Discovery of novel amino-pyrimidine inhibitors of the insulin-like growth factor 1 (IGF1R) and insulin receptor (INSR) kinases; parallel optimization of cell potency and hERG inhibition

Article abstract of DOI:10.1039/c5md00097a

The insulin-like growth factor-1 receptor (IGF1R) and closely related insulin receptor (INSR) are receptor tyrosine kinases which have been postulated to play a role in the tumorigenesis of certain cancers. Strategies for inhibiting oncogenic signaling via the IGF1R and INSR include IGF1R antibodies, IGF1/2 antibodies and dual IGF1R/INSR tyrosine kinase inhibitors (TKIs). IGF1R/INSR TKIs linsitinib (OSI-906) and BMS-754807 have progressed to phase II/III clinical studies in cancer patients. We describe here our efforts to develop small molecule dual inhibitors of the IGF1R/INSR receptor kinases based on an amino-pyrimidine structural class. Our main focus was the parallel optimization of cellular potency and off target activity (principally hERG inhibition) through modulation of physicochemical properties and introduction of key structural motifs using a matched molecular pairs approach and hERG homology model.

Full text of DOI:10.1039/c5md00097a

View Article Online
View Journal
MedChemComm
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: H. Tye, S. Zahn,
M. Hofmann, M. Sanderson, O. Schaff, M. Troy, M. Mayer, G. Rast, S. Pal and U. Guertler, Med. Chem.
Commun., 2015, DOI: 10.1039/C5MD00097A.
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Accepted Manuscripts are published online shortly after
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
Information for Authors.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the Ethical guidelines still
apply. In no event shall the Royal Society of Chemistry be held
responsible for any errors or omissions in this Accepted Manuscript
or any consequences arising from the use of any information it
contains.
www.rsc.org/medchemcomm

Page 1 of 14
MedChemComm
View Article Online
DOI: 10.1039/C5MD00097A
Title
Discovery of novel amino-pyrimidine inhibitors of the insulin-like growth factor 1 (IGF1R) and insulin
receptor (INSR) kinases; parallel optimization of cell potency and hERG inhibition
Authors
Heather Tyea*, Ulrich Guertlerc, Marco H. Hofmannb, Moriz Mayerb, Sandeep Pala, Georg Rastc,
Michael P. Sandersonb, Otmar Schaafb, Matthias Treub, Stephan K. Zahnb
(a) Evotec (UK) Ltd, 114 Innovation Drive, Milton Park, Abingdon, Oxfordshire OX14 4SA, UK
(b) Boehringer Ingelheim RCV GmbH & Co KG, Dr.-Boehringer-Gasse 5-11, 1120 Vienna, Austria
(c) Boehringer Ingelheim Pharma GmbH & Co. KG, Birkendorfer Str. 65, 88397 Biberach / Riß,
Germany
Abstract
The insulin-like growth factor-1 receptor (IGF1R) and closely related insulin receptor (INSR) are
receptor tyrosine kinases which have been postulated to play a role in the tumorigenesis of certain
cancers. Strategies for inhibiting oncogenic signaling via the IGF1R and INSR include IGF1R antibodies,
IGF1/2 antibodies and dual IGF1R/INSR tyrosine kinase inhibitors (TKIs). IGF1R/INSR TKIs linsitinib
(OSI-906) and BMS-754807 have progressed to phase II/III clinical studies in cancer patients. We
describe here our efforts to develop small molecule dual inhibitors of the IGF1R/INSR receptor
kinases based on an amino-pyrimidine structural class. Our main focus was the parallel optimization
of cellular potency and off target activity (principally hERG inhibition) through modulation of
physicochemical properties and introduction of key structural motifs using a matched molecular
pairs approach and hERG homology model.
Introduction
The insulin-like growth factor-1 receptor (IGF1R) and closely related homolog the insulin receptor
(INSR) are receptor tyrosine kinases which have been postulated to play a role in the tumorigenesis
of certain cancers.1 Engagement of the receptors by their cognate ligands (IGF1, IGF2 and insulin)
leads to auto-phosphorylation of specific intracellular domain tyrosine residues. The docking of
effector proteins, such as IRS-1, to these phospho-tyrosines activates signal transduction via the
Ras/Raf/MAPK and PI3K/AKT/mTOR pathways and promotion of oncogenic cellular responses
including proliferation, cell growth and survival.2 As such, much interest has been devoted towards
the development of therapeutic agents targeting the function of the IGF1R and INSR.
Strategies for inhibiting oncogenic signaling via the IGF1R and INSR include IGF1R antibodies3,IGF1/2
antibodies4 and dual IGF1R/INSR tyrosine kinase inhibitors (TKIs).5 The efficacy of IGF1R antibodies in
clinical studies, which target the IGF1R meanwhile sparing the INSR, has been mostly disappointing.6
Preclinical data suggests that co-targeting of the IGF1R and INSR, using small molecule kinase
inhibitors, more broadly inhibits the pathway and is associated with greater anti-tumor efficacy.7
Amongst the IGF1R/INSR TKI inhibitor class, linsitinib (OSI-906) and BMS-754807 (Figure 1a) have
progressed to phase II/III clinical studies in cancer patients. Preclinical data have indicated that BMS-

MedChemComm
Page 2 of 14
View Article Online
754807 is a potent inhibitor of the IGF1R and INSR with IC50 values of 1.8 and 1.7 nM, rDeOspI:e10c.t10iv3e9/lCy5.MD00097A
Additional kinases are also inhibited with IC50 values below 50 nM including TrkA, TrkB, c-Met, RON,
Aurora A and Aurora B. Thus BMS-754807 is somewhat less specific than linsitinib (OSI-906), which
displays potent and selective inhibition of only IGF1R and INSR.5,8
Figure 1a – Literature IGF1R Inhibitors
N
HN
NH2
N
NH
N
N
O
N
N
H
N
N
N
N
N
F
OH
Linsitinib
BMS-754807
We describe here our efforts to develop small molecule dual inhibitors of the IGF1R/INSR receptor
kinases based on an amino-pyrimidine structural class. Our main focus was the modulation of
physicochemical properties and introduction of structural motifs which would enable optimization of
cellular potency and off target activity (principally hERG inhibition) in parallel.
Results and discussion
During follow-up activities from an HTS campaign to identify small molecule dual inhibitors of the
IGF1R/INSR kinases, we identified a novel class of amino-pyrimidine analogues bearing an amino
benzyl oxindole substituent on the pyrimidine ring (Figure 1b). Compounds of this type were found
to be dual IGF1R/INSR inhibitors (assessment of a number of early compounds showed no selectivity
between IGF1R and INSR activity, data not shown) with modest activity in cell lines dependent on
IGF1R/INSR signaling (e.g. TC-177 and TC-71) and little activity in cells lacking IGF1R/INSR
dependence (e.g. HCT 116) (Table 1). The binding mode of this class of compound (as determined for
a closely related structural class, unpublished results) was as expected with the amino-pyrimidine
motif making key hydrogen bonding interactions with the hinge region of the ATP binding pocket.
Figure 1b – Pyrimidine analogues 1-8

Page 3 of 14
MedChemComm
View Article Online
DOI: 10.1039/C5MD00097A
CF3
CF3
CF3
H
N
H
N
CF3
H
N
H
N
NH
NH
O
NH
O
NH
O
N
N
N
N
N
N
N
N
O
HN
HN
O
HN
HN
H
N
O
N
N
O
O
N
N
N
O
N
N
3
4
1
2
CF3
CF3
CF3
H
Cl
H
H
N
H
N
N
N
NH
O
NH
O
NH
NH
N
N
N
N
N
N
N
N
O
O
HN
HN
HN
HN
O
O
N
O
O
N
N
N
N
O
N
N
N
H
O
6
5
7
8
Table 1 – Kinase and cell potencies for analogues 1-4
Compound Number IGF1Ra IC50 [nM] TC177b EC50 [nM]
HCT116c EC50 [ nM]
(1)
(2)
(3)
(4)
14
9
1
216
97
61
3475
895
3222
1
66
1968
Table 1 footnote:
(a) Mean IC50 values. Variation typically +/- 50% based on control compound IC50 69 nM, SD 34,
n of 80. Control compound was a non-specific ATP competitive kinase inhibitor tool
compound from Boehringer Ingelheim
(b) Mean ECC50 values. Variation typically +/- 50% based on control compound EC50 31 nM, SD
18, n of 5. The control compound used for all cell proliferation assays was a broadly
cytotoxic tool compound from Boehringer Ingelheim
(c) Mean EC50 values. Variation typically +/- 50% based on control compound EC50 5 nM, SD 3, n
of 307. The control compound used for all cell proliferation assays was a broadly cytotoxic
tool compound from Boehringer Ingelheim
An early optimization objective was to improve the cellular potency of this class of compounds.
Replacement of the polar amide motif in compound 1 with a piperazine resulted in an improvement
in cell potency (compound 2) but this was accompanied by a loss in cellular window with respect to
HCT 116 activity. A relatively high dose of 100 mg/kg p.o. (qd) was necessary to achieve a modest
tumor growth inhibition (TGI) of 59% in a subcutaneous allograft model (3T3-hIGF1R). The cellular
window could be regained by modification of the oxindole motif giving rise to a series of compounds
with a good cell window and acceptable cell potency (e.g. compounds 3 and 4). PK profiling revealed

MedChemComm
Page 4 of 14
View Article Online
compound 4 as a suitable tool compound with good oral exposure which could be usedDfoOIr: 1p0r.1o0f3i9li/nCg5MD00097A
in an efficacy model (see Table 4). A dose of 75 mg/kg p.o. (qd) of compound 4 delivered TGI values
of 71% and 103% in the 3T3-hIGF1R allograft and subcutaneous GEO xenograft models, respectively
(Table 5).
Encouraged by this result we sought to further optimize this compound with the goal of increasing
efficacy at a reduced dose by further improving the cellular potency. However, following additional
profiling of compound 4 we discovered that it was a potent inhibitor of the hERG potassium channel
(IC50 = 160 nM). From a CV safety perspective we considered a window between the hERG IC50 and
the unbound Cmax of >30 fold to be a minimum requirement. For compound 4 with an unbound
Cmax of 16 nM (based on Cmax 2600 nM, PPB 99.4%, see table 4) this window was only 10 fold.
With the expectation that the PK profiles of compounds in the series would be broadly similar we set
an objective of increasing the hERG IC50 to at least 1 M to be sure of achieving the desired window.
This led to the dual objective of optimizing cellular potency and hERG inhibition in parallel.
Compound 4 and the related 5-trifluoromethyl-2,4-diaminopyrimidines can be synthetically accessed
via the procedure shown in Figure 2 starting from 5-trifluoromethyl-2,4-dichloro-pyrimidine and
subsequent nucleophilic aromatic substitution as described by co-workers at Pfizer.9 For the 5-
trifluoromethyl series the regioselective introduction of the R2-aniline is catalyzed and facilitated by
zinc (II)-chloride. In the 5-halogen series the R4-amine can be introduced first with high
regioselectivity under basic conditions. Further details of synthetic routes are included in the
supporting information.
Table 2 – Kinase, cell, hERG and physicochemical properties for analogues 4-8
Compound
Number
IGF1Ra IC50 TC177b EC50
HCT116c
EC50 [ nM]
1968
hERG IC50
±SD [nM]
166±13
996±65
> 3000
logDd pH
7.4
Calc. pKae
[nM]
[nM]
(4)
(5)
(6)
(7)
1
66
4.3
4.0
2.2
2.3
7.8
7.8
n.a.
9.7
9.6
2.5
1.5
0.2
0.4
122
140
36
2843
3160
1637
1663
3186±171
n.d.
(8)
59
3.9
Table 2 footnote:
(a) Mean IC50 values. Variation typically +/- 50% based on control compound IC50 69 nM, SD 34,
n of 80. Control compound was a non-specific ATP competitive kinase inhibitor tool
compound from Boehringer Ingelheim
(b) Mean ECC50 values. Variation typically +/- 50% based on control compound EC50 31 nM, SD
18, n of 5. The control compound used for all cell proliferation assays was a broadly
cytotoxic tool compound from Boehringer Ingelheim
(c) Mean EC50 values. Variation typically +/- 50% based on control compound EC50 5 nM, SD 3, n
of 307. The control compound used for all cell proliferation assays was a broadly cytotoxic
tool compound from Boehringer Ingelheim
(d) logD values calculated using Chemaxon software
(e) pKa values calculated using Chemaxon software

Page 5 of 14
MedChemComm
View Article Online
DOI: 10.1039/C5MD00097A
Figure 2 – Synthetic routes
CF3
CF3
CF3
Cl
R2-NH2,
ZnCl2
H
R4-NH2
N
Cl
R1
N
N
N
N
N
N
HN
HN
Cl
R2
R2
X2
X1
compound
1,2,3,4,6,7,8, 9, 16, 18
Cl, Br
H
Cl, Br
N
Cl, Br
N
H
N
N
R4-NH2
base
R2-NH2
cat. H+
Cl
R1
R4
N
N
N
N
HN
Cl
Cl
R2
Y2
Y1
compound
5, 10-15, 17,19, 20
Figure 2 footnote:
For conversion X1 to X2: typical scale 1 – 15 g, yields approx. 50 – 90%; X2 to compound: typical
scale0.05 to 5 g, yields approx. 50 to 80%.
For conversion Y1 to Y2: typical scale 0.5 to 5 g, yields approx. 60 - 85%; Y2 to compound or
intermediate for final decoration: typical scale 0.05 to 1 g, yields 50 – 90%
Common strategies for reducing the hERG inhibition potential of compounds include reducing
lipophilicity, reducing basicity and/or altering the position of the basic center relative to the
lipophilic motif(s).10 We decided to explore all of these options in parallel by preparing a limited
number of compounds in which either lipophilicity (replacing CF3 with Cl in compound 5) or the
nature of the basic solubilizing group was altered (compounds 6 and 7). Both approaches resulted in
a reduction in hERG inhibition (see Table 2) however the incorporation of an amide motif
(compound 6) led to an undesirable reduction in cell potency. Switching from a piperazine to an
amino piperidine (compound 7) proved to be favorable in terms of both cell potency and hERG
inhibition.
At this point in our optimization campaign it was decided that it would be preferable to change the
linkage (X in figure 4) of the solubilizing motif from a nitrogen atom to a carbon atom. This was done
to mitigate the risk of releasing a potentially mutagenic aniline upon metabolism of the piperazine or
amino piperidine ring. Compound 8 was prepared to verify that such a change would be tolerated
with respect to IGF1R inhibition. This was indeed found to be the case with compound 8 retaining a
good level of cell potency and cellular window (see table 2).

MedChemComm
Page 6 of 14
View Article Online
To assist in the design of further analogues we decided to employ a homology model ofDtOhIe: 10h.E10R3G9/Cio5MnD00097A
channel. To this end we reproduced a model of the open form of the hERG channel described by
Aqvist11 into which we docked compound 4 (see Figure 3). The resultant docked pose suggested that
a key interaction was made between the basic center of compound 4 and serine residues in the
selectivity pocket. This was in agreement with our experimental observation that altering the
basicity or position of this motif had an impact on the level of hERG inhibition. The side chain from
the basic motif protruded into a strongly electronegative region of the hERG channel suggesting that
introduction of electronegative groups at this position may further disrupt binding to the hERG
channel.
Figure 3 – Schematic of hERG model for compound 4
Tyr
Ser
CF3
H
N
NH
N
N
O
Phe
HN
Ser
O
N
N
Phe
Figure 4 – Scaffold and R groups for analogues 9-20
O
N
R1
R3a
H
N
NH
OH
R3b
R3c
R3d
N
N
O
HN
O
o
X
m
R2
OH
N
R3
Based on the results of the hERG modelling and the data obtained so far the best approach was
considered to be variation of R1, R2 and R3 in parallel to provide molecules with reduced logD and

Page 7 of 14
MedChemComm
View Article Online
bearing motifs at R3 which could reduce the hERG inhibition potential either by modulaDtiOnI:g10b.1a0s3i9c/iCty5MD00097A
or introducing electronegative side chains (Figure 4). A number of analogues were prepared
following a matched molecular pairs approach12 to aid in data interpretation; Plot 1 shows the
relationship between pIC50 and logD for either IGF1R or hERG inhibition of these compounds. Whilst
there was a clear trend towards reduced hERG activity at reduced logD there was no apparent
relationship between IGF1R activity and logD. This gave us confidence that it should be possible to
identify potent IGF1R inhibitors with reduced hERG liability within this chemical series. The
remaining challenge was then to achieve this whilst maintaining cellular potency.
Plot 1 – Relating hERG and IGF1R potency to logD
Plot 1 Footnote:
Plot relating pIC50 IGF1R to logD shows no correlation (R2 = 0); Plot relating pIC50 hERG and logD
shows some correlation (R2 = 0.36), indicating that optimization of hERG inhibition can be achieved
by modulation of logD.
Table 3 summarizes the data for a subset of analogues for which cell data was obtained. For each R3
modification the matched molecular pairs (varying R1) were prepared. In all cases the less lipophilic
R1 = Cl analogue proved to have the better hERG profile, this came at the expense of a slight
reduction in potency in the IGF1R kinase assay, although the effect on cell potency was relatively
neutral (see data for pairs 9 &11, 16&17, 18&20). In fact only the Cl analogues were found to inhibit
hERG in the desirable range of IC50 > 1 M.

MedChemComm
Page 8 of 14
View Article Online
DOI: 10.1039/C5MD00097A
Table 3 – Kinase, cell, hERG and physicochemical properties for analogues 4 and 9-20
Compound
Number
X
R1
R2
R3 IGF1Ra TC177b HCT116c hERG IC50
logD
pH
Calc.
pKae
IC50
[nM]
1
EC50
[nM]
66
(22*)
330
80
260
118
(54*)
135
94
EC50
[ nM]
1968
±SD [nM]
7.4d
4.3
(4)
N
CF3
o-
OMe
H
H
H
Me
166±13
7.8
(9)
C
C
C
C
CF3
Br
Cl
a
a
a
b
1.5
0.8
7.4
1.0
1755
2772
>2500
1820
749±12
736±49
1541±70
898±46
4.3
4.2
4.0
3.1
n.a.
n.a.
n.a.
9.0
(10)
(11)
(12)
Br
H
(13)
(14)
C
C
Cl
Br
H
H
b
c
2.1
0.7
1413
1080
1165±54
267±14
2.9
3.7
9.0
9.1
(57*)
107
(51*)
9*
(15)
C
Cl
H
c
2.4
3.5
9.1
1127
325±27
(16)
(17)
(18)
C
C
C
CF3
Cl
CF3
o-Me
m-Me
o-
OMe
o-
OMe
o-
OMe
d
d
d
0.3
0.6
1704
649
387±15
1179±60
3.8
3.5
3.3
9.5
9.5
9.5
9*
0.5
0.9
1.1
8*
9*
1745
1353
2302
867±46
858±33
(19)
(20)
C
C
Br
Cl
d
d
3.2
3.0
9.5
9.5
15*
3229±113
Table 3 footnote:
(a) Mean IC50 values. Variation typically +/- 50% based on control compound IC50 69 nM, SD 34,
n of 80. Control compound was a non-specific ATP competitive kinase inhibitor tool
compound from Boehringer Ingelheim
(b) Mean ECC50 values. Variation typically +/- 50% based on control compound EC50 31 nM, SD
18, n of 5. The control compound used for all cell proliferation assays was a broadly
cytotoxic tool compound from Boehringer Ingelheim
(c) Mean EC50 values. Variation typically +/- 50% based on control compound EC50 5 nM, SD 3, n
of 307. The control compound used for all cell proliferation assays was a broadly cytotoxic
tool compound from Boehringer Ingelheim
(d) logD values calculated using Chemaxon software
(e) pKa values calculated using Chemaxon software
Based on earlier data for compound 6 the introduction of an amide group at R3 (R3a) was expected
to lead to reduced hERG inhibition. This was indeed the case, with compounds 9-11 showing the
better profile; however, as with compound 6, this came at the expense of reduced cellular potency.
An alternative approach was to incorporate a polar group at R3 (e.g. R3b) resulting in a reduction in
logD. Compounds 12 and 13 showed a reduction in hERG inhibition and retained IGF1R activity but
again the cellular potency was reduced. Initially, we believed this to be as a result of the H-bond
donor motif, however, the methyl ether R3c (compounds 14 and 15) showed similar activity in the

Page 9 of 14
MedChemComm
View Article Online
DOI: 10.1039/C5MD00097A
cellular assay and in fact had dramatically increased hERG inhibition. Based on our earlier modelling
we concluded that the OH group in R3b may be giving rise to an electrostatic repulsion in the hERG
selectivity pocket. We decided to retain this feature in R3 and use substitution alpha to the OH (R3d)
along with variations at R1 and R2 to tune the lipophilicity in the hope of achieving improved cellular
potency.
Compound 16 gave the first indication that such an approach could provide compounds with cell
potency in the 10 nM range. Incorporation of the R1 = Cl group (compound 17) lead to a reduction in
logD which resulted in good cellular potency combined with acceptable hERG inhibition.
Modification of the R2 group to methoxy provided compounds with a further reduced logD which
retained cellular potency and had a good cellular window (with respect to HCT 116) and reduced
hERG inhibition. Compounds 17 and 20 had the best overall profile close to our objective of single
digit nanomolar cellular potency combined with hERG inhibition at IC50 > 1 µM.
PK profiling of compounds 17 and 20 revealed that they had good oral exposure in mice with a
similar profile to compound 4 (Table 4). Both compounds were considered to have suitable CV safety
profiles with hERG to unbound Cmax ratios >100 fold. In the 3T3 allograft compound 20 resulted in a
TGI of only 52% following 22 days at the 75 mg/kg daily dose; for this reason only compound 17 was
subjected to further profiling.
Table 4 - PK data for analogues 4, 17 and 20
Compound
Number
PK Mouse p.o.
PPB Mouse
Dose (mg/kg)
Cmax (nM)
2600
tmax (h)
1.5
AUC (nM·h)
34000
(% bound)
99.4
≥ 99.9
(4)
(17)
(20)
50
40
50
2000
4500
1.2
0.5
24000
45000
≥ 99.9
Table 5 – Subcutaneous Allo- and Xenograft studies
Compound
Number
(2)
Dose (mg/kg)
TGI in 3T3 (%)
TGI in GEO (%)
TGI in RD-ES (%)
qd, p.o.
100
75
59
71
-
69
-
-
103
63
-
98
-
-
-
-
-
82
-
(4)
(4)
(17)
(17)
25
75
50
75
(20)
52
Table 5 Footnote:
Treatment of mice bearing subcutaneous 3T3, GEO or RD-ES tumors. The highest tested dose is
defined as maximum tolerated dose (MTD) of a compound and was defined in a two weeks
tolerability study prior to the efficacy experiment. Compound 4, which was considered as suitable
tool compound and compound 17 were tested in more than one Allo- or xenograft model.

MedChemComm
Page 10 of 14
View Article Online
Compound 17 was tested in a kinase panel to determine its selectivity profile (see tableD6O)I.: O10n.10ly396/Co5fMD00097A
the kinases in panel showed >50% inhibition at 1M and for all of these a >100 fold selectivity
window for IGF1R relative to the kinase in question was obtained.
Table 6 – Kinase panel data for compound 17
Kinase
ABL1
ACVR1B
AKT2
IC50 [nM]
227
-
-
AMPKA1B1G1
CDK2
462
-
-
CHEK1
CSNK1A1
CSNK1A2
EGFR
-
-
-
-
EPHB2
FGFR1
FRAP1 (MTOR)
GSK3B
244
-
-
LCK
142
-
-
-
-
-
-
MAP2K1
MAP3K8
MAPK14
MAPKAPK2
MYLK2
NEK2
PAK4
PDK1
-
-
PRKACA
RAF1
ROCK2
CAMK2D
SRPK2
-
-
-
150
-
SRK3
STK6
TBK1
-
117
-
Table 6 Footnote:
Compounds tested in panel of kinases; IC50s obtained where compound 17 showed greater than 50%
inhibition at 1M
Daily dosing of compound 17 with 50 mg/kg dose in the GEO model resulted almost in tumor stasis
with a TGI of 98% on day 24 (Table 5, Plot 2i). Compared to the control group the mice showed mild
diabetes related phenotypes (polyurea and polydipsia) with this dose starting from day 8 but this
was not connected with body weight loss (Plot 2ii). Efficacy was also observed in the IGF1R
dependent Ewing’s Sarcoma model (RD-ES, 82% TGI at the 50 mg/kg dose, table 5) and in the 3T3
allograft (69% TGI at the 75 mg/kg dose, table 5).
Plot 2 – GEO xenograft data for compounds 4 and 17

Page 11 of 14
MedChemComm
View Article Online
DOI: 10.1039/C5MD00097A
i)
Avg Tumor Volumes
1000
900
800
700
600
500
400
300
200
100
0
Natrosol 0.5 %
compound 17 (50 mg/kg)
compound 4 (25 mg/kg)
0
5
10
15
20
25
30
Day
ii)

MedChemComm
Page 12 of 14
View Article Online
DOI: 10.1039/C5MD00097A
Average Body Weight Changes
20
15
10
5
Natrosol 0.5 %
compound 17 (50 mg/kg)"
compound 4 (25 mg/kg)"
0
-5
-10
-15
-20
0
5
10
15
20
25
Day
Plot 2 Footnote:
GEO tumor bearing mice were daily treated (p.o.) for 25 days with either 25 mg/kg of compound 4
or 50 mg of compound 17 (i). Both doses were well tolerated as visible by the body weight curves (ii).
In order to confirm that the in vivo anti-tumor efficacy of the IGF1R inhibitors was related to target
modulation, the effect of an efficacious dose of compound 17 on IGF1R phosphorylation was
investigated. GEO tumor-bearing mice were treated once with vehicle or compound 17 (50 mg/kg,
p.o.) and the levels of phospho-IGF1R (pIGF1R) in tumors were quantified. Compound 17 led to a
statistically significant (p=0.0005) reduction (76% of control) of the pIGF1R signal in GEO tumors at 6
hours after dosing, confirming effective target modulation at this dose level (see figure 5).
Figure 5 – pIGF1R tumor level data for Compound 17

Page 13 of 14
MedChemComm
View Article Online
DOI: 10.1039/C5MD00097A
140
120
100
80
60
40
20
0
Vehicle
Compound 17
50 mg/kg
Figure 5 Footnote:
Mice (n=4) bearing GEO tumors were dosed with compound 17 (50 mg/kg po) and sacrificed after 6
hours. Tumors were excised and pIGF1R levels determined and compared to vehicle control.
Conclusion
During our optimization campaign we were able to identify modifications which allowed for the
simultaneous improvement of cellular potency and reduction in hERG inhibition. This was enabled
by the application of a systematic matched molecular pairs approach coupled with fine tuning of
physicochemical properties (principally logD) and the use of a hERG homology model to identify the
optimal decoration of the pyrimidine scaffold. The outcome was the identification of compound 17
which met our objectives of having an improved CV safety profile (unbound Cmax to hERG ratio >
100 fold) combined with a cell potency in the nanomolar range. Compound 17 had a good selectivity
profile with respect to a panel of kinases and was shown to modulate pIGF levels in tumour tissue
which was consistent with the high levels of efficacy achieved in several IGF1R responsive xenograft
models.
Acknowledgements
Patricia Amouzegh, Jens Bruchhaus, Sarah Ellam, Thomas Fett, Edward Glenn, Thomas Karner, Stefan
Kornigg, Shi Li, Sophie Mitzner, Gurubaran Raju Guido Scholz, Patrick Werni, who made valuable
contributions to the synthesis of the compounds.
Thomas Gerstberger, Reiner Meyer, Andreas Schrenk for measuring enzyme and cell activity data.
Stefan Fischer, Astrid Jeschko for performing the in vivo studies.
Christian Salamon for performing PK bioanalytics.

MedChemComm
Page 14 of 14
View Article Online
DOI: 10.1039/C5MD00097A
Notes and references
1. R. Baserga, Expert Opin. Ther. Targets, 2005, 9, 753-768
2. R. Baserga, F. Peruzzi, K. Reiss, Int. J. Cancer, 2003, 107, 873-877
3. P.J. Beltran, P. Mitchell, Y. A. Chung, E. Cajulis, J. Lu, B. Belmontes, J. Ho, M. M. Tsai, M. Zhu,
S. Vonderfecht, R. Baserga, R. Kendall, R. Radinsky, F. J. Calzone, Mol. Cancer Ther., 2009, 8,
1095-1105
4. K. Friedbichler, M. H. Hofmann, M. Kroez, E. Ostermann, H. R. Lamche, C. Koessl, E. Borges,
M. N. Pollak, G. Adolf , P. G. Adam, Mol. Cancer Ther., 2014, 13, 399-409
5. (a) J. M. Carboni, M. Wittman, Z. Yang, F. Lee, A. Greer, W. Hurlburt, S. Hillerman, C. Cao, G.
H. Cantor, J. Dell-John, C. Chen, L. Discenza, K. Menard, A. Li, G. Trainor , D. Vyas, R. Kramer,
R. M. Attar, M. M. Gottardis, Mol. Cancer Ther. 2009, 8, 3341-3349; (b) M. Jin, J. Wang, E.
Buck, M. J. Mulvihill, Future Med. Chem., 2012, 4, 315-328; (c) A. Negi, P. Ramana, R. Kumar,
Mini Rev. Med. Chem., 2013, 13, 653-681.
6. (a) D. Yee, J. Natl. Cancer Inst., 2012, 104, 975-981; (b) R. Osborne , Nat. Biotechnol., 2008,
26, 719-720
7. E. Buck, P. C. Gokhale, S. Koujak, E. Brown, A. Eyzaguirre, N. Tao, M. Rosenfeld-Franklin, L.
Lerner, M. I. Chiu, R. Wild, D. Epstein, J. A. Pachter, M. R. Miglarese, Mol. Cancer Ther., 2010,
9, 2652-2664
8. (a) E. Buck, M. Mulvihill, Expert Opin. Investig. Drugs, 2011, 20, 605-621; (b) Q. Ji, M. J.
Mulvihill, M. Rosenfeld-Franklin, A. Cooke, L. Feng, G. Mak, M. O’Connor, Y. Yao, C. Pirritt, E.
Buck, A. Eyzaguirre, L. D. Arnold, N. W. Gibson, J. A. Pachter, Mol Cancer Ther., 2007, 6,
2158-2167
9. WO2005023780
10. (a) A. M. Aronov, Drug Discovery Today, 2005, 10, 149-155; (b) A. M. Aronov, Current Topics
in Medicinal Chemistry, 2008, 8, 1113-1127
11. F. Osterberg, J. Aqvist, FEBS Lett., 2005, 579, 2939-2944
12. A. G. Dossetter, E. J. Griffen, A. G. Leach, Drug Discovery Today, 2013, 18, 724-731
Supporting Information
Details of synthetic procedures, in vitro assays and in vivo experiments.