Identification of a novel NAMPT inhibitor by combinatorial click chemistry and chemical refinement

Article abstract of DOI:10.1039/c5md00261c

The identification of compounds able to inhibit the NAD salvage pathway is experiencing a growing popularity as it has been proposed to be a novel target for antitumoral and anti-inflammatory drugs. In this manuscript, we used the copper-catalyzed [3+2] cycloaddition between azides and alkynes (click chemistry) to identify novel NAMPT inhibitors with a triazole-containing tail group. 720 compounds were synthesized in the first round, allowing the identification of 17 hit compounds. The second round of optimization brought about the discovery of compound 43 which displayed a cytotoxicity of 20 nM on neuroblastoma cancer cells and an inhibition of NAMPT of 114 nM.

Full text of DOI:10.1039/c5md00261c

MedChemComm
View Article Online
View Journal | View Issue
CONCISE ARTICLE
Identification of a novel NAMPT inhibitor by
combinatorial click chemistry and chemical
refinement†
Cite this: Med. Chem. Commun.,
015, 6, 1891
2
a
a
a
a
a
a
S. Theeramunkong,‡ U. Galli, A. A. Grolla, A. Caldarelli, C. Travelli, A. Massarotti,
b
b
c
a
a
M. P. Troiani, M. A. Alisi, G. Orsomando, A. A. Genazzani* and G. C. Tron*
The identification of compounds able to inhibit the NAD salvage pathway is experiencing a growing
popularity as it has been proposed to be a novel target for antitumoral and anti-inflammatory drugs. In this
manuscript, we used the copper-catalyzed [3+2] cycloaddition between azides and alkynes (click chemis-
try) to identify novel NAMPT inhibitors with a triazole-containing tail group. 720 compounds were synthe-
sized in the first round, allowing the identification of 17 hit compounds. The second round of optimization
brought about the discovery of compound 43 which displayed a cytotoxicity of 20 nM on neuroblastoma
cancer cells and an inhibition of NAMPT of 114 nM.
Received 19th June 2015,
Accepted 3rd September 2015
DOI: 10.1039/c5md00261c
www.rsc.org/medchemcomm
Nicotinamide phosphoribosyltransferase (NAMPT) is a key,
rate-limiting enzyme for the synthesis in cells of nicotinamide
for FK866 (ref. 16) and, in addition, both CHS828 and FK866
give rise to thrombocytopenia and various gastrointestinal
1
,2
16,17
adenine dinucleotide (NAD) from nicotinamide. Although
other pathways that lead to NAD synthesis exist in mammals,
NAMPT appears to be the enzyme in NAD metabolism most
frequently up-regulated in metabolic, inflammatory and
symptoms (phase I clinical trials).
A recent pre-clinical
1
8
study also suggests that the retinal toxicity may be relevant
for most NAMPT inhibitors as it appears to be an on-target
side effect. Furthermore, metabolism may also be an issue.
19
3–6
tumoral states. Indeed, while NAD is relatively long-lived in
cells as it is mainly used in redox reactions, a number of
enzymes (e.g. sirtuins, PARPs) may consume it and active cells
For this reason, the search and development of safer NAMPT
inhibitors is still a topic of interest. All the NAMPT inhibitors
1
3,14
reported to date, which were recently reviewed,
share a
7
–9
are in constant requirement of new NAD. The identification
common pharmacophoric working model: a meta or para
in 2002 of FK866 (Fig. 1), the first one-digit nanomolar
1
0,11
NAMPT inhibitor,
followed shortly after by the resolution
1
2
of the crystal structure of the enzyme bound to its inhibitor,
has opened the way for the medical exploitation of this bio-
logical target. After the disclosure of FK866, several groups
both in industry and academia have searched and found
1
3,14
small organic molecules able to inhibit NAMPT.
To date,
only two compounds, FK866 and CHS828, (Fig. 1) as its water
1
5
soluble pro-drug teglarinad, have reached clinical trials in
cancer. In vivo studies have anyway shown low bioavailability
a
Dipartimento di Scienze del Farmaco, Università degli Studi del Piemonte
Orientale “A. Avogadro”, Largo Donegani 2, 28100 Novara, Italy.
E-mail: armando.genazzani@uniupo.it, giancesare.tron@uniupo.it;
Fax: +39 0321375821; Tel: +39 0321375827
b
R&D, Angelini Research Center, Piazzale della Stazione, 00040 S. Palomba-
Pomezia, Italy
c
Department of Clinical Sciences (DISCO), Section of Biochemistry, Polytechnic
University of Marche, Via Ranieri 67, Ancona 60131, Italy
†
Electronic supplementary information (ESI) available: Materials and methods.
See DOI: 10.1039/c5md00261c
‡
Current address: Faculty of Pharmacy, Thammasat University (Rangsit Campus),
9 Phahon Yothin Rd., Khlong Luang, Pathum Thani, 12120, Thailand.
9
Fig. 1 Structures of FK866, CHS828, and GPP78.
Med. Chem. Commun., 2015, 6, 1891–1897 | 1891
This journal is © The Royal Society of Chemistry 2015

View Article Online
Concise Article
MedChemComm
Scheme 3 Coupling reaction between carboxylic acids and amines to
obtain amide compounds.
Fig. 2 The three main parts of the starting compounds: an alkyne, a
methylene chain and a carboxylic acid moiety.
With the goal of identifying novel NAMPT inhibitors,
always using click chemistry and exploiting a combinatorial
approach, we investigated the possibility of identifying
novel tail groups. To accomplish this task, a set of ω-alkyl
carboxylic acids were used as functional building blocks.
These structures contain three main parts: an alkyne, a
methylene chain and an acid moiety. The alkyne part
undergoes click chemistry reaction with different azide
derivatives to create different tail groups. The linker, five to
eight methylene units long, allows the cap group to enter the
catalytic tunnel, and finally the carboxylic acid can be
coupled with different amino pyridines forming the cap
group (Fig. 2).
The ω-alkyl carboxylic acids (8–10) were prepared starting
from the commercially available alcohols containing an
2
2
internal alkyne (1–3) using the acetylene zipper reaction,
Scheme
1
Acetylene zipper reaction of alcohol derivatives with
while for the alkyl chain with eight methylene groups,
10-undecynoic acid (7) was commercially available (Scheme 1).
The alcoholic function of 4–6 was then oxidized to carboxylic
acid under Jones conditions (Scheme 2). The yields of these
reactions were modest (25–42%) as it was not possible to sup-
press the side reaction leading to the esterification of the car-
boxylic product with the alcohol.
internal alkynes and the structure of 10-undecynoic acid (7).
The 4 carboxylic acids were then coupled with 6 commer-
cial amines containing a pyridine ring to obtain 24 different
amides (Scheme 3). The structures of amides (A–X) are shown
in Fig. 3.
Scheme 2 Oxidation of alcohols by the Jones reaction.
The resulting terminal alkyne-amides and 30 in-house
azides (11–40, Fig. 4) were used for the copperIJI) catalyzed
[3+2] cycloaddition to give 720 triazole products (Scheme 4).
All the compounds were subjected to MS analysis to verify
the presence of the desired triazoles and used as such with-
out purification after evaporation of the solvent. Given the
elevated number of compounds, the screening procedure did
not evaluate the purity by LC-MS but only the presence of the
desired compound. The 720 crude compounds were then
screened for cytotoxicity via the MTT method using SH-SY5Y
neuroblastoma cancer cells at a fixed concentration (10 μM).
This cell line was chosen as we have gathered considerable
substituted pyridine or an heterocycle containing pyridine as
a cap group, a connecting unit with a hydrogen bond accep-
tor, an alkyl chain linker and a hydrophobic tail group able to
interact with the hydrophobic amino acids at the rim of the
enzyme. In 2010, our group, using click chemistry, discovered
a one-digit nanomolar NAMPT inhibitor (EC50 for cytotoxicity
in vitro of 3.8 ± 0.3 nM and IC50 for NAD depletion of 3.0 ±
2
0
0
.4 nM on the SH-SY5Y cell line), named GPP78 (Fig. 1),
exploiting the bioisosterism between the amide of FK866 and
2
1
the 1,2,3-triazole ring.
1892 | Med. Chem. Commun., 2015, 6, 1891–1897
This journal is © The Royal Society of Chemistry 2015

View Article Online
MedChemComm
Concise Article
Fig. 3 Structures of the amides (A–X) synthesized.
2
3
evidence that it is sensitive to NAMPT inhibitors. While an
enzyme screening using recombinant NAMPT could also
have been considered to highlight possible hits, we preferred
this biological surrogate assay as important discrepancies
have been described between cellular and enzymatic activi-
ties and have been attributed to the ability of some inhibi-
tors to form phosphoribosyl adducts with the enzyme that
The cytotoxic effects of these seventeen purified com-
pounds are reported in Table 1. Only 8 compounds retained
activity once purified and, of these, two displayed consider-
able activity also at 1 μM. The presence of false positives in
the initial screening is not surprising and might be due to
inaccuracies in the biological assay as well as to the presence
of contaminants in the crude precipitate obtained. A full con-
centration–response curve of these two compounds (the
biphenyl-triazoles A12 and D12, Fig. 5) was obtained and the
EC50s for cytotoxicity were 456.2 ± 3.2 nM and 680.5 ± 30 nM,
respectively. As cytotoxicity is not necessarily a surrogate
marker for NAMPT inhibition, we also obtained a concentra-
tion–response curve for enzyme inhibition in vitro. A12
displayed an IC50 of 267 ± 94 nM and D12 displayed an IC50
of 170 ± 21 nM. In the assay conditions used, FK866 yielded
an EC50 for enzyme inhibition of 61 ± 7 nM.
2
4
can also lead to cellular accumulation. The enzymatic
assay, unlike the cytotoxicity assay, does not discriminate
2
4
between inhibitors that can or cannot forms adducts. It
should also be noted that the enzyme assay is more sensitive
to possible contaminants and compounds would have
needed to undergo purification; moreover the biological
assay is extremely more economical. Data for this screening
is presented in the ESI (Table S1†). As trace amounts of cop-
per salts could be present in the precipitated products, we
also evaluated the cytotoxicity of these salts in our cancer cell
line. Neither copperIJII) sulphate nor copperIJI) iodide
displayed cytotoxic effects up to 10 μM. Similar data on the
lack of confounding effect of copper salts had been pub-
Molecular modeling studies were then performed to
investigate the potential pose of A12 and D12 in the
FK866 binding site of NAMPT (PDB id: 2GVJ). Docking
studies were carried out using the software OMEGA2 (ref.
2
5
28,29
lished previously.
26 and 27) and FRED,
showing that the two structures
The first screening yielded seventeen compounds that
were able to reduce viability by >60% compared to the con-
trol: A12, C13, D12, G12, M14, M28, O14, P14, P34, Q12, Q39,
Q40, R12, S31, S39, W31, and W39 (Fig. S1†). These com-
pounds were then subject to chromatographic purification
and tested at increasingly lower concentrations.
lay with the same orientation as FK866. The structural
water molecule appears to be irrelevant to reproduce the
FK866 crystal pose, therefore it was not considered in the
docking simulations. The pyridine ring was stabilized in
the active site by a π–π stacking interaction with Phe193
and Tyr18′.
This journal is © The Royal Society of Chemistry 2015
Med. Chem. Commun., 2015, 6, 1891–1897 | 1893

View Article Online
Concise Article
MedChemComm
Fig. 4 Structures of in-house azides (11–40).
The carbonyl moiety was positioned in the binding cavity
near Ser275, accepting a hydrogen bond for both compounds,
while only the amide nitrogen of A12 was able to form an
H-bond interaction with Val242. The biphenyl group occupies
the same cavity around Ala379 and Lys189 in the case of com-
pound A12, while it has a different orientation involving a
π–π stacking interaction with Tyr188 in the case of D12
In addition, we synthesized the analogue with an E-double
bond between the amide and the pyridine group (compound 43),
Table 1 Viability of SH-SY5Y cells treated with the purified 17 triazole
products and enzyme inhibition activity. Data represent %cell viability in
mean ± SD of at least 4 determinations and %inhibition in mean ± SD of
at least 3 determinations
(
Fig. 6).
%
Cell viability
EC50 (nM)
for cell
viability
IC50 (nM)
for NAMPT
activity
This molecular modeling study suggests that both A12
(mean ± SD)
and D12 have longer H-bonding interactions compared to
FK866. For this reason, in an attempt to improve the potency
of the compounds, we decided to reverse the amide group
Comp.
10 μM
1 μM
A12
C13
D12
G12
M14
M28
O14
P14
P34
Q12
Q39
Q40
R12
S31
12.5 ± 0.5
100 ± 10
25 ± 1
47 ± 6
94 ± 3
96 ± 13
99 ± 13
101 ± 13
109 ± 9
32 ± 1
103 ± 10
125 ± 18
111 ± 6
29 ± 1
13.4 ± 0.5
456 ± 3.2
680 ± 30
267 ± 94
170 ± 21
(compounds 41 and 42, Fig. 7) as it appears to give a better
29.2 ± 1.1
98 ± 7
H-bonding interaction.
82 ± 3
102 ± 8
100 ± 2
104 ± 6
93 ± 2
S39
W31
W39
74 ± 3
53 ± 3
28 ± 1
Scheme 4 [3+2] cycloaddition between alkyne and azide to obtain
1
,4-disubstituted triazoles.
1894 | Med. Chem. Commun., 2015, 6, 1891–1897
This journal is © The Royal Society of Chemistry 2015

View Article Online
MedChemComm
Concise Article
Fig. 5 Structures compounds able to display considerable activity at 1
μM.
Fig. 8 Complex structure of hNAMPT enzyme and docked ligands (II).
The docking pose of 43 is shown in brown sticks. FK866 pose is
superimposed as a green wire model. The backbone of hNAMPT is
shown as a ribbon representation (yellow and cyan). The amino acids
of hNAMPT within 4.0 Å from FK866 are shown as wire models.
Hydrogen bonds (distance <3 Å) are shown as dotted yellow lines.
To prepare these three compounds, we started from com-
mercially available 6-heptynenitrile (44) which was coupled
with azide 12 via click reaction. The resulting nitrile 45 was
reduced to amine 46 using LiAlH and coupled with three dif-
4
ferent carboxylic acids (47, 48, and 49) to obtain the desired
amide products 41–43 (Scheme 5).
Fig. 6 Complex structure of hNAMPT enzyme and docked ligands (I).
Docking pose of A12 (a) and D12 (b) are shown in pink sticks. FK866 is
superimposed as a green wire model. The backbone of hNAMPT is
shown as a ribbon representation (yellow and cyan). The amino acids
of hNAMPT within 4.0 Å from FK866 are shown as wire models.
Hydrogen bonds (distance <3 Å) are shown as dotted yellow lines.
The cytotoxic results are reported in Table 2. Compounds
41 and 42 were inactive at 100 nM while compound 43 was
the most active with an IC50 of 20.0 ± 0.7 nM on neuroblas-
toma cells. The in vitro enzyme assay supported the activity
of this compound, with an IC50 of approx. 114 ± 12 nM in the
enzymatic assay for NAMPT activity. FK866, in our assay,
inhibited recombinant NAMPT with an IC50 of 61 ± 7 nM.
Last, in order to confirm the rank order of potency of
these compounds, we evaluated cellular NAD depletion in
SH-SY5Y cells treated with increasing concentrations of
FK866, A12, D12, and 43 for 16 hours. The rank order of
potency was maintained with respect to the other assays
performed, with A12 and D12 significantly less potent (IC50
s
of 1.1 ± 0.3 nM and 7.6 + 1.6 nM, respectively) compared to
FK866 and 43 (30.1 ± 5.9 pM and 50.9 ± 21.5 pM, respectively)
(
Fig. 9).
Conclusions
In conclusion, in the present study we employed click chem-
istry to synthesize 720 new compounds starting from simple
building blocks. We screened them for cytotoxicity and
NAMPT inhibition and then attempted to optimize them after
obtaining the initial biological data. This strategy led to a
compound with an EC50 for cytotoxicity of approx. 20 nM and
an IC50 for enzyme inhibition of approx. 100 nM that
displayed a 2-biphenyl-1,2,3-triazole as a novel tail group.
Fig. 7 Structure of compounds 41, 42 and 43.
mimicking the structure of FK866; as with our molecular
modeling studies we found that the interaction pattern of 43
was similar to that of FK866 (Fig. 8).
This journal is © The Royal Society of Chemistry 2015
Med. Chem. Commun., 2015, 6, 1891–1897 | 1895

View Article Online
Concise Article
MedChemComm
Scheme 5 Synthesis of three modified compounds (41–43).
Table 2 Viability of SH-SY5Y cells treated with 41, 42 and 43 at 1 μM
and 100 nM for 48 hours and enzyme inhibition of 43. Data represent
mean ± SD of at least 4 determinations
between the Department of Pharmaceutical Sciences and
Angelini Research Center.
%
Cell viability
EC50 (nM)
for cell
viability
IC50 (nM)
for NAMPT
activity
Notes and references
(mean ± SD)
Comp.
1 μM
100 nM
1 G. Magni, A. Amici, M. Emanuelli, N. Raffaelli and S.
Ruggieri, Adv. Enzymol. Relat. Areas Mol. Biol., 1999, 73,
135–182, xi.
4
4
4
1
2
3
99 ± 1
54 ± 1
32 ± 1
106 ± 2
112 ± 1
49 ± 3
20 ± 0.7
3.2 ± 0.4
114 ± 12
61 ± 7
2
A. Chiarugi, C. Dolle, R. Felici and M. Ziegler, Nat. Rev.
Cancer, 2012, 12, 741–752.
FK866
3
A. Garten, S. Petzold, A. Korner, S. Imai and W. Kiess, Trends
Endocrinol. Metab., 2009, 20, 130–138.
4
5
S. Imai, Curr. Pharm. Des., 2009, 15, 20–28.
M. Galli, F. Van Gool, A. Rongvaux, F. Andris and O. Leo,
Cancer Res., 2010, 70, 8–11.
6
L. Q. Zhang, D. P. Heruth and S. Q. Ye, J. Bioanal. Biomed.,
2
011, 3, 13–25.
7
8
9
M. Ziegler, Eur. J. Biochem., 2000, 267, 1550–1564.
H. C. Lee, Annu. Rev. Pharmacol. Toxicol., 2001, 41, 317–345.
D. Corda and M. Di Girolamo, EMBO J., 2003, 22,
1
953–1958.
1
1
1
0 K. Wosikowski, K. Mattern, I. Schemainda, M. Hasmann, B.
Rattel and R. Loser, Cancer Res., 2002, 62, 1057–1062.
1 M. Hasmann and I. Schemainda, Cancer Res., 2003, 63,
7
436–7442.
2 J. A. Khan, X. Tao and L. Tong, Nat. Struct. Mol. Biol.,
2006, 13, 582–588.
Fig. 9 Effect of NAMPT inhibitors on cellular NAD levels. Cells were
treated with the indicated compounds for 16 hours. n = 6 from two
separate experiments. IC50s are indicated in the main text.
13 U. Galli, C. Travelli, A. Massarotti, G. Fakhfouri, R.
Rahimian, G. C. Tron and A. A. Genazzani, J. Med. Chem.,
2
013, 56, 6279–6296.
1
4 D. Sampath, T. S. Zabka, D. L. Misner, T. O'Brien and P. S.
Dragovich, Pharmacol. Ther., 2015, 151, 16–31.
Acknowledgements
1
5 P. Beauparlant, D. Bedard, C. Bernier, H. Chan, K. Gilbert,
D. Goulet, M. O. Gratton, M. Lavoie, A. Roulston, E. Turcotte
and M. Watson, Anti-Cancer Drugs, 2009, 20, 346–354.
Financial support from AIRC to AAG (IG10509) and to CT
(Leonino Fontana e Maria Lionello Fellowship, 14832), as well
as financial support from the Fondazione San Paolo to AAG
16 K. Holen, L. B. Saltz, E. Hollywood, K. Burk and A. R.
Hanauske, Invest. New Drugs, 2008, 26, 45–51.
17 A. von Heideman, A. Berglund, R. Larsson and P. Nygren,
Cancer Chemother. Pharmacol., 2010, 65, 1165–1172.
(
(
oncoNAMPT) and the Italian Ministry of Education to GCT
PRIN-2010W7YRLZ_002), is gratefully acknowledged. This
work was also made possible by a research contract grant
1896 | Med. Chem. Commun., 2015, 6, 1891–1897
This journal is © The Royal Society of Chemistry 2015

View Article Online
MedChemComm
Concise Article
1
8 T. S. Zabka, J. Singh, P. Dhawan, B. M. Liederer, J. Oeh,
M. A. Kauss, Y. Xiao, M. Zak, T. Lin, B. McCray, N. La, T.
Nguyen, J. Beyer, C. Farman, H. Uppal, P. S. Dragovich, T.
O'Brien, D. Sampath and D. L. Misner, Toxicol. Sci.,
23 U. Galli, E. Ercolano, L. Carraro, C. R. Blasi Roman, G.
Sorba, P. L. Canonico, A. A. Genazzani, G. C. Tron and R. A.
Billington, ChemMedChem, 2008, 3, 771–779.
24 A. Oh, Y. C. Ho, M. Zak, Y. Liu, X. Chen, P. W. Yuen, X.
Zheng, Y. Liu, P. S. Dragovich and W. Wang, ChemBioChem,
2014, 15, 1121–1130.
25 T. Pirali, F. Pagliai, C. Mercurio, R. Boggio, P. L. Canonico,
G. Sorba, G. C. Tron and A. A. Genazzani, J. Comb. Chem.,
2008, 10, 624–627.
26 OMEGA, version 2.4.6, OpenEye Scientific Software, Santa Fe,
NM, http://www.eyesopen.com.
2
015, 144, 163–172.
1
9 M. Zak, B. M. Liederer, D. Sampath, P. W. Yuen, K. W. Bair,
T. Baumeister, A. J. Buckmelter, K. H. Clodfelter, E. Cheng,
L. Crocker, B. Fu, B. Han, G. Li, Y. C. Ho, J. Lin, X. Liu, J. Ly,
T. O'Brien, D. J. Reynolds, N. Skelton, C. C. Smith, S. Tay, W.
Wang, Z. Wang, Y. Xiao, L. Zhang, G. Zhao, X. Zheng and
P. S. Dragovich, Bioorg. Med. Chem. Lett., 2015, 25, 529–541.
0 G. Colombano, C. Travelli, U. Galli, A. Caldarelli, M. G.
Chini, P. L. Canonico, G. Sorba, G. Bifulco, G. C. Tron and
A. A. Genazzani, J. Med. Chem., 2010, 53, 616–623.
2
27 P. C. D. Hawkins, A. G. Skillman, G. L. Warren, B. A.
Ellingson and M. T. Stahl, J. Chem. Inf. Model., 2010, 50,
572–584.
2
1 A. Massarotti, S. Aprile, V. Mercalli, E. Del Grosso, G. Grosa,
G. Sorba and G. C. Tron, ChemMedChem, 2014, 9, 2497–2508.
2 S. E. Denmark and S.-M. Yang, Tetrahedron, 2004, 60, 9695–9708.
28 FRED, version 3.0.0, OpenEye Scientific Software, Santa Fe,
NM, http://www.eyesopen.com.
29 M. McGann, J. Chem. Inf. Model., 2011, 51, 578–596.
2
This journal is © The Royal Society of Chemistry 2015
Med. Chem. Commun., 2015, 6, 1891–1897 | 1897