N6-methyladenosine modification of circNSUN2 facilitates cytoplasmic export and stabilizes HMGA2 to promote colorectal liver metastasis

Abstract

Circular RNAs (circRNAs) have been implicated in cancer progression through largely unknown mechanisms. Herein, we identify an N⁶-methyladenosine (m⁶A) modified circRNA, circNSUN2, frequently upregulated in tumor tissues and serum samples from colorectal carcinoma (CRC) patients with liver metastasis (LM) and predicts poorer patient survival. The upregulated expression of circNSUN2 promotes LM in PDX metastasis models in vivo and accelerates cancer cells invasion in vitro. Importantly, N⁶-methyladenosine modification of circNSUN2 increases export to the cytoplasm. By forming a circNSUN2/IGF2BP2/HMGA2 RNA-protein ternary complex in the cytoplasm, circNSUN2 enhances the stability of HMGA2 mRNA to promote CRC metastasis progression. Clinically, the upregulated expressions of circNSUN2 and HMGA2 are more prevalent in LM tissues than in primary CRC tissues. These findings elucidate that N⁶-methyladenosine modification of circNSUN2 modulates cytoplasmic export and stabilizes HMGA2 to promote CRC LM, and suggest that circNSUN2 could represent a critical prognostic marker and/or therapeutic target for the disease.

Fig. 1

CircNSUN2 was upregulated in CRCs with liver metastasis (LM). a Top, flowchart illustrating the screening criteria of potential regulatory circRNAs enriched in CRCs. Bottom, clustered heatmap showing the dysregulated expression of circRNAs occurrent in both two CRC samples (the |average normalized fold change| ≥ 1.3) within susceptibility loci of CRC analyzed by CircRNAs Microarray. b qRT-PCR analysis of circNSUN2 expression in 97 CRC tissues and matched adjacent normal tissue. Data represent mean ± S.D., the P value was determined by a two-tailed paired Student’s t test. c Kaplan−Meier analysis of OS in CRC patients with low versus high expression the circNSUN2 from SYSUCC cohorts. The P value was determined by a Log-rank test. d qRT-PCR analysis of circNSUN2 expression from 18 normal colorectal tissues, 22 colorectal adenomas, 97 CRC patients without liver metastasis and 25 CRC patients with liver metastasis. Data represent mean ± S.D., the P values were determined by an unpaired Student’s t test. e qRT-PCR analysis of circNSUN2 expression in 20 pairs of primary colorectal cancer tissues (PC) and matched liver metastasis tissues (LM) surgically obtained from the same patients. Data represent mean ± S.D., the P value was determined by a two-tailed paired Student’s t test. f qRT-PCR analysis of circNSUN2 expression in serum from 18 normal control, 20 CRC patients without LM and 20 CRC patients with LM from the SYSUCC. Data represent mean ± S.D., the P values were determined by an unpaired Student’s t test

Fig. 2

Characterization of circNSUN2 in CRC. a The genomic locus of circNSUN2. Left, the expression of circNSUN2 was detected by qRT-PCR followed by Sanger sequencing. Arrows represent divergent primers binding to the genome region of circNSUN2. Right, qRT-PCR products with divergent primers showing circularization of circNSUN2. cDNA complementary DNA. gDNA genomic DNA. b qRT–PCR analysis for the expression of circNSUN2 and NSUN2 mRNA after treatment with RNase R in HCT116 cells. Data represent mean ± S.D. from five independent experiments; dot plot reflects data points from independent experiment. The P value was determined by a two-tailed unpaired Student’s t test. c Northern blotting of circNSUN2 and NSUN2 transcripts by hybridization with exon5 (top, left) and exon 5−exon 4 junction (top, right) probes in the presence or absence of RNase R treatment. Right, RNA marker. GAPDH mRNA (bottom) was also blotted as an internal control. The samples for blotting GAPDH were aliquoted before RNase R treatment and loaded separately in two wells for validation. d qRT–PCR analysis for the expression of circNSUN2 and NSUN2 mRNAs after treatment with Actinomycin D at the indicated time points in HCT116 cells. Data represent mean ± S.D. from five independent experiments; dot plot reflects data points from independent experiment. The P value was determined by a two-way ANOVA. e Cytoplasmic and Nuclear mRNA Fractionation experiment showing that circNSUN2 localized in the nucleus and the cytoplasm. β-actin and U3 were applied as positive controls in the cytoplasm and nucleus, respectively. Data represent mean ± S.D. from five independent experiments; dot plot reflects data points from independent experiment. f RNA fluorescence in situ hybridization for circNSUN2. Nuclei were stained with DAPI. Scale bar, 10 µm. For c and f, junction probe is complementary to the junction sequence of circNSUN2. Source data are provided as a Source Data file

Fig. 3

CircNSUN2 promotes metastasis of CRC. a, b Decreased (a) or increased (b) tumor metastasis formed in the livers of mice through the inferior hemispleen implantation of circNSUN2-konckdown (a) or circNSUN2-overexpression (b) TC71 PDX cells. Top, left, representative bioluminescent images of livers for each experimental group at 8 weeks. Top, right, statistical analysis of bioluminescent tracking plots. Bottom, left, representative liver. Bottom, right, the number of metastatic nodules formed in the lungs of mice for each group (n = 6 mice / group). Data represent mean ± S.D.; dot plot reflects data points from independent experiment. The P values were determined by a two-tailed unpaired Student’s t test. c, d Decreased (c) or increased (d) tumor metastasis formed in the lungs of mice through vein tail injection of circNSUN2-konckdown (c) or circNSUN2-overexpression (d) PDX cells. Top, left, representative bioluminescent images of lungs for each experimental group during 6 weeks. Top, right, statistical analysis of bioluminescent tracking plots. Bottom, left, representative lung and representative HE staining of lung metastatic lesions, Original magnification, ×4, scale bar, 100 µm. Bottom, right, the number of metastatic nodules formed in the lungs of mice for each group (n = 6 mice/group). Data represent mean ± S.D.; dot plot reflects data points from independent experiment. The P values were determined by a two-tailed unpaired Student’s t test. Source data are provided as a Source Data file

Fig. 4

YTHDC1 promotes cytoplasmic export of m⁶A methylated circNSUN2. a Left, identification of the circNSUN2-protein complex pulled down by circNSUN2 junction probe with protein extracts from HCT116 cells. The arrows indicating the additional band presented in circNSUN2-protein complex. Right, immunoblot analysis of YTHDC1 after pulldown assay showing its specific association with circNSUN2. b RIP assays showing the association of YTHDC1 with circNSUN2. Top, IP efficiency of YTHDC1-antibody shown in western blotting. Bottom, relative enrichment representing RNA levels associated with YTHDC1 relative to an input control. IgG antibody served as a control. Data represent mean ± S.D. from three independent experiments; dot plot reflects data points from independent experiment. c MeRIP assay showing that circNSUN2 was highly recruited in m⁶A precipitated fraction. Data represent mean ± S.D. from three independent experiments; dot plot reflects data points from independent experiment. The P values were determined by a two-tailed paired Student’s t test. d Top, schematic illustration showing the GAACU m⁶A motif located at exon 5−exon 4 junction site of circNSUN2. Bottom, the sequence of RNA probe for RNA-EMSA assay. e RNA-EMSA assay showing the binding ability of purified YTHDC1 with biotin-labeled oligonucleotides containing GAACU motif from circNSUN2. f Cytoplasmic and Nuclear mRNA Fractionation experiment showing that knockdown of YTHDC1 increased the nuclear circNSUN2 content, whereas the dysregulation of nuclear circNSUN2 caused by YTHDC1 RNAi was recovered by overexpression of WT but not the mutant of YTHDC1. Data represent mean ± S.D. from five independent experiments; dot plot reflects data points from independent experiment. The P values were determined by a two-tailed unpaired Student’s t test. g RNA-FISH showing that the increased nuclear staining of circNSUN2 caused by YTHDC1 RNAi was rescued by overexpression of WT, but not the mutant YTHDC1. Scale bar, 10 µm. Data represent mean ± S.D. from three independent experiments; dot plot reflects data points from independent experiment. The P values were determined by a two-tailed unpaired Student’s t test. Source data are provided as a Source Data file. Unprocessed original scans of blots are shown in Supplementary Fig. 10

Fig. 5

CircNSUN2 interacts with IGF2BP2 through CAUCAU motif. a Immunoblot analysis of IGF2BP2 after RNA-pulldown assay showing its specific association with circNSUN2. b RIP assays showing the association of IGF2BP2 with circNSUN2. Relative enrichment representing RNA levels associated with IGF2BP2 compared to an input control. IgG antibody served as a control. Data represent mean ± S.D. from three independent experiments; dot plot reflects data points from independent experiment. c IF-FISH assay showing that circNSUN2 is colocalized with IGF2BP2 protein in the cytoplasm. Scale bar, 10 µm. d Top, schematic structures showing RNA-binding domains within IGF2BP2 protein and a summary of IGF2BP2 truncations. Middle, relative enrichment representing circNSUN2 levels associated with truncated IGF2BP2 relative to an input control. Bottom, immunoblot analysis with anti-FLAG of HCT116 cells transfected with plasmids encoding FLAG-tagged WT or truncated IGF2BP2s. Data represent mean ± S.D. from three independent experiments; dot plot reflects data points from independent experiment. e Top, schematic illustration showing the CAUCAU motif located at exon 5−exon 4 junction site of circNSUN2 and the RNA probe for RNA-EMSA assay. Bottom, RNA-EMSA assay showing the binding ability of purified IGF2BP2 with biotin-labeled oligonucleotides containing CAUCAU motif from circNSUN2. Source data are provided as a Source Data file. Unprocessed original scans of blots are shown in Supplementary Fig. 10

Fig. 6

CircNSUN2/IGF2BP2/HMGA2 ternary complex stabilized HMGA2 mRNA. a Top, sequence BLAST analysis showing that circNSUN2 directly targets the 3′UTR of HMGA2 with high AU content. Bottom, relative enrichment representing HMGA2 (left panel) and circNSUN2 (right panel) RNA levels associated with circNSUN2 junction compared to control. Data represent mean ± S.D. from three independent experiments; dot plot reflects data points from independent experiment. The P values were determined by a two-tailed unpaired Student’s t test. b CircNSUN2 knockdown in HCT116 cells significantly downregulated HMGA2 mRNA abundance. Data represent mean ± S.D. from three independent experiments; dot plot reflects data points from independent experiment. The P values were determined by a two-way ANOVA. c Left, luciferase mRNA expression of luciferase reporter gene with HMGA2-WT or HMGA2-Mut in control and circNSUN2-knockdown or circNSUN2-overexpression HCT116 cells. Right, relative luciferase activity of luciferase reporter gene with HMGA2-WT or HMGA2-Mut in control and circNSUN2-knockdown or circNSUN2-overexpression HCT116 cells. Data represent mean ± S.D. from three independent experiments; dot plot reflects data points from independent experiment. The P values were determined by a two-tailed unpaired Student’s t test. d IF-FISH assay showing that the colocalization of circNSUN2/IGF2BP2/HMGA2 was decreased upon knockdown of circNSUN2. Scale bar, 10 µm. e Relative enrichment representing the enrichment of HMGA2 associated with truncated IGF2BP2 protein complex compared to an input control. IgG antibody served as a control. Data represent mean ± S.D. from three independent experiments; dot plot reflects data points from independent experiment. f RIP assays showing the association of IGF2BP2 with HMGA2 upon circNSUN2 silencing or overexpression. Data represent mean ± S.D. from three independent experiments; dot plot reflects data points from independent experiment. The P values were determined by a two-tailed unpaired Student’s t test. Source data are provided as a Source Data file

Fig. 7

CircNSUN2 promotes LM metastasis of CRC through HMGA2 pathway. a, b Decreased tumor metastasis formed in the livers of mice through the inferior hemispleen implantation of circNSUN2-konckdown TC71 PDX cells was rescued by overexpression of HMGA2. a Left, representative bioluminescent images of livers for each experimental group at 8 weeks. Right, statistical analysis of bioluminescent tracking plots. The P values were determined by a two-tailed unpaired Student’s t test. b Left, representative liver. Right, the number of metastatic nodules formed in the livers of mice for each group (n = 6 mice/group). The P values were determined by a two-tailed unpaired Student’s t test. c−e Transwell assay (c), inverted invasion assay (d) and 3D multicellular tumor spheroids invasion assay (e) showing that decreased cell invasion in circNSUN2 knockdown HCT116 cells was rescued by overexpression of HMGA2. Scale bar, 100 µm (c), 100 µm (d) and 10 µm (e). Data represent mean ± S.D. from five independent experiments; dot plot reflects data points from independent experiment. The P values were determined by a two-tailed unpaired Student’s t test. f Left, circNSUN2 expression showing positively correlated with HMGA2 expression in CRC patients derived from SYSUCC. Right, circNSUN2 expression showing positively correlated with CXCR4 expression in CRC patients derived from SYSUCC. The P values were determined by Pearson correlation analysis. g qRT–PCR analysis for the RNA expression of HMGA2 and CXCR4 in 20 pairs of primary colorectal cancer (PC) and matched liver metastasis (LM) surgically obtained from the same patients. Data represent mean ± S.D., the P values were determined by a two-tailed paired Student’s t test. h A proposed model for the regulatory landscape of the circNSUN2/IGF2BP2/HMGA2 signaling axis in promoting the metastasis of CRC. Source data are provided as a Source Data file