Identification of amides derived from 1H-pyrazolo[3,4-b]pyridine-5- carboxylic acid as potent inhibitors of human nicotinamide phosphoribosyltransferase (NAMPT)

Article abstract of DOI:10.1016/j.bmcl.2013.08.074

Potent, 1H-pyrazolo[3,4-b]pyridine-containing inhibitors of the human nicotinamide phosphoribosyltransferase (NAMPT) enzyme were identified using structure-based design techniques. Many of these compounds exhibited nanomolar antiproliferation activities a

Full text of DOI:10.1016/j.bmcl.2013.08.074

Accepted Manuscript
Identification of Amides Derived From 1H-Pyrazolo[3,4-b]pyridine-5-carbox‐
ylic Acid as Potent Inhibitors of Human Nicotinamide Phosphoribosyltransfer‐
ase (NAMPT)
Xiaozhang Zheng, Kenneth W. Bair, Paul Bauer, Timm Baumeister, Krista K.
Bowman, Alexandre J. Buckmelter, Maureen Caligiuri, Karl H. Clodfelter,
Yezhen Feng, Bingsong Han, Yen-Ching Ho, Nikolai Kley, Hong Li, Xiaorong
Liang, Bianca M. Liederer, Jian Lin, Justin Ly, Thomas O’Brien, Jason Oeh,
Angela Oh, Dominic J. Reynolds, Deepak Sampath, Geeta Sharma, Nicholas
Skelton, Chase C. Smith, Jarrod Tremayne, Leslie Wang, Weiru Wang,
Zhongguo Wang, Hongxing Wu, Jiansheng Wu, Yang Xiao, Guangxing Yang,
Po-wai Yuen, Mark Zak, Peter S. Dragovich
PII:
S0960-894X(13)01014-7
DOI:
http://dx.doi.org/10.1016/j.bmcl.2013.08.074
BMCL 20812
Reference:
To appear in:
Bioorganic & Medicinal Chemistry Letters
Received Date:
Revised Date:
Accepted Date:
14 July 2013
11 August 2013
15 August 2013
Please cite this article as: Zheng, X., Bair, K.W., Bauer, P., Baumeister, T., Bowman, K.K., Buckmelter, A.J.,
Caligiuri, M., Clodfelter, K.H., Feng, Y., Han, B., Ho, Y-C., Kley, N., Li, H., Liang, X., Liederer, B.M., Lin, J., Ly,
J., O’Brien, T., Oeh, J., Oh, A., Reynolds, D.J., Sampath, D., Sharma, G., Skelton, N., Smith, C.C., Tremayne, J.,
Wang, L., Wang, W., Wang, Z., Wu, H., Wu, J., Xiao, Y., Yang, G., Yuen, P-w., Zak, M., Dragovich, P.S.,
Identification of Amides Derived From 1H-Pyrazolo[3,4-b]pyridine-5-carboxylic Acid as Potent Inhibitors of
Human Nicotinamide Phosphoribosyltransferase (NAMPT), Bioorganic & Medicinal Chemistry Letters (2013),
doi: http://dx.doi.org/10.1016/j.bmcl.2013.08.074
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.

Graphical Abstract.
1

Identification of Amides Derived From 1H-Pyrazolo[3,4-b]pyridine-
5-carboxylic Acid as Potent Inhibitors of Human Nicotinamide
Phosphoribosyltransferase (NAMPT)
Xiaozhang Zhengb, Kenneth W. Bairb, Paul Bauerb, Timm Baumeisterb, Krista K. Bowmana,
Alexandre J. Buckmelterb, Maureen Caligiurib, Karl H. Clodfelterb, Yezhen Fengc, Bingsong
Hanb, Yen-Ching Hob, Nikolai Kleyb, Hong Lia, Xiaorong Lianga, Bianca M. Liederera, Jian
Linb, Justin Lya, Thomas O’Briena, Jason Oeha, Angela Oha, Dominic J. Reynoldsb, Deepak
Sampatha, Geeta Sharmab, Nicholas Skeltona, Chase C. Smithb, Jarrod Tremaynea, Leslie Wanga,
Weiru Wanga, Zhongguo Wangb, Hongxing Wuc, Jiansheng Wua, Yang Xiaoa, Guangxing Yangc,
Po-wai Yuenc, Mark Zaka, Peter S. Dragovicha,*
aGenentech, Inc., 1 DNA Way, South San Francisco, California 94080, USA
bForma Therapeutics, Inc., 500 Arsenal Street, Watertown, Massachusetts 02472, USA.
cPharmaron Beijing, Co. Ltd., 6 Taihe Road, BDA, Beijing, 100176, P.R. China.
Abstract. Potent, 1H-pyrazolo[3,4-b]pyridine-containing inhibitors of the human nicotinamide
phosphoribosyltransferase (NAMPT) enzyme were identified using structure-based design
techniques. Many of these compounds exhibited nanomolar antiproliferation activities against
human tumor lines in in vitro cell culture experiments, and a representative example (compound
26) demonstrated encouraging in vivo efficacy in a mouse xenograft tumor model derived from
the A2780 cell line. This molecule also exhibited reduced rat retinal exposures relative to a
previously studied imidazo-pyridine-containing NAMPT inhibitor. Somewhat surprisingly,
compound 26 was only weakly active in vitro against mouse and monkey tumor cell lines even
though it was a potent inhibitor of NAMPT enzymes derived from these species. The compound
also exhibited only minimal effects on in vivo NAD levels in mice, and these changes were
considerably less profound than those produced by an imidazo-pyridine-containing NAMPT
1

inhibitor. The crystal structures of compound 26 and the corresponding PRPP-derived ribose
adduct in complex with NAMPT were also obtained.
Keywords: nicotinamide phosphoribosyltransferase, NAMPT, X-ray crystal structure, tumor
metabolism.
2

The conversion of nicotinamide (NAM) to the biologically important enzyme co-factor
nicotinamide adenine dinucleotide (NAD) is accomplished in mammalian cells by the two-step
process depicted in Figure 1.1 Nicotinamide phosphoribosyltransferase (NAMPT; also known in
the literature as pre-B cell colony-enhancing factor/PBEF and visfatin) catalyzes the first and
rate-limiting step in this transformation in which nicotinamide (NAM) is condensed with
phosphoribosyl pyrophosphate (PRPP) to produce nicotinamide mononucleotide (NMN; Figure
1).2 Although covalent modification of NAM and/or PRPP by NAMPT is not believed to occur
during this conversion, efficient production of NMN is dependent on the phosphorylation of a
histidine residue in the NAMPT active site by ATP.2 NAM is produced by the normal function
of NAD-consuming enzymes such as PARPs and sirtuins, and the described NAM-NAD
conversion represents the only known mechanism for mammalian cells to efficiently recycle
NAM back to NAD.3 Accordingly, blocking NAMPT activity may impair the maintenance of
intracellular NAD levels, and this disruption may negatively impact the growth and/or function
of cells which are highly reliant on NAD-dependent processes for survival. As many tumors
and/or cancer cell lines exhibit elevated metabolisms and high proliferation rates, NAMPT
inhibition is currently viewed as a novel strategy for the development of new anticancer agents.4
Multiple examples of NAMPT inhibitors have been previously described in the scientific
and patent literature.2a,2b The most advanced of these agents, GMX-17785 (1), its prodrug GMX-
17776 (not shown), and APO-8667 (2) progressed to clinical trials during the past decade (Figure
2). In addition, our own NAMPT inhibitor discovery efforts recently identified amide-containing
sulfonamides and sulfones exemplified by compounds 3 and 4 (Figure 2).8,9 One concern
associated with the development of NAMPT inhibitors is that they may exhibit undesired
toxicities toward normal (non-cancerous) tissues, particularly those which are metabolically
3

active. Indeed, retinal atrophy was reported in pre-clinical rodent safety assessments with APO-
866, although such effects were not observed in related monkey studies.7b In addition, mutations
in NMNAT, the enzyme which catalyzes the second step in the NAM-NAD conversion (c.f.,
Figure 1), were recently shown to cause serious retinal defects in humans (Leber congenital
amaurosis; LCA).10 While it is currently uncertain whether the mutant NMNAT outcomes are
related to the protein’s ability to produce NAD,11 the collective data described above suggest that
inhibiting the NAM-NAD conversion has the possibility to detrimentally impact the health of
retinal tissue.12
We therefore sought to identify NAMPT inhibitors which exhibited reduced retinal
penetration and/or retention relative to the molecules we had previously studied. If mechanism-
based retinal toxicities were to be encountered with the previously-studied compounds, the new
entities might be able to attenuate these unwanted effects.13 Accordingly, we implemented a
physiochemical-based design strategy to minimize the blood-retina barrier penetration and/or the
retinal retention times of newly-synthesized NAMPT inhibitors.14 In particular, we sought to
identify compounds devoid of basic centers (even weakly basic ones) as protonation of such sites
was believed to contribute greatly to long retinal retention.14 This aspect of the new inhibitor
design posed a significant challenge since all potent, cell-active NAMPT inhibitors described to
date contain a weakly basic heterocyclic nitrogen atom believed to participate in an enzyme-
catalyzed condensation with PRPP in the protein’s active site (e.g., imidazo-pyridine-containing
inhibitors such as 5 are converted to PRPP-adducts 7; Figure 3).15 We hypothesized that
isosteric (and less-basic) indazole-derived inhibitors might form similar PRPP adducts in the
NAMPT active site and thereby also function as potent cell-active inhibitors of the enzyme (e.g.,
conversion of 6 to 8; Figure 3). In this report, we describe the successful identification of
4

indazole-derived NAMPT inhibitors along with their detailed biological and structural
characterization.
As shown in Table 1, incorporation of a simple indazole into the inhibitor design afforded
a compound that was a significantly weaker NAMPT inhibitor than the corresponding imidazo-
pyridine-containing molecule (compare 10 with 3). Introduction of nitrogen atoms at the 4 and
6-positions on the 5-carboxy-indazole ring system also afforded weakly active NAMPT
inhibitors (compare 11 and 12 with 9, Table 1). However, a molecule containing an amide
derived from 1H-pyrazolo[3,4-b]pyridine-5-carboxylic acid (13; i.e., incorporation of a nitrogen
at the 7-position of the 5-carboxy-indazole ring system) displayed relatively potent biochemical
NAMPT inhibition and was also active in an A2780 cell-based assessment.16,17 Similar
inhibition activities were observed when a phenyl sulfone moiety was incorporated at the
inhibitor terminus in lieu of the piperidine-derived sulfonamide (14). However, methylation of
the indazole NH group present in 14 or removal of the other indazole nitrogen atom was highly
detrimental to anti-NAMPT activity (compare compounds 15 and 16 with 14).
To facilitate further optimization of the 1H-pyrazolo[3,4-b]pyridine-containing
compounds, an X-ray crystal structure of compound 14 in complex with the NAMPT protein was
determined.18 As shown in Figure 4, the compound bound in a cleft formed between two
monomers of the NAMPT protein and extended from this binding site out towards solvent. The
1H-pyrazolo[3,4-b]pyridine ring system formed a face-to-face pi-stack with Phe193 of one
NAMPT monomer and with Tyr18’ of the other in a manner that likely mimics nicotinamide
binding to the same region. The N-2 nitrogen atom in the 1H-pyrazolo[3,4-b]pyridine ring was
hydrogen bonded to a water molecule in the active site, which also formed a hydrogen bond to
the backbone carbonyl oxygen of Phe293. A phosphate ion, likely originating from the
5

crystallization buffer, was also observed in the active site cleft, and this entity was coordinated
by the side chains of Arg392', Ser398' and Lys400' (for clarity, these side chains are not depicted
in Figure 4). In addition, nine water molecules were noted within hydrogen-bonding distance of
the phosphate oxygen atoms. The amide portion of 14 occupied a more polar NAMPT pocket
and formed hydrogen bonds with the side chain of Asp219 and a neighboring water molecule
(which itself was hydrogen bonded to the side chains of Arg311 and Ser275). The bisphenyl
sulfone moiety present in 14 made non-bonded contacts with the side chains of Ile351, Ile309
and Val242 in a much more hydrophobic portion of the binding site. The phenyl group proximal
to the amide also formed an edge-to-face pi-stack with His191, whereas the distal phenyl moiety
projected into the solvent exposed entrance of the binding cleft. A hydrogen bond between one
of the 14 sulfone oxygen atoms and a solvent water molecule was also observed.
With the above structural information in hand, we next sought to further improve the
biochemical and cell-based inhibition properties of the new inhibitor series. As shown in Table
2, incorporation of sulfonamides derived from substituted piperidines or piperazines (17-22),
morpholine (23), or a spirocyclic azetidine (24) into the inhibitor design was generally tolerated.
However, none of these entities was significantly more active than the lead compound containing
an unsubstituted piperidine-derived sulfonamide (13). In contrast, inhibitors incorporating
various substituted phenyl sulfones displayed dramatically improved anti-NAMPT properties in
both biochemical and cell-based assessments relative to the unsubstituted analog (compare 25-27
with 14, Table 2). Similar, incorporation of terminal sulfones containing various heteroaromatic
groups also typically afforded very potent biochemical and cell-based NAMPT inhibitors (28-35,
Table 2; exception = 33). Encouragingly, the majority of the compounds depicted in Table 2,
6

especially the diaryl-sulfones, exhibited acceptable stability (stable to moderate) toward human
liver microsomes when tested in an in vitro assay.
In preparation for conducting in vivo experiments with the new class of NAMPT
inhibitors, we profiled several of the most potent examples in a variety of additional in vitro
DMPK assessments. As shown in Table 3, all compounds examined were relatively labile
toward mouse liver microsomes, but all exhibited improved metabolic stability properties when
the assessment was performed using mouse hepatocytes. The molecules displayed a range of
permeabilities as determined in an MDCK cell monolayer assay with the more polar entities
unsurprisingly exhibiting the lower values. The compounds were typically highly bound to
mouse plasma proteins but one example (25) exhibited higher free fraction than the others. The
solubility of the molecules depicted in Table 3 was generally very poor with the values for many
compounds residing below the limit of the high throughput assessment used to measure this
property. However, compound 26 was an exception to this trend, and it exhibited measurable
(albeit low) solubility in the high throughput assay. Importantly, the other Table 3 data
associated with 26 collectively indicated that it possessed an ideal combination of potent
NAMPT inhibition activity and attractive in vitro DMPK properties. Based on these favorable
characteristics, compound 26 was therefore progressed to in vivo mouse PK and xenograft
efficacy studies.
As shown in Figure 5 and Table 4, inhibitor 26 exhibited encouraging plasma exposures
and PK parameters following IV and oral administration to female NCR nude mice. Single-dose
IV administration of the compound determined that the molecule’s in vivo mouse CL value was
lower than that predicted from the in vitro mouse microsome and hepatocyte stability
assessments (c.f., Table 3). This difference is potentially due to the high plasma protein binding
7

exhibited by the molecule. Importantly, QD administration of 26 at 100 mg/kg afforded free
mouse plasma levels which exceeded the compound’s A2780 antiproliferation cell culture IC50
values for approximately 24 h (Figure 5). Accordingly, compound 26 displayed robust efficacy
in an A2780 ovarian cancer xenograft model with relatively minor decreases in body weights
when dosed orally at 100 mg/kg for 5 consecutive days (Figure 6A; see supplementary data for
additional details). Notably, durable tumor regressions were observed in this experiment that
were maintained even after compound dosing was discontinued. In addition, tumor NAD levels
measured 1 hour after the day-5 dose of compound 26 were significantly reduced relative to the
levels observed in tumors from vehicle treated animals (Figure 6B). Similar reductions in tumor
NAD levels were also noted 24 hours after the day-5 compound 26 dose (Figure 6B).
Encouragingly, and consistent with the design strategy outlined in the beginning of this work,
compound 26 exhibited reduced retinal exposures relative to the imidazo-pyridine-containing
inhibitor 4 when this property was assessed at both 1 and 6.5 hours following separate oral
administration of the compounds to rats (Table 5; the differences were more apparent when the
retina levels were normalized with respect to the corresponding plasma concentrations).19 As
described above, these reduced exposures likely result from the attenuated basicity of compound
26 relative to 4 (measured pKa’s = <2.0 and 5.3, respectively), but other factors such as logD (26
= 3.4; 4 = 3.2) and/or TPSA (26 = 104; 4 = 80) differences may also be responsible.14
In anticipation of conducting animal toxicology experiments with compound 26, the
molecule was biochemically profiled against NAMPT enzymes derived from animal species
commonly used in such assessments. The imidazo-pyridine-containing inhibitor 4 was similarly
examined, since its use in related toxicology studies was also being considered. As shown in
Table 6, the biochemical IC50 inhibition values for each compound varied <12-fold across the
8

different enzymes tested, with the most potent inhibition consistently observed against human
NAMPT. A <3.5-fold variation in IC50 values was also noted when the two compounds were
tested against the same NAMPT isoform. However, since our prior experience indicated that
potent biochemical NAMPT inhibitors did not always function as potent cell-active
molecules,6,8,15 we also examined compounds 4 and 26 in a variety of in vitro assessments
employing cell-lines from the same animal species. As shown in Table 7, compound 4 displayed
potent or moderate anti-NAMPT cell activity in all lines against which it was tested. In stark
contrast, however, compound 26 displayed negligible activity in mouse and monkey-derived cell
lines and only exhibited potent antiproliferation effects in human and dog cells. This unexpected
finding prompted us to conduct several additional experiments to verify and better understand
the species activity differences associated with compound 26.
As shown in Table 8, compound 26 failed to significantly reduce NAD levels in mouse
KPP cells unless relatively high concentrations were employed while large mouse cell NAD
reductions were induced by comparably low levels of compound 4 (see also Figure S1 in the
supplementary data). In contrast, both compounds potently inhibited NAD levels in the human-
derived PC3 cell line (Table 8, Figure S1). Importantly, intracellular quantitation of each
compound in the murine KPP and human PC3 lines showed similar concentrations of 4 and 26
with no obvious difference in the levels of 26 present in the KPP (insensitive) and PC3
(sensitive) cells (Table 8). The ability of each compound to form NAMPT-catalyzed PRPP
adducts in biochemical experiments was also assessed,15 and no distinction among them was
observed (each compound formed the corresponding PRPP adduct with NAMPT enzyme from
all five species; see supplementary data). Thus, examination of two obvious explanations for the
described cell-based discrepancies failed to improve our understanding of the phenomenon.
9

To confirm that the cell-culture results depicted in Table 7 translated into in vivo
outcomes, we performed a 5-day PK-PD study with compounds 4 and 26 in nude mice. In order
to maximize the observed in vivo effects, we employed the highest tolerable doses of each
molecule in this experiment (4 = 30 mg/kg QD; 26 = 100 mg/kg QD). This dosing regimen
afforded much higher mouse exposures of inhibitor 26 relative to compound 4 (supplementary
data, Figure S5A).20 In spite of these exposure differences, administration of inhibitor 4 resulted
in more pronounced reductions in day-5 mouse blood NAD levels relative to those observed for
compound 26 (Figure S5B). Differing effects on NAD levels were also noted in eye tissue
sampled from animals in both dosing groups with compound 26 having minimal impact (Figure
S5C). The attenuated effect of 26 in eye tissue relative to its activity in blood is consistent with
the low retina/plasma ratio observed for the molecule in rat PK experiments (c.f., Table 5).
Collectively, the PK-PD data show that high exposures of compound 26 afforded less
pronounced effects on NAD levels in mice relative to those produced by much lower exposures
of inhibitor 4. Thus, the in vivo mouse PK-PD results are consistent with the Table 7 cell culture
data which indicate that compound 26 exhibits dramatically attenuated potency in mouse cell
lines relative to compound 4.21
We also obtained a crystal structure of the PRPP-ribosylated adduct corresponding to
compound 26 in complex with human NAMPT to determine whether analysis of its binding to
the protein might help explain the above biological outcomes.22 As shown in Figure 7, the non-
ribose atoms of the adduct were observed in locations very similar to those occupied by the
corresponding atoms of inhibitor 14 (c.f., Figure 4; RMSD = 0.29 Å for 28 common ligand
heavy atoms after superposition of the dimer protein chains). The backbone fold of NAMPT was
also very similar in complex with compound 14 and compound 26 (RMSD 0.15 Å for 444 C-
10

alpha atoms in a single NAMPT protomer, as defined by the protein alignment tool in PyMOL),
and the side chains lining the inhibitor binding cleft were located in very similar positions in both
structures (compare Figures 4 and 7). Thus, the protein-ligand interactions involving non-ribose
portions of the ligands were identical in the two structures. A similar binding orientation was
also observed for compound 4 and its corresponding ribose adduct.8,23 In contrast to the complex
with inhibitor 14, there was no electron density observed for water molecules adjacent to the N-2
indazole nitrogen or the amide carbonyl oxygen of the 26-ribose adduct. This difference may be
due to the different resolutions of the two structures (1.75 vs 2.40 Å). Importantly, the 1.5 Å
separation between the N-1 indazole nitrogen of 26 and the 1' ribose carbon atom clearly
indicated the formation of a covalent bond between the two entities. The phospho-ribose
portion of the 26-ribose adduct formed several hydrogen bonds with the NAMPT protein
involving the Arg311 side chain (to the ribose 2' hydroxyl), the Asp313 side chain (also to the
ribose 2' hydroxyl), and the backbone amide NH of Gly384 (to an oxygen of the 5' phosphate
group). There were also five water molecules observed within hydrogen bonding distance of the
adduct (one to the 2' hydroxyl, one bridging between the 2' and 3' hydroxyl groups, and three
proximal to the 5'-phosphate oxygen atoms).
Interestingly, the ligand binding site in the above crystal structure contained a well-
resolved pyrophosphate group held in place by hydrogen bond interactions with the side chains
of NAMPT residues Arg40', Arg392', Ser398', Lys400' and Lys423' (Figure 7; side chains not
shown for clarity). One of the terminal oxygen atoms of this pyrophosphate group was located
3.5 Å from the C1' ribose atom of the 26-ribose adduct (red dotted line in Figure 7). Given the
nearly linear relationship between the pyrophosphate oxygen, ribose carbon, and inhibitor 1H-
pyrazolo[3,4-b]pyridine nitrogen atoms (155 °), it is possible that Figure 7 depicts the conclusion
11

of a NAMPT-catalyzed process in which pyrophosphate is displaced from PRPP by the 1H-
pyrazolo[3,4-b]pyridine N-1 nitrogen atom (c.f., Figure 3). The dense network of H-bond
interactions formed between NAMPT and the pyrophosphate moiety suggest that the enzyme
preorders the PRPP co-factor and a suitable substrate (an inhibitor such as 26 or the natural
ligand NAM) and thereby facilitates the nucleophilic displacement of pyrophosphate from PRPP.
Recent kinetic isotope measurements and quantum calculations performed on the PRPP-NAM
transition state geometry also support this possibility.24
In this work, we describe the identification of potent NAMPT inhibitors which contain
the bicyclic 1H-pyrazolo[3,4-b]pyridine ring system. Many of these compounds exhibit
nanomolar antiproliferation activity against human tumor lines in in vitro cell culture
experiments, and a representative example (compound 26) demonstrated encouraging in vivo
efficacy in an A2780 mouse xenograft tumor model. However, compound 26 was also shown to
lack meaningful in vitro activity against mouse and monkey cell lines even though it inhibited
the NAMPT enzymes derived from these species. Such cell culture differences were not
observed for a previously described NAMPT inhibitor which contains an imidazo-pyridine in
lieu of the 1H-pyrazolo[3,4-b]pyridine ring system (compound 4). Importantly, the mouse cell
potencies observed for 4 and 26 correlated well with the ability of each compound to alter in vivo
NAD levels in both blood and eye tissue. The observed cell potency differences may thus
impact other in vivo outcomes in the associated animal species (e.g., safety assessments), and it
is therefore suggested that appropriate cell culture profiling of promising NAMPT inhibitors be
completed prior to conducting detailed in vivo studies with them.
12

Acknowledgements
We thank the members of the Genentech protein expression group for their help in
expressing the NAMPT proteins used in this work. We also thank the DMPK group at
Genentech, especially Jonathan Cheong, and Ning Liu, for their contribution to the generation of
the DMPK data, Rashida Garcia-Dancey and Shohini Ganguly for helping to generate the
cellular data, and Agilent Technologies (Woburn, MA) for generating the Rapidfire LCMS data.
In addition, we acknowledge the use of the synchrotron X-ray source at ALS supported by the
Department of Energy's (DOE) Office of Science. We also appreciate many helpful discussions
regarding NAMPT with Drs. Peter Jackson and Marcia Belvin.
Supplementary data
Supplementary data associated with this article can be found in the online version at
http://XXXXX
13

References and notes
1. (a) Houtkooper, R. H.; Cantó, C.; Wanders, R. J.; Auwerx, J. Endocrine Rev. 2010, 31, 194.
(b) Sauve, A. A. J. Pharmacol. Exp. Ther. 2008, 324, 883.
2. (a) Galli, U.; Travelli, C.; Massarrotti, A.; Fakhfouri, G.; Rahimian, R.; Tron, G. C.;
Genazzani, A. A. J. Med. Chem. 2013, DOI: 10.1021/jm4001049. (b) He, J.; Tu, C.; Li, M.;
Wang, S.; Guan, X.; Lin, J.; Li, Z. Curr. Pharm. Des. 2012, 18, 6123. (c) Burgos, E. S. Curr.
Med. Chem. 2011, 18, 1947. (d) Bi, T.-q.; Che, X.-m. Cancer Biol. Ther. 2010, 10, 119. (e)
Garten, A.; Petzold, S.; Körner, A.; Imai, S.-i.; Kiess, W. Trends Endocrin. Metab. 2008, 20,
130.
3. In contrast, the de novo NAD synthesis pathway requires eight enzyme-catalyzed steps to
produce the co-factor from tryptophan. See ref. 1 for additional details.
4. (a) Schulze, A.; Harris, A. L. Nature 2012, 491, 364. (b) Chiarugi, A.; Dölle, C.; Felici, R.;
Ziegler, M. Nat. Rev. Cancer 2012, 12, 741. (c) Zhang, L. Q.; Heruth, D. P.; Ye, S. Q. J.
Bioanal. Biomed. 2011, 3, 13. (d) Khan, J. A.; Forouhar, F.; Tao, X.; Tong, L. Expert Opin.
Ther. Targets 2007, 11, 695.
5. (a) Watson, M.; Roulston, A.; Bélec, L.; Billot, X.; Marcellus, R.; Bédard, D.; Bernier, C.;
Branchaud, S.; Chan, H.; Dairi, K.; Gilbert, K.; Goulet, D.; Gratton, M.-O.; Isakau, H.; Jang, A.;
Khadir, A.; Koch, E.; Lavoie, M.; Lawless, M.; Nguyen, M.; Paquette, D.; Turcotte, E.; Berger,
A.; Mitchell, M.; Shore, G. C.; Beauparlant, P. Mol. Cell. Biol. 2009, 29, 5872. (b) Ravaud, A.;
Cerny, T.; Terret, C.; Wanders, J.; Bui, B. N.; Hess, D.; Droz, J.-P.; Fumoleau, P.; Twelves, C.
Eur. J. Cancer 2005, 41, 702. (c) Hovstadius, P.; Larsson, R.; Jonsson, E.; Skov, T.; Kissmeyer,
A.-M.; Krasilnikoff, K.; Bergh, J.; Karlsson, M. O.; Lönnebo, A.; Ahlgren, J. Clin. Cancer Res.
2002, 8, 2843. Compound 1 is also known in the literature as CHS828.
14

6. Beauparlant, P.; Bédard, D.; Bernier, C.; Chan, H.; Gilbert, K.; Goulet, D.; Gratton, M.-O.;
Lavoie, M.; Roulston, A.; Turcotte, E.; Watson, M. Anti-Cancer Drugs, 2009, 20, 346.
7. (a) Olesen, U. H.; Thougaard, A. V.; Jensen, P. B.; Sehested, M. Mol. Cancer Ther. 2010, 9,
1609. (b) Holen, K.; Saltz, L. B.; Hollywood, E.; Burk, K.; Hanauske, A.-R. Invest. New Drugs
2008, 26, 45. Compound 2 is also known in the literature as FK866.
8. Zheng, X.; Bauer, P.; Baumeister, T.; Buckmelter, A. J.; Caligiuri, M.; Clodfelter, K. H.; Han,
B.; Ho, Y.-C.; Kley, N.; Lin, J.; Reynolds, D. J.; Sharma, G.; Smith, C. C.; Wang, Z.; Dragovich,
P. S.; Gunzner-Toste, J.; Liederer, B. M.; Ly, J.; O’Brien, T.; Oh, A.; Wang, L.; Wang, W.;
Xiao, Y.; Zak, M.; Zhao, G.; Yuen, P.-w.; Bair, K. W. J. Med. Chem. 2013, DOI:
10.1021/jm4008664.
9. For other recently disclosed examples of NAMPT inhibitors, see: (a) Dragovich, P. S.; Bair,
K. W.; Baumeister, T.; Ho, Y.-C.; Liederer, B. M.; Liu, X.; Liu, Y.; O’Brien, T.; Oeh, J.;
Sampath, D.; Skelton, N.; Wang, L.; Wang, W.; Wu, H.; Xiao, Y.; Yuen, P.-w.; Zak, M.; Zhang,
L.; Zheng, X. Bioorg. Med. Chem. Lett. 2013, 23, 4875. (b) Zheng, X.; Bauer, P.; Baumeister,
T.; Buckmelter, A. J.; Caligiuri, M.; Clodfelter, K. H.; Han, B.; Ho, Y.-C.; Kley, N.; Lin, J.;
Reynolds, D. J.; Sharma, G.; Smith, C. C.; Wang, Z.; Dragovich, P. S.; Oh, A.; Wang, W.; Zak,
M.; Gunzner-Toste, J.; Zhao, G.; Yuen, P.-w.; Bair, K. W. J. Med. Chem. 2013, 56, 4921. (c)
Gunzner-Toste, J.; Zhao, G.; Bauer, P.; Baumeister, T.; Buckmelter, A. J.; Caligiuri, M.;
Clodfelter, K. H.; Fu, B.; Han, B.; Ho, Y.-C.; Kley, N.; Liang, X.; Liederer, B. M.; Lin, J.;
Mukadam, S.; O’Brien, T.; Oh, A.; Reynolds, D. J.; Sharma, G.; Skelton, N.; Smith, C. C.;
Sodhi, J.; Wang, W.; Wang, Z.; Xiao, Y.; Yuen, P.-w.; Zak, M.; Zhang, L.; Zheng, X.; Bair, K.
W.; Dragovich, P. S. Bioorg. Med. Chem. Lett. 2013, 23, 3531. (d) Curtin, M. L.; Sorensen, B.
K.; Heyman, H. R.; Clark, R. F.; Woller, K. R.; Shah, O. J.; Michaelides, M.; Tse, C.;
15

Vasudevan, A.; Mack, H.; Hansen, T. M.; Sweis, R.; Pliushchev, M. A. US 2012/0122842. (e)
Kumar, D. V.; Slattum, P. M.; Yager, K. M.; Shenderovich, M. D.; Tangallapally, R.; Kim, S.-H.
WO 2012/177782. (f) You, H.; Youn, H.-S.; Im, I.; Bae, M.-H.; Lee, S.-K.; Ko, H.; Eom, S. H.;
Kim, Y.-C. Eur. J. Med. Chem. 2011, 46, 1153.
10. (a) Chiang, P.-W.; Wang, J.; Chen, Y.; Fu, Q.; Zhong, J.; Chen, Y.; Yi, X.; Wu, R.; Gan, H.;
Shi, Y.; Chen, Y.; Barnett, C.; Wheaton, D.; Day, M.; Sutherland, J.; Heon, E.; Weleber, R. G.;
Gabriel, L. A. R.; Cong, P.; Chuang, K.; Ye, S.; Sallum, J. M. F.; Qi, M. Nat. Genet. 2012, 44,
972. (b) Perrault, I.; Hanein, S.; Zanlonghi, X.; Serre, V.; Nicouleau, M.; Defoort-Delhemmes,
S.; Delphin, N.; Fares-Taie, L.; Gerber, S.; Xerri, O.; Edelson, C.; Goldenberg, A.; Alice
Duncombe, A.; Le Meur, G.; Hamel, C.; Silva, E.; Nitschke, P.; Calvas, P.; Munnich, A.; Roche,
O.; Dollfus, H.; Kaplan1 J.; Rozet, J.-M. Nat. Genet. 2012, 44, 975. (c) Koenekoop, R. K.;
Wang, H.; Majewski, J.; Wang, X.; Lopez, I.; Ren, H.; Chen, Y.; Li, Y.; Fishman, G. A.;
Genead, M.; Schwartzentruber, J.; Solanki, N.; Traboulsi, E. I.; Cheng, J.; Logan, C. V.;
McKibbin, M.; Hayward, B. E.; Parry, D. A.; Johnson, C. A.; Nageeb, M.; Poulter, J. A.;
Mohamed, M. D.; Jafri, H.; Rashid, Y.; Taylor, G. R.; Keser, V.; Mardon, G.; Xu, H.;
Inglehearn, C. F.; Fu, Q.; Toomes C.; Chen, R. Nat. Genet. 2012, 44, 1035. (d) Falk, M. J.;
Zhang, Q.; Nakamaru-Ogiso, E.; Kannabiran, C.; Fonseca-Kelly, Z.; Chakarova, C.; Audo, I.;
Mackay, D. S.; Zeitz, C.; Borman, A. D.; Staniszewska, M.; Shukla, R.; Palavalli, L.; Mohand-
Said, S.; Waseem, N. H.; Jalali, S.; Perin, J. C.; Place, E.; Ostrovsky, J.; Xiao, R.; Bhattacharya,
S. S.; Consugar, M.; Webster, A. R.; Sahel, J.-A.; Moore, A. T.; Berson, E. L.; Liu, Q.; Gai, X.;
Pierce E. A. Nat. Genet. 2012, 44, 1040.
11. Humans who are heterozygous carriers of the described NMNAT mutations do not exhibit
the LCA phenotype. In addition, the enzymatic activity of the mutant NMNAT proteins is only
16

partially impaired relative to wild type enzyme, and NAD levels measured in RBCs from a
homozygous mutation carrier were approximatley 70% of those observed in RBCs from a
heterozygous comparator (see ref. 10a-d for additional details).
12. Ophthalmologic evaluations (ERG and visual acuity) were a component of patient
monitoring performed during the clinical trials of compound 2. No concerns regarding the
retinal health of participating patients were noted. See ref. 7b for additional details.
13. HSP90 inhibitors which exhibit reduced retinal penetration and/or retention cause minimal
(none) retinal damage in rodents relative to molecules with higher retinal exposures. See: Zhou,
D.; Teofilovici, F.; Liu, Y.; Ye, J.; Ying, W.; Ogawa, L. S.; Inoue, T.; Lee, W.; Adjiri-Awere,
A.; Kolodzieyski, L.; Tatsuta, N.; Wada, Y.; Sonderfan, A. J. American Society of Clinical
Oncology Annual Meeting, Chicago, IL, June 1-5, 2012; Abstract #3086.
14. (a) Hosoya, K.-i.; Tomi, M.; Tachikawa, M. Expert Opin. Drug Deliv. 2011, 8, 1571. (b)
Toda, R.; Kawazu, K.; Oyabu, M.; Miyazaki, T.; Kiuchi, Y. J. Pharm. Sci. 2011, 100, 3904. (c)
Kadam, R. S.; Kompella, U. B. J. Pharmacol. Exp. Ther. 2010, 332, 1107.
15.
We currently believe that many potent NAMPT inhibitors form PRPP-derived
phosphoribosylated adducts in the protein’s active site which block the function of the enzyme.
This belief is consistent with the repeated observation of these adducts by mass spectrometry in
biochemical experiments (see refs. 5a, and 8) and with crystallographic observation of these
entities in complex with NAMPT (ref. 22 and this work). Once formed, the PRPP-adducts may
accumulate intracellularly and thereby enhance cell culture antiproliferation effects (see ref. 5a
for additional information and discussion). However, there are many other factors that also
likely influence NAMPT inhibitor cell potency including: biochemical potency, the ability of a
given inhibitor and/or its corresponding PRPP-derived ribose adduct to effectively compete with
17

the NAM substrate, cell membreane permeability, and/or protein binding. In addition,
performing the PRPP-adduct assessments was a relatively resource-intensive task and such
determinations were therefore made for only a small number of compounds.
16. The antiproliferative effects exhibited in cell culture by 13 were completely eliminated
("reversed") when the compound was tested in the presence of 0.33 mM NMN (the product of
the NAMPT-catalyzed condensation of nicotinamide and PRPP; c.f., Figure 1). This result
strongly suggests that the observed cell-based effects result from NAMPT inhibition. Related
NMN experiments were conducted in cell-based assessments with all inhibitors reported in this
work and all displayed similar reversals of their antiproliferative effects.
17. All compounds which contain the bicyclic heterocycle present in inhibitor 13 are referrd to
as 1H-pyrazolo[3,4-b]pyridines throughout this work.
18. NAMPT is believed to function as a symmetrical homodimer with two active sites formed at
opposite ends of the dimer interface. Accordingly, the protein crystalized with such a dimer
present in the asymmetric unit. The two monomer chains were highly similar (RMSD 0.13 Å for
413 C-alpha atoms, as defined by the protein alignment tool in PyMOL) and both active sites
contained a molecule of compound 14 in a similar orientation. In the crystallography discussion,
the NAMPT residues are designated with prime and non-prime notation (e.g., Tyr18', Phe193) to
distinguish the monomer chain in which a given residue resides. PDB code = 4M6P.
19. A caveat with comparing the retinal exposures of compounds 4 and 26 is that the described
results were obtained from experiments that employed different dosing regimens. Compound 4
was dosed for two consecutive days and was co-administered with nicotinic acid while
compound 26 was administered once as a single-agent. However, based on the measured
disappearance of 4 from rat retina over time (Table 5), we do not believe that significant
18

accumulation of the compound occurred during the 2-day experiment relative to what would be
observed following a single dose. We also do not believe that nicotinic acid co-administration
impacted the retinal exposures of 4, although we did not generate data to explore this hypothesis.
20. Day-5 exposures of 4 and 26 are depicted in Figure S5A. The day-1 exposure of each
compound was very similar to the corresponding day-5 exposure (data not shown). Compound
exposures on other study days were not assessed.
21. The mouse plasma protein binding properties of inhibitors 4 and 26 differ by <6-fold (fu =
0.067 and 0.013, respectively) and thus do not easily explain the NAD PK-PD study outcomes.
22. PDB code = 4M6Q. The two NAMPT monomer chains which formed the symmetrical
protein homodimer were highly similar (RMSD 0.17 Å for 426 C-alpha atoms as defined by the
protein alignment tool in PyMOL) and both active sites contained a molecule of the 26-PRPP
adduct in a similar orientation
23. Wang, W. manuscript in preparation.
24. Burgos, E. S.; Vetticatt, M. J.; Schramm, V. L. J. Am. Chem. Soc. 2013, 135, 3485.
19

Table 1. Initial structure-activity relationships.
R1
R2
A
NAMPT IC50 A2780 IC50
Cmpd.
(uM)a
(uM)b
0.028
0.006
3
9
B
A
0.018
>2.0
0.013
>2.0
10
11
12
13
B
B
A
B
1.3
>2.0
0.10
0.054
>2.0
>2.0
0.088
0.066
14
15
B
B
1.7
1.9
>2.0
>2.0
16
See supplementary data for experimental details associated with each assessment. All
biochemical and cell-based assay results are reported as the arithmetic mean of at least 2
separate runs.
b
aNAMPT biochemical inhibition. Antiproliferation activity determined in cell culture
experiments using A2780 cell line. This inhibition can be reversed by addition of 0.33
mM of NMN.
1

Table 2. Structure-activity relationships of 1H-pyrazolo[3,4-b]pyridine-containing NAMPT
inhibitors.
NAMPT IC50 A2780 IC50 HLM CLhep
Cmpd.
R
(uM)a
(uM)b
(mL/min/kg)c
0.10
0.088
18
13
17
18
0.41
0.40
>2.0
16
17
0.039
0.41
0.39
0.21
0.33
11
19
20
5.4
0.30
0.035
8.6
21
0.44
0.37
0.045
0.28
7.6
6.7
22
23
0.11
0.045
0.066
11
24
14
0.054
4.4
0.0079
0.0061
0.0077
0.0060
0.0043
0.0048
9.1
8.8
7.6
25
26
27
0.010
0.0045
7.2
28
2

0.011
0.0045
0.012
0.0024
0.0013
0.0032
13
11
14
29
30
31
0.0082
0.19
0.0022
0.031
13
6.0
6.3
4.4
32
33
34
35
0.010
0.099
0.0069
0.012
See supplementary data for experimental details associated with each assessment. All
biochemical, cell-based assay, and HLM results are reported as the arithmetic mean of at
least 2 separate runs.
aNAMPT biochemical inhibition. bAntiproliferation activity determined in cell culture
experiments using A2780 cell line. This inhibition can be reversed by addition of 0.33
mM of NMN. cHuman hepatic clearance predicted from human liver microsomes (stable,
moderate, labile = <6, 6-15, >15 mL/min/kg, respectively).
3

Table 3. In vitro DMPK properties of selected NAMPT inhibitors.
MLM CLhep
MH CLhep
MDCK Papp,A-B
;
Mouse
PPB (fu)d
0.114
0.013
0.008
0.050
0.074
0.019
0.018
0.009
0.071
Sol.
Cmpd.
(uM)e
(mL/min/kg)a (mL/min/kg)b
x10-6 cm/sc
7.8
69
42
42
65
76
65
77
73
63
22
13
8
18
24
31
45
44
27
<1.0
1.4
25
26
27
28
29
30
31
32
34
11
4.0
4.4
2.5
0.85
1.5
1.6
<1.0
<1.0
<1.0
<1.0
<1.0
<1.0
8.8
1.9
See supplementary data for experimental details associated with each assessment. All
assay results are reported as the arithmetic mean of at least 2 separate runs (exception:
solubility determinations are n = 1).
aMouse hepatic clearance predicted from mouse liver microsomes (stable, moderate,
b
labile = <27, 27-63, >63 mL/min/kg, respectively). Mouse hepatic clearance predicted
from mouse hepatocytes (stable, moderate, labile = <27, 27-63, >63 mL/min/kg,
respectively). Apparent permeability coefficients (Papp,A-B) across MDCK cell monolayers
e
dMouse plasma protein binding; fraction unbound (fu). Aqueous solubility.
4

Table 4. In vivo DMPK properties of compound 26 in female nude mice.
Route
Dose
Cmax
Tmax
(h)c
Vss
CL
T1/2
(mg/kg)a (µM)b
(L/kg)d (mL/min/kg)e (h)f
IV
PO
5
100
NA
53
NA
0.25
0.45
NA
3.3
NA
2.6
3.4
aFormulations: IV = 35:10:15:40 PEG400:EtOH:PG:Water (solution); PO = 60:10:30
b
c
PEG400:EtOH:Water (solution). Maximum plasma concentration. Time at which maximum
d
e
f
plasma concentration observed. Volume of distribution (steady-state). Clearance. Half-life.
NA = not applicable. See supplementary data for experimental descriptions.
5

Table 5. Retina concentration and retina/plasma ratio (both uncorrected for protein binding
effects) determined in rats for compounds 4 and 26 at indicated timepoints following a 30 mg/kg
oral dose of each molecule (n = 4 for 4; n = 3 for 26). Standard deviations are also provided.
See supplementary data for additional information.
[Retina] (µM)
Retina/Plasma Ratio
Cmpd.
1 h
6.5 h
1 h
6.5 h
4a
26
1.9 ± 0.82
0.94 ± 0.23
1.3 ± 0.31
0.68 ± 0.23
0.32 ± 0.11
0.30 ± 0.12
0.071 ± 0.006 0.093 ± 0.036
aCompound was dosed QD for 2 consecutive days in combination with 75
mg/kg nicotinic acid that was administered orally BID. Data were obtained
following day 2 dose.
6

Table 6. Biochemical inhibition activity (IC50 in nM) of compounds 4 and 26 as determined
against NAMPT enzymes from indicated species.a,b
NAMPT Biochemical IC50 Values (nM)
Cmpd.
Human
Mouse
24
Rat
23
Dog
7
Monkey
42
5
6
4
26
59
35
22
73
b
aReported as the arithmetic mean of at least 2 separate runs. PRPP-ribose adduct formation
detected for both compounds in every species. See supplementary data for additional
information.
7

Table 7. Antiproliferation effects of compounds 4 and 26 determined in various cell lines from
different animal species.a
CyQuant IC50 values in nM
Species
Human
Cell Line
HT1080
PC3
MiaPaCa2
HCT116
Calu6
Cmpd. 4
Cmpd. 26
2
3
7
2
5
9
9
30
8
21
KPP
KPR
B16F10
A-72
16
41
71
4
775
>1000
>1000
23
Mouse
Dog
D-17
DH-82
Cos1
Cos7
RF/6A
12
7
5
9
4
98
41
501
863
386
Monkey
aReported as the arithmetic mean of 3 separate runs. See supplementary data for additional
information.
8

Table 8. Ability of compounds 4 and 26 to reduce NAD levels in human and mouse cell lines.
PC3 (human)
KPP (mouse)
Cmpd.
EC50 (nM)a [Cell-1]b [Cell-2]c EC50 (nM)a [Cell-1]b [Cell-2]c
1.1
1.2
9
10
30
54
1.3
70
8
12
31
74
4
26
Data are reported as average of two independent experiments which utilized triplicate samples in
each run. See supplementary data for additional information.
aNAD inhibition activity in given cell line. bIntracellular concentration (nM) of 4 and 26
measured during NAD inhibition experiment in given cell line; extracellular concentration in
c
well = 200 nM. Intracellular concentration (nM) of 4 and 26 measured during NAD inhibition
experiment in given cell line; extracellular concentration in well = 1000 nM.
9