The m6A methyltransferase METTL3 promotes bladder cancer progression via AFF4/NF-κB/MYC signaling network

Abstract

N6-methyladenosine, known as m6A, represents the most prevalent modification found within eukaryotic messenger RNAs and plays crucial roles in numerous biological processes. However, its specific functions within bladder cancer remain largely unclear. In this study, we discovered that methyltransferase-like 3, a major RNA N6-adenosine methyltransferase often referred to as METTL3, was significantly elevated in human bladder cancer samples. When METTL3 expression was reduced through knockdown techniques, we observed a substantial decrease in bladder cancer cell proliferation, invasion capabilities, and survival both in laboratory settings and in animal models of tumor growth. Conversely, when METTL3 was overexpressed, we found a significant promotion of bladder cancer cell growth and invasion. Through a combination of transcriptome sequencing, m6A sequencing, and m6A methylated RNA immunoprecipitation followed by quantitative reverse-transcription polymerase chain reaction, we successfully mapped the profile of METTL3-mediated m6A modification in bladder cancer cells for the first time. Our analysis further identified AF4/FMR2 family member 4, two key regulators of the NF-κB pathway (IKBKB and RELA), and MYC as direct targets of METTL3-mediated m6A modification. Additionally, we demonstrated that beyond its role in the NF-κB pathway, AFF4 binds to the promoter region of MYC and enhances its expression, suggesting a novel, multi-layered regulatory network operating downstream of METTL3. Overall, our findings have revealed an AFF4/NF-κB/MYC signaling network that is modulated by METTL3-mediated m6A modification, providing new insights into the mechanisms underlying bladder cancer progression.

Introduction

Urothelial carcinoma of the bladder stands as the fourth most frequently diagnosed malignancy among men, with an estimated 81,190 new cases and 17,240 deaths projected for the year 2018 in the United States. A significant proportion of these cases are diagnosed at advanced stages, rendering patients ineligible for curative surgical interventions. Chemotherapy has served as the primary treatment modality, employing agents such as gemcitabine, cisplatin, oxaliplatin, capecitabine, and 5-fluorouracil. However, despite recent advancements in incorporating targeted therapies and immunotherapy into clinical practice, the responses to these treatments remain limited, exhibiting only a modest impact on overall survival. This limitation is largely attributed to the development of both inherent and acquired drug resistance. Consequently, a deeper understanding of the mechanisms driving bladder cancer progression and the identification of novel therapeutic strategies for this disease represent urgent and unmet medical needs. N6-Methyladenosine, abbreviated as m6A, is recognized as the most prevalent internal chemical modification within messenger RNAs in eukaryotic organisms. In mammalian cells, the introduction of this modification is catalyzed by a methyltransferase complex composed of several proteins, including methyltransferase-like 3, METTL14, Wilms tumor 1 associated protein, VIRMA, and RBM15. Conversely, the removal of m6A methylation is facilitated by two mammalian RNA demethylases: the fat mass and obesity-associated protein and alkylation repair homolog protein 5, highlighting the dynamic nature of m6A methylation. In human cells, thousands of messenger RNAs undergo m6A modification, which has been implicated in various aspects of RNA metabolism, including splicing, nuclear export, stability, and translation, as well as microRNA processing. Recent investigations have demonstrated that m6A regulation plays significant roles in a multitude of biological processes, such as development, metabolism, fertility, osteoporosis, stemness maintenance, and differentiation. It also contributes to synaptic function, axon regeneration in neurons, and the facilitation of hippocampus-dependent learning and memory. Furthermore, the suppressive functions of regulatory T cells and innate immunity within the immune system have also been shown to be regulated by messenger RNA m6A modification. As the critical methyltransferase responsible for RNA m6A modification, METTL3 has been reported to promote the progression of various cancers, including lung cancer, liver cancer, breast cancer, and myeloid leukemia, as well as contributing to chemo- and radioresistance in pancreatic cancer cells in recent studies. Additionally, METTL3-mediated m6A modification is also involved in the maintenance of embryonic stem cells and, perhaps not surprisingly, cancer stem cells. Nevertheless, there is also evidence suggesting that METTL3 can act as a tumor suppressor in glioblastoma, indicating that its role in cancer progression may be context-dependent. However, the specific role of METTL3 in bladder cancer has not been previously investigated. In the present study, our aim was to elucidate the biological function of METTL3 in the development and progression of bladder cancer and to explore the underlying molecular mechanisms by identifying its critical messenger RNA targets.

Results

METTL3 expression is elevated in bladder cancer tissue

To gain insight into the role of m6A modification in bladder cancer progression in living organisms, we measured messenger RNA m6A levels in two human bladder cancer samples. When compared to the corresponding para-tumor bladder urothelial tissues, the m6A levels in both tumor tissues were significantly higher. We then examined the expression of m6A writer and eraser enzymes in the Cancer Genome Atlas datasets to determine which subunit might play a crucial role in the deregulation of m6A in bladder cancer. We found that METTL3 messenger RNA expression was significantly elevated in bladder cancer compared to normal tissues. In contrast, the expression levels of METTL14, WTAP, FTO, or ALKBH5 did not show significant changes in these patient samples. We further examined the protein level of METTL3 in samples from a cohort of 22 bladder cancer patients using immunohistochemistry. Consistent with the Cancer Genome Atlas data, METTL3 protein expression was significantly higher in tumor samples compared to para-tumor samples. Additionally, we examined tissue samples from a previously established N-butyl-N-4-hydroxybutyl Nitrosamine-induced bladder carcinogenesis mouse model and found that METTL3 expression in mice with invasive bladder cancer was dramatically elevated compared to normal bladder tissue. Based on this evidence, we concluded that METTL3 is frequently upregulated in bladder cancer and may be involved in the development and progression of this disease.

METTL3 Promotes Bladder Cancer Cell Proliferation, Migration and Suppresses Apoptosis in vitro

Given our findings that METTL3 is abnormally expressed in bladder cancer tissue, we were motivated to determine if METTL3-regulated m6A modification plays a role in bladder cancer tumorigenesis. We reduced METTL3 expression using two distinct small interfering RNAs in the bladder cancer cell line 5637. The reduction in METTL3 expression by both small interfering RNAs was confirmed using quantitative reverse-transcription polymerase chain reaction and western blot analysis. As anticipated, the depletion of METTL3 in bladder cancer cells led to a decrease in the overall messenger RNA m6A level and also resulted in a strong inhibition of cancer cell growth, an increase in cell apoptosis, and a decrease in the invasive ability of bladder cancer cells. The effects of METTL3 were also confirmed in another bladder cancer cell line, UM-UC-3. Similarly, cell growth and invasion were decreased, while cell apoptosis was increased upon METTL3 knockdown. We next investigated how overexpressed METTL3 affects the growth and invasion of bladder cells. Immortalized human uroepithelial cells were transduced with a lentivirus expressing either a control vector or METTL3. Messenger RNAs and proteins were isolated from both cell types, and reverse transcription polymerase chain reaction analysis and western blot confirmed the overexpression of METTL3 in cells transduced with the METTL3-expressing lentivirus. In contrast to the knockdown experiments, ectopic expression of METTL3 promoted cell growth and invasion. Taken together, our data demonstrated the oncogenic role of METTL3 in bladder cancer cell growth, survival, and invasion.

Identification of METTL3 targets by high-throughput RNA-Seq and m6A-Seq

To identify potential messenger RNA targets whose m6A levels are influenced by METTL3 in bladder cancer cells, we conducted transcriptome sequencing and m6A-sequencing to investigate gene expression changes and map m6A modifications in 5637 cells where METTL3 expression was reduced. RNA-seq analysis revealed that the expression of 1759 genes was significantly decreased, while 1793 genes were significantly increased (fold change greater than 2.0) upon METTL3 knockdown. Gene set enrichment analysis indicated that multiple signaling pathways were positively or negatively correlated with METTL3 depletion. Notably, the MYC target genes and the TNF-α/NF-κB pathway target genes showed a significant negative correlation with METTL3 knockdown. Global profiling of m6A-modified genes using methylated RNA immunoprecipitation sequencing identified 3,588 m6A peaks within 2,537 genes in 5637 cells. Consistent with prior research, the GGAC motif was found to be highly enriched in the immunoprecipitated RNA, and metagene analysis showed that m6A peaks were predominantly located near stop codons, with a subset found in the 5′-untranslated region and internal exons. When examining m6A peaks in the key regulators of pathways identified, we found significant m6A enrichment near the stop codon and downstream 3′-untranslated region of MYC, AFF4, RELA, and IKBKB, which are involved in the MYC and NF-κB pathways, respectively. Quantitative reverse-transcription polymerase chain reaction using primers designed to amplify either the m6A peak region or a control region without an m6A peak was used to validate the methylated RNA immunoprecipitation sequencing data. Further quantitative reverse-transcription polymerase chain reaction analysis confirmed that reducing METTL3 expression using small interfering RNA resulted in significantly decreased levels of MYC and AFF4 messenger RNAs, while having minimal effect on the abundance of RELA or IKBKB messenger RNA. However, protein levels determined by western blot analysis showed that the expression of all four genes was substantially reduced upon METTL3 knockdown. Additionally, protein levels of MYC, AFF4, and RELA were also elevated in both human and mouse bladder cancer samples. These findings strongly suggest that MYC, AFF4, RELA, and IKBKB are direct targets of METTL3.

MYC, AFF4, RELA and IKBKB are functionally important target genes of METTL3 in bladder cancer

When we introduced specific small interfering RNAs targeting METTL3 into human bladder cancer cells, we observed a significant inhibition of the activities of the MYC and NF-κB pathways, as determined by dual luciferase reporter assays. To further understand the mechanism by which METTL3 regulates cell survival, proliferation, and invasion of bladder cancer cells through NF-κB signaling, we investigated the expression of several NF-κB target genes known to be associated with these cellular behaviors in bladder cancer. Specifically, the messenger RNA levels of BCL2A1, BIRC3, KITLG, MMP9, and PLAU were reduced after METTL3 knockdown. Furthermore, similar to the effects of METTL3 knockdown, when 5637 cells were transfected with small interfering RNAs targeting MYC, AFF4, RELA, and IKBKB individually, their growth and proliferation, as well as their ability to invade, were inhibited, and apoptosis was induced. Taken together, our results demonstrated that MYC, AFF4, RELA, and IKBKB are functionally important targets of METTL3 and play a significant role in METTL3-mediated promotion of bladder cancer cell proliferation, invasion, and survival.

AFF4 directly regulates MYC gene expression in bladder cancer cells

The roles of both the MYC and NF-κB pathways in bladder cancer have been well-established. However, the involvement of AFF4 in bladder cancer initiation and progression has not been previously investigated. AFF4 is a crucial component of the super elongation complex, which is involved in regulating the transcription elongation of many oncogenic genes, including MYC. To identify potential targets of AFF4 and the super elongation complex in bladder cancer cells, we performed transcriptome sequencing in 5637 cells where AFF4 expression was reduced. Interestingly, gene set enrichment analysis suggested an enrichment of MYC target genes that were downregulated in AFF4 knockdown cells, which aligns with the results observed upon METTL3 knockdown in bladder cancer cells. To determine whether MYC is directly regulated by AFF4 in bladder cancer cells, we further confirmed the expression of MYC in response to AFF4 knockdown using quantitative reverse-transcription polymerase chain reaction and western blot analysis. The results showed that reducing AFF4 expression led to a decrease in MYC expression at both the messenger RNA and protein levels. Furthermore, we found that AFF4 directly bound to the MYC promoter, as determined by chromatin immunoprecipitation assays, and this binding was significantly decreased after AFF4 knockdown. The above results suggest that AFF4 may be involved in bladder cancer tumorigenesis by acting as a direct upstream regulator of MYC.

Oncogenic role of METTL3 relies on its methyltransferase activity

METTL3 has been reported to function independently of METTL14 and can promote the translation of specific messenger RNAs in vitro regardless of its catalytic activity. To investigate whether the m6A catalytic activity is responsible for the oncogenic role of METTL3, we established a stable METTL3 knockdown model in 5637 cells using a specific short hairpin RNA. We then attempted to restore its function by transfecting a METTL3 complementary DNA containing nucleotide substitutions that made it resistant to the short hairpin RNA, allowing for the expression of Flag-tagged wild-type METTL3 or a catalytically inactive mutant METTL3 protein in these stable knockdown cells. As expected, western blotting and quantitative reverse-transcription polymerase chain reaction showed a significant reduction in the endogenous expression of MYC, AFF4, RELA, and IKBKB upon METTL3 knockdown, without affecting the messenger RNA abundance of RELA and IKBKB. Expression of the wild-type METTL3 resistant to the short hairpin RNA rescued the expression of all four target genes, whereas the catalytically inactive mutant METTL3 failed to restore target gene expression. Moreover, the phenotypes of cell growth, invasion, and apoptosis could also be rescued by the wild-type METTL3 resistant to the short hairpin RNA but not by the catalytically inactive mutant METTL3. These data indicate that the oncogenic role of METTL3 in bladder cancer cells is dependent on its methyltransferase activity.

Knock down of METTL3 diminish tumorigenicity of bladder cancer cells in vivo

To investigate the functional roles of METTL3 in living organisms, we performed a subcutaneous implantation experiment in immunodeficient mice. To minimize bias between individual mice, 5637 control cells or 5637 cells with stable METTL3 knockdown were subcutaneously injected at two sites on the backs of eight mice in each group. Tumor volume was monitored weekly, and tumor mass was weighed on day 35 at the end of the study. Stable knockdown of METTL3 effectively suppressed tumor growth, as evidenced by the significant reduction in tumor size and weight compared to the control group with non-targeting short hairpin RNA. Further confirmation of METTL3′s role in bladder cancer came from immunohistological analysis of Ki67, an indicator of cell proliferation, active Caspase-3, an indicator of cell apoptosis, and the METTL3 targets MYC, AFF4, and RELA in the tumor tissues. Knockdown of METTL3 reduced cell proliferation but induced cell apoptosis in bladder cancer tumors, and also led to the expected changes in the protein levels of MYC, AFF4, and RELA in the tumor tissue sections. All these results strongly support the conclusion that METTL3 and its downstream targets play a pivotal role in promoting bladder cancer progression in living organisms.

Discussion

M6A, the most prevalent chemical modification in human messenger RNA, has been reported to be crucial for cancer progression. However, the functions of m6A in cancer progression can be complex and sometimes contradictory due to the varied distribution of m6A on messenger RNAs, the diverse readers that respond to these m6A modifications, the specific cellular context of different cancers, and the target genes that regulate different cellular processes. For example, both the m6A writers (METTL3 and METTL14) and erasers (FTO and ALKBH5) have been reported to act as oncogenes in certain cancers like acute myeloid leukemia and glioblastoma, whereas METTL3 and METTL14 can also function as tumor suppressors in glioblastoma and hepatocellular carcinoma, respectively. Therefore, a detailed understanding of the mechanisms underlying m6A modification is essential to clarify its role in specific cancer types. In this study, we first demonstrated that METTL3, a major m6A writer, rather than METTL14 or m6A erasers, was abnormally expressed in both human and mouse bladder cancer. We then functionally established the essential role of METTL3 in promoting bladder cancer growth and survival using both in vitro and in vivo models. Importantly, we identified key regulators in the NF-κB pathway (IKBKB and RELA), the Myc pathway (MYC), and RNA elongation (AFF4) as direct downstream targets of METTL3-mediated m6A modification in bladder cancer. Collectively, our study highlights the functional significance of m6A methylation and the associated proteins in bladder cancer, suggesting that the m6A messenger RNA methylation machinery represents promising therapeutic targets for this disease. METTL3-mediated m6A modification has recently been found to regulate numerous genes in various cancer types, yet its targets in bladder cancer remained unknown. In this study, we employed a combination of m6A sequencing and transcriptome messenger RNA sequencing to provide the first comprehensive profiling of METTL3-mediated m6A modification in bladder cancer. Consistent with reported m6A modification of MYC messenger RNA in normal hematopoietic cells, leukemia cells, and glioma, the expression of MYC was also regulated by METTL3-mediated m6A modification in bladder cancer. However, unlike METTL14 and FTO, which promote the stability of MYC messenger RNA, METTL3-mediated m6A modification was thought to enhance the translational efficiency of MYC messenger RNA by affecting m6A abundance at different sites. Our results indicated that the reduction of METTL3 in bladder cancer cells decreased the stability of MYC transcripts by altering m6A abundance primarily around the stop codon and 3′-untranslated regions. It has been reported that a specific cis-acting element of approximately 250 nucleotides, termed the coding region instability determinant, exists in the 3′-region of MYC and is required for regulating MYC messenger RNA stability. The m6A reader IGF2BP preferentially recognizes and binds to the m6A-modified coding region instability determinant region of MYC messenger RNA, thereby stabilizing MYC messenger RNA and promoting translation. Meanwhile, AFF4 might share a similar regulatory mechanism involving METTL3-mediated m6A modification with MYC, as both exhibited reduced messenger RNA and protein expression upon METTL3 depletion. On the other hand, although the protein levels of IKBKB and RELA were efficiently reduced upon METTL3 knockdown, minimal effect was observed on their messenger RNA levels. Indeed, it has been reported that the m6A reader YTHDF1 can selectively recognize m6A-modified messenger RNAs and enhance translation efficiency, and the messenger RNAs of both IKBKB and RELA have been identified as direct targets of YTHDF1 via a photoactivatable ribonucleoside crosslinking and immunoprecipitation assay, suggesting that METTL3 might primarily promote the expression of IKBKB and RELA by regulating their translational efficiency. The oncogene MYC is known to be aberrantly expressed in bladder cancer, and the antitumor effects of MYC inhibitors have already been investigated. In fact, MYC is upregulated in the majority of all human cancers. The broad oncogenic effects of MYC rely on its ability to trigger the expression of target genes that promote cell proliferation, cell survival, and stemness maintenance. Mechanisms of MYC deregulation in bladder cancer include DNA mutation, signal transduction and transcriptional regulation, and microRNA-mediated post-transcriptional regulation. NF-κB is a well-known upstream regulator of MYC expression, through which NF-κB signaling enhances the proliferation and survival of cancer cells during the development and recurrence of bladder cancer. Furthermore, previous studies have provided evidence that MYC is one of the direct targets of AFF4/SEC, and the recruitment of SEC to the MYC gene regulates its expression in different cancer cells. Inspired by the strong correlation between AFF4 and cancer progression in acute lymphoblastic leukemia and head and neck squamous cell carcinoma, we hypothesized that AFF4 might also be involved in bladder cancer progression. In the current study, AFF4 expression was found to be elevated in bladder cancer samples from both a mouse model and clinical samples. Moreover, AFF4 directly bound to the promoter of MYC to facilitate transcription elongation. Hence, signals originating from METTL3-mediated m6A modification ultimately converge on MYC expression. Collectively, our observations demonstrated that METTL3-mediated m6A modification upregulates MYC expression at multiple levels, including activating NF-κB signaling through IKBKB and RELA to induce MYC transcription, promoting MYC messenger RNA elongation through AFF4, and influencing the m6A modification abundance of MYC messenger RNA. This complex regulatory network coordinated by METTL3-mediated m6A modification efficiently increases MYC protein levels in bladder cancer and likely makes MYC difficult to target by blocking a single signaling pathway. It is well-established that cancer is driven by hundreds or even thousands of dysregulated genes. AFF4/SEC and NF-κB also exert broad effects on the gene expression of various cancer-related genes. Therefore, the oncogenic role of METTL3 might not solely depend on MYC. Other potential target genes of METTL3 involved in bladder cancer progression warrant further investigation. Among these candidates, SOX2 was also identified as a functionally important target of METTL3 in glioblastoma, where METTL3-mediated m6A modification of SOX2 messenger RNA transcripts increases their stability. Intriguingly, a recent study found that AFF4 could also be recruited to the SOX2 promoter and promote SOX2 expression in head and neck squamous cell carcinoma. In our previous study, we found that SOX2 is a marker of stemness in bladder cancer stem cells both in living organisms and in laboratory settings. Taken together, these clues suggest the possibility that METTL3 might be involved in maintaining the stemness of bladder cancer stem cells by inducing m6A modification of SOX2. In summary, our studies demonstrate the critical role of METTL3 in bladder cancer progression, characterized by promoting cancer cell growth, survival, and invasion. Importantly, we uncovered that METTL3 operates a regulatory network involving AFF4, NF-κB, and MYC signaling. Thus, we provide the first insight into METTL3-mediated cancer progression in bladder cancer and propose that targeting METTL3 might be an effective therapeutic strategy for treating this disease.

Methods and materials

Samples, plasmids, and cell lines

We utilized human bladder cancer samples obtained from the Department of Urology, Huadong Hospital, Fudan University, with informed consent from the patients. The pathological condition of the samples was determined by experienced urologists at Huadong Hospital. Patient characteristics for the samples used for m6A level analysis are detailed in the supplementary information. The study received approval from the local ethics committee. The METTL3-expressing lentivirus vector was purchased from Hanbio, Shanghai. Plasmids for the expression of Flag-tagged wild-type METTL3 and a catalytically inactive mutant METTL3 were kindly provided by Dr. Shuibin Lin, Sun Yat-sen University. Mouse bladder cancer samples were prepared in our previous study. The bladder cancer cell lines 5637 and UM-UC-3, and the immortalized uroepithelial cell line SV-HUC-1 were obtained from the Chinese Academy of Cell Resource Center, Shanghai, China, and maintained according to previously established protocols. Cell lines were routinely tested for mycoplasma contamination and were not cultured for more than 20 passages.

LC-MS/MS analysis of m6A level

RNA was extracted from patient samples and 5637 cell lines using TRIzol reagent. The integrity and quantity of each RNA sample were assessed using agarose gel electrophoresis and a Nanodrop instrument. Messenger RNA was isolated and purified from the total RNA using a magnetic isolation module and then hydrolyzed into single nucleosides. These nucleosides were further dephosphorylated using an enzyme mixture. The pretreated nucleoside solution was then deproteinized using a spin filter with a molecular weight cutoff of 10,000 Daltons. The analysis of the nucleoside mixtures was performed on an Agilent 6460 QQQ mass spectrometer coupled with an Agilent 1290 HPLC system. Multi reaction monitoring mode was employed due to its high selectivity and sensitivity for parent-to-product ion transitions. LC-MS/MS data were acquired using Agilent Qualitative Analysis software. The multi reaction monitoring peaks for each modified nucleoside were extracted and normalized to the peak areas of normal adenosine in each sample. Samples were run in duplicate, and the ratios of m6A to adenosine were calculated.

Detect gene expression

For messenger RNA level examination, total RNA from bladder cancer cells was extracted using Trizol reagent. Complementary DNA synthesis was performed using a reverse transcription polymerase chain reaction kit with 1 microgram of RNA per sample. Quantitative polymerase chain reaction reactions were performed using a fluorescent dye-based master mix to determine messenger RNA transcript levels. The primers used for quantitative reverse-transcription polymerase chain reaction are listed in the supplementary information. For Western blotting, 5637 cells were lysed with RIPA buffer following a standard protocol. Cell lysates were then mixed with loading buffer, incubated at 100 degrees Celsius for 5 minutes, and subjected to conventional Western blot analysis. The antibodies used are listed in the supplementary information. For paraffin-embedded sections of bladder cancer tissue samples from mice and human patients, antigen retrieval, blocking, and processing were performed as previously described. Hematoxylin–eosin stains were performed using standard histology procedures. The intensity of immunostaining was measured using image analysis software. The intensity of each image was calculated by normalizing the average integrated optical density to the total selected area of interest.

In vitro cell proliferation, invasion, and apoptosis assay

For the cell proliferation assay, 5 × 103 cells per well were seeded onto a 96-well plate on day 0. Absorbances at 490 nanometers were measured using a cell proliferation assay kit for 4 consecutive days. Triplicate samples were counted. For the cell invasion assay, a specific invasion chamber with a Matrigel matrix was used according to the manufacturer’s instructions. Briefly, 4 × 104 cells transfected with small interfering RNAs were resuspended in 100 microliters of culture medium and seeded in the upper compartment of the invasion chamber. The lower compartment of the chamber contained 500 microliters of medium supplemented with fetal bovine serum and glutamine, serving as a chemoattractant. After 24 hours, non-invasive cells were removed from the upper surface of the membrane using a cotton swab. The invasive cells on the lower surface of the membrane were stained with crystal violet and counted in four separate areas using an inverted microscope. The numbers of apoptotic cells were quantified by flow cytometric assays using a specific apoptosis detection kit following the manufacturer’s instructions.

m6A meRIP-Seq, mRNA sequencing and gene set enrichment analysis

Methylated RNA immunoprecipitation sequencing and data analysis were performed as previously described. Briefly, messenger RNA was purified from total RNA of 5637 cells using a messenger RNA isolation system. 5 micrograms of messenger RNA were then fragmented and immunoprecipitated with an anti-m6A antibody. The immunoprecipitated RNA was washed and eluted by competition with N6-methyladenosine, and analyzed either through quantitative reverse-transcription polymerase chain reaction or by high-throughput sequencing. For high-throughput sequencing, purified RNA fragments from methylated RNA immunoprecipitation were used for library construction using a specific messenger RNA sample preparation kit. Samples were then sequenced using an Illumina HiSeq 2000. Reads mapping, m6A peak calling, motif search, and subsequent analyses were performed as previously described. For RNA-sequencing, RNA from 5637 cells transfected with different small interfering RNAs was extracted using TRIzol reagent. RNA libraries were prepared using a specific RNA-Seq Library Preparation Kit. RNAs were single-end sequenced on Illumina HiSeq 3000 machines. Alignment of reads was performed using Tophat with the Hg38 build of the human genome. Transcript assembly and differential expression were determined using Cufflinks with Refseq messenger RNAs to guide assembly. Analysis of RNA-seq data was performed using a specific package in R. Transcripts with extremely low expression in both cell lines were filtered out, and all remaining genes were preranked by expression fold change and subjected to gene set enrichment analysis to identify enriched functional annotations. Gene set enrichment analyses were performed using specific software, which is a computational method to determine if there is a statistically significant difference in a priori defined sets of genes between two biological states. P values were computed using a bootstrap distribution created by resampling gene sets of the same size. Details of the gene lists used in the analysis are described in the supplementary information.

The pathway luciferase reporter assay

The activities of the NF-κB and MYC signaling pathways were determined using commercially available pathway reporter systems. The analysis was conducted according to the manufacturer’s instructions. Briefly, the cells were transfected in triplicate with each firefly luciferase reporter construct in combination with a Renilla luciferase construct, and both luciferase activities in cell extracts at 24 hours after transfection were measured using a dual-luciferase reporter assay kit and a luminometer. Firefly luciferase activities from each set were normalized to the activity of Renilla luciferase to control for inter-transfection variability. The relative luciferase activities of the pathway reporter over the negative control in the transfected cells were calculated as a measure of pathway activity.

In vivo xenografts model

Specific lentiviral short hairpin RNA constructs targeting human METTL3 and a control short hairpin RNA targeting green fluorescent protein were purchased from a commercial source. Male immunodeficient nude mice, 8 weeks of age, were used for this study, without randomization or blinding of groups. METTL3 stable knockdown cells or control cells were embedded in a specific matrix and subcutaneously injected into the left or right flank of mice, respectively, with 8 mice per group. Tumor volumes were assessed weekly using a specific formula. Mice were humanely sacrificed on day 35, and the tumors were weighed and photographed. Tumor weight was presented as the mean ± standard deviation. Tumor samples were paraffin-embedded and sectioned for further immunohistochemical staining. All animal procedures were performed under a protocol approved by the Laboratory Animal Center of Anhui Medical University and in accordance with the National Institutes of Health guide for the care and use of Laboratory animals.

Statistics

Data are presented as the mean ± standard deviation or standard error. All statistical analyses were performed using specific software. The two-tailed Student’s t-test and one-way analysis of variance were used to calculate statistical significance. A p-value of less than 0.05 was considered statistically significant.