A novel series of IKKβ inhibitors part II: Description of a potent and pharmacologically active series of analogs

Article abstract of DOI:10.1016/j.bmcl.2010.10.125

A novel series of (E)-1-((2-(1-methyl-1H-imidazol-5-yl) quinolin-4-yl) methylene) thiosemicarbazides was discovered as potent inhibitors of IKKβ. In this Letter we document our efforts at further optimization of this series, culminating in 2 with submicromolar potency in a HWB assay and efficacy in a CIA mouse model.

Full text of DOI:10.1016/j.bmcl.2010.10.125

Bioorganic & Medicinal Chemistry Letters 21 (2011) 423–426
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
A novel series of IKKb inhibitors part II: Description of a potent
and pharmacologically active series of analogs
Timothy D. Cushing a, , Vijay Baichwal b, Karen Berry d, Roland Billedeau e, Viola Bordunov e, Chris Broka e,
⇑
Michelle F. Browner g, Mario Cardozo c, Peng Cheng a, David Clark a, Stacie Dalrymple f,
Michael DeGraffenreid a, Adrian Gill e, Xiaolin Hao a, Ronald C. Hawley e, Xiao He a, Sharada S. Labadie e,
Marc Labelle a, Csaba Lehel b, Pu-Ping Lu a, Joel McIntosh e, Shichang Miao d, Camran Parast b,
Youngsook Shin a, Eric B. Sjogren e, Marie-Louise Smith a, Francisco X. Talamas e, George Tonn d,
Keith M. Walker e, Nigel P. C. Walker c, Holger Wesche b, Chris Whitehead e, Matt Wright d, Juan C. Jaen a
a Medicinal Chemistry, Amgen Inc., 1120 Veterans Blvd., South San Francisco, CA 94080, USA
b Inflammation Biology, Amgen Inc., 1120 Veterans Blvd., South San Francisco, CA 94080, USA
c Molecular Structure, Amgen Inc., 1120 Veterans Blvd., South San Francisco, CA 94080, USA
d Pharmacokinetics and Drug Metabolism, Amgen Inc., 1120 Veterans Blvd., South San Francisco, CA 94080, USA
e Chemistry, Roche Palo Alto, Hillview Ave., Palo Alto, CA 94304, USA
f Biology, Roche Palo Alto, Hillview Ave., Palo Alto, CA 94304, USA
g Discovery Technologies, Roche Palo Alto, Hillview Ave., Palo Alto, CA 94304, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A novel series of (E)-1-((2-(1-methyl-1H-imidazol-5-yl) quinolin-4-yl) methylene) thiosemicarbazides
was discovered as potent inhibitors of IKKb. In this Letter we document our efforts at further optimization
of this series, culminating in 2 with submicromolar potency in a HWB assay and efficacy in a CIA mouse
model.
Received 8 September 2010
Revised 22 October 2010
Accepted 25 October 2010
Available online 29 October 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
IKKb
IKKa
Semithiocarbazones
Collagen
Arthritis
Previously, we described some preliminary SAR centered on 1 a
compound found through high through put screening (HTS).1 Lead
compound 1 is ATP competitive, reversible, potent and selective
tion factor Nuclear Factor kappa B NF-jB. NF-jB controls much
of the innate immune system and is instrumental in the induction
of genes leading to a pro-inflammatory response.
over IKK
represented a promising lead with respect to its inhibition of IKKb
it demonstrated poor aqueous solubility (3.1 g/mL, PBS buffer)
a
, (IC50 IKKb = 0.25
l
M, IC50 IKK
a
= 9.19
l
M). Although 1
Assays used to develop the SAR in this program are a pair of
chemiluminescent ELISA assays against IKKb and IKKa, cellular
l
and cell potency. Described herein is our efforts focused on further
increasing the cellular potency and improving the physicochemical
properties culminating in the pre-development candidate
(Fig. 1).
2
The highly homologous IKKb and IKK
a are members of a small
class of protein kinases known as I
j
B Kinases (IKKs).2 They have a
distinct role in the regulation of both the adaptive and innate
immunity and are critical for most aspects of host defense.3 IKKb
and IKK
a activity are required for the activation of the transcrip-
⇑
Corresponding author.
E-mail address: tcushing@amgen.com (T.D. Cushing).
Figure 1. Lead compound 1, and advanced analog 2.
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.10.125

424
T. D. Cushing et al. / Bioorg. Med. Chem. Lett. 21 (2011) 423–426
Table 1
Table 3
SAR of the quinoline core
Preliminary PK
Compds
X
R1
R2
IKKb IC50 (lM)
No R3
R1
H
R2 HWB IC50
M)
Sol.
M)
Rat Cl (L/h/ Rat
(
l
(
l
kg)
%F
1
3
4
5
6
7
8
9
10
11
12
13
14
15
16
S
S
S
S
S
S
S
S
S
S
S
S
O
O
O
H
Cl
H
Cl
H
H
Cl
Cl
Me
H
Me
OMe
OCF3
CF3
CF3
Br
H
Br
CF3
0.25
0.35
0.08
0.018
0.12
0.23
0.04
0.37
0.38
0.08
0.16
0.05
2.95
0.24
0.36
27 n-
H
>30
>100
0.36
20
Bu
NC
NC
18 Me
20 Me
H
3.4
1.5
20
20
2.9
3.3
17
15
H
Me
Me
Me
Br
Me
OMe
OMe
OMe
OMe
Me
Cl
(4) had a threefold boost relative to 1 and the 6,7-dichloro analog
(5) was 12-fold more potent than the parent. Similarly, while the
mono-methyl analogs (6) and (7) had similar potencies, the 6,7-di-
methyl analog (8) was sixfold more potent than 1. Increasing the
steric demand at the 7-position with a methoxy group (9) was
not well tolerated. A similar effect was seen with a bulky EWD
group, –OCF3 (10). Using a less bulky EWD group, –CF3 for com-
pound 11 showed only a slight loss in potency relative to 8. In this
context, at the 6-position a methoxy group (12) was tolerated with
only a minimal loss in activity relative to 11. Exchanging the tri-
fluoromethyl group with the more electronegative bromine (13)
provided a threefold boost in potency relative to 12. As noted ear-
lier,1 semicarbazones are less potent than the corresponding
semithiocarbazones, but potency SAR follows a similar pattern as
noted for the semicarbazones 14–16 relative to compounds 13
and 11. Despite the enhanced biochemical potency of 6,7-disubsti-
tuted analogs such as 8, 11 and 13, cellular activities of all com-
assays to examine NF-jB activation via gel-shift induced TNF and a
human whole blood (HWB) induced by lipopolysaccharide (LPS).
We previously reported the effect of single substitutions at the
6 or 7 position of the quinoline core. In general, small lipophilic or
EWD groups enhanced potency when located at the 7-position but
substitution of the 6-position with EWD or EDG failed to improve
potency.1 We now report the impact of combined substitutions at
positions 6 and 7. For example, although 6-chloro analog (3) was
less potent than the parent (1), the corresponding 7-chloro analog
Table 2
SAR of the quinoline core
pounds in Table 1 were >1 lM.
Table 4
Improvements in solubility
No R1
R2
Cl
IKKb
IC50
IKK
IC50
a
NF
IC50
j
B
(l
HWB
M) IC50
M)
(lM)
(l
M)
(l
O
O
No R3
IKKb
IC50
IKK
IC50
M)
a
NF
IC50
M)
j
B
HWB
IC50
Sol. PBS
buffer
(lM)
17
0.003
0.041
0.4
1.85
S
(
lM)
(
l
(
l
(lM)
NC
NC
18
19
20
21
22
H
0.006
0.002
0.002
1.4
0.32
0.14
0.2
3.4
0.42
1.5
6
Me
Cl
0.14
0.071
28
2.7
30
30
—
—
>100
>100
NC
NC
NC
HO
CF3
F
—
—
—
0.8
29
0.24
4.8
10
HO
HO
—
0.3
0.7
7.8
30
0.28
0.24
6.2
30
5
16
9
>100
>100
23
24
Me
H
0.031
6.7
>30
1.4
1.8
0.6
5.5
0.5
0.3
MeO
MeO
O
31
3.1
0.9
MeO
0.094
0.012
0.013
>30
1.3
3.7
25 HO
Me
Me
32
0.026
4.3
2
>100
N
HO
26

T. D. Cushing et al. / Bioorg. Med. Chem. Lett. 21 (2011) 423–426
425
Figure 2. Synthesis of the key 6-cyanoethyl-7-chloroquinoline (2).
In our investigations, we found that replacing the semithiocar-
bazone N–H with a methyl group had a positive effect on the PK
properties and that flexible bulkier groups at the quinoline 6-posi-
tion enhanced the cellular activity.1 With this in mind we set out to
explore this SAR more fully. Several analogs containing longer,
flexible moieties at the quinoline 6-position coupled with small
EWG at the quinoline 7-position had sub-micromolar potency in
the NF-jB assay. These included a methylethylsulfone (17) a
methoxypropyl (23) a propanol (25) and an n-butanol (26). Inter-
estingly, analogs 18, 19, 20, 21 and 22 with an ethyl cyano moiety
Our attempts at improving the physicochemical properties were
encouraged by the PK profile of the n-butyl-semithiocarbazone
(27). Although this compound suffered from poor cellular potency
it had increased solubility, lowered rat clearance and reasonable
oral bioavailability. Unfortunately, the potent analogs 18 and 20
had lower solubility and the rat clearance was quite high, though
the effect on oral bioavailability was muted (Table 3). To explore
and extend these observations, a series of substituted semithiocar-
bazones were prepared. While simple alkyl side chains such as 28
did improve solubility, the effect on potency was minimal.
Improvements in potency came from alcohols 29, 30, and ethers
such as 31. However, amine containing analogs like 32 proved to
have the greatest increase in cellular activity while retaining solu-
bility (Table 4).
We then undertook an SAR exploration to determine this opti-
mal amine side chain. We chose the 6-cyanoethyl-7-chloroquino-
line scaffold because of it relatively superior physicochemical
properties and ease of synthesis (Fig. 2). The cellular activity was
only slightly less than the 7-methyl or 7-fluoro substitutions (20
vs 19 and 22). The incorporation of an amino-alkyl semithiocarbaz-
one gave many compounds with potency in the HWB assay below
at the 6-quinoline position were all <1.0
lM in the NF-kB assay and
19 and 22 had HWB potency <1.0 M. Generally, a 7-methyl group
l
was most favored relative to a single hydrogen, with the overall or-
der for potency in cells being Me > F > Cl > CF3 > H. For example,
while 18 and 19 were only threefold apart in the IKKb biochemical
assay and NF-jB cellular assay, 19 was eightfold more potent than
18 in the HWB assay. The methylpropyl ether analogs (23) and (24)
were found to be threefold apart in the biochemical assay but 23
was still ninefold more potent in the NF-
the HWB assay (Table 2).
jB assay and >4-fold in
0.5 lM, presumably by decreasing the lipophilicity thereby
decreasing the plasma protein binding (Table 5).
Table 5
As a comparison of 35 versus 36 and 33 versus 34 indicates,
there was an increase in solubility upon moving to a slightly longer
side chain. Only a few analogs had rat clearance values near or be-
low 50% liver blood flow (Table 5).
Improvements in cellular potency
No
R3
NF
j
B IC50
HWB IC50
M)
Sol. PBS
M)
Rat Cl
(L/h/kg)
(
lM)
(
l
(
l
N
N
33
34
0.05
0.08
0.4 0.1
0.2 0.1
>100
10
3.6–2.9
2.4
35
36
0.04 0.02
0.1
0.2 0.1
0.2
15
4.4
3.1
N
N
>100
2
0.02 0.01
0.05 0.02
0.095 0.07
0.2 0.1
>100
75
1.9
3.3
N
Figure 3. CIA mouse model with 2: collagen/CFA: administered (id) on day 0.
Analog 2 administered (100 mg/kg) orally on day 22–35. Clinical scoring: Visual
clinical score (4 paws) day 23–35: edema, erythema, severe joint damage. Scoring
for each paw 0–4 (normal–very severe).
37
N

426
T. D. Cushing et al. / Bioorg. Med. Chem. Lett. 21 (2011) 423–426
Table 6
PK parameters of 2
Rat (mean n = 3)
1.9
4.6
2.2
2.2
@3 mg/kg
15
Dog (mean n = 3)
Rhesus (mean n = 3)
Cl (L/h/kg)
Vdss (L/kg)
MRT (h)
1.3 0.4
4.4 3.5
3.3 1.7
3.3 0.4
@ 4 mg/kg
37 34
1.4 0.2
5.0 1.4
3.6 0.4
14.8 1.4
@ 4 mg/kg
Cmax
(l
g/L)
F%
5
0.5
Although 2 showed in vivo efficacy and excellent HWB activity, it
suffered from high intrinsic clearance (CLint) as judged by human
hepatocytes (high metabolizing donor) in the absence and presence
of plasma proteins7 and CYP P450 inhibition (2D6 = 0.54
l
M,
3A4 = 1.7
lM). Additionally, 2 had medium to high clearance across
three species and variable oral bioavailability (Table 6). In any case,
the results from the in vivo models coupled with the potency as
found in the HWB assay demonstrated a robust proof of concept
for this series. Further elaboration on the development of these ana-
logs will be described subsequently.
Figure 4. IBD mouse model with 2: N = 5 mice. DSS administered in H2O for 7 days.
2 administered orally 25 mg/kg and 10 mg/kg BID. Vehicle: 1% Tween, 80 + 1%
methycellulose. Colon erosion: reflects loss of surface epithelium, generally
associated with mucosal hemorrhage. Clinical scoring: 0 = normal, 1 = minimal,
*
2 tailed T test to
2 = mild, 3 = moderate, 4 = marked, 5 = severe. p 6 0.05
‘dss + vehicle’.
Acknowledgment
Lawrence McGee for discussions on the manuscript.
References and notes
Based on the HWB potency, the relatively low rat clearance, and
the solubility, compound 2 was chosen for testing in vivo. In our
primary in vivo model of NF-jB activation, (TNFa induction by
LPS stimulation in mice) analog 2 had an ED50 <0.1 mg/kg.4 Com-
pound 2 was further evaluated in a mouse collagen induced arthri-
tis (CIA) model.5 Analog 2 elicited a sharp MCS reduction relative to
the vehicle control with only a reversal to the severe state (score of
4) at day 33, still considerably below the vehicle treated group
(Fig. 3). As part of our program targeting diseases of inflammation
we were very interested in inflammatory bowel diseases (IBD)
such as ulcerative Colitis and Crohn’s disease. One of the most reli-
able models to determine drug effectiveness in these diseases is
the mouse dextran sulfate sodium (DSS) model of IBD.6 At the
10 mg/kg dose 2 lowered the colon erosion score by 38% and at
25 mg/kg it lowered the score by 56% relative to the DSS-vehicle
control (Fig. 4). Additional histopathology determinations were
made by scoring the colon inflammation and colon glandular epi-
thelial loss. With 2 the colon inflammation inhibition relative to
the DSS-vehicle control was 19% and 31% for doses of 10 and
25 mg/kg, respectively. The colon glandular epithelial loss was
28% and 36% for 10 and 25 mg/kg, respectively. The total of all
three histopathology scored parameters provided an inhibition
relative to the DSS-vehicle control of 27% and 40% for the 10
and 25 mg/kg doses, respectively. The PK exposure post 16–18 h
was 1.4 0.94 ng/mL and 8.5 15 ng/mL for the 10 mg/kg and
25 mg/kg doses, respectively.
1. See Cushing, T. D.; Baichwal, V.; Berry, K.; Billedeau, R.; Bordunov, V.; Broka, C.;
Cardozo, M.; Cheng, P.; Clark, D.; Dalrymple, S.; DeGraffenreid, M.; Gill, G.; Hao,
X.; Hawley, R. C.; He, X.; Jaen, J. C.; Labadie, S. S.; Labelle, M.; Lehel, C.; Lu, P.-P.;
McIntosh, J.; Miao, S.; Parast, C.; Shin, Y.; Sjogren, E. B.; Smith, M.-L.; Talamas, F.
X.; Tonn, G.; Walker, K. M.; Walker, N. C. P.; Wesche, H.; Whitehead, C.; Wright,
M.; Browner, M. F. Bioorg. Med. Chem. Lett. 2010, 21, 417. for details pertaining to
the assays, synthesis and initial SAR.
2. (a) Solt, L. A.; May, M. J. Immunol. Res. 2008, 42, 3; (b) Hacker, H.; Karin, M. Sci.
STKE 2006, 357; (c) Woronicz, J. D.; Gao, X.; Cao, Z.; Rothe, M.; Goeddel, D. V.
Science 1997, 278, 866; (d) Regnier, C. H.; Song, H. Y.; Gao, X.; Goeddel, D. V.; Cao,
Z.; Rothe, M. Cell 1997, 90, 373.
3. Bonizzi, G.; Karin, M. Trends Immunol. 2004, 25, 280.
4. BALB/C mice were treated with compound 2 (PO) at various doses, followed by
IP delivery of LPS in 0.5 h. Mice were bled 1.5 h later and their blood plasma
analyzed for TNFa levels by ELISA.
5. Collagen/CFA was delivered subcutaneously at day zero, followed by LPS (50 mg,
IP) stimulation on day 21. On day 22, compound 2 (100 mg/kg) in a HPMC
vehicle was administered orally (QD) until day 35. The study was blinded with
respect to treatment. Mean clinical score (MCS) was based on observed edema,
erythema and severe joint damage.
6. This experiment was performed by Alison Bendele at Bolder BIOPATH Inc. Swiss-
Webster female mice were administered 5% DSS in H2O for 7 days, followed by
plain H2O thereafter. The compound 2 was dosed at 25 mg/kg and 10 mg/kg
orally (BID) from day 0 until day 9. At sacrifice (day 10) the 3 mm colon tissues
biopsy was incubated for 24 h and the resulting supernatant assayed for TNF
and IL-1B via ELISA. The colons were clinically scored and tissues evaluated
a
histopathologically.
*
7. CLint
:
ÀHSA/AAG = 3.4/6.8
; +HSA/AAG = 2.1/2/4 l
L/min 106 Cells. {[HSA
(human serum albumin)] = 40 mg/mL, [AAG (a1-acid glycoprotein) = 2 mg/mL]}.