Background
N
6-Methyladenosine (m
6A), the most prevalent modification in mRNA [
1,
2], is a dynamic RNA modification installed by “writer” methyltransferase-like 3 (METTL3), methyltransferase-like 14 (METTL14) and Wilms tumor associated protein (WTAP) [
3‐
5]; erased by “eraser” fat-mass and obesity-associated protein (FTO) and α-ketoglutarate-dependent dioxygenase alkB homolog 5 (ALKBH5) [
6,
7]; and recognized by “readers”. Dynamic m
6A modification affects a variety of cellular processes, such as RNA stability, export, splicing or translation [
2]. For instance, m
6A modification promotes the degradation of
Notch1a mRNA in the earliest hematopoietic progenitor cells [
3] while promoting the translation of immediate-early genes in long-term memory [
8]. Therefore, m
6A RNA modifications have attracted increasing attention in the pathogenesis of human disease.
As m
6A modifications play a key role in the maintenance of homeostasis, aberrant m
6A modifications may be an important inducer of tumorigenesis [
2]. Disturbance of m
6A modifications was reported to contribute to the tumorigenesis of glioblastoma, breast cancer and hepatocellular carcinoma [
9]. For example, decreased
METTL3 expression or
METTL14 mutation in endometrial cancer reduces the m
6A modification of AKT pathway-related genes, resulting in the activation of the AKT signaling pathway and contributing to tumorigenesis [
10]. In addition, FTO erases m
6A modification of tumor suppressor genes
MYC/CEBPA, which contributes to the tumor formation of leukemia. Therapeutically, R-2-hydroxyglutarate inhibits the activity of FTO, thereby suppressing leukemia progression. Thus, exploration of the novel m
6A RNA methylation drivers of tumorigenesis is potentially interesting.
Ocular melanoma, including uveal melanoma (UM) and conjunctival melanoma (CM), is the most common primary eye tumor in adults and the 2nd most common melanoma, with a high rate of recurrence and poor prognosis [
11‐
13]. The loss of one copy of chromosome 3 has been frequently identified in ocular melanoma. Previous studies have revealed that mutations in G protein subunit alpha Q (GNAQ) or G protein subunit alpha 11 (GNA11) result in the promotion of cell proliferation and sensitize cells to mitogen-activated protein kinase (MAPK) inhibitors. Furthermore, epigenetic drivers, such as DNA methylation, histone modifications, microRNAs and lncRNAs, also participate in tumorigenesis of ocular melanoma [
14‐
16]. For instance, lncRNA
ROR serves as a decoy oncoRNA that blocks G9a (a key enzyme of histone methylation) binding to the surfaces of target DNA, thereby promoting UM tumorigenesis, while lncRNA CASC15-New-Transcript 1 (
CANT1) inhibits UM progression by simultaneously activating other lncRNAs,
JPX and
FTX. However, the functional role of dynamic tuning of m
6A in ocular melanoma tumorigenesis remains unclear.
We thus aimed to identify the functional role of m6A methylation in malignant ocular melanoma and reveal its potential mechanism in tumorigenesis. We show that m6A methylation significantly inhibits the progression of ocular melanoma cells. Mechanistically, m6A methylation recognized by YTH N6-methyladenosine RNA binding protein 1 (YTHDF1) promotes translation of histidine triad nucleotide-binding protein 2 (HINT2), a tumor suppressor in ocular melanoma. Our study reveals the functional importance of RNA m6A methylation and thereby presents a novel dynamic mechanism of m6A RNA modifications.
Methods
Patient samples
A total of 88 human ocular melanoma tissues and 28 human normal melanocyte tissues were collected for immunofluorescence (IF) from patients of Ninth People’s Hospital, Shanghai JiaoTong University School of Medicine from 2007 to 2017. The histological features of all specimens were evaluated by pathologists according to the standard criteria, and the clinic pathological characteristics of ocular melanoma patients are listed in Additional file
2: Table S2 and Additional file
3: Table S3.
m6A RNA methylation assay
Total RNA was extracted from samples using an EZ-press RNA Purification Kit (B0004). The change in m6A level relative to total mRNA was measured using the m6A RNA Methylation Quantification Kit (Colorimetric) (ab185912) following the manufacturer’s protocol. Each sample was analyzed using 200 ng of RNA isolated from different cells.
Cell lines
The PIG1 human normal melanocyte cell line was kindly provided by the Department of Ophthalmology, Peking University Third Hospital. The human ocular melanoma cell lines, OCM1, OCM1a and OM431, were kindly supplied by Professor John F. Marshall (Tumor Biology Laboratory, Cancer Research UK Clinical Center, John Vane Science Centre, London, UK). The human conjunctival melanoma cell lines, CRMM1 and CM2005.1, were kindly supplied by Prof. Martine J. Jager (Leiden University Medical Center, Leiden, The Netherlands). The HEK293T human embryonic kidney cell line was purchased from the American Type Culture Collection (Manassas, VA, USA). The cell lines used in this study were authenticated by STR profiling.
Cell culture
Human OCM1, OCM1a, OM431 and HEK293T cells were cultured in DMEM (GIBCO). PIG1 and CM2005.1 cells were cultured in RPMI 1640 medium (GIBCO). CRMM1 cells were cultured in Ham’s F-12 K (Kaighn’s) Medium (GIBCO). All mediums are supplemented with 10% certified heat-inactivated fetal bovine serum (FBS; GIBCO), penicillin (100 U/mL), and streptomycin (100 mg/mL) and cells are all cultured at 37 °C in a humidified 5% CO2 atmosphere.
RNA isolation and quantitative real-time PCR
Total RNA was extracted from samples using the EZ-press RNA Purification Kit (B0004), and cDNA was generated using the PrimeScript RT Reagent Kit (Takara). Quantitative real-time PCR using Powerup SYBR Green PCR Master Mix (Life Technologies) was performed using a real-time PCR system (Applied Biosystems).
Western blot analysis
Cells were harvested at the indicated times and rinsed three times with PBS. Cell extracts were prepared with lysis buffer and centrifuged at 13,000 xg for 30 min at 4 °C. Protein samples were separated by 7.5% (wt/vol) sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene fluoride membranes. After blocking with 5% milk for 1 h at room temperature, the membrane was incubated with 2.5 μg/mL of antibody in 5% BSA overnight at 4 °C. The membrane was then incubated with a secondary antibody conjugated to a fluorescent tag (Invitrogen). The band signals were visualized and quantified using the Odyssey Infrared Imaging System (LI-COR, USA).
Plasmid construction
pLKO.1, pCDH and pCMV were used in our study. ShRNA sequences were generated by PCR and then cloned into the pLKO.1 vector. The HINT2 overexpression cassette was generated by PCR and cloned into the pCDH vector and verified by DNA sequencing. The YTHDF1 overexpression cassette was generated by PCR and cloned into the pCMV vector and verified by DNA sequencing.
Lentivirus packaging and generation of stable cell lines
Lipofectamine 3000 reagent (Invitrogen) was incubated with Opti-MEM I Reduced Serum Medium (GIBCO), and HEK239T cells were transfected with 3 mg of plasmid or 6.0 mg of the PsPax plasmid. Eight hours after transfection, the medium was replaced with 10 mL of fresh medium. The supernatant containing the viruses was collected at 48 and 72 h, filtered through a 0.45-mm cellulose acetate filter and used immediately. Viruses carrying a given plasmid were premixed 1:1, and 50 μL of virus was added to 1 mL of serum. Twenty-four hours prior to transfection, tumor cells were seeded at 2.0 × 105 cells per well in a 6-cm dish, and the medium was replaced with virus-containing supernatant supplemented with 10 ng/mL polybrene (Sigma-Aldrich). After 48 h, the medium was replaced with fresh medium. Cells were selected by incubation with 4 mg/mL puromycin (InvivoGen) for 2 weeks and maintained in 1 mg/mL puromycin (InvivoGen).
Cell proliferation/growth assays
Cell proliferation/growth was assessed by CCK8 assays (HY-K0301, MCE) following the manufacturer’s instructions. Briefly, cells were seeded in triplicate in 96-well plates at a density of 2000–10,000 cells/100 mL. Dye solution was added at the indicated time points, and the plates were incubated at 37 °C for 3–4 h before the absorbance was detected at 570 nm.
Apoptosis assays
FITC-Annexin V Apoptosis Detection Kit 1 (BD Biosciences, San Diego, CA) was used following the manufacturer’s instructions. Briefly, cells were washed twice with cold PBS, stained with FITC-Annexin V and PI on ice for 5 min, and subjected to flow cytometric analysis using a BD LSRFortessa analyzer (BD Biosciences).
A volume of 1 mL of complete medium containing 1000 cells was placed in each well of a six-well plate. The plate was stained with 0.25% crystal violet after 1–2 weeks.
Transwell assay
A 24-well transwell system (Corning) with polycarbonate filters (8-μm pores, Corning) was used. The upper compartment contained 10,000 cells suspended in the appropriate medium with 2% FBS; the lower chamber contained 10% FBS. After 1 day of incubation at 37 °C, the cells in the transwell system were stained with 0.25% crystal violet. The cells remaining in the upper transwell chamber were removed, and those that migrated to the lower chamber were photographed and counted.
Cell cycle analysis
A sample of approximately 106 cells was centrifuged at 800 xg for 4 min, washed twice with PBS, resuspended in 1 mL of PBS and fixed in 75% ethanol overnight at − 20 °C. The fixed cells were washed three times with 10 mL of ice-cold PBS, resuspended in 200–400 mL of PBS containing 10 mL of RNase (Qiagen) and incubated at 37 °C for 30 min in the dark. The cells were then subjected to FACS at the Flow Cytometry Facility.
Total RNA was extracted from samples using the EZ-press RNA Purification Kit (B0004). We confirmed the integrity of the RNA using the 2100 Bioanalyzer (Agilent Technologies, USA) and measured the RNA concentration using a Qubit 2.0 fluorometer with a Qubit RNA Assay Kit (Life Technologies, Carlsbad, CA, USA). We then prepared libraries from 100 ng of total RNA using the Illumina TruSeq RNA Sample Prep Kit (San Diego, CA, USA) following the manufacturer’s protocol.
Methylated RNA immunoprecipitation sequencing (MeRIP-seq)
MeRIP was performed as previously described [
17]. Briefly, purified mRNA was randomly fragmented to approximately 100 nucleotides using Ambion RNA fragmentation reagents and then subjected to IP with an anti-m
6A antibody (202,003, Synaptic Systems) and protein A magnetic beads (88,845, Pierce) in MeRIP buffer (150 mM NaCl, 10 mM Tris-HCl, pH 7.4, 0.1% NP-40) supplemented with RNase inhibitor. m
6A-containing mRNA fragments were eluted with m
6A in MeRIP buffer and purified using TRIzol reagent. For MeRIP-seq, two sets of samples were collected for duplicate biological repeats. For MeRIP-qRT-PCR, the same procedures were followed, except that the purified mRNA was fragmented to approximately 200 nucleotides; three biological repeats were conducted. The samples were sequenced with an Illumina HiSeq 4000 platform.
Cross-linking and methylated RNA immunoprecipitation (miCLIP) -SMARTer-m6A-seq
Small-scale, single base-resolution m6A methylome detection was carried out following procedures modified from a previous report. Briefly, 100 ng of mRNA was isolated from the tumor (OCM1, OCM1a, OM431, CRMM1 and CM2005.1) and normal (PIG1) cell lines using the Dynabeads mRNA Purification Kit (Life Technologies, 61,006), fragmented to ~ 100 nucleotides using fragmentation reagent (Life Technologies, AM8740), and incubated with 5 μg of an anti-m6A antibody (Abcam, ab151230) in 450 μL of IP buffer (50 mM Tris, 100 mM NaCl, 0.05% NP-40, adjusted to pH 7.4) under gentle rotation at 4 °C for 2 h. The mixture was transferred to a clear flat-bottom 96-well plate (Corning) on ice and irradiated three times with 0.15 J/cm2 at 254 nm using a CL-1000 Ultraviolet Crosslinker (UVP). After irradiation, the mixture was collected and incubated with 50 μL of prewashed Dynabeads Protein A (Life Technologies, 1001D) at 4 °C for 2 h. After extensive washing twice with high-salt buffer (50 mM Tris, 1 M NaCl, 1 mM EDTA, 1% NP-40, 0.1% SDS, adjusted to pH 7.4) and twice with IP buffer, the samples on the beads were subjected to dephosphorylation with T4 PNK (NEB, M0201 L) at 37 °C for 20 min. The RNA was then eluted from the beads by proteinase K (Sigma, P2308) treatment at 55 °C for 1 h, followed by phenol-chloroform extraction and ethanol precipitation. The purified RNA was subjected to library construction using the SMARTer smRNA-Seq Kit for Illumina (Clontech Laboratories, 635,030) according to the manufacturer’s instructions and sequenced using the Illumina HiSeq X Ten platform.
RNA-binding protein immunoprecipitation (RIP)-qPCR
To examine m6A modification or RNA-binding proteins on individual genes, the Magna RIP™ Quad RNA-Binding Protein Immunoprecipitation Kit (17–704, Millipore, Billerica, MA) was used according to the manufacturer’s instructions. Briefly, 200 mg of total RNA was enriched with antibody- or rabbit IgG-conjugated Protein A/G Magnetic Beads in 500 mL of 1x IP buffer supplemented with RNase inhibitors at 4 °C overnight. RNA of interest was immunoprecipitated with the beads. One-tenth of each fragmented RNA sample was saved as the input control and further analyzed by qPCR.
Sequencing data analysis
For all untreated tumor and normal cell lines, RNA-seq generated paired-end reads with a length of 151 bp, and MeRIP-seq generated single-end reads with a length of 151 bp. Cutadapt software (version 1.18) [
18] was used to trim off the adapter sequences for all raw reads. Reads that contained an ambiguous nucleotide or with lengths less than 18 nt were discarded by Trimmomatic (version 0.36) [
19].
The remaining reads were aligned to the human genome (version hg19) using Hisat2 (version 2.1.0) [
20]. Only uniquely mapped reads with a mapping quality score ≥ 20 were kept for the subsequent analysis for each sample.
For MeRIP-seq, MACS2 software (version 2.0.10) [
21] was used for m
6A peak calling in each MeRIP sample with the corresponding input sample serving as a control. The default options expected for ‘-nomodel, -keep-dul all’ to turn off fragment size estimation and to keep all uniquely mapped reads in MACS2 were set. Software BEDTools’ intersectBed (version 2.27.1) [
22] was used to annotate each peak on the Ensembl (release 72) gene annotation information.
For RNA-seq, the number of reads mapped to each Ensembl (release 72) gene was counted using the software featureCounts (version 1.6.3) [
23] from Subread package. FeatureCounts was applied with default options expect for ‘-s 2, -p’ to inform strand-specific library construction and fragment counting for paired-end reads.
Statistical analysis of differentially expressed genes and genes with differential protein levels
Differentially expressed genes between normal (PIG1) and tumor (OCM1, OCM1a, OM431, CRMM1 and CM2005.1) cell lines were determined using the R-package DEseq2 [
24]. Transcripts with a fold change cutoff > 1.5 or < − 1.5 and a
p-value cutoff < 0.05 were considered significantly differentially expressed genes. Genes with differential protein levels were also determined with the same cutoffs as above.
Gene ontology analysis
Gene Ontology (GO) analysis of specific genes was performed using DAVID (
http://david.abcc.
ncifcrf.gov/). GO terms with
P < 0.05 were statistically significant. Enrichment maps (Fig.
3e, Fig.
4b) were constructed using Cytoscape 3.7.0 installed with the Enrichment Map plugin. Within the enrichment maps, each node represents a GO pathway, and the node size is proportional to the total number of genes in each pathway. The edge thickness represents the number of overlapping genes between nodes. GO pathways of similar functions are sorted into one group, marked with labels and cycles. The number of genes in each cluster is labeled [
3].
Analysis of miCLIP sequencing data
As previously reported [
25], raw read preprocessing was performed. Fastx_clipper from fastx_toolkit (
http://hannonlab.cshl.edu/fastx_toolkit) was used to trim off the adapter sequence of the raw reads. Fastq_filter.pl, a Perl script from the CLIP Tool Kit (CTK) [
26] was used to filter out the low-quality bases, and reads with lengths shorter than 18 nt were discarded. We processed the paired-end data according to the previously reported approach [
25] with the same criteria. For individual replicates, we demultiplexed the forward reads based on 5′ barcodes by performing fastq2 collapse to remove PCR-amplified reads, and we reverse complemented the reverse reads and processed them in the same way as the forward counterparts. Finally, stripBarcode.pl was performed to strip the random barcodes of the remaining reads and move to read headers for subsequent analysis processed by the CIMS pipeline.
BWA software (version 0.7.10) [
27] was used to map the remaining reads to the human genome (version hg19), and an error rate (substitutions, insertions, or deletions) of ≤0.06 per read was allowed by setting parameter ‘bwa aln –n 0.06 –q20’, following the CTK Online Documentation (
https://zhanglab.c2b2.columbia.edu/index.php/CTK_Documentation). The mode of mutation calling was performed as previously reported with minor modifications [
28] to identify the m
6A locus. Program CIMS.pl [
29] was used to determine the coverage of the tag number (k) and mutations (m) for each mutation position. The mutation position with m > 1, m/k ≥ 0.01 and m/k ≤ 0.5 were kept, and only mutation positions within the RRACH motif were determined as m
6A for the subsequent analysis to remove the potential m
6Am modification [
28,
30].
Motif identification within m6A peaks
m
6A peaks were identified by extending 25 nt both downstream and upstream of the m
6A sites. The motifs enriched in m
6A peaks were analyzed by HOMER (version 4.10.3) [
31]. Motif length was restricted to 5 nucleotides. The nearby peaks were merged to one peak by performing BEDTools’ merge (version 2.27.1) [
22]. These peaks were used as target sequences, and background sequences were constructed by randomly shuffling peaks onto total mRNAs in the genome using BEDTools’ shuffleBed (version 2.27.1) [
22].
Immunofluorescence (IF)
Cells adhering to a glass slide were fixed with 4% formaldehyde (Fisher) for 15 min and then blocked with 5% normal goat serum (Vector) with or without 0.1% Triton X-100 in PBS for 60 min at room temperature. Immunostaining was performed using the appropriate primary and secondary antibodies. Nuclei and the cytoskeleton were counterstained with DAPI and phalloidin, respectively. IF staining was performed with the appropriate Alexa Fluor 488 or Alexa Fluor 546 secondary antibody (Invitrogen, 1:1000 dilution). Images were taken with a ZEISS Axio Scope A1 Upright Microscope.
RNA pull-down
Biotin-labeled ssRNA probes were synthesized in vitro by Sangon Biotin (Shanghai) Co., Ltd. In vitro RNA-protein pull-down assay were performed using Pierce™ Magnetic RNA-Protein Pull-Down Kit (Thermo Scientific, 20,164) according to the manufacturer’s instructions. The cell lysate was prepared using standard lysis buffers (Thermo Scientific Pierce IP Lysis Buffer) as suggested. 100 pmol of RNA and 50 μL of magnetic beads were used per sample. Input RNA of each sample was mixed with 1 μl 50% glycerol, separated on the 8% native 1x TBE gel, and visualized by phosphorimaging using the Personal Molecular Imager (Bio-Rad) [
32‐
37].
Luciferase reporter assay
Cells seeded in 6-well plates were transfected with the psiCHECK.2-based luciferase vector fused or not fused to the wild-type or mutated HINT2–3’UTR. Transfection efficiency was quantified by cotransfection with an actin promoter-driven Renilla luciferase reporter. The activities of firefly and Renilla luciferase in each well were calculated using a dual luciferase reporter assay system (Promega). The relative luciferase activity of the HINT2–3’UTR plasmid was further normalized to the signal in cells transfected with the firefly luciferase vector control under the same treatment conditions.
Polysome profiling
We followed previously reported protocols (
https://www.jove.com/pdf/51455/jove-protocol-51455-polysome-fractionation-analysis-mammalian-translatomes-on-genome-wide) with the following modifications. Before collection, 0.1 mg/mL cycloheximide (CHX) was added to the culture medium for 5 min. A sample of 150 million cells from each group was harvested, rinsed in cold PBS with 0.1 mg/mL CHX and quickly frozen in liquid nitrogen before lysis. The lysis buffer was formulated as 10 mM Tris-HCl (pH 7.5), 100 mM NaCl, 30 mM MgCl2, and 0.1 mg/mL CHX with freshly added 1:100 protease inhibitor (Roche) and 40 U/mL SUPERase in RNase Inhibitor (Ambion). The sample was homogenized using a liquid nitrogen grinder. We chose a linear 10 to 40% sucrose gradient according to cell type, and the sample was fractionated into 23 fractions (0.5 mL per fraction) and analyzed using a Gradient Station (BioCamp) equipped with an ECONOUV monitor (BioRad, Hercules, CA) and a Gilson FC203B fraction collector (Mandel Scientific, Guelph, Canada). RNA was purified from fractions 5–22 and subjected to qPCR analysis.
Mice and animal housing
The animal experiments were approved by the Shanghai Jiao Tong University Animal Care and Use Committee and conducted in accordance with the animal policies of Shanghai JiaoTong University and the guidelines established by the National Health and Family Planning Commission of China. Cells were harvested by trypsinization and washed twice with PBS (GIBCO). BALB/c nude mice (male, 6 weeks old) were used for the study.
Tumorigenesis assay in vivo
Approximately 1 × 106 melanoma cells from each group were injected subcutaneously into the right side of the abdomen of BALB/c nude mice (male, 6 weeks old). At 30 days after cell injection, the mice were humanely killed, and the tumors were harvested. Each tumor was fixed in 4% formaldehyde, embedded in paraffin, and examined for tumor formation by histologic analysis of hematoxylin and eosin (H&E)-stained sections. Tumor volume was calculated by the formula V = ab2/2, where a and b are tumor length and width, respectively.
Immunohistochemistry (IHC)
Tissue slides were deparaffinized and rehydrated through an alcohol series, followed by antigen retrieval with sodium citrate buffer. Tumor sections were blocked with 5% normal goat serum (Vector) with 0.1% Triton X-100 and 3% H2O2 in PBS for 60 min at room temperature and then incubated with appropriate primary antibodies at 4 °C overnight. IHC was performed with horseradish peroxidase (HRP) conjugates using DAB detection.
Discussion
Here, we discovered that m
6A mRNA methylation regulates the translation of the tumor suppressor gene
HINT2 and thereby regulates ocular melanoma tumorigenesis (Fig.
7i). One-third of mRNAs are m
6A modified, which may regulate mRNA stability and translation [
2]. A previous study discovered that the “writers” of m
6A methylation, METTL3 and METTL14, could play both oncogenic and tumor suppressor roles in numerous cancers [
40,
41], while oncogenic roles have been reported for “erasers” such as FTO and ALKBH5 [
7,
42,
43]. Here, we found that the decreased global m
6A modification in ocular melanoma promoted tumorigenesis, which resulted from decreased “writers” and upregulated “erasers”. Our study indicates that the disturbance of global m
6A homeostasis is a key driver in the regulation of tumor formation, which depends on the balance of “writers” and “erasers” of m
6A modification.
Notably, the balance of “writers” and “erasers” regulates clusters of gene function, and the AKT pathway is one of the most important targets regulated by m
6A modifications. The expression of some key genes of the AKT pathway, such as
PHLPP2,
PTEN,
Bcl2 and
c-Myc, has been reported to be regulated by their m
6A modifications. These modifications can be written by METTL3 or METTL14 and can be erased by ALKBH5 and regulate the tumorigenesis of leukemia, pancreatic cancer, endometrial cancer and gastric cancer [
10,
30,
44‐
46]. Although we did not find a significant regulation of the AKT pathway by m
6A modifications in ocular melanoma, genes in other biological processes, such as mRNA processing, translation, Hippo-YAP signaling and MAPK signaling, were differentially expressed (Fig.
4b, Additional file
8: Figure S5A-C). These signaling pathways have been reported to be involved in m
6A targets in colorectal cancer and lung cancer and in the inflammatory response [
47‐
49]. Other tumor-related genes, such as
ACAT2,
PIR and
CTBP1, were potentially regulated by m
6A modifications in ocular melanoma. Further studies should be performed to understand m
6A modification systematically.
Notably, the regulation mechanisms of m
6A modifications are complicated, reflected not only in multitudinous targets but also in different regulations in various cells. It should be noted that the key enzyme of m
6A modification may have opposite effects on tumorigenesis in different systems, even within one specific type of tumor. For example, Somasundaram et al. found that
METTL3 acts as an oncogene that promotes tumorigenesis, glioblastoma stem cell maintenance, and radio resistance by regulating SRY-Box 2 (
SOX2) [
50], whereas Shi et al. reported opposite effects, including inhibition of tumorigenesis and glioblastoma stem cell self-renewal/ proliferation via regulation of ADAM metallopeptidase domain 19 (
ADAM19) [
43]. METTL14 inhibits tumor invasion and metastasis by regulating
miR-126 [
51] while promoting tumor cell proliferation and migration by regulating suppressor of cytokine signaling 2 (
SOCS2) [
41]. In addition, the key enzyme of m
6A modification could also be involved in other non-m
6A modification activities. For example, METTL3 and YTH N
6-methyladenosine RNA binding protein 3 (YTHDF3) directly regulate translation by cooperating with eukaryotic translation initiation factor 3 (eIF3), independent of m
6A [
52,
53]. Future systematic studies could focus on the detailed mechanism of m
6A-related proteins regulating tumor fate.
m
6A-modified RNA recognized by different readers could present with different functions. For instance, YTHDF1 mediates nuclear export and translation [
10,
54], YTH N
6-methyladenosine RNA binding protein 2 (YTHDF2) regulates mRNA stability with m
6A [
10,
55] and RNA structural remodeling [
33], and YT521-B regulates sex-specific alternative splicing and subcellular localization of mRNAs with m
6A [
56,
57]. Here, we revealed for the first time that the m
6A modification of
HINT2 is recognized by YTHDF1. Further study could concentrate on the oncogenic role of other readers, such as YTHDF2/3, in regulating tumor formation of ocular melanoma.
HINT2 is a member of the superfamily of histidine triad AMP-lysine hydrolase proteins, which are closely associated with mitochondrial metabolism and tumor suppression [
58]. HINT2 has been reported to expedite Ca
2+ influx into mitochondria from the cytoplasm, which is a hallmark of the mitochondrial apoptosis pathway and an essential condition for early apoptosis [
58]. Loss of HINT2 disturbs mitochondrial lipid metabolism, glucose homeostasis and mitochondrial deformity [
59]. As a tumor suppressor gene,
HINT2 is downregulated in pancreatic cancer and hepatocellular cancer, endometrial cancer and colorectal cancer, preventing the mitochondrial apoptosis pathway and leading to poor survival [
38,
39,
60,
61]. Here, we revealed for the first time that
HINT2 is regulated by m
6A modification, indicating that disturbances in RNA methylation homeostasis result in dysregulation of proliferation and mitochondrial apoptosis, which contributes to the progression of ocular melanoma. Notably, we found that knockdown of
HINT2 only partially rescued the effect of knockdown of
ALKBH5 on ocular melanoma cells, which indicates that there are other cofactors involved in tumorigenesis that are regulated by m
6A modifications.
Uveal melanoma is the most common adult intraocular tumor, with a 5-year survival rate ranging from 71 to 76% [
11]. Additionally, more than half of patients develop metastasis after 5 years [
11], and the median survival time for metastatic UM is only 12 months [
62]. Conjunctival melanoma is also a rare but lethal cancer; nearly 30% of patients die within 10 years [
63]. Although B-raf mutations are rare in UM, activation changes in other MAPK pathway components, such as
GNAQ or
GNA11, are found in 85–95% of patients and are partly responsible for tumorigenesis [
64,
65]. In CM, the
B-raf mutation is present in up to 50% of patients [
66]. Ocular melanoma cells with the activated MAPK pathway have been reported to be modestly sensitive to MAPK/ERK kinase (MEK) inhibitors with or without combination treatment with the protein kinase C (PKC) inhibitor [
67]. Unfortunately, PKC targeting is limited by toxicity, and a completed phase 3 trial showed no clinical benefit [
68]. To date, structure-based selective inhibitors of the m
6A key enzyme have been discovered [
69,
70]. These drugs provide a novel pathway for targeted therapy of ocular melanoma.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.