Synthesis and evaluation of gallocyanine dyes as potential agents for the treatment of Alzheimer's disease and related neurodegenerative tauopathies

Article abstract of DOI:10.1016/j.ejmech.2015.11.024

In search of safe and effective anti-Alzheimer disease agents a series of gallocyanine dyes have been synthesized and evaluated for their ability to inhibit LRPs/DKK1 interactions. Modulation of the interactions between LRP_S and DKK1, regulate

Full text of DOI:10.1016/j.ejmech.2015.11.024

European Journal of Medicinal Chemistry 108 (2016) 28e38
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Research paper
Synthesis and evaluation of gallocyanine dyes as potential agents for
the treatment of Alzheimer's disease and related neurodegenerative
tauopathies
,
,
,
Spyros Mpousis a 1, Savvas Thysiadis a 1, Nicolaos Avramidis b 1, Sotirios Katsamakas c,
Spiros Efthimiopoulos b **, Vasiliki Sarli a
,
, *
a Department of Chemistry, Aristotle University of Thessaloniki, University Campus, 54124, Thessaloniki, Greece
b Division of Animal and Human Physiology, Department of Biology, National & Kapodistrian University of Athens, Panepistimiopolis, Ilisia, Greece
c Department of Pharmaceutical Chemistry, School of Pharmacy, Aristotle University of Thessaloniki, University Campus, 54124, Thessaloniki, Greece
a r t i c l e i n f o
a b s t r a c t
Article history:
In search of safe and effective anti-Alzheimer disease agents a series of gallocyanine dyes have been
synthesized and evaluated for their ability to inhibit LRPs/DKK1 interactions. Modulation of the in-
teractions between LRPS and DKK1, regulate Wnt signaling pathway and affect Tau phosphorylation. The
current efforts resulted in the identification of potent DKK1 inhibitors which are able to inhibit pros-
taglandin J2-induced tau phosphorylation at serine 396.
Received 16 June 2015
Received in revised form
27 October 2015
Accepted 17 November 2015
Available online 22 November 2015
© 2015 Elsevier Masson SAS. All rights reserved.
Keywords:
Gallocyanine
Phenoxazine
DKK1 inhibitors
DKK1/LRP6 interaction inhibitors
Tau phosphorylation inhibition
Alzheimer's disease
Tauopathies
1. Introduction
tangles of hyperphosphorylated microtubule-associated Tau pro-
tein [2]. Recent findings have established that deregulation of Wnt
Alzheimer's disease (AD) is the most common neurodegenera-
tive disease characterized by excessive accumulation of senile
plaques and neurofibrillary tangles (NFT) [1]. Senile plaques are
signaling contributes to the pathophysiology of AD, while activa-
tion of the pathway is neuroprotective [3]. According to Hanger and
Noble the kinase activity of GSK3b is inhibited following activation
composed mainly of beta-amyloid peptide (A
b) and neurofibrillary
of Wnt signaling, implicating this pathway in the hyper-
phosphorylation of Tau and the formation of NFTs [4].
Wnt signaling is initiated when Wnt proteins simultaneously
interact with their receptor and co-receptor, Frizzled (FRZ) and
LRP5/6. Through several cytoplasmic components, the signal is
Abbreviations: AD, Alzheimer's disease; NFT, neurofibrillary tangles; A
b, beta-
amyloid peptide; FRZ, frizzled; DKK1, Dickkopf-1; GSK3 glycogen synthase
b
,
transduced to GSK3
b with a subsequent stabilization of beta-
kinase-3 beta; TCF, T-cell factor; LEF, lymphoid enhancer factor; BBB, bloodebrain
barrier; PGJ2, prostaglandin J2; LRP6, low-density lipoprotein receptor-related
protein 6; Wnt, wingless; CNS, central nervous system; p53, tumor protein p53;
catenin, which in turn enters the nucleus and forms a complex
with TCF/LEFs to activate transcription of target genes [5a]. In the
absence of Wnt, phosphorylation of beta-catenin by GSK3b leads to
ubiquitination and subsequent proteasomal degradation, resulting
in low levels of cytoplasmic beta-catenin [5b]. The canonical Wnt
pathway is also negatively modulated by the extracellular protein
Dickkopf-1 (DKK1), which binds to LRPs preventing their interac-
tion with Wnts and Frizzled membrane receptor [6]. The Dickkopf
family encodes secreted proteins of 255e350 aminoacids and
IL-1
hydrogen bond donor; HBA, hydrogen bond acceptor; DMSO, dimethyl sulfoxide;
MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; PK,
b, interleukin-1 beta; NCI, National Cancer Institute; HB, hydrogen bond; HBD,
pharmacokinetic.
* Corresponding author.
** Corresponding author.
E-mail addresses: efthis@biol.uoa.gr (S. Efthimiopoulos), sarli@chem.auth.gr
(V. Sarli).
1
These authors contributed equally.
http://dx.doi.org/10.1016/j.ejmech.2015.11.024
0223-5234/© 2015 Elsevier Masson SAS. All rights reserved.

S. Mpousis et al. / European Journal of Medicinal Chemistry 108 (2016) 28e38
29
consists of four main members (DKK-1 to -4) in mammals [7]. Sokol
and coworkers have reported that the DKK C-terminal domain
mainly interacts with the third and fourth propellers of LRP5/6 and
these interactions are necessary and sufficient for Wnt inhibition
[8].
The expression of DKK1 is increased in brains of AD patients and
in AD transgenic mouse models and this induction is dependent on
p53 and associated with neurodegeneration [9,10]. In cultured
neurons, DKK1 is able to elicit cell death associated with loss of
BCL-2 expression, induction of BAX, and Tau hyperphosphorylation
[11]. Along these lines, a recent study demonstrated that DKK1-
Scheme 2.
2. Results and discussion
2.1. Docking studies
neutralizing antibodies suppress Ab-induced synapse loss in mice
[12]. Thus, modulation of the interactions between LRPs and DKK1,
could regulate Wnt-beta-catenin signaling offering an attractive
therapeutic approach for treating AD. In fact the LRP5-DKK1
interaction has been recently targeted with monoclonal anti-DKK1
antibodies [13]. However, the lack of blood brain penetration of the
antibodies limits their application for the treatment of CNS
diseases.
In this study, a small molecule approach for regulating Wnt
signaling has been developed. Based on NCI8642, a recently
discovered compound which blocks DKK1 inhibitory activity by
disrupting DKK1/LRP6 interactions, a series of NCI8642 derivatives
have been synthesized and evaluated. Importantly, this study
shows that NCI8642 and related analogues induce inhibition of tau
phosphorylation in vivo. Since tau hyperphosphorylation is a hall-
mark of Alzheimer's disease these results may find application in
the treatment of AD and other tauopathies.
NCI8642 (1a, gallocyanine hydrochloride salt), a synthetic blue
dyestuff synthesized by Otto in 1888 [14], is a phenoxazinone (see
Scheme 1) independently discovered as an inhibitor of DKK1-LRP5/
6 interactions by the groups of Remelli [15], Zheng and Wu [16].
Phenoxazines are privileged compounds, present in several bioac-
tive molecules and natural products that display diverse biological
activities. Well-known examples are the natural products actino-
mycins, which are polypeptide antibiotics produced by Strepto-
myces species that show chemotherapeutic activity in a number of
neoplastic diseases [17]. The phenoxazine template is also present
in the plectosphaeroic acids isolated from the fungus Plectos-
phaerella cucumerina, which are inhibitors of indoleamine 2,3-
dioxygenase [18]. Simpler phenoxazines and benzophenoxazines
have well documented anticancer [19e21], antiviral [22], antimi-
crobial [23] and antimalarial [24,25] properties. Importantly,
phenothiazine compounds have also been reported to inhibit
heparin-induced assembly of tau protein [26].
In an attempt to gain insight into the structural basis of the
observed activity, NCI8642 and related analogues were docked
against LRP6. Previous publications have demonstrated that
NCI8642 binds to LRP6 and inhibits DKK1-LRP6 binding by using
biolayer interferometry and surface plasmon resonance technology
(SPR) [15,16]. Accordingly, the heterotetramer structure 3S8V in the
absence of DKK1c module (PDB code: 3S8V) was used in the pre-
sent studies [28]. The docking results show top scoring poses in
LRP6-E3E4/DKK1 interfaces consisting of the residues Ser851,
Asp874 in interface A and Trp850, Asp874, Tyr875 in interface B. For
the screened compounds a repeated motif for their calculated
binding was witnessed. The phenoxazine was proved a privileged
scaffold for LRP6, due to a versatile interaction either through the
nitrogen or oxygen atom with Tyr875 of interface A, in a binding
pocket which is mainly hydrophobic and occupied by Leu231 of the
DKK1c module in the LRP6/DKK1 complex. In Fig. 1 three major
binding modes of NCI8642 are illustrated: two poses with the
dimethylamino group to extend towards Ser851 and one pose with
the ammonium ion to display ionic interactions with Asp874.
In an effort to extract a qualitative probability profile for each
solution, we decided to synthesize analogues that added hindering
effects towards the hydrophobic region of the pocket (e.g. 7b, 7e,
Scheme 3) or analogues that lacked characteristic groups (e.g. 4a, 4j,
Scheme 3) which appeared to participate through HB forming in-
teractions, while constantly maintaining the original scaffold. In
addition, a series of ester derivatives were chosen to be tested in an
effort to increase hydrophobicity and mimic Leu231 of DKK1. Be-
sides the phenoxazine privileged scaffold, of almost equal signifi-
cance to the binding orientation of the tested compounds appeared
to be the presence of the ammonium salt. Notably, the tested
structures displayed several HBs that are generally maintained in
different poses, while the salts exhibited the usual HBs with an
extra possible ion interaction with length bellow a 4 Å3 threshold
[30], between the positively charged nitrogen and the anionic re-
gion of the pocket formed by the two aspartic residues, Asp874A
and Asp874B. These docking results have to be verified from real
biological data.
Among the known Wnt modulators, NCI8642 is the only DKK1
inhibitor described so far, but its use as a clinically suitable drug is
limited, because its association with LRP6 is weak (IC50 of about
3 mM in the inhibition of DKK1 binding). Nevertheless, data show
that NCI8642 is effective in vivo as it reduces basal blood-glucose
concentrations and improves glucose tolerance in mice [16].
NCI8642 possess both acidic and basic groups and a mesomeric
chromophoric system which is depicted in Scheme 2 [27].
2.2. Synthesis of gallocyanine dyes
The synthesis of gallocyanine derivatives was performed as
depicted in Scheme 3. Initially, the preparation of 4a took place
from the decarboxylation of commercially available gallocyanine 1a
was achieved using CH3COONa in water and heating according to
ꢀ
Simek and coworkers [27]. Next, N,N-dialkyl-4-nitrosoaniline hy-
drochloride salts 2 and 6 were allowed to react with commercially
available phenols 3i, 3j or with gallate esters to give the corre-
sponding phenoxazinones 4bej, 7b and 7e. It should be noted that,
the nitroso compounds 2 and 6 were prepared following the
method of G. M. Bennett and E. V. Bell [31]. Compounds 3bef were
synthesized according to known procedures [32e35] whereas the
synthesis of 3g and 3h is described in the Supporting Information
Scheme 1.

Fig. 1. (A) Crystal structure of LRP6-E3E4/DKK1 (PDB: 3S8V). The docking studies were performed in the highlighted LRP6-E3E4/DKK1 interfaces. (B) In silico docking results. Three
highest rank poses of NCI8642 (gold sticks) docked into LRP6 (pink-interface A, grey-interface B). Critical hydrogen bonding residues are labeled. The figures were obtained by
PyMol version 1.4.1 [29].

S. Mpousis et al. / European Journal of Medicinal Chemistry 108 (2016) 28e38
31
Scheme 3. Synthesis of gallocyanine dyes.
section. Ammonium salts 5c, 5e, 5g and 5h were synthesized from
the corresponding amines after treatment with HCl in iPrOH.
3. Biological evaluation
3.1. The effect of NCI8642 and related derivatives on the binding of
DKK1 to the surface of HEK-293 cells overexpressing LRP6
In order to examine the effect of NCI8642 and related derivatives
on the binding of DKK1 on the surface of HEK-293 cells over-
expressing LRP6 (HEK-293LRP6), sub-confluent cultures were
incubated with DKK1 conditioned medium (DKK1-CM) plus DMSO,
100 mM NCI8642 or 100 m
M NCI8642 derivatives for 2 h at 4 ꢀC [15].
Cell surface binding of DKK1 was visualized by immunocyto-
chemistry using a primary anti-DKK1 antibody and fluorescent
secondary antibody. Fluorescence images were analyzed using Zen
software and quantified results are shown in Fig. 2. The results
derive from three independent experiments performed in dupli-
cates. The average values represent measures of 750 cells for each
case.
As it can been seen in Fig. 2 the compounds tested had the
ability to inhibit DKK1-LRP6 interactions and most of them were
equipotent or more potent than NCI8642. The presence of the hy-
drochloride salt was critical for inhibition as it was indicated in the
increased activity of the free amino derivative 1b. Compound 4a
lacking the carboxyl group on the phenoxazine template was found
slightly more potent than 1a. Next, the effects of ester compounds
Fig. 2. The effect of NCI8642 and related derivatives on the binding of DKK1 to the
surface of HEK-293 cells overexpressing LRP6 (HEK-293LRP6). HEK-293LRP6 cells were
incubated with DKK1 conditioned medium (DKK1-CM) plus DMSO, NCI8642 or
NCI8642 derivatives at a concentration of 100 m
M for 2 h at 4 ꢀC. Immunofluorescence
images were used to quantify DKK1 binding at the cell surface using Zen software. The
results are expressed as percentage of DKK1 binding. The experiment was repeated
three times in duplicates and 750 cells were analyzed in each case. The differences are
statistically significant (p value < 0.01). If not, they are indicated as (þ). 7f ¼ ethyl
gallate.
which include various lengths of alkyl side chains on the inhibition
of DKK1-LRP6 interactions were examined. The data show that the
inhibitory activity was dependent on the length of the alkyl chain
and minor increases in potency were observed with higher alkyl

32
S. Mpousis et al. / European Journal of Medicinal Chemistry 108 (2016) 28e38
homologues such as the hexyl derivative 4h. The methoxyethyl
ester 4g was equally potent inhibitor with the butyl analogue 4e,
while transformation of the carboxylate moiety to a benzyl ester
provided a poor DKK1 inhibitor. In addition, compound 4j lacking
the carboxyl group and having a hydroxyl group at position 1 of
phenoxazine template was less potent than 1a, while compound 4i
lacking the hydroxyl group and carrying methyl group in the place
of carboxyl group was an active compound.
To clarify the structureeactivity relationships and improve the
water solubility of the key analogues a series of simple ammonium
salts were tested. Only, the hydrochloride salt 5g afforded a sig-
nificant improvement in inhibition over 4g. The other salts tested
5c, 5e and 5h were equally or less active than the corresponding
free amines 4c, 4e and 4h. Replacement of the dimethylamino by
the dipropylamino group resulted in increased activity when the
phenoxazine was carrying a methyl ester (7b) and in decreased
activity in the case of the n-butyl derivative (7e). It should be noted
that the ethyl gallate 7f was also tested in this assay and was found
completely inactive.
Fig. 4. Primary mouse cortex neurons were treated with 20
mM PGJ2 plus DMSO,
10 mM NCI8642 or 10
m
M NCI8642 derivatives for 1 h at 37 ꢀC. Cell lysates were
analysed by western blot using antibodies against tau phosphorylated at serine 396 or
actin. The density of the bands was measured using ImageJ software and the levels of
tau phosphorylation at serine 396 were normalized to the levels of actin. The levels of
phosphorylation upon combined treatment with PGJ2 plus the various NCI8642 de-
rivatives are depicted as percent of tau phosphorylation upon treatment only with
PGJ2. The experiments were repeated at least three times. The differences are statis-
tically significant (***p value < 0.01, **p value < 0.025).
(1a) is able to reduce PGJ2-induced tau phosphorylation at serine
396. Ester derivatives 4beh were effective inhibitors of tau phos-
phorylation, with most potent the ethyl derivative 4c which is 2.5
times as potent as NCI8642. The decarboxylated derivative 4a and
the hexyl ester 4h were found less active than 1a. Evidently, there is
no good correlation between the effect of NCI derivatives on DKK1
binding and tau phosphorylation. This outcome may possibly be
related to the fact that the tested NCI8642 derivatives may display
diverse binding modes as was also initially observed from the
docking studies. Some derivatives may bind to the DKK1 binding
site and prevent DKK1 from binding. Some others may bind to the
DKK1 binding site and act as an agonist of DKK1 while at the same
time prevent binding of DKK1. The different potencies can also be
associated to a non-selective activity of the tested compounds.
In an effort to evaluate the drug like profile of the synthesized
compounds several PK properties were calculated as depicted in
Table S1 (See Supporting Information). Generally, the phenoxazine
scaffold poses a volume of 166 Å3, and its substitution pattern is the
key for a better DKK1 inhibition. The overall drug like profile shows
a good partitioning and distribution coefficients, except from some
violations of the lipophilicity mainly regarding to the ammonium
salt derivatives. The polar surface area varied from approximately
60-100, HBD from 0 to 3, HBA from 4 to 8 and they appeared flat
with Fsp3 values between 0.13 and 0.43. This flexibility is mostly
attributed to the presence of the long aliphatic substituents.
3.2. Cytotoxicity of the new compounds
The synthesized compounds were screened for their cytotoxic
effects using the colorimetric MTT assay in order to determine the
percentage of viable cells remaining following compound incuba-
tion at two different concentrations (10 and 100 mМ) [36] After 24 h
of cell culture, neuronal cells were not affected by the gallocyanine
dyes since their viability was almost 100% (See Supporting
Information Figure S1).
3.3. Prostaglandin J2 (PGJ2) induces tau phosphorylation at serine
396
In order to set up an assay of induced tau phosphorylation, we
examined the effect of various immune factors IL-1
charide and PGJ2 on tau phosphorylation in mouse cortex primary
neurons. IL-1 , and lipopolysaccharide did not induce tau phos-
phorylation at several sites that were examined. However, as
indicated in Fig. 3, treatment of primary neurons with 10 or 20
b, lipopolysac-
b
mM
PGJ2 resulted in a significant increase in tau phosphorylation at
serine 396. Treatment with PGJ2 resulted also in tau hyper-
phosphorylation at other residues including serine 404, serine 199
and threonine 205 (data not shown).
The effect of selected NCI8642 derivatives on PGJ2-induced tau
phosphorylation at serine 396 was subsequently examined. Pri-
mary cortical neurons were incubated with 20
NCI8642 (10 M) or NCI8642 derivatives (10
m
m
M PGJ2 plus DMSO,
m
M) for 1 h at 37 ꢀC.
4. Conclusion
Phosphorylation of tau was assayed by western blot using a
pSer396 antibody. Quantification of the results normalized for actin
is shown in Fig. 4. Importantly, the experiment shows that NCI8642
A series of novel gallocyanine dyes have been synthesized and
evaluated for their ability to inhibit LRP/DKK1 interactions. Several
Fig. 3. A. Primary mouse cortex neurons were treated with 0, 1, 10 or 20 m
M of PGJ2 for 1 h at 37 ꢀC. Cell lysates were analysed by western blot using antibodies against tau
phosphorylated at serine 396, N-terminus of tau (total tau) or actin. B. The density of the bands was measured using ImageJ software and the levels of tau phosphorylation at serine
396 were normalized to the levels of actin. The increase in the phosphorylation upon treatment with various concentrations of PGJ2 is expressed as percent of the levels of tau
phosphorylation in untreated samples (0 mМ PGJ2). The differences are statistically significant (***p value < 0.01).

S. Mpousis et al. / European Journal of Medicinal Chemistry 108 (2016) 28e38
33
compounds have been identified as more potent inhibitors than
NCI8642 and important features for activity have been recognized.
The present study further showed that NCI8642 and derivatives are
able to inhibit prostaglandin J2-induced tau phosphorylation at
serine 396. This finding is of particular importance and provides
evidence that DKK1 may be a potential therapeutic target for
treating AD and other tauopathies. Since, an x-ray structure of the
NCI8642 with LRP6 is currently not available, a more detailed
description of the relation between the chemical structures and the
observed biological activities based only on docking simulations are
not provided here. A binding mode validation of the inhibitors on
LRP6 by X-ray structure determination and a larger number of in-
hibitors accompanied by their biological data are required for a
detailed and valid SAR analysis. This project is on progress and will
be reported in due course.
for 5 h. The reaction mixture was allowed to cool at 0 ꢀC. The
unreacted gallocyanine was taken up in of sodium carbonate so-
lution, while the decarboxylated gallocyanine is insoluble. The
precipitate was collected by filtration. To increase the purity, the
solid was refluxed in ethanol (9 mL) for 10 min and stored at 8 ꢀC
overnight. The product was then collected by filtration as a dark
blue solid (640 mg, yield 84%) [17]. 4a: 1H NMR (500 MHz, pyridine-
d5)
J ¼ 9.6 Hz, 1H), 6.62 (d, J ¼ 9.0 Hz, 1H), 6.54 (d, J ¼ 1.7 Hz, 1H), 2.88
(s, 6H); 13C NMR (126 MHz, pyridine)
180.9, 154.3, 147.6, 144.4,
d
7.67 (d, J ¼ 8.9 Hz, 1H), 7.42 (d, J ¼ 9.6 Hz, 1H), 7.03 (d,
d
133.2, 133.0, 132.5, 132.2, 130.2, 127.2, 110.6, 97.8, 40.4. FT-IR (KBr):
3436, 2913, 2847, 2809, 1616, 1583, 1546, 1454, 1437, 1356, 1276,
1198, 1132, 1081, 1002; HRMS m/z for C14H13N2O3 [MþH]þ calcd
257.0921, found 257.0920.
5.3. Methyl 7-(dimethylamino)-4-hydroxy-3-oxo-3H-phenoxazine-
5. Experimental section
1-carboxylate, 4b
5.1. General experimental details
200 mg of methyl gallate (1.09 mmol) and 245 mg of freshly
prepared
N,N-dimethyl-4-nitrosoaniline
hydrochloride
All reactions were carried out under an atmosphere of Ar unless
otherwise specified. Commercial reagents of high purity were
purchased and used without further purification, unless otherwise
noted. Reactions were monitored by TLC and using UV light as a
visualizing agent and aqueous ceric sulfate/phosphomolybdic acid,
ethanolic p-anisaldehyde solution, potassium permanganate solu-
tion, and heat as developing agents. The 1H and 13C NMR spectra
were recorded at 500 and 125 MHz, and tetramethylsilane was
(1.31 mmol) were added in 10 mL of methanol. The reaction
mixture was refluxed for 6 h, while the solution was turning into
purple color. Then, the reaction mixture was cooled, the crystals
were collected by suction filtration on a filter paper and were
washed with chilled MeOH. The solid was dried under reduced
pressure to give 192 mg of 4b as a dark purple solid in 56% yield. 4b:
mp > 300 ꢀC, 1H NMR (500 MHz, DMSO-d6/pyridine-d5)
d 7.52 (d,
J ¼ 9.1 Hz, 1H), 7.01 (s, 1H), 6.81 (dd, J ¼ 9.1, 2.5 Hz, 1H), 6.62 (d,
used as an internal standard. Chemical shifts are indicated in
J ¼ 2.5 Hz, 1H), 3.86 (s, 3H), 3.09 (s, 6H); 13C NMR (126 MHz, DMSO-
1
d
values (ppm) from internal reference peaks (TMS H 0.00; CDCl3
d6/pyridine-d5) d 176.9, 165.3, 154.2, 146.3, 137.9, 135.1, 133.6, 131.7,
1H 7.26, 13C 77.00; DMSO-d6 1H 2.50, 13C 39.51, pyridine-d5 1H 8.74,
7.58, 7.22; 13C 150.35, 135.91, 123.87). Melting points (mp) are un-
corrected. High-resolution mass spectra (HRMS) were recorded by
131.74, 127.3, 125.9, 111.1, 96.3, 52.5 (1 carbon missing due to
overlapping with DMSO-d6); FT-IR (KBr): 3435, 3046, 1707, 1599,
1547, 1433, 1413, 1383, 1315, 1247, 1190, 1140, 985, 850; HRMS m/z
for C16H15N2O5 [MþH]þ calcd 315.0975, found 315.0975.
direct injection of a 2
mM (2 ml) solution of the compounds in
watereacetonitrile (1/1; v/v) and 0.1% formic acid on a mass
spectrometer (hybrid ion trap-orbitrap mass spectrometer)
equipped with an electrospray ion source in positive mode (source
voltage 3.5 kV, sheath gas flow 10, capillary temperature 275 ꢀC)
with resolution R ¼ 60.000 at m/z ¼ 400 (mass range ¼ 150e2000)
and dioctylphthalate (m/z ¼ 391.28428) as the “lock mass”.
5.4. Ethyl 7-(dimethylamino)-4-hydroxy-3-oxo-3H-phenoxazine-
1-carboxylate, 4c
200 mg of ethyl gallate (1.01 mmol) and 225 mg of freshly
prepared
N,N-dimethyl-4-nitrosoaniline
hydrochloride
The polyclonal/monoclonal Rabbit Anti-Mouse antibody kindly
provided from Dr. Luc Buee (Insern, Fac de Medicine, Univ. Lille 2,
Lille France) directed against the N-terminus of tau and the poly-
clonal Rabbit Anti-Mouse antibody detecting tau phosphorylated at
serine 396 were purchased from Acris Antibodies GmbH (Ger-
many). Goat Anti-Rabbit IgG-HRP secondary antibody was pur-
chased from Jackson ImmunoResearch (Baltimore Pike, USA). The
monoclonal Mouse antibody used to detect actin by immunoblot
was purchased from Merk Millipore (Darmstatd, Germany). Goat
Anti-Mouse IgG-HRP secondary antibody was purchased from
Santa Cruz Biotechnology (Dallas/Texas, USA). The polyclonal Rab-
bit Anti-Human antibody against DKK1 was obtained from Santa
Cruz Biotechnology (Dallas/Texas, USA). Goat anti-Rabbit IgG
(H þ L) Alexa Fluor 488-conjugated and/or Goat anti-Rabbit IgG
(H þ L) Cy 3-conjugated secondary was purchased from Jackson
ImmunoResearch (Baltimore Pike, USA). B27 and Neurobasal used
to culture mouse primary neuronal culture were purchased from
Life Technologies (Grand Island/NY, USA). Prostaglandin J2 was
purchased from Santa Cruz Biotechnology (Dallas/Texas, USA).
Common chemicals were from Sigma Aldrich (Athens, Greece).
(1.21 mmol) were added in 3.5 mL of ethanol. The reaction mixture
was refluxed for 1 h, while the solution was turning into purple
color. Then, the reaction mixture was cooled, the crystals were
collected by suction filtration on a filter paper and were washed
with chilled EtOH. The solid was recrystallized with EtOH and was
dried under reduced pressure to give 106 mg of 4c as dark purple
solid (32% yield). 4c: mp > 300 ꢀC, 1H NMR (500 MHz, pyridine-d5)
d
7.68 (d, J ¼ 9.1 Hz, 1H), 7.44 (s, 1H), 6.61 (dd, J ¼ 9.1, 2.5 Hz, 1H),
6.50 (d, J ¼ 2.5 Hz, 1H), 4.45 (q, J ¼ 7.1 Hz, 2H), 2.84 (s, 6H), 1.30 (t,
J ¼ 7.1 Hz, 3H); 13C NMR (126 MHz, pyridine-d5)
d 179.2, 166.2,
154.8, 147.6, 140.6, 137.5, 135.5, 133.1, 132.7, 129.3, 127.3, 111.3, 97.4,
62.3, 40.4, 14.8; FT-IR (KBr): 3437, 3071, 2994, 2928, 1699, 1601,
1547, 1491, 1463, 1420, 1383, 1315, 1266, 1241, 1196; HRMS m/z for
C17H17N2O5 [MþH]þ calcd 329.1132, found 329.1131.
5.5. N-(9-(Ethoxycarbonyl)-6,7-dihydroxy-3H-phenoxazin-3-
ylidene)-N-methylmethanaminium chloride, 5c
30 mL of HCl in iPrOH (5N) solution were added dropwise in a
suspension of ethyl 7-(dimethylamino)-4-hydroxy-3-oxo-3H-phe-
noxazine-1-carboxylate, 4c (10 mg, 0.03 mmol) in 0.6 mL of
ethanol. The mixture was stirred at room temperature for 30 min.
Then, the mixture was concentrated to furnish 5c (quantitative
yield) as a dark blue-green solid. 5c: mp > 300 ꢀC; 1H NMR
5.2. 7-(Dimethylamino)-4-hydroxy-3H-phenoxazin-3-one, 4a
Gallocyanine hydrochloride (1 g, 2.97 mmol) and sodium ace-
tate (1 g, 12.19 mmol) were heated under reflux in water (100 mL)
(500 MHz, DMSO-d6)
d
7.52 (d, J ¼ 9.6 Hz, 1H), 6.98 (s, 2H), 6.72 (s,

34
S. Mpousis et al. / European Journal of Medicinal Chemistry 108 (2016) 28e38
1H), 4.33 (q, J ¼ 7.1 Hz, 2H), 3.20 (s, 6H), 1.31 (t, J ¼ 7.1 Hz, 3H);
HRMS m/z for C17H17N2O5 [M]þ calcd 329.1132, found 329.1132.
(0.223 mmol) were dissolved in 1.7 mL of ethanol. The reaction
mixture was refluxed for 3 h, while the solution was turning into
purple color. The crude product was crystallized with the slow
addition of the reaction mixture into 48 mL of water. The crystals
were collected by suction filtration on a filter paper and were
further purified by flash column chromatography (eluent; ethyl
acetate/MeOH ¼ 100/5) to afford 26 mg of 4f (35% yield) as a dark
purple solid. 4f: mp ¼ 221e224 ꢀC; 1H NMR (500 MHz, DMSO-d6)
5.6. Isopropyl 7-(dimethylamino)-4-hydroxy-3-oxo-3H-
phenoxazine-1-carboxylate, 4d
100 mg of isopropyl gallate (0.475 mmol) and 106 mg of freshly
prepared
N,N-dimethyl-4-nitrosoaniline
hydrochloride
(0.569 mmol) were added in 4.3 mL of ethanol. The reaction
mixture was refluxed for 4 h, while the solution was turning into
purple color. The crude product was crystallized with the slow
addition of the reaction mixture into 120 mL of water. The crystals
were collected by suction filtration on a filter paper and were
further purified by flash column chromatography (eluent; ethyl
acetate/MeOH ¼ 100/3) to afford 41 mg of 4d (25% yield) as dark
purple solid. 4d: mp ¼ 232e235 ꢀC; 1H NMR (500 MHz, DMSO-d6,
d
9.49 (s, 1H), 7.53 (d, J ¼ 7.3 Hz, 2H), 7.47 (d, J ¼ 9.1 Hz, 1H), 7.42 (t,
J ¼ 7.4 Hz, 2H), 7.36 (t, J ¼ 7.4 Hz, 1H), 6.97 (s, 1H), 6.89 (d, J ¼ 8.4 Hz,
1H), 6.66 (s, 1H), 5.38 (s, 2H), 3.16 (s, 6H); 13C NMR (126 MHz,
pyridine-d5, 40 ꢀC)
d 179.0, 166.2, 154.9, 147.6, 140.6, 137.5, 137.1,
135.1, 133.1, 132.6, 129.6, 129.3, 129.0, 127.3, 111.4, 97.4, 67.9, 40.4
(one carbon missing due to overlapping); FT-IR (KBr): 3205, 2923,
2847,1728, 1618,1577,1521,1493,1456,1387,1362,1232,1209,1178,
1137,1115,1020; HRMS m/z for C22H19N2O5 [MþH]þ calcd 391.1288,
found 391.1294.
49 ꢀC)
d
7.46 (d, J ¼ 9.1 Hz, 1H), 7.07e6.75 (m, 2H), 6.64 (d,
J ¼ 2.0 Hz, 1H), 5.19 (dt, J ¼ 12.5, 6.3 Hz, 1H), 3.15 (s, 6H), 1.34 (d,
J ¼ 6.2 Hz, 6H); 13C NMR (126 MHz, DMSO-d6)
d
176.8, 164.1, 154.1,
5.10. 2-Methoxyethyl 7-(dimethylamino)-4-hydroxy-3-oxo-3H-
phenoxazine-1-carboxylate, 4g
146.1, 137.9, 134.6, 133.9, 131.5, 131.3, 126.6, 125.6, 110.9, 96.2, 68.9,
(1 carbon missing due to overlapping with DMSO-d6), 21.3; FT-IR
(KBr): 3434, 3232, 2980, 2926, 1703, 1618, 1587, 1533, 1455, 1373,
1332, 1265, 1188, 1123, 1099; HRMS m/z for C18H19N2O5 [MþH]þ
calcd 343.1288, found 343.1288.
50 mg of 2-methoxyethyl gallate (0.219 mmol) and 47 mg of
freshly prepared N,N-dimethyl-4-nitrosoaniline hydrochloride
(0.254 mmol) were dissolved in 2 mL of ethanol. The reaction
mixture was refluxed for 3.30 h, while the solution was turning into
purple color. Then, the reaction mixture was cooled, the crystals
were collected by suction filtration on a filter paper and dried under
reduced pressure to give 25 mg of 4g as a dark purple solid in 32%
5.7. Butyl 7-(dimethylamino)-4-hydroxy-3-oxo-3H-phenoxazine-
1-carboxylate, 4e
200 mg of butyl gallate (0.884 mmol) and 165 mg of freshly
prepared N,N-dimethyl-4-nitrosoaniline hydrochloride (0.884
mmol) were dissolved in 8 mL of methanol. The reaction mixture
was refluxed for 12 h, while the solution was turning into purple-
red color. The crystals were collected by suction filtration on a fil-
ter paper and were dried under reduced pressure to give 79 mg of
4e. The filtrate was concentrated and the product was purified by
dry silica chromatography (eluent; ethyl acetate/methanol ¼ 100/
3) to afford 91 mg of 4e (total yield 54%). 4e: mp ¼ 285e287 ꢀC; 1H
yield. 4g: mp > 300 ꢀC, 1H NMR (500 MHz, DMSO-d6)
J ¼ 9.5 Hz, 1H), 7.33 (m, 1H), 7.25 (s, 1H), 6.95 (d, J ¼ 1.8 Hz, 1H),
4.50e4.39 (m, 2H), 3.72e3.62 (m, 2H), 3.37 (s, 6H), 3.31 (s, 3H); 13
NMR (126 MHz, DMSO-d6/pyridine-d5, 49 ꢀC)
176.8, 164.5, 154.1,
d 7.66 (d,
C
d
146.1, 137.9, 133.4, 133.1, 131.5, 127.3, 125.8, 114.1, 110.9, 96.1, 69.5,
64.2, 57.9 (one carbon missing due to overlapping); FT-IR (KBr):
3433, 2923,1719, 1599, 1548, 1485, 1425, 1382, 1316, 1279, 1247,
1212, 1191, 1122, 1094; HRMS m/z for C18H19N2O6 [MþH]þ calcd
359.1238, found 359.1237.
NMR (500 MHz, DMSO-d6)
d
9.46 (s, 1H), 7.45 (d, J ¼ 8.4 Hz, 1H),
6.91 (s, 1H), 6.89 (m, 1H), 6.65 (s, 1H), 4.29 (t, J ¼ 6.3 Hz, 2H), 3.15 (s,
5.11. N-(6,7-Dihydroxy-9-((2-methoxyethoxy)carbonyl)-3H-
phenoxazin-3-ylidene)-N-methylmethanaminium chloride, 5g
6H),1.83e1.54 (m, 2H),1.46 (m, 2H), 0.94 (t, J ¼ 7.4 Hz, 3H); 13C NMR
(126 MHz, DMSO-d6)
d 177.0, 165.0, 154.3, 146.3, 137.9, 133.8, 131.7,
131.5, 127.0, 125.9, 111.3, 108.5, 96.3, 65.0, 40.2, 30.1, 18.6, 13.6; FT-IR
(KBr): 3247, 2957, 2924, 2860, 1716, 1614, 1586, 1536, 1489, 1455,
1375, 1330, 1260, 1186, 1117; HRMS m/z for C19H21N2O5 [MþH]þ
calcd 357.1445, found 357.1444.
28 mL of HCl in iPrOH (5N) solution were added dropwise in a
suspension of 2-methoxyethyl 7-(dimethylamino)-4-hydroxy-3-
oxo-3H-phenoxazine-1-carboxylate 4g (10 mg, 0.028 mmol) in
0.6 mL of ethanol. The mixture was stirred at room temperature for
30 min. Then, the mixture was concentrated to furnish 5g (quan-
titative yield) as a dark blue-green solid. 5g: mp ¼ 188e191 ꢀC 1H
5.8. N-(9-(Butoxycarbonyl)-6,7-dihydroxy-3H-phenoxazin-3-
ylidene)-N-methylmethanaminium chloride, 5e
NMR (500 MHz, DMSO-d6, 49 ꢀC)
7.09e6.91 (m, 2H), 6.74 (d, J ¼ 2.5 Hz, 1H), 4.61e4.25 (t, J ¼ 5.0 Hz,
2H), 3.88e3.58 (t, J ¼ 5.0 Hz, 2H), 3.32 (s, 3H), 3.22 (s, 6H); HRMS m/
z for C18H19N2O6 [M]þ calcd 359.1238, found 359.1238.
d
7.54 (d, J ¼ 9.2 Hz, 1H),
28 mL of HCl in iPrOH (5N) solution were added dropwise in a
suspension of butyl 7-(dimethylamino)-4-hydroxy-3-oxo-3H-phe-
noxazine-1-carboxylate 4e (10 mg, 0.028 mmol) in 0.6 mL of
ethanol. The mixture was stirred at room temperature for 30 min.
Then, the mixture was concentrated to furnish 5e (quantitative
yield) as a dark blue-green solid. 5e: mp ¼ 163e166 ꢀC; 1H NMR
5.12. Hexyl 7-(dimethylamino)-4-hydroxy-3-oxo-3H-phenoxazine-
1-carboxylate, 4h
(500 MHz, DMSO-d6, 49 ꢀC)
d
7.55 (d, J ¼ 9.3 Hz, 1H), 7.16 (d,
59 mg of hexyl gallate (0.233 mmol) and 52 mg of freshly pre-
pared N,N-dimethyl-4-nitrosoaniline hydrochloride (0.28 mmol)
were dissolved in 0.8 mL of ethanol. The reaction mixture was
refluxed for 1 h, while the solution was turning into purple color.
The crude product was crystallized with the slow addition of the
reaction mixture into 100 mL of water. The crystals were collected
by suction filtration on a filter paper and were further purified by
flash column chromatography (eluent; ethyl acetate/MeOH ¼ 100/
3) to afford 30 mg of 4h (33% yield) as a dark purple solid. 4h:
J ¼ 7.8 Hz, 1H), 7.10 (s, 1H), 6.83 (s, 1H), 4.32 (t, J ¼ 6.3 Hz, 2H), 3.29
(s, 6H), 1.70 (m, 2H), 1.48 (m, 2H), 0.96 (t, J ¼ 7.4 Hz, 3H); HRMS m/z
for C19H21N2O5 [M]þ calcd 357.1445, found 357.1442.
5.9. Benzyl 7-(dimethylamino)-4-hydroxy-3-oxo-3H-phenoxazine-
1-carboxylate, 4f
50 mg of benzyl gallate (0.192 mmol) and 42 mg of freshly
prepared
N,N-dimethyl-4-nitrosoaniline
hydrochloride
mp ¼ 186e189 ꢀC,1H NMR (500 MHz, DMSO-d6/pyridine-d5)
d 7.47

S. Mpousis et al. / European Journal of Medicinal Chemistry 108 (2016) 28e38
35
(d, J ¼ 9.1 Hz, 1H), 6.98 (s, 1H), 6.83 (d, J ¼ 8.6 Hz, 1H), 6.63 (s, 1H),
4.29 (t, J ¼ 6.4 Hz, 2H), 3.10 (s, 6H), 1.81e1.59 (m, 2H), 1.38 (m, 2H),
1.24 (m, 4H), 0.81 (s, 3H); 13C NMR (126 MHz, DMSO-d6/pyridine-
the reaction mixture into 84 mL of water. The crystals were
collected by suction filtration on a filter paper and were further
purified by flash column chromatography (eluent; ethyl acetate/
MeOH ¼ 100/3) to afford 32 mg of 7b (33% yield) as a dark purple
d5)
d 177.0, 165.0, 154.2, 146.3, 138.0, 133.8, 131.6, 127.1, 125.9, 111.1,
96.3, 65.3, 45.7, 30.8, 28.0, 25.0, 22.0, 13.8 (two carbons are missing
due to overlapping with pyridine-d5); FT-IR (KBr): 3454, 2955,
2844, 1726, 1582, 1457, 1428, 1336, 1242, 1188, 1111; HRMS m/z for
solid. 7b: mp ¼ 223e225 ꢀC; 1H NMR (500 MHz, CDCl3)
d 7.62 (d,
J ¼ 9.3 Hz, 1H), 7.21 (s, 1H), 6.73 (dd, J ¼ 9.3, 2.5 Hz, 1H), 6.58 (d,
J ¼ 2.5 Hz, 1H), 3.98 (s, 3H), 3.77e3.03 (m, 4H), 1.72 (m, 4H), 1.00 (t,
C
21H25N2O5 [MþH]þ calcd 385.1758, found 385.1762.
J ¼ 7.4 Hz, 6H); 13C NMR (126 MHz, CDCl3)
d 176.5, 165.5, 153.4,
146.9, 137.5, 134.0, 133.4, 133.3, 130.9, 127.4, 127.3, 111.3, 96.3, 53.4,
52.8, 20.6, 11.3; FT-IR (KBr): 3451, 3224, 2970, 2872, 1728, 1612,
1585,1534, 1457, 1401, 1364,1335, 1305, 1232, 1203, 1172, 1114;
HRMS m/z for C20H23N2O5 [MþH]þ calcd 371.1601, found 371.1601.
5.13. N-(9-((Hexyloxy)carbonyl)-6,7-dihydroxy-3H-phenoxazin-3-
ylidene)-N-methylmethan-aminium chloride, 5h
26 mL of HCl in iPrOH (5N) solution were added dropwise in a
suspension of hexyl 7-(dimethylamino)-4-hydroxy-3-oxo-3H-
phenoxazine-1-carboxylate 4h (10 mg, 0.026 mmol) in 0.6 mL of
ethanol. The mixture was stirred at room temperature for 30 min.
Then, the mixture was concentrated to furnish 5h (quantitative
yield) as a dark blue-green solid. 5h: mp ¼ 175e178 ꢀC, 1H NMR
5.17. Butyl 7-(dipropylamino)-4-hydroxy-3-oxo-3H-phenoxazine-
1-carboxylate 7e
117 mg of butyl gallate (0.515 mmol) and 150 mg of freshly
prepared N,N-dipropyl-4-nitrosoaniline hydrochloride (0.628
mmol) were dissolved in 1.8 mL of ethanol. The reaction mixture
was refluxed for 1.5 h, while the solution was turning into purple
color. The crude product was crystallized with the slow addition of
the reaction mixture into 126 mL of water. The crystals were
collected by suction filtration on a filter paper and were further
purified by flash column chromatography (eluent; ethyl acetate/
MeOH ¼ 100/3) to afford 63 mg of 7e (30% yield) as a dark purple
(500 MHz, DMSO-d6, 49 ꢀC)
d
7.50 (d, J ¼ 9.3 Hz, 1H), 7.00 (dd,
J ¼ 9.4, 2.2 Hz, 1H), 6.97 (s, 1H), 6.73 (d, J ¼ 2.7 Hz, 1H), 4.30 (t,
J ¼ 6.4 Hz, 2H), 3.22 (s, 6H), 1.78e1.68 (m, 2H), 1.55e1.37 (m, 2H),
1.33 (s, 4H), 0.89 (d, J ¼ 6.9 Hz, 3H); HRMS m/z for C21H25N2O5 [M]þ
calcd 385.1758, found 385.1760.
5.14. 7-(Dimethylamino)-1-methyl-3H-phenoxazine-3-one, 4i
solid. 7e: mp ¼ 102e105 ꢀC; 1H NMR (500 MHz, CDCl3)
d 7.58 (d,
100 mg of monohydrate orcinol (0.704 mmol) and 160 mg of
freshly prepared N,N-dimethyl-4-nitrosoaniline hydrochloride
(0.844 mmol) were dissolved in 7 mL of methanol. The reaction
mixture was refluxed for 12 h, while the solution was turning into
purple-red color. The reaction mixture was concentrated under the
reduced pressure and was purified by flash column chromatog-
raphy (eluent; ethyl acetate/MeOH ¼ 100/3) to obtain 82 mg of 4i as
a dark red solid (46% yield). 4i: mp ¼ 229e232 ꢀC; 1H NMR
J ¼ 9.2 Hz, 1H), 7.17 (s, 1H), 6.69 (d, J ¼ 9.2 Hz, 1H), 6.56 (s, 1H), 4.39
(t, J ¼ 8.0 Hz, 2H), 3.46e3.29 (t, J ¼ 10.0 Hz 4H),1.74 (m, 6H),1.52 (m,
2H), 0.99 (m, 9H); 13C NMR (126 MHz, CDCl3)
d 176.7, 165.2, 153.3,
146.9, 137.7, 133.7, 133.2, 130.9, 127.3, 127.2, 111.2, 96.3, 65.6, 53.4,
30.6, 20.6, 19.2, 13.7, 11.3 (one carbon is missing due to over-
lapping); FT-IR (KBr): 3442, 2961, 2927, 2850, 1628, 1587, 1520,
1451, 1396, 1356, 1232, 1115; HRMS m/z for C23H29N2O5 [MþH]þ
calcd 413.2071, found 413.2070.
(500 MHz, CDCl3)
6.61 (s, 1H), 6.43 (s, 1H), 6.18 (s, 1H), 3.14 (s, 6H), 2.40 (s, 3H); 13C
NMR (126 MHz, CDCl3) 185.9, 153.5, 150.8, 146.6, 142.9, 142.0,
d
7.59 (d, J ¼ 9.1 Hz, 1H), 6.69 (d, J ¼ 8.9 Hz, 1H),
5.18. Cell culture and preparation of mouse primary cortical
neurons
d
131.6, 130.7, 125.8, 110.3, 105.2, 96.4, 40.4, 16.8; FT-IR (KBr): 2880,
1614, 1547, 1483, 1459, 1413, 1378, 1335, 1287, 1198, 1175, 1135, 900;
HRMS m/z for C15H15N2O2 [MþH]þ calcd 255.1128, found 255.1127.
HEK-293 human embryonic kidney cells overexpressing LRP6
(HEK-293LRP6) or DKK1 (HEK-293DKK1) were cultured in a 5% CO2
atmosphere at 37 ꢀC in Dulbecco's Modified Eagle's Medium
(DMEM) (Sigma Aldrich, Athens, Greece) supplemented with 10%
heat inactivated fetal bovine serum (FBS) (PAA Laboratories GmbH,
Linz, Austria), 1% penicillin/streptomycin (Sigma Aldrich, Athens,
5.15. 7-(Dimethylamino)-1-hydroxy-3H-phenoxazin-3-one, 4j
N,N-dimethyl-4-nitroso aniline (345 mg, 1.85 mmol) and
phloroglucinol (200 mg, 1.23 mmol) were heated under reflux in
ethanol (15.5 mL) for 2.5 h. The mixture was allowed to cool at 0 ꢀC.
The precipitate was collected by filtration and washed with ethanol
(3 ꢁ 1 ml), to obtain 128 mg of 4j as a dark blue-purple product
(40% yield) [37]. 4j: mp > 300 ꢀC, 1H NMR (500 MHz, pyridine-d5)
Greece) and 10 mg/ml Blasticidine for HEK-293DKK1 and 100 mg/ml
Zeocine for HEK-293LRP6. Confluent HEK-293DKK1 cells were
cultured for 72 h and the conditioned medium was collected,
cleared from cell debris and floating cells by centrifugation and
stored at ꢂ80 ꢀC until use. Primary mouse cortex neuronal culture
were prepared from 16 days embryos as described and were treated
with PG J2 after 6 days in vitro [38].
d
7.56 (d, J ¼ 9.0 Hz, 1H), 6.63 (d, J ¼ 2.0 Hz, 1H), 6.61 (dd, J ¼ 9.1,
2.6 Hz, 1H), 6.46 (d, J ¼ 2.6 Hz, 1H), 2.91 (s, 6H); 13C NMR (126 MHz,
pyridine-d5)
d 186.4, 159.3, 154.2, 147.6, 138.7, 131.5, 125.3, 110.7,
109.5, 104.2, 97.4, 40.2 (one carbon is missing due to overlapping);
FT-IR (KBr): 3389, 2879, 1649, 1612, 1576, 1536, 1480, 1390, 1327,
1279, 1213, 1157, 1083, 900, 846; HRMS m/z for C15H13N2O3 [MþH]þ
calcd 257.0921, found 257.0920.
5.19. DKK1 binding assay
Binding of DKK1 was performed as previously described by Iozzi
and coworkers [15]. Briefly, HEK-293-LRP6 cells were seeded onto
poly-
were incubated with DKK1 conditioned media (DKK1-CM) plus
DMSO, DKK1-CM plus NCI8642 (100 M) or DKK1-CM plus
NCI8642 derivatives (100
M) for 2 h at 4 ꢀC. After washing, the
D-lysine-coated coverslips 24 h before the experiment and
5.16. Methyl 7-(dipropylamino)-4-hydroxy-3-oxo-3H-
phenoxazine-1-carboxylate, 7b
m
m
63 mg of methyl gallate (0.343 mmol) and 100 mg of freshly
prepared N,N-dipropyl-4-nitrosoaniline hydrochloride (0.412
mmol) were dissolved in 1.2 mL of ethanol. The reaction mixture
was refluxed for 1.5 h, while the solution was turning into purple
color. The crude product was crystallized with the slow addition of
cells was fixed in 3% paraformaldehyde (15 min, room tempera-
ture), blocked with PBS/0.1% BSA, and incubated with the primary
rabbit-anti-DKK1 antibody, and fluorescent secondary antibody
Alexa488 diluted into blocking solution. This antibody which emits
fluorescence at 519 nM was very carefully chosen following the

36
S. Mpousis et al. / European Journal of Medicinal Chemistry 108 (2016) 28e38
observation that some of the NCI8642 derivatives emit fluorescence
at 570 nM. Cell nuclei were counterstained with DAPI. Images were
collected using a fluorescence microscope and 750 cells were
analyzed each time using ZEN software.
5.24. Molecular docking analysis
Novel leads based on the hit compound NCI8642were designed
and rationalized in silico prior to synthesis [39]. These compounds
form a library of a smile formatted file (*.smi) with the use of the
program Open Babel [40], generating their conformers with the
Omega v.2.5.1.4 software (OpenEye Scientific Software, Inc., Santa
Fe, NM, USA; www.eyesopen.com) [41]. All the docking experi-
ments were performed on a typical desktop pc running Windows 7
64-bit operating system (Dual Core Intel Pentium 3.2 GHz CPU
processors, RAM 8 GB), using the OEDocking suite v. 3.0.1 programs
(OpenEye Scientific Software, Inc., Santa Fe, NM, USA; www.
eyesopen.com) [42e44] and the PyRx v. 0.8 program (The Scripps
Research Institute) [45]. Based on the fact that the programs run
with different algorithms, a greater probability of the predict model
is offered. The visualization of the docking solutions is given by the
software PyMol v. 1.4.1 [29].
5.20. Cell treatments and cell lysis
Mouse cortex primary neurons were treated with 20
mМ pros-
taglandin J2 and DMSO, 10
mM NCI8642 or 10 mМ NCI8642 de-
rivatives for 1 h at 37 ꢀC.
Following treatment, cells were lysed at 4 ꢀC in lysis buffer
[50 mM TriseHCl, 150 mM NaCl, 2 mM EDTA, pH 7.6, 1% Triton X-
100 (v/v)] supplemented with complete protease inhibitor cocktail
and phosphatase inhibitor cocktail (Roche Applied Science, Man-
heim, Germany). Cell lysates were left on ice for 30 min and then
centrifuged at 14,000 g for 12 min at 4 ꢀC. Protein concentrations in
supernatants were determined by using the Bradford protein assay.
Molecular descriptor values for the synthesized compounds
were also calculated in an effort to export the drug-like profile for
them. Those include the factor of sp3 molecular orbitals (Fsp3),
resembling the compound saturation levels (flatness) [46], Lip-
inski's rule of five [47,48] descriptors (where calculated logarithmic
lipophilicity “clogPMB00 was achieved with the use of program
Marvin Beans v. 14.9.29) [49,50] and spatial characteristics (volume
and polar surface area, see Table 1 in supporting information) were
calculated online with the freely available Molinspiration property
calculation service (www.molinspiration.com).
5.21. Western blot and densitometry
Appropriate protein concentrations of cell lysates were pre-
pared in Laemmli solubilisation buffer containing
b-mercaptoe-
thanol and analyzed by sodium dodecyl sulfate polyacrylamide
(SDS-PAGE) gel electrophoresis. Separated proteins were trans-
ferred to polyvinyldiene difluoride membranes (Roth GmbH,
Karlsruhe, Germany) and immunoblotted with the appropriate
dilutions of the primary antibody overnight at 4 ꢀC. Primary anti-
bodies were: Polyclonal rabbit antibody detecting mouse tau
phosphorylated at serine396 (1:1000) and a polyclonal antibody
directed against the N-terminus of tau (1:1000) that was a kind gift
of Dr. Luc Buee (Inserm, Lille, France). The membranes were washed
with Tris-buffered Saline-Tween 20 (TBS-T) three times for 5 min
and incubated with a 1:5000 dilution of goat anti-rabbit or anti-
mouse horseradish peroxidise-conjugated secondary antibodies
for 1 h at room temperature. Following washing with TBS-T, pro-
teins were detected by chemiluminescence using the enhanced
chemiluminescence system (GenScript Piscataway, NJ, USA) on a
Fluorochem 8800 imaging station (AlphaInnotech, CA, USA). TIFF
files of the images were viewed in ImageJ and band intensity was
quantified with ImageJ software. The intensity measurements of
the bands were normalized to the levels of actin. The results from
three independent experiments were presented in graphs as
mean SD.
5.25. Ligand and protein preparation
The crystal structure of LRP6-E3E4/DKK1 complex (PDB code:
3S8V) was obtained from the protein databank [28]. DKK1 peptide
was removed with the use of OEDocking suite program MAKE
RECEPTOR v. 3.0.1 (OpenEye Scientific Software, Inc., Santa Fe, NM,
USA; www.eyesopen.com) [42e44] and the prepared structure was
eventually used as input file for the docking. The search space was
centered on the LRP6-E3E4/DKK1 interfaces between subunits A
and B. This process generated an initial 40,025 Å3 box, which
resulted after a balanced site-shape potential implementation to an
outer contour docking space of 13,515 Å3, narrowed by an inner
contour docking space of 10,559 Å3. No further residue modifica-
tions were made and the docking was performed without any
present constraints. The in-house formed library of NCI8642 ana-
logues was collected and the structures were transformed to their
smile formatted file (*.smi) with the use of the program Open Babel
[40]. Optimal parameters for conformer generation were deter-
mined using the Omega v.2.5.1.4 software (OpenEye Scientific
Software, Inc., Santa Fe, NM, USA; www.eyesopen.com) [41,42] and
the docking experiments were accomplished using of OEDocking
suite program FRED v. 3.0.1 (OpenEye Scientific Software, Inc., Santa
Fe, NM, USA; www.eyesopen.com) [43e45], utilizing an Exhaustive
Search Algorithm. All the results were additionally refined by the
use of OEDocking suite program FRED rescore v. 3.0.1 (OpenEye
Scientific Software, Inc., Santa Fe, NM, USA; www.eyesopen.com)
[43e45] giving the sorting of poses based on its standard options.
5.22. MTT assay
Primary neuronal cultures were seeded in 96-well plates at a
density of 25,000 cells/well and 15,000 cells/well. The cultures
were grown for 6 days at 37 ꢀC with 5% CO2. Then medium was
changed to that containing various concentrations (10
100 M) of NCI8642 and NCI8642 derivatives and incubated for
24 h at 37 ꢀC with 5% CO2. In all cases the final concentration of
DMSO was 0.1%. 20 l of MTT reagent (2.5 mg/ml MTT in PBS) was
added to each well and incubated for 4 h. The resulting formazan
dye was extracted with 100 isopropanol/HCl (100 ml
l HCl) and the absorbance was measured
mМ or
m
m
m
l
isopropanol þ 833
m
spectrophotometrically at a wavelength of 545 nm.
5.26. Validation of the docking protocol and ligand docking
5.23. Statistical analysis
The docking was reproduced for the compiled library also with
the use of the program AutoDock Vina (The Scripps Research
Institute) [51], utilizing this program's Genetic Algorithm. Whilst
the protein is treated as rigid and the solely parameter adjustment
was observed on the “exhaustiveness”, which value was set to 100
instead of the default 8.
All experiments were repeated three times. One-way ANOVA
with Bonferroni's Multiple Comparison Test was used to evaluate
the statistical significance of the differences. Statistical significance
was defined as p < 0.05.

S. Mpousis et al. / European Journal of Medicinal Chemistry 108 (2016) 28e38
Cancer Res. 7 (2001) 704e708.
37
Acknowledgment
[20] F. Suzuki, K. Hashimoto, M. Ishihara, G. Westman, K. Samuelsson, M. Kawase,
N. Motohashi, H. Sakagami, Tumor-specificity and type of cell death induced
by phenoxazines, Anticancer Res. 27 (2007) 4233e4238.
This programme was implemented within the framework of the
Operational Programme «Education and Lifelong Learning» of
NSRF: ARISTEIA II. It was co-funded by the European Union and
national resources. Furthermore the authors would like to thank
Openeye Scientific Software, Inc., Santa Fe, NM, USA; www.
eyesopen.com, for providing an academic license of their programs.
€
[21] H. Prinz, B. Chamasmani, K. Vogel, K.J. Bohm, B. Aicher, M. Gerlach,
E.G. Günther, P. Amon, I. Ivanov, K. Müller, N-benzoylated phenoxazines and
phenothiazines: synthesis, antiproliferative activity, and inhibition of tubulin
polymerization, J. Med. Chem. 54 (2011) 4247e4263.
[22] K. Hayashi, T. Hayashi, A. Tomoda, Phenoxazine derivatives inactivate human
cytomegalovirus, herpes simplex virus-1, and herpes simplex virus-2 in vitro,
J. Pharmacol. Sci. 106 (2008) 369e375.
[23] V.H. Frade, M.J. Sousa, J.C. Moura, M.S. Gonçalves, Synthesis of naphtho[2,3-a]
phenoxazinium chlorides: structureeactivity relationships of these hetero-
cycles and benzo[a]phenoxazinium chlorides as new antimicrobial, Bioorg.
Med. Chem. 16 (2008) 3274e3282.
[24] K. Takasu, T. Shimogama, C. Satoh, M. Kaiser, R. Brun, M. Ihara, Synthesis and
antimalarial property of orally active phenoxazinium salts, J. Med. Chem. 50
(2007) 2281e2284.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.ejmech.2015.11.024.
[25] J.F. Ge, C. Arai, M. Yang, Md A. Bakar, J. Lu, N.S. Ismail, S. Wittlin, M. Kaiser,
R. Brun, S.A. Charman, T. Nguyen, J. Morizzi, I. Itoh, M. Ihara, Discovery of
novel benzo[a]phenoxazine SSJ-183 as a drug candidate for malaria, ACS Med.
Chem. Lett. 1 (2010) 360e364.
References
[1] D.J. Selkoe, Alzheimer's disease, Cold Spring Harb. Perspect. Biol. (2011),
http://dx.doi.org/10.1101/cshper spect.a004457.
[2] M. Goedert, M.G. Spillantini, R. Jakes, D. Rutherford, R.A. Crowther, Multiple
isoforms of human microtubule-associated protein tau: sequences and
localization in neurofibrillary tangles of Alzheimer's disease, Neuron 3 (1989)
519e526.
[3] W. Cerpa, J.A. Godoy, I. Alfaro, G.G. Farías, M.J. Metcalfe, R. Fuentealba,
C. Bonansco, N.C. Inestrosa, Wnt-7a modulates the synaptic vesicle cycle and
synaptic transmission in hippocampal neurons, J. Biol. Chem. 283 (2008)
5918e5927.
[4] D.P. Hanger, W. Noble, Functional implications of glycogen synthase kinase-3-
mediated tau phosphorylation, Int. J. Alzheimers Dis. 2011 (2011) 352805.
[5] a) K.M. Cadigan, M.L. Waterman, TCF/LEFs and Wnt signaling in the nucleus,
Cold Spring Harb. Perspect. Biol. 4 (2012) 1e23;
[26] S. Taniguchi, N. Suzuki, M. Masuda, S. Hisanaga, T. Iwatsubo, M. Goedert,
M. Hasegawa, Inhibition of heparin-induced tau filament formation by phe-
nothiazines, polyphenols, and porphyrins, J. Biol. Chem. 280 (2005)
7614e7623.
ꢀ
ꢁ
ꢀ
ꢀ
ꢁ
[27] V. Dostal, M. Kotoucek, H. Kalasova, V. Bryndova, J. Simek, Acid-base equilibria
of some 7-dimethylamino-3-phenoxazone derivatives, Collect. Czech. Chem.
Commun. 47 (1982) 1588e1596.
[28] Z. Cheng, T. Biechele, Z. Wei, S. Morrone, R.T. Moon, L. Wang, W. Xu, Crystal
structures of the extracellular domain of LRP6 and its complex with DKK1,
Nat. Struct. Mol. Biol. 18 (2011) 1204e1210.
€
[29] The PyMOL Molecular Graphics System, Version 1.4.1, Schrodinger, LLC.
[30] S. Kumar, R. Nussinov, Close-range electrostatic interactions in proteins,
Chembiochem 3 (2002) 604e617.
[31] G.M. Bennett, E.V. Bell, 2.4-Dinitrobenzaldehyde, Org. Synth. 2 (1943) 223.
[32] A. Alama, Y. Takaguchi, H. Ito, T. Yoshida, S. Tsuboi, Multi-functionalization of
b) C. Liu, Y. Kato, Z. Zhang, V.M. Do, B.A. Yankner, X. He, b-Trcp couples b-
catenin phosphorylation-degradation and regulates Xenopus axis formation,
Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 6273e6278.
gallic acid towards improved synthesis of
a- and b-DDB, Tetrahedron 61
(2005) 1909e1918.
[6] A.M. Zorn, Wnt signalling: antagonistic Dickkopfs, Curr. Biol. 11 (2001)
R592eR595.
[7] Y. Kawano, R. Kypta, Secreted antagonists of the Wnt signalling pathway,
J. Cell Sci. 116 (2003) 2627e2634.
[8] B.K. Brott, S.Y. Sokol, Regulation of Wnt/LRP signaling by distinct domains of
Dickkopf proteins, Mol. Cell Biol. 22 (2002) 6100e6110.
[9] M.C. Rosi, I. Luccarini, C. Grossi, A. Fiorentini, M.G. Spillantini, A. Prisco, C. Scali,
M. Gianfriddo, A. Caricasole, G.C. Terstappen, F. Casamenti, Increased
Dickkopf-1 expression in transgenic mouse models of neurodegenerative
disease, J. Neurochem. 112 (2010) 1539e1551.
[33] S. Takaoka, N. Takaoka, Y. Minoshima, J.-M. Huang, M. Kubo, K. Harada,
H. Hioki, Y. Fukuyama, Isolation, synthesis, and neurite outgrowth-promoting
activity of illicinin A from the flowers of Illicium anisatum, Tetrahedron 65
(2009) 8354e8361.
[34] K. Cheng, X. Wang, S. Zhang, H. Yin, Discovery of small-molecule inhibitors of
the TLR1/TLR2 complex, Angew. Chem. 51 (2012) 12246e12249.
[35] P.P. Ambika Singh, S.M.S. Chauhan, Chemoselective esterification of phenolic
acids in the presence of sodium bicarbonate in ionic liquids, Synth. Commun.
38 (2008) 928e936.
[36] T. Mosmann, Rapid colorimetric assay for cellular growth and survival:
application to proliferation and cytotoxicity assays, J. Immunol. Methods 65
(1983) 55e63.
[37] P. Kele, X. Li, M. Link, K. Na, A. Herner, K. Lrincz, S. Beni, O.S. Wolfbeis,
Clickable fluorophores for biological labeling with or without copper, Org.
Biomol. Chem. 7 (2009) 3486e3490.
[10] A. Caricasole, A. Copani, F. Caraci, E. Aronica, A.J. Rozemuller, A. Caruso,
M. Storto, G. Gaviraghi, G.C. Terstappen, F. Nicoletti, Induction of Dickkopf-1, a
negative modulator of the Wnt pathway, is associated with neuronal
degeneration in Alzheimer's brain, J. Neurosci. 24 (2004) 6021e6027.
[11] I. Cappuccio, A. Calderone, C.L. Busceti, F. Biagioni, F. Pontarelli, V. Bruno,
M. Storto, G.T. Terstappen, G. Gaviraghi, F. Fornai, G. Battaglia, D. Melchiorri,
[38] H.J. Rideout, P. Dietrich, Q. Wang, W.T. Dauer, L. Stefanis, alpha-synuclein is
required for the fibrillar nature of ubiquitinated inclusions induced by pro-
teasomal inhibition in primary neurons, J. Biol. Chem. 279 (2004)
46915e46920.
S. Zukin, F. Nicoletti, A. Caricasole, Induction of Dickkopf-1,
modulator of the Wnt pathway, is required for the development of ischemic
neuronal death, J. Neurosci. 25 (2005) 2647e2657.
a negative
[12] S.A. Purro, E.M. Dickins, P.C. Salinas, The secreted Wnt antagonist Dickkopf-1
[39] D. Wu, Y. Zhang, P. Liu, X. Li, J. Zhang, J. Shan, D. Engelhardt, Compositions and
Methods for Bone Formation and Remodeling European Patent, 2012, p. 19
(EP1758562 B1).
[40] N.M. O'Boyle, M. Banck, C.A. James, C. Morley, T. Vandermeersch,
G.R. Hutchison, Open babel: an open chemical toolbox, J. Cheminform. 3
(2011) 33.
[41] P.C. Hawkins, A.G. Skillman, G.L. Warren, B.A. Ellingson, M.T. Stahl, Conformer
generation with OMEGA: algorithm and validation using high quality struc-
tures from the protein databank and Cambridge structural database, J. Chem.
Inf. Model 50 (2010) 572e584.
[42] M.R. McGann, H.R. Almond, A. Nicholls, J.A. Grant, F.K. Brown, Gaussian
docking functions, Biopolymers 68 (2003) 76e90.
[43] M. McGann, FRED pose prediction and virtual screening accuracy, J. Chem. Inf.
Model 51 (2011) 578e596.
[44] G.B. McGaughey, R.P. Sheridan, C.I. Bayly, J.C. Culberson, C. Kreatsoulas,
S. Lindsley, V. Maiorov, J.F. Truchon, W.D. Cornell, Comparison of topological,
shape, and docking methods in virtual screening, J. Chem. Inf. Model 47
(2007) 1504e1519.
[45] S. Dallakyan, A. Olson, Small-molecule library screening by docking with PyRx,
in: J.E. Hempel, C.H. Williams, C.C. Hong (Eds.), Chemical Biology, Springer,
New York, 2015, pp. 243e250.
is required for amyloid
3492e3498.
b-mediated synaptic loss, J. Neurosci. 32 (2012)
[13] H. Glantschnig, R.A. Hampton, P. Lu, J.Z. Zhao, S. Vitelli, L. Huang, P. Haytko,
T. Cusick, C. Ireland, S.W. Jarantow, R. Ernst, N. Wei, P. Nantermet, K.R. Scott,
J.E. Fisher, F. Talamo, L. Orsatti, A.A. Reszka, P. Sandhu, D. Kimmel, O. Flores,
W. Strohl, Z. An, F. Wang, Generation and selection of novel fully human
monoclonal antibodies that neutralize Dickkopf-1 (DKK1) inhibitory function
in vitro and increase bone mass in vivo, J. Biol. Chem. 285 (2010)
40135e40147.
[14] R. Nietzki, R. Otto, Zur Kenntniss der Indamine und Indophenole, Chem.
Berichte 21 (1888) 1736e1745.
[15] S. Iozzi, R. Remelli, B. Lelli, D. Diamanti, S. Pileri, L. Bracci, R. Roncarati,
A. Caricasole, S. Bernocco, Functional characterization of a small-molecule
inhibitor of the DKK1-LRP6 interaction, ISRN Mol. Biol. 2012 (2012), http://
dx.doi.org/10.5402/2012/823875. Article ID 823875.
[16] X. Li, J. Shan, W. Chang, I. Kim, J. Bao, H.J. Lee, X. Zhang, V.T. Samuel,
G.I. Shulman, D. Liu, J.J. Zheng, D. Wu, Chemical and genetic evidence for the
involvement of Wnt antagonist Dickkopf2 in regulation of glucose meta-
bolism, Proc. Natl. Acad. Sci. U. S. A. 109 (2012) 11402e11407.
[17] U. Hollstein, Actinomycin. Chemistry and mechanism of action, Chem. Rev. 74
(1974) 625e652.
[46] F. Lovering, J. Bikker, C. Humblet, Escape from flatland: increasing saturation
as an approach to improving clinical success, J. Med. Chem. 52 (2009)
6752e6756.
[47] C.A. Lipinski, Lead- and drug-like compounds: the rule-of-five revolution,
Drug Discov. Today Technol. 1 (2004) 337e341.
[18] G. Carr, W. Tay, H. Bottriell, S.K. Andersen, A.G. Mauk, R.J. Andersen, Plec-
tosphaeroic acids A, B, and C, indoleamine 2,3-dioxygenase inhibitors pro-
duced in culture by
a marine isolate of the fungus Plectosphaerella
cucumerina, Org. Lett. 11 (2009) 2996e2999.
[19] T. Shimamoto, A. Tomoda, R. Ishida, K. Ohyashiki, Antitumor effects of a novel
phenoxazine derivative on human leukemia cell lines in vitro and in vivo, Clin.
[48] C.A. Lipinski, F. Lombardo, B.W. Dominy, P.J. Feeney, Experimental and

38
S. Mpousis et al. / European Journal of Medicinal Chemistry 108 (2016) 28e38
computational approaches to estimate solubility and permeability in drug
discovery and development settings, Adv. Drug Deliv. Rev. 46 (2001) 3e26.
[49] Marvin, Marvin was used for Drawing, Displaying and Characterizing Chem-
ical Structures, Substructures and Reactions, Marvin 14.9.29, ChemAxon,
2014. http://www.chemaxon.com.
[50] C. Plugins, Calculator Plugins were used for Structure Property Prediction and
Calculation, Marvin 14.9.29, ChemAxon, 2014. http://www.chemaxon.com.
[51] O. Trott, A.J. Olson, AutoDock Vina: improving the speed and accuracy of
docking with
a new scoring function, efficient optimization, and multi-
threading, J. Comput. Chem. 31 (2010) 455e461.