Synthesis of novel S-linked dihydroartemisinin derivatives and evaluation of their anticancer activity

Article abstract of DOI:10.1016/j.ejmech.2019.06.018

We report herein the synthesis and anticancer activity of a set of novel S-linked artemisinins bearing an aliphatic/aromatic/heterocyclic nucleus as a substituent on the sulfur. The compounds were prepared from artemisinin via its lactol-form by an acid-catalyzed condensation of the desired thiol with the lactol. Both the C-10-α- and β-configured thiol ethers were synthesized with a view to making them available for the anticancer activity evaluation using a variety of cell lines. The results show that many of the synthetic derivatives studied possessed good potential as anticancer agents. In order to draw more information on the origin of the anticancer activity, one of the compounds (9a), that showed a broad-spectrum activity in terms of reducing the viability of most of the cell lines studied, in particular proven to be most effective against Prostate (PC-3) cells, was studied in detail to find the underlying mechanism of its action by MTT assay, immunoblotting, flow cytometry and microscopy. Pretreatment of the PC-3 cells with N-acetyl cysteine affected the efficacy of 9a, suggesting the role of reactive oxygen species in reducing their viability. Cell cycle analysis showed increase in G1 phase that was indicative of G1 cell cycle arrest. Wound healing assay revealed anti-migratory effect of 9a Quantitative PCR and western blot analysis showed changes in the gene expression of PCNA, E2F1, Pin1, cyclinD1, phospho-c-jun, c-Myc, eIF4E and other genes involved in proliferation and maintaining the transformed phenotype of prostate cancer cells. Here we report the anti–proliferative property of 9a with a vital and potent target(s) in prostate cancer cells with one of such targets being Pin1 belonging to the parvulin family of PPIases. The results suggest that 9a could be a promising agent in combating prostate cancer.

Full text of DOI:10.1016/j.ejmech.2019.06.018

Accepted Manuscript
Synthesis of novel S-linked dihydroartemisinin derivatives and evaluation of their
anticancer activity
Rajesh Gour, Faiz Ahmad, Santosh Kumar Prajapati, Santosh Kumar Giri, Shibendra
Kumar Lal Karna, K.P. Ravindranathan Kartha, Yuba Raj Pokharel
PII:
S0223-5234(19)30537-9
DOI:
https://doi.org/10.1016/j.ejmech.2019.06.018
EJMECH 11421
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 12 March 2019
Revised Date: 24 May 2019
Accepted Date: 5 June 2019
Please cite this article as: R. Gour, F. Ahmad, S.K. Prajapati, S.K. Giri, S.K. Lal Karna, K.P.R. Kartha,
Y.R. Pokharel, Synthesis of novel S-linked dihydroartemisinin derivatives and evaluation of their
anticancer activity, European Journal of Medicinal Chemistry (2019), doi: https://doi.org/10.1016/
j.ejmech.2019.06.018.
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
Graphical Abstract
Synthesis of novel S-linked dihydroartemisinin derivatives and
evaluation of their anticancer activity
a
b
a
a
Rajesh Gour, Faiz Ahmad, Santosh Kumar Prajapati, Santosh Kumar Giri, Shibendra
b
b*
a*
Kumar Lal Karna, Yuba Raj Pokharel, K. P. Ravindranathan Kartha

ACCEPTED MANUSCRIPT
Synthesis of novel S-linked dihydroartemisinin derivatives and
evaluation of their anticancer activity
#
a #
b
a
a
Rajesh Gour, Faiz Ahmad, Santosh Kumar Prajapati, Santosh Kumar Giri,
b
a*
b*
Shibendra Kumar Lal Karna, K. P. Ravindranathan Kartha, Yuba Raj Pokharel
a
Department of Medicinal Chemistry, National Institute of Pharmaceutical Education and Research, S.A.S Nagar,
Punjab-160062. India and
b
Faculty of Life Sciences and Biotechnology, South Asian University, New Delhi- 110021, India *E-mail:
rkartha@niper.ac.in; yrp@sau.ac.in
a
b
#
Rajesh Gour & Faiz Ahmad equally contributed to this manuscript
Abstract
We report herein the synthesis and anticancer activity of a set of novel S-linked artemisinins
bearing an aliphatic/aromatic/heterocyclic nucleus as a substituent on the sulfur. The compounds
were prepared from artemisinin via its lactol-form by an acid-catalyzed condensation of the
desired thiol with the lactol. Both the C-10-α- and β-configured thiol ethers were synthesized
with a view to making them available for the anticancer activity evaluation using a variety of cell
lines. The results show that many of the synthetic derivatives studied possessed good potential as
anticancer agents. In order to draw more information on the origin of the anticancer activity, one
of the compounds (9a), that showed a broad-spectrum activity in terms of reducing the viability
of most of the cell lines studied, in particular proven to be most effective against Prostate (PC-3)
cells, was studied in detail to find the underlying mechanism of its action by MTT assay,
immunoblotting, flow cytometry and microscopy. Pretreatment of the PC-3 cells with N-acetyl
cysteine affected the efficacy of 9a, suggesting the role of reactive oxygen species in reducing
their viability. Cell cycle analysis showed increase in G1 phase that was indicative of G1 cell
cycle arrest. Wound healing assay revealed anti-migratory effect of 9a Quantitative PCR and
western blot analysis showed changes in the gene expression of PCNA, E2F1, Pin1, cyclinD1,
phospho-c-jun, c-Myc,eIF4E and other genes involved in proliferation and maintaining the
transformed phenotype of prostate cancer cells. Here we report the anti–proliferative property of
9
a with a vital and potent target(s) in prostate cancer cells with one of such targets being Pin1

ACCEPTED MANUSCRIPT
belonging to the parvulin family of PPIases. The results suggest that 9a could be a promising
agent in combating prostate cancer.
Keywords: Artemisinin thioacetal; S-Linked dihydroartemisinin; Anticancer activity; Pin1
1
. INTRODUCTION
Artemisinin (1, ART, see Fig.1 below) and its derivatives (artemisinins, ARTs) have been a set
of extremely successful drugs used for combating malaria, in particular, resistant malaria [1].
Some of the important attributes of ARTs such as their high efficacy, quick action, and minimum
side effects have greatly influenced researchers around the globe prompting them to study their
pharmacological properties beyond the antimalarial activity. Thus, anti-helminthic, antifungal,
antiviral as well as anticancer activities are some of the other reported properties of ARTs [2].
Since the anticancer action of ART-related endoperoxides was reported in the early 1990s by
Woerdenbag et al [3], a number of ARTs have been developed in attempts to derive the best-
desired effect against cancer cells.
The anticancer effects of ARTs have been found to be majorly associated with ROS (reactive
oxygen species) production following cleavage of the endoperoxide moiety present in their ring
system which gets reduced by the ferrous ions present inside the cancer cells in the high amount
owing to the need for rapid proliferation [[4], [5], [6], [7]]. But the debate on the exclusivity of
this mode of anticancer mechanism got started seriously when ROS-independent apoptosis in
HL-60 leukemia cells was reported by Lu et al [8], followed later by Qin et al in Hep G2 cells by
artesunate [9]. These and other reports suggest unequivocally that ARTs affect a spectrum of
pathways involved in proliferation, metastasis, and apoptosis [[10], [11], [12], [13]]. Thus, the
versatility of ART and its derivatives indicated above led us to the synthesis and anticancer
screening of a library of novel sulfur-linked ARTs as reported herein.
In the current study, we used S-linked artemisinin derivatives (4a-13b) to study their anti-cancer
property on seven cancer cell lines (LNCaP, PC-3, A-431, HeLa, Hep G2, MDA-MB-231, and
A549). Addition of sulfur atom aids in the enhanced uptake of molecules [14], and this may
enhance the anti-cancer property of these derivatives. In our study, we found 5b and 9a affecting
more than 50% cell viability against most of the cell lines studied. Thus, for a further
mechanistic study on the mode of action of this class of compounds, compound 9a was chosen. It

ACCEPTED MANUSCRIPT
was found in this study that 9a inhibited cell proliferation, affected/influenced the colony
formation nature and induced apoptosis in metastatic PC-3 cells.
Further, in our study, we discovered that the compound had vital targets, Pin1 being one of them,
which importantly is a pleiotropic effector, having a number of interacting partners influencing a
number of signaling cascades including cell proliferation and apoptosis. Pin1 is a small enzyme
known for catalyzing the cis-trans isomerization of proteins phosphorylated on serine or
threonine, at the proline amide bond, thereby altering the conformation and consequently
regulating the protein function [[15], [16]]. Subsequent studies reported in the literature
highlighted the role of Pin1 in various cancers, in which either it is overexpressed and or over
activated [[17], [18]]. A clinical study on prostate cancer patients has demonstrated Pin1 as an
independent prognostic marker [19]. Although ART derivatives have been shown to inhibit
cancer cell proliferation including prostate cancer cells, there are no reports of any ART
derivative inhibiting prostate cancer cell proliferation involving Pin1. In the present study using
9
a, we have shown the apoptosis of metastatic prostate cancer cell line PC-3 with the down-
regulation of Pin1 at the protein level.
2
. RESULTS AND DISCUSSION
Chemistry. Fishwick et al described that dihydroartemisinin (2, DHA) was responsible for the
inhibition of neurite outgrowth from differentiating NB2a neuroblastoma cells. Subsequent
studies confirmed the cytotoxic effects towards neurons [20]. It is well known that ART
derivatives bearing an alkyl group, as in artemether (2a), arteether (2b), etc, at C-10 through
oxygen-linkage are readily turned over by cytochrome P-450 present in the liver to produce 2
that has been proven to be neurotoxic (Fig.1) [21]. The elimination half-life of the water-soluble
2
is less than 1 h after intravenous administration in rats [22]. Therefore, we wished to synthesize
derivatives of artemisinin that are potentially more stable to cytochrome P-450. S-Linked
residues at C-10 of artemisinin seemed to serve this purpose. This derivative may also emerge as
relatively more stable to hydrolysis. Moreover, additional substituent groups can also be
introduced in the residue to change the solubility characteristics of the compound derived.

ACCEPTED MANUSCRIPT
Fig. 1. Artemisinin (1) and some of its known derivatives 2 & 3. In the inset is shown the anomeric center of a
pyranosugar for comparison of the C-10 center of 2 & 3
For the synthesis of novel S-linked artemisinins, DHA (2) was used as the starting material.
Reduction of the commercially available ART (1) with sodium borohydride in dry methanol at 0
°
C afforded an equilibrium mixture of α and β isomers of 2 in 30 min in 95% yield as shown in
1
13
Scheme 1a. The stereochemistry of 2 [23], at C-10 was determined by H and C NMR
spectroscopy. The signals for the C-10-α- and β-hydrogens appeared at 4.75 ppm (J = 9.2 Hz)
and 5.30 ppm (J = 3.2 Hz) respectively. Presence of the hydroxyl group was confirmed by
-1
absorption at 3369 cm in the IR spectrum and further confirmation of the product was obtained
+
by MS (MALDI-TOF) for C15
H
24
O
5
, calculated m/z 284.16. Found m/z 307.65 [M+Na] ,
+
3
23.65[M+K] . The lactol 2 thus obtained was then acetylated using acetic anhydride as the acyl
donor in the presence of pyridine (8 h, rt), or alternatively, DABCO (2 h, rt), as the acyl transfer
agent. The lactol acetate 3 was obtained as its α- isomer in excellent yield the structure of which
was, again, proven by NMR spectral characterization as for 2. Thus, compared to a value of 4.75
ppm in 2, the H-10 signal shifted downfield (5.78 ppm, J = 9.9 Hz) in the case of 3. Moreover, in
1
3
the C NMR spectrum of 3, the carbonyl carbon appeared at 169.9 ppm. In its IR spectra,
-1
likewise, the 1752 cm absorption obtained was also typical for the carbonyl group. The MS
spectrum of the compound (MALDI-TOF: for C H O , calculated m/z 326.17; found m/z
17
26
6
+
+)
3
49.51 [M+Na] , 365.59 [M+K] was also confirmatory of the expected acetate’s structure.

ACCEPTED MANUSCRIPT
Scheme 1a. Synthesis of dihydroartemisinin and dihydroartemisinin acetate
An examination of the nature of the lactol residue in 2/the lactol acetate residue in 3, readily
reveals their resemblance to the anomeric center of a pyranosugar, as its hemiacetal/the derived
glycosyl acetate, respectively (please see the structure shown inset under (Fig.1) for a general
representation of this feature). Therefore it could not be surprising that when 3 was treated with
an alkyl thiol in the presence of molecular iodine, a method that is used for the synthesis of alkyl
thioglycosides of sugars from their respective per-O-acetylated glycoses [24], the desired lactol
thioether was successfully obtained in excellent yields (Method 1, Scheme 1b). In the same
manner, treatment of 3, with alkyl thiols in the presence of BF .Et O [25], a traditional Lewis
3
2
acid employed for the electrophilic activation of the glycosyl acetate residue for S-glycosylation,
also led to the formation of the expected lactol thioethers (Method 2, Scheme 1b).
H
H
3
14
O
O
Method 1
I2/MeCN, -15 °C
O
O
O
O
O
O
+
H
H
H
H
1
0-20 min
75%
R SH
4 - 13)
>
(
S
S
R
R
Method 2
BF .Et O/CH Cl
(4b - 13b)
3
2
2
2
(4a - 13a)
-20 °C
2/3
H
O
R = aliphatic/aromatic/heterocyclic residue; (See Table 1 for details)
O
O
O
H
H
14
Scheme 1b. Synthesis of various thioethers of DHA
In the case of the BF -catalyzed reaction it was observed that a significant quantity of the β-
3
product accompanied the major α-compound formed in the reaction. They could be separated

ACCEPTED MANUSCRIPT
chromatographically on a column of silica for their unambiguous characterization by various
spectroscopic techniques. The stereochemistry of the products was ascertained by the H-10
1
chemical shift observed in their respective H NMR spectra. The large coupling constant
between H-9 and H-10 was indicative of trans-diaxial relationship and was assigned the α-
configuration (4a in the case of n-BuSH), while the small coupling constant was indicative of cis
relationship and was assigned the β-configuration. Interestingly, in the case of the iodine-
catalyzed reaction of 3 with alkane thiols, an exclusive formation of the α-configured product
could be achieved. Thus in a model reaction when DHAA taken up in MeCN was treated with n-
BuSH (4, entry 1, Table 1) at -15 °C the desired S-butyl derivative 4a was obtained in a yield of
7
5% in 15 min of reaction. The isomer 4a was the only product obtained under the above
conditions when a 4 mol equiv of the thiol was used for the reaction. In the presence of lower
mol equiv of the thiol, the corresponding β-isomer 4b was also obtained along with 4a. Thus,
under the conditions of DHAA:n-BuSH of 1:1, 4a:4b was obtained in a ratio of approximately
1
:1 whereas at a DHAA:n-BuSH ratio of 1:2, the set of isomeric compounds were obtained in a
ratio of 3:2. It was also observed that the iodine-method was best suited for alkyl thiols such as 4
and 5. In the case of N-containing thiol acceptors such as 12, 13, etc the reaction did not take
place at -15 °C. But when it was attempted at room temperature, was found to lead to the
formation of degradation products such as the anhydro compound (14 [26], Scheme 1b). Thus,
although the iodine-catalyzed method described above was well-suited for the preparation of α-
S-alkyl-linked artemisinin derivatives, the BF .Et O-assisted method was finally adopted in the
3
2
current context in view of its wider applicability as well as the requirement for both the α- and β-
linked compounds for the biological evaluation later.
Based on the above-described method various S-linked compounds, 4a [27], 4b [27], 5a, 5b, 6a,
6
4
b, 7b, 8a, 8b [28], 9a, 9ab, 10a, 10b, 11a, 12a, 12b, 13a and 13b were obtained by using thiols
, 5, 6, 7, 8, 9, 10, 11, 12 and 13, respectively, as the starting materials as shown in Table 1; and
were characterized by standard spectroscopic techniques as described in the foregoing.

ACCEPTED MANUSCRIPT
Table 1. S-Linked artemisinin derivatives synthesized having aliphatic/aromatic/heterocyclic
substitution at C-10 position
Entry
R
2
-SH
Products
1
2
3
4
5

ACCEPTED MANUSCRIPT
6
7
8
9
1
0

ACCEPTED MANUSCRIPT
Computational study. Various pharmacokinetic parameters of the synthesized derivatives were
examined by ADMET study. The results revealed that the synthesized compounds possess the
drug-likeness properties governed by the Lipinski’s rule of five. The LogP values, that denotes
the hydrophilicity/lipophilicity of compounds, are in the permissible range for the synthesized
compounds as shown in Table 2. As low hydrophilicity leads to a high logP value and vice
versa, governing the absorption or permeation of compounds, a logP value lower than 5 is a good
indicator for the absorption property of compounds. The numbers of HB donor and HB acceptor
in the molecule are the other important parameters for an expected better absorption and
permeation of compounds that also turned out to be in the allowed range for the synthesized
compounds. In addition, the synthesized compounds fell under 500 daltons in their molecular
weight, within the recommended limits (by the Lipinski’s rule) showing them to be acceptable in
terms of the solubility/cell membrane permeability parameters.
Table 2. Drug likeness property calculated using Qikprop software.
Molecular
weight
HB
HB
ADME
Entry
logPo/w
(< 5)
descriptors
Donor
Acceptor
Standard Range
<500 Da
(< 5)
(< 10)
1
2
3
4
5
6
7
8
9
0
1
2
3
4
Arteether
312.40
298.37
282.33
356.51
356.51
384.57
384.57
372.47
372.47
420.52
420.52
420.52
448.57
448.57
0
0
0
0
0
0
0
0
0
1
1
1
0
0
5.65
5.65
5.25
4.45
4.45
4.45
4.45
6.45
6.45
6.45
6.45
6.45
6.45
6.45
2.82
2.36
1.73
4.00
3.8
Artemether
Artemisinin
4a
4b
5a
4.78
4.51
3.03
2.9
5b
6a
6b
7b
8a
1
3.90
3.87
3.52
4.65
4.09
1
1
1
1
8b
9a
9ab

ACCEPTED MANUSCRIPT
1
5
10a
410.95
0
4.45
5.12
1
1
1
1
2
2
6
7
8
9
0
1
10b
11a
12a
12b
13a
13b
410.95
421.50
427.55
427.55
382.47
382.47
0
0
0
0
0
0
4.45
5.45
4.95
4.95
6.95
6.95
4.93
3.87
5.24
5.01
2.54
2.39
The aqueous solubility of a compound is important in determining its absorption and distribution
properties, usually indicated as LogS value of the compounds (Table 3). On this count also, the
synthesized compounds exhibited their respective values in the permissible range. Central
nervous system (CNS) activity parameter for any compound allows the range from -2 to +2. A
value of -2 for a compound is considered as inactive and a value of +2 as active. As shown in
Table 3 our compounds have CNS score ranging from -1 to 2. Out of these compounds, two have
shown a score of +2 and therefore these two compounds, namely 10a and 10b, could be CNS-
toxic. Similarly values of other parameters like apparent Caco-2 cell permeability, blood-brain
barrier (logBB), apparent MDCK permeability, skin permeability (logKp) and serum-protein
binding (logKhsa) are also in the acceptable range for synthesized compounds (Table 3).
Table 3. Pharmacokinetic parameters calculated using Qikprop Software
Apparent
Caco-2
Apparent
MDCK
logKhsa
Serum
Protein
Binding
Predicted logS for
logKp for
skin
permeability
ADME
descriptors
logBB for
Entry
CNS
aqueous
permeability Brain/Blood permeability
activity
solubility
(
nm/sec)
(nm/sec)
Standard
Range
(<25 poor,
>500 great)
(<25 poor,
>500 great)
(
-2,+2)
(-6.5/0.5)
(-3/1.2)
(-8/-1)
(-1.5/1.5)
1
2
3
Arteether
Artemether
Artemisinin
1
1
0
-3.08
-2.44
-2.26
5875.74
5878.13
1811.52
0.24
0.31
-.031
3354.40
3355.88
940.29
-1.76
-1.86
-2.95
-0.11
-0.27
-0.32
4
5
6
4a
4b
5a
1
1
1
-3.91
-4.50
-4.82
6711.94
8538.22
6712.21
0.21
0.29
0.08
4800.64
6025.41
4800.85
-1.46
-1.26
-1.27
0.34
0.33
0.60

ACCEPTED MANUSCRIPT
7
8
9
5b
6a
1
0
-5.3
8455.58
1885.35
1712.89
209.13
0.16
-0.23
-0.27
-0.46
-0.59
-0.65
-0.17
-0.18
0.64
6000.93
1439.10
1273.34
173.99
165.60
129.76
2189.78
1665.76
10000
-1.07
-2.63
-2.71
-2.99
-3.03
-3.16
-1.92
-2.09
-1.04
-1.29
-2.78
-0.85
-1.04
-2.75
-3.02
0.54
0.01
-0.03
0.27
0.24
0.13
0.58
0.36
0.68
0.64
0.45
0.80
0.73
-0.18
-0.21
-4.04
-3.87
-4.40
-4.91
4.56
6b
0
1
0
1
2
3
4
5
6
7
8
9
0
1
7b
-1
-1
-1
0
1
1
1
1
1
1
1
1
1
2
2
8a
185.07
8b
166.73
9a
-5.69
-4.20
-5.51
-5.22
-4.78
-5.77
-5.30
-3.13
-3.07
2640
9ab
10a
10b
11a
12a
12b
13a
13b
0
2116.63
8539.19
6511.72
1022.69
6977.64
5726.85
1781.75
1298.90
2
2
0.53
10000
0
-0.44
0.38
784.24
6360.93
4942.01
1395.80
971.88
1
1
0.30
0
-0.12
-0.24
0
#
Schrodinger Release 2014-1: QikProp, Schrodinger, LLC, New York, NY, 2014
Biology. The compounds thus synthesized (Table 1) were evaluated for their anticancer potential
as pure α- and β-isomers and/ their 1:1 mixture, as applicable). For the evaluation of the
(
compounds, the cell viability MTT assay was conducted using the following cancer cell lines,
namely LNCaP & PC-3 (prostate cancer cell lines), A-431 (epidermoid carcinoma cell line),
HeLa (cervical cancer cell line), Hep G2 (hepato-carcinoma cell line), MDA-MB-231 (breast
cancer cell line) & A549 (Lung cancer cell line) and the percentage cell viability of the
synthesized S-Linked artemisinin derivatives was calculated and is shown in Table 4. Further, 12
of the above compounds, namely, 4b, 5a, 5b, 7b, 8a, 8b, 9a, 9ab, 10a, 10b, 11a and 12b showed
less than 50% cell viability on LNCaP cell line. Likewise, thirteen of them, namely, 4b, 5b, 6a,
6
b, 8b, 9a, 9ab, 7b, 11a, 12a, 12b and 13a, 13b, showed less than 50% cell viability on PC-3
cell line. In the same manner, while compounds 4b, 5a, 5b, 8b, 9a, 10a, 10b, 11a, and 12b were
active against on A-431 cell line, compounds 4b, 5a, 5b, 9a, 9ab, 10a, 10b and 11a were active

ACCEPTED MANUSCRIPT
against HeLa cells; compounds 5a, 5b, 8b, 9a, 10a and 10b were active against Hep G2 cells;
and compounds 5b and 9a were active against MDA-MB-231 as depicted in (Fig. S1 A and B).
Significantly, these compounds, as can be seen from Table 4, were only marginally active, if at
all, in A549 cell line. Following the encouraging results obtained in the above screening
experiments, the IC values for 5b and 9a were calculated for four different cell lines and the
5
0
results are shown in Table 5. The significant reduction in the cell viability of various cell lines at
0 µM with these derivatives can be seen from the results, which is in sharp contrast to that of
1
artemisinin that requires > 200 µM concentration to be effective as reported earlier against
different cancer cell lines [[29], [30], [31]].
Table 4. Percentage cell viability of different cancer cell lines against synthesized S-linked artemisinin
derivatives
MDA-
Entry
1
Compound
LNCaP
PC-3
A-431
HeLa
Hep G2
A549
MB-231
Control
100±5.0
90±11.4
100±18.8 100±11.7 100±11.5 100±6.5 100±20.0 100±9.6
2
3
4
5
6
7
8
9
4a
92±4.7
86±21.9 82±11.8
92±5.8
90±4.1
96±5.6
4b
5a
47±1.7
36±2.1
46±4.4
60±9.7
71±2.4
43±2.2
47±5.1
41±1.9
38±2.7
31±1.0
37±2.3
37±4.3
33±3.9
90±2.9
39±3.0
32±2.4
44±9.9
41±5.7
84±2.3
32±3.2
28±5.1
31±15.8
70±13.4
99±2.1
47±13.3
35±9.0
45±4.0
75±12.0
49±4.0
44±4.1
49±2.8
79±4.7
52±2.2
35±0.8
41±5.6
80±5.5
94±3.0
60±8.7
64±5.1
45±3.5
50±2.2
67±7.9
42±1.2
36±1.7
71±20.5
72±13.3
39±6.6
98±4.6
99±4.3
97±7.0
95±13.2
64±13.1
45±4.8
96±6.4
97±8.1
70±15.9
79±7.6
70±5.3
55±6.1
91±5.7
97±3.4
89±3.0
79±6.0
76±3.2
68±3.1
74±6.3
55±10.5
52±6.0
5b
6a
6b
77±17.4 80±10.6
7b
90±8.9
58±10.5
31±6.2
35±5.4
52±12.7
42±6.6
36±7.2
58±5.0
60±3.0
55±10.1
43±3.2
43±3.9
48±5.0
41±2.7
8a
1
1
1
1
1
0
1
2
3
4
8b
9a
9ab
10a
10b

ACCEPTED MANUSCRIPT
1
1
1
1
1
5
6
7
8
9
11a
12a
12b
13a
13b
46±1.2
77±20.7
41±1.2
78±4.1
97±5.3
50±10.7
33±5.0
39±3.4
47±7.7
28±2.8
44±11.6 50±10.0
56±3.5
96±2.4
72±7.2
95±5.7
93±6.4
98±3.9
90±12.1
78±17.1
75±27.2
86±18.8
70±7.1
100±7.1
87±3.4
89±4.3
100±3.8
93±6.5
44±4.5
87±14.4
97±7.5
87±4.9
63±6.8
95±9.2
93±8.4
LNCaP, PC-3 (prostate cancer cell lines), A-431 (epidermoid carcinoma cell line), HeLa (cervical cancer cell line),
Hep G2 (hepato-carcinoma cell line), MDA-MB-231 (breast cancer cell line), A549 (Lung cancer cell line).
Table 5. IC50 values for the S-linked artemisinin derivatives 5b and 9a against different cancer
cell lines (µM)
Compounds PC3
A431
3.99
13.3
Hep G2
10.72
6.08
MDA-MB-231
3.94
5
9
b
a
5.8
6.13
9.03
Results of the primary screening thus proved compounds 5b and 9a to be of comparable broad-
spectrum anticancer activity effective on six (namely, LNCaP, PC-3, A-431, HeLa, HepG2, and
MDA-MB-231) of the seven cancer cell lines evaluated. Thus, based on the fact that the
synthetic protocol was consistently α-stereoselective, making the preparation of crystalline 9a in
the required quantities extremely facile, compound 9a was chosen for further detailed studies
using the prostate cancer cell line PC-3 (against which they had the best activity with IC of 6.13
5
0
µM) as the cell line of choice. The efficacy of 9a was also compared with natural artemisinin
ART) on the PC-3 cells by the MTT assay and results so obtained proved the superiority of 9a
(
as can be seen from (Fig. 3 A).The results obtained on the effect of compound 9a, as a function
of concentration (0.1, 0.3, 1.0, 3 and 10 µM) and treatment time (24, 48 and 72 h), on the
viability of PC-3 cells as evaluated by MTT assay are shown in (Fig. 3 B-D). Thus, a treatment
time of 24 h had no significant effect on the viability even up to 10 µM. Concentrations of 1 to
1
0 µM, however, showed a significant effect when treated for 72 h.

ACCEPTED MANUSCRIPT
Fig. 2. Reduction of the viability and cell proliferation of prostate cancer (PC-3) cells by compound 9a: (A) Efficacy
of 9a compared with artemisinin against treated PC-3 cells at 10 µΜ for 72 h. (B, C and D) PC-3 cells was treated
with 9a in time (24, 48 and 72 h) and dose dependent manner and assayed, post-incubation, with MTT reagent for
the cell viability. DMSO and doxorubicin (Dox)-exposed cells were used as a vehicle and positive control group
respectively. (E) Representative image of the colony formation assay. (F) Bar graph showing the colony numbers in
each treated group, as analyzed by Image J software. Results are expressed as fold change ± s. d. and are
representative of three separate experiments each performed in quadruple. (*) indicates P< 0.05 and (**) indicates
P<0.01 and (***) indicates P<0.001- significant difference, control Vs respective treatment group..
For evaluating the long-term cell survival and the ability of the cells to form colonies in the
presence of 9a, and for comparison, separately with artemisinin (at 3 and 10 µM), colony assay
was performed in a dose-dependent manner. The outcome of experiment was in accordance with
the results obtained in the MTT assay and thus, while a reduced clonogenic growth was observed

ACCEPTED MANUSCRIPT
in a dose-dependent manner with 9a (as shown in Fig. 2 E and F), a much less effect was
observed with artemisinin even at 10 µM concentration (Fig. S 2A) . Both of these assays
proved that the anti-proliferative activity of 9a is significantly higher than that of artemisinin
itself against the PC-3 cells. As reported by Li. et al. the introduction of the sulfur atom in
compounds, besides other possible factors, may have led to the enhancement of the cellular
uptake, which may magnify the effect of compounds by accumulating around the target proteins
[14].
For investigating the mechanism, lying behind the antiproliferative activity of this compound
9a) and to know whether ROS, if produced by the compound-treated cells, has any role in
(
reducing the cell proliferation (as the anti-proliferative properties of artemisinin derivatives have
been reported to be both ROS-dependent [[7], [13], [32]] and ROS-independent). We measured
ROS level inside the cells following treatment of the cells with compound 9a and performed
fluorescence microscopy, with CM-H2DCFDA staining. A gradual increase of ROS producing
cells on treatment with 9a, as compared to the vehicle control cells, was observed as shown in
(
Fig. 3 A and B). Next the cells were pretreated with ROS scavenger NAC (20 µM) for 2 h
followed by treating them with 9a for 72 h. The cytotoxicity of 9a in the presence and absence of
NAC was evaluated by MTT assay and it was found that the anti-proliferative property of 9a
significantly depended upon the ROS activity as shown in (Fig. 3 C).One of direct consequences
of the abrupt increase in the ROS level is the alteration of mitochondrial membrane potential
[
33]. Thus, we next evaluated ∆Ψ using JC-1 probe and found a decreasing red/green ratio in
m
the 9a-treated cells as compared to the control as shown in (Fig. 3 D), which is a sign of early
stage of apoptosis. In contrast, a nearly equal red/green ratio was observed for the artemisinin-
treated cells (Fig. S 2B). Taken together, these experiments thus revealed a significant
production as well as involvement of the ROS in the antiproliferative activity of compound 9a.

ACCEPTED MANUSCRIPT
Fig. 3. (A) Representative photographs are shown for CM-H2DCFDA-stained PC-3 cells observed using
fluorescence microscopy. (B) Bar graph showing the percentage of ROS producing cells for the representative
image in (A). Involvement of ROS in the anti-proliferative activity of compound 9a: (C) PC-3 cells were treated
with 9a at 3 and 10 µM alone and in the presence of the ROS scavenger NAC to explore the participation of any
ROS activity and graph shows the cell viability of 9a-treated cells in the presence of 20 µM NAC. Results are
expressed as fold change ± s.d. and are representative of three separate experiments; (**) indicates P<0.01 and (***)
indicates P<0.001. (D) ∆Ψ was evaluated using JC-1 dye by observing change in red/green fluorescence under
m
fluorescence microscope.
We recalled, a study on the prostate cancer cells, including PC-3 cells, had found the existence of
an inherent oxidative stress in these cells, for proliferation and survival [34], and a
disproportional increase in the ROS level could lead to the cell cycle arrest and apoptosis [35].
Therefore, we next sought to check the status of different phases of the cell cycle and apoptosis
in the 9a-treated cells in the following sets of experiments. Cell cycle analysis revealed that PC-3
cells treated with 3 and 10 µM of 9a were arrested in the G1 phase (54.49% and 53%) as
compared to the control (42%) and was found to be consistent with the decrease in the
percentage of cell population in S and G2M phase as shown in (Fig. 4 A, B and C). On the other
hand treatment with 3 and 10 µM of artemisinin had a non-significant effect on the G1 phase and
showing a 2 and 6 %, respectively, arrest in the S-phase of cell cycle (Fig. S 2C). Thus, the result
so obtained in the cell cycle analysis led us to envisage the induction of apoptosis if any as to
establish a link between the losses of viability via cell cycle arrest. So firstly, phase contrast
microscopy was performed in order to detect any changes associated with apoptosis. As shown

ACCEPTED MANUSCRIPT
in Fig.5 A, the images thus captured showed shrinkage in the cell size, which is one of the
features associated with the early process of apoptosis. In contrast, the artemisinin-treated cells
barely lost their cellular morphology (Fig. S 2D). Next, Hoechst 33342 staining and
fluorescence microscopy were also carried out to assess any changes in the nuclear morphology,
such as the nuclear condensation or nuclear fragmentation of the treated cells. It was found that
the nuclei of the treated cells were condensed as compared to the vehicle control cells, as shown
in (Fig. 5 B) as well as the artemisinin-treated cells (Fig. S 2D). To further confirm the induction
of apoptosis following the treatment of cells for 72 h in a dose-dependent manner, caspase-9 and
caspase-3 activation were checked by western blot. A decrease in the pro-caspase form of
caspase 9 and 3, (Fig. 5 C), was found which further strengthened the finding of apoptosis
pathway and the cause of reduction in the viability of 9a-treated cells. This decrease in the pro-
caspase 9 and 3 was accompanied by the down-regulation of the anti-apoptotic protein Bcl-xL
and up-regulation of pro-apoptotic protein Bax in a dose-dependent manner, as shown in (Fig. 5
C), suggesting the activation of an intrinsic pathway following the treatment with 9a. The down-
regulation of Bcl-xL is a promising indicator for the future, as Bcl-xL is overexpressed in ~100%
of the hormone-refractory prostate cancers and is associated with advanced disease, poor
prognosis, recurrence, and metastasis [[36], [37]]. Finally, the cells treated with 9a at 3 and 10
µM levels of concentration were stained with propidium iodide to examine the cell death
following apoptosis. Compound 9a at these concentration levels significantly activated cell death
in a dose-dependent manner following a 72 h-treatment, as shown in (Fig. 5 D and E). However,
it may be noted that the maximum cell death observed was only 28.9%, approximately, in
repeated experiments.

ACCEPTED MANUSCRIPT
Fig. 4. 9a causes G1 cell cycle arrest in PC-3 cells (A, B) - Cells treated with 3 and 10 µM of 9a were fixed and
permeabilized with 80% chilled ethanol and stored for overnight at 4 ºC. Then cells were washed and subjected to
RNase treatment and PI staining for 30 min at room temperature for flow cytometry to analyze DNA content. (A)
Histogram of a representative experiment. (B) Data are represented as mean ± SD of three independent experiments,
where (**) indicates P < 0.01 compared to vehicle control as determined by t-test. (C) Table showing percentage of
cells in different phases of cell cycle following treatment with 9a as compared to control.

ACCEPTED MANUSCRIPT
Fig. 5. 9a-Induced apoptosis and cell death in PC-3 Cells: (A) Morphology alteration in treated cells. Exponentially
growing PC-3 cells were exposed to 9a at 3 and 10 µM levels of concentration for 72 h, gently washed with 1×PBS
and then photographed with phase contrast microscopy. (B) Fluorescence micrographs of PC-3 prostate cancer cells
stained with Hoechst 33342. Cells were treated with 3 and 10 µM solutions of 9a for 72 h. It shows that 9a induces
apoptosis in PC-3 cells and are characterized by nuclear condensation. (C) PC-3 cells treated with 9a for 72 h shows
a decreased pro-form of caspases 9 and 3, the initiator and executioner caspases, and inhibition of the anti-apoptotic
protein Bcl-xL whereas upregulation of pro-apoptotic Bax. (D) Cell death analysis after treatment with 9a. PC-3
cells were cultured in a six-well culture plate, incubated with concentrations of 3 and 10 µM of 9a for 72 h. Cells
were harvested, washed with PBS and incubated with propidium iodide for 20 min at 4ºC. Samples were acquired
for cell death using flow cytometry. Cells with DMSO treatment were taken as the control. Data represent
percentage cell death after treatment with different concentrations of compound 9a. (E) Bar graph showing the
percentage of cell death for the representative image in (D), where (***) indicates -p < 0.001 compared to vehicle
control as determined by t-test.
Next, besides the apoptotic effect of 9a against the PC-3 cells, we performed the scratch assay to
assess the migration potential of these highly metastatic PC-3 cell lines following treatment with
9
a at 3 and 10 µM. As expected, we found that 9a significantly inhibited migration of the PC-3
cells as compared to the control (Fig. 6 A and B). Contrary to this the artemisinin-treated cells
retained the migration ability and filled the wound at the same time point as filled by the control
group (Fig. S 2E). In addition, the anti-migration effect of 9a was also assessed in the presence
of pan caspase inhibitor Z-Vad-Fmk to inhibit any apoptotic process during the wound healing
assay. It was found that blocking of apoptosis did not affect the anti migration effect of 9a (Fig 6
C and D). Thus both of these experiments suggest that 9a can inhibit PC-3 cell migration
independent of its apoptotic effect. In connection with this inhibitory action of 9a on the
migration potential of PC-3 cells, Gelatin zymography was done to assess the level of gelatinases
like MMP-2 and MMP-9 following treatment with 9a at 3 and 10 µM levels. While a dose-
dependent inhibition in the activity of MMP-9 was observed, MMP-2 remained unaffected as
shown in (Fig. 6 E). MMPs are matrix metalloproteinases and are involved in matrix degradation
that lead to cell migration causing metastasis and invasion [38]. In the context of prostate cancer,
MMP-2 and MMP-9 have been reported to be associated with metastatic potential leading to
severity of malignancy [39]. Therefore, the inhibition of MMP-9 by 9a calls for further detailed
study for evaluating its anti-metastatic potential against PC-3 cells.

ACCEPTED MANUSCRIPT
Fig. 6. 9a-Inhibited cell migration of PC-3 cells: (A and B) The inhibitory effect of 9a on PC-3 cells migration
detected using wound healing assay as shown in the representative image. Treatment with 3 and 10 µM 9a resulted
in decreased cell migration of PC-3 cells as compared to DMSO treated group. A distinguishable difference was
observed on the migration ability of PC-3 cells under the influence of 9a at different time intervals (0, 4, 8 and 12 h)
as evident from phase contrast micrographs. (C and D) Wound healing assay in the presence of pan caspase
inhibitor Z-Vad-Fmk- cells were pretreated with Z-Vad-Fmk(10 µM) for 2 h followed by treatment with 9a, wound
at 0, 12 and 24, are shown. E Gelatin zymography- Effect of 9a on MMPs abundance and activity in serum-free
media following treatment of PC-3 cells with 3 and 10 µM concentrations. Briefly, post treatment of cells in serum-
containing media for 48 h cells were further kept for 24 h in serum-free media. Media was collected and
electrophoresed with gelatin (4mg/mL) for the analysis of MMP 9 and 2. Representative gelatinase zymograph,
detected as white bands against blue background as well as black bands against grey background of repeated
independent experiments, each of which showed similar results. (F and G) Transwell assay the inhibitory effect of
9
a on PC-3 cells migration detected using Transwell assay as shown in the representative images. Hoechst 33342
stained nuclei present number of migrating cells in control and respective treatments. Treatment with 3 and 10 µM
a resulted in decreased cell migration of PC-3 cells as compared to DMSO treated group. Bar graph showing
migrating cells/field. (***) indicates -p < 0.001 compared to vehicle control as determined by t-test.
9

ACCEPTED MANUSCRIPT
Migration potential of PC-3 was also challenged by Transwell assay in the presence of 3 and 10
µM of 9a and similar results were obtained as in scratch assay. The number of cells migrating
through pores significantly decreased at both concentrations of 9a as shown in (Fig. 6 F and G).
Fig. 7. (A) Effect of 9a on mRNA expression evaluated by RT-PCR- A relative expression of PCNA,E2F1, Cyclin
B1,Bcl-xL,p21 and Noxa at mRNA level following treatment of PC-3 Cells for 72 h at 3 µM (blue bar) and 10
µM(red bar) concentrations. Results are expressed as means ± s.d. and are representative of three different
experiments. (B) 9a upregulates E2F1whereas it inhibits Pin 1, Cyclin D1, phospho-c-jun, eIF4E, c-Myc and PCNA:
PC-3 cells treated with (0.1-10 µM) of 9a for 72 h were subjected to western blotting and β-actin was used as
loading control. (C) Pin1 siRNA and 9a together significantly down-regulated Pin1. PC-3 cells were transfected
with Pin1-siRNA alone and with 9a for 72 h and were subjected to western blotting and β-actin was used as loading
control. Wortmannin, an autophagy inhibitor, unaltered the Pin1 inhibition by 9a: PC-3 cells treated with 10 µM of

ACCEPTED MANUSCRIPT
9
a in the presence and absence of Wortmannin for 72 h were subjected to western blotting. Pretreatment with
MG132 rescued Pin1 inhibition by 9a.
Further, to study the mechanism responsible for the reduction observed in the viability of cells
following treatment with 9a, a set of genes (not shown) involved in cell division, apoptosis and
anti-apoptotic processes were chosen and the following experiments were carried out. Towards
this end, essentially, the mRNA expression analysis by real-time PCR was performed. Cells
treated for 72 h with 3 µM and 10 µM concentrations of 9a were analyzed for the changes in the
mRNA level of the following genes – E2F1,PCNA, CyclinB1,Bcl-XL that were found to be
down-regulated, and p21 and Noxa, that were found up-regulated, as shown in (Fig. 7 A). Only
marginal changes in other selected genes were observed. E2F1 belonging to the E2F family of
transcription factors act as a transcriptional activator for progression through the G1/S transition
[40]. Activated E2F1 can induce apoptosis as well depending on the cumulative signal operating
in the cell [41]. Here we have found the inhibition of E2F1 at transcript level using 9a at 3 and
1
7
0 µM levels but the upregulation at protein level in a dose-dependent manner as shown in (Fig.
A and B, respectively). This negative correlation between mRNA and protein abundance could
exist due to the difference in the rate of protein synthesis and degradation for a given protein
42]. Thus it could be deduced that the up-regulation of E2F1 by 9a steers the PC-3 cells towards
[
apoptosis by modulating apoptotic signals.
Real-time analysis paved the way for further mechanistic study in which either the proteins of
these mRNA or their targets were analyzed by western blotting. Pin1 is one of the E2F1 target
genes [43], that has been found aberrantly upregulated or overactivated in many cancers and in
prostate cancer it has been reported as an independent prognostic marker [19]. In our study, Pin1
was significantly down regulated at 3 µM and 10 µM of 9a as shown in Fig. 7 B even with the
up-regulation of E2F1, which indicates Pin1 as a target of 9a independent of the requirement of
E2F1 down-regulation. However, notably, the Pin 1 expression remained unaffected in the
artemisinin-treated PC-3 cells (Fig. S 2F). The inhibition of Pin1 thus observed prompted us to
evaluate cyclin D1 level, as an overexpression of the latter has been linked to the progression and
development of cancer [44], as well as the fact that overexpression of Pin1 increases the cellular
cyclin D1 levels and the activation of its promoter [[45], [46]]. Indeed we found that a down-
regulation of cyclin D1 at the protein level occurred in a dose-dependent manner as depicted in
(
Fig. 7 B) with the inhibition of Pin1. This finding correlated well with a previous finding where

ACCEPTED MANUSCRIPT
the loss of Pin1 function in mouse mimicked the cyclin-D1-null phenotypes [47]. Moreover,
expression of cyclin D1 is also under the control of phosphorylated c-Jun [48]. Therefore we
next sought to check the expression of the phosphorylated c-Jun, which is one of the interacting
partners of Pin1 and also the transcription factor, for a number of genes involved in proliferation,
including cyclin D1 [[49], [50], [51]] . Indeed, it was found inhibited significantly at both the
concentration levels (3 µM and 10 µM) as shown in (Fig. 7 B). Thus, the modulation of
E2F1/Pin1/Phospho-c-Jun/cyclin D1 suggests the possible mechanism by which 9a is able to
inhibit proliferation of prostate cancer PC-3 cells.
In parallel with the inhibition of the above mentioned proliferating signals, the expression of
PCNA, a vital histologic marker for the G1/S phase in the cell cycle and whose expression level
has been suggested for an early diagnosis of prostate cancer [52], was examined by real-time
PCR as well as by western blot analysis. Indeed, it was found inhibited at both the transcript as
well as the protein levels in a dose-dependent manner as shown in (Fig. 7 B). With the previous
reports on other artemisinin derivatives such as DHA and artesunate down-regulating the c-Myc
expression [[53], [54]] in other cancer cell lines, we investigated the expression of c-Myc,
another proliferation regulator whose over expression is also one of the earliest molecular
alterations in prostate cancer [55]. c-Myc is a transcription regulator for many cell proliferation
genes including eIF4E [56], and the latter a component of mRNA 5’ cap binding protein
complex has been reported for aberrant expression in many cancers [57]. Both c-Myc and eIf4E
are oncogenes and have been reported to exhibit a cooperation in an Eµ-Myc transgenic model,
contributing to aggressive lymphomagenesis and leading to chemotherapy resistance [58].
Down-regulation of c-Myc led us to check the expression of eIF4E and it was indeed found
inhibited at 3 and10 µM of 9a as shown in (Fig. 7 B). Moreover, the down-regulation of cyclin
D1 could also be the consequence of inhibition of both c-Myc and eIF4E by 9a as the
expression of cyclin D1 has been reported to be eIF4E [56] as well as c-Myc-dependent [59].
In addition, we found the up regulation, at the transcript level as shown in (Fig. 7 A), of
p21(WAF1/Cip1) that functions as an effector of multiple anti-proliferative signals and acting in
both p53-dependent and p53-independent pathways and is mostly controlled at the transcription
level [60]. In our study, we found a significant up-regulation of p21 mRNA, which shows that 9a
up-regulates the level of p21 in p53 null PC-3 cells by some mechanism independent of p53. In
addition, we also found a significant down-regulation of cyclin B1 at the mRNA level which

ACCEPTED MANUSCRIPT
could be a consequence of the p21 up-regulation, as p21 checks the expression of cyclin B1 in
p53-dependent and independent ways by inhibiting CDK2 [[61], [62]]. The up-regulation of p21
at mRNA level and the down-regulation of PCNA at both the mRNA and the protein levels
suggest a strong anti-proliferative effect of 9a on the PC-3 cells as p21 competes for PCNA
binding with DNA polymerase δ and inhibits the DNA synthesis [63]. Investigating the apoptotic
effect of this derivative, a dose-dependent decrease of pro-caspase 9 and pro-caspase 3 was
observed as shown in Fig. 5 C, and in search of changes in another apoptotic marker, we found
an up-regulation of the pro-apoptotic Noxa at transcript level as in Fig. 7 A. Noxa is a BH3 only
protein known to activate pro-apoptotic Bak and Bax [64], and indeed we found an upregulation
of Bax at the protein level as in Fig. 5 C. On the other hand Noxa promotes the anti-apoptotic
Mcl-1 degradation [65],and although we did not check Noxa at the protein level, a study on
prostate cancer PC-3 cells treated with BH3 mimetic (-)-gossypol found an increase in the Noxa
protein without a significant change of the mRNA level [66]. In our study, we found a dose-
dependent increase in mRNA level of Noxa, which suggests an upregulation of Noxa protein,
and involvement in the apoptotic pathway.
We carried out siRNA-mediated knockdown of Pin1, alone and in the presence of 9a, and
evaluated the expression of Pin1 by western blotting. The level of Pin1 was significantly down-
regulated when the cells were treated with both the siRNA and Pin1 as compared to Pin1, siRNA
or 9a alone, as can be seen from Fig. 7 C, further showing that 9a inhibited Pin1. To further
validate Pin1 as one of the targets of 9a, PC-3 cells were pre-treated for 2 h with Wortmannin
(200 nM) and then with 10 µM of 9a for 72 h. Following incubation, the cell lysate was prepared
and subjected to western blotting to check the inhibition of Pin1 in the presence of autophagy
inhibitor Wortmannin. In the presence of Wortmannin with 9a absent, the Pin1 level was found
to be augmented whereas in the presence of both Wortmannin and 9a, it was found to be down-
regulated as in Fig. 7 C, which indicates that the inhibition of Pin1 is not autophagy-dependent
but follows some other inhibitory mechanism induced by 9a. This led us to investigate the
proteasome-mediated inhibition of Pin1, if any. For this we assessed the 9a-mediated inhibition
of Pin1 in the presence of proteasome inhibitor MG132 (100 nM). Treatment with 9a alone (3
and 10 µM) reduced the expression of Pin1 whereas pretreatment with MG132 followed by
treatment with 9a rescued Pin1 inhibition (Fig. 7 C), which suggest that proteasomal

ACCEPTED MANUSCRIPT
degradation of Pin1 is a predominant mode of its inhibition. In summary, the overall possible
mode of action by compound 9a is shown in Fig. 8.
Cell Proliferation
Apoptosis
Inhibition
Pin 1
Pro-caspase9
Pro-caspase3
Bcl-xL
E2F1
Bax
Noxa
p-c-Jun
c-Myc
eIF4E
PCNA
Effect on Migration
MMP-9
Cell Cycle Arrest
Cyclin D1
Cyclin B1
p21
Fig. 8 Possible mode of action of compound 9a
3
. CONCLUSIONS
Efficient synthesis of a library of dihydroartemisinin thiolactol ethers has been achieved. Their
structures were elucidated by standard spectroscopic techniques. It has been found that the
iodine-mediated method for the preparation of this class of compounds is best suited when thiols
to be used are alkane thiols, the N-containing thiols, in contrast, being less (or not at all) suitable
for the purpose. It is, however, highly α-stereoselective and therefore holds good potential for
application when the product desired is the neat α-thiolactol ether derivative. The BF .Et O-
3
2
mediated method is, on the other hand, more widely applicable and as a consequence a broad set
of thiols can be used for the preparation of the corresponding thiolactol ether derivatives of
dihydroartemisinin. The S-linked artemisinins reported in this work were proved highly effective
against various cancer cell lines as evident from the MTT assay. Compound 9a was chosen for
further mechanistic study and was found to be exhibiting its action partly by generating ROS and
partly by impairing the proliferating signals, thus affecting the cell viability of PC-3 cells and
causing their arrest in G1 phase that may have lead to apoptosis by activating pro-caspase-9 and
3
and upregulating Bax and ultimately to the cell death. In the attempts made to elucidate its
mode of action, it was found to be inhibiting Pin1, a pleiotropic regulator of various cellular

ACCEPTED MANUSCRIPT
processes and has been found three- to four-fold higher in prostate cancer cell lines. In addition,
we also found that the 9a-treated PC-3 cells have an inhibiting effect on Pin1-interacting partners
like cyclin D1, and phosphorylated-c-jun with parallel inhibition of c-Myc, eIF4E and PCNA
which further quells the proliferation signals. Along with hampering the viability of PC-3 cells,
9
a was also found affecting the migration potential of these highly metastatic cells. In
conclusion, this study presents the possible mechanisms by which artemisinin derivatives such as
9
a act as anti-proliferative, apoptotic as well as anti-metastatic agents for prostate metastatic PC-
3
cancer cells.
4
. EXPERIMENTAL SECTION
Materials and methods
All reagents and chemicals were purchased from Aldrich Chemical Co (Milwaukee, WI, USA)
or Alfa Aesar (a Johnson Matthey Company). TLC analyses were performed on 0.2 mm Merck
pre-coated silica gel 60 F254 aluminum sheets and the spots were visualized under UV lamp
and/or by immersion in an ethanolic solution of sulphuric acid (5%, v/v) followed by heating.
Final purification was performed using silica gel 200-400 mesh size. Melting points were
determined on a Buchi melting point (B-540) apparatus and are uncorrected. Specific rotations
were recorded on a Rudolph Autopol IV Polarimeter. NMR spectra were recorded on a Bruker
1
13
Avance DPX (400 MHz) spectrometer. H NMR and C NMR spectra were referenced to the
internal standard tetramethylsilane, in the respective deuterated solvents. Coupling constants (J)
1
1
are reported in Hertz. All assignments were confirmed with the aid of two-dimensional H- H
1
13
(
COSY) and/or H- C (HSQC) experiments using standard pulse programs. Processing of the
spectra was performed with Topspin software. IR spectra were recorded on a Nicolet FT-IR
Impact 410 instrument either as a thin film (neat) or as KBr pellet. Mass spectra were recorded
on MALDI (Bruker Daltonics, Ultraflex TOF/TOF) spectrometer and High-resolution mass
spectra (HRMS) on a Bruker Maxis spectrometer.
Dihydroartemisinin (2) [23]
Artemisinin (1, 2.0 g, 0.007 mol) was taken in distilled methanol (20 mL) and was cooled over
an ice bath to 0-5 °C. To the solution, an excess of sodium-borohydride was added in small

ACCEPTED MANUSCRIPT
portions during a period of 30 min. The solution was left to be stirred for 45 min at 0-5 °C after
which TLC (EtOAc/Hexane, 1:3) showed completion of the reaction After that, it was
neutralized by 5% cold AcOH/water to get a white precipitate of dihydroartemisinin (2). White
crystalline solid (yield, 1.91 g, 95%); IR (Neat) νmax 3369, 2924, 2854, 1446, 1377, 1174, 1134,
-1 1
1
092, 983, 875, 845, 783 cm ; H NMR (400 MHz, CDCl ) δ 5.60 (s, 1H, β H-12), 5.39 (s, 1H,
3 H
α H-12), 5.30 (d, 1H, J = 3.2 Hz, β H-10), 4.75 (d, 1H, J = 9.2 Hz, α H-10), 2.66-2.60 (m, 3H,
OH, β H-9), 2.42-2.29 (m, 3H, α H-9, α/β H-4a ), 2.07-2.01 (m, 2H, α/β H-4a ), 1.95-1.75 (m,
1
2
4
H, α/β H-5a , β H-8a , β H-8a ), 1.72-1.63 (m, 3H, α H-8a , α/β H-7a ), 1.58-1.46 (m, 4H, α/β
1 1 2 1 1
H-8a, α/β H-5a
2
), 1.44, 1.43 (2s, 6H, α/β H-14), 1.39-1.29 (m, 4H, α H-8a
2
, α/β H-6, β H-5a),
1
3
1
.25-1.22 (m, 1H, α H-5a), 1.05-0.94 (m, 14H, α/β H-7a , α/β H-15, α/β H-16); C NMR (100
2
MHz, CDCl
3
C
) δ 104.39, 104.13 (α/β C-3), 96.42, 94.63, 91.20, 87.77 (α/β C-12, α/β C-10),
8
6
1
1
3
1.12, 80.36 (α/β C-12a), 52.48, 51.52 (α/β C-5a), 45.42, 44.30 (α/β C-8a), 37.48, 37.37 (α/β C-
), 36.35, 36.26 (α/β C-4), 34.79, 34.69 (α/β C-7), 34.19, 30.77 (α/β C-9), 26.07, 25.97 (α/β C-
4), 24.71, 24.66 (α/β C-5), 24.57, 22.13 (α/β C-8), 20.38, 20.25 (α/β C-15), 13.19, 12.72 (α/β C-
+
6); MS (MALDI-TOF) for C H O , calculated m/z 284.16. Found m/z 307.65 [M+Na] ,
1
5
24
5
+
23.65[M+K] .
Dihydroartemisinin acetate (3) [23]
Method A: To a solution of dihydroartemisinin (2, 1.0 g, 3.5 mmol) in pyridine (12 mL) at room
temperature was added acetic anhydride (8 mL) and the solution was left at room temperature for
8
h. Completion of the reaction was ascertained by TLC (EtOAc/hexane, 1:3). It was then co-
evaporated thrice with toluene and was concentrated under reduced pressure. The resulting
mixture taken up in CH Cl was then successively washed, twice with cold dilute hydrochloric
2
2
acid (5%), saturated sodium bicarbonate solution, and finally brine. The organic portion was
collected and dried over anhydrous sodium sulfate and was then concentrated under reduced
pressure to give product 3 (yield, 1.06 g, 92%).
Method B: To a solution of 2 (1.0 g, 3.5 mmol) in acetic anhydride (2 mL), DABCO (200 mg,
1
.7 mmol) was added and the solution was stirred at room temperature. The reaction was
complete after 2 h as evidenced by TLC (EtOAc: Hexane, 1:3). The resulting mixture was
extracted with CH Cl and was successively washed, thrice with cold dilute hydrochloric acid
2
2
(5%), saturated sodium bicarbonate solution and finally with brine. The organic portion was

ACCEPTED MANUSCRIPT
collected and after drying over anhydrous sodium sulfate and was concentrated under reduced
pressure to give the product 3. White crystalline solid (yield, 1.03 g, 90%); IR (Neat) νmax 2927,
-1 1
2
869, 2086, 1752, 1634, 1453, 1375, 1228, 1130, 1103, 1029, 918, 877, 848, 825 cm ; H NMR
(
400 MHz, CDCl ) δ 5.78 (d, 1H, J = 9.9 Hz, H-10), 5.44 (s, 1H, H-12), 2.59-2.53 (m, 1H, H-
3
H
9
), 2.38 (dt, 1H, J = 3.9 Hz, J = 14.5 Hz, J = 13.5 Hz, H-4a ), 2.13 (s, 3H, COCH ), 2.08-2.00
1
3
(
m, 1H, H-4a
2
), 1.92-1.87 (m, 1H, H-5a
1
), 1.80-1.69 (m, 2H, H-8a
1
, H-7a
1
), 1.65-1.59 (m, 1H, H-
8
1
7
1
3
1
3
a), 1.52-1.45 (m, 1H, H-5a ), 1.44-1.40 (m, 4H, H-14, H-6), 1.38-1.32 (m, 1H, H-8a ), 1.31-
2 2
.26 (m, 1H, H-5a), 1.06-0.98 (m, 1H, H-7a ), 0.96 (d, 3H, J = 6.0 Hz, H-16), 0.85 (d, 3H, J =
2
1
3
.1 Hz, H-15); C NMR (100 MHz, CDCl
3 C 3
) δ 169.86 (COCH ), 104.68 (C-3), 91.85, 91.49 (C-
2, C-10), 80.13 (C-12a), 51.55 (C-5a), 45.22 (C-8a), 37.27 (C-6), 36.20 (C-4), 34.07 (C-7),
1.76 (C-9), 25.99 (C-14), 24.57 (C-5), 21.99 (C-8), 21.15 (C-15), 20.23 (COCH
3
), 12.09 (C-
+
6); MS (MALDI-TOF) for C H O , calculated m/z 326.17. Found m/z 349.51 [M+Na] ,
1
7
26
6
+
65.59 [M+K] .
General procedure for the synthesis of S-linked artemisinin derivatives having
aliphatic/aromatic/heterocyclic substitution on the C-10 sulfur: 4a [27], 4b [27], 5a, 5b, 10a,
1
0b, 11a, 12a, 12b, 13a, 13b
To a solution of dihydroartemisinin (2, 500 mg, 1.76 mmol) in dry CH
C, the desired thiol (aliphatic/aromatic or heterocyclic substituted, 1.2 mol equiv) was added.
BF .Et O (0.22 mL, 1 mol equiv) was added after 5 min of holding the solution stirred at -20 °C
2 2
Cl (10 mL) cooled to -20
°
3
2
and the stirring was continued at the same temperature until TLC showed the completion of the
reaction (10 to 20 min). The reaction mixture was diluted with CH Cl and was washed
2
2
3
successively with saturated aqueous NaHCO , ice-cold water, and brine. The organic layer was
dried over anhydrous sodium sulfate and was then concentrated to dryness under reduced
pressure. It was then chromatographed on silica gel column (230-400 mesh size) using
EtOAc/Hex as the eluent. The respective product was characterized by standard physical
methods.
General procedure for the synthesis of S-linked artemisinin derivatives having an aliphatic
2
7
27
substitution on the C-10 sulfur: 4a, 4b, 5a, 5b