Synthesis, biological evaluation and structure-activity relationship studies of isoflavene based Mannich bases with potent anti-cancer activity

Article abstract of DOI:10.1016/j.bmcl.2015.09.027

Phenoxodiol, an analogue of the isoflavone natural product daidzein, is a potent anti-cancer agent that has been investigated for the treatment of hormone dependent cancers. This molecular scaffold was reacted with different primary amines and secondary a

Full text of DOI:10.1016/j.bmcl.2015.09.027

Bioorganic & Medicinal Chemistry Letters xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis, biological evaluation and structure–activity relationship
studies of isoflavene based Mannich bases with potent anti-cancer
activity
Yilin Chen a,b, Shelley L. Cass a, Samuel K. Kutty a, Eugene M. H. Yee a,b, Daniel S. H. Chan a,
Christopher R. Gardner a,b, Orazio Vittorio b,c, Eddy Pasquier b,d, David StC. Black a, Naresh Kumar a,c,
⇑
a School of Chemistry, The University of New South Wales, Sydney, NSW 2052, Australia
b Children’s Cancer Institute Australia, Lowy Cancer Research Centre, UNSW, Randwick, NSW, Australia
c Australian Centre for NanoMedicine, University of New South Wales, Sydney, NSW 2052, Australia
d Metronomics Global Health Initiative, Marseille, France
a r t i c l e i n f o
a b s t r a c t
Article history:
Phenoxodiol, an analogue of the isoflavone natural product daidzein, is a potent anti-cancer agent that
has been investigated for the treatment of hormone dependent cancers. This molecular scaffold was
reacted with different primary amines and secondary amines under different Mannich conditions to yield
either benzoxazine or aminomethyl substituted analogues. These processes enabled the generation of a
diverse range of analogues that were required for structure–activity relationship (SAR) studies. The
resulting Mannich bases exhibited prominent anti-proliferative effects against SHEP neuroblastoma
and MDA-MB-231 breast adenocarcinoma cell lines. Further cytotoxicity studies against MRC-5 normal
lung fibroblast cells showed that the isoflavene analogues were selective towards cancer cells.
Ó 2015 Elsevier Ltd. All rights reserved.
Received 6 August 2015
Revised 7 September 2015
Accepted 10 September 2015
Available online xxxx
Keywords:
Isoflavene
Phenoxodiol
Anti-cancer
Mannich base
Benzoxazine
Hormone-dependent cancers such as breast and prostate can-
cers are on the increase in most Western countries. One in eight
Australian women will be diagnosed with breast cancer by the
time they turn 85. The risk of breast cancer in women increases
with age, with the majority of women diagnosed being between
50 and 69 years of age. In 2013, approximately 15,000 women
were diagnosed with breast cancer and 2700 women lost their
lives to breast cancer.1–4 Prostate cancer is a common cancer in
men, accounting for 32.4% of all male cancers in Australia.2–5 The
risk of developing prostate cancer increases with age, with 85%
of cases diagnosed in men over 65 years of age. In 2012, prostate
cancer was estimated to account for 15% of the total burden of can-
cer in men in Australia, second only to lung cancer.2–5
Among the various classes of phytoestrogens, isoflavones have
been among the most widely studied due to their marked estro-
genic activities.6 Phenoxodiol 1, a synthetic isoflavene derivative,
has been tested in clinical trials for the treatment of drug-resistant
ovarian (NCT00382811) and prostate (NCT00557037) cancer. Its
broad spectrum anti-cancer activity has been attributed to its abil-
ity to induce mitotic arrest, terminal differentiation and apoptosis
in cancer cells.7
Phenols undergo the Mannich reaction quite readily, and the
resultant product is typically ortho substituted relative to the
hydroxyl group. The Mannich reactions of isoflavones containing
a hydroxyl group at C7 give products substituted at either the C6
or C8 position, as the pendant ring B is less activated and does
not compete with ring A under these reaction conditions. However,
for phenoxodiol 1, the combination of the electron-donating
hydroxyl group on C7, the activating para-ether oxygen and the
absence of a C4 ketone group make the C6 position of phenoxodiol
1 more reactive than the C8 position.
Phytoestrogens are estrogen-like substances found in plants
such as beans, grains, vegetables, fruit and seeds. The high dietary
intake of phytoestrogens in Asian populations, due in part to their
high soy food consumption, has been linked to lower incidences of
hormone-dependent cancers such as prostate and breast cancer.5
Khilya and co-workers generated the 6,8-bis-piperdin-1-yl-
methyl- and 6,8-bis-piperazin-1-ylmethyl-substituted analogues
of isoflavones, biochanin A and orobol through Mannich bases.8
However, the synthesis of Mannich bases of isoflavenes has not
been investigated. Recently, Wang et al. chose isoflavones as the
⇑
Corresponding author at present address: School of Chemistry, The University
of New South Wales, Sydney, NSW 2052, Australia. Tel.: +61 2 9385 4698; fax: +61 2
9385 6141.
E-mail address: n.kumar@unsw.edu.au (N. Kumar).
http://dx.doi.org/10.1016/j.bmcl.2015.09.027
0960-894X/Ó 2015 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Chen, Y.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.09.027

2
Y. Chen et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
O
O
main scaffold to synthesize C8 substituted Mannich bases and oxa-
zinyl isoflavonoid compounds where the oxazinyl ring is fused at
C8 and C7 hydroxyl group via the Mannich reaction with good
in vitro anti-tumour activities.9 These studies have shown that iso-
flavones readily undergo the Mannich reaction and are able to form
the corresponding Mannich bases with high chemo- and regiose-
lectivities. Due to its potent anti-proliferative activities, and in con-
tinuation of our extensive research into isoflavenes,10–12 it is of
interest to synthesise new aminomethyl and oxazinyl fused analogues
of phenoxodiol and evaluate their anti-cancer properties.13,14
Initial attempts towards the Mannich reaction involved treating
the isoflavene phenoxodiol 1 with different primary amines.
Phenoxodiol 1 was reacted with formaldehyde and primary amines
in a molar ratio of (1:12:2) in EtOH.9 The progress of the reaction
was monitored by TLC and the crude products were recrystallized
from DCM and n-hexane. Generally, the reaction proceeded
smoothly in all cases with aminomethylation occurring at the C6
position followed by benzoxazine formation (Scheme 1).
HO
O
(a)
N
N
O
1
OH
2f
80%
Scheme 2. Reagents and conditions: (a) benzylamine (2 equiv), formaldehyde
(12 equiv), EtOH, 24 h, reflux.
O
O
HO
O
(a)
N
R
OH
OH
1
2g: R = Benzyl
2h: R = 2-Hydroxyethyl
2i
70%
80%
54%
82%
: R = 3-Butoxylpropyl
2j : R = Phenylethyl
Scheme 3. Reagents and conditions: (a) primary amines (2 equiv), formaldehyde
(12 equiv), EtOH, 48 h, rt.
Interestingly, when this reaction was carried out with benzy-
lamine, NMR analysis of the reaction product revealed a different
structure than expected. This was determined to be the di-substi-
tuted benzoxazine 2f (Scheme 2).
HO
O
HO
O
(a)
H
N
In order to synthesize a mono-substituted benzylamine benzox-
azine product 2g, the reaction was attempted at room temperature
and monitored via TLC. It was observed that the reaction needed a
longer time period of 48 h and that the mono-benzoxazine product
2g precipitated cleanly from the reaction mixture and was col-
lected by filtration in 70% yield and no recrystallization was
needed. This indicates that the duration of the reaction and the
temperature had a significant effect on the type of product formed
and the yield. This further highlights the different reactivity of the
phenoxodiol 1 to Mannich reactions at different temperatures.
Based on these findings, further reactions with primary amines
were carried out under these conditions to give the various desired
analogues in 54–82% yield (Scheme 3).
When the ratio of the phenoxodiol 1, formaldehyde and benzy-
lamine changed from 1:12:6 to 1:12:2, aminomethyl compound 3a
was formed instead of the expected benzoxazine product 2g
(Scheme 4). It was hypothesized that the excess of benzylamine
led to complete consumption of formaldehyde, preventing further
reaction to the benzoxazine. As a result, it was clear that by varying
the molar ratio of the reactants the formation of the products could
be controlled. Furthermore, literature reports of Mannich reactions
confirm that the structure of the reaction product, whether a ben-
zoxazine or aminomethyl species, depends on the ratio of the
reagents used.15–17
1
OH
OH
3a
20%
Scheme 4. Reagents and conditions: (a) benzylamine (6 equiv), formaldehyde
(12 equiv), EtOH, 48 h, rt.
HO
O
HO
O
H
(a)
N
R
OH
1
OH
3b
: R = Isopropyl
3c: R = Heptyl
18%
66%
84%
60%
55%
3d
: R = Butyl
3e: R = Phenethyl
3f
: R = 2-Hydroxyethyl
Scheme 5. Reagents and conditions: (a) primary amines (4 equiv), formaldehyde
(5 equiv), EtOH, 48 h, rt.
of the benzoxazine side product. In order to improve the yield for
the isopropylamine reaction, parameters such as temperature
and solvent were investigated. Initially the reaction was carried
out in 1,4-dioxane at room temperature, but with very little pro-
duct formation. Then, the reaction was heated to 70 °C and after
48 h, the aminomethyl-substituted compound precipitated cleanly
from the reaction mixture, but in poor yield. However, TLC analysis
of the filtrate showed that some benzoxazine had also formed in
the reaction. It was evident that the molar ratio of the formalde-
hyde was difficult to control and the formaldehyde left after the
generation of the iminium ions reacted with the aminomethyl-
substituted compound to form the benzoxazine.
Subsequently, a series of primary amines was reacted with
lower concentration of formaldehyde to form aminomethyl-substi-
tuted phenoxodiols 3b–f (Scheme 5). Briefly, the formation of 3f
was achieved by reacting phenoxodiol
1 with 37% aqueous
formaldehyde and ethanolamine in a molar ratio of 1:5:4 in EtOH
which proceeded smoothly with aminomethylation occurring at
the C6 position to give 3f in 55% yield (Scheme 5).
To overcome this problem, other methods were also investi-
gated for the synthesis of aminomethyl-substituted isoflavenes.
One possibility was to use a formyl-substituted analogue of phe-
noxodiol 1, which under reductive amination conditions could
yield the desired aminomethyl-substituted phenoxodiol.
The low yield of the isopropylamine-derived compound 3b was
attributed to the steric bulk of isopropylamine and the formation
O
O
HO
O
(a)
N
R
Hofslokken and Skattebol have reported a method for prepara-
tion of ortho formyl derivatives of phenols, which could be applied
to the phenolic phenoxodiol.18 Following the literature method,
paraformaldehyde was added into the mixture of phenoxodiol 1,
anhydrous magnesium dichloride and triethylamine in the ratio
3:1:2:2 in THF. The reaction was heated at reflux for 4 h. Then
the crude product was washed with 5% aq HCl to yield a C6-formyl
phenoxodiol 4 (Scheme 6).
OH
OH
1
2a: R = Heptyl
20%
40%
44%
55%
22%
2b
: R = Isopropyl
2c: R = Methyl
2d: R = Butyl
2e
: R = Allyl
Scheme 1. Reagents and conditions: (a) primary amines (2 equiv), formaldehyde
(12 equiv), EtOH, 24 h, reflux.
Please cite this article in press as: Chen, Y.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.09.027

Y. Chen et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
3
HO
O
HO
O
(a)
O
OH
4
82%
(b)
OH
1
HO
O
H
O
N
OH
3g
20%
Figure 1. ORTEP diagram of 3-(4-hydroxy-3-(piperidin-1-ylmethyl)phenyl)-6-
(piperidin-1-ylmethyl)-2H-chromen-7-ol 5b.
Scheme 6. Reagents and conditions: (a) paraformaldehyde (3 equiv), magnesium
dichloride (2 equiv), NEt3 (2 equiv), 4 h, reflux; (b) 3-butoxypropylamine, Na(CH3-
COO)3BH (1.5 equiv), 3-butoxypropylamine (1 equiv), THF, 48 h, rt, under N2.
HO
R1
O
Abdel-Magid et al.19 reported the use of sodium triacetoxyboro-
hydride as a reducing agent for the reductive amination of aldehy-
des. Following this literature procedure, the 6-formyl-phenoxodiol
was reacted with 3-butoxypropylamine in the same molar ratio of
1:1 in THF at room temperature, under N2. Sodium triacetoxyboro-
hydride (1.5 equiv) was added to the reaction mixture and stirring
continued at room temperature under a N2 atmosphere for 48 h.
The reaction mixture was quenched by adding aqueous 3 N NaOH,
which changed the colourless solution to dark green. The mixture
was extracted with ethyl acetate, dried and evaporated to give the
desired product 3g.
OH
5c: R1 = Diethylamine
5e
16%
30%
: R1 = Morpholine
HO
R1
O
HO
O
(a)
R2
OH
5d: R1 & R2 = Diethylamine
5%
8%
1
OH
: R1 & R2 = Morpholine
R3
5f
Mannich reactions of phenoxodiol 1 were also carried out with
the secondary amines, piperidine, diethylamine and morpholine.
The initial reaction of phenoxodiol 1 with a secondary amine and
formaldehyde in molar ratio of 1:2:12 in dioxane, gave multiple
products. Subsequently, the quantity of the amine and formalde-
hyde were reduced to 1.5 and 5 equiv, respectively, but this too
resulted in the formation of multiple products. For example, piper-
idine and formaldehyde were mixed in dioxane at reflux for
30 min. Phenoxodiol 1 was then added to the solution, which
was maintained at reflux overnight. Extensive column chromatog-
raphy was required and led to the isolation of two major products:
5a (23%) and 5b (10%) (Scheme 7).
HO
R1
O
R2
OH
5g: R1, R2 & R3 = Morpholine
50%
Scheme 8. Reagents and conditions: (a) amine (1.5 equiv), formaldehyde (5 equiv),
dioxane, overnight, reflux.
ceeds quickly and produces fewer unwanted side products.8 The
reaction of phenoxodiol 1 with bis(dimethylamine)methane under
reflux conditions generated the aminomethyl-substituted Mannich
base product 6 in 90% yield (Scheme 9).
The main objective of the synthesis of Mannich bases was to
enhance the potency and specificity against cancer cells of parental
compound phenoxodiol 1. To assess the anti-proliferative proper-
ties, the synthesized Mannich bases were grouped into 3 classes
(Fig. 2), and their effects on cancer cell proliferation were assessed
in vitro using MDA-MB-231 triple-negative breast cancer and SHEP
neuroblastoma cell lines.
As shown in Figure 3, all 25 compounds inhibited cancer cell
proliferation in a dose-dependent manner. When compared with
the parental compound phenoxodiol 1, both Class 1 and Class 2
displayed improved anti-proliferative activities against MDA-
MB-231 breast cancer cells except compound 2b, which showed
similar potency as phenoxodiol 1. For Class 3, only compounds
5a, 5b and 5d displayed improved anti-proliferative activity
against MDA-MB-231 breast cancer cells. Against SHEP neuroblas-
toma cells, the majority of Class 1 and 2 also showed improved
The structure of compound 5b was further confirmed by X-ray
crystallography to be a di-substituted aminomethyl piperidine
phenoxodiol (Fig. 1).20
The reaction of phenoxodiol 1 with diethylamine yielded simi-
lar products 5c (16 %) and 5d (5 %). Once again, extensive column
chromatography was required to isolate these products (Scheme 8).
However, for the morphine reaction, extensive chromatography of
the reaction mixture led to the isolation of three products, com-
pound 5e (30%), compound 5f (8%) and compound 5g (50%).
It has been shown in the literature that aminals can function as
preformed amine substrates in the Mannich reaction. These highly
reactive bis-(dialkylamino)methane derivatives can behave as
strong electrophiles and do not require formaldehyde in the
aminoalkylation process.16,17 The aminoalkylation reaction pro-
HO
O
N
HO
O
(a)
5a 23%
OH
HO
N
O
HO
O
(a)
HO
O
OH
1
N
OH
N
OH
1
5b
10%
6
90%
OH
Scheme 7. Reagents and conditions: (a) piperidine (1.5 equiv), formaldehyde
Scheme 9. Reagents and conditions: (a) bis(dimethylamine)methane, formalde-
(5 equiv), dioxane, overnight, reflux.
hyde (5 equiv), EtOH, 3 h, reflux.
Please cite this article in press as: Chen, Y.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.09.027

4
Y. Chen et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
Class 2
Class 3
Class 1
HO
O
HO
O
O
O
R1
R
N
R
OH
OH
OH
5c: R1 = Diethylamine
5e
: R1 = Morpholine
3a
3b
: R = Benzylamine
2a: R = Heptyl
2b
HO
R1
O
: R = Isopropylamine
: R = Isopropyl
3c: R = Heptylamine
2c: R = Methyl
3d
: R = Butylamine
R2
OH
2d
2e
: R = Butyl
3e: R = Phenethylamine
: R = Allyl
3f
: R = Ethanolamine
2g: R = Benzyl
: R1 & R2 = Diethylamine
: R1 & R2 = Morpholine
3g
: R = 3-Butoxypropylamine
2h
: R = Ethanol
2i : R = 3-Butoxylpropyl
2j
5d
5f
6 : R = Dimethylamine
: R = 2-Phenylethyl
R3
O
O
HO
R1
O
N
R
R
N
R2
OH
O
2f
: R = Benzyl
1
: R , R2 & R3 = Morpholine
5g
Figure 2. Class 1 Aminomethyl substituted isoflavenes, Class 2 benzoxazine substituted isoflavenes, Class 3 isoflavenes with secondary amine.
anti-proliferative activity with the exception of compounds 2b, 2e,
3b, 2h, and 3f. For Class 3, only compounds 5b and 5d displayed
slightly improved anti-proliferative activity compared to phenoxo-
diol 1 (Table 1).
Chemotherapy agents target rapidly dividing cells and com-
pounds that are toxic to normal and non-dividing cells are said
to be non-specific. To assess the specificity of the Mannich bases,
several compounds were selected and tested for their toxicity
against normal quiescent cells in vitro using MRC-5 normal lung
fibroblasts (Fig. 4).
The specificity of the Mannich bases was calculated using the
IC50 value for normal cells (MRC-5) divided by the IC50 value for
the other cell types, according to Eq. 1.
The specificity of phenoxodiol 1 appeared to be cell dependent,
with a good specificity against SHEP neuroblastoma cells but a rel-
atively poor specificity against MDA-MB-231 breast cancer cells
(Table 2).
compound 3c with an n-heptyl side chain and compound 3e with a
2-phenylethyl side chain exhibited the most potent anti-prolifera-
tive activity. In contrast, short chain alkyl amine derivatives such
as compounds 6 and 3f showed the weakest anti-proliferative
activity against MDA-MB-231, and compounds 3b and 3f showed
weak anti-proliferative activity against SHEP. These results demon-
strate that the formation of Mannich bases with long chain and
bulkier aryl amine groups significantly improved the anti-prolifer-
ative activity, whereas, short alkyl side chains or a terminal alcohol
did not enhance the anti-proliferative properties of the parent iso-
flavene 1.
For benzoxazines (Class 2), compounds 2a, 2d, 2j, 2f, 2i and 2g
exhibited superior anti-proliferative activity against cancer cells
compared to phenoxodiol 1. Compounds 2a and 2j, which were
formed from the same amines as compounds 3c and 3e, respec-
tively, exhibited similarly potent anti-proliferative activity sug-
gesting no significant difference in activity between
benzoxazines and phenolic aminomethyl analogues. However,
compound 2a a benzoxazine of n-heptyl amine was more specific
for MDA-MB-231 than compound 3c, an aminomethyl derivative
of n-heptyl amine. The difference in specificity between Class 1
and Class 2 compounds is interesting and needs further investiga-
tion. Benzylamine analogues, such as compound 2f, the di-substi-
tuted benzoxazine and compound 2g a mono-benzoxazine had
similar anti-proliferative activity, which suggests that both pheno-
lic groups can be substituted. Compounds 2b and 2h showed weak
anti-proliferative properties, similar to their analogues in Class 1
(compounds 3b and 3f, respectively). Additionally, compounds 2c
and 2e showed only slight improvement in their anti-proliferative
activity compared to phenoxodiol 1. Once again, this demonstrates
that short alkyl side chains do not significantly improve the activ-
ity. Thus, Class 2 showed a similar trend to that of Class 1 and com-
pounds with long chain and bulkier aryl amine derivatives were
more potent anti-proliferative agents.
Determination of cell line selectivity values,
IC50 of compound against normal cell
IC50 of compound against cell line of interest
Specificity ¼
ð1Þ
Selected compounds 2a, d, j, 3a, c, d, e and 5d were tested
against MRC-5 fibroblasts. All eight compounds showed improved
specificity towards MDA-MB-231 breast cancer cells but varying
specificity against SHEP neuroblastoma cells compared to parental
compound phenoxodiol 1. Phenylethylamine analogue 2j followed
by 3e showed highest specificity value against the two cell lines
(Table 2).
Phenoxodiol 1 is an isoflavene with potent anti-cancer and
chemoresistance modulation properties.21,22 Our results have
shown that the Mannich bases derived from phenoxodiol 1 dis-
played up to 30 times more potent anti-proliferative activity
against the cancer cell lines MDA-MB-231 and SHEP. The synthe-
sized molecules are analogues of phenoxodiol 1 with varying
For isoflavenes with secondary amines (Class 3), only com-
pounds 5d, a diethylamine analogue, and 5b, a piperidine analo-
gue, showed more potent anti-proliferative activity than
phenoxodiol 1. Both compounds are di-substituted at C6 and C30
and are more potent than their mono-substituted counterparts.
amine substitutions. Hence,
a structure–activity relationship
(SAR) can be established by comparison of the biological activities
of the different phenoxodiol Mannich bases (Table 3).
In the case of the aminomethyl-substituted isoflavenes (Class
1), compounds 3a, 3c, 3d, 3e and 3g exhibited improved anti-pro-
liferative activity against both MDA-MB 231 and SHEP cancer cell
lines compared to phenoxodiol 1 (Fig. 3). Among these compounds,
Interestingly, compound 5f,
a morpholine analogue, showed
almost no anti-proliferative effect. This may be due to the poor sol-
ubility of this compound. Also, compounds 5e and 5g had poor
Please cite this article in press as: Chen, Y.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.09.027

Y. Chen et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
5
Figure 3. In vitro anti-proliferative properties of phenoxodiol 1 and Mannich bases. MDA-MB-231 or SHEP cells were incubated with a range of concentrations of drug
compound for 72 h. (A1–3) Cell proliferation as a function of compound concentration. Points show % of cell proliferation as compared to untreated control cells. Error bars
show SE of at least four independent experiments. (B) Histogram representation of IC50 values of drug compounds. Error bars show the SE of at least four independent
experiments.
activity, which suggests that morpholine has a detrimental impact
on anti-proliferative activity.
Steric effects and lipophilicity are two important factors that
influence the potency of drug molecules. Lipophilicity affects the
Please cite this article in press as: Chen, Y.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.09.027

6
Y. Chen et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
Table 1
Table 2
IC50 values of compounds for MDA-MB-231 and SHEP cells
IC50 values of selected compounds for MRC-5 cells and specificity values for MDA-MB-
231 and SHEP cells versus MRC-5 cells
Class
Compound
IC50
MDA-MB-231
(lM)
Compound
IC50
(
l
M)
Specificity value
MDA-MB-231
SHEP
8.4
MRC-5
SHEP
Nil
1
Phenoxodiol 1
35.6
Phenoxodiol 1
109
3.0
19.8
7.6
266
7.8
9.0
8.9
45.3
7.0
13.0
8.7
8.7
274
7.1
5.8
8.5
20.9
6.3
6
21.4
5.3
9.5
1.8
3.8
0.8
22.2
5.8
5.8
5.7
15.1
2.8
4.0
1.8
2a
2d
2j
3a
3c
3d
3e
5d
11.9
33.8
197.0
40.9
16.0
33.8
36.7
21.2
3a
3b
3c
3d
3e
3f
27.5
5.1
3g
2
2a
2b
2c
2d
2e
2f
0.6
35.8
9.7
4.5
20.1
5.7
1.4
12.2
6.6
3.9
9.1
Table 3
LogP value of Mannich bases (calculated from ChemBioDrawUltra 12.0.3)
3.7
2g
2h
2i
4.9
25.7
4.9
5.7
24.6
4.9
Class 1
LogP
Class 2
LogP
Class 3
Log P
6
2.65
4.15
2.93
4.76
3.51
4.28
1.75
4.10
2a
2b
2c
2d
2e
2f
2g
2h
2i
5.44
3.60
2.94
4.19
3.64
6.65
4.68
2.43
4.14
4.96
5a
5b
5c
5d
5e
5f
3.38
4.06
3.32
3.94
2.25
1.79
1.33
2j
0.7
0.7
3a
3b
3c
3d
3e
3f
3
5a
5b
5c
5d
5e
5f
32.8
5.4
68.3
3.0
126.2
500.0
89.1
19.9
4.2
35.7
3.4
86.8
171.0
69.0
5g
3g
5g
2j
ability of the drugs to cross the plasma membrane while steric hin-
drance affects the ability of molecules to bind to certain receptors
and induce biological responses. Additionally, the synthesized
compounds have functional groups such as amine and hydroxyl
groups that are prone to metabolism. With increased steric hin-
drance, the metabolism of the Mannich bases by cellular enzymes
may be reduced, thus potentially enhancing the lifetime of the
molecules within cancer cells.
Investigations of all analogues were pursued to determine the
effect of branched and longer alkyl chains on the potency and
specificity of compounds. However, excessive lipophilicity may
become an issue as the alkyl chain is increased due to the increase
in cellular uptake and retention, potentiating non-specific toxicity.
Thus, it would be interesting to understand the effect of various
alkyl substituents on lipophilicity, logP and anti-proliferative
activity.
In order to investigate potential correlations, lipophilicity was
plotted against the IC50 values of the compounds against MDA-
MB-231 and SHEP (Fig. 5).
As shown in Figure 5, there is a significant correlation observed
between the lipophilicity of the compounds and their anti-prolifer-
ative activity against MDA-MB-231 (R2 = 0.23; p value = 0.015) and
SHEP cancer cells (R2 = 0.37; p value = 0.0013). As logP increases,
the IC50 value decreases and therefore the anti-proliferative activ-
ity increases. This suggests that lipophilicity plays an important
role in determining the activity and selectivity of these Mannich
bases against cancer cells.
In summary, a range of isoflavene Mannich bases was synthe-
sized by the reaction of phenoxodiol 1 with various primary and
secondary amines. Interestingly, it was observed that varying the
concentration of amine, the temperature or the time of reaction
led to the formation of different products. Biological assays showed
Figure 4. In vitro anti-proliferative properties of phenoxodiol 1⁄ and compounds 2a, d, 3a, c, d, e, j, and 5d against MRC-5. (A) Cell proliferation as a function of compound
concentration. Points show % of cell proliferation as compared to untreated control cells. Error bars show SE of at least four independent experiments. *At concentrations
above 100 lM the media becomes saturated with phenoxodiol 1, resulting in a plateauing of the activity. (B) Histogram representation of IC50 values of drug compounds.
Error bars show the SE of at least four independent experiments.
Please cite this article in press as: Chen, Y.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.09.027

Y. Chen et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
7
Figure 5. Plot of logP versus IC50 (lM) for MDA-MB-231 and SHEP.
5. Mense, S. M.; Hei, T. K.; Ganju, R. K.; Bhat, H. K. Environ. Health Perspect. 2008,
116, 426.
6. Gruber, C. J.; Tschugguel, W.; Schneeberger, C.; Huber, J. C. N. Engl. J. Med. 2002,
346, 340.
7. Gamble, J. R.; Xia, P.; Hahn, C. N.; Drew, J. J.; Drogemuller, C. J.; Brown, D.;
Vadas, M. A. Int. J. Cancer 2006, 118, 2412.
8. Bondarenko, S. P.; Levenets, A. V.; Frasinyuk, M. S.; Khilya, V. P. Chem. Nat.
Compd. 2003, 39, 271.
9. Wang, D.; Hou, L.; Wu, L.; Yu, X. Chem. Pharm. Bull. (Tokyo) 2012, 60, 513.
10. Heaton, A.; Kumar, N. Eur. Patent WO 2000AU00103 20000215, 2000.
11. Heaton, A.; Kumar, N.; Kelly, G. E.; Husband, A. Eur. Patent WO2000AU01056
20000906, 2001.
that the isoflavene Mannich bases exhibited higher anti-
proliferative activities against MDA-MB-231 breast cancer and
SHEP neuroblastoma cancer cells compared to the parent com-
pound phenoxodiol 1. Notably, the most potent molecule 2j
showed strong potency against both cancer cell lines and low
toxicity against MRC-5 normal human lung fibroblasts, with speci-
ficity values of 266 and 274 for MDA-MB-231 and SHEP cell lines,
respectively. Further analysis is therefore needed to determine the
mode of action of compound 2j.
12. Heaton, A.; Kumar, N.; Kelly, G. E.; Husband, A. U.S. Patent 20090233999, 2009.
13. Heaton, A.; Kumar, N.; Walker, C. Eur. Patent WO2008AU00267 20080229,
2008.
Acknowledgements
14. Husband, A. J.; James, M.; Kumar, N. Eur. Patent WO2009AU00973 20090730,
2010.
15. Kukhareva, T. S.; Krasnova, V. A.; Koroteev, M. P.; Kaziev, G. Z.; Kuleshova, L. N.;
Korlyukov, A. A.; Antipin, M. Y.; Nifant’ev, E. E. Russ. J. Org. Chem. 2004, 40,
1190.
We thank Dr. Mohan M. Bhadbhade for the single crystal X-ray
studies, NMR and BMSF facility, University of New South Wales,
Australia. We also thank the University of New South Wales and
the Australian Research Council for their financial support.
16. Garazd, Y. L.; Panteleimonova, T. N.; Garazd, M. M.; Khilya, V. P. Chem. Nat.
Compd. 2002, 38, 532.
17. Garazd, Y. L.; Panteleimonova, T. N.; Garazd, M. M.; Khilya, V. P. Chem. Nat.
Compd. 2003, 39, 330.
Supplementary data
18. Hofslokken, N.; Skattebol, L. Acta Chem. Scand. 1999, 53, 258.
19. Abdel-Magid, A. F.; Carson, K. G.; Harris, B. D.; Maryanoff, C. A.; Shah, R. D. J.
Org. Chem. 1996, 61, 3849.
20. Crystal data: C27H31N2O3, M.W. = 435.26, P21/c, cell dimensions: a = 12.518(3),
b = 6.3300(13), c = 14.363(3). Crystallographic data excluding structure factors
have been deposited with the Cambridge Crystallographic Data Centre as
supplementary publication number CCDC 1407654. A copy of the data can be
obtained free of charge from CCDC, 12 Union Road, Cambridge CB2 IEZ, UK or
email: deposit@ccdc.cam.ac.uk.
21. Howes, J. B.; de Souza, P. L.; West, L.; Huang, L. J.; Howes, L. G. BMC Clin.
Pharmacol. 2011, 11, 1.
22. Kamsteeg, M.; Rutherford, T.; Sapi, E.; Hanczaruk, B.; Shahabi, S.; Flick, M.;
Brown, D.; Mor, G. Oncogene 2003, 22, 2611.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2015.09.
027.
References and notes
1. Kennecke, H.; Yerushalmi, R.; Woods, R.; Cheang, M. C.; Voduc, D.; Speers, C. H.;
Nielsen, T. O.; Gelmon, K. J. Clin. Oncol. 2010, 28, 3271.
2. Bray, F.; Ren, J. S.; Masuyer, E.; Ferlay, J. Int. J. Cancer 2013, 132, 1133.
3. Ferlay, J.; Soerjomataram, I.; Dikshit, R.; Eser, S.; Mathers, C.; Rebelo, M.; Parkin,
D. M.; Forman, D.; Bray, F. Int. J. Cancer 2015, 136, E359.
4. Torre, L. A.; Bray, F.; Siegel, R. L.; Ferlay, J.; Lortet-Tieulent, J.; Jemal, A. CA
Cancer J. Clin. 2015, 65, 87.
Please cite this article in press as: Chen, Y.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.09.027