Solving time-dependent CYP3A4 inhibition for a series of indole-phenylacetic acid dual antagonists of the PGD2 receptors CRTH2 and DP

Article abstract of DOI:10.1016/j.bmcl.2014.04.092

Based on their structural similarity to previously described compound AMG 009, indole-phenyl acetic acids were proposed to be potent dual inhibitors of chemoattractant receptor-homologous molecule expressed on Th2 cells (CRTH2 or DP₂) and prostanoid D receptor (DP or DP₁). This series was equipotent to AMG 009 in binding assays against both receptors but exhibited decreased serum shift. We discovered early in the optimization of these indole-phenylacetic acid compounds that they demonstrated CYP3A4 time-dependent inhibition (TDI). Hypothesizing that the source of TDI was the indole core we modified the 1,2,3-substitution to eventually afford a highly potent modulator of CRTH2 and DP which did not exhibit TDI.

Full text of DOI:10.1016/j.bmcl.2014.04.092

Bioorganic & Medicinal Chemistry Letters xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Solving time-dependent CYP3A4 inhibition for a series of
indole-phenylacetic acid dual antagonists of the PGD2 receptors
CRTH2 and DP
a
a
a
a
Michael G. Johnson a, , Jim (Jiwen) Liu , An-Rong Li , Bettina van Lengerich , Sophie Wang ,
Julio C. Medina a, Tassie L. Collins b, Jay Danao b, Lisa Seitz b, Angela Willee b, Warren D’Souza b,
Alison L. Budelsky b, Peter W. Fan c, Simon G. W. Wong c
⇑
a Department of Therapeutic Discovery, Amgen, Inc., 1120 Veterans Blvd, South San Francisco, CA 94080, USA
b Department of Inflammation Research, Amgen, Inc., 1201 Amgen Court West, Seattle, WA 98119, USA
c Department of Pharmacokinetics, Metabolism and Distribution, Amgen, Inc., 1120 Veterans Blvd, South San Francisco, CA 94080, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Based on their structural similarity to previously described compound AMG 009, indole-phenyl acetic
acids were proposed to be potent dual inhibitors of chemoattractant receptor-homologous molecule
expressed on Th2 cells (CRTH2 or DP2) and prostanoid D receptor (DP or DP1). This series was equipotent
to AMG 009 in binding assays against both receptors but exhibited decreased serum shift. We discovered
early in the optimization of these indole-phenylacetic acid compounds that they demonstrated CYP3A4
time-dependent inhibition (TDI). Hypothesizing that the source of TDI was the indole core we modified
the 1,2,3-substitution to eventually afford a highly potent modulator of CRTH2 and DP which did not
exhibit TDI.
Received 20 March 2014
Revised 21 April 2014
Accepted 23 April 2014
Available online xxxx
Keywords:
CRTH2
DP
CYP3A4
Indole
Ó 2014 Elsevier Ltd. All rights reserved.
Phenylacetic acid
Cl
Cl
Activation of the G protein-coupled receptors prostanoid D
receptor (DP or DP1) and chemoattractant receptor-homologous
molecule expressed on Th2 cells (CRTH2 or DP2) by their endoge-
nous ligand prostaglandin D2 (PGD2) leads to mast cell mediated
allergic responses observed in allergic rhinitis, allergic conjuctivitis
asthma and atopic dermatitis.1 As such, these receptors have
emerged as targets for small-molecule antagonists that may mod-
ulate allergic responses. Our group as well as others have reported
the discovery of CRTH2 antagonists, DP antagonists and CRTH2/DP
dual antagonists.2 Over the past several years clinical trials with
CRTH2 antagonists have afforded mixed results with multiple pro-
grams discontinued due to lack of efficacy. Few trials are underway
with DP antagonists, and the dual CRTH2/DP antagonists AMG 009
and AMG 853 have been discontinued.
O O
S
O O
S
Cl
Cl
NH
OMe
NH
OMe
O
O
O
O
nBuHN
Me
OH
N
OH
H
O
AMG 009
1
Figure 1. CRTH2/DP dual antagonists AMG 009 (clinical candidate) and indole 1.
afford compounds with distinct potency and PKDM profiles
(Fig. 1). Indeed, the series exemplified by 1 was similar in binding
potency to AMG 009 against both CRTH2 and DP. Compound 1,
however, demonstrated time-dependent inhibition (TDI) of the
cytochrome P450 isoform 3A4. Through modification of the 1-, 2-
and 3-positions of the indole nucleus, we determined which sub-
stitutions mitigated TDI and discovered indole-phenylacetic acids
In this Letter we report on the discovery of a series of CRTH2/DP
dual antagonists that was inspired by the clinical compound AMG
009.3 We proposed that the replacement of the phenylacetamido
core of AMG 009 with an indole core structure found in 1 would
that maintained the binding potency of
1 but lacked TDI.
Subsequent to our resolution of the TDI problem, 3A4 metabolite
identification both in vitro and in vivo confirmed our hypothesis
that the indole nucleus was the liability for TDI.4
⇑
Corresponding author. Tel.: +1 650 244 2408.
E-mail address: mgjohnso@amgen.com (M.G. Johnson).
http://dx.doi.org/10.1016/j.bmcl.2014.04.092
0960-894X/Ó 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Johnson, M. G.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.04.092

2
M. G. Johnson et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
The first compound we made in the indole-phenylacetic acid
NO2
OMe
NO2
OMe
O
a,b
series was 1. We were gratified to see comparable potency to
AMG 009 in binding for both CRTH2 (1 IC50 = 18 nM and AMG
009 IC50 = 3 nM) and DP (1 IC50 = 9 nM and AMG 009 IC50 = 12 nM)
when measured in buffer. Moreover, functional inhibitory activity
of 1 against both receptors was confirmed in both the PGD2-med-
iated CRTH2 receptor internalization assay (Kb = 55 nM) and the DP
PGD2-stimulated cAMP production assay (Ki = 2.6 nM).5 Compound
1 CRTH2 binding potency decreases two to three-fold in the pres-
ence of 50% plasma compared with a nine-fold potency decrease
for AMG 009.3 The in vitro potency of 1 against CRTH2 and DP com-
pared favorably also with the later clinical candidate vidupiprant
(AMG 853) but the serum shift was less beneficial.5 While we
expected that the decreased plasma shift for 1 versus AMG 009
should translate to increased target coverage in vivo at equivalent
doses we anticipated that the inhibitory activity for the indole ser-
ies needed to be improved and began a campaign to interrogate
binding potency.
O
O
O
R2
N
OEt
H2N
OH
R1
2
3a (R1=H; R2=Me)
3b (R1=H; R2=cPr)
3c (R1=H; R2=iBu)
3d (R1=H; R2=tBu)
3e (R1=H; R2=nPr)
3f (R1=Me; R2=Me)
c
Cl
Phx=
Cl
d-f
1 (R1=H; R2=Me; R3=H)
4 (R1=H; R2=cPr; R3=H)
5 (R1=H; R2=iBu; R3=H)
Phx
O2S
NH
OMe
R3
6 (R1=H; R2=tBu; R3=H)
7 (R1=H; R2=nPr; R3=H)
8 (R1=Me; R2=Me; R3=H)
9 (R1=H; R2=Me; R3=Cl)
O
g
O
Very early in thir development it became clear that 1 and its
analogs were potent time-dependent in vitro inhibitors of CYP3A4.
When 1 was incubated at 15 lM with rat liver microsomes in the
R2
N
R1
OH
presence of NADPH, 34% of the 3A4 activity to metabolize testos-
terone was observed after 15 min. No TDI was observed with
AMG 009. Because the only structural difference between AMG
009 and 1 is the replacement of the phenyl acetamide unit with
the indole core it was apparent that the indole was responsible
for the observed TDI. We planned to mitigate TDI by focusing on
substitutions around the indole ring system with the goal of main-
taining or improving potency against CRTH2 and DP while keeping
the serum shift low.
The known potential for oxidation of the indole across the 2–3
double bond guided us toward modifications at the 1-, 2- and 3-
positions on the indole which might mitigate oxidation. We pro-
posed to reduce oxidation both by increasing steric demand close
to the 2–3 bond and by decreasing the nucleophilicity of the bond
inductively by 2- and 3-substitution with electronegative func-
tional groups. As such, we synthesized a set of indole compounds
and discovered that modifying the 1-, 2- and 3-positions did in fact
have a significant impact on the in vitro TDI (Table 1).
Scheme 1. Reagents and conditions: (a) R2COMe, KOtBu, DMSO, rt, 10–30% yield;
(b) EtOH, cat. H2SO4, >90% yield; (c) Cs2CO3, MeI, DMF, rt, 50% yield; (d) SnCl2,
EtOAc, reflux, 40–70% yield; (e) 2,4-dichlorobenzene-sulfonyl chloride, pyridine, rt,
70–95% yield; (f) LiOH, MeOH–H2O, rt, >90% yield; (g) NCS, DMF, rt, 45% yield.
and 4-fluoro-3-nitroaniline. From 2 the 2-alkyl substituted indoles
were synthesized according to the procedure of Makosza then
esterified to afford indole-ethyl phenylacetates 3a–e.6 The 1-N-
methyl analog 3f was synthesized by methylation of the sodium
salt of the indole, generated by treatment of the indole 3a with
cesium carbonate and iodomethane. The 4-nitroindole intermedi-
ates 3a–f were then reduced to the corresponding 4-aminoindoles
using tin(II) chloride, sulfonylated with 2,4-dichlorobenzenesul-
fonylchloride in pyridine and the ester hydrolyzed to the final
product acids 1 and 4–8. The 3-chloro indole compound 9 was syn-
thesized by chlorination of 1 with N-chlorosuccinimide.
Synthesis of the 2-cyano indole compound 12 followed the
sequence outlined in Scheme 2. The starting ethyl phenylacetate
10 was prepared by nucleophilic aromatic substitution of
commercially available 1-fluoro-4-iodo-2-nitrobenzene by ethyl
The indole compounds 1–9 were synthesized as shown in
Scheme 1. The starting material 2 was synthesized conveniently
by SNAr reaction between 4-hydroxy-3-methoxyphenylacetic acid
Table 1
CRTH2 and DP in vitro binding potencies and CYP3A4 in vitro time-dependent inhibition of indole-phenylacetic acid compounds
Phx
SO2
HN
OMe
R3
O
Cl
Cl
O
R2
Phx =
N
R1
OH
a
a
No.
R1
R2
R3
CRTH2 buffer
IC50
(l
M) plasma
DP buffer
IC50
(
lM) plasma
CYP3A4 TDIb
1
4
5
6
7
8
9
12
14
H
H
H
H
H
Me
H
H
Me
cPr
iBu
tBu
nPr
Me
Me
CN
H
H
H
H
H
H
Cl
H
0.018
0.008
0.032
0.032
0.057
0.022
0.22
0.056
0.025
0.048
ND
0.11
0.056
ND
0.009
0.074
0.062
8.4
0.045
0.027
1.1
0.040
0.35
0.11
ND
0.13
0.076
ND
34
41
38
81
79
28
87
73
103
0.027
0.003
0.13
0.031
0.082
0.017
0.13
1.3
Oxindole
All values are an average of at least two experiments and standard deviations less than 30%; ND = not determined; TDI = time-dependent inhibition.
a
Binding displacement of 3H-PGD2 from CRTH2 or DP expressed on HEK 293 cells in buffer solution with 0.5% bovine serum albumin or in 50% human plasma (see Ref. 3).
b
% remaining 3A4 activity after 15 min incubation with 15 l
M compound concentration.8
Please cite this article in press as: Johnson, M. G.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.04.092

M. G. Johnson et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
NO2 OMe
3
NO2
OMe
Phx
HN
HN
GS
O
O
a-e
O
O
O
O
3A4
+GSH
O
O
1
Me
Me
H2N
N
H
OMe
I
OEt
NADPH
N
N
H
11
H
10
1-GSH
O2
S
proposed 2,3-epoxy
reactive intermediate
NH
OMe
f-j
O
O
Figure 2. Proposed 2,3 oxidation of indole and following conjugation with
glutathione.
NC
N
H
OH
12
Scheme 2. Reagents and conditions: (a) NH2NHBoc, CuI, DMF, 90 °C, 45% yield; (b)
ethyl 2-oxopropanoate, Eaton’s reagent, DCM, 45 °C, 33% yield; (c) LiOH, THF–H2O,
rt, 100% yield; (d) MeOH, TMSCl, rt, 100%; (e) NH4OH, EDC, NMM, rt, 80% yield; (f)
H2, Pd/C, EtOAc, 100% yield; (g) 2,4-dichlorobenzene sulfonyl chloride, pyridine,
DCM, RT, 65% yield; (h) LiOH, THF–H2O, rt, 100% yield; (i) POCl3, DCM, 60 °C, 65%
yield; (j) LiOH, THF–H2O, rt, 100% yield.
2-alkyl substitution whereas binding to DP can be dramatically
influenced by these replacements. The trend appears to be that lar-
ger 2-alkyl groups afford less potent DP binders, especially in the
presence of plasma. In light of the divergence between TDI and DP
binding potency, we concluded that substitution at the 2-position
of the indole was not advantageous.
Our next approach to mitigating oxidation at the 2–3 bond of
the indole was to introduce substituents that reduce oxidation
potential of the bond through inductive effects, for example with
2-cyano compound 12 and 3-chloro compound 9. Indeed, both
12 and 9 lacked TDI presumably as a consequence of the decreased
electron density of the indole 2–3 double bond. Whereas these
substitutions seemed to confirm our hypothesis regarding the
implication of the indole 2–3 bond in TDI, we were disappointed
to see that 3-chlorination decreased binding potency versus both
receptors and 2-cyanation negatively impacted binding potency
versus DP (Table 1).
The most convincing evidence that the indole was responsible
for TDI was substitution of the ring with a structurally related
oxindole 14 which completely lacked TDI. With respect to receptor
binding, compound 14 retained potency against CRTH2 relative to
1, though with an increased plasma shift and loss of potency
against DP in plasma.
NO2
OMe
NO2
OMe
MeS
a,b
O
O
O
O
O
N
OMe
H2N
OH
H
13
2
O2
S
c,d
NH
OMe
Phx
O
O
O
N
H
OH
14
Scheme 3. Reagents and conditions: (a) TMSCHN2, MeOH, toluene, rt, 95% yield; (b)
tBuClO, ethyl methylthioacetate, DCM, À65 °C, 45% yield; (c) Raney nickel, EtOH–
H2O, rt, 75% yield; (d) (i) 2,4-dichlorobenzenesulfonyl chloride, pyridine, DCM, rt,
(ii) LiOH, MeOH–THF–H2O, rt, 85% yield (two steps).
Based upon the effects of indole 1-,2- and 3-substitution on 3A4
activity, we were convinced that this group was the culprit in time-
dependent inhibition. Subsequent to our SAR efforts around this
issue, our PKDM group established that oxidation of the 2–3 bond
in 1 appeared to be the cause of TDI.4 Incubation of 1 in the pres-
ence of CYP 3A4 from microsomes followed by addition of glutathi-
one afforded the glutathione adduct 1-GSH (Fig. 2). It is proposed
that the glutathione adduct arises from 2,3 epoxidation by 3A4.
In conclusion, we identified a series of indole-phenylacetic acid
antagonists of CRTH2 and DP with binding potencies similar to
those of the clinical compound AMG 009. Early in the optimization,
we discovered that these compounds were time-dependent inhib-
itors of CYP 3A4. We proposed that oxidation of the indole was
responsible for the TDI and showed that altering the 2- and 3-sub-
stitution of the indole could eliminate TDI and maintain binding
potency versus both CRTH2 and DP. Our hypothesis that the indole
was the cause of TDI was further strengthened by identification of
glutathione adducts in in vitro CYP3A4 metabolite trapping exper-
iments. The oxindole compound 14 has no time-dependent
CYP3A4 inhibition and is potent against both CRTH2 and DP,
although a significant plasma shift for DP prevents advancement
of this series into in vivo studies.
2-(4-hydroxy-3-methoxyphenyl)acetate in 79% yield. The aryl
iodide 10 was then transformed in five steps to the 2-carboxamidino
indole 11 via a Fischer indole synthesis. In five steps the amide was
dehydrated to the nitrile, the nitro reduced and sulfonylated and the
ester hydrolyzed to ultimately afford 2-cyano indole 12.
Synthesis of the oxindole compound 14 followed the synthetic
route portrayed in Scheme 3. The 3-methylthiooxindole 13 was
formed in two steps from 2 via Gassman’s oxindole synthesis.7
Then 13 was transformed into oxindole product 14 in three steps,
reducing the sulfurAcarbon bond and the aryl nitro with Raney
nickel followed by sulfonylation of the aniline and hydrolysis of
the ester.
Simple methylation of the 1-N-position in 2 increased TDI rela-
tive to 1 (Table 1). We hypothesized that the alkylation of the
indole nitrogen increases electron density across the 2–3 double
bond, thus increasing susceptibility to oxidation at this position.
The increased TDI of 1-N-methylation in combination with the
neutral impact on CRTH2 and DP potency steered us away from
further substitution at the indole nitrogen.
Substitution of the 2-methyl group in 1 with alkyl groups of dif-
ferent sizes had a neutral or positive influence on TDI. The 2-cyclo-
propyl and 2-isobutyl replacements in 4 and 5, respectively had no
effect on TDI. The 2-n-propyl and 2-tert-butyl replacements had pro-
nounced positive effects on TDI as observed with compounds 7 and
6. Neither compound inhibits CYP3A4 in a time-dependent manner.
While there is not a clear correlation between size of the 2-alkyl
group and TDI in this series, the largest alkyl group that was exam-
ined, 2-tert-butyl compound 5, is the least inhibitory with respect to
TDI. Focusing on CRTH2 and DP binding potency it is evident that
binding to CRTH2 in buffer and plasma is essentially unaffected by
References and notes
1. (a) Schuligoi, R.; Sturm, E.; Lusching, P.; Konya, V.; Philipose, S.; Sedej, M.;
Waldhoer, M.; Peskar, B. A.; Heinemann, A. Pharmacology 2010, 85, 372; (b)
Pettipher, R. Br. J. Pharmacol. 2008, 153, s191; (c) Pettipher, R.; Hansel, T. T.;
Armer, R. Nat. Rev. Drug Disc. 2007, 6, 313; (d) Mitsumori, S. Curr. Pharm. Des.
2004, 10, 3533; (e) Matsuoka, T.; Hirata, M.; Tanaka, H.; Takahashi, Y.; Murata,
T.; Kabashima, K.; Sugimoto, Y.; Kobayashi, T.; Ushikubi, F.; Aze, Y.; Eguchi, N.;
Urade, Y.; Yoshida, N.; Kimura, K.; Mizoguchi, A.; Honda, Y.; Nagai, H.;
Narumiya, S. Science 2013, 2000, 287.
Please cite this article in press as: Johnson, M. G.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.04.092

4
M. G. Johnson et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
2. (a) Norman, P. Expert Opin. Invest. Drugs 2014, 23, 55; (b) Lamers, C.; Flesch, D.;
Schubert-Zsilavecz, M.; Merk, D. Expert Opin. Ther. Patents 2013, 23, 47; (c)
Pettipher, R.; Whittaker, M. J. Med. Chem. 2012, 55, 2915; (d) Burgess, L. E. Annu.
Rep. Med. Chem. 2011, 46, 119; (e) Norman, P. Expert Opin. Invest. Drugs 2010, 19,
947; (f) Pettipher, R.; Hansel, T. T. Drug News Perspect. 2008, 21, 317.
3. Liu, J.; Fu, Z.; Wang, Y.; Schmitt, M.; Huang, A.; Marshall, D.; Tonn, G.; Seitz, L.;
Sullivan, T.; Tang, H. L.; Collins, T.; Medina, J. Bioorg. Med. Chem. Lett. 2009, 19,
6419.
7. Gassman, P. G.; Van Bergen, T. J. J. Am. Chem. Soc. 1974, 96, 5508.
8. To evaluate the potential for time-dependent CYP inactivation, incubations
containing test compound (15 lM), pooled human liver microsomes (1.0 mg/
mL) in potassium phosphate buffer (100 mM, pH 7.4) were initiated with
NADPH (1 mM); final concentration. After 15 min of incubation, aliquots of this
primary incubation mixture were transferred to a secondary incubation (1:10
dilution) containing
a saturating concentration (200 lM) of the CYP3A4
substrate testosterone, NADPH (1 mM) and K2PO4 buffer (0.1 M, pH 7.4) and
incubated for an additional 20 min. Formation of 6-b-OH-testosterone under
these conditions was confirmed to be linear with protein concentration and
incubation time (data not shown). The secondary reaction was terminated by
addition of one volume of acetonitrile containing deuterated 6-b-OH-
testosterone as the internal standard. The samples were centrifuged and
CYP3A activity was determined by measuring 6-b-OH-testosterone formation
by an LC/MS/MS method.
4. Wong, S. G.; Fan, P. W.; Subramanian, R.; Tonn, G. R.; Henne, K. R.; Johnson, M.
G.; Lohr, M. T.; Wong, B. K. Drug Metab. Dispos. 2010, 38, 841.
5. Liu, J.; Li, A.-R.; Wang, Y.; Johnson, M. G.; Su, Y.; Shen, W.; Wang, X.; Lively, S.;
Brown, M.; Lai, S.; Gonzalez Lopez de Turiso, F.; Xu, Q.; van Lengerich, B.;
Schmitt, M.; Fu, Z.; Sun, Y.; Lawlis, S.; Seitz, L.; Danao, J.; Wait, J.; Ye, Q.; Tang, H.
L.; Grillo, M.; Collins, T. L.; Sullivan, T. J.; Medina, J. C. ACS Med. Chem. Lett. 2011,
2, 326.
6. Moskalev, N.; Makosza, M. Tetrahedron Lett. 1999, 40, 5395.
Please cite this article in press as: Johnson, M. G.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.04.092