Targeting the kinases MNK1 and MNK2 has emerged as a valuable strategy in oncology. However, most of the advanced inhibitors are acting in an adenosine triphosphate (ATP)-competitive mode, precluding the evaluation of different binding modes in preclinical settings. Using rational design, we identified and validated the 4,6-diaryl-pyrazolo[3,4-b]pyridin-3-amine scaffold as the core for MNK inhibitors. Signaling pathway analysis confirmed a direct effect of the hit compound EB1 on MNKs, and in line with the reported function of these kinases, EB1 only affects the growth of tumor but not normal cells. Molecular modeling revealed the binding of EB1 to the inactive conformation of MNK1 and the interaction with the specific DFD motif. This novel mode of action appears to be superior to the ATP-competitive inhibitors, which render the protein in a pseudo-active state. Overcoming this paradoxical activation of MNKs by EB1 represents therefore a promising starting point for the development of a novel generation of MNK inhibitors.
Targeting the kinases MNK1 and MNK2 has emerged as a valuable strategy in oncology. However, most of the advanced inhibitors are acting in an adenosine triphosphate (ATP)-competitive mode, precluding the evaluation of different binding modes in preclinical settings. Using rational design, we identified and validated the 4,6-diaryl-pyrazolo[3,4-b]pyridin-3-amine scaffold as the core for MNK inhibitors. Signaling pathway analysis confirmed a direct effect of the hit compound EB1 on MNKs, and in line with the reported function of these kinases, EB1 only affects the growth of tumor but not normal cells. Molecular modeling revealed the binding of EB1 to the inactive conformation of MNK1 and the interaction with the specific DFD motif. This novel mode of action appears to be superior to the ATP-competitive inhibitors, which render the protein in a pseudo-active state. Overcoming this paradoxical activation of MNKs by EB1 represents therefore a promising starting point for the development of a novel generation of MNK inhibitors.
Altered regulation
of mRNA translation is a frequent feature of
cancer. Cell signaling pathways regulating mRNA translation are in
many cancers not only affected by genetic alterations but also by
the tumor environment, and aberrant mRNA translation is therefore
a widespread characteristic of tumor cells.[1,2] Consequently,
therapeutic targeting of the protein synthesis apparatus is emerging
as a valuable strategy in oncology.A key player in the regulation
of cap-dependent mRNA translation
is the eukaryotic translation initiation complex 4F (eIF4F), and eIF4E
is the rate-limiting component of this complex.[3] Many studies over the last years have revealed that deregulation
of eIF4E is implicated in tumorigenesis and cancer progression.[4,5] In line with this, we and others have described eIF4E as an independent
prognostic factor in several tumors.[1]The MAP kinase interacting kinases (MNK1/2) execute a unique function
in regulating eIF4E through phosphorylation at the conserved residue
Ser 209. This phosphorylation of eIF4E is necessary for oncogenic
transformation but appears to be dispensable for normal development.[6] Furthermore, MNKs phosphorylation of eIF4E has
been shown to be involved in tumor metastasis.[7−9] The increased
requirement for MNK-dependent phosphorylation of eIF4E with malignancy
is thought to selectively upregulate the translation of certain mRNAs
involved in tumor cell survival and dissemination.[4] With the implementation of cancer immunotherapy into the
clinics, the previously described immune-regulatory functions of MNKs
have also regained attention in the field of oncology. Those functions
may either be achieved through the regulation of cytokine signaling
and response[10,11] or even directly through the
regulation of immune checkpoint proteins like PD-L1 in tumor cells.
Importantly, MNK inhibition in this context has been shown to have
a major impact on tumor development and metastasis.[12−14]In the
tumor environment, cancer cells are exposed to various stresses,
including hypoxia and immune responses as well as treatment-induced
alterations in hormone levels or DNA damage. In response to those
alterations, activation of MNKs can be achieved through the MAPK stress
response pathways (ERK and p38α/β) and promote cellular
survival.[1,15] In this context, experiments with cell lines
from breast and pancreatic cancer as well as with glioblastoma revealed
that phosphorylation of eIF4E is required for survival from genotoxic
stress, caused by DNA damaging agents.[16−18] Furthermore, resistance
to hormone deprivation therapies in breast and prostate cancers has
been linked to elevated levels of p-eIF4E and MNK inhibition has been
shown to increase the sensitivity to those treatments.[19,20]MNKs belong to the family of Ser/Thr kinases, are classified
as
members of Ca2+/calmodulin-dependent kinases, and are present
in two isoforms: MNK1 and MNK2.[21,22] The catalytic domains
of MNKs, while being overall conserved to other protein kinases, contain
three unusual elements: two short sequence insertions and a DFD instead
of the classical DFG-motif.[23,24] These elements provide
the base for an MNK specific regulatory mechanism, in which the kinase
activity is autoinhibited by conversion of the activation segment
into an autoinhibitory module.[24] Although
this regulatory mechanism accounts for both kinases, the autoinhibition
is more stringent in the case of MNK1, while MNK2 possesses higher
constitutive basal activity. While both kinases can be activated by
ERK, only MNK1 contains a shared docking motif for p38α/β
and ERK, which allows its activation by both kinases.[25]Although several investigational compounds have been
used over
the last years for the study of MNKs function, agents such as cercosporamide
(1)[26] and CGP57380 (2)[27] (Figure A) also inhibit several other protein kinases
with similar or greater potency compared to MNKs.[28,29] Effort has been therefore devoted to the identification of novel
selective MNK inhibitors. In recent years, compounds like eFT508 (3),[30] BAY1143269,[31] SEL-201,[7] ETC-206,[32] MNK-I1,[9] and the
MNK-I1-derived MNK2 selective inhibitor MNK-7g[33] have been developed and tested in preclinical settings.
Among those, currently, eFT508, BAY1143269, and ETC-206 are tested
in clinical trials in oncology.[13,34] Notably, all of those
inhibitors, even though having a different chemical scaffold, are
categorized as Type I inhibitors, acting in an adenosine triphosphate
(ATP)-competitive manner on the active conformation of the kinase.[29] Studies on inhibitors of other protein kinases
have however revealed that some Type I kinase inhibitors are prone
to activate noncatalytic functions of the target protein.[35−37] For the clinical translation of the mainly genetic-based proposal
of targeting MNKs as a promising new strategy in oncology, it is therefore
desirable to identify non-ATP-competitive inhibitors to evaluate different
modes of kinase inhibition in preclinical and clinical settings.
Figure 1
Study
of the pyrazolo[3,4-b]pyridin-3-amine scaffold
as an MNK inhibitor. (A) Structure of the pyrazolo[3,4-b]pyridin-3-amine scaffold (4) and the known inhibitors
used as references (1, 2, 3). (B) Inhibitory activity of pyrazolo[3,4-b]pyridine
derivatives. Results from the enzymatic radiometric assay. Values
indicated as residual activity of kinases MNK1 and MNK2 after treatment
with the indicated compounds at 10 μM. (C) Western blot analysis
of MDA-MB-231 cells treated with pyrazolo[3,4-b]pyridine
derivatives for 24 h. Dimethyl sulfoxide (DMSO) and CGP (CGP57380)
are used as negative and positive controls, respectively.
Study
of the pyrazolo[3,4-b]pyridin-3-amine scaffold
as an MNK inhibitor. (A) Structure of the pyrazolo[3,4-b]pyridin-3-amine scaffold (4) and the known inhibitors
used as references (1, 2, 3). (B) Inhibitory activity of pyrazolo[3,4-b]pyridine
derivatives. Results from the enzymatic radiometric assay. Values
indicated as residual activity of kinases MNK1 and MNK2 after treatment
with the indicated compounds at 10 μM. (C) Western blot analysis
of MDA-MB-231 cells treated with pyrazolo[3,4-b]pyridine
derivatives for 24 h. Dimethyl sulfoxide (DMSO) and CGP (CGP57380)
are used as negative and positive controls, respectively.
Results and Discussion
Identification of New Scaffolds as MNK Inhibitors
In
recent years, the development of new MNK inhibitors has been centered
on small synthetic structures. Among those, bicyclic nitrogenated
compounds appear to be a privileged scaffold for the inhibition of
these kinases. Here, we propose the use of the pyrazolo[3,4-b]pyridin-3-amine core (4) (Figure A) as a new alternative scaffold
as an MNK inhibitor. Detailed analysis of CGP57380 (2) indicated that the pyridine- and pyridone-based scaffolds could
be an alternative to the widely reported pyrimidine-based structures
as the 5-N is not involved in the direct interaction of the molecule
with the ATP binding site.[38] In addition,
work by Yu et al. suggested that the 3-amino group of this part of
the molecule would form an additional interaction with Phe227 in the
ATP binding pocket of MNK2.[39]To
study the suitability of new scaffolds, several families of compounds
were prepared and the inhibition of MNK1 and MNK2 kinase activity
was tested with a radiometric protein kinase assay (33PanQinase
Activity Assay, Reaction Biology).The first stage of the study
(see the Supporting Information) was focused on exploring the suitability of 3-amino-pyrazolo[3,4-b]pyridin-6-one scaffold (S11, Figure S1A). These structures were obtained by
the cyclization of 2-methoxy-6-oxo-1,4,5,6-tetrahydropyridin-3-carbonitriles
(S13), synthesized by the reaction of an α,β-unsaturated
ester (S12) and malononitrile in NaOMe/MeOH, with hydrazine
or substituted hydrazines (Figure S1B).[40] We synthesized 37 compounds with a broad combination
of substituents in the most interesting points of this scaffold (R2,
R4, and R5). The compounds under study combined bulky and small substituents
in the pyridone ring (H, Me, Ph, and substituted aromatic rings) and
in the pyrazole ring (H, Me, Ph, substituted Ph and acyl groups) to
explore the maximum number of possibilities (compounds S14a–g, S15a–x, and S16a–f). However,
none of the newly developed molecules was able to inhibit MNK1/2 kinase
activity in vitro (Figure S2).Next, we included a set of modifications of the original
scaffold
by oxidizing and methylating the lactam ring (compounds S17a–c, S18a–d, and S19) (Figure S1C), substituting the pyrazole ring in
the N1 position[41] and testing the importance
of the NH2 group[42] (compounds S20 to S25). The set of 14 new compounds prepared
did not show any significant activity in the kinase assay in vitro (Figure S3). Therefore,
the pyrazolo[3,4-b]pyridin-6-one scaffold did not
appear suitable for the development of new MNK kinase inhibitors.While the lactam ring is a central feature of several kinase inhibitors
and was proposed to have a positive effect on the potency and selectivity
of MNK inhibitors,[43−46] none of our synthesized compounds was able to inhibit MNKs kinase
activity (Figures S2 and S3). Therefore,
we focused next on molecules containing a pyridine ring instead of
the pyridone ring to increase aromaticity and to obtain a more planar
structure (4, Figure A). A new set of compounds 5a–b and 6a–d with a pyrazolo[3,4-b]pyridin-3-amine core was synthesized. Modifications included: (a)
aromatization of the lactam ring by capturing the enolic form of the
lactam with a halogenating agent forming a 6-chloropyridine ring and
(b) inclusion of substituents of different bulkiness (H, Me, Ph) at
positions R1/R2, R4, and R6. Among those compounds, 6b was able to inhibit the MNK1 and MNK2 kinase activity in
vitro (Figure B).Having identified a biochemically active compound, we next
tested
its activity in a cell-based assay. Since phosphorylation of eIF4E
at S209 is uniquely carried out by MNK kinases, the assessment of
p-eIF4E by western blot analysis directly reflects MNK1/2 activity.
Treatment of MDA-MB-231 cells with compounds bearing a pyrazolo[3,4-b]pyridin-3-amine core revealed that only 6b resulted in a clear inhibition of eIF4E phosphorylation (Figure C). The structure–activity
relationship (SAR) analysis of these results revealed two requirements
for the activity of the first hit compound: (1) the presence of the
two phenyl rings at the C4 and C6 positions of the pyrazolo[3,4-b]pyridine scaffold and (2) the absence of substituents
(methyl) at the pyrazole ring.In summary, out of the nearly
60 synthesized and tested compounds,
we identified 6b as a first hit compound, which was able
to inhibit the kinase activity of MNK1/2 in vitro and eIF4E phosphorylation in cells.
4,6-Diaryl-pyrazolo[3,4-b]pyridin-3-amine scaffold
as a Core for MNK Inhibitors
To confirm the activity of the
core structure and to test the effect of substituents on the biological
activity, we generated derivatives of 6b. A family of
4,6-diphenyl-1H-pyrazolo[3,4-b]pyridin-3-amines
was designed with different substituents in both phenyl rings and
in the pyrazole ring, and a total of 11 derivatives of 6b (6e–o) were synthesized. The synthesis starts
by the condensation of the corresponding substituted acetophenone
(7) with the appropriate substituted benzaldehyde (8) in NaOMe/MeOH at room temperature followed by the Michael
addition with malononitrile under microwave irradiation to afford
upon cyclization of the Michael intermediate the corresponding cyanomethoxypyridine
(9). The conversion of the methoxy group of 9 into a better leaving group is achieved upon treatment with POBr3 in 1,4-dioxane in the presence of Pyridine·HBr to yield
the bromo-substituted pyridine (10). Finally, the cyclization
with the appropriate hydrazine in MeOH affords the corresponding pyrazole[3,4-b]pyridines 6e–o (Figure A). Test for their inhibitory effect on MNK1
and MNK2 kinase assay revealed that several derivatives were able
to inhibit the kinase activity of MNKs. Remarkably, the more active
compounds in the in vitro kinase assays were also
able to inhibit p-eIF4E phosphorylation in MDA-MB-231 cells, highlighting
the biological activity of this new compound family of MNK inhibitors
(Figure B,C). Therefore,
the 4,6-diaryl-1H-pyrazolo[3,4-b]pyridin-3-amine represents a new core structure for MNK1/2 inhibitors
with biological activity.
Figure 2
Study of the 4,6-diaryl pyrazolo[3,4-b]pyridin-3-amines
as MNK inhibitors. (A) General synthesis of 4,6-diaryl pyrazolo[3,4-b]pyridin-3-amine derivatives. (B) Inhibitory activity of
4,6-diaryl pyrazolo[3,4-b]pyridin-3-amine derivatives.
Results from the enzymatic radiometric assay. Values indicated as
residual activity of kinases MNK1 and MNK2 after treatment with the
indicated compounds at 10 μM. Cercosporamide is used as a positive
control. (C) Western blot analysis of MDA-MB-231 cells treated with
the indicated concentrations of the 4,6-diaryl pyrazolo[3,4-b]pyridin-3-amine derivatives for 24 h. DMSO and CGP (CGP57380)
are used as negative and positive controls, respectively.
Study of the 4,6-diaryl pyrazolo[3,4-b]pyridin-3-amines
as MNK inhibitors. (A) General synthesis of 4,6-diaryl pyrazolo[3,4-b]pyridin-3-amine derivatives. (B) Inhibitory activity of
4,6-diaryl pyrazolo[3,4-b]pyridin-3-amine derivatives.
Results from the enzymatic radiometric assay. Values indicated as
residual activity of kinases MNK1 and MNK2 after treatment with the
indicated compounds at 10 μM. Cercosporamide is used as a positive
control. (C) Western blot analysis of MDA-MB-231 cells treated with
the indicated concentrations of the 4,6-diaryl pyrazolo[3,4-b]pyridin-3-amine derivatives for 24 h. DMSO and CGP (CGP57380)
are used as negative and positive controls, respectively.SAR analysis revealed that the aromatic and planar structure
of
the pyrazolo[3,4-b]pyridine system is essential for
biological activity. This is in particular exemplified by the initial
compounds presenting a pyrazolo[3,4-b]pyridin-6-one
scaffold (Figures S2 and S3, compounds S14a–g, S15a–x, S16a–f, S17a–c, S18a–d, and S19), which did not present any MNK1/2 inhibitory activity
despite containing equivalent substituents. In line with the results
obtained above (Figure B,C), the lack of activity of compound 6c indicates
that the C4 and C6 phenyl rings are essential substituents of the
pyrazolo[3,4-b]pyridine-3-amine core structure, to
achieve inhibitory activity on MNK1/2. Moreover, the substitution
of the pyrazole ring scaffold appears to render the compounds inactive
as shown in Figures B,C and 2B,C. The drop in the activity of
the methyl-substituted compounds 6a, 6m,
and 6n in comparison with the corresponding unsubstituted
compounds 6b, 6e, and 6k supports
this. Therefore, the presence of an unsubstituted pyrazole moiety
is a structural requirement of 4,6-diphenyl-1H-pyrazolo[3,4-b]pyridin-3-amines as inhibitors of MNK1 and MNK2, indicating
that the presence of a H-bond donor in such a ring is essential for
the activity of these scaffolds. Supporting this further, all compounds
presenting the unsubstituted pyrazole ring showed inhibition of at
least one of the MNK kinases.In summary, we have identified
and validated 4,6-diaryl-1H-pyrazolo[3,4-b]pyridin-3-amines as a
novel core structure for MNK1/2 inhibitors and 6b, the
more potent compound of this family, was chosen for further detailed
analysis.
Characterization of the EB Family of Compounds
We finally
selected the compounds 6b, 6g, and 6o, from now on named EB1, EB2,
and EB3, respectively, for quantitative analysis and
determined their IC50 for MNK1 and MNK2 in protein kinase
assays. As summarized in Figure A, all three compounds inhibited MNK1 and MNK2 activity
and displayed a 4- to 14-fold lower IC50 for MNK1. The
first hit compound EB1 was maintained as the more active
and selective MNK1 inhibitor with an IC50 of 0.69 μM
for MNK1 compared to an IC50 of 9.4 μM for MNK2.
Moreover, treatment of MDA-MB-231 cells revealed a dose-dependent
inhibition of MNK activity as judged by the reduction of eIF4E phosphorylation
(Figure B).
Figure 3
4,6-Diaryl
pyrazolo[3,4-b]pyridin-3-amines compounds
as MNK inhibitors. (A) IC50 values obtained for the selected
4,6-diaryl pyrazolo[3,4-b]pyridin-3-amine derivatives.
CGP (CGP57380) is used as a positive control. (B) Titration curves
of the selected 4,6-diaryl pyrazolo[3,4-b]pyridin-3-amine
derivatives in MDA-MB-231 cells. DMSO and CGP (CGP57380) are used
as negative and positive controls, respectively. Cells were treated
for 24 h with the inhibitors, and western bot analysis with the indicated
antibodies was carried out.
4,6-Diaryl
pyrazolo[3,4-b]pyridin-3-amines compounds
as MNK inhibitors. (A) IC50 values obtained for the selected
4,6-diaryl pyrazolo[3,4-b]pyridin-3-amine derivatives.
CGP (CGP57380) is used as a positive control. (B) Titration curves
of the selected 4,6-diaryl pyrazolo[3,4-b]pyridin-3-amine
derivatives in MDA-MB-231 cells. DMSO and CGP (CGP57380) are used
as negative and positive controls, respectively. Cells were treated
for 24 h with the inhibitors, and western bot analysis with the indicated
antibodies was carried out.In summary, the 4,6-diphenyl-1H-pyrazolo[3,4-b]pyridin-3-amine-based compounds EB1, EB2, and EB3 represent a new class of biologically
active MNK1/2 inhibitors with a higher affinity for MNK1 compared
to MNK2.
Characterization of EB1
Focusing on EB1, we next carried out a detailed characterization of its
biological activity. To exclude cell-line-specific effects of EB1 activity, we assessed the inhibition of eIF4E phosphorylation
in other cell lines of solid tumors like MDA-MB-468 and MCF7 (breast
cancer), 22RV1 (prostate cancer), A375M (melanoma), as well as in
the leukemia cell line (MV4-11) (Figures and S4). In line
with the results described above for MDA-MB-231 (Figure B), treatment of cells for
24 h with EB1 also resulted in a dose-dependent inhibition
of eIF4E phosphorylation in all other cell lines tested, which therefore
excludes cell-line-specific effects of EB1.
Figure 4
Characterization
of EB1. (A) Analysis of the inhibition
of eIF4E phosphorylation in MDA-MB-231 cells treated with 10 μM EB1 for the indicated times. (B) Titration curve of EB1 in the cell lines MDA-MB-231, MDA-MB-468, MCF7, and IMR90.
Treatment was carried out for 72 h, and inhibition of eIF4E phosphorylation
was analyzed by western blotting with the indicated antibodies. (C)
Effect of EB1 on cell growth in the same cell lines as
in (B). Cells were treated for 72 h with the indicated concentrations
of EB1 and cell growth was analyzed by crystal violet
staining. The GI50 values are 15.08 (+/- 1.2) μM
(MDA-MB-231), 14.46 (+/- 1.2) μM (MDA-MB-468), 8.46 (+/- 1.1)
μM (MCF7), and >40 μM for IMR90. (D) Cytotoxicity of EB1 on different cell lines. Cells were treated as described
in (C), Hoechst and Propidium iodide (PI) was added to the cell culture
medium, and the fraction of PI-positive cells was determined microscopically. Cis-platin and DMSO are used as positive and negative controls,
respectively. Error bars represent the standard deviation of three
independent experiments.
Characterization
of EB1. (A) Analysis of the inhibition
of eIF4E phosphorylation in MDA-MB-231 cells treated with 10 μM EB1 for the indicated times. (B) Titration curve of EB1 in the cell lines MDA-MB-231, MDA-MB-468, MCF7, and IMR90.
Treatment was carried out for 72 h, and inhibition of eIF4E phosphorylation
was analyzed by western blotting with the indicated antibodies. (C)
Effect of EB1 on cell growth in the same cell lines as
in (B). Cells were treated for 72 h with the indicated concentrations
of EB1 and cell growth was analyzed by crystal violet
staining. The GI50 values are 15.08 (+/- 1.2) μM
(MDA-MB-231), 14.46 (+/- 1.2) μM (MDA-MB-468), 8.46 (+/- 1.1)
μM (MCF7), and >40 μM for IMR90. (D) Cytotoxicity of EB1 on different cell lines. Cells were treated as described
in (C), Hoechst and Propidium iodide (PI) was added to the cell culture
medium, and the fraction of PI-positive cells was determined microscopically. Cis-platin and DMSO are used as positive and negative controls,
respectively. Error bars represent the standard deviation of three
independent experiments.Next, we studied the
dynamics of EB1 inhibition in
MDA-MB-231 cells. Time course experiments revealed a fast inhibition
of MNK activity (<30 min) (Figure A). Moreover, this inhibition was maintained in long-term
treatments (>72 h), arguing for the stability of the compound under
physiological conditions (Figure B).To study the phenotypic consequences of MNK
inhibition by EB1, we focused on cell growth. Previous
studies in MEFs obtained
from the MNK1/2 knock-out mice have revealed that the function of
those kinases is not required for normal cell growth.[47] In contrast, various studies have reported the requirement
for phosphorylation of eIF4E by MNKs for tumor cell growth in vitro and in vivo.[48−51] To test for drug selectivity,
we therefore used fibroblast (IMR90) and breast cancer cell lines
(MDA-MB-231, MDA-MB-468, and MCF7) as models for normal and transformed
cells, respectively, and analyzed cell growth after the treatment
with EB1. As shown in Figure C, cell growth was not affected in IMR90
cells up to 20 μM and only a slight but not significant effect
could be observed at higher concentrations. In contrast, a dose-dependent
inhibition of cell growth could be observed in all tested breast cancer
cell lines (Figure C). Propidium iodide (PI) staining carried out in parallel did not
reveal a significant increase in the number of dead cells in all tested
cell lines compared to the cytotoxic agent cis-platin (Figure D). FACS analysis in MDA-MB-231
cells (Figure S5) could finally confirm
these results, as in contrast to the positive control doxorubicin,
an increase in the proportion of “sub G1” cells was
not detected in EB1-treated cells.In summary,
analysis of cell growth and cell death did not reveal
adverse effects of EB1 in the noncancerous cell line
IMR90 but caused a reduction of cell growth in breast cancer cell
lines. This is in line with the requirement on MNKs for the growth
of cancer cells but not of normal cells. This also supports the selectivity
of the compound in a cellular context, as an unperturbed cell growth
in the presence of EB1 (up to 20 μM) precludes
significant targeting of any cellular protein required for this.
EB1 Represents a Novel Type of MNK Inhibitor
To determine the mode of action of EB1, we employed
molecular docking in combination with molecular dynamics (MD) to study
the binding mode of EB1 to the active and inactive conformations
of MNK1. Based on the available crystal structures of MNK1 (PDB ID: 2HW6) and MNK2 (PDB ID: 2HW7), we generated a
model of MNK1 in its inactive conformation (DFD-out) and a model of
MNK1 in the active conformation (DFD-in) including the phosphorylation
at T209/214 (detailed in the Experimental Section). To identify the most probable binding mode of EB1 to MNK1, we first identified the possible binding sites by molecular
docking and then determined the stability of the ligand–protein
complexes by MD simulation.In the active form of MNK1, the EB1-MNK1 complex was broken during the simulation, indicating
that the interaction found in the docking study was not stable over
time. The ligand did not remain in the ATP binding site and the EB1 molecule left the initial position before completion of
the simulation (<40 ns). In contrast, the interactions formed in
the inactive model of MNK1 were stable over time (Figure A). Based on these results,
we conclude that the most probable binding of EB1 occurs
to an allosteric binding site adjacent to the ATP binding pocket,
presenting the inactive conformation of MNK1. Based on the classification
of protein kinase inhibitors,[29,52] the compound would
therefore represent a Type II MNK1 inhibitor.
Figure 5
Computational study of
the binding mode of EB1 to
MNK1. (A) Model of EB1 (green) bound to the inactive
conformation of MNK1. (B) Interactions formed between the 3-aminopyrazole
ring of EB1 (green) and residues Glu125 (blue), Gln128
(yellow), and Leu55 (orange). (C) Surface representation of the pocket.
The phenyl ring in position C4 of EB1 (green) fits the
inner hydrophobic part of the pocket. (D) π–π stacking
between the Phenyl ring in position C6 of EB1 (green)
and Phe 192 (blue). The interaction is maintained during the whole
simulation. Image taken after 36 ns of simulation. (E) Summary of
the interactions of EB1 with the inactive conformation
of MNK1.
Computational study of
the binding mode of EB1 to
MNK1. (A) Model of EB1 (green) bound to the inactive
conformation of MNK1. (B) Interactions formed between the 3-aminopyrazole
ring of EB1 (green) and residues Glu125 (blue), Gln128
(yellow), and Leu55 (orange). (C) Surface representation of the pocket.
The phenyl ring in position C4 of EB1 (green) fits the
inner hydrophobic part of the pocket. (D) π–π stacking
between the Phenyl ring in position C6 of EB1 (green)
and Phe 192 (blue). The interaction is maintained during the whole
simulation. Image taken after 36 ns of simulation. (E) Summary of
the interactions of EB1 with the inactive conformation
of MNK1.Finally, we performed interaction
studies based on the molecular
model of EB1 in complex with MNK1. Hydrogen bonds are
formed between the 3-aminopyrazole moiety and MNK1 (Glu125 and Gln128),
confirming the importance of the presence of the unsubstituted 3-aminopyrazole
ring (Figure B). Besides
these hydrogen bonds, a hydrophobic interaction with Leu55 of MNK1
may in addition stabilize this interaction. Moreover, the phenyl ring
found in position C4 of the pyrazolo[3,4-b]pyridine
scaffold occupies the hydrophobic cleft in the most internal part
of the ATP binding pocket increasing the affinity and stability of
the complex (Figure C).In addition to these stabilizing interactions, EB1 interacts with the phenylalanine residue (Phe192) of the DFD motif.
This π–π interaction is formed between the Phe192
and the phenyl on position C6 of the pyridine ring and is maintained
during all the simulation (Figure D). As previously reported, the change from the inactive
to the active form of MNK1 results in a slightly different shape of
the ATP binding pocket but is mainly characterized by the flip of
the Phe192 residue from the ATP binding site. While the change in
the shape of the ATP binding pocket should not interfere with the
formation of the other interactions described above, the π–π
interaction between Phe192 and the phenyl on position C6 of the pyridine
ring will not be possible in the active conformation.Therefore,
our modeling data demonstrate a stable interaction with
MNK1 in its inactive conformation (Figure E) and the interaction with Phe192 indicates
that EB1 might prevent the flip-out of the DFD motif
and thereby prevent ATP binding and kinase activity. Remarkably, the
DFD motif is a unique feature of MNKs that induces the proteins to
adopt an unusual stable inactive conformation, which is not common
for other kinases.
EB1 Blocks MNK Activity without
Paradoxical Activation
of MNK Functions
Having identified a biologically active
inhibitor of MNK1 with a novel model of action, we finally compared
the activity and selectivity of EB1 to the classical
ATP-competitive inhibitors CGP57380 (2) and eFT508 (3) in cell-based assays. Based on eIF4E phosphorylation as
a readout for MNK activity, EB1 appears to be more potent
than CGP5738 but less active than the eFT508 (Figure A). The activating phosphorylation of MNK1
at T209/214 by the upstream kinases p38 and ERK was not affected by EB1, supporting that reduction in eIF4E phosphorylation in
cells is a direct consequence of targeting the kinase activity of
MNKs. These results are in striking difference from both Type I inhibitors,
which strongly induced phosphorylation of MNKs at T209/214, rendering
the kinase in an activated state. This paradoxical feature of certain
Type I kinase inhibitors—induction of the activated state of
the kinase while blocking the kinase activity—has been described
before.[35,36] In the case of CGP57380 (2),
increased phosphorylation at T209/214 has been reported previously.[29,53] Having EB1 as an inhibitor with a novel mode of action,
we aimed next at deciphering the consequences of inhibiting the kinase
activity of MNK1 by different modes of action.
Figure 6
Comparison of EB1 to Type-I MNK inhibitors. (A) Western
blot analysis of MDA-MB-231 cells treated for 24 h with the indicated
concentrations of EB1, eFT508 (eFT), and CGP57380 (CGP).
(B) Treatment of MDA-MB-231 cells with the indicated concentrations
of EB1, eFT508 (eFT), and CGP57380 (CGP) for 24 h followed
by treatment with UV irradiation and recovery for 40 min. Western
blot analysis was carried out with the indicated antibodies. (C) Immunoprecipitation
of endogenous eIF4G from cell extracts of HEK-293 cells treated for
24 h with DMSO, 20 μM EB1, or 100 nM eFT508 followed
by UV irradiation and recovery for 40 min. Western blot analysis of
Immunoprecipitate (IP) and total cell extracts (input) are shown.
(D) Schematic drawing indicating the activity status of MNK1 and the
MNK1–eIF4G interaction. Proteins are described in the figure.
Phosphorylation events are indicated by -P. The activation loop of
MNK1 is depicted in gray, and active MNK is indicated by bound ATP.
Top left: inactive MNK1 is not bound to eIF4G; top right: activation
of MNK1 by upstream kinase signaling and binding to eIF4G; bottom
left: Type-II inhibitor binding maintains MNK1 in the inactive conformation
and MNK1 does not bind to eIF4G; bottom right: Type-I inhibitor binding
renters MNK1 into the active conformation allows phosphorylation by
upstream kinases and binding to eIF4G. (E) MDA-MB-231 cells treated
for 72 h with the indicated concentrations of EB1 and
eFT508. Cell growth was determined by crystal violet staining. Error
bars represent the standard deviation of three independent experiments.
Comparison of EB1 to Type-I MNK inhibitors. (A) Western
blot analysis of MDA-MB-231 cells treated for 24 h with the indicated
concentrations of EB1, eFT508 (eFT), and CGP57380 (CGP).
(B) Treatment of MDA-MB-231 cells with the indicated concentrations
of EB1, eFT508 (eFT), and CGP57380 (CGP) for 24 h followed
by treatment with UV irradiation and recovery for 40 min. Western
blot analysis was carried out with the indicated antibodies. (C) Immunoprecipitation
of endogenous eIF4G from cell extracts of HEK-293 cells treated for
24 h with DMSO, 20 μM EB1, or 100 nM eFT508 followed
by UV irradiation and recovery for 40 min. Western blot analysis of
Immunoprecipitate (IP) and total cell extracts (input) are shown.
(D) Schematic drawing indicating the activity status of MNK1 and the
MNK1–eIF4G interaction. Proteins are described in the figure.
Phosphorylation events are indicated by -P. The activation loop of
MNK1 is depicted in gray, and active MNK is indicated by bound ATP.
Top left: inactive MNK1 is not bound to eIF4G; top right: activation
of MNK1 by upstream kinase signaling and binding to eIF4G; bottom
left: Type-II inhibitor binding maintains MNK1 in the inactive conformation
and MNK1 does not bind to eIF4G; bottom right: Type-I inhibitor binding
renters MNK1 into the active conformation allows phosphorylation by
upstream kinases and binding to eIF4G. (E) MDA-MB-231 cells treated
for 72 h with the indicated concentrations of EB1 and
eFT508. Cell growth was determined by crystal violet staining. Error
bars represent the standard deviation of three independent experiments.As detailed above, MNKs possess a specific regulatory
mechanism,
in which the kinase activity is autoinhibited by the conversion of
the activation segment into an autoinhibitory module.[24] The release of this autoinhibition requires phosphorylation
of MNKs by p38 or ERK.[23,24]To test the efficacy of EB1 after stimulation of MNK
kinase activity, MDA-MB-231 cells were pretreated for 3 h with EB1 or eFT508 (3), exposed to UV irradiation,
and recovered for 40 min in the presence of the inhibitors. In line
with previous reports, UV irradiation strongly increases the activating
MNK1 phosphorylation at T209/214 and consequently increases the phosphorylation
of eIF4E (Figure B).
Also, under these conditions, EB1 strongly inhibited
eIF4E phosphorylation and did not affect the phosphorylation of MNK1
at T209/214. Moreover, activation of p38 (p-p38) and phosphorylation
of its downstream effector HSP27 were not affected. Although p-ERK
and its substrate p-p90-RSK appeared slightly elevated in EB1-treated cells, this effect would rather counteract inhibition of
MNKs and similar effects were observed with the other MNK inhibitors
tested. These results therefore indicate that EB1 acts
directly on MNK kinases without perturbing activating upstream signaling.
Moreover, pretreatment of cells with EB1 appears to maintain
MNK1 in the inactive conformation even when the signaling of the activating
upstream kinases is enhanced, which supports our molecular dynamics
studies.In comparison to EB1, striking differences
in the
activation state of MNK1 could be also after UV irradiation, when
cells were treated with eFT508 (3). While EB1 did not alter the phosphorylation of MNK1 at T209/214, MNK1 was
hyper-phosphorylated in eFT508-treated cells (Figure A). Similar to EB1, upstream
p38 signaling was not affected by eFT508. The increase in ERK activity
after eFT508 treatment may account for the activation of MNK1. However,
as the same occurs with EB1, this possibility appears
to be rather unlikely but cannot be ruled out rigorously at this point.In summary, ATP-competitive MNK inhibitors, while being effective
in inhibiting the kinase activity, strongly alter the turnover of
the phosphorylation event at T209/214 and induce hyper-phosphorylation
of the activation loop, which might be caused by promoting the adaption
of the kinase into the active conformation. According to our data, EB1 inhibits the kinase activity by preventing the adaption
into the active conformation and might thereby also prevent MNKs from
being hyper-phosphorylated.Importantly, phosphorylation of
MNK1 has been shown to not only
activate its kinase activity but also to enhance its binding to the
eukaryotic initiation factor 4G (eIF4G), a scaffold protein of the
translation initiation complex.[54] We therefore
hypothesized that the hyper-phosphorylation of MNK1 upon eFT508 treatment
should result in increased binding to eIF4G. To test this, we immuno-precipitated
endogenous eIF4G from HEK-293 cells and tested for interaction with
MNK1 by western blotting (Figure C). The treatment scheme combining UV and MNK inhibitors
was the same as described above. Also, in HEK-293 cells, treatment
with EB1 did not alter the phosphorylation of state of
MNK1 at T209/T214 compared to the respective control (Figure C, input). Moreover, stronger
phosphorylation of MNK1 could be detected after treatment with eFT508
and the combination with UV irradiation resulted in hyper-phosphorylation
of MNK1 at T209/T214. Under all tested conditions, MNK1 binding to
eIF4G strictly correlated with the phosphorylation of MNK1 at T209/214,
and consequently, the strongest interaction of eIF4G with MNK1 could
be detected with hyper-phosphorylated MNK1 after UV irradiation of
eFT508-treated cells. On the contrary, EB1 treatment,
while inhibiting eIF4E phosphorylation to a similar extent to eFT508,
did not alter the interaction between MNK1 and eIF4G following UV
irradiation. This furthermore supports our hypothesis that EB1 acts different from the classical ATP-competitive inhibitors, by
preventing the kinase to adopt its active conformation, which is not
only a prerequisite for the induction of the kinase activity but also
for the interaction with eIF4G.Finally, we aimed at addressing
if the different mode of action
of MNK1 inhibition causes distinct cellular phenotypes. Previous reports
indicate that MNK1 degradation by VNLG-152R but not inhibition of
the kinase activity by eFT508 caused a reduced number of viable MDA-MB-231
cells in vitro.[51] In line
with this, knock-down of MNK1 has been reported to affect cell growth
in different cancer cell line models in vitro.[49,50] We therefore compared cell growth inhibition between the eFT508
and EB1. In line with the above results, EB1 treatment caused a dose-dependent growth defect in MDA-MB-231 with
a maximal reduction to 60% at 20 μM (Figure E). Importantly, this concentration does
not affect cell growth of normal IMR90 cells (Figures C and S4C). In
line with previous reports, treatment with eFT508 up to 200 nM was
not able to inhibit the growth of MDA-MB-231 cells (Figures C and S4C).[51] However, complete inhibition
of eIF4E phosphorylation can be already observed at low nanomolar
concentrations of eFT508 (Figure A).In summary, our molecular docking studies
revealed that EB1 binds with the highest affinity to
the inactive form of MNK1. The
accompanied inhibition of the MNK1 kinase activity therefore suggests
that EB1 acts as Type II kinase inhibitor and prevents
the adoption of the active state of the MNK1 kinase domain. Acquisition
of the active state is a prerequisite for not only MNK1 kinase activity
but also interaction with eIF4G. While the kinase activity can be
efficiently blocked with different types of kinase inhibitors, only
ATP-competitive inhibitors enhance the interaction of MNK1 with eIF4G
(Figure D). The differences
in the mode of action of MNK1 inhibitors might ultimately result in
different cellular phenotypes as exemplified by the inhibition of
cellular growth of tumor cells, which is not affected by ATP-competitive
inhibitors but by EB1.
Conclusions
We
describe here the identification and validation of the novel
MNK inhibitor EB1 and compare its mode of action to the
currently available ATP-competitive inhibitors.Starting from
the scaffolds of previously reported MNK inhibitors,
we systematically designed and synthesized three families of compounds
and tested a total of 68 compounds in in vitro kinase
assays. Within the family of pyrazolo[3,4-b]pyridin-3-amines,
we identified the first hit compound EB1. Subsequently,
several derivatives were designed, synthesized, and tested in in vitro and in cell-culture-based assays. The activity
profile of this set of compounds confirmed the activity of the core
structure as MNK inhibitors and allowed us to perform a first SAR,
which revealed that the unsubstituted 3-aminopyrazole ring and aryl
groups on positions C4 and C6 of the pyridine ring are important for
the activity of the compound.A detailed characterization of EB1 revealed a higher
affinity for MNK1 compared to MNK2, a fast and sustained mode of action
of EB1 in cell culture and a selective inhibition of
tumor cells growth, while normal cells were not affected. Based on
the previously described genetic models, this selectivity profile
should be expected when blocking the function of MNKs and therefore
support specificity of EB1. Studies combining molecular
docking and molecular dynamics (MD) revealed that EB1 binds to the inactive form of MNK1 (DFD-out), which is typically
seen for Type-II kinase inhibitors. Based on these data, we set out
to compare EB1 as the first validated Type-II MNK inhibitor
to the classical ATP-competitive inhibitors. This comparison finally
revealed that while both types of MNK inhibitors efficiently block
the kinase activity, ATP-competitive inhibitors appear to induce a
pseudo-activated state of the kinase, defined by T209/214 hyper-phosphorylation
through the upstream priming kinases. Yet, phosphorylation of MNKs
at T209/214 has been reported to be required for eIF4G binding,[54] and consequently, an enhanced binding of MNK1
to eIF4G can be observed only in the presence of Type-I inhibitors.
While the structural roles of MNK1 beyond its catalytic activity have
not yet been studied in detail, it is reasonable to believe that this
might have implications on the functionality of the eIF4F complex.
Supporting this, many other protein kinases, including the MNK upstream
kinases ERK and p38, have been reported to carry out additional functions
such as scaffolding.[55] Moreover, genetic
ablation of MNK1, in which both structural and catalytic activities
are impaired, has been reported to block tumor cell growth[49,50] and only the novel MNK inhibitor EB1 is able to phenocopy
this.Kinases are attractive, “druggable” targets,
and
after the clinical approval of the first kinase inhibitor Imatinib
in 2001, 73 kinase inhibitors have received US FDA by 2021.[56,57] Remarkably though, while Imatinib is a Type II kinase inhibitor,
previous analyses have revealed that the vast majority of inhibitors
(33 of 42) are classified as Type I inhibitors and only six (including
Imatinib) are Type II.[58] Despite the remarkable
clinical success of several of those drugs, resistance to kinase inhibitors
through paradoxical activation of the target pathway currently limits
the success of some of those inhibitors. Paradoxical activation of
the target pathway was first described by Okuzumi et al.[59] as unintended activation following inhibitor
binding to the kinase. In recent years, inhibitors of many other protein
kinases, including the mutated form of BRAF (BRAFV600E),[60,61] PKB,[62] PKC,[63] PKD,[64] AMPK,[65] c-Src,[66] and ERK5,[67] have been described. Remarkably, in most of these cases,
ATP-competitive binding of the inhibitors induce conformational changes,
which activate noncatalytic functions of these kinases.[37,68]Our work has revealed that in the case of ATP-competitive
MNK inhibitors,
binding to the catalytic site promotes MNK priming by the upstream
kinases and recruitment to eIF4G, while inactivation of MNKs by EB1 induces neither kinase priming nor recruitment to eIF4G,
which might thereby prevent unintended activation or priming of MNK
functions.In recent years, the eIF4E phosphorylating kinases
MNK1/2 have
emerged as eligible targets for drug discovery in oncology. Interference
with the eIF4E/MNK axis, through either genetic models or the use
of knock-down techniques, resulted in promising antitumor effects
and the lack of adverse effects in dual MNK knock-out animals opened
up the possibility to develop nontoxic anticancer drugs. Based on
the wide knowledge on ATP-competitive inhibitors from other protein
kinases and the available crystal structures of MNKs, highly efficient
and selective inhibitors have been developed in recent years and some
of those, like eFT508 (3), BAY1143269, and ETC-206, have
entered clinical trial phases. The results of the Phase I studies
confirmed their safety, as predicted for targeting MNKs.We
have identified and validated now a novel MNK inhibitor EB1. Based on homology modeling and molecular simulations,
we predict EB1 to be a Type II kinase inhibitor because
it interacts with the DFD motif, a unique feature of MNKs and selectively
binds to the inactive form of MNK1. This binding mode of EB1 might ultimately resemble closer the well-studied knock-out models
of MNKs than Type I inhibitors because structural functions of the
activated kinase might be blocked in addition. Dissecting this aspect
in preclinical models will be key to elucidate in the future, if the
mainly genetic-based proposal of targeting MNKs as a promising new
strategy in oncology can be better achieved with Type I or Type II
inhibitors of MNK1/2.
Experimental Section
Chemistry
General
Information
All solvents and chemicals were
of reagent grade. Unless otherwise mentioned, all solvents and chemicals
were purchased from commercial vendors (Sigma-Aldrich, ABCR, Fluorochem,
and Acros Organics) and used without purification. 1H-
and 13C NMR spectra were recorded on a Varian 400-MR spectrometer
(1H NMR at 400 MHz and 13C NMR at 100.6 MHz).
Chemical shifts were reported in parts per million (δ) and are
referenced to the residual signal of the solvent DMSO-d6 (2.5 ppm in 1H NMR and 39.5 ppm in 13C NMR). Coupling constants are reported in hertz (Hz). Standard and
peak multiplicities are designed as follows: s, singlet; d, doublet;
dd, doublet of doublets; dt, doublet of triplets t, triplet; q, quadruplet;
qn, quintuplet; br, broad signal. HRMS data were obtained using a
micrOTOF (Bunker) high-resolution spectrometer (EI, ESI or APCI mode).
Elemental microanalyses were performed on a EuroVector Instruments
Euro EA 3000 elemental analyzer. The melting points were determined
with an SMP3 melting point apparatus (Stuart Scientific) and are uncorrected.
Automatic flash chromatography was performed in an Isco Combiflash
medium-pressure liquid chromatograph with RediSep silica gel columns
(35–70 μm) or basic alumina columns using a suitable
mixture of solvents as an eluent. Microwave irradiation experiments
were carried out in an Initiator (Biotage) microwave apparatus, operating
at a frequency of 2.45 GHz with continuous irradiation power from
0 to 400 W. Reactions were carried out in 2.5, 5, and 20 mL glass
tubes sealed with aluminum/Teflon crimp tops, which can be exposed
up to 250 °C and 20 bar internal pressure. Temperature was measured
with an IR sensor on the outer surface of the process vial. After
the irradiation period, the reaction vessel was cooled rapidly to
50 °C by air jet cooling.
Purity Statement
Purity (≥95%) of the compounds
for this study was confirmed by elemental analysis (within ±0.4%
of the calculated value). Additionally, lead compounds were tested
by HPLC (see the Supporting Information).
3-Amino-5-(2,6-dichlorophenyl)-2-phenyl-2,4,5,7-tetrahydro-6H-pyrazolo [3,4-b]pyridine-6-one (S15e) (50 mg, 0.1 mmol) was suspended in 3 mL of POCl3. The mixture was refluxed for 3 h at 120 °C. The POCl3 was removed under reduced pressure, and the crude was resuspended
in water and neutralized with NaHCO3. The solid was filtered
and purified by column chromatography (alumina column, Cy:AcOEt gradient
0 to 60% in 30 min) to give 13 mg (10%) of 6-chloro-5-(2,6-dichlorophenyl)-2-phenyl-2H-pyrazolo[3,4-b]pyridin-3-amine (5a) as a yellowish solid.Mp: >250 °C. 1H NMR (400 MHz, DMSO-d6): δ
8.16
(s, 1H), 7.73–7.69 (m, 2H), 7.67–7.59 (m, 4H), 7.55–7.48
(m, 2H), 6.95 (s, 2H). 13C NMR (100 MHz, DMSO-d6): δ 155.8, 151.3, 142.9, 138.5, 136.5, 135.8,
135.6, 131.5, 129.9, 128.7, 128.7, 125.1, 120.0, 101.1.
3-Amino-4-phenyl-2-phenyl-2,4,5,7-tetrahydro-6H-pyrazolo[3,4-b]pyridine-6-one (S15h) (100 mg, 0.33 mmol) and 341 mg (1.64 mmol) of PCl5 were dissolved in 2.5 mL of POCl3, and the mixture
was refluxed for 18 h. The mixture was poured into ice, and the precipitate
was filtered and washed with water. The solid was then resuspended
in water, and the suspension was neutralized with NaHCO3. The solid was filtered and dried in vacuo over
P2O5 to yield 5,6-dichloro-2,4-diphenyl-2H-pyrazolo[3,4-b]pyridin-3-amine (5b) in 72% yield.Mp: 130–133 °C. 1H NMR (400 MHz, DMSO-d6): δ 7.67–7.56
(m, 7H), 7.55–7.48 (m, 3H), 5.02 (s, 2H). 13C NMR
(100 MHz, DMSO-d6): δ 153.4, 149.8,
146.0, 140.8, 137.6, 133.3, 129.7, 129.6, 129.1, 128.7, 128.3, 124.9,
115.2, 100.1. HRMS (ESI) m/z: calculated
for C18H13Cl2N4 [M + 1]+: 355.0512; found [M + 1]+: 355.0513.
General
Method for the Synthesis of 2-Methoxynicotinonitriles
(9)
NaOMe (101 mg, 1.87 mmol) was dissolved
in 5 mL of MeOH in a microwave vial. Arylketone (7) (1.7
mmol) and arylaldehyde (8) (1.7 mmol) were added and
the mixture was stirred for 5 min at room temperature. Then, 1.7 mmol
of malononitrile were added and the reaction was heated under microwave
irradiation for 5 min at 140 °C. The solvent was removed under
reduced pressure. The solid obtained was resuspended in MeOH and filtered
to afford the corresponding 2-methoxynicotinonitrile (9).
General Method for the
Synthesis of 2-Bromonicotinonitriles
(10)
The corresponding 2-methoxy-nicotinonitrile
(9) (0.6 mmol) was dissolved in 4 mL of 1,4-dioxane together
with POBr3 (1.34 mmol), pyridinium HBr (0.015 mmol), and
H3PO4 (0.026 mmol). The mixture was heated at
60 °C for 18 h under argon atmosphere and then quenched with
cold water. Then, the mixture was neutralized with NaOH (6 M) and
the precipitate was filtered and washed with cold water. The solid
was dried in vacuo over P2O5 to yield the corresponding 2-bromonicotinonitrile (10).
General Method for the Synthesis of Pyrazolo[3,4-b]pyridine-3-amines (6)
The corresponding 2-bromonicotinonitrile
(10) (0.18 mmol) and hydrazine (0.36 mmol) were dissolved
in 3 mL of methanol and sealed in a 5 mL microwave vial. The mixture
was heated under microwave irradiation for 2 h at 140 °C and
cooled rapidly. The solvent was removed under reduced pressure, and
the crude was resuspended in methanol. The solid was filtered, washed
with methanol, and dried in vacuo over P2O5 to yield the corresponding pyrazolo[3,4-b]pyridine-3-amine (6).
The inactive unphosphorylated
model was created using the available crystal structure of MNK1 (PDB
ID: 2HW6(24)). The tertiary structure of the missing fragments
was predicted de novo by applying the loop modeler
module available in MOE2016.08 [Molecular Operating Environment (MOE),
2016.08. Chemical Computing Group ULC, 1010 Sherbooke St. West, Suite
#910, Montreal, QC, Canada, H3A 2R7 (2016)]. The best loop candidate
was energy-minimized by performing a short molecular dynamics (MD)
simulation, using AMBER software [Case, D. A. et al. AMBER 14. University
of California, San Francisco (2014)].The tridimensional structure
of the active phosphorylated MNK1 model was obtained by homology modeling
using the active MNK2 model as a template, which was obtained from
the MNK2 D228G crystal structure (PDB ID: 2HW7), manually reverted the mutation G228D
and prepared with the same protocol used for MNK1. The sequence of
MNK1 and the MNK2 model were aligned in MOE2016.08, and the tridimensional
structure was created based on the tertiary structure of the MNK2
model obtaining the active MNK1 structure. Finally, Thr209 and Thr214
residues were manually phosphorylated.The two models were then
submitted to an MD simulation. The system
was subjected to a first minimization including a 5000-step minimization
of TIP3P water molecules followed by a 20 000-step energy minimization
of the entire system. Then, heated to 300 K in 200 ps using the Langevin
thermostat restraining the backbone atoms except the loop atoms using
an 8 kcal·mol–1·Å–2 force constant. Pressure equilibration (1 atm) was performed for
1000 ps maintaining the restraints previously described. The SHAKE
algorithm[69] was used throughout to restrain
the bonds involving hydrogens and the particle mesh Ewald[70] method for long-range electrostatic, while the
short-range interactions had an 8 Å cutoff radius. The production
stage was extended to at least 70 ns defining a 2 fs time-step (until
root-mean-square deviation (RMSD) was stable). Amber ff13 force field
was used for the parameters of standard amino acids. The parameters
described by Homeyer et al. were used for phosphorylated amino acids.[71]Finally, an accelerated MD (aMD) simulation
was performed to explore
the accessible conformational space. The parameters used in each of
the aMD simulations were calculated from the corresponding MD simulation
and are the following for the inactive and active model, respectively:
average total potential energy threshold (E(tot)):
−205630.72 and −147028; inverse strength boost factor
for the total potential energy (α(tot)): 14664.6 and 10465;
average dihedral energy threshold (E(dih)): 383.8
and 343.2; inverse strength boost factor for the dihedral energy (α(dih)):
209.3 and 206.The simulations were extended until the RMSD
values were stable:
90000 steps for the inactive model and 120 000 steps for the
active model. All MD simulations in this study were performed in a
4.20 GHz × 8 Intel Core i7-7700K processor with 16 GB RAM and
a GeForce GTX 1050 Ti GPU. Cluster analysis was performed with CPPTRAJ.[72] For each simulation, the most populated and
stable cluster was selected as the representative structure of each
state of the protein.
Complex Preparation and MD Simulation
For each model,
the most plausible active site was identified by blind docking, using
MOE2016.08 on a 2.6 GHz Intel Core i7-6700HQ with 8 GB RAM. Docking
poses were generated using the triangle matcher placement method (1000
poses) with the induced fit protocol and scored using London ΔG
scoring function and GBVI/WSA ΔG for rescoring
(100 poses). The most populated cluster was defined as the active
site. For each model, a conformation of the ligand found in the identified
active site was selected to study the interaction with the protein.An MD simulation was performed for each ligand–protein complex
using the same conditions previously described. The production stage
was extended to at least 50 ns defining a 2 fs time-step (until RMSD
was stable).
Characterization of the Interactions
The study of the
hydrogen-bond pattern was performed using CPPTRAJ.[72] Hydrogen bonds were determined using simple geometric criteria
such as the donor to acceptor heavy atom distance.Ligand–protein
interactions were visualized with MOE2016.08 (MOE2016.08. Chemical
Computing Group ULC, 1010 Sherbooke St. West, Suite #910, Montreal,
QC, Canada, H3A 2R7 (2016)).
MNK1 and MNK2 Enzymatic
In Vitro Assay
Measurement of the Residual Activity
Determination
of the effect of the compounds on the kinase activity of MNKs (and
other kinases) in vitro was performed with a radiometric
protein kinase assay (33PanQinase Activity Assay) at Reaction
Biology (www.reactionbiology.com). Protein kinases were purchased from Life Technologies (Invitrogen
Corporation). The compounds were provided as solids and were dissolved
to 1 × 10–3 M stock solutions in 100% DMSO
(100 × stock). The 10 × stock solution was prepared by dissolving
10 μL of the 100 × stock in 90 μL of water resulting
in a 1 × 10–4 M/10% DMSO solution.The
final working concentration was 1 × 10–5 M,
resulting in a final concentration of 1% DMSO. Active compounds were
tested in replicates as indicated.
Protein Kinase Assay
A radiometric protein kinase assay
(33PanQinase Activity Assay) was used for measuring the
kinase activity of the two protein kinases MNK1 and MNK2. All kinase
assays were performed in 96-well FlashPlatesTM from PerkinElmer (Boston,
MA) in a 50 μL reaction volume. The reaction cocktail was pipetted
in four steps in the following order: (1) 20 μL of assay buffer,
(2) 5 μL of ATP solution (in H2O), (3) 5 μL
of test compound (in 10% DMSO), and (4) 20 μL of enzyme–substrate
mixture.The assay for all protein kinases contained 70 mM HEPES–NaOH
pH 7.5, 3 mM MgCl2, 3 mM MnCl2, 3 μM Na-orthovanadate,
1.2 mM DTT, 50 μg/mL PEG20000, ATP (variable amounts,
corresponding to the apparent ATP-Km of
the respective kinase, 1 μM for MNK1 and 0.3 μM for MNK2),
[γ-33P]-ATP (approx. 1.2 × 106 cpm
per well), protein kinase (variable amounts depending on the stock),
and substrate (S6 peptide, 2 μg/50 μL).The reaction
cocktails were incubated at 30 °C for 60 min.
The reaction was stopped with 50 μL of 2% (v/v) H3PO4, and the plates were aspirated and washed two times
with 200 μL of 0.9% (w/v) NaCl. The incorporation of 33Pi was determined with a microplate scintillation counter (Microbeta,
Wallac). All assays were performed with a Beckman Coulter/SAGIAN Core
System.
Evaluation of Raw Data
The residual activity (in %)
for each well of a particular plate was calculated using the following
formulawhere cpm is the counts
per minute measured
for the mix containing the compound, “low control” reflects
unspecific binding of radioactivity to the plate in the absence of
a protein kinase but in the presence of the substrate, and “high
control” is full activity in the absence of any inhibitor.
The difference between high and low controls was taken as 100% activity.
Quality Controls
As a parameter for assay quality,
the Z′-factor 10838414[73] for the low and high controls of each assay plate (n = 8) was used. Reaction Biology’s criterion for
repetition of an assay plate is a Z′-factor
below 0.4.[74]
Determination of the IC50
The IC50 profile of the compounds was
determined using the previously described
radiometric protein kinase assay (33PanQinase Activity
Assay). IC50 values were measured by testing 10 concentrations
of the compounds in the range of 5 × 10–5 to
1.5 × 10–9 M.
Cell-Culture-Based Experiments
Cell
Culture and Reagents
Breast cancer cell lines
were purchased from the American Type Culture Collection (ATCC). MDA-MB-231,
MDA-MB-468, MCF7, A375M, IMR90, and HEK-293 cells were grown in DMEM
(Dulbecco’s modified Eagle’s medium: 4.5 g/L d-glucose; 580 mg/L l-Glutamine (Gibco)) supplemented with
10% heat-inactivated fetal bovine serum (FBS) (Life Technologies),
100 U/mL penicillin, and 100 μg/mL streptomycin (Life Technologies).MV4-11 cells were grown in IMDM medium (Iscove’s modified
Dulbecco’s medium, Gibco) supplemented with 10% FBS, 1% l-glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin.Prostate cancer cell line 22Rv1 was obtained from ATCC. Cells were
cultured in RPMI-1640 (Biowest) supplemented with 10% heat-inactivated
FBS, 2 mM l-glutamine, 1% penicillin-streptomycin solution,
1% MEM non-essential amino acids, and 1% sodium pyruvate (all from
Biowest). The cells were maintained at 37 °C in a 5% CO2 humidified incubator.
Inhibitors and Cell Treatment
For
inhibitor treatments,
the cells were seed in at a density of 6000 cells/cm2 in
the respective plates 24 h before the treatment. Compounds were diluted
to the final concentration in a medium maintaining DMSO at a concentration
below 0.5%. DMSO was used as the negative control, and CGP57380 (Merck)
and eFT508 (Selleckchem) were used for comparison.
UV Treatment
MDA-MB-231 or HEK293T cells were incubated
with the indicated concentrations of MNK1/2 inhibitors for 24 h and
then irradiated with UV (60 J/m2) followed by 40 min incubation
at 37 °C. Total cell lysates were analyzed by immunoblotting
with the indicated antibodies.
Cristal Violet Growth Assay
The cells were seeded in
a 24- or 96-well plate at a density of 6000 cells/cm2,
and after 24 h, treatments were carried out. After the corresponding
time, the medium was removed and the cells were fixed with 4% formaldehyde
solution for 30 min and washed twice with phosphate-buffered saline
(PBS). The cells were stained with crystal violet solution (0.5% in
water) for 15 min, washed extensively with water, and air-dried. Crystal
violet was dissolved in 200 μL of 15% AcOH, and the optical
density of the solution was measured at 595 nm (Epoch Microplate Spectrophotometer,
BioTek).
Propidium Iodide-Hoechst-Based Analysis of
Cell Death
Prior to fixation of cells for the crystal violet
assay, propidium
iodide (2.5 μM) and Hoechst (0.05 mg/mL) in complete culture
medium was added to the cells without removal of the culture medium,
incubated for 30 min, and representative pictures were taken from
each well using a Nikon Eclipse TE2000-S microscope. A minimum of
100 cells were counted per experimental condition, and the percentage
of PI-positive cells among the Hoechst positive cells was determined.
Sulforhodamine B (SRB) Cytotoxicity Assay
Cells were
seeded at a density of 5000 cells per well in a 96-well tissue culture
plate. After 24 h, the cells were treated with different concentrations
of MNK1/2 inhibitors. Control cultures received the same amount of
DMSO as did the treated cultures. Doxorubicin at a 1 μM concentration
was used as a positive control. On day 3, cell numbers were calculated
using the SRB assay. The SRB assay was performed following protocols
adapted from Euglossa Orellana (2016) with minor modifications.[75] Briefly, the medium was discarded and the adherent
cells were fixed in situ by adding to each well 100
μL of cold trichloroacetic acid 10% (w/v) and incubating for
60 min at 4 °C. The plates were then washed four times with deionized
water and air-dried. Each well then received 100 μL of SRB (Merck,
S1402-5G) solution (0.04% w/v in 1% acetic acid), and the plates were
incubated for 1 h at room temperature. Unbound SRB was removed by
washing four times with 1% acetic acid. Then, the plates were air-dried.
Bound stain was solubilized with 100 μL of 10 mM unbuffered
Tris base (pH 10.5), and the optical densities were read using a microtiter
plate reader at 510 nm. Four replicates were used for each analysis.
The percentage of growth inhibition was calculated using the equation:
% Growth Inhibition = (1 – Nt/Nc) × 100, where Nt and Nc represent the absorbance
in treated and untreated cultures, respectively.
Flow Cytometry
The effect of EB1 on the
cell cycle of MDA-MB-231 cells was analyzed by flow cytometry measurement
of propidium iodide (PI)-stained cells. In brief, 1 × 106 cells were seeded in 10 cm plates and treated with the indicated
concentrations of compounds for 72 h. The cells were trypsinized,
washed twice with PBS, resuspended in 0.3 mL of PBS, and fixed by
the addition of 0.7 mL of ice-cold absolute EtOH under continuous
agitation. The cells were fixed for 2 h on ice. For PI staining, the
cells were pelleted and the supernatant was discarded and washed twice
with PBS. The cells were suspended in staining solution (RNAse A (300
μg/mL, Sigma), sodium citrate (1.14 mM), and propidium iodide
(15 μg/mL, Sigma)) and incubated at 4 °C overnight. DNA
content was measured using a FACS Fortessa instrument integrated with
FACS Diva (BD Biosciences).
Protein Extraction and
Immunoblotting
Total protein
extracts were generated using lysis buffer (50 mM Tris-HCl, pH 7.4,
150 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 1
mM EDTA) supplemented with PhosSTOP and Complete Phosphatase/Protease
Inhibitor Cocktails (Roche Diagnostics GmbH, Mannheim, Germany). Protein
extracts (20–25 μg per sample) were loaded onto sodium
dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels
and transferred electrophoretically to poly(vinylidene fluoride) (PVDF)
membranes. The following primary antibodies were used at a dilution
of 1:1000: ERK1/2 (p44/42) (ab17942, Abcam), p-ERK1/2 (p-p44/42) (Thr202/Tyr204)
(9101, Cell Signaling), MNK1 (2195, Cell Signaling), p-MNK1 (Thr197/202)
(2111, Cell Signaling), eIF4E (9742, Cell Signaling), p-eIF4E (p-Ser
209) (9741, Cell Signaling or NBP2-66802, Novus Biologicals), p38alpha
(9218, Cell Signaling), p38beta (2339, Cell Signaling), p-p38 (4511,
Cell Signaling), p-ATF2 (27934, Cell Signaling), Hsp27 (2402, Cell
Signaling), p-Hsp27 (S82) (2401, Cell Signaling), and p-p90 RSK (Thr573)
(9346, Cell Signaling). The primary HRP-conjugated antibody anti-β-actin
(Calbiochem) was used at a dilution of 1:20 000. Anti-mouse
and anti-rabbit HRP secondary antibodies were from Pierce and used
at a dilution of 1:10 000. Immunodetection of proteins was
performed using ECL Western Blotting Detection Reagents (GE Healthcare,
Buckinghamshire, U.K.).
Immunoprecipitation
Cells were lysed
with lysis buffer
(50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% NP-40, 5 mM EGTA, 5 mM EDTA)
including protease and phosphatase inhibitor cocktails (PhosSTOP and
Complete). Immunoprecipitation of eIF4G was performed using anti-eIF4G
antibodies (Santa Cruz Biotechnology) and Pierce Protein A agarose
beads (Thermo Scientific) following the manufacturer’s indications.
Briefly, beads were incubated in 300 μL of ice-cold PBS containing
3 μL of anti-eIF4G antibodies for 4 h rotating at 4 °C.
Then, the beads were washed three times in cold PBS and incubated
in 500 μL of ice-cold precleared lysates containing approximately
2 mg of cell lysate overnight rotating at 4 °C. The next day
beads were recovered by washing twice in ice-cold lysis buffer and
resuspended in 1× loading buffer, boiled at 95 °C for 5
min, and analyzed by SDS-PAGE and immunoblotting.
Authors: Susann Santag; Franziska Siegel; Antje M Wengner; Claudia Lange; Ulf Bömer; Knut Eis; Florian Pühler; Philip Lienau; Linda Bergemann; Martin Michels; Franz von Nussbaum; Dominik Mumberg; Kirstin Petersen Journal: Cancer Lett Date: 2016-12-31 Impact factor: 8.679
Authors: Takeshi Ueda; Masato Sasaki; Andrew J Elia; Iok In Christine Chio; Koichi Hamada; Rikiro Fukunaga; Tak W Mak Journal: Proc Natl Acad Sci U S A Date: 2010-08-02 Impact factor: 11.205
Authors: Sarah Diab; Theodosia Teo; Malika Kumarasiri; Peng Li; Mingfeng Yu; Frankie Lam; Sunita K C Basnet; Matthew J Sykes; Hugo Albrecht; Robert Milne; Shudong Wang Journal: ChemMedChem Date: 2014-02-12 Impact factor: 3.466
Authors: Hans-Guido Wendel; Ricardo L A Silva; Abba Malina; John R Mills; Hong Zhu; Takeshi Ueda; Rie Watanabe-Fukunaga; Rikiro Fukunaga; Julie Teruya-Feldstein; Jerry Pelletier; Scott W Lowe Journal: Genes Dev Date: 2007-11-30 Impact factor: 11.361
Authors: Phillip A Geter; Amanda W Ernlund; Sofia Bakogianni; Amandine Alard; Rezina Arju; Shah Giashuddin; Abhilash Gadi; Jacqueline Bromberg; Robert J Schneider Journal: Genes Dev Date: 2017-12-21 Impact factor: 11.361
Authors: Fiona A Ross; Simon A Hawley; F Romana Auciello; Graeme J Gowans; Abdelmadjid Atrih; Douglas J Lamont; D Grahame Hardie Journal: Cell Chem Biol Date: 2017-06-15 Impact factor: 8.116
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6