Structure based design of nicotinamide phosphoribosyltransferase (NAMPT) inhibitors from a phenotypic screen

Article abstract of DOI:10.1016/j.bmcl.2017.12.037

Nicotinamide phosphoribosyltransferase is a key metabolic enzyme that is a potential target for oncology. Utilizing publicly available crystal structures of NAMPT and in silico docking of our internal compound library, a NAMPT inhibitor, 1, obtained from a phenotypic screening effort was replaced with a more synthetically tractable scaffold. This compound then provided an excellent foundation for further optimization using crystallography driven structure based drug design. From this approach, two key motifs were identified, the (S,S) cyclopropyl carboxamide and the (S)-1-N-phenylethylamide that endowed compounds with excellent cell based potency. As exemplified by compound 27e such compounds could be useful tools to explore NAMPT biology in vivo.

Full text of DOI:10.1016/j.bmcl.2017.12.037

Bioorganic & Medicinal Chemistry Letters xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Structure based design of nicotinamide phosphoribosyltransferase
(NAMPT) inhibitors from a phenotypic screen
a
a
a
a
Daniel S. Palacios a, , Erik Meredith , Toshio Kawanami , Christopher Adams , Xin Chen ,
Veronique Darsigny a, Erin Geno a, Mark Palermo a, Daniel Baird b, Geoffrey Boynton b, Scott A. Busby b,
Elizabeth L. George b, Chantale Guy b, Jeffrey Hewett b, Laryssa Tierney b, Sachin Thigale b,
Wilhelm Weihofen b, Louis Wang b, Nicole White b, Ming Yin b, Upendra A. Argikar c
⇑
a Global Discovery Chemistry, Novartis Institute for Biomedical Research, 22 Windsor Street, Cambridge, MA 02139, USA
b Chemical Biology and Therapeutics, Novartis Institute for Biomedical Research, 181 Massachusetts Avenue, Cambridge, MA 02139, USA
c PK Sciences, Novartis Institute for Biomedical Research, 250 Massachusetts Avenue, Cambridge, MA 02139, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Nicotinamide phosphoribosyltransferase is a key metabolic enzyme that is a potential target for oncol-
ogy. Utilizing publicly available crystal structures of NAMPT and in silico docking of our internal com-
pound library, a NAMPT inhibitor, 1, obtained from a phenotypic screening effort was replaced with a
more synthetically tractable scaffold. This compound then provided an excellent foundation for further
optimization using crystallography driven structure based drug design. From this approach, two key
motifs were identified, the (S,S) cyclopropyl carboxamide and the (S)-1-N-phenylethylamide that
endowed compounds with excellent cell based potency. As exemplified by compound 27e such com-
pounds could be useful tools to explore NAMPT biology in vivo.
Received 17 October 2017
Revised 13 December 2017
Accepted 16 December 2017
Available online xxxx
Keywords:
NAMPT
In silico screen
Structure based drug design
Cyclopropane
Ó 2017 Elsevier Ltd. All rights reserved.
Nicotinamide phosphoribosyltransferase
Regulation of nicotinamide adenine dinucleotide (NAD+) levels
within cells is critical given the importance of NAD+ in both main-
taining cellular energy homeostasis and its role as an enzymatic
cofactor.1 There are multiple mechanisms within eukaryotic cells
to sustain cellular NAD+ levels but the rate determining source is
the conversion of nicotinamide (NAM) to nicotinamide mononu-
cleotide (NMN) by the enzyme nicotinamide phosphoribosyltrans-
ferase (NAMPT). The NMN produced by NAMPT is then further
processed to NAD+ by the enzyme Nicotinamide mononucleotide
adenylyltransferase (NMNAT, Fig. 1A).2 Several reports have
demonstrated that cancerous cells express higher levels of NAMPT
and have a higher NAD+ demand than normal cells.3 This evidence
therefore suggests that blocking NAD+ production through the
inhibition of NAMPT enzymatic activity could be selectively lethal
to cancerous cells and consequently provide a therapeutic benefit.4
Recently, we disclosed the pyrrolo-pyrimidine NAMPT inhibitor
1 (Fig. 1B), which was discovered through a phenotypic screening
campaign.5 While 1 is very potent in an A2780 Cell-Titer GloTM
(CTG) assay (IC50: 6 nM) the construction of the pyrrolo-pyrim-
idine core requires a significant investment of synthetic effort.6
Given the structures of known NAMPT inhibitors,7 we hypothe-
sized that this complex core is not necessary for NAMPT activity
and we could identify an alternative to 1 that was more syntheti-
cally tractable and therefore more amenable to iterative optimiza-
tion. In this vein, we used the crystal structure of NAMPT/FK866
(PDB code: 2GVJ)8 as the template for an in silico screen for alter-
native chemical matter. The pyridine of FK866 is in a pi-stacking
position between Tyr180 and Phe193 which mimics the position
of the natural ligand NMN to competitively inhibit enzyme func-
tion. Docking 1 into this pocket suggested that the pyridine nitro-
gen could form the same hydrogen-bond with the phenol of Tyr18
as observed for the analogous nitrogen in the FK866 crystal struc-
ture. To potentially enhance this interaction, a 2-aminopyridine
was selected as a new head group because of the enhanced basicity
of this pyridine nitrogen relative to an unsubstituted pyridine. The
2-aminopyridine was then used as the substructure search query
to filter our internal compound collection, followed by docking into
the FK866 binding pocket. A dozen top-scoring docking hits were
selected and submitted for biological activity testing in the
A2780 CTG assay. From this effort, the most promising ligand iden-
tified was 2-aminopyridine 2, which we found to be equipotent to
1 in both the A2780 and CORL23 CTG assays.9 Furthermore, the cel-
lular activity of 2 was rescued by concurrent treatment with 100
⇑
Corresponding author.
E-mail address: daniel.palacios@novartis.com (D.S. Palacios).
https://doi.org/10.1016/j.bmcl.2017.12.037
0960-894X/Ó 2017 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Palacios D.S., et al. Bioorg. Med. Chem. Lett. (2017), https://doi.org/10.1016/j.bmcl.2017.12.037

2
D.S. Palacios et al. / Bioorganic & Medicinal Chemistry Letters xxx (2017) xxx–xxx
Fig. 1. Identification of cyclopropyl carboxamide NAMPT inhibitors from phenotypic hit 1. A. NAMPT converts NAM to NMN on the path to NAD+. B. Structure of phenotypic
hit 1 and in silico derived 2.
l
M of the NAMPT product NMN. Compound 2 was also found to
have been previously reported in the literature,15 this effort was
limited to a phenylsulfone core. With this in mind, we were intri-
gued by the potential of the phthalimide moiety to provide new
directions for the optimization of 6.
The absolute configuration of the cyclopropane was determined
through co-crystallization with NAMPT, as shown in Fig. 2. In addi-
tion to determining the absolute stereochemistry, the crystal struc-
ture also suggests the basis for the enhanced cellular activity of the
inhibit NAMPT in a biochemical assay (Fig. 1B). This data strongly
supports the conclusion that 2 is decreasing the viability of A2780
and CORL23 cells through direct inhibition of NAMPT. In addition
to the potent activity of 2, this small molecule has promising phys-
iochemical properties (clogP: 2.2, MW: 401) and can be readily
synthesized in a convergent fashion. Thus, 2 appeared to be a
promising foundation to build a lead optimization campaign.
Examining the structure of 2, we initially sought to explore the
structure activity relationship (SAR) of the urea motif. We hypoth-
esized that replacing the urea functionality could potentially
improve the aqueous solubility and membrane permeability of
the scaffold by decreasing the number of NH bonds. Furthermore,
Table 1
SAR of the amide linker.
O
given the
a
,b unsaturated amide in FK86610 and the bicyclic
NAM mimetics that are known in the literature,11,12 we postulated
that the benzylic urea nitrogen was not engaged in a direct inter-
action with the target protein and could be replaced. As shown
in Table 1 we first removed the aniline on the 2 position of the pyr-
idine to simplify analog synthesis and found that pyridine 3 was
approximately ten fold less potent in the A2780 CTG assay.13
Next, deleting the benzylic nitrogen gave amide 4, which was a
potent NAMPT inhibitor (IC50: 25 nM) but was inactive in the
CORL23 CTG assay. This type of disconnect between biochemical
and cell based assays has also been reported for other NAMPT
inhibitors.11 We suspected that the increased flexibility of the ethyl
bridge in 4 led to its greatly attenuated activity and we therefore
rigidified the system by introducing unsaturation to give amide
N
R
N
H
O
Cmpd
R
A2780 CTG
(nM)
CORL23 CTG
(nM)
NAMPT IC50
(nM)
3
100
NTa
3.1
25
10
O
O
O
O
N
H
R
N
N
N
N
4
5
>10,000
NT
>10,000
2000
R
5. However, we were surprised to find that a,b unsaturated amide
5 was only moderately active with an IC50 of 2000 nM in the
CORL23 CTG assay. To further explore conformational constraints
at this position, we utilized the Corey-Chaykovsky14 reaction to
synthesize cyclopropyl carboxamide 6 and separated the two trans
cyclopropane enantiomers. Upon testing, we found the (S,S) enan-
tiomer afforded significantly improved activity over the parent
urea 3. We also found that the cellular activity of the cyclopropane
motif is highly stereoselective as the corresponding (R,R) enan-
tiomer 7 was inactive in both the A2780 and CORL23 CTG assays.
Confirmation that NAMPT is the target of 6 was achieved through
demonstration of the biochemical inhibition of NAMPT (Table 1).
Though 3-pyridine cyclopropyl carboxamide NAMPT inhibitors
R
R
6
7
27
35
<1
>2500
>10,000
>2500
a
Not tested.
Please cite this article in press as: Palacios D.S., et al. Bioorg. Med. Chem. Lett. (2017), https://doi.org/10.1016/j.bmcl.2017.12.037

D.S. Palacios et al. / Bioorganic & Medicinal Chemistry Letters xxx (2017) xxx–xxx
3
Table 2
Combining the (S)-N-1-phenylethylamine and (S,S) cyclopropane.
Cmpd
A2780 CTG (nM)
CORL23 CTG (nM)
8
9
22
8
8
630
12
340
405
5
92
137
8
>2500
6
1330
>2500
22
10
11
12
13
14
15
reduces the entropy of binding of the cyclopropane relative to the
urea.17
Concurrent to the exploration of the urea, we were also probing
the SAR of phthalimide region of 2. Given our aim to identify in vivo
active molecules we decided to drive the optimization with cell
based activity. As mentioned above, urea 2 has very poor aqueous
solubility (<0.005 mM). In order to improve the solubility, we took
advantage of the flexible SAR of the phthalimide portion of the
molecule and appended solubilizing groups to the scaffold as
shown in Fig. 3A and Table 2.
As demonstrated by morpholine 8 and piperizine 9, these
groups are well tolerated and substantially improve upon the sol-
ubility of the parent compound (Solubility at pH 6.8 for 8 = 0.13
mM, 9 = >1 mM). When we obtained an X-ray co-crystal structure
of piperizine 9 bound to NAMPT we observed a hydrophobic
pocket located close to the benzyl methylene of the central phenyl
ring (Fig. 3C). Hypothesizing that occupying this pocket with a
hydrophobic group could increase compound affinity, we synthe-
sized the two alpha substituted amides 10 and 11 and we found
that filling this pocket with the (S)-N-1-phenylethylamide
Fig. 2. Overlay of urea 3 (PDB:) and cyclopropane 6 (PDB:) in the NAMPT crystal
structure showing the similar position of the pyridine nitrogen.
(S,S) cyclopropane relative to the urea. As described by others, the
heterocyclic warhead of cellularly active NAMPT inhibitors acts as
a
nicotinamide mimetic and is ribophosphorylated by the
enzyme.16 The position of the pyridine nitrogen within the cat-
alytic pocket is therefore critical for this substitution reaction to
proceed.
Strikingly, as shown in Fig. 2, there is a high degree of structural
overlap between the aromatic rings of urea 3 and cyclopropane 6
and the pyridine nitrogens for the two ligands are in almost the
exact same position within the binding pocket. Presuming that
the binding pose is related to the reactive conformation of the
ligand, we postulate that the cyclopropane preconfigures the posi-
tion of the pyridine nitrogen into this conformation and therefore
Fig. 3. The cyclopropane and (S)-benzyl amide enhance NAMPT potency. A. Combining the (S)-phenylethylamide and (S,S) cyclopropane does not effect potency on higher
molecular weight NAMPT inhibitors. B. Combining the (S,S) cyclopropane and (S)-phenylethylamide on lower molecular weight compounds enabled the minimal
pharmacophore 15. C. Overlay of the co-crystal structure of 9 bound to NAMPT (PDB:) and 15 (PDB:).
Please cite this article in press as: Palacios D.S., et al. Bioorg. Med. Chem. Lett. (2017), https://doi.org/10.1016/j.bmcl.2017.12.037

4
D.S. Palacios et al. / Bioorganic & Medicinal Chemistry Letters xxx (2017) xxx–xxx
(compound 10) significantly increases the activity of the scaffold in
the CORL23 CTG assay. This interaction is also highly stereoselec-
tive as the corresponding (R)-N-1-phenylethylamide (compound
11) is approximately 100 fold less active than the (S) enantiomer.
Next, we were interested to discover if the combination of the (S,
S)-cyclopropane and the (S)-N-1-phenylethylamide would further
enhance cellular activity but we found, contrary to our expectation,
that urea 10 and cyclopropane 12 are approximately equipotent.
While cyclopropane 12 displays excellent in vitro potency, we
found 12 to have poor solubility, permeability and in vitro clear-
ance. To address these poor properties, we set out to reduce the
overall size and hydrophobicity of the molecule. However, as
shown in Fig. 3B, both the (S)-N-1-phenylethylamide 13 and
cyclopropane 14 were very weakly active in our cell based assays.
Next, despite the equipotency of urea 10 and cyclopropane 12,
when we merged the analogous compounds 13 and 14 we were
surprised to find that cyclopropane compound 15 gained greater
than 50-fold potency in both the A2780 and CORL23 CTG assays
versus the parent compounds. In addition, as we had hypothesized,
the smaller ethylamide 15 is both more soluble and metabolically
stable than benzyl morpholine 12. Thus, ethylamide 15 revealed
the utility of combining the cyclopropane and (S)-N-1-phenylethy-
lamide to enable the installation of relatively small groups on the
eastern portion of the molecule to improve compound properties.
Furthermore, this result was achieved by challenging our assump-
tion based on cyclopropane 12 that the (S,S) cyclopropane and (S)-
N-1-phenylethylamide would not be additive in terms of com-
pound activity. For our chemistry strategy, this was a lesson
learned that SAR between closely related series may not always
track and one has to always question preconceived wisdom regard-
ing scaffold optimization.
Based on the SAR shown in Table 3, we decided to keep the
methyl group fixed and commenced with an interrogation of the
amide SAR. As shown in Scheme 1, the synthesis of these analogues
began with the Heck reaction of 3-bromo pyridine 20 and t-butyl
acrylate to give the alpha-beta unsaturated ester 21. The t-butyl
ester is critical because we found that the corresponding methyl
ester was hydrolyzed during the Corey-Chaykovsky cyclopropana-
tion and the resulting acid proved difficult to isolate. Deprotection
of the ester with TFA followed by chiral chromatography (see Sup-
porting information for details) then afforded enantiomerically
pure 23. Through this sequence, we were able to generate multi-
gram quantities of the key non-racemic intermediate 23 in signif-
icantly fewer steps than was previously reported.14 Diversification
of the amide began with the commercially available secondary
amine 24 which was coupled to a range of acids using standard
amide bond forming conditions. Reduction of the aryl nitro group
then afforded the corresponding anilines 26a–i and the ultimate
step was a second amide bond formation to give compounds
27a–i.
As shown in Table 4, the eastern amide was broadly tolerant of
diverse substructures, enabling the identification of a number of
potent compounds with diverse physiochemical properties. Ali-
phatic heterocycles such as tetrahydropyran 27b and N-methyl
proline 27d have single digit nanomolar potency and good solubil-
ity. In addition, N-methyl proline 27d shows an improved perme-
ability profile, likely due to the formation of an intramolecular
hydrogen bond between the amide proton and the proline nitro-
gen.18 An unsubstituted phenyl group was active and though the
in vitro clearance was increased, this could be attenuated through
the addition of electron withdrawing groups to the four position of
the ring, as demonstrated by fluorine 27f and nitrile 27g.
Given the utility of the (S)-N-1-phenylethylamide, we next
probed the SAR of this position to determine if the methyl group
was ideal for activity and, as shown in Table 3, there is limited tol-
erance at this position with small non-polar groups being pre-
ferred. Even adding an additional carbon with ethyl 17 reduces
the activity of the compound about tenfold relative to the methyl
matched pair (compound 27e Table 4). In addition, adding a polar
atom, as demonstrated by primary alcohol 18, deleteriously
impacts the cell based activity of the scaffold (200 nM A2780 CTG).
In addition, heterocycles such pyridine 27h and pyrimidine 27i
can also replace the phenyl group with a small loss in activity but a
significantly decreased in vitro clearance. Significantly, the CYP3A4
inhibition of these compounds is attenuated relative to previously
reported cyclopropylcarboxamide NAMPT inhibitors14 (Table 4).
This may in part be due to the potency combination of the cyclo-
propane and (S)-N-1-phenylethylamide which enables the installa-
tion of relatively small amide appendages.
Given the promising in vitro properties of phenyl amide 27e, we
decided to investigate the in vivo metabolic stability of this com-
pound. As shown in Fig. S2 (supporting information), 27e has good
in vivo properties with a half-life of 90 min and a markedly reduced
intrinsic clearance than was predicted by the mouse liver
microsomes.
Table 3
SAR of the (S) benzyl position.
R
O
The good in vivo exposure of 27e combined with its concise syn-
thesis suggests that this small molecule would be an excellent tool
to explore NAMPT biology both in vivo and in vitro. With this in
mind, we employed a mouse CORL23 xenograft model and found
that phenylamide 27e, when dosed orally at 10 mg/kg bid, signifi-
cantly reduced the growth of the implanted tumor (Fig. 4A). The
antitumor efficacy that we observe for 27e is comparable to the
result shown in Fig. 4B for the known NAMPT inhibitor 28 (see
Supporting information). Furthermore, we also found that the
observed efficacy of 27e correlates to the blockade of glycolytic
flux, which requires NAD for turnover, as measured by the accu-
mulation of the glycolysis intermediates fructose-6-phosphate
and glucose-6-phosphate (Supporting information).19
To conclude, our pursuit of compounds that inhibit the function
of the NAD+ producing enzyme NAMPT began with phenotypic
screening hit 1. The poor physiochemical properties and complex
synthesis of pyrrolo-pyrimidine 1 catalyzed the search for alterna-
tive chemical matter and a consequent in silico screen identified 2-
amino pyridine 2. The promising cellular activity and much
improved access to analogues of 2 enabled a structure based lead
optimization effort to improve overall physiochemical properties.
N
H
N
Cmpd
R
A2780 CTG (nM)
832
CORL23 CTG (nM)
2310
16
17
18
19
43
200
1
137
630
4
Please cite this article in press as: Palacios D.S., et al. Bioorg. Med. Chem. Lett. (2017), https://doi.org/10.1016/j.bmcl.2017.12.037

D.S. Palacios et al. / Bioorganic & Medicinal Chemistry Letters xxx (2017) xxx–xxx
5
Table 4
SAR of the amide appendage.
Me
R
O
N
H
N
H
N
Cmpd
R
A2780 CTG (nM)
5
CORL23 CTG (nM)
18
SolpH6.8
(mM)
PERM classa
L
MLM
l
CYP3A4 (
lM)
L minꢀ1 mgꢀ1b
27a
0.041
83.7
4.2
27b
3
7
0.09
L
25.1
15.8
27c
27d
30
1
111
11
0.52
0.73
L
43.2
65.4
9.4
M
10.7
27e
27f
1
2
8
<0.004
0.006
M
M
148
17.1
4.7
12
13.7
27g
4
19
0.175
L
53
0.5
27h
27i
14
14
46
17
0.01
H
<3.42
38.4
0.95
7.2
0.027
M
a
Permeability class, based on PAMPA profiling.
Mouse liver microsome.
b
Scheme 1. Synthesis reagents and conditions. (a) tert-Butyl acrylate (1.5 eq), Pd(OAc)2 (0.05 eq), tri-o-tolylphosphine (0.1 eq), DIPEA (1.5 eq), DMF, 110 °C, 72%; (b)
trimethylsulfoxonium iodide (2.0 eq), sodium hydride (2 eq), DMSO, 50 °C, 95%; (c) TFA (6 eq), DCM, 23 °C; chiral separation, 45%; (d) carboxylic acid (1.05 eq) HATU (1.05
eq), DIPEA (3 eq), DMF, 23 °C; zinc (5 eq), ammonium chloride (5 eq), EtOH, water, 50 °C, 30 min, 91%, two steps; (e) 23 (1 eq) 26a–i (1 eq) HATU (1.05 eq), DIPEA (4 eq), DMF,
23 °C, 65%.
Please cite this article in press as: Palacios D.S., et al. Bioorg. Med. Chem. Lett. (2017), https://doi.org/10.1016/j.bmcl.2017.12.037

6
D.S. Palacios et al. / Bioorganic & Medicinal Chemistry Letters xxx (2017) xxx–xxx
Fig. 4. In vivo activity in a CORL23 xenograft model. A. In vivo efficacy of 27e dosed orally in a CORL23 xenograft. B. Efficacy data for the oral administration of compound 28.
Two insights obtained from this work were the identification of
the (S,S) cyclopropylcarboxamide and the (S)-N-1-phenylethy-
lamide as key functional groups that enhance the activity of these
NAMPT inhibitors. When combined, these two motifs can syner-
gize to yield highly ligand efficient inhibitors, as exemplified by
ethylamide 15. The cellular potency of the cyclopropyl carboxam-
ides tolerates a broad range of structures in the eastern portion of
the molecule, which can be utilized to tune compound properties
or to potentially generate probe compounds to interrogate NAMPT
biology. These optimization efforts culminated in phenyl amide
27e, a compound with excellent in vivo characteristics that was
efficacious in a CORL23 mouse xenograft experiment.
References
1. Cantó C, Menzies KJ, Auwerx J. Cell Metab. 2015;22:31–53.
2. Revollo JR, Grimm AA, Imai S. J Biol Chem. 2004;279:50754–50763.
3. (a) Wang B, Hasan MK, Alavarado E, Yuan H, Wu H, Chen WY. Oncogene.
2011;30:907–921;
(b) Bi T-Q, Che X-M, Liao X-H, et al. Oncol Rep. 2011;26:1251–1257;
(c) Shackleford RE, Bui MM, Coppola D, Hakam A. Int
J Clin Exp Med.
2010;3:522–527;
(d) Maldi E, Travelli C, Caldarelli A, et al. Pigm Cell Melanoma Res.
2013;26:144–146.
4. (a) Hasmann M, Schemainda I. Cancer Res. 2003;63:7436–7442;
(b) Chiarugi A, Dölle C, Felizi R, Ziegler M. Nature Rev. Cancer.
2012;12:741–752.
5. Estoppey D, Hewett J, Guy CT, et al. Sci Rep. 2017;7:42728.
6. Irie O, Ehara T, Iwasaki A, et al. Bioorg Med Chem Lett. 2008;18:3959–3962.
7. Galli U, Travelli C, Massarotti A, et al. J Med Chem. 2013;56:6279–6296.
8. Khan JA, Tao X, Tong L. Nat Struct Biol. 2006;13:582–588.
9. Gunzner-Toste J, Zhao G, Bauer P, et al. Bioorg Med Chem Lett.
2013;23:3531–3538.
10. Hasmann M, Schemainda I. Cancer Res. 2003;63:7436–7442.
11. Zheng X, Bauer P, Baumeister T, et al. J Med Chem. 2013;56:6413–6433.
12. Dragovich PS, Bair KW, Baumeister T, et al. Bioorg Med Chem Lett.
2013;23:4875–4885.
13. Zheng X, Bauer P, Baumeister T, et al. J Med Chem. 2013;56:4921–4937.
14. Li JJ. Named Reactions in Heterocyclic Chemistry. 2nd ed. Hoboken, New
Jersey: John Wiley & Sons; 2005.
15. Giannetti AM, Zheng X, Skelton NJ, et al. J Med Chem. 2014;57:770–792.
16. Oh A, Ho Y-C, Zak M, et al. ChemBioChem. 2014;15:1121–1130.
17. Boström J, Grant A. In: Mannhold R, ed. Exploiting Ligand Conformations in Drug
Design, Vol. 37. Weinheim, Germany: Wiley-VCH; 2007:183–205.
18. Kuhn B, Mohr P, Stahl M. J Med Chem. 2010;53:2601–2611.
19. Tan B, Young DA, Lu ZH, et al. J Biol Chem. 2013;288:3500–3511.
Acknowledgments
We thank the GDC separations group for the asymmetric purifi-
cations and Frank Cook in Analytical Sciences for Fig. S4.This
research was funded by the Novartis Institutes for Biomedical
Research.
A. Supplementary data
Supplementary data (experimental details, representative
curves for the biochemical inhibition of NAMPT, pharmacokinetic
and pharmacodynamic data of the compounds) associated with
this article can be found, in the online version, at https://doi.org/
10.1016/j.bmcl.2017.12.037.
Please cite this article in press as: Palacios D.S., et al. Bioorg. Med. Chem. Lett. (2017), https://doi.org/10.1016/j.bmcl.2017.12.037