Synthesis of 9,9-dialkyl-4,5-diazafluorene derivatives and their structure-activity relationships toward human carcinoma cell lines

Article abstract of DOI:10.1002/cmdc.201000034

A homologous set of 9,9-dialkyl-4,5-diazafluorene compounds were prepared by alkylation of 4,5-diazafluorene with the appropriate alkyl bromide and under basic conditions. The structures of these simple organic compounds were confirmed by spectroscopic techniques (FTIR, NMR, and FABMS). Their biological effects toward a panel of human carcinoma cells, including Hep3B hepatocellular carcinoma, MDAMB-231 breast carcinoma, and SKHep-1 hepatoma cells, were investigated; a structure-activity correlation was established with respect to the length of the alkyl chain and the fluorene ring structure. The relationship between the mean potency [log(1/IC50)] and alkyl chain length was systematically studied. The results show that compounds with butyl, hexyl, and octyl chains exhibit good growth inhibitory effects toward these three human carcinoma cell lines, and the 9,9-dihexyl-4,5-diazafluorene further exhibits antitumor activity in athymic nude mice Hep3B xenograft models. For the structurally related dialkylfluorenes that lack the diaza functionality, in vitro cytotoxicity was not observed at clinically relevant concentrations.

Full text of DOI:10.1002/cmdc.201000034

DOI: 10.1002/cmdc.201000034
Synthesis of 9,9-Dialkyl-4,5-diazafluorene Derivatives and
Their Structure–Activity Relationships Toward Human
Carcinoma Cell Lines
Qiwei Wang,[a] Marcus Chun-Wah Yuen,[b] Guo-Liang Lu,[a] Cheuk-Lam Ho,[a] Gui-
Jiang Zhou,[a] Oi-Mei Keung,[a] Kim-Hung Lam,[b] Roberto Gambari,[c] Xiao-Ming Tao,[b]
Raymond Siu-Ming Wong,[d] See-Wai Tong,[d] Kit-Wah Chan,[d] Fung-Yi Lau,[d] Filly Cheung,[b]
Gregory Yin-Ming Cheng,*[d] Chung-Hin Chui,*[b, d] and Wai-Yeung Wong*[a]
A homologous set of 9,9-dialkyl-4,5-diazafluorene compounds
were prepared by alkylation of 4,5-diazafluorene with the ap-
propriate alkyl bromide and under basic conditions. The struc-
tures of these simple organic compounds were confirmed by
spectroscopic techniques (FTIR, NMR, and FABMS). Their bio-
logical effects toward a panel of human carcinoma cells, in-
cluding Hep3B hepatocellular carcinoma, MDAMB-231 breast
carcinoma, and SKHep-1 hepatoma cells, were investigated; a
structure–activity correlation was established with respect to
the length of the alkyl chain and the fluorene ring structure.
The relationship between the mean potency [log(1/IC50)] and
alkyl chain length was systematically studied. The results show
that compounds with butyl, hexyl, and octyl chains exhibit
good growth inhibitory effects toward these three human car-
cinoma cell lines, and the 9,9-dihexyl-4,5-diazafluorene further
exhibits antitumor activity in athymic nude mice Hep3B xeno-
graft models. For the structurally related dialkylfluorenes that
lack the diaza functionality, in vitro cytotoxicity was not ob-
served at clinically relevant concentrations.
Introduction
With respect to the classical fluorene-derived segments (9H-flu-
orene I and fluoren-9-one II), the bidentate ligands 4,5-diaza-
fluorene (9H-cyclopenta[2,1-b:3,4-b’]bipyridine) and 4,5-diaza-
fluoren-9-one (III and IV, respectively), in which the CH groups
of the former are replaced by electron-donating nitrogen
atoms, have significant structural differences. In analogy with
the fluorene structure, the angle interior to the five-membered
ring of III, centered at C2, is compressed to ~1088, and the
angle exterior to the two fused rings is expanded to ~1318.[1]
This type of in-plane distortion in the diazafluorene moiety in-
creases the distance between the pyridyl nitrogen lone-pair
electrons and thus influences their cooperativity.[2] Both III and
IV show unique and distinctive properties relative to their
structurally related 2,2’-bipyridine (bipy) and 1,10-phenanthro-
line (phen) analogues due to the fluorene moieties (Figure 1).[3]
The fluorene skeletons can affect the metal chelation proper-
ties of these compounds, and there are clear differences in the
coordinating and bonding properties between III and bipy (or
phen).[4] Scaffold III remains essentially coplanar, and the meth-
ylene bridge in such 3,3’-annelated 2,2’-bipyridines distorts the
bipyridine portion of the molecule; the resulting increase in
the bite angle decreases the metal–nitrogen overlap.[3a,5]
Hence, III is a weaker s-bonding ligand than bipy, and is lower
than bipy in the spectrochemical series which causes a lower-
[a] Q. Wang, Dr. G.-L. Lu, Dr. C.-L. Ho, Dr. G.-J. Zhou, O.-M. Keung,
Prof. W.-Y. Wong
Department of Chemistry and Centre for Advanced Luminescence Materials,
Hong Kong Baptist University, Waterloo Road, Kowloon Tong, Hong Kong
(P.R. China)
Fax: (+852)3411-7348
E-mail: rwywong@hkbu.edu.hk
[b] Prof. M. C.-W. Yuen, Dr. K.-H. Lam, Prof. X.-M. Tao, Dr. F. Cheung,
Dr. C.-H. Chui
Institute of Textiles and Clothing, The Hong Kong Polytechnic University,
Hung Hom, Hong Kong, (P.R. China)
Fax: (+852)2773-1432
E-mail: chchui@graduate.hku.hk
[c] Prof. R. Gambari
BioPharmaNet, Department of Biochemistry and Molecular Biology
The University of Ferrara, Ferrara (Italy)
[d] Dr. R. S.-M. Wong, S.-W. Tong, K.-W. Chan, Dr. F.-Y. Lau, Prof. G. Y.-M. Cheng,
Dr. C.-H. Chui
Haematology Division, Department of Medicine and Therapeutics, Prince of
Wales Hospital, The Chinese University of Hong Kong, Shatin, Hong Kong
(P.R. China)
Figure 1. Structures of various 3,3’-annealed 2,2’-bipyridine ligands and relat-
Fax: (+852)2637-5396
ed compounds.
E-mail: gcheng@cuhk.edu.hk
ChemMedChem 2010, 5, 559 – 566
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
559

W.-Y. Wong, G. Y.-M. Cheng, C.-H. Chui, et al.
MED
ing of any ligand-field states. This could promote the dissocia-
tion of one of the donor nitrogen atoms, and provides a
unique pathway for the formation of labile binuclear com-
plexes that could be catalytically active.[5] For instance, the two
coordinating nitrogen atoms in IV have a larger bite distance
(2.99 ꢁ) relative to phen (2.65 ꢁ), which would result in unusu-
al coordination modes and would give rise to different redox
and biological properties.[6] Specifically, the diazafluorene-type
ligand was preferred over the bipyridyl analogue in some
cases for its planar configuration.
Results and Discussion
Synthesis and characterization
The structures and synthesis protocols of the studied set of
novel diazafluorene-based compounds are shown in Scheme 1.
4,5-Diazafluoren-9-one was prepared in one pot by oxidative
While derivatives of bipy exhibit a rich coordination chemis-
try, their functional properties remain to be explored. Indeed,
their bridged diquaternary derivatives have been studied ex-
tensively, with the potent herbicide diquat as an illustrative ex-
ample.[7] Rebek et al. have also bridged bipy at the 3- and 3’-
positions with ethylenedioxy units to create macrocycles capa-
ble of eliciting an allosteric effect by site-specific metal bind-
ing.[8] In another context, 2,7-bis[2-(N,N-diethylamino)ethoxy]-
fluoren-9-one (tilorone) is known to be a low-molecular-weight
interferon inducer,[9] and in view of its potential clinical applica-
tions, various biological activities of tilorone and its analogues
have been widely studied.[10] It is believed that the in vitro in-
teraction of tilorone with double-stranded DNA could be re-
sponsible for its various biological activities, including interfer-
on induction and independent antiviral activity.[11] Novel classes
of 1,8-diazafluorene and 4,7-phenanthroline derivatives were
also investigated as DNA-complexing agents, and their cellular
activities as well as interactions with nucleic acids in vitro were
examined.[12] Recently, the metallated analogues of diazafluor-
ene ligands[4] have also gathered attention in biochemical and
biomedical applications, and Zaleski and co-workers saliently
reported the use of bis(9-diazo-4,5-diazafluorene)copper(II) ni-
trate as an effective DNA photocleaving agent under anaerobic
conditions using visible light. This complex can serve as a po-
tential model for the action of kinamycin antitumor antibiotics
such as kinamycin C.[13] The study of 4,5-diazafluorene deriva-
tives is thus expected to continue to attract considerable re-
search interest.
Scheme 1. Synthesis of the new 9,9-dialkylated diazafluorene derivatives. Re-
agents and conditions: a) NH2NH2·H2O, 1308C, 24 h; b) Br(CH2)nCH3, NaH,
THF, reflux, 2 h; c) Br(CH2)nCH3, NaOH, (PhCH2)Et3NCl, DMSO, 608C, 24 h.
ring contraction of phen with aqueous potassium permanga-
nate in the presence of potassium hydroxide in water.[3a,14] In
this step, it is important to increase the solution basicity and
to add a dilute solution of potassium permanganate more
slowly in order to favor ring contraction. This enables a more
rapid benzyl–benzylic acid ring contraction of 1,10-phenan-
throline-5,6-quinone to optimize the synthetic yield of 4,5-dia-
zafluoren-9-one. 4,5-Diazafluorene 1 was obtained by reduc-
tion of 4,5-diazafluoren-9-one with neat hydrazine hydrate at
1308C for 24 h.[1,3a,15] Two different methods were attempted
to afford 2–8 in various overall yields via the alkylation of 1
with a suitable alkyl bromide using sodium hydride or sodium
hydroxide as the base.[16] We initially used sodium hydride for
several compounds, but the yields were quite unsatisfactory;
however, the amount obtained was still sufficient for our stud-
ies. The use of sodium hydroxide gave a better yield in the ma-
jority of cases. These air-stable compounds were isolated in
high purity as sticky oils or solids (depending on alkyl chain
length) by column chromatography on silica gel with the
proper eluent. These compounds were fully characterized by
1H and 13C NMR spectroscopy and mass spectrometry (both
fast atom bombardment (FAB) and matrix-assisted laser de-
sorption/ionization time-of-flight (MALDI-TOF) modes). Howev-
er, with the exception of 1, all attempts to get good-quality
single crystals for X-ray diffraction analysis from solid samples
have met with little success so far. We tested their possible
biological effects toward a panel of human carcinoma cells, in-
cluding Hep3B hepatocellular carcinoma, MDAMB-231 breast
carcinoma, and SKHep-1 hepatoma cell lines.
Herein we describe the preparation, chemical characteriza-
tion and biological evaluation of the antitumor activity of com-
pounds 1–8 (Figure 2) comprising a set of 9,9-dialkyl-4,5-diaza-
fluorenes and their parent compound 4,5-diazafluorene. The
growth inhibition potential of these compounds was screened
with respect to alkyl chain length toward three human carcino-
ma cell lines: Hep3B hepatocellular carcinoma (HB-8064),
MDAMB-231 breast carcinoma (HTB-26), and SKHep-1 hepato-
ma (HTB-52). Remarkably, the in vitro cytotoxicity of these
compounds depends on the length of the alkyl substituents.
Figure 2. Structures of the diazafluorene derivatives 1–8.
560
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 559 – 566

9,9-Dialkyl-4,5-diazafluorene Derivatives
Antitumor activities and structure–activity relationships
volved in the mitochondria-dependent programmed cell death
pathway. Accordingly, we determined whether compound 5 is
able to stimulate caspase-9 activation. As shown in Figure 5, in-
cubation of cancer cells with compound 5 at 15 mm for 48 h
significantly induced caspase-9 activity. Moreover, pre-incuba-
tion of the cancer cells with the pan-caspase inhibitor
zVADfmk at 20 mm significantly reversed its activity;
to determine whether caspase inhibition could re-
verse apoptosis induced by compound 5, Hep3B,
MDAMB-231, and SKHep-1 human cancer cells were
pre-incubated with zVADfmk prior to the addition of
compound 5 at 10 mm. After 48 h incubation, cell via-
bility was again determined by the sulforhodamine B
protein staining assay. As shown in Figure 6, pre-incu-
bation of this pan-caspase inhibitor could only parti-
ally reverse the growth inhibitory effect of compound
5 on all the three cancer cell lines. We therefore hy-
pothesize that the caspase-dependent pathway is
only one of the possible ways to mediate the signal
that is stimulated by treatment with compound 5.
As shown in Figure 3, while the screening test was carried out
starting from an initial concentration of 50 mm, we noted an in-
crease in cytotoxic activity of these compounds with increasing
Figure 3. Survival of Hep3B, MDAMB-231, and SKHep-1 human carcinoma cells after
treatment with compounds 1–8 (all at a final concentration of 50 mm) for 48 h shown as
a percentage of survival of untreated control cells. Biological activity was determined by
the sulforhodamine B protein staining assay. Cisplatin (CDDP) was used as a positive ref-
erence compound; DMSO was used at 0.05%. Each determination was made in triplicate;
three independent experiments were performed, and similar results were obtained. Re-
sults represent the mean ÆSD from one representative experiment; *p<0.05 relative to
untreated control.
Athymic nude mice xenografted with human
Hep3B hepatocellular carcinoma cells were employed
to further demonstrate the possible antitumor activi-
ty of compound 5 in vivo.[22] Fourteen mice xeno-
grafted with Hep3B cancer cells exhibiting an aver-
age tumor volume of ~200 mm3 received daily intra-
alkyl chain length from n=1 to n=6; there is an optimum at
n=6, and cytotoxicity then decreases for homologues with
alkyl chains of n=8 or greater. The reason behind this ob-
served increase in potency with increasing alkyl chain length is
not yet known, but it may be due to an increase in lipophilicity
and hence greater intracellular accumulation of the com-
pounds.[17] Further lengthening of the alkyl chains beyond n=
6 may be detrimental to cellular drug uptake through steric
and solubility factors (see below).
Because only compounds 4, 5, and 6 exhibit inhibition of
cell growth to an extent >50% relative to untreated controls,
we also determined their IC50 values.[18] Among them, com-
pound 5 showed the highest cytotoxicity. Morphological inves-
tigation under inverted microscopy after staining with sulfo-
rhodamine B[19] revealed the formation of apoptotic bodies
with both Hep3B and SKHep-1 cancer cells after treatment
with compound 5 (Figure 4). However, under the same condi-
tions, compound 5 showed a lower cytotoxic effect toward pri-
mary culture bone marrow cells obtained from two patients
with a nonmalignant hematological disorder. This suggests
that compound 5 may exhibit greater cytotoxicity toward ma-
lignant cells than nonmalignant bone marrow cells (Table 1).[20]
We also tested the cytotoxicity of the non-nitrogen-containing
model compounds of 4 and 5 (namely, 9,9-dibutylfluorene and
9,9-dihexylfluorene),[16b,21] and remarkably, both fluorene spe-
cies are not cytotoxic, even at 100 mm, suggesting that the
chelating nitrogen atoms play a key role in antitumor activity.
The caspase family plays an essential role in apoptosis,[20]
among which, caspase-9 is an essential family member in-
Figure 4. Morphological changes of Hep3B [a) and c)] and SKHep-1 [b) and
d)] human carcinoma cells after treatment with compound 5, staining with
sulforhodamine B, and comparison with DMSO vehicle control [a) for Hep3B
and b) for SKHep-1]. Compound 5 was added at 12.5 mm for a cell culture
period of 48 h.
peritoneal injections of either buffer carrier (50 mL) or an equal
volume of compound 5 at 10 mgkgÀ1 for a period of 9 days.
As shown in Figure 7, on day 9, mean tumor size of the test
group was roughly half that of the vehicle control group. Rep-
resentative mice are also shown in Figure 8a.
ChemMedChem 2010, 5, 559 – 566
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
561

W.-Y. Wong, G. Y.-M. Cheng, C.-H. Chui, et al.
MED
Table 1. IC50 values of compounds 4, 5, 6, cisplatin (CDDP), and doxorubi-
cin (DOX) toward the three human carcinoma cell lines tested after 48 h
incubation.[a]
Cell line
Hep3B
4
5
6
CDDP
DOX
13.3Æ3.2
6.6Æ1.5
21.8Æ4.3 38.2Æ2.9 3.1Æ1.2
MDAMB-231 18.6Æ2.4 10.3Æ0.8 26.4Æ2.1 79.3Æ5.3 9.7Æ0.3
SKHep-1
BM1
BM2
14.7Æ2.6
ND
ND
7.9Æ1.3
14.7Æ0.6
13.2Æ2.2
23.6Æ0.7 34.7Æ2.6 3.4Æ1.4
ND
ND
ND
ND
ND
ND
[a] Results are based on the sulforhodamine B assay. CDDP and DOX were
used as positive reference compounds. Each determination was per-
formed in triplicate. Three independent experiments were performed,
and similar results were obtained. Data represent the mean ÆSD from
one representative experiment; ND=Not determined; BM=Nonmalig-
nant hematological disorder bone marrow cells.
Figure 7. Representative results showing the changes of tumor volume
(mm3) as a function of time for mice treated with vehicle or compound 5. In-
traperitoneal injection started for both groups when the mean tumor
volume of mice reached ~200 mm3 on day 0. A total of 14 mice were ran-
domly divided into two groups. On day 9, after measuring individual tumor
volume, all mice were sacrificed for plasma collection. Results represent the
mean ÆSD from seven animals; *p<0.05 relative to untreated controls.
of kidney sections were investigated. We found no observable
necrotic features from autopsy samples of all the 14 mice ana-
lyzed (Figure 8b). Furthermore, analysis of plasma liver func-
tional enzymes, including alanine aminotransferase (ALT) and
aspartate aminotransferase (AST) showed normal liver func-
tions in the mice treated with compound 5 (Table 2). H and E
staining of tumor sections showed high cellular integrity,
which suggests that 5 does not induce necrosis in the xeno-
graft under the therapeutic dosage used (Figure 8c).
Figure 5. Caspase-9 activity assay with Hep3B, MDAMB-231, and SKHep-1
human cancer cells to study the apoptosis-inducing potential of compound
5 after 48 h incubation at a concentration of 15 mm; zVADfmk (20 mm) is an
inhibitor of caspase-9 activation. Results are expressed as a percentage of
the control and represent the mean ÆSD from triplicate tests; the data
shown are from a representative experiment taken from three independent
experiments giving similar results.
For comparison, the in vitro antitumor activities of all candi-
dates in the homologous series against Hep3B, MDAMB-231,
and SKHep-1 cell lines were expressed as the concentration of
compounds that exhibit 50% inhibition (IC50). The relationship
between the mean potency [log(1/IC50)] and n (excluding the
ethyl and decyl analogues, which exhibit potencies much
lower than 4.3) is illustrated in Figure 9. The hexyl analogue
appears to be the best candidate for antitumor activity, with
the highest cytotoxicity toward the selected cell lines. These
results suggest a better permeability of this candidate across
the cell membrane, which in turn explains its greater cytotoxic-
ity, consistent with some published reports.[17b,c] However, with
n>6, the compounds may tend to aggregate in the culture
medium, probably owing to stronger dipole–dipole interac-
tions among the side chains.[23] Hence, the variations in solubil-
ity and aggregation tendencies in the culture medium within
this compound series could explain the observed trend in cy-
totoxicity with n=6 and greater. It appears that the hexyl com-
pound shows the best balance between solubility and lipophi-
licity.
Figure 6. Analysis of the possible reversal effects of the pan-caspase inhibi-
tor zVADfmk (20 mm) on the cytotoxicity of compound 5 after 48 h incuba-
tion at a concentration of 15 mm toward Hep3B, MDAMB-231, and SKHep-1
cancer cells. Results shown represent the mean ÆSD from triplicate tests;
the data shown are from a representative experiment taken from three inde-
pendent experiments giving similar results.
Conclusions
We have developed a novel class of 9,9-dialkylated derivatives
of 4,5-diazafluorene that show a good therapeutic potential to
be exploited in bioorganic and medicinal chemistry studies.
To further determine whether compound 5 exerts any toxic
effects in the kidney, hematoxylin and eosin (H and E) staining
562
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 559 – 566

9,9-Dialkyl-4,5-diazafluorene Derivatives
Figure 9. Cytotoxic potency of the homologous series plotted as a function
of alkyl chain length (n).
length increases beyond n=6, the compounds show a de-
creased cellular uptake because of solubility/lipophilicity issues
and/or a greater aggregation tendency. In the absence of the
diaza functionality, cytotoxicity is abolished for the dialkylfluor-
enes at clinically relevant concentrations. The particular action
of these compounds toward different cancer cell lines of
human origin could be of great importance in the future
search for novel antitumor drugs. Research is ongoing to inves-
tigate the factors that may modify the biological properties of
these compounds. The influence of metal complexation with
these diazafluorene-type ligands on their biological effects is
worthy of future examination.
Figure 8. a) Representative Hep3B xenografted nude mice treated with vehi-
cle (left column) or compound 5 (right column) at 10 mgkgÀ1 for nine con-
secutive days. The tumor size of the Hep3B xenograft treated with vehicle is
larger than that of the animal treated with compound 5. b) Histochemistry
study of kidney sections from sacrificed mice to observe any possible kidney
toxicity; neither necrotic nor damaged tissue was observed. c) Histochem-
istry study of Hep3B tumor sections from sacrificed mice to detect any possi-
ble xenograft tumor necrosis; no systemic necrosis was observed.
Experimental Section
General
All reactions were performed under N2 atmosphere. Solvents were
carefully dried and distilled from appropriate drying agents prior
to use. Commercially available reagents were used without further
purification unless otherwise stated. 4,5-Diazafluoren-9-one was
prepared according to published procedures.[3a,14] All reactions
were monitored by thin-layer chromatography (TLC) with Merck
pre-coated glass plates. Flash-column chromatography was carried
out with silica gel from Merck (230–400 mesh). FABMS data were
recorded in m-nitrobenzyl alcohol matrices on a Finnigan MAT
SSQ710 system, and high-resolution (HR) MALDI-TOF mass spectra
were obtained on an Autoflex Bruker MALDI-TOF MS instrument.
1H and 13C NMR spectra were measured in CDCl3 on a Bruker
400 MHz FT-NMR spectrometer, and chemical shifts (d) are reported
in ppm relative to (CH3)4Si.
Table 2. Plasma liver enzyme assays for vehicle control and compound 5-
treated Hep3B xenografted athymic nude mice.[a]
Enzyme
Vehicle
5
Reference value
ALT
AST
67.9Æ12.8
57.1Æ14.7
28-132
59-246
197.0Æ41.5
161.9Æ45.3
[a] Number of mice=7 for both vehicle control group and test group
treated with compound 5. Enzyme levels were determined by Vet bio-
chemistry assay kits (IDEXX Laboratories instrument) and are expressed as
units per liter. Results represent the mean ÆSD.
Synthesis of 1: This compound was prepared according to a slight
modification of published procedures.[1,3a,15] A mixture of 4,5-diaza-
fluoren-9-one (500 mg, 2.74 mmol) and NH2NH2·H2O (5 mL) was
heated at 1308C for 24 h. After cooling, the mixture was extracted
with CH2Cl2 (3ꢂ25 mL). The organic layers were combined and
dried over anhydrous Na2SO4; solvents were removed in vacuo.
The crude product was purified by column chromatography on
silica gel, eluting with EtOAc to provide 1 (295 mg, 1.75 mmol) as a
white solid in 64% yield. Rf =0.28 (EtOAc); 1H NMR (400 MHz,
CDCl3): d=8.70–8.75 (d, J=3.6 Hz, 2H), 7.89–7.90 (d, J=7.2 Hz,
2H), 7.29–7.32 (m, 2H), 3.88 ppm (s, 2H); 13C NMR (100 MHz,
CDCl3): d=159.01, 149.73, 137.55, 133.07, 122.84, 32.45 ppm; FAB-
We evaluated their possible utility as anticancer agents and
also studied the structure–activity correlation between antitu-
mor efficacy, alkyl chain length, and the nature of fluorene
ring. There is a clear relationship between the mean potency
[log(1/IC50)] and alkyl chain length. The results obtained on the
in vitro antitumor activity of compounds 1–8 toward the three
different human carcinoma cell lines demonstrate that the
hexyl congener is the best candidate for the selected cell lines;
these results are compatible with improved lipophilic entry of
this compound across the cell membrane. As alkyl chain
ChemMedChem 2010, 5, 559 – 566
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
563

W.-Y. Wong, G. Y.-M. Cheng, C.-H. Chui, et al.
MED
MS: m/z 168 [M]+; MALDI-TOF HRMS: m/z calcd for C11H8N2:
MALDI-TOF HRMS: m/z calcd for C23H32N2: 337.2643, found:
169.0766, found: 169.0757.
337.2667.
9,9-Dioctyl-4,5-diazafluorene (6): Dark oil (25% based on meth-
od A and 63% based on method B). Rf =0.54 (CH2Cl2/EtOAc 3:1);
1H NMR (400 MHz, CDCl3): d=8.70–8.68 (m, 2H), 7.72–7.70 (m, 2H),
7.30–7.28 (m, 2H), 2.00–1.98 (m, 4H), 1.21–1.04 (m, 20H), 0.83–0.80
(t, J=7.2 Hz, 6H), 0.64 ppm (m, 4H); 13C NMR (100 MHz, CDCl3): d=
158.73, 149.71, 145.15, 130.81, 123.14, 51.54, 39.43, 31.94, 30.10,
29.36, 29.32, 24.21, 22.78, 14.27 ppm; FAB-MS: m/z 393 [M]+;
MALDI-TOF HRMS: m/z calcd for C27H40N2: 393.3270, found:
393.3276.
General procedures for the synthesis of 9,9-dialkylated diaza-
fluorene derivatives 2–8
Method A: 4,5-Diazafluorene (0.20 mmol) and NaH (60%,
0.60 mmol) were dissolved in THF (5 mL) in a round-bottomed
flask. The appropriate alkyl bromide (0.60 mmol) was added in one
portion. The mixture was held at reflux for 3 h, and then the reac-
tion was quenched with H2O. The solvent was evaporated, and the
reaction mixture was extracted with Et2O and dried over anhydrous
Na2SO4. The target product was isolated by purification of the
crude product using column chromatography on silica gel (first
using hexane as eluent to remove excess alkyl bromide, and then
CH2Cl2/EtOAc (3:1 v/v) to elute the product).
9,9-Didodecyl-4,5-diazafluorene (7): Pale-yellow oil (12% based
on method A). Rf =0.67 (CH2Cl2/EtOAc 3:1); 1H NMR (400 MHz,
CDCl3): d=8.70–8.68 (m, 2H), 7.72–7.70 (m, 2H), 7.30–7.27 (m, 2H),
2.01–1.98 (m, 4H), 1.28–1.04 (m, 36H), 0.90–0.85 (t, J=7.1 Hz, 6H),
0.67–0.63 ppm (m, 4H); 13C NMR (100 MHz, CDCl3): d=158.51,
149.49, 144.92, 130.58, 122.91, 51.31, 39.20, 31.89, 29.88, 29.58,
29.51, 29.45, 29.31, 29.20, 24.00, 22.68, 14.12 ppm; FAB-MS: m/z
505 [M]+; MALDI-TOF HRMS: m/z calcd for C35H56N2: 505.4522,
found: 505.4515.
Method B: 4,5-Diazafluorene (0.20 mmol) was added to a mixture
of DMSO (5 mL) and aqueous 50% NaOH (0.5 mL) in a round-bot-
tomed flask. A catalytic amount of benzyltriethylammonium chlo-
ride was added to the reaction mixture followed by the addition of
alkyl bromide (0.80 mmol). The mixture was stirred at 608C for
24 h. The resulting mixture was poured into EtOAc (30 mL), and
the slurry was filtered to remove the solid bromide salt. The filtrate
was washed with 1m HCl and H2O (2ꢂ10 mL), and then dried over
anhydrous Na2SO4. The desired product was obtained by purifica-
tion via column chromatography on silica gel (initially using
hexane as eluent to remove the excess organic bromide, followed
by CH2Cl2/EtOAc (3:1 v/v) to isolate the product). It was found that
this method gave better synthetic yields than method A. We
expect that the product yields for those compounds prepared
below based on method A only could be improved greatly with
the use of method B instead.
9,9-Dihexadecyl-4,5-diazafluorene (8): Pale-yellow solid (56%
based on method B). Rf =0.71 (CH2Cl2/EtOAc 3:1); 1H NMR
(400 MHz, CDCl3): d=8.70–8.68 (m, 2H), 7.72–7.69 (m, 2H), 7.78–
7.30 (m, 2H), 2.00–1.96 (m, 4H), 1.29–1.04 (m, 52H), 0.89–0.86 (t,
J=7.0 Hz, 6H), 0.67–0.64 ppm (m, 4H); 13C NMR (100 MHz, CDCl3):
d=158.52, 149.49, 144.91, 130.57, 122.90, 51.31, 39.20, 32.81,
31.92, 29.87, 29.65, 29.57, 29.44, 29.35, 29.19, 25.73, 24.00, 22.69,
18.44, 14.12 ppm; FAB-MS: m/z 617 [M]+; MALDI-TOF HRMS: m/z
calcd for C43H72N2: 617.5773, found: 617.5812.
Cancer cell culture
The human cancer cell lines Hep3B, MDAMB-231, and SKHep-1
were cultured in RPMI-1640 medium with 5% fetal bovine serum
(complete cell culture medium) at 378C under an atmosphere con-
taining 5% CO2 in a humidified incubator.
9,9-Diethyl-4,5-diazafluorene (2): Sticky dark solid (10% based on
method A). Rf =0.27 (CH2Cl2/EtOAc 3:1); H NMR (400 MHz, CDCl3):
1
d=8.71–8.69 (m, 2H), 7.72–7.70 (m, 2H), 7.31–7.28 (m, 2H), 2.10–
2.04 (m, 4H), 0.40–0.36 ppm (t, J=7.4 Hz, 6H); 13C NMR (100 MHz,
CDCl3): d=158.84, 149.57, 144.15, 130.67, 122.93, 52.27, 31.64,
8.60 ppm; FAB-MS: m/z 225 [M]+; MALDI-TOF HRMS: m/z calcd for
C15H16N2: 225.1392, found: 225.1390.
Bone marrow aspirate
Bone marrow cells from patients with nonmalignant hematological
disorder during presentation were obtained after informed consent
with approval obtained from the ethical committee of the institute
(Haematology Division, Department of Medicine and Therapeutics,
Prince of Wales Hospital, The Chinese University of Hong Kong).
Mononuclear cells were further isolated after Ficoll–Paque separa-
tion, and cells were re-suspended in complete cell culture medium.
9,9-Dipropyl-4,5-diazafluorene (3): Brown solid (14% based on
method A). Rf =0.37 (CH2Cl2/EtOAc 3:1); H NMR (400 MHz, CDCl3):
1
d=8.69–8.68 (m, 2H), 7.73–7.71 (m, 2H), 7.30–7.28 (m, 2H), 2.00–
1.96 (m, 4H), 0.70–0.68 ppm (m, 10H); 13C NMR (100 MHz, CDCl3):
d=158.46, 149.48, 144.87, 130.58, 122.86, 51.44, 41.50, 17.35,
14.33 ppm; FAB-MS: m/z 252 [M]+; MALDI-TOF HRMS: m/z calcd
for C17H20N2: 253.1705, found: 253.1700.
9,9-Dibutyl-4,5-diazafluorene (4): Dark oil (11% based on meth-
od A). Rf =0.47 (CH2Cl2/EtOAc 3:1); 1H NMR (400 MHz, CDCl3): d=
8.70–8.68 (m, 2H), 7.73–7.70 (m, 2H), 7.31–7.28 (m, 2H), 2.02–1.98
(m, 4H), 1.11–1.06 (m, 4H), 0.70–0.62 ppm (m, 10H); 13C NMR
(100 MHz, CDCl3): d=158.52, 149.50, 144.89, 130.59, 122.93, 51.23,
38.98, 26.18, 22.94, 13.74 ppm; FAB-MS: m/z 281 [M]+; MALDI-TOF
HRMS: m/z calcd for C19H24N2: 281.2018, found: 281.2026.
Sulforhodamine B assays for cell viability
Cancer cells were removed from sterile cell culture flasks with tryp-
sin and neutralized with fetal bovine serum. After washing twice
with phosphate buffered saline and centrifugation, cells were re-
suspended in complete cell culture medium at a concentration of
~1ꢂ105 cellsmLÀ1 and counted manually using a hemocytometer
under an inverted microscope. Cancer cells seeded in 96-well mi-
crotiter plates for 24 h were prepared for the screening of our syn-
thesized compounds. The compounds were dissolved in molecu-
lar-biology-grade DMSO. Cisplatin (CDDP) and doxorubicin (DOX)
were used as the positive reference compounds, and were added
at a starting concentration of 50 mm. Test compounds were added
at a starting concentration of 50 mm followed by twofold serial di-
9,9-Dihexyl-4,5-diazafluorene (5): Brown oil (22% based on meth-
od A and 60% based on method B). Rf =0.54 (CH2Cl2/EtOAc 3:1);
1H NMR (400 MHz, CDCl3): d=8.70–8.68 (m, 2H), 7.72–7.70 (m, 2H),
7.30–7.27 (m, 2H), 2.01–1.97 (m, 4H), 1.12–1.01 (m, 12H), 0.78–0.74
(t, J=7.2 Hz, 6H), 0.67–0.63 ppm (m, 4H); 13C NMR (100 MHz,
CDCl3): d=158.49, 149.47, 144.90, 130.58, 122.92, 51.31, 39.21,
31.39, 29.54, 23.96, 22.47, 13.95 ppm; FAB-MS: m/z 337 [M]+;
564
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 559 – 566

9,9-Dialkyl-4,5-diazafluorene Derivatives
lutions and incubated for a further period of 48 h. The maximum
concentration of DMSO employed was 0.1% by volume. Untreated
controls received either total complete cell culture medium or
0.1% DMSO. Afterward, evaluation of possible antiproliferative ac-
tivity of the compounds was performed by sulforhodamine B pro-
tein staining methods. Briefly, cancer cells were fixed with trichloro-
acetic acid (TCA), washed, and stained with sulforhodamine B. Cells
were then washed again with acetic acid, and stained cells were re-
suspended in un-buffered Tris base. Finally, optical absorptions
were measured at l 575 nm using a microplate reader (Victor V,
PerkinElmer, Life Sciences). The IC50 values of these compounds
and of CDDP were calculated from these experimental results.
08-09/I-20). C.-H.C. is supported by the assistant professorship
kindly offered by X.-M.T.
Keywords: cytotoxicity
synthesis · xenografts
·
diazafluorenes
·
hepatoma
·
[1] R. P. Thummel, F. Lefoulon, R. Mahadevan, J. Org. Chem. 1985, 50,
3824–3838.
[2] R. P. Thummel, D. K. Kohli, J. Org. Chem. 1977, 42, 2742–2747.
[3] a) L. J. Henderson Jr. , F. R. Fronczek, W. R. Cherry, J. Am. Chem. Soc.
1984, 106, 5876–5879; b) Y. Wang, W. Perez, G. Y. Zheng, D. P. Rillema,
Inorg. Chem. 1998, 37, 2051–2059; c) R. P. Thummel, F. Lefoulon, J. D.
Korp, Inorg. Chem. 1987, 26, 2370–2376; d) R. A. Klein, C. J. Klein, C. J.
Elsevier, F. Hartl, Organometallics 1997, 16, 1284–1291; e) K. Ono, K.
Saito, Heterocycles 2008, 75, 2381–2413.
Caspase inhibition assays
Cancer cells were seeded at a concentration of ~1ꢂ105 cellsmLÀ1
and counted manually using a hemocytometer under an inverted
microscope. After 24 h, the culture medium was changed, and cells
were pre-incubated with a pan-caspase inhibitor (zVADfmk) for 2 h
before addition of test compounds. After a further incubation of
48 h, cellular viability was determined by the sulforhodamine B
protein staining methods as described above.
[4] W.-Y. Wong, Coord. Chem. Rev. 2005, 249, 971–997 and references cited
therein.
[5] R. A. Klein, P. Witte, R. van Belzen, J. Fraanje, K. Goubitz, M. Numan, H.
Schenk, J. M. Ernsting, C. J. Elsevier, Eur. J. Inorg. Chem. 1998, 319–330.
[6] a) M. Riklin, A. von Zelewsky, A. Bashall, M. McPartlin, A. Baysal, J. A.
Connor, J. D. Wallis, Helv. Chim. Acta 1999, 82, 1666–1680; b) M. H. Chis-
holm, J. C. Huffman, I. P. Rothwell, P. G. Bradley, N. Kress, W. H. Woodruff,
J. Am. Chem. Soc. 1981, 103, 4945–4947.
[7] L. A. Summers, The Bipyridinium Herbicides, Academic Press, New York,
1980.
[8] a) J. Rebek, Jr., J. E. Trends, R. V. Wattley, S. Chakravorti, J. Am. Chem.
Soc. 1979, 101, 4333–4337; b) J. Rebek, Jr., Acc. Chem. Res. 1984, 17,
258–264.
[9] G. D. Mayer, R. F. Krueger, Science 1970, 169, 1214–1215.
[10] a) P. Chandra, M. Woltersdorf, G. J. Wright in Mechanism of Action of An-
tieukaryotic and Antiviral Compounds (Ed.: F. H. Han), Springer, Berlin–
Heidelberg–New York, 1979, pp. 385–413; b) B. G. Maiya, C. V. Ramana,
S. Arounaguiri, M. Nagarajan, Bioorg. Med. Chem. Lett. 1997, 7, 2141–
2144.
[11] a) J. Sturm, L. Schreiber, M. Daune, Biopolymers 1981, 20, 765–785;
b) R. G. M. Wright, L. P. G. Wakelin, A. Fieldes, R. M. Acheson, M. J.
Waring, Biochemistry 1980, 19, 5825–5836; c) G. D. Mayer, Pharmacol.
Ther. 1980, 8, 173–192.
[12] Z. Szulc, J. Mlochowski, M. Fikus, A. D. Inglot, Heterocycles 1984, 22, 73–
78.
In vivo athymic nude mice experiments
Eight-week-old athymic nude mice, weighing ~15–20 g, were pur-
chased from the animal unit of The Chinese University of Hong
Kong and maintained in a sterile facility, in accordance with the in-
stitutional guidelines on animal care, with the required consistent
temperature and relative humidity. All the procedures were ap-
proved by the Animal Research Ethics Committee [accreditation
number: (10-17)inDH/HA&P/8/2/4Pt.3]. Fourteen athymic nude
mice were injected subcutaneously with human hepatocellular car-
cinoma Hep3B cells. They were housed under sterile conditions.
Tumor size was measured daily by electronic calipers. When the
tumor size reached a mean volume of ~200 mm3 [where tumor
volume is determined by: (lꢂwꢂh)/2], they were randomly divided
into two groups. Compounds at a concentration of 10 mgkgÀ1
body weight were administered intraperitoneally each day for a
continuous 9-day period starting from day 1. The control group re-
ceived only buffer carrier [physiological saline for injection with
2% poly(ethylene glycol)]. Each group consisted of seven mice. On
day 10, mice were sacrificed, and H and E staining of autopsy anal-
ysis including kidneys from all animals were investigated for any
significant toxicology effects, including the occurrence of any ne-
crotic tissue. Tumors were also collected for investigation. Whole
blood was also collected, and plasma liver enzymes including ALT
and AST were measured by Vet biochemistry assay kits (for the
IDEXX Laboratories instrument) to detect any possible liver failure
phenomena from both groups of mice by comparison with normal
control ranges.
[13] H. J. Eppley, S. M. Lato, A. D. Ellington, J. M. Zaleski, Chem. Commun.
1999, 2405–2406.
[14] M. J. Plater, S. Kemp, E. Lattmann, J. Chem. Soc. Perkin Trans. 1 2000,
971–979.
[15] J. E. Dickeson, L. A. Summers, Aust. J. Chem. 1970, 23, 1023–1027.
[16] a) H.-C. Su, H.-F. Chen, C.-C. Wu, K.-T. Wong, Chem. Asian J. 2008, 3,
1922–1928; b) H.-C. Su, H.-F. Chen, F.-C. Fang, C.-C. Liu, C.-C. Wu, K.-T.
Wong, Y.-H. Liu, S.-M. Peng, J. Am. Chem. Soc. 2008, 130, 3413–3419;
c) S. P. Dudek, M. Pouderoijen, R. Abbel, A. P. H. J. Schenning, E. W.
Meijer, J. Am. Chem. Soc. 2005, 127, 11763–11768.
[17] a) A. R. Khokhar, Y. Deng, S. Al-Baker, M. Yoshida, Z. H. Siddik, J. Inorg.
Biochem. 1993, 51, 677–687; b) Medicinal Chemistry: Principles and Prac-
tice, 2nd Ed. (Ed.: F. D. King), RSC, Cambridge, 2002, pp. 215–242; c) R. B.
Silverman, The Organic Chemistry of Drug Design and Drug Action, Aca-
demic Press, London, 2004, pp. 11–50.
[18] G. L. Lu, C. L. Ho, Q. Wang, W. Y. Wong, C. H. Chui, R. S. M. Wong, R.
Gambari, F. Y. Lau, M. C. W. Yuen, C. S. W. Tong, A. K. W. Chan, J. C. O.
Tang, K. P. Ho, G. Y. M. Cheng, Aust. J. Chem. 2008, 61, 975–980.
[19] S. H. Chan, K. H. Lam, C. H. Chui, R. Gambari, M. C. W. Yuen, R. S. M.
Wong, G. Y. M. Cheng, F. Y. Lau, Y. K. Au, C. H. Cheng, P. B. S. Lai, C. W.
Kan, S. H. L. Kok, J. C. O. Tang, A. S. C. Chan, Eur. J. Med. Chem. 2009, 44,
2736–2740.
[20] K. H. Lam, R. Gambari, M. C. W. Yuen, C. W. Kan, P. Chan, L, Xu, W. Tang,
C. H. Chui, G. Y. M. Cheng, R. S. M. Wong, F. Y. Lau, C. S. W. Tong, A. K. W.
Chan, P. B. S. Lai, S. H. L. Kok, C. H. Cheng, A. S. C. Chan, J. C. O. Tang,
Bioorg. Med. Chem. Lett. 2009, 19, 2266–2269.
Acknowledgements
The authors acknowledge financial support for this work from
the junior academic staff starting fund offered by X.-M.T. from
the Institute of Textiles and Clothing, The Hong Kong Polytechnic
University (to C.-H.C.; HKPU: 1ZV-5L). R.G. is supported by the Ital-
ian Association for Cancer Research (AIRC), and W.-Y.W. acknowl-
edges financial support from the Hong Kong Research Grants
Council (HKBU202508) and Hong Kong Baptist University (FRG/
[21] a) Z. H. Li, M. S. Wong, Y. Yao, J. Lu, Chem. Eur. J. 2005, 11, 3285–3293;
b) M. Hanif, P. Lu, M. Li, Y. Zheng, Z. Xie, Y. Ma, D. Li, J. Li, Polym. Int.
2007, 56, 1507–1513.
ChemMedChem 2010, 5, 559 – 566
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
565

W.-Y. Wong, G. Y.-M. Cheng, C.-H. Chui, et al.
MED
[22] C. H Chui, R. S. M. Wong, R. Gambari, G. Y. M. Cheng, M. C. W. Yuen,
K. W. Chan, S. W. Tong, F. Y. Lau, P. B. S. Lai, K. H. Lam, C. L. Ho, C. W. Kan,
K. S. Y. Leung, W.-Y. Wong, Bioorg. Med. Chem. 2009, 17, 7872–7877.
[23] J.-Y. Liu, X.-J. Jiang, W.-P. Fong, D. K. P. Ng, Org. Biomol. Chem. 2008, 6,
4560–4566.
Received: December 28, 2009
Revised: February 5, 2010
Published online on March 5, 2010
566
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 559 – 566