Substituted oxines inhibit endothelial cell proliferation and angiogenesis

Article abstract of DOI:10.1039/c2ob06978d

Two substituted oxines, nitroxoline (5) and 5-chloroquinolin-8-yl phenylcarbamate (22), were identified as hits in a high-throughput screen aimed at finding new anti-angiogenic agents. In a previous study, we have elucidated the molecular mechanism of antiproliferative activity of nitroxoline in endothelial cells, which comprises of a dual inhibition of type 2 human methionine aminopeptidase (MetAP2) and sirtuin 1 (SIRT1). Structure-activity relationship study (SAR) of nitroxoline offered many surprises where minor modifications yielded oxine derivatives with increased potency against human umbilical vein endothelial cells (HUVEC), but with entirely different as yet unknown mechanisms. For example, 5-nitrosoquinolin-8-ol (33) inhibited HUVEC growth with sub-micromolar IC₅₀, but did not affect MetAP2 or MetAP1, and it only showed weak inhibition against SIRT1. Other sub-micromolar inhibitors were derivatives of 5-aminoquinolin-8-ol (34) and 8-sulfonamidoquinoline (32). A sulfamate derivative of nitroxoline (48) was found to be more potent than nitroxoline with the retention of activities against MetAP2 and SIRT1. The bioactivity of the second hit, micromolar HUVEC and MetAP2 inhibitor carbamate 22 was improved further with an SAR study culminating in carbamate 24 which is a nanomolar inhibitor of HUVEC and MetAP2. The Royal Society of Chemistry 2012.

Full text of DOI:10.1039/c2ob06978d

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Organic &
Biomolecular
Chemistry
Cite this: Org. Biomol. Chem., 2012, 10, 2979
www.rsc.org/obc
PAPER
Substituted oxines inhibit endothelial cell proliferation and angiogenesis†
Shridhar Bhat,a Joong Sup Shim,a Feiran Zhang,a Curtis Robert Chong‡a and Jun O. Liu*a,b
Received 23rd November 2011, Accepted 8th February 2012
DOI: 10.1039/c2ob06978d
Two substituted oxines, nitroxoline (5) and 5-chloroquinolin-8-yl phenylcarbamate (22), were identified
as hits in a high-throughput screen aimed at finding new anti-angiogenic agents. In a previous study, we
have elucidated the molecular mechanism of antiproliferative activity of nitroxoline in endothelial cells,
which comprises of a dual inhibition of type 2 human methionine aminopeptidase (MetAP2) and sirtuin 1
(SIRT1). Structure–activity relationship study (SAR) of nitroxoline offered many surprises where minor
modifications yielded oxine derivatives with increased potency against human umbilical vein endothelial
cells (HUVEC), but with entirely different as yet unknown mechanisms. For example, 5-nitrosoquinolin-
8-ol (33) inhibited HUVEC growth with sub-micromolar IC50, but did not affect MetAP2 or MetAP1,
and it only showed weak inhibition against SIRT1. Other sub-micromolar inhibitors were derivatives of
5-aminoquinolin-8-ol (34) and 8-sulfonamidoquinoline (32). A sulfamate derivative of nitroxoline (48)
was found to be more potent than nitroxoline with the retention of activities against MetAP2 and SIRT1.
The bioactivity of the second hit, micromolar HUVEC and MetAP2 inhibitor carbamate 22 was improved
further with an SAR study culminating in carbamate 24 which is a nanomolar inhibitor of HUVEC and
MetAP2.
cascade, other drug candidates including endostatin3 and meta-
Introduction
stat4 disrupt both the processes of extracellular matrix breakdown
Angiogenesis has been implicated in tumor growth and metasta-
sis as well as other human disorders such as age-related macular
degeneration and rheumatoid arthritis. Inhibitors of angiogenesis
are starting to emerge as a new type of therapeutic agents against
cancer and age-related macular degeneration. The known inhibi-
tors of angiogenesis currently used in the clinic and at various
stages of preclinical and clinical development target different
aspects of endothelial cell biology. While drugs such as bevaci-
zumab1 and sunitinib2 act by blocking angiogenesis signaling
and endothelial cell migration. Another group of anti-angiogenic
agents in clinical development, including cilengitide,5 ombrabu-
lin,6 squalamine,7 and lodamin,8 directly inhibit the growth of
endothelial cells, which has also proven to be a viable strategy
for inhibiting angiogenesis.
One of the most potent classes of inhibitors of endothelial cell
proliferation is the natural products fumagillin and ovalicin along
with their synthetic derivatives like TNP-470.9 After disappoint-
ing results in phase II clinical trials, the late Folkman and col-
leagues successfully resurrected TNP-470 using a nanomedicinal
approach and the new product named lodamin was demonstrated
to be not only efficacious but also orally bioavailable. TNP-470
is a potent inhibitor of human umbilical vein endothelial cell
(HUVEC) proliferation with an IC50 value below 1 nM. The
principal molecular target of the fumagillin family of angio-
genesis inhibitors was elucidated to be methionine aminopepti-
dase type 2 (MetAP2).10 The constitutively expressed isoform
MetAP1 and the inducible isoform MetAP2 are both riboso-
mally-associated metalloproteases performing the vital function
of cotranslational removal of initiator methionine from the
growing polypeptides.11 Despite a significant overlap of sub-
strates between the two isoforms, a specific set of substrates
appear to exist that are processed solely by MetAP2.12 But why
exactly endothelial cells are so uniquely sensitive to MetAP2
inhibition is still an unanswered question. Small-molecules like
TNP-470 work by inducing transcriptional activation of p53,
aDepartment of Pharmacology and Molecular Sciences, Johns Hopkins
University School of Medicine, 725 North Wolfe Street, Baltimore, MD
21205, USA
bDepartment of Oncology, Johns Hopkins University School of
Medicine, 725 North Wolfe Street, Baltimore, MD 21205, USA.
E-mail: joliu@jhu.edu; Fax: +1 410-955-4620; Tel: +1 410-955-4619
†Electronic supplementary information (ESI) available: Synthesis and
analytical data for the substituted quinolines 12–16, 20, 22, 27, 29,
32–34, 38, 40, 43–45, and 47. Expression, purification, and assay proto-
cols pertaining to human MetAPs. Experimental details for the cell-
based assays: (1) 3H-thymidine incorporation in HUVEC, HeLa, and
Jurkat T cells and (2) Calcein-AM in HUVEC and HFF cell lines.
Immunoblotting with compound treated HUVEC lysates. Representative
dose–response curves of MetAP, MetAP2, HUVEC, and HFF inhibition.
Representative UV–vis spectra of compounds with or without metal
ions. 1H-NMR and HPLC traces of the previously unknown compounds.
See DOI: 10.1039/c2ob06978d
‡Current address: Dana-Farber Cancer Institute, Harvard Medical
School, Boston, MA 02115, USA.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2979–2992 | 2979

View Article Online
which in turn leads to the expression of p21 and the latter
inhibits cyclinE·Cdk2 setting forth a cell cycle arrest in the late
G1 phase.13 Besides the understanding of these final downstream
events, the steps in between MetAP2 dysfunction and p53 acti-
vation are as yet unknown.
primary cell line, human foreskin fibroblast (HFF), to evaluate
any selectivity the compounds possessed. We were able to ident-
ify new oxine derivatives that exhibited improved inhibitory
activity against endothelial cell proliferation. Furthermore, we
made the surprising finding that subtle structural alterations can
lead to dramatic changes in the mechanism of action of the
resulting derivatives.
From a drug development perspective, one unwelcome feature
of fumagillin analogs is their irreversible mode of inhibition,
where the reactive spirocyclic epoxide covalently modifies the
His-231 side-chain of MetAP2.14 We have thus actively pursued
the discovery of new reversible inhibitors of MetAP2. Through
modification of fumagillin, we identified an analog named
fumarranol15 which was found to interact with MetAP2 reversi-
bly. In a complementary approach, we performed high-through-
put screen (HTS) of a 175 000-compound library (acquired from
ASDI, Inc., Newark, DE) using MetAP2 as a target and ident-
ified two different chemotypes of oxines, 5-substituted-8-hydro-
xyquinoline (5, Chart 1) and 8-quinolinyl carbamate (22,
Chart 1). Coincidentally, nitroxoline (5) was also a hit when the
Johns Hopkins Drug Library (JHDL), comprising of drugs in
clinical use, was screened for inhibition of HUVEC. Nitroxoline
is prescribed to treat acute microbial urinary tract infections16 in
several countries including the European Union. We recently
published our findings on the unique mode of anti-angiogenic
activity of nitroxoline and demonstrated its potential for appli-
cation in cancer chemotherapy using mouse xenograft models of
human breast and bladder cancers.17 We showed that unlike
Results and discussion
Library screen and identification of substituted oxines as hits
A target-based HTS of a library of 175 000 compounds provided
by ASDI, Inc. was performed to find novel inhibitors of MetAP2
using a coupled enzyme assay with Met-Pro-pNA as the sub-
strate.25 Among 294 hits that exhibited at least 50% inhibition of
MetAP2 activity at a final concentration of 10 μM, nitroxoline
(5, Chart 1), 5-chloroquinolin-8-yl phenylcarbamate (22,
Chart 1), and 5-chloro-8-allyloxyquinoline (27, Chart 1) stood
out for their potency of inhibition (>90%). We validated the hits
by measuring their IC50 values for MetAP2 and found them to
be 55 nM and 5.26 μM for quinolines 5 and 22 respectively. The
IC50 value for allyloxyquinoline 27 was found to be higher than
50 μM. In a HUVEC proliferation assay, the IC50 values for
nitroxoline, 22 and 27 are 1.9 μM, 1.2 μM and >50 μM, respect-
ively. Since, allyloxyquinoline 27 could not be validated in both
the enzyme and cellular assays; it was not pursued further in
ensuing SAR studies. Divalent manganese has been demon-
strated to be the physiologically relevant metal cofactor of
human MetAP2;26 however, this issue has not been settled in the
case of human MetAP1. Cobalt(II) is by far the best metal ion
activator of the human MetAP1 catalytic reaction in vitro, and
hence we and others have been carrying out MetAP1 assays
using Co2+ as the cofactor.27 Thus, we used 10 μM Mn2+ and
Co2+ in the MetAP2 and MetAP1 assays, respectively. In the
cases where an oxine derivative was found to be active against
both the isoforms (e.g., 4, 18, and 34), a likelihood of the coup-
ling enzyme inhibition in the MetAP assay was ruled out by per-
forming an additional control experiment (see ESI†).
fumagillin and TNP-470, nitroxoline inhibits MetAP2 (IC50
=
55 nM) and sirtuin 1 (SIRT1, IC50 = 20.2 μM) in a synergistic
fashion to induce premature senescence in endothelial
cells. Sirtuins are nicotinamide adenine dinucleotide (NAD+)-
dependent deacetylases and they play a crucial role in many cel-
lular processes including gene silencing and DNA repair.18
Among the seven human sirtuins (SIRT1–7), it has been shown
that SIRT1 plays a role in cellular senescence and angiogenesis19
and thus, SIRT1 inhibition leads to the suppression of
angiogenesis.20
Oxines are historically known for their extraordinary metal
chelating abilities. Notwithstanding a continued use in analytical
applications and the most notable utilization in the manufactur-
ing of organic light-emitting diodes (OLEDs),21 substituted
oxines such as cloxyquin22 and clioquinol23 besides nitroxoline
have also been in use as antimicrobial agents. The metal chela-
tion property is thought to be the underlying cause for the bioac-
tivity of cloxyquin and clioquinol, and it has been shown that
nitroxoline shares that feature to manifest its antibacterial
activity.24 Herein, we report the structure–activity relationship
(SAR) studies of the two oxines we identified as hits in the HTS
by obtaining and evaluating different analogs through either
purchasing from commercial source or synthesizing in-house.
We characterized each analog in HUVEC proliferation assay
using [3H]-thymidine incorporation, and enzymatic assays with
MetAP1 and MetAP2. In addition, we also determined the
effects of a subset of oxine derivatives on MetAP and SIRT1
activities in endothelial cells by assessing the processing of N-
terminal methionine of 14-3-3γ, a common substrate of both
MetAP1 and MetAP2; and by measuring acetylation status of
lysine-382 on p53, a known SIRT1 substrate. A selection of the
potent inhibitors of HUVEC proliferation were screened against
two other malignant cell lines, HeLa and Jurkat T, and a human
Synthesis of oxine derivatives
Substituted 8-hydroxyquinolines. Commercially available 8-
hydroxyquinolines 1 through 11 were acquired and used as such
except nitroxoline (5), which was of technical grade and hence
was recrystallized before use (see ESI†). The 7-aminomethyl-8-
hydroxyquinolines 12–19 were synthesized using Mannich con-
densation between an 8-hydroxyquinoline and an aldehyde by
following the procedure of Chi et al.28 with some modifications
(Chart 1, eqn 1).
Quinolin-8-yl carbamates. The carbamates 20–25 were pre-
pared by a simple addition reaction of the commercially available
isocyanates with 8-hydroxyquinolines (Chart 1, eqn 2).
8-Substituted quinolines. Quinolin-8-yl phosphate 26 was
made by reacting cloxyquin (4) with diethyl chlorophosphate
using a combination of triethylamine and DMAP as the base.
The quinolines 27, 28, and 30 were synthesized by carrying out
2980 | Org. Biomol. Chem., 2012, 10, 2979–2992
This journal is © The Royal Society of Chemistry 2012

View Article Online
Chart 1 Synthesis of oxine derivatives.
Modifications to the nitro group of nitroxoline. Nitrosoquino-
linol 33 was prepared according to the reported procedure by
treating 8-hydroxyquinoline 1 with nitrous acid.29 Aminoquino-
lines 34 and 45 were prepared by subjecting nitroquinolines 5
and 44, respectively, to catalytic hydrogenation.30
allylation of quinolin-8-ol (3, 4 or 9) under phase-transfer con-
ditions (Chart 1, eqn 3). The phenoxide anion generated from
quinolinol 4 was reacted with ethylene carbonate at 120 °C to
force a nucleophilic ring opening to provide quinoline 29 in
78% yield (Chart 1, eqn 4).
Initially we attempted monoalkylation of aminoquinoline 34
using 1.2 equivalents of an aldehyde as shown in Chart 1, eqn 5;
but we obtained a mixture of mono and dialkylated products.
Determined to press on with these examples, we added
Commercially available 8-aminoquinoline (31) was converted
to 8-quinolinyl p-toluenesulfonamide (32) by treating the amine
(31) with p-toluenesulfonyl chloride in dichloromethane using
triethylamine as the base.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2979–2992 | 2981

View Article Online
additional equivalents of aldehyde and drove the reaction to
completion forming the dialkylated products. The dialkylamino-
quinolines 35–37 were obtained by sodium triacetoxyborohy-
dride mediated reductive alkylation of the dihydrochloride salt
of aminoquinoline 34 with an appropriate aldehyde (Chart 1,
eqn 5).31
The dihydrochloride salt of 5-aminoquinolin-8-ol (34) was
dissolved in pyridine and treated with different acid chlorides to
yield oxines 38–42 (Chart 1, eqn 6).32 Following a literature pro-
cedure33 quinaldine 2 and cloxyquin (4) were nitrated using a
mixture of concentrated HNO3–H2SO4 to afford quinolinols 43
and 44, respectively.
clioquinol to exhibit anticancer activity by several different
mechanisms.35 That both cloxyquin and clioquinol have been
used in the clinic raised the possibility that their paths to the
clinic may be shortened through expedient clinical evaluation.
The MetAP isoform-selectivity exhibited by cloxyquin, nitroxo-
line, and clioquinol (Table 1) is notable. Cloxyquin, nitroxoline,
and clioquinol are known to chelate metal ions including Mn2+
and Co2+ ions with varying avidities but they only inhibited
MetAP2 with Mn2+ as the cofactor (Table 1), and none of them
inhibited either metalloform of MetAP1 (data not shown, IC50s:
12.9 μM for 4 with Mn2+; and >50 μM for 4 with Co2+, 5 and 7
with Mn2+ or Co2+). In fact nitroxoline also inhibited the cobalt
(II) form of MetAP2 (data not shown, IC50 = 1.48 μM). Even
though oxines 8–13 and 17 are equipped with the same LX
type36 chelating motif like 4, 5, and 7, they did not inhibit either
of the metalloforms of MetAP enzymes.
The anomalies pointed out here cannot be adequately
explained by a mere metal chelation model. While the HUVEC
inhibition by cloxyquin (4) and clioquinol (7) can be ascribed to
their activity against MetAP2 as one of the cellular targets, how
swapping the halogens in clioquinol for other halogens (8–11)
leads to the loss of inhibition of MetAP enzymes but not the
HUVEC activity remains to be elucidated.
Taking advantage of the Mannich reaction that is known to be
facile with oxines,37 we prepared several analogs bearing substi-
tutions at C-7 (12 to 19) and measured their IC50 values for
HUVEC, MetAP1, and MetAP2 inhibition. Among the Mannich
adducts tested, only 12 bearing a proton at C-5 turned out to be
inactive; however, the remaining adducts carrying a 5-chloro
substituent did show inhibition of HUVEC. Except 13 and 17
with polar diethanolamino and bulky phenylpiperazino groups
respectively, all the other Mannich adducts (14, 15, 16, 18, and
19) inhibited MetAP2, where the adduct 18 was the most potent
with an IC50 of 330 nM. It is noteworthy that none of the
hydroxyquinolines (1–32) showed any significant inhibition of
MetAP1, exhibiting isoform specificity. In the case of these
Mannich adducts, MetAP2 inhibition can at least partly account
for their antiproliferative activity in HUVEC. As a side note, the
oxines 4, 14, and 15 were also found to reactivate latent HIV-1
without inducing global T-cell activation, and thus they might
serve as leads in the development of a therapeutic strategy for
the eradication of AIDS.38
8-Hydroxy derivatized nitroxolines. To probe the role of the
8-hydroxy group in nitroxoline, we synthesized ester and ether
derivatives. Due to the presence of the strongly electron-with-
drawing nitro group in the para position, we could only make
pivalate ester 46, phosphate 47, and sulfamate 48, all of which
were found to be stable. The esters were prepared by treating
nitroxoline (5) with the corresponding acid chloride and DMAP/
Et3N or DIEA (in the case of 48) in dichloromethane (Chart 1,
eqn 7).
The 8-alkoxy derivatives were synthesized by reacting
nitroxoline with either ethyl bromoacetate and potassium carbon-
ate in the case of 49, or iodoacetamide and caesium carbonate in
the case of 50; where, in both the cases the reaction mixture in
DMSO was held at 60 °C for 14 h (Chart 1, eqn 7).
8-Aminoquinoline derivatives. Aminoquinolines 51–53 were
obtained from NCI/DTP repository. Nitrosoquinoline 54 was
prepared from sulfonamide 32, in the same manner as the syn-
thesis of nitrosoquinoline 33. 5-Nitroquinolin-8-yl sulfonamide
55 was obtained by oxidizing the corresponding nitrosoquinoline
54 using concentrated nitric acid in glacial acetic acid.
Oxine derivatives: MetAP and HUVEC assays
In our initial campaign on the SAR of nitroxoline, we decided to
replace the electron-withdrawing nitro group with other electron-
donating or electron-withdrawing groups and measure their
activities against HUVEC proliferation and the two isoforms of
MetAP (see Table 1).
Except for the cyanoquinoline 3 and quinoline-5-sulfonic acid
6, oxines 1 to 11 inhibited HUVEC proliferation with IC50
values in the single-digit micromolar range. Among those that
inhibited HUVEC proliferation, only the known antimicrobials
cloxyquin (4), nitroxoline (5), and clioquinol (7) affected
MetAP2 activity in vitro. The lack of inhibition of either
HUVEC or MetAPs by oxines 3 and 6 may be attributed to the
poor coordination ability of the pyridine nitrogen. The strongly
electron-withdrawing cyano group would decrease the electron
density on the pyridine nitrogen in quinolinol 3 and the strongly
acidic sulfonic acid in quinolinol 6, under our assay conditions,
would lock-up that nitrogen in the form of a zwitterion. In
contrast with quinolinol 3, quinolinol 2 carrying an electron-
donating methyl group, did inhibit HUVEC with an IC50 of
2.7 μM.
Quinolinyl carbamate 22 was the other validated hit in our
HTS and a limited SAR was performed with this structure as
well. All the carbamates (20–25) turned out to be active against
HUVEC proliferation. Three of the six carbamates showed little
inhibitory activity towards either of the MetAPs (IC50 >15 μM).
Two of them are derived from 5,7-dichloroquinolin-8-ol (21 and
25) and the third non-MetAP inhibitor is the 3,5-difluorophenyl-
carbamate 23. Carbamate 24 is the most potent MetAP2 inhibitor
(IC50 = 24 nM) among all the oxines we have tested in this
study, and the enzyme activity is also reflected in its potency
against HUVEC (IC50 = 324 nM).
In order to interrogate the role of the 8-hydroxy group in
the oxine inhibitors, we prepared quinoline derivatives 26–30,
and 32. Only the derivatives 28, 29, and 32 inhibited HUVEC
and all were inactive in the MetAP2 assay. Sulfonamide 32
was actually quite potent at annihilating HUVEC (308 nM),
however, it did not achieve that through inhibition of
Clioquinol (7) has been administered to Alzheimer’s disease
(AD) patients in clinical trials34 and studies have also shown
2982 | Org. Biomol. Chem., 2012, 10, 2979–2992
This journal is © The Royal Society of Chemistry 2012

View Article Online
Table 1 Oxine derivatives: MetAP and HUVEC assay results
IC50,a μM
Entry
Substituents (when not specified, R# = H)
hMetAP1
hMetAP2
HUVEC
1
2
3
4
5
6
7
8
—
R2 = Me
R2 = CN
>15
>15
>15
12.9 1.0
>15
>15
>15
>15
>15
2.03 0.3
>15
>15
1.27 0.6
0.055 0.02
>15
2.39 0.31
>15
>15
>15
>15
>15
6.21 1.1
2.70 0.61
>50
1.41 0.1
1.90 0.3
>50
2.42 0.5
2.21 0.6
3.64 0.53
2.53 0.71
2.62 0.08
>20
R5 = Cl (Cloxyquin)
R5 = NO2 (Nitroxoline)
R5 = SO3H
R5 = Cl, R7 = I (Clioquinol)
R5 = Cl, R7 = Br
R5, R7 = Cl
9
10
11
12
R5, R7 = Br
>15
>15
>50
R5, R7 = I
R7 = CH2N(CH2CH2)2O
13
14
15
16
17
18
19
R9 = N(CH2CH2OH)2
>50
>50
>50
>50
>15
14.01 0.8
>50
>15
10.73 0.2
13.51 1.4
5.57 0.16
9.56 0.8
1.46 0.33
1.92 0.6
2.53 0.13
R9 = N(CH2CH2)2CH2
7.68 1.2
2.66 0.2
3.81 0.74
>15
0.33 0.05
2.75 0.4
R9 = N(CH2CH2)2O
R9 = N(CH2CH2)2N–Me
R9 = N(CH2CH2)2NPh(4-OMe)
R9 = N(CH2CH2)2O, R10 = Ph
R9 = NH-2-thiazolyl, R10 = 2-furanyl
20
21
22
23
24
25
R9 = CH2CH2Cl
>15
>15
>15
>15
>15
>15
0.365 0.04
>15
5.26 0.6 (b = 24%)
>15
0.024 0.007
>15
2.11 0.25
2.62 0.5
1.21 0.51
2.20 0.1
0.324 0.06
3.58 0.9
R7 = Cl, R9 = CH2CH2Cl
R9 = Ph
R9 = –(3,5-difluoro)Ph
R9 = –(3,5-dichloro)Ph
R7 = Cl, R9 = –(4-fluoro)Ph
26
27
28
29
30
31
32
R5 = Cl, R8 = OPO(OEt)2
R5 = Cl, R8 = OCH2CHvCH2
R5, R7 = Cl, R8 = OCH2CHvCH2
R5 = Cl, R8 = OCH2CH2OH
R2 = CN, R8 = OCH2CHvCH2
R8 = NH2
>15
>15
>15
>15
>15
>15
>15
>50
>15
>15
>15
>15
>15
>50
>50
2.32 0.23
2.40 0.3
>50
>50
0.308 0.04
R8 = NHSO2-(4-Me)Ph
6.97 0.25 (b = 21%)
a IC50 values are the average of at least three independent experiments consisting of triplicates and standard deviation is given. b = the bottom value
in the dose–response curve, noted in parentheses only when it was above 15%.
either MetAP1 or MetAP2 and the real target needs to be
elucidated.
On the other hand, acylated compounds (38–42) derived from
aminoquinoline 34, produced mixed results in HUVEC prolifer-
ation experiments. Only benzamide 40 showed an inhibition of
HUVEC below 1 μM. Unlike nitroxoline, the derivatives 33–45
either lost their ability to inhibit the MetAP isoforms completely
or became weak inhibitors. A second nitro group on the nucleus
(43) or scrambling of the position of the nitro group (44) resulted
in a loss of potency against HUVEC. Despite the amino group
being in a scrambled position in relation to aminoquinoline 34,
quinolinol 45 inhibited HUVEC proliferation with very high
potency but in a manner that is independent of either of the
MetAP isoforms as targets.
An often invoked role of the 8-hydroxy group in nitroxoline is
its participation in chelation besides a well documented intra-
molecular hydrogen bonding with the pyridine nitrogen. We
decided to fathom the importance of the 8-hydroxy group in
nitroxoline using ester derivatives 46–48. All the three esters
(46–48) retained the MetAP2 inhibitory activity of nitroxoline,
with sulfamate ester 48 displaying improved potency against
Nitroxoline analogs: MetAP and HUVEC assays
In the second SAR campaign (see Table 2) at probing the
nitroxoline structure (5), we first made the nitroso derivative 33.
When assayed for the ability to inhibit HUVEC proliferation, the
nitrosoquinoline (33) was about five times more potent, and the
potency improved with complete reduction of the nitro group
(34). The potency for HUVEC inhibition was retained even after
dialkylation (quinolines 35–37) of the amino group on 34.
Alkylated amines 35–37 showed weak inhibition of MetAP1 and
MetAP2, but the trend has been towards MetAP1 selectivity. In
the case of 37, the selectivity totally switched over to MetAP1.
With long-chain alkyl substitution, 36 did not affect either of the
MetAP isoforms and hence it must be acting on a different
target altogether to inhibit HUVEC proliferation.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2979–2992 | 2983

View Article Online
Table 2 Nitroxoline analogs: MetAP and HUVEC assays
IC50,a μM
Entry
Substituents (when not specified, R# = H)
hMetAP1
hMetAP2
HUVEC
5
R5 = NO2 (Nitroxoline)
>15
>15
0.055 0.02
1.90 0.3
33
34
35
36
37
38
39
40
41
42
43
44
45
R5 = NO
>50
0.364 0.2
0.218 0.04
0.329 0.17
0.789 0.21
0.551 0.08
10.2 1.1 (b = 28%)
8.45 0.91
0.789 0.11
12.94 0.13
1.88 0.3
R5 = NH2
8.92 1.2
7.21 1.3 (b = 29%)
>40
1.46 0.3 (b = 32%)
>40
11.88b 1.1 (b = 26%)
R5 = N(CH2CH3)2
R5 = N([CH2]5CH3)2
R5 = N(CH2(4-OMe)Ph)2
R5 = NHCOCH3
20.13b 2.2 (b = 25%)
>40
>40
>40
>40
>40
>40
>40
>15
>15
>15
R5 = NHCO(CH2)4CH3
R5 = NHCOPh
>40
6.45b 1.1
>40
R5 = NHSO2CH(Me)2
R5 = NHCON(CH2CH2)2O
R2 = CH3, R5, R7 = NO2
R5 = Cl, R7 = NO2
R5 = Cl, R7 = NH2
>40
17.31 1.7
>15
15.74 1.3
7.73 0.77
0.273 0.07
>40
46
47
48
49
50
R9 = COC(Me)3
R9 = PO(OPh)2
R9 = SO2N(Me)2
R9 = CH2CO2Et
R9 = CH2CONH2
>15
>15
>15
>15
>15
0.434 0.1
0.117b 0.07
0.057 0.015
>15
1.6 0.05
2.01 0.06
0.77 0.1
7.3 0.81
6.3 1.4
1.057 0.5
51
52
53
54
55
R5 = Cl
>40
>40
>40
>15
>15
>15
>15
>15
0.721 0.3
1.66 0.06
>20
0.204 0.02
0.305 0.07
R7 = Cl
R5 = Cl, R6 = OMe
R5 = NO, R9 = SO2(4-Me)Ph
R5 = NO2, R9 = SO2(4-Me)Ph
10.51 0.9 (b = 18%)
21.05b 1.8 (b = 19%)
a IC50 values are the average of at least three independent experiments with each consisting of triplicates and standard deviation is given. b The curve
was non-sigmoidal. b = the bottom value in the dose-response curve, given in parentheses only when it was above 15%.
Effect of substituted oxines on MetAP and SIRT1 activities in
HUVEC
HUVEC. The alkylation of nitroxoline (49, 50) resulted in a sub-
stantial loss of potency against HUVEC. Except the aminoquino-
line 53, all the quinolines 51–55 inhibited HUVEC proliferation.
Interestingly, they did not affect the activity of MetAP2.
As we have demonstrated earlier, nitroxoline inhibits HUVEC
by a concurrent inhibition of two targets: MetAP2 and sirtuins
(SIRT1 and SIRT2). Furthermore, inhibition of either of these
two enzymes alone is sufficient to arrest the growth of HUVEC.
To evaluate the activity of a selected subset of substituted oxines
in a cellular context, we treated HUVEC with different oxines at
the same dose (5 μM) at which nitroxoline elicited strong signals
on the immunoblot. And in the case of nitroquinoline 48, which
is more potent than nitroxoline, the cells were treated with a
range of concentrations (0.5–10 μM) of the compound. After a
24 h treatment, the levels of Met-14-3-3γ (the marker for MetAP
inhibition), 14-3-3γ, Ac-p53 (the marker for suppression of
SIRT1 function), and total p53 in the cells were probed and visu-
alized using immunoblotting (Fig. 1).
When compared to nitroxoline, nitrosoquinoline 33 at the
same concentration (5 μM) had very little effect on the acety-
lation status of p53, and neither did it exert any effect on MetAP
activity in cells (Fig. 1). The latter result is in agreement with the
enzyme assay where we had seen insignificant inhibition of
MetAP1 or MetAP2 by 33 at concentrations up to 15 or 50 μM,
respectively. Hence, we surmise this nitroso derivative is distinct
Assessment of cell type selectivity of oxine analogs
A subset of compounds representing different chemotypes of
oxines with high potency for inhibition of HUVEC proliferation
were further examined in proliferation assays using two malig-
nant cell lines, HeLa and Jurkat T cells. We deemed it necessary
to assess these compounds for selectivity against other human
primary cell lines and we chose HFF for this purpose. Since
thymidine uptake in HFF is known to be very feeble, we per-
formed cell-viability assays on HUVEC and HFF side-by-side,
using calcein-AM (see ESI†). Of the derivatives tested, only the
aminoquinoline 34 showed a good selectivity profile; it was
nearly 30 times more selective for HUVEC inhibition than HFF,
and about 8 times more potent against HUVEC than HeLa and
Jurkat T cell lines (Table 3). The lack of selectivity in other cases
may be indicative of a target or process that is very fundamental
to the four cell types that we examined.
2984 | Org. Biomol. Chem., 2012, 10, 2979–2992
This journal is © The Royal Society of Chemistry 2012

View Article Online
Table 3 Evaluation of a subset of oxine derivatives: selectivity against other cell lines and in vivo (HUVEC) activity on MetAPs and SIRT1
IC50,a μM; 3H-T Uptake
Entry HUVECc HeLa
IC50,a μM; Calcein-AM
IC50,a μM; hMetAPs
Immunoblotb
Jurkat-T
HUVEC
HFF
hMetAP1c hMetAP2c Ac-p53 Met-14-3-3γ
17
18
24
32
1.46
1.92
0.324
0.308
ND
ND
ND
ND
ND
ND
ND
3.55 0.9
0.746 0.19
ND
>15
14
>15
>15
+
−
3.66 0.7 (b = 18%)
2.69 0.5 (b = 17%)
0.909 0.12 (b = 21%) 6.97
0.33
0.024
>15
ND
ND
−
ND
ND
−
1.12 0.2 0.79 0.08 1.47 0.2
5
1.90
3.97 0.9 1.58 0.3
0.64 0.2 0.36 0.07 3.72 0.8
2.36 0.5 1.28 0.3
5.13 1.2
3.69 1.0 (b = 21%)
9.08 1.2 (b = 17%)
>15
>15
0.055
>50
++
+
−
+++
+
33
34
40
0.364
0.218
0.789
−
−
+
1.84d 0.15 (b = 19%) 37.3 2.1 (b = 38%)
8.92
11.88d
>40
ND
ND
ND
ND
6.45d
45
46
47
48
0.273
1.6
2.01
0.77
ND
ND
ND
ND
ND
ND
ND
ND
ND
4.21 1.4
ND
ND
ND
>40
>15
>15
>15
>15
−
++
++
+++
−
+
+
+
0.434
0.117
0.057
2.41 0.8 1.73 0.2
2.26 0.3 (b = 23%)
51
54
55
0.721
0.204
0.305
ND ND
0.59 0.1 0.45 0.05 1.32 0.2
1.03 0.2 0.82 0.15 1.46 0.21
3.51 1.1
4.77 1.0 (b = 21%)
>40
>15
>15
>15
−
−
−
−
+
−
0.782 0.13 (b = 22%) 10.51
0.783 0.11 (b = 21%) 21.05d
a Note: IC50 values are the average of at least three independent experiments, each one consisting of triplicates, and standard deviation is given.
b Immunoblotting of compound treated HUVEC lysates (see Fig. 1): The inducers of p53 acetylation are grouped into four categories (−, +, ++, and
+++) based on the immunoblot signals of negative control (DMSO, 1.0) and the positive control (nitroxoline at 5 μM, 11.6), thus −, <2.0; +, 2.0–7.0;
++, 7.1–11.6; +++, >11.6. In the case of Met-14-3-3γ the compounds are grouped into two categories (− and +) based on the immunoblot signals
relative to the control (DMSO, 1.0), hence, −, <1.5; and +, ≥1.5. c see Tables 1 or 2 for standard deviation and other notes. ND: Not Determined.
d Dose-response curve was non-sigmoidal; b = bottom value of the curve.
Fig. 1 Effect of substituted oxines on SIRT1 and MetAP activities in HUVEC. A representative immunoblot is shown and the relative pixel densities
(Ac-p53 band versus p53 and Met-14-3-3γ versus 14-3-3γ band in every lane) were calculated using ImageJ software. The signals are being compared
to the control (relative density of DMSO = 1.0).
from nitroxoline in its mechanism of action, interacting with as
yet unidentified target(s) in HUVEC. The esters of nitroxoline
46–48 all had retained high inhibitory potency against MetAP2
in vitro, but are more potent HUVEC inhibitors than nitroxoline,
which was reaffirmed by the immunoblot analyses. With sulfa-
mate ester 48, we observed a dose-dependent effect on the level
of Ac-p53. It is clear from the immunoblot that the signal for
Ac-p53 in cells treated with 2 μM sulfamate 48 is comparable to
that obtained from cells treated with 5 μM nitroxoline and
perhaps this also explains why sulfamate 48 is superior to
nitroxoline at inhibiting HUVEC proliferation. Similarly to
nitroxoline, a biphasic mode of sirtuin inhibition was also seen
with sulfamate 48 with the maximal inhibition peaking at around
2 μM (for nitroxoline it does at 5 μM). Mannich adduct 17 at
10 μM did increase the level of Ac-p53 when compared to the
DMSO treated cells. Although 17 is a weak inhibitor of SIRT1,
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2979–2992 | 2985

View Article Online
it is interesting because it has a novel structure with elaborate
appendage on the 7th position of quinolinol and thus it is amen-
able to further structural refinements. Benzamidoquinolinol 40
appears to be a good inhibitor of SIRT1 and MetAP in the cell.
Curiously, in the enzyme assays it did not inhibit MetAP2 (IC50
>40 μM), but did inhibit MetAP1 moderately (6.45 μM). Benza-
mide 40 inhibited HUVEC proliferation with an IC50 of 789 nM,
and this compound deserves further study. Unlike benzamide 40,
the nitrosoquinoline 54 whose IC50 for HUVEC proliferation
was 204 nM, did not much alter the levels of Ac-p53 or Met-14-
3-3γ, suggesting that its inhibition of endothelial cell prolifer-
ation is independent of SIRT1 and MetAP enzymes. Up to
10 μM of the substituted oxines 32, 34, 45, 51, and 55 had no
effect on the levels of Ac-p53 or Met-14-3-3γ (data not shown),
and the latter is also in accordance with the MetAP enzyme inhi-
bition data. Due to the cytotoxicity of carbamate 24, not enough
cells were left even after treatment with only 0.5 μM of the com-
pound for 24 h (this concentration is a log unit lower than the
concentration we set out to use to compare against cellular
activity of nitroxoline), and thus we did not attempt further
experiments with carbamate 24.
Perhaps this provides a clue as to why nitroxoline can inhibit
MetAP2 and nitroso derivative 33 cannot. On the other hand,
Co2+ did cause both a bathochromic and a hyperchromic shift in
the UV–vis spectrum of 33. However, the ability of oxine 33 to
chelate Co2+ has not translated into inhibition of either isozyme
of MetAP. Benzamide 40 on the other hand exhibited signs of
complex formation (hypsochromic shifts) with both Mn2+ and
Co2+, but it only inhibited MetAP1 moderately. Addition of
Mn2+ or Co2+ chloride to nitroso derivative 54 gave rise to huge
hypsochromic shifts suggesting the formation of chelates, but
this has not made sulfonamides 54 or 55 into strong inhibitors
of MetAPs. The only conclusion we can draw from all these
results is that metal chelation or lack thereof is not a dependable
parameter to predict or adequately explain MetAP or cellular
activities of the oxine derivatives.
Implications of the SAR
There are ongoing clinical trials on clioquinol (7) for relapsed or
refractory hematological malignancies, eczema, and Parkinson’s
disease. Our finding that clioquinol is anti-angiogenic through
inhibition of MetAP2 could prompt a close scrutiny of clioqui-
nol’s anti-angiogenic effects.
Metal complexation studies
Mannich adducts (12–19) in general inhibited HUVEC pro-
liferation through specifically affecting MetAP2 activity.
Mannich adduct 18 was the most potent inhibitor while main-
taining high potency against MetAP2, and hence this should
serve as a starting point for further SAR. An outlier among the
Mannich adducts, 17, did not inhibit MetAPs and weakly inhib-
ited SIRT1. Further studies are needed to reveal the molecular
mechanism of 17 in HUVEC. The second hit in our HTS, carba-
mate 22 was improved upon by SAR to yield highly cytotoxic
24. Given the number of chlorine atoms on the structure and its
indiscriminate cytotoxicity, carbamate 24 does not appear to be
an attractive candidate for further development.
In the second SAR campaign at probing the nitroxoline struc-
ture (5), a minor modification from nitro to nitroso (33) resulted
in five-fold enhancement in the potency against HUVEC pro-
liferation, but MetAP2 inhibitory activity was abolished. As we
have demonstrated earlier, nitroxoline inhibits HUVEC by con-
current inhibition of two targets: MetAP2 and sirtuins (SIRT1
and SIRT2). Furthermore, inhibition of either of these two
enzymes alone is sufficient to arrest the growth of HUVEC.
Nitroso compound 33 showed a weak inhibition of SIRT1 and
no inhibition of MetAP in vivo (HUVEC). Thus, nitro to nitroso
functional group conversion (5 → 33); which is, as such a small
structural change, led to a complete change in the mechanism of
action.
In order to gain some insight into the metal complexation prop-
erties, we recorded UV–vis spectra of selected oxine derivatives
(5, 24, 33, 40, 48, and 54) at 10 μM in MetAP assay buffer in
the absence and presence of Co2+ and Mn2+ ions (see ESI†). The
choice of metal ion concentration and the buffer was very delib-
erate on our part to at least emulate the enzyme assay conditions.
Carbamate 24 exhibited one absorption maximum at 393 nm,
and this peak neither shifted nor a new peak emerged upon incu-
bation with Mn2+ or Co2+ ions; suggesting a lack of complexa-
tion. This result is in full agreement with an earlier report where
a carbamate similar to 24 used as a prochelator (targeting acetyl-
choline esterase and monoamine oxidase; a project on AD
therapy) was demonstrated to have very little affinity for Fe2+,
Cu2+, or Zn2+ ions.39 Carbamate 24 happens to be the strongest
inhibitor of MetAP2 (24 nM) in this SAR study and it is obvious
now that metal ion sequestration by chelation is not the modus
operandi at least in this case.
Nitroxoline is known to form a chelate with Mn2+ (21 μM 5 +
10 mM MnCl2 was used in this study; bathochromic and hypso-
chromic shifts were seen, λmax = 448 nm shifted to 430 nm),24
and we also observed a measurable shift even at 10 μM (both 5
and MnCl2). However, under the same conditions CoCl2 caused
a huge hypochromicity coupled with a bathochromic shift
(450 nm to 435 nm). Interestingly, nitroxoline inhibited both
Mn2+- and Co2+-metalloforms of MetAP2 (0.055 and 1.48 μM,
respectively), but inhibited neither metalloforms of MetAP1.
Sulfamate 48 also registered similar shifts as nitroxoline did,
except that 48 exhibited two absorption maxima (λmax = 319,
448 nm) and both of them were changed in the presence of
Mn2+ and Co2+. Although, we cannot judge the stabilities of
these complexes with sulfamate 48, we can confirm that it does
form chelates with Mn2+ and Co2+. Interestingly nitrosoquinoline
33, showed very little affinity for Mn2+, because the spectral
traces of 33 with or without Mn2+ completely overlapped.
In an earlier study, Strobl et al. showed that nitrosoquinoline
33 inhibits the proliferation of MCF-7 human breast cancer cell
line through induction of differentiation and apoptosis;40 impor-
tantly, the cellular activity of 33 was attributed to the formation
of reactive oxygen species (ROS). However, much higher con-
centrations of 33 (10 μM) were required to observe ROS
response in comparison to its IC50 for HUVEC inhibition
(364 nM). We also performed dihydrodichlorofluorescein diace-
tate (DCF-DA) assay in HUVEC treated with 10 μM 33, but
failed to observe any significant increase in ROS levels (data not
2986 | Org. Biomol. Chem., 2012, 10, 2979–2992
This journal is © The Royal Society of Chemistry 2012

View Article Online
Fig. 2 Nitroxoline, summary of SAR.
shown). We surmise that nitrosoquinoline 33 acts through a
different mechanism to inhibit HUVEC proliferation. When we
continued along the reduction path all the way to aminoquinoline
34 (i.e. 5 → 33 → 34), it became even more potent against
HUVEC proliferation. Although the aminoquinoline 34 regained
weak activity against MetAPs in vitro, it did not affect MetAP
activity in HUVEC or SIRT1 (data not shown). The fact that the
HUVEC inhibitory activity was not significantly sacrificed by
the long-chain alkyl substitution on amine 34 offers a clue for
the design of molecular probes to facilitate target identification
pertaining to this series of compounds. Among the acylated
analogs of aminoquinoline 34, most were inactive towards
MetAPs or HUVEC proliferation, except the benzamide 40 and
urea 42. The most interesting compound in this set is 40 which
inhibited SIRT1 slightly better than nitroxoline and also caused
N-terminal methionine retention on 14-3-3γ due to the inhibition
of MetAP1. Despite a modest in vitro MetAP1 inhibition, benza-
mide 40 showed a two-fold higher potency against HUVEC pro-
liferation than nitroxoline. Since, MetAP1 and SIRT1 activities
do not appear to be synergistic (Zhang, F. unpublished results:
MetAP1 knockdown in HUVEC in conjunction with a SIRT1
inhibitor, EX-527 treatment did not increase EX-527’s IC50),
benzamide 40 is likely to be acting on a separate target besides
SIRT1.
Based on structural intuition, we were expecting that derivati-
zation of 8-hydroxy group of nitroxoline to its ester would lead
to inactive analogs due to altered metal ion biding properties. On
the contrary, all the three nitroxoline esters (46–48) retained the
activity against HUVEC and MetAP2. Sulfamate 48 was about
twice as potent as nitroxoline at HUVEC inhibition and its IC50
for MetAP2 inhibition was nearly identical to that of nitroxoline.
The UV–vis studies also revealed that sulfamate 48 can chelate
both Mn2+ and Co2+. Sulfamate 48 was very effective in block-
ing SIRT1 and MetAP2 activities in endothelial cells, and in fact
48 was much more potent at inhibiting SIRT1. Fortuitously, this
SAR establishes that the only readily amenable functional group
in nitroxoline can be utilized for derivatization without being
detrimental to nitroxoline’s bioactivity.
Another structural variation where the 8-hydroxy group in the
nitroxoline structure was replaced with an amino group (54, 55)
yielded a different type of endothelial cell inhibitors. Quinoline
55 potently inhibited HUVEC proliferation without affecting
MetAPs or SIRT1. Further SAR and target identification studies
are warranted to demystify the mechanism of action of these
8-aminoquinoline derivatives. To recapitulate the SAR of
nitroxoline, we have uncovered three mechanistically divergent
chemotypes: (1) 5-aminoquinolinols (e.g. 34 and 35); (2) 5-
amidoquinolinols (e.g. 40); and (3) 5-nitro-8-amidoquinolines
(e.g. 55).
Conclusions
In this SAR study of nitroxoline, we found that incremental
functional group alterations can lead to dramatic changes in the
mechanism of action of the analogs. As summarized in Fig. 2,
conversion from 5 to 48 led to preservation and bolstering of the
existing enzymatic and cellular activities of nitroxoline. Simply
removing a single oxygen atom from the structure, that is trans-
formation of 5 to 33, completely abolished the MetAP2 inhibi-
tory activity and decreased activity towards SIRT1 as well.
Reduction of the nitro group to an amino group (34) gave rise to
a mechanistically new inhibitor of endothelial cells. Amazingly,
further acylation of the amino group in 34 (40), restored most of
the SIRT1 and MetAP inhibitory activity. If the 8-hydroxy group
of 5 was transformed into sulfonamido derivative 55, all the
bioactivities against SIRT1 and MetAPs vanished, and yet the
antiproliferative activity against HUVEC increased. Interestingly,
a number of analogs were uncovered with enhanced potency
against HUVEC proliferation without affecting MetAP2 and
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2979–2992 | 2987

View Article Online
SIRT1, suggesting that a rather fragment sized (in a fragment-
based drug discovery approach, a typical fragment has an MW
≤250; nitroxoline, MW = 190) nitroxoline-based pharmacophore
transmutes into a putative ligand for other distinct targets upon
structural modifications. This SAR phenomenon may be attribu-
table to the small sizes of nitroxoline and the analogs. Nitroxo-
line contains only 14 heavy atoms while the analogs except
sulfonamide 55 possess no more than 20 atoms. As previously
suggested by Reynolds et al.,41 this is the linear range for the
‘maximal affinity’ whereby each heavy atom can make a high
impact. Thus, even minor manipulations of the functional groups
could have a huge impact on the overall ligand surface character-
istics and target specificity. In addition to identifying more
potent but mechanistically distinct inhibitors of endothelial cell
proliferation as novel leads for the development of anti-angio-
genic agents, the many surprises we encountered in the medic-
inal chemistry of nitroxoline serves as a reminder that while
dealing with the SAR of small bioactive molecules, minor
changes in structure may lead to major changes in bioactivities
and mechanisms of action.
methanol and 2,5-dihydroxybenzoic acid solution in acetone,
where the latter served as the matrix. Data are reported in the
form m/z (molecular ion).
When necessary, products were purified using a JASCO
HPLC fitted with a semi-preparative VyDAC C18 column
(7 micron, 10 × 250 mm). A mixture of water, acetonitrile, and
2-propanol was used as a mobile phase while eluting at a flow
rate of 3 mL min−1. Samples were analyzed for purity on a
JASCO HPLC equipped with a Rainin Microsorb-MV C18
column (5 μm, 4.6 × 250 mm). The mobile phase set at a flow
rate of 0.8 mL min−1 was programmed for a gradient elution
starting from a 4 : 4 : 2 mixture of H2O–MeCN–MeOH to a final
less polar ratio of 0 : 9 : 1 over six minutes, maintained at that
ratio for three more minutes, and in the next 0.2 min the ratio
was reverted back to 4 : 4 : 2 and the column was equilibrated
over 0.8 min, making each run a total of 10 min long. Purity of
the final compounds was determined to be >95%, using a 10 μL
injection (approximately 1 mM in acetonitrile, most of the time
from a DMSO stock solution) with quantitation by area under
the curve (AUC) at 256 and 277 nm (JASCO Diode Array
Detector). The retention time (tR) is reported in minutes with the
AUC given in parentheses.
While each one of the three chemotypes exemplified by 34,
40 and 55 will be the subject of future studies, a much improved
version of nitroxoline, sulfamate 48 is a prime candidate for
in vivo angiogenesis and xenograft studies in mice.
General procedure for the synthesis of 7-(2-aminomethyl)-8-
hydroxyquinolines
Experimental section
The Mannich bases 12 through 19 were prepared by modifying
a known procedure.26 To a solution of 8-hydroxy-quinoline
(1 mmol) in absolute ethanol (15 mL), triethylamine (1.2 mmol,
170 μL) and paraformaldehyde (1.2 mmol, 36.2 mg) were added
and the reaction mixture was refluxed for 12 h and cooled to
room temperature. The reaction mixture was filtered and the
filtrate was concentrated. Recrystallization of the residue from
1 : 1 EtOH–H2O was the way of purification unless stated
otherwise.
General
Unless stated otherwise, all non-aqueous reactions were carried
out under ambient atmosphere in oven-dried glassware. Indicated
reaction temperatures refer to those of the reaction bath, while
room temperature (rt) is noted as 25 °C. All solvents were of
reagent grade purchased from Fisher Scientific or VWR and
used as received. Commercially available starting materials and
reagents were purchased from Acros, Aldrich, or TCI America
and were used as received, unless stated otherwise.
5-Chloro-7-((4-(4-methoxyphenyl)piperazin-1-yl)methyl)-
quinolin-8-ol (17). Yield: 244 mg (91%). Rf: 0.26
(7 : 3 hexanes–EtOAc); mp: 129 °C; tR: 6.347 min (98%);
1H-NMR (400 MHz, CDCl3): δ 8.88 (dd, J = 3.7, 1.2 Hz, 1H,
H-2), 8.51 (dd, J = 8.3, 1.4 Hz, 1H, H-4), 7.56 (dd, J = 8.3, 3.7
Hz, 1H, H-3), 7.38 (s, 1H, H-6), 6.91 (dd, J = 9.2, 1.6 Hz, 2H,
Ph), 6.85 (dd, J = 9.2, 1.6 Hz, 2H, Ph), 3.92 (s, 2H, CH2-7),
3.77 (s, 3H, OMe), 3.19 (app t, J = 4.8 Hz, 4H, piperazine), 2.82
(app t, J = 4.8 Hz, 4H, piperazine). 13C-NMR (100.6 MHz,
CDCl3): δ 154.32, 149.97, 146.72, 141.87, 139.28, 132.19,
128.23, 126.54, 125.56, 122.07, 121.14, 115.20, 115.51, 56.11,
53. 45, 52.34, 50.03. MALDI-TOF: m/z 384 (M + H)+, 406
(M + Na)+.
Analytical thin layer chromatography (TLC) was performed
using Analtech Uniplates (silica gel HLF, W/UV254, 250 μm).
Visualization was achieved using 254 nm UV light or addition-
ally by staining with iodine, or ceric ammonium molybdate
stain. The retention factor is reported as follows: Rf (eluent
system). Crude products were purified by air-flashed column
chromatography over silica gel (0.06–0.2 mm, 60 Å, from
Acros) using indicated eluents. Melting points were recorded on
a Mel-Temp II apparatus and are uncorrected. NMR data were
collected on a Varian Unity-400 (400 MHz H, 100.6 MHz 13C)
1
or Bruker-Spectrospin 400 MHz spectrometers. Chemical shifts
are reported in ppm (δ) with the solvent resonance or 0.1% tetra-
methylsilane contained in the deuterated solvent as the internal
reference. Data are reported as follows: chemical shift, multi-
plicity (s = singlet, d = doublet, t = triplet, q = quartet, dd =
doublet of doublet, m = multiplet, br = broad, app = apparent,
exch = exchangeable), coupling constants (J, reported in Hertz,
Hz), and number of protons. Low resolution mass spectra were
acquired on a Thermo-Finnigan MAT, LCQ Classic ESI-mass
spectrometer or Voyager DE-STR, MALDI-TOF (Applied
Biosystems) instruments. The MALDI-samples were prepared by
mixing the droplets of the sample solutions in chloroform or
General procedure for the synthesis of 7-(2-amino-2-aryl-
methyl)-8-hydroxy-quinolines
To a solution of 5-chloro-8-hydroxyquinoline (1 mmol, 180 mg)
in absolute ethanol (15 mL), triethylamine (1 mmol, 140 μL)
and aromatic aldehyde (1 mmol) were added and the reaction
mixture was stirred at room temperature for 2 days. The reaction
mixture was concentrated and the residue was recrystallized from
1 : 1 EtOH–H2O.
2988 | Org. Biomol. Chem., 2012, 10, 2979–2992
This journal is © The Royal Society of Chemistry 2012

View Article Online
5-Chloro-7-(morpholino(phenyl)methyl)quinolin-8-ol (18).42
121.98, 112.47, 109.07 & 108.78 (pair of d’s), 103.02 (t).
MALDI-TOF: m/z 335 (M + H)+, 357 (M + Na)+.
Yield: 315 mg (89%). Rf: 0.35 (7 : 3 hexanes–EtOAc). mp:
1
89 °C. tR: 6.080 min (97%). H-NMR (400 MHz, acetone-d6): δ
5-Chloroquinolin-8-yl 3,5-dichlorophenylcarbamate (24).
8.91 (pair of dd’s, J = 3.4, 1.2 Hz, 1H, H-2), 8.53 (pair of dd’s,
J = 6.5, 1.4 Hz, 1H, H-4), 7.71 (pair of dd’s, J = 6.5, 4.4 Hz,
1H, H-3), 7.61 (pair of d’s, J = 8.0 Hz, 2H, 2,6-Ph), 7.27 (pair
of t’s, J = 8.0 Hz, 2H, 3,5-Ph), 7.15 (d, 1H, 4-Ph), 3.71 (br t, J =
4.8 Hz, 2H, 2,6-morpholine), 2,88 (br s, 3H, 2,6-morpholine,
OH), 2.85 (br s, 1H, -N–CH(Ph)-), 2.47 (br s, 4H, 3,5-morpho-
line). 13C-NMR (100 MHz, acetone-d6): δ 153.35, 150.43,
149.97, 149.87, 142.38, 139.99, 133.73, 133.54 (C-4), 129.56 &
128.97 (3,5-Ph), 128.54 & 128.20 (2,6-Ph), 127.26 (4-Ph),
125.92, 125.42, 124.00 (C-3), 123.69 (C-2), 120.86, 111.28,
Yield: 300 mg (82%). Rf: 0.61 (7 : 3 hexanes–EtOAc). mp:
1
159 °C. tR: 8.480 min (96%). H-NMR (400 MHz, acetone-d6):
δ 9.88 (br s, 1H, NH), 8.99 (dd, J = 4.4, 1.6 Hz, 1H, H-2), 8.65
(dd, J = 8.6, 1.4 Hz, 1H, H-4), 7.81 (d, J = 8.2 Hz, 1H, H-7),
7.75 (dd, J = 8.6, 3.8 Hz, 1H, H-3), 7.70 (app. t, J = 2.0 Hz, 2H,
H-2,6-aniline), 7.65 (d, J = 8.2 Hz, 1H, H-6), 7.20 (app. t, J =
2.0 Hz, H-4-aniline). 13C-NMR (100 MHz, acetone-d6): δ
152.13, 143.35, 142.40, 135.91, 133.61, 128.96, 127.88, 127.34,
124.02, 123.46, 122.85, 117.63. MALDI-TOF: m/z 367
(M + H)+, 389 (M + Na)+.
69.09
& 67.55 (2,6-morpholine), 53.51 (3,5-morpholine).
ESI-MS: m/z 355 (M + H)+, 377 (M + Na)+.
5,7-Dichloroquinolin-8-yl
4-fluorophenylcarbamate
(25).
Yield: 312 mg (89%). Rf: 0.68 (7 : 3 hexanes–EtOAc). mp:
148 °C. tR: 5.973 min (96%). H-NMR (400 MHz, CDCl3): δ
5-Chloro-7-(furan-2-yl(thiazole-2-ylamino)methyl)quinolin-8-
ol (19).42 Yield: 258 mg (72%). Rf: 0.32 (7 : 3 hexanes–EtOAc).
mp: 126 °C. tR: 6.147 min (97%). 1H-NMR (400 MHz, CDCl3):
δ 8.85 (dd, J = 4.0, 1.6 Hz, 1H, H-2), 8.51 (dd, J = 8.4, 1.6 Hz,
1H, H-4), 7.69 (s, 1H, H-6), 7.57 (dd, J = 8.4, 4.4 Hz, 1H, H-3),
7.41 (dd, J = 2.8, 0.8 Hz, 1H, furyl-H-5), 7.15 (d, J = 3.6 Hz,
1H, thiazole-H-5), 6.55 (d, J = 3.6 Hz, 1H, thiazole-H-4), 6.40
(s exch., 1H, –OH), 6.34 (dd, J = 3.2, 2.0 Hz, 1H, furyl-H-3),
6.27 (dd, J = 2.8, 0.8 Hz, 1H, furyl-H-4), 1.62 (br s, 1H, –NH–).
13C-NMR (100 MHz, CDCl3): δ 168.93, 152.62, 149.11,
148.72, 142.47, 138.84, 138.60, 133.13, 126.32, 125.85, 122.53,
121.55, 120.61, 110.42, 107.74, 107.11, 52.02. MALDI-TOF:
m/z 358 (M + H)+, 380 (M + Na)+.
1
8.85 (br s, 1H, NH), 7.71 (dd, J = 5.1, 3.3 Hz, 1H, H-2), 7.53
(dd, J = 5.7, 3.5 Hz, 1H, H-4), 7.43 (dd, J = 8.8, 4.8 Hz, 1H,
H-3), 7.35 (s, 1H, H-6), 7.28 (t, J = 8.3 Hz, 2H, H-2,6-aniline),
7.01 (t, J = 8.8, 2H, H-3,5-aniline). 13C-NMR (100 MHz,
CDCl3): δ 161.14 (d), 149.05, 133.93, 131.99, 131.90, 131.16,
129.07, 122.90, 122.82, 118.03, 117.98, 116.12, 115.90.
MALDI-TOF: m/z 351 (M + H)+, 373 (M + Na)+.
Synthesis of 5-chloroquinolin-8-yl diethyl phosphate (26).
Diethyl chlorophosphate (1 mmol, 145 μL) was added to a
mixture of DMAP (1 mmol, 122 mg), Et3N (2.5 mmol, 350 μL),
and 5-chloro-8-hydroxyquinoline (1 mmol, 180 mg) in dichloro-
methane (20 mL) and stirred at rt for 24 h. The reaction mixture
was washed with saturated NaHCO3 (2 × 20 mL) and brine
(1 × 15 mL) and concentrated to dryness. The crude mixture was
purified using flash-column chromatography over silica gel
(eluent: 20% EtOAc in CH2Cl2). Yield: 291 mg (92%). Rf: 0.43
(7 : 3 hexanes–EtOAc). tR: 7.107 min (98%). 1H-NMR
(400 MHz, CDCl3): δ 9.06 (d, J = 1.6 Hz, 1H, H-2), 8.57 (d, J =
4.8 Hz, 1H, H-4), 7.66 (dd, J = 4,8, 1.6 Hz, 1H, H-3), 7.59 (app.
t, J = 5.2 Hz, 2H, H-6,7), 4.38 (q, J = 4.6 Hz, 4H, OEt), 1.41
(t, J = 4.6 Hz, 6H, Et). 13C-NMR (100 MHz, CDCl3): δ 154.3,
153.2, 138.7, 132.5, 128.2, 127.6, 127.4, 122.9, 115.1, 51.4,
12.7. MALDI-TOF: m/z 316 (M + H)+, 338 (M + Na)+.
General procedure for the synthesis of 8-quinolinyl carbamates
Triethylamine (3 drops) was added to a suspension of 8-hydro-
xyquinoline (1 mmol) and an isocyanate (1 mmol) in dichloro-
methane (15 mL). The reaction mixture was stirred for 2 days at
room temperature and the solvent was removed using a rotary
evaporator and the residue was subjected to flash-column chrom-
atography over silica gel (eluent: hexanes or 5% EtOAc in
hexanes) to afford the respective quinolinyl carbamate.
5,7-Dichloroquinolin-8-yl 2-chloroethylcarbamate (21).42
Yield: 233 mg (73%). Rf: 0.65 (7 : 3 hexanes–EtOAc). mp:
112 °C. tR: 4.747 (96%); 1H-NMR (400 MHz, acetone-d6): δ
9.00 (dd, J = 4.0, 1.2 Hz, 1H, H-2), 8.60 (dd, J = 8.4, 1.1 Hz,
1H, H-4), 7.88 (s, 1H, H-6), 7.75 (dd, J = 8.4, 4.0 Hz, 1H, H-3),
7.54 (br s, 1H, NH), 3.76 (t, J = 6 Hz, 2H, H-2′), 3.61 (app. q, J
= 6 Hz, 2H, H-1′). 13C-NMR (100 MHz, acetone-d6): δ 154.29,
152.90, 150.87, 139.78, 133.98, 133.61, 129.00, 127.97, 126.62,
124.08, 44.16, 43.72. MALDI-TOF: m/z 319 (M + H)+, 341
(M + Na)+.
General procedure for the synthesis of 8-allyloxyquinolines
A solution of 8-hydroxyquinoline (1 mmol) in aqueous NaOH
(0.25 M, 8 mL; 2 mmol) containing tetrabutylammonium
bromide as the phase-transfer catalyst (0.1 mmol, 33 mg) was
stirred for 30 min., and allyl bromide (1 mmol, 85 μL) and
benzene (8 mL) were added and the reaction mixture was stirred
vigorously at 50 °C for 14 h. The reaction mixture was parti-
tioned between water (15 mL) and EtOAc (15 mL) after cooling
to room temperature and the extraction was repeated twice. The
organic layers were combined and washed with 1 M NaOH solu-
tion (15 mL) and concentrated. The crude product was purified
by flash-column chromatography over silica gel using
1 : 2 hexanes–ether as eluent to afford the allyloxyquinolines.
5-Chloroquinolin-8-yl 3,5-difluorophenylcarbamate (23).42
Yield: 188 mg (56%). Rf: 0.67 (7 : 3 hexanes–EtOAc). mp:
1
152 °C. tR: 6.333 min (96%). H-NMR (400 MHz, CDCl3): δ
8.98 (dd, J = 4.4, 1.3 Hz, 1H, H-2); 8.56 (dd, J = 5.5, 1.3 Hz,
1H, H-4); 8.32 (br s, 1H, NH); 7.59 (dd, J = 5.5, 4.4 Hz, 1H,
H-3); 7.53 (d, J = 5.8 Hz, 1H, H-7); 7.41 (d, J = 6.3 Hz, 2,6-
aniline); 7.01 (d, J = 5.8 Hz, 1H, H-6); 6.42 (d, J = 6.3 Hz, 1H,
4-aniline). 13C-NMR (100 MHz, CDCl3): δ 165.01 (d), 153.11,
151.23, 150.89, 140.31, 139.46, 132.92, 128.45, 126.27, 122.30,
8-(Allyloxy)-5,7-dichloroquinoline (28). Yield: 218 mg (86%).
Rf: 0.59 (7 : 3 hexanes–EtOAc). mp: 64 °C. tR: 5.427 min
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2979–2992 | 2989

View Article Online
1
5-(N,N-Di[(4-methoxy)benzyl])amino-8-hydroxyquinoline (37).
p-Methoxybenzaldehyde (2.5 mmol, 305 μL) was used, and the
column was eluted with 10% EtOAc–CH2Cl2 to obtain quinoline
37 as a lemon yellow syrup. Yield: 264 mg (91%). tR: 7.080 min
(97%). H-NMR (400 MHz, CDCl3): δ 9.03 (d, J = 1.7 Hz, 1H,
H-2), 8.54 (dd, J = 4.8, 1.7 Hz, 1H, H-4), 7.63 (s, 1H, H-6),
7.53 (dd, J = 4,8, 1.7 Hz, 1H, H-3), 6.22 (m, 1H, CHvCH2),
5.39 and 5.22 (two dd’s, J = 14.3, 1.2 and J = 6.9, 1.2 Hz; 2H,
CHvCH2), 4.98 (dd, J = 6.5, 1.1 Hz, 2H, CH2–CHvCH2).
13C-NMR (100 MHz, CDCl3): δ 155.2, 152.8, 140.7, 133.8,
133.6, 129.1, 127.9, 126.6, 129.1, 114.5, 109.8, 72.5. MALDI-
TOF: m/z 255 (M + H)+, 277 (M + Na)+.
1
(96%). H-NMR (400 MHz, CDCl3): δ 8.81 (d, J = 4.0 Hz, 1H,
H-2), 8.76 (d, J = 8 Hz, 1H, H-4), 7.37 (d, J = 8.0 Hz, 1H, H-3),
7.28 (d, J = 7.0 Hz, 2H, PMB), 7.21 (d, J = 8.0 Hz, 1H, H-6),
6.95 (d, J = 8.0 Hz, 1H, H-7), 6.83 (d, J = 7.0 Hz, 2H, PMB),
4.15 (s, 3H, OCH3), 3.77 (s, 2H, CH2). 13C-NMR (100 MHz,
CDCl3): δ 158.6, 147.7, 144.6, 138.9, 138.7, 132.8, 130.1,
129.2, 121.3, 120.9, 120.3, 114.1, 110.1, 109.1, 55.2, 48.8.
MALDI-TOF: m/z 401 (M + H)+, 422 (M + Na)+.
8-(Allyloxy)quinoline-2-carbonitrile (30). Yield: 172 mg
(82%). Rf: 0.51 (7 : 3 hexanes–EtOAc). tR: 5.707 min (100%).
1H-NMR (400 MHz, CDCl3): δ 8.27 (d, J = 6.8 Hz, 1H, H-4),
7.73 (d, J = 6.8 Hz, 1H, H-3), 7.61 (t, J = 6.9 Hz, 1H, H-6),
7.43 (dd, J = 6.9, 1.2 Hz, 1H, H-5), 7.18 (dd, J = 6.9, 1.2 Hz,
1H, H-7), 6.11 (m, 1H, CHvCH2), 5.56 and 5.41 (pair of dd’s,
J = 15.2, 1.4; 11.0, 1.4 Hz, 2H, CHvCH2), 4.91 (dd, J = 5.5,
1.1 Hz, 2H, CH2CHvCH2). 13C-NMR (100 MHz, CDCl3): δ
156.2, 137.1, 136.9, 133.6, 132.5, 130.2, 129.7, 123.3, 120.3,
116.6, 116.2, 108.9, 71.5. MALDI-TOF: m/z 211 (M + H)+, 233
(M + Na)+.
General procedure for the synthesis of 5-acylamino-8-
hydroxyquinolines
5-Amino-8-hydroxyquinoline dihydrochloride salt was dissolved
in pyridine and treated with carbonyl halides given below
according to the procedure described by Ghedini et al.32 The
crude products were purified by flash-column chromatography
over silica gel.
General procedure for the alkylation of 5-amino-8-
hydroxyquinoline
N-(8-Hydroxyquinolin-5-yl)hexanamide (39). Yield: 56%. tR:
6.067 min (96%). H-NMR (400 MHz, CDCl3) δ: 8.82 (d, J =
1
The reductive amination procedure we used was adapted from
the one described by Röhrig et al.31 5-Amino-8-hydroxyquino-
line dihydrochloride salt (1 mmol, 233 mg) was suspended in
CH2Cl2 (20 mL) and AcOH (0.75 mL) and an aldehyde
(2.5 mmol) were added and stirred vigorously at rt for 15 min.
Sodium triacetoxyborohydride (4 mmol, 850 mg) was added in
one portion and the mixture was stirred overnight at rt. The reac-
tion mixture was diluted with water (30 mL) and extracted with
CH2Cl2 (2 × 25 mL) and the residue left after evaporation of
CH2Cl2 was slurried with silica gel and the desired products
were obtained after flash-column chromatography.
4.0 Hz, 1H, H-2), 8.59 (br s, 1H, OH), 8.31 (d, J = 8 Hz, 1H,
H-4), 7.67 (d, J = 8.0 Hz, 1H, H-6), 7.45 (dd, J = 8.0, 4.0 Hz,
1H, H-3), 7.54 (br s, 1H, NH), 7.37 (d, J = 8.0 Hz, 1H, H-7),
2.74 (t, J = 8 Hz, 2H, COCH2), 1.72 (m, 2H, COCH2CH2), 1.34
(m, 4H), 0.86 (t, J = 8.0 Hz, 3H, CH3). 13C-NMR (100 MHz,
CDCl3) δ: 173.4, 152.4, 143.6, 134.5, 138.2, 130.1, 122.1,
120.1, 116.6, 112.0, 34.7, 28.4, 26.3, 22.1, 17.8. MALDI-TOF:
m/z 259 (M + H)+, 281 (M + Na)+.
N-(8-Hydroxyquinolin-5-yl)propane-2-sulfonamide
(41).
1
Yield: 87%. tR: 4.373 min (97%). H-NMR (400 MHz, CDCl3):
δ 8.84 (dd, J = 8.0, 4.0 Hz, 1H, H-2), 8.05 (dd, J = 8.0, 4.0 Hz,
1H, H-4), 7.43 (d, J = 8.0 Hz, 1H, H-6), 7.25 (dd, J = 8.0, 4.0
Hz, 1H, H-3), 6.57 (d, J = 8.0 Hz, 1H, H-7), 4.61 (br s, 2H, NH,
OH), 3.91 (sept, J = 8 Hz, 1H, iPr), 1.62 (d, J = 8 Hz, 6H, Me).
13C-NMR (100 MHz, CDCl3): δ 150.5, 142.2, 141.7, 136.9,
130.3, 123.9, 120.2, 119.1, 108.2, 53.4, 17.0. MALDI-TOF: m/z
267 (M + H)+.
5-(N,N-Diethyl)amino-8-hydroxyquinoline (35). Acetaldehyde
(2.5 mmol, 150 μL) was used, and the column was eluted with
5% MeOH–CH2Cl2 to obtain quinoline 35 as a dark-green wax.
Yield: 207 mg (96%). tR: 3.867 min (96%). 1H-NMR
(400 MHz, CDCl3): δ 8.80 (d, J = 4.0 Hz, 1H, H-2), 8.65 (d, J =
8 Hz, 1H, H-4), 7.45 (d, J = 8.0 Hz, 1H, H-3), 7.20 (d, J = 8.0,
4.0 Hz, 1H, H-6), 7.15 (d, J = 8.0 Hz, 1H, H-7), 3.10 (q, J = 7.0
Hz, 4H, CH2), 1.01 (t, J = 7.0 Hz, 6H, CH3). 13C-NMR
(100 MHz, CDCl3): δ 157.1, 153.7, 148.2, 133.3, 132.2,
131.1, 121.3, 120.3, 109.2, 44.1, 13.5. MALDI-TOF: m/z 217
(M + H)+, 239 (M + Na)+.
N-(8-Hydroxyquinolin-5-yl)morpholine-4-carboxamide (42).
Yield: 72%. tR: 3.613 min (100%). 1H-NMR (400 MHz,
CDCl3): δ 8.91 (d, J = 4.0 Hz, 1H, H-2), 8.31 (d, J = 8.0 Hz,
1H, H-4), 7.38 (dd, J = 8.0, 4.0 Hz, 1H, H-3), 7.19 (d, J = 8.0
Hz, 1H, H-7), 7.05 (d, J = 8.0 Hz, 1H, H-6), 3.80 (br s, 2H),
3.65 (m, 4H, OCH2), 3.45 (m, 4H, NCH2). 13C-NMR
(100 MHz, CDCl3): δ 158.1, 150.4, 142.9, 138.7, 134.5, 130.3,
123.8, 120.5, 118.1, 113.1, 66.7, 48.6. MALDI-TOF: m/z 274
(M + H)+, 296 (M + Na)+.
5-(N,N-Di-n-hexyl)amino-8-hydroxyquinoline (36). Hexanal
(2.5 mmol, 307 μL) was used, and the column was eluted with
20% EtOAc–CH2Cl2 to obtain quinoline 36 as a yellow-green
syrup. Yield: 310 mg (94%). tR: 8.613 min (96%). 1H-NMR
(400 MHz, CDCl3): δ 8.78 (dd, J = 4.0, 2.0 Hz, 1H, H-2), 8.68
(d, J = 8.0 Hz, 1H, H-4), 7.44 (dd, J = 4.0, 8.0 Hz, 1H, H-3),
7.22 (d, J = 8.0, 4.0 Hz, 1H, H-6), 7.14 (d, J = 8.0 Hz, 1H,
H-7), 3.03 (t, J = 5.5 Hz, 4H, NCH2), 1.25 (m, 16H, CH2), 0.84
(t, J = 5.0 Hz, 6H, CH3). 13C-NMR (100 MHz, CDCl3): δ 148.6,
147.7, 138.8, 133.2, 127.2, 121.1, 120.3, 109.4, 55.6, 31.7,
27.3, 27.1, 22.7, 14.1. MALDI-TOF: m/z 329 (M + H)+, 351
(M + Na)+.
Synthesis of 5-nitroquinolin-8-yl pivalate (46). Pivaloyl chlor-
ide (2 mmol, 250 μL) was added to an ice-cold mixture of
nitroxoline (2 mmol, 380 mg), DMAP (0.8 mmol, 98 mg), and
Et3N (8 mmol, 1.1 mL) in CH2Cl2 (20 mL), and the mixture was
stirred vigorously for 14 h at rt. The reaction mixture was
washed with 10% aqueous NaHCO3 and the CH2Cl2 layer was
concentrated. The crude residue was subjected to flash-column
2990 | Org. Biomol. Chem., 2012, 10, 2979–2992
This journal is © The Royal Society of Chemistry 2012

View Article Online
chromatography (eluent: hexanes–CH2Cl2 = 7 : 3) to afford the
ester 46 as a yellow solid. Yield: 465 mg (85%). Rf: 0.67 (2 : 3
EtOAc–hexanes). tR: 7.613 min (99%). 1H-NMR (400 MHz,
CDCl3): δ 8.99 (d, J = 8.8 Hz, 1H, H-4), 8.93 (d, J = 3.6 Hz,
1H, H-2), 8.38 (d, J = 8.4 Hz, 1H, H-6), 7.59 (dd, J = 8.8, 3.6
Hz, 1H, H-3), 7.46 (d, J = 8.8 Hz, 1H, H-7), 1.49 (s, 9H, t-Bu).
13C-NMR (100 MHz, CDCl3) δ: 178.0, 153.7, 151.4, 142.9,
141.2, 132.4, 125.5, 124.6, 122.8, 120.0, 39.7, 27.6. MALDI-
TOF: m/z 275 (M + H)+, 297 (M + Na)+.
such. Yield: 595 mg (91%). Rf: 0.37 (EtOH–CHCl3 = 1 : 19). tR:
6.840 min (100%). H-NMR (400 MHz, CDCl3) δ: 9.40 (s, 1H,
1
H-2), 8.76 (d, J = 3.0 Hz, 1H, H-4), 8.15 (d, J = 7.6 Hz, 1H,
H-6), 7.87 (s, 1H, H-3), 7.81 (d, J = 8.0 Hz, 2H, Ts), 7.48
(m, 2H, H-7), 7.15 (d, J = 8.0 Hz, 2H, Ts), 2.29 (s, 3H, Me).
13C-NMR (100 MHz, CDCl3) δ: 156.1, 150.3, 149.4, 142.5,
137.2, 136.9, 136.4, 129.5, 127.6, 122.8, 121.3, 120.6, 118.7,
26.6. MALDI-TOF: m/z 327 (M + H)+, 350 (M + Na)+.
Synthesis of 4-methyl-N-(5-nitroquinolin-8-yl)benzene-sulfo-
namide (55). Concentrated nitric acid (0.32 mL) was added
dropwise to a suspension of nitrosoquinoline 51 (1 mmol,
326 mg) in glacial acetic acid (6 mL) at rt. The reaction mixture
was stirred at rt for 2 h and then cooled to 0 °C and basified with
the addition of 10% NaOH to pH 10–11. The resulting brown
precipitate was filtered, resuspended in water, and acetic acid
was added to adjust the pH again to 5–6. The light-brown pre-
cipitate was finally filtered, washed with water and air-dried.
Yield: 264 mg (77%). Rf: 0.39 (EtOH–CHCl3 = 1 : 19). tR:
Synthesis of 5-nitroquinolin-8-yl dimethylsulfamate (48). To
an ice-cold mixture of nitroxoline (1 mmol, 190 mg) and DIEA
(2 mmol, 190 μL) in CH2Cl2 (15 mL), N,N-dimethylamino-sul-
fonyl chloride (1 mmol, 107 μL) was added and the stirring was
continued at rt for 16 h. The reaction mixture was washed with
10% aqueous NaHCO3 (1 × 15 mL) and the CH2Cl2 layer was
concentrated to dryness. The crude residue was loaded onto a
silica gel column and eluted with 1% EtOH in CH2Cl2 to yield
sulfonamide 48 as a yellow solid. Yield: 222 mg (88%). Rf: 0.27
(EtOH–CHCl3 = 1 : 19). tR: 5.600 min (100%). 1H-NMR
(400 MHz, CDCl3): δ 9.10 (d, J = 4.1 Hz, 1H, H-2), 8.45 (d, J =
7.9 Hz, 1H, H-4), 7.90 (d, J = 8.2 Hz, 1H, H-6), 7.75 (dd, J =
7.9, 4.1 Hz, 1H, H-3), 7.20 (d, J = 8.2 Hz, 1H, H-7), 3.15
(s, 6H, NMe2). 13C-NMR (100 MHz, CDCl3): δ 162.3, 152.1,
143.7, 138.4, 131.0, 126.2, 125.4, 122.5, 109.9, 42.8. MALDI-
TOF: m/z 254 (M + H)+, 276 (M + Na)+.
1
6.827 min (100%). H-NMR (400 MHz, CDCl3): δ 9.33 (s, 1H,
H-2), 8.76 (s, 1H, H-4), 8.13 (d, J = 7.2 Hz, 1H, H-6), 7.82
(s, 1H, H-3), 7.80 (d, J = 7.6 Hz, 2H, Ts), 7.46 (s, 1H, H-7),
7.15 (d, J = 7.6 Hz, 2H, Ts), 2.29 (s, 3H, Me). 13C-NMR
(100 MHz, CDCl3): δ 150.3, 149.8, 142.5, 141.1, 137.2, 136.9,
136.4, 129.5, 127.6, 126.9, 124.9, 121.3, 118.7, 26.6. MALDI-
TOF: m/z 344 (M + H)+, 366 (M + Na)+.
Synthesis of ethyl 2-(5-nitroquinolin-8-yloxy)acetate (49).
Nitroxoline (5.3 mmol, 1 g) was heated at 60 °C with ethyl bro-
moacetate (5.3 mmol, 0.6 mL) and potassium carbonate
(7.3 mmol, 1 g) in DMSO (20 mL) for 18 h, cooled to rt, diluted
with water (100 mL) and extracted with a mixture of 2 : 3
EtOAc–Et2O. Solvent was evaporated and the crude product was
purified by column chromatography on silica gel (eluent: 1 : 3
EtOAc–CH2Cl2). Yield: 980 mg (67%). Rf: 0.38 (EtOAc–
hexanes = 2 : 3). tR: 8.893 min (98%). 1H-NMR (400 MHz,
CDCl3): δ 9.15 (d, J = 8.8 Hz, 1H, H-4), 9.01 (d, J = 3.5 Hz,
1H, H-2), 8.42 (d, J = 8.8 Hz, 1H, H-6), 7.67 (dd, J = 8.8, 3.5
Hz, 1H, H-3), 6.93 (d, J = 8.8 Hz, 1H, H-7), 5.03 (s, 2H,
OCH2CO), 4.25 (q, J = 7.5 Hz, 2H, Et), 1.25 (t, J = 7.5 Hz, 3H,
Et). MALDI-TOF: m/z 277 (M + H)+, 299 (M + Na)+.
Acknowledgements
We would like to thank Dr. Jon Lorsch for providing us with
some oxine derivatives he had in turn procured from NCI/DTP
repository. This work was supported by National Cancer Institute
(CA78743).
Notes and references
1 N. Ferrara, K. J. Hillan, H. Gerber and W. Novotny, Discovery and devel-
opment of bevacizumab, an anti-VEGF antibody for treating cancer, Nat.
Rev. Drug Discovery, 2004, 3, 391–400.
2 C. Le Tourneau, E. Raymond and S. Faivre, Sunitinib: a novel tyrosine
kinase inhibitor. A brief review of its therapeutic potential in the treat-
ment of renal carcinoma and gastrointestinal stromal tumors (GIST),
Ther. Clin. Risk Manag., 2007, 3, 341–348.
Synthesis of 2-(5-nitroquinolin-8-yloxy)acetamide (50). Aceta-
mide 50 was also prepared according to the procedure described
above for making the acetate 49, except the reaction was run
with 1 mmol of iodoacetamide and potassium carbonate was
supplanted with caesium carbonate. Yield: 193 mg (78%). Rf:
0.31 (2 : 3 EtOAc–hexanes). tR: 5.802 min (98%). 1H-NMR
(400 MHz, CDCl3): δ 9.40 (d, J = 8.8 Hz, 1H, H-4), 8.81 (d, J =
3.5 Hz, 1H, H-2), 8.55 (d, J = 8.8 Hz, 1H, H-6), 7.68 (dd, J =
8.8, 3.5 Hz, 1H, H-3), 7.51 (d, J = 8.8 Hz, 1H, H-7), 6.80 (br s,
2H, NH2), 4.20 (s, 2H, OCH2CO). MALDI-TOF: m/z 248 (M +
H)+, 270 (M + Na)+.
3 J. Folkman, Antiangiogenesis in cancer therapy–endostatin and its mech-
anisms of action, Exp. Cell Res., 2006, 312, 594–607.
4 C. Gialeli, A. D. Theocharis and N. K. Karamanos, Roles of matrix
metalloproteinases in cancer progression and their pharmacological
targeting, FEBS J., 2011, 278, 16–27.
5 D. A. Reardon, L. B. Nabors, R. Stupp and T. Mikkelsen, Cilengitide: an
integrin-targeting arginine-glycine-aspartic acid peptide with promising
activity for glioblastoma multiforme, Expert Opin. Invest. Drugs, 2008,
17, 1225–1235.
6 A. Delmonte and C. Sessa, AVE8062: a new combretastatin derivative
vascular disrupting agent, Expert Opin. Invest. Drugs, 2009, 18, 1541–
1548.
7 B. Connolly, A. Desai, C. A. Garcia, E. Thomas and M. J. Gast, Squala-
mine lactate for exudative age-related macular degeneration, Ophthalmol.
Clin. North Am., 2006, 19, 381–391.
Synthesis of 4-methyl-N-(5-nitrosoquinolin-8-yl)benzene-sulfo-
namide (54). The sulfonamide derivative of 8-aminoquinoline
(32) was subjected to nitrosation (2 mmol scale) in a manner
similar to the synthesis of nitrosoquinoline 33, and sulfonamide
54 was collected as a beige colored solid after filtration from the
reaction mixture which was spectroscopically pure and used as
8 O. Benny, O. Fainaru, A. Adini, F. Cassiola, L. Bazinet, I. Adini,
E. Pravda, Y. Nahmias, S. Koirala, G. Corfas, R. J. D’Amato and
J. Folkman, An orally delivered small-molecule formulation with antian-
giogenic and anticancer activity, Nat. Biotechnol., 2008, 26, 799–807.
9 (a) B. Lefkove, B. Govindarajan and J. L. Arbiser, Fumagillin: an anti-
infective as a parent molecule for novel angiogenesis inhibitors, Expert
Rev. Anti. Infect. Ther., 2007, 5, 573–579; (b) R. Mazitschek, A. Huwe
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2979–2992 | 2991

View Article Online
and A. Giannis, Synthesis and biological evaluation of novel fumagillin
and ovalicin analogues, Org. Biomol. Chem., 2005, 3, 2150–2154.
10 (a) E. C. Griffith, Z. Su, B. E. Turk, S. Chen, Y. H. Chang, Z. Wu,
K. Biemann and J. O. Liu, Methionine aminopeptidase (type 2) is the
common target for angiogenesis inhibitors AGM-1470 and ovalicin,
Chem. Biol., 1997, 4, 461–471; (b) N. Sin, L. Meng, M. Q. Wang,
J. J. Wen, W. G. Bornmann and C. M. Crews, The anti-angiogenic agent
fumagillin covalently binds and inhibits the methionine aminopeptidase,
MetAP-2, Proc. Natl. Acad. Sci. U. S. A., 1997, 94, 6099–6103.
11 R. A. Bradshaw and E. Yi, Methionine aminopeptidases and angiogen-
esis, Essays Biochem., 2002, 38, 65–78.
12 S. E. Warder, L. A. Tucker, S. M. McLoughlin, T. J. Strelitzer,
J. L. Meuth, Q. Zhang, G. S. Sheppard, P. L. Richardson, R. Lesniewski,
S. K. Davidsen, R. L. Bell and J. C. Rogers, et al. Discovery, identifi-
cation, and characterization of candidate pharmacodynamic markers of
methionine aminopeptidase-2 inhibition, J. Proteome Res., 2008, 7,
4807–4820.
13 Y. Zhang, E. C. Griffith, J. Sage, T. Jacks and J. O. Liu, Cell cycle inhi-
bition by the anti-angiogenic agent TNP-470 is mediated by p53 and
p21WAF1/CIP1, Proc. Natl. Acad. Sci. U. S. A., 2000, 97, 6427–6432.
14 S. Liu, J. Widom, C. W. Kemp, C. M. Crews and J. Clardy, Structure of
human methionine aminopeptidase-2 complexed with fumagillin, Science
(Washington, D. C.), 1998, 282, 1324–1327.
15 J. Lu, C. R. Chong, X. Hu and J. O. Liu, Fumarranol, a rearranged fuma-
gillin analogue that inhibits angiogenesis in vivo, J. Med. Chem., 2006,
49, 5645–5648.
14, 764–774; (c) X. Hu, A. Addlagatta, B. W. Matthews and J. O. Liu,
Identification of pyridinylpyrimidines as inhibitors of human methionine
aminopeptidases, Angew. Chem., Int. Ed., 2006, 45, 3772–3775;
(d) A. Addlagatta, X. Hu, J. O. Liu and B. W. Matthews, Structural basis
for the functional differences between type I and type II human methion-
ine aminopeptidases, Biochemistry, 2005, 44, 14741–14749; (e) X. Hu,
J. Zhu, S. Srivathsan and D. Pei, Peptidyl hydroxamic acids as methion-
ine aminopeptidase inhibitors, Bioorg. Med. Chem. Lett., 2004, 14,
77–79.
28 K. Chi, Y. S. Ahn, K. T. Shim, T. H. Park and J. S. Ahn, Synthesis of
Mannich bases using substituted aromatic alcohols with secondary
amines: relative reactivity and regioselectivity depending on substrates,
J. Korean Chem. Soc., 2001, 45, 51–60.
29 A. A. Isaev, I. I. Lomovskii, K. G. Korolev and R. K. Karimov, Technol-
ogy of preparing 8-hydroxy-5-nitroquinoline, Chem. Heterocycl. Compd.,
2005, 41, 1027–1030.
30 P. R. Sleath, B. J. Noar, G. A. Eberlein and T. C. Bruice, Synthesis of
7,9-didecarboxymethoxatin (4,5-dihydro-4,5-dioxo-1H-pyrrolo[2,3-f]qui-
noline-2-carboxylic acid) and comparison of its chemical properties with
those of methoxatin and analogous o-quinones. Model studies directed
toward the action of PQQ requiring bacterial oxidoreductases and mam-
malian plasma amine oxidase, J. Am. Chem. Soc., 1985, 107, 3328–3338.
31 U. F. Röhrig, L. Awad, A. Grosdidier, P. Larrieu, V. Stroobant, D. Colau,
V. Cerundolo, A. J. G. Simpson, P. Vogel, B. J. Van den Eynde, V. Zoete
and O. Michielin, Rational design of indoleamine 2,3-dioxygenase inhibi-
tors, J. Med. Chem., 2010, 53, 1172–1189.
16 F. M. E. Wagenlehner, A. Pilatz, K. G. Naber, G. Perletti,
C. M. Wagenlehner and W. Weidner, Anti-infective treatment of bacterial
urinary tract infections, Curr. Med. Chem., 2008, 15, 1412–1427.
17 J. S. Shim, Y. Matsui, S. Bhat, B. A. Nacev, J. Xu, H. C. Bhang,
S. Dhara, K. C. Han, C. R. Chong, M. G. Pomper, A. So and J. O. Liu,
Effect of nitroxoline on angiogenesis and growth of human bladder
cancer, J. Natl. Cancer Inst., 2010, 102, 1855–1873.
32 M. Ghedini, M. La Deda, I. Aiello and A. Grisolia, Fine-tuning the lumi-
nescent properties of metal-chelating 8-hydroxyquinolines through amido
substituents in 5-position, Inorg. Chim. Acta, 2004, 357, 33–40.
33 K. Majerz-Maniecka, R. Musiol, A. Skórska-Stania, D. Tabak, P. Mazur,
B. J. Oleksyn and J. Polanski, X-ray and molecular modelling in frag-
ment-based design of three small quinoline scaffolds for HIV integrase
inhibitors, Bioorg. Med. Chem., 2011, 19, 1606–1612.
18 L. Guarente, Diverse and dynamic functions of the Sir silencing complex,
Nat. Genet., 1999, 23, 281–285.
34 Y. Wang, R. Branicky, Z. Stepanyan, M. Carroll, M. Guimond, A. Hihi,
S. Hayes, K. McBride and S. Hekimi, The anti-neurodegeneration drug
clioquinol inhibits the aging-associated protein CLK-1, J. Biol. Chem.,
2009, 284, 314–323.
35 H. Yu, J. R. Lou and W. Ding, Clioquinol independently targets
NF-kappaB and lysosome pathways in human cancer cells, Anticancer
Res., 2010, 30, 2087–2092.
36 M. L. H. Green, A new approach to the formal classification of covalent
compounds of the elements, J. Organomet. Chem., 1995, 500, 127–148.
37 R. T. Bronson, J. S. Bradshaw, P. B. Savage, S. Fuangswasdi, S. C. Lee,
K. E. Krakowiak and R. M. Izatt, Bis-8-hydroxyquinoline-armed diaza-
trithia-15-crown-5 and diazatrithia-16-crown-5 ligands: possible fluoro-
phoric metal ion sensors, J. Org. Chem., 2001, 66, 4752–4758.
38 S. Xing, S. Bhat, N. Shroff, H. Zhang, J. Lopez, J. Margolick, J. O. Liu
and R. F. Siliciano, Novel structurally-related compounds reactivate latent
HIV-1 in a bcl-2 transduced primary CD4+ T cell model without inducing
global T cell activation, J. Antimicrob. Chemother., 2012, 67, 398–403.
39 H. Zheng, M. B. H. Youdim and M. Fridkin, Site-activated chelators tar-
geting acetylcholinesterase and monoamine oxidase for Alzheimer’s
therapy, ACS Chem. Biol., 2010, 5, 603–610.
19 H. Vaziri, S. K. Dessain, E. Ng Eaton, S. I. Imai, R. A. Frye,
T. K. Pandita, L. Guarente and R. A. Weinberg, hSIR2(SIRT1) functions
as an NAD-dependent p53 deacetylase, Cell, 2001, 107, 149–159.
20 (a) C. L. Brooks and W. Gu, How does SIRT1 affect metabolism, senes-
cence and cancer?, Nat. Rev. Cancer, 2009, 9, 123–128; (b) T. Liu,
P. Y. Liu and G. M. Marshall, The critical role of the class III histone dea-
cetylase SIRT1 in cancer, Cancer Res., 2009, 69, 1702–1705.
21 M. Albrecht, M. Feige and O. Osetska, 8-Hydroxyquinolines in metallo-
supramolecular chemistry, Coord. Chem. Rev., 2008, 252, 812–824.
22 P. Hongmanee, K. Rukseree, B. Buabut, B. Somsri and
P. Palittapongarnpim, In vitro activities of cloxyquin (5-chloroquinolin-8-
ol) against mycobacterium tuberculosis, Antimicrob. Agents Chemother.,
2007, 51, 1105–1106.
23 W. Rohde, P. Mikelens, J. Jackson, J. Blackman, J. Whitcher and
W. Levinson, Hydroxyquinolines inhibit ribonucleic acid-dependent
deoxyribonucleic acid polymerase and inactivate Rous sarcoma virus and
herpes simplex virus, Antimicrob. Agents Chemother., 1976, 10, 234–
240.
24 C. Pelletier, P. Prognon and P. Bourlioux, Roles of divalent cations and
pH in mechanism of action of nitroxoline against Escherichia coli strains,
Antimicrob. Agents Chemother., 1995, 39, 707–713.
25 Y. Zhou, X. C. Guo, T. Yi, T. Yoshimoto and D. Pei, Two continuous
spectrophotometric assays for methionine aminopeptidase, Anal.
Biochem., 2000, 280, 159–165.
26 J. Wang, G. S. Sheppard, P. Lou, M. Kawai, C. Park, D. A. Egan,
A. Schneider, J. Bouska, R. Lesniewski and J. Henkin, Physiologically
relevant metal cofactor for methionine aminopeptidase-2 is manganese,
Biochemistry, 2003, 42, 5035–5042.
27 (a) V. M. D’souza, B. Bennett, A. J. Copik and R. C. Holz, Divalent
metal binding properties of the methionyl aminopeptidase from Escheri-
chia coli, Biochemistry, 2000, 39, 3817–3826; (b) X. Hu, Y. Dang,
K. Tenney, P. Crews, C. W. Tsai, K. M. Sixt, P. A. Cole and J. O. Liu,
Regulation of c-Src nonreceptor tyrosine kinase activity by bengamide
A through inhibition of methionine aminopeptidases, Chem. Biol., 2007,
40 (a) A. Martirosyan, S. Leonard, X. Shi, B. Griffith, P. Gannett and
J. Strobl, Actions of a histone deacetylase inhibitor NSC3852 (5-nitroso-
8-quinolinol) link reactive oxygen species to cell differentiation and
apoptosis in MCF-7 human mammary tumor cells, J. Pharmacol. Exp.
Ther., 2006, 317, 546–552; (b) J. S. Strobl, C. W. Seibert, Y. Li,
R. Nagarkatti, S. M. Mitchell, A. C. Rosypal, D. Rathore and
D. S. Lindsay, Inhibition of toxoplasma gondii and plasmodium falci-
parum infections in vitro by NSC3852, a redox active antiproliferative
and tumor cell differentiation agent, J. Parasitol., 2009, 95, 215–223.
41 C. H. Reynolds, S. D. Bembenek and B. A. Tounge, The role of molecu-
lar size in ligand efficiency, Bioorg. Med. Chem. Lett., 2007, 17, 4258–
4261.
42 J. O. Liu, J. S. Shim, C. R. Chong and S. Bhat, Quinoline compounds as
inhibitors of angiogenesis, human methionine aminopeptidase, and
SIRT1, and methods of treating disorders, World Pt., 042163, pp. 112,
2010.
2992 | Org. Biomol. Chem., 2012, 10, 2979–2992
This journal is © The Royal Society of Chemistry 2012