Discovery of 9-(1-phenoxyethyl)-2-morpholino-4-oxo-pyrido[1,2-a]pyrimidine- 7-carboxamides as oral PI3Kβ inhibitors, useful as antiplatelet agents

Article abstract of DOI:10.1016/j.bmcl.2014.07.007

Optimization of AZD6482 (2), the first antiplatelet PI3Kβ inhibitor evaluated in man, focused on improving the pharmacokinetic profile to a level compatible with once daily oral dosing as well as achieving adequate selectivity towards PI3Kα to minimize the risk for insulin resistance. Structure-based design and optimization of DMPK properties resulted in (R)-16, a novel, orally bioavailable PI3Kβ inhibitor with potent in vivo anti-thrombotic effect with excellent separation to bleeding risk and insulin resistance.

Full text of DOI:10.1016/j.bmcl.2014.07.007

Bioorganic & Medicinal Chemistry Letters 24 (2014) 3936–3943
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of 9-(1-phenoxyethyl)-2-morpholino-4-oxo-pyrido
[1,2-a]pyrimidine-7-carboxamides as oral PI3Kb inhibitors, useful
as antiplatelet agents
Fabrizio Giordanetto a, , , Bernard Barlaam b, Susanne Berglund a, Karl Edman c, Olle Karlsson a,
⇑
Jan Lindberg a, Sven Nylander d, Tord Inghardt a,
⇑
a Medicinal Chemistry, AstraZeneca R&D, CVMD iMed, Mölndal, Pepparedsleden 1, SE-431 83 Mölndal, Sweden
b Medicinal Chemistry, AstraZeneca R&D, Oncology iMed, Alderley Park, Macclesfield, Cheshire SK10 4TG, United Kingdom
c Discovery Sciences, AstraZeneca R&D, Pepparedsleden 1, SE-431 83 Mölndal, Sweden
d Bioscience, AstraZeneca R&D, CVMD iMed, Pepparedsleden 1, SE-431 83 Mölndal, Sweden
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 9 July 2014
Optimization of AZD6482 (2), the first antiplatelet PI3Kb inhibitor evaluated in man, focused on improving
the pharmacokinetic profile to a level compatible with once daily oral dosing as well as achieving adequate
selectivity towards PI3Ka to minimize the risk for insulin resistance. Structure-based design and optimi-
Keywords:
Phosphoinositide 3-kinase
p110b
zation of DMPK properties resulted in (R)-16, a novel, orally bioavailable PI3Kb inhibitor with potent
in vivo anti-thrombotic effect with excellent separation to bleeding risk and insulin resistance.
Ó 2014 Elsevier Ltd. All rights reserved.
PI3Kb
Inhibitor
Thromboembolism
Anti-thrombotic agent
Antiplatelet
Insulin resistance
Phosphoinositide 3-kinases (PI3K) are lipid kinases that regulate
signal transduction pathways critical for cell regulation and func-
tion. Three different classes of PI3K’s (Class I, II and III) have been
reported to date. Class I PI3K’s phosphorylate the 30-hydroxy posi-
tion of the inositol ring of phosphatidylinositol-4,5-biphosphate
(PIP2) to the corresponding phosphatidylinositol-3,4,5-triphosphate
(PIP3).1 Class I PI3K’s have been further classified into Class IA,
Additionally, PI3Kb is a key regulator of platelet activation via
19
multiple pathways, including P2Y12
,
glycoprotein VI20,21 and
integrin
a
IIbb3 outside-in signaling.22,23 Importantly, treatment
with the selective p110b inhibitor TGX-221 (1, Fig. 1) resulted in
a significant reduction of arterial thrombosis.24 This is especially
noteworthy as no increase in bleeding time was observed with
TGX-221, in contrast to aspirin and clopidogrel.24
comprising the p110
a
, b and d catalytic subunits, and Class IB
AZD6482 (2, Fig. 1) is an ATP-competitive PI3Kb inhibitor orig-
inally discovered by Shaun Jackson and colleagues as a racemic
mixture named KN-309.20 We recently reported the first human
target validation of PI3Kb inhibition with AZD6482.25 Following a
3 h infusion of seven different doses of AZD6482, a wide separation
between anti-thrombotic effect and bleeding was observed,
demonstrating that previous pharmacodynamic findings in dog
translated well to man.25
Whereas AZD6482 (2) was well tolerated in man, a weak but
significant concentration-dependent increase in plasma insulin
and corresponding homeostasis model analysis (HOMA) index
was recorded. In the plasma concentration range tested (Css up to
(p110 catalytic subunit). PI3Ka was demonstrated to be involved
c
in cell proliferation and cancer by genetic analysis,2–6 while also
playing a key role in regulating insulin-mediated signaling and
glucose metabolism.7,8
PI3Kb has been implicated in the development of phosphatase
and TENsin homolog-(PTEN) deficient tumours9,10 and several
chemical classes of PI3Kb inhibitors have recently been reported
by different groups as having anti-proliferative properties11–18
.
⇑
Corresponding authors. Tel.: +49 (0)231 9742 7224; fax: +49 (0)231 9742 7219
(F.G.); tel.: +46 31 7761954; fax: +46 31 7763724 (T.I.).
E-mail addresses: fgiordanetto@tarosdiscovery.com (F. Giordanetto), tord.inghardt
@astrazeneca.com (T. Inghardt).
5.7
lM), such an effect could be ascribed to the compound’s ability
to inhibit PI3K
a
(IC50 = 0.87 lM) although other mechanisms
cannot be ruled out.7,8,25 AZD6482 had a short plasma half-life
(5–43 min) due to high metabolic clearance and a relatively small
Present address: Taros Chemicals GmbH & Co. KG, Emil-Figge-Str. 76a, 44227
Dortmund, Germany.
http://dx.doi.org/10.1016/j.bmcl.2014.07.007
0960-894X/Ó 2014 Elsevier Ltd. All rights reserved.

F. Giordanetto et al. / Bioorg. Med. Chem. Lett. 24 (2014) 3936–3943
3937
O
O
N
to the backbone of the hinge region are critical for PI3K activity.7,29
The pyrimidinone core is tightly packed against the side chain of
N
N
Y867 and I879 in the so-called ‘affinity pocket’ of PI3K
c, a key
N
N
N
potency determinant for PI3K inhibitors. The anthranilic side chain
O
O
of 2 induces the so-called PI3Kc ‘specificity’ pocket that is absent in
NH
O
NH
the apo enzyme.29 mainly by shifting the side chain of M804. The
ability of compounds to induce this conformational switch was
shown to be related to their degree of selectivity across different iso-
OH
1. TGX-221
2. AZD6482
forms.7,30 According to the present PI3K
c structure, the carboxylic
Figure 1. Selected PI3Kb inhibitors as chemical starting points.
acid of 2 is not involved in any specific interaction with the enzyme,
although the side chains of K802 and K890 are relatively proximal
(<3.6 Å). Lastly, the methyl group at position 7 of the pyridopyrimid-
inone core is lined up with the side chains of K833 and D964, as
shown in Figure 2. Interestingly, these two amino acids are
conserved in PI3Kb, and position 7 of the scaffold could represent
a suitable substitution vector towards the PI3K channel normally
occupied by the triphosphate group of ATP.29
distribution volume (40–68 L) 25. Thus, the pharmacokinetic and
pharmacodynamic profile of AZD6482 limits its use to parenteral
administration in situations where a low bleeding risk is desirable,
such as during acute stroke, during percutaneous coronary inter-
vention or as bridging therapy when oral antiplatelet treatment
needs to be interrupted for surgery. Furthermore, while the effect
of AZD6482 on insulin homeostasis would not be clinically rele-
vant during short-term infusion, it could potentially cause insulin
resistance during chronic administration.
Based on these premises, a number of programs with the aim of
developing oral PI3Kb inhibitors for the treatment of various dis-
eases were initiated in our laboratories. With a view to address
arterial thrombosis, these included a fragment-based approach,
as previously reported,26 as well as a follow-up optimization of
AZD6482, as discussed in the present study. This latter work also
informed medicinal chemistry efforts to discover PI3Kb inhibitors
for the treatment of PTEN-deficient tumours, as detailed in the
accompanying Letter.27
Based on the PI3Kc-2 complex, we focused on two main areas
for modification: (a) the anthranilic acid side chain and (b) the
methyl group at position 7 of the pyridopyrimidinone core. Consid-
ering the short half-life of 2 in humans, metabolic clearance repre-
sented a key optimization parameter. When incubated with human
hepatocytes, 2 was rapidly metabolized (Clint = 17.5 l
L/min/106
cells, Table 1) by UDP-glucuronosyl transferases to the correspond-
ing acyl glucuronide at its anthranilic acid side chain. An initial
attempt to modify the acidic functionality was thus made in order
to verify the effects on metabolic turnover, as well as to evaluate
the resulting PI3Kb SAR. Accordingly, compounds 3–8 were
31
synthesized based on Scheme 1. Starting from alcohol 25
compounds 3, 4, 6 and 8 were made through bromination and
subsequent displacement with the corresponding aniline in one
pot. Compounds 5 and 7 were made from 2 through coupling with
hydroxylamine and methylsulfonamide, respectively.
Starting from AZD6482, the overall strategy consisted of two
main objectives: (a) to improve the pharmacokinetic profile to a
level that would enable once daily oral dosing in an eventual chronic
therapy setting and (b) to improve selectivity over the PI3K
a
Replacement of the anthranilic acid group of 2 with a 3-amino-
benzoic acid regioisomer (3) drastically improved metabolic stabil-
ity, although this was counterbalanced by a significant potency loss
against PI3Kb, as shown in Table 1. Likewise, the sulfonic acid ana-
log 6 was very stable in human hepatocytes, although its minimal
cellular permeability (Papp <0.05) might confound such result.
Alteration of the pKa of the acid by installment of an electron
withdrawing chlorine atom (4), as well as typical carboxylic acid
bioisosters 5, 7 and 8 did not offer any advantage in terms of intrin-
sic clearance, despite maintaining favorable PI3Kb potencies
isoform as a way to minimize the risk of interference with insulin
signaling. As an aid to the generation of structure activity relation-
ship (SAR) hypotheses, the crystal structure of PI3K
with 2 was solved at a final resolution of 2.9 Å.28 Analogously to
the crystal structure of PI3K
in complex with LY294002,29 the mor-
pholino ring oxygen atom of 2 is within hydrogen bonding distance
of the amide nitrogen of V882 in the hinge region of PI3K as shown
c in complex
c
c
in Figure 2. All residues highlighted in the discussion are identical in
the PI3Kb isoform except for K890 which is replaced by D856 in
PI3Kb. Previous studies have demonstrated that the interactions
Figure 2. Crystal structure of 2 (capped sticks, green carbon atoms) bound to PI3K
according to PI3K
c
(capped sticks, white carbon atoms). Selected side chains (residue identity and numbering
c
amino acid sequence) are displayed. Putative hydrogen bonds are shown as solid yellow lines.

3938
F. Giordanetto et al. / Bioorg. Med. Chem. Lett. 24 (2014) 3936–3943
Table 1
Modification and bioisosteric replacement of carboxylic acid in AZD6482 (2)
O
N
N
N
O
NH
R1
Compounda
R1
PI3Kb
PI3K
IC50
a
(
Papp
hHEP CLint
L/min/106 cells)
LogD7.4
c
d
e
b
b
IC50
(
lM)
lM)
(10ꢁ6 cm/s)
(l
*
*
*
1
2
0.049
5.8
66
20
3.6
O
0.010
4.8
0.87
18
17.5
0.75
OH
3
4
12.1
0.9
<4
0.3
1.2
OH
O
O
O
*
OH
0.031
1.46
15
11.9
Cl
*
*
*
OH
5
6
7
0.084
0.030
0.012
2.88
0.33
0.16
52
30.2
<4
1.5
0.3
0.9
N
H
O
O
S
0.05
0.11
OH
O
O
O
S
N
H
15
N
N
*
N
8
0.051
1.41
0.08
18.1
0.8
N
H
a
b
c
All compounds except 2, 5 and 7 are racemic.
Results are mean of at least two experiments. Experimental errors within 20% of value.32
Permeability measured in Caco-2 cells in the A to B direction, pH 6.5.28
Intrinsic clearance of test compounds after incubation with human hepatocytes.28
Experimental LogD at pH 7.4 using HPLC.33
d
e
O
O
N
N
N
NH
N
O
N
N
i
O
OH
R
25
3, 4, 6, 8
O
N
O
N
N
N
N
N
O
O
O
NH
O
NH
2
ii
iii
NHSO2Me
NHOH
5
7
Scheme 1. Synthesis of compounds 3–8. Reagents and conditions: (i) PBr3 (1 equiv), CH2Cl2, reflux, 7 h, then ArNH2 (2.4 equiv), NEt3 (4.5 equiv), reflux, 20 h (27–66%) (ii)
H2N–OHꢀHCl, EtOCOCl, NEt3, 2 h (15%) (iii) H2NSO2Me, EDC, DMAP, 3 d (99%).

F. Giordanetto et al. / Bioorg. Med. Chem. Lett. 24 (2014) 3936–3943
3939
O
O
MeO2C
MeO2C
HO2C
N
N
N
NH2
N
N
N
N
i-ii
iii
Br
26
O
Br
O
Br
27
28
iv
O
N
O
O
O
N
HO2C
R
HO2C
N
O
N
N
N
N
N
N
N
vi
v
O
O
O
NH
NH
29
10-15
9
Scheme 2. Synthesis of compounds 9–15. Reagents and conditions: (i) malonic acid bis(2,4,6-trichlorophenyl) ester (1.2 equiv), toluene, 90 °C (93%); (ii) MsCl (1.1 equiv),
NEt3 (1.2 equiv), THF, 10–20 °C then morpholine (3 equiv), THF, 50 °C, (82%); (iii) NaOH (1.5 equiv), water, 70 °C, 1 h, (92%); (iv) 4-(vinyloxy)butan-1-ol (5 equiv), 1,3-
bis(diphenylphosphino)propane (0.1 equiv), Pd(OAc)2 (0.025 equiv), K2CO3 (2.5 equiv), DMF–water (9:1), 80 °C, then 28, 135 °C, 3 h, then HCl(aq.) (quant); (v) PhNH2
(5 equiv), AcOH, polystyrene supported trimethylammonium cyanoborohydride, DMF–water (3:1), rt, 18 h, (48%); (vi) RNHMe (2 equiv), NEtiPr2 (5 equiv), HATU (1.5 equiv),
CH2Cl2, rt.
(IC50 = 0.012–0.084
chemical variation of the acid group of 2 did not indicate any obvi-
ous avenue for further exploration.
Additionally, direct comparison with 1, indicated that the acid
group could be removed without a substantial deterioration in
potency, selectivity and metabolic stability (Table 1).
l
M), as summarized in Table 1. Preliminary
to form reactive metabolites after oxidative metabolism and would
require additional assessment of its genotoxicity potential. The
perceived toxicological risks associated with the development of
aniline-based compounds, especially when considering the high
demands for patient safety in cardiovascular indications, prompted
us to evaluate structural alternatives. The nitrogen-to-oxygen
isosteric replacement was thought to preserve most of the confor-
mational and physicochemical properties of 11, whilst removing
the aniline group. A set of phenol-containing derivatives was thus
synthesized according to Scheme 3. Selective Luche reduction of
the methyl ketone 29 followed by amide coupling with TBDMS-
protected 2-(methylamino)ethanol gave alcohol 31. Mitsunobu
reaction with ADDP and the corresponding phenol followed by
deprotection of the silylated alcohol gave compounds 16–24.
As shown in Table 3, replacing the anilinic nitrogen atom of 11
with an oxygen (16) decreased PI3Kb potency but increased selec-
The relatively high lipophilicity of 1 (LogD7.4 = 3.6) could be
partly held responsible for the high turnover in hepatocytes
(Clint = 20 l
L/min/106 cells, Table 1) and could represent a general
threat to further development.34 Based on this hypothesis, we set
out to introduce polar groups at position 7 of the pyridopyrimidi-
none scaffold. This strategy also offered the opportunity to evaluate
additional molecular interaction with PI3Kb and PI3Ka, as predicted
from the structural information available (Fig. 2), and their effects on
potency and selectivity. Compounds 9–15 were thus synthesized
according to Scheme 2. Starting from aminopyridine 26, the
pyridopyrimidone core was obtained through cyclization with
bis-trichlorophenyl malonate. Mesylation of the resulting
3-hydroxypyridopyrimidone, followed by displacement with
morpholine and hydrolysis of the ester afforded acid 28. This was
subjected to Heck coupling with 4-(vinyloxy)butan-1-ol which
afforded methyl ketone 29 after hydrolysis. Reductive amination
with aniline gave 9 which after amide couplings gave compounds
10–15.
tivity towards PI3K
a
(38ꢂ vs 53ꢂ, respectively). At this point in the
program, we had observed a lack of linear correlation between the
PI3Kb enzyme assay and functional activity (total or unbound) in
the human platelet rich plasma aggregation assay (PRP). Knowing
that PRP activity was a good predictor of in vivo efficacy in dog
and man,25 we introduced this assay earlier in the screening cas-
cade and payed more attention to the ‘safety margin’ between
PRP and PI3Ka. As a reference point, compound 2 (AZD6482) has
Oxidation of the methyl group of 1 to the corresponding carbox-
a PRP IC50 = 0.28
l
M and PI3K IC50 = 0.87 M, that is, a ratio of
a
l
ylic acid derivative 9 was tolerated by PI3Kb (IC50: 0.059
significantly improved metabolic stability (Clint: <4
L/min/106
cells, Table 2). Nevertheless, selectivity towards PI3K was
reduced by a factor of 10 and, as expected, cellular permeability
was limited, as shown in Table 2. Neutralization of the permanent
negative charge of 9 (in a physiological pH range of 6.5–7.4) with
the corresponding N,N-dimethylamide analog 10 restored PI3Kb/
l
M) and
3.1 whereas compound 16 gave a significantly improved ratio of
25 (Table 3). Further mono- and di-substitution of the phenol side
chain (17–24) had varying effects on potency and selectivity but
always afforded weaker platelet inhibition than the unsubstituted
compound (Table 1). As an example, the meta-methoxy analog 22
displayed the best PI3Kb potency and selectivity (PI3Kb
l
a
IC50 = 0.047 /PI3Kb = 957ꢂ) but its mediocre platelet
l
M, PI3K
a
PI3Ka
selectivity and passive diffusion to adequate levels (Table 2).
inhibition (IC50 = 1.9 lM) and high metabolic instability (data not
Among the different tertiary amides evaluated (11–15) in this ser-
ies, the N-(2-hydroxymethyl)-N-methyl-amide derivative 11
offered the best compromise of potency, selectivity, permeability
and stability, especially when considering its moderate lipophilic-
ity (cf. 10 and 11, 11 and 12, Table 2).
shown) prevented any further elaboration.
Based on the available results, separation of 16 into its constitu-
ent enantiomers was carried out for further profiling.35 PI3Kb, PI3K
a
and PRP potency resided mainly in the (R) enantiomer (e.g. PI3Kb
IC50 = 0.1 vs 0.8 M). As summarized in Table 4, (R)-16 offered a
l
While 11 afforded the right balance of in vitro properties for
further characterization, concerns existed over the presence of an
unsubstituted aniline moiety in its structure. This had the potential
favorable overall pharmacological and pharmacokinetic profile
when compared to compound 2. It potently inhibited platelet acti-
vation in plasma and whole blood, was highly soluble, metabolically

3940
F. Giordanetto et al. / Bioorg. Med. Chem. Lett. 24 (2014) 3936–3943
Table 2
Modification of position 7 in TGX-221 (1)
O
R1
N
N
N
O
NH
Compounda
R1
PI3Kb
PI3Kb
Papp
hHEP CLint
L/min/106 cells)
LogD7.4
c
d
e
(10ꢁ6 cm/s)
(l
b
b
IC50
(
lM)
IC50
(
lM)
1
9
0.049
5.8
65
19.8
<4
3.6
0.6
*
O
0.059
0.068
0.61
2.5
0.04
35.4
HO
*
O
10
11
4.3
<4
2.7
2.0
N
*
O
O
HO
0.058
0.14
2.2
6.9
4.3
N
N
*
*
O
12
13
26.9
<4
<4
2.7
0.6
N
*
O
HO
0.34
3.8
1.0
O
O
N
14
15
0.035
0.076
2.3
4.8
0.99
0.05
<4
<4
1.7
1.5
N
*
H
HN
O
N
N
H
*
a
b
c
All compounds are racemic.
Results are mean of at least two experiments. Experimental errors within 20% of value.32
Permeability measured in Caco-2 cells in the A to B direction, pH 6.5.28
Intrinsic clearance of test compounds after incubation with human hepatocytes.28
Experimental LogD at pH 7.4 using HPLC.33
d
e
O
O
N
HO2C
HO2C
N
O
N
N
N
N
O
ii
O
i
OH
30
29
O
N
O
N
HO
O
O
N
O
TBDMSO
N
N
N
O
N
N
iii-iv
O
OH
31
R
16-24
Scheme 3. Synthesis of compounds 16–24. Reagents and conditions: (i) CeCl3ꢀ7H2O (1.05 equiv), MeOH–CH2Cl2 (2:1), then NaBH4 (1.3 equiv), 15 °C, 15 min, (84%); (ii)
TBDMSOCH2CH2NHMe, TSTU (1.1 equiv), NEtiPr2 (3 equiv), CH2Cl2, (59%); (iii) ArOH, ADDP (1.2 equiv), PPh3 (1.2 equiv), CH2Cl2, reflux, 16 h; (iv) TFA, CH2Cl2 (29–57% over 2
steps).
stable, cellular permeable and did not generate any detectable reac-
tive metabolites (Table 4). In vivo pharmacokinetics in different
species (exemplified by the pharmacologically relevant dog species
in Table 4) reflected well the observed in vitro profile. (R)-16 was
indeed orally bioavailable (F = 31%) with a limited volume of
distribution (1.1 L/kg) and moderate clearance (8 mL/min/kg).
Furthermore, no significant inhibition of Cyp450 enzymes and ion
channels involved in cardiac function39 was recorded at the highest

F. Giordanetto et al. / Bioorg. Med. Chem. Lett. 24 (2014) 3936–3943
3941
Table 3
Phenol analogs of 11
O
O
N
HO
N
N
X
N
O
R1
Compounda
R1
X
PI3Kb
PI3K
IC50
a
(
PRP
IC50
LogD7.4
d
b
b
c
IC50
(
lM)
lM)
(lM)
11
16
17
18
19
20
21
22
23
24
H
H
o-Me
m-Me
o-F
m-CN
o-OMe
m-OMe
m-Me, p-F
o-Me, m-F
NH
O
O
O
O
O
O
O
O
0.058
0.19
0.11
0.18
0.66
1.7
0.4
0.047
0.2
2.2
10.1
1.3
9.4
3.4
8.2
2.5
4.5
2.4
0.6
0.14
0.4
0.7
5
2.1
1.5
9.6
1.9
2.5
0.7
2.0
2.3
3
2.8
2.3
2
1.9
2.3
2.9
3.3
O
0.046
a
All compounds are racemic.
b
Results are mean of at least two experiments. Experimental errors within 20% of value.32
Human collagen-induced platelet rich plasma (PRP) aggregation.36
Experimental LogD at pH 7.4 using HPLC.33
c
d
Table 4
In vitro/vivo profile of (R)-16 compared to 2 (AZD6482)
(R)-16
2
a
PI3Kb/
a
/
c
/d IC50
(
l
M)
0.1/3.5/83/0.4
210
1.2
0.01/0.87/1.09/0.08
b
Hu collagen-induced PRP aggregation IC50 (nM)
Hu ADP-induced whole blood platelet aggregation IC50
Dog ADP-induced whole blood platelet aggregation IC50
280
0.27
1.4
100
40
c
(
l
M)
M)
c
(l
0.95
Solubilityd
(
lM)
>1796
>173
11.1
No (0.0)
26/42
1/2
8
1.2
31
e
hHEP T1/2 (min)
Papp (10ꢁ6 cm/s)
18
f
RMg
No (0.0)
8/10
0.5/nd
17
14
nd
h
Human/dog PPB Fu (%)
Dog dose iv/poi (
lmol/kg)
Dog CLi (mL/min/kg)
Dog Vss (L/kg)
Dog Fi (%)
a
b
c
d
e
f
Results are mean of at least two experiments. Experimental errors within 20% of value.32
N = 6. SD = 0.03.36
N >3, SD <0.2.37
DMSO/HBSS solubility measured at pH 7.4.33
Intrinsic clearance of test compounds after incubation with human hepatocytes.28
Permeability measured in Caco-2 cells in the A to B direction, pH 6.5.28
Ratio of reactive metabolite (RM) formation measured from 300 incubation with human liver microsomes (1 mg/mL) in the presence of glutathione.38
% fraction unbound in plasma measured by equilibrium dialysis (18 h at 37 °C).
g
h
i
Pharmacokinetic parameters calculated from noncompartmental analysis concentrations in fasted Beagle male dogs (iv/po N = 5:4).
tested concentrations (20 and 33
ing panel including 55 different protein kinases did not reveal any
significant inhibition when tested at 1 M compound concentra-
tion. No significant binding was observed against more than 100 dif-
l
M, respectively). A large screen-
reductions, CFRs), bleeding time, blood loss, ex vivo platelet aggre-
gation (Multiplate) and drug exposure. As shown in Figure 3, (R)-16
elicited a concentration-dependent inhibition of platelet aggrega-
l
tion ex vivo (EC80 = 0.69 0.06
lM), which well predicted its inhibi-
ferent enzyme, receptors and ion channels at 30
concentration.
(R)-16 achieved the two original optimization parameters:—it
afforded a pharmacokinetic profile compatible with once daily oral
l
M compound
tion of thrombosis in vivo (EC80: 0.6 0.05 l
M). Importantly, no
significant increase in bleeding time and blood loss (here defined
as a fold increase greater than 3.5, a cut off used previously in this
model for P2Y12 antagonists41) was recorded at the observed max-
dosing and an improved safety margin towards PI3K
a
. It was thus
imum compound concentrations (24.1 2.3
tigate if the improved margin between PRP and PI3K
l
M). In order to inves-
translated
characterized in vivo to evaluate its anti-thrombotic potential and
any potential risks associated with bleeding and insulin resistance.
A modified Folts’ model25,40 was used to evaluate efficacy versus
bleeding and anaesthetized dogs received vehicle (saline) followed
by consecutive doses of (R)-16 intravenously over 30 min periods
a
into less impact on insulin signaling relative to AZD6482, (R)-16
was evaluated in terms of relative homeostasis model analysis
(HOMA) insulin resistance index (the product of plasma glucose
and plasma insulin) in SD rats, as summarized in Figure 4. Gratify-
ingly, no significant increase of the HOMA-index from baseline was
apparent at the maximum compound concentration sampled
(bolus 0.03–1.3
lg/kg and infusion 0.005–0.24 lg/kg/min). The
main parameters measured were blood flow (cyclic flow

3942
F. Giordanetto et al. / Bioorg. Med. Chem. Lett. 24 (2014) 3936–3943
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2014.07.007.
References and notes
1. Vanhaesebroeck, B.; Waterfield, M. D. Exp. Cell Res. 1999, 253, 239.
2. Samuels, Y.; Wang, Z.; Bardelli, A.; Silliman, N.; Ptak, J.; Szabo, S.; Yan, H.;
Gazdar, A.; Powell, S. M.; Riggins, G. J.; Willson, J. K.; Markowitz, S.; Kinzler, K.
W.; Vogelstein, B.; Velculescu, V. E. Science 2004, 304, 554.
3. Levine, D. A.; Bogomolniy, F.; Yee, C. J.; Lash, A.; Barakat, R. R.; Borgen, P. I.;
Boyd, J. Clin. Cancer Res. 2005, 11, 2875.
4. Whyte, D. B.; Holbeck, S. L. Biochem. Biophys. Res. Commun. 2006, 340, 469.
5. Shayesteh, L.; Lu, Y.; Kuo, W. L.; Baldocchi, R.; Godfrey, T.; Collins, C.; Pinkel, D.;
Powell, B.; Mills, G. B.; Gray, J. W. Nat. Genet. 1999, 21, 99.
6. Campbell, I. G.; Russell, S. E.; Choong, D. Y.; Montgomery, K. G.; Ciavarella, M.
L.; Hooi, C. S.; Cristiano, B. E.; Pearson, R. B.; Phillips, W. A. Cancer Res. 2004, 64,
7678.
7. Knight, Z. A.; Gonzalez, B.; Feldman, M. E.; Zunder, E. R.; Goldenberg, D. D.;
Williams, O.; Loewith, R.; Stokoe, D.; Balla, A.; Toth, B.; Balla, T.; Weiss, W. A.;
Williams, R. L.; Shokat, K. M. Cell 2006, 125, 733.
8. Chaussade, C.; Rewcastle, G. W.; Kendall, J. D.; Denny, W. A.; Cho, K.; Grønning,
L. M.; Chong, M. L.; Anagnostou, S. H.; Jackson, S. P.; Daniele, N.; Shepherd, P. R.
Biochem. J. 2007, 404, 449.
Figure 3. PKPD relationship between the plasma concentration of (R)-16 (lM),
ex vivo inhibition of whole blood ADP-induced platelet aggregation (red squares
and fit line), in vivo inhibition of thrombus formation (pink circles and fit line), fold
increase in blood loss (light blue squares) and bleeding time (dark blue squares),
following administration to Beagle dogs. Data represents means SEM (N = 4).25,41
9. Jia, S.; Liu, Z.; Zhang, S.; Liu, P.; Zhang, L.; Hyun, S.; Zhang, L. J.; Signoretti, S.;
Loda, M.; Roberts, T. M.; Zhao, J. J. Nature 2008, 45, 776.
10. Wee, S.; Wiederschain, D.; Maira, S. M.; Loo, A.; Miller, C.; deBeaumont, R.;
Stegmeier, F.; Yao, Y. M.; Lengauer, C. Proc. Natl. Acad. Sci. U.S.A. 2008, 105,
13057.
11. Certal, V.; Halley, F.; Virone-Oddos, A.; Filoche-Romme, B.; Carry, J.-C.; Gruss-
Leleu, F.; Bertin, L.; Guizani, H.; Pilorge, F.; Richepin, P., et al. Bioorg. Med. Chem.
Lett. 2014, 24, 1506.
12. Certal, V.; Carry, J.-C.; Halley, F.; Virone-Oddos, A.; Thompson, F.; Filoche-
Romme, B.; El-Ahmad, Y.; Karlsson, A.; Charrier, V.; Delorme, C., et al. J. Med.
Chem. 2014, 57, 903.
13. Yu, H.; Moore, M. L.; Erhard, K.; Hardwicke, M. A.; Lin, H.; Luengo, J. I.; McSurdy-
Freed, J.; Plant, R.; Qu, J.; Raha, K., et al. ACS Med. Chem. Lett. 2013, 4, 230.
14. Certal, V.; Halley, F.; Virone-Oddos, A.; Delorme, C.; Karlsson, A.; Rak, A.;
Thompson, F.; Filoche-Rommé, B.; El-Ahmad, Y.; Carry, J. C.; Abecassis, P. Y.;
Lejeune, P.; Vincent, L.; Bonnevaux, H.; Nicolas, J. P.; Bertrand, T.; Marquette, J.
P.; Michot, N.; Benard, T.; Below, P.; Vade, I.; Chatreaux, F.; Lebourg, G.; Pilorge,
F.; Angouillant-Boniface, O.; Louboutin, A.; Lengauer, C.; Schio, L. J. Med. Chem.
2012, 55, 4788.
15. Sanchez, R. M.; Erhard, K.; Hardwicke, M. A.; Lin, H.; McSurdy-Freed, J.; Plant,
R.; Raha, K.; Rominger, C. M.; Schaber, M. D.; Spengler, M. D.; Moore, M. L.; Yu,
H.; Luengo, J. I.; Tedesco, R.; Rivero, R. A. Bioorg. Med. Chem. Lett. 2012, 22, 3198.
16. Lin, H.; Erhard, K.; Hardwicke, M. A.; Luengo, J. I.; Mack, J. F.; McSurdy-Freed, J.;
Plant, R.; Raha, K.; Rominger, C. M.; Sanchez, R. M.; Schaber, M. D.; Schulz, M. J.;
Spengler, M. D.; Tedesco, R.; Xie, R.; Zeng, J. J.; Rivero, R. A. Bioorg. Med. Chem.
Lett. 2012, 22, 2230.
17. Lin, H.; Schulz, M. J.; Xie, R.; Zeng, J.; Luengo, J. I.; Squire, M. D.; Tedesco, R.; Qu,
J.; Erhard, K.; Mack, J. F.; Raha, K.; Plant, R.; Rominger, C. M.; Ariazi, J. L.; Sherk,
C. S.; Schaber, M. D.; McSurdy-Freed, J.; Spengler, M. D.; Davis, C. B.;
Hardwicke, M. A.; Rivero, R. A. ACS Med. Chem. Lett. 2012, 3, 524.
18. Kim, J.; Hong, S.; Hong, S. Bioorg. Med. Chem. Lett. 2011, 21, 6977.
19. Garcia, A.; Kim, S.; Bhavaraju, K.; Schoenwaelder, S.; Kunapuli, S. Biochem J.
2010, 429, 369.
Figure 4. PKPD relationship between compound plasma concentrations (lM) and
the relative homeostasis model analysis (HOMA) insulin resistance index (the
product of plasma glucose and plasma insulin) in SD rats. Data represents
means SEM (N P5).25 AZD6482 (2) is depicted as red squares, (R)-16 as black
squares.
A relative HOMA index of 3 (dashed line) represents a statistically
significant change from baseline.
20. Kim, S.; Mangin, P.; Dangelmaier, C.; Lillian, R.; Jackson, S. P.; Daniel, J. L.;
Kunapuli, S. P. J. Biol. Chem. 2009, 284, 33763.
21. Gilio, K.; Munnix, I. C.; Mangin, P.; Cosemans, J. M.; Feijge, M. A.; van der
Meijden, P. E.; Olieslagers, S.; Chrzanowska-Wodnicka, M. B.; Lillian, R.;
Schoenwaelder, S.; Koyasu, S.; Sage, S. O.; Jackson, S. P.; Heemskerk, J. W. J. Biol.
Chem. 2009, 284, 33750.
22. Schoenwaelder, S. M.; Ono, A.; Sturgeon, S.; Chan, S. M.; Mangin, P.; Maxwell,
M. J.; Turnbull, S.; Mulchandani, M.; Anderson, K.; Kauffenstein, G.; Rewcastle,
G. W.; Kendall, J.; Gachet, C.; Salem, H. H.; Jackson, S. P. J. Biol. Chem. 2007, 282,
28648.
23. Canobbio, I.; Stefanini, L.; Cipolla, L.; Ciraolo, E.; Gruppi, C.; Balduini, C.; Hirsch,
E.; Torti, M. Blood 2009, 114, 2193.
24. Jackson, S. P.; Schoenwaelder, S. M.; Goncalves, I.; Nesbitt, W. S.; Yap, C. L.;
Wright, C. E.; Kenche, V.; Anderson, K. E.; Dopheide, S. M.; Yuan, Y.; Sturgeon, S.
A.; Prabaharan, H.; Thompson, P. E.; Smith, G. E.; Shepherd, P. R.; Daniele, N.;
Kulkarni, S.; Abbott, B.; Saylik, D.; Jones, C.; Lu, L.; Giuliano, S.; Hughan, S. C.;
Angus, J. A.; Robertson, A. D.; Salem, H. H. Nat. Med. 2005, 11, 507.
25. Nylander, S.; Kull, B.; Björkman, J. A.; Ulvinge, J.-C.; Oakes, N.; Emanuelsson, B.
M.; Andersson, M.; Skärby, T.; Inghardt, T.; Fjellström, O.; Gustafsson, D. J.
Thromb. Haemost. 2012, 10, 2127.
26. (a) Giordanetto, F.; Kull, B.; Dellsén, A. Bioorg. Med. Chem. Lett. 2011, 21, 829;
(b) Giordanetto, F.; Wållberg, A.; Cassel, J.; Ghosal, S.; Kossenjans, M.; Yuan, Z.-
Q.; Wang, X.; Lub, F.; Zhoub, X.; Dianqing, L.; Liang, L. Bioorg. Med. Chem. Lett.
2012, 22, 6665; (c) Giordanetto, F.; Wållberg, A.; Ghosal, S.; Iliefski, T.; Cassel,
J.; Yuan, Z. Q.; von Wachenfeldt, H.; Andersen, S. M.; Inghardt, T.; Tunek, A.;
Nylander, S. Bioorg. Med. Chem. Lett. 2012, 22, 6671.
(20.4 2.1 lM). A significant effect on insulin resistance at higher
compound concentrations cannot be excluded. Nevertheless, the
safety margin to compound concentrations resulting in full anti-
thrombotic effect appears to be acceptable for (R)-16.
In summary, optimization of AZD6482, aimed at improving its
PK profile and increase the margin to PI3Ka inhibition, resulted
in the design, synthesis and characterization of (R)-16, a novel
PI3Kb inhibitor well suited for oral administration. Based on its
potent in vivo anti-thrombotic effect and minimized risk of bleed-
ing and insulin resistance, (R)-16 was selected as a preclinical
candidate for further development.
Acknowledgements
We thank the following AstraZeneca R&D colleagues: Jan-Arne
Björkman and Helen Zachrisson for Folts dog experiments; Elham
Nikookhesal and Göran Wahlund for rat HOMA index experiments;
Anna-Karin Asztély and Jan-Christer Ulvinge for in vitro platelet
function experiments; Anders Tunek for DMPK input; Bryan Egner
and Joakim Gradén for synthesis contributions.

F. Giordanetto et al. / Bioorg. Med. Chem. Lett. 24 (2014) 3936–3943
3943
27. Barlaam, B. et al Bioorg. Med. Chem. Lett. 2014, 24, 3928.
28. For experimental details, see Supporting information.
29. Walker, E. H.; Pacold, M. E.; Perisic, O.; Stephens, L.; Hawkins, P. T.; Wymann,
M. P.; Williams, R. L. Mol. Cell. 2000, 6, 909.
30. Berndt, A.; Miller, S.; Williams, O.; Le, D. D.; Houseman, B. T.; Pacold, J. I.;
Gorrec, F.; Hon, W. C.; Liu, Y.; Rommel, C.; Gaillard, P.; Ruckle, T.; Schwarz, M.
K.; Shokat, K. M.; Shaw, J. P.; Williams, R. L. Nat. Chem. Biol. 2010, 6, 117.
31. Fjellström, O; Gustafsson, D.; Jackson, S; Lindberg, J. PCT Int. Appl.
WO2009093972 (2009).
36. Blood was collected by venipuncture from healthy donors into vacutainer
tubes containing sodium citrate (final concentration in blood 11 mM).
Collagen-induced platelet rich plasma (PRP) light transmission aggregometry
(LTA) was evaluated using an in house developed 96 well plate assay. For
experimental details, see Ref. 25.
37. Blood was collected by venipuncture from healthy donors or beagle dogs into
vacutainer tubes containing sodium citrate (final concentration in blood
11 mM). Whole blood impedance aggregometry was evaluated using the
Multiplate (Dynabyte, Munich, Germany) impedance aggregometer. For
additional details, see Ref. 25.
32. PI3Kabcd enzyme isoform selectivity was evaluated in vitro using human
recombinant enzyme and an AlphaScreen assay to measure PIP3 production, as
described in Bailey, J. P.; Giles, M. B.; Pass, M. PCT Int. Appl. WO2006005914,
(2006).
38. Ratio obtained by dividing the total integrated peak area of glutathione
adducts of the test compound by the integrated peak area of the major
glutathione adduct of the control compound clozapine (average value, N = 3).
RM formation is ranked as follows: ratio >1 = ‘high’; ratio between 0.25–
1 = ‘medium’; ratio <0.25 = ‘low’.
33. Wan, H.; Holmén, A. G. Comb. Chem. High Throughput Screening 2009, 12, 315.
34. Waring, M. J. Expert Opin. Drug Discovery 2010, 5, 235.
35. The (R) configuration was assigned to the more potent enantiomer, in analogy
39. Patch clamp assay using IONWORKS™ technology in CHO cells expressing the
relevant human ion channel (hERG, Nav1.5, Nav1.2, Kv4.3, IKs, Cav3.2, Cav2.2).
For Cav1.2: Patch clamp assay using IONWORKS™ technology in CHO cells
expressing CACNA1C, CACNB2 and CACNA2D1 (ChanTest, Cleveland, OH), as in
Dilbaghi, S.; Abi-Gerges, N.; Morton, M. J.; Bridgland-Taylor, M. H.; Pollard, C.
E.; Valentin, J.-P. J. Pharmacol. Toxicol. Methods 2010, 62, 1.
40. Folts, J. Circulation 1991, 83, 3.
41. van Giezen, J. J.; Berntsson, P.; Zachrisson, H.; Björkman, J. A. Thromb. Res. 2009,
124, 565.
with the crystal structure of 2 bound in PI3K
chiral preparative HPLC of racemic 16 on
50 ꢂ 250 mm, particle size 20 m, using EtOH/Et3N 100:0.1 as eluent, flow
c
. (R)-16 was obtained through
a
Chiralpak AD column,
l
20
rate 120 mL/min at 40 °C. Detection PDA 280 nm.
[
a
]
ꢁ174.8° (c = 1,
D
CH3CN), 1H NMR (400 MHz, DMSO-d6) d 1.62 (d, 3H), 2.89 (br s, 3H), 3.1–
3.25 (m, 1H), 3.34–3.70 (m, 11H), 4.68–4.83 (m, 1H), 5.68 (s, 1H), 5.95 (q, 1H),
6.87 (m, 3H), 7.21 (m, 2H), 7.74 (br s, 1H), 8.79 (d, 1H). (ESI) m/z calcd for
C
24H28N4O5 [M+H]+ 453.2138; found 453.2120.