Challenging clinically unresponsive medullary thyroid cancer: Discovery and pharmacological activity of novel RET inhibitors

Article abstract of DOI:10.1016/j.ejmech.2018.02.080

It is now known that “gain of function” mutations of RET (REarranged during Transfection) kinase are specific and key oncogenic events in the onset of thyroid gland cancers such as the Medullary Thyroid Carcinoma (MTC). Although a number of RET inhibitors exist and are capable of inhibiting RET variants, in which mutations are outside the enzyme active site, the majority becomes dramatically ineffective when mutations are within the protein active site (V804L and V804M). Pursuing a receptor-based virtual screening against the kinase domain of RET, we found that compound 5 is able to inhibit efficiently both wild type and V804L mutant RET. Compound 5 was able to significantly reduce proliferation of both commercially available TT cell lines and surgical thyroid tissues obtained from patients with MTC and displayed a suitable drug-like profile, thus standing out as a promising candidate for further development towards the treatment of clinically unresponsive MTC.

Full text of DOI:10.1016/j.ejmech.2018.02.080

Accepted Manuscript
Challenging clinically unresponsive medullary thyroid cancer: Discovery and
pharmacological activity of novel RET inhibitors
Valeria La Pietra, Stefania Sartini, Lorenzo Botta, Alessandro Antonelli, Silvia Martina
Ferrari, Poupak Fallahi, Alessio Moriconi, Vito Coviello, Luca Quattrini, Yi-Yu Ke,
Hsieh Hsing-Pang, Federico Da Settimo, Ettore Novellino, Concettina La Motta,
Luciana Marinelli
PII:
S0223-5234(18)30223-X
10.1016/j.ejmech.2018.02.080
EJMECH 10257
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 21 December 2017
Revised Date: 24 February 2018
Accepted Date: 26 February 2018
Please cite this article as: V. La Pietra, S. Sartini, L. Botta, A. Antonelli, S.M. Ferrari, P. Fallahi, A.
Moriconi, V. Coviello, L. Quattrini, Y.-Y. Ke, H. Hsing-Pang, F. Da Settimo, E. Novellino, C. La Motta, L.
Marinelli, Challenging clinically unresponsive medullary thyroid cancer: Discovery and pharmacological
activity of novel RET inhibitors, European Journal of Medicinal Chemistry (2018), doi: 10.1016/
j.ejmech.2018.02.080.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Challenging Clinically Unresponsive Medullary Thyroid Cancer: Discovery and
Pharmacological Activity of Novel RET Inhibitors.
Valeria La Pietra,1§ Stefania Sartini, 2§ Lorenzo Botta,3§ Alessandro Antonelli,4 Silvia Martina
Ferrari,4 Poupak Fallahi,4 Alessio Moriconi,5 Vito Coviello,2 Luca Quattrini,2 Yi-Yu Ke,6 Hsieh
Hsing-Pang,6 Federico Da Settimo,2 Ettore Novellino,1 Concettina La Motta,2* and
Luciana Marinelli1*
1Dipartimento di Farmacia, Università di Napoli "Federico II", Via D. Montesano 49, 80131 Napoli,
Italy.
2Dipartimento di Farmacia, Università di Pisa, Via Bonanno 6, 56126 Pisa, Italy.
3Biological and Ecological Sciences Department, University of Tuscia, Viterbo 01100, Italy.
4Dipartimento di Medicina Clinica e Sperimentale, Università di Pisa, Via Savi 10, 56126 Pisa,
Italy.
5Department of Discovery, Dompé SpA Research Center, 67100 L'Aquila, Italy.
6Institute of Biotechnology and Pharmaceutical Research, National Health Research Institutes, 35,
Keyan Road, Zhunan Town, Miaoli County 350, Taiwan.
Corresponding Author
* For L.M.: phone, 0039-328-6795645; E-mail, lmarinel@unina.it
* For C.L.M.: phone, 0039-050-2219593; Email, concettina.lamotta@unipi.it
HIGHLIGHTS
•
•
•
•
A tailored ligand-based filtering step generated an heterogeneous focused library
Twenty-two compounds selected from a VS campaign were found active against RET
Compound 5 efficiently inhibits the wild type and also the V804L mutant RET kinase
Compound 5 greatly reduces proliferation of Medullary thyroid carcinoma cells
1

ACCEPTED MANUSCRIPT
ABSTRACT.
It is now known that “gain of function” mutations of RET (REarranged during Transfection) kinase
are specific and key oncogenic events in the onset of thyroid gland cancers such as the Medullary
Thyroid Carcinoma (MTC). Although a number of RET inhibitors exist and are capable of
inhibiting RET variants, in which mutations are outside the enzyme active site, the majority
becomes dramatically ineffective when mutations are within the protein active site (V804L and
V804M). Pursuing a receptor-based virtual screening against the kinase domain of RET, we found
that compound 5 is able to inhibit efficiently both wild type and V804L mutant RET. Compound 5
was able to significantly reduce proliferation of both commercially available TT cell lines and
surgical thyroid tissues obtained from patients with MTC and displayed a suitable drug-like profile,
thus standing out as a promising candidate for further development towards the treatment of
clinically unresponsive MTC.
KEYWORDS. Virtual Screening, Docking, Focused library, RET Kinase, Medullary thyroid
carcinoma
1. Introduction
Medullary thyroid carcinoma (MTC) is a malignant endocrine tumor originating from parafollicular
calcitonin-producing C cells. Mainly sporadic (75% of MTC forms), it may be also present in an
inherited form (25% of cases) as the component of both autosomal-dominant Multiple Endocrine
Neoplasia type-2 syndromes and familial medullary thyroid carcinoma (FMTC) [1,2]. At present,
the treatment of choice for different forms of MTCs relies in their complete surgical resection.
However, disease can persist or recur, with local and distant metastases, and the limited
effectiveness of both external beam radiation and conventional cytotoxic agents suggests the need
of novel therapeutic strategies [3-6].
2

ACCEPTED MANUSCRIPT
Recent advances in the knowledge of pathogenic mechanisms leading to MTCs clearly identified
the proto-oncogene RET, encoding the homologous receptor tyrosine kinase RET, as the main actor
in MTCs development and outbreak. Actually, ‘gain of function’ mutations and rearrangement of
RET activate the kinase activity of the encoded receptor, thus providing mitogenic and survival
signals to calcitonin-producing C cells. As mutations in RET have been identified in about 98% of
inherited MTC cases, at the germ-line level, and in 30% to 50% of the sporadic cases, at the somatic
level, this protein represents a sound target for the molecular therapy of most people affected by
MTC [7-10]. Accordingly, a number of small heterocyclic compounds, able to block the kinase
activity of RET receptor, have been proposing as a viable therapeutic strategy to treat this
pathology. Among them, vandetanib (1, Chart 1) [11] and cabozantinib (2, Chart 1) [12] have been
recently approved by both FDA and EMA for the treatment of symptomatic or progressive,
advanced and metastatic MTC. Their efficacy is mainly due to the combined blockade of
functionally linked and relevant kinases, like RET, VEGFRs, EGFR for vandetanib and MET,
VEGFRs, RET, c-KIT for cabozantinib, which in turn lead to the concurrent blockade of pathways
active in both tumor parenchyma and stroma [13].
However, although survival gains associated with these inhibitors are clear, their use is not devoid
of mild to heavy toxicities, and treatment discontinuation is often required [6].
Moreover, in the case of vandetanib, the ability to inhibit all the RET variants harbouring mutations
outside the enzyme active site, like M918T, E768D, L790F, Y791F, A883F, dramatically fades out
when a mutation affects the protein active site, like in the V804L and V804M variants. However,
vandetanib resistance to V804 alteration is not surprising. Actually, from a structural point of view,
V804 in the RET backbone corresponds to the gate-keeper residues of different kinases like Abl,
PDGFR, c-KIT, and EGFR, and mutations at these residues are known to confer resistance to a
large number of inhibitors [14].
In fact, besides vandetanib, almost all the known RET inhibitors have been demonstrated to be
ineffective when the V804L mutation occurs [13]. Accordingly, novel compounds able to block the
3

ACCEPTED MANUSCRIPT
kinase activity of RET are urgently needed to strike effectively this protein in all its mutant forms,
especially the V804L and V804M variants which have been recently classified as those conferring a
high risk of poor clinical course and survival [15,16].
Br
F
NH
O
O
H
S
Cl
N
N
N
N
N
N
1, Vandetanib
Cl
5, DP01920
H
N
H
H
X
R2
N
N
N
N
O
O
O
N
F
O
O
R1
43a-n
2, Cabozantinib
Chart 1. Chemical Structures of RET Inhibitors
Herein, we present the discovery of novel RET inhibitors by means of a receptor-based virtual
screening campaign inclusive of a Receptor-Ligand-interaction-based filtering step that allowed us
to reach a hit rate of ~50%. Moreover, we succeeded in discovery a novel heterocyclic compound,
5-(4-chlorophenyl)-3-((4-chlorophenylthio)methyl)-1H-1,2,4-triazole, (DP01920, compound 5,
Chart 1), which proved to inhibit efficiently both wild type and V804L mutant RET and showed a
significant antitumor activity against both commercially available TT cell lines and primary cell
lines obtained from surgical thyroid tissues of patients with MTC. Investigated for its physical-
chemical and pharmacokinetic properties, compound 5 displayed a suitable drug-like profile, thus
standing out as a promising candidate to develop for the treatment of clinically unresponsive MTC.
Accordingly, a number of parent compounds of 5 were designed and synthesized, characterized by
4

ACCEPTED MANUSCRIPT
different linker moieties connecting its main liphophilic areas and bearing distinctive substitution
patterns in both positions 3 and 5 of the heterocyclic core. Functional evaluation of the novel
derivatives helped to sketch out the key structural fragments of RET molecular recognition, which
were corroborated and rationalized through molecular docking simulations.
2. Results and Discussion
Building up the Filter for the Virtual Screening. From a structural point of view, RET kinase
domain consists of two lobes, N and C, connected through a hinge region. The cleft between the
lobes forms the highly flexible active site where the ATP binds. As happens for other kinases, in the
RET hinge region, two H-bond acceptors (herein referred as “UP” and “DOWN”, and a donor
(“MIDDLE”) are present (Figure 1A). As highlighted by Dubinina and co-workers [17], known
tyrosine kinase inhibitors are able to form at least one H-bond with the MIDDLE spot with the
majority of them (87% of the kinase inhibitor database from MDDR) establishing an additional H-
bond either with the UP or the DOWN counter-group of the hinge backbone (these two patterns of
H-bonds interactions are herein named 2-UP and 2-DOWN, Figures 1b and 1c, respectively).
Occasionally, the inhibitor scaffold is even able to establish three H-bonds (TRIPLE pattern, Figure
1d).
5

ACCEPTED MANUSCRIPT
Figure 1. a) Representation of the RET kinase catalytic domain: the N-lobe, the C-lobe and the
hinge region are in grey, yellow and orange ribbon, respectively. The rectangular selection
highlights the hinge region with its three H-bonding zones: up, middle and down. In b, c, and d,
prototypes of RET inhibitors establishing a 2-UP, 2-DOWN and TRIPLE interaction patterns are
shown. Specifically, the PP1, PD-173955 and Hymenialdisine inhibitors are depicted in panels b, c,
and d, respectively.
Accordingly, in an attempt to reduce the false positives coming from a conventional VS approach,
as well as the number of compounds to be actually screened, we filtered the Maybridge Hit Finder
Data Base retaining just any scaffolds potentially able to form two H-bonds with the hinge region,
namely having an H-bond donor and acceptor within 3-7 bonds distance (“hinge filter”, see Figure
2) [17].
6

ACCEPTED MANUSCRIPT
Figure 2. Representation of 2D structural filters used to refine the Hit Finder Maybridge database
Importantly, such a filter permitted to reduce the number of compounds to be screened from 14,400
to 6,248 without compromising the chemical heterogeneity of the database. In this regard, the
molecular similarity between the compounds of our collection has been evaluated on the base of the
Tanimoto coefficient (MOLPRINT2D2D fingerprint type) [18] with the aid of the “Canvas
Similarity and Clustering” tool of the Schrodinger package [19]. Table 1 shows the percentage of
compounds within different Tanimoto cutoffs, ranging from 0.01 (minimal similarity) to 0.99
(maximum similarity). Interestingly, the 11,86% of the molecules possess a Tanimoto similarity
index smaller than 0.01, and more than the 90% smaller than 0.05. These numbers indicate that the
molecules enclosed in our focused library are certainly very dissimilar and a large variety of
chemical chemotypes could possibly be exploited by the virtual screen. For comparison, also some
commercially available kinase-focused databases have been analysed: Life Chemical Kinase (LCK)
Docking, LCK Similarity, Asinex “Type I”, Chembridge KINACore and Chembridge KINASet.
Looking at Table 1 and considering, for example, the Tanimoto cutoff of 0.05, it emerges that the
94,7% of our molecules are included, while the commercial databases show a percentage of
compounds of 28.8%, 32.8%, 14.7%, 62.2% and 24.5% respectively. In line with published studies
[20,21], the data herein reported highlight a certain correlation between the filtering procedure and
the scaffold redundancy: better results in terms of diversity are obtained when using 3D
pharmacophore matching (compare KINACore and LCK Similarity) and large queries set (compare
7

ACCEPTED MANUSCRIPT
KINACore and KINASet or LCK Similarity and KINASet). However, it is evident that none of
these libraries possess the same level of chemical heterogeneity than ours and that the molecular
descriptors used in this work are not biased toward known molecular substructures.
Finally, this focused library, prepared on the base of the kinase-inhibitor interactions pattern, might
be potentially screened on every 3D kinase structure representing an universal tool in the kinase
inhibitor discovery process.
Table 1. Percentage of Compounds at different Tanimoto Coefficients Cutoffa
Tanimoto Hinge Filter
LCK
LCK
Asinex
Chembridge Chembridge
cutoff
Database
(6248)
Docking Similarity “Type I” KINACore
KINASet
(10993)
(13803)
0.4
2.1
7.5
28.8
95.7
100
(2671)
0.04
1.8
(8771)
0
0.01
0.6
14.7
98.3
100
100
(12254)
0.5
<0.01
<0.02
<0.03
<0.05
<0.1
11.9
28.3
54.1
94.7
100
100
100
1
2.8
6.1
24.5
75.5
100
100
5.5
7
16.7
62.2
100
32.8
97.2
100
100
<0.5
<0.99
100
100
100
aThe Tanimoto values were calculated using a n × n matrix (all compounds from set compared to
one another). The number in parentheses indicates the total number of compounds for each
database.
Receptor-Based VS. Docking-based virtual screening was then carried out within the Glide
software on the active conformation of RET (PDB code 2IVV [22], see Materials and Methods for
the choice criteria). Noteworthy, differently from other kinases, in which the active and the inactive
forms show different conformations, in the case of RET, X-ray analysis demonstrated that the non-
phosphorylated and phosphorylated forms show the same interlobe orientation, characteristic of the
active tyrosine kinases, and have essentially identical A-loop conformations [22]. Thus, targeting
the RET active conformation should provide ligands which bind both phosphorylated and non-
phosphorylated forms, as already observed for vandetanib [22].The best 1000 hits (in terms of free
energy of binding, see Table SI1), as resulted from VS, were subjected to a careful visual inspection
of the predicted binding poses in which, as expected, all the compounds were able to establish at
least two H-bonds with the RET hinge region (E805, Y806, A807). The predicted binding modes of
these compounds were then analysed and only few molecules were selected on the basis of these
8

ACCEPTED MANUSCRIPT
criteria i) the ligand proximity to the residues known to be important for the conformational motion
of the two lobes (e.g. K758 and E775); ii) the ligand ability to form hydrophobic interactions within
the buried part of the N-lobe (eg. L772, L779, I788, L790, L802); iii) the shape complementarity of
compounds with respect to the RET active site. Forty-four candidates were finally chosen and
purchased, to test their inhibitory efficacy against a human recombinant kinase domain of wild type
RET. The RET kinase inhibitory activity of the test compounds was determined following a
previously reported protocol [23-25], by monitoring the increase in fluorescence resulting from
phosphorylation of a peptide substrate, catalysed by the target enzyme in the presence of ATP. Five
compounds were discarded, being insoluble in the assay medium, while twenty two, out of the
remaining thirty nine, proved to inhibit the target protein when tested at a concentration of 100 µM
(hit ratio of 56.4%; Table 2 and Table SI2), with fourteen molecules displaying more than 50% of
inhibition at 100 µM (final hit ratio of 36%). Thus, the experimental success rate obtained in our VS
campaign is above the 20-30% generally obtained when a VS conventional protocol is used.
Remarkably, four compounds, 3-6 (Table 2), showed IC50 values against the target protein in the
lower micromolar range. Due to the known importance of RETV804L in the clinically unresponsive
MTC, the efficacy of derivatives 3-5 against the mutant protein was tested as well. Notably,
compounds 3 and 5 proved to inhibit also the mutated form, being 5 almost equally potent on both
wild type and RETV804L
Table 2. Structure, Maybridge Codes and Inhibitory Activity of Test Compounds 3-6.
.
wt-RET RETV804L VEGFR2 EGFR
N°
3
Compound Structure
Code
(IC50)a
(IC50)a
(IC50)a
(IC50)a
F
HTS04502
6.16
66.2
51.6
n.a.b
N
O
N
H
O
NH2
JFD01063
6.47
n.a.b
n.a.b
n.a.b
4
O
Cl
9

ACCEPTED MANUSCRIPT
H
S
Cl
N
N
N
DP01920
9.38
10.3
n.t.c
7.73
1.13
45.5
10.1
5
6
Cl
S
S
SEW04666
9.4
O
N
H
aIC50 values (µM) represent the concentration required to produce 50% enzyme inhibition. Values
are the average from at least two independent dose-response curves. Standard error of the mean
b
(SEM) are ≤ 10 %. Not active. No inhibition was observed up to 100 µM of the test compound.
cNot tested.
The observed inhibitory profile of 5 was clarified through an additional docking study, performed
comparing its binding modes to the active sites of both wild type and V804L mutated protein. As
highlighted in Figure 3, compound 5 establishes two H-bonds with the hinge backbone through its
triazole ring (in both the 1H-1,2,4- and 2H-1,2,4- tautomeric forms, see also Figure SI1). The p-Cl-
phenyl ring of the tested compound is able to hold different positions thanks to the presence of the
flexible thio-methyl linker, thus it can easily shift its profitable interactions established with L779
and L802 side chains of the wild type protein (Figure 3a) to the hydrophobic contacts with V738
and L881 side chains in the case of RETV804L (Figure 3b).
10

ACCEPTED MANUSCRIPT
Figure 3. Binding mode of compound 5 into the catalytic domain of WT (a) and V804L mutant (b)
RET kinase. The ligand is depicted in magenta sticks. Enzymes are represented in grey ribbons,
while the interacting residues are shown as white sticks. V804 and L804 are shown as balls and
sticks.
Functional evaluation of compound 5. As 5 turned out to be the most promising candidate among
the identified hits, its functional efficacy was deepened through in vitro studies, to highlight both its
inhibitory profile and its anti-tumor efficacy. Investigated against other relevant kinases, compound
5 proved to inhibit VEGFR-2 and -3, EGFR and PDGFR-β, thus showing a remarkably favorable
activity (Table 2 and Figure 4). Actually, over-expression of VEGFRs, together with a
constitutively activated RET, have been shown to be present in MTC tissues, thus making these
proteins primary targets in the treatment of MTCs [26]. Moreover, it has been demonstrated that a
combined inhibition of both PDGFR-β and VEGFR signaling is very efficacious even in late-stage
cancers, disrupting the established tumor vasculature and effecting tumor regression [27].
11

ACCEPTED MANUSCRIPT
Figure 4. Kinase selectivity profile of compound 5.
The anti-tumor efficacy of 5 was tested on commercially available TT cell lines, derived from
human MTC and representing the most reliable model system for the parafollicular C cells available
to date [28]. Moreover, 5 was also tested on surgical thyroid tissues, obtained from five patients
with MTC at the time of surgery, to establish its efficacy on primary cell cultures. As regards the
TT cell lines, the viability and proliferation assays (Cell Proliferation Reagent WST-1) were used to
assess the number of viable cells. The results highlighted that, when TT cells were treated with 5,
30 and 50 µM of compound 5, a reduction of proliferation with respect to the control both at 1 h
(data not shown) and at 2 h was observed (Figure 5), mainly with 30 (P = 0.001, by ANOVA) and
50 µM (P < 0.0001, by ANOVA). Results observed with surgical thyroid tissues, obtained from the
five patients with MTC, provided comparable results (Figure 6). Notably, among the five MTC
specimens, two RET mutations were observed (C634G, C634R, respectively). Results about the
inhibition of proliferation, obtained in primary MTC cell cultures with RET mutation, were not
statistically different from those obtained in MTC without RET mutations (data not shown).
Unfortunately, the V804L mutation was not found in none of the selected patients, thus available
data are limited to in vitro tests (Table 2).
12

ACCEPTED MANUSCRIPT
Figure 5. Results of WST-1 assay (mean ± S.E.M. of all samples at 2 h from the start of tetrazolium
reaction) showed a significant reduction of proliferation (after 48 h of incubation) with respect to
the control (Ctrl) with 5 at 30 and 50 µM (P < 0.05, by ANOVA) in TT cells. Bars are mean ±
S.E.M. *P < 0.05 or less versus Ctrl by Bonferroni-Dunn test.
Figure 6. Results of WST-1 assay (mean ± S.E.M. of all samples at 2 h from the start of tetrazolium
reaction) showed a significant reduction of proliferation (after 48 h of incubation) with respect to
the control (Ctrl) with 5 at 30 and 50 µM in primary medullary thyroid cancer cells. Bars are mean
± S.E.M. *P < 0.05 or less versus Ctrl by Bonferroni-Dunn test.
13

ACCEPTED MANUSCRIPT
LogD, Metabolic Stability, hERG Theoretical Binding. Prompted by the significant anti-tumour
results, we than pursued our studies on 5 investigating its physical-chemical properties, metabolic
stability in plasma and liver microsomes, and potential hERG-related toxicity of the compound, to
probe its soundness as novel hit candidate for the treatment of MTCs. The logD value was
experimentally determined at pH = 7.4 and resulted to be within the range of the drug-likeness
(LogD7.4 = 4.1) [29]. The thermodynamic solubility turned out to be somewhat low (1 nM), but this
limit was profitably overcome by conversion of 5 into the corresponding potassium salt, 42. This
latter, obtained treating 5 with KOH/EtOH, kept unaltered the inhibitory profile and turned out to be
equipotent to the parent acid, 5, against the wild type RET (data not shown). Metabolic stability of 5
was assessed as well, by using liver human microsomes and human plasma. The test compound
showed a good stability in microsomes (40% of remaining compound after 15 min of incubation at
1 µM) and an excellent plasma stability (85% of remaining compound after 1h of incubation).
Furthermore, 5 was predicted not to be an hERG inhibitor by using the rule of thumb of an in silico
model recently described [30]. Accordingly, 5 should display, in principle, a safe cardiac profile,
thus turning out to be more favorable with respect to vandetanib, for which a high rates of hERG-
mediated QT interval prolongation has been reported so that FDA required a restricted distribution
program [31]. All together, these data indicate that 5 represent a viable hit to optimize.
Structural optimization of compound 5. To identify novel promising lead compounds, we
embarked in a structure-guided synthetic campaign exploring the key pharmacophore elements of
compound 5 and their relevance in the molecular recognition of RET active site.
Leaving unaltered the 1,2,4-triazole core, required to hook the hinge region through H-bond
interactions, we first investigated the role of the thio-methyl linker, which was replaced either with
an oxy-methyl or an ethyl chain, or even oxidized to a sulfonyl-methyl fragment, as in compounds
43a-c. Then, focusing on the pendant phenyl rings in positions 3 and 5 of the nucleus, different
substituents were exploited to replace the chlorine atoms, one at a time, in order to verify the
influence of electron-withdrawal, electron-release, steric bulk and orientation on the interaction
14

ACCEPTED MANUSCRIPT
with both the buried region of the site and the area exposed to the solvent. Accordingly, we
designed and synthesized both compounds 43d-i, bearing either electron-withdrawing or electron-
releasing groups in position para of the 5-phenyl ring, and compounds 43j-n, where the aromatic
area in position 3 of the hit was expanded through the insertion of phenyl-ureido, benzamido, and
phenyl-sulfonamido fragments.
Chemistry. Derivative 43a, 3-((4-chlorophenoxy)methyl)-5-phenyl-1H-1,2,4-triazole, was obtained
as reported in Scheme 1. Reaction of 2-(4-chlorophenoxy)acetonitrile, 44, with benzohydrazide, 45,
in the presence of sodium ethoxide, gave the key intermediate 46, N’-(2-(4-chlorophenoxy)-1-
aminoethylidene)benzohydrazide, which cyclized to the target inhibitor, 43a, by treatment with
POCl3. Compound 43b, 3-(4-chlorophenethyl)-5-(4-chlorophenyl)-1H-1,2,4-triazole, was prepared
according to Scheme 2. Treatment of the commercially available ethyl 3-(4-
chlorophenyl)propanoate, 47, with hydrazine in methanol solution afforded the corresponding
propanehydrazide 48, which was converted into the desired inhibitor 43b by reaction with the
commercially available benzamidine hydrochloride 49, in the presence of sodium ethoxide.
Derivative 43c, 3-(4-chlorophenyl)-5-(((4-chlorophenyl)sulfonyl)methyl)-1H-1,2,4-triazole, was
synthesized as depicted in Scheme 3. The commercial acetohydrazide 50 was reacted with 4-
chlorobenzonitrile 51, at 150°C in a sealed tube and in the presence of K2CO3, to achieve the 1,2,,4-
triazole 5 according to the procedure of Yeung and co-workers [32]. Then, 5 was oxidized to 43c by
treatment with m-chloroperbenzoic acid.
Scheme 1. Synthesis of 3-((4-Chlorophenoxy)methyl)-5-phenyl-1H-1,2,4-triazole, 43a
NH2
O
CN
O
O
NHNH2
N
H
O
Cl
N
N
N
O
NH
Cl
a
b
46
43a
Cl
45
44
a) EtONa, EtOH, reflux, 12h, 25%; b) POCl3, 130°C, 8h, 33%.
15

ACCEPTED MANUSCRIPT
Scheme 2. Synthesis of 3-(4-chlorophenethyl)-5-(4-chlorophenyl)-1H-1,2,4-triazole, 43b
OEt
NHNH2
H
Cl
Cl
N
N
N
O
O
a
b
43b
Cl
Cl
Cl
H2N
NHHCl
47
48
49
a) NH2NH2, MeOH, reflux, 2h, 89%; b) EtONa, EtOH, reflux, 8h, 30%.
Inhibitors 43d-i, keeping unaltered the p-chlorophenyl-thiomethyl residue of the hit, were
synthesized following a similar procedure (Scheme 4), from acetohydrazide 50 and commercially
available or previously prepared benzonitriles (52a-f, Supporting Information).
16

ACCEPTED MANUSCRIPT
Compounds 43j-n, characterized by a different substitution pattern on the phenyl-thiomethyl
fragment, were obtained as described in Scheme 5. Reaction of 4-aminobenzenethiol 53 with ethyl
chloroacetate and K2CO3, followed by treatment with hydrazine in methanol solution, gave the
intermediate acetohydrazide 54, which cyclized to the 1,2,4-triazole derivative 55 by reaction with
the suitable benzonitrile, in the presence of K2CO3. The key 1,2,4-triazole intermediate afforded the
ureic inhibitors 43j-l by treatment with differently substituted phenyl isocyanates, the benzamide
derivative 43m by reaction with benzoyl chloride, and the benzenesulfonamide derivative 43n by
using benzenesulfonyl chloride.
17

ACCEPTED MANUSCRIPT
18

ACCEPTED MANUSCRIPT
Functional Evaluation of Compounds 43a-n. All the synthesized derivatives were tested for their
inhibitory properties against the target protein RET, starting from the reference concentration of 100
µM exploited in the lead off screening study. Compounds 43a-c, featuring different di-atomic
linkers connecting the main cyclic portions of the inhibitors, did not show any relevant functional
efficacy (Table 3), thus emphasizing the prominent role played by the thio-methyl fragment in the
interaction with the active site of the protein. Modification of the p-chloro substitution pattern on
the 5-phenyl ring of the hit, like in compounds 43d-i, gave rise to contrasting results. Replacing the
chloro atom with the bulkier, less electron-attracting bromine one turned out to be detrimental, as
the inhibitory activity of the resulting 43d was strongly reduced (Table 3). On the contrary, the
presence of the electron-donating hydroxy group, as in derivative 43e, retained an appreciable
functional efficacy. Further functionalization of the hydroxy group, performed either with a benzyl
residue, as in 43f, or with a pyridine ring, like in 43g, drastically reduced or even nullified the
activity. The aromatic area in position 3 of the 1,2,4-triazole core was then widen through the
insertion of phenyl-ureido (43j-43l), benzamido (43m) and phenyl-sulfonamido (43n) moieties.
Among others, the insertion of both a meta-acetyl residue and a dimethoxy substitution on the distal
phenyl ring, as in 43k and 43l respectively, allowed to increase the inhibitory potency of the hit.
In order to rationalize the better activity of some phenyl-ureido containing compounds (eg. 43k and
43l), molecular docking of the most active derivative 43l was performed with the aid of the Glide
software. The docking calculations clearly suggest that the novel analogue is no longer able to fit
the widened aromatic area in position 3 in the buried part of the N-lobe like in the case of 5. In fact
the program converged toward a pose (Figure 7) where the triazole core, still able to H-bonds with
E805 and A807, is flipped so that the p-Cl-phenyl ring is oriented toward the N-lobe pocket, while
the thio-methyl linker bends in order to direct the extended phenyl-ureido moiety under the
nucleotide-binding loop (K728-K737). This pose, already found for other bis-aryl urea kinase
inhibitors [33], is stabilized by an H-bond between the catalytic D892 carboxylate and the urea
group, and allows the formation of several hydrophobic contacts between the two 43l aromatic rings
19

ACCEPTED MANUSCRIPT
and the protein (F735, V738, L881, L895) and of a cation-pi interaction between the terminal
phenyl ring and the K758 side chain, which is strengthened by the two methoxy electron donor
groups.
On the other hand, the inactive compounds 43m and 43n, possessing an amido and a sulfonamido
group, which are endowed with a different geometry and length with respect to the urea, would not
accommodate in the ATP pocket in the same fashion of 43l. Indeed, docking of 43m, shows that in
order to place the amido-aromatic portion under the nucleotide-binding loop, the molecule shifts its
position and the two H-bonds with the hinge region are totally lost (Figure SI2).
These results provide useful and clear clues about the future synthetic development of this class of
RET inhibitors. Actually, compounds 43k and 43l clearly point out that substitution on the distal
aromatic area in position 3 of the 1,2,4-triazole core could be a valuable staring point to obtain
novel and highly potent inhibitors, and that phenyl-ureido moieties substituted with large and polar
groups are required in this region.
Figure 7. Binding mode of compound 43l into the catalytic domain of WT RET kinase. The ligand
is depicted in blue sticks. The protein is represented in grey ribbons, while the interacting residues
are shown as white sticks. H-bonds are highlighted in dashed black lines and apolar hydrogens are
omitted for sick of clarity.
20

ACCEPTED MANUSCRIPT
Table 3. RET Inhibitory Activity of 1,2,4-Triazole derivatives 43a-n.
a
% Inhibition
IC50
N°
X
R1
R2
(100 µM)
(µM)
O
CH2
SO2
S
H
Cl
Cl
Cl
Cl
Cl
Cl
n.a.b
n.a.b
n.a.b
30
43a
43b
43c
43d
43e
Cl
Br
OH
S
32.5
S
S
Cl
Cl
n.a.b
43f
33.0
45.7
43g
S
S
NH2
Cl
Cl
43h
43i
19.0
S
S
Cl
H
15.2c
43j
6.07
5.51
43k
S
Cl
43l
S
S
Cl
Cl
11.5c
33.9c
43m
43n
21

ACCEPTED MANUSCRIPT
S
Cl
Cl
9.38
5
aIC50 values (µM) represent the concentration required to produce 50% enzyme inhibition.
Values are the average from at least two independent dose-response curves. Standard error of
the mean (SEM) are ≤ 10 %. bNot active. No inhibition was observed up to 100 µM of the test
compound. cCompound tested at 10 µM.
3. Conclusion
In this study, a virtual screening campaign was conducted to identify novel RET inhibitors.
Particularly, taking advantage of the knowledge regarding the interactions pattern between ATP-
mimetic inhibitors and the kinase hinge region, we applied a 2D filtering step that allowed us to
reduce the number of molecules to screen and to obtain a focused library where, possibly, a
generous number of false positives were removed. Then, this ligand-based procedure was followed
by a receptor-based VS on the RET X-ray structure. Through the herein described campaign,
twenty-two out of thirty-nine compounds proved to inhibit the target protein yielding an
experimental hit ratio higher than the 20-30% generally obtained by a conventional VS protocol.
Moreover, we succeeded in discovering a novel 1,2,4-triazole derivative, 5, which proved to inhibit
effectively both wild type and V804L mutant RET, thus overcoming in principle the problem of
resistance displayed by drugs currently approved for MTC, vandetanib and cabozantinib. The
combined blockade of RET, EGFR, VEGFR-2 and -3, and PDGFR-β, all playing a pivotal role in
MTC onset and progression, allowed 5 to reduce viability and proliferation of both TT cells and
primary cell cultures obtained from surgical thyroid specimens. This functional behavior, together
with favourable early ADME properties, made 5 an ideal candidate for a lead optimization
campaign. Accordingly, a series of derivatives were synthesized and tested, which helped us to
sketch out the key structural fragments for a fruitful RET molecular recognition. Interestingly,
derivatives 43k and 43l, characterized by the widest and most polar substitution pattern on position
3 of the triazole core, turned out to be more active than 5, thus paving the way for future structure-
activity investigation in this direction.
22

ACCEPTED MANUSCRIPT
4. Materials and Methods
Chemistry. Melting points were determined using a Reichert Köfler hot-stage apparatus and are
uncorrected. Routine 1H-NMR spectra were recorded in DMSO-d6 solution on a Varian Gemini 200
spectrometer operating at 200 MHz. Evaporation was performed in vacuo (rotary evaporator).
Analytical TLC was carried out on Merck 0.2 mm precoated silica gel aluminium sheets (60 F-254).
Flash chromatography was performed on silica gel (70–230 mesh ASTM) using the reported
eluents. High-resolution ESI-MS spectra were performed on a Thermo LTQ Orbitrap XL mass
spectrometer. Purity of the target inhibitors, 43a-n, was determined by HPLC analysis, performed
exploiting a Merck Hitachi D-7000 liquid chromatograph (UV detection at 242 nm) and a
Discovery C18 column (250 mm x 4.6 mm, 5 µm, Supelco), with a gradient of 40% water and 60%
acetonitrile and a flow rate of 1.5 mL/min. All the compounds showed percent purity values ≥ 95%.
2-(4-Chlorophenoxy)acetonitrile, benzohydrazide, ethyl 3-(4-chlorophenyl)propanoate, 4-
chlorobenzene-1-carboximidamide hydrochloride, 4-aminothiophenol, 4-chlorobenzonitrile, 4-
bromobenzonitrile, 4-hydroxybenzonitrile, 2-[(4-chlorophenyl)sulfanyl]acetohydrazide and
chloroethylacetate, used to obtain the target inhibitors, were from Alfa Aesar, Aldrich and Fluka.
Synthesis of N’-(2-(4-chlorophenoxy)-1-aminoethylidene)benzohydrazide, 46. A solution of 2-
(4-chlorophenoxy)acetonitrile 44, (1.00 mmol), and sodium ethoxide (0.10 mmol) in absolute EtOH
(20 mL) was left under stirring at room temperature for 1 hour, than benzohydrazide 45, (1.00
mmol), was added and the resulting mixture was refluxed under stirring until the disappearance of
the starting materials (TLC analysis). After cooling to room temperature, the solid precipitated was
1
collected by filtration and purified by crystallization from EtOH. M.p. 195 °C. Yield: 25%. H-
NMR (δ, ppm): 9.88 (s, 1H, exc.), 7.80 (d, 2H, J=7.8 Hz), 7.44 (d, 3H, J=7.3 Hz), 7.33 (d, 2H,
J=8.8 Hz), 7.04 (d, 2H, J=8.9 Hz), 6.56 (s, 2H, exc.), 4.52 (s, 2H).
Synthesis of 3-((4-chlorophenoxy)methyl)-5-phenyl-1H-1,2,4-triazole, 43a. A suspension of N'-
(2-(4-chlorophenoxy)-1-aminoethylidene)benzohydrazide 46, (1.00 mmol) in POCl3 (5 mL) was
23

ACCEPTED MANUSCRIPT
heated under stirring at T = 130 °C until the disappearance of the starting material (TLC analysis).
Once the reaction was complete, the resulting mixture was carefully poured into crushed ice and the
separated solid was collected by filtration and purified by crystallization from EtOH. M.p. 154-155
1
°C. Yield: 33%. H-NMR (δ, ppm): 8.01-7.97 (m, 2H), 7.47 (s, 3H), 7.37-7.32 (m, 2H), 7.11-7.07
(m, 2H), 5.20 (s, 2H). 13C-NMR (δ, ppm): 159.65, 157.15, 146.28, 132.43, 131.92, 131.12, 129.41,
129.20, 127.54, 126.74, 117.50, 60.34. HRMS (ESI) [M + H]+ calculated for C15H12ClN3O:
286.0669, found: 286.0607.
Synthesis of 3-(4-chlorophenyl)propanehydrazide, 48 [34]. The commercially available ethyl 3-
(4-chlorophenyl)propanoate 47 (1.00 mmol) was allowed to react with hydrazine (3.00 mmol) in
MeOH solution (10 mL), at reflux, until the disappearance of the starting material (TLC analysis).
The solvent was removed in vacuo and the solid obtained was purified by crystallization from
EtOH. M.p. 121-122 °C. Yield: 89%. 1H-NMR (δ, ppm): 8.30 (s, 1H, exc.), 7.30-7.16 (m, 6H), 2.76
(t, 2H, J=8.0 Hz), 2.27 (t, 2H, J=8.0 Hz), 5.20 (s, 2H).
Synthesis of 3-(4-chlorophenethyl)-5-(4-chlorophenyl)-1H-1,2,4-triazole, 43b. A solution of 3-
(4-chlorophenyl)propanehydrazide 48, (1.00 mmol), and sodium ethoxide (2.00 mmol) in absolute
EtOH (20 mL) was left under stirring at room temperature for 1 hour. The commercially available
4-chlorobenzene-1-carboximidamide hydrochloride 46 (1.00 mmol) was then added and the
resulting mixture was refluxed until the disappearance of the starting materials (TLC analysis). The
solvent was removed in vacuo and the solid obtained was purified by crystallization from toluene.
M.p. 116-118 °C. Yield: 30%. 1H-NMR (δ, ppm): 8.03 (bs, 1H, exc.), 7.80 (d, 2H, J=6.6 Hz), 7.52
13
(d, 2H, J=6.8 Hz), 7.33-7.19 (m, 4H), 2.82 (t, 2H, J=7.8 Hz), 2.36 (t, 2H, J=7.8 Hz). C-NMR (δ,
ppm): 173.65, 167.25, 140.61, 136.54, 133.50, 130.97, 130.88, 130,63, 129.28, 128.76, 128.59,
36,86, 30.54. HRMS (ESI) [M + H]+ calculated for C16H14Cl2N3: 317.0487, found: 318.0408.
General synthesis of 3,5-disubstituted-1,2,4-triazoles, 5, 43d-i and 55a,b. A mixture of 2-((4-
chlorophenyl)thio)acetohydrazide 50 (1.00 mmol), the suitably substituted nitrile (3.00 mmol), and
K2CO3 (0.5 mmol) in n-BuOH (2 mL) was stirred for 16 h in a reusable sealed tube at 150 °C, in an
24

ACCEPTED MANUSCRIPT
oil bath. The progress of the reaction was monitored by the disappearance of the hydrazide by TLC
analysis. The solvent was removed in vacuo and the residue obtained was purified by flash
chromatography (hexanes/EtOAc 1:1).
5-(4-Chlorophenyl)-3-(((4-chlorophenyl)thio)methyl)-1H-1,2,4-triazole, 5. White solid. M.p.
1
121-122 °C. Yield: 51%. H-NMR (δ, ppm): 10.63 (bs, 1H, exc.), 7.93 (d, 2H, J=8.4 Hz), 7.42 (d,
2H, J=8.0 Hz), 7.31-7.25 (m, 4H), 4.27 (s, 2H). 13C-NMR (δ, ppm): 159.52, 156.95, 136.28, 133.43,
132.72, 131.22, 129.41, 129.13, 128.54, 127.74, 127.49, 30.34. HRMS (ESI) [M + H]+ calculated
for C15H11Cl2N3S: 335.0051, found: 335.0233.
5-(4-Bromophenyl)-3-(((4-chlorophenyl)thio)methyl)-1H-1,2,4-triazole, 43d. White solid. M.p.
1
156-158 °C. Yield: 33%. H-NMR (δ, ppm): 7.84 (d, 2H, J=8.0 Hz), 7.53 (d, 2H, J=8.4 Hz), 7.26-
13
7.19 (m, 4H), 4.27 (s, 2H). C-NMR (δ, ppm): 159.32, 157.65, 135.88, 132.93, 132.43, 131.62,
129.27, 129.06, 128.82, 127.78, 127.49, 29.34. HRMS (ESI) [M + H]+ calculated for
C15H11BrClN3S: 378.9546, found: 378.9681.
4-(3-(((4-Chlorophenyl)thio)methyl)-1H-1,2,4-triazol-5-yl)phenol, 43e. White solid. M.p. 139-
1
141 °C. Yield: 21%. H-NMR (δ, ppm): 7.78 (d, 2H, J=8.4 Hz), 7.40-7.28 (m, 4H), 7.01 (d, 2H,
J=8.8 Hz), 5.08 (s, 2H). 13C-NMR (δ, ppm): 159.19, 158.13, 157.05, 134.38, 132.96, 131.08,
128.71, 125.13, 118.01, 34.21. HRMS (ESI) [M + H]+ calculated for C15H11ClN3OS: 317.0390,
found: 317.0712.
5-(4-(Benzyloxy)phenyl)-3-(((4-chlorophenyl)thio)methyl)-1H-1,2,4-triazole, 43f. White solid.
1
M.p. 161-162 °C. Yield: 63%. H-NMR (δ, ppm): 7.88 (d, 2H, J=8.4 Hz), 7.41-7.22 (m, 9H), 7.04
13
(d, 2H, J=86.8 Hz), 5.11 (s, 2H), 4.24 (s, 2H). C-NMR (δ, ppm): 158.98, 158.15, 157.09, 136.47,
134.38, 132.51, 131.33, 129.78, 128.97, 127.85, 127.01, 125.09, 116.12, 69.78, 34.45. HRMS (ESI)
[M + H]+ calculated for C22H18ClN3OS: 407.0859, found: 407.0903.
4-(4-(3-(((4-Chlorophenyl)thio)methyl)-1H-1,2,4-triazol-5-yl)phenoxy)pyridine, 43g. Colorless
solid. M.p. 157-157 °C. Yield: 38%. 1H-NMR (δ, ppm): 8.18-8.13 (m, 4H), 7.65 (d, 2H, J=8.4 Hz),
13
7.38 (d, 2H, J=8.4 Hz), 7.29 (d, 2H, J=8.4 Hz), 6.58 (d, 2H, J=7.2 Hz), 4.59 (s, 2H). C-NMR (δ,
25

ACCEPTED MANUSCRIPT
ppm): 162.81, 159.24, 158.67, 156.99, 163.28, 134.43, 132.31, 131.37, 130.41, 129.28, 128.63,
121.74, 103.49, 31.87. HRMS (ESI) [M + H]+ calculated for C20H15ClN4OS: 394.0655, found:
394.0787.
4-(3-(((4-Chlorophenyl)thio)methyl)-1H-1,2,4-triazol-5-yl)aniline, 43h. Yellow solid. M.p. 145-
1
147 °C. Yield: 49%. H-NMR (δ, ppm): 7.69 (d, 2H, J=8.4 Hz), 7.31-7.19 (m, 4H), 6.68 (d, 2H,
J=8.4 Hz), 4.20 (s, 2H). 13C-NMR (δ, ppm): 150.69, 149.79, 133.80, 129.57, 126.73, 120.35,
114.45, 112.95, 99.76, 30.34. HRMS (ESI) [M + H]+ calculated for C15H13ClN4S: 316.0549, found:
316.0557.
1-(4-(3-(((4-Chlorophenyl)thio)methyl)-1H-1,2,4-triazol-5-yl)phenyl)-4-methylpiperazine, 43i.
1
White solid. M.p. 172-174 °C. Yield: 67%. H-NMR (δ, ppm): 7.82 (d, 2H, J=8.4 Hz), 7.33-7.18
(m, 4H), 6.88 (d, 2H, J=8.4 Hz), 4.22 (s, 2H), 3.32-3.29 (m, 4H), 2.71-2.69 (s, 4H), 2.43 (s, 3H).
13C-NMR (δ, ppm): 158.93, 158.14, 151.98, 133.90, 132.71, 131.03, 129.62, 129.11, 127.59,
125.70, 118.98, 117.86, 115.42, 101.42, 54.47, 47.56, 45.48, 30.94. HRMS (ESI) [M + H]+
calculated for C20H22ClN5S: 399.1284, found: 399.1343.
4-(((5-phenyl-1H-1,2,4-triazol-3-yl)methyl)thio)aniline, 55a. Colorless oil. Yield: 52%. 1H-NMR
(δ, ppm): 14.01 (s, 1H, exc.), 7.96 (d, 2H, J=8.6 Hz), 7.41-7.48 (m, 3H), 7.08 (d, 2H, J=8.5), 6.50
13
(d, 2H, J=8.4 Hz), 5.26 (s, 2H, exc.), 4.03 (s, 2H). C-NMR (δ, ppm): 169,80, 142,52, 139.90,
132.75, 132,15, 131.13, 129,90, 128.62, 127.41, 116.98, 117.86, 115.42, 36.24. HRMS (ESI) [M +
H]+ calculated for C15H15N4S: 283,0939, found: 283.0999.
4-(((5-(4-Chlorophenyl)-1H-1,2,4-triazol-3-yl)methyl)thio)aniline, 55b. White solid. M.p. 168-
1
170 °C. Yield: 72%. H-NMR (δ, ppm): 14.03 (s, 1H, exc.), 7.96 (d, 2H, J=8.6 Hz), 7.54 (d, 2H,
13
J=8.5), 7.07 (d, 2H, J=8.4 Hz), 6.47 (d, 2H, J=8.5), 5.30 (s, 2H, exc.), 4.04 (s, 2H). C-NMR (δ,
ppm): 169,80, 142,52, 139.90, 132.75, 132,15, 131.13, 129,90, 128.62, 127.41, 116.98, 117.86,
115.42, 36.24. HRMS (ESI) [M + H]+ calculated for C15H14ClN4S: 317,0549, found: 317.0576.
Synthesis of 5-(4-chlorophenyl)-3-(((4-chlorophenyl)sulfonyl)methyl)-1H-1,2,4-triazole, 43c. A
solution of 5 (1.00 mmol) in DCM (10 mL) was treated with m-chloroperbenzoic acid (3.00 mmol)
26

ACCEPTED MANUSCRIPT
and stirred at room temperature for 5 h. The solid formed was removed by filtration and the filtrate
was washed with sodium-bicarbonate solution, water and brine. The organic layer was dried over
Na2SO4 and evaporated under vacuum. The residue obtained was purified by Flash
1
Chromatography (hexanes/EtOAc 1:1). Yellow solid. M.p. 163-165 °C. Yield: 81%. H-NMR (δ,
ppm): 7.82 (d, 2H, J=8.4 Hz), 7.73 (d, 2H, J=7.2 Hz), 7.49 (d, 2H, J=8.4 Hz), 7.43 (d, 2H, J=6.6
13
Hz), 4.66 (s, 2H). C-NMR (δ, ppm): 158.31, 156.03, 135.17, 129.98, 129.71, 129.28, 127.66,
127.08, 125.87, 124.62, 29.70. HRMS (ESI) [M + H]+ calculated for C15H11Cl2N3O2S: 366.9949,
found: 366.9957.
Synthesis of 2-((4-aminophenyl)thio)acetohydrazide, 54. A solution of 4-aminobenzenethiol
(1.00 mmol), chloroethylacetate (1.00 mmol), and K2CO3 (1.00 mmol) in DMF (5 mL) was left
under stirring at room temperature until the disappearance of the starting material (TLC analysis).
Once the reaction was complete, the solvent was removed in vacuo and the resulting residue was
washed with water to obtain the key intermediate ethyl 2-((4-aminophenyl)thio)acetate. This latter
was then allowed to react with hydrazine (3.00 mmol) in MeOH solution (10 mL), at reflux, until
the disappearance of the starting material (TLC analysis). The solvent was removed in vacuo and
the solid obtained was purified by crystallization from EtOH. M.p. 94-96 °C. Yield: 90%. 1H-NMR
(δ, ppm): 9.03 (s, 1H, exc.), 7.11 (d, 2H, J=8.5), 6.51 (d, 6H, J=8.5), 5.24 (s, 2H, exc.), 3.33 (s, 2H,
exc.), 3.30 (s, 2H).
General synthesis of 1-(4-(((5-(4-chlorophenyl)-1H-1,2,4-triazol-3-yl)methyl)thio)phenyl)-3-
(substituted)phenylurea, 43j-l. A solution of the 1,2,4-triazole derivative 55 (1.00 mmol) and the
appropriate phenyl isocyanate (1.00 mmol) in THF (20 mL) was left under stirring at room
temperature until the disappearance of the starting material (TLC analysis). Once the reaction was
complete, the solvent was removed in vacuo and the residue obtained was purified by crystallization
from the suitable solvent.
1-(4-(((5-(4-Chlorophenyl)-1H-1,2,4-triazol-3-yl)methyl)thio)phenyl)-3-phenylurea, 43j. White
solid. M.p. 186-189 °C. Cryst. solvent: MeOH. Yield: 24%. 1H-NMR (δ, ppm): 14.10 (s, 1H, exc.),
27

ACCEPTED MANUSCRIPT
8.80 (s, 1H, exc.), 8.75 (s, 1H, exc), 7.97 (dd, 2H, J=8.6 Hz, J=4.8 Hz), 7.54 (d, 2H, J=8.6 Hz),
7.45-7.41 (m, 4H), 7.34 (dd, 2H, J=8.7 Hz, J=4.7 Hz), 7.28 (t, 2H, J=7.7), 6.97 (t, 1H, J=7.3 Hz),
13
4.23 (s, 2H). C-NMR (δ, ppm): 156,50, 152,88, 140,06, 139,48, 134,29, 131,31, 129,39, 129,24,
129,07, 127,98, 126,97, 122,35, 119,17, 118,69, 30,63. HRMS (ESI) [M + H]+ calculated for
C22H19ClN5OS: 436.0921, found: 436,0999.
1-(3-acetylphenyl)-3-(4-(((5-(4-chlorophenyl)-1H-1,2,4-triazol-3-yl)methyl)thio)phenyl)urea,
43k. White solid. M.p. 158-160°C. Yield: 25%. 1H-NMR (δ, ppm): 14.10 (s, 1H, exc.), 8.97 (s, 1H,
exc.), 8.79 (s, 1H, exc), 8.06 (t, 1H, J=1.9), 7.97 (dd, 2H, J=8.6 Hz, J=1.5 Hz), 7.67 (d, 1H, J=8.0
13
Hz), 7.61-7.58 (m, 2H), 7.48-7.41 (m, 4H), 7.26 (d, 2H, J=2.0 Hz), 4.10 (s, 2H), 2.69 (s, 3H). C-
NMR (δ, ppm): 161,15, 152,91, 140,48, 137,84, 130,67, 129,80, 129,65, 129,56, 129,34, 129,11,
126,48, 126,18, 123,33, 122,49, 121,91, 118,63, 117,86, 117,02, 116,42, 27,26. HRMS (ESI) [M +
H]+ calculated for C24H22N5O2S: 444.1416, found: 444,1494.
1-(4-(((5-(4-chlorophenyl)-1H-1,2,4-triazol-3-yl)methyl)thio)phenyl)-3-(2,5-dimethoxyphenyl)
1
urea, 43l. White solid. M.p. 178-180 °C. Cryst. solvent: MeOH. Yield: 25%. H-NMR (δ, ppm):
14.13 (s, 1H, exc.), 9.42 (s, 1H, exc.), 8.26 (s, 1H, exc.), 7.97 (dd, 2H, J=8.6, J=2.4), 7.84 (d, 1H,
J=3.0), 7.55 (d, 2H, J=8.4), 7.42 (d, 2H, J=8.7), 7.35 (d, 2H, J=8.7), 6.92 (d, 1H, J=8.8), 6.49 (dd,
13
1H, J=8.4, J=3.0), 4.23 (s, 2H), 3.82 (s, 3H), 3.69 (s, 3H). C-NMR (δ, ppm): 153,74, 152,63,
131,96, 129,85, 129,62, 129,42, 129,14, 128,00, 126,48, 126,18, 123,33, 122,49, 121,91, 118,63,
117,86, 117,02, 116,42, 63,95, 56,79, 29,35. HRMS (ESI) [M - H]- calculated for C24H21ClN5O3S:
494.1132, found: 494,1054.
Synthesis of 4-chloro-N-(4-(((5-(4-chlorophenyl)-1H-1,2,4-triazol-3-yl)methyl)thio)phenyl)
benzamide, 43m. A solution of the 1,2,4-triazole derivative 55 (1.00 mmol) and 4-chlorobenzoyl
chloride (1.20 mmol) in toluene (20 mL) was left under stirring at room temperature until the
disappearance of the starting material (TLC analysis). Once the reaction was complete, the solvent
was removed in vacuo and the residue obtained was washed with water and purified by
recrystallization from MeOH. M.p. 228-230 °C. Yield: 46%. 1H-NMR (δ, ppm): 14.12 (s, 1H, exc.),
28

ACCEPTED MANUSCRIPT
10.35 (s, 1H, exc.), 7.97 (dd, 4H, J=10.2, J=4.5), 7.73 (d, 2H, J=8.7), 7.62-7.60 (m, 2H), 7.56-7.54
13
(m, 2H), 7.42 (d, 2H, J=8.7), 4.31 (s, 2H). C-NMR (δ, ppm): 164,87, 136,93, 133,94, 130,09,
129,94, 128,95, 128,00, 121,32, 27,26. HRMS (ESI) [M - H]- calculated for C22H15Cl2N4OS:
453.0422, found: 453,0344.
Synthesis of N-(4-(((5-(4-chlorophenyl)-1H-1,2,4-triazol-3-yl)methyl)thio)phenyl)benzene
sulfonamide, 43n. A solution of the 1,2,4-triazole derivative 55 (1.00 mmol), benzenesulfonyl
chloride (1.20 mmol) and and triethyl amine (1.20 mmol) in toluene (20 mL) was left under stirring
at room temperature until the disappearance of the starting material (TLC analysis). Once the
reaction was complete, the solvent was removed in vacuo and the residue obtained was purified by
recrystallization from EtOH. M.p. 228-230 °C. Yield: 46%. 1H-NMR (δ, ppm): 14,10 (s, 1H, exc.),
10,35 (s, 1H, exc.), 7.95 (d, 2H, J=8.7), 7.73 (d, 2H, J=7.6), 7.59-7.51 (m, 5H), 7.28 (d, 2H, J=8.7),
7.02 (d, 2H, J=8.7), 4.48 (s, 2H). 13C-NMR (δ, ppm): 164,87, 136,93, 133,94, 130,09, 129,94,
128,95, 128,00, 121,32, 27,26. HRMS (ESI) [M + H]+ calculated for C21H17ClN4O2S2: 457.0481,
found: 457,0404.
Synthesis of 5-(4-chlorophenyl)-3-((4-chlorophenylthio)methyl)-1H-1,2,4-triazole potassium
salt, 42. Anhydrous ethyl ether was added dropwise to an ice-cooled, stirred solution of 5-(4-
chlorophenyl)-3-((4-chlorophenylthio)methyl)-1H-1,2,4-triazole 5 (1.00 mmol) in 0.10 M
KOH/EtOH (10 mL). The resulting solid was collected by filtration. M.p.: >300 °C. Yield: 95%.
1H-NMR (δ, ppm): 7.90 (d, 2H, J=8.60 Hz), 7.46 (d, 2H, J=8.60 Hz), 7.33 (s, 4H), 4.18 (s, 2H).
Biology. Tyrosine Kinase Assays. Compounds selected by VS were purchased from Maybridge,
and subjected to NMR analysis prior to testing for their inhibitory activity, to verify their
correspondence to the chemical structures shown in Tables 2 and SI2. In addition, HPLC analyses
were performed, to verify their consistence with a purity of at least 95%. Analyses were performed
using a Merck Hitachi D-7000 liquid chromatograph (UV detection at 242 nm) and a Discovery
C18 column (250 mm x 4.6 mm, 5 µm, Supelco), with a gradient of water and acetonitrile, suitably
fitted to each compound, and a flow rate of 1.4 mL/min. All the compounds showed percent purity
29