Hedgehog pathway inhibitors of the acylthiourea and acylguanidine class show antitumor activity on colon cancer in vitro and in vivo

Article abstract of DOI:10.1016/j.ejmech.2018.07.053

Small series of acylguanidine and acylthiourea derivatives were synthesized in gram-scale and assayed for their ability to modulate the Hh signalling pathway. In vitro studies showed a low micromolar inhibitory activity toward tumor cell lines, while the oral administration revealed an excellent ADME profile in vivo. Compound 5 emerged as the most active and safe inhibitor of colon cancer cells both in vitro and in a xenograft mouse model. Based on these data, 5 could be prioritized to further development with the perspective of clinical studies.

Full text of DOI:10.1016/j.ejmech.2018.07.053

Accepted Manuscript
Hedgehog pathway inhibitors of the acylthiourea and acylguanidine class show
antitumor activity on colon cancer in vitro and in vivo
Loredana Vesci, Ferdinando Maria Milazzo, Maria Antonietta Stasi, Silvia Pace,
Francesco Manera, Carlo Tallarico, Elena Cini, Elena Petricci, Fabrizio Manetti, Rita
De Santis, Giuseppe Giannini
PII:
S0223-5234(18)30619-6
10.1016/j.ejmech.2018.07.053
EJMECH 10588
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 7 June 2018
Revised Date: 19 July 2018
Accepted Date: 22 July 2018
Please cite this article as: L. Vesci, F.M. Milazzo, M.A. Stasi, S. Pace, F. Manera, C. Tallarico, E. Cini,
E. Petricci, F. Manetti, R. De Santis, G. Giannini, Hedgehog pathway inhibitors of the acylthiourea and
acylguanidine class show antitumor activity on colon cancer in vitro and in vivo, European Journal of
Medicinal Chemistry (2018), doi: 10.1016/j.ejmech.2018.07.053.
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
O
NH
MeO
MeO
N
N
NH
H
H
O
OMe
5
Luc inhibition (Shh-lightII cells): IC50 = 20 nM
ALP inhibition (C3H10T1/2 cells): IC50 = 40 nM
Growth inhibition (LS180 cells): EC50 = 2.5 µM
Growth inhibition (HT1080 cells): EC50 = 2.0 µM
Good PK parameters

ACCEPTED MANUSCRIPT
Hedgehog pathway inhibitors of the acylthiourea and acylguanidine class show antitumor
activity on colon cancer in vitro and in vivo
Loredana Vesci,a^ Ferdinando Maria Milazzo,a^ Maria Antonietta Stasi,a Silvia Pace,a Francesco
Manera,a Carlo Tallarico,a Elena Cini,b,c Elena Petricci,c* Fabrizio Manetti,b,c Rita De Santis,a and
Giuseppe Gianninia*
aResearch & Development, Alfasigma SpA, via Pontina, km 30.400, I-00040 Pomezia, Italy
bLead Discovery Siena Srl, via Fiorentina 1, I-53100 Siena, Italy
cDipartimento di Biotecnologie, Chimica e Farmacia, Università di Siena, Dipartimento di
Eccellenza 2018-2022, via A. Moro 2, I-53100 Siena, Italy
Loredana Vesci: loredana.vesci@alfasigma.com
Ferdinando Maria Milazzo: ferdinando.milazzo@alfasigma.com
Maria Antonietta Stasi: mariaantonietta.stasi@alfasigma.com
Silvia Pace: silvia.pace@alfasigma.com
Francesco Manera: francesco.manera@alfasigma.com
Carlo Tallarico: carlo.tallarico@alfasigma.com
Elena Cini: cinielena@libero.it
Elena Petricci: elena.petricci@unisi.it
Fabrizio Manetti: fabrizio.manetti@unisi.it
Rita De Santis: rita.desantis@alfasigma.com
Giuseppe Giannini: giuseppe.giannini@alfasigma.com
^These two authors equally contributed to this work
*Corresponding author: Dr. Giuseppe Giannini, phone: +39 06 91393640, email:
giuseppe.giannini@alfasigma.com; Dr. Elena Petricci, phone: +39 0577 234276, email:
elena.petricci@unisi.it.
1

ACCEPTED MANUSCRIPT
Abstract. Small series of acylguanidine and acylthiourea derivatives were synthesized in gram-
scale and assayed for their ability to modulate the Hh signalling pathway. In vitro studies showed
a low micromolar inhibitory activity toward tumor cell lines, while the oral administration
revealed an excellent ADME profile in vivo. Compound 5 emerged as the most active and safe
inhibitor of colon cancer cells both in vitro and in a xenograft mouse model. Based on these data,
5 could be prioritized to further development with the perspective of clinical studies.
Keywords: Hedgehog inhibitor; in vivo pharmacokinetics; gram-scale synthesis; ADME-Tox;
LS180 xenograft; acylguanidine.
2

ACCEPTED MANUSCRIPT
1. Introduction1
The Hedgehog (Hh) signalling pathway, which is active in early embryonic development, plays a
pivotal role in regulation of cell proliferation and differentiation, by controlling the correct
maturation of developing tissues [1, 2]. In adulthood, Hh signalling is active in early progenitor
cells, whereas aberrant activation of the Hh pathway is associated with the development of
malignant transformation in a variety of human cancers [3]. The pathway can be stimulated by
three ligands: Sonic Hedgehog (Shh), Indian Hedgehog, and Desert Hedgehog. When one of
these ligands binds to Patched (Ptch), a 12-pass transmembrane protein that functions as a
repressor of the pathway, the inhibition of the downstream transmembrane receptor Smoothened
(Smo) by Ptch itself is relieved. Activated Smo in turn initiates a downstream signalling cascade
leading to the activation of Gli transcription factors (Gli-1, -2, -3). They translocate into the
nucleus and induce transcription of Gli target genes, such as Gli-1 itself, Ptch1 (the gene
encoding the Hh-interacting protein), Bcl2, D-type cyclins, c-Myc, Snail and BMI1, all having a
pro-tumorigenic role [4]. Ligand-independent mechanisms can also be responsible for the
constitutive Hh pathway activity in cancers. Ligand-independent constitutive activation of the
Hh pathway, found in basal cell carcinoma (BCC) and medulloblastoma, is characterized by the
presence of somatic mutations in Ptch1, Smo or SUFU genes [5]. On the other hand, the ligand-
dependent activation observed in a wide range of cancers including melanoma, pancreatic, lung,
breast, renal, and colorectal cancers [1, 3], is associated with elevated production of Hh ligands
by tumor or stromal cells [2]. In all cases, the activated Hh signalling leads to increased cell
proliferation and neoplastic transformation.
Anticancer therapy based on the inhibition of Hh signalling has been clinically validated with
two new drugs approved by FDA in 2012 and 2015 for the treatment of adult BCC [6, 7]: GDC-
0449 (vismodegib, Erivedge™ by Roche, 1, Figure 1) [8] and NVP-LDE225 (erismodegib, also
known as sonidegib, OdomzoTM by Novartis, 2) [9]. Both of them are Smo inhibitors, act by
direct interaction with the protein, and share a common diaryl carboxamide chemical scaffold.
1Abbreviations: Hh, Hedgehog; Shh, Sonic Hedgehog; Ptch, Patched; Smo, Smoothened; BCC,
basal cell carcinoma; PE, petroleum ether; SRB, sulforhodamine B; Shh-CM, Shh-conditioned
media; TV, tumor volume; TUI, tumor volume inhibition; BWL, body weight loss; HBSS,
Hank’s balanced salt solution; RLU, Relative Light Unit; ALP, alkaline phosphatase; SAG, Smo
agonist; MW, microwave; 7-ETC, 7-ethoxy coumarin; Gli, glioma-associated oncogene; I/E,
influx/efflux.
3

ACCEPTED MANUSCRIPT
Cl
O
Cl
O
N
F3CO
N
N
N
O
H
S
O O
N
H
GDC-0449, vismodegib (ErivedgeTM), 1
erismodegib, sonidegib (OdomzoTM), 2
Figure 1. FDA-approved Smo antagonists.
Compound 1 is a small Smo antagonist able to block intracellular signalling pathway upon
binding to Smo, thus resulting in inhibition of Gli-1 and down-modulation of related target
genes. Because the activated Hh pathway found in BCC is shared with other types of tumors, 1 is
currently under investigation in 61 active and completed clinical trials targeting
medulloblastoma, chondrosarcoma, pancreatic, small-cell lung, gastric cancer and colorectal
cancer [10, 11].
In a similar way, 2 is involved in 28 active and completed clinical trials for prostatic, ovarian,
pancreatic, breast, and esofageal cancer, as well as medulloblastoma and other tumors [11].
Moreover, there are currently three additional Smo inhibitors in clinical trials (namely, PF-
04449913, LY2940680, and itraconazole) [11]. Extensive pre-clinical investigations are ongoing
worldwide on Smo because it is recognized as a particularly appealing target for its central role
in the regulation of the Hh signalling pathway and for its easy accessibility on the tumor cell
surface [12]. Unfortunately, cell lines that exhibit the presence of Smo variants and are resistant
to 1 already appeared [13-15]. In this context, although 2 seems to represent an effective
alternative treatment to 1 in resistant tumors, new additional small molecules able to modulate
the Hh pathway are urgently needed.
Our research group is strongly involved in the search of novel inhibitors of the Hh pathway [16-
23] and/or Smo agonists [24, 25]. In this paper, we show that preliminary in vitro experiments on
acylguanidine and acylthiourea Smo inhibitors, synthesized at lab scale, allowed us to identify
four candidates (Table 1) that were then prepared in larger scale (5-7 g) for further
pharmacological characterization. An ADME-Tox study was then performed to evaluate their
drug-likeness, followed by an in vivo pharmacokinetic profiling and administration in a
xenograft colon cancer mouse model. Following a classical medicinal chemistry approach, the
overall experimental protocol was applied to eventually prioritize one or more compounds to
access clinical trials. Three of these compounds (namely, 3-5) belong to the acylguanidine class
previously disclosed [16, 17, 21], while 6 is a new acylthiourea derivative. They have a common
3,4,5-trimethoxybenzoyl terminal fragment (thereafter referred to as region A, Figure 2), and an
4

ACCEPTED MANUSCRIPT
acylguanidine or an acylthiourea moiety as the core portion linked to a second terminal aryl
carboxamide fragment (region B).
The acylthiourea 6 and its corresponding acylguanidine analogue 3 were specifically selected to
investigate the influence of the core portion on ADME properties and in vivo activity. Moreover,
for a better understanding of the impact of the fluorine atom on the biological profile, the
unsubstituted analogue 4 was also selected. Finally, previous results on leukemia and melanoma
cells showed a remarkable in vitro [17, 18] and in vivo activity [19] of 5 (the guanidine
derivative bearing a phenylethyl group in 4 position of the region B), thus prompting us to
include this compound in our present investigation.
Figure 2. General structure of the studied compounds.
Table 1. Selected compounds tested as putative Smo inhibitors.
O
X
O
MeO
MeO
N
N
N
H
H
H
R
ST8115, 3: X = NH·HCl, R = pF-Ph
ST8117, 4: X = NH·HCl, R = Ph
ST8116, 5: X = NH·HCl, R = -(CH2)2-Ph
ST8113, 6: X = S, R = pF-Ph
OMe
Experimental results presented here confirm that compounds with an acylguanidine core are
more active than the corresponding acylthiourea analogues (compare 3 with 6), and a flexible
molecular edge as in 5 is preferred to a rigid biphenyl moiety. Overall, 5 emerges as the best lead
candidate among test compounds and deserves further development efforts as a new oral Hh
inhibitor.
2. Materials and Methods
2.1. Chemistry
All reagents were used as purchased from commercial suppliers without further purification.
Reactions were carried out in oven-dried or flamed vessels. Solvents were dried and purified by
conventional methods prior use. All MW-assisted reactions were performed in a CEM Discover
microwave oven equipped with a 35 mL tube for reactions under pressure and an external IR
sensor to detect the reaction temperature during the irradiation (CEM Corporation, Cologno al
Serio, Italy). Flash column chromatography was performed with Aldrich silica gel 60, 0.040-
0.063 mm (230-400 mesh). Aldrich aluminum backed plates pre-coated with silica gel 60
(UV254) were used for analytical thin layer chromatography and were visualized by staining
5

ACCEPTED MANUSCRIPT
with a KMnO4, Pancaldi, or ninidrine solution. NMR spectra were recorded at 25 °C and 400 or
300 MHz for 1H and 100 or 75 MHz for 13C. The solvent is specified for each spectrum. Splitting
patterns are designated as s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet; br, broad.
Chemical shifts (δ) are given in ppm relative to the resonance of their respective residual solvent
peaks. High and low resolution mass spectroscopy analyses were recorded by electrospray
ionization.
Synthesis of (E)-4-styrylbenzoic acid 10. To a solution of 4-iodobenzoic acid 7 (1.5 g, 6 mmol, 1
equiv) and styrene 8 (2.7 mL, 24 mmol, 4 equiv) in DMF (10 mL), 9 (39 mg, 0.06 mmol, 0.01
equiv) and Et3N (1.2 mL, 8.4 mmol, 1.4 equiv) were added. The reaction mixture was irradiated
with MW in open vessel conditions, at 110 °C for 20 min (maximum power 150 W). Compound
8 (0.67 mL x 2, 6 mmol x 2, 1 equiv x 2) was added and the reaction was irradiated with MW in
the same conditions twice. Water (50 mL) was added, followed by 1N HCl until pH 2. The white
precipitate formed was filtered over Büchner, washing with H2O (50 mL) and petroleum ether
(PE, 2 x 50 mL). The pale rose solid obtained (1.5 g) was purified by flash chromatography
1
using CHCl3/CH3OH (9:1) as the eluent thus leading to a white solid (1.1 g) in 82% yield. H
NMR (400 MHz, DMSO-d6) δ 12.75 (bs, 1H), 7.96 (d, J = 8 Hz, 2H), 7.72 (d, J = 8 Hz, 2H),
7.64 (d, J = 7.6 Hz, 2H), 7.45–7.26 (m, 11H). 13C NMR (100 MHz, DMSO-d6) δ 131.41, 130.21,
129.23, 128.63, 127.77, 127.27, 126.92. ES-MS: 223 [M - H]-.
Synthesis of (E)-4-styrylbenzoyl chloride 11. The acid 10 (3 g, 13.4 mmol, 1 equiv) was
dissolved in CH2Cl2 (220 mL), and oxalyl chloride (5.6 mL, 67 mmol, 5 equiv) and DMF (730
µL) were added. The reaction mixture was stirred at rt for 3 h. The solvent was removed under
reduced pressure and the solid obtained was directly used for the next step without further
purification and characterization.
General procedure for the synthesis of nitro derivatives 13 and 17a-b. The proper acylchloride
(11 or 16a-b, 9.2 mmol, 1.2 equiv) was dissolved in dry CH2Cl2 (50 mL) and the solution was
cooled down to 0 °C. 2-Methyl-5-nitro aniline 12 (1.17 g, 7.7 mmol, 1 equiv) and pyridine (1.2
mL, 15.4 mmol, 2 equiv) were added. After 10 min, the reaction mixture was heated up to rt and
stirred overnight under N2. After dilution with CH2Cl2 (50 mL), the reaction mixture was washed
with 1N HCl (3 x 30 mL), a saturated solution of NaHCO3 (3 x 30 mL), and brine (2 x 30 mL).
The organic phase was dried over dry Na2SO4 and, after filtration and evaporation under reduced
pressure, the crude product was purified by flash chromatography.
General procedure for the synthesis of 14 and 18a-b. To a solution of 13 or 17a-b (5 mmol, 1
equiv) in MeOH (30 mL), Pd/C 5% (0.2 mmol, 0.05 equiv) was added under N2. After 3
N2/vacuum cycles, the reaction mixture was stirred at rt under H2 overnight. The suspension was
6

ACCEPTED MANUSCRIPT
filtered over a Celite pad and washed with MeOH (3 x 50 mL). The solvent was removed under
reduced pressure and the product directly used without further purification.
General procedure for the synthesis of 15 and 19a-b. To solution of the proper aniline (14 or
18a-b, 4.1 mmol, 1 equiv) in MeOH (2.5 mL), a 10% HCl solution in MeOH (7.5 mL) was
added. The mixture was stirred at rt under N2 overnight. The precipitate formed during the night
was directly used without further purification and characterization.
Synthesis of 4’-fluoro-N-(2-methyl-5-(3-(3,4,5-trimethoxybenzoyl)thioureido)phenyl)-[1,1’-
biphenyl]-4-carboxamide (6). 3,4,5-Trimethoxybenzoyl chloride 20 (2.4 g, 10.4 mmol, 1 equiv)
was dissolved in dry acetone (200 mL) and NH4SCN (858 mg, 11.2 mmol, 1.2 equiv) was added.
The suspension obtained was stirred at reflux for 2 h and 18b (3 g, 9.4 mmol, 1 equiv) was
added. The reaction mixture was refluxed for additional 3 h under N2 and directly poured into a
backer previously filled with ice. The solid formed was filtered in a Büchner funnel and washed
with H2O (2 x 40 mL), and PE (2 x 40 mL). The solid was finally purified by flash
1
chromatography (AcOEt/PE: 4:1) obtaining a pale yellow solid in 67% yield. H NMR (400
MHz, CDCl3) δ 12.54 (bs, 1H), 9.05 (bs, 1H), 8.32 (ds, J = 2.2 Hz, 1H), 7.93 (d, J = 8.0 Hz, 2H),
7.78 (s, 1H), 7.63 (d, J = 7.9 Hz, 2H), 7.57-7.54 (m, 2H), 7.47, 7.46 (dd, J1 = 8.2 Hz, J2 = 2.2 Hz,
1H), 7.25 (s, 1H), 7.12 (t, J = 8.5 Hz, 2H), 7.05 (s, 2H), 3.91 (s, 9H), 2.32 (s, 3H). 13C NMR (100
MHz, DMSO, d6) δ = 179.1, 167.5, 1656.1, 160.5, 152.8, 142.5, 141.7, 136.7, 135.6, 135.5,
133.1, 131.7, 130.3, 129.0, 128.9, 128.8, 126.6, 122.0, 121.8, 116.0, 115.9, 115.7, 106.4, 60.2,
56.2, 17.6. Mp: 215-217 °C. Elemental analysis for C31H28FN3O5S calc.: C, 64.91; H, 4.92; F,
3.31; N, 7.33; O, 13.95; S, 5.59; found: C, 64.94; H, 4.94; N, 7.32; S, 5.55.
Synthesis of N-cyano-3,4,5-trimethoxybenzamide 21. 3,4,5-Trimethoxybenzoyl chloride 20
(0.026 mol, 1 equiv) was suspended in Et2O (30 mL), and a solution of NH2CN (0.026 mol, 1
equiv) in 10% aqueous NaOH (26 mL) was added at 0 °C. The reaction mixture was stirred at rt
for 1 h, then cooled down to 0 °C, and 1N HCl was added until pH 2. The white solid obtained
was dissolved in AcOEt at reflux, then PE was added until full precipitation. After filtration over
Büchner and washing with PE, a white solid was obtained. 1H NMR (400 MHz, CD3OD) δ 7.21
(s, 2H); 3.87 (s, 6H), 3.81 (s, 3H). 222 [M + H]+. Molecular formula C11H12N2O4. Yield 75%.
General procedure for the synthesis of acylguanidines 3-5. The aniline hydrochloride 15 or 19a-
b (13.5 mmol, 1 equiv) was dissolved in toluene (800 mL) in the presence of 21 (5.2 g, 14.9
mmol, 1.2 equiv). The reaction mixture was stirred at reflux overnight. The solvent was removed
under vacuum and the solid obtained was dissolved in MeOH (35 mL), then a 10% HCl in
MeOH solution (30 mL) was added. The mixture was stirred at rt under N2 overnight to yield the
final compounds 3-5.
7

ACCEPTED MANUSCRIPT
4'-Fluoro-N-(2-methyl-5-(3-(3,4,5-trimethoxybenzoyl)guanidino)phenyl)biphenyl-4-
1
carboxamide, HCl (3). H NMR (400 MHz, DMSO-d6) δ 12.40 (bs, 1H), 11.52 (bs, 1H), 10.05
(s, 1H), 9.47 (bs, 1H), 9.43 (bs, 1H), 8.95 (bs, 1H), 8.10 (d, J = 6.4 Hz, 2H), 7.85-7.80 (m, 4H),
7.60 (s, 1H), 7.55-7.45 (m, 2 H), 7.45-7.24 (m, 4H), 3.92 (s, 6 H), 3.79 (s, 3H), 2.34 (s, 3H). 13C
NMR (75 MHz, DMSO, d6) δ 176.4, 165.6, 164.7, 161.5, 159.9, 153.1, 142.9, 140.8, 137.4,
136.4, 134.8, 133.9, 131.1, 129.8, 129.7, 129.4, 129.2, 127.4,120.2, 120.1, 116.8, 116.5, 106.7,
60.8, 56.4. Mp: 240-245 °C. Elemental analysis for C31H30ClFN4O5 calcd.: C, 62.78; H, 5.10; Cl,
5.98; F, 3.20; N, 9.45; O, 13.49; found: C, 62.74; H, 5.05; N, 9.47. Yield 82%.
N-(2-Methyl-5-(3-(3,4,5-trimethoxybenzoyl)guanidino)phenyl)biphenyl-4-carboxamide, HCl (4):
1H-NMR (400 MHz, DMSO, d6) δ = 9.18 (bs, 1 H), 9.00 (bs, 1H), 8.06 (bs, 2H), 7.80-7.10 (m,
13
14H), 3.70 (s, 6H), 3.68 (s, 3H), 2.77 (s, 3 H). C NMR (100 MHz, DMSO, d6) δ 176.0, 168.2,
165.0, 160.0, 156.5, 150.0, 146.3, 139.0, 137.4, 133.1, 131.8, 131.4, 131.0, 128.9, 128.6, 127.1,
127.0, 126.4, 123.5, 121.0, 119.3, 118.5, 116.0, 61.3, 56.0, 34.9. Mp: 250-255 °C. Elemental
analysis for C31H31ClN4O5 calcd.: C, 64.75; H, 5.43; Cl, 6.16; N, 9.74; O, 13.91; found: C,
64.72; H, 5.40; N, 9.79. Yield 92%.
3,4,5-trimethoxy-N-(N-(4-methyl-3-(4-phenethylbenzamido)phenyl)carbamimidoyl)benzamide,
HCl (5): 1H NMR (400 MHz, Methanol-d4) δ 12.31 (bs, 1H), 11.46 (bs, 1H), 9.90 (bs, 1H), 9.43
(bs, 1H), 8.95 (bs, 1H), 7.92 (s, 1H), 7.75 (d, J = 8.0 Hz, 2H), 7.67 (s, 1H), 7.26 (s, 1H), 7.24 (t,
J = 7.6 Hz, 2H), 7.17 (d, J = 6.8 Hz, 2H), 7.12 (d, J = 8.0 Hz, 2H), 3.82 (s, 9H), 2.96-2.90 (t, J =
2.0 Hz, 4H), 2.22 (s, 3H). 13C NMR (101 MHz, Methanol-d4) δ 176.35, 165.47, 159.01, 146.00,
140.59, 140.31, 136.35, 134.35, 134.78, 132.85, 131.51, 131.23, 131.23, 128.56, 126.01, 127.48,
126.84, 125.71, 120.62, 118.85, 105.54, 60.41, 55.53, 37.21, 36.98, 16.94. ES-MS: 567 [M +
H]+, 589 [M + Na]+. Mp: 118-120 °C. Elemental analysis for C33H35ClN4O5 calcd.: C, 65.72; H,
5.85; Cl, 5.88; N, 9.29; O, 13.26; found: C, 65.75; H, 5.82; N, 9.31. Yield 87%.
2.2. Biology
2.2.1. Drugs. For in vitro experiments, 10 mM stock solutions of 1 and 3-6 were prepared in
DMSO and stored at -20 °C until use. For oral administration, the compounds were dissolved in
a solutol HS15:water mixture (1:20, v/v) and delivered in a volume of 10 mL/kg.
2.2.2. Cell lines and cell cultures. LS180 (human colon adenocarcinoma), HT1080 (human
fibrosarcoma) and C3H10T1/2 (Clone 8) (mouse embryonic fibroblast) cells were purchased
from ATCC. All cells were routinely cultured in their appropriate growth medium and
maintained in a humidified atmosphere with 5% CO2 at 37 °C. All experiments were performed
8

ACCEPTED MANUSCRIPT
starting from frozen cells. Upon thawing, cells were characterized in house by assessing cell
morphology, cell growth kinetics curve, and absence of mycoplasma. The human cell lines
purchased from accredited biological resource centers (i.e. ATCC) have been originally
authenticated and characterized directly by the providers (short tandem repeats profiling). All the
experiments have been then performed using cells within 6-8 passages (in the case of human
tumor cell lines) or 2 passages (in the case of C3H10T1/2 cells) since thawing from an internal
cell bank.
2.2.3. Gli-luciferase reporter assay. Shh-Light II cells, stably incorporating the Gli-luc reporter
and the prL-TK Renilla reniformis control sequences, were seeded in three 96-well plates at a
density of 4·104 cells/well. Four days later, they were washed once with PBS (calcium-
magnesium free) and incubated for 30 h in 0.5% FBS supplemented medium containing the
reference agonist SAG (100 nM), in the presence or in the absence of test compounds (at
different concentrations, ranging from 1 to 10000 nM). After incubation, the medium was
removed and plates were sealed and frozen at -80 °C. A few days later, plates were thawed and
firefly luciferase (expressed under control of a Gli-responsive promoter) and Renilla luciferase
(expressed under control of a constitutive promoter) were measured by means of the Dual-Glo
Luciferase Assay System (Promega, Madison, WI), according to the manufacturer’s instructions,
using a microplate reader (Victor2, Perkin Elmer). Percentage of inhibition of the SAG-induced
Gli-dependent luciferase activity was ultimately calculated with respect to DMSO-treated
(control) cells and values were plotted using GraphPad Prism (version 5.02) to determine the
IC50 ± SE values. The compounds did not modify significantly the Renilla activity at 10 µM.
2.2.4. Alkaline phosphatase (ALP)-based cell differentiation assay. A well-suited ALP assay,
enabling to assess the effect of putative Smo inhibitors on differentiation of the mesenchymal
pluripotent C3H10T1/2 cells into osteoblasts, was used. C3H10T1/2 cells were seeded into 96-
well plates (3·103 cells/well) in complete growth medium. Twenty-four h later, medium was
replaced by one supplemented with 2% FBS and containing 0.1 µM SAG as cell differentiation
inducer, in the presence or the absence of increasing concentrations of test compounds. Cells
were incubated for additional 6 days, then washed once with PBS and lysed by means of a buffer
solution (0.1% Triton-X 100, in 150 mM NaCl and 50 mM Tris-HCl, pH 9.5). ALP activity (that
is a typical osteoblast marker) was ultimately quantified by adding 10 µL of cell lysates to 50 µL
of the CDP-Star chemiluminescence substrate with Sapphire-II Enhancer (molecular probes -
Thermo Fisher Scientific), and measuring the resulting chemiluminescence by means of a
multiplate reader. Results were expressed, after normalization with respect to protein content of
9

ACCEPTED MANUSCRIPT
cell lysates, as percent residual ALP activity compared to cells induced only by SAG. Data were
finally plotted and the IC50 ± SE values were calculated using GraphPad Prism software.
2.2.5. Proliferation assay. Antiproliferative activity was assessed on two human tumor cell lines:
LS180 (colon adenocarcinoma) and HT1080 (fibrosarcoma). Cells were seeded in triplicate in
96-well tissue culture plates in complete medium, and 24 h after seeding they were exposed to
increasing concentrations of the test compounds for 72 h. Inhibition of cell proliferation was
ultimately assessed by the sulforhodamine B (SRB) assay. The drug potency was evaluated by
GraphPad Prism software (version 5.02) and defined as IC50 (drug concentration required for
50% inhibition of cell survival) ± SE values.
2.2.6. To evaluate possible off-target effects, 5 has been recently characterized for its activity on
a panel of 45 kinases (the Express Diversity Kinase Profile, a 45 kinase activity-based assay
utilizing TR-FRET technology, provided by Eurofins Pharma Discovery Services).
2.2.7. Ethics statement. In vivo studies were performed in accordance with the “Directive
2010/63/UE” on the protection of animals used for scientific purposes, made effective in Italy by
the Legislative Decree 4 March 2014, n. 26, and applying the principles of 3Rs (namely, replace,
reduce, refine). Mice were purchased at Harlan Laboratories (Udine, Italy). All procedures
performed on the animals were approved by Animal Welfare Body and authorized by the Italian
Ministry of Health, 46/2014-PR. At the end of the treatment period and before necropsy, mice
were euthanized by compressed CO2 gas in cylinder as indicated in the American Veterinary
Medical Association Panel on Euthanasia and according to the United Kingdom Co-ordinating
Committee on Cancer Research guidelines (UKCCR 1998).
2.2.8. In vivo xenograft models. In vivo experiments on solid tumor models were carried out
using 5-6 week-old female athymic nude mice (Harlan Laboratories). Mice were maintained in
laminar flow rooms with constant temperature and humidity according to the NIH guidelines.
The human tumor xenograft model LS180 (derived from colon carcinoma) was used for
antitumor activity study. Exponentially growing tumor cells were s.c. inoculated (5·106/mouse)
in the right flank of mice. Groups of eight mice for each experimental group were employed to
assess antitumor activity. Drug treatments started 6 days after tumor injection. The compounds
were given daily for five days per week for 2 weeks. Tumor growth was followed by
measurement of tumor diameter with a Vernier caliper. Tumor volume (TV) was calculated
using the formula: TV (mm3) = [d2·D]/2, where d and D are the shortest and the longest
diameter, respectively. The efficacy of the drug treatment was assessed as: TV inhibition
percentage (TVI%) in treated versus control mice, calculated as: TVI% = 100 – (mean TV
treated/mean TV control · 100). When tumors reached a volume of about 1200 mm3, mice were
10

ACCEPTED MANUSCRIPT
sacrificed by cervical dislocation. To examine the possible toxicity of treatment, body weight
was recorded throughout the study.
2.2.9. P-glycoprotein 1 assay. To evaluate the interaction with P-glycoprotein 1 (P-gp), the
luminescent P-gp-GloTM ATP-ase Assay (Promega) was used with recombinant human P-gp in
a cell membrane fraction. The assay was carried out in quadrupled 96-well plate, according to
assay’s instruction. Hh inhibitors (10 µM) were tested in comparison to the reference compound
verapamil (200 µM). The values were expressed as amount of ATP consumed and fold
stimulation.
2.2.10. Caco-2 permeability assay. Caco-2 cell line derived from a human colorectal carcinoma
were used. Caco-2 cells (ATCC) were cultured for 20 days in DMEM, media was changed every
two or three days. The cells were used between passage numbers 40-60 and seeded on to
Millipore Multiscreen Caco-2 plates (1·105 cells/cm2). On day 20, the monolayers were prepared
by rinsing both basolateral and apical surfaces twice with Hank’s Balanced Salt Solution (HBSS)
at 37 °C. Cells were then incubated with HBSS (atmosphere of 5% CO2 with a relative humidity
of 95% at 37 °C) in both apical and basolateral compartments for 40 min to stabilize
physiological parameters. HBSS was then removed from the apical compartment and replaced
with test compound solutions for apical to basolateral (A-B) permeability assessment. The
fluorescent integrity marker lucifer yellow was also included in the testing solution. Analytical
standards were made from testing solutions. The apical compartment inserts were then placed
into companion plates containing fresh HBSS. For basolateral to apical (B-A) permeability
determination, the experiment was initiated by replacing buffer in the inserts, then placing them
in companion plates containing dosing solutions. The companion plate was removed at 120 min,
and apical and basolateral samples were diluted for analysis. Test compound permeability was
assessed in duplicate. The A-B and B-A permeability coefficient for each compound (Papp) and
the influx/efflux ratio (I/E ratio) were calculated. Compounds of known permeability were run as
controls on each plate. The Papp A-B and B-A values of test compounds were compared to those
of control compounds atenolol and propranolol which have human absorption of approximately
50 and 90%, respectively. The MS/MS analysis for Hh inhibitors resulted in 573.25, 556.27,
538.28, and 566.32 amu for 3, 4, 5, and 6, respectively, with positive ionization and collision
energy of 40 eV for 6 and 30 eV for the other compounds.
2.2.11. CYP450 isoforms assay. To investigate the potential drug-drug interaction capabilities of
the compounds, the effects of the drugs on CYP1A2, CYP2C9, CYP2C19, CYP2D6, and
CYP3A4 activity were examined using P450-GLO-TM assay kits (Promega) according to the
manufacturer’s instructions. The positive controls (inhibitors of enzyme activity) included α-
11

ACCEPTED MANUSCRIPT
naphthoflavone (1 µM) for CYP1A2, sulfaphenazole (10 µM) for CYP2C9, troglitazone (10
µM) for CYP2C19, quinidine (1 µM) for CYP2D6, and ketoconazole (5 µM) for CYP3A4. The
compounds were studied using recombinant CYPs, with respect to their capacity to inhibit the
formation of luminescent metabolites from CYPs specific luminogenic (proluciferine) substrates.
Different concentrations of the compounds (1, 5 and 10 µM), together with reference
compounds, positive and negative (blank) controls, were tested in two independent experiments.
The assay was performed by incubating a luminogenic CYP substrate with a CYP enzyme and
NADPH regeneration system. The luminogenic P450-GloTM substrates are derivatives of beetle
luciferin [(4S)-4,5-dihydro-2-(6-hydroxybenzothiazolyl)-4-thiazole carboxylic acid], a substrate
of firefly luciferase. Light production was evaluated at a luminometer (Glomax-Promega), and
values were displayed as Relative Light Unit (RLU). The net CYP-dependent luminescence was
calculated by subtracting the average luminescence of the blank from that of the CYP-containing
reactions. The percent inhibition of test compound versus the positive control (total activity) was
also calculated.
2.2.12. In vitro metabolism in mouse hepatocytes. Cryopreserved mouse hepatocytes were
thawed at 37 °C, centrifuged at rt (50g for 3 min) and re-suspended at a final concentration of
3·105 cells/mL. The viability of the hepatocytes was determined using trypan blue exclusion.
Test compounds and reference (7-ethoxy coumarin, 7-ETC) were incubated at 1 µM with a
hepatocyte suspension in 24 wells at 37 °C in a CO2 incubator. Test compounds in buffer (Krebs-
Henseleit) and in inactivated hepatocyte suspensions were also included as negative controls. A
50 µL aliquot of the incubation suspension was taken from each well at 0, 5, 15, 30, 60, and 120
min, quenched with methanol or acetonitrile (150 µL), centrifuged at 13000 rpm for 3 min, and
analysed by HPLC/MS/MS, using Phenomenex Onyx C18 100 x 4.6 mm chromatographic
column and AB-SCIEX API3000 tandem mass spectrometry. The extracts were assayed against
a calibration curve ranging from 50 to 1000 ng/mL for all compounds. Run time was
approximately 3.5 min. The data were acquired and processed by the data acquisition system
Analyst, version 1.5.2. Each time-point was performed in triplicate. The intrinsic clearance
(CLint) was calculated using the half-life (t1/2) approach. The t1/2 and the CLint were determined
from the concentration remaining at the different sampling points.
2.2.13. Metabolic stability in mouse plasma. Stability of each compound in murine plasma (200
ng/mL) was investigated at 37 °C. Compounds were tested in triplicate and the samples were
collected at 0, 15, 30, 45, 60, and 120 min. The same compounds in deproteinized plasma were
used as metabolic stability controls. Collected samples were deproteinized by adding acetonitrile
12

ACCEPTED MANUSCRIPT
and analyzed by HPLC/MS/MS in the same conditions described for the in vitro metabolism in
mouse hepatocytes.
2.2.14. Pharmacokinetic study. CD1 male mice were treated orally with a single dose of the
compounds at 200 mg/kg using 20% solutol HS 15 in water. Blood samples were collected at
0.25, 0.5, 1, 2, 4, 6, and 24 h post treatment from 4 animals per time point. Levels of the
compounds were determined in plasma by quantitative HPLC-MS/MS having a method range
from 50 to 1000 ng/mL for all compounds. The PK parameters Cmax (maximum plasma
concentration), Tmax (time of maximum plasma concentration), Clast (last quantifiable
concentration), Tlast (time of last quantifiable plasma concentration), AUClast (area under the
concentration versus time curve from 0 to Tlast), AUCINF (area under the concentration versus
time curve from 0 to infinity), t1/2 (terminal half-life), CL/F (apparent systemic clearance), and
Vz/F (apparent terminal volume of distribution), were derived from the analyte plasma
concentration versus time data according to a model-independent approach for sparse data
sampling by using Phoenix WinNonlin software, version 6.3 (Pharsight, Certara, Princeton, NJ).
Statistical analysis. Data are expressed as the mean ± SD or S.E.M. Statistical analysis was
performed using Mann-Whitney’s test. A P-value of ≤ 0.05 was considered statistically
significant.
3. Results
3.1. Medium scale synthesis of the compounds
A medium-scale (5-7 g), more sustainable synthesis of the acylthiourea (6) and acylguanidine (3-
5) derivatives was developed for the first time to allow for a possible large scale-up and for the in
vivo studies. A convergent approach seemed to be the most valuable for this purpose, coupling a
proper functionalized aniline (A) with acyl chloride (20) or acyl cyanamide (21) to obtain
acylthiourea and acylguanidine derivatives, respectively (Figure 3). Moreover, a classical amide
synthesis, using a properly functionalized acyl chloride (B) and the nitro aniline 12, was planned
for the synthesis of aniline derivative A.
13

ACCEPTED MANUSCRIPT
O
S
O
O
NH
O
MeO
MeO
MeO
MeO
N
N
N
N
H
N
N
H
H
H
H
H
R'
R'
6
3-5
OMe
OMe
O
O
O
MeO
MeO
MeO
MeO
Cl
NHCN
H2N
N
H
R'
OMe
OMe
20
(A)
21
O
Cl
O2N
NH2
R'
12
(B)
Figure 3. Retrosynthetic approach for the synthesis of the studied acylguanidines and
acylthioureas.
The acyl chlorides (B) required for the synthesis of 3, 4, and 6 are commercially available, while
the preparation of a suitable chloride was required for 5. Previously, we have applied a Wittig
reaction to obtain a small-scale amount (about 100 mg) of 10 [17]. However, while this reaction
was a good opportunity for the early hit identification/optimization steps, the same protocol was
not suitable for the large-scale synthesis of this compound as the reaction lacked for poor atom
economy (i.e. production of 1 equivalent of triphenylphosphine oxide as by-product, and use of
dangerous bases such as BuLi in overstoichiometric amounts), long reaction times, change in
reaction temperature (from -78 °C to rt), and low yield.
14

ACCEPTED MANUSCRIPT
Cl
O
CO2H
HO2C
N
HO
OH
Pd
Cl
9
I
ClCOCOCl
DMF, CH2Cl2
2
7
DMF, Et3N
+
3 h, rt
MW, 110 °C, 3 x 20 min
>99%
8
10
11
O2N
H2N
H2N
HCl
O
O
HN
HN
HN
O
H2, Pd/C, MeOH, 12
HCl, MeOH,
12 h, rt
h, rt
O2N
NH2
98%
12
Py, CH2Cl2, 12 h, rt
61%
13
14
15
Scheme 1. Synthesis of 15.
For these reasons, the use of a more sustainable catalytic protocol involving an Heck coupling
between 4-iodobenzoic acid 7 and styrene 8 in the presence of freshly prepared Pd catalyst 9
[26], in DMF and Et3N taking advantage from MW irradiation at 110 °C for 20 min, was
investigated (Scheme 1). The irradiation was repeated for three times. The conversion into 10
was finally quantitative and the product was directly crystallized from H2O. To avoid side
reactions and by-products during the synthesis of 11, a purification of 10 by flash
chromatography demonstrated to be a mandatory step for removing Pd traces responsible, in
medium-scale, for aniline 12 demethylation during the coupling. A summary of the advantages
in terms of sustainability of the developed Heck coupling is reported in Table 2 [27].
The acid 10 was transformed into its acyl chloride derivative (11) by treatment with oxalyl
chloride in the presence of a catalytic amount of DMF and directly reacted with 12 in the
presence of pyridine as the base. The nitro derivative 13 was obtained in 61% yield after a simple
acid-base extraction during the work-up. In the small-scale protocol already reported [16], a
MW-assisted hydrogenation reaction using ammonium formiate as the hydrogen source in the
presence of Pd/C was applied to obtain 14. Nevertheless, this protocol failed in medium-scale
preparation because of several security issues: i) the reaction mixture is completely
heterogeneous as none of the reactants, reagents, and the catalyst (13, HCO2NH4 and Pd/C,
15

ACCEPTED MANUSCRIPT
respectively) are soluble in MeOH; ii) the reaction is really exothermic, and iii) working on a
500 mg scale, an explosion occurs after 30 s MW irradiation. For these reasons, we finally
decided to simultaneously reduce both the nitro group and the double bond by treatment of 7
with 10% Pd/C in MeOH directly using the more atom economic H2 atmosphere (1 atm) at rt
overnight. The expected product, obtained in quantitative yield, was directly transformed into the
hydrochloric salt 15 for the preparation of the corresponding acyl guanidine.
A similar approach, starting from commercially available biphenyl-4-carbonyl chloride (16a)
and 4’-fluorobiphenyl-4-carbonyl chloride (16b), was successfully applied to the synthesis of
anilines 18a-b and their hydrochloric salts 19a-b, as reported in Scheme 2.
Table 2. Higher sustainability of the Heck coupling in comparison to a Wittig reaction for the
synthesis of the studied compounds.
Parameters
Wittig reaction
THF
(20 mL for 2 mmol SMa
Heck coupling
DMF
Solvent
(15 mL for 12 mmol SM)
or 180 mL for 12 mmol SM)
Et3N
BuLi
(1.4 equiv = 1.7 g for 12 mmol SM)
Pd cat.
Additives
(3.5 equiv = 448 mg for 2 mmol SM
or 2.6 g for 12 mmol SM)
(0.01 equiv = 78 mg for 12 mmol SM)
POPh3
(1 equiv = 556 mg for 2 mmol SM
or 3.4 g for 12 mmol SM)
LiBr
Et3NHI
By-products
(1 equiv = 2.7 g for 12 mmol)
(1 equiv = 166 mg for 2 mmol SM
or 996 mg for 12 mmol SM)
-78 °C (Cryostate) and rt
1 h + 12 h
Temperature
Time
110 °C (MW)
1.5 h
DMF (+)
Flammable reagents
Purification
BuLi (+++), THF (+)
Extraction (AcOEt), column
70%
Precipitation (1N HCl), column
>99%
Yield
aSM = Starting material.
16

ACCEPTED MANUSCRIPT
O2N
H2N
HCl
H2N
O
Cl
HN
O
HN
O
HN
HCl, MeOH
12 h, rt
O
O2N
NH2
H2, Pd/C, MeOH
12
12 h, rt
Py, CH2Cl2,12 h, rt
17a-b
18a-b
19a-b
16a-b
a: R = H
b: R = F
R
R
R
R
Scheme 2. Synthesis of 19a-b.
The final thiourea 6 was obtained by treatment of 18b with ammonium thiocyanate and
trimethoxy benzoyl chloride 20 in acetone at reflux for 3 h. A white solid (86% yield) resulted
after crystallization (Scheme 3).
Aniline hydrochlorides 15 and 19a-b reacted with acylcyanamide 21 in toluene at reflux to give
the acylguanidines that were finally transformed into their hydrochloric acid salts 3-5. All these
derivatives were synthesized in 5 to 7 g scale with purity >97%.
The protocol developed demonstrates to be effective for the large-scale synthesis of both
acylthiourea and acylguanidine compounds, avoiding the risk of using procedures that are known
to lack for sustainability and safety issues.
O
S
O
O
MeO
MeO
MeO
MeO
N
N
N
H
NH4SCN, 18b
Cl
H
H
acetone, reflux, 3 h
6
OMe
OMe
F
20
NH2CN
NaOH, H2O
Et2O, 1 h, rt
O
H2N
N
O
NH
O
O
H
HCl
MeO
MeO
R
N
N
MeO
MeO
N
15 or 19a-b
NHCN
H
H
H
HCl
R
toluene, reflux, 3 h
then HCl, MeOH, 12 h, rt
3: R = pF-Ph
4: R = Ph
OMe
OMe
21
5: R = -CH2-CH2-Ph
Scheme 3. Synthesis of the final compounds.
17

ACCEPTED MANUSCRIPT
3.2. In vitro pharmacology
In vitro treatment with the compounds resulted in Hh signaling inhibition and decreased viability
of colon cancer and fibrosarcoma cell lines.
To assess whether 3-6 could inhibit the Hh signaling pathway, a Gli-luciferase gene reporter
assay and a cell differentiation assay were performed including 1 as positive control. Although a
similar evaluation of 3-5 were previously reported by us [16, 17, 21], the Gli-luc and the alkaline
phosphatase-based assays (see below) were repeated to collect homogeneous biological data,
including those of the new compound 6.
The Shh-light II cell line, a clonal NIH-3T3 cell line expressing Gli-responsive firefly luciferase
and constitutive Renilla luciferase reporters, was stimulated for 30 h with 0.1 µM SAG in the
presence or absence of test compounds at various concentrations (from 0.001 to 10 µM). Overall,
the compounds inhibited the Gli-responsive reporter activity induced by SAG in a dose-
dependent manner, with IC50 values ranging from 20 to 100 nM (Table 3, Figure 4).
Compound 5 resulted the most active and its inhibitory activity (IC50 = 0.02 µM) was also
stronger than that of 1 (IC50 = 0.05 µM).
Figure 4. Inhibition of SAG-induced Gli-dependent luciferase activity in Shh-Light II cells by
compounds 3-6 in comparison to 1 used as reference compound. Inhibition curves were
generated using increasing concentrations of compounds in the presence of SAG (0.1 µM). The
values are expressed as a percentage of the residual luciferase activity, and are the means ± S.E.
of triplicates. The data shown are representative of 2 independent experiments.
To confirm the data, a second orthogonal assay was performed on the mesenchymal pluripotent
C3H10T1/2 cells that, once induced with Hh ligands, differentiate into ALP-positive osteoblasts
[28, 29]. Exposure of these cells to the compounds for 6 days suppressed, in a dose-dependent
18

ACCEPTED MANUSCRIPT
manner, ALP activity evoked by SAG (Table 3, Figure 5). As in the previous reporter assays, 5
(IC50 = 0.04 µM) was shown to be the most active compound.
Figure 5. Inhibition of ALP activity in SAG-induced C3H10T1/2 cells by the compounds in
comparison to 1 used as reference compound. Inhibition curves were generated using increasing
concentrations of compounds in the presence of SAG (0.1 µM). The values are expressed as a
percentage of the maximal response induced by SAG, and are the means ± S.E. of triplicates.
The data shown are representative of two independent experiments.
Figure 6. Inhibition of LS180 (left) and HT1080 (right) cell proliferation upon a 72 h treatment
with the compounds in comparison to 1 used as reference compound. Results, expressed as
tumor cell proliferation percent with respect to DMSO-treated (control) cells, are the means ±
S.E. of triplicates.
In order to test the ability of the compounds to inhibit growth and viability of tumor cells, LS180
human colon carcinoma and HT1080 human fibrosarcoma cell lines were used in SRB assay,
after 72 h incubation (Figure 6). All the compounds potently inhibited cell proliferation of both
cell lines, in a dose-dependent manner. Anti-proliferative activity was 5-18 and 3-50 fold better
than that of 1 against LS180 and HT1080 cells, respectively. Consistently with previous data, 5
19

ACCEPTED MANUSCRIPT
resulted the most active compound and showed a similar activity against both cell lines (IC50
values of 2.5 and 2.0 µM, respectively). These activities were at least three-fold higher than that
of the other compounds, including 1.
Overall in vitro pharmacology data reported in Table 3 confirm that the compounds significantly
inhibited Hh signaling pathway and tumor cell proliferation with 5 being the most active
compound.
Moreover, evaluation of 5 on a panel of 45 kinases resulted in percent inhibition of control
values lower than 15% in all cases, thus demonstrating that 5 did not affect significantly the
activity of these targets (no off-target effects).
Table 3. Summary of comparative effects of the compounds and the reference Smo inhibitor 1 in
different cell-based assays.a
Tumor Cell Viabilityb
Luciferase activity
(Shh-Light II cells)c
ALP activity
Compound
(C3H10T1/2 cells)d
LS180
HT1080
45 ± 5
>100
0.05 ± 0.01
0.05 ± 0.01
0.07 ± 0.01
0.02 ± 0.01
0.10 ± 0.01
0.05 ± 0.01
0.07 ± 0.01
0.06 ± 0.01
0.04 ± 0.01
0.67 ± 0.02
1
3
4
5
6
8.1 ± 1.0
9.7 ± 1.2
2.5 ± 0.3
7.5 ± 1.1
20 ± 1
29 ± 1
2.0 ± 0.5
27 ± 1
aValues are reported as IC50 ± SE (expressed in micromolar concentrations) and were calculated
as mean of at least two experiments.
bInhibition of tumor cells proliferation, by SRB assay; upon a 72 h treatment.
cInhibition of the Gli-luciferase reporter activity in Shh-Light II cells after treatment with various
concentrations of the compounds for 30 h. The firefly luciferase activity (expressed under control
of a Gli-responsive promoter) and the Renilla luciferase activity (expressed under control of a
constitutive promoter) were measured using the Dual-Glo Luciferase Assay System. Percentage
of inhibition of the SAG-induced Gli-dependent luciferase activity was ultimately calculated
with respect to DMSO-treated (control) cells.
dInhibition of ALP activity induced by SAG (0.1 µM) in C3H10T1/2 cells. Cells were treated
with various concentrations of the compounds for 6 days.
3.3. In vivo activity
The oral administration of 5 inhibited the tumor growth of a colon carcinoma xenograft cancer in
nude mice. The LS180 colon cancer xenograft model was used to explore the effect of the
compounds on Hh-dependent tumors in vivo. LS180 tumor cells secrete Hh ligands, thus leading
20

ACCEPTED MANUSCRIPT
to enhanced Gli1 transcription. Mice bearing established LS180 tumors were administered by
oral route with 200 mg/kg of test compounds (3-5) daily (5 days/week) for 2 weeks and tumor
growth was monitored. Only 5 significantly inhibited tumor growth by 48% compared to
vehicle-treated controls (P<0.01, Mann-Whitney’s test) (Figure 7). No toxicity in terms of body
weight loss was observed.
Figure 7. In vivo tumor growth inhibition induced by the compounds in comparison to 1 used as
reference compound. LS180 colon cancer cells were injected in nude mice and 6 days after
tumor implantation, the test compounds were delivered by oral route. Tumor lesions were
measured by Vernier caliper.
3.4. In vitro ADME parameters and pharmacokinetics of the compounds orally
administered in mice
3.4.1. P-glycoprotein 1 assay. A linear standard curve of ATP ranging from 37.5 to 300 nM was
built to calculate P-gp ATPase activity through ATP consumption. Compound 3 and 6 showed a
low effect on P-gp ATPase activity (fold induction of 3.0 and 2.3, respectively, Table 4), while 4
appeared to be a potential substrate of P-gp (fold induction of 4.8). Compound 5 did not have
any interaction with the P-gp. Verapamil, used as reference compound, induced a 8.5-fold
stimulation of P-gp ATPase activity.
21

ACCEPTED MANUSCRIPT
Table 4. In vitro effect of the compounds on P-gp.
ATPa
Fold stimulationb
Compd
Basal activity
13
30
2.3
4.8
1.4
3.0
8.5
3
63
4
18
5
39
6
verapamil
111
aP-gp ATPase activity was expressed as picomoles of ATP consumed per µg P-gp per minute
and was calculated converting RLU value to ATP concentrations by a standard curve of ATP.
bFold stimulation was the ratio of ATP over basal value.
3.4.2. Caco-2 Permeability assay. Influx/Efflux (I/E) ratio of the compounds in Caco-2 colon
cancer cells is shown in Table 5. Compound 6 was not detectable in receiver compartment with a
high value of starting concentration. Compounds 3 and 5 had an I/E ratio ~1, suggesting that they
could not affect the human absorption. Compound 4 showed a low recovery that could be
indicative of non-specific binding issues. However, its I/E ratio <2 suggested a low potential to
be a P-gp substrate. Atenolol (paracellular transport) and propranolol (passive transcellular
transport) with known human absorption of 50% and 90% respectively, as well as talinolol (a P-
gp substrate), were used as reference compounds to confirm that tumor cells were expressing
functional influx/efflux proteins.
Table 5. Evaluation of bi-directional permeability of Hh inhibitors.
Compd
Papp A-Ba
0.366±0.044
0.428±0.063
0.701±0.166
% recovery
33.0
Papp B-Aa
0.370±0.006
0.651±0.016
0.737±0.189
% recovery I/E ratiob
30.6
31.9
44.6
42.9
1.01
1.52
1.05
3
4
5
6
16.5
41.1
65.4
atenolol
propranolol
talinolol
0.316±0.045
30.9±0.7
94.7
70.0
94.8
0.721±0.018
21.1±2.3
102
89.0
88.6
2.28
0.68
78.2
0.162±0.020
12.6±0.3
22

ACCEPTED MANUSCRIPT
aValues are the mean ± SD of two replicates, expressed in 10-6·cm·s-1. Atenolol (paracellular
transport), propranolol (passive transcellular transport), and talinolol (P-gp substrate) were used
as reference compounds.
bI/E ratio = (mean Papp B-A/mean Papp A-B).
3.4.3. CYP450 isoform assay. Except 6, all the compounds interfered (as competitive substrates or
inhibitors) with the activity of CYP1A2 and CYP3A4, sometimes in a concentration-dependent
manner. All the compounds interfered with the activity of CYP2C9 and CYP2C19 without a
clear dependency on their concentration. Finally, none of the compounds interfered significantly
with CYP2D6. Interfering effects (%) are shown in Table 6 as mean (±S.E.M) of triplicate
assays.
Table 6. Interfering effects (% ± S.E.M.) of the studied and reference compounds on cytochrome
CYP450 enzymes.a
CYP
1A2
5
3A4
5
2C9
5
(µM)
1
10
1
10
1
10
46±1
49±1
29±2
2±3
58±2
64±2
43±1
5±4
59±2
64±2
40±2
1±2
52±9
38±8
76±3
4±11
78±5
76±3
89±2
7±15
2D6
5
-82±8
-87±3
-92±2
15±21
63±1
43±3
68±1
63±1
76±1
71±0
81±0
69±1
79±2
74±3
83±1
71±0
3
4
5
6
CYP
2C19
5
(µM)
1
10
1
10
59±2
57±1
67±1
25±2
76±0
79±0
81±1
33±1
76±1
81±1
83±1
27±2
20±4
19±4
15±5
18±9
25±5
29±2
12±4
1±8
21±5
23±5
11±3
6±5
3
4
5
6
aReference compounds. CYP1A2: α-naphthoflavone (1 µM, 98%); CYP3A4: ketoconazole (5
µM, 96%); CYP2C9: sulfaphenazole (10 µM, 98%); CYP2C19: troglitazone (10 µM, 84%);
CYP2D6: quinidine (1 µM, 91%).
3.4.4. In vitro metabolism in mouse hepatocytes. Compound 3 and 4 showed chemical instability,
so the CLint could not be calculated. By contrast, 5 had a CLint of 756 mL/min/kg with a t1/2 value
of about 8 min (Table 7). The reference compound, 7-ETC, revealed a t1/2 of about 5 min and a
CLint corresponding to 1260 mL/min/kg, as described in the literature.
23

ACCEPTED MANUSCRIPT
Table 7. In vitro mouse hepatocytes metabolism.
Compounda
t1/2 (min)
CLint
CLint
(mL/min/kg)
(µL/min/106 cells)
8.07
4.84
286
756
5
7-ETC
477
1260
aThe compounds and 7-ETC (reference compound) were incubated at the concentration of 1 µM
with mouse hepatocytes. ND = not determined.
3.4.5. Metabolic stability in murine plasma. The compounds showed different metabolic stability
in murine plasma at 37 °C up to 120 min. Compound 6 was highly metabolized, while 3 and 5
were moderately metabolized. Conversely, 4 was metabolically stable (Figure 8). When
incubated in deproteinized plasma, the recovery of 4 was about 83%, the same value observed
after incubation with mouse plasma, thus indicating chemical instability instead of metabolism.
All the other compounds were stable when incubated over 120 min in deproteinized plasma at 37
°C.
Figure 8. Percent values of 3-6 recovered after a 120 min incubation at 37 °C in mouse plasma.
Noteworthy, when incubated in deproteinized plasma over 120 min at 37 °C, the recovery of 4
was about 83%, the same observed after incubation with fresh mouse plasma. For Vismodegib
plasma stability see Ref. 30.
3.4.6. In vivo oral pharmacokinetics in mouse. The PK profile in healthy mice was determined
after a single dose of 200 mg/kg of each compound administered by oral route. Compound 6 was
sporadically quantified in mouse plasma (2 out the 32 samples collected). The remaining
compounds were detected in all the plasma samples collected up to last collection time (24 h).
Compounds 3 and 5 showed multiple peaks in plasma profiles, the highest of which occurred at
24

ACCEPTED MANUSCRIPT
1.0 and 4.0 h, respectively. Conversely, 4 was rapidly absorbed and showed a Cmax of about 30
µg/mL 2 h after dosing. Subsequently, its plasma concentration declined according to a mono-
exponential profile (Figure 9).
Figure 9. Mean (± S.E.M) plasma concentration versus time profiles (log-linear scale) of 3-6 in
CD-1 male nude mice receiving a 200 mg/kg single oral dose of each test compound. For
Vismodegib in vivo stability see Ref. 30.
All compounds were cleared from plasma with a t1/2 of about 2.5 h. Compounds 3 and 4 showed
similar CL/F (893 and 843 mL/h/kg), lower than that of 5 (1460 mL/h/kg). Accordingly, the
systemic exposur of 3 and 4 was higher than that of 5. PK parameters are summarized in Table 8.
Table 8. Pharmacokinetic parameters of the studied compounds determined after a 200 mg/kg
single dose to CD-1 male nude mice.a
b
b
Cmax
AUClast
Compd
Tmax Clast Tlast
AUCINF t1/2 CL/F Vz/F
17.3±2.3
29.6±3.9
11.0±2.2
1
2
4
90.7 24
74.2 24
50.7 24
223±31
237±17
137±23
223970 2.6
237226 2.4
893
843
3323
2963
5280
3
4
5
136943 2.5 1460
aAbbreviations used. Cmax: maximum plasma concentration, in µg/mL; Tmax: time of Cmax, in h;
Clast: last quantifiable concentration, in µg/mL; Tlast: time of Clast, in h; AUClast: area under the
concentration versus time curve from 0 to Tlast, in µg·h/mL; AUCINF: area under the
concentration versus time curve from 0 to infinity, in µg·h/mL; t1/2: terminal half-life, in h; CL/F:
apparent systemic clearance, in mL/h/kg; Vz/F: apparent terminal volume of distribution, in
mL/kg.
25

ACCEPTED MANUSCRIPT
bMean ± S.E.M.
4. Discussion
A preliminary in vitro investigation on acylthiourea and acylguanidine derivatives showed their
ability to impair Smo activity at nanomolar concentrations also toward Smo mutants resistant to
already known antagonists [17]. These results prompted us to carry out a preliminary non-
clinical characterization of their in vitro ADME and in vivo pharmacokinetic profile, and to
eventually prioritize the best compound eligible for further clinical development.
For this purpose, the medium-scale synthesis of 3-6 was for the first time approached, taking also
into account sustainable and safety issues for a future scale-up of the overall process. All the
compounds were obtained in large quantity (5-7 g) with very high purity, using the more
sustainable protocols developed so far for the synthesis of this family of compounds.
In addition, we further support the hypothesis that these compounds, in particular 5, selectively
suppressed the Hh signalling pathway activity by potentially targeting the critical component
Smo and consequently inhibiting the Hh-dependent cancer growth.
In vitro ADME properties and in vivo pharmacokinetic studies in mouse, after a single oral dose,
were investigated. In vitro studies suggested that none of the compounds were potential P-pg
substrates. However, they underwent extra-hepatic (6 only) or moderate hepatic metabolism.
Except 6, the compounds were potential substrates and/or inhibitors of CYP1A1 and CYP3A4,
all interfered with the activity of CYP2C9 and CYP2C19, and none of them with the activity of
CYP2D6. The in vitro ADME profiles were in good agreement with the in vivo pharmacokinetic
profiles observed in nude mice after a single oral 200 mg/kg dose of each compound. Compound
6 showed an extremely low bioavailability, probably due to its extensive extra-hepatic
metabolism, rather than to an incomplete absorption. Compounds 3-5 showed similar plasma
profiles in nude mice having a terminal half-life of about 2.5 h. However, 5 had a lower systemic
exposure compared to 3 and 4, possibly because of its higher hepatic metabolism and/or larger
volume of distribution (assuming a similar bioavailability as argued by in vitro tests).
Considering that 5 effectively inhibited in vivo tumor growth of colorectal cancer LS180, thus
indicating that such a compound is distributed outside the systemic circulation to reach the target
tissues, we hypothesize that a large volume of distribution rather than a high clearance could be
responsible for its low systemic exposure.
Moreover, 5 effectively inhibited tumor growth without affecting body weight, and significantly
suppressed the Hh signalling pathway in tumor lesions.
26

ACCEPTED MANUSCRIPT
In the present study, we identified 5 as a strong inhibitor of the Hh pathway activity by targeting
Smo. Using a colon cancer model overexpressing the Hh ligand, we also demonstrated that the
same compound significantly impaired Hh-dependent tumor growth by inhibiting Hh pathway
activity. The ability of this compound to inhibit Smo activity also in mutant cells resistant to 1
and other known Smo inhibitors, could give support to the future usage of 5 for treating cancers
driven by aberrant Hh pathway. Given that 5 showed a good PK and metabolic profile, it could
be further investigated (alone or in combination with other treatment strategies) for a clinical
efficacy on patients with tumors dependent on the Hh pathway disregulation.
5. Disclosure of potential conflicts of interest
LV, FMM, MAS, SP, FrM, CT, RDS, and GG are ALFASIGMA employees. The remaining
authors disclosed no conflicts of interest. All authors have approved the final form of this article.
6. Author contribution
EP planned the large scale synthesis. EC and EP synthesized the compounds, LV, FMM, RDS
coordinated the in vitro and in vivo studies. LV, FaM, MAS, EP analyzed data and wrote the
manuscript. MAS, SP, FrM and CT performed PK studies. GG coordinated the project.
7. Acknowledgements
The authors would like to thank Giuseppe Roscilli, Emanuele Marra, Fabiana Fosca Ferrara, and
Luigi Aurisicchio of Takis-Biotech s.r.l. (Castel Romano, Rome, Italy) for their support for in
vitro and in vivo studies. The Italian Ministry of Education, University, and Research (MIUR) is
also acknowledged for a Grant as Dipartimento di Eccellenza 2018-2022.
8. Appendix A. Supplementary data
Supplementary data related to this article can be found at
9. References
[1] N.S. Chari, T.J. McDonnell, The sonic hedgehog signaling network in development and
neoplasia, Adv. Anat. Pathol. 14 (2007) 344-352. DOI: 10.1097/PAP.0b013e3180ca8a1d.
[2] J. Briscoe, P.P. Therond, The mechanisms of Hedgehog signaling and its roles in
development and disease, Nat. Rev. Mol. Cell Biol. 14 (2013) 416-429. DOI: 10.1038/nrm3598.
27

ACCEPTED MANUSCRIPT
[3] S.J. Scales, F.J. de Sauvage, Mechanisms of Hedgehog pathway activation in cancer and
implications for therapy, Trends Pharmacol. Sci. 30 (2009) 303-312. DOI:
10.1016/j.tips.2009.03.007.
[4] S. Teglund, R. Toftgard, Hedgehog beyond medulloblastoma and basal cell carcinoma,
Biochim. Biophys. Acta 1805 (2010) 181-208. DOI: 10.1016/j.bbcan.2010.01.003.
[5] Y. Bilir, E. Gokce, B. Ozturk, F.A. Deresoy, R. Yuksekkaya, E. Yaman, Metastatic basal cell
carcinoma accompanying Gorlin syndrome, Case Rep. Oncol. Med. 2014 (2014) 362932. DOI:
10.1155/2014/362932.
[6] M. Guha, Hedgehog inhibitor gets landmark skin cancer approval, but questions remain for
wider potential, Nat. Rev. Drug Discov. 11 (2012) 257-258. DOI: 10.1038/nrd3714.
[7] T.K. Rimkus, R.L. Carpenter, S. Qasem, M. Chan, H.-W. Lo, Targeting the Sonic Hedgehog
signaling pathway: review of Smoothened and GLI inhibitors, Cancers 8 (2016) 22. DOI:
10.3390/cancers8020022.
[8] https://www.accessdata.fda.gov/drugsatfda_docs/label/2012/203388lbl.pdf (accessed 16 July
2018).
[9] https://www.accessdata.fda.gov/drugsatfda_docs/label/2015/205266s000lbl.pdf (accessed 16
July 2018).
[10] A. Sekulic, M.R. Migden, A.E. Oro, L. Dirix, K.D. Lewis, J.D. Hainsworth, J.A. Solomon,
S. Yoo, S.T. Arron, P.A. Friedlander, E. Marmur, C.M. Rudin, A.L. Chang, J.A. Low, H.M.
Mackey, R.L. Yauch, R.A. Graham, J.C. Reddy, A. Hauschild, Efficacy and safety of
vismodegib in advanced basal-cell carcinoma. N. Engl. J. Med. 366 (2012) 2171-2179. DOI:
10.1056/NEJMoa1113713.
[11] https://clinicaltrials.gov (accessed 16 July 2018).
[12] D. Amakye, Z. Jagani, M. Dorsch, Unraveling the therapeutic potential of the Hedgehog
pathway in cancer. Nat. Med. 19 (2013) 1410-1422. DOI: 10.1038/nm.3389.
[13] C. Danial, K.Y. Sarin, A.E. Oro, A.L. Chang, An investigator-initiated openl trial of
sonidegib in advanced basal cell carcinoma patients resistant to vismodegib, Clin. Cancer Res.
22 (2016) 1325-1329. doi: 10.1158/1078-0432.CCR-15-1588.
[14] T.W. Ridky, G. Cotsarelis, Vismodegib resistance in basal cell carcinoma: not a smooth fit,
Cancer Cell 7 (2015) 315-316. doi: 10.1016/j.ccell.2015.02.009.
[15] S.X. Atwood, K.Y. Sarin, R.J. Whitson, J.R. Li, G. Kim, M. Rezaee, M.S. Ally, J. Kim, C.
Yao, A.L. Chang, A.E. Oro, J.Y. Tang, Smoothened variants explain the majority of drug
resistance in basal cell carcinoma, Cancer Cell 27 (2015) 342-353. doi:
10.1016/j.ccell.2015.02.002.
28