Aberrant N6-methyladenosine (m6A) modification has emerged as a driver of tumor initiation and progression, yet how long noncoding RNAs (lncRNA) are involved in the regulation of m6A remains unknown. Here we utilize data from 12 cancer types from The Cancer Genome Atlas to comprehensively map lncRNAs that are potentially deregulated by DNA methylation. A novel DNA methylation–deregulated and RNA m6A reader–cooperating lncRNA (DMDRMR) facilitated tumor growth and metastasis in clear cell renal cell carcinoma (ccRCC). Mechanistically, DMDRMR bound insulin-like growth factor 2 mRNA-binding protein 3 (IGF2BP3) to stabilize target genes, including the cell-cycle kinase CDK4 and three extracellular matrix components (COL6A1, LAMA5, and FN1), by specifically enhancing IGF2BP3 activity on them in an m6A-dependent manner. Consequently, DMDRMR and IGF2BP3 enhanced the G1–S transition, thus promoting cell proliferation in ccRCC. In patients with ccRCC, high coexpression of DMDRMR and IGF2BP3 was associated with poor outcomes. Our findings reveal that DMDRMR cooperates with IGF2BP3 to regulate target genes in an m6A-dependent manner and may represent a potential diagnostic, prognostic, and therapeutic target in ccRCC.

Significance:

This study demonstrates that the lncRNA DMDRMR acts as a cofactor for IGF2BP3 to stabilize target genes in an m6A-dependent manner, thus exerting essential oncogenic roles in ccRCC.

Long noncoding RNAs (lncRNA) are non–protein coding transcripts longer than 200 nucleotides (nt; ref. 1). The deregulation of lncRNAs by somatic/germline or epigenetic alterations plays critical roles in tumor initiation and progression (2, 3). DNA methylation is one of most prevalent epigenetic modifications in regulating gene expression (4). A number of DNA methylation–deregulated (DMD) protein-coding genes (PCG) have been shown to be molecular drivers in tumor initiation and progression (5). However, the DMD lncRNAs and their consequences remain largely unexplored.

The molecular mechanisms of lncRNA function are multilayered complexity, as lncRNAs serve as guides or scaffolds to regulate DNA–RNA protein interactions (6). N6-methyladenosine (m6A) is the most abundant modification of mRNAs and lncRNAs in eukaryotes and plays critical roles in gene expression and cell fate (7). This modification is regulated by m6A modulators, including “writer,” “eraser,” and “reader” proteins (8). For example, insulin-like growth factor 2 mRNA-binding proteins (IGF2BP), including IGF2BP1/2/3, are a family of m6A reader that recruit cofactor proteins to stabilize m6A-modified transcripts (9). However, the lncRNAs that cooperate with m6A readers to regulate their functions and the underlying mechanisms of their action remain poorly elucidated.

In this study, we provide the comprehensive landscape of DMD lncRNAs in human cancers and reveal that a DMD and RNA m6A reader–cooperating lncRNA (DMDRMR) cooperates with IGF2BP3 to regulate their target transcripts, for example, cyclin-dependent kinase 4 (CDK4), in an m6A-dependent manner, thus exhibiting essential oncogenic roles in clear cell renal cell carcinoma (ccRCC). Collectively, our work illustrates the posttranscriptional regulation of CDK4 and highlights the functional importance of the DMDRMR/IGF2BP3/CDK4 axis in ccRCC cells.

DMD pattern analysis

The differentially methylated promoters (DMP) were analyzed according to previously published article (10). Twelve cancer types, in which cases of adjacent tissues are greater than 10, were fitted into the analysis. Data from The Cancer Genome Atlas (TCGA) HumanMethylation450 (HM450) BeadChip Array was processed and analyzed using RnBeads version 2.2.0 (11) in the R software 3.5.0. In the analysis, the reference human genome 19 was used to annotate all genes and CpGs. The linear models for microarray data package (limma) were used to identify DMPs between tumor and adjacent tissues (12). Promoters with false discovery rate (FDR) < 0.05 and delta-beta methylation difference between tumor and adjacent normal tissues less than or equal to −0.2 were considered hypomethylation, promoters with FDR < 0.05 and delta-beta greater than or equal to 0.2 were considered hypermethylation. Tumor-specific methylated promoters were observed in only one cancer. Tumor-common methylated promoters occurred at least eight cancers with consistent trend.

Cell lines and clinical specimens

786-O, 769-P, ACHN, and Caki-1 cell lines were maintained in RPMI1640 medium (Gibco) supplemented with 10% FBS. HK2 was maintained in DMEM/F-12 medium (Gibco) containing 10% FBS. The human embryonic kidney HEK293T cell was cultured in DMEM (Gibco) supplemented with 10% FBS. These cell lines were purchased from the Shanghai Cell Bank Type Culture Collection Committee, verified by short tandem repeat assays for their identification, and cultured at 37°C in a humidified incubator with 5% CO2. These cell lines were tested negative for Mycoplasma contamination. A cDNA microarray including 15 pairs of ccRCC and adjacent tissues (Supplementary Table S1), and a tissue array including 90 pairs of ccRCC and adjacent tissues (Supplementary Table S2), were purchased from Shanghai Outdo Biotech CO. All clinical information of patients with ccRCC was provided by the company. Forty-eight pairs of frozen ccRCC and adjacent tissues (Supplementary Table S3) were randomly obtained with written informed consent from patients who underwent radical resections in The General Hospital of the People's Liberation Army (Beijing, China). Our study was approved by the Medical Ethics Committees of The General Hospital of the People's Liberation Army (Beijing, China). Clinicopathologic characteristics of patients with ccRCC from TCGA cohort were also analyzed (Supplementary Table S4).

In vivo assays

Stable DMDRMR knockdown (KD) and control cell lines were injected subcutaneously (s.c.; 1 × 107 cells/inoculum) into the flanks of recipient NOD/SCID/IL2Rγ-null (NSG) mice (Shanghai Model Organisms Center). Tumor formation/growth was assessed until the experimental endpoint, and tumor volume was calculated by the formula: (width)2 × length/2. For tail vein injection, stable DMDRMR KD and control cell lines (5 × 106 cells/0.1 mL PBS) were injected into the lateral tail vein of age-matched NSG mice. Mice were euthanized after 4–5 weeks. The number of metastatic foci in the lung was determined using the hematoxylin and eosin (H&E) staining (Beyotime Biotechnology) in tissue sections under a binocular microscope (Leica). All protocols involving animals were previously approved by the Ethics Committee for the Use of Experimental Animals of the Suzhou Institute of Biomedical Engineering and Technology, Chinese Academy of Sciences (Suzhou, Jiangsu, China).

Sequences and antibodies

Primer sequences, probe sequences, sequences of short hairpin RNA (shRNA), and small interfering RNA (siRNA) for KD and antibodies used are listed in Supplementary Tables S5–S11.

Statistical analysis

Data was presented as the mean ± SEM or SD. Two-tailed Student t tests were performed to assess the statistical significance of differences between groups. Kaplan–Meier estimate was performed to analyze the correlation between the RNA or protein levels and the overall survival. A two-sided χ2 test was used to assess the statistical significance of the association between the expression of RNA or protein levels and the clinicopathologic parameters of patients. Spearman correlation was performed to analyze the correlation. “pROC” package in R software was used to construct receiver operating characteristic (ROC) curves and then to calculate area under curve (AUC; ref. 13). P < 0.05 was considered statistically significant. All statistical analyses were performed using GraphPad Prism 5.0 or R software (version 3.5.2).

Additional methods can be found in the Supplementary Material and Methods.

Demethylation-mediated DMDRMR activation

To study functional DMD lncRNAs in tumorigenesis, we first identified the differential DNA methylation levels in tumors and adjacent tissues in 12 cancer types identified in TCGA (Supplementary Fig. S1). Given that DNA methylation in a gene promoter is an essential epigenetic regulator of the gene expression (14), our analysis focused on DMPs (Supplementary Dataset S1). From these DNA methylation pattern data, we identified tumor-common and tumor-specific methylated promoters (Supplementary Fig. S2A and S2B; Supplementary Dataset S2) and found that the DMDRMR (Ensembl: ENSG00000256128) promoter was the most frequently hypomethylated across nearly all cancer types (Supplementary Fig. S2C), in consistent with previous reports (3, 15, 16). We also analyzed the DMDRMR expression data from the MiTranscriptome database for the 12 cancer types with adjacent tissues available and found that DMDRMR was significantly upregulated in the tumor tissues of 8 cancer types compared with its level in the adjacent tissues (Supplementary Fig. S2D; ref. 17). Integrated analyses revealed that promoter methylation and expression levels of DMDRMR were negatively correlated in 7 of 9 cancer types (Supplementary Fig. S3A). Altogether, these data suggest that DMDRMR is a potential functional DNA hypomethylation–driven lncRNA in cancers.

DMDRMR association with clinical outcomes for patients with ccRCC

We next assessed whether DMDRMR expression was correlated with clinical outcomes in the 8 TCGA cancer types, in which it was upregulated (Supplementary Fig. S2D). For ccRCC, higher DMDRMR expression was significantly associated with a low survival rate (Supplementary Fig. S3B) and poor outcomes, including pathologic stage, tumor size, metastatic status, and Fuhrman grade in patients with ccRCC (Supplementary Fig. S3C and S3D; Supplementary Table S4), but not for the other seven cancer types (Supplementary Fig. S3D and S3E). Similarly, a low methylation level of DMDRMR was associated with poor outcome for patients with ccRCC (Supplementary Fig. S4A). It has been reported that the von Hippel Lindau (VHL) tumor suppressor gene is frequently mutated in ccRCC (18); therefore, we also analyzed the relationships between the expression and promoter methylation levels of DMDRMR and the mutation status of VHL. Neither the DMDRMR expression nor the methylation levels were associated with the mutation status of VHL in the patients with ccRCC (Supplementary Fig. S3C and S4A). ROC analysis showed that both DMDRMR expression and promoter methylation levels displayed highly AUCs, clearly distinguishing ccRCC tissues from the adjacent tissues (Supplementary Fig. S4B and S4C). Using the beta values of two probes (cg04396850 and cg00440032) in HM450 BeadChip mapping to the DMDRMR promoter (19), we classified patients with ccRCC into three subgroups by hierarchical clustering analysis. The hypermethylated subgroup exhibited a high DMDRMR methylation level similar to that in adjacent normal tissues (Fig. 1A). This subgroup was characterized by reduced DMDRMR expression and improved overall survival in comparison with the other two groups (Fig. 1A; Supplementary Fig. S4D). In contrast, the hypomethylated subgroup had elevated DMDRMR expression and the worst survival (Fig. 1A; Supplementary Fig. S4D). Other publicly available databases also revealed that DMDRMR was overexpressed (Supplementary Fig. S4E) and hypomethylated (Supplementary Fig. S4F) in ccRCC tissues versus adjacent normal tissues. To further validate the clinical significance of DMDRMR in ccRCC, we examined the DMDRMR expression or methylation level in two independent ccRCC cohorts. Quantitative real-time PCR (qRT-PCR) analysis showed higher expression levels of DMDRMR (Fig. 1B and C; Supplementary Tables S1 and S3), and pyrosequencing analysis showed a lower cg00440032 methylation level in the ccRCC tissues compared with adjacent normal tissues (Fig. 1D; Supplementary Table S3). However, no specific primers were identified for the detection of the cg04396850 methylation level. Moreover, a correlation analysis showed that the cg00440032 methylation level was negatively correlated with the DMDRMR expression level (Fig. 1E), and a similar result was observed in the TCGA ccRCC cohort (Supplementary Fig. S4G), suggesting that high levels of DMDRMR expression might be activated by its promoter DNA demethylation in patients with ccRCC. Overall, these data demonstrate that DMDRMR expression is elevated in ccRCC and is associated with poor outcomes for patients with ccRCC.

Figure 1.

DMDRMR functions as an oncogenic lncRNA in ccRCC. A, Heatmap with beta values for the two DMDRMR HM450 probes in the ccRCC tumors and adjacent tissues. Three subgroups were identified: black, hypermethylation; green, intermediate; and red, hypomethylation. DMDRMR DNA methylation in the adjacent tissues (blue) is shown as a control. Relative DMDRMR expression in these three subgroups compared with the level in adjacent tissues, respectively. B and C, qRT-PCR analysis of DMDRMR expression in 15 paired (B) and 48 paired (C) ccRCC and adjacent tissues. DMDRMR expression was normalized to the reference gene GAPDH. Delta threshold cycle (ΔCt). D, Pyrosequencing analysis of cg00440032 methylation levels from the same cohort as shown in C. E, Correlation analysis between cg00440032 methylation and DMDRMR expression levels in the same cohort of C. F, Representative images of DMDRMR expression and localization in the indicated cells using RNA FISH assays. The nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI). Scale bars, 10 μm. GL, Cell proliferation assay assessing DMDRMR KD (G) and OE (J) cells (n = 3). Representative micrographs and quantification of the DMDRMR KD (H and I) and OE (K and L) cells in Matrigel-coated or noncoated Transwell assays (n = 3). M, Effects of 786-O KD cells on tumor growth in s.c.-implanted NSG mice (n = 6). N and O, Endpoint tumor size (N) and mass (O) in the NSG mice s.c.-implanted with 786-O KD cells. P and Q, Representative images of brightfield (P) and H&E-stained (Q) samples from the tumor foci of the lungs obtained from NSG mice, the lateral tail vein of which were injected by 786-O KD cells. R, Number of metastatic lesions in mice (n = 6), as determined by H&E staining. Data present the mean ± SD or SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 1.

DMDRMR functions as an oncogenic lncRNA in ccRCC. A, Heatmap with beta values for the two DMDRMR HM450 probes in the ccRCC tumors and adjacent tissues. Three subgroups were identified: black, hypermethylation; green, intermediate; and red, hypomethylation. DMDRMR DNA methylation in the adjacent tissues (blue) is shown as a control. Relative DMDRMR expression in these three subgroups compared with the level in adjacent tissues, respectively. B and C, qRT-PCR analysis of DMDRMR expression in 15 paired (B) and 48 paired (C) ccRCC and adjacent tissues. DMDRMR expression was normalized to the reference gene GAPDH. Delta threshold cycle (ΔCt). D, Pyrosequencing analysis of cg00440032 methylation levels from the same cohort as shown in C. E, Correlation analysis between cg00440032 methylation and DMDRMR expression levels in the same cohort of C. F, Representative images of DMDRMR expression and localization in the indicated cells using RNA FISH assays. The nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI). Scale bars, 10 μm. GL, Cell proliferation assay assessing DMDRMR KD (G) and OE (J) cells (n = 3). Representative micrographs and quantification of the DMDRMR KD (H and I) and OE (K and L) cells in Matrigel-coated or noncoated Transwell assays (n = 3). M, Effects of 786-O KD cells on tumor growth in s.c.-implanted NSG mice (n = 6). N and O, Endpoint tumor size (N) and mass (O) in the NSG mice s.c.-implanted with 786-O KD cells. P and Q, Representative images of brightfield (P) and H&E-stained (Q) samples from the tumor foci of the lungs obtained from NSG mice, the lateral tail vein of which were injected by 786-O KD cells. R, Number of metastatic lesions in mice (n = 6), as determined by H&E staining. Data present the mean ± SD or SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Close modal

DMDRMR is an oncogenic lncRNA in ccRCC cells

To evaluate the potentially oncogenic role of DMDRMR in ccRCC cells, we first performed 5′- and 3′- rapid amplification of cDNA ends (RACE) and found that DMDRMR was a 669-nt transcript with 5 exons (Supplementary Fig. S5A and S5B; Supplementary Table S12), the expected size of which was confirmed by northern blot analysis (Supplementary Fig. S5C). Coding Potential Calculator software indicated that DMDRMR is noncoding (Supplementary Fig. S5D). RNA FISH using two independent probes and cellular fractionation assays revealed that DMDRMR was located in both the cytoplasm and nucleus of 769-P cells, which was further verified by shRNA-mediated DMDRMR KD (Fig. 1F; Supplementary Fig. S5E and S5F), using ACTB as a cytoplasmic control (Supplementary Fig. S5E). The MassARRAY assay showed that methylation levels of the DMDRMR promoter in the 786-O and 769-P cell lines were significantly lower than the immortalized human HK2 tubular epithelial cell line (Supplementary Fig. S5G). Furthermore, inhibition of DNA methyltransferase led to a dose-dependent increase in DMDRMR expression in 786-O cells (Supplementary Fig. S5H). Moreover, the copy number and expression levels of DMDRMR were higher in ccRCC cell lines than the HK2 cell line (Supplementary Fig. S5I–S5K). These results indicate that DMDRMR is expressed and regulated by DNA methylation in ccRCC cells.

To further define the molecular mechanisms driving DMDRMR expression in ccRCC cells, we predicted the potential transcription factors of DMDRMR using JASPAR (20), AliBaba2.1 (21), and PROMO software (22). The putative transcription factor c-Jun was identified (Supplementary Fig. S6A). We therefore further confirmed the effects of c-Jun on DMDRMR expression. Overexpression (OE) of c-Jun increased and KD of c-Jun decreased DMDRMR expression in 786-O cells (Supplementary Fig. S6B and S6C). Two putative c-Jun–binding sites (c-Jun-BS1: -640 to -627 and c-Jun-BS2: -571 to -558 bps) on upstream of the DMDRMR transcriptional start site (TSS) were predicted (Supplementary Fig. S6A). Luciferase reporter assay showed that c-Jun OE increased the luciferase activity of the DMDRMR promoter wild-type reporter but not the mutated BS reporter (Supplementary Fig. S6D). Moreover, chromatin immunoprecipitation (ChIP) qRT-PCR analyses showed that c-Jun directly bound to both of the identified BSs in the DMDRMR promoter, and the OE of c-Jun further increased the occupancy (Supplementary Fig. S6E). Consistently, ENCODE ChIP-sequence data also confirmed the presence of c-Jun on the DMDRMR promoter in both K562 and HeLa cell lines (Supplementary Fig. S6F). Taken together, these data suggest that DMDRMR is transcribed by c-Jun.

Given the association of DMDRMR expression with poor outcomes for patients with ccRCC, we explored the potential function of DMDRMR in ccRCC. We used shRNA and CRISPR-Cas9 system technologies to stably KD and knockout (KO) endogenous DMDRMR in 786-O cells expressing high DMDRMR level (Supplementary Fig. S7A–S7C). The KD of DMDRMR by two independent shRNAs led to a significant decrease in the proliferation, migration, and invasion of the 786-O cells (Fig. 1G,I), which was further confirmed by the KO of DMDRMR (Supplementary Fig. S7E and S7G). Moreover, DMDRMR KO resulted in reduced colony formation in a soft agar assay (Supplementary Fig. S7H and S7I). In contrast, DMDRMR OE accelerated the proliferation, migration, and invasion of the 769-P cells (Fig. 1JL; Supplementary Fig. S7D). Furthermore, we investigated the roles of DMDRMR in tumor growth and metastasis in vivo. DMDRMR KD was shown to inhibit in vivo xenograft tumor growth and mass of 786-O and ACHN cells (Fig. 1MO; Supplementary Fig. S8A–S8F). Moreover, H&E staining showed that number of metastatic nodules in the lungs were significantly decreased in the DMDRMR KD group (Fig. 1PR; Supplementary Fig. S8G–S8I). Taken together, these data suggest oncogenic activity of DMDRMR in ccRCC.

DMDRMR and IGF2BP3 cooperate to play oncogenic roles

As lncRNAs bind to specific proteins to exert their functions (6), thus we performed RNA pull-down assay coupled with mass spectrometry to screen DMDRMR-interacting proteins (Fig. 2A; Supplementary Dataset S3). We verified that IGF2BP3 is a putative DMDRMR-binding protein (Fig. 2B). Furthermore, RNA immunoprecipitation (RIP) assay confirmed that IGF2BP3 was associated with DMDRMR but not the U1 control, the IgG as a negative control (Fig. 2C). We also determined the regions of DMDRMR required for this interaction. Truncated mapping revealed that the 51–115 nt region within DMDRMR exon 1 interacts with IGF2BP3 (Fig. 2D and E). IGF2BP3 possesses 2 RNA-recognition-motif (RRM) domains and 4 K homology (KH) domains (Fig. 2F; ref. 23). RIP assays for FLAG-tagged full-length and truncated IGF2BP3 showed that the KH1 and KH2 domains were essential for recruiting DMDRMR (Fig. 2G and H). In addition, DMDRMR FISH followed by immunofluorescence of IGF2BP3 demonstrated the colocalization of DMDRMR and IGF2BP3 in the cytoplasm of 786-O cells, further supporting their interaction (Fig. 2I). Overall, these data demonstrate that DMDRMR physically interacts with IGF2BP3.

Figure 2.

DMDRMR directly binds with IGF2BP3. A, Visualization of silver-stained protein bands by biotin-labeled DMDRMR sense and antisense RNA probes incubated with total protein extracts from 769-P cells. Red arrow, approximate 72-kilodalton (KDa) DMDRMR sense–specific band. B, Immunoblotting to determine the specific association of IGF2BP3 with biotinylated DMDRMR. C, qRT-PCR analysis of DMDRMR enriched by IGF2BP3 in 769-P cells (top). Immunoblot of IGF2BP3 is shown (bottom; n = 6). IP, immunoprecipitation. D, Secondary structure of DMDRMR analyzed by RNAfold web server and deletion mapping of biotinylated DMDRMR RNA motifs, as indicated. The red boxes represent the remaining fragments of DMDRMR, with the corresponding number label in the corner. E, Immunoblot showing the association of IGF2BP3 with biotinylated DMDRMR sense and antisense RNA strands and the above-mentioned biotinylated DMDRMR RNA motifs. F, Schematic structures showing six domains in IGF2BP3. G and H, Immunoblot of the FLAG (G) and RIP (H) analysis for DMDRMR enrichment in HEK293T cells transfected with the FLAG-tagged full-length or truncated IGF2BP3 constructs (n = 3). IB, immunoblot; aa, amino acid. I, Confocal images showing the colocalization of DMDRMR and IGF2BP3 in the IGF2BP3 KD and DMDRMR KD cells. Scale bars, 10 μm. J, qRT-PCR analysis of IGF2BP3 mRNA levels (top) and immunoblot of IGF2BP3 (bottom) in the DMDRMR KD cells (n = 3). K, qRT-PCR analysis of DMDRMR RNA levels (top) and immunoblot of IGF2BP3 (bottom) in the IGF2BP3 KD cells (n = 3). The results are presented as the mean ± SEM. ***, P < 0.001; ns, not significant.

Figure 2.

DMDRMR directly binds with IGF2BP3. A, Visualization of silver-stained protein bands by biotin-labeled DMDRMR sense and antisense RNA probes incubated with total protein extracts from 769-P cells. Red arrow, approximate 72-kilodalton (KDa) DMDRMR sense–specific band. B, Immunoblotting to determine the specific association of IGF2BP3 with biotinylated DMDRMR. C, qRT-PCR analysis of DMDRMR enriched by IGF2BP3 in 769-P cells (top). Immunoblot of IGF2BP3 is shown (bottom; n = 6). IP, immunoprecipitation. D, Secondary structure of DMDRMR analyzed by RNAfold web server and deletion mapping of biotinylated DMDRMR RNA motifs, as indicated. The red boxes represent the remaining fragments of DMDRMR, with the corresponding number label in the corner. E, Immunoblot showing the association of IGF2BP3 with biotinylated DMDRMR sense and antisense RNA strands and the above-mentioned biotinylated DMDRMR RNA motifs. F, Schematic structures showing six domains in IGF2BP3. G and H, Immunoblot of the FLAG (G) and RIP (H) analysis for DMDRMR enrichment in HEK293T cells transfected with the FLAG-tagged full-length or truncated IGF2BP3 constructs (n = 3). IB, immunoblot; aa, amino acid. I, Confocal images showing the colocalization of DMDRMR and IGF2BP3 in the IGF2BP3 KD and DMDRMR KD cells. Scale bars, 10 μm. J, qRT-PCR analysis of IGF2BP3 mRNA levels (top) and immunoblot of IGF2BP3 (bottom) in the DMDRMR KD cells (n = 3). K, qRT-PCR analysis of DMDRMR RNA levels (top) and immunoblot of IGF2BP3 (bottom) in the IGF2BP3 KD cells (n = 3). The results are presented as the mean ± SEM. ***, P < 0.001; ns, not significant.

Close modal

To further characterize the molecular consequences of the DMDRMR and IGF2BP3 interaction, we performed qRT-PCR and immunoblot assays and confirmed that DMDRMR and IGF2BP3 failed to alter each other's expression level (Fig. 2J and K; Supplementary Fig. S9A and S9B). Moreover, DMDRMR and IGF2BP3 KD did not affect each other's cellular localization and fluorescence intensity (Fig. 2I). We next explored the role of the DMDRMR/IGF2BP3 axis in ccRCC and performed rescue experiments. Strikingly, IGF2BP3 KD abolished the induction of cell proliferation, migration, and invasion elicited by the DMDRMR OE (Fig. 3AC; Supplementary Fig. S9C). Conversely, DMDRMR KD abrogated the promoting effects of IGF2BP3 OE on cell proliferation, migration, and invasion (Fig. 3DF; Supplementary Fig. S9D). These results suggest that DMDRMR and IGF2BP3 may cooperate to play oncogenic roles in ccRCC.

Figure 3.

DMDRMR and IGF2BP3 cooperative to play oncogenic roles. A, Cell proliferation assay assessing IGF2BP3 KD 769-P cells with ectopically expressed DMDRMR on the fifth day (top). Immunoblot of IGF2BP3 from the above-mentioned cells (n = 3). B and C, Representative micrographs (B) and quantification (C) of the above-mentioned cells in the Matrigel-coated or noncoated Transwell assays (n = 4). D, Cell proliferation assay assessing DMDRMR KD 786-O cells with ectopically expressed IGF2BP3 on the fifth day. Immunoblot of IGF2BP3 from the above-mentioned cells (n = 4). E and F, Representative micrographs (E) and quantification (F) of the above-mentioned cells in the Matrigel-coated or noncoated Transwell assays (n = 6). Results are presented as the mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant.

Figure 3.

DMDRMR and IGF2BP3 cooperative to play oncogenic roles. A, Cell proliferation assay assessing IGF2BP3 KD 769-P cells with ectopically expressed DMDRMR on the fifth day (top). Immunoblot of IGF2BP3 from the above-mentioned cells (n = 3). B and C, Representative micrographs (B) and quantification (C) of the above-mentioned cells in the Matrigel-coated or noncoated Transwell assays (n = 4). D, Cell proliferation assay assessing DMDRMR KD 786-O cells with ectopically expressed IGF2BP3 on the fifth day. Immunoblot of IGF2BP3 from the above-mentioned cells (n = 4). E and F, Representative micrographs (E) and quantification (F) of the above-mentioned cells in the Matrigel-coated or noncoated Transwell assays (n = 6). Results are presented as the mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant.

Close modal

DMDRMR cooperates with IGF2BP3 to stabilize their targets

IGF2BP3, which belongs to a conserved family of RNA-binding proteins (RBP), has been reported to stabilize a large repertoire of target mRNA transcripts, including MYC, which is critical for oncogenic function (9). Thus, we hypothesized that DMDRMR modulates IGF2BP3-dependent gene regulation. To confirm this hypothesis, we first performed an integrated analysis on the RNA-sequencing data of DMDRMR KD and two published IGF2BP3 RIP- and cross-linking immunoprecipitation (CLIP)-sequencing data (9, 24) to define the target genes potentially regulated by DMDRMR and IGF2BP3. Gene expression profiling after DMDRMR KD showed the downregulation of 1,039 genes and upregulation of 749 genes (Supplementary Fig. S9E; Supplementary Dataset S4), which were indeed enriched in cell growth- and cell metastasis–related pathways such as “cell cycle” and “focal adhesion” (Fig. 4A), according to a Kyoto Encyclopedia of Genes and Genomes (KEGG) database analysis (25). Among 1,039 downregulated genes after DMDRMR KD, 145 genes had IGF2BP3′s enrichment on their transcripts through two independent IGF2BP3 RIP- and CLIP-sequencing data (Fig. 4B; Supplementary Dataset S5). By further performing KEGG pathway analysis of the 145 genes, we identified three pathways (ecm–receptor interaction, focal adhesion, pathways in cancer) that contained 10 genes as possible key targets of the DMDRMR/IGF2BP3 axis (Fig. 4C). Strikingly, the transcript levels of the cell-cycle kinase CDK4 and three extracellular matrix (ECM) components, collagen type VI alpha 1 chain (COL6A1), laminin alpha 5 (LAMA5), and fibronectin 1 (FN1), were decreased in KD of both DMDRMR and IGF2BP3 (Fig. 4D and E), suggesting that CDK4, COL6A1, LAMA5, and FN1 are cotarget genes of DMDRMR and IGF2BP3. Similar to the transcript levels, the protein levels of CDK4 were reduced in the KD and KO of DMDRMR and the KD of IGF2BP3 cells (Supplementary Fig. S9F). Conversely, OE of DMDRMR and IGF2BP3 upregulated CDK4 expression (Supplementary Fig. S9G). Consistent with a previous study (9), IGF2BP3 KD reduced the transcript level of MYC, while DMDRMR KD failed to alter MYC expression (Fig. 4D and E). These results suggest that DMDRMR specifically regulates a subset of IGF2BP3 targets, consistent with recent study that even direct modulation of RBPs only affects transcript abundance of some targets (26).

Figure 4.

DMDRMR cooperates with IGF2BP3 to stabilize CDK4. A, The top 20 enriched KEGG terms for differentially expressed genes. The color intensities represent the Q values. The circle sizes represent the number of differentially expressed genes. B, Venn diagram showing the 145 overlapping genes by downregulated genes of DMDRMR KD RNA-sequence data and IGF2BP3 binding genes of IGF2BP3 RIP- and CLIP-sequencing data. C, Chord graph representing the association of 10 overlapping genes, with the indicated KEGG terms. D and E, qRT-PCR analysis of the 10 overlapping genes and MYC transcript levels in the DMDRMR KD (D) and IGF2BP3 KD (E) 786-O cells (n = 3). F and G, The half-life of CDK4 after treatment with 5 μmol/L actinomycin D for the indicated times in the IGF2BP3 KD 769-P cells with ectopically expressed DMDRMR (F) and in the DMDRMR KD 786-O cells with ectopically expressed IGF2BP3 (n = 5; G). Results are presented as the mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant.

Figure 4.

DMDRMR cooperates with IGF2BP3 to stabilize CDK4. A, The top 20 enriched KEGG terms for differentially expressed genes. The color intensities represent the Q values. The circle sizes represent the number of differentially expressed genes. B, Venn diagram showing the 145 overlapping genes by downregulated genes of DMDRMR KD RNA-sequence data and IGF2BP3 binding genes of IGF2BP3 RIP- and CLIP-sequencing data. C, Chord graph representing the association of 10 overlapping genes, with the indicated KEGG terms. D and E, qRT-PCR analysis of the 10 overlapping genes and MYC transcript levels in the DMDRMR KD (D) and IGF2BP3 KD (E) 786-O cells (n = 3). F and G, The half-life of CDK4 after treatment with 5 μmol/L actinomycin D for the indicated times in the IGF2BP3 KD 769-P cells with ectopically expressed DMDRMR (F) and in the DMDRMR KD 786-O cells with ectopically expressed IGF2BP3 (n = 5; G). Results are presented as the mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant.

Close modal

To further investigate whether DMDRMR mediates the stabilization of IGF2BP3 targets, we chose CDK4 for further study. IGF2BP3 KD abolished the half-life and mRNA level of CDK4 increased by the OE of DMDRMR (Fig. 4F; Supplementary Fig. S9H), and DMDRMR KD abrogated the stable effect and mRNA level of IGF2BP3 OE on the CDK4 transcript (Fig. 4G; Supplementary Fig. S9I), suggesting that the stabilizing effects of DMDRMR and IGF2BP3 on the CDK4 transcript are likely interdependent. Collectively, these data indicate that DMDRMR cooperates with IGF2BP3 to regulate CDK4 by enhancing mRNA stability.

DMDRMR mediates IGF2BP3-dependent gene regulation in an m6A-dependent manner

IGF2BP3 has been characterized as an m6A reader that regulates m6A-modified genes (9), which led us to hypothesize that DMDRMR mediates IGF2BP3-dependent gene regulation in an m6A-dependent manner. To test this hypothesis, we first identified the potential m6A-modified regions of CDK4, COL6A1, LAMA5, and FN1 based on the IGF2BP3 RIP- and m6A RIP-sequencing data (9), which were enriched by IGF2BP3 and m6A methylation that markedly reduced upon methyltransferase-like 14 (METTL14) KD (Fig. 5A; Supplementary Fig. S10A–S10C). According to the “GGAC” m6A core motif across the m6A peaks, we further confirmed whether these targets are modified by m6A methylation. The gene-specific m6A RIP qRT-PCR assays showed that in methyltransferase-like 3 (METTL3) and METTL14 KD cells (Supplementary Fig. S11A), the m6A methylation of the 5′ untranslated region (UTR), but not the exon 1 and 3′ UTR of the CDK4 transcript, were substantially decreased (Fig. 5B). A similar result was found for the exon 15 and 3′ UTR in COL6A1, the exon 20 in FN1 and the exon 4 in LAMA5 (Supplementary Fig. S11B). As expected, a similar reduction in the coding region instability determinant (CRD) of MYC was also observed, but it was not found in the negative controls of HPRT1 or DMDRMR (Fig. 5B; Supplementary Fig. S11B and S11C), confirming that the targets of DMDRMR and IGF2BP3 are modified by m6A. Considering that IGF2BP3 preferentially binds the “GGAC” m6A core motif of its target mRNAs (9), we next explored whether DMDRMR affects IGF2BP3 binding to these m6A-modified regions. The IGF2BP3 RIP qRT-PCR assays showed that DMDRMR KD significantly reduced IGF2BP3 binding to the m6A-modified regions of CDK4, COL6A1, LAMA5, and FN1 but not to the MYC CRD and CDK4 exon 1 and 3′ UTR (Fig. 5C; Supplementary Fig. S11D), supporting that DMDRMR specifically enhances the occupancy of IGF2BP3 at the m6A-modified regions of their targets. In addition, in vitro–transcribed DMDRMR pull-down assay revealed that, except for the COL6A1 3′ UTR, MYC CRD and CDK4 exon 1 and 3′ UTR, the binding abilities of DMDRMR to the m6A-modified regions of CDK4, COL6A1, LAMA5, and FN1 were nearly abolished by IGF2BP3 KD (Fig. 5D; Supplementary Fig. S11E), suggesting that DMDRMR may bind the m6A-modified regions of their targets in an IGF2BP3-dependent manner. To further define the binding abilities, we then examined whether IGF2BP3 was required for the engagement of CDK4 5′ UTR by DMDRMR. Protease K treatment remarkably weakened the interaction of CDK4 5′ UTR and DMDRMR (Supplementary Fig. S11F). Moreover, truncated DMDRMR pull-down assay showed that 51–115 nt of DMDRMR bound to CDK4 5′ UTR (Supplementary Fig. S11G), suggesting that the binding of DMDRMR to the m6A-modified regions of their targets are dependent on IGF2BP3. These data demonstrate that DMDRMR contributes to IGF2BP3 reading of the m6A-modified regions in their targets.

Figure 5.

DMDRMR cooperates with IGF2BP3 to regulate CDK4 in an m6A-dependent manner. A, Distribution of m6A peaks across the CDK4 transcript based on m6A RIP- and IGF2BP3 RIP-sequencing data. The potential m6A-modified and IGF2BP3-binding regions in CDK4 are marked by rectangles. B, RIP qRT-PCR showing the enrichment of m6A modification in the CDK4 5′ UTR/exon 1/3′ UTR regions in the METTL3 and METTL14 KD 769-P cells (n = 3). MYC CRD was used as a positive control. HPRT1 was used as a negative control. C and D, RIP qRT-PCR detecting the enrichment of IGF2BP3 (C) and biotin-labeled DMDRMR (D) in the CDK4 5′ UTR and MYC CRD in DMDRMR KD (C) and IGF2BP3 KD (D) 769-P cells (n = 3). E, Schematic representation of wild-type (WT) and mutated (MUT; GGAC to AAGT) CDK4 5′ UTR of the pmirGLO vector. F and H, RIP qRT-PCR detection of the enrichment of IGF2BP3 (F) and m6A (H) in the CDK4 5′ UTR WT and MUT luciferase reporters in the DMDRMR and IGF2BP3 OE cells (n = 3). G and I, RIP qRT-PCR detection of the enrichment of IGF2BP3 (G) and m6A (I) in the CDK4 5′ UTR WT and MUT luciferase reporters in the DMDRMR and IGF2BP3 KD cells (n = 3). J, Relative luciferase activity levels of CDK4 5′ UTR WT and MUT reporters in the DMDRMR and IGF2BP3 OE cells (n = 3). K, Relative luciferase activity levels of CDK4 5′ UTR WT and MUT reporters in the DMDRMR and IGF2BP3 KD cells (n = 3). L, Relative luciferase activity levels of CDK4 5′ UTR WT in the METTL3 KD, METTL14 KD, or IGF2BP3 KD cells with ectopically expressed DMDRMR (n = 3). Results are presented as the mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant.

Figure 5.

DMDRMR cooperates with IGF2BP3 to regulate CDK4 in an m6A-dependent manner. A, Distribution of m6A peaks across the CDK4 transcript based on m6A RIP- and IGF2BP3 RIP-sequencing data. The potential m6A-modified and IGF2BP3-binding regions in CDK4 are marked by rectangles. B, RIP qRT-PCR showing the enrichment of m6A modification in the CDK4 5′ UTR/exon 1/3′ UTR regions in the METTL3 and METTL14 KD 769-P cells (n = 3). MYC CRD was used as a positive control. HPRT1 was used as a negative control. C and D, RIP qRT-PCR detecting the enrichment of IGF2BP3 (C) and biotin-labeled DMDRMR (D) in the CDK4 5′ UTR and MYC CRD in DMDRMR KD (C) and IGF2BP3 KD (D) 769-P cells (n = 3). E, Schematic representation of wild-type (WT) and mutated (MUT; GGAC to AAGT) CDK4 5′ UTR of the pmirGLO vector. F and H, RIP qRT-PCR detection of the enrichment of IGF2BP3 (F) and m6A (H) in the CDK4 5′ UTR WT and MUT luciferase reporters in the DMDRMR and IGF2BP3 OE cells (n = 3). G and I, RIP qRT-PCR detection of the enrichment of IGF2BP3 (G) and m6A (I) in the CDK4 5′ UTR WT and MUT luciferase reporters in the DMDRMR and IGF2BP3 KD cells (n = 3). J, Relative luciferase activity levels of CDK4 5′ UTR WT and MUT reporters in the DMDRMR and IGF2BP3 OE cells (n = 3). K, Relative luciferase activity levels of CDK4 5′ UTR WT and MUT reporters in the DMDRMR and IGF2BP3 KD cells (n = 3). L, Relative luciferase activity levels of CDK4 5′ UTR WT in the METTL3 KD, METTL14 KD, or IGF2BP3 KD cells with ectopically expressed DMDRMR (n = 3). Results are presented as the mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant.

Close modal

To determine the effect of DMDRMR on the binding of IGF2BP3 to CDK4 m6A-modified sites, we firstly constructed luciferase reporters for the 5′ UTR in CDK4 that contain the wild-type or mutated “GGAC” m6A motif (Fig. 5E). IGF2BP3 and m6A RIP qRT-PCR assays showed the higher enrichment of both IGF2BP3 and m6A with the wild-type reporter, compared with the mutated reporter (Fig. 5F,I). Moreover, OE of DMDRMR or IGF2BP3 remarkably increased, and KD of DMDRMR or IGF2BP3 remarkably reduced the binding of IGF2BP3 and m6A in the wild-type reporter (Fig. 5F,I). In the mutated reporter, we also confirmed the endogenous IGF2BP3 and m6A binding (Fig. 5F,I). Furthermore, luciferase reporter assays showed that DMDRMR OE increased and DMDRMR KD reduced the luciferase activity of the 5′ UTR wild-type reporter but not the mutant reporter, the similar results were observed for IGF2BP3 (Fig. 5J and K). The elevated luciferase activity was also observed in the 51–115 nt region within DMDRMR, supporting that the 51–115 nt region of DMDRMR is functionally important by interacting with IGF2BP3 and CDK4 (Supplementary Fig. S11H). Moreover, the KD of METTL3, METTL14, or IGF2BP3 significantly restored the effects of DMDRMR OE on the 5′ UTR wild-type reporter (Fig. 5L), indicating that DMDRMR cooperates with IGF2BP3 to regulate the CDK4 transcript by enhancing the reading of the m6A-modified site in the CDK4 5′ UTR by IGF2BP3. Altogether, these data reveal that DMDRMR cooperates with IGF2BP3 to regulate their targets in an m6A-dependent manner.

DMDRMR and IGF2BP3 coordinate cell proliferation through the CDK4-mediated cell cycle

It has been previously shown that CDK4 is required for the G1–S transition in the cell cycle (27). Given the above findings, we thus analyzed the effect of the DMDRMR/IGF2BP3 axis on the cell cycle and performed flow cytometry assays. The DMDRMR KD and KO inhibited the G1–S transition (Fig. 6A and B), which was also observed in the KD of IGF2BP3 cells (Fig. 6C), indicating that the DMDRMR/IGF2BP3 axis promoted cell proliferation by accelerating the G1–S transition. Consequently, we further examined whether DMDRMR and IGF2BP3 mediates CDK4/6 inhibitor resistance in ccRCC. Cell proliferation assay showed that both DMDRMR and IGF2BP3 KD cells significantly reduced palbociclib resistance compared with that of the control cells (Supplementary Fig. S12A). In contrast, DMDRMR and IGF2BP3 OE cells increased the resistance to palbociclib (Supplementary Fig. S12B). To further determine whether CDK4 is a key downstream effector of the DMDRMR/IGF2BP3 axis promoting the proliferation of ccRCC cells, we also performed rescue experiments. As expected, the inhibitory effects of DMDRMR KD and IGF2BP3 KD on cell proliferation could be greatly reversed by CDK4 OE (Fig. 6D and E). Together, these data suggest that DMDRMR and IGF2BP3 accelerate ccRCC cell proliferation by stabilizing CDK4 and mediates the resistance of CDK4/6 inhibitor.

Figure 6.

DMDRMR and IGF2BP3 promote tumor cell progression via CDK4. Flow cytometry analysis assays showing the percentage of DMDRMR KD (A) and DMDRMR KO (B) and IGF2BP3 KD (C) 786-O cells in the sub-G0–G1, S, and G2–M cell-cycle phases (n = 3). Representative cell-cycle profiles (left) and quantification (right) analysis are shown. PI-A, propidium iodide-A. Cell proliferation assay assessing 786-O DMDRMR KD (D) and IGF2BP3 KD (E) cells with ectopically expressed CDK4 on the fifth day (top; n = 3). Immunoblots of CDK4 from the above-mentioned cells. Results are presented as the mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 6.

DMDRMR and IGF2BP3 promote tumor cell progression via CDK4. Flow cytometry analysis assays showing the percentage of DMDRMR KD (A) and DMDRMR KO (B) and IGF2BP3 KD (C) 786-O cells in the sub-G0–G1, S, and G2–M cell-cycle phases (n = 3). Representative cell-cycle profiles (left) and quantification (right) analysis are shown. PI-A, propidium iodide-A. Cell proliferation assay assessing 786-O DMDRMR KD (D) and IGF2BP3 KD (E) cells with ectopically expressed CDK4 on the fifth day (top; n = 3). Immunoblots of CDK4 from the above-mentioned cells. Results are presented as the mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Close modal

DMDRMR and IGF2BP3 coordinate cell invasion and metastasis partially through FN1

FN1, an abundant and ubiquitous ECM component, has been shown to be responsible for cancer cell movement, including cell metastasis and adhesion (28, 29). Thus, we examined whether FN1 is a downstream effector of DMDRMR/IGF2BP3 axis to mediate the invasion and metastasis of ccRCC cells. The transwell assay showed that, FN1 KD partially reversed the increased invasion and metastasis of ccRCC cell by the OE of DMDRMR and IGF2BP3 (Supplementary Fig. S12C–S12F), suggesting that DMDRMR and IGF2BP3 promote invasion and metastasis of ccRCC partially through FN1.

DMDRMR is clinically associated with IGF2BP3 in ccRCC

To further investigate the clinical relevance of these findings, a tissue microarray containing 90 pairs of ccRCC and matched adjacent tissues (Supplementary Table S2) was evaluated to characterize the relationship of DMDRMR and IGF2BP3 expression. In situ hybridization (ISH) results revealed that, in both the cytoplasm and nucleus, DMDRMR expression significantly increased in the ccRCC tissues versus the adjacent tissues and in patients at a pathologic high stage compared with those at a pathologic low stage (Fig. 7A and B; Supplementary Fig. S13A). High expression levels of DMDRMR were associated with poor overall survival of patients with ccRCC (Fig. 7E; Supplementary Fig. S13B). Similar results were observed by IGF2BP3 IHC staining (Fig. 7C, D, and F). Further correlation analysis results showed that the cytoplasmic expression of DMDRMR was positively correlated with the levels of IGF2BP3 (Fig. 7G). In addition, the combination of highly expressed DMDRMR and IGF2BP3 predicted the poorest overall survival of patients (Fig. 7H). These findings support the oncogenic role of DMDRMR in ccRCC.

Figure 7.

The correlation between cytoplasmic DMDRMR and IGF2BP3 in the ccRCC tissues. AD, Representative stained tissue microarray cores of DMDRMR (A) and IGF2BP3 (C). Quantification of cytoplasmic DMDRMR ISH staining (B) and IGF2BP3 IHC staining (D) in the ccRCC tissues versus the matched adjacent tissues (left) and in patients with pathologic stage greater than II (n = 25) versus patients with pathologic stage I+II (n = 65; right). Scale bars, 50 μm. Kaplan–Meier curves represent the percentages of surviving patients with ccRCC with high- and low-DMDRMR (E)/IGF2BP3 (F) scores. The score one for the cytoplasmic DMDRMR/IGF2BP3 expression was used as the cut-off value. G, Correlation analysis of cytoplasmic DMDRMR and IGF2BP3 expression. H, Kaplan–Meier analysis indicating the correlation of the combination of cytoplasmic DMDRMR and IGF2BP3 expression with percentages of surviving patients with ccRCC. I, Proposed model for the DMDRMR/IGF2BP3 axis promoting the pathogenesis of ccRCC. DMDRMR, which is activated by DNA demethylation and transcribed by c-Jun, cooperates with IGF2BP3 to regulate their m6A-modified target transcripts, including CDK4, FN1, COL6A1, and LAMA5, thus promoting cell proliferation, invasion, and migration in ccRCC. **, P < 0.01; ***, P < 0.001.

Figure 7.

The correlation between cytoplasmic DMDRMR and IGF2BP3 in the ccRCC tissues. AD, Representative stained tissue microarray cores of DMDRMR (A) and IGF2BP3 (C). Quantification of cytoplasmic DMDRMR ISH staining (B) and IGF2BP3 IHC staining (D) in the ccRCC tissues versus the matched adjacent tissues (left) and in patients with pathologic stage greater than II (n = 25) versus patients with pathologic stage I+II (n = 65; right). Scale bars, 50 μm. Kaplan–Meier curves represent the percentages of surviving patients with ccRCC with high- and low-DMDRMR (E)/IGF2BP3 (F) scores. The score one for the cytoplasmic DMDRMR/IGF2BP3 expression was used as the cut-off value. G, Correlation analysis of cytoplasmic DMDRMR and IGF2BP3 expression. H, Kaplan–Meier analysis indicating the correlation of the combination of cytoplasmic DMDRMR and IGF2BP3 expression with percentages of surviving patients with ccRCC. I, Proposed model for the DMDRMR/IGF2BP3 axis promoting the pathogenesis of ccRCC. DMDRMR, which is activated by DNA demethylation and transcribed by c-Jun, cooperates with IGF2BP3 to regulate their m6A-modified target transcripts, including CDK4, FN1, COL6A1, and LAMA5, thus promoting cell proliferation, invasion, and migration in ccRCC. **, P < 0.01; ***, P < 0.001.

Close modal

More importantly, TCGA ccRCC expression data revealed positive correlations of DMDRMR with IGF2BP3, CDK4, FN1, LAMA5, and COL6A1 and of IGF2BP3 with CDK4, FN1, and COL6A1 (Supplementary Fig. S13C). The high expression levels of IGF2BP3, CDK4, COL6A1, and FN1 in the patients with ccRCC were correlated with poor outcomes (Supplementary Fig. S14A–S14D). The high expression level of LAMA5 in ccRCC cells failed to show clinical relevance (Supplementary Fig. S14E). These data suggest that the DMDRMR/IGF2BP3 axis is critical to ccRCC tumor pathogenesis and the prognosis of patients with ccRCC.

DNA methylation is one of the most common epigenetic modifications that regulate gene expression (4, 30). The deregulated expression of lncRNAs has been linked to the initiation, progression, and drug responses of cancer (31). Thus, the identification of new DMD lncRNAs with driver functions is important for the comprehensive understanding of tumorigenesis. In this study, we identified numerous DMD lncRNAs, including well-characterized lncRNAs, such as EPIC1 (3), LINC00968 (32), and WT1-AS (33). Our study reveals that elevated DMDRMR expression is regulated by its hypomethylated promotor in patients with ccRCC.

Genomic studies have revealed critical gene mutations that drive the pathogenesis of ccRCC, such as frequent VHL inactivation and mTOR mutation, which support the current use of first-line therapies for advanced ccRCC, which primarily target VEGF and mTOR (18). However, the fact that the efficacious targeted treatments were quite limited in clinic impel to identify novel ccRCC drivers. In recent years, the expansion of knowledge of lncRNA biology has revealed their critical roles in ccRCC tumorigenic processes (34, 35). m6A modification has been linked to diverse effects in human cancers through the regulation of oncogene expression (36). Nevertheless, few lncRNAs and m6A-regulations have well-characterized functions in ccRCC. Here, we extended new insights into ccRCC tumorigenesis by showing that the lncRNA DMDRMR interacts with the IGF2BP3 protein and cooperatively stabilizes the cell-cycle kinase CDK4 and three ECM components in an m6A-dependent manner, potentially promoting ccRCC tumorigenesis.

It had been reported that IGF2BP3 represents an independent predictor for aggressive ccRCC (37). The mechanism of IGF2BP3 overexpression in the ccRCC and other tumors have not been elucidated. We showed that the expression of IGF2BP3 is not regulated by c-Jun and DNA methylation (Supplementary Fig. S15A–S15C). The mechanism of IGF2BP3 deregulation in ccRCC need to be further investigated, which might contribute to target IGF2BP3. IGF2BP3 has been shown to promote cell invasion and migration in ccRCC (38); however, the molecular mechanism for the acquisition of these phenotypes remains enigmatic. Here, we discovered that the ability of IGF2BP3 to drive ccRCC is dependent on DMDRMR, which regulates some common targets, including CDK4, COL6A1, FN1, and LAMA5. All of these targeted genes are involved in proliferative or metastatic tumor progression (28, 39–41). Furthermore, DMDRMR did not alter IGF2BP3 binding to or the expression of MYC, a well-known target of IGF2BP3 (9). In contrast, DMDRMR but not IGF2BP3 regulates VEGFA, which is a critical angiogenic cytokine and may be a key therapeutic target for renal cell carcinoma (Fig. 4D and E; ref. 42), thus implying that DMDRMR and IGF2BP3 also selectively regulate the targets, except regulating cotargets. The specificity of DMDRMR function on IGF2BP3 and underlying mechanism need to be further studied. Although the importance of lncRNAs in mediating the stabilization of IGF2BP3 targets in cancers has been established by numerous studies (43–45), it is not yet clear how lncRNAs cooperate with IGF2BP3 in regulating these targets. Our finding not only support that DMDRMR can bind IGF2BP3 and enhance its function, but also propose a precise mechanism by which the action of DMDRMR and IGF2BP3 on their targets is likely dose-cooperative and m6A-dependent. IGF2BP3 also regulates expression of target genes through RNA-induced silencing complex (46); however, whether DMDRMR is involved in this process through m6A remains elusive.

CDK4 is a therapeutic target in various cancers, including breast cancer and melanoma, resulting from amplification and upregulation of the CDK4 gene (47). Our study sheds light on posttranscriptional m6A-mediated CDK4 expression. Numerous m6A mapping studies have displayed a pronounced accumulation of m6A in the 3′ UTR and near stop codons, only a few of which were in the 5′ UTR (48). Our study confirmed that rather than the 3′ UTR, the CDK4 5′ UTR is modified by m6A and interacts with IGF2BP3 and DMDRMR, resulting in the stabilization of CDK4. Together, 5′ UTR m6A is functionally important, but whether it widely stabilizes transcripts remains unknown and further studies are needed. Clearly, our study used the published IGF2BP3 CLIP- and RIP-sequencing data, which are not from ccRCC cell lines. We need further explore the IGF2BP3-binding targets based on ccRCC cells to improve the identification of genes regulated by DMDRMR and IGF2BP3 in future.

Overall, our study reveals that DMDRMR is a protumorigenic lncRNA that mediates the stabilization of IGF2BP3 targets in an m6A-dependent manner in ccRCC (Fig. 7I).

No disclosures were reported.

S. Gao: Designed research;wrote the paper. Y. Gu: Performed research;designed research;wrote the paper;analyzed data. S. Niu: Provided samples. Y. Wang: Performed research. L. Duan: Performed research. Y. Pan: Validation. Z. Tong: Performed research. X. Zhang: Provided samples. Z. Yang: Performed research. B. Peng: Performed research. X. Wang: Performed research. X. Han: Performed research. Y. Li: Performed research. T. Cheng: Performed research. Y. Liu: Performed research. L. Shang: Performed research. T. Liu: Performed research. X. Yang: Performed research. M. Sun: Validation. S. Jiang: Performed research. C. Zhang: Performed research. N. Zhang: Performed research. Q. Ye: Performed research.

This work was supported by grants from the National Natural Science Foundation of China (82025029, 81773023, and 81802526); Strategic Pilot Science and Technology Project (XDB29040103) and Frontier Research Program (QYZDB-SSW-SMC038) of Chinese Academy of Sciences; National Key R&D Program of China (2016YFC1302103); Technological Innovation Project of Shanxi Transformation and Comprehensive Reform Demonstration Zone (2017KJCX01); Scientific Research and Equipment Development Project of Chinese Academy of Sciences (YJKYYQ20180032).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Mattick
JS
,
Rinn
JL
. 
Discovery and annotation of long noncoding RNAs
.
Nat Struct Mol Biol
2015
;
22
:
5
7
.
2.
Li
Z
,
Zhang
J
,
Liu
X
,
Li
S
,
Wang
Q
,
Di
C
, et al
The LINC01138 drives malignancies via activating arginine methyltransferase 5 in hepatocellular carcinoma
.
Nat Commun
2018
;
9
:
1572
.
3.
Wang
Z
,
Yang
B
,
Zhang
M
,
Guo
W
,
Wu
Z
,
Wang
Y
, et al
lncRNA Epigenetic landscape analysis identifies EPIC1 as an oncogenic lncRNA that interacts with MYC and promotes cell-cycle progression in cancer
.
Cancer Cell
2018
;
33
:
706
20
.
4.
Shen
H
,
Laird
PW
. 
Interplay between the cancer genome and epigenome
.
Cell
2013
;
153
:
38
55
.
5.
Nephew
KP
,
Huang
TH
. 
Epigenetic gene silencing in cancer initiation and progression
.
Cancer Lett
2003
;
190
:
125
33
.
6.
Yao
RW
,
Wang
Y
,
Chen
LL
. 
Cellular functions of long noncoding RNAs
.
Nat Cell Biol
2019
;
21
:
542
51
.
7.
Zhao
BS
,
Roundtree
IA
,
He
C
. 
Post-transcriptional gene regulation by mRNA modifications
.
Nat Rev Mol Cell Biol
2017
;
18
:
31
42
.
8.
Roundtree
IA
,
Evans
ME
,
Pan
T
,
He
C
. 
Dynamic RNA modifications in gene expression regulation
.
Cell
2017
;
169
:
1187
200
.
9.
Huang
H
,
Weng
H
,
Sun
W
,
Qin
X
,
Shi
H
,
Wu
H
, et al
Recognition of RNA N6-methyladenosine by IGF2BP proteins enhances mRNA stability and translation
.
Nat Cell Biol
2018
;
20
:
285
95
.
10.
Gu
Y
,
Zou
YM
,
Lei
D
,
Huang
Y
,
Li
W
,
Mo
Z
, et al
Promoter DNA methylation analysis reveals a novel diagnostic CpG-based biomarker and RAB25 hypermethylation in clear cell renel cell carcinoma
.
Sci Rep
2017
;
7
:
14200
.
11.
Assenov
Y
,
Müller
F
,
Lutsik
P
,
Walter
J
,
Lengauer
T
,
Bock
C
. 
Comprehensive analysis of DNA methylation data with RnBeads
.
Nat Methods
2014
;
11
:
1138
40
.
12.
Smyth
GK
. 
Linear models and empirical bayes methods for assessing differential expression in microarray experiments
.
Stat Appl Genet Mol Biol
2004
;
3
:
Article3
.
13.
Robin
X
,
Turck
N
,
Hainard
A
,
Tiberti
N
,
Lisacek
F
,
Sanchez
JC
, et al
pROC: an open-source package for R and S+ to analyze and compare ROC curves
.
BMC Bioinformatics
2011
;
12
:
77
.
14.
Singal
R
,
Wang
SZ
,
Sargent
T
,
Zhu
SZ
,
Ginder
GD
. 
Methylation of promoter proximal-transcribed sequences of an embryonic globin gene inhibits transcription in primary erythroid cells and promotes formation of a cell type-specific methyl cytosine binding complex
.
J Biol Chem
2002
;
277
:
1897
905
.
15.
Yang
X
,
Gao
L
,
Zhang
S
. 
Comparative pan-cancer DNA methylation analysis reveals cancer common and specific patterns
.
Brief Bioinform
2017
;
18
:
761
73
.
16.
Ding
W
,
Chen
G
,
Shi
T
. 
Integrative analysis identifies potential DNA methylation biomarkers for pan-cancer diagnosis and prognosis
.
Epigenetics
2019
;
14
:
67
80
.
17.
Iyer
MK
,
Niknafs
YS
,
Malik
R
,
Singhal
U
,
Sahu
A
,
Hosono
Y
, et al
The landscape of long noncoding RNAs in the human transcriptome
.
Nat Genet
2015
;
47
:
199
208
.
18.
Hsieh
JJ
,
Purdue
MP
,
Signoretti
S
,
Swanton
C
,
Albiges
L
,
Schmidinger
M
, et al
Renal cell carcinoma
.
Nat Rev Dis Primers
2017
;
3
:
17009
.
19.
Bibikova
M
,
Barnes
B
,
Tsan
C
,
Ho
V
,
Klotzle
B
,
Le
JM
, et al
High density DNA methylation array with single CpG site resolution
.
Genomics
2011
;
98
:
288
95
.
20.
Fornes
O
,
Castro-Mondragon
JA
,
Khan
A
,
van der Lee
R
,
Zhang
X
,
Richmond
PA
, et al
JASPAR 2020: update of the open-access database of transcription factor binding profiles
.
Nucleic Acids Res
2020
;
48
:
D87
-
D92
.
21.
Grabe
N
. 
AliBaba2: context specific identification of transcription factor binding sites
.
In Silico Biol
2002
;
2
:
S1
15
.
22.
Messeguer
X
,
Escudero
R
,
Farré
D
,
Núñez
O
,
Martínez
J
,
Albà
MM
. 
PROMO: detection of known transcription regulatory elements using species-tailored searches
.
Bioinformatics
2002
;
18
:
333
4
.
23.
Bell
JL
,
Wächter
K
,
Mühleck
B
,
Pazaitis
N
,
Köhn
M
,
Lederer
M
, et al
Insulin-like growth factor 2 mRNA-binding proteins (IGF2BPs): post-transcriptional drivers of cancer progression?
Cell Mol Life Sci
2013
;
70
:
2657
75
.
24.
ENCODE Project Consortium
. 
An integrated encyclopedia of DNA elements in the human genome
.
Nature
2012
;
489
:
57
74
.
25.
Kanehisa
M
,
Goto
S
. 
KEGG: kyoto encyclopedia of genes and genomes
.
Nucleic Acids Res
2000
;
28
:
27
30
.
26.
Van Nostrand
EL
,
Freese
P
,
Pratt
GA
,
Wang
X
,
Wei
X
,
Xiao
R
, et al
A large-scale binding and functional map of human RNA-binding proteins
.
Nature
2020
;
583
:
711
9
.
27.
Goel
S
,
DeCristo
MJ
,
McAllister
SS
,
Zhao
JJ
. 
CDK4/6 Inhibition in cancer: beyond cell cycle arrest
.
Trends Cell Biol
2018
;
28
:
911
25
.
28.
Cai
X
,
Liu
C
,
Zhang
TN
,
Zhu
YW
,
Dong
X
,
Xue
P
. 
Down-regulation of FN1 inhibits colorectal carcinogenesis by suppressing proliferation, migration, and invasion
.
J Cell Biochem
2018
;
119
:
4717
28
.
29.
Yousif
NG
. 
Fibronectin promotes migration and invasion of ovarian cancer cells through up-regulation of FAK-PI3K/Akt pathway
.
Cell Biol Int
2014
;
38
:
85
91
.
30.
Jones
PA
,
Baylin
SB
. 
The fundamental role of epigenetic events in cancer
.
Nat Rev Genet
2002
;
3
:
415
28
.
31.
Schmitt
AM
,
Chang
HY
. 
Long Noncoding RNAs in Cancer Pathways
.
Cancer Cell
2016
;
29
:
452
63
.
32.
Wang
Y
,
Zhou
J
,
Xu
YJ
,
Hu
HB
. 
Long non-coding RNA LINC00968 acts as oncogene in NSCLC by activating the Wnt signaling pathway
.
J Cell Physiol
2018
;
233
:
3397
406
.
33.
Lv
L
,
Chen
G
,
Zhou
J
,
Li
J
,
Gong
J
. 
WT1-AS promotes cell apoptosis in hepatocellular carcinoma through down-regulating of WT1
.
J Exp Clin Cancer Res
2015
;
34
:
119
.
34.
Zhai
W
,
Sun
Y
,
Jiang
M
,
Wang
M
,
Gasiewicz
TA
,
Zheng
J
, et al
Differential regulation of LncRNA-SARCC suppresses VHL-mutant RCC cell proliferation yet promotes VHL-normal RCC cell proliferation via modulating androgen receptor/HIF-2α/C-MYC axis under hypoxia
.
Oncogene
2017
;
36
:
4525
.
35.
Qu
L
,
Ding
J
,
Chen
C
,
Wu
ZJ
,
Liu
B
,
Gao
Y
, et al
Exosome-transmitted lncARSR promotes sunitinib resistance in renal cancer by acting as a competing endogenous RNA
.
Cancer Cell
2016
;
29
:
653
68
.
36.
Lan
Q
,
Liu
PY
,
Haase
J
,
Bell
JL
,
Hüttelmaier
S
,
Liu
T
. 
The critical role of RNA m6A methylation in cancer
.
Cancer Res
2019
;
79
:
1285
92
.
37.
Jiang
Z
,
Chu
PG
,
Woda
BA
,
Rock
KL
,
Liu
Q
,
Hsieh
CC
, et al
Analysis of RNA-binding protein IMP3 to predict metastasis and prognosis of renal-cell carcinoma: a retrospective study
.
Lancet Oncol
2006
;
7
:
556
64
.
38.
Pei
X
,
Li
M
,
Zhan
J
,
Yu
Y
,
Wei
X
,
Guan
L
, et al
Enhanced IMP3 expression activates NF-кB pathway and promotes renal cell carcinoma progression
.
PLoS One
2015
;
10
:
e0124338
.
39.
Baker
SJ
,
Reddy
EP
. 
CDK4: A key player in the cell cycle, development, and cancer
.
Genes Cancer
2012
;
3
:
658
69
.
40.
Chiu
KH
,
Chang
YH
,
Wu
YS
,
Lee
SH
,
Liao
PC
. 
Quantitative secretome analysis reveals that COL6A1 is a metastasis-associated protein using stacking gel-aided purification combined with iTRAQ labeling
.
J Proteome Res
2011
;
10
:
1110
25
.
41.
Gordon-Weeks
A
,
Lim
SY
,
Yuzhalin
A
,
Lucotti
S
,
Vermeer
JAF
,
Jones
K
, et al
Tumour-derived laminin α5 (LAMA5) promotes colorectal liver metastasis growth, branching angiogenesis and Notch pathway inhibition
.
Cancers
2019
;
11
:
630
.
42.
Dvorak
HF
. 
Vascular permeability factor/vascular endothelial growth factor: a critical cytokine in tumor angiogenesis and a potential target for diagnosis and therapy
.
J Clin Oncol
2002
;
20
:
4368
80
.
43.
Ma
F
,
Liu
X
,
Zhou
S
,
Li
W
,
Liu
C
,
Chadwick
M
, et al
Long non-coding RNA FGF13-AS1 inhibits glycolysis and stemness properties of breast cancer cells through FGF13-AS1/IGF2BPs/Myc feedback loop
.
Cancer Lett
2019
;
450
:
63
75
.
44.
Jiang
W
,
Cheng
X
,
Wang
T
,
Song
X
,
Zheng
Y
,
Wang
L
. 
LINC00467 promotes cell proliferation and metastasis by binding with IGF2BP3 to enhance the mRNA stability of TRAF5 in hepatocellular carcinoma
.
J Gene Med
2020
;
22
:
e3134
.
45.
Bao
G
,
Huang
J
,
Pan
W
,
Li
X
,
Zhou
T
. 
Long noncoding RNA CERS6-AS1 functions as a malignancy promoter in breast cancer by binding to IGF2BP3 to enhance the stability of CERS6 mRNA
.
Cancer Med
2020
;
9
:
278
89
.
46.
Ennajdaoui
H
,
Howard
JM
,
Sterne-Weiler
T
,
Jahanbani
F
,
Coyne
DJ
,
Uren
PJ
, et al
IGF2BP3 modulates the interaction of invasion-associated transcripts with RISC
.
Cell Rep
2016
;
15
:
1876
83
.
47.
Sherr
CJ
,
Beach
D
,
Shapiro
GI
. 
Targeting CDK4 and CDK
6
: from discovery to therapy
.
Cancer Discov
2016
;
6
:
353
67
.
48.
Meyer
KD
,
Saletore
Y
,
Zumbo
P
,
Elemento
O
,
Mason
CE
,
Jaffrey
SR
. 
Comprehensive analysis of mRNA methylation reveals enrichment in 3′ UTRs and near stop codons
.
Cell
2012
;
149
:
1635
46
.