier Ltd. All rights reserved.
In the preceding article, we reported the results of our
investigations into the inhibition of inosine monophos-
phate dehydrogenase (IMPDH) with low molecular
weight indole fragments. As well as providing valuable
insights into the requirements for potency against
IMPDH II, two hit templates were highlighted for fur-
ther investigation, namely 3-cyanoindoles and 3-(4-
pyridyl)indoles.
sition.1,2 Based on binding mode hypotheses this was
deemed necessary to allow formation of key H-bonds
from the linker to Asp 274. From the observed SAR
of the indole fragments, coupled with docking studies
(see preceding paper), it was hoped that in the case of
1a and 2 the indole NH could perform this function,
thereby potentially allowing a greater range of substitu-
tion in the carbocyclic ring. Additionally, modelling
studies suggested the indole 6- and 7-positions to be
the optimum substitution points. Consequently, a range
of 6- and 7-substituted indoles were prepared to test our
hypotheses, and their syntheses are described in Schemes
1–5.
Herein, we report our efforts to further develop repre-
sentatives of these templates, namely the cyanoindoles
1a and 1b, and pyridylindole 2, which have IMPDH II
potencies (IC50) of 20.9, 7.66 and 1.15 lM, respectively.
N
NC
4
Indoles 3, 7 and 8 were prepared via direct cyanation of
their commercially available precursors with chlorosulf-
onylisocyanate (CSI) (Scheme 1). Catalytic hydrogena-
tion of 3 afforded aminoindole 4, which was further
transformed into the urea 5 and the benzoxazole 6. In
the synthesis of N-1 alkylated cyanoindoles, nitroindoles
(3) were alkylated and then reduced to versatile 6 or 7-
aminoindole intermediates I. Employing isocyanate
chemistry, indoles I were then converted to ureas (9,
10 and 16), an N-hydroxy urea (11) and an O-amino car-
bamate (12). Sulfonamides (13–15) and the benzoxazole
(17) were also prepared from aminoindoles I (Scheme 2).
On the pyridylindole template, a phenyl group was
introduced in the 6-position via Suzuki reaction upon
bromoindole II to give 18 (Scheme 3). The ester 19
3
5
6
2
1 N
R
7
N
H
1a R=H
1b R=Me
2
Although SAR around a 3-(oxazol-5-yl)indole and 3-
cyanoindole template has been previously reported, in
both cases compounds were restricted to incorporation
of a NHCONH or (2-oxazolyl)amino linker at the 6-po-
Keywords: IMPDH; Inosine monophosphate dehydrogenase; Frag-
ment elaboration.
*
Corresponding author. Fax: +44 1223 896400; e-mail: duncan.
hannah@ucb-group.com
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.01.090

2540
R. E. Beevers et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2539–2542
NC a, b
a
N
N
CO2H
CO2Et
3 R=6-NO2
7,8 R=OBn
H
N
N
PhO2S
H
R
H
R
c
b
N
I
NC
NC
c
d
O
N
N
Ph
CO2Et
NH2
N
N
N
N
H
CO2Et
H
PhO2S
H
H
4
5
19
d
Scheme 4. Reagents and conditions: (a) EtOH, H2SO4, reflux; (b)
PhSO2Cl, K2CO3, methylethyl-ketone, reflux (51%, 2 steps); (c) N-
iodosuccinimide, MeCN, rt (13%); (d) Pyridin-4-yl-boronic acid,
Pd(PPh3)4, K3PO4, DME, H2O, 80 °C (17%).
NC
O
N
N
N
H
H
6
N
N
Scheme 1. Reagents and conditions: (a) CSI, MeCN, rt; then DMF, rt
(22–100%); (b) (R = 6-NO2) H2, 10% Pd–C, MeOH, rt (100%); (c)
triphosgene, Et3N, DCM, 0 °C; then PhNHMe, rt (19%); (d) 2-
chlorobenzoxazole, DMF, microwave, 120 °C (7%).
a
R
R
N
N
H
NH2
H
N
H
SO2R'
III
b
20, 25, 29
c
NC
N
NC
N
O
N
N
N
H
NO2
N
R
3
H
O
N
R''
O
N
N
N
R'
H
N
H
H
a, b
H
R''
9-11, 16
c
N
H
OR
R'
21, 22, 26, 27
23, 24, 28
NC
NC
d, e
Scheme 5. Reagents and conditions: (a) PhSO2Cl, DCM, pyridine, rt
(8–49%); (b) triphosgene, THF, ꢀ78 to 0 °C; then R0R00NH, rt (7–
34%); (c) RONH2.HCl, triphosgene, Et3N, DCM, ꢀ78–0 °C; then III,
rt (4–33%).
O
H
N
N
R
NH2
N
I
N
O
H
f
12
g
NC
NC
was prepared from indole-6-carboxylic acid (Scheme 4).
The aminoindole precursor III, described in the preced-
ing paper, was elaborated to sulfonamides (20, 25 and
29), ureas (21, 22, 26 and 27) and N-alkoxyureas (23,
24 and 28) (Scheme 5).
SO2R'
N
R
N
N
H
HN
O
13-15
N
17
The in vitro enzyme potencies and activity in a periph-
eral blood mononuclear cell (PBMC) proliferation as-
say are shown in Table 1. The first observation is
that compared to the pyridylindoles, only moderate
improvements in potency were achieved with N-meth-
ylcyanoindoles, and with the NH cyanoindoles no real
improvements were found with the small set prepared.
This may reflect a sub-optimal binding of the cyanoin-
dole cores, as well as highlighting the sensitivity of the
enzyme to the precise orientation of the substituents at-
tached. Though simple 6-amino substitution on the NH
cyanoindole in 4 was not tolerated, the urea 5 retained
the potency of the fragment core. The benzoxazole 6
appeared to lose activity, failing to give an IC50, and
the 6-benzyloxyindole 7 was only weakly active. The
7-benzyloxyindole 8 was not tolerated, modelling
predicting this position to be sterically restricted in
the active site. In the case of N-methylcyanoindoles,
Scheme 2. Reagents and conditions: (a) MeI or EtI, KOH, acetone, rt
(87%); (b) H2, 10% Pd–C, MeOH, rt (100%); (c) PhNCO, DCM, rt
(50%); or, triphosgene, Et3N, DCM, 0 °C; then PhNHMe, rt (10–53%);
or, PhCH2ONH2ÆHCl, triphosgene, Et3N, DCM, ꢀ78 to 0 °C; then I,
rt (25%); (d) triphosgene, Et3N, DCM, 0 °C; then HONMeBoc (65%);
(e) TFA, DCM, rt (89%); (f) R0SO2Cl, DCM, pyridine, rt (9–35%); (g)
2-chlorobenzoxazole, DMF, microwave, 120 °C (33%).
N
N
a, b
N
N
Br
Ph
H
PhO2S
II
18
Scheme 3. Reagents and conditions: (a) PhB(OH)2, Pd(PPh3)4,
Na2CO3, DME, H2O, microwave, 150 °C; (b) KOH, MeOH, reflux
(62%, 2 steps).

R. E. Beevers et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2539–2542
2541
Table 1. SAR of indoles
R3
R2
R6
N
R1
R7
Compound
R1
R2
R3
R6
R7
IMPDH II3 IC50 (lM)
PBMC3 IC50 (lM)
1a
1b
2
—
—
—
H
—
—
—
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Me
Me
H
H
H
—
—
—
—
—
—
—
H
H
H
H
20.900
7.660
1.150
IAa
—
—
—
CN
—
NH2
28.0
—
4
5
H
CN
CN
NHCON(Me)Ph
Benzoxazol-2-ylamino
OBn
15.400
48% at 30 lM
71.800
IAa
—
—
6
H
7
H
CN
CN
—
—
8
H
H
OBn
H
9
Me
Me
Me
Me
Me
Me
Et
Me
Me
H
CN
CN
NHCONHPh
NHCON(Me)Ph
NHCONHOBn
NHC(O)ONHMe
NHSO2Ph
NHSO2Bn
NHSO2Bn
H
0.722
17.000
1.170
4.040
7.680
21.000
IAa
14.0
—
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
H
CN
CN
H
17.0
—
H
CN
CN
H
—
—
H
CN
CN
H
NHCON(Me)Ph
—
—
7.850
58.400
0.456
0.586
0.364
IAa
CN
H
Ph
Benzoxazol-2-ylamino
H
—
4-Pyridyl
4-Pyridyl
4-Pyridyl
4-Pyridyl
4-Pyridyl
4-Pyridyl
4-Pyridyl
4-Pyridyl
4-Pyridyl
4-Pyridyl
4-Pyridyl
4-Pyridyl
10.0
—
H
CO2Et
H
H
NHSO2Ph
NHCONHPh
NHCON(Me)Ph
NHCONHOMe
NHCONHOBn
NHSO2Ph
NHCON(Me)Ph
H
H
14.0
—
H
H
H
H
0.084
1.110
0.076
0.222
0.128
19.500
1.360
6.210
12.0
—
H
H
H
H
4.6
—
H
H
H
H
—
—
H
NHCON(Me)Ph
NHCONHOBn
NHSO2Ph
H
H
—
—
H
H
a Inactive at 30 lM.
substitution at the 6-position with the monosubstituted
urea 9 gave IMPDH potency in line with that previous-
ly reported,2 but addition of a methyl in 10 resulted in a
23-fold reduction in activity. The hydroxyurea 11 was
comparable to the urea 9, however, attempts to signif-
icantly improve the potency through variation of the
benzyl group were unsuccessful (data not shown). The
O-aminocarbamate 12 and sulfonamides 13 and 14
gave low micrololar potency. Variation of the phenyl
group of 13 for substituted phenyl/heteroaromatics also
failed to improve potency (data not shown). Replace-
ment of the N-methyl with an N-ethyl in 15 abolished
activity, in agreement with expected steric restrictions
at this position in our binding hypothesis. Only a small
number of 7-substituted-N-methyl indoles were pre-
pared. The urea 16 retained activity at a level similar
to that of the fragment 1b, whereas the benzoxazolyla-
mino group in 17 lost activity.
the cyanoindoles, in that the simple urea 21 was inactive
but the methyl-substituted analogue 22 showed potent
inhibition (IC50 = 84 nM). In a further twist, although
the simple methoxy urea 23 showed micromolar activity,
the benzyloxy 24 again gave potent inhibition. In the
case of the pyridylindole fragment 2 of the preceding
article, addition of a 2-methyl on the indole increased
affinity roughly threefold, however, as can be seen with
25 and 26 this effect was variable (compare to 20 and 22,
respectively). As seen with cyanoindoles, substitution at
the 7-position of pyridylindoles in ureas 27 and 28, and
sulfonamide 29 led to a reduction in potency.
The activities in a PBMC proliferation assay showed a
consistently high drop-off from the enzyme potency.
The core fragments 1b and 2 show good transport in a
Caco-2 membrane permeability assay (Table 2), and
the reasons for this observed drop-off remain unclear.
The sub-micromolar IMPDH potency of the 6-phenyl
and 6-ethoxycarbonyl pyridylindoles 18 and 19, respec-
tively, lends support to the hypothesis of the indole NH
forming the key interaction with Asp 274. In this case, a
functionalised 6-amino group leads to templates with
consistantly better affinity. For example, sulfonamide
20 showed promising enzyme potency. In the case of
ureas, the SAR appeared to completely contrast with
A number of physicochemical properties and Caco-2
transport were determined for cyanoindole 11 and
pyridylindoles 20, 22 and 24 (Table 2). All four com-
pounds exhibited acceptable log Ds and good Caco-2
transport. As expected, the solubility of the pyridyl
derivatives was greatest at lower pH. The low molecular
weights and PSAs allow significant scope for further
optimisation of each template.

2542
R. E. Beevers et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2539–2542
Table 2. Drug-like properties
e
PSA (A )
2
Log D7.4a
Solb (lg/ml)
Caco-2c (10ꢀ7 cm/s)
A-B B-A
MWd
˚
Compound
1b
2
2.30
2.81
2.74
2.01
2.93
3.54
—
—
367
262
193
65
374
239
191
93
156
194
320
349
342
358
30
32
88
79
57
87
11
20
22
24
4.0 (4.5)
ND (8.0)
ND (42)
3.2 (17.5)
143
65
67
73
a Log D, measured at pH 7.4—Ref. 4.
b Aqueous solubility at pH 7.4, solubility at pH 5.0 in parentheses—Ref. 5. ND, not detected.
c Caco-2 Papp (apparent permeability), A-B (apical to basolateral) and B-A (basolateral to apical)—Ref. 6.
d Molecular weight.
e Polar surface area.
Whitworth, J. M.; Williams, S. C. Bioorg. Med. Chem. Lett.
2005, 15, 751.
In summary, starting from a focused fragment based ap-
proach around an indole core we have gained valuable
insights into the requirements for activity against
IMPDH II. From this position, 3-cyanoindole and 3-
(4-pyridyl)indole have been further elaborated, with
the pyridylindole template providing sub-micromolar
inhibitors of this enzyme with good drug-like properties.
This represents a further departure from the 3-methoxy-
4-(5-oxazolyl)aniline moiety of known IMPDH inhibi-
tors, and efforts to optimise this interesting class will
be reported elsewhere.
4. Log DpH7.4 was measured as follows. Compound (1–2 mg)
was shaken with 1 ml of pH 7.4 buffer (0.1 M KH2PO4
adjusted to pH7.4 with 0.1 M NaOH, presaturated with
octanol over 18 h) and 1 ml of octanol (presaturated with
pH 7.4 buffer over 18 h) for 1 hour at 750 osc/min. The
separate layers were filtered through 0.2 lm PTFE and
analysed by HPLC. The log D was determined according
to: log DpH 7.4 = log (normalised peak area in octanol/
normalised peak area in bufferpH 7.4).
5. The aqueous solubility was measured as follows. Com-
pound calibration standards were made up from DMSO
stock solutions to produce a three-point calibration curve.
Buffer (0.5 ml, 150 mM phosphate buffer) was added to
0.5 mg of compound sample and left to shake at room
temperature for 90 mins. After this, the samples were
filtered and the concentration of sample in the filtrate was
measured using the calibration graph. Analysis of both
calibration standards and filtrate samples was carried out
by HPLC (UV detection at 254 nm). The solubility was
measured in duplicate at two pHs.
6. Caco-2 transepithelial transport studies were performed
using 24-well Transwell (Beckman Dickinson) format (10–
21 day culture/passage 30–40). Once cells had grown to a
confluent monolayer, confirmed by TEER measurement
readings, the growth media (DMEM) were removed and
replaced with Hanks’ Buffer Salt Solutions (HBSS)
pH = 7.4 (0.2 ml and 0.7 ml in apical and basolateral wells,
respectively). Test compound (1 lM) cassetted with 2
reference compounds (10 lM) was spiked into the apical
or basolateral wells (n = 4), and plates were incubated for
2 h at 37 °C, 5% CO2 and 95% humidity. The basolateral or
apical media were removed (if apically or basolaterally
spiked, respectively) and transferred to a 96-well plate,
containing calibration standards, and samples and stan-
dards were analysed by LC–MSMS. The measured con-
centrations of compound and reference compounds in each
well were used to give a normalised compound Papp, based
on the known Papps of the reference compounds.
Acknowledgments
The authors thank the Molecular and Cellular Systems
department at Granta Park for in vitro activity measure-
ments, the Pharmaceutical Sciences department at
Granta Park for the log D measurements and colleagues
at UCB Celltech in Slough for solubility and Caco-2
measurements.
References and notes
1. Dhar, T. G. M.; Shen, Z.; Fleener, C. A.; Rouleau, K. A.;
Barrish, J. C.; Hollenbaugh, D. L.; Iwanowicz, E. J. Bioorg.
Med. Chem. Lett. 2002, 12, 3305.
2. Dhar, T. G. M.; Shen, Z.; Gu, H. H.; Chen, P.; Norris, D.;
Watterson, S. H.; Ballentine, S. K.; Fleener, C. A.;
Rouleau, K. A.; Barrish, J. C.; Townsend, R.; Hollenb-
augh, D. L.; Iwanowicz, E. J. Bioorg. Med. Chem. Lett.
2003, 13, 3557.
3. The IMPDH II enzymatic inhibition and PBMC prolifer-
ation assay protocols are described in: Buckley, G. M.;
Davies, N.; Dyke, H. J.; Gilbert, P. J.; Hannah, D. R.;
Haughan, A. F.; Hunt, C. A.; Pitt, W. R.; Profit, R. H.;
Ray, N. C.; Richard, M. D.; Sharpe, A.; Taylor, A. J.;