Discovery of novel PI3Kγ/δ inhibitors as potential agents for inflammation

Article abstract of DOI:10.1016/j.bmcl.2012.05.121

Dual PI3Kγ/δ inhibitors have recently been shown to be suitable targets for inflammatory and respiratory diseases. In a recent study we described the discovery of selective PI3Kγ inhibitors based on a triazolopyridine scaffold. Herein, we describe the elaboration of this structural class into dual PI3Kγ/δ inhibitors with excellent selectivity over the other PI3K isoforms and the general kinome. Structural optimization led to the identification of two derivatives which showed significant efficacy in an acute model of lung inflammation.

Full text of DOI:10.1016/j.bmcl.2012.05.121

Bioorganic & Medicinal Chemistry Letters 22 (2012) 4546–4549
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of novel PI3Kc/d inhibitors as potential agents for inflammation
Katie Ellard a, , Mihiro Sunose a, Kathryn Bell a, Nigel Ramsden a, Giovanna Bergamini b, Gitte Neubauer b
⇑
a Cellzome Ltd, Chesterford Research Park, Little Chesterford CB10 1XL, UK
b Cellzome AG, Meyerhofstr. 1, 69117 Heidelberg, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Dual PI3K
diseases. In a recent study we described the discovery of selective PI3K
pyridine scaffold. Herein, we describe the elaboration of this structural class into dual PI3K
with excellent selectivity over the other PI3K isoforms and the general kinome. Structural optimization
led to the identification of two derivatives which showed significant efficacy in an acute model of lung
inflammation.
c
/d inhibitors have recently been shown to be suitable targets for inflammatory and respiratory
inhibitors based on a triazolo-
/d inhibitors
Received 8 May 2012
Revised 29 May 2012
Accepted 30 May 2012
Available online 9 June 2012
c
c
Keywords:
PI3Kc/d
Ó 2012 Elsevier Ltd. All rights reserved.
Triazolopyridine
Ureas
Inflammation
Asthma
Phosphatidylinostol-3 kinases (PI3Ks) are a family of enzymes
which phosphorylate the 3-hydroxyl position of the inositol ring
of phosphatidylinositol upon activation by various cell surface
receptors.1 The resultant phosphoinositides play important roles
in lipid and cell signaling and membrane trafficking, hence the
PI3Ks have been implicated in many cellular functions such as
growth, proliferation, survival, apoptosis, adhesion and migra-
tion.2,3 As such the development of inhibitors has generated vast
interest in many areas such as oncology, metabolic, cardiovascular
and, in particular, inflammatory diseases where they can be poten-
tially targeted towards conditions such as systemic lupus erythe-
matosus, psoriasis, rheumatoid arthritis, asthma and chronic
obstructive pulmonary disorder (COPD).3,4
Figure 1. TG100-115 with published PI3K activity (pIC50).
infarction7 and has been investigated in inhaled models of asthma
and COPD.11
In a recent publication we described the elaboration of lead
compound CZC19091 2 to CZC19945 4 (Fig. 2). The former suffered
from poor cellular activity and bioavailability whereas 4 had good
cellular potency, pharmacokinetics and in vivo efficacy in inflam-
matory models. Although it was only moderately selective over
the PI3K family 4 was selective over the general kinome.12 This
account describes the use of this lead in order to develop potent
The most extensively investigated are the Class I PI3Ks, which
can be further subdivided into class IA isoforms (
the sole class IB member ( ). PI3K and b are ubiquitously
expressed and mice deficient in their catalytic sub-units are embry-
onically lethal. In contrast, the expression of d and are limited to
a, b, and d) and
c
a
c
hematopoietic and endothelial cells and loss of either or both
isoforms yields viable offspring, albeit with compromised immu-
PI3Kc/d inhibitors that would be suitable for indications such as
asthma and COPD.
nity.3,5 The limited expression pattern of PI3K
c and PI3Kd as well
as the viability of the genetically modified, kinase dead mice makes
the dual inhibitor approach attractive and yet there are only a few
examples of such compounds in the literature.6–10 One compound,
TG100-115 1 (Fig. 1) has entered clinical trials as an intravenously
administered drug to restore blood flow following acute myocardial
Analysis of the crystal structure that we obtained of CZC19091 2
complexed to PI3Kc c inhibitors
13 (Fig. 3) revealed that, as in all PI3K
with structural data, the compound made a back-bone interaction
with the hinge region (Val 882)6,14 and that the acetyl group
pointed toward bulk solvent, on the periphery of the ATP pocket.
It has been reported that although the interior of the ATP binding
pocket is highly conserved between the PI3K family members, there
are major differences in the residues on the boundary, giving
significant changes in both shape and charge in this region.14 We
⇑
Corresponding author.
E-mail address: catherineellard@aol.com (K. Ellard).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.05.121

K. Ellard et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4546–4549
4547
suitable for further characterization. Ureas were selected in prefer-
ence to amides to circumvent the in vivo instability that had been
observed for CZC19091.12 Our approach was complimented by
our Kinobeads technology where binding to the target protein is
evaluated in a cell lysate, avoiding the need for recombinant protein
and allowing access to all required isoforms in a single experiment.
This allowed selectivity to be evaluated early in the assay cascade
so that it could be rapidly incorporated into inhibitor design. Cellu-
lar activity was assessed in an fMLP-induced human neutrophil
migration assay.12
We chose the CZC19945 4 in preference to the CZC19463 3 sub-
structure as our starting scaffold as this was the most potent
(in vitro and in vivo) and had the most desirable selectivity profile
with regards to undesired off-targets. We had also developed a
reliable and scalable route to this intermediate devoid of any
column chromatography (Scheme 1). Preparation of sulfonamide
6 was achieved by reaction of 5-bromopyridine-3-sulfonyl chloride
5 with tert-butylamine, in pyridine. Subsequent Suzuki coupling
with 2-aminopyridine-5-boronic acid pinacol ester gave aminopyr-
idine 7 which underwent reaction with ethoxycarbonyl isothiocy-
anate to give the corresponding thiourea 8. Cyclization with
hydroxylamine in a 1:1 mixture of refluxing ethanol/methanol
afforded 4 which could be isolated directly from the reaction mix-
ture, with high purity, by simple filtration.
Figure 2. PI3K activity (pIC50) of triazolopyridine lead compounds.
In a facile two step urea formation process, 4 was first reacted
with trisphosgene in pyridine to afford the intermediate isocya-
nate, which was subsequently dissolved in DMF and transferred
into reaction vials containing the desired amines. The reactions
were then heated at 65 °C overnight and the desired ureas isolated
by preparative HPLC directly from the reaction mixture. This meth-
od enabled the rapid, high-purity synthesis of the desired ureas in
a parallel fashion. The glycinamide ureas were accessed via inter-
mediate acid 10 which was synthesized via reaction of the in-situ
generated isocyanate with glycine. Standard amide bond formation
with HATU then furnished the desired products.15
Figure 3. Crystal structure of CZC19091 with PI3Kc showing the acetyl group on
the periphery of the ATP binding pocket, pointing toward bulk solvent.
To probe the SAR a diverse range of amines were selected; key
examples are shown in Table 1. The simple methyl urea 9a was
tolerated whereas the morpholine 9b showed a drop in both
PI3K potency and cell-based activity. This was also observed for
other simple trisubstituted ureas (data not shown) and therefore
disubstituted ureas became the focus of our synthetic efforts.
therefore postulated that the acetyl moiety, as well as being an
amenable synthetic handle, would be a logical area to modify in
order to obtain PI3K inhibitors with diverse selectivity profiles
Scheme 1. Reagents and conditions: (a) tert-Butylamine, pyridine, 0–40 °C; (b) 2-aminopyridine-5-boronic acid pinacol ester, Pd(dppf)(Cl)2ÁDCM, Na2CO3, DME:EtOH:H2O
(5:3:2),
wave, 120 °C, 1 h; (c) ethoxycarbonyl isothiocyanate, DCM, 35 °C, 24 h; (d) NH2OHÁHCl, DIPEA, MeOH:EtOH (1:1), 80 °C, 18 h; (e) triphosgene, THF:pyridine, 0–35 h,
2 h; (f) R1R2NH, DMF:pyridine (10:1), 65 °C, 18 h; (g) R3R4NH, HATU, DIPEA, DMF, 65 °C, 18 h.
l

4548
K. Ellard et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4546–4549
Table 1
Table 2
Ureas synthesized via Scheme 1 with PI3K potency as measured in the Kinobeads
assay
Heterocyclic and glycinamide ureas synthesized via Scheme 1 with PI3K potency, as
measured in the Kinobeads assay
O
R1
O
R1
N
N
N
N
N
N
NH R2
NH R2
N
N
N
N
O
S
O
O S O
HN
HN
Compound
NR1,R2
n/a
PI3K (pIC50
)
Neu
mig
(pIC50
Compound NR1, R2
PI3K (pIC50
)
Neu
mig
(pIC50)
ca
da
aa
ba
ca
da
aa
ba
a
a
)
4
7.6
7.9
5.8
6.1
5.6
6.7
6.4
6.0
6.1
N
N
N
H
9j
N
H
7.6 6.7 6.2
8.1 7.8 6.1
8.2 7.6 5.9
6.2
6.8
9a
<5.3
<4.6
N
S
9b
9c
6.9
6.9
5.5
5.9
5.4
5.1
4.9
O
9k
<6.7 6.4
N
N
H
N
N
<5.3
5.9
<6.0
N
H
N
9l
N
6.9
6.6
7.2
6.9
6.5
O
N
H
9d
7.8
7.7
6.7
7.3
N
H
N
N
9m
8.0 7.1 5.5
8.6 8.2 6.1
O
N
H
9e
9f
6.6
7.2
7.5
<4.6
5.8
<4.0
<4.7
<5.5
<4.0
5.5
—
N
H
N
N
N
9n
7.1
6.6
N
H
N
N
5.8
6.4
H
N
N
N
N
9o
7.2 7.3 <5.2 5.6
H
9g
7.6
6.0
N
N
H
H
N
O
N
N
H
11b
11c
7.8 6.1 <4.8 <5.7 4.7
9h
7.6
8.5
7.9
7.1
5.7
4.9
7.1
5.9
6.6
6.1
O
N
H
N
8.2 7.2 6.0
8.4 7.8 6.4
5.9
6.4
5.8
6.3
N
H
N
11a
N
H
O
O
O
O
O
N
a
N
N
The values are averages of at least two independent experiments.
11d
11e
N
H
Although the dimethylamino ethyl moiety 9c showed a drop in
potency, this was restored with the corresponding morpholine
9d which also showed increased affinity for PI3Kd as well as mod-
erate selectivity over the other isoforms. Aryl 9e and benzyl ureas
9f showed a drop in PI3Kd activity, but the latter was restored with
analogues 9g and 9h with the ethyl linker which again showed a
8.6 7.9 6.1
6.0
5.9
N
H
a
The values are averages of at least two independent experiments.
PI3K
c
/d dual inhibition profile. Finally glycinamide 11a, was our
homologation to the propyl spacer 9o gave a drop in activity across
all isoforms.
first example with PI3K
c
pIC50 greater than eight. PI3Kd activity
was moderate and selectivity over the other isoforms was more
than 100 fold.
The glycinamides (11a, Table 1 and 11b–e, Table 2) generally
showed enhanced selectivity over PI3Kb than the heterocyclic
Encouraged by the PI3Kc/d profile obtained and the potential
subseries and retained excellent potency for PI3Kc and d. Tertiary
therapeutic applications for dual inhibitors,2,9,16 we decided to
elaborate on two of the urea sub-series optimizing activity and
selectivity for both isoforms. Data for selected compounds is
shown in Table 2, initially for the alkyl heterocycles (9j–9o) and
then the glycinamides (11b–11e).
amides, 11a, c–e, showed superior cell-based activity relative to
their secondary counterparts (11b), but in general there were no
gross differences in PI3K activity between them. To further assess
selectivity, in a competition binding experiment in which over
100 kinases were detected, aside from the Class I PI3Ks, 11d only
The pyrazole 9j retained similar PI3Kc potency to the parent 4,
exhibited off-target activity against 1 kinase (PIK4Ca) demonstrat-
ing that the high selectivity extended into the general kinome.17
Compounds from each sub-series were then further investigated
for microsomal stability, solubility and permeability (Table 3).
Perhaps unsurprising, given the high PSA and molecular weight of
the compounds, Caco-2 data suggested low permeability and high
whereas thiazole 9k, pyrazine 9l and oxadiazole 9m showed a
moderate increase (pIC50 >8) with d activity within 10 fold. In
general, selectivity for the aminoethyl heterocycles was 100 fold
over PI3K
a but moderate over PI3Kb. The most potent example
was tetrazole 9n which also showed good cell based activity;

K. Ellard et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4546–4549
4549
Table 3
All compounds showed a reduction in neutrophil count, with
statistical significance demonstrated for 9n and 11d which showed
a 37% and 34% reduction in neutrophil accumulation resepectively
(P <0.05), thus providing validation for these chemotypes in a
respiratory inflammation model.
In conclusion, a series of urea triazolopyrimidines were synthe-
sized with excellent c/d potency and high selectivity over the other
isoforms and the general kinome. Significant efficacy for two com-
pounds from different sub-series was observed in a preliminary,
inhaled, asthma model. The physicochemical properties of these
molecules preclude them from oral absorption but lend themselves
to an inhaled method of delivery and therefore respiratory diseases
which represent a considerable area of unmet medical need.
Physicochemical data and in vitro PK for selected ureas
a
a
Mwt
PSA
clogP
Clint
Clint
Caco-2b
Kin
solc
(human)
(rat)
9d
9n
11a
11d
503
526
474
517
142
176
161
160
2.3
2.3
1.5
1.6
1.4
0.5
0
0.9
0.9
0.4
0.8
n.d.
65
0.54/109
0.08/252
n.d.
<6.5
>100
>100
0
a
b
c
Microsomal clearance (mL/min/g liver).
Papp, A–B (Â10À6 cm sÀ1)/efflux ratio.
Kinetic solubility precipitation range (
l
M), 1% DMSO.
efflux.18 When 11d was evaluated in rat pharmacokinetics no
measurable exposure was obtained upon oral administration, and
rapid elimination (>200 mL/min/kg) was observed on intravenous
dosing. Plasma samples were analyzed for parent 4 or possible
hydrolysis product, acid 10 but no trace of either, potentially active
metabolite, was detected. As this was not in-line with the in vitro
microsomal stability, the in vivo clearance is most likely attributed
to a transporter mediated clearance mechanism. Of all of the
urea sub-series, the glycinamides exhibited the highest kinetic
solubility.
We postulated that the physico-chemical properties of the
ureas made them suitable for inhaled delivery as this would enable
the drug to be delivered directly to the airways, thus bypassing the
GI tract. Low absorption and high clearance is also desirable in this
delivery method to minimize systemic exposure and thus potential
side effects.19 Efficacy of PI3K inhibitors had also been previously
demonstrated in murine respiratory models.11
A moleculefrom each subserieswas progressed to an LPS induced
pulmonary neutrophilia model as a preliminary in vivo surrogate for
asthma.20 The LPS challenge rapidly produces an inflammatory
response, dominated by neutrophils. Compounds were assessed by
their ability to reduce the accumulation of these immune cells.
9d, 9n, 11d and dexamethasone (0.5 mg) were dosed intrat-
racheally, 1 hour prior to LPS challenge to BalbC, non-fasted mice
(eight mice per group). Eight hours later the bronchoalveolar lung
fluid (BALF) was collected and neutrophils were counted (Fig. 4).
Acknowledgments
The authors gratefully acknowledge Thilo Werner and the
biochemistry group for profiling; Warren Miller for computational
chemistry and graphics preparation; Paloma Diaz for purification
and analysis; Pneumolabs for performing the in vivo experiments;
and Tammy Ladduwahetty for invaluable advice on manuscript
preparation.
References and notes
1. Wymann, M. P.; Pirola, L. Biochim. Biophys. Acta 1998, 1436, 127.
2. Thomas, M.; Owen, C. Curr. Opin. Pharmacol. 2008, 8, 267.
3. Marone, R.; Cmiljanovic, V.; Giese, B.; Wymann, M. P. Biochim. Biophys. Acta
2008, 1784, 159.
4. Ward, S.; Sotosios, Y.; Dowden, J.; Bruce, I.; Finan, P. Chem. Biol. 2003, 10, 207.
5. Vanhaesbroeck, B.; Ali, K.; Bilancio, A.; Geering, B.; Foukas, L. C. Trends Biochem.
Sci. 2005, 30, 194.
6. Ameriks, M. K.; Venable, J. D. Curr. Top. Med. Chem. 2009, 9, 738.
7. Palanki, M. S. S.; Dneprovskaia, E.; Doukas, J.; Fine, R. M.; Hood, J.; Kang, X.;
Lohse, D.; Martin, M.; Noronha, G.; Soll, R. M.; Wrasidlo, W.; Yee, S.; Zhu, H. J.
Med. Chem. 2007, 50, 4279.
8. Williams, O.; Houseman, B. T.; Kunkel, E. J.; Aizenstein, B.; Hoffman, R.; Knight,
A.; Shokat, K. M. Chem. Biol. 2010, 17, 123.
9. Subramaniam, P. S.; Whye, D. W.; Efimenko, E.; Chen, J.; Tosello, V.; De
Keersmaecker, K.; Kashisian, A.; Thompson, M. A.; Castillo, M.; Cordon-Cardo,
C.; Dave, U. P.; Ferrando, A.; Lannutti, B. J.; Diacovo, T. G. Cancer Cell 2012, 21,
459.
10. Perry, B.; Beevers, R.; Bennett, G.; Buckley, G.; Crabbe, T.; Gowers, L.; James, L.;
Jenkins, K.; Lock, C.; Sabin, V.; Wright, S. Bioorg. Med. Chem. Lett. 2008, 18, 5299.
11. Doukas, J.; Eide, L.; Stabbins, K.; Racanelli-Layton, A.; Dellamry, L.; Martin, M.;
Dneprovskaia, E.; Noronha, G.; Soll, R.; Wrasidlo, W.; Acevedo, L. M.; Cheresh,
D. A. J. Pharmacol. Exp. Ther. 2009, 328, 758.
12. Bergamini, G.; Bell, K.; Shimamura, S.; Werner, T.; Cansfield, A., Müller, K.,
Perrin, J.; Rau, C.; Ellard, K.; Hopf, C.; Doce, C.; Leggate, D.; Mangano, R.;
Mathieson, T.; O’Mahony, A.; Plavec, I.; Rharbaoui, F.; Reinhard, F.; Savitski, M.
M.; Ramsden, N.; Hirsch, E.; Drewes, G.; Rausch, O.; Bantscheff, M.; Neubauer,
G. Nat. Chem. Biol. doi: http://dx.doi.org/10.1038/nchembio.957.
13. PDB code: 4AOF.
14. Knight, Z. A.; Chiang, G. G.; Alaimo, P. J.; Kenski, D. M.; Ho, C. B.; Coan, K.;
Abraham, R. T.; Shokat, K. M. Bioorg. Med. Chem. 2004, 12, 4749.
15. For experimental procedures see: Ramsden, N.; Bell, K.; Taylor, J.; Sunose, M.;
Middlesmiss, D.; Ellard, K. PCT Int. Appl. WO2010092015, 2010.
16. Rommel, C.; Camps, M.; Ji, H. Nat. Rev. Immunol. 2007, 7, 191.
17. Assessed by a <20% inhibition of binding to Kinobeads at 1 lM, for materials
and method see: Bantscheff, M.; Eberhard, D.; Abraham, Y.; Bastuck, S.;
Boesche, M.; Hobson, S.; Mathieson, T.; Perrin, J.; Raida, M.; Rau, C.; Reader, V.;
Sweetman, G.; Bauer, A.; Bouwmeester, T.; Hopf, C.; Kruse, U.; Neubauer, G.;
Ramsden, N.; Rick, J.; Kuster, B.; Drewes, G. Nat. Biotechnol. 2007, 25, 1035.
18. Veber, D. F.; Johnson, S. R.; Cheng, H.-Y.; Smith, B. R.; Ward, K. W.; Kopple, K. D.
J. Med. Chem. 2002, 45, 2615.
19. Millan, D. S.; Bunnage, M. E.; Burrows, J. L.; Butcher, K. J.; Dodd, P. G.; Evans, T.
J.; Fairman, D. A.; Hughes, S. J.; Kilty, I. C.; Lemaitre, A.; Lewthwaite, R. A.;
Mahnke, A.; Mathias, J. P.; Philip, J.; Smith, R. T.; Stefaniak, M. H.; Yeadon, M.;
Phillips, C. J. Med. Chem. 2011, 54, 7797.
20. The LPS induced pulmonary neutrophilia model was carried out by
Pneumolabs (UK) Ltd, London.
Figure 4. Bar graphs representing the total number of neutrophils in BALF of LPS
exposed mice pre-treated with test compounds or dexamethasone. Each column
represents the mean of eight animals. Changes were compared to the vehicle
control animals using ANOVA followed by Dunnets test: ⁄P <0.05 and ⁄⁄P <0.01.