Synthesis, Structure-Activity Relationship, and Pharmacophore Modeling Studies of Pyrazole-3-Carbohydrazone Derivatives as Dipeptidyl Peptidase IV Inhibitors

Article abstract of DOI:10.1111/j.1747-0285.2012.01365.x

Type 2 diabetes mellitus (T2DM) is a metabolic disease and a major challenge to healthcare systems around the world. Dipeptidyl peptidase IV (DPP-4), a serine protease, has been rapidly emerging as an effective therapeutic target for the treatment for T2DM. In this study, a series of novel DPP-4 inhibitors, featuring the pyrazole-3-carbohydrazone scaffold, have been discovered using an integrated approach of structure-based virtual screening, chemical synthesis, and bioassay. Virtual screening of SPECS Database, followed by enzymatic activity assay, resulted in five micromolar or low-to-mid-micromolar inhibitory level compounds (1-5) with different scaffold. Compound 1 was selected for the further structure modifications in considering inhibitory activity, structural variability, and synthetic accessibility. Seventeen new compounds were synthesized and tested with biological assays. Nine compounds (6e, 6g, 6k-l, and 7a-e) were found to show inhibitory effects against DPP-4. Molecular docking models give rational explanation about structure-activity relationships. Based on eight DPP-4 inhibitors (1-5, 6e, 6k, and 7d), the best pharmacophore model hypo1 was obtained, consisting of one hydrogen bond donor (HBD), one hydrogen bond acceptor (HBA), and two hydrophobic (HY) features. Both docking models and pharmacophore mapping results are in agreement with pharmacological results. The present studies give some guiding information for further structural optimization and are helpful for future DPP-4 inhibitors design.

Full text of DOI:10.1111/j.1747-0285.2012.01365.x

ª 2012 John Wiley & Sons A/S
doi: 10.1111/j.1747-0285.2012.01365.x
Chem Biol Drug Des 2012; 79: 897–906
Research Article
Synthesis, Structure–Activity Relationship, and
Pharmacophore Modeling Studies of Pyrazole-3-
Carbohydrazone Derivatives as Dipeptidyl
Peptidase IV Inhibitors
Deyan Wu1,†, Fangfang Jin1,†, Weiqiang
Key words: dipeptidyl peptidase IV, inhibitors, molecular docking,
pharmacophore modeling, synthesis
Lu1,†, Jin Zhu1, Cui Li1, Wei Wang1, Yun
Tang1, Hualiang Jiang1,2, Jin Huang1,*,
Guixia Liu1,* and Jian Li1,*
Received 19 October 2011, revised 11 January 2012 and accepted for
publication 17 February 2012
1Shanghai Key Laboratory of New Drug Design, School of
Type 2 diabetes mellitus (T2DM) is a metabolic disease and has
become a worldwide epidemic. Currently, approximate 194 million
people in worldwide are affected by this disease and this number is
forecasted to increase to 366 million by 2030a'.b The health, social,
and economic burden of T2DM is significant; consequently, T2DM
presents a major challenge to healthcare systems around the world.c
Pharmacy, East China University of Science and Technology, 130
Mei Long Road, Shanghai 200237, China
2Drug Discovery and Design Center, Shanghai Institute of Materia
Medica, Chinese Academy of Sciences, 555 Zu Chong Zhi Road,
Shanghai 201203, China
*Corresponding authors: Jin Huang, huangjin@ecust.edu.cn; Guixia
Liu, gxliu@ecust.edu.cn; Jian Li, jianli@ecust.edu.cn
ꢀThese authors contributed equally to this work.
The current oral treatment options for T2DM include metformin, sul-
fonylurea (SU) or thiazolidinedione (TZD) derivatives, glycosidase
inhibitors, and the recently introduced dipeptidyl peptidase IV (DPP-
4) inhibitors. DPP-4 is a serine protease that is ubiquitously
expressed as both a membrane-bound protein and a soluble protein
in plasma (1,2). It mediates the activities of regulatory peptides by
cleaving dipeptides from the N-terminus of glucagon-like peptide-1
amide (GLP-1-NH2) and glucose-dependent insulinotropic polypep-
tide (GIP) to yield inactive GLP-1-NH2 peptide and GIP, respectively
(3,4). Consequently, inhibition of DPP-4 is rapidly emerging as a
novel therapeutic approach for the treatment for type 2 diabetes.
To data, a large number of DPP-4 inhibitors have been reported in
the literature. Among them, MK-0431 (Sitagliptin) (5), which is the
first approved DPP-4 inhibitor, has been prescribed in the United
States since 2006. LAF-237 (Vildagliptin) (6) was approved in Europe
in 2007, and BMS-477118 (Saxagliptin) (7) has been approved and
on the market in the United States in 2009. SYR-322 (Alogliptin) (8)
was approved in Japan in 2010, and BI-1356 (Linagliptin) (9) was
approved by the US FAD in May 2011.
Type 2 diabetes mellitus (T2DM) is a metabolic dis-
ease and a major challenge to healthcare systems
around the world. Dipeptidyl peptidase IV (DPP-4),
a serine protease, has been rapidly emerging as an
effective therapeutic target for the treatment for
T2DM. In this study, a series of novel DPP-4 inhibi-
tors, featuring the pyrazole-3-carbohydrazone scaf-
fold, have been discovered using an integrated
approach of structure-based virtual screening,
chemical synthesis, and bioassay. Virtual screening
of SPECS Database, followed by enzymatic activity
assay, resulted in five micromolar or low-to-mid-
micromolar inhibitory level compounds (1–5) with
different scaffold. Compound 1 was selected for
the further structure modifications in considering
inhibitory activity, structural variability, and syn-
thetic accessibility. Seventeen new compounds
were synthesized and tested with biological assays.
Nine compounds (6e, 6g, 6k–l, and 7a–e) were found
to show inhibitory effects against DPP-4. Molecular
docking models give rational explanation about
structure–activity relationships. Based on eight
DPP-4 inhibitors (1–5, 6e, 6k, and 7d), the best phar-
macophore model hypo1 was obtained, consisting
of one hydrogen bond donor (HBD), one hydrogen
bond acceptor (HBA), and two hydrophobic (HY)
features. Both docking models and pharmacophore
mapping results are in agreement with pharmaco-
logical results. The present studies give some guid-
ing information for further structural optimization
and are helpful for future DPP-4 inhibitors design.
With the aim to find new scaffold inhibitors of DPP-4, structure-
based virtual screening had been employed in our laboratory to
screen the SPECS Database against the crystal structure of DPP-4
(PDB code: 2P8S) (10). And five micromolar or low-to-mid-micromo-
lar inhibitory level compounds (1–5) with different scaffold were
successfully identified (Figure 1) after bioassay verification. Consid-
ering inhibitory activity, structural variability, and synthetic accessi-
bility of compound 1, it may serve as a reasonable lead compound
for the further structure modifications. To provide expedient and
significant structure–activity relationship (SAR) information and
improve inhibitory activity of the lead compound 1, chemical
897

Wu et al.
Figure 1: The structures and
their DPP-4 inhibitory activities of
compounds 1–5 selected from the
candidates by virtual screening and
bioassay.
modifications were performed in two cycles, by maintaining the pyr-
azole-3-carbohydrazone scaffold. Totally, seventeen analogs (6a–l
and 7a–e) of compound 1 have been synthesized and tested
against DPP-4. Finally, nine compounds (6e, 6g, 6k–l, and 7a–e)
were found to show micromolar or low-to-mid-micromolar level
inhibitory effects against DPP-4, generally 2–35 times lower than
the lead compound 1. To reasonably explain the activity loss of
compounds 6–7, subsequently, the SARs of the pyrazole-3-carbo-
hydrazone scaffold (1 and 6–7) were summarized tentatively; the
binding modes of inhibitors 6k and 7c to DPP-4 were compared
with that of inhibitor 1 by means of the molecular docking; eight
compounds (1–5, 6e, 6k, and 7d) were employed to generate the
pharmacophore model of DPP-4 inhibitors. These results give some
guiding information for further structural optimization and are help-
ful for future DPP-4 inhibitors design.
pounds were built within Catalyst (Accelrys Inc.).f Compound 1 was
considered as the reference compound specifying a Principal value of
2 and a MaxOmitFeat value of 0, meaning its structure and conforma-
tion would have the strongest influence on the model building phase.
The Principal value and MaxOmitFeat value for the remaining com-
pounds were set to 1 and 1, respectively. The Poling algorithm imple-
mented within Catalyst was used to generate conformations for all of
the compounds. For each compound, possible diverse sets of confor-
mations were generated over a 20 kcal ⁄ mol range using the BEST
flexible conformation generation option available in Catalyst. The
chemical features considered in the pharmacophore model generation
run were H-bond acceptor (HBA), H-bond donor (HBD), hydrophobic
(HY), and positively ionizable (PI) features. HipHop was set to consider
these features in the generation of the pharmacophore hypotheses.
Synthesis
Experimental Section
General methods
Computational modeling
Anhydrous ethyl alcohol was dried by distillation from calcium
hydride. Other solvents and commercial chemicals were purchased
at the highest commercial quality and we reused without further
purification, unless otherwise indicated. Reactions requiring anhy-
drous conditions were performed under nitrogen or a calcium chlo-
ride tube. Melting points were determined on a SGW X-4 melting
point apparatus without correction. Nuclear magnetic resonance
(NMR) spectroscopy was obtained on a Bruker-400, in CDCl3 or
DMSO-d6 solutions, and chemical shifts were measured in ppm
downfield from an internal tetramethylsilane (TMS) standard. The
following abbreviations are used: s (singlet), d (doublet), t (triplet), q
(quartet), bs (broad singlet), m (multiplet), etc. Coupling constants
were reported in Hz. Low- and high-resolution mass spectra (LRMS
and HRMS) were given with electric and matrix-assisted laser
desorption ionization (EI and MALDI) produced by Finnigan MAT-95
and IonSpec 4.7 T. The products were purified by recrystallization or
column chromatography on silica gel (200–300 mesh). Thin-layer
chromatography (TLC) was carried out with HSGF 254 (150–200 lm
thickness; Yantai Huiyou Co., China), and components were visual-
ized by observation under UV light (254 and 365 nm).
Molecular docking
Docking experiments were performed using GLIDE 5.0 included in
SCHRꢀDINGER Package, MASTRO INTERFACE version 9.0.d The co-ordinates
for the X-ray crystal structure of DPP-4 (PDB code: 2P8S) (10) was
obtained from the RCSB Protein Data Bank. Before the molecular
docking process, the ligand of the crystal structure was removed
and then the free protein was modified by adding hydrogen atoms
using the protein preparation workflow. Molecules to be docked
were prepared and energy-minimized according to standard proce-
dure of LigPrep module implemented in MAESTRO 9.0.e
In the molecular docking process, standard-precision (SP) docking
method was adopted to generate the minimized pose, and the Glide
scoring function (Glide Score) was used to select the final poses for
each ligand.
Ligand-based pharmacophore modeling
Common feature pharmacophore hypotheses were generated using
the HipHopRefine algorithm based on eight DPP-4 inhibitors (1–5,
6e, 6k, and 7d). The structures and conformations of the eight com-
Ethyl 2-oxo-2-(2-oxocyclohexyl)acetate (9a): A solution of Sodium
ethoxide was prepared by the cautious addition of 2.55 g
898
Chem Biol Drug Des 2012; 79: 897–906

Studies of Pyrazole-3-Carbohydrazone Derivatives
(110 mmol) of sodium to 40 mL of anhydrous ethyl alcohol in a
250 mL three-necked flask equipped with a dropping funnel, and a
reflux condenser carrying a calcium chloride tube. The flask was
then immersed in an ice bath and the stirrer was started. Then, a
cold solution of 10.5 mL (100 mmol) of cyclohexanone in 13.5 mL
(100 mmol) of ethyl oxalate was added from the dropping funnel
over a period of about 15 min. When the addition was complete,
the ice bath was retained for 1 h, and then the mixture was stirred
at room temperature for another 6 h (11). The reaction mixture was
added to ice water and acidified by the addition of 2 N HCl. The
aqueous solution was extracted with CH2Cl2 (3 · 40 mL), washed
with brine, dried over anhydrous MgSO4, and evaporated to give
the residue, which was purified by flash column chromatography on
silica gel, eluted with a mixture of EtOAc ⁄ petroleum ether (1:25,
v ⁄ v), to afford 9a (11.5 g, 59%) as a solid. 1HNMR (400 MHz,
CDCl3) d 6.27 (s, 1H), 5.32 (t, J = 4.8 Hz, 1H), 3.77 (q, J = 7.2 Hz,
2H), 3.00 (t, J = 5.6 Hz, 2H), 2.26 (q, J = 5.6 Hz, 2H), 1.68 (m, 2H),
1.34 (t, J = 7.2 Hz, 3H).
N'-[(4-Cyano-phenyl)methylene]-4,5,6,7-tetrahydro-1H-indazole-3-carbo-
hydrazone (6b): Starting with 11a and 4-cyanobenzaldehyde, the
same procedures were followed as in the synthesis of 1.
Yield = 84%; mp = 314–328 ꢀC; 1HNMR (400 MHz, DMSO-d6) d
12.98 (s, 1 H), 11.80 (s, 1H), 8.55 (s, 1H), 7.90 (d, J = 8.4 Hz, 2H),
7.83 (d, J = 8.4 Hz, 2H), 2.69 (t, J = 6.0 Hz, 2H), 2.63 (t, J = 6.0 Hz,
2H), 1.75–1.69 (m, 4H); EI-MS m ⁄ z 293.1 (M+), 149.1 (100%); HRMS
(EI) m ⁄ z calcd C16H15N5O (M+) 293.1277; found 293.1278.
N'-[(4-Chloro-3-fluorophenyl)methylene]-4,5,6,7-tetrahydro-1H-indazole-
3-carbohydrazone(6c): Starting with 11a and 4-chloro-3-fluoro-
benzaldehyde, the same procedures were followed as in the synthe-
sis of 1. Yield = 80%; mp = 311–315 ꢀC; 1HNMR (400 MHz,
DMSO-d6) d 12.96 (s, 1H), 11.73 (s, 1H), 8.45 (s, 1H), 7.66 (m, 2H),
7.52 (d, J = 8.4 Hz, 1H), 2.68 (t, J = 6.0 Hz, 2H), 2.63 (t, J = 6.0 Hz,
2H), 1.75–1.68 (m, 4H); EI-MS m ⁄ z 320.1 (M+), 149.1 (100%); HRMS
(EI) m ⁄ z calcd C15H14ClFN4O (M+) 320.0840; found 320.0844.
N'-[(4-Chloro-3-nitrophenyl)methylene]-4,5,6,7-tetrahydro-1H-indazole-
3-carbohydrazone (6d): Starting with 11a and 4-chloro-3-nitro-
benzaldehyde, the same procedures were followed as in the synthe-
sis of 1. Yield = 65%; mp = 299–302 ꢀC; 1HNMR (400 MHz,
DMSO-d6) d 13.00 (s, 1H), 11.89 (s, 1H), 8.55 (s, 1H), 8.31 (s, 1H),
7.97 (d, J = 8.4Hz, 1H), 7.84 (d, J = 8.4 Hz, 1H), 2.68 (t, J = 5.6 Hz,
2H), 2.63 (t, J = 6.0 Hz, 2H), 1.76–1.69 (m, 4H); EI-MS m ⁄ z 347.1
(M+), 149.1 (100%); HRMS (EI) m ⁄ z calcd C15H14ClN5O3 (M+)
347.0785; found 347.0786.
Ethyl 4,5,6,7-tetrahydro-1H-indazole-3-carboxylate(10a): Ten grams
(50.45 mmol) 9a was dissolved in 20 mL of acetic acid at 0 ꢀC. A
total of 3.2 mL (56.50 mmol) of 85% hydrazine hydrate was slowly
added. The mixture was heated to reflux for 10 h. After cooling of
the sample, the solid matter was filtered and dried in vacuo, result-
ing in 10a (9.6 g, 98%) of white crystals (12); HNMR (400 MHz,
CDCl3) d 4.38 (q, J = 7.2 Hz, 2H), 2.76 (t, J = 6.0 Hz, 2H), 2.70 (t,
J = 6.0 Hz, 2H), 1.84–1.75 (m, 4H), 1.39 (t, J = 7.2 Hz, 3H).
1
4,5,6,7-Tetrahydro-1H-indazole-3-carbohydrazide (11a): A mixture
of 5.0 g (25.7 mmol) 10a and 24.0 mL (350 mmol) of 85% hydrazine
hydrate in 100 mL of ethanol was refluxed for 36 h then cooled. The
solid collected by filtration, washed with water, and dried in vacuo,
gave 11a (4.6 g, 100%) (13); HNMR (400 MHz, DMSO-d6) d 2.74 (t,
J = 5.6 Hz, 2H), 2.66 (t, J = 5.6 Hz, 2H), 1.83–1.77 (m, 4H).
N'-[1-(2,4-Dihydroxyphenyl)ethylidene]-4,5,6,7-tetrahydro-1H-indazole-
3-carbohydrazone (6e): A mixture of 11a (180 mg, 1 mmol), 2,4-
dihydroxyacetophenone (167 mg, 1.1 mmol), and acetic acid (20 lL)
in 5 mL of ethanol was refluxed for 16 h then cooled. The solid col-
lected by filtration, washed with ethanol, and dried in vacuo, gave
1
1
6e (85 mg, 27%); mp = 292–304 ꢀC; HNMR (400 MHz, DMSO-d6)
d 13.44 (s, 1H), 12.97 (s, 1H), 10.58 (s, 1H), 9.56 (s, 1H), 7.43 (d,
J = 8.8 Hz, 1H), 6.32 (d, J = 8.8 Hz, 1H), 6.27 (s, 1H), 2.69 (t,
J = 5.2 Hz, 2H), 2.62 (t, J = 5.6 Hz, 2H), 2.35 (s, 3H), 1.75–1.64 (m,
4H); EI-MS m ⁄ z 314.1 (M+), 149.1 (100%); HRMS (EI) m ⁄ z calcd
C16H18N4O3 (M+) 314.1379; found 314.1380.
N'-[(2-Hydroxy-naphthalen-1-yl)methylene]-4,5,6,7-tetrahydro-1H-indazole-
3-carbohydrazone (1): A mixture of 11a (180 mg, 1 mmol), 2-
Hydroxy-1-naphthaldehyde (190 mg, 1.1 mmol),and acetic acid
(20 lL) in 5 mL of ethanol was refluxed for 6 h, then cooled. The
solid collected by filtration, washed with ethanol, and dried in
vacuo, gave
1
(318 mg, 95%); mp = 298–317 ꢀC; 1HNMR
N'-[3,4-Dihydro-1(2H)-naphthalen-1-yl]-4,5,6,7-tetrahydro-1H-indazole-
3-carbohydrazone (6f): Starting with 11a and 3,4-dihydro-1(2H)-
naphthalenone, the same procedures were followed as in the
synthesis of 6e. Yield = 64%; mp = 269–274 ꢀC; 1HNMR
(400 MHz, DMSO-d6) d 12.93 (s, 1H), 10.12 (s, 1H), 8.05 (d,
J = 7.2 Hz, 1H), 7.31 (t, J = 7.2 Hz, 1H), 7.27 (t, J = 7.6 Hz, 1H),
7.21 (d, J = 7.6 Hz, 1H), 2.77(t, J = 6.0 Hz, 2H), 2.74–2.69 (m, 4H),
2.68 (s, 3H), 2.64 (t, J = 5.6 Hz, 2H), 1.91–1.84 (m, 2H), 1.76–1.69
(m, 4H); EI-MS m ⁄ z 308.1 (M+), 149.1 (100%); HRMS (EI) m ⁄ z calcd
C19H18ClN5O (M+) 308.1637; found 308.1636.
(400 MHz, DMSO-d6) d 13.05 (s, 1H), 12.98 (s, 1H), 11.97 (s, 1H),
9.66 (s, 1H), 8.12 (d, J = 8.4 Hz, 1H), 7.91 (t, J = 7.2 Hz, 2H), 7.60
(t, J = 7.6 Hz, 1H), 7.41 (t, J = 7.2 Hz, 1H), 7.22 (d, J = 8.8 Hz, 1H),
2.72 (t, J = 5.2 Hz, 2H), 2.65 (t, J = 5.6 Hz, 2H), 1.77–1.71 (m, 4H);
EI-MS m ⁄ z 334.1 (M+), 149.1 (100%); HRMS (EI) m ⁄ z calcd
C19H18N4O2 (M+) 334.1430, found 334.1436.
N'-[(2-Hydroxy-phenyl)methylene]-4,5,6,7-tetrahydro-1H-indazole-
3-carbohydrazone (6a): Starting with 11a and 2-Hydroxybenzalde-
hyde, the same procedures were followed as in the synthesis of 1.
Yield = 48%; mp = 280–287 ꢀC; 1HNMR (400 MHz, DMSO-d6) d
12.99 (s, 1H), 11.92 (s, 1H), 11.54 (s, 1H), 8.66 (s, 1H), 7.39 (d,
J = 7.6 Hz, 1H), 7.29 (t, J = 8.8 Hz, 1H), 6.94 (d, J = 7.6 Hz, 1H),
6.90 (t, J = 7.2 Hz, 1H), 2.69 (t, J = 5.6 Hz, 2H), 2.63 (t, J = 6.0 Hz,
2H), 1.76–1.69 (m, 4H); EI-MS m ⁄ z 284.1 (M+), 149.1 (100%); HRMS
(EI) m ⁄ z calcd C15H16N4O2 (M+) 284.1273; found 284.1272.
N'-[(1-methyl-1H-pyrrol-2-yl)methylene]-4,5,6,7-tetrahydro-1H-indazole-
3-carbohydrazone (6g): Starting with 11a and N-methylpyrrole-2-
carboxaldehyde, the same procedures were followed as in the
1
synthesis of 1. Yield = 73%; HNMR (400 MHz, DMSO-d6) d 12.86
(s, 1H), 11.17 (s, 1H), 8.42 (s, 1H), 6.92 (s, 1H), 6.42 (s, 1H), 6.08 (s,
1H), 3.83 (s, 3H), 2.67 (t, J = 6.0 Hz, 2H), 2.60 (t, J = 6.0 Hz, 2H),
Chem Biol Drug Des 2012; 79: 897–906
899

Wu et al.
1.78–1.69 (m, 4H); EI-MS m ⁄ z 271 (M+), 149 (100%); HRMS (EI)
m ⁄ z calcd C14H17N5O (M+) 271.1433; found 271.1433.
Ethyl 1,4,5,6-tetrahydro-cyclopenta[c]pyrazole-3-carboxylate (10b): Start-
ing with 9b, the same procedures were followed as in the synthesis
of 10a. Yield = 33%; 1HNMR (400 MHz, CDCl3)
d 4.38 (q,
N'-[(2,4-dihydroxyphenyl)methylene]-4,5,6,7-tetrahydro-1H-indazole-3-
carbohydrazone (6h): Starting with 11a and 3,4-dihydroxybenzal-
dehyde, the same procedures were followed as in the synthesis of
J = 7.2 Hz, 2H), 2.85–2.77 (m, 4H), 2.54–2.46 (m, 2H), 1.40 (t,
J = 7.2 Hz, 3H).
1
1. Yield = 73%; mp = 268–272 ꢀC; HNMR (400 MHz, DMSO-d6) d
1,4,5,6-Tetrahydro-cyclopenta[c]pyrazole-3-carbohydrazide (11b): Start-
ing with 10b, the same procedures were followed as in the synthesis
12.87 (s, 1H), 11.23 (s, 1H), 9.32 (s, 1H), 9.24 (s, 1H), 8.29 (s, 1H),
7.19 (s, 1H), 6.86 (d, J = 8.4 Hz, 1H), 6.76 (d, J = 8.4 Hz, 1H), 2.66
(t, J = 5.6 Hz, 2H), 2.62 (t, J = 6.0 Hz, 2H), 2.35 (s, 3H), 1.75–1.68
(m, 4H); EI-MS m ⁄ z 300.1 (M+), 149.1 (100%); HRMS (EI) m ⁄ z calcd
C15H16N4O3 (M+) 300.1222; found 300.1224.
1
)
of 11a. Yield = 53%; HNMR (400 MHz, DMSO-d6 d 8.97 (s, 1H),
4.38 (bs, 2H), 2.68–2.61 (m, 4H), 2.42–2.38 (m, 2H).
N'-[(2-Hydroxy-naphthalen-1-yl)methylene]-1,4,5,6-tetrahydro-cyclo-
penta[c]pyrazole-3- carbohydrazone (7a): Starting with 11b and 2-
hydroxy-1-naphthaldehyde, the same procedures were followed as
in the synthesis of 1. Yield = 78%; mp = 287–298 ꢀC; 1HNMR
(400 MHz, DMSO-d6) d 13.08 (s, 1H), 12.94 (s, 1H), 11.96 (s, 1H),
9.64 (s, 1H), 8.13 (d, J = 8.0 Hz, 1H), 7.91 (t, J = 9.2 Hz, 2H), 7.60
(t, J = 7.6 Hz, 1H), 7.41 (t, J = 7.6 Hz, 1H), 7.21 (d, J = 9.2 Hz, 1H),
2.74–2.66 (m, 4H), 2.53–2.49 (m, 2H); EI-MS m ⁄z 320.1 (M+), 135.1 (100%);
HRMS (EI) m ⁄ z calcd C18H16N4O2 (M+) 320.1273; found 320.1271.
N'-cycloheptylidene-4,5,6,7-tetrahydro-1H-indazole-3-carbohydrazone
(6i): Starting with 11a and cycloheptanone, the same procedures
were followed as in the synthesis of 6e. Yield = 83%; mp = 258–
1
264 ꢀC; HNMR (400 MHz, DMSO-d6) d 12.83 (s, 1H), 9.68 (s, 1H),
2.67–2.56 (m, 4H), 2.47–2.45 (m, 4H), 1.73–1.67 (m, 6H), 1.60–1.54
(m, 6H); EI-MS m ⁄ z 274.2 (M+), 149.1 (100%); HRMS (EI) m ⁄ z calcd
C15H22N4O (M+) 274.1794; found 274.1792.
N'-[(3-Chloro-4-hydroxy-5-methoxyphenyl)methylene]-4,5,6,7-tetrahydro-
1H-indazole-3-carbohydrazone (6j): Starting with 11a and 3-
chloro-4-hydroxy-5-methoxy-benzaldehyde, the same procedures
were followed as in the synthesis of 1. Yield = 82%; mp = 313–
Ethyl 2-oxo-2-(2-oxocycloheptyl)acetate (9c): Starting with cyclo-
heptanone and ethyl oxalate, the same procedures were followed
1
as in the synthesis of 9a. Yield = 87%; HNMR (400 MHz, CDCl3) d
15.50 (s, 1H), 4.35 (q, J = 7.2 Hz, 2H), 2.65 (t, J = 6.4 Hz, 2H),
2.51–2.48 (m, 2H), 1.79–1.63 (m, 6H), 1.37 (t, J = 7.2 Hz, 3H).
1
316 ꢀC; HNMR (400 MHz, DMSO-d6) d 12.91 (s, 1H), 11.47 (s, 1H),
9.91 (s, 1H), 8.35 (s, 1H), 7.24 (s, 1H), 7.21 (s, 1H), 3.90 (s, 3H),
2.68 (t, J = 4.8 Hz, 2H), 2.63 (t, J = 5.6 Hz, 2H), 1.75–1.68 (m, 4H);
EI-MS m ⁄ z 348.1 (M+), 149.1 (100%); HRMS (EI) m ⁄ z calcd
C16H17ClN4O3 (M+) 348.0989; found 348.0990.
Ethyl 1,4,5,6,7,8-hexahydro-cyclohepta[c]pyrazole-3-carboxylate (10c):
Starting with 9c, the same procedures were followed as in the
synthesis of 10a. Yield = 88%; HNMR (400 MHz, CDCl3) d 8.37 (s,
1
1H), 4.38 (q, J = 7.2 Hz, 2H), 2.95 (t, J = 5.6 Hz, 2H), 2.82 (t,
J = 5.6 Hz, 2H), 1.90–1.84 (m, 2H), 1.72–1.65 (m, 4H), 1.40 (t,
J = 7.2 Hz, 3H).
N'-[(Naphthalen-1-yl)methylene]-4,5,6,7-tetrahydro-1H-indazole-3-carbo-
hydrazone (6k): Starting with 11a and 1-naphthaldehyde, the same
procedures were followed as in the synthesis of 1. Yield = 90%;
mp = 289–293 ꢀC; 1HNMR (400 MHz, DMSO-d6) d 12.98 (s, 1H),
11.61 (s, 1H), 9.23 (s, 1H), 8.42 (d, J = 8.4 Hz, 1H), 8.00 (d,
J = 8.4 Hz, 2H), 7.91 (d, J = 7.2 Hz, 1H), 7.65 (t, J = 7.2 Hz, 1H),
7.61 (t, J = 7.6 Hz, 2H), 2.73 (t, J = 5.2 Hz, 2H), 2.65 (t, J = 5.6 Hz,
2H), 1.77–1.71 (m, 4H); EI-MS m⁄z 318.1 (M+), 149.1 (100%); HRMS
(EI) m ⁄ z calcd C19H18N4O (M+) 318.1481; found 318.1480.
1,4,5,6,7,8-Hexahydro-cyclohepta[c]pyrazole-3-carbohydrazide (11c):
Starting with 10c, the same procedures were followed as in the
synthesis of 11a. Yield = 57%; 1HNMR (400 MHz, DMSO-d6) d
8.96 (s, 1H), 4.28 (bs, 2H), 2.84 (s, 2H), 2.67 (t, J = 5.6 Hz, 2H),
1.75 (t, J = 5.2 Hz, 2H), 1.59–1.52 (m, 4H).
N'-[(2-Hydroxy-naphthalen-1-yl)methylene]-1,4,5,6,7,8-hexahydro-cyclo-
hepta[c]pyrazole-3- carbohydrazone (7b): Starting with 11c and 2-
hydroxy-1-naphthaldehyde, the same procedures were followed as
in the synthesis of 1. Yield = 86%; mp = 321–327 ꢀC; 1HNMR
(400 MHz, DMSO-d6) d 13.00 (s, 1H), 12.98 (s, 1H), 11.93 (s, 1H),
9.65 (s, 1H), 8.12 (d, J = 8.8 Hz, 1H), 7.91 (t, J = 8.0 Hz, 2H), 7.60
(t, J = 7.2 Hz, 1H), 7.41 (t, J = 7.2 Hz, 1H), 7.22 (d, J = 8.8 Hz, 1H),
2.99 (t, J = 5.2 Hz, 2H), 2.77 (t, J = 5.6 Hz, 2H), 1.85–1.75 (m, 2H),
1.66–1.61(m, 4H); EI-MS m ⁄ z 348.1 (M+), 163.1 (100%); HRMS (EI)
m ⁄ z calcd C20H20N4O2 (M+) 348.1586; found 348.1588.
N'-[(4-Methoxy-naphthalen-1-yl)methylene]-4,5,6,7-tetrahydro-1H-in-
dazole-3-carbohydrazone (6l): Starting with 11a and 4-methoxy-
1-naphthaldehyde, the same procedures were followed as in the
1
synthesis of 1. Yield = 92%; mp = 262–265 ꢀC; HNMR (400 MHz,
DMSO-d6) d 12.94 (s, 1H), 11.44 (s, 1H), 9.07 (s, 1H), 8.93 (d,
J = 8.4 Hz, 1H), 8.25 (d, J = 8.4 Hz, 1H), 7.82 (d, J = 8.4 Hz, 1H),
7.68 (t, J = 7.6 Hz, 1H), 7.59 (t, J = 7.6 Hz, 1H), 7.10 (d, J = 8.4 Hz,
1H), 4.05 (s, 3H), 2.72 (t, J = 5.2 Hz, 2H), 2.65 (t, J = 5.6 Hz, 2H),
1.78–1.71 (m, 4H); EI-MS m ⁄ z 348.1 (M+), 183.1 (100%); HRMS (EI)
m ⁄ z calcd C20H20N4O2 (M+) 348.1586; found 348.1583.
5-(3-Methoxyphenyl)-1H-pyrazole-3-carbohydrazide (11d): Starting
with 3-methoxyacetophenone and ethyl oxalate, the same proce-
dures were followed as in the synthesis of 11a. Yield = 60%;
1HNMR (400 MHz, DMSO-d6) d 13.63 (s, 1H), 9.41 (s, 1H), 7.38–
7.34 (m, 3H), 7.14 (s, 1H), 6.91 (d, J = 7.2 Hz, 1H), 4.50 (bs, 2H),
3.81 (s, 3H).
Ethyl 2-oxo-2-(2-oxocyclopentyl)acetate (9b): Starting with cyclo-
pentanone and ethyl oxalate, the same procedures were followed
1
as in the synthesis of 9a. Yield = 64%; HNMR (400 MHz, CDCl3) d
4.35 (q, J = 7.2 Hz, 2H), 2.96 (t, J = 7.2 Hz, 2H), 2.48 (t, J = 8.0 Hz,
2H), 2.03–1.95 (m, 2H), 1.37 (t, J = 7.2 Hz, 3H).
900
Chem Biol Drug Des 2012; 79: 897–906

Studies of Pyrazole-3-Carbohydrazone Derivatives
N'-[(2-Hydroxy-naphthalen-1-yl)methylene]-5-(3-methoxyphenyl)-1H-pyra-
zole-3-carbohydrazone (7c): Starting with 11d and 2-hydroxy-1-
naphthaldehyde, the same procedures were followed as in the
5-(4-Chlorophenyl)-1H-pyrazole-3-carbohydrazide (11f): Starting with
4-chloroacetophenone and ethyl oxalate, the same procedures were
followed as in the synthesis of 11a. Yield = 100%; 1HNMR
(400 MHz, DMSO-d6) d 7.79 (d, J = 8.4 Hz, 2H), 7.51 (d, J = 8.8 Hz,
2H), 7.15 (s, 1H).
1
synthesis of 1. Yield = 85%; mp = 283–286 ꢀC; HNMR (400 MHz,
DMSO-d6) d 13.88 (s, 1H), 12.93 (s, 1H), 12.15 (s, 1H), 9.69 (s, 1H),
8.18 (d, J = 8.4 Hz, 1H), 7.95 (d, J = 8.8 Hz, 1H), 7.92 (d,
J = 8.4 Hz, 1H), 7.62 (t, J = 7.6 Hz, 1H), 7.47–7.40 (m, 4H), 7.31 (s,
1H), 7.25 (d, J = 9.2 Hz, 1H), 7.00 (d, J = 7.6 Hz, 1H), 3.85 (s, 3H);
MALDI-MS m ⁄ z 387.1 [M + H]+; HRMS (MALDI) m ⁄ z calcd
C22H19N4O3 [M + H]+ 387.1457, found 387.1442.
N'-[(2-Hydroxy-naphthalen-1-yl)methylene]-5-(4-chlorophenyl)-1H-pyra-
zole-3-carbohydrazone (7e): Starting with 11f and 2-hydroxy-1-
naphthaldehyde, the same procedures were followed as in the
1
synthesis of 1. Yield = 48%; mp = 300–308 ꢀC; HNMR (400 MHz,
DMSO-d6) d 13.95 (s, 1H), 12.93 (s, 1H), 12.19 (s, 1H), 9.69 (s, 1H),
8.17 (d, J = 8.4 Hz, 1H), 8.04 (d, J = 8.8 Hz, 1H), 7.94–7.89 (m, 4H),
7.63–7.57 (m, 3H), 7.41 (t, J = 7.6 Hz, 1H), 7.22 (t, J = 8.8 Hz, 1H);
MALDI-MS m ⁄ z 391.1 [M + H]+; HRMS (MALDI) m ⁄ z calcd
C21H16ClN4O2 [M + H]+ 391.0962, found 391.0958.
5-(3-Nitrophenyl)-1H-pyrazole-3-carbohydrazide (11e): Starting with
3-nitroacetophenone and ethyl oxalate, the same procedures were
followed as in the synthesis of 11a. Yield = 59%; 1HNMR (400 MHz,
DMSO-d6) d 13.93 (s, 1H), 9.84 (s, 1H), 8.53 (s, 1H), 8.20 (d,
J = 7.2 Hz, 2H), 7.75 (t, J = 8.0 Hz, 1H), 7.37 (s, 1H), 4.55 (bs, 2H).
N'-[(2-Hydroxy-naphthalen-1-yl)methylene]-5-(3-nitrophenyl)-1H-pyra-
zole-3-carbohydrazone (7d): Starting with 11e and 2-hydroxy-1-
naphthaldehyde, the same procedures were followed as in the
synthesis of 1. Yield = 69%; mp >300 ꢀC; 1HNMR (400 MHz,
DMSO-d6) d 14.20 (s, 1H), 12.90 (s, 1H), 12.23 (s, 1H), 9.69 (s, 1H),
8.78 (s, 1H), 8.36 (d, J = 6.8 Hz, 1H), 8.26 (d, J = 8.4 Hz, 1H), 8.18
(d, J = 7.6 Hz, 1H), 7.95 (d, J = 8.8 Hz, 1H), 7.91 (d, J = 8.4 Hz,
1H), 7.80 (t, J = 8.0 Hz, 1H), 7.62 (t, J = 8.0 Hz, 1H), 7.52 (s,
1H),7.42 (t, J = 7.2Hz, 1H); MALDI-MS m ⁄ z 402.1 [M + H]+; HRMS
(MALDI) m ⁄ z calcd C21H16N5O4 [M + H]+ 402.1202, found 402.1192.
Biological assay
The ability of the compounds 1–7 to inhibit human recombinant DPP-
4 was determined using a DPP-4 Drug Discovery Kit (Catalog No.
BML-AK499, biomol) according to the manufacturer's instructions. A
commercial DPP-4 Drug Discovery Kit (Biomol) was performed in 96-
well flat-bottom plates in 100 lL reaction volume. The recombinant
soluble human DPP-4 in 50 lL assay buffer (50 mM Tris, pH 7.5) was
incubated for 15 min at room temperature with indicated concentra-
tions of the compounds to be tested. Reactions were initiated by the
addition of 50 lL of H-Gly-Pro-AMC (H-Gly-Pro-7-amino-4-methyl-
O
H
N
O
O
O
O
O
NH2
R2
R1
R2
R1
a
c
b
O
R1
N
N
R1
N
N
R2
R2
9a-f
O
H
H
8a
: R1, R2= (CH2)4
8b: R1, R2=
8c: R1, R2=
: R1=3-MeO-phenyl, R2=H
8e: R1=3-NO2-phenyl, R2=H
: R1=4-Cl-phenyl, R2=H
10a-f
11a-f
(CH2)3
(CH2)5
1
:
R3=2-OH-naphthalen-1-yl, R4=H
6a: R3=2-OH-phenyl, R4=H
6b
8d
: R3=4-CN-phenyl, R4=H
6c: R3=4-Cl-3-F-phenyl, R4=H
6d
8f
H
O
R4
R3
N
: R3=4-Cl-3-NO2-phenyl, R4=H
6e: R3=2,4-Di-OH-phenyl, R4=CH3
6f
N
d
11a
N
: R3, R4=
N
H
6g
: R3=1-methyl-1H-pyrrol-2-yl,R4=H
6h: R3=2,4-OH-phenyl, R4=H
6i: R3, R4= (CH2)6
1 and 6a-l
6j
: R3=3-Cl-4-OH-5-MeO-phenyl, R4=H
6k: R3=naphthalen-1-yl, R4=H
6l
H
N
O
OH
: R3=4-MeO-naphthalen-1-yl, R4=H
7a: R1, R2= (CH2)3
(CH2)5
e
N
R2
R1
11b-f
N
7b
: R1, R2=
N
7c: R1=3-MeO-phenyl, R2=H
7d: R1=3-NO2-phenyl, R2=H
H
7e
: R1=4-Cl-phenyl, R2=H
7a-e
Scheme 1: Reagents: (a) Diethyl oxalate, CH3CH2ONa, 0 ꢀC to room temperature, 6h; (b) NH2–NH2ÆH2O, CH3COOH, reflux, 8 h; (c) NH2–
NH2ÆH2O, CH3CH2OH, reflux, 8h; (d) R3COR4, CH3COOH, reflux, 4 h; (e) 2-Hydroxy-1-naphthaldehyde, CH3COOH, reflux, 4h.
Chem Biol Drug Des 2012; 79: 897–906
901

Wu et al.
coumarin, GPAMC) to a final concentration of 5 lM. Isoleucine thiaz-
olidide (P32 ⁄ 98), a competitive transition-state substrate analog
inhibitor of DPP-4, was used as a positive control (14), and DMSO
(0.50% w ⁄ v) was used as a native control (NC). Compounds dilutions
were prepared from stock in DMSO and diluted with assay buffer for
inhibition assay. The increased fluorescence signal of 7-amino-4-
methylcoumarin (AMC) (excitation at 380 nM and emission at 460 nM)
hydrophobic groups at N'-positions of carbohydrazone in compound
1, and twelve analogues (6a–l) were designed (Scheme 1) to
determine whether 2-hydroxy-naphthalene moiety is necessary for
inhibitor ⁄ DPP-4 interaction. Second, replacing cyclohexyl ring in
compound 1 with rings of different sizes (cyclopentyl or cycloheptyl
ring) and open-ring substituents, we synthesized five analogues
(7a–e) (Scheme 1) to estimate whether the binding pocket of the
DPP-4 around cyclohexyl ring has enough spatial volume available
to accommodate large chemical moieties.
was monitored for 20 min with 1-min interval using SynergyTM
2
Multi-Mode Microplate Reader (BioTek, Winooski, Vermont, USA). The
inhibitory rate (IR) was calculated by using the following equation.
Scheme 1 depicts the sequence of reactions that led to the prepa-
ration of compounds 1, 6a–l, and 7a–e using R1, R2-di-substituted
ethanone (8) as the starting material. Treatment of 8 with diethyl
oxalate and sodium ethoxide in anhydrous ethyl alcohol afforded a,
c-diketo ester 9, followed by the twice condensation with hydrazine
hydrate to obtain ethyl pyrazole-3-carboxylate 10 and carbohydraz-
ide 11, respectively. Finally, the condensation of compound 11 with
different aldehydes or ketones giving the target compounds 1 and
6–7 (Table 1).
ꢀ
ꢁ
Scompound
SNC
IRð%Þ ¼ 1 ꢀ
ꢁ 100
ð1Þ
where S represents slope that indicates the rate of reaction.
IC50 values were determined from the results of three independent
experiments and calculated from the inhibition curves using the
GRAPHPAD PRISM software (GraphPad Software Inc. La Jolla, CA, USA).
Results and Discussion
Biological activities and SARs
For the primary assay, using P32 ⁄ 98 (Isoleucine thiazolidide) as a
positive control, the inhibition rates of the compounds 1, 6a–l, and
7a–e at 10 lM were measured. The results are summarized in
Table 1. Of the synthetic derivatives tested, nine compounds (e.g.,
Analogue design and synthesis
Compound 1-directed chemical modifications were performed in
two cycles. First, we incorporated various steric, electronic, and
Table 1: Chemical structures of compounds 1, 6a–l, 7a–e and their activities
Compound
R1
R2
R3
R4
Inhibitory rate at 10 lM (%)
IC50 (lM)
1
–(CH2)4–
–(CH2)4–
–(CH2)4–
–(CH2)4–
–(CH2)4–
–(CH2)4–
–(CH2)4–
2-OH-naphthalen-1-yl
2-OH-phenyl
4-CN-phenyl
4-Cl-3-F-phenyl
4-Cl-3-NO2-phenyl
2,4-Di-OH-phenyl
H
H
H
H
H
CH3
76.7
18.0
12.1
19.6
2.5
2.12
–
–
–
–
6a
6b
6c
6d
6e
6f
24.1
14.5
70.80
6g
6h
6i
–(CH2)4–
–(CH2)4–
–(CH2)4–
1-methyl-1H-pyrrol-2-yl
2,4-Di-OH-phenyl
-(CH2)6-
H
H
31.9
22.0
0
52.48
6j
–(CH2)4–
–(CH2)4–
–(CH2)4–
–(CH2)3–
3-Cl-4-OH-5-MeO-phenyl
Naphthalen-1-yl
H
H
H
H
H
H
H
H
3.4
6k
6l
68.3
52.8
38.2
53.0
47.9
38.3
42.3
92.70
3.44
18.20
43.65
19.50
16.59
35.62
18.61
0.14
4-MeO-naphthalen-1-yl
2-OH-naphthalen-1-yl
2-OH-naphthalen-1-yl
2-OH-naphthalen-1-yl
2-OH-naphthalen-1-yl
2-OH-naphthalen-1-yl
7a
7b
7c
–(CH2)5–
3-methoxyphenyl
3-nitrophenyl
4-chlorophenyl
H
H
H
7d
7e
P32 ⁄ 98
902
Chem Biol Drug Des 2012; 79: 897–906

Studies of Pyrazole-3-Carbohydrazone Derivatives
6e, 6g, 6k–l, and 7a–e) displayed good inhibitory activities
against DPP-4 (expressed as inhibitory rate at 10 lM ‡ 25%), indi-
cating that these compounds were good candidate inhibitors of
DPP-4. The half-maximal inhibitory concentration (IC50) values, rang-
ing from 2.12 to 70.80 lM (Table 1), were calculated by fitting the
dose–response curve using the GRAPHPAD PRISM software with three
independent determinations.
important to the inhibitory activities; (ii) displacement of 2-OH sub-
stitution on naphthalene ring with 4-MeO (1 versus 6l) or removal
of 2-OH substitution on naphthalene ring (1 versus 6k) substantially
decreased the inhibitory activity of the derivatives, which implied
the polar substituent has some contributions to the activity of the
analogs; (iii) introduction of 4-MeO substitution on the naphthalene
ring (6k versus 6l) could not be tolerated, leading to a fivefold
reduction in activity, which proposed that steric groups at 4-position
of naphthalene ring were disadvantaged to the activity; (iv) replace-
ment of the 2-OH-naphthalene ring with the 2,4-di-OH-phenyl (1
versus 6e) or the 1-methyl-pyrrolyl group (1 versus 6g) substan-
tially decreases potency, which confirmed the significance of the
naphthalene ring and the 2-OH substitution on it, respectively, and
also suggested that the hydrophobic interaction was more important
than the polar interaction (6g versus 6e). (v) Among cyclohexane
(1), cyclopentane (7a), cycloheptane (7b), and open-ring (7c–e)
substituents, cyclohexane substituent was optimal, and significantly
improved the biological activities (1 versus 7a, 1 versus 7b and 1
versus 7c–e).
For further study on the DPP-4 selectivity for nine active compounds
(6e, 6g, 6k–l, and 7a–e), we calculated the ligand-receptor bind-
ing free energy for the active compounds with the three models of
DPP-4, DPP-8, and DPP-9, respectively (Table S1, Supporting Infor-
mation). According to the data listed in Table S1, we found that
the values of the binding free energy of nine active compounds
with the model of DPP-4 were all lower than the values of the
binding free energy of them with the models of DPP-8 and DPP-9.
These results indicated tentatively that our active compounds are
DPP-4 selectivity over DPP-8 and DPP-9.
The SAR analysis of a set of eighteen compounds provided impor-
tant insights into the essential structural requirements for effective
DPP-4 inhibition. An analysis of the data shown in Table 1 revealed
some noteworthy observations of the SAR for compounds 1, 6a–l,
and 7a–e: (i) compared with non-naphthalene-containing com-
pounds (6a–d, 6f, 6h, 6i, and 6j), the naphthalene-containing
analogs (1, 6k–l and 7a-e) all showed good activity (ranging from
2.12 to 43.65 lM), this means the naphthalene moiety was crucially
Binding models
To address more information for SARs of the discovered DPP-4
inhibitors and to gain clues for further structural optimization, the
binding models of the lead compound (1) and the two representa-
tive analogs (6k and 7c) to DPP-4 were generated by using GLIDE
5.0.d Figure 2 showed the predicted binding poses of 1 (2A), 6k
A
B
C
Figure 2: Three-dimensional (3D) interaction schemes of docked poses of 1 (A), 6k (B), and 7c (C) in the active sites of DPP-4. The
ligands and critical residues of the binding pocket are shown as sticks, and the non-carbon atoms are colored by atom types (receptor carbon
in gray and ligand carbon in orange). Hydrogen bonds are shown as dotted lines. All hydrogen atoms were not displayed.
Chem Biol Drug Des 2012; 79: 897–906
903

Wu et al.
(2B), and 7c (2C) in the DPP-4 active site. It was obvious that the
Pharmacophore mapping
4,5,6,7-tetrahydro-indazole moiety of compound
1
got tightly
To further elaborate the SARs based on the ligand information,
common feature pharmacophore models were generated by HipHop
present in Catalyst 4.10e based on eight DPP-4 inhibitors (1-5, 6e,
6k, and 7d, Table 2) and then top-ranking pharmacophore models
were exported for further studies. As an internal validation of the
pharmacophore models, the training set compounds were mapped
onto the top four pharmacophore models. Based on the best-fit val-
ues and DE values of the training set compounds, hypo1 was cho-
sen as the best pharmacophore model, which consisted of one
HBD, one HBA, and two hydrophobic (HY) features (Figure 3). Our
model hypo1 partially matched with the best pharmacophore model
generated by HypoGen in 2008 (16), which further verified the accu-
racy of our model. Furthermore, the features of hypo1 were consis-
tent with the 3D-QSAR model of DPP-4 inhibitors reported by our
laboratory (17), which also proved our model was reasonable.
trapped in the hydrophobic cage formed by residues Tyr631, Val656,
Trp659, Tyr662, Tyr666, Val711 (named as S1 pocket), and the a-
amino group of carbohydrazone established the hydrogen bond
interaction with the special residue Glu205 (Figure 2A). In addition,
the 2-hydroxyl substitution on naphthalene ring formed hydrogen
bond interactions with residues Glu206 and Arg669, and the
naphthalene ring of compound 1 made the potent hydrophobic
interaction with the residue Phe357, which was consistent with the
result of dynamics performed by our group (15) (Figure 2A). For
compound 6k, its binding mode (Figure 2B) to DPP-4 was very simi-
lar to that of 1 to DPP-4, such as also forming hydrophobic interac-
tions with the S1 pocket and the residue Phe357. These
hydrophobic interactions play critical roles for maintaining the activ-
ity of analogs. Although compound 6k was lack of 2-OH
substitution on naphthalene ring, which led to the loss of the
hydrogen bond interactions with residues Glu206 and Arg669
observed in the case of 1, and was regarded as one of the key
factors leading to ꢂ2 times potency decrease in compound 6k
compared with compound 1, compound 6k still exhibited the most
potent inhibitory effect than other analogs. Unlike compounds 1
and 6k, compound 7c adopted a different docking conformation
(Figure 2C); the 5-phenyl-pyrazole group deviated from the hydro-
phobic S1 pocket. The S1 pocket was spatially constrained and
could not accommodate larger groups, especially polar groups,
which caused the lower inhibitory activity of 7c–e compared with
1. All the experimental results were in agreement with the molecu-
lar-binding model.
The 3D space and distance constraints of these pharmacophore fea-
tures of hypo1 are shown in Figure 3A,B presented the hypo1
aligned with the most active compound 1. It was revealed that the
HBA matched 2-hydroxyl substitution on naphthalene ring; the
hydrogen bond donor mapped with a-amino group of carbohydraz-
one; the two HY features mapped with cyclohexane ring of 4,5,6,7-
tetrahydro-indazole moiety and benzene ring of naphthalene group,
respectively. This pharmacophore mapping result was highly consis-
tent with the docking result of the crystal structure of DPP-4. For
example, the HBA and HBD features perfectly directed to the resi-
dues Arg669 and Glu205, respectively. One HY feature was close to
the Phe357 and the other HY feature was corresponding to hydro-
Table 2: Structures and inhibitory activities of the training set compounds used to generate qualitative pharmacophore models
Compound
Structure
IC50 (lM)
Compoundd
Structure
IC50 (lM)
H
N
O
1
2.12
5
36.35
OH
N
NH
H
N
N
N
O
H
O
O
O
H
N
H
N
2
3
15.31
17.87
6e
6k
70.80
3.44
N
O
N
N
N
N
NH2
HO
N
H
OH
H
N
O
O
N
N
N
NH
O
N
N
N
H
N
H
N
O
4
35.64
7d
35.62
OH
N
S
N
H2N
F
N
F
F
O
N
N
N
H
H
NO2
904
Chem Biol Drug Des 2012; 79: 897–906

Studies of Pyrazole-3-Carbohydrazone Derivatives
A
B
C
D
Figure 3: The best model Hypo1 (A) with distance constraints (in angstroms). Compound 1 (B), 6k (C), and 7c (D) mapped onto the phar-
macophore model Hypo1. Pharmacophore features are color-coded with blue for HY, green for hydrogen bond acceptor, and purple for hydro-
gen bond donar.
phobic S1 pocket (Figure 2). The mapping results of the other
active compounds were also encouraged and in accord with the
docking results (Figure 3C–D). Compound 6k missed the HBA fea-
ture of hypo1 so that it cannot form hydrogen bond interactions
with the residues Glu206 and Arg669. For compound 7c, one of the
HY features was lost, which made it impossible to form the hydro-
phobic interaction with the S1 pocket. Pharmacophore mapping
results were in good agreement with docking results. Besides, some
guiding information was obtained for further structural optimization:
based on the two key hydrophobic moieties (4,5,6,7-tetrahydro-in-
dazole group and naphthalene ring), we may add some electric frag-
ments to make interaction with surrounding polar residues, such as
Glu205, Arg669, Asn710, His740 (Figure 2).
some guiding information for further structural optimization and are
helpful for future inhibitors design.
Acknowledgments
We gratefully acknowledge the financial supports from the National
Natural Science Foundation of China (Grants 21002028, 20902022
and 30973642), National S&T Major Project, China (Grant
2011ZX09102-005-02), the 111 Project (Grant B07023), the Funda-
mental Research Funds for the Central Universities, the Innovation
Program of Shanghai Municipal Education Commission (Grant
10ZZ41), and the Shanghai Committee of Science and Technology
(Grant 11DZ2260600).
Conclusion
References
In summary, we have discovered novel inhibitors of DPP-4, featuring
the pyrazole-3-carbohydrazone scaffold, using a structure-based vir-
tual screening approach in conjunction with chemical synthesis and
bioassay. Based on the structure of lead compound 1, chemical
modifications were performed in two cycles. Totally, seventeen new
compounds have been synthesized and tested with biological
assays. Finally, nine compounds (6e, 6g, 6k–l, and 7a–e) were
found to show micromolar or low-to-mid-micromolar level inhibitory
effects against DPP-4. Molecular docking models give rational
explanation about SARs. Based on eight DPP-4 inhibitors (1–5, 6e,
6k, and 7d), the best pharmacophore model hypo1 was obtained,
composed of one HBD, one HBA and two hydrophobic (HY) features.
Both docking model and pharmacophore mapping results are in
agreement with pharmacological results. The present studies give
1. Gorrell M.D., Gysbers V., McCaughan G.W. (2001) CD26: a multi-
functional integral membrane and secreted protein of activated
lymphocytes. Scand J Immunol;54:249–264.
2. Meester I.D., Durinx C., Proost P., Scharpe S., Lambier A.M.
(2002) Ectopeptidases: CD13 ⁄ Aminopeptidase N and CD26 ⁄ Dip-
eptidylpeptidase IV in Medicine and Biology. New York: Kluwer
Academic ⁄ Pleneu Publisher; pp. 223–257.
3. Mentlein R., Gallwitz B., Schmidt W.E. (1993) Dipeptidylpepti-
dase IV hydrolyses gastric inhibitory polypeptide, glucagon-like
peptide-1 (7–36) amide, peptide histidine methionine and is
responsible for their degradation in human serum. Eur J Bio-
chem;214:829–835.
Chem Biol Drug Des 2012; 79: 897–906
905

Wu et al.
4. Kieffer T.J., McIntosh C.H., Pederson R.A. (1995) Degradation of
glucose-dependent insulinotropic polypeptide and truncated glu-
cagon-like peptide 1 in vitro and in vivo by dipeptidyl peptidase
IV. Endocrinology;358:5–3596.
5. Kim D., Wang L.P., Beconi M., Eiermann G.J., Fisher M.H., He
H.B., Hickey G.J. et al. (2005) (2R)-4-Oxo-4-[3-(trifluoromethyl)-
5,6-dihydro[1,2,4]triazolo[4,3-a]pyrazin-7(8H)-yl]-1-(2,4,5-trifluor-
ophenyl)butan-2-amine: a potent, orally active dipeptidyl pepti-
dase IV inhibitor for the treatment of type 2 diabetes. J Med
Chem;48:141–151.
13. Kirchner F.K., Bethlehem N.Y. (1961) 1-(3-Indazolecarbonyl)-hydra-
zines and their preparation. Patent No. US3007938.
14. Demuth H.U., McIntosh C.H., Pederson R.A. (2005) Type 2 diabe-
tes-therapy with dipeptidyl peptidase IV inhibitors. Biochim Bio-
phys Acta;1751:33–44.
15. Li C., Shen J., Li W.H., Lu C.H., Liu G.X., Tang Y. (2011) Possible
ligand release pathway of dipeptidyl peptidase IV investigated
by molecular dynamics simulations. Proteins;79:1800–1809.
16. Lu I.L., Tsai K.C., Chiang Y.K., Jiaang W.T., Wu S.H., Mahindroo
N., Chien C.H., Lee S.J., Che X., Chao Y.S., Wu S.Y. (2008) A
three-dimensional pharmacophore model for dipeptidyl peptidase
IV inhibitors. Eur J Med Chem;43:1603–1611.
17. Zeng J., Liu G.X., Tang Y., Jiang H.L. (2007) 3D-QSAR studies on
fluoropyrrolidine amides as dipeptidyl peptidase IV inhibitors by
CoMFA and CoMSIA. J Mol Model;13:993–1000.
6. Villhauer E.B., Brinkman J.A., Naderi G.B., Burkey B.F., Dunning
B.E., Prasad K., Mangold B.L., Russell M.E., Hughes T.E. (2003)
1-[[(3-hydroxy-1-adamantyl)amino] acetyl]-2-cyano-(S)-pyrrolidine:
a potent, selective, and orally bioavailable dipeptidyl peptidase
IV inhibitor with antihyperglycemic properties.
Chem;46:2774–2789.
J
Med
Notes
7. Augeri D.J., Robl J.A., Betebenner D.A., Magnin D.R., Khanna
A., Robertson J.G., Wang A.Y. et al. (2005) Discovery and pre-
clinical profile of saxagliptin (BMS-477118): a highly potent,
long-acting, orally active dipeptidyl peptidase IV inhibitor for the
treatment of type 2 diabetes. J Med Chem;48:5025–5037.
8. Feng J., Zhang Z., Wallace M.B., Stafford J.A., Kaldor S.W.,
Kassel D.B., Navre M., Shi L., Skene R.J., Asakawa T., Takeuchi
K., Xu R., Webb D.R., Gwaltney S.L. (2007) Discovery of aloglip-
tin: a potent, selective, bioavailable, and efficacious inhibitor of
dipeptidyl peptidase IV. J Med Chem;50:2297–2300.
9. Eckhardt M., Langkop E., Mark M., Tadayyon M., Thomas L., Nar
H., Pfrengle W., Guth B., Lotz R., Sieger P., Fuchs H., Himmels-
bach F. (2007) 8-(3-(R)-aminopiperidin-1-yl)-7-but-2-ynyl-3-methyl-
1-(4-methyl-quinazolin-2-ylmethyl)-3,7-dihydropurine-2,6-dione (BI
1356), a highly potent, selective, long-acting, and orally bioavail-
able DPP-4 inhibitor for the treatment of type 2 diabetes. J
Med Chem;50:6450–6453.
ahttp: ⁄ ⁄ www.cdc.gov ⁄ Features ⁄ Livingwithdiabetes ⁄ .
bWorld Health Organization at http://www.who.int/diabetes/facts/
world_figures/en/.
chttp://www.diabetesatlas.org/.
dGLIDE version 5.0. New York, NY: Schrçdinger, LLC; 2009. http://
www.schrodinger.com/.
eLIGPREP version 2.2. New York, NY: Schrçdinger, LLC; 2005. http://
www.schrodinger.com/.
fCATALYST User Guide, Version 4.10 (Software Package). San Diego,
CA, USA: Accelrys Inc. http://accelrys.com/.
Supporting Information
Additional Supporting Information may be found in the online ver-
sion of this article:
10. Biftu T., Scapin G., Singh S., Feng D., Becker J.W., Eiermann G.,
He H. et al. (2007) Rational design of a novel, potent, and orally
bioavailable cyclohexylamine DPP-4 inhibitor by application of
molecular modeling and X-ray crystallography of sitagliptin. Bio-
org Med Chem Lett;17:3384–3387.
11. Snyder H.R., Brooks L.A., Shapiro S.H. (1943) Pimelic acid. Org
Synth;2:531.
12. van Herk T., Brussee J., van den Nieuwendijk A.M.C.H., van der
Klein P.A.M., IJzerman A.P., Stannek C., Burmeister A., Lorenzen
A. (2003) Pyrazole derivatives as partial agonists for the nico-
tinic acid receptor. J Med Chem;46:3945–3951.
Table S1. The predicted binding free energies of the active com-
pounds (1, 6e, 6g, 6k–l, and 7a–e) with the receptors (DPP-4,
DPP-8 and DPP-9) by MM-GBSA method.
Please note: Wiley-Blackwell is not responsible for the content or
functionality of any supporting materials supplied by the authors.
Any queries (other than missing material) should be directed to the
corresponding author for the article.
906
Chem Biol Drug Des 2012; 79: 897–906