Identification and optimisation of 4,5-dihydrobenzo[1,2-d:3,4-d]bisthiazole and 4,5-dihydrothiazolo[4,5-h]quinazoline series of selective phosphatidylinositol-3 kinase alpha inhibitors

Article abstract of DOI:10.1016/j.bmcl.2015.06.067

Abstract A cyclisation within a 4′,5-bisthiazole (S)-proline-amide-urea series of selective PI3Kα inhibitors led to a novel 4,5-dihydrobenzo[1,2-d:3,4-d]bisthiazole tricyclic sub-series. The synthesis and optimisation of this 4,5-dihydrobenzo[1,2-d:3,4-d]bisthiazole sub-series and the expansion to a related tricyclic 4,5-dihydrothiazolo[4,5-h]quinazoline sub-series are described. From this work analogues including 11, 12, 19 and 23 were identified as potent and selective PI3Kα inhibitor in vivo tool compounds.

Full text of DOI:10.1016/j.bmcl.2015.06.067

Bioorganic & Medicinal Chemistry Letters 25 (2015) 3575–3581
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Identification and optimisation of 4,5-dihydrobenzo
[1,2-d:3,4-d]bisthiazole and 4,5-dihydrothiazolo[4,5-h]quinazoline
series of selective phosphatidylinositol-3 kinase alpha inhibitors
⇑
Robin A. Fairhurst , Marc Gerspacher, Patricia Imbach-Weese, Robert Mah, Giorgio Caravatti,
Pascal Furet, Christine Fritsch, Christian Schnell, Joachim Blanz, Francesca Blasco, Sandrine Desrayaud,
Daniel A. Guthy, Mark Knapp, Dorothee Arz, Jasmin Wirth, Esther Roehn-Carnemolla, Van Huy Luu
Novartis Institutes for BioMedical Research, Global Discovery Chemistry, Novartis Pharma AG, Werk Klybeck, Postfach, CH-4002 Basel, Switzerland
a r t i c l e i n f o
a b s t r a c t
A cyclisation within a 40,5-bisthiazole (S)-proline-amide-urea series of selective PI3K
a inhibitors led to a
novel 4,5-dihydrobenzo[1,2-d:3,4-d]bisthiazole tricyclic sub-series. The synthesis and optimisation of
this 4,5-dihydrobenzo[1,2-d:3,4-d]bisthiazole sub-series and the expansion to a related tricyclic 4,5-di-
hydrothiazolo[4,5-h]quinazoline sub-series are described. From this work analogues including 11, 12,
Article history:
Available online 26 June 2015
Keywords:
Oncology
Kinase inhibitor
Phosphatidylinositol-3-kinase-alpha
19 and 23 were identified as potent and selective PI3K
a
inhibitor in vivo tool compounds.
Ó 2015 Elsevier Ltd. All rights reserved.
Phosphatidylinositol-3-kinase alpha (PI3K
a
) is a member of the
moiety to close a 6-membered ring onto the unsubstituted position
of the second thiazole ring yielded the general structure 2. This
cyclisation was anticipated to introduce additional favorable
hydrophobic contacts within the affinity pocket with the side-
chain of tyrosine 836, as shown in Figure 3. Additionally, the cycli-
sation also locks the inhibitor in the conformation required for
optimally interacting within the PI3K ATP-pocket. However, the
low energy conformation of the 40,5-bisthiazole biaryl-bond corre-
sponds closely to the locked conformation, and a negligible contri-
bution was anticipated from limiting the conformational mobility
about the bisthiazole linkage. At the initiation of this work, a sim-
ilar cyclisation had also been reported for a related series of
aminothiazole tricyclic PI3K inhibitors.7
To prepare the analogues 2, a synthesis was developed starting
from 2-bromo-1,3-cyclohexanedione.8 Annulation with thiourea
introduced the 2-aminothiazole moiety which was N-acylated to
give 3. Bromination of the remaining ketone functionality was fol-
lowed by formation of the second thiazole ring upon heating with a
range of thioamides, to give the analogues 4, in a reaction catalysed
by ammonium phosphomolybdate.9 Hydrolysis of the acetamide
moiety was followed by preparation of the acylimidazole interme-
diates, which were isolated and then reacted with the appropriate
proline-amide derivatives to give the targeted compounds 5–12, of
the general structure 2.
class 1 PI3K family which has been strongly implicated as a key dri-
ver in a number of cancers through amplification, overexpression
and mutation.1 In particular, PI3K
a
is one of the most commonly
mutated genes in human cancer, and these mutations typically
lead to enhanced, or constitutive, PI3K
kinase-activity.2 As a
result several groups have sought to identify selective PI3K inhi-
a
a
bitors to explore as anticancer treatments. Of these the (S)-proline-
amide aminothiazole-urea derivative alpelisib is currently one of
the most advanced with Phase I/II clinical studies currently
ongoing.3
In the preceding Letter we described the identification of a 40,5-
bisthiazole variant of an (S)-proline-amide aminothiazole-urea ser-
4
ies, as exemplified by A66 in Figure 1. This formed part of a
broader piece of work which started from the hit structures, exem-
plified by 1 in Figure 2, and culminated in the identification of
alpelisib.5,6 In this Letter we describe an additional extension of
this work in which a further annulation was introduced to generate
the tricyclic 4,5-dihydrobenzo[1,2-d:3,4-d]bisthiazole core, as
exemplified by the general structure 2.
Expanding upon the homology modeling that was carried out
with the 40,5-bisthiazole analogues, such as A66, suggested an
opportunity to increase the interaction within the ATP pocket of
PI3Ka 4
.
A cyclization from the 4-methyl of the aminothiazole
Table 1 shows the biochemical and cellular data for A66 and the
cyclised analogues 5–12.10 Comparison of A66 with the corre-
sponding cyclised analogue 5 shows the impact of the cyclisation
on class 1 PI3K activity: similar levels of biochemical activity and
⇑
Corresponding author. Tel.: +41 61 69 64294; fax: +41 61 6966246.
E-mail address: robin.fairhurst@novartis.com (R.A. Fairhurst).
http://dx.doi.org/10.1016/j.bmcl.2015.06.067
0960-894X/Ó 2015 Elsevier Ltd. All rights reserved.

3576
R. A. Fairhurst et al. / Bioorg. Med. Chem. Lett. 25 (2015) 3575–3581
activity relationships (SAR) for the affinity-pocket 2-thiazole ring
N
N
was found to be comparable to that observed for the parent series.
Small quaternary alkyl residues led to the highest levels of activity,
as exemplified by comparing the isomers 5 and 6, consistent with
optimally filling the affinity pocket.4 Similarly, the substituted pro-
line SAR was mirrored in the cyclised analogues, as exemplified by
compounds 7–11. In particular, the trans- and cis-3-methyl proline
analogues led to some of the most highly potent and selective
S
N
H
N
H
N
S
S
N
N
N
CF3
O
O
O
O
H2N
H2N
alpelisib
A66
examples prepared in the entire series, with cellular PI3Ka activi-
Figure 1. Structures of the PI3K
alpelisib and A66.
a (S)-proline-amide aminothiazole-urea derivatives
ties below 10 nM and selectivities of greater than 50-fold versus
the other class 1 PI3K’s. Building on this observation, the potential
for incorporating a tertiary amine into the proline nucleus, as a sol-
ubility enhancing functionality, was explored. One successful
example was the cis-3-dimethylaminomethyl proline analogue
12, the key building block for which was prepared via an extension
of a reported amino-zinc-enolate-cyclisation.12,8 Compound 12
N
N
O
S
H
H
N
SO2Me
N
N
R2
S
S
N
R1
resulted in a potent and selective PI3Ka inhibitor substituted with
F
N
O
O
a tertiary amine, something which had proven to be difficult to
achieve in the parent 40,5-bisthazole sub-series.4
O
Evaluating the physical properties of the 4,5-dihydrobenzo[1,2-
d:3,4-d]bisthiazole analogues 2, the increase in PI3Ka potency led
H2N
H2N
1
2
to a favorable impact upon the perceived pharmacokinetic (PK)
risk by reducing the anticipated level of exposure required to
achieve efficacy. This was helpful as the low solubility of the
40,5-bisthiazole series, which had been attributed as one of the pri-
mary reasons for the low oral bioavailabilities, was not signifi-
cantly improved upon by the cyclisation, when direct analogues
were compared between the series.4 Solubilities determined at
pH 6.8 for the most potent 3-methyl analogues 9–11 were all
Figure 2. Structures of the proline-amide-urea series starting point 1 and the
proposed cyclised 4,5-dihydrobenzo[1,2-d:3,4-d]bisthiazole analogues 2.
below 14
l
M.13 However, for the direct A66 analogue, compound
M was determined, in combi-
5, a slightly higher solubility of 65
l
nation with high permeability and also a good level of metabolic
stability.14,15 A rat PK study with 5 resulted in an encouraging
exposure profile, and a respectable oral bioavailability of 38%,
when administered as a suspension at a dose of 3.0 mg kgꢀ1. The
PK data for selected compounds are summarised in Table 3.16
The more soluble basic analogue 12 also showed a good level of
metabolic stability with a moderate level of permeability that
was associated with a modest efflux signal in a Caco-2 monolayer
assay.14 Exploring further, a rat PK study with 12 also resulted in
an encouraging level of oral bioavailability of 56%, and a long
half-life, driven by the higher volume of distribution. However, in
contrast to the other analogues in Table 1, compound 12 showed
significant inhibition of dofetilide binding to the human ether-à-
go-go related gene ion-channel (hERG), with an IC50 of 5.7 lM,
Figure 3. Model of a cyclised version of A66 corresponding to the proposed series 2,
compound 5, docked in the ATP binding site of PI3Ka based on the crystal structure
of the enzyme in complex with alpelisib (PDB code: 4JPS). Key hydrogen bonds are
represented as dashed lines.
which deprioritised further follow up for this compound. Overall,
these data suggested improved PK profiles were possible for the
cyclised analogues 2, at least at low dose levels below that
expected to deliver efficacy, which encouraged further follow up
studies for the most interesting examples, such as compound 11.
On the basis of the favorable outcome of the cyclisation leading
to the 4,5-dihydrobenzo[1,2-d:3,4-d]bisthiazole series 2, the same
strategy was also applied to the proline-amide-urea sub-series
containing a pyrimidyl affinity-pocket binder, as exemplified by
compound 13.3,5 Applying the same annulation leads to the 4,5-di-
hydrothiazolo[4,5-h]quinazoline analogues of the general struc-
ture 14, as depicted in Figure 4.
selectivity were determined for both compounds, but a greater
level of cellular PI3Ka activity was measured for the cyclised ana-
logue 5. This observation, of increased PI3K activity for the cyclised
analogues, particularly at the cellular level, is seen throughout the
series and supports the hypothesised favorable interaction of the
additional ring-methylene with the conserved Y836 residue
(PI3K
a
numbering). The increased cellular activity was especially
The analogues 14 were readily prepared, and starting from the
N-acylated 2-aminothiazole-fused cyclohexanone 3 was found to
be the most efficient route, as outlined in Scheme 2.8
Formylation of the lithium enolate of 3 with methyl formate gave
the b-ketoaldehyde 15 in high yield. Subsequent reaction of 15
with a range of amidines generated the pyrimidine containing tri-
cyclic core, which was followed by acetamide hydrolysis to give
the key intermediates 16. Conversion of these 2-aminothiazole
intermediates 16 into the final proline-amide-urea targets was
interesting because during the optimisation of the series it was
found to be the best predictor of in vivo potency and selectivity.
Evaluating the level of plasma protein binding (PPB) for the
cyclised series showed similar levels to the acyclic analogues when
predicted from the human and rat serum albumin binding (HSA
and RSA), Table 2.11 Thus, indicating the increase in cellular activ-
ity, relative to the biochemical activity, between the two series not
to be driven by a change in unbound fraction.4 The structure

R. A. Fairhurst et al. / Bioorg. Med. Chem. Lett. 25 (2015) 3575–3581
3577
Table 1
PI3K , PI3Kb, PI3K
a
c
and PI3Kd activities for A66 and the 4,5-dihydrobenzo[1,2-d:3,4-d]bisthiazole analogues 5–12
N
S
N
S
N
S
N
S
H
N
H
N
H
N
H
N
S
N
S
N
S
N
S
N
N
N
N
N
O
O
O
O
O
O
O
O
5
9
6
7
8
H2N
H2N
H2N
H2N
N
S
N
S
N
S
N
S
H
H
H
N
CF3
H
S
N
S
N
S
N
S
N
N
N
N
N
O
O
N
O
O
N
O
O
N
O
O
10
11
12
N
H2N
H2N
H2N
H2N
Compound
Biochemical IC50
PI3Kb
(
l
M)
Cellular IC50 (lM)
PI3K
a
PI3K
c
PI3Kd
PI3K
0.20
0.019
0.19
0.028
0.011
0.010
0.010
0.005
0.025
a
PI3Kb
PI3Kd
A66
5
6
7
8
0.015
0.013
0.070
0.008
0.009
0.013
0.011
0.009
0.010
6.5
2.5
5.1
0.72
4.5
1.8
2.3
0.93
5.5
8.0
0.16
1.1
n.d.
8.0
0.33
0.20
0.16
0.99
1.5
0.70
1.6
n.d.
5.6
0.91
1.0
>10
3.6
1.3
>10
1.4
3.4
4.0
>10
1.0
1.5
0.59
4.0
1.2
>10
0.45
1.0
0.98
>10
9
10
11
12
0.23
0.97
n.d. not determined.
Table 2
Physicochemical and in vitro PK data for selected reference compounds and proline-amide-urea aminothiazole derivatives
Compound
clogP/PSA (Å2)
Sol. pH 6.8 (
l
M)
HDM FA (%)
Caco-2 A-B/B-A (10ꢀ6 cm sꢀ1
)
Rat microsome Cl (
l
l minꢀ1 mgꢀ1
)
HSA/RSA (%)
Alpelisib
A66
5
6
7
2.1/101
2.5/101
2.8/101
2.4/101
3.3//101
2.5/101
3.3/101
3.3/101
2.9/101
2.1/104
2.0/114
2.5/114
1.8/114
2.5/114
2.2/14
53
90
65
55
310
351
< 4
9
89
96
93
57
93
76
91
95
86
30
63
84
50
45
52
n.d.
80
29
3.8/18.3
12.3/15.6
n.d.
n.d.
12.1/15.6
n.d.
n.d.
n.d.
n.d.
0.76/6.87
n.d.
3.4/25.5
n.d.
3.4/21.7
n.d.
n.d.
5.2/15.8
n.d.
29
126
37
26
73
96
100
80
34
18
34
33
31
n.d.
18
3
90.0/92.7
n.d.
91.9/92.6
95.3/95.0
91.9/92.1
n.d.
n.d.
n.d.
n.d.
n.d.
8
9
10
11
12
17
18
19
20
21
22
23
24
14
n.d.
125
60
51
71
n.d.
90.3/n.d.
87.8/94.9
90.4/94.2
n.d.
n.d.
78.7/87.9
89.7/94.5
9
1.3/117
2.5/117
1.9/117
276
>1000
500
58
53
clogP was calculated using clogP version 7.4, BioByte Corporation.
Table 3
Pharmacokinetic data for selected proline-amide-urea aminothiazole derivatives
Compound
CL (mL minꢀ1 kgꢀ1
)
Vss (L kgꢀ1
)
Terminal half-life (h)
iv AUC d.n. (nmol h Lꢀ1
)
p.o. AUC d.n. (nmol h Lꢀ1
)
Bioavailability (%)
2.0 1.0
A66
5
74 14
2.0 0.4
0.4 0.0
24 11
1.3 0.3
0.6 0.1
0.5 0.2
3.2 0.1
0.3 0.1
1.5 0.2
9.6 2.8
1.9 0.9
3.7 0.3
1.1 0.2
2.0 0.1
561 101
4132 304
956 227
1128 376
16472 3211
5241 589
1444 190
70 16
1585 388
531
199 111
13599 4065
1874 917
1116 234
10
1
39
56
9
12
13
19
20
23
39 11
42 14
18 10
83 25
36 17
77 16
3
8
26
0
1
3
d.n. = dose normalised to 1.0 mg kgꢀ1
.

3578
R. A. Fairhurst et al. / Bioorg. Med. Chem. Lett. 25 (2015) 3575–3581
achieved following the same two-step protocol outlined in
Scheme 1.
Table 4 shows the biochemical and cellular data for compound
13 and the cyclised analogues 17–24, corresponding to the general
structure 14.10 Consistent with the observations made for series 2,
the SAR was also found to track well from the biaryl analogues to
the tricyclic series 14, and as anticipated a significant increase in
potency, in particular cellular potency, was associated with the
cyclised series. For example, the direct cyclised analogue of 13,
N
O
N
O
N
N
H
N
H
N
S
S
N
N
N
N
R2
R1
O
O
H2N
H2N
13
14
Figure 4. Expansion to a 4,5-dihydrothiazolo[4,5-h]quinazoline series 14: compar-
ison with the acyclic pyrimidine series, as exemplified by compound 13.
compound 17, showed a 7-fold increase in cellular PI3Ka potency.
However, some erosion of the PI3K
a
selectivity (<5-fold) was
N
O
S
N
O
i), ii)
iii), iv)
v), vi), vii)
O
H
2
R2
S
N
N
S
HN
O
Br
O
3
4
Scheme 1. Synthesis of the 4,5-dihydrobenzo[1,2-d:3,4-d]bisthiazole analogues 2. Reagents and conditions: (i) thiourea, EtOH, 80 °C, 4 h (32–52%); (ii) Ac2O, reflux, 2 h (81–
84%); (iii) Br2, AcOH, 75 °C, 18 h (82–86%); (iv) R2C(S)NH2, 0.02 equiv. (NH4)3PMo12O40, EtOH, reflux, 2.5 h (11–72%); (v) concd HCl, EtOH, reflux, 4 h (90–98%); (vi) carbonyl
diimidazole (CDI), DCM, 25 °C, 36 h (49–95%); (vii) (S)-proline-amide derivative, Et3N, DMF, 25 °C, 18 h (46–86%).
N
O
i)
ii), iii)
iv), v)
N
N
O
OH
N
14
S
HN
S
O
H2N
S
N
HN
O
R2
3
15
16
Scheme 2. Synthesis of the 4,5-dihydrothiazolo[4,5-h]quinazoline analogues 14. Reagents and conditions: (i) 3 equiv LiHMDS, ꢀ78 °C, then HCO2Me ꢀ78 to 25 °C, 24 h (81–
95%); (ii) R2C(NH)NH2, pyridine, sealed-tube, 160 °C, 17 h (29–38%); (iii) concd HCl(aq), EtOH, 85 °C, 2 h (90–95%); (iv) CDI, DCM, 25 °C, 2 h (35–92%); (v) (S)-proline-amide
derivative, Et3N, 2:1 CH2Cl2/DMF, 25 °C, 18 h (20–93%).
Table 4
PI3Ka, PI3Kb, PI3Kc and PI3Kd activities for the reference compound 13 and the 4,5-dihydrobenzo[1,2-d:3,4-d]bisthiazole analogues 17–24
N
N
N
N
N
N
N
N
H
N
H
N
H
N
H
N
S
S
S
S
N
N
N
N
N
N
N
N
O
O
O
O
O
O
O
O
17
18
19
20
H2N
H2N
H2N
H2N
N
N
N
N
N
N
N
N
H
H
N
H
N
H
N
S
S
S
S
N
N
N
N
N
N
N
N
CF3
N
O
O
O
O
O
O
O
O
N
21
22
23
24
N
N
H2N
H2N
H2N
H2N
Compound
Biochemical IC50
PI3Kb
(
l
M)
Cellular IC50 (lM)
PI3K
a
PI3K
c
PI3Kd
PI3Ka
PI3Kb
PI3Kd
13
17
18
19
20
21
22
23
24
0.007
0.004
0.005
0.006
0.006
0.005
0.005
0.014
0.004
1.9
0.23
0.38
0.039
0.006
0.004
0.013
0.004
0.004
0.015
0.025
0.26
3.1
1.5
0.22
0.075
0.29
0.32
0.081
0.77
1.4
0.016
0.071
0.18
0.005
0.015
n.d.
0.064
0.070
0.11
0.11
0.021
n.d.
0.22
0.47
0.73
0.28
0.22
3.7
0.29
0.43
0.91
0.20
0.082
1.2
1.5
0.23
0.069
0.40
2.3
>10
0.66
4.4
1.0

R. A. Fairhurst et al. / Bioorg. Med. Chem. Lett. 25 (2015) 3575–3581
3579
observed for 7 versus 13, when compared to the other class 1 PI3K
isoforms, and this was also apparent for the majority of the sub-
series of the general structure 14. Small quaternary alkyl groups
in the pyrimidyl 2-position, and the 2- and 3-methyl proline-amide
more an absence of mTor activity for the sub series 2 and 14 was
determined, which was also evident in the parent biaryl series.
To increase the understanding of the interactions between the
tricyclic series 2 and 14, an X-ray structure was solved for com-
variants were again found to provide the most potent PI3K
a
inhi-
pound 20 bound into the ATP-pocket of PI3Ka, as shown in
bitors, as exemplified by 18–21. The cis-3-methylproline ana-
logues, 20 and 21, provided the highest level of potency for the
sub-series 14, as well as from our investigations into the whole
2-aminothaizole proline-amide-urea series. Incorporation of a ter-
tiary amine, as a solubility enhancing moiety, was again explored
at the proline 2-, 3- and 4-positions as exemplified by 22–24.8
The 2- and cis-3-dimethylaminomethylproline analogues, 22 and
23, in particular retained a high level of potency, with an accept-
able level of selectivity versus the other class 1 PI3K isoforms.
Evaluating the physical properties of the 4,5-dihydrothia-
zolo[4,5-h]quinazoline analogues 14, solubilities were again mod-
Figure 5.19 The structure revealed the expected binding interac-
tions of the aminothiazole moiety within the hinge region, and
for the proline-amide with the non-conserved PI3Ka-specific
Q859 residue.6 In addition, the starting hypothesis was validated
with the methylene, introduced to close the central ring, making
hydrophobic contacts with Y836.
To better understand the potential of the sub-series 2 and 14 as
PI3K
a
inhibitors, pharmacodynamic (PD) studies were conducted
in nude mice bearing Rat1-myr-p110
a
xenografts.20 The data from
two separate studies with compounds 11 and 19 are shown in
Figure 6. Similar mouse in vitro PK parameters were determined
for both compounds: microsome extraction ratios of 69% and
47%; PPB levels of 89% and 90%, respectively, for 11 and 19. In each
study plasma levels and the extent of the inhibition of phosphory-
lation of Akt, on residue Ser473 (p-Akt) as a downstream readout of
PI3Ka inhibition, were measured at the indicated time points.
Compound 11, dosed as a suspension, produced a PK profile
which strongly suggested an extended absorption phase, consis-
est in the absence of the tertiary amino group (<125 lM). More
favorable was the combination of a moderately high permeability
with high metabolic stability for these analogues, enabling good
levels of rat oral exposure to be obtained, as exemplified by com-
pounds 19 and 20 from Table 4. In particular, a low dose rat PK
study with 19 resulted in an excellent oral bioavailability of 83%
and particularly high exposure levels, as shown in Table 3.
However, plasma protein binding measurements indicated lower
free fractions in the rat for these analogues (<1%) compared to
the other sub series: for example, free fractions of 0.4% and 6.0%
were determined for compounds 19 and 13 respectively.17 As a
consequence the increases in exposure, when expressed as free
drug levels, were less dramatic, but were still significantly
improved over the biaryl parent compounds. Interestingly, the
serum albumin affinities indicated a higher level of binding in
the rat compared to human and this was established to be the case
(human PPB for 19: 91.3%), but the extent of the high degree of
binding in the rat was underpredicted by this method. The tertiary
amine containing analogues 22–24 all showed higher solubilities.
Compound 23 in particular proved interesting: the nitrogen atom
in the b-position to the tertiary amine lowering the pKa to 6.6.18
The reduced basicity of compound 23 was considered a key factor
in reducing the hERG activity, with no significant inhibition at con-
tent with the low solubility. Plasma concentrations above 1 lM
were measured up to 8 h post dosing for compound 11, and were
associated with a high level of inhibition of p-Akt. Compound 19,
dosed as a solution, produced a high plasma concentration at 8 h
post dosing, which had fallen to 1 lM after 16 h. The high level
of variability seen in the plasma levels of 19, at both time points,
can also be assigned to the low solubility: potentially arising as a
result of the compound precipitating in the gut, and leading to high
variability in the extent of absorption. Inhibition of p-Akt for com-
pound 19 was complete at 8 h after dosing and remained high after
16 h. For both compounds the correlation between free plasma
concentration and the extent of p-Akt inhibition, in the context
of the Rat1-myr-p110
a cellular IC50 values, supported that the
anticipated PK/PD relationship was operating.3 Furthermore, com-
parison of the levels of p-Akt inhibition obtained for 11 and 19,
with those observed for alpelisib, would predict both compounds
centrations up to 30
lM, whilst the solubility at pH 6.8 remained
to have the potential to inhibit the growth of PI3Ka-dependent
high at >1 mM. High permeability and a reasonable level of
in vitro metabolic stability were also measured for compound 23,
which resulted in good exposure and an oral bioavailability of
77% being obtained in a low dose rat PK study.
tumor-xenografts following chronic twice-daily dosing at, or close
to, the above dose levels, assuming consistent exposure could be
To further characterise the series 2 and 14, the analogues were
tested against an internal panel of 35 kinase assays, including the
lipid kinases phosphatidylinositol-4-kinase beta (PI4Kb), Vps34
and mTor. All the compounds exhibited no significant inhibition
at concentrations up to 10
the exception of PI4Kb and Vps34, as exemplified for the analogues
5, 10, 12, 18, 20 and 23 in Table 5. These data support a high PI3K
lM in these biochemical assays, with
a
selectivity for the series versus other kinases, with the PI4Kb selec-
tivity at a similar level to that within the class 1 PI3K family. Once
Table 5
PI4Kb and Vps34 activities for selected analogues from the series 2 and 14
Compound
Biochemical IC50 (lM)
PI4Kb
Vps34
5
0.32
0.21
3.2
0.065
0.16
4.0
>9.1
9.6
>9.1
7.0
>9.1
>9.1
10
12
18
20
23
Figure 5. X-Ray structure of compound 20 bound into the ATP pocket of PI3Ka,
solved with a resolution of 2.8 Å. Only residues proximal to the inhibitor are
represented. Key hydrogen bonds are represented as dashed lines.

3580
R. A. Fairhurst et al. / Bioorg. Med. Chem. Lett. 25 (2015) 3575–3581
favourable potency and selectivity profiles of these tricyclic series
whilst tackling the above limitation.21
Acknowledgements
The authors would like to thank Dorothea Haasen and Joerg
Trappe for conducting biochemical assays, Peter Wipfli for per-
forming bioanalysis and Werner Gertsch, Marijana Petrovic,
Philippe Ramstein and Martin Kessler for their excellent technical
assistance.
References and notes
1. Polivka, J., Jr.; Janku, F. Pharmacol. Ther. 2014, 142, 164.
2. Flanagan, J. U.; Shepherd, P. R. Biochem. Soc. Trans. 2014, 42, 120.
3. 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.; Brachmann, S. M.; Maira, S.-M.; Wilson, C.; Boehm, M.;
García-Echheverría, C.; Chène, P.; Wiesmann, M.; Cozens, R.; Lehar, J.; Schlegel,
R.; Caravatti, G.; Hofmann, F.; Sellers, W. R. Mol. Cancer Ther. 2014, 13, 1117.
4. Fairhurst, R. A.; Imbach-Weese, P.; Gerspacher, M.; Caravatti, G.; Furet, P.;
Zoller, T.; Fritsch, C.; Haasen, D.; Trappe, J.; Guthy, D. A.; Arz, D.; Wirth, J. Bioorg.
Med. Chem. Lett. 2015, 25, 3569.
5. Bruce, I.; Akhlaq, M.; Bloomfield, G. C.; Budd, E.; Cox, B.; Cuenoud, B.; Finan, P.;
Gedeck, P.; Hatto, J.; Hayler, J. F.; Head, D.; Keller, T.; Kirman, L.; Leblanc, C.; Le
Grand, D.; McCarthy, C.; O’Connor, D.; Owen, C.; Oza, M. S.; Pilgrim, G.; Press, N.
E.; Sviridenko, L.; Whitehead, L. Bioorg. Med. Chem. Lett. 2012, 22, 5445.
6. Furet, P.; Guagnano, V.; Fairhurst, R. A.; Imbach-Weese, P.; Bruce, I.; Knapp, M.;
Fritsch, C.; Blasco, F.; Blanz, J.; Aicholz, R.; Hamon, J.; Fabbro, D.; Caravatti, G.
Bioorg. Med. Chem. Lett. 2013, 23, 3741.
7. (a) Brandl, T.; Maier, U.; Hoffmann, M.; Scheuerer, S.; Joergensen, A. T.; Pautsch,
A.; Breitfelder, S.; Grauert, M.; Hoenke, C.; Erb, K.; Pieper, M.; Pragst, I.
Preparation of dihydrothiazoloquinazolines as PI-3 kinase inhibitors. WO 2007/
115929, Oct. 18, 2007.; (b) Betzemeier, B.; Brandl, T.; Breitfelder, S.; Brueckner,
R.; Gerstberger, T.; Gmachl, M.; Grauert, M.; Hilberg, F.; Hoenke, C.; Hoffmann,
M.; Impagnatiello, M.; Kessler, D.; Klein, C.; Krist, B.; Maier, U.; McConnell, D.;
Reither, C.; Scheuerer, S.; Schoop, A.; Schweifer, N.; Simon, O.; Steegmaier, M.;
Steurer, S.; Waizenegger, I.; Weyer-Czemilofsky, U.; Zoephel, A. Preparation of
thiazoloindazoles for treatment and prevention of cancer. WO 2006/040281,
April 20, 2006.; (c) Breitfelder, S.; Maier, U.; Brandl, T.; Hoenke, C.; Grauert, M.;
Pautsch, A.; Hoffmann, M.; Kalkbrenner, F.; Joergensen, A.; Schaenzle, G.;
Peters, S.; Büttner, F.; Bauer, E. Preparation of fused thiazoles such as
pyrazolobenzothiazoles, thiazolocyclopentapyrazoles, thiazolocycloheptapyrazoles,
thiazoloquinazolines, and naphthothiazoles as PI3 kinase inhibitors. WO 2006/
040279, April 20, 2006.
8. Experimental details of the syntheses of the final products and intermediates
are described. In: Fairhurst, R. A.; Gerspacher, M.; Mah, R. Preparation of
carboxamidecycloaminoureathiazole derivatives for use as PI3-K inhibitors.
WO 2011/000855, Jan. 6, 2011.
9. Das, B.; Reddy, V. S.; Ramu, R. J. Mol. Cat. 2006, 252, 235.
10. The biochemical (all four isoforms) and cellular (a, b and c isoforms) PI3K
Figure 6. PK/PD relationships for compounds 11 and 19. Female athymic mice
bearing subcutaneous xenotransplants of Rat1-myr-p110
a tumors were treated
with a single dose of 50 mg kgꢀ1, p.o of compound 11 (A) or 19 (B). At the indicated
time points, the groups of mice (n = 4) were sacrificed, blood and tissues collected.
Each tumor tissue was flash-frozen then pulverized and analyzed by Western blot
to determine Ser473 p-Akt levels and in parallel the concentration of each compound
was analyzed and quantified in plasma.
activities were determined in the same manner as described in Ref. 4.
11. Affinity for rat or human serum albumin, as a predictor of PPB, was measured
using a chromatography method as described in: Reilly, J.; Etheridge, D.;
Everatt, B.; Jiang, Z.; Aldcroft, C.; Wright, P.; Clemens, I.; Cox, B.; Press, N. J.;
Watson, S.; Porter, D.; Springer, C.; Fairhurst, R. A. J. Liq. Chromatogr. Relat.
Technol. 2011, 34, 317.
12. Lorthiois, E.; Marek, I.; Normant, J. F. J. Org. Chem. 1998, 63, 2442.
13. Passive membrane permeability was assessed using a hexadecane membrane
(HDM) permeability assay, as described in: Wohnsland, F.; Faller, B. J. Med.
Chem. 2001, 44, 923; Data are expressed as the predicted fraction absorbed (FA)
from the gastrointestinal tract following oral administration. The permeability
of selected compounds was measured using Caco-2 monolayers, as described
in: Pade, V.; Stavchansky, S. J. Pharm. Sci. 1998, 87, 1604.
14. Solubility was measured using a high throughput equilibrium method at pH
6.8, as described in: Glomme, A.; März, J.; Dressman, J. B. J. Pharm. Sci. 2005, 94,
1.
15. In vitro metabolic stability was measured, using the compound depletion
approach, in rat liver microsomes from pooled rat hepatic microsome
preparations; microsomes (protein concentration 0.5 mg/mL) were incubated
maintained.3 However, indications of solubility limited phenom-
ena are apparent at the efficacious dose levels for both compounds
in these and in other studies.
In conclusion, annulation of two biaryl series of (S)-proline-
amide aminothiazole-urea PI3Ka inhibitors generated the tricyclic
series 2 and 14 which maintained the favorable selectivity profiles
of the parent compounds, and additionally led to an increase in
PI3Ka potency. The rationale for this annulation was to strengthen
the interaction with Y836, as a way to increase the affinity for the
PI3K ATP-pocket, and the validity of this starting hypothesis has
been supported by an X-ray structure of the cyclised analogue 20
within PI3Ka. In addition, low dose rat PK studies have shown
improved oral exposures for the annulated series, when compared
to the biaryl parent compounds. PK/PD studies in the mouse with
analogues from both of the sub-series 2 and 14 have shown them
to be capable of inhibiting signaling through the PI3K pathway for
greater than 8 hours, demonstrating their potential as in vivo
at 37 °C with the test compounds (1 lM) and the co-factor NADPH. Aliquots
were removed at 0, 5, 15 and 30 min. Reactions were stopped by addition of
acetonitrile and samples were subsequently analysed by LC–MS/MS after
protein precipitation. The data were analysed as percentage disappearance of
parent relative to the zero time sample, from which hepatic extraction ratios
were determined.
16. Rat pharmacokinetics: were conducted in female animals of the Sprague-
Dawley strain; assessed blood levels from 5 minutes to 48 h post dosing;
animals were conscious, permanently-cannulated with free access to food;
doses of 1.0 and 3.0 mg kgꢀ1 were applied in the intravenous and oral arms
respectively.
PI3K
a inhibitor tool compounds. However, a non-basic example
which was not associated with some type of solubility-limited
exposure was not identified from the sub-series 2 and 14, and
resulted in future efforts being focused upon retaining the

R. A. Fairhurst et al. / Bioorg. Med. Chem. Lett. 25 (2015) 3575–3581
3581
17. Plasma protein binding was measured by a rapid equilibrium dialysis method.
18. pKa values were determined by UV-absorbance changes as the test compounds
are exposed to a calibrated pH gradient.
Burger, M.; Knapp, M.; Sterker, D.; Schnell, C.; Guthy, D.; Nagel, T.; Wiesmann,
M.; Brachmann, S.; Fritsch, C.; Dorsch, M.; Chène, P.; Shoemaker, K.; De Pover,
A.; Menezes, D.; Martiny-Baron, G.; Fabbro, D.; Wilson, C. J.; Schlegel, R.;
Hofmann, F.; García-Echeverría, C.; Sellers, W. R.; Voliva, C. F. Mol. Cancer Ther.
2012, 11, 317.
19. The single-crystal X-ray structure of compound 20 bound into PI3Ka has been
deposited with the Protein Data Bank ID: 4ZOP.
20. PK/PD studies were conducted with female athymic nude mice bearing Rat1-
myr-p110 xenografts, as described in: Maira, S.-M.; Pecchi, S.; Huang, A.;
21. Gerspacher, M.; Fairhurst, R. A.; Mah, R.; Roehn-Carnemolla, E.; Furet, P.;
Fritsch, C.; Guthy, D. A. Bioorg. Med. Chem. Lett. 2015, 25, 3582.
a