Discovery of 1-(4-(5-(5-amino-6-(5-tert-butyl-1,3,4-oxadiazol-2-yl)pyrazin-2-yl)-1-ethyl-1,2,4-triazol-3-yl)piperidin-1-yl)-3-hydroxypropan-1-one (AZD8835): A potent and selective inhibitor of PI3Kα and PI3Kδ for the treatment of cancers

Article abstract of DOI:10.1016/j.bmcl.2015.10.002

Starting from potent inhibitors of PI3Kα having poor general kinase selectivity (e.g., 1 and 2), optimisation of this series led to the identification of 25, a potent inhibitor of PI3Kα (wild type, E545K and H1047R mutations) and PI3Kδ, selective versus PI3Kβ and PI3Kγ, with excellent general kinase selectivity. Compound 25 displayed low metabolic turnover and suitable physical properties for oral administration. In vivo, compound 25 showed pharmacodynamic modulation of AKT phosphorylation and near complete inhibition of tumour growth (93% tumour growth inhibition) in a murine H1047R PI3Kα mutated SKOV-3 xenograft tumour model after chronic oral administration at 25 mg/kg b.i.d. Compound 25, also known as AZD8835, is currently in phase I clinical trials.

Full text of DOI:10.1016/j.bmcl.2015.10.002

Accepted Manuscript
Discovery of 1-(4-(5-(5-Amino-6-(5-tert-butyl-1,3,4-oxadiazol-2-yl)pyra-
zin-2-yl)-1-ethyl-1,2,4-triazol-3-yl)piperidin-1-yl)-3-hydroxypropan-1-one
(AZD8835): A Potent and Selective Inhibitor of PI3Kα and PI3Kδ for the
Treatment of Cancers
Bernard Barlaam, Sabina Cosulich, Bénédicte Delouvrié, Rebecca Ellston,
Martina Fitzek, Hervé Germain, Stephen Green, Urs Hancox, Craig S. Harris,
Kevin Hudson, Christine Lambert-van der Brempt, Honorine Lebraud,
Françoise Magnien, Maryannick Lamorlette, Antoine Le Griffon, Rémy
Morgentin, Gilles Ouvry, Ken Page, Georges Pasquet, Urszula Polanska, Linette
Ruston, Twana Saleh, Michel Vautier, Lara Ward
PII:
S0960-894X(15)30111-6
DOI:
http://dx.doi.org/10.1016/j.bmcl.2015.10.002
BMCL 23158
Reference:
To appear in:
Bioorganic & Medicinal Chemistry Letters
Received Date:
Revised Date:
Accepted Date:
11 September 2015
30 September 2015
1 October 2015
Please cite this article as: Barlaam, B., Cosulich, S., Delouvrié, B., Ellston, R., Fitzek, M., Germain, H., Green, S.,
Hancox, U., Harris, C.S., Hudson, K., Brempt, C.L-v., Lebraud, H., Magnien, F., Lamorlette, M., Griffon, A.L.,
Morgentin, R., Ouvry, G., Page, K., Pasquet, G., Polanska, U., Ruston, L., Saleh, T., Vautier, M., Ward, L., Discovery
of
1-(4-(5-(5-Amino-6-(5-tert-butyl-1,3,4-oxadiazol-2-yl)pyrazin-2-yl)-1-ethyl-1,2,4-triazol-3-yl)piperidin-1-
yl)-3-hydroxypropan-1-one (AZD8835): A Potent and Selective Inhibitor of PI3Kα and PI3Kδ for the Treatment
of Cancers, Bioorganic & Medicinal Chemistry Letters (2015), doi: http://dx.doi.org/10.1016/j.bmcl.2015.10.002
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Bioorganic & Medicinal Chemistry Letters
Discovery of 1-(4-(5-(5-Amino-6-(5-tert-butyl-1,3,4-oxadiazol-2-yl)pyrazin-2-yl)-1-
ethyl-1,2,4-triazol-3-yl)piperidin-1-yl)-3-hydroxypropan-1-one (AZD8835): A Potent
and Selective Inhibitor of PI3K and PI3Kfor the Treatment of Cancers
b
b
a
b
b
Bernard Barlaam*, Sabina Cosulich, Bénédicte Delouvrié, Rebecca Ellston, Martina Fitzek, Hervé
a
b
b
a,1
b
Germain, Stephen Green, Urs Hancox, Craig S. Harris, Kevin Hudson, Christine Lambert-van der
Brempt, Honorine Lebraud, Françoise Magnien, Maryannick Lamorlette, Antoine Le Griffon, Rémy
Morgentin, Gilles Ouvry, Ken Page, Georges Pasquet, Urszula Polanska, Linette Ruston, Twana
Saleh, Michel Vautier, Lara Ward.
a
a
a
a
a
a,2
a,1
b
a
b
b
a
a
b
a
AstraZeneca, Centre de Recherches, Z.I. La Pompelle, B.P. 1050, Chemin de Vrilly, 51689 Reims, Cedex 2, France
AstraZeneca, Oncology iMed, Alderley Park, Macclesfield, Cheshire SK10 4TG, United Kingdom
b
corresponding author: B.Barlaam e-mail: bernard.barlaam2@astrazeneca.com
1
current address: Galderma R&D, les Templiers, 2400 route des Colles, 06410 Sophia-Antipolis, France
current address: Edelris, 115 avenue Lacassagne, 69003 Lyon, France
2
ARTICLE INFO
ABSTRACT
Article history:
Received
Revised
Accepted
Available online
Starting from potent inhibitors of PI3K having poor general kinase selectivity (e.g. 1 and 2), optimisation of this
series led to the identification of 25, a potent inhibitor of PI3K (wild type, E545K and H1047R mutations) and
PI3K, selective vs. PI3K and PI3K, with excellent general kinase selectivity. Compound 25 displayed low
metabolic turnover and suitable physical properties for oral administration. In vivo, compound 25 showed
pharmacodynamic modulation of AKT phosphorylation and near complete inhibition of tumour growth (93% tumour
growth inhibition) in a murine H1047R PI3K mutated SKOV-3 xenograft tumour model after chronic oral
administration at 25 mg/kg b.i.d. Compound 25, also known as AZD8835, is currently in phase I clinical trials.
Keywords:
PI3K
PI3K
2
015 Elsevier Ltd. All rights reserved.
Kinase
PIK3CA mutation
The phosphoinositide 3-kinases (PI3Ks) are a group of lipid kinases which phosphorylate the 3-hydroxyl group of phosphatidylinositol
1,2
(4,5)-bisphosphate (PIP2) to phosphatidylinositol (3,4,5)-triphosphate (PIP3), while the phosphatase PTEN reverses this process.
PIP3 accumulation at the plasma membrane results in recruitment of AKT which is phosphorylated and activated, leading to a cascade
of cell signalling which controls a range of cellular processes like cell proliferation, growth and survival. In mammals, there are eight
PI3K family members, divided into three classes (I, II and III) based on structure, regulation and lipid substrate specificity. The PI3K
Class I contains four isoforms, PI3K and PI3K being ubiquitously expressed, whereas the expression of PI3K and PI3K is mainly
restricted to the immune system. All four isoforms are heterodimers comprising a p110α, p110β and p110δ (Class IA) or p110γ (Class
IB) catalytic subunit (for PI3K, PI3K, PI3K and PI3K respectively) together with a regulatory subunit. Class IA and the Class IB
3-5
are respectively activated by receptor tyrosine kinases and G-protein coupled receptors.
The PI3K/AKT/m-TOR pathway is very commonly activated in a wide range of human cancers as a result of hyperactivation of
6
upstream growth factor signalling, mutations, or loss of tumour suppressor genes such as PTEN. Amplification and mutation of the
7,8
PIK3CA gene encoding the p110α catalytic unit occurs frequently in common human cancers, including ca. 25% of breast cancers.
The three most common mutations (E542K, E545K and H1047R) in the PIK3CA gene have been confirmed as activating mutations,
9
whilst activating mutations have not been reported for p110,and have only very recently been reported for p110. The strong
evidence pointing at the oncogenic nature of the PIK3CA gene mutations and their high frequency have fuelled a strong interest in
1
0,11
developing selective inhibitors of PI3K.
The first PI3K inhibitors in clinical trials have been pan-PI3K inhibitors, inhibiting multiple PI3K isoforms.
12,13
The main concern
associated with the pan-PI3K inhibitors has been their tolerability profile due to the inhibition of different PI3K enzymes. Moreover,
some of the early pan-PI3K clinical candidates also have off-targets activity, especially on other members of the PI3K-related kinase
1
4
(
PIKK) family. On the other hand, isoform-selective PI3K inhibitors may offer potential strategies for maximising the therapeutic
window with added potential for combination, while eliminating the more deleterious effects which may arise from pan-isoform
15
inhibition.

At the start of this work, very few isoform-selective PI3K inhibitors exhibiting appropriate drug-like properties had been described.
Idelalisib (also known as GS-1101 and CAL-101; launched in 2014) was the first isoform selective compound entering clinical trial for
1
6,17
the treatment of cancers and, based on its selective PI3K pharmacology, was developed in hematological malignancies.
More
18,19
isoform selective agents are in the clinical phase, including PI3K isoform selective inhibitor NVP-BYL719,
PI3K-sparing
20
21
22
inhibitor GDC-0032, PI3K isoform selective inhibitors GSK263771 and SAR260301, and finally dual PI3Kinhibitor
23
AZD8186 (Figure 1).
N
N
N
N
CF3
N
O
S
N
N
O
N
H
N
H N
2
O
O
N
H N
2
N
NVP-BYL719
O
GDC-0032
O
F
O
N
N
O
N
N
O
NH
HN
N
N
N
F
F
NH
AZD8186
Idelalisib
Figure 1. Structure of selected isoform selective PI3K inhibitors.
We previously reported the identification of a promising novel series of PI3K inhibitors selective for PI3K versus PI3Kand having
excellent general protein kinase selectivity (exemplified by compound 3), by designing out undesired protein kinase activity associated
24
with our initial aminopyrazine lead while keeping the selectivity for PI3K versus PI3K. An essential aspect of that work was to
improve overall general kinase selectivity from the notoriously promiscuous aminopyrazine core. During our hit identification
activities, we had also identified a series of 3-oxadiazolyl-2-aminopyridines, exemplified by 1, as potent inhibitors of PI3K both at the
enzyme and the cellular level (Figure 2). Further exploration of the substitution on the oxadiazole identified the tert-butyl analogue 2
showing increased potency for PI3K. However, as anticipated from the similarity of this series to our initial aminopyrazine lead,
aminopyridines 1 and 2 displayed poor general kinase selectivity. Again, the identity of some of those anti-targets and their known
pharmacological effects might lead to an unacceptable safety profile.
R
N
N
N
O
N
N
N
N
N
N
N
N
N
N
N
N
O
NH2
N
NH2
1
2
(R= H)
(R= Me)
3
Figure 2. Structure of compounds 1-3.
In this Letter, we describe the optimisation work from 1 and 2, building potency against PI3K and selectivity versus PI3K and more
generally versus the anti-targets such as KDR, Aurora B, FGFR1 and GSK3 as surrogates of general protein kinase selectivity.
Selectivity across the PI3K isoforms was routinely measured with biochemical enzyme assays and the  selectivity was also
monitored in cell assays (inhibition of AKT phosphorylation at Thr308 in PIK3CA mutant human breast ductal carcinoma BT474 cells,
sensitive to PI3Kinhibition, and at Ser473 in PTEN-null breast adenocarcinoma MDA-MB-468 cells, sensitive to PI3Kinhibition;
23
providing measurements of PI3K and PI3Kinhibition respectively. Kinase promiscuity was routinely measured using several
kinase enzyme assays: KDR, Aurora B, FGFR1, GSK3, but for simplicity, only the KDR and/or FGFR1 enzyme results are reported in
this paper. Broader kinase selectivity was confirmed for prioritized compound(s) using additional panel screens.
Compound 1-2, 4-25 were made according to Schemes 1-3:
25
Suzuki coupling of commercially available 1a with known boronic ester 26 followed by functional group manipulation afforded ester
b, which was converted into oxadiazole 1 in three steps. Alternatively, starting from 2a, the oxadiazole group was first assembled in
the synthesis, followed by bromination of the aminopyridine, to give 2b. Coupling of 2b with 26, followed by further functional group
1
24
manipulation gave 2. Compound 4 was made from previously described 4a. Synthesis of 5 is based on the key stannane 5a, which
2
4
24
was prepared from 2b in two steps. Stille coupling of 5a with dibromotriazole 27, followed by Suzuki coupling with boronic ester 28
and reduction of the double bond, provided the Boc-protected desired regioisomeric triazole, which was then readily converted to 5
Scheme 1).
(

N
N
O
N
Br
CO Me
N
CO Me c-d
2
N
N
2
N
a-b
N
N
NH2
O
B
N
NH2
N
NH2
1a
Boc
N
O
1b
1
N
N
26
N
O
O
CO Me
N
N
N
2
Br
i-j
N
e-h
N
N
N
NH2
N
NH2
O
B
N
NH2
2
a
2b
Boc
N
N
O
2
N
26
N
N
N
N
N
N
AcN
N
N
N
Boc
N
N
N
N
k
N
N
N
NH2
N
NH2
4a
4
N
O
N
O
N
N
l-m
Me Sn
3
n-r
AcN
N
N
N
2b
Br
N
NBoc2
N
NH2
O
O
N
N
5a
Boc
N
B
N
5
Br
27
28
Scheme 1. Synthesis of compounds 1, 2, 4 and 5. Reagents and conditions: (a) 26, Pd(PPh
3
)
4
, Cs
2
CO
3
, dioxane-water (4:1), 80 °C, 2 h, 70%; (b) 4N HCl,
i
i
i
dioxane, rt, 4 h; CH
(1 eq.), MeSO
30 °C, 4 h, 76%; (h) NBS, THF, rt, 2 h, 88%; (i) 26, (PPh
MeOH (1:1), 25 °C, 10 min, 43%; (k) TFA-CH Cl (1:1), rt, 1 h, 84%; AcCl, NEt Pr
Sn Me , Pd(PPh , toluene, 110 °C, 2 h, 66%; (n) 27, (PPh PdCl , LiCl, DME, 130-140 °C (µwave), 58%; (o) 28, (PPh
0 min, 68%; (p) Pd, HCO NH , DMF, 70 °C, 3 h, 59%; (q) TFA, CH
2
O, NEt Pr
H, 70 °C, 18 h, 4%; (e) NH
2
, NaBH(OAc)
3
, MeOH, rt, 21 h, 55%; (c) NH
NH . H O, EtOH, 90 °C, 15 h, 96%; (f) BuCO
PdCl , CsF, MeOH, 120 °C (µwave), 30 min, then TFA, rt, 67%; (j) CH
2
NH
2
. H
2
O, EtOH, 80 °C, 5 h, 94%; (d) HATU, PrCO
2
H, NEt Pr
2
, NMP, rt, 48 h;
t
i
i
P
2
O
5
3
2
2
2
2
H, EDCI, HOPO, DMF, rt, 15 h, 73%; (g) Pr-N=C=N- Pr, DMA,
1
3
)
2
2
2
O, NaBH(OAc)
, rt, 2 h, 85%; (l) DMAP, NEt Pr O, DMF, rt, 60 h, 75%; (m)
2
PdCl , CsF, MeOH, 140 °C (µwave),
3 2 2
, CH Cl -
i
i
2
2
2
, CH
2
Cl
2
2 2
, Boc
2
6
3
)
4
3
)
2
2
3
)
2
7
2
4
2
Cl
2
, rt, 1 h, 90%; (r) AcCl, NEt
3
, CH
2
Cl
2
, 0 °C, 1 h, 72%.
O
N
N
O
N
N
Br
N
N
CO Me
Br
N
N
Me Sn
N
2
3
a-b
c
NH2
NH2
N
NH2
29
30
31
N
N
Boc
N
N
Br
d
32
2
R
2
R
N
1
R
N
O
N
N
N
O
N
N
f
N
N
N
X
N
N
N
g
O
N
N
N
NH2
NH2
6
-21; 25
3
3
3a (R = Me, X= Boc)
2
2
3b (R = Et, X= Boc) from 36 only
e
k
2
3
3
9a (R = Me, X= H)
2
9b (R = Et, X=H)
O
N
NH2
O
N
H
N
N
N
N
N
H N
N
N
O
2
h
N
N
N
i-j
Boc
N
N
N
N
N
N
NH2
NH2
NH2
34
35
36

t
Scheme 2. Synthesis of compounds 6-21 and 25. Reagents and conditions: (a) NH
2
NH
2
. H
2
O, EtOH, 50 °C, 40 h, quant.; (b) ( BuCO)
PdCl , LiCl, 4-Me-2-pentanol, 140
, MeCN, rt, 1 h, 97% (for 10) and quant. (for
, (for 6-7); or AcOCH COCl, NEt , CH Cl , then aq. LiOH,
, TBTU, DMF or NMP (for 11-14, 16-21); (g) Pd (dba) (0.025 eq),
O, PrOH, 50 °C, 16
h, 79% ; (i) N-BOC isonipecotic acid, HATU, N-Me morpholine, DMA, rt, 95%; (j) AcOH, dioxane, 70-100 °C, 18 h, 77%; (k) DBU, MeI, 2-methylTHF, rt,
2
O, MeCN, 80 °C, 1 h,
i
9
°
2
6%; NEt Pr
C, 2 h, 79% (for 33a, R = Me); (e) TFA - CH
5), then PPTS, MeOH, 60 °C, 90 min, 78% (for 10) and 85% (for 25); or AcCl, NEt Pr
2
, TsCl (1.2 eq.), MeCN, 70 °C, 2 h, quant.; (c) Sn
2
Me
6
, (PPh
3
)
2
PdCl
2
, toluene, 80 °C, 2 h, 39% ; (d) 32, (PPh
3
)
2
2
2
i
2
Cl
2
(1:1), rt, quantitative; (f) THP-OCHi
2
CH
, CH
i
2
CO
2
H, HATU, NEt Pr
Cl
2
2
2
2
2
3
2
2
3
3
MeOH (for 8); or R CO
2
H, EDCI, HOPO, CH
2
Cl
2
or NMP (for 9, 15); or R CO
2
H, NEt Pr
2 2 2
(0.6 eq), DMA, 100 °C, 1 h, 95% ; (h) NH NH . H
2
2
3
i
dicyclohexyl(2’,4’,6’-triisopropylbiphenyl-2-yl)phosphine (0.1 eq), Zn (0.1 eq), ZnCN
2
2
2
7
6% (for 33a, R = Me) or DBU, EtI, 2-methylTHF, 35 °C, 80% (for 33b, R = Et).
Compound 6-21 and 25 were made from compounds 33a or 33b by Boc deprotection followed by coupling with the corresponding
carboxylic acid (Scheme 2). For hydroxyl containing side chains, protection with an acetyl (for 8) or with a tetrahydropyranyl (for
compounds 10 and 25) was used during the amide coupling. Two routes to compound 33a were developed, both starting from
commercially available methyl 2-amino-5-bromopyrazine-2-carboxylate 29. Ester conversion into the tert-butyl oxadiazole afforded the
key common intermediate 30. The first synthesis relied on the conversion of 30 into the stannane 31 using stoichiometric amounts of
24
hexamethyldistannane, followed by Stille coupling with previously described 32 to give 33a. Therefore, a second route better suited
for larger scale synthesis was developed, avoiding use of toxic tin derivatives or by-products. The bromo derivative 30 was converted
into the nitrile 34 using zinc dicyanide and palladium catalysis. Addition of hydrazine onto 34 gave 35. Coupling with N-Boc
isonipecotic acid, followed by cyclisation under acidic conditions gave the key triazole 36. Finally, regioselective methylation with
methyl iodide and DBU gave 33a in excellent yield. The ethyl analogue 33b was obtained from 36 by reaction with ethyl iodide under
similar conditions.
3
R
3
R
O
N
N
CO Me
N
N
N
2
Me Sn
N
e-g
O
O
a-d
3
N
N
N
N
N
N
N
NH2
N
NH2
HO
37
22-24
NH2
38a-c
3
i
Scheme 3. Synthesis of compounds 22-24. Reagents and conditions: (a) NH
2
NH
, TsCl (1 eq.), 0 °C - reflux 45-64%; (c) NBS, THF, rt, 16 h, 79-88%; (d) Sn
DME, 130 °C (µwave), 36-54%; (f) TFA, rt, 30 min, 88-100%; (g) THP-OCH CH CO H, EDCI, HOPO, CH
2
. H
2
O, EtOH, 60 °C, 2 h, 97%; (b) R CO
Me , (PPh PdCl , dioxane, 80 °C, 47-64%; (e) 32, (PPh
Cl , rt, 2 h; PPTS, MeOH, 60 °C, 3 h, 47-61%.
2
H, TBTU, NEt Pr
2
, MeCN, 80 °C, 20
i
min; NEt Pr
2
2
6
3
)
2
2
3
)
2
PdCl , LiCl,
2
2
2
2
2
2
A variation on the first synthesis of the aminopyrazines was developed for the variation of the tert-butyl group on the oxadiazole
Scheme 3). Starting from methyl 2-aminopyrazine-2-carboxylate 37, hydrazide formation followed by coupling with the corresponding
carboxylic acid and cyclisation afforded the oxadiazoles 38a-c. Bromination followed by stannylation and Stille coupling with 32 gave
(
2
6
the Boc-protected derivatives. Boc deprotection, followed by amide coupling with 3-(tetrahydropyran-2-yloxy)propanoic acid and
final deprotection with PPTS in methanol, afforded compounds 22-24.
Table 1.
Structure, PI3K, KDR and FGFR1 enzyme and PI3K cell inhibitory potencies of compounds 1-6.
PI3K
enz. IC50
0.100
PI3K
enz. IC50
5.7
KDR
enz. IC50
0.13
FGFR1
enz. IC50
0.49
PI3K
cell IC50
0.200
PI3K
cell IC50
13
Cpd
substituent
Structure
a
a
a
a
a
a
3
i
3
R
1
R = Pr
N
O
N
N
N
3
t
N
2
3
R = Bu
0.014
0.029
3.1
1
0.079
9.0
0.15
10
0.074
0.27
5.3
6.9
20
N
NH2
1
R CO =
2 2
Me NCH CO
N
N
1
R
N
N
N
N
N
N
1
O
b
b
4
5
6
R CO = Ac
A = CH
A = N
0.028
0.410
0.068
1.4
24
8.6
17
0.79
7.6
N
NH2
b
b
b
83
>100
>30
N
N
O
N
N
A
N
N
N
b
b
9.6
>100
>100
1.0
>30
O
NH2
a
Geometric means of at least two IC50 determinations per compound, M.
n=1.
b
24
We previously reported the identification of 3 as a selective PI3K inhibitor versus PI3K having excellent general kinase selectivity
Figure 2). We showed that replacement of the pyrazole group on the left hand side of the molecule (present in the initial lead) by the 1-
(
methyl-1,2,4-triazole group (present in 3) dramatically increased general kinase selectivity. However, one shortcoming identified with 3
was poor oral exposure in rat (AUC: 0.7 M.h following an oral dose of 5 mg/kg in rat, bioavailability 10%). One possible explanation
-6
is the modest permeability measured in Caco-2 (Papp, A-B 1.6 x 10 cm/s) associated with efflux potential (efflux ratio of 47 in the MDR-

1
overexpressing mdck cell line), probably linked to the overall low lipophilicity of 3 (logD7.4 0.9) in conjunction with high molecular
27
weight and the presence of a basic side chain. The neutral acetamide 4 had much higher permeability and was less prone to efflux
-6
(Caco-2 Papp, A-B 36 x 10 cm/s; efflux ratio of 5 in the MDR-1 overexpressing mdck cell line). 4 retained good general kinase
selectivity, but unfortunately showed a significant reduction of potency (see Table 1).
28
Docking of compound 2 in the publicly available crystal structure of PI3K was performed manually, based on the binding mode of
aminopyrazine and aminopyridine derivatives observed in various in-house kinase crystal structures (unpublished results). In this model
(see Figure 3), 2 makes a double donor/acceptor interaction with the hinge residues Val851 and Glu849 and the sp2 nitrogen of the
pyrazole ring is located at H-bond distance of the Gln859 residue that is unique to the PI3K isoform. This binding mode is very
24
similar to the proposed binding mode of 3, where one nitrogen of the triazole interacts with Gln859. During the course of this work,
another group reported that hydrogen bond interactions with the Gln859 side chain also was the structural determinant for PI3K
1
8
selectivity in their aminothiazole series.
2
7
Figure 3. Compound 2 docked into the publicly available crystal structure of PI3K.
Therefore, we thought we could apply the same strategy to solve the general kinase selectivity issue seen with 2. But to our surprise,
pyridine 5 lost most of its PI3K activity. Although it is difficult to fully rationalise it, we cannot exclude subtle conformational
differences between the pyridine and the pyrazine cores. Indeed, we were pleased to see that the corresponding pyrazine 6 regained
some potency against PI3Kwhilst it lost activity against the two anti-targets KDR and FGFR1. The PI3Kcell activity was still
modest, but again improvement of general kinase selectivity was confirmed in a panel of kinases (see Figure 4, compound 6 vs.
compound 2). Compound 6 displayed excellent selectivity on PI3K vs. PI3K, good physical and metabolic properties (low
-6
lipophilicity: LogD7.4 2.1; high permeability: Caco-2 Papp, A-B 39 x 10 cm/s; low rate of metabolic turnover: rat hepatocyte and human
6
microsomal Clint: 4.4 L/min/10 cell and 9 L/min/mg, respectively; low human plasma protein binding: fraction unbound f
u
32%), but
the overall PI3K potency would need to be improved.
Figure 4. Heatmap showing the inhibitory activity of compound 2 (on top) and compound 6 (below) against a panel of protein kinases (‘Dundee University
protein kinase panel’) at 1 M. The percentage of inhibition is colored from green (0%) to yellow (60%) and red (90%).

Table 2 illustrates the modifications we explored to further optimize compound 6. General kinase selectivity remained excellent for
compounds 7-24, as illustrated by the very modest KDR activity reported in Table 2 for these compounds, and will not be discussed
further. Variation of the substitution on the triazole showed a significant potency change, with ethyl 7 being more potent than methyl 6.
But metabolic stability in human microsomes was also reduced, probably linked to increased lipophilicity. We then looked at the
modification of the amide. Larger amides than acetamide 6 were tolerated, either linear such as 9 or 11, or cyclic such as 12 or 13. But
again metabolic stability was an issue; some hydroxyl containing amides, such as 8 or 10, showed acceptable metabolic stability, but
others, such as 14 or to a lesser degree 15, did not. Increasing the size of the amide (9-13) also increased PI3K potency. Overall, 10
showed the combination of potency, selectivity, metabolic stability we were looking for (see Table 2). However, one immediate concern
was the potential -elimination of the hydroxyl group, leading to the reactive acrylamide. To test this, we performed chemical stability
measurements in aqueous conditions and trapping experiments with glutathione in human liver microsomes and did not see any
evidence of formation of the acrylamide. Instead, one of the metabolites we identified in rodent, dog and human hepatocytes was the
carboxylic acid resulting from oxidation of the primary alcohol, and chemical stability measurements showed traces of ring opening of
the oxadiazole in aqueous solution at very acidic pH, i.e. pH 1. Because of the serious issues that formation of the acrylamide in humans
would cause, we nevertheless tried to block potential -elimination (compounds 16-18), but again poor metabolic stability was an issue,
although it is worth noting that the cyclopropyl 18 was the most metabolically stable of the three. Compounds 19-21, containing an
extra methyl on the amide portion compared to 10 showed good potency and/or microsome stability. But overall, the 3-
hydroxypropanamide present in 10 was considered as the best amide.
Table 2.
Structure, PI3K enzyme and cell inhibitory potencies and other properties of compounds 6-25.
3
R
2
R
N
N
O
1
R
N
N
N
N
N
N
O
NH2
1
2
3
PI3K
enz. IC50
0.068
0.008
0.045
0.031
0.023
0.034
0.025
0.022
PI3K
enz. IC50
9.6
0.95
10
KDR b
PI3K
cell IC50
1.0
0.22
1.1
0.37
0.34
0.26
0.4
PI3K
cell IC50
>30
7
HLM
Clint
d
Compound
R
R
R
a
a
a
a
c
LogD7.4
enz. IC50
t
t
t
t
t
t
t
t
6
7
8
9
Me
Me
OH
Me
Et
Bu
Bu
Bu
Bu
Bu
Bu
Bu
Bu
>100
10
47
30
20
26
20
24
9.4
25
4
25
3
35
53
46
2.1
2.5
1.4
1.8
1.5
2.1
-
CH
CH
CH
CH
2
Me
Me
Me
Me
Me
Me
27
14
13
2
OMe
1.9
2.2
3.5
2.1
1
1
1
1
0
1
2
3
2
CH
CH
2
OH
OMe
2
2
24
e
4-tetrahydropyranyl
7.8
c
Pr
2.6
0.4
15
-
5
1
t
e
e
1
4
Me
Bu
0.011
19
0.62
19
29
-
OH
t
t
t
1
1
1
5
6
7
C(Me)
CH C(Me)
C(Me) CH
2
OH
Me
Me
Me
Bu
Bu
Bu
0.037
0.046
0.032
6.3
4.8
3.4
24
18
27
0.42
0.35
0.32
17
15
16
13
21
60
-
-
-
2
2
OH
OH
2
2
2
6
t
1
8
OH
Me
Bu
0.031
2.2
0.41
15
11
-
t
t
t
1
2
2
9
0
1
(R)-CH(Me)CH
(R)-CH CH(Me)OH
CH(Me)OH
2
OH
Me
Me
Me
Bu
Bu
Bu
0.032
0.025
0.030
3.4
2.9
3.3
24
25
20
0.53
0.31
0.27
24
21
15
3.5
13
10
1.8
1.8
1.8
2
(S)-CH
2
2
2
CH
2
CH
2
OH
Me
0.024
0.102
2.5
8.1
3.3
47
0.42
1.2
16
4
1.4
F
2
3
CH
2
CH
2
OH
Me
>30
5.5
1.1
i
2
2
4
5
CH
CH
2
2
CH
CH
2
2
OH
OH
Me
Et
Pr
0.074
0.0062
7.1
0.431
29
6.4
0.77
0.057
24
3.5
4.9
8.7
1.2
1.9
t
Bu
a
b
c
d
e


M; unless stated otherwise, geometric means of at least two IC50 determinations per compound.
M; single IC50 determination or geometric means of at least two IC50 determinations per compound.
-1
-1
human liver microsome intrinsic clearance (HLM Clint); L.min .mg .
Measured using shake-flask methodology with a buffer:octanol volume ratio of 100:1.
n=1 value.
Next, we turned our attention to the oxadiazole. We considered potential replacement of the oxadiazole by other 5-membered
heteroaromatics. But the proposed binding mode of these compounds in PI3Ksuggests that both nitrogens of the oxadiazole are
necessary, one making an internal hydrogen bond with the aminopyrazine and maintaining some degree of planarity, and the second
being involved in a H-bond acceptor, presumably with Asp933 via a water molecule, in a similar fashion as the 4-pyridine of NVP-

18
t
BYL719 interacts with PI3K. We considered the substitution on the oxadiazole, but other groups were at best equivalent to the Bu
i.e. 22), with less lipophilic groups being less potent (i.e. 23 or 24).
(
Finally combining the SAR elements of the different portions (best amide, substitution on the oxadiazole and the triazole) gave the
increased potency we were looking for. Compound 25 displayed superior PI3K potency, yet with good selectivity and metabolic
stability, and was selected for further profiling.
TM
Figure 5. Kinase selectivity of compound 25 (KINOMEScan , DiscoveRx); percentage of control at 1 M against a panel of 456 kinases (picture a); against
atypical kinases, including lipid kinases and mutant kinases (picture b).
The broader kinase selectivity of 25 was evaluated in several panels of kinase assays. 25 showed very limited activity in a panel of 71
protein kinases (‘Dundee University protein kinase’ panel) when tested at 1 M, with only one kinase, GCK (germinal center kinase),
showing >70% inhibition. The kinase selectivity of 25 was confirmed in the commercial KINOMEscan (DiscoveRx) against a panel of
456 kinases, where it showed no or limited competitive binding at 1 M on the majority of the kinases (see Figure 5). The 22 kinases
out of the 456 tested, where probe binding was determined to be significant at 1 M, were further investigated. K determination on
d
those 22 kinases confirmed the high degree of kinase selectivity of 25, with only one other enzyme (ALK6, also known as BMPR1B)
outside PI3K class I enzymes being within 10 fold of PI3K wild type (Table 10 in Supplementary data). It is worth noting that 25
showed similar potency on the mutated PI3K enzymes (E542K, E545K and H1047R) when compared to wild type PI3Kenzyme.
Confirmation of selectivity among the lipid kinases was assessed in a panel of 15 kinases (‘Dundee University lipid kinase’ panel),
where an IC50 below 1 M was seen for only two kinases outside PI3K class I enzymes, i.e. PI4Kand PI4Kbut with more than 50
fold selectivity margin vs. PI3KTable 9 in Supplementary data). Bespoke kinase assays confirmed selectivity against selected
kinases from the closely related PIKK family (m-TOR IC50 12.1 M, DNAPK IC50 4.0 M). 25 was screened for inhibition of other
members of the PIKK family, ATR and ATM, in cell based assays. Very weak inhibitory activity was observed against ATM (endpoint:
phosphorylation of ATM at Ser1981 in HT29 cells, IC50 >30 M) or ATR (endpoint: phosphorylation of Chk-1 at Ser345 in HT29 cells,
IC50 >21 M), consistent with the desired target selectivity profile of 25.
Table 3
Class I PI3K kinase enzyme and cell inhibition of compound 25.
a
Assay details
Enzyme inhibition
IC50 (M)
0.0062
0.0060
0.0058
0.431
0.0057
0.090
0.057
3.5
Class I PI3K enzyme
PI3K
Mutated PI3K – E545K
Mutated PI3K – H1047R
PI3K
PI3K
PI3K
BT474
MDA-MB-468
Jeko-1
Cellular inhibition
p-AKT endpoint
PI3K
PI3K
PI3K
PI3K
0.049
0.53
RAW264
a
Geometric means of multiple IC50 determinations.
The selectivity profile of 25 among the class I PI3K isoforms was tested in enzyme and cell based assays (Table 3). At the enzyme
level, 25 is a potent mixed inhibitor of PI3K(IC50 0.0062 M) and PI3K(IC50 0.0057 M), with selectivity against PI3K(IC50
0.431 M) and PI3K(IC50 0.090 M). 25 is also a potent inhibitor of the commonly occurring PI3Kmutants, PI3K- E545K(IC50
0.0060 M) and PI3K- H1047R(IC50 0.0058 M). In cell-based assays assessing the ability to inhibit AKT phosphorylation, 25 was
a potent inhibitor in cells sensitive to PI3Kinhibition (IC50 0.057 M in PIK3CA mutant human breast ductal carcinoma BT474 cell
line) and in cells sensitive to PI3K inhibition(IC50 0.049 M in Jeko-1 B cell line, but not to cells sensitive to PI3Kinhibition (IC50

3
.5 M in PTEN null breast adenocarcinoma MDA-MB-468 cells) or to PI3Kinhibition (IC50 0.53 M in monocytic RAW264 cell
line).
Table 4
Physical, pharmacokinetic and other properties of compound 25.
Properties
Value
14 ; 24 ; 26; 17
8.6 ; < 3.4 ; < 2.0
a
Plasma protein binding Hu / Rat / Dog / Mo: f
Intrinsic Clearance in hepatocytes, rat / dog / hu: Clint (L/min/10 cell)
u
(%)
6
b
Rat / Dog Pharmacokinetic parameters
c
Clearance: Cl (mL/min/kg)
11.6 / 6.7
1.1 / 2.1
47 / 96
196
39
>30 (all)
c
Volume of distribution: V
d
(L/kg)
c
Bioavailability: F (%)
d
Aqueous solubility (M)
-
6
e
Permeability Caco-2: Papp, A-B (10 cm/s)
CYP inhibition – 1A2, 2C9, 2C19, 2D6, 3A4 IC50 (M)
hERG (% inhibition at 100 M)
f
g
37.9
a
u
Protein binding in human, rat, dog and mouse plasma, fraction unbound f .
Intrinsic clearance measured from fresh rat/dog hepatocytes and cryopreserved human hepatocytes, Clint
From plasma concentrations in male Han Wistar rats (at least n=2), compound administered at 4 mol/kg i.v. and 10 mol/kg p.o. for 25 as a formulation in
b
c
.
HPMC/Tween for the oral arm and as a solution formulation in 10% DMA , 90% Captisol (30% w/v in water) for the iv arm; from plasma concentrations in
Beagle dogs (at least n=2), compound administered orally at a dose around 2 mol/kg as a formulation in 0.5% (w/v) HPMC, 0.1% (w/v) Tween 80 in water and
i.v. at a dose around 2 mol/kg as a solution in 5% (v/v) DMSO, 35% (v/v) tetraethylene glycol, 60% (v/v) citrate buffer. Intravenous parameters calculated from
an intravenous bolus; bioavailability, from oral and i.v. AUC.
d
Solubility in aqueous phosphate buffer, pH 6.8; from crystalline material.
Permeability in Caco-2 cell line, A pH 6.5, B pH 7.4.
e
f
Inhibition of cytochrome P450, IC50
.
g
Inhibition of hERG potassium channel, % inhibition at 100 M
Compound 25 displayed good solubility, good permeability and low turnover in hepatocytes from various species. As expected from the
in vitro data, low in vivo clearance associated with high bioavailability was seen in both rat and dog (Table 4). Compound 25 showed
high exposure following oral administration to SCID mice (AUC: 137 M.h and cmax 34 M at 50 mg/kg p.o.) and was selected for
further in vivo evaluation.
In a pharmacodynamic experiment following chronic oral dosing (25 mg/kg b.i.d. or 6 mg/kg b.i.d. of 25) in nude mice bearing mutant
H1047R PI3K SKOV-3 tumour xenografts, target modulation was assessed by measuring AKT phosphorylation levels at Ser473 at 30
minutes and 8 hours. At both doses, strong inhibition of AKT phosphorylation was observed at the 30 minute timepoint. At 8 hours,
significant inhibition was still seen at the 25 mg/kg dose, whereas no inhibition was seen at the lower dose of 6 mg/kg, consistent with
the lower plasma concentrations observed (see Figure 6).
Figure 6. Inhibition of AKT phosphorylation (black bars) in the SKOV-3 xenograft model in nude mice following chronic oral administration of 25 (25 mg/kg
b.i.d. and 6 mg/kg b.i.d.) and associated free plasma levels (red squares) at 30 minutes and 8 hours.
The antitumour activity of 25 was evaluated in the same model following chronic oral dosing at the same doses of 25 mg/kg b.i.d. and 6
mg/kg b.i.d. (Figure 7). Near complete inhibition of tumour growth (93% tumour growth inhibition) was seen at 25 mg/kg b.i.d., whilst
the lower dose of 6 mg/kg b.i.d gave only partial inhibition of tumour growth (32% tumour growth inhibition), again consistent with the
reduced inhibition of AKT phosphorylation observed 8 hours after the lower dose.
In summary, starting from a series of 2-aminopyridines exemplified by 1 and 2 as potent inhibitors of PI3K both at the enzymatic and
the cellular level, but having modest general kinase selectivity, optimisation of this series led to the identification of the clinical
candidate 25, a potent inhibitor of PI3K (wild type, E545K and H1047R mutations) and PI3K, selective vs. PI3K and PI3K, and
with excellent general kinase selectivity. Compound 25 displayed low metabolic turnover and suitable physical properties for oral
administration. In vivo, compound 25 showed pharmacodynamic modulation of AKT phosphorylation in a SKOV-3 xenograft tumour
model (H1047R mutant PI3K grown in mouse, after chronic oral administration at 25 mg/kg b.i.d. at the 30 minute and 8 hour
timepoints. Compound 25 showed near complete inhibition of tumour growth (93% tumour growth inhibition) in the same model after
chronic oral administration at 25 mg/kg b.i.d. Compound 25 (also known as AZD8835) was selected as a clinical candidate for the

treatment of PIK3CA-dependent cancers and is currently in phase I clinical trials. Further in vitro and in vivo biological characterization
2
9
of 25 will be reported in due course.
Control
Compound 25 - 25mg/kg bid
Compound 25 - 6mg/kg bid
1.2
1.0
0.8
0.6
0.4
0.2
0
5
10
15
20
Days of treatment
Figure 7. Tumour growth inhibition in the SKOV-3 xenograft model in nude mice following chronic oral administration of 25 (25 mg/kg b.i.d. and 6 mg/kg
b.i.d.).
Acknowledgments
We acknowledge the excellent technical expertise of the scientists at AstraZeneca in particular: Vanessa Gonnot, Patrice Koza,
Jonathan Lecoq, Fabrice Renaud, Gordon Currie and Stuart Wells for synthesis of compounds; Paul Davey, Christian Delvare, Delphine
Dorison Duval for characterization of compounds; the members of the SAR Screening Groups for generating the cell and biochemical
data; Neil James, Dan Owen, Samantha Waite nee Johnston and Cath Trigwell for biological evaluation.
Supplementary data
Supplementary data (synthetic procedures and characterization for all compounds, biological protocols and procedures for
determination of physiochemical and DMPK properties) associated with this article can be found in the on-line version.
References and notes
1
.
Vanhaesebroeck, B.; Leevers, S. J.; Ahmadi, K.; Timms, J.; Katso, R.; Driscoll, P. C.; Woscholski, R.; Parker, P. J.; Waterfield, M. D. Annu. Rev. Biochem.
001, 70, 535.
2
2
3
4
5
6
7
.
.
.
.
.
.
Engelman, J. A.; Luo, J.; Cantley, L. C. Nat. Rev. Genet. 2006, 7, 606.
Cantley, L. C. Science 2002, 296, 1655.
Foster, F. M.; Traer, C. J.; Abraham, S. M.; Fry, M. J. J. Cell Sci. 2003, 116, 3037.
Vanhaesebroeck, B.; Guillermet-Guibert, J.; Graupera, M.; Bilanges, B. Nat. Rev. Mol. Cell. Biol. 2010, 11, 329.
Liu, P.; Cheng, H.; Roberts, T. M.; Zhao, J. J. Nat. Rev. Drug Discov. 2009, 8, 627.
Bachman, K. E.; Argani, P.; Samuels, Y.; Silliman, N.; Ptak, J.; Szabo, S.; Konishi, H.; Karakas, B.; Blair, B. G.; Lin, C.; Peters, B. A.; Velculescu, V. E.,
Park, B. H. Cancer Biol. Ther. 2004, 3, 776.
8
.
Wu, G.; Xing, M.; Mambo, E.; Huang, X.; Liu, X.; Guo, Z.; Chaterjee, A.; Goldenberg, D.; Gollin, S. M.; Sukumar, S.; Trink, B.; Sidransky, D. Breast
Cancer Res. 2005, 7, R609.
9
1
1
1
1
1
1
1
.
Choi, M. R.; Yoo, N. J.; An, C. H.; Lee, S. H. Human Pathology 2015, 46, 753.
0. Denny, W. A. Expert Opin. Ther. Pat. 2013, 7, 789.
1. Lee J. Y.; Engelman, J. A., Cantley L. C. Science 2007, 317, 206.
2. Liu, P.; Cheng, H.; Roberts, T. M.; Zhao, J. J. Nat. Rev. Drug Discov. 2009, 8, 627.
3. Courtney, K. D.; Ryan, B.; Engelman, J. A. J. Clin. Oncol. 2010, 28, 1075.
4. Fruman, D.A.; Rommel, C. Nat. Rev. Drug Discov. 2014, 13, 140.
5. Thorpe, L. M.; Yuzugullu, H.; Zhao, J. J. Nat. Rev. Cancer 2015, 15, 7.
6. Lannutti, B. J.; Meadows, S. A.; Herman, S. E. M.; Kashishian, A.; Steiner, B.; Johnson, A. J.; Byrd, J. C.; Tyner, J. W.; Loriaux, M. M.; Deininger, M.;
Druker, B. J.; Puri, K. D.; Ulrich, R. G.; Giese, N. A. Blood 2011, 117, 591.
1
1
7. Norman, P. Expert Opin. Ther. Pat. 2011, 11, 1773.
8. Furet, P.; Guagnano, V.; Fairhurst, R.A.; Imbach-Weese, P.; Bruce, I.; Knapp, M.; Fritsch, C.; Blasco, F.; Blanz, J.; Aichholz, R.; Hamon, J.; Fabbro, D.;
Caravatti, G. Bioorg. Med. Chem. Lett. 2013, 23, 3741.
1
2
2
2
9. Fritsch, C.; Huang, A.; Chatenay-Rivauday, C.; Schnell, C.; Reddy, A.; Liu, M.; Kauffmann, A.; Guthy, D.; Erdmann, D.; De Pover, A.; Furet, P.; Gao, H.;
Ferretti, S.; Wang, Y.; Trappe, J.; Bracmann, S. M.; Sauveur-Michel M.; Wilson, C.; Boehm, M.; Garcia-Etcheverria, C.; Chene, P.; Wiesmann, M.; Cozens,
R.; Lehar, J.; Schlegel, R.; Caravatti, G.; Hofmann, F.; Sellers, W. R. Mol. Cancer Ther. 2014, 13, 1117.
0. Ndubaku, C. O.; Heffron, T. P.; Staben, S. T.; Baumgardner, M.; Blaquiere, N.; Erin Bradley, E.; Bull, R.; Do, S.; Dotson, J.; Dudley, D.; Edgar, K. A.;
Friedman, L.S.; Goldsmith, R.; Heald, R.A.; Kolesnikov, A.; Lee, L.; Lewis, C.; Nannini, M.; Nonomiya, J.; Pang, J.; Price, S.; Prior, W.W.; Salphati, L.;
Sideris, S.; Wallin, J. J.; Wang, L.; Wei, B.; Sampath, D.; Olivero, A. G. J. Med. Chem. 2013, 56, 4597.
rd
1. a) Blackman, S. C.; Gainer, S. D.; Suttle, B. B.; Skordos, K.W.; Greshock, J. D.; Motwani, M.; Roadcap, L. T.; Hardwicke, M. A. A.; Wooster, R. F. 103
Ann. Meet. Am. Assoc. Cancer Res. (AACR) 2012, Abst 1752; b) A phase I/IIa, first time in human, study of GSK2636771 in subjects with advanced solid
tumors with phosphatase and tensin homolog (PTEN) deficiency (NCT01458067) ClinicalTrials.gov Web Site 2011, October 20.
2. Certal, V.; Carry, J. C.; Halley, F.; Virone-Oddos, A.; Thompson, F.; Filoche-Romme, B.; Al-Ahmad, Y.; Karlsson, A.; Charrier, V.; Delorme, C.; Rak, A.;
Abecassis, P.-Y.; Amara., C.; Vincent, L.; Bonnevaux, H.; Nicolas, J.-P.; Mathieu, M.; Bertrand, T.; Marquette, J. P.; Michot, N.; Benard, T.; Perrin, M.-A.;
Lemaitre, O.; Guerif, S.; Perron, S.; Monget, S.; Gruss-Leleu, F.; Doerflinger, G.; Guizani, H.; Brollo, M.; Delvarre, L.; Bertin, L.; Richepin, P.; Loyau, V.;
Garcia-Echeverria, C.; Lengauer, C.; Schio, L. J. Med. Chem. 2014, 57, 903.

2
2
2
2
3. Barlaam, B.; Cosulich, S.; Degorce, S.; Fitzek, M.; Green, S.; Hancox, U.; Lambert-van der Brempt, C.; Lohmann, J.-J.; Maudet, M.; Morgentin, R.;
Pasquet, M.-J.; Péru, A.; Plé, P.; Saleh, T.; Vautier, M.; Walker, M.; Ward, L.; Warin, N. J. Med. Chem. 2015, 58, 943.
4. Barlaam, B.; Cosulich, S.; Fitzek, M.; Green, S.; Harris, C. S.; Hudson, K.; Lambert-van der Brempt, C.; Ouvry, G.; Page, K.; Ruston, L.; Ward, L.;
Delouvrié, B. Bioorg. Med. Chem. Lett. 2015, 25, 2679.
5. de Koning, P. D.; McAndrew, D.; Moore, R.; Moses, I. B.; Boyles, D. C.; Kissick, K.; Stanchina, C. L.; Cuthbertson, T.; Kamatani, A.; Rahman, L.;
Rodriguez, R.; Urbina, A.; Sandoval, A.; Rose, P. R. Org. Proc. Res. Dev. 2011, 15, 1018.
6. Barlaam, B. C.; Delouvrie, B.; Ouvry, G.; Lambert-Van der Brempt, C. M. P.; Harris, C. S.; Berry, D.; Tomkinson, G. P.; Reid, G. P. U.S. Pat. Appl. (2014),
US 20140206700.
2
2
7. Gleeson, M. P. J. Med. Chem. 2008, 51, 817.
8. Mandelker, D.; Gabelli, S. B; Schmidt-Kittler, O.; Zhu, J.; Cheong, I.; Huang, C. H.; Kinzler, K. W.; Vogelstein, B.; Amzel, L. M. Proc. Natl. Acad. Sci.
USA 2009, 106, 16996. The publicly available crystal structure of PI3K in complex with wortmannin was used to dock compound 2 (PBD ID code: 3HHM;
coordinates available).
2
9. Hudson, K; Hancox, U.; Trigwell, C.; McEwen, R.; Polanska, U.; Nikolaou, M; Morentin Gutierrez, P.; Avivar-Valderas, A.; Delpuech, O.; Dudley, P.;
Hanson, L.; Ellston, R.; Jones, A.; Cumberbatch, M.; Cosulich, S.; Ward, L.; Cruzalegui, F.; Green S. Intermittent high dose scheduling of AZD8835, a
novel selective inhibitor of PI3Kα and PI3Kδ demonstrates treatment strategies for PIK3CA-dependent breast cancers. Submitted for publication.

Graphical Abstract
N
N N
N
O
O
O
N
N
N
N
N
N
N
N
N
HO
N
NH2
NH2
25
1
PI3K enz. IC50 0.0062 M
PI3K cell IC50 0.057 M
PI3K cell IC50 3.5 M
PI3K enz. IC50 0.100 M
PI3K cell IC50 0.200 M
poor kinase selectivity
PI3K cell IC50 0.53 M
PI3K cell IC50 0.049 M
excellent kinase selectivity and physical properties