Systematic diversification of benzylidene heterocycles yields novel inhibitor scaffolds selective for Dyrk1A, Clk1 and CK2

Article abstract of DOI:10.1016/j.ejmech.2016.02.017

The dual-specificity tyrosine-regulated kinase 1A (Dyrk1A) has gathered much interest as a pharmacological target in Alzheimer's disease (AD), but it plays a role in malignant brain tumors as well. As both diseases are multi-factorial, further protein kin

Full text of DOI:10.1016/j.ejmech.2016.02.017

Accepted Manuscript
Systematic diversification of benzylidene heterocycles yields novel inhibitor scaffolds
selective for Dyrk1A, Clk1 and CK2
Marica Mariano, Rolf W. Hartmann, Matthias Engel
PII:
S0223-5234(16)30087-3
10.1016/j.ejmech.2016.02.017
EJMECH 8368
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 14 December 2015
Revised Date: 4 February 2016
Accepted Date: 5 February 2016
Please cite this article as: M. Mariano, R.W. Hartmann, M. Engel, Systematic diversification of
benzylidene heterocycles yields novel inhibitor scaffolds selective for Dyrk1A, Clk1 and CK2, European
Journal of Medicinal Chemistry (2016), doi: 10.1016/j.ejmech.2016.02.017.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Graphical Abstract
Systematic diversification of benzylidene heterocycles yields
novel inhibitor scaffolds selective for Dyrk1A, Clk1 and
CK2
Leave this area blank for abstract info.
Marica Mariano, Rolf W. Hartmann, Matthias Engel
Pharmaceutical and Medicinal Chemistry, Saarbrücken, Germany
1
4
7
13
18
Dyrk1A
Clk1
Dyrk1A
CK2
Clk1

ACCEPTED MANUSCRIPT
European Journal of Medicinal Chemistry
journal homepage: www.elsevier.com
Systematic diversification of benzylidene heterocycles yields novel inhibitor scaffolds
selective for Dyrk1A, Clk1 and CK2
Marica Marianoa , Rolf W. Hartmanna,b and Matthias Engela,
a
Department of Pharmaceutical and Medicinal Chemistry, Saarland University, Saarbrücken, Germany.
b Helmholtz Institute for Pharmaceutical Research Saarland (HIPS), Campus C2-3, D 66123 Saarbrücken, Germany
ARTICLE INFO
ABSTRACT
Article history:
Received
Received in revised form
Accepted
The dual-specificity tyrosine-regulated kinase 1A (Dyrk1A) has gathered much interest as a
pharmacological target in Alzheimer´s disease (AD), but it plays a role in malignant brain
tumors as well. As both diseases are multi-factorial, further protein kinases, such as Clk1 and
CK2, were proposed to contribute to the pathogenesis. We designed a new class of α-
benzylidene–γ-butyrolactone inhibitors that showed low micromolar potencies against Dyrk1A
and/or Clk1 and a good selectivity profile among the most frequently reported off-target kinases.
A systematic replacement of the heterocyclic moiety gave access to further inhibitor classes with
interesting selectivity profiles, demonstrating that the benzylidene heterocycles provide a
versatile tool box for developing inhibitors of the CMGC kinase family members Dyr1A/1B,
Clk1/4 and CK2. Efficacy for the inhibition of Dyrk1A–mediated tau phosphorylation was
demonstrated in a cell-based assay. Multi-targeted but not non-specific kinase inhibitors were
also obtained, that co-inhibited the lipid kinases PI3Kα/γ. These compounds were shown to
inhibit the proliferation of U87MG cells in the low micromolar range. Based on the molecular
properties, the inhibitors described here hold promise for CNS activity.
Available online
Keywords:
Dyrk1A
PI3 kinase
Alzheimer
glioma
multi-targeted inhibitor
2009 Elsevier Ltd. All rights reserved
.
———
Corresponding author. E-mail: ma.engel@mx.uni-saarland.de; phone: +4968130270312.

vivo-studies. Therefore, a selective inhibitor of Dyrk1A is a
ACCEPTED MANUSCRIPT
1. Introduction
primary goal of drug development efforts, in particular with
respect to Down Syndrome–related neurodegeneration, for which
the overexpression of Dyrk1A is a main cause [24].
Rather selective Dyrk1A inhibitors have been described
previously, such as harmine, INDY, KH-CB19, Leucettines and
pyridinylthiophenes [25–29], however, they still inhibited Clk1/4
with equal potency. Recently, Falke et al. [30] reported some 10-
iodo-11H-indolo[3,2-c]quinoline-6-carboxylic acid derivatives as
highly selective inhibitors of Dyrk1A, but the carboxylic acid
function may impair CNS penetration [31].
The dual-specificity tyrosine-regulated kinases (Dyrk) and
CDC2-like kinases (Clk) both belong to the CMGC branch of the
kinome and share the ability to phosphorylate tyrosine as well as
serine/threonine residues; however, only the activating
autophosphorylation occurs actually at tyrosine.
The Dyrk1A isoform is expressed ubiquitously while the most
closely related homolog, Dyrk1B, is mainly found in testes and
muscle tissue [1]. In the last decade, Dyrk1A, but not Dyrk1B,
received much attention due to its implication in the development
of Down Syndrome–related and sporadic Alzheimer’s disease
(AD) [2,3], where it promotes neurodegeneration by hyper-
phosphorylating the microtubule-associated protein tau [4].
Furthermore, Dyrk1A was also shown to phosphorylate amyloid
precursor protein (APP) and presenilin-1, a key component of the
γ-secretase complex, thereby accelerating the formation of
neurotoxic Aβ peptides [5].
From a medicinal chemistry point of view, it is advantageous
that Dyrk1A can potently be inhibited by very small molecules
with molecular weights below 300 g/mol [25,26], suggesting that
the binding site can adapt to small ligands. This property
contributes to the attractiveness of Dyrk1A as a target for CNS
diseases, because as a rule of thumb, the smaller the inhibitor, the
higher are the chances to pass the blood brain barrier [31].
In the following, we describe the development of small
benzylidene derivatives, which cover a broad range of selectivity
and potency profiles, depending on the combined heterocycles.
Recently, Dyrk1A was also proposed as a new target in
glioblastoma [6]. Glioma are among the most malignant types of
solid tumors and survival rates for patients are still very poor:
less than 5% of patients survive for 5 years after having been
diagnosed with glioblastoma multiforme, the most aggressive
subtype [7]. Hence, more effective therapeutic agents are
urgently needed. Of note, Dyrk1A was reported to act as an
enhancer of receptor tyrosine kinase signaling; known examples
include FGFR [8] and EGFR [6,9]. Pozo and coworkers found
Dyrk1A to be strongly overexpressed in glioblastoma, paralleling
EGFR expression [6]. Dyrk1A prevents endocytosis–mediated
2. Results and discussion
2.1. Design strategy and synthesis
Two of the most important factors that determine the CNS
permeability of small molecules are the molecular mass, which
showed a median of 305 Da in a thorough analysis of CNS drugs,
and the number of H-bond donors, which should be ≤ 1 [31].
Therefore, our design of new Dyrk1A inhibitors was based on
previously published inhibitors that fulfilled the following
criteria: i) low mass, ii) no requirement for an H-bond donor
contacting the hinge region, and iii) good potency in cellular
systems. One of the most compact but yet potent and orally
available inhibitor of Dyrk1A is the β-carboline alkaloid harmine
(Figure 1). In vitro studies showed good activity against Dyrk1A,
particularly in cells, and a slight selectivity toward other
members of the Dyrk family. However, it also exerts several
adverse CNS effects due to its potent monoaminoxidase-A
inhibition [32]. The benzothiazole TG003 is a synthetic inhibitor
of Dyrk1A which was identified in a high-throughput screening
[26]. This compound is able to block the negative regulatory
activity of Dyrk1A in the NFAT pathway. Unfortunately, its
potency is not sufficient for in vivo applications (cell-free IC50 for
Dyrk1A: 0.8 µM) [21]. Another rather small compound with
inhibitory activity against Dyrks and Clk1 is leucettamine B,
which was isolated from a marine sponge [22].
degradation of EGFR by
a
mechanism that requires
phosphorylation of the EGFR signaling modulator Sprouty2 [9].
Moreover, Dyrk1A–catalyzed phosphorylation of FOXO (also
called FKHR) was shown to promote nuclear exclusion, thus
maintaining cell cycle progression and suppressing the activation
of pro-apoptotic genes [10,11]. Compared with healthy tissue, the
mRNA of Dyrk1A, but not of the homologous kinases Dyrk1B or
Dyrk2, was strongly overexpressed in glioblastoma, medullo-
blastoma, astrocytoma and oligodendroglioma (analyzed using
the Oncomine portal, www.oncomine.org) [12,13]. This suggests
that Dyrk1A could be an interesting anti-cancer target in EGFR–
overexpressing tumors.
The Clk1 and -4 isoforms do not only share a strong structural
homology with the Dyrk family but also some physiological
functions: Both are involved in the regulation of alternative
mRNA splicing through the phosphorylation of serine-arginine-
rich (SR) proteins, including AF2/ASF, SC35 and SRp55 [14–
18].
SR proteins mostly promote the inclusion of exons via binding to
exonic splicing enhancers, and their activity is mainly regulated
by multiple phosphorylations [19,20]. Upon dysregulation of the
kinase activities in the case of Alzheimer´s disease, the
alternative splicing of tau is influenced, which results in a
pathogenic imbalance between 3R-tau and 4R-tau isoforms
through skipping of exon 10 [17]. Thus, co-inhibition of Dyrk
and Clk kinases might efficiently correct the tau splicing isoform
imbalance. Indeed, we could show that dual inhibitors of Dyr1A
and Clk1 are more efficient pre-mRNA splicing modulators than
selective Dyrk1A or Clk1 inhibitors [21]. Furthermore, Leucet-
tine L41, a dual Dyrk1A/Clk1 inhibitor [22] was recently
reported to prevent memory impairments and neurotoxicity
induced by oligomeric Aβ25-35 peptide administration in mice
[23].
A superimposition of
harmine, INDY (the 5-hydroxy
derivative of TG003) and Leucettine L41 (a synthetic derivative
of Leucettamine B [22]) as bound to Dyrk1A in published
cocrystal structures revealed some common pharmacophoric
features, which were used as starting points for the design of new
Dyrk1A inhibitors (Figure 2). Obviously, two H-bond acceptor
(HBA) functions at the molecule ends and an aromatic system
connected to one HBA were sufficient for strong binding affinity
to the ATP binding cleft of Dyrk1A. The smallest possible linker
to connect two suitable moieties, which preserved the planar
overall geometry, was a methine bridge. This was already present
in the Leucettamine derivatives (Figure 1), in which the 2-
aminoimidazolone carbonyl oxygen was functioning as HBA for
Lys188. Employing Claisen–Schmidt chemistry, a large variety
of different HBA ring combinations was accessible, which were
expected to have a strong impact on the potency and selectivity
profile due to differences in lipophilicity, acidity and molecular
However, whether the benefit of such multi-targeted
approaches outweighs the potential accumulation of side effects
related to target inhibitions, will only be seen in long term in

Figure 2. In order to maintain a good probability of CNS
electrostatic potentials. Since a systematic sAtudCyCofEtPheTEmoDst MANUSCRIPT
straightforward combinations had not been done before, we
decided to synthesize and explore a set of new scaffolds
exhibiting the minimum pharmacophor features according to
penetration, the molecular weight was kept below 280 g/mol.
Table 1. Inhibitory activity of α-benzylidene-γ-butyrolactone derivatives against Dyrk1A and selectivity profiling against a panel of
selected kinases.
D1A
D1B
D2
CK2
Clk1
EGFR
PKCβ
#
Structure
(IC50, µM)a
1
1.81
0.93
1.10
1.20
0.96
0.17
0.13
0.88
0.60
1.08
0.13
3.80
6.70
5.80
10.5
2.17
2.64
1.05
4.29
2.63
1.68
0.18
>10
>10
>10
n.i.
n.i.
>10
>10
>10
>10
>10
>10
>10
>10
>10
n.d.
>10
n.i.
>10
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
n.d.
2
3
3.90
7.52
4.37
>10
>10
>10
>10
>10
0.17
0.22
4
n.d.
>10
>10
>10
>10
>10
>10
n.d.
5
0.82
0.43
0.06
2.04
0.99
>10
3.00
6
7
8
9
10
Hrm
aS.D.≤18%, [ATP]=100µM; D1A=Dyrk1A, D1B=Dyrk1B, D2=Dyrk2; Hrm=Harmine; n.i., no inhibition; n.d., not determined.
2.2. Structure-Activity relationships (SAR)
The first synthesized compound 1 was moderately active as
Dyrk1A inhibitor (Table 1) and quite selective against our panel
of kinases. It did not appreciably inhibit Dyrk2 and showed
weaker activity even against Dyrk1B. In order to increase its
activity and selectivity, we decided to rigidify the p-methoxy-
benzene into a chromane (compound 2). This non-planar,
partially saturated system was favorable for the inhibition of
Dyrk1A but less well tolerated by Dyrk1B, resulting in increased
selectivity towards the latter isoform. This modification may take
advantage – at least indirectly – of the only difference between
the ATP binding pockets of Dyrk1A and Dyrk1B, which is the
exchange of the linear side chain of hinge region residue Met240
in Dyrk1A by a branched leucine in Dyrk1B. Although pointing
away from the ATP pocket, the Leu side chain might restrict the
flexibility of some β-sheet residues that form the lid of the pocket
to adapt to sterically more demanding cyclic moieties. An
extension of the lactone ring to a six-membered cycle (3) or by a
methyl substituent (4) confirmed the trend that the ATP-binding
First, several γ-butyrolactone moieties were combined with
different benzylidene units, and the influence on potency and
selectivity was tested. The kinases chosen for the primary
screening panel were (i) frequently reported off-targets for
diverse classes of Dyrk1A inhibitors (first five kinases in Tables
1 and 2, all from the CMGC family) [21], and (ii) kinases from
more distant families (last two in Tables 1 and 2). In this way,
conclusive selectivity data were immediately obtained.
Figure 1. 2D structures of harmine, INDY and Leucettamine B.

pocket of Dyrk1A accepts more bulky structures than that of
phosphorylation assays. The next most strongly inhibited kinase
was PI3Kα, with an IC50 of 0.65 µM (determined by a substrate
phosphorylation assay in the presence of 100 µM ATP). The PI3
lipid kinases had been included in the screening panel for reasons
stated below.
ACCEPTED MANUSCRIPT
Dyrk1B, as indicated by the 9 times higher potency of 4 for
Dyrk1A. Consistent with this finding, the less bulky coumaran
moiety in 5 caused a drop of the selectivity towards Dyrk1B
(compared with 2) – and particularly towards Dyrk2 –, while the
inhibition of Dyrk1A was conserved. Reintroducing a methyl
substituent in 6 not only recovered the selectivity toward Dyrk1B
but markedly increased the potency toward Dyrk1A (IC50: 0.17
µM). Thus, the potency was similar to that of harmine but the
selectivity towards Dyrk1B was considerably improved, reaching
a remarkable selectivity factor of 15.5 (cf. Table 1). The 5-methyl
substituent in 6 increased the selectivity in general, which
became particularly evident from the negligible inhibition of the
atypical kinase haspin by 6 (43% at 5 µM) compared to 7 (75%
at 5 µM) (Table 3). Haspin was reported as a notorious off-target
even for some of the most selective Dyrk1A inhibitors, including
TG003 and harmine [27,33]. An extended selectivity profiling for
compound 6 revealed that besides Dyrk1A and Clk1, only Clk4
was affected in our panel (Table S1, Supplementary Material).
The IC50 for Clk4 as determined by the LanthaScreen® Kinase
Binding Assay was 32 nM. However, this assay is based on the
competition of the test compound with an Alexa Fluor®–
conjugated tracer compound for binding to the ATP site, thus the
IC50 is difficult to compare with IC50s obtained by peptide
Because 6 still significantly inhibited Clk1 (and probably
Clk4), we tried to further optimize the selectivity by modifying
the benzylidene part of the molecule. Our strategy of including a
second oxygen atom was unsuccessful as both potency and
selectivity decreased (cf. compounds 8 and 9). Subsequently, we
decided to replace the coumaran of compound
5 by a
benzothiazole moiety, leading to a potent inhibitor (7) with dual
selectivity for Dyrk1A and Clk1. Surprisingly, our attempt to
further increase the potency and selectivity for Dyrk1A by
introducing
a methyl on the γ-butyrolactone ring failed
(compound 10). Probably the methyl group forced the compound
into a slightly different binding orientation, which negatively
affected the H-bond interaction between the benzothiazole
nitrogen and the hinge region. The distinct behavior of compound
6 might be explained by the different spatial orientation of the
oxygen lone pairs. However, the affinity to Clk1 was less
affected, thus rendering compound 10 moderately selective for
Clk1.
Table 2. Inhibitory activity of benzylidene thioxothiazolidinone–based derivatives against Dyrk1A and selectivity
profiling against a panel of selected kinases.
D1A
D1B
D2
CK2
Clk1
EGFR
PKCβ
#
A
B
(IC50, µM)a
11
>10
6.4
n.d.
>10
5.2
n.d.
n.i.
n.d.
n.i.
n.d.
n.d.
n.d.
n.i.
n.d.
n.i.
n.i.
n.i.
>10
9.2
12
13
14
15
16
17
10
3.6
5.7
0.5
0.1
0.2
1.3
9.2
0.17
0.22
0.1
0.25
0.1
1.5
>10
1.0
n.i.
0.3
>10
>10
>10
0.04
0.02
0.07
0.06
0.20
0.06
0.06
0.04
>10
18
0.7
0.9
1.5
2.4
0.02
>10
>10
>10
n.i.
AS
0.03
0.08
0.1
0.02
0.07
aS.D.≤18%, [ATP]=100µM; D1A=Dyrk1A, D1B=Dyrk1B, D2=Dyrk2, AS= AS605240; n.d., not determined; n.i., no inhibition.

benzothiazole (16) and quinoline (17) rings were introduced. The
Our second set of compounds (Table 2) AwCasCdeEsiPgnTedEDby MANUSCRIPT
modifying the lactone part of the molecule in order to establish
stronger interactions in the binding site and thus boost the
potency. For the first compounds, we decided to use again 4-
methoxybenzaldehyde as coupling partner, since the plain
methoxy substituent on the benzene had conferred the highest
selectivity towards Clk1 so far (cf. compounds 1 and 3 in Table
1). The first designed compound 11 was not active, suggesting
that the combination of two carbonyls in the pyrrolidinedione
was too polar for the binding site (Table 2). Similarly, the
hydantoin derivative 12 exhibited only a weak inhibitory activity
toward Dyrk1A. Nonetheless, when the NH heteroatom unit was
replaced by sulfur leading to the thiazolidinedione ring,
appreciable inhibition was observed (compound 13).
Interestingly, 13 was mainly selective for CK2, exhibiting an IC50
of 0.17 µM. Since CK2 is known to preferably bind acidic
compounds in the ATP-binding site [34,35], the relatively low
pKa value of the thiazolidinedione group (reported to be 6.14 for
the unsubstituted analog [36]) may account for this selectivity.
Indeed, the NH acidity of the CK2-inactive analog 12 is
considerably lower (pka value: 8.9 [37]). We further explored the
effects of sulfur atoms by incorporating the rhodanine system (2-
thioxothiazolidin-4-one). This resulted in 5-fold increase in the
Dyrk1A inhibitory potency (compare compound 14 with 13),
while the CK2 inhibition remained unchanged. This was again in
accordance with the acidic pka of 14, which was previously
reported to be 5.8 [38]. Thus, 14 was a dual inhibitor of Dyrk1A
and CK2; of note, the overall selectivity did not drop, even
though the rhodanine system was frequently present in hits
identified for many different enzymes [39]. In particular, the high
selectivity (23-fold) of 14 over Clk1 was remarkable. CK2 is also
believed to be involved in the pathology of Alzheimer´s disease
[40–42]. Thus, 14 might allow to evaluate whether simultaneous
inhibition of Dyrk1A and CK2 has any benefit in Alzheimer´s
disease models without significantly affecting Clk1.
resulting nanomolar inhibitors 16 and 17 were multi-targeted
within the CMGC kinase family, but not promiscuous, since
PKCβ and EGFR kinase were not affected.
Because the benzothiazole moiety in 16 generated a moderate
selectivity over Dyrk2, it was also combined with thiohydantoin
(cpd. 18) to test whether this favorable inhibitory profile and the
potency of the rhodanine analog 16 might be retained. While the
activities against Dyrk1A and CK2 dropped, compound 18
displayed an impressive selectivity for Clk1 (35 fold over
Dyrk1A). Thus, our scaffold diversification identified
thiohydantoin as a novel moiety mediating strong and selective
affinity for Clk1.
Because of the obvious structural similarity of our compounds
16 and 17 with the thiazolidinedione AS605240 (Table 2), which
was previously described as a rather selective PI3Kγ inhibitor
(IC50: 8 nM) [43], it was straightforward to enlarge our panel of
screened kinases by some of the PI3K isoforms; in turn, we
tested AS605240 against our kinase targets. Indeed, AS605240
turned out to be a nanomolar inhibitor of Dyrk1A and CK2 as
well (Table 2), and not surprisingly, both 16 and 17 were highly
active against PI3Kγ and –α (Table 3). Compound 17 was
slightly more selective than 16, exhibiting lower activity toward
PI3Kδ and Clk1 (Tables 2 and 3; see also extended screening in
Table S1, Supplementary Material). Of note, both for 16 and 17,
a reduced activity against the atypical kinase haspin was noted
(Table 3), corroborating that the rhodanine moiety does not a
priori confer non-specific behaviour. Collectively, we identified
CK2α, Clk1, STK17A and PI3Kα/γ as additional targets for 16
and 17 (Tables 2, 3 and S1), of which, however, PI3Kα/γ turned
out to be most potently inhibited. Literature data on the
pathogenic roles of these kinases in glioma suggest that they can
all be considered as potential targets in this brain malignancy
(Table S2, Supplementary Material). In light of this, the
compounds presented here might display a favorable profile as
multi-target agents for the treatment of glioma. Indeed, we found
that 16, 17 and AS605240 inhibited the growth of U87MG
glioma cells with submicromolar IC50s, being more potent than
the rather Dyrk1A/1B–selective reference compound harmine
(Table S3, Supplementary Material).
Table 3. Extended selectivity profiling of 16 and 17.
16
17
Kinase
IC50 (nM)b
STK17A (DRAK1)
PIM1
n.d.
12.8
n.d.
103
Haspin
255 (93%a) 259 (93%a)
PI3K-C2β
56%a
62%a
Figure 2. Superimposition of harmine (green), INDY (cyan)
and Leucettine L41 (magenta) cocrystallized with Dyrk1A
(PDB codes 3ANR, 3ANQ and 4AZE, respectively). All
ligands are bound in the ATP-binding cleft via two H-bonds,
formed with Leu241-NH at the hinge region and Lys188
(indicated by the grey dashed lines). The measured ranges of
the N–O distances are given in Å. Additional contacts
comprise van der Waals-interactions with the side chains of
Ala186, Val173 and Leu294, which mainly involve the left
aromatic rings.
PI3K-p110-α
PI3K-p110-δ
PI3K-p110-γ
2.2
2.6
12.4
2.3
16.1
2.3
a% inhibition at 1 µM; shown are mean values of two different
measurements performed in duplicates (S.D. ≤ 2.5%); bIC50s were only
determined when the % inhibition at 1 µM was greater than 85%. Data
were calculated based on mean values of duplicates that differed by
less than 7%.
In the course of our study, we found that the selectivity of the
rhodanine moiety was strongly modulated by the extent of the
aromatic system. The selectivity toward Dyrk1B and Dyrk2 was
reduced stepwise when the coumaran (15) and particularly the
To the best of our knowledge, this is the first report showing
that AS605240 is a strong inhibitor of Dyrk1A, CK2 and Clk1,

and that dual inhibitors of PI3 lipid kinases and protein kinases
were chosen, possessing a plasma-stable [50] and a labile ester
function [51], respectively. As can be seen from Table 4, all
analyzed compounds exhibited a long half-life in rat plasma,
similar to or better than that of diltiazem (t1/2 = 125 min),
indicating a high metabolic stability in plasma. In contrast, N-
acetyl-L-phenylalanine ethyl ester was completely metabolized
within a few minutes, as it was expected. Thus, it can be
concluded that the lactone and lactam rings of our compounds are
not hydrolyzed by plasma enzymes.
ACCEPTED MANUSCRIPT
can be developed. However, inhibition of multiple kinases might
be associated with severe side effects. In this regard, it is
noteworthy that AS605240, which exhibits an in vitro inhibitory
profile similar to 16 and 17, found a widespread use in many
murine models of inflammatory diseases, including rheumatoid
arthritis, systemic lupus erythematosus, atherosclerosis, colitis,
liver fibrosis and pulmonary fibrosis [44–46]. No signs of
toxicity were observed in any of these studies, even though some
used high oral dosage, e. g., 50 mg/kg twice per day over 14 days
[44]. Moreover, in a xenograft model using PI3Kγ–expressing
neuroblastoma cells in SCID mice, AS605240 effectively
suppressed tumor growth at 20 mg/kg [47]. In the light of our
findings – although the inhibition of PI3Kγ may account for a
large part of the observed therapeutic effects – it is fair to ask
retrospectively what contribution was made by the inhibition of
the kinase targets that were identified for AS605240 in our
screening, Dyrk1A, CK2α and Clk1. At any rate, the absence of
toxic effects of AS605240 proves that in rodents, multi-targeted
inhibitions involving even a pleiotropic kinase such as CK2, are
not harmful to the organism, at least in short-term treatments.
Figure 3. Inhibition of Dyrk1A–catalyzed tau-Thr212 phosphorylation in
stably transfected HEK293 cells. The test compounds, DMSO or the
reference inhibitor harmine [54] were added to the cell medium and the
Dyrk1A expression induced as described. After immunoblotting of the total
cell proteins, the level of phospho-tau-Thr212 was detected using a
phosphospecific antibody. To normalize the signals, total recombinant GFP-
tau protein was quantified on the same blot using an anti-GFP antibody
(fluorescence image converted to grey scales). Initial screenings were
Having discovered new Dyrk1A inhibitory scaffolds with
diverse selectivity profiles, we were interested to evaluate their
potency to inhibit the Dyrk1A–catalyzed tau phosphorylation in
intact cells. To this end, we used a HEK293 cell line which
overexpressed both Dyrk1A and human tau protein. Under the
chosen conditions, phosphorylation of tau at Thr212 is solely
dependent on the Dyrk1A activity [48]. Testing of the most
potent Dyrk1A inhibitors from Tables 1 and 2, 6, 7, 16 and 17,
along with AS605240, revealed that the tau phosphorylation was
suppressed at low micromolar and sub-micromolar
concentrations (Figure 3). There was no obvious correlation with
the inhibitory activities of the compounds against purified
Dyrk1A, since the most potent compounds in the cell free assay,
16, 17 and AS506240 were inferior to harmine and compound 7.
Probably the distinct physicochemical properties of the scaffolds
accounted for variable degrees in cell permeability and further
pharmacokinetic properties. The high cellular potency of 7 was
encouraging, suggesting that benzothiazolylmethylene-γ-
butyrolactone might be a suitable core structure for the
development of novel therapeutic agents against Alzheimer´s
disease and also against diseases involving abnormal pre-mRNA
splicing, since Clk1 was strongly co-inhibited by 7 in the cell free
assay (Table 1). Given the low molecular weight (231.3 g/mol),
the small polar surface area (tPSA= 38.66 Å2) and the absence of
H-bond donors, the compound fulfills the major criteria for CNS
availability [31].
performed at
1 µM compound concentrations (lower panel; one
representative experiment out of two is shown). aIC50 values were
determined after cell treatment with different concentrations in triplicates,
referring to the signal derived from DMSO–treated cells as 100%. Values
represent averages from at least two independent determinations (±S.D.);
AS, AS605240.
3. Conclusions
Systematic variation of benzylidene–coupled heterocycles,
differing in acceptor strength, lipophilicity and electrostatic
potential, was highly effective to generate new scaffolds with
different kinase selectivities and potencies, while keeping the
molecular weight below 280 g/mol. Although we could not
provide experimental evidence for
a blood–brain–barrier
penetration of our inhibitors within the scope of this study, the
close structural similarity to some benzylidene derivatives
previously reported to be brain–permeable suggests that at least
some of our lead compounds might share this property, including
the benzylidene hydantoins (similar to compound 28 in [52]) and
the thiazolidinediones that are structurally similar to AS605240
[53].
However, most of our new scaffolds exhibited two
metabolically weak spots, which might impede in vivo
applications. One was the lactone ring structure, which might be
hydrolyzed by esterases present in the serum and/or cells, and the
other one was the potential Michael acceptor motif, which may
also have toxicological implications. The Michael-type acceptor
reactivity of some of the scaffolds disclosed as kinase inhibitors
in this study, including 5-benzylidene thiazolidinediones,
hydantoins, thiohydantoins and rhodanines, had been analyzed
previously by Arsovska et al [49]. The authors found that the 5-
benzylidene five-membered oxo-heterocycles revealed almost
insignificant reactivity toward cysteamine as an exemplary
biological nucleophile.
With respect to novel agents against neurodegenerative
diseases, compound 7 displayed the most promising activity in
our cell–based tau phosphorylation assay. In a cytotoxicity assay,
7 was not toxic toward HEK293 cells at concentrations up to 20
µM (data not shown). The compound might be useful as a tool to
further explore the effects of co-inhibition of Dyrk1A and Clk1/4
on the pathogenic mechanisms of Alzheimer´s disease. Other
compounds, such as 3 and 4, were not very potent but showed an
appreciable selectivity towards Dyrk1B and Clk1. These
compounds might be considered as elaborated, selective
fragments with still good ligand efficiencies (0.55 and 0.54,
respectively). In order to optimize the potency, they might be
Hydrolytic instability was mainly a concern for the lactone
derivatives in Table 1. Therefore we decided to test the stability
of some selected compounds in rat plasma. As reference
compounds, diltiazem and N-acetyl-L-phenylalanine ethyl ester
extended
analogously
to
the
leucettines,
through
functionalization e. g., at position 5 of the γ-butyrolactone in 4.

Synthesis of (E)-3-((2,3-dihydrobenzofuran-5-yl)methylene)-5-
methyldihydrofuran-2(3H)-one (6)
Within our panel of frequent off-target kinases for Dyrk1A
inhibitors, the thiazolidinedione 13 and thiohydantion 18 were
selective for CK2, probably because of their elevated acidity. The
presence of thione sulfur in the five-membered ring generally
increased the potency, however, mostly at the expense of
selectivity. The loss of selectivity remained limited when the
aromatic portion of the benzylidene moiety was restricted to a
single ring (cf. compounds 14 and 16). Upon introducing a
bicyclic aromatic system, however, the selectivity toward
Dyrk1B, Dyrk2 and CK2 was abolished, yielding potent, multi-
targeted inhibitors with enhanced anti-proliferative effect on the
U87MG glioma cell line (Table S3, Supplementary Material).
Interestingly, this class of compounds included the widely
studied AS605240, formerly reported as rather selective PI3Kγ
inhibitor. Altogether, our results showed that inhibitors which
simultaneously target PI3 kinases and oncogenic protein kinases
can be developed, with potential benefit in anti-cancer
applications.
ACCEPTED MANUSCRIPT
To a solution of NaH in dry toluene (0.162 g, 4.05 mmol), a
solution of the δ-valerolactone (0.324 g, 3.24 mmol, 0.309 mL)
was added, and the mixture stirred overnight at room
temperature. On the next day, the mixture was cooled in an ice
bath and 2,3-dihydrobenzo-furan-5-carboxaldehyde (0.300 g,
2.025 mmol) dissolved in dry toluene was added dropwise. The
temperature was slowly increased to reflux until the starting
materials were consumed. After cooling down, a solution of 10%
H2SO4 was added and the solution stirred at room temperature for
few minutes. Then 10 mL of water were added and the mixture
extracted 3 times with ethyl acetate (10 mL). The organic phase
was washed with brine and dried over Na2SO4. After evaporating
the solvent in vacuo, crude product was purified the column
1
chromatography. Yield: 31%. Mp: 131°C. H NMR (500 MHz,
DMSO-d6) δ ppm 1.36 (d, J=6.31 Hz, 3 H) 2.72 - 2.84 (m, 1 H)
3.23 (t, J=8.67 Hz, 2 H) 3.33 - 3.40 (m, 1 H) 4.60 (t, J=8.67 Hz, 2
H) 4.69 - 4.79 (m, 1 H) 6.87 (d, J=8.20 Hz, 1 H) 7.35 (s, 1 H)
7.37 - 7.41 (m, 1 H) 7.51 (s, 1 H). 13C NMR (126 MHz, DMSO-
d6) δ ppm 22.03 (s, 1 C) 28.66 (s, 1 C) 34.63 (s, 1 C) 71.66 (s, 1
C) 73.64 (s, 1 C) 109.37 (s, 1 C) 122.20 (s, 1 C) 126.72 (s, 1 C)
127.18 (s, 1 C) 128.47 (s, 1 C) 131.34 (s, 1 C) 135.15 (s, 1 C)
161.17 (s, 1 C) 171.65 (s, 1 C). LC-MS (ESI): m/z MH+= 231.
Table 4. Metabolic stability in rat plasma.
t
1/2 in rat plasma
Compound
[min]a
6
≫ 120b
≫ 120b
186
5
General procedure A: rhodanine or hydantoine (1 eq), the
corresponding aldehyde (1 eq) and sodium acetate (3 eq) were
dissolved in 5 ml of acetic acid and the solution heated at 110°C,
if necessary under microwave conditions (130W for ~10
minutes). After cooling, the precipitate was filtered off, washed
to remove all the acetic acid and dried in vacuo. The reaction has
been described in Mendgen et al[39] to be stereospecific
producing the Z-isomer as confirmed by the single peak in the
HPLC.
7
16
101
Diltiazem
125
N-acetyl-L-phenylalanine
< 5
ethyl ester
b
amean values from two experiments, S.D.< 17%; no degradation
Synthesis
of
(Z)-5-(benzo[d]thiazol-6-ylmethylene)-2-
detected after 120 min
thioxothiazolidin-4-one (16)
Rhodanine (0.254 g, 1.84 mmol), 1,3-benzothiazole-6-
carbaldehyde (0.300 g, 1.84 mmol) and sodium acetate (0.452 g,
5.51 mmol) were used following the general procedure A
described above. Yield: 2%. Mp: > 300°C. 1H NMR (500 MHz,
DMSO-d6) δ ppm 7.76 (d, J=8.51 Hz, 1 H) 7.78 (s, 1 H) 8.21 (d,
J=8.51 Hz, 1 H) 8.47 (s, 1 H) 9.50 - 9.58 (m, 1 H) 13.88 (br. s., 1
H). 13C NMR (126 MHz, DMSO-d6) δ ppm 123.74 (s, 1 C)
125.09 (s, 1 C) 125.97 (s, 1 C) 128.30 (s, 1 C) 130.26 (s, 1 C)
131.13 (s, 1 C) 134.94 (s, 1 C) 153.93 (s, 1 C) 159.56 (s, 1 C)
169.43 (s, 1 C) 195.71 (s, 1 C). LC-MS (ESI): m/z MH+= 279.
4. Methods
4.1. Chemistry. Synthesis of the key compounds 6, 16 and 17
Synthesis
of
(Z)-5-(quinolin-6-ylmethylene)-2-
thioxothiazolidin-4-one (17)
Rhodanine (0.254 g, 1.909 mmol), 6-quinoline-carbaldehyde
(0.300 g, 1.909 mmol) and sodium acetate (0.470 g, 5.73 mmol)
were used following the general procedure A described above.
Yield: 90%. Mp: >300°C. 1H NMR (500 MHz, DMSO-d6) δ ppm
7.62 (dd, J=8.35, 4.26 Hz, 1 H) 7.82 (s, 1 H) 7.95 (dd, J=8.83,
2.21 Hz, 1 H) 8.13 (d, J=8.83 Hz, 1 H) 8.24 (d, J=2.21 Hz, 1 H)
8.52 - 8.56 (m, 1 H) 8.98 (dd, J=4.26, 1.73 Hz, 1 H) 13.69 - 14.22
(m, 1 H). 13C NMR (126 MHz, DMSO-d6) δ ppm 122.53 (s, 1 C)
126.93 (s, 1 C) 127.13 - 128.39 (m, 1 C) 130.06 (s, 1 C) 130.25 -
130.44 (m, 1 C) 130.55 - 130.76 (m, 1 C) 131.03 (s, 1 C) 131.15
(s, 1 C) 136.95 (s, 1 C) 147.74 (s, 1 C) 152.34 (s, 1 C) 169.45 (s,
1 C) 195.72 (s, 1 C) LC-MS (ESI): m/z MH+= 273.
4.2. Biology

EGFR and reduces EGFR-dependent glioblastoma growth, J. Clin.
Invest. 123 (2013) 2475–2487. doi:10.1172/JCI63623.
Kinase assay. Inhibition assays for Dyrk1A, Dyrk1B, Dyrk2,
CK2alpha, Clk1, EGFR and PKCβ were performed using
purified recombinant kinase protein in the presence of 100 µM
ATP as described previously [29].
ACCEPTED MANUSCRIPT
[7]
S. Agarwal, R.K. Mittapalli, D.M. Zellmer, J.L. Gallardo, R.
Donelson, C. Seiler, S.A. Decker, K.S. Santacruz, J.L. Pokorny,
J.N. Sarkaria, W.F. Elmquist, J.R. Ohlfest, Active efflux of
Dasatinib from the brain limits efficacy against murine
glioblastoma: broad implications for the clinical use of molecularly
targeted agents, Mol. Cancer Ther. 11 (2012) 2183–2192.
doi:10.1158/1535-7163.MCT-12-0552.
HEK293-tau-Dyrk1A cell assay. The cell line was developed,
and the assays for the inhibition of cellular Dyrk1A activity were
performed as described previously[48]. In brief, HEK293-tau-
Dyrk1A cells which stably overexpress EGFP-fused human full
length tau protein and tetracycline–inducible EGFP–fused
Dyrk1A were grown in 12-well plates to 70% confluency; then
tetracycline (0.3 µg/mL) and test compounds were added from
stock solutions. The next day, the cells were lysed and the lysates
analyzed by immunoblotting using an anti-phospho-tau-Thr212
antibody (pT212, Life Technologies GmbH, Cat No. 444-740g,
dil. 1:1000) and a mouse anti-GFP antibody (Life Technologies
GmbH, Cat. No. 33-2600, dil. 1:1000) for normalization of the
signals. After incubation with dye-labeled secondary antibodies,
the signals for pT212 and EGFP were simultaneously detected
using an Odyssey infrared imager (LI-COR). Figure 3 shows the
image converted to grey scales.
[8]
[9]
S. Aranda, M. Alvarez, S. Turró, A. Laguna, S. de la Luna,
Sprouty2-mediated inhibition of fibroblast growth factor signaling
is modulated by the protein kinase DYRK1A, Mol. Cell. Biol. 28
(2008) 5899–5911. doi:10.1128/MCB.00394-08.
S.R. Ferron, N. Pozo, A. Laguna, S. Aranda, E. Porlan, M. Moreno,
C. Fillat, S. de la Luna, P. Sánchez, M.L. Arbonés, I. Fariñas,
Regulated segregation of kinase Dyrk1A during asymmetric neural
stem cell division is critical for EGFR-mediated biased signaling,
Cell Stem Cell. 7 (2010) 367–379. doi:10.1016/j.stem.2010.06.021.
[10]
[11]
[12]
Y.L. Woods, G. Rena, N. Morrice, A. Barthel, W. Becker, S. Guo,
T.G. Unterman, P. Cohen, The kinase DYRK1A phosphorylates the
transcription factor FKHR at Ser329 in vitro, a novel in vivo
phosphorylation site, Biochem. J. 355 (2001) 597–607.
H.S. Chang, C.H. Lin, C.H. Yang, M.S. Yen, C.R. Lai, Y.R. Chen,
Y.L. Liang, W.C. Yu, Increased expression of Dyrk1a in HPV16
immortalized keratinocytes enable evasion of apoptosis, Int. J.
Cancer. 120 (2007) 2377–2385. doi:10.1002/ijc.22573.
Metabolic stability tests in rat plasma. The test or reference
compounds (1 µM) were incubated with rat plasma (pooled,
heparinized) at 37 °C for 0, 5, 15, 30, 60, and 120 min. The
incubation was stopped by precipitation of plasma proteins with
5 volumes of cold acetonitrile containing an internal standard,
and the remaining compound concentration was analyzed by LC-
MS/MS. N-Acetyl-L-phenylalanine ethyl ester (Cat. No. A4251)
and diltiazem hydrochloride (Cat. No. D2521) were from Sigma.
S.L. Pomeroy, P. Tamayo, M. Gaasenbeek, L.M. Sturla, M.
Angelo, M.E. McLaughlin, J.Y.H. Kim, L.C. Goumnerova, P.M.
Black, C. Lau, J.C. Allen, D. Zagzag, J.M. Olson, T. Curran, C.
Wetmore, J.A. Biegel, T. Poggio, S. Mukherjee, R. Rifkin, A.
Califano, G. Stolovitzky, D. N. Louis, J.P. Mesirov, E.S. Lander,
T.R. Golub, Prediction of central nervous system embryonal
tumour outcome based on gene expression, Nature 415 (2002) 436–
442. doi:10.1038/415436a.
Calculation of ligand efficiency. The calculations were done
by dividing the Gibbs free energy of binding by the number of
non-hydrogen atoms. Ki values were calculated from the IC50s
using the Cheng-Prussoff equation. The KM value of our Dyrk1A
preparation for ATP was previously determined to be 64 µM
[29].
[13]
[14]
R. Shai, T. Shi, T.J. Kremen, S. Horvath, L.M. Liau, T.F.
Cloughesy, P.S. Mischel, S.F. Nelson, Gene expression profiling
identifies molecular subtypes of gliomas, Oncogene 22 (2003)
4918–4923. doi:10.1038/sj.onc.1206753.
J. Shi, T. Zhang, C. Zhou, M.O. Chohan, X. Gu, J. Wegiel, J. Zhou,
Y.W. Hwang, K. Iqbal, I. Grundke-Iqbal, C.X. Gong, F. Liu,
Increased dosage of Dyrk1A alters alternative splicing factor
(ASF)-regulated alternative splicing of tau in Down syndrome, J.
Biol. Chem. 283 (2008) 28660–28669.
Bibliography
[1]
[2]
S. Aranda, A. Laguna, S. de la Luna, DYRK family of protein
kinases: evolutionary relationships, biochemical properties, and
functional roles, FASEB J. 25 (2011) 449–462. doi:10.1096/fj.10-
165837.
doi:10.1074/jbc.M802645200.
[15]
[16]
W. Qian, H. Liang, J. Shi, N. Jin, I. Grundke-Iqbal, K. Iqbal, C.X.
Gong, F. Liu, Regulation of the alternative splicing of tau exon 10
by SC35 and Dyrk1A, Nucleic Acids Res. 39 (2011) 6161–6171.
doi:10.1093/nar/gkr195.
W.J. Song, L.R. Sternberg, C. Kasten-Sportès, M.L. Keuren, S.H.
Chung, A.C. Slack, D.E. Miller, T.W. Glover, P.W. Chiang, L.
Lou, D.M. Kurnit, Isolation of human and murine homologues of
the Drosophila minibrain gene: human homologue maps to 21q22.2
in the Down syndrome “critical region”, Genomics 38 (1996) 331–
339. doi:10.1006/geno.1996.0636.
X. Yin, N. Jin, J. Gu, J. Shi, J. Zhou, C.-X. Gong, K. Iqbal, I.
Grundke-Iqbal, F. Liu, Dual-specificity tyrosine phosphorylation-
regulated kinase 1A (Dyrk1A) modulates serine/arginine-rich
protein 55 (SRp55)-promoted Tau exon 10 inclusion, J. Biol.
Chem. 287 (2012) 30497–30506. doi:10.1074/jbc.M112.355412.
[3]
[4]
R. Kimura, K. Kamino, M. Yamamoto, A. Nuripa, T. Kida, H.
Kazui, R. Hashimoto, T. Tanaka, T. Kudo, H. Yamagata, Y.
Tabara, T. Miki, H. Akatsu, K. Kosaka, E. Funakoshi, K.
Nishitomi, G. Sakaguchi, A. Kato, H. Hattori, T. Uema, M. Takeda,
The DYRK1A gene, encoded in chromosome 21 Down syndrome
critical region, bridges between beta-amyloid production and tau
phosphorylation in Alzheimer disease, Hum. Mol. Genet. 16 (2007)
15–23. doi:10.1093/hmg/ddl437.
[17]
[18]
A.M. Hartmann, D. Rujescu, T. Giannakouros, E. Nikolakaki, M.
Goedert, E.M. Mandelkow, Q.S. Gao, A. Andreadis, S. Stamm,
Regulation of alternative splicing of human tau exon 10 by
phosphorylation of splicing factors, Mol. Cell. Neurosci. 18 (2001)
80–90. doi:10.1006/mcne.2001.1000.
P. Jain, C. Karthikeyan, N.S. Hari Narayana Moorthy, D. Kumar
Waiker, A. Kumar Jain, P. Trivedi, Human CDC2-Like Kinase 1
(CLK1): A Novel Target for Alzheimer’s Disease, Curr. Drug
Targets 15 (2014) 539–550.
Y.L. Woods, P. Cohen, W. Becker, R. Jakes, M. Goedert, X. Wang,
G.C. Proud, The kinase DYRK phosphorylates protein-synthesis
initiation factor eIF2Bepsilon at Ser539 and the microtubule-
associated protein tau at Thr212: potential role for DYRK as a
glycogen synthase kinase 3-priming kinase, Biochem. J. 355 (2001)
609–615.
[19]
[20]
S. Stamm, S. Ben-Ari, I. Rafalska, Y. Tang, Z. Zhang, D. Toiber,
T.A. Thanaraj, H. Soreq, Function of alternative splicing, Gene 344
(2005) 1–20. doi:10.1016/j.gene.2004.10.022.
[5]
[6]
F.J. Tejedor, B. Hämmerle, MNB/DYRK1A as a multiple regulator
of neuronal development, FEBS J. 278 (2011) 223–235.
doi:10.1111/j.1742-4658.2010.07954.x.
T. Misteli, J.F. Cáceres, J.Q. Clement, A.R. Krainer, M.F.
Wilkinson, D.L. Spector, Serine phosphorylation of SR proteins is
required for their recruitment to sites of transcription in vivo, J. Cell
Biol. 143 (1998) 297–307.
N. Pozo, C. Zahonero, P. Fernández, J.M. Liñares, A. Ayuso, M.
Hagiwara, A. Pérez, J.R. Ricoy, A. Hernández-Laín, J.M.
Sepúlveda, P. Sánchez-Gómez, Inhibition of DYRK1A destabilizes
[21]
C. Schmitt, P. Miralinaghi, M. Mariano, R.W. Hartmann, M. Engel,
Hydroxybenzothiophene Ketones Are Efficient Pre-mRNA

ACCEPTED MANUSOC.PR. KIuPkhTarenko, I.M. Kotey, O.V. Ostrynska, S.M. Yarmoluk,
Splicing Modulators Due to Dual Inhibition of Dyrk1A and Clk1/4,
ACS Med. Chem. Lett. 5 (2014) (9) 963–967.
doi:10.1021/ml500059y.
Synthesis and biological evaluation of substituted (thieno[2,3-
d]pyrimidin-4-ylthio)carboxylic acids as inhibitors of human
protein kinase CK2, Eur. J. Med. Chem. 46 (2011) 870–876.
doi:10.1016/j.ejmech.2010.12.025.
[22]
[23]
M. Debdab, F. Carreaux, S. Renault, M. Soundararajan, O.
Fedorov, P. Filippakopoulos, O. Lozach, L. Babault, T. Tahtouh, B.
Baratte, Y. Ogawa, M. Hagiwara, A. Eisenreich, U. Rauch, S.
Knapp, L. Meijer, J.P. Bazureau, Leucettines, a class of potent
inhibitors of cdc2-like kinases and dual specificity, tyrosine
phosphorylation regulated kinases derived from the marine sponge
leucettamine B: modulation of alternative pre-RNA splicing, J.
Med. Chem. 54 (2011) 4172–4186. doi:10.1021/jm200274d.
[36]
L.A. Dakin, M.H. Block, H. Chen, E. Code, J.E. Dowling, X. Feng,
A.D. Ferguson, I. Green, A.W. Hird, T. Howard, E.K. Keeton, M.L.
Lamb, P.D. Lyne, H. Pollard, J. Read, A.J. Wu, T. Zhang, X.
Zheng, Discovery of novel benzylidene-1,3-thiazolidine-2,4-diones
as potent and selective inhibitors of the PIM-1, PIM-2, and PIM-3
protein kinases, Bioorg. Med. Chem. Lett. 22 (2012) 4599–4604.
doi:10.1016/j.bmcl.2012.05.098.
G. Naert, V. Ferré, J. Meunier, E. Keller, S. Malmström, L.
Givalois, F. Carreaux, J.P. Bazureau, T. Maurice, Leucettine L41, a
DYRK1A-preferential DYRKs/CLKs inhibitor, prevents memory
impairments and neurotoxicity induced by oligomeric Aβ25-35
peptide administration in mice, Eur. Neuropsychopharmacol. 25
(2015) 2170–2182. doi:10.1016/j.euroneuro.2015.03.018.
[37]
[38]
[39]
H. de Koning, W.N. Speckamp, Succinimide, in: Encycl. Reagents
Org. Synth., John Wiley & Sons, Ltd, 2001.
doi:10.1002/047084289X.rs126.
G.G. Alfonso, J.L.G. Ariza, Derivatives of rhodanine as
spectrophotometric analytical reagents, Microchem. J. 26 (1981)
574–585. doi:10.1016/0026-265X(81)90144-2.
[24]
[25]
W. Becker, U. Soppa, F.J. Tejedor, DYRK1A: a potential drug
target for multiple Down syndrome neuropathologies, CNS Neurol.
Disord. Drug Targets. 13 (2014) 26–33.
T. Mendgen, C. Steuer, C.D. Klein, Privileged scaffolds or
promiscuous binders: a comparative study on rhodanines and
related heterocycles in medicinal chemistry, J. Med. Chem. 55
(2012) 743–753. doi:10.1021/jm201243p.
C. Kuhn, D. Frank, R. Will, C. Jaschinski, R. Frauen, H.A. Katus,
N. Frey, DYRK1A is a novel negative regulator of cardiomyocyte
hypertrophy, J. Biol. Chem. 284 (2009) 17320–17327.
doi:10.1074/jbc.M109.006759.
[40]
[41]
[42]
M.A. Pagano, L. Cesaro, F. Meggio, L.A. Pinna, Protein kinase
CK2: a newcomer in the “druggable kinome”, Biochem. Soc.
Trans. 34 (2006) 1303–1306. doi:10.1042/BST0341303.
[26]
[27]
[28]
Y. Ogawa, Y. Nonaka, T. Goto, E. Ohnishi, T. Hiramatsu, I. Kii,
M. Yoshida, T. Ikura, H. Onogi, H. Shibuya, T. Hosoya, N. Ito, M.
Hagiwara, Development of a novel selective inhibitor of the Down
syndrome-related kinase Dyrk1A, Nat. Commun. 1 (2010) 86.
doi:10.1038/ncomms1090.
D.I. Perez, C. Gil, A. Martinez, Protein kinases CK1 and CK2 as
new targets for neurodegenerative diseases, Med. Res. Rev. 31
(2011) 924–954. doi:10.1002/med.20207.
S. Côté, P.-H. Carmichael, R. Verreault, J. Lindsay, J. Lefebvre, D.
Laurin, Nonsteroidal anti-inflammatory drug use and the risk of
cognitive impairment and Alzheimer’s disease, Alzheimers.
Dement. 8 (2012) 219–226. doi:10.1016/j.jalz.2011.03.012.
O. Fedorov, K. Huber, A. Eisenreich, P. Filippakopoulos, O. King,
A.N. Bullock, D. Szklarczyk, L.J. Jensen, D. Fabbro, J. Trappe, U.
Rauch, F. Bracher, S. Knapp, Specific CLK inhibitors from a novel
chemotype for regulation of alternative splicing, Chem. Biol. 18
(2011) 67–76. doi:10.1016/j.chembiol.2010.11.009.
[43]
M. Camps, T. Rückle, H. Ji, V. Ardissone, F. Rintelen, J. Shaw, C.
Ferrandi, C. Chabert, C. Gillieron, B. Françon, T. Martin, D.
Gretener, D. Perrin, D. Leroy, P.A. Vitte, E. Hirsch, M.P. Wymann,
R. Cirillo, M.K. Schwarz, C. Rommel, Blockade of PI3Kgamma
suppresses joint inflammation and damage in mouse models of
rheumatoid arthritis, Nat. Med. 11 (2005) 936–943.
doi:10.1038/nm1284.
T. Tahtouh, J.M. Elkins, P. Filippakopoulos, M. Soundararajan, G.
Burgy, E. Durieu, C. Cochet, R.S. Schmid, D.C. Lo, F.
Delhommel, A.E. Oberholzer, L.H. Pearl, F. Carreaux, J.P.
Bazureau, S. Knapp, L. Meijer, Selectivity, cocrystal structures, and
neuroprotective properties of leucettines, a family of protein kinase
inhibitors derived from the marine sponge alkaloid leucettamine B,
J. Med. Chem. 55 (2012) 9312–9330. doi:10.1021/jm301034u.
[44]
[45]
X. Peng, X. Wu, L. Chen, Z. Wang, X. Hu, L. Song, C.M. He, Y.F.
Luo, Z.Z. Chen, K. Jin, H.G. Lin, X.L. Li, Y.S. Wang, Y.Q. Wei,
Inhibition of phosphoinositide 3-kinase ameliorates dextran sodium
sulfate-induced colitis in mice, J. Pharmacol. Exp. Ther. 332 (2010)
46–56. doi:10.1124/jpet.109.153494.
[29]
[30]
[31]
C. Schmitt, D. Kail, M. Mariano, M. Empting, N. Weber, T. Paul,
R.W. Hartmann, M. Engel, Design and synthesis of a library of
lead-like 2,4-bisheterocyclic substituted thiophenes as selective
Dyrk/Clk inhibitors, PLoS One. 9 (2014) e87851.
doi:10.1371/journal.pone.0087851.
Z. Wang, X. Wu, L. Song, Y. Wang, X. Hu, Y.F. Luo, Z.Z. Chen, J.
Ke, X.D. Peng, C.M. He, W. Zhang, L.J. Chen, Y.Q. Wei,
Phosphoinositide 3-kinase gamma inhibitor ameliorates
concanavalin A-induced hepatic injury in mice, Biochem. Biophys.
Res. Commun. 386 (2009) 569–574.
H. Falke, A. Chaikuad, A. Becker, N. Loaëc, O. Lozach, S. Abu
Jhaisha, W. Becker, P.G. Jones, L. Preu, K. Baumann, S. Knapp, L.
Meijer, C. Kunick, 10-iodo-11H-indolo[3,2-c]quinoline-6-
carboxylic acids are selective inhibitors of DYRK1A, J. Med.
Chem. 58 (2015) 3131–3143. doi:10.1021/jm501994d.
doi:10.1016/j.bbrc.2009.06.060.
[46]
X. Wei, J. Han, Z. Chen, B. Qi, G. Wang, Y.H. Ma, H. Zheng, Y.F.
Luo, Y.Q. Wei, L.J. Chen, A phosphoinositide 3-kinase-gamma
inhibitor, AS605240 prevents bleomycin-induced pulmonary
fibrosis in rats, Biochem. Biophys. Res. Commun. 397 (2010) 311–
317. doi:10.1016/j.bbrc.2010.05.109.
T.T. Wager, R.Y. Chandrasekaran, X. Hou, M.D. Troutman, P.R.
Verhoest, A. Villalobos,Y. Will, Defining desirable central nervous
system drug space through the alignment of molecular properties, in
vitro ADME, and safety attributes, ACS Chem. Neurosci. 1 (2010)
420–434. doi:10.1021/cn100007x.
[47]
[48]
V. Spitzenberg, C. König, S. Ulm, R. Marone, L. Röpke, J.P.
Müller, M. Grün, R. Bauer, I. Rubio, M.P. Wymann, A. Voigt, R.
Wetzker, Targeting PI3K in neuroblastoma, J. Cancer Res. Clin.
Oncol. 136 (2010) 1881–1890. doi:10.1007/s00432-010-0847-2.
[32]
[33]
S. Udenfriend, B. Witkop, B.G. Redfield, H. Weissbach, Studies
with reversible inhibitors of monoamine oxidase: Harmaline and
related compounds, Biochem. Pharmacol. 1 (1958) 160–165.
doi:10.1016/0006-2952(58)90025-X.
M. Mariano, C. Schmitt, P. Miralinaghi, M. Catto, R.W. Hartmann,
A. Carotti, M. Engel, First Selective Dual Inhibitors of Tau
Phosphorylation and Beta-Amyloid Aggregation, Two Major
Pathogenic Mechanisms in Alzheimer’s Disease, ACS Chem.
Neurosci. 5 (2014) 1198–1202. doi:10.1021/cn5001815.
G.D. Cuny, N.P. Ulyanova, D. Patnaik, J.-F. Liu, X. Lin, K.
Auerbach, S.S. Ray, J. Xian, M.A. Glicksman, R.L. Stein, J.M.
Higgins, Structure-activity relationship study of beta-carboline
derivatives as haspin kinase inhibitors, Bioorg. Med. Chem. Lett.
22 (2012) 2015–2019. doi:10.1016/j.bmcl.2012.01.028.
[49]
[50]
E. Arsovska, J. Trontelj, N. Zidar, T. Tomašić, L.P. Mašič, D.
Kikelj, J. Plavec, A. Zega, Evaluation of Michael-type acceptor
reactivity of 5-benzylidenebarbiturates, 5-benzylidenerhodanines,
and related heterocycles using NMR, Acta Chim. Slov. 61 (2014)
637–644.
[34]
[35]
R. Prudent, V. Moucadel, M. López-Ramos, S. Aci, B. Laudet, L.
Mouawad, C. Barette, J. Einhorn, C. Einhorn, J.N. Denis, G.
Bisson, F. Schmidt, S. Roy, L. Lafanechere, J.C. Florent, C.
Cochet, Expanding the chemical diversity of CK2 inhibitors, Mol.
Cell. Biochem. 316 (2008) 71–85. doi:10.1007/s11010-008-9828-z.
G. Caillé, L.M. Dubé, Y. Théorêt, F. Varin, N. Mousseau, I.J.
McGilveray, Stability study of diltiazem and two of its metabolites
A.G. Golub, V.G. Bdzhola, N. V Briukhovetska, A.O. Balanda,

using a high performance liquid chromatographic method,
Biopharm. Drug Dispos. 10 (1989) 107–114. ACCEPTED MANUSCRIPT
[51]
[52]
O.D. Ratnoff, I.H. Lepow, Some properties of an esterase derived
from preparations of the first component of complement, J. Exp.
Med. 106 (1957) 327–343.
Acknowledgments
The authors thank Mrs. Nadja Weber and Dr. Christian
M.A. Khanfar, R.A. Hill, A. Kaddoumi, K.A. El Sayed, Discovery
of novel GSK-3β inhibitors with potent in vitro and in vivo
activities and excellent brain permeability using combined ligand-
and structure-based virtual screening, J. Med. Chem. 53 (2010)
8534–8545. doi:10.1021/jm100941j.
Schmitt for performing some kinase inhibition assays. The
financial support by the Deutsche Forschungsgemeinschaft
(DFG) to M.E. (grant EN 381/2-3) is also greatly
acknowledged.
[53]
[54]
G.F. Passos, C.P. Figueiredo, R.D.S. Prediger, K.A.B.S. Silva, J.M.
Siqueira, F.S. Duarte, P.C. Leal, R. Medeiros, J.B. Calixto,
Involvement of phosphoinositide 3-kinase gamma in the neuro-
inflammatory response and cognitive impairments induced by beta-
amyloid 1-40 peptide in mice, Brain. Behav. Immun. 24 (2010)
493–501. doi:10.1016/j.bbi.2009.12.003.
Supplementary Material
Supplementary data associated with this article can be found
in the online version.
These data include general synthetic methods, further
compound syntheses and analytical data, Tables S1–S3 and the
MTT assay method.
D. Frost, B. Meechoovet, T. Wang, S. Gately, M. Giorgetti, I.
Shcherbakova, T. Dunckley, β-carboline compounds, including
harmine, inhibit DYRK1A and tau phosphorylation at multiple
Alzheimer’s disease-related sites, PLoS One. 6 (2011) e19264.
doi:10.1371/journal.pone.0019264.

List of captions:
ACCEPTED MANUSCRIPT
Figure 1. 2D structures of harmine, INDY and Leucettamine B.
Figure 2. Superimposition of harmine (green), INDY (cyan) and Leucettine L41 (magenta) cocrystallized with Dyrk1A (PDB codes
3ANR, 3ANQ and 4AZE, respectively). All ligands are bound in the ATP-binding cleft via two H-bonds, formed with Leu241-NH
at the hinge region and Lys188 (indicated by the grey dashed lines). The measured ranges of the N–O distances are given in Å.
Additional contacts comprise van der Waals-interactions with the side chains of Ala186, Val173 and Leu294, which mainly involve
the left aromatic rings.
Figure 3. Inhibition of Dyrk1A–catalyzed tau-Thr212 phosphorylation in stably transfected HEK293 cells. The test compounds,
DMSO or the reference inhibitor harmine [54] were added to the cell medium and the Dyrk1A expression induced as described.
After immunoblotting of the total cell proteins, the level of phospho-tau-Thr212 was detected using a phosphospecific antibody. To
normalize the signals, total recombinant GFP-tau protein was quantified on the same blot using an anti-GFP antibody (fluorescence
image converted to grey scales). Initial screenings were performed at 1 µM compound concentrations (lower panel; one
a
representative experiment out of two is shown). IC50 values were determined after cell treatment with different concentrations in
triplicates, referring to the signal derived from DMSO–treated cells as 100%. Values represent averages from at least two
independent determinations (±S.D.); AS, AS605240.
Table 1. Inhibitory activity of α-benzylidene-γ-butyrolactone derivatives against Dyrk1A and selectivity profiling against a panel of
selected kinases.
aS.D.≤18%, [ATP]=100µM; D1A=Dyrk1A, D1B=Dyrk1B, D2=Dyrk2; Hrm=Harmine; n.i., no inhibition; n.d., not determined.
Table 2. Inhibitory activity of benzylidene thioxothiazolidinone–based derivatives against Dyrk1A and selectivity
profiling against a panel of selected kinases.
aS.D.≤18%, [ATP]=100µM; D1A=Dyrk1A, D1B=Dyrk1B, D2=Dyrk2, AS= AS605240; n.d., not determined; n.i., no inhibition.
Table 3. Extended selectivity profiling of 16 and 17.
a% inhibition at 1 µM; shown are mean values of two different measurements performed in duplicates (S.D. ≤ 2.5%); bIC50s were only determined when the
% inhibition at 1 µM was greater than 85%. Data were calculated based on mean values of duplicates that differed by less than 7%.
Table 4. Metabolic stability in rat plasma.
amean values from two experiments, S.D.< 17%; bno degradation detected after 120 min

ACCEPTED MANUSCRIPT
Systematic diversification of benzylidene heterocycles yields
novel inhibitor scaffolds selective for Dyrk1A, Clk1 and CK2
Marica Marianoa , Rolf W. Hartmanna,b and Matthias Engela,
a
Department of Pharmaceutical and Medicinal Chemistry, Saarland University, Saarbrücken, Germany.
b Helmholtz Institute for Pharmaceutical Research Saarland (HIPS), Campus C2-3, D 66123 Saarbrücken, Germany
Highlights
•
•
•
•
A new α-benzylidene–γ-butyrolactone scaffold was designed and systematically
diversified
Different scaffolds exhibiting selectivity for Dyrk1A, Clk1 and CK2 were identified,
with IC50s from the low micromolar to the nanomolar range
Multi-targeted kinase inhibitors were also obtained, that co-inhibited the lipid kinases
PI3Kα/γ
Compounds possessing different scaffolds inhibited tau phosphorylation in intact cells
Corresponding author. E-mail: ma.engel@mx.uni-saarland.de; phone: +4968130270312.
1