IRAK-4 inhibitors. Part III: A series of imidazo[1,2-a]pyridines

Article abstract of DOI:10.1016/j.bmcl.2008.04.042

Following the identification of a potent IRAK inhibitor through routine project cross screening, a novel class of IRAK-4 inhibitor was established. The SAR of imidazo[1,2-a]pyridino-pyridines and benzimidazolo-pyridines was explored.

Full text of DOI:10.1016/j.bmcl.2008.04.042

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 3656–3660
IRAK-4 inhibitors. Part III: A series of imidazo[1,2-a]pyridines
George M. Buckley, Richard Fosbeary, Joanne L. Fraser, Lewis Gowers,
Alicia P. Higueruelo, Lynwen A. James, Kerry Jenkins,* Stephen R. Mack,
Trevor Morgan, David M. Parry, William R. Pitt, Oliver Rausch,
Marianna D. Richard and Verity Sabin
UCB, Granta Park, Great Abington, Cambridge CB21 6GS, UK
Received 12 March 2008; revised 15 April 2008; accepted 18 April 2008
Available online 24 April 2008
Dedicated to the memory of our friend and colleague David Rainey.
Abstract—Following the identification of a potent IRAK inhibitor through routine project cross screening, a novel class of IRAK-4
inhibitor was established. The SAR of imidazo[1,2-a]pyridino-pyridines and benzimidazolo-pyridines was explored.
ꢀ 2008 Elsevier Ltd. All rights reserved.
In the preceding article we demonstrated how a cross
screening hit from another kinase programme yielded
the potent imidazo[1,2-a]pyridino-pyrimidine IRAK-4
inhibitor 1.1 This led to lead compounds 2 (imi-
dazo[1,2-a]pyridino-pyridine) and 3 (benzimidazolo-
pyridine) as our preferred scaffolds for optimisation.
Substitution in the 6-position (R2) resulted in significant
increase in IRAK-4 potency (5- to 25-fold) whereas sub-
stitution at the 5-position (R1) reduced potency irrespec-
tive of whether the 6-position was substituted (Table 1).
With this in mind we embarked on the more involved
synthesis of imidazo[1,2-a]pyridino-pyridines substi-
tuted in the regio-equivalent position (Scheme 2 and Ta-
ble 2). A variety of groups were tolerated in this
position, many giving low nanomolar enzyme inhibition
potencies (selected examples shown, 12–24). The SAR
was not very discernable except that the bulky tertiary
amides were less potent than the primary amide
(23 < 24 < 22). Homology model docking experiments
of this series imply that the 6-substituent points towards
the ATP-ribose region.1 It is plausible that the flat SAR
might be a result of the displacement of a bound water
molecule from this site, thus gaining an entropic advan-
tage in binding which would be largely independent of
spatial complementarily or bonding interactions.
N
N
N
N
N
N
N
NH
N
NH
N
NH
N
N
H
N
H
N
H
1, IRAK-4, IC50 216nM
3, IRAK-4, IC50 70nM
2, IRAK-4, IC50 35nM
Initially the SAR of the 5,6-fused bicyclic binding group
was studied on the benzimidazole analogue series due to
the simplicity of the chemistry and readily available
substituted starting materials.2 A series of analogues
were prepared from commercially available benzimidaz-
oles and 2,6-difluoro or 2,6-dibromopyridine by sequen-
tial SNAr displacement reactions and separation of
regioisomers as required (Scheme 1).
We also focused attention on the pyridine 2-position
substituent. Much of the SAR could be probed by
employing parallel synthesis (Scheme 3). The key bro-
mopyridine intermediate 25 was prepared via Stille cou-
pling3 of the tri-n-butylstannane formed in situ from
trans-metallation of the Grignard derived from 3-bromo
imidazo[1,2-a]pyridine. A variety of N-linked deriva-
tives were prepared using an automated serial micro-
wave reactor,4 either by direct SNAr chemistry
(method b) or by using Buchwald palladium catalysed
Keywords: IRAK; IRAK-4; IRAK-4 inhibitor; Kinase; Kinase inhib-
itor; Imidazo[1,2-a]pyridine; Inflammation; Anti-inflammatory; Liga-
nd efficiency; IL-1; TNF-alpha; TNFa; Immunity.
*
Corresponding author.
drkerryjenkins@yahoo.co.uk
Tel.:
+44 1223 896400.; e-mail:
0960-894X/$ - see front matter ꢀ 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.04.042

G. M. Buckley et al. / Bioorg. Med. Chem. Lett. 18 (2008) 3656–3660
3657
R1
H2N
N
R1
R2
5
R1
R2
R2
N
boc
N
N
H
N
X
6
N
N
N
c, d
a, b
N
N
X
NH
N
H
X
X = Br or F
3-11
Scheme 1. Synthesis of substituted benzimidazolo-pyridines 3–11. Reagents and conditions: (a) NaH, NMP, 180 ꢁC, microwave (33–90%); (b) HPLC
separation of regioisomers as required; (c) 4-amino-1-Boc piperidine, NMP, 140 ꢁC, microwave (13–75%); (d) HCl, Et2O, DCM.
Table 1. IRAK-4 enzyme inhibition for benzimidazolo-pyridines 3–11
potencies (Table 3). Some examples of secondary linked
alkylamines indicated a preference for a free –NH–/NH2
spaced approximately 2–3 carbon atoms from the link-
ing amino group (e.g., 32 > 31 and 34 > 33). The
IRAK-4 potency varied markedly with quite subtle
changes to the positioning of the distal amino group.
For example, the bicyclic analogue 34 was an order of
magnitude less potent than 4-aminopiperidine, 2, whilst
azetidine 37 and 3-(S)-piperidyl 39 had comparable po-
tency. The 3-linked pyrrolidine rac–38 and methylene
spaced 2-linked pyrrolidine rac-36 were ꢀ10-fold less
potent than 2 and the methylene spaced 4-linked piper-
idine 35 is 100-fold less potent. Tertiary N-linked alkyl-
amines (40–43) demonstrated that a 2-aminopyridine
free NH was not an essential prerequisite to binding,
consistent with our earlier findings.1 However, these
compounds do not represent especially potent IRAK-4
inhibitors. Once again, the precise positioning and con-
straints of the distal free amino NH/NH2 appeared to be
important for good IRAK potency (rac-42 > 41).
Compound
R1
R2
IRAK-4 inhibition,
IC50 (nM)
3
H
H
70
4
5
6
H
OMe
H
OMe
15
208
181
OMe
OMe
7
8
9
H
Cl
H
3
112
166
Cl
Cl
Cl
10
11
H
Me
Me
H
10
1085
couplings (method c).5 Bifunctional bis-amines were
suitably protected with tert-butoxycarbonyl groups.
The products were purified by mass directed preparative
HPLC6 following deprotection step (d) where required.
Having assessed the SAR of substituents at different
ends of the molecule the obvious next step was to com-
bine preferred substituents in the same structure (Table
The series of aniline and heteroaryl analogues (26–29)
showed some interesting SAR (e.g., 2-PhCN > Ph ꢀ 2-
N-methylpyrazol-3-yl > 4-PhCN), albeit with modest
R1 = Cl; X = I; a, b (tBu3PHBF4), c
12, R1 = Cl, R2 = H
N
R1
N
R1 = OMe; X = Br; a, b (Cy2biphen), c
Br
X
d
13, R1 = OMe, R2 = H
boc
N
N
14
k, c
k, c
k, c
k, c
18, R1 = 2-Methoxypyridin-5-yl, R2 = H
19, R1 = 1-Methylpyrazol-4-yl, R2 = H
N
H
N
R1
N
20, R1 = Pyrazol-3-yl, R2 = H
21, R1 = Pyrrol-2-yl, R2 = H
R2
N
N
e
15, R1 = Br, R2 = Boc
16, R1 = CN, R2 = H ref.7
17, R1 = CN, R2 = Boc
N
H
j
f
22, R1 = CONH2, R2 = H
g, h, i, c
g, h, i, c
O
23, R1 =
2
N
O
, R = H
24, R1 = CONMe2, R2 = H
Scheme 2. Synthesis of 6-substituted imidazo[1,2-a]pyridin-3-yl-pyridin-2-ylamines, 12–24. Reagents and conditions: (a) i-PrMgCl, THF, À78 ꢁC; n-
Bu3SnCl, À78 ꢁC to rt; 2,6-dibromopyridine, Pd(PPh3)4, 140 ꢁC, microwave (11–21%); (b) 4-amino-1-Boc piperidine, NaOtBu, Pd(OAc)2;
t-Bu3PHBF4 DME, 140 ꢁC, microwave (47%) or 2-(dicyclohexylphosphino)biphenyl (Cy2biphen), toluene, 140 ꢁC, microwave (44%); (c) HCl, Et2O,
DCM, MeOH (40–63%); (d) i-PrMgCl, THF, À78 ꢁC; n-Bu3SnCl, À78 ꢁC to rt; 14, Pd(PPh3)4, reflux or 140 ꢁC, microwave;7 (e) Boc2O, Et3N, DCM
(61%); (f) KOH, THF, H2O, 140 ꢁC, microwave; (g) 20% (aq) H2SO4, 140 ꢁC, microwave (57%); (h) Boc2O, NaOH, dioxane, H2O (96%); (i)
morpholine or dimethylamine (2.0 M in THF), EDCI, HOBt, DCM (38–43%); (j) TFA, DCM (57%); (k) ArB(OH)2 or ArB(OCMe2)2, Na2CO3,
n-Bu4NBr, Pd(PPh3)4, DME, H2O, 140 ꢁC, microwave (47–50%).

3658
G. M. Buckley et al. / Bioorg. Med. Chem. Lett. 18 (2008) 3656–3660
Table 2. IRAK-4 enzyme inhibition for 6-substituted imidazo-[1,2-
a]pyridin-3-yl-pyridin-2-ylamines, 12–24
Table 3. IRAK-4 enzyme inhibition for 2-amino substituted imidazo-
pyridino-pyridines 26–43
Compound
R1
IRAK-4 inhibition,
IC50 (nM)
Compound
NR1R2
IRAK-4 inhibition,
IC50 (nM)
12
13
16
Cl
OMe
CN
1
6
4
26
27
28
1156
433
N
H
18
19
8
7
N
H
MeO
N
N
CN
CN
(12% inhibition
at 10 lM)
N
N
H
20
21
2
N
N
N
N
H
29
1134
N
H
63
NH
N
H
N
N
30
31
32
497
N
H2N
O
22
5
O
N
N
8100
1300
H
N
H
23
24
263
49
NH2
O
H
H
N
N
33
34
914
225
H
O
N
H
NH
N
N
N
H
N
H
b or c
N
a
N
N
N
then d
N
R1
N
H
(if appropriate)
35
3030
313
Br
N
NH
Br
R2
25
26 - 43
H
N
rac-36
N
H
Scheme 3. Parallel synthesis of 2-amino substituted imidazopyridino-
pyridines 26–43. Reagents and conditions: (a) i-PrMgCl, THF, À78 ꢁC;
n-Bu3SnCl, À78 ꢁC to rt; 2,6-dibromopyridine, Pd(PPh3)4; 66 ꢁC
(71%); (b) R1R2NH, Et3N, n-BuOH, 150 ꢁC, microwave; or (c)
R1R2NH, Pd(OAc)2, Cy2biphen, NaOtBu, toluene, 150 ꢁC, micro-
wave; (d) HCl, Et2O, DCM (77–97%).
NH
37
39
N
H
H
N
rac-38
545
N
H
4). The compounds were prepared by constructing the
substituted imidazopyridine from condensing function-
alised 2-aminopyridines with chloroacetaldehyde8 and
halogenating regioselectively in the 3-position with N-
iodo or N-bromosuccinmide.9 Once again, Stille cross-
coupling and Buchwald conditions were used to con-
struct the final compounds (Scheme 4).
H
N
abs-39
abs-40
19
N
H
H
H
N
126
NH
In the examples given it was apparent that the SAR was
not additive. Having a substituent in the imidazopyri-
dine 6-position generally increased potency by 4- to
500-fold (44–47). This perhaps suggested the relative
dominance of the 6-substituent in terms of binding over
the pendant 2-amino heterocycles. However, compound
48 is an example where this is not the case. In the ab-
sence of a 6-substituent (R1 = H, 39), potency is a
respectable 19 nM. Introduction of the nitrile group in
the 6-position dropped potency by ꢀ5-fold. This high-
N
41
1880
535
NH
2
NH
2
N
N
rac-42
43
487
NH

G. M. Buckley et al. / Bioorg. Med. Chem. Lett. 18 (2008) 3656–3660
3659
with few solubility issues,13 and the in vitro DMPK
looked promising with low to moderate microsomal
clearance in both human and rat. Cytochrome P450 lia-
bilities were not a major issue, with many compounds
exhibiting only very weak CYP inhibition.14
Table 4. IRAK-4 inhibition of hybridized substituted imidazo[1,2-a]-
pyridino-pyridines 44–48
Compound
R1
R2
IRAK-4, IC50 (nM)
6
44
Cl
NH
NH
rac-45
rac-46
47
Cl
1
9
7
We were encouraged by the fact that this series of com-
pounds exhibited high enzyme potencies (frequently
<10 nM at 10 lM ATP), whilst still retaining low molec-
ular weights (300–350). This represents high Ligand Effi-
ciency (LE)15 and suggests further optimisation could be
achieved, designing in selectivity elements whilst retain-
ing IRAK potency and desirable physicochemical prop-
erties. For example, compound 16 has a LE of 0.52
(Table 5). As a point of reference the average kinase
inhibitor has a LE of 0.3.16 The calculated LogP and
Cl
NH
NH
OMe
NH
48
CN
95
H
2
˚
PSA are all respectably low (aLogP < 3, PSA < 85A ),
giving plenty of scope for further elaboration.17
lights subtleties in SAR, where optimising two remote
parts of a molecule independently sometimes results in
non-additive effects.
We have described how an active hit identified through
routine cross screening was developed into a new class
of IRAK-4 inhibitor. The SAR of the new lead series
was explored with guidance from the IRAK-4 homology
model developed in-house. Some highly potent (low
nanomolar) IRAK-4 inhibitors with good cellular
TNFa inhibition were prepared and were shown to have
good drug-like properties and encouraging in vitro
DMPK profiles. Initial kinase panel profiling and cell
screening has indicated that some degree of off-target
activity is still evident with this series, despite departing
from the aminopyrimidine motif. However, the scope
for fine tuning and optimising this potent and efficient
class of IRAK inhibitor could lead to the generation
of new therapeutic agents or useful tool compounds
based on the blockade of the IL-1 driven immune
response.
Compounds that exhibited significant IRAK-4 potency
were submitted into a screening cascade to build up a
fuller profile and to prioritise compounds for in vivo
studies (Table 5). This included an IRAK-1 enzyme as-
say.10 IRAK-1 is a related kinase homologue that is
implicated in the same signalling pathway and hence a
dual inhibitor of both IRAK-4 and IRAK-1 could offer
synergistic benefits.11 A variety of functional cell assays
were also used to assess potency, such as LPS-induced
TNFa cytokine release assay.12 Potency in the IRAK
enzyme assays tracked reasonably well with cytokine re-
lease in the cell assays. Good cellular IC50s of 10s–100s
nanomolar were often observed. The physicochemical
profile of this series of compounds was generally good,
N
R1
N
R1
N
N
NH2
N
c
a, b
d, e
N
R1
N
R1
N
N
R2
N
X
Br
H
X = I,Br
44-48
Scheme 4. Synthesis of hybridised substituted imidazo[1,2-a]pyridino-pyridines 44–48. Reagents and conditions: (a) chloroacetaldehyde, NaHCO3,
H2O (45–55%); (b) NIS or NBS, MeCN (18–64%); (c) i-PrMgCl, THF, À78 ꢁC; n-Bu3SnCl, 66 ꢁC; 2,6-dibromopyridine, Pd(PPh3)4, 150 ꢁC,
microwave; (28–44%); (d) Boc-R2–NH2, Pd(OAc)2, dicyclohexyl-1-biphenylphosphine, NaOtBu, microwave 140 ꢁC toluene; (e) HCl, DCM, Et2O,
MeOH.
Table 5. Selected in vitro/cellular potency, in vitro DMPK data and calculated properties
Compound MWt 2D
aLogPa IRAK-4 Ligand
IRAK-1 TNFa
Mic. Clint, Mic. Clint, CYP 2D6
(rat) at
CYP 3A4
inhibition
(lM)
2
PSAa (A )
IC50
(nM)
efficiency IC50
(LE)b
inhibition (hu) at
0.5 lMc
inhibition
(lM)
˚
(nM)
(nM)
0.5 lMc
13
16
323
318
300
314
328
68.1
82.6
55.5
60.1
60.1
1.7
1.6
2.3
2.3
2.9
6
4
6
1
9
0.52
0.54
0.48
0.48
0.47
293
96
220
200
54
19.5
8.3
5.3
39.9
0.01
49.3
78.1
41% at 50 lM 31
42 22
21% at 10 lM 11
44
rac-45
rac-46
195
23
13
16
65
19.2
29.5
15
1.8
12
175
8.6
a alogP and 2D PSA calculated with Scitegic Pipeline Pilot software. http://www.scitegic.com.
b Ligand efficiencies approximated from IC50 values: LE (kcal molÀ1) = {[ÀRT lnIC50 (mol LÀ1)]/# of non-hydrogen atoms}.15
c Units for micosomal intrinsic clearance, lL/min/mg of protein.

3660
G. M. Buckley et al. / Bioorg. Med. Chem. Lett. 18 (2008) 3656–3660
8. Lo¨ber, S.; Hubner, H.; Gmeiner, P. Bioorg. Med. Chem.
¨
Lett. 1999, 9, 97.
Acknowledgements
9. Enguehard, C.; Renou, J.-L.; Collot, V.; Hervet, M.;
Rault, S.; Gueiffier, A. J. Org. Chem. 2000, 65, 6572,
Compound 47 derived via NBS bromination (X = Br)
whilst compounds 44–46 and 48 derived via analogous
NIS iodination (X = I).
The authors thank the analytical support group. In par-
ticular Jags Matharoo (mass directed preparative
´
´
HPLC), Paloma Fernandez Dıaz (LC–MS) and Phil Gil-
bert (NMR). We also thank the DMPK group for per-
forming the microsome stability and CYP inhibition
assays (Andy Sykes and Helen Brown) and the protein
expression group (David Rainey and Alistair Durrant)
for supplying IRAK-1 and IRAK-4 for the enzyme as-
says and crystallographic trials.
10. The IRAK-1 in vitro enzyme assay determined the effect of
test compounds on the phosphorylation of a biotinylated 32
amino acid peptide based on the activation loop of IRAK1,
using FlashPlates to detect incorporation of 33P. It was
performed in 96-well streptavidin coated FlashPlates in a
volume of 100 ll, comprising of reaction buffer (50 mM
Hepes, 10 mM MgCl2, 5 mM EGTA, 1 mM DTT, pH 7.2),
1 lM ATP, 0.5lCi/well 33P-ATP, 2 lM peptide substrate
(Biotin-DFGLARFSRFAGSSPSQSSMVARTQ-TVRG
TLA, Peptide Protein Research Ltd), 8 nM full length GST-
His-IRAK1 (in-house) and test compound in 2% DMSO.
The kinase reaction was incubated for 90 min at room
temperature, then terminated by addition of 100 ll 100 mM
EDTA. After a further 30 min incubation to maximise
peptide capture, the plate was washed three times with 0.1%
Tween 20 in PBS. Incorporation of 33P into the peptide
substrate was measured using a TopCount plate reader.
11. Lye, E.; Mirtsos, C.; Suzuki, N.; Suzuki, S.; Yeh, W. C. J.
Biol. Chem. 2004, 279, 40653.
12. Peripheral blood mononuclearcytes (PBMC) were isolated
form human blood by density gradient centrifugation.
Cells were treated with IL-1b in the presence of test
compounds for 24 h. Supernatants were collected and the
levels of soluble cytokines analysed via multiplexed
particle based flow cytometric assay (Luminex).
13. Solubility measured by AKAS determination from DMSO
stock solutions. Of 44 compounds from this article that
were tested, 40 had >999 lM solubility at pH 7.4.
14. Hollenberg, P. F. Drug Metab. Rev. 2002, 34, 17.
15. Hopkins, A. L.; Groom, C. R.; Alex, A. Drug Discovery
Today 2004, 9, 430.
References and notes
1. Buckley, G. M. et al. Bioorg. Med. Chem. Lett. 2008,
doi:101016/j.bmcl.2008.04.039.
2. An example of a similar chemotype to the benzimidazolo
pyridines has been reported by Pfizer as a PDGFR
inhibitor (CP-673,451) but no binding hypothesis is
offered. Cancer Res. 2005, 65, 957.
3. (a) Stille, J. K. Angew. Chem. Int. Ed. Engl. 1986, 25, 508;
(b) Farina, V.; Krishnamurthy, V.; Scott, W. J. Org.
React. 1998, 50, 1; (c) Mitchell, T. N. Synthesis 1992, 803.
4. Biotage (formerly Personal Chemistry) Emrys Opti-
mizerTM. Reaction times varied from 5 min to >1 h.
Individual reactions were periodically monitored by LC–
MS until complete conversion or emergence of undesired
by-products.
5. Wolfe, J. P.; Buchwald, S. L. Org. Synth. 2004, 10, 423;
Wolfe, J. P.; Buchwald, S. L. Org. Synth. 2002, 78, 23.
6. Waters Autoprep. Mass (ESI positive ion mode) and u.v.
directed fraction collection. Phenomenex Lunaꢂ station-
ary phase (5l C18(2); 250 · 21.2 mm) against acetonitrile
gradient with either formic acid (0.8 vol %) or ammonium
acetate (10 mM) mobile phase.
7. Under the high temperature (140 ꢁC) Suzuki coupling
conditions compound 16 (R1 = CN) was formed directly
with loss of the Boc protecting group whilst compound 15
(R1 = Br) retained the Boc group. This may be due to
relative substrate stabilities or subtle experimental varia-
tions between the two reactions. In any case, reinstallation
and removal of the Boc groups was synthetically trivial.
16. Paolini, G. V.; Shapland, R. H.; van Hoorn, W. P.;
Mason, J. S.; Hopkins, A. L. Nat. Biotechnol. 2006, 24,
805.
17. (a) Leeson, P. D.; Springthorpe, B. Nat. Rev. Drug Disc.
2007, 6, 881; (b) Lipinski, C.; Lombardo, F.; Dominy, B.;
Feeney, P. Adv. Drug Delivery Rev. 1997, 23, 3.