METTL14 promotes migration and invasion of choroidal melanoma by targeting RUNX2 mRNA via m6A modification

Abstract The modification of N6‐methyladenosine is involved in the progression of various cancers. This study aimed to clarify its regulatory mechanism in the pathogenesis of choroidal melanoma. Expression of methyltransferase‐like 14 in choroidal melanoma or normal choroidal tissues was determined by Western blot and immunohistochemistry. The impacts of methyltransferase‐like 14 on invasion and migration of choroidal melanoma cells were determined using functional and animal experiments. The interaction between methyltransferase‐like 14 and its downstream target was identified by methylated RNA immunoprecipitation and a dual‐luciferase reporter assay. Additionally, Wnt/β‐catenin signalling pathway was evaluated by Western blot. Methyltransferase‐like 14 was upregulated in choroidal melanoma compared to the normal choroidal tissues. Overexpression or knockdown of methyltransferase‐like 14 enhanced or inhibited the invasion and migration of choroidal melanoma cells, respectively, both in vivo and in vitro. Methyltransferase‐like 14 directly targeted downstream runt‐related transcription factor 2 mRNA, depending on N6‐methyladenosine. Additionally, the Wnt/β‐catenin signalling pathway was activated by methyltransferase‐like 14 in choroidal melanoma cells. Our study identified a novel RNA regulatory mechanism in which runt‐related transcription factor 2 was upregulated by enhanced expression of methyltransferase‐like 14 via N6‐methyladenosine modification, thus facilitating migration and invasion of choroidal melanoma cells.

N6-methyladenosine (m6A) is a widely observed modification in various cancers, and abnormal m6A modifications are associated with tumour progression. [4][5][6] Liu et al. reported that mutations in methyltransferase-like 14 (METTL14) or downregulation of METTL3 reduced the m6A modification of genes related to the AKT pathway in endometrial cancer. As a result, the AKT signalling pathway was activated, promoting tumorigenesis. 7 Yu et al reported that insulin-like growth factor 2 mRNA-binding protein 2 (IGF2BP2), an m6A reader in RNA, stabilizes Slug mRNA and that this effect is dependent on m6A. IGF2BP2 showed a positive correlation with epithelial-mesenchymal transition as well as lymphatic metastasis of head and neck squamous cell carcinoma (HNSCCs). 8 Li et al. demonstrated that DIAPH1-AS1 m6A methylation mediated by WTAP was involved in nasopharyngeal carcinoma tumorigenesis and metastasis. 9 Thus, m6A is closely associated with cancer metastasis. CM is highly malignant and easily metastasizes, thus severely threatening health and vision of most cases. Notably, one of the methylated transferases, METTL3, plays a dominant role in ocular melanoma. 10 However, the role of METTL14, another important methyltransferase, in CM has not yet been studied. Thus, the expression and function of METTL14 in CM remain unknown, which, given the severity of the disease, warrants further investigation.
In this study, we compared the expression of m6A methyltransferase METTL14 in CM tissues with normal choroidal tissues. The oncogenic role of METTL14 in CM was evaluated both in vivo and in vitro, while the underlying molecular mechanism was clarified by identifying the critical target of METTL14.  Table S1 and File S1. All participants provided written informed consent, and ethical approval was issued by the ethics committee of The First Hospital of China Medical University (approval No.: AF-SOP-07-1.1-01). The samples were divided into two portions, one was soaked in formalin immediately, and the other kept at −80°C.

| Western blot (WB) assay
Total protein from tissues or cell lysates was extracted using a buffer containing 1% phenylmethylsulfonyl fluoride (PMSF), and protein concentrations were determined using a bicinchoninic acid (BCA) assay kit (Beyotime Institute of Biotechnology). Protein (40 μg/lane) was denatured and isolated by SDS (10%)-polyacrylamide (PAL) electrophoresis (140 V, 50 min). The protein was transferred to a polyvinyl formal (PVF) membrane (350 mA, 90 min) blocked with 5% milk (0% fat), which was then sealed in a container at 37 °C for 60 min. The membrane was subsequently incubated with 5% milk (0% fat) containing the following primary antibodies at 4 °C for 12 h: Membranes were rinsed three times using Tween-20 buffer for 5 min each and incubated using the corresponding secondary antibodies at 37°C for 1 h. Density was measured using ImageJ software (National Institutes of Health), and protein bands were normalized to β-actin.

| Immunohistochemistry (IHC)
Sections of CM and normal tissue samples were embedded in paraffin, followed by deparaffinization and rehydration. The slides were incubated overnight at 4 °C with the primary antibodies anti-METTL14 (1:1000, HPA038002, Sigma-Aldrich) and anti-RUNX2 (1:1000, MAB2006, RD), followed by addition of secondary antibodies to an avidin-biotin-peroxidase complex (e.g. biotinylated anti-rabbit and anti-mouse antibodies in goats). The slides were kept at 25°C for 60 min and successively stained with DAB reagent and counterstained with haematoxylin. Images were captured using an inverted microscope (EVOS XL system, AMEX1200; Life Technologies Corp) (magnifications = 200× and 40×).

| Cell lines and cell culture
The human CM cell lines OCM1 and MUM-2B were purchased from the Chinese Academy of Sciences Type Culture Collection Cell Bank. The passage time was ≤180 days with mycoplasma elimination performed by the company. Using six-well plates, we cultured OCM1 cells in RPMI medium (Hyclone; GE Healthcare) and 10% FBS, and MUM-2B cells were allowed to proliferate in DMEM medium (Hyclone; GE Healthcare) and 10% FBS. All CM cell lines were maintained at 37°C in 5% CO 2 . At a cell fusion rate of 90%, cells were incubated with trypsin (1 ml) for 5 min and neutralized with growth medium (1 ml). Finally, they were centrifuged and passaged.

| RNA extraction and qRT-PCR
Extraction of total RNA from OCM1 and MUM-2B cells was con-

| Transwell assay
Transfected cells (1 × 10 6 ) in growth medium (200 μl, containing no serum) were transferred to the upper chamber of a transwell plate, and another 600 μl of the medium was added to the bottom chamber. Cells were incubated at 37°C in 5% CO 2 for 1 day. Using a swab, we removed residual cells from the upper chamber and fixed the cells by incubation with 4% paraformaldehyde for 600 s. Next, a crystal violet stain was added, and the cells were further incubated for 10 min. Images were collected using an optical microscope and analysed using ImageJ (National Institutes of Health). Additionally, the transwell filter was coated with Matrigel (BD) and dried overnight for invasion assays.

| Lentivirus transfection
METTL14 and RUNX2 knockdown and overexpression plasmids were purchased from GeneChem. All the procedures were performed in accordance with manufacturer's instructions. Selection of stable transformants was conducted over two weeks using puromycin (5 μg/ml) and confirmed by WB assay for METTL14 and RUNX2.
Cells with confirmed knockdown and overexpression were stored for further tests.

| Animal experiments
Female BALB/c nude mice (n = 15) aged 4-6 weeks were provided by a commercial contractor and housed in a pathogen-free environment.
In the analysis of METTL14-mediated metastasis, we injected 150 μl

| Characterization by PET/CT imaging
Tumour-bearing mice (n = 5 per group), identified using METIS PET, were injected with ~100 μCi [18F] fluoro-D-glucose via the tail vein. Anaesthesia was conducted using isoflurane (3% and 2% for induction and maintenance, respectively, in pure O 2 ), and mice were placed in a prone position in the centre of the scanner.
The obtained PET and CT metadata were reconstructed, and each image was statistically analysed. An organ shape was contoured on each slide, which comprises a part of this organ in the fused PET-CT image.

| The Cancer Genome Atlas (TCGA) data
We conducted a bioinformatic analysis using data from the TCGA database. Survival curves for CM patients with different levels of RUNX2 were selected using a gene expression profiling interactive analysis (GEPIA) (http://gepia.cance r-pku.cn/) based on a suitable expression threshold.

| Methylated RNA immunoprecipitation (MeRIP)
We evaluated the abundance of specific mRNA transcripts in m6A immunoprecipitation and input groups by qPCR using a MeRIP kit (BersinBio). RNA was randomly divided into 100 nucleotide fragments, and immunoprecipitation was performed with anti-m6A or anti-rabbit IgG antibodies linked by A/G magnetic beads.
A magnetic frame was used to elute the m6A-precipitated RNA.

| Plasmid development and dual-luciferase reporter assay
Cells were seeded into individual wells of a 24-well plate. Wild-type/

| Statistical analysis
All tests were executed in triplicates, and all data were described as mean ± standard deviation (SD). Software GraphPad Prism of version 8.0 was employed for statistical analysis. Inter-group differences were assessed using a Student's t-test. The differences in expression levels between paired samples were assessed using a Wilcoxon signed-rank test. Survival rates were estimated using the Kaplan-Meier log-rank method.

| METTL14 is overexpressed in CM tissues
The results of WB assay revealed that the protein of METTL14 was drastically upregulated in CM tissues than that in normal choroidal tissues ( Figure 1A). Expressions of METTL14 in CM and normal choroidal tissues were detected by IHC assays. According to Figure 1B

| METTL14 promotes migration and invasion of CM cells both in vivo and in vitro
To assess the role of METTL14 in CM, METTL14 was knocked down in human CM cell lines; qRT-PCR revealed that METTL14 mRNA was downregulated as a result of transfection with si-METTL14-1 and si-METTL14-2 ( Figure 2A). Additionally, WB analysis revealed that METTL14 protein levels were downregulated as a result of transfection ( Figure 2B). In the transwell assays, METTL14 depletion markedly reduced the migratory and invasive abilities of OCM1 and MUM-2B ( Figure 2C). The WB analysis demonstrated that METTL14

| RUNX2 is an essential METTL14 target gene in CM
The expression of transcription factor RUNX2 is upregulated in melanoma cells 14 and mediates cell migration and invasion in many cancers, including prostate, colorectal and gastric cancer. [15][16][17] However, the expression of RUNX2 has not been described for CM and RUNX2 potentially has m6A sites. 18 In the TCGA database, METTL14 expres-  Figure 6A). According to data from m6A databases, SRAMP was located at the base of the 3′-UTR; 1660A is the m6A motif; and GGAC in the base sequence was mutated to GGCC. A schematic diagram showing the methylation site on RUNX2 and mutation at this site is presented in Figure 6B-D.
The RUNX2 3′-UTR-reporter luciferase assay revealed that overexpression of METTL14 enhanced the luciferase activity of constructs in the wild-type RUNX2 3′-UTR, but not in the mutated RUNX2 3′-UTR sequence ( Figure 6E). It is also demonstrated that RUNX2 mRNA can be methylated by METTL14.

| RUNX2 affects migration and invasion of CM cells in vitro and knockdown of RUNX2 inhibits lung metastasis in nude mice models
Low overall survival was associated with high levels of RUNX2 ( Figure 6F). The protein expression of RUNX2 in CM tissues was upregulated compared to normal choroidal tissues ( Figure 6G).
According to the IHC assay, RUNX2 expression in CM tissues was upregulated compared to normal choroidal tissues ( Figure 6H). The transwell assay revealed that the depletion of RUNX2 reduced

| RUNX2 overexpression partially restores invasion and migration of cells induced by decreased METTL14
We performed transwell assays to evaluate the effects of RUNX2 overexpression in cell invasion and migration ability under reduced levels of METTL14. RUNX2 overexpression partially restored the migration and invasion abilities of CM cells ( Figure 7B,C), suggesting that RUNX2 is a dominant effector by which METTL14 promotes migration and invasion.

| DISCUSS ION
CM is a complex disease with high rates of metastasis, and no effective therapies have been described to date. 10 m6A is known to affect pathological processes, as well as physiological functions, particularly during tumorigenesis, 19,20 with m6A dysregulation observed in multiple cancers. 6,21,22 It has been demonstrated that abnormal epigenetic regulation of gene function via m6A modification plays a crucial role in human tumorigenesis and cancer progression. 6,9,23  reported that the expression of METTL3 was comparatively reduced in ocular melanoma tissues. 10 Our study demonstrated that METTL14 favours migration and invasion of CM cells, which is not contradictory to the activity of METTL3; for example, METTL3 and METTL14 can play opposing roles in the regulation of hepatocellular carcinoma. 36 Ours is the first study to demonstrate that METTL14 has an oncogenic role and favours migration and invasion as a tumour promoter in CM, and thus could serve as a novel biomarker.
RUNX2 is a key regulator of osteogenesis and essential for normal skeletal and bone development. 37 RUNX2 expression is tightly regulated in normal osteoblastic cells. [38][39][40] In details, expression of RUNX2 protein is dramatically high in G0 stage, low at G1/S transition, low during G2 and M phases, and high in early G1 stage. 39,40 RUNX2 favours a quiescent state and inhibits cell proliferation. This reveals a novel aspect of epigenetic alterations associated with CM and provides promising targets for novel interventional therapies. RUNX2 upregulation favours the migration of T-ALL cells and F I G U R E 7 Role of METTL14 and RUNX2 in metastasis. (A) Knockdown of RUNX2 inhibits lung metastasis in nude mice models. WB verified the successful construction of knockdown stabilized MUM-2B cells. WT and RUNX2 knockdown MUM-2B stable cells were injected via the tail vein. Representative images of metastatic lung tumours. (B, C) Transwell assay demonstrates the effects of METTL14 and RUNX2 on the migration and invasion capability of CM cells. Overexpression of RUNX2 partially reversed the reduction in cell migration and invasion caused by decreased METTL14 expression. All data are presented as mean ± SD from three independent replicates. Student's t-test was used to assess inter-group differences.
progression in leukaemia, 42 while it regulates progression in renal cell carcinoma. 43 In addition, FTO regulates RUNX2 mRNA expression via its demethylation of m6A. 18 In the present study, RUNX2 served as an essential target gene for METTL14 in CM and was dependent on m6A. The role of RUNX2 in tumour metastasis, development and progression has been widely reported. Since the first report describing the role of RUNX2 in the differentiation and migration of osteoblasts into chondrocytes and its engagement in the proinvasive/promigratory behaviour of various tumour cells, especially in melanoma cells, RUNX2 has been established as a key factor in metastasis. 44

ACK N OWLED G EM ENT
None.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

CO N S E NT FO R PU B LI C ATI O N
Written informed consent for publication was obtained from all participants.

DATA AVA I L A B I L I T Y S TAT E M E N T
The datasets generated/analysed in the present study are available upon reasonable request from the corresponding author.