Abstract

   N6-Methyladenosine (m^6A) modification and its regulators play critical
   roles in human cancers, but their functions and regulatory mechanisms
   in adenocarcinoma of the esophagogastric junction (AEG) remain unclear.
   Here, we identified that IGF2BP3 is the most significantly up-regulated
   m^6A regulator in AEG tumors versus paired normal adjacent tissues from
   the expression profile of m^6A regulators in a large cohort of AEG
   patients. Silencing IGF2BP3 inhibits AEG progression in vitro and in
   vivo. By profiling transcriptome-wide targets of IGF2BP3 and the m^6A
   methylome in AEG, we found that IGF2BP3-mediated stabilization and
   enhanced expression of m^6A-modified targets, including targets of the
   cell cycle pathway, such as CDC25A, CDK4, and E2F1, are critical for
   AEG progression. Mechanistically, the increased m^6A modification of
   CDC25A accelerates the G1-S transition. Clinically, up-regulated
   IGF2BP3, METTL3, and CDC25A show a strong positive correlation in TCGA
   pan-cancer, including AEG. In conclusion, our study highlights the role
   of post-transcriptional regulation in modulating AEG tumor progression
   and elucidates the functional importance of the m^6A/IGF2BP3/CDC25A
   axis in AEG cells.

   Keywords: Adenocarcinoma of the esophagogastric junction, m^6A,
   IGF2BP3, CDC25A, Cell cycle

Introduction

   Adenocarcinoma of the esophagogastric junction (AEG) is defined as the
   adenocarcinoma that occurs in the esophagogastric junction within the
   range of 5 cm in both directions[41]^1. The incidence of AEG is
   increasing worldwide with unclear pathogenesis[42]^2^, [43]^3. Despite
   recent advances in genomic, transcriptomic, proteomic and
   phosphoproteomic profiling of large cohorts of AEG patients, and in
   multimodal treatments, AEG still lacks effective therapeutic targets
   and prognostic markers[44]^4^-[45]^7. Therefore, it is imperative to
   understand the molecular mechanisms underlying AEG carcinogenesis and
   to identify potential prognostic markers and therapeutic targets.

   N6-methyladenosine (m^6A) modification plays a critical role in
   regulating mRNA fate in both physiology and pathology, including
   cancer[46]^8^-[47]^11. This modification is installed by the m^6A
   methyltransferase complex consisting of METTL3/METTL14/WTAP and other
   cofactors, and can be removed by the demethylases FTO and
   ALKBH5[48]^12^-[49]^14. The functional effects of m^6A on mRNA fate are
   mediated through m^6A reader proteins, such as the YTH
   domain-containing proteins (YTHDF1/2/3) and insulin-like growth factor
   2 mRNA-binding proteins (IGF2BP1/2/3)[50]^15^-[51]^17. These m^6A
   regulators have been shown to play critical roles in multiple
   cancers[52]^10^, [53]^18^, [54]^19. However, the role of m^6A in
   regulating AEG tumor progression remains unexplored.

   The function of IGF2BPs is well studied in multiple tumors[55]^15^,
   [56]^20^-[57]^22, but its function and mechanism in AEG have not been
   elucidated. Here, we found that high IGF2BP3 promotes AEG progression
   by increasing CDC25A expression in an m^6A dependent manner to
   accelerate the G1-S transition. Our study highlights the importance of
   post-transcriptional regulation in AEG tumor progression and the
   clinical significance of the m^6A/IGF2BP3/CDC25A axis in AEG.

Materials and Methods

Patients and sample collection

   Thirty treatment-naïve patients, who were newly diagnosed with AEG,
   were enrolled in this study. Prior to their surgeries, written informed
   consent was obtained from each patient, and fresh samples were
   collected post-surgery. The clinical characteristics of these patients
   are listed in [58]Table S1.

Plasmids

   The wild-type (WT) IGF2BP3, METTL3, FTO, ALKBH5, and CDC25A, along with
   the mutant METTL3 (D377A, D395A, N539A, E532A), FTO (H231A/D233A,
   R316Q/R322Q), and ALKBH5 (H204A) expression plasmids were cloned into
   the pLVX-IRES-Neo vector. The dCas13b-ALKBH5/FTO and non-targeting gRNA
   plasmids were reported in our previous study[59]^11. Two gRNAs
   targeting the CDS of CDC25A were designed and cloned into the gRNA
   plasmid (Addgene plasmid #103854). The sequences of the gRNAs were:

   gRNA1: 5′-CATACCGGCACATGCGGGGACCTCTCTCAGAAGAAAACTC-3′; gRNA2:
   5′-GTTCATCCCACTGTGGCTCAGAGCAGCTTGACACGGTGCT-3′.

   Mature antisense sequences of shRNA are listed in [60]Table S2.

Cell culture and transfection

   Human AEG cell lines, including OE-19 (from a gastric cardia
   adenocarcinoma of a 72-year-old male patient) and SK-GT4 (from the
   primary tumor of an 89-year-old Caucasian man with an adenocarcinoma of
   the distal esophagus), were obtained from the Shanghai Cell Bank Type
   Culture Collection Committee (Shanghai, China). They were cultured in
   RPMI1640 medium (Gibco) with 10% fetal bovine serum (FBS) and 1%
   penicillin/streptomycin (Invitrogen). These cell lines were verified by
   short tandem repeat assays for their identification and tested negative
   for Mycoplasma contamination. The cells were cultured at 37^oC in a
   humidified incubator with 5% CO[2].

RNA isolation and qRT-PCR

   Total RNA was isolated using RNAiso Plus (Takara) and subjected to
   reverse transcription using the HiScript III All-in-One RT SuperMix
   Perfect for qPCR kit (Vazyme, China). qRT-PCR was performed using the
   Taq Pro Universal SYBR qPCR Master Mix kit (Vazyme, China) and analyzed
   with a QuantStudio 7 Flex Real-Time PCR System (Thermo Fisher). PCR
   results, recorded as Ct numbers, were normalized against GAPDH. The
   relative gene expression levels were analyzed using the 2^-ΔΔCT method.
   The primers used are listed in [61]Table S3.

Immunoblot

   Proteins were extracted with lysis buffer (150 mM KCl, 10 mM HEPES pH
   7.6, 2 mM EDTA, 0.5% NP-40, 0.5 mM DTT, 1:100 protease inhibitor
   cocktail), separated by SDS-PAGE, transferred onto PVDF membranes,
   blocked in 5% nonfat milk, and then blotted with specific antibodies
   ([62]Table S4). Immunoblot results were normalized against GAPDH.

Cell proliferation, migration, and colony formation assay

   Cells were seeded in 96-well plates, with each well containing
   1500-2000 cells in 100 μl of cell suspension. After a certain period in
   culture, cell viability was measured using the CellTiter-Glo
   Luminescent™ Cell Viability Assay (Promega). Migration assays were
   performed using transwell inserts with polyethylene terephthalate
   membranes (24-well inserts, 8.0 μm, Corning). 4×10^4 cells in
   serum-free medium were added to the upper chamber, while 500 μl of
   medium supplemented with 10% FBS was added to the lower chamber to
   serve as a chemotactic agent. After incubation, cells that migrated
   through the filters were fixed with methanol and stained with crystal
   violet. The stained cells were destained with 33% acetic acid, and
   migration ability was assessed by measuring absorbance at 560 nm. For
   the colony formation assay, 1000 cells were cultured in a six-well
   plate at 37^oC in a 5% CO[2] humidified environment. After two weeks,
   cells were fixed, stained with crystal violet, and photographed.

m^6A-RIP

   Total RNA was extracted using Trizol and treated with TURBO DNase (2
   U/µl) (Invitrogen, AM2239), then fragmented by RNA fragmentation
   reagents (Thermo, AM8740). 100 μg purified RNA was incubated with A/G
   magnetic beads (Millipore) coated with 3 μg anti-m^6A antibody
   (#202003, Synaptic Systems) at 4°C for 6 hours in 500 μl of NT2 buffer
   (200 mM NaCl, 50 mM HEPES pH 7.6, 2 mM EDTA, 0.05% NP-40, 0.5 mM DTT,
   200 U ml^-1 RNase inhibitor). 10 μg purified RNA was saved as an input
   sample. The beads were then washed eight times with 500 μl ice-cold NT2
   buffer, and mixed with 1 ml TRIzol to save as the IP sample. Input and
   co-immunoprecipitated RNAs were recovered by TRIzol, extraction and
   analyzed by qPCR or RNA-seq.

RIP

   1×10^7 cells were washed twice with 10 ml ice-cold PBS, then lysed in
   400 μl lysis buffer with 400 U ml^-1 RNase inhibitor, and supernatants
   were collected after centrifugation. Cell lysates were incubated with
   A/G magnetic beads (Millipore) coated with 3 μg anti-IGF2BP3 antibody
   (Proteintech, #14642-1-AP) at 4°C for 6 hours. 10% of the supernatant
   from the RIP lysates was saved as input. The following procedure was as
   described in the above m^6A-RIP method.

Luciferase reporter assay

   To evaluate the effect of the 3′ UTR on CDC25A expression, the WT or
   mutant-m^6A sites of the 3′ UTR of CDC25A were inserted behind the
   firefly luciferase (F-Luc) coding region. The pmirGLO-CDC25A-3′ UTR-WT
   and pmirGLO-CDC25A -3′ UTR-mutant were transfected into WT or
   METTL3/IGF2BP3 KD cells for 48 hours. The luciferase activities were
   detected using the Dual-Luciferase Reporter Assay System (Promega,
   USA). Renilla Luciferase (R-Luc) was used to normalize F-Luc activity.

SELECT qRT-PCR

   SELECT qRT-PCR was conducted as previously reported [63]^23. Briefly,
   total RNA was quantified using the Qubit with Qubit™ RNA HS Assay Kit
   (Thermo Fisher Scientific). Total RNA (1000 ng) was mixed with 40 nM up
   primer, 40 nM down primer and 5 µM dNTPs in 17 µl of 1× CutSmart buffer
   (NEB). The RNA and primers were incubated at a temperature gradient:
   90^oC for 1 min, 80^oC for 1 min, 70^oC for 1 min, 60^oC for 1 min,
   50^oC for 1 min and 40^oC for 6 min. The mixture was then incubated
   with 3 µl of 0.01 U Bst 2.0 DNA polymerase, 0.5 U SplintR ligase, and
   10 nM ATP at 40^oC for 20 min, followed by denaturation at 80^oC for 20
   min. Subsequently, a 20 µl qPCR reaction was set up containing 2 µl of
   the final reaction mixture, 200 nM SELECT primers, and 1x Taq Pro
   Universal SYBR qPCR Master Mix (Vazyme, China). Primers for SELECT qPCR
   are listed in Supplementary [64]Table S5. Ct values of samples were
   normalized to Control Ct values. All assays were performed in
   triplicate.

RNA pulldown assay

   OE-19 and SK-GT4 cells were lysed in lysis buffer containing 1%
   protease inhibitor cocktail. 2 nmol biotinylated RNA probes were
   incubated with 200 µg protein from cell extracts at 4^oC for 4 hours in
   with 1× Protein-RNA Binding Buffer (20 mM Tris (pH 7.5), 50 mM NaCl, 2
   mM MgCl[2], 1% Tween-20). 30 μl of washed Pierce™ Streptavidin Magnetic
   Beads (Thermo Scientific™) were added to each binding reaction and
   further incubated at 4^oC for 2 hours and then washed six times with
   Wash Buffer (20 mM Tris (pH 7.5), 10 mM NaCl, 0.1% Tween-20). The
   pull-down proteins were used for immunoblot analyses.

mRNA stability

   The stability of RNA in WT and METTL3 or IGF2BP3 KD cells was assessed
   by incubating cells with 5 μM actinomycin D (Act-D, Catalog #A9415,
   Sigma, USA). Cells were collected at the indicated times and RNA was
   isolated for qRT-PCR.

Immunohistochemistry (IHC)

   All slides were placed in a 60°C incubator for 20 min, de-paraffinized
   in xylene, and rehydrated in gradient ethanol. The slides were
   incubated with 3% hydrogen peroxide for 10 min, followed by antigen
   retrieval using 0.01 M citrate buffer (pH 6.0) for 30 min. After
   blocking with 5% BSA, slides were incubated overnight at 4°C with the
   relevant primary antibody. The secondary biotin-conjugated antibody was
   applied for 1 hour at room temperature. The IHC staining was visualized
   using diaminobenzidine reaction, counterstained with hematoxylin.

RNA-seq

   Total RNA was isolated from IGF2BP3 knockdown or control OE-19 cells
   using TRIzol. Poly(A) RNA was subsequently purified from 50-100 ng
   total RNA using the NEBNext Poly(A) mRNA Magnetic Isolation Module. The
   NEBNext Ultra Directional RNA Library Prep Kit (New England BioLabs)
   was used for library preparation. Each group was sequenced in
   triplicate.

Linear amplification of cDNA ends and sequencing (LACE-seq)

   IGF2BP3 LACE-seq was performed following the previously reported
   protocol[65]^24. Briefly, cells seeded in a 15-cm dish at 80-90%
   confluency were crosslinked by UV on ice at 400 mJ twice. 15 µg of
   IGF2BP3 antibody was used for RNA immunoprecipitation. The
   immunoprecipitated RNA was then fragmented by MNase, dephosphorylated,
   then subjected to reverse transcription, library preparation and deep
   sequencing.

Sequencing data analysis

   For RNA-seq data: RNA-seq was sequenced by Illumina NovaSeq 6000 with
   pair-end 150-bp read length. All clean reads were mapped to the human
   genome version hg38 by using hisat2[66]^25 with default settings. Read
   counts were calculated using featureCounts[67]^26. Differentially
   expressed genes between conditions were statistically assessed using
   the R/Bioconductor package DESeq2[68]^27.

   For RIP/meRIP-seq data: Samples were sequenced by Illumina NovaSeq 6000
   with pair-end 150-bp read length. The RIP-seq reads were mapped to the
   human genome version hg38 using Tophat2[69]^28 with default settings.
   RIP targets were called using MACS2[70]^29 with default settings. Motif
   analysis was performed using HOMER.

   For LACE-seq: Clean reads were aligned to human pre-rRNA using
   Bowtie[71]^30, and the remaining unmapped reads were then aligned to
   the human genome version hg38 using Bowtie with parameters: -v 2 -m 10
   -best -strata; -v 2 -k 10 -best -strata. Peaks were identified by
   Piranha[72]^31 with parameters: -s -b 20 -p 0.01. Motif analysis was
   performed using HOMER.

In vivo animal assays

   OE-19 cells with IGF2BP3 KD or CDC25A OE were injected subcutaneously
   (1 x 10^7 cells/inoculum) into the flanks of five-week-old nude mice.
   Tumor formation/growth was assessed until the experimental endpoint,
   and tumor volume was calculated using the formula: (width)^2 ×
   length/2.

Statistical analysis

   Data were presented as the mean ± Standard Error of the Mean (SEM) or
   Standard Deviation (SD). Two-tailed Student's t tests were performed to
   assess the statistical significance of differences between groups.
   Spearman correlation was performed to analyze the correlation. p < 0.05
   was considered statistically significant. All statistical analyses were
   performed using GraphPad Prism 8.0 or R software (version 4.2.2). *p <
   0.05, **p < 0.01, ***p < 0.001; ns, not significant.

Results

Elevated IGF2BP3 expression correlates with poor prognosis of patients with
AEG

   To explore the expression profile of m^6A regulators in AEG, we
   analyzed published RNA-seq data from 83 AEG patients with paired tumor
   and adjacent normal tissues[73]^6. Compared with methyltransferase-like
   proteins (METTL3, METTL14 and METTL16), demethylases (FTO and ALKBH5),
   and YTHDF proteins (YTHDF1, YTHDF2 and YTHDF3), IGF2BP proteins had a
   higher fold change (FC) in their mRNA levels in AEG tumors versus
   paired normal adjacent tissues (Fig. [74]1A). Notably, among the IGF2BP
   proteins, IGF2BP3 had the highest FC (Fig. [75]1A) and mRNA level (Fig.
   [76]1B). Receiver operating characteristic (ROC) analysis also revealed
   that IGF2BP3 had the highest area under the curve (AUC) (Fig.
   [77]1C-E), which was chosen for further study. qRT-PCR and immunoblot
   also confirmed that the RNA and protein levels of IGF2BP3 were higher
   in 4 out of 5 (80%) AEG tumors than in their normal adjacent tissues
   (Fig. [78]1F-G). Furthermore, high expression of IGF2BP3 was observed
   in multiple cancers in The Cancer Genome Atlas (TCGA) and The Human
   Protein Atlas (HPA) datasets (Fig. [79]1H and I). Kaplan-Meier analysis
   showed that high IGF2BP3 expression was associated with poor survival
   prognosis in patients with AEG and TCGA pan-cancer (Fig. [80]1J and K).
   Taken together, these results revealed that IGF2BP3 is highly expressed
   in AEG and associated with poorer prognosis.

Figure 1.

   [81]Figure 1
   [82]Open in a new tab

   Elevated IGF2BP3 expression correlates with poor prognosis of patients
   with AEG. (A-B) Differential expression analysis of m^6A modulators (A)
   and IGF2BPs (B) in 83 AEG tumors relative to adjacent normal tissue
   (PRJNA788008). (C-E) ROC analysis of IGF2BPs (C-D) in 83 AEG patients.
   FC, fold change. (F-G) mRNA and protein expression levels of IGF2BP3 in
   paired AEG tumors (T) and adjacent normal tissue (N) (n = 5). (H)
   Differential analysis of IGF2BP3 mRNA levels across 23 cancer types in
   TCGA datasets. (I) Expression levels of IGF2BP3 protein across 16
   cancer types in HPA datasets. (J-K) Kaplan-Meier survival analysis of
   overall survival rates with low and high IGF2BP3 mRNA expression in AEG
   patients and TCGA Pan-cancer. *p < 0.05, **p < 0.01, and ***p < 0.001.

IGF2BP3 plays oncogenic role in AEG

   We next determined the role of IGF2BP3 in AEG progression in vitro and
   in vivo. Cell proliferation and transwell assays showed that IGF2BP3 KD
   significantly decreased proliferation, migration, and colony formation
   of OE-19 and SK-GT4 cells (Fig. [83]2A-F). Furthermore, a subcutaneous
   tumorigenesis mouse model showed that IGF2BP3 KD inhibited in vivo
   xenograft tumor growth and mass of OE-19 cells (Fig. [84]2G-I),
   suggesting that IGF2BP3 KD can suppress tumor progression in AEG
   cancer.

Figure 2.

   [85]Figure 2
   [86]Open in a new tab

   IGF2BP3 plays oncogenic role in AEG. (A-B) Cell proliferation assays
   assessing IGF2BP3 KD OE-19 (A) and SK-GT4 (B) cells. (C-D)
   Quantification (left) and representative micrographs (right) of IGF2BP3
   KD OE-19 (C) and SK-GT4 (D) cells in non-coated Transwell assays. (E-F)
   Quantification (left) and representative images (right) of IGF2BP3 KD
   OE-19 (E) and SK-GT4 (F) cells in colony formation assay. (G-I) Effects
   of IGF2BP3 KD OE-19 cells on tumor size (G), tumor growth (H), and mass
   (I) in nude mice (n = 6). (J) KEGG analysis showing the top 20
   significantly enriched pathways from down-regulated genes in IGF2BP3 KD
   OE-19 cells. (K-L) Flow cytometry analysis assays showing the
   percentage of IGF2BP3 KD OE-19 (K) and SK-GT4 (L) cells in sub-G0-G1,
   S, and G2-M cell-cycle phases (n = 3). Representative cell-cycle
   profiles (left) and quantification analysis (right) are shown. PI-A,
   propidium iodide-A. Scale bars, 20 µm. *p < 0.05, **p < 0.01, and ***p
   < 0.001.

   To explore the functional pathways of IGF2BP3 in driving AEG, we
   conducted RNA-seq in IGF2BP3 KD OE-19 cells. The results of
   differential analysis showed 2966 significantly down-regulated and 2715
   significantly up-regulated genes in IGF2BP3 KD cells ([87]Fig. S1A-B).
   Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis
   revealed that the “Cell cycle” is the most significantly enriched
   signaling pathway among down-regulated genes (Fig. [88]2J), with a
   total of 75 genes associated with this pathway ([89]Fig. S1C). Notably,
   23 of these genes have been confirmed to be associated with the
   regulation of the G1-S transition ([90]Fig. S1C and [91]Table S6), of
   which 8 genes, including CCNE2, CDC6, CDK4, CDC25A, CHEK2, E2F1, E2F3
   and WEE1 were selected for qPCR assays and were significantly reduced
   in the IGF2BP3 KD OE-19 cells ([92]Fig. S1D). Indeed, flow cytometry
   assays showed that IGF2BP3 KD inhibited the G1-S transition in both
   OE-19 and SK-GT4 cells (Fig. [93]2K-L). Consistently, based on the
   median value of IGF2BP3 expression in the RNA-seq data of 83 AEG
   patients[94]^6, Gene Set Enrichment Analysis (GSEA) revealed that the
   “Cell cycle” signaling pathway was significantly enriched in the
   high-expression group ([95]Fig. S2A-B). These results indicated that
   IGF2BP3 promoted AEG progression by accelerating the G1-S transition.

IGF2BP3 KD globally down-regulates target gene expression

   IGF2BP3 has been defined as an m^6A reader recognizing and stabilizing
   a large repertoire of mRNA transcripts[96]^15. To understand the
   underlying mechanism of IGF2BP3 in AEG, we performed RIP-seq, LACE-seq
   and meRIP-seq in OE-19 cells. Firstly, we calculated the correlation
   between biological replicates, and confirmed good reproducibility of
   RIP-seq (r = 0.89, P < 2.2e-16), LACE-seq (r = 0.76, P < 2.2e-16), and
   meRIP-seq (r = 0.74, P < 2.2e-16, [97]Fig. S3A). Sequencing purified
   RNA identified 6821 and 7745 potential target genes from RIP and LACE
   samples, respectively ([98]Fig. S3B). Among them, 4240 genes overlapped
   with RIP and LACE-seq targets (Fig. [99]3A), which were considered as
   high-confidence targets. Further, meRIP-seq identified 7557 potential
   m^6A modification genes ([100]Fig. S3B). Both RIP and LACE targets
   preferentially bind to the 'UGGAC' consensus sequence containing the
   'GGAC' m^6A core motif[101]^32 (Fig. [102]3B), and 65% of the high
   confidence targets contain at least one m^6A peak as detected by
   meRIP-seq (Fig. [103]3C). Moreover, most of the IGF2BP3 binding sites
   (79%) are located in protein-coding transcripts (Fig. [104]3D) and are
   highly enriched near stop codons and in 3′ UTRs, coinciding with the
   m^6A distribution[105]^33 (Fig. [106]3E and F). We next integrated the
   meRIP-seq, RIP-seq, LACE-seq, and RNA-seq data. The global transcripts
   were grouped into non-targets and RIP+LACE targets (high confidence
   targets). IGF2BP3 KD globally inhibited the expression of its target
   gene (Fig. [107]3G), while IGF2BP3 KD had no significant effect on the
   expression of targets with or without m^6A group (Fig. [108]3H).

Figure 3.

   [109]Figure 3
   [110]Open in a new tab

   IGF2BP3 KD globally down-regulates target gene expression. (A) Overlap
   of IGF2BP3 target genes identified by RIP-seq and LACE-seq in OE-19
   cells. (B) Top consensus sequences of IGF2BP3 binding sites and the
   m^6A motif detected by HOMER motif analysis with meRIP-seq, RIP-seq and
   LACE-seq data. (C) Pie charts showing the numbers and percentages of
   IGF2BP3 high-confidence target genes containing m^6A peaks. (D)
   Percentages of various RNA species in RIP-seq, LACE-seq, and meRIP-seq.
   (E) Metagene profiles showing enrichment of IGF2BP3 binding sites and
   m^6A modifications across the mRNA transcriptome. CDS, coding sequence.
   UTR, untranslated region. (F) Distribution of IGF2BP3 binding peaks and
   m^6A peaks within different gene regions. (G-H) Cumulative frequency of
   mRNA log[2]FC in non-target genes, IGF2BP3 high-confidence target genes
   (G), and IGF2BP3 high-confidence target genes with or without m^6A (H)
   upon IGF2BP3 KD. P values were calculated using two-sided Wilcoxon and
   Mann-Whitney tests. (I-J) KEGG analysis showing the top 20
   significantly enriched pathways from IGF2BP3 high-confidence target
   genes with m^6A (I) and the overlapping genes from IGF2BP3
   high-confidence target genes with m^6A and downregulated genes in
   IGF2BP3 KD (J).

   To explore the major function of IGF2BP3 targets in AEG, we performed
   KEGG pathway enrichment analysis. Results showed that IGF2BP3 targets
   with m^6A were significantly enriched in the “Cell cycle” signaling
   pathway (Fig. [111]3I). Furthermore, m^6A down-regulated targets in
   IGF2BP3 KD were also significantly associated with the “Cell cycle”
   signaling pathway (Fig. [112]3J). Taken together, these results
   indicated that IGF2BP3 regulated the cell cycle in an m^6A-dependent
   manner.

IGF2BP3 stabilizes the CDC25A transcript in an m^6A-dependent manner

   To investigate whether and how IGF2BP3 affects the cell cycle of AEG
   cells in an m^6A-dependent manner, we first performed an analysis for
   IGF2BP3 m^6A-dependent-bound genes that are involved in the cell cycle
   and down-regulated in IGF2BP3 KD, and found that CDC25A was the most
   significantly down-regulated gene (Fig. [113]4A). Moreover, CDC25A has
   been shown to be a critical regulator of the G1-S transition of
   mammalian cells[114]^34, thus we focused on the CDC25A that as a
   potential m^6A target of IGF2BP3. qRT-PCR and immunoblot assays showed
   that both mRNA and protein levels of CDC25A were significantly
   decreased in both IGF2BP3 KD OE-19 and SK-GT4 cells (Fig.[115]4 B-C).
   In contrast, the OE of IGF2BP3 markedly increased CDC25A protein levels
   (Fig. [116]4D). We further assessed whether the methyltransferase
   METTL3 and demethylases ALKBH5/FTO are involved in the regulation of
   CDC25A expression. Strikingly, METTL3 KD down-regulated CDC25A
   expression (Fig. [117]4E-F), whereas the OE of WT METTL3, but not the
   catalytically inactive METTL3 mutant, markedly increased CDC25A
   expression levels (Fig. [118]4G). In contrast, ALKBH5 and FTO KD
   up-regulated CDC25A expression ([119]Fig. S4A), whereas the OE of WT
   ALKBH5 and FTO, but not the catalytically inactive mutant, markedly
   decreased CDC25A expression levels ([120]Fig. S4B). These results
   suggest that m^6A regulates the expression of CDC25A.

Figure 4.

   [121]Figure 4
   [122]Open in a new tab

   IGF2BP3 stabilizes the CDC25A transcript in an m^6A-dependent manner.
   (A) Differential expression analysis of cell cycle related genes bound
   by IGF2BP3 in an m^6A-dependent manner upon IGF2BP3 KD. (B) qRT-PCR
   analysis of CDC25A in IGF2BP3 KD OE-19 and SK-GT4 cells. (C-D)
   Immunoblot analysis of CDC25A in IGF2BP3 KD (C), and IGF2BP3 OE (D)
   OE-19 (top) and SK-GT4 (bottom) cells. (E) qRT-PCR analysis of CDC25A
   in METTL3 KD OE-19 and SK-GT4 cells. (F-G) Immunoblot analysis of
   CDC25A in METTL3 KD (F), and METTL3 OE (G) OE-19 and SK-GT4 cells.
   (H-I) Half-life of CDC25A after treatment with 5 μM actinomycin D for
   the indicated times in IGF2BP3 (H) or METTL3 (I) KD OE-19 cells. *p <
   0.05, **p < 0.01, and ***p < 0.001.

   To further determine how IGF2BP3 and METTL3 regulate CDC25A expression,
   we treated OE-19 cells with the transcriptional inhibitor actinomycin D
   and assessed the mRNA stability of CDC25A, and found that both IGF2BP3
   and METTL3 KD significantly decreased the half-lives of CDC25A mRNA
   (Fig. [123]4H-I). Collectively, these data demonstrate that IGF2BP3
   stabilizes the CDC25A transcript and increases its expression in an
   m^6A-dependent manner.

Identification of CDC25A m^6A sites

   To determine how IGF2BP3 regulates CDC25A expression through m^6A
   modification, we analyzed IGF2BP3 RIP-seq, LACE-seq, and meRIP-seq
   data. We identified that the m^6A peaks, which coincide well with
   IGF2BP3-binding sites, are enriched in the 3′ UTR of the CDC25A
   transcript (Fig. [124]5A). To verify this finding, we performed
   m^6A-RIP qPCR analysis and found significant m^6A modifications in the
   3′ UTR of CDC25A in OE-19 and SK-GT4 cells ([125]Fig. S5A). These
   modifications were significantly decreased in METTL3 KD OE-19 and
   SK-GT4 cells (Fig. [126]5B). Consistent with these results, RIP qPCR
   analysis revealed that IGF2BP3 is remarkably enriched in the 3′ UTR of
   CDC25A ([127]Fig. S5B), and these enrichments were significantly
   decreased in METTL3 KD OE-19 and SK-GT4 cells (Fig. [128]5C),
   suggesting that IGF2BP3 binds to the 3′ UTR of the CDC25A transcript in
   an m^6A-dependent manner.

Figure 5.

   [129]Figure 5
   [130]Open in a new tab

   Identification of CDC25A m^6A sites. (A) Distribution of peaks across
   the CDC25A transcript based on meRIP-seq, IGF2BP3 RIP-seq and LACE-seq
   data (n = 2). (B) m^6A RIP-qPCR showing the enrichment of m^6A
   modification in the CDC25A 3' UTR in METTL3 KD OE-19 (left) and SK-GT4
   (right) cells. (C) RIP-qPCR showing the enrichment of IGF2BP3 in the
   CDC25A 3' UTR in METTL3 KD OE-19 (left) and SK-GT4 (right) cells. (D)
   CLIP-seq datasets reveal m^6A modification sites in the CDC25A 3' UTR.
   (E) Threshold cycle (Ct) of qPCR showing SELECT results of A2131, A2142
   and A2164 sites of CDC25A in METTL3 KD OE-19 cells. (F) Immunoblot
   detecting the enrichment of IGF2BP3 in RNA pulldown results using
   non-m^6A or m^6A CDC25A 3' UTR probes in OE-19 (top) and SK-GT4
   (bottom) cells. *p < 0.05, **p < 0.01, and ***p < 0.001.

   To further identify the m^6A sites on the CDC25A 3′ UTR, we analyzed
   published single-nucleotide resolution crosslinking-immunoprecipitation
   followed by high-throughput sequencing datasets and identified three
   RRACH sites in the 3′ UTR of CDC25A that are potentially modified by
   m^6A (Fig. [131]5D). Next, we verified the three potential methylation
   sites in vivo using the SELECT method achieving single-base resolution
   accuracy[132]^11^, [133]^23, and found that m^6A levels at these sites
   significantly decreased in METTL3 KD OE-19 cells (Fig. [134]5E). To
   confirm IGF2BP3 binding to these m^6A sites, we performed RNA pulldown
   using biotinylated and m^6A-modified RNA probes. Immunoblot results
   showed specific IGF2BP3 enrichment in the m^6A probe pulldown fraction,
   but not in non-m^6A and mutant probes (Fig. [135]5F). Together, these
   data suggest that the CDC25A 3' UTR contains three m^6A sites required
   for IGF2BP3 binding.

IGF2BP3 increases CDC25A expression through three m^6A sites in its 3' UTR

   To define the effect of the three m^6A sites on CDC25A expression, we
   constructed a pmirGLO-CDC25A 3' UTR luciferase reporter ([136]Fig.
   S6A). mRNA stability and dual-luciferase assays showed that IGF2BP3 and
   METTL3 KD significantly decreased the half-lives of F-Luc mRNA and
   luciferase activities of the CDC25A 3' UTR WT reporter (Fig.
   [137]6A-C). Subsequently, we designed reporters containing mutations
   (MUT) in the three m^6A sites ([138]Fig. S6A) and found that IGF2BP3 or
   METTL3 KD did not affect the half-lives of F-luc mRNA of the CDC25A 3'
   UTR MUT reporter (Fig. [139]6D). However, the mutations in the three
   m^6A sites significantly reduced the half-lives of F-luc mRNA and
   luciferase activities of the reporter compared to the WT reporter (Fig.
   [140]6E-F). These results suggest that IGF2BP3 stabilizes the CDC25A
   transcript through three m^6A sites in its 3' UTR.

Figure 6.

   [141]Figure 6
   [142]Open in a new tab

   IGF2BP3 increases CDC25A expression through three m^6A sites in its 3'
   UTR. (A and D) Half-lives of F-Luc coupling 3' UTR WT (A) and mutant
   (D) of CDC25A after treatment with 5 μM actinomycin D for the indicated
   times in IGF2BP3 and METTL3 KD 293T cells. (B and C) Dual-luciferase
   assays showing luciferase activities of CDC25A 3' UTR in IGF2BP3 or
   METTL3 KD 293T cells. (E) Half-lives of F-Luc coupling 3' UTR WT and
   mutant of CDC25A after treatment with 5 μM actinomycin D for the
   indicated times in 293T cells. (F) Dual-luciferase assays showing
   luciferase activities of the CDC25A 3' UTR WT and mutant in the 293T
   cells. (G-H) RIP-qPCR showing the enrichment of m^6A modification (G)
   and IGF2BP3 (H) in the CDC25A 3' UTR in OE-19 cells transfected with
   dm^6ACRISPR systems of FTO and ALKBH5. (I-J) qRT-PCR and immunoblot
   analysis of mRNA (I) and protein (J) expression of CDC25A in OE-19
   cells transfected with dm^6ACRISPR systems of FTO and ALKBH5. *p <
   0.05, **p < 0.01, and ***p < 0.001.

   We next specifically demethylated the m^6A modification of CDC25A using
   the dm^6ACRISPR-FTO/ALKBH5 system to confirm the effects of m^6A on
   CDC25A expression. Two gRNAs targeting sequences around the three m^6A
   sites were designed to target CDC25A ([143]Fig. S6B). To test the
   efficiency of the gRNAs, we analyzed the mRNA levels of CDC25A in cells
   transfected with gRNAs and WT Cas13b. qRT-PCR analysis showed that the
   two gRNAs co-transfected with WT Cas13b, but not the transfection of
   gRNAs alone or combined with dCas13b, significantly decreased the mRNA
   level of CDC25A ([144]Fig. S6C), suggesting that the two gRNAs can
   efficiently recognize CDC25A. Next, we verified the demethylation
   effects of the dm^6ACRISPR system on CDC25A. m^6A-RIP-qPCR analysis
   showed that both dCas13b-ALKBH5 and dCas13b-FTO combined with gRNAs
   decreased the m^6A level of the CDC25A 3' UTR (Fig. [145]6G). RIP qPCR
   analysis showed that the binding ability of IGF2BP3 to CDC25A was
   significantly decreased after the m^6A of CDC25A were demethylated by
   the dm^6ACRISPR system (Fig. [146]6H). We then tested the effect of
   CDC25A 3′ UTR demethylation on its expression and found that
   dm^6ACRISPR targeting CDC25A led to significant downregulation of
   CDC25A mRNA and protein levels in OE-19 cells (Fig. [147]6I-J).
   Together, these data suggest that the CDC25A 3' UTR contains three m^6A
   sites required for IGF2BP3 binding and CDC25A expression.

IGF2BP3 regulates cell cycle and cancer growth through CDC25A

   Given these findings, we analyzed whether IGF2BP3 regulates the cell
   cycle and cancer growth through CDC25A by performing rescue
   experiments. Flow cytometry assays showed that CDC25A OE prevented the
   decreased G1-S transition in the cell cycle induced by IGF2BP3 KD (Fig.
   [148]7A-B). Furthermore, CDC25A OE partially reversed the reduction in
   cell proliferation and migration caused by IGF2BP3 KD in OE-19 and
   SK-GT4 cells (Fig. [149]7C-H). To further confirm the effects of the
   IGF2BP3/CDC25A axis on AEG progression, we performed in vivo xenograft
   experiments and found that CDC25A OE partially reversed the tumor
   growth reduction caused by IGF2BP3 KD (Fig. [150]7I-K), suggesting that
   CDC25A is involved in IGF2BP3-mediated cell cycle and cancer growth in
   AEG.

Figure 7.

   [151]Figure 7
   [152]Open in a new tab

   IGF2BP3 regulates cell cycle and cancer growth through CDC25A. (A and
   B) Flow cytometry analysis showing the percentage of IGF2BP3 KD with
   CDC25A OE in OE-19 cells in sub-G0-G1, S, and G2-M cell-cycle phases (n
   = 3). Representative cell-cycle profiles (A) and quantification
   analysis (B) are shown. (C-H) Cell proliferation and Transwell assays
   assessing IGF2BP3 KD with CDC25A OE in OE-19 (C-E) and SK-GT4 (F-H)
   cells. (I-K) Effects of IGF2BP3 KD with CDC25A OE in OE-19 cells on
   tumor size (I), tumor growth (J) and mass (K) in nude mice (n = 5).
   (L-M) Correlation between CDC25A and IGF2BP3 (L) or METTL3 (M) in 83
   AEG patients. (N-P) IHC analysis of CDC25A, IGF2BP3 and METTL3 in AEG
   tumor tissues. The quantification analysis is shown (N). Spearman
   correlation analysis between CDC25A and IGF2BP3 (O) or METTL3 (P).
   Normal adjacent tissues (N) = 15; Tumor tissues (T) = 30. (Q-R)
   Spearman correlation between CDC25A and IGF2BP3 (Q) or METTL3 (R)
   across 33 cancer types in TCGA datasets. (S) Proposed model for the
   METTL3/IGF2BP3/CDC25A axis regulates cancer cell growth. Scale bars, 20
   µm. *p < 0.05, **p < 0.01, and ***p < 0.001.

   To further determine the clinical relevance of the
   METTL3/IGF2BP3/CDC25A axis, we analyzed the RNA correlation of these
   genes in RNA-seq data from 83 AEG patients[153]^6. Results showed that
   CDC25A had a strong positive correlation with both IGF2BP3 (r = 0.634,
   P < 0.001) and METTL3 (r = 0.469, P < 0.001) (Fig. [154]7L-M). Next, we
   performed IHC staining in 30 AEG tumors, and 15 paired adjacent normal
   tissues. IHC results showed that METTL3, IGF2BP3, and CDC25A expression
   significantly increased in cancer tissues compared to normal adjacent
   tissues (Fig. [155]7N, and [156]Fig. S7). In particular, a strong
   positive correlation was observed between IGF2BP3 and CDC25A expression
   (r = 0.649, P < 0.001), and between METTL3 and CDC25A expression (r =
   0.683, P < 0.001) (Fig. [157]7O-P). More importantly, CDC25A expression
   was positively correlated with IGF2BP3 and METTL3 expression in most
   cancer types in TCGA datasets (Fig. [158]7Q-R). These results support
   that CDC25A is clinically associated with METTL3 and IGF2BP3 in AEG
   cancer.

Discussion

   The genomic, transcriptomic, proteomic, and phosphoproteomic profiling
   of large cohort of AEG patients had been studied, here, our study
   provides the post-transcriptional regulation of AEG. By performing a
   transcriptome-wide m^6A methylome analysis in AEG, we found that m^6A
   modification plays a key role in regulating AEG tumor progression.
   Specifically, IGF2BP3 mediates m^6A modification on the expression of
   critical genes, such as CDC25A. By positively regulating the mRNA
   stability of these transcripts, IGF2BP3 promotes the cell cycle,
   allowing for the rapid proliferation and survival of AEG cells (Fig.
   [159]7S).

   Among m^6A regulators, METTL3/14, FTO, ALKBH5, and members of the YTHDF
   family have been extensively studied for their contributions to RNA
   regulation and cancer progression[160]^35^-[161]^38. We found that
   IGF2BP3 is the most up-regulated m^6A regulator in AEG tumors,
   suggesting that IGF2BP3 may play a major role in AEG. Recent studies
   have reported that IGF2BP3 binds to certain target genes, such as
   CDK4[162]^21, CCND1[163]^39, and MYC [164]^40, in an m^6A-dependent
   manner to govern cell cycle progression. In this study, we identified a
   broad spectrum of target genes regulated by IGF2BP3 in the cell cycle,
   further elucidating the significant role of IGF2BP3 in regulating the
   cell cycle. Importantly, silencing IGF2BP3 results in a significant
   decrease in AEG progression in vitro and in vivo, highlighting a
   therapeutic window for targeting IGF2BP3 in AEG treatment.

   The cell cycle plays a crucial role in cancer development and
   progression, and aberrant cell cycle regulation is a hallmark of
   cancer[165]^41. CDC25A is a phosphatase enzyme that activates
   cyclin-dependent kinases by removing inhibitory phosphate
   groups[166]^42. This activation leads to cell cycle progression from
   the G1 to the S phase[167]^34. Overexpression of CDC25A has been
   frequently documented in multiple cancer cell lines, and is highly
   associated with malignancy and poor prognosis in cancer
   patients[168]^43^, [169]^44. Several mechanisms regulate the expression
   and activity of CDC25A, including transcriptional[170]^45^, [171]^46,
   post-transcriptional[172]^47^, [173]^48, and
   post-translational[174]^49^, [175]^50 regulation. However, the
   post-transcriptional modification and regulation of the CDC25A gene
   remain unclear. Here, we identified three m^6A sites in the CDC25A 3'
   UTR, that are required for IGF2BP3 binding and CDC25A expression.
   Importantly, these specific m^6A modification sites of CDC25A,
   associated with its expression, could be promising therapeutic targets
   for CDC25A-associated malignancies.

Supplementary Material

   Supplementary figures and tables.
   [176]ijbsv20p4209s1.pdf^ (1.1MB, pdf)

Acknowledgments