Sarah Huff1, Indrasena Reddy Kummetha1, Lingzhi Zhang1, Lingling Wang1, William Bray1, Jiekai Yin2, Vanessa Kelley1, Yinsheng Wang2, Tariq M Rana1,3. 1. Division of Genetics, Department of Pediatrics, Center for Drug Discovery Innovation, Program in Immunology, Institute for Genomic Medicine, University of California San Diego, 9500 Gilman Drive MC 0762, La Jolla, California 92093, United States. 2. Environmental Toxicology Graduate Program and Department of Chemistry, University of California, Riverside, California 92521, United States. 3. San Diego Center for Precision Immunotherapy, Moores Cancer Center 3855 Health Sciences Drive, University of California San Diego, La Jolla, California 92093, United States.
Abstract
Aberrant regulation of N6-methyladenosine (m6A) RNA modification has been implicated in the progression of multiple diseases, including cancer. Previously, we identified a small molecule inhibitor of the m6A demethylase fat mass- and obesity-associated protein (FTO), which removes both m6A and N6,2'-O-dimethyladenosine (m6Am) RNA modifications. In this work, we describe the rational design and optimization of a new class of FTO inhibitors derived from our previous lead FTO-04 with nanomolar potency and high selectivity against the homologous m6A RNA demethylase ALKBH5. The oxetanyl class of compounds comprise competitive inhibitors of FTO with potent antiproliferative effects in glioblastoma, acute myeloid leukemia, and gastric cancer models where lead FTO-43 demonstrated potency comparable to clinical chemotherapeutic 5-fluorouracil. Furthermore, FTO-43 increased m6A and m6Am levels in a manner comparable to FTO knockdown in gastric cancer cells and regulated Wnt/PI3K-Akt signaling pathways. The oxetanyl class contains significantly improved anticancer agents with a variety of applications beyond glioblastoma.
Aberrant regulation of N6-methyladenosine (m6A) RNA modification has been implicated in the progression of multiple diseases, including cancer. Previously, we identified a small molecule inhibitor of the m6A demethylase fat mass- and obesity-associated protein (FTO), which removes both m6A and N6,2'-O-dimethyladenosine (m6Am) RNA modifications. In this work, we describe the rational design and optimization of a new class of FTO inhibitors derived from our previous lead FTO-04 with nanomolar potency and high selectivity against the homologous m6A RNA demethylase ALKBH5. The oxetanyl class of compounds comprise competitive inhibitors of FTO with potent antiproliferative effects in glioblastoma, acute myeloid leukemia, and gastric cancer models where lead FTO-43 demonstrated potency comparable to clinical chemotherapeutic 5-fluorouracil. Furthermore, FTO-43 increased m6A and m6Am levels in a manner comparable to FTO knockdown in gastric cancer cells and regulated Wnt/PI3K-Akt signaling pathways. The oxetanyl class contains significantly improved anticancer agents with a variety of applications beyond glioblastoma.
N6-methyladenosine
is the most common
internal modification in eukaryotic RNA and is known to regulate RNA
stability and translation, particularly in mRNAs.[1−4] This dynamic modification is regulated
by three classes of proteins, known as “writers”, “readers”,
and “erasers”. The multiprotein “writer”
complex including METTL3, METTL14, and WTAP installs the methyl modification
to mRNAs[5,6] while “reader” proteins, predominantly
comprising the YTH domain-containing family (YTHDF1–3 and YTHDC1–2),
recognize and bind to the m6A-modified transcripts and
mediate downstream signaling events including degradation.[7−10] The “eraser” class of proteins, including both m6A RNA demethylase ALKBH5 (ALKBH5) and fat mass- and obesity-associated
protein (FTO), remove the m6A modifications.[11−13] While ALKBH5 has only been demonstrated to recognize and remove
m6A methyl groups, FTO is known to also recognize and remove N6,2′-O-dimethyladenosine
modifications.[14−18]After the identification of FTO as an m6A demethylase
in 2011, its role in tumorigenesis and poor prognosis of multiple
cancers, including GBM and acute myeloid leukemia (AML), has gained
widespread interest.[19−29] This interest has led to the identification of several small-molecule
inhibitors including rhein, which binds to FTO and its homologue ALKBH5
indiscriminately, and meclofenamic acid (MFA).[30,31] As MFA was identified to increase m6A levels in cells
by inhibiting FTO preferentially over ALKBH5, a variety of derivative
small-molecule inhibitors were inspired by this structure.[26] One such derivative was recently determined
to suppress the proliferation of human-derived AML cell lines in xenotransplanted
mice, validating FTO as a druggable cancer target.[26] Interest in the discovery of small molecule inhibitors
of FTO is growing rapidly with the publication of several excellent
reviews of existing inhibitors and their applications in 2021.[32,33] However, the cellular efficacy of these analogues is modest, and
their use in vivo is limited by poor ADME and PK
profiles. To progress the development of FTO inhibitors as anticancer
therapeutics, it is essential to identify chemically diverse inhibitors
with improved cellular efficacy and physicochemical properties.Recently, we reported the design and biochemical characterization
of a novel class of FTO inhibitors with improved physicochemical properties
relative to existing inhibitors MFA and FB23-2.[29] Lead compound FTO-04 was determined to inhibit FTO preferentially
over its homologue demethylase ALKBH5 with improved potency relative
to MFA (Figure A).
Treatment of glioblastoma stem cell (GSC)-derived neurospheres with
FTO-04 resulted in a significant impairment of neurosphere self-renewal
and sphere formation, while neurospheres grown from healthy neural
stem cells (hNSCs) were not impaired by FTO-04.[29] In this study, we sought to extend our previous work by
further optimizing the potency of FTO-04 using a combination of rational
design and medicinal chemistry techniques and to broaden the range
of applications as a potential chemotherapy.
Figure 1
Rational design of oxetanyl
FTO inhibitors. (A) Optimization strategy
identifying lead FTO-43 N. (B) Docking pose of hit FTO-04. FTO-04
forms a hydrogen bond with Arg 96, and a water-mediated hydrogen bond
is observed with Lys 216. Optimization of the benzothiazole ring was
directed toward catalytic residues His 231 and Glu 234. (C) Docking
pose of FTO-04 with the benzothiazole ring replaced by 5-fluoroindole.
New interactions are observed with catalytic His 231, Glu 234, and
Arg 322, while interactions are maintained with Arg 96. (D) Docking
pose of a pyrrolidine amide replacement of the pyrimidine ring. The
amide forms a weak hydrogen bond to the backbone of Ser 229. (E) Comparison
of the pyrrolidine amide (green) and oxetane (gray) scaffolds. The
oxetane substitution is observed to strengthen the hydrogen bond with
Ser 229 without significantly altering the interactions with residues
Arg 96, His 231, Glu 234, or Arg 322.
Rational design of oxetanyl
FTO inhibitors. (A) Optimization strategy
identifying lead FTO-43 N. (B) Docking pose of hit FTO-04. FTO-04
forms a hydrogen bond with Arg 96, and a water-mediated hydrogen bond
is observed with Lys 216. Optimization of the benzothiazole ring was
directed toward catalytic residues His 231 and Glu 234. (C) Docking
pose of FTO-04 with the benzothiazole ring replaced by 5-fluoroindole.
New interactions are observed with catalytic His 231, Glu 234, and
Arg 322, while interactions are maintained with Arg 96. (D) Docking
pose of a pyrrolidine amide replacement of the pyrimidine ring. The
amide forms a weak hydrogen bond to the backbone of Ser 229. (E) Comparison
of the pyrrolidine amide (green) and oxetane (gray) scaffolds. The
oxetane substitution is observed to strengthen the hydrogen bond with
Ser 229 without significantly altering the interactions with residues
Arg 96, His 231, Glu 234, or Arg 322.In this work, we report the identification and
characterization
of a new class of oxetanyl FTO inhibitors derived from our previous
lead FTO-04. This class demonstrated nanomolar potency and high selectivity
against the homologue m6A RNA demethylase ALKBH5 with significantly
improved LLE values over FTO-04. We determined that like FTO-04, the
oxetanyl class remains to be competitive inhibitors of FTO in agreement
with our molecular docking experiments directed at the meclofenamic
acid (MA) binding site. This class demonstrated potent antiproliferative
effects in a variety of cancer cell models, including glioblastoma,
acute myeloid leukemia, and gastric cancer, where FTO is highly upregulated.
Lead FTO-43 N was demonstrated to impair growth in multiple gastric
cancer cell lines with potency comparable to clinical chemotherapeutic
agent 5-fluorouracil and no significant toxicity toward healthy colon
cells. Furthermore, FTO-43 N was demonstrated to increase m6A and m6Am levels in a manner comparable to
FTO knockdown in gastric cancer cells, confirming that FTO is a relevant
cellular target. These data indicate that the oxetanyl class of FTO
inhibitors and FTO-43 N specifically are significantly improved antiproliferatives
with a variety of applications beyond glioblastoma.
Results
Rational Design and In Silico Modeling of Oxetanyl
FTO Inhibitors
Previously, we identified a pyrimidine inhibitor
that was demonstrated to inhibit FTO competitively with a ∼10×
preference for FTO over its homologue demethylase ALKBH5. Importantly,
FTO-04 exhibited significant impairment of GSC-derived neurosphere
formation without impeding neurosphere formation in hNSCs.[29] While this compound presented an interesting
lead for optimization, the potency remained modest with an enzymatic
IC50 of ∼3 μM. While FTO-04 showed some selectivity
against the homologue ALKBH5 (IC50 ≈ 40 μM),
we theorized that selectivity could be further optimized via interactions
with the nucleotide recognition lid, which is distinct between FTO
and ALKBH5. We therefore sought to improve these parameters using
a combination of rational design and medicinal chemistry techniques.
Proposed modifications were first docked with Glide XP in the Schrödinger
software suite, and compounds were ranked according to their expected
docking interactions. Molecular docking was directed toward the catalytic
site where we previously predicted FTO-04 would bind.[29] Previous docking indicated FTO-04 was likely to interact
with catalytic His 231 and Arg 96 (Figure B); in this round of design, we sought to
maintain these interactions while forming new interactions with Glu
234 and Ser 229 at the periphery of the binding pocket (Figure B–E). Representative
docking poses are presented in Figures B–E and and Figure S1.
Figure 2
Docking poses for select
oxetanyl FTO inhibitors. (A) Optimization
strategy. Docking pose of hit FTO-04 with FTO-11 N overlaid. FTO-04
is shown in magenta, and FTO-11 N is shown in green. FTO-04 forms
a hydrogen bond with Arg 96 and π–π interactions
with His 231. Optimization sought to strengthen interactions with
Ser 229 and Glu 234 while maintaining π–π interactions
with His 231. (B) Docking pose of FTO-09 N. Hydrogen bonds are observed
with Ser 229 and Glu 234, and π–π interactions
are observed with His 231. (C) Docking pose of FTO-11 N. Hydrogen
bonds are observed with Ser 229 and Glu 234, and π–π
interactions are observed with His 231. (D) Docking pose of FTO-38
N. A hydrogen bond is observed with Ser 229, and π–π
interactions are observed with His 231.
Docking poses for select
oxetanyl FTO inhibitors. (A) Optimization
strategy. Docking pose of hit FTO-04 with FTO-11 N overlaid. FTO-04
is shown in magenta, and FTO-11 N is shown in green. FTO-04 forms
a hydrogen bond with Arg 96 and π–π interactions
with His 231. Optimization sought to strengthen interactions with
Ser 229 and Glu 234 while maintaining π–π interactions
with His 231. (B) Docking pose of FTO-09 N. Hydrogen bonds are observed
with Ser 229 and Glu 234, and π–π interactions
are observed with His 231. (C) Docking pose of FTO-11 N. Hydrogen
bonds are observed with Ser 229 and Glu 234, and π–π
interactions are observed with His 231. (D) Docking pose of FTO-38
N. A hydrogen bond is observed with Ser 229, and π–π
interactions are observed with His 231.Initial design focused on replacement of the benzothiazole,
where
a wide variety of heterocyclic and aromatic ring systems were docked
to the catalytic site of FTO. Modeling indicated that substituted
benzene, pyridine, indole, and biaryl variants were most likely to
form favorable interactions with catalytic residues His 231 and Glu
234 while maintaining interactions with Arg 96 (Figure C). This modification positioned the pyrimidine
ring deeper within the binding pocket such that the expected water-mediated
hydrogen bond to Lys 216 observed for FTO-04 is no longer expected.
An amine linker was introduced between the ring systems so that interactions
with the periphery of the binding site, specifically Ser 229, could
be achieved while maintaining the new interactions with His 231 and
Glu 234. Subsequent modeling was focused on optimization of the pyrimidine
ring, which was modified with a variety of aromatic rings, as well
as saturated ring systems intended to introduce more three-dimensional
character to the scaffold. Of these, a pyrrolidine amide scaffold
was expected to effectively maximize the unoccupied space in the binding
pocket and form a hydrogen bond with the amide backbone of Ser 229
(Figure D).In an effort to limit the facile hydrolysis often observed for
amide groups, we chose to substitute this feature for an oxetane moiety.
Oxetane ring systems have only recently become synthetically feasible
groups for drug design, although initial studies indicate that there
are several advantages to this system over ketone, ester, and amide
variants. Oxetanes are able to form stronger hydrogen bonds than most
cyclic ethers and can compete with carbonyl groups as H-bond acceptors.
This is due in part to the strained C–O–C bond angle
(90.2°, unsubstituted ring) exposing the lone pair electrons
and encouraging H-bond interactions. Replacement of amide groups with
oxetanes has been shown to decrease amide basicity and improve chemical
stability at a broad pH range (1–10) while maintaining similar
dipoles and H-bonding properties.[34−39] Molecular docking of this substituent in place of the proposed amide
linker was observed to strengthen the desired hydrogen bonding interactions
with Ser 229 without significantly altering the position of the remainder
of the compound scaffold (Figure E). As such, we chose to prioritize the oxetanyl pyrrolidine
scaffold and focus our in vitro evaluation on substitution
to the other heterocyclic systems.Prior to synthesis, physicochemical
properties were calculated
for each proposed inhibitor using QikProp, including measures of lipophilicity
(clogP), membrane permeability (Caco-2 and MDCK model diffusion rates),
and solubility (polar surface area and topological polar surface area; Table S1). Preference was given to compounds
that were expected to show similar physicochemical properties to those
observed for clinical candidates with CNS exposure (Table S2);[40−43] however, no compound was excluded from synthesis based on these
criteria alone. In total, 53 compounds expected to target the MA binding
site with favorable physicochemical properties were selected for synthesis
and biochemical evaluation.
Synthesis of the Oxetanyl Library
The synthetic scheme
is presented in Scheme , where synthesis of the key intermediate 4 was performed
as described in McLaughlin et al.(39) First, commercially available 3-oxetanone was treated with
nitromethane and a catalytic amount of triethylamine to generate a
β-nitroalcohol intermediate. Subsequent dehydration with methanesulfonylchloride
and excess triethylamine at −78 °C yielded the 3-(nitromethylene)oxetane 2. Pyrrolidine was coupled to the nitroolefin proceeded with
NaHCO3 in THF (3), and the nitro group was
reduced to a primary amine with Raney Ni under a hydrogen atmosphere
to obtain key intermediate 4. Intermediate 4 was treated with the appropriate aldehyde in methanol to form a
Schiff base, which was then reduced by sodium borohydride in situ to afford the final product. After extraction with
ethyl acetate and water (3 × 15 mL), the crude product was purified
by flash column chromatography (MeOH:DCM, 1:9 → 2:8) to afford
the final product at >95% purity as determined by analytical HPLC.
All compounds were characterized by 1H and 13C NMR and HRMS prior to biochemical evaluation.
Scheme 1
Synthesis of Oxetanyl
FTO Inhibitors
Rationally Designed Oxetanyl Pyrrolidines Inhibit Recombinant
FTO with Nanomolar Potency and High Selectivity
To evaluate
their ability to inhibit FTO, the compounds were screened against
recombinant FTO using the high-throughput fluorescence-based inhibition
assay developed by Svensen et al.(44) and previously described in Huff et al.(29) Briefly, a nonfluorescent methylated
RNA substrate termed “m6A7-Broccoli”
is incubated with FTO in the presence of 2-oxoglutarate (300 μM),
(NH4)2Fe(SO4)2·6H2O (300 μM), and l-ascorbate (2 mM) for 2 h
at room temperature in reaction buffer (50 mM NaHEPES (pH 6)). Read
buffer (250 mM NaHEPES (pH 9), 1 M KCl, and 40 mM MgCl2) containing the small molecule 3,5-difluoro-4-hydroxybenzylidene
imidazolinone (DFHBI-1 T, 2.2 μM) was added to the reaction
mixture, and DFHBI-1 T binds preferentially to demethylated Broccoli
to produce a fluorescent signal after incubation for 2 h at room temperature.
The enzymatic activity of FTO was tested at six concentrations of
each inhibitor, ranging from 0–40 μM in triplicate. To
exclude false positives, the assays were repeated with demethylated
Broccoli to ensure that any change in fluorescence was not due to
interference with the A7 Broccoli-DHBI-1 T complex (Figure S2); no compounds were observed to significantly
alter fluorescence signal at concentrations up to 40 μM. Previously,
we determined that DMSO does not impair enzyme activity until the
concentrations exceed 1% total volume;[29] hence, all inhibitors were dissolved in assay buffer to a final
concentration of less than 0.2% DMSO.The IC50 values
of the oxetane compounds ranged from 0.07 to >20 μM (Table and Table S1). Approximately half of the compounds showed IC50 values below 5 μM, and 11 exhibited submicromolar
IC50s. The 30 most potent inhibitors were also screened
against ALKBH5 to determine if the compounds showed specificity toward
either homologue (Table ).
Table 1
In Vitro Data Summary
for Select Oxetane Inhibitors
Representative dose–response curves are presented
in Figure B,C. Of
these, three
showed >10× greater potency toward FTO than ALKBH5 (FTO-09,
11,
13, 14, 18, 30, 31, 35, 38, 40, 44, 45, and 47 N). Four compounds,
FTO-09, 13, 14, and 47 N exhibited no inhibition toward ALKBH5 at
any concentration tested. Seven compounds showed equivalent potency
toward either homologue (defined as <5× difference in IC50; FTO-12, 16, 17, 37, 39, 42, and 48 N).
Figure 3
Oxetanyls are potent
inhibitors of recombinant FTO in vitro. (A)
LLE plot for select FTO inhibitors. pIC50 is determined
from in vitro inhibition assay against recombinant
FTO. Oxetanyl libraries (red squares) show a significant
improvement in pIC50 and LLE from pyrimidine libraries
(blue spheres). Hit FTO-04 is shown as a yellow triangle. (B) Dose–response
curves for FTO-09 N against recombinant FTO and ALKBH5. (C) Dose–response
curves for FTO-43 N against recombinant FTO and ALKBH5. (D) Double-reciprocal
plot for FTO-11 N. All data sets converge upon a common y-intercept, consistent with competitive inhibition. (E) Double-reciprocal
plot for FTO-35 N. All data sets converge upon a common y-intercept, consistent with competitive inhibition. (F) Double-reciprocal
plot for FTO-38 N. All data sets converge upon a common y-intercept,
consistent with competitive inhibition. (G) Double-reciprocal plot
for FTO-43 N. All data sets converge upon a common y-intercept, consistent with competitive inhibition. (H) Double-reciprocal
plot for FTO-49 N. All data sets converge upon a common y-intercept, consistent with competitive inhibition.
Oxetanyls are potent
inhibitors of recombinant FTO in vitro. (A)
LLE plot for select FTO inhibitors. pIC50 is determined
from in vitro inhibition assay against recombinant
FTO. Oxetanyl libraries (red squares) show a significant
improvement in pIC50 and LLE from pyrimidine libraries
(blue spheres). Hit FTO-04 is shown as a yellow triangle. (B) Dose–response
curves for FTO-09 N against recombinant FTO and ALKBH5. (C) Dose–response
curves for FTO-43 N against recombinant FTO and ALKBH5. (D) Double-reciprocal
plot for FTO-11 N. All data sets converge upon a common y-intercept, consistent with competitive inhibition. (E) Double-reciprocal
plot for FTO-35 N. All data sets converge upon a common y-intercept, consistent with competitive inhibition. (F) Double-reciprocal
plot for FTO-38 N. All data sets converge upon a common y-intercept,
consistent with competitive inhibition. (G) Double-reciprocal plot
for FTO-43 N. All data sets converge upon a common y-intercept, consistent with competitive inhibition. (H) Double-reciprocal
plot for FTO-49 N. All data sets converge upon a common y-intercept, consistent with competitive inhibition.An examination of fluoroindole substituents revealed
that the specificity
of the inhibitors is highly dependent on the ring localization of
the substituent. FTO-11 N, where the fluorine atom is located on the
5-position of the ring, showed a strong preference toward FTO inhibition
relative to ALKBH5 (0.11 μM vs 6.6 μM). However, moving
the fluorine atom to either the 3, 4, or 6 position of the indole
ring resulted in equivalent IC50s between the two homologues
(FTO-12, 16, and 17 N). The docking poses for these compounds indicate
that while they are expected to occupy the same region in the binding
pocket, placement of the 5-fluoroindole variant binds within hydrogen
bonding distance of two arginine residues Arg 96 and Arg 322 (Figure S3). As the fluorine substituent is oriented
further away from these residues, the potency against FTO is observed
to decrease and any specificity relative to ALKBH5 is lost. Similarly,
the 5-methylindole variant FTO-09 N showed no inhibition against ALKBH5
at any concentration tested while inhibiting FTO with submicromolar
potency (Table ).
Shifting the methyl group to the 4-position (FTO-42 N) led to a decrease
in potency against FTO and a loss of specificity against ALKBH5. The
6-methylindole FTO-18 N, however, retained its selectivity toward
FTO, although this substitution was less potent that the 5-methylindole
variant.
The Improved Potency of Oxetanyl Inhibitors Is Not Dependent
on Increasing Lipophilicity
As part of our optimization strategy,
the logD and lipophilic ligand efficiency were determined
for the 30 most potent and selective inhibitors. We theorized that
compounds exhibiting improved potency with modest alterations in lipophilicity
were more likely to show adequate solubility properties for cellular
and in vivo studies. The logD was
measured for the 30 most potent compounds (Table ). The logD value for previously
reported lead compound FTO-04 was also determined for comparison. Figure A shows the ligand
trajectory plot, where the elogD of the 20 most potent compounds is
plotted against the pc as determined by the in vitro inhibition assay against recombinant FTO (red squares). For comparison,
seven of the most potent compounds from our previous pyrimidine library
are included as blue spheres. FTO-04 is included as a yellow triangle.
In general, the oxetane class of inhibitors displayed significantly
higher pIC50s with average logD values
similar to lead FTO-04 (logD = 1.0). Of the 30 compounds
evaluated, 15 compounds showed logD <1, and 15
returned a logD of <1. Of these, 14 were within
the range of 1–3. Compounds with a pIC50 of >6
were
clustered between logD 0.71 and 1.57. These data
demonstrate that the optimization of the oxetanyl inhibitors was achieved
with minimal increases to lipophilicity and the majority of this class
is likely to show favorable solubility properties.
The Oxetanyl Class Is a Competitive Inhibitor of Recombinant
FTO
Previously, lead inhibitor FTO-04 was determined to inhibit
FTO through a competitive mechanism.[29] To
determine if the oxetane class of FTO inhibitors also follows a competitive
mechanism, five representative oxetanes were chosen for analysis by
steady-state enzyme kinetics. The reaction velocity was determined
for FTO in the presence of varying concentrations of FTO-11, 35, 38,
43, and 49 N with a range of 10 substrate concentrations between 0
and 10 μM. A plot of the reaction velocity versus substrate
concentration shows that all concentrations of inhibitors approach
a common vmax when substrate concentrations
exceed 10 μM (Figure S4). The double-reciprocal
plots show all concentrations of either FTO-11, 35, 38, 43, or 49
N converge on a common y-intercept, indicating vmax is independent of the concentration of inhibitor,
supporting a competitive mechanism of inhibition (Figure D–H). This mechanism
is consistent with the in silico modeling targeted
toward the MA binding site and the competitive mechanism previously
reported for both MA and lead compound FTO-04.[29,31]
The Oxetane Class of FTO Inhibitors Is a Potent Antiproliferative
in Multiple Cancer Models
The role of m6A dysregulation
in cancer progression and therapeutic resistance is still emerging,
and FTO has been identified as a potential oncogene in a variety of
cancers, including AML and GBM.[23,24,26,27,29,45,46] Several recent
reports have also indicated that FTO is overexpressed in gastric cancer
tissues versus normal adjacent tissues, and this overexpression is
correlated with poor overall survival and progression-free survival
rates in patients.[47,48] FTO was also found to promote
stem-like characteristics in gastric cancer models and promote lymph
node metastasis, while knockdown of FTO impaired proliferation, migration,
and invasion of gastric cancer cells in vitro.[47] For this study, we elected to evaluate the antiproliferative
effects of our oxetanyl inhibitors in all three cancer models.Initially, the 16 most potent and selective compounds identified
by our in vitro studies were screened in a panel
of GBM and GSC cell lines at a single dose of 30 μM. Compounds
were also evaluated in HEK cells to estimate their off-target toxicity. Figure A illustrates the
effects of oxetane inhibitors as a heat map, where cell viability
is presented as a gradient from pale yellow (100% viability) to black
(0% viability) (data presented in Table S3). Compounds FTO-38, 43, and 49 N were observed to impair cell viability
in the majority of glioblastoma lines tested with minimal effects
in HEK cells. Interestingly, these compounds also reported logD values near 1.5. At this stage, FTO-38, 43, and 49 N were
chosen for further evaluation, while the remainder of the compounds
was triaged.
Figure 4
Oxetanyls impair proliferation in multiple cancer subtypes.
(A)
Heat map depicting the antiproliferative effects of 16 oxetanyl inhibitors
in glioblastoma (T98G, U87, TS576, GSC23, GBM6, and A172) and HEK
cells. Color indicates cell viability normalized to DMSO control.
All cells were treated with a single dose of 30 μM. Cell viability
was recorded by MTS assay after 48 h exposure. (B) Single-dose screen
(30 μM) of FTO-38 N, FTO-43 N, and FTO-49 N in a broad panel
of cancer cells. FTO-43 N shows significant inhibition of AML and
gastric cancer subtypes. Cell viability was recorded by MTS assay
after 48 h exposure. N > 10, *p <
0.05, **p < 0.01, ***p < 0.001
by Student’s t test. (C) Dose–response
curves for FTO-43 N in gastric cancer (AGS, SNU16, and KATOIII) and
healthy colon (CCD841) cell lines.
Oxetanyls impair proliferation in multiple cancer subtypes.
(A)
Heat map depicting the antiproliferative effects of 16 oxetanyl inhibitors
in glioblastoma (T98G, U87, TS576, GSC23, GBM6, and A172) and HEK
cells. Color indicates cell viability normalized to DMSO control.
All cells were treated with a single dose of 30 μM. Cell viability
was recorded by MTS assay after 48 h exposure. (B) Single-dose screen
(30 μM) of FTO-38 N, FTO-43 N, and FTO-49 N in a broad panel
of cancer cells. FTO-43 N shows significant inhibition of AML and
gastric cancer subtypes. Cell viability was recorded by MTS assay
after 48 h exposure. N > 10, *p <
0.05, **p < 0.01, ***p < 0.001
by Student’s t test. (C) Dose–response
curves for FTO-43 N in gastric cancer (AGS, SNU16, and KATOIII) and
healthy colon (CCD841) cell lines.Having identified the three most promising antiproliferative
agents
in GBM, we next sought to expand our screen to include AML and gastric
cancer models for which FTO is an emerging target. A broad panel of
cell lines representing glioblastoma, AML, and gastric cancers were
treated with 30 μM of FTO-38, 43, and 49 N for 48 h, and cell
viability was assessed by MTS assay. As observed in Figure B, antiproliferative effects
are consistently observed in all cancer types. FTO-43 N displayed
strong growth inhibition in NB4, AGS, and SNU-16 cell lines. Additionally,
FTO-43 N showed significantly less toxicity toward HEK cells than
FTO-38 and 49 N, indicating that this compound might have a broader
therapeutic index. Therefore, we elected to further evaluate FTO-43
N in both AML and gastric cancer cell models, where this compound
was most potent.To establish the time dependency of inhibition,
AGS, KATOIII, and
SNU-16 cells were treated with varying doses of FTO-43 N, and the
cell viability was assessed at 24, 48, and 72 h post-treatment by
Cell Titre Glo. As observed in Figure S5, FTO-43 N showed little cytotoxicity until 48 h post-treatment across
AGS, SNU-16, and KATOIII cell lines. While some improvement in antiproliferative
effects were observed at 72 h, this increase in growth inhibition
was modest relative to that observed after 48 h. As such, the remainder
of our cell-based experiments were performed 48 h post-exposure.To determine the IC50, AGS, KATOIII, and SNU-16 cells
were treated with eight doses of FTO-43 N (0, 0.5, 1, 2.5, 5, 10,
25, and 50 μM), and the cell viability was examined (Figure C). The experiment
was also repeated in CCD841-CoN normal colon cells to estimate the
therapeutic index. FTO-43 N was also observed to inhibit the growth
of gastric cancer cell lines with EC50s ranging from 17.7
to 35.9 μM, and no significant growth inhibition observed in
CCD841-CoN normal colon cells (Table ). These values are within range of the reported IC50 for clinical chemotherapeutic agent 5-fluorouracil in these
cell lines (5FU, Table ).[49,50] These data indicate that FTO-43 N is capable
of impairing proliferation of gastric cancer cells with little cytotoxicity
toward normal colon cells.
Table 2
EC50 Values for FTO-43
and 5FU in Gastric Cancer Cellsa
CCD 841
AGS
KATOIII
SNU16
FTO-43
>50
20.3
35.9
17.7
5FU
∼218
7.25–22
21.4
6.4
IC50 values for 5FU averaged
from literature reports.[45,46]
IC50 values for 5FU averaged
from literature reports.[45,46]
FTO-43 N Increases m6A Levels and Regulates Wnt/PI3K-Akt
Signaling in Gastric Cancer Cells Comparable to FTO Knockdown
Intracellular inhibition of FTO is known to increase the levels of
both m6A and m6Am; Yang et
al. previously reported that shRNA silencing of FTO in AGS
cells led to an observable increase in m6A levels as observed
by dot-blot assays.[48] To determine the
effects of FTO-43 N on m6A modifications, AGS cells were
first treated with shControl, shFTO1, or shFTO2 to establish the relative
change in m6A levels due to FTO knockdown (knockdown efficiency
presented in Figure S6). Dot-blot assay
results showed that as expected, both shFTOs were observed to increase
m6A levels (Figure S7A). Similarly,
cells treated with FTO-43 N exhibited significantly elevated levels
of m6A relative to DMSO control, indicating that FTO-43
N is able to alter m6A mRNA levels, which is consistent
with FTO knockdown (Figure S7B).As an orthogonal assay control, the relative levels of m6A and m6Am were further quantified by high-performance
liquid chromatography–tandem mass spectrometry (HPLC–MS/MS/MS)
(Figure ). AGS cells
were treated with shControl (NTC or ctrl), shFTO1, DMSO, or FTO-43
N, and the polyadenylated messenger RNA were isolated, decapped, and
digested to single nucleosides prior to quantification. As observed
in Figure A, FTO knockdown
by shFTO1 led to increased m6A and m6Am levels in AGS cells relative to shControl. Treatment with FTO-43
N also resulted in augmented m6A and m6Am levels with a larger increase for m6Am levels, which is in agreement with FTO inhibition (Figure B). While one of our datasets
does not show a statistically significant increase in m6Am levels post-treatment with FTO-43 N, this is likely
due to an outlier in the dataset, and the trend is readily apparent
in both replicates.
Figure 5
FTO-43 N increases m6A and m6Am levels in gastric cancer cells consistent with FTO knockdown.
(A)
FTO knockdown by shFTO1 increases both m6A and m6Am levels relative to shNTC control (ctrl or NTC) in AGS
gastric cancer cells. (B) Treatment of AGS gastric cancer cells with
FTO-43 N increases both m6A and m6Am levels. The data represent the mean ± SEM of duplicate LC–MS/MS
measurement results from three biological replicates. The two bars
in each data series in (A) and (B) represent results obtained from
two separate LC-MS/MS measurements. *p < 0.05,
**p < 0.01, unless otherwise noted, by Student’s t test. (C) Western blot analysis of the expression of E-cadherin,
Axin1, β-catenin, Akt, Phospho-Akt (Ser473), FTO (top band),
and GAPDH. GAPDH served as a loading control.
FTO-43 N increases m6A and m6Am levels in gastric cancer cells consistent with FTO knockdown.
(A)
FTO knockdown by shFTO1 increases both m6A and m6Am levels relative to shNTC control (ctrl or NTC) in AGS
gastric cancer cells. (B) Treatment of AGS gastric cancer cells with
FTO-43 N increases both m6A and m6Am levels. The data represent the mean ± SEM of duplicate LC–MS/MS
measurement results from three biological replicates. The two bars
in each data series in (A) and (B) represent results obtained from
two separate LC-MS/MS measurements. *p < 0.05,
**p < 0.01, unless otherwise noted, by Student’s t test. (C) Western blot analysis of the expression of E-cadherin,
Axin1, β-catenin, Akt, Phospho-Akt (Ser473), FTO (top band),
and GAPDH. GAPDH served as a loading control.Furthermore, we investigated the mechanism of FTO-43
N inhibition
in gastric cancer by analyzing the Wnt/PI3K-Akt signaling pathways.
Previous studies have shown that a reduction of m6A RNA
methylation predicted adverse clinical features of gastric cancer
and activated oncogenic Wnt/PI3K-Akt signaling to promote malignant
phenotypes of gastric cancer cells.[51]Increasing the levels of m6A RNA methylation by FTO
knockdown reversed these phenotypical and molecular changes in gastric
cancer.[51] As shown in Figure C, Wnt (marked by Axin 1 and
β-catenin expression) and PI3K-Akt (marked by Ser473-Akt phosphorylation)
signaling were downregulated by FTO inhibition by shRNA knockdown
in AGS gastric cancer cells. Conversely, the expression of E-cadherin
(CDH1) was upregulated when FTO was inhibited. Treatment with FTO-43
N replicated this phenotype of downregulated Wnt/PI3K-Akt signaling.
Collectively, these data indicate that FTO is likely a cellular target
of FTO-43 N.
Discussion
In this work, we sought to improve upon
the previously identified
FTO inhibitor FTO-04 by increasing potency and selectivity while maintaining
favorable physicochemical properties, such as logD. Rational design of the oxetanyl class of compounds identified multiple
compounds with nanomolar IC50s against recombinant FTO
with little inhibition observed against the homologue demethylase
ALKBH5 (IC50 > 40 M). LogD measurement
for the most potent and selective compounds indicated that this optimization
was independent of large increases in lipophilicity, suggesting that
most of the compounds identified in this study are likely to show
favorable solubility properties for further development as antiproliferative
agents. This class was further found to inhibit FTO competitively,
consistent with our in silico modeling, and lead
FTO-43 N was demonstrated to increase both m6A and m6Am levels in a manner comparable to FTO knockdown
in gastric cancer cells, confirming that FTO is a relevant cellular
target. Analysis of the Wnt and PI3K-Akt signaling pathways post-treatment
with FTO-43 N demonstrated that this compound downregulated both oncogenic
signaling pathways in a manner consistent with FTO knockdown. Collectively,
these results indicate a successful optimization of our previous hit
and the identification of an improved class of FTO inhibitors.While the role of FTO as an oncogene in AML has been established,
its role in the progression and tumorigenesis of other cancer types
is still emerging. In this work, we expanded our evaluation of oxetanyl
inhibitors to include glioblastoma, which we previously demonstrated
was susceptible to FTO inhibition, and gastric cancer. Currently,
there are very few reports regarding the role of FTO in gastric cancer;
evaluation of TCGA and GTex datasets indicate that FTO is overexpressed
in gastric cancer tissues versus normal adjacent tissues, and this
overexpression is correlated with poor overall survival and progression
free survival rates in patients. Xu et al. further
indicated that FTO expression is correlated with tumor stage, where
higher expression was observed in advanced disease stages (clinical
stage III and T3–4).[47] Recently,
Yang et al. reported that FTO stabilized MYC transcripts,
leading to overexpression of MYC and accelerated proliferation, invasion,
and migration of gastric cancer cells.[48] In our study, we demonstrated that FTO-43 N was able to impair the
proliferation of gastric cancer cells with comparable potency to the
clinical chemotherapy 5-fluorouracil. These data suggest that FTO
inhibition could play a role in gastric cancer progression, although
the biological mechanisms for such a role are yet to be established.Gastric cancer is the fifth most common cancer and the third leading
cause of cancer death worldwide. Gastric tumors are highly heterogeneous
and resistant to treatments in large part because most cases are diagnosed
at advanced stages. Life expectancy for this disease is low with a
5 year survival rate of just 20–30%, and the need for novel
methods to target resistant and advanced stage gastric cancers remains
urgent. We believe that our data in combination with emerging reports
regarding the high upregulation of FTO in gastric cancer suggest further
study of the role of FTO in gastric cancer progression is warranted.While we demonstrate that the oxetanyl class of FTO inhibitors
is significantly improved upon our previous hit FTO-04 in accordance
with the goals of our study, the applications of these compounds are
yet to be fully explored. While FTO-43 N demonstrated antiproliferative
effects comparable to the clinical chemotherapy 5-fluorouracil in
gastric cancer cells, it is important to note that 5-fluorouracil
is often administered to patients in combination with other chemotherapies
such as cisplatin. Additional studies to evaluate whether FTO-43 N
can confer synergistic effects when combined with other FDA-approved
chemotherapies would provide vital information regarding the applications
of this compound as a potential chemotherapeutic agent. Furthermore,
the cell lines used to evaluate FTO-43 N are not enriched in gastric
cancer stem cells, where FTO inhibition is most likely to impair cell
proliferation and tumorigenesis. Additional studies are warranted
to assess FTO-43 N in cancer models with more physiological relevance
to the disease stage most likely to respond to FTO inhibition. Nevertheless,
we believe that this work presents a successful optimization of our
previous hit FTO-04 with many potential antiproliferative applications.
Experimental Section
Detailed procedures for in silico screening and
docking of the inhibitors can be found in the SI. Protocols for protein
expression and purification, in vitro inhibition
assays, and steady-state enzyme kinetics can be found in the SI. Detailed synthetic procedures are presented
in the SI. All cell culture procedures can also be found in the SI.
Molecular Modeling with Schrödinger
In silico modeling of FTO inhibitors was performed using
the Glide docking module of the Schrödinger 11.5 modeling software
suite. A crystal structure of FTO bound to meclofenamic acid (MA)
(PDB ID: 4QKN)[1] was first refined using Prime.[31,52,53] Missing side chains and hydrogen
atoms were resolved before docking, and the Optimized Potentials for
Liquid Simulations All-Atom (OPLS) force field and the surface-generalized
Born (SGB) continuum solution model was used to optimize and minimize
the crystal structures. The docking grid was generated as a 5 Å
× 5 Å × 5 Å cube centered on MA. Glycerol and
α-ketoglutarate were removed from the docking site prior to
grid generation. Ligprep was used to generate a minimized 3D structure
for all prospective FTO inhibitors using the OPLS 2001 force field.
Docking was performed with Glide XP. QikProp was used to predict physicochemical
properties such as clogP and membrane permeability in Caco-2 and MDCK
cell lines for the most promising compounds.
Protein Expression and Purification
The protein expression
and purification protocol was adapted from Svensen and Jaffrey and
previously published in Huff et al.[44,29]Escherichia coli BL21 competent cells
(New England Biolabs) were transformed with a pET28-SUMO-His10-FTO
plasmid (a generous gift from the Jaffrey lab) by heat shock and spread
on a LB Kanamycin agar plate and then incubated overnight at 37 °C.
Two to three colonies were picked and transferred to 5 mL of LB media
treated with kanamycin (0.5 mg/mL final concentration) and then grown
overnight shaking at 37 °C. The overnight culture was then transferred
to 2 L of LB kanamycin medium and incubated at 37 °C until an
OD of 0.8 was reached. The culture was cooled at 4 °C for 20
min. induced with 0.5 mM isopropyl β-d-1-thiogalactopyranoside
(IPTG), and then grown while shaking at 16 °C. Cell pellets were
collected by centrifugation (5000g for 10 min at
4 °C), and the supernatant was discarded. The pellets were resuspended
in B-PER bacterial protein extraction reagent (6 mL per gram) with
DNase 1 (5 U per mL, RNase-free) and incubated at 4 °C for 1
h. The suspension was centrifuged at 10,000g for
20 min, and the supernatant was transferred to a Talon Metal Affinity
Resin column that had been pre-equilibrized with binding buffer (50
mM NaH2PO4 (pH 7.2), 300 mM NaCl, 20 mM imidazole,
and 1 mM β-mercaptoethanol in RNase-free water). The supernatant
was incubated with the affinity resin column at 4 °C for 1 h
with end-over-end rotation. After incubation, the column was washed
with 5 bed volumes of binding buffer and then incubated with 1 bed
volume of elution buffer (50 mM NaH2PO4 (pH
7.2), 300 mM NaCl, 500 mM imidazole, and 5 mM β-mercaptoethanol
in RNase-free water) for 20 mins. After incubation, the eluant was
collected, and the column was incubated again with 1 bed volume of
elution buffer; the elution process was repeated until no further
protein was collected (3–5 bed volumes total). The eluant was
combined and transferred to a Slyde-A-Lyzer Dialysis Cassette (20,000
MWCO, Thermo Scientific) and dialyzed overnight at 4 °C against
dialysis buffer (50 mM Tris–HCl (pH 7.4), 100 mM NaCl, 5 mM
B-mercaptoethanol, and 5% (v/v) glycerol in RNase-free water). Protein
concentration was measured by absorbance at 280 nm and calculated
by Beer–Lambert’s Law (A = ε/C, εFTO = 95,340). ALKBH5 was expressed
and purified from pET28-SUMO-His10-ALKBH5 plasmid by the same procedure
described above.
In Vitro Inhibition Assay Method
The in vitro inhibition assay method was adapted from Svenson
and Jaffrey and previously reported in Huff et al.44,29 All reactions were performed in a 96-well plate
with 200 μL of assay buffer (50 mM HEPES (pH 6), 300 μM
2-oxoglutarate, 300 μM (NH4)2Fe(SO4)2·6 H2O, and 2 mM ascorbic acid
in RNase-free water) with 7.5 μM m6A7-Broccoli
RNA and 0.250 μM FTO. Inhibitors were added in concentrations
ranging from 0.008 to 40 μM; all inhibitors were dissolved in
DMSO and added to a final concentration of 0.2% DMSO. Previously,
we demonstrated that FTO activity is stable at concentrations of DMSO
up to 1% (v/v).[29] Prior to incubation,
40 μL of read buffer (250 mM HEPES (pH 9.0), 1 M KCl, 40 mM
MgCl2, and 2.2 μM DFHBI-1 T in RNase-free water)
was added to bring the final well volume to 200 μL. After incubation
at room temperature for 2 h, the plates were left at 4 °C overnight
(16 h) to allow DFHBI-1 T to bind to A7-Broccoli RNA. Specificity
assays were performed by the same method with 0.250 μM ALKBH5.
Fluorescence intensity was measured with a BioTek Synergy plate reader
with FITC filters (excitation, 485 nm; emission, 510 nm). Sigmoidal
dose–response curves were fitted in GraphPad Prism 8. All assays
were performed in triplicate, with additional repetitions added as
necessary.Michealis–Menton kinetics was performed using
the inhibition assay procedure described above; the activity of FTO
concentrations of 0, 0.250, 0.385. 0.500, 0.625, 0.750, 1.25, 2.5,
5, and 10 μM m6A Broccoli were recorded for the following
concentrations of FTO-11 N: 0, 0.5, 1.5, 5, and 10 μM, and for
FTO-35 N: 0, 0.03125, 11, and 15 μM, FTO-38 N: 0, 0.125, 0.25,
and 0.5 μM, FTO-43 N: 0, 10, 20, and 40 μM, FTO-49 N:
0, 5, 10, 20, and 25 μM. The data were fitted in GraphPad Prism
8.
General Synthetic Procedures
Synthesis of 3-(Nitromethylene)oxetane (2)
Synthesis of 3-(nitromethylene)oxetane (2) was performed
as described in Wuitschik et al.(9) and McLaughlin et al.(10) 3-Oxetanone (130 μL, 2.03 mmol), nitromethane (154
μL, 2.85 mmol), and NEt3 (57 μL, 0.41 mmol)
were stirred at RT for 30 min and then diluted with CH2Cl2 (10 mL) and cooled to −78 °C. To this
solution was added NEt3 (565 μL, 4.05 mmol), followed
by MsCl (157 μL, 2.03 mmol) dropwise over 10 min. The reaction
mixture was stirred at −78 °C for 40 min. The reaction
mixture was allowed to warm to RT and directly poured on a column
(15% → 25% EtOAc in hexane), providing compound (2) as a yellow oil.
Synthesis of 1-(3-(Nitromethyl)oxetan-3-yl)pyrrolidine (3)
Synthesis of 1-(3-(nitromethyl)oxetan-3-yl)pyrrolidine (3) was adapted from procedures described in McLaughlin et al.(10)A stirred solution of
SM (2) (1
eq) in THF was treated with NaHCO3 (1 eq) and pyrrolidine
(1 eq) at room temperature. The reaction mixture was stirred for 1
h at room temperature. After completion of the reaction, the reaction
mixture was filtered through a Celite bed. The solvent was evaporated
under vacuum, and the crude product was purified by column chromatography
(EtOH:hexane, 1:9 → 2:8) to give compound (3) as
a liquid.
Synthesis of (3-(Pyrrolidin-1-yl)oxetan-3-yl)methanamine (4)
Reduction of 1-(3-(nitromethyl)oxetan-3-yl)pyrrolidine (3) was performed as described in McLaughlin et al.(10) to a stirred solution of SM (3) (1 eq) in THF and followed by the addition of Raney nickel.
The reaction mixture was stirred for 3 h under a hydrogen balloon.
After completion of the reaction, the reaction mixture was filtered
through a Celite bed. The solvent was evaporated under a vacuum, and
the crude product was purified by column chromatography (MeOH:DCM,
1:9 → 2:8) to give compound (4) as a liquid.
General Procedure for the Synthesis of General Compound (5) (FTO-1-53 N)
A two-necked round-bottom flask was
charged with compound (4) (1 eq) in MeOH and followed
by the addition of corresponding aldehyde (1 eq). The reaction mixture
was stirred for 3 h at room temperature. After the starting materials
had been consumed, the reaction mixture was slowly treated with NaBH4 (1.5 eq). The reaction mass was again stirred for 2 h at
room temperature. After completion of the reaction, the crude product
was concentrated under vacuum, dissolved in water, and extracted with
ethyl acetate (3 × 15 mL). The organic layers were combined,
dried with NaSO4, and filtered. The filtrate was concentrated
under vacuum to afford the crude product. The crude material was purified
by flash column chromatography (MeOH:DCM, 1:9 → 2:8) to give
the final compound (5) as an oil.
Compound Characterization
The Supporting Information provides a detailed description of the solvent
purification and general methods for the synthesis and purification
of each analogue. All compounds were purified by column chromatography
and characterized by 1H and 13C NMR and HRMS.
Purity was determined by HPLC at >95% for all compounds. Spectra
can
be found in the Supporting Information.
AGS (CRL-1739; Human Gastric Adenocarcinoma),
SNU-16 (CRL-5974; Human Gastric Carcinoma), and KATOIII (HTB-103;
Human Gastric Carcinoma) were all purchased from the American Type
Culture Collection (ATCC). These cell lines were cultured in F-12K
medium (ATCC, 30-2004), RPMI-1640 (ATCC, 30-2001) and IMDM (ATCC,
30-2005), respectively. HEK293T cells and MRC-5 cells were cultured
in DMEM. U937 cells, NB4 cells, and THP-1 cells were cultured in RPMI-1640
(Gibco), and CCD 841 CoN cells were cultured in Eagle’s minimum
essential medium (EMEM) (ATCC, 30–2003). All media were supplemented
with 10% fetal bovine serum (Gibco). All cells were cultured at 37
°C in a humidified 5% CO2 atmosphere.
Cell Viability Assays
The procedures were approved
by the University of California San Diego Institutional Review Board.
To evaluate the antiproliferative effects of FTO inhibitors in various
cancerous and noncancerous cell lines, cytotoxicity was measured using
a Cell Counting Kit-8 (CCK-8) assay (Apexbio Technology LLC K1018,
Fisher Scientific 50-190-5564) following the manufacturer’s
instructions. The evaluation of FTO inhibitors was performed in glioblastoma
cells (U87, A172, T98, TS576, GSC23, and GBM6), gastric cancer cells
(AGS, SNU-16, KATOIII), acute myeloid leukemia cells (U937, THP-1,
NB4), and normal epithelial cells (CCD 841). GSC23, GBM6, and TS576
cell lines were obtained from the Frank Furnari lab at UCSD and cultured
in DMEM/F12 medium supplemented with 1:100 B27 without vitamin A,
EGF (20 ng/mL), FGF (10 ng/mL), and penicillin–streptomycin
(100 IU/mL) as previously described.[11,12]A total
of 10,000 cells were plated in 96-well plates for cell viability assays.
Initial screening was performed for a single dose of 30 μM 48
h post-exposure. Tested cell lines were exposed to a series of concentrations
(0, 5, 10, 20, 25, and 50 μM) of FTO-43 N for 24, 48, and 72
h. All data were normalized to culture medium for background control.
Two hours before the end of the culture period, 10 μL of CCK-8
was added to each well including medium-only wells (blanks). After
2 h of incubation with CCK-8 at 37 °C, the optical density of
the viable cells was recorded at 450 nm using a BioTek Synergy plate
reader. Relative cell viability was expressed as a percentage relative
to the DMSO treated control cells and 0 time points. FTO knockdown
cell lines were expressed as a percentage-relative shRNA control (shNTC).
Each experiment was repeated three times and performed in triplicate
or quadruplicate.
Lentiviral Generation and Infection
Lentiviral particles
were prepared for shControl, shFTO1, and shFTO2 as previously described
in Huff et al.[8] Lentiviruses
were generated by co-transfecting HEK293T cells with the shRNA-expressing
vectors (carrying a puromycin resistance gene), a packaging plasmid
(psPAX2), and an envelope plasmid (pMD2.G) using Opti-MEM and Lipofectamine
3000 (Life Technologies, 11,668,027) according to the manufacturer’s
instructions. The medium was replaced with fresh completed DMEM after
4–6 h. The virus-containing supernatant was harvested after
48 h of transduction, filtered at 0.45 μm, and stored at −80
°C. Generated shRNA control (NTC), shFTO-249, and shFTO-250 lentivirus
particles were used to infect AGS cells, SNU-16 cells, and KATOIII
cells in the presence of Polybrene (8 μg/mL) (Millipore) by
spin transduction (centrifuge at 750g for 45 min
at RT). After 12 h of infection, the lentivirus-containing medium
was replaced with fresh medium, and transduced cells were selected
by culture with puromycin (Sigma TR-1003) at 1.5 μg/mL for AGS
and SNU-16, 4 ug/ml for KATOIII for 7 days to generate TFO knock down
(KD) stable cell lines, and KD efficiency was determined by qPCR or
Western blot analysis.
Immunoblotting
Proteins from cells were extracted using
Pierce IP Lysis Buffer supplemented with protease inhibitors (Life
Technologies, 87786) followed by centrifugation to remove insoluble
material, and clarified supernatant was measured using a BCA protein
assay kit (Bio-Rad). Thirty micrograms of protein was resolved by
NuPAGE 4–12% Bis-Tris gels and transferred to PVDF membranes
(Bio-Rad) according to the manufacturer’s instructions. Membranes
were blocked in 5% non-fat dry milk in PBS buffer and then incubated
with the indicated antibodies including FTO (Cell Signaling Technology,
14386), GAPDH (Proteintech, HRP-60004), Axin1 (Cell Signaling Technology,
2087), β-catenin (Cell Signaling Technology, 8480), Akt (Cell
Signaling Technology, 4691), Phospho-Akt (Ser473) (Cell Signaling
Technology, 4060) and E-cadherin (Cell Signaling Technology, 3195)
overnight at 4 °C. After being washed, the membranes were incubated
with HRP-conjugated secondary antibodies at room temperature (RT)
for 15 min and then washed 3 times as before (Thermo Scientific, Pierce
Fast Western Blotting Blot Kit). Protein was visualized on autoradiography
film (Genesee Scientific Inc., 30-100) using the enhanced chemiluminescence
(ECL) detection system (Thermo Scientific).
RNA Isolation
Total RNA was extracted from culture
cells using Direct-zol RNA MiniPrep Kit (Zymo Research, 11–331);
all RNAs were treated with DNase I. mRNA was isolated using the Magnetic
mRNA Isolation Kit (New England Biolabs, S1550S) following the manufacturer’s
instructions.
Quantitative Real-Time PCR
Reverse transcription of
total RNA to cDNA was performed
using the iScript Reverse Transcription Synthesis Kit (Bio-Rad, 1,708,841),
and qPCR reactions were set up in 10 μL volumes, consisting
of 3 μL of cDNA (6–15 ng), 2 μL of mixed forward
and reverse (300 nM) primers and 5 μL of SsoAdvanced Universal
SYBR Green PCR SuperMix (Bio-Rad, 1725270). qPCR was run on 384 well-PCR
machine. The primers used in this study were listed below. The transcript
level of FTO was quantified using the ΔΔ
Ct method using GAPDH as a housekeeping reference
gene.FTO forward: 5′-TTGCCCGAACATTACCTGCT-3′.FTO reverse: 5′-TGTGAGGTCAAACGGCAGAG-3′.
m6A Dot-Blot Assay
m6A dot-blot
assays were performed as described in Huff et al.(29) mRNA or total RNA were isolated from
AGS cells treated with either shControl, shFTO1, shFTO2, DMSO, or
FTO-43 N using Magnetic mRNA Isolation Kit (New England Biolabs, S1550S).
RNA samples were quantified, serially diluted, and denatured at 95
°C for 3 min and then chilled on ice to prevent the reformation
of the secondary structure of mRNA. Denatured mRNA samples were spotted
on an Amersham Hybond-N+ membrane (GE Healthcare, RPN3050B) and cross-linked
to the membrane with UV radiation. After cross-linking, the membrane
was washed with phosphate-buffered saline with 0.02% Tween 20 (PBST),
blocked in 5% non-fat milk in PBST buffer for 1 h at room temperature
(RT), and then incubated with anti-m6A antibody (1: 1000;
Abcam) overnight at 4 °C. The membrane was then washed as before
and incubated in HRP-conjugated secondary antibodies for 1 h at RT.
The signaling was detected by SuperSignal West Pico Chemiluminescent
Substrate (Thermo Scientific, 34080) and developed using autoradiography
film (Genesee Scientific Inc., 30-100). For RNA loading control, after
imaging, the membrane was stained with 0.2% methylene blue in 0.4
M sodium acetate and 0.4 M acetic acid for 1 h and washed with water
for until the background is clean.
Statistical Analysis
All values are expressed as the
means ± SEM of at least three independent experiments if no additional
information was indicated. Statistical differences among groups were
determined using either Student’s t test or
one-way ANOVA. P values of less than 0.05 were considered
statistically significant. The analyses were performed using GraphPad
Prism 8 software.
Quantification of m6A and m6Am by LC–MS/MS/MS
Polyadenylated RNA was analyzed by
LC–MS/MS/MS as described previously.[13,29] Briefly, polyadenylated RNA (100 ng) was incubated with 5 units
of RppH in ThermoPol buffer at 37 °C for 3 h. The mixture was
then digested with 1 unit of nuclease P1 in 25 μL buffer containing
25 mM of NaCl and 2.5 mM of ZnCl2 at 37 °C for 2 h
followed by the addition of 3 μL of 1 M NH4HCO3 and 1 U alkaline phosphatase. After an additional incubation
at 37 °C for 2 h, uniformly 15N-labeled rA and D3-labeled m6A internal standards were added to the
mixture. The enzymes in the digestion mixture were removed by extraction
with chloroform, and the salts in the digestion mixture were removed
by CH3CN precipitation. The ensuing supernatant was dried
by Speed-vac, redissolved in doubly distilled water, and subjected
to LC–MS/MS analysis.The LC–MS/MS/MS experiments
were conducted on an LTQ-XL linear ion trap mass spectrometer (Thermo
Fisher Scientific) coupled with an EASY-nLC II system (Thermo Fisher
Scientific). Mobile phase A was 0.1% formic acid in H2O,
and mobile phase B was 0.1% formic acid in acetonitrile. The samples
were loaded onto a trapping column (150 μm i.d.) packed with
porous graphitic carbon (PGC, 5 μm particle size, Thermo Fisher
Scientific) at a flow rate of 2.5 μL/min within 8 min. The samples
were then eluted onto an analytical column (75 μm i.d.) packed
with Zorbax SB-C18 (5 μm in particle size, 100 Å in pore
size, Agilent) using a nonlinear gradient consisting of 0–16%
B in 5 min, 16–22% B in 23 min, 22–50% B in 17 min,
50–90% B in 5 min, and finally at 90% B for 30 min at a flow
rate of 300 nL/min. Quantification was performed by comparison with
the calibration curves obtained from pure nucleoside standards analyzed
under the same experimental conditions. Based on the peak area ratios
in the selected-ion chromatograms for the analytes over their corresponding
isotope-labeled standards, the amounts of the labeled standards added
and the equations derived from the calibration curves, the moles of
m6A, m6Am, and rA in the nucleoside
mixtures were calculated. The amounts of m6A and m6Am were normalized to that of rA in the same sample
to give the levels of these two modified nucleosides.
Authors: Georg Wuitschik; Mark Rogers-Evans; Klaus Müller; Holger Fischer; Björn Wagner; Franz Schuler; Liudmila Polonchuk; Erick M Carreira Journal: Angew Chem Int Ed Engl Date: 2006-11-27 Impact factor: 15.336
Authors: Kate D Meyer; Deepak P Patil; Jun Zhou; Alexandra Zinoviev; Maxim A Skabkin; Olivier Elemento; Tatyana V Pestova; Shu-Bing Qian; Samie R Jaffrey Journal: Cell Date: 2015-10-22 Impact factor: 41.582
Authors: Guanqun Zheng; John Arne Dahl; Yamei Niu; Peter Fedorcsak; Chun-Min Huang; Charles J Li; Cathrine B Vågbø; Yue Shi; Wen-Ling Wang; Shu-Hui Song; Zhike Lu; Ralph P G Bosmans; Qing Dai; Ya-Juan Hao; Xin Yang; Wen-Ming Zhao; Wei-Min Tong; Xiu-Jie Wang; Florian Bogdan; Kari Furu; Ye Fu; Guifang Jia; Xu Zhao; Jun Liu; Hans E Krokan; Arne Klungland; Yun-Gui Yang; Chuan He Journal: Mol Cell Date: 2012-11-21 Impact factor: 17.970
Authors: Isaia Barbieri; Konstantinos Tzelepis; Luca Pandolfini; Junwei Shi; Gonzalo Millán-Zambrano; Samuel C Robson; Demetrios Aspris; Valentina Migliori; Andrew J Bannister; Namshik Han; Etienne De Braekeleer; Hannes Ponstingl; Alan Hendrick; Christopher R Vakoc; George S Vassiliou; Tony Kouzarides Journal: Nature Date: 2017-11-27 Impact factor: 49.962
Figure 1
Figure 2
Scheme 1
Figure 3
Figure 4
Figure 5