Synthesis of new derivatives of boehmeriasin A and their biological evaluation in liver cancer

Article abstract of DOI:10.1016/j.ejmech.2019.01.056

Two series of boehmeriasin A analogs have been synthesized in short and high yielding processes providing derivatives differing either in the alkaloid's pentacyclic scaffold or its peripheral substitution pattern. These series have enabled, for the first

Full text of DOI:10.1016/j.ejmech.2019.01.056

Accepted Manuscript
Synthesis of new derivatives of boehmeriasin A and their biological evaluation in liver
cancer
Ece Akhan Güzelcan, Ian R. Baxendale, Rengul Cetin-Atalay, Marcus Baumann
PII:
S0223-5234(19)30069-8
DOI:
https://doi.org/10.1016/j.ejmech.2019.01.056
EJMECH 11063
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 9 September 2018
Revised Date: 2 January 2019
Accepted Date: 23 January 2019
Please cite this article as: E.A. Güzelcan, I.R. Baxendale, R. Cetin-Atalay, M. Baumann, Synthesis of
new derivatives of boehmeriasin A and their biological evaluation in liver cancer, European Journal of
Medicinal Chemistry (2019), doi: https://doi.org/10.1016/j.ejmech.2019.01.056.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Synthesis of New Derivatives of Boehmeriasin A
and their Biological Evaluation in Liver Cancer
a
b,*
a,*
c,*
Ece Akhan Güzelcan, Ian R. Baxendale, Rengul Cetin-Atalay, Marcus Baumann,
a
Graduate School of Informatics, Cancer Systems Biology Laboratory, METU, 06800 Ankara,
Turkey
b
Department of Chemistry, University of Durham, South Road, DH1 3LE, Durham, UK
c
School of Chemistry, University College Dublin, Science Centre South, Belfield, Dublin 4,
Ireland
Key words: boehmeriasin A, natural product analogs, liver cancer, cell cycle analysis
1

ACCEPTED MANUSCRIPT
ABSTRACT
Two series of boehmeriasin A analogs have been synthesized in short and high yielding
processes providing derivatives differing either in the alkaloid’s pentacyclic scaffold or its
peripheral substitution pattern. These series have enabled, for the first time, comparative studies
into key biological properties revealing a new lead compound with exceptionally high activity
against liver cancer cell lines in the picomolar range for both well (Huh7, Hep3B and HepG2)
and poorly (Mahlavu, FOCUS and SNU475) differentiated cells. The cell death was
characterized as apoptosis by cytochrome-C release, PARP protein cleavage and SubG1 cell
cycle arrest. Subsequent testing associated apoptosis via oxidative stress with in situ formation of
reactive oxygen species (ROS) and altered phospho-protein levels. Compound 19 decreased Akt
protein phosphorylation which is crucially involved in liver cancer tumorigenesis. Given its
simple synthetic accessibility and intriguing biological properties this new lead compound could
address unmet challenges within liver cancer therapy.
2

ACCEPTED MANUSCRIPT
Introduction
Primary liver cancer (hepatocellular carcinoma, HCC), is the sixth most frequent cancer type
and the second highest cause of cancer-related death worldwide [1]. Viral infections with
hepatitis B or C, alcoholic injury, obesity and other factors inducing cirrhosis and chronic liver
disease are additional major risk contributors for HCC [2-4]. It is also expected that the rate of
liver cancer and associated deaths will increase in the coming years due to the global epidemic of
non-alcoholic fatty liver disease (NAFLD) [5]. Despite these well-known etiological factors,
liver cancer remains one of the most lethal types of cancer, due to the very limited therapeutic
options that offer significant clinical benefits [6]. Liver resection, transplantation and
chemoembolization are still the most relied upon options, however, not all the patients can meet
the criteria for these treatments [7]. Meanwhile, patients with advanced HCC suffer from the lack
of effective therapy. The first FDA approved drug Sorafenib (5, Figure 1), which is a multi-
kinase inhibitor, can only improve patients’ median survival for about 3 months [8]. Very
recently, the closely related multi-kinase inhibitor Regorafenib was used in patients as a second
line treatment. The mean survival period was 10 months compared to a placebo (~8 months) [9].
To address these challenges in the arena of liver cancer [10], we decided to investigate the
chemical synthesis and biological evaluation of new analogs of the known plant alkaloid
boehmeriasin A.
Boehmeriasin A (1, Figure 1) is a pentacyclic phenanthroquinolizidine alkaloid recently
isolated from the ethanolic extract of boehmeria siamensis Craib [11], a plant that has long been
used in south-east Asia for the treatment of rheumatism as well as skin diseases such as urticaria
(hives). Through extensive biological studies, it has been revealed that 1 possesses significant
3

ACCEPTED MANUSCRIPT
activity against several cancer cell lines commonly exceeding the potency of Taxol by at least
tenfold. The nanomolar cytotoxic activity of boehmeriasin A was established for various cancer
cells, originating from lung, colon, breast, prostate, kidney cancers, and leukemia [12]. In
addition to this anti-cancer activity boehmeriasin A also causes G1 cell cycle arrest, cellular
differentiation, affects cellular morphology and results in lipid droplet accumulation in breast
cancer cells [13]. It was furthermore found that boehmeriasin A shares this promising anti-cancer
activity with other members of the phenanthroquinolizidine and -pyrrolizidine alkaloids such as
tylophorine (2), antofine (3) and cryptopleurine (4) (Figure 1) [14-21]. Topoisomerases and
SIRT2 were proposed as potential biological targets of Boehmeriasin A [24].
Figure 1: Structures of racemic boehmeriasin A 1, related alkaloids (2-4) and sorafenib.
Driven by the promising bioactivity reports, many successful total syntheses of these intriguing
cytotoxic alkaloids have been disclosed [12,22-24]. Despite these efforts neither the molecular
target nor the mode of action of these important structures have been identified and detailed SAR
studies are lacking, which severely hamper development of further boehmeriasin A based leads.
To establish this missing key data, we wished to evaluate boehmeriasin A as well as several new
analogs in the context of human epithelial cancers especially in liver cancer.
Analog Synthesis: Upon analysis of the parent pentacyclic scaffold of boehmeriasin A (1) and
its closely related natural products (2-4) we decided to design and synthesize two series of target
4

ACCEPTED MANUSCRIPT
compounds to study their structure-activity relationship. The first series would retain the
oxygenation pattern of boehmeriasin A but introduce alterations in the connectivity of the fused
ring system, whereas the second series would conserve the original pentacyclic framework while
modifications to the embedded quinolizidine system would be made.
To create these two series, we decided to exploit a synthetic route to boehmeriasin A itself which
we had previously developed (Scheme 1) [24]. The route is based on an efficient 7-step sequence
in which a Perkin reaction is followed by an esterification to yield building block 8 that was then
converted into the phenanthrene unit 9 by a FeCl -mediated oxidation. The pendant ester
3
functionality was then reduced to an alcohol and activated as the corresponding chloride (10) to
allow coupling with pipecholic acid (11). Treatment of 12 with polyphosphoric acid (PPA) leads
to ring closure via a Friedel-Crafts acylation reaction and furnishes ketone 13 that upon full
reduction gives the desired natural product 1. This approach was deemed very attractive as it had
been shown to be robust and scalable and furthermore allows the introduction of the desirable
variations on different moieties of the parent structure of 1.
Scheme 1: Previous synthesis of racemic boehmeriasin A (1).
5

ACCEPTED MANUSCRIPT
The synthesis of the first analog series utilized intermediates 8 and 9 that were independently
elaborated by amide coupling with aminopropyl imidazole (14, Scheme 2). The aminopropyl
imidazole was selected as a hydrophylic appendage to mimic the quinolizidine substructure of 1
and to improve bioavailability. Furthermore, it was intended to modulate the planarity of the
phenanthrene system by omitting the central C-C bond in some of the analogs, which would
enable us to evaluate the likelihood of these entities acting via DNA interchelation.
MeO
MeO
MeO
1
. KOH, MeOH, 4 h, 50 o
. CDI, MeCN, 1 h, 50 o
then
C
LAH, THF,
rt, 2 h
2
C
N
N
OMe
H
H
N
N
N
H N
N
2
N
O
N
h, 50 oC;
14
O
MeO
2
MeO
MeO
MeO
MeO
OMe
OMe
OMe
8
15, 78%
17, 92%
MeO
1
. KOH, MeOH, 4 h, 50 o
C
N
LAH, THF,
rt, 2 h
N
2. CDI, MeCN, 1 h, 50 oC
OMe
H
N
H
N
N
N
then
H N
O
2
N
MeO
O
N
2 h, 50 oC;
MeO
14
MeO
OMe 9
16, 81%
18, 94%
OMe
OMe
Scheme 2: Synthesis of analogs 15-18 (Series 1).
Compounds 8 and 9 were hydrolyzed under basic conditions rendering the corresponding
carboxylic acid derivatives that were subsequently activated by CDI and coupled with
aminopropyl imidazole (14) to furnish the desired amides 15 and 16 in good yield. In addition,
both amides were treated with LAH (2.5 equiv.) to obtain the corresponding amines 17 and 18
allowing to establish the effect of having an amide versus an amine incorporated into these
structures, which could have implications due to altered H-bonding patterns.
With the first analog series in hand we turned our attention to a second series in which we
envisaged studying alterations on the pentacyclic framework of boehmerisain A. Specifically, we
6

ACCEPTED MANUSCRIPT
decided to explore manipulations of the carbonyl of 13 by means of its reduction to the
corresponding alcohol derivatives. The synthesis of the desired alcohols was readily
accomplished by reduction of 13 by LAH (2.0 equiv.) giving a diastereomeric mixture of alcohol
products 19a and 19b in a 1:2 ratio. Pleasingly, it was possible to separate this mixture by
preparative TLC to undertake characterization and subsequent biological evaluation of the
individual diastereomers. Additionally, single crystal X-ray diffraction experiments were used to
identify the major diastereomer 19b as the trans-alcohol product (Scheme 3).
Scheme 3: Synthesis of analogs 19a and 19b from ketone 13 (Series 2).
Together with boehmeriasin A (1) and its carbonyl derivative 13, these individual alcohol
derivatives 19a and 19b therefore represent the second series of analogs available for detailed
biological testing (Figure 2).
7

ACCEPTED MANUSCRIPT
Figure 2: Overview of both series of new boehmeriasin A analogs.
Biological Studies:
2
.1 Biological Activities of First Series of Boehmeriasin A Analogs
Commencing our studies with boehmeriasin A analogs 15-18 we analyzed their cellular toxicity
against different hepatocellular carcinoma cells, namely Huh7 and Mahlavu. As pointed out
earlier we wished to establish whether simplified structures possessing a conformationally open
ring system would still impart significant cytotoxicity, whilst improving solubility by
introduction of an imidazole appendage. However, to our surprise it was found that these analogs
displayed either very moderate activity (compounds 16 and 17 IC 10-54 µM) or no activity
5
0
(compounds 15 and 18, Table 1). Additionally, the presence of an amide or amine linkage as
well as the central biaryl bond appears to be irrelevant with regards to activity against these cell
lines. Whilst somewhat unexpected this finding establishes clearly that the pentacyclic
framework of the boehmeriasins is crucial for their sub-micromolar biological activity which had
8

ACCEPTED MANUSCRIPT
not been demonstrated before. This furthermore highlights the importance of the embedded
quinolizidine substructure and prompted us to evaluate analogs of our second series next.
Table 1: Cytotoxicity of compounds 15-18 evaluated in HCC cells.
IC50 Values (µM)
Compound
Huh7
Mahlavu
1
1
5
6
NI
NI
19.2
54
1
1
7
8
9.3
NI
13.8
NI
*
NI: No inhibition
2
.2 Cytotoxicity of Second Series of Boehmeriasin A Derivatives
To verify the anticipated biological activity of the parent boehmeriasin A (1) and its analogs (12,
3, 19) an initial screen against different epithelial cancer cells, namely liver (Huh7), breast
MCF7) and colon (HCT116) carcinoma cells was performed. The cytotoxicity of each
1
(
compound was assessed by sulforhodamine B (SRB) assay. IC50 values of the synthesized
compounds were calculated for each cell line following treatment for 72 hours (Table 2). The
high cytotoxicity found in all cancer cell types may address the possible role of these compounds
on common cancer activated pathways.
Table 2: Initial screening of compounds 1, 12, 13 and 19.
IC Values (µM)
5
0
HCT116
.7
Compound
1
Huh7
0.008
0.002
NI
MCF7
0.005
0.001
NI
0
<
0.001
NI
19
12
13
8
.1
12
7
9

ACCEPTED MANUSCRIPT
*
NI: No inhibition
In general, boehmeriasin A (1) and its hydroxy analog 19 (as 1:2 mixture of diastereomers) were
identified as promising anti-cancer compounds due to their low nanomolar IC values. Although
5
0
ketone derivative 13 showed moderate cytotoxicity, when compared to 1 and 19 its bioactivity
was low despite the only minor alteration of its structure. Additionally, the carboxylate species
1
2 had no cytotoxic bioactivity against these cancer cells, possibly because it does not easily
permeate into cells due to its ionic nature. Furthermore, this result is in line with the loss of
activity seen for the other ‘open-structured’ analogs 15-18.
Considering the specific lack of any HCC targeted chemotherapeutic agents and based upon our
specific aim in this study, we furthermore tested these compounds (1, 12, 13 and 19) against
various liver cancer cell lines and normal transformed breast epithelial cells since there are no
transformed normal liver hepatocytes available. Therefore, we assessed the IC values of these
5
0
compounds in a HCC cell panel (Huh7, Mahlavu, SNU475, FOCUS, Hep3B and HepG2) (Table
).
3
Table 3: Cytotoxicity of compounds in HCC cell panel.
IC50 Values (µM)
Mahlavu SNU475 FOCUS MCF12A MCF10A
Huh7
Hep3B HepG2
Compound
1
0.008
0.002
NI
0.016
0.06
0.016
0.01
NI
0.014
0.015
NI
0.02
0.012
NI
0.07
2<
0.14
0.2
1
1
1
9
2
3
0.017 2.7 ± 0.9
NI
45
NI
NI
NI
NT
NT
12
35
19.1
5
NI
*
NI: No inhibition, NT: Not tested
10

ACCEPTED MANUSCRIPT
Similar to our results presented in Table 2, compounds 1 and 19 were found to be significantly
cytotoxic in all cell types of HCCs. Crucially, 1 and 19 displayed nanomolar IC values not only
5
0
for well differentiated HCC cells such as Huh7, Hep3B and HepG2 but also for poorly
differentiated and more aggressive cells as tested here, namely, Mahlavu, FOCUS and SNU475.
Furthermore, normal transformed epithelial cells were less sensitive to compounds 1 and 19. Our
data showed significant cytotoxic effects of compounds in the nanomolar range in cancer cells
while the cytotoxic effect on normal transformed cells was diminished (Table 3). The higher IC50
found in normal like HepG2 cells further supported the cancer-specific activity of these
compounds. Therefore, 1 and 19 can be considered potential anti-cancer drug candidates in liver
cancer cells. On the other hand, 13 displayed limited cytotoxic activity and 12 had no activity in
any type of HCC cells (data not shown).
2
.3. Real-Time Assessment of Hepatocellular Carcinoma Cell Growth upon Treatment with
Compounds 1 and 19
A real-time cell electronic sensing (xCELLigence RTCA analysis) assay was used to evaluate the
bioactivities of Boehmeriasin A and its analog 19 on well differentiated Huh7 cells and poorly
differentiated aggressive mesenchymal Mahlavu cells (25). According to the data obtained from
this analysis, the time dependent IC50 values were calculated (Table 4). Compounds 1 and 19
caused significant growth inhibition in both cells. Drug resistant Mahlavu cells were prominently
more sensitive to 19 (Figure 3). The real-time growth inhibition pattern suggested a cell cycle
arrest in cells treated with 1 and 19, as DMSO treated cells retain their proliferation profile until
they become confluent.
11

ACCEPTED MANUSCRIPT
The curvature of the real time cell death data and the measured IC50 values imply that the cell
death associated with these compounds could be apoptosis. The differences of IC50 values
between SRB end point colorimetric assay and the real time cell growth assays can be due to the
evaluation of distinct cellular elements. While SRB detects the total protein levels, RTCA is
dependent on the cell surface attachment properties of each different cell. During apoptosis there
is a high destruction of cellular proteins which leads to less colorimetric data collection.
Therefore, it is common to observe IC50 differences between cell death detection techniques.
Hence in this study, we established the bioactivities of the new boehmeriasin A analogs using
two complementary techniques.
Table 4: Time dependent IC50 values of compounds 1 and 19.
IC Values (µM)
5
0
Compound
1
Huh7
48h
Mahlavu
48h
2
4h
72h
24h
72h
1.4 µM
0.8 µM
0.7 µM
0.5 µM
0.7 µM
0.4 µM
0.2 µM
0.01 µM
0.4 µM
0.004 µM
0.6 µM
0.01 µM
19
12

ACCEPTED MANUSCRIPT
Figure 3. xCELLigence RTCA analysis. Cell growth index with various concentrations of the
compounds 1, 19 and their corresponding DMSO solvent controls on Huh7 and Mahlavu cell lines.
2
.4. Characterization of Cell Death Mechanism Induced by Boehmeriasin Derivatives
The cell death mechanism of the cells treated with boehmeriasin A or its derivatives was initially
analyzed by observing the release of cytochrome-C which was examined via fluorescence
microscopy (Figure 4A). Camptothecin (CPT) was used as positive control for cytochrome-C
release, showing that cytochrome-C levels in cells treated with 1 and 19 were comparable to the
positive control. A significant increase in the release of cytochrome-C in cells treated with 1 and
1
9 indicates apoptotic cell death.
The activation of apoptotic pathways by boehmeriasin A derivatives was further studied by
evaluating the status of one of the most well-known apoptotic proteins, PARP. Here a significant
increase in PARP cleavage in the cells treated with 19 was identified and became significant
13

ACCEPTED MANUSCRIPT
within 72 h in cells. Compared to boehmeriasin A itself, the hydroxy-derivative 19 displayed a
more significant effect on PARP cleavage that is paralleled by its increased cytotoxic effect on
liver cancer cells (Figure 4B).
During cellular apoptosis, many biochemical changes occur. Once the apoptosis mechanism is
triggered DNA fragmentation, phosphatidylserine flipping to the outer surface of the cells and
some morphological changes such as smaller cell size, chromatin condensation and formation of
apoptotic bodies eventually occur. Hoechst is a fluorescence dye that attaches to DNA of both
dead and living cells and makes their nucleus visible. The intensity of Hoechst staining
(chromatin condensation), structure of nucleus and formation of apoptotic bodies are the initial
indicators of cellular apoptosis. The Hoechst staining of 1 and 19 treated cells showed many
indicators of apoptosis (chromatin condensation and apoptotic body formation). The effect was
increased at 72 h for all samples.
Moreover, the apoptosis of cells treated with 1 and 19 was further demonstrated with Annexin V
assay via flow cytometry. Normally, phosphatidylserine (PS) residues are hidden in the
membrane of living cells. Upon apoptosis, PS flips from the inner cell to the outer surface which
then can be detected by Annexin V. This assay showed the sharp increase in apoptosis of cells
treated with 1 and 19 for epithelial Huh7 cells at 72 h. The drug response was similar in Mahlavu
cells at 72 h but 1 showed less dramatic effect on these cells. The Annexin V results thus
supported the previous findings.
A
14

ACCEPTED MANUSCRIPT
B
C
15

ACCEPTED MANUSCRIPT
D)
16

ACCEPTED MANUSCRIPT
Figure 4: Analysis of apoptotic processes in cells treated with Boehmeriasin A derivatives. (A) The
evaluation of Cytochrome C (green) release in Huh7 and Mahlavu cells treated with 1 µM 1 and 19. The
nuclei of cells were stained with Hoechst (blue). (B) PARP cleavage in Huh7 and Mahlavu cells treated
with 1 µM 1 and 19 for 24 and 72 h. DMSO is used as solvent control. (C) The nucleus staining with
Hoechst dye (blue) to identify the cellular apoptosis. White arrows indicate the apoptotic cells. (D) The
Annexin V analysis of Huh7 and Mahlavu cells treated with 1 µM 1 and 19 for 24, 48 and 72 h.
2.5. Cell Cycle Analysis of HCC Cells treated with Boehmeriasin Derivatives
To identify the effect of the boehmeriasin derivatives on cell cycle progress, flow cytometry
analysis was performed using propidium iodide (PI) staining in Huh7 and Mahlavu liver cancer
cells. Upon treatment with compounds 1 and 19, increasing levels of apoptotic cells along with
SubG1 arrest were observed. The most significant increase was observed at 72 h. Derivative 19
resulted in the higher number of cells that accumulated at SubG1 phase. Additionally, HCC cells
treated with 1 also led to an increase in SubG1, however, this effect became more notable after
7
2 h (Figure 5 A, B). In addition, the effect of boehmeriasin A derivatives on the cell cycle was
further investigated in Huh7 and Mahlavu cells by western blot analysis of the principal proteins
which are involved in this process. Compounds 1 and 19 down-regulated Cyclin B1 (CDK1
regulator and mitotic initiator in the late G2 phase) and its companion CDK1 (essential for G1/S
and G2/M phase transitions) levels in Huh7 cells. Treatment with 1 and 19 also reduced the
CDK2 (G1 to S phase transition) and its regulator Cyclin E level, which became more significant
at 72 h. Combined, these results indicate that boehmeriasin A derivatives induce cell cycle arrest
at G2/M and SubG1 phases which is then followed by apoptosis of the cells (Figure 5A).
17

ACCEPTED MANUSCRIPT
The effect of these compounds on cell cycle protein expression is more pronounced on Huh7
cells (Figure 5B). Mahlavu cells are reported as more resistant to small molecule inhibitors due
to PTEN deletion in PI3K/Akt signaling pathway [25]. However, the effect of compounds 1 and
1
9 can still be observed even in these cells. Furthermore, the active phosphorylated form of Akt
protein is significantly down-regulated with compound 19 at 72 h (Figure 5B).
A
B
18

ACCEPTED MANUSCRIPT
Figure 5: The assessment of the cell cycle status of cells treated with boehmeriasin A derivatives. (A)
The time dependent cell cycle analysis of treated Huh7 and Mahlavu cells with flow cytometry. (B)
Comparative analysis of cell cycle protein expression in the presence of 1 and 19. DMSO is used as the
solvent control.
2
.6. Release of Reactive Oxygen Species and Stress Mechanism Induced by Boehmeriasin
Derivatives
Reactive oxygen species (ROS), depending on their dose, can alter cellular pathways and
promote cell cycle arrest and apoptosis in liver cancer cells [26]. The released ROS can be
visualized using the 2,7-dichlorofluorescein diacetate (DCFH-DA) assay. The presence of
fluorescence stain, which is detected via fluorescence microscopy, indicates ROS activity and
thus cellular stress. Evaluating the presence of ROS in liver cancer cells that have been treated
with boehmeriasin A and its derivatives showed that the ROS activity in cells incubated with 1
and 19 was significantly increased and became comparable to the positive control. The ROS
activity increased with the time of incubation of 1 and 19 (Figure 6A).
19

ACCEPTED MANUSCRIPT
In addition, these Boehmeriasin derivatives induced oxidative stress, that was assessed by
evaluating the phosphorylation status of the proteins involved in the ROS pathway via western
blot [25]. It was found that Mahlavu and Huh7 cells treated with 1 and 19 led to a decrease of the
phosphorylated Ser-966-ASK1 levels which is associated with cellular oxidative stress, though
the effect on Mahlavu is being more prominent [25] (Figure 6B). Mahlavu cells are characterized
by a hyperactive PI3K/AKT pathway due to PTEN protein deletion and therefore we observe
differential downstream P-SAPK/JNK protein levels. The stress protein SAPK/JNK
phosphorylation was increased in Huh7 cells treated with both 1 and 19. In Mahlavu cells the P-
SAPK/JNK levels were differentially altered with 1 and 19 treatments in parallel to AKT
phosphorylation status. The compound 19 decreases the p-AKT levels which is the active from
of this protein required for cell survival. Our data indicates that the novel compound 19 gains its
intriguing bioactivity through PI3K/AKT pathway modulation which is strategically involved in
liver cancer tumorigenesis.
A
20

ACCEPTED MANUSCRIPT
B
Figure 6: The analysis of cellular stress mechanism. (A) The evaluation of ROS activity by DCHF-DA
staining in Huh7 cells treated with boehmeriasin derivatives. DCFH-DA is a cell permeable dye, and,
after uptake, it is cleaved by intracellular esterases to 2,7-dichlorodihydrofluorescin (DCFH2), which is
trapped within the cells, and is oxidized to fluorescent 2,7-dichlorofluorescein (DCF) in the presence of
ROS. The cells that grow in serum free medium were shown to induce ROS activity, therefore, we used
serum free culture media as positive control. (B) The western blot analysis of the stress related proteins of
boehmeriasin derivatives treated Huh7 and Mahlavu cells.
2
.7. The Bioactivity of Individual Hydroxy-Analogs 19a and 19b on Epithelial Cancer and
Normal Epithelial Immortalized Cells
Having determined that 19 (originally tested as a 1:2 mixture of diastereomers 19a and 19b)
possessed the highest bioactivity, we decided to further study its individual diastereomers to
establish whether a preference of the relative stereochemistry manifests in differentiated activity.
The separation of this mixture of diastereomers was accomplished by preparative thin layer
chromatography yielding 19a as the minor (lower R ) and 19b as the major (higher R )
f
f
diastereomer in a 1:2 ratio. The individual hydroxy-analogs 19a and 19b were initially screened
in human liver, colon and breast cancer cells along with immortalized transformed normal breast
21

ACCEPTED MANUSCRIPT
epithelial cells (Table 5). The data was verified with SRB results confirming not only that both
hydroxy-analogs were highly active anti-cancer compounds but also 19b displayed superior
activity in the nanomolar range.
Table 5: Cytotoxicity of compounds 19a and 19b evaluated in different human cancer cells and normal
transformed epithelial cells.
IC50 Values (µM)
Huh7
0.306
Mahlavu
.05>
0.05>
HCT116
0.043
MCF7
MCF10A
MCF12A
0.05>
1
1
9a (minor)
9b (major)
0.016
0.8
0
0.017
0.007
0.021
0.14
0.1
Correlating this intriguing finding with the single crystal X-ray structure secured for the major
diastereomer 19b (see Scheme 3) demonstrates that the highest activity results from the
diastereomer having the hydroxy substituent trans to the methine proton and syn to the ring
nitrogen. Although, the reason of this is not entirely clear at this point this finding might point to
a specific binding mode of 19b within its molecular target.
2
.8. The Induced Apoptosis of Individual Hydroxy-Analogs 19a and 19b
After we showed the cytotoxicity of 19a and 19b in nanomolar range, we further studied the cell
death mechanism of these compounds in Huh7 and Mahlavu cells. Initially, treated cells were
stained with Hoechst dye and their apoptotic morphologies were analyzed. The Hoechst staining
of 19a and 19b treated cells showed a high number of apoptotic cells which possessed condensed
chromatin and apoptotic bodies, compared to the DMSO control. The effect was increased for
1
9b treated Huh7 cells which showed more apoptotic features compared to 19a at 72 h. The
apoptotic effect of compounds was similar in Mahlavu cells. A reduced number of remaining
22

ACCEPTED MANUSCRIPT
cells was observed after 19b treatment in Mahlavu cells, therefore, the percentage of apoptotic
cells had increased (Figure 7A).
Furthermore, the Annexin V assay was utilized to analyze the apoptotic effect of 19a and 19b in
more detail. Annexin V results showed the significant increase of apoptotic cells in 19b treated
cells, especially at 72 h. Similar to previous results, this assay further demonstrated the major
compound 19b had a severe effect on cellular apoptosis when compared to 19a (Figure 7B).
Together these results supported the findings of SRB assay which revealed their high
cytotoxicity.
A
23

ACCEPTED MANUSCRIPT
B
Figure 7: The analysis of apoptosis. (A) The nucleus staining with Hoechst dye (blue) to identify the
cellular apoptosis. White arrows indicate the apoptotic cells. (B) The Annexin V analysis of Huh7 and
Mahlavu cells treated with 1 µM 19a and 19b for 24, 48 and 72 h.
In conclusion, we have successfully prepared two sets of new boehmeriasin A analogs featuring
either a simplified structure or a complete pentacyclic framework as found in the parent alkaloid.
Whilst the former was found to be significantly less active (IC > 10 µM) than boehmeriasin A,
5
0
the latter series displayed superb anti-cancer activity across a range of cell lines. Minute
structural changes from a ketone (13, H-bond acceptor) to a secondary alcohol (19, H-bond
donor) have resulted in a 10,000-fold increase in activity, which is more pronounced in the trans-
diastereomer (19b) and surpasses the activity of boehmeriasin A itself. Subsequent in-depth
biological studies that were performed for the first time in HCC cells (well differentiated Huh7
and poorly differentiated aggressive Mahlavu cells) and have revealed a broad spectrum of
activities including cell cycle arrest in the sub-G1 stadium, the generation of reactive oxygen
species as well as the activation of apoptotic pathways through PARP cleavage. Moreover,
Annexin V and Hoechst staining results further demonstrated the induction of apoptosis in drug
24

ACCEPTED MANUSCRIPT
treated cells. Due to the straightforward synthetic accessibility and the intriguing biological
properties of our new lead compound 19b we foresee its potential for future developments to
address unmet challenges within liver cancer therapy.
Experimental section
3
.1. Synthesis Procedures for New Compounds
Purity of all compounds was established via HPLC using an Agilent 1100 instrument with a
Zorbax C18 column (4.6 x 100 mm) and UV-detection at 220 nm. An isocratic eluent system
was used (60:40 MeCN:water, containing 0.1% TFA; 1 mL/min). All compounds displayed
purities above 95% unless stated otherwise.
Rac-Boehmeriasin A, 1 and 3,6,7-trimethoxy-12,13,14,14a-tetrahydro-9
pyrido[1,2-b]isoquinolin-15(11 )-one, 13 were prepared as reported previously [24] and
displayed purities of 97% and 96% respectively by HP-LC.
)- -(3-(1 -Imidazol-1-yl)propyl)-2-(3,4-dimethoxyphenyl)-3-(4-methoxyphenyl)acryl-
amide, 15: To a solution of methyl (E)-2-(3,4-dimethoxyphenyl)-3-(4-methoxyphenyl)acrylate
8, 1.0 g, 3.05 mmol) in methanol (10 mL, 0.3 M) was added an aqueous solution of KOH (5.0
H-dibenzo[f,h]-
H
(E
N
H
(
M, 5 mL, 8.3 equiv. KOH). The resulting mixture was stirred at 50 °C for 4 h when tlc analysis
indicated full conversion of 8. After neutralizing with aqueous HCl (1 M) and extraction
(DCM/water, 2 x 50 mL) the organic layers were combined, dried over anhydrous sodium
sulfate, filtered and evaporated to dryness under reduced pressure. The solid residue was
dissolved in MeCN (15 mL) and combined with carbonyldiimidazole (CDI, 500 mg, 3.09 mmol,
1
equiv.). After heating this mixture for 1 h at 50 °C aminopropylimidazole (275 mg, 3.0 mmol,
25

ACCEPTED MANUSCRIPT
1
.0 equiv.) was added and the mixture was stirred at 50 °C for a further 2 h. After evaporation of
the volatiles the crude product was extracted (DCM/water, 2 x 20 mL) and isolated as an off-
white solid (1.0 g, 2.4 mmol, 78 %) after drying over anhydrous sodium sulfate, filtration,
evaporation and trituration from cold ethyl acetate.
1
H-NMR (400 MHz, CDCl
H), 6.91 – 6.87 (m, 3H), 6.83 (d, J = 1.3 Hz, 1H), 6.72 (dd, J = 8.2, 1.9 Hz, 1H), 6.64 (d, J = 1.9
Hz, 1H), 6.60 (d, J = 8.9 Hz, 2H), 5.73 (t, J = 6.1 Hz, 1H), 3.89 (t, J = 7.0 Hz, 2H), 3.86 (s, 3H),
3
): δ/ppm 7.68 (s, 1H), 7.34 (d, J = 1.2 Hz, 1H), 6.92 (d, J = 1.1 Hz,
1
1
3
3
.72 (s, 3H), 3.66 (s, 3H), 3.23 (app q, J = 6.7 Hz, 2H), 1.90 (app p, J = 6.9 Hz, 2H). C-NMR
(
100 MHz, CDCl ): δ/ppm 167.9 (C), 159.9 (C), 149.9 (C), 149.1 (C), 137.0 (CH), 136.8 (CH),
3
1
32.0 (2CH), 131.3 (C), 129.5 (CH), 128.5 (C), 127.4 (C), 122.0 (CH), 118.8 (CH), 113.7
(
2CH), 112.6 (CH), 112.2 (CH), 56.0 (CH ), 55.9 (CH ), 55.1 (CH ), 44.5 (CH ), 37.2 (CH ),
3
3
3
2
2
-
1
3
2
1.3 (CH ). IR (neat, ν/cm ): 3375 (w), 2981 (s), 1655 (m), 1602 (m), 1509 (s), 1446 (m), 1252
(
s), 1176 (m), 1139 (s), 1080 (m), 1019 (m), 959 (m), 833 (s), 738 (m), 620 (m). HR-MS (TOF-
AP+) calculated for C24 422.2080, found 422.2079. HP-LC (220 nm) Rt = 1.72 min
97% purity).
-(3-(1 -Imidazol-1-yl)propyl)-3,6,7-trimethoxyphenanthrene-9-carboxamide,
28 3 4
H N O
(
N
H
16:
In
analogy to the preparation of 15, methyl 3,6,7-trimethoxyphenanthrene-9-carboxylate (9, 700
mg, 2.15 mmol) was converted into the target compound 16 (730 mg, 1.74 mmol, 81%) which
was obtained as a yellow waxy solid.
1
H-NMR (400 MHz, CDCl
d, J = 8.8 Hz, 1H), 7.50 (s, 1H), 7.49 (s, 1H), 7.10 (dd, J = 8.7, 2.4 Hz, 1H), 7.06 (t, J = 5.9, 5.9
Hz, 1H), 6.99 (t, J = 1.1 Hz, 1H), 6.93 (t, J = 1.3 Hz, 1H), 4.03 (s, 3H), 4.01 (t, J = 6.8 Hz, 2H),
3
): δ/ppm 7.71 (s, 1H), 7.65 (s, 1H), 7.63 (d, J = 2.5 Hz, 1H), 7.61
(
1
3
3
.97 (s, 3H), 3.92 (s, 3H), 3.43 (app q, J = 6.8 Hz, 2H), 2.07 (p, J = 6.8 Hz, 2H). C-NMR (100
26

ACCEPTED MANUSCRIPT
MHz, CDCl
3
): δ/ppm 170.5 (C), 159.2 (C), 149.6 (C), 149.1 (C), 137.2 (CH), 131.9 (C), 130.7
(CH), 129.4 (CH), 129.1 (C), 124.8 (C), 124.7 (CH), 124.2 (C), 124.1 (C), 118.9 (CH), 115.9
(
CH), 106.4 (CH), 103.7 (CH), 103.2 (CH), 55.9 (CH ), 55.8, (CH ), 55.5 (CH ), 44.8 (CH ),
3
3
3
2
-
1
3
1
7.2 (CH ), 31.2 (CH ). IR (neat,
ν
/cm ): 3259 (broad), 2939 (m), 1618 (m), 1522 (s), 1508 (s),
2
2
473 (s), 1270 (s), 1228 (s), 1206 (s), 1158 (m), 1033 (m), 832 (w), 734 (w). HR-MS (TOF
AP+) calculated for C H N O 420.1923, found 420.1921. HP-LC (220 nm) Rt = 1.64 min
2
4
26
3
4
(
98% purity).
)- -(3-(1
en-1-amine, 17: To a solution of (E)-N-(3-(1H-imidazol-1-yl)propyl)-2-(3,4-dimethoxyphenyl)-
-(4-methoxyphenyl)acryl amide (15, 500 mg, 1.19 mmol) in dry THF (10 mL, 0.12 M) was
(E
N
H-Imidazol-1-yl)propyl)-2-(3,4-dimethoxyphenyl)-3-(4-methoxyphenyl)prop-2-
3
added lithium aluminium hydride (LAH, 200 mg, 5.26 mmol) in portions at room temperature.
After 2 hours the reaction mixture was carefully quenched by addition of ethyl acetate (1 mL)
and water (3 mL). The resulting emulsion was filtered through a pad of celite (ca. 10 g) and
washed with ethyl acetate (30 mL). The resulting pale-yellow solution was evaporated to dryness
yielding a yellow oil that was purified by column chromatography on silica (20-40%
EtOAc/hexanes). The desired product 17 was obtained as a colorless oil that solidified upon
standing (445 mg, 1.09 mmol, 92%).
1
H-NMR (400 MHz, CDCl ): δ/ppm 7.40 (s, 1H), 7.33 – 7.29 (m, 2H), 7.06 – 7.03 (m, 2H),
3
7
3
.02 (s, 1H), 6.92 – 6.86 (m, 3H), 6.82 (s, 1H), 6.77 (s, 1H), 3.96 (t, J = 6.9 Hz, 2H), 3.91 (s,
H), 3.89 (s, 3H), 3.82 (s, 3H), 3.75 (s, 2H), 2.54 (t, J = 6.6 Hz, 2H), 1.82 (p, J = 6.8 Hz, 2H).
1
3
3
C-NMR (100 MHz, CDCl ): δ/ppm 158.6 (C), 149.0 (C), 148.7 (C), 138.3 (C), 137.2 (CH),
1
1
34.5 (C), 130.0 (2CH+C), 129.4 (CH), 129.1 (CH), 118.9 (CH), 118.7 (CH), 113.8 (2CH),
11.2 (CH), 109.9 (CH), 56.0 (2CH ), 55.3 (CH ), 48.3 (CH ), 45.8 (CH ), 44.6 (CH ), 31.1
3
3
2
2
2
27

ACCEPTED MANUSCRIPT
-
1
(
CH
176 (m), 1143 (m), 1024 (s), 812 (m), 765 (m), 731 (s), 683 (m), 531 (m). HR-MS (TOF-AP+)
calculated for C H N O 408.2287, found 408.2285. HP-LC (220 nm) Rt = 1.52 min (96%
2
). IR (neat, ν/cm ): 2935 (w), 2836 (w), 1670 (m), 1604 (m), 1509 (s), 1463 (m), 1246 (s),
1
2
4
30
3
3
purity).
-(1 -Imidazol-1-yl)-N-((3,6,7-trimethoxyphenanthren-9-yl)methyl)propan-1-amine, 18: In
3
H
analogy to the preparation of compound 17, N-(3-(1H-imidazol-1-yl)propyl)-3,6,7-
trimethoxyphenanthrene-9-carboxamide (16, 500 mg, 1.19 mmol) was converted to the target
compound that was isolated as a waxy solid (454 mg, 1.12 mmol, 94%).
1
H-NMR (400 MHz, CDCl ): δ/ppm 7.80 (s, 1H), 7.75 (d, J = 2.4 Hz, 1H), 7.68 (d, J = 8.7 Hz,
3
1
H), 7.40-7.45 (m, 2H), 7.36 (s, 1H), 7.13 (dd, J = 8.7, 2.4 Hz, 1H), 6.96 (t, J = 1.1 Hz, 1H), 6.78
(
(
(
t, J = 1.1 Hz, 1H), 4.04 (s, 2H), 4.02 (s, 3H), 3.97 (s, 3H), 3.94 (s, 3H), 3.91-3.95 (m, 2H), 2.64
1
3
t, J = 6.7 Hz, 2H), 1.87 (app p, J = 6.7 Hz, 2H). C-NMR (100 MHz, CDCl
3
): δ/ppm 158.1
C), 149.2 (C), 148.7 (C), 137.1 (CH), 130.8 (C), 130.5 (C), 129.9 (CH), 129.3 (CH), 126.2 (C),
1
5
25.7 (C), 124.9 (C), 124.6 (CH), 118.9 (CH), 115.5 (CH), 104.8 (CH), 103.8 (CH), 103.7 (CH),
6.0 (CH ), 55.9 (CH ), 55.5 (CH ), 52.5 (CH ), 46.3 (CH ), 44.7 (CH ), 31.5 (CH ). IR (neat,
3
3
3
2
2
2
2
-
1
ν
/cm ): 3002 (w), 2935 (w), 2832 (w), 1609 (s), 1523 (s), 1509 (s), 1467 (s), 1269 (s), 1205 (s),
161 (m), 1070 (m), 1032 (m), 834 (w), 731 (m). HR-MS (TOF AP+) calculated for C H N O
3
1
4
3
1
2
4
28
3
06.2131, found 406.2137. HP-LC (220 nm) Rt = 1.46 min (95% purity).
,6,7-Trimethoxy-11,12,13,14,14a,15-hexahydro-9 -dibenzo[ ]pyrido[1,2-
5-ol (19a and 19b): To a solution of 3,6,7-trimethoxy-12,13,14,14a-tetrahydro-9H-
H
f,
h
b]isoquinolin-
dibenzo[f,h]pyrido[1,2-b]isoquinolin-15(11H)-one (13, 250 mg, 0.64 mmol) in dry THF (10 mL,
.06 M) was slowly added lithium aluminium hydride (73 mg, 1.92 mmol, 3.0 equiv.). After
stirring at room temperature for 1 h the mixture was carefully quenched by addition of ethyl
0
28

ACCEPTED MANUSCRIPT
acetate (1 mL) and water (1 mL). Filtration over a pad of celite (ca. 5 g) and elution with ethyl
acetate (20 mL) yielded after evaporation the crude mixture of 19a and 19b (213 mg, 0.54 mmol,
8
5%) in a 1:2 ratio as evidenced by 1H-NMR. The separation of this mixture was accomplished
by preparative thin layer chromatography (2 runs) using EtOAc/hexanes (30:70) as solvent
system.
Rac-(14aR,15
S
)-3,6,7-Trimethoxy-11,12,13,14,14
]isoquinolin-15-ol, 19a:
): /ppm 8.30 (d, J = 9.1 Hz, 1H), 7.82 (d, J = 2.6 Hz, 1H), 7.80 (s, 1H), 7.21
dd, J = 9.1, 2.5 Hz, 1H), 6.69 (s, 1H), 4.99 (d, J = 5.6 Hz, 1H), 4.10 (s, 3H), 4.01 (s, 3H), 3.93 (d, J =
5.0 Hz, 1H), 3.89 (s, 3H), 3.03 (br s, 1H), 2.44 (br s, 2H), 2.10 (s, 1H), 1.83 (d, J = 12.7 Hz, 1H), 1.75
d, J = 14.3 Hz, 1H), 1.64 (d, J = 13.8 Hz, 1H), 1.41 (q, J = 12.7 Hz, 1H). 3 Resonances were not
a,16-hexahydro-9H-dibenzo[f,h]pyrido-
[1,2-b
1
H-NMR (700 MHz, CDCl
3
δ
(
1
(
1
3
observed due to broadening effects. C-NMR (176 MHz, CDCl ):
δ
/ppm 157.7 (C), 149.4 (C), 148.9
C), 131.1 (C), 127.5 (CH), 124.1 (C), 124.0 (C), 123.8 (C), 114.8 (CH), 104.6 (CH), 103.7 (CH), 103.2
CH), 70.3 (CH), 65.5 (CH), 55.9 (CH ), 55.8 (CH ), 55.5 (CH ), 54.5 (CH ), 23.3 (CH ). 5 Resonances
/cm ): 3300 (broad), 2936 (m), 1732 (w), 1615 (m), 1514 (s),
471 (s), 1426 (m), 1262 (s), 1235 (m), 1206 (s), 1172 (m), 1147 (m), 1039 (s), 914 (m), 840 (m), 731
m). HR-MS (ESI-TOF+) calculated for C24 394.2018, found 394.2028. HP-LC (220 nm) Rt =
3
(
(
3
3
3
2
2
-
1
were not observed (2 C, 3 CH ). IR (neat,
ν
2
1
(
H28NO
4
1
.76 min (95% purity).
Rac-(14aR,15R)-3,6,7-Trimethoxy-11,12,13,14,14a,16-hexahydro-9H-dibenzo[f,h]pyrido-
[
1,2-
b
]isoquinolin-15-ol, 19b:
/ppm 8.35 (d, J = 9.0 Hz, 1H), 7.68 (d, J = 2.6 Hz, 1H), 7.50 (s, 1H),
1
H-NMR (600 MHz, CDCl ):
δ
3
7
3
1
.21 (dd, J = 9.0, 2.5 Hz, 1H), 5.91 (s, 1H), 4.59 (d, J = 2.5 Hz, 1H), 4.07 (s, 3H), 4.02 (s, 3H), 3.71 (s,
H), 2.84-3.00 (m, 2H), 2.80 (d, J = 14.9 Hz, 1H), 2.24 (qd, J = 13.3, 3.8 Hz, 1H), 2.07 (d, J = 11.2 Hz,
H), 1.90-2.00 (m, 2H), 1.82 (q, J = 13.1 Hz, 1H), 1.73 (d, J = 13.1, Hz, 1H), 1.67 (d, J = 12.2 Hz, 1H),
29