The homologous cytokines macrophage migration inhibitory factor (MIF) and d-dopachrome tautomerase (d-DT or MIF2) play key roles in cancers. Molecules binding to the MIF tautomerase active site interfere with its biological activity. In contrast, the lack of potent MIF2 inhibitors hinders the exploration of MIF2 as a drug target. In this work, screening of a focused compound collection enabled the identification of a MIF2 tautomerase inhibitor R110. Subsequent optimization provided inhibitor 5d with an IC50 of 1.0 μM for MIF2 tautomerase activity and a high selectivity over MIF. 5d suppressed the proliferation of non-small cell lung cancer cells in two-dimensional (2D) and three-dimensional (3D) cell cultures, which can be explained by the induction of cell cycle arrest via deactivation of the mitogen-activated protein kinase (MAPK) pathway. Thus, we discovered and characterized MIF2 inhibitors (5d) with improved antiproliferative activity in cellular models systems, which indicates the potential of targeting MIF2 in cancer treatment.
The homologous cytokines macrophage migration inhibitory factor (MIF) and d-dopachrome tautomerase (d-DT or MIF2) play key roles in cancers. Molecules binding to the MIF tautomerase active site interfere with its biological activity. In contrast, the lack of potent MIF2 inhibitors hinders the exploration of MIF2 as a drug target. In this work, screening of a focused compound collection enabled the identification of a MIF2 tautomerase inhibitor R110. Subsequent optimization provided inhibitor 5d with an IC50 of 1.0 μM for MIF2 tautomerase activity and a high selectivity over MIF. 5d suppressed the proliferation of non-small cell lung cancer cells in two-dimensional (2D) and three-dimensional (3D) cell cultures, which can be explained by the induction of cell cycle arrest via deactivation of the mitogen-activated protein kinase (MAPK) pathway. Thus, we discovered and characterized MIF2 inhibitors (5d) with improved antiproliferative activity in cellular models systems, which indicates the potential of targeting MIF2 in cancer treatment.
Cancer is one of the
major public health challenges and contributes
to a currently estimated annual death toll of 10 million worldwide.[1] Although targeted cancer treatment has achieved
enormous progress over the last decades, its effectiveness is limited
by the heterogeneity and acquired therapy resistance of cancers.[2] Therefore, it is important to explore novel anticancer
drug targets and to develop new therapeutic agents to target them.
This could expand the possibilities to employ targeted therapeutic
approaches and also increase the possibilities to develop combination
therapy regimens.[3] The macrophage migration
inhibitory factor (MIF) family proteins are implicated in the development
of cancers, which is demonstrated by the overexpression of MIF family
proteins in several cancer types, such as genitourinary cancer,[4] melanoma,[5] neuroblastoma,[6] and lung carcinoma.[7] Notably, downregulation of MIF family proteins by gene-knockout[8] or gene-knockdown[9,10] exhibited
reduction of tumor progression and induction of antitumor immune responses.
Thus, MIF family proteins could be promising targets of novel cancer
therapeutics.The most studied member of the MIF family proteins
is macrophage
migration inhibitory factor (MIF), which was initially discovered
as an inflammatory cytokine. Currently, there is accumulating evidence
revealing a key role in the proliferation of cancer cells. MIF exerts
its proliferative effect through various mechanisms, for example,
binding to the cognate receptor cluster of differentiation 74 (CD74).[11,12] CD74 binding induces the activation of the MAPK pathway, which stabilizes
cyclin D1 and consequently regulates cell cycle progression to enhance
cell proliferation. In addition, activation of the MAPK pathway also
suppresses p53 activity, which results in the inhibition of cell apoptosis.[4] Other binding partners of MIF, such as CXCR4[13] and JAB1 also play central roles in cancer growth.
Several MIF-targeting reagents showed substantial potency on the deactivation
of MIF-related signaling pathway and the inhibition of cancer cell
proliferation.[14,15]d-Dopachrome tautomerase
(d-DT or MIF2) is a
structural and functional homologue but not a backup of MIF.[16] Most studies on MIF2 were comparisons with MIF,
as MIF2 is a relatively recently identified family member of MIF.[17] MIF2 and MIF share a high similarity in several
aspects. The 3D structure shows that the overall folding and subunit
topology of MIF2 and MIF are almost identical, with two β-α-β
motifs related by pseudo-2-fold symmetry and similar
trimeric β-sheet packing (Figure A).[18] Both MIF2 and MIF
harbor enzyme activity and catalyze the keto-enol tautomerization
of 4-hydroxylphenylpyruvate (4-HPP) in an active site centered around
proline-1. Moreover, both MIF and MIF2 are ligands of CD74 and JAB1[17] that could consequently endow these two proteins
with a similar effect on cell growth and tumorigenesis. Despite their
high structure similarity, MIF and MIF2 share just 34% percent amino
acid sequence identity (Figure B).[19] These sequence differences
may provide differences in interaction sites. For instance, the difference
of amino acids inside the tautomerase active sites causes differences
in their activity toward the keto-enol tautomerization of 4-HPP.[17] Moreover, MIF2 does not bind to MIF receptors
CXCR2/4 because it lacks a pseudo(E)LR motif, which mediates the interactions.[20,21] Thus, understanding the similarities and differences between MIF2
and MIF is important for the discovery of therapeutics to selectively
target these MIF family members.
Figure 1
Tertiary and secondary structures of MIF
and MIF2 and some key
inhibitors. (A) Similar 3D structures of MIF (PDB code: 3ijj)[31] and MIF2 (PDB code: 6c5f)[32] monomers.
(B) Comparison of the MIF and MIF2 amino acid sequences. (C) Some
representative MIF or MIF2 tautomerase inhibitors. ISO-1 is one of
the most studied MIF inhibitors.[25] 4-IPP
is a covalent inhibitor for both MIF and MIF2.[19] 4-CPPC is an MIF2 inhibitor with an IC50 value
of 27 μM with selectivity for MIF2 over MIF.[30]
Tertiary and secondary structures of MIF
and MIF2 and some key
inhibitors. (A) Similar 3D structures of MIF (PDB code: 3ijj)[31] and MIF2 (PDB code: 6c5f)[32] monomers.
(B) Comparison of the MIF and MIF2 amino acid sequences. (C) Some
representative MIF or MIF2 tautomerase inhibitors. ISO-1 is one of
the most studied MIF inhibitors.[25] 4-IPP
is a covalent inhibitor for both MIF and MIF2.[19] 4-CPPC is an MIF2 inhibitor with an IC50 value
of 27 μM with selectivity for MIF2 over MIF.[30]Considerable efforts have been
devoted to the discovery of small-molecule
inhibitors to target MIF family proteins in cancers. For the last
two decades, researchers have reported dozens of MIF-targeting inhibitors,
most of which bind to the tautomerase active site of MIF. As the tautomerase
active sites are located in close proximity to amino acid residues
playing key roles in binding to the CD74 receptor,[22,23] small-molecule binders of the MIF tautomerase site can also interfere
with the structural or dynamic features of MIF that are essential
for CD74 binding and MIF-related cell signaling.[24] For instance, the most widely investigated MIF inhibitor
ISO-1 (Figure C) not
only inhibits MIF tautomerase activity[25] but also shows significant suppression of prostate cancer growth
in cellular and animal models, which can be explained by the attenuation
of MIF-triggered activation of the MAPK pathway.[26] In a recent study, we demonstrated that a 7-hydroxycoumarin
derivative can interfere with MIF-CD74 interaction through binding
to the MIF tautomerase active site.[15] Additionally,
the Jorgensen Lab discovered several promising biaryltriazole- or
pyrazole-containing MIF inhibitors, which showed inhibitory potency
at the nanomolar level.[27,28] Moreover, we developed
a potent MIF-targeted proteolysis targeting chimera (PROTAC) to remove
the MIF protein from its interaction network, which further expanded
the toolbox to study MIF functions.[29] However,
little is known about the effect of MIF2 inhibitors on cancer development
due to a lack of potent MIF2 inhibitors and other effective molecular
tools. 4-Iodo-6-phenylpyrimidine (4-IPP) is the firstly discovered
MIF2 inhibitor, which covalently binds to proline-1 of MIF2 to interfere
with its tautomerase enzyme activity and its biological function.[19] However, 4-IPP shows low potency on MIF2 inhibition
with an IC50 value larger than 100 μM. In contrast,
4-IPP inhibits MIF with micromolar potency and binds covalently to
the active site proline. As reported in 2019, Bucala and co-workers
discovered the selective MIF2 inhibitor 4-(3-carboxyphenyl)-2,5-pyridinedicarboxylic
acid (4-CPPC) through virtual screening, which displays an IC50 value of 27 μM on MIF2 tautomerase activity.[30] More importantly, 4-CPPC can also inhibit MIF2-CD74
binding and MIF2-mediated pERK activation. To further investigate
the biological role of MIF2 in cancer and exploit it to develop cancer
therapy, more potent MIF2 selective inhibitors are needed.In
this study, we describe the discovery and exploration of a potent
and selective MIF2 inhibitor 5d with a low micromolar-level
potency for inhibition of both MIF2 tautomerase activity and cancer
cell proliferation (Figure ). First, the tautomerase activity assay was optimized for
assessing the potency of MIF2 inhibitors in competition binding studies
and enzyme kinetic analysis. Subsequently, several in-house available
compound collections were screened and a promising hit compound was
chosen for further optimization. Toward this aim, 33 analogues were
prepared and tested for their inhibitory potency on MIF2 tautomerase
activity. The most potent inhibitor 5d was tested in
two-dimensional (2D) and three-dimensional (3D) cultures of non-small
cell lung cancer (NSCLC) cell lines. Altogether, the development of 5d provides a powerful tool for MIF2-oriented research and
provides a perspective toward MIF2-directed therapeutics.
Figure 2
Workflow for
the discovery of MIF2 inhibitors. In this study, a
MIF2 tautomerase activity assay was established using phenylpyruvate
as a substrate. This assay was employed to screen several structurally
diverse compound collections. A hit compound was discovered and further
optimized. The most potent MIF2 tautomerase inhibitor was also tested
on non-small cell lung cancer cells for its effect on cell proliferation.
Workflow for
the discovery of MIF2 inhibitors. In this study, a
MIF2 tautomerase activity assay was established using phenylpyruvate
as a substrate. This assay was employed to screen several structurally
diverse compound collections. A hit compound was discovered and further
optimized. The most potent MIF2 tautomerase inhibitor was also tested
on non-small cell lung cancer cells for its effect on cell proliferation.
Results
Optimization MIF2 Tautomerase
Activity Assay
To facilitate
the effective assessment of binding potency of MIF2 inhibitors, a
convenient and reliable assay is needed. The most widely used assay
to evaluate MIF tautomerase activity inhibition is the 4-HPP-based
tautomerization assay, in which the potency on inhibition of MIF-catalyzed
4-HPP tautomerization is applied to reflect the binding affinity of
the tested compound to MIF.[33] This assay
was also applied to assess MIF2 binding in previous studies.[30] However, the catalytic activity of MIF2 on the
4-HPP keto-enol conversion is 10 times lower compared to MIF (Figure A).[17] This renders 4-HPP unsuitable as a substrate in the MIF2
tautomerase activity assay because high enzyme concentrations and
long measurement times are needed. This indicates the need for a different
substrate.
Figure 3
Setup of the phenylpyruvate tautomerization-based MIF2 binding
assay. (A) Rate of the MIF- or MIF2-catalyzed 4-HPP tautomerization
reaction. The rate of the MIF-catalyzed reaction is 10-fold faster
than that of the MIF2-catalyzed reaction.[17] (B) Superposition of MIF2 (PDB code: 6c5f)[32] and complex
(PDB: 6ijj)[31] of MIF and 4-HPP reveals the differences in
the tautomerase active site pocket. (C) UV spectra of 1 mM phenylpyruvate
in its keto-form (gray) or enol-form (blue). (D) The signal in the
MIF2 tautomerase activity assay is halved in the presence of 10 μM
Cu2+ and can be rescued by 0.5 mM ethylenediamine tetraacetic
acid (EDTA).
Setup of the phenylpyruvate tautomerization-based MIF2 binding
assay. (A) Rate of the MIF- or MIF2-catalyzed 4-HPP tautomerization
reaction. The rate of the MIF-catalyzed reaction is 10-fold faster
than that of the MIF2-catalyzed reaction.[17] (B) Superposition of MIF2 (PDB code: 6c5f)[32] and complex
(PDB: 6ijj)[31] of MIF and 4-HPP reveals the differences in
the tautomerase active site pocket. (C) UV spectra of 1 mM phenylpyruvate
in its keto-form (gray) or enol-form (blue). (D) The signal in the
MIF2 tautomerase activity assay is halved in the presence of 10 μM
Cu2+ and can be rescued by 0.5 mM ethylenediamine tetraacetic
acid (EDTA).The structural features of the
tautomerase active site of MIF2
were compared to the tautomerase active site of MIF. A superimposition
of the crystal structure of MIF2[32] on a
complex[31] of MIF and 4-HPP showed that
the proline-1 of MIF2 overlaps well with the proline-1 of MIF, which
indicates that the catalytic centers of MIF2 and MIF occupy the same
spatial location in the active sites. Nevertheless, MIF2 is not able
to form the two hydrogen bonds observed between Asn97 of MIF and the
4-position hydroxyl group of 4-HPP, as the Arg98 locates at the corresponding
position of MIF2 (Figure B). Accordingly, we hypothesized that MIF2 could exhibit different
catalytic activity for keto-enol tautomerization of 4-HPP analogues
with different substituents at the phenyl 4-position. Therefore, we
set out to find a more active substrate for MIF2 to establish a convenient
and sensitive enzymatic assay. Three 4-HPP analogues, 4-methoxylphenylpyruvate
(4-MPP), 4-chlorophenylpyruvate (4-CPP), and phenylpyruvate (PP),
were synthesized and the catalytic activities of MIF and MIF2 toward
these “artificial substrates” were tested. The measured
parameters show that KM values of MIF
on catalyzing PP, 4-CPP, 4-HPP, and 4-MPP keto-enol conversion are
0.76, 0.60, 0.94, and 1.3 mM (Table ), respectively. These values are lower than the corresponding
values of MIF2. Nevertheless, the catalytic efficiencies for MIF and
MIF2 on each substrate are diverse. For instance, the catalytic efficiency
of MIF on 4-HPP is 7 times higher than that of MIF2, making 4-HPP
a less effective MIF2 substrate. In contrast, MIF2 can catalyze the
keto-enol conversion of PP around 3 times more efficient than MIF,
which provides MIF2 with the highest catalytic efficiency for PP among
this series of substrates. Interestingly, PP is a known MIF2 substrate
and MIF2 was even named as phenylpyruvate tautomerase (PPT II) because
of this property as early as 1997.[34] Altogether,
this identifies PP as a substrate for a MIF2 tautomerase assay.
Table 1
Enzyme Kinetic Parameters for the
Conversion of PP and Its Derivatives by MIF and MIF2
enzyme
parameter
PP
4-CPP
4-HPP
4-MPP
MIF
KM (mM)
0.76 ± 0.28
0.60 ± 0.24
0.94 ± 0.14
1.3 ± 0.32
kcat (s–1)#
14 ± 2.5
25 ± 4.1
25 ± 1.8[37]
13 ± 1.7
kcat/KM (104 M–1 s–1)
1.9
4.3
2.6
0.98
MIF2
KM (μM)
2.4 ± 0.56
2.17 ± 0.22
3.62 ± 0.47
6.7 ± 0.78
kcat (s–1)
132 ± 17
68 ± 3.7
13 ± 1.0
28 ± 2.3
kcat/KM (104 M–1 s–1)
5.5
3.2
0.37
0.43
Next, PP was used to establish the
MIF2 tautomerase assay in the
same way as done for MIF. In this assay format, a MIF2 solution in
assay buffer was mixed with the compound stock solution and an aqueous
EDTA solution followed by 10 min preincubation.[35] The assay was started by adding a PP solution to the inhibitor–enzyme
mixture. In the final reaction mixture, there were 50 nM MIF2, 0.5
mM PP, 2.5% (v/v) dimethyl sulfoxide (DMSO), and variable concentrations
of the inhibitor. The formation of the reaction product was monitored
at 300 nm and corrected for the blank in which the enzyme was excluded
(Figure C). The linear
regression parameters were utilized to determine the inhibitory potency
or IC50 using GraphPad Prism. The window coefficient (Z′-factor) of this optimized assay was determined
as 0.75 in this setup, indicating that the quality of this assay is
sufficient for application in medium- to high-throughput screenings
(0.5–1).[36] Similar to MIF, we observed
that Cu2+ ions interfere with this MIF2 tautomerase activity
assay and that including EDTA in the assay buffer prevents this (Figure D).[35]
Compounds Screening and Hit Confirmation
The MIF2 tautomerase
activity assay was employed to screen an in-house available compound
collection containing 305 distinct chemical entities at compound concentrations
of 50 μM (Figure A and Supporting Table 1). The performance
of the MIF2 activity assay in this screening campaign proved to be
excellent with an average Z′ factor (a screening
window coefficient) of 0.87. The hit identification criterion was
defined as beyond 75% inhibition (Table S1). The inhibitory potency, potential binding efficiency, and drug-likeness
were used to select hit compounds for follow-up investigation. R110
(3a) exhibited good inhibitory potency against MIF2 tautomerization
activity (IC50 = 15 μM) and good ligand efficiency
(LE = 0.34) (Figure B). To verify the binding between R110 (3a) and MIF2,
the thermal stability assay was performed as an orthogonal assay.
The apparent melting temperature (ΔTm) of MIF2
increased with 2.5 °C in the presence of 100 μM R110 (3a) (Figure C), which confirms binding of R110 (3a) to MIF2. Subsequently,
the enzyme kinetics for binding of R110 (3a) to MIF2
were investigated. The Lineweaver–Burk plots show intersection
at the ordinate (y-axis), demonstrating that R110
(3a) binds in competition with the substrate phenylpyruvate
to the MIF2 tautomerase active site (Figure D). Taken together, R110 (3a) is a competitive inhibitor of MIF2 tautomerase activity with micromolar
potency, indicating that R110 (3a) is a promising chemical
starting point for the development of a more potent MIF2 inhibitor.
Figure 4
Library
screening and hit confirmation. (A) 305 compounds from
an in-house library were screened at a single concentration of 50
μM for their inhibition on MIF2 tautomerase activity. The inhibition
potency of each compound was shown as an average of three measurements.
(B) Structure of selected hit compound R110 (3a). (C)
Fluorescence-based thermal shift assay result from the interaction
between R110 and MIF2. ΔTm depicts the difference
between the apparent melting temperature of MIF2 with or without R110.
(D) Lineweaver–Burk plot of MIF2 activity in the presence of
R110 (3a). Results are shown as mean ± standard
deviation (SD) of three experiments.
Library
screening and hit confirmation. (A) 305 compounds from
an in-house library were screened at a single concentration of 50
μM for their inhibition on MIF2 tautomerase activity. The inhibition
potency of each compound was shown as an average of three measurements.
(B) Structure of selected hit compound R110 (3a). (C)
Fluorescence-based thermal shift assay result from the interaction
between R110 and MIF2. ΔTm depicts the difference
between the apparent melting temperature of MIF2 with or without R110.
(D) Lineweaver–Burk plot of MIF2 activity in the presence of
R110 (3a). Results are shown as mean ± standard
deviation (SD) of three experiments.
Design and Synthesis
To rationalize the binding between
R110 (3a) and MIF2, a docking study was performed by
docking R110 (3a) into the crystal structure of MIF2
(PDB code: 6c5f).[32] Modeling was performed using the
software Discovery Studio 3.0. R110 (3a) was docked into
the tautomerase active site of the crystal structure of MIF2 and energy-minimized
(Figure ). The five
highest-scoring poses occupy similar positions. The optimal binding
pose indicates that the chlorophenyl group of R110 (3a) is embedded into the hydrophobic active site pocket of MIF2, while
the n-butyl group is exposed to water. The interactions
between R110 (3a) and MIF2 are mainly formed through
van der Waals forces, including the interactions between the chlorophenyl
group and Phe2, Arg98, Leu102, and Ile107 of MIF2, and the interactions
between the thiophene and Arg36, Met114 of MIF2. A hydrogen bond is
formed between R110 (3a) and Lys109 of MIF2. This model
indicates that there is still some unoccupied space around the chlorophenyl
fragment and the butyl group of R110 (3a). Therefore,
we set out to exploit these regions in our structure–activity
relationship (SAR) with the aim to enhance the potency of the MIF2
inhibitor.
Figure 5
Schematic representation of the binding mode of R110 (3a) (thick sticks) with MIF2 (PDB code: 6c5f).[32] The residues
of the binding pocket are shown as thin sticks. The hydrogen bond
formed with Lys109 is shown as the red dash line. Discovery Studio
was employed for docking, and PyMOL was used for graph preparation.
Schematic representation of the binding mode of R110 (3a) (thick sticks) with MIF2 (PDB code: 6c5f).[32] The residues
of the binding pocket are shown as thin sticks. The hydrogen bond
formed with Lys109 is shown as the red dash line. Discovery Studio
was employed for docking, and PyMOL was used for graph preparation.To optimize R110 (3a), phenylthiophene
derivatives
were constructed employing the Gewald three-component reaction as
a key step using routes as depicted in scheme –. Two
different methods were employed to prepare compounds 3a–k. Compounds 3a–c and 3e–h were obtained using the Gewald-reaction employing
different aldehydes, cyanoacetamides, and elementary sulfur as starting
materials with yields of 18–56%. Compounds 3d and 3i–k were synthesized by condensation of the corresponding
ketones and cyanoacetamides in the presence of SnCl4 and
Et3N to provide the corresponding intermediates, which
were cyclized with S8 to afford the products with overall
yields of 16–42%. The resulting 2-aminothiophenes were acylated
by different acylchlorides to prepare the desired 2-amide-substituted
thiophene products 4a–c with yields of 44–58%.
The 2-aminothiophenes were also employed to synthesize thieno[2,3-d]pyrimidine-2,4(1H,3H)-diones 5a–e using 1,1′-carbonyldiimidazole
(CDI) as a coupling reagent with yields of 40–91%. The profiles
of 5d were further investigated. 5d and 5e show different retention times on chiral chromatography
and opposite optical rotation in optical characterization (Figure S3). The results together with the NMR
spectra and high-resolution mass spectrometry (HRMS) data indicate
that 5d and 5e are chirally pure enantiomers.
The same scaffold in 7a–i and 11a–b was synthesized using a different method, in which the respective
isocyanates were reacted with 2-aminothiophenes (6a–g) to the corresponding ureas that were cyclized by treatment
with MeONa to provide overall yields of 20–62%. Compounds 8a–d were constructed by the condensation
of 2-aminothiophenes with nitriles in 4 N HCl to provide the corresponding
products in yields of 32–60%. All final products were purified
with chromatography, and their structures were characterized by 1H and 13C NMR spectroscopy and liquid chromatography
high-resolution mass spectrometry (LC-HRMS) (Supporting information). The purity of all tested compounds is >95%
as
determined by high-performance liquid chromatography (HPLC).
Scheme 1
Synthesis
of Thiophene Derivatives
Reagents and conditions: a. (i)
S8, TEA, EtOH, reflux; b. (i) SnCl4, TEA, THF;
(ii) S8, EtOH, TEA, reflux; c. Acyl chloride, pyridine, N,N-dimethylformamide (DMF), rt; d. CDI, CH2Cl2.
Scheme 3
Synthesis of Thiophene
Derivatives
Reagents and conditions: a. Ethyl
cyanoacetate, S8, TEA, EtOH, reflux; b. (i) Ethyl cyanoacetate,
acetic acid, toluene, reflux; (ii) S8, EtOH, reflux; c.
CDI, CH2Cl2.
Synthesis
of Thiophene Derivatives
Reagents and conditions: a. (i)
S8, TEA, EtOH, reflux; b. (i) SnCl4, TEA, THF;
(ii) S8, EtOH, TEA, reflux; c. Acyl chloride, pyridine, N,N-dimethylformamide (DMF), rt; d. CDI, CH2Cl2.
Synthesis of Thiophene Derivatives
Reagents and conditions: a. (i)
Ammonium acetate, AcOH, toluene, reflux; (ii) S8, TEA,
EtOH, reflux; b. (i) R-NCO, pyridine, reflux; (ii) 30% MeONa, reflux;
c. R-CN, HCl, dioxane, reflux.
Synthesis of Thiophene
Derivatives
Reagents and conditions: a. Ethyl
cyanoacetate, S8, TEA, EtOH, reflux; b. (i) Ethyl cyanoacetate,
acetic acid, toluene, reflux; (ii) S8, EtOH, reflux; c.
CDI, CH2Cl2.
Structure–Activity
Relationships
The inhibitory
potency toward MIF2 tautomerase activity of this focused compound
collection was evaluated using the MIF2 tautomerase activity assay
to provide IC50 values as shown in Tables –4. The positive control compound 4-CPPC exhibits an IC50 of 47 ± 7.2 μM, which is in line with the result
of 27 μM from the literature.[30] Compared
with inhibitor 3a (R110) with an IC50 of 15
± 0.8 μM, the bromo-substituted analogue 3b exhibits enhanced potency with an IC50 value of 7.2 ±
0.6 μM. Neither adding a butyl group to the 4-position of thiophene
(3d) nor elongating the butyl to an octyl (3f) influenced the activity of R110 (3a). In contrast,
replacement of butyl group with phenethyl (3c), (tetrahydrofuran-3-yl)methyl
(3e), ethoxyethyl (3g), or 4-chlorophenethyl
(3h) diminished the MIF2 inhibitory potency. Amidation
of the 2-amino group with different acyls (4a–c) also failed to increase the potency. However, the derivatives with
a CF3 group and a naphthalene substitution (3i–k) showed increased activity, with an IC50 of 2.6 ±
0.2 μM for the most potent compound 3i.
Table 2
Potency of MIF2 Tautomerase Inhibition
by Thiophene Derivatives Determined Using the MIF2-Catalyzed PP Conversion
Assaya
# (n = 3, values are shown
as IC50 ± SD).
Table 4
Potency of MIF2 Tautomerase Inhibition
by Thieno[2,3-d]pyrimidin-4(1H)-ones
Determined Using MIF2-Catalyzed PP Conversion Assaya
# (n = 3,
values are shown as IC50 ± SD).
# (n = 3, values are shown
as IC50 ± SD).# (n = 3, values are shown
as IC50 ± SD).# (n = 3,
values are shown as IC50 ± SD).To investigate the importance
of the thiophene core of inhibitors,
we changed the thiophene to thieno[2,3-d]pyrimidine-2,4(1H,3H)-diones using ring closure strategies
to bridge the 2-amino and 3-amide functionalities. The activities
of 5a and 5b, which have aliphatic chains
at R3 position, are similar to R110 (3a).
By analyzing the SARs of 7a–d, we
observed that a chlorophenyl group, compared with a phenyl group,
at the R2 position and a benzyl group at the R3 position is beneficial for MIF2 inhibitory potency within this series
of inhibitors. Thus, 7d, which has an IC50 of 5.1 ± 0.5 μM, was employed as a new starting point
for further exploring the SARs. First, a CF3 group installed
on the para-position of the benzyl group afforded 7e, which shows similar potency to 7d. Subsequently,
we moved the phenyl or 4-chlorophenyl group to the neighboring position
to prepare 11a and 11b, which both show
impaired activity. Therefore, we started again from 7d to explore the space around the 4′-chlorophenyl group. Interestingly,
all four patterns of substitutions (7f–i) on the
phenyl ring provide increased potency. Among them, the 3′-CF3 substitution (7h) results in a 3-fold improvement
of potency with an IC50 of 1.7 ± 0.1 μM. By
replacing the benzyl group of 7h with a (naphthalen-1-yl)methyl
group, we achieved 5c, which affords another a 2-fold
increase of potency on MIF2 tautomerase inhibition with an IC50 of 0.8 ± 0.1 μM. However, the solubility of 5c is only 3.3 μg/mL (7.4 μM), which would be
a limitation for further application in cell-based assays. Therefore,
we synthesized a pair of methyl-branched enantiomers (5d and 5e) to break the planarity of 5c and
to improve water solubility.[38] Favorably,
both 5d and 5e have improved solubility
in aqueous solution with saturated concentrations of 16 μg/mL
(36 μM) and 15 μg/mL (33 μM) for 5d and 5e, respectively. Moreover, 5d exhibits
a comparable potency as 5c, while 5e is
less active. This identified 5d as one of the most potent
inhibitors of MIF2 tautomerase with a water solubility that is sufficient
for further studies.In addition, a scaffold hopping from thieno[2,3-d]pyrimidine-2,4(1H,3H)-dione (7d) to thieno[2,3-d]pyrimidin-4(1H)-one (8b) led to a decrease in potency with
an IC50 of 15 ± 1.4 μM (Table ). Adding a methoxy group to the meta-position
of benzyl (8d) increases the inhibitory potency of 8b by 3 times to an IC50 of 4.3 ± 0.4 μM,
but modifying the benzyl to chloromethyl group (8a) or
Cl to F (8c) failed to improve the potency of compound 8b.
Characterization of 5d
We next confirmed
that 5d inhibits MIF2 through a direct interaction by
monitoring MIF2 thermal stability using the thermal shift assay. Incubation
with 5d stabilizes MIF2 from heat-induced unfolding and
increased its melting temperature in a dose-dependent manner over
the DMSO control (Figure A). The melting temperature of MIF2 increases by 8.5 °C
in the presence of 50 μM 5d, indicating a binding
between 5d and MIF2. Subsequently, 5d was
subjected to enzyme kinetic analysis for MIF2 inhibition. The Lineweaver–Burk
plot shows that the Vmax is unaffected
by 5d, while KM increases
with the increase of 5d concentration (Figure B), indicating that 5d binds in competition with the tautomerase substrate to the active
site. Moreover, we also interrogated the selectivity of 5d toward MIF2 tautomerase activity by testing its inhibitory potential
for MIF tautomerase activity. 5d provided less than 50%
inhibition of MIF tautomerase activity at 50 μM, which indicates
a more than 50-fold difference in potency (Figure C). Taken together, 5d is a
potent, competitive, and selective binder of MIF2.
Figure 6
Characterization of 5d. (A) Fluorescence-based thermal
shift assay for the binding of 5d to MIF2. The melting
temperature shows a concentration-dependent shift to a higher temperature
in the presence of 5d. (B) Lineweaver–Burk plot
representation. (C) Incubation of MIF with micromolar concentrations
of 5d provided less than 50% inhibition compared to the
vehicle control. The data shown are the average of triplicate samples
with SD. (n = 3, *p < 0.05, **p < 0.01 and ***p < 0.001 vs control).
Characterization of 5d. (A) Fluorescence-based thermal
shift assay for the binding of 5d to MIF2. The melting
temperature shows a concentration-dependent shift to a higher temperature
in the presence of 5d. (B) Lineweaver–Burk plot
representation. (C) Incubation of MIF with micromolar concentrations
of 5d provided less than 50% inhibition compared to the
vehicle control. The data shown are the average of triplicate samples
with SD. (n = 3, *p < 0.05, **p < 0.01 and ***p < 0.001 vs control).
Molecular Modeling of the
Interaction between 5d and MIF2
To gain insight
into the structure–activity
relationships, a modeling study was done for the binding of inhibitor 5d to MIF2. 5d was docked into the tautomerase
active site of the crystal structure of MIF2 (PDB code: 6c5f),[32] and energy was minimized (Figure ). The five highest-scoring poses of 5d occupy similar positions. In this model, the π–π
stacking between Phe2 and the phenyl group of R110 (3a) is preserved for 5d. Furthermore, 5d and
Ser50 interact through a hydrogen bond, which is also found between
R110 (3a) and Lys109. Instead, 5d has two
π–cation interactions with Lys109. Additionally, the
CF3 group of 5d has several interactions with
MIF2 via Phe2, Arg98, and Leu102. Besides, Pro1 and Leu117 are two
newly involved interactive residues that can contribute to 5d-MIF2 binding. These newly formed interactions could explain the
increased potency of 5d compared with R110 (3a). Taken together, the putative binding mode rationalizes the binding
of 5d to MIF2 and also sheds light on the increased binding
affinity of 5d.
Figure 7
Schematic representation of the binding mode
of 5d (thick sticks) with MIF2 (PDB code: 6c5f).[32] The residues
of the binding pocket are shown as thin sticks. The hydrogen bond
formed with Ser50 is shown as the red dashed line. The cation−π
interactions formed with Lys109 are shown as the blue dash line. Discovery
Studio was employed for docking and PyMOL was used for graph preparation.
Schematic representation of the binding mode
of 5d (thick sticks) with MIF2 (PDB code: 6c5f).[32] The residues
of the binding pocket are shown as thin sticks. The hydrogen bond
formed with Ser50 is shown as the red dashed line. The cation−π
interactions formed with Lys109 are shown as the blue dash line. Discovery
Studio was employed for docking and PyMOL was used for graph preparation.
Antiproliferative Activity of 5d
After
the identification of 5d as a potent MIF2 tautomerase
inhibitor, we tested its effect on the proliferation of non-small
cell lung cancer (NSCLC) cells. First, we investigated the toxicity
of 5d using a cell viability assay (MTS assay). The results
(Figure S4) indicated that 5d did not show cell viability inhibition up to a concentration of
10 μM in A549, H1650, H1299, and HCC827 cells upon 24 h of exposure.
To evaluate the antiproliferative effects of inhibitor 5d, we treated the cells with 5d in different concentrations
for 72 h before quantifying cell numbers by measuring the DNA content
using a CyQUANT assay. The results indicate that 5d inhibited
the growth of several types of NSCLC cell lines dose-dependently (Figure ). 5d showed a visible inhibitory effect at 1 μM and reached about
90% inhibition of cell growth at 7.5 μM on A549 cells, showing
an IC50 of 3.0 μM. 5d also inhibited
the proliferation of H1650, H1299, and HCC827 cells with IC50 values of 5.3, 7.6, and 5.2 μM, respectively. Taken together,
these results demonstrate that 5d inhibits the proliferation
of several types of NSCLC cell lines.
Figure 8
Effect of inhibitor 5d on
the proliferation of the
NSCLC cells lines A549, H1650, H1299, and HCC827. The cells were seeded
at a density of about 1000 cells per well in a 96-well plate. After
overnight culturing, the cells were treated with various concentrations
of 5d for 72 h. Afterward, relative cell numbers were
determined by CyQUANT cell proliferation assays and compared with
a DMSO-treated control. The data shown are the average of three experiments
with SD (n = 3).
Effect of inhibitor 5d on
the proliferation of the
NSCLC cells lines A549, H1650, H1299, and HCC827. The cells were seeded
at a density of about 1000 cells per well in a 96-well plate. After
overnight culturing, the cells were treated with various concentrations
of 5d for 72 h. Afterward, relative cell numbers were
determined by CyQUANT cell proliferation assays and compared with
a DMSO-treated control. The data shown are the average of three experiments
with SD (n = 3).The inhibitory effect of 5d on the proliferation of
NSCLC cell lines was further evaluated using a clonogenic assay. The
results of this colony formation assay showed that 5d potently inhibited cell growth of the four NSCLC cell lines tested.
We observed a dose-dependent reduction in both colony number and size
in 5d-treated cells compared to vehicle-treated controls
(Figure A). The number
of colonies in each well was quantified by the absorption of crystal
violet. 5d suppressed colony formation in a dose-dependent
manner by around 90% in all four NSCLC cell lines tested at a concentration
of 10 μM (Figure B). These results confirm the efficacy of 5d on inhibition
of NSCLC cell proliferation.
Figure 9
Treatment with inhibitor 5d inhibits
colony formation
of NSCLC cell lines. (A) A549 or other NSCLC cell lines were seeded
in a 12-well plate (1000 cells/well). The cells were preincubated
overnight and were then treated with compounds or vehicle (DMSO) for
5 days. Afterward, the cells were fixed and stained with 0.5% crystal
violet solution. Image of representative wells were scanned and are
shown. (B) Stained colonies were dissolved in 30% acetic acid and
then quantified by measuring the absorbance at 590 nm. The relative
colony number is normalized to the DMSO-treated control. The data
shown are the average of three experiments with SD (n = 3, *p < 0.05, **p < 0.01
and ***p < 0.001 vs control).
Treatment with inhibitor 5d inhibits
colony formation
of NSCLC cell lines. (A) A549 or other NSCLC cell lines were seeded
in a 12-well plate (1000 cells/well). The cells were preincubated
overnight and were then treated with compounds or vehicle (DMSO) for
5 days. Afterward, the cells were fixed and stained with 0.5% crystal
violet solution. Image of representative wells were scanned and are
shown. (B) Stained colonies were dissolved in 30% acetic acid and
then quantified by measuring the absorbance at 590 nm. The relative
colony number is normalized to the DMSO-treated control. The data
shown are the average of three experiments with SD (n = 3, *p < 0.05, **p < 0.01
and ***p < 0.001 vs control).To gain insights into the effect of longer-term 5d treatment in a more elaborate model, we employed a 3D spheroid
tumor
growth model, which was established using A549 cancer cells,[39] with each spheroid containing about 1000 A549
cells. After 5-day culturing, different concentrations of 5d were used to treat these spheroids with 72 h intervals over 12 days.
The diameters of spheroid were monitored and compared to day 0 of
the treatment. Treatment of 5d significantly reduced
the size of tumor spheroids (Figure ). With exposure to 1, 2, and 5 μM concentrations
of 5d for 12 days, the volume of spheroid was reduced
by 40, 63, and 79%, respectively. These results indicate that the
MIF2 tautomerase inhibitor 5d effectively reduces the
growth of A549 cancer cells in a 3D tumor model.
Figure 10
5d treatment
reduces the growth of A549 cancer cells
in a spheroid model. A549 cells were seeded in ultralow attachment
96-well U bottom plate (1000 cells/well) to generate tumor spheroids
(a single spheroid per well). After initiation, the spheroids were
treated with 5d at the indicated concentrations every
3 days. DMSO was used as a vehicle control. The day of the first treatment
was indicated as day 0. (A) Representative images were obtained at
the indicated intervals using an inverted microscope. Scale bar: 500
μm. (B) Analysis was carried out using NIS-Elements AR 3.1 software,
and growth curves were obtained relative to the volume of untreated
spheroids (day 0) and plotted with GraphPad Prism8. Values are displayed
as mean ± SD (n = 3 spheroids for each time
point, *p < 0.05, **p < 0.01
and ***p < 0.001 vs control).
5d treatment
reduces the growth of A549 cancer cells
in a spheroid model. A549 cells were seeded in ultralow attachment
96-well U bottom plate (1000 cells/well) to generate tumor spheroids
(a single spheroid per well). After initiation, the spheroids were
treated with 5d at the indicated concentrations every
3 days. DMSO was used as a vehicle control. The day of the first treatment
was indicated as day 0. (A) Representative images were obtained at
the indicated intervals using an inverted microscope. Scale bar: 500
μm. (B) Analysis was carried out using NIS-Elements AR 3.1 software,
and growth curves were obtained relative to the volume of untreated
spheroids (day 0) and plotted with GraphPad Prism8. Values are displayed
as mean ± SD (n = 3 spheroids for each time
point, *p < 0.05, **p < 0.01
and ***p < 0.001 vs control).
5d Arrests the Cell Cycle of
A549 Cells at the G0/G1 Phase
The cell cycle progression of 5d-treated cancer cells
was investigated using flow cytometry. A549 cells were exposed to
different concentrations of 5d for a duration of 2 days
before analysis. The data show that 5d induced cell cycle
arrest at the G0/G1 phase dose-dependently (Figure ). The percentage of A549 cells in the G0/G1 phases was
56% for the control group. Upon treatment with 5, 7.5, and 10 μM 5d, this percentage increased to 58, 63, and 67%, respectively.
This result indicates that 5d inhibited cell cycle progression,
which provided an explanation for the observed inhibition of cancer
cell growth.
Figure 11
Treatment with 5d induces cell cycle arrest
in A549
cells. (A) A549 cells were treated with 5d at the indicated
concentrations for 48 h. Representative cell cycle distributions graphs
obtained using propidium iodide staining-based flow cytometry. (B)
Bar graph representatives of the distribution of cells in each phase
of the cell cycle (G0/G1, S, and G2/M phases) were analyzed using FlowJo. Results are
displayed as mean ± SD (n = 3, t-test analysis was performed between G2/M phase of treated
groups and control group. *p < 0.05 and **p < 0.01 vs control.).
Treatment with 5d induces cell cycle arrest
in A549
cells. (A) A549 cells were treated with 5d at the indicated
concentrations for 48 h. Representative cell cycle distributions graphs
obtained using propidium iodide staining-based flow cytometry. (B)
Bar graph representatives of the distribution of cells in each phase
of the cell cycle (G0/G1, S, and G2/M phases) were analyzed using FlowJo. Results are
displayed as mean ± SD (n = 3, t-test analysis was performed between G2/M phase of treated
groups and control group. *p < 0.05 and **p < 0.01 vs control.).
5d Inhibits ERK Phosphorylation
The effect
of 5d treatment on MIF2-related signaling pathways was
studied by assessing MIF2-induced ERK phosphorylation. Towards this
aim, MIF2 or 5d preincubated MIF2 was employed to stimulate
A549 cells for 15 min; subsequently, ERK phosphorylation of treated
cells was analyzed using western blot detection. MIF2 treatment stimulated
ERK phosphorylation in A549 cells to about 4.5-fold of the control.
Treatment with 5d attenuated this stimulation in a dose-dependent
manner (Figure ).
These data demonstrate that 5d treatment inhibits ERK
phosphorylation in response to MIF2-stimulation.
Figure 12
Inhibition of MIF2-induced
ERK phosphorylation in A549 cells upon
treatment with 5d. (A) A549 cells were stimulated with
recombinant MIF2 with or without 5d preincubation at
indicated concentrations for 15 min; the concentrations of pERK and
total ERK were examined by immunoblots with anti-pERK, anti-ERK using
glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as a loading control.
A representative western blot is shown (n = 2). (B)
Relative pERK level was quantified by calculating the pERK:ERK ratio
and normalized to the control group. Data are displayed as the mean
of two experiments.
Inhibition of MIF2-induced
ERK phosphorylation in A549 cells upon
treatment with 5d. (A) A549 cells were stimulated with
recombinant MIF2 with or without 5d preincubation at
indicated concentrations for 15 min; the concentrations of pERK and
total ERK were examined by immunoblots with anti-pERK, anti-ERK using
glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as a loading control.
A representative western blot is shown (n = 2). (B)
Relative pERK level was quantified by calculating the pERK:ERK ratio
and normalized to the control group. Data are displayed as the mean
of two experiments.
Discussion and Conclusions
The overexpression of MIF family proteins, MIF and MIF2, has been
implicated in several types of cancer.[40,41] Therefore,
MIF and MIF2 are promising drug targets for cancer treatment. A significant
effort has been invested into the development of inhibitory strategies
to target MIF.[42] Nevertheless, the drug
discovery efforts to target MIF2 are lagging behind. Thus, the aim
of this study was to develop potent small-molecule MIF2 inhibitors
that could be applied in cancer therapy.To discover a potent
small-molecule inhibitor to target MIF2, we
set up a MIF2-targeted drug discovery campaign, which includes library
screening, hit optimization, and a relevant cellular model. Toward
this aim, we firstly optimized the 4-HPP based enzyme assay. Although
MIF2 possesses an enzymatic activity on catalyzing the keto-enol tautomerization
of 4-HPP, its catalytic efficiency is much lower than that of MIF.[17] The structural information shows that the differences
inside the enzyme active sites of MIF and MIF2 not only contribute
to their unique substrate preference but also facilitate the possibility
to develop selective inhibitors such as ISO-1, which is a representative
MIF inhibitor that does not inhibit MIF2.[19] Therefore, we set up a MIF2-based tautomerization assay using phenylpyruvate
as a substrate. This PP-based assay was employed to screen an in-house
library, which afforded several compounds with IC50’s
in the low micromolar range. 3a was selected as the hit
for further optimization on the basis of its potency and druglike
properties.The putative binding mode of 3a with
MIF2 reveals
that the chlorophenyl group of 3a is embedded in the
pocket and the butyl group is exposed to water. There is vacancy around
the chloro and thiophene groups, as well as a water-exposed rim. To
exploit the unoccupied space in the binding site, a focused collection
of thiophenes was designed and prepared and MIF2 binding was assessed
using a PP-tautomerization assay. We observed that the potency of 3a analogues can be tuned by different substituents in the
thiophene 3-position, and a (naphthalen-1-yl)methyl group proved to
be the favorable modification in this series of derivatives. Ring
closure of the respective 2-aminothiophene precursor afforded the
corresponding thienopyrimidione 5c, which shows an IC50 of 0.81 μM. Thus, inhibitor 5c gained
50-fold in potency compared to the previously identified MIF2 inhibitor
4-CPPC that proved to have an IC50 of 47 μM in our
assay format.[30] To improve the solubility
of 5c, we disrupted the planarity of 5c by
introducing a methyl group to provide 5d,[38] which has improved solubility and preserved
MIF2 binding potency. Notably, 5d exhibits weak inhibition
on MIF thus demonstrating a more than 50-fold selectivity for MIF2
compared to MIF. Taken together, we obtained 5d as a
selective MIF2 tautomerase inhibitor with submicromolar potency and
optimal solubility, which enable its use in cellular model systems.The effect of 5d on proliferation was evaluated using
cell culture assays on several NSCLC cell lines including A549, H1299,
HCC827, and H1299. 5d showed more than 80% inhibition
of the proliferation of all four types of tested cancer cells at 10
μM in both a DNA content-based assay and a clonogenic assay.
The effectiveness of 5d was also confirmed on a spheroid
cell culture assay, which is able to mimic some of the main features
of solid human tumors, such as their cellular layered assembly, hypoxia
state, structural organization, and nutrient gradients.[43] In this assay, the spheroid volume growth was
reduced by 63% upon treatment with 2 μM 5d. The
data demonstrate that 5d treatment provides a strong
inhibition of cell proliferation, which is in line with the previously
reported antiproliferation effect of siRNA-mediated MIF2 silencing.[41,44] We also noticed, compared with our previous studies on MIF tautomerase
inhibitors,[15,35] inhibition of MIF2 tautomerase
activity in this study yielded a more substantial impact on cancer
cell proliferation. This observation is in line with previous results
that MIF2-silencing siRNA inhibited cell growth more potently than
MIF-silencing siRNA.[41] Taken together,
our results indicate that MIF2 inhibitor 5d inhibits
the cell proliferation of several model systems, which is in line
with previous siRNA-based studies on similar model systems.[41,44]A main feature of the growth of cancer cells is their continuous
and ordered progression of the cell cycle.[45] Treatment of A549 cancer cells with 5d arrested cells
in the G0/G1 phase, indicating interference of cell cycle progression. MIF2 is
a growth factor that is involved in the modulation of cell cycle progression
through the MAPK pathway.[41,46] We also observed that 5d treatment attenuates the MAPK signaling as we found less
MIF2-induced ERK phosphorylation. The 5d-mediated MAPK
pathway deactivation is also in line with the effect of siRNA-mediated
downregulation of MIF2 protein levels.[41] It is also consistent with the inhibition of ERK phosphorylation
by MIF2 inhibitor 4-CPPC.[30] Collectively,
these data show that the growth inhibition of cancer cell proliferation
by 5d treatment can be explained by ERK phosphorylation
blockade and cell cycle progression arrest.In conclusion, screening
of focused compound collections and structure-guided
hit optimization enabled the discovery of the potent MIF2 inhibitor 5d. This inhibitor has submicromolar potency toward MIF2 tautomerase
activity, whereas MIF tautomerase activity is not affected. 5d binds in competition with the MIF2 tautomerase substrate,
and its binding was verified by the thermal shift assay in which 5d improves the MIF2 thermal stability. Furthermore, we have
demonstrated that 5d-mediated inhibition of MIF2 activity
suppresses the growth of NSCLC cells in 2D and 3D cell cultures, which
can be explained by the inhibition of the MAPK signaling and subsequent
cell cycle arrest, thus validating MIF2 as a target with therapeutic
potential for NSCLC patients. Taken together, the thienopyrimidione
reported here represents a novel molecule for the inhibition of MIF2
with a clear potential for use in cellular model systems.
Experimental Section
General
All of the chemicals and
solvents were purchased
from Sigma-Aldrich, AK Scientific, Fluorochem, or Acros Organics and
were used without further processing unless stated otherwise. Thin-layer
chromatography (TLC) of Merck silica gel 60 F254 plates
was used for reaction monitoring. Column chromatography was conducted
using MP Ecochrom silica 32–63 (60 Å). 1H NMR
(500 MHz) and 13C NMR (126 MHz) were recorded on a Bruker
Avance 500 spectrometer. 1H and 13C NMR spectra
were reported in parts per million (ppm) referenced to deuterated
solvents, for example, CDCl3: δ = 7.26 ppm (1H) and 77.05 ppm (13C) or DMSO-d6: δ = 2.50 ppm (1H) and 39.52 ppm (13C). To report spin multiplicity, the following abbreviations
were used: s (singlet), d (doublet), dd (doublet of doublets), t (triplet),
q (quartet), and m (multiplet). Coupling constants were reported in
hertz (Hz). Fourier transform mass spectrometry (FTMS) and electrospray
ionization (ESI) on an Applied Biosystems/SCIEX API3000-triple quadrupole
mass spectrometer were applied for high-resolution mass spectra. Purity
of the compounds was confirmed to be >95% by C18 HPLC analysis.
The
analogues of 4-HPP were synthesized by following a published method.[47]
The starting materials 1a–e and 2a–j were either purchased or prepared using
our previously published methods.[48] To
synthesize R110 (3a), 2-(4-chlorophenyl)acetaldehyde
(1a, 0.32 g, 2.0 mmol) in EtOH (5 mL) was added into
a solution of N-butyl-2-cyanoacetamide (2a, 0.31 g, 2.1 mmol). To the resulting suspension, triethylamine (0.30
mL, 2.1 mmol) was added, followed by the addition of S8 (80 mg, 0.60 mmol). Then, the reaction mixture was refluxed overnight.
After cooled down to room temperature, the mixture was diluted using
EtOAc (50 mL) and washed with water (2 × 30 mL) and brine (2
× 30 mL). The EtOAc layer was collected and dried over MgSO4. After filtration, the solvent was collected and removed
under reduced pressure and purification was conducted using flash
chromatography, with petroleum ether/EtOAc 5:1 (v/v) as eluent. The
product (260 mg) was obtained as a light brown solid, yield 42%. 1H NMR (500 MHz, DMSO-d6) δ
7.70 (t, J = 5.6 Hz, 1H), 7.64 (s, 1H), 7.50 (s,
2H), 7.43–7.35 (m, 4H), 3.18 (q, J = 6.9 Hz,
2H), 1.47 (p, J = 7.3 Hz, 2H), 1.35–1.29 (m,
2H), 0.90 (t, J = 7.3 Hz, 3H). 13C NMR
(126 MHz, DMSO-d6) δ 165.15, 161.12,
133.28, 129.92, 128.94, 125.09, 121.76, 119.86, 108.33, 38.04, 31.58,
19.71, 13.80. HRMS, calculated for C15H18ON2ClS [M + H]+: 309.0823, found 309.0824.
1-(4-Chlorophenyl)hexan-2-one
(1c, 0.20 g, 1.0 mmol) and 2-(4-chlorophenyl)acetaldehyde
(1a, 0.15 g, 1.0 mmol) were dissolved in dry THF (5 mL).
To the stirred solution, SnCl4 (0.22 mL, 2.0 mmol) was
added dropwisely. Then, Et3N (0.30 mL) was added. The reaction
mixture was stirred at 40 °C for 16 h, followed by the addition
of HCl solution (1 N, 25 mL) and extraction with EtOAc (3 × 20
mL). The combined organic layers were washed with a NaOH solution
(1 N, 25 mL) and dried over MgSO4. The solid was filtrated,
and the solvent was removed using evaporation under reduced pressure.
The resulting mixture was dissolved in EtOH (5 mL), and S8 (32 mg, 0.25 mmol) and Et3N (0.2 mL) were added. The
reaction mixture was refluxed overnight. Then, the mixture was diluted
with EtOAc (25 mL) and washed with water (2 × 50 mL) and brine
(2 × 50 mL). The organic layers were dried over MgSO4, filtrated, and the solvent was removed using evaporation under
reduced pressure. The product was obtained as a brown solid (151 mg,
yield 41%) after chromatography using petroleum ether:EtOAc 30:1 (v/v)
as eluent. 1H NMR (500 MHz, DMSO-d6) δ 7.64 (t, J = 5.6 Hz, 1H), 7.45
(d, J = 8.5 Hz, 2H), 7.28 (d, J =
8.5 Hz, 2H), 6.21 (s, 2H), 3.19 (q, J = 6.8 Hz, 2H),
2.66–2.58 (m, 2H), 1.49–1.43 (m, 2H), 1.33 (dt, J = 14.9, 7.2 Hz, 4H), 1.16 (q, J = 7.3
Hz, 2H), 0.89 (t, J = 7.3 Hz, 3H), 0.76 (t, J = 7.3 Hz, 3H). 13C NMR (126 MHz, DMSO-d6) δ 165.47, 155.55, 135.69, 133.52, 131.11,
130.51, 128.70, 117.17, 113.51, 38.36, 32.22, 31.39, 27.12, 22.10,
19.79, 13.79, 13.65. HRMS, calculated for C19H26ON2ClS [M + H]+: 365.1449, found 365.1449.
To
a solution of 2-amino-N-butyl-5-phenylthiophene-3-carboxamide
(0.3g, 1.0 mmol) in DCM (10
mL), 1,1-carbonyldiimidazole (CDI, 0.50 g, 3.1 mmol) was added, and
the solution was refluxed for 16 h. Subsequently, the solvent was
removed under reduce pressure and EtOAc (25 mL) was added to dissolve
the residual mixture. The organic solution was then washed with water
(25 mL) and brine (25 mL). The EtOAc layer was collected and dried
over MgSO4. After the removal of MgSO4 by filtration,
the solvent was removed using evaporation under reduced pressure.
The purification was done using chromatography with petroleum ether:EtOAc
3:1 (v/v) as an eluent, and a 159 mg light yellow solid was obtained
as the product. Yield 52%. 1H NMR (500 MHz, DMSO-d6) δ 7.66 (d, J = 7.4
Hz, 2H), 7.58 (s, 1H), 7.41 (t, J = 7.7 Hz, 2H),
7.32 (t, J = 7.4 Hz, 1H), 3.96–3.73 (m, 2H),
1.53 (p, J = 7.5 Hz, 2H), 1.30 (h, J = 7.4 Hz, 2H), 0.90 (t, J = 7.4 Hz, 3H). 13C NMR (126 MHz, DMSO-d6) δ 158.32,
150.26, 149.70, 133.44, 132.60, 129.25, 127.95, 125.29, 118.10, 115.60,
29.48, 22.39, 19.68, 13.76. HRMS, calculated for C16H17O2N2 [M + H]+: 301.1005,
found 301.1006.
A mixture
of phenolacetone
(1f, 0.7 mL, 5 mmol), ethyl cyanoacetate (0.5 mL, 5 mmol),
ammonium acetate (0.1 g, 1 mmol), and acetic acid (0.2 mL, 4 mmol)
in toluene (5 mL) was heated under reflux for 20 h, while water was
removed using a molecular sieve. After the mixture was cooled to room
temperature, the mixture was concentrated in vacuo. The residue was diluted with saturated NaHCO3 (20 mL)
and extracted with CHCl3 (3 × 25 mL). The extract
was washed with brine (25 mL) and dried (MgSO4). After
evaporation of the solvent in vacuo, the residue
was purified by flash column chromatography to give an oil-like intermediate,
which was dissolved in EtOH (9 mL). To the solution were added sulfur
powder (0.13 g, 4.0 mmol) and triethylamine (0.6 mL, 4 mmol), and
the resulting mixture was stirred at 100 °C for 2 h. After removal
of the solvent in vacuo, the residue was diluted
with brine and extracted with CHCl3 (3 × 25 mL). The
organic layers were collected and washed with brine (25 mL) and dried
(MgSO4). The solution was concentrated in vacuo, and the residue was recrystallized from CHCl3 to afford 6a (1.1 g, 81%) as dark brown crystals. 1H NMR
(500 MHz, DMSO-d6) δ 7.82 (t, J = 5.8 Hz, 1H), 7.46 (s, 2H), 7.40 (d, J = 8.0 Hz, 2H), 7.34 (t, J = 7.8 Hz, 2H), 4.36 (q, J = 7.0 Hz, 3H), 1.34 (t, J = 7.1 Hz, 3H).
To a stirred solution of 6a (0.3 g, 1 mmol) in pyridine
(3 mL), phenyl isocyanate (0.21 mL, 1.5 mmol) was added and stirred
at 45 °C for overnight. The resulting mixture was concentrated
reduced pressure, and the residue was suspended in MeOH (5 mL). To
the suspension, 25% sodium methoxide (0.6 mL, 3 mmol) was added and
the resulting mixture was stirred for 6 h at room temperature. Subsequently,
the mixture was acidified with 1 N HCl (10 ml) at 0 °C and the
pH was regulated to 4. The organic solvent was evaporated to a volume
of about 5 mL, and the precipitate was collected. The obtained solid
was dissolved in methanol and purified with chromatography using DCM:MeOH
(100:1) as a solvent. The product was obtained as light brown crystals
(135 mg, yield 40%). 1H NMR (500 MHz, DMSO-d6) δ 12.37 (s, 1H), 7.51–7.39 (m, 8H), 7.31–7.27
(m, 2H), 2.40 (s, 3H). 13C NMR (126 MHz, DMSO-d6) δ 159.44, 150.86, 150.15, 135.79, 132.38, 129.83,
129.22, 129.07, 128.99, 128.86, 128.71, 127.97, 126.74, 114.19, 14.15.
HRMS, calculated for C19H15O2N2S [M + H]+: 335.0849, found 335.0849.
6b (0.33 g, 1.0 mmol) and 2-chloroacetonitrile (0.1
mL, 1.2 mmol) were diluted into HCl (4 N, 6 mL) in 1,4-dioxane (2
mL) and stirred for 1 h at room temperature. Subsequently, the reaction
mixture was heated to 100 °C 12 h until precipitation formed.
After cooled down to room temperature, the reaction mixture was filtered
and the precipitate was washed with n-hexane. The
final product was purified with column chromatography with DCM:MeOH
10:1 (v/v). The product was obtained as a 195 mg white solid. Yield
60%. 1H NMR (500 MHz, DMSO-d6) δ 12.81 (s, 1H), 7.55 (m, 4H), 4.57 (s, 2H), 2.53 (s, 3H). 13C NMR (126 MHz, DMSO-d6) δ
162.8, 158.8, 153.1, 133.1, 132.4, 131.5, 131.1, 130.9, 130.0, 129.1,
128.9, 123.0, 42.5, 14.3. HRMS, calculated for C14H11ON2Cl2S [M + H]+: 324.9964,
found 324.9964.
6g (0.32 g, 1 mmol) was reacted with 2-phenylacetonitrile
(0.21 g, 1.6 mmol) following similar procedure of synthesis of 8a to afford 8d (203 mg, 53%) as a white solid. 1H NMR (500 MHz, DMSO-d6) δ
12.6 (s, 1H), 7.6–7.4 (m, 4H), 7.2 (t, J =
7.8 Hz, 1H), 7.0–6.9 (m, 1H), 6.9 (d, J =
7.8 Hz, 1H), 6.8 (ddd, J = 8.4, 2.7, 1.0 Hz, 1H),
3.9 (s, 2H), 3.7 (s, 3H), 2.5 (s, 3H) ppm. 13C NMR (126
MHz, DMSO-d6) δ 163.8, 159.3, 159.1,
156.9, 137.7, 132.8, 131.7, 131.0, 130.8, 130.7, 129.7, 129.6, 129.0,
128.8, 121.9, 121.0, 114.8, 112.2, 55.1, 54.9, 14.2. HRMS, calculated
for C21H18O2N2ClS [M +
H]+: 397.0772, found 397.0770.Synthesis of 10a was performed by following a reported method[50] (ethyl 2-amino-4-phenylthiophene-3-carboxylate, 10a). An equimolar mixture of powdered sulfur (160 mg, 5.0
mmol) and morpholine (0.50 mL) was stirred until total dissolution
of the sulfur. Then, ethyl cyanoacetate (0.60 mL, 5.0 mmol) and acetophenone
(0.60 mL, 5 mmol) were added to the reaction mixture, which was stirred
at room temperature for 18 h. After completion of the reaction, as
monitored by TLC, the crude product was chromatographed on silica
with CH2Cl2 to afford a white solid, yield 34%.
10b was synthesized by following
the same method as the preparation of 6a using 1-(4-chlorophenyl)ethan-1-one
(5 mmol) as a starting material. A white powder (250 mg) was obtained
as the product with a yield of 18%.
The detailed
procedures to produce recombinant human MIF2 were reported in our
previous publications.[46] Gene sequences
of the human MIF2 gene (Invitrogen) were adapted to bacterial expression.
After subcloning into a pET20b(+) expression vector, IPTG-induced
expression was performed in E.coli strain
BL21(DE3). MIF2 protein was overproduced overnight, and harvested
cells were resuspended and sonicated. The soluble fraction was purified
using a Q sepharose column (GE Healthcare) with a gradient of NaCl.
The fractions containing MIF2 were brought to 1.7 M ammonium sulfate
and loaded on a phenyl sepharose column (GE Healthcare) and eluted
with a gradient to 0 M ammonium sulfate in a 20 mM sodium phosphate
buffer, pH 8.0. Finally, the proteins were purified by size exclusion
chromatography on a Superdex75 column (GE Healthcare) in 20 mM sodium
phosphate buffer, pH 8.0, with an elution volume characteristic for
trimeric MIF2. The collected protein was concentrated using a VivaSpin
centrifugation column with a molecular-weight cutoff at 5000 Da (Sartorius
Stedim Biotech GmbH). Purified proteins were aliquoted, snap-frozen
in liquid nitrogen, and stored at −80 °C. The purity of
the obtained protein was tested by sodium dodecyl sulfate polyacrylamide
gel electrophoresis (SDS-PAGE) and coomassie staining.
Tautomerase
Assay
The MIF tautomerase activity and
enzyme kinetics were measured by a previously described protocol.[35] The methods to study MIF2 enzyme activity were
adapted from protocols to study MIF. Briefly, 10 μL of EDTA
(20 mM) solution in demiwater was added into 180 μL of MIF2
solution (100 nM) in boric acid buffer (435 mM, pH 6.2), followed
by the addition of 10 μL of the desired compound in DMSO or
vehicle. This mixture was homogenized and preincubated at room temperature
for 10 min. Subsequently, 50 μL of this enzyme–inhibitor
mixture was mixed with 50 μL of phenylpyruvate solution (1 mM)
in ammonium acetate buffer (50 mM, pH 6.0). The increase of UV absorbance
at 300 nm over time represents MIF2 tautomerase activity. MIF2 tautomerase
activity in the group containing blank DMSO dilution was set to 100%.
Noncatalyzed conversion of the PP without the MIF presence was set
to 0%. Data from the first 3 min were linear and used to calculate
the conversion velocities. GraphPad Prism was employed for the calculation.
Cell Culture
Four different human lung cancer cell
lines including A549 (ATCC-CCL-185), H1650 (ATCC-CRL-5883), H1299
(ATCC-CRL-5803), and HCC827 (ATCC-CRL-2868) were cultured in RPMI-1640
Medium (GibcoTM #61870-010) containing 10% (v/v) fetal bovine serum
(FBS) and 100 U/mL penicillin/streptomycin (GibcoTM#10378016) at 37
°C with 5% CO2 in humidified air.
Cell Proliferation
Assay
Cell proliferation was measured
with the CyQUANT Direct Cell Proliferation Assay Kit (Thermo Fisher,
#C35011) by following the protocol. Cells were cultured in 96-well
plates at a density of 1000 cells/well and treated with different
concentrations of 5d (0.25–10 μM) for 72
h. The cells were incubated with a detection reagent (100 μL)
for 60 min at 37 °C with 5% CO2. The fluorescence
of each well was read at 485/535 nm by a plate reader (BioTek).
Clonogenic Assay
Cells were seeded in 12-well plates
(1000 cells per well in 2 mL of the same medium used for cell proliferation
assay) and incubated overnight. Corresponding inhibitors were added
to the cell culture medium, and the cells were exposed to the treatment
for 5 days. Subsequently, the medium was carefully removed and the
cells were washed with PBS before they were fixed with 4% (v/v) paraformaldehyde
for 20 min at room temperature. After removal of the fixation reagent,
the cells were stained with 0.5% (w/v) crystal violet in the dark
for 20 min. After washing, the image of stained cancer cell colonies
in each well was photographed. To quantify the staining, 10% acetic
acid was utilized to dissolve the colonies. The absorbance at a wavelength
of 590 nm was measured to represent the relative cell number by comparing
with the DMSO-treated group.
Tumor Spheroid Assay
A549 cells
(1000 cells/well) were
seeded onto a 96-well U bottom ultralow attachment plate (Corning).
After 2 days of incubation without disturbance, the spheroid was treated
with indicated compound every 3 days. Images were captured, and the
diameter of each tumor spheroid was measured on the indicated days
post-treatment using an inverted microscope (Nikon Eclipse Ti) connected
with NIS-Elements software. The data were analyzed and plotted with
GraphPad Prism8.
ERK Phosphorylation Study[15]
A549 cells were seeded into each well of a six-well
plate in a density
of 3 × 105 cells per well with 2 mL of RPMI-1640 medium
containing 0.5% (v/v) FBS (Costar Europe, Badhoevedorp, The Netherlands),
and 1% (v/v) penicillin/streptomycin solution (Corning). After overnight
culturing, the cells were stimulated with MIF2 (100 ng/mL in FBS-free
medium) or a mixture of MIF2 and different concentrations of 5d for 15 min. Subsequently, the cells were washed with cold
PBS and lysed using RIPA buffer containing PhosSTOP and protease inhibitor
(PI) cocktail (Roche, Mannheim, Germany). After determination of the
protein concentration of each sample using the BCA Protein Assay Kit
(Pierce, Rockford IL), 20 μg protein was loaded onto and separated
by a precast 10% NuPAGE Bis-Tris gel (Invitrogen). The proteins in
the gel were then transferred to a poly(vinylidene difluoride) (PVDF)
membrane. After blocking with 5% (w/v) of skimmed milk for 1 h at
room temperature and incubation with the appropriate primary antibody
(pERK, #9101, Cell Signaling, 1:1000; ERK, #9102, Cell Signaling,
1:1000; GAPDH, #97166, Cell Signaling, 1:10 000) overnight
at 4 °C, the membrane was treated with HRP-conjugated secondary
antibodies and the protein bands were visualized with enhanced chemiluminescence
(ECL) solution (GE Healthcare)and quantified with ImageJ software
based on grayscale.
Flow Cytometry
For cell cycle analysis,
A549 cells
were seeded in six-well plates at a density of 1 × 105 per well. The next day, the cells were treated with 5d or vehicle for 48 h. After the treatment, the cells were washed
with PBS (3×) and then harvested after trypsinization. After
centrifugation at 300g, the cells were incubated
with a solution containing 20 μg/mL propidium iodide (PI) (Sigma,
P4864) and 0.1% (v/v) Triton-X100 (Sigma, T8787) for 15 min at room
temperature. Fluorescence was detected by a Cytoflex flow cytometer
(Beckman Coulter, Woerden, the Netherlands) immediately. A total of
30 000 cells were collected for each sample. Data were analyzed
using FlowJo software (Tree Start, Ashland).
Table 3
Potency of MIF2 Tautomerase Inhibition
by Thieno[2,3-d]pyrimidine-2,4(1H,3H)-diones Determined Using MIF2-Catalyzed PP Conversion
Assaya
Authors: Katherine L Meyer-Siegler; Kenneth A Iczkowski; Lin Leng; Richard Bucala; Pedro L Vera Journal: J Immunol Date: 2006-12-15 Impact factor: 5.422
Authors: Pawel Dziedzic; José A Cisneros; Michael J Robertson; Alissa A Hare; Nadia E Danford; Richard H G Baxter; William L Jorgensen Journal: J Am Chem Soc Date: 2015-02-20 Impact factor: 15.419
Authors: Christian Weber; Sandra Kraemer; Maik Drechsler; Hongqi Lue; Rory R Koenen; Aphrodite Kapurniotu; Alma Zernecke; Jürgen Bernhagen Journal: Proc Natl Acad Sci U S A Date: 2008-10-13 Impact factor: 11.205
Authors: Georgios Pantouris; Mansoor Ali Syed; Chengpeng Fan; Deepa Rajasekaran; Thomas Yoonsang Cho; Eric M Rosenberg; Richard Bucala; Vineet Bhandari; Elias J Lolis Journal: Chem Biol Date: 2015-09-10
Authors: Maria Vinci; Sharon Gowan; Frances Boxall; Lisa Patterson; Miriam Zimmermann; William Court; Cara Lomas; Marta Mendiola; David Hardisson; Suzanne A Eccles Journal: BMC Biol Date: 2012-03-22 Impact factor: 7.431
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Scheme 1
Scheme 3
Figure 6
Figure 7
Figure 8
Figure 9
Figure 10
Figure 11
Figure 12