Discovery of wtRET and V804MRET Inhibitors: From Hit to Lead

Article abstract of DOI:10.1002/cmdc.201700243

Oncogenic activation of RET kinase has been found in several neoplastic diseases, like medullary thyroid carcinoma, multiple endocrine neoplasia, papillary thyroid carcinoma, and non-small-cell lung cancer. Currently approved RET inhibitors were not originally designed to be RET inhibitors, and their potency against RET kinase has not been optimized. Hence, novel compounds able to inhibit both wild-type RET (^wtRET) and its mutants (e.g., ^V804MRET) are needed. Herein we present the development and the preliminary evaluation of a new sub-micromolar ^wtRET/^V804MRET inhibitor, N-(2-fluoro-5-trifluoromethylphenyl)-N′-{4′-[(2′′-benzamido)pyridin-4′′-ylamino]phenyl}urea (69), endowed with a 4-anilinopyridine structure, starting from our previously identified 4-anilinopyrimidine hit compound. Profiling against a panel of kinases indicated 69 as a multi cKIT/^wtRET/^V804MRET inhibitor.

Full text of DOI:10.1002/cmdc.201700243

Accepted Article
Title: Discovery of novel wtRET and V804MRET inhibitors: from hit to
lead
Authors: luca mologni, martina dalla via, adriana chilin, manlio
palumbo, and Giovanni Marzaro
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemMedChem 10.1002/cmdc.201700243
Link to VoR: http://dx.doi.org/10.1002/cmdc.201700243
A Journal of
www.chemmedchem.org

10.1002/cmdc.201700243
ChemMedChem
Discovery of novel wtRET and V804MRET inhibitors: from hit to lead.
Luca Mologni,[a]‡ Martina Dalla Via,[b]‡, Adriana Chilin,[b] Manlio Palumbo,[b] and Giovanni
Marzaro*[b]
[a]Dr. L. Mologni
School of Medicine and Surgery, University of Milano-Bicocca, via Cadore 48, 20900 Monza
(Italy)
[b]Dr. M. Dalla Via, Prof. A. Chilin, Prof. M. Palumbo, Dr. G. Marzaro
Department of Pharmaceutical and Pharmacological Sciences, University of Padova, via Marzolo
5, 35131 Padova (Italy).
e-mail: giovanni.marzaro@unipd.it
‡First two authors contributed equally to this work and should be considered as co-first authors.
Abstract
Oncogenic activation of the RET kinase has been found in several neoplastic diseases, like
medullary thyroid carcinoma, multiple endocrine neoplasia, papillary thyroid carcinoma and non-
small cells lung cancer. Currently approved RET inhibitors were not originally designed to be RET
inhibitors and their potency against RET kinase has not been optimized. Hence, novel compounds
able to inhibit both wtRET and its mutants (e.g. V804MRET) are needed. Herein we present the
development and the preliminary evaluation of a new sub-micromolar wtRET/V804MRET inhibitor (69)
endowed with 4-anilinopyridine structure, starting from our previously identified 4-anilinopyrimidine
hit compound. Profiling against a panel of kinases indicated 69 as a multi cKIT/wtRET/ V804MRET
inhibitor.
Graphical Abstract
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201700243
ChemMedChem
Introduction
Protein kinases (PKs) are among the most investigated targets for cancer therapy.[1] This class of
enzymes regulates almost all cell events, by transferring a phosphate group from ATP to the target
protein substrate.[2, 3] The human kinome comprises 518 members, grouped in families, subfamilies
and classes, according to structure and function similarities.[3] Of particular interest is the family of
the tyrosine kinases (TKs): the deregulation, through mutations or over-expression of TKs, is often
related to cancer onset and progression[4] and in the last two decades a number of ATP-mimic TKs
inhibitors (TKIs) have been developed.[5] TKIs share some common pharmacophore features, like
a nitrogen containing heterocycle (for the interaction with the kinase hinge region), a lipophilic
moiety (for the interaction with a hydrophobic region not exploited by ATP) and a bridge connecting
the two.[1, 6] According to the bound kinase conformation, TKIs are classified in type I, I½, II, III and
IV inhibitors. In particular, type II inhibitors bind the inactive kinase conformation, that bears a
larger lipophilic cleft than the active conformation. Accordingly, type II inhibitors shows a larger
lipophilic moiety than type I inhibitors.[7]
As part of our kinase inhibitors discovery project,[8, 9] we have recently reported that some N-
phenyl-N’-[4-(pyrimidin-4-ylamino)phenyl]urea derivatives behave as subfamily selective class III
multi-target kinase inhibitors.[9] To further investigate the usefulness of these compounds in
developing novel TKIs, we tested compound 1 (selected as hit compound in our previous study;
Figure 1) against some kinases phylogenetically related to the class III receptor TKs, i.e. the FGFR
family and the kinase RET. From this very preliminary screening we found that 1 possessed a low
but not negligible inhibitory activity against RET (IC50 = 11.9 µM).
Figure 1. Structure of hit compound 1.
The RET proto-oncogene has been recognized as an important therapeutic target in thyroid
neoplasia and in a small subset of non-small cell lung cancer (NSCLC) patients.[10, 11] Oncogenic
activation of RET occurs through either of two mechanisms: point mutations in the extracellular or
the kinase domain of RET are frequently found in medullary thyroid carcinoma (MTC) and in
multiple endocrine neoplasia, while chromosomal rearrangements involving the catalytic domain of
RET are predominant in papillary thyroid carcinoma (PTC) and in NSCLC.[10] In both cases,
aberrant RET-driven signaling ensues, leading to increased cell proliferation and survival. Thus,
actual inhibition of RET kinase is paramount to an effective treatment of these diseases. Four
drugs (vandetanib, cabozantinib, sorafenib and lenvatinib) have been recently approved for the
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201700243
ChemMedChem
treatment of thyroid cancer.[12] However, these compounds were not originally designed to be RET
inhibitors. A few more compounds are under investigation[12] including a potent inhibitor of the
V804M mutant.[13] Different strategies have been pursued for the identification of novel RET
inhibitors, as drug repurposing,[14] kinase profiling,[15] use of multicomponent reaction libraries.[16] In
this work we report our efforts in the development of novel analogs of compound 1 endowed with
sub-micromolar RET inhibitory potency. According to our previous results,[9] we hypothesized that
1 would bind the inactive kinase conformation. Since no crystallographic structure has been to date
deposited in the Protein Data Bank for the inactive RET conformation, we decided to face a
classical medicinal chemistry approach rather than a structure-based approach. However, in order
to suggest a plausible binding mode supported by the experimental evidences, molecular docking
study based on homology modelling is also presented.
Results and Discussion
Optimization of urea moiety
The first optimization step on hit compound 1 regarded the evaluation of different substituents on
both the phenyl rings in the urea function (Table 1).
Table 1. Structure and IC50 (isolated wtRET) of 1-12.
IC50
ID
R1
R2
(mean±SEM, µM)
11.9±0.7
3.3±0.5
4.8±0.2
6.1±0.7
4.1±0.1
6.2±0.5
1
2
H
H
H
3-Me
3
H
3-CF3
4
H
2-F
5
H
H
2-F,5-CF3
2,5-diF
6
7
H
2-F,5-Me
3-OMe
7.1±0.7
19.8±0.8
>100
8
H
9
H
4-Cl,3-CF3
2-F,5-CF3
2-F,5-CF3
2-F,5-CF3
10
11
12
2’-OMe
2’-Br
3’-OMe
10.7±0.7
42.3±5.5
>100
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201700243
ChemMedChem
The diphenyl urea motif is present in several known type II kinase inhibitors and commonly
interacts with the lipophilic cleft typical of tyrosine kinases.[6] Hence, we decided to functionalize the
terminal phenyl ring with lipophilic substituents, differing from the steric and the electronic features.
In particular, we focused our attention on substituents already evaluated for the development of
kinase inhibitors, as for example the 2-fluoro-5-methyl,[17] the 2-fluoro-5-trifluoromethyl,[18] the 4-
chloro-3-trifluoromethyl[19] or the 2,5-difluoro[20] substituents. The central aniline ring was
functionalized with bulky functions in order to obtain indirect information about the size of the
hydrophobic cleft in inactive RET conformation. Inhibitory potency against recombinant wtRET
kinase (expressed as IC50, µM) allowed for the derivation of useful SARs. The functionalization of
the terminal ring was mandatory for enhancing the activity. In particular, the presence of a methyl
or a trifluoromethyl function at 5(3) position led to the most potent compounds in the series (see 2,
3 and 5). A 2-fluoro function was useful in the absence of any other substituent (compare 4 with 1)
or when a 5-trifluoromethyl function was also present (compare 5 with 3). Conversely, when a 5-
methyl function was present, the 2-fluoro substituent did not improve the activity (compare 7 with
3). Bulky substituents were not well tolerated (see 8). The presence of a 4-chloro-3-trifluoromethyl
function led to completely inactive compound. The functionalization of the central aniline ring was
detrimental for the activity (see 10-12). On the basis of the inhibitory data and on the synthetic
accessibility, compound 5 (bearing the 2-fluoro-5-trifluoromethyl pattern in the terminal phenyl) was
chosen as model for further improvement. Compounds listed in Table 1 were synthesized
according to Scheme 1.
Scheme 1. Synthesis of compounds 2-12: a) phenylboronic acid, Pd(PPh3)4, Na2CO3, H2O, DME,
reflux, 16 h; b) aniline derivative, iPrOH, MW, 80°C, 3x20 min; c) phenylisocyanate derivative,
CH2Cl2, RT, 16 h; d) NaOH, H2O, RT, 16 h; e) iron, acetic acid, EtOH, H2O, reflux, 2h.
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201700243
ChemMedChem
Briefly, 4,6-dichloropyrimidine was first condensed with phenylboronic acid and then with a slight
excess of the appropriate substituted aniline. Compounds 16 and 17 were submitted to alkaline
hydrolysis, whereas compound 20 was reduced with elemental iron in an aqueous/ethanolic
solution, made slightly acid with acetic acid. All the amino derivatives were finally condensed with
the appropriate isocyanate derivatives in DCM at room temperature affording the desired ureas in
moderate to good yields.
Optimization of nitrogen heterocycle
The second optimization step regarded the evaluation of different substituents at the 6’’ position of
the pyrimidine ring (Table 2).
Table 2. Structure and IC50 (isolated wtRET) of 22-26.
IC50
ID
R
(mean±SEM, µM)
5
4.1±0.1
22
23
3.8±0.2
5.2±0.1
24
36.0±2.4
25
26
>100
1.2±0.3
Here we mainly chose to replace the phenyl ring with simple heterocycles (i.e. furan and
thiophene, to verify the effect of electron density in this part of the moiety) or to insert a spacer
(potentially able to establish hydrogen bonds) between the aryl and the pyrimidine moiety. The
substitution of the 6’’-phenyl ring with heteroaryls retained similar potency (compare 22 and 23 with
5). Conversely to what previously observed for the inhibition of class III RTK members[9] the
introduction of a polar function at the 6’’-phenyl reduced the compound potency (see 24). The
highest effect was observed when a spacer was introduced between the pyrimidine and the aryl:
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201700243
ChemMedChem
the NH function led to totally inactive compound (25), whereas a carboxamido function (26)
furnished the overall most active compound.
Compounds listed in Table 2 were synthesized according to Scheme 2.
Scheme 2. Synthesis of compounds 22-26: a) arylboronic acid derivative, Pd(PPh3)4, Na2CO3,
H2O, DME, reflux; b) p-phenylenediamine, iPrOH, MW, 80°C, 3x20 min; c) 2-fluoro-5-
trifluoromethylphenylisocyanate, CH2Cl2, RT, 16 h; d) iron, AcOH, EtOH, H2O, reflux, 2 h; e)
chloroacetyl chloride, TEA, DME, RT, 16 h; f) N-methylpiperazine, KI, TEA, DMF, RT, 48 h; g)
aniline, TEA, iPrOH, MW, 150°C, 30 min; h) NH4OH, iPrOH, 80°C (sealed tube), 16 h; i) benzoyl
chloride, DMAP, MW, 100°C, 5 min.
Compounds 22 and 23 were synthesized according to the same synthetic strategy as 5 (see
Scheme 1), with the only difference being the nature of arylboronic acid used in the first step.
Compounds 24 and 25 were synthesized following our previously reported strategy.[9] For the
synthesis of compound 26, the starting dichloropyrimidine was first converted to the 6-amino-4-
chloropyrimdine. Intermediate 38 was then condensed with benzoyl chloride in presence of DMAP,
taking advantage from the microwave assisted organic synthesis technique. Compound 39 was
then submitted to the usual synthetic steps for obtaining the desired urea derivative.
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201700243
ChemMedChem
In parallel to the evaluation of different substituents at the 6’’ position of the pyrimidine ring we also
evaluated different pyrimidines, in terms of both the relative position of nitrogen atoms with respect
to the phenyl ring and of the presence of substituents (Table 3).
Table 3. Structure and IC50 (isolated wtRET) of 41-45.
IC50
ID
X
Y
R
(mean±SEM, µM)
4.1±0.1
>100
5
CH
C-CN
CH
CH
N
N
N
H
H
41
42
43
44
45
N
Me
S-Me
H
>100
N
>100
CH
C-CN
1.5±0.1
>100
N
H
As aforementioned, we decided to hinder the pyrimidine nucleous with different substituents (i.e.,
CN, Me, SMe) in order to explore the steric features of the ATP binding pocket of RET. Besides,
we modified the number and the relative position of the nitrogen atoms in the heterocycle in order
to investigate the importance the of the H-bond acceptor features of the scaffold. The addition of a
further substituent in the pyrimidine ring led to absolutely inactive compounds (see 41-43 and 45).
Conversely, the change of the pyrimidine ring geometry from the 4-anilino-6-phenyl to the 4-
anilino-2-phenyl led to an interesting activity improvement (compare 44 with 5). At this stage, a
preliminary cytotoxicity test was run on RET+ thyroid carcinoma cells at a fixed compound
concentration (10 μM), using the molecules that showed activity on the recombinant enzyme. The
analysis confirmed 22, 26 and 44 as the most active compounds, showing on average >75% cell
growth inhibition (Figure S2 in Supplementary Information). Compounds listed in Table 3 were
synthesized according to Schemes 3-5.
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201700243
ChemMedChem
Scheme 3. Synthesis of compound 41: a) ethyl cyanoacetate, formamidine acetate, K2CO3, EtOH,
reflux, 16 h; b) POCl3, reflux, 2 h; c) p-phenylenediamine, iPrOH, 80°C, 20 min; d) 2-fluoro-5-
trifluoromethylphenylisocyanate, CH2Cl2, RT, 16 h.
Compound 41 was synthesized starting from benzaldehyde that was condensed in a single-pot
reaction with ethyl cyanoacetate and formamidine, yielding the hydroxypyrimidne 47 (Scheme 3).
Then, usual chlorination, condensation with p-henylenediamine and with isocyanate furnished the
desired final product.
Scheme 4. Synthesis of compounds 42 (A) and 43 (B): (A) a) HCl 3M, nBuOH, MW, 125°C, 10
min; b) phenylboronic acid, Pd(PPh3)4, K2CO3, DME, H2O, MW, 130°C, 5 min; c) POCl3, 80°C, 4 h;
d) p-phenylenediamine, iPrOH, MW, 150°C, 20 min; e) 2-fluoro-5-trifluoromethylphenylisocyanate,
CH2Cl2, RT, 16 h. (B) a) NaOEt 20% in EtOH, RT, 20 min, then thiourea, reflux, 24 h; b) MeI,
NaOH, H2O, RT, 16 h; c) POCl3, 80°C, 4 h; d) p-phenylenediamine, iPrOH, MW, 150°C, 20 min; e)
2-fluoro-5-trifluoromethylphenylisocyanate, CH2Cl2, RT, 16 h.
The 2-substituted pyrimidine derivatives were prepared in two different ways. Compound 42
(Scheme 4A) was prepared starting from dichloropyrimidine 50 that was selectively mono-
hydrolized to compound 51 before being submitted to Suzuki reaction and subsequent chlorination.
Unfortunately, any attempt to perform the Suzuki reaction on 50 to directly obtain compound 53
was unsuccessful. For compound 43 (Scheme 4B) the thiopyrimidine ring was obtained by treating
ethyl benzoylacetate with thiourea. Compounds 44 and 45 were synthesized starting from
benzamidine 60 that was condensed with either ethyl propiolate or with 2-cyano-3-ethoxyacrylate
(Scheme 5).
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201700243
ChemMedChem
Scheme 5. Synthesis of compounds 44 and 45: a) ethyl propiolate, K2CO3, EtOH, reflux, 16 h; b)
POCl3, 80°C, 4 h; c) p-phenylenediamine, iPrOH, MW, 80°C, 40 min; d) 2-fluoro-5-
trifluoromethylphenylisocyanate, CH2Cl2, RT, 16 h; e) ethyl (ethoxymethylene)cyanoacetate, DMF,
reflux, 96 h.
In all the cases, the hydroxypyrimidine intermediates (i.e. 52, 57, 61 and 64) were chlorinated and
submitted to the usual condensation with p-phenylenediamine and then with the isocyanate.
Development of lead compound
The acquired biological data were then used to design three novel compounds bearing the features
of the most active derivatives (Table 4).
Table 4. Structure and IC50 (isolated wtRET) of 67-69.
IC50
ID
X
Y
R
(mean±SEM, µM)
4.1±0.1
5
CH
CH
N
N
N
26
44
68
1.2±0.3
1.5±0.1
31.8±1.3
CH
CH
N
69
CH
CH
0.24±0.01
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201700243
ChemMedChem
The optimization steps previously described suggested the importance of a benzamido substituent
at the pyrimidine ring (see compound 26) and of the nitrogen heterocycle geometry (see compound
44). Starting from these observations, two novel compounds were finally designed and synthesized
that shared two common features: the presence of the benzamido function and the absence of a
nitrogen atom at the “Y” position of the nitrogen heterocycle (see Table 4 for structure clarification).
Indeed, compound 68 was a 4-anilino-2-substituted pyrimidine, whereas compound 69 was a 4-
anilinopyridine. Almost surprisingly compound 68 showed poor activity, although bearing the
structural features of the two most active compounds identified during our work. On the contrary,
the replacement of the pyrimidine ring with a pyridine one furnished the first sub-micromolar RET
inhibitor in our series. Compound 69 was profiled against a panel of selected kinases selected
within the ScanEDGE subset of DiscoverX (http://www.discoverx.com/services/drug-discovery-
development-services/kinase-profiling/kinomescan/scanedge). In particular, we tested our lead
compound at a single concentration (1 µM) against the target kinases of the starting hit compound
1 (i.e.FLT3, cKIT and PDGFRβ) and against the most common targets of other RET inhibitors (i.e.
Abl1, BRAF, ERK1, MET, Src and VEGFR2) using the KinomeScanTM platform. The KinomeScan
technology measures the ability of the tested compound to displace the target kinase from a high
affinity ATP-mimic probe. The result of the assay is expressed as “Percent of Control” (POC)
value, that is statistically correlated with the dissociation constant. The lower the POC, the higher
the affinity of the test compound for the target kinase. For a more exhaustive description of POC
and of Kinomescan technology see https://www.discoverx.com/tools-resources/leadhunter-study-
reports-data-analysis. The results are listed in Table 5.
Table 5. Selectivity profile of compound 69
Kinase
Abl1
POC (1 µM)
12
BRAF
ERK1
FLT3
67
100
1.5
<0.5
76
cKIT
MET
PDGFRβ
RET
0.95
<0.5
28
Src
VEGFR2
8
Taken together, our results indicated 69 as a multi cKIT/wtRET/V804MRET inhibitor. Notably, hit
compound 1 was previously identified as a dual cKIT/PDGFRβ inhibitor (POC values <0.5 against
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201700243
ChemMedChem
both kinases under the same experimental conditions).[9] Hence, our efforts to improve the activity
against RET kinases led also to a slightly lower potency against PDGFRβ. Compounds listed in
Table 4 were synthesized according to Scheme 6, starting from the commercially available 2-
amino-4-chloro- pyrimidine and pyridine.
Scheme 6. Synthesis of compounds 67-68 (A) and 69 (B): a) p-phenylenediamine, iPrOH, MW,
80°C, 40 min; b) 2-fluoro-5-trifluoromethylphenylisocyanate, CH2Cl2, RT, 16 h; c) benzoic acid,
PyBOP, TEA, DMF, RT, 30 min; d) benzoyl chloride, DMAP, MeCN, MW, 100°C, 5 min.
Compound 70 was condensed with p-phenylenediamine and then with the substituted isocyanate.
The last synthetic step was the PyBOP mediated condensation of the free amino function with
benzoic acid.For the synthesis of compound 69, conversely, the formation of the benzamido
function was obtained as the first synthetic step.
In vitro testing
The most active compounds (22, 26, 44, 69) were then tested against the recombinant V804MRET
mutant, and in cellular assays against selected RET+ and RET- cell lines (Table 6). Vandetanib
was used for comparison purposes. All compounds (with the only exception of 22 bearing a 6’’-furyl
substituent) were more potent against the mutant than the wild type RET kinase. The novel RET
inhibitors showed promising cytotoxic effects on RET+ cancer cell lines, of both thyroid and lung
origin. Interestingly, three of four tested compounds showed selective cytotoxicity in RET+ versus
RET- thyroid carcinoma cells. In particular, activity and selectivity of compound 69 was comparable
to that of vandetanib. We then analyzed RET autophosphorylation in the presence of 69, in TPC-1
cells harboring the CCDC6/RET fusion and in HEK293 ectopically expressing the V804M
mutation.[13] Compound 69 was able to inhibit RET activation at low micromolar concentrations, in
line with cell growth data (Figure 2). The inhibition of RET autophosphorylation has also been
determined in NIH3T3-PTC2 (see Figure S3).
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201700243
ChemMedChem
Table 6. Biological evaluation of selected compounds (IC50, µM).
* Selectivity index = Ratio of non-target (RET-neg, ARO and NPA) to target cell average IC50 values
TPC-1
[CCDC6/
RET]
Lc2/ad
[CCDC6/
RET]
TT
MZ-CRC-1
[
ARO
NPA
Selectivity
index*
ID
wtRET
3.8±0.2
V804MRET
60.4
[
C634WRET]
M918TRET]
15.8±1.4
4.6±0.5
[
V600EBRAF]
[
V600EBRAF]
22
8.5±0.8
6.3±0.8
2.6±0.02
6.3±0.4
8.6±2.8
7.8±0.02
7.9±2.3
1
4
5
8
7
26
1.2±0.3
0.65±0.04
0.74±0.19
0.11±0.02
>10
3.6±0.3
5.2±0.03
1.9±0.3
0.46±0.01
3.7±0.1
0.62±0.01
4.1±0.1
44
69
1.5±0.1
12.1±0.2
1.3±0.06
1.2±0.3
37.8±13.7
7.0±0.6
20.5±3.9
7.2±0.02
7.0±0.9
0.24±0.01
0.10±0.01
0.26±0.01
0.86±0.03
0.16±0.01
0.22±0.02
Vandetanib
0.41±0.04
2.6±0.1
Figure 2. Western blot analysis of RET phosphorylation inhibition by compound 69, at the indicated micromolar concentrations, in TPC-1 (A)
and HEK-RETV804M cells (B). Total RET is shown for loading control.
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201700243
ChemMedChem
Molecular docking simulations
Preliminary molecular modelling studies were used to propose the binding mode for compound 69
in both wt and mutant RET. As mentioned above, no crystallographic structure has been to date
deposited in the Protein Data Bank for the inactive RET conformation. Hence homology models for
both wtRET and V804MRET kinases were computed using Swiss-Model.[21] The PDB:3WZE
coordinates for VEGFR2 in complex with sorafenib[22] were used as a template. The comparison
between the active and inactive conformation for RET are reported in Figure S4.
AutoDock Vina was used to predict the binding modes for compound 69 with wtRET (Figure 3A-C)
and V804MRET (Figure 3D-F).
Figure 3. A) Predicted binding mode for 69 and wtRET. B) Details of the wtRET amino acids
wt
surrounding 69. C) Schematic representation of the interactions between 69 and RET. H-bonds
are reported as dashed red lines; arene-arene interaction is reported as dashed green line;
hydrophobic interactions are reported as black solid curved lines. D) Predicted binding mode for 69
and V804MRET. E) Details of the V804MRET amino acids surrounding 69. F) Schematic representation
of the interactions between 69 and V804MRET; H-bonds are reported as dashed red lines; arene-
arene interaction is reported as dashed green line; hydrophobic interactions are reported as black
solid curved lines; sulfur-arene interaction is reported as dashed blue line.
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201700243
ChemMedChem
Compound 69 was predicted to establish two H-bonds with the A807 in the hinge region. In
particular, the H-bond between the carboxamide NH of 69 with the carboxylic function of A807
could justify the importance of an amidic substituent. The urea moiety was supposed to interact
both with the carboxylate of the conserved E775 from the αC helix and with the backbone amide of
D892 (that belongs to the DFG motif). The terminal substituted ring fits well in the hydrophobic
pocket, establishing lipophilic interactions with several amino acids. In the case of RET, the hinge
region and the gatekeeper residue are highly hydrophobic. Hence, we concluded that 69 was more
active than 26 due to the presence of a less hydrophilic moiety (the pyridine in place of a
pyrimidine). Finally, as hypothesized in our previous work with ponatinib, the observed activity of
69 against the V804M mutant may be due to the ability to induce the DFG-out conformation of the
kinase: in this case the compound would fit in the active site without clashing with the bulky mutant
gatekeeper.[13] Moreover, the slightly improved activity of the compound against the V804M mutant
compared to wild-type enzyme is possibly due to the ability of 69 to establish a sulfur-arene
interaction with the mutated gatekeeper. With respect to the nitrogen heterocycle, molecular
docking simulations failed in suggesting the reason for the higher activity of the pyridine derivative
69 with respect to the pyrimidine analogs 68. Docking simulations of 68 in RET furnished
absolutely comparable results than with 69, in terms either of binding pose and Vina Score (data
not shown). Hence, our findings constitute an example on how sometimes a classical medicinal
chemistry approach can still offer new (poorly predictable) insights into protein-ligand binding.
Conclusions
Starting from our previously identified hit compound 1 (dual cKIT/PDGFRβ inhibitor) we have
synthesized and tested several novel analogs with the aim to identify novel promising RET
inhibitors. Taken together our data suggest that the 4-anilinopyridine scaffold rather than the 4-
anilinopyrimidine one can be used for the development of RET inhibitors. Molecular modeling
simulations were not useful for justifying this observation, suggesting the importance of facing a
classical medicinal chemistry in hit-to-lead and lead-optimization processes. The compounds
reported herein showed a slight preference toward the V804MRET mutant, making them attractive
lead compounds. In particular, compound 69 (multi cKIT/wtRET/V804MRET inhibitor) can be
considered as the most promising lead in the series for further derivatization, to obtain new potent
inhibitors of RET tyrosine kinase, for the specific treatment of RET+ cancer. Future work
addressing its efficacy and toxicity in vivo will be needed.
Experimental Section
See the Supporting Information for the synthesis and characterization of compounds.
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201700243
ChemMedChem
Biochemical and cellular assays
The recombinant RET catalytic domain (WT and V804M mutant, aa 700-1020) was expressed as a
[23]
His-tagged protein in Baculovirus and purified by affinity chromatography as described.
[24]
Inhibitors were tested for activity using an ELISA-based kinase assay as described previously.
All cell lines were purchased from the American Type Culture Collection and maintained in DMEM
medium (except Lc2/ad, which grow in Ham's F12 medium) supplemented with 10% fetal bovine
serum, 2 mmol/L of L-glutamine, 100 units/mL of penicillin-streptomycin, and 100 units/mL of
gentamicin. TT, TPC-1, and MZ-CRC-1 are thyroid carcinoma cell lines expressing oncogenic
RET;[13] ARO and NPA thyroid cancer cells are not dependent on RET activity, rather they express
the V600EBRAF oncogene;[25] Lc2/ad are a NSCLC cell line harboring the CCDC6/RET fusion.[26]
Cell growth assays were performed using MTS assays (CellTiter 96® AQueous One Solution Cell
Proliferation Assay, Promega) after 72 hours incubation of cells (10,000/well) with increasing
compounds concentrations, in 96-well plates. Dose-response curves were generated by GraphPad
Prism software, by plotting log inhibitor concentration versus normalized kinase activity, and IC50
values were calculated as the concentration causing half-maximal response relative to control.
Data were generated in triplicate and repeated at least in two independent experiments. Ponatinib
and vandetanib were purchased from Selleck Chemicals, dissolved in DMSO, aliquoted and stored
at -20°C.
Computational methodologies
Computational studies were carried out on a 4 CPU (Intel Core2 Quad CPU Q9550, 2.83ꢀGHz)
ACPI ×64 Linux workstation with Ubuntu 16.10 operating system. The tridimensional structure of
the template for homology modeling was downloaded from Protein Data Bank (PDB ID: 3WZE).
wt
The sequences of the kinase domain of RET and of V804MRET were downloaded from Pubmed
wt
Protein (https://www.ncbi.nlm.nih.gov/protein). The RET and V804MRET homology models were
prepared with SwissModel Automatic Modelling Mode (swissmodel.expasy.org)[21] as previously
reported.[16]
The structure of compound 69 was prepared with MarvinSketch 5.5.0.1 software
(www.chemaxon.com/products). The lowest energy conformations and the degree of protonation at
pH 7.4 were determined with OpenBabel software[27] using the MMFF94s force field. For all the
molecules, the appropriate “.pdbqt” files were prepared with the AutoDockTools graphical interface
of AutoDock 4 software.[28] All the docking studies have been performed with AutoDock Vina,[29]
using a docking box of 22.5ꢀ×ꢀ16.5ꢀ× 20.25ꢀÅ dimensions and exhaustiveness = 20. The lowest
energy conformation for both 69/ wtRET and 69/ V804MRET was finally analyzed.
Acknowledgments
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201700243
ChemMedChem
The present work has been carried out with the financial support of the University of Padova
‘Progetto Giovani Studiosi 2012’ to G.M. M.D.V. thanks the Fondazione Cariparo for a PhD student
grant.
Conflict of interest
The authors declare no conflict of interest
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201700243
ChemMedChem
References
[1]
[2]
J. Zhang, P. L. Yang, N. S. Gray, Nat Rev Cancer 2009, 9(1), 28-39.
D. R. Robinson, Y. M. Wu, S. F. Lin, Oncogene 2000, 19(49), 5548-5557.
[3]
G. Manning, D. B. Whyte, R. Martinez, T. Hunter, S. Sudarsanam, Science 2002, 298(5600), 1912-
1934.
[4]
A. Arora, E. M. Scholar, J Pharmacol Exp Ther 2005, 315(3), 971-979; M. K. Paul, A. K.
Mukhopadhyay, Int J Med Sci 2004, 1(2), 101-115.
[5]
[6]
[7]
[8]
S. Gross, R. Rahal, N. Stransky, C. Lengauer, K. P. Hoeflich, J Clin Invest 2015, 125(5), 1780-1789.
Y. Liu, N. S. Gray, Nat Chem Biol 2006, 2(7), 358-364.
R. Roskoski, Pharmacol Res 2016, 103, 26-48.
A. Chilin, M. T. Conconi, G. Marzaro, A. Guiotto, L. Urbani, F. Tonus, P. Parnigotto, J Med Chem
2010, 53(4), 1862-1866; M. T. Conconi, G. Marzaro, L. Urbani, I. Zanusso, R. Di Liddo, I. Castagliuolo, P.
Brun, F. Tonus, A. Ferrarese, A. Guiotto, A. Chilin, Eur J Med Chem 2013, 67, 373-383; G. Marzaro, I.
Castagliuolo, G. Schirato, G. Palu', M. Dalla Via, A. Chilin, P. Brun, Eur J Med Chem 2016, 115, 416-425.
[9]
V. Gandin, A. Ferrarese, M. Dalla Via, C. Marzano, A. Chilin, G. Marzaro, Sci Rep 2015, 5, 16750.
I. Plaza-Menacho, L. Mologni, N. Q. McDonald, Cell Signal 2014, 26(8), 1743-1752.
M. Song, J Med Chem 2015, 58(9), 3672-3681.
[10]
[11]
[12]
99.
L. Mologni, C. Gambacorti-Passerini, P. Goekjian, L. Scapozza, Expert Opin Ther Pat 2017, 27(1), 91-
[13]
L. Mologni, S. Redaelli, A. Morandi, I. Plaza-Menacho, C. Gambacorti-Passerini, Mol Cell Endocrinol
2013, 377(1-2), 1-6.
[14]
L. M. Mulligan, Expert Rev Proteomics 2016, 13(7), 631-633; D. Viola, L. Valerio, E. Molinaro, L.
Agate, V. Bottici, A. Biagini, L. Lorusso, V. Cappagli, L. Pieruzzi, C. Giani, E. Sabini, P. Passannati, L. Puleo, A.
Matrone, B. Pontillo-Contillo, V. Battaglia, S. Mazzeo, P. Vitti, R. Elisei, Endocr Relat Cancer 2016, 23(4),
R185-205.
[15]
K. C. Duong-Ly, K. Devarajan, S. Liang, K. Y. Horiuchi, Y. Wang, H. Ma, J. R. Peterson, Cell Rep 2016,
14(4), 772-781.
[16]
[17]
B. Frett, M. Moccia, F. Carlomagno, M. Santoro, H. Y. Li, Eur J Med Chem 2014, 86, 714-723.
Y. Dai, K. Hartandi, Z. Ji, A. A. Ahmed, D. H. Albert, J. L. Bauch, J. J. Bouska, P. F. Bousquet, G. A.
Cunha, K. B. Glaser, C. M. Harris, D. Hickman, J. Guo, J. Li, P. A. Marcotte, K. C. Marsh, M. D. Moskey, R. L.
Martin, A. M. Olson, D. J. Osterling, L. J. Pease, N. B. Soni, K. D. Stewart, V. S. Stoll, P. Tapang, D. R. Reuter,
S. K. Davidsen, M. R. Michaelides, J Med Chem 2007, 50(7), 1584-1597.
[18]
Shibahara, H. Toyoda, L. Wang, R. T. Nolte, J. M. Veal, M. Cheung, J Med Chem 2007, 50(18), 4453-4470.
[19] S. Wilhelm, C. Carter, M. Lynch, T. Lowinger, J. Dumas, R. A. Smith, B. Schwartz, R. Simantov, S.
Kelley, Nat Rev Drug Discov 2006, 5(10), 835-844.
[20] K. Kubo, T. Shimizu, S. Ohyama, H. Murooka, A. Iwai, K. Nakamura, K. Hasegawa, Y. Kobayashi, N.
Takahashi, K. Takahashi, S. Kato, T. Izawa, T. Isoe, J Med Chem 2005, 48(5), 1359-1366.
M. Hasegawa, N. Nishigaki, Y. Washio, K. Kano, P. A. Harris, H. Sato, I. Mori, R. I. West, M.
[21]
[22]
K. Arnold, L. Bordoli, J. Kopp, T. Schwede, Bioinformatics 2006, 22(2), 195-201.
K. Okamoto, M. Ikemori-Kawada, A. Jestel, K. von König, Y. Funahashi, T. Matsushima, A. Tsuruoka,
A. Inoue, J. Matsui, ACS Med Chem Lett 2015, 6(1), 89-94.
[23] E. Sala, L. Mologni, S. Cazzaniga, E. Papinutto, C. Gambacorti-Passerini, Int J Biol Macromol 2006,
39(1-3), 60-65.
[24]
L. Mologni, E. Sala, B. Riva, L. Cesaro, S. Cazzaniga, S. Redaelli, O. Marin, N. Pasquato, A. Donella-
Deana, C. Gambacorti-Passerini, Protein Expr Purif 2005, 41(1), 177-185.
[25]
E. Sala, L. Mologni, S. Truffa, C. Gaetano, G. E. Bollag, C. Gambacorti-Passerini, Mol Cancer Res
2008, 6(5), 751-759.
[26]
D. Matsubara, Y. Kanai, S. Ishikawa, S. Ohara, T. Yoshimoto, T. Sakatani, S. Oguni, T. Tamura, H.
Kataoka, S. Endo, Y. Murakami, H. Aburatani, M. Fukayama, T. Niki, J Thorac Oncol 2012, 7(12), 1872-1876.
[27]
N. M. O'Boyle, M. Banck, C. A. James, C. Morley, T. Vandermeersch, G. R. Hutchison, J Cheminform
2011, 3, 33.
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201700243
ChemMedChem
[28]
Chem 2009, 30(16), 2785-2791.
[29] O. Trott, A. J. Olson, J Comput Chem 2010, 31(2), 455-461.
G. M. Morris, R. Huey, W. Lindstrom, M. F. Sanner, R. K. Belew, D. S. Goodsell, A. J. Olson, J Comput
This article is protected by copyright. All rights reserved.