Evaluating the effects of fluorine on biological properties and metabolic stability of some antitubulin 3-substituted 7-phenyl-pyrroloquinolinones

Article abstract of DOI:10.1016/j.ejmech.2019.05.092

A small number of fluorinated 7-phenyl-pyrroloquinolinone (7-PPyQ) derivatives was synthesized in an attempt to improve the metabolic stability of 3N-ethyl-7-PPyQ and 3N-benzoyl-7-PPyQ. The possible impacts of the fluorine-hydrogen isosterism on both biol

Full text of DOI:10.1016/j.ejmech.2019.05.092

Accepted Manuscript
Evaluating the effects of fluorine on biological properties and metabolic stability of
some antitubulin 3-substituted 7-phenyl-pyrroloquinolinones
Roberta Bortolozzi, Davide Carta, Matteo Dal Prà, Giuseppe Antoniazzi, Elena
Mattiuzzo, Mattia Sturlese, Veronica Di Paolo, Laura Calderan, Stefano Moro, Ernest
Hamel, Luigi Quintieri, Roberto Ronca, Giampietro Viola, Maria Grazia Ferlin
PII:
S0223-5234(19)30516-1
DOI:
https://doi.org/10.1016/j.ejmech.2019.05.092
EJMECH 11402
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 28 March 2019
Revised Date: 31 May 2019
Accepted Date: 31 May 2019
Please cite this article as: R. Bortolozzi, D. Carta, M.D. Prà, G. Antoniazzi, E. Mattiuzzo, M. Sturlese,
V. Di Paolo, L. Calderan, S. Moro, E. Hamel, L. Quintieri, R. Ronca, G. Viola, M.G. Ferlin, Evaluating
the effects of fluorine on biological properties and metabolic stability of some antitubulin 3-substituted
7-phenyl-pyrroloquinolinones, European Journal of Medicinal Chemistry (2019), doi: https://
doi.org/10.1016/j.ejmech.2019.05.092.
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
GRAPHYCAL ABSTRACT
Evaluating the effects of fluorine on biological properties and metabolic stability of
some antitubulin 3-substituted 7-phenyl-pyrroloquinolinones
1
Roberta Bortolozzi, Davide Carta , Matteo Dal Prà, Giuseppe Antoniazzi, Elena Mattiuzzo, Mattia Sturlese,
Veronica Di Paolo, Laura Calderan, Stefano Moro, Ernest Hamel, Luigi Quintieri, Roberto Ronca,
Giampietro Viola, Maria Grazia Ferlin

ACCEPTED MANUSCRIPT
Evaluating the effects of fluorine on biological properties and metabolic stability
of some antitubulin 3-substituted 7-phenyl-pyrroloquinolinones
2
1
1
1
2
Roberta Bortolozzi , Davide Carta , Matteo Dal Prà , Giuseppe Antoniazzi , Elena Mattiuzzo , Mattia
1
1
1
1
3
1
Sturlese , Veronica Di Paolo , Laura Calderan , Stefano Moro , Ernest Hamel , Luigi Quintieri ,
4
2
1
Roberto Ronca , Giampietro Viola , Maria Grazia Ferlin
1
2
3
. Department of Pharmaceutical and Pharmacological Sciences, University of Padova,
3
5131 Padova, Italy.
. Department of Woman’s and Child’s Health, University of Padova, Laboratory of
Oncohematology, 35128 Padova, Italy.
. Screening Technologies Branch, Developmental Therapeutics Program, Division of
Cancer Treatment and Diagnosis, Frederick National Laboratory for Cancer
Research, National Cancer Institute, National Institutes of Health, Frederick,
Maryland 21702, USA.
4
. Department of Molecular and Translational Medicine, Oncology and Immunology
Section, University of Brescia, 29881 Brescia, Italy.
Abstract
A small number of fluorinated 7-phenyl-pyrroloquinolinone (7-PPyQ) derivatives was
synthesized in an attempt to improve the metabolic stability of 3N-ethyl-7-PPyQ and 3N-
benzoyl-7-PPyQ. The possible impacts of the fluorine-hydrogen isosterism on both biological
activity and metabolic stability were evaluated. Introduction of a fluorine atom in the 2 or 3
position of the 7-phenyl ring yielded the 7-PPyQ derivatives 12, 13 and 15, which showed
potent cytotoxicity (low micromolar and sub-nanomolar GI s) both in human leukemic and
5
0
solid tumor cell lines. None of them induced significant cell death in quiescent and
1

ACCEPTED MANUSCRIPT
proliferating human lymphocytes. Moreover, 12, 13 and 15 exhibited remarkable cytotoxic
Vbl100
activity in the multidrug-resistant cell line CEM
, suggesting that they are not substrates
for P-glycoprotein. All compounds inhibited tubulin assembly and the binding of
3
[
H]colchicine to tubulin, with the best activity occurring with compound 15. Mechanistic
studies carried out on compound 12 indicated that it caused (a) a strong G2/M arrest; (b)
apoptosis in a time- and concentration-dependent manner; (c) a significant production of ROS
(
in good agreement with the observed mitochondrial depolarization); (d) caspase-3 and poly
ADP-ribose) polymerase activation; and (e) a decrease in the expression of anti-apoptotic
(
proteins. In vivo experiments in a murine syngeneic tumor model demonstrated that
compounds 12 and 15 significantly reduced tumor mass at doses four times lower than that
required for the reference compound combretastatin A-4 phosphate. Neither monofluorination
of the 7-phenyl ring of 3N-ethyl-7-PPyQ nor replacement of the benzoyl function of 3N-
benzoyl-7-PPyQ with a 2-fluorobenzoyl moiety led to any improvement in the metabolic
stability.
Keywords. Anti-tubulin, fluoro-phenylpyrroloquinolinone, apoptosis, molecular docking,
structure-activity relationships, metabolic stability
1
. Introduction
The introduction of fluorine atoms in organic molecule drug candidates has become a
common practice in modern small-molecule drug discovery. Indeed, the presence of the small
and highly electronegative fluorine is capable of modulating various molecular properties,
including pharmacological potency, physicochemical features and pharmacokinetics [1]. In
particular, replacement of hydrogen with fluorine in a compound subject to cytochrome P450
2

ACCEPTED MANUSCRIPT
(CYP)-catalyzed oxidative metabolism has been frequently successful in decreasing the rate
of metabolic elimination of a chemical entity, thereby improving its oral bioavailability and
prolonging its in vivo half-life [2]. Furthermore, the substitution of fluorine in place of
hydrogen has become an approach commonly used by the medicinal chemist in an attempt to
improve the binding affinity of a bioactive compound to a target protein and thus,
conceivably, to increase its potency and target selectivity [3-6].
Tricycle phenylpyrroloquinolinones (PPyQs) are a class of compounds that have shown
interesting in vitro and in vivo antiproliferative properties acting as tubulin polymerization
inhibitors by binding at the colchicine site in β-tubulin. In particular, 7-PPyQs have shown
very high affinity for the colchicine site and high potency in inhibiting tubulin polymerization
comparable to that of reference compounds, such as combretastatin A-4 (CA-4)[7-11].
Structural optimization efforts produced highly anti-tumorigenic PPyQs that showed
nanomolar and subnanomolar GI s towards a broad spectrum of human tumor cell lines, with
5
0
induction of tumor cell death by an apoptotic mechanism. These compounds also overcome
resistance to taxol and vincristine, have limited cytotoxicity in non-tumor cell lines and act
synergistically with conventional chemotherapeutic agents in inhibiting leukemic cell
proliferation [10]. Among the synthesized compounds, 3N-ethyl-7-PPyQ (20) [9] and 3N-
benzoyl-7-PPyQ (21) [11], as representative of 3N-alkyl and 3N-acyl analogues of 7-PPyQs,
respectively (Figure 1), had shown the best antiproliferative profile in their respective series.
In agreement with tubulin polymerization inhibition assay data, docking simulations with
compounds 20 and 21 into tubulin suggested that they have a high affinity for the colchicine
site (as shown for compound 21 in Figure 1) and that their binding mode is compatible with a
competitive mechanism of action [11].
3

ACCEPTED MANUSCRIPT
docking pose of compound 21
3N-ethyl-7-PPyQ 20
3
N-benzoyl-7-PPyQ 21
Figure 1. Molecular docking simulation of compound 21 in the colchicine site of β-tubulin
and structures of 7-PPyQs 20 and 21.
Previous in vitro studies in human liver microsomes (HLMs) demonstrated that the amide 3N-
benzoyl-7-PPyQ (21) is quite stable to metabolism by NADPH-dependent enzymes (i.e., CYPs
and flavin monooxygenases, [6]), but it was, however, susceptible to slow microsomal
hydrolysis to 3N-unsubstituted 7-PPyQ [10]. In contrast, 3N-ethyl-7-PPyQ (20) was relatively
unstable when incubated with NADPH-supplemented HLMs (0.5 mg of protein/mL), with
less than 70% compound remaining after a 30-min incubation at 37°C (Di Paolo, unpublished
data). Based on these findings, and in order to possibly improve the metabolic stability and
anticancer properties of 7-PPyQs, we recently synthesized a small library of untargeted
monofluoro-phenyl derivatives of compounds 20 and 21. In particular, based on data from a
metabolite profiling experiment indicating conversion of 20 to oxygenated metabolites by
HLMs (see Results section), and, being aware that aromatic carbon oxygenation is a quite
common CYP-catalyzed reaction [6] a fluorine atom was introduced in the ortho, meta or
para position of the 7-phenyl ring of compound 20 giving compounds 12-14. Moreover, in an
attempt to further improve the stability of amide 21 to hydrolytic metabolism, we replaced its
benzoyl function with a 2-fluorobenzoyl moiety, thus obtaining compound 19. Finally, to get
more insight into the possible impact of fluorination on the biological properties of PPyQs,
4

ACCEPTED MANUSCRIPT
our library was completed with the PPyQ derivatives 15 and 18. Since fluorination might
influence both therapeutic efficacy and target affinity [5], we evaluated the synthesized
fluorinated PPyQs both in vitro and in vivo by biological assays aimed at characterizing their
anticancer properties. Furthermore, molecular modelling studies and metabolic stability
experiments were performed with compounds 12-15 and 19. The results of these
investigations are presented and discussed in this paper.
2
. Results and discussion
2
.1 Chemistry
The 4-step method leading to 7-PPyQ compounds was previously described [8] (Scheme 1).
First, commercially available 5-nitroindole was subjected to an N-alkylation and acylation
reactions using appropriate halogenated compounds to obtain the nitroindole derivatives 1-3.
Compound 1 was obtained using ethyl bromide in anhydrous DMF in the presence of NaH at
room temperature, while for compounds 2 and 3, a reaction with benzoyl chloride or 2-F
benzoyl chloride, respectively, in presence of DMAP and pyridine at room temperature for 24
h, was preferred. The 2-fluorobenzoyl chloride (17) was prepared by reacting 2-fluorobenzoic
acid with SOCl at reflux. The catalytic reduction (Pd/C 10%, H at atmospheric pressure,
2
2
ethyl acetate) of intermediate 1 and (Pt/C, H at atmospheric pressure, ethyl acetate) of
2
intermediates 2 and 3 gave the corresponding aminoindoles 4-6 in almost quantitative yields.
Compound 4 was reacted with commercial ethyl-2-, -3- and -4-fluoro benzoyl acetate and
compound 5 with ethyl-3-fluoro-benzoyl acetate, whereas compound 6 with ethyl benzoyl
acetate. All the reactions were carried out in absolute ethanol at reflux in the presence of a
catalytic amount of acetic acid, yielding the acrylate derivatives as crude material 7-11. These
were purified by silica gel column chromatography before being submitted to thermal
cyclization in boiling diphenyl ether (250 °C) to obtain the final products 12-15 in good
5

ACCEPTED MANUSCRIPT
yields. Compound 11 cyclized but with loss of the 2-fluoro-benzoyl moiety, giving the known
7
-PPyQ 16.[7] For this reason, in order to obtain the 3N-2-fluorobenzoyl 7-PPQ, compound
6 was directly reacted with 2-fluorobenzoyl chloride 17 in DMF in the presence of NaH,
1
obtaining two products: the main reaction product was identified as the isomeric ester 18
78%), and the desired 3-amide compound 19 was obtained in lower yield (22%).
(
Scheme 1. a) Br-ethane, NaH 60%, DMF, 50°C, 6h, 99% (1); benzoyl chloride, pyridine,
DMAP, rt, 24h, 54% (2), or 2-F-benzoyl chloride, pyridine, DMAP, rt, 24h, 62% (3) b) H2,
Pd/C 10%, EtOAc, 50 °C, 12h, 98% (4) or H , Pt/C 15%, EtOAc, rt, 2 h, 84% (5, 6); c) ethyl
2
(2-fluorobenzoyl) acetate (7), ethyl (3-fluorobenzoyl) acetate (8, 10), ethyl (4-fluorobenzoyl)
acetate (9), ethyl benzoyl acetate (11), absolute ethanol, CH COOH cat, drierite, refluxing,
3
2
4-48 h, 35-70%; d) diphenyl ether, 250 °C, 15 min, 27-56%; e) SOCl , DCM, 80 °C,
2
refluxing, 2 h, 80%.
O N
O N
H N
2
2
a
2
b
c
N
H
N
R
N
R
1-3
4-6
O
O
H
R1
O
HN
N
R1
R
d
R
R1
N
H
1, 4
C H
2 5
benzoyl
2-F-benzoyl
--
--
--
N
R
2
, 5
3
7
8
9
1
, 6
7
-11
12-15
, 12
, 13
, 14
C H
2-F-phenyl
3-F-phenyl
4-F-phenyl
3-F-phenyl
phenyl
2
5
5
5
C H
2
C H
2
0, 15 benzoyl
2-F-benzoyl
COOH
F
COCl
F
e
11
17
F
F
O
O
O
O
NH
NH
N
a
O
N
N
N
H
H
16
18
19
6

ACCEPTED MANUSCRIPT
2
.2 Biological evaluation
2
.2.1 Metabolic pattern of compound 20 in human liver microsomes
An exploratory in vitro drug metabolism experiment was conducted to get preliminary
information on the possible metabolic fate of the known compound 20 (Figure 1) in humans.
The HPLC trace of compound 20 incubated with NADPH-supplemented HLMs demonstrated
three major new fluorescence peaks (Figure 2), which were undetectable when NADPH, an
essential cofactor for both CYP- and flavin monooxygenase-mediated oxidations [6], was
omitted from the incubation mixture (not shown). Compound 20 constituted a peak eluting at
1
2.9 min, while the three compound 20 metabolites (M1-M3) eluted at 11.2, 12.3 and 14.7
min, respectively; the retention time of M1 corresponded exactly to that of authentic 3N-
unsubstituted 7-PPyQ (scheme 1, compound 16), a recently identified metabolite arising from
hydrolysis of 3N-benzoyl-7-PPyQ (compound 21) catalyzed by HLMs [11]. Fractions
corresponding to the M1-M3 and compound 20 peaks were collected and subjected to positive
ion ESI-mass spectrometry analysis. As expected, the mass spectrum of peak M1 was
compatible with that of authentic 7-PPyQ 16 and showed a prominent protonated molecule
+
ion, [M+H] , at m/z 261.0996. These findings demonstrated that compound 20 underwent
oxidative N-demethylation catalyzed by a human liver microsomal enzyme(s) to form 7-
+
PPyQ. The material in peak M2 showed a [M+H] at m/z 277.0930, which was 12.0419 Da
lower than that of protonated parent compound 20 (measured m/z = 289.1349). This finding
suggested the loss of C H and the introduction of an oxygen atom into the parent molecule,
2
4
i.e., deethylation and oxygenation of compound 20. Finally, the mass spectrum scan of the
+
most prominent chromatographic peak, M3, showed a [M+H] ion at m/z 305.1273, which
7

ACCEPTED MANUSCRIPT
was 15.9924 Da higher than that of 20, suggesting that an oxygen atom has been introduced
into the parent drug (oxygenation of compound 20).
Figure 2. HPLC separation of compound 20 and its human liver microsomal metabolites M1-
M3. HLMs (0.25 mg/mL) were incubated with 10 µM compound 20 and NADPH (1 mM) at
3
7 °C for 60 min. The sample was processed and analyzed by HPLC-fluorescence as
described in Materials and Methods. A peak of unmodified compound 20 and three major
peaks corresponding to metabolites M1-M3 are present. All metabolite peaks were absent
when NADPH was omitted from the reaction mixture. The reported m/z values were obtained
+
from ESI-mass spectra of collected peaks and refer to [M+H] ions.
2
.2.2 Metabolic stability of compounds 12-15 and 19-21 in human liver microsomes
The metabolic stability in HLMs of fluorinated 7-PPyQs 12-14 was compared with that of
their non-fluorinated counterpart, compound 20. Like 20, compounds 12-14 were quite stable
when incubated with HLMs in phosphate buffer (pH 7.4) or phosphate buffer only, but time-
dependently disappeared when both microsomes and NADPH were contained in the mixture
8

ACCEPTED MANUSCRIPT
(Figure 3). In particular, the depletion half-lives in mixtures containing HLMs and NADPH
were 11.4, 11.8, 10.6 and 17.3 min for 12, 13, 14 and 20, respectively. Therefore, introduction
of fluorine in the ortho, meta or para position of the 7-phenyl ring failed to decrease but
instead slightly increased the metabolic lability of compound 20. These findings may suggest
that the 7-phenyl ring of compound 20 is not a site of carbon oxygenation by a NADPH-
dependent enzyme(s) in HLMs. On the other hand, as stressed by Obach et al. [12], for
compounds metabolized at multiple sites, as in the case of 20, fluorination at a single site may
only result in a “metabolic switching”, with no decrease in the overall rate of metabolic
elimination. Moreover, introduction of fluorine at a metabolically labile site can also
unexpectedly increase the rate of metabolism by CYP enzymes, as observed for substitution
of fluorine in place of hydrogen at the tolyl methyl group of celecoxib or at position 2 of
ramelteon [12]. Proper interpretation of the results of this set of experiments thus requires a
more in-depth metabolic profiling (i.e., identification of the metabolic “hot spots”) of
compound 20.
HLMs + NADPH
HLMs
Buffer only
1
00
1
00
8
6
4
0
0
0
compound 13
compound 12
8
6
4
2
0
0
0
0
0
20
0
0
5
10
15
20
25
30
0
5
10
15
20
25
30
1
00
100
80
60
40
20
0
compound 20
8
6
4
2
0
0
0
0
0
compound 14
0
5
10
15
20
25
30
0
5
10
15
20
25
30
Incubation time (min)
9

ACCEPTED MANUSCRIPT
Figure 3. Assessment of metabolic stability of 12, 13, 14 and 20 in human liver microsomal
incubations. Each compound (10 µM) was incubated at 37 °C with HLMs (0.5 mg/mL; -∆-),
HLMs plus 1 mM NADPH (-■-), or buffer only (0.1 M KH PO , pH 7.4; -□-). Incubation
2
4
time points were 0, 2, 5, 15 and 30 min (“HLMs”), or 0 and 30 min (“HLMs plus NADPH”
and “buffer only”). The data are expressed as percentage of parent compound remaining at
each time compared with time 0 min and represent the mean ± SD of n = 3 independent
determinations. Error bars smaller than the symbols are not shown.
Subsequent experiments examined the impact of quercetin, a potent inhibitor of various
human drug-metabolizing CYP enzymes, including CYP1A2, CYP2C9, CYP2D6 and
CYP3A4 [13-15] on the metabolic stability of compounds 12-14 and 20 in NADPH
supplemented-HLMs. As shown in Figure 4, metabolic depletion of all the investigated 7-
PPyQ derivatives was significantly inhibited by 100 µM quercetin. These findings indicate
the involvement of one or more human liver microsomal CYPs in the metabolism of
compounds 12-14 and 20.
compound 12
compound 13
1
00
**
100
80
60
40
20
0
*
*
8
6
4
2
0
0
0
0
0
HLMs
NADPH
Quercetin
+
+
+
+
+
-
+
+
+
+
+
-
compound 14
compound 20
1
00
***
100
80
60
40
20
0
*
8
6
4
2
0
0
0
0
0
HLMs
NADPH
Quercetin
+
+
+
+
+
-
+
+
+
+
+
-
1
0

ACCEPTED MANUSCRIPT
Figure 4. Effect of the CYP inhibitor quercetin on depletion of 12, 13, 14 and 20 in NADPH-
supplemented human liver microsomal incubations. Each compound (10 µM) was incubated
with HLMs (1 mg/mL) and NADPH (1 mM) at 37 °C for 0 and 15 min, both in the absence
and in the presence of 100 µM quercetin. Each bar represents the percentage (mean ± SD; n =
3
) of compound remaining at 15 min, compared with time 0 min. *P<0.05; **P<0.01;
**P<0.001
*
A last set of drug metabolism experiments compared the metabolic stability in HLMs of the
monofluorinated 7-PPyQs 15 and 19 with that of their non-fluorinated counterpart 3N-
benzoyl-7-PPyQ (21), an amide compound known to be quite stable to NADPH-dependent
metabolism, despite being susceptible to slow hydrolysis catalyzed by HLMs [12]. As shown
in Figure 5, for all the tested compounds, a 60-min incubation with HLMs (1 mg of
protein/mL) resulted in a modest decline (~ 20%) in the level of compound remaining,
regardless of the presence of NADPH in the incubation mixture. Thus, fluorine substitution at
the meta position of the 7-phenyl ring or the ortho position of the benzoyl moiety did not
modulate the metabolic stability of compound 21.
1
1

ACCEPTED MANUSCRIPT
1
00
7
5
2
5
0
5
0
HLMs
-
-
+
+
+
-
NADPH
1
00
7
5
2
5
0
5
0
HLMs
-
+
+
NADPH
-
+
-
1
00
7
5
2
5
0
5
0
HLMs
-
-
+
+
+
-
NADPH
Fig. 5. Assessment of metabolic stability of 15, 19 and 21 in human liver microsomal incubations.
Each compound (10 µM) was incubated at 37 °C with HLMs (1 mg/mL), HLMs plus 1 mM NADPH,
or buffer only (0.1 M KH PO , pH 7.4) for 0 or 60 min. Each bar represents the percentage (mean ±
2
4
SD; n = 3) of compound remaining at 60 min, compared with time 0 min.
2
.2.3 In vitro antiproliferative activity of compounds 12-15, 18 and 19
Evaluation of antiproliferative activities of 12-15, 18 and 19 was performed with the 3-(4,5-
dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay [16] against a panel of
1
1 human tumor cell lines (DND41, Jurkat, HL-60, RS4;11, MCF-7, MDA-MB-231, MDA-
MB-468, HeLa, A375, A2058, HT-29). GI values, the concentrations that inhibit cell growth
5
0
1
2

ACCEPTED MANUSCRIPT
by 50%, are presented in Table 1. From the cytotoxicity data reported in Table 1, we can
make some observations on fluoro-derivatives 12-15, 18 and 19 in comparison with the parent
compounds 20 and 21 lacking the fluorine and previously evaluated, as well as on the
substitution pattern of the fluorine atom in the 7-PPyQ structure. The N-ethyl compounds 12
and 13 have the fluorine atom in the ortho and meta position, respectively, and they are the
most cytotoxic compounds, with nano and sub-nanomolar concentrations. These compounds
are comparable with the reference compound 20. Analog 14, having the fluorine in the para
position, is less active with micromolar or high nanomolar GI s. Between the two N-benzoyl
5
0
compounds 15 and 19, the derivative 15 with a fluorine in position 3 of the side phenyl is
much more cytotoxic than derivative 19 with a fluorine in the ortho position of the benzoyl
side chain. Finally, among all the fluoro-derivatives, compound 18, which is unsubstituted at
the pyrrolic N and with a fluoro-benzoyl ester in position 9, is the least active. We note that
the benzoyl fluoro-derivatives 15 and 19 are less cytotoxic than the reference compound 21.
In conclusion, based on the antiproliferative data shown in Table 1, the preferred substitution
sites for a fluorine atom are the meta and ortho positions of the side phenyl ring, as in
compounds 12, 13 and 15.
1
3

ACCEPTED MANUSCRIPT
Table 1. In vitro cell growth inhibitory effects of compounds 12-15, 18-21.
a
GI50 (nM)
MDA-MB-
31
MDA-MB-
468
Cd
DND-41
Jurkat
HL-60
RS4;11
MCF-7
HeLa
A375
A2058
HT-29
2
1
1
1
1
1
1
2
2
2
3
4
5
8
9
0
1
0.26±0.088
0.13±0.020
476.7±128.1
10.2±0.160
2425±425
240.0±15.3
nd
0.19±0.01
0.11±0.02
333.5±12.1
32.4±2.2
323.5±29.1
53.7±20.2
0.5±0.2
0.17±0.07
0.08±0.002
1440±42.4
15.1±0.090
2307±77
0.09±0.005
0.07±0.006
293.3±61.7
56±12
246.6±12.0
52.7±10.7
51.3±9.9
103±50
164±43.7
69±6.0
22.3±2.2
15.0±1.1
43.0±12.6
871±60
1318±68.4
70±15.7
nd
1.2±0.1
0.7±0.05
340±18.4
13.3±0.9
501±34.4
9.3±1.1
15.5±1.5
11.4±2.0
69.8±5.9
n.d.
4.8±0.2
3.5±0.4
1180±90.7
n.d.
1.1±0.2
0.9±0.1
447±12.0
402±37
216±17.1
26.6±2.0
32±1.2
206±8.3
454±48
5591±500.4
3031±340.4
nd
696.7±72.6
119.7±25.4
2±0.3
3464±231.5
1082±24.9
40±10
2553±139.6
332±13.6
nd
1493±127
322±12.5
Nd
276.0±20.9
0.5±0.02
11±8
nd
16±0.6
2±0.8
0.3±0.0001
0.2±0.0.1
nd
nd
0.2±0.0.4
nd
Nd
0.1±0.1
a
GI = compound concentration required to inhibit tumor cell proliferation by 50%. Data are presented as the mean ± SE from the dose-response
5
0
curves of at least three independent experiments. N.d. not determined.
.
1
4

ACCEPTED MANUSCRIPT
2
.2.4 Evaluation of cytotoxicity of compounds 12, 13, 15, 20 and 21 in human non-cancer
cells
To obtain a preliminary indication of the cytotoxic potential of these derivatives in normal
human cells, three of the most active compounds (12, 13 and 15) along with compounds 20
and 21 chosen for comparison as non-fluorinated compounds, were evaluated in vitro against
peripheral blood lymphocytes (PBL) from healthy donors (Table 2). All compounds showed
very low activity in quiescent lymphocytes (GI >10 µM), while in the presence of the
5
0
mitogenic stimulus phytohematoaglutinin (PHA), the GI values were slightly decreased but
5
0
remained in the higher micromolar range. These values were almost 10000-30000 times
higher than those observed against the lymphoblastic cell lines Jurkat, DND-41 and CEM
(
Tables 1 and 3). Thus, these compounds have only a modest inhibitory effect in primary
lymphocytes, as previously observed for other antimitotic derivatives developed by our group
10].
[
Table 2. Cytotoxicity of compounds 12, 13, 15, 20 and 21 for human peripheral blood
lymphocytes (PBL)
a
GI (µM)
5
0
b
c
Cd
PBLresting
PBLPHA
1
1
1
2
2
2
3
5
0
1
28.0±2.3
15.2±6.9
10.8±3.8
17.6±3.9
7.6±0.06
2.7±1.3
31.3±10.3
45.6±6.3
>100
>100
a
Compound concentration required to inhibit cell growth by 50%.
Values are the mean ± SEM for three separate experiments
b
PBL not stimulated with PHA.
1
5

ACCEPTED MANUSCRIPT
c
PBL stimulated with PHA.
2
.2.5 Effect of compounds12, 13, 15, 20 and 21 on multidrug resistant cells
To investigate whether these derivatives are substrates of drug efflux pumps, compounds 12,
Vbl-100
1
3 and 15 were tested against the CEM
cell line that is a multidrug-resistant line selected
against vinblastine [17] and that overexpress P-glycoprotein (P-gp). This membrane protein
acts as a drug efflux pump and exhibits resistance to a wide variety of structurally unrelated
anticancer drugs and other compounds. As shown in Table 3, all the compounds exhibited
Vbl100
cytotoxic activity in the CEM
cell line that was even higher than their activity against the
parental line, indicating that these derivatives are not substrates for P-gp.
1
6

ACCEPTED MANUSCRIPT
Table 3. Cytotoxicity of 12, 13, 15, 20 and 21 in multidrug resistant cell lines
a
GI (nM)
5
0
wt
Vbl100
Cd
CEM
CEM
12
13
15
20
21
3.1±0.2
2.2±0.3
44±6
1.1±0.4
0.8±0.3
36±10
14.0±0.4
2.5±0.6
4.0±0.4
3.1±0.4
a
Compound concentration required to reduce cell
growth by 50%. Values are the mean ± SEM for three
separate experiments.
2
.2.6 Inhibition of tubulin polymerization and colchicine binding
To evaluate the antitubulin properties of the new compounds, we investigated their effects on
3
the inhibition of tubulin polymerization and on the binding of [ H]colchicine to tubulin (Table
4
) [18, 19]. For comparison, CSA-4 was examined in contemporaneous experiments as a
reference compound. Among the test compounds, 12, 13 and 15 strongly inhibited tubulin
assembly with IC values below 1 µM (0.96, 0.84 and 0.38 µM, respectively), with 12 and 13
5
0
slightly less active than CSA-4 (IC =0.64 µM) and compound 15 more active than CSA-4.
5
0
These results with tubulin correlate well with the growth inhibitory effects exhibited by these
compounds, indicating that their antiproliferative activity derives from an interaction with
tubulin. The other compounds also inhibited tubulin assembly, with IC values in the range
5
0
3
of 1.3-3.7 µM. All compounds inhibited the binding of [ H]colchicine to tubulin, with the best
1
7

ACCEPTED MANUSCRIPT
activity occurring with 12, 13 and 15, but none approached CSA-4 in its potency as an
inhibitor of colchicine binding.
Table 4. Inhibition of tubulin polymerization and colchicine binding by compounds 12-15
and 19-21
Inhibition of tubulin assembly
Inhibition of colchicine binding
Cd
a
b
IC (µM) ± SD
% Inhibition ± SD
5
0
1
1
1
1
1
2
3
4
5
9
0.96±0.08
77±3
83±5
0.78±0.03
3.7± 0.4
12±4
0.38±0.03
1.3±0.07
0.64±0.1
0.57±0.02
0.89±0.04
1.2±0.1
70±0.8
34±4
CSA-4
100±2
73±0.7
70±0.20
98±0.7
c
20
d
21
c, d
CSA-4
a Inhibition of tubulin polymerization. Tubulin was at 10 µM.
3
b Inhibition of [ H]colchicine binding. Tubulin and colchicine were at 1 and 5 µM
concentrations, respectively.
c Data taken from reference [10]
d Data taken from reference [11]
1
8

ACCEPTED MANUSCRIPT
2
.2.7 Influence of test compound 12 on the cell cycle
The effect of compound 12, taken as representative of a fluoro-PPyQ derivative, on cell cycle
progression was examined by flow cytometry in two melanoma cell lines (A375 and A2380).
After a 24 h treatment, 12, evaluated at the concentrations of 50 and 100 nM, induced a strong
G2/M arrest at both concentrations in both cell lines (Figure 6, Panels A and B). A
concomitant reduction of both the S and G1 phases was also observed. In order to determine
whether 12 was able to block cells at the mitotic phase (M), cells were stained with an
immunofluorescent antibody to p-histone H3, a well-known mitotic marker [20], as well as
PI, and analyzed by flow cytometry. As shown in Figure 6 (Panel C), HeLa cells arrested in
M phase by treatment with 12 are readily distinguished from G2 cells by the higher level of p-
histone H3. In particular, treatment with 12 induced an increase in the percentage of mitotic
cells from 1.5% observed in untreated cells to about 10% and 44 % with 50 and 100 nM
concentrations, respectively. We also studied the association between 12-induced G2/M arrest
and alterations in G2/M regulatory protein expression in HeLa cells. As shown in Figure 7
compound 12 caused an increase in cyclin B1 expression after 24 and 48 h in a concentration-
dependent manner, indicating an activation of the mitotic checkpoint following drug
exposure.
This effect was confirmed by a reduction in the expression of phosphatase cdc25c at 24 h,
followed by a disappearance in its expression at 48 h. This was associated with the
appearance of slower migrating forms of phosphatase cdc25c indicating its phosphorylation.
The phosphorylation of cdc25c directly stimulates its phosphatase activity, and this is
necessary to activate cdc2/cyclin B on entry into mitosis [21, 22]. We also observed a
decrease of the phosphorylated form of cdc2 kinase, in particular after the 48 h treatment.
1
9

ACCEPTED MANUSCRIPT
A
B
C
Figure 6. Percentage of cells in each phase of the cell cycle in A375 (Panel A) and A2058
Panel B) cells treated with compound 12 at the indicated concentrations for 24 h. Cells were
fixed and labeled with PI and analyzed by flow cytometry as described in the Experimental
(
2
0

ACCEPTED MANUSCRIPT
Section. Data are represented as mean of two independent experiments ± SEM. Panel C.
Mitotic index evaluated in Hela cells with compound 12 after a 24 h treatment with 12 at the
indicated concentrations.
2
.2.8 Compound 12 induces apoptosis in Hela cells
To evaluate the mode of cell death induced by test compounds, we performed a bi-parametric
cytofluorimetric analysis using propidium iodide (PI) and annexin-V-FITC, which stain DNA
and phosphatidylserine (PS) residues, respectively [23, 24]. We used Hela cells to evaluate
these effects of compound 12. As shown in Figure 8, the compound induced apoptosis in a
time and concentration dependent manner, even at the lower concentration used (50 nM).
Figure 7. Effect of compound 12 on cell cycle checkpoint proteins. HeLa cells were treated
for 24 or 48 h with the indicated concentrations of 12. The cells were harvested and lysed for
detection of the expression of the indicated protein by western blot analysis. To confirm equal
protein loading, each membrane was stripped and reprobed with anti-γ-actin antibody.
2
1

ACCEPTED MANUSCRIPT
2
.2.9 Compound 12 induced mitochondrial depolarization and reactive oxygen species (ROS)
production
Mitochondria play an essential role in the propagation of apoptosis [25, 26]. It is well
established that, at an early stage, apoptotic stimuli alter the mitochondrial transmembrane
potential (∆ψ ). ∆ψ was monitored by the fluorescence of the dye JC-1 [27]. HeLa cells
mt
mt
treated with compound 12 (10-100 nM) showed a time-dependent increase in the percentage
of cells with low ∆ψ (Figure 9, Panel A). The depolarization of the mitochondrial membrane
mt
is associated with the appearance of annexin-V positivity in the treated cells when they are in
an early apoptotic stage. In fact, the dissipation of ∆ψ is characteristic of apoptosis and has
mt
been observed with both microtubule stabilizing and destabilizing agents, including other 7-
PPyQ derivatives, in different cell types [28-31]. It is well known that mitochondrial
membrane depolarization is associated with mitochondrial production of ROS [32, 33].
Therefore, we investigated whether ROS production increased after treatment with the test
compounds. We utilized the fluorescence indicator 2,7-dichlorodihydrofluorescein diacetate
(
H -DCFDA)[33]. As shown in Figure 9 (Panel B) compound 12 induced significant
2
production of ROS starting after 12-24 h of treatment at 100 nM, in good agreement with the
mitochondrial depolarization described above.
A
2
2

ACCEPTED MANUSCRIPT
B
Figure 8. Flow cytometric analysis of apoptotic cells after treatment of HeLa (Panels A and
B) cells with compound 12 at the indicated concentrations after incubation for 24 (A) or 48 h (B).
The cells were harvested and labeled with annexin-V-FITC and PI and analyzed by flow
cytometry. Dual staining for annexin-V and with PI permits discrimination between live cells
-
-
+
-
+
+
(
annexin-V /PI ), early apoptotic cells (annexin-V /PI ), late apoptotic cells (annexin-V /PI )
-
+
and necrotic cells (annexin-V /PI ). Data are represented as mean±SEM of three independent
experiments.
2
3

ACCEPTED MANUSCRIPT
A
B
Figure 9. Panel A. Assessment of mitochondrial membrane potential (∆ψ ) after treatment of
mt
HeLa cells with compound 12. Cells were treated with the indicated concentrations of
compound for 6 , 12 , 24 and 48 h and then stained with the fluorescent probe JC-1 for
analysis of mitochondrial potential. Cells were then analyzed by flow cytometry as described
in the experimental section. Data are presented as mean±SEM of three independent
experiments. Panel B. Assessment of ROS production after treatment of HeLa cells with
compound 12. Cells were treated with the indicated concentrations of compound for 6, 12, 24,
and 48 h and then stained with H -DCFDA for the evaluation of ROS levels. Cells were then
2
analyzed by flow cytometry as described in the experimental section. Data are represented as
mean±SEM of three independent experiments.
2
4

ACCEPTED MANUSCRIPT
2
.2.10 Compound 12 induced caspase-3 and poly (ADP-Ribose) polymerase (PARP)
activation and caused a decrease in the expression of anti-apoptotic proteins
As shown in Figure 10, compound 12 in HeLa cells caused a concentration and time-
dependent increase of the cleaved fragment of caspase-3 and concomitantly the cleavage of
PARP, confirming its pro-apoptotic activity.
We also investigated the expression of anti-apoptotic proteins, such as Bcl-2 and Mcl-1. Bcl-2
plays a major role in controlling apoptosis through the regulation of mitochondrial processes
and the release of mitochondrial proapoptotic molecules that are important for the cell death
pathway [35-37]. Our results (Figure 10) showed that the expression of the anti-apoptotic
protein Bcl-2 was decreased starting after a 24 h treatment at both concentrations of 12 used
(10 and 100 nM). The decrease in expression of Mcl-1 was even greater.
Interestingly, as observed for other antimitotics, we observed after treatment with 12 a
significant increase in the expression of the phosphohistone H2AX, a well-known marker of
DNA damage. In this context, it is worthwhile noting that prolonged mitotic arrest may
induce DNA damage that ultimately leads to apoptosis.
2
5

ACCEPTED MANUSCRIPT
Figure 10. Western blot analysis of Bcl-2, Mcl-1 and PARP after treatment of HeLa cells
with 12 at the indicated concentrations and for the indicated times. To confirm equal protein
loading, each membrane was stripped and reprobed with anti-γ-actin antibody.
2
.2.11 Evaluation of the antitumor activity of compounds 12 and 15 in vivo.
To determine the in vivo antitumor activity of 12 and 15, a syngeneic murine model was used
38], the BL6-B16 mouse melanoma cell line. Preliminary in vitro experiments were carried
out to evaluate the cytotoxicity of these compounds in this cell line. The GI obtained after
[
5
0
7
2 h was 10 ± 4 nM and 18 ± 6 nM, for 12 and 15, respectively.
BL6-B16 cells were injected subcutaneously in syngeneic C57BL/6 mice. Once the tumor
3
reached a measurable size (about 100 mm ), mice were randomly assigned to different
experimental groups and treated intraperitoneally every other day with vehicle (DMSO) or
one of the two compounds (at doses of 7.5 or 3.0 mg/kg) or with combretastatin A4 phosphate
(CA4P) ((at 30 mg/kg) as a reference compound.
As shown in Figure 11 (Panel A), both compounds 12 and 15 caused a similar and significant
reduction in the growth of BL6-B16 melanoma cells at the dose of 7.5 mg/kg (60% for 12 and
2
6

ACCEPTED MANUSCRIPT
5
8.4% for 15) in comparison with the vehicle-treated group. Also, at the dose of 3 mg/kg we
observed a significant reduction of the tumor mass only for compound 12 (28.6%), whereas
the result with compound 15 did not reach statistical significance (16.7%). Notably, treatment
with the reference compound CA4P at 30 mg/kg caused only a small reduction in tumor
volume (31.3 %) if compared with both doses of compound 12. From the safety/tolerability
point of view, no significant variation in body weight occurred in animals treated with either
1
2 and 15 at the higher concentration (Figure 11, panel C).
A
B
2
7

ACCEPTED MANUSCRIPT
C
Figure 11. Inhibition of mouse allograft growth in vivo by compounds 12 and 15. Panel A:
7
male C57BL/6 mice were injected subcutaneously at their dorsal region with 10 BL6-B16
murine melanoma cells. Tumor-bearing mice were administered the vehicle, as control, or 7.5 and
3.0 mg/kg of 12 and 15 or CA4P as reference compound at 30 mg/kg. Injections were given
intraperitoneally on the days indicated by the arrows. Panel B: tumor weight registered at the end of
the treatment. Panel C: body weight variation after treatment with CA-4P, 12 and 15. Data are
presented as mean ± SEM of tumor volume at each time point for 5 animals per group. **p<0.01 vs.
control.
2
.3 Molecular modelling
Molecular docking studies were carried out to investigate the possible binding mode of the
synthesized compounds with the aim of interpreting the biological data. Recently, we reported
a study to identify the best docking protocol and the protein conformation most suitable to
accommodate the 7-PPyQ scaffold [11]. Taking advantage of our previous findings, we
docked derivatives: 12-15, 18 and 19. According with the previously reported 7-PPyQ
2
8